Photoelectron Spectroscopy of α- and δ-Plutonium · Photoelectron Spectroscopy of α- and...

168 Los Alamos ScienceNumber 26 2000

Photoelectron Spectroscopy of α- and δ-Plutonium

Aloysius J. Arko, John J. Joyce, Luis A. Morales, Jeffrey H. Terry, and Roland K. Schulze

Number 26 2000 Los Alamos Science 169

After decades of speculation about the role of plutonium’s 5f electrons in bonding and hybridization, photoelectron spectroscopy (PES)

is now providing the first detailed results indicatingthat some of the 5f electrons are itinerant and form

a narrow conduction band. This measurement technique probes electronic structure directly

and with great accuracy, defining a new path into the complex world of

actinide behavior.

The electronic structure of the 5f series of metals in the periodic table represents a relatively unexplored area in con-

densed-matter physics. It is also one of the most interesting areasof research because the 5f electrons in all these elements tend toward

highly correlated behavior as opposed to the independent-particle behav-ior prevalent in other systems. The 5f electrons in the actinides behave in

more varied ways than the electrons in any other series of elements—from local-moment magnetism, through enhanced specific heat like heavy fermions, tonarrow-band behavior.

Among the metals in this group, plutonium appears to be the most exotic. It displays six allotropic phases, and its 5f electrons are at the transition point between being localized, as they are in americium, and being delocalized, as theyare in neptunium. The resistivities of both α- and δ-plutonium are large and have a negative temperature coefficient, which is highly unusual for an elemental mater-ial. The delicate balance between 5f electron bonding and localization is thoughtto be responsible for many of plutonium’s unusual properties, including its propen-sity for phase changes, low-symmetry crystal structures, and low melting point bycomparison with other actinides. Because of its many allotropic forms, plutoniumallows us to observe the changes in electronic structure and their relationship tothe striking volume and structural differences of those allotropes.

Plutonium’s low-temperature phase (α-phase) has a simple monoclinic crystalstructure and a high density—about 19 grams per cubic centimeter (g/cm3). Typically, local-density approximation (LDA) band calculations suggest that the f-electron wave functions have direct overlap, yielding narrow but robust 5f bands.As a result, α-plutonium can be described as having transition-metal properties(see the article “Actinide Ground-State Properties” on page 128). By contrast, thehigh-temperature δ-phase is an fcc structure with a density of 16 g/cm3. Direct f-foverlap is considered unlikely for most of the 5f electrons. Small residual hy-bridization with s and d band electrons, as proposed by the Periodic Anderson



Ground-state density of states Measurement of photoelectron spectrum in vacuum

Detector

Sample

Vacuum

EB

Satellite

0

Emptystates

Occupied valence states

Corelevel

Relaxationenergy

Secondary electrons

Primaryelectrons

EF

EK

EK = hυ − EB − Φhυ

hυ

Φ

e−

hυ − Φ

hυ − Φ

EF

Onsetof

signal

Binding energy increasing

Kinetic energy decreasing

Figure 1. Schematic Representation of the Photoemission Process The experimental setup for photoemission is shown in the inset at lower left. Monochromatic ultraviolet or soft x-ray light of energy

hν is focused onto a clean sample surface in high vacuum (typically, 10 –10 torr), and a detector records the number of photoelec-

trons emitted from the surface as a function of their kinetic energy EK. The ground-state electronic structure is shown by the

schematic density of states on the upper left. The filled states are shown in green and labeled by EB, where the binding energy EB

is defined to be positive. The higher the binding energy EB, the farther is the electron below the Fermi level at EB = 0. Empty states

above the Fermi level are not colored. Electrons at EB absorb a photon of energy h ν and are raised in energy by that same amount.

If they are emitted into the vacuum, they lose not only the energy EB needed to raise them to the Fermi level, but also the energy

associated with the work function Φ due to an image charge at the surface. Thus, their kinetic energy in the vacuum is

EK = hν – EB – Φ, as shown on the right. The primary electrons are those at each EK (this is what the spectrometer actually mea-

sures) transmitted into the vacuum without a scattering event. Their number is, to a first approximation, directly related to th e

number of electrons in the states EB. Those electrons that are inelastically scattered before emission through the surface are called

secondary electrons and contribute to a featureless background (indicated in gray). A core-level electron at ground-state energ y EBC

may be emitted with energy EK = hν – EBC – Φ and produces the sharp dashed peak, provided the core hole left by the electron is

well screened by the conduction electrons. Emission of core-level electrons may also lead to one or more satellites at lower EK if

the screening of the core hole is poor.

Model, or perhaps itinerancy fromsome fraction of the 5f electrons mayyield some 5f density of states at theFermi energy (see the articles “A Possi-ble Model for δ-Plutonium” and“Actinide Ground-State Properties” onpages 154 and 128, respectively). Indeed, both resistivity data and recentmagnetic-susceptibility measurementssuggest that δ-plutonium has a narrow(very flat) conduction band pinned tothe Fermi energy. Moreover, low-temperature specific-heat measurementsindicate that the crystalline effectivemass of conduction electrons at theFermi energy is many times the restmass of electrons, a property suggestiveof heavy-fermion materials. Clearly, adetailed understanding of the plutonium5f electronic structure is fundamentallyimportant to the overall understandingof plutonium properties and to con-densed matter physics as a whole.

In this article, we present measure-ments of the electronic structure ofplutonium metal done by photoemissionspectroscopy. Both α- and δ-plutoniumwere examined with several methods ofsurface preparation and in two indepen-dent photoelectron-spectrometersystems. The qualitative agreement between the two measurements is good.Spectra from both phases display a narrow 5f-related feature at the Fermienergy. That feature is narrower in δ- than in α-plutonium—a full width at half maximum (FWHM) of about 70 milli-electron-volts (meV) vs one of about 150 meV. The 70-meVFWHM is also typical of f electronpeaks near the Fermi energy in heavyfermions. In both plutonium allotropes,the photon energy dependence of thisfeature suggests nonnegligible 6d admixture—albeit, somewhat less in δ-plutonium. These results were madepossible by the design, development,and implementation of a unique laser-plasma light source at Los Alamos (by Arko, Joyce, and Morales) and by the first use of a synchrotron sourceto acquire photoemission data on pluto-nium metal (Terry and Schulze).

