Handbook on the Physics and

Chemistry of Rare Earths, volume 20

Elsevier, 1995
Edited by: Karl A. Gschneidner, Jr. and LeRoy Eyring
ISBN: 978-0444820143

Handbook on the Physics and Chemistry of Rare Earths

VoL 20
edited by K.A. Gschneidner, Jr. and L. Eyring
1995 Elsevier Science B.V. All rights reserved

PREFACE
Karl A. G S C H N E I D N E R , Jr., and L e R o y E Y R I N G

These elements perplex us in our rearches [sic], baffle us in our speculations, and haunt
us in our very dreams. They stretch like an unknown sea before us - mocking, mystifying,
and murmuring strange revelations and possibilities.

Sir William Crookes (February 16, 1887)

This volume, which completes the second decade of volumes in this series, is focused
on physical aspects of metallic compounds. Research efforts on metallic rare earth
compounds started in earnest about 50 years ago and received a significant boost with
the discovery of the RCo5 permanent magnets about 12 3(ears later. In this time much
has been learned about the structure as well as the electrical, magnetic and thermal
properties of ~2500 binary rare earth metallic compounds. But considering the possible
true ternary compounds, and possible pseudo-binary ternary alloys formed by mixing two
binary compounds, we have just begun to scratch the surface of the wealth of knowledge
these yet-to-be discovered materials will bring to mankind. The information contained in
this volume will be but a few footsteps in our journey to unravel the mysteries hidden in
these unknown materials. Hopefully, the four chapters will serve as useful guides in our
quest for this new knowledge.
The first chapter (135) in this volume, by Onuki and Hasegawa, deals with the Fermi
surfaces of rare earth (Y, La, Ce, Pr, Nd, Sm, Gd and Yb) intermetallic compounds.
Initially, the reader is introduced to the relevant theories required to describe the electronic
behavior of the electrons near the Fermi surface, and then to the basic experimental
techniques to study these surfaces. The main portion of the chapter is devoted to
a comparison of the experimental results with the band structure calculations for a
large number of compounds. The authors have found some systematics in the observed
behaviors, and have grouped the compounds into several different classes: i.e. non-4f
behaviors, including some cerium compounds; valence fluctuation compounds; Kondo
regime materials; and magnetic materials with magnetic energy gaps etc. They also note
that the more complicated the crystal structure the larger the discrepancy between theory
and experiment.
Next, Gasgnier examines the world of thin films of rare earth metals, alloys, and
compounds in chapter 136. The three main topics are the pure metals themselves, metallic

vi

PREFACE

alloys and compounds, and metalloid compounds. Many of the metallic elements exhibit
a "valence" change when they change from the vapor to solid or vice-versa and this has
some strong influences on the physical properties, especially the vapor pressure. Another
major problem is the easy way which the rare earth metal films can become contaminated
especially by the strongly electronegative non-metallic elements, i.e. H, N, O. The reported
results are critically examined by Gasgnier. Included in his coverage are the results
on metallic alloy/compound thin films include permanent magnets, R-Ni and R-Co
hydrogen storage alloys, modulated and multilayered structures, and superconducting
materials. The metalloid films include those with the chalcogenides, bismuth, lead and
their combinations.
The third chapter (137) in this volume, by Vajda, is devoted to hydrogen in metals
and their binary compounds RH2 and R H 3 . One of the critical problems is the purity
of the starting rare earth metal itself because phase relations can be greatly affected by
impurities. Therefore, Vajda devotes some time discussing the preparation of specimens
and the phase diagrams. The interesting structural properties, kinetics and thermodynamic
behavior, as well as electronic, magnetic and thermal properties are reviewed. The
occurrence of H-H pairs in zig-zag chains along the c-axis in the heavy lanthanides
and Sc and Y terminal solid solution alloys is one of the unusual structural behaviors
observed in these materials. The formation or annihilation of these pairs lead to some
interesting kinetic effects and phase transitions during heating and cooling. Also examined
is the profound influence of hydrogen on the magnetic properties by the mediation of
RKKY interaction in these materials.
The final chapter (138) of this volume, by Gignoux and Schmitt, is an update on
the magnetic behaviors of lanthanide intermetatlic compounds. This chapter builds on
the review of Kirchmayr and Poldy in chapter 14 of volume 2 of this Handbook
series. When one examines chapter 138, it will immediately be apparent that a great
deal of science and technology has occurred in the past 15 years in the magnetic
behaviors of lanthanide compounds. Gignoux and Schmitt divide their chapter into
two main parts: one is devoted to 3d magnetism where both the 3d metal and the
lanthanide element contribute to the magnetic behavior; and the second is concerned with
lanthanide magnetism by itself. In the first group the major emphasis is on systems which
exhibit collective electron metamagnetism. In addition, magnetocrystalline anistropy, and
topological frustration and magnetic instability are reviewed. The section devoted to
compounds with the lanthanide as a magnetic atom, deals with materials which exhibit
metamagnetic processes from different origins. It is found that the majority of these
compounds order antiferromagnetically with complex magnetic field vs. temperature
phase diagrams. This is due to the long range and oscillatory nature of the RKKY
exchange interaction.

CONTENTS

Preface
Contents

vii

Contents of Volumes 1-19

ix

135. Y. 0nuki and A. Hasegawa

Fermi surfaces of intermetallic compounds

136. M. Gasgnier
The intricate world of rare earth thin films." metals, alloys, intermetallics,
chemical compounds . . . .
105
137. P. Vajda
Hydrogen in rare-earth metals, including RH2+x phases
138. D. Gignoux and D. Schmitt
Magnetic properties of intermetallic compounds

Author Index

425

Subject Index

457

vii

293

207

CONTENTS OF VOLUMES 1-19

V O L U M E 1: Metals
1978, 1st repr. 1982, 2nd repr. 1991; ISBN 0-444-85020-1
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.

Z.B. Goldschmidt, Atomic properties (free atom) 1

B.J. Beaudry and K.A. Gschneidner Jr, Preparation and basic properties of the rare earth metals 173
S.H. Liu, Electronic structure of rare earth metals 233
D.C. Koskenmaki and K.A. Gschneidner Jr, Cerium 337
L.J. Sundstr6m, Low temperature heat capacity of the rare earth metals 379
K.A. McEwen, Magnetic and transport properties of the rare earths 411
S.K. Sinha, Magnetic structures and inelastic neutron scattering." metals, alloys and compounds 489
T.E. Scott, Elastic and mechanical properties 591
A. Jayaraman, High pressure studies: metals, alloys and compounds 707
C. Probst and J. Wittig, Superconductivity: metals, alloys and compounds 749
M.B. Maple, L.E. DeLong and B.C. Sales, Kondo effect." alloys and compounds 797
M.E Dariel, Diffusion in rare earth metals 847
Subject index 877

V O L U M E 2: Alloys and intermetallics

1979, 1st repr. 1982, 2nd repr. 1991; ISBN 0-444-85021-X
13.
14.
15.
16.
17.
18.
19.
20.

A. Iandelli and A. Palenzona, Crystal chemistry ofintermetallic compounds l

H.R. Kirchmayr and C.A. Poldy, Magnetic properties of intermetallic compounds of rare earth
metals 55
A.E. Clark, Magnetostrictive RFe2 intermetallic compounds 231
J.J. Rhyne, Amorphous magnetic rare earth alloys 259
E Fulde, Crystalfields 295
R.G. Barnes, NMR, EPR and M6ssbauer effect: metals, alloys and compounds 387
E Wachter, Europium chalcogenides: EuO, EuS, EuSe and EuTe 507
A. Jayaraman, Valence changes in compounds 575
Subject Index 613

V O L U M E 3: Non-metallic compounds - I
1979, 1st repr. 1984; ISBN 0-444-85215-8
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.

L.A. Haskin and T.P. Paster, Geochemistry and mineralogy of the rare earths 1
J.E. Powell, Separation chemistry 81
C.K. Jorgensen, Theoretical chemistry of rare earths 111
W.T. Carnall, The absorption and fluorescence spectra of rare earth ions in solution
L.C. Thompson, Complexes 209
G.G. Libowitz and A.J. Maeland, Hydrides 299
L. Eyring, The binary rare earth oxides 337
D.J.M. Bevan and E. Summerville, Mixed rare earth oxides 401
C.P. Khattak and EEY. Wang, Perovskites and garnets 525
L.H. Brixne~ J.R. Barkley and W. Jeitschko, Rare earth molybdates (VI) 609
Subject index 655
ix

171

CONTENTS OF VOLUMES 1-19

V O L U M E 4: N o n - m e t a l l i c c o m p o u n d s - I I
1979, 1st repr. 1984; ISBN 0-444-85216-6

31.
32.
33.
34.
35.
36.
37A.
37B.
37C.
37D.
37E.
37E
37G.
38.
39.
40.

J. Flahaut, Sulfides, selenides and tellurides 1

J.M. H~chke, Halides 89
E Hulliger, Rare earth pnictides 153
G. Blasse, Chemistry and physics of R-activated phosphors 237
M.J. Weber, Rare earth lasers 275
EK. Fong, Nonradiative processes of rare-earth ions in crystals 317
J.W. O'Laughlin, Chemical spectrophotometric and polarographic methods 341
S.R. Taylor, Trace element analysis of rare earth elements by spark source mass spectroscopy 359
R.J. Conzemius, Analysis of rare earth matrices by spark source mass spectrometry 377
E.L. DeKalb and V.A. Fassel, Optical atomic emission and absorption methods 405
A.P. D'Silva and V.A. Fassel, X-ray excited optical luminescence of the rare earths 441
EW.V. Boynton, Neutron activation analysis 457
S. Schuhmann and J.A. Philpotts, Mass-spectrometric stable-isotope dilution analysis for lanthanides
in geochemical materials 471
J. Reuben and G.A. Elgavish, Shift reagents and NMR of paramagnetic lanthanide complexes 483
J. Reuben, Bioinorganic chemistry: lanthanides as probes in systems of biological interest 515
T.J. Haley, Toxicity 553
Subject index 587

VOLUME 5
1982, 1st repr. 1984; ISBN 0-444-86375-3
41.
42.
43.
44.
45.
46.

M. Gasgnier, Rare earth alloys and compounds as thin films 1

E. Gratz and M.J. Zuckermann, Transport properties (electrical resitivity, thermoelectric power and
thermal conductivity) of rare earth intermetallic compounds 117
EP. Netzer and E. Bertel, Adsorption and catalysis on rare earth surfaces 217
C. Boulesteix, Defects and phase transformation near room temperature in rare earth sesquioxides 321
O. Greis and J.M. Haschke, Rare earth fluorides 387
C.A. Morrison and R.P. Leavitt, Spectroscopic properties of triply ionized lanthanides in transparent
host crystals 461
Subject index 693

VOLUME 6
1984; ISBN 0-444-86592-6
47.
48.
49.
50.

K.H.J. Busehow, Hydrogen absorption in intermetallic compounds 1

E. Parth~ and B. Chabot, Crystal structures and crystal chemistry of ternary rare earth-transition metal
borides, silicides and homologues 113
P. Rogl, Phase equilibria in ternary and higher order systems with rare earth elements and boron 335
H.B. Kagan and J.L. Namy, Preparation of divalent ytterbium and samarium derivatives and their use
in organic chemistry 525
Subject index 567

VOLUME 7
1984; ISBN 0-444-86851-8

51.
52.
53.

P. Rogl, Phase equilibria in ternary and higher order systems with rare earth elements and silicon
K.H.J. Buschow, Amorphous alloys 265
H. Schumann and W Genthe, OrganametalIic compounds of the rare earths 446
Subject index 573

CONTENTS OF VOLUMES 1-19

xi

VOLUME 8
1986; ISBN 0-444-86971-9
54.
55.
56.
57.

K.A. Gschneidner Jr and EW. Calderwood, Intra rare earth binary alloys: phase relationships, lattice
parameters and systematics 1
X. Gao, Polarographie analysis of the rare earths 163
M. Leskel~t and L. Niinist6, Inorganic complex compounds I 203
J.R. Long, Implications in organic synthesis 335
Errata 375
Subject index

379

VOLUME 9
1987; ISBN 0-444-87045-8
58.
59.
60.
61.

R. Reisfeld and C.K. Jorgensen, Excited state phenomena in vitreous materials 1

L. Niinist6 and M. Leskel~i, Inorganic complex compounds II 91
J.-C.G. Biinzli, Complexes with synthetic ionophores 321
Zhiquan Shen and Jun Ouyang, Rare earth coordination catalysis in stereospecific polymerization
Errata 429
Subject index 431

395

VOLUME 10: High energy spectroscopy

1988; ISBN 0-444-87063-6
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.

Y. Baer and W.-D. Schneider, High-energy spectroscopy of lanthanide materials - An overview 1

M. Campagna and EU. Hillebrecht, f-electron hybridization and dynamical screening of core holes in
intermetallic compounds 75
O. Gunnarsson and K. Sch6nhammer, Many-body formulation of spectra of mixed valence systems 103
A.J. Freeman, B.I. Min and M.R. Norman, Local density supercell theory ofphotoemission and inverse
photoemission spectra 165
D.W.Lynch and J.H. Weaver, Photoemission of Ce and its compounds 231
S. Hfifner, Photoemission in chalcogenides 301
J.E Herbst and J.W Wilkins, Calculation of 4f excitation energies in the metals and relevance to mixed
valence systems 321
B. Johansson and N. M~trtensson, Thermodynamic aspects of 4f levels in metals and compounds 361
EU. Hillebrecht and M. Campagna, Bremsstrahlung isochromat spectroscopy of alloys and mixed valent
compounds 425
J. R6hler, X-ray absorption and emission spectra 453
EP. Netzer and J.A.D. Matthew, Inelastic electron scattering measurements 547
Subject index 601

VOLUME 11: Two-hundred-year impact of rare earths on science

t988; ISBN 0-444-87080-6
73.
74.
75.
76.
77.
78.
79.

H.J. Svec, Prologue 1

E Szabadv~ry, The history of the discovery and separation of the rare earths 33
B.R. Judd, Atomic theory and optical spectroscopy 81
C.K. Jorgensen, Influence of rare earths on chemical understanding and classification 197
J.J. Rhyne, Highlights from the exotic phenomena of lanthanide magnetism 293
B. Bleaney, Magnetic resonance spectroscopy and hyperfine interactions 323
K.A. Gschneidner Jr and A.H. Daane, Physical metallurgy 409
S.R. Taylor and S.M. McLennan, The significance of the rare earths in geochemistry and
cosmochemistry 485
Errata 579
Subject index 581

xii

CONTENTS OF VOLUMES 1-19

V O L U M E 12
1989; ISBN 0-444-87105-5
80.
81.
82.
83.
84.
85.
86.
87.

J.S. Abell, Preparation and crystal growth of rare earth elements and intermetallic compounds 1
Z. Fisk and J.E Remeika, Growth of single crystals from molten metal fluxes 53
E. Burzo and H.R. Kirchmayr, Physical properties of R2Fel4B-based alloys 71
A. Szytuta and J. Leciejewicz, Magnetic properties of ternary intermetallie eompounds of the RT2X2
type 133
H. Maletta and W. Zinn, Spin glasses 213
J. van Zytveld, Liquid metals and alloys 357
M.S. Chandrasekharaiah and K.A. Gingerich, Thermodynamic properties ofgaseoas species 409
W.M. Yen, Laser spectroscopy 433
Subject index 479

V O L U M E 13
1990; ISBN 0-444-88547-1
88.
89.
90.
91,
92,

E.I. Gladyshevsky, O.I. Bodak and V.K. Pecharsky, Phase equilibria and crystal chemistry in ternary
rare earth systems with metallic elements 1
A.A. Eliseev and G.M. Kuzmichyeva, Phase equilibrium and crystal chemistry in ternary rare earth
systems with chalcogenide elements 191
N. Kimizuka, E. Takayama-Muromachi and K. Siratori, The systems R203-M203-MtO 283
R.S. Houk, Elemental analysis by atomic emission and mass spectrometry with inductively coupled
plasmas 385
P.H, Brown, A.H. Rathjen, R.D, Graham and D.E. Tribe, Rare earth elements in biologicalsystems 423
Errata 453
Subject index 455

V O L U M E 14
1991; ISBN 0-444-88743-1
93.
94.
95.
96.
97.

R. Osborn, S.W. Lovesey, A.D. Taylor and E. Balcar, Intermultiplet transitions using neutron
spectroscopy 1
E. Dormann, NMR in intermetallic compounds 63
E. Zirngiebl and G. Giintherodt, Light scattering in intermetallic compounds 163
E Thalmeier and B. Liithi, The electron-phonon interaction in intermetallic compounds 225
N. Grewe and E Steglich, Heaoyfermions 343
Subject index 475

V O L U M E 15
1991; ISBN 0-444-88966-3
98.
99.
100.
101.
102.
103.
104.

J.G. Sereni, Low-temperature behaviour of cerium compounds 1

G.-y. Adachi, N. Imanaka and Zhang Fuzhong, Rare earth carbides 61
A. Simon, Hj. Mattausch, G.J. Miller, W. Bauhofer and R.K. Kremer, Metal-rich halides 191
R.M. Almeida, Fluoride glasses 287
K.L. Nash and J.C. Sullivan, Kinetics of complexation and redox reactions of the lanthanides in aqueous
solutions 347
E.N. Rizkalla and G.R. Choppin, Hydration and hydrolysis oflanthanides 393
L.M. Vallarino, Macroeycle complexes of the lanthanide(llI) yttrium(IIl) and dioxouranium(VI) ions
from metal-templated syntheses 443
Errata 513
Subject index 515

CONTENTS OF VOLUMES 1-19

xiii

M A S T E R INDEX, Vols. 1 - 1 5
1993; ISBN 0-444-89965-0
V O L U M E 16
1993; ISBN 0-444-89782-8
105.
106.
107.
108.
109.

M. Loewenhaupt and K.H. Fischer, Valence-fluctuation and heavy-fermion 4fsystems 1

I.A. Smirnov and V.S. Oskotski, Thermal conductivity of rare earth compounds 107
M.A. Subramanian and A.W. Sleight, Rare earthspyrochlores 225
R. Miyawaki and I. Nakai, Crystal structures of rare earth minerals 249
D.R. Chopra, Appearance potential spectroscopy of lanthanides and their intermetallics 519
Author index 547
Subject index 579

V O L U M E 17: Lanthanides/Actinides: Physics - I

1993; ISBN 0-444-81502-3
M.R. Norman and D.D. Koelling, Electronic structure, Fermi surfaces, and superconductivity in
f electron metals 1
111. S.H. Liu, Phenomenological approach to heavy-fermion systems 87
112. B. Johansson and M.S.S. Brooks, Theory of cohesion in rare earths and actinides 149
113. U. Benedict and W.B. Holzapfel, High-pressure studies - Structural aspects 245
114. O. Vogt and K. Mattenberger, Magnetic measurements on rare earth and aetinide monopnictides and
monochaleogenides 301
115. J.M. Fournier and E. Gratz, Transport properties of rare earth and actinide intermetallies 409
116. W. Potzel, G.M. Kalvius and J. Gal, M6ssbauer studies on electronic structure of intermetallie
compounds 539
117. G.H. Lander, Neutron elastic scattering from actinides and anomalous lanthanides 635
Author index 711
Subject index 753
110.

V O L U M E 18: Lanthanides/Actinides: Chemistry

1994; ISBN 0-444-81724-7
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.

G.T. Seaborg, Origin of the aetinide concept 1

K. Ba|asubramanian, Relativistic effects and electronic structure of lanthanide and actinide
molecules 29
J.V. Beitz, Similarities and differences in trivalent lanthanide- and actinide-ion solution absorption
spectra and luminescence studies 159
K.L. Nash, Separation chemistry for lanthanides and trivalent actinides 197
L.R. Morss, Comparative thermochemical and oxidation-reduction properties of lanthanides and
actinides 239
J.W.Ward and J.M. Haschke, Comparison o f 4 f a n d 5felement hydride properties 293
H.A. Eiek, Lanthanide and aetinide halides 365
R.G. Haire and L. Eyring, Comparisons of the binary oxides 413
S.A. Kinkead, K.D. Abney and T.A. O'Donnell,f-element speciation in strongly acidic media: lanthanide
and mid-actinide metals, oxides, fluorides and oxide fluorides in superacids 507
E.N. Rizkalla and G.R. Choppin, Lanthanides and aetinides hydration and hydrolysis 529
G.R. Choppin and E.N. Rizkalla, Solution chemistry ofactinides and lanthanides 559
J.R. Duffield, D.M. Taylor and D.R. Williams, The biochemistry of the f-elements 591
Author index 623
Subject index 659

xiv

CONTENTS OF VOLUMES 1-19

VOLUME

19: L a n t h a n i d e s / A c t i n i d e s : P h y s i c s - II

1994; ISBN 0-444-82015-9

130.
131.
132.
133.
134.

E. Holland-Moritz and G.H. Lander, Neutron inelastic scattering from actinides and anomalous
lanthanides 1
G. Aeppli and C. Broholm, Magnetic correlations in heavy-fermion systerr~." neutron scattering from
single crystals 123
P. Wachter, Intermediate valence and heavy fermions 177
J.D. Thompson and J.M. Lawrence, High pressure studies - Physical properties of anomalous Ce, Yb
and U compounds 383
C. Colinet and A. Pasturel, Thermodynamic properties of metallic systems 479
Author Index 649
Subject Index 693

Handbook on the Physics and Chemistry of Rare Earths

Vol. 20
edited by K.A. Gschneidner, Jr. and L. Eyring
1995 Elsevier Science B.V. All rights reserved

Chapter 135
F E R M I SURFACES OF INTERMETALLIC C O M P O U N D S
Yoshichika Onuki

Department of Physics, Faculty of Science, Osaka University,

Toyonaka, Osaka 560, Japan
Akira Hasegawa

Department of Physics, Faculty of Science, Niigata University, Niigata 950-21,

Japan

Co~e~s
List of symbols and abbreviations
1. Introduction
2. Theory of energy band structure
2.1. Relativistic effect in the lanthanide
atoms
2.2, Luttinger's theorem on the Fermi surface
2.3. Relativistic band theory
2.3.1. Kohn-Sham-Dirac one-electron
equation
2.3.2. Self-consistent, symmetrized
relativistic APW approach
2.3.2.1. APW matrix elements in
a symmetrized form
2,3.2.2. Determination of
eigenvalues and
eigenfunctions
2.3.2.3. Electron density function
2.3.2.4. Self-consistent
calculation
2.4. Determination of the density of states,
the Fermi surface and the cyclotron
effective mass
2.5. Mass enhancement factors
3. Transverse magnetoresistance and de Haasvan Alphen effect
3.1. Transverse magnetoresistance
3.2. de Haas-van Alphen effect

1
2
8
8
11
12
14
16
16

19
20
21

22
24
26
26
27

4. Experimental results and comparisons with

band calculations
30
4.1. Fermi surfaces in the simple cubic
Brillouin zone
30
4.1.1. RB 6
30
4.1.2. LaAg and YZn
38
4.1.3. RIn3
40
4.1.4. RSn 3
52
4.2. Fermi surfaces in the bcc Brillouin zone 59
4.2.1. RX
59
4.2.2. RAt 2
67
4.3. Fermi surfaces in the hexagonal Brillouin
zone
70
4.3.1. RG%
70
4.4. Fermi surfaces in the tetragonal Bfillouin
zone
74
4.4.1. RRu2Si 2 and RRu2Ge2
74
4.4.2. CeCu2Si2
81
4.5. Fermi surface in the orthorhombic
Brillouin zone
82
4.5.1. RNi
82
4.5.2, RCu 2
87
4.5.3. RCu 6
90
5. Conclusions
95
Acknowledgement
98
References
98

List of symbols and abbreviations

a

lattice constant

AF1,2

antiferromagnetic states

2
APW
c
C
dHvA
e
EF
Ef
Eex
Ei
F
FFT
FS
g

Y. 0NUKI and A. HASEGAWA

KKR
LAFW

augmented plane wave

phase velocity of light
specific heat
de Haas-van Alphen
electronic charge
Fermi energy
f level
exchange splitting energy
eigenvalue of ~pi
de Haas-van Alphen frequency
fast Fourier transformation
Fermi surface
g factor for the spin of the conduction
electron
Land6 g factor
magnetic field
critical field for the metamagnetic
transition
Planck constant divided by 2~
effective H including the exchange field
current
total angular momentum
z-component of J
wave vector
magnitude of k
Boltzmann constant
magnitude of wave vector along the
field direction
Fermi vector along the three principal
axes (i = 1,2, 3)
Korringa-Kohn-Rostoker
linearized augmented plane wave

LDA
LMTO
m0
mb

local density approximation

linearized muffin-tin orbital
free electron mass
band mass

gj
H
Hc
h
Hex
J
J
J~
k
k
kB
kH
kv~

m*
m~
Mosc
MBZ

N(EF)
ne
nh
RKKY
S
T
Tc
TD
TK
TN
TQ
U
v
vF
v

V(r)
V
7
7b
)~
~'m
~,p
/tB
#x~[p(r)]

Ap/p
a(r)
z
wc

effective mass
cyclotron effective mass
oscillatory component of magnetization
magnetic Brillouin zone
density of states at E F
number of electron carriers
number of hole carriers
Ruderman-Kittel-Kasuya-Yosida
extremal cross-sectional area of the
Fermi surface
absolute temperature
Curie temperature
Dingle temperature
Kondo temperature
N6el temperature
quadrupolar ordering temperature
Coulomb repulsive force
velocity of an electron
Fermi velocity
velocity component perpendicular to the
Fermi surface or cyclotron orbit
external potential
hybridized coupling constant
electronic specific heat coefficient
7 calculated from the band model
mass enhancement factor
~, due to electron-magnon interaction
), due to electron-phonon interaction
Bohr magneton
exchange-correlation potential
magnetoresistance,
Ap/p = [p(H)-p(O)]/p(O)
local spin density at r
scattering lifetime
cyclotron frequency

1. Introduction
T h e l a n t h a n i d e c o m p o u n d s are u s u a l l y t r e a t e d in m a g n e t i s m b y a n f - l o c a l i z e d m o d e l ,
b u t s h o w v a r i o u s i n t e r e s t i n g p h e n o m e n a s u c h as v a l e n c e fluctuations, gap states, K o n d o
lattice, a n d h e a v y electrons. T h e s e o r i g i n a t e f r o m t h e 4 f e l e c t r o n s i n t h e l a n t h a n i d e
c o m p o u n d s , w h i c h are e i t h e r b o u n d to t h e l a n t h a n i d e a t o m s or delocalized, i n d i c a t i n g

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

an itinerant nature. The 4f electrons in the atom are pushed deep into the interior of the
closed 5s and 5p shells because of the strong centrifugal potential l(l + 1)/r 2, where l = 3
holds for the f electrons. This is why the 4f electrons possess an atomic-like character in
the crystal. On the other hand, the tail of their wave function spreads to the outside of the
closed 5s and 5p shells, which is highly influenced by the potential energy, the relativistic
effect and the distance between the lanthanide atoms; this results in hybridization of the
4f electrons with the conduction electrons. These cause the various phenomena mentioned
above.
The Coulomb repulsive force of the 4f electron (or the intra-atomic correlation energy)
U at the same atomic site is so strong, for example U ~ 5 eV in Ce compounds, that
occupancy of a same site by two 4f electrons is usually prohibited. The 4f partial density
of states determined by resonant photoemission experiments shows a maximum below the
Fermi level which corresponds to the binding energy or the level of the 4f electrons El.
For example, the distance from this 4f level to the Fermi level is 7.0 eV and 1.2 eV in
SmCu6, 5.5 eV in NdCu6, 3.5 eV in PrCu6 and 2.4eV in CeCu6 (Ishii et al. 1987). The
tail of the 4f partial density of states extends to the Fermi level even at room temperature
in CeCu6 and also slightly in PrCu6, while no trend of 4f states is observed around the
Fermi level in NdCu6.
Near the Ce or Yb end of the R series, the 4f level thus approaches the Fermi level in
energy and the 4f electrons hybridize more strongly with the conduction electrons with
the kinetic energy Ek. This f-hybridized coupling constant is denoted by V. A theoretical
treatment for such a system is called the periodic Anderson model (Anderson 1961).
The parameters Ek, V, Ef and U predominantly control the dynamics of the system.
These values depend actually on the crystal structure. The relation between the magnetic
ordering temperature and the distance between the Ce (or U) atoms is known as a Hill
plot (Hill 1970).
When U is strong and/or V is ignored, the freedom of the charge in the 4f electron
is suppressed, while the freedom of the spin is retained, representing the 4f-localized
state. Naturally, the degree of localization depends on El, where larger Ef helps to
increase the localization. This situation is applied to most of the lanthanide compounds
in which the RKKY interaction (Ruderman and Kittel 1954, Kasuya 1956, Yosida 1957)
plays a predominant role in magnetism. Therefore, the mutual magnetic interaction
between the 4f electrons occupying different atomic sites cannot be of a direct type,
such as in 3d metal magnetism, but should be indirect, which occurs only through the
conduction electrons.
In the RKKY interaction, a localized spin Si interacts with a conduction electron with
spin s, which leads to a spin polarization of the conduction electron. This polarization
interacts with another spin Sj localized on ion j and therefore creates an indirect
interaction between the spins Si and Sj. This indirect interaction extends to the far distance
and damps with a sinusoidal 2kv oscillation, where kv is half of the caliper dimension
of the Fermi surface. When the number of 4f electrons increases in such a way that the
lanthanide element changes from Ce to Gd or reversely from Yb to Gd in the compound,
the magnetic moment becomes larger and the RKKY interaction stronger, leading to

Y. ONUKI and A. HASEGAWA

i00

"~"1

,,,,,,,i

'

'

'''"'1

'

_X=

0.50 . . . . . . . . . . . . . . . . . . . . _
-o.73

~ ' ~ 1

',,,,q

CexLal-xCu6
J//b-axis
.

o
50

,/

9 0 .........
0.01

0,1

Temperature (K)

10

100

Fig. 1. Temperature dependence of

the electric resistivity in CexLa~_xCu6
(Sumiyama et al. 1986).

magnetic order of which the ordering temperature roughly follows the de Gennes relation,
(gj-1)2j(j + 1). Here gj and J are the Land6 g factor and the total angular momentum,
respectively.
Contrary to what happens at large U, higher V tends to enhance the hybridization
of 4f electrons with conduction electrons, thus accelerating the delocalization of the
4f electrons (Koelling et al. 1985). The delocalization of 4f electrons tends to make the
4fband wide. When Ef > V, we have still better localization and expect the Kondo regime
in the Ce (or Yb) compounds.
The Kondo effect was studied for the first time in a dilute alloy where a ppm range
of the 3d transition metal is dissolved in a pure metal of copper. Kondo (1964) showed
that the third-order scattering of the conduction electron with the localized moment of
the transition impurity diverges logarithmically with decreasing temperature, and clarified
the origin of the long standing problem of the resistivity minimum. This became the start
of the Kondo problem, and it took ten years for theorists to solve this divergence problem
at the Fermi energy (Wilson 1975).
The many-body Kondo bound state is now understood as follows. For the simplest
case of no orbital degeneracy, the localized spin S(T) is coupled antiferromagnetitally with the spin of the conduction electron s(+). Consequently the singlet state
{S(i") s(J,) + S(1)" s(T)} is formed with binding energy kBTK. Here the Kondo temperature TK is the single energy scale. In other words, disappearance of the localized moment
is thought to be due to the formation of a spin-compensating cloud of the conduction
electron around the impurity moment.
Kondo-like behavior was observed in the ianthanide compounds, typically in Ce and
Yb compounds (Buschow et al. 1971, Parks 1977, Falicov et al. 1981). For example, the
electric resistivity in CexLal-xCu6 increases logarithmically with decreasing temperature
for all the x-values (Sumiyama et al. 1986), as shown in fig. 1. The Kondo effect occurs
independently at each cerium site even in a dense system. Therefore, this phenomenon
was called the dense Kondo effect.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

The Kondo temperature in the Ce (or Yb) compound is large compared to the
magnetic ordering temperature based on the RKKY interaction. For example, the cerium
ion is trivalent (J = 5/2), and the 4f energy level is split into three doublets by the
crystalline electric field, namely possessing the splitting energies of A1 and A2. The Kondo
temperature is given as follows (Yamada et al. 1984):

T~z=Dexp

- 3

[JexlD(EF)

when T > AI, A2,

(1)

when T < At, A2.

(2)

and
TK = A---~Dexp

I&xlb(EF)

Here D, [Jex [ and D(Ev) are the band width, exchange energy and density of states,
respectively. If we postulate TK ~ 5 K, for D = 104 K, zll = 100 K and A2 = 200 K, the value
of T~ ~ 50K is obtained, which is compared to the S = -Kondo temperature of 10 3 K
defined as T = D exp(-1/IJex I D(EF)). These large values of Kondo temperatures shown
in eqs. (1) and (2) are due to the orbital degeneracy of the 4f levels. Therefore, even at
low temperatures the Kondo temperature is not T but TK shown in eq. (2).
On the other hand, the magnetic ordering temperature is about 5 K in the Ce (or
Yb) compound, which can be simply estimated from the de Gennes relation under the
consideration of the Curie temperature of about 300K in Gd. Therefore, T~ is much
higher than the magnetic ordering temperature, but TK is close to it. Therefore, it
depends on the compound whether or not magnetic ordering occurs at low temperatures
(Brandt and Moshchalkov 1984). As shown in table 1, some compounds such as CeB6
or CeAI2 order antiferromagnetically below 5 K, while CeCu6 (TK = 4 K) does not order
magnetically.
The ground-state properties of dense Kondo systems are interesting in magnetism,
which is highly different from the dilute Kondo effect. In the cerium intermetallic
compounds such as CeCu6, cerium ions are periodically aligned whose ground state
cannot be a scattering state but becomes a coherent Kondo-lattice state. The electric
resistivity p decreases steeply with decreasing temperature, following p ~ A T 2 with a
large value of the coefficient A. The v/A-value is proportional to the effective mass
of the carrier and thus inversely proportional to the Kondo temperature (Kadowaki
and Woods 1986). Correspondingly, the electronic specific heat coefficient y roughly
follows the simple relation y ~ 104/TK (mJ/K2mol). It reaches 1600mJ/K2mol for
CeCu6 because of a small Kondo temperature (Satoh et al. 1989). The Ce Kondolattice compound with magnetic ordering also possesses the large y value even if the
RKKY interaction overcomes the Kondo effect at low temperatures. For example, the
~, value of CeB6 is 250 mJ/K 2 mol, which is roughly one hundred times larger than that
of LAB6, 2.6 mJ/K 2 tool. The conduction electrons possess large effective masses and thus
move slowly in the crystal. These heavy electrons become superconductive in CeCuaSi2

Y. ONUKI and A. HASEGAWA

Table 1
Characteristic properties of Ce compounds~
Compound

Crystal
structure

a (.&)

Tn (K)

(a) Non-Kondo lattice compounds with magnetic ordering

CeGa2
hexa
4.32
11.4
CeRu2Ge2
tetra
4.27
8.5

Tc (K)

y (mJ/K2mol)

8.2

7.5

20

H c (kOe)

(b) Kondo lattice compounds with magnetic ordering

CeSb
cubic
4.54
16.2
Celn3
cubic
4.69
10.2
CeA12
cubic
3.49
3.8
CeCu2
ortho
3.57
3.4
CeB6
cubic
4.14
2.3 b
CeCu2Si2
tetra
4.11
0.7 c

20
130
135
82
250
1000

38
>150
53
18
15
70

(c) Kondo lattice compounds without magnetic ordering

CeCu6
ortho
4.83
CeRu2Si2
tetra
4.19
CeNi
ortho
3.59
CeSn3
cubic
4.72

1600
350
65-85
53

20
80

" Symbols: a, distance between nearest Ce atoms; TN, N6el temperature; Tc, Curie temperature; TQ, quadrupolar
ordering temperature; To, superconducting transition temperature; ~/, electronic specific heat coefficient;
He, critical field for metamagnetic transition.
b TQ=3.2K.
e T0=0.7K '

(Steglich et al. 1980). Therefore, the Kondo-lattice system is called a heavy-electron or

heavy-Fermion system.
W h e n E f < V, the 4 f electrons may tend to be delocalized, manifesting the valencefluctuation regime. CeSn3 and CeNi were once called valence-fluctuation compounds
or mixed-valent compounds. The magnetic susceptibility in these compounds follows
the Curie-Weiss law at higher temperatures than room temperature, possessing the
magnetic moment near Ce 3+, while it becomes approximately temperature-independent
with decreasing temperature, showing a broad m a x i m u m around 150-200 K (Gschneidner
et al. 1985). Thus the valence o f Ce atoms seems to change from Ce 3+ into Ce 4+ (nonmagnetic state) with decreasing temperature.
The ionic radius o f the lanthanide atom decreases with increasing number o f
4 f electrons, which is well known as lanthanide contraction. A plot o f the lattice
constant o f the lanthanide compound versus the atomic number o f the lanthanide element
shows a nearly straight line, except for some Ce, Sm, Eu, Tm and Yb compounds,
where the lanthanides can take integral valencies different from 3 as in Ce 4 and
Sm 2+, Eu 2+, Tm 2+, Yb z+. The valence change in these compounds is brought about
by changing the constitution x (such as in Sml-xLaxB6 Kasaya et al. 1980) or by

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

introducing pressure (for SmS Jayaraman et al. 1970) or a magnetic field (YbB12
Sugiyama et al. 1988), as well as by changing the temperature as mentioned above for
CeSn3 and CeNi.
The first insulating valence fluctuations were studied by Jayaraman et al. (1970) for
SINS. An insulating black phase of SmS at ambient pressure changes into a metallic
golden phase at high pressures. In other words, divalence of Sm 2+ changes into the
intermediate valence between 2+ and 3+. The application of hydrostatic pressure is
associated with a smaller volume, which introduces the 4f valence transition and
consequently delocalization of 4f electrons.
Fermi surface studies are very important to know the ground-state properties of these
various magnetic compounds (Norman and Koelling 1993, Onuki et al. 1991a). Even
in the localized system, the presence of 4f electrons alters the Fermi surface through
the 4f-electron contribution to the crystal potential and through the introduction of new
Brillouin zone boundaries and magnetic energy gaps which occur when 4f electron
moments order. The latter effect may be approximated by a band-folding procedure where
the paramagnetic Fermi surface, which is roughly similar to the Fermi surface of the
corresponding La compound, is folded into a smaller Brillouin zone based on the magnetic
unit cell, which is larger than the chemical unit cell.
If the magnetic energy gaps associated with the magnetic structure are small enough,
conduction electrons undergoing cyclotron motion in the presence of a magnetic field can
tunnel through these gaps and circulate the orbits on the paramagnetic Fermi surface. If
this magnetic breakthrough (or breakdown) occurs, the paramagnetic Fermi surface may
be observed in the de Haas-van Alphen (dHvA) effect even in the presence of magnetic
order.
For Kondo-lattice compounds with magnetic ordering, the Kondo effect is expected
to have minor influence on the topology of the Fermi surface, representing that Fermi
surfaces of the Ce compounds are roughly similar to those of the corresponding
La compounds, but are altered by the magnetic Brillouin zone boundaries mentioned
above. Nevertheless, the effective masses of the conduction carriers are extremely large
compared to those of La compounds mentioned above. In this system a small amount of
4f electron most likely contributes to make a sharp density of states at the Fermi energy.
Thus the energy band becomes flat around the Fermi energy, which brings about the large
mass.
There is a big difference in f-electron character between the Kondo regime and the
valence-fluctuation regime. One may be tempted to think that the 4f electrons in a Kondo
lattice compound with a large value of TK are itinerant. This seems to be true, as shown
later in detail for CeSn3 and CeNi or CeRu2 Si2.
In the following sections we present the dHvA results of the lanthanide compounds
shown in table 2, which are compared to the results of energy band calculations.
Comparisons of the dHvA experiments with band calculations are essentially important to
determine the f character, namely whether the 4f electrons are itinerant or localized. These
Fermi surface properties should shed light on the basic understanding of the strongly
correlated 4f-electron system.

Y. (3NUKI and A. HASEGAWA

Table 2
Rare earth compounds for which dHvA results and energy band calculations are presented in the text.
Elements

La

Ce

Pr

Nd

LaB 6

CeB 6

PrB 6

NdB 6

Ag

LaAg
Prln 3

Ndln 3

Zn

Lain3

Celn 3

Sn

LaSh 3

CeSn 3

Bi

LaBi

CeBi

Sb

LaSb

CeSb

As

Smln 3

Gdln 3

PrSb

SmSb

GdSb

CeAs
YA12

LaA1z

Yb

Ga

LaGa2

CeG%

LaRu2Siz

CeRu2Si2

Ru/Ge

LaRu2Ge2 CeRu2Ge z

Cu/Si

YbAs

CeAIz

Ru/Si

SmGa 2

CeCu2Si2

Ni
Cu

Gd

YZn

In

A1

Sm

LaNi
YCu z

CeNi

PrNi

CeCu 2
LaCu 6

CeCu 6

SmCu 2
PrCu 6

NdCu 6

SmCu 6

2. Theory of energy band structure

2.1. Relativistic effect in the lanthanide atoms
The lanthanide atoms have fairly large atomic numbers and their compounds contain other
heavy atoms as the constituent elements. Therefore, it is essential to take into account
the relativistic effect in calculations of the energy band structures for the lanthanide
compounds. In this section, we explain how the energy and the wave function of an
electron in these compounds may be influenced by relativity.
We treat the neutral cerium atom as an example, and explain the important effect
of relativity on the electrons in its outer shells such as the 4f, 5d and 6s electrons.
The electrons in the s states in both inner (the Xe core) and outer shells have finite
probability amplitudes at the nucleus. As the nuclear potential is deep in the vicinity of the
nucleus, electron velocity approaches light velocity and consequently the relativistic effect
becomes appreciably large. Compared to the non-relativistic theory, the corresponding
energy of all the s (l = 0) electrons decreases significantly, because the s electrons have
relatively large probability amplitudes at the nucleus and their wave functions contract
toward the nucleus. This direct relativistic effect on the s electrons induces an indirect
effect on the other (l ~ 0) electrons. Namely, the s electrons tend to screen more effectively
the nuclear potential which the 4f and 5d electrons feel, and therefore the latter would be
bound more loosely. As a result, their energies increase and their wave functions tend to
spread outward in contrast to the 6s electrons.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

.72_
5d

'<3

i~
J
i

eve

i
i
ir

~ - 0 ,

~x7

2
3
4
D i s t a n c e ( a. u. )

Fig. 2. Electronic charge densities of the 4f,

5p, 5d and 6s orbits for the neutral Ce atom.
The solid and dashed lines show the results
of relativistic and non-relativistic calculations, respectively. In each orbit, except the
6s orbit, two solid curves distinguish the
charge densities having different total angular momentum j, where the charge density
with higher j spreads out more widely than
with lower j. Units are arbitrary, but the
normalization constant is the same for all
states. Both calculations are performed for
the configuration of 4f25p65d6s2, i.e. the
5d orbit is not occupied, although the result
for the 5d state is added in the figure for the
sake of comparison.

,I
-1.0

La Ce Pr Nd Pm Sm Eu Gd Tb DyHo Er Tm Yb Lu

Fig. 3. Relativistic (solid line) and non-relativistic

(dashed line) calculations for the eigenvalues of the 4f,
5d and 6s states of the neutral lanthanide atoms.

It is now clear that in the many-electron atoms relativity causes a different effect on
the energies and wave functions of electrons with different quantum numbers. Actually,
they are so complicated that they cannot be taken into account correctly by the Pauli
approximation nor by any other perturbation theories.
For the sake of illustration, we explain the electronic structure of the cerium atom by
using the self-consistent field method of Liberman et al. (1965), which is based on the
Dirac one-electron equation with the exchange and correlation interactions in the localdensity approximation (Callaway and March 1984). A spherically symmetric potential is
assumed. Thus the solid lines in fig. 2 represent calculated charge densities of the outershell orbits, which are compared to those of the non-relativistic calculations shown by
dashed lines (Herman and Skillman 1963). The corresponding results of the relativistic
and non-relativistic calculations for the eigenvalues of the 4f, 5d and 6s states of the
neutral lanthanide atoms are shown in fig. 3 by the solid and dashed lines, respectively.
Electronic configurations in the outer shells, which are used in calculations, are listed
in table 3. It is remarkable in fig. 3 that magnitudes of the relativistic energy shift and

10

Y. 0 N U K I and A. H A S E G A W A

Table 3
Electronic configuration in the outer shells o f the neutral lanthanide atoms
Z

Element
4f

Neutral atom
5d
5d I

Trivalent ion
6s

57

La

4f

6s ~

4f

58

Ce

4f z

6s 2

4f l

59

Pr

4f 3

6s 2

4f 2

60

Nd

4f 4

6s 2

4f 3

61

Pm

4f s

6s 2

4f 4

62

Sm

4f 6

6s 2

4f 5

63

Eu

4f 7

64

Gd

4f 7

65

Tb

66
67

6s 2

4f 6

6s 2

4f 7

4f 9

6s 2

4f 8

Dy

4 f l

6s z

4f 9

Ho

4 f 11

6s 2

4 f 1
4 f 11

5d 1

68

Er

4 f 12

6s 2

69

Tm

4 f 13

6s 2

4 f 12

70

Yb

4 f TM

6s 2

4 f 13

71

Lu

4 f TM

6s 2

4 f TM

5d 1

the spin-orbit splitting in the 4f states are far larger than those in the 5d or 6s states,
and increase monotonically as the lanthanide element changes from La to Lu. At the
Gd atom having the 4f 7 configuration, a large energy dip of the 4f states is found in
both relativistic and non-relativistic results. Such a large dip would not be found for the
4f 8 configuration. Therefore, the depth of the dip provides an approximate estimatign
of the intra-atomic correlation energy of the 4f electrons in the lanthanide atoms, i.e.
U ~ 400 mRyd (= 5 eV).
The relativistic energy shift and the spin--orbit splitting in the 4f states of the
cerium atom are 200 mRyd and 31 mRyd, respectively. These energies are comparable
to and/or larger than the width of typical 4f bands in the cerium compounds. The spinorbit splitting in the 5d electrons of the cerium atom is 20mRyd in energy, which
is not negligibly small either. Therefore, the relativistic effect should be taken into
account in quantitative calculations of the energy band structures for the lanthanide
compounds, and the relativistic energy band structure calculations should be carried
out self-consistently because the relativistic effect on the 4f electrons is essentially
indirect.
In various cerium compounds, the nearest-neighbor distance between the cerium atoms
ace-ce ranges from 6 to 9 a.u. (Hill 1970), and its intermediate distance a c e - C e / 2 coincides
roughly with the outer-shell region of the cerium atom. As seen in fig. 2, the wave function
of the 4f electron decreases rapidly in the outer-shell region. The overlap between the wave
functions of the 4f electrons in the neighboring cerium atoms is actually so small that
they cannot be itinerant due to that alone. Nevertheless, the 4f electrons can be itinerant

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

11

in a crystal, but only if they hybridize strongly with the 6s, 5p or 5d electrons of the
neighboring cerium atoms and/or the valence electrons of the constituent atoms. If this
hybridization effect is not strong enough, the 4f electrons cannot be itinerant but will be
localized at each cerium ion in the crystal.
It is now clear that the relativistic screening effect of the 6s electrons as well as the
other s electrons tends to promote an itinerant nature of the 4f electrons in a crystal.
It is difficult at present, however, to predict correctly in which cerium compounds the
4f electrons may be itinerant, because the strong correlations between the 4f electrons
make it difficult to derive a quantitative criterion for the itinerancy of the 4f electrons
theoretically.

2.2. Luttinger's theorem on the Fermi surface

The concept of the Fermi surface was introduced originally in the one-electron theory of
metals in which the electron-electron interactions (correlations) are neglected. The Fermi
surface represents a limit of the electronic momentum distribution in the ground state, and
is defined as a constant energy surface in momentum space whose energy is equal to the
Fermi energy. It divides momentum space sharply into two regions which the electrons
occupy or do not occupy.
If a magnetic field is applied to the conduction electrons, the energy levels of the
electrons converge to discrete energy levels called the Landau levels. The interval of the
Landau levels is proportional to the magnitude of the magnetic field. As the magnetic
field is increased, the Landau levels cross the Fermi level successively and cause various
physical quantities to oscillate as a function of field. The de Haas-van Alphen effect is
a typical example of such magnetic oscillatory phenomena (Shoenberg 1984).
In real metals, the conduction electrons interact with each other and therefore the
Fermi surface cannot be defined so simply as in the one-electron theory. Instead of
the one-electron energy, the one-particle excitation energy should be considered. In
momentum space for such interacting electrons or quasi-particles, it is meaningful to
define a constant-energy surface whose energy is equal to the chemical potential. Luttinger
called such a constant-energy surface the true Fermi surface (Luttinger 1960, 1961). In
the ground state, quasiparticles have a long lifetime on the true Fermi surface and the
momentum distribution is discontinuous across it. Luttinger asserted that the volume of
the true Fermi surface is equal to the volume enclosed by the Fermi surface in a noninteracting electron system. Thus the volume of the Fermi surface is independent of
the interactions. He also showed that the true Fermi surface causes various magnetic
oscillatory phenomena, just as the Fermi surface does in the one-electron theory.
Even though the electron-electron interactions may change the topology of the Fermi
surface, the volume of the Fermi surface is kept invariant. This is the essence of the
Luttinger theorem on the Fermi surface. Therefore, the true Fermi surface is observed in
measurements of the de Haas-van Alphen effect. In the following, we will call it simply
the Fermi surface.

12

Y. 0NUKI and A. HASEGAWA

2.3. Relativistic band theory

Among many methods that were proposed for calculations of the electronic energy band
structure in a given periodic potential, the augmented plane wave (APW) method (Slater
1937) and the Green function (KKR) method (Korringa 1947, Kohn and Rostoker 1954)
were useful to produce quantitatively accurate results for simple metals, transition metals
and their compounds, and these methods contributed much to the understanding of various
physical properties of these materials (Dimmock 1971). In both methods, a spatial form
of the potential is assumed in the so-called muffin-tin approximation, in which it is
spherically symmetric within the sphere (the APW sphere) inscribed around each nucleus
and is constant in the interstitial region between the spheres. Inside the APW sphere, the
Bloch wave function is expanded in terms of the solutions of the Schr6dinger equation.
This ensures high accuracy of the eigenvalues, but causes both methods to consume much
computational time especially when applied to compounds with complex crystal structures
having many atoms per primitive cell, because the eigenvalues should be found as zeros
of a determinant with a large dimension.
To improve on this drawback in the APW and the KKR methods, linearized versions for
these methods such as the LAPW method and the linearized muffin-tin orbital (LMTO)
method were invented (Andersen 1975, Koelling and Arbman 1975). In these linearized
methods, it is essential that the electron wave function behaves quite accurately as a
linear function of energy around an energy appropriately chosen in the band so that the
eigenvalues can be determined much faster by a usual matrix diagonalization technique.
The relativistic effect is taken into account by perturbation theories, such as the Pauli
approximation. These methods proved to work well in the understanding of systematic
variations of many ground-state properties of a series of metals and compounds.
However, the Fermi surface depends sensitively on details of the E - k relation, and
the determination of the Fermi surface demands a more precise calculation for the
eigenvalues than that for the cohesive energy. Therefore, for studies of the Fermi surface,
these l'inearized methods should be applied carefully. Errors originating from linearized
approximations should be minimized. Nevertheless, the LAPW method is valuable for
calculations of the electronic structure in lanthanide compounds with complex crystal
structures, such as LaCu6.
For studies of the Fermi surface in the lanthanide compounds, it is necessary to develop
a reliable theoretical method in which hybridization of the 4f electrons with other electrons
as well as the relativistic effect can be taken into account quantitatively. For that purpose,
the relativistic APW method proposed by Loucks (1967) provides a good starting basis.
Loucks derived his original method from the Dirac one-electron equation, which is a
natural extension of Slater's non-relativistic APW method (Slater 1937). It proved to
be a powerful method comparable to a relativistic KKR method (Onodera and Okazaki
1966, Takada 1966). Loucks' method does not accocunt for the symmetrization of the
wave functions by group theory, nor it is a self-consistent method. These shortcomings
are serious limitations for calculations of the energy band structure in the lanthanide
compounds.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

13

To achieve symmetrization, a direct product of the space operator and the spin
operator was constructed as a new operator, under which the variational expression of
Loucks' method is kept invariant (Yamagami and Hasegawa 1990). By the projection
operator technique, it is straightforward to derive a symmetrized form of the relativistic
APW method, which covers both symmorphic and non-symmorphic space group.
Koelling (1969) proposed a similar method, but in his method he derived a symmetrized
form by using a basis function of the Pauli-spinor type and the Foldy-Wouthusen
transformation. An approach proposed by Yamagami and Hasegawa(1990) is more general
and has wider applicability than his method because the relativistic effect can be taken
into account both inside and outside the APW spheres consistently without using any
approximations such as the Foldy-Wouthusen transformation. Koelling's method was not
self-consistent.
Yamagami and Hasegawa carried out a self-consistent calculation of the energy
band structure by solving the Kohn-Sham-Dirac one-electron equation by the densityfunctional theory in a local-density approximation (LDA). This self-consistent, symmetrized relativistic APW approach was applied to many lanthanide compounds and
proved to give quite accurate results for the Fermi surface.
In LDA, the electron correlations are taken into account only by a mean field
approximation which utilizes the correlation energy of the uniform electron gas. In the
Ce compounds where the 4f electrons are believed to be itinerant in the ground state,
such as in CeSn3, the topology of the Fermi surface can be described by the band
structure calculated in LDA. However, the strong intra-atomic correlation effect between
the 4f electrons should be considered for consistent explanations of the Fermi surface, the
electronic specific heat coefficient and the cyclotron effective mass. Beyond LDA, there
are two approaches by which the correlation effect between the 4f electrons is taken into
account in an explicit way. One is p - f mixing theory and the other is renormalized band
theory.
p - f mixing theory was first proposed for explaining complicated magnetic properties
of the Ce pnictides (Kasuya et al. 1987). In this theory, the 5d and 6s electrons in the
Ce atom and the valence electrons in the pnictide atom form usual bands just as calculated
for the La pnictides in LDA. At each Ce atom, one 4f electron is localized and the 4f state
is split by spin-orbit interaction. The 4f energy levels relative to the Fermi energy can be
determined phenomenologically from the observed XPS spectrum. Hybridization of the
4f state with other states is then introduced in an LMTO scheme. There are no adjustable
parameters in calculating the Fermi surface. It was found that the occupied 4f state lies
far below the Fermi level, and weakly affects the Fermi surface. This p - f mixing theory
explained well various magnetic properties of the Ce pnictides, especially experimental
results for the de Haas-van Alphen effect in CeSb.
The Fermi surface of CeSb was thus calculated on the basis of p - f mixing theory,
although it was not easy. This is because the Ce pnictides are semimetals and their energy
band structures and Fermi surfaces are simple. In the usual Ce compounds where the
4f electrons are believed to be localized, it is difficult to carry out such calculations.
Therefore, simplified calculations were carried out for such Ce compounds as CeA12

14

. ()NUKI and A. HASEGAWA

(Guo 1990) and CeCu2Si2 (Harima and Yanase 1991a, 1992a) in which the 4f states
are assumed to belong to the core states and the 4f bands are removed artificially if they
appear in the conduction band.
For the antiferromagnetic compound NdB6 another simplified calculation was tried
by Kubo et al. (1993). The antiferromagnetie band structure was first determined selfconsistently with the standard KKR method by assuming that the 4f electrons are itinerant,
and then the 4f bands were removed from the band structure obtained from the final
potential. In these simplified calculations, the electron correlations are not taken into
account.
The second theory beyond LDA is called renormalized band theory; this was applied
to CeSn3 (Strange and Newns 1986) and CeRu2Si2 (Fulde 1991, Zwicknagl et al. 1990).
Like p - f mixing theory, the correlation interactions between the 4f electrons are taken
into account phenomenologically in the LMTO scheme. The set of phase shifts for the s,
p and d states in the constituent atoms is given by a usual LDA calculation, and that for
the 4f state in the Ce atom should be determined so as to include strong correlations. To
calculate the electronic structure close to the Fermi energy, the values of the phase shifts
and their derivatives at the Fermi energy are necessary. For CeRu2 Si2 Zwicknagl et al.
(1990) determined the values of the phase shifts at the Fermi energy so that one 4f electron
per Ce atom participates in forming the Fermi surface, and fitted their derivatives so that
the density of states at the Fermi energy agrees quantitatively with the observed electronic
specific heat constant y. Therefore, all necessary phase shifts are fixed and there are no
adjustable parameters. Zwicknagl et al. (1990) explained the origins of the major observed
dHvA branches and the magnitude of the cyclotron effective masses consistently.
In the following sections we outline the self-consistent, symmetrized relativistic
APW approach.

2.3.1.

Kohn-Sham-Diracone-electronequation

Here, we describe the framework of band theory in a local-density approximation

(Hohenberg and Kohn 1964, Kohn and Sham 1965). The usual band theory is based on
the Kohn-Sham one-electron equation. The wave function ~Pi and the eigenvalue Ei of
an electron in the state i in a crystal is given as a solution of the equation,

(-A+v(r)+ f ~_rqdr2n(r') , +btxc[n(r)])~pi = Eil~i,

(3)

where n(r) is the electron density. In the left-hand side of eq. (3), the first term is the
kinetic energy, v(r) is the nuclear potential, the third term is a direct Coulomb potential,
and ~c[n(r)] is the exchange-correlation potential defined by

#xc(n)

d(nExc(n))
dn

'

(4)

FERMI SURFACES OF INTERMETALLIC C O M P O U N D S

15

where Exc(n) is the exchange-correlation energy as a function of the density n in a

homogeneous electron gas. The electron density in a crystal is calculated by

n(r) = ~-'~fi [~Pi(r)]2 ,

(5)

wherefi is the occupation number of state i. Therefore, eqs. (3) and (5) should be solved
in a self-consistent way. The exchange-correlation energy of a homogeneous electron gas
is used under the assumption that the electron density is a slowly varying function of
space.
The exchange-correlation potential should be a complicated function of n(r) and many
formulae were proposed for practical purposes. Here we show the formula proposed by
Gunnarsson and Lundqvist (1976). This formula is frequently used in calculations for the
Fermi surfaces which we shall explain in this chapter. It is expressed as a product of the
exchange part and the correlation part as
#xc(n) = 13(n) #x(n),

(6)

where #x(n) is the Kohn-Sham exchange potential,

1

#[n(r)] --

-y

2 [3zc2n(r)] 5

(7)

and

fi(n) = 1 + 0.0545rs In 1 +

1 1.4

rs /

(8)

r3
",1/3
where rs = t~svn)
.
Relativity affects the kinetic term and the exchange-correlation potential in the KohnSham equation. As investigated in detail for the uranium atom and the cerium atom, the
relativistic effect on the exchange correlation potential is rather small and therefore we use
#xc[n(r)] in a relativistic band structure calculation. The relativistic effect on the kinetic
term is appreciably large and can be taken into account by adopting the Kohn-ShamDirac one-electron equation instead of eq. (3) as follows:

{ ap + fi + Io(r) + I J __2n(r')
]~_rqar' ' + I#xc[n(r)] _'~1~Pi = WlPi,

/3=[g

01] (1 is the 2 x 2 unit matrix),

(9)

(11)

where W is the total energy of state i including the rest energy. In eq. (9), ~Pi is the fourcomponent wave function. In a relativistic energy band theory, eqs. (9) and (5) should be
solved in a self-consistent way.

16

Y. ONUKI and A. HASEGAWA

Spatial forms of the electron density and the potential are determined in the muffin-tin
approximation. Then, the wave function is expanded by the solutions of the Dirac equation
inside the APW spheres and by the plane waves in the interstitial region. To determine the
expansion coefficients and the eigenvalues, a variational principle is derived from eq. (9)
as

Wf+ii OtOd~2=j!i+iI'IH~dS-'2fspheres)(OI+OII)ta'/5(I-OII) dS,

(13)

where regions I and II are the inside and the outside of the APW spheres, respectively, and
/5 is the unit vector in the normal direction outward the APW spheres. In this variational
principle, the surface integral performed on the APW spheres is introduced to ensure
good eigenvalues for the composite trial wave functions.

2.3.2. Self-consistent, symmetrized relativistic APW approach

In a relativistic APW approach, symmetrization with double space groups is necessary
for the following reasons:
(1) The symmetry of the electronic states in a crystal can be identified.
(2) The dimension of the matrix in the eigenvalue problem can be reduced and
its determinant behaves linearly as a function of energy in the vicinity of the
eigenvalues. This result is important, because it becomes easy to determine
the eigenvalues without loss of computational accuracy. Note that in a nonsymmetrized form it behaves parabolically because of the double degeneracy due
to spin.
(3) A self-consistent calculation can be carried out for compounds with a complicated
crystal structure.
Here it is assumed that a crystal consists of different kinds of atoms, and the number
of atoms of the sth kind in a primitive cell is ns. As for the shape of the one-electron
potential and the electron density, the muffin-tin shape approximation is adopted. It is
assumed that the space group is non-symmorphic. A non-symmorphic space group cannot
be divided into translational and rotational parts, i.e., every rotation of the point group of
the non-symmorphic space may be accompanied by a non-primitive translation, while a
symmorphic space group can be divided neatly into translational and rotational parts. For
example, CeNi has the non-symmorphic space group Cmcm, and CeSna the symmorphic
space group Pm3m. For the application of the following formulae to the symmorphic
space groups, the non-primitive translation can be set to be zero. In this section, we use
atomic units in which h = 1, the rest mass of the free electron is m0 = and the velocity
of light is c = 274.072.
2.3.2.1. APW matrix elements in a symmetrized form. N sets of plane waves with spin
are prepared first. Then, the Bloch function q)(k,r) is expanded in terms of symmetrized

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

17

plane waves which transform according to the vth column of the Xth extra representation
F;~(Rj) of the sub-group of order g associated with wave vector k as

O(k,r) = E ~ ~ E UXv~)7"*~W(pa, r).

a

3,

(14)

Here, t/J(p~,r) is the four-component plane wave with wave vector k~a) and spin m = + l
which is selected as the progenitor from the ath plane wave set, andp~ is an abbreviation
for (k~a), m). 7)xv~ is the full projection operator defined by
=

(15)

O(Rj),

.
J

where l~ is the dimension of the extra representation X and the sum is taken over all
elements of the single group. In eq. (15), the operator O(Rj) is defined by the direct
product of the space operator O1 (Rj) and the spin operator O:(Rj),
o(Rj) = OI(Rj) O O2(Rj),

(16)

where

o (Rj)

(Rjl,,(Rs)),

(17)

0
02(Rj)= (D1/; (Rj) D1/2(Rj)
),

(18)

In eq. (17), v(Rj) is the non-primitive translation associated with Rj. The term DI/2(Rj)
in eq. (18) is the spin representation of the full rotation group The sum over # in
eq. (14) is executed when linearly independent plane waves with the same symmetry
are projected out. It is easy to show that the variational expression (13) is kept invariant
under the operation O(Rj). Because of its similar symmetry property, we may use the
four-component relativistic APW function.
By the theory of irreducible representation, it is possible to determine the eigenstates
belonging to each extra representation separately: A variational calculation leads to a set
of linear equations for the expansion coefficients which hold irrespective of v,

~
3

~ (k, E) '~v~'
rrZ(3)= 0,
A//at~3~,

a = 1,2,. .. , N,

/~ = 1, 2,. .., n,

(19)

#'

with the matrix element

Adxau3~,(k, E) = E
j

exp

[-iRjkla, v(Rj)]

x. (Rj) Dn,,
1/2(Rj) M lj (n')m '
F,,u

(20)

n'

wherep3 = (kl 3), n). In eq. (20) M lj (m)

n' represents the matrix element calculated between
the single relativistic APW functions with (k~a),m) and (k)3),n'), where k) 3)= Rjkl ~).

18

Y. ()NUKI and A. HASEGAWA

Therefore, the eigenvalues of the Xth irreducible representation at k are determined by

the condition
det AdxauB~,(k, E) = 0.

(21)

For each eigenvalue, the expansion coefficients ~t;.~(a)

v , are determined by eq. (19) together
with the normalization condition.
On account of the phase factor which appears in the matrix element of eq. (20), it
is convenient to use the irreducible multiplier representation for non-symmorphic space
groups instead of F ~ , (Rj), which is defined by

A~,~,(Rj) = exp [ik . v(Rj)] F~,(Rj),

;~

(22)

because Rjk~ ~)- k is equal to a reciprocal lattice vector. Following a group theoretical
prescription proposed by Sahni and Venkataraman (1970), we calculate the irreducible
multiplier representations for the necessary sub-groups of both symmorphic and nonsymmorphic double space groups.
Actually, the condition that the relativistic effect may be negligibly small in the
interstitial region (i.e. hkj/mo <<c) holds fairly well in the energy band structure
calculations for various lanthanide compounds. For computational purposes, it is
practical to define the energy-independent components in the matrix elements for each
representation as .Aau8~, , 13at~#,,, Cat~u,(s, l), and Dau~u,(s, l) and then rewrite eq. (20) in
the form
Z
A.tauS,,(k
, E) = - ( E x
z
+/3a~tt,

+~

[ Catcu,(s
, l) {lrh(as) + (l + 1)tl-t+l(as)}
Z

~
s

(23)

+Da~,(s, l) {-~/l(as) + ~-I+1(a,)}],

where
~:(a~) - cf(as)
- -

t + 1
(24)
g(a~)
as
In eq. (23), V0 is the average potential in the interstitial region of space. In eq. (24),
f(as) and g(a~) denote the small and large components of the electron wave function at
the sphere radius as, respectively, which depend on E. They are obtained by solving the
radial Kohn-Sham-Dirac equations for a given E,
d(cfr)dr - ( r - 1)~-frr - (E - Vs(r))gr,

(25)

dgx _ c_ 2 (E - V~(r) + c2) cf x - (x + 1) g_yr,

(26)
dr
r
where Vs(r) is the potential inside the s-type APW sphere consisting of the nuclear
potential, the direct Coulomb potential and the exchange-correlation potential.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

19

Here are the explicit expressions for the energy-independent components of the matrix
elements in the symmetrized form that appear in eq. (23):

AX~.,e.,= ~

exp [-i (~) -

k).

v(Rj)] A./(Rj)
~"
D '/2.,.Rj)

(27)

Or~c~ V _

exp [ikl~~). r~

4 g g a s J l ( k V a~)
t'/s

'~U

[k{ ~') . k~~'] exp [-i<~ ~,- k). v(Rj)] A~//<&)

BL, e,,, : ~
J

(28)

xO:/'Z(RJ)[6~a61J-E 4"ga~J'
kk{~e)
f:'J],za,
.~ ~)'+exp[i
(~)
s

GL~~,(s, l) = E exp [-i ,(k!~>J-

~V

k).

i=l

v(Rj)] Az*./,.,p''D
o v2`D'.,n,.:,

(29)
(k(l'~)a,)j,

x [~j,
D~.,~/(s,

l) :

exp [-i(<~>-k)

i~-~ ~
]

(~'~)a~)~,=, exp like;'). 'r;] P, (k(,+

.

v(Rj)]A~*/(R,)(mlaln) '" D ,/.t.;)(kl

+
'/~ " " "

nt

[4JraZ

x [~-~j,(kl~)a,)j~(k~)a,)

"
,vq "'J / j ,
~,=t exp [ikl~~). ~::]P'{~(~,.l-(~)~]

J .

x k <~)]

(30)

where

k/>:Rjk?),

~(~) : k/) <+,

i.|j

--

(31)

and r~ is the position vector of the ith atom of the sth kind, g2 is the primitive cell volume,
is the Pauli matrix and jr(x), Pt(x) and P~(x) are the spherical Bessel function, the
Legendre polynomials and derivative of the Legendre polynomial, respectively. Moreover,
the notation (m lal n') is defined by

(+ I< ) = e~,

( I< m ) = ex T ie.

(32)

where ex, ey and ez are the unit vectors in the x, y and z directions, respectively. The
energy-independent components C~W~W(s, l) and D ~ u , (s, l) are also used to calculate the
electron radial density function within the APW spheres.
2.3.2.2. Determination of eigenualues and eigenfunctions. The irreducible representations for the symmetry points, axes and planes of the Brillouin zone are classified
into three types according to the Herring criterion (Herring 1937). These three types

20

Y. 0NUKI and A. HASEGAWA

are defined as (a), (b), and (c), in which the Herring criterion is equal to -1, 0 and 1,
respectively.
In type (a), there is no degeneracy between irreducible representations, and the
twofold degeneracy due to spin can be resolved by symmetrization. For each irreducible
representation, the symmetrized APW determination behaves linearly as a function of
energy in the vicinity of eigenstates, so that both eigenvalues and eigenfunctions can be
determined easily.
In type (b), two different irreducible representations are degenerate. This type can be
treated by merely calculating one of the two representations.
In type (c), the twofold degeneracy due to spin cannot be resolved by symmetrization.
Both eigenvalues and eigenfunctions can be determined by using Soven's technique to
overcome the problems of the double roots which result from the twofold degeneracy due
to spin (Loucks 1967). If the subgroup contains the inversion, the following relations,
4
4.
Ada~/3u,(k
, E) = Ad_a,_/3,,(k,
E),

(33)

4 (k, E) and Ad_at~_/3

4. u, (k, E)
hold between the symmetrized APW matrix elements A/tau~u,
in which both wave vector and spin of the progenitors p~ and p~ are inverse, where the
plus sign is applied if the progenitors have the same spin and the minus sign is applied
if they have opposite spins. In such a case, it is possible to transform the matrix into a
form suitable for Soven's technique. At general points of the Brillouin zone, calculations
can be carried out by using the original, unsymmetrized relativistic APW method of
Loucks (1967).
2.3.2.3. Electron density function. In order to carry out a self-consistent energy band
calculation, it is necessary to calculate the electron density function associated with
a symmetrized relativistic APW function. For the Bloch state (k,)~), the spherically
averaged radial density function within the s-type APW sphere is given by
~r~Z(k, r) = 1 ~
ns

[OZ2(s' l) {l~t(r) + (1 + 1)~_,_l(r)}

(34)

+ Q2(
Z s , l) (~l(r) - ~-l-l(r)}],

where
1

TTz(a)r r4(~):4
a

Z S
Q2(,
l)= ~l Z

a
=

fi

Z
iz

[g~:(as)as] 2

(3

iS l),

(35)

~'

r r4(a) r)~(t3)~Z
(~
t-/ vl~ u r i c 1 z~ apf3g'~.,

l),

(36)

(37)

FERMI SURFACES

OF INTERMETALLIC

COMPOUNDS

21

The normalization constant N z for the wave function belonging to the irreducible
representation F ~ is written explicitly as

N z _ x+ ~ E { Nzc(s' l)[l~ +(l+

-- N a

1)~s,_,]

+N~(s, l ) ( ~ -

~s_l_,)}

(38)

where

NXa= E ~ E Z ~'ufr~(aT
c'~,t~,"'a'(/3)AZc,#,/3u',
a

/3

g'

rrX(a)* rrX(/3)t'X
a

N:(s,

l) = ~

~t

E
a

(39)

/3

tS l),

(40)

tJvu"
~z(a)* r,~(/3)~Zuvt
r ~, ua~,/3,,to,' l),

(41)

tt'

/3

#'

:~ = foa~ [/r(r) 2 +gr(r) 2] r2dr

[g~(aAaA 2

(42)

The component Ba#,/3#,x does not appear in the normalization constant. The ratio N~/Nz
is the probability amplitude in the interstitial region of the primitive cell, and the ratio of
each term in the sum over s and l in eq. (38) to N x gives the probability amplitude
partitioned according to the angular momentum l in the s-type APW sphere. These
probability amplitudes are used to calculate the partial density of states. The spherically
averaged radical density a,(r) within the s-type APW sphere is obtained by summing
a)(k, r) over all occupied states as follows,

s,coro(r)+

Wok :(k, r)
n

(43)

The first term on the right-hand side of eq. (43) represents the radial density of all the
core electrons within the s-type APW sphere. In the second term on the right-hand side
of eq. (43), the notation is changed and the Bloch states are identified by (n, k) instead
of ()~, k), where n is a band index. This term represents the average of cry(k, r) with a
weighting factor Wnk. A constant electron density in the region between the APW spheres
is determined so that the unit cell is kept electrically neutral.
2.3.2.4. Self-cons&tent calculation. It is now possible to carry out a relativistic band
structure calculation self-consistently. Following Mattheiss' method (Mattheiss 1964), a
starting electron density for a crystal is constructed by superposing the self-consistent
atomic electron densities, which are calculated for the neutral atoms using the method of
Liberman et al. (1965). In calculations both for the atom and the crystal, the exchange

22

Y. ONUKI and A. HASEGAWA

and correlation interactions of electrons are taken into account by the density functional
theory in the local-density approximation. As for the charge densities of the core states
in compotmds, a frozen-core approximation is adopted, i.e. the atomic charge densities
of the core states are assumed to represent the core electrons in the compounds. In
calculations of the energy band structure and the Fermi surface for CeSn3 (Hasegawa
et al. 1990), for example, the atomic charge densities for Ce (the Xe core except for the
5p state) and for Sn (the Kr core plus the 4d state) were assumed to be frozen in CeSn3.
The 5p state in the Ce atom should be included in the itinerant states and should be
determined self-consistently together with other states in the outer shell states, because
hybridization between them must be accounted for. A new crystal density function is
calculated by eq. (43). A self-consistent calculation can be carried out in a manner similar
to the ordinary non-relativistic band structure calculation (Mattheiss et al. 1968). In fig. 4,
a flow chart is represented for a self-consistent energy band structure calculation by a
symmetrized relativistic APW method in LDA.
2.4. Determination of the density of states, the Fermi surface and the cyclotron

effective mass
To grasp the nature of the energy band structure visually, it is useful to make a drawing
of the energy dispersion curves. Using a final potential of a self-consistent calculation,
many eigenvalues are calculated for each band at symmetry points and along symmetry
axes in the irreducible part of the Brillouin zone, and smooth energy dispersion curves
are obtained by an interpolation scheme based on a eubic-spline method.
The density of states at the energy E per unit volume is related to the band structure
by the well-known relation
2 j ' dS
N(E) = (2z) 3
h-~vk'
(44)
where vk is the absolute magnitude of velocity and the surface integral is performed over
the constant-energy surface with energy E. Many eigenvalues should be calculated for
each band at fine mesh points in the irreducible part of the Brillouin zone. If necessary,
eigenvalues can be interpolated further at finer mesh points by an interpolation scheme
based on a Fourier series expansion. The integration is performed by a tetrahedral method
(Jepson and Anderson 1971, Rath and Freeman 1975). Each band can Contain two
electrons per primitive cell in the Brillouin zone, and the total number of Bloch electrons
determines the Fermi energy EF. It is given by
NBIoCh =

N(E) dE.

(45)

The density of states is also calculated for each band as a function of energy. The
density of states at the energy E for the nth band is further partitioned according to the
s-type APW sphere and the angular momentum I. Such partial density is defined by
Nnl(E)

2 [ Pn,(k),~
aa
(2jr)3 J
v~

(46)

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

23

Initial crystal charge density

Solve the Kohn-Sham-

Superpose tlae atomic charge

Dirac equation in the LDA.

densities.

O'in(i+ 1)(r) =

O.in(i)(r)

1 - f ) Oin(i)(r) + f Gout(i)(r)

Potential in the LDA

I
VT(r) = v(r) + VH(r) + gxc(r)

E =E +AE

>~

Extra renresentation F t
Symmetrlze plane waves.

Matrix ~lemen~
-~ ~o~,pp,(/,,E)

Eigenvalges
det Idf~c,p,,.(k g ) l = 0

~W tutrge densltv 1
IJout(i)(r)

Convergence te~t,
I O'out(il(r)- O'i.(i)(r) I < 5

No

~Yes

E-k curve
Density of states
Fermi surface
Cyclotron mass

Fig. 4. A flow chart for a self-consistent energy band structure calculation by a symmetrized relativistic
APW method in LDA. In each cycle of the iteration, the total electron potential VT(r ) is given by the sum of
the nuclear potential v(r), the direct Coulomb potential VH(r ) and the exchange-correlation potential #xc(r).
To construct a new input potential, the parameter f ( < 1) is introduced to optimize the convergence.

24

Y. 0NUKI and A. ttASEGAWA

where the probability amplitude P~t(k) is obtained by integrating an(k, r) from r = 0 to

r = as. In particular, the number of electrons in the f state contained in the Ce APW sphere,
nf, is given by
nf = Z

f E F N~e,~=3(E) dE"

(47)

The electronic contribution to the specific heat of a metal at low temperature varies
linearly with temperature, Ce~ = ~'bT, with the coefficient Yb given in one-electron theory
by the density of states at EF as
= 2k N(EF).

(48)

Since the density of states on the free-electron model is proportional to the electron mass,
this coefficient is usually expressed by a specific-heat effective mass or a thermal effective
mass mtb, which is defined by
(49)

y_..bb= mtb,

7o

m0

where 70 and m0 are the free-electron values of the two constants.

The interpolation scheme for each band is used to display a perspective view of the
Fermi surface and to calculate the extremal (maximum or minimum) cross-sectional area
of the Fermi surface. It is also used to calculate a cyclotron effective mass mob which is
defined for a given orbit on the Fermi surface by
mcb = zzh

J --'vdk

(50)

where v is the component of velocity normal to magnetic field and the integral is
performed around the orbit. This formula can readily be transformed into

mcb- h20S(kH)

(51)

where S(kn) is the cross-sectional area of the Fermi surface which is perpendicular to
field and kH is the wave number along the field direction. In dHvA experiments we can
determine the cyclotron mass m2 at the extremal cross-sectional area, which is compared
to the band mass mcb.
2.5. Mass enhancement factors
The experimental values for the low-temperature electronic specific heat coefficient ~, and
equivalently the thermal effective mass mt are usually larger than the theoretical values of

FERMI SURFACES OF INTER_METALLIC COMPOUNDS

25

Yb and mtb which are determined by the band structure alone, as mentioned above. The
mass enhancement factor 3, is defined by the ratio (Gschneidner and Ikeda 1983)
y _ m~' _ 1 + ~ .
~b

(52)

mtb

Origins for 2 are ascribed to the many-body effects which cannot be taken into account
in the usual band theory. As the most probable origins, the electron-phonon interaction
and the magnetic interaction are considered, and their contributions are denoted by 2p
and 3,rn, respectively. Therefore, 3, is expressed as a sum of two contributions,
)~ = /~p nt- Am.

(53)

The experimental cyclotron effective mass mc is also usually larger than the theoretical
value mcb defined in eq. (51). Therefore, the enhancement factor for the cyclotron effective
mass can be defined in the same way, such as
rn~' _ 1 + )~.

(54)

mcb

It should be noted, however, that the magnitude of this enhancement factor may be
different from orbit to orbit on the Fermi surface. In sect. 4 we present values of Y/~'b
and me~rob for many compounds, where mb is simply used instead of mcb.
Properties of the electron-phonon term ,~p in normal metals such as Pb, including its
temperature dependence, were thoroughly studied and are well understood at present (Goy
and Castaing 1973). Its magnitude is significantly smaller than 1. If it were large, it might
cause lattice instability. In contrast to this small value of ~p, the magnetic contribution
)~m can take a huge value in the heavy-electron compounds. For example, it amounts to
more than 100 in CeCu6.
The magnetic contribution or the magnon enhancement mechanism Xm can be divided
into two terms according to its origins. The first term occurs in many lanthanide
compounds in which the 4f electrons are localized at lanthanide ions and their spin
fluctuations enhance the effective mass of the conduction electrons via c - f interactions
such as RKaKY interaction and the many-body Kondo effect, where c stands for
conduction electrons and f for 4f electrons. A large ~ value, which is observed in CeBr,
as shown in table 1, is an example of this term. In this cerium compound, however, the
4f electron affects the Fermi surface only weakly. The Fermi surface of CeB6 is quite
similar to that of LAB6, although the value of ~m reaches 100. In this case, the y value
of LaB6 is assumed to be the ]7b value of CeB6. In P r B 6 and N d B 6 , the 4f electrons are
localized, but do not form a Kondo lattice. Therefore, only a small mass enhancement
factor of Xm = 1-2 is observed in these lanthanide compounds. Nevertheless, this mass
enhancement is larger than the mass enhancement due to the electron-phonon interaction
mentioned above.

26

Y. ONUKI and A. HASEGAWA

Apart from the electron-phonon case,/~m depends in some cases strongly on magnetic
field as shown for Pr (Forgan 1981, Fulde and Jensen 1983). For the ferromagnetic
paramagnon or spin waves, the effect of magnetic field is direct and reduces J.m strongly
when the Zeeman energy exceeds the characteristic energy of the spin fluctuation.
For the antiferromagnetic case, the effect is more complicated and in some cases, for
example in CeB6, as we will see later, ~m increases with increasing field because of an
induced magnon softening. However, when the Zeeman energy is much stronger than
the characteristic energy of the magnetic fluctuation, ~,m decreases rather rapidly with
increasing field.
Another magnetic contribution to ~-mmanifests itself in the iron-series transition metals
and their compounds. In these compounds, the 3d electrons are itinerant and their spins
are fluctuating. Their magnetic properties stem predominantly from the spin fluctuations.
In calculations of the band effective mass mb for these 3d electron compounds, a greater
part of the contribution of the magnetic interaction can be taken into account, and thus the
magnitude of ~-mis not very large. For example, MnSi is typical in this system, possessing
a relatively large y value of 36 mJ/K2 mol. In MnSi, ~m is found to be only about 4,
although the conduction carriers are heavy electrons with the large cyclotron mass of
15m0 (Taillefer et al. 1986, Fawcett et al. 1970).
As will be explained in sect. 4, the 4f electrons in cerium compounds such as CeSn3
and CeNi, which belong to the valence-fluctuation regime, are itinerant in the ground
state and contribute directly to the formation of the Fermi surfaces. Their Fermi surfaces
can be well described by band theory. The enhancement factor in CeSn3 and CeNi is not
large, less than 5, as in the 3d electrons of the iron-series transition metals.

3. Transverse magnetoresistance and de Haas-van Alphen effect

3.1. Transoerse magnetoresistance
The high-field transverse magnetoresistance A p / p = [p(H) - p(0)]/p(O), in which the
directions of magnetic field and current are perpendicular to each other, provides
important information on the overall topology of the Fermi surface (Fawcett 1964). Under
the high-field condition of ~ocr >> 1, it is possible to know whether the compound under
investigation is a compensated metal with an equal carrier number of electrons and holes,
ne = nh, or an uncompensated metal, n e e nh, and whether or not open orbits exist. Here,
coc = ell~roSe is the cyclotron frequency, r is the scattering lifetime, mS is the effective
cyclotron mass and cocr/2~r means the number of the cyclotron motions performed
by the carrier without being scattered. The characteristic features of the high-field
magnetoresistance are summarized as follows: (1) when all cyclotron orbits are closed,
(a) for the uncompensated metal the magnetoresistance saturates, A p / p ~ H , and (b) for
the compensated metal the magnetoresistance increases quadratically, A p / p ~ (wcz) 2.
(2) When some of the cyclotron orbits are not closed but form open orbits, the
magnetoresistance increases quadratically and depends on the current direction as

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

27

A p / p ~ H2cos2c~, where a is the angle between the current direction and the open orbit
direction in k-space. This is true regardless of the state of compensation.
If we count the number of valence electrons in a primitive cell, most of the lanthanide
compounds are even in number, meaning that they are compensated metals. In this case
the transverse magnetoresistance increases as H n (1 < n ~<2) for a general direction of
the field. Note that the integer n is not equal to 2 because the high-field condition is
not fully satisfied in the real compounds. When the magnetoresistance saturates for a
particular field direction, often a symmetrical direction, there exist some open orbits
whose directions are parallel to ar x H , namely a = ~/2 in k-space.
Experimentally the current direction is fixed to a crystal symmetry axis of the sample
and the sample is slowly rotated under constant magnetic field which is perpendicular to
the current direction. The presence of open orbits is revealed by (a) spikes against a low
background for the uncompensated metal and (b) dips against a large background for the
compensated metal.

3.2. de Haas-van Alphen effect

Under a strong magnetic field the orbital motion of conduction electrons is quantized and
forms Landau levels. Therefore various physical quantities show a periodic variation with
H -1 since increasing field strength causes a sharp change in the free energy of the electron
system when a Landau level crosses the Fermi level. In the three-dimensional system this
sharp structure is observed at the extremal (maximum or minimum) cross-sectional area
of the Fermi surface perpendicular to the field direction because the density of states also
becomes extremal.
The magnetization or the magnetic susceptibility is the most common one of these
physical quantities, and its periodic variation is called the de Haas-van Alphen (dHvA)
effect (de Haas and van Alphen 1930, 1932). It provides one of the best tools for the
investigation of Fermi surface properties such as the extremal cross-sectional area S,
the cyclotron mass mc and the scattering lifetime r of metals (Shoenberg 1984).
Sometimes other physical quantities are also measured; for example, torque, static strain,
ultrasonic velocity, and magnetoresistance, etc. The last type of measurement is called
the Shubnikov-de Haas effect.
The theoretical expression for the oscillatory component of magnetization Most
due to the conduction electron was given by Lifshitz and Kosevich (1955). The
dHvA experiments are usually conducted by the field modulation method. The detected
signal or effective output voltage Vosc in the usual 2co-technique is proportional to Most
and is given as follows:
2~crF
Vsc = Z Arsin( T
-b ~r),
r

(55)

02 S -1/2 exp(_arm c TD/H) co (szgrm~ ~

Ar ~ J2(x)TH -1/2 ~
~
s \ 2m0 /

(56)

28

Y. 0NUKI and A. HASEGAWA

or

Ar ~ J2(x) T m 1/2

02S

e~'

exp [-

rmZ(V + VD)m]

1-exp(-2arm~T/H)

2jr2ckB
a-

-1/2

cos\

2m0 J '

(57)

2~Fh
x-

H2 ,

(58,59)

where h is the modulation field, J2(x) is the Bessel function, possessing a maximum
at x=3.1, and r (r = 1, 2, ... ) is a positive integer showing higher harmonics of the
oscillation. The output voltage or the magnetization is periodic on 1/H and has a
dHvA frequency F:
F = ---~-hS
2~e

(60)

which is directly proportional to the extremal cross-sectional area S at zero field. The
amplitude factor A is related to the thermal damping at finite temperatures T and the
Landau level broadening TD which is caused by impurities, crystalline imperfections or
strains. TD is called the Dingle temperature and is given by
TD=~h
~-1,
2JrkB

(61)

where r is the scattering lifetime of the carrier.

For the fundamental oscillation (r = 1), we can determine the cyclotron mass me from
the temperature dependence of the dHvA amplitude A(=A 1) in eq. (57), namely from the
slope of a plot of In {A [ 1 - exp(-2am~T/H)]/T} vs T at constant H and h by using a
method of successive approximations, and we can obtain the Dingle temperature TD or the
scattering lifetime r from the field dependence of the amplitude in eq. (56), namely from
the slope of a plot of In [AH1/2sinh(am~T/H)] vs H -1 at constant J2(x) and temperature.
The amplitudes of the higher harmonics (r ~>2) become vanishingly small, and the
fundamental one (r= 1) becomes dominant in the usual dHvA measurements. However,
when the cyclotron mass is not large and the temperature becomes lower than 1 K,
the higher harmonics become detectable. To distinguish the higher harmonics from the
fundamental one, it is necessary to check carefully the magnitude, intensity and angular
dependences of the dHvA frequencies and their cyclotron masses.
The quantity ]OzS/Ok~i]-1/2 is the inverse square root of the curvature factor 02S/Ok~.
The rapid change of the cross-sectional area around the extremal cross-sectional area
along the field direction diminishes the dHvA amplitude for this extremal area.
The term cos(~grmS/2mo) is called the spin factor. When g = 2 (free-electron value)
and m* =0.5m0, this term becomes zero for the fundamental oscillation (r= 1) and the
dHvA oscillation vanishes for all values of magnetic field. This is called the zero spinsplitting situation in which the up and down spin contributions to the oscillation cancel

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

29

LaSh3

(a)

Magnetic Field
Fig. 5. Schematic picture of the change in
the extremal cross-sectional areas depending
on the up and down spin states. AF~ and
AF2 mean the different antiferromagnetic
states, and H c is the critical field showing
the metamagnetic transition. St, and S~ are
the extremal areas for the up and down
spin electrons, respectively, obtained from
the dHvA measurements in the AF2 region
(Harima 1988).

'Y,Ti

80

90kOe

H//<III>
(b)

0.SK
~7z-- Or

2,e

L, ojhlll,

7,+2z!

II

0
5
10
15x1070e
Fig. 6. dHvA oscillation and its FFT spectrum for LaSn3 (Umehara et al. 1991a). The Greek letters in the
FFT spectrum designate the various orbits.

out, and this can be useful for determining the g value. Note that in this situation the
second harmonics for r = 2 should have a full amplitude.
When the extremal area changes linearly with increasing external field, the dHvA frequencies of the up and down spin electrons coincide, giving the extremal crosssectional area for zero field mentioned above. Many lanthanide compounds show a
magnetically ordered state at low temperatures. Conduction electrons in this system
have different Zeeman and exchange energies, depending on the up and down spin
electrons. For example, the antiferromagnetic AF~ state of these compounds often changes
into a different antiferromagnetic AF2 state or into the field-induced ferromagnetic
(paramagnetic) state. In this case, we usually get different Fermi surface areas for the
up and down spin electrons, S T and St, when the field is increased above the critical
field showing the metamagnetic transition He, as shown in fig. 5. The spin factor Sr
becomes

y'grm e

=cS 2mo

(62)

30

Y. (3NUKIand A. HASEGAWA

where Hex is defined by the exchange splitting energy Eex = ~BHex. In ferromagnetic
compounds, it is possible to obtain different Fermi surface areas associated with the up
and down spin electrons in zero field.
Simply thinking, the dHvA oscillation is detected when the high-field condition is
almost satisfied; ~oer/2zr > 1 and the spacing between the Landau levels is larger than the
thermal broadening kBT; hcoc > kBT. I f the magnetic field H is 100kOe or 10T and the
carrier possesses a cyclotron mass of 10m0, the following conditions for the temperature
and the scattering lifetime are required: T < 1.3 K and 7: > 3.6 10-11 s or TD < 0.03 K.
A temperature of 0.4 K can be attained in the He3-cryostat (Windmiller and Ketterson
1968), and much lower temperatures are obtained in a dilution refrigerator (Reinders et al.
1987). Values of 77= 10-12-10 -11 s or TD=0.1-1 K are usual in samples.
The exact dHvA oscillation contains many dHvA frequencies Fi (i = 1, 2, 3, ... ) or
cross-sectional areas Si and becomes a sum of their contributions, which are analyzed
by the fast Fourier transformation (FFT) method. The amplitude Ai corresponds to the
amplitude in the FFT spectrum. Figure 6 shows the dHvA oscillation and its FFT spectrum
for a field along the (111) direction of the cubic crystal LaSn3 at 0.5 K (Umehara et al.
1991a). From the FFT spectrum we can see many dHvA oscillations due to harmonics
or sums and differences of the several dHvA frequencies.

4. Experimental results and comparisons with band calculations

4.1. Fermi surfaces in the simple cubic Brillouin zone
4.1.1. RB 6
The rare earth hexaborides RB 6 crystallize in the cubic (CaB 6 type) structure which
possesses a CsCI type arrangement of R atoms and B6 octahedra. Figure 7 shows the
crystal structure of RB6 and its simple cubic Brillouin zone. LaB6 is a reference non-f
compound. CeB6 is a typical Kondo-lattice compound undergoing two magnetic ordering
(a)

(b)

()

Fig. 7. (a) RBt-cubic crystal structure. Large spheres without pattern and small spheres with pattern show the
R atoms and the B atoms, respectively.(b) Brillouin zone of the simple cubic crystal lattice.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

31

Table 4
Characteristic properties of the RB 6 compounds with a cubic crystal structure~
Compound

T N (K)

7 (m J/K2 tool)
2.6

LaB 6

Pauli para

ne

CeB6

2.3
TN = 3.2 K
Kondo lattice

n~

PrB6

7.0

ne

Ndt36

7.8

n~ = nh

250

Fermi surface
three ellipsoids connected by necks
similar to LaBr, but spin-split

similar to LaBr, but spin-split

strongly altered by MBZ

Symbols: n~, number of electron carders; nh, number of hole carders; when no=nh the carders are
compensated.
MBZ, magnetic Brillouin zone.

transitions at the quadrupolar ordering temperature TQ = 3.2 K and at the Nrel temperature
TN=2.3 K (Effantin et al. 1985, Komatsubara et al. 1983). Existence of quadrupolar
ordering is due to the quartet F8 ground state in the 4flevels. PrB6 (TN = 7.0 K) and NdB6
(TN = 7.8 K) are typical localized 4f systems with magnetic ordering. Their characteristic
properties are summarized in table 4.
The measurements of the dHvA effect in LaB6 (Suzuki et al. 1988, Ishizawa et al.
1977, 1980, Arko et al. 1976), shown in fig. 8, revealed that the Fermi surface consists of
a set of three equivalent nearly spherical ellipsoids, denoted by ai (i = 1, 2 and 3), which
i

i--~

LaB6 i
lZ { 1 0 0 }

3.

{110}

10 8

o~
{D

o}

~i0 7

6
>

10 5

80
<ii0>

0
<i00>

30

60
<iii>

Field Angle (Degrees)

90
<ii0>

Fig. 8. Angular dependence of the dHvA frequency

in LaB 6 (Ishizawa et al. 1977, 1980, Suzuki et al.
1988). The solid and dashed lines connecting the
data are guidelines. The Greek letters designate the
various orbits.

32

Y. 0NUKI and A. HASEGAWA

R

LaB6

(b)
M

(a)

<i00> /<110>

(e)

"q---------- ~"<i00> R
Fig. 9. (a) Cross-sections of the multiply connected ellipsoidal Fermi surfaces (Ishizawa et al. 1977). Co) Main
three multiply connected ellipsoidal Fermi surfaces (Hasegawa and Yanase 1977b). (c) Twelve pocket Fermi
surfaces in LaB 6 (Harima et al. 1988). The pocket Fermi surface is enlarged for visual convenience. The Greek
letters designate the various orbits.

are connected by necks. This topology of the Fermi surface was constructed from the
magnitude of the dHvA frequencies and the angle range where the dHvA branches a i
were detected.
The energy band structure of LaB6 is characterized by the wide B 2s-2p bands which
are split into the bonding and antibonding bands and by the La 5d bands which lie across
the energy gap between the bonding and antibonding bands. This feature originates from
a particular configuration of atoms in the CeB6 crystal structure. The six s states of the
B atoms in an octahedron form d-like orbitals with F12 symmetry about the center of the
octahedron, and the six p states of these B atoms also form d-like orbitals with both
F12 and F25 symmetries about the same center (Longuet-Higgins and Roberts 1954).
These d-like orbitals and the La 5d states have nearly equal energies, and therefore strong
hybridization occurs between them. The strong hybridization causes a large wave-vector
dependence of the d bands, and the Fermi surface is formed by one such d(eg) band
(Hasegawa and Yanase 1977b). The Fermi surface of LaB6 is shown in figs. 9a,b. It
consists of three equivalent electron sheets which are centered at the X point and are
connected by small necks which intersect the E axes in the simple cubic Brillouin zone.
The total number of carriers is almost equal to one electron per primitive cell. The
electrons on the Fermi surface have dominantly La d character, and on the average the
magnitude of their cyclotron effective masses is smaller than the free electron mass.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

33

The neck orbit was not detected, however, in the above mentioned dHvA measurements
done by Ishizawa et al. (1977) and Arko et al. (1976). Later, branches Pi (i = 1. . . . . 6)
shown in fig. 8 by open circles were detected by the torque method (Ishizawa et al.
1980) and were attributed to the necks because the angular dependence of branches Pi are
consistent with the topologies of the necks. This Fermi surface of the neck is, however,
thin and rather cylindrical, which is inconsistent with the short and thick neck constructed
from the unobserved region of the ellipsoidal branches cti and also the results of band
calculations. This puzzle was solved later by a combination of the improved ultrasonic
dHvA measurements done by Suzuki et al. (1988) and the careful band calculations done
by Harima et al. (1988). In fig. 9c twelve pocket Fermi surfaces calculated by Harima
et al. are shown as an enlarged scale by a factor of ten. It was shown that branches p; are
not due to the necks but due to the small and flat electron Fermi surfaces. The data shown
as triangles in fig. 8 were obtained by the ultrasonic dHvA measurements (Suzuki et al.
1988). The complete observation of branches P3 and p5 is a clear evidence for existence
of the small closed Fermi surfaces.
New band calculations done by Harima et al. were made by shifting the unoccupied
La 4f levels upwards by an amount of 0.10 Ry, which leads to a new band which crosses
the Fermi energy very slightly. Langford et al. (1990) also confirmed existence of the
pocket Fermi surface by the LMTO band calculations. The neck orbit is, however, not
detected experimentally because of the rapid variation of the cross-sectional area around
the extremal neck orbit, implying a large curvature factor 02S/Ok~.
Similar Fermi surface topologies were obtained in CeB6 (0nuki et al. 1989a, Joss
et al. 1987, 1989, Goto et al. 1988a, Suzuki et al. 1987, van Deursen et al. 1985) and
PrB6 (Onuki et al. 1985b, 1989d, van Deursen et al. 1985), as shown in figs. I0 and 11,
respectively. Judging from the values of branches ai for field along the (100) direction,
the main Fermi surface is more spherical in CeB6 than in LaB6 and PrB6. The ratio of
the maximum to minimum areas of the ellipsoidal Fermi surface is about 1.16 in CeB6,
1.24 in PrB6 and 1.27 in LAB6.
We also show in fig. 12 the cross-sectional area of the small and flat Fermi surface
deduced from branch pi for RB6. It is approximated as an ellipsoidal Fermi surface. The
Fermi wave vectors kFi (i = x, y and z) along the three principal axes are kvx = 0.012 (2~v/a),
kFy = 0.023 (2st/a) and kvz = 0.0044 (2Jr/a) in LAB6.
In PrB6 two kinds of pockets Pi and pf as well as the ellipsoids a3 and a~ are found, as
shown in fig. 11. The cross-sectional areas ofpi and pf are 118 and 28 times larger than
in LAB6, respectively. The existence of two kinds of Fermi surfaces in PrB6 is explained
by an exchange splitting of the up and down spin states of the conduction electrons as
shown in fig. 5. This is due to a change of the antiferromagnetic spin structure at about
10kOe in PrB6 (Galera et al. 1992). The up and down spin states have different effective
Fermi surface areas and cyclotron masses. For example, branches a3 and a~ in PrB6
have the values of 8.19x 1070e (1.95mo) and 7.25x 1070e (2.52m0) for field along the
(100) direction, respectively.
A similar spin splitting of the Fermi surfaces is expected in CeB6 because the
antiferromagnetic state of the so-called phase III changes into that of phase II

34

. 0NUKI and A. HASEGAWA

I

'

' --

CeB6

{lOO}
0~i

10 8

+4--

'

'

'

'

PrB8

{11o}

(i00}

! ~ ~ o {Ii0}

0~

108

L+-:~-o-~-o-+ ~

2 _~_~

~__a~j~-~-~F~-

_ .~

. . . . . .

%,
O
o)
O

"

~=-~-+~-o.__~_o~ . . . . .

107

>
~'Lx
-

P3~, P4

'

~P3 ~ D5

p,

106
-.e.-,a--',,~o~ ~
I

30

<110>

s
,

0
30
60
<100>
<111>
F i e l d Angle (Degrees)

90
<110>

Fig. 10. Angular dependence of the dHvA frequency

in CeB 6, Data shown by circles, crosses, squares and
triangles are cited from ()nuki et al. (1989a), van
Deursen et al. (1985), Joss et al. (1987, 1989) and
Suzuki et al. (1987) and Goto et al. (1988a) Goto
(1992), respectively. The solid lines connecting the data
are guidelines. The Greek letters designate the various
orbits.

LaB6

CeB6

30
<ii0>

0
30
60
<i00>
<iii>
F i e l d Angle (Degrees)

90
<ii0>

Fig. 11. Angular dependence of the dHvA frequency

in PrB~ (Onuki et al. 1985b, 1989d). The solid and
dashed lines connecting the data are guidelines. The
Greek letters designate the various orbits.

PrB 6

O!

p'
- 0.2 @rc/a~

Fig. 12. Cross-sections of the pocket-Fermi surfaces in

LAB6, CeB~ and PrB 6. These electron pockets are fiat
in character (()nuki et al. 1989a).

(quadrupolar ordering) at about 15 kOe. Goto (1992) confirmed that branches P6 and p~
shown in fig. 10 are due to two kinds of Fermi surfaces with different spin states. The
Fermi surface of branch Pi in CeB6 accidentally possesses the same size as branch pf
in PrB6.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

35

Table 5
dHvA frequencies F and cyclotron masses m~ in RB 6 compounds (Ishizawa et al. !977, 1980, Onuki et al.
1989a,d, Goto 1992) a
RB 6

a3
F

a~
m~

y
m~

LaB 6

7.89

0.64

3.22

CeB 6

8.67

14-21

2.19

PrB 6

8.19

1.95

7.25

a
F
NdB 6

9.90

2.52

2.00

F
3.44

2.43

F
1.27

p'

m~

0.85

0.49

0.0052

0.046

1.30
0.188

9.2
5.5

0.120

1.7-3.5

0.080

4.6

1.94

0.94
0.79

0.94
0.66

0.590

0.64

0.150

0.28

m~

g
m~

15.5

3.27

c
m~

e
m~

m~

m~

hi

1.47

0.95

m*
1.08

F values in 107 Oe, m~ expressed in m 0.

Here, we note the occurrence of dHvA branches with frequencies of (1-2)x 107 0 e
in CeB6 and (3-4)x 107 0 e in PrB6. These branches are not present in LAB6, and are
probably produced by the small antiferromagnetic Brillouin zone boundaries in CeB6 and
PrB6.
The cyclotron masses in LaBr, CeB6 and PrB6 are summarized in table 5. All masses
in CeB6 are heavily renormalized by the many-body Kondo effect compared to those of
LaB 6 and PrB6. The cyclotron masses in PrB6 are also three times larger than those in
LAB6, which should be attributed to the usual electron-magnon interaction.
The cyclotron mass of branch a3 in CeB6 shows a striking variation as a function
of magnetic field, as shown in fig. 13. Quite a different field dependence is observed
in the electronic specific heat coefficient y (Mfiller et al. 1988), as shown also in
fig. 13. It varies from 250mJ/K2mol in zero field to 50mJ/K2mol in 220kOe. Here,
the masses shown by the solid line through the crosses in fig. 13 were estimated from the
y value of CeB6 by using the relation ofmc(CeB6 ) = me(LaB6) [y(CeB6)/y(LaBr)], where
mc(LaB6) = 0.6 lm0 for a3 and y (LaBr)= 2.6 mJ/K 2 tool. There exists a clear discrepancy
between the results of two types of experiments. The reason for this is thought to be
as follows. When one compares the Fermi surface of CeB6 to that of PrBr, one is led
to believe that the observed a3-Fermi surface in CeB6 corresponds to branch a3 in
PrB6. In PrB6, the mass of branch a~ is larger than that of branch a3. Therefore, it is
natural to assume that branch a~ in CeB6 has also a larger mass and hence could not
be observed experimentally. Therefore, the missing branch af is the main origin of the
present discrepancy. We note that the field dependence of the cyclotron mass or the yvalue was discussed briefly in sect. 2. The theoretical treatment for CeB6 was done by
Wasserman et al. (1989).
The dHvA frequencies observed in NdB6 (Onuki et al. 1989d) are substantially different
from those of LAB6, PrB6 and CeBr, as shown in fig. 14. One main reason for the
discrepancy between the dHvA branches in NdB6 and in LaB 6 or PrB 6 seems to be the

36

Y. 0NUKI and A. HASEGAWA

'

7--

'

{100} [ {110l
/

CeB6
H//<I00>

i00

10 8 o o o o o O O O

NdB6

~OOOoo O O o o o

D .................... .; .................

++

c
g
phase

50

g @'15 b

~i~6o-o-o d~

hs .........

f8 ............

::"<he _g_~x-,,~

II

10 7

+ +
+
++

...........: - - - - - - ~ : ~

- ~ _ o 6 ~ ~ ~

:o_~j I

fld2

o o g o . q o o o o o ~o~.a,i

oo~%oooov

~ OOOoo

hl

o~

J2;:

100

200

300

Magnetic Field (kOe)

Fig. 13. Field dependence of the cyclotronmass of the
a 3 orbit. The data are represented by circles (0nuki
et al. 1989a), triangles (Joss et al. 1987, 1989) and
a square (van Deursen et al. 1985). The cyclotron
masses estimated from the low-temperature specific
heat coefficientsat various fields are shown by crosses
(Mfiller et al. 1988). The solid and dashed lines
connecting the data are guidelines.

10 6

<110>

I
30

30
<100>

60
<111>

90
<110>

Field Angle (Degrees)

Fig. 14. Angular dependence of the dHvA frequency
in NdB 6 (Onuki et al. 1989d). The solid and dotted
lines are the results of band calculations (Kubo et al.
1993). The letters designate the various orbits.

large magnetic gaps in NdB 6 due to a larger number of the occupied 4f electrons in a
new magnetic Brillouin zone.
Min and Jang (1991) and Kubo et al. (1993) calculated the Fermi surfaces both in the
paramagnetic and antiferromagnetie states of NdB6 by the LMTO and K K R methods,
respectively, treating the 4f electrons as the core electrons. The Fermi surface in the
paramagnetic state o f NdB6 is similar to that o f LaB 6. On the other hand, the Fermi
surface in the antiferromagnetic state, which was calculated by Kubo et al. (1993), can
be roughly understood by the band-folding o f the paramagnetic Fermi surface into a
smaller tetragonal Brillouin zone based on the magnetic unit cell which contains the
double chemical unit cells.
Figure 15a shows a multiply connected hole Fermi surface in band 27. A closed orbit
centered at the M point on the M - X - F plane is denoted by branch c. This branch is
equivalent to branch e in the paramagnetic state. Figure 15b shows two ellipsoidal electron
Fermi surfaces centered at the R point and a multiply connected electron one centered at
the F point in band 28. NdB6 in the antiferromagnetic state is thus a compensated metal,
as opposed to the uncompensated metal LAB6, o r NdB6 in the paramagnetic state. Here,
the electron Fermi surface o f NdB6 in the paramagnetic state is essentially a sphere with
a hollow in the center.
The solid and dotted lines in fig. 14 represent the theoretical results based on the
Fermi surface in fig. 15 (Kubo et al. 1993). Namely, the solid lines represent the field

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

37

NdB6 ( A n t i f e r r o m a g n e t i e )
X

~r

(a)

[uollT

(band27)

Z
A

(b)

M
A.._

[010]

/
[10

......

'on (band28)

Fig. 15. (a) Hole and (b) electron Fermi

surfaces in the antiferromagneticstate of
NdB6 (Kubo et al. 1993). The letters
designate the variousorbits.

configuration of [110] to [100] in the (001) plane and [100] to [011] in the (0il) plane,
while the dotted lines represent that of [011 ] to [001 ] in (100) and [001 ] to [ T10] in (110).
To explain the experimental results, it is necessary to consider three kinds of magnetic
domains, whose tetragonal axes are directed along the three equivalent (100) axes. For
example, the solid lines denoted by g and hi in fig. 14 correspond to orbits g and hi in
fig. 15 for field along the [100] direction. On the other hand, orbit a cannot be closed
for this field direction, but can be closed for field along the [001] direction. This orbit
corresponds to branch a in fig. 14, as shown by a dotted line. Detected dHvA branches
are thus explained by considering the three equivalent domains.
The cyclotron masses are in the range of I-2.4m0, roughly three times larger than
those of LAB6. Mass enhancement due to the electron-magnon interaction in NdB 6 is
comparable to that in PrB6, in contrast to extremely large enhancement in CeB6, as shown
in table 5.
Magnetoresistance measurements were done for RB6. Reflecting the multiply connected
Fermi surface, open orbits were observed for LAB6, PrB 6 and Ndl3 6 (Onuki et al. 1989d).

38

Y. ONUKI and A. HASEGAWA

LaAg

CD

()

<>
0
30

Fig. 16. CsC1 type cubic crystal

structure.

<110>

80

60

<i00>
<iii>
Field Angle (Degrees)

90

<ii0>

Fig. 17. Angular dependence of the dHvA frequency in LaAg (Niksch et al. 1987). The solid lines show the
results of band calculations. The Greek letters designate the various orbits.

Table 6
Characteristic properties of LaAg and YZn with the CsCl-type cubic crystal structure
Compound
LaAg
YZn

TN (K)

1,
(mJ/K2mol)

Pauli para
Pauli para

n~= nh
nh

9.9

Fermi surface
ellipsoids
multiply connected hole Fermi surface

4.1.2. LaAg and YZn

LaAg and YZn are Pauli paramagnetic compounds with the CsC1 structure as shown in
fig. 16. Their characteristic properties are listed in table 6.
Magnetoacoustic quantum oscillations were obtained for LaAg (Niksch et al. 1987).
Two kinds o f ellipsoidal Fermi surfaces, denoted by a and fi, were observed, as shown in
fig. 17. The solid lines are theoretical ones based on relativistic self-consistent Augmented
Spherical Wave (ASW) calculations.
LaAg is a compensated metal. This is simply expected from the number o f valence
electrons; La 3+ and Ag 1+ and the primitive cell contains one molecule of LaAg. Just as
in LAB6, the La atoms make a simple cubic lattice, and therefore the La d(eg) band is
expected to play an important role in determining the Fermi surface. Unlike the B atoms,
however, the Ag atoms make a simple cubic lattice, and thus hybridization between the
L a d band and the Ag sp bands cannot be so strong as the La d-B sp hybridization in
LaB 6. The energy band structure of LaAg is rather similar to those of simple transition
metals. The density of states as a function of energy has a sharp peak which originates
from a flat part o f the La d(eg) band, and the Fermi level lies slightly below the sharp
peak. Consequently the cyclotron effective masses are small in LaAg.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

39

LaAg

(a) hole

(b) electron

Fig. 18. (a) Hole and (b) electronFermi surfaces in LaAg

(Niksch et al. 1987).The Greekletters a and/3 designate
the orbits.

Figure 18 shows theoretical Fermi surfaces calculated by Niksch et al. (1987).

Branches a and fl correspond to small hole ellipsoids centered at the M point, possessing
a dHvA frequency of 1.86x 107 Oe for field along the (110) direction and small electron
ellipsoids centered at the X point, possessing a dHvA frequency of 7.20x 106 Oe for
field along the (111) direction, respectively. A large hole Fermi surface centered at the
F point and a large electron one centered at the R point were not observed experimentally,
probably because of a resolution problem.
The conduction carriers are mainly due to the La 5d states, as mentioned above.
However, the cyclotron masses are 0.92m0 for the small hole and 1.10m0 for the small
electron, which are roughly twice as large as the corresponding band masses of 0,44m0
and 0.87m0, respectively. This is consistent with the ratio of the measured electronic
specific heat coefficient y, 9.9 mJ/K 2 tool to the calculated one Yb, 5.1 mJ/K 2 mol (Hill
et al. 1976). This large mass enhancement for LaAg is mainly due to electron-phonon
interaction because the band Jahn-Teller effect is found in LaAg and also slightly due to
hybridization of the La 4f electrons with the conduction 5d electrons.
dHvA experiments were done by Jan (1973) for YZn with the CsCl-type crystal
structure. Two kinds of branches a and y were observed. Figure 19 shows the angular
dependence of the dHvA frequency for branch y. The solid line in fig. 19 shows the
results of APW band calculations done by Hasegawa and Yanase (1977a), which is based
on the Xc~ method with an exchange parameter of 0.6.
Branch y is due to a multiply connected hole Fermi surface with the topology of
intersecting cylinders directed along the FX direction in band 3, as shown in fig. 20, and
branch a is due to a small electron-Fermi surface at the R point in band 4. In contrast
to LaAg, YZn is an uncompensated metal with a hole character. This is simply expected
from the number of valence electrons in a primitive cell; y3+ and Zn 2+. Compared to
LaAg, YZn has one more conduction electron per primitive cell, and therefore the Fermi
level approaches closely the sharp peak in the density of states which originates from a
flat part of the Y 4d(eg) band (Hasegawa and Yanase 1977a). As a natural consequence,
the electrons on the main Fermi surface (the y sheet) dominantly have the Y 4d character,
and their cyclotron effective masses are larger than those in LaAg.

40

Y. 0NUKI and A. HASEGAWA

0.4

R
(100}

{110}

YZn

OO"
O

Fig. 20. IntersectingcylindricalFermi

surfacesin band 3 ofYZn (Hasegawa
and Yanase 1977a). The Greek letter }, designates the orbit.
0.2

30
<ii0>

30
<i00>

60
<iii>

90
<Ii0>

Field Angle (Degrees)

Fig. 19. Angular dependence of the dHvA frequency in YZn (Jan 1973). The solid lines show the results of
band calculations (Hasegawaand Yanase 1977a). The Greek letter 7 designates the orbit.
The cyclotron mass is 2.08m0 for 7 and 0.237m0 for a, compared to the band masses
of 1.21m0 for 7 and 0.14m0 for a. The mass enhancement is relatively large, which is
mainly due to electron-phonon interaction as in LaAg.
Magnetoresistance experiments (Sellmyer et al. 1975) confirmed that YZn is an
uncompensated metal with a multiply connected Fermi surface which favors open orbits,
at least, along the (110) direction. These results are consistent with the results of band
calculations mentioned above.
4.1.3. RIn3
There are many RX3 compounds with the AuCu3 type cubic structure. Figure 21 shows
the crystal structure of RX3 where the corner-sites are occupied by the R atoms and the
face-centered sites are occupied by the X atoms. RIn3 and RSn3 are typical examples.
Celn3 is a well known Kondo-lattice compound showing antiferromagnetic ordering at
10K. The ordered moment 0.65/~B per cerium atom is comparable to the value 0.71 ~R
expected from the F7 ground state (Lawrence and Shapiro 1980). Nevertheless, CeIn3
possesses a large electronic specific heat coefficient of 130 mJ/K 2 mol at low temperature,
indicating a heavy-electron system (Nasu et al. 1971). Therefore Celn3 is considered to
be in a similar situation as CeB6, possessing a low Kondo temperature and displaying
magnetic order at low temperatures. The ground crystalline field state is, however,
the F7 doublet and thus no quadrupolar ordering is expected. Prln3 is paramagnetic,

FERMI SURFACES OF INTERMETALLICCOMPOUNDS

4I

Fig. 21. The AuCu3 type cubic crystal structure of RIn3. Large spheres
without pattern and small spheres with pattern show the R atoms and the
In atoms, respectively.
Table 7
Characteristic properties of the RIn3 compounds with the AuCu3-typecubic crystal structure

RIn3

TN (K)

y
(mJ/K2mol)

Lain3

Pauli para

ne = nh

5.3-6.3

CeIn3

10
Kondo lattice
singlet
5.9
14.7, 15.2
(TQ=15.9K)
42

n~= nh

130

ne = nh
ne = nh

11.4

PrIn3
NdIn3
Stain3
Gdln3

ne=nh

n ~n h (?)

Fermi surface
nearly spherical electron Fermi surface (band 7);
complicated hole Fermi surface (band 6)
similar to Lain3 (branch d), but strongly altered
by MBZ
similar to Laln3
similar to Lain3, but affected by MBZ
similar to Laln3, but strongly altered by MBZ
similar to LaIna, but affected by MBZ

possessing a singlet ground state; NdIn3 ( T N = 5 . 9 K ) , Smln3 ( T N = 1 4 . 7 and 15.2K,

T Q = 1 5 . 9 K ) and GdIn3 ( T N = 4 2 K ) are localized 4 f systems. Their characteristic
properties are summarized in table 7.
As a proper reference for the RIn3 compounds, the electronic structure was investigated
for Lain3 by various theoretical methods. A n analysis o f the partial density o f states
revealed that the Bloch states in the lowest three bands consist dominantly o f the In s state,
and those in the upper part o f the valence band consist mainly o f the In p and L a d states.
The Fermi level lies in the vicinity o f a sharp peak in the density o f states. The Fermi
surface consists o f two sheets o f bands 6 and 7 (Hasegawa 1982). The band 7 electron
Fermi surface is essentially a sphere, which is centered at the R point and bulges slightly
toward the M point and contains 0.38 electrons per primitive cell. The band 6 hole Fermi
surface looks quite complex and contains the same number o f holes as electrons, because
Lain3 has an even number o f Bloch electrons per primitive cell and thus is a compensated
metal.
The angular dependence o f the d H v A frequency in Lain3 is shown in fig. 22 (Umehara
et al. 1991b). The solid lines represent the results o f A P W band calculations (Kletowski
et al. 1987, Kitazawa et al. 1985). Branch a originates from the band 7 electron Fermi
surface, while the others originate from the band 6 hole Fermi surface, as shown in fig. 23.

42

Y. 0NUKI and A. HASEGAWA

10 8

-e
__c d _ ~ g
~

eo~o~o~eeeoo~o_

(D

! o ek o e

e eoo

10 7

e e

h~_'

~o
eoeeo

e,

e~ooOee

10 6

30
<ii0>

30
<I00>

60
<Iii>

Field Angle (Degrees)

-90
<ii0>

Fig. 22. Angular dependence of the dHvA frequency in Lain 3 (Umehara et al. 1991b). The solid
lines represent the results of band calculations
(Kletowski et al. 1987, Kitazawa et al. 1985,
Hasegawa 1982). The letters designate the various
orbits.

The latter Fermi surface consists of three kinds of major parts, which are centered at the
F, R and X points. Among them, a Fermi surface centered at F, denoted by d, is electron
in dispersion and is spherical in topology, bulges slightly along the (100) direction and
connects with another part of the Fermi surface centered at R by the slender arm elongated
along the (111) direction. The arm is denoted b y j . The topology of Fermi surface is
similar to that of Cu, although the volume is small compared to that of Cu.
Conduction electrons are mainly due to La 5d and In 5p electrons, as mentioned above.
The cyclotron masses for branches a and d for field along the (100) direction are 0.57m0
and 0.40m0, respectively, which are almost the same as the band masses of 0.58mo and
0.53m0, respectively.
dHvA measurements in CeIn3 were done by Kurosawa et al. and others (Kurosawa
et al. 1990, Satoh et al. 1992, Ebihara et al. 1993). The results are shown in fig. 24.
Many dHvA branches are observed in the range of 4x 105 Oe to 3 x 107 Oe, which are
highly different from those of Lain3 in the angular dependence.
Among them, a branch with the dHvA frequency of about 3x 107 Oe, denoted by d,
is similar to branch d in Lain3, although this branch is observed in the whole angle
region for CeIn3. Correspondingly, the arm orbit, denoted by j, is not observed in CeIn3.
The branch d Fermi surface in CeIn3 is thus roughly spherical, possessing no arms. The
angular dependence of the amplitude and the cyclotron mass for branch d is strange
(Ebihara et al. 1993). As shown in fig. 25, the dHvA amplitude for branch d in CeIn3

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

fg

43

k
.

J
(a)

(b)
F

Fig. 23. (a) Holeand (b) electronFermi

surfacesin Lain3 (Kletowskiet al. 1987,
Kitazawa et al. 1985, Hasegawa 1982).
The stippledregions indicatethe crosssections of the hole Fermi surfaces in
the 100 and 110 planes. The letters a,
b, ..., designatethe various orbits.

is strongly reduced in two angle regions where branch d in Lain3 is not observed by
existence of arms. This is mainly due to the large cyclotron mass in these regions; for
example, it is about 12m0 for field along the (110) direction, as shown in fig. 26, while
it is rather light in the angular regions around (100} and (111). The broken line in
fig. 25 shows the angular dependence of the dHvA amplitude under consideration of the
reduction factor [sinh(ctm2T/H)] -1, roughly in good agreement with the experimental
data. The strong angular dependence of the amplitude may be explained much better if
we furthermore consider the curvature factor of this closed Fermi surface.
The cyclotron masses 2.02m0 for field along (100) and 2.88m0 along (111) in Celn3
are five to eight times larger than the corresponding values 0.40rn0 and 0.37m0 in
Lain3, respectively. The detected largest mass is 20.7m0 for a branch with the frequency
of 7.6x 106 Oe. This may be compared to the electronic specific beat coefficients of
130mJ/K 2 tool in CeIn3 and 5.3-6.3 mJ/K 2 mol in Lain3.
Here we note that CeIn3 is antiferromagnetic at the usual applied field of 150 kOe.
Fermi surfaces in CeIn3 are thus strongly modified by the magnetic Brillouin zone
boundaries. A spherical electron Fermi surface in band 7, denoted by a in Lain3, is not
observed in CeIn3. This is most likely ascribed to the influence of the magnetic Brillouin
zone boundaries because the renormalized mass for this orbit is not extremely large,

44

Y. 0NUKI and A. HASEGAWA

108

{110} CeIn3

i0 7

10 6

o~ooOOO

30
<ii0>

oo

0
80
60
<i00>
<iii>
Field ~ g l e (Degrees)

90
<ii0>

Fig. 24. Angular dependence of the dHvA frequency in Celn3 (Kurosawa et al. 1990, Satoh
et al. 1992, Ebihara et al. 1993). The solid
lines connecting the data are guidelines. The
letter d designates the orbit.

about 10m0, following the relation m*(Celn3)= m*(Laln3)[7(Celn3)/7(Laln3)], where

mc(Laln3) is about 0.5m0 for branch a, y (Celn3) is 130mJ/K2mol and y (Lain3) is
5.3-6.3 mJ/K 2 tool.
The conduction electrons with large masses in the heavy-electron system are believed
to move slowly in the crystal compared to the usual carriers in the s and p electrons for
a similar Fermi surface topology and mean free path. This is one of the fundamental
understandings of the heavy-electron system. Ebihara et al. (1992b) experimentally
confirmed that the product of the cyclotron mass and the inverse scattering lifetime or
the Dingle temperature is constant by using two orbits of branch d in Celn3.
The relation between the effective mass m* and the scattering lifetime r for the
conduction electrons in metals is defined as follows;
hkF = m'or;

(63)

l = vvr,

(64)

and

where the wave number kv is half of the caliper dimension of the cross-sectional area S of
the Fermi surface, vF is the Fermi velocity and l is the mean free path. For example, the
equation S = ;rkv applies to a simple circular area.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

15

104

CeIn8
{100} } { 1 1 0 ~ ,

i0

10s

=o
o

10g

~D

0
~q

45

lo~

{100}I

"i'i10}

<ii0>

30

0
30
60
90
<i00>
<iii>
<ii0>
Field Angle (Degrees)

Fig. 26. Angular dependence ofthe cyclotron mass

for branch d in CeI~ (Ebihara et N. 1993).

10o

i0-1

Fig. 25. Angular dependenceof the dHvA amplitude

for branch d in Celn~ at 0.5 K and 129 kOe (Ebihara
et al. 1993). The solid line connectingthe data is a
0
30
60
90 guideline. The dashed line shows the angular depen<i00>
<111> <110> dence of the dHvA amplitude under considerationof
Field Angle (Degrees)
the dHvA-reductionfactor [sinh(am~T/H)]-~.

80

<ii0>

From eqs. (63) and (64) the following equation is obtained:

m*

hkF

(65)

Thus m*/r is expected to be the same for two similar orbits with the same mean free path.
When the effective mass m* is enhanced by a factor of (1 + ~) from spin fluctuations,
eq. (65) indicates that r should also be enhanced by a factor of (1 +,1,). This occurs
because a large mass is translated into a small velocity from eq. (64) so that the scattering
lifetime becomes (1 +,t,) times larger than that obtained by eq. (64) if the mean free path
is the same. In terms o f the Dingle temperature, eq. (65) becomes
h2kF

m*TD- 2JckBl"

(66)

The effective mass m* is assumed to be the cyclotron mass m e in the dHvA experiments.
Ebihara et al. (1992b) tested whether meT D is the same for similar orbits.
The above relation between the cyclotron mass and the Dingle temperature was applied
to the heavy electrons in CeIn3. There are two orbits for branch d which possess the same
dHvA frequency and therefore the same cross-sectional area of the Fermi surface but a
different cyclotron mass, as mentioned above. Thus, hkv is the same for the two orbits.
Moreover, the mean free path is considered to be the same because it is approximately
equal to the average distance between impurities.

46

Y. 0NUKI and A. HASEGAWA

+~
i

Celn8

102

orbit d
3.15x1070e
H=70kOe

x ~
~-.10

orbit

D ~o~

~2~44m0='

3.29x107Oe

H=129kOe
m~=12mo

71

0.5

1.0

.5

Temperature (K)
Fig. 27. Temperaturedependence of the dHvA amplitude for orbits d and D in CeIn3 (Ebiharaet al. 1992b).
The dHvA frequency and cyclotron mass for field along the (111) direction in
branch d are 3.15x 107 Oe and 2.44m0, respectively. The other orbit for field approximately along the (110) direction, which is now denoted by D, has 3.29x107 Oe and
12m0. Here the cyclotron masses were determined from the temperature dependence of
the dHvA amplitude A, as shown in fig. 27.
For these two Orbits, fig. 28 shows a Dingle plot of In
vs H -].
The slope in fig. 28 gives the product of the cyclotron mass and the Dingle temperature.
The slope for orbit d is almost the same as that for orbit D, indicating that the product
of m~TD is constant for the two orbits. Orbit d possesses TD = 0 . 1 9 K or l:=6.5x I0 -12 s,
while T D = 0 . 0 4 K or ~:=3.0x10 -ll s for orbit D. We can thus conclude that the larger
the mass becomes, the larger the scattering lifetime is such that the value of m c r - I or
m~ TD is constant.

[AHmsinh(am~T/H)]

+~

Celn3
orbit D -~g~.~^
o
3.29x1070e o% oo ~o~yg~,-~
12m0
0.SK
orbit

TD=0,04K

TD=0.19K

3.15 X107

'

2.44m0

0.5K
I

__

1.0
H-1 (xl0-5Oe-l)

Fig. 28. Dingle plot for orbits d and D in CeIn3 (Ebiharaet al. 1992b).

1.5

FERMI SURFACES OF INTERMETALLICCOMPOUNDS

{100}

{110}

47

Prln3

i08

id
o)
O

o~ i0 ~
o

gr~

N
J

I06

30
<Ii0>

0
<100>
Field

~gle

30

60

<111>
(Degrees)

9O

<110>

Fig. 29. Angular dependence of the dHvA frequency in Prln3 (Umeharaet al. 1991c). The solid
lines connectingthe data are guidelines. The letters
designate the various orbits.

Figure 29 shows the angular dependence of the dHvA frequency in PrIn3 (Umehara
et al. 1991c). The common dHvA branches are denoted by the letters defined for Lain3.
The Fermi surfaces of PrIn3, which has a singlet ground state, are similar to those of
Lain3, although each cyclotron mass of PrIn3 is twice as large as the corresponding one
for Lain3, as shown in table 8.
NdIn3 becomes antiferromagnetic below 5.9 K. The magnetization exhibits metamagnetic transitions at 76, 85 and 110kOe for field along the (100) direction, and at 73 and
90kOe for field along I110) (Czopnik et al. 1991). With increasing field, this compound
goes from an antiferromagnetic to a paramagnetic state, passing through an intermediate
magnetic state.
Umehara et al. (1992b) measured the dHvA effect in each magnetic state. The Fen'hi
surfaces of NdIn3 in the paramagnetic (or field-induced ferromagnetic) state, which
are obtained at higher fields than l l 0 k O e , are similar to those of Lain3, as shown in
fig. 30a. Branches a and d are characteristic Fermi surfaces in bands 7 and 6 of Lain3,
respectively. The main branch a consists of two closely separated frequencies, reflecting
the up and down spin states of the electrons. We note that it is difficult in general to
detect dHvA frequencies lower than 107 Oe at high fields.
However, in the antiferromagnetic state, there are many differences between the Fermi
surfaces of NdIn3 and those of Lain3, as shown in fig. 30b. This is due to the magnetic
Brillouin zone boundaries. This situation is similar to that of NdB6.

48

Y. 0NUKI and A. HASEGAWA

Table 8
dHvA frequencies and cyclotron massesa in RIn3 compounds for field along the (100) direction (Umehara
et al. 1991b,c, 1992a,b, Satoh et al. 1992, Ebiharaet al. 1992a)
RIn3
Lain3

m~

7.47
6.07

0.57
0.69

Celn3

Prln3

7.57

0.99

Ndln3 (paramagnetic)
Ndln~
(anfiferromagnetic)

7.75
8.09

1.14
1.05

Stain3

7.81

1.30~

Gdln3

6.63

0.61

a F values in 107 Oe, m~ expressed in m0.

b Field along the (111) direction.

F
2.78
0.330
2.93
3.30~

jb

m~

m;

0.40
0.49
2.02
12~

1.38

0.26

3.02
2.57
2.67
3.17
1.70

0.60
0.87
1.39
0.80
0.69

3.08
1.10
3.07
1.64
1.11

0.80
0.80
0.49
0.52
0.31

0.76c
0.079c
0.051 c
1.28
0.27
1.20
1.29
0.49
0.26
0.22
0.19
0.95
0.857
0.261
0.193

20.7 c
4.8 c
1.5 c
0.36
0.68
0.91
0.60
0.60
0.42
0.30
0.31
0.24
0.32
0.17
0.16

m~

0.096

0.17

0.128

0.28

0.12

0.31

0.096

0.29

c Field along the (110) direction.

According to the results of neutron experiments (Lethuillier et al. 1973), the magnetic
unit cell possesses a tetragonal structure which contains double chemical unit cells.
Therefore, the volume of the magnetic Brillouin zone becomes half the chemical one.
A m o n g the observed dHvA branches, the branch d Fermi surface remains in the
antiferromagnetic state because it is fully contained in the tetragonal magnetic Brillouin
zone.
Branch a is also observed in the antiferromagnetic state, although the dHvA amplitude
is strongly damped when the field is tilted by about 10 from (100) to (110 / in the
{100} and {110} planes, as shown in fig. 31. Umehara et al. (1992b) constructed the
Fermi surface such that a nearly spherical Fermi surface in the paramagnefic state, which
corresponds to the band 7 electron Fermi surface of Lain3, is changed into a multiply
connected Fermi surface with necks, as shown in fig. 32. Note that no necks exist along
the [001] and [00i] directions in the cubic notation. Therefore, cubic symmetry is broken
in the antiferromagnetic state of Ndln3. This new Fermi surface in the antiferromagnetic
state favors an open orbit along the [110] direction when field is directed along the

49

FERMI SURFACES OF INTERMETALLICCOMPOUNDS

(a) P a r a m a g n e t i c

{100}

(b) Antiferromagnetic

State
i

NdI~3

{110}

State

,
{i00}

{110}

NdIn3

10 8

I0 s

d
o

o o

o~

107

o~

i0 7

I0 ~
I

30

<ii0>

30

60
<i00)
<ii]>
Field Angle (Degrees)

10 6

90
<110>

30

<iI0~

0
30
60
<I00>
<iii>
Field Angle (Degrees)

90
<110>

Fig. 30. Angular dependence of the dHvA frequency in (a) paramagneticand (b) antiferromagneticstates for
Ndln3 (Umehara et al. 1992b). The solid lines connecting the data are guidelines. The letters designate the
various orbits.
[110] direction, which is confirmed in magnetoresistance experiments (Umehara et al.
1992b).
I f we follow the Fermi surface as shown in fig. 32, the branch a oscillation is not
expected to be detectable for field along (110}; but it was experimentally observed
even in the antiferromagnetic and intermediate magnetic states, although its amplitude
is extremely small. This is because an electron can circulate around a closed orbit by
breaking through the neck-Fermi surfaces. These necks, which are introduced by magnetic
ordering, are broken through by the circulating electron in rather low field of 50 kOe.
The antiferromagnetic Fermi surface is thus reconstructed on the basis of the magnetic
Brillouin zone. The same procedure mentioned above should be applied to the band 6
hole Fermi surfaces, probably producing small closed Fermi surfaces. They may possibly
correspond to dHvA branches with frequencies of 1.5x 107 Oe and 5 106 Oe as shown
in fig. 30b.
The cyclotron mass of NdIn3 is twice that of Lain3. Differences of the cyclotron mass
are small between the antiferromagnetic and paramagnetic states, as shown in table 8.
For Stain3, a large number of small spherical Fermi surfaces were also observed, which
are not present in Lain3 (Ebihara et al. 1992a), as shown in fig. 33. Branches a and
d in fig. 33 are similar to those of Lain3 in the magnitude of dHvA frequency. However,

50

Y. 0NUKI and A. HASEGAWA

'

'

T - - ~

Io
30 -

NdIna
{1oo}

o I

63-78kOe

I {110}

(a)

0.SK
20

i(

:>

[ioo]

10

t
: . ~

.~ .= . q

.= z

. :

il?: 5 "
2
: 2
i l i ? i ~ ' ; 2 l ? ! ( b
2!?2
~illl

"q'--r-l

~'~'s

80
<110>

'~

i|f~r

0
<I00>

Field Angle (Degrees)

o n

30

~
iii.3113.

--[010]

,[i~1:11

Fig. 31. Angular dependence of the dHvA amplitude Fig. 32. (a) Proposed electron Fermi surface for
for branch a in the antiferromagnetie state of NdIn 3 band 7; (b) cross-sections of this Fermi surface in
(Umehara et al. 1992b). The solid and dotted lines the R - M - X plane in the antiferromagnetic state of
connecting the data are guidelines.
NdIn 3 (Umehara et al. 1992b).

branch a for Smln3 is observed only around the (110) direction. The Fermi surfaces
are probably modified in topology by the magnetic Brillouin zone boundaries. If we
consider the spherical Fermi surface with arms or bumps along the [100], [100], [010] and
[050] directions, the electron cannot circulate around (100) but can circulate around both
(110) and (111). However, no signal is observed experimentally around (111>. Probably,
the (111 ) direction is less detectable compared to (110), which originates from strong
damping of the signal due to the curvature factor of the Fermi surface with arms or
bumps.
Branch d is also supposed to correspond to branch d in Lain3, although the arms,
denoted by j, are not present in Smln3. Its Fermi surface is probably nearly spherical as
in Celn3, although the dHvA signal is not observed over the whole angular region.
The cyclotron masses of Smln3 are also roughly two or three times larger than those
of Lain3, as shown in table 8.
Branches d and j are observed in the antiferromagnetic state of Gdln3, as shown in
fig. 34 (Umehara et al. 1992a). This Fermi surface is probably contained in the magnetic
Brillouin zone. As the magnetic structures of SmIn3 and Gdln3 are unknown, further
analyses are not yet done.

51

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

1

'

'

{ii0}

{100}

'

I ~

'

{zoo}

Smln3

'

{ii0}

@dln3

10 8

i08

o2

0)
0

oo

C)

o~o
o o_~o_~o-o-~

i0 7

107
- o ~ - -~-~-

0..~.~.~

~.~__~

!~

o~g~a~-a&'o-g-o~

4 - 8 4

I o04-/~.~- o
o~

106

10 6
I

30
<ii0>

0
30
60
<i00>
<iii>
F i e l d Angle (Degrees)

0
<ii0>

Fig. 33. Angular dependence of the dHvA frequency

in SmIn~ (Ebihara et al. 1992a). The solid lines
connecting the data are guidelines. The letters
designate the various orbits.

30

<ii0>

30

60

<i00>
<iii>
Field Angle (Degrees)

90

<ii0>

Fig. 34. Angular dependence of the dHvA frequency

in GdIn 3 (Umehara et al. 1992a). The solid lines
connecting the data are guidelines. The letters
designate the various orbits.

The cyclotron masses for branches d and j in Gdln3 are almost the same as those of
Lain3. We note that the dHvA branch with the frequency of 7.28 x 107 Oe in Gdln3, which
is observed in a narrow region around ( 111 >, is the same frequency as branch a in Lain3.
This branch possesses a relatively large mass of 1.70m0, which is three to four times
larger than that of Lain3, 0.47m0 if it corresponds to branch a.
In powdered YbIn3, one dHvA branch of 4.50x 106 Oe is detected at high fields up to
30 T, suggesting that there exists a small spherical portion of the Fermi surface (Meyer
et al. 1973).
The magnetoresistances of RIn3 (La, Ce and Nd) increase over a wide field region
(Umehara et al. 1991b, 1992b, Satoh et al. 1992, Kurosawa et al. 1990). This result
suggests that these compounds are compensated metals with an equal carrier number of
electrons and holes. This is expected simply from the total number of valence electrons, 12
per primitive cell because of R 3+ and In3+. However, the magnetoresistances in CeIn3 and
Ndln3 saturate for fields along the characteristic field directions, indicating the existence
of open orbits in these Fermi surfaces. For example, the multiply connected Fermi
surface with necks in the antiferromagnetic state of NdIn3 favors an open orbit along

52

Y. ONUKI and A. HASEGAWA

the (110) direction, as mentioned above. These behaviors are compared to those of Lain3
where the magnetoresistance increases over the whole angular region, indicating no open
orbits. A different behavior of the magnetoresistance is observed in Gdln3, where the magnetoresistance saturates in all field directions in the {100} plane (Kletowski et al. 1985).
For Gdln3, it is necessary to consider the Fermi surface in the antiferromagnetic state.

4.1.4. RSn3
CeSna at low temperature is thought to belong to the so-called valence-fluctuation regime
with a Kondo temperature of about 200 K. Therefore, it is interesting to study the alloy
system Ce(Inl_xLax)3, clarifying how the Kondo regime changes to the valence-fluctuating
regime. Actually, an experimental work was reported (Benoit et al. 1985). There are some
mysteries hidden in CeSn3 itself (Gschneidner et al. 1985). Some anomalies are attributed
to off-stoichiometry effects and/or to the inclusion of other phases such as Ce2Sns, but
many of the anomalies are not yet well understood. The characteristic properties of the
RSn3 compounds are summarized in table 9.
Figure 35 shows the angular dependence of the dHvA frequency in LaSn3 (Settai
et al. 1995, Umehara et al. 1991a, Johanson et al. 1981, 1983, Boulet et al. 1982). Main
branches are two branches denoted by a and/3 which exist in the whole range of angles
in the low frequency region, and three branches with high frequencies denoted by 71, Y2,
and 73 centered at (111).
Compared to Lain3, LaSn3 has three electrons more per primitive cell, and hence the
electronic structure of LaSn3 is that of an uncompensated metal having a Fermi surface
which occupies half of the Brillouin zone (Hasegawa and Yamagami 1991, Hasegawa
1981, Koelling 1982). All sheets of the Fermi surface in Lain3 disappear in LaSn3.
In LaSn3, band 8 yields a main sheet of the Fermi surface which looks like a composite
sheet, and two small hole pockets lie at the F point in bands 7 and 8, as shown in fig. 36.
The solid lines in fig. 35 are the results of band calculations done by Hasegawa and
Yamagami(1991). A strong resemblance between the experimental and the theoretical
angular dependences of the dHvA branches suggests that the experimental branches a and
/3 may originate from the small hole Fermi surfaces in bands 7 and 8, respectively.
Hasegawa and Yamagami (1991) revealed that the composite sheet in band 8 actually
consists of a large distorted hole sphere centered at the R point and a network which
Table 9
Characteristic propertiesof the R S n 3 compoundswith the A u C u 3-type cubic crystal structure
RSn3

TN (K)

LaSh3
CeSn3

Pauli para
valence
fluctuation

ne=nh

nh

7
Fermi surface
(mJ/K2tool)
11
53

distortedspherewith necks; and network(hole, band 8)

distortedsphere(hole,band 8); and complicatedelectron
Fermi surface (band 9)

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

'

fl {loo}
[

'

{110}

(e)

LaSh3 1

6"

7,~

,c~

107

,/
:~,~,~<~

30

<II0>

8)

(band 7)

O~
~ ~ , , ~>o ~ ~ ~

)AAj

10 6 I

ao

h o l e (band

(a) h o l e

10 8

LaSna

53

0 J

<i00>

30

~i

A~

~,

60
<111>

A
i

(b) h o l e
(band 8)

90
<110>

Fig. 35. Angular dependence of the dHvA frequency

in LaSh3 shown by circles (Umehara et al. 1991a) and
triangles (Settai et al. 1995). The solid lines show the
results of band calculations (Hasegawa and Yamagami
1991). The Greek letters designate the various orbits.

P
(d) h o l e (band 8)
Fig. 36. Fermi surfaces of LaSn3. (a) Small hole sphere centered at the F point in band 7; (b) small distorted
hole sphere centered at the F point in band 8. The large hole surface in band 8 is shown in two separate parts:
(c) large distorted sphere; (d) network (Hasegawa and Yamagami 1991). The Greek letters designate the various
orbits.

surrounds the distorted hole sphere with many slender arms, as shown separately in
figs. 36c, d, respectively. The distorted hole sphere contains about 0.6 holes per primitive
cell and is connected with the network near the M point. Each connection part consists
of a set of four thin necks. Complex structures of band 8 Fermi surface produce many
dHvA branches, as shown in fig. 36c, d. The detected dHvA branches are thus well
explained by the present band calculations.
Figure 37 shows the angular dependence of the dHvA frequency in CeSn3 (Umehara
et al. 1990, Johanson et al. 1983). Some Fermi surface parts seem to be similar but other
parts are considerably different from those of LaSn3, and the detected cyclotron masses
are roughly five times larger than those in LaSn3, in agreement with the y values, as
shown in tables 10 and 11.

54

Y. ONUKI and A. HASEGAWA

CeSn3
CeSn3

{loo}

{11o}

F
F

10 8

Ib

~ ~ e

:f

(a)

107

F
P

<

F
106

R
R

~
105

poo
i

30
(110>

0
(100>
Field ~gle

60
(111>
(Degrees)

30

(b)

90
(110>

Fig. 37. Angular dependence of the dHvA frequency in

CeSn3 (Umehara et al. 1990). The solid lines connecting
the data are guidelines. The letters designate the various
orbits.

Fig. 38. (a) Large hole Fermi surface in band 8; (b) large electron Fermi surface in band 9 of CeSn3 (Hasegawa
et al. 1990). The letters designate the various orbits.

Table 10
Cyclotron masses and band masses in LaSn 3 (Umehara et al. 1991a, Hasegawa and Yamagami 1991) a

mass
m~ (too)
mb (m0)

m*dmu

a
(111)

/3
(111)

71
(111)

72
(111)

73
(111)

6
(110)

e
(27 )

~
(100)

p
(110)

0,38
0.16
2.4

0.74
0.42
1.8

1.07
0.67
1.6

0.91
0.42
2.2

1.91
0.67
2.9

1.16
1.2
1.0

2.33
1.6
1.5

0.78
0.69
1.1

0.45
0.40
1.1

a Note that the g value of 11 mJ/K 2 mol is roughly twice as large as the theoretical Yb value of 5.56 mJ/K2 tool.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

55

Table 11
Cyclotronmasses m~ and their band masses mb in CeSn3 (Umeharaet ai. 1990, Hasegawaet al. 1990). Note
that the 7 value of 53 mJ/K2tool is roughlythree times larger than the theoreticalvalueof 15.4mJ/K2mol
Mass
(100)
m~0)
mb~0)
m~ b

2.1

a
(110)

(111)

4.15
1.7
2.4

3.83
0.9
4.3

b (100)

c (110)

d (111)

g (110)

2.72
1.7
1.6

6.30
2.3
2.7

6.25
2.3
2.7

3.6

1.4

4.07
2.1
1.9

Hasegawa et al. (1990) calculated the energy band structure of CeSn3 under the
assumption that the 4f electron is itinerant. As shown in fig. 38, the calculated Fermi
surfaces mainly consist of a large band 8 hole Fermi surface centered at R and a large
band 9 electron surface centered at F. The origin of the analogy in Fermi surfaces between
CeSn3 and LaSn3 is as follows. Namely, the large distorted spherical hole Fermi surface of
band 8 in LaSn3 is similar to the band 8 hole Fermi surface in CeSn3. Small spherical hole
Fermi surfaces in bands 7 and 8, denoted by a and/3 in LaSh3, are also present in CeSn3,
corresponding to branches p and q, respectively in fig. 37, although their volumes of the
Fermi surfaces in CeSn3 are smaller than those of LaSn3. The difference between LaSn3
and CeSn3 is ascribed to the large electron Fermi surface in CeSn3. Note that this Fermi
surface has no occupied states along the (111) direction, which is different from the results
of band calculations done by Koelling (1982) and the renormalized band calculations done
by Strange and Newns (1986). Later, Norman and Koelling (1993) confirmed the validity
of the electron Fermi surface calculated by Hasegawa and coworkers.
The solid lines in fig. 39 indicate the results of the band calculations done by Hasegawa
et al. (1990). Branch a corresponds to the large hole Fermi surface, while the other main
branches are due to the electron one. Both in the magnitude of the cross-sectional area
and in the observed range of angle, the theoretical branches agree reasonably well with
those of the experimental ones, supporting the validity of both the large hole and the large
electron Fermi surfaces predicted by the calculations.
Here we note that branch a in fig. 39 disappears at angles in the vicinity of the
(100) direction. The reason for the disappearance of branch a is due to the combined
effect of the curvature factor and cyclotron mass, as shown in fig. 40. The hole Fermi
surface is not a perfect sphere, but bulges appreciably towards the (100) direction. In the
(111 ) direction, the orbit does not pass on any bulges. Therefore, as shown by the broken
line in fig. 40, 102S/Ok2 [-,/2 has a sharp peak along ( I l l ) and becomes small along (100)
and (110). The minimum value is about 1/20 of the maximum. Another possible origin is
an increase of the cyclotron mass m* for branch a, because it becomes a maximum along
(100) and a minimum along (111). Here we note that the probability amplitude of the
4f state averaged over the orbits perpendicular to the (100), (110) and (111) directions
is found to be 70%, 65% and 60%, respectively. The intensity of the oscillation of the de
Haas-van Alphen effect is thus damped by the factor [sinh(am2 T/H)] -1. The solid line

56

Y. ONUKI and A. HASEGAWA

8

{I00}

'

'

CeSn3

r~
4~

{I00}

108 ~

{ii0}

;J
I
I

'

II

hi

/',

4~

r~.

LI

'

<

30

<ii0>
i07

30
60
<i00>
<iii>
Field Angle (Degrees)

90
<ii0>

Fig. 39. Angular dependence of the high dHvA

frequency in CeSn3 (Umehara et al. 1990). The solid
lines show the results of band calculations (Hasegawa
et al. 1990). The letters designate the various orbits,
'

40

'

'

'

'

'

30

60

90

<i00>
<iii>
Field Angle (Degrees)

I ,

<Ii0>

~\

<ii0>

Fig. 40. Angulardependenceof the dHvA amplitude

for branch a (Hasegawa et al. 1990). The dashed curve
shows the curvature factor 102S/02kr, I-'/2, and the
solid curve includes another factor [sinh(a m*eT/H)] -I"
[02S/Ok~l-'/2. Data are for T - 0 . 5 K and H = 150kOe.
The circles are the experimental results (Umehara
et al. 1990 ).

'

LaSn3
J//<ll0>

40

(b)

I-

I I ~ I E I I I

LaSn3
J//<100 >
0.5K

30

3O

Q.

<L

<3

<L

<3

20

2O

i0

10

30

0
0
30
60
Field Angle (Degraes)

90

HH4.5

~ ~ H f f <

100,

50
I00
Magnetic Field (kOe)

150

Fig. 41. (a) Angular and (b) field dependencies of the magnetoresistance in LaSn3 (Umehara et al. 1991a).
Open orbits occur in the stippled region in (a). The solid lines connecting the data are guidelines.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

57

in fig. 40 shows the angular dependence of the curvature factor multiplied by this factor
which is calculated for branch a by assuming that T = 0 . 5 K and H = 150kOe. We see
that this combined damping factor qualitatively explains the experimental result for the
intensity of branch a.
The large hole Fermi surface in band 8 contains 0.44 holes per primitive cell, and the
large electron one in band 9 contains the compensating number of electrons. The number
of holes and electrons contained in other small Fermi surfaces is fairly small, in total
about 0.1 per cell. Every carrier possesses larger masses than the corresponding band
masses. The ratio of the cyclotron mass me to the band mass m2/mbfor each branch is in
the range of 2 to 4, as shown in table 11. This is consistent with the ratio of y to Yb, 3.4.
The magnetoresistance is also useful to determine the character of the 4f electrons
in CeSn3. Figure 41 shows the angular and field dependences of the magnetoresistance
in LaSna (Umehara et al. 1991a). The sharp dips and spikes around the (100) direction
indicate that the (100) direction is a singular field direction and the spikes mean open
orbits, resulting from the uncompensated nature of LaSh3. This is simply expected from
the total number of valence electrons, 15 for La 3+ and Sn4+. These open orbits are
explained well by the results of band calculations done by Hasegawa and Yamagami
(1991). Figure 42 is a typical example of the (100) open orbits running on the large Fermi
surface in band 8. Magnetic field is at 4.5 from the (100) direction on the { 100} plane.
On the other hand, the magnetoresistance of CeSn3 in fig. 43 increases in the all field
directions (Umehara et al. 1990), indicating compensated carriers and the absence of
open orbits in CeSn3, which may be compared to the uncompensated state of LaSn3.
These results indicate that Ce is tetravalent, Ce4+, confirming the itinerant nature of the
4f electron.
The results of magnetoresistances for LaSn3 and CeSn3 are consistent with the
results of the band calculations mentioned above (Hasegawa and Yamagami 1991,
Hasegawa et al. 1990). The multiply connected band 8 Fermi surface in LaSn3 favors
the experimentally observed open orbits. All of the Fermi surfaces in CeSn3 are closed
ones, in good agreement with the compensated metal of CeSn3 without open orbits.
It is remarkable that the itinerant 4fband model fits the dHvA results so well. From the
viewpoint of the energy band structure we summarize the f-electron behaviors of LaSn3
and CeSn3. In LaSn3, band 8 forms a large distorted hole sphere around the R point,
and each of bands 7 and 8 forms a small hole sheet at the F point. These characteristic
features of the energy band structure in the vicinity of the Fermi level (EF) are shown
schematically in figs. 44a, b. The center of the unoccupied 4f bands is located at about
0.2 Ryd. above EF. The energy band structure around the R point in band 8 is similar
to the nearly-free-electron band and the Fermi surface of band 8 is quite insensitive to
the location of the 4f bands (E4f). In contrast, however, since the Bloch states at the
F point in bands 7, 8 and 9, denoted as F 7 and F~, consist dominantly of the Sn 5p states
and have the same symmetry as in the 4f bands, they can hybridize strongly with the
4f states. Hence, the position of their energy levels depends sensitively on the location
of the 4f bands. The size of the small hole sheets is apt to vary easily by a change of the
location of the 4f bands.

58

Y. O N U K I and A. HASEGAWA

LaSn3

50

I,

It i''' ' '

, , ,-

I J//<lO0>

40I'l5K
,

150k0e

(
20

30
<Ii0>

0
<i00>

10
,.-

~t

t-0.07~.0.997,01

open orbits

Fig. 42. (100) open orbits running on the large hole

Fermi surface in band 8 of LaSn3 (Hasegawa and
Yarnagami 1991). Magnetic field is applied at an angle
of 4.5 from the (100) direction on the {100} plane.
The center of this figure is located at (0.6, 0.5,
0.5)(2n:/a). The hatched regions show the unoccupied
electron states.

0
<i00>

30

60
90
<iii> <ii0>

Field Angle (Degrees)

Fig. 43. Angular dependence of the magnetoresistance
in CeSn3 (Umehara et al. 1990). The solid lines
connecting the data are guidelines.

(a)

(b)
E4f

E4f

./F~-

r s,r--

./- F7- F8-

-.,%%

-1
EF

'1

:'q-~ 7
8 --t a

Fig. 44. Schematic representations of the relationship between the energy band structures (bands 7, 8, 9) for
LaSn3 and CeSn 3. Solid curves show the energy bands, and solid lines show the Fermi level (EF) and the center
of the f bands (E4r) for LaSn 3, and dashed curves and lines show those for CeSn3. (a) Energy band scheme
around the R point; (b) energy band scheme around the F point.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

59

Based on the itinerant-electron model for the 4f electrons, CeSn3 has one more Bloch
electron per primitive cell. As electrons are added to the 4f bands, the 4f bands shift
downwards and the Fermi level rises. In contrast to a small change in the band 8 large
hole sphere shown in fig. 44a, bands 7, 8 and 9 around the F point are modified drastically,
as shown in fig. 44b. The F~ state shifts down to just above EF, and the two small hole
surfaces diminish considerably. Band 9 around the F point also sinks so deeply, i.e., the
I'~ state shifts toward the same position of the F 8 state, that it can form an electron
sheet around the F point which is dominant enough to compensate the large hole sphere
around the R point. Only a slight rising of the Fermi level is necessary. Therefore, the
large hole sheet contracts a little, and is separated completely from the network. The
energy band structure for CeSn3 thus obtained is shown by broken curves in figs. 44a, b.
This is a simple explanation for the fact that CeSn3 has an isolated, distorted hole sphere
in band 8 which looks quite similar to the distorted hole sphere in band 8 in LaSn3. In
spite of the resemblance between these hole sheets, the Bloch states on the large hole
sheet in CeSn3 have a large probability of 4f components, about 60% on the average.
A distinct difference between LaSn3 and CeSn3 is the existence of a large electron sheet
in band 9 in CeSn3 which appears to compensate the large hole sheet in band 8.
4.2. Fermi surfaces in the bcc Brillouin zone
4.2.1. RX
Rare earth monopnictides RX possess the NaC1 type crystal structure. The primitive cell
contains one molecule of RX. Figure 45 shows the crystal structure and its Brillouin zone.
Especially the CeX compounds, in which X stands for N, P, As, Sb and Bi, have attracted
a particular interest because of various anomalous magnetic and transport properties
(Kasuya et al. 1987). LaX is expected to be a semimetal with a small and equal number
of electrons and holes. A similar semimetallic character is observed in CeX, except
CeN. CeX is a Kondo-latfice compound, having anomalous magnetic properties at low
temperatures. Even though the carrier number is small, the Kondo effect is strong and the
kz

t~

(a)

A ~

(b)

Fig. 45. (a) NaC1 type cubic crystal structure of the RX compounds. Spheres with and without pattern show
the X and R atoms, respectively; (b) Brillouin zone of the face-centered cubic crystal lattice.

Y. ONUKI and A. HASEGAWA

60

Table 12
Characteristic properties of the RX compounds
Compound

TN (K)

LaSb

Pauli p a r a

no=nh

0.80

CeSb

16.5
Kondo lattice
singlet
2.1
27
Pauli para
25
Kondo lattice
7-8
0.5

n~= nh

20

PrSb
SmSb
GdSb
LaBi
CeBi
CeAs
YbAs

7
Fermi Surface
(mJ/K2mol)
three ellipsoidal electron Fermi surfaces (band 4) and
compensated two spherical hole Fermi surfaces
(band 2 and 3). n~=0.014mo1-1
similar to LaSb but spin-split

no= nh
n~= n~
no= nh
ne = nh
n~= nh

0.95
20

similar to LaSb
similar to LaSb
similar to LaSb
similar to LaSb
similar to LaBi

ne = n~
n~= n~

200

similar to LaSb, but n~=0.001 mo1-1

similar to LaSb

), value is large, about 20 mJ/K 2 mol for both CeBi and CeSb. The characteristic properties
o f RX compounds are summarized in table 12.
The angular dependence of the dHvA frequency was clearly observed in LaSb (Settai
et al. 1993, Kasuya et al. 1987, Kitazawa et al. 1983). It is expected from a simple
consideration o f the valencies of constituent atoms that LaSb may be an insulator,
but in reality there is a slight overlapping between the Sb 5p valence bands and the
La 5d conduction bands. Thus LaSb becomes a semimetal. Although a self-consistent
relativistic calculation in LDA actually yielded a good fit with the experimental results, it
was found that a better fit can be achieved by shifting the La 5d bands and the unoccupied
4 f b a n d s upwards relative to the Sb 5p valence bands (Hasegawa 1985). It was also found
that a correct Fermi surface can be obtained for LaSb by a non-self-consistent calculation
with a full Slater exchange potential. Results o f band calculations thus obtained are shown
by solid lines in fig. 46, which are compared with the experimental dHvA results (Settai
et al. 1993).
In later calculations due to an LMTO method performed by Sakai et al. (1985) and
Kasuya et al. (1987), the relative positions o f the Sn 5p bands and the La 5d bands were
adjusted artificially so as to reach a proper overlap between them, and the 4 f bands were
shifted upward to a level suggested by a BIS experiment. The results of their calculations
are the same as those shown in fig. 46. There is, however, an obvious shortcoming which
is inherent in LDA; the overlap between the valence and conduction bands is too deep.
Such a shortcoming manifests itself most seriously in small Fermi surfaces of semimetals.
Most o f this may be removed by taking the self-interaction correction into account.
The Fermi surface consists of three kinds o f sheet (Hasegawa 1985), as shown in fig. 47.
Band 2 forms a small and nearly spherical hole Fermi surface centered at the F point,

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

151

II

,-

{100}

61

LaSb

[110}

o, oo

o)
5q

i --l--- i

30

30

<11o)

<100)
Field

I
60

<111)

90

<110)

Angle (Degrees)

(b)

Fig. 46. Angular dependence of the dHvA frequency

in LaSb (Settaiet al. 1993).The solidlines showthe

results of band calculations (Hasegawa 1985). The
Greek letters designate the various orbits.

Fig. 47. Fermi surfaces of LaSb (Hasegawa 1985):

(a) hole Fermi surface in band 2, centered at F;
(b) hole Fermi surface in band 3, centered at F and
the three electron Fermi surfaces in band 4, centered
atX.

denoted by ft. Band 3 also forms a hole Fermi surface centered at the F point, denoted
by fiq It is slightly stretched into the (100) direction. On the other hand, band 4 consists
of three equivalent, nearly ellipsoidal electron Fermi surfaces centered at the X point,
denoted by ai (i = 1, 2 and 3).
LaSb is thus a semimetal. Namely, the top of the valence band with the dominant
Sb 5p character is located at F. The bottom of the conduction band with mostly
La 5d character is located at X. They slightly overlap. The narrow 4f bands lie a few
eV above the Fermi level, and thus the mixing of the 4f states into the valence band
states is not large. The cyclotron masses are thus small; 0.14m0 for branch a3, 0.15m0
for branch fl and 0.49m0 for branch fl~ for field along (100) direction (Settai et al. 1993).
LaBi is also a semimetal. The Fermi surface is quite similar to that of LaSb
(Kitazawa 1982). The cyclotron masses are in the range of 0.18m0 to 0.36m0. The
experimental electronic specific heat coefficients of LaSb and LaBi are 0.80 and
0.95 mJ/K 2 tool, respectively. This is in good agreement with the calculated values, 0.50
and 0.85 mJ/K 2 mol (Hasegawa 1985, Sakai et al. 1985).

62

Y. 0NUKI and A, HASEGAWA

70

II

CeSb
(o10)
r/3

6O

<
>

~D
0

F(~

?2 50
g

36

30
[101]

IL

S~t~)

30

30
[001]

II

42

Fig. 49. dHvA oscillation on the high- and lowfield sides of the magnetic transition from the S'
to F phases in CeSb (Crabtreeet al. 1987).

40
v

38
40
Magnetic Field (kOe)

60
[111]

Field ~gle (Degrees)

90
[110]

Fig. 48. Angular dependenceof the transition field

betweenthe S' and F phases in CeSb (Crabtreeet al.
1987).

CeSb and CeBi, which show anomalously large magnetic anisotropy and a very
complex antiferromagnetic phase diagram called "devil's staircase" (Rossat-Mignod et al.
1985), belong to a Kondo-lattice system with a small carrier number. Even though the
cubic 1"7 state isthe ground state in the paramagnetic state, anisotropy in the ordered state
is very strong.
CeSb exhibits at least 15 magnetic phases as a function of temperature and
magnetic field. The low-temperature phases consist of stacks of ferromagnetically aligned
(001) planes with magnetization parallel (T) or antiparallel (+) to field along the
[001] direction; S2(TT$+Ti"+), S'(TT$) and F(Ti"T). The ordered state thus shows an Isinglike character with ferromagnetic ordering within (001) or z-plane, while the nearly full
moment of Jz = 5/2 is oriented along the [001] or z-axis with 4f(Fs) character. These
interesting properties are explained by the p - f mixing model (Takahashi and Kasuya
1985). Because of the strong p - f mixing between the 4f(F8) and Sb 5p states at the
F point, the 4f(Fs) state becomes more stable than the 4f(F7) state at low temperatures
in the ferromagnetic ordered state. This feature is strongly enhanced by the non-linear
effect due to the small Fermi energy. Here the 4f(F7) state mixes rather weakly with the
electrons in the Ce 5d conduction band.
The strong magnetic anisotropy leads to domain effects in the presence of a field. In
the absence of fields, the magnetic moment has an equal probability of pointing along
any of the three crystallographically equivalent (100) directions. In field the magnetic
energy is lowered if the moment chooses to point along the <100/ direction closest to the
applied field direction, resulting in a single domain sample.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

63

Figure 48 shows the transition field between S I and F phases as a function of angle in
the (010) and (510) planes (Crabtree et al. 1987). Domain effects are evident in the sharp
cusp in the transition field occurring at [111]. For field directions between [001] and [111]
in the (510) plane and between [001] and [101] in the (010) plane, the favored moment
direction is [001]. However for directions between [111] and [110] in the (110) plane, the
favored direction switches to either [010] or [100]. The cusp at [l 11] reflects this switch
in the magnetic domain. The strong magnetic anisotropy can also be seen in the shape
of the transition field. The solid line in fig. 48 is the form Hc/cosO where 0 is the angle
between the field and the closest (100) type direction. The excellent fit of this curve to the
transition field implies that for an arbitrary direction the transition occurs whenever the
component of field along the closest (100) direction equals Ho. This behavior is expected
if the moment is pinned along the closest (100) direction.
Figure 49 shows the dHvA oscillation on the high and low-field sides of the magnetic
transition from S r to F (Crabtree et al. 1987). The Fermi surface changes significantly
between these two phases. The low-dHvA frequency oscillations with large amplitude
in the high-field region are abruptly replaced by the higher-dHvA frequency oscillations
with smaller amplitude in the low-field region.
dHvA studies in the F phase were reported first by Kitazawa et al. (1983); more
complete studies in the two phases o f S ~ and F phases were done by Aoki and coworkers
(Aoki et al. 1985, 199l, Crabtree et al. 1987). Figures 50 and 51 show the angular
dependence of the dHvA frequency in the S ~ and F phases, respectively.
In fig. 51, the solid lines show the results of band calculations done by Kasuya and
coworkers (Kasuya et al. 1987, Sakai et al. 1985). The band structure calculations were
carried out for a ferromagnetic state by an LMTO method. To represent the Sb 5p bands
and the Ce 5d bands, potential parameters in the LMTO method were determined so as
to reproduce the Fermi surface of LaSb properly. The CeSb 4f state with J z = 5/2 was
located at 1 eV below the Fermi level so as to fit the photoemission spectra (Hillebrecht
et al. 1985). One of the 4f electrons is localized at this level. Hence, there are no
adjustable parameters in calculations of the Fermi surface for CeSb. The 4f electrons
cause ferromagnetism, and affect the Fermi surface, especially the small hole sheets at the
F point through the 5p-4f hybridization effect. As shown fig. 51, the agreement between
the experimental data and the theoretical results is surprisingly good, supporting validity
of the p - f mixing model. Due to ferromagnetic ordering, two hole bands split into four
different hole bands, named fit,/32,/33 and/34. On the other hand, the electron bands at X
are split by the 5d-4f intra-atomic exchange interaction. The different sizes of the three
electron ellipsoids explain the separation in the dHvA branches ai, as shown in fig. 52
(Crabtree et al. 1987). The smaller ellipsoid whose axis is along the moment direction of
the [001] axis gives rise to branch a3, while the larger ellipsoids produce branches a1 and
a2. The cusp in branch ai at [111] is due to the domain switching effect. As mentioned
above, when field is directed from [111] to [110] in (110), the favored magnetic moment
switches from [001] to either [010] or [100]. Figures 51 and 52 are tentatively illustrated
as the [100] moment. Therefore, branch al corresponds to the smaller ellipsoid. Here all
branches eti are experimentally split into the up and down electron states.

