Full text of DOI:10.1557/mrs2001.178

www.mrs.org/publications/bulletin
this last, and least explored, row of the
periodic table. We emphasize our efforts
on Np, Pu, and Am, and how such studies
can enhance our understanding of the
complete series. We start with the impor-
tance of preparation and then describe
measurement techniques including those
to characterize structural, magnetic, trans-
port, and surface properties.
Understanding
Actinides through
the Role of
Preparation andTransport Studies
Our effort starts with preparation. Many
different materials science techniques exist,
but our laboratory has devoted a consid-
erable effort to producing high-purity single
crystals of actinide compounds,1 which
these high-level physics experiments de-
mand. Techniques such as mineralization,
zone-refining and the Czochralski process
for the metallic compounds, and vapor
transport for the oxides, can be used to
produce crystals and to increase the purity,
thus assuring the stoichiometry. Some of
these materials are shown in Figure 1. The
crystals are then characterized by x-ray,
scanning electron microscopy, and micro-
probe analysis techniques.
5f Electrons
T. Gouder, F. Wastin, J. Rebizant,
and G.H. Lander
Introduction
Studies of the actinide elements and
compounds were (and are) motivated by
the need to characterize their structural
and thermodynamic properties for the de-
velopment of nuclear fuels and the treat-
ment of waste, whether it be for long-term
storage or ideas involving transmutation
in high-powered accelerators. For the most
part, tables giving these data exist, although
the data for transuranium compounds are
rather sparse. A much more difficult task
is to understand the data and develop theo-
ries that have predictive power in this part
of the periodic table. In doing this, how-
ever, we are confronted with the extremely
complicated electronic structure of the 5f
shell and the great paucity of high-quality
data on materials containing transuranium
isotopes.
article focuses particularly on work from
our own laboratory. We give a brief review
of the basic research activities on actinide
elements and compounds, revealing the
varying nature of the 5f electrons across
Akey transport property is the electrical
resistivity of a material, measured both as
a function of temperature and pressure.
Research on uranium, both the element
and compounds of it, is conducted in many
parts of the world because uranium is a
nontoxic material (at least in the quantities
used for basic research) and can be handled
in university laboratories. Our lab has proj-
ects on uranium because of our collabora-
tions with European laboratories (as well
as laboratories in the United States and
Japan); however, our primary focus is on
transuranium materials, which cannot be
investigated in many places.
Because our European Commission lab
is situated in Karlsruhe at an institute
that specializes in transuranium materials,
mostly connected with the nuclear-fuel in-
dustry in the areas of fuel development, fuel
characterization, or aspects of the nuclear-
waste problem, we have an infrastructure
for handling transuranium isotopes (prin-
cipally Np, Pu, Am, and Cm). Thus, this
Figure 1. (a) Single crystals of mixed (U, Np)O2 oxide obtained by the vapor-transport
technique. (b) Single crystals of PuSb grown by mineralization. (c) Czochralski growth
of crystals of the intermetallic PuFe2. (d) Pu-Am cast for resistivity and magnetization
measurements.
684
MRS BULLETIN/SEPTEMBER 2001

Understanding Actinides through the Role of 5f Electrons
Figure 2 shows the resistivity of PuTe as
a function of temperature and pressure.2
These experiments require very small
samples, for example, a small flake of the
order of 100 ꢀm in size, which has to be
cleaved inside a glovebox from a larger
piece, and mounted between diamond
anvils with the necessary leads. PuTe is
particularly interesting from a basic point
of view, as the 5f states form a narrow
band with strong correlations between
them. Surprisingly, none of the Pu mono-
chalcogenides, including PuTe, show
magnetic order at low temperature, and
this aspect of their behavior, and the fun-
damental difference between them and
comparable lanthanide compounds, has
interested theorists for many years.3,4
More recent theories4 are able to explain
the initial increase in the resistivity (see
Figure 2) with increasing pressure and
even the sudden decrease in the resistance
when PuTe transforms5 from the NaCl
(B1) to the CsCl (B2) crystal structure at
about 11 GPa. Moreover, they predict that
Figure 3. Relative volume V (where V0 is the volume at zero pressure) versus pressure
curve for americium. The inset shows the atomic volume at ambient pressure across the
actinide series. (From Reference 8.)
the high-pressure phase of PuTe will be
magnetic. There is a chance that this has
been seen2 as a slight inflection point in
the resistance versus temperature R(T)
curve (see the arrow marked at 15 K in
Figure 2), but this needs to be confirmed
by other measurements.
and Pu is commonly attributed to the fact
that the 5f electrons, which are added as
one progresses across the series from Pa
onward, participate in the bonding of the
element and thus contribute to the cohe-
sive properties and decrease the volume.