Basics of Photoelectron Spectroscopy (PES)

PES may be the most direct and detailed tool for measuring the electron-ic structure of metals. Photons ofknown energy (ultraviolet to soft x-ray)incident on a sample surface in vacuumare absorbed by the metal’s electrons.The kinetic energy of these electronsincreases by the amount of the photonenergy. Electrons that are near the sam-ple surface and that, prior to absorption,have a momentum component perpen-dicular to the surface1 may escape fromthe material into the vacuum, where adetector records the number of elec-trons collected as a function of kineticenergy. As shown in the Figure 1schematic, the measured photoelectronspectrum gives, to a first approxima-tion, the shape of the occupied electrondensity of states in the sample (scaledby the orbital cross sections).

PES is especially effective when thephoton source is tunable. We can thenmeasure the photoelectron spectra atdifferent incoming photon energies. Because the cross section for photoab-sorption varies strongly with electronorbital symmetry at vacuum-ultraviolet(VUV) energies (Yeh and Lindau1985), comparison of the differentspectra at different photon energies allows identifying those spectral features that are due to electrons of agiven orbital symmetry (s, p, d, or f).We can therefore compare such contri-butions as those of plutonium’s 5f and6d electrons.

In the very near future, when high-quality crystals with mirror surfacesbecome available, one could utilize precise sample-orientation devices to determine accurately the incoming photon direction, as well as the direc-tion of the exiting electrons with respectto the sample surface. This angle-resolved PES technique, or ARPES,would allow development of a completemap of the energy bands—electron

energies as a function of crystal momentum. Certainly, the latter possibility represents the future for research into the electronic structure of plutonium. Below, however, we limitthe discussion to the simpler angle-integrated PES.

The PES Spectrum. Figure 1 defines the essential parameters of PES.Monochromatic ultraviolet or soft x-raylight of energy hν is focused onto aclean sample surface in high vacuum(typically, 10–10 torr). The electron energy in the material is specified byEB, the binding energy, which is mea-sured relative to the Fermi level, EF.When an electron in the material absorbs a photon (the process may bedescribed by dipole selection rules), itsenergy is raised by an amount hν. Toescape the surface into the vacuum, theelectron must have a momentum com-ponent normal to the sample surface, itmust originate close enough to the sur-face—at the most, a few electron meanfree paths (MFPs)—so that it should notbe easily scattered, and it must gain suf-ficient energy from the photon in thenormal component of momentum toovercome the electric dipole energy ofthe surface (usually referred to as thework function, or Φ). Once this photo-electron is in the vacuum chamber, itsenergy and momentum can be deter-mined by an electron energy analyzer.

Because energy and, to some degree,momentum are conserved in the trans-mission through the surface, there is asimple relationship between the photo-electron’s kinetic energy, EK, in thevacuum and its binding energy, EB, inthe sample: EK = hν – EB – Φ. (Actual-ly, Φ is the work function of theanalyzer because the measured kineticenergy of the electron is relative to thelevel of vacuum in the analyzer.) To measure a typical spectrum, we count the electrons detected at anygiven kinetic energy in the vacuum, and the number detected is directly related to the number (density) of elec-trons within a given energy interval inthe sample.



1Note that photoabsorption does not alter theelectron momentum significantly.

The schematic view in Figure 1,however, greatly simplifies the true situation. In reality, many complicationsprevent the PES spectrum at a certainphoton energy from being a replica ofthe occupied density of states, or theelectronic band structure. Here, we con-sider two major complications: surfacesensitivity and photoabsorption cross-section dependence on photon energy.

Surface Sensitivity. One limitationof PES is that it probes only the firstfew atomic layers of a material’s sur-face. Although the surface and bulkelectronic structures may differ, it hasoften been demonstrated that the bulkelectronic structure of homogeneoussolids is already well established at oneor two monolayers below the surface.The universal curve in Figure 2 showsthat the probing depth (related to theMFP of the outgoing electron) variesstrongly with the kinetic energy of theoutgoing electron, and the deepestprobe (the greatest bulk sensitivity)

occurs at very low and again at veryhigh kinetic energies. This continuousvariation allows us to separate surfacefrom bulk features. We can measure aset of PES spectra at photon energiesspanning a significant portion of theelectron kinetic-energy range and then,by comparing them directly, distinguishsurface from bulk features.

The great sensitivity to the surfacelayer also implies that atomically cleansurfaces prepared in vacuum are an absolute requirement if one is to mea-sure true bulk properties.

Photoelectron Cross Sections. AsYeh and Lindau point out (1985), the probability of an electron absorbinga photon—that is, the photoelectroncross section—varies strongly with theincident photon energy (hν), the elec-tronic orbital symmetry (s, p, d, or f),and the atomic configuration (for exam-ple, Z-number and occupied energylevels). These variations complicate interpretation of the measured PES

spectra and of the material’s electronicstructure. To obtain a complete pictureof the density of states, one needs totake many spectra, using different pho-ton energies, and then take into accountthe effects of the photoabsorption crosssections on the measured intensities of the photoelectron features.

Figure 3 shows photoabsorptioncross sections for a plutonium atom cal-culated with the code of Yeh andLindau (1985). At low photon ener-gies—1 to about 30 electron-volts(eV)—the 6d states have the highestcross section and dominate the photoe-mission spectrum, whereas above abouthν = 40 eV, the 5f states begin to dom-inate. In principle, one can thereforeeasily separate the 5f from the 6d con-tributions to the density of states bycomparing several photoemission spec-tra taken at different energies in therange of 20 to 50 eV. Figure 4 showssuch a set of spectra for the compoundneptunium-arsenide (NpAs). The 6demission predominates in the 21-eVspectrum. In Figure 4, we have overlaidcopies of the 21-eV spectrum on thespectra taken at higher photon energies.Clearly, the 5f contribution grows dramatically with hν and may be deconvolved from the 6d features. Sucha separation is possible for plutoniumand any of its compounds or alloys.