64

Y. ONUKI and A. HASEGAWA

9,0

'

'

'

9,0

'

'

'

'

(010)

I (ii0)

(010)

CeSb
S' ()

(ii0)

F(t'1't') I

09

~
o

o o oX
- 80O0oO~

i 0 -

....

o
o

a~/

Fx.
>

30
[101]

0
[001]

30

60
[111]

,__

9O

[110]

0
[001]

30

[i01]

Field Angle (Degrees)

30

60
[111]

90
[110]

Field Angle (Degrees)

Fig. 51. Angular dependence of the dHvA frequency

Fig. 50. Angular dependence of the dHvA frequency in the F phase of CeSb (Aoki et al. 1991, 1985).
in the S' phase of CeSb (Aoki et al. 1985). The solid The solid lines show the results of band calculations
lines connecting the data are guidelines. The Greek (Kasuya et al. 1987, Sakai et al. 1985). The Greek
letters designate the various orbits,
letters designate the various orbits.

HI[O01 ]

[IO0]

[001].

[100]

[110]

Fig. 52. Ellipsoidal Fermi surfaces of CeSb for the magnetic moment in the [001] and [100] directions (Crabtree
et al. 1987).
According to the p - f mixing model, branch/34 is the most important band. In earlier
experiments, however, branch/34 was not observed. This fact caused some investigators to
doubt the strong p - f mixing model and to apply a weak p - f mixing model (Norman and
Koelling 1986). Later, branch f14 was observed by acoustic and usual d H v A measurements
(Aoki et al. 1991, Kasuya et al. 1987). Generally thinking, it may be difficult to detect
branch/34 by the usual d H v A experiments because the small amplitude o f F F T spectrum

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

65

Table 13
Cyclotron masses m~ for the field along (100) in the RX compounds (Settai et al. 1993, Aoki et al. 1985, Goto
et al. 1993, Ozeki et al. 1991, Tanaka et al. 1993, Takeda 1992, Takeda et al. 1993)
Compound
LaSb
CeSb

a3

0.14
0.23

PrSb

0.19

SmSb

0.21
0.28

GdSb
CeAs
YbAs

0.24
0.35
0.18

al, az

0.94
0.82

0.15
0.50 (/30
0.97 (./32)

0.82

0.26

0.60
0.52

0.27
0.23
0.47

[Y

0.49

0.89 (/33)
(/34)
0.40
0.74
0.64

for branch/34 with its large mass is masked by the huge amplitudes of higher harmonics
for the other branches with small masses. The difficulty in observing branch/34 is thus
ascribed to its large mass. The mass enhancement in CeSb is quite large, more than twenty.
However, the cyclotron masses in CeSb, except branch/34, are not large, as shown in
table 13. They are in the range of 0.3m0 to 0.9m0 mass enhancement factor of two or three
(Kitazawa et al. 1988, Aoki et al. 1985). Therefore the mass &branch/34 should be large,
it is estimated to be about 10m0. Unfortunately, it has not yet been clearly determined by
experiments.
We mention dHvA measurements in other RX compounds, dHvA measurements on
CeBi were made by Kasuya et al. (1987). The result is not perfect, but the observed
branches ai, [31 and/32 agree very well with the results of band calculations.
dHvA measurements were also done for PrSb, SmSb and GdSb (Goto et al. 1993,
Ozeki et al. 1991, Tanaka et al. 1993). Figures 53-55 show the angular dependence of
the dHvA frequency in PrSb, SmSb and GdSb, respectively. The Fermi surfaces obtained
are similar to that of LaSb. For the antiferromagnetic compound SmSb, the dimensions of
the Fermi surfaces are the same in both paramagnetic and antiferromagnetic states. The
cyclotron masses in the paramagnetic (antiferromagnetic) states are 0.21m0 (0.28m0) for
branches a3, 0.82m0 (0.60m0) for branch al and a2, 0.26m0 (0.27 m0) for branch/3, and
0.74m0 (0.64rn0) for branch y for field along the (100) direction, which are 1.5-2 times
larger than those of LaSb.
Shubnikov-de Haas oscillations were detected in CeAs (Kwon et al. 1991, Takeda
et al. 1993). For a good stoichiometric sample of CeAs, the carrier number is one order
of magnitude smaller than that of CeSb. Branches ai and 3 were observed, although
branch/31 due to a hole Fermi surface in band 3 was not observed. The topology of the
Fermi surface is similar to that of LaSb, although the volume is small compared to LaSb,
as shown in fig. 56.

66

Y. 0NUKI and A. HASEGAWA

15
PrSb

{100}

{110}

10
o
o)

r+,
<

30

60

<i00>

<ii0>

Field

20

80

<iii>

Angle

'

{100}

90
<Ii0>

(Degrees)

'

{110}

Fig. 53. Angular dependence of the dHvA frequency

in PrSb (Goto et al. 1993). The solid lines connecting
the data are guidelines. The Greek letters designate
the various orbits.

20

'

'

'

'

SmSb

&

bi

+ 1o i+ ' w !

/'

2
N

I
0

30
<110>

O/a

I
I
~

0
30
60
<100>
<111>
F i e l d Angle (Degrees)

90
<110>

Fig. 54. Angular dependence of the dHvA frequency

in SmSb (Ozeki et al. 1991, Tanaka et al. 1993). The
solid lines connecting the data are guidelines. The
Greek letters designate the various orbits.

30
<110>

0
30
60
<100>
<111>
F i e l d Angle (Degrees)

90
<110>

Fig. 55. Angular dependence of the dHvA frequency

in GdSb (Goto et al. 1993). The solid lines connecting
the data are guidelines. The Greek letters designate the
various orbits.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

'

'

{100}

{110}

'

'

67
p

'

V-

teAs

{100}

rz~

o
x

{110}

VbAs

10

~4
r,-.
<I

0
<110>

30

__,

0
30
60
<100>
<111>
F i e l d Angle (Degrees)

90
<110>

Fig. 56. Angular dependence of the dHvA frequency

in CeAs (Kwon et al. 1991, Takeda et al. 1993). The
solid lines connecting the data are guidelines. The
Greek letters designate the various orbits.

<1
>
OL~

0
30
<110 >

0
30
60
<100>
<111>
F i e l d Angle (Degrees)

90
<110>

Fig. 57. Angular dependence of the dHvA frequency in YbAs (Takeda et al. 1990, Yakeda 1992), The solid
lines connecting the data are guidelines, The Greek letters designate the various orbits.

dHvA measurements were also done for YbAs, as shown in fig. 57 (Takeda et al.
1990, Takeda 1992). It is an interesting compound in the sense that it is a Kondo-lattice
compound which competes strongly with magnetic fluctuations. At low temperatures of
0.6 K, it orders antiferromagnetically. The dHvA signals were observed both above and
below the N6el temperature, without a change of the dHvA frequency but with a decrease
of the dHvA amplitude below the N6el temperature. The Fermi surface of YbAs is similar
to that of LaSb. The ~ value is estimated at the extremely large value of 200 mJ/K2 mol
(Suzuki 1993).
4.2.2. RA/2
The RA12 rare earth compounds possess the cubic Laves-phase structure, which is shown
in fig. 58. CeAI2 provides a good example of a Kondo-lattice system with the F7 ground
state and an incommensurate sinusoidally modulated antiferromagnetic structure below
3.8K (Barbara et al. 1979), as shown in table 14.
Figure 59 shows the angular dependence of the dHvA frequency in LaA12 (Seitz and
Legeler 1979, Reichelt and Winzer 1978). These dHvA branches are well explained by
the results of APW band calculations by Hasegawa and Yanase (1980a), as shown by solid
lines in fig. 59. The La atoms in fig. 58 form a diamond sublattice, and the tetrahedra of
the A1 atoms occupy vacant comers in the La sublattice. Owing to this closely packed
crystal structure, the electronic structure is characterized by strong hybridization between
the La 5d and the A1 3p bands (Hasegawa and Y~anase 1980a). In the vicinity of the
Fermi level, the 4f components in the Bloch states are small, less than 3% only. The
Fermi surface consists of various sheets distributed in bands 8, 9 and 10.

Y. 0NUKI and A. HASEGAWA

68

109
{100 }

108

O
10~

Fig. 58. The RAI2 compounds

with the cubic Laves phase crystal structure. Spheres with and
without pattern show the A1atoms
and the R atoms, respectively.

c~

CD

38 0 0

oO

10c

Fig. 59. Angular dependence of

the dHvA frequency in LaA12
(Seitz and Legeler 1979). The
solid lines are the results of
band calculations(Hasegawa and
Yanase 1980a). The Greek letters
designate the various orbits.

oOOoo~o
o o o

10s ~
30

30
<110>

<100>
Field angle

60
<111>

90
<110>

(Degrees)

Table 14
Characteristic properties of the RA12compounds
Compound

YA1z
LaA12
CeAI2

TN (K)

y
(mJ/K2tool)

Pauli para

ne= nh

5.4

Pauli para
3.8
Kondo lattice

no= nh
ne= nh

11
135

Fermi surface
sphere (electron,band 10); and a
"jungle-gym"Fermi surface (hole, band 9)
similar to YAlz
similar to LaA12,but spin-split

A main spherical branch ~ is due to a band 10 electron Fermi surface which is a sphere
centered at the F point but has bumps in the (100 / direction, as show in fig. 60. The other
branches are ascribed to the multiply connected hole Fermi surface associated with band 9.
The latter "jungle-gym"-Fermi surface favors the (100> and <110) open orbits, consistent
with magnetoresistance results (Reichelt and Winzer 1978). LaA12 is thus a compensated
metal. This is simply understood because the primitive cell contains two molecules of
LaA12. The cyclotron mass is in the range of 1.7m0 to 0.2m0, as shown in table 15.
dHvA experiments and band calculations were carried out for YA12 by Seitz (1978)
and Hasegawa and Yanase (1980b), respectively. Although the data are limited only in the

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

69

LaAI 2

(a)

(b)

Fig. 60. (a) Electron Fermi surface in band 10; (b) hole Fermi surface in band 9 of LaAI2 (Hasegawa and
Yanase 1980a). The Greek letters designate the various orbits.
Table 15
dHvA frequencies F, cyclotron masses rn~ and band masses rnb in the RA12compoundsa
Compound
LaA12
CeA12

~(110)
m~

7.23
7.61
6.98

1.68
16.2
15.0

mb
0.94

F
0.362
0.318

e(111)
m~
0.23
1.1

mb
0.19

" F values in 107 Oe, m* expressed in mo; mb expressed in m0.

References: Seitz and Legeler (1979), Hasegawa and Yanase (1980a), Reinders and Springford (1989).
{ 100} plane, the Fermi surface o f YA12 is similar to LaAlz. A small difference between
YAI2 and LaA12 is found in the band 9 hole Fermi surface.
Figure 61 shows the typical dHvA oscillation for field tilted 45 from the (100) direction in the { 110} plane o f CeA12 (Reinders and Springford 1989). The metamagnetic
transition from the antiferromagnetic to field-induced ferromagnetic (or paramagnetic)
state occurs at 52.7 kOe at this orientation. One dHvA frequency of 3.10 x 106 Oe denoted
by e is observed in the antiferromagnetic state. The same frequency is also seen above
the transition, changing its amplitude to a huge value.
The angular dependence o f the dHvA frequency in CeA12 is shown in fig. 62 (Reinders
and Springford 1989, Lonzarich 1988). The dHvA branches of CeAl2 are similar to those
o f LaAl2. The electron branch ~ is clearly split into up and down spin branches due to the
ferromagnetic exchange interaction. The exchange splitting energy defined by eq. (62) is
estimated to be 10meV.
These spin-split ~ branches are explained by LMTO band calculations (Guo 1990)
which treat the 4 f electron as a spin-polarized core state. The dHvA branch with the
frequency o f 1.3x 108 Oe, centered at <111>, was assigned from a band 9 hole Fermi
surface.

70

Y. 0NUKI and A. HASEGAWA

10 9

CeAl2

'

'

{100}

{110}

CeAl2

bO
r~

10 8 ,

~8-"J

~ ' ~ - o n ~ ~. .-.~. -. .o . . ~ ~ o oo

Hc
50

100
Magnetic Field (k0e)

Fig. 61. dHvA oscillation taken with the field

oriented 45 from (100) in the {110} plane of

CeA12 (Reindersand Springford 1989). The dHvA
oscillations are seen above and slightly below the
magneticphase transition, which occursat 52.7kOe
at this field orientation.
Fig. 62. Angular dependence of the dHvA frequency
in CeA12 (Reinders and Springford 1989, Lonzarich
1988). The Greek letters designate the various

orbits.

~(+)

J
I

107
r~.

e(+)

>

I
I

lO6

105

j6(~)

I
30

0
30
60
<100>
<111>
<ii0>
F i e l d Angle (Degrees)

9O
<110>

The measured cyclotron masses in CeAla are 1. lm0 for branch e and 16m0 for branch ~.
The enhancement factor for the mc value of LaAlz is seen to vary and falls between 5 and
10. Here, the y value of CeA12 at low temperatures (Bredl et al. 1978) is 135 mJ/K 2 mol,
while that of LaAI2 is 11 mJ/K 2 tool (Hungsberg and Gschneidner 1972). Therefore,
the enhancements of mS and y in CeA12 over their values in LaAI: are of comparable
magnitudes.
4.3. Fermi surfaces in the hexagonal Brillouin zone
4.3.1. RGa2
RGa2 compounds crystallize in the simple hexagonal A1B2-type structure. The primitive
cell contains one molecule of RGa2. Figure 63 shows the crystal structure of RGa2 and its
hexagonal Brillouin zone. Among them, CeGa2 is a highly anisotropie ferromagnet with
the easy axis in the basal plane. The magnetic phase diagram is not simple, indicating
the ferromagnetic state below 8.4 K and two or three complicated antiferromagnetic states
in the temperature region between 8.4 K and 11.4 K (Jerjini et al. 1988, Takahashi et al.
1988). From resistivity data and other magnetic properties, it is concluded that CeGa2 is
not a Kondo lattice compound but a usual f-localized one. SmGa2 is also an f-localized
compound with a N6el temperature of 20K. We summarize in table 16 the characteristic
properties of RGa2 (Umehara et al. 1992c, Tang and Gschneidner 1989, Tang 1989,
Dijkrnan 1982).
We show in figs. 64 and 65 the angular dependences of dHvA frequency in LaGa2
(Sakamoto et al. 1990) and CeGa2 (Umehara et al. 1991d, 1992c), respectively. The dHvA

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

71

kz

(b)

Fig. 63. (a) The hexagonal A1B2 type crystal structure of the RGa2 compounds. Spheres with and without
pattern show the Ga atoms and the R atoms, respectively; (b) Brillouin zone of the body-centered tetragonal
crystal lattice for c/a > v'~.
Table 16
Characteristic properties of the RGa2 compounds
Compotmd

TN (K)

LaGa2

Pauli
para

CeGa2
SmGa~

8.4
20

Tc (K)

11.4

7 (mJ/K2moO

nh

4.1-5.6

nh
nh

8.9-9.1

Fermisurface
two kinds of ellipsolidal hole Fermi surfaces
(bands 3 and 4) and a multiply connected hole
Fermi stu-face (band 5).
similar to LaGa2, but spin-split
similar to LaGa2, but affected by MBZ

branches in CeGa2 are similar to those o f LaGa2, although all branches in CeGa2 are
split into the up and down spin states o f the electrons due to the ferromagnetic exchange
interaction. The exchange energy is estimated to be about l 7 m e V from the spin factor.
Branches a and/3 correspond to the ellipsoidal Fermi surfaces o f revolution along the
c-axis ( F A direction). We note that the branch denoted b y / 3 actually consists o f two
separated branches in LaGa2. Its topology has dumb bell characteristics. It is not easy to
clarify the topologies o f branches 7, e and 6 from the present data.
The magnetoresistance was measured for CeGa2 (Umehara et al. 1991d, 1992c). The
magnetoresistance increases in a wide angle range, with a H 13-1'8 dependence. W h e n field
is applied along the [1010] direction, the magnetoresistance saturates. It was concluded
that CeGa2 possesses open orbits in a wide range o f angles, namely, in the basal plane
and arotmd the [0001 ] direction.
Owing to the simple hexagonal crystal structure with one molecule per primitive cell
belonging to the symmorphie space group (P6/mmm), LaGa2 is a compound like LaSb for
which the energy band structure can be calculated most easily i f the one-electron potential

72

Y. O N U K I a n d A. H A S E G A W A

108_ ' ' I ' ' I ' '

-

LaGa2

' ' ! ''

t/

"

"'--J

' I ' '

] 0 8 ~

' 1 ' ' 1 ' '

''

' ' 1 ~ ' 1

''

', 6
d

CeGa2

o "~

.....

"

107

FA
FM
FK
FA
1061 , , I ~ , I , ~ I ~ , I ~ , I , , I , , I
90
60
30
0
0
30
60
90

[0001]

[10i0] [1150]
Field

Angle

[0001]

(Degrees)

Fig. 64. A n g u l a r dependence o f the dHvA frequency

in LaGa 2 (Sakamoto et al. 1990). The solid lines
connecting the data are guidelines. The dashed lines
show the results o f band calculations for band 5
( H a r i m a and Yanase 1991b). The Greek letters
designate the various orbits.

FA
90

[0001]

60

30

FM

F'K

FA
30

[10i0] [1150]
Field Angle (Degrees)

60

90

[0001]

Fig. 65. Angular dependence o f the dHvA frequency

in CeGa 2 (Umehara et al. 1991d). The solid lines
connecting the data are guidelines. The Greek letters
designate the various orbits.

is constructed appropriately. Nevertheless, its Fermi surface proved to possess a sheet

with a fairly complicated topology. Harima and Yanase (1991b) performed LAPW band
calculations for LaGa2. An ellipsoidal hole Fermi surfaces exist in bands 3 and 4, centered
at the A point, which correspond to branches a and fi, respectively. A mukiply connected
hole Fermi surface, which favors the open orbits, exists in band 5, as shown in fig. 66. The
broken lines in fig. 64 are the results of band calculations for band 5. The band 5 Fermi
surface consists of twelve multiply connected arms which stretch from A to K, possessing
a hollow in the center F. Detected branches 7, 6 and e are ascribed to this Fermi surface.
The cyclotron mass of LaGa2 is in the range of 0.15m0 to 1.83m0, which is similar
to that of CeGa2, 0.32m0 to 1.44m0. The masses of branches a and fi in CeGa2 are
twice as large as those of LaGa2. On the other hand, the masses of the other branches
in CeGa2 are almost the same as those in LaGa2. The mass enhancement is anisotropic,
depending on the band, as shown in table 17. Here we note the 7 value of CeGa2. The
7 value determined at temperatures lower than the magnetic ordering temperature is about
9 mJ/K 2 mol (Umehara et al. 1992c, Dijkman 1982), while Tang and Gschneidner (1989)
estimated it as 32.6 mJ/K2 mol from the paramagnetic state. Gsehneidner (1993) suggests

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

[0001]

'

73
!

s.~a~2 ~ 1

108

0
o

i07

[ioio]

[1120]

Fig. 66. Multiply connected hole Fermi

surface in band 5 of LaG% (Harima
and Yanase 1991b). The Greek letters
designate the various orbits.

FA
i061, , ,
90

30

G0

[oooi]

FM

FK

FA
30

60

[10i0] [1120]
Field AngLe (Degrees)

90

[0001]

Fig. 67. Angular dependence of the dHvA frequency in SmGa2 (Sakamoto et al. 1990, 1992). The solid lines
connecting the data are guidelines. The Greek letters designate the various orbits.

Table 17
dHvA frequencies F and their cyclotron masses m~ a in the RGa 2 compounds for the field along [0001] except
in branches 0, r/, V for the field along [100] (Sakarnoto et al. 1990, 1992, Umehara et al. 1991d)
RG~

0
F

q
m e*

6
m~

LaG%

71.7

CeGa 2
SmG~2 128.3

66.0
64.0
2.5

77.15

1.6

y
m*e

1.83 37.6
1.42

mc*

me*

1.30

18.3

0.94
1.33
1.28

33.0

1.41

28.85

1.2

18.0
16.0

a
me*

me*

4.98

0.27

2.21

0.15

4.50
4.00
3.75
4.96

0.25
0.41
0.44
0.25

1.88

0.32

2.28

0.11

a F values in 106 Oe, m~ expressed in m0.

t h a t b o t h v a l u e s a r e c o r r e c t . T h a t is, o n o r d e r i n g , t h e d e n s i t y o f s t a t e s at t h e F e r m i e n e r g y
is r e d u c e d b y a f a c t o r o f a b o u t 3.5 i n C e G a 2 .
S m G a 2 is a n a n t i f e r r o m a g n e t . T h e e - a x i s c o r r e s p o n d s to a n e a s y a x i s a n d m e t a m a g n e t i c
t r a n s i t i o n s o c c u r at 13, 3 2 . 5 , 35, 36.5 k O e a n d m o s t likely at h i g h e r fields. F i g u r e 67

74

Y. ()NUKIand A. HASEGAWA

shows the angular dependence of the dHvA frequency (Sakamoto et al. 1990, 1992).
Branches a and/3 are split into the up and down spin states of the electrons. The splitting
energies are 10meV at fields of 17kOe to 27kOe and 26meV at fields of 45kOe to
60kOe. The cyclotron mass in SmGa2 is in the range of 0.1m0 to 2.5m0, the same
magnitude as in CeGa2. The mass enhancements of CeGa2 and SmGa2 are due to the
usual electron-magnon interaction.
4.4. Fermi surfaces in the tetragonal Brillouin zone
4.4.1. RRu2Si2 and RRu2Ge2
RRu2(Si, Ge)2 compounds possess the tetragonal ThCrzSi2-type crystal structure with
one molecule per primitive cell. Figure 68 shows the crystal structure and its tetragonal
Brillouin zone. LaRuzSi2 or LaRu2Ge2 is a non-4f reference compound. CeRu2Si2 is
thought to be a non-magnetic Kondo-lattice compound. Reflecting a rather low value
of Kondo temperature of about 20 K, the electronic specific heat coefficient y is large:
350mJ/K 2 mol (Besnus et al. 1985). CeRuzGe2 is not a Kondo lattice compound but an
f-localized ferromagnetic compound with a Curie temperature of 8 K, similar to CeGa2,
as shown in tables 1 and 18.
Figures 69 and 70 show the angular dependence of the dHvA frequencies in LaRu2Si2
(Onuki et al. 1992) and LaRu2Ge2 (Fukuhara et al. 1993). The detected branches are
simple in their angular dependences. Branch a is the main one. This Fermi surface is
assumed to be an ellipsoid of revolution shrunk along the [001] direction. The volume
of the Fermi surface becomes 1.3 x 1024 cm -3 for LaRu2Si2, which is about half of the
volume of the Brillouin zone. Here, the volume of the Brillouin zone is (2Jr/a)2(4Jr/c)
of 2.912x1024cm-3 if we use the lattice parameters a = 4 . 2 1 5 A and c = 9 . 9 3 0 A for

[001]

kz
r~ ...................

ElOOlli
'

(a)

kx

(b)

010 ]

Fig. 68. (a) Body-centeredtetragonal ThCr2Siz type crystal structure. Spheres without pattern, with simple
cross pattern and with complexpattern show the Th, Cr and Si atoms, respectively;(b) Brillouin zone of the
body-centered tetragonal crystal lattice for c/a > v/2. The rectangularparallelepipedwhich is equivalent to the
Brillouin zone is indicated by the dashed lines.

75

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

Table 18
Characteristic properties of the RRu2Si 2 and RRuzGe 2 compounds
Compound

T (K)

LaRuzSi2
CeRuzSi z

Pauli para
valence fluctuation

n~,
n = n h

LaRuzGe 2

Pauli para

nh

CeRuzGe2

nh

'

'

'

'

'

7 (m J/K2 tool) Fermi surface

large hole Fermi surface (band 14)
large hole Fermi surface (band 14); and a
multiply connected electron Fermi surface
(band 15) with open orbits
similar to LaRuzSi2
similar to LaRu2Si 2

6.5
350

20

'

'-

Ru2Si2

'

LaRu2Ge~
O~

108

1o 8

o
o)

g
o
'eeoeoooo~
o

oosm~e

o~

I~

#i~

oeo

107

90
[001]

60

30

[100]
F i e l d Angle (Degrees)

30

90

[110]

Fig. 69. Angular dependence of the dHvA frequency

in LaRuzSi 2 (Onuki et al. 1992). The solid lines are the
results of band calculations (Yamagami and Hasegawa
1992). The Greek letters designate the various orbits.

60

30

[001]
Field Angle

[100l
(Degrees)

30

[n0]

Fig. 70. Angular dependence of the dHvA frequency

in LaRu2Ge 2 (Fukuhara et al. 1993). The solid lines
connecting the data are guidelines. The Greek letters
designate the various orbits.

LaRu2Si2 in calculation. The other branches fib /32 and [33 also originate from small
ellipsoids, similar to rugby balls in topology.
The cyclotron masses for these branches are in the range of 0.5mo-2.4mo, as shown in
table 19.
It was revealed by a relativistic APW calculation (Yamagami and Hasegawa 1992) that
the main Fermi surface of LaRu2Si2 is a large closed hole sheet which is centered at the
Z point in band 14. It is essentially a sphere but highly distorted. Although its volume is

76

Y. ()NUKI and A. HASEGAWA

Table 19
Cyclotron masses for the field along [100]

Compound

fll

f12

f13

m~

mb

m~

mb

m~

mb

LaRu2Siz

1.44
2.37 b

1.20
2.59b

0.84
0.67b

0.85
0.40b

0.53
0.55b

0.58
0.30b

LaRu2Ge2
CeRuzGe2

1.95
5.6

0.76
1.5
0.97 b

5.7

0.43
0.87

~4
mb

0.62 b
0.81

mb

1.9 b
0.84

1.5 b

m~
0.52
0.51 b
0.47
0.63
2.3

0.43
0.85

1.5 b
1.5

0.97 b

m~

1.5

0.62 b

1.1 b

m~
CeRu2Si2~

120

2.3
2.6 b

References: 0nuki (1993), Aoki et al. (1992),

Onuki et al. (1992), Fukuhara et al. (1993), King
and Lonzarich (1991), Lonzarich (1988).

1.8
1.0b

19.7

20

12.3

b Values for the field along [001].

c Field along [100] for ~;, along [110] otherwise.

nearly equal to h a l f o f the Brillouin zone, it has an extraordinarily large cross-sectional

area in the (001) plane which exceeds the size o f the Brillouin zone itself. The Fermi
surface is a flat ellipsoid with steep concave and convex curvatures as shown in fig. 71.
The solid lines in fig. 69 are the results o f band calculations. Additionally, three rugby
balls are present in bands 11, 12 and 13 centered at Z, which correspond to/33,/32 and/31,
respectively. Branch/34 s e e m s to be an orbit circulating along the tube o f the doughnut in a
band 15 electron Fermi surface. The calculated )'b value is 6.95 mJ/K 2 mol for LaRu2Si2,
which is nearly the same as the experimental y value o f 6.5 mJ/K 2 mol. Therefore the
band masses are the same as those o f the experimental cyclotron masses, as shown in
table 19.
The field dependence o f magnetoresistance for LaRu2Si2 shows a saturated behavior
for the whole angle region, indicating that LaRu2Si2 (LaRu2Ge2) is an uncompensated
metal and possesses no open orbits ((3nuki et al. 1992). This is simply expected because
the primitive cell contains one molecule o f LaRu2Si2 and the La ion is trivalent, La 3,
which is consistent with the results o f band calculations.
The d H v A data for the ferromagnetic compound CeRu2Ge2 in fig. 72 (King and
Lonzarich 1991) are similar to those o f LaRu2Ge2 or LaRu2Si2, although every branch
is split into the up and down spin states. The cyclotron masses for branch a is 5 . 6 5.7m0 which is roughly three to four times larger than 1.95m0 o f LaRu2Ge2 or 1.44m0
o f LaRu2Si2, as shown in table 19. We note that branch a is not detected around [001]
in LaRu2Ge2 and CeRu2Ge2. This is mainly due to a resolution problem because the

FERMI SURFACES OF INTERMETALL1C COMPOUNDS

77

LaRu2Si 2
band 11 band 12 band 13

~ ~ ~h01e
/~3

band
hole

14
centered

at Z

/~

~, centered at Z

b a n d 15
electron c e n t e r e d

at F

Fig. 71. Fermi surfaces of LaRu2Si2 (Yamagami and Hasegawa 1992). The Greek letters designate the various
orbits.

dHvA signal for field along [001] is strongly reduced by both the curvature factor and
the cyclotron mass compared to the signal around [100] in LaRu2Si2.
Figure 73 shows the angular dependence of the dHvA frequency in CeRu2Si2 (0nuki
1993, Aoki et al. 1992, Onuki et al. 1992, Lonzarich 1988). Two kinds of rugby balls
denoted by 7 and/3 are present, which are similar to branches/31 and/32 of LaRu2Si2,
respectively. On the other hand, branches to, e and a are not present in LaRu2Si2. The
cyclotron masses are of the order of the free-electron mass for branches )' and/3, while the
other branches possess extremely large values, 120m0 for branch tp, 20too for branch to,
19.7m0 for branch e and 12.3m0 for branch a, as shown in table 19.
Under the assumption that the 4f electrons are itinerant, Zwicknagl et al. (1990) carried
out band structure calculations for CeRuzSi2 by a relativistic LMTO method both with
the LDA and the RB theory, and showed that the Fermi surface consists of four closed
hole sheets centered at the Z point and one complicated multiply-connected electron
sheet. Under the same assumption, Yamagami and Hasegawa (1993) also calculated an
energy band structure for CeRuzSi2 by a relativistic APW method with LDA. In fig. 73,
the calculated results for the dHvA frequencies by Yamagami and Hasegawa (1993) are
shown by solid lines which are compared to the experimental results. The theoretical

78

Y. 0NUKI and A. HASEGAWA

I

60

30

'

CeRuiGe~
108

io 8

i
c)

107

107

90

60

[001]

Field

30

0
[100]
Angle (Degrees)

30

[110]

Fig. 72. Angular dependence of the dHvA frequency

in CeRuzGe2 (King and Lonzarich 1991). The solid

90
[001]

[100]

30

[110]

Field Angle (Degrees)

Fig. 73. Angular dependence of the dHvA frequency
in CeRu2Si2 (Onuki 1993, Aoki et al. 1992, Onuki

et al. 1992, Lonzarieh 1988). The solid lines are the

lines connecting the data are guidelines. The Greek results of band calculations(Yamagamiand Hasegawa
letters designate the various orbits.
1993). The letters designatethe various orbits.

Fermi surface sheets which were used to calculate the dHvA frequencies in fig. 73 are
shown in fig. 74. The Fermi surfaces proposed by Zwicknagl et al. (1990) and Yamagami
and Hasegawa (1993) actually look similar to each other. In calculations of the former,
adjustable parameters were used to fit the theoretical dHvA frequencies to the observed
ones, while no such phenomenological procedure was employed in calculations of the
latter. Therefore, we discuss the origins of dHvA branches on the basis of the Fermi
surfaces calculated by Yamagami and Hasegawa.
The hole Fermi surfaces in CeRu2Si2, centered at Z, are similar to those of LaRu2Si2.
The large hole Fermi surface denoted by a in figs. 73 and 74 corresponds to branch t/t,
although it is observed in a limited angle region. The cyclotron mass of ~ is
extremely large. The band mass is theoretically calculated as 1.93m0 for field along
the [100] direction and as 4.12m0 for [001] (Yamagami and Hasegawa 1993). The
experimental mass of 120m0 is about 60 times larger than the band mass. The cyclotron
mass is thus estimated as 260m0 for [001]. This extraordinarily large mass is consistent
with the large 7 value and also the prediction of the renormalized band structure
calculations (Zwicknagl et al. 1990), which gives a mass exceeding 200m0 for [001].
Branches/3, y, to, e and a most likely correspond to hole bands 12 (denoted by c) and
13 (b), band 15 electron orbits j, h and k, respectively.

FERMI SURFACES OF INTERMETALLIC COMPOLrNDS

CeRu~Si~
(a) hole (centered at Z)
bandll bandl2
bandlS

79

(b) electron (band15)

centered
at Z

bandl4

centered
at X

Fig. 74. Fermi surfaces of CeRu2Si 2 (Yamagami and Hasegawa 1992). The letters designate the various orbits.

The magnetoresistance of CeRu2Si2 is also useful to determine the nature of the

4f electrons as in CeSn3. In figs. 75 and 76a we show the angular dependence of
the magnetoresistanee zlp/p = [p(H)- p(0)]/19(0) under constant fields and in fig. 76b
the field dependence of the magnetoresistanee for currents along [110] and [001],
respectively. The magnetoresistance shows a sharp peak at the critical field Hc of
75 kOe for field along the [001] (c-axis) direction. This corresponds to the metamagnetic
transition.
We will now discuss the magnetoresistances at lower fields than the critical field.
In the configuration of J [I [110] and/-/[1 [001], the magnetoresistance Ap/p follows
a Hl'7-dependence below 65kOe, while in the configuration of J II [110], H II [1t0]
it saturates with increasing fields, reflecting a Shubnikov-de Haas oscillation with
the dHvA frequency of 2.8x105Oe. The angular dependence of magnetoresistance
under field of 65kOe shows a maximum and a minimum at H II [001] and H II [110],
respectively. Moreover, when the direction of field is rotated from the [100] direction
to [110] direction for current along the [001] direction, the magnetoresistance increases
over the whole field direction, typically showing a Hl8-dependence in the J [I [001] and
//It [110] configuration as shown in fig. 76b. These results indicate that CeRu2Si2 in
the ground state is a compensated metal with an equal carrier number of electrons and
holes and that open orbits exist along the J x H direction, namely [001 ] (c-axis) direction.
The open orbits originate from the multiply connected electron Fermi surface mentioned
above.

Y. ONUKI and A. HASEGAWA

80
'

(a)

(b)

J//[ilO]
0.5K

/150kOe ~

q~

<1
.

CeRu28i~

CeRu2Si2

J//Ill0]

H//[IIO]

q~
Q

I
J

30

0
[001]

60

l i l l ] l l l i l i 1 1 1

90

50
100
Magnetic Field (kOe)

[110]
Field Angle (Degrees)

150

Fig. 75. (a) Angular and (b) field dependences of the magnetoresistance for the current along [110] in CeRuzSi2
(Onuki et al. 1992).

1.5

{a)l

(b)

CeRu2Si2
J//[001]
H=150kOe
0.5K

CeRu2Si2
J//[OOl]
H//[ll0]
o, 5K

\
Q
<~

/ 1

0.5

_ i

[loo]

45

[11o]
Field Angle (Degrees)

50
100
Magnetic Field (kOe)

150

Fig. 76. (a) Angular and (b) field dependences of the magnetoresistance for the current along [001] in CeRu2Siz
(Onuki et al. 1992).

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

81

We note that the magnetoresistance under field of 65 kOe approximately follows the
angular dependence of A p / p = a cos20 + b cos20, where 0 means the field angle between
the field and open orbit directions, and a and b are constant but are functions of
H. This angular dependence is expected for a compensated metal with open orbits.
If the metamagnetic behavior is not present in this compound, we expect a similar
angular dependence even under field of 150 kOe. Here, a sharp peak reflected in the
angular dependence of the magnetoresistance at 150 kOe at 0 = 60 , as shown in fig. 75a,
corresponds to the metamagnetic transition. It follows a 75kOe/cos0-dependence.
Therefore the transition occurs whenever the component of field along the [001 ] direction
equals 75 kOe and thus it becomes 150 kOe at 0 = 60.
As mentioned above, LaRu2Si2 is an uncompensated metal, while CeRu2Si2 is expected
to be a compensated metal from the results of magnetoresistance. As the primitive cell
contains one molecule of LaRu2Si2 or CeRu2Si2, we can conclude that the Ce ion becomes
tetravalent, namely implying an itinerant 4f-electron character. Therefore, the situation is
similar to the relationship between LaSn3 and CeSn3.
According to the inelastic neutron-scattering experiments on CeRu2 Si2, the ground state
originates from two low-temperature contributions: on-site interactions (Kondo) and intersite interactions (antiferromagnetic correlations). The inter-site contribution is drastically
reduced or suppressed at H >> Hc (Rossat-Mignod et al. 1988). Correspondingly the lowtemperature specific heat coefficient of 350mJ/K2mol is reduced t o 80mJ/K2mol at
200kOe (van der Meulen et al. 1991). There remains a possibility that the main Fermi
surfaces with a dominant 4f component disappear in topology when field crosses the
critical field. This is open to future studies.

4.4.2. CeCu2Si2
dHvA oscillations have been observed in the typical Kondo lattice compound CeCu2Si2
(Hunt et al. 1990, Springford 1991), having the same crystal structure as CeRu2Si2.
Superconductivity and very weak antiferromagnetic order coexist in this material below
0.7 K. The characteristic properties of CeCu2Si2 are shown in table 1.
Figure 77 shows the angular dependence of the dHvA frequency. Although only part of
the Fermi surfaces is observed, the detected carriers possess relatively large masses of 5m0
for field along the a-axis. Judging from the large value of the electronic specific heat coefficient, 1000 mJ/K2 mol, carriers with much larger masses should exist in this material.
Band calculations were made for CeCu2Si2 by Sticht et al. (1986), using a Kondo-lattice
ansatz for the cerium 4f state and LDA potential parameters. Harima and Yanase (1991 a,
1992a) also calculated the energy band structure for LaCu2Si2 and CeCu2Si2. As regards
the observed part of Fermi surface, the calculated Fermi surface of LaCu2Si2 seems to fit
the experimental results of CeCu2Si2 better than the itinerant f-electron Fermi surface for
CeCu2Si2. Figure 78 shows the multiply connected Fermi surface in band 19 for LaRu2Si2
which corresponds to the observed dHvA branches. The 4f electrons in CeCu2Si2 are thus
most likely localized and do not affect the Fermi surface significantly. Precise experiments
are necessary to clarify the real situation.

82

Y. ONUKI and A. HASEGAWA

I

o/

T-- I

CeCu2Si2

LaCu2Si 2

o
o

30

60

Fig. 78. Multiply connected Fermi surfaces in band 19 of LaCu2Si2 (Harima

and Yanase 1991a, 1992a).

90

c-axis
a
FieldAngle(Degrees)

Fig. 77. Angular dependence of the dHvA frequency of CeCuzSi z (Hunt et al. 1990). The solid lines connecting
the data are guidelines.

4.5. Fermi surface in the orthorhombic Brillouin zone

4.5.1. RNi
RNi crystallizes in an orthorhombic structure. Figure 79 shows the CrB crystal structure
and its Brillouin zone. The primitive cell contains two molecules of RNi. CeNi can be
characterized as a valence-fluctuating compound, similar to CeSn3. Its Kondo temperature
kz

(a)

""-,

D~ "',r'>j

"x

-]

,1,.

! I

_LJ

Fig. 79. (a) c-base centered orthorhombic CrB type-crystal structure. Spheres with and without pattern show
the B atoms and the Cr atoms, respectively. (b) Brillouin zone of the e-base centered orthorhombic crystal
lattice for a < b. The rectangular parallelepiped which is equivalent to the Brillouin zone is shown by the dashed
lines.

83

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

Table 20
Characteristic properties of the RNi compounds
T~ (K)

Compound
LaNi
CeNi
PrNi

g
Fermi surface
(mJ/K2 tool)

Pauli para
n= nh
valence fluctuation n= nh
20
n~= nh

multiply connected Fermi surfaces with open orbits

multiply connected Fermi surfaces with open orbits
similar to LaNi but should be spin-split

5.0
65-85

(a)
LaNi
J//e-axis
10

<

-30

30

H//b-axis

\
q

~ . ~ ~ -

Jffe-axis/
0.5K

"

,\

LaNi

4t

150k0e / / ~ , ~

o.sK

\
q~
<1

(b)

H//b-axi s
,

60

90

50

100
Magnetic Field (k0e)

120

Field Angle (Degrees)

150

Fig. 80. Angular and field dependencies of the magnetoresistance in LaNi (Onuki et aI. 1989b). The solid lines
connecting the data are guidelines.
,

(a)

'

'

CeNi
J//e-axis
0.5K
g/

I l l l l t l - - l t l t l l l

'

(b)

oj~'~'~o~.
~.~-'~
150kOe

CeNi
JHe-axis

0.5K

~2

Hffb-axi s /
/

Q2
H//a-axis

0 ' ' '

-30
0

b
'
30

' '
60

'
90

Field Angle (Degrees)

120

'

50

100

150

Magnetic Field (k0e)

Fig. 81. Angular and field dependencies of the magnetoresistance in CeNi (Onuki et al. 1989b). The solid lines
connecting the data are guidelines.
is about 150 K ( G i g n o u x et al. 1983). PrNi is an f-localized anisotropic ferromagnet with
an easy axis along the c-axis, w h i c h orders at Te = 20 K. The characteristic properties o f
RNi are s u m m a r i z e d in table 20.
Figures 80 and 8t show the typical angular and field dependences o f the transverse
m a g n e t o r e s i s t a n e e in LaNi and CeNi, respectively (Maezawa et al. 1989, Onuki et al.

84

Y. ONUKIand A. HASEGAWA
IL'I

'' ]

CeNi
108

108

mm~

0
o

0~

107

10~
<

o
m m m e)m
O

10 8

,,

[i00]
a-axis

I,,

[010]
[001]
O
b
Field Angle (Degrees)

Fig. 82. Angular dependenceof the dHvAfrequency

in LaNi (Maezawa et al. 1989). The solid lines
representthe resultsof band calculations(Yamagami
and Hasegawa 1991). The Greek letters designate
the various orbits.

[zoo]
a-axis

E_

[OLO]
[ooz]
b
e
F i e l d Angle (Degrees)

Fig. 83. Angulardependenceof the dHvAfrequency

in CeNi (Onuki et al. 1989b). The solid lines
representthe resultsof band calculations(Yamagami
and Hasegawa 1991). The letters designate the
various orbits.

1989b). The magnetoresistance of CeNi is found to be almost the same as in LaNi. The
magnetoresistance of LaNi at 75kOe is similar to that in CeNi at 150kOe, regarding
its shape as well as its magnitude. In other planes the behavior between LaNi and CeNi
is also similar. The magnetoresistance increases with increasing field over a wide angle
region, except for several particular configurations of field and current. These behaviors
suggest that LaNi and CeNi are compensated metals with similar Fermi surfaces. The
open orbits exist along the b- and c-axes.
Figures 82 and 83 show the angular dependence of dHvA frequency in LaNi and CeNi,
respectively (Maezawa et al. 1989, Onuki et al. 1989b). The two branches a (4.10 x 107 Oe)
and /3 (1.40 x107 Oe) for field along the b-axis in LaNi may correspond to those
a (3.60x 107 Oe) and c (1.25x 107 Oe) in CeNi, although the angular region for these
branches are different in both compounds. However, two other branches, indicated as
f and h, are observed along the b-axis in CeNi, which are not present in LaNi. From these
experimental results, Onuki and coworkers concluded that the Fermi surface of CeNi is
similar to that of LaNi.
The cyclotron masses of branches a and c in CeNi are 10.3mo and 8.91m0, respectively.
The masses of branches a and /3 in LaNi are 1.73m0 and 0.93m0, respectively. The

85

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

e

LaNi

CeNi

I~

(a)

(a)
b

(b)
,v

(b)

Jz

Fig. 84. (a) Hole Fermi surfaces in band 19;

(b) electron Fermi surfaces in band 20 of LaNi,
respectively (Yamagami and Hasegawa 1991). The
Greek letters designate the various orbits.

Fig. 85. (a) Hole Fermi surfaces in band 20;

(b) electron Fermi surfaces in band 21 of CeNi,
respectively (Yamagami and Hasegawa 1991). The
letters a, c, h, . . . , designate the various orbits.

cyclotron masses of CeNi are about ten times larger than those of LaNi. This is roughly
consistent with the electronic specific heat coefficient. It is 65 or 85 mJ/K 2 tool in CeNi,
while it is only 5mJ/K 2 mol in LaNi (Gignoux et al. 1983, Isikawa et al. 1987).
Yamagami and Hasegawa (1991) calculated the energy band structures of LaNi and
CeNi by the relativistic APW method. The solid curves in figs. 82 and 83 are theoretical
results. The experimentally observed branches are almost in agreement with the calculated
ones. Here the f electron in CeNi is treated as an itinerant electron. The origins of the
branches for LaNi and CeNi are shown in figs. 84 and 85, respectively. The Fermi surfaces
in bands 19 and 20 in LaNi refer to hole and electron, respectively. In CeNi bands 20 and
21 refer to hole and electron, respectively. Both compounds are compensated metals with
an equal carrier number of electrons and holes and possess multiply connected Fermi
surfaces which favor the experimentally observed open orbits along the b- and c-axes for
both compounds. The fact that LaNi and CeNi have almost the same dHvA frequency

86

Y. 0NUKI and A. HASEGAWA

Table 21
dHvA frequencies F, cyclotron masses rn~ and band masses mb in the RNi compounds for the field along the
[010] direction (b-axis) a
Compound

a
F

m~

LaNi
CeNi

4.10
3.60 (a)

PrNi

4.30

1.73
10.3 (a)

2.2

mb

m:

mb

1.04
2.62 (a)

1.40
1.25 (c)
0.658 (f)
0.215 (h)
1.30

0.93
8.91 (c)
3.06 (0
2.58 (tl)
1.06

0.50
3.68 (c)
0.91 (f)
0.98 (11)

a F values in 107 Oe, rn~ expressed in m 0.

References: 0nuki et al. (1989b), Maezawa et al. (1989, 1992), Yamagami and Hasegawa (1991).

for field along the b-axis is accidental because the topologies of the Fermi surfaces are
different between LaNi and CeNi. Since the main dHvA branches are explained by the
present band calculations, the 4f electron in CeNi is considered to form an itinerant
4f band in the same way as in CeSn3.
Every carrier possesses larger masses than the corresponding band masses, as shown
in table 21. The ratio of the cyclotron mass m~* to the band mass mb is in the range of
3-4, which is roughly consistent with the ratio of Y/Yb, 4.6-6.0.
dHvA experiments on the anisotropic ferromagnet PrNi (Tc = 20 K) show that the Fermi
surface seems to be roughly similar to that of LaNi (Maezawa et al. 1992), as shown in
fig. 86. The results ofmagnetoresistance are also almost similar to those of LaNi, although

PrNi
10 8

10 7

10 6

~ , , ~ , , ,

, , , , , , , ,

[i00]
Field

[010]

[001]

Angle

(Degrees)

Fig. 86. Angular dependence of the dHvA frequency in PrNi

(Maezawa et al. 1992). The solid lines connecting the data are
guidelines. The Greek letters designate the various orbits.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

87

open orbits exist only along the b-axis in PrNi. The cyclotron mass is also comparable
to that of LaNi, as shown in table 21.
The dHvA study in LaNi and CeNi provides an important lesson. When the number
of Ce atoms is even per primitive cell, the band structures of the La and Ce compounds
can be similar even in the itinerant 4f band model. Also the transverse magnetoresistance
has similar characteristics. Careful band calculations and precise experimental studies are
essential.
4.5.2. RCu2
The crystal structure of RCu2 is orthorhombic (Larson and Cromer 1961) as shown
in fig. 87. The primitive cell contains two molecules of RCu2. This structure can be
thought of as a distorted hexagonal A1B2 structure because the orthorhombic b-axis
approximately corresponds to the hexagonal c-axis and the relation c ~ v ~ a holds.
In fact, only LaCu2 possesses the hexagonal A1B2 structure. Therefore, instead of
LaCu2, YCu2 becomes a non-f reference material for RCu2. CeCu2 can be classified
as an antiferromagnetic Kondo-lattice substance with the Nrel temperature TN = 3.4 K
(Onuki et al. 1990a, 1985a, Gratz et al. 1985). Below TN, the magnetization of CeCu2
shows a metamagnetic behavior around 17 kOe when magnetic field is applied along
the a-axis. The magnetic susceptibility and magnetization show a large anisotropy
at low temperatures, reflecting the orthorhombic structure. SmCu2 is an f-localized
antiferromagnet with Nrel temperature of 21.7 K, as shown in table 22.
Figure 88 shows the angular dependence of the dHvA frequency in YCu2 (Settai
et al. 1990, 1992, Onuki et al. 1989c). About twenty branches are observed in YCu2.
Many cylindrical arms are detected, showing a nearly 1/cos0 behavior of the angular
dependence. Here 0 is the angle between the direction of the cylinder axis and the field
direction. We note that the cylinder axes of branches a and 6 do not coincide with the
symmetry axes. They deviate by 15 and 4 from the b-axis in the plane perpendicular to
the a-axis (a-plane). Branches/3 and g also deviate by 40 in the b-plane and by I0 in
the a-plane from the c-axis, respectively. Theoretically, three closed Fermi surfaces are
C

~ a

Fig. 87. Orthorhombic crystal structure of RCu 2. Large and small

spheres show the R atoms and the Cu atoms, respectively.

88

Y. ONUKI and A. HASEGAWA

Table 22
Characteristic properties of the RCu2 compounds

Compound
YCu 2
CeCu 2
SmCuz

T N (K)

Pauli para
3.5
Kondo lattice
21.7

ne =nh
ne = n h

7 (mJ/K~ mol)
6.7
82

Fermi surface
multiply connected Fermi surfaces

nc = nh

present in bands 30, 31 and 33. They correspond to branches ~, 0 and ~/, respectively,
which are nearly ellipsoidal Fermi surfaces elongated along the b-axis.
The conduction carriers due to branches a and 6 are thought to play an important role
in the band Jahn-Teller effect of the C44 and C66 modes from the results of the acoustic
dHvA measurements (Settai et al. 1992).
The magnetoresistance was measured for YCu2 (Onuki et al. 1989c). The magnetoresistance increases with increasing field over a wide angle range, following H 1'1-1-4. This
behavior indicates that YCu2 is a compensated metal. However, the magnetoresistance
saturates only when field is directed along the a-axis with current along the b- and caxes. Therefore, open orbits exist along the b- and c-axes. These results suggest that the
Fermi surface is multiply connected.
IO s YCu~

o
o~

i.o
10 7

.,

Field Angle (Degrees)

Fig. 88. Angular dependence of the dHvA

frequency shown by squares (0nuki et al.
1989c) and by circles, diamonds and
triangles (Settai et al. 1990, 1992) in YCu 2.
The solid lines connecting the data are
guidelines. The Greek letters designate the
various orbits.

FERMI SURFACES OF INTERMETALLICCOMPOUNDS

89

YCu2
1st . . . .

~n+~

hole

32nd electron

(a)

(c)

SIst hole

,)
b,

33rd electron

<d)
b
a

Fig. 89. Hole and electron Fermi surfaces in YCu/: (a) hole in band 30; (b) hole in band 31; (c) electron in
band 32; (d) electron in band 33. The solid lines show the first Brillouin zone (Harima and Yanase 1992b).
The thick solid lines correspond to the observedorbits.

The LAPW band calculations were done by Harima and coworkers (Harima and Yanase
1992b, Harima et al. 1990a). The origin of the dHvA branches are shown in fig. 89. YCu2
is a compensated metal, Band 30 consists of small hole ellipsoids. Band 31 consists of
ellipsoidal hole Fermi surfaces centered at the W point and a multiply connected Fermi
surface, stretching in the a-plane. The latter Fermi surface favors the presence of open
orbits along the b- and c-axes. Band 32 is also a multiply connected electron Fermi
surface, which favors open orbits along the b-axis. Finally, band 33 consists of small
electron-ellipsoids centered at the F point. These results are consistent with the results
of magnetoresistance measurements (Onuki et al. 1989c). The fact that the minima of
branches are off-symmetrical is due to the non-cubic crystal structure.
On the other hand, as shown in fig. 90, about ten kinds of dHvA branches are observed
around the a-axis in CeCu2, whereas only one branch is detected around the c-axis (Settai
et al. 1990, 1992, Satoh et al. 1990). As the dHvA experiments were done above 60 kOe,
this compound is in the paramagnetic (or field-induced ferromagnetic) state for the field
along the a-axis. On the other hand, it is antiferromagnetic in the a-plane. In general, the

90

Y. ()NUKI and A. HASEGAWA

10 8

CeCu2

,-i

107

106

...............

H//b-axis

o, L, . . . . . .
e

F i e l d Angle (Degrees)

Fig. 90. Angular dependence of the dHvA frequency

in CeC (Satoh et al. 1990). The solid lines
connecting the data are guidelines. The Greek letters
designate the various orbits.

detectable number of dHvA branches is small in the antiferromagnetic state due to the
magnetic Brillouin zone boundaries, as seen in CeA12 and Celn3.
When we compare the Fermi surfaces of CeCu2 to those of the reference material
YCu2, it is difficult to say whether they are similar or not because too many branches are
thought not to be observed in CeCu2. The effective cyclotron masses in CeCu2 range from
0.5m0 to 5.3m0, and are larger than those in YCu2, 0.1-0.7m0. This mass enhancement is
roughly consistent with that of the low-temperature specific heat coefficient ratio. Namely,
the y value of CeCu2 is 82 mJ/K 2 mol but is reduced to 50 mJ/K 2 mol at 80 kOe (Bredl
1987), while it is 6.7mJ/K 2 mol in YCu2 (Luong et al. 1985).
Magnetoresistance measurements were done for SmCu2, indicating that it is a
compensated metal with open orbits, at least, along the b-axis (Maezawa et al. 1986).
Shubnikov-de Haas oscillations were observed in the magnetoresistance. The detected
dHvA frequencies are in the range of (0.68-1.25)x 106 Oe.
4.5.3. RCu6
The f electrons in RCu6 (R = Ce, Pr, Nd and Sm) show a variety of magnetic behaviors.
C e C u 6 is a typical non-magnetic Kondo-lattice compound with a Kondo temperature of
about 4 K (Onuki and Komatsubara 1987a). PrCu6 is a nuclear cooling material with
a singlet ground state (Takayanagi et al. 1988). NdCu6 (TN =6.1 K) is a metamagnetic
compound with four discontinuous steps in the magnetization curve (Takayanagi et al.
1991, Onuki et al. 1986), and SmCu6 (TN = 9.6 K) shows a Van Vleck susceptibility due to

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

91

Table 23
Characteristic properties of the RCu6 compounds
Compound

T N (K>)

LaCu6

Paul/para

ne = nh

CeCu 6

Kondo lattice

PrCu6

singlet

ne = n h

NdCu 6

6.1

n~ = nh

SmCu6

9.6

ne ~ nh(?)

y (mJIK2 mol)

Fermi surface
four complicated Fermi surfaces

8.0

similar to LaCu6(? )

1600

similar to LaCu 6

16

similar to LaCu 6
strongly altered by MBZ

10 8

_O~

~10 7
u2

"*-i'~"~"
~_.d.I,

1/

62

/ G

o~

e
N 10 6

b\

7..' L-72

t~.J J , , , c_
b

,a.

Fig. 91. Orthorhombic CeCu 6 type crystal structure.

Large and small spheres show the Ce atoms and the
Cu atoms, respectively.

Hl/b-axis

, ,a~_s~_-7-3
e
Field Angle

(Degrees)

Fig. 92. Angular dependence of the dHvA frequency in LaCu6 (Onuki et al. 1987b, 1991b). The solid lines
connecting the data are guidelines. The Greek letters designate the various orbits.

the Sm 3+ ion (Onuki et al. 1990b). The characteristic properties of R C u 6 a r e summarized

in table 23.
The RCu6 compounds possess the monoclinic structure at low temperatures, which is a
slight modification of the orthorhombic CeCu6 structure. For example, the orthorhombic
structure of CeCu6 with a = 8.105 A, b = 5.105 A and c = 10.159 A at room temperature,
as shown in fig. 91, changes into the monoelinic one with a=5.080,~, b = 10.121 A,
c = 8 . 0 6 7 A and /3=91.36 at 65K (Asano et al. 1986). The structural transition
temperature is roughly 200 K. This transition is associated with a complete softening of
the transverse elastic constant C66 (Suzuki et al. 1985). Here, a, b, c and Y (= 90) in the

92

Y. ONUKI and A. HASEGAWA

108

-~-~

J '

'

'

'

'

' TI

10 i

r ,

i , , , ,

I
CeCu

PrCu6

t-"
~io 7

~,L. ,,~....~,

~10'

* " ~ --'7:,,'" . "'.

~o,,,

"'"

-6~ ~ J ~ 4-

lO~

.--~,~r

.~

..O~o
oi.

.-

***
*

I : -:'?'"

*
-

o*

....

oo*

i0 c

7
-

?:"

"i, ~

,f

I
C

....

I
10~ t I I
H//b-axis

I I

Field Angle (Degrees)

.~--~__.__-

10 5

~ i
b

H//b-axis

i i

i i

i
b

Field M~gle (Degrees)

Fig. 94. Angular dependence of the dHvA frequency

Fig, 93. Angular dependence of the dHvA frequency in PrCu6 (Onuki et al. 1987b). The solid lines
in CeCu6 (Springford 1991, Chapman et al. 1990a). connectingthe data are guidelines. The Greek letters
The solid lines connecting the data are guidelines,
designatethe various orbits.
orthorhombic notation are converted into c, a, b and/3, respectively, in the monoclinic
one. As the crystal distortion is small, we use the orthorhombic notation. The primitive
cell contains four molecules of RCu6.
Figures 92-96 show the angular dependence of the dHvA frequency in RCu6 where
R represents La, Ce, Pr, Nd and Sm (Onuki et al. 1987b, 1990b, 1991b, Springford
1991, Chapman et al. 1990a, Springford and Reinders 1988, Endoh et al. 1987). The
band calculations for LaCu6 made by Harima et al. (1990b) show a lot of extremal crosssections. The observed signals in LaCu6 are really quite numerous but they represent only
a part of the signals expected from the band calculations. However, the main observed
signals do correspond to results of band calculations. In the same sense, the signals in
PrCu6 and NdCu6 seem to correspond to the band calculations. The common dHvA
branches are denoted by Greek letters. The detected dHvA branches are many in number,
concentrating at the symmetry axes. The largest branch is a which consists of two or three
branches with the largest dHvA frequencies of about 2x 107 Oe.
The Fermi surface of SmCu6 is quite different from that of LaCu6, probably reflecting
the small magnetic Brillouin zone boundaries. In C e C u 6 , w e find both similar but also
different branches along the c-axis, compared to NdCu6 or LaCu6.
As mentioned above, Harima et al. (1990b) calculated the band structure of LaCu6 with
the orthorhombic structure by an LAPW method. The Cu-3d band and La-4f band are

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

108 . . . . . . . . . . . . . . . . . . . . . . . .

107

r
I

, c~

NdCu6

"-~-

93

SmCu6

x2;2U

----~ o~

'.

.
10 5 f ........
tt//b-axis

y.o.o

...o.71
i ........

', ........

104

Field Angle (Degrees)

Fig.95.Angulardependenceof thedHvA frequency
in NdCu6 ((3nuidet al. 1991b).The solidlines
connecting the data are guidelines. The Greek letters
designate the various orbits.

,,[i,I

I-I//b-axi s

....

, I i , i ,

Field Angle (Degrees)

Fig. 96. Angular dependence of the dHvA frequency
in SmCu 6 (Endoh et al. 1987, Onuki et al. 1990b).
The solid lines connecting the data are guidelines.

well localized and are separated from the Fermi level. The conduction bands are mainly
due to the 4s components of Cu. We show in fig. 97 the calculated Fermi surfaces. As
the non-cubic unit cell contains four molecules of LaCu6, about 1200 basis functions for
each point are needed especially for 120 d-bands due to the 24 Cu atoms. LaCu6 is a
compensated metal with hole Fermi surfaces in bands 149 and 150, and electron Fermi
surfaces in bands 151 and 152. The theoretical Fermi surfaces are many in number and are
characterized as corrugated Fermi surfaces with concave and convex curvatures, elongated
along the c-axis. From the magnitudes and angular dependences of the extremal orbits,
most of the experimental dHvA branches can be identified, as shown in fig. 97. The largest
orbit, which is a belly orbit on a "catcher mitt" due to the band 151 electron, corresponds
to branch a. The theoretical dHvA frequency of 2.06 107 Oe in the direction along the
b-axis is in good agreement with the experimental value of about 2 x 107 Oe.
The calculated electronic specific heat coefficient of 6.2 mJ/K 2 tool is almost the same
as the experimental value of 8 mJ/K 2 mol for LaCu6. The detected masses are thus small
and fall into the range of 0.076m0 to 2.50m0, reflecting the main 4s components of Cu.
Branch a is the main Fermi surface, which possess the cyclotron mass of 1.89m0.
On the other hand, the cyclotron masses of PrCu6 and CeCu6 are twice and forty times
larger than that of LaCu6, respectively. No mass enhancement is found in NdCu6 and
SmCu6, as shown in fig. 98. Here, the electronic specific heat coefficients of PrCu6 and

94

Y. ONUKIand A. HASEGAWA
LaCu 6
149th h o l e

151st electron

(a)

R
e

,
l~fl'th

hnlo

!~

152nd electron

b~
]

Fig. 97. Hole and electron Fermi surfacesin L a C u 6 (Harimaet al. 1990b): (a) hole band 149; (b) hole band 150;
(c) electron band 151; (d) electron band 152. The thick solid lines correspondto the observedorbits.

CeCu6 at zero external field are 16 and 1600 mJ/K 2 mol, respectively. These values may
be compared to 8 mJ/K 2 mol of LaCu6 (Satoh et al. 1989, Takayanagi et al. 1988).
The large mass due to the Kondo-lattice character of CeCu6 seems to be strongly
reduced by magnetic field. The specific heat coefficient y at low temperatures strongly
depends on the magnitude and direction of the field (Satoh et al. 1989, Amato et al.
1987). Experimental results obtained by Amato and coworkers for field along the c-axis,
which is the most sensitive direction, indicate a metamagnetic transition at about 20 kOe.
Note that the field effect is weak along the a- and b-axes, which are the hard axes. For
the highest applied field of 75 kOe, which almost corresponds to the initial field for the
dHvA measurement, the ], value is reduced to 500 mJ/K2 mol, which is larger than the
g value of LaCu6 by a factor of 63. This value is roughly consistent with the cyclotron
mass enhancement ratio of 40.
The field dependence of the cyclotron mass was studied by Chapman et al. (1990b)
for field along the c-axis. The mass of 1 lm0 measured at 41 kOe is reduced to 6.0m0 at
106 kOe for dHvA branch of 1.2 x 106 Oe. For another branch of 1.03 x 107 Oe, 34.6m0 at
103 kOe is also reduced to 29.6m0 at 122 kOe. These results are shown in fig. 99. Here, a
zero-field cyclotron mass was assumed for which the ratio m*(H)/m*(O)fits with the field
dependence of y(H). Because of the limited range of the field, it is difficult to establish

FERMI SURFACES OF INTERMETALLICCOMPOUNDS

i00

10

....

"1

LaCu6
o CeCu6
o PrCu6
NdCu6
SmCu6

~ , ~ , , i

~
O

'

'

'

'

~ '

'

' - -

CeCu6

H//c-axs

~ ~- 0.5
D
o

0.1

'

95

/:

0 [. . . . .
0

~ ....

I ....
100

Magnetie Field (kOe)

0.01 ~ r ~,..i , ,~ .....i ........ i . . . . .

104
105
106
107
108
dHvA Frequency (Oe)
Fig. 98. Cyclotron masses vs dHvA frequencies in
(Onuki et al. 1988). The solid straight lines
are guidelines.

RCH 6

Fig. 99. Field dependence of the normalized

cyclotron masses m*(H)/m*(O) (Chapman et al.
1990a) and the specific heat coefficients y(H)/y(0)
(Amato et al. 1987). The cyclotron masses at
zero field are estimated as m*(0)=20m0 for a
dHvA frequency F=l.20xl06Oe (open circles)
and m*(0)= 120m0 for F = 1.034x 107 Oe (squares).
The results of specific heat coefficients are denoted
by crosses, and the solid line is a guideline.

whether both field dependences are the same or not. Note that the field effect on the
cyclotron mass for the hard axes has to be weak, in order to be consistent for y(H).
Chapman et al. (1990b) concluded that the Fermi surface o f CeCu6 is not well described
by L M T O band calculations performed within the local density approximations, in which
the f electrons are included as either band or core states. Other kinds o f band calculations
and more experimental d H v A data, in particular around the b-axis, are desirable.

5. Conclusions
Many lanthanide compounds are clarified in Fermi surface properties. Most o f them
are light lanthanide compounds. They most likely order antiferromagnetically and the
4 f electrons are localized. Therefore, the Fermi surface in the antiferromagnetically
ordered state should be similar to that o f the corresponding non-4f La compound.
The presence o f f electrons, however, alters the Fermi surface through the f-electron
contribution to the crystal potential and through the introduction o f magnetic Brillouin
zone boundaries. On the other hand, 4 f electrons are delocalized, representing an itinerant
character, in some compounds such as CeSn3, CeNi and most likely CeRu2Si2. The many-

96

Y. ONUKI and A. HASEGAWA

body Kondo effect strongly enhances the effective mass in the Ce (or Yb) Kondo-lattice
compounds.
We summarize the Fermi surface properties of the lanthanide compounds:
(1) As the 4f bands in Y and La compounds are located far above the Fermi level, their
contribution to the conduction electrons is less than 10%. For example, the conduction
electrons are mainly La 5d electrons in LAB6, while they are Cu 4s electrons in LaCu6.
The magnetoresistance and dHvA data are in good agreement with the results of band
calculations for the compounds with the cubic, hexagonal and tetragonal structures.
However, even in the cubic case, the spin-orbit interaction is essential to correctly
calculate the energy bands and Fermi surfaces, as shown in LaSn3. As for LaNi, YCu2 and
LaCu6 with the orthorhombic crystal structure, the characteristic topology of the Fermi
surface such as open orbits for LaNi is also explained by the results of band calculations.
(2) Many dHvA measurements were performed in Ce compounds. The 4f-localized
compounds such as CeGa2 and CeRu2Ge2, which do not show the many-body Kondo
effect, are similar to the corresponding La compounds in topology of the Fermi surface.
The detected cyclotron mass is not large, comparable to those of the other lanthanide
magnetic compounds.
For the compounds belonging to the valence-fluctuation regime, such as CeSn3 and
CeNi, the dHvA data can be fitted very well by the results of band calculations based
on the itinerant 4f model. The mass enhancement is fairly homogeneous for each orbit
and not very large, less than 5, because the 4f band mass is already fairly large. Even
in CeRu2Si2, the 4f electrons are most likely itinerant in the ground state. A main
Fermi surface with the dominant 4f component is observed by the dHvA experiments,
possessing the large cyclotron mass of 120m0 for field along the [100] direction. The
mass enhancement is anisotropic for each Fermi surface in CeRuzSi2.
The 4f character in CeSn3 and CeRu2Si2 can be determined by magnetoresistance
measurements because the primitive cell contains one molecule and the valence electrons
are odd in number for LaSn3 and LaRu2Si2. If the 4f electron is itinerant, the valence
electrons are even in number for CeSn3 and CeRu2Si> This difference can be clearly
distinguished by the magnetoresistance. The experimental results are as follows. The
magnetoresistance in LaSn3 and LaRu2 Si2 saturates over a wide angle region, representing
uncompensated metals. The magnetoresistance in CeSn3 and CeRu2Si2 increases over a
wide angle region. This result shows that CeSn3 and CeRuzSi2 are compensated metals.
Therefore, it is concluded that the 4f electron is itinerant in CeSn3 and CeRu2Si2 in the
ground state.
For the compounds belonging to the Kondo regime, the dHvA data can be explained
by putting one 4f state below the Fermi energy and the other unoccupied 4f bands
above the Fermi energy. Therefore the Fermi surface is essentially similar to that of the
La compound, as well as to those of the well localized 4f systems. The mass enhancement
factor for the conduction electrons in the Kondo regime is quite large, of the order of 10
to 100 which becomes anisotropic in the some compounds such as in CeSb.

FERMI SURFACESOF INTERMETALLICCOMPOUNDS

97

The large cyclotron mass of 80m0 is detected in the non-magnetic Kondo-lattice

compound CeCu6. It is, however, not clear whether the 4f-electron character of CeCu6 is
itinerant or localized.
(3) The 4f electron character of Yb compounds is supposed to be similar to that of
Ce compounds. The dHvA measurements for Yb compounds are small in number. The
Fermi surface of YbAs with magnetic ordering is similar to that of LaSb.
(4) Fermi surfaces of Pr compounds with the singlet ground state such as PrIn3,
PrSb and PrCu6 are almost the same as those of the corresponding the La compounds,
although their cyclotron masses are twice as large as those of La compounds. Even in the
antiferromagnetically ordered Pr compounds such as PrB6, most of the dHvA branches are
roughly similar to those of La compounds, while some dHvA branches are not present
in the La compounds. They are probably produced by the antiferromagnetic Brillouin
zone boundaries. The cyclotron mass of PrB6 is roughly three times larger than that of
LAB6. The Fermi surface in the ferromagnetic compound PrNi is similar to that of LaNi,
although the directions of open orbits are a little different between them.
(5) The magnetic energy gaps associated with the magnetic ordering are small enough
and thus the electrons undergoing cyclotron motions can break through these energy gaps
and circulate on orbits of the paramagnetic Fermi surface as in Pr compounds, even if the
magnetic field is not strong. The occurrence of this effect becomes more difficult as the
number of f electrons increases. The electrons follow the orbits on the antiferromagnetic
Fermi surfaces in Nd, Sm and Gd compounds. Typical examples are shown in NdIn3
and NdB 6. For example, the paramagnetic (or field-induced ferromagnetic) Fermi surface
of NdIn3 is almost the same as that of Lain3. A nearly spherical Fermi surface in the
paramagnetic state, which corresponds to the band 7 electron Fermi surface of Lain3, is
changed in the antiferromagnetic state into a multiply connected one with necks. On the
other hand, if a Fermi surface is small and is located at the center of the Brillouin zone,
it remains in the antiferromagnetic state because it is contained in the magnetic Brillouin
zone. Branch d in Lain3 is most likely observed in the magnetic RIn3 compounds. The
small Fermi surfaces of the monopnictides, RX, also are not changed by the magnetic
Brillouin zone.
Mass enhancement due to the electron-magnon interaction in magnetically ordered
compounds is large compared to the one due to the electron-phonon interaction. The
mass enhancement factor is in the range of 1-2.
(6) Sm compounds have interesting magnetic behaviors, dHvA measurements have been
performed on some compounds such as SmIn3, SmSb, SmGa2, and SmCu6. They are,
however, not valence-fluctuation compounds but f-localized ones. Therefore, the situation
is similar to the Nd system. Fermi surface studies for valence-fluctuating Sm compounds
are open to future studies.
(7) We found that the band theory worked fairly well for studies of the electronic
structures in various lanthanide compounds. The energy band structures and the Fermi
surfaces were clarified for many La compounds, and the theoretical results were used as a
good starting point for understanding of the electronic structures of the Ce and other light
lanthanide compounds, in which the 4f electrons are believed to be localized. Moreover,

98

Y. ONUKI and A. HASEGAWA

the band theory could provide clear evidence that the 4f electrons in some Ce compounds,
especially CeSn3, belonging to the valence-fluctuation regime, are itinerant in the ground
state, and contribute directly to the formation of the Fermi surface. The agreement
between theory and experiment is as good as that attained for the 3d electron system in
the iron series transition metals, as far as the topology of the Fermi surface is concerned.
(8) Some drawbacks of the band theory are apparent. The disagreement between theory
and experiment becomes larger when it is applied to the complicated crystal structures,
like in LaNi, CeNi, YCu2 and LaCu6. Even in the La compounds with cubic crystal
structures, some of the dHvA frequency branches cannot be explained well quantitatively,
especially in the low-frequency region. The theory can certainly be improved by taking
into account non-muffin-tin shape corrections via a full-potential scheme, and some
disagreements may be removed for these La and Ce compounds. It seems to us, however,
that a treatment of the exchange-correlation interactions is more serious, and it is essential
to improve the theory so as to go beyond LDA. Such an improvement should be carried
out in a new framework of band theory in which the exchange interaction is taken into
account explicitly with the screening effect.
(9) As for an appropriate band theory for the localized 4f-electron system, an attractive
approach based on the p - f mixing model was proposed, and was applied to CeSb.
A future problem is to refine the approach so as to carry out quantitative calculations
in a self-consistent way. The anomalously large enhancement factors for the cyclotron
effective masses and the y values observed in the Ce compounds cannot be explained
by band structure alone. Quantitative analysis of the mass enhancement factor is a
problem challenging to many-body theory. There is still much room for improvement
for a complete understanding of the electronic structures of lanthanide compounds.

Acknowledgement
We are very grateful to T. Kasuya, O. Sakai, K. Ueda, A. Yanase, H. Harima,
H. Yamagami, Y. Kubo, T. Komatsubara, T. Suzuki, T. Goto, K. Satoh, H. Aoki, H. Sato,
I. Umehara, I. Sakamoto, K. Maezawa, M. Springford and G. Crabtree for helpful
discussions, and to T. Yamazald for typing the manuscript.

References
Amato, A., D. Jaccard, J. Flouquet, E Lapierre, J.L.
Tholence, R.A. Fisher, S.E. Lacy, J.A. Olsen and
N.E. Phillips, 1987, J. Low Temp. Phys. 68, 371.
Andersen, O.K., 1975, Phys. Rev. B 12, 3060.
Anderson, P.W., 1961, Phys. Rev. 124, 41.
Aoki, H., G.W. Crabtree, W. Joss and E Hulliger, 1985,
J. Magn. & Magn. Mater. 52, 389.

Aoki, H., G.W. Crabtree, W. Joss and E Hulliger, 1991,

J. Magn. & Magn. Mater. 97, 169.
Aoki, H., S. Uji, A.K. Albessard and Y. 0nuld, 1992,
J. Phys. Soc. 61, 3457.
Arko, A.J., G.W. Crabtree, D. Karim, EM. Mtieller,
L.R. Windmiller, J.B. Ketterson and Z. Fisk, 1976,
Phys. Rev. B 13, 5240.
Asano, H., M. Umino, Y. ()nuki, T. Komatsubara, E

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

Izumi and N. Watanabe, 1986, J. Phys. Soc. Jpn.
55, 454.
Barbara, B., M.E Rossignol, J.X. Boucherle, J.
Schweizer and J.L. Buevoz, 1979, J. Appl. Phys.
50, 2300.
Benoit, A., J.X. Boucherle, J. Flouquet, J. Sakurai
and J. Schweizer, 1985, J. Magn. & Magn. Mater.
47&48, 149.
Besnus, M.J., J.P. Kappler, P. Lehmann and A. Meyer,
1985, Solid State Commun. 55, 779.
Boulet, R.M., J.-P. Jan and H.L. Skriver, 1982, J. Phys.
F 12, 293.
Brandt, N.B., and VV Moshchalkov, 1984, Adv. Phys.
33, 373.
Bredl, C.D., 1987, L Magn. & Magn. Mater. 63&64,
355.
Bredl, C.D., E Steglich and K.D. Schotte, 1978, Z.
Phys. B 29, 327.
Buschow, K.H.J., H.J. van Daal, EE. Maranzana and
EB. van Aken, 1971, Phys. Rev. 3, 1662.
Callaway, J., and N.H. March, 1984, Solid State Phys.
38, 135.
Chapman, S.B., M. Hunt, E Meeson, RH.P. Reinders,
M. Springford and M. Norman, 1990a, J. Phys.:
Condens. Matter 2, 8123.
Chapman, S.B., M. Hunt, R Meeson and M.
Springford, 1990b, Physica B 163, 361.
Crabtree, G.W., H. Aoki, W. Joss and E Hulliger,
1987, in: Theoretical and Experimental Aspects
of Valence Fluctuations and Heavy Fermions, eds
L.C. Gupta and S.K. Malik (Plenum Press, New
York), p. 197.
Czopnik, A., J. Kowalewski and M. Hackemer, 1991,
Phys. Status Solidi a 127, 243.
de Haas, W.J., and P.M. van Alphen, 1930, Leiden
Comm. 208d, 212a.
de Haas, W.J., and RM. van Alphen, 1932, Leiden
Comm. 220d.
Dijkrnan, W.H., 1982, Thesis (University of Amsterdam).
Dimmock, J.O., 1971, in: Solid State Physics, Vol.
26, eds. H. Ehrenreich, E Seitz and D. Tttrnbull
(Academic Press, New York) p. 103.
Ebihara, T., I. Umehara, A.K. Albessard, Yr. Fujimaki,
K. Satoh, Y. Onuki, B. Liu and M. Kasuya, 1992a,
L Phys. Soc. Jpn. 61, 2104.
Ebihara, T., I. Umehara, A.K. Albessard, K. Satoh and
Y. 0nuki, 1992b, J. Phys. Soc. Jpn. 61, 1472.
Ebihara, T., I. Umehara, A.K. Albessard, K. Satoh and
Y. Onuki, 1993, Physica B 186-188, 123.

99

Effantin, J.M., J. Rossat-Mignod, R Burlet, H.

Bartholin, S. Kunii and T. Kasuya, 1985, J. Magn.
& Magn. Mater. 47&48, 145.
Endoh, D., T. Goto, T. Suzuki, T. Fujimura, Y. 0nuki
and T. Komatsubara, 1987, J. Phys. Soc. Jpn. 56,
4489.
Falicov, L.M., W. Hanke and M.B. Maple, 1981,
Valence Fluctuations in Solids (North-Holland,
Amsterdam).
Fawcett, E., 1964, Adv. Phys. 13, 139.
Fawcett, E., J.E Maita and J.H. Wernick, 1970, Int. J.
Magn. 1, 29.
Forgan, G.M., 1981, Physica 107 B, 65.
Fukuhara, T., K. Maezawa and J. Sakurai, 1993,
Physica B 186-188, 143.
Fulde, P., 1991, Electron Correlations in Molecules
and Solids (Springer, New York).
Fulde, P., and J. Jensen, 1983, Phys. Rev. B 27, 4085.
Galer~ R.M., E Morin, S. Kunii and T. Kasuya, 1992,
J. Magn. & Magn. Mater. 104-107, 336.
Gignoux, D., E Givord, R. Lemaire and E Tasset,
1983, J. Less-Common Met. 94, 165.
Goto, T., 1992, private communication.
Goto, T., T. Suzuki, Y. Ohe, S. Sakatsume, S. Kunii,
32 Fujimura and T. Kasuya, 1988a, J. Phys. Soc.
Jpn. 57, 2885.
Goto, T., H. Matsui, R. Settai and S. Sakatsume, 1993,
Physica B 186--188, 107.
Goy, P., and B. Castaing, 1973, Phys. Rev. B 7, 4409.
Gratz, E., E. Bauer, B. Barbara, S. Zemirli, E Steglich,
C.D. Bredl and W. Lieke, 1985, J. Phys. F 15, 1975.
Gschneidner Jr, K.A., 1993, private commun.
Gschneidner Jr, K.A., and K. Ikeda, 1983, J. Magn.
& Magn. Mater. 31-34, 265.
Gschneidner Jr, K.A., S.K. Dhar, R.J. Stierman, T.W.E. Tsang and O.D. McMasters, 1985, J. Magn.
& Magn. Mater. 47&48, 51.
Gunnarsson, O., and B.I. Lundqvist, 1976, Phys. Rev.
B 13, 4274.
Guo, G.Y., 1990, Physica B 165&166, 335.
Harima, H., 1988, Thesis (Tohoku University).
Harima, H., and A. Yanase, 1991a, J. Phys. Soc. Jpn.
60, 21.
Harima, H., and A. Yanase, 1991b, J. Phys. Soc. Jpn.
60, 2718.
Harima, H., and A. Yanase, 1992a, J. Magn. & Magn.
Mater. 108, 145.
Harima, H., and A. Yanase, 1992b, J. Phys. Soc. Jpn.
61,602.
Harima, H., O. Sakai, T. Kasuya and A. Yanase, 1988,
Solid State Commun. 66, 603.

100

Y. ONUKI and A. HASEGAWA

Hailma, H., S. Miyahara and A. Yanase, 1990a,

Physica B 163, 205; 1990.
Hailma, H., A. Yanase and A. Hasegawa, 1990b,
J. Phys. Soc. Jpn. 59, 4054.
Hasegawa, A., 1981, J. Phys. Soc. Jpn. 50, 3313.
Hasegawa, A., 1982, in: Crystalline Electric Field
Effects in f-Electron Magnetism, eds. R.E Guetin,
W, Suski and Z. Zolnierek (Plenum, New York) p.
201.
Hasegawa, A., 1985, J. Phys. Soc. Jpn. 54, 677.
Hasegawa, A., and H. Yamagami, 1991, J. Phys. Soc.
Jpn. 60, 1654.
Hasegawa, A., and A. Yanase, 1977a, J. Phys. F 7,
1229.
Hasegawa, A., and A. Yanase, 1977b, J. Phys. F 7,
1245.
Hasegawa, A., and A. Yanase, 1980a, J. Phys. F 10,
847.
Hasegawa, A., and A. Yanase, 1980b, J. Phys. F 10,
2207.
Hasegawa, A., H. Yamagami and H. Johbettoh, 1990,
J. Phys. Soc. Jpn. 59, 2457.
Herman, E, and S. Skillman, 1963, Atomic Structure
Calculations (Prentice-Hall, Englewood Cliffs,

NJ).
Herring, C., 1937, Phys. Rev. 52, 361.
Hill, H.H., 1970, in: Plutonium 1970 and other
Aetinides, ed. W.N. Minder (Metal Soc. AIME, New
York) p. 2.
Hill, R.W., J. Cosier and D.A. Hukin, 1976, J. Phys.
F 6, 1731.
Hillebrecht, EU., W. Gudat, N. M~rtensson, D.D.
Sarma and M. Campagna, 1985, J. Magn. & Magn.
Mater. 47&48, 221.
Hohenberg, P., and W. Kohn, 1964, Phys Rev. B 136,
864.
Hungsberg, R.E., and K.A. Gschneidner Jr, 1972,
J. Phys. & Chem. Solids 33, 401.
Hunt, M., P. Meeson, E-A. Probst, P.H.P. Reinders, M.
Springford, W. Assmus and W. Sun, 1990, Physica
B 165&166, 323.
Ishii, T., K. Soda, K. Naito, T. Miyahara, H. Kato, S.
Sato, T. Moil, M. Taniguchi, A. Kakizaki, Y. (3nuki
and T. Komatsubara, 1987, Phys. Scr. 35, 603.
lshizawa, Y., T. Tanaka, E. Bannai and S. Kawai, 1977,
J. Phys. Soc. Jpn. 42, 112.
Ishizawa, Y., H. Nozaki, T. Tanaka and T. Nakajima,
1980, J. Phys. Soc. Jpn. 48, 1439.
Isikawa, Y., K. Mori, T. Mizushima, A. Fujii, H.
Takeda and K. Sato, 1987, J. Magn. & Magn. Mater.
70, 385.

Jan, J.-P., 1973, Phys. Rev. B 8, 3590.

Jayaraman, A., V. Narayanamurti, E. Bucher and R.G.
Maines, 1970, Phys. Rev. Lett. 25, 1430.
Jepson, J., and O.K. Anderson, 1971, Solid State
Commun. 9, 1763.
Jerjini, M., M. Bonnet, P. Burlet, G. Lapertot, L
Rossat-Mignod, J.Y. Henry and D. Gignoux, 1988,
J. Magn. & Magn. Mater. 76&77, 405.
Johanson, W.R., G.W. Crabtree, A.S. Edelstein and
O.D. McMasters, 1981, Phys. Rev. Lett. 46, 504.
Johanson, W.R., G.W. Crabtree, A.S. Edelstein and
O.D. McMasters, 1983, J. Magn. & Magn. Mater.
31-34, 377.
Joss, W., J.M. van Ruitenbeek, G.W. Crabtree, J.L.
Tholence, A.J.E van Deursen and Z. Fisk, 1987,
Phys. Rev. Lett. 59, 1609.
Joss, W., J.M. van Ruitenbeek, G.W. Crabtree, J.L.
Tholence, A.J.P. van Deursen and Z. Fisk, 1989,
J. Phys. C 8, 747.
Kadowaki, K., and S.B. Woods, 1986, Solid State
Commun. 58, 507.
Kasaya, M., J.M. Tarascon and J. Etourneau, 1980,
Solid State Commun. 33, 1005.
Kasuya, T., 1956, Prog. Theor. Phys. 16, 45.
Kasuya, T., O. Sakai, J. Tanaka, H. Kitazawa and T.
Suzuki, 1987, J. Magn. & Magn. Mater. 63&64,
9.
King, C.A., and G.G. Lonzarich, 1991, Physica B 171,
161.
Kitazawa, H., 1982, Thesis (Tohoku University).
Kitazawa, H., T. Suzuki, M. Sera, I. Oguro, A. Yanase,
A. Hasegawa and T. Kasuya, 1983, J. Magn. &
Magn. Mater. 31-34, 421.
Kitazawa, H., Q.Z. Gao, H. Shida, T. Suzuki, A.
Hasegawa and T. Kasuya, 1985, J. Magn. & Magn.
Mater, 52, 286.
Kitazawa, H., Y.S. Kwon, A. Oyamada, N. Takeda, H.
Suzuki, S. Sakatsume, T. Satoh, T. Suzuki and T.
Kasuya, 1988, J. Magn. & Magn. Mater. 76&77,
40.
Kletowski, Z., N. Iliew and B. Stalinski, 1985, Physica
130 B, 84.
Kletowski, Z., M. Glinski and A. Hasegawa, 1987,
J. Phys. F 17, 993.
Koelling, D.D., 1969, Phys. Rev. 188, 1049.
Koelling, D.D., 1982, Solid State Commun. 43, 247.
Koelling, D.D., and G.O. Arbman, 1975, J. Phys. F 5,
2041.
Koelling, D.D., B.D. Dunlap and G.W. Crabtree, 1985,
Phys. Rev. B 31, 1966.

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

Kohn, W., and R. Rostoker, 1954, Phys. Rev. 94,
1111.
Kohn, W., and L.J. Sham, 1965, Phys. Rev. A 140,
1133.
Komatsubara, T., N. Sato, S. Kunii, I. Oguro, Y.
Fun~awa, Y. Onuki and T. Kusuya, 1983, J. Magn.
& Magn. Mater. 31-34, 368.
Kondo, J., 1964, Prog. Theor. Phys. 32, 37.
Korringa, J., 1947, Physica 13, 392.
Kubo, Y., A. Asano, H. Harima and A. Yanase, 1993,
J. Phys. Soc. Jpn. 62, 205.
Kurosawa, Y., I. Umehara, M. Kikuchi, N. Nagai, K.
Satoh and Y. 0nuki, 1990, J. Phys. Soc. Jpn. 59,
1545.
Kwon, Y.S., T. Suzuki and T. Kasuya, 1991, Physiea
B 169, 497.
Langford, H.D., W.M. Temmerman and G.A. Gehring,
1990, J. Phys. Condens. Matter. 2, 559.
Larson, A.C., and D.T. Cromer, 1961, Acta Crystallogr. 14, 73.
Lawrence, J.M., and S.M. Shapiro, 1980, Phys. Rev.
B 22, 4379.
Lethuillier, R, J. Pierre, G. Fillion and B. Barbara,
1973, Phys. Status Solidi 15, 613.
Liberman, D., J.T. Waber and D.T. Cromer, 1965, Phys.
Rev. 137, A27.
Lifshitz, I.M., and R.M. Kosevich, 1955, Zh. Eksp.
Teor. Fiz. 29, 730 [1956, Soy. Phys.-JETP 2, 636].
Longuet-Higgins, H.C., and M. de V. Roberts, 1954,
Proc. R. Soc. 224, 336.
Lonzarich, G.G., 1988, J. Magn. & Magn. Mater.
76&77, 1.
Loucks, T.L., 1967, Augmented Plane Wave Method
(Benjamin, New York).
Luong, N.H., J.J.M. Franse and T.D. Hien, 1985,
J. Magn. & Magn. Mater. 50, 153.
Luttlnger, J.M., 1960, Phys. Rev. 119, 1153.
Luttinger, J.M., 1961, Phys. Rev. 121, 1251.
Maezawa, K., S. Wakabayashi, Y. Isikawa, K. Sato,
Y. ()nuki and T. Komatsubara, 1986, J. Phys. Soc.
Jpn. 55, 4135.
Maezawa, K., T. Kato, Y. Isikawa, K. Sato, I. Umehara,
Y. Kurosawa and Y. Onuki, I989, J. Phys. Soe. Jpn.
58, 4098.
Maezawa, K., T. Kato, Y. Isikawa and K. Sato, 1992,
J. Magn. & Magn. Mater. 104-107, 1365.
Mattheiss, L.E, 1964, Phys. Rev. 133, A1399.
Mattheiss, L.E, J.H. Wood and A.C. Switendiek, 1968,
in: Methods in Computational Physics, Vol. 8, eds.
S. Fembach and M. Rotenberg (Academic Press,
New York) p. 63.

101

Meyer, R.T.W., L.W. Roeland, ER. de Boer and J.C.R

Klaasse, 1973, Solid State Commun. 12, 923.
Min, B.I., and Y.-R. Jang, 1991, Phys. Rev. B 44,
13270.
Mfiller, T., W. Joss, J.M. van Ruitenbeek, U. Welp and
R Wyder, 1988, J. Magn. & Magn. Mater. 76&77,
35.
Nasu, S., A.M. Van Diepen, HM. Neumann and R.S.
Craig, 1971, J. Phys. & Chem. Solids 32, 2773.
Niksch, M., B. Lfithi and J. Ktibler, 1987, Z. Phys. B
68, 291.
Norman, M.R., and D.D. Koelling, 1986, Phys. Rev.
33, 6730.
Norman, M.R., and D.D. Koelling, 1993, in: Handbook
on the Physics and Chemistry of Rare Earths,
Vol. 17, eds K.A. Gsehneidner Jr., L. Eyring,
G.H. Lander and G.R. Choppin (North-Holland,
Amsterdam) p. 1.
Onodera, Y., and M. Okazaki, 1966, J. Phys. Soc. Jpn.
21, 1273.
Onuki, Y., 1993, Physica B, to be published.
Onuki, Y., and T. Komatsubara, 1987a, J. Magn. &
Magn. Mater. 63&64, 281.
0nuki, Y., Y. Machii, Y. Shimizu, T. Komatsubara and
T. Fujita, 1985a, Jr. Phys. Soc. Jpn. 54, 3562.
Onuki, Y., M. Nishihara, M. Sato and T. Komatsubara,
1985b, J. Magn. & Magn. Mater. 52, 317.
0nuki, Y., K. Ina, M. Nishihara, T. Komatsubara, S.
Takayanagi, K. Kameda and N. Wada, 1986, J. Phys.
Soe. Jpn. 55, 1818.
Onuki, Y., A. Umezawa, W.K. Kwok, G.W. Crabtree,
M. Nishihara, T. Komatsubara, K. Maezawa and S.
Wakabayashi, 1987b, Jap. J. Appl. Phys. 26, 509.
0nuki, Y., Y. Kurosawa, T. Omi, T. Komatsubara, R.
Yoshizaki, H. Ikeda, K. Maezawa, S. Wakabayashi,
A. Umezawa, W.K. Kwok and G.W. Crabtree, 1988,
J. Magn. & Magn. Mater. 76&77, 37.
Onuki, Y., T. Komatsubara, RH.E Reinders and M.
Springford, 1989a, J. Phys. Soc. Jpn. 58, 3698.
0nuki, Y., Y. Kurosawa, K. Maezawa, I. Umehara, Y.
Isikawa and K. Sato, 1989b, J. Phys. Soc. Jpn. 58,
3705.
0nuki, Y., T. Omi, Y. Kurosawa, K. Satoh and T.
Komatsubara, 1989c, J. Phys. Soc. Jpn. 58, 4552.
0nuki, Y., A. Umezawa, W.K. Kwok, G.W. Crabtree,
M. Nishihara, T. Yamazaki, T. Omi and T.
Komatsubara, 1989d, Phys. Rev. B 40, 11195.
Onuki, Y., A. Fukada, I. Ukon, I. Umehara, K. Satoh
and Y. Kurosawa, 1990a, Physica B 163, 600.
Onuki, Y., A. Umezawa, W.K. Kwok, G.W. Crabtree,
M. Nishihara, K. Ina, T. Yamazaki, T. Omi, T.

102

Y. 0NUKI and A. HASEGAWA

Komatsubara, K. Maezawa, S. Wakabayashi, S.

Takayanagi and N. Wada, 1990b, Phys. Rev. B 41,
568.
Onuki, Y., T. Goto and T. Kasuya, 1991a, in:
Materials Science and Technology, Vol. 3A, ed.
K.H.J. Buschow (VCH, Weinheim) oh. 7, p. 545.
0nuki, Y., I. Umehara, N. Nagai, Y. Kurosawa and K.
Satoh, 1991b, J. Phys. Soc. Jpn. 60, 1022.
Onuki, Y., I. Umehara, A.K. Albessard, T. Ebihara and
K. Satoh, 1992, J. Phys. Soc. Jpn. 61, 960.
Ozeki, S., Y.S. Kwon, Y. Haga, T. Suzuki, G. Kido
and T. Kasuya, 1991, Physica B 169, 499.
Parks, R.D., 1977, Valence Instabilities and Related
Narrow-Band Phenomena (Plenum Press, New
York).
Rath, J., and A.J. Freeman, 1975, Phys. Rev. B 11,
2109.
Reichelt, J., and K. Winzer, 1978, Phys. Status Solidi
b 89, 489.
Reinders, EH.P., and M. Springford, 1989, J. Magn.
& Magn. Mater. 79, 295.
Reinders, P.H.P., M. Springford, E Hilton, N. Kerley
and N. Killoran, 1987, Cryogenics 27, 689.
Rossat-Mignod, J., J.M. Effantin, P. Burlet, T.
Chattopadhyay, L.P. Regnault, H. Bartholin, C.
Vettier, O. Vogt, D. Ravot and J.C. Achard, 1985,
J. Magn. & Magn. Mater. 52, 111.
Rossat-Mignod, J., L.P. Regnault, J.L. Jacoud, C.
Vettier, P. Lejay, J. Flouquet, E. Walker, D. Jaccard
and A. Amato, 1988, J. Magn. & Magn. Mater.
76&77, 376.
Ruderman, M.A., and C. Kittel, 1954, Plays. Rev. 96,
99.
Sahni, V.C., and G. Venkataraman, 1970, J. Phys.:
Condens. Matter 11, 199.
Sakai, O., M. Takeshige, H. Harima, K. Otaki and T.
Kasuya, 1985, J. Magn. & Magn. Mater. 52, 18.
Sakamoto, I., T. Miura, K. Miyoshi and H. Sato, 1990,
Physica B 165&166, 339.
Sakamoto, I., T. Miura, H. Satoh, T. Miyamoto, I.
Shiozaki, I. Oguro and S. Maruno, 1992, J. Magn.
& Magn. Mater. 108, 125.
Satoh, K., T. Fujita, Y. Maeno, Y. (3nuki and T.
Komatsubara, 1989, J. Phys. Soc. Jpn. 58, 1012.
Satoh, K., I. Umehara, Y. Kurosawa and Y. ()nuki,
1990, Physica B 165&166, 329.
Satoh, K., I. Umehara, N. Nagai, Y. Onuki, I.
Sakamoto, M. Hunt, P. Meeson, E-A. Probst and
M. Springford, 1992, J. Magn. & Magn. Mater.
104-107, 1411.
Seitz, E., 1978, J. Phys. F 8, L189.

Seitz, E., and B. Legeler, 1979, J. Phys. Colloq. C 5,

suppl. 5, 40, 76.
Sellmyer, D.J., ER. Szofran and J.W. Weymouth, 1975,
J. Low Temp. Phys. 18, 85.
Settai, R., H. Natsui, Y. Ohe, T. Goto, T. Fujimura,
Y. ()nuki, E Iga and M. Kasaya, 1990, J. Magn. &
Magn. Mater. 90&91,446.
Settai, R., T. Goto and Y. 0nuki, 1992, J. Phys. Soc.
Jpn. 61,609.
Settai, R., T. Goto, S. Sakatsume, Y.S. Kwon, T. Suzuki
and T. Kasuya, 1993, Physica B 186--188, 176.
Settai, R., T. Goto and Y. ()nuki, 1995, unpublished.
Shoenberg, D., 1984, Magnetic Oscillations in Metals
(Cambridge University Press, Cambridge).
Slater, J.C., 1937, Phys. Rev. 51,846.
Springford, M., 1991, Physica B 171, 151.
Springford, M., and P.H.P. Reinders, 1988, J. Magn.
& Magn. Mater. 76&77, 11.
Steglich, E, J. Aarts, C.D. Bredl, W Lieke, D.
Meschede, W. Franz and H. Schafer, 1980, J. Magn.
& Magn. Mater. 15-18, 889.
Sticht, J., N. d'Ambrumenil and J. Kfibler, 1986, Z.
Phys. B 65, 149.
Strange, D., and D.M. Newns, 1986, J. Phys. F 16,
335.
Sugiyama, K., F. Iga, M. Kasaya, T. Kasuya and M.
Date, 1988, J. Phys. Soc. Jpn. 57, 3946.
Sumiyama, A., Y. Oda, H. Nagano, Y. 0nuki, K.
Shibutani and T. Komatsubara, 1986, J. Phys. Soc.
Jpn. 55, 1294.
Suzuki, T., 1993, Physica B 186-188, 347.
Suzuki, T., T. Goto, A. Tamaki, T. Fujimura, Y. (3nuki
and T. Komatsubara, 1985, J. Phys. Soc. Jpn. 54,
2367.
Suzuki, T., T. Goto, S. Sakatsume, A. Tamaki, S. Kunii
and T. Kasuya, 1987, Jap. J. Appl. Phys. 26, 511.
Suzuki, T., T. Goto, Y. Ohe, T. Fujimura and S. Kunii,
1988, J. Phys. C 8, 799.
Taillefer, L., G.G. Lonzarich and P. Strange, 1986,
J. Magn. & Magn. Mater. 54-57, 957.
Takada, S., 1966, Prog. Theor. Phys. Suppl. 36, 224.
Takahashi, H., and T. Kasuya, 1985, J. Phys. C 18,
2697, 2709, 2721, 2731, 2745, 2755.
Takahashi, M., H. Tanaka, T. Satoh, M. Kohgi, Y.
Ishikawa, T. Miura and H. Takei, 1988, J. Phys.
Soc. Jpn. 57, 1377.
Takayanagi, S., N. Wada, T. Watanabe, Y. 0nuki and
T. Komatsubara, 1988, J. Phys. Soc. Jpn. 57, 3552.
Takayanagi, S., N. Wada, Y. Onuki and T. Komatsubara, 1991, J. Phys. Soc. Jpn. 60, 975.
Takeda, N., 1992, Thesis (Tohoku University).

FERMI SURFACES OF INTERMETALLIC COMPOUNDS

Takeda, N., M. Kagawa, K. Tanaka, A. Oyamada, N.
Sato, S. Sakatsume, 1". Suzuki and T. Komatsubara,
1990, J. Magn. & Magn. Mater. 90&91,425.
Takeda, N., Y.S. Kwon, Y. Haga, N. Sato, T. Suzuki and
T. Komatsubara, 1993, Physica B 186-188, 153.
Tanaka, K., N. Takeda, Y.S. Kwon, N. Sato, T. Suzuki
and T. Komatsubara, 1993, Physica B 186-188,
150.
Tang, J., 1989, PhD Thesis (Iowa State University).
Tang, J., and K.A. Gschneidner Jr, 1989, J. LessCommon Met. 149, 341.
Umehara, I., Y. Kurosawa, N. Nagai, M. Kikuchi, K.
Sato and Y. Onuki, 1990, J. Phys. Soc. Jpn. 59,
2848.
Umehara, I., N. Nagai and Y. 0nuki, 1991a, J. Phys.
Soc. Jpn. 60, 1294.
Umehara, I., N. Nagai and Y. Onuki, 1991b, J. Phys.
Soc. Jpn. 60, 591.
Umehara, I., N. Nagai and Y. 0nuki, 1991c, J. Phys.
Soc. Jpn. 60, 3150.
Umehara, I., N. Nagai and Y. Onuki, 1991d, J. Phys.
Soc. Jpn. 60, 1464.
Umehara, I., T. Ebihara, N. Nagai, Y. Fujimaki, K.
Satoh and Y. 0nuki, 1992a, J. Phys. Soc. Jpn. 61,
19.
Umehara, I., T. Ebihara, N. Nagai, K. Satoh, Y.
Fujimaki and Y. Onuki, 1992b, J. Phys. Soc. Jpn.
61, 1633.

103

Umehara, I., N. Nagai, A. Fukada, K. Satoh, Y.

Fujimaki and Y. 0nuki, 1992c, J. Magn. & Magn.
Mater. 104-107, 1407.
van der Meulen, H.E, A. de Visser, J.JM. Franse,
T.T.J.M. Berendschot, J.A.AJ. Perenboom, H. van
Kempen, A. Lacerda, R Lejay and J. Flouquet, 1991,
Phys. Rev. B 44, 814.
van Deursen, A.J.E, R.E. Pols, A.R. de Vroomen and
Z. Fisk, 1985, J. Less-Common Met. 111,331.
Wasserman, A., M. Springford and A.C. Hewson,
1989, J. Phys. Condens Matter. 1, 2669.
Wilson, K.G., 1975, Rev. Mod. Phys. 47, 773.
Windmiller, L.R., and LB. Ketterson, 1968, Rev. Sci.
Instrum. 39, 1672.
Yamada, K., K. Yosida and K. Hanzawa, 1984, Prog.
Theor. Phys. 71, 450.
Yamagami, H., and A. Hasegawa, 1990, J. Phys. Soc.
Jpn. 59, 2426.
Yamagami, H., and A. Hasegawa, 1991, J. Phys. Soc.
Jpn. 60, 1011.
Yamagami, H., and A. Hasegawa, 1992, J. Phys. Soc.
Jpn. 61, 2388.
Yamagami, H., and A. Hasegawa, 1993, J. Phys. Soc.
Jpn. 62, 592.
Yosida, K., 1957, Phys. Rev. 106, 893.
Zwicknagl, G., E. Rtmge and N.E. Christensen, 1990,
Physica B 163 97.

Handbook on the Physics and Chemistry of Rare Earths

Vol. 20
edited by K.A. Gschneidner, Jr. and L. Eyring
1995 Elsevier Science B.V.. All rights reserved

Chapter 136
THE INTRICATE WORLD OF RARE EARTH THIN FILMS:
Metals, Alloys, Intermetallics, Chemical compounds, ...
M. G A S G N I E R

Centre National de la Recherche Scientifique,

Laboratoire de Chimie Mdtallurgique et Spectroscopic des Terms Rares,
UPR 209 CNRS Bellevue, 1, place A. Briand, 92195 Meudon Cedex, France

Co~e~s
Symbols, abbreviations and units
1. Introduction
2. Rare earth metals
2.1. Introductory remarks
2.2. Valence changes in rare earths: from
vapor to solid state
2.2.1. Introduction
2.2.2. R atomic vapor
2.2.3. R clusters and surface
characterizations
2.2.4. R sub-, mono-, and multi-layers
on oriented crystals
2.2.4.1. Samarium
2.2.4.2. Ytterbium
2.2.4.2.1. The Yb/Ni system
2.2.4.2.2. The Yb/A1 system
2.2.4.2.3. The Yb/Mo system
2.3. Crystallographic properties
2.3.1. Introduction
2.3.2. Epitaxial crystal growth
2.3.2.1. R/W systems
2.3.2.2. R/Re systems
2.3.2.3. R/V systems
2.4. Electrical properties
2.4.1. Crystalline films
2.4.2. Amorphous samples
2.5. Magnetic properties
2.5.1. Crystalline state
2.5.1.1. Thin oriented and
polycrystalline R films
on oriented substrates
2.5.1.1.1. W (110) substrates
2.5.1.1.2. Fe (100) substrates

108
108
111
111
111
111
112
114
t 15
115
116
116
116
117
118
118
118
118
119
119
120
120
121
122
122

122
122
124
105

2.5.1.t.3. Cu (100) and

Ni (111) substrates
2.5.1.2. Polycrystalline films
2.5.1.2.1. Ferromagnetic order
at R surface
2.5.1.2.2. Ferromagnetic films
2.5.1.2.3. Multilayer
(Er, Tm/Lu),-type
films
2.5.1.2.4. Theoretical models
2.5.2. Amorphous films
2.5.3. Magnetic domain structures
2.6. Optical properties
2.6.1. Very low energy range (below
7eV)
2.6.1.1. Scandium films and
single crystals
2.6.1.2. Yttrium
2.6.1.3. Europium, gadolinium,
terbium, dysprosium,
holmium, erbium,
thulium and ytterbium
2.6.2. Low energy range (7-70eV)
2.6.2.1. Scandium films and
single crystals
2.6.2.2. Yttrium
2.6.2.3. Lanthanum and cerium
2.6.2.4. Gadolinium, dysprosium,
thulium, and lutetium
2.6.2.5. Samarium
2.6.2.6. Europium and ytterbium
2.6.3. Medium energy range (70600 eV)

124
124
124
125

125
125
125
126
126
126
126
127

127
128
128
128
129
129
130
130
130

106

M. GASGNIER
2.6.3.1.
2.6.3.2.
2.6.3.3.
2.6.3.4.

Scandium and yttrium

Lanthanum
Cerium
Gadolinium, thulium and
lutetium
2.6.3.5. Ytterbium
2.6.4. High energy range (beyond
800 eV)
2.7. Applications relative to metallic films
and foils
2.7.1. The piezoresistance response of
Yb foils
2.7.2. Diffusion barriers: Sm and Yb
metal interfaces
2.7.3. Tunnel junctions: Yb/oxide/pInAs
2.7.4. Magnetic devices
2.8. Metallic or contaminated films or fine
particles? Remarks about the fcc phase
2.9. Formation and properties of the rare
earth hydrides
2.9.1. Formation of the hydrides
2.9.1.1. Contamination of thin R
films
2.9.1.2. Residual hydrogen
2.9.1.3. Reaction of R elements
with water vapor
2.9.2. Physical characteristics
2.9.2.1. Electric resistance and
resistivity
2.9.2.2. Work function
2.9.2.3. Surface magnetization
2.10. Rare earth deuteride and tritide
2.10.1. Bulk materials
2.10.1.1. Introduction
2.10.1.2. Gettering
2.10.2. Thin films
2.10.2.1. Scandium and yttrium
2.10.2.2. Erbium
2.11. Formation of the rare earth nitrides
2.12. Reactivity with CO, C Q and CnH.
gases
2.12.1. Rare earth/CO and C,H n
interactions
2.12.2. Lanthanide/CQ reactions
3. Metallic R-alloys
3.1. Permanent magnets
3.1.1. Introduction
3.1.2. [R(Pr, Nd),R'(Dy)]xFeyBz alloys
3.1.2.1. Magnetic properties
3.1.2.1.1. Thick and thin films

130
130
131
131
131
131
132
132
132
133
133
133
135
135
135
135
137
137
137
140
140
140
140
140
140
141
141
141
142
143
143
144
144
144
144
145
145
145

3.1.2.1.2. Multilayers and

modulated films
3.1.2.1.3. Magnetic domains
3.1.2.2. Hydrogen treatments
3.1.2.3. Microstructural
properties
3.1.2.4. Other compounds:
RFe(N or C)
3.1.2.5. NdH Fe4B4
3.1.2.6. AI, Cr, Mn, Ga, Nb,
Zr, Cu and W addition
effects
3.1.2.6.1. AI addition
3.1.2.6.2. Cr and Mn additions
3.1.2.6.3. Ga addition
3.1.2.6.4. Nb addition
3.1.2.6.5. Zr addition
3.1.2.6.6. Si addition
3.1.2.6.7. W addition
3.1.2.7. Applications
3.1.3. RTiFe(Co) alloys (R=Sm, Nd,
Dy)
3.1.3.1. (Sm, Nd)TiFe magnets
3.1.3.2. (Sin, Dy)TiCo magnets
3.1.4. Smx(Co,M)y substituted alloys
(M=Fe, Cu and Zr)
3.1.5. RCo(B) amorphous films
3.1.5.1. Sm~Col0o_xalloys
3.1.5.2. Amorphous NdCoB
magnetic thin films
3.1.5.3. Applications
3.2. (TM/R)n modulated and multilayered
films. (R = Y, La, Ce, Nd, Gd, Tb, Dy;
TM=Fe, Co, Ni)
3.2.1. fie/R), films
3.2.1.1. (Fe/Y),,
3.2.1.2. (Fe/La), layers
3.2.1.3. (Fe/Nd), and (Fe/Dy),
films
3.2.1.4. (Fe/Gd), films
3.2.1.5. (Fe/Tb), films
3.2.1.6. (Fe/Nd, Er, Tm), films
3.2.2. (Co/R), films
3.2.2.1. (Co/(Pr, Nd)), films
3.2.2.2. (Co/Gd), films
3.2.2.3. (Co/Tb), films
3.2.2.4. (Co/Dy), films
3.2.3. (Ni/R), films
3.2.3.1. (Ni/Ce), layers
3.2.3.2. (Ni/Dy), layers
3.2.4. Other multilayer systems

146
147
148
150
151
152

152
152
153
153
153
153
153
155
155
155
155
155
156
156
156
157
157

157
157
157
158
158
159
160
161
162
162
162
164
164
165
165
165
165

THE INTRICATE WORLD OF RARE EARTH TH1N FILMS

3.2.4.1. YCo/(Nd, Gd, Er)Co/YCo
sandwiches
3.2.4.2. (TbFe/Ni-Fe-Mo)
bilayer films
3.2.4.3. (PtMnSb/TbFe) bilayer
films
3.2.4.4. (Fe/TbFe) bilayers
3.2.4.5. (Zr-Fe/TbFe)~
amorphous multilayers
3.2.4.6. (ZnS/GdFe) multilayers
3.2.4.7. (TbFe/SiO), muttilayers
3.2.4.8. Insulating/magnetooptic/insulating systems
3.2.4.8.1. Si3N4
3.2.4.8.2. SiO2/GdTbFe
3.2.4.8.3. Dielectric/Nd-DyFe~2o-Ti/Si3 N 4A1N on plastic
3.2.4.9. Amorphous

165
165
165
166
I66
166
166
166
166
166

166

(TbxFel_x/A1203)n

3.3.

3.4.

3.5.

3.6.

multilayers
3.2.5. Formation of sandwiched layers:
R(Dy, Tb) ions implanted in
Fe and Ni films. Structure and
magnetic properties
R-Fe,Co/hydrogen and other gaseous
contaminants: changes in magnetic
properties
RNi 5 and RC% as hydrogen storage
films
3.4.1. LaNi 5 alloy
3.4.1.1. Hydrogen absorption Crystalline states
3.4.1.2. Electric resistivity
3.4.1.3. Hydrogen separation and
permeation
3.4.2. Other RNi 5 alloys
3.4.3. RC% alloy (R = La, Sm)
3.4.4. Remark
Polytypic structures
3.5.1. Ferromagnetic samarium-nickel
alloys
3.5.2. Samarium-cobalt alloys
3.5.3. Rare earth trialuminides
R-noble metal alloys
3.6.1. The R-Cu system
3.6.1.1. Amorphous R-Cu films
(R = Y, Tb, Dy)
3.6.1.1.1. CUxYI_~
3.6.1.1.2. Dy-Cu and Tb-Cu
3.6.1.1.3. Cu addition to TbFe-Co

166

167

167
169
169
169
170
171
172
172
172
172
173
174
177
177
177
177
177
178
178

3.6.1.2. Crystalline R-Cu films

(R =Y, Ce, Dy, Ho, Er
and Yb)
3.6.1.2.1. (Cu/Y),
3.6.1.2.2. Ce/Cu interface
3.6.1.2.3. Cu-Dy, Cu-Ho, CuEr
3.6.1.2.4. Yb/Cu
3.6.1.3. R-Cu samples as
precursors for synthesis
catalysts
3.6.1.4. Ion-beam mixing
3.6.2. The R-Au system
3.6.2.1. Amorphous GdAu and
GdFeAu films
3.6.2.2. Crystalline films
3.7. R-Pd alloys (R=Ce, Eu, Er, Yb)
3.7.1. Valence change in R-Pd alloys
3.7.1.1. Ce-Pd amorphous
crystalline samples.
Comparison to other
materials
3.7.1.2. Eu and Yb-Pd
amorphous specimens
3.7.2. Rx-Pdy crystallographic
properties
3.8. Superconducting materials
3.9. Miscellaneous R-M alloy films and thin
crystals
3.9.1. Y, La-Pb multilayer samples
3.9.2. Transport properties in U-R films
(R = ad, Yb)
3.9.3. Quasicrystal Co-Er alloys
4. Metalloid R-compounds
4.1. Rare-earth chalcogenide and bismuth
systems
4.1.1. RAs films (R=Tb, Dy, Er, Yb
and Lu)
4.1.1.1. ErAs, ErxScl_xAs and
ErP, As~_x specimens
4.1.1.2. YbAs films
4.1.1.3. Yb-As2S3 films
4.1.2. RSb films (R=Sm, Tm, Yb)
4.1.3. RBi films (R=Gd)
4.2. Pb(Eu, Yb)(Te, Se, S) thin films
4.2.1. Lead--europium films
4.2.2. Lead-ytterbium films
5. Conclusion
References

107

178
178
178
179
179

179
180
180
180
180
18I
181

18 I
182
182
184
185
185
185
185
186
186
186
186
187
187
188
188
188
188
190
190
190

108

M. GASGNIER

Symbols, abbreviations and traits

AES
bee
dhcp
Eg
e-beam
e-gun
EDP
EDX
EELS
ESR
fcc
Hc
hcp
HREM
kAm-1
Ku
L
LEED

Auger electron spectttm

body centered cubic
double hexagonal close packed
energy gap
electronbeam
electrongun
electrondiffraction pattern
energy dispersive X-rays
electronenergy loss spectrum
electronspin resonance
face centered cubic
coercive force or coercivity
hexagonalclose packed
high resolution electron microscopy
12.5 Oe (1 kOe=80kAm-1)
uniaxial magnetic anisotropy (uniaxial
perpendicular energy)
Langmuir (deposit at
10-6 Torr s = 1.33x 10-4 Pa s)
low-energyelectron diffraction

MBE
Ms
Pa
R
RBS
rf
RHEED
SQUID
STEM
T
Tc
TCR
TEM
TM
UPS
XAS
XPS
XRD

molecular beam epitaxy

saturation magnetization
133.33 Torr
rare earth elements
Rutherford backscattering
radio frequency (sputtering)
reflection high-energyelectron diffraction
superconducting quantum interference
device
scanning transmission electron
microscope
Tesla (10 kGauss)
superconducting transition temperature
temperature coefficientof resistance
transmission electron microscope
transition metal
ultraviolet photoemission spectroscopy
X-ray absorption spectroscopy
X-ray photoemission spectroscopy
X-ray diffraction pattern

1. Introduction
During the last decade numerous papers have reviewed and commented the main physical
and chemical properties o f the rare earth elements. Eyring (1979), Gasgnier (1980, 1989)
and Boulesteix (1982, 1984) have reported the main characteristics o f the sesquioxides (R2
03) and o f the intermediate oxides (ROx, with R = Ce, Pr and Tb, and 1.5 ~<x ~ 2). Netzer
and Bertel (1982) and Netzer and Matthew (1987) have provided much informations
relative to the surface reaction (adsorption), valence and electronic states and catalysis
properties o f rare earth metal alloy, hydride and oxide as thin samples. The characteristics
o f rare earth compounds (sulfides, silicides, borides,... ) have been reviewed by Gasgnier
(1989). Eschenfelder (1980a, 1980b), Gasgnier (1982), Buschow et al. (1982) and
Busehow (1984) have presented a large number o f results on the rare earth-3d transition
metal alloys as amorphous or crystallized thin and bulk materials. The properties of
permanent magnets, superconductors (other than those o f the new high-To ceramics) and
magnetic multilayers have been reviewed by Burzo and Kirchmayr (1989), Coey (1991)
and Gasgnier (1991).
Before investigating the results given in the literature, one must emphasize three main
problems which seem to be always in competition. Indeed, and in spite of the numerous
reliable experimental results and theoretical studies published by several research groups,

THE INTRICATEWORLDOF RARE EARTHTHIN FILMS

!09

"strange" and "extraordinary" new properties are always being reported now and again
in different articles.
The first problem is relative to the oxidation state, and we should like to exclaim: "poor
rare earth valency!". So, recently, Savrin et al. (1988) have reported an tmexpected series
of oxides as Pr2Os, Nd205 and Nd6Oll. However, a careful survey of the published Xray diffraction patterns shows that these "compounds" correspond to Pr(OH)3, Nd(OH)3
and C-Nd203 respectively. The misappreciation of the oxidation states of the R elements
is hardly understandable. Indeeed, the authors claim that the Pr and Nd oxidation states
correspond to the electronic 4f3(6s 2) and 4f4(6s 2) structures, respectively, and therefore
they conclude that these elements have a +5 valence, includingf electrons in the valence
band. Such electronic configurations are known to be related to the neutral atoms, but in
the elemental metallic state the basic configuration becomes 4fn(5d16s 2) except for Eu
and Yb which retain the 4fn(6s 2) configurations (n = 7 or 14). To explain the valences of
the R elements, one must recall the works of D.A. Johnson (1969, 1977) who has studied
the stability of the RC12 compounds with respect to the disproportionation reaction:
3RC12(s) + R(s) --+ 2RC13(s).
The dichlorides are stable if the change in free energy is positive. The following order for
the relative reducibility of the lanthanides has been deduced: Eu, Yb, Sin, Tm, Dy, Nd,
Ho, Er, Pr, Tb, Ce, Gd and La. In this way, one can remember the useful classification
of the R elements:
pure +2 (EuO)
Eu
Yb, Sm, Tm

+2 with S, Se, Te (very unstable with 02)

La to Lu (incl. Sc + Y)

+3 (R203)

Ce, Pr, Tb

+4 (Ro2)

+3 --+ +4 (ROx)
Ce, Pr, Tb
Under standard conditions of temperature and pressure it is not possible to obtain other
oxides.
The second problem concerns the reactivity of the lanthanides. Numerous research
groups have demonstrated that they are excellent getters for hydrogen and that they can
absorb (OH)- radicals. But such properties often remain misappreciated, and numerous
physical and chemical results and interpretations are then hidden or altered. Indeed, the
hydrides or the hydroxides can be formed as chemical compounds either at the surface
of a sample or at grain boundaries. This getter effect is strongly enhanced in the case of
thin films which are characterized by a large surface/volume ratio. Besides, in the case
of multilayer sequences, the interface between two different materials must be carefully
studied. Two phenomena can occur concurrently: one is relative to the cation and/or anion
diffusion and the other to the possible formation of chemical compounds and alloys. So,
in the case ofW/C/W/C.., as a stacking sequence, the formation of a WxCyOz compound
(0.5-1 nm thick) has been observed at ambient temperature (Gasgnier et al. 1983). In the
same way, the formation of passivation thin films, even if their thickness does not exceed

110

M. GASGNIER

some 10nm, becomes a "poison" for some physical measurements. A rapid calculation,
in the case of a film with a thickness of about 100nm, demontrates that about 10% of
the total thickness is far from the "ideal" metallic material. Therefore, many times it is
difficult to neglect the formation of thin contamination films at the interfaces.
The last problem is relative to the crystallographic phases. They are often deduced from
electron diffraction patterns (EDP's). Various phenomena which occurred in the course
of the study of thin samples are sometimes badly known, or else the patterns can be
misinterpreted and the indexation becomes wholly wrong. Recently Z. Li et al. (1988)
have claimed to the formation of new polymorphic erbium oxide phases. These were in
fact the well-known ErH2, C- and B-Er203 compounds (Gasgnier 1980, 1990). Other
misinterpretations result from decided opinions on chemical reactivity, phase transitions,
compound formation (as Lu(OH)4 for example) (Gasgnier 1991) ... and/or on disorder
between two crystallographic phases. The rare earth series display basic chemical and
physical properties which are now well established. Moreover, the new micro- (and
even nano-) analysis apparatus should be used in a systematic way to insure accurate
determination of the specific properties of the materials.
Another remark must be made: it is appropriate that authors should report all the
pertinent references in their articles, but many times this is not the case. So, it is often
observed that compounds like SmO (this one is something of a "myth" for numerous
authors) and Sm20 (Eick et al. 1956, Ellinger and Zachariasen 1953) are always taken
into account. However a later paper of Felmlee and Eyring (1968), which demonstrated
that these compounds are actually SmNl_xOx and SmH2, respectively, is never listed. As
an example, one can mention the results of Tkach et al. (1980) who claim to the formation
of SmO (with a fcc parameter as a = 0.502 nm) instead of the SmNl_xOx compound. In the
same way one can add that these authors have "deduced a new Sm-hcp metallic phase",
which is in fact the dhcp phase (Gasgnier 1980) which seems fully unknown to them.
A second example is also noteworthy: Curzon and Chlebeck (1972, 1973) had, at first,
claimed that rare metal films crystallized as a new fcc phase. However, later on (see
Gasgnier 1980 for accurate comments) they agreed to have been confused and concluded
that this phase was in fact the rare earth dihydride. These authors have published a
large number of papers which refer to the dihydride films, but others only reference
the first articles. So the following questions can be put: "Why is the gettering action of
R elements for hydrogen neglected and/or ignored by some research groups? Is it to fit
some theoretical results?" In conclusion one can notice that the second example overlaps
the three problems discussed in this Introduction. These remarks allow us to understand
why some properties of rare earth elements have always been the subject of hard polemics
and debates.
Therefore, our purpose is to demonstrate that the physical and chemical properties of
rare earth metallic and rare earth alloy thin films or single-crystals must be cautiously
analyzed. Particularly, the lack of crystallographic spectra (X-ray or electron diffraction),
of chemical analyses (absorbed or adsorbed gases, surface contamination, impurities,... ),
of structural investigations (grain size, defects. . . . ) for example, is truly detrimental
to precise characterization of the materials. In this way one can claim that numerous

THE INTRICATEWORLD OF RARE EARTHTHIN FILMS

111

polemics become unfruitful because the arguments of the different authors are often
strictly not convincing.

2. R a r e earth metals

2.1. Introductory remarks

For about twenty years the problem of the purity of rare earth metals either as thin films,
or thin particles, or thinned materials is a matter of fact. These metals are quite sensitive
to various gaseous contaminants. So, processes like vacuum deposition, annealing, aging,
and thinning treatments either under vacuum or at ambient atmosphere, must be carried
out very cautiously. Consequently the possible formation of chemical compounds and/or
surface or volume contamination might be proven. So, one can point out that the
experiments of Onsgaard et al. (1979), Onsgaard and Ignatiev (1982) and Tougaard
and Ignatiev (1981) have shown that the (10.0) and (00.1) single crystal surfaces of
scandium exhibit segregation of impurities like S, C1, C and O. After heating (between 813
and 1213 K) and/or krypton ion bombardment a clean Sc surface has been obtained, as
demonstrated by the Auger spectra. Another example is from SIMS measurements by
Reichl and Gaukler (1986). At the surface of a Y evaporated film the following species
have been observed: YH +, YHO +, YO-, YO] and Y203. It is unlikely that the presence of
such poisons, and also that of composite materials (e.g. metal + hydride and/or oxide) are
always verified, and thereby the physical and chemical properties are changed compared
to the ones of the pure bulk metal.
2.2. Valence changes in rare earths: from vapor to solid state

2.2.1. Introduction
The problem of mixed-valence behavior which is sensitive to particle size has been studied
for Sm, Tm and Yb metals. A review paper by Connerade and Karnatak (1990) points out,
for vapors and clusters, the main calculations and interpretations which have been done
to explain this property. The authors report numerous XAS results which demonstrate
experimentally the valence change with cluster size.
Netzer and Bertel (1982) have given numerous results relative to the mechanism
of surface oxidation. A theoretical approach of valence fluctuations has been done by
B. Johansson (1979) and Rosengren and Johansson (1982 and refs. therein). The latter
is a review paper which allows to have a good understanding of the main results
and hypotheses published up to this date. Before discussing the results, one must add
the following remarks. The first was made by B. Johansson (1979): "This means that
experimental results, obtained from a surface-sensitive technique, must be carefully
analyzed before a specific structure of the spectrum can be said to originate from the
bulk or the surface, respectively". The second one is that the divalent state of Tm has not
been convincingly demonstrated up to this date.

112

M. GASGNIER

2.2.2. R atomic vapor

The problem of the electronic (or ionic) configurations for the three different states of the
R elements: vapor, metal and oxide, has been described in detail by Gsehneidner (1971),
Gschneidner and Daane (1988), Gschneidner et al. (1990), and Beaudry and Gsehneidner
(1978). In the gaseous state two types of configuration are encountered:
(1) 4fn(5d16s 2) for La, Ce, Gd and Lu [Sc with 3da4s 2, and Y with 4d15s 2 must be
added]. This is the normal configuration whieh leads to a difficult reducibility of
the elements (see D.A. Johnson 1969, 1977).
(2) 4fn+d(6s2) which is the most common configuration for Pr--+ Eu and Tb ~ Yb.
In the metallic state the configuration change leads to the following electronic states:
(1) 4fn(5d16s2), i.e., the normal trivalent state for Sc, Y, La --+ Sm, Gd ~ Tm and Lu.
(2) 4fn(6s 2) for Eu and Yb, i.e., the divalent state and the more reducible metals of
the R series (see D.A. Johnson 1969, 1977).
The first experimental absorption spectra have been reported by Materlik et al. (1983b).
The LIII (2p --+ 5d transition) absorption edges of Ce, Sm, Gd and Er as vapors and metals
are compared. From previous works (Materlik et al. 1983a), it is asserted that Sm and Gd
to Lu metals exhibit similar white lines well described by the ground-state potential and
that no major redistribution of the valence electrons occurs. (The authors demonstrated
that the white line and the near-edge structure can be well described by a single-particle
band-structure calculation.) In the case of metallic vapor of Ce and Gd the Lni edges
almost coincide with those of the metal. In contrast, for Sm, and Er, the peak of the white
line is shifted towards higher energies by about 8 eV. In the same way the LI spectra give
identical results (shift of about 5 eV). The authors conclude that a configuration change
occurs for Sm and Er (4f6--* 4f 5 and 4f~2--~ 4f 11, respectively) upon solidification.
Blancard et al. (1989, 1990a,b), Sarpal et al. (1991) and Connerade and Karnatak
(1990) have carried out accurate experiments to establish the valence of Sm and Tm
vapors. The Mw-v (3d--+ 4f transitions) edges recorded by XAS show that Sm and
Tm atoms have characteristics of the +2 valence state. This is well demonstrated by a
remarkable evolution in the spectra as one changes from the atomic vapor to matrix-

......

ib

......

ib~ . . . . .
1i

ib 3 ......
i

Fig. 1. The evolution of valence (by adding 2 to IAnl)

with clusterconcentrationfor Sm (solid squares) and Tm
(opensquares). (Thetermclustersizerelatesto the average
number of atoms in the first coordinationcell, as revealed
by EXAFS.) (By courtesy of Drs. R.C. Karnatak and
J.M. Esteva, Univ. Paris-Sud, LURE, Orsay,France).

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Samarium

Tm CLUSTERS
NIv

,.~
:%.:

....

__1

113

;.

M v

".

M IV

SOLID

....--."v.......__

/"x.,"

~(~/Tm) = 8o:1

'-

____.j.....".
,;" X._,."

/".
:'
"" ""

"<

R'(Ar/Tm)=221:l

" ".

1070

1080

1090

1100

1110

1120

PHOTONENERGY(eV)

R(Ar/Tm)=5085:1
.,.,~'.,'v~':,:...."

455

1060

R'(Ar/Tm)= 3346:1

"

1460

,.,.,~,:

1465

Fig. 2. (a) The 3d (Mv) spectrum of Tm showing the

evolution from the vapor to the solid via intermediate
concentrations of trapped clusters for Tm in At. (b) The
3d (Mw_v) spectra of Sm: comparison between the
(1) vapor and (2) condensed phases. (By courtesy of Drs.
R.C. Karnatak and J.M. Esteva).

1470

Photon Energy (eV)

isolated atoms, clusters and solids. In the case of atomic Sm vapor, it is observed that
the 3d spectrum is close to that of SINS. It is concluded (figs. 1-3) that the localized
Mv ::=

Sm2+

MIv

L~]~ ~'~ Vaponr (1300K) / ~

am

SmS(lO0) ( 2 9 3 o K ~
,i

: %

: ".."" r=2300:1 (10K).'~

""" ....

1070

'

,""
"

1090
Photon Energy (eV)

"':"

1110

Fig. 3. Comparison between the 3d spectra

of Sm vapor, SInS solid and Sm in clusters
for two different concentrations ~ (argon
to samarium ratio). Notice that at very
low concentrations muttiplets are present.
(By courtesy of Drs. R.C. Karnatak and
J.M. Esteva).

114

M. GASGNIER

excited states possess properties similar to those in the solid. The spectra are interpreted
according to the multiconfigurational Dirac-Fock calculations.

2.2.3. R clusters and surface characterizations

Mason et al. (1981) have studied Sm clusters prepared by vapor deposition onto
amorphous carbon. XPS and UPS data show that Sm is primarily divalent in the smallest
particle sizes. The trivalent state appears as the size increases and becomes the dominant
state for the bulk metal. For Sm, Nd and Pr clusters isolated in matrices of Ne, Ar and
Kr (freezing at 5 K), Niemann et al. (1986, 1987) and Liibcke et al. (1986) show, from
the Lm edge spectra, that for Sm a valence transition occurs at 1% metal concentration
(from +2.0 to +2.6) with a critical cluster size of 13 atoms. For clusters consisting of
more than 20 atoms the average valence approaches +3. For Pr and Nd a complete
valence change occurs within an atomic concentration range of 0.10-0.25% metal content,
and clusters containing more than 5 atoms are totally trivalent. Chorkendorff et al.
(1983) have studied AES and energy-loss spectra of Yb vapors. It is demonstrated that
the N4,sN6,7N6,7 Auger transition can be interpreted by using an independent particle
description. Experimental and calculated results agree reasonably between them. The
16.10 eV metal-vapor Auger energy shift can be explained by a thermochemical model.
The relative energies and intensities of the individual Auger lines are found to be
essentially the same for the vapor and the solid phases. The quasiatomic rather than the
band-like model allows to describe correctly the solid-state spectra. Brrchignac et al.
(1991) have studied the generation of Sm, Eu and Yb metal clusters by means of the
gas-aggregation technique. They draw attention to the transfer of a 4f-localized electron
to the 5d-delocalized orbital as cluster size increases, and second to the interplay between
the electronic structure of the clusters and their stability. As a function of temperature
(4.2-10 K) the mass spectra exhibit different nucleation conditions. It is suggested that a
rapid cooling favors icosahedral clusters structures (see also Rayane et al. (1989) in the
case of Sm clusters), whereas smooth cluster growth induces the R bulk structure. The
authors conclude that it is not straigthforward to correlate cluster stabilities and electronic
structures.
The experiments of Blancard et al. (1989, 1990a,b) and Connerade and Kamatak
(1990) show clearly that Sm and Tm clusters imbedded in an Ar matrix (10K) exhibit
a +2 valence for very low dilutions (figs. 1-3). Two main conclusions arise from these
results: First, surface divalency in metallic Tm occurs only on rough surfaces where the
atoms have reduced coordination numbers, as observed by Domke et al. (1986) for Tm
sputtered films (100nm thick) deposited onto copper substrates at 300 or 13 K. Second,
the previous interpretation of Thole et al. (1985) for Sm is not consistent with the new
results. Indeed, the Hartree-Fock calculation cannot explain transitions from the Sm 4f 6
ground state and the observed 3d-4f multiplet structure for Sm atoms and small clusters
at low teperature. So, multiconfigurational Dirac-Fock calculations (unsealed "ab initio"
calculations) have been performed to understand the 3d-4f spectra. It is also concluded
that for Sm clusters the divalence is found to stabilize for a coordination of 8-10,

THE INTRICATEWORLDOF RARE EARTHTHIN FILMS

115

whereas it is only 6-7 for Tm. The latter result has been determined in the case of rough
surfaces by Domke et al. (1986). They report from photoemission measurements that
films deposited onto smooth surfaces are purely divalent, whereas in the case of a rough
surface small divalent components (low-coordinated atoms) are well observed.
Previously B. Johansson (1979), L.I. Johansson et al. (1981), Lang et al. (1981), Kaindl
et al. (1983) and Gerken et al. (1985) concluded that, whatever the substrates, thin Tm
films stay trivalent. A systematic study of the lanthanide series from Ce to Lu has been
carried out by Gerken et al. (1982, 1985) and Kammerer et al. (1982), who report XPS and
100 eV phonon energy spectra to investigate the surface core level shifts. The 4f emission
from the surface layer atoms and the bulk atoms exhibits clearly separated structures with
different binding energies. For Sm they conclude to a well-defined surface divalent state.
One must point out the work of Strasser (1988) who has studied the 4d core electron and
valence excitations by EELS. It is shown that Sm is divalent at the surface and trivalent
in the bulk. Wieliczka and Olson (1990) have observed, for Sm (2-200 nm)/Cu bilayers, a
heterogeneous mixed valence state in Sm. From energy distribution curves it is concluded
that the surface layer is completely divalent, with the underlying layers being trivalent. The
divalent intensities recorded increase with the initial stages of Cu diffusion throughout
the Sm film. The anomalous diffusion indicates that the Cu distorts the Sm lattice to an
extent which allows for a trivalent to divalent transformation. It is also demonstrated that
no Cu-Sm alloys are formed.
One can conclude, first that for Sm the divalent and mixed-valence characteristic
surfaces have been well recognized (Wertheim and Creeelius 1978, J.W. Allen et al. 1978,
1980, Gerken et al. 1982, 1985, den Boer et al. 1988); and second that this character is
only observed in Tm for some particular cases (vapor, clusters, and films deposited onto
rough surfaces).

2.2.4. R sub-, mono-, and multi-layers on oriented crystals

2.2.4.1. Samarium. F~ildt and Myers (1983, 1984a,b, 1985, 1986) and F~ildt et al. (1988)
have studied the mixed-valence state of Sm onto A1 [(001) and (111)]-, Cu (001)- and
Pd (001)-oriented single-crystals. From LEED and XPS (3d5/2 spectra) measurements it
is shown that isolated Sm atoms adsorbed onto A1 (111) have an intermediate valence
of+2.40. For a Sm monolayer the valence reaches +3. On A1 (001) and Cu (001) surfaces
the valence states of one monolyer are 2.95 and 2.70 respectively. The mixed valence is
homogeneous for Sm onto Cu (001). Surprisingly a weak component [Sm on A1 (001)]
has been interpreted to be associated with a Sm4+ valence state, which is not possible of
course (see the Introduction and Connerade and Karnatak 1990). In the case of Pd (001)
substrates XPS shows that Sm exhibits a completely +3 state, similar to A1 [(001) and
(111)] substrates. The valence decreases with coverage and saturates at the value 2.75,
that might be due to charge transfer between adsorbate and substrate.
On the contrary, Andersen et al. (1988) report, from XPS study of Sm on a Cu (100)
single-crystal surface, that the mixed valence of Sm is of a heterogeneous nature. This
suggests that such a conclusion also applies to A1 [(111) and (100)] surfaces. This is

116

M. GASGNIER

also supported by Wieliczka and Olson (1990). The heterogeneity results from a divalent
surface layer and a trivalent intermetallic compound. However, according to Jaffey et
al. (1989), the observation of mixed valence associated with a single overlayer structure
as found for the SrrdCu (111) system, indicates that a homogeneous mixed valence can
exist. But, at other coverages, and after annealing, more than one phase will form, (SmCu
and SmCus) and heterogeneous mixed valence will be observed. These authors assert
that valency variations may have consequences for both the energy and intensity of the
Auger transitions. A correlation between LEED intensity and Auger intensity ratio during
Sm uptake at 300K is reported. A scheme of the structures observed in LEED after
annealing is given.
Andersen et al. (1988) explain the results of Ffildt andMyers (1983, 1984b, 1986) by
the fact that Sm and Cu or AI form intermetallic compounds. Stenborg and Bauer (1987a)
also discuss these results in terms of the average coverage and not of atomic distribution.
They observe that Sm has a pronounced tendency to form islands on Mo (110) surfaces.
They conclude, from low EELS measurements, that changes in the electronic structrure
do occur as a function of coverage but these cannot be interpreted unambiguously in terms
of a valency change. In the case of Sm monolayers deposited onto Mo (110) substrates,
Stenborg and Bauer (1987a, 1988) have observed, from EELS measurements, that the
two volume plasmon losses are respectively attributed to divalent and trivalent Sm.
2.2.4.2. Ytterbium.
2.2.4.2.1. The Yb/Ni system. Heterogeneous mixed-valence and compound formation
in ordered Yb/Ni (100) and (110) overlayers have been studied by Chorkendorff et
al. (1985a,b), Andersen et al. (1987) and Nilsson et al. (1987, 1988). From various
spectroscopic analyses the authors report structural investigations for films grown
at 300K or at 673 K. The results demonstrate the importance of surface compound
formation. It is concluded that, at 300 K the interface exhibits a simple layer-by-layer
growth, whereas at 673 K three types of coverage are observed: at low coverages Yb is
found to be absorbed on the Ni surface (hexagonal overlayer incommensurate with the
Ni (100) surface); at higher Yb coverages (less than a full monolayer) there is formation
of a surface intermetallic compound; and with further deposition a three-dimensional
crystallographic system growth (formation of the compound). As a function of the
temperature treatments it is demonstrated that Yb changes valency from +2 to +3 in the
bulk, whereas a stable divalent layer remains at the interface. Lastly, XPS measurements
show that the 5p level resonates strongly at 181 eV as a consequence of the 4d-4f
giant resonance.
2.2.4.2.2. The Yb/AI system. Spectroscopic and structural investigations of Yb/A1 (110)
and Yb/A1 (100) interfaces have been studied by Chorkendorff et al. (1985a), Nyholm et
al. (1984), Onsgaard and Ellegaard (1984), Onsgaard et al. (1984), Tibbetts and Egelhoff
(1980), and Egelhoff and Tibbetts (1980). In the case of the Yb/A1 (110) system it is
concluded that Yb in Yb-A1 alloys forms homogeneous-mixed-valence compounds, and
that island growth and/or diffusion play an important role. It is also found that Yb persists
in a divalent state at the surface independently of the concentrations. Two compounds as

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

117

YbAI2 and YbA13 can be formed. For the Yb/A1 (100) system it has been found that the
pure Yb surface layers exhibit only the valence two, whereas a mixed-valent state (YbA13)
begins after heating above 423 K. Further annealing above 623 K causes the Yb to revert
to the divalent state. In the case of single crystals Kaindl et al. (1982, 1983) report that
YbA12 has a mean valence of 2.4 and that the two surface layers are divalent.
i

~055

-CLOSEPAEKED~ "~ e
FROMCRYSTALDATAF

_-

@too~]l

<

20.50

,/

Yb / Mo(110)

045

i [1~ol

<~

>
LJ

cz 0.40

~(10x2)

LaJ
LLJ

10
1L5
2~0
DEPOSITION TlME(min)

25

Fig. 4. (a) Coverage as deduced from a LEED pattern as a function of deposition time: crosses, n x 2 structure;
solid circles, hexagonal structure. (b) Model of the 10x2 structure (letters a~t). One half of the unit mesh
is shown with the correct relative size of Yb and Mo atoms. (By courtesy of Prof. E. Bauer, Physikalisches
Institut, Univ. Clausthal, 3392 Clausthal-Zellerfeld, Germany).

2.2.4.2.3. The Yb/Mo system. Stenborg and Bauer (1987a,b, 1988) and Stenborg et al.
(1990) have studied the adsorption of Yb onto Mo (110) crystals. It is deduced from
LEED and different spectroscopic measurements that the number of nearest neighbors
may be changed from 0 to 12 going from a two-dimensional gas of individual atoms (onedimensional system), to a two-dimensional condensate with adjustable density and finally
to a metallic multilayer system. This permits one to link dimensionality and electronic
structure. From LEED patterns, the "LEED coverages" are obtained by assuming that cell
sites in the adsorbate lattice are occupied. The n x 2 and 10x2 structures are observed for
a "coverage" from 0.40 to 0.45, and then the hexagonal structure is formed (fig. 4a).
The coexistence range of the two structures is small. Different possible models of the
10x2 structures are shown in fig. 4b. The authors conclude that intermixing with the
substrate does not occur in the submonolayer range. Some results are given in terms of
the comparison & t h e 4fphotoelectron spectra relative to one Yb monolayer and to a thick
film: differences in line widths for various temperatures are caused only by vibrational
(of phonon) broadening. From EELS data the volume plasmon loss at 9.40 eV shows that
Yb is divalent in agreement with the results of Colliex et al. (1976) and Onsgaard and
Chorkendorff (1986).

118

M. GASGNIER

2.3. Crystallographicproperties
2.3.1. Introduction
Since about ten years few new data have been reported about the rare earth phase diagrams
(Gasgnier 1980). One can only mention the case of cerium for which some new transitions
have been observed at high pressure and high temperature (Gschneidner 1985).
Among the broad literature which deals with rare earth metals, as thin films, few
of them exhibit clear XRD and/or EDP results. One can notice, for instance, that the
patterns exhibited by O. Nakamura et al. (1988) and Hauser (1985) define pure Gd and
Tb thin films respectively. However, W.D. Schneider et al. (1983) assert that their Yb films
crystallize in the metallic (hop + fcc) phases without XRD proof. Likewise Kosak et al.
(1987) do not report crystallographic data in the case of Sc thin films. One should only
admit that the films are purely metallic as claimed by some authors (R.E Miller et al.
(1979) for instance). However, that is not always the case: Dudfis et al. (1985) have
proved from XRD, that the Dy and DyH2 phases can coexist. Moreover, the formation of
the sesquioxide cannot be rejected. Mahmoud and el Mandouh (1987) did not find any
evidence for the formation ofSc203 from XRD measurements on thick Sc films (see sect.
2.8).
Remark: In the field of structural investigations one must mention the work of
Kuzmenko et al. (1972, 1976, 1980, 1984a,b, 1986a,b) and Kuzmenko and Melnikov
(1982, 1988) who have studied the amorphous ~ crystalline transition in Yb films. It is
observed to occur under particular experimental conditions: application of a perpendicular
magnetic field, a flow of helium gas, an oblique deposition; and a short current pulse. The
most important parameter is the thickness of the sample. The transition is explained in
terms of avalanche (shock) crystallization. However no XRD or EDP data are exhibited
and the crystalline phases of the metal (Ice or hcp) are not defined. One must notice
that Haussler and Bauman (1980) using a parallel (to the surface) magnetic field do not
observe this transition at 4.2 K.
2.3.2. Epitaxial crystal growth
2.3.2.1. R/W systems. Melmed et al. (1984) and Ciszewski and Mehned (1984a-c) have
studied surface diffusion of Sin, Eu, Gd and Dy on tungsten by means of field emission
microscopy. The crystalline layers have been grown by vapor deposition onto either
W emitters or (001)-oriented W single crystal. Optimum growth is studied as a function of
the temperature (between 730 K for Eu and 1200 K for Dy). The nucleation sites and the
commonly epitaxial relationships are well defined by various field electron micrographs.
The epitaxial relationship most commonly observed is (00.1) R (Sm, Gd, Dy) [I (011) W
and (110) Eu [I (110) W. Reshentnikova and Yuldasheva (1979) have studied the emission
areas of Dy, Ho and Er deposited onto a point single crystal of W.
The adsorption of Gd, Dy and Sc on the (112) surface of a W single crystal has
been studied by EM. Gonchar et al. (1987) and Gorodetskii and Shevlyakov (1980),
respectively. Various kinds of structure growth have been observed as a function of the
ratio of the density of adatoms to the surface density of W atoms. They also found a

THE INTRICATEWORLD OF RARE EARTHTHIN FILMS

t 19

large difference in the initial heats of adsorption of Gd (close to that of La) and of
Dy. V.V. Gonchar et al. (1985) report the atomic structure of submonolayer films fo La
absorbed on the (001) or (111) faces of W. The authors have constructed a phase diagram
demonstrating the evolution from first-kind phase transitions to the hexagonal structures.
[It is also shown that films grow on (011) Cr surface exhibit rather isotropic structures,
whereas for (011) W and Nb surfaces they present anisotropic structures. Such differences
are discussed in terms of the electronic properties of the three different systems as
interactions of adatoms through d-electrons of the substrate.] Gorodetskii and Martynyuk
(1989) report that the threshold potential methods allow to detect Sc on polycrystalline
W surface, in amounts as small as about 5% of a monolayer.
Other work has been carried out by Mittsev and Mukhuchev (1982, 1983) who
investigate the lifetime dependence of Cs adatoms on film thickness of Tm and Sm
monolayers deposited onto W substrates. The adsorption of Cs atoms yields data on the
growth mechanism of R films on W surfaces. It is assumed first that the binding energy
of Sm atoms decreases weakly in the transition from a two-layer film to a three-layer one,
and second that the growth of a Tm crystalline phase occurs on the surface of the first
layer of atoms and that the concentration in the second layer required for the starting of
the growth is low. It is also observed that the lifetime of Cs atoms presents a nonmonotic
variation. The surface-coverage dependences of the heat of adsorption and lifetime of
Tm atoms on a (100) face of W have been studied by Burmistrova et al. (1983). They
conclude that theory and experiment are satisfactory in agreement and that the lifetime
is strongly influenced by binary and ternary collisions in the adsorbed film.
2.3.2.2. R/Re systems. Melmed et al. (1984, 1987) have investigated the epitaxial growth
from the vapor of Eu/Re systems by means of the field emission microscopy. They
assume that besides the classical bcc phase, a metastable fce phase can form. However
the latter, consisting of very small crystallites, is metastable. In the case of lanthanum
films on a (1010) face of rhenium, Lozovyi et al. (1984 and refs. therein), show that
there is a correlation between the surface-density dependence of the heat of adsorption
and structural changes. They compare the results with those obtained for other substrates
such as (112) W and Mo (Lozovyi et al. 1982).
2.3.2.3. R/V systems. From RHEED measurements Homma et al. (1987a,b), and Yang et
al. (1988) have shown that Ce (111) and Dy (00.1) single crystals deposited by MBE onto
V (ll0)-oriented films (-100nm thick), present structural changes as a function of the
thickness (from 0.3 up to 15 nm). The surface reconstruction patterns allow to define how
these patterns are correlated to the in- and out-of-plane crystal structures. It is observed
that the in-plane lattice constant of Dy is approximately 4% expanded relative to the bulk
one in the case of 0.4-5 nm thick films. The epitaxy is such that the (11.0) Dy axis is
found to be parallel to the (110) V axis. The same orientation is reported for the Gd/Nb
system (Kwo et al. 1986).

120

M. GASGNIER
5O0
450
4OO
350
3O0

B
d

250

200
150

q)

100
BULK
50
0

500

I000

!500
Thickness

2000

Fig. 5. Dependence of resistivity p on the thickness of Y films deposited at 5 x 10- 7 Torr. (By courtesy of Prof.
A.E. Curzon, Physics Dept., Simon Fraser Univ., Bumaby, Canada).

2.4. Electrical properties

2.4.1. Crystallinefilms
Curzon and Singh (1979) report anomalous resistivity behavior of yttrium thin films as a
function of the thickness (fig. 5). The results are explained in terms of faint contamination
(H2, 02) and of defect scattering. So, the results of Kosak et al. (1987), Loboda and
Protsenko (1981a,b), Loboda et al. (1980, 1982) and Protsenko (1988), relative to Sc
thin layers, must be examined with much wariness (see sect. 2.8).
The temperature and thickness dependence of the electric resistivity of Sin, Dy and
Tm thin films (25-370 nm thick) was measured between 4.2 and 300 K by Dud~is and
Fehrr (1984, 1987), Dud~is et al. (1985, 1986, 1987a-d, 1990), and J~inos et al. (1987).
For Dy films the Nrel temperature and the residual resistivity ratio increase both as
thickness increases. The thickness dependence of the spin-disorder resistivity is reported
in fig. 6. In the case of Tm samples it is observed that the Nrel temperature increases
(from 49 to 54K) and spin-disorder resistivity decreases (from 35 to 10~tff2m) as
thickness increases (fig. 7). Such results, and correlations with the crystalline orientations
(basal plane of crystallites parallel or perpendicular to the substrate surface) correct the
errors made by other investigators (see Gasgnier 1980). On the other hand the resistance
ratio as a function of the temperature exhibits different kinds of curves as the thickness
varies as measured for Tm films (fig. 8). Different magnetic transitions have been also
observed. One can conclude that numerous kinds of anomalies were caused by hydrogen
in solution in the metallic matric, in agreement with Dud~is (1991), and also by structure

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

121

~- 60
E

E
pj-,
5O
BULK

t40
I

iO0

200

300

t[nm]

Fig. 6. Thickness dependence of the spin-disorder resistivity of thin DX films. (By courtesy of Dr. J. Dud~s,
Technical Univ., Kosice, Czechoslovakia).
4O

60

-g

BULK

%
20

5O

/BULK

tO0

20O

3OO

-12

10

tInml

Fig. 7. Thickness dependence of N~el temperature TN and spin-disorder resistivity Pm of Trn

thin films (By courtesy of Dr. J. Dudfis).

100

T[K]

200

300

Fig. 8. Resistance ratio R]R4. 2 as a function of temperature

for 4 Tm films: (1) 16nm thick; (2) 89nm, (3) 189nm;
(4) 364nm. (By courtesy of Dr. J. Dud~s).
defects a n d internal stress. Such parameters, and above all the hydrogen c o n t a m i n a t i o n
are either omitted or taken into account separately. Therefore the interpretations are often
in d i s a g r e e m e n t with each other.
2.4.2. Amorphoussamples
K u z m e n k o a n d M e l n i k o v (1982) and K u z m e n k o et al. (1986b) have shown that on heating
a m o r p h o u s Y b films the electric resistivity falls abruptly at 15 K, corresponding to the

122

M. GASGNIER

amorphous --~ crystalline transition. When the sample reaches a critical thickness the
resistance also decreases rapidly (Kuzmenko et al. 1984a). In another paper Kuzmenko
et al. (1984b) report that amorphous Yb films exhibit a positive magnetostriction which
decreases sharply with increasing temperature. They conclude that a superconducting
transition may occur at low temperature.
2.5. Magnetic properties
2.5.1. Crystalline state
2.5.1.1. Thin oriented and polycrystalline R films on oriented substrates.
2.5.1.1.1. W (110) substrates. Structural magnetic and electronic excitations of epitaxial
ferromagnetic Gd (00.1) films deposited onto W (110) clean single crystals have been
studied by Weller and Alvarado (1985, 1986, 1988), Weller and Sarma (1986) and Weller
et al. (1985a-c). Such substrates provide high surface mobility of adsorbate atoms, and
no Gd-W compounds are formed even at 723 K. The growth habit and the surface
geometry are generally characterized by AES and LEED for film thicknesses from 0
up to 100 nm. Spin-polarized LEED and magneto-optic Kerr effect allow to demonstrate
the existence of a surface-enhanced magnetic order. Clean Gd surfaces are found to
order magnetically at temperatures up to 22K above the bulk Curie point (293 K).
However this surface temperature falls to the bulk one when the Gd surface is
contaminated by 02, H2, CO or CO2 (even under a vacuum of about 10-8 Pa). Therefore,
as also reported by Kolaczkiewicz and Bauer (1986), extreme care must be taken to clean
the substrate and the evaporant. The main results given by Weller and Alvarado (1985,
1986, 1988) are as follows. First, the 4f spins of the surface are not ferrimagnetieally
coupled to the bulk moments. Second, films grown onto substrates heated at 723 K exhibit
a remanent surface magnetization at 200 K. Third, the 4fT(5d6s)3 --~ 4f8(5d6s) 2 interband
transition has been detected. This 4f ~ 5d mechanism has been previouly investigated
by Mauri and Landolt (1981) in the case of polycrystalline films and Reihl and Himpsel
(1982) for (00.1) films by means of spin-polarized photoemission. Mauri and Landolt
(1981) claim that the transition is a strong quasielastic one-electron transition from the
4f 8 state into conduction states (resonant scattering mechanism), while Reihl and Himpsel
(1982) explain the data by taking into account the Gd band structure and/or escape cone
rules.
On another part, Farle and Baberschke (1987) and Baberschke et al. (1987), from ESR
studies, claim that the ferromagnetic ordering temperature is a function of the nearestneighbor coordinations. For N = 12 they obtain the bulk Curie point; for N = 9, as a semiinfinite system, the surface Curie temperature is 22 K higher; and for N = 6, as a single
monolayer, this temperature is 22 K lower. Such results allow one to detect a fraction of a
Gd monolayer on a metal surface in the paramagnetic state. However, the Ising model of
phase transitions in ultrathin films reported by Aguilera-Granja and Mor~n-L6pez (1990)
fails to explain such results.
The theoretical determination of antiferromagnetic coupling and surface state for
Gd (00.1) has been studied by Wu and Freeman (1991). It is shown that the surface Gd

]23

THE INTRICATE W O R L D OF RARE EARTH THIN FILMS

co

'

'

'

'

'

'

'

'

_o

' |

270

300

20 -

'

Tb (O001)/W (IlO)

15--

_N
~

10--

ft.

TCbTNb
12o

15o

18o

21o

240

TEMPERATURE (K)

Fig. 9. Electron spin polarization P (%) of the topmost surface layer of 5 nm-thin Tb (00.1)/W (110) films as
a fimction o f temperature. (By courtesy of Dr. C. Rau, Dept. Physics, Rice Univ., Houston, TX, USA).

layer expands outwardly by 6% and is coupled antiferromagnetically with the underlying

ferromagnetic bulk. The existence of this surface state has been established by D. Li et
al. (1991) for 2-4 Gd monolayers (0.6-1 nm) onto W (110) substrates. This indicates the
spatial locality of the surface state. It is added that this surface state can be correlated
with the development of the Gd bulk band structure which occurs with increasing
film thickness as Gd films grow. Salas and Mirabal-Garcia (1989, 1990a) have studied
another parameter as the Hopkinson temperature. The "Hopkinson maximum" is the local
maximum which the initial magnetic susceptibility presents just below the Curie point
(~3 K). The authors claim that Weller and Alvarado (1988) have observed this maximum.
But it cannot be detected whether the samples are contaminated. From other works,
relative to "butterfly" susceptibility measurements, Salas and Mirabal-Garcia (1990b)
concluded to the incoherent character of the magnetization processes at low field in the
vicinity of the Hopkinson temperature.
Farle et al. (1985, 1989) have investigated the uniaxial magnetic anisotropy of
monolayers and a 8 nm thick epitaxial Gd layer near the Curie point. For monolayers the
magnetization is aligned normal to the surface and is up to 10 times larger than the bulk
anisotropy. In zero external field, magnetization lies completely in the surface plane.
In the case of Tb (00.1) films (5 um thick) epitaxially and homogeneously grown onto
W (110) crystals, Rau and Jin (1988) and Rau et al. (1990) find that at the topmost surface
layer a long-ranged ferromagnetic order exists up to a surface Curie temperature of 250 K
(above the Curie and Nrel points, 220 and 228 K respectively). The results are shown
in fig. 9 for electron spin polarization (% as a function of the temperature). Moreover a
short-range ferromagnetic order at Tb surfaces extends up to 300 K.

124

M. GASGNIER

2.5.1.1.2. Fe (100) substrates. Taborelli et al. (1986a,b) and Landolt et al. (1986) have
studied the magnetic coupling of a very thin Gd adlayer on a Fe (100) surface. They
conclude that individual Gd atoms chemisorbed on Fe strongly couple antiferromagnetically to the substrate. A Gd monolayer induces a Gd-Fe surface Curie temperature
near 800 K (far from those of Gd and Fe, -1040K). A film with a thickness of 3 nm
(12 Gd atomic layers) is found to order magnetically below the Curie point (Gd-Gd
coupling), with an opposite magnetization to the Fe substrate. Spin-polarized AES allows
one to investigate the 4d core-hole transitions. Carbone and Kisker (1987) have used
spin-resolved photoemission to study the electronic structure of the ferrimagnetic Gd/Fe
interface. They observe for the first time a high Gd 4f-electron polarization for Gd films
up to one monolayer thickness. The main effect of the Gd adsorption on the Fe derived
valence bands is a depolarization. For thicker films (3.5 nm) the 4f polarization is small.
In the case of Nd and Dy atoms adsorbed onto Fe surfaces, Carbone et al. (1990) have
invoked Hund's rule to demonstrate that the Dy moment is antiparallel to the Fe moment,
whereas the Nd moment is parallel to the Fe moment.
2.5.1.1.3. Cu (100) and Ni (111) substrates. LaGraffe et al. (1989a-c) and Dowben et
al. (1989) have investigated the magnetic properties of Gd/Cu (100) systems. It is shown
that Gd (1-6 monolayers) exhibits a local magnetization component along the surface that
decreases for thicker overlayers. The 5d exchange splitting changes from 1.1 to 0.6 eV as
thickness increases. From 170 up to 340 K there is little or no interdiffusion; that occurs
only at higher temperature (above 370 K). Using valence-band emission it is observed that
the binding energies of both the Cu 3d and the state symmetry of the 5p3/2 and 5d levels
of Gd have been determined. It is concluded that a number of many-electron processes
occur, which makes photoemission of R overlayers particularly complex and substantially
alters the photoemission cross-section.
LaGraffe et al. (1990a) report that the Tb/Ni (111) system exhibits a magnetic
ordering of the Tb overlayers as measured from the Tb 5p3/2 branching ratio. This
ordering increases, reversibly, at temperatures well below the Curie point. The Tb
magnetization axis is widely in-plane for very thick films, whereas it is found normal to
the film for the Gd/Cu (100) system. In the Gd/Ni (111) system LaGraffe et al. (1990b)
observed a pronounced interdiffusion of Ni at tempera~res as low as 150K. Such a
phenomenon is equivalent to that noted previously for the Yb/Ni (111) interface (LaGraffe
et al. 1985, 1988).
2.5.1.2. Polycrystalline films.
2.5.1.2.1. Ferromagnetic order at R surface. Rau (1982, 1983) and Rau and Eichner
(1981a,b, 1986) have shown that 500nm thick Gd films, deposited onto various
substrates, show evidence for ferromagnetic order at the surface. From electron-spin
polarization measurements they observe an almost linear variation of the electron-spin
polarization between 160 and 315K (polarization = 0). It is found that the surface
Curie temperature (310K) is far above that of the bulk (293 K). This indicates uniaxial
anisotropic magnetic exchange interactions at the topmost surface of the film. Such results
are in agreement with those of Weller et al. (1985c).