A similar effect is found in the transition-
metal series, where the cohesion involves
d electrons. There is, however, a sudden
increase in the volume at Am and then a
progression that appears to suggest a rare-
earth-like series, in which the 5f states are
localized and do not participate in the bond-
ing.6 Pu and Am might then be capable of
exhibiting a transition to the “other” state,
larger in the case of Pu and smaller in the
case of Am.
We shall discuss in the following section
how one can make Pu larger (by examin-
ing Pu atoms on the surface), but here we
concentrate on making Am smaller, which
is best done with pressure experiments. Of
course, this has been attempted for many
years,7 but recently8 new synchrotron ex-
periments have shown the richness of the
phase diagram of Am under pressure. The
details of the volume versus pressure curve
are shown in Figure 3. The delocalization
of the 5f states in Am actually starts at
about 15 GPa (150 kbar), where there is
a transformation from Am(II), which is
fcc, to Am(III), which has the same face-
centered orthorhombic structure as the
high-temperature form of ꢁ-Pu. There is
then a further contraction of the volume
(by ꢀ7%) when entering the Am(IV) phase,
which remains up to 100 GPa. These re-
sults are in disagreement with theory,9 but
work is in progress to reconcile these dif-
ferences. It should be realized that experi-
ments of this sort, which are now under
way on Cm and Cf, provide an enormously
important test for first-principles theories.
These theories must take into account all
the complications of the relativistic nature
of the 5f electrons, together with the strong
spin-orbit corrections, and they are at the
frontier of condensed-matter theory today.
Structural Studies at
High Pressures
As mentioned in the article by Hecker
in this issue, across the elements in the
actinide series there is an extraordinary
change of volume (shown in the inset of
Figure 3), which has been known for more
than 50 years. The minimum for U, Np,
Surface Studies
Our surface characterization is based
mainly on photoelectron spectroscopy tech-
niques, namely, XPS (x-ray photoelectron
spectroscopy) and high-resolution UPS
(ultraviolet photoelectron spectroscopy),
which also provide direct information
about basic features of electronic structure,
such as the degree of delocalization of the
5f states. The spectroscopy study can be
performed on bulk materials, but a real
breakthrough was the implementation of
a sputter-deposition technique that allows
the preparation of thin layers of trans-
uranium materials. This technique yields
materials that do not suffer from surface
segregation of oxygen, which seriously af-
fects surfaces of bulk samples. Moreover,
it provides larger flexibility in variations
of stoichiometry and microstructure, and
it allows us to follow the variations of
Figure 2. Temperature-dependence
of the electrical resistance of PuTe at
selected pressures of up to 24.3 GPa.
(From Reference 2.)
MRS BULLETIN/SEPTEMBER 2001
685

Understanding Actinides through the Role of 5f Electrons
properties when the thickness is reduced
down to a monoatomic layer.
ing. The comparison with spectra taken at
lower photon energy (hꢂ ꢀ 21.2 eV), for
which the 5f photoexcitation cross section
is negligible, demonstrates that all three
features are related to the 5f electronic
states. But the mechanism responsible for
the sharp features still remains unclear. As
they also appear in some other Pu-based
systems (at identical energies but various
intensities), it seems that they are related
not to details of a particular band struc-
ture, but more probably to a rather general
many-electron phenomenon.11
The interesting position of Pu metal just
at the verge of the 5f localization is re-
flected in dramatic variations of its elec-
tronic structure with the crossover from
three-dimensional to two-dimensional sys-
tems. Figure 5 shows the spectra of the
valence band, recorded for a photon en-
ergy of 40.8 eV for Pu deposited on a Mg
substrate.11 As Mg does not react or inter-
mix at the interface during the sputter-
deposition process with Pu, ultrathin Pu
layers of controlled thickness can be pre-
pared. The features of this dramatic locali-
zation process can be observed in the
electronic structure below five monolayers’
thickness. For the “thick” layer, one can
see the spectral intensity representing a
5f band, increasing up to the Fermi level
cutoff. The reduced thickness leads to a new
spectral feature around 1.7 eV below the
Fermi energy EF; at the same time, the itin-
erant 5f states leading to the high intensity
at EF are gradually suppressed. This broad
maximum is undoubtedly of 5f origin, as
shown by comparison with spectra of dif-
ferent photon energies, and corresponds
roughly to the 5f emission found in PuSb.