Resonant PES. Another method toidentify specific orbital contributions to the spectrum is resonant PES (Fano1961). In this technique (see Figure 5),one works at a photon energy near the5d absorption edge (approximately 115 eV for plutonium) to promote a 5d core-level electron into an empty 5f state just above the Fermi energy.Because this is primarily an atomicprocess, it is the final-state electronicconfiguration 5d95fn+1 that determinesthe exact photon energy (energies) atwhich resonance occurs (Lynch andWeaver 1987). As this electron decaysto refill the core hole, it gives up its excess energy via an Auger-like processto a 5f (or, to a lesser extent, a 6d)electron near the Fermi energy. That



Atte

nuat

ion

leng

th (

Å)

Photoelectron energy (eV)

20

10

05 50

HeI

LPLS range

ALS range

HeI

I

Ym

ζ

Mg

Kα

Al K

α

500 500010 100 1000

Figure 2. Universal Curve of the Probing Depth of PESThe photon energy range of the laser-plasma light source (LPLS) system is depicted

by the dotted rectangle. The range of Beamline 7.0.1 at the Advanced Light Source

(ALS) is shown by the dashed rectangle. Also shown are the energies of various line

sources commonly used in PES. The attenuation length, or probing depth, is plotted

as a function of the photoelectron energy. The probing depth is proportional to the

MFP of the photoelectron, which has not yet exited the sample. Approximately 95%

of the photoelectrons come from a probe depth within 3 MFPs of the surface.



Figure 3. Calculated Photoabsorption Cross Sections for Plutonium The photoabsorption cross sections for different atomic states in plutonium calculated using the methodology of Yeh and Lindau

(1985) are shown here as representative of the cross sections in plutonium. (a) At low photon energies, the 6d cross section do mi-

nates over the 5f, but above approximately 30 eV, the opposite is true. The cross sections for the 5f 56d1 configurations are shown

because they are more relevant to plutonium metal than the 5f 66d0 configuration of isolated atoms. (b) The cross sections for pluto-

nium valence and core-level electrons down to the 4f state are plotted on a log scale up to 1600 eV.

0.01

0.1

1

10

Part

ial photo

ioniz

ation c

ross s

ection (

Mb)

Part

ial photo

ioniz

ation c

ross s

ection (

Mb)

604002000

P

5f6

6s6p

5d

7s 6s 7s6d

16

14

12

10

8

6

4

2

160140120100806040200

Photon energy (eV)

5f6

6d1 5f5

��

��

��

��

��

Pho

toel

ectr

on in

tens

ity (

arbi

trar

y un

its)

15 20 25 30Kinetic energy (eV)

35 40 45

Energy of incomingphoton

21 eV

E = EF

Np-6d

30 eV

NpAs40 eV

50 eV

Increase in Np-5fintensity

Figure 4. Photoemission Spectra of NpAs for Different Photon EnergiesThis plot shows several measurements of the d- and f-electron photoemission peaks as a function of photoelectron kinetic energy

EK. Because EK increases with the incoming photon energy h ν, the Fermi edges of the different spectra, indicated by the sharp

and sudden onset of emission, are offset from each other by h ν. The spectrum taken with h ν = 21 eV, the lowest photon energy,

has been superimposed on the spectra taken at higher photon energies. In each spectrum, the broad feature at –3 eV relative to

the Fermi edge is due to the 6d electrons. The sharp peak at –0.5 eV relative to the Fermi edge increases in intensity as the p hoton

energy increases. This increase, as well as the change in line shape, is due to increased 5f cross section and subsequent emiss ion

at higher h ν. The increase in 5f intensity is represented by the hatched areas.

0.01

0.1

1

10

Par

tial p

hoto

ioni

zatio

n cr

oss

sect

ion

(Mb)

Par

tial p

hoto

ioni

zatio

n cr

oss

sect

ion

(Mb)

16001400120010008006004002000

Photon energy (eV)

4f

5f6

5d

6s6p6p

5d

7s 6s 7s6d

16

14

12

10

8

6

4

2

160140120100806040200

Photon energy (eV)

5f6

6d1 5f5

electron is ejected into the continuumof states above the vacuum level. (The refilling of the hole may producea photon of commensurate energy instead of an Auger-electron, but theradiative process is much less favored.)

The Auger-like process yields afinal state that is identical to that resulting from normal PES and there-fore interferes either constructively or destructively with the normal PESphotoabsorption. Because the resonantcross section (constructive interfer-ence) is very large (it is proportionalto the cross section for the initial core-level process promoting a 5d electroninto a 5f state), it greatly increases the 5f emission intensity in a narrowrange of photon energies near the absorption edge. By contrast, destruc-tive interference, or antiresonance,may yield almost zero 5f emission atthe so-called antiresonance (approxi-mately 104 eV for plutonium). Spectraof actinide compounds taken at resonance and antiresonance are partic-ularly useful for separating the 5f and, to some extent, the 6d com-ponents from other electronic states.

Experimental Details

Powerful synchrotrons have beendeveloped into intense tunable lightsources for photoemission studies. TheAdvanced Light Source (ALS) at theLawrence Berkeley National Laboratoryrepresents the cutting edge of today’smachines. Although potential hazardsand political sensitivity surroundingplutonium have often precluded study-ing this material at synchrotron sources,the situation is being remedied, and aswe will describe, Schulze and Terryhave performed successful PES mea-surements of plutonium at the ALS.Arko and Joyce, on the other hand, de-signed, constructed, and currentlyoperate a one-of-a-kind tunable laser-plasma light source (LPLS)2, whichmimics the output of a synchrotron. Inaddition, the LPLS system has an insitu capability for automatically produc-

ing clean actinide surfaces, an essentialfeature for work at VUV photon ener-gies. The LPLS and ALS are describedin the short articles “Laser-PlasmaLight Source” and “Actinide Photoe-mission Measurements at the AdvancedLight Source” on pages 186 and 188.