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

125

Cerri et al. (1983) report first that the spin polarization of photoelectron decreases
linearly between 20 and 300 K and that this parameter strongly deviates from the bulk
magnetization. As hydrogen influences the magnetism of Gd, the authors observe that a
submonolayer of hydrogen strongly reduces the spin polarization at 20 K, from 70% for
the clean surface to 45% and 30% for 0.5LH2 and 1.0LH2, respectively.
2.5.1.2.2. Ferromagnetic films. The magnetic properties or rf sputtered Gd films have
been studied by O. Nakamura et al. (1988). They report that the variations of the estimated
Curie temperature (from 273 to 293 K, i.e., below the bulk value) of films (900 nm thick)
annealed up to 830K, are correlated with the relaxation of lattice imperfections. The
Curie point of 273 K is measured for unannealed films. However it is noted, first that
after annealing the lattice constants of the hep structure become smaller than those of
the bulk; second, that at interfaces the oxygen content increases; and third, that the
hydrogen amount is not known (see Cerri et al. 1983). Ferromagnetic Gd foils for
transient field experiments have been prepared by Maier-Komor (1989). It is observed that
samples either contaminated by hydrogen, or grown onto substrates (overall Ta substrates)
heated at 700 K have no or poor magnetic quality. To obtain good ferromagnetic foils
a refrigerator cryopump is used. Bachurin et al. (1981) in the case of sputtered Gd
films, observe the ferromagnetic --+ paramagnetic transition near 300 K. Such a result is
observed from temperature dependence of the relative sputtering coefficient of Gd: at the
Curie point this parameter increases by 10-15% in the course of the annealing from 233
up to 323 K. It is observed that in the case of Dy films no transition occurs.
2.5.1.2.3. Multilayer (Er, Tm/Lu)n-typefilms. Coherent multilayers have been synthetized
by sputtering by Lowe et al. (1985). Coherence has been determined from XDP
for film thicknesses between 0.8 and 1.8 ~tm. Lowe et al. observed suppression of
the antiferromagnetie-ferromagnetic transition temperature and a decrease of the Nrel
temperature with decreasing modulation thickness. These phenomena have not been
elucidated. It is also observed that magnetic interactions between Tm layers separated
by Lu layers is small compared to the interactions within the Tm layers.
2.5.1.2.4. Theoretical models. Few new papers have appeared on this subject. UrbaniakKucharczyk (1986, 1988) has studied the magnetic contribution to electric resistivity
in thin Gd films. The numerical results show the resistivity dependence on the film
thickness (as monoatomic layers) and the surface properties. The surface deformation
of helimagnetic order in Dy, Ho and Tb0.9Ho0.1 films with a thickness of 10-25 atomic
(00.1) planes has been investigated by Zajac (1977).
2.5.2. Amorphous films
Hauser (1985) has established that a spin-glass transition occurs in disordered Tb films
(0.27-1.50 ~tm thick). The spin-glass interaction depends strongly on the degree of
disorder of the films as shown by the wide variation of the spin-glass freezing temperature
and by anomalies in the susceptibility versus temperature curves. However, above the spinglass freezing temperature there is no sign of ferromagnetism, and films deposited at 77 K
do not exhibit any significant spontaneous magnetization. One notes that surface samples,

126

M. GASGNIER

after exposure to air, are considered to have transformed to TbO2. But no crystallographic
or spectrographic proofs were given to support such an assertion.
2.5.3. Magnetic domain structures
The magnetic domain structures of Gd and Tb single crystals, between 77 and 293 K,
have been studied by Comer and Saad (1977), Comer et al. (1978, 1980) and R.L. Smith
et al. (1980a,b). The conventional Bitter wet colloid technique was used for the Curie
temperature observations and a dry colloid method (carried out by evaporation of Fe on
the sample) for the low-temperature experiments. Various kinds of domains are observed
as a function of the applied magnetic fields, of the crystallographic planes (a, b or c axes)
and of the temperature. Some are clearly identified as parallel ferromagnetic domains with
well-defined contrast, other ones are complex and difficult to interpret. Particularly, some
patterns observed at 77 K exhibit, in the basal plane (00.1), a network of domains with
hexagonal symmetry (Comer and Saad 1977, Comer et al. 1978). However, later on,
this configuration was never observed. Chapman et al. (1981) have studied by Lorentz
microscopy the domain structures of thinned Tb single crystals. The samples have been
observed in a range of temperatures between 150 and 230K, i.e. in the range of the
magnetic transitions of Tb. By heating and cooling, the configuration of the magnetic
domains (parallel spaced 180 domain walls) seems to be well linked to these transitions:
the paramagnetic --+ helical antiferromagnetic transition at 230K and a ferromagnetic
transition at 220 K.
2.6. Optical properties
2.6.1. ~ r y low energy range (below 7 eV)
The results reported by numerous research groups show broad differences. In fact, as
noticed previously (Gasgnier 1980) only a few papers deal with contamination of the
films (see, for instance, R.E Miller et al. 1979, Krizek and Taylor 1975, and Apostolov
et al. 1983).
2.6.1.1. Scandium films and single crystals. The optical properties of single-crystal and
thin hcp Sc samples have been performed from 0.20 to 5.50 eV by Weaver and Olson
(1977a,b) who notice that there are close likenesses of the electronic properties of Sc,
, and heavy lanthanides. Sigrist et al. (1987) report the main optical characteristics
(reflectivity, real and imaginary parts of the dielectric function, electron energy loss
function) carried out for vacuum measured freshly deposited and air exposed films. These
parameters are compared to those obtained for a single crystal at 4.2 K (Weaver and Olson
1977a,b). Some slight discrepancies appeared between the results, but it is asserted that
the four structures observed below 5.50 eV are assigned to pure Sc metal. On another part,
Tougaard and Ignatiev (1981), from EELS studies, report that Sc (00.1) single crystals
exhibit a peak at 5.80eV and a strong shoulder at 4.00eV; these are interpreted as
excitation of valence band electrons. Onsgaard et al. (1980) using reflection energy-loss

THE INTRICATEWORLDOF RARE EARTHTHIN FILMS

i27

spectra observed two structures for Sc (00. l) single crystals. They are located at 2.30 and
5.50 eV and are interpreted as interband transitions. Lastly, it is difficult to analyze the
results of Mahmoud and el Mandouh (1987) because their films are highly contaminated.
2.6.1.2. Yttrium. Weaver and Olson (1977a) notice remarkable resemblance of the
electronic properties of Sc and Y as single crystals and films (Weaver and Olson 1977b,
Weaver and Lynch 1973). In the case of thin films, Apostolov et al. (1983) take care
to note that hydrogenation and oxidation can play an important role in the energy range
above 3 eV. They observe four direct optical transitions between 1.60 and 3.10 eV, in good
agreement with electron band structure calculations.
2.6.1.3. Europium, gadolinium, terbium, dysprosium, holmium, erbium, thulium and
ytterbium. Apostolov et al. (1981) have studied the transmission spectra of Tb films
deposited at 77K onto glass substrates over the energy range 1.60-3.10eV. During
evaporation and measurements the pressure was around 10-6 Pa. A direct evidence for
seven optical transitions is explained in terms of electron band structure calculations.
The magnetic ordering effects in the transmission spectra of ferromagnetic Tb are also
discussed.
The optical conductivity, established by polarimetry, for Tb, Dy and Ho films has been
reported by Ahmed-Mokhtar et al. (1982) over the energy range 0.50-5.00 eV at different
temperatures (from 30 to 300 K). The results are compared to those given by Krizek and
Taylor (1975) carried out by ellipsometry. It is evident that wide differences appear, and
that it is difficult to fit the results. However, Ahmed-Mokhtar et al. (1982) report that for
Tb, changes in the magnetic state are observed below 1.50 eV which give rise to an optical
resonance (effective s - f exchange energy); while for Dy such changes have little influence
on the optical properties; and for Ho unexplained anomalies occur in the conductivity
and the refraction indices. One must add that the thickness of the films does not exceed
some 30 nm, and under these conditions probably contaminations occur, but they are never
taken into account. Reflectance, dielectric constants, optical conductivity and interband
transitions of Gd and Dy films have been studied by Quemerais et al. (1981) between 2
and 30eV. The authors compare their results to those reported previously. Between 2
and 10 eV they conclude that the principal interband transitions are p-+ d, and d-+ p
with a slight participation from s ~ p.
The thermally modulated reflectance of Gd, Tb and Dy films (100 ~m thick) has been
measured by R.E Miller et al. (1979), over the range 1.45-3.20 eV, at 315 and 97K. The
experimental procedure (evaporation under a vacuum of 10-8 Pa) excludes the formation
of the hydride and the oxide (Rahrnan Khan and Miller 1979). At 97K, pronounced
structure develops: 1.60, 2.00 and 2.80 eV for Gd; 1.70 eV for Tb; and 2.20 eV for Dy.
They noted that thermoreflectance anomalies occur at the ferromagnetic Curie point of
Gd and the N~e! points of Tb and Dy. The structures are interpreted according to different
hypotheses and compared to those of Krizek and Taylor (1975). For Tm films evaporated
under a vacuum of 10-7 Pa, Onsgaard et al. 0983) report EELS measurements at low
energy range. At 4.50 eV they observe an intense peak attributed to a valence band to

128

M. GASGNIER

conduction band transition. Idczak and Zukowska (1981) give some optical properties
(reflection factor, refractive index) of thick (~350um) thermally evaporated Yb films.
They found that the optical constants increase monotically with increasing wavelength of
the light.
Bauer and Kolaczkiewicz (1985) have investigated by low EELS the edge structures of
Eu, Gd and Tb single-crystal films or atomically dispersed layers adsorbed on W (110) and
W (112) surfaces. Several few intense structures and a sharp dominant loss peak at
about 4 eV have been recorded. The latter has been attributed to spin exchange excitation
of the 4f level. A comparison to XPS results led to a re-consideration of the interpretation
of the 4f XPS binding energies.
2.6.2. Low energy range (7-70eV)
2.6.2.1. Scandium films and single crystals. In the case of clean single crystals Tougaard
and Ignatiev (1981) and Onsgaard et al. (1980) report from EELS experiments that the
surface and volume plasma are located at about 8 and 12 eV respectively. (The bulk
plasmon corresponds to the collective excitation of the conduction electrons). Other
experiments carried out with thin films yield higher energy values for the volume
plasmon: 14.00 eV (Brousseau-Lahaye et al. 1975); 13.50 eV (Frandon et al. 1980), and
12.90 eV (Cukier et al. 1980). The calculated value is reported to be equal to 12.90 eV; it
seems clear that the discrepancy proceeds from the purity of the films. (The bulk plasmon
is shifted towards higher energies for hydrided and oxidized films as demonstrated by
Colliex et al. (1976) and Onsgaard et al. (1980).) At higher energies, Tougaard and
Ignatiev (1981) and Onsgaard et al. (1980) report for excitations of 3p (M2-3 threshold)
and 3s electrons to empty states in the valence band, and collective excitations of
"inner" 3p electrons the following values: 29-31 eV, 54 eV and 42.50 eV, respectively. The
M2-3 threshold has been located at similar values by Brousseau-Lahaye et al. (1975) and
Frandon et al. (1980), whereas Cukier et al. (1980) report a higher energy: 38 eV. The real
and imaginary parts of the dielectric constant and the absorption coefficient have been
reported by Brousseau-Lahaye et al. (1975) and Frandon et al. (1980).
In the case of Sc films exposed to air, Sigrist et al. (1987) observed in the vacuum
ultraviolet region several structures located at 7.60, 10.30 and 21.00eV. These authors
discuss such peaks according to the formation of a thin transparent surface oxide layer.
In this way, it is not possible to assign the peaks at 7.60 and 10.30eV to the metal or to
the oxide layer. However, the peak at 21.00eV, observed also by Frandon et al. (1980)
and Cukier et al. (1980) for Sc203 thin films, is related to this oxide layer.
2.6.2.2. Yttrium. Chee et al. (1979) have studied the optical behavior of Y films (200 nm
thick) in ultrahigh vacuum. They report the refractive index and extinction coefficient, the
reflectance, the real and imaginary parts of the dielectric constant and the volume and
surface loss functions as a function of the photon energy. They observe that oxidation
(or hydrogenation?) and nitrogen gas (as admitted deliberately in the vacuum chamber)
can strongly affect the optical properties. The results are compared to those of Weaver

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

129

and Olson (1977b), and it is concluded that there is no resemblance between them: the
interband transition at 7.00 eV, the surface plasmon at 9.40 eV and the volume plasmon at
11.80 eV reported by Chee et al. (1979) do not correspond to the values given by Weaver
and Olson (1977b) as: 7.70 eV, 11.50 eV and 12.50 eV respectively. Other researchers
report the following values for the volume plasmon: 12.50 eV (Brousseau-Lahaye et al.
1975), 12.80eV (Frandon et al. 1980); 13.00eV (Cukier et al. 1980). (The calculated
value is 11.20 eV.) At higher energies the 1'42-3 threshold has been recorded at 36.00eV
(Cukier et al. 1980) and 35.00eV (Brousseau-Lahaye et al. 1975, Frandon et al. 1980).
The real and imaginary parts of the dielectric constant and the absorption coefficient as
a function of the energy have been deduced by the latter authors.
2.6.2.3. Lanthanum and cerium. Cukier et al. (1980) have reported on LaF3 thin
films studied by EELS, that the volume plasmon of La is located at about 12eV. The
02-3 threshold (5p electron excitations) occurs at 22-23 eV. Miyahara et al. (1982) assert
that the absorption spectra of metallic La and Ce films present two maxima close to 10 eV
and 25 eV. The authors compare their results to those published previously and discuss
the origins of small structures in terms of intrinsic properties or of contamination of the
samples. The higher energy peaks are linked to the 5p--+ 5d transition.
2.6.2.4. Gadolinium, dysprosium, thulium, and lutetium. The positions of volume
plasmon energy of Gd and Dy (calculated and experimental) reported by Quemerais et
al. (1981) differ from those given in other works. Their calculated values of 11.20 and
ll.40eV, and 12.10 and 12.20eV as measured for Gd and Dy, respectively, are lower
than the values reported by Colliex et al. (1976) and Cukier et al. (1980). This indicates
that the samples of Quemerais et al. (1981), evaporated under a vacuum of 10-6pa,
are purer. In the same way the volume plasmon energy (~11 eV) given by Miyahara
et al. (1982) reflects a nearly pure metallic state. The peak relative to the 5p electron
excitations is located at 33-35 eV according to the above-mentioned authors. However,
different interpretations have been given to explain this edge. So, Quemerais et al. (1981)
show that the shake-up structure comes from a 5p hole and a 4f open shell coupling,
whereas Murgai et al. (1988) from photoemission measurements, report that the 5p-5d
intershell interaction results in an enhancement of the 5d emission (at 32 eV) above the
5p3/2 and 5pl/2 core level thresholds at 21 and 27 eV respectively. The 02-3 threshold has
been observed by Cukier et al. (1980) and Miyahara et al. (1982) between 20 and 30eV;
this one is more intensive for GdF3 and DyF3 than for LaF3, due to the presence of the
4f electrons. For Tm films evaporated at a pressure below 10-7 Pa, Onsgaard et al. (1983)
observe four main peaks, located at 8.90 eV (surface plasmon), 12.50 eV (bulk plasmon),
40.50 eV (5p ~ 5d transition) and 57.00 eV (5s ~ conduction band transition). One must
note that the higher volume plasmon value given by Colliex et al. (1976) is surely caused
by contamination of the film. (The calculated value is estimated at ~11.90 eV.)
In the case of clean Lu (00.1) single crystals, surface plasmon (9.80 eV), bulk plasmon
(13.20eV), 02-3 threshold (30-33 eV) and 5p--+ds transition (43.30eV) have been
recorded by Onsgaard et al. (1980). The bulk plasmon value is in agreement with the

130

M. GASGNIER

theoretically expected values and the results of Colliex et al. (1976). (One must add
that lutetium films are less sensitive to contaminants than those of other R elements).
Onsgaard et al. (1980) notice that the Lu 5p ~ 5d strong resonant transition is similar to
the Sc 3p --~ 3d one. Such a result has been also suggested by Murgai et al. (1988) in the
case of Gd thin films, and appears to occur in all R elements possessing d character in
the valence band.
2.6.2.5. Samarium. The ultraviolet absorption spectrum of Sm thin films exhibits three
peaks at 9.00, 16.00 and 30.00eV (Miyahara et al. 1982). The origin of the second one
has not been assigned. (However it might be linked to that for the bulk plasmon of the
hydride.)
Remark: The Sm203 bulk plasmon is located at 15.20 eV (Gasgnier and Brown 1983),
and that relative to SmH2 is located at higher energy (Colliex et al. 1976).
On their part den Boer et al. (1988) observe that Sm monolayers deposited onto
Nb substrates are initially purely divalent. The photoemission spectrum exhibits two peaks
at about 5 eV and 25 eV, which are characteristic of the Sm2 valency state. One hour after
deposition and exposition to oxygen, the spectrum fully changes and two peaks (at about
10 and 20 eV) are observed, which are related to mixed-valent or trivalent Sm. One must
notice that after exposition to 02 the volume plasmon peak is shifted towards higher
energy (~14 eV). The 5p --~ 5d resonance has been measured, and it is evidently sensitive
to the occupancy of the Sm 5d state.
2.6.2.6. Europium and ytterbium. The 5p--* 5d transition (5p3/2-5pl/2 peaks) for Eu and
Yb has been studied from photoelectron energy distribution curves by G. Rossi (1987),
G. Rossi and Barski (1985, 1986), G. Rossi et al. (1984, 1985) and Yeh et al. (1984). As
a preliminary remark one must note that, first, Yb films (~50 nm) are considered to be
purely fcc, and second, possible contaminations are never taken into account. However, the
main results indicate that strong 5p hole --* 5d electron interactions occur in the electron
emission processes. The two metals show the same resonance effect above the threshold
of the 5p multiplet. Miyahara et al. (1982) for Eu thin films observe the presence of
small structures near the onset of the 5p absorption. These are discussed in terms of EuO
contamination.
2.6.3. Medium energy range (70-600 eV)
2.6.3.1. Scandium and yttrium. Frandon et al. (1980) have recorded the L2-3 edge
(2p electron excitations) of Sc at about 405 eV, and the M4-5 (3d electrons) and M2-3
(3p electons) edges of Y at 155 and 300 eV. The spin-orbit splitting of the 2p levels in
Sc is measured to be about 5 eV, while for Y the spin--orbit splitting of the 3d electrons
(2 eV) is not observed.
2.6.3.2. Lanthanum. Moser et al. (1984) have compared the N4-5 edge (excitation of
the 4f3/2 and 4f5/2 electrons) recorded by EELS and XPS. The former exhibits a rich

THE INTRICATEWORLDOF RARE EARTHTHIN FILMS

131

collection of structures extending from 97 up to 140 eV (large electrostatic coupling

between the 4d and 4f shells in the final 4d94f 1 state). The intensity of these multiple
structures varies as a function of the primary electron energy; i.e., the intensity increases
with decreasing excitation energy. This is due to dipole-forbidden transitions. From LaF3
thin films, Cukier et al. (1980), locate the N4-5 edge at about l l 9 e V
2.6.3.3. Cerium. For cerium films crystallized as the y-phase (fcc), L.I. Johansson et
al. (1978) report a giant enhancement of the valence-band photoemission intensity. The
photoemission peak is located at 120eV. The effect is interpreted as being due to
autoionizafion decay following the 4d ~ 4f transitions.
2.6.3.4. Gadolinium, thulium and lutetium. 4d-+ 4f resonant photoemission and Auger
spectra of Gd have been carried out by Gerken et al. (1980, 1981). Several multiplet
structures are observed close to the main photoemission peak located at 148 eV Some
Auger structures are identified as due to the direct recombination of 4d94f 8 states with
the 4f and valence electrons, whereas for photoemission they are identified as originating
from excited quintet configurations of the 4f 6 state reached by the Auger decay of the
resonantly excited 4d94f 8 configurations.
For Tm this 4 d - 4 f resonance has been investigated by Egelhoff et al. (1981). With
quite good resolution the authors observe a shoulder at 172 eV and a peak at 177 eV in
the photoelectron spectra. The autoionization resonance is explained according to a decay
channel processes. Onsgaard et al. (1983) observe two structures at 170 and 178 eV and
notice that theoretical calculations lead to an energy splitting equal to 11 eV. For Lu,
Onsgaard et al. (1980) report that the 4d5/2 and 4d3/2 transitions occur at 203 and 215 eV
respectively. The energy separation reflects the spin--orbit splitting.
2.6.3.5. Ytterbium. Another kind of experiment has been reported by Hecht et al. (1984).
They investigate the 4f5/2 and 4f7/2 photoemission peaks which can be resolved into two
pairs of doublets at a kinetic energy of 100eV. They conclude that the doublet results
from surface and bulk emission. The former is interpreted as an intrinsic binding energy
shift and is not due to a contaminant as claimed previously by Alvarado et al. (1980).
Different models are discussed to describe this result.
2.6.4. High energy range (beyond 800 eV)
From XPS measurements the 2pl/2, 2p3/2 and 3d3/2, 3d5/2 doublets of Sc and Y have been
recorded at 850-855 eV and 1090-1098 eV, respectively, by Reichl and Gaukler (1986,
1988).
The major part of these studies has been carried out for La thin films by Kanski
and Nilsson (1979), Kanski and Wendin (1981), Esteva et al. (1983a,b) and Moser
et al. (1984); they were extended to Ce and Gd by Esteva et al. (1983a,b). The
3d---~4f (3d5/2 and 3d3/2 main lines or M4-5 threshold) spectra have been recorded by
various experimental methods: XAS, XPS, electron-excited-Auger (and stimulated-X-ray)

132

M. GASGNIER

appearance and EELS. One of the most interesting results looks at the disagreement of
2-3 eV in the energies between XPS and XAS (Esteva et al. 1983a). Such a discrepancy
is discussed in terms of strongly localized valence electrons. In the same way, Moser
et al. (1984) notice that EELS and XPS exhibit a similar shift which is interpreted as
an indication for the breakdown of dipole selection rules in EELS when the excitation
threshold is approached (Matthew et al. 1983). Esteva et al. (1983b) have shown that the
multiplet structures of La and Gd are only detected over a limit thickness of about 30 urn.
Remark: The M4-5 thresholds of the R elements have been studied by EELS and XAS,
by the way of the sesquioxides, intermediate oxides and dioxides (Ce, Pr and Tb). (Colliex
et al. 1985, Manoubi et al. 1990, Gasgnier et al. 1989.)
2.7. Applications relative to metallic films and foils

2.7.1. The piezoresistance response of Yb foils

To measure the amplitude of the pressure induced inside a material submitted to a shock
wave, it is possible to use the piezoresistivity response of metals (as ytterbium) or alloys
(as manganin). In the ease of Yb numerous studies have been done by Y.M. Gupta
(1983a,b), S.C. Gupta and Gupta (1985), Chen et al. (1984), Brat and Gnpta (1986), Grady
and Ginsberg (1977), S.A. Miller (1988), Aidun (1986) and Bosca et al. (1981). The metal
is used either as a thin deposited film or a foil (50 ~tm thick). Although the laboratory
results promised to be very interesting for the fabrication of gauges, it is noteworthy
that to our knowledge no patent was taken out. It seems that several parameters restrict
industrial application: formation of the dihydride (YbH2), coexistence of the hcp and
fce phases (Bosca et al. 1981), stress measurements below 200 Pa, difficulty to interpret
lateral measurements, resistance change for parallel and perpendicular measurements
(S.C. Gupta and Gupta 1985). The whole of these results show that simple analyses are
not adequate and, therefore, further developments are required (Y.M. Gupta, 1983a). One
can add that the lack of crystallographic data and chemical analyses is truly prejudicial
to interpret the results. In this way, numerous authors do not refer to the high-pressure
crystallographic transition of ytterbium.

2.7.2. Diffusion barriers: Sm and Yb metal interfaces

It has been shown that ytterbium (and also samarium) have a unique reactive behavior
on mercury-cadmium-telluride surfaces, due to the large rare earth affinity for both Hg
and Te. They act as effective diffusion barriers at the interface between HgCdTe and
prototypical reactive metals (A1, In or Cr). Ytterbium was found to prevent A1-Te reaction,
and to reduce the Al-induced mercury depletion of the substrates, and the outdiffusion
of Te. The results of Raisanen et al. (1988), Wall et al. (1987, 1988) and Franciosi et al.
(1988) show that the diffusion barrier effect is the result of the superior thermodynamic
stability of the HgCdTe/Yb reaction products at the interface.

THE INTRICATEWORLDOF RARE EARTHTHIN FILMS

133

2.7,3. Tunnel junctions: Yb/oxide/p-InAs

In order to obtain devices with negative differential conductance, Kunze and Kowalsky
(1988) have used ytterbium as a film gate, The low work function of Yb compared to the
electron affinity of InAs increases the surface field. And, the low resistive tunnel junction
serves as ohmic contact to the inversion layer.
2.7.4. Magnetic deuices
A Japanese patent (K. Takeda et al. 1989) mentions the use of polycrystalline Gd film
(with a lower Curie temperature than a single crystal) after annealing treatments, as heatsensing switches or thin film inductance devices.
2.8. Metallic or contaminated films or fine particles? Remarks about the fcc phase
Once again one must draw attention to the high reactivity of rare earth metal, as thin
films, with gases as H2, 1-120, 02, CO, CO2, CnHn, N2, .... In this way the purpose
of this section is to indicate that a lot of results published by several research groups,
seem to be doubtful. One may refer to a paper of Curzon (1984) who demonstrates, by
means of a quadrupole gas analyzer, that the films are at first hydrided and not oxidized
(see sect. 2.9).
The first series of results concerns the electrical properties of three R elements (Y,
Sm and Yb) reported by Ashrit and Angadi (1980a, b, 1981, 1982, 1983), Ashrit et al.
(1980, 1984) and Angadi and Ashrit (1981a--c, 1983a,b). The main purpose of these
papers was to study the electric resistivity, the temperature coefficient of the resistance
(TCR) and the thermoelectric power of films deposited trader a poor vacuum of about
10-3-10-4pa. The variations of these parameters were studied as a function of the
thickness, annealing, aging, substrate temperature, deposition rate and pressure in the
vacuum chamber. Since the authors do not report any XRD or EDP they cannot define
the crystallographic structure of their films. The results relative to electric resistivity
were analyzed using the Fucks-Sondheimer and Mayadar-Shatzkes theories, i.e., in
terms of pure metallic films. One must notice that these authors sometimes mention
that possible contamination by hydrogen is possible (Angadi and Ashrit 1983a, Ashrit
and Angadi 1982). Rajora and Curzon (1984) have commented on the whole of these
results and demonstrated that the variations of resistivity, by whatever the characterization
method, are clearly due to hydride formation in the films, with the additional possibility
of oxygen contamination. They conclude: "it is certain that the films studied were
composite films (metal+hydride) thus agreement between theory and experiment is
fortuitous".
The next series is concerned with fine particles. Yamaguchi (1982a-d) has studied
the crystallographic properties of tribophysically pulverised R-metals (Gd, Tb and Dy).
This author observed a strange hcp +-~fcc martensitic transformation for Gd and Tb
small particles (about 100 nm in size), but not for Dy small grains (about 5 nm in size).
Moreover the cubic particles of Gd are not attracted to a magnet, whereas coarser grains

134

M. GASGNIER

having the hexagonal structure are ferromagnetic. The interplanar spacings reported in
various tables do not correspond to those of the metal. In this way Curzon (1982) and
Curzon and Gasgnier (1985) have shown that the fcc structure may be due either to
R203 or RH2 chemical compounds. Exactly in the same way Chizhov et al. (1982) and
previously Morozov et al. (1975, 1976) and Kostygov et al. (1977) report the presence
of the fcc phase in very small particles (25-40 nm in size) for Gd, Tb and Ho materials.
From XRD Chizhov et al. (1982) show that, as the grain size decreases, two fcc structures
are formed. As demonstrated by Singh and Curzon (1982) it is clear that the first one
is due to the dihydride and the second one to the sesquioxide (for holmium the lattice
parameter is 0.529 urn, i.e., exactly half the value of that for the bcc C-Ho203). In
addition, Chizhov et al. (1982) observe that the magnetic properties undergo a drastic
change for small particles (below 100nm) in that no magnetic ordering is observed
down to the helium temperature and they become no longer ferromagnetic. Singh and
Curzon (1982) notice that the magnetic ordering disappears for hydrided particles and
conclude that: "the results must be interpreted in terms of contaminants rather than in
terms of new size dependent forms of the R metals" (see also sect. 2.9.1.1 for other
details).
Some other papers report the formation of the fcc metallic phase for the R elements.
In a first paper Kozlovskii et al. (1981) disclaimed the formation of the Gd dihydride
compound, but later on (Kozlovskii 1989) asserted the contrary; Parmigiani (1982) reports
a strange lattice contraction for Yb and Eu films and does not link this change to
hydridization of the metals; McMinn et al. (1990) claim, in the case of rio thin films, that
the fcc structure is a non-equilibrium metallic phase. Hussain and A1-Bassam (1979) and
Hussain (1980, 1981) report that electrical properties of Gd thin films are closely linked
to microstructural changes. Particularly the authors observe that a granular to porous and
to continuous crystalline change corresponds to the fcc to hcp crystallographic change.
The resistivity curves and EDP show that the films are hydrided (the vacuum deposition
carried out under 10-4 Pa).
To conclude, and in order to introduce the next section, relative to rare earth hydride
compounds, one must point out the work of Loboda et al. (1980), Loboda and Protsenko
(1981a,b) and Loboda et al. (1982). These authors have studied the structure and electric
resistance of scandium films (30 up to 280nm) deposited under a vacuum of 10 - 3 10 .4 Pa. Loboda et al. (1980) claim that the chemical transformations, observed inside
a TEM, lead to the formation of ScH2 and Sc203. They report a good agreeement for the
lattice parameters. Then, Loboda and Protsenko (1981a) disclaim such an assertion, and
assign the fcc lattice to a pure metallic phase. In a following paper Loboda and Protsenko
(1981b) come back to their first assertion and claim the formation of the ScH2 compound.
Lastly, Loboda et al. (1982) do not discuss the electric resistivity in terms of chemical
contamination, but as a function of the defect ratio, thickness and grain size. This is
an exceptional change of ideas which demonstrates how it is sometimes difficult to give
sterling conclusions without carrying out experiments which lead to a focused network
of results.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

135

2.9. Formation and properties of the rw~e earth hydrides

2.9.1. Formation of the hydrides

2.9.1.1. Contamination of thin R films. Contamination of R thin films is too much
important to be neglected. Under various experimental procedures the formation of the
dihydride (RH2) can readily occur. It is now well-established that thinner films (1040 nm thick) consist mainly of hydride, whereas for thicker layers (40-80 nm) there is
coexistence of the metal and hydride. Curzon (1984) has tentatively explained why R films
are so contaminated. When the R metal is evaporated, hydrogen is generated at the
filament and/or the crucible. Another H2 source is due to the reduction of water at the
vapor source, and this one cannot be eliminated unless I-t20 is removed from the system.
So, hydrogen is difficult to remove efficiently by diffusion pumps, and during metal
condensation it is easily absorbed. Therefore, the thinner films consist mainly of hydride,
whereas in thicker layers only the surface would be affected. The surface substrate can
also act as a contaminant. The experiments of Curzon (1984) have been carried out with
17120 and D20, and it is shown that erbium, at the evaporation source, reduces water to
produce H2 and D2. Such an article is very important because it demonstrates, without
ambiguity, the high power "getter" effect of R for hydrogen. In this way, the new phases
such as the new fcc metal structure, the strange RxOy compositions, the new chemical
compounds,... (of which the lattice parameters are close to the one of the bulk hydrides),
reported by different research groups, cannot be taken into account.
In this way, Gschneidner (1980, 1992) has clarified the problem of the purity of
R metals. It is shown, first, that H, O, C, N and F are usual contaminants for commercial
samples, and, second, that physical properties (magnetic, low-temperature capacity,
electronic, thermal, ... ) are often flawed by such impurities. Therefore these samples
must be carefully outgassed before making films. However, it is difficult to completely
eliminate the hydrogen inside the vacuum chamber. Whatever the deposition mode during
evaporation, numerous parts of the equipment (crucible, target, walls of the chamber,
... ) are more or less heated. They outgas various gases including water vapor which
is dissociated by the rare earth to form the hydride. In this way one must mention the
conclusion given by Surplice and Kandasamy (1982): "The experiments show that Er, Yb
and Sc films prepared in ultra high vacuum will sorb H2 readily at room temperature".
2.9.1.2. Residual hydrogen. Rahman Khan (1977, 1978, 1981, 1984, 1987a,b) and
Rahman Khan and Miller (1979) show that the formation of RH2 is strongly dependent of
the residual gases in the vacuum chamber and of the adsorbed gases within the substrate.
The results are given for Gd, Tb, Ho and Er thin films. The gettering action of Gd is
reported in figure 10. From the curve it is seen that effective gettering happens within
45 seconds, and then saturates. On heating in He there is formation of RH3. In the same
way Curzon and Singh (1979, 1981a,b) have studied the formation of the hydrides for Y,
Ce, Pr and Nd samples. They observe the formation of various compounds like C- and
A-Nd203, CeO2 and also C-Ce203. However, one must notice that in the case of Ce, two
fcc structures (al ~ 0.543 nm and a2 ~ 0.555 nm) have been formed. This is explained

136

M. GASGNIER

5 x 10_5

0
I'Z

-- 5 x l O - 6
bJ
0
o')
bJ
n--

5 x l O -7
0

I0

20

50

TIME

IN

40

50

SECONDS

60
>

Fig. 10. The gettering action of Gd (formation of GdH2 in the early stages of the
evaporation). (By courtesy of Dr. M.S. Rahman Khan, Dep. Appl. Phys., Univ. of
Rajshahi, Rajshahi, Bangladesh).

in terms of metastable solid solutions of y-Ce and Cell2. In the case of Y layers these
authors add that it is difficult to form the YH3 compound even under exposure to 2670 Pa
of hydrogen. Kozlovskii (1989) reports that under sputtering deposition in Ne, Ar and Kr
gases, there is formation of the GdH2 and DyH2 compounds.
The hydrogen absorption rate of thin R films has been studied by Eley and Needham
(1984), Wulz and Fromm (1986a), Varkanova and Morozova (1984), Zhavoronkova and
Peshkov (1978, 1979), and Zhavoronkova and Boeva (1989). The former authors have
drawn particular attention to interstitial H2 and to catalytic activity of the dihydride
films for the H2 +D2 ~ 2HD reaction. Calorimetric measurements of the interaction of
H2 with R (Ce, Gd, Dy, Er, Tm, Yb and Lu) films (20-50nm thick) have been carried
out by Smutek and Cern~, (1985) and Boeva et al. (1986). They conclude that the heats
of H2 dissolution and RH2 formation become closer to each other as the temperature
decreases. They demonstrated that, at 295 K, Ce, Dy, Er, Tm and Lu give heats of
dihydride formation values 6-14% lower than those calculated from the data at high
temperature. Only Gd (which has an anomalously high specific heat) does not present a
significant difference. One can also notice that the effect of oxygen preadsorption layers
strongly influences the H2 absorption. So, Wulz and Fromm (1986a) have determined
that, in the case of La films, a thickness of 20 monolayers of oxygen at the layer surface
drastically reduces the H2 reaction. Hosoda et al. (1988) add that 02 precoverage has
less effect than H20 precoverage. Kawahata et al. (1990) compare the rate of the H2
adsorption on La and Ni films.
Kumar et al. (1984a,b) have investigated the chemisorption of H2 upon clean La and
Yb polycrystalline bulk materials. Time-of-flight analyses, surface recoiling of adsorbate
atoms, XPS and UPS measurements show clearly the evolution of the recorded spectra
upon exposure of the metals to H2. It is observed that LaH2 is formed in the outermost
sufaee layer at 50L H2 exposure. However, the top layers (~2 nm) are not saturated for
~2000L exposure.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

137

AES and EELS methods have been used by Bracconi et al. (1987, 1988) and Bracconi
and Laesser (1987, 1988) to compare the electronic structures of Y and Nd metal films
to those of their hydrides (YH2, PrH2 and PrH3). As reported by Colliex et al. (1976),
EELS reveals that the volume plasmon of Pr and PrH2 are shifted. Also special attention
is given to 02 contamination.
2.9.1.3. Reaction of R elements with water vapor. The problem of the reaction of the R
elements with water vapor is very important (Dexpert-Ghys et al. 1975, Curzon 1984).
H.K. Smith et al. (1980) and Oesterreicher et al. (1979) have demonstrated that the watersplitting process generates the formation of the dihydride according to the reaction:
2R + 2H20(g)

-+

2RH2

+ 02.

Kumar et al. (1984b) observed that during initial H20 exposure upon La specimens,
both oxide and hydride are formed, which is followed by hydroxide formation at
>35L exposures. Gimzewski et al. (1979) observe, from XPS experiments, that reaction
of Sc with water at 293 K gives initially the oxide (Sc203), and above ~50L the hydroxide.
However, at 80 K only hydroxide and chemisorbed water are produced, which is in contrast
to the behavior of the lanthanides. But, since no experiments show the plasmon energy
loss spectra (below 50 eV) the possible formation of the dihydride is not observed (see
Gimzewski et al. 1977 and Gasgnier 1980).
2.9.2. Physical characteristics
2.9.2.1. Electric resistance and resistivity. From the results reported by Curzon and Singh
(1979, 1981a,b) it appears that the electric resistance of RH2 films (R = Pr, Nd and Y)
is complex. As a function of thickness, temperature, H2 partial pressure, and aging, the
variations of resistance differ from one element to the other with a nearly similar but not
systematic identical appearance. For Y thin layers, the complex variation of the electric
resistance as a function of time at constant H2 pressure is shown in fig. 11. The resistivity
curves, for Y, YH2 and YH3, as a function of temperature are given in fig. 12. Rahrnan
Khan (1981, 1984, 1987a) has shown that the formation of the trihydride involved a rapid
fall in conductivity and a negative value of the TCR. This indicates that the metallic
dihydrides have become converted to semiconducting trihydrides. This transition has been
also observed by Curzon and Singh (1979, 1981a,b) in the course of over hydrogenation
of their samples (under H2 pressures of 2670 Pa). The increasing in resistivity for ScH2
thin films has been observed by Loboda et al. (1980) and Loboda and Protsenko (1981b).
Rahman Khan (1977, 1987b) explains the variation of resistivity with temperature
and thickness in terms of structural phase changes. It is observed that the temperature
dependence of the resistance and TCR values indicate metallic conduction characteristics
of the hydrides. The thickness dependence of resistivity of HoH2 films (7-90 rim) at 293
and 77K is given in fig. 13. Surplice and Kandasamy (1982), Kandasamy and Surplice
(1981, 1982, 1985) have studied H-Sc, H E r and H-Yb systems and deduced phase
boundaries from changes of film resistance with atomic ratio H/R. The slow variation of

138

M. GASGNIER
18

16

14

12

I0

Time

rains

Fig. 11. The effect of 20 Torr of H 2 at 293 K on the electric resistance of a 180 n m Y film. At point A hydrogen
is admitted to the system: an ordered phase develops and the resistance drops. At point B the hydrogen is
pumped away and the resistance remains constant up to point C. (By courtesy of Prof. A.E. Cnrzon).
9O

70

...~-~"~p

n -c m

60 ~
5O

T4
2O
iJ~l-cm
10
r

I00

150

200

250

300

TK-~

Fig. 12. A plot of resistivity versus T for films containing predominantly Y (circles), YH 2 (squares) and
YH 3 (triangles). (By courtesy of Prof. A.E. Curzon).

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

139

200
Room

. . . . .

150

of

77

temperature
o

<
IO0
I--

~_

>

Vff)

~..e . . . .

"

__

Bulk

50

bJ

I00

2 0

3 0

4.00

500

THICKNESS,

600

700

800

900

.~

Fig. 13. Thickness dependence of resistivity of HoH z films (7-90nm). (By courtesy of Dr. M.S. Rahman Khan).

109

108

107.

10~-

a-_
10~'.

I0~
1

d7
~ / xz/
V..~,

1o7]

100

Fig. 14. Change of resistance with atomic ratio

r = H/Yb. Solid and dashed curves are for films
that were hydrided slowly and quickly, respectively.
(After Surplice and Kandasamy 1982.)

resistivity in RH2 indicates that most of the modifications of the conduction band occur
in the mixed phase; this could be due to the R-RH2 structural transformation between
the metal and the metal + dihydride phases. In the case of H-Yb system, in which the
hydride is a poor conductor, the solubility limit (the position of the first phase boundary)
has been estimated. Changes of resistance with atomic ratio are reported in fig. 14.

140

M. GASGNIER

2.9.2.2. Workfunction. The changes in work function in the course of hydrogenation have
been determined for Sc and Er thin films by Miiller and Surplice (1977) and Kandasamy
and Surplice (1985). As for the electric resistance they correlate the variations with
changes of phase in the R/H system. In a general manner, first, small doses of H2 reduce
the work function to about 0.!0 eV below the value for the clean metal; second, larger
doses increase it to about 0.20 eV above this value; and third, further doses quickly raise
the work function to a maximum of about 0.55 eV above the one of the clean metal. The
results relative to the metallic phase can be interpreted as changes of surface potential;
but for the other phases they depend on changes of the Fermi level of the H/R system as
well as on the surface potential. Moreover these results are compared to those obtained
for other metal hydrides (La, Ta, Ti, U, Pd, Nb and Zr). Similar observations have been
reported by Eley and Needham (1984) for Gd films.
2.9.2.3. Surface magnetization. Cerri et al. (1983) have studied the spin-polarized
emission of clean and hydrided (submonolayer coverages) polycrystalline Gd thin films.
At 20 K, hydrogen coverage (chemisorption) drastically reduces the spin polarization from
70% (clean surface) to 45% (N0.5L H2) and 30% (~I.0L H2). With increasing temperature
the polarization decreases linearly, up to 130K, then rises to a maximum near 200K,
and finally decreases again. Extrapolation of the linear part of the curves shows that
chemisorbed hydrogen strongly reduces the ordering temperature and induces a canted
or disordered spin structure at the Gd surface. This experiment is a sensitive monitor of
hydrogen contamination.
2.10. Rare earth deuteride and tritide
2.10.1. Bulk materials
2.10.1.1. Introduction. As a short introduction, one must emphasize the latest results of
Adachi et a1.(1992) who report the possibility of "cold fusion" using the system D2LaNis. This has been deduced from analysis of H2 in gases resulting from D2 absorption
with an LaNi5 ingot at 100-300 K under a pressure of ~5x 105 Pa. Such an assertion calls
for new experiments and will certainly be the subject of new polemics.
2.10.1.2. Gettering. Maienschein (1978) has reported that cerium is one of the best
chemical getters which can be used to scavenge tritium from inert gases. Indeed,
this metal, as well as Sc, Y and Er, has low dissociation pressures for temperatures
between 298 and 523 K. Maienschein assumes that dissociation of the tritium molecule
to two tritium atoms and diffusion of tritium atoms through the solid tritide (due to the
flaking and spalling effects of the material during tritiding) are both rapid. As cerium
hydrides are stable and exhibit plateau regions, it is also assumed that the gas-phase mass
transfer is the controlling step in the gettering process.
Hubberstey et al. (1976) have used yttrium sponge as a getter for hydrogen isotope
removal from liquid lithium. The rate of gettering is remarkably rapid at 673K.
Hydrogen isotope concentration in Li can be reduced from 1.00 to less than 0.05 tool%

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

141

x (x = H or D). Such a result should be extrapolated to lithium-tritide solutions. In the

same way Buxbaum (1982) have used yttrium for the separation of tritium from the liquid
lithium breeder-blanket of a fusion reactor.
2.10.2. Thinfilms
2.10.2.1. Scandium and yttrium. Malinowski (1981)has shown, from AES spectra, that
the so-called LMV peak of Sc might be a useful indicator of film deuterium content.
Indeed, he reports a linear dependence of the Sc LMV peak height on the deuterium
content of the film (100nm thick). This is observed as a function of the fractional
decomposition of the film such as
S c D 2 ----r ScD0.5 --+ Scmetal

(at 673

K).

One must notice that films were contaminated by C and O (from evaporator), and S (from
Mo substrates).
Cowgill (1979) has studied dynamic deuteron implant effects (at 40keV) using
Sc targets (previous experiments gave results at 200keV, see Cowgill (1977)). The
experimental system is useful for studying D2 retention and mobility in materials under
deuteron bombardment. Isotopic hydrogen exchange has been observed dynamically
at 413 K. It is also shown that oxygen implants deplete the target of D2 within the implant
range due to the formation of ScaO3. Cowgill (1981) has used the same technique for
measuring D2 diffusion in films. He notices that no significantly different behavior is
observed for films of different thicknesses (0.5-5.5 ~tm), the diffusion being identical to
the one observed in bulk samples.
Singleton and Yannopoulos (1975) have used Sc and Y films (500-1200 nm thick) for
fabrication of radioactive electron emitters (tritriated electron sources). The stability of
the sources in flow gas streams increases from Ti to Y to Sc. Once again, it is shown
that the surface contamination has a profound influence on both the loading and loss of
tritium in the films. Bacon et al. (1984) have manufactured SoD2 and ScDT thin film
targets (10-50gm) for neutron protection inside an intense neutron source for use in
cancer therapy. The films must not be heated at temperatures over 723 K to maintain
their chemical stability.
2.10.2.2. Erbium. Thin ErD2 or ErT2 films are used for neutron generator targets (or
tubes), for high-intensity rotating target neutron sources, for radiotherapy, and for highintensity neutron sources for cancer research.
Provo (1979) has studied the hydriding process for the following system: Er (400600nm thiek)/Cr (100-500nm thick)/Cu (substrate). He determines that to achieve an
E r D 2 / T 2 occhider film gas-to-metal atomic ratio of 1.7, a minimum of 150nm of Cr
underlay is required for an in situ hydriding process, whereas such a minimum is 300 nm
for an air-exposed hydriding process. The formation of oxides at each interface and
the interfacial metal loss (diffusion) at the Er/Cr edge can limit the optimum hydriding
process. Another characteristic of ErT2 films has been studied by Mitchell and Patrick

142

M. GASGNIER

(1981) and Mitchell and Provo (1985). They report the temperature dependence and
irregularities of helium release rates from EfT2 films. Some fragmented results had
been reported previously by Beavis (1980), Beavis and Kass (1977) and Kass (1977).
Mitchell and Patrick (1981) give details on He release fraction as a function of aging
and of temperature. The samples are aged over very long periods (70-2587 days), either
maintained at different temperatures (from 77 to 500K) or sometimes annealed in the
course of aging. The main results indicate first that large changes in He release rates
follow immediately upon temperature changes and second that the accelerated release
process is reversible. Then, Mitchell and Provo (1985) observe that nonuniformity of the
release is greatest for samples undergoing the transition into accelerated release, which
occurs when the oecluders (EFT2 films, 0.5-2 ~tm thick) approach the maximum quantities
of helium that they can retain. It is concluded that the variability in He release rates might
be due to a bursting activity: 3He is released in bursts of at least 109 atoms; this release
is also stimulated by vibrating or flexing the film substrates.
Holloway et al. (1978) and Antepenko and Holloway (1980) have studied the
degradation of ErD2 and ErT3 films, ranging in areal density from 0.012 to 0.534 mg/cm2,
and deposited by e-beam onto Mo substrates. It is shown that the total unavailable
metal for the nonannealed films, if converted to total oxide, would be 11.5 nm. For in
situ hydrided films this thickness does not excess 5 nm. It has been also determined
that 0.01 mg/cm2 of Er is not hydrided for films that were e-beam deposited and in
situ hydrided. Lastly samples subjected to vacuum annealing at 773 K (1 h) exhibit a
total nonhydrided metal quantity of ~0.02 mg/cm2, therefore a total surface oxide level
of 16.5 nm.
2.11. Formation o f the rare earth nitrides

The formation of pure rare earth nitrides is often difficult to carried out, because the
presence of oxygen can lead to the formation of oxynitrides. However under certain
experimental conditions it is possible to form pure nitrides.
One of these procedures consists of carrying out the interaction of nitrogen with
continuously renewed films of rare earth metals as reported by Varkanova and Nazarov
(1977), Varkanova and Morozova (1981), Varkanova (1982) and Varkanova et al. (1982).
These studies have been done for Sc, Y, Sm, Gd, Er and Yb metals. The absorption
of nitrogen has been investigated as a function either of the rate of condensation of
the metals, or of the temperature (between 298 and 473 K), and/or of the nitrogen
pressure (between 10-6 and 5x 10-4pa). The most important study is relative to the
sticking coefficient of nitrogen against the various parameters above-mentioned. In a
general manner, the functions obey linear, increasing or decreasing, laws. Two kinds of
composition have been observed. For scandium in the gas-excess regions it forms a singlephase nitride system (ScN0.9, fec with a = 0.453 rim). In the metal-excess region there is
a mixture of two phases: one relative to the metal and the other to the nitride. Varkanova
(1982) has also established, as a first approximation, that the changes in the sticking
coefficient are due to the heat of sublimation of the metal: higher heat of sublimation

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

143

leads to higher sticking. The nitrogen absorption of rare earths is always compared to
that of titanium which seems to be similar.
Another technique has been carried out by Ma et al. (1987): thin films (150nm
thick) are deposited onto Si or NaC1 substrates, and then irradiated at high dose with
nitrogen ions (2-51017N+cm-2). In the case of gadolinium there is, at lower doses
(2 1017N+cm-2), formation of a compound of which the crystallographic parameters do
not exactly correspond to the lattice standard value of GdN due, probably, to a deficiency
in nitrogen content. In the same way, XRD indicates that the gadolinium lattice is also not
correct. This is possibly due to a metastable structure of gadolinium supersaturated with
nitrogen. At higher doses the GdN compound becomes predominant and the interplanar
spacings are close to the standard values. However, one must remark that the correct
XRD interplanar spacings show clearly that the "so-called" Gd (110) line (with a low
intensity) with a spacing value of 0.187 um is in fact the (220) line of GdH2. In this
way the formation of a new metastable compound seems less probable. The presence of
hydrogen throughout the starting material is once again not surprising (the authors do not
give their vacuum evaporation parameters).
2.12. Reaetioity with CO, C02 and CnHn gases
2.12.1. Rare earth~CO and CnHn interactions
In an earlier chapter of this Handbook series, Netzer and Bertel (1982) have reviewed
work carried out in order to study the reaction of rare earth metal surface with carbon
monoxide. Affrossman (1981) has studied the reactions of CO with clean scandium film.
He reports, from XPS experiments, that this metal dissociates CO to form a carbide and
sorbed oxygen. The behavior of scandium shows strong similarities with that of tungsten.
In the same way methanol and ethanol dissociate to leave oxygen preferentially at the
Sc surface at low exposures and to form carbides at higher doses. In this case, there is
formation o f a "~-CO" type layer. Cern~ and Pientka (1987) and Cern~, and Smutek (1990)
for thin Dy films (150 nm) report, from calorimetric experiments (heat measurements),
that CO dissociates on the metal surface (see also Surplice and Brearley 1978). This is
followed by rapid penetration of the oxygen atoms (formation of the oxide), while the
carbon atoms remain on the surface and cause a gradual blocking. They do not observe
the formation of carbides. Moreover, it is determined that the rate of heat production is
fairly high, but lower than with hydrogen at the same H/Dy ratios (Boeva et al. 1986).
From mass spectrometry studies Curzon (1984) reports that the reduction of CO, inside
the vacuum chamber, leads to the formation of CH4 (and consequently of CH3, CH2 and
CH). The results indicate that the CH4 (and also H2) arises from the reduction of the
CO (and also H20, which reacts with CO). That is another way to explain the formation
of hydrides with rare earth thin films. The adsorption of C3H6, C2H2 and CH4 on clean
polyerystalline Dy films at 295K has been studied by Cern~ and Smutek (1990). The
experiments suggest that at low doses, the gases are completely dissociated into C and
H atoms. The bonding of these atoms to Dy is assumed to be equivalent to that which

144

M. GASGNIER

occurs in chemisorption of CO. That leads to a surface-blocking effect of the formed

species, and Dy films possess a higher chemisorption capacity for C2H2 than for C3H6.
2.12.2. Lanthanide/C02 reactions
Little research has been carried out to study lanthanide/CO2 reactions (see Netzer and
Bertel 1982). Mehrhoff (1980) has studied the gettering of CO2 by erbium thin films
deposited onto Mo substrate discs heated at about 700 K. The reaction begins near 570 K
in an abrupt manner (no reaction is detected with Mo). Mehrhoff (1980) also reports
isothermal measurements, ratios of CO2 pumped by the films at various temperatures
(from 890 to 1185 K), sticking coefficients versus exposure to CO2, and monolayers of
CO2 absorbed as a function of the temperature. It is also shown that there is formation of
CO near 770 K, that could be due to the interaction of CO2 and erbium metal. Arakawa et
al. (1988) have reported the dissociation of CO2 on some rare earth (Pr, Nd, Sm, Dy and
Er) films (1 ~tm thick). The oxidation of the metal is investigated by measurement of the
resistivity of the films. Under CO2 atmosphere the resistivity increases strongly: for Dy, at
603 K, it is 14 times that at 298 K. Above 823 K the films have insulating properties due
to the metal --+ sesquioxide transformation. For praseodymium the resistivity behaviour
is fully different, which is possibly due to the formation of a non-stoichiometric oxide.

3. Metallic R-alloys
Metallic R-alloys are now being studied increasingly in order to obtain numerous
manufactured products. However, as for pure metallic thin films, the cross-shaped problem
is the contamination by atmospheric gases. This is a dramatic feature, not always pointed
out by numerous research groups, and it can be on the contrary a benefit in the case of
hydrogen storage materials like RNi5 or RCos.
3.1. Permanent magnets
3.1.1. Introduction
Numerous papers have been published on these permanent magnet materials; we cannot
list all of them in this chapter (see Burzo and Kirchmayr 1989 and Gasgnier 1991 for
example). A publication of the Gorham Advanced Materials Institute (May 10, 1991)
showed that the global permanent magnet market was clearly influenced by the increasing
NdFeB sales ($2.7 billion or 12.652 metric tons, therefore 17% of the market), and by
the continued penetration of RCo alloys (11% of the market).
One of the most important problems relative to the synthesis of these materials remains
the formation of new phases and structural inhomogeneities inside the matrices. From
XRD after heat treatment at 740K of NdFeB melt-spun ribbons, Strzeszewski et al.
(1990) have pointed out four phases: Nd, Nd203, and two (Nd, Fe)O iron-rich phases
(with tetragonal and hexagonal structures, 70 at.% and 85 at.% Fe, respectively). TEM
imaging showed that such phases were present as spherical groins. As different kinds of

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

145

phases can be formed, it follows that the magnetic properties can be different from one
sample to the other. The aim of this section is to report the new experiments and results
obtained since about 1987.
3.1.2. [R(Pr, Nd),R' (Dy)]xFeyBz alloys
In a previous paper (Gasgnier 1991) we have shown that, since 1984, the research and
development on these materials was growing more and more. Thus, there are a lot of
recent papers which give details on new experimental methods and unusual synthesis
treatments.

3.1.2.1. Magnetic properties.

3.1.2.1.1. Thick and thin films. Sputtering deposition methods of R-TM permanent
magnet films (SmCos, Sm2(Co, Fe, Zr)17, Sm-Yi-Fe and Nd2Fel4B) which exhibit high
intrinsic coercive forces, large remanent moment values, and special anisotropies have
been studied by Cadieu (1988). In the case of thick crystallized Nd2FeI4B films (thickness
0.8-3.2 ~xm), Cadieu claims that, as a function of the sputtering rates, it is possible to
synthetize in-plane or perpendicular easy direction of magnetization. So, at deposition
rates <0.18 nrn/s the films are easier to magnetize perpendicularly to the plane. Such
specimens possess higher remanent magnetization (~1 T) than in-plane magnetic films
(~0.6 T). An XRD trace (between 10 and 18) shows only the presence of the reflections
relative to the tetragonal structure (a = 0.880 nm and c = 1.218 nm) &the Nd2Fe14B alloy.
The crystallization behavior of rf sputtered amorphous films (4 and 8 ~tm thick) have
been studied by Homburg et al. (1990) and Gu et al. (1990). The films when annealed
at 886 K begin to crystallize and are fully crystallized at 1023 K as the NdzFe14B phase.
At 1123 K a small amount of (x-Fe was observed. At a slow sputtering rate deposition
and after annealing at 1023 K, an unknown phase was formed. Such a sample exhibited
a higher coercivity (>1256kAm-1). The sputtered and annealed (up to 823K) films
are magnetically soft with a low coercivity (<800kAm-1). The former present an inplane magnetization suggesting the existence of anisotropy. After annealing above the
crystallization temperature, a hard magnetic phase is formed. The coereivity reaches a
maximum at 1023 K and then decreases rapidly (due possibly to the formation of large
grains containing many domains). SEM observations show that films deposited at a rate
of 0.8 nm/s have a dense packed structure (block-like grains with a laminar substructure)
whereas films sputtered at a rate of >l.5nm/s exhibit a columnar structure with voids
between the columns. The latter possess an effective vertical anisotropy. It is concluded
that grain size and orientation are the most important factors determining the coercivity.
Other parameters are the degree of isolation between grains and the density of faults
which increases with the sputtering rate.
For the thinnest amorphous films (100nm) deposited by triode sputtering onto
substrates maintained at 293 and 77K, and then annealed upto 773 K, Alameda et al.
(1990) show that the in-plane magnetic (induced) anisotropy, measured at the glass-film
interface, increases at low temperature and falls to low values after annealing. The same

146

M. GASGNIER

result is valid for Ku. The authors give a clear correlation between the behaviors of these
macroscopic and microscopic anisotropies (i.e. the product of the magnitude of local
anisotropy and the volume where the principal axes of these anisotropies are correlated).
3.1.2.1.2. Multilayers and modulated films. Aylesworth et al. (1988, 1989) report the
properties of Nd17(Fe0.9Co0.1)76B7/Fe or Ag sputtered multilayer specimens deposited
onto mica or tantalum substrates at different temperatures (between 293 and 993 K). The
individual layer thicknesses are 10-50urn for the alloy and 0.5-20 nm for the Fe or Ag,
the total thickness being 1 ~tm. After annealing at 873 K, XRD reveals first the formation
of the Nd2Fe17 and AgNd alloys, and second the presence of a contaminant labelled as
"NdO" with a fcc structrue (a0 ~ 0.510 nm). One must point out that this compound does
not exist. Moreover, although the measured interplanar spacings (0.288 and 0.250urn
for the two first diffraction lines) are consistent with a fee lattice (dm/d2oo = 1.15),
they have been indexed as a bcc structure, i.e., according to the (110) and (200) planes
(dllo/d2oo = 1.415). In fact, the value of the fcc lattice parameter corresponds to that of
NdNxOy which is formed after moderate annealing of R thin films (Gasgnier et al. 1976).
This compound cannot be confused with NdH2, NdN, or C- and A-Nd203. The authors
observe that the "NdO" contaminant diminishes in amount when the mica substrate is
covered by an Fe layer (50nm) and disappears completely if a Ta substrate is used.
However, in the case of (ProrNd)2Fe14B/Ta cosputtered multilayers this contaminant
coexists with the Nd2Fe14B and Fe2Ta alloys (Aylesworth et al. 1991). But, in the case of
Pr2Fe14B/Pr or Ta multilayers this contaminant is not observed (Aylesworth et al. 1990,
1991). Multilayers which are contaminated have larger coercivity, lower magnetization and
are more randomly aligned than clean films prepared under similar conditions. The grain
orientations and the anisotropies strongly depend on the substrate material and the insitu applied magnetic field. So, non-multilayer samples deposited onto mica tend to have
larger Ku than similar films deposited onto Ta. Moreover a magnetic field (104 kA m -1)
applied parallel to the film plane during Nd2(FeCo)14B formation, produced films with
in-plane anisotropy. For these samples, after annealing between 723 and 873 K, He can
reach values as 800kAm -I at 293 K. Variations of Hc versus maximum applied field
(NdFeB/Fe samples), temperature (PrFeB/Pr samples) and nominal Ta thickness ((Pr or
Nd)FeB/Ta samples) are shown in figs. 15, 16 and 17, respectively.
Martinez et al. (1988a,b) have investigated Nda6Fe68B6 or Nd12FesoB6/Fe92B8 compositionally modulated films with a modulation length varying between 0.34 to 5.47 um for
the Nd-rich alloy, and a thicker Fe-rich layer (200 urn). At helium temperature and as
a function of the applied magnetic fields (perpendicular or parallel to the substrate), the
magnetization increases as these fields increase. From hysteresis loops it is shown that the
spin-wave modes (collective excitations) depend on both the thickness and the modulation
of the multilayers (geometry of the samples). The perpendicular magnetization values
decrease as the modulation lengths increase. This phenomenon may be attributed to
the increase of the number of paramagnetic Fe atoms, as deduced from Mtssbauer
spectra which show an increase of the paramagnetic doublet contribution with increasing
modulation length. From these spectra it is shown that both the hyperfine fields and the
easy magnetization magnitude are not correlated.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

I

147

perpendicular

T=2OOK X

~fl~

I ~ 1 1 1 ~

C3
c
Fig. 15. Coercivity versus maximarn
applied field at T=200K for a NdFeB
(20nm)/Fe(0.5nm) sample. The result
indicates that magnetization reversal in
the easy direction is controlled by a
domain-wall pinning mechanism. (By
courtesy of Dr. D. Sellmyer,Behlen Lab.
of Physics, Univ. Nebraska, Lincoln,
NB, USA).

:o

210

O
tO

'

410
Hma ( k O e )

'

610

810

18

NFB(200A):Ta(XA)
'random~
/

-$O 9
-rO

(U
C)

NFB(200A):Ta(XA)
'aligned'

PFB(200A):Ta(XA)

3~

--

zoo ~: 40 ~,

+-A--

200

--

~: l o

iO0 $,: I0
I

",,,%,.',,,
\",~

~
A

Temperature

,,IJ~,,I,~,,l,~l,,,~l,~,,lJ,,,

lam

100

200

300

(K)

10 15 20 25 30 35
X

Fig, 17. Coercivity versus nominal Ta thickness

with the field applied perpendicular to the film for
"aligned" PrFeB (20nm)/Fe (xnm) and "aligned"
and "randomized" NdFeB (20nm)/Ta (xnm) films
(By courtesy of Dr. D. Sellmyer).

Fig. 16. Summary of coercivity versus temperature data with the field applied perpendicular to PrFeB film
plane. (By courtesy of Dr. D. Sellmyer).

3.1.2.1.3. Magnetic domains. In a general manner magnetic domains are observed either
by means o f Lorentz electron microscopy investigations (thin specimens) or by decorative
methods (thick samples). In the first case, Bras et al. (1988a,b, 1990) have improved
the suitable conditions to investigate magnetic domain nucleation, domain wall motion,
domain size and wall energy in highly uniaxial magnetocrystalline anisotropy materials.
This technique is available by using the magnetic field o f the standard objective lens

148

M. GASGNIER

which induces magnetic structure variations and by optimizing the grain orientation (i.e.,
a small angle between the anisotropy axis and the normal to the foil). The "optimal
orientation" is obtained by means of a rotating specimen holder. This allows separate
study of small grains and their interactions. In the case of a thinned Nd2Fe14B ribbon
the authors have determined from the unusual Foucault mode images the wall width,
the diameter of bubbles near stripe domains (70-100nm) and their stability, the global
grain magnetization which permits one to deduce that the reversal domain nucleation
arises under the influence of a demagnetizing field, and the neighbouring grain influence
(abrupt change in the grain magnetization). Examples of strip domains obtained by 200 kV
Foucault images are shown in figure 18.
In another set of experiments Griitter et al. (1988, 1990a,b) have used magnetic force
microscopy (Heinzelmann et al. 1987) to study Nda4FeglB5 materials having optimum
magnetic properties. This allows them to observe magnetic domains in air, and to achieve
a high lateral magnetic resolution (<10 nm). The authors notice that Lorentz microscopy
images look similar to those obtained by this technique; They observe the formation of
polyhedral-shaped domains (600-700 nm in size) and of double-peak structures separated
by less than 10 nm. But a conclusive interpretation has not been given up to now. In the
same way, domain walls are observed as being very sharp due to efficient pinning of
which the exact nature remains unclear.
3.1.2.2. Hydrogen treatments. To produce coercive powders for the processing of
permanent magnets McGuiness et al. (1986, 1989a,b, 1990a,b) have used hydrogen. In
a first route they have produced N16Fe76B8 and Nd14.sDyl.sFe76B7 alloys by a hydrogen
decrepitation or attritor/jet milling. The former induces larger magnetic parameters, e.g.,
the coercivity is 1040kAm -1 and 1600kAm -1 for NdFeB and NdDyFeB, instead of
720 kA m -1 for a sample prepared by attritor milling. This coercivity is achieved for an
average grain size of <14 ~tm (for larger grain sizes the value of He decreases). To avoid
excessive grain growth the sintering temperature must be reduced to below 1313 K for
a 1 h sinter. In addition, a Nd-rich phase has been observed at grain boundaries, and
the NdFe4B4 phase is also detected. This research group has used another treatment
as an effective processing route to produce polymer-bonded magnets. These are made
by a hydrogenation=disproportionation-desorption process which leads to the production
of fine-grained homogeneous isotropic materials which exhibit appreciable coercivities
(604-1040kAm-1). The optimum temperature was found to be 1058K. Thermal
treatment strongly affect, here again, the grain size, and therefore the magnetic properties.
This process leads to roughly spherical grain powders (~500 gm in diameter), whereas
hydrogen decrepitation leads to flake-like powders with a great proportion of fine debris.
From Ndl6Fe76B8 and Nd2.1FelaB materials, heated at -1073 K, and then quickly cooled
in hydrogen to 298 K, the formation of different crystallographic structures has been
observed. These have been indexed from XRD as a-Fe (bcc, a = 0.287 rim), NdHx (fcc,
a=0.546nm), Fe2B (ferromagnet tetragonal with a=0.511nm and c=0.425nm) and
NdFe4B4. The formation of the hydrided variant NdEFelaBHz7 has also been observed
(MeGuiness et al. 1989b). The authors conclude, first, that disproportionation takes place

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

149

Fig. 18. 200 kV Foucault images of strip domains due to demagnetizing field relative to Nd2Fe14B monocrystals.
(a, b) Opposite domain contrasts due to change of sign of the angle between lens field and anisotropy axis.
(c) Softening contrast due to the decrease of the magnetization component (the electron beam deviation becomes
weak for low angles). (d) No domain contrast (angle close to 0). (By courtesy of Drs. J. Bras et al., Lab.
Physique du Solide, Univ. R Sabatier, Toulouse, France).

150

M. GASGNIER

during the production of the high-coercivity powder (for the sample not subjected to a
vacuum annealing treatment), and second, that the development of this high coercivity
can be ascribed to the reforming of crystals of Nd2Fel4B and the establishment of a fine
optimum grain size.
3.1.2.3. Microstructuralproperties. As discussed earlier (Gasgnier 1991) disorder seems
to rule over the nature of the various mierostructural phases which are formed either
during synthesis or in the course of the preparation of the samples for TEM investigations.
For example Hole et al. (1990) noticed that the phases are different according to the
mode &preparation: ion erosion or electropolishing. Ramesh et al. (1987) observed for
a (Nd, Pr, Ce)2Fe14B thinned alloy, the formation of a fee phase (a = 0.524 nm) and of
a bee one (a-0.290nm). They conclude, on the basis of results published in 1961(!),
that the former is a "NdO" phase wich contains some dissolved Fe, and that the latter
is a ion-milling artifact. Such structures are observed at the grain-boundaries. Fidler
(1987) and Fidler and Tawara (1988) for electropolished and ion-milled NdlsFeyTB8
specimens report the formation of the following structures: Nd2Fel4B, NdFe4B4, four
Nd-rieh compounds located at grain boundaries, and a-Fe precipitates. One of the Ndrich phases, NdzFeB3, has been recognized within the phase diagram. Other ones also
identified are: A-Nd203, NdC12, Nd chlorate and Nd(OH)2C1. Such contaminations are
very prejudicial to magnetic properties, more especially the Nd-ehlorine phases which
contain a high density of dislocations. The problem of the Nd-rich phases has also been
studied by Tang et al. (1988) who report the formation of a fee structure (a -- 0.510 nm)
as do Aylesworth et al. (1989, 1990). From Auger spectra, Tang et al. (1988) conclude
that this phase is a Nd-Fe-B-O quaternary liquid phase. These authors observe another
Nd-rich phase with a dhep structure (a=0.365 nm and c = 1.180nm) which is that of Nd
metal. Hole et al. (1990) observe also at grain-boundaries the formation of such Nd-rich
phase with Fe and B additions. They identify, from Auger spectra, the presence of a-Fe
mierocrystallites and Nd203 and B203 inclusions. They conclude that during annealing
the quantity of hard magnetic phase increases due to the reaction of Fe excess with the Ndrich phase; on fast cooling, precipitation of ct-Fe could contribute to magnetic hardening.
Iwamura et al. (1990) observe a grain-boundary phase with a tetragonal structure
(a=0.512nm and c=0.800nm) in NdzFe14B ribbons (45-100~tm thick) with high H
(1840 kAm-1). It is metastable and decomposes into "fee Nd" and NdHFe4B4 on heating
at 963 K. On the contrary, G.X. Huang et al. (1990) claim that their Nd13.sFes1.75B4.75
melt-spun thinned ribbons are single phase which explains their high coercivity. After
annealing at 573 K an unknown intergranular grain-bonndary phase is formed. Koestler
et al. (1989), from TEM observations and EDX microanalyses, have deduced that the
grain-boundary fee phase was Nd75Fe25. The a0 parameter (a0 ~ 0.560 nm) is however
incompatible with any of the NdFe (O?) known structures, but it could possibly be linked
to a [Nd(Fe)]203 bcc compound. Indeed the C-Nd203 compound with ac = 1.116 nm (i.e.
a0 2) can be confused with a fee compound. In this way, the weak (211) reflection
does not appear in EDP, and only the (222), (400), (440), (622) and (444) reflections are
observed.

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

151

In the case of Co-substituted Ndls(Fel-xCox)77B8 samples, Koestler et al. (1989) report

the formation of two fcc phases. The first one is defined as Nd(Fe, Co)2 with a fcc
structure (a=0.730nm). The second has a fcc lattice (a=0.540nm) and composition
Nd20(Ndl-xCox)80. The authors claim that this phase is close to that of Nd6Fe23 (fee,
a = 1.210nm), but there are few correlations between the lattice parameters. They have
not observed the Ndl.lFenB4 compound. Yamamoto et al. (1987) for NdlsFe77-xCoxB8
(x = 0-77) and Nd24Fe58Co10B8 materials observed the following NdCo alloys: 1:3, 2:7,
5:19, 1:5, 1:2, 7:3, and 3:1; and the Ndl.I(Fe, Co)aB4, Nd(Fe, Co)2 and Nd3(Fe0.sCo0.5)
intermetallics. The Nd(Fe, Co)2 phase may be formed peritectically and acts as a
nucleation site of reversed domains. In the case of the Nd24Fe58Co10B8 sample the
Nd2(Fe, Co)14B, Nd-rich and B-rich phases are not observed. Strzeszewski et al. (1988)
have studied rf sputtered Nd16.5(Fe0.9Co0.1)76.sB7films (1-2 gm thick). XRD reveals the
coexistence of the 2:14:1 hard phase with the fcc (a = 0.512 nm) phase. The amorphouscrystalline transition is observed at 878 K. Thin samples observed by TEM show the
coexistence of ct-Fe and an amorphous phase. After annealing up to 1003 K Nd203 is
formed. To prevent oxidation and obtain the hard magnetic phase, the films must be
covered with a thin gold layer (20 nm). Lorentz microscopy reveals the presence of various
domains predominantly parallel to the film plane.
Cochet-Muchy and Paidassi (1988) have studied the problem of the effects of oxygen
for different solidification conditions of NdlsFe77B8 samples. For as-cast materials they
observe the following phases: Nd2Fe~4B, ct-Fe, NdL1FeaB4, Nd-rich phase, C-Nd203,
Nd(OH)3, Nd82.sFe2.5Ox (as do Ramesh et al. (1987)), and Nd55Fe25020 or Nd7Fe55031.
One must also point out the Auger analysis results of Jaswal et al. (1988) who report
for different R2Fel4B (R = Y, Nd, Gd) materials, that the surfaces present a high degree
of oxidation after exposure to the atmosphere. Reduction of these oxides is achieved by
heating in the presence of H2 or CO (C, S and N as contaminants are also detected).
In conclusion one must point out that the synthesis of Nd2Fel4B pure hard magnet
films has not been achieved. The formation of various phases, often labeled with chemical
formulae which are not known to exist, is observed as a function of the thermal treatments,
the purity of the starting materials, the contamination by various gases . . . . The most likely
contaminant seems to be oxygen. However, numerous authors do not know the chemical
properties of the transition metals and deduce new metallic or oxide compounds either
from the value of the lattice parameters of from "tentative" identifications. So, Aylesworth
et al. (1991) for Pr2Fel4B (20nm)/Pr (4rim) multilayers claim the formation of a "PRO"
oxide (as for "NdO", Aylesworth et al. (1989)). One must also notice that these phases are
generally located at grain boundaries, or are present as small particles, or are observed as
amorphous states. This problem is also encountered in the new high-Te superconductors:
it is very difficult, during synthetization, to avoid the formation of isolated phases; they
always affect the physical properties of the superconductors as well as the supermagnets.
3.1.2.4. Other compounds: RFe(N or C). Recently, new types of compounds have been
prepared. Sellmyer (1991) has studied the magneto-optic properties of R2Fel7Nx (x ~ 4)
alloys as thin films. Such a material crystallizes according to a rhombohedral structure.

152

M. GASGNIER

In the solid state, Jaswal et al. (1991) have studied the structural parameters and the
magnetic moments of (Nd, Y, Lu)2FeI7 and (Nd, Y, Lu)2Fel7Nx alloys. They observe,
for instance, a significant increase in Curie temperature upon nitrogenation. Coene et
al. (1990) have investigated the magnetic and crystallographic properties of Y2FetTCx
(0 ~<x ~< 16) compounds. As a function of the increasing carbon content they observe
a structural change from the hexagonal Th2Nil7 type to the rhombohedral Th2Zn17
type. Near the phase transition there appears a partial disordering of the lattice leading
to an X-ray pattern which has CaCus-type base reflections. EDP also reveals such a
disorder. It is observed that the lattice constants increase as the C content increases. The
crystallographic properties are reported in terms of stacking sequences.
3.1.2.5. Ndl.IFe4B4. The Ndl.IFe4B4 (B-rich) compound, observed by different authors
mentioned above, has been studied in detail by Zhao et al. (1989a-d). They have
demonstrated from EDP and TEM observations that thinned materials are tetragonal
(a=0.711 nm and c = 0.389 nm) with an incommensurate structure in the c direction of
the tetragonal lattice (Zhao et al. 1989a,b). The specimens are highly planar-faulted.
Refinements of the results have led to the conclusion that, in fact, this compound exhibits
the characteristics of infinitely adaptive structures (Zhao et al. 1989c,d). Several different
superstructures have been observed, which coexist throughout the matrices and may be
related to local fluctuations of the composition. In the case of Co-substituted samples
the crystal structures are built up of an interpenetrated Nd-substructure and a (Fe, Co)B-substructure. Similar results have been reported for Cel.IFe4B4 alloys by L. Li et al.
(1990).
3.1.2.6. AI, Cr, Mn, Ga, Nb, Zr, Cu and W addition effects.
3.1.2.6.1. AI addition. For NdlsFe66Co10BsAll magnets, Yamamoto et al. (1987) have
observed that the Ndx(Fe,Co)y Laves phases become unstable with A1 addition,
decomposing into the Nd3(Fe, Co) (non-magnetic) and Nd(Fe, Co)3 (magnetic) alloys.
The former is supposed to be peritectically formed around Nd2(Fe, Co)24B grains to
magnetically separate each grain and suppress domain walls in the matrix. Therefore,
A1 addition increases coercivity but weakens intrinsic magnetic properties (Rodewald and
Fernengel 1988); consequently the improvement must be due to microstructural changes.
Such an increase of He has been measured by G. Schneider (1988), Grieb et al.
(1989), Knoch et al. (1989a,b, 1990a,b) and Fidler et al. (1989) for A1, A1N or A1203
additions. For instance, the maximum recorded Hc value can reach 1400kAm -1 after
annealing at 873 K of a Nd2FeT1B6A12.3 material. The as-sintered material has a coercivity
of 800kAm -1, and an undoped material exhibited a He value of 60kAm -1. The
microstructural properties studied by EDX and EDP vary from one sample to another:
probably due to a contaminant whether or not it has been detected. According to Fidler
et al. (1989) and Knoch et al. (1990b) five different structures have been recognized:
(1) fce (a = 0.520 nm) labeled "~'-Nd" (Nd-rich);
(2) rhombohedral, as NdsFe3B6;
(3) Laves phase, as Nd(Fe, A1)2 (low A1 content, stable up to 873 K);

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

153

(4) hexagonal (a=0.380nm and c= 0.600 nm), as Nd203;

(5) hexagonal (a=0.390nm and c=0.610nm), as "(zI-Nd''.
(Other phases, such as NdsFe2(B, O)x, Nd36Fe63Alx (2.5 ~<x ~<5), Nd33Fe67Alx (7
<x <25), have also been detected, by EDX experiments). One must remark that "(~and ~-Nd" do not exist in the phase diagram. The first one (hexagonal) is observed upon
A1203 addition and it is characterized by a low Fe content (4.5 at.%). Its parameters,
close to those of A-Nd203, are possibly slightly enlarged by the presence of Fe atoms in
the Nd203 lattice. The second (ice with 7 at.% Fe) remains fully unknown, and cannot
be compared to the fcc/3-La structure (Fidler et al. 1989). From these results the abovementioned authors find a correlation between the microstructure of their materials and the
wetting behavior of the liquid at the sintering temperature. Particularly it is responsible
for the separation of the Nd2Fe14B grains by a nonmagnetic phase which would suppress
the magnetic interaction between the grains. Therefore enhancement of He results from a
higher contiguity of the intergranular region and from the suppression of the nucleation
sites for reversed domains (surface hardening of hard magnetic grains).
3.1.2.6.2. Cr and Mn additions. Rodewald and Fernengel (1988) have observed that
Cr additions (up to 5%) increase Hc, do not change the anisotropy field and facilitate
numerous magnetic parameters (particularly the nucleation of reversed domains). On the
contrary, Mn addition decreases He.
3.1.2.6.3. Ga addition. Fidler et al. (1990) and Knoch et al. (1990a) report that Ga
additions enhance He. So, for a NdlsFe61-xCo17BTGax material for x--O, 1.7 and 4.24,
He has been measured as 584, 1308 and 1456 kAm -1 respectively. Various phases such
as NdCo2, Nd(Co, Fe)3, Nd(Co, Fe)4 B4, Nd(Ga, Fe, Co), Nd(Ga, Fe, Co)2 and Nd-rich
(hexagonal) have been detected as intergranular compounds. The hard phase showed the
presence of domain walls (Lorentz microscopy). In the bulk state Grrssinger et al. (1990)
reported a large number of magnetic properties for the Nd(Fe0.9-xB0.08Gax)5.5materials.
3.1.2.6.4. Nb addition. The magnetic properties and the microstructure of Nb-containing
(Nd, Dy)FeB materials are reported by Parker et al. (1987a,b). As Nb content increases
the He values are enhanced (1040kAm -1 without Nb and 1280kAm -1 with Nb). Four
phases, other than the hard magnetic one, have been detected: NdFe4B4, Nd-rich (at grain
boundaries, as small coherent precipitates) Fe2Nb (as large inclusions) and (x-Fe. The
presence of these precipitates would increase H e in sintering and melt-spun samples.
Allibert (1989) for NdlnNblFe78B7 alloys observed, as a function of the annealing
temperature, the formation of different alloys and compounds: Nb3Fe3B4, and Nd-rich
at 1273 K; Ndl.IFe4B4, Nd-rich (containing chlorine), pure Nd and Nd2FeB3 (as needles)
at 873 K (figs. 19a,b). It is noted that under equilibrium conditions Nb does not dissolve
in the Nd2FeI4B phase and cannot alter its intrinsic properties.
3.1.2.6.5. Zr addition. Identical conclusions have been reported for Zr-containing alloys
by Allibert (1989). At 1273 K, three phases have been detected: Nd2Fe14B (without Zr),
ZrB2 (as needles) and Nd-rieh (fig. 19c). As for Nb, it is concluded that the influence of
Zr on NdFeB magnets is typically a microstructural effect.
3.1.2.6.6. Si addition. The interaction between ion-plasma sprayed NdxFe93-xSi2B5 films
and Cu substrates has been studied by Linetski et al. (1990). The layers (40-70 ~tm thick)

154

M. GASGNIER

Nb3Fe3B 4

Nd2Fel4B
Nd rich

phase

Ndl+eFe4B 4
Nd2FeI4B
Y

Nd
X

Nd2FeI4B
ZrB 2
Nd rich phase

Fig. 19. (a,b) Microstructures and chemical compounds in the Nd14NblFe78B7 samples annealed (a) at
1273K for 30h and Co) at 873K for 800h. (The Y phase contains chlorine and the X needles are close to
the Nd2FeB 3 composition). (c) Microstructure and chemical compounds of a Ndj4Zr1Fe78B 7 sample annealed
at 873 K for 800h. (By courtesy of Dr. C.H. Allibert, Lab. Thermodynamique et Physico-Chimie M&allurgie
INPG, St. Martin d'H~res, France).

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

155

have a low Hc due to the formation of Nd-Cu phases. FiNer (1987) reports that under
production conditions a large amount of Si is found and that Si atoms replace Fe atoms
within the Nd2FeI4B grains.
3.1.2.6.7. Waddition. Bulk Nd~3Dy9Fe72-xCosWxB8alloys (x = 0--1.0) have been studied
by Rodewald and Schrey (1990). At 1373 K only 0.5-0.6 at.% W is dissolved, whereas
at 1123 K the solubility drops to 0.4 at.%. For thinned samples TEM observations reveal
the formation of rod-shaped particles (100-600 nm in diameter, 1-3 ~tm in length) relative
to the orthorhombic WFeB compounds. Hc can reach 1360 kA m -1 for sintered magnets
annealed at 803 K (x = 0.2 at.% W). The coercivity is determined by nucleation of reversed
domains.
3.1.2.7. Applications. To our knowledge few patents (see also Gasgnier 1991) have
been taken out in the field of thin films during the last years. Yamashita et al. (1988)
have prepared NdxFexoo-xByfilms (x = 13-27; y = 3-17) formed by sputtering under welldefined experimental conditions. The patent is concerned with preparing perpendicular
magnetic films.
3.1.3. RTiFe(Co) alloys (R=Sm, Nd, Dy)
3.1.3.1. (Sm,Nd)TiFe magnets. The works ofKamprath et al. (1988a, b, 1990), Liu et al.
(1988) and Cadieu (1988) deal with the synthesis of a (SmYiFe) phase with a high inplane intrinsic Hc, e.g. 2800 kA m -1 at 293 K (4640 kA m -1 at 5 K). (No corresponding
hard phase has been observed for the (NdTiFe) system.) The composition of this alloy has
been assigned to be Sm(Ti, Fe)ll. As for SmxCoy sputtered films, the magnetic properties
depend upon the experimental procedures. However, if such properties are of a high
quality, it seems that the crystalline characteristics of this system are very difficult to
define. Kamprath et al. (1988a) report that the magnetically hard phase coexisted with
Sm2Fe17, (Sm+Ti)Fe5 and Sm(Ti, Fe)12 alloys, and assigned to the Sm(Ti, Fe)11 phase a
tetragonal structure (a = 0.839 nm and c = 1.230 nm). In a second paper (Kamprath et al.
1990) this phase became hexagonal (a = 0.876 nm and c = 2.562 nm). Other alloys, such as
SrnFe3, Sm2Fe7 and Sm(Fe, Ti)2, have been also observed. The last one is indexed from
XRD to have a tetragonal structure (a = 0.709 nm and c = 0.781 nm). Demczyk and Cheng
(1990) report in the case of ion-beam milled ingots, that convergent e-beam diffraction
and EDP allow them to show that the SmTiFell structure is tetragonal (space group:
I4/mmm) with a=0.830nm and c=0.475nm. (Stadelmaier et al. 1987, reported, from
XRD, these parameters as a=0.857nm and c=0.476nm). EDX spectra (Demczyk and
Cheng 1990, Cheng et al. 1988) showed that the stoichiometry of the magnetic alloy is
close to 1:1:11.
3.1.3.2. (Sm,Dy)TiCo magnets. In the case of RTiCol~ alloys, Demczyk and Cheng
(1990) and Cheng et al. (1990) have shown the presence of 1:12 and 2:17 phase mixtures.
Convergent e-beam diffraction showed that the Sm2Fel7 phase is rhombohedral (space
group: R3m) (a = 0.841 nm and c = 1.211 nm referred to the hexagonal axes). In the case

156

M. GASGNIER

of DyTiColl a hexagonal 2:17 form (space group: P63/mmc) was found (a=0.796nm
and c = 0.795 nm).
Therefore, the crystallographic problems are not yet clearly resolved. This results
assuredly from the great difficulty to synthesize pure stoichiometric alloys. So, the
samples are constituted by the coexistence of several phases which are formed either
in the course of the deposition of the materials, or during annealing, or yet within
the amorphous--crystalline transition. This assertion can be illustrated by the results
obtained by Sahashi et al. (1983) on Sm(Co, Cu, Fe, Ti)7 magnets, which crystallize in
three polymorphic phases as Th2N17 (stabilized by Fe), Th2Nil7 (ordered states) and
TbCu7 (stabilized by Cu and Ti). The hard magnetic properties depend, according to
these authors, on the TbCuT-type phase stability.
3.1.4. Smx(Co,M)y substituted alloys (M=Fe, Cu and Zr)
Cadieu (1988) and Cadieu et al. (1989, 1990) have synthesized SmCos, Sm2(Co, Fe, Z r ) l 7
and Sm(Co, Fe, Cu, Zr)7 based sputtered films (2-10 ~tm thick). Such materials can exhibit
an in-plane (static) energy product of 15-30 MGOe. The authors have demonstrated that
this parameter strongly depends mainly on the texture of the films. In this case the caxes of the crystallites lie in the film plane. Such a configuration can be obtained by
strictly following the experimental procedures: sputtering, rare gas pressures, substrate
temperature, substrate materials, applied magnetic field during sputter deposition and
crystallization as a result. SmCo5 films with a (110) texture possess a substantially greater
intrinsic He (1120 kA m -1) than (200)-oriented layers. Moreover, a lesser degree of infilm-plane anisotropy was observed in films directly crystallized than for films crystallized
from an amorphous deposit.
In the case of substituted Co films Cadieu et al. (1989, 1990) have obtained a high static
energy product of 30 MGOe for specimens deposited onto sapphire substrate precoated
with A1203, and sputtered with a two-component Ar + 25% Xe sputtering gas. From XRD
the authors conclude that the 2:17 H phase allows the magnetization to remain in the
film plane (for the 2:17 R phase it does not). The layer must not have any a columnar
growth pattern during deposition, because this introduces porosity and lowers the effective
magnetization values. Lastly as a function of the sputter conditions, the Sm content can
vary (films could be richer or deficient in Sm when compared to the target) and this can
also change the magnetic properties of the films.
3.1.5. RCo(B) amorphousfilms
3.1.5.1. SmxCo~oo-xalloys. It now seems established that the main magnetic properties
of these materials depend on the cation content. That is what is observed by Numata
et al. (1988) and Lii et al. (1983) who discuss the origin of the anisotropy (see also
Buschow 1984, Gasgnier 1982). It is shown that among the amorphous SmxCol00-x
(9~<x~<44) magnetron sputtered films deposited onto substrates held at 313-413K,
perpendicular magnetization is observed for a composition range of 18 ~<x ~<26 and for
a narrow range of substrate temperatures of 333-363 K. In this region, the anisotropy is

THE INTRICATEWORLDOF RAREEARTHTHIN FILMS

!57

positive and possibly due to a microstructure mechanism. Between 363 and 413 K the
anisotropy is negative. It is concluded that no correlation occurs between stress-induced
and growth-induced anisotropy and uniaxial anisotropy. Fang et al. (1988) report on the
thermomagnetic effect of films subjected to high magnetic fields at low temperature. The
results are discussed in terms of Sm moments to the magnetization.
3.1.5.2. Amorphous NdCoB magnetic thinfilms. Amorphous NdCoB magnetic thin films
have been studied by Ratajczak et al. (1989, 1990a,b) and Ratajczak and G6sciafiska
(1990). The amorphous layers (14-100 nm thick) are deposited by flash evaporation onto
substrates cooled with liquid nitrogen. The starting material is a NdzCo14B compound.
Some films are covered with a SiO protective layer. EDP reveals the amorphous or poorlycrystallized character of the films. It is unlikely that chemical microanalyses have been
carried out. The authors report on the influence of annealing (from 523 K to 623 K)
on saturation magnetization, Hall resistivity, spontaneous Hall coefficient (these three
parameters decrease as T increases), and kinetics of structural relaxation. They notice
that film thickness plays an important role in saturation magnetization and coercive field
measured in the easy direction.
3.1.5.3. Applications. Thin (Sm, Pr, Ce, Nd, Y)Co65-85 ferromagnetic films have been
manufactured by Harada et al. (1988) useful for recording materials and miniature
motors. Miyamoto and Tazaki (1985) have fabricated magnetic discs, cards or tapes with
Sml-xCox films deposited by coevaporation onto polyimide substrates.

3.2. (TM/R)n modulated and multilayered films. (R = Y, La, Ce, Nd, Gd, Tb, Dy;
TM= Fe, Co, Ni)
3.2.1. (Fe/R)n films
3.2. l. 1. (Fe/Y),. By means of e-gun evaporation Badia et al. (1990) have investigated the
magnetic properties of various Y(~3.8 nm thick)/Fe (1.3-4.7 nm thick) modulated films.
By using a SQUID magnetometer they observed that Ms (at constant high applied field)
decreases as the substrate temperature increases (from 0 to 200 K) and increases as the
iron layer thickness increases. It is also shown that remanence and Hc are both small,
and that the magnetization is in-plane. M6ssbauer spectra revealed different magnetic
structure relative to each sample (i.e. the number of the Y-Fe sequences). It is concluded
that both polarization of the Fe layers by the diffused Y atoms and the surface roughness
are to be considered as a source for the random anisotropy.
From a dual e-beam gun system, Morishita et al. (1986, 1987) and Kajiura et al.
(1987) observe that various kinds of sequences lead to various magnetic behaviors. For
Fe thickness less than 0.6 nm magnetization disappears and an amorphous alloy forms;
for thickness between 0.6 and 1.2 nm the iron magnetization is reduced; and for thickness
more than 1.2. nm the iron bulk property is re-established.

158

M. GASGNIER

3.2.1.2. (Fe/La)n layers. Ferrater et al. (1990) have investigated the approach to Ms and
thermal demagnetizing processes for several (Fe/La)n modulated films (300 nm thick). It
is shown that Ms at 4.2 K is lower than for (Y/Fe)n films (except for larger Fe thicknesses)
and that magnetization decreases as the temperature increases. Moreover, it is asserted that
the La/Fe interfaceis constituted by mixing of Fe in the La which contributes to a small
paramagnetic state, and that as a consequence of the interplay between the ferromagnetic
layers and the Fe-La interfaces, the moments freeze over a wide range of temperatures
in a spin glass-like state with a great deal of short-range ferromagnetic order.
3.2.1.3. (Fe/Nd)n and (Fe/Dy)n films. These two systems are often compared for
magnetization measurements. In order to have a clear idea of the magnetic properties one
must point out a short paper by White (1990) who discusses the results of studies relative
to the alignment of R moments opposite to Fe moments. Indeed, Yoden et al. (1988),
Hosoito et al. (1988a,b, 1989), Mibu et al. (1989, 1990) and Shinjo et al. (1989a,b) find
that Dy moments align antiparallel to Fe moments, whereas Nd moments align parallel to
Fe moments. This might lead one to conclude that there is an antiferromagnetic exchange
interaction for Fe-Dy, whereas the F e N d exchange interaction should be ferromagnetic.
According to White (1990), who invokes Hund's rule, the exchange interaction in both
cases is, in fact, antiferromagnetic (see also Carbone et al. 1990). The main results given
by oden et al. (1988), Hosoito et al. (1988a,b, 1989), Mibu et al. (1989) and Shinjo
et al. (1989a,b) do indicate that, as a function of the temperature, the magnetization
changes direction, from in-plane at high temperature to perpendicular at low temperature.
The magnetic properties are discussed as a function of the Dy and Fe thicknesses. For
instance, with decreasing thickness of the Dy layer (below 3 nm), Ku of the Fe layer
becomes dominant, which probably originates from the magnetic coupling of the Fe/Dy
interface. In the case of thin Dy layers (below 2 nm) the Fe films are amorphous with
a magnetic transition at about 270 K, close to that of the bulk amorphous Fe. In the
case of Fe/Nd samples a gradual spin reorientation takes place which is discussed as
a function of the Nd moment. The experiments have been carried out from neutron
diffraction, MSssbauer, and SQUID magnetometer measurements. XRD clearly shows the
crystallographic changes of the samples, prepared with a dual e-gun system, according
to the thickness of the Nd layer: the dhcp phase of Nd metal and the bcc phase of Fe are
well-resolved.
Sellmyer and Shah (1987), Tiwald et al. (1988a,b), Shan et al. (1988) and Shan and
Sellmyer (1988, 1990) have investigated the structural, magnetic and magneto-optic properties of (Fe/Nd, Dy)n layers (300 nm thick) prepared by a multiple-gun system. The main
results show that Ku depends on the thicknesses of the Nd, Dy or Fe layers. The origin of
the parameter is discussed in terms of high density of R-Fe pairs oriented perpendicular
to the films. The authors report on a great number of magnetic properties as a function of
the thicknesses of the R and Fe layers: magnetization, volume anisotropy energy (which
is found to be larger than that of crystalline Fe), coercive field, and also magneto-optic
properties such as Kerr rotation and real part of the dielectric constant. XRD indicates that
a 4 nm Fe/5.6 nm Dy film has a crystalline structure, whereas a 1.5 nm Fe/1.4 nm Dy film

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

159

is badly crystallized. The indexation of the XRD patterns would allow one to conclude that
the c~-Fe and Dy phases coexist (Shan and Sellmyer 1988). For (Fe/Nd), layers Sellmyer
and Shah 1987 report that the samples are well crystallized for stacking layers 10 nm thick
and nearly amorphous for stackings 2 nm thick. This is unlikely: in the first case an
accurate re-examination of the XRD pattern does not allow one to conclude that Nd metal
is present. Indeed, the so-called (10.0) and (00.4) diffraction lines located normally at
20=28.12 and 30.25 respectively, are measured at 20=27.90 and 29.55 . The first
one probably corresponds to the (222) or (11.0) diffraction line of C-Nd203 or Nd(OH)3,
whereas the second one does not correspond to any known compotmd of Nd or Fe.
Other results reported by Brouder et al. (1986, 1988a) for (Fe/Nd)n films show that,
from XAS measurements, the local environment of the Fe atoms is strongly modified
between 77 and 293 K. So, at 77 K, the iron layer is strongly disordered, whereas at
298 K the pure spectrum of bcc Fe is recorded. The films were prepared by alternate
evaporation deposited onto various substrates. As a function of the thickness it is observed
that thick Fe layers induce a large magnetic moment in Nd. Moreover Ku is reported
as clearly induced by the structure of the film. Therefore the magnetic properties are
overall influenced by the mutual coupling of Nd and Fe: at low temperature Fe should
induce ferromagnetic coupling in Nd. Such a result is, thereby, not in agreement with
that reported by White (1990). On another part Pieeuch et al. (1988) and Baczewski et
al. (1989, 1990) have compared (Nd/Fe)n and (Tm/Fe)n structures (see sect. 3.2.1.6.)
3.2.1.4. (Fe/Gd)nfilms. The magnetism and structure of Fe/Gd)n films have been first
studied by Morishita et al. (1985). In the case of (3.30-4.6nm Fe/3.50-4.90nm Gd)50
multilayers a new magnetic layer having a "compensation temperature" has been found.
This temperature decreases as the thickness ratio of Gd and Fe layers decreases. This
was called a new "ferrimagnet". Later on Kamiguchi et al. (1989) and Fujimori et
al. (1990) have also pointed out this compensation phenomenon of ferrimagnetism.
They report magnetization and magnetoresistance variations as a function of the applied
magnetic fields and temperature. From magnetoresistance measurements, they observe
an anomalous field dependence which is discussed in terms of the spin-flop type
magnetization of artificial ferromagnetism. No structural results are reported. This is not
the case for films studied by Morishita et al. (1985) who report that the bce and hcp
structures of Fe and Gd respectively, are well observed with the thicker layers, whereas
for periodicity of about 1 nm the samples are anaorphous, and exhibit magnetization quite
similar to amorphous Gd30FeT0 coevaporated films.
One must mention AES analyses reported by Umemura et al. (1985) in the case
of (10nm Fe/10nm Gd)12 multilayers, which show that the atomic content of oxygen
throughout the Gd layers is almost the same as that of Gd. It is concluded, as for (Fe/Sm).
layers where the presence of Sm203 was detected, that the degree of oxidation clearly
affects Ku.
Theoretical studies of the phase transitions, magnetization, susceptibility and compensation points have been carried out by Camley (1989), Camley and Tilley (1988),
and LePage and Camley (1990). The results are discussed in terms of equilibrium

160

M. GASGNIER

structure of a magnetic superlattice formed from two ferromagnetic materials which

couple antiferromagnetically at the interfaces. It is concluded that a variety of phases
can exist in such a system.
3.2.1.5. (Fe/Tb)nfilms. As a preliminary remark one must draw attention to the fact that
the indexing of XRD and EDP reported by different authors is somewhat erroneous.
First, Sato and Habu (1987) indexed two X-ray diffraction lines, recorded at 19.15 and
26.20 , as the (10.1) and (110) planes relative to the rare earth and iron, respectively.
However these two lines, with interplanar spacings as 0.235 nm and 0.1745nm, do not
correspond to the (10.1) plane of the hcp rare earth (d10.1 ~0.275nm) nor to the bcc
iron (dll0 = 0.203 nm). Such lines should be linked to the formation of an intermetallic
TbxFey alloy. On the other hand, one intense line has been forgotten: located at 16.90 ,
it corresponds to the (200) plane of the rare-earth dihydride (d200 = 0.264 nm).
Second, Shin et al. (1987) for Tb/Fe95Co5 multilayers, claim that their XRD shows the
(10.2) plane of the terbium. This peak is located at 15, with a corresponding interplanar
spacing of 0.298nm, but the dl0.2 value of Tb metal is reported to be 0.210urn; The
indexing is clearly wrong, and this X-ray line should correspond either to the (200) plane
of TbO2 or to the (11.3) plane of Tb2Fe17.
Third, Cherifi et al. (1990) from EDP report that their films present three kinds of
crystals. In fact they observe the classical hydrogenation of rare-earth thin films: the TEM
image shows clearly the formation of large hydride crystals (with pole figures) which
have grown throughout the metallic Tb matrix. Moreover, the parameter of the dihydride
seems to fluctuate from one sample to one another, which is probably due to a small
amount of iron diffusion into the TbH2 lattice. Finally, another pattern can be indexed
as a fcc lattice with a parameter equal to 0.495 nm, and not as 0.525 nm which is the
value for the dihydride. This parameter (0.495 urn) corresponds to the oxinitride TbNzOx
(Gasgnier et al. 1976). Therefore the three types of crystals correspond to: Tb, TbH2 and
TbNxOy.
Fourth, N. Sato (1986) reports that after annealing two broad peaks at 17.50 and
25.50 should correspond to the (00.1?) and (110) planes of Tb and Fe. The deduced
interplanar spacings, 0.256nm and 0.179urn, should correspond to a TbxFey alloy, and
to the (200) plane of the y-Fe (fcc) phase respectively. Moreover, for small-angle-X-ray
diffraction two peaks are observed, whereas Shah and Sellmyer (1990) and Tanaka et al.
(1987) observe only one peak. AES profiles reported by N. Sato (1986) indicate a large
quantity of oxygen inside the Tb layers. In the same way Yamauchi et al. (1988) report
that the (00.1.9) diffraction line of the Tb is clearly observed. Consequently the magnetic
results reported by Shin et al. (1987) from ferromagnetic resonance measurements,
Yamauchi et al. (1988), N. Sato (1986), N. Sato and Habu (1987) and N. Sato et al. (1987)
must be regarded with great care. However the latter authors observe that for Fe/Gd, Tb
or Dy a compensation point occurs for a R content close to 25%. Ku, He, Ms and the
Curie temperature show clear changes at the compensation compositions. Although the
crystallographic results of Yamauchi et al. (1988) are not dearly described, these authors
observe a systematic change of the temperature dependence of magnetization. It also

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

t61

depends on the periodicity and the thickness ratio. Very thin films (below 1 nm Fe and Tb)
are composed of only ferromagnetically coupled Fe-Tb regions. However, ferrimagnetic
coupling Fe-Tb, ferromagnetic Fe and Tb, and/or magnetically compensated Tb regions
are found in films with 1 nm Fe/1 nm Tb layers.
Shah and Sellmyer (1990) study the layer-thickness dependence of hysteresis loops
and conclude that magnetization and Ku, which depend on the Fe layer thickness, can
be interpreted in terms of both the compositionally modulated distribution of Tb and
Fe atoms and their ferrimagnetic coupling. The authors report a modeling analysis which
offers information about the micromagnetic structure. The behavior of the anisotropy and
constituent magnetization near the compensation point are discussed. Tanaka et al. (1987)
in the case of FeCo/Tb films (240 nm thick), deposited by two-source magnetron sputtering, report on the magnetic properties as a function of the [(FeCo)+ (Tb)] compositional
modulation. The AES profile shows that oxygen is localized at the sample surface. It
is observed that with increasing compositional modulation the Kerr rotation and Ms
increase, whereas He, Ku and the Curie temperature decrease. Above a thickness of
2.9 nm, the magnetic anisotropy changes from perpendicular to in-plane. This indicates
a decrease of the ferrimagnetic coupling and an increase of the ferromagnetic coupling
between the same kind of atoms.
3.2.1.6. (Fe/Nd,Er, Tm)nfilms. Piecuch et al. (1988) and Baczewski et al. (1989, 1990)
have investigated the in-plane and perpendicular anisotropy directions of (Tm/Fe)n
and (Nd/Fe)n multilayers. The samples are prepared by alternate evaporation of the
two metals under a vacuum of 10-6pa, onto a substrate held at 410K. M6ssbauer
spectra (figs. 20, 21) show clearly that the two structures are quite different. Additional
magnetization measurements (hysteresis loops) allow one to conclude that (Nd/Fe)n
samples exhibit strong perpendicular anisotropy while (Ym/Fe)n (and also (Er/Fe)n)
samples exhibit in-plane anisotropy. It is concluded that it is not possible to consider
the anisotropy pair ordering with associated magnetic dipolar interactions as a major
source of the anisotropy. This is explained in terms of modifications of the R crystal

)
t;"

-..

"
..

Io

"o
*

~,

13 AFe
38ANd

13 ~ Fe

-4
a

VELOCITY (MM/S)

-~

o'

VELOCITY

,'

(MM/S)

Fig. 20. M6ssbauer spectra of (a) Nd/Fe and (b) Tm/Fe multilayers at T = 4 . 2 K and H = 0 . The Fe layer is
1.3 nm thick.

162

M. GASGNIER
.

,=,,

T= 4.2K
31 AFe

Tz 4.2K

31AFe

38 A Nd

0
VELOCITY

6
(MM/S)

12

i
-8

t
.4

I
I
0
4
VELOCITY(M M]S)

I
8

Fig. 21. Mtssbauer spectra of (a) NdlFe and (b) Tm/Fe multilayers at T=4.2K and H=0. The Fe layer
3.1 nm thick. (By courtesyof Dr. G. Marchat,LaboratoirePhysiquedu Solide,Universit~NancyI, Vandeeuvreles-Nancy,France).
field, i.e., according to the sign of the second-order Stevens factor which seems to induce
the anisotropy direction.
3.2.2. (Co/R)nfilms
3.2.2.1. (Co/(Pr,Nd))n films. Suzuki et al. (1987a,b) have investigated the variations
of Ms and Ku as a function of the modulation length and temperature. The films are
prepared by means of a dual e-beam system. The two magnetic properties increase as
modulation and temperature both decrease. It is deduced that Ku stems from Nd atoms
which are adjacent to Co atoms in the vicinity of the interface. A model of the magnetic
structure is reported. It is shown that Ku does not exceed that of coevaporated Nd0.25Co0.75
films. At 293 K anisotropy is low and films show an in-plane magnetization. XRD and
RHEED patterns reveal that Nd and Co atoms do not form a coherent interface and
that the films are amorphous. At 293 K anisotropy is low and films exhibit an in-plane
magnetization whereas at 77 K the films are perpendicularly magnetized. Miyazaki et al.
(1989) also observe an enhancement of the magnetization at 77 K which is explained
by the ferromagnetic coupling between Co and Pr or Nd at the interfaces. According to
these authors, Ku originates from the anisotropic pair ordering at the adjusting interfaces.
Takahashi et al. (1989) investigate the magnetic properties of (Co/Pr), films prepared by
e-beam evaporation under oxygen. Ms, Ku, polar Kerr rotation and hysteresis loops are
studied as functions of oxygen partial pressure and of temperature. It is observed that
polar Kerr rotations are large.
3.2.2.2. (Co/Gd)n films. T. Nakamura et al. (1987) have carried out syntheses of
amorphous (Co/Gd), sputtered films with anisotropy modulated superstructures. Webb et
al. (1985) have investigated the structural and magnetic properties of such specimens.
The modulated films were sputtered onto a thin Cr layer deposited previously itself
onto oriented substrates. From XRD it is observed that the longer sequences exhibit
both fcc and hcp Co structures, whereas smaller sequences are amorphous. Deposition

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

163

onto substrates kept at 653 K, leads to the formation of the fcc GdCo2 alloy. The
probable explanation for the absence of GdCo3 (the stable phase at the average
composition of the symmetric modulated films) is the kinetic preference for GdCo2 in
these highly nonequilibrium samples. Such an explanation is quite probable (Gasgnier
1982). It is observed that films are strongly textured with the (220) plane parallel
to the substrate, and that the d220 spacing decreases slowly as the period increases.
The magnetic measurements carried out at 293 K do reveal that the magnitude of the
Co moment opposite to that of Gd decreases as the bilayer thickness increases. Thus,
as the magnetization, with a field parallel to the film surface, quickly saturates, it is
concluded that a high degree of in-plane anisotropy exists. After deposition at 653 K
it is observed that coercivity reaches a maximum value (400kAm -1) as the bilayer
thickness increases from 4 to 8 nm. This is tentatively correlated to an increase of the
grain size of the films and not to the temperature. The magnetization data are analyzed
assuming the films consist of a mixture of GdCo2 and Co. It is found that as the
periodic thickness increases as the compensation temperature decreases, and that the
GdCo2 and Co moments exhibit an antiferromagnetic coupling. Webb et al. (1988b)
defined the Co/GdCo2 system as a "macroscopic ferrimagnet". They have investigated the
giant exchange anisotropy in Co/GdCo2 amorphous multilayers deposited at 293 K.
The exchange anisotropy is produced by heating near 770 K and rapid cooling. One
finds a temperature dependence of the exchange anisotropy with a compensation point
at l l 0 K , which can be controlled by changing the layering. It is also found that after
annealing the Gd2Co17 permanent magnet forms at the interface. Webb et al. (1988a)
have shown that Hc decreases near the compensation temperature (120K) instead of
diverging. Toxen et al. (1988) have studied multilayers (l-2 ~tm thick) Co/GdCox (x--2,
3, 5 and 8.5) in order to define the shifting of the hysteresis loops. It is observed that
the loops are unidirectionally displaced for x = 3-5, and not for x = 5-8.5. XRD indicates
that the samples are microcrystalline or amorphous. The former should be relative to
the GdCo5 alloy which should be at the origin of the unidirectional anisotropy, and not
Gd~Co17 as reported by Webb et al. (1988a, b).
The magnetization process of (Co/Gd)~ multilayers prepared by e-beam evaporation in a
vacuum of 10-5 Pa has been reported by Tsunashima et al. (1988). It is shown that around
150K a peculiar magnetization process occurs, which is attributed to a transition of
Gd spins from aligned to twisted states in agreement with the surface spin reorientations
described by Camley (1987). Okuno et al. (1986) have studied the electric resistivity and
the magnetization of modulated (Gd + Co) films prepared by rf magnetron sputtering with
double targets and a revolving substrate. As a function of exposure time (in seconds per
revolution) compositionally modulated homogeneous, amorphous and crystalline films
can be prepared. In the same way electric resistivity, Ms and Hc vary with the structure
variation and exposure time.
The effect of oxygen on modulated (Co/Gd)n rf diode sputtered films, has been
investigated by Choe et al. (1988). It is shown from AES that Gd, Co and O2 profiles
vary widely according to the layer thicknesses, and the application or absence of substrate
bias (65 V and 0 V). XRD showed the coexistence of the fcc and hcp Co phases, hcp

164

M. GASGNIER

Gd phase, and GdCo2 after annealing. As the period thickness increases the average
magnetic moment per Co atom increases. It is deduced that the 0 V bias substrates lead
to better coherent films, and that Gd atoms are magnetically deactivated by the formation
of Gd203.
3.2.2.3. (Co/Tb)n films. From RBS spectra Honda et al. (1986) have shown that
multilayered structures have a higher resistance to oxidation. The films (300nm thick)
are prepared by rf sputtering and a negative bias voltage applied to the substrate is
changed alternatively between 0 V and 100 or 200 V periodically. Anisotropy field, Ms
and He exhibit a compensation point with a number of layers up to about 20 layers.
With decreasing Ms, perpendicular He and Ku increase, which helps to stabilize the
written bits (down to 2 ~tm in diameter). Honda et al. (1987) report the variation of
the same magnetic parameters and Kerr rotation as a function of temperature, applied
field and thickness. It is observed that films prepared by modulating bias voltage are
modulated compositionally between 14.4 and 20.1 at.% Tb. The short periodic modulation
film (0.4 nm thick) realizes the ideal perpendicular rectangular hysteresis loop with the
highest Ku, and exhibits a high Kerr effect. N. Sato et al. (1987) report that Ms, He and
Ku have maximum values for Tb thickness close to 0.4 nm (these values are somewhat
smaller than those of (Fe/Tb)n films). Moreover Co substitution for Fe increases these
values. This is also clearly observed in the Kerr rotation angle. Therefore, (FeCo/Tb)n
multilayers with large Ms and He and low Curie temperature seem suitable for magnetooptical recording media.
One must point out the experiments of Marshall and Dobbertin (1986) who have studied
the cross-section of (Co/Tb)n films (400 nm thick), deposited by magnetron sputtering, by
means of a TEM. The Tb layers are amorphous and vary randomly in thickness, whereas
Co layers are polycrystalline and relatively uniform in thickness. The interfaces appear
to be fairly sharp but not smooth.
3.2.2.4. (Co/Dy)n films. Shan and Sellmyer (1988) have studied the magnetic and
microstructural characteristics of these multilayers prepared by multiple-gun sputtering.
An accurate indexing of the pattern leads to conclude that for thicker layers there
is coexistence of the hop Co [(10.0) and (10.2) diffraction lines] and of the DyH2
[(111) diffraction line] phases; and that for thinner layers there is probably formation
of a disordered alloy (the intense line measured at 20--36.25 does not correspond to
any known compound relative to Co or Dy). One must point out that, from magnetic
measurements, the interface is considered as the main source of the perpendicular
anisotropy. It is found to be described according to the single-ion model (due to spinorbit interactions) associated with the rare-earth ions (Shan et al. 1989, Sellmyer and Shah
1989).
N. Sato et al. (1987) report for (Co/Dy)n films results which are equivalent to those
found for (Co/Tb)n multilayers. The main difference is that the magnetic property values
are smaller for (Co/Dy)~ layers: for instance Ku is twice as small.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

t65

3.2.3. (2gi/R)nfilms
3.2.3.1. (Ni/Ce)n layers. The solid state reaction in (Ni/Ce)n films has been studied
extensively by Brouder et al. (1988a-d) and Guilmin et al. (1988). Amorphous samples
are prepared by alternate evaporation and then investigated by XAS in a dispersive mode.
Ni K and Ce Lnl edges are recorded as the temperature increases from 77 to 300 K. The
average valence of Ce is deduced as a function of the diffusion parameter according to two
different models. This parameter reaches a maximum value of about 3.25 in both cases.
For the thicker Ni and Ce films, the formation of the CeNi2 alloy has been observed. The
great similarity between the spectra leads to the conclusion that the local environment of
Ni in the amorphous phase is reminiscent of that of Ni in CeNi2.
3.2.3.2. ~i/Dy)n layers. Shan and Sellmyer (1988) report that magnetizations, at 293 K,
of these systems are much weaker than those of (Fe/Dy)n and (Co/Dy)n films. This is due
to the weaker exchange coupling in Ni as compared to Fe and Co.
3.2.4. Other multilayer systems
3.2.4.1. YCo/(Nd, Gd, Er)Co/YCo sandwiches. Amorphous samples prepared by dc triode
sputtering have been studied at 4.2 K, by Dieny et al. (1990). In the case of (Gd-Co) layer
a transition towards ferromagnetic arrangement takes place at a certain critical field; it is
associated with the formation of Bloch walls within the Co sublattices at the interfaces.
In the ease of (Nd-Co) layer anisotropy, magnetization and coercivity are not the same
in films of different natures. Under applied field, the magnetization reversal does not
occur at the same field values for the different layers. Consequently, Bloch walls within
the Co sublattices are again created and lie almost entirely inside the (Y-Co) films.
However, walls appear only when the magnetization in the less coercive layer reverses.
Both phenomena are involved in the case of (Er-Co) films. Such a study should permit
to create samples with unique planar Bloch wall parallel to the film surface.
3.2.4.2. (TbFe/Ni-Fe-Mo) bilayer films. The exchange interaction between amorphous
ferrimagnetic TbFe and crystalline soft magnetic material (Permalloy) Ni-Fe--Mo has
been investigated by Hellman et al. (1987). It is shown that an unexpected unidirectional
anisotropy occurs in such a system. It is observed that the direction of the easy axis in TbFe is in fact tilted at a small angle with respect to the normal of the Ni-Fe-Mo moment
which lies in the film plane. The tilt depends critically on the deposition parameters. Thus,
the unidirectional anisotropy induced in the Permalloy by the TbFe film is unique in that
it is large, stable to high in-plane fields, but reversible by application of a perpendicular
field.
3.2.4.3. (PtMnSb/TbFe) bilayerfilms. A PtMnSb/TbFe bilayer, prepared by rf sputtering,
allows one to improve the magnetic and magneto-optical properties of the PtMnSb film
as reported by Koyama et al. (1987). Compared to a PtMnSb monolayer the Kerr rotation
peak shifts to a shorter wavelength, and Hc is 6 times larger (~56kAm-1).

166

M. GASGNIER

3.2.4.4. (Fe/TbFe) bilayers. The magnetic exchange coupling phenomena have been
studied for Fe/TbFe bilayer systems by Marinero et al. (1988). The Fe moment causes an
uniaxial alignment in the plane of the substrate. Variations of Hc of the Fe layer and of
the shape of the hysteresis loops are discussed in terms of coupling interactions between
the ferromagnetic layer and the ferrimagnetic alloy.
3.2.4.5. (Zr-Fe/TbFe)n amorphous multilayers. Katayama et al. (1987) have studied the
variation of saturation magnetization as a function of the composition modulation period.
As the multilayers becomes thinner, Ms increases. This is explained by the fact that
magnetization of Fe within the ZrFe film is induced by the TbFe film.
3.2.4.6. (ZnS/GdFe) multilayers. The magneto-optic recording performance of the system
GdFe/ZnS/glass is improved by inserted a dielectric film (MgF2, SiO) between ZnS and
glass as reported by Sagasaki et al. (1985a,b).
3.2.4.7. (TbFe/SiO)n multilayers. Inoue et al. (1985), for TbFe/SiO/TbFe sandwich films
formed by evaporation, observed a considerable improvement in Kerr angle readout, and
little enhancement of the Faraday angle. The magnetic Kerr rotations of these amorphous
layers prepared by e-beam evaporation has been studied by R. Sato et al. (1988). It is
shown that such modulated films enhance the Kerr rotation to 25 degrees for a total
thickness of 650 nm. An important result is that the refractive index averaged over the
multilayers forms an effective refractive index of the film for the light with wavelengths
considerably large compared to the layer thicknesses. This parameter is thus effectively
modulated.
3.2.4.8. Insulating~magneto-optic~insulating systems.
3.2.4.8.1. Si3N4. Coulman and Anthony (1986) have shown that Si3N4 has better
characteristics than SiO or SiO2 to prevent interdiffusion between layers and oxidation of
TbFeCo films.
3.2.4.8.2. Si02/GdTbFe. A. Suzuki et al. (1988) have investigated the interface of Sirich SiO2/GdTbFe bilayers. From AES and XPS measurements they observe that the
formation of FeSiz, as an interface layer, works as a blocking layer against O2 penetration.
Moreover any excess Si plays an important role in the improvement of the magneto-optical
properties.
3.2.4.8.3. Dieleetric/Nd-Dy-Fe-Co-Ti/Si~N4-AlN on plastic. Aoyama (1988) has observed that Hc and Kerr rotation are enhanced in the case of the following sequence:
dielectric film/Nd-Dy-Fe-Co-Ti film/Si3N4-A1N film/plastic substrate. Such a system
prevents the magnetic film from degassing the substrate.
3.2.4.9. Amorphous (TbxFel-x/Al203)n multilayers. Such a system has been studied by
Krishnan et al. (1988). Samples are prepared by sequential rf sputtering. The A1203
thickness is kept constant at 4nm and the TbFe layer varies from 3 to 100nm. For
thicknesses <4nm the magnetization of the TbxFel-x (x=20 and 22) layer increases

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

167

strongly, indicating some partial oxidation of Tb. For larger thicknesses, magnetization
remains constant, but He and Ku increase. Variation of coercivity is controlled by various
parameters such as the microstructure of the film. The fact that Ku reaches a high
value, while He remains small, indicates that these two parameters are not simply
related.
3.2.5. Formation of sandwiched layers: R(Dy, Tb) ions implanted in Fe and Ni films.

Structure and magnetic properties

The effect of high-dose ion implantation on structure and magnetic properties of 3dmetal films has been investigated by Suezawa et al. (1983, 1985), Gond6 and Suezawa
(1986) and Gond6 et al. (1983, 1984), and by Tsukahara and Kanayama (1983) and
Tsukahara et al. (1983). It has been shown previously (Grant 1978, Poate 1978) that
implantation of metal ions into crystalline metal cause the formation of an amorphous
layer and alloying of the metals takes place at the same time. Such a phenomenon has
been observed whatever the energy implantation (70 or 400 keV) and for high doses
(between 1016 and 1017 ions/cm2). Gond6 et al. (1984), Gond6 and Suezawa (1986)
and Suezawa et al. (1985) report, for Tb+/Fe films, that the magnetic moment of the
Fe decreases as the Tb ion dose increases. Moreover an uniaxial magnetic anisotropy
is induced in Fe films (20-70nm thick) at 293 K with an easy axis parallel to an
applied static magnetic field. The magnetization behaviour is suggested to be close to
a multilayer structure consisting of a Tb-Fe layer sandwiched by Fe layers. Gond6 and
Suezawa (1987) have shown that magneto-optical rotations of Fe films are changed by
the implantation of Tb+ ions. For dosages up to 31015 ions/cm2 Faraday and static
polar Kerr rotations increase slowly, and then decrease rapidly for higher doses. Such
an increase has not been observed in the case of Tb-Fe films prepared by conventional
methods. In the case of Dy+/Ni films, Suezawa et al. (1983) and Gond6 et al. (1983)
conclude to the formation of a paramagnetic Dy-Ni alloy layer formed at high Dy+ doses.
Concurrently the magnetic moment of the Ni film decreases. Tsukahara and Kanayama
(1983) and Tsukahara et al. (1983) have implanted Dy+ ions (at 400keV) into Fe films
(50-150 nm thick). It is explained that Ms decreases monotonically with increasing Dy
concentration. It takes about 8 at.% to make a Fe film completely amorphous. They also
observed by Lorentz microscopy that implantation-induced changes in the domains and
domain-wall structures due to the reduction and destruction of anisotropies, decreased the
magnetization; and caused a possible variation of exchange constant and the distribution
of the alloy concentration.
3.3. R-Fe, Co/hydrogen and other gaseous contaminants: changes in magnetic

properties
In a previous chapter of this Handbook series (Gasgnier 1982) it has been demonstrated
that magnetic characteristics of amorphous R-TM thin films might be dramatically
changed when the samples are contaminated by absorbed (or adsorbed) gases. In this
connection, one must mention some other interesting results. Schelleng et al. (1984)

168

M. GASGNIER

and Forester et al. (1984) report that the inclusion of hydrogen sharply reduces the
compensation temperatures by about 200 K and the Curie temperatures by 50 K in the
case of Gdl_xFex:Hy films (x= 0.0-0.9, y = 0 , 1.25, 3.0; 1 ~tm thick). The exchange
parameters JFe-Fe and JFe-rd show small relative decreases whereas Jrd-Gd undergoes
a large change [from ferromagnetic (y = 0) to antiferromagnetic coupling (y = 3)], which
is explained in terms of anionic behavior of the hydrogen. Vittoria et al. (1984) notice
from ferromagnetic resonance field measurements that the Gd sublattice is more affected
than the Fe sublattice by the presence of hydrogen. The magnetic moment of the
former is reduced in comparison to the unhydrided material. Naoe et al. (1982) have
studied the influence of H2 addition on Ku for GdCo amorphous films. They observe
that this parameter increases when Co content is under 75at.% (Ms < 4 x 10-2T) and
decreases when Co content is over 75 at.% (Ms > 4 10-2 T). Moreover, the compensation
composition shifts to higher Gd content. Niihara et al. (1984) report similar results
for TbCo films. They add that H2 additions lead to the formation of rectangular Kerr
hysteresis loops. XRD (Forester et al. 1984) and EDP (Naoe et al. 1982) results
indicate that the RH2 compound is formed during hydrogenation. In the case of
Dyl_y(Or Hol_y)TbyFeHx(y = 0.85, 0.75; x = 0-3.3) compounds Annapoorni et al. (1989,
1990) report that Curie temperature, exchange field, magnetic moments decrease as the
hydrogen ratio increases.
The influence of contaminant gases on the magnetic properties of R-TM films is very
significant. One can mention, for example, that nitrogen decreases Ku as demonstrated
by Hoshi et al. (1982) and Hashimoto et al. (1987). The latter noted that N atoms
affect Tb atoms by decreasing the Tb sublattice magnetization in Tb-Fe-Co films, and
consequently remarkably decrease the magnetostriction. Heitmann et al. (1987) notice,
for GdTbFe films, that compensation temperature, Curie temperature and Ku decrease
strongly with increasing impurity concentration in the films. However the influence of
nitrogen is about a factor of 3 less than that of oxygen and water. They add that it
is due to the deactivation of Tb atoms. In the same way, Frausto et al. (1986) have
observed that carbon and oxygen impurities increase Ms and decrease He. However,
oxygen contamination seems to be more complex as reported by Tsunashima et al. (1980),
van Dover et al. (1986), and Frankenthal et al. (1986) for GdFe, GdCo, and TbF2 films.
To prevent oxidation of R-TM films it is now well known that protective layers as SIO2,
SiO or Si3N4 are used. However the problems due either to the diffusion of oxygen or
to the formation of compounds at the interfaces remain complex. For instance, consider
the formation of either R203 or FeSi2: if the first is formed the magnetic properties of
the films are destroyed (Anthony et al. 1986), while the second works as a blocking layer
against oxygen penetration (Suzuki et al. 1988).
The protection of amorphous R-TM films is very difficult to carry out. Forester et
al. (1983) have taken out a patent which demonstrates that a thin palladium layer (23 nm), overlaying a GdFe film, increases resistance to poisoning by atmospheric gases.
However the sample rapidly absorbed hydrogen. In this case the films did not fracture or
disintegrate and remained firmly adherent to the substrate.

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

169

3.4. RNi5 and RCos as hydrogen storage films

3.4.1. LaNis alloy
LaNi5 alloy is a well-known hydrogen storage medium. It is a suitable material for energy
storage, heat pumps, hydrogen purification (used as filters), H2/D2 separation and battery
electrode. However the problem of H2 absorption-desorption is very sensitive in the case
of crystalline alloys. Indeed, upon H2 absorption the LaNi5 lattice expands, and this
expansion leads to the pulverization of the material into particles, which complicates
technological applications. This has been reported by H. Uchida et al. (1984) who
observe particle pulverization under the lower stresses induced by the repeated volume
change of the alloy during the H2 absorption-desorption cycles. To prevent this scattering
phenomenon H. Uchida et al. (1987) have mixed LaNi5 powder with silicone rubber
and the mixture is formed into a sheet. Various procedures have been used to define the
optimal properties of thin and thick films. The crystalline state is an important parameter:
the amorphous or crystalline character plays a leading role in the absorption-desorption
mechanism.
3.4.1.1. Hydrogen absorption - Crystalline states. The determination of the H2 content in
thermally evaporated amorphous films (0.6-2 gm thick) has been reported by Sakaguchi
et al. (1985c). They observe, first, that the H2 concentration increases as film thickness
increases, and second, that the H2 content is higher at lower temperature (333 K). The
maximum amounts of H2 reaches 1.8-2.0 H2 atoms for a formula weight of LaNis. The
authors notice that the formation of a thin oxide layer at the surface of the films hardly
affects the H2 absorption. Wulz and Fromm (1986b) for films deposited by successive
evaporation of the two components, layer by layer, observe that the H2 absorption rate
for a microcrystalline film is 0.5-1.5 orders of magnitude higher than that for a crystalline
film. Preadsorption of oxygen reduces the reaction rate of the microcrystalline sample by
2 orders of magnitude. Moreover, the absence of H20 molecules at the surface could
explain the formation of La203 as surface segregation. S.K. Singh et al. (1985) report
that microcrystalline and amorphous films deposited by thermal evaporation have a higher
absorption capacity than that of the bulk. In the case of amorphous and microcrystalline
(about 80% crystallinity) sputtered films (0.53 and 0.57 ~tm thick respectively), Sakaguchi
et at. (1988, 1989a, 1990) have shown that the thicker films are more suitable to
absorb hydrogen. Moreover the concentration increases monotonously with increasing
H2 pressure whatever the variation in temperature and film thickness. In crystalline films
it is concluded that H2 atoms seem to occupy only sites strongly influenced by La,
whereas in amorphous layers they occupy two types of sites: one influenced by La and the
other comprised of more Ni atoms. The absence or the narrowing of a plateau, recorded
from pressure-composition isotherms in the case of powders, could be due to impurity
occlusions, or to H-H interactions affected by the film thickness, or by stresses induced
at the film-substrate boundary as explain by Larsen et al. (1981) for thin films (150nrn
thick), H. Uchida et al. (1989) and Kojima et al. (1989) for flash-evaporated layers, and
Sakaguchi et al. (1988) for sputtered samples.

170

M. GASGNIER

In the course of hydrogen absorption-desorption cycles Adachi et al. (1985a) observe

that amorphous films do not form powders even after 150 cycles. H. Uehida et al.
(1989) report that after 80 cycles, the crystalline films become amorphous by generating
dislocations. Their crystalline samples exhibit a large 1-12storage capacity (H2/LaNi5 > 5)
and were found markedly pulverized and peeled off from the glass substrate, whereas the
amorphous specimens with a lower solubility (H2/LaNi5 < 0.5) show neither pulverization
nor the peeling effect. By using aluminium and copper foil substrates, Sakaguchi et al.
(1986a) show that amorphous flash-evaporated films do not disintegrate after 75 cyles.
S.K. Singh et al. (1985) have observed by TEM that hydrogenation leads to enhancement
of amorphosity.
For sputtered films Sakaguehi et al. (1988) report larger thermal conductivities than for
bulk samples. The mechanical properties show that the film-substrate adhesion occurs in
the following order for the substrate used: Ni foil > Ni > AI foil > Cu foil > Cu > A1 >
glass. Such an order appears to be due to the difference of the thermal expansivity at the
film-substrate interface.
3.4.1.2. Electric resistivity. The effect of H2 absorption on the electric resistivity of LaNi5
films has been studied by Adachi et al. (1981, 1982, 1985a,b) and Sakaguchi et al.
(1985b,e). The films (0.3-1.8 ~tm thick) prepared by evaporation of the powder placed
on a tungsten filament and deposited onto quartz plate, are amorphous. As a function of
temperature, thickness, and H2 pressure, the variation of the resistivity with the time of
H2 absorption-desorption cycles presents an initial increase and a sharp decrease during
H2 absorption. Figures 22a--c show the variations of the electric resistance as a function of
the time of application of 1-12pressure. The resistivity decreases as the number of cycles is
increased; increases as the temperature is raised to 363 K; and saturates at a H2 pressure
of 1.2 106 Pa. However, Adaehi et al. (1985a) noticed that the low 1-12 uptake for the

6.5

..%

uE9

~8.C

,~ 6.0

6
5.5i

0
a

30
Time/rain

60

30
Time/min

3b
Time/rnin

6b

60

Fig. 22. Variation of the resistivity versus time during activation (first and second cycles at 313 K, others
at 363 K; hydrogen pressure 2.5106 Pa): (a) thick LaNi 5 film (0.63 btm); (b) thin LaNi 5 film (0.20 p.m);
(e) very thin LaNi 5 film (0.038 btm). (By courtesy of Prof. G.Y. Adachi, Dept. Applied Chemistry, Fac.
Engineering, Osaka Univ., Osaka, Japan).

THE INTRICATEWORLDOF RAREEARTHTHIN FILMS

17i

films in comparison with that for the bulk, is due to the amorphous character and to the
presence of a surface oxide such as Ni-free La203 which is a passivation layer for H2
absorption. For thinner films (50-150 nm) Ramakrisna and Srivastava (1987) claim that
the dependence of resistivity on hydrogenation time is originates solely from the surface
characteristics of the films. The resistivity, for aged layers, drops rapidly after exposure to
air. Such a phenomenon is explained to be due to the formation ofa La203-free Ni surface
layer. S.K. Singh et al. (1985) observe a curious variation of the electric resistivity which
decreases to nearly zero during the early exposure on hydrogenation and then increases
and reaches a saturation value. Larsen et al. (1981) report measurements of resistance as
a function of H2 pressure.
3.4.1.3. Hydrogen separation and permeation. Adachi et al. (1984) have investigated the
property of flash-evaporated LaNi5 films for the separation of hydrogen. This material is
less expensive than thin palladium films. The LaNi5 film (10 gm thick) is deposited onto
a stainless steel disc which is inserted into a special apparatus, where Hz-Ar and H2-N2
gas mixtures are blown. It is observed that above 333 K hydrogen does not permeate the
films. For Hz-C3Hs, Hz-CH4, Hz-N2 and Hz-Ar gas mixtures, Sakaguchi et al. (1986b)
have studied hydrogen separation in the case of stainless steel discs/Ni or A1 films/LaNi5
flash-evaporated films. The results are discussed first in terms of microcracks formed
during H2 absorption, second as a function of the film thickness, and third according
to the nature of the intermediate metallic layer. In particular it is observed that A1
allows the highest value of H2 concentration. Sakaguchi and Adachi (1990) have studied
the influence of CO on the hydrogenation of amorphous LaNi5 films and on hydrogen
separation. The H2-CO gas separation is performed by using films deposited onto Nicoated polyimide membranes. In the case of multilayer Ni (1.4 gm)/LaNi5 (0.1 gin) films
the H2 permeability (penetration rate) is greater than that for Ni films up to 373 K.
Moreover, LaNi5 films are found to have excellent resistance to harmful CO in comparison
with the crystalline bulk material (H2 is concentrated to more than 98 tool% in the
permeated gas).
Another interesting experiment has been reported for Hz-D2 gas mixtures by Sakaguehi
et al. (1989b). The rf magnetron sputtered LaNi5 films are deposited onto teflon and
polyimide membranes. Isotope separation is mainly influenced by the difference in
solubility of H2 and Dz atoms in metals, the difference in diffusivity in metals, and the
polymer's own isotope effects. LaNi5 films have a permeability coefficient about twice
as large as that of Ni films, and less than one-thousandth that of Pd films. Previously,
Adaehi et al. (1986, 1987), in the ease of amorphous WO3/Metal/LaNi5 sandwich-type
films, have studied the hydrogen permeability for various metals. The amorphous WO3
films, which have electrochromic properties, become blue only under the area covered
by the LaNi5 layer when H2 is introduced into the system. The authors conclude on the
following order for hydrogen permeability:
LaNis, Pt, Pd > Fe > Ni, Co, Ti > Mn > Cu > Mg, Cr > A1 > Au, Ag, Zn.

172

M. GASGNIER

In the same way, Shirai et al. (1990) have studied the H2 penetration into amorphous
V205 films. For a LaNis/VzOs/WO3 multilayer system, the diffusivity at the VzOs/WO3
interface seems to be much smaller than that at the Cu/WO3 interface.
3.4.2. Other RNis alloys
Other studies report that MmNi4.sMn0.5 films (Mm--mischmetal, generally as a (La, Pr,
Ce, Nd) mixture) have resistivities almost identical to that of LaNi5 films. However, the
former are more easily oxidized (Adachi et al. 1982). S.K. Singh et al. (1985) prepared
RNi5 films (R = Sm, Gd, Ho, Mm and Mm cerium free) by thermal vapor deposition;
these are initially amorphous, and then crystallize on annealing. S.K. Singh et al. (1985)
conclude that the amorphous state absorbs the hydrogen more easily.
3.4.3. RCos alloy (R = La, Sin)
Sakaguchi et al. (1985a, 1987) have studied the effects of hydrogen absorption on the
electric resistivity of LaCo5 films deposited as LaNi5 layers (Adachi et al. 1985a).
The resistivity behavior of LaCo5 specimens during H2 absorption-desorption cycles
resembles that for the LaNi5 films. However, LaCo5 films absorb less 1-12.But the amount
of absorbed H2 increases with increasing film thickness. The pressure-composition
isotherms indicate a monotonous increase with increasing pressure, while a plateau
pressure is absent.
For SmCo5 films Sakaguehi et al. (1985a,b) report that, as a function of thickness, the
resistivity varies differently depending on the 1-12absorption-desorption cycles. Hydrogen
molecules are absorbed on the surface and then dissociate into atoms. The dissolved
H2 anions diffuse into the films and then react with Sin, giving a highly conductive hydride
which lowers resistivity. It is asserted that the H2 concentration in SmCo5 is of an order
of magnitude 5 times smaller than that of LaNi5 films (in the bulk the ratio is close to
3/7).
3.4.4. Remark
R(Ni, Co)5 thin and thick films are available as materials for hydrogen storage. The
different authors referred to throughout sect. 3.4 report the possibility of a lot of
applications. However, to our knowledge, no patent has been taken out during the last
decade. This should be explained from the fact that these materials do not present
the perfect characteristics of stability during aging, annealing, H2 absorption-desorption
cycling, etc. The formation of compounds such as RH2, R203, ... seem to attest this
explanation.
3.5. Polytypic structures
Verma and Krishna (1966) have reported the main properties of polytypic materials.
Different notations have been used to describe the stacking sequence of successive packed
layers in the hexagonal unit cell. This has been chosen as the basic lattice because it

THE INTRICATEWORLD OF RARE EARTH THIN FILMS

173

includes at one and at the same time the cubic (C), rhombohedral (R) and hexagonal (H)
structures. The different structures are characterized by the following law relative to the
lattice parameters: a = constant, c = n x h (n = number of block layers, h = spacing between
two layers).

3.5.1. Ferromagnetic samarium-nickel alloys

Polytypic and intergrowth properties of SmxNiy alloys have been studied by S. Takeda
et al. (1982, 1983), S. Takeda (1983), Horikoshi et al. (1985) and Komura (1989). This
system appears somewhat complex. So, the following intermetallic compounds have been
observed:
3C;
SmNi2:
SmNi3:

3R;

Sm2NiT:

2H, 3R, 4H, 5T (trigonal) 9R and 12R"

2H, 3R, 4H, 5T, 6T, 7T, 9R, 12R, 15R and 18R.
SmsNi19:
A model of stacking sequences is shown in fig. 23.
HREM studies show, first, that such structures can coexist inside thin crystals
electrolytically polished, and second, that anomalous intergrowths can occur. For instance,
in the first case small "blocks" of Sm2Ni7 can grow in the matrix of SmsNi19, and in the
second case small "blocks" of Sm7Ni29 and SmNi4 can grow in the same matrix. Such
intergrowths can be caused by a small amount of concentration fluctuation in one part
of the specimen. Moreover, such crystals are characterized by various complicated defect
structures: non-periodic stacking sequences and stacking faults of the block layers. The
oT

OSrn

T .,I-T

)0~

X
X

~o0~o

~ ~

oo
c

~ X
)oo0

A
X

oo

')-I-~ B"

Oo

)0 Y

V I

)~~v

_G I

Fig. 23. Projections onto (1;20) plane along [010] direction of (a) SmTm2, (b) 3R polytype of SmzTm7,
(c) 2H polytype of Sm2TmT,(d) 3R polytype of SmsTm19, (e) 2H polytype of SmsTmj9, and (f) SrnTm5. (By
courtesy of Prof. S. Takeda, College of General Education, Osaka Univ., Osaka, Japan).

174

M. GASGNIER

presence of intergrowth phases and of defects could affect the physical properties of these
materials, but apparently the above-mentioned authors have not resolved this problem up
to now. So, the problem of layer sequences, intergrowths, syntaxies, . . . , becomes very
complex because a great number of compounds can be formed but also each of them
can be characterized by a wide range of crystalline structures (ten at least, in the case of
SmsNi19).
3.5.2. Samarium-cobalt alloys
The approach of using crystallographic polytypic structures was been reported first by
Cromer and Larson (1959) and later on by Parth6 and Moreau (1977) who have discussed
the problem in terms of stacking-blocks of various rare earth alloys. C.W. Allen et al.
(1974a,b, 1977); Melton and Perkins (1976), Melton and Nagel (1977) and Fidler and
Skalicky (1978, 1981) have investigated by electron microscopy the polytypic stacking
faults and defect structures in R2Co17 and SmCo5 materials. C.W. Allen et al. (1974a) have
reported the first atomic arrangements for six structural modifications (2H, 1R, 4H, 5H,
6H1 and 61-12) of R2Coa7. They correlate such structures, by the way of R sites, to magnetic
properties. For example they conclude that an intrinsic fault in the 1R structure could
provide a source for domain-wall nucleation. The authors illustrate the various stacking
sequences for different stacking faults in 2H and 1R R2Col7 alloys.
More detailed results have been given by Komura et al. (1981), S. Takeda (1983),
and Sahashi et al. (1983). The studies were carried out on thinned materials by electron
microscopy (TEM, EDP and HREM). This has allowed them to observe the classical
structures as: SmCo3-3R, Sm2CoT-2H and 3R and SmsCo19-2H and 3R; and to find new
long-period structures such as Sm2Co7-4H and 15R and SmsCo19-18R types. The lattice
parameters for the 15R and 18R structures are respectively:
15R:
Sm2Co7:
a=0.50nm and c= 19.50nm, i.e., 15l.3nm
18R:
S m 5 C o 1 9 : a=0.50nm and c=31.00nm, i.e., ~18 1.72nm
(1.30 and 1.72nm being, respectively, the values of the c parameters of the 2H basic
structure for each alloy). HREM imaging permits one to observe stacking faults inside
the stacking sequences. This also allows one to build these sequences, recognize the
different polytypic structures, and determine the possible stacking sequences. Thus, the
18R polytype relative to the Sm5Co19 alloy can be depicted according to three different
sequences. An interesting conclusion comes from the experiments by S. Takeda (1983).
He notices that, for electrolytically polished specimens, the axial ratios co/ao obtained
from EDP are about 6-8% greater than those obtained by the X-ray method. Such a
discrepancy was not observed in the case of crushed specimens. The author concludes
that possibly a chemical reaction of inclusions occurred in the course of polishing.
Other studies have dealt with the eutectic decomposition of these alloys. Linetski
and Salo (1989) report, from X-ray investigations, that thick (20-200~tm) sputtered
amorphous layers, as SmxCoy (18-34% Sm), can form a great number of alloys:
Sm2C017, Sm2C024, Sm2C026, SmC05, SmC07, SmCo8, SmCo8.5, SmCo9, SmCo9.5,
SmC012, SmC013 and other unidentified phases. The results are discussed on the basis of

THE INTRICATEWORLD OF RARE EARTHTHIN FILMS

175

the equilibrium diagram. It is concluded that the formation of a wide range of continuous
metastable solid solutions on crystallization from the amorphous state is probably due
to the similarity of the free energies and related crystal structures of the SmC05 and
Sm2Co17 phases. Such observations seem to be a general phenomenon as reported by
Gasgnier (1982) for other binary amorphous R-TM alloys. The amorphous system can
be also taken as thermodynamically far from equilibrium and the amorphous ~ crystalline
transition and the reerystallization which follow as the temperature increases, are often
very complex. Moreover one cannot exclude the important role of impurities (overall
oxygen) which can act as catalysts. Pan et al. (1989) have studied polished and ion-beam
thinned SmCo5 platelets between 673 and 1023 K. They observed first, the formation of
Sm2Co17 precipitates (as homogeneous centers formed at 693 K) and second, the eutectic
decomposition below 1023 K (formation of Sm2Co7 and Sm2Co17 phases). It is noticed
that at 1023 K the coercivity degraded abruptly, but it can be restored after annealing
at 1123-1273 K. Such a phenomenon has been attributed to the various microstructure
changes as a function of the temperature.
Shen and Laughlin (1990) have prepared near stoichiometric (Sm0.75Pr0.25)sCo19
materials. TEM studies were carried out after ion-milling of the specimens. Thermomagnetic analysis indicates that this alloy consists mainly of the 5:19 phase (>85%),
2:7 phase and 1:5 phase (<3%). The authors have observed the coexistence of the
5:19 (rhombohedral) and 2:7 (either rhombohedral or hexagonal)phases (faulted regions),
the presence of stacking faults and strained zones. They observed that within high-strain
zones the supersatured 5:19 phase transforms to less-strained 5:19R and 2:7H phases.
The transformation seems to be heterogeneous. Figure 24 shows HREM images of these

Fig. 24. Polytypephases relativeto a (Sm,Pr)sCo19alloy:(a) latticeimage of the 5:19Rphase; (b) latticeimage
of the 2:7H phase. (By courtesyof Prof. E. Laughlin, CarnegieMellonUniv., Pittsburg, PA, USA).

176

M. GASGNIER

Fig. 25. Electron diffraction patterns of the two (Sm, Pr)5Co19 phases: (a) [1010] zone axis for the 5:19R phase;
(b) [2110] zone axis for the 5:19R phase; (c) [i010] zone axis for the 2:7H phase; (d) [2110] zone axis for the
2:7H phase. (By courtesy of Prof. E. Laughlin).

phases. The patterns in figs. 2 5 a - d are representative o f each structure. In the interface
region they show a clear orientation relationship b e t w e e n the two phases.

THE INTRICATEWORLD OF RARE EARTHTHINFILMS

177

3.5.3. Rare earth trialuminides

The characteristics of rare arth trialuminides, as thin films, have been described previously
(Gasgnier 1982). However, a structural problem has arisen since the conclusions of this
review. Indeed, Schiffmacher (1983) has shown that EDP computer simulations do not fit
the patterns obtained by photography. The discrepancies may possibly be due to deviations
from either the exact stacking positions of the atomic planes as described in the bulk
(Buschow and van Vucht 1965), or from the sites of the cationic atoms in each plane, or
yet of the stacking order of layers themselves. This problem is illustrated by the presence
on TEM photographs of the (00.l) diffraction spots with 1= 1, 2 . . . . , 6 for instance in
the case of the 6H polytype (stacking sequence = 6), in contrast to computer simulations
where the spots with l = 1, 2 . . . . , 5 do not appear.
A.K. Singh and Srivastava (1979-1983, 1985) have obtained results similar to those
of Schiffmacher (1983) and Gasgnier (1982). They showed that such polytypic structures
migh be not only formed with pure Gd, Tb, Ho, Er and Y elements, but also with (Gd, Tb),
(Gd, Ho) and (Gd, Er) intermetallics. However the conclusions of Singh and Srivastava
do not appear to be clearly established. Indeed, they explain the formation of modulated
phases in terms of the reduction of electron energies which is manifested in the type of
contact between the Fermi surface and the Brillouin zone. But since the films are initially
amorphous and crystallize instantaneously under a focused e-beam such a hypothesis
seems irrelevant: first because the Fermi surface does not exist for an amorphous material,
and second, because in their theory the authors claim that this Fermi surface is spherical,
which is far from being true.
In another study Harris et al. (1984) have observed the formation of ERA13 by
interaction of thin erbium films deposited onto aluminium (non-anodized and anodized)
substrates. At low temperature (550 K) there is formation of Er2A1 and Er3A12, at 573 K
the ERA13 alloy is formed. At 673 K XRD reveals that this alloy is the only intermetallic
compound present and that it is stoichiometrically uniform. One must notice that in the
case of an ErD2 thin film significant interdiffusion takes place with formation at 673 K of
the ERA13 alloy. It is also observed that erbium does not alloy with A1203 between 673 K
and 773 K. This oxide inhibits the reaction of Er with A1.
3.6. R-noble metal alloys

3.6.1. The R-Cu system

3.6.1.1. Amorphous R-Cu films (R = Y,, Tb, Dy).
3.6.1.1.1. CUxYl-x. Martin et al. (1985) and Frigerio et al. (1985) have studied the
electronic and chemical structure of amorphous CUxYl-x films grown by triode sputtering.
XPS and UPS valence band spectra are characteristics of the Y and C u d bands. The
two bands overlap with significant electron transfer from Y to Cu atoms. The binding
energy of the Cu core levels increases on alloying, in apparent contradiction with chargetransfer arguments. From the Auger LVV transition, it is asserted that the effective
Coulomb interaction on Cu sites decreases by alloying with Y. Valence band, core-level

178

M. GASGNIER

spectra and optical properties give a realistic picture of the electronic structure of these
alloys.
3.6.1.1.2. Dy-Cu and Tb-Cu. Coey et al. (1981) and von Molnar et al. (1980, 1982a,b)
have studied the random anisotropy and spin freezing properties of Dy--Cu thick films
(0.6-160 ~t thick). It is shown that such amorphous alloys, as highly anisotropic random
magnets, do not develop a spontaneous moment at any temperature. These systems are not
spin glasses. The spin-freezing transition at 18 K and the asperomagnetic model proposed
by Coey et al. (1981) have been discussed by Pickart et al. (1984) and Moch et al. (1986)
in terms of spin-like glass behavior for Dy-Cu and Tb-Cu alloys.
3.6.1.1.3. Cu addition to Tb-Fe-Co. The addition of Cu to TbFe, TbFeCo and TbCo
alloy films improved the magneto-optical characteristics and overall He and Kerr rotation
angle (K. Uchida et al. 1987, Nakajima and Sato 1986, Told 1986).
3.6.1.2. Crystalline R-Cu films (t? = Y, Ce, Dy, Ho, Er and Yb).
3.6.1.2.1. (Cu/Y)n. R.W. Johnson et al. (1989a,b) have shown that compositionally
modulated (Cu/Y)n films form an amorphous CuxYl-x phase at 293 K. The amorphization
reaction occurs both at the original Cu/Y interface and within the Y grain boundaries
(fig. 26). The authors refer to the Y hcp --+ Y fcc transition in the case of the thinner

Fig. 26. High-resolution phase-contrast image showing

grain-boundary amorphization reaction in Y/Cu thin films.
The reaction results in a nonatomically smooth reaction
front. The 0.3 nm interplanar distance corresponds to the
(111) plane of YH2 (see text). (By courtesy of Dr. C. Ahn,
California Institute of Technology, Pasadena, CA, USA).

Y films, which is unlikely. However, XRD clearly shows that the interplanar spacing,
which is close to 0.3 nm, is well correlated with the (111) reflection of the YH2 compound.
Once again this misunderstanding is greatly prejudicial to the theoretical models given
by the authors.
3.6.1.2.2. Ce/Cu interface. The Ce/Cu interface has been studied by means of XPS and
UPS measurements by Braaten et al. (1989). The cerium metal is deposited onto thin Cu
film (30nm thick). It is observed that, at 293 K, a thick interface is formed (>10nm)

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

179

which contains about 80 at.% Ce ions. These Ce ions are in their trivalent state. After
exposure to oxygen Ce 4+ ions are formed and the copper film is not oxidized.
3.6.1.2.3. Cu-Dy, Cu-Ho, Cu-Er. The crystallization behavior in amorphous vacuumdeposited Cu-Dy, Cu-Ho and Cu-Er films has been investigated by Shikhmanter et al.
(1982, 1983a) and Venkert et al. (1987). Either Dy(or Er)0.a0Cu0.60 films or (Cu/Ho)n
(45 at.% Ho) multilayers have been studied. The nucleation and crystallization processes
are determined by TEM and EDP. As expected from the R-Cu phase diagram (Franceschi
1982), the DyCu and ErCu alloys (CsCl-type structure) are formed at about 430K.
Kinetics of crystallization and nucleation sites are discussed and analyzed with simple
models. The presence of the R203 compound has been observed in all cases. For
(Cu/Ho)n samples it is reported that the initially crystallized films (160 nm thick) become
nearly amorphous after heating at 373 K, and become crystalline again at 533 K due to the
formation of HoCu2 and Ho203. The mixing process and the depth profiling have been
analyzed by XPS, UPS and AES measurements. It is observed that, first, the topmost
layer formed is Ho203, which is induced by surface segregation of Ho, and second, the
interfaces are not sharp due to interdiffusion between the two metals.
3.6.1.2.4. Yb/Cu. Y.S. Huang and Murgai (1989) report that Yb/Cu films annealed up to
873 K form the YbCu2 alloy. In the course of this reaction the ytterbium valence changes
from 3 to 2.2. It is asserted that the YbCu2 alloy is characterized by a homogeneous
mixed valence state. The results are discussed in terms of the number of f holes and 4ff
contributions to the photoemission spectra.

3.6.1.3. R-Cu samples as precursors for synthesis catalysts. Nix and coworkers (Nix and
Lambert 1987, Nix et al. 1988a, 1989a) and Jaffey et al. (1989) have shown that at 3001100K pure Nd and Sm overlayers on Cu (100) and Cu (11 l) substrates transform to
Nd(Sm)Cu and Nd(Sm)Cu5 intermetallie phases. The interaction of Nd with Cu has been
widely studied by AES, XPS, UPS and LEED. It is shown that at 300 K pure Nd films
grow on the Cu substrate by a layer-by-layer mechanism. At higher temperatures (800 K)
rearrangement occurs with the formation of alloy phases. Nix et al. (1988a) compared the
results obtained for Cu (111) and Cu (100) substrates, and those reported by Jaffey et al.
(1989) for the Sm/Cu (111) system. Such studies have been carried out in order to broadly
investigate the properties of the Nd(or Ce)-Cu intermetallic catalyst precursors as reported
by Nix and Lambert (1989a,b), Nix et al. (1987, 1988b, 1989b), Owen et al. (1987),
Bryan et al. (1988), Hay et al. (1988) and Jennings et al. (1989). The oxidation of Nd (or
Ce) overlayers; oxidation by dissociative chemisorption of CO; sorption, chemisorption
and desorption of H2; treatments with CO2, CO/Hz, N2 and N20/H2; and, overall the
activation and performance of methanol synthesis catalysts have been studied. The whole
of these results, obtained by different experimental procedures, shows that the mechanism
by which low-temperature methanol synthesis occurs on these catalysts is quite different
from that which operates on commercial Cu/ZnO/A1203 catalysts (Nix et al. 1989b).
Jennings et al. (1989) have shown that for R/Cu precursors, C Q causes strong irreversible
deactiviation, whereas inclusion of Ti, or A1, or Zr enhances poison resistance.

180

M. GASGNIER

3.6.1.4. Ion-beam mixing. In order to prepare the new high-Te superconductors as thin
films, a large number of methods have been used. One of them, ion-beam mixing, has been
investigated by Borgensen and Lilienfeld (1989) and Mathevet et al. (1990). The former
have irradiated their Cu/Y/Cu samples ~vith 600 keV Xe ions. The results are analyzed by
RBS spectra. The formation of YCu2, YCu4, YCu5 and YCu7 alloys has been determined.
It is concluded that, as a function of temperature, the mixing phenomenon varies strongly.
At 100K it is dominated by thermal spike effects, whereas at 323 K the rapid growth of
the hexagonal YCu7 phase becomes predominant. In another investigation, Mathevet et al.
(1990) have irradiated La(OH)3/Cu bilayers with 3.65 MeV Au ions, at 300 K and 700 K.
The results are analyzed by RBS spectra and XRD. It is concluded that a homogeneous
depth distribution is not reached, and that ion irradiation simultaneously breaks down
the La-O-H bonds and pushes in the Cu atoms. This induces the formation of unknown
alloys and/or compounds. These results seem to indicate that the above methods of ionbeam mixing do not form a preferential route for fabricating the ternary RBa2Cu307 and
La2_xSrxCuO4 ceramics.
3.6.2. The R-Au system
3.6.2.1. Amorphous GdAu and GdFeAufilms. The magnetic properties, Hall conductivity,
Curie temperature and resistivity of GdxAUl-x films (x = 0.26-0.72) have been studied
by Gambino et al. (1981). It is suggested that conduction-electron spin polarization
determines the magnitude of the spontaneous Hall effect. Gambino and McGuire (I 984)
have determined that the addition of a non-S-state heavy lanthanide (Tb, Ho, Er or Tin)
to ferromagnetic Gd-Au does not increase the spontaneous Hall effect to a greater extent
than would be caused by a similar increase in Gd. In contrast, the addition of Nd leads
to a significant increase in this transport parameter, von Molnar et al. (1982b) have
shown that Dy-Au amorphous alloys are highly anisotropic magnets. They develop a
spontaneous moment at any temperature, and never reach infinite susceptibility. One must
also point out the work of Hartmann and McGuire (1983) and Hansen and Hartmann
(1986) relative to the magnetic and magneto-optics properties of GdFe-Au films. We
also note the investigations of De Luca et al. (1981) relative to the bias-field dependence
of domain drag propagation velocities in GdCoAu bubble films.
3.6.2.2. Crystallinefilms. Schwarz and Johnson (1983) have studied (La/Au)n multilayers
at temperatures of 323-353K. The unannealed samples are characterized by the
coexistence of the well-crystallized Au and ~-La (fcc) and/or (~-La or LaH2 phases (the
interplanar spacings being equal). The most interesting result is the formation, after
heating at 353 K for 4 h, of a single amorphous phase. This is essentially due to the
fast diffusion behavior of Au in La, and the existence of a negative heat of mixing in
the amorphous alloys. The latter provides the necessary chemical driving force for the
reaction.
Raaen (1990) has investigated the Ce/Au system by means of XPS. Ce films of various
thicknesses are deposited onto Au films (>20nm thick). XPS core-level intensities for

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

181

Au 4f and Ce 3d emissions indicate the formation of a mixed amorphous interface (5 nm

thick) with an average volume content of Ce estimated at -35%. The Au 4f level shifts
by 0.8 eV to higher binding energies which may be explained by changes in final-state
screening caused by alloying of Ce and Au.
Shikhmanter et al. (1983b) have carried out TEM experiments in order to study
the crystallization behavior of some R-Au (R = Gd, Tb, Dy and Er) vapor-deposited
amorphous films (120nm thick). Crystallization takes places in the temperature range
of 463-513K, and further heating by an additional 50K leads to the formation of
the RAu alloys (CsC1 structure type). Further annealing at 533 K induces an allotropie
transformation such as CsC1 type (cubic structure) ~ CrB type (orthorhombic structure).
The former is metastable, while the latter is, as in the bulk, more stable at low
temperatures. The presence of R203 crystallites can act as catalyst for the transformation.
It is concluded that conditions amenable to heterogeneous nucleation will appear on
the R-Au films at higher temperatures than in the R-Cu films (413-423 K) or R A g
films (388-398 K).
L.I. Johansson et al. (1982a,b) have studied the chemically shifted surface core-level
binding energies and surface segregation in Eu-Au and Yb-Au alloys. Photoemission
spectra show that Au atoms deposited on the top of an Eu film dissolve into the film
more readily at 293 K than for an Yb film. The authors studied mainly the intensity ratio
[surf/Ibulk of Au, Eu and Yb 4f lines, and the chemical shifts upon alloying for different
Au, Eu and Yb thicknesses and annealing temperatures. The most important feature is
the persistence of the rare-earth surface 4f signal with increasing Au content.
3.7. R-Pd alloys (R=Ce, Eu, Er, Yb)

3.7.1. Valence change in R-Pd alloys

3.7.1.1. Ce-Pd amorphous crystalline samples. Comparison to other materials. The Ce
valence-state disparity between crystalline and amorphous Ce-TM alloy films has been
studied by Lu et al. (1985, 1986) and Croft et al. (1984, 1985). For CePd3 and CePdt.5
films (700 nm thick) Lm absorption spectra show that crystalline samples possess a lower
valence state (3.17) than amorphous films (3.30). Moreover, Ag substitution for Pd in
crystalline CePd3 acts to stabilize nearly pure Ce 3+ behavior. In the case of Cel-xMxPd3
materials (M=Y, La), the Ce valence decreases from 3.17 at x = 0 to 3.09 at x=0.75
(M = La), and increases to 3.23 (M = La) and 3.29 (M = Y) for x = 0.40. (In the case of Th 4+
and U 4+ substitutions the valence decreases to 3.10 at x = 0.40.) The results are discussed
in terms of lattice-parameter changes, electronegativity arguments and band filling. These
valence changes have also been determined for other metals, such as Cu, Ni, Co, Fe, Mn,
Mo, Ru, Rh and A1 (or AlSo).
Parks et al. (1983) have established a linear relation between the Lm-based valence estimates (from 3.00 to 3.21) and the bulk-property based valence estimates
(from 3.00 to 4.00) in the Ce(Ag or Rh, Pd)3 systems. Then, Parks et al. (1984) in the case
of fracturing clean RPd3 samples (R = La, Ce, Pr, Nd) have observed a 4f-derived resonant
photoemission. The authors speculate that a mixed-valence state can occur in Ce systems

182

M. GASGNIER

only if the 4f holes are locally screened. The contraction of a valence electron may explain
the anomalously large lattice contraction observed in these systems.
3.7.1.2. Eu and Yb-Pd amorphous specimens. For amorphous films of EuxPd~_x
(0.16 ~<x ~<0.35) and YbxPdl-x (0.29 ~<x ~<0.66), Malterre et al. (1986) have shown that
Lm edges demonstrate a concentration-induced valence change. The samples (1 ~tm thick)
are prepared by co-evaporation. Two peaks indicating a mixed-valence state are recorded.
As a function of the concentration it is shown that Eu valency decreases from 3.00
to 2.00 as x increases (in the same way as in crystalline EuPd5 and EuPd2 samples).
The local Eu environments in amorphous samples look like those of the bulk. For Yb this
valence decrease is observed between x = 0.29 (valence 2.90) and x = 0.66 (valence 2.12).

3.7.2. Rx-Pdy crystallographicproperties

The phase diagrams of the R-Pd alloys (R = Y, Sin, Gd, Dy, Ho and Er) have been first
investigated by Loebich and Raub (1973a,b). They report the formation of 7 different
alloys. However, Moreau and Parth6 (1973) and Palenzona and Iandelli (1974) have
shown that the SmsPd2 and Sm4Pd5 compounds are of Sm4Pd3 and Sm3Pd4 stoichiometry.
Such results have been also corroborated by Jordan and Loebich (1975) from XRD and
magnetic measurements. They also deduce an effective valence of +3 for Sm.
In the range of low R content (below 20 at.%) extensive work has been done by Brooks
et al. (1976), Hirst et al. (1976) and Hughes et al. (1980a,b). Physical and mechanical
properties exhibit numerous anomalies. These seem to be correlated to the presence of
long-range order in the Pd-R solid-solution alloys close to the solid solubility limit. In
the case o f a Pd-12.5 at.% Ce electropolished disc, D.A. Smith et al. (1982) have shown
by TEM the existence of the Pd7Ce superlattice. The crystallographic structure would
be a stacking fec sequence such as A(4Pd)B(Pd3Ce)C(4Pd). The presence of this phase
should explain the physical anomalies reported by the above-mentioned authors. In the
same way, Kuwano et al. (1981) have observed equivalent EDP, and conclude that the
PdTCe alloy is formed as a superstructure and that it is stable below 1170K. They also
found another phase, PdsCe, an ordered stacking structure with a rhombohedral symmetry.
(The crystallographic structure of the PdsR alloy has not been reported by Jordan and
Loebich (1975).) Then, Kuwano et al. (1985) report that the PdsCe phase was in fact
defined by a hexagonal symmetry with two polytypes of stacking structure, i.e., L-PdsCe
and H-PdsCe types, stable at low and high temperature respectively.
The Pd-rich side (5-21 at.%) of the Pd-R systems (R=Y, Ce, Eu, Sm, Gd, Dy, Ho,
Er) has been extensively investigated by Sakamoto et al. (1985, 1986, 1988, 1990a,b,
1991) and Takao et al. (1989, 1990). The phase diagrams, the electric resistance versus
temperature, and the XRD and EDP results (fig. 27) are systematically reported for alloys
prepared by arc-melting. The hydrogen solubility has been measured in ordered and
disordered PdTCe alloys. The main results for the yttric lanthanides appear to have the
following characteristics:

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

183

Fig. 27. Electron diffraction patterns of annealed Pd-x at.% R alloys. (a,b) Pd-12.5 at.% Sm: the zone axes are
[001] and [010], respectively (PdsSm phase). (c) Pd-14.3 at.% Eu: the zone axis is [101] (coexistence of the
Pd7Eu and Pd3Eu phases). (d) Pd-17.5 at.% Gd: the zone axis is [101] (coexistence of the PdvGd and Pd3Gd
phases). (By courtesy of Profs. Y. Sakamoto and K. Takao, Nagasaki Univ., Nagasaki, Japan).

(1) F o r m a t i o n o f an ordered structure o f the LI2-type Pd3R phase and/or o f P d s R

(R = Sm) and the formation o f Pd7R;
(2) It is a c c o m p a n i e d b y a peritectoid reaction as: c~-Pd + L I 2 - t y p e Pd3R ~-~ Pd7R;

184

M. GASGNIER

(3) The transition temperature (order-disorder transition) decreases with decreasing

R content, it corresponds to the solid solubility limit of R in the (t-Pd solid solution
below the peritectoid reaction temperature,
(4) The magnitude of this transition temperature may be attributed to the metallic
radius of the R metals and decreases from Eu to Er.
Figures 27a,c illustrate the PdsSm phase for the Pd-12.5 at.% Sm alloy. The pattern
in fig. 27a, relative to the Pd-14.3 at.% Eu alloy, shows the coexistence of the Pd7Eu
and LI2-type Pd3Eu phases. Lastly, EDP (fig. 27d) of the Pd-17.5 at.% Gd alloy, is
commonly observed for all the PdTR alloys. It is characterized by two types of superlattice
reflections.
One must point out that in the case of Pd-Ce alloys Sakamoto et al. (1990a)
have greatly refined the results given by D. Rossi et al. (1975) and Kuwano et al.
(1985). It is reported that two PdsCe phases, H-PdsCe and L-PdsCe (i.e., formed
at high and low temperatures), with hexagonal (CusCa-type layers) and fcc (longperiod antiphase superstructure) structures can be observed. The formation of the Pd7Ce
phase is accompanied by a peritectoid reaction: (x-Pd+L-PdsCe--+~Pd7Ce for higher
Ce content, whereas a phase transition a-Pd ~ (~-Pd+ PdTCe occurs for lower Ce content.
At 1074-t-10 K the L-Pd5 Ce ~ H-Pd5 Ce transition occurs. This temperature is higher than
those determined for heavy lanthanides.
3.8. Superconducting materials
The main results obtained in the field of thin superconducting films and crystals
have recently been summarized by Gasgnier (1991). This review paper brought into
perspective the numerous efforts which have been made, since 1960, to try to obtain
better superconducting performance with various alloys and compounds. But, since
the discovery of the new high-Te superconducting ceramics (RBaCuO, BiSrCaCuO,
PbSrYCaCuO, YSrCuO, NdCeCuO, ... ) the importance of the other materials has
diminished abruptly. It is not feasible in this chapter to deal with the properties of the
thin films of these new materials, because this subject is nowadays much too extensive.
(The reader should refer to specialized review papers and Handbooks).
However, one must draw atteotion to articles which report on some new (nonhigh-temperature) supercrnducting compounds. Cendlewska and Rojek (1987) and
Bohdziewiez et al. (1989) have prepared, by cathode sputtering, thin LuRhl.2Sm 4 films.
Their Tc is 3.72 K, i.e., somewhat lower than for the bulk (4.08 K). Structures on the
resistivity curve near 4.70 K indicate, possibly, the existence of another superconducting
phase with higher Tc. In the case of ErRhnB4 and Er0.sHo0.sRh4B4 ferromagnetic
superconducting films, Mancini and Noee (1986, 1987) have studied the electromagnetic
properties in the presence of an external homogeneous magnetic field and the reflection,
transmission and absorption coefficients of an electromagnetic wave normally incident
on the surface of the layer. It has been found that there exists a critical temperature,
below which spontaneous magnetization occurs as a function of thickness. It is
concluded that ErRh4B4 exhibits a sharp decrease of the reflection coefficient near

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

185

0.8K (first-order transition to a normal state with ferromagnetic ordering) which

seems to be a manifestation of spin-fluctuation effects due to the ferromagnetism. For
Er0.sHo0.sRh4B4 this coefficient shows a very small deviation, therefore superconductivity
and magnetic order cannot coexist in this compound. These results confirm previous
investigations relative to the surface impedance penetration depth (C.Y. Huang et al.
1985).
One must also point out that Xu and Shelton (1987) and Xu (1987) have synthetized
Lu2Fe3Si5 alloys with a Tc of6.11 K (the addition of transition metals reduced Te). In the
same way RsOs4Si15 (R = Y, Er, Tb) and Y4Os4Sil3 bulk compounds are superconductors
with a Te of 10.8, 5.5, 3.2 and 9.8 K, respectively. Under hydrostatic pressure Tc degrades
significantly (Kuboth et al. 1989). One can conclude that the synthesis of new materials
is always possible in the near or remote future.
3.9. Miscellaneous R-M alloy films and thin crystals

3.9.1. Y, La-Pb multilayer samples

The formation of the RPb3 alloy has been studied by Ronay et al. (1982). This investigation has been carried out by means of multilayer samples such as: Y, La(5-20 nm)/Y, LaPb("phase-in" layer)/Pb(200 nm)/Al(100 nm). After aging, RBS experiments show that no
compounds are formed between La and Pb, nor between Y and Pb. Moreover the grain
size of lead is not reduced. The formation of holes and hillocks around the holes is clearly
observed, and is related to the oxidation of the interface.