We interpret this broad maximum, with
the support of the Pu-4f core level spectra,
as being due to the emission from the 5f
states, which become localized as a result
of the reduced Pu coordination number in
the monoatomic layer. For such a Pu layer,
one can thus expect very different cohesion,
lattice parameters, and other physical
properties as compared with the bulk. For
the intermediate state of the localization
process, one can clearly distinguish the
three features close to the Fermi level, ob-
served already in PuSe.
The 5f localization due to the reduced
thickness, which diminishes the number
of nearest neighbors and thus the effective
5f–5f overlap, can be understood as the
opposite process to the loss of localization
of Am under pressure (at which the com-
pression increases the overlap of the 5f
wave functions centered on nearest neigh-
bors). The reduced number of nearest Pu
To gain information on the electronic
structure of Pu monochalcogenides, we
studied PuSe,10 which has properties very
similar to PuTe, as discussed earlier. Fig-
ure 4 shows the valence-band spectra of
PuSe layers deposited on Si, in compari-
son with PuSb and Am, the latter having
the 5f states localized, yielding the maxi-
mum intensity at about 2-eV binding en-
ergy. The low density of states at the Fermi
level is indicated by a weak emission
around zero energy. In contrast to this situ-
ation, PuSe shows three sharp features at
low binding energies, the highest one lo-
cated at the Fermi level. These three features
are strongly dependent on the quality of
the surface. They are reduced at an excess
of Pu and disappear for deposition on a
substrate cooled to T ꢀ 77 K (i.e., for pre-
sumably amorphous PuSe). The same varia-
tions are observed for bulk PuSe and for a
surface depleted by Se due to Ar ion etch-
Figure 4. Ultraviolet photoelectron
spectroscopy valence-band spectra
of (a) PuSb, (b) PuSe, and (c) Am.
Blue lines represent spectra for
hꢂ ꢀ 40.8 eV, black lines represent
spectra for hꢂ ꢀ 21.2 eV. For PuSb (a),
the red line represents the difference
spectrum. For higher binding energies
than 5 eV, the difference spectrum is
negative due to a higher secondary-
electron background for hꢂ ꢀ 21.2 eV
in this energy range.
Figure 5. Photoelectron spectra of valence band of Pu layers of variable thickness, recorded
with 40.8-eV photon energy. The spectral feature originating from the localized 5f states is
emphasized by the darker shading. (From Reference 11.)
686
MRS BULLETIN/SEPTEMBER 2001

Understanding Actinides through the Role of 5f Electrons
neighbors should, however, affect the 5f
localization, even in the surface layer of
bulk Pu metal. We indeed observed, due
to the extreme surface sensitivity of photo-
electron spectroscopy, that the freshly pre-
pared ꢃ-Pu surface adopts the character of
ꢄ-Pu with more localized 5f states. The fact
that the kinetics of this process is increas-
ing with increasing temperature points to
a real reconstruction of surface crystallog-
raphy in favor of a structure accommo-
dating more localized (i.e., bigger) surface
Pu atoms. Interestingly, the stability of the
ꢄ-Pu surface at the surface of ꢃ-Pu was
predicted theoretically.12
availability of single crystals; this is a major
motivation for crystal production in our
laboratory.
Characterization of magnetic properties
initially requires a magnetometer. Detailed
magnetization curves for crystals of less
than 1 mg have been obtained with a
SQUID (semiconducting quantum interfer-
ence device) magnetometer. For the mag-
netism of Np compounds, the Mössbauer
The capabilities of reactive sputtering,
during which the deposited material is ex-
posed to a gas in ionized atomic form, can
be used to prepare and study systems that
cannot be prepared by standard vacuum
techniques. Thus, for example, layers of
higher nitrides and oxides could be syn-
thesized for U and Th (U2N3, Th3N4). In
analogy to uranium, for which UO2 pre-
pared by exposure of U layers to molecu-
lar oxygen can be further oxidized to UO3
by atomic oxygen, we were able to follow
how such an oxidation process increases
the O content in PuO2. The existence of a
higher Pu oxide (presumably water-soluble)
has serious consequences for nuclear-waste
storage considerations.13
Magnetism and Its Many Forms
The filling of the 5f shell and the subse-
quent application of Hund’s rules lead to
the prediction of magnetism in the actinide
series, just as is found in the lanthanides.