Surface preparation is crucial to aPES experiment because, as shown inFigure 2, the probing depth of PES isonly a few atomic layers (in most met-als, about 1 to 5 atomic layers, or 5 to25 angstroms). Atomically clean sur-faces are thus essential to accurate PESresults. In the synchrotron experiments,surfaces are cleaned by sputter anneal,and in the LPLS experiments, by laser

ablation—the laser beam is redirectedinto the vacuum chamber and cleansthe sample. The latter technique is simple to use and allows cleaning the sample at low temperatures. Lowtemperatures are significant becausethey limit diffusion and reduce degrada-tion reactions at the sample surface.

Figure 6 shows a sequence of PESspectra taken at five stages of in situcleaning. The top spectrum is of the α-plutonium sample with a thick oxideoverlayer, and it shows a broad oxidepeak centered around an electron kineticenergy of 29 eV (EB = –6 eV), whichis due to the electrons in the oxygen 2pband states, and a second peak centeredat 33.5 eV, which is due to the local-ized 5f electrons in the plutoniumoxide. These LPLS data were taken at aphoton energy of 40 eV to enhance the



hυ

EF

Excited5f electron

Excited5f electron

Final state 2Final state 1

Refilling 5f hole

5d core level

5d 5f resonancetransition

5f emptystates

5f band states

Figure 5. Schematic of Resonant PhotoemissionIn resonant photoemission in plutonium, a photon of energy h ν is sufficient to promote

a core 5d electron into an empty state, just above the Fermi level. As this electron de-

cays to refill the core hole, it gives up its energy to a 5f electron in the occupied states

via an Auger process, and the resulting state is final state 1. However, the identical

final state can also be reached by direct photoemission and is depicted as final state 2.

The two final states can interfere with each other constructively (resonance) or

destructively (antiresonance). In plutonium, the resonance and antiresonance are at

120 eV and 104 eV, respectively.

2Since 1992, this project has been funded by theOffice of Basic Energy Sciences, Division of Materials Sciences, of the Department of Energy.

PES intensity of the oxygen 2p elec-trons relative to that of the 5f electrons.We used only a few laser pulses to ablate the sample surface between taking each of the subsequent spectra.Note that, with each cleaning, the Fermi edge grows in intensitywhereas the oxide peak at a binding energy of –6 eV decreases and changesin shape (residual surface oxygen isstill present in cycle 5, but by compari-son with the starting oxide, it is clearlydifferent in shape and reduced in inten-sity). The final metallic spectrum isdramatically different from the startingspectrum, but it is similar to the finalspectra obtained by the more-traditionalsputter-anneal (or scraping) methods,thus demonstrating the effectiveness of laser ablation.

In preparation for the synchrotronexperiments at the ALS, the sampleswere first passed through numeroussputter-anneal cycles in the Chemistryand Metallurgy Research (CMR) Facility at Los Alamos, and they wereexamined by traditional x-ray photoe-mission and Auger surfacespectroscopies. The samples were thentransferred under vacuum into vesselsmaintained at a pressure of 10–8 torr bya portable battery-powered ion pumpsystem. Being isolated from oxygenand air in these vessels, the sampleswere shipped to the Lawrence Berkeley National Laboratory, where they weretransferred under vacuum to a prepara-tion system appended to the x-rayphotoelectron-diffraction analysis cham-ber of the Spectromicroscopy FacilityBeamline 7.0.1 (see the article “Actinide Photoemission Measurementsat the Advanced Light Source” on page 188). In the preparation chamber,the samples underwent final sputter-anneal cleaning cycles before beingtransferred into the chamber for synchrotron PES measurements.

Using these techniques, we have obtained the world’s first resonant-photoemission data on α- andδ-plutonium. The higher resolutions and cleaner surfaces made possible bylaser ablation have yielded dramatic,



28 30Kinetic energy (eV)

32 34 36

Pu oxide

Adsorbed O-2p

hν = 40 eV using the LPLS

Pu metal

(1)

(2)

(3)

(4)

(5)

Laser ablationcleaning cycle

Figure 6. Cleaning of α-Plutonium by Laser AblationThis sequence of spectra from an α-plutonium sample was taken at various stages

of the laser-ablation cleaning cycle. At the start of the cycle shown in (1), the

α-plutonium sample has an oxide surface, and the broad peaks are due to plutonium

oxide molecular states. The features of the oxygen 2p electrons centered at EK =

29 eV are a double peak due to oxygen band states. The large peak centered at

33.5 eV is associated with localized 5f states in plutonium oxide. As the ablation cycle

proceeds, the line shape of the oxygen feature at EK = 29 eV changes because

the oxygen is no longer in a molecular state, and the density of states at the Fermi

edge from the pure plutonium metal grows in intensity. By stage (5), the greatly

reduced oxygen 2p signal is entirely due to adsorbed oxygen, while the localized

5f plutonium oxide state at 33.5 eV has been replaced by the metallic 5f band states.

Only a few laser pulses were applied between each of the spectra shown.

unexpected results, as we show next.Lower-energy data were obtained with the LPLS system, whereas higher-energy data, including core levels, wereobtained at the ALS.

PES Results forα- and δ-Plutonium

ALS Resonance Data at 300 Kelvins.Figures 7(a) and 7(b)show three-dimensional plots of reso-nant photoemission from α- andδ-plutonium, respectively. The α-pluto-nium sample was high-puritypolycrystalline. The δ-plutonium sam-ple was a small single-crystal grain(about 1 millimeter on a side) stabilizedwith nominally 1 weight percent (wt %)gallium and surrounded by a matrix ofsmaller crystallites—see Figure 5(b)in the article “Preparing Single Crystalsof Gallium-Stabilized Plutonium” onpage 226). Both surfaces had beencleaned to a level of oxygen on theorder of 0.5 monolayer or less based ona simple calculation using peak intensi-ties and the photoemission crosssections. The data were taken at a tem-perature of 300 kelvins.