3.9.2. Transport properties in U-R films (R = Gd, Yb)

The competing magnetic, localization and phonon effects on the transport properties of
amorphous magnetron-sputtered U-R films (200 nm thick) have been described by Freitas
et al. (1988). While UxYbl-x films are well nonmagnetic, it is clearly shown that UxTbl_x
and UxGdl-x samples are, respectively, spin glasses and random-anisotropy dominated
systems. For both these systems U addition leads to a linear decrease of the freezing
temperatures. For x = 0.40 these temperatures are high enough to indicate positive average
exchange. The signature of U moment is not found. The magnetoresistance is negative
(as it is for UxDyl-x and UxNdl-x systems). Coexistence between spin disorder and
weak localization is evidenced from magnetoresistance and resistivity measurements in
UxGdl-x samples.

3.9.3. Quasicrystal Co-Er alloys

An icosahedral dominant phase was reported by H. Li et al. (1988) to exist in rapidly
quenched CosoEr2o alloys, made by melt spinning in He. The quasicrystals are strongly
magnetic, with sperimagnetic characteristic similar to those observed in amorphous films
of the same composition. The quasicrystal--+ crystal transition occurs at 660-900 K.

186

M. GASGNIER

4. Metalloid R-compounds

4.1. Rare-earth chalcogenide and bismuth systems

4.1.1. RAs films (R = Tb, Dy, Er, Yb and Lu)
The growth of RAs single crystals as epitaxial films on GaAs is important for
forming semiconductor/metal/semiconductor heterostructures. Numerous researches have
been done in the field of the GaAs/RAs/GaAs structures (R=Er, Yb, Lu) and then,
more recently, in the field of GaAs/ErPl_xAsx/GaAs, GaAs/Scl_xErxAs/GaAs and
InP/ErPxSbl_x/InP structures. The aim of these experiments is to obtain a nearly perfect
lattice match to GaAs. One of the main electrical characteristics of the RAs materials
is that they exhibit a metallic behaviour with resistivities of 50-90 ~tf2cm. Therefore the
high melting point of RAs materials and their excellent electric conductivity make them
promising candidates for stable contacts to GaAs, and for use as base metallization in
metal-base transistors, resonant tunnelling/transistors or quantum-well systems.
4.1.1.1. ErAs, ErxSel_xAs and ErPxAsl-x specimens. These materials seem to exhibit
the highest performances. To our knowledge the first experiments were carried out by
Waldrop (1985) in order to realize rare-earth metal (Tb, Dy, Er) Schottky barrier contacts
to GaAs. (The Schottky-barrier height is 0.85 eV). However, a direct contact cannot be
made on this basis since the R-GaAs interface is thermodynamically unstable. At this
interface GaAs dissociates and forms an interfacial RAs layer of more than 1 nm in
thickness.
Since then, Palmstrorn et al. (1988-1990), S.J. Allen et al. (1989) and Zhu et al.
(1990) have studied the ErAs and ErxSCl_xAs epitaxial layers buried in GaAs prepared
by MBE. Palmstrom et al. (1988, 1989) have characterized the single-crystal film
growth of (100) ErAs on (100) GaAs by RHEED, LEED and RBS. The structure
GaAs (50 nm)/ErAs (1-1.5 nm)/GaAs gave minimum yields of ~7%, slightly higher than
that obtained from a perfect single crystal. This is due to the presence of stacking faults
in the overgrown GaAs layer and also to a slight misorientation of the ErAs film with
respect to the substrate (attributed to a partial relaxation of the ErAs thin layer). The
resistivity of such structures decreases from 70 ~tf~cm at 293 K to 17 ~tff2cm at 1.5 K. At
low temperature, Hall measurements show that conduction is dominated by electrons with
an effective n-type mobility. However, it has been noticed that the overgrowth of ultrathin
GaAs films on ErAs is difficult due to the poor GaAs wetting of the RAs surface, which
results in three-dimensional growth.
Remark: Nearly identical results have been obtained with LuAs films (Palmstrom et
al. 1989). For a 6.5nm thin layer the resistivity decreases from 90~tf2cm at 293K to
47 ~tf2cm at 1.5 K. It is noticed that the presence of Lu203 prevents the film growth.
Zhu et al. (1990) report epitaxial microstructures characterized by HREM and RBS.
It is observed that ErAs can have an excellent epitaxy on (100) GaAs substrates. On
the other hand the overgrown GaAs layer can present twins, threading dislocations (for
thick ErAs layers), different epitaxial orientations (such as (111) GaAs on (100) ErAs)

THE INTRICATEWORLDOF RAREEARTHTHINFILMS

187

and different phase boundary (such as {122} GaAs/(100) ErAs). It is also confirmed
that this overgrown GaAs layer develops by an island growth mechanism. S.J. Allen
et al. (1989) have studied the magnetotransport and spin-disorder scattering of these
structures. They observe that ErAs epitaxial layers are a relatively low-density semimetal
ordering magnetically at low temperatures (<5K). At low temperatures (-5 K) spindisordered scattering produces a resistive anomaly at the antiferromagnetic-paramagnetic
phase boundary.
Palmstrom et al. (1990) have prepared a new kind of material: a multilayer GaAs
(50nm)/Sc0.32Er0.68As(20nm)/(100) GaAs that seems to be more attractive than the
pure GaAs/ErAs/GaAs heterostructures. They report room-temperature resistivities of
50 ~tf2cm and a lattice matching to GaAs estimated to be ~<0.1%. This low resistivity
could be due to less strain and defects in the lattice-matched sample. RHEED oscillations
during ScAs, ErAs and ScErAs growth indicate a monolayer-by-monolayer growth. Other
kinds of materials have been prepared by MBE by Le Correet al. (1989a,b, 1990),
Guivarc'h et al. (1989) and Caulet et al. (1990). Such materials as ErPxAsl-x (or Sbl-x)
have been chosen because perfectly matched layers on GaAs and InP substrates can be
obtained. Indeed, since the ErAs lattice parameter is larger than that of GaAs (+1.4%)
and the ErP one is smaller (-1%), the ErPxASl_x alloy is used to match the lattice GaAs.
Such a hypothesis has been verified by HREM which reveals an abrupt ErP0.6As0.4/GaAs
interface with a very low density of dislocations. However, dislocations and microtwins
are observed in the overgrown GaAs layer. The crystalline quality of this layer is better for
(111) GaAs substrates than for (001) GaAs substrates. Compared to ErAs films, it seems
that ErP0.6As0.4 layers do not degrade after several months in the atmosphere, whereas the
former are oxidized after one month exposure. The resistivity of the latter (~85 ~tf2cm)
is intermediate between those of ErAs (60 gf2 cm) and ErP (150 ~t~cm). Moreover, a
Schottky barrier fabricated from ErP0.6As0.4 films on n-type (001) GaAs exhibits a barrier
height of 0.88 eV. Hall measurements show that such materials are low-density semimetals
with predominantly electron conduction. One notes that Le Correet al. (1989b) have taken
out a patent.
4.1.1.2. YbAsfilms. Richter et al. (1988) have investigated thin single-crystals YbAs films
grown epitaxially on GaAs by MBE. XPS indicates the absence of oxide impurities and
the presence of only trivalent Yb. The epitaxial film is composed entirely of YbAs which
has a moderate imperfection of its crystalline quality and a lattice misfit o f ~ 8 x 10-3 with
respect to GaAs.

4.1.1.3. gb-As2S3films. Efendiev et al. (1988, 1989) have studied the formation of ternary
compounds as YbAs4S7 and Yb3As4S9 during thermal evaporation and interaction of
amorphous As2S3 and polycrystalline ytterbium. They claim that after annealing at 573 K
there is formation of C-Yb203 and of YbAs4S7 with an orthorhombic lattice.

188

M. GASGNIER

4.1.2. RSb films (R =Sm, Tm, Yb)

Only a few RSb materials have been few studied. However as reported by Le Corre
et al. (1989a,b) these materials are comparable to the RAs ones. As to pure thin
films, Dzhabua et al. (1987), Dadiani et al. (1989) and Glurdzhidze et al. (1990) have
reported the temperature dependences of the resistivity, Hall coefficient, Hall mobility
and thermoeleetromotive force at 80-300 K. The layers are deposited by vacuum thermal
evaporation from two independent sources. The values of the physical parameters agree
with those obtained for a variable valence systems. In this way TmSb, SmSb2 and YbSb2
compounds were formed.
4.1.3. RBi films (R=Gd)
Baba et al. (1989) have studied the formation of the GdxBiy compounds at various
annealing temperatures. Sputtered thin films of Gd-Bi (4:3 in atomic ratio) show that
after annealing under vacuum at 803 K the GdnBi3 phase is formed. At lower temperature
this phase is not observed, whereas at higher temperature GdsBi3 and GdBi phases grow.
Compared to the bulk, Gcl4Bi3 films have the same Curie temperature of 330K, but the
residual resistance is several times larger.
4.2. Pb(Eu, Yb)(Te,Se, S) thin films
These materials are useful for IR lasers of 2.5-6 ~tm range, which are themselves useful
for an optical source for a gas sensing system (molecules in a gas or adsorbed onto
surfaces). Moreover, a 3 ~tm laser is feasible for long-distance optical communication
using fluoride glass. These materials have the highest continuum wavelength operating
temperature ever recorded with lead salt diode lasers.
4.2.1. Lead-europium films
Partin (1983a,b, 1984, 1985, 1987, 1988) and Partin and Thrush (1984) have prepared
Pbl_xEuxTe~-ySey (x = 0.015, y = 0.020) films by MBE. This alloy is either covered with a
PbTe film (20 nm thick) to form a single-quantum-well laser, or deposited onto PbTe (100)
or (111) substrates for diode lasers. It is assumed that there is a good lattice match to PbTe.
It is shown that the laser emission energy as a function of composition increases as the
relative Eu content increases. The photoluminence properties at 4.2 K, in PbTe/PbEuTeSe
multiquantum wells have been reported by Goltros et al. (1985).
Another procedure to prepare Pbl_xEuxTe films has been used by Krost et al. (1985).
This technique, called hot-wall epitaxy, allows one to obtain single-phase material. The
layers are deposited onto BaF2 (111) or KC1 (100) substrates. Krost et al. (1985),
however, have not obtained good crystalline films with Eu content over 5%. They
report numerous optical and magneto-optical data for x=0.01 and 0.06. Ishida et al.
(1986, 1987, 1988a,b) have succeeded in preparing Pbl_xEux(Te, S) films with high
EuTe and EuS content (x ~<0.26) and PbTe/Pbl_xEuxTe superlattices. Multilayers in the

THE INTRICATEWORLDOF RARE EARTHTHIN FILMS

189

form of PbTe wells (4-25 nm)/Po.95Eu0.05Te barrier layers (14nm)/PbEuTe buffers (49 ~tm)/KC1 (100) substrates are available to obtain high-operating-temperature lasers in
the 3-6 ~tm range. Experimental absorption edges agree well with the theoretical ones
calculated assuming that the conduction band offset is equal to that of the valence band.
In the case of PbTe/Pbl_xEuxTe (2-4 ~tm)/BaF2 (111) systems the energy band gaps,
obtained from optical transmission spectra, vary nonlinearly beyond a EuTe ratio close
to 0.075 and vary linearly with tempearature between 50 and 300 K. Single-phase films
with Eg (energy gap) ~<0.83 eV and x ~<0.26 are obtained. Electrical properties (carrier
concentration and mobility) are measured at 77 K. High carrier concentration and low
resistivity are obtained up to x = 0.06 (Eg = 0.55 eV) and high resistivities are measured for
x > 0.07 and Eg > 0.60 eV In the case of PbS/Pbl-xEuxS/BaF2(111) heterojunctions, the
energy band gap can reach a maximum of 0.92 eV (x = 0.22) and Eg increases rapidly with
EuS content up to x = 0.05. Suryanarayanan and Das (1990) have used co-evaporation to
prepare thin samples deposited onto BaF2 (111) substrates with 0 < x < 0.55. It is observed
that the band gap increases more or less linearly with increasing x up to x = 0.1. The data
indicate that the Eu 2+ is stable for all values of x studied.
The above-mentioned authors have observed that both optical and electrical properties
depend on crystallographic characteristics. In this way the choice of PbTe (111) buffers,
with a lattice constant close to that of a PbEuTe film, leads to an important strain effect,
whereas BaF2 substrates, with a lattice mismatch of 4%, induce very small strain effect.
On another part, it clearly appears that there is a strong deviation from Vegard's law.
Moreover, the dependence of the lattice constants of Pbl_xEuxTe films on EuTe content
(0 < x < 0.55) shows a discrepancy between the results of Partin (1987, 1988), Ishida et
al. (1988a), Krost et al. (1985) and Suryanarayanan and Das (1990). Such a deviation is
also observed in the case of Pbl_xEuxS materials (Ishida et al. 1988b). This result has not
been understood up to date; it could be explained if the affinity of Te to Pb is different
from that to Eu.
TEM observations carried out by Salamanca-Young et al. (1988) reveal new crystallographic properties. As a function of Eu content (0.20 ~<x ~<0.75) and of substrate (BaF2)
temperature the following results have been obtained for films prepared by MBE. One
must recall that Pbl-xEuxTe is defined by a NaCl-type structure (a0 = 0.646-0.652 nm as a
function of Eu content). At low substrate temperatures (~573-673 K) and low Eu content
(x = 0.20) no ordered compounds are formed. (The multiphase formation observed by
Krost et al. (1985) at very low concentration was the result of the higher substrate
temperature (723 K) and of the hot-wall epitaxy method.) For x = 0.35 it is asserted that
an ordered compound or a superlattice appears of which the lattice constant given is
equal either to 1.125 nm or to 1.083 nm according to text and table, respectively. However
these values do not correspond to 3 x3 x3 unit cells of PbTe as built by the authors.
But the corresponding EDP exhibits weak spots located at 1/3 and 2/3 between (000)
and {422} spots. Such extra spots can also be indexed according to a hcp deriving
from the fcc one (as for Au, Ag, RH2). In this way one can argue that a hcp surface
superstructure with ah = 0.460 nm (or ao/v~) and Ch = 0.750 nm (Ch/ah= 1.633) might
be formed. For x=0.51, another fcc superstructure is formed by a 2 x 2 x 2 unit cell (the

190

M. GASGNIER

cubic parameter being well equal to 1.30nm). In this case the EDP results correspond
to this structure. For x=0.75 no disordered structures are observed. However, alloys
with 0.40 ~<x ~<0.51 exhibit a spinodal decomposition characterized by long-wavelength
periodicities (1.44 up to 2.10 nm). Such a decomposition indicates that the solid solution
for x ~ 0.50 is unstable at room temperature.
4.2.2. Lead-ytterbiumfilms
Pbl_xYbxTe films have been prepared by MBE onto BaF2 substrates by Partin (1983b,
1988). Ytterbium bonds strongly to the PbTe coverage film and it has a small diffusion
coefficient. Ytterbium introduces a donor level into PbTe near the valence-band edge
caused by the yb2+-+ Yb 3+ + e - transition. The lattice constant versus Yb content
(0 ~<x ~<0.30) does not obey Vegard's law. The optical and electrical properties of samples
prepared by co-evaporation have been reported by Das and Suryanarayanan (1989a-c).

5. Conclusion

In an earlier chapter (Gasgnier 1982) concerning R alloys and compounds as thin films,
three main wishes were emphasized. The first related to accurate characterization of the
materials. This way is now broadly opened: microstructural properties and microchemical
analyses are effective tools used by a great number of laboratories. The second one seems,
for its part, more difficult to realize. Indeed, the fabrication of "ideal" and "reproducible"
thin films always seems to be a critical problem. The presence of narrow and/or extended
defects, of amorphous or unexpected compounds or alloys, of passivating layers, etc., is
difficult to monitor. The last wish, and maybe the most important, was to draw attention
to the exchange of information between physicists and chemists, as well as to the accurate
knowledge of the complex properties of the rare earth elements. Unfortunately, as revealed
by numerous articles, these problems remain often unknown to various researchers.
Lastly, in the present state, one can readily predict the future of R elements for industrial
applications. Two of them seem to be a sure thing: supermagnets and superconductors.
Another appears to be quite likely: multilayer devices such as (R/TM)n, (R/R~)n and
heterostructures, in spite of the interfacial problems. However, as these new materials
have been synthesized in the course of the last ten years, it is not impossible to think that
new other alloys, or intermetallics, or compounds will arise in the near future. Rare earths
are multifaceted elements and can be adapted accordingly to various unexpected routes.

References

Adachi, G., K. Niki and J. Shiokawa, 1981, J. LessCommon Met. 81, 345.
Adachi, G., K. Nieki, H. Nagai and J. Shiokawa, 1982,
J. Less-CommonMet. 88, 213.

Adachi, G., H. Nagai and J. Shiokawa, 1984, J. LessCommonMet. 97, L9.

Adachi, G., H. Sakaguchi, K. Niki, N. Nagai and J.
Shiokawa, 1985a,L Less-CommonMet. 108, 107.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Adachi, G., H, Sakaguchi, K. Niki, H. Nagai and J.
Shiokawa, 1985b, Bull. Chem. Soc. Jpn. 58, 885.
In Japanese.
Adachi, G., H. Sakaguchi, T. Shimoghori and J.
Shiokawa, 1986, J. Less-Common Met. 116, L13.
Adachi, G., H. Sakaguchi, T. Shimoghori and J.
Shiokawa, 1987, J. Less-Common Met. 133, 271.
Adachi, G., H. Sakaguchi and K. Nagao, 1992,
J. Alloys & Compounds 181, 469.
Affrossman, S., 1981, Surf. Sci. 111, 1.
Aguilera-Granja, E, and J.L. Morfin-L6pez, 1990,
Solid State Commtm. 74, 155.
Ahmed-Mokhtar, N., J.R P6trakian, R. Philip, R.
Fraisse and B. Lazarides, 1982, Thin Solid Films
88, 177.
Aidun, J.B., 1986, Shock Waves in Condensed Matter
(Plenum Press, New York) p. 519.
Alameda, J.M., M.C. Contreras, J.E Fuertes, A.
Li6nard and EH. Salas, 1990, J. Magn. & Magn.
Mater. 83, 75.
Allen, C.W., D.L. Kuruzar and A.E. Miller, 1974a,
IEEE Trans, Magn. MAG-10, 716.
Allen, C.W., A.E. Miller and D.L. Kuruzar, 1974b, in:
Proc. 8th Int. Congr. on Electron Microscopy, Vol.
1 (Canberra, Australia) p. 638.
Allen, C.W., K.C. Liao and A.E. Miller, 1977, J. LessCommon Met. 52, 109.
Allen, J.W., L.I. Johansson, R.S. Bauer, I. Lindau and
S.B.M. Hagstr6m, 1978, Phys, Rev. Lett. 41, 1499.
Allen, J.W., L.I. Johansson, I. Lindau and S.B.M.
Hagstr6m, 1980, Phys. Rev. B 21, 1335.
Allen Jr, S.J., N. Tabatabaie, C.J. Palmstrom, G.W.
Hull, T. Sands, E De Rosa, H.L. Gilchrist and K.C.
Garrison, 1989, Phys. Rev. Lett. 62, 2309.
Allibert, C.H., 1989, J. Less-Common Met. 152, L1.
Alvarado, S.E, M. Campagna and W. Gudat, 1980,
J. Electron. Spectrosc. Relat. Phenom. 18, 13.
Andersen, J.N, J. Onsgaard, A. Nilsson, B. Eriksson,
E. Zdansky and N. Mhrtensson, 1987, Surf. Sci.
189/190, 399.
Andersen, J.N., I. Chorkendorff, J. Onsgaard, J.
Ghijsen, R.L. Johnson and E Grey, 1988, Phys.
Rev. B 37, 4809.
Angadi, M.A., and EV. Ashrit, 1981a, Phys. Status
Solidi a 67, K119.
Angadi, M.A., and RV. Ashrit, 1981b, J. Phys. D 14,
L125.
Angadi, M.A., and RV. Ashrit, 1981c, Vacuum 31,
161.
Angadi, M.A., and EV. Ashrit, 1983a, J. Mater. Sci.
18, 3177.

i91

Angadi, M.A., and EV. Ashrit, 1983b, Phys. Status

Solidi a 77, 685.
Annapoomi, S., G. Markandeylu and K.xZS. Rama
Rao, 1989, J. Appl. Phys. 65, 4955.
Annapoomi, S., G. Markandeylu and K.V.S. Rama
Rao, 1990, J. Appl. Phys. 68, 1394.
Antepenko, R.J., and D.M. Holtoway, 1980, Appl.
Spectrosc. 34, 206.
Anthony, T.C., J. Brag, S. Naberhuis and H. Birecki,
1986, J. Appl. Phys. 59, 213.
Aoyama, A., 1988, Jpn. Kokai Tokkyo Koho
JP 88 80 408. Patent in Japanese.
Apostolov, A.V., K.I. Grozdev, V.S. Shopov and I.V.
Velichkov, 1981, L Phys. F 11, L269.
Apostolov, A.V., K.I. Grozdev and LV. Velichkov,
1983, Czech. J. Phys. B 33, 1154.
Arakawa, T., N. Suezawa, G. Adachi and J. Shiokawa,
1988, Appl. Surf. Sci. 33/34, 286.
Ashrit, EV., and M.A. Angadi, 1980a, J. Less-Common
Met. 75 147.
Ashrit, P.V., and M.A. Angadi, 1980b, J. Phys. D 14,
L125.
Ashrit, EV., and M.A. Angadi, 1981, Phys. Status
Solidi a 63, K77.
Ashrit, P.V., and M.A. Angadi, 1982, Vacuum 32,
101.
Ashrit, RV., and M.A. Angadi, 1983, Proc. Nuel. Phys.
Solid State, Phys. Syrup. 23(C), 656.
Ashrit, EV., S.M. Shivaprasad and M.A. Angadi, 1980,
Thin Solid Films 72, L5.
Ashrit, EV, EE. Girouard and V.V. Truong, 1984,
J. Less-Common Met. 102, L15.
Aylesworth, K.D., Z.R. Zhao, D.J. Sellmyer and G.C.
Hadjipanayis, 1988, J. Appl. Phys. 64, 5742.
Aylesworth, K.D., Z.R. Zhao, D.J. Sellmyer and G.C.
Hadjipanayis, 1989, J. Magn. & Magn. Mater 82,
48.
Aylesworth, K.D., D.J. Sellmyer and G.C. Hadjipanayis, 1990, J. Appl. Phys. 67, 5725.
Aylesworth, K.D., D.J. Sellmyer and G.C. Hadjipanayis, 1991, J. Magn. & Magn. Mater. 98, 65.
Baba, K., H. Ishii, I. Yamaguchi, O. Nakamura and T.
Takeda, 1989, Mater. Res. Soc. Syrup. Proc. 151,
289.
Baberschke, K., M. Farle and M. Zomack, 1987, Appl.
Phys. A 44, 13.
Bachurin, V.I., V.E. Zhuravski, E.S. Kharlamockhin
and V.E. Yurasova, 1981, Sov. Techn. Phys. Lett. 7,
311.

192

M. GASGNIER

Bacon, EM., D.E Cowgill, C.E. Hickox, R.J. Walco,

S.R. Subia Jr and A.A. Riedel, 1984, Rev. Sci.
Instrum. 55, 42.
Baczewski, L.T., M. Piecuch, J. Durand, G. Marehal
and R Delcroix, 1989, Phys. Rev. B 40, 11237.
Baczewski, L.T., M. Piecuch, J. Durand, G. Marchal
and P. Delcroix, 1990, J. Magn. & Magn. Mater.
83, 68.
Badia, E, C. Ferrater, A. Lousa, B. Martinez, A.
Labarta and J. Tejada, 1990, J. Appl. Phys. 67,
5652.
Bauer, E., and J. Kolaczkiewicz, 1985, Phys. Status
Solidi b 131, 699.
Beaudry, B.J., and K.A. Gschneidner Jr, 1978,
Preparation and basic properties of the rare earth
metals, in: Handbook on the Physics and Chemistry
of Rare Earths, Vol. 1, eds K.A. Gschneidner Jr and
L. Eyring (North-Holland, Amsterdam) ch. 2.
Beavis, L.C., 1980, Metal Tritides Helium Emission,
microfilm DAND79-0645 (Sandia National Laboratory, Albuquerque, NM).
Beavis, L.C., and W.J. Kass, 1977, J. Vac. Sci. &
Technol. 14, 509.
Blancard, C., J.M. Esteva, R.C. Karnatak, LR
Connerade, U. Kuentgens and J. Hormes, 1989,
J. Phys. B 22, L575.
Blancard, C., J.M. Esteva, R.C. Karnatak, J.E
Connerade, U. Kuentgens and J. Hormes, 1990a, in:
2nd Eur. Conf. Progr. X-Ray Synchrotron Radiat.
Res., Nuovo Cimento, Spec. Vol., p. 113.
Blancard, C., J.M. Esteva, R.C. Karnatak, B.K. Sarpal,
J.E Connerade, U. Kuentgens and J. Hormes, 1990b,
2nd. Eur. Conf. Progr. X-Ray Synchrotron Radiat.
Res., Publ.: Nuovo Cimento, Spec. Vol., p. 117.
Boeva, O.A., K.N. Zhavoronkova, M. Smutek and S.
Cern~,, 1986, J. Less-Common Met. 118, 1.
Bohdziewicz, A., B. Cendlewska, M. Kazimierzki, J.
Szymazek and J.J. Wnuk, 1989, Phys. Status Solidi
a 114, K49.
Borgensen, P., and D.A. Lilienfeld, 1989, J. Appl. Phys.
66, 3298.
Bosca, G., J. David, L. Allay, J.M. Darmon and Y.
Victor, 1981, Rev. Phys. Appl. 16, 387. In French.
Boulesteix, C., 1982, Defects and phase transformations in rare earth sesquioxides, in: Handbook on the
Physics and Chemistry of Rare Earths, Vol. 5, eds
K.A. Gschneidner Jr and L. Eyring (North-Holland,
Amsterdam) oh. 44.
Boulesteix, C., 1984, Phys. Status Solidi a 86, 11.
Braaten, N.A., J.K. Grepstad and S. Raaen, 1989, Phys.
Rev. B 40, 7969.

Bracconi, P., and R. Laesser, 1987, Appl. Surf. Sci.

28, 204.
Bracconi, P., and R. Laesser, 1988, Z. Phys. Chem.
N.E 163, 615.
Braceoni, P., E. P6rschke, K.H. Klatt and R. Laesser,
1987, J. Vac. Sci. & Technol. A 5, 235.
Bracconi, P., E. P6rschke and R. Laesser, 1988, Appl.
Surf. Sci. 32, 392.
Brat, N.S., and Y.M. Gupta, 1986, in: Shock waves
in Condensed Matter (Plenum Press, New York) p.
513.
Bras, J., M. Fagot and J. Dejanque, 1988a, J. Microsc.
Spectrosc. Electron. 13, 439.
Bras, J., J. Dejauque and M. Fagot, 1988b, Solid State
Commun. 68, 117.
Bras, J., J. Dejauque, M. Fagot and K. Biyadi, 1990,
J. Magn. & Magn. Mater. 83, 203.
Br6chignae, C., Ph. Cahuzac, E Carlier, M. de Frutos,
A. Massov and J.-Ph. Roux, 1991, Z. Phys. D 19,
195.
Brooks, J.W., M.H. Loretto and I.R. Harris, 1976, Met.
Sci. 10, 897.
Brouder, C., M. Piecuch, G. Marchal, P. Delcroix and
A. Bruson, 1986, Proc. Eur. Mat. Res., Ed. Phys.
(Paris) p. 85.
Brouder, C., G. Krill, E. Dartyge, A. Fontaine, G.
Tourillon, P. Guilmin and G. Marchal, 1988a, Mat.
Sci. & Eng. 97, 109.
Brouder, C., G. Krill, P. Guilmin, G. Marchal, E.
Dartyge, A. Fontaine and G. Tourillon, 1988b, Phys.
Rev. B 37, 2433.
Brouder, C., G. Krill and M. Piecuch, 1988c, J. LessCommon Met. 140, 67.
Brouder, C., G. Krill, W. Felsch, P. Guilmin, G.
Marchal, E. Dartyge, A. Fontaine and G. Tourillon,
1988d, J. Less-Common Met. 145, 123.
Brousseau-Lahaye, B., C. Colliex, J. Frandon, M.
Gasgnier and P. Trebbia, 1975, Phys. Status Solidi
a 69, 257.
Bryan, S.I., J.R. Jennings, S.J. Kipling, G. Owen, R.M.
Lambert and R.M. Nix, 1988, Appl. Catal. 40, 173.
Burmistrova, O.P., M.A. Mittsev and A.M. Mukhuchev,
1983, Sov. Phys. Solid State 25, 25.
Burzo, E., and H.R. Kirchmayr, 1989, Physical
properties of R2Fe14B-based alloys, in: Handbook
on the Physics and Chemistry of Rare Earths, Vol.
12, eds K.A. Gschneidner Jr and L. Eyring (NorthHolland, Amsterdam) oh. 82.
Buschow, K.H.J., 1984, Amorphous alloys, in:
Handbook on the Physics and Chemistry of Rare

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Earths, Vol. 7, eds K.A. Gschneidner Jr and
L. Eyring (North-Holland, Amsterdam) oh. 52.
Buschow, K.H.J., and J.H.N. van Vucht, 1965, Philips
Res. Rep. 22, 233.
Buschow, K.H.J., EC.E Bouten and A.B. Miedema,
1982, Rep. Progr. Phys. 45, 937.
Buxbanm, R.E., 1982, Univ. Microfilms Int., Order
No. DA 8209248, 161pp.
Cadieu, EL, 1988, J. Vac. Sci. & Technol. A 6, 1668.
Cadieu, EJ., H. Hedge and K. Chen, 1989, IEEE Trans.
Magn. MAG-25, 3788.
Cadieu, EJ., H. Hedge and K. Chen, 1990, J. Appl.
Phys. 67, 4969.
Camley, R.E., 1987, Plays. Rev. B 35, 3608.
Camley, R.E., 1989, Phys. Rev. B 39, 12316.
Camley, R.E., and D.R. Tilley, 1988, Phys. Rev. B 37,
3413.
Carbone, C., and E. Kisker, 1987, Phys. Rev. B 36,
1280.
Carbone, C., R. Rochow, L. Braicovich, R. Jungblut,
T. Kachel, D. Tillman and E. Kisker, 1990, Phys.
Rev. B 41, 3866.
Caulet, J., A. Le Corre, B. Guesnais, G. Ropars and
A. Guivarc'h, 1990, J, Cryst. Growth 102, 309.
Cendlewska, B., and A, Rojek, 1987, Jpn. J. Appl.
Phys. Part I, Suppl. 26-3 (Part 2), 1295.
Cem), S., and Z. Pientka, 1987, Surf. Sci. 191,449.
Cem~r, S., and M. Smutek, 1990, J. Catal. 123,275.
Cerri, A., D. Mauri and M. Landolt, 1983, Phys. Rev.
B 27, 6526.
Chapman, J.W,, G.R. Morrison, D. Fort and D.W.
Jones, 1981, J. Magn. & Magn. Mater. 22, 412.
Chee, K.T., EE. Girouard and V.V. Truong, 1979, Appl.
Opt. 18, 1702.
Chen, D.Y., Y.M. Gupta and M.H. Miles, 1984, J. Appl.
Phys. 55, 5984.
Cheng, S.E, V.L. Sinha, Y. Xu, J.M. Elbicki, E.B.
Boltich, W.E. Williams, S.G, Sankar and D.E.
Laughlin, 1988, J. Magn. & Magn. Mater. 75, 330.
Cheng, S.E, B.G. Demczyk, D.E. Laughlin and W.E.
Wallace, 1990, J. Magn. & Magn. Mater. 84, 162.
Cherifi, K., P. Donovan, C. Dufour, Ph. Mangin and
G. Marchal, 1990, Phys. Status Solidi a 122, 311.
Chizhov, EE., A.N. Kostygov and VI. Petinov, 1982,
Solid State Commun. 42, 323.
Choe, G., M.E Becker and R.M. Walser, 1988, J. Appl.
Phys. 64, 6104.
Chorkendorff, I., J. Onsgaard, H. Aksela and S. Aksela,
1983, Phys. Rev. B 27, 945.
Chorkendorff, I., J. Kofoed and J. Onsgaard, 1985a,
Surf. Sci. 152/153, 749.

193

Chorkendorff, I., J. Onsgaard, J. Schmidt-May and R.

Nyholm, 1985b, Surf. Sci. 160, 587.
Ciszewski, A., and A.J. Melmed, 1984a, J. Cryst.
Growth 69, 253.
Ciszewski, A., and A.J. Melmed, 1984b, Surf. Sci.
145, L471.
Ciszewski, A., and A.J. Melmed, 1984c, J. Phys.
(Paris) Colloq. 45(Suppl. 12), C9-39.
Cochet-Muchy, D., and S. Pa'idassi, 1988, J. Plays.
(Paris) Colloq. 49(Suppl. 12), C8-661.
Coene, W., E Hakkens, T.H. Jacobs, D.B. de Mooij
and K.H.J. Buschow, 1990, J. Less-Common Met.
157, 255.
Coey, J.M.D., 1991, Permanent magnets, in: Handbook
on the Physics and Chemistry of Rare Earths, Vol.
16, eds K.A. Gschneidner Jr and L. Eyring (NorthHolland, Amsterdam) ch. 112.
Coey, J.M.D., T.R. McGuire and B. Tissier, 1981, Phys.
Rev. B 24, 1261.
Colliex, C., M. Gasgnier and E Trebbia, 1976, J. Phys.
(Paris) 37, 397.
Colliex, C., T. Manoubi, M. Gasgnier and L.M. Brown,
1985, Vol. II, Scanning Electron Microscopy, SEM
Inc., Ed. A.M.E O'Hare (Chicago, IL, USA) p.
489.
Connerade, J.P., and R.C. Kamatak, 1990, Comm. At.
Mol. Phys. 24, 1.
Comer, W.D., and EM. Saad, 1977, Physica 86/87,
1331.
Comer, W.D., EM. Saad, D.W. Jones and R.G. Jordan,
1978, Inst. Phys. Conf. Ser. 37, chp. 8, 222.
Comer, W.D., H. Bareham, R.L. Smith, EM. Saad,
B.K. Tauner, S. Farraut, D.W. Jones, B.J. Beaudry
and K.A. Gschneidner Jr, 1980, J. Magn. & Magn.
Mater. 15118, 1488.
Coulman, D., and T. Anthony, 1986, Mater. Res. Soc.
Symp. Proc. 54 (Thin Films-Interfaces Phenom.) p.
547.
Cowgill, D.E, 1977, Nucl. Instmm. Methods 145,
507.
Cowgill, D.E, 1979, IEEE Trans. Nucl. Sci. NS-26,
1697.
Cowgill, D.E, 1981, IEEE Trans. Nucl. Sci. NS-28,
1851.
Croft, M., R. Neifield, C.U. Segre, S. Raaen and R.D.
Parks, 1984, Phys. Rev. B 30, 4164.
Croft, M., E Lu, M.E. Melczer, A. Zolandz, G. Hall
and E.G. Spencer, 1985, J. Magn. & Magn. Mater.
47/48, 115.
Cromer, D.T., and A.C. Larson, 1959, Acta Crystallogr. 12, 855.

194

M. GASGNIER

Cukier, M., B. Gauth6 and C. Wehenkel, 1980, L Phys.

(Paris) 41,603.
Curzon, A.E., 1982, L Phys. D 15, L69.
Curzon, A.E., 1984, J. Less-Common Met. 98, 149.
Curzon, A.E., and H.G. Chlebeck, 1972, J. LessCommon Met. 27, 411.
Curzon, A.E., and H.G. Chlebeck, 1973, J. Phys. F 3,
1.
Curzon, A.E., and M. Gasgnier, 1985, L Electrochem.
Soc. 132, 1255.
Curzon, A.E., and O. Singh, 1979, Thin Solid Films
57, 157.
Curzon, A.E., and O. Singh, 1981a, Thin Solid Films
76, 185.
Curzon, A.E., and O. Singh, 1981b, J. Less-Common
Met. 78, 261.
Dadiani, T.O., E.V. Dokadze, Z.U. Dzhabua, T.S.
Lochoshvili, L.N. Glurdzhidze and V.V. Sanadze,
1989, Soobeshch. Akad. Nauk. Gruz. SSR 134,
365.
Das, S.K., and R. Suryanarayanan, 1989a, J. Appl.
Phys. 66, 4843.
Das, S.K., and R. Suryanarayanan, 1989b, Thin Solid
Films 175, 221.
Das, S.K., and R. Suryanarayanan, 1989c, C.R. Acad.
Sci. (Paris) 308(II), 481.
De Luca, LC., R.J. Gambino, A.E Malozemoff and L.
Berger, 1981, J. Appl. Phys. 52, 6188.
Demczyk, G.B., and S.E Cheng, 1990, J. Magn. &
Magn. Mater. 88, 376.
den Boer, M.L., C.L. Chang, S. Horn and V. Murgai,
1988, Phys. Rev. B 37, 6605.
Dexpert-Ghys, J., C. Loier, Ch. Henry La Blanchetais
and E Caro, 1975, J. Less-Common Met. 41, 105.
Dieny, B., D. Givord, J.M.B. Ndjaka and J.M. Alameda,
1990, L Appl. Phys. 67, 5677.
Domke, M., C. Laubsehat, M. Prietsch, T. Mandel, G.
Kaindl and W.D. Schneider, 1986, Phys. Rev. Lett.
56, 1287.
Dowben, EA., D. LaGraffe and M. Onellion, 1989,
J. Phys.: Condens. Matter 1, 6571.
Dudfis, J., 1991, private communication (Fak.
Electroteeh. Informatcky, Ko~ice, Slovakia).
Dudfis, L, and A. Fehrr, 1984, in: Proc. 17th Conf.
Low Temperature Physics (LT17), Vol. 1 (ALl4),
eds U. Eckern, A. Schmid, W. Weber and H. Wtihl
(North-Holland, Amsterdam) p. 159.
Dudfis, J., and A. Fehrr, 1987, Jpn. L Appl. Phys. 26
(Suppl. 26-3) 823.
Dudfis, J., A. Fehrr and V. Kavecansk~,, 1985, Phys.
Lett. 109A, 113.

Dudfis, J., A. Feh~r and M. Konc, 1986, Acta Phys.

Slov. 36, 301.
Dudfis, J., A, Fehrr and S. Jfinos, 1987a, J. LessCommon Met. 134, L9.
Dudfis, J., A. Feh~r and S. J~nos, 1987b, Acta Phys.
Polon. A 72, 263.
Dudks, L, A. Feh~r and S. Jfinos, 1987c, Acta Phys.
Hung. 62, 77.
Dudfis, J., A. Fehrr and J. Daniel-Szab6, 1987d, Acta
Phys. Sol. 37, 50.
Dud~s, J., D. Knezo, A. Feh~r and H. Ratajczak, 1990,
Physica B 165/166, 217.
Dzhabua, Z.U., T.O. Dadiani, L.N. Glurdzhidze, E.V.
Dokadze and V.V. Sanadze, 1987, Soobshch. Akad.
Nauk. Gruz. SSR 127, 605:
Efendiev, E.G., A.I. Manedov, T.M. Ilyasov and R.G.
Rustamov, 1988, Inorg. Mater. 23, 183.
Efendiev, E.G., T.M. Ilyasov, A.I. Manedov and EG.
Rustamov, 1989, Inorg. Mater. 24, 1080.
Egelhoff Jr, W.E, and G.G. Tibbetts, 1980, Phys. Rev.
Lett. 44, 482.
Egelhoff Jr, W.E, G.G. Tibbetts, M.H. Hecht and I.
Lindau, 1981, Phys. Rev. Left. 46, 1071.
Eick, H.A., C. Baenziger and L. Eyring, 1956, J. Am.
Chem. Soc. 78, 5147.
Eley, D.D., and D. Needham, 1984, Proc. Roy. Soc.
London, A 393, 257.
Ellinger, EH., and W.H. Zachariasen, 1953, L Am.
Chem. Soc. 75, 5650.
Eschenfelder, A.H., 1980a, in: Ferromagnetic Materials, ed. E.E Wohlfarth (North-Holland, Amsterdam)
ch. 6, p. 345.
Eschenfelder, A.H., 1980b, Springer Series in Solid
State Sciences 14 (Springer Verlag, Berlin, New
York) oh. 6-4, p. 210.
Esteva, J.M., R.C. Karnatak, LC. Fuggle and G.A.
Sawatzky, 1983a, Phys. Rev. Lett. 50, 910.
Esteva, J.M., R.C. Karnatak and J.E Connerade, 1983b,
J. Electron Spectrosc. Related Phenom. 31, 1.
Eyfing, L., 1979, The binary rare earth oxides,
in: Handbook on the Physics and Chemistry of
Rare Earths, Vol. 3, eds K.A. Gschneidner Jr and
L. Eyring (North-Holland, Amsterdam) oh. 27.
F/ildt, A., and H.E Myers, 1983, Solid State Commun.
48, 253.
F~ildt, A., and H.E Myers, 1984a, Phys. Rev. Lett. 52,
1315.
F~ildt, it., and H.E Myers, 1984b, Phys. Rev. B 30,
5481.
Ffildt, A., and H.E Myers, 1985, J. Magn. & Magn.
Mater. 47/48, 225.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Fgldt, A., and H.E Myers, 1986, Phys. Rev. B 34,
6675.
F~ildt, /k., D.K. Kristensson and H.E Myers, 1988,
Phys. Rev. B 37, 2682.
Fang, R.Y., D.S. Dai, X.L. Rao, Z.X. Liu, J. Lan and
H. Wan, 1988, Wuli Xuebao 37, 1065. In Chinese.
Farle, M., and K. Baberschke, 1987, Phys. Rev. Lett.
58, 511.
Fade, M., M. Zomack and K. Baberschke, 1985, Surf.
Sci. 160, 205.
Farle, M., A. Berghaus and K. Baberschke, 1989, Phys.
Rev. B 39, 4838.
Felmlee, T.L., and L. Eyring, 1968, Inorg. Chem. 7,
660.
Ferrater, C., B. Martinez, E Badia, R. Ribas, A. Lousa
and J. Tejada, 1990, J. Magn. & Magn. Mater. 83,
69.
Fidler, J., 1987, IEEE Trans. Magn. MAG-23, 2106.
Fidler, J., and P. Skalicky, 1978, Phys. Status Solidi a
50, 73.
Fidler, J., and P. Skalicky, 1981, Philos. Mag. A 44,
427.
Fidler, J., and Y. Tawara, 1988, IEEE Trans. Magn.
MAG-24, 1951.
Fidler, J., K.G. Knoch, H. Kronmgller and G.
Schneider, 1989, J. Mater. Res. 4, 806.
Fidler, J., C. Groiss and M. Tokunaga, 1990, IEEE
Trans. Magn. MAG-26, 1948.
Forester, D.W, P. Lubitz and C. Vittoria, 1983, US
Patent 386 465.
Forester, D.W., E Lubitz, J.H. Schelleng and C.
Vittoria, 1984, J. Non-Cryst. Solids 61162, 685.
Franceschi, E., 1982, J. Less-Common Met. 87, 249.
Franciosi, A., A. Raisanen, A. Wall, S. Chang, E
Philip, N. Trouiller and D.J. Peterman, 1988, Appl.
Phys. Lett. 52, 1490.
Frandon, J., B. Brousseau and E Pradal, 1980, Phys.
Status Solidi b 98, 379.
Frankenthal, R.P., R.B. van Dover and D.J. Siconolfi,
1986, J. Vac. Sci. & Technol. A 4, 1631.
Frausto, RR., W.B. Johnson and C.E. Wickersham,
1986, Proc. SPIE, Inst. Soc. Opt. Eng. 695 (Opt.
Mass Data Storage 2) p. 56.
Freitas, P.R, T.S. Plaskett, J.M. Moreira and VS.
Amaral, 1988, J. Appl. Phys. 64, 5453.
Frigerio, J.M., M. Martin, J. Rivory and Tran Mirth
Duc, 1985, in: Rapidly Quenched Metals, Vol. 1,
ed. H. Warlimont (North-Holland, Amsterdam) p.
1003.
Fujimori, H., Y. Kamiguchi and Y. Hayakawa, 1990,
J. Appl. Phys. 67, 5716.

195

Gambino, R.J., and T.R. McGuire, 1984, IEEE Trans.

Magn. MAG-20, 1341.
Gambino, R.J., T.R. McGuire and K. Fukamichi, 1981,
J. Appl. Phys. 52, 2190.
Gasgnier, M., 1980, Phys. Status Solidi a 57, 11.
Gasgnier, M., 1982, Rare earth alloys and compounds
as thin films, in: Handbook on the Physics
and Chemistry of Rare Earths, Vol. 5, eds
K.A. Gschneidner Jr and L. Eyring (North-Holland,
Amsterdam) ch. 41.
Gasgnier, M., 1989, Phys. Status Solidi a 114, 11.
Gasgnier, M,, 1990, Mater. Lett. 9, 161.
Gasgnier, M,, 1991, J. Mater. Sci. 26, 1989.
Gasgnier, M., and L.M. Brown, 1983, unpublished
data (Univ. Paris-Sud, France).
Gasgnier, M., Ch. Henry La Blanchetais and P Caro,
1976, Thin Solid Films 31, 283.
Gasgnier, M., L. N6vot, P Baillif and J. Bardolle,
1983, Phys. Status Solidi a 79, 531.
Gasgnier, M., G. Schiffmacher, L. Albert, E Cam,
H. Dexpert, J.M. Esteva, C. Blancard and R.C.
Karnatak, 1989, J. Less-Common Met. 156, 59.
Gerken, E, J. Barth, K.L.I. Kobayashi and C. Kunz,
1980, Solid State Commun. 35, 179.
Gerken, E, J. Barth and C. Kunz, 1981, Phys. Rev.
Lett. 47, 993.
Gerken, E, J. Barth, R. Kammerer, L.I. Johansson and
A.S. Flodstr6m, 1982, Surf. Sci. 17, 468.
Gerken, E, A.S. Flodstr6m, J. Barth, L.I. Johansson
and C. Kunz, 1985, Physica Scripta 32, 43.
Gimzewski, J.K., D.J. Fabian, L.M. Watson and S.A.
Hrossman, 1977, J. Phys. E 7, L305.
Gimzewski, J.K., S. Affrossman, M.T. Gibson, L.M.
Watson and D.J. Fabian, 1979, Surf. Sci. 80, 298.
Glurdzhidze, L.N., M.Yu. Stamateli, T.S. Lochoshvili,
E.V. Dokadze, ZM. Dzhabua and T.O. Dadiani,
1990, Soobshch. Akad. Nauk Gruz SSR 138, 533.
Goltros, W, J. Nakahara, A.V. Nurmiko and D.L.
Partin, 1985, Appl Phys. Lett. 46, 1173.
Gonchar, EM., V.K. Medveded, T.R Smereka, Ya.B.
Lozovyi and G.V. Babkin, 1987, Soy. Phys. Sol.
State 29, 1629.
Gonchar, VV., O.K. Kanash and A.G. Fedorus, 1985,
Phys. Chem. Mech. Surf. 2, 1621.
GondS, Y., and Y. Suezawa, 1986, J. Magn. & Magn.
Mater. 54/57, 283.
Gond6, Y., and Y. Suezawa, 1987, IEEE Trans. Magn.
MAG-23, 3317.
Gond6, Y., Y. Suezawa, T. Nakao and K. Kakuno,
1983, J. Magn. & Magn. Mater. 31/34, 329.

196

M. GASGNIER

Gondr, Y., Y. Suezawa and R. Tsuchiya, 1984, IEEE

Trans. Magn. MAG-20, 1314.
Gorodetskii, D.A., and A.V. Martynyuk, 1989, Phys.
Chem. Mechn. Surf. 5, 503.
Gorodetskii, D.A., and S.A. Shevlyakov, 1980, Sov.
Phys. Crystallogr. 25, 377.
Grady, D.E., and M.J. Ginsberg, 1977, J. Appl. Phys.
48, 2179.
Grant, WA., 1978, J. Vac. Sci. & Technol. 15, 1644.
Grieb, B., K.G. Knoch, E.-Th. Henig and G. Petzow,
1989, J. Magn. & Magn. Mater. 80, 75.
Grrssinger, R., X.C. Kou, R. Krewenka, H.R.
Kirchmayr and M. Tokunaga, 1990, IEEE Trans.
Magn. MAG-26, 1954.
Grfitter, E, E. Meyer, H. Heinzelmarm, L. Rosenthaler,
H.-R. Hidber and H.-J. Giintherodt, 1988, J. Vac.
Sci. & Technol. A 6, 279.
Grfitter, P., Th. Jung, H. Heinzelmann, A. Wadas, E.
Meyer, H.-R. Hidber and H.-J. Gtintherodt, 1990a,
J. Appl. Phys. 67, 1437.
Grfitter, E, A. Wadas, E. Meyer, H. Heinzelmann, H.R. Hidber and H.-J. Gfintherodt, 1990b, J. Vac. Sci.
& Technol. A 8, 406.
Gschneidner Jr, K.A., 1971, J. Less-Common Met. 25,
405.
Gsehneidner Jr, K.A., 1980, in: Science and Technology of Rare Earth Materials, eds E.C. Subbarao and
W.E. Wallace (Academic Press, New York) p. 25.
Gschneidner Jr, K.A., 1985, J. Less-Common Met.
110, 1.
Gschneidner Jr, K.A., 1992, J. Alloys & Compounds
180, 1.
Gschneidner Jr, K.A., and A.H. Daane, 1988,
Physical Metallurgy, in: Handbook on the Physics
and Chemistry of Rare Earths, Vol. 11, eds
K.A. Gschneidner Jr and L. Eyring (North-Holland,
Amsterdam) ch. 78.
Gschneidner Jr, K.A., B.J. Beaudry and J. Capellen,
1990, Metals Handbook, 10th Ed., Vol. 2, (ASM
International, Metals Park, OH) p. 720.
Gu, B.X., H. Homburg, S. Methfessel and H.R. Zhai,
1990, Phys. Status Solidi a 120, 159.
Guilmin, E, C. Brouder, G. Krill, W Felseh, G.
Marchal, E. Dartyge, A. Fontaine and G. Tourillon,
1988, J. Phys. (Paris) Colloq. 48(Suppl. 12), C81757.
Guivarc'h, A., J. Caulet and A. Le Corre, 1989,
Electron. Lett. 25, 1051.
Gupta, S.C., and Y.M. Gupta, 1985, J. Appl. Phys. 57,
2464.
Gupta, Y.M., 1983a, J. Appl. Phys. 54, 6094.

Gupta, Y.M., 1983b, J. Appl. Phys. 54, 6256.

Hansen, P., and M. Hartmann, 1986, J. Appl. Phys.
59, 859.
Harada, T., M. Ikeda, S. Yamashita and K. Hara, 1988,
Jpn. Kokai Tokkyo Koho JP 8865611. Patent in
Japanese.
Harris, J.M., E.P. Boespflug, D.M. Holloway and J.L.
Provo, 1984, Sandia National Lab. Rpt. SAN831054, UC-25.
Hartmann, M., and T.R. McGuire, 1983, Phys. Rev.
Lett. 51, 1194.
Hashimoto, S., Y. Ochiai, M. Kaneko, K. Watanabe
and K. Aso, 1987, IEEE Trans. Magn. MAG-23,
2278.
Hauser, J.J., 1985, Solid State Commtm. 55, 163.
Haussler, P., and E Battman, 1980, Z. Phys. B 38,
43.
Hay, C.M., J.R. Jennings, R.M. Lambert, R.M. Nix,
G. Owen and T. Rayment, 1988, Appl. Catal. 37,
291.
Hecht, M.H., A.J. Viescas, I. Lindau, J.W. Allen and
L.I. Johansson, 1984, J. Electron. Spectrosc. Relat.
Phenom. 34, 343.
Heinzelmann, H., E Griitter, E. Meyer, L. Rosenthaler,
H.-R. Hidber and H.-J. Giintherodt, 1987, Surf. Sci.
189/190, 29.
Heitmann, H., A.M.J. Spruijt, E Willich and H.
Wilting, 1987, J. Appl. Phys. 61, 3343.
Hellman, E, R.B. van Dover and E.M. Gyorgy, 1987,
Appl. Phys. Lett. 50, 296.
Hirst, J.R., M.L.H. Wise, D. Fort, J.EG. Farr and I.R.
Harris, 1976, J. Less-Common Met. 49, 193.
Hole, J., S. Besenicar and D. Kolar, 1990, J. Mater.
Sci. 25, 215.
Holloway, D.M., R.J. Antepenko and WE. Swartz Jr,
1978, Appl. Spectrosc. 32, 106.
Homburg, H., Th. Sinnemarm, S. Methfessel, M.
Rosenberg and B.X. Gu, 1990, J. Magn. & Magn.
Mater. 83, 231.
Homma, H., K.Y. Yang and I.K. Schuller, 1987a, Phys.
Rev. B 36, 9435.
Homma, H., K.Y. Yang and I.K. Schuller, 1987b,
Interfaces, Superlattices, Thin Films, eds J.D. Dow
and I.K. Schuller (Mater. Sci. Soc. Pittsburg R.A.)
p. 557.

Honda, S., M. Ohkoshi and T. Kusuda, 1986, IEEE

Trans. Magn. MAG-22, 1221.
Honda, S., N. Nawate, M. Yashiyama and T. Kusuda,
1987, IEEE Trans. Mag. MAG-23, 2952.
Horikoshi, H., S. Takeda and Y. Komura, 1985, Jpn.
J. Appl. Phys. 24, L210.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Hoshi, Y., M. Naoe and S. Yamanaka, 1982, J. Appl.
Phys. 53, 2344.
Hosoda, N., I. Nishiguchi, Y. Ohtani, H. Uchida and
T. Takahashi, 1988, Kidoriu 12, 36. In Japanese.
Hosoito, N., K. Yoden and T. Shijo, 1988a, Hyperfine
Interact. 41,583.
Hosoito, N., K. Yoden, K. Mibu and T. Shinjo, 1988b,
J. Phys. (Paris) Colloq. 48(Suppl. 12), C8-1777.
Hosoito, N., K. Yoden, K. Mibu, T. Shinjo and Y.
Endoh, 1989, L Phys. Soc. Jpn. 58, 1775.
Huang, C.Y., C.E. Olsen, G. Kozlowski, H. Matsumoto, H. Umezawa, E Mancini, M.B. Maple,
M.C. Hamaker, M.S. Torikachvili, J.P. Whitehead
and EE. Wang, 1985, J. Appl. Phys. 57, 3104.
Huang, G.X., Q.Z. Lu, L.H. Wu, S.Y. Shi, W.M. Gao,
S.E Yu, H.C. Hua, G.A. Wang and C.H. Zheng,
1990, Phys. Scripta 41,372.
Huang, Y.S., and V. Murgai, 1989, Solid State
Commun. 69, 873.
Hubberstey, P., RE Adams and R.J. Pulham, 1976,
Radial Effects Triitittm Technol. Fusion React.,
Proc. Int. Conf. 1975, eds J.S. Watson and EW.
Wiffen (NTIS, Springfield, Va, USA), Ill-270.
Hughes, D.T., J. Evans and I.R. Harris, 1980a, J. LessCommon Met. 74, 255.
Hughes, D.T., L Evans and I.R. Harris, 1980b, J. LessCommon Met. 76, 119.
Hussain, A.A.A., 1980, J. Less-Common Met. 72, 71.
Hussain, A.A.A., 1981, Thin Solid Films 77, 331.
Hussain, A.A.A., and T.S. A1-Bassam, 1979, Thin
Solid Films 59, 77.
Idczak, E., and K. Zukowska, 1981, Thin Solid Films
75, 139.
Inoue, A., M. Ikematsu, T. Miyama and T. Fujii, 1985,
IEEE Transl. J. Magn. Jpn. TJMJ-1-3, 363.
Ishida, A., S. Matsuura, H. Fujiyasu, H. Ebe and K.
Shimohara, 1986, Superlattiees & Microstructttres
2, 575.
Ishida, A., S. Matsuura, H. Mizuno and H. Fujiyasu,
1987, Appl. Phys. Lett. 51, 478.
Ishida, A , S. Matsuura, H. Mizuno, Y. Sase and H.
Fujiyasu, 1988a, J. AppI. Phys. 63, 4572.
Ishida, A., N. Nakahara, T. Okamura, Y. Sase and H.
Fujiyasu, 1988b, Appl. Phys. Lett. 53, 274.
Iwamura, E., K. Nagahama, T. Suzuki and T. Umeda,
1990, J. Jpn. Inst. Met. 54, 111. In Japanese.
Jaffey, D.M., A. Gellman and R.M. Lambert, 1989,
Surf. Sci. 214, 407.
Jfinos, S., A. Feh6r and J. Dudfis, 1987, Czch. J. Phys.
B 37, 5.

!97

Jaswal, S.S., M.A. Langell, Y.G. Ren, M.A. Engelhardt

and D.J. Sellmyer, 1988, J. Appl. Phys. 64, 5577.
Jaswal, S.S., WB. Yelon, G.C. Hadjipanayis, Y.Z.
Wang and D.J. Sellmyer, i991, Phys. Rev. Lett.
67, 644.
Jennings, J.R., R.M. Lambert, R.M. Nix, G. Owen and
D.G. Parker, 1989, Appl. Catal. 50, 157.
Johansson, B., 1979, Phys. Rev. B 19, 6615.
J'ohansson, LJ., J.W. Allen, T. Gustafsson, I. Lindau
and S.B.M. Hagstr6m, 1978, Solid State Commun.
28, 53.
Johansson, L.I., J.W. Allen and I. Lindau, 1981, Phys.
Lett. 86 A, 442.
Johansson, L.I., A. Flodstr6m, S.E. H6mstr6m, B.
Johansson, J. Barth and E Gerken, 1982a, Solid
State Commun. 41,427.
Johansson, L.I., A. Flodstr6m, S.E. H6mstr6m, B.
Johansson, J. Barth and E Gerken, 1982b0 Surf.
Sci. 117, 475.
Johnson, D.A., 1969, J. Chem. Soc. A 1525 and 2578.
Johnson, D.A., 1977, Adv. Inorg. Chem. Radioelem.
20, 1.
Johnson, R.W., C.C. Ahn and E.R. Rather, 1989a,
Appl. Phys. Lett. 54, 795.
Johnson, R.W., C.C. Ahn and E.R. RaUler, 1989b,
Phys. Rev. B 40, 8139.
Jordan, R.G., and O. Loebich Jr, 1975, J. LessCommon Met. 39, 55.
Kaindl, G., B. Reihl, D.E. Eastman, R.A. Pollack,
N. M~rtensson, B. Barbara, T. Penncey and T.S.
Plaskett, 1982, Solid State Commun. 41, 157.
Kaindl, G., W.D. Schneider, C. Laubschat, B. Reihl
and N. MS.rtensson, 1983, Surf. Sci. 126, 105.
Kajiura, M., T. Morishita, Y. Togami and K. Tsushima,
1987, IEEE Transl. J. Magn. Jpn. TJMJ-2-2, 171.
Kamiguchi, Y., Y. Hayakawa and N. Fujimori, 1989,
Appl. Phys. Lett. 55, 1918.
Kammerer, R., J. Barth, E Gerken, A. Flodstr6m and
L.I. Johansson, 1982, Solid State Commun. 41,
435.
Kamprath, N., L. Wickramasekara, H. Hedge, N.C.
Liu, J.K.D. Jayanetti and EJ. Cadieu, 1988a, J. Appl.
Phys. 63, 3696.
Kamprath, N., N.C. Liu, H. Hedge and EJ. Cadieu,
1988b, J. Appl. Phys. 64, 5720.
Kamprath, N., X.R. Qian, H. Hedge and EJ. Cadieu,
1990, J. Appl. Phys. 67, 4948.
Kandasamy, K., and N.A. Surplice, 1981, Phys.
Transition Metals, Inst. Phys. Conf. Ser. 55, p. 587.
Kandasamy, K., and N.A. Surplice, 1982, J. Phys. C
14, L61.

198

M. GASGNIER

Kandasamy, K., and N.A. Surplice, 1985, J. Phys. D

18, 1377.
Kanski, J., and P.O. Nilsson, 1979, Phys. Rev. Lett.
43, 1185.
Kanski, J., and G. Wendin, 1981, Phys. Rev. B 24,
4977.
Kass, W.J., 1977, J. Vac. Sci. & Technol. 14, 518.
Katayama, T., H. Awano and Y. Nishihara, 1987, IEEE
Transl. J. Magn. Jpn. TJMJ-2-2, 188.
Kawahata, T., N. Ninomiya, Y. Ohtani, H. Uchida and
T. Takahashi, 1990, Kidoriu 16, 104. in Japanese.
Knoch, K.G., J. Fidler, E. Bischoff and E.-Th. Henig,
1989a, Mater. Lett. 8, 301.
Knoch, K.G., G. Schneider, J. Fidler, E.-Th. Henig and
H. Kronmfiller, 1989b, IEEE Trans. Magn. MAG25, 3426.
Knoch, K.G., B. Grieb, E.-Th. Henig, K. Kronmiiller
and G. Petzow, 1990a, IEEE Trans. Magn. MAG26, 1951.
Knoch, K.G., E.-Th. Henig and J. Fidler, 1990b,
J. Magn. & Magn. Mater. 83, 209.
Koestler, C., R. Ramesh, C.J. Ether, G. Thomas and
J. Wecker, 1989, Acta Metall. 37, 1945.
Kojima, T., A. Denda and H. Uchida, 1989, Kidorui
14, 96. In Japanese.
Kolaczldewicz, J., and E. Bauer, 1986, Surf. Sci. 175,
487.
Komura, Y., 1989, Phase transitions 16/17, 495.
Komura, Y., S. Takeda and Y. Kitano, 1981, Sci. Rep.
RITU A 29(Suppl. 1), 13. In English.
Kosak, M.M., R.S. Pan6i~in and Z.V. Stasfik, 1987,
Ukr. Fiz. Zh. 32, 751. In Russian.
Kostygov, A.N., V.E Degtyareva, Yu.G. Morozov, P.E.
Chizhov and V.I. Petinov, 1977, Sov. Phys. Solid
State 19, 1553.
Koyama, M., R. Ohyama, T. Kovanagi, Y. Watanabe
and K. Matsubara, 1987, Nippon Oyo. Giki
Gakkaishi 11, 173. In Japanese.
Kozlovskii, L.V., 1989, Latv. PSR. Zinat. Akad Vestes.
Fiz. Teh. Zinat. Ser. 6, 65. In Russian.
Kozlovskii, L.V., E.M. Reikhrudel and G.V. Smirnitskaya, 1981, Sov. Phys. Techn. Phys. 26, 991.
Krishnan, R., M. Porte, M. Tessier, J.P. Vitton and Y.
Le Corre, 1988, J. Less-Common Met. 145, 613.
Krizek, J., and K.N.R. Taylor, 1975, J. Phys. F 5, 774.
Krost, A., B. Harbecke, R. Fraymonville, H. Schlegel,
E.J. Fantner, K.E. Ambroth and G. Bauer, 1985,
J. Phys. C 18, 2119.
Kuboth, M., H.E Braun, L. Schellenberg and J. Muller,
1989, Physica C 162/164, 407.

Kumar, R., M.H. Mintz and J.W. Rabalais, 1984a, Surf.

Sci. 147, 15.
Kumar, R., M.H. Mintz and J.W. Rabalais, 1984b,
Surf. Sci. 147, 37.
Kunze, U., and W. Kowalsky, 1988, Appl. Phys. Lett.
53, 367.
Kuwano, N., T. Shiwaku, Y. Tomokiyo and T. Eguchi,
1981, Jpn. J. Appl. Phys. 20, 1603.
Kuwano, N., S. Higo, K. Yamamoto, K. Oki and T.
Eguchi, 1985, Jpn. J. Appl. Phys. 24, L663.
Kuzmenko, V.M., and V.I. Melnikov, 1982, Sov. Phys.
JETP 55, 474.
Kuzmenko, V.M., and VI. Melnikov, 1988, Sov. J. Low
Temp. Phys. 14, 533.
Kuzmenko, V.M., B.G. Lazarev, V.I. Melnikov and
A.I. Sudovtsov, 1972, Ukr. Fiz. Zh. 17, 1923. In
Russian.
Kuzmenko, V.M., B.G. Lazarev, V.I. Melnikov and A.I.
Sudovtsov, 1976, Ukr. Fiz. Zh. 21,883. In Russian.
Kuzmenko, V,M., V.I. Melnikov and A.I. Sudovtsov,
1980, Phys. Met. Metallogr. 54(2), 165.
Kuzmenko, V.M., V.I. Melnikov and VA. Rakhubovskii,
1984a, Sov. Phys. JETP 59, 612.
Kuzmenko, V.M., A.N. Vladychkin, V.I. Melnikov and
A.I. Sudovtsov, 1984b, Sov. Phys. JETP 59, 102.
Kuzmenko, V.M., B.G. Lazarev and V.I. Melnikov,
1986a, Soy. J. Low Temp. Phys. 12, 431.
Kuzmenko, V.M., A.N. Vladychkin and A.I. Sudovtsov, 1986b, Sov. J. Low Temp. Phys. 12, 714.
Kwo, J., M. Hong and S. Nakahara, 1986, Appl. Phys.
Lett. 49, 319.
LaGraffe, D., EA. Dowben and M. Onellion, 1985,
Surf. Sci. 160, 587.
LaGraffe, D., EA. Dowben and M. Onellion, 1988,
Phys. Rev. B 38, 10357.
LaGraffe, D., P.A. Dowben and M. Onellion, 1989a,
Phys. Rev. B 40, 970.
LaGraffe, D., P.A. Dowben and M. Onellion, 1989b,
Phys. Rev. B 40, 3348.
LaGraffe, D., P.A. Dowben and M. Onellion, 1989c, in:
Growth Characterization and Properties of Ultrathin
Magnetic Films and Multilayers, Mater. Res. Soc.
Symp. Proe. 151, 71.
LaGraffe, D., PA. Dowben and M. Onellion, 1990a,
Phys. Lett. A 147, 240.
LaGraffe, D., P.A. Dowben and M. Onellion, 1990b,
J. Vac. Sci. & Technol. A 8, 2738.
Landolt, M., R. Allenspach and M. Taborelli, 1986,
Surf. Sci. 178, 311.
Lang, J.K., Y. Baer and P.A. Cox, 1981, J. Phys. F 11,
121.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Larsen, J.W., M.L. Fuller and B.R. Livesay, 1981,
Alternative Energy Sources 2(8), 3533.
Le Corre, A., J. Caulet and A. Guivarc'h, 1989a, Appl.
Phys. Lett. 55, 2298.
Le Corre, A., J. Caulet and A. Guivarc'h, 1989b,
French Patent No. 89 04899.
Le Corre, A., B. Guesnais, A. Guivarc'h, D.
Lecrosnier, J. Caulet, M. Minier and G. Ropars,
1990, J. Cryst. Growth 105, 234.
LePage, J.G., and R.E. Camley, 1990, Phys. Rev. Lett.
65, 1152.
Li, D., C.W. Hutchings, P.A. Dowben, C. Hwang, R.T.
Wu, M. Onellion, B. Andrews and J.L. Erskine,
1991, J. Magn. & Magn. Mater. 99, 85.
Li, H., S. Jiang, M. Qiu, Y. Guo and R.C. O'Handley,
1988, Jinshu Xuebao 24, A131. In Chinese.
Li, L., ELi, D. Yang, L Tian and Z. Lin, 1990, Wuli
Xuebao 39, 788. In Chinese.
Li, Z., H. Hahn and R.W. Siegel, 1988, Mat. Lett. 6,
342.
Lii, Z.Y., H. Okuno, T. Numata and Y. Sakurai, 1983,
Nippon Oyo Jiki Gakkaishi 7, 47.
Linetski, Ya.L., and I.R Salo, 1989, Phys. Met. Metall.
68(0, 115.
Linetsld, Ya.L., R.M. Raigorodskii, E.L. Khatsernova
and V.Yu. Tsvetkov, 1990, Izv. Vyssh. Uchebn.
Zaved. Chem. Mater. 7, 64.
Liu, N.C., N. Kamprath, L. Wickramasekara and EJ.
Cadieu, 1988, J. Appl. Phys. 63, 3589.
Loboda, V.B., and I.E. Protsenko, 1981a, Cryst. Res.
Technol. 16, 357.
Loboda, V.B., and I.E. Protsenko, 1981b, Cryst. Res.
Technol. 16, 489.
Loboda, V.B., I.E. Protsenko and A.V. Yaremenko,
1980, Kristall und Teknik 15, 43.
Loboda, V.B., I.E. Protsenko and V.G. Sharnonja,
1982, Ukr. Fiz. Zh. 27, 1343. In Russian.
Loebich Jr, O., and E. Raub, 1973a, J. Less-Common
Met. 30, 47.
Loebich Jr, O., and E. Raub, 1973b, J. Less-Common
Met. 31, 111.
Lowe, W.R, E.M. Gyorgy, D.B. McWhan, L.H. Greene,
W.L. Feldman and J.M. Rowell, 1985, J, Appl. Phys.
58, 1615.
Lozovyi, Ya.B., V.K. Medveded, T.P. Smereka and G.V.
Babkin, 1982, Sov. Phys. Solid State 24, 1213.
Lozovyi, Ya.B., V.K. Medveded, T.P. Smereka, G.V.
Babldn and B.M. Palyukh, 1984, Soy. Phys. Solid
State 26, 738.
Lu, E, M. Croft, A. Leiberich and G. Hall, I985,
in: New frontiers in Rare Earth Science and

199

Applications, eds Xu Guangxian and Xiao Jimei

(Science Press, Beijing) p. 374.
Lu, E, M. Croft and E.G. Spencer, 1986, Phys. Rev.
B 33, 5950.
Ltibcke, M., B. Sonntag, W. Niemaml and E Rabe,
1986, Phys. Rev. B 34, 5184.
Ma, E., B.X. Liu, X. Chen and H.D. Li, 1987, Thin
Solid Films 147, 49.
Mahmoud, S., and Z.S. el Mandouh, 1987, J. Mater.
Sci. 22, 651.
Maienschein, J.L., 1978, Nuct. Technol. 38, 387.
Maier-Komor, P., 1989, Nucl. Instr. Meth. Phys. Rev.
A 282, 133.
Malinowski, M.E., 1981, J. Vac. Sci. & Technol. 19,
120.
Malterre, D., A. Siari, J. Durand, G. Krilt and
G. Marchal, 1986, J. Phys. (Paris) Colloq.
47(Suppl. 12), C8-991.
Mancini, E, and C. Noce, 1986, Nuovo Cimento D 7,
1.
Mancini, E, and C. Noce, 1987, Physica 145 B, 342.
Manoubi, T., C. Colliex and P. Rez, 1990, J. Electron.
Spectrosc. Relat. Phenom. 50, 1.
Marinero, E.E., G.S. Sprokel and H. Notarys, 1988,
Mater. Res. Soc. Symp. Proc. (Multilayers) 245,
1033.
Marshall, A.N., and D.C. Dobbertin, 1986, Ultramicroscopy 19, 69.
Martin, M., Tran Minh Duc, J. Rivory and J.M.
Frigerio, 1985, Solid State Commun. 55, 683.
Martinez, B., M.A. Moreu and J. Tejada, 1988a,
J. Phys. (Paris) Colloq. 49(Suppl. 12), C8-1807.
Martinez, B., M.A. Moreu, A. Ruiz, A. Labarta, X.
Obradors, E Briones and J. Tejada, 1988b, IEEE
Trans. Magn. MAG-24, 1694.
Mason, M.G., S.T. Lee, G. Apai, R.E Davis, D.A.
Shirley, A. Franciosi and J.H. Weaver, I981, Phys.
Rev. Lett. 47, 730.
Materlik, G., J.E. Miiller and J.W. Wilkins, 1983a,
Phys. Rev. Lett. 50, 267.
Materlik, G., B. Sonntag and M. Taush, 1983b, Phys.
Rev. Lett. 51, 1300.
Mathevet, J.R, A. Traverse, J. Chaumont and S.
Megtert, 1990, J. Mater. Res. 5, 1970.
Matthew, J.A.D., G. Strasser and EP. Netzer, 1983,
Phys. Rev. B 27, 5839.
Mauri, D., and M. LandoR, 1981, Phys. Rev. Lett. 47,
1322.
McGuiness, RJ., I.E. Harris, E. Rozendaal, J. Onnerod
and M. Ward, 1986, J. Mater. Sci. 21, 4107.

200

M. GASGNIER

McGuiness, P.J., E. Devlin, I.R. Harris, E. Rozendaal

and J. Ormerod, 1989a, J. Mater. Sci. 24, 2541.
McGuiness, P.J., I.R. Harris, U.D. Scholz and H. Nagel,
1989b, Z. Phys. Chem. N.E 163/165, 587.
McGuiness, P.J., X.J. Zhang, X.J. Yin and I.R. Harris,
1990a, J. Less-Common Met. 158, 359.
McGuiness, P.L, X.J. Zhang, H. Forsyth and I.R.
Harris, 1990b, J. Less-Common Met. 162, 379.
McMinn, R., A. Aknetey and P.E. Donovan, 1990, in:
Proc. Int. Phys. Electron. Microsc. Analysis Group.
Roy. Microsc. Soc. Conf. (Institute of Physics,
Bristol, UK) Vol. 1, p. 183.
Mehrhoff, T.K., 1980, GEPP-TIS-482 (Microfilm),
24pp.
Melmed, A.J., V. Maurice, O. Franck and J.H. Block,
1984, J. Phys. (Paris) Colloq. 49(Suppl. 12), C947.
Melmed, A.J., V. Maurice, O. Franck and J.H. Block,
1987, J. Cryst. Growth 84, 123.
Melton, K.N., and H. Nagel, 1977, J. Appl. Phys. 48,
2608.
Melton, K.N., and R.S. Perkins, 1976, J. Appl. Phys.
47, 2671.
Mibu, K., N. Hosoito and T. Shinjo, 1989, J. Phys.
Soc. Jpn. 58, 2916.
Mibu, K., N. Hosoito and T. Shinjo, 1990, Hyperfine
Interact. 54, 831.
Miller, R.E, M. Syms and L.S. Julien, 1979, J. Phys.
D.: Appl. Phys. 12, 1985.
Miller, S.A., 1988, Shock Waves in Condensed Matter
(North-Holland, Amsterdam) p. 605.
Mitchell, D.J., and R.C. Patrick, 1981, J. Vac. Sci. &
Technol. 19, 236.
Mitchell, D.J., and J.L. Provo, 1985, L Appl. Phys. 57,
1855.
Mittsev, M.A., and A.M. Mukhuchev, 1982, Soy. Phys.
Solid State 24, 1284.
Mittsev, M.A., and A.M. Mthkhuchev, 1983, Sov.
Techn. Phys. Lett. 9, 154.
Miyahara, T., H. Ishii, T. Hanyu, H. Ohkuma and S.
Yamaguchi, 1982, J. Phys. Soc. Jpn. 51, 1834.
Miyamoto, K., and A. Tazaki, 1985, Jpn. Kokai
Tokkyo Koho JP 85 602 526. Patent in Japanese.
Miyazaki, T., M. Tukahashi, T. Shimamori and T.
Wakiyama, 1989, in: Proc. MRS Intern. Meet. Adv.
Mater. Multilayers, Vol. 10 (Materials Research
Society, Pittsburgh, PA) p. 379.
Moth, Th., B. Lippold and A. Handstein, 1986, Phys.
Status Solidi a 98, K161.
Moreau, J.M., and E. Parthr, 1973, J. Less-Common
Met. 32, 91.

Morishita, T., Y. Togami and K. Tsushima, 1985,

J. Phys. Soc. Jpn. 54, 37.
Morishita, T., Y. Togami and K. Tsushima, 1986,
J. Magn. & Magn. Mater. 54/57, 789.
Morishita, T., Y. Togami and K. Tsushima, 1987, IEEE
Transl. J. Magn. Jpn. TJMJ-2-2, 169.
Morozov, Yu.G., A.N. Kostygov, VJ. Petinov and EE.
Chizhov, 1975, Phys. Status Solidi a 32, K119.
Morozov, Yu.G., A.N. Kostygov, A.E. Petrov, P.E.
Chizhov and V.I. Petinov, 1976, Sov. Phys. Solid
State 18, 804.
Moser, H.R., B. Delley, W.D. Schneider and Y. Baer,
1984, Phys. Rev. B 29, 2947.
Mfiller, J., and N.A. Surplice, 1977, J. Phys. D 10,
213.
Murgai, V., Y.S. Huang, M.L. den Boer and S. Horn,
1988, Solid State Commun. 66, 329.
Nakajima, K., and T. Sato, 1986, Jpn. Kokai Tokkyo
Koho JP 86 117 749. Patent in Japanese.
Nakamura, O., K. Baba, H. Ishii and T. Takeda, 1988,
J. Appl. Phys. 64, 3614.
Nakamura, T., S. Kagami, H. Mariko and H.
Yamamoto, 1987, IEEE Transl. J. Magn. Jpn.
TJMJ-2-2, 175.
Naoe, M., Y. Hoshi, T. Ishiwatari and S. Yamanaka,
1982, J. Appl. Phys. 53, 7807.
Netzer, EP., and E. Bertel, 1982, Adsorption and
catalysis on rare earth surfaces, in: Handbook on the
Physics and Chemistry of Rare Earths, Vol. 5, eds
K.A. Gschneidner Jr and L. Eyring (North-Holland,
Amsterdam) ch. 43.
Netzer, EP., and J.A.D. Matthew, 1987, Inelastic
electron scattering measurements, in: Handbook on
the Physics and Chemistry of Rare Earths, Vol. 10,
eds K.A. Gschneidner Jr and L. Eyring (NorthHolland, Amsterdam) ch. 72.
Niemann, W., W. Malzfeldt, P. Rabe and M. Liibcke,
1986, J. Phys. (Paris) Colloq. 47(Suppl. 12), C8931.
Niemann, W., W. Malzfeldt, P. Rabe, R. Haensel and
M. Lfibcke, 1987, Phys. Rev. B 35, 1099.
Niihara, T., S. Takayama, K. Kaneko and Y. Sugita,
1984, Appl. Phys. Lett. 45, 872.
Nilsson, A., B. Eriksson, N, M~rtensson, J.N.
Andersen and J. Onsgaard, 1987, Phys. Rev. B 36,
9308.
Nilsson, A., B. Eriksson, N. M~trtensson, J.N.
Andersen and J. Onsgaard, 1988, Phys. Rev. B 38,
10357.
Nix, R.M., and R.M. Lambert, 1987, Surf. Sci. 186,
163.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Nix, R.M., and R.M. Lambert, 1989a, J. Phys. Chem.
93, 4189.
Nix, R.M., and R.M. Lambert, 1989b, Surf. Sci. 220,
L657.
Nix, R.M., T. Rayment, R.M. Lambert, J.R. Jennings
and G. Owen, 1987, J. Catal. 106, 216.
Nix, R.M., R.W. Judd and R.M. Lambert, 1988a, Surf.
Sci. 203, 307.
Nix, R.M., R.W. Judd and R.M. Lambert, 1988b, Surf.
Sci. 205, 59.
Nix, R.M., R.W Judd and R.M. Lambert, 1989a, Surf.
Sci. 215, L316.
Nix, R.M., R.W. Judd, R.M. Lambert, J.R. Jennings
and G. Owen, 1989b, J. Catal. 118, 175.
Numata, T., H. Kiriyama and S. Inokuchi, 1988,
J. Appl. Phys. 64, 5501.
Nyholm, R., I. Chorkendorff and J. Schmidt-May,
1984, Surf. Sci. 143, 177.
Oesterreicher, H., H. Bittner and K. Shuler, 1979,
J. Solid State Chem. 29, 191.
Okuno, H., Y. Sakaki and Y. Sakurai, 1986, J. Phys.
D 19, 873.
Onsgaard, J., and I. Chorkendorff, 1986, Phys. Rev. B
33, 3503.
Onsgaard, J., and O. Ellegaard, 1984, Nucl. Instr. Meth.
Phys. Res. B 2, 666.
Onsgaard, J., and A. Ignatiev, 1982, Surf. Sci. 115,
270.
Onsgaard, J., S. Tougaard and P. Morgen, 1979, Appl.
Surf. Sei. 3, 113.
Onsgaard, J., S. Tougaard, E Morgan and F. Ryborg,
1980, J. Electron Spectmsc. Relat. Phenom. 18, 29.
Onsgaard, J., I. Chorkendorff and O. Sorensen, 1983,
Physica Scfipta T 4, 169.
Onsgaard, J., I. Chorkendorff, O. Ellegaard and O.
Sorensen, 1984, Surf. Sci. 138, 148.
Owen, G., C.M. Hawkes, D. Lloyd, J.R. Jennings,
R.M. Lambert and R.M. Nix, 1987, Appl. Catal.
33, 405.
Palenzona, A., and A. Iandelli, 1974, J. Less-Common
Met. 34, 121.
Palmstrom, C.J., N. Tabatabaie and S.J. Allen Jr, 1988,
Appl. Phys. Lett. 53, 2608.
Palmstrom, C.J., K.C. Garrisson, S. Mounier, T. Sands,
C.L. Schwartz, N. Tabatabaie, S.J. Allen Jr, H.L.
Gilchrist and EE Miceli, 1989, J. Vac. Sci. &
Technol. B 7, 747.
Palmstrom, C.J., S. Mounier, T.G. Finstad and RE
Miceli, 1990, Appl. Phys. Lett. 56, 382.
Pan, S.M., Z.B. Zhao, H. Jin, E Tian and R.Z. Ma,
1989, J. Magn. & Magn. Mater. 78, 371.

201

Parker, S.EH., R.J. Grundy and J. Fidler, 1987a,

J. Magn. & Magn. Mater. 66, 74.
Parker, S.EH., RJ. Pollard, D.G. Lord and R.J. Grundy,
1987b, IEEE Trans. Magn. MAG-23, 2103.
Parks, R.D., S. Raaen, M.L. den Boer, V. Murgai and
T. Mihalisin, 1983, Phys. Rev. B 28, 3556.
Parks, R.D., S. Raaen, M.L. den Boer, Y.S. Chang and
G.E Williams, 1984, Phys. Rev. Lett. 52, 2176.
Parmigiani, E, 1982, Phys, Lett. A 92, 419.
Parth~, E., and J.M. Moreau, 1977, J, Less-Common
Met. 53, 1.
Partin, D.L, 1983a, Appl. Phys. Lett. 43, 996.
Partin, D.L, 1983b, J. Vac. Sci. & Technol. B 1, 174.
Partin, D.L., 1984, J. Electron. Mater. 14, 493.
Partin, D.L., 1985, Supeflattices & Microstructures 1,
131.
Partin, D.L., 1987, Mater. Res. Soc. Syrup. Proc. 90,
47.
Partin, D.L, 1988, IEEE Trans. Quantum Electron.
QE-24, 1716.
Partin, D.L., and C.M. Thrush, 1984, Appl. Phys. Lett.
45, 193.
Pickart, S.J., H.A. Halpern, R.J. Gambino and T.R.
McGuire, 1984, J. Appl. Phys. 55, 1763.
Piecuch, M., L.T. Baczewski, J. Durand, G. Marchal,
E Deleroix and H. Nabli, 1988, J. Phys. (Paris)
Colloq. 49(Suppl. 12), C8-1755.
Poate, J.M., 1978, J. Vac. Sci. & Technol. 15, 1636.
Protsenko, I.E., 1988, Soy. Phys. J. 6, 460.
Provo, J.L., 1979, J. Vac. Sci. & Teehnol. 16, 230.
Quemerais, A., B. Loisel, G. Jezequel, J. Thomas and
J.C. Lemonnier, 1981, J. Phys. F 11, 293.
Raaen, S., 1990, Solid State Commun. 73, 389.
Rahman Khan, M.S., 1977, J. Bangladesh Acad. Sci.
1,1.
Rahrnan Khan, M.S., 1978, Indian J. Phys. 52 A, 561.
Rahman Khan, M.S., 1981, Indian J. Phys. 55 A, 23.
Rahman Khan, M.S., 1984, Thin Solid Films 113,
207.
Rahman Khan, M.S., 1987a, J. Mater. Sci. Lett. 6,
160.
Rahman Khan, M.S., 1987b, Bull. Mater. Sci. 9, 181.
Rahman Khan, M.S., and R.E Miller, 1979, J. Phys.
D 12, 271.
Raisanen, A., A. Wall, S. Chang, R Philip, N. Trouiller,
A. Franciosi and D.J. Peterman, 1988, J. Vac. Sci.
& Technol. A 6, 2741.
Rajora, O.S., and A.E. Curzon, 1984, J. Mater. Sci.
Lett. 3, 627.
Ramakrisna, K., and O.N. Srivastava, 1987, J. Mater.
Sci. Lett. 6, 15.

202

M. GASGNIER

Ramesh, R., J.K. Chen and G. Thomas, 1987, J. Appl.

Phys. 61, 2993.
Ratajczak, H., and I. Grsciafiska, 1990, Acta Phys.
Polonica A 77, 723.
Ratajczak, H., T. Lucifiski, A. Kowalczyk, I.
Grsciafiska and P. SovLk, 1989, J. Magn. & Magn.
Mater. 81,341.
Ratajczak, H., I. Grsciafiska and A. Kowalczyk,
1990a, Acta Phys. Slov. 40, 223.
Ratajczak, H., I. G6sciafiska and T. Lucifiski, 1990b,
Phys. Status Solidi a 118, 189.
Ran, C., 1982, J. Magn. & Magn. Mater. 30, 141.
Ran, C., 1983, J. Magn. & Magn. Mater. 31/34, 874.
Ran, C., and S. Eichner, 1981a, Phys. Rev. Lett. 47,
939.
Rau, C., and S. Eichner, 1981b, Springer Ser. Chem.
Phys. (Springer N.Y.) Vol. 17, p. 138.
Ran, C., and S. Eichner, 1986, Phys. Rev. B 34, 6347.
Ran, C., and C. Jin, 1988, J. Phys. (Paris) Colloq.
49(Suppl. 12), C8-1627.
Ran, C., K. Waters and N. Chen, 1990, J. Elect.
Spectrosc. Relat. Phenom. 51,291.
Rayane, D., D. Mrlinon, P. Cabuad, B. Hoareau, H.
Tribollet and B. Broyer, 1989, Phys. Rev. A 39,
6056.
Reichl, R., and K.H. Gaukler, 1986, Appl. Surf. Sci.
26, 196.
Reichl, R., and K.H. Gaukler, 1988, Appl. Surf. Sci.
91, 460.
Reihl, R., and EJ. Himpsel, 1982, Solid State
Commun. 44, 1131.
Reshentnikova, L.V., and M.Kh. Yuldasheva, 1979,
Bull. Acad. Sci. USSR, Set. Phys. 43, 74.
Richter, H.J., R.S. Smith, N. Herses, M. SeehmanEggebert and E Wennekers, 1988, Appl. Phys. Lett.
53, 99.
Rodewald, W, and W. Femengel, 1988, IEEE Trans.
Magn. MAG-24, 1638.
Rodewald, W., and P. Schrey, 1990, J. Magn. & Magn.
Mater. 83, 206.
Ronay, M., C.E Aliotta and C.M. Serano, 1982,
J. Appl. Phys. 53, 4919.
Rosengren, A., and B. Johansson, 1982, Phys. Rev. B
26, 3068.
Rossi, D., R. Ferro and R. Marazza, 1975, J. LessCommon Met. 40, 345.
Rossi, G., 1987, Physica Seripta 35, 590.
Rossi, G., and A. Barski, 1985, Phys. Rev. B 32,
5492.
Rossi, G., and A. Barsll, 1986, Solid State Commun.
57, 277.

Rossi, G., J.J. Yeh, I. Lindau and J. Nogami, 1984,

Phys. Rev. B 29, 5389.
Rossi, G., J.J. Yeh, I. Lindau and J. Nogami, 1985,
Surf. Sci. 152/153, 743.
Sagasaki, M., T. Kobayashi, S. Tsunashima and S.
Uchiyama, 1985a, IEEE Transl. J. Magn. Jpn.
TJMJ-1-3, 369.
Sagasaki, M., T. Kobayashi, S, Tsunashima and S.
Uchiyama, 1985b, Nippon Oyo Gill Gakkaishi 9,
105. In Japanese.
Sahashi, M., T. Mizoguchi and K. Inomata, 1983,
J. Magn. & Magn. Mater. 31/34, 225.
Sakaguchi, H., and G. Adachi, 1990, J. Less-Common
Met. 162, 231.
Sakaguehi, H., H. Nagai, G. Adachi and L Schokawa,
1985a, J. Mater. Sci. Lett. 4, 347.
Sakaguchi, H., N. Taniguehi, G. Adachi and J.
Shiokawa, 1985b, in: New Frontiers in Rare Earth
Science and Applications, Proc. Int. Conf. on Rare
Earths Developments and Applications, eds Xu
Guangxian and Xiao Jimei (Science Press, Beijing)
Vol. 2, p. 1076.
Sakaguehi, H., N. Taniguchi, H. Nagai, K. Niki, G.
Adachi and J. Shiokawa, 1985c, J. Plays. Chem. 89,
5550.
Sakaguchi, H., G. Adachi and J. Shiokawa, 1986a,
J. Less-Common Met. 118, L1.
Sakaguchi, H., H. Nagai, G. Adachi and J. Shiokawa,
1986b, J. Less-Common Met. 126, 83.
Sakaguchi, H., Y. Yagi, N. Taniguchi, G. Adaehi and J.
Shiokawa, 1987, J. Less-Common Met. 135, 137.
Sakaguchi, H., N. Taniguchi, H. Seri, J. Shiokawa and
G. Adachi, 1988, J. Appl. Phys. 64, 888.
Sakaguchi, H., H. Seri and G. Adachi, 1989a, Kidorui
14, 92. In Japanese.
Sakaguchi, H., Y. Yagi, J. Shiokawa and G. Adachi,
1989b, J. Less-Common Met. 149, 185.
Sakaguchi, H., H. Seri and G. Adachi, 1990, J. Phys.
Chem. 94, 5313.
Sakamoto, Y., T.B. Flanagan and T. Kuji, 1985, Z.
Phys. Chem. N.E 143, 61.
Sakamoto, Y., M. Yoshida and T.B. Flanagan, 1986,
J. Mater. Res. 1, 781.
Sakamoto, Y., K. Takao, M. oshida and T.B.
Flanagan, 1988, J. Less-Common Met. 143, 207.
Sakamoto, Y., K. Takao and M. Ohmall, 1990a,
J. Less-Common Met. 162, 343.
Sakamoto, Y., EL. Chert, M. Furukawa and K. Minc,
1990b, J. Less-Common Met. 159, 191.
Sakamoto, Y., K. Takao and Y. Nagaoka, 1991,
J. Mater. Sci. Lett. 10, 341.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

203

Salamanca-Young, L., D.L. Partin and J. Heremans,

Shikhmanter, L., M. Talianker and M.E Dariel, 1983b,
1988, J. Appl. Phys. 63, 1504.
J. Less-Commnn Met. 92, 247.
Salas, EH., and M. Mirabal-Garcia, 1989, J. Appl. Shin, S.C., M.L. Cofield and R.H.D. Nuttall, 1987,
Phys. 66, 5523.
J. Appl. Phys. 61, 4326.
Salas, EH., and M. Mirabal-Garcia, 1990a, Phys. Rev. Shinjo, T., K. Mibu, S. Ogawa and N. Hosoito, 1989a,
B 41, 10859.
MRS Symp. Proc. (Growth Charact. Prop. Ultrathin
Satas, EH., and M. Mirabal-Garcia, 1990b, J. Magn.
Magn. Films-Multilayers) 151, p. 87.
& Magn. Mater. 83, 109.
Shinjo, T., K. Yoden, N. Hosoito, J.R Sanchez and
Sarpal, B.K., C. Blancard, J.R Connerade, J.M. Esteva,
J.M. Friedt, 1989b, J. Phys. Soc. Jpn. 58, 4255.
J. Hormes, R.C. Karnatak and U. Kuentgens, 1991, Shirai, H., H. Sakaguchi and G. Adachi, 1990, J. LessJ. Phys. B 24, 1593.
Common Met. 159, L17.
Sato, N., 1986, J. Appl. Phys. 59, 2514.
Sigrist, M., G. Chassaing, J.C. Francois, E AntomanSato, N., and K. Habu, 1987, J. Appl. Phys. 61, 4287.
geli, N. Zema and M. Piacentini, 1987, Phys. Rev.
Sato, N., K. Habu and T. Onaya, 1987, IEEE Trans.
B 35, 3760.
Magn. MAG-23, 2614.
Singh, A.K., and O.N. Srivastava, 1979, Jpn. J. Appl.
Sato, R., N. Saito and T. Morishita, 1988, IEEE Trans.
Phys. 18, 1603.
Magn. MAG-24, 2458.
Sing/a, A.K., and O.N. Srivastava, 1980, Z. Metallkd.
Savrin, V.D., A.V. Elyutin, N.E Mikhailova and Z.V.
71, 57.
Eremenko, 1988, Russ. J. Inorg. Chem. 33, 1250.
Singh, A.K., and O.N. Srivastava, 1981, Thin Solid
Schelleng, J.H., D.W. Forester, E Lubitz and C.
Films 81,367.
Vittoria, 1984, J. Appl. Phys. 55, 1805.
Singh, A.K., and O.N. Srivastava, 1982, J. LessSchiffmacher, G., 1983, Progr. Cryst. Growth Charact.
Common Met. 88, 97.
17(I/4), 379.
Singh, A.K., and O.N. Srivastava, 1983, Z. Metallkd.
Schneider, G., 1988, P h . D . Thesis, Universit~it
74, 440.
Stuttgart (Germany).
Singh, A.K., and O.N. Srivastava, 1985, Z. Metallkd.
Schneider, W.D., C. Laubschat and B. Reihl, 1983,
76, 760.
Phys. Rev. B 27, 6538.
Singh, O., and A.E. Curzon, 1982, Solid State
Schwarz, R.B., and W.L. Johnson, 1983, Phys. Rev.
Commun. 44, 1121.
Lett. 51,415.
Singh, S.K., A.K. Singh and O.N. Srivastava, 1985,
Sellmyer, D.J., 1991, plenary lecture, 19th Rare Earth
Int. J. Hydrogen Energy 10, 523.
Conf. Res., Lexington, KN, USA.
Singleton, J.H., and L.N. Yannopoulos, 1975, J. Vac.
Sellmyer, D.J., and Z.S. Shah, 1987, J. Appl. Phys. 61,
Sci. & Technol. 12, 414.
4323.
Smith, D.A., I.E Jones and I.R. Harris, 1982, J. Mater.
Sellmyer, D.J., and Z.S. Shan, 1989, in: Proc. MRS
Sci. Lett. 1, 463.
Intern. Meet. Adv. Mater. (Multilayers), Vol. 10 Smith, H.K., A.G. Moldovan, R.S. Craig, W.E. Wallace
(Materials Research Society, Pittsburgh, PA) p.
and S.G. Sankar, I980, J. Solid State Chem. 32,
379.
239.
Shan, Z.S., and D.J. Setlmyer, 1988, J. Appl. Phys. 64, Smith, R.L., W.D. Corner and B.K. Tanner, 1980a,
5745.
J. Phys. E 13, 620.
Shah, Z.S., and D.J. Sellmyer, 1990, J. Appl. Phys. 67, Smith, R.L., W.D. Corner and B.K. Tanner, 1980b,
5713.
J. Magn. & Magn. Mater. 20, 265.
Shan, Z.S., S. Naris, K.D. Aylesworth and D.J. Smutek, M., and S. Cern), 1985, Thermodyn. Acta
Sellmyer, 1988, J. Appl. Phys. 63, 3218.
91, 31.
Shah, Z.S., D.J. Sellmyer, S.S. Jaswal, Y.J. Wang and Stadelmaier, H.H., EJ. Cadieu and N.C. Liu, 1987,
J.X. Shen, 1989, Phys. Rev. Lett. 63, 449.
Mater. Lett. 6, 80.
Shen, Y., and D.E. Laughlin, 1990, J. Appl. Phys. 67, Stenborg, A., and E. Bauer, 1987a, Surf. Sci. 185,
7510.
394.
Shikhmanter, L., M. Talianker and M.R Dariel, 1982, Stenborg, A., and E. Bauer, 1987b, Phys. Rev. B 36,
Thin Solid Films 90, 51.
5840.
Shikhmanter, L., M. Tlianker and M.E Dariel, 1983a, Stenborg, A., and E. Bauer, 1988, Solid State
J. Phys. & Chem. Solids 44, 745.
Commun. 66, 561.

204

M. GASGNIER

Stenborg, A., J.N. Andersen, O. Bj6rneholm, A.

Nilsson and N. Mfirtensson, 1990, J. Electron.
Spectrosc. Relat. Phenom. 52, 47.
Strasser, G., 1988, Vak-Technik 37, 213. In German.
Strzeszewski, J., A. Nazareth, G.C. Hadjipanayis, K.D.
Aylesworth, Z.R. Zhao and D.J. Sellmyer, 1988,
Mat. Sci. & Eng. 99, 153.
Strzeszewski, J., T. Tsoukatos and G.C. Hadjipanayis,
1990, J. Appl. Phys. 67, 4966.
Suezawa, Y., Y. Gond6, T. Nakao and K. Kakuno,
1983, J. Magn. & Magn. Mater. 35, 323.
Suezawa, Y., R. Tsuchiya and Y. Gond6, 1985, Jpn. J.
Appl. Phys. 24, L140.
Surplice, N.A., and W. Brearley, 1978, Surf. Sci. 72,
84.
Surplice, N.A., and K. Kandasamy, 1982, J. Phys. D
15, 1117.
Suryanarayanan, R., and S.K. Das, 1990, J. Appl. Phys.
67, 1612.
Suzuki, A., R. Matsuda, K. Taki, Y. Watanabe and S.
Mizutani, 1988, Jpn. J. Appl. Phys. 27, 25.
Suzuki, Y., H. Masuda, T. Shibata and N. Koshizuka,
1987a, IEEE Transl. J. Magn. Jpn. TJMJ-2-2, 173.
Suzuki, Y., H. Masuda, T. Shibata and N. Koshizuka,
1987b, IEEE Trans. Magn. MAG-23, 3704.
Taborelli, M., R. Allenspach, G. Boffa and M. Landolt,
1986a, Phys. Rev. Lett. 56, 2869.
Taborelli, M., R. Allenspach and M. Landolt, 1986b,
Phys. Rev. B 34, 6112.
Takahashi, M., T. Takeda, H. Watanabe and T.
Wakiyama, 1989, Nippon Oyo Jiki Gakkaishi 13,
323. In Japanese.
Takao, K., Y. Sakamoto, M. Yoshida and T.B.
Flanagan, 1989, J. Less-Common Met. 152, 115.
Takao, K., K.L. Zhao and Y. Sakamoto, 1990, J. Mater.
Sci. 12, 1255.
Takeda, K., H. Ishii and O. Nakamura, 1989, Jpn.
Kokai Tokkyo Koho JP 89122105. Patent.
Takeda, S., 1983, J. Sci. Hiroshima Univ. A 46, 149.
In English.
Takeda, S., K. Kitano and Y. Komura, 1982, J. LessCommon Met. 84, 317.
Takeda, S., H. Horikoshi and Y. Komura, 1983,
J. Micros. 129, 347.
Tanaka, M., H. Yuzurihara and T. Tokita, 1987, IEEE
Trans. Magn. MAG-23, 2955.
Tang, W., S. Zhou, R. Wang and C.D. Graham Jr,
1988, J. Appl. Phys. 64, 5516.
Thole, B.T., G. van der Laan, J.C. Fuggle, G.A.
Sawatsky, R.C. Karnatak and J.M. Esteva, 1985,
Phys. Rev. B 32, 5107.

Tibbetts, G.G., and W.E Egelhoff Jr, 1980, L Vac. Sci.

& Technol. 17, 458.
Tiwald, T.E., J.A. Woollam and D.J. Sellmyer, 1988a,
J. Appl. Phys. 63, 3215.
Tiwald, T.E., J.A. Woollam, Z.S. Shan and D.J.
Sellmyer, 1988b, Mater. Res. Soc. Symp. Proc. 103
(Multilayers) 239.
Tkach, VI., V.E Bashev, A.B. Lysenko and I.S.
Miroshnichenko, 1980, Phys. Met. Metall. 48(1),
183.
Toki, K., 1986, Jpn. Kokai Tokkyo Koho JP 86104 445.
Patent in Japanese.
Tougaard, S., and A. Ignatiev, 1981, Surf. Interface
Anal. 3, 3.
Toxen, A.M., T.H. Geballe and R.M. White, 1988,
J. Appl. Phys. 64, 5431.
Tsukahara, S., and T. Kanayama, 1983, J. Magn. &
Magn. Mater. 31/34, 927.
Tsukahara, S., T. Kanayama and H. Tanoue, 1983,
J. Magn. & Magn. Mater. 35, 319.
Tsunashima, S., T. Shinoda, H. Miyatake and S.
Uchiyama, 1980, J. Appl. Phys. 51, 5901.
Tsunashima, S., T. Ichikawa, M. Nawate and
S. Uchiyama, 1988, J. Phys. (Paris) Colloq.
49(Suppl. 12), C8-1803.
Uchida, H., H. Uchida and Y.C. Huang, 1984, J. Lesscommon Met. 101,459.
Uchida, H., T. Ebisawa, K. Terao, N. Hosoda and Y.C.
Huang, 1987, J. Less-Common Met. 131, 365.
Uchida, H., T. Ebisawa, A. Denda, T. Itoh and S.
Tsuzuki, 1989, Z. Phys. Chem. N.E 164S, 1313.
Uchida, K., M. Murakami and H. Kawabata, 1987,
Jpn. Kokai Tokkyo Koho JP 87 287 453. Patent in
Japanese.
Umemura, S., H. Tajika, E. Kita and A. Tasaki, 1985,
IEEE Trans. Mag. MAG-21, 1985.
Urbaniak-Kucharczyk, A., 1986, Phys. Status Solidi b
134, 751.
Urbaniak-Kucharczyk, A., 1988, Phys. Status Solidi b
149, 543.
van Dover, R.B., E.M. Gyorgy, R.P. Frankenthal, M.
Hong and D.J. Siconolfi, 1986, J. Appl. Phys. 59,
1291.
Varkanova, R.G., 1982, Russ. J. Phys. Chem. 56, 420.
Varkanova, R.G., and L.V. Morozova, 1981, Russ. J.
Phys. Chem. 55, 1777.
Varkanova, R.G., and L.V. Morozova, 1984, Russ. J.
Phys. Chem. 58, 1760.
Varkanova, R.G., and A.S. Nazarov, 1977, Russ. J.
Phys. Chem. 51, 397.

THE INTRICATE WORLD OF RARE EARTH THIN FILMS

Varkanova, R.G., L.V. Morozova and Kh.Kh. Khanudova,
1982, Russ. J. Phys. Chem. 56, 933.
Venkert, A., N. Shamir, M. Talianker, U. Atzmony
and M.E Dariel, 1987, J. Vac. Sci. & Technol. A 5,
1730.
Verma, A.R., and P. Krishna, 1966, Polymorphism and
Polytypism in Crystals, J. Wiley and Sons, NewYork.
Vittoria, C., J.H. Schelleng, P. Lubitz and D.W.
Forester, 1984, J. Appl. Phys. 55, 2450.
von Molnar, S., C.N. Guy, R.J. Gambino and T.R.
McGuire, 1980, J. Magn. & Magn. Mater. 15/18,
I391.
yon Molnar, S., B. Barbara, T.R. McGuire and R.J.
Gambino, 1982a, J. Appl. Phys. 53, 2350.
von Molnar, S., T.R. McGuire, R.J. Gambino and B.
Barbara, 1982b, J. Appl. Phys. 53, 7666.
Waldrop, J.R., 1985, Appl. Phys. Lett. 46, 864.
Wall, A., A. Raisanen, S. Chang, P. Philip, N. Trouiller,
A. Franciosi and D.J. Peterman, 1987, J. Vac. Sci.
& Teehnol. A 5, 3193.
Wall, A., A. Raisanen, S. Chang, P. Philip, N. Trouiller,
A. Franciosi and DJ. Peterman, 1988, J. Vac. Sci.
& Teehnol. A 6, 1575.
Weaver, J.H., and D.W. Lynch, 1973, Phys. Rev. B 7,
4737.
Weaver, J.H., and C.G. Olson, 1977a, Phys. Rev. B
16, 731.
Weaver, J.H., and C.G. Olson, 1977b, Phys. Rev. B
15, 590.
Webb, D.J., R.G. Walmsley, K. Parvin, RH. Dickinson,
T.H. Geballe and R.M. White, 1985, Phys. Rev. B
32, 4667.
Webb, DJ., A.N. Marshall, Z. Sun, T.H. Geballe and
R.M. White, 1988a, IEEE Trans. Magn. MAG-24,
588.

Webb, D.J., A.N. Marshall, A.M. Toxen, T.H. Geballe

and R.M. White, 1988b, IEEE Trans. Magn. MAG24, 2013.
Weller, D., and S.E Alvarado, 1985, Z. Phys. B 58,
261.
Weller, D., and S.E Alvarado, 1986, J. Appl. Phys. 59,
2908.
WeUer, D., and S.R Alvarado, 1988, Phys. Rev. B 37,
9911.
Weller, D., and D.D. Sarma, 1986, Surf. Sci. 171,
L425.
Weller, D., S.E Alvarado and M. Campagna, 1985a,
Physica 130B, 72.
Weller, D., S.E Alvarado, M. Campagna, W. Gudat

205

and D.D. Sarma, 1985b, J. Less-Common Met. 111,

277.
Weller, D., S.E Alvarado, W. Gudat, K. Schr6der and
M. Campagna, 1985c, Phys. Rev. Lett. 54, 1555.
Wertheim, G.K., and G. Crecelius, 1978, Phys. Rev.
Lett. 40, 813.
White, R.L., 1990, J. Phys. Soc. Jpn. 59, 3018.
Wieliezka, D.M., and C.G. Olson, 1990, J. Vac. Sci.
& Technol. A 8, 891.
Wu, R., and A.J. Freeman, 1991, J. Magn & Magn.
Mater. 99, 81.
Wulz, H.G., and E. Fromm, 1986a, J. Less-Common
Met. 118, 293.
Wulz, H.G., and E. Fromm, 1986b, J. Less-Common
Met. 118, 315.
Xu, Y., 1987, Rep. IS-T-1280, Order No. DE87013602,
102 pp.
Xu, Y., and R.N. Shelton, 1987, Jpn. J. Appl. Phys. I
26(Suppl. 26-3), 1269.
Yamaguchi, S., 1982a, J. Electrochem. Soc. 129,
1145.

Yamaguchi, S., 1982b, Soy. Phys. Crystallogr. 27,

480.
Yamaguchi, S., 1982c, Exp. Technik Physics 30, 251.
In German.
Yamaguchi, S., 1982d, J. Phys. D 15, 347.
Yamamoto, H., S. Hirosawa, S. Fujimura, K. Tokuhara,
H. Nagata and M. Sagawa, 1987, IEEE Trans. Mag.
MAG-23, 2100.
Yamashita, S., M. Mokeda and K. Hara, 1988,
Jpn. Kokai Tokkyo Koho JP 88 84 005. Patent in
Japanese.
Yamauchi, K., K. Habu and N. Sato, 1988, J. Appl.
Phys. 64, 5748.
Yang, K.A., H. Homma and I.K. Sehuller, 1988,
J. Appl. Phys. 63, 4066.
Yeh, J.J., J. Nogami, G. Rossi and I. Lindau, 1984,
J. Vac. Sci. & Technol. A 2, 969.
Yoden, K., N. Hosoito, K. Kawaguchi, K. Mibu and
T. Shinjo, 1988, J. Appl. Phys. 27, 1680.
Zajac, S., 1977, Physica B 86-88, 51.
Zhao, Z.B., S.K. Xia, R.Z. Ma and S.M. Pan, 1989a,
J. Less-Common Met. 155, 75.
Zhao, Z.B., S.K. Xia, R.Z. Ma, J. Ping, M. Baihe and
S.M. Pan, 1989b, Solid State Commun. 69, 1011.
Zhao, Z.B., S.K. Xia, C.R. Wang and R.Z. Ma, 1989c,
J. Phys.: Condens. Matter 1, 7513.
Zhao, Z.B., R.Z. Ma and M. Baihe, 1989d, Scripta
Metall. 23, 743.
Zhavoronkova, K.N., and O.A. Boeva, 1989, React.
Kinet. Catal. Lett. 40, 291.

206

M. GASGNIER

Zhavoronkova, K.N., and A.V. Peshkov, 1978, Dep.

VINITI publ. No. 480-78 (23pp).
Zhavoronkova, K.N., and A.V. Peshkov, 1979, React.

Kinet. Catal. Lett. 12, 259.

Zhu, J.G., C.B. Carter, C.J. Palmstr~m and S. Mounier,
1990, Appl. Phys. Lett. 56, 1323.

Handbook on the Physics and Chemistry of Rare Earths

VoL 20
edited by K.A. Gschneidner, Jr. and L. Eyring
1995 Elsevier Science B.V. All rights reserved

Chapter 137
HYDROGEN IN RARE-EARTI-I METALS,
INCLUDING RIt2+x PItASES*
E VAJDA

Laboratoire des Solides Irradis, CNRS, Ecole Polytechnique,

91128 PaIaiseau, France

Co~e~s
Symbols and abbreviations
1. Introduction
2. Hydride preparation and phase diagrams
2.1. Specimen preparation
2.1.1. Metallic or a-phase
2.1.2. Dihydride or [3-phase
2.1.3. Trihydride or y-phase
2.2. Phase diagrams
3. Structural properties
3.1. a-Phase solid solntions
3.2. [3-Phase dihydrides
3.2.1. The hydrides of Eu and Yb
3.2.2. ,/-Phase trihydrides
4. Kinetics and thermodynamics

207
208
210
211
211
212
214
214
216
216
221
231
231
232

4.1. a-Phase solid solutions

4.2. [3-Phase dihydrides
5. Electronic properties
5.1. Solid solutions
5.2. (Mainly) t-phase hydrides
5.2.1. Metal-serniconduetor transitions
5.2.2. Discussion of specific hydrides
5.2.3. Spectroscopic studies
6. Magnetic properties
6.1. *-Phase solid solutions
6.2. (Mainly) [3-phase systems
6.2.1. Speeific RH2+x hydrides
7. Summary and outlook
References

Symbols and abbreviations

A~ac

ep

specific heat

C~ Cl3

elastic constants for transverse wave

propagation in a hexagonal crystal
c-axis modulated

AF
c

acoustic eleetron-phonon coupling

parameter
optic electron-phonon coupling parameter
lattice parameter
static lattice expansion (due to H)
dynamic (thermal) lattice expansion
antiferromagnetic (ism)
commensurate

CAM
CF
Ea
Ed
Ef

e
c
c,~
Cm~x

lattice parameter
concentration
concentrations in the and [3-phase
limiting concentration

Em
Eb
AE
EPR

A,~,opt
a

Aa/Ax
Aa~AT

crystal electric field

activation energy
activation energy for diffusion
activation energy for formation
activation energy for migration
binding energy
electronic energy gap
electron paramagnetic resonance

* This review is dedicated to the memory of J.N. Daou, a pioneer of the rare-earth hydrides.
207

232
236
243
244
248
251
255
265
266
267
271
273
286
287

208

P. VAJDA

ferromagnetic (ism)
critical field parallel to the c-axis
Herllc
critieal field perpendicular to the c-axis
HerA_c
enthalpy of formation
HF
Hs
heat of solution
ic
incommensurate
INS
inelastic neutron scattering
long-range order
LRO
metal-insulator transition
M-I
metal-semieonductor transition
M-S
3N, 4N 99.9%, 99.99% purity
conduetion electron density
ne
nuclear magnetic resonance
NMR
octahedral site
O
perturbed angular correlation
PAC
quasielastic neutron scattering
QENS
short-range order
SRO
tetrahedral site
T
anomaly temperature
Tan
Curie temperature
Tc
critical temperat~re
T~
helieoidal temperature
TH
intermediate temperature
Tr
temperature of resistivity maximum
TM
temperamre of resistivity minimum
Tm
melting temperature
Tm
TM_s(T) temperature of metal-insulator transition
in the heating run
FM

TM-s(D temperature of metal-insulator transition

TN
Tp
X

XPS
Y
YeJ
Fi
A
6
B+
P, Pr,~~
GS
Dm,mag

Ap

Ap
T

rm
690
O~

Op
Z

in the cooling run

Nel temperature
peak temperature (for recovery)
deviation from stoichiometry
X-ray photoelectron spectroscopy
reaction order
electronic specific-heat coefficient
electronic angular momentum stares in a
crystal electric field
energy gap in the magnon spectrum
stoichiometric deficit
positive muon
effective paramagnetie moment
resistivity, residual
magnetic resistivity term, in the ground
stare
additional resistivity due to hydrogen
additional resistivity due to quenching
relaxation time
magnetic propagation vector
Debye temperature
Einstein temperature
paramagnetic Curie temperature
magnetic susceptibility

1. Introduction

The rare earths absorb hydrogen readily and form solid solutions and/or hydrides
exothermally at temperatures of several hundred C. Their phase diagrams consist, in
general, of three basic parts: (a) the metallic solid solution, or c-phase, with the H atoms
inserted in the tetrahedral interstices of the host-metal lattice; (b) the equally metallic
dihydride B-phase, where the two H atoms occupy ideally the two available tetrahedral
sites; this phase crystallizes in the fcc fluorite system; (c) the insulating trihydride, or yphase, which possesses an hcp unit cell with both tetrahedral sites and the one octahedral
site filled up. A schematic view is given in fig. 1. Exeeptions are the divalent lanthanides
Eu and Yb, whose dihydrides are already insulators and exhibit an orthorhombic structure,
and Sc whose very small unit cell does not normally accept more than two H atoms.
An interesting feature of the phase diagram in fig. I is the relatively large existence
region of the pure and, in partieular, of the [3-phase (whieh extends up to the trihydride

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

209

\
(X

i f

I--

2
3
at.H / at.R

Fig. 1. Sehematic phase diagram for R-H systems, et, solid

solution; et*, ordered H-sublattice in some R-H sotid solutions;
B, dihydride; , trihydride.

for the light lanthanides La, Ce and possibly Pr). This means that the excess H atoms in
solution will, from a certain concentration on, interact with each other and form hydrogen
sublattices, which are stable at low temperature and might, in some cases, deform the host
lattice. This hydrogen ordering is observed as an anomaly in several physical properties
and has been characterized by neutron scattering. Thus, it was shown that, in the hcp cphase of several rare earths including Sc, Y and the heavy trivalent lanthanides from Ho
to Lu, the hydrogen atoms, at temperatures below N200 K, form zig-zag ehains of H - H
pairs along the c-axis. In the super-stoichiometric dihydrides RHz+x, on the other hand,
the octahedral (x) hydrogens can sometimes order in a Ni3Mo-type structure such that, for
a stoichiometric composition with x = 0.25, one observes one filled (420) plane followed
by three empty ones.
Another exciting feature is the possibility of observing a continuous decrease in the
earrier density when absorbing hydrogen, which gives rise to a metal-insulator (M-I)
transition somewhere in the interval between 2 and 3 atoms H per atom R. Moreover,
temperature dependent M-I transitions have been observed in the B-phase between
250 and 300K both in sub-stoichiometric LaH3-x and CeH3_x as weil as in superstoichiometric YH2+x and in RH2+x (R = Gd, Ho, Er) with x ,~ 0.1-0.3, which are driven
by the order-disorder transformation in the H sublattice mentioned above.
A further physical property which is strongly influenced by the hydrogen eoneentratin
in solution, x, is the magnetism both of the R-metals and of their dihydrides. Thus, not
only the number of the x-atoms direetly aets upon the RKKY interaction responsible for
the magnetism but, also, their eventual ordering into an H sublattiee leads to modifications
of the crystal-field symmetry with the appearance of new magnetic phases. It is also interesting, though in the end not too surprising, that the cerium hydride CeH2+x is a magnetically ordering (except for x = 0.05, which is paramagnetie) heavy-fermion system with a
Kondo lattice, like many other Ce-based eompounds due to their hybridized 4f-eleetrons.
In view of what was said above, one can already see the importanee of a reliable
eharaeterization of the H eoneentration in the specimens. In partieular, the distinetion
between the two tetrahedral H atoms and the octahedral (x) excess H atoms in the [3-

210

R VAJDA

RH2+x systems, is essential for any comparison between results from various laboratories.
We shall, therefore, always try to separate the true concentration in the dihydride
(somewhere from 1.85 to 2.0 atoms H per atom R) from the x-concentration in a particular
compositin, thus specifying e.g. RH1.9+010 instead of just summing up RHz05. A special
subsection will be devoted to specimen preparation and characterization to emphasize the
importance of the subject.
The field is at present in full development and the last decade has seen the production
of a large amount of new and reliable data mainly in the non-stoichiometric - and
B-phase systems. This is due to the availability of purer rare earths (there is orten
a qualitative property change when going from 3N to 4N material), to more reliable
specimen characterization, and to the application of modern measuring techniques, e.g.
low-temperature specific-heat and resistivity, NMR, internal friction, neutron- and other
spectroscopies. We feel, therefore, that a critical review of these data in the present volume
of this Handbook fills a certain gap after the "Hydride-chapter" (26) by Libowitz and
Maeland (1979) in volume 3 (devoted, by the way, to Non-metallic compounds). More
recent reviews in the field were those by Arons (1982) (new version Arons 1991) mostly
on magnetic properties; Wiesinger and Hilscher (1991) also on magnetism and mainly
on H in intermetallic compounds; and a general review by Vajda and Daou (1995) in a
Metal-Hydrogen-Systems compendium. A survey of surface and activation properties of
hydrogen on rare earths has been given by Schlapbach (1992). The outline of the present
chapter is as follows:
Sect. 2: Phase diagrams, including a subsection on specimen preparation.
Sect. 3: Structural properties, both of the alloy and of possible H sublattices.
Sect. 4: Kinetics and thermodynamics, comprising various thermally activated processes.
Sect. 5: Electronic properties, including results from electric resistivity, specific heat,
NMR, EPS etc., but leaving magnetism for the next section.
Sect. 6: Magnetic properties, including the Kondo effect.
Sect. 7: Summary and outlook.

2. Hydride preparation and phase diagrams

The establishment of reliable T-c phase diagrams in the R-H systems is intimately connected with an unambiguous specimen characterization and, hence, with a strict control
of the preparation conditions. The lack of such a control in the past has frequently led to
contradictory experimental results, in particular where non-stoichiometric compositions
are concemed. It is, therefore, necessary to include in this section a relatively broad
subsection on specimen preparation in order to optimize the sample quality, to determine
the hydrogen concentration in the particular phase and, last but not least, to compare the
data of different experiments.

HYDROGENIN RARE-EARTHMETALS,INCLUDINGRH2+x PHASES

211

2.1. Specimen preparation

Hydrogen is absorbed exothermally by rare earths and the specimens can be readily
prepared by direct action of the gas at a temperature of several hundred C. Nevertheless,
for a successful preparation of weil characterized samples without mechanical and/or
chemical defects, several conditions should be satisfied:
(1) A specimen should consist of a single, undivided piece (foil or bulk) to assure
a uniform H concentration in all parts; when divided samples are needed, as for
powder-diffraction or susceptibility experiments, it is preferable to hydrogenate a
bulk specimen and crush it afterwards.
(2) The original metal specimen should be of high purity and good metallurgical
quality: stress-free, without textures and microcraeks; high-temperature annealing
(at half the melting point) and degassing is usually required.
(3) The preparation chamber should permit the maintenance of the specimen at
rather high temperatures (above 500C) in a good vacuum (4 10-7 Torr) before
hydrogenation, because of the strong affinity of the rare earths towards oxygen
and nitrogen. Quartz tubes with ttmgsten crucibles for the specimen a r e a good
choice. A hydrogen filter (e.g. PdAg tubes) might be necessary for purification.
(4) One should possess good instrumentation for precise pressure measurements, to
warrant a reliable volumetric concentration determination. Capacitance manometers are the ideal pressure sensor-transducers for absolute readings to 10-3
between 0 and 10bars.
In the following, we shall describe the preparation conditions for the three principal
phases, a, ~ and y, distinct by their solubility limits and the critical temperatures for phase
transitions or for hydrogen-configuration changes.

2.1.1. Metallic or a-phase

The ~t-phase is a solid solution where the hydrogen atoms are distributed statistically
in the tetrahedral (T) interstices of the metal lattice and behave as impurities. For the
heavy lanthanides from Ho to Lu, as welt as for Sc and Y, there exist two varieties in the
metallic phase, c~and ct*, the latter being metastable at low temperatures and representing,
B*
in fact, an ordered configuration of the H sublattice; its limiting concentration, Cmax,
is independent of T (see sect. 3 for more details). For La through Dy, no *-phase is
observed (though there might be some traces in Dy), and the c~-phase does not form
below a temperature T of several hundred C. The hydrogen solubility in the existence
range of the c-phase increases with the atomic number of the lanthanide and seems to
be related with the lanthanide contraction: it is, e.g. at 500C, ten times larger in Lu than
in La (Sc and Y behave in this context like heavy lanthanides); it increases linearly with
the melting temperature of the rare earth.
In table 1, the various critical temperatures and concentrations in the c-phase of the
rare earths are collected. We note that, in view of the high mobility of hydrogen in the ctphase compared to the *- or ~-phases, temperatures of T >~500C are most appropriate
for charging. Finally, it is important to know that the characteristics listed in table 1 are

212

P. V A J D A

Table 1
C r i t i c a l t e m p e r a t u r e s a n d e o n c e n t r a t i o n s in t h e c - p h a s e a

Vat
(cm3/mol)

T m (C)

T~~ b (C)

T ....

(C)

T ....

(oc)

c~x (at.H/R)

C~oo*c
(at.H/R)

Sc

15.039

1541

n.m. c

19.893

1522

n.m.

n.m.

0.35

n.m.

400

0.20

/>0.32

La

22.602

918

350

0.04

Ce

20.696

798

380

0.06

Pr

20.803

931

380

0.06

Nd

20.583

1021

280

0.08

Sm

20.000

1074

200

0.15

Gd

19.903

1313

180

0.24

Tb

19.310

1356

140

Dy

19.004

1412

(80)

Ho

18.752

1474

90

Er

18.449

1529

120

340

0.06

0.32

Tm

18.124

1545

140

380

0.11

0.34

Lu

17.779

1663

160

420

0.20

0.36

a F r o m D a o u a n d V i a l l a r d (1972).

n.m.

0.26

<0.01

0.28

0.03

0.30

c n.m., n o t m c a s u r e d .

b T~~ is t h e (ct ~ [3) t r a n s i t i o n t e m p e r a t u r e

e x t r a p o l a t e d to c a -- 0.

hardly modified by the impurity contents of the metal when going from 4N to 3N material,
in contrast to the B-phase described below.
2.1.2. Dihydride or il-phase
The dihydride phase crystallizes in general in the fcc fluorite-type structure, with an ideal
composition of RH The frequently noted stoichiometric deficit, RH2_6, is caused by
impurities and structural defects of various kinds such as surfaces and grain boundaries.
Thus, 6 is smaller for bulk specimens than for powder and for specimens made of
originally 4N material than for those from 3N; it can reach values of up to 0.2, leading
to "pure" dihydrides with a formula RH1.8.
The -4 ~ transformation from hcp to fee is accompanied by a shift of atomic planes
and can cause severe damage unless certain precautions are taken. Thus, an anneal
of several hours at T = 500-800C (depending on the rare earth) is a good procedure
before the introduction of H to form the dihydride. In fact, at these temperatures the
hydrogen will readily diffuse through the whole specimen, without the inhomogeneous
formation of H-rich ~-RH2+x regions near the surface and H-deficient two-phase regions
comprising - and ~-areas in the bulk causing pressure gradients and possible destruction
of the sample. The ~--~ ~ transformation will occur smoothly and will be total for the
whole specimen. Single-crystal work on hydrides was limited until now to the system
CeH2+x because of the favourably low melting point of Ce-metal (Libowitz 1972, Abeln

HYDROGEN IN RARE-EARTHMETALS, INCLUDINGRH2 PHASES

213

1987), which permits hydrogenation in the liquid phase and the crystal growth from
the charged melt. Otherwise, the loss o f monocrystallinity during the a--* 13 transition
is difficult to avoid, and the only relatively successful test known to us was made
on TbD2 where a Tb-crystal was deuterated very slowly at T ~ 0 . 8 Tm (Daou and
Vajda, unpublished). But even with the above precautions, the resulting specimens are
brittle and all mechanical manipulafions such as cutting or fixing o f leads should be
done on the metal before hydrogenation. In partieular, eleetric contacts are normally
made from platinum wire spot-welded on the metal ends before hydrogen is introduced
into the sample. An additional problem that occurs sometimes is the low chemical
stability o f the hydrides in air (especially for the light lanthanides), as they can form
hydroxides and decompose; they should always be stored in vaeuum or, at least, in
dry air.
The other important problem in the preparation o f p-phase specimens is their
eharacterization. As already mentioned above, the particular properties induced by the
octahedrally dissolved excess H atoms in a superstoichiometric system [3-RH2x, require
a precise determination o f their concentration and, therefore, o f the true concentration
for the saturated dihydride, 13-RH2_6, i.e. o f the value o f 6. To achieve this, one can use
the fact that the equilibrium pressure at the isothermal plateau o f the dihydride RH2_6
is very low (10 -6 Torr at 500-600C) while, above its upper limit, it begins to increase
rapidly (orte needs e.g. - 1 Torr to absorb x = 0.01 H/R in excess o f the upper limit at the
hydrogenation temperature o f ~400C). This means that, at the same temperature, the
Table 2
Lower boundaries of the pure [3-phase RH2_6; solubility lirnit x~x of their solid solutions ~-RH2_6+x;and
transition temperatures Tl (13~ ~/) and Tu (~,--~13);for Y, Tb and Lu
R

Ref.

X~max(at.H/R) Ref.

1
2
3
4

0.2, 0.33

7, 8

Y 4N
Tb 3N

1.80
1.90
1.97-1.99
1.92-1.94

0.10
0.35

3
4

Tb 4N

1.96-2.00

0.25

Lu 3N

1.82
1.85
1.97-1.99

1
7
6

0.25
0.38
0.03

8
7
6

Y 3N

Lu 4N

2-6

References
(1) Daou et al. (1984a)
(2) Juckum (1980)
(3) Vajda and Daou (1991), Daou and Vajda (1992)
(4) Vajda et al. (1985)

T~ (C)
300
300
300
250
285
250
285
320
326
320
326

Ref.

r (c)

8
8
8
8
7
8
7
8
7
8
7

(5) Vajda et al. (1987a)

(6) Daou et al. (1988b)
(7) Mintz et al. (1976)
(8) Daou and Vajda, unpublished.

400
400
400
350
362
350
362
410
416
410
416

Ref.
8
8
8
8
7
8
7
8
7
8
7

214

R VAJDA

stability of the H atoms at the T sites will be much higher than that of the x-atoms at
the O sites, giving an easy means to separate the two H species by simply pumping oft
the less stable x-atoms at e.g. 500-600C in a vacuum of < 10-6 Torr. After that, one
is certain to have a saturated "pure" dihydride leit; and the super-stoiehiometric RI-I2+x
compound ean then be formed by just selecting the right P - T value, for example 10100 Torr at a temperature which is high enough to absorb the hydrogen but which should
not be too low in order avoid the formation of the y-phase (cf. table 2). In table 2, we are
also giving examples of the maximum solubility X~maxof the octahedral hydrogens in the
~-phase before transformation into the y-phase; one notes the strong purity dependence
ofx~max attaining a factor 10 in the ease of Lu.
2.1.3. Trihydride or y-phase
Most R-H systems exhibit a seeond phase transformation [3--+y in the interval RH2RH3. The y-phase possesses an hcp structure, but with a bigger unit cell than the metal
because of symmetry loss due to an off-center displacement of the H atoms occupying the
T and O sites (see seet. 3). Those exhibiting only one (a ~ [3) transformation are the two
divalent lanthanides Eu and Yb, Se (too small to incorporate a third H atom under normal
pressure conditions) as well as the systems La-H, Ce-H and Pr-H, with a big enough unit
cell to permit a complete occupation of the O sites in the [3-phase. The latter, however,
exhibit second-order transformations with slight cubie-to-tetragonal deformations of the
host lattice.
Thus, the problems encountered during the first transformation ~ [3 are aggravated
in the ~ ~ 7 one: the brittleness increases due to growing H coneentration and also
because of the generally lower absorption temperatures (ef. table 2). The extreme chemical
instability requires that a sample be maintained under hydrogen atmosphere, with no air
contact, whatsoever. It is advised to perform the desired measurements in situ.
2.2. Phase diagrams
In fig. 1, we presented a schematic phase diagram showing the three principal phases
of the general R-H system. The actual phase diagrams had been determined in the past
for some of the light rare earths (La, Ce, Nd) and described among others by Libowitz
and Maeland (1979) in the earlier review on this subject in this Handbook series. In
the meantime, the known phase diagrams, such as La-H and Ce-H, have been given
additional attention and new ones, Pr-H, Y-H and Lu-H, have been, at least in part,
established. Among the more remarkable new phenomena is the confirmed existence of a
high-temperature peritectic in the light R-H systems (Bischof et al. 1985) permitting the
hope of growing single crystals of these hydrides [in addition to those already fabricated
in the Ce-H system by Libowitz (1972)]. Most interesting, however, is the observation
of hydrogen sublattice ordering at low temperatures (mostly below 300K) in both aand B-phase systems. This seems to be a general phenomenon, which is sometimes
accompanied by a distortion of the host lattice. This will be treated in more detail in

HYDROGENIN RARE-EARTHMETALS,1NCLUDINGRH2+x PHASES

Weight P e r c e n t Hydrogen
04
0
08
1
15

0z

1530

215

1478oC::

(o-.p)
1330-

0
0
0
0
o
0

It30

0
0
0
0
0
0
,,>0o

/0

..)

930

(Y)

730

"%,

xlt~

~3

*,x

xla
e

530

E
E--,

e~aa

Oe &

/&

330

.a
,IP'*

13o

...............................................,,y2..............
,
~'

-70

~,

Il

t
!
!
!

-270

.........

, . . . . . . . . . . . . .~
...
20

I0

30

. . . . . . . . .

40

. . . . . . . . .

. . . . . . . . .

50

60

. . . . . . . . .

i . . . . . . . . .

70

80

Atomic P e r c e n t Hydrogen
Fig. 2. Phase diagram in the Y-H system assessed from thermodynamic and structural data. Note the abscissa
scale with x = H/(Y + H) in at.% and the designations (x', 6 and for the et*, [3 and used in this review. The
various symbols refer to different investigations, details are found in Khatamian and Manchester (1988).

sect. 3, where we shall also present partial phase diagrams as proposed from various
physical measurements.
In figs. 2 and 3, we are exhibiting the phase diagrams for the systems Y-H and La-H,
respectively, which resulted from a critical assessment by Khatamian and Manchester
(1988, 1990) of the available data. Note in particular the tetragonally distorted lowtemperature phase in the La-H system near the trihydride discovered by Klavins et al.
(1984) and the ordered *-phase in a-YHx. The latter has been seen in several c~-RHx
systems (cf. also the Lu-H system in fig. 4) and has been determined by neutron scattering
(Blaschko et al. 1985) to be due to H - H pairs ordered in chains along the c-axis. Figure 5
gives the Pr-H phase diagram, which resembles the La-H diagram at high temperature.
For all other systems, reliable data are still lacking, in particular in the region of the
y-phase existence (i.e. in the heavy R-H systems from Nd-H on), though there is some
progress noted in the RH2+x systems at low temperatures (cf. sect. 3).

216

E VAJDA
Weight Percent Hydrogen
0.5

1.5

25

1200

1100

1000

0
0

81111
700

~7

..~ 600

~7
400
310
3OO

283'C

200
100
0
-t00
0

I0

20

La

30
40
50
Atom/e Percent Hydrogen

60

70

BO

Fig. 3. Phase diagram in the La-H system assessed from thermodynamic and structural data. Note the abseissa
seale with x=H/(La+ H) in at.% and the designation 6 for the 13used in this review. The various symbols refer
to different investigations, details are found in Khatamian and Manchester (1990).

3. Structural properties
3.1. a-Phase solid solutions

The site occupation by hydrogen in solid solution in the rare earth metals has been
determined in the early eighties, after some controversy, through unambiguous ionchanneling experiments on single crystals of c-LuDx (Danielou et al. 1981) as well as by
neutron diffraction on c-YDx (Khatamian et al. 1981), on c-TmDx (Daou et al. 1983),
and on a-ScDx (Saw et al. 1983). It has been established that the hydrogen atoms were
localized on the T sites of the hcp lattice over the entire measured temperature range
between 4 K and 300 K. Later quasi-elastic neutron-scattering experiments (e.g. Anderson
et al. 1989) and NMR studies (Barnes et al. 1989) at high temperatures extended the
validity of the T-site occupation to the whole domain of c-phase existence where a site
can still be specified (several hundred C).
In table 3, we present the lattice parameters in the hcp a-phase for various hydrogen (or
deuterium) concentrations and at various temperatures, together with their expansivities,

HYDROGEN IN RARE-EARTH METALS, 1NCLUDING RH2+~ PHASES

0

1000

217

Lu- H

800

600

6"
o
1-

40O

200

F
/
-200 ~0 0.2

1 1.2
at.H/Lu

0.6

1.6

1100

li:t
2 2.2

Fig. 4. Phase diagram in the Lu-H system constructed

from resistivity and thermodynamic data. Open circles,
Daou (1965) and unpubl, data; solid cireles, Beaudry and
Spedding (1975); open squares, Subramanian and Smith
(1982); solid squares, Jones et al. (1964); crosses, Daou
et al. (1988b).

Pr(i )

p1000
q

(s) + PrH x(s )

900

// -Pr (s,bcc)

j__

__.

8001 "w~~-~---~d-pr ' ~ Y

700
600

(s,fcc)/
/

Pr(s) + PrHx(s)

10

20

30

40
50
60
70
Atomic Percent Hydrogen

Fig. 5. Phase diagram in the Pr-H system constructed from thermodynamic data. Note the abscissa scale with
x = H/(Pr+ H) in at.%. The various symbols refer to different investigations, details are found in Bischof et al.
(1985).

both static (caused by H insertion) and dynamic (thermal), determined in most cases by Xray crystallography. We note that the c/a-ratio of the high-temperature c-phase decreases

R VAJDA

218

Table 3
Crystallographic parameters of the tl.c.p, a-phase, RH(D)x

Sc

Gd
Tb

Dy

Ho

Er
Tm

Lu

Aa
a~x
(10-4/at%)

Ac

Aa

Ac

cAx

aAT

c&T

(10-4/at%)

(10-6/K)

(10~/K)

x (H/R
ratio)

T(K)

0.33D

300

3.338

5.299

2.68

1.77

0.15
0.15

80
300

3.318
3.3215

5.2775
5.286

3.47
2.71

2.85
1.96

5.0
5.0

2.0
8.5

2
2

0.30

80

3.3295

5.2845

2.91

1.87

3.6

6.7

0.30

300

3.3340

5.2945

2.63

1.52

6.6

7.75

0.19

300

3.6636

5.7900

2.22

5.33

a ()

c (A)

Ref.

2
3

0.176D

300

3.662

5.778

2.15

4.57

0.12

300

3.6542

5.7654

1.37

4.59

5.3

20.2

0.12

775

3.670

5.8175

2.28

3.81

11.9

22.9

0.22

300

3.6637

5.7940

1.93

4.76

2.9

16.4

0.22

775

3.678

5.838

2.23

3.68

12.2

14.9

0.2
0.2

775
775

c/a = 1.5890

3.6515

5.7735

5.52

3.57

7.1

16.3

0.2
0.2D

1000
800

3.6573
3.6511

5.7947
5.7752

5.33
5.36

3.22

7.1

16.3

6,7

3.57

7.8

16.7

6,7

0.2D

6,7

6
6,7

1000

3.6568

5.7945

5.23

3.21

7.8

16.7

0.2

675

3.6269

5.7255

4.15

3.21

6.5

18.7

6,7

0.2

950

3.6334

5.7550

4.11

2.97

6.5

18.7

6,7

0.2D

700

3.6304

5.7318

4.57

3.49

7.5

18.3

6,7

0.2D

950

3.6372

5.7580

4.64

3.23

7.5

18.3

6,7

0.2

645

3.6087

5.6985

3.38

3.47

9.9

16.2

6,7

0.2

900

3.6178

5.7220

3.94

3.16

9.9

16.2

6,7

0.2D

635

3.6103

5.7021

3.58

3.78

10.8

15.8

6,7

0.2D

875

3.6197

5.7237

4.19

10.8

15.8

0.035
0.27
0.08

300
775
300

3.5636
3.6035
3.5462

5.5960
5.700
5.5795

3.43
3.63
3.20

3.31
5.47

6,7
3

3.87

13.8

16.7

0.09D

300

3.5466

5.5868

2.73

5.95

8
3

0.11

300

3.5478

5.5890

2.90

4.36
5.09

6.2

25.0

9
9

0.175

775

3.572

5.656

3.53

4.06

0.205

300

3.5267

5.6071

2.98

5.05

8.7

14.8

0.20

300

3.5250

5.6082

2.76

5.18

9.2

13.2

0.20

775

3.5412

5.6385

2.73

2.78

10.9

10.9

10

0.19D

300

3.5222

5.6040

2.68

0.19D

775

3.5400

5.6386

2.78

5.18
2.88

9.9
12.3

14.9
12.2

10
10

3
10

continued on next p a g e

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+ PHASES

219

Table 3, notes

References
(1) Saw et al. (1983)
(2) Daou and Vajda, unpublished.
(3) Beaudry and Spedding (1975)
(4) Khatarnian et al. (1981)

(5) Juckum(1980)
(6) Daou et al. (1981a)
(7) Chiheb (1980)

(8) Bormet (1975)

(9) Boanet and Daou (1979)
(10) Daou and Bonnet (1974)

with increasing x for Gd, Tb, Dy and Ho, while it increases slightly for Er, Tm and Lu
(Daou et al. 1981a), reflecting the evolution of the anisotropy and of the anharmonicity
of the interatomic potential upon H insertion. Another interesting point is the exceptional
behaviour of the static expansivity in the low-temperature a*-ScH(D)x: it is larger in the
basal plane than along the c-axis leading to a decreasing c/a-ratio with increasing x, in
contrast to all other *-phase systems, probably a consequence of the unusually large
elastie anisotropy ratio C *= C12/C13 in Sc (cf. Vajda et al. 1990b, 1991e).
As already mentioned several times before, the hydrogen atoms in certain RHx systems have a tendency to organize themselves in short-range-ordered (SRO) or
long-range order (LRO) structures when cooled below temperatures corresponding to
their mobility limit. The earliest manifestation of such low-temperature ordering was
the appearanee of a resistivity anomaly near 170K in ct*-LuHx (Daou et al. 1976),
which had been ascribed to a condensation into H-H pairs after interpreting the results
of low-temperature electron irradiation experiments (Daou et al. 1977) and internalfriction measurements (Vajda et al. 1983a). The same anomaly was detected in all
a*-RH(D)x systems, with R = Sc, Y, Ho, Er, Tm and Lu (for a review see Daou and
Vajda 1988). Strong anisotropy manifestations in the magnetic system ~t*-ErH(D)x gave
a further early indication as to a c-axis alignment of these H-H pairs (Daou et al.
1980a). The ordered configuration was finally identified through diffuse-neutron scattering
experiments on single crystals of *-LuDx (Blaschko et al. 1985, 1989a) as secondneighbor D-D pairs on T sites along the c-axis surrounding a Lu-atom (fig. 6). Several
of these pairs then condense into zig-zagging quasi-linear chains parallel to the c-axis
and shift along the b~axis in the basal plane, while the chains, on the other hand,
interact repulsively to form a three-dimensional superstructure in the metal lattice (fig. 7).
The above configuration was afferwards confirmed on the systems *-ScDx (Blaschko
et al. 1989c) and ct*-YDx (Fairclough et al. 1993). The number of pairs in a chain
(i.e. the chain length) differs from one system to another and seems to correlate
with the elastic anisotropy of the metal; the chain density varies as a function of
H concentration.
In table 4 we have collected various characteristics of the ordered structures in the
a*-phase, such as the anomaly temperatures (Tan) in electric resistivity or the binding
energies as determined from a kinetic analysis of the resistivity data in the vicinity of the
anomaly via an Arrhenius plot (cf. Daou et al. 1988a). An immediately visible result
(fig. 8) is the correlation between Tan and the c/a-ratio, indicating a lower ordering
temperature (i.e. a higher mobility of H) for a c/a-ratio closer to the ideal (8/3) 1/2 = 1.63
(i.e. in systems with less distortion). Note also the isotope effect, with a 4-5 K higher Tan

220

P. VAJDA
0

1:3
0

:o
o

o~

0
D

I1

o
0

~o

3~

o
o

~
l
0

a3q
a/~

D
o

o
O

Fig. 7. Various two-chain arrangements made of D-D pairs

Fig. 6. HCP unit cell of LuD x in the *-phase on adjacent c-axes along the b-direction in a*-LttI)x. Open
showing two adjacent D-D pairs (dashed lines) circles, metal atoms; open squares, empty T-sites; solid
shifted along the b-axis. To, T l and T 2 are first squares, occupied T-sites (Blaschko et al. 1989a).
and second neighbour tetrahedral sites along the
c-axis (Vajda et al. 1986).

Table 4
Characteristics of the ordered H-configurations in ct*-RH(D, T)x

c/a
1.592
1.571
1.570
1.569
1.571
1.583

Sc

Y
Ho
Er
Tm
Lu
Lu a (lla)
Lu a

(llc)

Lu ~ (T for D)'

T~ (K)

E b (meV)

Xmax

Rel.

152.5(5)
168.0(5)
174(1)
173(2)
169(1)
166(1)

160.5(5)
173.0(5)
179(1)
177(2)
174.5(10)
172(1)
174.0(5)
174.0(5)
180(1)

50(2)
61(2)

57(2)
70(2)
72(2)

0.35(2)
0.20(1)
0.03
0.065

1
2
3
4

0.11(1)

73(3)
65(3)
74(3)

0.20(1)

6
7
7
8

64(2)
63(2)

a Single crystal.

References
(1) Daun et al. (1986b)
(2) Vajda et al. (1987c)
(3) Daou et al. (1987)

(4) Daou et al. (1981d)

(5) Daou et al. (1981e)
(6) Daou et al. (1986e)

(7) Vajda et al. (1986)

(8) Jung and Lsser (1988)

HYDR.OGEN IN RARE-EARTH METALS, INCLUDING RH2x PHASES

190

221

180

v
"~"~ 1 7 0
I.-160

150
1.57

1.58

c/a

1.59

Fig. 8. The anomaly temperature as a function of c/aratio of the metal R in c-RH(D)x systems (Daou and
Vajda 1988).

in the case of deuterium and another 5-6 K for tritium (in the case of ct*-LuTx, Jung and
Lsser 1988), again stressing the H mobility as an essential parameter for ordering; the
same is true for the binding energy Eb.
3.2. il-Phase dihydrides
As already discussed in sect. 2, it has been a problem for some time to separate the
excess hydrogen atoms, x on O sites, from the "pure" dihydride H atoms on T sites in the
non-stoichiometric dihydrides RH2+x. Thus, the assumption was that, for therrnodynamic
reasons, the O sites began to be occupied before a complete filling of the T sites: that,
for example, RHzl0 contained 1.95 H atoms on T sites and 0.15 H atoms on O sites. As
has been shown above, the substoichiometry of the "pure" dihydride depended more on
the metallurgical state of the specimen than on thermodynamics. It is, in faet, possible
to obtain practically ideal T-site occupation in the dihydride, to +0.01 H/R, when using
high-quality samples and working at high temperatures. In the following tables, we shall,
therefore, either specify the "pure" dihydride concentration (when possible) or at least
separate the x-coneentration, avoiding the sum composition, e.g. giving RH1.97+0 ' 13 instead
of RH2.10.
In table 5 the lattice parameters, a, of the cubic [3-RH2+x phases are presented as a
funetion of x at several temperatures. The static and therrnal lattice expansions have
been added when available. We note the well-known general contraetion (negative Aa/Axvalues) of the dihydride lattice with increasing x, which is an expression of the strong
ionic character ofits interaction with the excess H atoms on O sites; an example is given
in fig. 9 for the oase of YH2+x, where the break in the a(x)-curve at x = 0.10 H/Y indicates
the limit of the pure [3-phase.
Another phenomenon which can be observed through the lattice parameter is the lowtemperature ordering within the x-sublattice. It can either lead to a tetragonal distortion
of the originally cubie unit cell given as a and e in table 5, such as in the cases of LaH2+x
(fig. 10, for x = 0.80, Klavins et al. 1984), CeH2+x (fig. 11, for x = 0.16, Boroch and Kaldis
1989) and SmH2+x (for x=0.3, Greis et al. 1981, Zogal and UHritier 1991) or just give
rise to anomalies in the thermal behavior of the lattice parameter a such as e.g. in the case
of GdH2+x for x ~>0.2 (fig. 12, Chiheb et al. 1993). Such ordering was first established in

222

P. VAJDA
Table 5
Crystallographic parameters of the [3-phase, RH(D)2+x
x (H(D)/R ratio)

Sc
Y

0
0

Aa/Ax
104a
(at.% H, D) -l

Aa/AT
a

Ref.

(104/K)

4.784
5.2095

873

5.248

298

5.1996

873

5.242

298

5.2082

-0.50

5.0

0.05

298

5.2070

-0.50

5.5

0.10

298

5.2056

-0.50

6.2

2.3
18
2.2
20

2
2
2
2
3

90

5.2032

-0.45

4.2

0.05

90

5.2019

-0.45

2.6

0.10

90

5.2002

- 1.15

2.0

298
773

5.6698
5.7005

6.0

16.5

298

5.6656

-1.3

0.27

298

5.6498

-1.3

0.53

298

5.6270

-1.3

0.60

298

5.6239

30 a

5.604 (a)
5.619 (c)

298
30 a

5.6218
5.598 (a)
5.654 (e)

0.80

0.92

298

5.6186

-0.5

0.28D
0.41D

298
298

5.6295
5.6164

-1.8
-1.8

0.70D

298
30 a

5.6176
5.600 (a)
5.616 (c)

0.78D

298

5.6111

30 a

Ce

a ()

298
298

0D

La

T (K)

5.587 (a)
5.649 (c)

0.91D
1.00D

298
298

5.6100
5.6064

0
0.16

298
298

5.581
5.650

200 a

5.5500 (a)
5.5727 (c)

340

5.5612

200 a

5.5455 (a)
5.5715 (c)

0.24 b

5
5
6

-1.8

5.0

7
7

continued on next p a g e

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

223

Table 5, continued
R

x (H(D)/R ratio)

Ce (cont'd)

T(K)

Aa/Ax
104a
(at.% H, D) -~

a ()

Aa~AT
a

0.63 b

285

5.5375

10

0.90

300

5.5364

13.5

0.18D

295

5.565

4.2 a
0.29D

370
4.2 ~

ReL

( 10-6/K)

5.547 (a)
11.13 (e)
5.564

5.544 (a)
11.125 (c)

er

Nd

Sm

0.9D

295

5.546

0.96D

295

5.536

298

5.518

-1.35

10

0.25

298

5.498

-1.35

10

0.47

298

5.483

-1.35

10

0D

298

5.505

-1.25

10

9
9

0.29D

298

5.484

-1.25

10

0.48D

298

5.472

-1.25

I0

298

5.4689

8.1

773

19.5

-1.4

10

-1.4

10

298

5.5005
5.465

0.27

298

5.444

0.47
0D

298

5.430

-1.4

10

298

5.455

-1.5

10

0.09D

298

5.447

-1.5

10

0.29D

298

5.431

-1.5

10

295

5.3773

4.3

85

5.3738

2.1

11

0.15

295

5.3693

-1.0

5.6

11

85
295

5.3648

-1.1

3.4

0.26

4.5

11
11

85
295

3.2

11

0.45

5.3637
5.3556

11

-1.0
-1.35

11
11

0D

85
295

5.3521
5.3430
5.365

-1.2

12

0.20D

295

5.3518

-1.2

12

0.40D

295

5.3400

-1.2

12

0.33

295 a

0.33D

295"

3.7780
5.3647
3.7716
5.3503

(a)
(e)
(a)
(c)

12
12

continued on next page

E VAJDA

224
Table 5, continued
R

x(H(D)/Rrafio)

2" (~)

a ()

Aa/Ax
104a
(at.% H, D) -~

Aa/AT
a

Rel.

(lO-~/K)
13
13

Sm (cont~)

0.3

295
98 a

5.365
5.344 (a)
5.362 (c)

Gd

298

5.3022

773

5.3305

295

5.3025

4.4

14

95

5.2978

4.4

14

295

5.2990

5.0

14

95

5.2935

5.0

14

295

5.2960

-0.65

5.3

14

95

5.2905

-0.7

5.3

295

5.2926

95

5.2845

3.8

14

295

5.2485

90

5.2430

0.09

295
90

5.2433
5.2375

7.6
3.4
7.15

14
14
14

7.15

14

0.24

295
90

5.2380
5.2308

0.10
0.19
0.25
Tb

Dy

14

-0.7
-0.85

4
4

14
14

8.0

14

3.8

14

8.0

298

5.2060

773

5.2365

295

5.2056

5.6

14

90

5.1996

0.10

295

5.2020

5.6
6.6

14
14

90

5.1950

6.6

14

0.19

295
90

5.2000
5.1931

7.7

14

295
90

5.1988
5A911

5.8
14

14
14

0.27
Ho
Er

8.3
16

0
0

17

-0.5
-0.7

4.9

14
15

295
298

5.165
5.1290

7.8

773

5.158

15.5

7.8

Tm

298
773

5.0925
5.122

Yb

0.55

295

5.192

4
16

0,27
0,44

500
517

5.253
5.192

16
17

0.70

295

5.178

23.5

-1.8

continued on next page

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

225

Table 5, continued
R

x (H(D)/R ratio)

Yb (cont'd)

0.5
0.6

473-573
295
295
473-573
298
773
298
773

0D

a ()

295
295

0 <x < 0.5

0.5D
0.6D
0 <x < 0.5D
Lu

T (K)

Aa/Ax

Aa/AT

(at.% H, D) -I

( 10-6/K)

104a

Rel.

5.192
5.180

18
18

5.253
5.182
5.168
5.236

18
18
18
18
6.5
14.5
7.6
18

5.0338
5.062
5.022
5.054

cubic --~ tetragonal transformation below ~200-350 K.

4
4
4
4

b Corrected data.

References
(7) Boroeh and Kaldis (1989)
(8) Boroch and Kaldis (1987)
(9) Schefer et al. (1984)
(10) Mller et al. (1979)
(11) Daou et al. (1989b)
(12) Greis et al. (t981)

(1) Bashldn et al. (1978)

(2) Juekum (1980)

(3) Daou and Vajda (1992)
(4) Bonnet and Daou (1977)
(5) Klavins et al. (1984)
(6) Kaldis et al. (1987)

YH 2+x

(13) Zogal and UHritier (1991)

(14) Chiheb et al. (1993)

(15) Pebler and Wallaee (1962)
(16) Drulis et al. (1988a)
(17) Iwasieczko et al. (1992)
(18) Knappe and Mller (1986)

v
m

5.203

"X~ ,~~~a

5.208

5.207

5.206

295K

5.202

5.201

5.205

5.200

I
0.1

J
x ( a t . H / Y )

0.2

Fig. 9. Lattice parameter of YH2+x as a function of

x at room tempera~re and at 90 K (Daou and Vajda

1992).

t h e s y s t e m CeD2+x ( a n d l a t e r c o n f i r m e d f o r R D 2 + x w i t h R = L a a n d P r ) w h e r e s u p e r l a t t i c e
r e f l e c t i o n s h a v e b e e n f o u n d b y n e u t r o n s c a t t e r i n g ( T i t c o m b et al. 1972, K n o r r et al. 1978
a n d F e d o t o v et al. 1982), i n d i c a t i n g d i f f e r e n t o r d e r e d s t r u c t u r e s a s a f u n c t i o n o f x.

226

E VAJDA
I

La H2.8o

5.67

5.65
Z
o3
Z

o 5,65
W
I--

~5.61 QT,,~..~r. ~

5.59

x x I

I
I
I
I00
150
200
TEMPERATURE (K)

50

Cell parameters

I
250

Fig. 10. Temperature dependence of

the lattice parameters for LaHz80,
showing the cubic---*tetragonal transformation below 248 K (Klavins et al.
1984).

I
500

vs. temperature

CEH2.16

CUB., a

--~ 5.5750

TETR., a

1-~

TETR., c

5.5700

- 5.5650

-15.6600

5.5550

5.5500
I
110

130 150

170

190

210

230

250 270

290

310

I
330

350 370

J
Temperature

Fig. 11. Temperamre dependence of the lattice parameters for CeHz~6, showing the cubic--*tetragonal
transformation below 290 K (Boroch and Kaldis 1989).

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

a()

227

j'><-><

Gd H 2.x

5.300
X=0

5.295

Fig. 13. Ni3Mo-type structure (DO2z) of the octahedral Hsublattice in TbD2+x as determined by neutron diffraction by
Andr et al. (1992). Solid circles (small), Tb atoms; dashed
solid circles, H atoms on T sites; solid circles (large), H atoms
on O sites.

5.290

5.285

20

100

200 T(K)

I_

300

Fig. 12. Temperature dependence of the lattice parameters

for GdH2+x, for 0~<x~<0.25. The anomaly near 180K for
x = 0.22 and the strong non-linearity near room temperature
are indications for x-superlattice ordering (Chiheb et al.
1993).

In heavy ~-RH2+x systems, the first manifestations of a possible structural transformation (SRO and LRO) in the H sublattice were signalled as resistivity anomalies in
[3-TbH(D)2+x in the range 150-250K (Vajda et al. 1985, 1987a). These manifestations
were purity dependent, in the sense that hydrides from originally 3N-metal exhibited
an LRO H sublattice starting with higher x-concentrations than for those from 4Nmaterial, indicating hydrogen trapping by impurities. Detailed neutron scattering studies
on this system (Andr et al. 1992) have shown that the octahedral D-atoms were ordering
for 0.1 ~<x ~<0.25 in a DO22 structure (Ni3Mo-type) such that every occupied (420)plane was followed by three empty ones (fig. 13) corresponding to a stoichiometric
composition of TbD2.25. Similar ordering manifestations were observed in most other
~-RI-I2+x systems, and in table 6 we have listed the anomaly temperatures for typical
selected concentrations as determined from resistivity, specific heat or differential thermal
analysis (DTA) measurements. Figures 14-16 illustrate the situation by presenting partial
phase diagrams in the (T, x)-ranges of the ordered H-configurations. We note maxima
close to partieular compositions, e.g. CeHz5 or GdHz25, possibly corresponding to
stoichiometric compound formation and, in several cases, the presence of more than
one anomaly. The one at lower temperature is possibly due to a SRO-mardfestation.
Furthermore, one remarks that the ordering temperatures are roughly the same throughout
the R-series, which means that the H mobility for the same concentration x is independent
of R and the lanthanide contraction is only manifest as a diminishing solubility limit in the
~-phase, Xmax.A theoretical treatment of the phenomenon was undertaken by Ratishvili
and Vajda (1993) and Ratishvili et al. (1993, 1994) using the mean-field approximation
of the static-concentration-waves theory, who were able to reproduce parts of the phase
diagram in the TbH2x and in the CeH2+x systems.

la. VAJDA

228

Table 6
Anomaly temperatures due to H-ordering in ~5-RH(D)z+x-systems

La

x (H/R ratio)

T~ (K)

Pr

Nd

R
Nd

0.05

150

0.085

150
200

x (H/R ratio)

Tan (K)

Ref.

0.64

210

11

0.65

245
280

11

0.70

240

0.075

t80

12

0.10

200

0.6

197

0.7

239

0.16

180

12

0.26

240

12

0.08

155

13,14

0.14
0.17

155
185

13,14
13,14

0.16D

210

13,14

0.215

285

13,14

0.25

255

13,14

0.25D

260

13,14

0.305

200

13,14

0.045

150

15

0.10D

150

15

0.15

150

15

0.215

170

15

0.07

190

16

0.15

215

16

0.20

250

16

0.25

>300

16

0.055

155

17

0.10

150

17

0.13
0.18

175
170

17
17

0.22

260

17

0.27

210

17

0.065

150

18

0.11

18

0.75

241

0.76D

251

0.8

248

0.91D
1

251
241
270

5
6

211
230.5
233.5
274

0.I

215

1D

Ce

Rel.

0.2

250

0.25

300

0.34

340

0.6

275

0.65

240

0.7D a

250

0.75D a

317

0.86D a

292

0.81

240

0.08

170

10

0.23

>300

I0

0.53

>300

10

0.61

>300

10

0.65

253

3,10

0.76

250

10

0.06
0,1

185
200

11
11

0.19

265

11

0.32
0.44

>300
>300

11
11

0.56

165
295

11

Sm

Gd

Tb-3N

Tb-4N

Dy

Ho

Er

0.14

160
230
240

0.045

150

19

0.05D

19

0.088

150
175
180

0.09D

230

19

0.091

225

19

18

19

continued on next p a g e

HYDROGEN 1N RARE-EARTH METALS, INCLUDING RH2+x PHASES

229

Table 6, notes
a Corrected data.

References
(8) Abeln (1987)
(9) Libowitz (1972)
(10) Burger et al. (1988)
(I1) Daou et al. (1992)
(12) Vajda et aI. (1989b)
(13) Vajda et al. (1991a)
(14) Vajda and Daou (1993)

(1) Daou and Vajda (1992)

(2) Klavins et al. (1984)
(3) Bieganski and Drulis (1977)
(4) Chowdhury et al. (1989)
(5) Ito et al. (1982)
(6) Ito et al. (1983)
(7) Burger et al. (1991)

300

(15) Daou et al. (1986a)

(16) Vajda et al. (1987a)
(17) Vajda and Daou (1992a)
(18) Daou and Vajda (1994)
(19) Vajda and Daou (1994)

\
o

280
cub.

260

l*

Iv

~_

240

220

/o

180

0.5

I ~o

cu b!
200

ffI

0.6

0.7

0.8

0.9

X (D. H/La)

1.0

Fig. 14. Partia! phase diagram of LaH(D)2+x, for

0.5 <~x ~< 1, constructed from specific-heat and Xray diffraction data. Solid line, deuterium; dashdotted line, hydrogen; dash double-dotted line,
phase boundary between the compositions LaD2.50
and LAD2.75; dotted line, phase boundary between
LaDz75 and LaD3 (Kai et al. 1989).

0.2

Ce H 2,x

T(K

400

0.4
I

0.6
'

0.8 x

300
p,,

200

~I

,o,ra0o,

201

100 -I
0

I
J

(a)!

0.5 x( at.H / Ce)

100

i
0

I
O.2

(b)
I

0.4

0.6
0.8
x(at.H/at.Nd)

Fig. 15. (a) Partial phase diagram of CeH(D)2+x, for 0 ~<x ~<0.96, constructed from resistivity (solid circles)
(Burger et al. 1991), differential thermal analysis (solid squares), and neutron data (erosses) (Abeln 1987).
(b) Partial phase diagram of NdH2+x, for 0 ~<x ~<0.65, constructed from resistivity data (Daou et al. 1992).

230

P. VAJDA
300 0

0.1

0.2

'

'

0.3 x
'

(a)

A 25O
200
150
I

0.1

0.2

0.3
x(at.H(D)/at.Gd)

I
300

I3

(b)

/oI--~-

200

p..

100

IJ
5

10

300

'

15

20

x (at%)

25

'

(c)

D y H 2+x
1-

2OO

B'SROi I

100
0

0.1

r
0.2
0.3
x(at.H/at.Dy)

Fig. 16. (a) Partial phase diagram of GdH2+x (open circles) and GdD2+x (solid circles) for 0 ~<x ~<0.3, constructed
from resistivity data (Vajda et al. 1991a). (b) Partial phase diagram of TbD2+x constructed from neutron scattering
data (Andr et al. 1992), indicating the ordered DO22 structure in the low-T [3-phase for x > 0 . 1 . (c) partial
phase diagram in DyH2+~, for 0 ~<x ~<0.3, constructed from resistivity data (Vajda and Daou 1993).

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

231

Table 7
Lattice parameters of the orthorhombic dihydrides of Eu and Yb at 295 K (from Haschke and Clark 1975)

a ()

b ()

c (A)

Eu
Yb

6.254
5.889

3.806
3.576

7.212
6.789

3.2.1. The hydrides of Eu and Yb

The pure dihydrides of the divalent metals Eu and Yb are orthorhombic and their lattice
parameters are given in table 7. In the case of Yb, however, the metal ion in the higher
hydrides undergoes a valency change (cf. sect. 5) permitting the reappearance of the fcc
B-phase (see table 5).
Table 8
Lattice parameters of the h.c.p, trihydrides ~/-RHz+x at 295 K
R
Sc
Y
Nd
Sm
Gd
Tb
Dy
Ho
Er
Tm
Lu

x (H/R ratio)
0.65 a
1.00
0.82
0.82
0.84D
0.91
1
>0.9
>0.9
>0.9
>0.9
>0.9
>0.9

a ()
3.380
3.672
3.849
3.7870
3.7726
3.73
3.76
3.700
3.671
3.642
3.621
3.599
3.558

a = ax/~
5.854
6.360
6.667
6.559
6.534
6.46
6.51
6.409
6.359
6.308
6.272
6.234
6.163

c ()
6.135
6.657
6.893
6.7926
6.7632
6.71
6.705
6.658
6.615
6.560
6.526
6.489
6.443

Refl
1
2
3
4
4
5
6
7
7
7
7
7
7

a Under a pressure of 10 kbar.

References
(1) Bashkin et al. (1978)
(2) Forker et al. (1994)
(3) Bashkin et al. (1977)

(4) Greis et al. (1981)

(5) Sturdy and Mulford (t956)

(6) Zogal (1987)

(7) Pebler and Wallace (1962)

3.2.2. y-Phase trihydrides

Since the original experiments of Pebler and Wallace (1962) very little dditionat work
on the structural properties of the hexagonal trihydrides has been performed. Mansmann
and Wallace (1964) have established that the deuterium atoms in HoD3 occupied three
different non-equivalent sites, attributing its structure to the space group P 3cl (D4d). This
led to a reassessment of its lattice parameters as well as ofthose for the other R-trihydrides
by introducing a larger basal-plane constant, an = a . As to newly discovered systems,
there are the trihydrides of Sc produced under high pressure (Bashkin et al. 1978) and
those o f N d (Bashkin et al. 1977), which we added to the older values listed in table 8.