Surprisingly, none of the elements are mag-
netic until one reaches Cm. This is now
understood exactly in the same manner as
already explained in the section on “Struc-
tural Studies at High Pressures” for the
participation of the 5f states in the cohesive
properties. The 5fstates in the early elements
form bands, rather than localized states,
and these bands are too wide to sustain
spontaneous magnetism. Arguments to
understand this are based on work by
Stoner (and others) on the 3d series and
date back 50 years.6 However, the situa-
tion is more complicated in the actinides
as, in addition to the spin moment, the
5f states also have a large orbital contribu-
tion to the magnetism simply because, as
they are f electrons, they have a large an-
gular momentum term (l ꢀ 3).14 Magnet-
ism does appear in many U, Np, and Pu
compounds, and fundamental studies of
such magnetic properties15 represent an-
other perspective on our efforts to under-
stand the 5f electrons and their interaction
both with each other and with other elec-
tron states in the solid. An important re-
quirement for magnetic studies is the
Figure 6. The upper diagrams show the triple-k (left) and single-k (right) magnetic
configurations. The main figure shows a schematic phase diagram for the (U0.25Pu0.75)Sb
sample as a function of field (H) and temperature (T). Note that the Néel temperature TN
is ꢀ90 K. Tꢅ and Hc refer to the temperature and critical field, respectively, at which the
configuration changes from single-k to triple-k. Red (blue) areas imply that the triple-k
(single-k) configuration dominates. (From Reference 19.)
MRS BULLETIN/SEPTEMBER 2001
687

Understanding Actinides through the Role of 5f Electrons
Earths, Vol. 17, edited by K.A. Gschneidner, L.
Eyring, G.H. Lander, and G. Choppin (Elsevier,
Amsterdam, 1993) p. 245.
8. S. Heathman, R.G. Haire, T. Le Bihan, A.
Lindbaum, K. Litfin, Y. Meresse, and H. Libotte,
Phys. Rev. Lett. 85 (2000) p. 2961.
9. P. Soderlind, R. Ahuja, O. Eriksson, B. Johans-
son, and J.M. Wills, Phys. Rev. B 61 (2000) p. 8119.
10. T. Gouder, F. Wastin, J. Rebizant, and
L. Havela, Phys. Rev. Lett. 84 (2000) p. 3378.
11. T. Gouder, F. Wastin, J. Rebizant, and
L. Havela, Europhys. Lett. (2001) in press.
12. O. Eriksson, L.E. Cox, B.R. Cooper, J.M.
Wills, G.W. Fernando, Y.G. Hao, and A.M. Boring,
Phys. Rev. B 46 (1992) p. 13576.
13. J.M. Haschke, T.H. Allen, and L.A. Morales,
Science 287 (2000) p. 285.
14. M.S.S. Brooks and P.J. Kelly, Phys. Rev. Lett.
51 (1983) p. 1708.
15. J.M. Fournier and R. Troc, in Handbook on
the Physics and Chemistry of the Actinides, Vol. 2,
edited by A.J. Freeman and G.H. Lander
(Elsevier, Amsterdam, 1985) p. 29.
16. W. Potzel, G.M. Kalvius, and J. Gal, in
Handbook on the Physics and Chemistry of the Rare
Earths, Vol. 17, edited by K.A. Gschneidner, L.
Eyring, G.H. Lander, and G. Choppin (Elsevier,
Amsterdam, 1993) p. 539.
interactions drastically change the mag-
effect with the 237Np nucleus16 may be ef-
netic properties.
fectively used to gather useful information,
and this is used routinely with Np com-
pounds. Other techniques important in
magnetic studies are resistance and mag-
netoresistance. A specific-heat apparatus
will be added to our lab later in 2001. This
will greatly expand our capability to char-
acterize magnetic properties and is essen-
tial for the research into “nearly” magnetic
properties such as many of the so-called
heavy-Fermion compounds. These have
been found widely in uranium compounds
and surely exist also in transuranium com-
pounds. More sophisticated experiments
involve the use of neutrons, x-rays, or
muons that are produced at large facilities
(e.g., at synchrotron sources). For these,
special encapsulation procedures have
been developed, and the experiments are
allowed to take place in some user facili-
ties, provided that the amounts are small
and all the regulations are followed.
Future Perspectives
Much remains to be done. Theory is now
actively involved, especially in the pres-
sure experiments on resistivity as well as
on the atomic structures. Completely new
perspectives have opened up with the
production of thin layers in the photo-
emission effort, not only for basic research,
but also in the potential for applications
to actinide surface corrosion. New results
with the SQUID magnetometer on small
samples (less than a milligram) show the
complexity of the magnetic interactions in
the transuranium compounds, and we shall
need all the power of complementary tech-
niques, such as Mössbauer (for Np), neu-
trons, photons, and muons, to unravel the
phase diagrams. In addition, with the com-
pletion of the first stage of our “user labo-
ratory,” we will have available a facility
where outside scientists can come and de-
fine their own projects on transuranium
materials.