The three-dimensional plots are acompilation of the standard valence-band photoemission spectra (intensityas a function of electron binding energy) for different photon energiesaround the primary resonance energy.Thus, at a fixed photon energy, the pho-toemission spectrum yields a measureof the valence-band density of states(convoluted with orbital cross sections).At a fixed electron binding energy (orfixed initial electronic state), the inten-sity as a function of photon energy is a measure of the photoemission crosssection for that electronic state as a function of the photon energy, whichis modulated by the resonant process.

From a cursory examination of the data, we can identify the salient features of the spectrum. Along thebinding-energy axis, the huge intensitypeak from EF (where EF ≡ EB = 0) to about EB = –4 eV constitutes the



0

–2

–4

–6

–8

–10–1

2

100

110

120

Proto

n en

ergy

(eV)

130

140

150

160

Binding energy (eV)

0

–2

–4

–6

–8

–10–1

2

100

110

120

Proto

n en

ergy

(eV)

130

140

150

160

Binding energy (eV)

(a)

(b)

Figure 7. Valence-Band Resonant Photoemission Spectra for α- and δ-Plutonium These three-dimensional plots of valence-band resonant photoemission spectra for

(a) α-plutonium and (b) δ-plutonium were obtained at the ALS at the Lawrence Berkeley

National Laboratory. The usual photoemission spectra showing intensity as a function

of binding energy are plotted along a third axis, that of photon energy. The photon

energy was scanned across the 5d absorption edge at about 120 eV so that one can

follow intensity as a function of photon energy for a fixed binding energy (also called

a CIS curve for constant initial state) through both the antiresonance and resonance

energies. The spectra in both (a) and (b) show strong prominent resonance features.

α-Pu

δ-Pu

contributions from the valence bands,specifically the 7s, 5f, and 6d states of plutonium. The feature at a bindingenergy of about –6 eV is probablycaused by the residual oxygen at the sample surface. Along the photonenergy axis, the intensity minimum (about 104 eV) and maximum (approxi-mately 120 eV) correspond to the5d→5f antiresonance and primary reso-nance, respectively. These generalfeatures are similar for the α- and δ-plutonium PES measurements.

A more detailed examination revealssubtle differences in resonant behaviorbetween α- and δ-plutonium. These dif-ferences become most evident if oneexamines the intensity as a function of photon energy for a fixed bindingenergy (usually referred to as a constantinitial state, or CIS, curve). To a roughfirst approximation (within the one-electron approximation), CIS spectrarepresent the empty-state density ofstates convoluted with the 5d photo-electron cross section, includingfinal-state effects. Thus, to a limited degree (governed by dipole selectionrules, ∆l = ±1, wherel is the orbital angular momentum), they give a hint ofthe density of states and, particularly,the 5f multiplet structure for the(5d95fn+1) configuration above the Fermi energy.

By examining the CIS spectra nearthe Fermi energy in Figure 7, we seethat, although both the α- and the δ-allotrope show the antiresonance min-imum at virtually the same point inphoton energy (104 eV), the CIS curvesfor α-plutonium show much morestructure than those for δ-plutonium.For example, the α-plutonium datahave a shoulder at a photon energy of 110 eV, rising to a sharp peak at 117 eV followed by well-defined secondary features at photon energiesof 123 eV, 135 eV, and 148 eV. In contrast, the δ-plutonium data showa much less pronounced shoulder at110 eV, rising to a relatively broad,featureless peak at a 120-eV resonanceenergy followed by relatively less-pronounced secondary features.

The interpretation of all the secondaryfeatures (for example, their relation tothe density of states) will require supporting theoretical work, which is ongoing.

The CIS curves show a distinctFano-line shape (see Figure 8), espe-cially near antiresonance (104 eV). The symmetry of the curve and themagnitude of q (basically, the differ-ence between the intensities atresonance and antiresonance) are highlydependent on the relative probabilitiesof resonance vs direct photoemission.The experimental CIS spectrum of δ-plutonium has a q value of 2, whereas that of α-plutonium has a q value of 1. The CIS spectrum of δ-plutonium has also a greater Fano-like symmetry. These features indicatethat the relative probability for the res-onant process is larger in δ- than inα-plutonium. This increase may be ascribed to an increase in electroniclocalization in the 5f states for the fol-lowing reason. The photoabsorption

cross section for the resonant 5d→5ftransition (Da) is not likely to be signif-icantly enhanced for itinerant versuslocalized 5f. However, the cross sectionfor direct 5f photoemission (De) is increased for itinerant electrons relativeto localized electrons. The q value isgiven byq = Da/(πVDe), where D isthe matrix element for the specifiedtransition and V is the relaxation matrixelement. Thus, the Fano symmetry (andthe magnitude of q) of the CIS spec-trum is very sensitive to the degree of5f electron localization. Compared withα-plutonium, δ-plutonium exhibits agreater Fano symmetry with a largervalue of q, consistent with a greater degree of f electron localization.

The valence-band photoemissionspectra at fixed photon energies (whichare reflective of the density of states inthe material to within the limitationsstated in the introduction) are surpris-ingly similar for the two plutoniumallotropes. Again, however, the datashow some subtle differences. Figure 9



16

12

8

4

0

Inte

nsity

(ar

bitr

ary

units

)

–8 –6 –4 –2 0Reduced energy (arbitrary units)

2 4 6 8

q = 4

q = 3

q = 2

q = 1

q = 0.5q = 0

Figure 8. Calculated Fano-Line ShapesThis plot of standard Fano-line shapes as a function of the parameter q (basically,

the intensity at resonance minus that at antiresonance) shows that the resonance

peak gets sharper (the difference between the resonance and antiresonance) as

the electrons involved become more localized. The shape with q = 1 is similar to the

CIS curve for α-plutonium, and the shape with q = 2 is similar to the CIS curve for

δ-plutonium. Thus, the resonance data suggest that the 5f electrons are more localized

in δ- than in α-plutonium.

shows the plutonium valence-bandspectra for the α- and δ-phase near res-onance (hν = 125 eV) and at theCooper minimum3 (hν = 220 eV). At resonance (spectra at 125 eV), the δ-plutonium valence-band photoemissionwidth extends about 200 milli-electron-volts (meV) farther below the Fermilevel than the α-plutonium valence-band photoemission width.