232

R VAJDA

4. Kinetics and thermodynamics

In this section, we shall present the available data conceming the dynamical properties
of hydrogen in its various ordered and disordered configurations, both in the (x- and in
the B-phase. This applies to different activation energies (or enthalpies) and to more
specific vibrational properties such as hydrogen local modes, Debye temperatures OD
of the metal-host lattice, and Einstein temperatures OE of the H sublattice. The data
are taken from nuclear magnetic resonance (NMR) experiments, inelastic relaxation
(internal friction and Gorsky effect), inelastic and quasi-elastic neutron scattering (INS
and QENS, respectively), H-"defect" recovery after quenching or electron irradiation, and,
generally from thermal dependeneies of various physical properties (e.g. resistivity, lattice
parameters, specific heat) measured at elevated temperatures.
4.1. a-Phase sod solutions
In table 9 we give the activation energies for diffusion Ed, and the corresponding
relaxation times r0, considering ~ = r0 exp(Ed/kBT), mainly for the three systems which
have a low-temperature solid solution *-phase with R = Sc, Y, and Lu. All techniques
consistently yield Ed-values in the range 0.5-0.6 eV for the four measured systems. The
relaxation times exhibit relatively large fluctuations, but show a tendency, at least in
the case of ct-ScHx, of increasing r0 with increasing concentration x. The latter seems
reasonable considering the correlation between the attempt jump frequency to an adjacent
site, f 0 = r 1, and the probability x, that such a site is occupied, f 0 e( (1 - x ) .
The H(D)-concentration dependence of the internal-friction amplitude Q-l, in the
ct*-RH(D)x systems with R=Sc, Y, and Lu (Vajda et al. 1990b, 1991c, 1983a) is very
close to linear; this raised some questions as to the origin of the observed peaks, since
a H - H pair interaction could be expected to yield a square x-dependence. The problem
was solved by Cordero (1993) who calculated the relaxation strength of tetrahedral H H pairs, taking into account blocking effects and the formation of complexes; he found
a roughly linear x-dependence for concentrations c > 0.01.
Low-temperature results of experiments with positive muons (Gygax et al. 1993) have
given an activation energy of 0.054 eV in SCH0.0, which is an order of magnitude lower
than the Ed'S of table 9, but seems to correspond nicely to the pair-binding energies Eb,
obtained from resistivity studies (=0.05 eV, cf. table 4 in sect. 3) and f-rom neutron
scattering (=0.06 eV, Berk et al. 1991), as weil as to the energy for fast localized motion
observed below room temperature in NMR (=0.052eV, Lichty et al. 1989). The same
mechanism could be responsible for the very low-T peak (< 1.4 K) in the elastic energy
loss in YH0.1, attributed to tunneling between T sites (Cannelli et al. 1991, Leisure
et al. 1993a).
Quenching across the ordering temperature in the *-phase (cf. sect. 3) or energetic
electron irradiation at low temperature creates disorder in the ordered configurations,
i.e. hydrogen "defects" which, upon heating, migrate and recombine in a recovery stage
centered at a peak temperature Tp. Applying a chemical rate equation to the recovery

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+~ PHASES

233

Table 9
Diffusion energies and relaxation times of H in ct-RH(D)~
R

Se

0.057
0.11
0.27
0.1
0.2
0.2D
0.3
0.20
0.20
0.10
0.14
0.20
0.20D
0.05~0.09
0.20
0.20D
0.15
0.05
0.05D

Gd
Lu

E (eV)

r o (10-14s)

0.54(1)
0.54
0.505
0.61 (3)
0.57(3)
0.58(3)
0.54(3)
0.51 (2)
0.57(3)
0.57(3)
0.62(3)
0.56(5)
0.55(3)
0.52(2)
0.56(5)
0.64(3)
0.52(2)
0.57(2)
0.63(2)

Rel.

0.7
1.2
3.3
0.07
0.7
1.0
3.6
9.1
0,3
9
0,6
10
50
6.1
10
0.37
20

1
1
1
2
2
2
2
3
4
5
5
5
5
6
7
7
8
9
9

References
(7) Vajda et al. (1983a)
(8) Torgeson et al. (1989)
(9) Vlkl et al. (1987)

(4) Anderson et al. (1989)

(5) Vajda et al. (1991e)
(6) Herbst and Kronmller (1981)

(1) Han et al. (1987)

(2) Vajda et al. (1990b)
(3) Lichty et al. (1987)

Table 10
"Defect" migration energies, Ere, at a peak temperature, Tp, after a quench or electron irradiation, in -RH(D)x
Em (er)
Sc
Y
Er
Lu

To ~ )

0.27(2)
0.20(2)
0.26(3)
0.26(2)

0.30(1)
0.27(2)
0.28(3)
0.27(2)

Ref.
D

152
163
172
164

160.5
169
177
170

1
2
3
4

Refernces
(1) Daou et al. (1986a)
(2) Vajda et al. (1987c)

(3) from Daou et al. (1981d)

(4) Daou et al. (1986c)

p r o c e s s ( a f t e r d e t e r m i n a t i o n o f its r e a c t i o n order, 7 ~ 1), o n e c a n d e t e r m i n e t h e m i g r a t i o n

e n e r g y E m o f t h e r e c o v e r i n g d e f e c t s p e c i e s , g i v e n i n t a b l e 10. O n e n o t e s t h a t E m ,,~Ea/2,

234

P. VAJDA
Table 11
Enthalpies of solution in the heavy lanthanides for ~t-RH(D)2-systems

H s (eV)
H

Gd
Tb
Dy
Ho
Er
Tm
Lu

Rel.
D

0.19
0.15(1)
0.16(1)
0.14(1)
0.05-0.08
0.037(5)
0.025(5)

0.15(1)
0.15(1)
0.125(1)

Beaudry and Spedding

Daou et al. (1981a)
Daou et al. (1981a)
Daou et al. (1981a)
Beaudry and Spedding
Beaudry and Spedding
Beaudry and Spedding

(1975)

(1975)
(1975)
(1975)

Table 12
Hydrogen local mode energies in ct-RH(D)x
B(meV)
[Ic
Sc

Ce f.c.c.
Tm
Lu

0.19D
0.33D
0.05
0.16
0.25
0.34
0.10-0.18
0.03-0.06
0.18D
0.04
0.06
0.06D
0.19D
0.19

77
78
101
103
103
103.5, 106
99, 102.5
96
73, 77
96(1)
I00 (97, 101)
75
68, 76
102, 105

Ref.
.Le

107
148
147.5
147.5
134
96
134
103
100
143.7

References
(1) Blaschko et al. (1989b)
(2) Udovicet al. (1990)
(3) Udovicet al. (1992)

(4) Anderson et al. (1986, 1988)

(5) Wegeneret al. (1980)
(6) Benningtonet al. (1991)

(7) Pleschiutschniget al. (1991)

(8) Blaschkoet al. (1988)
(9) Udovicet al. (1993)

indicating that Em refers to the s a m e diffusion process as Ed, the latter containing in
addition E r ~ E m , an energy for (thermal) " d e f e c t " formation, a parameter w h i c h has
already b e e n disposed o f in the q u e n e h or irradiation proeess. The Tp-values c o r r e s p o n d
to the a n o m a l y temperatures Tan f r o m table 4, e m p h a s i z i n g that the s a m e o r d e r - d i s o r d e r
process in the H sublattice is at the origin o f both manifestations.

H Y D R O G E N IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

I

O
r,D

c - a x i s mode

x=0.18

* c~-ScHo.a4
**

+
+ ,~'

~-~ ' ~ x = O . O 3

+ ,

(8K)
*%,

eoooe

't~ "+'~~ ~,~o.o~

c ~ - Y H x (80K)

~ / ~ .~~~o

.o

235

+
+
**

* ,

a-LuHo19

,:. ~ '. +++++;2:;

+
m

I
90

I
100

L
110

I
120

Energy Loss (meV)

Fig. 17. Inelastic energy loss spectra of ct-YH x for
various x, obtained at 80 K by neutron scattering.
Note the mode splitting for higher x, indicative of
H - H pair interaction (Udovic et al. 1993).

nU nn
i
90

a-YHo,la

(8K)

=,=m=

I
100

I
110

I
120

Energy Loss (meV)

Fig. 18. Inelastic energy loss spectra along the c-axis at
low temperature for c-ScH0.34, -LuH0.~9 and -YHats,
indicating progressive mode splitting with increasing
lattice constant of the system (Udovic et al. 1993).

The heat of solution, Hs, of hydrogen in the heavy lanthanides has been determined
for the system RH(D)0.2 from equilibrium measurements in the a ~ ~ reaction and is
given in table 11. The strongly decreasing tendency of Hs with increasing Z of the
rare earth correlates with the decreasing atomic volume and with the inereasing melting
temperature of R.
INS and QENS inelastic loss spectra have been measured to determine the local mode
energies of H and/or D in a few c~-phase systems (table 12). They generally exhibit a
soft vibration parallel to the c-axis and a 25-30% harder mode perpendicular to it. Highresolution spectra of the e-axis modes reveal a temperature- and concentration-dependent
splitting (fig. 17) representing the dynamical coupling of the H-H pairs in the SRO chain
structure of the *-phase systems (cf. figs. 6, 7). A comparison of the three systems
measured shows a sharpening of these bands in the order of increasing atomic volume
(Se ~ Lu ~ Y), implying an increasing extent of the chain ordering (fig. 18).
The Debye temperatures OD in these solid solutions increase, as expeeted, with
increasing H coneentration, such as roughly linearly in the ease of c~-LuH(D)x:
OD(a-LuHx) = (157 + 75x) K in the interval 0 ~<x ~<0.133, as determined from energy-

236

E VAJDA
280
260

'

~ LuD'm

'~

~240

200

,~~~____~~I
I-w

uJ 200

,~

___i.~

/ ///~/

160

18 0

!6C
0

,~
140

50

I00

150

200

250

300

TEMPERATURE (K}
Fig. 19. Debye temperature of Lu (dotted line) and of LuDIs3 (solid line) as a function of temperature. Note
the shifted ordinate scales (Vajda et al. 1986).

dispersive X-ray diffraction on a single crystal (Metzger et al. 1985) and from heatcapacity measurements (Vajda et al. 1986, fig. 19). A ~gD determination of a-LuHx with
0 ~<x ~<0.18 from low-T specific-heat measurements (Thome et al. 1978) yielded a sharp
maximum at 203.8 K for x = 0.015, attributed by the authors to H tunneling (ef. also sect. 5
for the electronic specific-heat eoefficient ~'eh and the above diseussion of T-T tunneling
observed in inelastie relaxation measurements). Ultrasound spectroscopy measurements
by Leisure et al. (1993b) yielded a general increase of all elastic constants upon addition of
H into scandium" Ob(SeD0.18) was determined to 371 K as eompared to Ob(Sc)= 355 K,
at room temperature. Finally, in the case of a-ErH(D)x, Cp measurements have given the
same OD = 192-t-5 K for all x in the interval 0 ~<x ~<0.03 (Schmitzer et al. 1987).
4.2. il-Phase dihydrides
In table 13, we present for four RH(D)2+x systems the diffusion energies and relaxation
times of hydrogen atoms thermally excited out of the oceupied T sites in the pure
dihydrides or of those occupying oetahedral sites in superstoiehiometric compounds. In
LaH2+x, which remains the only system where extensive E determinations as a function
of the excess-hydrogen eoneentration x have been made (Schreiber and Cotts 1963), one
notes a remarkable deerease of Ea with inereasing x, implying a lowering of the potential
barrier for O-O and O-T jumps. The reason for this is not quite clear but might be
related to the increasing ionicity, i.e. inereasing Coulomb repulsion between the x-atoms.
We have also included the results of a ~t+-diffusion experiment, for completeness.
The "defect" migration eharacteristies following a queneh across the temperature
region where possible H sublattiee ordering oeeurs are given in table 14. The wealth
of data collected during the recent years in resistivity studies by the group at Orsay
has permitted the establishment of two general phenomena. First, Em seems to follow

237

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

Table 13
Diffusion energies and relaxafion times in [3-RH(D)2+x
R

2+x

Sc
Y

1.98
1.98
1.98
1.92-2.09
2.00
2.15
2.36
2.27
2.47
2.61
2.75
2.85
3D
2.03

La

Lu

E a (eV)

r o (s)

0.35(3)
0.435(5)
0.34(9)
0.39
0.415(4 )
0.49(5)
0.40(4)
0.35
0.30(4)
0.22(3)
0.185(5) a
0.14(2)
0.20
0.33(3)~).38(4)

Ref.

8 10-~
2x 10-12

1
2
3
4
5
5
5
2
5
5
6
5
7
1

3 10-11
3.5 10-14
2 10-13
1.310 -12
7 10-12
4.5 X 10-12
3 10-1"
4.3 10-12
1.5 10-~2
10TM to 5 10-~

a W_diffusion

References
(4) Sttdar et al. (1993)
(5) Schreiber and Cotts (1963)

(1) Zogal et al. (1984)

(2) Phua et al. (1983)
(3) Bamfather et al. (1989)

(6) Chowdhuryet al. (1989)

(7) Borsa et al. (1982)

the same x-dependence as Ed (cf. table 13 for LaH2+x), confirming the same potentialbarrier mechanism for the two processes. Second, there exists, above a certain x-value
(whieh depends on the rare earth under study), more than one "defeet"-recovery process,
centered at different temperatures Tp and governed by different Ere. Figure 20 gives as an

15

o,~

: 10
:=L

0.3

2"
<~

t,

~~~,

o.2

"

?
0.1

100

150

200

~1/

T(K)

140

180

I I

220

Fig. 20. Annealing of the quenched-in resistivity Apq, with its derivative on the right hand side, of GdH(D)2+x
with x=0.25, showing an isotope effect on the peak temperamre near 160K and a second recovery process
near 200K (Vajda et al. 1991a).

P. VAJDA

238

"Defect" migration energies,

Ere,

Table 14
at a peak temperature, Tp, after a quench, in [3-RH(D)~+x

E m (eV)

Tp (K)

0.095
0.1
0.10
0.20
0.25
0.08
0.23
0.53
0.61 a
0.10
0.19 a
0.24
0.32
0.60"
0.64 a
0.075
0.16
0.08
0.12
0.16D
0.17
0.25 a

0.35(2)
0.31(2)
0.245
0.17
0.14
0.19
0.175
0.13

172.5
180
183
177
180
175
151
130
131
182
176.5
167
156.5
140
148.5
170
153
144.5
146
174
172
156

Ce

Pr

Nd

Sm
Gd

0.11
0.10
0.12
0.14
0.24
0.24
0.17(1)
0.13(1)
0.125(2 )
0.13(2)
0.12(2)
0.15(2)
0.09(1)

Rel.

1
1
2
2
2
3
3
3
3
4
4
4
4
4
4
5
5
6
6
6
6
6

Gd (ct'd)
Tb 3N

Tb 4N
Dy

Ho

Er

0.25D"
0.10D
0.15
0.15
0.20
0.055
0.10
0.13
0.18 a
0.22 a
0.065
0.11 a
0.14
0.048
0.045D
0.054 a
0.05D a
0.09
0.09D

E m (eV)
0.095(1 )
0.28(2)
0.24(2)
0.15(2)
0.4(2)
0.16(2)
0.16(2)
0.145(1 )
0.20
0.14
0.15
0.27
0.25
0.18
0.16
0.18
0.31
0.17
0.30
0.29
0.28

Tp (K)
161
150
146
176
173
148
146
167
165
155
146.5
183
183
141
148.5
141
177
148
177.5
177
178

Rel.
6
7
7
8
8
9
9
9
9
9
10
10
10
11
11
11
11
11
11
11
11

a Complex recovery,

References
(5) Vajda et al. (1989b)
(6) Vajda et al. (1991a)
(7) Daou et al. (1986a)
(8) Vajda et al. (1987a)

(1) Danu and Vajda (1992)

(2) Burger et al. (1991)
(3) Burger et al. (1988)
(4) Daou et al. (1992)

(9) Vajda and Daou (1992a)

(10) Daou and Vajda (1994)
(11) Vajda and Daou (1994)

0.2

O.1

170

150
130
I

0.1

0.2
0.3
x(at.H(D)/at.Gd)

Fig. 21. Activation energies E a, of the recovery

process for the quenched-in resistivity in GdH2+x
(open symbols) and GdD2+~ (solid symbols) and
the peak temperatures Tp, where this process takes
place (Vajda et al. 1991a).

HYDROGENIN RARE-EARTHMETALS,INCLUDINGRH2+~ PHASES

239

example the annealing curves of quenched GdH(D)I.%+0.2, exhibiting two recovery stages
at Tpl = 156K (161 K for the deuterated specimen) and at Tp2 --- 195 K, respectively. This
behaviour is most probably due to the recombination of defects created in different ordered
configurations, of SRO-type for the former and of LRO-type for the latter. Figure 21 shows
the x-dependence of the temperature for the principal recovery stage Tpl , and of its Em
for all measured GdH(D)2+x specimens, indicating the appearance of a new process near
x = 0.15, which corresponds nicely to the suggested phase diagram shown in fig. 16a. The
same phenomenon is also clearly visible in the systems HoH2+x and ErH2+x (table 14),
with simple recovery for low and for high x-values occurring at a low and at a high Tp,
respectively, but with complex recovery containing both processes at intermediate x.
Table 15 presents the enthalpies for dihydride formation Hf; thse for solution of
hydrogen Hs in the ~-phase, obtained from phase-equilibria smdies; and the defectformation energies Ef, as determined from the thermal expansion coefficients. Breaks
in the x-dependencies of Hs are manifestations of structural transformations or of metalinsulator transitions or both (figs. 22, 23).
e~
-1-

-I O 0

-I

iii IIIOlll

O0

'

'

o o q o o o o o

ZE
"~
i

o absorption

desorption

-50

e4 - 2 0 0
-1-

r"o"

ooO

._c
u~ -1 O0
"w
0.3

oO e ~ .
oo

-~
.~~

U3

desorption
I

1.0
ratio

000000
o

absorption
desorption

o coo

O0 0 0 0 0

-I O 0

0.5
atom

oo

desorption

-150

o absorption
l

o absorption

-50

~o

-150

"I"~
i

e4 - 2 0 0

O e 00

e.,
13

OoO O

O0

.E

H/Pr

Fig. 22. Variation of the solution enthalpy and entropy

of H in the PrH2+x system, for 0.55 ~<x ~< 0.97 (Ohki
et al. 1989).

0.3

0.5
atom

.0
ratio

H/Nd

Fig. 23. Variation of the solution enthalpy and entropy

of H in the NdH2. x system, for 0.32 ~<x~< 0.95. The
break near x = 0.65 indicates a phase transformation
(cf. fig. 15b) (Ohki et al. 1989).

In table 16 the Debye temperatures OD and the Grneisen temperatures On are given,
as characteristics of the acoustic dihydride lattice vibrations. On is generally obtained
from specific-heat measurements, while On results from the fit of a Grneisen law to
low-T resistivity measurements (e.g. Daou et al. 1984a). The simultaneously obtained
electron-phonon coupling constants will be given in the next section (sect. 5, table 19).
An empirical formula connecting the Debye temperatures of a hydride and its metal

240

R VAJDA

Table 15
Enthalpies for dihydride formation, Hf; for H-solution in I]-RH2:systems, H~; for the transformation
RH 2 ~ RH3, Htr; and activation energies for defect formation in the dihydride, E r
R

2+x

-H (eV)

- H s (eV)

Htr (eV)

Ef (eV)

Rel.

Sc

1.04

1.18

La

1.08

2.2-2.6
Ce

2
2.2-2.5

0.51-0.39

Nd

0.435-0.39

0.42-0.36

0.39
0.38

1.08

2
2.3-2.45

1.05

2
2

0.96
1.02

>2.25
2
2

Tm

2
>2.10
2
2
2

7
3,7

0.36

2,7

0.38 b

0.35

7,8

0.405-0.34

0.40; 0.42

1
2,6,8
0.49

9
1,3

0.48

0.50

8
4,3

1.11

1.06(1)

0.53
0.48

0.40

1.17

1.06(1)

0.60

4,3

0.66

8
4,3

0.43 s
1.00 s

>2.05
Yb
Lu

2,5

0.385-0.32

>2.15
Er

0.30
0.44

>2.25
Ho

0.42 a

0.35

>2.3

Dy

0.48

2.4-2.6

2.3-2.65

Tb

2,3
4
2,5

1.11

2.4-2.6

Eu
Gd

0.25

0.37-0.26

2.55-2.75
2

Sm

0.41

1.00(1)

2.4-2.9
Pr

0.56 s
0.94
0.96(I)

8
0.75

2.05

0.54

1
4,3
8

a See fig. 22.

b See fig. 23.

References
(1) Libowitz and Maeland (1979)
(2) Messer and Park (1972)
(3) Bonnet and Daou (1977)

(4) Boeva et al. (1986)

(5) Viallard and Daou (1972)
(6) Ohki et al. (1984)

(7) Ohki et al. (1989)

(8) Mintz et al. (1976)
(9) Toguchi et al. (1982)

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

241

Table 16
Debye, O D, and Grneisen, O, temperatures of RH(D)z+x hydrides
R

2+x

O D (K)

Se

1.88
1.99

759

710

O G (K)
465

340
537

2
La

1.9-3.0
(1.91-3.0)D

1
2
3

1.81
1.93

Rel.

1
2

525

319-382

341-379

Pr
Eu

2.65
1.85 b

Tb

1.93

240

Ho

1.86

240

Er
Yb

1.81
1.80 ~
1.90 b

381
239"
243 a

2
6
6

2.37
2.41

350 ~
352 a

8
8

Lu

356 a
170

1.82
2

5
6

240
361

7
7

1
9

b orthorhombic.

a corrected data.

Referenees
(1)
(2)
(3)
(4)
(5)

Daou et al. (1984a)

Beattie 1972)
from Kai et al. (1989)
Kai et al. (1989)
Drulis and Bieganski 1979)

host, OD(MH) and

Jacob et al. (1981):

(6)
(7)
(8)
(9)

Drulis and Stalinski (1989)

Lucasson et al. (1985)
Drulls et al. (1988b)
Thome et al. (1978)

OD(M), with the formation enthalpy HF, has been proposed by

OD(MH) ]2
[ OD(M) J = 3.7HF (eV/atom) 0"87.
Finally, neutron-scattering expeilments on several RH(D)2+x-systems revealed loealized
vibrational modes ascribed to H vibrations on O and on T sites, which are summarized
in table 17. A third (higher-energy) peak is sometimes observed in superstoichiometric
samples and attributed to the lowered symmetry of a tetrahedral site (T*) in the
neighborhood of an x-atom. The role of SRO and LRO in the Ho-sublattice upon the
evolution of the local modes (both their concentration- and temperature dependence) in
a seiles of RH(D)2+x hydrides (R = Y, La, Ce, Tb, Dy) was demonstrated recently by

P. VAJDA

242

Table 17
Hydrogen local mode energies in [5-RH(D)2+x
2+x
O
Sc

E v (meV)
T

OE0~)

1.88

1.85D
Y

La

2D
1.81
1.80D
1.94
2.18
2.28
2.99

Ce

2.98
2.98D
1.98
2.01D
2.12D

Pr

Nd
Eu
Gd
Tb
Dy
Ho

Er

43(2)
75
54

2.72D
2.96

-45
63(2)

2.97D
1.94
1.98
1.7D
2.3D
2D

47(2)

2.5D
2.06
1.85
2D
1.89
1.85D
1.93
1.93D
1.83
1.93D
1.86
1.86D
1.98
1.88
1.76D

1570
1110

1475
1043

2
2
1
1

127
90

73(2)
72
65
61(2)

78(2)
52
55
52
67; 82
77; 103
60; 72

103
109(3)
100
120
127(4)
93(3)
106
75.5
75

3
4
5
5
4

127

4
6
6
7
6
4
4

83; 95

128(4)
92(3)
108
120(3)
76

Ref.

T*

3
4
3
3
8
8
4
12
12

86

76
75

96

69
123(3)

80-85

129 a
91"
1392
984
1415
1000
1438
1017
1460
1032
126
1485
1050

9
9
10
10
9
9
I0
10
3
9
9

continued on next p a g e

243

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+~ PHASES

Table 17, continued
R

2 +x

E,, (meV)
T

Tm
Yb

Ref.

1.96

1508

1,90D

1066

2
1,98

Lu

oB (K)
T*

89

130

89; 115

131; 138

11
12

1.82

1531

1.82D

1083

Average of three infrared lines.

References
(1) Daou et al. (1984a)
(2) Rush et al. (1966)
(3) Hunt and Ross 1976)
(4) Parshin et al. (1978)

(5) Kamitakaharaand Crawford(1982)

(6) Vorderwischet al. (1980)
(7) Glinkaetal. (1977)
(8) Knorr et al. (1978)

(9) Burger et at. (1984)

(10) Lucassonetal. (1985)
(11) Maeland (1970)
(12) Drulis (1993)

Udovic et al. (1995). We have added the Einstein temperatures OE, characteristic for
the optical vibrations in the pure dihydride lattice, corresponding to hcoT= kBOE, which
has been determined (Daou et al. 1984a) from an analysis of the high-T resistivities of
RH2 and RD2 in the harmonic approximation, i.e. under the assumption of OE(H) =
BOE(D). They increase with decreasing lattice parameter, i.e. with decreasing M-H
distance, resulting in higher optieal mode energies for the heavier lanthanides and for Sc.

5. Electronic properties
The controversy concerning the electronic state of hydrogen in rare earths between the
protonic and the aninic model has been settled in favour of an elaborate form of the latter,
namely the "hydridic" model proposed by Switendick (1970-1972) in the early 1970s
and developed later by Gupta (Gupta 1978, Gupta and Burger 1980), Kulikov (1985) and
others. In this model, confirmed by numerous experiments, especially during the past two
decades, hydrogen forms within the metal bonding and antibonding R-H and H-H (for
RHy with y >~2) bonds located below the d-band of the host. A direct consequence is
the decrease of the carrier density upon hydrogenation, and a modification of the Fermi
surface, where the Fermi level Er is pushed downwards. This results, for example, in an
apparent diminution of the Fermi-surface anisotropy as observed in resistivity studies on
single crystals of c~-LuDx (Vajda et al. 1986) and in changes of its nesting features in the
case of non-stoichiometrie [~-phase hydrides as suggested from magnetic manifestations
in ~-GdH2+x (Vajda et al. 1991a). This will be treated in detail in the next section. The
decrease of the conduction-electron density is responsible for a general decrease of all
related processes such as s-d scattering, electron-phonon coupling and, in particular, the
magnetic interactions mediated through the Ruderman-Kittel-Kasuya-Yosida (RKKY)
electron polarization.

244

P. VAJDA

The eleetrie resistivity p, which is the principal subject ofthis seetion, evolves in several
stages. First, p inereases when dissolving H in the a-phase, -RHx, where the hydrogen
atoms aet as impurity scattering eenters for the eonduetion eleetrons; it deereases in the
dihydride, ~-RH2, reaehing lower values than in the pure metal, mainly due to reduced
eleetron-phonon eoupling; then the resistivity increases again upon adding of excess (x)
H atoms in solution on octahedral sites, [3-RH2+x;fmally, the hydrides end up as insulators,
going through a metal-semieonduetor (M-S) transition somewhere between RH2 and RH3
(in the B-phase for La to Pr, in the y-phase for all heavier lanthanides and Y):
R ~ a-RHx + [3-RH2(_o) --~ ~-RH2+x --- RH3(_~).
In the following we shall present detailed diseussions of the various specifie cand B-phase systems, treating - in addition to the resistivity data - heat-eapaeity
measurements and, when available, results from optieal and photoeleetron speetroscopy.
Specific phenomen such as superconductivity and the Kondo effects will be treated, while
a speeial subseetion shall be devoted to the field of the M-S transitions where reeently a
new impetus has been given.
51 Solid solutions
The effeet of dissolving hydrogen atoms as scattering centers for the conduction electrons
can be measured either in the metastable low-T a*-phase or in the high-T solid-solution
a-phase. Thus, the specifie resistivity of an added H atom is measured near 0 K as a
residual-resistivity increase Aprresin the former ease, and at some higher temperature in
the latter. The condensation of hydrogen in the ordered H-H ehain eonfiguration of the
a*-phase at 150-180K (cf. seet. 3) does not seem to influenee Ap~ s signifieantly (it is
praetically the same at 4.2 K and at room temperature, cf. figs. 24, 25, 28), though a slight
decrease of the eleetronie charge density was observed in NMR-measurements in a-ScHx
and a-YHx upon H ordering (Liehty et al. 1989). In table 18 we present the values for
Ap~ s in the six existing a*-phase systems, both at zero eoneentrations (i.e. determined
from the initial slope of the experimental p(x)-eurves) and at maximum eoneentrations,
the latter being usually smaller due to a saturation effect. The few investigations of a
possible isotope effeet on Ap~ s were not eonclusive in the case ofa*-YH(D)x (Vajda et al.
1987e) but showed a deerease with inereasing isotope mass in the ease of a*-LuH(D, T)x,
possibly implying a deereasing vibration amplitude of the interstitial hydrogen isotope
(Daou et al. 1986e, Jung and Lsser 1988)
In figs. 24-28 we show several isothermals, p(x), of different RH(D)x-systems, which
are useful for the determination of the solubility llmlt Xmax (ef. table 1), sinee the lower
resistivity of the ~-phase (see next subseetion) leads to a decrease ofAp~ s when exeeeding
the a-phase limit into the two-phase (a+ [3)-region; we also list these Xma
x values in
table 18.
The eleetronic speeific-heat coeffieient Yel is extraeted from an extrapolation to zero
of the low-temperature linear part in the ep/T=f(T) plot. It usually inereases with

245

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+~ PHASES

Table 18
Residual resistivity increase per at.%H, Ap~s,nfor a-RH(D, T)~
R

Sc

Tm

0.30
0.35
0.19
0.20
0.03
0.03
0.06
0.11

Lu

0.205

Y
Ho
Er

~pH~(0) a (~tff2crrdat.%H)

Xmax

Ap,o~(Xm=)H
b (p~ff2cm/at.%H)

3.6

2.3
2.6

4.2
3.8
3.6

3.4; 3.1; 3.2(D)

2.4
3.3

Rel.
1
2
3
4,5
6
3,7

3.51Ic

3.111c; 3.0Hb

4.2

3.25

3,9

9.011c; 3.911b

6.0llc; 3.911b

10

3.75
3.3; 3.1(D)

3.15
3.05; 2.95(D)

3,11,12
13,14

2.30")
2.15(D)11c; 3.75(D)llb

15
16

0.1839
a Determined from the initial slope of the ,o(x)-dependence.
b Determined at the maximum concentration in the c-phase.

References
(1) Jensen and Zalesky (1980)
(2) Daou et al. (1986b)
(3) Beaudry and Spedding 1975)
(4) Vajda et al. (1987c)
(5) Juckum (1980)
(6) Daou et al. (1987)

(7) Daou et al. (1981d)

(8) Vajda et al. (1987b)
(9) Daou et al. (1981c)
(10) Vajda et al. (1989c)
(11) Daou and Bonnet (1974)

(12) Daou et al. (1976)

(13) Daou et al. (1986c)
(14) Daou et al. (1981e)
(15) Jung and Lsser (1988)
(16) Vajda et al. (1986)

15(:

lOO
E
0

c
Z

50
42K

0.1

0.2

0.3

0.4

0.5

x=at,H(D)/at.Sc

Fig. 24. Resistivity isothermals for ScHx (open symbols) and ScDx (solid symbols) at 4.2 K and 0C as a
function of H(D)-concentration x, indicating maximum solubility for Xm~=0.35(2)atM(D)/Sc (Daou et al.
1986b).

246

E VAJDA

+
120
80

50

1oo

~o ~

4 40

'~

6O

'0
Q

I
0.05

I
0.10

I
0.15

I
0.20

I
0.25
x.at.H(D)/at.Y

Fig. 25. Resistivity isothermals for YHx (open symbols) and YDx (solid symbols) at 4.2 K and 0C as a function
of x, indicating maximum solubility at Xmax= 0.20(1)at.H(D)/Y (Vajda et al. 1987c).
15

15
140

.~," ~ ~ ~ - ~' ~

'~

-~"~

.=

10
E
o

c- t

lO
c~

5
t
1

o:os
x(at.H/at.

I/

120/
0

~-~*~+

I
.01

1
.02

I
.03

I
.04

I
.05

se"
E r)

Fig. 27. Residual resistivity of ErH~ crystals

]Ic as a function of x, indicating maximum
solubility for Xmx= 0.065(5) at.H/Er (Vajda et al.
1987b).

x/at.H(D)/at.Ho

Fig. 26. Nel temperature TN, and residual resistivity of HoH~ (open symbols) and HoDx (solid symbols) as a
function of x, indicating maximum solubility at Xm~x= 0.03 at.H(D)/Ho (Daou et al. 1987).

increasing x, sometimes passing through a maximum, such as in the case of ct*-LuHx,

where it rises sharply from 7el=8.35mJ/gatK 2 at x = 0 to 11.34 for x=0.015 and
then decreases continuously to 6.96 at the c-phase limit (fig. 29); Thome et al. (1978)
and Stierman and Gschneidner (1984) attribute the sharp peak to hydrogen tunneling.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

P42K et PSOK
mu. Qcm

9C

P296K~
in u.Qcm J120

8C

7(

~o

247

t 1]0
J

" ~ "296K

-~.~~

4100

~o

30

60

20

50

10

1,0

I
0.1

I
0.2

I
ot.H/otLu

Fig. 28. Resistivity isothermals tor LuH x at 4.2 K, 90 K and 296 K as a fimction of x, indicatJng maximum
so]ubi]ity for x=a* = 0.20(])at.H/Lu (Daou et a]. ] 976).

12

T~

il
I0

2
Qt%H

I.[xiO-6

Q9

8
g-

7
\

~
0

10

20

30

40

1
50

I
60

%
t
70

ATOMIC PERCENTHYDROGEN

80

Fig. 29. Variation of the electronic

specific-heat coefficient in the Lu-H system as a function of the Hconcentration. The inset shows the variation of the room temperature magnetic
susceptibility. Note the abscissa scale,
with c n = H / ( L u + H ) (Thome et al.
1978).

248

E VAJDA

The determination of Yel in magnetically ordering systems is more delicate because

of the superposition of a magnetic term in Cp, which has to be subtracted. The only
two existing experiments on tx-ErHx and ct-TmHx gave, for the former: )'el= 10.413.0mJ/gatK 2 for 0 ~<x ~<0.03 at.H/Er (Schmitzer et al. 1987), and for the latter (which
is uncertain): )'el = 10.5-17.5 mJ/g atK 2 in the interval 0 ~<x ~<0.03 at.H/Tm (Daou et al.
1988e).
5.2. (Mainly) t~-phase hydrides
The mobility (and sometimes also the number) of the charge carriers in the dihydrides
depends strongly on the possible presence of ordered and/or disordered electronic
and/or atomic configurations. The electronic configurations concern the magnetically
ordering dihydrides and will be treated in detail in sect. 6. The atomic configurations
occur in the superstoichiometric RH2+x systems where the octahedral excess hydrogen
atoms x exhibit a tendency to order in SRO or LRO structures for concentrations
exceeding a certain R-dependent value Xcrit, somewhere in the range 0.05 ~<Xent~<0.15
(for a discussion see sect. 3). The ordering transitions occur at various temperatures
between 150 and 350 K, depending on the system and on the concentration x, and exert
an increasing influence on the conduction-electron scattering as measured through the
electric resistivity. Therefore, in table 19 the electric resistivities have been listed for
judieiously selected concentrations in all measured RH2+x systems, at three different
temperatures: pres, obtained, where feasible, by extrapolation to zero K, in order to
eliminate as much as possible the influence of spin-disorder scattering; Pls0n, showing
the role of the phonon contribution to scattering; P295K, adding the part due to structural
disorder of the hydrogen superlattice (where existing), which has already broken down
in most cases at room temperature. Systems with concentrations close to the B-phase
limit somefimes undergo temperature-dependent M-S transitions, which are marked by
an asterisk in table 19, and shall be discussed in more detail in a special subsection
below.
In table 20 the electronic specific-heat coefficients )'el of the measured RH2+x systems
are presented. The high values in the cases of R = Ce and Yb are indicative of heavyfermion behaviour in these valence-fluctuating compounds (see below). Otherwise, the
hydrides have rauch lower yel-values eompared to the et-phase systems (el. sect. 5.1),
which is mainly a consequenee of the strongly decreased density of stares at the Fermi
surface N(Ev). In the same table the acoustic and optic electron-phonon coupling
parameters, A).ae and A).opt, are given, obtained from an analysis of resistivity data (see
sect. 4). The coupling to the optical (hydrogen) phonons is found to be much weaker
than that to the acoustic (host-metal) phonons, due to much smaller partial densities
of states in the former~case. In addition, the authors (Daou et al. 1984a) observed
a strong reduction of the acoustic coupling parameter when going from the metal to
the dihydride, A~ac(R)A~ac(RH2), confirming the relation between N(EF) and )'el
mentioned above.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

249

Table 19
Residual resistivities, p~~, and the resistivities at 150K and at 295K, for ~-RH2+x systems
R
Sc
Y

2+x

Ce

Rel,

0.49

3.0

7.7

0.42

2.9

8.5

1.97

0.38

4.7

12.2

2,3
2,3

0.21

4.9

12.8

1.97 + 0.04

11.9

17.7

26.5

2,3

1.98 + 0.085

22.3

28.4

37.5

2,3

1.97+0.10 a

143.7

2.00

1.37

2.00D

0.50

141
8.7
8.4

1551, 175T

2,3

19.0

19.5

2.70 a

190

700

2.80 a

350

3000

2.90 a
2.93

2500
5x 107

9 x 104
5 105

5
5

2.00

0.8

19.2

26.4

2.02

5.5

21.5

29.2

8.0
27.8

23.3
44.0

31.5
59.2

6
6

122.5

2.25

16.0

46.5

2.40

120

365

735

2.65

980

2500

5850

2.70 a

1100

5000

2.78 a
2.81

800
lxl07

1000
8104

5
7

~109

3 10

2.85
1.97

-4.0

17.6

1.98+0.08
1.97 + 0.23

30.7
17.4

41.5
46.2

59
145

8,9
8,9

1.98 + 0.53

22

8,9

1.96+0.61

83.8

1.97 + 0.76
Nd

P295K (~t~cm)

1.85D

2.02D
2.10

Pr

PlSOK (Ixf~cm)

1.88

1.97D

La

p~~ (gQcm)

112

29.0

8,9

87

400

210

620

8,9

460

860

8,9

2.00

0.18

12,8

23.1

2.00D

0.32

12.5

22.9

10
10

2.06

29.65

43,0

59.5

10
10

2.19

23.4

47.3

115.5

2.32

30.0

60.4

147

10

2.44
2.56

39.3
47.8

70
121

160
420

10
10

465

900

2300

10

1900

3400

6600

10

2.63
2.65 a

continued on hext p a g e

250

E VAJDA

Table 19, continued

R
Sm

Gd

2+x
1.98

Ho

29.5

39.0

11

42.8

55.7

11

1.98 + 0.075

25.0

45

58.5

11

2.00+0.16

61.6

76

96.5

11

1.96+ 0.26

86.7

117

1.99

35.5

44.05

11
12,13

0.4

35.6

45.0

12,13

14.65

43.5

52.3

13

1.98+0.14
1.98+0.25

72
51.3

70.0
88.4

81.4
138.2

13
13

498

14

139

415

1.96

0.5

24.9

33.3

1.91D

0.8

25.3

34.1

14,15,16

1.97 + 0.07

29.3

32.3

39.0

14,15,16

14,15,16

1.96+0.15

46.5

48.5

67.5

14,15,16

1.97+0.20

50.2

57.4

92.1

14,15,16

2.00+0.245

23.8

85

1.99

0.45

155

16.6

24.5

14,15,16
17

1.93D

0.5

16.7

25.4

12

1.97 + 0.02

5.45

20.8

29.65

18

1.98+0.10

38.8

48.8

59.8

18

1.98 + 0.18
1.99 + 0.22

52.75
43.4

62.6
63.7

77.2
117.

18
18

1.99 + 0.27

47.4

61.3

93

18

2.00

0.4

11.5

19.5

17

1.86D

2.1

13.3

23.4

19

28

40.0

20

23.5

74

20

1.98+0.12 a

190

171

1991, 204T

20

57

60.5

1.99+0.14 a

455

355

3701, 380]"

20

1.95

<0.5

7.9

15.0

1.76D

< 1.5

9.4

17.6

12

12
29.5
167

20.5
36
165

28.8
47
1971, 223]"

21
21
21

360

295

3201, 340T

21

1.96 + 0.03
1.985 + 0.07
(1.98 + 0.09)D a
1.985 +0.091 a

Yb

0.20

177

1.97+0.04

1.98 + 0.08

Tm

Rel.

8.8

1.99 + 0.02

Er

P295K (~tff2cm)

16.7

1.97+0.305 a

Dy

PlSOK (~tff2cm)

2.00+0.03

1.85D

Tb

Pres (~tQem)

1.96
1.90D
1.90
2.33
2.62

0.19
0.19
5 x 10s
>109

9.15
9.25
102
5103

17

17.2
18.2
1013

12
12
7

5
1

22
22

eontinued on next p a g e

25t

H Y D R O G E N IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

Table 19, continued
R

2+x

Lu

Pres (p~~'2cm)

P,50K (~t~'2cm)

P~951~ (gff~cm)

Rel.

1.98

0.10

4.6

11.0

23

1.93D

0.12

4.7

12.0

23

(1.99+ 0.03)D
(1.99 + 0.06)D

3.45
50

18.8

41.7

280

612

23
23

1.98+0.15

112

649

t335

23

1.97 + 0.20

502

2165

3640

23

a M - S transition present.

References
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)

Daou et al. (1984a)

Daou and Vajda (1992)
Vajda and Daou (1991)
Vajda et al. (1991b)
Shinar et al. (1988, 1989)
Burger et al. (1991)
Libowitz (1972)
Burger et al. (1988)

(9) Daou et at. (1989a)

(10) Daou et al. (1992)
(11) Vajda et al. (1989b)
(12) Burger et al. (1984)
(13) Vajda et al. (1991a)
(14) Vajda and Daou (1993)
(15) Daou et al. (1986a)
(16) Vajda et al. (1987a)

(17)
(18)
(19)
(20)
(21)
(22)
(23)

Daou et al. (1988b)

Vajda and Daou (1992a)
Lucasson et al. (1985)
Daou and Vajda (1994)
Vajda and Daou (1994)
Srnirnov et al. (1992)
Daou et al. (1988c)

5.2.1. Metal-semiconductor transitions

The first unambiguous observation of a temperature-dependent M-S transition was made
on CeH2+x crystals for 0.7<~x~<0.8 by Libowitz and eoworkers (Libowitz and Pack
1969, Libowitz 1972). This transition was inverse to the classical behavior in that
the specimens were metallic at low temperature (below a TM-S ~ 250K) and turned
semiconducting above it. The authors related this electronic transition to a structural
cubic--+ tetragonal transformation in the Ce lattice upon cooling. The experiments were
recently repeated using a contact-free high-frequency technique (Shinar et al. 1988, 1989)
eonfirming Libowitz' data on CeH2x and yielding new results on the system LaH2+x
(fig. 30). The authors interpreted their data in the framework of a variable-range-hopping
model (Shinar et al. 1990), reformulating the original proposal by Libowitz (1972) in
particular. At low temperature, the conductivity is governed by a delocalized band at the
Fermi energy associated with the superlattice which is formed by interacting vacancies
on octahedral sites Vo. At higher temperature (and at lower Vo-concentrations, i.e.
at higher x), this band breaks down leading to loealized states and loss of metallic
conduction. The resulting effective energy gap varies between 36 and 280 meV in the
range 0.80 ~<x ~<0.90. LaH2.93 is already a pure semiconductor over the entire measured
temperature range, with an intrinsic gap of 480meV, in the same way as CeH2.81 and
CEH2.85 (Libowitz 1972).
Recently, M-S transitions with strong hysteresis effects have been observed by the
group at Orsay in the heavy rare-earth dihydrides GdH2+x, HoH2+x, ErH2+x and in YH2+x,
with a surprisingly low superstoiehiometry, of the order of x ~ 0.1-0.3 (Vajda and Daou
1991, 1994, Daou and Vajda 1994, see e.g. fig. 31). Again, it is clear that the MS transitions are driven by the collapse of a superlattice but, this time, the structural

252

P. VAJDA

Table 20
Electronic specific-heat coefficient, V,j, and acoustic and optic electron-phonon coupling parameters, A~,a~ and
A~,opt, for RH(D)2x compounds

References
(1) Daou et al. (1984a)
(2) From Kai et al. (1989)
(3) Kai et al. (1989)
(4) Abeln (1987)

(5) Sehlapbach et al. (1986)

(6) Drulis and Bieganski (1979)
(7) Drulis and Stalinski (1989)
(8) Burger et al. (1984)

(9) Lueasson et al. (1985)

(I0) Drulis et al. (1988b)
(11) Thome et al. (1978)

o r d e r i n g c o n c e r n s t h e direct o c t a h e d r a l H o l a t t i c e o f t h e x - a t o m s a n d not t h e i r V o v a c a n c i e s , s u c h as i n t h e s u b s t o i c h i o m e t r i c t r i h y d r i d e s , d e s c r i b e d a b o v e . T h e f a c t t h a t

HYDROGEN 1N RARE-EARTH METALS, INCLUDING RH2+~ PHASES

253

Fig. 30. Normalized resistivity pn(T)=,o(T)/p(295K) of various LaHx specimerls as a function of reciprocal temperature, showing the metal-semiconductor
transition (Shinar et al. 1990).

Fig. 31. Temperature dependence of the electric resistivity of three YH2+~ specimens, in the relaxed (R) and
in the quenched (Q) state, with x = 0.085, 0.095 and 0.10. Lower inset: enlarged view of the low-temperature
minimum in the case of x = 0.085. Upper inset: a fit of the low-temperature part of the resistivity for x = 0.10
by an expression of the type p(T) =po(T/To) 1/2exp(T/To)-v4, indicating a Mott mechanism as origin for the
S-M transition (Vajda and Daou 1992b).

the transitions occur in these dihydrides at x-values close to the 13/(~+ y) phase boundary,
is an indication that two factors are simultaneously required in order for the transitions to
take place: (1) the existence of an order-disorder transformation in the x-sublattice and
(2) a significant decrease of the carrier concentration, due e.g. to the presence of nuclei
of insulating y-phase microdomains. This transition is manifest by a diverging resistivity
(fig. 32, Daou and Vajda 1992). Furthermore, the higher the sample resistivity, the lower
the TM-S (i.e. the weaker the Ho-superlattice); or, in other words, specimens with lower
cartier density exhibit lower band stability. This rule seems to be valid for all investigated

254

R VAJDA

0.05

0.1

I
i

150

p
5.208

5.203

,
~ l ~ j

El 295K

5.207

5.202

-'-'-"-"~'-P

5,206
=

Q.

5.201

100

5,200

I
0.1 x (~tt.H / y )

50~
0
0

5.205

a,S \
I
0,2

P i ~
~~"POK
I
I
0.05
0.1
x(at.%H/at.%Y)

Fig. 32. Resistivityisothermals at 0 K and at 295 K as a function ofx in the YH2+x system.Note the diverging
resistivity when approachingx=0.1. The inset gives the lattice parameteras a function ofx for T=90K and
295 K; the break at x = 0. I 0 indicates the ~-phase limit (Vajda and Daou 1992b).

systems, including LaH2+x and CeH2+x, covering three orders of magnitude in resistivity
(Vajda 1995).
Finally, and most interestingly, these compounds exhibit an additional S - M transition
(upon heating) at low temperatures, fig. 31 (Vajda and Daou 1992b), which has not
been observed in the light RU2+x systems. The p-minimum at TS-M occurs at 30120K in the various hydrides; its depth and position increase with increasing sample
resistivity, with TS-M and TM-S approaching each other. This S-M transition might
therefore be related to weak localization of the carriers due to scattering on atomic
disorder of the excess H atoms. A fit of the experimental points to an expression of the
type p(T)c((T/To) 1/2 exp(T/To) -1/4, where the prefactor eould be due to multiphonon
scattering, is not unreasonable (note the inset concerning YH21 in fig. 31). The
temperatures TS-M, TM-Sl and TM-sT (taken in cooling and heating runs, respectively)
as weil as the eorresponding resistivities PM-S~ and PM-sT are collected in table 21 for

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2x PHASES

255

Tabie 21
Temperatures for the various metal-semiconductor transifions, Ts_M and TM_S, and the resistivities PM-S at
TM_S, for RH(D)2+x systems
R

2 +x

La

Ce

Nd
Gd
Ho

Er

Ts_M (K)

TM-S I (K)

TM_s ~ (K)

PM-S I

PM-S T

(~t~cm)

(~f2cm)

Ref.

1.97+0.095

37

n.o. a

283(1)

n.o.

84

1,2

1.97+0.10

79

235(1)

256(1)

170

185

1,2

(1.96 + 0.095)D

n.m. b

n.o.

315(2)

n.o.

60

(1.98 +0.10)D

n.m.

248(1)

275(1)

120

138

n.m.

n.m.

260(10)

n.m.

850

4,5

2.70

2.80

n.m.

n.m.

230(10)

n.m.

5000

4,5

2.84

n.m.

n.m.

220(5)

n.m.

3 104

4,5

2.90

n.m.

n.m.

260(3)

n.m.

6 x 105

4,5

2.70

n.m.

n.m.

260(10)

n.m.

5500

2.78
2.715

n.m.
n.m.

n.m.
245(5)

230(10)
n.m.

n.m.
1104

2400
n.m.

4
6

2.74

n.m.

200(10)

n.m.

1.7 x 104

n.m.

2.77

n.m.

220(5)

n.m.

2.3 104

n.m.

2.00 + 0.65
1.97+0.305

n.o.
105

290(3)
260(1)

.n.o.
265(1)

6500
484

n.o.
501

7
8

1.99+0.106

40.5

264(1)

273(1)

106

108

1.99+0.11

55

264(1)

285(1)

126

129

1.98+0.12

100

255.(1)

260.5(1 )

210

212

1.99+0.14

132.5

252(1)

255(1)

425

428

9
10

1.985+0.088

32

242(1)

260(1)

160

170

1.985+0.091

120

237(1)

245(1)

383

395

262(2)

290(1)

20.5

(1.955 + 0.09)D

79

244(1)

260(1)

223

239

10

(1.97+ 0.09)D

39

246(1)

257(1)

264

275

10

n.o., not observed.

61.5

92.5

10

(1.98 + 0.08)D

10

b n . m , not measured.

References
(1) Vajda and Daou (1991)
(2) Daou and Vajda (1992)
(3) Vajda and Daou, unpubl.
(4) Shinar et al. (1988)

(5) Shinar et al. (1990)

(6) Libowitz (1972)
(7) Daou et al. (1992)

(8) Vajda and Daou (1993)

(9) Daou and Vajda (1994)
(10) Vajda and Daou (1994)

all of the measured systems. (Also, note the remarks concerning surface M-S transitins
at the end of sect. 5.2.3 and in the diseussion of CeH2+x in seet. 5.2.2.)
5.2.2. Discussion of specific hydrides
The formation of the hydrogen superlattice in several LaD2+x compounds was clearly
seen in the form of specific-heat peaks at 250-270 K (fig. 33) and also as breaks in the x-

256

R
i

VAJDA
I

30

X=2.65

E 20

2,75

/"

2.80

<5-~

2.90

?
x=2.55

2.78

--

10

1 O0
T

(K)

200

300

Fig. 33. Heat capacities of four LaDx

specimens as a function of T, indicating
a transition near 250K (Kai et al.
1989).

dependence of the electronic specific-heat coefficient ~/el, at concentrations corresponding

to particular ordered configurations: x = 0.25, 0.5, and 0.75, by Kai et al. (1989). The small
but definitely non-zero value for Yel in LaH(D)3.0 (see table 20) indicates the presenee
of some charge carriers and is interpreted by the authors as due to a semi-metallic band
overlapping the top of the valence band and the bottom of the conduetion band. In an
EPR experiment detecting an Er3+-ion in LaH2+x, Drulis and Hoffmann (1985) have
determined mainly a cubic Er environment in the case o f x ~ 0, but only Er sites of axial
and orthorhombic symmetry for x > 0. The observation of superconductivity (for the first
time in a rare-earth hydride) was reported in LaHzl0, with a Te = 0 . 1 7 K (from Kai et al.
1989); this small value is due to a small electron-phonon coupling term and is consistent
with the electronic structure calculations of Gupta and Burger (1980), who predicted a
superconducting transition, if any, at a T e ~ 0 K. On the other hand, measurements in the
two-phase system LaH2_x, in the range 0 ~<x ~<2 between La and LaH2, have indicated
percolating superconductivity of the La-metal, with a Tc = 6 K (corresponding to that of
the fcc [5-La, stable above 300C), up to concentrations of LaH1.75 and traces even in
LaH2 (Vajda et al. 1991b).
With its broad R-phase range (extending up to x = I) and its hybridizing f-eleetron,
CeH2+x is a fascinating system. It was here that Libowitz and Pack (1969) grew
the first hydride single crystals, in which the first M-S transitions were observed in
resistivity experiments (Libowitz and Pack 1969, Libowitz 1972) for 0.7 ~<x ~<0.8, see
sect. 5.2.1 above. Shinar and coworkers, who later repeated Libowitz' experiments, give
a nice illustration of the very x-sensitive behavior of CeH2+x specimens in fig. 34,
where two samples exhibit a M - S transition near 250 K while the third one is a pure
semiconductor with a gap of A E = 2 1 0 m e V below 240K and 100meV above. Zamir

HYDROGEN 1N RARE-EARTH METALS, INCLUDING RH2+x PHASES

257

T(K)
500

500

200

160
xX~
x
x
x

x
xx

x CeH z.81

10.0

CeHz.82

xx

x
GeH 2.78

.
1.o
%

%
oO

O.li

IO00/T(1/T)

Fig. 34. Normalized resistivity ph(T) =_ p(T)/

p(295 K) as a hmction of reciprocal temperature for three CeH2+x specimens, showing M-S
transltlons and pure semiconductor behaviour.
The actual H content of the CeH2.81 sample
was probably somewhat higher than indicated
(Shinar et al. 1989).

et al. (1984) observed concentration-dependent M-S transitions in NMR-experiments

for x > 0.65 attributed to a Mott mechanism. Schlapbach et al. (1986) reported peaks
near the Fermi surface in photo-eleetron spectra below 70K for x=0.7, the latter
interpreted as due to low-temperature removal of H atoms from the surface to the bulk
and consequently pushing back of d-states towards EF to form a metallic band: a typical
surface phenomenon. The complicated magnetic phase diagram due to the interplay
between antiferromagnetic (AF) and ferromagnetic (FM) interactions will be treated in
sect. 6, but here we have to mention the observation of an incoherent Kondo effect for
0 ~<x ~<0.34 transforming into a coherent lattice below T ~ 20K (fig. 35, Vajda et al.
1990a, Burger et al. 1991). It is encouraging that the low-temperature behavior (for
T ~<2.5K, cf. inset in fig. 35) follows a T2-1aw, which is expected according to the
heavy-fermion model, as postulated from the high value for the specific-heat 7eI (see
table 20). At the same time, the quenehed CeH210 specimen (see inset in fig. 35) exhibits
a flatter T2-dependence than the slowly cooled one, indicating an additional spin-wave
contribution in the magnetically ordered lattice. In general, the order~lisorder situation of
the octahedral (x) H atoms exerts a strong influence upon the electric resistivity, as seen
from the x-dependence of the residual resistivity (fig. 36): two plateau-like regions for
0.1 ~<x ~<0.3 and for x > 0.5 indicate the oecurrence of structural ordering of the x-atoms
below a temperature Ts given in the inset (cf. also sect. 3). The effeet of a quench from
room temperature across the ordering interval on p is seen in fig. 37 for three CeH2+x
specimens, showing its low efficiency for the two x-richer samples where the ordering
fange extends up to 300K and above; the recovery of the introduced Apq takes place at
the low-temperature end of the ordering interval emphasizing the beginning mobility of
the x-atoms (cf. seet. 4).

258

P.

i I ~ ll~l

I ~ llJ[

VAJDA

10

!
!
/
/

o
40

o
C

35

~ ~....,I

10'
A

E
o

c=

4-

Q.'- 3 0

40G

~~ \

25

~20C

20
1

i I l lltl

10

20

I J ~l~l

50

100

10(~

300

T(K)

Fig. 35. Resistivity of CeHz05 and CeHz10, after

subtraction of the phonon contribution, as a function
of log T in the relaxed (R, open symbols) and in the
quenched (Q, solid symbols) stare. The inset shows the
p = p r + A T 2 dependence at low temperature. The break
near 200K is due to CF-splitting (Vajda et al. 1990a).

2.5
t

1
2

2.2

2.4

H/Ce

2.6

Fig. 36. Residual resistivity of CeH2+x and the

structural ordering temperatures (inset) as a
fimetion ofx (Burger et al. 1991).

Praseodymium forms, like La and Ce, B-phase hydrides up to x = 1, and this permits
one to observe the approach towards a potential M - S transition. The resistivities o f three
x-rich PrH2+x specimens (Burger et al. 1988) shown in fig. 38 indicate a strongly hysteretic
structural transformation at 200 ~< T ~< 250 K for x = 0.76 which is probably a precursor
for a M - S transition. The p-isothermals seen in fig. 39 show an ordering plateau for
0.1 ~<x ~< 0.5 at T ~< 200 K and probably another one for x > 0.6.
The resistivity isothermals of NdH2x exhibited in fig. 40 show an ordering plateau
between x ~ 0 . 1 and 0.5 (just as PrH2+x) and a strong increase, up to 6000~tf2 cm, for
x = 0.65 (close to the ~/(~ + y) phase boundary) as the concentration approaches the M S transition (Daou et al. 1992).
In SmH2+x, where the phase boundary is at X~ax ~ 0 . 4 5 (Daou et al. 1989b), the
resistivity curves (fig. 41) show that SRO already begins below x = 0.1, and that a clear
LRO-stage exists for x = 0 . 2 6 (Vajda et al. 1989b).
The heat capacity of the semiconducting orthorhombic europium hydride was measured
by Drulis and Stalinski (1989) for EuHl.5 and by Drulis (1993) for EuD2, who observed
a k-type anomaly for the magnetic transition at Tc = 15.8K for the former and at
Tc = 16.3 K for the latter (see sect. 6). The T-dependence of Cp above ~ 5 0 K exhibits
a marked isotope effect which was analysed by Drulis (1993) on the basis of hydrogen
optical modes as determined from infrared absorption spectra.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

Bi

,oo

259

120
110

/j_T,~o

6050-

~4o

~~o!___~~_~_~~~
20
4
30
2O

50

40 -

90
80
70

~o

o_:i~~~~ ~

20
10

50

100

150

200

250

300 T(K)

Fig. 37. Thermal variation of the resistivity for three CeH2+x specimens in the relaxed (open symbols) and
quenehed (solid symbols) stare; Apq is the quenched-in resistivity inerease reeovering near 170 K (Burger et al.
1991).

The resistivity isothermals at zero K, Pres, and at room temperature/0295 K, of GdH2+x

(fig. 42) are an indication of the LRO starting in the x-sublattice near x = 0.15 at low
temperatures (Pres-maximum) and the continuous dissolution ofx-hydrogen atoms withut
ordering at 295 K (Vajda et al. 1991a). Fig. 43 shows the effect of quenching upon the
x-sublattice and the magnetic transitions (see sect. 6) for x=0.16D and 0.21 as well as
the M-S transition at 260 K for x = 0.305, at the [3-phase limit (Vajda and Daou 1993).
Figure 44 indicates, as in GdH2+x, the presence of LRO in the x-sublattice of TbH2+x
for x ~>0.15, an increasing ordering amplitude Apa~, and ordering temperature Ta~, with
increasing x, while a small SRO-anomaly is visible for x=0.07 near 180K (Vajda and
Daou 1993, Vajda et al. 1993). No M-S transition was observed in the sample with
x =0.245 up to room temperature despite the proximity of the [3-phase boundary and

260

E VAJDA
900

801-

,o+
+

800

20

700
0

0,20

040

0"60

080

x(at.H/at.Pr)

o+
+
+o

600
oO

+
o
+,
Bx=O'76

500

o.

400

..

o~

o OoO
,

o+
+ +

200

.'"

.
-o-

.o

~"
,

, ..

""

x=0"61

;~

o-

" .""
o
oo

IO0"Z

o,~,

o"

oO

oO

oo

o
o

,o

. ~oO

300

,a'
.o " "

."

o ~

:il/o~2 .... ::o.:...o.o.O.o x=o.~~

I

50

100

150

200

250

300

T(K)
Fig. 38, Thermal variation of the resistivity for three PrHz+x specimens in the relaxed (open symbols) and
quenched (solid symbols) stare; Apq is the quenched-in p inerease measured at 80 K and plotted as a function
of x in the inset. Note the hysteresis between 200 and 250 K for x = 0.76 (Burger et al. 1988).

the large Apan; a transition is expected above ~320K. The evolution of the magnetic
ordering will be treated in sect. 6.
The situation is similar in DyH2+x as for the other heavy lanthanide dihydrides (fig. 45).
LRO occurs above x ~ 0.1, with the maximum effect apparently being attained before
x=0.27 (Vajda and Daou 1992a). The latter alloy must already contain some y-phase
precipitates, since its resistivity and the ordering amplitude are smaller than those for
x=0.22 (el. also sect. 3). Ferreira et al. (1987) performed the first perturbed angular

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2x PHASES

800~

261

Pr H 2,x

700

OOK
600

500

.i

400

300

200

100~-

Or
O

I
.20

I
.40

///*~4.2K

I
.60

I
.80
x(at.H/at.tat

Fig. 39. Resistivity isothermals of PrH2+, at various temperatures as a function of x (Burger et al. t 988).

correlation (PAC) measurements on a rare-earth hydride with DyH1.83 and DyH2.95, in

the pure [~- and y-phase, respectively. Along with the other data obtained, the authors
estimated that the electronic contribution to the electric field gradient in the hcp y-phase
is about three times that of the lattice contribution.
HoH2x (together with ErH2+x) is another superstoiehiometrie dihydride system
exhibiting clear M-S transitions as observed in YH2+x (cf. fig. 32). They were measured
for coneentrations 0.106 ~<x ~<0.14, yielding effective gaps of a width AE ~ 25-45 meV,
immediately after the appearance of LRO in the x-sublattice (Daou and Vajda 1994).
This is seen in fig. 46, where the recovery of quench-induced disorder occurs in two

262
.^4
]u

P. VAJDA
0.1
I

0.2
J

0.3
]

0.4
I

0.5
I

180

0.6
J

Sm H2+ x

160

140

10 3

120

B,

110

~2

.Y~

90

2OOK

6C

wo

295 K

c~

~S-.,

.o

40

x~O

10
5

}'18 fi]r x = I
0
0.1
0.2

I
0.3

I
0.4

I
I
0~5 0.6 .7
x ( a t . H / at. N d )

510

110

150

k~

200

2;0

T/K 3 0 0

Fig. 41. Thermal variation of the resistivity for various

SmH2~ specimens in the relaxed (R) and quenched (Q)
state (Vajda et al. 1989b).

~1_ ~

1OO

x--o.~o~~

400

o.

300'
30O

50
00

2O

20

Fig. 40. Resistivity isothermals of NdH2+x

at various temperatures as a function of
x (Daou et al. 1992).

.o

0.1

0.2

0.3

x(at.H(D)/at.Gd)

Fig. 42. Resistivity isothermals at 0 K and at 295 K

of GdH(D)2x. The maximum of Pr near x=0.15
indieates low-temperature ordering (Vajda et al.
1991a).

110
IO0

200

!00 ~ " ~ ~ 6

D t 90

100I
100

54K
f0L l ~ I
0
100

Gd H 2.x

70

60
r
200

300 T(K)

Fig. 43. Resistivity of GdH(D)zx, for x=0.16D, 0.21 and 0.305 in the relaxed (R) and in the quenched (Q)
stare. Note the magnetlc transitions below ~100 K, the structural transformations around 200 K for all three
alloys and the M-S transition near 260K for x=0.305 (Vajda and Daou 1993).

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

140

/,

TbH

120

2,x

100

Dy

x--0.245

H2.x

80-

,~ ~

.~

x--0.27 120

"K

"-K
_

.e'Z:P'W-

x=O.15

40-

J
40

80

60

100 -

263

"~x=O.07

20

100

200

x=0,22

~ ~

1~~

300 T(K)

Fig. 44. Resistivity of TbHz+x for various x-values.

Note the magnetic transitions below -40 K and the
structural anomalies around 200 K. The inset shows
the presence of an anomaly already for x = 0.07 due
to SRO (Vajda and Daou 1993).

100

200

T(K) 300

50

Fig. 45. Resistivity of DyHzx for various x-values

in the relaxed (R) and quenched (Q) state. Note the
magnetic transitions below ~15 K (cf. fig. 64) and the
stmctural anomalies around 200 K (Vajda and Daou
1993).

stages: the first corresponding to that seen for x = 0.08 (SRO), and the second being
preponderant above x---0.1 (LRO). A similar situation has been described earlier in
sect. 4 in the annealing of GdH2+x with x = 0.25 (fig. 20). Interesting enough, the strongly
decreasing p in the semiconducting region above TM-S (~250 K) becomes smaller above
T ~ 300 K than the lowest values in the metallic region at TS-M (~130 K). This is also seen
in fig. 47, where the two isothermals p(0 K) and p(Tmax) cross each other near x = 0.13.
A similar situation is observed in ErH2+x (fig. 48), showing the transition from the
purely metallic behaviour at low x to the M-S and S-M transitions at higher x, as soon
as LRO becomes important (Vajda and Daou 1994). The effective gaps on the hightemperature side vary in the range 40 ~<AE ~<60meV, while the low-temperature part
shows localization behaviour down to the magnetically ordering domain (cf. sect. 6) where
the separation between the two phenomena becomes difficult. The isothermal p(x) at room
temperature (fig. 49) resembles that of YH2+x (fig. 32). The diverging part near x = 0.09
indicates the proximity of the ~/(~ + y) phase boundary.
The dihydride of Yb (like that of Eu) is non-metallic, since the Yb (and the Eu) ion is
found in its divalent form. For higher hydrides, YbH2+x with x ~ 0.4, fluctuating valencies
seem to prevail leading to heavy-fermion behaviour and high ye~-values (Drulis et al.

264

P. VAJDA
[

500

Hort 2+x

500 ~ ,

" ~

x=0.14

500

Hort 2+x

400i

400

400
A

E 300
o

132.5K

~[
<3.

100 ~

20O

2~4K

x=0.08

60

100

50

2C ~::~2T3K

o:

10 2 _ , ~ ~ * - ~ - * X "
0'

I
100

I
200

x=0
I
300

or

0.05

p (295K)
"T "
P (M-S)

(0K)
I

0.1
x(at.H/Ho)

0.15

T(K)
Fig. 46. Resistivity of Hort2x for various x-values
in the relaxed (R) and quenched (Q) stare. Note the
magnetic transitions below ~10K (cf. fig. 66), the
structural anomalies around 200K (hysteresis!) and
the M-S and S-M transitions for high x-values (Daou
and Vajda 1994).

Fig. 47. Resistivity isothermals at 0 K and at 295 K (of

at TM-S in the case of M-S transitions) for Hort2+ x.
Note the diverging p for x > 0 . 1 3 (Daou and Vajda
1994).

1988b and table 20). The latest work by Smirnov et al. (1992), who have measured
the thermal conduetivity te, the eleetric resistivity as well as the thermoelectric power
of YbH2+x in the range 0 ~<x ~<0.7, indicates that the transition from the orthorhombic
to the cubic phase (el. seet. 3.2.1) for x > 0.25 is accompanied by a valence change from
2 to 2.6. The authors attribute the thermal behaviour of t and o f p in these compounds
to an inhomogeneous mixed-valence system, aeting as a defeet-rich glass-like solid with
carrier hopping between ions of different valencies, rather than to heavy fermions. The
large ]tel would then be caused by the so-called non-magnetic atomic disorder (as analyzed
by Gschneidner et al. 1990) due to the x-hydrogens, which gives rise to random exchange
interactions between the Yb-ions and to spin-glass behavior.
For LuH2+x with 4N-pure starting material, the composition limit of the metallic ~phase is rather low (cf. also seet. 2). This results in a percolating conduetivity due to the
presence of the insulating y-phase as soon as x exceeds a few at.%, and at the same time
blocking the ordering of the x-atoms (Daou et al. 1988b). An analysis of the resistivity
data has given a large percolation faetor showing a tendency for the insulating phase to
coat the metallic one along the grain boundaries. A small localization mimimum is seen
for the samples with the highest p (but much lower in amplitude than in corresponding
YH2+x, ErH2+x or HoH2+x samples, see above) and no M-S transition at high temperature,

265

HYDROGEN 1N RARE-EARTHMETALS, INCLUDINGRH2+~ PHASES

100
k

200
Q

300

= ./,~-- 2,~SK

--400

ErH (D) 2+x

300

HT~
JD

E
o

C
,~~

100 ~

Q~"~~~

ErH,.
=~n2+x

x=0.03

30

200

100

~Y

20

X=0

10

I
0.05

i
I
0.1 X

Fig. 49. Resistivityisothermalsat 295 K for ErH(D)2+x,

taken with increasing and decreasing temperature,
Fig. 48. Resistivity of ErH2+ for various x-values showing diverging p for x > 0.09 (Vajda and Daou
in the relaxed (R) and quenched (Q) state. Note the 1994).
magnetic transitions at low temperature (el. fig. 67),
the structural anomalies around 200K (hysteresis!)
and the M-S and S-M transitions for high x-values
(Vajda and Daou 1994).

100

200

300 T(K

whatsoever, even in samples with a p ~ 3500 ~tf2 cm, stressing the necessary existence of
LRO in the x-sublattice as one of the conditions for its occurrence.
5.2.3. Spectroscopic studies
Experimental investigation of the electronic structure of dihydrides of Sc, Y and Lu using
optical spectroscopy (Weaver et al. 1979) has shown several specific features dominated
by intraband absorption at low energies (~< 1.5 eV) and by interband absorption above.
The structures have been related to specific bands; it was shown partieularly that dband absorption was broadened by increasing disorder due to the occupation of O sites.
Similarly, XPS-studies of the Y - H system (Fujimori and Schlapbach 1984) have revealed
a large chemical shift of the core levels going from Y to YH2, which indicates a charge
transfer from Y to H; the valence-band spectra of YI-I2 and YH3 showed strong H-induced
bands near 6 eV suggesting a mixing of Y 5sp states into the H ls band. A comparison
of Compton profiles taken for metallic Lu and for LuH2 (Lsser et al. 1979) yields
good agreement with Switendick's model, accommodating the two extra electrons in
antibonding ls hydrogen wavefunctions. Finally, a surface S-M transition was observed in
several hydrides of light and heavy rare earths at low temperatures using photoelectron
spectroscopy as shown in fig. 50 (Schlapbaeh et al. 1987); cf. also the discussion for
CeH2+x above.

266

P. V A J D A
I

I
I

I
I

r~

300 K

25 K

I
!

L
I

I
I
/

I
I

i/
iI

,/

/
t

)-

\.

iI~~

,/

iI

/^\

/I

I
!

' S
/!/',,

I,z

/
/
I

' \

I
/
/

I f

_CoH2,

,/ / ' i
_/

!
!
CeH2.2

"\ "'1

I
6

G.. dH~.o

4
2
O:EF 2
O:EF
-.,w- BINDtNG ENERGY (eV)

Fig. 50. Photoelectron spectra of the valence bands

of several RH2+x compounds taken at 25 K and at
3 0 0 K , s h o w i n g a sttrface S - M transition at low
temperature (Schlapbach et al. 1987).

6. Magnetic properties
The incompletely filled 4f-shell confers a magnetic moment to the lanthanide ions and
it is the Ruderman-Kittel-Kasuya-Yosida (RKKY) meehanism via polarization of the
conduetion electrons which is responsible for the oceurrence of long-range ordered
magnetic structures in the lanthanide metals. Details can be found in numerous reviews,
among others in vol. 1, ehs. 3, 6 and 7 (1978) ofthis Handbook series and a reeent one by
Jensen and Mackintosh (1991). The addition of hydrogen, both in solution or resulting in
hydride formation, leads to a decrease of the conduction electron density ne (of. sect. 5)

HYDROGENIN RARE-EARTHMETALS,1NCLUDINGRH2+x PHASES

267

and to a general diminution of the magnetic manifestations dependent on ne through the

RKKY-interaction. In fact, the observed decrease of the magnetic susceptibility and of the
magnetic ordering temperatures upon hydrogenation was one of the early arguments in
favour of an "anionic" versus a "protonic" mechanism for hydride formation. The eventual
ordering of hydrogen at low temperatures (cf. sect. 3), which presents new symmetry
situations such as the quasi-linear H-H chains along the c-axis in a*-phase systems or
the tetragonal sublattices of the excess H atoms on octahedral sites in superstoichiometric
dihydrides, can change the magnetism dramatically, leading sometimes to its complete
suppression and sometimes to new magnetic transitions (e.g. Vajda and Daou 1984, 1993,
Vajda et al. 1989a). The reason for this is a variation of the magnetic anisotropy and the
magnetoelastic interactions, as weil as a modification of the crystal field (CF) or, more
fundamentally, that of the Fermi surface.
Again, as for the electrical properties (sect. 5), the knowledge of the exact concentration
of the excess H atoms x, dissolved in the dihydrides RH2+x is essential for a correct
interpretation of the experimental data and for their comparison. We shall, therefore, stress
this point in the following by separating, when possible, the pure dihydride (cf. sect. 2)
from that of the excess atoms x. Below we have listed the relevant data on the magnetic
manifestations in R-H systems resulting from investigations of the electric resistivity,
specific heat, susceptibility, neutron diffraction, Mssbauer effect etc., concentrating on
the most recent experiments, performed mainly since the eighties; for earlier compilations,
the review by Arons (1982) may be usefully consulted. As before, we shall separate
the discussion into two subsections on ct- and [3-phases and discuss the specific systems
individually within each subsection.
6.1. ct*-Phase solid solutions
Among the magnetieally ordering lanthanide metals, only Ho, Er and Tm form a solid
solution (in the metastable a*-phase, cf. sect. 3) at temperatures low enough for the
occurrence ofmagnetic transitions. Hence, a particular effort toward the understanding of
the influence of hydrogen upon their magnetic properties has been made in the last 1015 years, mairdy by the group at Orsay. The various magnetic structures are qualitatively
the same for the pure metals on the one hand and for their solid solutions on the other. We
shall present, in the following, the quantitative evolution of their different characteristic
transition temperatures Tc, TN, etc., as weil as their paramagnetic Curie temperatures Op,
and the corresponding effective magnetic moments/~ePff, when available.
Holmium possesses two magnetic transitions: below TN = 133 K the magnetic structure
is a basal-plane helix, which transforms into a cone along the c-axis at Tc = 20 K. The
only studies of the magnetism in a*-HoHx were done by resistivity measurements of
polycrystals (Daou et al. 1987). These authors observed a regular decrease of TN with
increasing x by ATN/AX - 2 K/at.%H(D) up to Xamax= 0.03 at.H(D)/Ho (fig. 26 in sect. 5).
The constancy of TN for higher x-values is another indication for Xamaxto be the solubility
limit and the beginning of the two-phase (ct + [3)-region.

268

E VAJDA
0.05
I

0.10

at.H/at.Er

56

86

54

84

52

82

8 0 ~.z

~'=: 5 0
48

78

46

76
B

24

(a)

./

22
2O
18

I
0.05

I
0.1o

0.15

x(at.H/at.Er)

Fig. 51. Variation of the main transition temperatures Tc, TH and T N in relaxed (open symbols)
and quenehed (solid symbols) ErHx crystals I[c, as a
function of x. The break indicates a-phase samration
and the beginning precipitation of the 13-phase (Vajda
et al. 1987b).

Erbium exhibits a complex magnetic behaviour, with at least four observable characteristic temperatures: below TN=85K, the moments order in a sinusoidal c-axis
modulated (CAM) strue~are; at TH = 54 K, there appears a component perpendicular to
the c-axis resulting in a helieoidal struemre; this intermediate phase exhibits a sequence
of lock-in transitions of commensurate phases (spin-slip struetures) with wave vectors
4 T~,
5 ~6, ~1 and ~ as observed by X-ray scattering (Gibbs et al. 1986), the
rm = 2, 3 , T3,
latter loeking into a ferromagnetic cone along the c-axis at Tc = 19 K. Finally, Jensen and
Cowley (1992) found that these higher harmonics were distorted by two-ion couplings of
trigonal symmetry. Some of the intermediate spin-slip structures have also been observed
in resistivity and suseeptibility experiments, in particular the rm =~1 configuration at
Tel =27 K, and the infiuenee of hydrogen upon them has been studied. Thus, the two
FM transitions at Tc and Tel shift to higher temperatures by ~0.5 K/at.%H upon the
absorption of hydrogen, the two AF-transitions at TN and at TH to lower ones by
~1.5 K/at%H (Daou et al. 1981d, 1984b, Vajda et al. 1983b, 1987b, Ito et al. 1984, Burger
et al. 1986b, Schmitzer et al. 1987), see also fig. 51. An additional increase of Tc and
Tel (by ~2 K in the case of x = 0.06) oceurred in quenehed solid solutions, which was
particularly striking in c-axis crystals (Vajda et al. 1987b). This shows the influence of
structural order (H-H pairs oriented along the c-axis) on the magnetic one (c-axis cone)
and stressing the magnetoelastie origin of the latter). The same effect, though smaller,
is observed for TN and TH (eounteraeting the negative effect due to earrier removal)
and is caused by the ehange in magnetie anisotropy. The energy gap A in the magnon
speetrum, determined from an exponential fit of the magnetic contributions to resistivity
Pm oc exp(-A/kBT), (Vajda et al. 1987b) and to speeific heat, cm c T3exp(-A/kBT),
(Sehmitzer et al. 1987) deereased by ~0.5-1 K/at.%H from a value A/kB= 11-14 K in
pure Er. Magnetisation measurements on a*-ErHx single crystals (Burger et al. 1986b)
exhibited strong anisotropy effects on the critical field for the FM transition Her. Thus,

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+~ PHASES

269

TmHx
H//c

3
O

~= 20
0

%
10

O~ ~
0

10

I
20

I
30

I
40

50

Hlnt(kOe)
Fig. 52. Field dependence of the magnetisation of ct-TmI-Ix crysta[s [Ic at 4.3 K, showing the strong decrease
of the cfitical field for the ferri- to ferromagnetic transition with x (Daou et at. 1990).

Hcrlle decreased upon hydrogenation while Herz inereased with increasing x; the latter
was, in addition, anisotropic within the basal plane, Herllb > Hcrlla, for x > 0. Hcr generally
decreased with increasing x in the case of polycrystals, while the paramagnetic Op
dropped simultaneously from 40 to 31.5 K, and/tePfr remained roughly constant, 9.5 ~tB
(Vajda et al. 1983b).
Thulium orders in an AF CAM structure below TN = 57.5 K and mrns "ferrimagnefic"
(three spins up and four spins down) below a Tc=39.5 K. Contrary to Er, the addition
of H diminishes both transition temperatures: TN and Tc decrease by ~1K/at.%H
(Daou et al. 1980b, 1981b, 1983, 1988e, Ito et al. 1984, Vajda et al. 1989c), while
magnetisation measurements parallel to the c-axis of monocrystalline specimens gave a
2K/at.%H decrease in Tc (Daou et al. 1990), stressing the anisotropy of the magnetic
structure. This indicates that the process at Tc is governed by an electronic rather
than a magnetoelastic mechanism, the variation of the electronic structure probably
overwhelming the magnetoelastic interaction. At the same time, the critical field
for the ferri-to-ferromagnetic transition parallel to c, Her[ic, decreases strongly upon
hydrogenation (fig. 52). Furthermore, the anisotropy between c-axis and basal plane
favours the basal-plane magnetisation and, possibly, even the appearance of basal-plane
AF ordering (Daou et al. 1990). 6)0 decreases from 19.1 K to 12.2 K and #Pfr from 7.61 ptB
to 6.80~tB when going from pure Tm to -TmH0.x0 (Daou et al. 1981b). Bennington
et al. (1991) have noted, in inelastic neutron-seattering experiments, changes in the CF
induced by hydrogen ordering, in particular a dispersionless peak near 15 meV, as welt
as a coupling of magnetic excitations to H-vibration stares (cf. table 12 in sect. 4).
A surprising phenomenon was observed in resistivity experiments on single erystals of
ct-TmHx (Vajda et al. 1989c). While the introduetion of hydrogen essentially resulted in

270

E VAJDA

(~
TmH x//b

100

~o
o

70

!...

60
O.
50

"ot~..~~

...

.,"~. "

o o

;.f.~. ~

i........~-o

40

."

30

' ""

"

."

"

" ." " t " " "

~o iii;.. il"

42.6

."

~~ " " ~ " " . . . - "

,~ ~ ~ ~ ,

1.3J

/
50

701

I 2
100

" " ~ ~ " - 30.6

.....
I

I"

6I
150

"~

-31.0

.......

_t. .,.,. ,.....,. .o. , ' . ~

1.5~ .~........ . o...- ...

10
0 .... "
0

T(K)

o.

,oo
10.4

lOK
200

TmHx / / C

(b)

0.10

::k
O.

x=O
0

I
50

I
100

I
150

T(K)

I
200

Fig. 53. (a) Thermal dependence of the resistivity of ct-TmHx crystals [Ib; the arrows indicate TN and
the structural anomaly near 170K. The inset shows the low-temperature part demonstrating the growing
contribution of the magnetic excitations with x. (b) Same as (a) for a-TmHx crystals ]1c; the low-temperature
part demonstrates the suppression of the ferrimagnetic ordering manifestation with x (Vajda et al. 1989c).

an increase o f the residual resistivity in the b-axis crystals (fig. 53a), the picture changed
qualitatively for the c-axis crystals (fig. 53b). The low-temperature decrease in p due to
ferrimagnetic ordering is strongly suppressed and vanishes completely for x > 0.05. The
latter is attributed to a modification (by ~30% for x = 0 . 1 ! ) o f the magnetic superzone
gap produced b y the periodic structures along the c-axis, due to the increase o f Pres.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

271

The temperature dependence of the magnetic resistivity term in the basal plane Pm,b(T),
is best fitted through a combination Pm,b(T) e( ATn+ BT 2 exp(-A/kBT), where the first
term predominates at low temperatures (T A/kB) and possibly signifies the presence
of an AF excitation, while the second term is due to magnon excitation in anisotropic
ferromagnets. Both the power n and the spin-wave excitation gap A decrease in the
hydrogenated crystals, i.e. from n ~ 5 to 2.5 and from A/kB ~ 38 K to 6 K, for x ging
from 0 to 0.1.
There also exist susceptibility measurements on two non-magnetic *-phase solid
solutions. In (x*-ScHx, Volkenshtein et al. (1983) fotmd X decreasing by nearly a factor
of 2 between pure Sc and x = 0.36, illustrating the decrease of the density of electrons
responsible for the measured paramagnetism. In ct*-LuHx, Stierman and Gschneidner
(1984) observed, for the same reasons, a regular decrease of x for 0.01 ~<x ~<0.06; at
low temperatures, the behaviour is explained by superposition of spin fluctuations and
hydrogen tunneling (cf. fig. 29 in sect. 5).
6.2. (Mainly) O-phase systems
A systematic study of the paramagnetic spin-disorder resistivity, pm(T), has been carried
out on the heavy RH2 compounds (R = Gd through Tm) by Burger et al. (1986a) and Daou
Table 22
Spin-disorder resistivities, p o , measured at

T=Tr~, and ground-state magnetic moments, #Gs, for pure

hydrides RH 2

p~(T~) (~t~2cm)

#as (/~B)

Ref.

Ce

14

0.91

1,2,3

Pr

~4

1.54

2,4

1.06; 1.09

5,6,7

7.1

9
10,11

Nd
Sm
Eu

4.3
12
s.c. a

Gd

28

6.5

Tb

15.3

7.4

12,13

6.4

13,14
13,15

Dy

4.25

Ho

5.8

34

Er

1.5

2.75

15,16

Tm

17

a S.C., semiconductor.

References
(1) Vajda et al. (1990a)
(2) Abeln (1987); Arons et al. (1987b)
(3) Burger et al. (1991)
(4) Daou et al. (1989a)
(5) Senoussi et al. (1987)
(6) Daou et al. (1992)

(7) Carlin et al. (1982)

(8) Vajda et al. (1989b)
(9) Bischof et al. (1983)
(10) Vajda et al. (1991a)
(11) Arons and Schweizer (1982)
(12) Vajda et al. (1987a)

(13) Shaked et al. (1984)

(14) Vajda and Daou (1992a)
(15) Daou et al. (1988b)
(16) Arons (1982)
(17) Burger et al. (1986a)

P. VAJDA

272

Table 23a
Characteristic magnetic temperatures in light RHz+x-systems
R
Ce

Pr

Nd

Sm

Eu

2+x
2.00
2,00+0.02
2.00+0.05
2.00+0.10
2.00+ 0.25
2.00+0.40
2.00 + 0.60
1.95D
(1.95 + 0.05)D
(1.95 + 0.44)D
(1.95 + 0.67)D
(1.95 + 0.82)D
(1.95 + 0.96)D
1.97
1.96 + 0.03
1.96 + 0.08
1.95+0.15
1.97+0.23
1.98+0.53
1.97+0.76
1.95D
2.25D
2.00
1.99 + 0.06
2.00 + 0.32
2.00 + 0.44
2.00+0.56
2.00+0.60
2.00+0.65
2.7
1.99
2.00+ 0.03
1.98+0.16
1.96+0.26
1.85
2
2D

TC,N~a (IC)
(2.4); 4.2; 6.9
(2.4); 4.2; 6.9
n.o. ~> 1.3 b
4.1; 6.8
3.1; 6.2
3.3; 4.3
3.3
(1.7); 5; 6.2
n.o./> 0.35
4.0
3.1
n.o. ~> 1.8
4.1
3.35
2.0
2.5
n.o. ~> 1.5
n.o. ~> 1.5
n.o. ~> 1.5
3.3
n.o./> 2
6.6
6.0
5.3
2.65
2.9; 4.5
(1.75); 2.95
(2.15); 2.25
3.5
9.6
9.6
8.0
n.o. >/1.5
15.8
18.3
16.3

Tral (IC)
14.5
16
17.5
21
16

6.0
(~5)
5.5

Tm~ (IC)
42.5
42
42
33
26

28
16.5
12.5

4.5
4.3
(5.5)

9.5
12.2

10.2
10.8

13.5
12.6

Ref.
1,2
1,2
1,2
1,2
1,2
1,2
1,2
3,4
3,4
3,4
3,4
3,4
3,4
5,6
5,6
5,6
5,6
5,6
5,6
5,6
3,7
3,7
8,9
8,9
8,9
8,9
8,9
8,9
8,9
8,10
5,11
5,11
5,11
5,11
12,13
12,13
14

continued on next p a g e

HYDROGEN 1N RARE-EARTH METALS, 1NCLUDING RH2+~ PHASES

273

Table 23a, notes

a TNj stands for several TN (TN1, TN2, etc.).

b n.o. >/ . . . . not observed above . . . .

References
(1) Vajda et al. (1990a)
(2) Burger et al. (1991)
(3) Abeln (1987); Arons et al. (1987b)
(4) Schefer et al. (1984)
(5) Vajda et al. (1989a)

(6) Daou et al. (198%)

(7) Arons et al. (1987a,c)
(8) Senoussi et al. (1987)
(9) Daou et al. (1992)
(10) Drulis and Stalinski (1982)

(11)
(12)
(13)
(14)

Vajda et al. (1989b)

Drulis and Stalinski (1989)
Bischof et al. (1983)
Drulis (1993)

et al. (1988a). It was noted that the overall variation ofpm followed the de Germes factor
but that their absolute values were much smaller than those for the corresponding pure
R-metals. This implies not only modifieations of the electronic band structure but also a
significant reduction of the exchange interaction between conduction and f-electron spins.
Deviations of the ground-state resistivities Pm,
GS at the ordering temperature T = TN, from
the theoretical free-ion values are shown to originate from orten large crystal-field (CF)
effects. The values of Pm and of the ground-state magnetic moments #es are given in
table 22. A theoretical basis for the CF-analysis in the RH2+x-systems has been given in
the work of Lea et al. (1962) who had calculated the splitting of the (2J+ 1) degeneracy
of the ionic ground state in a cubic environment. The interaction of these CF-exeitations
with magnetism ean be observed principally by inelastic neutron scattering and also in the
elassical measurements of specific heat, resistivity or susceptibility. (A detailed discussion
is given by Arons 1991.)
Table 23a assembles the various characteristic temperatures for magnetie transitions as
observed by different experimental techniques. We have separated the light RH2+x-systems
(table 23a) from the heavy ones (table 23b), since the situation in the latter is often more
complicated. Furthermore, this allows us to present a typical selection of coherent data
covering a representative x-interval using the most recent and reliable results. Tc,N are
the Curie and the Nel-temperatures (sometimes several TNi) for transition towards the
FM and the AF stares, respectively; TM1 and Tml are the positions of the low-temperature
maxima and minima in resistivity, corresponding to incommensurate ordering transifions
and/or short-range ordered spin fluctuations (or Kondo effects for CeH2+x or YbH2+x);
TM2 and Tm2 are the respective positions for a second possible higher-temperature
magnetic transformation. The various TM,m are sometimes seen as additional kinks or
shoulders in susceptibility and specific-heat measurements.
6.2.1. Specic RH2+xhydrides
The magnetic phase diagram of the CeH2+x-system is rather eomplex and has been established mainly from neutron diffraction (Schefer et al. 1984, Abeln 1987), susceptibility
(Abeln 1987, Arons et al. 1987b) and resistivity measurements (Vajda et al. 1990a, Burger
et al. 1991). There are two AF transitions in the pure dihydride CED1.95 (Abeln 1987)
or CeH2.00 (Vajda et al. 1990a, Burger et al. 1991): TN1 =6.2K or 6.9K, TN2=5K or
4.2 K, possibly followed by another magnetic configuration at even lower temperatures,
1.7K or 2.4K. Neutron-seattering experiments on single crystals (Arons et al. 1987b)

274

P. VAJDA
Table 23b
Characteristic magnetic temperatures in heavy RH2+x-systems
TM1 (K)

Tml (K)

23.7

53.7

1.98+0.14

11.5

60.3

1.96+0.17

14.5

36.6

1.97 + 0.21

(16)

R
Gd

2+x

Tc,Nia (K)

1.94

18.3

1.97 + 0.04

17.75

Tb

1.96

9.5
12

3.7; 5.0

1.97+0.02

3.7; 5.0

1.98+0.10
1.97+0.13

65

105

30.5
32.5

38.5

52

38

52.5

2.55

32.5

42

36

42

5.5

12
(10)

5
5

8.6

(10.5)

16

10.7

20.5

10.7

18.7

5
6

5.4

2.00 + 0.095
2.00+0.175

6
8.2

2.00 + 0,23
3.2
3.5; 4.7

3
4
4

3.4; 5.5

2.00 + 0.047

5
~<1.3
1.8

1.99+0.27

25

1.99+0.22

Ho

16; 19

1.99

2.98
1.98

80

2.00 + 0.245

2.00

74

2
12.5

1.97+0.20

Dy

54.3

18; 21

1.96+0.15

Rel.

57

1.8

1.97 + 0.07

2.00D
(2.00 + 0.18)D

Tm2 (K)

1.97 + 0.305
3

TM2 (K)

6.3

11.2

6
6

15.3

6
7
8

24

1.99 + 0.02

3.75; 4.65

(2.65)

23

1.98 + 0.03

~3.5

(1.45)

23

1.98 + 0.064

5.3

23

1.99+0.106
1.98+0.12
1.99 + 0.14

6.3
6.3
6.3

(~15)
(~13)

25
32

8
8
8

2.00

3.5; 5.2

6.9

2.00+0.035
2.00 + 0.07

3.0; 4.2

5.9
6.3

6
6

2.00+0.133

6.7

2.00+0.16
2D

7.2
~9

6
9

4.0

continued on next page

HYDROGEN 1N RARE-EARTH METALS, INCLUDING RH2+x PHASES

275

Table 23b, continued

R
Er

2+x
1.98
(1.94+ 0.01)D

T,N: (K)

TM1 (K)

2.30

7.2

14.0

I0

1.98

2.15

14.0

10

Tm2 (K)

Ref.

8.6

10

1.96+0.03

1.85
1.53

(2.1)

14.0

10

1.98+0.07

1.75

(2.1)

18.5

10

(2.15)

(~5)

(2.2)

(-5)

1.98 + 0.091

Tm

TMz (K)

1.98 + 0.055
(1.97 + 0.09)D

Yb

Trl (K)

1.98
1.96 + 0.025

2.30
2.20

1.98 + 0.048

2.17

2.43

1.98+0.088

2.32

2.54

2D
1.96

2.15
n.o. >~ 1.5 b

1.80

n.o./>4

10
30

10

11
11
11
11
9
12
13

2.25

3.9

13

2.37

3.9

13

2.41

13

a T~i stands for several T N (Tm, Tr~ etc.),

b n.o. >/ . . . . not observed above . . . .

References
(1) Vajda et al. (1991a)
(2) Carlin et al. (1980)
(3) Vajda et al. (1987a)
(4) Vajda et al. (1993)
(5) Vajda and Daou (1992a)

(6) Boukraa et al. (1993a)

(7) Carlin and Krause (1981a)
(8) Daou and Vajda (1994)
(9) Shaked et al. (1984)

(10) Vajda and Daou (1994)

(11) Boukraa et al. (1993b)
(12) Burger et al. (1986a)
(13) Drulis et ak (1988b)

permitted the analysis of the AF structure below Tm as MnO-type, with AF-coupted

FM-planes modulated along [1101; the propagation vector is rm = 11111] + [110], cf.
fig. 54a. The introduetion of excess H atoms x on O sites first induces the disappearance
of the AF order (for x = 0.05), and then the appearance of FM order (for 0.05 ~<x < 0.6),
followed by another AF order up to the highest concentrations measured (x = 0.96). The
latter AF strucmre is close to that observed for x = 0, simplified by the lack of the
[110] modulation (fig. 54b) caused by the absence of the RKKY-mediating conduction
electrons. All magnetic transformatins are strongly influenced by the ordering state of
the x-atoms (cf. sect. 3). Thus, a quench from room temperature into liquid nitrogen,
conserving the random distribution of the octahedral H atoms, drastically modifies the
magnetic manifestations, sometimes even leading to their disappearance, cf. e.g. fig. 55
for x = 0.1 and 0.2 (Burger et al. 199 t). The same is true for the Kondo transition observed
in these systems (Vajda et al. 1990a): the coherent-incoherent transformation temperature
shifts from 17.5 to 2 2 K after a quench, while the two magnetic transitions at T N I and

276

P. VAJDA

~.-'-~5~--I 4--/Y

Fig. 54. (a) Magnetic structure of CeDI.95. The AF-coupled (111) planes show a modulation along [li0].
(b) Magnetic structure of CeDz91 . The modulation shown in (a) is suppressed (Abeln 1987).

ss ~ j ~

i,f----"

~0.2

50

o~~'

I, " ~ ~ - ~

45

',-o.,

40

35

35

3O

25
5

II

10

30

50

7 0 T(K)

Fig. 55. Thermal variation of the resistivity of CeHzl and CeH2.2 in the relaxed (open symbols) and quenched
(solid symbols) state. The arrows indicate the Kondo minima and maxima between 15 and 35 K, together with
the magnetic anomalies at lower temperatures. Note the suppression of magnetism in the quenched specimens
(Burger et al. 1991).

TN2 vanish (cf. fig. 35 in sect. 5.21 and the related discussion). Figure 56 summarizes
the data in the form of a magnetic phase diagram.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RHz+x PHASES

0.2

0.4

0.6

0.8

277

"

I
0.2

I
0.4

I
0.6

_1

I
0.8

x(at.H/at.Ce)

Fig. 56. Magnetic phase diagram for CeH2+x,

constructed from resistivity (Burger et al.
1991), susceptibility (Abeln 1987) and
neutron scattering data (Schefer et al.
1984).

The pure Pr dihydride PrH1.97 possesses a TN = 3.35 K (Abeln 1987, Vajda et al. 1989a)
and exhibits - like CeH2 - a resistivity minimum (Daou et al. 1989b) which is not related
to an incommensurate magnetic structure. In the case of CeH2.00, it has been connected
to a Kondo effect (Vajda et al. 1990a), while in PrH1.97 it is due to a CF effect. It
was shown (Burger et al. 1990) that the experimental situation can be well reproduced
assuming a magnetic ground stare, F5 in this case, and a non-magnetic first excited
stare, F1, so that the contribution of spin-disorder scattering to p decreases first in a
certain interval with increasing temperature, before increasing again due to the taking
over by the phonon scattering. The progressive disappearance of the p-minimum and of
the magnetic ordering for x > 0 (cf. fig. 57 and table 23a) can thus be explained by a
splitting of the degenerate F5-triplet through the non-cubic CF symmetry. Investigations
of CF-levels in PrH(D)2+x were performed by inelastic neutron scattering (Knorr et al.
1978, Arons et al. 1987c), specific heat (Drulis 1985) and by proton NMR (Belhoul et al.
1991); the latter results nieely confirm the above interpretation of the p-data through
an orthorhombic CF distortion giving a singlet ground state. The magnetic structure
of PrD2 was determined by neutron diffraction (Arons and Cable 1985, Abeln 1987)
and is similar to that of CeD2 (fig. 54a), except for a different propagation vector:
rm = 11111]- [112].
NdH2, apart from the orthorhombic semiconducting EuH2 (see below), is the only pure
(x = 0) R-dihydride ordering ferromagnetically (Te ,-~6.5 K), see Carlin et al. (1982), Bohrt
and Arons (1982), Senoussi et al. (1987). Upon introduction of x-hydrogen atoms, the
FM is gradually suppressed, evolving for x ) 0 . 5 towards complex AF structures (Daou
et al. 1992) and behaving for still higher x (>0.65-0.7) like a spin glass (Senoussi
et al. 1987). This effect is due to an interplay between RKKY-exchange interactions
and the random uniaxial anisotropy energy, which appears for x > 0, and is sensitive
to the concentration and the ordering state of the exeess hydrogens x. Thus, similar to
CeH2+x (see above, fig. 55) a quench aeross the 200K region can completely suppress
the magnetic ordering and turn the sample spin-glass like (e.g. fig. 58 for x=0.65).

278

P. VAJDA
0

10

20

30

40

55

EB

T2

T1
~1oo~ ....
/

54,'

Pr H2. x
0.15

32

T3

54

28

oo!.

~F

651
0

~1

oo

--

I
10

4K

[
20

~~~~_

T I
30

10.5

T/K

40

Fig. 57. Thermal variation of the resistivity of several PrH2+x speeimens, showing the characteristic temperatures
for magnetic transitions (Vajda et al. 1989a).

This effect becomes more pronounced when coming close to the M-S transition (and
the y-phase region) near x=0.7 (of. sect. 5.2.2), a clear indication for a modification
of the electronic band structure. A similar flattening of the initially k-type peaks for
the magnetic transitions was seen in specific-heat measurements (Drulis and Stalinski
1982) when approaching x ,~ 0.7. A tentative magnetic phase diagram in fig. 59 (Daou
et al. 1992) summarizes the experimental situation, possibly involving a two-phase region
around x ~ 0.6 (cf. sect. 3, fig. 15).
The pure dihydride SmH2 has an AF structure of the MnO type, with a propagation
vector r m = [ l l l ] (Arons and Schweizer 1982); it orders at TN=9.6K. For x > 0 ,
TN decreases progressively (down to 8K for x=0.16) and disappears completely for
x > 0.25. Again, like the lighter RH2+x-systems (see above), structural disorder in the
x-sublattice exerts a strong influence upon the magnetism. A striking example is given in
fig. 60 where the (now possibly incommensurate) magnetism seems to have reappeared in
the quenched x = 0.26 specimen. These phenomena are probably related to perturbations
of the CF, which may lead to a splitting of the F8 ground-level quartet. Indications for this
were noted, for x > 0, in the anomalous temperature variations of the resistivity (Vajda
et al. 1989b) and the complex behaviour of the specific heat (Opyrchal and Bieganski
1978). In the latter experiment, their SmHz00 speeimen contained a non-negligible amount

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+ PHASES

I

T(K

PM

279

NdH2+ x

--4.0
4

3.9

2.1
B

X = 0.65 R

~L

2.0
1.94

~
,
I
2 ~, 2.5 K
!
12"17K
I

10

o.2

I ?"~
0.4

S
0.6
0.8
x (at.H / at.Nd)

Ndl-I2+x system, from resistivity (solid circles, Daou

et al. 1992)and magnetisationdata (crosses, Senoussi
et al. 1987).

2.25K

I
1.5

Fig. 59. Tentative magnetic phase diagram in the

1.9
1.90

2 -- FM I

15 T(K)

Fig. 58. Thermal variation of the resistivity of relaxed

(R, solid circles) and quenehed (Q, solid triangles)

NdHz65 in the magneticregion (Daou et al. 1992).

ofx-hydrogen atoms, which is reflected in a rather low TN ---6.75 K and a bump near 10 K,
whieh was not mentioned by the authors, and possibly corresponding to the TM1 of the
p-measurements (cf. table 23a).
The orthorhombic EuH2 is (like YbH2) an insulator, since the divalent europium has
lost both its valence electrons upon hydriding. The Eu 2+ ion has seven unpaired spins
and is strongly magnetic, resulting in a FM transition at Tc = 16-18K (Mustachi 1974,
Bischof et al. 1983, Drulis and Stalinski 1989, Drulis 1993).
The magnetic structure of pure GdH2 (like that of SmH2) is AF of MnO type, with
a propagation vector ~m = [111] (Arons and Schweizer 1982). Additional x-hydrogens
on O sites lead to incomrnensuracy, the structure becomes helicoidal, with an axis
along [111]. Simultaneously, the TN, which was 18.3 K for x = 0, decreases regularly to
TN = 11.5 K for x=0.14, the situation becoming more complicated for higher x (Vajda
et al. 1991a). Thus, a second maximum at TM2 appears in the p-curves for x=0.16 (cf.
fig. 43 in sect. 5.2.2) which grows in amplitude at the expense of the first maximum at TM1
until the complete vanishing of the latter for x near the phase boundary at x~max~ 0.30.
The extreme sensitivity to x-sublattice ordering is shown in the quench experiments. The
maximum at TM2, which is evidently related to the ordered H structure, has disappeared to
the benefit of a strong emergence of the maximum at TM1, which is obviously conneeted
to the disordered (random) x-hydrogen atoms. Gd being a pure S ground state ion (L = 0),
one does not expect any significant interaction with the CF; therefore, the strong variations
of the magnetic manifestations have to be attributed to profound modifications of the
Fermi surface and its nesting conditions. Figure 61 presents, as a summary, the tentative
magnetic phase diagram of GdH2+x in the relaxed and in the quenched stare of the x-

280

P. VAJDA
0

10

105f

20

T/K

30

11 Sm H2+x I. ."

104 5'

: ..

1;:I~-::i~oio:~~'~~i~
i:

95
90

~~

5oooo
~,~-/-~

69.

.."

~oooo~.. x~o,

,Fr

4.5

37"

* o*

36.5

35

34.5

36

2~

oO

~ o o
.

~ x = O

o:R

20.

201
0

10

20

T/K

30

Fig. 60. Thermal variation of the resistivity of

several SmH2+x specimens in the relaxed (open
symbols) and quenched (solid symbols) stare.
Note the reappearance of magnetism after a
quench for x=0.26 (Vajda et al. 1989a,b).

/
-

/ /

~~ f T : ~

80

+/4/ / /

iC 2

60
p-

40

IC

.~ Tml
"~+ i"' icl

20
-

AF

hel
]

0.1

o/

0.2
0.3
x(at,H(D)/at.Gd)

Fig. 61. Tentative magnetic phase diagram of

GdH2+x(open symbols) and GdD2+x (solid symbols)
in the relaxed (solid lines) and in the quenched
(dotted lines) stare obtained from resistivity studies
(Vajda et al. 1991a). icl, incommensurate or
magnetically SRO phases; hel, helieoidal phase
For the various critical temperatures, see text and
table 23b.

HYDROGEN IN RARE-EARTHMETALS, INCLUDINGRHz+x PHASES

(a)
/

;/

(b)
k= [1/4, 1/4 ,3/4] [113]
/

,/
,

281

k'= [1/8. 1/8,3/4 ]//[116]

Fig. 62. Magnetic structures of TbDz00; (a) commensuratephase observed for T < 16 K; (b) incommensurate
phase in the interval 14.8~<T ~<19 K. The projections of their magnetic momentsupon the directions of the
corresponding propagationvectorsk are shown to the left of the structures (Vajda et al. 1993).

hydrogens in the interval 0 ~ x ~<0.30 (Vajda et al. 1991a). The hexagonal y-phase GdH3
orders AF at TN = 1.8 K (Carlin et al. 1980).
The structure of pure TbH(D)2 below TN = 16 to 18 K as determined by neutron
diffraetion (Arons et al. 1982, Shaked et al. 1984) is commensurate AF, modulated with
a propagation vector r~n= [113]; the spin axis is oriented parallel to [001]. Between
TN (and even overlapping it by AT ,~ 2 K) and an intermediate transition temperature,
TI = 19-21 K, one observes a sinusoidal incommensurate magnetic structure, with a
propagation vector rim
c ~ [0.12, 0.14, 0.76], close to the [116] direction (Vajda et al.
1993).The two configurations are shown in fig. 62. The superstoichiometrie compounds
studied neutrono-graphically by Vajda et al. (1993) for the cases x = 0.18 (and 0.245), also
exhibit two magnetie phases. One exists below Ty = 32.5 (and 36 K), and is eommensurate
and modulated with a propagation vector r ~ = [114], i.e. it is different from that in pure
TbD2. The other exists between TN and T1 = 42 K (and 43 K), and is incommensurate with
a ~ime~ [0.24, 0.18, 1] and slightly varying with temperature. The overlap interval with
the commensurate phase increases, with a roughly constant relative width AT~T2 ~ O. 1.
The intricate interplay between structural ordering of the x-atoms (determined as DO22
for TbD2+x by neutron scattering by Andr et al. 1992, cf. sect. 3.2) and the magnetie
ordering, which increases the ordered intervals and modifies the configurations, has also
been observed through quenching effects in resistivity studies (Vajda et al. 1987a). This
behavior is similar to that discussed above for GdH2+x (cf. fig. 43). It is encouraging that
the configurations (both the commensurate and the incommensurate ones) are different
for the hydrides with x = 0 and with x > 0, especially in view of the presenee of different
H ordering in the two cases. In the same context, we also mention the existence of two
types of Tb-atoms with regard to their surroundings in an ordered DO22 stmeture (fig. 13):
25% of them have two oceupied O sites as nearest neighbours and 75% only one, giving

282

l
40

P. VAJDA

TbH(D) 2+x

~-

2O

0.1
0.2
x(at,H(D)/at,Tb)

Fig. 63. Magnetic phase diagram of TbH(D)zx exhibiting the

existence region for the two commensurate phases (a) and (b)
and the ineommensurate phase (c) (note the overlap regions!).
Crosses, neutron data; open circles, resistivity data; open squares,
susceptibility data, solid diamonds, specific heat data (Vajda et al.
1993).

a clear indication as to the different CF symmetries as a possible origin for the various
magnetic phases. An analysis of the CF splitting in TbH2+x up to x ,~ 0.15 was made using
specific-heat measurements by Drulis et al. (1984), who established F2 as the ground state
singlet for small x and a contribution of a doublet (or triplet) for larger x, due to the axial
symmetry of some Tb 3+ ions. The magnetic phase diagram in fig. 63 is made up from
relevant neutron-diffraction, resistivity, susceptibility and heat capacity data and shows, in
particular, the overlap region between the commensurate and incommensurate phases.
The magnetic structure of DyH2 is similar to that of TbH2. It is AF commensurate
c _ 1 [113] (Shaked et al. 1984). A second
below TN = 3.7 K, with a propagation vector Zm-~
transition due to the appearance of an as yet undetermined intermediate structure (but
probably similar to that oceurring in TbD2), is seen at TI=5.0K (Shaked et al. 1984,
Bieganski et al. 1975, Vajda and Daou 1992a). The addition ofx-hydrogens suppresses the
manifestations of the commensurate AF order: TN is no longer visible for 0.05 < x < 0.15
in the resistivity (Vajda and Daou 1992a), suseeptibility (Carlin and Krause 1981a)
and Mssbauer measurements (Friedt et al. 1979). The presence of some structure in
susceptibility curves (broad, flat peaks) could be evidence for short-range AF order
(Boukraa et al. 1993a). For still higher x-values, x>0.15, however, new transitions
appear, with a striking p-peak eentered at 10.7K (a precursor hump is observed at
10.5K for x=0.13) and a minimum at 19-20K (Vajda and Daou 1992a). The extreme
sensitivity to x-hydrogen ordering is shown in fig. 64, clearly assigning the 10.7K peak
as representative for the ordered H-configuration. A tentative magnetic phase diagram
constructed from resistivity and susceptibility data (Vajda and Daou 1992a, Boukraa et
al. 1993a) is shown in fig. 65. The break near x=0.1 is related to the phase boundary
between SRO and LRO x-hydrogen ordering in the structural phase diagram (fig. 16 in
sect. 3.2). Finally, the y-phase trihydride DyHz98 orders AF at 3.2 K (Carlin and Krause
1981a). Recent diffraction studies on DyD2 using cold neutrons (to compensate for the
high thermal-neutron absorption of Dy) undertaken by Vajda et al. (1995) permitted to
specify the low-temperature structure below TN as incommensurate, with a propagation
vector r~ ~ [0.28, 0.28, 0.745], close to but not exactly 11113]. Surprisingly enough, this

HYDROGEN 1N RARE-EARTH METALS, INCLUDING RH~+~ PHASES

I

x = 0.27

47.5 ~-

" ~,72.5

47"4 ~ ~ 1 ~ ~ ~ ~
1

72'4 j

4K

10

AF-sao [

50

ic

;o 45
c

:::L
CL

2~

[ -~'~~~~'/'~-]72.6

2.5K

283

AF-c

\,

AF- e

41{

9K

0.1

72.8,

10.7K ~N 72.61~1
2

50 --

45

I
3

I
4

70

]
5K

65

0.2
0.3
x(at.H/at.Dy)

Fig. 65. Tentative magnetic phase diagram of DyHz+x

indicating AF-commensurate and incommensurate
and SRO phases, constructed from resistivity (solid
eircles) and susceptibility (erosses) data (Vajda and
Daou 1993).

20.5K

B,,~..

DY H2+x

~o.71- ~,'-.
~~-~~
2
01 ~"
0

5.1

x = 0"22 .~60

2k ~: x01~
11
10

51 7 "L' I
20 T(K) 30

Fig. 64. Resistivity of DyH2+x specimens in the

magnetic region. Note the striking changes introduced
by a quench (Q, solid inverted triangles) when
compared to the slowly cooled samples (R, solid
circles) (Vajda and Daou 1993).

incommensurate AF structure seems to coexist with a FM component, the latter vanishing

at Tc = 5,8 I~, close .to the above mentioned TI.
The magnetic structure of HoD2 was determined by neutron diffraction to be of the
same type as TbD2 (cf. fig. 62), except for the spin axis oriented close to [110] (Shaked
c _- 1~[113] below T~ ~ 5 K, and exhibits
et al. 1984). It is AF, with a propagation vector r m
an undetermined intermediate structure between TN and Tl ~ 6.5-8 K. T1 is nicely seen
as a maximum in the resistivity (fig. 66) and in the susceptibility (Boukraa et al. 1993a),
while TN is just a break in the resistivity drop below Tb Only TN has been observed in
specific-heat measurements (Bieganski and Stalinski 1976). An additional break in p is
noted at 3.5K. T h e resistivity minimum near 23 K could be due to magnetic fluctuations
but also, possibly, to CF effects, i.e. to excitations from a magnetic ground state (Fs)
to a non-magnetic first excited state (F3), similar to PrH2 (Daou et al. 1989b, Burger
et al. 1990). The situation in the superstoichiometric dihydrides evolves rapidly (fig. 66).
Thus, the addition of only 0.02 at.H/Ho on octahedral sites leads to a nearly complete
0 , the resistivity drop due to ordering below TI, TN. For x ~>0.05, only
suppression OfPmag

284

P. VAJDA

57

B',~ /\~.~\,/x=o.o~ ,~

23

25

"~ 7

x=o.o2

4,65 J~'\

~ , 2 \ T\

4 /x=O

~
2

~.~~
i
1

1
2

I
3

I ~
4
T(K)

i
6

I
7

I
8

Fig. 66. Resistivity of various Hort2+ specimens in the

magnetic region showing the evolution of the critieal
temperatures as a function ofx (Daou and Vajda 1994).

TI is l e r as a manifestation of magnetism in the form of a slope maximum near 6 K

in p and as a flat maximum in X (Boukraa et al. 1993a). On the other hand, a second
structure develops as a hump above T ~ 25 K for concentrations x > 0.12 (el. also fig. 46
in sect. 5) which could be due to the appearance of a new magnetic phase caused by the
LRO in the x-sublattice, similar to GdH2+x and TbH2+x.
The magnetic structure of ErH2, the heaviest of the magnetically ordering dihydrides,
is a complex AF below TN = 2.15-2.30 K (Bieganski and Stalinski 1976), with a mixture
of both eommensurate and incommensurate components down to 1.5 K (Shaked et al.
1984). The incommensurate structure lines in the neutron diffraction pattern are at the
same positions as those of the intermediate structures in TbD2, DyD2 and HoD2, showing
their close relationship to each other. It should, in fact, be interesting to investigate ErH2
down to even lower temperatures in view of an eventual observation, by analogy, of the
pure commensurate structure, since the measured "mixed" spectra might be representative
only for the overlap region between the two phases such as noted in TbD2 (Vajda et al.
1993). This close relationship is also evident in the behaviour of the superstoichiometrie
ErH2+x-system, in that it is similar to other heavy RH2+x compounds: a decrease of TN
for small x and the appearance of a new magnetic structure at higher temperatures for
x high enough to form a H sublattice (figs. 67 and 48). The tentative phase diagram
for the characteristic temperatures in the magnetically ordering range (fig. 68) was
constructed from the available resistivity (Vajda and Daou 1994) and susceptibility
data (Boukraa et al. 1993b, Carlin and Krause 1981b). Again, as for DyH2+x, neutronscattering measurements are required to determine the precise SRO and LRO magnetic
configurations.

285

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

0.02
121
x = 0.088 " ~

29.5

~~

29

119
117

L
3!

0,1
I ....

Er H2+x

g ~ pM

~ \x=oo7
2.5

~-2~

12.0

1.5

1
0

0.6
0

0.06

I
1

I
2

I
3

1
_l
0.04
x(at.H/at.Er)

f
0.08

I
4 T(K)

Fig. 67. Resistivity of various ErH2+x speeimens in

the magnetie region showing the evolution of the
transition temperature with increasing x (Vajda and
Daou 1994).

Fig. 68. Tentative magnetic phase diagram of ErH2+

indicating the different coexisting eommensurate and
incommensurate phases as well as the SRO domains,
constructed from resistivity (solid triangles, solid
inverted tringles, Vajda and Daou 1994) and suseeptibility data (solid cireles, solid squares, Boukraa et al.
1993b; crosses, Carlin and Krause 1981b).

No magnetic transitions have been observed in TmH2 down to 1.5 K. An analysis of

the spin-disorder resistivity, Pm, shows that the ground stare (F2) is non-magnetic but
separated from the first excited magnetic state (F~2)) by AE ~ 170K - a Van Vleck
paramagnet (Burger et al. 1986a, Shaltiel et al. 1991a). The only low-temperature
investigation undertaken up to now in the TmI-I2+x system coneerned Tm-169 Mssbauer
studies on samples with ill-defined eompositions by Waibel et al. (1980) who interpreted
the spectra as evidence for the presence of two non-magnetic phases. Further thorough
experiments should, however, be interesting in view of the great sensitivity of such a
Van Vleck compound to minor changes in the surroundings of the Tm-ions. For example,
it was shown in EPR-linewidth measurements of substitutional Gd and Er impurities
(0.01 at.%) in TmH2 that the latter reduced the first CF-level separation AE by nearly
a factor of two (Shaltiel et al. 1991b). Even greater effeets should be expected from
interstitial type defects such as the oetahedral x-hydrogens.
The orthorhombic YbH2 is non-magnetic because of the divalency of the yb2+-ion,
which becomes an analog of Lu 3+. In the fee superstoiehiometric dihydrides with x > 0.25,
the presence of yb3+-ions, yielding a mixed-valence system, seems to lead to Kondolattiee behaviour, with a eoherent-incoherent transition near 4 K (Drulis et al. 1988b,
but: cf. also the note in sect. 5.2.2).

286

R VAJDA

7. Summary and outlook

This review of experimental data concerning hydrogen in rare earths, both in the form of
solid solutions and of hydrides, shows - as a first general result - the great influence of
the initial metal purity upon the characterization of the fmal specimen and its physical
and physico-chemical properties. A thorough control of the preparation conditions is a
further (related) requirement for obtaining unambiguous and reproducible results.
Among the general regularities along the R-series, one can note the evolution of
eertain dimension-dependent properties, for example a decreasing solubility in the
B-phase (i.e. a narrowing of the pure R-phase region). On the other hand and for the
same reasons, the solubility in the low-temperature *-phase increases, as the eventual
preeipitation of the B-phase is retarded in lattices with smaller unit cells. Similarly, for
the J-dependent magnetic properties, such as the transition temperatures TN and the spindisorder resistivities Pmag, the B-RH2hydrides follow the de Gennes factor, just as in the
pure metals.
As to the more prominent properties, one has to mention the hydrogen sublattice
ordering occurring below room temperature both in the *-phase solid solutions and in
the [3-phase RHz+x systems. In the former case one observes zig-zagging chains along
the c-axis made of H-H pairs on tetrahedral sites of the hcp lattice, and in the latter,
the excess hydrogen atoms x on octahedral sites of the fluorite type fee lattice form a
tetragonal DO22 configuration, at least in the heavy RH2+x compounds. This ordering has
a striking influence on the magnetie manifestations, via modification of the Fermi surfaee
and/or the crystal-field symmetry.
Also related to the above x-sublattice ordering are the recently observed metalsemiconductor transitions near room temperature in the heavy superstoichiometric
dihydrides GdH2+x, HoH2+x, ErH2+x and in YH2+x for x dose to the R-phase boundary
(0.1< Xmax
~ < 0.3). Another M-S transition oecurring in these systems at lower temperatures (20-12010 seems to be due to carrier localization caused by atomic disorder.
Analogous effects were noted in the light substoichiometric trihydrides (also in the
B-phase) LaHz+x and CeH2+x in the range 0,7 < x < 0.9.
As to future developments, one expects a more intense neutron-scattering work both for
the determination of structural phase diagrams (mainly concerning the H sublattice) and
of magnetic phase diagrams, in particular the complex ineommensurate configurations
in the heavy RHz+x compotmds. Inelastic neutron-scattering studies are required for the
investigation of phonon-dispersion relations and of H local modes. Also, further work is
needed for the determination of the precise electronic mechanism responsible for the M-S
transitions. And, finally, little is known of the low-T properties of the y-phase trihydrides
because of their inherent chemical instability.
Last but not least, a major effort should be devoted to the growth of monoerystalline
hydride specimens other than CeH2+x, a prerequisite for the study of short-range-ordered
configurations.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2+x PHASES

287

References
Abeln, A., 1987, Report no. Jl.-2152 (Jlich).
Anderson, I.S., J.J. Rush, T.J. Udovic and J.M. Rowe,
1986, Phys. Rev. Lett. 57, 2822.
Anderson, I.S., N.E Berk, J.J. Rush and T.J. Udovic,
1988, Phys. Rev. B 37, 4358.
Anderson, I.S., D.K. Ross and J.E. Bonnet, 1989, Z.
Phys. Chem. NF 164, 923.
Andr, G., O. Blaschko, W. Schwarz, J.N. Daou and
E Vajda, 1992, Phys. Rev. B 46, 8644.
Arons, R.R., 1982, Rare earth hydrides, in: LandoltBmstein New Seiles, Vol. 12c, ed. K.H. Hellwege,
eh. 6.3.
Arons, R.R., 1991, Rare earth hydrides, in: LandoltBmstein New Series, Vol. 19dl, ed. H.EJ. Wijn,
ch. 2.3.
Arons, R.R., and J.W. Cable, 1985, Solid State
Commun. 55, 835.
Arons, R.R., and J. Schweizer, 1982, J. Appl. Phys.
53, 2645.
Arons, R.R., W. Schaefer and J. Schweizer, 1982,
J. Appl. Phys. 53, 2631.
Arons, R.R., A. Abeln-Heidel and H. Mutka, 1987a,
ILL Exptl. Report (Grenoble) p. 81.
Arons, R.R., A. Abeln-Heidel and C. Ritter, 1987b,
ILL Exptl. Report (Grenoble) p. 186.
Arons, R.R., J.W. Cable and R.M. Nicldow, 1987c, J.
Appl. Phys. 61, 3423.
Barnes, R.G., D.R. Torgeson, T.J.M. Bastow, G.W.
West, E.EW. Seymour and M.E. Smith, 1989, Z.
Phys. Chem. NF 164, 867.
Barnfather, K.J., E.EW. Seymour, G.A. Styles, A.J.
Dianoux, R.G. Barnes and D.R. Torgeson, 1989, Z.
Phys. Chem. NF 164, 935.
Bashkin, I.O., E.G. Ponyatovskii and M.E. Kost, 1977,
Phys. Status Solidi b 83, 517.
Bashkin, I.O., E.G. Ponyatovskii and M.E. Kost, 1978,
Phys. Status Solidi b 87, 369.
Beattie, A.G., 1972, J. Appl. Phys. 43, 3219.
Beaudry, B.J., and EH. Spedding, 1975, Metall. Trans.
B 6, 419.
Belhoul, M., R.L Schoenberger, D.R. Torgeson and
R.G. Barnes, 1991, J. Less-Common Met. 172-174,
366.
Bermington, S.M., R. Osborn, D.K. Ross and M.J.
Benham, 1991, J. Less-Common Met. 172-174,
440.
Berk, N.E, J.J. Rush, T.J. Udovic and I.S. Anderson,
1991, J. Less-Cornmon Mer. 172, 496.

Bieganski, Z., and M. Drulis, 1977, Phys. Status Solidi

(a) 44, 91.
Bieganski, Z., and B. Stalinski, 1976, J. Less-Common
Mer. 49, 421.
Bieganski, Z., J. Opyrehal and M. Drulis, 1975, Phys.
Status Solidi (a) 28, 217.
Bischof, R., E. Kaldis and E Waehter, 1983, J. Magn.
& Magn. Mater. 31-34, 255.
Bischof, R., M. Tellefsen and E. Kaldis, 1985, J. LessCommon Mer. 110, 99.
Blaschko, O., G. Krexner, J.N. Daou and E Vajda,
1985, Phys. Rev. Lett. 55, 2876.
Blaschko, O., G. Krexner, L. Pintsohovius, E Vajda
and J.N. Daou, 1988, Phys. Rev. B 38, 9612.
Blasehko, O., G. Krexner, J. Plesehiutschnig, G. Ernst,
J.N. Daou and P. Vajda, 1989a, Phys. Rer. B 39,
5605.
Blaschko, O., J. Pleschiutsehnig, L. Pintschovius, LP.
Burger, J.N. Daou and P. Vajda, 1989b, Phys. Rer.
B 40, 907.
Blaschko, O., J. Pleschiutschnig, E Vajda, LE Burger
and J.N. Daou, 1989c, Phys. Rev. B 40, 5344.
Boeva, O.A., K.N. Zhavoronkova, M. Smutek and S.
Cern~, 1986, J. Less-Common Mer. 118, 1.
Bohrt, H.G., and R.R. Arons, 1982, J. Appl. Phys. 53,
2072.
Bonnet, J.E., 1975, Thesis (University of Paris).
Bonnet, J.E., and J.N. Daou, 1977, J. Appl. Phys. 48,
964.
Bonnet, J.E., and J.N. Daou, 1979, J. Phys. & Chem.
Solids 40, 421.
Boroch, E., and E. Kaldis, 1987, Inorg. Chim. Acta
140, 89.
Boroeh, E., and E. Kaldis, 1989, Z. Phys. Chem. NF
163, 117.
Borsa, E, R.G. Barnes, B.J. Beaudry and D.R.
Torgeson, 1982, Phys. Rev. B 26, 1471.
Boukraa, A., E Vajda and J.N. Daou, 1993a, in: Proc.
lnt. Conf. on Metal-Hydrogen Systems, Uppsala,
Sweden, 1992, Z. Phys. Chem. NF 179, 413.
Boukraa, A., P. Vajda and J.N. Daou, 1993b, J. Magn.
& Magn. Mater. 123, L5.
Burger, J.R, J.N. Daou, P. Vajda and A. Lucasson,
1984, J. Less-Common Met. 103, 381.
Burger, J.R, J.N. Daou, A. Lucasson and R Vajda,
1986a, Phys. Rev. B 34, 4782.
Burger, J.P., P. Vajda, J.N. Daou and G. Chouteau,
1986b, J. Phys. F 16, 1275.

288

R VAJDA

Burger, J.P., J.N. Daou and P. Vajda, 1988, Philos.

Mag. B 58, 349.
Burger, J.P., J.N. Daou and R Vajda, 1990, Z. Phys. B
80, 233.
Burger, J.R, P. Vajda and J.N. Daou, 1991, J. Phys. &
Chem. Solids 52, 779.
Cannelli, G., R. Cantelli, E Cordero, E Trequattrini,
I.S. Anderson and J.J. Rush, 1991, Phys. Rev. Lett.
67, 2682.
Carlin, R.L., and L.J. Krause, 1981% Chem. Phys.
Lett. 82, 323.
Carlin, R.L., and L.J. Krause, 1981b, Phys. Rer. B 23,
6149.
Carlin, R.L., R.D. Chirico, K.O. Joung, G.K. Shenoy
and D.G. Westlake, 1980, Phys. Lett. A 75, 413.
Carlin, R.L., L.J. Krause, A. Lambreeht and H. Claus,
1982, J. Appl. Phys. 53, 2634.
Chiheb, M., 1980, Thesis (University of Paris-Sud).
Chiheb, M., LN. Daou and E Vajda, 1993, in: Proe.
Int. Conf. on Metal-Hydrogen Systems, Uppsala,
Sweden, 1992, Z. Phys. Chem. NF 179, 255.
Chowdhury, M.R., G.A. Styles, E.EW. Seymour, S.EL
Cox, C.A. Scott and G.H. Eaton, 1989, J. Phys.:
Condens. Marter 1, 4659.
Cordero, E, 1993, Phys. Rev. B 47, 7674.
Danielou, R., J.N. Daou, E. Ligeon and P. Vajda, 1981,
Phys. Status Solidi a 67, 453.
Daou, J.N., 1965, C.R. Aead. Sei. Paris, Ser. C 261,
1675.
Daou, J.N., and J.E. Bonner, 1974, J. Phys. & Chem.
Solids 35, 59.
Daou, J.N., and P, Vajda, 19??, unpublished.
Daou, J.N., and R Vajda, 1988, Arm. Chim. (Paris) 13,
567.
Daou, J.N., and P. Vajda, 1992, Phys. Rev. B 45,
10907.
Daou, J.N., and R. Viallard, 1972, in: Hydrogne dans
les Mtaux, Vol. 1 (Seience et Industrie, Paris) p.
123.
Daou, J.N., A. Lueasson and P. Lueasson, 1976, Solid
Stare Commtm. 19, 895.
Daou, J.N., P. Vajda, M. Biget, A. Lucasson and R
Lueasson, 1977, Phys. Status Solidi a 40, 101.
Daou, J.N., R Vajda, A. Lueasson and P. Lueasson,
1980a, J. Phys. F 10, L305.
Daou, J.N., R Vajda, A. Lueasson and P. Lucasson,
1980b, Solid State Commun. 34, 959. Erratum: 35,
809.
Daou, J.N., M. Chiheb and P. Lucasson, 1981a, J.
Less-Common Met. 79, 65.

Daou, J.N., P. Radhakrishna, R. Tur and R Vajda,

1981b, J. Phys. F 11, L263.
Daou, J.N., P. Vajda, A. Lucasson and E Lueasson,
1981e, J. Phys. C 14, 129.
Daou, J.N., R Vajda, A. Lueasson and P. Lueasson,
1981d, J. Phys. C 14, 3155.
Daou, LN., R Vajda, A. Lueasson and E Lucasson,
1981e, Solid State Commtm. 40, 261.
Daou, J.N., P. Radhakrishna, R Vajda and Y. Allain,
1983, J. Phys. F 13, 1093.
Daou, J.N., A. Lueasson, E Vajda and J.R Burger,
1984a, L Phys. F 14, 2983.
Daou, J.N., P. Vajda, E Radhakrishna and Y. Allain,
1984b, Solid State Commun. 52, 1051.
Daou, J.N., E Vajda, J.R Burger and A. Lueasson,
1986a, Phys, Status Solidi a 98, 183.
Daou, LN., P. Vajda, A. Lueasson and J.R Burger,
1986b, Phys. Status Solidi a 95, 543.
Daou, J,N., 13. Vajda, A. Lueasson, E Lueasson and
J.P. Burger, 1986e, Philos. Mag. A 53, 611.
Daou, J.N., R Vajda and LE Burger, 1987, Solid Stare
Commun. 64, 937.
Daou, J.N., E Vajda and J.P. Burger, 1988a, Phys. Rev.
B 37, 5236.
Daou, J.N., R Vajda, J.P. Burger and D. Shaltiel, 1988b,
Europhys. Lett. 6, 647. Erratum: 1989, 8, 587.
Daou, J,N., E Vajda, G. Hilseher and N. PilImayr,
1988e, J. Phys. (Paris) 49(Suppl.), C8-357.
Daou, J.N., J.E Burger and R Vajda, 1989a, Phys.
Status Solidi b 154, 321.
Daou, J.N., R Vajda and J.P. Burger, 1989b, Solid State
Commun. 71, 1145.
Daou, J.N., J.P. Burger, E Vajda, G. Chouteau and R.
Tur, 1990, J. Phys.: Condens. Marter 2, 7897.
Daou, J.N., J.E Burger and E Vajda, 1992, Philos.
Mag. B 65, 127.
Drulis, H., and K.P. Hoffmann, 1985, Z. Phys. Chem.
NF 145, 11.
Drulis, H., M. Drulis and W. lwasieezko, 1988a, J.
Less-Common Met. 141, 201.
Drulis, M., 1985, Z. Phys. Chem. NF 145, 19.
Drulis, M., and Z. Bieganski, 1979, Phys. Status Solidi
a 53, 277.
Drulis, M., and B. Stalinskl, 1982, Phys. Status Solidi
a 69, 385.
Drulis, M., and B. Stalinski, 1989, Z. Phys. Chem. NF
163, 155.
Drulis, M., J. Opyrehal and W. Borkowska, 1984,
J. Less-Common Met. 101, 211.
Drulis, M., H. Drulis and B. Stalinsld, 1988b, J. LessCommon Met. 141, 207.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2x PHASES

Fairclough, J.EA., D.K. Ross and S.M. Bermington,
1993, in: Proc. lnt. Conf. on Metal-Hydrogen
Systems, Uppsala, Sweden, 1992, Z. Phys. Chem.
NE 179, 281.
Fedotov, V.K., VG. Fedotov, M.E. Kost and E.G.
Ponyatovskii, 1982, Sov. Phys. Solid State 24,
1252.
Ferreira, R, G. Teisseron, E Vuillet, J.M. Gil, RJ.
Mendes, A.R de Lima and N. Ayres de Campos,
1987, J. Less-Common Met. 130, 155.
Friedt, J.M., G.K. Shenoy, B.D. Dunlap, D.G. Westlake
and A.T. Aldred, 1979, Phys. Rev. B 20, 251.
Fujimori, A., and L. Schlapbach, 1984, J. Phys. C 17,
341 and referenes therein.
Gibbs, D., J. Bohr, J.D. Axe, D.E. Moncton and K.L.
D'Amico, 1986, Phys. Rev. B 34, 8182.
Glinka, C.J., J.M. Rowe, J.J. Rush, G.G. Libowitz and
A. Maeland, 1977, Solid State Commun. 22, 541.
Greis, O., E Knappe and H. Mller, 1981, J. Solid
Stare Chem. 39, 49.
Gschneidner Jr, K.A., J. Tang, S.K. Dhar and A.
Goldman, 1990, Physiea B 163, 507.
Gupta, M., 1978, Solid State Commun. 27, 1355.
Gupta, M., and J.E Burger, 1980, Phys. Rev. B 22,
6074.
Gygax, EN., A. Amato, I.S. Anderson, J.J. Rush and
A. Schenk, 1993, in: Proc. Int. Conf. on MetalHydmgen Systems, Uppsala, Sweden, 1992, Z.
Phys. Chem. NF 179, 181.
Han, J.W., C.T. Chang, D.R. Torgeson, E.EW. Seymour
and R.G. Barnes, 1987, Phys. Rev. B 36, 615.
Haschke, J.M., and M.R. Clark, 1975, High-Temp. Sei.
7, 152.
Herbst, G., and H. KronmUer, 1981, Phys. Status
Solidi a 66, 255.
Hunt, D.G., and D.K. Ross, 1976, J. Less-Common
Met. 49, 169.
Ito, T., B.J. Beaudry, K.A. Gschneidner Jr and T.
Takeshita, 1983, Phys. Rer. B 27, 2830.
Ito, T., S. Legvold and B.J. Beaudry, 1984, Phys. Rev.
B 30, 240.
Iwasieezko, W., H. Drulis and R.M. Frak, 1992,
J. Alloys & Compounds 180, 265.
Jacob, I., A. Wolf and M.H. Mintz, 1981, Solid State
Commun. 40, 877.
Jensen, C.L., and M.E Zalesky, 1980, J. Less-Common
Met. 75, 197.
Jensen, J., and R.A. Cowley, 1992, J. Phys.: Condens.
Matter 4, 9673.
Jensen, J., and A.R. Mackintosh, 1991, Rare Earth
Magnetism (Clarendon Press, Oxford).

289

Jones, RM.S., J. Southall and K. Goodhead, 1964,

AWRE-Report 0-22/64 (Atomie Weapon Researeh
Establishment).
Juckum, C., 1980, Ph.D. Thesis (University of
Paris VI).
Jung, P., and R. Lsser, 1988, Phys. Rev. B 37, 2844.
Kai, K., K.A. Gsehneidner Jr, B.J. Beaudry and D.T.
Peterson, 1989, Phys. Rev. B 40, 6591.
Kaldis, E., E. Boroeh and M. Tellefsen, 1987, J. LessCommon Met. 129, 57.
Kamitakahara, W.A., and R.K. Crawford, 1982, Solid
State Commun. 41,843.
Khatamian, D., and ED. Manchester, 1988, Bull. Alloy
Phase Diagrams 9, 252.
Khatamian, D., and ED. Manchester, 1990, Bull. Alloy
Phase Diagrams 11, 90.
Khatamian, D., C. Stassis and B.J. Beaudry, 1981,
Phys. Rev. B 23, 624.
Klavins, E, R.N. Shelton, R.G. Barnes and B.J.
Beaudry, 1984, Phys. Rev. B 29, 5349.
Knappe, R, and H. Mller, 1986, J. Less-Common
Met. 124, 263.
Knorr, K., B.E.E Fender and W. Drexel, 1978, Z. Phys.
B 30, 265.
Kulikov, N.J., 1985, J. Less-Common Met. 107, 111.
Lsser, R., B. Lengeler, K.A. Gschneidner Jr and E
Palmer, 1979, Phys. Rev. B 20, 1390.
Lea, K.R., M.J.M. Leask and W.E Wolf, 1962, J. Phys.
& Chem. Solids 23, 1381.
Leisure, R.G, R.B. Schwarz, A. Migliori, D.R.
Torgeson, I. Svare and I.S. Anderson, I993, in: Proe.
Int. Conf. on Metal-Hydrogen Systems, Uppsala,
Sweden, 1992, Z. Phys. Chem. NF 179, 359.
Libowitz, G.G., 1972, Ber. Bunsenges. Phys. Chem.
76, 837 and rel. therein.
Libowitz, G.G., and A. Maeland, 1979, Hydrides,
in: Handbook on the Physies and Chemistry of
Rare Earths, Vol. 3, eds. K.A. Gschneidner Jr and
L. Eyring (North-Holland, Amsterdam) ch. 26.
Libowitz, G.G., and J.G. Pack, 1969, 3. Chem. Phys.
50, 3557.
Lichty, L.R., R.J. Sehoenberger, D.R. Torgeson and
R.G. Barnes, 1987, J. Less-Common Met. 129, 31.
Liehty, L.R, J.W. Han, R. Ibanez-Meier, D.R.
Torgeson, R.G. Bames, E.EW. Seymour and C,A.
Sholl, 1989, Phys. Rev. B 39, 2012.
Lueasson, A., R Vajda, J.E Burger and J.N. Daou,
1985, Solid State Commun. 54, 807.
Maeland, A.J., 1970, J. Chem. Phys. 52, 3952.
Mansmann, W., and W.E. Wallace, 1964, J. Phys.
(Paris) 25, 454.

290

R VAJDA

Messer, C.E., and M.K. Park, 1972, J. Less-Common

Met. 26, 235.
Metzger, T.H., P. Vajda and J.N. Daou, 1985, Z. Phys.
Chem. NF 143, 129.
Mintz, M.H., D. Hierschler and Z. Hadari, 1976,
J. Less-Common Met. 48, 241.
Mller, H., P. Knappe and O. Greis, 1979, Z. Phys.
Chem. NF 114, 45.
Mustaehi, A., 1974, J. Phys. & Chem. Solids 35,
1447.
Ohki, C., M. Tada, Y.C. Huang, H. Uchida and H.H.
Uchida, 1984, J. Less-Common Mer. 103, 103.
Ohki, C., H. Uchida and Y.C. Huang, 1989, Z. Phys.
Chem. NF 163, 149.
Opyrchal, J., and Z. Bieganski, 1978, Solid State
Commun. 26, 965.
Parshin, P.E, M.G. Zemlyanov, M.E. Kost, A.Yu.
Rumyantsev and N.A. Chemoplekov, 1978, Inorg.
Mater. 14, 1288.
Pebler, A., and W.E. Wallace, 1962, J. Phys. Chem.
66, 148.
Phua, T.T., B.J. Beaudry, D.T. Peterson, D.R. Torgeson,
R.G. Barnes, M. Belhoul, G.A. Styles and E.EW.
Seymour, 1983, Phys. Rev. B 28, 6227.
Pleschiutschnig, J., W. Schwarz, O. Blaschko, P. Vajda
and J.N. Daou, 1991, Phys. Rev. B 43, 5139.
Rush, J.J., H.E. Flotow, D.W. Connor and C.L. Thaper,
1966, J. Chem. Phys. 45, 3817.
Saw, C.K., B.J. Beaudry and C. Stassis, 1983, Phys.
Rev. B 27, 7013.
Schefer, J., P. Fischer, W. Hlg, J. Osterwalder, L.
Schlapbach and LD. Jorgensen, 1984, J. Phys. C
17, 1575.
Schlapbach, L., 1992, Surface properties and activation, in: Hydrogen in Intermetallic Compounds II,
ed. L. Schlapbaeh L., Vol. 67 of Topics in Applied
Physics (Springer, Berlin) eh. 2.
Schlapbach, L., H.R. Ott, E. Felder, H. Rudigier, P.
Thiry, J.E. Bonnet, Y. Petroff and J.P. Burger, 1986,
J. Less-Common Met. 130, 239.
Schlapbach, L., J.P. Burger, J.E. Bonner, P. Thiry and
Y. Petroff, 1987, Surf. Sci. 189/190, 747.
Schmitzer, C., G. Hilscher, P. Vajda and J.N, Daou,
1987, J. Phys. F 17, 865.
Schreiber, D.S., and R.M. Cotts, 1963, Phys. Rev. 131,
1118.
Senoussi, S., J.N. Daou, P. Vajda and J.P. Burger, 1987,
J. Less-Common Met. 130, 55.
Shaked, H., D.G. Westlake, J. Faber and M.H. Mueller,
1984, Phys. Rev. B 30, 328.

Shaltiel, D., J.P. Burger, J.N. Daou, P. Vajda and A.

Grayevsky, 1991a, Phys. Rev. B 43, 6022.
Shaltiel, D., H. Winter, E. Dormann, J.P. Burger, J.N.
Daou, P. Vajda and A. Grayevsky, 1991b, J. LessCommon Mer. 172-174, 293.
Shinar, J., B. Dehner, B.J. Beaudry and D.T. Peterson,
1988, Phys. Rev. B 37, 2066.
Shinar, J., B. Dehner and BA. Beaudry, 1989, Z. Phys.
Chem. NF 163, 591.
Shinar, J., B. Dehner, R.G. Barnes and B.J. Beandry,
1990, Phys. Rev. Lett. 64, 563.
Smimov, I.A., L.S. Parfen'eva, T.B. Zhukova, Kh.M.
Kholmedov, V.S. Oskotski, I.N. Kulikova, V.A.
Shaburov, L.G. Karpukhina, H. Drulis, M. Drulis
and W. Iwasieczko, 1992, Sov. Phys. Solid Stare 34,
281.
Stierman, R.J., and K.A. Gschneidner Jr, 1984,
J. Magn. & Magn. Mater. 42, 309.
Stuhr, U., D. Steinbinder, H. Wipf and B. Frick, 1993,
in: Proc. Int. Conf. on Metal-Hydrogen Systems,
Uppsala, Sweden, 1992, Z. Phys. Chem. NF 181,
1.
Sturdy, G.E., and R.N.R. Mulford, 1956, J. Am. Chem.
Soc. 78, 1083.
Subramanian, RR., and J.E Smith, 1982, J. LessCommon Met. 87, 205.
Switendick, A.C., 1970, Solid State Commtm. 8,
1463.
Switendick, A.C., 1971, Int. J. Quantum Chem. 5,
459.
Switendick, A.C., 1972, Ber. Bunsenges. Phys. Chem.
76, 535.
Thome, D.K., K.A. Gsehneidner Jr, G.S. Mowry and
J.E Smith, 1978, Solid State Commun. 25, 297.
Titcomb, C.G., A.K. Cheetham and B.E.E Fender,
1972, J. Phys. C 7, 2409.
Toguchi, K., M. Tada and Y.C. Huang, 1982, J. LessCommon Met. 88, 469.
Torgeson, D.R., J.W. Hart, C.T. Chang, L.R. Lichty,
R.G. Bames, E.EW. Seymour and G.W. West, 1989,
Z. Phys. Chem. NF 164, 853.
Udovic, T.J., J.J. Rush, I.S. Anderson and R.G. Barnes,
1990, Phys. Rev. B 41, 3460.
Udovic, T.J., JJ. Rush, N.E Berk and I.S. Anderson,
1992, Phys. Rer. B 45, 12573.
Udovic, T.J., J.L Rush, N.E Berk, I.S. Anderson, J.N.
Daou, P. Vajda and O. Blaschko, 1993, in: Proc.
Int. Conf. on Metal-Hydrogen Systems, Uppsala,
Sweden, 1992, Z. Phys. Chem. NF 179, 349.
Vajda, P., and J.N. Danu, 1984, J. Less-Common Met.
101, 269.

HYDROGEN IN RARE-EARTH METALS, INCLUDING RH2x PHASES

Vajda, R, and J.N. Daou, 1991, Phys. Rev. Lett. 66,
3176.
Vajda, P., and LN. Daou, 1992a, Phys. Rer. B 45,
9749.
Vajda, P., and J.N. Daou, 1992b, Mod. Phys. Lett. B
6, 251.
Vajda, R, and J.N. Daou, 1993a, in: Proc. Int. Conf. on
Metal-Hydrogen Systems, Uppsala, Sweden, i992,
Z. Phys. Chem. NF 179, 403.
Vajda, R, and J.N. Daou, 1993b, The rare-earth
hydrogen systems, in: Metal-Hydrogen Systems,
Vol. 1, eds A. Aladjem and RA. Lewis (VCH,
Weinheim) ch. 3a.
Vajda, R, and J.N. Daou, 1994, Phys. Rev. B, to be
published.
Vajda, E, J.N. Daou and P. Moser, 1983a, J. Phys.
(Paris) 44, 543.
Vajda, R, J.N. Daou, E Radhakrishna and G. Chouteau,
1983b, J. Phys. F 13, 2359.
Vajda, E, J.N. Daou, J.E Burger and A. Lucasson,
1985, Phys. Rer. B 31, 6900.
Vajda, E, J.N. Daou, J.R Burger, K. Kai, K.A.
Gsehneidner Jr and B.J. Beaudry, 1986, Phys. Rev.
B 34, 5154.
Vajda, E, J.N. Daou and J.R Burger, 1987a, Phys. Rev.
B 36, 8669.
Vajda, R, J.N. Daou, J.R Burger, C. Schmitzer and G.
Hilscher, 1987b, J. Phys. F 17, 2097.
Vajda, E, J.N. Daou, A. Lucasson and J.R Burger,
1987c, J. Phys. F 17, 1029.
Vajda, R, J.N. Daou and J.E Burger, 1989a, Z. Phys.
Chem. NF 163, 637.
Vajda, R, J.N. Daou and J.R Burger, 1989b, Phys. Rev.
B 40, 500.
Vajda, R, J.N. Daou, J.R Burger, G. Hilseher and N.
Pillmayr, 1989e, J. Phys. Cond. Matter 1, 4099.
Vajda, R, J.E Burger and J.N. Daou, 1990a, Europhys.
Lett. 11,567.
Vajda, R, J.N. Daou, R Moser and R Remy, 1990b, J.
Phys.: Condens. Marter 2, 3885.

291

Vajda, R, J.N. Daou and J.R Burger, 1991a, J. LessCommon Met. 172-174, 271.
Vajda, R, J.N. Daou and J.R Burger, 1991b, J. Phys.
Cond. Matter 3, 6267.
Vajda, R, J.N. Daou, R Moser and R Remy, 1991c, J.
Less-Common Met. 173, 522; Solid State Commun.
79, 383.
Vajda, R, J.N. Daou and G. Andr, 1993, Phys. Rer.
B, to be published.
Viallard, R., and J.N. Daou, 1972, in: Hydrogne dans
les Mtaux, Vol. 1 (Science et Industrie, Paris) p.
76.
Volkenshtein, N.V., E.V. Gatoshina, M.E. Kost and T.S.
Shubina, 1983, Phys. Status Solidi b 117, K47.
Vlkl, J., H. Wipf, B.J. Beaudry and K.A. Gschneidner
Jr, 1987, Phys. Status Solidi b, 144, 315.
Vorderwisch, P., S. Hautecler and W. Wegener, 1980,
J. Less-Common Met. 74, 117.
Waibel, F., A. Strauss, W. Potzel, EE. Wagner and G.
Wortmarm, 1980, J. Phys. (Paris) 41(Suppl.), C1231.
Weaver, J.H., R. Rosei and D.T. Peterson, 1979, Phys.
Rev. B 19, 4855.
Wegener, W., P. Vorderwisch and S. Hautecler, 1980,
Phys. Status Solidi b 98, Kl71.
Wiesinger, G., and G. Hilscher, 1991, Magnetism of
hydrides, in: Handbook of Magnetic Materials, Vol.
6, ed. K.H.J. Buschow (North-Holland, Amsterdam)
ch. 6.
Zarnir, D., R.G. Barnes, N. Salibi, R.M. Cotts, T.T.
Phua, D.R. Torgeson and D.T. Peterson, 1984, Phys.
Rev. B 29, 61.
Zogal, O.J., 1987, J. Less-Common Met. 130, 187.
Zogal, O.J., and R UHritier, 1991, J. Alloys &
Compounds 177, 83.
Zogal, O.J., Ch. Jger, H. Dhler and B. Schnabel,
1984, Phys. Status Solidi a 82, K153.

Handbook on the Physics and Chemistry of Rare Earths

VoL 20
edited by K.A. Gschneidner, Jr. and L. Eyring
1995 Elsevier Science B.V.. All rights reserved

Chapter 138
MAGNETIC PROPERTIES OF
INTERMETALLIC COMPOUNDS*
D. G I G N O U X

a n d D. S C H M I T T

Laboratoire de Magndtisme Louis N~el, C.N.R.S., B P 166,

38042 Grenoble Cddex 9, France

Contents
List of symbols and abbreviations
1. Introduction
2. 3d Magnetism
2.1. Onset of magnetism in Co- and Nibased alloys
2.1.1. Collective electron
metamagnetism (CEM)
2.1.1.1. The WohlfarthRhodes model
2.1.1.2. RCo2
2.1.1.3. ThCo 5
2.1.1.4. Further CEM
systems
2.1,1.5, Strong
magnetoelastic
effects
2.1. t .6. Spin fluctuation
effects
2.1.1.7. A new model
of CEM
2.1.2. Very weak itinerant
ferromagnetism (VWIF)
2.1.3. Cobalt antiferromagnetism
2.2. 3d Magnetocrystalline anisotropy
2.2.1. Experimental characteristics
2.2.2. Theoretical interpretations
2.3. Instability and frustration of Mn
magnetism in RMn 2 compounds
2.3.1. Topological frustration
2.3.2. Mn moment instability and
complex magnetic structures

294
295
296
297
298
298
299
301
302

303
304
306
308
309
310
310
313
316
316
317

* In memory of Remy Lemaire.

293

2.3.3. Large Mn anisotropy

2.3.4. Giant spin fluctuations
2.3.5. Theoretical approaches
2.4. Lanthanide-3d transition-metal
compounds where both carry a well
defined magnetic moment
2.4.1. General characteristics
2.4.2. Curie temperatures and
3d-4f exchange interactions
2.4.3. High-field magnetization
processes, 3d-4f interaction
and magnetocrystalline
anisotropy
3. 4f Magnetism
3.1. Crystal-field and exchange interactions
in ferromagnetic compounds
3.2. Metamagnetism and associated phase
diagrams
3.2.1, General considerations
3.2.1.1. Demagnetizing
field effects and
hysteresis
3.2.2. CEF metamagnetic systems
3.2.3. Quadrupolar metamagnetic
systems
3.2.4. Ferromagnetic
metamagnetic systems
3.2.5. Weakly anisotropic
metamagnetic systems

322
323
325

327
327
328

331
337
338
345
345

347
350
355
359
363

294

D. GIGNOUX and D. SCHMITT

3.2.6. Spin-flip metamagnetic

systems: simple
antiferromagnets
3.2.7. Modulated metamagnetic
systems
3.2.8. Spin-flip metamagnetism:
long-period commensurate
systems
3.2.9. Spin-slip metamagnetism:
long-period commensurate
systems
3.2.10. Complex multistep
metamagnetism: long-period
commensurate systems
3.2,11. Planar metamagnetic
systems
3.2.12. Multiaxial metamagnetic
systems
3.3. Quantitative analysis of
incommensurate magnetic systems
4. Summary and conclusion
5. Appendix: Definitions and/or descriptions of
magnetic terms or phenomena
5.1. Antiferroquadrupolar and
ferroquadrupolar ordering
5.2. Amplitude-modulated structure

5.3. Antiphase structure

5.4. Collinear and noncollinear structures
5.5. Commensurate and incommensurate
structures
5.6. Equal-moment structure
5.7. Exchange interaction J(/j) and Fourier
transform J(q)
5.8. Fan structure
5.9. Flopside structure
5. i 0. Frustration
5.11. Helical structure
5.12. Helifan
5.13. Multiaxial structure
5.14. Multistep metamagnetic process
5.15. Multi-Q structure (double, triple. . . . )
5.16. Quadrupolar moment
5.17. RKKY (Ruderman, Kittel, Kasuya,
Yosida) exchange interaction
5.18. Single-Q structure
5.19. Spin-flip transition
5.20. Spin-flop transition
5.21. Spin fluctuations
5.22. Spin-slips (or spin discommensurations)
5.23. Spin-reorientation transition
References

367
369

373

378

380
390
394
399
406
408
408
408

409
409
409
410
410
410
411
411
412
413
413
413
413
413
414
415
415
415
416
416
417
417

List of symbols and abbreviations

a, b, e
a*, b*, c*
AM
B
B~
C phase
CEF

basis vectors of the unit cell

basis vectors of the reciprocal unit
cell
amplitude modulated
magnetic induction
crystal field parameters
commensurate phase
crystalline electric field

CEM
EM
F
FOMR
Gj

collective electron metamagnetism

equal moment
free energy
first-order magnetic reorientation
tetragonal quadrupolar parameter

G2
gj
H

trigonal quadrupolar parameter

Land~ factor
magnetic field

7-(
I phase
INS

Hamiltonian
incommensurate phase
inelastic neutron scattering

J
J1

Jz
J(ij)
J(q)
J(0)
kB
K
K~
L
LMTO
M
M

total kinetic moment operator

exchange interaction between
nearest neighbours
exchange interaction between
second-nearest neighbours
exchange interaction between
moments i andj
Fourier transform of the exchange
interactions
paramagnetic exchange parameter
Boltzmann constant
reciprocal lattice vector
second-order anisotropy constant of
tmiaxial systems
orbital kinetic moment operator
linear muffin tin orbital
chemical symbol for 3d transition
metals
magnetization

MAGNETIC PROPERTIES OF INTERMETALLICCOMPOUNDS

Ms
M-NM
n, n
n(/?F)

Na
N
NMR
O~
PF model
Q
q
R
Ri
RKKY
RPA
S
S
SCR
Tc

component of the total kinetic

Tt~

moment

TQ

magnetic-nonmagnetic
bilinear exchange coefficient
density of states at the Fermi level
demagnetizing field factor
number of moments per magnetic
unit cell
nuclear magnetic resonance
Stevens operators
periodic field model
propagation vector of magnetic
structure
vector of the reciprocal space
chemical symbol for rare earth,
lanthanide
position of atom i
Ruderman-Kittel-Kasuya-Yosida
random phase approximation
Stoner enhancement factor
spin kinetic moment operator
self-consistent renormalization
Curie temperature

TR
TsF
Tt, TI, T2, ...
TEC
U
VWIF
W
Fi

y
eF
)~
~
#B
~tSR
r
X
Z~)

295

Nrel temperature
quadrupolar ordering temperature
spin reorientation temperature
spin fluctuation temperature
transition temperatures
thermal expansion coefficient
Coulomb repulsion or exchange
energy
very weak itinerant ferromagnetism
bandwidth
irreducible representation of a
symmetry group
electronic specific heat coefficient
Fermi energy
spin orbit coupling constant
wavelength
Bohr magneton
muon spin resonance
incommensurate component of the
magnetic propagation vector
magnetic susceptibility
third-order paramagnetic
susceptibility

1. Introduction
This chapter is devoted to the magnetic properties o f rare earth intermetallic compounds
investigated during the last fifteen years. The earlier works in this field have been
described by Kirchmayr and Poldy (1979) in a previous chapter o f this Handbook
series (volume 2, chapter 14), During the last 15 years, research in magnetism can be
characterized by a b o o m in the field o f rare-earth (R)-based materials, in particular
the metallic ones. Currently, rare earth intermetallics are in a prominent situation not
only from a fundamental point o f view but also for the large number o f technological
applications, in particular in the field o f permanent magnets. Rare-earth intermetallics play
an important role in a large range o f current research fields, in particular those devoted
to heavy fermions, valence fluctuations, Kondo lattices, magnetostrictive materials,
permanent-magnet materials, spin glasses and random anisotropy systems. Since these
aspects o f magnetism in rare-earth intermetallics are treated elsewhere, they will not
be discussed here. In this chapter we are mainly concerned with the basic properties o f
intermetallic compounds with normal lanthanides, i.e. those with a well localized 4 f shell.
This means that most o f the Ce and Yb materials are excluded. Furthermore, rather than
giving an exhaustive report on magnetic properties with many physical values reported

296

D. GIGNOUXand D. SCHMITT

in tables, we prefer to emphasize what we consider to be the major steps in this field of
research during the period under consideration.
This chapter is divided into two major parts. The first part is devoted to systems in
which the obtained results led to a better knowledge of 3d magnetism. Four major aspects
will be considered. First (sect. 2.1), the onset of magnetism in Co and Ni, in which
the major part is devoted to the systems where collective electron metamagnetism has
been observed. The understanding of the large 3d magnetocrystalline anisotropy of many
intermetallic compounds is treated in sect. 2.2. In the cubic Laves phase RMn2, the
instability of Mn magnetism together with the topological frustration of antiferromagnetic
interactions lead to quite original properties which are at the origin of new physical
concepts and theoretical approaches. The magnetic properties of these fascinating
compounds are presented in sect. 2.3. The last section of this part is devoted to rareearth-3d transition-metal compounds where both carry a well defined magnetic moment.
These compounds generally are excellent materials for permanent-magnet applications,
but only the last results obtained concerning their intrinsic properties are presented.
The second part of this chapter concerns the magnetic properties of compounds in
which only the lanthanide atom is magnetic. After a description (sect. 3. I) of recent
progress made in the quantitative knowledge of the main interactions (exchange and
crystal field) in the small number of lanthanide-based series which are ferromagnetic,
the main purpose of this part is devoted to the compounds which exhibit metamagnetic
processes of quite different origins (sect. 3.2). Because of the long range and oscillatory
character of the indirect RKKY exchange interaction the majority of these compounds
order antiferromagnetically with complex magnetic-field-temperature phase diagrams
often characterized by the competition between commensurate and incommensurate
magnetic structures. Dramatic progress has been made in the knowledge of these systems
during the last decade due to the improvement of experimental devices, the increasing
number of single crystals of good quality and the evolution of theoretical models. Parallel
to the experimental advances in this field, theoretical models have been proposed to
quantitatively analyze these complex phase diagrams, in particular the incommensurate
magnetic systems, which is the purpose of the last section (sect. 3.3) of this part.
In this chapter a large number of experimental results and theoretical approaches,
already presented in previous review papers, are assumed to be known. The reader will
find useful information in the reports on rare-earth-based intermetallic compounds by
Buschow (1977a, 1979, 1980, 1988) and by Kirchmayr and Poldy (1979).

2. 3d Magnetism
The R-M systems, where M is a 3d transition metal, form an outstanding tool for the study
of 3d band magnetism and in particular the interactions, instabilities and anisotropies of
such magnetism. In the majority of cases, for a given M element, a series of compounds
with different rare earths crystallize in the same crystallographic structure and thus have
practically the same band structure. It is then possible to study the 3d magnetism under

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

297

several conditions depending on the rare earth (nonmagnetic or magnetic, isotropic or

not, different sign of the magnetocrystalline anisotropy parameters . . . . ). Many studies
are devoted to these systems and it is impossible to describe all the results obtained. We
focus on the most original advances in 3d magnetism discovered in R-M intermetallic
compounds during the last decade.
2.1. Onset of magnetism in Co- and Ni-based alloys
These compounds are formed by the association of the 3d band of the M element with
the 5d band of R (4d for Y) with higher energy. The electronegativity difference between
the constituents causes a transfer of 5d (4d) electrons towards the unfilled 3d band.
Since the screening of the nuclear potentials by the electrons is modified, the two bands
approach each other leading to 3d-5d (or 3d-4d) hybridized states (Cyrot and Lavagna
1979, Shimizu et al. 1984). The Fermi level of the compounds often lies in this region of
the density of states. This itinerant description of 3d magnetism is the most appropriate
for Co and Ni in which, due to the width of the 3d band, the Un(ev) product (U is the
Coulomb repulsion or exchange energy between up and down spins, and n(eF) the density
of states at the Fermi level) is smaller than unity (with Mn and Fe this ratio is closer to
unity and accordingly magnetism is more localized). Starting from pure Ni or Co, the
progressive increase of the R percentage leads first to a decrease of the density of states
at the Fermi level n(eF). For a critical concentration range (around RCo2 for cobalt and
RNi5 for nickel) alloys are close to the conditions required for the onset of magnetism
(Stoner criterion) and magnetic instabilities can be observed, each behaviour strongly
depending on the fine structure of n(e) near eF. However, resurgence of 3d magnetism
appears for a slightly larger R amount and then disappears altogether as shown in fig. 1.
e--

o~ 6~-e,l
(.9
~.-__

1.5 "

.-~

go
I

Yxcol-x

,-~
~,',ec

~r
v

,
1.0

LaxCOl_ x

YxN,l_ x

g,

Iii

k)

~+I om
'I ~

...1

g
+'

t~

5-

c,4

I
+

0.5

~;
~,

'
t,
t--,

0.1

0.2

~
~

o
c9

0.3

0.4

0'.5 x ~-

Fig. 1. Mean value of the 3d moment as a

function of the rare earth concentration in the
compounds of the La-Co, Y-Co and Y-Ni
systems (Gignoux and Schmitt 1991).

298

D. GIGNOUX and D. SCHMITT

Three types of characteristic behaviours are mainly observed for R concentrations near
or larger than the critical ones, namely collective electron metamagnetism (CEM), very
weak itinerant ferromagnetism (VWIF) and Co antiferromagnetism.
2.1.1. Collective electron m e t a m a g n e t i s m (CEM)
2.1.1.1. The Wohlfarth-Rhodes model. Collective electron magnetism (CEM), predicted
in 1962 by Wohlfarth and Rhodes (1962), refers to the transition from a nonmagnetic to a
magnetic state when the field acting on the band is larger than a critical value HM. Using
a Landau-type expansion of the magnetic free energy of the d-electrons, such as
F = AM 2 + BM 4 + CM 6 + ....

(1)

MH,

the theory led to the following expressions for the first lower-order coefficients

1
(2)

A - 4n,gF,---~
o,ot)

1
B-=-64n3(eF)

2]
3n(eF------)- \ n(eF) J J '

(3)

where S = (1 - Un(eF)) -1 is the Stoner enhancement factor, n, n' and n" are the density
of states and its first and second derivatives at the Fermi level. CEM occurs when the
magnetization dependence of F has the upper variation shown in fig. 2a in zero field and
it becomes the lower curve of this figure above the critical field HM. The corresponding
expected low-temperature magnetization curve is shown in fig. 2b. The above formulae
show that such a behaviour can occur when: (1) A is weakly positive, i.e. when the Stoner
criterion for the onset of ferromagnetism is almost satisfied, and (2) B is negative, which
implies another minimum for a nonzero value of M. This latter condition requires n"(eF)
to be large enough, which means that the density of states at the Fermi level has a strong
positive curvature.
The system is now a Pauli paramagnet in low field and the thermal and field
dependences of its susceptibility are given by
X = SXo

[,7~2(.BH(EF) Rt(EF)2)
1 - --~S

T2+

S3(Bt"(EF)_nt(EF)2"~H2 ]
+ ...

(4)

This formula shows that a maximum in the thermal variation of the susceptibility, having
the same origin, is also predicted (fig. 3b). Because of this increase of the susceptibility
with temperature, the shape of the magnetization curves changes with temperature;
the magnetization discontinuity tends to decrease and even to disappear above a given
temperature. The high magnetization state is then reached continuously (fig. 3a).
This behaviour, which was assumed only to occur on Co in RCo2, but had not been
directly observed until the late seventies, has since clearly shown up in several rareearth-transition-metal alloys. The best examples are RCo2 and ThCo5 which are presented
below.

MAGNETIC PROPERTIESOF 1NTERMETALLICCOMPOUNDS

z~F (a)

299

(b)

H=0

HE
i

/M=
H
H~

HM

H~

Fig. 2. Collective Electron Metamagnetism. (a) Variation of the difference between the free energy of the
ferromagnetic and paramagnetic states as a function of magnetization at different fields; (b) variation of
magnetization with increasing field (Barbara et al. 1988).

M~ (a)

HM

I~

(b)

T2

Fig. 3. Schematic representation of the Collective Electron

Metamagnetism. (a) M vs. H at various temperatures; Co)thermal
variation of the initial susceptibility (Gignoux et al. 1983).