Magnetic ordering in the actinides takes
many forms, but in high-symmetry crystal
structures, such as those compounds with
the NaCl fcc structure, a common form
of the magnetic structure is the so-called
multi-k configuration, where k is a wave
vector defining a Fourier component of
the magnetic configuration.17 In the upper
part of Figure 6, we show the configura-
tions with the so-called single-k (1k) and
triple-k (3k) symmetry. The first has
tetragonal symmetry, whereas the second
has cubic symmetry. The precise configu-
ration that a compound develops is a sig-
nature of how the 5f electrons are arranged
on the Fermi surface and may be loosely
thought of as a measure of how the 5f elec-
trons mix with the valence electrons in the
compound.18 USb is known to have the
3k symmetry, whereas PuSb has the 1k
configuration.19 The question is, what
happens when we make a compound
(U, Pu)Sb where the actinide ions are dis-
tributed randomly over the one sublattice
of the NaCl structure, and the magnetic
anisotropies of U and Pu therefore com-
pete? We find that the anisotropy of the
uranium dominates the compounds, and
it is not until 75% Pu that unusual effects
are observed. This material attempts to
transform spontaneously from 3k (as in
USb) to 1k (as in PuSb) as the temperature
is lowered (see Figure 6). A magnetic field
rapidly stabilizes the 1k state. To establish
this phase diagram, we have used a com-
bination of SQUID, neutron diffraction,
and resonant x-ray magnetic scattering.
The diffraction experiments, especially the
high resolution of the x-ray experiments,
establish also that the correlation length
within these magnetic orderings is rela-
tively short-range. Indeed, the competing
17. G.H. Lander and P. Burlet, Physica B 215
(1995) p. 7.
18. H. Yamagami, Phys. Rev. B 61 (2000) p. 6246;
K. Knöpfle and L.M. Sandratskii, Phys. Rev. B 63
014411 (2001).
Acknowledgments
A great many people have contributed,
and continue to contribute, to these efforts.
It is appropriate to thank the groups at
ETH, Zurich, and Grenoble, France, for
collaborations over a long period. The
americium (and higher actinide) experi-
ments are done in close collaboration with
Dick Haire at Oak Ridge National Labora-
tory. Ladia Havela (Charles University,
Prague) has made major contributions to
the surface science studies. We thank Mike
Brooks for a continuous dialogue about
theory. Finally, we thank the management
and other groups of the Institute for
Transuranium Elements for supporting
our efforts in a variety of ways.
19. P. Normile (unpublished manuscript).
ꢀ
MRS Web site:
www.mrs.org
www.mrs.org
www.mrs.org
Advertisers in This Issue
References
Page No.
1. J.C. Spirlet and O. Vogt, in Handbook on the
Physics and Chemistry of the Actinides, Vol. 1, ed-
ited by A.J. Freeman and G.H. Lander (Elsevier,
Amsterdam, 1984) p. 79.
Chemat Technology, Inc.
714
High Voltage Engineering Inside front cover
2. V. Ichas, J.C. Griveau, J. Rebizant, and J.C.
Spirlet, Phys. Rev. B 63 045109 (2001).
3. P. Wachter, F. Marabelli, and B. Bucher, Phys.
Rev. B 43 (1991) p. 11136.
4. P.M. Oppeneer, T. Kraft, and M.S.S. Brooks,
Phys. Rev. B 61 (2000) p. 12825.
Huntington Mechanical
Laboratories, Inc.
IUMRS Facets
Outside back cover
660
5. S. Dabos-Seignon, U. Benedict, S. Heathman,
and J.C. Spirlet, J. Less-Common Met. 160 (1990)
p. 35.
National Electrostatics Corp.
Thermionics Vacuum Products
706
697
6. B. Johansson and M.S.S. Brooks, in Hand-
book on the Physics and Chemistry of the Rare
Earths, Vol. 17, edited by K.A. Gschneidner, L.
Eyring, G.H. Lander, and G. Choppin (Elsevier,
Amsterdam, 1993) p. 149.
7. U. Benedict and W.B. Holtzapfel, in Hand-
book on the Physics and Chemistry of the Rare
For free information about the products and services
offered in this issue, check www.mrs.org/publications/
bulletin/advertisers or fill out and fax the inside
back cover to 724-779-4397.
www.mrs.org/publications/bulletin
688
MRS BULLETIN/SEPTEMBER 2001