The sharp feature near the Fermilevel appears more intense in α- than in

δ-plutonium. At first glance, this difference seems inconsistent with the general trend of localization in δ-plutonium. The lower intensity of the δ-plutonium feature, however, is atleast partially a consequence of its nar-rowness—its FWHM is on the order of70 meV. At 300 kelvins, the Fermifunction is about 110 meV wide (theresolution is 10 to 90 percent of theFermi function value), and any spectralfeature of approximately this width or less located within kBT of the Fermienergy will be affected by the Fermifunction (truncated and broadened). The narrow feature in δ-plutonium fallsinto this category. However, tempera-ture alone may not account for all the missing intensity relative to the dataat 40 eV (see below). Some of the dif-ferences may be ascribed to differentgallium concentrations in the two speci-

mens or, perhaps, to differences in 6dadmixture.

LPLS Data at 80 Kelvins. To circumvent the issue of thermal broad-ening and aim for a more accurateapproximation of the density of statesin the PES spectrum, we have collecteddata at 80 kelvins and 60 meV in reso-lution by using the sharp discharge lineof a helium arc lamp at a photon energy of 41 eV. At this photon energy,the calculated 6d and 5f photoabsorp-tion cross sections per electron areapproximately equal. The surfaces were prepared by laser ablation. Figures 10(a) and 10(b) show a widescan and an expanded view, respective-ly, of the resulting PES spectra for bothα- and δ-plutonium. Here, the sec-ondary electron background has beensubtracted, and the integrated intensitiesbetween EF and –4 eV are set to beequal. For this normalization, there isanother slight difference in bandwidth.Globally, the two phases of plutoniumappear surprisingly similar except forthe narrow feature near the Fermi level,which, unlike the data at resonance inFigure 9, is now much narrower in δ-than in α-plutonium. All features observed in δ-plutonium are also foundin α-plutonium, but they are much bet-ter defined in δ-plutonium, perhapsreflecting greater localization. Note thatthe residual feature around –6 eV represents submonolayer surface-oxygen contamination, perhaps between0.1 and 0.2 monolayer, and is therefore ignored. For the most part, the 41-eV data are similar to those pre-sented in Figure 9, except for the sharpfeatures near the Fermi level in thelow-temperature data.

The expanded view in Figure 10(b)shows the only clearly understandabledifference between α- and δ-plutoniumin the valence band data. In particular,the sharp feature in gallium-stabilizedδ-plutonium is clearly much narrowerthan that found in α-plutonium and, towithin the resolution of our spectrome-ter (75 meV), it crosses EF. This featureis consistent with the larger specific-



35

30

25

20

15

10

5

0

Inte

nsity

(ar

bitr

ary

units

)

–5 −4 −3 −2 −1 0 1Energy below EF (eV)

Resonance = 220 eV

Resonance = 125 eV

Valence band

× 140δ-Pu

α-Pu

δ-Pu

α-Pu

Figure 9. Spectra for α- and δ-Plutonium Near Resonance and Near theCooper MinimumThese valence-band spectra for α- and δ-plutonium taken near resonance (h ν = 125 eV)

and at the Cooper minimum (h ν = 220 eV) are remarkably similar. At both photon ener-

gies, the width of the emission band is about 200 meV wider in δ-plutonium than in

α-plutonium, when normalized to the peak at a binding energy of ~1 eV. The increased

intensity of δ-plutonium at the high binding-energy edge of the valence band may be a

consequence of greater electron localization.

3The Cooper minimum provides a way for usinga tunable light source when there is a node in the radial wave function of the initial state. Bymatching this node to a final-state wave function,which would give equal contributions to the posi-tive and negative components of the transitionprobability, one obtains a minimum in the ab-sorption cross section. This matching depends onthe energy of the final state and thus on theincident photon energy.

heat γ-value of δ-plutonium and is simi-lar to the feature found in spectra ofuranium compounds identified as heavyfermions.

Heavy-fermion materials are metalsor compounds with an unfilled 4f or 5fshell, whose specific heat and other bulkproperties suggest (Lee et al. 1986) thatthe f electrons at the Fermi energy havea crystal effective mass larger by ordersof magnitude than the rest mass of theelectron and, indeed, much larger thanthe calculated band mass for the material.A long-standing (though simplistic)model to explain this behavior is the so-called single-impurity model (SIM)(Gunnarsson and Schonhammer 1987),which postulates that the f electrons behave as noninteracting magnetic impu-rities at high temperatures. As thetemperature is lowered, a cloud of con-duction electrons aligns antiparallel tothe f spin to form a singlet state, thusquenching the magnetic moment. Withinthe SIM, very weak residual f-electron

hybridization with this cloud produces anarrow resonance at the Fermi energy,sometimes called the Kondo resonance,which should manifest itself as a verynarrow sharp feature in a photoemissionspectrum. However, the applicability ofthe SIM to such periodic, crystalline sys-tems is tenuous at best since the conceptof an f electron “impurity” on every site(as would be required for δ-plutonium)is almost a contradiction in terms. Recent measurements in uranium heavy-fermion compounds suggest thatLDA-derived bands are well reproducedexperimentally, except at the Fermi energy, where the measured bands areflatter, or equivalently, narrower than the calculated bands. A comprehensivemodel that incorporates both lattice peri-odicity and electron correlations is yet to be quantified.