2.1.1.2. RCo2. As shown in fig. 1, RCo2 compounds are at the limit o f the onset
o f Co magnetism. In these cubic Laves-phase compounds Co atoms belong to one
crystallographic site. With magnetic rare earths the compounds are ferromagnetic (with
light lanthanides) or ferrimagnetic (with heavy lanthanides) and, below the Curie
temperature, Co is magnetic with a moment close to I#B. Conversely in YCo2 and
LuCo2, Co is nonmagnetic (Lemaire 1966). These latter compounds are enhanced Pauli

300

D. GIGNOUX and D. SCHMITT

20

A
18
ol

YCo2
...........
v- 12
,.J

-~ ~o
n

U8
~

,6o

26o

Fig. 4. Thermal variationof the susceptibility of YCo2 and LuCo2 (Gignoux

et al. 1983).

36o

TEMPERATURE (K)

Y(CI"xAtx)2
0.5

4.2K

'

,O0'6
..~

t/oo, I/t=
10

20
(T)

LuCU

J r-,..

~ ~ - - ~ " 1

LJ

'

30

Fig. 5. Magnetization curves of Y(Col_xAlx)2 at 4.2 K

(Sakakibara et al. 1987).

T=IO K

0,

z,o

20

40

60

80

100

H(T)

Fig. 6. Magnetization curves of YCo2 and LuCo2 at 10 K in pulsed ultra-high magnetic field up to 94 T
(Gignoux and Schmitt 1991, after Goto et al. 1990). The magnetizationdata measured in a long pulse field are
also plotted as dots.
paramagnets but the field and thermal effects indicate the possibility o f CEM. Indeed, in
YCo2 and LuCo2 the susceptibility exhibits a broad m a x i m u m around 230 K and 370 K,
respectively (fig. 4). Moreover, at 4.2 K the superimposed susceptibility o f YCo2 increases
by about 20% between 0 and 35 T (Bloch et al. 1975), whereas in LuCo2 this effect is
much smaller (Schinkel 1978). Actually, CEM was not observed because this m a x i m u m
magnetic field was smaller than the critical field HM o f the metamagnetic transition. With

MAGNETICPROPERTIESOF 1NTERMETALLICCOMPOUNDS

301

magnetic rare earths the high magnetization state was reached owing to the molecular
field contribution of the rare earth. Moreover, the first-order transition observed at the
Curie temperature in some of these compounds, namely DyCo2, HoCo2 and ErCo2, has
been ascribed to the collapse of the Co magnetic moment at this temperature (Lemaire
1966, Petrich and M5ssbauer 1968, Givord and Shah 1972). In 1977, polarized neutron
diffraction studies on TmCo2 and HoCo2 carried out by Gignoux et al. (1977) showed
that HM should be smaller than 100T (around 70T with Tin). From a theoretical point of
view, the first calculations (Bloch et al. 1975) led to a much larger value of Hr~ (142 T).
Later, more realistic band structure calculations led to H ~ values around 80 T (Cyrot
et al. 1979, Yamada et al. 1985, 1987, Yamada and Shimizu 1985, 1990). In order to
directly observe the metamagnetic transition in this system, a large effort was undertaken
to depress the critical field by substitution effects, and in 1987 this was obtained by
substituting a small amount of A1 for Co in YCo2 and LuCo2 (Sakakibara et al. 1987)
where the transition occurs below 40 T (fig. 5). In the Y(Col-xFex)2 system, although
less dramatic, the CEM has been observed for 0.04 ~<x <~0.07 (Yoshimura et al. 1987).
Finally, the most dramatic event in this fascinating story occurred quite recently thanks to
the availability of magnetic measurement up to 94 T: as shown in fig. 6 the metamagnetic
transition was directly observed in YCo2 and LuCo2 around 70T (Goto et al. 1990) in
rather good agreement with the values calculated and those experimentally determined
with a magnetic rare earth.
2.1.1.3. ThCo5 (Givord et al. 1979). As shown previously with trivalent rare earths, the
critical concentration for the onset of Co magnetism is achieved in the RCo2 compounds,
but this limit is close to the RCo5 composition with tetravalent R such as Th or
Ce. Although Th is not a rare earth, ThCo5 has the same hexagonal CaCus-type
crystallographic structure as the RCo5 series. In these compounds Co atoms are located
at two different crystallographic sites. Only those at the 2c site are ferromagnetic. Within
the homogeneity range of the ThCo5 phase, the magnetic properties strongly depend on
the composition. The Curie temperature increases from 410K for ThCo5 to 730K for
Th0.92Co5.16 (Van der Goot and Busehow 1971). As sketched in fig. 7c (lower part), in
the Co-poor compounds, the temperature dependence of the spontaneous magnetization
and that of the superimposed susceptibility show a maximum near 200 K (Ganapathy et
al. 1974). Below 100K, a transition towards a state of higher magnetization is induced by
a large applied field (fig. 7c, upper part). Above the transition, the magnetization reaches
a value close to that of the spontaneous magnetization in the Co-rich compounds. Study
of these two magnetization states by polarized neutron diffraction (Givord et al. 1977,
1979) has shown that: (i) in the low magnetization state, the magnetic moment is lower
by about 20% on the Co3g atoms (0.96gB) than on the Co2c ones (1.21/~B); (ii) in the
state of high magnetization, magnetic moments are the same on both sites (1.58/~B); and
(iii) on the crystal of low spontaneous magnetization, i.e. the compound with the exact
ThCo5 stoichiometry, the maximum of the temperature dependence of the spontaneous
magnetization and superimposed susceptibility originate from the Co3g atoms. It is
therefore concluded that these atoms exhibit CEM.

302

D. GIGNOUX and D. SCHMITT

M~ 4.2K
0

(a)

M~,~

appliedfield

total field

HA

(b)

HT

~r

4.2

i Hc--HM-HE
applied
field

HA

(c)

Fig. 7. Schematic field and temperature dependences of the spontaneous magnetization M s and the superimposed
susceptibility of (a) Co 2c atoms, (b) Co 3g atoms and (c) ThCo5. Abbreviations: HA, applied field; HE,
exchange field on the Co 3g atoms due to the Co 2c atoms; HT=HB+HA, the total field acting on Co
3g atoms; HM, the critical field for metamagnetism for Co 3g atoms; Hc, experimental transition field in
ThCo5 (Gignoux et al. 1983).

The magnetic properties of ThCo5 are schematized in fig. 7. The Co2c atoms are
ferromagnetic, they exhibit a spontaneous magnetization. The superimposed susceptibility
is almost constant below room temperature and close to that in YCo5 (fig. 7a). Co3g atoms
exhibit a large enhanced paramagnetism with a susceptibility maximum around 200 K and
are close to the condition for CEM. Magnetization, which is induced on these Co3g atoms
by the exchange field HE arising from Coac atoms, then exhibits a maximum at the same
temperature as the susceptibility (fig. 7b). The exchange field is slightly lower than the
critical field for CEM. The latter can be reached in the Co-poor compounds in an applied
field (as is observed in the stoichiometric compound), and in the Co-rich compounds
such as Th0.92C05.16 by an increase of the exchange field on the Co3g atoms, caused by
statistical substitutions of Co dumbbells on Th atoms, as usually occurs in the RCo5
system.
2.1.1.4. Further CEM systems. Within the homogeneity range of the CeCo5 phase, in
which Ce is tetravalent, Co atoms are in a high magnetization state. The exchange
interactions can be reduced by replacing the cobalt atoms by nonmagnetic nickel atoms
and in Ce(Co0.93Ni0.o7)5 the field dependence of magnetization exhibits the same transition
at low temperature as in ThCo5 as a result of the CEM of one part of the Co atoms (Givord
et al. 1983).
Figure 1 shows that Ni magnetism falls rapidly with the R content. It is then not
surprising that within their homogeneity range, the magnetic properties of the weak
Y2Nil7 ferromagnet strongly depend on the stoichiometry. In particular in Y2Ni16

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

303

the thermal variation of the spontaneous magnetization and the low-temperature field
dependence of the magnetization, although less pronounced, are analogous to those of
ThCo5 showing a CEM behaviour of one part of Ni atoms (Gignoux et al. 1980a).

2.1.1.5. Strong magnetoelastic effects. Besides their magnetic properties, large magnetoelastic effects associated with the transition between magnetic and nonmagnetic states of
Co have been observed in RCo2 and derived compounds. In the early seventies, large
and negative pressure effects were measured on the Curie temperatures of magnetic
RCo2 (Bloch et al. 1971, Voiron and Bloch 1971) and on the susceptibility of YCo2
(Voiron et al. 1973). As well, all the RCo2 compounds show an Invar-like thermal
expansion anomaly below their Curie temperature and the large observed positive
magnetic contribution to the spontaneous volume expansion (fig. 8) was found to be
proportional to the square of the Co magnetic moment (Minakata et al. 1976, Lee and
Pourarian 1976). Such a M:(Co) dependence of the spontaneous volume magnetostriction
was also observed in Gdl-xYxCo2 (Muraoka et al. t983, 1984) and Gd(Col_xNix)2
(Tanaka et al. 1980). Let us quote also that in the Y(Col-xAlx)a compounds with
0.095 ~ x ~<0.12 a large volume change associated with the metamagnetic transition
has been observed (fig. 9). It is also found to be proportional to the square of the
magnetization in both paramagnetic and ferromagnetic regions (Wada et al. 1988).
These magnetovolume effects can be treated theoretically in the framework of the
itinerant electron model and were, up to recently, interpreted as a consequence of the
kinetic energy cost arising from the splitting of the 3d band when passing from the
nonmagnetic to the magnetic state of Co, this energy cost being then minimized by
a volume expansion, the 3d energy bandwidth being highly dependent on the atomic
separation. In a first approximation, the increase in kinetic energy and, accordingly,
the volume change are proportional to the square of the magnetization. Self-consistent
spin-polarized energy band calculations performed by Janak and Williams (1976) led
to an overestimation of the volume change. Later a better account of the experimental
10

Spontaneous Magnetostriction of RCo2

"~..,,~"~-~.

A
B
C
D
E

GdCo2
TbCo 2
Dy Co 2
HoCo 2
ErCo2

100

200
300
T(K)

aO0

500

Fig. 8. Temperaturedependenceof the magnetic

contributionto the spontaneousvolumemagnetostrictionof the RCo2 compounds.The arrows
indicatethe Curietemperatures(Minakataet al.
1976).

304
15

D. GIGNOUX and D. SCHMITT

Y(CI' x AIx)'2

'

4.2K

x ~ = 0.0s5

10

~0
_J
.J 5
x=OA05

i
10

20

0
30

H (T)

Fig. 9. Longitudinal magnetostriction curves of Y(COl_xAlx)2

at 4.2K (Wada et al. 1988).

results has been obtained taking into account spin fluctuations (see Nakamura 1983 and
references therein).
2.1.1.6. Spin fluctuation effects. It is well known that the Stoner (or Hartree-Fock) theory,
which considers only individual excitations of electron-hole pairs with opposite spins,
fails to take into account thermal effects of itinerant electron magnetism. Collective
modes of spin fluctuations, particularly in nearly ferromagnetic (or antiferromagnetic)
systems play a crucial role in thermal properties such as specific heat and magnetic
susceptibility. For this purpose much attention has been paid to the RCo2 compounds to
test the different itinerant-electron theories of magnetism. The first approach, using the
random phase approximation (RPA) pointed out that spin fluctuations in the paramagnetic
phase (paramagnons) induce a logarithmic dependence on temperature of the electronic
specific heat coefficient 7 (Doniach and Engelsberg 1966, Berk and Schrieffer 1966) and
lead to an enhancement of the coefficient of the T 2 term in the magnetic susceptibility
(B6al-Monod and Lawrence 1980). These latter authors introduced a spin-fluctuation
temperature TSF = kBeF/S, where eF is the Fermi energy, to show that the thermal effects
have to appear at much lower temperatures than those expected in a model which takes
into account the individual excitations only. x(T, H) has almost the same dependence as

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

305

in eq. (4) with respect to n", n' and n, but the coefficient of the T2 term is proportional
to S 3 instead of $2:
6

2 n--~F)

l.Z~)

T 2 + ~ C H 2 + ...

(5)

where C is a coefficient which may be different from that in eq. (4) but which also depends
on the band structure (B~al-Monod 1982). This model led to a better simultaneous
quantitative account of field effect at low temperature and the susceptibility maximum
of YCo2 and LuCo2. Later Misawa (1988) has shown that the thermal variation o f z was
proportional to T 2 In T rather than T 2. It was pointed out by Brinkman and Engelsberg
(1968) and by B6al-Monod et al. (1968) that the application of a high magnetic field
quenches the spin fluctuations. This quenching has been evidenced in YCo2 and LuCo2
(fig. 10) from specific measurement where the y coefficient is reduced by a factor of 4%
and 10% respectively in a field of 10 T (Ikeda et al. 1984, Ikeda et al. 1991 and references
therein). Using a self-consistent renormalization (SCR) theory of spin fluctuations, a
unified picture has been proposed of itinerant magnetism ranging from the localized limit,
for which the spin fluctuations are localized in space, and the limit of weak itinerant
ferromagnetism (or antiferromagnetism), for which spin fluctuations are localized in the
reciprocal space (Moriya and Takahashi 1978, Moriya and Hasegawa 1980, Moriya 1991
and references therein). Among the results of this approach one can quote the quantitative
description of the nuclear spin relaxation rate and the high-temperature susceptibility. The
latter follows a Curie-Weiss law without localized moments and the Curie constant, which
is then a measure of the spin fluctuations, is the sum of two contributions, C = Ct + CI,
where Ct corresponds to transverse fluctuations of the localized limit and CI corresponds
to longitudinal fluctuations of the itinerant limit. The ratio between the effective moment
deduced from the Curie constant and the low-temperature saturation moment allows one
to evaluate the preponderant nature of spin fluctuations. High-temperature susceptibility
measurements performed on RCo2 allowed one to observe the Curie-Weiss behaviour
and to derive the effective moments which are reported in fig. 11 as a function of the
14

t3

Lu Co 2

(35.0 ot./o Lu)

,,,~T,~6,~,"'J

o 0.00 T

~,~..~..,e-~

I0

~@li~ i ~ , ~ ^

" 2.50

8
71

x 9.98
i

12

16

20

24

28

52

T2( K2)

36

Fig. 10. Heat capacity of LuC% at five

magnetic fields (Ikeda et al. 1984).

D. GIGNOUXand D. SCHMITT

306

z+ILu~2

l'%fflCo)spin only

73

o.

e_

tlt=
t t
TmCo21
PrCo~ NdCo~ Dy[o~

"~
1

CadCo2

TbCo2

ErCo2 HoCo2

too

200

300

Tc (K)

t,00

Fig. 11. The effectivecobalt moments as a function of Curie temperatures in RCo2 compotmds(Burzo and
Lemaire 1992).
Curie temperature Tc (Burzo and Lemaire 1992). These Co effective moments are well
above the saturation Co magnetic moment (~I#B) and steadily decrease from about 4#B
in YCo2 and LuCo2 to about 2.6/~B in GdCo2 which has the largest Curie temperature.
This decrease has been ascribed to the quenching of spin fluctuations by the internal
field, as does an applied field, leading then to a decrease of the Curie constant. Finally
it is worth noticing that the nuclear magnetic relaxation observed in YCo2 (Yoshimura
et al. 1984) is well accounted for with the SCR theory of Moriya of spin fluctuations in
weakly or nearly ferromagnetic metal.
2.1.1.7. A new model o f CEM. Up until now the CEM, in agreement with the prediction
of Wohlfarth and Rhodes (1962), was ascribed to a special shape of the density of
states near the Fermi level in systems where the Stoner criterion is almost satisfied.
As shown above the large magnetovolume effects were considered as a consequence of
the magnetic-nonmagnetic state transition. Band calculations were then performed in
order to check if in the compounds which exhibit CEM the density of states had the
expected shape (Yamada et al. 1984, 1985, 1987, Yamada and Shimizu 1985). Indeed,
band calculations were performed in the tight binding approximation. From the deduced
density of states and taking into account the effect of spin fluctuations, relevant parameters
were found for RCo2 which fit the thermal variation of the susceptibility of YCo2 and
LuCo2, and which lead to critical fields of the metamagnetic transition close to those
later observed by Goto et al. (1990). Actually, given the rather large number of systems
close to the onset of magnetism and exhibiting CEM, it is a little surprising that, in all

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

307

cases, the density of states presents a peculiar variation near the Fermi level. So in 1992,
Duc et al. (1992) proposed a new approach to CEM. For any shape of the density of
states this phenomenon would simply result from the interplay between the magnetic and
elastic energies. Considering the variation of the bandwidth with volume, it is shown
that the transition towards magnetism is determined by the balance between the elastic
energy, which increases with volume, and the magnetic energy, which decreases when
the volume increases, in the ferromagnetic phase. Assuming an elliptic density of states
n(e) = ( 1 / ~ W 2 ) ~ W -'~ - e 2 (2W =bandwidth) in order to exclude the type of instability
to which CEM was previously ascribed, the total energy in the Stoner model at 0 K is
written as Etot = Eel + Elat, where Eel is the electronic energy,
Eel =

rE+ e n(e) de + rE- e n(e) de UM2

rv

4It~3

MH,

(6)

and Elat the lattice energy:

Elat = Cg2 + K(22,

(7)

with 2 = ( V - Vo)/Vo (Vo being the volume in the paramagnetic state) and K is the
contribution to the elastic constant due to all terms other than the band contribution.
C is determined by the condition

OEel
o ~ para,

~c2--0

+ c = 0.

(8)

In the tight binding approximation the bandwidth decrease with the volume increase is
given by W = W0 exp(-qf2), where q is of the order of 1 to 5/3 for 3d electrons. Under
these conditions, when the filling of the band (about 9.6 3d electrons per atom) is such that
the Stoner criterion is almost satisfied, one can have the variation of Etot with g2 shown in
fig. 12, where the variation of the magnetization is also reported. In the presence of a field
(exchange and/or applied) one obtains the evolution of the total energy shown in fig. 13a.
A metamagnetie transition, associated with a large volume discontinuity, then occurs for
a critical field (fig. 13b). For fillings of the d band smaller or larger than that necessary
for CEM, ferromagnetic or paramagnetic states are stable, respectively. This is the reason
why, for an excess of Co in ThCos, both Co sites are in the high magnetization state,
the filling of the 3d band being smaller than in the stoichiometric compound. This model
gives account for alloying effects in Y(Col-xAlx)2 where the volume increases with the
A1 concentration, leading to a transition from Pauli paramagnetism to ferromagnetism for
a critical concentration x ~ 0.13. At larger AI concentration, ferromagnetism disappears
again due to the dilution of Co atoms by nonmagnetic A1. Note that in this model,
metamagnetism cannot be expected in materials where K is too small or too large. This
has been studied experimentally in Y(CoxAlyCuz)2 where, when ehoosing the A1 vs Cu
concentration in order to maintain the band structure, it is found that by adding Cu

308

D. GIGNOUX and D. SCHMITT

-2.003

0.5

0.4

-2.004

0.3
-2.005

-2.006
-2.007
-0.05

0.2
-Qc

o.1

-OM

I
~/
I
I
o
0.05 0.10 0.15 0.20
2

0.5

-2.002
-2.004
-2.006
~-2.008
-2.010
-2.012
-2.014
-2.016

Fig. 12. Calculated variations of the total energy

and magnetization as a function of volume 2 for
N=9.55, K=2.6eV, U=2.5eV and Wo=5eV
after Due et al. (1992). N, number of d electrons
per atom; K, contribution to the elastic constant
due to all terms except the band contribution; U,
exchange energy; 2Wo, bandwidth at 2=0.

00"20[

0.3

0.12[

9.2

2f

0.1

0
/

0.10

0.15

0.20

0.25

200
j

400

H(T),

200 300
400 500
H (T)
Fig. 13. (a) Calculated total energy for the same parameters as in fig. 12, for different magnetic fields;
(b) calculated curve M(H) for the same parameters (after Due et al. 1992).
0

0.05

11[

b)

3.4

100

the metamagnetic transition disappears (Due, unpublished). This can be attributed to

the decrease o f K since Cu is very stiff. Finally let us note that similar models for
nonmagnetic-magnetic transitions have been proposed for other systems where strong
volume effects are observed, in particular in cerium heavy-fermion compounds (Lavagna
et al. 1983, Ohkawa 1989).

2.1.2. Very weak itinerant ferromagnetism (VWIF)

Whereas YNi5 is a Pauli paramagnet, the resurgence o f 3d magnetism observed in YNi3
(Gignoux et al. 1980b) and Y2Ni7 (Ballou et al. 1990b), as shown in fig. 1, must be
associated with the fine structure o f n(e) in the region o f 3d-4d hybridized states, such
that n(eF) is large enough for the Stoner criterion to be fulfilled. Bulk magnetic properties
are similar to those o f ZrZn2 o f which the behaviour can quite well be described within
the model o f very weak itinerant ferromagnetism (Wohlfarth 1951). The magnetization
is weak and strongly field dependent at all temperatures. The Arrott plots (M 2 versus
H/M) show a linear variation, and the variation o f M2(H = O, T) as a function o f T 2 is
also linear. However, contrary to ZrZn2, magnetization is not diffuse but localized at the
Ni sites, as shown on the magnetization density map in YNi3 determined from polarized
neutron diffraction experiments on a single crystal (fig. 14). This localization difference
originates from the difference in the magnetic states in the band. Indeed, the magnetism
in ZrZn2 is due to 4d electrons with a bonding character at the bottom o f the 4d band

309

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

z

0.33: - -
@ Y
o Ni
o 2Ni

z,

k ~
<-'_'-_=._-z'~ c ,
i~--~-~--. _-~,,-2.~-L-,- 3~

0.90
-1.4o

,i
0

o@ o@
@

axis
0.161 - -

~olool,
0

@o

->

NilI~YI
"

\ t t t ~ c . ~ ) ) ? l t - - - A - o.m

, ~ - -.. - C.'>"

:-... ~'i

"~- ~ ~ ' - -

0,082 i

*
1

--0.45

L./

(a)

0.30

0.5

(b)

Fig. 14. Magnetization density map in YNi 3 at 4.2 K (projection along the a axis of the hexagonal cell). The
contours are in units of 10-'p,B/,~ 2 (Gignoux et al. 1980b).

while in YNi3 and Y2Ni7 the 3d electrons, which contribute to magnetism, lie at the top
of the 3d band and have antibonding character.
2.1.3.

Cobalt antiferromagnetism

In the LaCol_e (e=0.15) and La2Co3 compounds, although the La amount is larger
than the critical concentration (RCo2) for the onset of magnetism, one observes a
resurgence of magnetism characterized by an antiferromagnetic ordering with a large
Nrel temperature (146 K and 315 K, respectively) (Gignoux et al. 1985). This resurgence
as well as its antiferromagnetic character (quite unusual in a Co-based metallic system)
can be understood in the light of the crystallographic structure of these compounds.
In LaCol-e, La atoms form an ABAB-type hexagonal packing while Co atoms lie
along chains parallel to the 6-fold axis. The Co-Co distance (2.36A) in a chain is
much smaller than the distance between chains (4.89 A) and incommensurate with the
c parameter of the La skeleton. Moreover weak correlations in Co positions manifest
themselves between the chains. All this leads to the exciting phenomenon of quasiunidimensionality of the Co lattice. This compound has a triangular magnetic structure
with Co atoms, belonging to the same chain, ferromagnetically (Mco = 0.7 ~tB) coupled
with their moment perpendicular to c. These chains are divided into three sublattices
of which the magnetizations make an angle of 120 (fig. 15). Ferromagnetism inside
a chain arises from the quasi-one-dimensional character of Co. As shown by Weinert
and Freeman (1983), the large reduced dimensionality of linear chains gives rise to large
moments and exchange values as compared to those of the bulk. The triangular magnetic

310

D. GIGNOUXand D. SCHMITT

LaCo1_E

2V_:/-" V

o[O

/'A /

! V

,ka

Fig. 15. Magnetic structure of LaCol_~. Moments of

Co atoms of the samechain, which is parallelto the c axis,
are parallelwith each other (Gignouxet al. 1985).

structure between chains results from the frustration of negative interactions between
neighbouring chains in a triangular (hexagonal) lattice. These interactions cannot arise
from direct exchange which, on account of the large interchain distances, is expected to
be negligible.
This unusual negative interaction in metallic Co systems must therefore originate
from an indirect C o L a - C o exchange through the 3d-5d hybridization bearing strong
analogy with superexchange in insulators. When the 3d magnetism is established, the
3d-5d hybridization induces a polarization of the 5d electrons. The system takes then
advantage of a spin distribution which minimizes the 5d electron kinetic energy and gives
rise to an antiferromagnetic coupling between chains. This indirect Co-La-Co negative
exchange interaction is easily affected by exchange scattering due to RKKY interactions
which polarize the 5d electrons in the vicinity of the 4f shells as it is shown by the change
into collinear ferromagnetism when La is replaced by Pr or Nd (Ballou et al. 1986a,b).
For La2Co3 the same interpretations of resurgence as well of antiferromagnetism work as
the crystallographic structure can be described as a packing of alternating layers of Co
and La characterized by a Co-Co distance in the layer much smaller than that from layer
to layer.
2.2. 3d Magnetocrystalline anisotropy

2.2.1. Experimental characteristics

Although crystalline electric field (CEF) coupling is much larger on 3d ions than on
4f ions, 3d metals and alloys are known to have a small magnetocrystalline anisotropy.
This is particularly true in systems with cubic symmetry as can be observed in pure Fe
and Ni, whereas anisotropy can be one order of magnitude larger in pure Co which is
hexagonal and where the second-order phenomenological anisotropy constant Kl reaches
7.9x 105 J/m 3 at 4.2 K. This small anisotropy, compared to that of the lanthanide atom,

MAGNETICPROPERTIESOF 1NTERMETALLICCOMPOUNDS

311

arises from the band structure which gives rise to a small number of states having a welldefined orbital character and from a quenching of the orbital moment in particular in
cubic systems (Friedel 1969). As the magnetocrystalline anisotropy of itinerant electron
systems is difficult to describe with a theoretical model starting from first principles on
account of the great complexity of the collective character of the electronic states, the
weakness of the 3d anisotropy in metallic systems was commonly admitted and there
was no clear-cut knowledge of the mechanisms which indeed rule this anisotropy.
Contrary to pure 3d metal, in the early seventies evidence had emerged of a large
3d contribution to the uniaxial anisotropy in the hexagonal RCo5 compounds, in favour
of the 6-fold axis. This magnetocrystalline anisotropy is of the same order of magnitude as
that of the lanthanide at room temperature, giving rise to spin reorientation in compounds
where the crystal field favours the basal plane as easy axis of the lanthanide. At low
temperature the lanthanide anisotropy is larger than that of Co and, as exchange interaction
dominates, magnetization is perpendicular to c. At high temperature the Co anisotropy
overcomes that of the lanthanide and magnetization becomes parallel to c. This has been
studied extensively in TbCos, NdCo5 and DyCo5 (Lemaire 1966, Ohkoshi et ak 1976,
1977). In YCos, from torque measurements on a single crystal, an anisotropy constant
K I = 7.0 106 J/m 3 had been deduced (Klein et al. 1975), a value one order of magnitude
larger than in pure cobalt. The main results concerning the magnetocrystalline anisotropy
of this series, obtained after 1979, are the following:
(i) Polarized neutron diffraction studies on YCo5 (Schweizer and Tasset 1980) have
shown a large orbital contribution to the Co magnetization (26% and 16% of the
2c and 3g sites, respectively). These authors then suggested that this could be the
origin of the large magnetocrystalline anisotropy. Although only qualitative, nuclear
magnetic resonance (NMR) studies on different RCo5 compounds later confirmed
this large orbital contribution to the Co magnetism (Yoshie et al. 1985).
(ii) Magnetization measurements on a single crystal of YCos, shown in fig. 16, led
to an anisotropy field of 140kOe at 300K (Alameda et al. 1981) whereas it is
only 10kOe in pure Co, and to K1 =7.4x 106 J/nl3 at 4.2K. Moreover this study
has revealed a large anisotropy of the magnetization, between the c axis and the
basal plane, which reaches 4% of the total magnetization at 4.2 K. Presented in the
same paper (Alameda et al. 1981) a polarized neutron diffraction study on NdCos,
below and above the spin reorientation temperature range, confirmed this result and
showed that this anisotropy of the magnetization mainly arises from the 2c site.
Besides the RCo5 compounds, which have been particularly well studied, it has been
found that many other R-3d compounds where the 3d atoms are magnetic, i.e. those rich
in 3d, exhibit a large 3d magnetocrystalline anisotropy. This is particularly so in other uniaxial intermetallic compounds of the same type but with more complex crystallographic
structure, such as R2MI7 (hexagonal and rhombohedral), R2M7 (rhombohedral), and the
new material for permanent magnet applications R2M~4B (tetragonal). For instance, K1
reaches -3.2x 106 J/m 3 in Y2Fel7 (Perkins and Nagel 1975), 2.3x 106 J/m 3 in Y2Co7
(Kakol et al. 1984), -7.9x106J/m 3 and 7.1xl06j/m 3 in Y2CoI4B and Y2FeI4B,
respectively (Thuy et al. 1988). Actually the macroscopic anisotropy in these compounds

312

D. GIGNOUXand D. SCHMITT
//C

~6
o

4
Z

N
2

JL

.e.

.c / ' ~
YCo.,
4.2 K

5O

1OO

150

200

APPLIEDMA6NETICFIELD(KOel
Fig. 16. Variation of the magnetization in YCo5 at 4.2K in a field applied parallel and perpendicularto
c (Alamedaet al. 1981).

with several transition-metal sites results from the compromise between the different local
anisotropies which can be very different and even of opposite sign. A method to determine
the anisotropy constant of the different sites has been proposed by Thuy and Franse
(1986). Combining the results of neutron diffraction studies on the preferential occupation
of iron and cobalt atoms in the pseudobinary series Nd2(Fel-xCox)17 (Herbst et al. 1982),
and the experimental bulk anisotropy data of the pseudobinary series Y2(Fel-xCox)17
(Perkins and Nagel 1975, Perkins and Str~issler 1977), Thuy and Franse (1986) have
evaluated the contribution of the iron and cobalt atoms belonging to four inequivalent
sites. They have later extended this method to other series. Such an approach allows
for instance to understand why the bulk anisotropy of Y2Co17 (KI = - 0 . 5 106 J/m 3) is
much smaller than in the isomorphous Fe-based compound mentioned above. However
the results, based on the assumption that the local anisotropies of iron or cobalt on each
site are independent o f x, must be considered with caution. Indeed, Thuy et al. (1988)
concluded that in YCos the anisotropy constant of the Co at the 3g site was negative,
i.e. opposite to that of the 2c site with the largest anisotropy, whereas the magnetization
anisotropy, deduced from polarized neutron diffraction on NdCos (Alameda et al. 1981),
Mco, IIc- Mco, c is positive for both sites.

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

313

In the cubic Laves phases RM2, the 3d atoms are at a site with a three-fold symmetry
axis lying along one of the four (111) directions. In the RCo2 series NMR spectra have
revealed a large and anisotropic orbital moment leading to a strong local anisotropy
of each Co atom (Hirosawa and Nakamura 1982). However, because of the particular
symmetry of this cubic structure, contributions from respective 3d atoms in the unit cell
to the total anisotropy energy cancel each other, resulting in a relatively small anisotropy
energy of the order of 104x J/m3, as observed in GdCo2 (Gignoux et al. 1975). The same
conclusions have been obtained for Mn from NMR studies and from the knowledge of
the magnetic structures of the (Yl-xTbx)Mn2 compounds (Berthier et al. 1988).
Finally, it is worth noting the hexagonal RCoI_~ compounds (R = La, Pr, Nd), where the
strong Co anisotropy is, contrary to that in YCos, in favour of the basal plane (Ballou and
Lemaire 1988): the anisotropy constant K1, deduced from magnetization measurements on
single crystals, is of the order of -13 x 106 J/m3. NMR studies on these three compounds
led in LaCo~-e to a relative orbital contribution of 30% to the Co magnetic moment
(Ballou, unpublished). This confirms again that orbital magnetism is at the origin of the
3d magnetic anisotropy.

2.2.2. Theoretical interpretations

The origin of this large 3d anisotropy has been a puzzle until recently, and its
understanding led, in the few last years, to a better knowledge of the 3d magnetism in
metallic systems. As we will see below, this anisotropy arises precisely from the itinerant
character of magnetic electrons (or holes), and its phenomenological approach, first
proposed by Ballou and Lemaire (1988) and then more precisely developed by Pinettes
and Lacroix (1993a) in the case of LaCol_E, has been confirmed "by band calculations in
the case of YCos (Daalderop et al. 1992, Nordstrrm et al. 1992).
It is well known that magnetocrystalline anisotropy arises from spin-orbit coupling.
However the sign and magnitude of this anisotropy depend on the orbital selection
operated on the collective electron states before the spin-orbit coupling is switched on.
Such a selection has two origins (see for instance Friedel 1969): (i) The first is analogous
to what occurs in insulators or in 4f ions, and corresponds to the diagonal elements (the so
called a overlap integrals) of the band Hamiltonian in the individual state representation;
it gives rise to the well known crystal field effects. This effect is always rather small and
cannot explain the large observed anisotropy. (ii) The second, properly itinerant, is due
to the/3-type multi-centre resonance integrals, mixing the individual states and leading
to electron energy dispersion, i.e., the bandwidth. A strong dependence of this dispersion
on the azimuthal quantum number is to be expected for an atomic packing other than the
compact ones. This latter situation is encountered in pure 3d metals. On the contrary, in
the rare-earth-3d intermetallic compounds we are faced with a band structure with rather
well-pronounced one-dimensional or two-dimensional character. This latter situation is
precisely the one encountered in thin films on which band calculations, leading to similar
conclusions as in the compounds we are interested in, have been performed (Bruno 1989).

314

D. GIGNOUXand D. SCHMITT

YCo5 enters in this category whereas LaCol_e has a strongly pronounced one-dimensional
character.
Let us consider for instance the case of hexagonal RCol-e. In these compounds
Co atoms form chains parallel to c in which the Co-Co distance is much shorter
than the distance between chains. One is faced with a quasi one-dimensional itinerant
system. Then the dispersion of the 12, 0) states of 3d electrons, i.e. the 3z2 - r 2 orbitals,
which are elongated along c, is much larger than that of the 12,+2) states, i.e. the
x 2 _y2 and xy orbitals. Taking into account band calculations within the local spin-density
approximation of Weinert and Freeman (1983 ), Pinettes and Lacroix (1993 a) deduced the
band structure sketched in fig. 17. The width of the d~2 band is ten times larger than that of
the degenerate dxy and dx2_y2bands, and twice as large as the degenerate dxz and dyz ones.
Note that, due to the a integral, i.e. the crystal field, the center of the dxy and dx2_y2band is
lower than that of the other bands. The 3d band is almost filled up and the position of the
Fermi level lies in states corresponding to orbitals with a prolate character. The application
of spin-orbit coupling leads then to a strong basal-plane anisotropy in agreement with
experiment. Note that the sign and magnitude of the anisotropy energy strongly depends
on the position of the Fermi level. This is well illustrated in fig. 18 which shows the
anisotropy constant K1 as a function of the number, n, of d electrons. In RCol_e, with the
parameters used, one finds a negative Kl value about half that estimated experimentally,
but the model is too simple to make a quantitative comparison. However this calculated
anisotropy is still one order of magnitude larger than in pure Co. Figure 18 also shows
the calculated anisotropy A L = L [ I - L -L of the orbital moment as a function of the band
filling n (number of 3d electrons per atom). It is worth noting the close connection
between K1 and AL: the same relative amplitude, opposite signs when the band is less
than half filled and the same sign when the band is more than half filled. This is quite
consistent with experiment. The different signs for n < 5 and n > 5 arise from the fact that
in the former case one considers electrons whereas in the latter one considers positive
holes.
This phenomenological approach is less simple in the case of YCo5 where, opposite
to the situation above, the cobalt chemical bonds have a strong planar character (in
layers perpendicular to c). However it is easy to predict a situation opposite to that of
the RCOl-e compounds: the dispersion of prolate states is much smaller than that of
the dxy and dx2_y2 band. As the Fermi level is also near the top of the 3d band where
the states have a 12,-4-2) orbital character, the system has a strong positive anisotropy
constant. This has recently been studied quantitatively by ab initio calculation of the
magnetocrystalline anisotropy and orbital magnetism in YCo5 using the LMTO (Linear
Muffin Tin Orbital) method (Daalderop et al. 1992, Nordstrrm et al. 1992). As illustrated
in fig. 19 one can see again the close connection between the anisotropy energy and
the anisotropy of the orbital moment. Moreover, the lower part of the figure shows that,
in agreement with experiment, the anisotropy of the 3g site is smaller than that of the
2c site.
In a very simplified way one can understand the link between anisotropy energy and
orbital anisotropy. Indeed, the spin-orbit coupling of 3d electrons can be written as

315

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

15

q
,/

"Z',

dxy and dx2_y2 : 12,_+2>

10

40

~-

20

dxz and dyz : I2,_+1>

's
-~

dz2 : 12,0>

="

-5

_ ~ '

/1""~

S (a.u.)

-10

,m

-20
-40

g
._=
,4
<3

EF

Fig. 17. Schematic density of states of 3d electrons in a quasi one dimensional itinerant
system.

E
o
o9

~_C

a)

... .(.~

0
~

00

"'%..'""

2.--I

~ [~Jr

-2
44

46

r
48

50

-6
52

b)
."'"".. [

o.

-5

_J

Fig. 18. Model of Co anisotropy in hexagonal RCoI_~:

calculated anisotropy constant K~ and orbital moment
anisotropy AL as a function of the band filling n (number of 3d electrons per atom) for U=0.3 V/~ and
z l = - 0 . 4 e V (Pinettes and Lacroix 1993a). U, Coulomb
repulsion; gel, d,2 bandwidth; A, the crystal field
parameter which splits the two planar orbitals from
the three other orbitals. K~ is in Z2/W~ (per unit cell)
whereas AL is in )./W1 [xJat, where ). is the spin-orbit
coupling coefficient.

2C "', /
-10
44

46

48

50

Bandfilling q

52

Fig. 19. Ab initio calculation of the magnetocrystalline

anisotropy energy and anisotropy of the orbital angular
momentum in YCo 5 using the LMTO method (Daalderop
et al. 1992). (a) Magnetic anisotropy energy per 3d atom,
AEn(q), versus bandfilling q. Solid curve, calculation
including spin--orbit coupling together with orbital polarization; dotted curve, spin-orbit coupling only; dashed
curve (referring to the right-hand axis), anisotropy of
the orbital angular momentum per 3d atom (calculation
includes orbital polarization); vertical line, actual number
of valence electrons n; horizontal dashed line, experimental
anisotropy energy. (b) Anisotropy of the orbital angular
momentum per Co atom at the 2c site (dashed) and 3g site
(dotted), in units of h/2:~, and the average anisotropy per
Co atom (solid).

316

D. GIGNOUXand D. SCHMITT

Eso = ). L . S with )~ < 0 for electrons (n < 5) and ~ > 0 for holes (n > 5). The anisotropy
constant is
KI = Eso, - Eso, [I = ,~(L - Lrl) S = -~. z ~ . S,
if one assumes that the spin is independent of the direction. Then KI is proportional to
AL with the same sign for n > 5 and with opposite sign for n < 5. The large anisotropy
in many R-3d intermetallic compounds arises from the large orbital moment and hence
its large anisotropy.
In principle the Coulomb repulsion U between electrons depends on the orbitals
occupied by the two electrons, and this could be important for the calculation of orbital
effects. However it has been established that this has little effect on the anisotropy energy.
In conclusion, the large magnetocrystalline anisotropies in rare-earth-3d intermetallics
arise from the orbitally selective 3d band energy dispersion due to the particular character
of the 3d atoms surrounding, associated with the presence of the rare earth, this effect
being much larger than the usual crystal field effects.
2.3. Instability and frustration of Mn magnetism in RMn2 compounds
It is well known that exchange interactions in metallic Mn-based systems are negative.
Moreover, when two nearest neighbours of an atom can be nearest neighbours to each
other (e.g. triangular lattices or atoms on the tops of a regular tetrahedron) one is faced
with frustrated magnetic systems. Such a topological frustration occurs in the RMnl2,
R6Mn23 and RMnz series. Among them, the RMnz compounds are especially fascinating
because the Mn moment is close to the instability of band magnetism. So, magnetism
is complex and presents exotic features which have attracted much attention during the
last ten years and have been investigated by using macroscopic techniques (such as
magnetization and thermal expansion measurements) as well as microscopic techniques
(e.g. neutron and X-ray diffraction, nuclear magnetic resonance (NMR) and Mrssbauer
measurements).

2.3.1. Topologicalfrustration
For light lanthanides or for heavy ones, that is for large or small R atoms (R = Pr, Nd or
Er, Tm, Lu), the RMn2 compounds crystallize in the hexagonal C14 Laves phase. ThMn2
and ScMn: also crystallize in this C14 phase. For intermediate R atoms (R= Gd, Y, Tb,
Dy) the RMn2 compounds crystallize in the cubic (f.c.c.) C15 Laves phase. A dimorphism
is observed for R = Sm or Ho. As shown in fig. 20, in both structures Mn atoms are at
the tops of regular tetrahedra. These tetrahedra are stacked in chains along the six-fold
axis in the hexagonal phase and they are packed in the diamond arrangement, connected
by sharing vertices, in the cubic structure. In both structures, the topology of the packing
ensures that any antiferromagnetic ordering will be highly frustrated.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

C15

I .

....

o-- .-o 0

317

' - ~ C

14

Fig. 20. Crystallographic structures of f.c.c. CI5 and hexagonal C14 Laves phases. Left, projection in
(001) plane; right, stacking of Mn tetrahedrons.

2.3.2. Mn moment instability and complex magnetic structures

A characteristic signature of the magnetic instability of manganese in the RMn2
intermetallics is well illustrated by the large magnetovolume anomalies observed (fig. 21)
for some R atoms. In fact, as shown (Wada et al. 1987a and references therein, Shimizu
1985) through thermal expansion and NMR measurements (fig. 22), such anomalies occur
depending upon the Mn-Mn interatomie distances. A critical distance (de = 2.66 A) for
the onset of a Mn moment appears to exist, which allows us to divide the compounds
into three different subsets of magnetic behaviours.
When the Mn-Mn distance is significantly lower than de, as in ScMnz, ErMn2
and HoMn2, Mn is nonmagnetic. ScMn2 is a Pauli paramagnet and its thermal
expansion coefficient (TEC) is small. ErMn2 and HoMn2 exhibit ferromagnetic ordering
characteristic of the rare-earth moment only (Feleher et al. 1965). The paramagnetic TEC
is large but no volume discontinuity appears at TN. (The situation of HoMn2 is not quite
clear and NMR seems to indicate that, depending on the preparation, some Mn sites could
be magnetic.)
When the Mn-Mn distance is larger than de, as in PrMn2, NdMn2, SmMn2 and
GdMn2, the Mn magnetism is well stabilized with large values of the Mn moment
(about 2.7/~B). Magnetic Mn ordering is accompanied by a large volume discontinuity
and a paramagnetic TEC is as large as in the compounds with Er and Ho. Collinear
antiferromagnetic structures with moments parallel to [120] have been determined by
neutron diffraction in PrMn2 (TN = I15K) and NdMn2 (TN = 104K). However, in the
latter, below 50 K, some Mn moments and all Nd moments progressively rotate toward the
[100] axis due to the magnetocrystalline anisotropy ofNd (Ballou et al. 1988a, Ouladdiaf
1986). The magnetic structure of GdMn2 bears some similarities [same propagation vector

318

D. GIGNOUX and D. SCHMITT

YMn2

L
:17:--_-:;

PrM n 2

1,

....

100

200
TEMPERATURE (N)

'

ditcttometer

300

Fig. 21. Thermal expansion curves of RMn2 obtained by X-ray diffraction measurements (circles) and
dilatornetric measurements (dashed lines). Open and solid circles show the processes with decreasing and
increasing temperature, respectively. Solid lines are guides for the eye (Wada et al. 1987a).
2 3,
2 0)] with the virgin magnetic structure of TbMn2 (Ouladdiaf 1986). The magnetic
(3,
structure of SmMn2 is still unknown.
The most dramatic properties are observed in the other compounds, where Mn is close
to the magnetic instability and where magnetism is very sensitive to external parameters
such as temperature, pressure, magnetic field and alloying. When temperature is increased,
YMn2 shows a first-order transition (with a large hysteresis) at TN accompanied with
a giant volume drop of about 5%, which is ascribed to a substantial reduction of the
Mn moment at TN. The TEC above TN is the largest of the series (50 10-6 K -1 at 300 K).
The magnetic structure of YMn2, determined from neutron diffraction, is helimagnetic
(fig. 23) with a long period (~380 ]~). On the other hand NMR spectra have shown that the
helix is not regular but distorted owing to a large Mn magnetic anisotropy which favours

319

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

(o)

%
zL
~3

Fir

T
o

:d.
I

Sc.

2.5

2.6

Ho

im

2.8

d~n_t4n Qt 4.2 K2 "~,)

! ff6

(b)

~2-

2O

Tb{

<J

so.

0
/ 2.5

I
2.6

Er, H o 0#,
2.7

Nd Pr

10

2.8

INTERATOMIC DISTANCE, d~r~M#t R.T. (~,)

Fig. 22. Variations of Mn moment, volume change at T~ and thermal expansion coefficient (TEC) with
interatomic distance, dMn_Mn(Wada et al. 1987a). (a) Mn moment, /~Mn, at 4.2K in RMn2 as a function of
dMn_Mn; CO) volume change at T N (AV/V) at T= T~ (solid circles) and the average TEC between 200 K and
300 K (open circles) plotted against dMn-Mnat room temperature. Lal0 and La25 represent Y09oLaoloMnz and
Y0.75Lao25Mn2, respectively.

one of the local two-fold axes (110) (Ballou et al. 1987). The substitution of a small
amount of Sc for Y shrinks the lattice, leading to nonmagnetic Mn atoms (Nakamura et
al. 1988a). In Y0.97Sc0.03Mn2 a complete paramagnetic state at 4.2 K has been observed
and the absence of Mn moment has been confirmed by NMR. However, a large TEC,
as in YMn2, is observed. For a Sc amount smaller than 2% the coexistence of the
magnetic state with a Mn moment mMn ~ 2.7/UB and nonmagnetic state is observed. On
the contrary, the substitution of La for Y expands the lattice and enhances the stability of
the antiferromagnetism. In particular, in Y0.75La0.25Mn2 the volume discontinuity at TN is
about 2% and the TEC at room temperature is 40 10--6 K -I, smaller values than in pure
YMn2 (Nakamura et al. 1988b). In the Yl-xCexMn2 pseudo-binary alloys, the decrease
of the cell volume with Ce content and the destruction of the magnetic ground state for
x ~> 0.05 are surprising as Ce has a larger ionic volume than Y in both the 3+ and the

320

D. GIGNOUX and D. SCHMITT

2N-I

~"/

22N !

2N+2

O ,

~J~+

~' ~ >?:

b I

, / ~-/

')

y /'

>

! ,.i ....

0~. I .....

..............

....,j

Fig. 23. Magnetic structure of YMnz. Only Mn atoms are shown. (a) Collinear antiferromagnetic structure
previously reported (Nakamura et al. 1983), open and solid circles represent Mn atoms with up and down spins,
respectively. The propagation vector is Q = (0, 0,1); (b) helimagnetic structure deduced from neutron diffraction
studies using a long wavelength. In the tetrahedron layers (solid lines) the antiferromagnetic arrangement is
collinear as in (a). The propagation vector is Q=(r,0,1) with r=0.02. A and z~' are the two easy directions
of Mn moments. Inset: in both models, the Heisenberg exchange interactions do not cancel between pairs of
atomic layers 2N - 1,2N, but do cancel between pairs of atomic layers 2N, 2N + I. Magnetic structures are
then formed of highly correlated layers of tetrahedra with weak coupling between these layers. (Ballou et al.
1987).
4+ state. This property has b e e n ascribed to a strong hybridization o f the C e - 4 f and M n - 3 d
b a n d s ( M o n d a l et al. 1992).
In T b M n 2 , the m a g n e t i c field structure in zero applied field is complex, with all
m a n g a n e s e atoms i n the m a g n e t i c state. However, a " m i x e d structure" with coexisting
m a g n e t i c a n d n o n m a g n e t i c M n atoms (fig. 24) is i n d u c e d b y an applied field o f
4.5 T at 25 K or b y chemical pressure i n d u c e d b y the substitution o f smaller Fe atoms

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

321

Mn(lb)~ Mn(3d)~ Tb~

Fig. 24. Magnetic structure of Tb0Vlno.96Fe0.04)2below 30 K, or the S2-typestructure of TbMn2 induced by an
applied field of 4.5 T at 25 K (Brown et al. 1992).
for a small amount (~4%) of Mn atoms as shown by neutron diffraction (Brown et al.
1992) and by M6ssbauer measurements (Oddou et al. 1993). In this magnetic structure,
although all the Mn sites are chemically equivalent, only 25% of them bear a magnetic
moment. Note that in TbMn2, as well as in YMn2, the variation of the N6el temperature
with hydrostatic pressure reaches a huge value, namely - 3 6 K/kbar (Voiron et al. 1990,
Oomi et al. 1987).
DyMn2 is the first compound in which, in the absence of applied field, a mixed
magnetic-nonmagnetic Mn state was evidenced by NMR (Yoshimura and Nakamura
1984, Yoshimura et al. 1986a). The magnetic structure, as determined from neutron
diffraction (Ritter et al. 1991), is similar to that of TbMn2 in an applied field when
considering the Mn sublattiee only.
Because thorium is tetravalent, the M n - M n distance in ThMn2 cannot be compared
with dc determined for trivalent R elements; however in this compound, due to the
larger filling of the 3d band, Mn is also close to the magnetic instability. ThMn2 orders
antiferromagnetically at TN=ll5 K (Buschow 1977b) in a mixed structure, shown in
fig. 25, in which magnetic Mn atoms form a triangular structure resulting from the
topological frustration of antiferromagnetic interactions (D6portes et al. 1987a). Note that
in all these mixed states the nonmagnetic Mn atoms are those which are subjected to a

322

D. GIGNOUXand D. SCHMITT

In 2

Fig. 25. Projection into the basal plane of the ThMn2 magnetic structure. Only Mn atoms of the 6h site are
magnetic (Drportes et al. 1987a).
total field smaller than the critical value necessary to induce a magnetic state of higher
energy. In particular, in ThMn2 the molecular field on nonmagnetic atoms is strictly zero.
However the cancellation of Mn moment on some sites is really a nonmagnetic state
and not a paramagnetic one which would have resulted from a simple cancellation of
the field on localized moments. Indeed, the susceptibility does not increase at decreasing
temperatures as expected for local moments in paramagnetic state in an applied field.
Consequently the splitting of the local band is cancelled.
2.3.3. Large Mn anisotropy
Although it cannot be measured directly because of the complex frustrated structures,
a large local magnetocrystalline anisotropy of Mn moments has been shown in the
RMn2 compounds. It manifests itself clearly in YMn2 and NdMn2. In the former,
a quantitative analysis of the tilt of the moments with respect to a perfect helix,
inferred from NMR and neutron diffraction experiments (Ballou et al. 1987), leads to
a local anisotropy of the same order of magnitude as generally observed in uniaxial
3d intermetallics, in particular those containing cobalt (see sect. 2.2). In NdMn2 this
anisotropy, of the same order of magnitude as that of Nd in the plane perpendicular to c, is
responsible for the noncollinear magnetic structure observed at low temperature (Ballou
et al. 1988a).

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

323

2.3.4. Giant spin fluctuations

Most physical properties of the RMn2 series bear witness to the importance of
spin fluctuations in these frustrated systems, which are dramatically enhanced in the
compounds close to the Mn magnetic instability, in particular in YMn2. As cited above,
the large volume change at TN and the large TEC above this temperature in this compound
is ascribed to a substantial reduction of the Mn local moment mMn and then to its rapid
recovery with temperature in the paramagnetic state. Indeed, a phenomenological theory
of magnetovolume effects (Shiga 1981) has shown that the magnetic contribution to the
TEC is proportional to the square amplitude of the local spin fluctuations or, in other
words, the square of the local moment. The lattice contribution to the TEC has been
taken as that of the Y(Mnl-xAlx)2 compounds for x > 0.1; there it has been established
from different measurements (lattice parameter as a function of the AI content, NMR
experiments, Yoshimura et al. 1986b) that mMn is stable and takes the maximum local
value of about 3/~B which is temperature independent. From this weak contribution to
the TEC, it has been possible to deduce the thermal variations of local spin fluctuations,
i.e. (m2n), in the different RMn2 compounds. They are schematically shown in fig. 26.
Curve (e) is the local moment limit (Y(Mnl_xAlx)2 compounds with x > 0.1, Shiga et al.
1987, Motoya et al. 1988) whereas, at the opposite, curve (a) is the Pauli paramagnetic
type limit (ScMn2). In YMn2 the Mn moment, which reaches 2.7#B at low temperature,
is reduced to about 1/tB just above TN and recovers to 2#B at room temperature. This
result has been confirmed from neutron diffraction experiments below TN and from
paramagnetic scattering measurements using polarized neutrons (D@ortes et al. 1987b,
Freltoft et al. 1988). The results revealed that the amplitude of the local magnetic moment
of Mn atoms drops by more than 30% at TN and then slowly increases with increasing
temperature. Moreover, up to 6TN, the paramagnetie scattering was strongly enhanced
about the staggered antiferromagnetic wavevector. The thermal variation of the magnetic
susceptibility of YMn2 is also quite unusual (fig. 27) and its increase with temperature
above TN is further evidence for the increase of the local spin-fluctuation amplitude with
temperature. In the nonmagnetic Y0.97Sc0.03Mn2 compound, the TEC as large as in pure
(e__])

( PrMn~ Ndt,,tn
,j--F~--~, ts~Mr~2
'
~u t |GdMn 2
~
~c

YMn2

~(c)

/X
(

-I
V

Mn2

}HoMn)

Er~4n-

(O)

T EMPERATURE

i~ xSC2xi,An

ScMn 2

"ID

Fig. 26. Schematic representation of the

temperature variation of ( ~ 2 n ) of RMn2.
Curve (a) shows the Pauli paramagnetic
type and (e) the local moment limit. The
type of spin fluctuations changes from (b) to
(d) in RMn 2 with increasing dMn_Mn(Wada
et al. 1987a).

324

D. GIGNOUXand D. SCHMITT

...,.. x =0.05

e Xe 0

..~"

Jle
-

e 0*
101

x=O.02
' eev*"- =
oI

~
I*=l.

i~'
%e,

le
%

eI

e= I

.o e

=joe
~eg***oe

o
t,e

ie

leo lee l*

oe e

.e

,=

..........o.:..f.....

~**

........

ieoQ~

"x._AY
0

100

200

300
t~00
500
T(K}
Fig. 27. Temperaturedependenceof the susceptibilityof Y(MnI_~Alx)2 for 0 ~<x ~<0.05. Arrowsindicateheating
and coolingprocesses(H= 8.28kOe) (Shiga et al. 1987).
YMn2 is typical of a nearly antiferromagnet in which giant spin fluctuations are thermally
excited (Wada et al. 1987a). In systems which are not so critical as YMn2 the jump of
(m~n) at the ordering temperature and its thermal increase in the paramagnetic state are
slightly smaller (curves b, d in fig. 26). Muon spectroscopy (~tSR) on YMn2 and related
compounds (Y0.9Tb0.1Mn2, Y0.98Sc0.02Mn2 and TbMn2) confirmed that the dynamics of
Mn magnetic moments in the paramagnetic state is typical for an itinerant system (Kalvius
1994). This measurement also confirmed the observation of Ballou et al. (1990a) that
paramagnetic state and antiferromagnetic phase (spin-glass like) coexist within a certain
temperature range above TN as an outcome of frustration.
Note finally that other physical quantities testify the importance of spin fluctuations in
this frustrated series close to the itinerant magnetism instability: (i) The thermoelectric
power in YMn2 shows an anomalous behaviour with temperature (Kamimura et al.
1987). (ii) The electric resistivity of this compound (Kamimura et al. 1987) follows a
T 5/3 dependence up to 10 K in accordance with spin-fluctuation theory (Ueda and Moriya
1975). (iii) In YMn2 a pressure of 1.6 kbar is enough to suppress magnetic ordering, and
the electronic specific-heat coefficient reaches 160mJ/K 2 mol (Fisher et al. 1992). As
well, in Y0.97Sc0.o3Mn2, which is also paramagnetic, y reaches 140mJ/K 2 mol (fig. 28)
(Wada et al. 1987b). These values are very large among the 3d transition metals and
compounds, particularly in comparison with YCo2 ( y ~ 18mJ/K 2 mol) which is close

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

325

16I
1501

....

""~" "~"-"Z."-''--'''---""-~-"
"
........ ...-~...,-;-'-"
""
oJ 1401 --~.."
E
13C
E

12C
11C
( Yo.97SCo.o3)Mn 2
100
0

10
T2

20
( K2 )

30

40

Fig. 28. Low-temperature specific heat of Y0mSc0.03Mnz in the form of C / T against T 2. The solid line has
7~w = 140 mJ/K2 tool and 0D = 300 K (Wada et al. 1987b).

to the Co moment instability and with YNi2 (7 ~ 1.8 mJ/K2 mol). These features bear
resemblance to the heavy-fermion systems, such as UPt3 (Franse et al. 1985a).
2.3.5. Theoretical approaches
The possibility of two magnetic states on Mn atoms in R M n 2 , depending on the
lattice constant, has been discussed in 1988 by Yamada and Shimizu (1988) from band
calculations within the tight-binding approximation. However the magnetic instability
of highly frustrated crystallographic structures has been theoretically treated for the
first time more recently (Ballou et al. 1991, 1992, Nunez-Regueiro et al. 1992a,b)
in an approach which takes into account the four main experimental characteristics,
namely: (i) the itinerant character of magnetism; (ii) the frustration of interactions;
(iii) the magnetic instability; and (iv) the anisotropy of Mn sites. Since in the RMn2
compounds the R-R and R-Mn exchange interactions are one order of magnitude smaller
than the Mn-Mn interactions, the theory is limited to the Mn lattice which is described
using a Hubbard model, where the on-site Coulomb repulsion U is of the order of the
bandwidth W, close to the magnetic-nonmagnetic (M-NM) instability. Due to the large
Mn anisotropy, transverse fluctuations are neglected and only the longitudinal component
of the local moment,/4 = (niT - nil 1, is considered. The effective Hamiltonian can then
be mapped into a pseudo spin S = 1 Blume-Capel Hamiltonian:

7~ = A ~ S~ + ~ Jij siS/,
i

(9)

i ~j

where zl and Jo are related to the parameters of the Hubbard model. (Note that in
the convention adopted in eq. (9), a positive Jij corresponds to a negative interaction
and conversely). The M-NM transition can occur if A > 0 and then Si can take three

326

D. GIGNOUX and D. SCHMITT

',

:
-0.5

',

',

',

0.5

It-

O2/J1

Fig. 29. Phase diagram at T = 0 of a two-dimensional triangular lattice using an effective Hamiltonian mapped
into the pseudo spin S = 1 Blume-Capel Hamiltonian (Ballou et al. 1991). Open circles indicate nonmagnetic
sites. The borderlines between the different phases are determined by the following: phases I and III, J2 = A - J ~ ;
phases I and II, J2 = J 1 / 2 - A / 3 ; phases II and III, J2 =zl/9; phases II and IV, J2 =-A/3. Jl and J2 are nearest
and second-nearest neighbour interactions, respectively. Note that following equation (9), positive values of J~
correspond to negative interactions and conversely.

values: Si = + 1 if the site has a magnetic moment/~i = :L/~, and S i = 0 if #i = 0, the latter
corresponding to the ground state in the absence of exchange and/or applied field. Rather
than introducing the complexity of the Laves phase RMn2 structure, a two dimensional
triangular lattice has been used to model frustration. This Hamiltonian was studied by
two different methods, i.e. Monte Carlo simulations and analytical calculations where
correlations in the elementary magnetic cell are treated exactly. Both methods give the
same phase diagram at low temperature. As shown in fig. 29 this model leads to a rich
phase diagram that can be related to the different RMn2 compounds: (i) a nonmagnetic
phase (phase I) where/~i = 0 at all Mn sites (ScMn2 and ErMn2); (ii) magnetic phases
(phases III and IV) in which/~i ~ 0 at all Mn sites (YMn2 and NdMn2); and (iii) a new type
of mixed magnetic phase (phase II, ThMn2 and DyMn2) in which, because of frustration,
the molecular field at some Mn sites, but not all, is too small to compensate the
energy A necessary to stabilize the moment. The field effect on these different phases is
particularly interesting: generally several field-induced first-order transitions are obtained,
even when starting from the nonmagnetic phase in zero field. Moreover, when applying
the field on a fully magnetic phase, mixed intermediate phases appear (fig. 30). This
characteristic is exactly that observed in TbMn2. It is worth noting that, in this model, a
uniaxial anisotropy favouring an easy magnetization direction is essential in order for a
mixed phase to be stabilized. Note that a recent ab initio band calculation in the ordered
phase of ThMn2 gives a good prediction of the observed mixed structure in this compound
(Uhl et al. 1993).
More recently, some states obtained within the above Blume-Capel (pseudo spin S = 1)
model have been studied using a more realistic treatment of the Hubbard model which
takes into account charge transfer and then does not fix an ad hoe value of the magnetic
moments (Pinettes and Lacroix 1993b). This model confirms the results obtained with
the pseudo spin S = 1 model, in particular it shows that the mixed phase can be stable

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

327

0.7. =

z
o
m

IN
F_0.
UJ
Z

o
<
0.25

A P P L I E D FIELD ( u n i t s of

10

J1)

Fig. 30. Monte Carlo simulation for increasing

magnetic field of a two dimensional triangular
lattice, for A=J 1 and Jz =JJ2, correspondingto
an antiferromagneticphase in zero field (Ballou
et al. 1992). Note that following equation (9),
positive values of J~ correspond to negative
interactions and conversely.

in the triangular lattice if." (i) U is not too large, and (ii) there is a large anisotropy.
The introduction of spin fluctuations in this model allowed one to qualitatively study
the thermodynamical properties of these systems. Depending on the temperature range
studied, the random phase approximation (Pines and Nozi~res 1966, Landau and Lifshitz
1981) or the Murata-Doniach approximation (Murata and Doniach 1972) were used.
The main results obtained are the following: (i) the large pressure-induced decrease
of the ordering temperature mainly comes from the magnetic instability rather than
the frustration; (ii) the specific-heat coefficient )~ of the paramagnetic phase seriously
increases in the vicinity of an instability when the interactions are strongly frustrated.
This result accounts for the much larger value of ~, in the paramagnetic phase of YMn2
(obtained by applied or chemical pressure) than in YCo2 where no frustration is associated
with the instability.
2.4. Lanthanide-3d transition-metal compounds where both carry a well defined
magnetic moment

2.4.1. General characteristics

As discussed at the beginning of this section, which is devoted to 3d magnetism, the
magnetic moment of the 3d ions disappears when the rare-earth fraction is larger than a
critical value which depends on the 3d element (fig. 1). Whereas this amount is small with
Ni, it is much larger with Fe. So in rich 3d rare-earth-transition-metal alloys the 3d atoms
are magnetic. Most of these alloys have tmiaxial structures (hexagonal, rhombohedral,
tetragonal) which are derived from the hexagonal CaCu5 structure found in the RCo5

328

D. GIGNOUX and D. SCHMITT

and RNi5 phases. They are of the type R2M17, R2M7, R]V[3, RMs. These alloys together
with derived compounds such as the prominent R2M14B series and more recent interstitial
compounds (nitrides, carbides . . . . ) are well known for their technological applications
as permanent magnets. Several reviews devoted to these materials have been written in
the last few years: Buschow (1988), Strnat (1988), Li and Coey (1991), and Franse and
Radwanski (1993). Since permanent magnets are beyond the scope of this chapter, we will
not discuss too much of the magnetic properties of these compounds. We would like only
to,present some major steps made in the last fifteen years in the knowledge of intrinsic
properties of these alloys in which the 3d element is mainly Fe, Co, or Ni. Long since, it
is known that in these materials (Kirchmayr and Poldy 1979, and references therein):
(i) The ferromagnetic 3d-3d dominant interaction is at the origin of the large Curie
temperatures which are well above room temperature.
(ii) The smaller antiparallel coupling between the lanthanide spin and that of
the 3d element leads to ferromagnetic compounds for light lanthanides where
J = [ L - S] and to ferrimagnetic compounds for heavy lanthanides where

J=L+S.
(iii) The indirect 4f-4f exchange coupling is still one order of magnitude smaller than
the previous one.
(iv) Spin reorientations are observed in some compounds as a result of the competition of the 3d and lanthanide magnetocrystalline anisotropies, the latter being
preponderant at low temperature whereas the former dominates near the ordering
tempera~tre.
2.4.2. Curie temperatures and 3d-4f exchange interactions
The experimental determination of the parameters describing the main interactions,
namely the exchange and the crystal field, has considerably progressed in the last decade.
We will focus on the 3d-4f exchange interaction. This interaction can be written as
ERM ---- --nRMMRMM,

(lO)

w h e r e M R and M M a r e the lanthanide and transition metal magnetizations, respectively,

and nRM the relevant molecular field coefficient. Assuming that the 3d moment is of spin
origin, the R and M spin magnetizations can be expressed as M s = yRMR and M s = MM,
where YR = - 2 ( g j - 1 ) / g j . Expression (10) can then be rewritten as
_

~S

~l~S ] 4 S

ERM -- --ttRMZ,a R~,~ M,

(11)

where nSul(<O)=nr~M/),R is the negative molecular field coefficient between spin

magnetizations.
In a given series of compounds (M fixed and R varied) the highest value of Tc always
occurs in the Gd compound, where the spin is the largest of the lanthanides. This means,
as is generally expected, that the R-M interaction occurs between spins and can be written
as in eq. (11) (Belorizky et al. 1987). In a given series, due to the expected similarity of
the lanthanide 5d band structure, it is usually assumed that npSM is constant.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

329

Applying a ferrimagnetic molecular field treatment the following expression for the
Curie temperature Tc has been derived by Nrel:

Tc = [TM+ TRv/(TM - TR)2+4T~M] ,

(12)

where TM, TR and TRM represent the contributions to Tc arising from the M-M, R-R
and R-M interactions, respectively, and are given by:
_ nMMCM,
s
TM : nMMCM --

T~ =

2 s
nr~CR = yRnr~RCr~,

(13,

14)
(15)

where CM and CR are the Curie constant of the M and R sublattices, respectively.
Concerning the M sublattice, the assumption of a Curie susceptibility is not obvious due to
the itinerant character of the 3d magnetism. Actually experiments show that the temperature variation of the total susceptibility may be analyzed assuming XM= CM/T where
CM has been measured for all the series considered. The above expression for TRM
shows that TiM is proportional to the "de Gennes" factor of the lanthanide element,
G ( J ) = ( g j - 1 ) 2 j ( j + 1). Assuming a constant value of nSM in a series, in general a
satisfying correlation between the value of Tc and the de Gennes factor is found for
heavy lanthanides compounds, but for light lanthanide compounds deviations from simple
systematics are always observed, e.g., Buschow and van Stapele (1970) for RFe2, Pszczola
and Krop (1986) for the iron series and Sinnema et al. (1984) for R2Fel4B. Basing on a
careful and systematic study of n ~ of different series Belorizky et al. (1987) were the
first to ascribe this deviation to substantial increase of the 3d-4f interactions towards the
light lanthanides. Combination of the above equations (Belorizky et al. 1987) leads to
-

5-,/(re
I~J[ V

rM)(rc

rR)

(16)

CRCM

In the series studied, TM was deduced from the value of Tc of compounds with
nonmagnetic rare earths such as La and Lu, whereas TR was deduced from the ordering
temperature of compounds with Ni where it is nonmagnetic. The results for the R2Fel4B
and RCo2 series, shown in fig. 31 a, reveal a significant decrease of InSMIwith increasing
number of 4f electrons. Because of the spatial localization of the 4f shell the direct overlap
between the 3d and 4f spins is rather small and Campbell (1972) suggested that the
3d-4f coupling is indirect and involves two stages, namely the 3d-5d and 5d-4f direct
interactions. These interactions being distant dependent, they are expected to vary from
compound to compound through dl=dRM -rso -r3d and d2 = rsd --r4f respectively, where
dRM is the distance between R and M atoms and r3d, rsd and raf are the radii of the 3d,
5d and 4f shells, respectively. In their analysis Belorizky et al. (1987) have shown that the
lanthanide contraction (associated with rsa) is ten times smaller than the corresponding
contraction of the 4f shell (associated with raf ). As a result across a series, whereas the

330

D. GIGNOUX and D. SCHMITT

5000

t"3

RCo 2

,.000

(*)

R 2 Fe14B (-)
300C

200C

t
(a)

100

La Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er Tm Yb Lu

(b)
1,3

1.25
!

f,~

1,2

Le Ce Pr Nd PmSm Eu Gd Tb Dy Ho Er

Tm Yb Lu

Fig. 31. (a) Absolute value of n ~ in the R2Fel4B and RCo 2 series. (b) r4f-rsa across the 4f series (rsd is noted
r R in the figure) (Belorizky et al. 1987).

variation of dz is negligible, there is a large increase of d2 as shown in fig. 3lb. As

a consequence, the overlap of 5d and 4f wave functions decreases towards the heavy
lanthanide ions.
The decrease of InSMItherefore originates from the increase of the difference between
the spatial extent of the 4f and 5d electrons (see fig. 3 lb). It has been concluded that the
change of the effective 3d--4f exchange interaction through the 4f series mainly originates

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

331

~00

[....,

0
1

c~

o, *-["l-.-...~R2Fel

7' R2Fe17C
",,, "- ..... ~. R2Fe14B,R2FelC
".~
R6 Fe~
"~-.,
z~

2O0
Ox

,,

""

",8. ,lp"

"~--

.. R2Co17,RCo5

R2C14B

",,

"-o

RCo2
o - - R2Ni17

cePrNd

Sm

6dTboyHOErTmyb

R
Fig. 32. Values of the exchange field Be~,e,M
experienced by the 4f spin moment in a number of
RnMm series. Solid circles, crosses and open circles
are data obtained from magnetization studies,
inelastic neutron scattering and from T c values,
respectively. Broken lines are guides for the eye
(after Radwanski et al. 1990).

from the change of the 4f-5d part. This variation of nSM is quite general and has been
observed in other series, as shown in fig. 32 which reports (Radwanski et al. 1990) the
exchange field on the rare earth due to the M atom (Bex,
P. RM = nRMMM)
, s S derived from
different types of experiments.
2.4.3. High-field magnetization processes, 3 d - 4 f interaction and magnetocrystalline
anisotropy
The analysis of high-field magnetization processes provided a quantitative determination
of the different exchange and anisotropy parameters. In particular in fields up to 35 T it
is possible to induce the noncollinearity of the R and M magnetizations in ferrimagnets
and then to measure the 3d-4f exchange coupling although, due to the importance of
this interaction, much larger fields (of the order of 200 T) are needed to reach the full
ferromagnetic state even in compounds with small anisotropy such as those with Gd.
In a two-sublattice model, the phenomenological energy of a magnetic system in an
external field Bo has been written as (Verhoef et al. 1989, 1990)
ERM = EMa + E Ra - nRMMRMM - (MM + M R ) Bo,

(17)

where MM and MR are assumed to be constant in modulus. The first two terms represent
the magnetocrystalline anisotropy of the two sublattices. The M sublattice is not affected
by the applied field (all M moments are assumed to remain parallel) and, its intrinsic
energy being conserved, it does not appear in the expression of the energy. Besides, the
4f-4f coupling is very weak and is thus neglected.

332

D. G1GNOUX and D. SCHMITT

,

< ~

~-

MR ~

MM

: r :~

200

150 ~W

g
m

100
M=IM M -MRI/

...........
50

BCPit
i
Blcpit
IApplied Magnetic Field 2

Fig. 33. Calculated field dependence of the resultant magnetic moment of a two-sublattice system without
anisotropy, or for free to rotate single crystals when only one sublattice has a magnetoerystalline anisotropy
(solid curve and left-hand scale). The dashed curve (right-hand scale) is the angle e, 180 - e being the angle
between the magnetizations of the two sublattices; for this curve, M M was taken equal to 2M R and [nRMIequal to
0.5 kgT/A m 2. For the different field ranges the related sublattice moment configurations are indicated (Verhoef
et al. 1990).

In ferrimagnetic systems without magnetocrystalline anisotropy, it can be readily

deduced from the above expression of the energy that in a field region such as:

InRMI IMM --MR[ < B < InRMIIMM +MRI

(18)

the magnetization curve is linear with a slope given by I1/nRM].Below and above this field
range, magnetization is constant and equal to IMM--MR[and MM+MR respectively.
Figure 33 shows an overview of the expected magnetization curve as a function of the
applied field Bo together with the related configurations of the two sublattice moments.
The field dependence of the angle e, by which the moment configuration deviates from
a strict anti-parallel alignment, is shown as well in fig. 33. Actually the assumption of
zero magnetocrystalline anisotropy is not realistic in compounds that are known to have
a substantial anisotropy. However, in systems where it is assumed that only one sublattice
has a magnetocrystalline anisotropy, it has been shown that the magnetization curves of
single crystals (one spherical single crystal or a finely powdered polycrystalline sample)
free to rotate in the applied field, follow quite well the above variation (fig. 33) when the
field is sufficiently large (Verhoef et al. 1989). This arises because the anisotropy does not
enter into the energy expression, due to the fact that the single crystals always orient their
total magnetization along the applied field direction. The above assumption is satisfied
in Gd-based compounds. With the other magnetic lanthanides it is assumed that it is the
magnetic anisotropy of the 3d sublattice which is zero because, in most hard magnetic

MAGNETIC PROPERTIESOF INTERMETALLICCOMPOUNDS

70

'

Ho2Co:l 7

60

T = 4,2

free

single-crystallane

free

poiycrystailine

333

:::
:::::
sphere

:::.

:/::::5:

powder

2~

50
<:I:

.::::::

40
[i

30

I.

~0

2O
8 IT]

40

3O

Fig. 34. High-field magnetization curves at 4.2 K of a free single-crystalline sphere (open circles) and a
free powder sample (crosses) of the Ho2Co17 compound (Verhoef et al. 1990). Dashed lines represent the
experimental results of linearly-decreasing field pulses.
lO0

BO

:o.................................. :i!!!!!!!)

60

....

.-1111'.'.11211..--'"

..."'., .v'"

..-'

.,

40

,..,

o.

v
o
o
o

20

o
o

10

20
B [T]

30

Ho2CoI7
Ho2Co14Fe 3
Er2Col 7
-Er2Fe17
Dy2Co~ 7
,

40

Fig. 35. High-field magnetization curves at 4.2 K of several R2M17single-crystalline spheres that are free to
orient themselves in the applied magnetic field (Verhoefet al. 1992).
R - M intermetallic compounds, it is known that the anisotropy o f the 3d sublattice is
smaller than the anisotropy o f the 4 f sublattice by at least one order o f magnitude. As
illustrated in figs. 34 and 35, using this remarkable property the intersublattice interaction
has been measured directly from the high-field magnetization curve in a large number
o f R - M ferrimagnets. Note in fig. 34 that there is a remarkable agreement between the
magnetization curves performed on a free single crystalline sphere and a free powdered
polyerystalline sample. This type of measurements is then particularly interesting because
it allows a direct quantitative determination o f nRM by using polyerystalline materials.

334

D. GIGNOUX and D. SCHMITT

I C

It- ......

10
..-7.

GdC3

~ - -

B/

Mc
5

.."'/
Uc

.............

.. /

c[ B

M~

Mco[ N ~ z /

--

s/

(b)

40

100

300 B(T)

Fig. 36. Magnetization curves in the full-field range for the GdCo3 ferrimagnet calculated with relevant
parameters when the field is parallel and perpendicular to the c axis (Radwanski 1986). The moment
configurations for the field of 250T applied parallel and for the field of 100T perpendicularto the easy
hexagonal axis are shown.

In the above experiments the magnetization of the sublattice considered as anisotropic

(a lanthanide with a nonzero orbital moment or 3d atoms in Gd-based compounds)
remains parallel to its easy magnetization direction and no information about the magnetic
anisotropy can be obtained. In case of fixed single crystals both moments are pushed
out of their easy direction and therefore magnetization curves depend on the anisotropy
as well as exchange interactions. Whereas in a simple uniaxial ferromagnet with only
a single magnetic sublattice the magnetization curve along the hard axis depends only
on the anisotropy constants, in ferri- or ferromagnetic systems, composed of different
magnetic sublattices, the external field disturbs the magnetic configuration of the different
sublattices. Expression (17) for the phenomenological energy still applies and has been
used in order to analyze the magnetization processes along the different symmetry
directions.
Two typical examples of the magnetization process for Gd-based compounds are
reported in figs. 36 and 37 which show the calculation performed for GdCo3 (Radwanski
1986) and GdCo5 (Radwansld et al. 1992a) by using relevant parameters. For fields
applied along the easy c direction the curves for both cases are similar, with the
same characteristics as those of fig. 33 where the intermediate region corresponds
to noncollinearity. Note that, due to the large intersublattice interaction, the forced
ferromagnetic state is usually reached well above 100T. The magnetization processes
along the hard axis of both examples are rather different and depend, in general, on
the magnitude of the spontaneous magnetization of the ferrimagnetic system and on the
strength of the intersublattice coupling. For large spontaneous magnetization and strong
3d-4f coupling, the hard-axis magnetization process is that of GdCo3 in which there is an

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

20

335

~o

.'/""

:tOO

200

300

'400

B (T)

Fig. 37. Magnetization curves in the full

field range for the GdCo5 ferrimagnet
calculated with relevant parameters when
the fieldis parallel and perpendicularto the
c axis (Radwanskiet al. 1992a).

30

"-~ 2 0

D~DoaoOa

aQ

GdCo 5
,," J_c

4.2

A
A

~A

-0

'

I0

20
B

(T)

30

40

Fig. 38, Experimental magnetization curves

for GdCo5 at 4.2K measured for field
parallel and perpendicular to the hexagonal
c axis (Radwanski et al. 1992a).

intermediate state with magnetization close to that observed along the easy axis. In the
other case the process is that of GdCo5 where, even in small field, the internal magnetic
structure is modified and the magnetization continuously and almost linearly increases
toward the forced ferromagnetic configuration. In particular, the apparent anisotropy
field, i.e. the crossing of easy- and hard-axis magnetization curves, does not depend on
anisotropy only but is a complex function of all the magnetic parameters. Such behaviour
has been verified by magnetization experiments performed up to 35 T which clearly show
the crossing of magnetizations along the easy and hard magnetization directions (fig. 38)
(Radwanski et al. 1992a). The fit of the hard-axis magnetization curve yielded nRM and
a value of the Co-anisotropy constant Kt close to that determined in YCos.
Let us focus now on compounds where the magnetic lanthanide has a nonzero orbital
moment. The large lanthanide magnetocrystalline anisotropy is, in general, one order of
magnitude larger than that originating from the 3d sublattice. Interesting magnetization
processes are observed in the high-field region, in particular in ferrimagnetic hexagonal
easy-plane systems. In most cases, due to the large number of parameters necessary
to describe the magnetocrystalli,ne anisotropy of both sublattices in addition to the
intersublattice coupling, the analysis of magnetization curves is a complex problem.

336

D. GIGNOUX and D. SCHMITT

40

Fig. 39. Magnetization curves for Ho2Co17

calculated within the two sublattice model
in the full field range for fields applied
along the a and b directions of the
orthohexagonal cell ([100] and [120] axes
of the hexagonal cell, respectively) of the
basal plane (Radwanski et al. 1985, Franse
and Radwanski 1993).

Ho2Coz 7

20
~

II

4.2

III

IV

Fig. 40. Different configurations of the R and M moments for a ferrimagnetic easy-plane hexagonal compound,
when the field is applied parallel to the easy direction of this plane (after Franse et al. 1988).

However the task is simplified in easy-plane systems where the second-order anisotropy
constant is large enough so that magnetic moments of both sublattices remain in the basal
plane during the magnetization process, when the field is applied within this plane. In such
a case, only the in-plane K anisotropy parameter of the lanthanide and nRM enter into the
expression (17) used to analyze the magnetization curves. The in-plane anisotropy of the
3d sublattice can be considered as negligible. Calculated magnetization process, shown
in fig. 39 for H o 2 C o 1 7 (Franse et al. 1985b), along the easy symmetry [120] direction
of the basal plane exhibits three transitions before the collinear induced ferromagnetic
state is reached. The intermediate states correspond to noncollinear magnetic structures
associated with first-order magnetic reorientation (FOMR) as shown in fig. 40. For fields
applied along the hard axis of the easy plane one expects only two transitions (fig. 39).
This behaviour has been observed for the first time in Ho2Co17 for which, thanks to a
35 T high-field installation, the first transition of each symmetry axis of the easy plane
was observed (fig. 41) (Franse et al. 1985b). More recently the second easy-axis transition
has been observed at 44 T (Tomiyama et al. 1991) in good agreement with the calculated
value (Radwanski et al. 1985). Such dramatic behaviour has also shown up in Dy2Col7
(Sirmema et al. 1986), Ho2Col4Fe3 (Sinnema et al. 1987) and TbCo5 (Ballou et al. 1989).
R RM = nSRMMM),
The exchange fields experienced by the 4f spin moment, i.e nSM, nex,
deduced from the different low-temperature high-field magnetization measurements are
reported for different series in fig. 32. They agree quite well with the values determined

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

~0

337

Ho2COl 7

60

"

2O

10

20

30

H (T)

Fig. 41. The high-fieldmagnetization curves for Ho2Co17 at 4.2 K measured Mong the three crystallographic

directions (the [120] axis is the easy axis of the easy basal plane). Symbols:+, [100] axis; x, [120] axis; open
circles, [001] axis; full lines are the calculatedcurves (after Sirmemaet al. 1987).
from the Curie temperatures as described previously, which has ted to the same conclusion
concerning the origin of the variation of nSM with the lanthanide element (Beiorizky et
al. 1987).

3. 4f Magnetism

This part is devoted to rare-earth-based intermetallic compounds in which only the

rare earth is magnetic. Due to the oscillatory and long-range RKKY interaction
characteristic of these materials, in most cases there is a competition between positive and
negative interactions. This competition leads, at low temperature, to frustrated magnetic
structures, a majority of them being of antiferromagnetic type associated with complex
magnetic-field-temperature diagrams characterized by the competition between magnetic
incommensurability and commensurability. Moreover, crystal field effects play a crucial
role not only in the magnetocrystalline anisotropy, thus leading in uniaxial compounds to
either easy-axis or easy-plane systems, but also in the lanthanide moment itself when
CEF leads to a nonmagnetic singlet ground state. Magnetization processes are often
complex and of metamagnetic type. Therefore the major part (sect. 3.2) of this chapter is
devoted to the large variety of metamagnetie processes found in these systems and to the
associated phase diagrams. The number of ferromagnetic compounds is much smaller and
our knowledge of the evolution of the CEF and exchange parameters in such materials

338

D. GIGNOUXand D. SCHMITT

during the last decade is the subject of sect. 3.1. Finally in sect. 3.3. we present a short
review of the advances in the analysis of incommensurate magnetic systems.

3.1. Crystal-fieM and exchange interactions in ferromagnetic compounds

The first studies devoted to the quantitative determination of the main interactions,
namely exchange coupling and CEF, have been performed on simple cubic ferromagnets
and a few antiferromagnets, in particular the RA12 and RNi2 series (simple cubic
Laves phase structure), the RZn, RCd, RAg and RCu compounds (simple CsCl-type
structure) and RSb series (NaCl-type structure). Due to the high symmetry there are only
two CEF parameters, and inelastic neutron scattering (INS) experiments are generally
sufficient for their determination. The paramagnetic exchange parameter J(0) is then
obtained from the thermal variation of the paramagnetic susceptibility. In ferromagnets,
where not enough information was obtained from inelastic neutron experiments, CEF and
paramagnetic exchange parameters were refined from a joint analysis o f M vs H below Tc
on single crystals along the symmetry axes and of the paramagnetie susceptibility. A
rather good knowledge of these interactions had been obtained in the late seventies
(Busehow 1980 and references therein, Waind et al. 1983 and references therein) and from
that time many studies were devoted to the determination of the other interactions, which
are smaller by one or two orders of magnitude, in particular quadrupolar interactions and
magnetoelastic effects. These results have been gathered recently in an extensive review
paper (Morin and Schmitt 1990). So these aspects of lanthanide intermetallie compounds
will not be presented here.
For the lower symmetries the number of CEF parameters is larger: four and five
in hexagonal and tetragonal symmetries, respectively, and more in other systems. It is
then impossible to determine CEF parameters unambiguously from INS experiments
only. When the number of parameters is not too large, in hexagonal and tetragonal
compounds, they can be determined, together with J(0), only from a joint analysis of
different experiments such as INS, magnetization measurement on single crystals, and
specific heat. In compounds with low symmetry (orthorhombic, monoclinic, triclinie)
such a task is illusory because of the large number of parameters to be determined.
So the next step of this type of studies has been devoted to ferromagnetic compounds
with hexagonal symmetry. In particular, much effort has directed towards the RNi5
series for the following reasons: (i) It is the only series of hexagonal compounds
where only the lanthanide metal is magnetic and which are ferromagnetic for any
lanthanide. (ii) The structure is simple (one lanthanide atom per unit cell at a highsymmetry hexagonal site). (iii) They are isomorphous with RCo5 in which characteristics
of the lanthanide are difficult to determine because of the strong 3d contribution to
magnetocrystalline anistropy and exchange interaction. The RNi5 compounds then offer
the possibility of determining the lanthanide contribution to the anisotropy assuming
that there is not much difference between RNi5 and RCos. (iv) Much work has been
devoted to PrNi5 because, due to a nonmagnetic singlet CEF ground state and small
exchange interactions, it does not order down to very low temperature and in fact has been

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

339

successfully used in nuclear adiabatic demagnetization experiments which have yielded

end temperatures around 0.8 mK (Andres and Darack 1977).
In hexagonal ferromagnets the Hamiltonian acting on the R 3+ ion is the sum of the
CEF contribution,
0 o + , 0 o 0 +,0o
"]"~CEF= B202

6 6
+B606,

(19)

the Zeeman coupling with the effective magnetic field,

7-/Zee = gJ#BJ" Heft,

(20)

and the exchange interaction,

7-/ex = -n(gj/zB) 2 {J)' J,

(21)

where n is the isotropic ferromagnetic exchange interaction related to J(0), the Fourier
transform of the exchange interaction at Q=0. In RNis, although Ni is nonmagnetic its
small induced contribution has been taken into account because it plays an important role
in magnetic properties when the lanthanide contribution is small, i.e. at high temperature.
This Ni polarization contributes in two ways: (i) by enhancing the internal (applied
plus demagnetizing) field Hi and leading to an effective field Herr = Hi(1 + ~t), and
(ii) by enhancing the exchange interaction between the lanthanide ions n = nRR + anRNi.
Here a = ZNinltNi where Z~i is the exchange-enhanced susceptibility, and nRR and nRNi the
molecular field constants between moments as defined in sect. 2.4 (Barthem et al. 1988).
Contrary to cubic compounds, the number of crystal field parameters is too large to be
determined unambiguously from the fit of inelastic neutron scattering (INS) experiments
on powder alone. Actually these four parameters, together with the lanthanide exchange
interaction and the two parameters (ZNi and nRNi) describing the Ni contribution, were
refined from a simultaneous analysis of the latter experiments and low-temperature
magnetization processes along the three symmetry axes together with the anisotropy of
the thermal dependence of the paramagnetic susceptibility measured on single crystals.
Additional measurements such as specific heat were used to confirm the obtained
parameters. Useful information has also been obtained from M6ssbauer spectroscopy
(Gubbens et al. 1988). The different sets of parameters obtained have been proposed
by Andres et al. (1979), Alekseev et al. (1980) and Reiffers et al. (1989) for PrNi5,
Goremychkin et al. (1985a) and Barthem et al. (1989a) for NdNis, Ballou et al. (1988b)
for SmNis, Gignoux et al. (1979) and Goremychkin et al. (1984) for TbNis, Aubert et
al. (1981) for DyNis, Gignoux et al. (1979) and Goremychkin et al. (1985b) for HoNi5,
Goremychkin et al. (1984), Radwanski et al. (1992b) and Zhang et al. (1994) for ErNis,
and Barthem et al. (1989b) for TmNis.
The importance of analyzing the largest number of experiments is well illustrated in
the case of ErNi5 in which a set of parameters has been proposed which gives good fits
for INS, paramagnetic susceptibilities, the low-temperature magnetization curve along the

340

D. GIGNOUX and D. SCHMITT

I

/
/
/
/
o

I II

Pr-Ni 5
/

"T

.-,-

///

,'/

//

/ / i/.,

,,2,/

100
200
TEMPERATURE (K)

300

Fig. 42. Thermal variation of the reciprocal susceptibilities of PrNi5 parallel and perpendicular
to c. Solid circles are the experimental values,
solid lines are the calculated variations, and
dashed lines are the variations corrected for the
nickel contribution (Barthem et al. 1988).

easy [001] axis and the specific heat, but which fails to account for the very high field lowtemperature curves along the [100] and [120] directions of the basal plane (see sect. 3.2.4).
A rather different set of parameters was then found which accounts quite satisfactorily for
all the experimental results. At this moment, from the comparison of the different sets of
CEF parameters through the series it seems that the latest ones determined in compounds
with Pr, Nd, Tb, Dy and Er are reliable whereas improvement can still be obtained for
those of Sm, Ho and Tm compounds (Zhang et al. 1994).
The relative importance of Ni magnetism is the largest for the light lanthanides (Pr, Nd
and Sm) and yields thermal variations of the paramagnetic susceptibilities quite different
from that of the lanthanide ion alone as shown in fig. 42 for PrNis. The experimental
susceptibilities parallel and perpendicular to [001], corrected for Ni contribution, follow
the linear Curie-Weiss behaviour at high temperature with a slope corresponding to the
effective moment of the R 3 ion. Note that the shift of the reciprocal susceptibilities
parallel and perpendicular to the [001] direction corrected for the Ni contribution allows
a first determination of the second-order CEF parameter (Boutron 1973) which generally
is close to the value refined from all the experiments.
It is worth noting that the first direct measurement of the Zeeman splitting of the
CEF levels by point-contact spectroscopy was done using PrNi5 (Reiffers et al. 1989).
The significance of this experiment, illustrated in fig. 43, is that: (i) it has been possible
to follow the transitions between levels when the field is applied successively along
the three symmetry axes and (ii) transitions which were forbidden in zero applied

MAGNETIC PROPERTIES OF INTERMETALLICCOMPOUNDS

15

0.15 I
0.27 I

0114

|
|

o.55
1.28

1.28
0.74

10

341

o.8s I
0.82
1.37
~.
, J ,.--';f _

15

0.02
0.29
0.5(:
0.7e
0.

0.16
0.48
1.21
1.42

::D
Q

..~E1(]
>.
0
uJ
Z
IJJ

1(]

'

I
1.89
2.59

'

1.6(]
2.87
I

.... L . _ +
0

1'o
MAGNETIC

1'8
FIELD ( T )

2'0

Fig. 43. Magnetic field dependence of the CEF energies

obtained for PrNis~Cu point contacts for fields along the
[100], [120] and [001] axes (different symbols are used
for different contacts) after Reiffers et al. (1989). The
solid curves are the calculated variations with the refined
parameters. The sequence of the lowest CEF levels at
zero field is F3 (ground state), F 1 (38.3 K), F~I~(4I .6 K) and
F6 (42.2 K). The numbers are proportional to the calculated
transition probabilities from the ground state at 10 T and
20 T and are given in the same order as the calculated
energy levels. The dashed curves show the calculated
energy of the first excited F t level with the CEF parameters
given by Barthem et al. (1988). (Zero energy corresponds
to the ground state).

field were observed. In particular it has been possible to follow the field dependence
o f the first excited F] CEF level whereas the transition from the ground state to
this level cannot be observed directly in zero field. This experiment confirmed that
the sets o f parameters previously determined were reasonably correct (Andres et al.
1979, Alekseev et al. 1980, Barthem et al. 1988) and at the same time allowed them
to be refined. The knowledge o f CEF parameters allowed one to predict a highly
original high-field magnetization behaviour arising from the existence o f the nonmagnetic
singlet ground state, namely transitions associated with the field-induced "crossing" and
"anticrossing" o f the two lowest states along the [100] and [120] directions, respectively.
In quantitative agreement with prediction, metamagnetic transitions have been observed
at low temperature (T ~< 5 K ) from magnetization measurements in fields up to 38T;
associated with "crossing" these become steeper and steeper as temperature decreases,

342

D. GIGNOUX and D. SCHMITT

~: 15

20

40

60

80
T(K)

O0

Fig. 44. Specific heat of ErNis: black dots are the

experimentalpoints(Radwanskiet al. 1992b),the solid
line is the calculationwithparametersgivenby Zhang
et al. (1994).

whereas those associated with "anticrossing" are nearly temperature independent (Ball
et al. 1992a) (see sect. 3.2.2).
In TbNis, in addition to the above quoted experiments, the spin waves and single-ion
type magnetic excitations have been studied by inelastic neutron scattering at 4 K on a
single crystal along the (q, 0, 0) and (0, 0, q) propagation directions (Gignoux and Rhyne
1986). One dispersive accoustic mode and two nondispersive modes were observed, from
which J(q) along these two directions has been deduced; this resulted in improvement of
the CEF parameters proposed earlier (Gignoux et al. 1979, Goremychkin et al. 1984).
As quoted above, in ErNi5 the latest set of parameters (Zhang et al. 1994) gives
particularly good fits for all the experiments, in particular the low-temperature high-field
magnetization measurements along the hard axes of the basal plane (see sect. 3.2.4) and
the specific heat characterized by a large Schottky anomaly around 35 K (Radwanski et al.
1992b; fig. 44). The microscopic aspects of the magnetization processes along the [120]
and [100] axes, calculated with the refined set of parameters, exhibit quite interesting
features as shown in fig. 45: (i) There is a large decrease of the value of the Er moment
during the rotation process from the easy c axis toward the applied field direction, the
minimum value corresponding to the field at which moments become parallel to [120] and
[100] (i.e. 19 T and 22 T), respectively. (ii) Above 22 T, the reduction of the magnetization
measured in the basal plane (compared to that measured along [001]), arises from a huge
anisotropy of the magnetization and not from the fact that moments would not be parallel
to the applied field. As with the reduction of the Er moment during the rotation, this is a
consequence of the fact that the moments in the basal plane are induced by the applied
field.
As for ErNi5, the calculation with the refined parameters allows us to understand
the magnetization processes. For instance, in NdNi5 (fig. 46), the anisotropy energy
between the a ([100]) and b ([120]) axes (hatched area in the figure) is weak: a field
of 11 kOe applied along b is large enough to align the moments in this direction. It is
worth noting that in higher fields the magnetization difference between both axes arises
from a large anisotropy of the magnetization which strongly increases with the field
up to 130kOe and reaches 12% in 150kOe. When the field is applied along the hard

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

343

10
Q ~

~ 0

Q 0

Q I

Q Q ~ ~

Z
LU

,o

H//[120] . . . . . .

~3

ErNi

zL

v
4-a
C-

o~o~O~

'I.4K

d)

./.

-_

7
/

/".

~" ~

H//[100]

/"

/
/
/
/
/

100

o*lao*aee*eawmm*~ol

(19
(1)
Ob

80

(1)

q~

H//[120]

60

H//[100]
o
o

40

o
o
.o
~o

20

10

20

30

H(Tesla)

40

Fig. 45. ErNis: calculated field dependences of the moment amplitude (upper
part) and of the angle between moments
and the [001] axis (lower part) when
the field is applied along the [120] and
[100] axes of the hard plane. Dashed
lines in the upper part represent the
calculation of the moment amplitude
when moments are forced to be either
along [120] or [100] (after Zhang et al.
1994).

c axis, the observed variation of the magnetization corresponds to a slow rotation o f the
Nd moments associated with a gradual decrease o f their amplitude. Besides, in zero field
the Nd moments are perpendicular to c and reach 2.01/ZB, in an internal field o f 150 kOe
the moment amplitude is only 1.40/~B and moments still make an angle of 29 with the
field (Barthem et al. 1989a).
From these quantitative studies in this hexagonal series and earlier studies in cubic
ferromagnetic compounds it has been possible to compare the exchange parameters
along the series. As for the previously reported analysis of the R - M exchange coupling,
Belorizky et al. (1988) analyzed the R - R coupling in different series. The variation OfnRM
in all the studied series had been ascribed to a variation of the 4f-5d exchange interaction
arising from the increasing distance between the 4fand 5d shells going from light to heavy
R elements. As the same 4f-5d exchange interactions are involved in any metallic system
containing magnetic R elements, it was natural to study the R - R interactions in systems
where they are not masked by dominant R - M interactions, i.e. in the absence o f magnetic
M atoms. The RA12, RZn and RNi5 series are quite suitable for such a study. As shown

344

D. GIGNOUX and D. SCHMITT

%
z
113
z

._..,_____..-----J

rn
:3.

-~

lib

NdNi 5
1.5K

MNd = 1.&.0 PB

HIIc

~=29" "

Nd

L ~ M N d =2.01PI~

50

160

Hi(kOe )

150

Fig. 46. Internal field (applied field corrected for the demagnetizing field) dependences of the magnetization
at 1.5 K in NdNi 5. Black dots are the experimental values. Solid lines are the calculated variations (Barthem
et al, 1989a).

5 000

4 000
I

~ " 3 ooo

~l ~

m~
20OO

\\\\
k\

~x
\\ \
\ ".

U
1000
\

\
~,~

0
Lo.Ce

Pr Nd,

Sm

~O
I

Gal.Tb Dy Ho Er Tm

Lu

Fig. 47. Molecular field coefficients nSR in the

RA12 (open circles), RZn (solid circles) and
RNi 5 (open triangles) series (Belorizky et al.
1988).

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

345

in fig. 47 there is a large increase of the R-R interactions across a given series from
heavy to light R atoms. This increase is still larger than for the R-M interactions (see
figs. 31 and 32). In particular it has been observed that the increase of nSg is similar to that
of (nSM)2. Knowing that the ne,g coupling involves three stages, namely the 5d-4f, 5d-5d
and again the 5d-4f interactions, the observed variation has the same origin as that of nSM
and confirms that, according to Campbell (1972), the exchange interaction involving R
atoms are essentially mediated by the 4f-5d exchange. The same conclusions have been
derived from the analysis of the transferred hyperfine field at the nonmagnetic site like
A1 and Zn in the RAI2 and RZn series (Belorizky et al. 1988 and references therein).

3.2. Metamagnetism and associated phase diagrams

The concept of metamagnetism has spread considerably over the last decades, since
Kramers introduced the term for the original property of a first-order phase transition
between a simple antiferromagnetic structure with a low magnetization and the induced
ferromagnetic state with a high magnetization, this transition being induced by the
application of a magnetic field (see ref. 1 in Stryjewski and Giordano 1977). First limited
to such single-step phase transitions, this concept has been progressively extended to a
lot of processes breaking antiferromagnetic structures, and more generally to all types
of field-induced magnetic phase transitions (Date 1990, Gignoux and Schmitt 1991).
The aim of this section is to present an extensive review of the different types of
metamagnetic processes occurring in the rare-earth intermetallic compounds in which
only the rare earth is magnetic (the metamagnetism associated with 3d magnetism is not
considered here). For each process, a number of examples as large as possible will be
provided for illustrating the particular physics involved or the particular balance between
the different couplings giving rise to the metamagnetic behaviour. After some general
considerations, each type of metamagnetic process will be widely developed by order of
increasing complexity.

3.2.1. General considerations

First we have to specify exactly what the term metamagnetism will signify in the
following. We will consider as metamagnetie behaviour any anomaly occurring during
the magnetization process of a given compound, compared to the "normal" behaviour
of a magnetization curve exhibiting a downward curvature (usual paramagnetic or
ferromagnetic system). Such an anomaly is characterized by an upward curvature present
in a limited field range. Two main classes of metamagnetic transitions can then be
distinguished (see fig. 48): (i) a step-like behaviour where the magnetization jumps more
or less suddenly between two well defined plateaus: class (1); (ii) an S-shaped behaviour
where the magnetization increases smoother than in class-(1) systems, showing only an
inflexion point at a given critical field He: class (2). It can be noticed that class-(1)
transitions are often, but not always, associated with a first-order transition, which is never
the case for class-(2) transitions; moreover, class-(1) transitions are generally sharper as
the temperature is lowered, and they often evolve at high temperature towards class-(2)

346

D. GIGNOUX and D. SCHMITT

M

(a)

(b)

class (1)

Hext

2)

Hext

Fig. 48. Schematic representation of: (a) a discontinuous (class (1)), and Co) a continuous (class (2))
metamagnetic transition.

behaviour. More subtle particularities may occur in both classes: for class-(1) materials,
the transition can be discontinuous or can present a portion with a finite slope as a
function of the external (applied) magnetic field, and this will be related to demagnetizing
field effects in the presence of hysteresis (see below); for class-(2) compounds, the
magnetization curve may exhibit an angular point at the critical field, i.e. a discontinuity
of the slope dM/dH at the transition, or keep a smooth variation in the whole field range.
Note that this smooth behaviour may be intrinsic or due to faults or inhomogeneities in the
sample. It is worth noticing that some compounds may present both types of metamagnetic
transitions, on the same magnetization curve or at different temperatures.
The value of the transition field Hc generally depends on temperature, T, and
sometimes on the direction of variation of the field (increasing or decreasing field).
Moreover, some transitions may exist for one single direction of variation. By reporting
the corresponding H - T values, more or less complex phase diagrams can be constructed,
showing one or several distinct regions, each one associated with a particular spin
structure, and separated by boundary lines which are generally first- or second-order
transition lines (see fig. 49). In addition, these phase diagrams may exhibit several critical
points on the transition lines, such as tricritical points (junction between first- and secondorder transition lines), triple points (junction between three first-order transition lines) and
critical or bicritical endpoints (end of a second- or first-order transition line) (Stryjewski
and Giordano 1977). Numerous examples will be shown throughout this section. The
general topology of a phase diagram together with its associated details will reflect the
underlying microscopic properties at the origin of the metamagnetic behaviour. The two
main ingredients playing an essential role in these phenomena are the anisotropy and
the interionic interactions. The former property, i.e. the crystalline electric field (CEF)
coupling, plays a fundamental role by constraining the magnetic moments to lie along
a given direction (axial anisotropy), or within a given plane (planar anisotropy), or
anywhere in space (weak anisotropy). A particular situation of axial anisotropy will
occur when there are several equivalent noncollinear easy magnetization directions, as
for example the two-fold axes in hexagonal or tetragonal symmetry, or the three-fold
axes in cubic symmetry. The one-ion magnetoelastic coupling must be included in this
first category, being considered as the strain derivative of the CEF itself. The second

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

347

first order line

" ~ , , , ~ tricrifical point
X
\

triple point

bicritical
endpoint

',
I

second orderqine
I

Fig. 49. Schematic view of a H - T magnetic

phase diagram with some critical points.

fundamental ingredient includes all types of interionic interactions, namely the isotropic
or anisotropic bilinear exchange coupling and the two-ion quadrupolar coupling, as well
as their possible strain dependence, i.e. the interionic magnetoelastic couplings. These two
basic microscopic properties appear to be the essential characteristics of a metamagnet.
Their relative importance as well as the sign of the different parameters involved will lead
the considered compound to belong to one of the families described below, keeping in
mind that, sometimes, there is no obvious frontier between two different behaviours, and
that an observed transition can result from the combined action of several phenomena.
3.2.1.1. Demagnetizing field effects and hysteresis. Special attention must be paid to the
distinction between internal and external (applied) magnetic field, in particular as far as
a first-order transition is concerned. For an ellipsoidal sample, it is well known that the
magnetization is uniform as a function of the applied field Hext, but actually the magnetic
ions experience an effective internal field, Hi, instead of Hext:
Hi = Hext - Hd = Hext - NdM,

(22)

w h e r e N d is the demagnetizing field factor depending on the shape of the sample and
on the direction of application of the field. The presence of this demagnetizing field
Hd has an important consequence in the case of a first-order transition (Stryjewski
and Giordano 1977). Let us consider such a first-order magnetic transition for the
critical internal field H~ and without any hysteresis (see fig. 50a). Actually this field is
defined by the crossing of the Gibbs free energy variation for both magnetic states under
consideration. The magnetization M then should jump discontinuously from MB to Mc
when the internal field reaches the critical value H~ (vertical line BC in fig. 50a). In
external field, the transition starts for the critical value HCxt = H c + NdMm However, the
transition cannot proceed all at once, because an increase of M would lead to a decrease of
Hi below the value required to induce the transition. The transition therefore must take

348

D. GIGNOUX and D. SCHMITT

M
(a)

~B

A-

A'
C

Hi

(c)

(e)

J.,

Hexl

H~

D'

C;

Hi

(b)

H~

DF

(d)

A~

E'

D'

C'

F'

B~

Hi

Hext

!,'S

(f)

D' E

'

A'

Hext

M
(g)

G
A

(h)

E !

C I_ _

D'

G,I/t

A'

F
Hi

F'
Hext

Fig. 50. Schematic representation of a first-order metamagnetic transition in internal (a, c, e, g) or external (b,
d, f, h) field, and without hysteresis (a, b), with a large hysteresis (c, d, g, h) or with a small hysteresis (e,
f); the distinction between (c, d) and (g, h) arises from the way the domain structure is modified during the
transition (see text).

MAGNETIC PROPERTIESOF INTERMETALLICCOMPOUNDS

349

place gradually, any change 6M occurring only for the corresponding change 6Het which
keeps Hi constant, i.e.:
5M

6Hext

Nd"

(23)

The result is an inclined portion in the magnetization process as a function of the applied
field (line WC t in fig. 50b), the slope corresponding to the reciprocal demagnetizing
field factor 1/Nd. During this portion, the sample breaks into domains of the two phases
coexisting at the first-order transition, a behaviour similar to the initial magnetization
process of a ferromagnet below its Curie temperature. This mixed-phase region is then
associated with all first-order transitions between states of different magnetization but
without any hysteresis.
Let us consider now the case of a first-order magnetic transition with hysteresis. The
crossing of the free energy of both states still defines the theoretical critical field H~.
However, the magnetization does not jump for this critical field, but exhibits a hysteresis
cycle between the two limits H u and H d (see fig. 50c). These fields correspond to the
points where the change of magnetic configuration is initiated in increasing and decreasing
field, respectively. This phenomenon arises when an energy barrier between both states
prevents the system to carry out the transition at H~. The hysteresis width may range from
small to relatively large values, as we will see in the following, depending on the origin of
this barrier and on the temperature. The critical point H~ corresponds to the instability of
the low-field state which may occur either within the whole sample or in some nucleation
centers. In the former situation, and if the hysteresis is wide enough, the full system
suddenly jumps from B to C on the cycle, experiencing a corresponding reduction of the
internal field equal to N a ( M c - M B ) . This is possible because point C corresponds to a
state allowed for the system, i.e. a stable state. This results in a vertical portion WC p
in the magnetization process in external field (fig. 50d). The inverse process occurs in
decreasing field (lines EF and E~Ft). As a consequence, the hysteresis width AHCt in
external field is reduced compared to the theoretical value Zk/-/ic = (H ui - H di ) in internal
field:
AH~xt = AH~ - Nd(Mc - MB).

(24)

In the case of a narrow hysteresis, namely AH~ smaller than N d ( M c - M s ) , the situation
should be more complex, exhibiting both behaviours described above, and should lead
to an unusual hysteresis cycle in external field, where the increasing critical field HeUt is
lower than the decreasing one Hdxt (fig. 50e, f). Moreover, from the above considerations,
it follows that a transition which follows the demagnetizing field factor in external field
should always be associated with a first-order transition without hysteresis, while the
existence of hysteresis should always produce a transition with vertical portions, at least
in part. However this behaviour does not seem to be always the case in experiments,
where transitions following the demagnetizing field factor in external field are often

350

D. GIGNOUXand D. SCHMITT

associated with a strong hysteresis. This means that, generally, the sample does not transit
at once, i.e. the domains where the high-field state is nucleated do not extend suddenly
to the whole crystal: the transition occurs progressively as in the first case described
above, within a finite external field range where the sample is in a mixed-phase state
(fig. 50g, h). The nucleation centers around which the transition is initiated may be any
fault, inhomogeneous region or antiferromagnetic domain wall. In this latter case, the
situation is necessarily different by increasing or decreasing field since the magnetic
structure itself can be very different on both sides of the transition. This difference can
play an important role in the occurrence and width of the hysteresis. Finally, the effective
critical field H u for example may be smaller than the theoretical value corresponding to
the instability of the low-field state, because of the presence of these nucleation centers,
following the same process as in ferromagnets with a hysteresis, where the coercive field
is generally far below the anisotropy field.
3.2.2. C E F metamagnetic systems
The simplest systems where metamagnetism can occur are pure CEF ones, i.e. those
where the CEF is the only interaction acting on the 4f shell. Indeed this single coupling
is sufficient to induce field transitions under certain conditions of anisotropy and field
direction. This effect was first predicted by Cooper (1966) for TmSb. This cubic
compound exhibits a ['1 singlet as ground state, with a M j =4 main component. When
an external field is applied along the hard magnetization direction [001], a crossing of F1
and of one of the levels issued from the F4 excited state (having a Mj = 5 component) is
anticipated at about 40 T, giving rise to a step in the predicted magnetization becoming
more and more pronounced as the temperature is lowered down to 0 K. A second crossing
between F4 and one of the F~2) levels is predicted in still higher fields, resulting in a
second step toward the saturated magnetization (Mj = 6). Due to the low availability of
high magnetic fields at that time, this effect has not been observed experimentally1.
Such a crossing of CEF levels has been observed several years later in Pr metal at a field
of 31.5 T applied along the [001] hard direction of this double-hexagonal close-packed
material (MeEwen et al. 1973). In this system, the ground state at both the cubic and
hexagonal sites is a singlet, and the exchange interaction is undercritical so that there
is no magnetic ordering down to very low temperatures. At 4.2 K and for a field of
31.5 T, the magnetization process along the hard direction exhibits a well-defined step
(see fig. 51), which has been ascribed to a first-order transition on the hexagonal sites.
The size of the magnetization jump is 2.08/~B per hexagonal site, and there is a noticeable
hysteresis of about 0.2 T at 4.2 K, the transition being sharper and the hysteresis larger
at lower temperature. This transition has been explained by the field-induced crossing of
the F1 singlet ground state (Mj = 0) with the F3 excited level (Mj = 3 component) lying at
about 50K above F1, in zero field. This has been confirmed later (Franse et al. 1990).
1 Note added in proof A field-inducedmetamagnetictransition associatedto a CEF level crossing has been
observed very recentlyin TmSb along the [001] direction,for a field of about 48T, in reasonableagreement
with the theoreticalpredictions(Nimoriet al. 1995).

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

351

o II10]
.
[001]
o

c)
,-n
::::t.

F-..

Z
LJ

C3
5-

eeo
O~ooe

"

o
5

10

2O
30
z.0
INTERNAL FIELD (TESLA

Fig. 51. High-field magnetization data for

paramagnetic Pr metal at 4.2K along two
different directions; note the hysteresisof the
first order transition at 31.5T (after McEwen
et al. 1973).

A still more interesting example is provided by the hexagonal compound PrNi5 in

which two different types of CEF metamagnetic transition occur, namely a crossing of
CEF levels along the [100] direction and an anticrossing along the [120] direction (Ball
et al. 1992a). Indeed, in this paramagnet the ground state is the F3 singlet followed by
another singlet F1 around 30 K and two doublets F~1) and F6 at about 50 K. When the
magnetic field is applied in the basal plane, the energy of the first excited level decreases
much faster than that of the ground state so that these two levels should cross for a
given field. However, according to group theoretical considerations, the ground level I"3
transforms into a F2 or 1"5 level under the application of a field along the [100] or [120]
directions, respectively, while the first excited level keeps the Fl symmetry. It follows that
the two levels actually cross each other in the first case, because they belong to different
representations, while in the second case an anticrossing occurs because they both belong
to the same representation (see fig. 52). This leads to a metamagnetic transition which
becomes sharper and sharper in the first case when the temperature is lowered, finally
reaching a stepwise behaviour at 0 K, while in the second case the transition remains
smooth even at 0 K. Both transitions have been experimentally observed, for magnetic
fields of 18.5 T and 12 T, respectively (fig. 53). It turns out that such behaviours (crossing
or anticrossing) give useful information about the CEF level scheme because the value
of the critical field of the metamagnetic transitions is directly related to the zero-field
energy position of the excited level which crosses or anticrosses the ground state. This
is particularly true in PrNis, where the composition of the 1`1 excited level forbids one
to observe directly a transition from the ground state by neutron spectroscopy in zero
field. However, this transition becomes possible for nonzero magnetic field and has been
actually observed by point-contact spectroscopy (see sect. 3.1, fig. 43) (Reiffers et al.
1989). This allowed one to slightly adjust the CEF parameters determined previously
(Ball et al. 1992a).
In a general way, both types of CEF metamagnetic transitions occur when a nonor weakly magnetic ground state evolves into a stronger magnetic ground state as
a function of the magnetic field, the continuous or discontinuous character of the
transition depending on the symmetry of the levels, i.e. of the detailed composition in

352

D. GIGNOUXand D. SCHMITT
4

3 _

PrNi

PrNi s [1001

54.2K ......~~ i - 2 " 0

i ....

"

/~.~._~ 1.4K
4.2K / ~ . ~ . . . .

.f

"

1.4K and 0.3K

.__~~F

-100

(c)

[120]

e~ -120
-140
-160
-180

10
20
30
Applied Field (Tesla)

10
20
30
Applied Field (Tesla)

40

Fig. 52. Calculated field variations of the magnetizationand the energy of the first two CEF levels in PrNi5
along both symmetrydirections of the basal plane and at various temperatures; note the different behaviourof
the metamagnetictransitions associatedwith the crossing or anticrossingof the levels(after Ball et al. 1992a).
Mj components. According to the sequence of the CEF levels and to the direction of the
applied field, one or several successive stepwise transitions can be expected, associated
with several successive levels crossings or anticrossings (Mrachkov and Leyarovski 1988).
In fact, the metamagnetic behaviour should be present as soon as the CEF ground
state eigenfunction does n o t include the Mj = J component as its major part, for the
considered field direction. There always exists a direction fulfilling this condition, even
possibly along the easy magnetization direction, although the corresponding critical fields
often are much too high to be reached experimentally. In addition, as seen above, the
sharpness of the metamagnetic behaviour increases when the temperature decreases,
so that a low temperature is often needed to observe them. That explains why both
examples where these transitions were observed (see above) are paramagnets, having
a nonmagnetic singlet as ground state, because this particular CEF configuration more
easily allows the exchange interaction to be undercritical with regard to the occurrence
of magnetic ordering. A measurement at low temperature can then be made without
being disturbed by the latter. However, one can imagine that the increasing availability
of both high magnetic fields and low temperatures in the future will allow one to use
these properties to get information about the CEF level scheme under magnetic field,

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

i

353

PrNi 1.4K
5

t120]/'~i
/

................

.'

[100]

J
f

/j/

Go

PrNi 4.2K
5

[120]

#
o
o

[lOO]

o
o

o
o
8

[001]

s
o eoeo*

10

I
20

I
30

Applied Field (Tesla)

40

Fig. 53. High-field magnetization curves in

PrNi 5 along the two main symmetry directions
of the basal plane at 1.5 K and 4.2 K (after Ball
etal. 1992a).

similarly as neutron spectroscopy does in zero field. Moreover, and as already mentioned
by Cooper (1966), the field-induced transition is generally associated with an anomaly of
the temperature dependence of the magnetic moment, namely the presence of a maximum
which is particularly strong for a magnetic field just below the critical field: this effect
is obvious when the successive isothermal curves in fig. 52 are considered. It can be
noticed that this maximum is not always present on the susceptibility curve itself, i.e.
in the zero-field limit, in particular in the case of a level crossing. However, when
the susceptibility variation exhibits such a maximum, it is generally associated with a
change of sign of the third-order magnetic susceptibility (Morin et al. 1988): below the
temperature where this maximum occurs, the third-order susceptibility becomes positive,
i.e. the initial magnetization process exhibits a positive curvature which is the signature
of a metamagnetic transition in higher field.
Another interesting effect related to a CEF level crossing may appear in the presence
of nonzero magnetic interactions and has been studied by Tsuneto and Murao (1971)
in the case of a S = 1 system. R has been shown that, in the vicinity of the crossing
point, small ferromagnetic exchange interactions may be at the origin of a field-induced
second-order magnetic ordering. This can be explained as follows: within the mean-field
picture, the usual magnetic ordering occurs when the initial susceptibility increases, by

354

D. GIGNOUX and D. SCHMITT

(a)

(b)

T<T

AM

H(T) H

H~

Hc(T) Hc

H.
I

H
critical p ~ . ~

endpoint

critical point

endpoint

H(T)
metamagnetic
transition line
first order transition line

I
Te

metamagnetic
transition line
first order transition lines

(c)
T

l
T

(d)
T

Fig. 54. Schematic determination of the occurrence of a field-induced magnetic transition near a CEF level
crossing (see text); (a) situation at the critical temperature To; (b) situation below Tc. The slope of the straight
line which defines the critical field Hc(T ) is 1/n. Temperature dependence of the critical field in the case of
(c) ferromagnetic or (d) antiferromagnetic interactions.

decreasing temperature, above a certain threshold value proportional to the reciprocal

exchange coefficient 1 / n . In the present case, the superimposed susceptibility at the
crossing field diverges at 0 K so that, provided an applied field polarizes the system near
this point, there exists also a critical temperature Tc where this susceptibility reaches the
same threshold value 1 / n , resulting in a second-order transition (see fig. 54a). Below this
temperature, the order parameter M experiences an additional variation AM compared
to the expected field-induced dependence, for a critical field He(T) slightly smaller than
He(T~) (fig. 54b). It is noted that the critical field Hc(Tc) includes a contribution of the
ferromagnetic exchange coefficient n through the relation
Hc(Tc)

= H o
e - nMc,

(25)

where H is the metamagnetic critical field without interactions and Mc is the associated
magnetization. This shift of the critical field is effective in fact over the whole
temperature range of the transition. The corresponding temperature dependence of

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

355

He(T) is schematized in fig. 54c. Below the critical temperature To, a first-order
transition line separates two ferromagnetic phases with different magnetization, while
above Tc, the line corresponds to the normal metamagnetic field; this latter vanishes
above a given temperature where there is no longer any metamagnetic transition due
to the thermal effects on all the CEF levels. Note that this metamagnetic field is
temperature independent, because it Js related to a level crossing which does not depend
on temperature. In the case of predominant antiferromagnetic exchange interactions, the
behaviour should be still more exotic, because the same effect should lead to a two-site
ferromagnetic ordering below a critical temperature Tc and for a critical field He, where
the antiferromagnetic component of the magnetization should follow the corresponding
propagation vector maximizing J(q). A second transition line then should close this
intermediate phase above the previous one (fig. 54d). Although such phenomena have
not yet been observed experimentally, one can anticipate that the increasing availability
of high magnetic fields together with low temperatures will allow us to confirm the above
assumptions in the future. A possible example could be the PrNi5 compound, where
a small ferromagnetic Coupling is present (see above in this section): such an effect
should occur along the [100] direction (see fig. 52), but below a critical temperature Tc
estimated to be about 1 K.
3.2.3. Quadrupolar metamagnetic systems
The CEF interaction is not the only coupling able to induce metamagnetic processes
within the paramagnetic phase. In fact, any coupling of rank 2 or higher may potentially produce metamagnetic transitions, through nonlinear effects on the magnetization.
Among them, the best known is magnetoelastic coupling or, more generally, quadrupolar
interactions. By using a perturbation theory, it has been shown that these latter provide an
additional contribution to the third-order magnetic susceptibility X~) (Morin and Schmitt
1990). According to the sign of this term and to its magnitude compared to the pure
CEF contribution (see above), the total third-order susceptibility may become positive,
i.e. the initial curvature of the magnetization process may become positive, giving rise to
the occurrence of a metamagnetic transition at a certain field. This transition can be either
smooth, class (2), see sect. 3.2.1, when the quadrupolar coupling only speeds up slightly
the "natural" magnetization process, or more discontinuous, class (1), when the strength
of this coupling exceeds a critical value which depends on the magnetic and quadrupolar
character of the CEF ground state.
Examples where the quadrupolar metamagnetic transition is of class (2) include
mainly cubic compounds such as YmCu (Jaussaud et al. 1980), PrPb3 (Morin et al.
1982), TbP (Raffius and Krtzler 1983), DySb (Al~onard et al. 1984a), ErAI2 (del
Moral et al. 1987), TmGa3 (Morin et al. 1987a) CeMg and CeZn (A16onard and Morin
1990). The predominance of cubic compounds is explained by the fact that for other
symmetries the quadrupotar moment is already ordered by the crystal field; then the
effects of magnetoelastic couplings are more difficult to be observed. In all these
cubic compounds, the third-order magnetic susceptibility becomes positive in a limited

356

D. GIGNOUX and D. SCHMITT

to

bo

-5
0

10

r(K)

15

20

25

Fig. 55. Temperature variation of the thirdorder magnetic susceptibility for a magnetic
field applied along a four-fold axis in cubic
TmCu; solid lines are calculated taking into
account the quadrupolar parameter Gj indicated
(after Jaussaud et al. 1980).

temperature range above the ordering temperature, but its value remains quite weak.
Therefore, the associated positive curvature of the magnetization process is barely visible
on the curves themselves, and only a detailed analysis can provide a reliable value for
X~ ) and thus a confirmation of the metamagnetic behaviour. As an example, fig. 55
shows the dependence of this third-order susceptibility in TmCu. In this compound,
the metamagnetic transition occurs only along the four-fold axis, the corresponding
quadrupolar parameter G1 = 10inK being positive and large enough. Along the other
main symmetry direction, i.e. the three-fold one, the trigonal quadrupolar parameter
G2 is negative, preventing the metamagnetic transition to appear. This is a general
rule in lanthanide intermetallic compounds: the magnetic processes are anisotropic, at
least in high field, therefore searching for a metamagnetic transition requires one to
investigate successively each of the independent symmetry directions. In TmGa3, it
turns out that the direction where a slight metamagnetie transition occurs just above the
ordering temperature is the three-fold one, indicating that in this compound the trigonal
quadrupolar coupling is involved in this behaviour. The same is true for the pnietide TbP.
It is worth noting that the occurrence of this metamagnetic transition, i.e. of a positive
third-order magnetic susceptibility, in the close vicinity of the magnetic ordering leads in
principle to a first-order magnetic transition (Morin and Schmitt 1990). This is effectively
the case for TmCu, TbP, DySb, CeMg and CeZn, all ordering antiferromagnetically within
a first-order transition.
A more spectacular example is provided by the cubic compound TmZn (Morin et al.
1978) which exhibits a quadrupolar metamagnetic behaviour of class (1). This compound
orders ferromagnetically at Tc = 8.1 K, but the quadrupolar interactions are large enough to
induce a ferroquadrupolar ordering at the slightly higher temperature TQ = 8.55 K. Above
this latter temperature, i.e. in the cubic paramagnetic phase, a single-step metamagnetic
transition occurs along the [001] easy magnetization direction (see fig. 56). In this
compound, the CEF ground-state triplet F~1) consists of a magnetic doublet with a
dominant IMj = + 5 ) component and a nonmagnetic but strongly quadrupolar singlet
having a main IMj = 6) character. In low field, the ground state arises from the doublet

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

357

6
A

:P
4

O
)-<~
N

,<
l[

0
20
40
60
80
100
120
Fig. 56. Magnetizationcurvesin cubic TmZnalongthe [001] directionat varioustemperatures:the insetshows
the temperaturedependenceof the critical field(afterMorinet al. 1978).
state within the triplet. The observed field-induced first-order transition takes place when
the internal field exceeds the critical field necessary to induce the ferromagnetic ordering
on the singlet component of the triplet, this phenomenon being strengthened by the
quadrupolar character of this singlet together with the large positive quadrupolar coupling
present in this compound (Morin and Schmitt 1990). Moreover, this first-order transition
is replaced by a second-order one at a higher temperature, due to thermal averaging
effects on the involved crystal-field levels. A similar behaviour exists in the isomorphous
compound TmCd (A16onard and Morin 1979), although in this latter compound, the only
phase transition is a ferroquadrupolar ordering which takes place at a lower temperature
than in TmZn (TQ = 3.16 K). The temperature dependence of the metamagnetic critical
field of both compounds leads to the phase diagram shown in fig. 57 (L/ithi et al. 1979).
This kind of metamagnetic behaviour is characterized by a critical line starting from
the ordering temperature and exhibiting a critical endpoint at the temperature where the
third-order magnetic susceptibility becomes negative (these endpoints are out of scale in
fig. 57).
A very specific type of metamagnetic transition related to the quadrupotar interactions
occurs within the quadrupolar phase itself, i.e. when the quadrupolar moment is ordered
but the magnetic moment is not. This was observed for example in CeAg, where the wide

358

D. GIGNOUXand D. SCHMITT

2O
Ta(K)
o

15

TmZn

~.o

10

~ ~ T m C d

B (T)

Fig. 57. Magneticphasediagramfor TrnZnand TmCd;

the critical fields are determined by magnetization
(open circles), elastic constants (open triangles) and
electrical resistivity (solid circles); solid lines are
calculated (afterLfithiet al. 1979).

range of stability of this phase allowed a detailed investigation (Morin 1988). In this cubic
compound, the ground state is a F8 quartet which splits into two doublets at TQ = 15.85 K
due to the strong quadrupolar coupling associated with the tetragonal symmetry mode
(<O) is the order parameter). This cubic-tetragonal transition is accompanied by a
huge tetragonal distortion due to a noticeable magnetoelastic coefficient (c/a- 1 ~ 1.7%
at 1.5 K). The ferromagnetic interactions present in this compound lead to the ordering
of the magnetic moments far below TQ, namely Te = 5.2 K. In the absence of any external
stress, a strain domain structure takes place between Tc and TQ, i.e. three tetragonal
domains are present, corresponding to the three equivalent four-fold directions of the
initial cubic symmetry. When a magnetic field is applied along the [001] direction, it turns
out that it is parallel to the local c axis for one domain but perpendicular to it for the
two others. In this latter case, the field is then aligned with a local tetragonal a axis, and
a sudden magnetization jump occurs as soon as the Zeeman and exchange energies are
strong enough to destroy the initial quadrupolar ordering: at the corresponding critical
field the a axis transforms into a c axis (see fig. 58). An additional nonreproducible
hysteresis is associated with this strong metamagnetic behaviour, due to the actual
repartition of the virgin sample into the three tetragonal domains. In the corresponding
phase diagram, the critical field first increases between Tc and 10K, then decreases
up to TQ. This anomalous behaviour has been explained by the different temperature
dependences of the quadrupolar and bilinear energies: the latter, which decreases more
rapidly than the former, requires an additional amount of external field to produce the
transition in intermediate temperature range. Note that a similar metamagnetic behaviour
also occurs within the ferromagnetic phase, i.e. below Te, the only difference being the
presence of a spontaneous ferromagnetic component.

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS
1.5

/ /

1.0
//

.,

//

6K
~.~..--~,----

I
I

359

1,5

l
~1.0

.....

...= . . < ....-I./"~.~-~.

, , ~ . . t . .............. "...................
~---~ . -_-_-:4~
-[
.---.~~'~_............
~--~
1
.-41.25

0.5

--0

/2:: ....

.!t- j~

/..:"

/,-:."
,~::'"

--

:E

//

./

as
I

9K

~J

I - 1.0

0.75
0.5 -

/~<:;_

.....

_ . ~ . ~

........ "-:

).5
l///
/ " ' "
~ -'<'"
O~
q //""

OI,//

210

~
'

----{0011
----[101]
----[1111

'

LIo
H IkOe)

60

l__

Fig. 58. Magnetization curves in the

quadrupolarphase of cubic CeAg (after
Morin 1988).

Another case of metamagnetism should occur in antiferroquadrupolar phases. In

this case, only the quadrupolar moments are ordered, but within at least two different
sublattices having a different local c axis. A sufficiently large magnetic field is then
expected to destroy this antiferroquadrupolar arrangement, resulting in a metamagnetic
transition. Unfortunately, only a few examples are known in the literature, namely
CeB6, TmGa3 and PrPb3 (see Morin and Schmitt 1990). Moreover, the temperature
range of the considered phase is very narrow in TmGa3: between 4.26 and 4.29 K, and
the metamagnetic processes have been observed in fact within the antiferromagnefic
phase, i.e. below 4.26K (see sect. 3.2.12). The case of CeB6 is complicated by the
Kondo behaviour of the cerium ions, and the phase diagram suggests that a huge
critical field is necessary to destroy the antiferroquadrupolar state (Effantin et al. 1985).
In PrPb3, the ordering temperature is low (TQ=0.35K), but a preliminary magnetic
measurement seems to indicate the expected metamagnetic behaviour around 7 T at 0.1 K
(Morin 1993).
3.2.4. Ferromagnetic metamagnetic systems
In both preceding sections, only paramagnetic systems were concerned, namely either
strictly nonordered compounds or quadrupolarly ordered but still paramagnetic compounds. In the following sections we will focus on magnetically ordered systems. The first
group considered includes different types of ferromagnetic compounds. In the majority

360

D. GIGNOUXand D. SCHMITT

of cases, the metamagnetic behaviour is basically associated with crystal-field properties,

namely a crossing or an anticrossing of CEF levels. The only complication arises
from the presence of a ferromagnetic exchange field which implies that the magnetic
moments are not always parallel to the applied magnetic field. Actually, in most cases,
the metamagnetic process occurs along a hard magnetization direction, i.e. during the
rotation of the moments from the easy axis toward the applied field. To our knowledge,
there is no experimental evidence of CEF level crossing in a ferromagnet along an easy
magnetization direction. Examples where such metamagnetic processes have been well
studied included several cubic systems, such as the RZn and RA12 compounds. In ErZn
(Te=20.6K), the moments are ordered along the [001] four-fold direction. When the
magnetic field is applied along the [111] hard axis, the magnetization increases suddenly
at a critical value He which depends on the temperature (He = 20 kOe at 4.2 K, Morin and
Schmitt 1978). The associated jump in the magnetization is moderated (AM ~ 0.5/%/at),
because the field-induced transition occurs at the end of the rotation process. In HoZn, the
same behaviour is observed, but the metamagnetism takes place along the [111] direction
at low temperature (below TR =23 K) and along the [I01] direction at high temperature,
i.e. between TR and Te=74K (Morin and Schmitt 1978). This is related to the change of
easy magnetization direction which occurs precisely at TR from the [101] axis toward the
[111] axis. Here also, the jump of magnetization is small but relatively sudden, and the
critical fields remain less than 10kOe. In ERA12 (To = 14K), the same behaviour occurs
during the rotation of the moments from the [111] easy axis toward the [001] direction,
albeit at a much larger critical field (~80kOe at 4.2K, Purwins et al. 1976). In DyA12
(Tc=64K), the metamagnetic transition observed along the [111] direction exhibits a
clear first-order character with a noticeable discontinuous jump in the magnetization
(AM ~ 3/% at 4.2K, Barbara et al. 1978). Such first- or second-order transitions have
been widely studied and have been explained in the frame of a Hamiltonian which
includes the effects of the magnetocrystalline anisotropy in the presence of an effective
(applied plus exchange) magnetic field. In some cases, additional couplings have been
considered, e.g. the quadrupolar coupling in ErZn and HoZn, but it is worth noting that
the metamagnetic phenomena are basically related to the CEF anisotropy, i.e. to a strongly
distorted dependence of the energy on the spatial position of the moments due to the highorder CEF terms.
The same is true in lower symmetry compounds, such as in hexagonal systems
where, due to the uniaxial character of the anisotropy, the metamagnetic transitions can
be more spectacular. In HoRu2 (To = 18K), the moments rotate suddenly at a critical
field of about 15 kOe at 4.2 K from near the a easy axis towards the c direction, i.e.
over almost 90, resulting in a magnetization jump of more than 6/tB (see fig. 59;
Okamoto et al. 1985). As in cubic systems, the explanation of this phenomenon has
been given in terms of a cross-over between the ground state and the first excited state,
this latter evolving to a more and more magnetic state as the field increases along the
hard magnetization direction. According to the exact composition of the CEF levels,
this evolution may be more or less abrupt, as shown by the neighbouring compounds
DyRu2 (Te = 30 K) and ErRu2 (Te = 8 K) where the metamagnetic transition is smoother

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

361

10

:?:
m

..m

m.

"c

HRu2

Fig. 59. Magnetization curves of HoRu 2

at 4.2 K along the three main symmetry

2
0

10

20

40

30

50

H,,.. (kOe)

10

'

'

'

'

'

'

'

. . . .

directions of the hexagonal cell. Solid lines

are calculated curves including crystal field
and exchange couplings (after Okamoto
et al. 1985).

J ' ' '

Z
8
LLI
rfl

g6
~"

ErNi 5

/"
0
0

T=1.4K
,

10

20

30

H(Tesla)

1-

40

Fig. 60. Magnetization curves of ErNi s at 1.4 K along

the three main symmetry directions of the hexagonal
cell. Solid and dashed lines are calculated curves
for two different sets of crystal field and exchange
parameters (after Zhang et al. 1994).

at 4.2K than in HoRu2 (Andoh 1987). However, care should be taken to distinguish
between a true second-order transition and a smooth transition arising from thermal
effects. The relation between the sharpness of the transition and the exact composition of
the CEF levels is emphasized in the other hexagonal compound ErNi5 (Zhang et at. 1994).
In this latter compound (To = 9 K), most of the magnetic properties are well described
by two different sets of CEF parameters which differ only by the nature of the second
excited state lying at about 70 K above the ground state. This apparently minor difference
actually leads to a drastically different behaviour on the magnetization curves along the
[100] and [ 120] hard directions, the sharpness of the metamagnetic processes occurring in
a high field (around 20 T) being strongly reduced for one set of parameters, in agreement
with experiment (see fig. 60). This underlines the interest to use as much experimental
information as possible to obtain a reliable determination of the crystal field.
The same behaviour as in HoRu2 is present in the orthorhombic compound HoGa, when
the magnetic field is applied along the b hard magnetization direction (Shohata 1977).
In other compounds of the same series, i.e. PrGa, NdGa, SmGa, TbGa and DyGa, the
metamagnetic transition occurs along the easy axis and is accompanied with a hysteresis

362

D. GIGNOUXand D. SCHMITT

-40

~
-20

I0.

~.
b

-5"

[SmGo ]

20

40

Heft (kOe)

-I0-15-

Fig. 61. Magnetizationcurvesof SmGa at 4.2 K alongthe three main symmetrydirectionsof the orthorhombie
cell. Solid lines are guides for the eye (after Shohata 1977).

cycle and a remanent magnetization, indicative of domain-wall freezing by the large

anisotropy energy. The coercive field depends on the lanthanide, ranging from less than
1 kOe (Pr, Nd) to 47kOe (Sin) (fig. 61). This kind of metamagnetic behaviour is not
intrinsically related to microscopic properties but due to macroscopic effects associated
with the dynamics of the ferromagnetic domains. In ErGa, the metamagnetic process
observed along the a hard magnetization direction has a character different from that
of the other compounds within the series, and has been interpreted as the signature of
a noncollinear ferromagnetic structure (Shohata 1977). The same explanation is valid
in the orthorhombic RNi series (Gignoux and Lemaire 1974). In these low-symmetry
systems, the local easy axes are not all parallel to each other so that the magnetic
lanthanide ions can be divided into two sublattices. Ferromagnetic interactions then
result in a noncollinear (canted) moment arrangement, with both ferromagnetic and
antiferromagnetie components. When the magnetic field is applied along this latter one,
and if the magnetocrystalline anisotropy is large enough, a spin-flip mechanism may
occur, corresponding to the reversal of the moments of one sublattice. This is the case of
ErNi along the a axis, or of DyNi along the c axis (see fig. 62; Sato et al. 1986). Other
types of moment reorganization may appear in low-symmetry compounds with more than
one crystallographic site for the magnetic lanthanide ion, leading also to metamagnetic
behaviour. For example, in rhombohedral HoNi3, the lanthanide atoms are divided in two
sublattices with a different point symmetry and consequently a different ordered moment
(9.3 and 7.9ktB at 4.2K along the b axis; Hashimoto 1988). Along the a and c axes,
several steps have been observed on the magnetization processes, probably related to the

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

6 ~

......

Mc

~4.2K

=-

I bt;

..-"

"" "'56K

0
0

363

20

z,0
H(kOe

60
)

80

Fig. 62. Field dependences of the magnetization along the c axis in orthorhombic
DyNi (after Sato et al. 1986).

existence of the two sites with different anisotropy, but the exact nature of the intermediate
states has not been determined.
3.2.5. Weakly anisotropic metamagnetic systems
The gadolinium-based antiferromagnetic compounds constitute special metamagnetic
systems owing to their very weak magnetocrystalline anisotropy. Indeed, as the magnetic
moments are nearly free to rotate in any direction, the magnetic behaviour under field
reflects only the way the antiferromagnetic structure is destroyed, i.e. it characterizes the
properties of the exchange interactions alone. However, in general a small anisotropy
subsists (mainly of exchange origin), which fixes the zero-field orientation of the magnetic
structure with regard to the lattice, and which is revealed in low field through a spinflop metamagnetic transition. During this transition, there is typically a rotation of
the moments toward a direction perpendicular to the applied field, without drastically
changing the position of the spins relatively to each other. This occurs because, in
most cases, the perpendicular susceptibility is larger than the longitudinal one. The
associated metamagnetic field therefore is the critical field needed to overcome the
residual lattice-related anisotropy. A good example is the orthorhombic GdCu6 compound
(TN = 16K; Takayanagi et al. 1989), where the magnetization curves along the b and
c axes exhibit small discontinuous steps around 3.5 T before varying linearly up to the
saturated value reached at Hs = 24 T (see fig. 63). The absence of a transition along the
a direction suggests a planar symmetry for the magnetic structure, for example a helical
arrangement of the moments within the b-c plane, because it is the only direction for
which all the moments are initially perpendicular to the applied field. In the case of
a simple antiferromagnet, the spin-flop transition should occur along one single axis,
namely parallel to the direction of the moments, at least in low-symmetry systems. If the
symmetry is higher and if there are several equivalent directions for the antiferromagnetic

D. GIGNOUX and D. SCHMITT

364

./

/,
/'/
o<ioo
1.3K

<.~///
b~ . /
7--1/

//./"
:x2

/J
00

48
12
Magnetic Field { T

Fig. 63. Magnetization curves along each principal axis in

the orthorhombic GdCu 6 compound at 1.3 K; note the spinflop transition along the b and c axes (after Takayanagi
et al. 1989).

structure, some spin-flop transitions may correspond to change in the magnetic domain
configuration, i.e. without changing the local symmetry for the moments. A more detailed
analysis of the magnetization processes as well as a good knowledge of the magnetic
structure itself is then needed to fully explain the observed behaviours.
Once the perpendicular configuration is achieved, i.e. above the spin-flop transition,
any anomaly on the magnetization curves will be associated with the destruction of the
antiferromagnetic structure itself. It is worth noting that, for a simple antiferromagnet or
a helical structure, the magnetization process is, in principle, strictly linear. Therefore,
any observed metamagnetic transition will suggest a more complex structure or will be
the signature of anisotropic exchange interactions. For example, in the orthorhombic
GdCu2 compound (TN=42K; Borombaev et al. 1987), only a tiny lattice-related
anisotropy exists as shown by the small differences between the three main axes, but
the field-induced ferromagnetism is reached through two relatively sharp steps occurring
at 68 and 95 kOe (see fig. 64). These steps have been interpreted as transitions from a
spiral structure towards a ferrimagnetic structure and then to the induced ferromagnetic
state. Although the magnetic structure is not precisely known, it can be anticipated
GdCu2

4.2 K

j4'

:::,,:

[.<

b~

z
,<

7
~":'"

I
50

""i""
100

C laXls I
750

APPLIED MAGNETIC FIELD (kOe)

Fig. 64. Magnetization curves along each principal axis

in the orthorhombic GdCu2 compound at 4.2 K (after
Borombaev et al. 1987).

M A G N E T I C PROPERTIES OF INTERMETALLIC COMPOUNDS

i

GdGa2
Z
<

/:
Z
<

365

1.5 K
c axis

//"
/i
I
10

I
20

I
30

MAGNETIC FIELD ( T )

40

Fig. 65. High-fieldmagnetizationcurve along the

c axis in hexagonal GdGa~ at 1.5K (after Ball
1993).

that it should be incommensurate, as observed in several neighbouring compounds

within the RCu2 series (see sects. 3.2.8 and 3.2.10). However, the magnetic processes
observed, as well as the specific heat variation (Luong et al. 1985), are not consistent
with a spiral structure, at least in zero field. As described by Blanco et al. (1991a),
the reduction of the height of the ~, anomaly at TN is well explained if a collinear
(modulated) incommensurate structure is assumed rather than a helical one. It can
therefore be assumed that the observed metamagnetic process is more complex than
initially interpreted. The metamagnetic transitions may be also much more subtle than
in GdCu2, as in the hexagonal GdGa2 compound, where several successive anomalies
have been seen along each of the three main symmetry directions (see for example
the c axis in fig. 65), a situation quite surprising in a gadolinium compound (Ball
1993). Below TN =22 K, this compound exhibits an incommensurate magnetic structure
with propagation vector Q= (0.39, 0.39, 0). Thermodynamic considerations have shown
that this structure is not helimagnetic but amplitude modulated, then antiphase at low
temperature (Bouvier et al. 1991, Blanco et al. 1991 a). The anomalies in the magnetization
processes manifest themselves either as small discontinuous steps with a possible
hysteresis (AM ~ 0.2#B/Gd), or as a change of slope or curvature, and lead to a rather
complex magnetic phase diagram (fig. 66). This complex behaviour has been attributed
to slight modifications of the magnetic structure under field involving transitions from
antiphase to different successive noncollinear (fan) structures, and suggests the presence
of an anisotropie exchange coupling. In the cubic G d B 6 compound (TN = 15.4 K), apart
from a spin-flop transition occurring at about 6 kOe along the four-fold axis and related
to domain effects (Nozaki et al. 1980), small anomalous steps have been observed near
the saturation field Hs =490 kOe. They have been explained by the local spin-flip of
disordered Gd ions located at the B6 cluster site (Sugiyama et al. t988). Although they
are not intrinsic, such disorder effects have to be kept in mind when considering small
steps in the magnetization processes of these nearly isotropie systems.
Finally, in all the above examples single crystals were concerned, and the results
along the various crystallographic directions have shown that, even if there are some
minor differences between them, they all give the same principal features. As a

366

D. GIGNOUX and D. SCHMITT

80

. . . .

. . . .

,i

60-

~ L

40

400

,
i

~~
~h

GdGa 2
.
300 [ 0 0 1 ]

20

~L,

~ 200 "

",,

10

30

20

100
0

10

20

T(K)

30

Fig. 66. Magnetic phase diagram along

the c axis in GdGa2; open circles
correspond to the anomalies in the
magnetization process. Note the small
region near 4 K - 3 0 k O e (after Ball
1993).

consequence, measuring polycrystalline materials should already give a good picture of

the main characteristics of the magnetization processes. For example, some metamagnetic
transitions appear to be relatively smooth, as in GdNizSi2 and GdCu2Si2 (Barandiaran
et al. 1988), in GdRu2Si2 (Czjzek et al. 1989) or in GdPt2Si2 (Gignoux et al. 1991c),
while others are much sharper, as in GdMg under pressure (Liu et al. 1986) In the
hexagonal Gd2In compound, the Gd ions are at two different crystallographic sites, and
two magnetic phases have been established below Tc = 187 K (McAlister 1984). Between
Te and Tt = 99.5 K the compound is a ferromagnet, while below Tt the magnetic structure
has been interpreted as being spiral, with an abrupt metamagnetic transition at 10 kOe, the
amplitude of the magnetization jump being more than 5#B/Gd (see fig. 67). Moreover,
the slope of the magnetization curve during the transition is close to the initial slope
in the ferromagnetic phase (demagnetizing field factor), so that the transition should
be nearly vertical once drawn as a function of the internal field. It has been suggested
that this behaviour is due to the collapse of the spiral structure into a ferromagnetic
one, i.e. a first-order field-induced transition with a change of the propagation vector.

4.2 K

Gd2In

o . . . . .

oo

ee

oO
o

oo

~oo

140 K

<
eq

F.
z
<

oO

~a

. . . . ~
5

APPLIED

288 K

o
I0

MAGNETIC

15

F I E L D (kOe)

20

Fig. 67. Field dependence of the magnetization on

polycrystalline GdEIn at several temperatures (after
McAlister 1984).

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

367

It can be noticed that an analogous strong metamagnetic transition of similar amplitude

had been previously observed in orthorhombic GdA1 (B~cle et al. 1970). In conclusion,
it is clear that the absence of magnetocrystalline anisotropy in gadolinium compounds
does not necessarily facilitate the understanding of magnetic behaviour under field. In
fact, the situation is even more complicated because several types of coupling generally
neglected should now be taken into consideration. Moreover, it should be noticed that a
good determination of the magnetic structure remains important to describe the properties,
and this has been done in Gd compounds by carrying out neutron diffraction experiments
using a Gd isotope or at a very low wavelength ()~ ~ 0.5 A), because of the huge absorption
cross-section of natural gadolinium in normal conditions. This explains the small number
of such experiments in the literature. As a summary, the weak lattice-related anisotropy
in these systems manifests itself by : (i) spin-flop transitions in low field to position
the magnetic structure in the whole in a proper configuration relative to the magnetic
field; (ii) small differences in the magnetization along the various crystallographic
directions. Oppositely, the exchange-related anisotropy, even weak, deeply affects the way
the magnetic structure is destroyed, resulting in magnetization processes often far from
linearity.
3.2.6. Spin-flip metamagnetic systems: simple antiferromagnets
In all of the lanthanide intermetallic compounds except those with gadolinium, the
presence of a strong magnetocrystalline anisotropy favours an easy axis or an easy plane
for the magnetic moments, according to the crystal symmetry and the sign of the related
CEF parameters. As a consequence, in the case of an easy-axis anisotropy and in the
presence of simple antiferromagnetic interactions, spin-flip metamagnetic transitions may
occur when the magnetic field is applied along the direction of the moments. These
transitions are characterized by a simple reversal of all the moments antiparallel to the
field. The most spectacular case should correspond to simple antiferromagnets having
a strong uniaxial anisotropy, e.g. hexagonal or tetragonal systems with a c easy axis,
where the transition would be single-step from a low magnetization state toward a
fully saturated state. It turns out that examples showing such behaviour are scarce in
lanthanide intermetallic compounds, while several have been investigated for a long time
in semimetals or insulators, such as the lanthanide phosphates or vanadates. With the
increasing number of new compounds being studied using single crystals, in particular
in ternary systems, such examples should be available in the future in lanthanide
intermetallics.
Very surprisingly, spin-flip metamagnetic transitions showing two steps have been
found in few simple antiferromagnets. For example, the body-centered tetragonal
compound DyCo2Si2 orders below TN = 24 K within an antiferromagnetic structure with
the propagation vector Q= (0, 0, 1), the magnetic moments being along the c direction.
The sequence of the moments then is + - + - along the propagation vector, in zero field. At
low temperature, the magnetization process along the same direction exhibits a two-step
metamagnetic transition (see fig. 68), with an intermediate region between Hcl = 23 kOe

368

D. GIGNOUX and D. SCHMITT

i

I [o01]
12

~I

DyCo2Si2

o Dy

4
Uoo]

ErGo
2 / BK'~

HI/<O01>
7 '- '--~ "~

Co
* Si

10

1111

1111 ~,

~_

FoO-OO- o - o - -o -o J

T:4.2K

t.lrl-Ii

i~. i i . J I I 1 "

"-T~'~"M

'

20

'

40

He(kOe)

60

12

APPLIED

16
FIELD

20

24

28

(kOe)

Fig. 68. Magnetization curves along and perpendicular

to the c axis in the tetragonal compound DyCo2Si2
at 4,2 K; the inset represents the zero field magnetic Fig. 69. Magnetization along the easy e axis in the
structure; dashed lines are calculated (after Iwata et al. hexagonal ErGa: compound at various temperatures
1990).
(after Doukour6 and Gignoux 1982).

and He2 = 44 kOe where the magnetization reaches 5/~B/Dy, i.e. exactly one half of the
saturated value (Iwata et al. 1990). This plateau suggests that one half of the moments
antiparallel to the field have been flipped, and a sequence like + + + - for the moments
along the c axis has been proposed and found consistent with molecular-field calculations
involving four exchange parameters and crystal-field effects. The intermediate magnetic
structure remains nevertheless to be confirmed by neutron diffraction under field2.
Another similar behaviour has been observed in the simple hexagonal compound ErGaa
(Doukour6 and Gignoux 1982). In this latter compound, the zero-field magnetic structure
is characterized by the propagation vector Q = (0,1, 0), the moments being perpendicular
to the basal plane. When the field is applied along the direction of the moments, a two-step
process occurs, the critical fields being here about 7 and 20 kOe while the value of the
intermediate magnetization reaches one half of the saturated value, as in DyCo2Si2 (see
fig. 69). In the intermediate phase, as shown by neutron diffraction on a single crystal, one
half of the moments antiparallel to the field have been flipped, leading to a new magnetic
cell four times larger than the crystallographic one, i.e. (2a, 2b, c) (see fig. 70). It is worth
noting that this arrangement does not correspond to a new propagation vector such as
Q = (0, , 0) for example, but is actually the superposition of: (i) a triple-Q magnetic
structure with the same propagation vector as in zero field, and (ii) a ferromagnetic
component. The metamagnetic process of ErGa2 has been satisfactorily accounted for by
Note added in proof The intermediate magnetic structure of DyCoaSi 2 has been recently determined by
neutron diffraction: it consists of ferromagnetic (111) planes with the sequence + + + - along the additional
I
1
propagation vector Q2-- - (i,
i,1 i)
(Shageoka et al. 1994).

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

369

ErGa2
\

I
H < 6.8 kOe

Er
o Ga

1
6.8<

H <20

kOe

Fig. 70. Magnetic structures of ErGa 2 at 1.5 K in low field (left-hand side) and in the intermediate region
(fight-hand side) (after Doukour6 and Gignoux 1982).

a simple model involving a huge uniaxiat anisotropy and exchange interactions between
nearest and second-nearest neighbours in the plane perpendicular to c. The same situation
seems to occurs also in the hexagonal compound CeZn5 (TN = 3.8 K). This compound
crystallizes in another structure than ErGa2, but the lanthanide lattice is the same in both
compounds, namely simple hexagonal, and the propagation vector is identical as well
as the direction of the magnetic moments. Although magnetic measurements have been
performed on a polycrystal, the same two-step metamagnetic behaviour as in ErGa2 has
been observed, with surprisingly the same ratio between both critical fields (Gignoux et
al. 1987). The same feature has been also reported by Kirchmayr and Poldy (1979) for the
rhombohedral compound c~-ErA13 where the zero-field magnetic structure is reminiscent
of that in ErGa2 (Weber 1977). Finally, it can be noticed that, if there are several equivalent
directions for the magnetic moments, e.g. two-fold axes in the basal plane of tetragonal
or hexagonal systems, or high symmetry axes in cubic systems, the situation of the
simple antiferromagnets is more complicated, because several magnetic domains have
to be considered, and spin-flop rather than spin-flip processes may then occur.
3.2.7. Modulated metamagnetic systems
In the intermetallic compounds, the presence of conduction electrons allows the local
exchange interaction between themselves and the 4f shell to propagate from one site to the
other ones. This effective long-range RKKY-type indirect coupling is a common feature
in all the lanthanide compounds. It often leads to the existence of magnetic structures
more complex than ferromagnetic or simple antiferromagnetic. As a consequence, in the
presence of long-range and frustrated exchange interactions, and in the case of an easy-

370

D. GIGNOUXand D. SCHMITT

0.3

TbNi2Si 2

0. ~

O.2

"~,--;

en

).

HII [001]

k...,/ ',,.~.T = 9 K H /

I oJ j~..-",
"
J

i.

'

" " " = T=IO K

~'~.

4~,

O.lO 1 ~ : ~ - . .

0J2

" ...."'.

..... ....

"~'.; ~r- l l K

; ~k

""

"

"-..

"..k" -

.-/ ~.
o~;a"~
~.~ %,

0.12

"~o.

".

""",.~',,.. T=12 K '"'.

,'c-....

..,.,.,~_
%

"'.

-'a~..,,,
T=13 K "'".
: '='t~,,
0:10

--%
%.

o.o60

~o

~o

H (kOe)

Fig. 71. Field derivativesof the magnetization

curves of tetragonal TbNi2Si2 along the c axis
at varioustemperatures;the peak at 12 and 13 K
indicates the metamagneticcritical field in the
modulated phase (after Blanco et al. 1991b).

axis anisotropy, the magnetic moments may develop an incommensurate or long-period

commensurate amplitude-modulated magnetic structure instead of a spiral one, being
frozen along a single direction This is particularly true near the ordering temperature
TN if the Fourier transform of the exchange coupling, J(q), exhibits a maximum value
for a vector Q not in a high-symmetry position in the Brillouin zone, i.e. Q ~ 0 and
Q ~ ( a ) K , where K is a reciprocal lattice vector. In that case, the modulated structure
is mainly sinusoidal, the growing of higher-order harmonics in the periodic structure
occurring only at lower temperatures, on account of the squaring of the structure. The
metamagnetism associated with such modulated structures is simple when the magnetic
field is applied along the direction of the moments, exhibiting a single smooth transition
of class (2) (see sect. 3.21) However, it turns out that the anomaly on the magnetization
curves is often hardly visible due to the proximity of TN. It is worth noting that Arrott
plots (M 2 vs H/M) or plots of dM/dH vs H are generally a good way to amplify
these anomalies. On the other hand, the magnetization processes at low temperatures
are often much more spectacular, exhibiting much sharper anomalies (see the following
sections), so that the study of the modulated phase is generally neglected Thus, lanthanide
metals as Er and Tin, where such an amplitude modulated phase exists immediately
below TN and which have been considerably studied for more than 20 years, should
exhibit this kind of metamagnetic transition. Finally, in the case of modulated systems

MAGNETIC PROPERTIESOF INTERMETALLICCOMPOUNDS

. . . ~

8
"r"

'--m

371

HoAIGa//c

1.5K

. :"
.

6K

"" . . . . . . . "

2-

o
2O_
.....

0
0 TJ.,.~
0

..o.-"
....

" "i
20

40

60
H(kOe)

80

Fig. 72. Magnetizationprocesses in hexagonal HoA1Ga at different temperatures

along the c axis; note the modulated
metamagnetic behaviour at 20K (after
Gignoux et al. 1991b).

with a high crystallographic symmetry, e.g. a cubic symmetry with a modulation along
the four-fold axes, a difficulty arises concerning the magnetic domains, because it is not
obvious a priori whether the magnetic field favours the domain with moments parallel or
perpendicular to itself. In uniaxial-type modulated systems, this problem does not occur.
More recently, the increasing number of lanthanide intermetallic compounds showing
an incommensurate modulated structure, at least in a limited temperature range below TN,
allowed these transitions to be observed. In particular, numerous tetragonal compounds
having the ThCr2 Si2-type crystallographic structure order within an amplitude-modulated
structure with their magnetic moments pointing along the c direction. This is the
case for TbNi2Si2 between T t = 9 K and TN=15K, where the propagation vector is
Q = (0.426, 0.574, 0), while, below Tt, the structure changes towards simple commensurate
antiferromagnetie with Q =(0.5, 0.5, 0) (Blanco et at. 1991b). Depending on temperature,
several different incommensurate phases are induced by a magnetic field applied along the
c axis, but in the modulated phase, more precisely between 12 and 15 K, there is one single
metamagnetic transition around 10-15 kOe, as is clearly seen in fig. 71. The isomorphous
tetragonal compound PrCo2Si2 exhibits the same behaviour between T t = 1 7 K and
T N = 3 0 K (Sugiyama et al. 1992). In the simple hexagonal compound HoA1Ga, the
magnetic structure is antiphase at low temperature, with the basic propagation vector
Q=(, , ) together with its third harmonics 3Q=(0,0, ), but, between Tt=18.5K
and TN=31 K, an additional modulation of the moments appears along the c axis,
with a periodicity of about 53 unit cells (Gignoux et al. 1991b). In this latter phase,
a single metamagnetic transition occurs again, as shown in fig. 72. It is worth noting
that the corresponding anomaly in the magnetization curves is smoothed by thermal
effects. In the above examples as well as in the vast majority of the modulated systems,
the high-temperature structure is not stable down to 0 K, because of entropy effects,

372

D. GIGNOUX and D. SCHMITT

2

"-"

"

PrNi2Si2

..~

[1001

MAGNETIC

6
FIELD

( T )

Fig. 73. Magnetization processes in tetragonal PrNizSi 2 at 1.5 K along and perpendicular to the c easy direction;
note the cusp when the induced ferromagnetic state is reached along the c axis. Inset: corresponding phase
diagram. Lines are calculated by using the periodic field model (after Blanco et al. 1992a).

11
II
\~

\
I

II

"-.,- J

/t

II
t~
~,\

~../

tl
xj'

\~

ii

,t

,,//
-

~
x

H=4 T

X x

H=0 T

H=5 T

\j

Fig. 74. Schematic representation of

the magnetic structure of PrNi2Si 2 in
various magnetic fields at 1.5 K; note
the residual modulation at 5.8T, i.e.
immediately below the cusp where it
vanishes (after Blaneo et al. 1992a).

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

373

and the system transforms at an intermediate temperature into a more stable state,
namely a magnetic structure with a shorter period. One exception may occur when the
CEF ground state is a nonmagnetic state, as in tetragonal PrNi2Si2. In this compound,
a sine-wave modulated structure occurs at TN =20K, the propagation vector being
Q= (0, 0, 0.870) and the moments aligned along the c axis (Blanco et al. 1992a). This
structure persists down to OK without changing its periodicity owing to the existence of
a nonmagnetic CEF singlet as ground state well isolated from the excited levels. In these
conditions it is possible to study the modulated phase down to low temperatures, and in
particular the metamagnetic behaviour. In fig. 73, this latter manifests itself very clearly
as a positive curvature in the magnetization curve along the c axis up to about He ~ 6 T
where a cusp is observed. As depicted in fig. 74, this curvature corresponds to the
progressive vanishing of the modulation in the applied field, and the observed cusp
indicates precisely the completion of this evolution, i.e. the entrance into the induced
ferromagnetic state. It should be noted that above He, the large residual superimposed
susceptibility is reminiscent of the nonmagnetic character of the CEF ground state. The
temperature dependence of the critical field leads to a very simple phase diagram having
a single line of separation between the modulated phase and the induced ferromagnetic
state (see inset of fig. 73). This diagram has been perfectly accounted for by using the
periodic field model (see sect. 3.3), and by considering only the CEF parameters and the
exchange coupling coefficients J(0) and J(Q).

3.2.8. Spin-flip metamagnetism: long-period commensurate systems

As noted above, in the presence of an uniaxial anisotropy, incommensurate magnetic
structures are generally not stable at low temperature, and evolve or transform toward
a structure with equal moments. Moreover, their periodicity often becomes locked onto
the lattice, resulting in long-period commensurate (antiphase) structures (Gignoux and
Schmitt 1993). This situation may lead to a particular spin-flip metamagnetism, namely a
field-induced multistep behaviour. During this process, the magnetic moments antiparallel
to the field are successively flipped, giving rise to successive steps in the magnetization
curve, with intermediate plateaus which may remain relatively flat, at least at low
temperature. The corresponding values of the magnetization are then directly related to
the magnetic periodicity and to the number of moments which have been flipped. In the
same way, the amplitude of the steps depends on this periodicity and must be a multiple
of (2/N)Mo, where N is the number of moments per magnetic unit cell and Mo the value
of each moment. Consequently, the steps are as much smaller as the magnetic periodicity
is longer. Note that in the limit of a true incommensurate structure, the magnetization
should then increase continuously. However, the various intermediate configurations are
not all possible, due to free energy considerations, and the system may even change
its propagation vector during the magnetization process to minimize this free energy.
The simplest case corresponds to systems which keep the same propagation vector and
where the magnetic periodicity is short. Examples are provided within the orthorhombic
RCu2 series, through TbCu2 (TN = 55 K) and DyCu2 (TN = 31 K) (Lebech et al. 1987). At

374

D. GIGNOUX and D. SCHMITT

(a)

(b)

F
(c)

(d)

Fig. 75. Various moment configurations projected on the ab plane when the magnetic field is applied along the
a axis; the proposed sequences are (a)-(d) for TbC and (a)-(b)-(d) for DyCu 2.

low temperature, both compounds exhibit the same antiferromagnetic ordering, namely
an antiphase structure with a basic propagation vector Q = (, 0, 0), the moments being
parallel to the a axis. As there are four lanthanide atoms in the unit cell, 12 magnetic
moments, decomposed into four sublattices, have been considered to describe the possible
structures in the presence of a magnetic field applied along the easy direction (Iwata
et al. 1989). By using a self-consistent mean-field model, four possible configurations
have been found (see fig. 75), their free energy varying with the field in a specific
way which depends on the values of the exchange coefficients. The first configuration
(fig. 75a) is the fully compensated antiphase structure, with a net magnetization equal to
zero, observed in TbCu2 and DyCu2 in zero field. At each subsequent step (figs. 75b-d),
2 moments antiparallel to the field have been flipped, resulting in successive magnetization
1
2Mo, then Mo. Experimentally, it has been shown that the
values equal to gMo,
magnetization process along the a easy axis of TbCuz exhibits a single step at 19kOe
(fig. 76), while a two-step metamagnetic process is observed in DyCu2 with critical
fields of 16 and 19 kOe and an intermediate magnetization of about of the saturated
value (fig. 77). Therefore, the successive sequences of configurations (a) to (d) and
(a) to (b) to (d) have been proposed to occur in TbCu2 and DyCu2, respectively, and
appropriate exchange coefficients allow us to account for the observed single- or two-step
behaviours (Iwata et al. 1988, Iwata et al. 1989). These proposed configurations need to
be checked by means of neutron diffraction on a single crystal and in a magnetic field.
The above cases of TbCu2 and DyCuz are particularly pedagogical, because of the
short magnetic periodicity (3 crystallographic unit cells) which leads to large jumps
in the magnetization. Other examples where a similar behaviour has been proved
are not common, because it requires neutron diffraction experiments in a magnetic

MAGNETIC PROPERTIES OF 1NTERMETALLIC COMPOUNDS

lo
8

TbCu2

I
I oOO o

I
o

I
o

~
10

ao

~
6
[.-,

io

<
<

4.2 K j .I "
1

I
J

~: 2

Z
<

DyCu2 : .."-'-------'-~a
12

375

2
o

4.2K

20
10

20

30

40

50

40

60

H (kOe)

H (kOe)
Fig. 76. Magnetization process along the three main
symmetry directions at 4.2 K in orthorhombic TbCu2;
lines are calculated (after lwata et al. 1987).