Temperature Dependence of theData. One test for the applicability ofthe SIM to plutonium is a determina-

tion of the temperature dependence ofthe narrow 5f feature. In Figure 11, weshow data taken at 21 eV at both 300and 20 kelvins for α- and δ-plutonium.No change with temperature is evidentin α-plutonium. On the other hand, a decrease in intensity occurs in δ-plutonium at 300 kelvins. Before interpreting this temperature depen-dence as evidence for the applicabilityof SIM, one must take into account the role of the Fermi function, whichtruncates this narrow feature as thetemperature increases. One can approxi-mate the effect of the 300-kelvin Fermifunction by simply convoluting the datataken at 80 kelvins with a Gaussianwhose FWHM is about 100 meV. Thisconvolution broadens the narrow fea-tures at EF but has no effect on featureswider than 100 meV. With this convo-lution, we can fully account for thewhole temperature dependence betweenthe data at 20 and at 300 kelvins. Invoking the SIM or any other correla-



Pho

toel

ectr

on in

tens

ity (

arbi

trar

y un

its)

−8 −6

δ-Pu

α-Pu

hυ = 41 eVT = 80 K

hυ = 41 eVT = 80 K

Surface oxyen

−4Energy below EF (eV) Energy below EF (eV)

−2 −1.0 −0.8 −0.6 −0.4 −0.2 0 0.20

δ-Pu

α-Pu

(a) (b)

Figure 10. Spectra of α- and δ-Plutonium with Helium II Radiation(a) In the wide-scan spectra of α- and δ-plutonium at 80 K using helium II radiation, the feature at –6 eV is due to about 0.1 monolayer of

adsorbed oxygen. The secondary electron background has been removed from the spectra, and the normalization is such that the ar eas

under the spectra, between EF and –4eV, are equal. (b) In the expanded view of spectra near the Fermi energy, the most significant

difference between the spectra is the broader near- EF feature in α-plutonium.



2.0

1.5

1.0

0.5

0−450 −440

Inte

nsity

(ar

bitr

ary

units

)

Binding energy (eV)

−430 −420 −410

Broadmanifoldof states

Pu 4f Core Levelhυ = 850 eV

α-Puδ-Pu

Sharpmetallicstate

Figure 12. Plutonium 4f Core-LevelPhotoemission Spectra for α- andδ-PlutoniumThe 4f core-level spectra suggest that

electrons in δ-plutonium are more local-

ized than electrons in α-plutonium.

These spectra were taken at high resolu-

tion at the ALS, and the photon energy

was 850 eV. In both cases, the spectra

show a sharp peak at a binding energy

of about –421 eV, reflecting the 4f core-

level state in the metal. The core hole

produced in the photoemission process

is well screened from the departing pho-

toelectron. Thus, the photoelectron

reflects its initial core-level energy state.

The broad peak at higher binding

energies (–422 to –424 eV), by contrast,

reflects a manifold of final states pro-

duced when the departing electron is

poorly screened by the surrounding

conduction electrons.

Pho

toel

ectr

on in

tens

ity (

arbi

trar

y un

its)

Pho

toel

ectr

on in

tens

ity (

arbi

trar

y un

its)

15.0 15.5

α-Pu δ-Pu

16.0Kinetic energy (eV)

16.5 17.0 15.0 15.5 16.0Kinetic energy (eV)

16.5 17.0

hυ = 21 eV hυ = 21 eV

20 K

300 K

20 K

20 K(smoothed)

300 K

(a) (b)

Figure 11. Spectra of α- and δ-Plutonium at 20 and 300 K(a) Whereas α-plutonium is temperature independent, the temperature dependence of δ-plutonium (b) can be well accounted for by a

convolution with an approximately 100-meV FWHM Gaussian, which approximates the Fermi function broadening at 300 K.

tion effects to explain that dependencebecomes therefore unnecessary. Onemust use some other model (perhapsthe Periodic Anderson Model or arenormalized band calculation) to explain the correlations shown as thenarrow peak at the Fermi energy.

Evidence for 5f Localization in δ-Plutonium. Satellites in a 4f core-electron spectrum can shed light on thenature of the 5f states. When a coreelectron is ejected from an atom, thefinal state of the system is an atom witha core hole (a positively charged atom)and a photoelectron in the continuum.If the core hole is fully screened fromthe departing electron (perhaps by a refilling of the core hole on the timescale of the photoemission process),there is no electrostatic interaction between the departing electron and thecore hole (no loss of energy), and theelectron carries information from the initial state of the system. If, how-ever, the screening is incomplete (as, for example, in the case of screen-ing by a reconfiguration of the densityof states around the atom), the photo-electron loses energy because of the electrostatic interaction with thepositive core hole and displays charac-teristics that are a convolution of initialand final states.

The 4f core-level spectra in Figure 12 suggest that electrons in δ-plutonium are more localized thanthose in α-plutonium. These spectrawere taken at the ALS at high resolu-tion. The photon energy was 850 eV. In both cases, the spectra show a sharppeak at a binding energy of about –421 eV, reflecting the 4f core-levelstate in the metal. The core hole pro-duced in the photoemission process iswell screened from the departing photo-electron, and thus the photoelectronreflects its initial core-level energystate. The broad peak at higher bindingenergies (–422 to –424 eV), by con-trast, reflects a broad manifold of finalstates produced when the departingelectron is poorly screened by the sur-rounding conduction electrons. Poor

screening may result when the densityof states of the electrons at the Fermienergy is low or when the electron mobility is poor because these electronshave a large effective mass (that is,they exibit correlation behavior). By comparing the α- andδ-plutoniumspectra, we find that δ-plutonium showsa greater percentage of poorly screenedfinal-state events than α-plutonium.This result suggests that δ-plutoniumhas fewer mobile (delocalized) elec-trons available for screening of the corehole, which is consistent with δ-pluto-nium having a narrow f band and its felectrons having a larger effective mass

and hence being more localized. Thecore-level results are consistent with the valence-band PES data.