Fig. 77. Magnetization process along the three main

symmetry directions at 4.2 K in orthorhombic DyCu2;
lines are calculated (after Iwata et al. 1989).

field. Nevertheless, favorable cases should be compounds with a strong anisotropy and
which do not exhibit any change of propagation vector within their ordered phase
at zero field, meaning that their magnetic periodicity is very stable. The tetragonal
compound CeA12Ga2 (TN = 8.5 K) seems to be in this situation (Gignoux et al. 1988).
It orders with the propagation vector Q=(0,0, 6), with moments aligned along the
[110] direction, resulting at low temperature in a non-compensated antiphase structure
with a net ferromagnetic component ~3Mo ~ 0. I#B (see fig. 78). Although a noticeable
superimposed susceptibility is present on the magnetization curves, a clear plateau can
be observed around 25 kOe in the basal plane, followed by a sharp step at a critical
field He ~ 38 kOe. The values of the magnetization on both sides of this metamagnetic
transition, i.e. roughly 0.5 and 0.9#B, suggest that two moments have been flipped between
0 and 25 kOe, and two others at He. Another example where the magnetic periodicity does
not change within the ordered phase is provided by the hexagonal compound DyA1Ga
(TN = 50 K; Gignoux et al. 1993). Moreover, the propagation vector is relatively simple,
namely Q=(1, , 1), leading to the sequence ( + + + - - - ) , or (33), for the moments
along Q. A few steps with large amplitude can then be expected in the magnetization
process along the c easy axis, and are actually observed (fig. 79). It turns out that
two well-defined plateaus are observed at low temperature, the first one corresponding
to ~ of the saturated magnetization Mo = 9.67/z~, while the second one reaches Mo.
These metamagnetic transitions lead to a phase diagram which exhibits a quite original
feature (fig. 80). Although no neutron diffraction data are available for this compound,
the comparison with the isomorphous compound HoAIGa (Ball et al. 1992b) allows us
to assume that, in the first intermediate phase, region II, the magnetic cell is tripled
along c, and that, among the 18 corresponding moments, two antiparallel moments have
been flipped and one parallel moment has been "anti"-flipped. Apart from this region
where the propagation vector has changed, all the remaining regions in fig. 80 were
supposed to have the same fundamental propagation vector Q plus several harmonics nQ,
depending on the location in the diagram: the region I' corresponds to an amplitude-

376

D. GIGNOUX and D. SCHMITT

1.2

ID

12 K

0~

~0.2
"-a

,~

+
j

1.5 K
j~4" ~ K

C A

"

0.B

~.....t"

1.5K
.:---,--:10 ....... 20:.... -i30

/.0

i 50.... ",60

70

Hezt ( k O e )
Fig. 78. Magnetization curves of tetragonal CeA12Ga2 for the three main symmetry directions (after Gignoux
et al. 1988).

modulated structure, the region I to a pure antiphase structure (Q, 3Q), the region I"
to a field-induced non-compensated antiphase structure (Q, 2Q, 6Q). For this latter case,
the sixth harmonics corresponds to the ferromagnetic component, and the sequence of
the moments is (45) along Q. The original characteristic of this diagram consists in the
presence of the bicritical endpoint B (see sect. 3.2.1), which arises because the phases

,\

MAGNETI~ PROPERTIES OF INTERMETALLIC COMPOUNDS

10
8

DyAIGa//c

6
ca 4
~ 2
0
2

377

80

DyAIGa//c

~" 1oc

.,e,

1.4K

60 '

,~TN=50K

40 'T T ~ ' ~ A B

" II

20~II
~

~
20

40

60

80

100

120

140

"~

I'

, I ,
10

20

l
30

,- I

'

40

I
50

60

Temperature(K)

H(kOe)
Fig. 80. Magnetic phase diagram of DyA1Ga for the
c direction. Note the absence of boundary between
Fig. 79. Magnetization process along the c axis in the antiphase region I, the modulated phase I' and the
hexagonal DyA1Ga; at 1.4 K, the first (second) plateau field-induced non-compensated antiphase region I't,
reaches ~ (3) of the saturated value (after Gignoux leading to the presence of the bicritical endpoint B
et al. 1993).
(after Gignoux et al. 1993).

I, I' and I'~ all have the same propagation vector: this critical point is reached when the
thermal modulation of the moments is large enough to prevent a sudden flipping of the
moments involved in the transition and to allow a continuous modification of their value.
A single metamagnetic transition then remains at high temperature, associated with the
vanishing of the modulated structure (see sect. 3.2.7). Therefore, the magnetic structure
may continuously change between the different moment configurations I, I ~ and I", by
moving around B without crossing any borderline.
Another example where the spin-flip character of the metamagnetic transition has been
well established by a neutron diffraction experiment is provided by the simple hexagonal
system PrGa2 (TN = 7.2 K), in a limited region of its phase diagram. At 1.5K and in zero
field, this compound exhibits a non-compensated long-period commensurate antiphase
structure, with the propagation vector Q = ( 4 , 4 , 0), the moments being aligned along the
[100] axis of the basal plane (Ball et al. 1993a). Although this latter plane is expected
to be nearly isotropic due to the hexagonal symmetry, the compound actually presents
a uniaxial character along the [100] axis, at least as far as the low-field magnetization
processes are concerned (H < 9kOe, see fig. 81). More precisely, it has been shown by
a detailed neutron diffraction experiment on a single crystal that the first field-induced
metamagnetic transition is of spin-flip type (Ball et al. 1994a). Below a critical field
He ,.~2.3 kOe (phase II), the magnetic structure includes 27 moments, the sequence of
which is (43)(321)(33)(43) along the [100] direction, i.e. it consists of 4 groups of seven
moments (4 up, 3 down), with two faults over 28 sites, namely one inversion up-down in
the second group and a missing up moment in the third group. The induced magnetic
structure above He (phase III) corresponds to the vanishing of the first fault of the zero

378

D. GIGNOUX and D. SCHMITT

I

'

'

PrGa 2
VI

=~

:~

_.~..v'~ ..~'';i

0.~

.".::::::J

.
~a
ee~

~,i

":*:

::/"

p~" ..it

,~"~
'T

I
II

~
III

Fig. 81. Low-fieldmagnetization process along

the [100] axis at 1.5 K in hexagonal PrGa2; the
first transition at ~2.3 kOe (increasing field) is
of spin-flip type, the other ones of spin-slip
type; the propagation vector is Q = (4, ~, 0) in
phases II and III, and Q=(-~, -~,0) in phases IV
and VI (after Ball et al. 1994a).

IV

6
8
H (kOe)

10

12

PrGa 2

+ -

T=I.8K

++++---

H = 4 kOe - Phase VI
Spin-slip

++++

+ + .

'

H = 3 kOe - Phase HI

H = 0 kOe - Phase II

+++@
First

default

+I

Spin-flip

++++---

++

++++

Second default

Fig. 82. Low-fieldmagnetic structures of PrGa2 at 1.8 K; note the successive spin-flip and spin-slip behaviours
(after Ball et al. 1994a).
field structure, i.e. to a single spin-flip within the second block, without changing the
magnetic periodicity (see fig. 82). Therefore, the resulting magnetization changes from
1 M o to 3 M o , where Mo = 3. I#B is the Pr magnetic moment, as observed experimentally.
This spin-flip transition is associated with the growing o f the 2Q and 4Q even harmonics
in the neutron diffraction patterns, accompanied by the decrease in intensity o f the
odd harmonics. It is worth noting that the other transitions are not pure spin-flip ones,
and this will be discussed in the next section.
3.2.9. Spin-slip metamagnetism: long-period commensurate systems
The spin-slip concept is now well established since the detailed original investigation of
the magnetic structures o f holmium and erbium by X-ray scattering studies (Gibbs et
al. 1985, Gibbs et al. 1986). In both hexagonal metals, the magnetic propagation vector
depends on the temperature through a succession o f lock-in transitions on commensurate

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

i

379

Ho

3
B (T}

0 '
I0

20

30

40

T{K)

Fig. 83. Phase diagram of hexagonal holmium metal at 2 K with a magnetic field applied along the c axis; note
the successive values of the propagation vectors illustrating the spin-slip behaviour (after Jehan et al. 1992).

values close to each other. The formation of these spin-slips is principally driven
by the competition between the exchange coupling, favouring the formation of an
incommensurate structure, and the crystal field interaction, which favours a commensurate
structure. More precisely, there is a tendency for the basal-plane component of the
magnetic moments to lock close to the easy axes within the hexagonal plane, resulting in
periodic defects (spin slips) instead of a regular rotation of the spiral order. Although
pure lanthanide metals are not the main purpose of this review, it is interesting and
important to present some details about this spin-slip phenomenon and the associated
H - T phase diagrams, because it is associated with a remarkable renaissance in the
study of the magnetic structures of the heavy lanthanides (Jensen and Mackintosh 1992).
Furthermore, this phenomenon is quite similar to the behaviour found in numerous
lanthanide intermetallics (see the following sections). In holmium for example, it has
been shown by neutron diffraction at low temperature that the application of a magnetic
field along the c hard axis produces a resurgence of the propagation vectors observed
at higher temperature (see fig. 83). The sequence of the successive propagation vectors
underlines the slight rearrangement of the magnetic structure occurring at each spin-slip
transition, i.e. the small step-like increase of its periodicity. A similar behaviour has been
observed in erbium (Lin et al. 1992). All these effects have been mainly deduced from
X-ray or neutron diffraction, because of the sensitivity of these experiments with regard
to the magnetic periodicity. In the past few years these spin-slip transitions have been
observed in other types of measurements such as heat capacity, ac susceptibility and
electrical resistivity of erbium (see Pecharsky et al. 1993 and references therein).

380

D. GIGNOUX and D. SCHMITT

A recent example where spin-slip behaviour has been clearly seen is provided by
the intermetallic hexagonal compound PrGa2 in its low-field magnetization process. The
difference with the above case of pure lanthanide metals arises in the tmiaxial character of
PrGa2 in low field while the spiral part of the structure was concerned in Ho and Er. Apart
from the first field-induced transition between phases II and III, attributed to a spin-flip
mechanism (see previous section), it has been demonstrated by neutron diffraction that
the other transitions occurring below 9 kOe are of spin-slip type (Ball et al. 1994a). More
precisely, the well-defined plateau observed by decreasing field between 6 and 2 kOe (see
fig. 81, Phase VI) has been shown to correspond to a magnetic periodicity of exactly
seven unit cells, i.e. Q = ( 4 , 4 , 0 ) , where the second fault in the sequence of magnetic
blocks has disappeared (see fig. 82). Moreover, this periodicity appears by increasing
field at around 4 kOe and subsists up to 8 kOe (Phase IV), the width of the associated
neutron diffraction peaks varying in the same field range and being related to the large
superimposed susceptibility observed on the magnetization curve. This behaviour has
been explained by the difficulty for the system to establish a perfect long-range periodicity
of seven cells, possibly due to defects in the crystal.

3.2.10. Complex multistep metamagnetism: long-period commensurate systems

This section extends the simple metamagnetic behaviours described in the previous
sections for compounds having an uniaxial symmetry to more complex systems: these
also have a strong easy-axis anisotropy, but exhibit several different propagation vectors
as a function of temperature. As a consequence, the magnetization processes at low
temperature along the easy axis generally show a multistep metamagnetic behaviour,
mixing spin-flip, spin-slip mechanisms and/or more drastic changes of the magnetic
structure, resulting in more or less complex H - T phase diagrams. An increasing number
of lanthanide intermetallic compounds has been shown to belong to this category, but
only a few of them have been thoroughly investigated by means of detailed magnetic
and neutron diffraction measurements on single crystals. A wealthy family is the bodycentered tetragonal ThCr2Si2-type system where a strong magnetocrystalline anisotropy
often confines the magnetic moments along the c high-symmetry axis. Among this
family, PrCo2Si2 has been particularly thoroughly investigated. This compound orders
at TN = 30 K within a long-period commensurate structure with Q1 = (0, 0, 7) (Shigeoka
et al. 1987). As the temperature is lowered, a first-order transition occurs at T2 = 17 K, the
new propagation vector being Q 2 -(0,
0, ~),
13 while the periodicity locks below T1 = 9 K
into a simple antiferromagnetic structure with Q3 =(0, 0, 1). High-field magnetization
measurements performed at 1.3 K have shown a four-step metamagnetic behaviour along
the c axis, the value of the net magnetization at each step being a fractional multiple,
namely 14'
~ 14'
3 29 and 1, of the saturated value of Mo = 3.2#B/Pr (see fig. 84; Fujii
and Shigeoka 1990). These intermediate magnetization values are consistent with fieldinduced magnetic structures corresponding to the same successive propagation vectors
Q2 and Q1 as those stabilized at higher temperature. This has been confirmed by a neutron
diffraction study in pulsed high magnetic field (Motokawa et al. 1992). The temperature

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

38i

Hc~,

PrCo2Si 2
T=1.3 K

H a II c - a x

is

5
"-2
rn
v

Hc 2 Hc~
I
I

.
Ha.Lc -axis

O0

2'0

10

3'0

Ha (T)

Fig. 84. High-field magnetization curves

along the c and a axes at 1.3 K in tetragonal
PrCozSi2; the dashed line is calculated by
using a simple incommensurate mean field
model (after Fujii and Shigeoka 1990).

variation of the critical fields leads to a characteristic phase diagram where the hightemperature phases extend toward the low-temperature and high-field region by turning
around the origin (fig. 85). A simplified incommensurate exchange field model was first
proposed by Date (1988) to explain this behaviour by successive spin-slip and spin-flip
transitions. A more elaborate model was then developed by Iwata (1990), within the mean
field approximation, and taking into account a realistic shape for the Fourier transform
J(q) of the exchange interactions along the whole c* direction. This function is maximum
for q = Q1 with a weak variation between Ql and Q3. The free energy of several moment
configurations has then been calculated in a self-consistent way and the experimental
features of the multistep metamagnetic process could be well reproduced, in particular the
small third step at 89 kOe interpreted as a transition between Qz = 7 and Qz = 3 It is worth
noting that all the values for the propagation vectors and the net magnetizations differ only
PrCo2Si2

15

A
A

10

\
\
\
\
\

Be
5

Be1"

.
J

TT

10

T2

T(K)

TN

20

30

Fig. 85. Magnetic phase diagram ofPrCo2Si 2

along the c axis determined by high-field
neutron diffraction studies; the open circles,
solid circles, open triangles, and crosses
show the successive phases with Qz = 1,
0.926, 0.777 and 0, respectively (after Nojiri
et al. 1991).

382

D. GIGNOUX and D. SCHMITT

TbNi2Si 2
T=I.5K

:3_
:E

H.[O01]

..-"

H,EllO!
,.

H~' 'Hf

....

."

".

HIl[lO0]

. . , . . . , . ~

2'0

'

40

6'0 H(kOe) 80

Fig. 86. Magnetizationprocesses at 1.5K in tetragonalTbNi2Si2alongthe 3 main symmetrydirections(after

Blanco et al. 1991b).
slightly between both models, and correspond to very slight differences in the free energy.
However, the main features of the observed behaviour are well understood. A similar
situation has been seen in the neighbouring compound NdCo2Si2 where the sequence of
propagation vectors is Q = (0, 0, 1) for T < 15 K, Q = (0, 0, 0.928) between 15 and 24 K
and Q=(0,0,0.785) between 24K and TN=32K (Shigeoka et al. 1988). Low-field
magnetization measurements at 4.2 K have revealed the first two metamagnetic transitions
which can be presumably associated with the resurgence of the two high-temperature
propagation vectors, the last transition being expected at a much higher magnetic field.
The isomorphous compound NdCo2Ge2 exhibits only two magnetic phases in zero
field, i.e. Q=(0,0,1) below 10K and Q=(0,0,0.75) between 10K and TN=28.5K
(Vinoharova et al. !991). Magnetization measurements performed on a polycrystal show
at least 3 sharp metamagnetic steps at low temperature, suggesting also the existence of
a complex phase diagram.
A second analogous compound has been well studied, in particular by neutron
diffraction under field on a single crystal, namely TbNi2Si2. This compound belongs to
the same family as PrCo2Si2, and orders at TN = 15 K within a modulated structure with
Q1 = ( l - T , ~+z,
1
0) and 17= 0.074, while a first-order transition occurs at Tt = 9 K toward
(Blaneo et al. 1991b). At low
a simple antiferromagnetic structure with Q2 = (, ~,0)
1
temperature, a three-step metamagnetic process takes place along the easy c axis, with two
well defined plateaus (see fig. 86). The temperature variation of the metamagnetic critical
fields leads to the complex phase diagram shown in fig. 87. The various intermediate
phases have been investigated by neutron diffraction. Once again, both field-induced
phases at low temperature reveal a resurgence of an incommensurate or long-period
commensurate propagation vector close to Q1, n a m e l y Q3 - (21-- r, 1 + 1;, 0) and 17= 0.104,
the difference between phases III and IV arising from the relative magnitude of all
its associated high-order harmonics. Moreover, the high-temperature phase V exhibits

M A G N E T I C PROPERTIES OF INTERMETALLIC COMPOUNDS

o
-r-

383

TbNi2Si2
40

F ~ * ~

H II[00 I]
H/~

30

]~-

20

Ill

10

~o

~2

T(K)

Fig. 87. Magnetic phase diagram of TbNi2Si2 along the c axis; phase F is induced-ferromagnetic,phase I is
amplitude modulated, phase II is simple antiferromagnetic,phase III is long-periodcommensurate,phase IV
is complexamplitudemodulatedand zone V is a complexmagneticphase (after Blanco et al. 1991b).
a complex behaviour with the presence of several propagation vectors of the type
( 31 - r, ~1 + 3, 0), with values of r ranging from 0.025 to 0.104 according to the field and
the temperature. It is noted that an independent investigation of the same compound by
another group (Shigeoka et al. 1992a) led to a slightly more complex situation: there the
magnetization process at low temperature exhibited two additional intermediate steps,
and the neutron diffraction experiment gave another sequence for the propagation vectors
(fig. 88). The corresponding full magnetic phase diagram seems then more complex, with
at least 9 intermediate phases. Note however that the observed values of the propagation
vector always are on the same symmetry line of the Brillouin zone, with r ranging
between 0 and 0.125. This behaviour appears as a typical example of a spin-slip behaviour
(see sect. 3.2.9) or more generally a devil's staircase, and the slight differences between
both single crystals emphasize the coexistence of several similar magnetic phases within
the same reduced free energy range.
Apart from the above examples found in the tetragonal systems with Co and Ni,
other compounds of the same family present similar metamagnetic behaviours, however
they have been much less investigated. In the series with Ru, TbRu2Si2 exhibits a
single propagation vector Q=(0.23, 0, 0) below TN =56 K. Its magnetization process
at 4.2K shows a two-step metamagnetic behaviour with a strong hysteresis (fig. 89;
Shigeoka et al. 1992b). Neutron diffraction under field has revealed an increase of the
second harmonics (0.46, 0, 0) below the first critical field, and the appearance of the new
vector (0.23, 0.23, 0) in the intermediate region. A field-induced transition from a singleQ toward a double-Q structure seems then to occur in this compound, the new vector

384

D. GIGNOUX and D. SCHMITT

10

10

TbNi~

(o)

~6

T=I.3

--

TbRu2 Si2

T=4.2

~01~

R
~r
2
0

b)

( ~14.'
- T , ~ - ~ , ' 0)

l,J 0.I
0
0

= "

6
H ( Testa )

10

20

30
40
Ha ( k O e )

50

60

10

Fig. 88. (a) Magnetization process at 1.3 K along

the a and c axes in TbNi2Si2; (b) field dependence
of the propagation vector as determined by neutron
diffraction (after Fujii and Shigeoka 1990).

Fig. 89. Magnetization process at 4.2 K along the main

symmetry axes oftetragonal TbRu2Si 2 (after Shigeoka
et al. 1992b).

possibly resulting from the coupling between both equivalent basic vectors (0.23, 0, 0)
and (0, 0.23, 0). In the same series, NdRu2Si2 develops two different magnetic structures
below TN = 24 K: an amplitude-modulated structure with Qt = (0.13, 0.13, 0) above 10 K,
and a ferromagnetic structure below this temperature. The magnetization curves along
the easy c direction exhibit ferromagnetic behaviour below 10K, and a single-step or
two-step metamagnetic process with low critical fields above 10K (fig. 90; Shigeoka
et al. 1990). Due to the presence of the ferromagnetic phase at low temperature, which
is not completely surprising due to the proximity of Q1 and Q = 0, the corresponding
magnetic phase diagram should be quite original and rather different from that observed
in the previous examples. This is effectively the case for the isomorphous compound
NdRu2Ge2, although only polycrystalline materials have been investigated (Ivanov et
al. 1992b). In this compound, two successive magnetic phases are present in zero
field: below 10K the magnetic arrangement is ferromagnetic, while between 10K and
TN = 17 K the structure is amplitude modulated with two wavevectors Q1 = (0.12, 0.12, 0)
and Q2 = (0.19, 0.05, 0.125), the moments always being parallel to the c axis (Ivanov et al.
1992a). A metamagnetic process is observed only in the incommensurate phase, leading
to a magnetic phase diagram exhibiting a single antiferromagnetic region completely
surrounded by the spontaneous or induced ferromagnetic phase (fig. 91). Other tetragonal
lanthanide intermetallic series seem to be worth studying using single crystals, since
polycrystals already exhibit magnetization curves with a sharp metamagnetic behaviour.
For example, the compound TbRh2Si2 orders at TN = 94 K in a simple antiferromagnetic
structure, with Q=(0,0,1), while its magnetization process at 4.2K seems to show a
three-step metamagnetic behaviour (Szytuta et al. 1992). On account of the large number

MAGNETIC PROPERTIESOF INTERMETALLICCOMPOUNDS

385

NdRu2Si2

co>'T: ~+.+ ;
~.-o---+,--+~--e.--

+ ,,--,,e-0--

(b) T:'12 K'

o__..-0

<001>

<100>
.

v
~r

<110>

(el) T=29 K'

c)'T=l~ K '

20

4.0

T(K)T~"~

0
20
H~ ( kOo )

/+0

60

Fig. 90. Magnetization process at various

temperatures and along the main symmetry
axes in tetragonal NdRuzSi2 (after Shigeoka
et al. 1990).

P ____..o-----~

10

NdRu2Ge2
I

lo

H(kOe)

Fig. 91. Magnetic phase diagram of NdRu2Ge2

obtained on a polycrystal; AF refers to the modulated
phase, P and F to the para- and ferromagnetic
phases, respectively. Note that the borderline between
P and F has no physical meaning, both phases being
ferromagnetic (induced or spontaneous) (after Ivanov
et al. 1992b).

o f compounds having this tetragonal crystallographic structure and a c easy magnetization

direction, numerous new results can be expected in the future.
The other very simple crystallographic structure having a strong uniaxial character
is the hexagonal structure. However, few complex antiferromagnefic systems with this
symmetry and a c easy axis have been thoroughly investigated. One recent example is
provided by the simple A1B2-type compound HoA1Ga, which has been extensively studied
by magnetization and neutron diffraction measurements on a single crystal. As quoted
above (see sect. 3.2.7), this compound exhibits two different magnetic phases in zero
field (Gignoux et al. 1991b): (i) an antiphase one below Tt = 18.5K, characterized by
i 2)
1 and its third harmonics 3Ql =(0,0, ); (ii) an amplitude
the basic vector Q 1 - ( 3 I, 3,

386

D. GIGNOUX and D. SCHMITT

o
"r
:::L 6

oo

~tPI -~

\~

~ ~///

c axis

t(

T:I./Ga

/=~-~,

I
20

40

60

80

H(kOe)
60

. . . .

50

2Ql

. . . .

"15

"to

" 2Q2

. . . .

Fig. 92. Magnetization process at 1.7 K

in hexagonal HoA1Ga along the c axis
and associated magnetic structures as
determined by neutron diffraction in
fields of 0, 30 and 48 kOe; A, B and
C denote lines of moments [3n,3n, z],
[3n+l,3n+l,z] and [3n+2,3n+2,z],
respectively (n integer) (after Ball et al.
1992b).

. . . .

HoAIGaJ/c

' 2Q2

'1- 40
30

.m

20 ~ \

\
f-o'Q~'2Q, \
~,3)

~Tt :185K

,.

II

i TN:31K

10

lo

20

30 T(K) 40

Fig. 93. Magnetic H-T phase diagram of

HoA1Ga for the c direction; the propagation vectors observed by neutron diffraction are shown for each measured point
(solid squares); the hatched region indicates
the progressive vanishing of the lowtemperature periodicity associated with Q~
(see text) (after Ball et al. 1992b).

modulated structure between Tt and TN = 31 K with the single incommensurate vector

Q2 = (, , 0.481). It is worth noting that no anomaly could be detected by specific heat
or susceptibility experiments at Tt; only neutron diffraction allowed us to determine the
temperature of this transition. At low temperature, a three-step metamagnetic process is
observed along the c axis (see fig. 92), with a hysteresis decreasing in increasing field,
the two intermediate magnetizations reaching ~ i and of the full moment. From the
temperature dependence of the critical fields (see fig. 72, sect. 3.2.7), a magnetic phase
diagram has been constructed (see fig. 93). It shows three ordered regions, the magnetic
structures of which have been determined by neutron diffraction on a single crystal (Ball et
al. 1992b). The first intermediate phase (Phase III) is characterized by: (i) the vanishing of
the 3Q1 harmonics; (ii) the growing of the even harmonics 2Q1 and 6Q1 = Q0 = (0, 0, 0),
this latter corresponding to the ferromagnetic component; (iii) the appearance of new
propagation vectors, namely (0, 0, 1) and (, , ). The smallest common vector able to

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

387

generate the observed ones being (, , ~), a magnetic cell three times larger than the zerofield one and containing 18 moments has been considered. As shown in fig. 92, in the
corresponding structure, a single line B (or equivalently C) is involved: along this line, a
flipping of 32-of the initially antiparallel moments and an "antiflipping" of of the parallel
moments has occurred, in agreement with the value of the net magnetization. The second
field-induced state at 1.7 K is characterized by the flipping of all the remaining antiparallel
moments of line B, the magnetic unit cell recovering the same size as at zero field,
i.e. it includes 6 moments. However, this structure does not appear fully consistent with
the apparent absence of borderline detected by magnetic measurements within Phase II
(Gignoux et al. 1991b), since the other extremity of this phase is associated with an
incommensurate propagation vector. Effectively, such borderline has been determined by
neutron diffraction, although the transition is not as strong as expected (see the hatched
region in fig. 93). In fact, a broadening of the diffraction peaks at 1.7 K and 48 kOe is
already present and suggests that a perfect arrangement with wavevector Q1 and its even
harmonics is not fully achieved within this limited region, but presents faults reminiscent
of the high-temperature propagation vector and illustrating the competition between both
commensurate and incommensurate periodicities.
Another crystal symmetry showing complex multistep metamagnetism is the ort h o r h o m b i c symmetry. In the RCu2 series, HoCuz, ErCu2 (Hashimoto et at. 1979) and
more particularly NdCu2 exhibit a metamagnetic behaviour. In this latter compound, four
successive antiferromagnetic phases have been detected in zero field below TN = 6.3 K,
and magnetoresistance as well as magnetization measurements have revealed a quite
complex phase diagram along the b easy direction, showing numerous antiferromagnetic
and ferrimagnetic phases (see fig. 94; Svoboda et al. 1992). In the isomorphous
I

a MvsT
0 RvsH
D CysT
3

NdCu2
0

[-.

10

T [K]
Fig. 94. Proposed magnetic phase diagram of orthorhombic NdCu2 along the b axis, deduced from
magnetization,magnetoresistanceand specificheatmeasurements,A, antiferromagneticphases;F, ferrimagnetic
phases; P, paramagneticphases. (AfterSvobodaet al. 1992.)

388

D. GIGNOUX and D. SCHMITT

120

o H~

PrZn2
100

cD
_~

o H~
80

80

O
O

"I-~0
ZZ
20
0

8g

00

10

15

20

25

T(K)
Fig. 95. Magnetic phase diagram in orthorhombic PrZn 2 along the three main symmetry axes; a is the easy
axis, and c is the least favoured axis (after Kaneko et al. 1992b).

0,

o-oxi5

o ~ o ~ O -

O0

50

100
H (kOe)

150

Fig. 96. Magnetization processes of orthorhombic CeZn~ at low temperature along the three main symmetry

axes (afterYamashitaet al. 1987).

RZn2 series, several more or less sharp metamagnetic transitions have been observed
along the three different symmetry axes of NdZn2 (Kitai et al. 1990) and DyZn2 (Abe
et al. 1992). The same is true in PrZn2 for which the magnetic phase diagram has been
determined for the three main symmetry directions (fig. 95; Kaneko et al. 1992b). This

MAGNETIC PROPERTIES OF 1NTERMETALLIC COMPOUNDS

80

_ -*--~.~...~

CeZn z
b

60

389

axis

oa

.2

40

20

4
6
Temperature (K)

Fig. 97. Magnetic phase diagram of

CeZn2 along the b easy direction (after
Gignoux et al. 1992).

diagram suggests the existence of two zero-field magnetic phases, but no experimental
data is available at the present time. The simple antiferromagnet CeZn2 (TN = 7 K) appears
to be much more anisotropic than the other compounds of the same series, and exhibits
a three-step metamagnetic behaviour along the single b easy axis (see fig. 96). The
associated phase diagram is shown in fig. 97. Comparison with the substituted compound
Ce(Zn0.9Cu0.1)2, which exhibits a similar behaviour and which has been investigated
by neutron diffraction in magnetic fields using a single crystal (Gignoux et al. 1992),
suggests that both intermediate phases correspond to the successive flipping of blocks of
moments according to a magnetic periodicity first tripled, then doubled along the c axis.
These changes of magnetic cell are consistent with the values of the net magnetization in
both phases, i.e. g1 and of the saturation magnetization. Within the other orthorhombic
series RCu6, the NdCu6 compound appears as a complex metamagnetic system below
TN = 6.1 K (Takayanagi et al. 1991). In zero field, up to three incommensurate magnetic
phases have been found in a narrow temperature range below TN, the associated transition
temperatures being successively 5.83, 5.74 and 5.58 K, and all having a propagation vector
of the type (0, r, 1) (Mitsuda et al. 1991), while below 5.58 K the simple antiferromagnetic
structure with Q=(0, 0, 1) is stabilized. The magnetization process at 1.3 K along the
b easy axis shows four successive sharp metamagnetic transitions with noticeable
hysteresis (see fig. 98). It is noted that the hysteresis is anomalous for the last two
transitions, the critical fields being lower in increasing field than in decreasing field (see
sect. 3.2.1). Furthermore, the large superimposed susceptibility in the first plateau can
be resolved into two different phases with an additional intermediate step by lowering
the temperature down to 0.6 K. The low-field corresponding phase diagram has been
determined by high-resolution specific-heat measurements under field (see fig. 99). As
often observed, the incommensurate phases of the low-field and high-temperature region
seem to replace the simple antiferromagnetic one at high field and low temperature,
but this fascinating compound clearly needs a lot of additional investigations to be well
understood, in particular the determination of the structures of the numerous field-induced

390

D. GIGNOUXand D. SCHMITT
~

2.0

NdCue

1.3K

TI

HRb-axis

5.0~

4.0

E/

H//b

3.0

:Z

]E

1.0

JJ
~
,r

~ 1.0
a

//"

i[l
L

.~
2.0
m

/ j ~
...-/ " /. / "

t> j
0

4.5

5.5

Temperature [K]

6.5

Fig. 99. Magnetic phase diagram of NdCu6 along the b axis

deduced from specific heat measurements (after Takayanagi
et al. 1991).

50
100
I50
M'agne'tic Field ( kOe )

Fig. 98. Magnetizationcurves at 1.3 K in orthorhombicNdCu6 along the three main symmetrydirections. Note
the peculiar hysteresis associated with the last two transitions (after (Jnuki et al. 1986).
magnetic phases. As a conclusion of this long section devoted to the complex multistep
metamagnetic behaviour of compounds having a strong uniaxial-type anisotropy, one may
still expect a large number of new experimental results in the future, from magnetization
measurements as well as neutron diffraction experiments performed on single crystals,
allowing us to determine the microscopic aspects of the related complex phase diagrams.

3.2.11. Planar metamagnetic systems

The previous section has shown the variety of behaviours which can be observed in
uniaxial lanthanide systems having their magnetic moments confined along the highsymmetry direction in the presence of incommensurate magnetic structures. Much less
spectacular and a priori more complex are the same systems where the crystal field
anisotropy fixes the moments to remain perpendicular to this direction. Indeed, the onedimensional character of the possible structures has to be replaced by a two-dimensional
problem, which either can be isotropic or may reveal new features related to the inplane anisotropy. Again, although the pure lanthanide metals are not the main subject
of this review, it is worth making a few remarks about them, on account of the recent
remarkable renaissance in the study of their magnetic structures in various conditions
of field and temperature (Mackintosh and Jensen 1992). Most of the new features were
revealed in high-resolution studies with both X-rays and neutrons, performed using highquality single crystals. First, new types of magnetic structures have been identified during
the magnetization process of a helical structure with a magnetic field applied in the plane
of the helix, as in Ho metal (Jensen and Mackintosh 1990). These helifan structures do not

MAGNETIC PROPERTIES OF 1NTERMETALLICCOMPOUNDS

391

require specific anisotropic interactions, and occur between the distorted helix observed
in low field and the fan configuration present in high field. Secondly, as described
above (see sect. 3.2.9), spin-slip transitions have been observed as a function of field
or temperature, resulting from the in-plane anisotropy of Ho for example (Jensen and
Mackintosh 1992). Finally, the presence of several equivalent symmetry directions in
the plane may also favour multiple-Q magnetic structures as in Nd metal for example,
which exhibits a fascinating hierarchy of 1Q, 2Q, 3Q and 4Q structures below its
Nrel temperature TN = 19.9K, and where more than 10 phases have been revealed in
its complex phase diagram (McEwen and Zochowski 1990). Some of these phases
are characterized by the appearance of subharmonics of the fundamental modulation
wavevector and present some analogy with the helifan structure quoted above.
As a general rule, the existence of several equivalent easy directions within the
basal plane of planar intermetallic lanthanide systems allow the magnetic moment
configurations to become non-collinear during the magnetization processes. Such canted
magnetic structures noticeably multiply the number of possible magnetic phases. From
this point of view, the most favourable symmetry is the hexagonal one, showing 6 possible
directions for the moments. In this context, the RGa2 series appears particularly promising
to be studied, several compounds of this series exhibiting at least one incommensurate
magnetic phase together with an easy axis within the basal plane. For example, the
PrGa2 compound, which presents a low-field behaviour being of Ising-type along its
[100] easy direction at low temperature, as described above (see sects. 3.2.8, 3.2.9),
recovers a planar behaviour at higher fields and/or higher temperature, as shown in fig. 100
(Ball et al. 1993a). The associated metamagnetic process is characterized by a two-step
behaviour along the [100] direction, and a three-step one along the [120] hard direction
of the basal plane, in the field range 10-15kOe. Moreover, the intermediate phases
present a large superimposed susceptibility, a feature which seems to be quite general
in these planar processes. This feature can be definitely associated with the noncollinear
arrangement of the moments and their ability to rotate more or less continuously within
3.2
~ 2.4
=
~

1.5K

1.6

130

10

20

30 40 50
H (kOe)

613 70

80

Fig. 100. Magnetization curves of hexagonal PrGa~ at 1.5 K along the three main symmetry directions. Note the
large superimposed susceptibility of the high-field intermediate phases along the [100] and [120] axes (above
10kOe) (after Ball et al. 1993a).

392

D. GIGNOUXand D. SCHMITT
'

[120]

'

3
Ol

NdGa 2

/~.~

1.5 K
....

.......

" t

20

40
60
80
H (kOe)
Fig. 101. Magnetizationcurves of hexagonalNdGa2 at 1.5K alongthe three main symmetrydirections.Note
the largesuperimposedsusceptibilityof the intermediatephases alongthe [100] and [120] axes(afterBall et al.
1994b).
the basal plane. In addition, it can be seen that during such rotations, the moments
may be temporarily locked into intermediate easy directions. A preliminary neutrondiffraction experiment in this high-field phase at 1.5 K along the easy axis has shown
the resurgence of the incommensurate propagation vector Q = (0.148, 0.148, 0.023) which
characterizes the high-temperature phase of this compound in zero field (between 3.5 K
and TN = 7.2 K). The resulting H - T magnetic phase diagrams of PrGa2 along both
directions [100] and [120] of the basal plane then include 5 and 6 distinct phases,
respectively. In the neighbouring compound NdGa2 (TN = 9.5 K), the situation is quite
similar. This compound exhibits two different incommensurate magnetic structures in zero
field (Ball et al. 1992c), and a multistep metamagnetic behaviour along both directions
of the basal plane (see fig. 101). Again, the intermediate phases exhibit a superimposed
susceptibility which is still larger than in PrGa2, and the corresponding phase diagrams
include 4 (6) distinct regions along the [120] ([100]) directions (Ball et al. 1994b).
The compound CeGa2 as well as its derived solid solutions Ce(Gal-xAlx)2 exhibits also
planar metamagnetic properties, with the presence of ferromagnetic and incommensurate
magnetic structures in their H - T phase diagrams (Takahashi et al. 1988, Jerjini et al.
1988, Frrmy et al. 1989). However, the most complex system of the RGa2 series is
certainly DyGa2, where three successive magnetic phases have been observed in zero
field (Gignoux et al. 1991a). Below 6.1 K, the structure is simple antiferromagnetic,
the propagation vector being Q=(0, ,0). Between 6.1 K and' TN = 11 K, the structure
is incommensurate (amplitude modulated) with Q=(2r, - r, 0) and 3=0.134. Within
the modulated phase, there is an additional spin reorientation transition at 8.7 K, the
easy magnetization direction being the [120] and the [100] axis below and above
this temperature, respectively. At low temperature, the magnetization processes show
a particularly complex multistep metamagnetic behaviour, with numerous field-induced
transitions when the magnetic field is applied in the basal plane, suggesting a great
number of intermediate magnetic phases (see fig. 102). Some transitions are very sharp,
while others are more subtle. A hysteresis is associated with almost all transitions, their

MAGNETIC PROPERTIES OF INTERMETALLICCOMPOUNDS

393

10

~6

oi
0

7O

20

40

60

'

'

DyGa2 II b

60
"v 50

80

H(kOe)

'~ " 'VIII

Fig. 102. Magnetization curves of hexagonal DyGa2 at 1.5K along the three
main symmetry directions; b represents the
[120] easy direction and a the [100] hard
direction in the basal plane (after Gignoux
et al. 1991a).

4O
30
" t

IX",

2O
Fig. 103. Magnetic phase diagram of DyGa2

10

Oo

4 T(K) 6T1

8 T2 10 TN 12

along the [120] axis; the magnetic structure

is simple antiferromagnetic below T~ and
amplitude modulated between T Z and TN
(after Gignoux et al. 1991a).

amplitude decreasing with increasing field. Some intermediate phases are present only in
decreasing field. Although the general features of the curves are similar for both directions
in the basal plane, they slightly differ in detail, and lead to quite complicated magnetic
phase diagrams (see fig. 103 for the [120] axis). Very likely, all the observed fieldinduced transitions are associated with a mixing of the different types of metamagnetism
(spin-flip, spin-flop, change of periodicity, ... ), but only an extensive neutron study will
allow us to determine all the microscopic aspects of these phase diagrams. It is worth
noting that there are only a few examples of planar multistep systems available up to now
in other symmetries than the hexagonal one, this latter remaining the most favourable.
One orthorhombic compound however, namely Ce3Sn7 (TN=5.1 K), exhibits a two-step
metamagnetic process along the c easy axis (Fillion et al. 1992). The small discontinuous
field-induced transitions have been explained by the occurrence of various intermediate

394

D. GIGNOUXand D. SCHMITT

canted magnetic structures in the (a, c) plane, and these latter have been determined by
using polarized and nonpolarized neutron scattering.

3.2.12. Multiaxial metamagnetie systems

In the previous sections, a single axis or a single plane was involved in the metamagnetic
processes, the symmetry of the anisotropic lanthanide systems being of uniaxial type.
This last section is devoted to the systems where the whole space will be concerned,
i.e., where the equivalent symmetry directions for the moments do not all belong to the
same plane. This is the case of the cubic symmetry. The first example of metamagnetism
associated with this symmetry can be found in cubic NaCl-type lanthanide pnictides,
where some compounds exhibit, in an applied field, canted intermediate structures similar
to the "topside" antiferromagnetie structure evidenced in HoP (Fischer et al. 1985).
An early example is provided by DySb (Everett and Streit 1979), but similar behaviours
have been also reported in other pnictides, such as HoSb (Brunet al. 1980), HoAs and
HoBi (Hulliger 1978) and DyBi (Hulliger 1980). This latter compound undergoes a firstorder transition at TN = 11.2 K into a simple antiferromagnetic structure [Q = (, , )]
with magnetic moments aligned along the [001] direction. Under the application of
a magnetic field along the three main cubic symmetry directions, i.e. [100], [110]
and [111], there is in each case an abrupt metamagnetic transition, the corresponding
critical fields being approximately in the ratio 2:x/~:v~, and the resulting intermediate
magnetizations reaching respectively , l/x/2 and l/v/3 of the saturated value (see
fig. 104). A second and last step occurs at a slightly higher field along the four-fold axis,
while no further transition could be seen up to 200 kOe for the two other directions. All
the relative values of the critical fields and intermediate magnetizations are consistent
with the "topside" structure where the moments are oriented along two orthogonal
(100) directions, keeping the same propagation vector (bi-axial structure). This behaviour
requires first a strong anisotropy favouring the four-fold axes, and secondly strong
negative two-ion quadrupolar interactions in order to stabilize such a canted structure
(K_im and Levy 1982). A biquadratic coupling was also found to be important to
explain the similar two-step metamagnetic processes observed in the S-state isomorphous
compound EuSe (Fukuma et al. 1985).
A strong antiferroquadrupolar coupling has been also shown to be fundamental to
explain the spontaneous triple-Q magnetic structures observed at low temperature, as
well as the occurrence of well-defined field-induced multi-axis magnetic structures in
some other simple cubic CsCl-type systems, such as DyCu and DyAg (Morin et al.
1990). The compound DyCu orders at TN--62.8K within a simple antiferromagnetic
structure characterized by the propagation vector Q = (, , 0). A thorough study of its
magnetoelastic properties, in particular the absence of any spontaneous magnetostriction
at low temperature in spite of a large magnetoelastic coupling, led to the conclusion that
the cubic symmetry is preserved in the ordered phase by developing a triple-Q structure
with magnetic moments pointing along the four three-fold symmetry directions of the unit
cell (A16onard et al. 1984b). This configuration is favoured by the existence of a negative

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

395

M ( p B)
9

[lOO]

8
[110]

7
6

2"
[111]

..,

4
3
2

DyBi
l

,j,:"

15K

..

f~

20

40

60

J
H (kO)

80

Fig. 104 Magnetization curves in cubic

DyBi at 1.5 K (after Hulliger 1980)

4.2KK _ _ _ _ ~

z
O

12

7Lu 2

z
O
,<

50

100
I N T E R N A L FIELD (kOe)

150

200

Fig. 105 Magnetization curves in cubic DyCu along the [111] direction; the drawn magnetic structures
correspond to the configuration on both sides of the metamagnetic transition (after Al@onard et al. 1984b).

quadrupolar coupling for the appropriate trigonal symmetry, because of the corresponding
noncollinear mutual arrangement of the associated quadrupolar components. This multiaxis configuration then appears particularly rigid with regard to the field effects, leading
to the existence of very sharp steps defining one or several well-defined plateaus in the
magnetization processes along the three main cubic symmetry directions (see fig. 105
along the [111] axis for example). Several mechanisms occur during these metamagnetic

396

D. GIGNOUX and D. SCHMITT

I

120

ii I

~,.!

DyCu
H//r.o 01]

,..x ~ . / . /

~ d_.N.N

dT

80

v
"r-

\ \~o

40

0
30

40

50
TEMPERATURE (K)

~'

60

Fig. 106. Magnetic phase diagram in DyCu for a four-fold direction; in the representation of the structures,
the arrows correspond to the magnetic moments, the ellipsoids to the quadrupolar moments (after Morin et al.
1990).

10'
t111 ]

3o'
0

400

yj
0

'
100

'
200
H (kOe)

Fig. 107. High-field magnetization curves in cubic

DyAg at 4.2K along the three main symmetry
directions (after Morin et al. 1989).

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

397

transitions: some magnetic moments may flip preferentially along their own direction
in order to preserve the same quadrupolar energy (the quadrupolar moment is invariant
during such spin-flip transitions), or the triple-Q structure may be replaced by a doubleQ one, the moments being along (111) or (101) directions. Only at the very end of the
process, the magnetic moments leave a high-symmetry direction and rotate progressively
toward the applied field. This leads to the existence of phase diagrams which are
specific for each symmetry direction, as shown in fig. 106 for the [001] axis. The
case of the isomorphous compound DyAg is similar albeit more complex. Indeed, this
compound exhibits successively: (i) the same low-temperature structure as in DyCu
below T1 = 47.4 K; (ii) a double-Q structure between T1 and T2 = 49 K; (iii) an amplitudemodulated structure having the incommensurate propagation vector Q = ( - r , ,0),
with 7:~0.07 between T2 and TN=56K (Morin et al. 1989). As a consequence,
the magnetization processes along the three cubic symmetry directions present some
additional features (see fig. 107). In addition to the large metamagnetie transitions near
100 kOe analogous to those observed in DyCu, there exists a succession of several other
smaller steps in very high field, in particular along the [101] and [111] directions. From
the general behaviour described in the previous sections, it can then be assumed that these
high-field transitions are likely related to the resurgence of the incommensurate periodicity
existing in zero field close to TN. The corresponding phase diagrams then should exhibit
numerous field-induced magnetic phases. Multistep metamagnetic transitions are also
observed in several other isomorphous compounds presenting both antiferromagnetic
bilinear coupling and negative quadrupolar interactions favouring multi-axis (multi-Q)
structures, such as PrZn and NdZn (Morin and Pierre 1975), PrAg (Morin and Schmitt
1982), PrCd (Al~onard and Morin 1984), and ErAg and HoAg (Kaneko et al. 1992a).
Antiferroquadrupolar interactions are also present in other cubic systems, as in
the lanthanide hexaborides, where they drive the antiferroquadrupolar state occurring
in CeB6 at 3.3K (Effantin et al. 1985). In the neighbouring compound PrB6, an
incommensurate double-Q magnetic structure occurs through a first-order process at
TN =6.9 K, characterized by the propagation vector Q= (- r, , ) with r =0.05, while
it is replaced below Tt = 3.9 K by a commensurate double-Q structure through a lock-in
process of Q on the value (, , ) (Morin et al. 1991). Here the breaking of the multiQ structure by a magnetic field leads only to weak anomalies in the magnetization curves,
but they have been confirmed by neutron diffraction experiments. These transitions allow
us to construct phase diagrams which reveal how strong the quadrupolar interactions
are through the weakness of the field dependence of the commensurate-incommensurate
transition temperature. The last examples of cubic systems where multistep metamagnetic
transitions related to antiferroquadrupolar interactions have been seen are provided by the
AuCu3-type cubic compounds TmGa3 and Ndin3. In both compounds, the lanthanide
lattice is simple cubic as in DyAg and DyCu or in the lanthanide hexaboride. In
TmGa3, two first-order transitions occur at TN =4.26 K and TQ = 4.29 K, corresponding to
antiferromagnetic and antiferroquadrupolar orderings, respectively (Morin et al. 1987b).
In the low-temperature phase, the negative quadrupolar coupling associated with the
trigonal symmetry favours a multiaxial triple-Q spin arrangement, identical to the low-

398

D. GIGNOUX and D. SCHMITT

ooo o

ooooOO

,.

oAo

ojo

o~)

T= 1 , 5 4 , ' "
o

e."

"5K

J o"

oOO.

oo

a3

H#c1113 I

"5"

o,

"

:o"
:o

o
o

Fig. 108. Magnetization processes for a

~*oo o

I0

2'0

3'0
z,'O
H (kOe)

5~0

6'0

70

[111] magneticfield directionin cubic

TmGa3 (afterMorinet al. 1987b).

temperature structure of DyCu and DyAg (see above). It follows that the magnetization
processes are also similar, with for example a two-step metamagnetic behaviour along
the [111] direction (see fig. 108). The three magnetic phase diagrams corresponding
to the main symmetry directions are similar to each other, except for the values of
the critical fields (see fig. 109 for the three-fold axis). Owing to the range of these
critical fields, which are smaller than in the Dy compounds, neutron diffraction has been
performed on a single crystal allowing us to determine the intermediate phases, which
have a bi-axial or a quadri-axial character. The case of Ndln3 has been less studied but
it presents an analogous behaviour. This compound exhibits three successive transitions
at Tl = 4.65 K, T2 = 5.13 K and TN = 5.9 K, and a multistep metamagnetic behaviour has
been observed at low temperature along the three symmetry directions (Czopnik et al.
1985). The first two transitions at T] and T2 are first-order and are related to a change
of the magnetic structure. A recent neutron diffraction investigation on a single crystal
has proved that both high-temperature phases are incommensurate, with Q= (, , r)
and l:=0.017 and 0.037 below and above T2, respectively, while the low-temperature
phase is simple antiferromagnetic with Q -_(~,1 1, 0), the moments pointing along a fourfold axis at any temperature (Amara et al. 1994). Detailed magnetic measurements have
also revealed the existence of very complex magnetic phase diagrams along the three
symmetry axes, with numerous intermediate structures, some of them being triple-Q while
others are characterized by the resurgence of the incommensurate periodicity. Again in
this compound as in all the compounds described in this section, the competition between:
(i) a strong magnetocrystalline anisotropy with several equivalent directions in the
crystal, (ii) large antiferromagnetic interactions, and (iii) a noticeable antiferroquadrupolar
coupling, is very likely at the origin of the complex multistep metamagnetic behaviour
observed, through the stabilization of multi-axis magnetic structures.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

399

Hc2

15

TIO

;
TEMPERATURE (K)

Fig. 109. Magnetic phase diagram of TmGa s for a magnetic field along a three-fold direction (atter Morin et al.
1987b).

3.3. Quantitative analysis of incommensurate magnetic systems

During the last decade the theoretical approach of the complex magnetic phase diagram,
resulting from the competition of magnetic interactions in systems with strong magnetocrystalline anisotropy, has attracted much attention. In particular theoretical models
focused on the transition between commensurate (C) and incommensurate (I) phases, and
on the properties of the incommensurate phase.
A model which has been widely used is the axial next-nearest neighbour Ising (ANNNI)
model (Bak 1982 and references therein, Bak and von Boehm 1980, Selke and Duxbury
1984). It considers a simple spin- Ising variable, S = 4-1. The magnetic arrangements
are described as a stacking of ferromagnetic planes (this means one propagation vector
perpendicular to the planes) in which a positive interaction J0 between nearest neighbours
is considered. Along the propagation vector an interaction Jl > 0 and J2 < 0 between
nearest and next-nearest neighbours is assumed. On the basis of this model, phase
diagrams have been determined by using different methods, in particular Monte Carlo
techniques and various mean-field theories. The most prominent results are illustrated in
figs. 110 and 111 which show the ([Jz/Jl[, T) phase diagram constructed on the basis of a
self-consistent mean-field calculation, and the wavevector versus temperature calculated
for J2/J1 =-0.6, respectively. In fig. 110 dark areas indicate high-order commensurate
phases with incommensurate phases in-between. Note that, at the ordering temperature,
the wavevector changes continuously from q = 0 at the Lifshitz point to q--~z (the
+ + - - phase) for -J2/Jl --* oc. At T = 0 on the other hand the system can have only two
states, i.e. q = 0 f o r - J 2 / J l < 0.5 and q = ~ for-J2/J1 > 0.5. (In fact the existence of only
two possibilities at low temperature arises because of the severe limitation concerning

400

D. GIGNOUX and D. SCHMITT

i

'

-I

PN

FN

It,

0.2

~I

0.4

0.6

-J2/Ja
Fig. 110. Mean-field phase diagram for the 3D Ising model with competing interactions (Bak 1982). The dark
areas indicate high-order C phases with I phases in-between (incomplete devil's staircase). P is the Lifshitz
point.
I

tit,

0.25

219 311t. 115

0.20

3116 - -

2111

2111"

- 3117

I16
0.15

Tc

t
t,

Fig. 111. Propagation vector versus

temperature as calculatedby Bak (1982)
for J2/Jl = -0.6.

the signs o f J l and J2 and because longer-range interactions are neglected.) As shown in
fig. 111, the periodicity thus changes as the temperature is lowered, leading to the so called
"devil's staircase" behaviour. It has also been shown that, in addition to C and I phases,
there are chaotic structures which are at least metastable (Bak 1982 and references
therein). These theoretical studies allowed us to discover new quite exciting physical
concepts and to qualitatively account for experimental results, but, because of their great
simplifications, they fail to explain the experimental observations in any exact detail.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

401

T=4.2K

,.,.2

~2
::k

2._

o
o

lOO
H

200
(kOe)

300

Fig. 112. Comparison between experimental

(solid line) and calculated (dashed line) metamagnetic process for PrCo2Si2 at 4.2K (Iwata
1990).

In order to explain the complex phase diagram of CeSb and CeBi (Rossat-Mignod et al.
1983, Chattopadhyay et al. 1986), Date (1988) considered an Ising spin chain immersed in
a sinusoidally modulated exchange field. This model is less restrictive concerning the sign
and range of exchange interactions, however it is not self-consistent and, surprisingly, the
q dependence of the Fourier transform of the exchange interactions J(q) is not considered,
but instead an additional energy term of the type V(q-qe)2 is introduced. In spite of these
simplifications, the phase diagram of some compounds such as CeSb, CeBi, PrCo2Si2
(Shigeoka et al. 1987, 1989) and UPdln (Sugiura et al. 1990) were sketched by this
model.
Much more realistic is the approach of Iwata (1990). Within the mean-field approximation Iwata considers the q dependence of d(q) and solves the problem selfconsistently in the same way as will be described in more detail in a more recent model
presented below. The only limitation is that the CEF ground state is the Jz =1~/) doublet,
the higher-energy CEF levels being ignored, leading then to an Ising-like system. The
model then can be applied in systems with large uniaxial anisotropy imposing only one
direction for magnetic moments. These conditions are fulfilled in PrCo2Si2, where a
remarkable quantitative account of the H-T phase diagram, in particular the multistep
low-temperature metamagnetic process (figs. 112, 113) and the temperature dependence
of the propagation vector in zero applied field (see sect. 3.2.10). The q dependence of
J(q) along the [001] direction has been determined to account for these properties. Later
the complicated phase diagram of CeBi has been satisfactorily interpreted within the
framework of the same model (Iwata et al. 1991).
Recently, a more realistic and more general model has been proposed, and has been
successfully applied to describe the magnetic properties (e.g. magnetization processes,
magnetic susceptibility, phase diagram, specific heat . . . . ) of several frustrated magnetic
systems with any periodicity (commensurate as well as incommensurate) and any moment
direction (collinear as well as noncollinear structures) (Blanco et al. 1991a, 1992a,

402

D. GIGNOUX and D. SCHMITT

(cl) H<lOkOe, Q=l,rn=0.

LI! i
(b) lOkOe<~H<37kOe, Q=13/14, m=l/14.

T TTTIITTTT]
t[L
LLLLLLL

TI]I
iLL

(c) 37 kOe<_H< 89kOe, Q=7/9, m=2/9.

I tl ] 1] 111]T ._.
I t [ L LL I
(d) 89kOe<H<123kOe, Q=3/4, rn=l/Z.,.

T TT TT

I t t
Fig. 113. PrCo2Si2: predicted arrangements of the
Pr moments in various magnetic fields (Iwata 1990).
Notations Q and m represent the propagation vector
and reduced magnetization, respectively.

(e) 123kOe_<H, Q=0, m=l.

U!

Ball et al. 1993a). This self-consistent periodic field (PF) model is based on an N-site
Hamiltonian, N being the number of magnetic ions over one period of the magnetic
structure. This Hamiltonian can be written as
N

7-g = ~
i=1

~(i)
' ~'CEV+ ~
i=I

N
~-/g) + ~
i=1

1
~'~(B)+ 2 ~

(M(i)) H(i)x.

(26)

i=1

In this expression, the first term is the CEF coupling from which the anisotropy originates.
This term has generally two effects: (i) it may provide either a preferential (easy) direction
for the magnetic moments, leading to the occurrence of an amplitude-modulated structure
at TN, or a weak or easy-plane anisotropy, leading to a helical structure, or even a
mixing of both; (ii) by mixing the 4f wave functions, the CEF term, by raising the
(2J + 1)-degeneracy of the ground-state multiplet, then may produce either a magnetic or
a nonmagnetic level as the ground state (in the latter case a modulated magnetic structure
can be stable down to 0 K as observed in PrNi2Si2, see sect. 3.2.7) (Blanco et al. 1992b).
The second term, 7-/z, is the Zeeman coupling -H.M(i) between the 4f magnetic moment M(i)=- gj/tBar(i) at site i and the internal magnetic field H . The third term, ~B,
is the isotropic bilinear interaction -Hex(i)'M(i) written in the mean-field approximation
as a function of the effective exchange field acting on the ith site:
Hex(i) = (gj~tB)-2 ~
j=xi

J(ij) (M(j)) = ~ Hn~g~einQg'.

n

(27)

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

403

The second part of this equation arises from the periodicity of magnetic structure which
leads to a Fourier expansion of the magnetic moments:

(M(j)) = Z MnQ einQ/~j

(28)

which defines the basic propagation vector Q of the magnetic structure. Therefore it
follows that the exchange field is also periodic with the same periodicity as M, the Fourier
harmonics I-InO.being related in a simple way to the corresponding harmonics nnQ of
the magnetic moments through the Fourier transform J(q) of the interionic exchange
interaction J (ij):

HnQ =

(gJ/2B) -2

J(nQ) MnQ.

(29)

This equation shows that the periodic shape of the exchange field does not necessarily
follow that of the magnetic moment, depending on the relative values of the successive
couplings J(nQ). The last term in eq. (26) is a correction term necessary due to the meanfield treatment. From the self-consistent diagonalization of the Hamiltonian for the N ions
over one magnetic period, one can calculate: (i) the magnetic moment of each ion at any
temperature and in any internal magnetic field; (ii) the specific heat at any temperature.
The parameters which enter in the Hamiltonian are the CEF parameters and the
J(nQ)'s. The former, as well as J(0), can be determined in the paramagnetic state from
a joint analysis of a set of experimental results such as inelastic neutron scattering on
powder, thermal dependences of paramagnetic susceptibilities and field dependences of
magnetization measured on a single crystal, and specific heat. Moreover, J(Q) at TN is
given by
(gJIZB) 2

J(Q)- Zo(TN~'

(30)

where Z0 is the susceptibility without interaction which can be calculated as soon as

CEF parameters are known. The other J(nQ) coefficients are adjustable parameters (their
determination would require the analysis of the magnetic excitations measured by inelastic
neutron scattering on single crystal in order to obtain J(q) over the whole Brillouin zone).
In fact the role of J(Q) is preponderant (in particular higher-order terms have no effect
in sine-wave modulated and helical states) and it has been shown that the knowledge of
CEF parameters and of J(0) and J(Q) was generally enough to account quite satisfactorily
for the magnetic properties below TN.
From parameters determined in the paramagnetic state, the PF model has been
successfully applied in the analysis of the magnetic properties below TN of the tetragonal
PrNi2Si2 (Blanco et al. 1992a,b) and TbNi2Si2 (Blanco et al. 1992c), and the hexagonal
HoA1Ga (Ball et al. 1993a, Zhang 1992) and PrGa2 (Ball et al. 1993c, Ball 1993)
compounds, as illustrated in figs. 73 (sect. 3.2.7), 114 and 115. It is worth noting that

404

D. GIGNOUX and D. SCHMITT

6O

........

~" 50
O

II

,.

,,~ 40

HoA1Ga
[001]

-'X

iIi I
..i'

30

"\

20

'~\ T=18.5K

r
10

3Q,'Q,= I

10

1 6

14

c~

....

"\t; Q2=
',,,
20

z=u2

T(K)

~fz=7/47 ,' ''~101

t'c=1/44, / A',~" 8[

40

~ 1 :

1-

....

PrGa
5

~7

./,
0

30

TN=31K

"'i ~% ................. .5

1o "

Fig. 114. Boundaries of the phase

diagram of the hexagonal HoA1Ga
compound along the easy [001] axis
(Ball et al. 1993b, Zhang 1992). Dots
are experimental points. Solid lines are
calculations with the PF model Boundaries of phase III are not accounted
for by the model because this phase
is described with several noncollinear
propagation vectors.

10

Temperature (K)

I;

2
15

Fig. 115. Low-temperature variation of the

magnetic contribution to the specific heat in
PrGa2 (Ball et al. 1993c). The solid line
results from calculation with the PF model
(AM structure). For comparison the simple antiferromagnetic case (EM structure) is also drawn
(dashed line). Inset: corresponding variation of
the entropy.

the PF model does not apply only for collinear structures but also for noncollinear
arrangements such as those expected in easy-axis systems when the field is applied
perpendicular to the easy axis, for instance in PrNi2Si2 (Blanco et al. 1992a). Indeed,
the model allowed one to discuss the magnetization processes observed in GdGa2, in
particular the possibility of field-induced helical-to-fan structure transition (Ball 1993). As
well, in PrGa2, which is easy-plane, metamagnetic processes involving domain effects and
noncollinear field-induced structure have been fruitfully discussed, thanks to this model
(Ball 1993; fig. 116).
An interesting capability of the PF model concerns the microscopic description
of the phase diagrams, in particular the intermediate magnetic phases. Indeed, the
PF model may allow one to understand, in some cases, the occurrence of these new
magnetic configurations. As explained above, the periodicity of the exchange field
follows that of the magnetic moments. However, while in the majority of situations
these latter keep a constant magnitude over the whole magnetic unit cell at low
temperature (antiphase structure), the same is not true for the exchange field itself,
through eq. (29). Some magnetic sites, namely those situated near the nodes of the
modulation, then experience a reduced exchange field, so that the corresponding magnetic

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

"F

405

2o

E12ol

-:"<...

E
10

T/T N

00

20

40
H (kOe)

60

80

Fig. 116. Field dependence of the magnetization in hexagonal

PrGa 2 compound along the [120] direction at 5 K (Ball et al. 1993c,
Ball 1993). The circles are experimental; the solid line is calculated
by the PF model for the D3o- and Dgo-domains (see references).
Several noncollinear configurations for the magnetic moments are
also shown. Inset: initial moment configurations of the domains
with respect to the applied field for the calculations.

Fig. 117. Calculated magnetic contribution to the specific heat as a function of

the reduced temperature T / T N for (a)ferromagnetic and (b) AM structures. The
exchange parameters are: (a) J(0) = 10;
(b) J ( Q ) = 1 0 (arbitrary units) (Blanco
et al. 1991a).

moments will be flipped preferentially by an applied magnetic field, resulting in a onestep metamagnetic transition. Typical examples are provided by the spin-flip transition
observed in low field and at low temperature in PrGa2 (see fig. 82, sect. 3.2.8), or in
TbNizSi2 (Blanco et al. 1992c).
Another remarkable result obtained by the PF model concerns the thermodynamical
properties of the incommensurate or long-period commensurate systems. Indeed, it
has been shown that, in the case of a single-Q magnetic structure, the specific-heat
discontinuity AC at TN of a system having an amplitude-modulated (AM) structure is
only two third of that present in a compound with an equal-moment (EM) structure, i.e.
a ferromagnetic, a simple antiferromagnetic or a helimagnetic compound (fig. 117; Blaneo
et al. 1991a). This noticeable reduction occurs because of the modulated distribution
of the magnetic moments and accordingly of their internal energy just below TN. In
the case of a double-Q or a triple-Q structure, it can be easily extrapolated that the
reduction is still stronger, the ratio CAM~GEMthen reaching (2)2 or (2)3, respectively.
This unique feature, first derived in Gd-based systems (Blanco et al. 1991a), has been
extended later to compounds in the presence of CEF effects (Blanco et al. 1992a). In
these latter compounds however, the analysis is more complex and requires first a good
knowledge of the CEF parameters. The second important consequence of the magnetic
modulation on the thermodynamical properties is the dependence of the slope of C at
TN on the ratio J(3Q)/J(Q) for AM systems. In particular, under certain conditions, this
slope may become negative, leading to the original feature that a maximum of specific
heat must occur at a temperature below TN. It follows that the whole shape of the specific
heat variation below TN is affected by the modulation, and more specifically by the
relative amplitude of the successive J(nQ) parameters (fig. 118). These original results
throw a new light on the experimental temperature dependences of specific heat found in
the literature, in particular in Gd-based compounds (Bouvier et al. 1991 and references

406

D. GIGNOUX and D. SCHMITT

2c
O

E
-~ 1C

-S
d ...,.,,..
.....f -.&.
..--C

1
T/T N

Fig. I 18. Calculated magnetic contribution to the specific heat for various sets of exchange parameters ( J ( Q ) = 10
for all curves): (a) J(3Q) = - 8 , J(5Q) = 8; (b) J(5Q) = 8; (c) J(3Q) = 8; (d) J(3Q) = 8, J(5Q) = 8 (arbitrary units)
(Blaneo et al. 1991a).

therein). They should be quite useful to choose between a helical, i.e. EM, and an AM
structure near TN, from specific heat data.
One aspect of the low-temperature metamagnetic processes of easy-axis compounds
with large anisotropy concerns the hysteresis of the transitions which present the
following features: (i) The experimental as well as the calculated hysteresis decrease when
temperature is increased due to thermal effects as shown in the case of HoA1Ga (Ball
et al. 1993a). (ii) From the analysis of the relative amplitude of the hysteresis of the
three metamagnetie steps observed in the Zn-rich Ce(Znl-xCux)2 compounds it has been
concluded that hysteresis arises from the energy barrier associated with the exchange
interactions and not with the anisotropy (Gignoux et al. 1992). (iii) The calculated
hysteresis, in particular at low temperature, is much larger than the observed one. This
last point, which is still a puzzle, could be due to: (1) the crossing of the energy barrier
through thermal fluctuations; (2) the effect of higher-order J(nQ) harmonics not taken
into account in the calculation; and/or (3) a true distribution of exchange fields larger
than the calculated one due to the true incommensurability of the analyzed structure or
to the existence of defects.
In conclusion, the self-consistent PF model is a powerful tool for the analysis of a large
variety of long-period magnetic structures, collinear (antiphase, amplitude-modulated) as
well as noncollinear (helix, fan. . . . ). In the examples presented above this model is
limited to single-Q (and obviously to its harmonics and Q =0) systems but it can be
extended without problem to the case of multi-Q systems.

4. Summary and conclusion

This survey shows the extreme wealth of information and unusual behaviour of rare-earth
intermetallics in magnetism, including itinerant 3d magnetism in addition to localized
4f magnetism. During the last fifteen years crucial advances were obtained thanks to
experiments of higher performances (in particular magnetization measurements in intense
magnetic field and neutron diffraction experiments) and the simultaneous evolution of the
theoretical approaches.

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

407

Concerning 3d magnetism, during this period the most dramatic results, to our feeling,
have been obtained on: (i) collective electron metamagnetism; (ii) 3d magnetocrystalline
anisotropy; and (iii) the simultaneous effects of topological frustration and magnetic
instability as observed in the fascinating RMn2 series. This latter topic is expected to
have still further interesting experimental as well as theoretical developments in the future,
not only for the cubic RMn2 series but also for other Mn series such as the tetragonal
RMn2Si2 compounds for instance.
Important improvements have also been obtained in the knowledge of the compounds
where both 4f and 3d magnetism are well established. These compounds, rich in
3d elements, are of special interest not only for their fundamental aspects but also for
their technological applications, in particular as permanent magnets and magnetostrictive
materials. The complex magnetization processes in intense field as measured on
single crystals allowed a better knowledge of the main interactions involved, namely
exchange couplings and magnetocrystalline anisotropies. Further useful information
will surely be obtained in still larger applied fields, large enough to break the large
exchange coupling between the 3d and 4f sublattices.
In the intermetallic compounds where only the rare-earth element is magnetic, a
large number of metamagnetic behaviours of quite different origins (crystal field,
quadrupolar interactions and/or frustration of bilinear exchange interactions) have shown
up, and the often complex magnetic phase diagrams have been determined in some
compounds. In most cases these metamagnetic processes are rather well understood and
quantitatively analyzed. If we take into account the metamagnetic processes characterizing
3d magnetism in the above compounds, one can tell that metamagnetism is the most
common feature encountered in lanthanide-based intermetallic compounds. At this
moment, a large number of studies are devoted to the determination of complex
magnetic phase diagrams for a large number of compounds where the frustration of
RKKY-type exchange interactions associated with crystal-field anisotropy frequently leads
to incommensurate high-temperature structures which compete with low-temperature
commensurate ones. The number of systems where the microscopic aspects of the
phase diagrams have been investigated is rather limited, and a special effort should be
made in this direction. Simple uniaxial compounds, in particular those with hexagonal
and tetragonal structures, have been studied most extensively. Accurate magnetization
measurements and neutron diffraction in magnetic field on single crystals would allow
us in the future to obtain a better knowledge of this type of systems, in particular the
easy-plane ones where the phase diagrams are more complex than in easy-axis systems,
due to the degeneracy of the easy magnetization direction. Special attention will also
need to be paid to the weakly anisotropic Gd-based systems, in particular with respect
to their microscopic magnetic structures which have not, so far, been studied due to
the large neutron absorption of Gd. Finally let us note that, for each compound where
L ~ 0, a good knowledge of the crystal-field parameters is needed in order to perform an
accurate quantitative analysis of the complex magnetic properties in the ordered magnetic
state.

408

D. GIGNOUXand D. SCHMITT

5. Appendix: Definitions and/or descriptions of magnetic terms or phenomena

5.1. Antiferroquadrupolar and ferroquadrupolar ordering
In a form analogous to the bilinear exchange coupling between magnetic dipoles, a twoion coupling between 4f quadrupoles O~ (see quadrupolar moment) is possible and has
been found in several rare-earth intermetallic compounds. As an example, if we restrict
ourselves to the single component m--0 among the five quadrupolar components, this
coupling between two ions (i and j) can be written as

~Q = -K(ij) O(i) O(j).

A ferroquadrupolar (antiferroquadrupolar) coupling, i.e. K(ij) > 0 (< 0), favours electronic
distributions on sites i and j such that their quadrupoles have their local z axes
parallel (perpendicular). When the magnitude of this interionic coupling is large enough,
compared to the bilinear exchange coupling, ordering of the 4f quadrupoles can occur
without simultaneous magnetic ordering. Such quadrupolar ordering has only been
observed in few experimental situations (see Morin and Schmitt 1990). Couplings
K(ij) which are predominantly positive lead to a ferroquadrupolar ordering where all
quadrupoles are aligned (fig. 119a). On the contrary, if the K(ij) are mainly negative, an
, antiferroquadrupolar arrangement will be obtained (fig. 119b).

0@@@@
Ferroquadrupolarordering

(a)

Antiferroquadrupolarordering

(b)

Fig. 119. Schematicrepresentationsof (a) ferroquadrupolar and (b) antiferroquadmpolarorderings.

5.2. Amplitude-modulated structure

The simplest amplitude-modulated structure is the collinear sine-wave modulated structure, illustrated in fig. 74 (sect. 3.2.7) in the case of PrNi2Si2, in which the moment at
site i is given by
M(ri) = M0 cos(Q ri + (9).

MAGNETIC PROPERTIESOF INTERMETALLICCOMPOUNDS

409

Such structures are stabilized at TN when: (i) the magnetocrystalline anisotropy is strong
and uniaxial so that the moment always remains parallel to the easy axis; (ii) the Fourier
1 or ~K
1 (this
transform J(q) of the magnetic interactions is maximum for Q ~ 0, ~K
latter situation is particular because it can lead either to modulated or to equal-moment
structures, see Gignoux and Schmitt 1993), where K is a reciprocal lattice vector. The
periodicity associated with the propagation vector Q is then usually long-period or
incommensurate.
In this type of structure, the moment of a large number of atoms is reduced. When
the intrinsic magnetic moment, in presence of crystal field, is well defined, amplitudemodulated magnetic structure is not stable at low temperature, especially at OK. It
transforms into a structure with equal moments (ferromagnetic, simple antiferromagnetic,
antiphase, helicoidal). However, when in the absence of applied or exchange field,
the ground state is a nonmagnetic singlet, the magnetic moment at 0 K is induced
by the field and an amplitude-modulated magnetic structure can be stable down to
very low temperature.
5.3. Antiphase structure
An antiphase magnetic structure is a collinear structure with equal moments which
can be described with a long-period commensurate or incommensurate (see under
commensurate, incommensurate structures) propagation vector Q and its harmonics.
If Q is incommensurate or commensurate, Q = ( p / q ) K with q even, the resulting
magnetization is zero; the antiphase structure is said to be compensated. In case of one
magnetic atom per unit cell, if Q is commensurate with q odd, there is a small resulting
magnetization; the antiphase structure is said to be uncompensated. Examples of antiphase
structures are shown in figs. 82 (sect. 3.2.8) and 113.
5.4. Collinear and noncollinear structures
A collinear (noncollinear) structure has all magnetic moments parallel (not parallel).
Noncollinear and multiaxial have the same meaning.
5.5. Commensurate and incommensurate structures
The magnetic periodicity is commensurate when the propagation vector can be written
as Q = ( p / q ) K where p and q are integer. This means that the magnetic periodicity is
q times the crystallographic one.
An incommensurate magnetic periodicity has a propagation vector which can be written
as Q = r K with r ~p/q.

410

D. GIGNOUX and D. SCHMITT

5.6. Equal-moment structure

In contrast with amplitude-modulated structures, the magnetic moments in equalmoment structures all have the same amplitude; this includes ferromagnetic, simple
antiferromagnetic, antiphase, helical or triangular structures.
5.7. Exchange interaction J (ij) and Fourier transform J (q)
The exchange energy, which is of electrostatic origin, between magnetic moments at sites
i a n d j can be written as follows:
(i) For lanthanides: E U= - J ( i j ) (J(i)) (J(j)), where J(i) and J(j) are the total angular
momenta.
(ii) For 3d atoms: E O.=-J(ij) (S(i)) (S(j)), where S(i) and S(j) are the spin momenta.
J(i) [or S(i)] is proportional to the magnetic moment M(i), and according to the sign of
J(ij) this interaction favours either parallelism or antiparallelism of magnetic moments.
The Fourier transform of this interaction,
J(q) = Z

J(ij) exp[iq(rj - ri)],

ji

is a useful function. In particular, it can be shown that the Q vector of the reciprocal space
for which J(q) is maximum defines the propagation vector, i.e. the magnetic periodicity,
of the system just below the ordering temperature.
5.8. Fan structure
When a helimagnetic structure is submitted to a field applied along one direction of the
plane of the moments, for a field larger than a critical value the structure becomes a fan
structure (fig. 120). The regular rotation of the moment in the helimagnetic arrangement
/

i+4
i+l

k~i

i+5

\
\

Fig. 120. Schematic representation of a fan structure.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

411

transforms into a sine wave oscillation of the moments about the field direction between
two maximum values of the angle (+0M and --0M). At OK, moments have the same
amplitude and their angle with field direction are such that Oi = 0M sin(Q ri + 0).
5.9. Flopside structure
The word "topside structure" has been introduced to characterize a biaxial magnetic
structure observed in the cubic HoP compound (Fischer et al. 1985). As shown
in fig. 121 the propagation vector is Q=(, g,
l g)
l and the structure is formed as
a stacking of ferromagnetic planes with their moments successively parallel to two
different edges of the cube. This structure originates from the degeneracy of the
easy (100) magnetization direction and the compromise between ferromagnetic bilinear
exchange interactions and antiferroquadrupolar interactions. This structure observed
in zero field in HoP can be induced by a field in other compounds such as DySb
(Everett and Streit 1979).

(b)

Fig. 121. Flopside magnetic structure of

the HoP cubic compound.

Fig. 122. A simple picture of the frustration: (a) if two moments are
antiparallel in order to satisfy the negative interaction (J < 0), the
direction of the third cannot be antiparallel to both of the others;
(b) under certain conditions the system finds a compromise through
a triangular magnetic structure.

5.10. Frustration
In magnetism, this word is generally used for magnetic structures in which the interatomic
magnetic interactions are not all satisfied. A schematic example is given by three magnetic
moments at the corners of an equilateral triangle (fig. 122a) or by four magnetic moments
at the comers of a regular tetrahedron. If the nearest neighbours are antiferromagnetically
coupled (J(ij)< 0), any magnetic arrangement is frustrated, i.e. cannot simultaneously
satisfy all the couplings. The resulting structure should then be a compromise such as that
shown in fig. 122b, in which the moments make an angle of 2Jr/3 (triangular structure).
Helimagnetic or amplitude modulated structures are typical examples of such frustrated
arrangements. Most compounds with antiferromagnetic interactions are frustrated. This
generally occurs in rare-earth intermetallics on account of the RKKY long-range and
oscillatory character of the exchange interactions.

412

D. GIGNOUX and D. SCHMITT

5.11. Helical structure

In a helical or helimagnetie structure, as sketched in fig. 123, the magnetic moment at
site i is given by
M(ri) = M [u cos(Q ri) + v sin(Q ri)],
where Q is the propagation vector, while u and o are perpendicular unit vectors defining
the plane P in which the moments remain parallel. This structure can then be described as
a stacking o f collinear ferromagnetic planes perpendicular to Q. From plane to plane the
moments o f each plane rotate in plane P o f the same angle. The propagation vector Q is
generally long-period or incommensurate. Such structure can be stabilized when the magnetocrystalline anisotropy imposes the moments to lie in plane P and when the anisotropy
in the plane is small (for instance hexagonal systems). Note that in numerous compounds
Q is perpendicular to the plane P. When the propagation vector lies in the plane P o f the
moments, the structure is also called cycloidal. Classical examples o f helimagnetic structures are found in MnAua, MnO2 and in the lanthanide metals such as Tb, Dy and Ho.

HelifQn (3/2)

<Z-_-_
i'

Fig. 123. Helical structure.

Fig. 124. The helifan (3/2) structure in Ho at 50K. The moments lie in
planes normal to the c-axis and their relative orientations are indicated
by arrows. A magnetic field of 11 kOe is applied in the basal plane, and
moments with components respectively parallel and antiparallel to the
field are designated by solid and open arrowheads. This componentof
the moments has a periodicity which is 3/2 that of the corresponding
helix, and the helicity of the structure changes regularly (after Jensen
and Mackintosh 1992).

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

413

5.12. Helifan
When a magnetic field is applied in the plane of a helical structure, an intermediate phase
can be stabilized in a small field range before the fan structure is reached, this latter
closing progressively toward the ferromagnetic state. This intermediate phase is called
the helifan structure. Such a phase has been recently discovered in Ho (fig. 124) and Dy
from a careful neutron diffraction investigation (Jensen and Mackintosh 1992).

5.13. Multiaxial structure

See under collinear and noncollinear structure.

5.14. Multistep metamagnetic process

At low temperatures, a large number of compounds exhibit magnetization processes
called multistep metamagnetic processes, characterized by more than one metamagnetic
transition. Such behaviour is illustrated in figs. 68, 69, 72, 79, 81, 84, etc.
5.15. Multi-Q structure (double, triple, ...)
A multi-Q structure is necessarily described with different propagation vectors:

M(ri) = ~

MnQj exp(inQjrg).

n,j

In many multi-Q structures (in particular double and triple-Q structures) all Qj are
equivalent in reciprocal space and can be deduced from one another through pointsymmetry operations. They form a so-called star. The structure is then multiaxial with
equal moments. Such structures are often observed in cubic systems on account of the
high degeneracy of the symmetry axes. Note that Qi and Qj can be considered as different
only if Qi - Qj =~K, where K is a reciprocal lattice vector. For more details about multiQ structures see for instance Rossat-Mignod (1986).
5.16. Quadrupolar moment
The charge distribution of the electrons of the 4f shell can be expanded in successive
harmonics. The lowest-order terms of this expansion, i.e. the second-order ones, are the
5 components of the quadrupolar moment. For a charge distribution associated with a state

414

D. GIGNOUX and D. SCHMITT

In) = gt(1, 2 , . . . , i , . . . , n) of the n electrons of the 4f shell of a lanthanide ion these 5 (for
m = 1 to 5) components are given by:

Cm ~

~*(1, ..., i, ..., n ) f m(ri) ktt(1, ..., i, ..., n)d3ri,

i=1

where the

Cm

are constants and the fm are

f l ( r ) = 3z2 - r z,

f2(r) = x 2 _y2,

f3(r) ---xy,

f 4 ( r ) =yz,

f s ( r ) = zx.

As the In) state is a linear combination of [ J , Mj) states characterizing the total kinetic
momentum, so called Stevens Equivalent Operators have been introduced to calculate
these component. For instance, the first component can be rewritten as

C1 ~

I/-/*(1, ..., i, ..., n ) ( 3 z ~ - r 2) tp(1 . . . . . i . . . . , n)d3ri = C'aj (u

Iol u),

i=1

where C' is a constant, aj is the second-order so called Stevens coefficient characteristic

of the 4f shell of the considered lanthanide atom and 020 =3Jz2 - J ( J + 1) is an operator
acting on the ~/, M j ) states. In terms of operators used in a hamiltonian this quadrupolar
component can be written as C'ajO . The operators associated with the otherfi functions
are

0 2 = j2x _,12,

Pxy = l(JxJy -JyJx),

Pyz = l(jyJz -JzJy),

Pzx = l(JzJx -JxJz).

A positive value of the C'aj {u I0 I u) component of the quadrupole corresponds to

a distribution of the 4f electrons elongated along the z axis (prolate shape). On the
contrary a negative value of C laJ (u IO1 u) corresponds to an pancake-like distribution
(oblate shape).
Relevant references are Hutchings (1964), Schrnitt (1986), Morin and Schmitt (1990).
5.17. RKKY (Ruderman, Kittel, Kasuya, Yosida) exchange interaction
This term concerns the long-range and oscillatory character of the bilinear exchange
interaction between lanthanide magnetic moments in a metallic compound. Due to the
localization of the 4f shell, this interaction is not direct, as is the case between 3d shells of
transition metals such as Fe, Co or Ni, but indirect. This means that it is mediated through
the polarization of the conduction electron. This interaction between lanthanide atoms is
much smaller than the direct interaction in 3d metal, leading to ordering temperatures
in the former one order of magnitude smaller than in pure Fe, Co or Ni. On account

MAGNETICPROPERTIESOF INTERMETALLICCOMPOUNDS

415

of the itinerant character of conduction electrons, the effective interionic interaction is

long-range and oscillatory. This leads to a Fourier transform of this interaction as

J(q) = Z

J(ij)exp[iq(rj - ri)],

j ~ei

where J(ij) is the exchange interaction between atoms i and j, which frequently has
a maximum for a non-simple value of Q. The magnetic structure, at least near the
ordering temperature, is then long-period-commensurate or incommensurate with the
corresponding propagation vector Q.
Relevant references are Ruderman and Kittel (1954), Kasuya (1956), Yosida (1957)
and Coqblin (1977).

5.18. Single-Q structure

In the single-Q structure the magnetic moments can be described with a unique
propagation vector Q and possibly its harmonics:
M(ri) = Z

MnQ exp(inQ ri).

There is a large variety of single-Q antiferromagnetic structures: (sine-wave) amplitudemodulated, antiphase, simple antiferromagnetic, helimagnetic, etc.
5.19. Spin-flip transition
In compounds with large uniaxial anisotropy, when the field H is applied along the easy
axis of an equal-moment collinear antiferromagnetic structure, the magnetization process
(of metamagnetic type) involves, for H > He, the flipping of all or part of the moments
antiparallel to the field which then become parallel to the latter (fig. 125). The concept of
spin-flip transition was first introduced by Nrel to interprete the metamagnetism of this
type of compounds. In some cases the metamagnetic behaviour involves simultaneously
spin-flip and anti spin-flip processes (see for instance Gignoux et al. 1993).
H<Hc
M2
--

~2

H>Hc
M1
~

M2
.

M1
~

_ _ _

__

Fig. 125. Spin-fliptransition. M t and M z schematizethe directionsof the two sublatticesinvolved.

5.20. Spin-flop transition

In compounds with small magnetocrystalline anisotropy, when the field is applied along
the moment direction of a collinear antiferromagnetic structure, for a critical field Hc

D. GIGNOUX and D. SCHMITT

416

H<Hc

H>Hc

Fig. 126. Spin-flop transition. M1 and M 2 schematize the directions of the two sublattices involved.
the moments take the magnetic configuration schematized in fig. 126. The concept o f
spin-flop transition was first introduced by N6el to account for this type o f transition. For
larger fields the moments progressively rotate towards the field direction.
5.21.

Spin fluctuations

This quite general concept is used to specify any shift (in time and space) between the
instantaneous value o f magnetic moments and their mean value given by the molecular
field approximation and the Boltzmann statistics. Spin fluctuations can be individual
and/or collective. In the latter case processes such as spin waves or paramagnons are
considered and many models have been developed to interprete experiments with these
concepts.
5.22.

Spin-slips (or spin discommensurations)

The concept o f spin-slips or discommensurations was first introduced to explain the

observed lock-in transitions in the magnetic spirals o f lanthanide metals such pure Ho
and D y in terms o f simple commensurate structures (fig. 127). More generally this term
can be used to characterize structures which present periodic faults in a simple sequence
o f magnetic moments. For instance, let us consider, in an Ising chain, a sequence o f
4 moments up followed by 3 moments down (this is found in some compounds). The
propagation vector is ~. In certain regions of H-T space the propagation vector is slightly

Fig. 127. Self-consistent mean-field calculations of periodic structures in Ho. Each circle represents the
magnitude and direction of the ordered moment in a specific plane, relative to the size of the moment at
absolute zero (10/~B), indicated by the length of the horizontal lines. The orientation of moments in adjacent
planes is depicted by the positions of the neighbouring circles. (a) The 12-layer zero-spin-slip structure at 4 K.
The open circle in the centre indicates the ferromagnetic component in the cone structure; (b) the 1l-layer
one-spin-slip structure at 25 K. The bunched pairs of moments are disposed asymmetricallywith respect to the
easy axis in the vicinity of the spin slip (after Jensen and Mackintosh 1992).

MAGNETIC PROPERTIES OF INTERMETALLICCOMPOUNDS

417

different, for instance 4 instead of 4 . This indicates the presence of a fault each four
sequences.
The term spin-slip transition is used to characterize the change from one discommensuration to another. It is generally associated with a metamagnetic transition of small
amplitude in the magnetization process and a small shift of the propagation vector, as in
PrGa2 (fig. 81) at 1,5K in fields smaller than 9 k O e (Ball et al. 1994a).
Relevant references are Gibbs et al. (1985), Cowley and Bates (1988), Jensen and
Mackintosh (1992), Gibbs et al. (1986) and Bohr et al. (1986).
5.23. Spin-reorientation transition
This term specifies all transitions with temperature (or even pressure) which involve
a change in the magnetization direction associated with a symmetry change. These
transitions can be of first or second order. Many compounds exhibit such spin
reorientations. For instance, in the cubic HoZn compound the magnetization, which
is along a two-fold axis below 25K, aligns, through a first-order transition, along
a three-fold axis above 2 5 K (Morin and Schmitt 1990). As well, in the hexagonal
TbCo5 compound, magnetization is along the six-fold axis below 3 9 0 K and in the
basal plane above 440 K. Between these two temperatures a progressive rotation takes
place (Lemaire 1966). Under an applied magnetic field spin-reorientation transitions
are more specific phenomena in a sense that, when the applied field is not parallel
to the initial magnetization, a rotation of the latter always occurs. In that case,
spin-reorientation transitions, generally of first order, characterize drastic acceleration
o f such rotation process. Such transitions, called FOMR (first-order magnetization
reorientation), are frequently observed in lanthanide-3d intermetallics such as Ho2Co17
and Dy2COl7 (Franse and Radwanski 1993). Field-induced spin reorientations have been
reviewed by Asti (1990).

References
Abe, S., T. Kaneko, M. Ohashi, Y. Nakagawa and T.
Kitai, 1992, J. Magn. & Magn. Mater. 104-107,
1403.
Alameda, J.M., D. Givord, R. Lemaire and Q. Lu,
1981, J. Appl. Phys. 52, 2079.
Alekseev, P.A., A. Andreeff, H. Griessmann, L.P.
Kaun, B. Lippold, W. Matz, I.P. Sadikov, O.D.
Christyakov, I.A. Markova and E.M. Savitskii, 1980,
Phys. Status Solidi b 97, 87.
A16onard, R., and E Morin, 1979, Phys. Rev. B 19,
3868.
Al~onard, R., and E Morin, 1984, J. Magn. & Magn.
Mater. 42, 151.

A16onard, R., and P. Morin, 1990, L Magn. & Magn.

Mater. 84, 255.
A16onard, R., P. Morin, D. Schmitt and E Hulliger,
1984a, J. Phys. F 14, 2689.
A16onard, R., P. Morin and J. Rouchy, 1984b, J. Magn.
& Magn. Mater. 46, 233.
Amara, M., R.M. Gal~ra, R Morin, T. Veres and P.
Burlet, 1994, J. Magn. & Magn. Mater. 130, 127.
Andoh, Y., 1987, J. Phys. Soc. Jpn. 56, 4075.
Andres, K., and S. Darack, 1977, Physiea B 86-88,
1071.
Andres, K., S. Darack and H.R. Ott, 1979, Phys. Rev.
B 19, 5475.
Asti, G., 1990, in: FerromagneticMaterials, Vol. 5, eds

418

D. GIGNOUX and D. SCHMITT

K.H.J. Buschow and E.P. Wohlfarth (North-Holland,

Amsterdam) ch. 5, p. 397.
Aubert, G., D. Gignoux, B. Hermion, B. Michelutti
and A. Nait-Saada, 1981, Solid State Commtm. 37,
741.
Bak, P., 1982, Rep. Prog. Phys. 45, 587.
Bak, P., and J. von Boehm, 1980, Phys. Rev. B 21,
5297.
Ball, A.R., 1993, Thesis (University of Grenoble)
unpublished.
Ball, A.R., D. Gignoux, EE. Kayzel, D. Schmitt and
A. de Visser, 1992a, L Magn. & Magn. Mater. 110,
337.
Ball, A,R., D. Gignoux, D. Schmitt, EY. Zhang and
M. Reehuis, 1992b, J. Magn. & Magn. Mater. 110,
343.
Ball, A.R., D. Gignoux and D, Schmitt, 1992e, Physica
B 180-181, 58.
Ball, A.R., D. Gignoux and D. Schmitt, 1993a, J.
Magn. & Magn. Mater. 119, 96.
Ball, A,R., D. Gignoux, D. Sehmitt and EY. Zhang,
1993b, Phys. Rev. B 47, 11887.
Ball, A.R., D. Gignoux, A.E Murani and D. Schmitt,
1993c, Physica B 190, 214.
Ball, A.R., D. Gignoux, D. Sehmitt, EY. Zhang and
P. Burlet, 1994a, J. Magn. & Magn. Mater. 130,
317.
Ball, A.R., D. Gignoux, J. Rodriguez Fernandez and
D. Sehmitt, 1994b, J. Magn. & Magn. Mater. 137,
281.
Ballou, R., and R. Lemaire, 1988, J. Physique 49,
C8-523.
Ballou, R., D. Gignoux, R. Lemaire, R. MendiaMonterroso and J. Schweizer, 1986a, J. Magn, &
Magn. Mater. 54-57, 499.
Ballou, R., D. Gignoux, B. Gorges and R. Lemaire,
1986b, J. Magn. & Magn. Mater. 54-57, 497.
Ballou, R., J. Ddportes, R. Lemaire, Y. Nakamura and
B. Ouladdiaf, 1987, J. Magn. & Magn. Mater. 70,
129.
Ballou, R., J. Ddportes, R. Lemaire and B. Ouladdiaf,
1988a, J. Appl. Phys. 63, 3487.
Ballou, R., VM.T.S. Barthem and D. Gignoux, 1988b,
Physica 149B, 340.
Ballou, R., B. Gorges, R. Lemaire, H. Rakoto and J.C.
Ousset, 1989, Physica B 155, 266.
Ballou, R., J. Ddportes, R. Lemaire, R Rouault and
J.L. Soubeyroux, 1990a, J. Magn. & Magn. Mater.
90-91, 559.
Ballou, R., B. Gorges and P. Molho, 1990b, J. Magn.
& Magn. Mater. 84, L1.

Ballou, R., C. Lacroix and M.D. Nunez-Regueiro,

1991, Phys. Rev. Lett. 66, 1910.
Ballou, R., B. Ouladdiaf, P.J. Brown, M.D. NunezRegueiro and C. Lacroix, 1992, Phys. Rev. B 45,
3158.
Barandiaran, J.M., D. Gignoux, D. Schmitt, J.C.
Gomez-Sal, J. Rodriguez Fernandez, P. Chieux and
J. Schweizer, 1988, J. Magn. & Magn. Mater. 73,
233.
Barbara, B., M. Rossignol and P. Bak, 1978, J. Phys.
C 11, L183.
Barbara, B., D. Gignoux and C. Vettier, 1988,
in: Lectures on Modern Magnetism, ed. Li Yifa
(Springer, Heidelberg) ch. 10.
Barthem, V.M.T.S., D. Gignoux, A. Nait-Saada and D.
Schmitt, 1988, Phys. Rev. B 37, 1733.
Barthem, V.M.T.S., D. Gignoux, A. Nait-Saada, D.
Schmitt and A. Takeuchi, 1989a, J. Magn. & Magn.
Mater. 80, 142.
Barthem, V.M,T.S., D, Gignoux, D, Schmitt and G.
Creuzet, 1989b, J. Magn. & Magn. Mater. 78, 56.
Bdal-Monod, M.T., 1982, Physiea B 109-110, 1837.
B~al-Monod, M.T., and J.M. Lawrence, 1980, Phys.
Rev. B 21, 5400.
B~al-Monod, M.T., S.K. Ma and D.R. Fredkin, 1968,
Phys. Rev. Lett. 22, 929.
B~cle, C., R. Lemaire and D. Paccard, 1970, J. Appl.
Phys. 41, 855.
Belorizky, E., M.A. Frdmy, J.P. Gavigan, D. Givord
and H.S. Li, 1987, J. Appl. Phys. 61, 3971.
Belorizky, E., J.P. Gavigan, D. Givord and H.S. Li,
1988, Europhys. Lett. 5, 349.
Berk, N., and J.R. Schrieffer, 1966, Phys. Rev. Lett.
17, 433.
Berthier, Y., J. D~portes, M. Horvatic and P. Rouault,
1988, J. Phys. (Paris) 49, C8-261.
Blanco, J.A., D. Gignoux and D. Schmitt, 1991a, Phys.
Rev. B 43, 13145.
Blanco, J.A., D. Gignoux, D. Schmitt and C. Vettier,
1991b, J. Magn. & Magn. Mater. 97, 4.
Blanco, J.A., D. Schmitt and J.C. Gomez-Sal, 1992a,
J. Magn. & Magn. Mater. 116, 128.
Blanco, J.A., D. Gignoux and D. Sehrnitt, 1992b, Phys.
Rev. B 45, 2529.
Blanco, J.A., D. Gignoux and D. Schmitt, 1992c, Z.
Phys. B 89, 343.
Bloch, D., E Chaissd, E Givord, J. Voiron and E.
Burzo, 1971, J. Phys. (Paris) 32, Clq559.
Bloeh, D., D.M. Edwards, M. Shimizu and J. Voiron,
1975, J. Phys. F 5, 1217.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

Bohr, J., D. Gibbs, D.E. Moncton and K.L. D'Amico,
1986, Physica A 140, 349.
Borombaev, M.K., R.Z. Levitin, A.S. Markosyan, VA.
Reimer, E.V. Sinitsyn and Z. Smetana, 1987, Sov.
Phys.-JETP 66, 866.
Boutron, P., 1973, Phys. Rev. 7, 3226.
Bouvier, M., P. Lethuillier and D. Schmitt, 1991, Phys.
Rev. B 43, 13137.
Brinkman, W.E, and S. Engelsberg, 1968, Phys. Rev.
169, 417.
Brown, P.J., B. Ouladdiaf, R. Ballou, J. D@ortes and
A.S. Markosyan, 1992, J. Phys.: Condens. Matter
4, 1103.
Brtm, T.O., EW. Korty and J.S. Kouvel, 1980, L Magn.
& Magn. Mater. 15-18, 298.
Bruno, R, 1989, Phys. Rev. B 39, 865.
Burzo, E., and R. Lemaire, 1992, Solid State Commun.
84, 1145.
Buschow, K.H.J., 1977a, Rep. Prog. Phys. 40, 1179.
Busehow, K.H.J., 1977b, Solid State Commun. 21,
1031.
Buschow, K.H.J., 1979, Rep. Prog. Phys. 42, 1373.
Buschow, K.H.J., 1980, in Ferromagnetic Materials,
Vol. 1, ed. E.P. Wohlfarth (North-Holland, Amsterdam) oh. 4, p. 297.
Buschow, K.H.J., 1988, in: Ferromagnetic Materials,
Vol. 4, eds E.P. Wohlfarth and K.H.J. Buschow
(North-Holland, Amsterdam) ch. 1, p. 1.
Buschow, K.H.J., and R.P. van Stapele, 1970, J. Appl.
Phys. 41, 4066.
Campbell, I.A., 1972, J. Phys. F 2, L47.
Chattopadhyay, T., P. Burlet and J. Rossat-Mignod,
1986, J. Magn. & Magn. Mater. 54--57, 503.
Cooper, B.R., 1966, Phys. Lett. 22, 244.
Coqblin, B., 1977, The Electronic Structure of RareEarth Metals and Alloys: the Magnetic Heavy RareEarths (Academic Press, New York).
Cowley, R.A., and S. Bates, 1988, J. Phys. C 21,
4113.
Cyrot, M., and M. Lavagna, 1979, J. Phys. (Paris) 40,
763.
Cyrot, M., D. Gignoux, E Givord and M. Lavagna,
1979, J. Phys. (Paris) 40, C5-171.
Czjzek, G., V. Oestreich, H. Schmidt, K. Latka and K.
Tomala, 1989, J. Magn. & Magn. Mater. 79, 42.
Czopnik, A., N. Iliew, B. Stalinski, C. Bazan, H.
M~idge and R. Port, 1985, Physica B 130, 259.
Daalderop, G.H.O., RJ. Kelly and M.EH. Schuurrnans,
1992, J. Magn. & Magn. Mater. 104--107, 737.
Date, M., 1988, J. Phys. Soc. Jpn. 57, 3682.
Date, M., 1990, J. Magn. & Magn. Mater. 90-91, 1.

419

del Moral, A., M.R. Ibarra, LS. Abell and LED.

Montenegro, 1987, Phys. Rev. B 35, 6800.
D@ortes, J., R. Lemaire, B. Ouladdiaf, E. Roudaut
and E Sayetat, 1987a, J. Magn. & Magn. Mater. 70,
191.
Drportes, J., B. Ouladdiaf and K.R.A. Ziebeck, 1987b,
J. Magn. & Magn. Mater. 70, 14.
Doniach, S., and S. Engelsberg, 1966, Phys. Rev. Lett.
17, 750.
Doukourr, M., and D. Gignoux, 1982, J. Magn. &
Magn. Mater. 30, I 11.
Duc, N.H., D. Givord, C. Lacroix and C. Pinettes,
1992, Europhys. Lett. 20(1), 47.
Effantin, J.M., J. Rossat-Mignod, R Butler, H.
Bartholin, S. Kunii and T. Kasuya, 1985, J. Magn.
& Magn. Mater. 47-48, 145.
Everett, G.E., and P. Streit, 1979, J. Magn. & Magn.
Mater. 12, 277.
Felcher, G.R, L.M. Corliss and L.M. Hastings, 1965,
J. Appl. Phys. 36, 1001.
Fillion, G., E Givord, P. Lejay, E. Ressouche, J.
Schweizer and A. Stunault, 1992, Physica B 180181, 142.
Fischer, E, A. Furrer, E. Kaldis, D. Kim, J.K. Kjerns
and P.M. Levy, 1985, Phys. Rev. B 31, 456.
Fisher, R.A., R. Ballou, J.P. Emerson, E. Leli~vreBerna and N.E. Phillips, 1992, Int. J. Mod. Phys. B
7, 830.
Franse, J.J.M., and R.J. Radwanski, 1993, in:
Handbook of Magnetic Materials, Vol. 7, ed.
K.H.J. Buschow (North-Holland, Amsterdam) ch.
5, p. 307.
Franse, J.J.M., A.A. Menovsky, A. de Visser, C.D.
Bredl, U. Gottwich, W. Lieke, H.M. Mayer, U.
Rauchschwalbe, G. Sparn and E Steglich, 1985a,
Z. Phys. B 59, 15.
Franse, LJ.M., ER. de Boer, RH. Frings, R. Gersdoff,
A.A. Menovsky, EA. Muller, R.J. Radwanski and
S. Sinnema, 1985b, Phys. Rev. B 31, 4347.
Franse, J.J.M., N.R Thuy and N.M. Hong, 1988,
J. Magn. & Magn. Mater. 72, 361.
Franse, J.J;M., R.J. Radwanski and R. Verhoef, 1990,
J. Magn. & Magn. Mater. 81,299.
Freltoft, T., P. Brni, G. Shirane and K. Motoya, 1988,
Phys. Rev. B 37, 3454.
FrOmy,M.A., D. Gignoux, D. Schmitt and A. Takeuchi,
1989, J. Magn. & Magn. Mater. 82, 175.
Friedel, J., 1969, in: The Physics of Metals, 2. Electrons, ed. J.M. Zeeman (Cambridge University
Press, Cambridge) ch. 8.

420

D. GIGNOUX and D, SCHMITT

Fujii, H., and T. Shigeoka, 1990, J. Magn. & Magn.

Mater. 90-91, 115.
Fukauna, H., T. Komatsubara, T. Suzuki, E. Kaldis and
T. Kasuya, 1985, J. Phys. Soc. Jpn. 54, 3067.
Ganapathy, E.V., W.E. Wallace and R.S. Craig, 1974,
in: Proc. l lth Rare Earth Research Conf., eds
J.M. Hoschke and H.A. Eick (USAEC, Oak Ridge,
TN) p. 430.
Gibbs, D., D.E. Moncton, K.L. D'Amico, J. Bohr and
B.H. Grier, 1985, Phys. Rev. Lett. 55, 234.
Gibbs, D., J. Bohr, J.D. Axe, D.E. Moncton and K.L.
D'Amico, 1986, Phys. Rev. B 34, 8182.
Gignoux, D., and R. Lemaire, 1974, Solid State
Commun. 14, 877.
Gignoux, D., and J.J. Rhyne, 1986, J. Magn. & Magn.
Mater. 54-57, 1179.
Gignoux, D., and D. Schmitt, 1991, J. Magn. & Magn.
Mater. 100, 99.
Gignoux, D., and D. Schmitt, 1993, Phys. Rev. B 48,
12682.
Gignoux, D., E Givord and R. Lemaire, 1975, Phys.
Rev. B 9, 3878.
G-ignoux, D., E Givord and W.C. Koehler, 1977,
Physiea B 86-88, 165.
Gignoux, D., A. Nait-Saada and R. Perrier de la Bathie,
1979, J. Phys. (Paris) 40, C5-188.
Gignoux, D., R. Lemaire and P. Molho, 1980a, J.
Magn. & Magn. Mater. 21, 119.
Gignoux, D., R. Lemaire, P. Molho and E Tasset,
1980b, J. Magn. & Magn. Mater. 21, 307.
Gignoux, D., E Givord, R. Lemaire and E Tasset,
1983, J. Less-Common Met. 94, 1.
Gignoux, D., R. Lemaire, R. Mendia-Monterroso, J.M.
Moreau and J. Schweizer, 1985, Physiea B 130,
376.
Gignoux, D., D. Schmitt and M. Zerguine, 1987,
J. Magn. & Magn. Mater. 66, 373.
Gignoux, D., D. Schmitt and M. Zerguine, 1988,
J. Phys. (Paris) 49, C8-433.
Gignoux, D., D. Schmitt, A. Takeuchi and EY. Zhang,
1991a, J. Magn. & Magn. Mater. 97, 15.
Gignoux, D., D. Schmitt, T. Takeuchi, EY. Zhang, C.
Rouchon and E. Roudaut, 1991b, J. Magn. & Magn.
Mater. 98, 333.
Gignoux, D,, R Morin and D. Schmitt, 1991c, J. Magn.
& Magn. Mater. 102, 33.
Gignoux, D., E Morin, J. Voiron and P. Burlet, 1992,
Phys. Rev. B 46, 8877.
Gignoux, D., D. Schmitt, J. Voiron, EY. Zhang,
E. Bauer and G. Schandy, 1993, J. Alloys &
Compounds 191, 139.

Givord, D., L Laforest and R. Lemaire, 1977, Physica

B 86--88, 204.
Givord, D., J. Laforest and R. Lemaire, 1979, J. Appl.
Phys. 50, 7489.
Givord, D., J. Laforest, R. Lemaire and Q. Lu, 1983,
J. Magn. & Magn. Mater. 31-34, 191.
Givord, E, and LS. Shah, 1972, Comptes Rendus 274,
923.
Goremychkin, E.A., E. Mtihle, P.G. Ivanitskii, V.T.
Krotenko, M.V. Pasechnik, V.V. Slisenko, A.A.
Vasilkevich, B. Lippold, O.D. Chistyakov and E.M.
Savitskii, 1984, Phys. Status Solidi b 121, 623.
Goremychldn, E.A., E. Mtihle, I. Natkaniec, M.
Popescu and O.D. Christyakov, 1985a, Sov. Phys.
Solid State 27, 1195.
Goremychkin, E.A., E. Miihle, B. Lippold, O.D.
Christyakov and E.M. Savitskii, 1985b, Phys. Status
Solidi b 127, 371.
Goto, T., T, Sakakibara, K. Murata, H. Komatsu and K.
Fukamichi, 1990, J. Magn. & Magn. Mater. 90&91,
700.
Gubbens, EC.M., A.M. Van der Kraan and K.H.J.
Buschow, 1988, J. Magn. & Magn. Mater. 76-77,
187.
Hashimoto, Y., 1988, 1. Phys. (Paris) 49, C8-421.
Hashimoto, Y., H. Fujii, H. Fujiwara and T. Okamoto,
1979, J. Phys. Soc. Jpn. 47, 67.
Herbst, J.H., J.J. Croat, R.W. Lee and W.B. Yelon,
1982, J. Appl. Phys. 53, 250.
Hirosawa, S., and Y. Nakamura, 1982, J. Magn. &
Magn. Mater. 25, 284.
Hulliger, E, 1978, J. Magn. & Magn. Mater. 8, 183.
Hulliger, F., 1980, J. Magn. & Magn. Mater. 15--18,
1243.
Hutehings, M.T., 1964, Solid State Physics, Vol. 16
(Academic Press, New York) p. 227.
Ikeda, K., K.A. Gschneidner Jr, R.J. Stierman, T.-W.E.
Tsang and O.D. McMasters, 1984, Phys, Rev. B. 29,
5039.
Ikeda, K,, S.K. Dhar, M. Yoshizawa and K.A.
Gschneidner Jr, 1991, J. Magn. & Magn. Mater.
100, 292.
Ivanov, V., L. Vinokurova and A. Szytuta, 1992a, J.
Magn. & Magn. Mater. 110, L259.
Ivanov, V., L. Vinokurova, A. Szytuta and A. Zygmunt,
1992b, J. Magn. & Magn. Mater. 114, L225.
Iwata, N., 1990, J. Magn. & Magn. Mater. 86, 225.
Iwata, N., T. Kimura, T. Shigeoka and Y. Hashimoto,
1987, J. Magn. & Magn. Mater. 70, 279.
Iwata, N., T. Kimura, T. Shigeoka and Y. Hashimoto,
1988, J. Magn. & Magn. Mater. 73, 79.

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

Iwata, N., Y. Hashimoto, T. Kimura and T. Shigeoka,
1989, J. Magn. & Magn. Mater. 81, 354.
Iwata, N., K. Honda, T. Shigeoka, Y. Hashimoto and
H. Fujii, 1990, J. Magn. & Magn. Mater. 90-91,
63.
Iwata, N., M. Nishikawa and T. Shigeoka, 1991,
J. Magn. & Magn. Mater. 99, 209.
Janak, J.E, and A.R. Williams, 1976, Phys. Rev. B 14,
4199.
Jaussaud, C., P. Morin and D. Schmitt, 1980, J. Magn.
& Magn. Mater. 22, 98.
Jehan, D.A., D.E McMorrow, R.A. Cowley and G.J.
Melntyre, 1992, L Magn. & Magn. Mater. 104.107,
1523.
Jensen, J., and A.R. Mackintosh, 1990, Phys. Rev.
Lett. 64, 2699.
Jansen, J., and A.R. Mackintosh, 1992, J. Magn. &
Magn. Mater. 104-107, 1481.
Je~jini, M., M. Bonnet, P. Burlet, G. Lapertot, J.
Rossat-Mignod, J.Y. Henry and D. Gignoux, 1988,
J. Magn. & Magn. Mater. 76-77, 405.
Kakol, Z., H. Figiel and K. Turek, 1984, IEEE Trans.
Mug. MAG-20, 1605.
Ka/vius, G.M., 1994, Hyp. Int. 84, 249.
Kamimura, H., J. Sakurai, Y. Komura, H. Nakamura,
M. Shiga and Y. Nakamura, 1987, J. Magn. & Magn.
Mater. 70, 145.
Kaneko, T., S. Abe, S. Sakurada, H. Yoshida, G. Kido
and Y. Nakagawa, 1992a, J. Magn. & Magn. Mater.
104.107, 1401.
Kaneko, T., T. Kitai, S. Abe, M. Ohashi and Y.
Nakagawa, 1992b, Physica B 177, 295.
Kasuya, T., 1956, Progr. Theor. Phys. 16, 45.
Kim, D., and P.M. Levy, 1982, J. Magn. & Magn.
Mater. 27, 257.
Kirchmayr, H.R., and C.A. Poldy, 1979, in: Handbook
on the Physics and Chemistry of Rare Earths, Vol.
2, eds K.A. Gschneidner Jr and L. Eyring (NorthHolland, Amsterdam) oh. 14, p. 55.
Kitai, T., T. Kaneko, S. Abe, S. Tomiyoshi and Y.
Nakagawa, 1990, L Magn. & Magn. Mater. 90-91,
55.
Klein, H.P., A. Menth and R.S. Perkins, 1975, Physiea
B 80, 153.
Landau, L., and E. Lifshitz, 1981, Statistical Physics
(Pergamon Press, Oxford).
Lavagna, M., C. Lacroix and M. Cyrot, 1983, J. Phys.
F 13, 1007.
Lebech, B., Z. Smetana and V. Sima, 1987, J. Magn.
& Magn. Mater. 70, 97.

421

Lee, E.W., and E Pourarian, 1976, Phys. Status Solidi

a 33, 483.
Lemaire, R., 1966, Cobalt 32, 132.
Li, H.S., and J.M.D. Coey, 1991, in: Handbook of
Magnetic Materials, Vol. 6, ed. K.H.J. Buschow
(North-Holland, Amsterdam) ch. 1, p. 1.
Lin, H., M.E Collins, T.M. Holden and W. Wei, 1992,
J. Magn. & Magn. Mater. 104-107, 1551.
Liu, W.L., M. Kurisu, H. Kadomatsu and H. Fujiwara,
1986, J. Phys. Soc. Jpn. 55, 33.
Luong, N.H., J.J.M. Franse and TD. Hien, 1985,
J. Phys. F 15, 1751.
Liithi, B., R. Sommer and R Morin, 1979, J. Magn. &
Magn. Mater. 13, 198.
Mackintosh, A.R., and J. Jensen, 1992, Physica B
180-181, 1.
McAlister, S.E, 1984, J. Phys. F 14, 2167.
MeEwen, K.A., and S.W. Zochowski, 1990, J. Magn.
& Magn. Mater. 90-91, 94.
McEwen, K.A., G.J. Cock, L.W. Roeland and A.R.
Mackintosh, 1973, Phys. Rev. Lett. 30, 287.
Minakata, R., M. Shiga and Y. Nakamura, 1976,
J. Phys. Soc. Jpn. 41, 1435.
Misawa, S., 1988, Physica B 149, 162.
Mitsuda, S., H. Yoshizawa, N. Wada, Y. (3nuki and I".
Komatsubara, 1991, J. Phys. Soe. Jpn. 60, 1428.
Mondal, S., S.I-I. Kllcoyne, R. Cywinski, B.D.
Rainford and C. Ritter, 1992, J. Magn. & Magn.
Mater. 104-107, 1421.
Morin, R, 1988, J. Magn. & Magn. Mater. 71, 151.
Morin, R, 1993, private communication.
Morin, R, and J. Pierre, 1975, Phys. Status Solidi a
30, 549.
Morin, R, and D. Schmitt, 1978, J. Phys. F 8, 951.
Morin, P., and D. Sehmitt, 1982, Phys. Rev. B 26,
3891.
Morin, R, and D. Schmitt, 1990, in: Ferromagnetic
Materials, Vol. 5, eds K.H.J. Buschow and
E.P. Wohlfarth (North-Holland, Amsterdam) ch. 1,
p. 1.
Morin, P., J. Rouehy and D. Schrnitt, 1978, Phys. Rev.
B 17, 3684.
Morin, P., D. Schmitt and E. de Lacheisserie, 1982,
J. Magn. & Magn. Mater. 30, 257.
Morin, P., J. Rouchy, M. Giraud and A. Czopnik,
1987a, J. Magn. & Magn. Mater. 67, 95,
Morin, R, M. Giraud, P. Burlet and A. Czopnik, 1987b,
J. Magn. & Magn. Mater. 68, 107.
Morin, P., J. Rouchy and D. Sctunitt, 1988, Phys. Rev.
B 37, 5401.

422

D. GIGNOUX and D. SCHMITT

Morin, R, J. Rouchy, K. Yonenobu, A. Yamagishi and

M. Date, 1989, J. Magn. & Magn. Mater. 81,247.
Morin, E, J. Rouchy, D. Schmitt and E. de Laeheisserie,
1990, J. Magn. & Magn. Mater. 90-91, 105.
Morin, E, S. Kunii and T. Kasuya, 1991, J. Magn. &
Magn, Mater. 96, 145.
Moriya, T., 1991, J. Magn. & Magn. Mater. 100, 261.
Moriya, T., and H. Hasegawa, 1980, J. Phys. Soc. Jpn.
48, 1490.
Moriya, T., and Y. Takahashi, 1978, J. Phys. Soc. Jpn.
45, 397.
Motokawa, M., H. Nojiri and 5(. Endoh, 1992, Physica
B 177, 279.
Motoya, K., T. Freltoft, R B6ni and G. Shirane, 1988,
Phys. Rev. B 38, 4796.
Mrachkov, J., and E. Leyarovski, 1988, Physica B 150,
404.
Mttraoka, Y., M. Shiga and Y. Nakamura, 1983,
J. Magn. & Magn. Mater. 31-34, 121.
Muraoka, Y., H. Okuda, M. Shiga and Y. Nakarnura,
1984, J. Phys. Soc. Jpn. 53, 331.
Murata, K.K., and S. Doniach, 1972, Phys. Rev. Lett.
29, 285.
Nakamura, H., Y. Kitaoka, K. Yoshimura, Y. Kohori,
K. Asayama, M. Shiga and Y. Nakamura, 1988a, J.
Phys. (Pads) Colloq. 49, C8-257.
Nakamura, H., H. Wada, K. Yoshimura, M. Shiga,
Y. Nakamura, J. Sakurai and Y. Komura, 1988b, J.
Phys. F 18, 981.
Nakamura, Y., 1983, J. Magn. & Magn. Mater. 31-34,
829.
Nakamura, Y., M. Shiga and S. Kawano, 1983, Physica
B 120, 212.
Nimori, S., G. KJdo, D.X. Li, Y. Haga and T.
Suzuki, 1995, J. Magn. & Magn. Mater. 140-144,
to appear.
Nojiri, H., M. Uchi, S. Watamura, M. Motokawa, H.
Kawai, Y. Endoh and T. Shigeoka, 1991, J. Phys.
Soc. Jpn. 60, 2380.
NordstrOm, L., M.S.S. Brooks and B. Johansson, 1992,
J. Magn. & Magn. Mater. 104-107, 1942.
Nozaki, H., T. Tanaka and Y. Ishizawa, 1980, J. Phys.
C 13, 2751.
Nunez-Regueiro, M.D., C. Lacroix and R. Ballou,
1992a, Phys. Rev. B 46, 990.
Nunez-Regueiro, M.D., C. Lacroix, R. Ballou and E.
Leli~vre-Bema, 1992b, J. Magn. & Magn. Mater.
104-107, 285.
Oddou, J.L., C. Jeandey, R. Ballou, J. D6portes and B.
Ouladdiaf, 1993, Solid State Commun. 85, 419.
Ohkawa, EJ., 1989, Solid State Commun. 71,907.

Ohkoshi, M., H. Kobayashi, T. Katayama, M. Hirano

and T. Tsushima, 1976, AlP Conf. Proc. 29, 616.
Ohkoshi, M., H. Kobayashi, T. Katayama, M. Hirano
and T. Tsushima, 1977, Physica 86--88B, 195.
Okamoto, T., H. Fujii, Y. Andoh and H. Fujiwara,
1985, J. Magn. & Magn. Mater. 52, 208.
()nuki, Y., K. Ina, M. Nishihara, T. Komatsubara, S.
Takayanagi, K. Kameda and N. Wada, 1986, J. Phys.
Soc. Jpn. 55, 1818.
Oomi, G., T. Terada, M. Shiga and Y. Nakamura, 1987,
J. Magn. & Magn. Mater. 70, 137.
Ouladdiaf, B., 1986, Thesis (University of Grenoble)
unpublished.
Peeharsky, V.K., K.A. Gsehneidner Jr and D. Fort,
1993, Phys. Rev. B 47, 5063.
Perkins, R.S., and H. Nagel, 1975, Physica B 80, 143.
Perkins, R.S., and S. Str~issler, 1977, Phys. Rev. B 15,
477.
Petrieh, G., and R.L. M6ssbauer, 1968, Phys. Lett. A
26, 403.
Pines, D., and P. Nozi~res, 1966, The Theory of
Quantum Fluids, Vol. I (Benjamin, New York).
Pinettes, C., and C. Lacroix, 1993a, J. Magn. & Magn.
Mater. 123, 153.
Pinettes, C., and C. Lacroix, 1993b, Solid State
Commun. 85, 565.
Pszczola, J., and K. Krop, 1986, J. Magn. & Magn.
Mater. 59, 95.
Purwins, H.G., E. Walker, B. Barbara, M. Rossignol
and A. Furrer, 1976, J. Phys. C 9, 1025.
Radwanski, R.J., 1986, Physica B 142, 57.
Radwanski, R.J., J.J.M. Franse and S. Sinnema, 1985,
J. Phys. F 15, 969.
Radwanski, R.J., J.J.M. Franse and R. Verhoef, 1990,
J. Magn. & Magn. Mater. 83, 127.
Radwanski, R.J., J.J.M. Franse, EH. Quang and EE.
Kayzel, 1992a, J. Magn. & Magn. Mater. 104-107,
1321.
Radwanski, R.J., N.H. Kim-Ngan, EE. Kayzel, J.J.M.
Franse, D. Gignoux, D. Schmitt and F.Y. Zhang,
1992b, J. Phys.: Condens. Matter 4, 8853.
Raffius, G., and J. K6tzler, 1983, Phys. Lett. A 93,
423.
Reiffers, M., Y.G. Naidyuk, A.G.M. Jansen, P. Wyder,
I.K. Yanson, D. Gignoux and D. Schmitt, 1989,
Phys. Rev. Lett. 62, 1560.
Ritter, C., S.H. Kilcoyne and R. Cywinski, 1991,
J. Phys.: Condens. Matter 3, 727.
Rossat-Mignod, J., 1986, in: Neutron Scattering in
Condensed Matter Research, eds K. Skold and
D.L. Price (Academic Press, New York).

MAGNETIC PROPERTIES OF INTERMETALLIC COMPOUNDS

423

Rossat-Mignod, J., E Burlet, S. Quezel, J.M. Effantin,

Stmat, K.J., 1988, in: Ferromagnetic Materials, Vol.
D. Delacote, H. Bartholin, O. Vogt and D. Ravot,
4, eds E.P. Wohlfarth and K.H.J. Buschow (North1983, J. Magn. & Magn. Mater. 31-34, 398.
Holland, Amsterdam) ch. 2, p. 131.
Ruderman, M.A., and C. Kittel, 1954, Phys. Rev. 96, Stryjewski, E., and N. Cfiordano, 1977, Adv. Phys. 26,
487.
99.
Sakakibara, T., T. Goto, K. Yoshimura, M. Shiga, Y. Sugiura, E., K. Sugiyama, H. Kawanaka, T. Takabatake, H. Fujii and M. Date, 1990, J. Magn. &
Nakamura and K. Fukamichi, 1987, J. Magn. &
Magn. Mater. 90-91, 65.
Magn. Mater. 70, 126.
Sato, K., Y. Yosida, Y. Isikawa and K. Moil, 1986, Sugiyama, K., Y. Koyoshi, S. Kunii, T. Kasuya and
M. Date, 1988, J. Phys. Soc. Jpn. 57, 1762.
J. Magn. & Magn. Mater. 54-57, 467.
Sugiyama, K., K. Yonenobu, M. Date, T. Shigeoka
Schinkel, C., 1978, J. Phys. F 8, L87.
and H. Fujii, 1992, Physica B 177, 275.
Schmitt, D., 1986, J. Phys. (Paris) 47, 677.
Svoboda, P., M. Divis, A.V. Andreev, N.V. Baranov,
Schweizer, J., and E Tasset, 1980, J. Phys. F 10, 2799.
M.I. Bartashevich and P.E. Markin, 1992, J. Magn.
Selke, W., and RM. Duxbury, 1984, Z. Phys. B 57,
& Magn. Mater. 104-107, 1329.
49.
Szytuta,
A., R.J. Radwanski and ER. de Boer, 1992,
Shiga, M., 1981, J. Phys. Soc. Jpn. 50, 2573.
J. Magn. & Magn. Mater. 104-107, 1237.
Shiga, M., H. Wada, H. Nakamura, K. Yoshimura and
Takahashi, M., H. Tanaka, T. Satoh, M. Kohgi, Y.
Y. Nakamura, 1987, J. Phys. F 17, 1781.
Ishikawa, T. Miura and H. Takei, t988, J. Phys.
Shigeoka, T., N. Iwata, H. Fujii, T. Okamoto and Y.
Soc. Jpn. 57, 1377.
Hashimoto, 1987, J. Magn. & Magn. Mater. 70,
Takayanagi, S., Y. ()nuki, K. Ina, T. Komatsubara,
239.
N. Wada, T. Watanabe, T. Sakakibara and T. Goto,
Shigeoka, T., N. Iwata, Y. Hashimoto, Y. Andoh and
1989, J. Phys. Soe. Jpn. 58, 1031.
H. Fujii, 1988, J. Phys. (Paris) 49, C8-431.
Takayanagi,
S., N. Wada, Y. ()nuki and T. KomatsubShigeoka, I"., H. Fujii, K. Yonenobu, K. Sugiyama and
ara, 1991, J. Phys. Soc. Jpn. 60, 975.
M. Date, 1989, J. Phys. Soc. Jpn. 58, 394.
Shigeoka, T., M. Saeki, N. Iwata, T. Takabatake and Tanaka, Y., I". Tsuchida and Y. Nakamura, 1980,
J. Phys. Soc. Jpn. 48, 1092.
H. Fujii, 1990, J. Magn. & Magn. Mater. 90-91,
Thuy, N.P., and J.J.M. Franse, 1986, J. Magn. & Magn.
557.
Mater. 54-57, 915.
Shigeoka, T., H. Fujii, M. Nishi, Y. Uwatoko, T.
Thuy, N.P., J.J~M. Franse, N.M. Hong and R.D. Hien,
Takabatake, I. Oguro, K. Motoya, N. Iwata and
1988, J. Phys. (Paris) 49, C8-499.
Y. Ito, 1992a, J. Phys. Soc. Jpn. 61, 4559.
Tomiyama, F., M. Ono, M. Date, A. Yamagishi, R.
Shigeoka, T., S. Kawano, N. Iwata and H. Fujii, 1992b,
Verhoef, F.R. de Boer, LJ.M. Franse and X.P. Zhong,
Physica B 180-181, 82.
1991, J. Appl. Phys. 69, 5539.
Shigeoka, T., A. Gamier, D. Gignoux, D. Schmitt, EY.
Tsuneto, T., and T. Mumo, 1971, Physiea 51, 186.
Zhang and R Buffet, 1994, J. Phys. Soc. Jpn. 63,
Ueda, K., and T Moriya, 1975, J. Phys. Soc. Jpn. 39,
2731.
605.
Shimizu, K., 1985, J. Phys. Soc. Jpn. 54, 1154.
Uhl, M., L. Sandratskii and J. KObler, I993, Int. J.
Shimizu, M., J. Inoue and S. Nagasawa, 1984, J. Phys.
Mod. Phys. 7, 1027.
F 14, 2673.
Van der Goot, A.S., and K.H.J. Buschow, 1971, Phys.
Shohata, N., 1977, J. Phys. Soe. Jpn. 42, 1873.
Status Solidi a 5, 665.
Sinnema, S., R.J. Radwanski, JJ.M. Franse, D.B. de Verhoef, R., F.R. de Boer, J.J.M. Franse, C.J.M.
Mooij and K.H.J. Buschow, 1984, J. Magn. & Magn.
Denissen, T.H. Jacobs and K.H.J. Busehow, 1989,
Mater. 44, 333.
J. Magn. & Magn. Mater. 80, 41.
Sinnema, S., J.J.M. Franse, A.A. Menovsky and R.J. Verhoef, R., R.J. Radwanski and JJ.M. Franse, 1990,
Radwanski, 1986, J. Magn. & Magn. Mater. 54-57,
J. Magn. & Magn. Mater. 89, 176.
1639.
Verhoef, R., S. Sinnema, P.H. Quang and JJ.M. Franse,
Sinnema, S., J.J.M. Franse, R.J. Radwanski, A.A.
1992, J. Magn. & Magn. Mater. 104-107, 1325.
Menovsky and ER. de Boer, 1987, J. Phys. F 17, Vinokurova, L., V. Ivanov and A. Szytuta, 1991,
233.
J. Magn. & Magn. Mater. 99, 193.

424

D. GIGNOUX and D. SCHMITT

Voiron, J., and D. Bloch, 1971, J. Phys. (Paris) 32,

949.
Voiron, J., J. Beille, D. Bloch and C. Vettier, 1973,
Solid State Commun. 13, 201.
Voiron, J., R. Ballou, J. D6portes, R.M. Gal~ra and E.
Leli~vre-Bema, 1990, J. Appl. Phys. 69, 5678.
Wada, H., H. Nakamura, K. Yoshimura, M. Shiga and
Y. Nakamura, 1987a, J. Magn. & Magn. Mater. 70,
134.
Wada, H., H. Nakamura, E. Fukami, K. Yoshimura, M.
Shiga and Y. Nakamura, 1987b, J. Magn. & Magn.
Mater. 70, 17.
Wada, H., K. Yoshimura, G. Kido, M. Shiga, M.
Mekata and Y. Nakamura, 1988, Solid State
Commun. 65, 23.
Waind, ER., I,S. MacKenzie and M.A.H. McCausland,
1983, J. Phys. F 13, 1041.
Weber, K., 1977, Physiea B 86--88, 91.
Weinert, M., and A.J. Freeman, 1983, J. Magn. &
Magn. Mater. 38, 23.
Wohlfarth, E.E, 1951, Philos. Mag. 42, 374.
Wohlfarth, E.E, and E Rhodes, 1962, Philos. Mag. 7,
1817.
Yamada, H., and M. Shimizu, 1985, J. Phys. F 15,
L175.
Yamada, H., and M. Shimizu, 1988, J. Phys. (Pads)
49, C8-247.
Yamada, H., and M. Shimizu, 1990, J. Magn. & Magn.
Mater. 90&91, 703.

Yamada, H., J. Inoue, K. Terao, S. Kanda and M.

Shimizu, 1984, J. Phys. F 14, 1943.
Yamada, H., J. Inoue and M. Shimizu, 1985, J. Phys.
F 15, 169.
Yamada, H., T. Tohyama and M. Shimizu, 1987,
J. Magn. & Magn. Mater. 70, 44.
Yamashita, M., M. Kurisu, H. Kadomatsu, I. Oguro
and H. Fujiwara, 1987, J. Phys. Soe. Jpn. 56, 32.
Yoshie, H., M. Shiga and Y. Nakamura, 1985, J. Phys.
Soe. Jpn 54, 1116.
Yoshimura, K., and Y. Nakamura, 1984, J. Phys. Soe.
Jpn, 53, 3611.
Yoshimura, K., T. Shimizu, M. Takigawa, H. Yosuoka
and Y. Nakamttra, 1984, J. Phys. Soe. Jpn. 53, 503.
Yoshimura, K., M. Shiga and Y. Nakamura, 1986a, J.
Phys. Soc. Jpn. 55, 3585.
Yoshimura, K., M. Takigawa, H. Yasuoka, M. Shiga
and Y. Nakamura, 1986b, J. Magn. & Magn. Mater.
54-57, 1075.
Yoshimura, K., Y. Yoshimoto, M. Mekata, T.
Sakakibara and T. Goto, 1987, J. Magn. & Magn.
Mater. 70, 147.
Yosida, K., 1957, Phys. Rev. 106, 893.
Zhang, EY., 1992, Thesis (University of Grenoble)
unpublished.
Zhang, EY., D. Gignoux, D. Schmitt, J.J.M. Franse,
EE. Kayzel, N.H. Kim-Ngan and R.J. Radwanski,
1994, J. Magn. & Magn. Mater. 130, 108.

AUTHOR INDEX
Aarts, J., s e e Steglich, F. 6
Abe, S. 388
Abe, S., s e e Kaneko, T. 388, 397
Abe, S., s e e Kitai, T. 388
Abell, J.S., s e e del Moral, A. 355
Abeln, A. 212, 229, 252, 271,273, 276, 277
Abeln-Heidel, A., s e e Arons, R.R. 271,273
Aehard, J.C., s e e Rossat-Mignod, J. 62
Adachi, G. 140, 170-172
Adachi, G., s e e Arakawa, T. 144
Adachi, G., s e e Sakaguchi, H. 169-172
Adaehi, G., s e e Shirai, H. 172
Adams, RE, s e e Hubberstey, P. 140
Affrossman, S. 143
Affrossman, S., s e e Gimzewski, J.K. 137
Aguilera-Granja, E 122
Ahmed-Mokhtar, N. 127
Ahn, C.C., s e e Johnson, R.W. 178
Aidun, J.B. 132
Aknetey, A . , s e e McMinn, R. 134
Aksela, H., s e e Chorkendorff, I. 114
Aksela, S., s e e Chorkendorff, I. 114
A1-Bassam, T.S., s e e Hussain, A . A . A .
134
Alameda, J.M. 145, 311,312
Alameda, J.M., s e e Dieny, B. 165
Albert, L., s e e Gasgnier, M. 132
Albessard, A.K., s e e Aoki, H. 76-78
Albessard, A.K., s e e Ebihara, T. 42, 44-46, 48,
49, 51
Albessard, A.K., s e e Onuki, Y. 74-78, 80
Aldred, A.T., s e e Friedt, J.M. 282
Alekseev, P.A. 339, 341
Al~onard, R. 355, 357, 394, 395, 397
Aliotta, C.E, s e e Ronay, M. 185
Allain, Y., s e e Daou, J.N. 216, 268, 269
Allay, L., s e e Bosca, G. 132
Allen, C.W. 174
Allen, J.W. 115
Allen, J.W., s e e Heeht, M.H. 131
Allen, J.W., s e e Johansson, L.I. 115, 131
Allen Jr, S.J. 186, 187
Allen Jr, S.J., s e e Palmstrom, C.J. 186
Allenspach, R., s e e Landolt, M. 124
Allenspach, R., s e e Taburelli, M. 124
Allibert, C.H. 153

Alvarado, S.E 131

Alvarado, S.E, s e e Weller, D. 122-124
Amara, M. 398
Amaral, V.S., s e e Freitas, P.P. 185
Amato, A. 94, 95
Amato, A., s e e Gygax, EN. 232
Amato, A., s e e Rossat-Mignod, J. 81
Ambroth, K.E., s e e Krost, A. 188, 189
Andersen J.N. 115, 116
Andersen J.N., s e e Nilsson, A. 116
Andersen J.N., s e e Stenborg, A. 117
Andersen O.K. 12
Anderson I.S. 216, 233, 234
Anderson I.S., s e e Berk, N.E 232
Anderson I.S., s e e Carmelli, G. 232
Anderson I.S., s e e Gygax, EN. 232
Anderson I.S., s e e Leisure, R.G. 232
Anderson I.S., s e e Udovie, T.J. 234, 235, 243
Anderson O.K., s e e Jepson, J. 22
Anderson P.W. 3
Andoh, Y. 361
Andoh, Y., s e e Okamoto, T. 360, 361
Andoh, Y., s e e Shigeoka, T. 382
Andr6, G. 227, 230, 281
Andr6, G., s e e Vajda, P. 259, 275, 281,282
Andreeff, A., s e e Alekseev, P.A. 339, 341
Andreev, A.V., s e e Svoboda, P. 387
Andres, K. 339, 341
Andrews, B., s e e Li, D. 123
Angadi, M.A. 133
Angadi, M.A., s e e Ashrit, EV. 133
Annapoorni, S. 168
Antepenko, R.J. 142
Antepenko, R.J., s e e Holloway, D.M. 142
Anthony, T., s e e Coulman, D. 166
Anthony, T.C. 168
Antomangeli, F., s e e Sigrist, M. 126, 128
Aoki, H. 63~55, 76-78
Aoki, H., s e e Crabtree, G.W. 62-64
Aoyama, A. 166
Apai, G., s e e Mason, M.G. 114
Apostolov, A.V 126, 127
Arakawa, T. 144
Arbman, G.O., s e e Koelling, D.D. 12
Arko, A.J. 31, 33
425

426

AUTHOR INDEX

Arons, R.R. 210, 267, 271,273, 277-279, 281

Arons, R.R., s e e Bohn, H.G. 277
Asano, A., s e e Kubo, Y. 14, 36, 37
Asano, H. 91
Asayama, K., s e e Nakamura, H. 319
Ashrit, P.V. 133
Ashrit, EV., s e e Angadi, M.A. 133
Aso, K . , s e e Hashimoto, S. 168
Assmus, W., s e e Hunt, M. 81, 82
Asti, G. 417
Atzrnony, U., s e e Venkert, A. 179
Aubert, G. 339
Awano, H., s e e Katayama, T. 166
Axe, J.D., s e e Gibbs, D. 268, 378, 417
Aylesworth, K.D. 146, 150, 151
Aylesworth, K.D., s e e Shan, Z.S. 158
Aylesworth, K.D., s e e Strzeszewski, J. 151
Ayres de Campos, N., s e e Ferreira, P. 260
Baba, K. 188
Baba, K., s e e Nakamura, O. 118, 125
Baberschke, K. 122
Baberschke, K., s e e Farle, M. 122, 123
Babldn, G.V., s e e Gonchar, EM. 118
Babkin, G.V., s e e Lozovyi, Ya.B. 119
Bacburin, V.I. 125
Bacon, EM. 141
Baczewski, L.T. 159, 161
Baczewski, L.T., s e e Piecuch, M. 159, 161
Badia, E 157
Badia, E, s e e Ferrater, C. 158
Baenziger, C., s e e Eick, H.A. 110
Baer, Y., s e e Lang, J.K. 115
Baer, Y., s e e Moser, H.R. 130-132
Baihe, M., s e e Zhao, Z.B. 152
Baillif, P., s e e Gasgnier, M. 109
Bak, R 399, 400
Bak, P., s e e Barbara, B. 360
Ball, A.R. 342, 351-353, 365, 366, 375, 377,
378, 380, 386, 391,392, 402-406, 417
Ballou, R. 308, 310, 313, 317, 319, 320, 322,
324-327, 336, 339
Ballou, R., s e e Brown, P.J. 321
Ballou, R., s e e Fisher, R.A. 324
Ballou, R., s e e Nunez-Regueiro, M.D. 325
Ballou, R., s e e Oddou, J.L. 321
Ballou, R., s e e Voiron, J. 321
Bannai, E., s e e Ishizawa, Y. 31-33, 35
Barandiaran, J.M. 366
Baranov, N.V.., s e e Svoboda, R 387
Barbara, B. 67, 299, 360
Barbara, B., s e e Gratz, E. 87
Barbara, B., s e e Kaindl, G. 117
Barbara, B., s e e Lethuillier, R 48

Barbara, B., s e e Purwins, H.G. 360

Barbara, B., s e e von Molnar, S. 178, 180
Bardolle, J., s e e Gasgnier, M. 109
Bareham, H., s e e Corner, W.D. 126
Barnes, R.G. 216
Barnes, R.G., s e e Barnfather, K.J. 237
Barnes, R.G., s e e Belhoul, M. 277
Barnes, R.G., s e e Borsa, E 237
Barnes, R.G., s e e Han, J.W. 233
Barnes, R.G., s e e Klavins, P. 215, 221,225, 226,
229
Barnes, R.G., s e e Lichty, L.R. 232, 233,244
Barnes, R.G., s e e Phua, T.T. 237
Barnes, R.G., s e e Shinar, J. 251,253, 255
Barnes, R.G., s e e Torgeson, D.R. 233
Barnes, R.G., s e e Udovic, T.J. 234
Barnes, R.G., s e e Zamir, D. 256
Barnfather, K.J. 237
Barski, A., s e e Rossi, G. 130
Bartashevich, M.I., s e e Svoboda, E 387
Barth, J., s e e Gerken, F. 115, 131
Barth, J., s e e Johansson, L.I. 181
Barth, J., s e e Kammerer, R. 115
Barthem, VM.T.S. 339 341,343,344
Barthem, V.M.T.S., s e e Ballou, R. 339
Bartholin, H., s e e Effantin, J.M. 31,359, 397
Bartholin, H., s e e Rossat-Mignod, J. 62, 401
Bashev, V.E, s e e Tkach, V.I. 110
Bashkin, I.O. 225, 231
Bastow, T.J.M., s e e Barnes, R.G. 216
Bates, S., s e e Cowley, R.A. 417
Bauer, E. 128
Bauer, E., s e e Gignoux, D. 375, 377, 415
Bauer, E., s e e Gratz, E. 87
Bauer, E., s e e Kolaczkiewicz, J. 122
Bauer, E., s e e Stenborg, A. 116, 117
Bauer, G., s e e Krost, A. 188, 189
Bauer, R.S., s e e Allen, J.W. 115
Bauman, E, s e e Haussler, P. 118
Bazan, C., s e e Czopnik, A. 398
B6al-Monod, M.T. 304, 305
Beatfie, A.G. 241
Beaudry, B.J. 112, 217, 219, 234, 245
Beaudry, B.J., s e e Borsa, E 237
Beaudry, B.J., s e e Corner, W.D. 126
Beaudry, B.J., s e e Gschneidner Jr, K.A. 112
Beaudry, B.J., s e e Ito, T. 229, 268, 269
Beaudry, B.J., s e e Kai, K. 229, 241,252, 256
Beaudry, B.J., s e e Khatamian, D. 216, 219
Beaudry, B.J., s e e Klavins, P. 215, 221,225, 226,
229
Beaudry, B.J., s e e Phua, T.T. 237
Beaudry, B.J., s e e Saw, C.K. 216, 219
Beaudry, B.J., s e e Shinar, J. 251,253, 255, 257

AUTHOR INDEX
Beaudry, B.J., s e e Vajda, R 220, 236, 243, 245
Beaudry, B.J., s e e V611d, J. 233
Beavis, L.C. 142
Becker, M.E, s e e Choe, G. 163
B~cle, C. 367
Beille, J., s e e Voiron, J. 303
Belhoul, M. 277
Belhoul, M., s e e Phua, T.T. 237
Belorizky, E. 328-330, 337, 343-345
Benham, M.J., s e e Bennington, S.M. 234, 269
Bermington, S.M. 234, 269
Bennington, S.M., s e e Fairclough, J.EA. 219
Benoit, A. 52
Berendschot, T.T.J.M., s e e van der Meulen, H.E
81
Berger, L., s e e De Luca, J.C. 180
Berghaus, A., s e e Farle, M. 123
Berk, N. 304
Berk, N.E 232
Berk, N.E, s e e Anderson, I.S. 234
Berk, N.E, s e e Udovic, T.J. 234, 235
Bertel, E., s e e Netzer, EP. 108, I 11, 143, 144
Berthier, Y. 313
Besenicar, S., s e e Hole, J. 150
Besnus, M.J. 74
Bieganski, Z. 229, 282-284
Bieganski, Z., s e e Drulis, M. 241,252
Bieganski, Z., s e e Opyrchal, J. 278
Biget, M., s e e Daou, J.N. 219
Birecki, H., s e e Anthony, T.C. 168
Bischof, R. 214, 217, 271,273, 279
Bischoff, E., s e e Knoch, K.G. 152
Bittner, H., s e e Oesterreicher, H. 137
Biyadi, K., s e e Bras, J. 147
Bj6rneholm, O., s e e Stenborg, A. 117
Blancard, C. 112, 114
Blancard, C., s e e Gasgnier, M. 132
Blancard, C., s e e Sarpal, B.K. 112
Blaneo, J.A. 365, 370-373, 382, 383, 401-406
Blasehko, O. 215, 219, 220, 234
Blasehko, 0 . , s e e Andr6, G. 227, 230, 281
Blaschko, 0 . , s e e Pleschiutschnig, J. 234
Blaschko, O, s e e Udovie, T.J. 234, 235
Bloch, D. 300, 301,303
Bloch, D., s e e Voiron, J. 303
Block, J.H., s e e Melmed, A.J. 118, 119
Boespflug, E.R, s e e Harris, J.M. 177
Boeva, O.A. 136, 143, 240
Boeva, O.A., s e e Zhavoronkova, K.N. 136
Boffa, G., s e e Taborelli, M. 124
Bohdziewicz, A. 184
Bohn, H.G. 277
Bohr, J. 417
Bohr, J., s e e Gibbs, D. 268, 378, 417

427

Boltich, E.B., s e e Cheng, S.E 155

B6ni, P., s e e Freltoft, T. 323
B6ni, R, s e e Motoya, K. 323
Bonnet, J.E. 219, 225, 240
Bonnet, J.E., s e e Anderson, I.S. 216, 233
Bonnet, J.E., s e e Daou, J.N. 219, 245
Bonnet, J.E., s e e Schlapbach, L. 252, 257, 265,
266
Bonnet, M., s e e Jerjini, M. 70, 392
Borgensen, P. 180
Borkowska, W., s e e Drulis, M. 282
Boroch, E. 221,225, 226
Boroch, E., s e e Kaldis, E. 225
Borombaev, M.K. 364
Borsa, E 237
Bosca, G. 132
Boucherle, J.X., s e e Barbara, B. 67
Boucherle, LX., s e e Benoit, A. 52
Boukraa, A. 275, 282-285
Boukraa, A., s e e Ratishvili, I.G. 227
Boulesteix, C. 108
Boulet, R.M. 52
Bouten, RC.R, s e e Buschow, K.HJ. 108
Boutron, P. 340
Bouvier, M. 365, 405
Braaten, N.A. 178
Bracconi, R 137
Brag, J., s e e Anthony, T.C. 168
Braicovich, L., s e e Carbnne, C. 124, 158
Bran&, N.B. 5
Brar, N.S. 132
Bras, J. 147
Braun, H.E, s e e Kuboth, M. 185
Brearley, W., s e e Surplice, N.A. 143
Br~chignac, C. 114
Bredl, C.D. 70, 90
Bredl, C.D., s e e Franse, J.J.M. 325
Bredl, C.D., s e e Gratz, E. 87
Bredl, C.D., s e e Steglich, E 6
Brinkrnan, W.E 305
Briones, E, s e e Martinez, B. 146
Brooks, J.W. 182
Brooks, M.S.S., s e e Nordstr6m, L. 313, 314
Brouder, C. 159, 165
Brouder, C., s e e Guilmin, E 165
Brousseau, B., s e e Frandon, L 128-130
Brousseau-Lahaye, B. 128, 129
Brown, L.M., s e e Colliex, C. 132
Brown, L.M., s e e Gasgnier, M. 130
Brown, RJ. 321
Brown, RJ., s e e Ballou, R. 325, 327
Broyer, B., s e e Rayane, D. 114
Brun, T.O. 394
Bruno~ E 313

428

AUTHOR INDEX

Bruson, A., s e e Brouder, C. 159

Bryan, S.I. 179
Bucher, E., s e e Jayaraman, A. 7
Buevoz, J.L., s e e Barbara, B. 67
Burger, J.P. 229, 238, 243, 251,252, 257-261,
268, 271,273, 275-277, 283, 285
Burger, J.P., s e e Blaschko, O. 219, 234
Burger, J.P., s e e Daou, J.N. 213, 217, 219, 220,
225, 229, 233, 238, 239, 241,243-246, 248,
251,252, 255, 258, 262, 264, 267, 269, 271,
273, 277-279, 283
Burger, J.E, s e e Gupta, M. 243, 256
Burger, J.P., s e e Lucasson, A. 241,243, 251,252
Burger, J.P., s e e Schlapbach, L. 252, 257, 265,
266
Burger, J.P., s e e Senoussi, S. 271,273, 277, 279
Burger, J.P., s e e Shaltiel, D. 285
Burger, J.P., s e e Vajda, P. 213, 220, 227, 229,
230, 233, 236-238, 243-246, 251,256-259,
262, 267-271,273,275, 277-281
Buffet, P., s e e Amara, M. 398
Burlet, E, s e e Ball, A.R. 377, 378, 380, 417
Burlet, P., s e e Chattopadhyay, T. 401
Buffet, P., s e e Effantin, J.M. 31,359, 397
Burlet, P., s e e Gignoux, D. 389, 406
Buffet, P., s e e Jerjini, M. 70, 392
Burlet, P., s e e Morin, E 397-399
Burlet, P., s e e Rossat-Mignod, J. 62, 401
Buffet, P., s e e Shigeoka, T. 368
Burmistrova, O.P. 119
Burzo, E. 108, 144, 306
Burzo, E., s e e Bloch, D. 303
Buschow, K.H.J. 4, 108, 156, 177, 296, 321,328,
329, 338
Busehow, K.H.J., s e e Coene, W. 152
Buschow, K.H.J., s e e Gubbens, P.C.M. 339
Buschow, K.H.J., s e e Sirmema, S. 329
Buschow, K.H.J., s e e Van der Goot, A.S. 301
Busehow, K.H.J., s e e Verhoef, R. 331,332
Buxbaum, R.E. 141
Cable, J.W., s e e Arons, R.R. 273, 277
Cabuad, E, s e e Rayane, D. 114
Cadieu, EJ. 145, 155, 156
Cadieu, EJ., s e e Kamprath, N. 155
Cadieu, F.J., s e e Liu, N.C. 155
Cadieu, EJ., s e e Stadelmaier, H.H. 155
Cahuzae, Ph., s e e Br~chignac, C. 114
Callaway, J. 9
Camley, R.E. 159, 163
Camley, R.E., s e e LePage, J.G. 159
Campagna, M., s e e Alvarado, S.E 131
Campagna, M., s e e Hillebrecht, F.U. 63
Campagna, M., s e e Weller, D. 122, 124

Campbell, I.A. 329, 345

Cannelli, G. 232
Cantelli, R., s e e Cannelli, G. 232
Capellen, J., s e e Gschneidner Jr, K.A. 112
Carbone, C. 124, 158
Carlier, E, s e e Br6chignae, C. 114
Carlin, R.L. 271,275, 277, 281,282, 284, 285
Caro, P., s e e Dexpert-Ghys, J. 137
Caro, P., s e e Gasgnier, M. 132, 146, 160
Carter, C.B., s e e Zhu, J.G. 186
Castaing, B., s e e Goy, P. 25
Caulet, J. 187
Caulet, J., s e e Guivarc'h, A. 187
Caulet, J., s e e Le Corre, A. 187, 188
Cendlewska, B. 184
Cendlewska, B., s e e Bohdziewicz, A. 184
Cern~,, S, 143
Cern~,, S, s e e Boeva, O.A. 136, 143, 240
CereS,, S., s e e Smutek, M. 136
Cerri, A. 125, 140
Chaiss6, E, s e e Bloch, D. 303
Chang, C.L., s e e den Boer, M.L. 115, 130
Chang, C.T., s e e Han, J.W. 233
Chang, C.T., s e e Torgeson, D.R. 233
Chang, S., s e e Franciosi, A. 132
Chang, S., s e e Raisanen, A. 132
Chang, S., s e e Wall, A. 132
Chang, Y.S., s e e Parks, R.D. 181
Chapman, J.W. 126
Chapman, S.B. 92, 94, 95
Chassaing, G., s e e Sigrist, M. 126, 128
Chattopadhyay, T. 401
Chattopadhyay, T., s e e Rossat-Mignod, J. 62
Chaumont, J., s e e Mathevet, J.P. 180
Chee, K.T. 128, 129
Cheetham, A.K., s e e Titcomb, C.G. 225
Chen, D.Y. 132
Chen, EL., s e e Sakamoto, Y. 182
Chen, J.K., s e e Ramesh, R. 150, 151
Chen, K., s e e Cadieu, EJ. 156
Chen, N,, s e e Rau, C. 123
Chen, X., s e e Ma, E. 143
Cheng, S.E 155
Chang, S.E, s e e Demczyk, G.B. 155
Cherifi, K. 160
Chernoplekov, N.A., s e e Parshin, EP. 243
Chieux, E, s e e Barandiaran, J.M. 366
Chiheb, M. 219, 221,225, 227
Chiheb, M., s e e Daou, J.N. 219, 234
Chirico, R.D., s e e Carlin, R.L. 275, 281
Chistyakov, O.D., s e e Goremyehkin, E.A. 339,
342
Chizhov, P.E. 134
Chizhov, P.E., s e e Kostygov, A.N. 134

AUTHOR INDEX
Chizhov, RE., s e e Morozov, Yu.G. 134
Chlebeck, H.G., s e e Curzon, A.E. 110
Choe, G. 163
Chorkendorff, I. 114, 116
Chorkendorff, I., s e e Andersen, J.N. 115, 116
Chorkendorff, I., s e e Nyholm, R. 116
Chorkendorff, I., s e e Onsgaard, J. 116, 117, 127,
129, 131
Chouteau, G., s e e Burger, J.E 268
Chouteau, G., s e e Daou, J.N. 269
Chouteau, G., s e e Vajda, E 268, 269
Chowdhury, M.R. 229, 237
Christensen, N.E., s e e Zwieknagl, G. 14, 77, 78
Christyakov, O.D., s e e Alekseev, RA. 339, 341
Christyakov, O.D., s e e Goremyehkin, E.A. 339
Ciszewski, A. 118
Clark, M.R., s e e Haschke, J.M. 231
Claus, H., s e e Carlin, R.L. 271,277
Cochet-Muchy, D. 151
Cock, GJ., s e e McEwen, K.A. 350, 351
Coene, W. 152
Coey, J.M.D. 108, 178
Coey, J.M.D., s e e Li, H.S. 328
Cofield, M.L., s e e Shin, S.C. 160
Colliex, C. 117, 128-130, 132, 137
Colliex, C., s e e Brousseau-Lahaye, B. 128, 129
Colliex, C., s e e M a n o u b i , T. 132
Collins, M.E, s e e Lin, H. 379
Connerade, J.R 111, 112, 114, 115
Connerade, J.P., s e e Blancard, C. 112, 114
Connerade, J.R, s e e Esteva, JM. 131, 132
Cormerade, J.R, s e e Sarpal, B.K. 112
Cormor, D.W., s e e Rush, JA. 243
Contreras, M.C., s e e Alameda, J.M. 145
Cooper, B.R. 350, 353
Coqblin, B. 415
Cordero, E 232
Cordero, E, s e e Cannelli, G. 232
Corliss, L.M., s e e Feleher, G.R 317
Comer, W.D. 126
Comer, W.D., s e e Smith, R.L. 126
Cosier, J., s e e Hill, R.W. 39
Cotts, R.M., s e e Schreiber, D.S. 236, 237
Cotts, R.M., s e e Zamir, D. 256
Coulman, D. 166
Cowgill, D.F. 141
Cowgill, D.E, s e e Bacon, EM. 141
Cowley, R.A. 417
Cowley, R.A., s e e Jehan, D.A. 379
Cowley, R.A., s e e Jensen, J. 268
Cox, RA., s e e Lang, J.K. 115
Cox, S.RJ., s e e Chowdhury, M.R. 229, 237
Crabtree, G.W. 62-64
Crabtree, G.W., s e e Aoki, H. 63~55

429

Crabtree, G.W., s e e Arko, A.J. 31, 33

Crabtree, G.W., s e e Johanson, W.R. 52, 53
Crabtree, G.W., s e e Joss, W. 33, 34, 36
Crabtree, G.W., s e e Koelling, D.D. 4
Crabtree, G.W., s e e 0nuki, Y. 33-37, 91-93, 95
Craig, R.S., s e e Ganapathy, E.V 301
Craig, R.S., s e e Nasu, S. 40
Craig, R.S., s e e Smith, H.K. 137
Cmwford, R.K., s e e Kamitakahara, W.A. 243
Creeelius, G., s e e Wertheim, G.K. 115
Creuzet, G., s e e Barthem, V.M.T.S. 339
Croat, J.J., s e e Herbst, J.H. 312
Croft, M. 181
Croft, M., s e e Lu, F. 181
Cromer, D.T. 174
Cromer, D.T., s e e Larson, A.C. 87
Cromer, D.T., s e e Liberman, D. 9, 21
Cukier, M. 128, 129, 131
Curzon, A.E. 110, 120, 133-135, 137, 143
Curzon, A.E., s e e Rajora, O.S. 133
Curzon, A.E., s e e Singh, O. 134
Cyrot, M. 297, 301
Cyrot, M., s e e Lavagna, M. 308
Cywinski, R., s e e Mondal, S. 320
Cywinski, R, s e e Ritter, C. 321
Czjzek, G. 366
Czopnik, A. 47, 398
Czopnik, A., s e e Morin, P. 355, 397-399
Daalderop, G.H.O. 313-315
Daane, A.H., s e e Gschneidner Jr, K.A. 112
Dadiani, T.O. 188
Dadiani, T.O., s e e Dzhabua, Z.U. 188
Dadiani, T.O., s e e Glurdzhidze, L.N. 188
Dai, D.S., s e e Fang, R.Y. 157
d'Ambrumenil, N., s e e Stieht, J. 81
D'Amico, K.L., s e e Bohr, J. 417
D'Amieo, K.L., s e e Gibbs, D. 268, 378, 417
Daniel-Szab6, J., s e e Dudks, L 120
Danielou, R. 216
Daou, J.N. 212, 213, 216, 217, 219-221,225,
229, 233, 234, 238, 239, 241,243-248, 251-253,
255, 258, 261, 262, 264, 267-269, 271,273,
275, 277-279, 283, 284
Daou, J.N., s e e Andre, G. 227, 230, 281
Daou, J.N., s e e Blaschko, O. 215, 219, 220, 234
Daou, J.N., s e e Bonnet, J.E. 219, 225, 240
Daou, J.N., s e e Boukraa, A. 275, 282-285
Daou, J.N., s e e Burger, J.P. 229, 238, 243, 251,
252, 257-261,268, 271,273, 275-277, 283, 285
Daou, J.N,, s e e Chiheb, M. 221,225, 227
Daou, J.N., s e e Danielou, R. 216
Daou, LN., s e e Lucasson, A. 241,243, 251,252
Daou, J.N., s e e Metzger, T.H. 236

430

AUTHOR INDEX

Daou, J.N., s e e Pleschiutschnig, J. 234

Daou, J.N., s e e Schmitzer, C. 236, 248, 268
Daou, J.N., s e e Senoussi, S, 271,273, 277, 279
Daou, J.N., s e e Shaltiel, D. 285
Daou, J.N., s e e Udovic, T.J. 234, 235
Daou, J.N., s e e Vajda, P. 210, 213,219, 220, 227,
229, 230, 232, 233, 236-238, 243-246, 251,
253-260, 262, 263, 265, 267-271,273, 275,
277-285
Daou, J.N., s e e Viallard, R. 240
Daraek, S., s e e Andres, K. 339, 341
Dariel, M.E, s e e Shikhmanter, L. 179, 181
Dariel, M.R, s e e Venkert, A. 179
Darmon, J.M., s e e Bosca, G. 132
Dartyge, E., s e e Brouder, C. 159, 165
Dartyge, E., s e e Guilmin, R 165
Das, S.K. 190
Das, S.K., s e e Suryanarayanan, R. 189
Date, M. 345, 381,401
Date, M., s e e Morin, R 396, 397
Date, M., s e e Shigeoka, T. 401
Date, M., s e e Sugiura, E. 401
Date, M., s e e Sugiyama, K. 7, 365, 371
Date, M., s e e Tomiyama, E 336
David, J., s e e Bosca, G. 132
Davis, R.E, s e e Mason, M.G. 114
de Boer, ER., s e e Franse, J.J.M. 336
de Boer, ER., s e e Meyer, R.T.W. 51
de Boer, ER., s e e Sinnema, S. 336, 337
de Boer, ER., s e e Szytuta, A. 384
de Boer, ER., s e e Tomiyama, E 336
de Boer, ER., s e e Verhoef, R. 331,332
de Frutos, M., s e e Br~chignac, C. 114
de Haas, W.J. 27
de Lacheisserie, E., s e e Morin, R 355, 394, 396
de Lima, A.E, s e e Ferreira, R 260
De Luea, J.C. 180
de Mooij, D.B., s e e Coene, W. 152
de Mooij, D.B., s e e Sinnema, S. 329
De Rosa, E, s e e Allen Jr, S.J. 186, 187
de Visser, A., s e e Ball, A.R. 342, 351-353
de Visser, A., s e e Franse, J.J.M. 325
de Visser, A., s e e van der Meulen, H.E 81
de Vroomen, A.R., s e e van Deursen, A.J.R 33,
34, 36
Degtyareva, V.E, s e e Kostygov, A.N. 134
Dehner, B., s e e Shinar, J. 251,253, 255, 257
Dejauque, J., s e e Bras, J. 147
del Moral, A. 355
Delaeote, D., s e e Rossat-Mignod, J. 401
Deleroix, R, s e e Baezewski, L.T. 159, 161
Deleroix, P., s e e Brouder, C. 159
Delcroix, P., s e e Piecueh, M. 159, 161
Delley, B . , s e e Moser, H.R. 130-132

Demczyk, B.G., s e e Cheng, S.E 155

Demczyk, G.B. 155
den Boer, M.L. 115, 130
den Boer, M.L., s e e Murgai, V. 129, 130
den Boer, M.L., s e e Parks, R.D. 181
Denda, A., s e e Kojima, T. 169
Denda, A., s e e Uchida, H. 169, 170
Denissen, C.J.M., s e e Verhoef, R. 331,332
D~portes, J. 321-323
D6portes, J., s e e Ballou, R. 317, 319, 320, 322,
324
D6portes, J., s e e Berthier, Y. 313
D6portes, J., s e e Brown, EJ. 321
D6portes, J., s e e Oddou, J.L. 321
D6portes, J., s e e Voiron, J. 321
Devlin, E., s e e McGuiness, P.J. 148
Dexpert, H., s e e Gasgnier, M. 132
Dexpert-Ghys, J. 137
Dhar, S.K., s e e Gschneidner Jr, K.A. 6, 52, 264
Dhar, S.K., s e e Ikeda, K. 305
Dianoux, A.J., s e e Barnfather, K.J. 237
Dickinson, P.H., s e e Webb, D.J. 162
Dieny, B. 165
Dijkman, W.H. 70, 72
Dimmock, J.O. 12
Divis, M., s e e Svoboda, P. 387
Dobbertin, D.C., s e e Marshall, A.N. 164
D6hler, H., s e e Zogal, O.J. 237
Dokadze, E.V., s e e Dadiani, T.O. 188
Dokadze, E.V., s e e Dzhabua, Z.U. 188
Dokadze, E.V., s e e Glurdzhidze, L.N. 188
Domke, M. 114, 115
Doniach, S. 304
Doniach, S., s e e Murata, K.K. 327
Donovan, E, s e e Cherifi, K. 160
Donovan, P.E., s e e McMinn, R. 134
Dormann, E., s e e Shalfiel, D. 285
Doukour6, M. 368, 369
Dowben, P.A. 124
Dowben, P.A., s e e LaGraffe, D. 124
Dowben, P.A., s e e Li, D. 123
Drexel, W., s e e Knorr, K. 225, 243, 277
Drulis, H. 225, 256
Drulis, H., s e e Drulis, M. 241,252, 263, 275,
285
Drulis, H., s e e Iwasieczko, W. 225
Drulis, H., s e e Smirnov, I.A. 25t, 264
Drulis, M. 241,243, 252, 258, 263, 273, 275,
277-279, 282, 285
Drulis, M., s e e Bieganski, Z. 229, 282
Drulis, M., s e e Drulis, H. 225
Drulis, M., s e e Smirnov, I.A. 251,264
Due, N.H. 307, 308
Dud~s, J. 118, 120

AUTHOR INDEX
Dudfis, J., s e e Jinos, S. 120
Dufour, C., s e e Cherifi, K. 160
Dunlap, B.D., s e e Friedt, J.M. 282
Dunlap, B.D., s e e Koelling, D.D. 4
Durand, J., s e e Baczewski, L.T. 159, 161
Durand, J., s e e Malterre, D. 182
Durand, J., s e e Piecuch, M. 159, 161
Duxbury, RM., s e e Selke, W. 399
Dzhabua, Z.U. 188
Dzhabua, Z.U., s e e Dadiani, T.O. 188
Dzhabua, Z.U., s e e Glurdzhidze, L.N. 188
Eastman, D.E., s e e Kaindl, G. 117
Eaton, G.H., s e e Chowdhury, M.R. 229, 237
Ebe, H., s e e Ishida, A. 188
Ebihara, T. 42, 44-46, 48, 49, 51
Ebihara, T., s e e Umehara, I. 47-51
Ebihara, T., s e e Onuki, Y. 74-78, 80
Ebisawa, T., s e e Uehida, H. 169, 170
Eeher, C.J., s e e Koestler, C. 150, 151
Edelstein, A.S., s e e Johanson, W.R. 52, 53
Edwards, D.M., s e e Bloch, D. 300, 301
Efendiev, E.G. 187
Effantin, J.M. 31,359, 397
Effantin, J.M., s e e Rossat-Mignod, J. 62, 401
EgelhoffJr, W.E 116, 131
Egelhoff Jr, W.E, s e e Tibbetts, G.G. 116
Eguchi, T., s e e Kuwano, N. 182, 184
Eichner, S., s e e Rau, C. 124
Eick, H.A. 110
el Mandouh, Z.S., s e e Mahmoud, S. 118, 127
Elbicki, J.M., s e e Cheng, S.E 155
Eley, D.D. 136, 140
Ellegaard, 0 . , s e e Onsgaard, J. 116
Ellinger, EH. 110
Elyutin, A.V, s e e Savrin, V.D. 109
Emerson, J.P., s e e Fisher, R.A. 324
Endoh, D. 92, 93
Endoh, Y., s e e Hosoito, N. 158
Endoh, Y., s e e Motokawa, M. 380
Endoh, Y., s e e Nojiri, H. 381
Engelhardt, M.A., s e e Jaswal, S.S. 151
Engelsberg, S., s e e Brinkman, W.E 305
Engelsberg, S., s e e Doniach, S. 304
Eremenko, Z.V, s e e Savrin, V.D. 109
Eriksson, B., s e e Andersen, J.N. 116
Eriksson, B., s e e Nilsson, A. 116
Ernst, G., s e e Blaschko, O. 219, 220
Erskine, J.L., s e e Li, D. 123
Eschenfelder, A.H. 108
Esteva, J.M. 131, 132
Esteva, J.M., s e e Blancard, C. 112, 114
Esteva, J.M., s e e Gasgnier, M. 132
Esteva, J.M., s e e Sarpal, B.K. 112

Esteva, J.M., s e e Thole, B.T. t 14

Etourneau, J., s e e Kasaya, M. 6
Evans, J., s e e Hughes, D.T. 182
Everett, G.E. 394, 411
Eyring, L. 108
Eyring, L., s e e Eick, H.A. 110
Eyring, L., s e e Felmlee, T.L. 110

Faber, J., s e e Shaked, H. 271,275, 281-284

Fabian, D.J., s e e Gimzewski, J.K. 137
Fagot, M., s e e Bras, J. 147
Fairclough, J.P.A. 219
F/ildt, A. 115, 116
Falicov, L.M. 4
Fang, R.Y. 157
Fantner, E.J., s e e Krost, A. 188, 189
Farle, M. 122, 123
Farle, M., s e e Baberschke, K. 122
Farr, LEG., s e e Hirst, J.R. 182
Farraut, S., s e e Corner, W.D. 126
Fawcett, E. 26
Fedorus, A.G., s e e Gonchar, V.V 119
Fedotov, V.G., s e e Fedotov, V.K. 225
Fedotov, V.K. 225
Feh6r, A., s e e Dudfis, J. 118, 120
Feh~r, A., s e e J~inos, S. 120
Felcher, G.P. 317
Felder, E., s e e Schlapbach, L. 252, 257
Feldman, W.L., s e e Lowe, W.P. 125
Felmlee, T.L. 110
Felsch, W., s e e Brouder, C. 165
Felsch, W., s e e Guilmin, P. 165
Fender, B.E.E, s e e Knorr, K. 225, 243, 277
Fender, B.E.E, s e e Titcomb, C.G. 225
Fernengel, W., s e e Rodewald, W. 152, 153
Ferrater, C. 158
Ferrater, C., s e e Badia, E 157
Ferreira, P. 260
Ferro, R., s e e Rossi, D. 184
Fidler, J. 150, 152, 153, 155, 174
Fidler, J., s e e Knoch, K.G. 152
Fidler, J., s e e Parker, S.EH. 153
Figiel, H., s e e Kakol, Z. 311
Fillion, G. 393
Fillion, G., s e e Lethuillier, E 48
Finstad, T.G., s e e Palmstrom, C.J. 186, 187
Fischer, P. 394, 411
Fischer, R, s e e Schefer, J. 225, 273, 277
Fisher, R.A. 324
Fisher, R.A., s e e Amato, A. 94, 95
Fisk, Z., s e e Arko, A.J. 31, 33
Fisk, Z., s e e Joss, W. 33, 34, 36
Fisk, Z . , s e e van Deursen, A.J.R 33, 34, 36

431

432

AUTHOR INDEX

Flanagan, T.B., s e e Sakamoto, Y. 182

Flanagan, T.B., s e e Takao, K. 182
Flodstr6m, A., s e e Johansson, L.I. 181
Flodstr6m, A., s e e Kammerer, R. 115
Flodstr6m, A.S., s e e Gerken, E 115
Flotow, H.E., s e e Rush, J.J. 243
Flouquet, J., s e e Amato, A. 94, 95
Flouquet, J., s e e Benoit, A. 52
Flouquet, J., s e e Rossat-Mignod, J. 81
Flouquet, J., s e e van der Meulen, H.R 81
Fontaine, A., s e e Brouder, C. 159, 165
Fontaine, A., s e e Guilmin, P. 165
Forester, D.W. 168
Forester, D.W., s e e Schelleng, J.H. 167
Forester, D.W., s e e Vittoria, C. 168
Forgan, G.M. 26
Forker, M. 231
Forsyth, H., s e e McGuiness, P.J. 148
Fort, D., s e e Chapman, J.W. 126
Fort, D., s e e Hirst, J.R. 182
Fort, D., s e e Pecharsky, V.K. 379
Fraisse, R., s e e Ahmed-Mokhtar, N. 127
Frak, R.M., s e e Iwasieczko, W. 225
Franceschi, E. 179
Franciosi, A. 132
Franciosi, A., s e e Mason, M.G. 114
Franciosi, A., s e e Raisanen, A. 132
Franciosi, A., s e e Wall, A. 132
Franck, O., s e e Melmed, A.J. 118, 119
Franqois, J.C., s e e Sigrist, M. 126, 128
Frandon, J. 128-130
Frandon, J., s e e Brousseau-Lahaye, B. 128, 129
Frankenthal, R.E 168
Frankenthal, R.P., s e e van Dover, R.B. 168
Franse, J.J.M. 325, 328, 336, 350, 417
Franse, J.J.M., s e e Luong, N.H. 90, 365
Franse LJ.M., s e e Radwanski, R.J. 331,334-336,
339 342
Franse J.J.M., s e e Sinnema, S. 329, 336, 337
Franse J.J.M., s e e Thuy, N.P. 311,312
Franse J.J.M., s e e Tomiyama, E 336
Franse J.J.M., s e e Verhoef, R. 331-333
Franse J.J.M., s e e Zhang, EY. 339, 340, 342,
343 361
Franse J.J.M., s e e van der Meulen, H.P. 81
Franz W., s e e Steglich, E 6
Fraustc ER. 168
Fraymonville, R., s e e Krost, A. 188, 189
Fredkin, D.R., s e e B~al-Monod, M.T. 305
Freeman, A.J., s e e Rath, J. 22
Freeman, A.J., s e e Weinert, M. 309, 314
Freeman, A.J., s e e Wu, R. 122
Freitas, EP. 185
Freltoft, T. 323

Freltoft, T., s e e Motoya, K. 323

Fr~my, M.A. 392
Fr6my, M.A., s e e Belorizky, E. 328-330, 337
Frick, B., s e e Stuhr, U. 237
Friedel, J. 311,313
Friedt, J.M. 282
Friedt, J.M., s e e Shinjo, T. 158
Frigerio, J.M. 177
Ffigerio, J.M., s e e Martin, M. 177
Frings, P.H., s e e Franse, J.J.M. 336
Fromm, E., s e e Wulz, H.G. 136, 169
Fuertes, J.E, s e e Alameda, J.M. 145
Fuggle, J.C., s e e Esteva, J.M. 131, 132
Fuggle, J.C., s e e Thole, B.T. 114
Fujii, A., s e e Isikawa, Y. 85
Fujii, H. 380, 381,384
Fujii, H., s e e Hashimoto, Y. 387
Fujii, H., s e e Iwata, N. 368
Fujii, H., s e e Okamoto, T. 360, 361
Fujii, H., s e e Shigeoka, T. 380, 382-385, 401
Fujii, H., s e e Sugiura, E. 401
Fujii, H., s e e Sugiyama, K. 371
Fujii, T., s e e Inoue, A. 166
Fujimaki, Y., s e e Ebihara, T. 48, 49, 51
Fujimaki, Y., s e e Umehara, I. 47-51, 70-72
Fujimori, A. 265
Fujimori, H. 159
Fujimori, N., s e e Kamiguchi, 3(. 159
Fujimura, S., s e e Yamamoto, H. 151, 152
Fujimura, T., s e e Endoh, D. 92, 93
Fujimura, T., s e e Goto, T. 33, 34
Fujimura, T., s e e Settai, R. 87-89
Fujimura, T., s e e Suzuki, T. 31, 33, 91
Fujita, T., s e e Satoh, K. 5, 94
Fujita, T., s e e Onuki, Y. 87
Fujiwara, H., s e e Hashimoto, Y. 387
Fujiwara, H., s e e Liu, W.L. 366
Fujiwara, H., s e e Okamoto, T. 360, 361
Fujiwara, H., s e e Yamashita, M. 388
Fujiyasu, H., s e e Ishida, A. 188, 189
Fukada, A., s e e Umehara, I. 70-72
Fukada, A., s e e Onuki, Y. 87
Fukami, E., s e e Wada, H. 324, 325
Fukamichi, K., s e e Gambino, R.J. 180
Fukamichi, K., s e e Goto, T. 300, 301,306
Fukamichi, K., s e e Sakakibara, T. 300, 301
Fukuhara, T. 74-76
Fukuma, H. 394
Fulde, E 14, 26
Fuller, M.L., s e e Larsen, J.W. 169, 171
Furrer, A., s e e Fischer, E 394, 411
Furrer, A., s e e Purwins, H.G. 360
Furukawa, M., s e e Sakamoto, Y. 182
Furukawa, Y., s e e Komatsubara, T. 31

AUTHOR INDEX
Galera, R.M. 33
Gal6ra, R.M., s e e Amara, M. 398
Gal6ra, R.M., s e e Voimn, J. 321
Galoshina, E.V, s e e Volkenshtein, N.V. 271
Gambino, R.J. 180
Gambino, R.J., s e e De Luca, J.C. 180
Gambino, R.J., s e e Pickart, S.J. 178
Gambino, R.J., s e e yon Molnar, S. 178, 180
Ganapathy, E.V. 301
Gao, Q.Z., s e e Kitazawa, H. 41-43
G a o , W . M . , s e e Huang, G.X. 150
Gamier, A., s e e Shigeoka, T. 368
Garrison, K.C., s e e Allen Jr, S.J. 186, 187
Garrisson, K.C., s e e Palmstrom, C.J. 186
Gasgnier, M. 108-110, 118, 120, 126, 130, 132,
137, 144-146, 150, 155, 156, 160, 163, 167,
175, 177, 184, 190
Gasgnier, M., s e e Brousseau-Lahaye, B. 128, 129
Gasgnier, M., s e e Colliex, C. 117, 128-130, 132,
137
Gasgnier, M., s e e Curzon, A.E. 134
Gaukler, K.H., s e e Reiehl, R. 111, 131
Gauth~, B., s e e Cukier, M. 128, 129, 131
Gavigan, J.P., s e e Belorizky, E. 328-330, 337,
343-345
Geballe, T.H., s e e Toxen, A.M. 163
Geballe, T.H., s e e Webb, D.J. 162, 163
Gehring, G.A., s e e Langford, H.D. 33
Gellman, A., s e e Jaffey, D.M. 116, 179
Gerken, E 115, 131
Gerken, E, s e e Johansson, L.I. 181
Gerken, E, s e e Kammerer, R. 115
Gersdorf, R., s e e Franse, J.J.M. 336
Ghijsen, J., s e e Andersen, J.N. 115, 116
Gibbs, D. 268, 378, 417
Gibbs, D., s e e Bohr, J. 417
Gibson, M.T., s e e Gimzewski, J.K. 137
Gignoux, D. 83, 85, 297, 299-303, 308-310,
313, 339, 342, 345, 362, 366, 369, 371, 373,
375-377, 385, 387, 389, 392, 393, 406, 409, 415
Gignoux, D., s e e Aubert, G. 339
Gignotlx, D., s e e Bail, A.R. 342, 351-353, 375,
377, 378, 380, 386, 391,392, 402-406, 417
Gignoux, D., s e e Ballou, R. 310, 339
Gignoux, D., s e e Barandiaran, LM. 366
Gignoux, D., s e e Barbara, B. 299
Gignoux, D., s e e Barthem, V.M.T.S. 339-341,
343, 344
Gignoux, D., s e e Blaneo, J.A. 365, 370, 371,382,
383, 401-403, 405, 406
Gignoux, D., s e e Cyrot, M. 301
Gignoux, D., s e e Doukour6, M. 368, 369
Gignoux, D., s e e Fr6my, M.A. 392
Gignoux, D., s e e Jerjini, M. 70, 392

433

Gignoux, D., s e e Radwanski, R.J. 339, 342

Gignoux, D., s e e Reiffers, M. 339-341,351
Gignoux, D., s e e Shigeoka, T. 368
Gignoux, D., s e e Zhang, EY. 339, 340, 342, 343,
361
Gil, J.M., s e e Ferreira, R 260
Gilchrist, H.L., s e e Allen Jr, S.J. 186, 187
Gilchrist, H.L., s e e Palmstrom, C.J. 186
Gimzewski, J.K. 137
Ginsberg, M.J., s e e Grady, D.E. 132
Giordano, N., s e e Stryjewski, E. 345-347
Giraud, M., s e e Morin, E 355, 397-399
Girouard, EE., s e e Ashrit, RV. 133
Girouard, EE., s e e Chee, K.T. 128, 129
Givord, D. 301,302
Givord, D., s e e Alameda, J.M. 311, 312
Givord, D., s e e Belorizky, E. 328-330, 337,
343-345
Givord, D., s e e Dieny, B. 165
Givord, D., s e e Due, N.H. 307, 308
Givord, E 301
Givord, E, s e e Bloeh, D. 303
Givord, E, s e e Cyrot, M. 301
Givord, E, s e e Fillion, G. 393
Givord, E, s e e Gignoux, D. 83, 85, 299-302, 313
Glinka, C.J. 243
Glinski, M., s e e Kletowski, Z. 41-43
Glurdzhidze, L.N. 188
Glurdzhidze, L.N., s e e Dadiani, T.O. 188
Glurdzhidze, L.N., s e e Dzhabua, Z.U. 188
Goldman, A., s e e Gsehneidner Jr, K.A. 264
Goltros, W. 188
Gomez-Sal, J.C., s e e Barandiaran, J.M. 366
Gomez-Sal, J.C., s e e Blanco, J.A. 372, 373, 401,
403-405
Gonchar, EM. 118
Gonehar, V.V. 119
Gond6, Y. 167
Gond6, Y., s e e Suezawa, Y. 167
Goodhead, K., s e e Jones, P.M.S. 217
Goremychkin, E.A. 339, 342
Gorges, B., s e e Ballou, R. 308, 310, 336
Gorodetskii, D.A. 118, 119
G6seiafiska, I., s e e Ratajezak, H. 157
Goto, T. 33-35, 6 5 , 6 6 , 300, 301,306
Goto, T., s e e Endoh, D. 92, 93
Goto, T., s e e O n u l d , Y . 7
Goto, T . , s e e Sakakibara, T. 300, 301
Goto, T., s e e Settai, R. 52, 53, 60, 61, 65, 87-89
Goto, T., s e e Suzuki, T. 31, 33, 34, 91
Goto, Z, s e e Takayanagi, S. 363, 364
Goto, T., s e e Yoshimura, K. 301
Gottwich, U., s e e Franse, J.J.M. 325
Goy, P. 25

434

AUTHOR INDEX

Grady, D.E. 132

Graham Jr, C.D., s e e Tang, W. 150
Grant, W.A. 167
Gratz, E. 87
Grayevsky, A., s e e Shaltiel, D. 285
Greene, L.H., s e e Lowe, W.P. 125
Greis, O. 221,225, 231
Greis, O., s e e Mfiller, H. 225
Grepstad, J.K., s e e Bmaten, N.A. 178
Grey, F., s e e Andersen, J.N. 115, 116
Grieb, B. 152
Grieb, B., s e e Knoch, K.G. 152, 153
Grier, B.H., s e e Gibbs, D. 378, 417
Griessmann, H., s e e Alekseev, P.A. 339, 341
Groiss, C., s e e Fidler, J. 153
Gr6ssinger, R. 153
Grozdev, K.I., s e e Apostolov, A.V. 126, 127
Grundy, R.J., s e e Parker, S.EH. 153
Griitter, P. 148
G~tter, P., s e e Heinzelmann, H. 148
Gsehneidner Jr, K.A. 6, 25, 52, 72, 112, 118,
135, 264
Gsehneidner Jr, K.A., s e e Beaudry, B.J. 112
Gsehneidner Jr, K.A., s e e Corner, W.D. 126
Gsehneidner Jr, K.A., s e e Hungsberg, R.E. 70
Gsehneidner Jr, K.A., s e e Ikeda, K. 305
Gsehneidner Jr, K.A., s e e Ito, T. 229
Gselmeidner Jr, K.A., s e e Kai, K. 229, 241,252,
256
Gschneidner Jr, K.A., s e e Lgsser, R. 265
Gsehneidner Jr, K.A., s e e Pecharsky, V.K. 379
Gschneidner Jr, K.A., s e e Stierman, R.J. 246, 271
Gschneidner Jr, K.A., s e e Tang, J. 70, 72
Gsehneidner Jr, K.A., s e e Thome, D.K. 236, 241,
246, 247, 252
Gschneidner Jr, K.A., s e e Vajda, E 220, 236, 243,
245
Gsehneidner Jr, K.A., s e e Vglkl, J. 233
Gu, B.X. 145
Gu, B.X., s e e Homburg, H. 145
Gubbens, EC.M. 339
Gudat, W., s e e Alvarado, S.E 131
Gudat, W., s e e Hillebreebt, EU. 63
Gudat, W., s e e Weller, D. 122, 124
Guesnais, B., s e e Caulet, J. 187
Guesnais, B., s e e Le Corre, A. 187
Guilmin, E 165
Guilmin, E, s e e Brouder, C. 159, 165
Guivarc'h, A. 187
Guivarc'h, A., s e e Cadet, J. 187
Guivarc'h, A., s e e Le Corre, A. 187, 188
Gunnarsson, O. 15
Gt~ntherodt, H.-J., s e e Grtitter, E 148
Gantherodt, H.-J., s e e Heinzelmann, H. 148

Guo, G.Y. 14, 69

Guo, Y., s e e Li, H. 185
Gupta, M. 243, 256
Gupta, S.C. 132
Gupta, Y.M. 132
Gupta, Y.M., s e e Brar, N.S. 132
Gupta, Y.M., s e e Chen, D.Y. 132
Gupta, Y.M., s e e Gupta, S.C. 132
Gustafsson, T., s e e Johansson, L.I. 131
Guy, C.N., s e e yon Molnar, S. 178
Gygax, EN. 232
Gyorgy, E.M., s e e Hellman, E 165
Gyorgy, E.M., s e e Lowe, W.P. 125
Gyorgy, E.M., s e e van Dover, R.B. 168
Habu, K., s e e Sato, N. 160, 164
Habu, K., s e e Yamauehi, K. 160
Haekemer, M., s e e Czopnik, A. 47
Hadari, Z., see Mintz, M.H. 213, 240
Hadjipanayis, G.C., s e e Aylesworth, K.D. 146,
150, 151
Hadjipanayis, G.C., s e e Jaswal, S.S. 152
Hadjipanayis, G.C., s e e Strzeszewski, J. 144, 151
Haensel, R., s e e Niemann, W. 114
Haga, Y., s e e Nimori, S. 350
Haga, Y., s e e Ozeki, S. 65, 66
Haga, Y., s e e Takeda, N. 65, 67
Hagstr6m, S.B.M., s e e Allen, J.W. 115
Hagstr6m, S.B.M., s e e Johansson, L,I. 131
Hahn, H., s e e Li, Z. 110
Hakkens, E, s e e Coene, W. 152
H~ilg, W., s e e Schefer, J. 225, 273, 277
Hall, G., s e e Croft, M. 181
Hall, G., s e e Lu, E 181
Halpem, H.A., s e e Pickart, S.J. 178
Harnaker, M.C., s e e Huang, C.Y. 185
Hart, J.W. 233
Han, J.W., s e e Lichty, L.R. 232, 244
Hart, J . W . , s e e Torgeson, D.R. 233
Handstein, A., s e e Moch, Th. 178
Hanke, W., s e e Falieov, L.M. 4
Hansen, P. 180
H a n y u , T., s e e Miyahara, T.
129, 130
Hanzawa, K., s e e Yamada, K. 5
Hara, K., s e e Harada, T. 157
Hara, K., s e e Yamashita, S. 155
Harada, T. 157
Harbecke, B., s e e Krost, A. 188, 189
Harima, H. 14, 29, 32, 33, 72, 73, 81, 82, 89, 92,
94
Harima, H., s e e Kubo, Y. 14, 36, 37
Harima, H., s e e Sakai, O. 60, 61, 63, 64
Harris, I.R., s e e Brooks, J.W. 182
Harris, I.R., s e e Hirst, J.R. 182

AUTHOR INDEX
Harris, I.R., s e e Hughes, D.T. 182
Harris, I.R., s e e McGuiness, EJ. 148
Harris, I.R., s e e Smith, D.A. 182
Harris, J.M. 177
Hartmann, M. 180
Hartmann, M., s e e Hansen, P. 180
Haschke, J.M. 231
Hasegawa, A. 22, 32, 39-43, 52-58, 60, 61,
67~69
Hasegawa, A., s e e Harima, H. 92, 94
Hasegawa, A., s e e Kitazawa, H. 41-43, 60, 63
Hasegawa, A., s e e Kletowski, Z. 41-43
Hasegawa, A., s e e Yamagami, H. 13, 75, 77-79,
84-86
Hasegawa, H., s e e Moriya, T. 305
Hashimoto, S. 168
Hashimoto, Y. 362, 387
Hashimoto, Y., s e e Iwata, N. 368, 374, 375
Hashimoto, Y., s e e Shigeoka, T. 380, 382, 401
Hastings, L.M., s e e Felcher, G.E 317
Hauser, J.J. 118, 125
Haussler, E 118
Hautecler, S., s e e Vorderwisch, E 243
Hautecler, S., s e e Wegener, W. 234
Hawkes, C.M., s e e Owen, G. 179
Hay, C.M. 179
Hayakawa, Y., s e e Fujimori, H. 159
Hayakawa, Y., s e e Kamiguchi, Y. 159
Hecht, M.H. 131
Hecht, M.H., s e e Egelhoff Jr, W.E 131
Hedge, H., s e e Cadieu, EJ. 156
Hedge, H., s e e Kamprath, N. 155
Heinzelmann, H. 148
Heinzelmarm, H., s e e Grfitter, P. 148
Heitmann, H. 168
HelIman, E 165
Henig, E.-Th., s e e Grieb, B. 152
Henig, E.-Th., s e e Knoch, K.G. 152, 153
Hennion, B., s e e Aubert, G. 339
Henry, J.Y., s e e Jerjini, M. 70, 392
Henry La Blanehetais, Cb., s e e Dexpert-Ghys, J.
137
Henry La Blanchetais, Ch., s e e Gasgnier, M. 1 4 6 ,
160
Herbst, G. 233
Herbst, J.H. 312
Heremans, J., s e e Salamanca-Young, L. 189
Herman, F. 9
Herring, C. 19
Herses, N., s e e Richter, H.J. 187
Hewson, A.C., s e e Wasserman, A. 35
Hickox, C.E., s e e Bacon, EM. 141
Hidber, H.-R., s e e Grfitter, R 148
Hidber, H.-R., s e e Heinzelmann, H. 148

435

Hien, R.D., s e e Thuy, N.P. 311, 312

Hien, T.D., s e e Luong, N.H. 90, 365
Hierschler, D., s e e Mintz, M.H. 213, 240
Higo, S., s e e Kuwano, N. 182, 184
Hill, H.H. 3, 10
Hill, R.W. 39
Hillebrecht, EU. 63
Hilscher, G., s e e Daou, J.N. 248, 251,269
Hilscher, G., s e e Schmitzer, C. 236, 248, 268
Hilscher, G., s e e Vajda, E 245, 246, 268-270
Hilscher, G., s e e Wiesinger, G. 210
Hilton, P., s e e Reinders, P.H.E 30
Himpsel, EJ., s e e Reihl, R. 122
Hirano, M., s e e Ohkoshi, M. 311
Hirosawa, S. 313
Hirosawa, S., s e e Yamamoto, H. 151, 152
Hirst, J.R. 182
Hoareau, B., s e e Rayane, D. 114
Hoffmann, K.E, s e e Drulis, H. 256
Hohenberg, R 14
Holc, J. 150
Holden, T.M., s e e Lin, H. 379
Holloway, D.M. 142
Holloway, D.M., s e e Antepenko, R.J. t42
Holloway, D.M., s e e Harris, J.M. 177
Homburg, H. 145
Homburg, H., s e e Gu, B.X. 145
Homma, H. 119
Homma, H., s e e Yang, K.A. 119
Honda, K., s e e Iwata, N. 368
Honda, S. 164
Hong, M., s e e Kwo, J. 119
Hong, M., s e e van Dover, R.B. 168
Hong, N.M., s e e Franse, J.J.M. 336
Hong, N.M., s e e Thuy, N.E 311, 312
Horikoshi, H. 173
Horikoshi, H., s e e Takeda, S. 173
Hormes, J., s e e Blancard, C. 112, 114
Hormes, J., s e e Sarpal, B.K. 112
Horn, S., s e e den Boer, M.L. 115, 130
Horn, S., s e e Murgai, V. 129, 130
HSrnstr6m, S.E., s e e Johansson, L.I. 181
Horvatic, M., s e e Berthier, Y. 313
Hoshi, Y. 168
Hoshi, Y, s e e Naoe, M. 168
Hosoda, N. 136
Hosoda, N., s e e Uehida, H. 169
Hosoito, N. 158
Hosoito, N., s e e Mibu, K. 158
Hosoito, N., s e e Shinjo, T. 158
Hosoito, N., s e e Yoden, K. 158
Hrossman, S.A., s e e Gimzewski, J.K. 137
Hua, H.C., s e e Huang, G.X. 150
Huang, C.Y. 185

436

AUTHOR INDEX

Huang, G.X. 150

Huang, Y.C., s e e Ohki, C. 239, 240
Huang, Y.C., s e e Toguchi, K. 240
Huang, Y.C., s e e Uchida, H. 169
Huang, Y.S. 179
Huang, Y.S., s e e Murgai, V. 129, 130
Hubberstey, P. 140
Hughes, D.T. 182
Hukin, D.A., s e e Hill, R.W. 39
Hull, G.W., s e e Allen Jr, S.J. 186, 187
Hulliger, F. 394, 395
Hulliger, F., s e e A16onard, R. 355
Hulliger, E, s e e Aoki, H. 63-65
Hulliger, E, s e e Crabtree, G.W. 62-64
Hungsberg, R.E. 70
Hunt, D.G. 243
Hunt, M. 81, 82
Hunt, M., s e e Chapman, S.B. 92, 94, 95
Hunt, M., s e e Satoh, K. 42, 44, 48, 51
Hussain, A.A.A. 134
Hutchings, C.W., s e e Li, D. 123
Hutchings, M.T. 414
Hiitten, U., s e e Forker, M. 231
Hwang, C., s e e Li~ D. 123
Iandelli, A., s e e Palenzona, A. 182
Ibanez-Meier, R., s e e Liehty, L.R. 232, 244
Ibarra, M.R., s e e del Moral, A. 355
Ichikawa, T., s e e Tsunashima, S. 163
Idczak, E. 128
Iga, E, s e e Settai, R. 87-89
Iga, E, s e e Sugiyama, K. 7
Ignatiev, A . , s e e Onsgaard, J. 111
Ignatiev, A., s e e Tougaard, S. 111, 126, 128
Ikeda, H., s e e 0nuki, Y. 95
Ikeda, K. 305
Ikeda, K., s e e Gsctmeidner Jr, K.A. 25
Ikeda, M., s e e Harada, T. 157
Ikematsu, M., s e e Inoue, A. 166
Iliew, N., s e e Czopnik, A. 398
Iliew, N., s e e Kletowski, Z. 52
tlyasov, T.M., s e e Efendiev, E.G. 187
Ina, K., s e e ()nuki, Y. 90-93, 390
Ina, K., s e e Takayanagi, S. 363, 364
Inokuchi, S., s e e Numata, T. 156
Inomata, K., s e e Sahashi, M. 156, 174
Inoue, A. 166
Inoue, J., s e e Shimizu, M. 297
Inoue, J., s e e Yamada, H. 301,306
Ishida, A. 188, 189
Ishii, H., s e e Baba, K. 188
Ishli, H., s e e Miyahara, T. 129, 130
Ishii, H., s e e Nakamura, O. 118, 125
Ishii, H., s e e Takeda, K. 133

Ishii, T. 3
Ishikawa, Y., s e e Takahashi, M. 70, 392
Ishiwatari, T., s e e Naoe, M. 168
Ishizawa, Y. 31-33, 35
Ishizawa, Y., s e e Nozaki, H. 365
Isikawa, Y. 85
Isikawa, Y., s e e Maezawa, K. 83, 84, 86, 90
Isikawa, Y., s e e Sato, K. 362, 363
Isikawa, Y., s e e Onuki, Y. 83, 84, 86
Ito, T. 229, 268, 269
Ito, Y., s e e Shigeoka, T. 383
Itoh, T., s e e Uchida, H. 169, 170
Ivanitskii, P.G., s e e Goremychkin, E.A. 339, 342
Ivanov, V. 384, 385
Ivanov, V., s e e Vinokurova, L. 382
Iwamura, E. 150
lwasieczko, W. 225
Iwasieczko, W., s e e Drulis, H. 225
Iwasieczko, W., s e e Smirnov, I.A. 251,264
Iwata, N. 368, 374, 375, 381,401,402
lwata, N., s e e Shigeoka, T. 380, 382-385, 401
Izumi, E, s e e Asano, H. 91

Jaccard, D., s e e Amato, A. 94, 95

Jaccard, D., s e e Rossat-Mignod, J. 81
Jacob, I. 241
Jacobs, T.H., s e e Coene, W. 152
Jacobs, T.H., s e e Verhoef, R. 331,332
Jacoud, J.L., s e e Rossat-Mignod, J. 81
Jaffey, D.M. 116, 179
J~iger, Ch., s e e Zogal, O.J. 237
Jan, J.-P. 39, 40
Jan, J.-P., s e e Boulet, R.M. 52
Janak, J.E 303
Jang, Y.-R., s e e Min, B.I. 36
J~nos, S. 120
J~nos, S., s e e Dud~s, J. 120
Jansen, A.G.M., s e e Reiffers, M. 339-341,351
Jaswal, S.S. 151, 152
Jaswal, S.S., s e e Shan, Z.S. 164
Jaussaud, C. 355, 356
Jayanetti, J.K.D., s e e Kamprath, N. 155
Jayaraman, A. 7
Jeandey, C., s e e Oddou, J.L. 321
Jehan, D.A. 379
Jennings, J.R. 179
Jennings, J.R., s e e Bryan, ST 179
Jennings, J.R., s e e Hay, C.M. 179
Jeunings, J.R., s e e Nix, R.M. 179
Jennings, J.R., s e e Owen, G. 179
Jensen, C.L. 245
Jensen, J. 266, 268, 379, 390, 391,412, 413, 416,
417

AUTHOR INDEX
Jensen, J., s e e Fulde, P. 26
Jensen, J., s e e Mackintosh, A.R. 390
Jepson, J. 22
Jerjini, M. 70, 392
Jezequel, G., s e e Quemerais, A. 127, 129
Jiang, S., s e e Li, H. 185
Jin, C., s e e Rau, C. 123
Jin, H., s e e Pan, S.M. 175
Johanson, W.R. 52, 53
Johansson, B. 111, 115
Johansson, B., s e e Johansson, L.I. 181
Johansson, B., s e e Nordstr6m, L. 313, 314
Johansson, B., s e e Rosengren, A. 111
Johansson, L.I. 115, 131, 181
Johansson, L.I., s e e Allen, J.W. 115
Johansson, L.I., s e e Gerken, E 115
Johansson, L.I., s e e Hecht, M.H. 131
Johansson, L.I., s e e Kammerer, R. 115
Johbettoh, H., s e e Hasegawa, A. 22, 54-57
Johnson, D.A. 109, 112
Johnson, R.L., s e e Andersen, J.N. 115, 116
Johnson, R.W. 178
Johnson, W.B., s e e Frausto, RR. 168
Johnson, W.L., s e e Schwarz, R.B. 180
Jones, D.W., s e e Chapman, J.W. 126
Jones, D.W., s e e Corner, W.D. 126
Jones, I.R, s e e Smith, D.A. 182
Jones, RM.S. 217
Jordan, R.G. 182
Jordan, R.G., s e e Corner, W.D. 126
Jorgensen, J.D., s p e Schefer, J. 225, 273, 277
Joss, W. 33, 34, .,6
Joss, W., s e e Aoki, H. 63-65
Joss, W., s e e Crabtree, G.W. 62-64
Joss, W., s e e Muller, T. 35, 36
Joung, K.O., s e e Carlin, R.L. 275, 281
Juckum, C. 213, 219, 225, 245
Judd, R.W., s e e Nix, R.M. 179
Julien, L.S., s e e Miller, R.E 118, 126, 127
Jung, R 220, 221,244, 245
Jung, Th., s e e Grfitter, R 148
Jungblut, R., s e e Carbone, C. 124, 158
Kaehel, I"., s e e Carbone, C. 124, 158
Kadomatsu, H., s e e Liu, W.L. 366
Kadomatsu, H., s e e Yamashita, M. 388
Kadowaki, K. 5
Kagami, S., s e e Nakamura, T. 162
Kagawa, M., s e e Takeda, N. 67
Kai, K. 229, 241,252, 256
Kai, K., s e e Vajda, E 220, 236, 243,245
Kaindl, G. 115, 117
Kaindl, G., s e e Domke, M. 114, 115
Kajiura, M. 157

437

Kakizaki, A., s e e Ishii, T. 3

Kakol, Z. 311
Kakuno, K., s e e Gond6, Y. 167
Kakuno, K., s e e Suezawa, Y. 167
Kaldis, E. 225
Kaldis, E., s e e Bischof, R. 214, 217, 271,273,
279
Kaldis, E., s e e Boroch, E. 221,225, 226
Kaldis, E., s e e Fischer, E 394, 411
Kaldis, E., s e e Fukuma, H. 394
Kalvius, G.M. 324
Kameda, K., s e e Onuki, Y. 90, 390
Kamiguchi, Y. 159
Kamiguchi, Y., s e e Fujimori, H. 159
Kamimura, H. 324
Kamitakahara, W.A. 243
Kammerer, R. 115
Kammerer, R., s e e Gerken, E 115
Kamprath, N. 155
Kamprath, N., s e e Liu, N.C. 155
Kanash, O.K., s e e Gonchar, VV 119
Kanayama, T., s e e Tsukahara, S. 167
Kanda, S., s e e Yamada, H. 306
Kandasamy, K. 137, 140
Kandasamy, K., s e e Surplice, N.A. 135, 137, 139
Kaneko, K., s e e Niihara, T. 168
Kaneko, M., s e e Hashimoto, S. 168
Kaneko, T. 388, 397
Kaneko, T., s e e Abe, S. 388
Kaneko, T., s e e Kitai, T. 388
Kanski, J. 131
Kappler, J.R, s e e Besnus, M.J. 74
Karim, D., s e e Arko, A.J. 31, 33
Karnatak, R.C., s e e Blancard, C. 112, 114
Karnatak, R.C., s e e Connerade, J.R 111, 112,
114, 115
Karnatak, R.C., s e e Esteva, J.M. 131, 132
Karnatak, R.C., s e e Gasgnier, M. 132
Karnatak, R.C., s e e Sarpal, B.K. 112
Karnatak, R.C., s e e Thole, B.T. 114
Karpukhina, L.G., s e e Smirnov, I.A. 251,264
Kasaya, M. 6
Kasaya, M., s e e Settai, R. 87-89
Kasaya, M., s e e Sugiyama, K. 7
Kass, W.J. 142
Kass, W.J., s e e Beavis, L.C. 142
Kasuya, M., s e e Ebihara, T. 48, 49, 51
Kasuya, T. 3, 13, 59, 60, 63-65, 415
Kasuya, T., s e e Effantin, J.M. 31,359, 397
Kasuya, T., s e e Fukuma, H. 394
Kasuya, T., s e e Galera, R.M. 33
Kasuya, T., s e e Goto, T. 33, 34
Kasuya, T., s e e Harima, H. 32, 33
Kasuya, T., s e e Kitazawa, H. 41-43, 60, 63, 65

438

AUTHOR INDEX

Kasuya, T., s e e Kwon, Y.S. 65, 67

Kasuya, T., s e e Morin, E 397
Kasuya, T., s e e Ozeki, S. 65, 66
Kasuya, T., s e e Sakai, O. 60, 61, 63, 64
Kasuya, T., s e e Settai, R. 60, 61, 65
Kasuya, T., s e e Sugiyama, K. 7, 365
Kasuya, T., s e e Suzuki, T. 33, 34
Kasuya, T., s e e Takahashi, H. 62
Kasuya, T., s e e ()nuki, Y. 7
Katayama, T. 166
Katayama, T., s e e Ohkoshi, M. 311
Kato, H., s e e Ishii, T. 3
Kato, T., s e e Maezawa, K. 83, 84, 86
Kaun, L.E, s e e Alekseev, P.A. 339, 341
Kavecansk), V, s e e Dudfis, J. 118, 120
Kawabata, H., s e e Uchida, K. 178
Kawaguchi, K., s e e Yoden, K. 158
Kawahata, T. 136
Kawai, H., s e e Nojiri, H. 381
Kawai, S., s e e Ishizawa, Y. 31-33, 35
Kawanaka, H., s e e Sugiura, E. 401
Kawano, S., s e e Nakamura, Y. 320
Kawano, S., s e e Shigeoka, T. 383, 384
Kayzel, EE., s e e Ball, A.R. 342, 351-353
Kayzel, EE., s e e Radwanski, R.J. 334, 335, 339,
342
Kayzel, EE., s e e Zhang, EY. 339, 340, 342, 343,
361
Kazimierzki, M., s e e Bohdziewicz, A. 184
Kelly, EJ., s e e Daalderop, G.H.O. 313-315
Kerley, N., s e e Reinders, EH.E 30
Ketterson, J.B., s e e Arko, A.J. 31, 33
Ketterson, J.B., s e e Windmiller, L.R. 30
Khanudova, Kh.Kh., s e e Varkanova, R.G. 142
Kharlamockhin, E.S., s e e Bachurin, V.I. 125
Khatamian, D. 215, 216, 219
Khatsernova, E.L., s e e Linetski, Ya.L. 153
Kholmedov, Kh.M., s e e Smirnov, I.A. 251,264
Kido, G., s e e Kaneko, T. 397
Kido, G., s e e Nimori, S. 350
Kido, G., s e e Ozeki, S. 65, 66
Kido, G., s e e Wada, H. 303,304
Kikuchi, M., s e e Kurosawa, Y. 42, 44, 51
Kikuchi, M., s e e Umehara, I. 53-58
Kilcoyne, S.H., s e e Mondal, S. 320
Kilcoyne, S.H., s e e Ritter, C. 321
Killoran, N., s e e Reinders, EH.E 30
Kim, D. 394
Kim, D., s e e Fischer, E 394, 411
Kim-Ngan, N.H., s e e Radwanski, R.J. 339, 342
Kim-Ngan, N.H., s e e Zhang, EY. 339, 340, 342,
343, 361
Kimura, T., s e e Iwata, N. 374, 375
King, C.A. 76, 78

Kipling, S.J., s e e Bryan, S.I. 179

Kirchmayr, H.R. 295, 296, 328, 369
Kirchmayr, H.R., s e e Burzo, E. 108, 144
Kirchmayr, H.R., s e e Gr6ssinger, R. 153
Kiriyama, H., s e e Numata, T. 156
Kisker, E., s e e Carbone, C. 124, 158
Kita, E., s e e Umemura, S. 159
Kitai, T. 388
Kitai, T., s e e Abe, S. 388
Kitai, T., s e e Kaneko, T. 388
Kitano, K., s e e Takeda, S. 173
Kitano, Y., s e e Komura, Y. 174
Kitaoka, Y., s e e Nakamura, H. 319
Kitazawa, H. 41-43, 60, 61, 63, 65
Kitazawa, H., s e e Kasuya, T. 13, 59, 60, 63-65
Kittel, C., s e e Ruderman, M.A. 3,415
Kjems, J.K., s e e Fischer, P. 394, 411
Klaasse, J.C.E, s e e Meyer, R.T.W. 51
Kiatt, K.H., s e e Bracconi, E 137
Kiavins, E 215, 221,225, 226, 229
Klein, H.E 311
Kletowski, Z. 41-43, 52
Knappe, E 225
Knappe, E, s e e Greis, O. 221,225, 231
Knappe, E, s e e Mfiller, H. 225
Knezo, D., s e e Dudfis, J. 120
Knoch, K.G. 152, 153
Knoch, K.G., s e e Fidler, J. 152, 153
Knoch, K.G., s e e Grieb, B. 152
Knorr, K. 225, 243,277
Kobayashi, H., s e e Ohkoshi, M. 311
Kobayashi, K.L.I., s e e Gerken, E 131
Kobayashi, T., s e e Sagasaki, M. 166
Koehler, W.C., s e e Gignoux, D. 301
Koelling, D.D. 4, 12, 13, 52, 55
Koelling, D.D., s e e Norman, M.R. 7, 55, 64
Koestler, C. 150, 151
Kofoed, J., s e e Chorkendorff, I. 116
Kohgi, M., s e e Takahashi, M. 70, 392
Kohn, W. 12, 14
Kohn, W., s e e Hohenberg, E 14
Kohori, Y., s e e Nakamura, H. 319
Kojima, T. 169
Kolaczkiewicz J. 122
Kolaczkiewicz J., s e e Bauer, E. 128
Kolar, D., s e e Holc, J. 150
Komatsu, H., s e e Goto, T. 300, 301,306
Komatsubara T. 31
Komatsubara T., s e e Asano, H. 91
Komatsubara T., s e e Endoh, D. 92, 93
Komatsubara T., s e e Fukuma, H. 394
Komatsubara T., s e e Ishii, T. 3
Komatsubara T., s e e Maezawa, K. 90
Komatsubara T., s e e Mitsuda, S. 389

AUTHOR INDEX
Komatsubara, T., s e e Satoh, K. 5, 94
Komatsubara, T., s e e Sumiyama, A. 4
Komatsubara, T., s e e Suzuki; T. 91
Komatsubara, T., s e e Takayanagi, S. 90, 94, 363,
364, 389, 390
Komatsubara, T., s e e Takeda, N. 65, 67
Komatsubara, T., s e e Tanaka, K. 65, 66
Komatsubara, T., s e e 0nuki, Y. 33-37, 87-93,
95, 390
Komura, Y. 173, 174
Komura, Y., s e e Horikoshi, H. 173
Komura, Y., s e e Kamimura, H. 324
Komura, Y., s e e Nakamura, H. 319
Komura, Y., s e e Takeda, S. 173
Konc, M., s e e Dudfis, J. 120
Kondo, J. 4
Korringa, J. 12
Korty, EW., s e e Brun, T.O. 394
Kosak, M.M. 118, 120
Kosevich, R.M., s e e Lifshitz, I.M. 27
Koshizuka, N., s e e Suzuki, Y. 162
Kost, M.E., s e e Bashldn, I.O. 225, 231
Kost, M.E., s e e Fedotov, V.K. 225
Kost, M.E., s e e Parshin, EE 243
Kost, M.E., s e e Volkenshtein, N.V. 271
Kostygov, A.N. 134
Kostygov, A.N., s e e Chizhov, EE. 134
Kostygov, A.N., s e e Morozov, Yu.G. 134
K6tzler, J., s e e Raffius, G. 355
Kou, X.C., s e e Gr6ssinger, R. 153
Kouvel, J.S., s e e Brun, T.O. 394
Kovanagi, T., s e e Koyama, M. 165
Kowalczyk, A., s e e Ratajczak, H. 157
Kowalewski, J., s e e Czopnik, A. 47
Kowalsky, W., s e e Kunze, U. 133
Koyama, M. 165
Koyoshi, Y., s e e Sugiyama, K. 365
Kozlovskii, L.V. 134, I36
Kozlowski, G., s e e Huang, C.Y. 185
Krause, L.J., s e e Carlin, R.L. 271,275, 277, 282,
284, 285
Krewenka, R., s e e Gr6ssinger, R. 153
Krexner, G., s e e Blaschko, O. 215, 219, 220, 234
Krill, G., s e e Brouder, C. 159, 165
Krill, G., s e e Guilmin, R 165
Krill, G., s e e Malterre, D. 182
Krishna, E, s e e Verma, A.R. 172
Krishnan, R. 166
Kristensson, D.K., s e e F~ldt, A. 115
Krizek, J. 126, 127
Kronmiiller, H., s e e Fidler, J. 152, 153
Kronmfiller, H., s e e Herbst, G. 233
Kronmfiller, H., s e e Knoch, K.G. 152
Kronmiiller, K., s e e Knoch, K.G. 152, 153

439

Krop, K., s e e Pszczola, J. 329

Krost, A. 188, 189
Krotenko, V.T., s e e Goremyehkin, E.A. 339, 342
Kiibler, J., s e e Niksch, M. 38, 39
Kfibler, J., s e e Sficht, J. 81
K~ibler, J., s e e Uhl, M. 326
Kubo, Y. 14, 36, 37
Kuboth, M. 185
Kuentgens, U., s e e Blancard, C. 112, 114
Kuentgens, U., s e e Sarpal, B.K. 112
Kuji, T., s e e Sakamoto, Y. 182
Kulikov, N.J. 243
Kulikova, I.N., s e e Smirnov, I.A. 251,264
Kumar, R. 136, 137
Kunii, S., s e e Effantin, J.M. 31,359, 397
Kunii, S., s e e Galera, R.M. 33
Kunii, S., s e e Goto, T. 33, 34
Kunii, S., s e e Komatsubara, T. 31
Kunii, S., s e e Morin, P. 397
Kunii, S., s e e Suglyama, K. 365
Kunii, S., s e e Suzuki, 1". 31, 33, 34
Kunz, C., s e e Gerken, E 115, 131
Kunze, U. 133
Kurisu, M., s e e Liu, W.L. 366
Kurisu, M., s e e Yamashita, M. 388
Kurosawa, Y. 42, 44, 51
Kurosawa, Y., s e e Maezawa, K. 83, 84, 86
Kurosawa, Y., s e e Satoh, K. 89, 90
Kurosawa, Y., s e e Umehara, I. 53-58
Kurosawa, Y . , s e e ()nuki, Y. 83, 84, 86-89,
91-93, 95
Kuruzar, D.L., s e e Allen, C.W. 174
Kusuda, T., s e e Honda, S. 164
Kusuya, T., s e e Komatsubara, T. 31
Kuwano, N. 182, 184
Kuzmenko, V.M. 118, 121, 122
Kwo, J. 119
Kwok, W.K., s e e C)nuki, Y. 33-37, 91-93, 95
Kwon, Y.S. 65, 67
Kwon, Y.S., s e e Kitazawa, H. 65
Kwon, Y.S., s e e Ozeki, S. 65, 66
Kwon, Y.S., s e e Settai, R. 60, 61, 65
Kwon, Y.S., s e e Takeda, N. 65, 67
Kwon, Y.S., s e e Tanaka, K. 65, 66
Labarta, A., s e e Badia, E 157
Labarta, A., s e e Martinez, B. 146
Lacerda, A., s e e van der Meulen, H.E 81
Lacroix, C., s e e Ballou, R. 325-327
Lacroix, C., s e e Due, N.H. 307, 308
Lacroix, C., s e e Lavagna, M. 308
Lacroix, C., s e e Nunez-Regueiro, M.D. 325
Lacroix, C., s e e Pinettes, C. 313-315, 326
Lacy, S.E., s e e Amato, A. 94, 95

440
Laesser, R., s e e Bracconi, R 137
Laforest, J., s e e Givord, D. 301,302
LaGraffe, D. 124
LaGraffe, D., s e e Dowben, RA. 124
Lambert, R.M., s e e Bryan, S.I. 179
Lambert, R.M., s e e Hay, C.M. 179
Lambert, R.M., s e e Jaffey, D.M. 116, 179
Lambert, R.M., s e e Jennings, J.R. 179
Lambert, R.M., s e e Nix, R.M. 179
Lambert, R.M., s e e Owen, G. 179
Lambrecht, A., s e e Carlin, R.L. 271,277
L a n , J., s e e Fang, R.Y. 157
Landau, L. 327
Landolt, M. 124
Landolt, M., s e e Cerri, A. 125, 140
Landolt, M., s e e Mauri, D. 122
LandoR, M., s e e Taborelli, M. 124
Lang, J.K. 115
Langell, M.A., s e e Jaswal, S.S. 151
Langford, H.D. 33
Lapertot, G., s e e Jerjini, M. 70, 392
Lapierre, E, s e e Amato, A. 94, 95
Larsen, J.W. 169, 171
Larson, A.C. 87
Larson, A.C., s e e Cromer, D.T. 174
LSsser, R. 265
L~ser, R., s e e Jung, E 220, 221,244, 245
Latka, K., s e e Czjzek, G. 366
Laubschat, C., s e e Domke, M. 114, 115
Lanbschat, C., s e e Kaindl, G. 115, 117
Laubschat, C., s e e Schneider, W.D. 118
Laughlin, D.E., s e e Cheng, S.E 155
Laughlin, D.E., s e e Shen, Y. 175
Lavagna, M. 308
Lavagna, M., s e e Cyrot, M. 297, 301
Lawrence, J.M. 40
Lawrence, J.M., s e e B6al-Monod, M.T. 304
Lazarev, B.G., s e e Kuzmenko, VM. 118
Lazarides, B., s e e Ahmed-Mokhtar, N. 127
Le Corre, A. 187, 188
Le Corre, A., s e e Caulet, J. 187
Le Corre, A., s e e Guivarc'h, A. 187
Le Corre, Y., s e e Krishnan, R. 166
Lea, K.R. 273
Leask, M.J.M., s e e Lea, K.R. 273
Lebech, B. 373
Lecrosnier, D., s e e Le Corre, A. 187
Lee, E.W. 303
Lee, R.W., s e e Herbst, J.H. 312
Lee, S.T., s e e Mason, M.G. 114
Legeler, B., s e e Seitz, E. 67~59
Legvold, S., s e e Ito, T. 268, 269
Lehmann, P., s e e Besnus, M.J. 74
Leiberich, A., s e e Lu, E 181

AUTHOR INDEX
Leisure, R.G. 232, 236
Lejay, P., s e e Fillion, G. 393
Lejay, P., s e e Rossat-Mignod, J. 81
Lejay, E, s e e van der Meulen, H.R 81
Leli~vre-Bema, E., s e e Fisher, R.A. 324
Leli~vre-Berna, E., s e e Nunez-Regueiro, M.D.
325
Leli~vre-Berna, E., s e e Voiron, J. 321
Lemaire, R. 299, 301,311,417
Lemaire, R., s e e Alameda, J.M. 311,312
Lemaire, R., s e e Ballou, R. 310, 313, 317, 319,
320, 322, 324, 336
Lemaire, R., s e e Burzo, E. 306
Lemaire, R., s e e B&le, C. 367
Lemaire, R., s e e D~portes, J. 321,322
Lemaire, R., s e e Gignoux, D. 83, 85, 299, 300,
302, 303, 308-310, 313, 362
Lemaire, R., s e e Givord, D. 301,302
Lemonnier, J.C., s e e Quemerais, A. 127, 129
Lengeler, B., s e e Hisser, R. 265
LePage, J.G. 159
Lethuillier, R 48
Lethuillier, P., s e e Bouvier, M. 365, 405
Levitin, R.Z., s e e Borombaev, M.K. 364
Levy, EM., s e e Fischer, E 394, 411
Levy, EM., s e e Kim, D. 394
Leyarovski, E., s e e Mrachkov, J. 352
UH&itier, R, s e e Zogal, O.J. 221,225
Li, D. 123
Li, D.X., s e e Nimori, S. 350
Li, E, s e e Li, L. 152
Li, H. 185
Li, H.D., s e e Ma, E. 143
Li, H.S. 328
Li, H.S., s e e Belorizky, E. 328-330, 337,
343-345
Li, L. 152
Li, Z. 110
Lian, K.C., s e e Allen, C.W. 174
Liberman, D. 9, 21
Libowitz, G.G. 210, 212, 214, 229, 240, 251,
255, 256
Libowitz, G.G., s e e Glinka, C.J. 243
Lichty, L.R. 232, 233, 244
Lichty, L.R., s e e Torgeson, D.R. 233
Lieder, M., s e e Forker, M. 231
Lieke, W., s e e Franse, J.J.M. 325
Lieke, W., s e e Gratz, E. 87
Lieke, W., s e e Steglich, F. 6
Li~nard, A., s e e Alameda, J.M. 145
Lifshitz, E., s e e Landau, L. 327
Lifshitz, I.M. 27
Ligeon, E., s e e Danielou, R. 216
Lii, Z.Y. 156

AUTHOR INDEX
Lilienfeld, D.A., s e e Borgensen, P. 180
Lin, H. 379
Lin, Z., s e e Li, L. 152
Lindau, I., s e e Allen, J.W. 115
Lindau, I., s e e EgelhoffJr, W.E 131
Lindau, I., s e e Hecht, M.H. 131
Lindau, I., s e e Johansson, L.I. 115, 131
Lindau, I., s e e Rossi, G. 130
Lindau, I., s e e Yeh, J.J. 130
Linetski, Ya.L. 153, 174
Lippold, B., s e e Alekseev, RA. 339, 341
Lippold, B., s e e Goremychkin, E.A. 339, 342
Lippold, B., s e e Moth, Th. 178
Liu, B., s e e Ebihara, T. 48, 49, 51
Liu, B.X., s e e Ma, E. 143
Liu, N.C. 155
Liu, N.C., s e e Kamprath, N. 155
Liu, N.C., s e e Stadelmaier, H.H. 155
Liu, W.L. 366
Liu, Z.X., s e e Fang, R.Y. 157
Livesay, B.R., s e e Larsen, J.W. 169, 171
Lloyd, D., s e e Owen, G. 179
Loboda, VB. 120, 134, 137
Lochoshvili, T.S., s e e Dadiani, T.O. 188
Lochoshvili, T.S., s e e Glurdzhidze, L.N. 188
Loebich Jr, O. 182
Loebich Jr, O., s e e Jordan, R.G. 182
Loier, C., s e e Dexpert-Ghys, J. 137
Loisel, B., s e e Quemerais, A. t27, 129
Longuet-Higgins, H.C. 32
Lonzarich, G.G. 69, 70, 76-78
Lonzarich, G.G., s e e King, C.A. 76, 78
Lonzarich, G.G., s e e Taillefer, L. 26
Lord, D.G., s e e Parker, S.EH. 153
Loretto, M.H., s e e Brooks, J.W. 182
Loucks, T.L. 12, 20
Lousa, A., s e e Badia, E 157
Lousa, A., s e e Ferrater, C. 158
Lowe, W.R 125
Lozovyi, Ya.B. 119
Lozovyi, Ya.B., s e e Gonchar, EM. 118
Lu, E 181
Lu, E, s e e Croft, M. 181
Lu, Q., s e e Alameda, J.M. 311, 312
Lu, Q., s e e Givord, D. 302
Lu, Q.Z., s e e Huang, G.X. 150
Lfibcke, M. 114
Lfibcke, M., s e e Niemann, W. 114
Lubitz, R, s e e Forester, D.W. 168
Lubitz, E, s e e Schelleng, J.H. 167
Lubitz, R, s e e Vittoria, C. 168
Lucasson, A. 241,243,251,252
Lucasson, A., s e e Burger, J.E 243, 251,252, 271,
275, 285

441

Lucasson, A., s e e Daou, J.N. 213, 219, 220, 229,

233, 238, 239, 241,243-245, 247, 248, 251,
252, 268, 269
Lucasson, A., s e e Vajda, P. 213,220, 227, 233,
244-246
Lucasson, E, s e e Daou, J.N. 219, 220, 233,234,
244, 245, 247, 268, 269
Lucifiski, T., s e e Ratajczak, H. 157
Lundqvist, B.I., s e e Gunnarsson, O. 15
Luong, N.H. 90, 365
Ltithi, B. 357, 358
Ltithi, B., s e e Niksch, M. 38, 39
Luttinger, J.M. 11
Lynch, D.W., s e e Weaver, J.H. 127
Lysenko, A.B., s e e Tkach, VI. 110
Ma, E. 143
Ma, R.Z., s e e Pan, S.M. 175
Ma, R.Z., s e e Zhao, Z.B. 152
Ma, S.K., s e e B~al-Monod, M.T. 305
Machii, Y., s e e ()nuki, Y. 87
MacKenzie, I.S., s e e Waind, RR. 338
Mackintosh, A.R. 390
Mackintosh, A.R., s e e Jensen, J. 266, 379, 390,
391,412, 413, 416, 417
Mackintosh, A.R., s e e McEwen, K.A. 350, 351
Mgdge, H., s e e Czopnik, A. 398
Maeland, A., s e e Glinka, C.J. 243
Maeland, A., s e e Libowitz, G.G. 210, 214, 240
Maeland, A.J. 243
Maeno, Y., s e e Satoh, K. 5, 94
Maezawa, K. 83, 84, 86, 90
Maezawa, K., s e e Fukuhara, T. 74-76
Maezawa, K., s e e 0nuki, Y. 83, 84, 86, 91-93, 95
Mahmoud, S. 118, 127
Maienschein, J.L. 140
Maier-Komor, E 125
Maines, R.G., s e e Jayaraman, A. 7
Maita, J.R, s e e Fawcett, E. 26
Malinowski, M.E. 141
Malozemoff, A.R, s e e De Luca, J.C. 180
Malterre, D. 182
Malzfeldt, W . , s e e Niemann, W. 114
Manchester, ED., s e e Khatamian, D. 215, 216
Mancini, E 184
Mancini, E, s e e Huang, C.Y. 185
Mandel, T., s e e Dornke, M. 114, 115
Manedov, A.I., s e e Efendiev, E.G. 187
Mangin, Ph., s e e Cherifi, K. 160
Manoubi, T. 132
Manoubi, T., s e e Coltiex, C. 132
Mansmann, W. 231
Maple, M.B., s e e Falicov, L.M. 4
Maple, M.B., s e e Huang, C.Y. 185

442

AUTHOR INDEX

Maranzana, EE., s e e Buschow, K.H.J. 4

Marazza, R., s e e Rossi, D. 184
March, N.H., s e e Callaway, J. 9
Marchal, G., s e e Baczewski, L.T. 159, 161
Marchal, G., s e e Brouder, C. 159, 165
Marchal, G., s e e Cherifl, K. 160
Marchal, G., s e e Guilmin, P. 165
Marchal, G., s e e Malterre, D. 182
Marchal, G., s e e Piecuch, M. 159, 161
Mariko, H., s e e Nakamura, T. 162
Marinero, E.E. 166
Markandeylu, G., s e e Annapoorni, S. 168
Markin, P.E., s e e Svoboda, R 387
Markosyan, A.S., s e e Borombaev, M.K. 364
Markosyan, A.S., s e e Brown, P.J. 321
Markova, I.A., s e e Alekseev, P.A. 339, 341
Marshall, A.N. 164
Marshall, A.N., s e e Webb, D.J. 163
Mfirtensson, N., s e e Andersen, J.N. 116
Mfirtensson, N., s e e Hillebrecht, EU. 63
M~rtensson, N., s e e Kaindl, G. 115, 117
Mfirtensson, N., s e e Nilsson, A. 116
ML,'tensson, N., s e e Stenborg, A. 117
Martin, M. 177
Martin, M., s e e Frigerio, J.M. 177
Martinez, B. 146
Martinez, B., s e e Badia, E 157
Martinez, B., s e e Ferrater, C. 158
Martynyuk, A.V., s e e Gorodetskii, D.A. 119
Maruno, S., s e e Sakamoto, I. 73, 74
Mason, M.G. 114
Massov, A., s e e Br&hignac, C. 114
Masuda, H., s e e Suzuki, Y. 162
Materlik, G. 112
Mathevet, J.R 180
Matsubara, K., s e e Koyama, M. 165
Matsuda, R., s e e Suzuki, A. 166, 168
Matsui, H., s e e Goto, T. 65, 66
Matsumoto, H., s e e Huang, C.Y. 185
Matsuttra, S., s e e Ishida, A. 188, 189
Mattheiss, L.E 21, 22
Matthew, J.A.D. 132
Matthew, J.A.D., s e e Netzer, EE 108
Matz, W., s e e Alekseev, RA. 339, 341
Mauri, D. 122
Mauri, D., s e e Cerri, A. 125, 140
Maurice, V . , s e e Melmed, A.J. 118, 119
Mayer, H.M., s e e Franse, J.J.M. 325
MeAlister, S.R 366
McCausland, M.A.H., s e e Waind, ER. 338
McEwen, K.A. 350, 351,391
McGuiness, RJ. 148
McGuire, T.R., s e e Coey, J.M.D. 178
MeGuire, T.R., s e e Gambino, R.J. 180

McGuire, T.R., s e e Hartmann, M. 180

McGuire, T.R., s e e Pickart, S.J. 178
McGuire, T.R., s e e von Molnar, S. 178, 180
Mclntyre, GJ., s e e Jehan, D.A. 379
McMasters, O.D., s e e Gschneidner Jr, K.A. 6, 52
McMasters, O.D., s e e Ikeda, K. 305
McMasters, O.D., s e e Johanson, W.R. 52, 53
McMirm, R. 134
McMorrow, D.E, s e e Jehan, D.A. 379
McWhan, D.B., s e e Lowe, W.P. 125
Medveded, V.K., s e e Gonchar, EM. 118
Medveded, V.K., s e e Lozovyi, Ya.B. 119
Meeson, P., s e e Chapman, S.B. 92, 94, 95
Meeson, P., s e e Hunt, M. 81, 82
Meeson, P., s e e Satoh, K. 42, 44, 48, 51
Megtert, S., s e e Mathevet, J.R 180
Mehrhoff, T.K. 144
Mekata, M., s e e Wada, H. 303, 304
Mekata, M., s e e Yoshimura, K. 301
Melczer, M.E., s e e Croft, M. 181
M~linon, D., s e e Rayane, D. 114
Melmed, A.J. 118, 119
Melmed, A.J., s e e Ciszewski, A. 118
Melnikov, V.I., s e e Kuzmenko, V.M. 118, 121,
122
Melton, K.N. 174
Mendes, P.J., s e e Ferreira, R 260
Mendia-Monterroso, R., s e e Ballou, R. 310
Mendia-Monterroso, R., s e e Gignoux, D. 309,
310
Menovsky, A.A., s e e Franse, J.J.M. 325, 336
Menovsky, A.A., s e e Sinnema, S. 336, 337
Menth, A., s e e Klein, H.P. 311
Meschede, D., s e e Steglich, E 6
Messer, C.E. 240
Methfessel, S., s e e Gu, B.X. 145
Methfessel, S., s e e Homburg, H. 145
Metzger, T.H. 236
Meyer, A., s e e Besnus, M.J. 74
Meyer, E., s e e Grtitter, P. 148
Meyer, E., s e e Heirlzelmann, H. 148
Meyer, R.T.W. 51
Mibu, K. 158
Mibu, K., s e e Hosoito, N. 158
Mibu, K., s e e Shinjo, T. 158
Mibu, K., s e e Yoden, K. 158
Mieeli, RE, s e e Palmstrom, C.J. 186, 187
Michelutti, B., s e e Aubert, G. 339
Miedema, A.B., s e e Buschow, K.H.J. 108
Migliori, A., s e e Leisure, R.G. 232, 236
Mihalisin, T., s e e Parks, R.D. 181
Mikhailova, N.R, s e e Savrin, V.D. 109
Miles, M.H., s e e Chen, D.Y. 132
Miller, A.E., s e e Allen, C.W. 174

AUTHOR INDEX
Miller, R.E 118, 126, 127
Miller, R.E, s e e Rahman Khan, M.S. 127, 135
Miller, S.A. 132
Min, B.I. 36
Minakata, R. 303
Minc, K., s e e Sakamoto, Y. 182
Ming Lei, s e e Leisure, R.G. 236
Minier, M., s e e Le Corre, A. 187
Mintz, M.H. 213, 240
Mintz, M.H., s e e Jacob, I. 241
Mintz, M.H., s e e Kumar, R. 136, 137
Mirabal-Garcia, M., s e e Salas, EH. 123
Miroshnichenko, I.S., s e e Tkach, V.I. 110
Misawa, S. 305
Mitchell, DJ. 141, 142
Mitsuda, S. 389
Mittsev, M.A. 119
Mittsev, M.A., s e e Burmistrova, O.R 119
Miura, T., s e e Sakamoto, I. 70, 72-74
Miura, T., s e e Takahashi, M. 70, 392
Miyahara, S., s e e Harima, H. 89
Miyahara, T. 129, 130
Miyahara, T., s e e Ishii, T. 3
Miyarna, T., s e e Inoue, A. 166
Miyamoto, K. 157
Miyamoto, T., s e e Sakamoto, I. 73, 74
Miyatake, H., s e e Tsunashima, S. 168
Miyazaki, T. 162
Miyoshi, K., s e e Sakamoto, I. 70, 72-74
Mizoguchi, T., s e e Sahashi, M. 156, 174
Mizuno, H., s e e Ishida, A. 188, 189
Mizushima, T., s e e Isikawa, Y. 85
Mizutani, S., s e e Suzuki, A. t66, 168
Moch, Th. 178
Mokeda, M., s e e Yamashita, S. 155
Moldovan, A . G . , s e e Smith, H.K. 137
Molho, R, s e e Ballou, R. 308
Molho, P., s e e Gignoux, D. 303, 308, 309
Moncton, D.E., s e e Bohr, L 417
Moncton, D.E., s e e Gibbs, D. 268, 378, 417
Mondal, S. 320
Montenegro, J.ED., s e e del Moral, A. 355
Morfin-L6pez, J.L., s e e Aguilera-Granja, E 122
Moreau, J.M. 182
Moreau, J.M., s e e Gignoux, D. 309, 310
Moreau, LM., s e e Parth~, E. 174
Moreira, J.M., s e e Freitas, P.P. 185
Moreu, M.A., s e e Martinez, B. 146
Morgen, E, s e e Onsgaard, J. 111, 126, 128-131
Moil, K., s e e Isikaw~, Y. 85
Moil, K., s e e Sato, K. 362, 363
Mori, T., s e e Ishii, T. 3
Morin, R 338, 353, 355-360, 394, 396-399, 408,
414, 417

443

Morin, P., s e e Al~onard, R. 355, 357, 394, 395,

397
Mofin, P., s e e Amara, M. 398
Morin, P., s e e Galera, R.M. 33
Morin, E, s e e Gignoux, D. 366, 389, 406
Morin, P., s e e Jaussaud, C. 355, 356
Morin, E, s e e Ltithi, B. 357, 358
Morishita, T. 157, 159
Morishita, T., s e e Kajiura, M. 157
Morishita, T., s e e Sato, R. 166
Moriya, T. 305
Moriya, T., s e e Ueda, K. 324
Morozov, Yu.G. 134
Morozov, Yu.G., s e e Kostygov, A.N. 134
Morozova, L.V, s e e Varkanova, R.G. 136, !42
Morrison, G.R., s e e Chapman, J.W. 126
Moser, H.R. 130--132
Moser, E, s e e Vajda, E 219, 232, 233
Moshchalkov, V.V., s e e Brandt, N.B. 5
M6ssbauer, R.L., s e e Petrich, G. 301
Motokawa, M. 380
Motokawa, M., s e e Nojiri, H. 381
Motoya, K. 323
Motoya, K., s e e Freltoft, T. 323
Motoya, K., s e e Shigeoka, T. 383
Mounier, S., s e e Palmstram, C.J. 186, 187
Mounier, S., s e e Zhu, J.G. 186
Mowry, G.S., s e e Thome, D.K. 236, 241, 246,
247, 252
Mrachkov, J. 352
Miieller, EM., s e e Arko, A.J. 31, 33
Mueller, M.H., s e e Shaked, H. 27t, 275, 281-284
Mtihle, E., s e e Goremychkin, E.A. 339, 342
Mukhuchev, A.M., s e e Burmistrova, O.R 119
Mukhuchev, A.M., s e e Mittsev, M.A. 119
Mulford, R.N.R., s e e Sturdy, G.E. 231
Muller, EA., s e e Franse, J.J.M. 336
Mfiller, H. 225
Mtiller, H., s e e Greis, O. 221,225, 231
M/iller, H., s e e Knappe, R 225
Mtiller, J. 140
Muller, J., s e e Kuboth, M. 185
M~iller, J.E., s e e Materlik, G. 112
M~ller, T. 35, 36
Murakami, M., s e e Uchida, K. 178
Murani, A.P., s e e Ball, A.R. 403-405
Murao, T., s e e Tsuneto, T. 353
Muraoka, Y. 303
Murata, K., s e e Goto, T. 300, 301,306
Murata, K.K. 327
Murgai, V. 129, 130
Murgai, V, s e e Huang, Y.S. 179
Murgai, V., s e e Parks, R.D. 181
Murgai, V, s e e den Boer, M.L. 115, 130

444

AUTHOR1NDEX

Mustaehi, A. 279
Mutka, H., s e e Arons, R.R. 273
Myers, H.P., s e e Fgldt, ,~,. 115, 116
Naberhuis, S., s e e Anthony, T.C. 168
Nabli, H,, s e e Pieeueh, M. 159, 161
Naris, S., s e e Shan, Z.S. 158
Nagahama, K., s e e lwamura, E. 150
Nagai, H., s e e Adachi, G. 170-172
Nagai, H., s e e Sakaguchi, H. 169-172
Nagai, N., s e e Adaehi, G. 170, 172
Nagai, N., s e e Kurosawa, Y. 42, 44, 51
Nagai, N., s e e Satoh, K. 42, 44, 48, 51
Nagai, N., s e e Umehara, I. 29, 30, 41, 42, 47-58,
70-73
Nagai, N., s e e Onuki, Y. 91-93
Nagano, H., s e e Sttmiyama, A. 4
Nagao, K., s e e Adaehi, G. 140
Nagaoka, Y., s e e Sakamoto, Y. 182
Nagasawa, S., s e e Shimizu, M. 297
Nagata, H., s e e Yamamoto, H. 151, 152
Nagel, H., s e e MeGuiness, RJ. 148
Nagel, H., s e e Melton, K.N. 174
Nagel, H., s e e Perkins, R.S. 311, 312
Naidyuk, Y.G., s e e Reiffers, M. 339-341,351
Nait-Saada, A., s e e Aubert, G. 339
Nait-Saada, A., s e e Barthem, V.M.T.S. 339-341,
343, 344
Nait-Saada, A., s e e Gignoux, D. 339, 342
Naito, K., s e e Ishii, T. 3
Nakagawa, Y., s e e Abe, S. 388
Nakagawa, Y., s e e Kaneko, T. 388, 397
Nakagawa, Y., s e e Kitai, T. 388
Nakahara, J., s e e Goltros, W. 188
Nakahara, N., s e e Ishida, A. 188, 189
Nakahara, S., s e e Kwo, J. 119
Nakajima, K. 178
Nakajima, T., s e e Ishizawa, Y. 31, 33, 35
Nakamura, H. 319
Nakamura, H., s e e Kamimura, H. 324
Nakamura, H., s e e Shiga, M. 323, 324
Nakamura, H., s e e Wada, H. 317-319, 323-325
Nakamura, O. 118, 125
Nakamura, O., s e e Baba, K. 188
Nakamura, O., s e e Takeda, K. 133
Nakamura, T. 162
Nakamura, Y. 304, 320
Nakamura, Y., s e e Ballou, R. 319, 320, 322
Nakamura, Y., s e e Hirosawa, S. 313
Nakamura, Y., s e e Kamimura, H. 324
Nakamura, Y., s e e Minakata, R. 303
Nakamura, Y., s e e Mttraoka, Y. 303
Nakamura, Y., s e e Nakamura, H. 319
Nakamura, Y., s e e Oomi, G. 321

Nakamura, Y., s e e Sakakibara, T. 300, 301

Nakamura, Y., s e e Shiga, M. 323, 324
Nakamura, Y., s e e Tanaka, Y. 303
Nakamura, Y., s e e Wada, H. 303, 304, 317-319,
323-325
Nakamura, Y., s e e Yoshie, H. 311
Nakamura, Y., s e e Yoshimura, K. 306, 321,323
Nakao, T., s e e GondS, Y. 167
Nakao, T., s e e Suezawa, Y. 167
Namoradze, N.Z., s e e Ratishvili, I.G. 227
Naoe, M. 168
Naoe, M., s e e Hoshi, Y. 168
Narayanamurti, V., s e e Jayaraman, A. 7
Nasu, S. 40
Natkaniec, I., s e e Goremyehkin, E.A. 339
Natsui, H., s e e Settai, R. 87-89
Nawate, M., s e e Tsunashima, S. 163
Nawate, N., s e e Honda, S. 164
Nazareth, A., s e e Strzeszewski, J. 151
Nazarov, A.S., s e e Varkanova, R.G. 142
Ndjaka, J.M.B., s e e Dieny, B. 165
Needham, D., s e e Eley, D.D. 136, 140
Neifield, R., s e e Croft, M. 181
Netzer, EP. 108, 111, 143, 144
Netzer, ER, s e e Matthew, J.A.D. 132
Neumann, H.H., s e e Nasu, S. 40
N6vot, L., s e e Gasgnier, M. 109
Newns, D.M., s e e Strange, D. 14, 55
Nieki, K., s e e Adachi, G. 170, 172
Nieklow, R.M., s e e Arons, R.R. 273, 277
Niemann, W. 114
Niemann, W., s e e LiJbcke, M. 114
Niihara, T. 168
Niki, K., s e e Adaehi, G. 170, 172
Niki, K., s e e Sakaguehi, H. 169, 170
Nikseh, M. 38, 39
Nilsson, A. 116
Nilsson, A., s e e Andersen, J.N. 116
Nilsson, A., s e e Stenborg, A. 117
Nilsson, P.O., s e e Kanski, J. 131
Nimori, S. 350
Ninomiya, N., s e e Kawahata, T. 136
Nishi, M, s e e Shigeoka, T. 383
Nishiguehi, I., s e e Hosoda, N. 136
Nishihara, M., s e e 0nuki, Y. 33-37, 90-93, 390
Nishihara, Y., s e e Katayama, T. 166
Nishikawa, M., s e e Iwata, N. 401
Nix, R.M. 179
Nix, R.M., s e e Bryan, S.I. 179
Nix, R.M., s e e Hay, C.M. 179
Nix, R.M., s e e Jennings, J.R. 179
Nix, R.M., s e e Owen, G. 179
Noce, C., s e e Maneini, E 184
Nogami, J., s e e Rossi, G. 130

AUTHOR INDEX
Nogami, J., s e e Yeh, J.J. 130
Nojiri, H. 381
Nojiri, H., s e e Motokawa, M. 380
Nordstr6m, L. 313, 314
Norman, M., s e e Chapman, S.B. 92, 95
Norman, M.R. 7, 55, 64
Notarys, H., s e e Marinero, E.E. 166
Nozaki, H. 365
Nozaki, H., s e e Ishizawa, Y. 31, 33, 35
Nozi6res, E, s e e Pines, D. 327
Numata, T. 156
Numata, T., s e e Lii, Z.Y. 156
Nunez-Regueiro, M.D. 325
Ntmez-Regueiro, M.D., s e e Ballou, R. 325-327
Nurmiko, A.V, s e e Goltros, W. 188
Nuttall, R.H.D., s e e Shin, S.C. 160
Nyholm, R. 116
Nyholm, R., s e e Chorkendorff, I. 116
Obradors, X., s e e Martinez, B. 146
Ochiai, Y., s e e Hashimoto, S. 168
Oda, Y., s e e Sumiyama, A. 4
Oddou, J.L. 321
Oesterreicher, H. 137
Oestreieh, V., s e e Czjzek, G. 366
Ogawa, S., s e e Shinjo, T. 158
Oguro, I., s e e Kitazawa, H. 60, 63
Oguro, I., s e e Komatsubara, T. 31
O g u r o , I., s e e Sakamoto, I. 73, 74
Oguro, I., s e e Shigeoka, T. 383
O g u r o , I., s e e Yamashita, M. 388
O'Handley, R.C., s e e Li, H. 185
Ohashi, M., s e e Abe, S. 388
Ohashi, M., s e e Kaneko, T. 388
Ohe, Y., s e e Goto, T. 33, 34
Ohe, Y., s e e Settai, R. 87-89
Ohe, Y., s e e Suzuki, T. 31, 33
Ohkawa, EJ. 308
Ohki, C. 239, 240
Ohkoshi, M. 311
Ohkoshi, M., s e e Honda, S. 164
Ohkuma, H., s e e Miyahara, T. 129, 130
Ohmaki, M., s e e Sakamoto, Y. 182, 184
Ohtani, Y., s e e Hosoda, N. 136
Ohtani, Y., s e e Kawahata, T. 136
Ohyama, R., s e e Koyama, M. 165
Okamoto, T. 360, 361
Okamoto, T., s e e Hashimoto, Y. 387
Okamoto, T., s e e Shigeoka, T. 380, 401
Okamura, T., s e e Ishida, A. 188, 189
Okazaki, M., s e e Onodera, Y. 12
O l d , K . , s e e Kuwano, N.
182, 184
Okuda, H., s e e Muraoka, Y. 303
Okuno, H. 163

445

Okuno, H., s e e Lii, Z.Y. 156

Olsen, C.E., s e e Huang, C.Y. 185
Olsen, J.A., s e e Amato, A. 94, 95
Olson, C.G., s e e Weaver, J.H. 126-129
Olson, C.G., s e e Wieliczka, D.M. 115, 116
Omi, T., s e e 0nuki, Y. 33-37, 87-89, 91-93, 95
Onaya, T., s e e Sato, N. !60, 164
Onellion, M., s e e Dowben, P.A. 124
Onellion, M., s e e LaGraffe, D. 124
Onellion, M., s e e Li, D. 123
Ono, M., s e e Tomiyama, E 336
Onodera, Y. 12
Onsgaard, J. 111, 116, 117, 126-131
Onsgaard, J., s e e Andersen, J.N. 115, 116
Onsgaard, J., s e e Chorkendorff, I. 114, 116
Onsgaard, J., s e e Nilsson, A. 116
0nuki, Y. 7, 33-37, 74-78, 80, 83, 84, 86-93,
95, 390
0nuki, Y., s e e Aoki, H. 76-78
O n u k i , Y . , s e e Asano, H. 91
0nuki, Y., s e e Ebihara, T. 42, 44-46, 48, 49, 51
Onuki, Y., s e e Endoh, D. 92, 93
Onuki, Y., s e e Ishii, T. 3
Onuki, Y., s e e Komatsubara, T. 31
0nuki, Y., s e e Kurosawa, Y. 42, 44, 51
0nuki, Y., s e e Maezawa, K. 83, 84, 86, 90
0nuki, Y., s e e Mitsuda, S. 389
0nuki, Y., s e e Satoh, K. 5, 42, 44, 48, 51, 89, 90,
94
0nuki, Y., s e e Settai, R. 52, 53, 87-89
0nuki, Y., s e e Sumiyama, A. 4
0nuki, Y., s e e Suzuki, T. 91
0nuki, Y., s e e Takayanagi, S. 90, 94, 363, 364,
389, 390
0nuki, Y., s e e Umehara, I. 29, 30, 41, 42, 47-58,
70-73
Oomi, G. 321
Opyrchal, J. 278
Opyrchal, J., s e e Bieganski, Z. 282
Opyrchal, J., s e e Drulis, M. 282
Ormerod, J., s e e McGuiness, RJ. 148
Osborn, R., s e e Bennington, S.M. 234, 269
Oskotski, V.S., s e e Smirnov, I.A. 251,264
Osterwalder, J., s e e Schefer, J. 225, 273, 277
Otaki, K., s e e Sakai, O. 60, 61, 63, 64
Ott, H.R., s e e Andres, K. 339, 341
Ott, H.R., s e e Schlapbach, L. 252, 257
Ouladdiaf, B. 317, 318
Ouladdiaf, B., s e e Ballou, R. 317, 319, 320, 322,
325, 327
Ouladdiaf, B., s e e Brown, P.J. 321
Ouladdiaf, B., s e e D6portes, J. 321-323
Ouladdiaf, B., s e e Oddou, J.L. 321
Ousset, J.C., s e e Ballou, R. 336

446

AUTHOR INDEX

Owen, G. 179
Owen, G., s e e Bryan, S.I. 179
Owen, G., s e e Hay, C.M. 179
Owen, G., s e e Jennings, J.R. 179
Owen, G., s e e Nix, R.M. 179
Oyamada, A., s e e Kitazawa, H. 65
Oyamada, A . , s e e Takeda, N. 67
Ozeki, S. 65, 66
Paccard, D., s e e B~cle, C. 367
Pack, J.G., s e e Libowitz, G.G. 251,256
Pa'idassi, S., s e e Cochet-Muchy, D. 151
Palenzona, A. 182
Palmer, P., s e e L/isser, R. 265
Palmstrom, C.J., s e e Allen Jr, S.J. 186, 187
Palmstrom, C.J. 186, 187
Palmstrom, C.J., s e e Zhu, J.G. 186
Palyukh, B.M., s e e Lozovyi, Ya.B. 119
Pan, S.M. 175
Pan, S.M., s e e Zhao, Z.B. 152
Pan6i~in, R.S., s e e Kosak, M.M. 118, 120
Parfen'eva, L.S., s e e Smirnov, I.A. 251,264
Park, M.K., s e e Messer, C.E. 240
Parker, D.G., s e e Jennings, J.R. 179
Parker, S.EH. 153
Parks, R.D. 4, 181
Parks, R.D., s e e Croft, M. 181
Parmigiani, E 134
Parshin, P.P. 243
Parth~, E. 174
Parth~, E., s e e Moreau, J.M. 182
Partin, D.L. 188-190
Partin, D.L., s e e Goltros, W. 188
Partin, D.L., s e e Salamanca-Young, L. 189
Parvin, K., s e e Webb, D.J. 162
Pasechnik, M.V., s e e Goremychkin, E.A. 339,
342
Patrick, R.C., s e e Mitchell, D.J. 141, 142
Pebler, A. 225, 231
Pecharsky, V.K. 379
Peuncey, T., s e e Kaindl, G. 117
Perenboom, J.A.A.J., s e e van der Meulen, H.P. 81
Perkins, R.S. 311,312
Perkins, R.S., s e e Klein, H.P. 311
Perkins, R.S., s e e Melton, K.N. 174
Perrier de la Bathie, R., s e e Gignoux, D. 339, 342
Peshkov, A.V, s e e Zhavoronkova, K.N. 136
Peterman, D.J., s e e Franciosi, A. 132
Peterman, D.J., s e e Raisanen, A. 132
Peterman, D.J., s e e Wall, A. 132
Peterson, D.T., s e e Kai, K. 229, 241,252, 256
Peterson, D.T., s e e Phua, T.T. 237
Peterson, D.T., s e e Shinar, J. 251,255
Peterson, D.T., s e e Weaver, J.H. 265

Peterson, D.T., s e e Zamir, D. 256

Petinov, V.I., s e e Chizhov, RE. 134
Pefinov, VI., s e e Kostygov, A.N. 134
Petinov, VI., s e e Morozov, Yu.G. 134
P&rakian, J.P., s e e Ahmed-Mokhtar, N. 127
Petdch, G. 301
Petroff, Y., s e e Schlapbach, L. 252, 257, 265, 266
Petrov, A.E., s e e Morozov, Yu.G. 134
Petzow, G., s e e Grieb, B. 152
Petzow, G., s e e Knoch, K.G. 152, 153
Philip, P., s e e Franciosi, A. 132
Philip, P., s e e Raisanen, A. 132
Philip, E, s e e Wall, A. 132
Philip, R., s e e Ahmed-Mokhtar, N. 127
Phillips, N.E., s e e Amato, A. 94, 95
Phillips, N.E., s e e Fisher, R.A. 324
Phua, T.T. 237
Phua, T.T., s e e Zamir, D. 256
Piacentini, M., s e e Sigrist, M. 126, 128
Pickart, S.J. 178
Piecuch, M. 159, 161
Piecuch, M., s e e Baczewski, L.T. 159, 161
Piecueh, M., s e e Brouder, C. 159, 165
Pientka, Z., s e e CereS,, S. 143
Pierre, J., s e e Lethuillier, P. 48
Pierre, J., s e e Morin, E 397
Pillmayr, N., s e e Daou, J.N. 248, 251,269
Pillmayr, N., s e e Vajda, P. 245, 269, 270
Pines, D. 327
Pinettes, C. 313-315, 326
Pinettes, C., s e e Due, N.H. 307, 308
Ping, J., s e e Zhao, Z.B. 152
Pintschovius, L., s e e Blaschko, O. 234
Plaskett, T.S., s e e Freitas, ER 185
Plaskett, T.S., s e e Kaindl, G. 117
Pleschiutschnig, J. 234
Pleschiutschnig, J., s e e Blaschko, O. 219, 220,
234
Poate, J.M. 167
Poldy, C.A., s e e Klrchmayr, H.R. 295, 296, 328,
369
Pollack, R.A., s e e Kaindl, G. 117
Pollard, R.J., s e e Parker, S.EH. 153
Pols, R.E., s e e van Deursen, A.J.R 33, 34, 36
Ponyatovskii, E.G., s e e Bashkin, I.O. 225, 231
Ponyatovskii, E.G., s e e Fedotov, VK. 225
Popeseu, M., s e e Goremychkin, E.A. 339
P6rschke, E., s e e Bracconi, R 137
Porte, M., s e e Krishnan, R. 166
Port, R., s e e Czopnik, A. 398
Potzel, W., s e e Waibel, E 285
Pourarian, E, s e e Lee, E.W. 303
Pradal, E, s e e Frandon, J. 128-130
Prietsch, M., s e e Domke, M. 114, 115

AUTHOR INDEX
Probst, P.-A., s e e Hunt, M. 81, 82
Probst, P.-A., s e e Satoh, K. 42, 44, 48, 51
Protsenko, I.E. 120
Protsenko, I.E., s e e Loboda, V.B. 120, 134, 137
Provo, J.L. 141
Provo, J.L., s e e Harris, J.M. 177
Provo, J.L., s e e Mitchell, D.J. 142
Pszczola, J. 329
Pulham, R.J., s e e Hubberstey, R 140
Purwins, H.G. 360

Qian, X.R., s e e Kamprath, N. 155

Qiu, M., s e e Li, H. 185
Quang, RH., s e e Radwanski, R.J. 334, 335
Quang, EH., s e e Verhoef, R. 333
Quemerais, A. 127, 129
Quezel, S., s e e Rossat-Mignod, J. 401

Raaen, S. 180
Raaen, S., s e e Braaten, N.A. 178
Raaen, S., s e e Croft, M. 181
Raaen, S., s e e Parks, R.D. 181
Rabalais, J.W., s e e Kumar, R. 136, 137
Rabe, R, s e e Lfibcke, M. 114
Rabe, E, s e e Niemann, W. 114
Radhakrishna, R, s e e Daou, J.N. 216, 268, 269
Radhakrishna, R, s e e Vajda, E 268, 269
Radwanski, R.J. 331,334-336, 339, 342
Radwanski, R.J., s e e Franse, J.J.M. 328, 336, 350,
417
Radwanski, R.J., s e e Sinnema, S. 329, 336, 337
Radwanski, R.J., s e e Szytuta, A. 384
Radwanski, RJ., s e e Verhoef, R. 331-333
Radwanski, R.J., s e e Zhang, EY. 339, 340, 342,
343, 361
Raffius, G. 355
Rahman Khan, M.S. 127, 135, 137
Raigorodskii, R.M., s e e Linetski, Ya.L. 153
Rainford, B.D., s e e Mondal, S. 320
Raisanen, A. 132
Raisanen, A., s e e Franciosi, A. 132
Raisanen, A., s e e Wall, A. 132
Rajora, O.S. 133
Rakhubovskii, V.A., s e e Kuzmenko, V.M. 118,
122
Rakoto, H., s e e Ballou, R. 336
Rama Ran, K.MS., s e e Annapoorni, S. 168
Ramakrisna, K. 171
Ramesh, R. 150, 151
Ramesh, R., s e e Koestler, C. 150, 151
Ran, X.L., s e e Fang, R.Y. 157
Ratajczak, H. 157

447

Ratajczak, H., s e e Dud~s, J. 120

Rath, J. 22
Ratishvili, I.G. 227
Rather, E.R., s e e Johnson, R.W. 178
Rau, C. 123, 124
Raub, E., s e e Loebich Jr, O. 182
Rauchschwalbe, U., s e e Franse, J.J.M. 325
Ravot, D., s e e Rossat-Mignod, J. 62, 401
Rayane, D. 114
Rayment, T., s e e Hay, CM. 179
Rayment, T., s e e Nix, R.M. 179
Reehuis, M., s e e Ball, A.R. 375, 386
Regnault, L.P., s e e Rossat-Mignod, J. 62, 81
Reichelt, J. 67, 68
Reichl, R. 111,131
Reiffers, M. 339-341,351
Reihl, B., s e e Kaindl, G. 1t 5, 117
Reihl, B., s e e Schneider, W.D. 118
Reihl, R. 122
Reikhrudel, E.M., s e e Kozlovskii, L.V i34
Reimer, V.A., s e e Borombaev, M.K. 364
Reinders, EH.P. 30, 69, 70
Reinders, P.H.R, s e e Chapman, S.B. 92, 95
Reinders, P.H.E, s e e Hunt, M. 81, 82
Reinders, PH.P., s e e Springford, M. 92
Reinders, P.H.R, s e e ()unki, Y. 33-36
Remy, E, s e e Vajda, R 219, 232, 233
Ren, Y.G., s e e Jaswal, S.S. 151
Reshentnikova, L.V. 118
Ressouche, E., s e e Fillion, G. 393
Rez, P., s e e Manoubi, T. 132
Rhodes, E, s e e Wohlfarth, E.R 298, 306
Rhyne, J.J., s e e Gignoux, D. 342
Ribas, R., s e e Ferrater, C. 158
Richter, HJ. 187
Riedel, A.A., s e e Bacon, EM. 141
Ritter, C. 321
Ritter, C., s e e Arons, R.R. 271,273
Ritter, C., s e e Mondal, S. 320
Rivory, J., s e e Frigerio, J.M. 177
Rivory, J., s e e Martin, M. 177
Roberts, M. de V., s e e Longuet-Higgins, H.C. 32
Rochow, R., s e e Carbone, C. 124, 158
Rodewald, W. 152, 153, 155
Rodriguez Fernandez, J., s e e Ball, A.R. 392
Rodriguez Fernandez, J., s e e Barandiaran, J.M.
366
Roeland, L.W., s e e McEwen, K.A. 350, 351
Roeland, L.W., s e e Meyer, R.T.W. 51
Rojek, A., s e e Cendtewska, B. 184
Ronay, M. 185
Ropars, G., s e e Caulet, J. 187
Ropars, G., s e e Le Corre, A. 187
Rosei, R., s e e Weaver, J.H. 265

448

AUTHOR INDEX

Rosenberg, M., s e e Homburg, H. 145

Rosengren, A. 111
Rosenthaler, L., s e e Griitter, P. 148
Rosenthaler, L., s e e Heinzelmann, H. 148
Ross, D.K., s e e Anderson, I.S. 216, 233
Ross, D.K., s e e Beunington, S.M. 234, 269
Ross, D.K., s e e Fairclough, J.P.A. 219
Ross, D.K., s e e Hunt, D.G. 243
Rossat-Mignod, J. 62, 81,401,413
Rossat-Mignod, J., s e e Chattopadhyay, T. 401
Rossat-Mignod, J., s e e Effantin, J.M. 31,359,
397
Rossat-Mignod, J., s e e Jerjini, M. 70, 392
Rossi, D. 184
Rossi, G. 130
Rossi, G., s e e Yeh, J.J. 130
Rossignol, M., s e e Barbara, B. 360
Rossignol, M., s e e Purwins, H.G. 360
Rossignol, M.E, s e e Barbara, B. 67
Rostoker, R., s e e Kohn, W. 12
Rouault, R, s e e Ballou, R. 324
Rouault, P., s e e Berthier, Y. 313
Rouchon, C., s e e Gignoux, D. 371,385, 387
Rouchy, J., s e e Al6onard, R. 394, 395
Rouehy, J., s e e Morin, R 353, 355-357, 394, 396,
397
Roudaut, E., s e e D~portes, J. 321, 322
Roudaut, E., s e e Gignoux, D. 371,385, 387
Roux, J.-Ph., s e e Br~chignac, C. 114
Rowe, J.M., s e e Anderson, I.S. 234
Rowe, J.M., s e e Glinka, C.J. 243
Rowell, J.M., s e e Lowe, W.E 125
Rozendaal, E., s e e McGuiness, P.J. 148
Ruderman, M.A. 3, 415
Rudigier, H., s e e Schlapbach, L. 252, 257
Ruiz, A., s e e Martinez, B. 146
Rumyantsev, A.Yu., s e e Parshin, P.R 243
Runge, E., s e e Zwicknagl, G. 14, 77, 78
Rush, J.J. 243
Rush, J.J., s e e Anderson, I.S. 234
Rush, J.J., s e e Berk, N.E 232
Rush, J.J., s e e Cannelli, G. 232
Rush, J.J., s e e Glinka, C.J. 243
Rush, J.J., s e e Gygax, EN. 232
Rush, J.L, s e e Udovic, T.J. 234, 235, 243
Rustamov, P.G., s e e Efendiev, E.G. 187
Rustamov, R.G., s e e Efendiev, E.G. 187
Ryborg, E, s e e Onsgaard, J. 126, 128-131
Saad, EM., s e e Comer, W.D. 126
Sadikov, I.P., s e e Alekseev, P.A. 339, 341
Saeki, M., s e e Shigeoka, T. 384, 385
Sagasaki, M. 166
Sagawa, M., s e e Yamamoto, H. 151, 152

Sahashi, M. 156, 174

Sahni, VC. 18
Saito, N., s e e Sato, R. 166
Sakaguchi, H. 169-172
Sakaguchi, H., s e e Adachi, G. 140, 170-172
Sakaguchi, H., s e e Shirai, H. 172
Sakai, O. 60, 61, 63, 64
Sakai, O., s e e Harima, H. 32, 33
Sakai, O., s e e Kasuya, T. 13, 59, 60, 6345
Sakaki, Y., s e e Okuno, H. 163
Sakakibara, T. 300, 301
Sakakibara, T., s e e Goto, T. 300, 301,306
Sakakibara, T . , s e e Takayanagi, S. 363, 364
Sakakibara, T., s e e Yoshimura, K. 301
Sakamoto, I. 70, 72-74
Sakamoto, I., s e e Satoh, K. 42, 44, 48, 51
Sakamoto, Y. 182, 184
Sakamoto, Y., s e e Takao, K. 182
Sakatsume, S., s e e Goto, T. 33, 34, 65, 66
Sakatsume, S., s e e Kitazawa, H. 65
Sakatsume, S., s e e Settai, R. 60, 61, 65
Sakatsume, S., s e e Suzuki, T. 33, 34
Sakatsume, S., s e e Takeda, N. 67
Sakurada, S., s e e Kaneko, T. 397
Sakurai, J., s e e Benoit, A. 52
Sakurai, J., s e e Fukuhara, T. 74-76
Sakurai, J., s e e Kamimura, H. 324
Sakurai, J., s e e Nakamura, H. 319
Sakurai, Y., s e e Lii, Z.Y. 156
Sakurai, Y., s e e Okuno, H. 163
Salamanca-Young, L. 189
Salas, EH. 123
Salas, F.H., s e e Alameda, J.M. 145
Salibi, N., s e e Zamir, D. 256
Salo, I.P., s e e Linetski, Ya.L. 174
Sanadze, V.V., s e e Dadiani, T.O. 188
Sanadze, V.V., s e e Dzhabua, Z.U. 188
Sanchez, J.P., s e e Shinjo, T. 158
Sandratskii, L., s e e Uhl, M. 326
Sands, T., s e e Allen Jr, S.J. 186, 187
Sands, T., s e e Palmstrom, C.J. 186
Sankar, S.G., s e e Cheng, S.E 155
Sankar, S.G., s e e Smith, H.K. 137
Sarma, D.D., s e e Hillebrecht, EU. 63
Sarma, D.D., s e e Weller, D. 122
Sarpal, B.K. 112
Sarpal, B.K., s e e Blancard, C. 112, 114
Sase, Y., s e e Ishida, A. 188, 189
Sato, H., s e e Sakamoto, I. 70, 72-74
Sato, K. 362, 363
Sato, K., s e e Isikawa, Y. 85
Sato, K., s e e Maezawa, K. 83, 84, 86, 90
Sato, K., s e e 0nuki, Y. 83, 84, 86
Sato, K . , s e e Umehara, I. 53-58

AUTHOR INDEX
Sato, M., s e e 0nuki, Y. 33, 34
Sato, N. 160, 164
Sato, N., s e e Komatsubara, T. 31
Sato, N., s e e Takeda, N. 65, 67
Sato, N., s e e Tanaka, K. 65, 66
Sato, N., s e e Yarnauchi, K. 160
Sato, R. 166
Sato, S., s e e Ishii, T. 3
Sato, T., s e e Nakajima, K. 178
Satoh, H., s e e Sakamoto, I. 73, 74
Satoh, K. 5, 42, 44, 48, 51, 89, 90, 94
Satoh, K., s e e Ebihara, T. 42, 44-46, 48, 49, 51
Satoh, K., s e e Kurosawa, Y. 42, 44, 51
Satoh, K . , s e e Umehara, I. 47-51, 70-72
Satoh, K., s e e Onuki, Y. 74-78, 80, 87-89, 91-93
Satoh, T., s e e Kitazawa, H. 65
Satoh, T., s e e Takahashi, M. 70, 392
Savitskii, E.M., s e e Alekseev, P.A. 339, 341
Savitskii, E.M., s e e Goremychldn, E.A. 339, 342
Savrin, V.D. 109
Saw, C.K. 216, 219
Sawatsky, G.A., s e e Thole, B.T. 114
Sawatzky, G.A., s e e Esteva, J.M. 131, 132
Sayetat, E, s e e D~portes, J. 321,322
Schaefer, W . , s e e Arons, R.R. 281
Schafer, H., s e e Steglich, E 6
Schaudy, G., s e e Gignoux, D. 375, 377, 415
Schefer, J. 225, 273, 277
Schellenberg, L., s e e Kuboth, M. 185
Schelleng, J.H. 167
Schelleng, J.H., s e e Forester, D.W. 168
Schelleng, J.H., s e e Vittoria, C. 168
Schenk, A., s e e Gygax, F.N. 232
Schiffmacher, G. 177
Schiffmacher, G., s e e Gasgnier, M. 132
Sehinkel, C. 300
Schlapbach, L. 210, 252, 257, 265, 266
Schlapbach, L., s e e Fujimori, A. 265
Schlapbach, L., s e e Schefer, J. 225, 273, 277
SchlegeI, H., s e e Krost, A. 188, 189
Schmidt, H., s e e Czjzek, G. 366
Schmidt-May, J., s e e Chorkendorff, I. 116
Schmidt-May, J., s e e Nyholm, R. 116
Schmitt, D. 414
Schmitt, D., s e e Al6onard, R. 355
Schmitt, D., s e e Ball, A.R. 342, 351-353, 375,
377, 378, 380, 386, 391,392, 402-406, 417
Schmitt, D., s e e Barandiaran, J.M. 366
Schmitt, D., s e e Barthem, V.M.T.S. 339-341,
343, 344
Schmitt, D., s e e Blanco, J.A. 365, 370-373, 382,
383, 401-406
Schmitt, D., s e e Bouvier, M. 365, 405
Schmitt, D., s e e Fr6my, M.A. 392

449

Schmitt, D., s e e Gignoux, D. 297, 300, 345, 366,

369, 371,373, 375-377, 385, 387, 392, 393,
409, 415
Schmitt, D., s e e Jaussaud, C. 355, 356
Schmitt, D., s e e Morin, P. 338, 353, 355-357,
359, 360, 394, 396, 397, 408, 414, 417
Schmitt, D., s e e Radwanski, R.J. 339, 342
Schmitt, D., s e e Reiffers, M. 339-341,351
Schmitt, D., s e e Shigeoka, T. 368
Schrnitt, D., s e e Zhang, EY. 339, 340, 342, 343,
361
Schmitzer, C. 236, 248, 268
Schmitzer, C., s e e Vajda, P. 245, 246, 268
Schnabel, B., s e e Zogal, O.J. 237
Schneider, G. 152
Schneider, G., s e e Fidler, L 152, 153
Schneider, G., s e e Knoch, K.G. 152
Schneider, W.D. 118
Schneider, W.D., s e e Domke, M. 114, 115
Schneider, W.D., s e e Kaindl, G. 115, 117
Schneider, W.D., s e e Moser, H.R. 130-132
Schoenberger, R.J., s e e Belhoul, M. 277
Schoenberger, R.J., s e e Lichty, L.R. 233
Schokawa, J., s e e Sakaguchi, H. 172
Scholz, U.D., s e e McGuiness, P.J. 148
Schotte, K.D., s e e Bred1, C.D. 70
Schreiber, D.S. 236, 237
Schrey, E, s e e Rodewald, W. 155
Schrieffer, J.R., s e e Berk, N. 304
Schr6der, K., s e e W e l l e r , D. 122, 124
Schuller, I.K., s e e Homma, H. 119
Schuller, I.K, s e e Yang, K.A. 119
Schuurmans, M.EH., s e e Daalderop, G.H.O.
313-315
Schwartz, C.L., s e e Palmstrom, C.J. 186
Schwarz, R.B. 180
Schwarz, R.B., s e e Leisure, R.G. 232, 236
Schwarz, W., s e e Andr6, G. 227, 230, 281
Schwarz, W., s e e Pleschiutschnig, J. 234
Schweizer, J. 311
Schweizer, J., s e e Arons, R.R. 271,278, 279,
281
Schweizer, J., s e e Ballou, R. 310
Schweizer, J, s e e Barandiaran, LM. 366
Schweizer, J., s e e Barbara, B. 67
Schweizer, J., s e e Benoit, A. 52
Schweizer, J., s e e Fillion, G. 393
Schweizer, J., s e e Gignottx, D. 309, 310
Scott, C.A., s e e Chowdhury, M.R. 229, 237
Seehman-Eggebert, M., s e e Richter, H.J. 187
Segre, C.U., s e e Croft, M. 181
Seitz, E. 67-69
Selke, W. 399
Sellmyer, D.J. 40, 151, 158, 159, 164

450

AUTHOR INDEX

Sellmyer, D.J., s e e Aylesworth, K.D. 146, 150,

151
Sellmyer, D.J., s e e Jaswal, S.S. 151, 152
Sellmyer, D.J., s e e Shan, Z.S. 158-161, 164, 165
Sellmyer, D.J., s e e Strzeszewski, J. 151
Sellmyer, D.J., s e e Tiwald, T.E. 158
Senoussi, S. 271,273, 277, 279
Sera, M., s e e Kitazawa, H. 60, 63
Serano, C.M., s e e Ronay, M. 185
Seri, H., s e e Sakaguchi, H. 169, 170
Settai, R. 52, 53, 60, 61, 65, 87-89
Settai, R., s e e Goto, T. 65, 66
Seymour, E.EW., s e e Barnes, R.G. 216
Seymour, E.EW., s e e Bamfather, K.J. 237
Seymour, E.EW., s e e Chowdhury, M.R. 229, 237
Seymour, E.EW., s e e Hart, J.W. 233
Seymour, E.EW., s e e Lichty, L.R. 232, 244
Seymour, E.EW., s e e Phua, T.T. 237
Seymour, E.EW., s e e Torgeson, D.R. 233
Shaburov, V.A., s e e Smimov, I.A. 251,264
Shah, J.S., s e e Givord, E 301
Shaked, H. 271,275, 281-284
Shaltiel, D. 285
Shaltiel, D., s e e Daou, J.N. 213, 217, 251,264,
271
Sham, L.J., s e e Kohn, W. 14
Shamir, N., s e e Venkert, A. 179
Shah, Z.S. 158-161, 164, 165
Shah, Z.S., s e e Sellmyer, D.J. 158, 159, 164
Shah, Z.S., s e e Tiwald, T.E. 158
Shapiro, S.M., s e e Lawrence, J.M. 40
Sharnonja, V.G., s e e Loboda, V.B. 120, 134
Shelton, R.N., s e e Klavins, P. 215, 221,225, 226,
229
Shelton, R.N., s e e Xu, Y. 185
Shen, J.X., s e e Shah, Z.S. 164
Shen, 3(. 175
Shenoy, G.K., s e e Carlin, R.L. 275, 281
Shenoy, G.K., s e e Friedt, J.M. 282
Shevlyakov, S.A., s e e Gorodetskii, D.A. 118
Shi, S.Y., s e e Huang, G.X. 150
Shibata, T., s e e Suzuki, Y. 162
Shibutani, K., s e e Sumiyama, A. 4
Shida, H., s e e Kitazawa, H. 41-43
Shiga, M. 323, 324
Shiga, M., s e e Kamimura, H. 324
Shiga, M., s e e Minakata, R. 303
Shiga, M., s e e Muraoka, Y. 303
Shiga, M., s e e Nakamura, H. 319
Shiga, M., s e e Nakamura, Y. 320
Shiga, M., s e e Oomi, G. 321
Shiga, M., s e e Sakakibara, T. 300, 301
Shiga, M., s e e Wada, H. 303, 304, 317-319,
323-325

Shiga, M., s e e Yoshie, H. 311

Shiga, M., s e e Yoshimura, K. 321,323
Shigeoka, T. 368, 380, 382-385, 401
Shigeoka, T., s e e Fujii, H. 380, 381,384
Shigeoka, T., s e e Iwata, N. 368, 374, 375, 401
Shigeoka, T., s e e Nojiri, H. 381
Shigeoka, T., s e e Sugiyama, K. 371
Shijo, T., s e e Hosoito, N. 158
Shikhmanter, L. 179, 181
Shimamori, T., s e e Miyazaki, T. 162
Shimizu, K. 317
Shimizu, M. 297
Shimizu, M., s e e Bloch, D. 300, 301
Shimizu, M., s e e Yamada, H. 301,306, 325
Shimizu, T., s e e Yoshimura, K. 306
Shimizu, Y., s e e ()nuki, Y. 87
Shimoghori, T., s e e Adachi, G. 171
Shimohara, K., s e e Ishida, A. 188
Shin, S.C. 160
Shinar, J. 251,253, 255, 257
Shinjo, T. 158
Shinjo, T . , s e e Hosoito, N. 158
Shinjo, T., s e e Mibu, K. 158
Shinjo, T., s e e Yoden, K. 158
Shinoda, T., s e e Tsunashima, S. 168
Shiokawa, J., s e e Adachi, G. 170-172
Shiokawa, J., s e e Arakawa, T. 144
Shiokawa, J., s e e Sakaguchi, H. 169-172
Shiozaki, I., s e e Sakamoto, I. 73, 74
Shirai, H. 172
Shirane, G., s e e Freltoft, T. 323
Shirane, G., s e e Motoya, K. 323
Shirley, D.A., s e e Mason, M.G. 114
Shivaprasad, S.M., s e e Ashrit, P.V. 133
Shiwaku, T., s e e Kuwano, N. 182
Shoenberg, D. 11, 27
Shohata, N. 361,362
Sholl, C.A., s e e Lichty, L.R. 232, 244
Shopov, V.S., s e e Apostolov, A.V. 127
Shubina, T.S., s e e Volkenshtein, N.V. 271
Shuler, K., s e e Oesterreicher, H. 137
Siari, A., s e e Malterre, D. 182
Siconolfi, D.J., s e e Frankenthal, R.P. 168
Siconolfi, D.J., s e e van Dover, R.B. 168
Siegel, R.W., s e e Li, Z. 110
Sigrist, M. 126, 128
Sima, V., s e e Lebech, B. 373
Singh, A.K. 177
Singh, A.K., s e e Singh, S.K. 169-172
Singh, O. 134
Singh, O., s e e Curzon, A.E. 120, 135, 137
Singh, S.K. 169-172
Singleton, J.H. 141
Sinha, V.L., s e e Cheng, S.E 155

AUTHOR INDEX
Sinitsyn, E.V., s e e Borombaev, M.K. 364
Sinnema, S. 329, 336, 337
Sinnema, S., s e e Franse, J.J.M. 336
Sinnema, S., s e e Radwanski, R.J. 336
Sinnema, S., s e e Verhoef, R. 333
Sinnemann, Th., s e e Homburg, H. 145
Skalicky, R, s e e Fidler, J. 174
Skillman, S., s e e Herman, E 9
Skriver, H.L., s e e Boulet, R.M. 52
Slater, J.C. 12
Slisenko, V.V., s e e Goremychkin, E.A. 339, 342
Smereka, T.P., s e e Gonchar, EM. 118
Smereka, T.P., s e e Lozovyi, Ya.B. 119
Smetana, Z . , s e e Borombaev, M.K. 364
Smetana, Z., s e e Lebech, B. 373
Smirnitskaya, G.V., s e e Kozlovskii, L.V. 134
Smimov, I.A. 25 I, 264
Smith, D.A. 182
Smith, H.K. 137
Smith, J.E, s e e Subramanian, RR. 217
Smith, J.E, s e e Thome, D.K. 236, 241,246, 247,
252
Smith, M.E., s e e Barnes, R.G. 216
Smith, R.L. 126
Smith, R.L., s e e Comer, W.D. 126
Smith, R.S., s e e Richter, H.J. 187
Smutek, M. 136
Smutek, M., s e e Boeva, O.A. 136, 143, 240
Smutek, M., s e e CereS,, S. 143
Soda, K., s e e Ishii, T. 3
Sommer, R., s e e Lfithi, B. 357, 358
Sonntag, B., s e e Ltibcke, M. 114
Sonntag, B., s e e Materlik, G. 112
Sorensen, O., s e e Onsgaard, J. 116, 127, 129, 131
Soubeyroux, J.L, s e e BaIlou, R. 324
Southall, J., s e e Jones, P.M.S. 217
Sov~k, P., s e e Ratajczak, H. 157
Spain, G., s e e Franse, J.J.M. 325
Spedding, EH., s e e Beaudry, B.J. 217, 219, 234,
245
Spencer, E.G., s e e Croft, M. 181
Spencer, E.G., s e e Lu, E 181
Springford, M. 81, 92
Springford, M., s e e Chapman, S.B. 92, 94, 95
Springford, M., s e e Hunt, M. 81, 82
Springford, M., s e e Reinders, P.H.R 30, 69, 70
Springford, M., s e e Satoh, K. 42, 44, 48, 51
Springford, M., s e e Wasserman, A. 35
Springfurd, M., s e e 0nuki, Y. 33-36
Sprokel, G.S., s e e Marinero, E.E. 166
Spruijt, A.M.J., s e e Heitmann, H. 168
Srivastava, O.N., s e e Ramak.risna, K. 171
Srivastava, O.N., s e e Singh, A.K. 177
Srivastava, O.N., s e e Singh, S.K. 169-172

451

Stadelmaier, H.H. 155

Stalinski, B., s e e Bieganski, Z. 283, 284
Stalinski, B., s e e Czopnik, A. 398
Stalinski, B., s e e Drulis, M. 241, 252, 258, 263,
273, 275, 278, 279, 285
Stalinski, B., s e e Kletowski, Z. 52
Stamateli, M.Yu., s e e Glurdzhidze, L.N. 188
Stassis, C., s e e Khatamian, D. 216, 219
Stassis, C., s e e Saw, C.K. 216, 219
Stasfik, Z . V . , s e e Kosak, M.M. l 18, 120
Steglich, E 6
Steglich, E, s e e Bredl, C.D. 70
Steglich, E, s e e Franse, J.J.M. 325
Steglich, E, s e e Gratz, E. 87
Steinbinder, D., s e e Stuhr, U. 237
Stenborg, A. 116, 117
Sticht, J. 81
Stierman, R.J. 246, 271
Stierman, R.J., s e e Gschneidner Jr, K.A. 6, 52
Stierman, R.J., s e e Ikeda, K. 305
Strange, D. 14, 55
Strange, P., s e e Taillefer, L. 26
Strasser, G. 115
Strasser, G., s e e Matthew, J.A.D. 132
StrSssler, S., s e e Perkins, R.S. 312
Strauss, A., s e e Waibel, E 285
Streit, P., s e e Everett, G.E. 394, 411
Strnat, K.J. 328
Stryjewski, E. 345-347
Strzeszewski, J. 144, 151
Stuhr, U. 237
Stunault, A., s e e Fillion, G. 393
Sturdy, G.E. 231
Styles, G.A., s e e Bamfather, K.J. 237
Styles, G.A., s e e Chowdhury, M.R. 229, 237
Styles, G.A., s e e Phua, T.T. 237
Subia Jr, S.R., s e e Bacon, FM. 141
Subramanian, RR. 217
Sudovtsov, A.I., s e e Kuzmenko, V.M. 118, 121,
122
Suezawa, N., s e e Arakawa, T. 144
Suezawa, Y. 167
Suezawa, Y., s e e Gond6, Y. 167
Sugita, Y., s e e Niihara, T. 168
Sugiura, E. 401
Sugiyama, K. 7, 365, 371
Sugiyama, K., s e e Shigeoka, T. 401
Sugiyama, K., s e e Sugiura, E. 401
Sumiyama, A. 4
Sun, W., s e e Hunt, M. 81, 82
Sun, Z., s e e Webb, D.J. 163
Surplice, N.A. 135, 137, 139, 143
Surplice, N.A., s e e Kandasamy, K. 137, 140
Surplice, N.A., s e e Mfiller, J. 140

452

AUTHOR INDEX

Sttryanarayanan, R. 189
Suryanarayanan, R., s e e Das, S.K. 190
Suzuki, A. 166, 168
Suzuki, H., s e e Kitazawa, H. 65
Suzuki, T. 31, 33, 34, 67, 91
Suzuki, T., s e e Endoh, D. 92, 93
S u z u k i , T . , s e e Fukuma, H. 394
Suzuki, T., s e e Goto, T. 33, 34
S u z u k i , T . , s e e Iwamura, E. 150
S u z u k i , T . , s e e Kasuya, T. 13, 59, 60, 63~i5
Suzuki, T., s e e Kitazawa, H. 41-43, 60, 63, 65
Suzuki, T., s e e Kwon, Y.S. 65, 67
Suzuki, T., s e e Nirnori, S. 350
Suzuki, T., s e e Ozeki, S. 65, 66
S u z u k i , T . , s e e Settai, R. 60, 61, 65
Suzuki, T., s e e Takeda, N. 65, 67
S u z u k i , T., s e e Tanaka, K. 65, 66
Suzuki, Y. 162
Svare, I., s e e Leisure, R.G. 232
Svoboda, P. 387
Swartz Jr, W.E., s e e Holloway, D.M. 142
Switendick, A.C. 243
Switendick, A.C., s e e Mattheiss, L.E 22
Syms, M., s e e Miller, R.E 118, 126, 127
Szofran, ER., s e e Sellmyer, D.J. 40
Szymazek, J., s e e Bohdziewicz, A. 184
Szytula, A. 384
Szytuta, A., s e e Ivanov, V. 384, 385
Szytuta, A . , s e e Vinokurova, L. 382
Tabatabaie, N., s e e Allen Jr, S.J. 186, 187
Tabatabaie, N., s e e Palmstrom, C.J. 186
Taborelli, M. 124
Taborelli, M., s e e LandoR, M. 124
Tada, M., s e e Ohki, C. 240
Tada, M., s e e Toguchi, K. 240
Taillefer, L. 26
Tajika, H., s e e Umemura, S. 159
Takabatake, T., s e e Shigeoka, T. 383-385
Takabatake, T., s e e Sugiura, E. 401
Takada, S. 12
Takahashi, H. 62
Takahashi, M. 70, 162, 392
Takahashi, T., s e e Hosoda, N. 136
Takahashi, T., s e e Kawahata, T. 136
Takahashi, Y., s e e Moriya, T. 305
Takao, K. 182
Takao, K., s e e Sakamoto, Y. 182, 184
Takayama, S., s e e Niihara, T. 168
Takayanagi, S. 90, 94, 363, 364, 389, 390
Takayanagi, S., s e e 0nuki, Y. 90-93, 390
Takeda, H., s e e Isikawa, Y. 85
Takeda, K. 133
Takeda, N. 65, 67

Takeda, N., s e e Kitazawa, H. 65

Takeda, N., s e e Tanaka, K. 65, 66
Takeda, S. 173, 174
Takeda, S., s e e Horikoshi, H. 173
Takeda, S., s e e Komura, Y. 174
Takeda, T., s e e Baba, K. 188
Takeda, T., s e e Nakamura, O. 118, 125
Takeda, T., s e e Takahashi, M. 162
Takei, H., s e e Takahashi, M. 70, 392
Takeshige, M., s e e Sakai, O. 60, 61, 63, 64
Takeshita, T., s e e Ito, T. 229
Takeuchi, A., s e e Barthem, V.M.T.S. 339, 343,
344
Takeuchi, A., s e e Fr6my, M.A. 392
Takeuchi, A., s e e Gignoux, D. 392, 393
Takeuchi, T., s e e Gignoux, D. 371,385, 387
T a l d , K . , s e e Suzuki, A. 166, 168
Takigawa, M., s e e Yoshimura, K. 306, 323
Talianker, M., s e e Shikhmanter, L. 179, 181
Talianker, M., s e e Venkert, A. 179
T a m a k i , A . , s e e Suzuki, T. 33, 34, 91
Tanaka, H., s e e Takahashi, M. 70, 392
Tanaka, J., s e e Kasuya, T. 13, 59, 60, 63-65
Tanaka, K. 65, 66
Tanaka, K., s e e Takeda, N. 67
Tanaka, M. 160, 161
Tanaka, T., s e e Ishizawa, Y. 31-33, 35
Tanaka, T., s e e Nozaki, H. 365
Tanaka, Y. 303
Tang, J. 70, 72
Tang, J., s e e Gschneidner Jr, K.A. 264
Tang, W. 150
Taniguehi, M., s e e Ishii, T. 3
Taniguchi, N., s e e Sakaguchi, H. 169, 170, 172
Tanner, B.K., s e e Smith, R.L. 126
Tanoue, H., s e e Tsukahara, S. 167
Tarascon, J.M., s e e Kasaya, M. 6
Tasaki, A., s e e Umemura, S. 159
Tasset, F., s e e Gignoux, D. 83, 85, 299, 300, 302,
308, 309
Tasset, F., s e e Schweizer, J. 311
Tauner, B.K., s e e Corner, W.D. 126
Taush, M., s e e Materlik, G. 112
Tawara, Y., s e e Fidler, J. 150
Taylor, K.N.R., s e e Krizek, J. 126, 127
Tazaki, A., s e e Miyamoto, K. 157
Teisseron, G., s e e Ferreira, P. 260
Tejada, J., s e e Badia, F. 157
Tejada, J., s e e Ferrater, C. 158
Tejada, J., s e e Martinez, B. 146
Tellefsen, M., s e e Bischof, R. 214, 217
Tellefsen, M., s e e Kaldis, E. 225
Temmerman, W.M., s e e Langford, H.D. 33
Terada, T., s e e Oomi, G. 321

AUTHOR INDEX
Terao, K., s e e Uchida, H. t69
Terao, K., s e e Yamada, H. 306
Tessier, M., s e e Krishnan, R. 166
Thaper, C.L., s e e Rush, J.J. 243
Thiry, E, s e e Schlapbach, L. 252, 257, 265, 266
Thole, B.T. 114
Tholence, J.L., s e e Amato, A. 94, 95
Tholence, J.L., s e e Joss, W. 33, 34, 36
Thomas, G., s e e Koestler, C. 150, 151
Thomas, G., s e e Ramesh, R. 150, 151
Thomas, J., s e e Quemerais, A. 127, 129
Thome, D.K. 236, 241,246, 247, 252
Thrush, C.M., s e e Partin, D.L. 188
Thuy, N.R 311,312
Thuy, N.R, s e e Franse, JJ.M. 336
Tian, E, s e e Pan, S.M. 175
Tian, J., s e e Li, L. 152
Tibbetts, G.G. 116
Tibbetts, G.G., s e e Egelhoff Jr, W.E 116, 131
Tilley, D.R., s e e Camley, R.E. 159
Tillman, D., s e e Carbone, C. 124, 158
Tissier, B., s e e Coey, J.M.D. 178
Titcomb, C.G. 225
Tiwald, T.E. 158
Tkach, VI. 110
Tlianker, M., s e e Shikhmanter, L. 179
Togami, Y., s e e Kajiura, M. 157
Togami, Y., s e e Morishita, T. 157, 159
Toguchi, K. 240
Tohyama, T., s e e Yamada, H. 301,306
Toki, K. 178
Tokita, T., s e e Tanaka, M. 160, 161
Tokuhara, K., s e e Yamamoto, H. 151, 152
Toktmaga, M., s e e Fidler, J. 153
Tokunaga, M., s e e Gr6ssinger, R. 153
Tomala, K., s e e Czjzek, G. 366
Tomiyama, R 336
Tomiyoshi, S., s e e Kitai, T. 388
Tomokiyo, Y., s e e Kuwano, N. 182
Torgeson, D.R. 233
Torgeson, D.R., s e e Barnes, R.G. 216
Torgeson, D.R., s e e Bamfather, K.J. 237
Torgeson, D.R., s e e Belhoul, M. 277
Torgeson, D.R., s e e Borsa, F. 237
Torgeson, D.R., s e e Han, J.W. 233
Torgeson, D.R., s e e Leisure, R.G. 232
Torgeson, D.R., s e e Lichty, L.R. 232, 233, 244
Torgeson, D.R., s e e Phua, T.T. 237
Torgeson, D.R., s e e Zamir, D. 256
Torikaehvili, M.S., s e e Huang, C.Y. 185
Tougaard, S. 111, 126, 128
Tougaard, S., s e e Onsgaard, J. 111, 126, 128-131
Tourillon, G., s e e Brouder, C. 159, 165
Tourillon, G., s e e Guilmin, R 165

453

Toxen, A.M. 163

Toxen, A.M., s e e Webb, D.J. 163
Tran Minh Duc, s e e Frigerio, LM. 177
Tran Mirth Duc, s e e Martin, M. 177
Traverse, A., s e e Mathevet, J.R 180
Trebbia, R, s e e Bmusseau-Lahaye, B. 128, 129
Trebbia, R, s e e Colliex, C. 117, 128-130, 137
Trequattrini, F., s e e Cannelli, G. 232
Tribollet, H., s e e Rayane, D. 114
Trouiller, N., s e e Franciosi, A. 132
Trouiller, N., s e e Raisanen, A. 132
Trouiller, N., s e e Wall, A. 132
Truong, V.V., s e e Ashrit, RV. 133
Truong, V . V . , s e e Chee, K.T. 128, 129
Tsang, T.-W.E., s e e Gschneidner Jr, K.A. 6, 52
Tsang, T.-W.E., s e e Ikeda, K. 305
Tsoukatos, T., s e e Strzeszewski, J. 144
Tsuchida, T., s e e Tanaka, Y. 303
Tsuchiya, R., s e e Gond6, Y. 167
Tsuchiya, R., s e e Suezawa, Y. 167
Tsukahara, S. 167
Tsunashima~ S. 163, 168
Tsunashima, S., s e e Sagasaki, M. 166
Tsuneto, T. 353
Tsushima, K., s e e Kajiura, M. 157
Tsushima, K., s e e Morishita, T. 157, 159
Tsushima, T., s e e Ohkoshi, M. 311
Tsuzuki, S., s e e Uchida, H. 169, 170
Tsvetkov, V.Yu., s e e Linetski, Ya.L. 153
Tukahashi, M., s e e Miyazaki, T. 162
T u r , R . , s e e Daou, J.N. 269
Turek, K., s e e Kakol, Z. 311

Uchi, M., s e e Nojiri, H. 381

Uchida, H. 169, 170
Uchida, H., s e e Hosoda, N. 136
Uchida, H., s e e Kawahata, T. 136
Uchida, H., s e e Kojima, T. 169
Uchida, H., s e e Ohki, C. 239, 240
Uehida, H., s e e Uchida, H. 169
Uchida, H.H., s e e Ohki, C. 240
Uehida, K. 178
Uchiyama, S., s e e Sagasaki, M. 166
Uchiyama, S., s e e Tsun~shima, S. 163, 168
Udovic, T.J. 234, 235, 243
Udovic, T.J., s e e Anderson, I.S. 234
Udovic, T.J., s e e Berk, N.E 232
Ueda, K. 324
Uhl, M. 326
Uji, S., s e e Aoki, H. 76-78
Ukon, I., s e e 0nuki, Y. 87
Umeda, T., s e e Iwamura, E. 150
Umehara, I. 29, 30, 41, 42, 47-58, 70-73

454

AUTHOR INDEX

Umehara, I., s e e Ebihara, T. 42, 44-46, 48, 49,

51
Umehara, I., s e e Kurosawa, Y. 42, 44, 51
Umehara, I., s e e Maezawa, K. 83, 84, 86
Umehara, I., s e e Satoh, K. 42, 44, 48, 51, 89, 90
Umehara, I., s e e ()nuki, . 74-78, 80, 83, 84, 86,
87, 91-93
Umemura, S. 159
Umezawa, A., s e e Onuki, Y. 33-37, 91-93, 95
Umezawa, H., s e e Huang, C.Y. 185
Umino, M., s e e Asano, H. 91
Urbaniak-Kucharczyk, A. 125
Uwatoko, Y., s e e Shigeoka, T. 383
Vajda, E 210, 213, 219, 220, 227, 229, 230, 232,
233, 236-238, 243-246, 251,253-260, 262,
263, 265, 267-271,273, 275, 277-285
Vajda, R, s e e Andr6, G. 227, 230, 281
Vajda, R, s e e Blaschko, O. 215, 219, 220, 234
Vajda, E, s e e Boukraa, A. 275, 282-285
Vajda, R, s e e Burger, J.E 229, 238, 243, 251,
252, 257-261,268, 271,273, 275-277, 283, 285
Vajda, E, s e e Chiheb, M. 221,225, 227
Vajda, P., s e e Danielou, R. 216
Vajda, P., s e e Daou, J.N. 213, 216, 217, 219-221,
225, 229, 233, 238, 239, 241,243-246, 248,
251-253, 255, 258, 261,262, 264, 267-269,
271,273, 275, 277-279, 283, 284
Vajda, R, s e e Lucasson, A. 241,243, 251,252
Vajda, R, s e e Metzger, T.H. 236
Vajda, R, s e e Plesehiutschnig, J. 234
Vajda, R, s e e Ratishvili, I.G. 227
Vajda, R, s e e Schmitzer, C. 236, 248, 268
Vajda, R, s e e Senoussi, S. 271,273, 277, 279
Vajda, R, s e e Shaltiel, D. 285
Vajda, R, s e e Udovic, T.J. 234, 235
van Aken, EB., s e e Buschow, K.H.J. 4
van Alphen, EM., s e e de Haas, WJ. 27
van Daal, H.J., s e e Buschow, K.HJ. 4
Van der Goot, A.S. 301
Van der Kraan, A.M., s e e Gubbens, EC.M. 339
van der Laan, G., s e e Thole, B.T. 114
van der Meulen, H.E 81
van Deursen, A.J.R 33, 34, 36
van Deursen, A.J.E, s e e Joss, W. 33, 34, 36
Van Diepen, A.M., s e e Nasu, S. 40
van Dover, R.B. 168
van Dover, R.B., s e e Frankenthal, R.E 168
van Dover, R.B., s e e Hellman, E 165
van Kempen, H., s e e van der Meulen, H.E 81
van Ruitenbeek, J.M., s e e Joss, W. 33, 34, 36
van Ruitenbeek, J.M., s e e Mtiller, T. 35, 36
van Stapele, R.R, s e e Busehow, K.H.J. 329
van Vucht, J.H.N., s e e Busehow, K.HJ. 177

Varkanova, R.G. 136, 142

Vasilkevich, A.A., s e e Goremychkin, E.A. 339,
342
Velichkov, I.V., s e e Apostolov, A.V. 126, 127
Venkataraman, G.,see Sahni, V.C. 18
Venkert, A. 179
Veres, T., s e e Amara, M. 398
Verhoef, R. 331-333
Verhoef, R., s e e Franse, J.J.M. 350
Verhoef, R., s e e Radwanski, RJ. 331
Verhoef, R., s e e Tomiyama, E 336
Verma, A.R. 172
Vettier, C., s e e Barbara, B. 299
Vettier, C., s e e Blanco, J.A. 370, 371,382, 383
Vettier, C., s e e Rossat-Mignod, J. 62, 81
Vettier, C., s e e Voiron, J. 303
Viallard, R. 240
Viallard, R., s e e Daou, J.N. 212
Victor, Y., s e e Bosca, G. 132
Viescas, A.J., s e e Heeht, M.H. 131
Vinokurova, L. 382
Vinokurova, L., s e e Ivanov, V 384, 385
Vitton, J.P., s e e Krishnan, R. 166
Vittoria, C. 168
Vittoria, C., s e e Forester, D.W. 168
Vittoria, C., s e e Sehelleng, J.H. 167
Vladychkin, A.N., s e e Kuzmenko, V.M. 118, 121,
122
Vogt, O., s e e Rossat-Mignod, J. 62, 401
Voiron, J. 303, 321
Voiron, J., s e e Bloch, D. 300, 301,303
Voiron, J., s e e Gignoux, D. 375, 377, 389, 406,
415
Volkenshtein, N.V 271
V61kl, J. 233
von Boehm, J., s e e Bak, P. 399
yon Molnar, S. 178, 180
Vorderwisch, R 243
Vorderwisch, E, s e e Wegener, W. 234
Vuillet, E, s e e Ferreira, E 260
Waber, J.T., s e e Liberman, D. 9, 21
Wachter, R, s e e Bischof, R. 271,273, 279
Wada, H. 303, 304, 317-319, 323-325
Wada, H., s e e Nakamura, H. 319
Wada, H., s e e Shiga, M. 323,324
Wada, N., s e e Mitsuda, S. 389
Wada, N., s e e 0nuki, Y. 90-93, 390
Wada, N., s e e Takayanagi, S. 90, 94, 363,364,
389, 390
Wadas, A., s e e Griitter, P. 148
Wagner, EE., s e e Waibel, E 285
Waibel, E 285
Waind, P.R. 338

AUTHOR INDEX

455

Wakabayashi, S., s e e Maezawa, K. 90

Wakabayashi, S., s e e ()nuki, Y. 91-93, 95
Waldyama, T., s e e Miyazaki, T. 162
Wakiyama, T., s e e Takahashi, M. 162
Walco, R.J., s e e Bacon, EM. 141
Waldrop, J.R. 186
Walker, E., s e e Purwins, H.G. 360
Walker, E., s e e Rossat-Mignod, J. 81
Wall, A. 132
Wall, A., s e e Franciosi, A. 132
Wall, A., s e e Raisanen, A. 132
Wallace, WE., s e e Cheng, S.E 155
Wallace, W.E., s e e Ganapathy, E.V. 301
Wallace, W.E., s e e Mansmann, W. 231
Wallace, WE., s e e Pebler, A. 225, 231
Wallace, W.E., s e e Smith, H.K. 137
Walmsley, R.G., s e e Webb, D.J. 162
Walser, R.M., s e e Choe, G. 163
Wan, H., s e e Fang, R.Y. 157
Wang, C.R., s e e Zhao, Z.B. 152
Wang, E E . , s e e Huang, C.Y. 185
Wang, G.A., s e e Huang, G.X. 150
Wang, R., s e e Tang, W. 150
Wang, Y.J., s e e Shah, Z.S. 164
Wang, Y.Z., s e e Jaswal, S.S. 152
Ward, M., s e e McGuiness, P.J. I48
Wasserman, A. 35
Watamura, S., s e e Nojiri, H. 381
Watanabe, H., s e e Takahashi, M. 162
Watanabe, K., s e e Hashimoto, S. 168
Watanabe, N., s e e Asano, H. 91
Watanabe, T., s e e Takayanagi, S. 90, 94, 363, 364
Watanabe, Y., s e e Koyama, M. 165
Watanabe, Y., s e e Suzuki, A. 166, 168
Waters, K., s e e Rau, C. 123
Watson, L.M., s e e Gimzewski, J.K. 137
Weaver, J.H. 126-129, 265
Weaver, J.H., s e e Mason, M.G. 114
Webb, DJ. 162, 163
Weber, K. 369
Wecker, J., s e e Koestler, C. 150, 151
Wegener, W. 234
Wegener, W., s e e Vorderwisch, R 243
Wehenkel, C., s e e Cukier, M. 128, 129, 131
Wei, W., s e e Lin, H. 379
Weinert, M. 309, 314

Westlake, D.G., s e e Carlin, R.L. 275, 281

Westlake, D.G., s e e Friedt, J.M. 282
Westlake, D.G., s e e Shaked, H. 271, 275,
281-284
Weymouth, J.W., s e e Sellmyer, D.J. 40
White, R.L. 158, 159
White, R.M., s e e Toxen, A.M. 163
White, R.M., s e e Webb, D.J. 162, 163
Whitehead, J.R, s e e Huang, C.Y. 185
Wickersham, C.E., s e e Frausto, RR. 168
Wickramasekara, L., s e e Kamprath, N. 155
Wickramasekara, L., s e e Liu, N.C. 155
Wieliczka, D.M. 115, 116
Wiesinger, G. 210
Wilkins, J.W., s e e Materlik, G. 112
Williams, A.R., s e e Janak, J.E 303
Williams, G.E, s e e Parks, R.D. 181
Williams, W.E., s e e Cheng, S.E 155
Willich, E, s e e Heitmann, H. 168
Wilson, K.G. 4
Wilting, H., s e e Heitmann, H. 168
Windmiller, L.R. 30
Windmiller, L.R., s e e Arko, A.J. 31, 33
Winter, H., s e e Shaltiel, D. 285
Winzer, K., s e e Reichelt, J. 67, 68
Wipf, H., s e e Stuhr, U. 237
Wipf, H., s e e V61kl, J. 233
Wise, M.L.H., s e e Hirst, J.R. 182
Wnuk, J.J., s e e Bohdziewicz, A. 184
Wohlfarth, E.P. 298, 306, 308
Wolf, A . , s e e Jacob, I. 241
Wolf, WP., s e e Lea, K.R. 273
Wood, J.H., s e e Mattheiss, L.E 22
Woods, S.B., s e e Kadowaki, K. 5
Woollam, J.A., s e e Tiwald, T.E. 158
Wortmann, G., s e e Waibel, E 285
Wu, L.H., s e e Huang, G.X. 150
Wu, R. 122
Wu, R.T., s e e Li, D. 123
Wulz, H.G. 136, 169
Wyder, R, s e e Mtiller, T. 35, 36
Wyder, R, s e e Reiffers, M. 339-341,351

Weller, D. 122-124
Welp, U., s e e Mtiller, T. 35, 36
Wendin, G., s e e Kanski, J. 131
Wennekers, E, s e e Richter, H.J. 187
Wemick, J.H., s e e Fawcett, E. 26
Wertheim, G.K. 115
West, G.W., s e e Barnes, R.G. 216
West, G.W., s e e Torgeson, D.R. 233

Yagi, Y., s e e Sakaguchi, H. 171, 172

Yamada, H. 301,306, 325
Yamada, K. 5
Yamagami, H. 13, 75, 77-79, 84-86
Yamagami, H., s e e Hasegawa, A. 22, 52-58
Yamagishi, A., s e e Morin, E 396, 397
Yamagishi, A., s e e Tomiyama, E 336
Yamaguchi, I., s e e Baba, K. 188

Xia, S.K., s e e Zhao, Z.B.

Xu, Y. 185
Xu, Y., s e e Cheng, S.E

152
155

456

AUTHOR INDEX

Yamaguchi, S. 133
Yamaguchi, S., s e e Miyahara, T. 129, 130
Yamamoto, H. 151, 152
Yamamoto, H., s e e Nakamura, T. 162
Yamamoto, K . , s e e Kuwano, N. 182, 184
Yamanaka, S., s e e Hoshi, Y. 168
Yamanaka, S., s e e Naoe, M. 168
Yamashita, M. 388
Yamashita, S. 155
Yamashita, S., s e e Harada, T. 157
Yamauchi, K. 160
Yamazaki, T., s e e 0nuki, Y. 33-37, 91-93
Yanase, A . , s e e Harima, H. 14, 32, 33, 72, 73,
81, 82, 89, 92, 94
Yanase, A., s e e Hasegawa, A. 32, 39, 40, 67-69
Yanase, A., s e e Kitazawa, H. 60, 63
Yanase, A., s e e Kubo, . 14, 36, 37
Yang, D., s e e Li, L. 152
Yang, K.A. 119
Yang, K.Y., s e e Homma, H. 119
Yannopoulos, L.N., s e e Singleton, J.H. 141
Yanson, I.K., s e e Reiffers, M. 339-341,351
Yaremenko, A.V., s e e Loboda, V.B. 120, 134, 137
Yashiyama, M., s e e Honda, S. 164
Yasuoka, H., s e e Yoshimura, K. 323
Yeh, J.J. 130
Yeh, J.L, s e e Rossi, G. 130
Yelon, W.B., s e e Herbst, J.H. 312
Yelon, W.B., s e e Jaswal, S.S. 152
Yin, X.J., s e e McGuiness, P.J. 148
Yoden, K. 158
Yoden, K., s e e Hosoito, N. 158
Yoden, K., s e e Shinjo, T. 158
Yonenobu, K., s e e Morin, P. 396, 397
Yonenobu, K., s e e Shigeoka, T. 401
onenobu, K., s e e Sugiyama, K. 371
Yoshida, H., s e e Kaneko, T. 397
Yoshida, M., s e e Sakamoto, . 182
oshida, M., s e e Takao, K. 182
Yoshie, H. 311
Yoshimoto, Y., s e e Yoshimura, K. 301
Yoshimura, K. 301,306, 321,323
Yoshimura, K., s e e Nakamura, H. 319
Yoshimura, K., s e e Sakakibara, T. 300, 301
Yoshimura, K., s e e Shiga, M. 323, 324
Yoshimura, K., s e e Wada, H. 303, 304, 317-319,
323-325
Yoshizaki, R., s e e 0nuki, Y. 95
Yoshizawa, H., s e e Mitsuda, S. 389
Yoshizawa, M., s e e Ikeda, K. 305

Yosida, K. 3, 415
Yosida, K., s e e Yamada, K. 5
Yosida, Y., s e e Sato, K. 362, 363
Yosuoka, H., s e e Yoshimura, K. 306
Yu, S.E, s e e Huang, G.X. 150
Yuldasheva, M.Kh., s e e Reshentnikova, L.V. 118
Yurasova, V.E., s e e Bachurin, V.I. 125
Yuzurihara, H., s e e Tanaka, M. 160, 161
Zachariasen, W.H., s e e Ellinger, EH. 110
Zajac, S. 125
Zalesky, M.P., s e e Jensen, C.L. 245
Zamir, D. 256
Zdansky, E., s e e Andersen, J.N. 116
Zema, N., s e e Sigrist, M. 126, 128
Zemirli, S., s e e Gratz, E. 87
Zemlyanov, M.G., s e e Parshin, P.R 243
Zerguine, M., s e e Gignoux, D. 369, 375, 376
Zhai, H.R., s e e Gu, B.X. 145
Zhang, EY. 339, 340, 342, 343, 361,403, 404
Zhang, EY., s e e Ball, A.R. 375, 377, 378, 380,
386, 404, 417
Zhang, EY., s e e Gignoux, D. 371,375, 377, 385,
387, 392, 393, 415
Zhang, EY., s e e Radwanski, R.J. 339, 342
Zhang, EY., s e e Shigeoka, T. 368
Zhang, X.J., s e e McGuiness, P.J. 148
Zhao, K.L., s e e Takao, K. 182
Zhao, Z.B. 152
Zhao, Z.B., s e e Pan, S.M. 175
Zhao, Z.R., s e e Aylesworth, K.D. 146, 150, 151
Zhao, Z.R., s e e Strzeszewski, J. 151
Zhavoronkova, K.N. 136
Zhavoronkova, K.N., s e e Boeva, O.A. 136, 143,
240
Zheng, C.H., s e e Huang, G.X. 150
Zhong, X.P., s e e Tomiyama, E 336
Zhou, S., s e e Tang, W. 150
Zhu, J.G. 186
Zhukova, T.B., s e e Smimov, I.A. 251,264
Zhuravski, V.E., s e e Bachurin, V.I. 125
Ziebeck, K.R.A., s e e D6portes, J. 323
Zochowski, S.W., s e e MeEwen, K.A. 391
Zogal, O.J. 221,225, 231,237
Zolandz, A., s e e Croft, M. 181
Zomack, M., s e e Baberschke, K. 122
Zomaek, M., s e e Farle, M. 123
Zukowska, K., s e e Idczak, E. 128
Zwicknagl, G. 14, 77, 78
Zygmunt, A., s e e Ivanov, V. 384, 385

SUBJECT INDEX
B-TbH(D)2+x 227
B/(13 +'~) phase boundary 263
B ---+ -~ transformation 214

acoustic lattice vibrations 239

zoustic phonons 248
activation energy 232
- for diffusion 232, 233
c~-ErHx 248
c~*-ErHx 268
-ErH(D)x 236
*-ErH(D)x 219
c~*-HOHx 267
~-LuDx 216
c~*-LuDx 219
c~*-LuHx 219, 246, 271
c~-LuH(D)x 235
c~*-LuH(D, T)x 244
c~-phase 208
c~*-phase 211
c~-RHx 215, 244
cd-RH(D)x 219, 232
*-RH(D, T)x 220
c~-ScDx 216
cd-ScDx 219
c~-ScHx 232
c~*-SCHx 271
e~*-ScH(D)x 219
c~-TmDx 216
cx-TmHx 248
c~-YDx 216
(x*-YDx 219
~x-YHx 215
(x*-YH(D)x 244
--+ ~ transformation 212, 214
anionic model 243
ANNNI model 399
anomaly temperature 219
anticrossing 351,360
antiferromagnetic (AF) interactions 257
antiferroquadrupolar ordering 359, 394, 408
antiphase structure 373, 385, 409
band structure calculations,
calculations
B-phase 208
~3-RH2_ 6 213
-RH2 244
~-RH2+x 212, 244

see

Ce-H 214
CeAI2 68-70
CeAs 60, 65, 67
CeB6 31, 34-36
CeBi 60
CeCu2 88, 90
CeCu6 91, 92, 95
CeCu2Si2 82
CeD2+x 225
CeGa2 71-73
CeH2+x 209, 212, 221,256
CeH3_x 209
CeH(D)2+x 229
CeIn3 41, 44-46, 48
CeNi 83-86
CeRu2Ge2 75, 76, 78
CeRu2Si2 75, 76, 78-80
CeSb 60, 62, 64, 65
CeSn3 52, 54, 56, 58
(Co/(Pr, Nd))n films 162
(Co/Dy)n films 164
(Co/Gd)n films 162
(Co/R)n films 162
Co-Er quasicrystals 185
colleetive electron metamagnetism 298-308
commensurate structure 409
compensated metal 26
critical field 300, 332, 349
critical point 346
critical temperature 212
cross-sectional area of Fermi surfaee 28
crossing 350, 351,353, 360
crystalline electric field (CEF) 209, 267, 338,
350, 367, 373
crystalline state of R films on substrates 122
crystallographic parameters 218
crystallographic properties of R-Pd alloys 182
cyclotron mass 28

energy band

de Haas-van Alphen (dHvA) effect 27

- CeA12 8, 67-70
457

458

SUBJECT INDEX

de Haas-van Alphen (dHvA) effect (cont'd)

- CeAs 8, 60, 65, 67
CeB 6 8, 31, 33-35
Cei 8, 60, 62, 65
CeCu2 8, 87-90
- CeCu6 8, 90-95
CeCu2Si2 8, 81, 82
- CeGa 2 8, 70-73
- Celn 3 8, 40-46, 48
CeNi 8, 82-87
CeRu2Ge2 8, 75, 76
CeRu2Si2 8, 75-81
- CeSb 8, 60, 62-65
- CeSn 3 8, 52-59
frequency 28
Gdln 3 8, 41, 48, 50, 51
- GdSb 8, 60, 65, 66
- LaAg 8, 38, 39
- LaA12 8, 67-69
La6 8, 30-35
LaBi 8, 60, 61
- LaCu 6 8, 91-95
- LaCu2Si 2 82
- LaGa2 8, 70-73
- Laln3 8, 41-43, 48
LaNi 8, 82-87
- LaRu2Ge2 8, 74-76
- LaRu2Si 2 8, 74-76, 81
- LaSb 8, 60, 61, 65
- LaSn3 8, 29, 52-54, 56-59
NdB6 8 , 3 1 , 3 5 - 3 7
NdCu6 8, 90-93, 95
- Ndln 3 8, 41, 47-49
PrB 6 8, 31, 33-35
- PrCu 6 8, 90-92, 95
Prln 3 8, 41, 47, 48
PrNi 8, 82, 83, 86
PrSb 8, 60, 65, 66
- SmCu2 8, 87, 88, 90
- SmCu6 8, 90-93, 95
SmGa2 8, 71, 73, 74
Smln 3 8, 41, 48-50
SmSb 8, 60, 65, 66
- YA12 8, 68, 69
- YCu2 8, 87-89
- YZn 8, 38-40
- YbAs 8, 60, 65, 67
demagnetizing field 347
deuterides 140, 141,216-265, 272, 274-277,
280-283
devil's staircase 62, 400
dHvA, see de Haas-van Alphen effect
Dingle temperature 28
DyH2+x 230, 260
-

easy-axis anisotropy 367, 401

easy-plane anisotropy 335, 390
eleetric resistivity in LaNi 5 170
electrical properties of RH 2 137-139
electrieal properties of R metal films 120
electronic configurations
gaseous atom (vapor) 112
metallic 112
neutral atom 10
- trivalent ion 10
electronic specific-heat coefficient 6, 31, 38, 41,
52, 60, 68, 71, 75, 83, 88, 91,236, 244
endpoint 346, 376
energy band calculations
- CeA12 8, 67-70
- CeAs 8, 60, 65, 67
- Ce6 8, 31, 33-35
CeBi 8, 60, 62, 65
- CeCu 2 8, 87-90
- CeCu6 8, 90-95
- CeCu2Si2 8, 81, 82
- CeGa z 8, 70-73
- CeIn 3 8, 40-46, 48
- CeNi 8, 82-87
CeRu2Ge2 8, 75, 76
- CeRu2Si2 8, 75-81
- CeSb 8, 60, 62-65
- CeSn 3 8, 52-59
- Gdln 3 8, 41, 48, 50, 51
- GdSb 8, 60, 65, 66
- LaAg 8, 38, 39
- LaA12 8, 67-69
LaB 6 8, 30-35
LaBi 8, 60, 61
- LaCu 6 8, 91-95
- LaCu2Si2 82
- LaGa2 8, 70-73
- LaIn3 8, 41-43, 48
LaNi 8, 82-87
- LaRu2Ge2 8, 74-76
- LaRu2Si 2 8, 74-76, 81
- LaSb 8, 60, 61, 65
- LaSn 3 8, 29, 52-54, 56-59
- NdB6 8, 31, 35-37
- NdCu6 8, 90-93, 95
- Ndln 3 8, 41, 47-49
- PrB6 8, 31, 33-35
- PrCu6 8, 90-92, 95
PrIn 3 8, 41, 47, 48
PrNi 8, 82, 83, 86
- PrSb 8, 60, 65, 66
SmCu2 8, 87, 88, 90
- SmCu 6 8, 90-93, 95
SmGa2 8, 71, 73, 74
-

SUBJECT INDEX

energy band calculations (cont'd)

SmIn 3 8, 41, 48-50
SmSb 8, 60, 65, 66
- YA12 8, 68, 69
- YCuz 8, 87-89
- YZn 8, 38~40
YbAs 8, 60, 65, 67
energy gap 251
epitaxial crystal growth 118
ErHz+x 239, 2 5 1 , 2 6 1 , 2 6 3
Er0.sHo0.5 Rb.4B4 184
ErRh4B4 184
exchange interaction 330, 334, 414
external field 347

- carrier hopping 264

- characteristic magnetic temperatures 272, 274,
275
charge transfer 265
eoherent-incoherent transformation 275
eommensurate components 284
commensurate phase 268, 281
complex antiferromagnetic (AF) structures
277, 284
- Compton profiles 265
critical concentrations 211
- crystal field (CF) 209, 267
- - splitting 277
- cubic-to-tetragonal deformation 214
- de Germes faetor 273, 286
"defect" migration 236
- "defect"-recovery 237
deloealized band 251
density of stares 248
- differential thermal analysis (DTA) 227
dihydrides (f3-phase) 208-214, 221-232,
236-244, 248-267, 271-285
- EPR 256
effective gap 263
- electrical properties of RH2 137-139
electron-phonon coupling 239
- electronic properties 243-265
- enthalpy for dihydride formation 239
- f-electmn spins 273
- ferromagnetie (FM) interactions 257
fluctuating valence 263
formation enthalpy 241
5-phase 208, 214, 231
gap 251
- Gorsky effect 232
- ground state 277
- grormd-state magnetic moments 273
ground-state resistivities 273
Gffmeisen law 239
- Grneisen temperamre 239
heat capaeity 282
- heavy fermions 209
hydrogen sublattice 209
- hydrogen trapping 227
- hydroxides 213
- incommensuracy 273, 279
incommensurate
- components 284
- - magnetic structure 277
- phase 281
- intermediate structure 282
ionicity 236
isothermal plateau 213
kinetie properties 232-243
-

fan structure 410

fcc phase of R metals 133
(Fe/Gd)n films 159
(Fe/R)n films 157
(Fe/Tb) films 160
(FeFFm)n films 161
Fermi surface
intermetallie compounds 1-98
metals with dissolved H 243
ferromagnetic films 125
ferromagnetism 338, 359
ferroquadrupolar ordering 356, 357, 408
first-order magnetie transition 356
first-order transition 345, 347
flopside strueture 394, 411
frustration 310, 316, 411

gadolinium compounds 363

GdH2+x 221,230, 251,259
GdH(D)2+x 230, 237
GdIn 3 41, 48, 51
GdSb 60, 65, 66

heat capacity, see also electronic specific-heat

eoefficient 282
helical structure 364, 412
HoA1Ga 385, 403
HoD x 246
HoD 3 231
Hort x 246
HoH2+x 239, 251,261
hydrides 135-140, 208-215, 221-231,236-243,
248-265, 271-286
see also ~3-XX a n d the specific rare-earth
hydride
atomic configuration 248
atomic volume 235
CF-excitations 273

459

460

SUBJECT INDEX

hydrides (cont'd)
Kondo
- - effect 210, 273, 276
lattice 209
- - m a x i m a 276
- - m i n i m a 276
- - transition 275
- X-type peaks 278
lattice expansions 221
lattice parameters
- - 13-RH2 222-225
- "/-RH2+x 231
liquid phase 213
- localized states 251
localized vibrational modes 241
magnetic anomalies 276
- magnetie configuration 273
magnetic fluctuations 283
- magnetic ground stare 277
magnetie ordering 260
- magnetic phase diagram 257, 273
magnetie properties 271-285
- magnetic transformation 273
metal-insulator ( M - I ) transition 209, 239
metal-semiconductor (M-S) transition 244,
251
mixed valence 264
- modulation 275
- l~+-diffusion 236
muons 232
- neutron scattering 273
- nonmagnetic first excited stare 277
O sites 214
octahedral H atoms 209
octahedral site 208
optical phonons 248
- optieal spectroscopy 265
- optical vibrations 243
- order--disorder transformation 209
- ordering temperatures 227
- orthorhombic CF distortion 277
- orthorhombie dihydrides 231
orthorhombic structure 208
overlap region 284
paramagnetism 209
percolation 264
- perturbed angular correlation 261
- phase diagram 209, 214, 215
- phonon scattering 277
- photoelectron spectroseopy 265
- potential barrier 237
preparation 2 1 0 - 2 1 4
- propagation vector 275
- proton N M R 277
-

quench-induced disorder 261

quenching effects 281
- R K K Y exchartge interaction 209, 277
- resistivity isothermals 247
semiconduetor-metal (S-M) transition 254
- short-range ordered spin fluctuations 273
single erystals 212
sinusoidal incommensurate magnetic structure
281
solubility limit 213
solution enthalpy 239
- solution entropy 239
- spin axis 283
spin disorder 248
spin-disorder resistivity 271
- spin glass 264, 277
- spin wave 257
- stoichiometric composition 227
stoichiometrie defieit 212
structural disorder 278
structural transformations 239
- structure 216-231
- sublatfice ordering 214
- superconductivity 256
- superstoichiometric dihydrides RH2 +x 209
T sites 214
tetragonal distortion 221
tetrahedral H atoms 209
- tetrahedral site 208
- thermal conduetivity 264
- thermodynamie properties 236-243
- thermoelectric power 264
- trihydride (',(-phase) 208, 214, 231
- uniaxial anisotropy energy 277
- Van Vleck compound 285
- Van Vleck paramagnet 285
variable-range hopping 251
- weak localization 254
x-atoms 214
x-sublattice 221
hydrogen absorption in LaNi5 169
hydrogen dissolved in R (a-phase) 208-221,
232-236, 244-248, 2 6 7 - 2 7 1 , 2 8 6
see also (x-XX and the speeific rare-earth metal
(e.g. La-H)
- b-axis 219
basal plane 219
- binding energy 219, 221
- blocking effects 232
c-axis 219
- c-axis modulated (CAM) structure 268
c/a-ratio 217
- earrier density 243
- chain ordering 235
-

SUBJECT INDEX
hydrogen dissolved in R (a-phase) (cont'd)
chains 219
- conduetion-electron density 243
eritieal temperature for phase transitions 211
erystallographic parameters 218
D - D pairs 219
- Debye temperature 232
- deuterium 221
diffuslon process 234
effective magnetie moments 267
Einstein temperature 232
eleetron density 266
electron irradiation 232
electron-phonon coupling 244
electronie density 244
electronic properties 244-248
electronie speeifie-heat coeffieient 236, 244
electronic structure 265
- energy for "defect" formation 234
expansivity 216
- Fermi surface 243
- ferrimagnetie ordering 270
ferrimagnetic-ferromagnetie transition 269
- ferrimagnetism 269
ferromagnetic eone 268
Gorsky effect 232
H - H pairs 215
H-eonfigurations 220
- H-"defeet" reeovery 232
heat ofsolution 235
helieoidal strueture 268
- "hydridie" model 243
- hydrogen "defects" 232
hydrogen isotope 244
- hydrogen loeal mode energies 242
- hydrogen ordering 209
- hydrogen solubility 211
- hydrogen ttmneling 271
- hydrogenation 211,213
- inelastie energy loss 235
inelastic relaxation 232
interatomie potential 219
intermediate phase 268
- internal frietion 210
isothermals 244
isotope effect 219
- kinetic properties 232-236
lanthanide contraction 211
- local mode energy 235
- Iocalized motion 232
lock-in transitions 268
long-range-ordered (LRO) struetures 219
- magnetic anomaly 267, 268
- magnetic properties 267-271
-

461

- magnetic structures 267

- magnetic superzone 270
- magnetic susceptibility 247, 267
magnetic transitions 267
magnetization 268
magnetoelastic interactions 267, 268
magnon 268
magnon excitation 271
- migration energy 233
Nel temperature 246
- nesting 243
neutron seattering 209
order~tisorder process 234
- pair-binding energy 232
- paramagnetic Curie temperature 267
paramagnetism 269, 271
peak temperature 232
phase diagram 209, 214, 215
photoelectron spectra 266
- preparation 210-212
- protonic model 243
- quenehing 232
RKKY exchange interaction 209, 266
- recovery stage 232
relaxation times 232
residual resistivity 245
resistivity anomaly 219
- short-range order (SRO) 219
solid solution 208
solubility limit 211
- spectroseopy 244
spin fluctuations 271
spin-slip structure 268
spin-wave exeitation gap 271
- structure 216-22t
- superconductivity 244
superstructure 219
- thermodynamic properties 232-236
- tritium 221
- tunneling 232
hydrogen gaseous contarninants in R - T M films
167
hydrogen separation and permeation in LaNi 5 171
hysteresis 349, 389, 406
-

incommensurate magnetic structure

370, 399-406, 409
internal field 347
ion implantation 167
itinerant ferromagnetism 308

277, 365,

Kondo
- effect 4 ~ , 210, 273, 275, 276, 285
- lattice 6, 209, 285

462

SUBJECT INDEX

Kondo (cont)
- maxima 276
minima 276
regime 4-6
transition 275
Kondo lattice compounds
- CeA12 6, 69, 70
- Ce6 6, 30, 31, 33-35
- CeCu2 6, 34, 87, 89, 90
- CeCu6 6, 90-95
- CeCu2Si 2 6, 81
Celn3 6, 40-46, 48
CeNi 6, 83-86
CeRu2Si 2 6, 75-81
- CeSb 6, 59-65
- CeSn 3 6, 52-59
-

La-H 214
LaAg 38, 39
LaAlz 68, 69
La 6 31, 32, 34, 35
Lai 60
LaCo 1_ ~ 309
LaCu 6 91, 94, 95
LaCu2 Si2 82
LAD2+ x 255
LaGa2 71-73
LaH2_x 256
LaH 2 256
LaH2+ x 2 2 1 , 2 5 1 , 2 5 6
LaH3 - x 209
LaH(D)2+x 229
Laln3 41-43, 48
LaNi 83-86
LaNi5 169
- electric resistivity 170
- hydrogen absorption 169
hydrogen separation and permeation 171
LaRu2Ge2 75, 76
LaRu2Si2 75-77
LaSb 60, 61, 65, 67
LaSn 3 29, 52, 53, 56, 58
lead-europium films 188
lead-ytterbium films 190
long-period commensurate structure 370
Lu-H 214
Lu2Fe 3Si 5 185
LuH2+x 264
LuRhl.2Sm4 184

magnetic
magnetic
magnetic
magnetic

anisotropy 310-316, 322, 331-337

domain structure 126, 147
instability 297, 317, 318
phase diagram 257, 273, 345-398, 401

magnetic properties
see also specific topic
- of H in R 267-271
- of HoA1Ga 385, 403
- of incommensurate magnetic systems 277,
399-406
ofintermetallic compounds 6, 31, 38, 41, 52,
60, 68, 71, 75, 83, 88, 91,293-417
- of LaCol_e 309
ofmetamagnets 298-308, 336, 345-398
of PrCo2Si2 380, 401
- of PrGa2 377, 380, 391
- ofPrNi 5 351
- of PrNi2Si 2 373, 403
o f R alloys 145
- o f R C o 2 299, 303, 304
- of RCo5 311
o f R H 2 140, 271-285
of RMn2 316
o f R N i 5 338
- of TbNi2Si 2 403
- of ThCo5 301
of thin films
alloys 144-167
- metals 122-126
- of TmGa3 397
- of TmSb 350
- o f Y N i 3 308
magnetic state 298, 319, 325
magnetic structures 267, 345-398
- amplitude-modulated 370, 373, 386, 408
- specific heat 405
- anisotropy 269
magnetlc transitions, see also metamagnetic
transition and metamagnetism 267
magnetocrystalline anisotropy 310-316, 322-337
magnetoelastic effects 303, 317
magnetostriction 303
mass enhancement factors 24
metallic R alloys 144
metalloid R-compounds 186
metamagnetic transitlon 298, 300, 306, 336
metamagnetism 345-398
collective electron 298-308
- multistep 380, 392, 401,413
mixed structure 320, 326
modulated and multilayered films 157
multi-Q structure 413
multiaxial structure 394, 413
multilayer (Er, Tm/Lu)n films 125
multilayer systems, other 165
-

NdB6 31, 35-37

NdCu6 91-93, 95

SUBJECT INDEX
NdFeB
crystallographic properties 150
hydrogen treatment 148
NdklFe4B4 150
Nd2Fe14B 145
NdI-I2+x 229, 239, 258
NdIn3 41, 48-50
nesting conditions 279
neutron diffraction 273
(Ni/Ce)n layers 165
(Ni/Dy)n layers 165
(Ni/R), films 165
Ni3Mo-type structure 209, 227
noncollinear structure 336, 362, 365, 391,395,
409
noncompensated antiphase structure 375, 377
nonmagnetic state 298, 317, 319, 325
-

optical properties of R metal films

optieal spectra (0-t200eV) 126
periodic field model 402
permanent magnets 144
phase diagram 208, 210
magnetic 346
piezoresistance of foils 132
polytypic structures 172
Pr-H 214
PrB6 31, 34, 35
PrCo2Si 2 380, 401
PrCu6 91, 92, 95
PrGa2 377, 380, 391
PrH2+x 239, 258
PrIn3 41, 47, 48
PrNi 83, 86
PrNi 5 351
PrNi2Si 2 373, 403
PrSb 60, 65, 66
-

quadrupolar interactions 355

quadrupolar moment 413
quasicrystal Co Er 185
R-H 207-286
R-hoble metal alloys (Cu, Au) 177
R-Pd alloys 181
- crystallographic properties 182
valence change in 181
RA12 68
RAs films 186
RBi films 188
R clusters 114
RCo2 299, 303, 304
RCo5 144,311
- hydrogen storage films 169

126

R2Co17 333
RCo(B) amorphous films 156
R crystallographic properties 118
RH(D)0.2 235
RKKY exchange interaction 209, 266, 277
R layers on crystals 115
RMn2 316
RNi5 144, 338
- hydrogen storage films 169
RsOs4Si15 185
R/Re systems 119
RSb films 188
RTiFe(Co) alloys 155
R/V systems 119
R/W systems 118
rare earth metal films 111
reactivity with CO, CO2 and CnHn 143
rare earth nitrides 142
rare earth trialuminides 177
-

samarium-cobalt polytypes 174

samarium-nickel polytypes 173
ScH(DL 245
scattering lifetime 28
second-order transition 346
singlet ground stare 350, 373
Sm2(Co, Fe, Zr)17 145
SmCu2 88
SmCu6 91-93, 95
SmGa2 71, 73
SmH2+x 221,258
SmIn3 41, 48, 51
SmSb 60, 65, 66
spin-flip transition 367, 373, 380, 415
spin-flop transition 363, 416
spin fluctuations 304, 323
spin-slip transition 379, 380, 417
Stoner factor 298
structural properties 210
superconducting materials 184
TbD2 213
TbD2+x 230
TbH2 +x 259
TbNi2Si2 403
ThCo5 301
thin films 105-190
- addition effects of elements 152-155
- amorphous 121, 122, 125, I26, 156, 157
- hydrides 135-140, 169-172
metallic alloys
- n o b l e metal 177-181
palladium 181-184
- permanent magnet 144-157
-

463

464

SUBJECT INDEX

thin films (cont'd)

metalloid 143, 144, 186-190
metals
- - crystallography 118, 119
- - electrical properties 120-122
- - magnetic properties 122-126
- - optical properties 126-132
- - reactivity 133, 134
third-order magnetic suseeptibility 353, 355
TmGa3 397
TmSb 350
transverse magnetoresistance 26
triangular structure 309
tritides 140, 141,221
-

U-R films 185

uncompensated metal

26

valence change in R-Pd alloys 181

valence-fluctuation regime 6
valence in vapor and solid stare 111
very weak itinerant ferromagnetism 308

Y, La-Pb multilayers 185

Y-H 214
YA12 68
YCu2 88, 89
YH2+ x 209, 221,251
YH(D)x 246
YNi 3 308
Y4Os4SiI3 185
YZn 38, 40
YbAs 60, 65, 67
YbH2+x 263