Ruling out Surface Effects. In thepast, questions were raised about the possibility that α-plutonium maycontain a δ-like surface structure. Indeed, the similarities of the PES spec-tra would seem to suggest such a possibility. However, the more-bulk-sensitive spectra at 220 eV appearsimilar to those at resonance (125 eV)and at 40 eV and would indicate that,in all cases, we are observing bulk fea-tures and that the small differences



Pho

toel

ectr

on in

tens

ity (

arbi

trar

y un

its)

−3.5 −3.0 −2.5 −1.5−2.0 −1.0

Binding energy (eV)

−0.5 0 0.5

α-Pu T ~ 80 K

60 eV

41 eV

30 eV

21 eV

Figure 13. Spectra of α-Plutonium at 80 K and Several Photon Energies The narrow feature near EF is persistent at all photon energies and must thus contain

substantial 6d admixture in its orbital symmetry. At 21 eV, the photoemission should

be dominated by 6d emission (see Figure 4). Moreover, it is most likely bulk-derived

based on the larger escape depth at 21 eV. The purer 5f states are in the region of

–1 eV. That region clearly grows faster with photon energy than the near- EF peak.

between α- and δ-plutonium are bulkderived. Further, this assumption sug-gests that α-plutonium does not exhibita δ-like surface. This conclusion is additionally corroborated by spectra atlower photon energies. At these ener-gies, PES measurements using 21-eVphotons yield even greater bulk sensi-tivity than those using 220-eV photons(see Figure 2). In Figure 13, we showα-plutonium PES spectra taken at 80 kelvins with 21-, 30-, 41-, and 60-eV photons. The sharp feature at EFis, in fact, somewhat more intense rela-

tive to the other features in the 21-eVspectrum than in the spectra at higherenergies, which indicates not only thatthis peak derives from the bulk, butalso that it contains more 6d admixturein its orbital character than the remain-ing features, which may be more 5f in character.

A similar plot for δ-plutonium (seeFigure 14) reveals that we are againdealing with bulk phenomena becausethe 21-eV data are similar to the 41-eVdata. In this case, however, the intensityof the sharp peak at EF relative to other

features increases significantly from the21-eV spectrum to the 41-eV spectrum.This result indicates that the near-EFpeak contains more 5f character in δ-than in α-plutonium. The near invari-ance of the relative intensities of thefeatures with photon energy points to arelatively uniform admixture of 6dcharacter throughout the valence bandregion (compare these data to the NpAsdata illustrated in Figure 4). Irrespec-tive of the amount of 6d admixture, the spectrum is consistent with the verynarrow band behavior observed in bulk properties and with the conven-tional wisdom that the 5f states inα-plutonium are more hybridized andband-like than those in δ-plutonium.

Surface reconstruction is commonfor metals throughout the periodictable. In general, however, it is onlyobservable in high-quality single crys-tals, and even then, one mustdistinguish between a reconstructed sur-face and simply a surface terminationlayer. Although interpretations of previ-ous data suggested that the crystalstructure on the surface of α-materialwould become δ-like, the dependenceof normal emission spectra on photo-electron escape depth measured byvarying the photon energy was notavailable. Our new measurements ofthe α-plutonium spectra as a functionof escape depth yield nearly invariantspectra, strongly suggesting a lack ofreconstruction (from α- to δ-material).

Current understanding would indi-cate that we have indeed measured the bulk electronic structure of α- andδ-plutonium. All data are consistentwith the concept of much narrower itin-erant states in the δ- than in the α-phase of plutonium. Indeed, bulkproperty data for δ-plutonium suggestbehavior like that of heavy fermions.However, very narrow bands are already obtained within a constrainedband calculation (see the article “Actinide Ground-State Properties” onpage 128). An additional electric-fieldcomponent acts on the 5f electronswithin this generalized-gradient approxi-mation (GGA), which is a refinement of



Pho

toel

ectr

on in

tens

ity (

arbi

trar

y un

its)

−2.0 −1.5 −1.0

α-Pu

−0.5Energy below EF (eV)

0 0.5

41 eV

41 eV

21 eV

21 eV

T = 80 K

δ-Pu

Figure 14. Spectra of α- and δ-Plutonium at 21 and 41 eV and at 80 KData at 21 eV (black lines) and 41 eV (red lines) and a temperature of 80 K for α- and

δ-plutonium are overlaid to show relative peak intensities of the near- EF features.

There is a noticeable increase in the 5f vs 6d character in δ-plutonium.

the traditional local-density approxi-mation (LDA), the workhorse ofdensity functional theory. These band-widths in the constrained calculationagree, to some extent, with the experi-mental bandwidths in δ-plutonium.Thus, there may be no need for fur-ther band narrowing such as mightresult from additional correlation effects (see the comparison with cal-culations in the article “ElectronicStructure of α- and δ-Plutonium” onpage 152). Note that, by constrainingsome of the 5f electrons to be local-ized in the above-mentioned GGAcalculation, one has already intro-duced substantial correlation effects.

The Need for Single Crystals.Many of the remaining questions on plutonium and its compounds oralloys will be resolved once data are obtained on high-quality single-crystal materials. For photoemissionexperiments, these crystals shouldhave mirror surfaces and well-definedaxes normal to the surface. A glimpseof this future can be observed in Figure 15, in which we show angle-resolved spectra on a high-qualitysingle crystal of the uranium-antimo-ny compound USb2 grown4 byprecipitation from a flux. In additionto reflecting the high quality of thecrystal, the data in Figure 15 representstate-of-the-art resolution (15 meV)obtained with a state-of-the-art detector (the Scienta), which simulta-neously preserves energy and angleinformation. Each scan of the Scientasimultaneously measures up to 100 electron emission angles (relativeto the sample surface) and, hence, 100 spectra. Using this apparatus at ahigh-intensity undulator beamline (atthe Synchrotron Radiation Center inMadison, Wisconsin) produces the dra-matic spectra of Figure 15. The figureshows only the first 200 meV of the valence bands below the Fermi energy.The data at 40-eV photon energy in (a)

contain contributions mainly from d states, whereas the data at 60-eV pho-ton energy in (b) derive primarily fromthe 5f electrons. The three-dimensionalplot in (b) clearly shows that the 5f intensity in antiferromagnetic USb2 isconfined to a narrow portion of the Bril-louin zone. As the 5f band hybridizes

with the 6d electrons, it disperses awayfrom the Fermi energy, and the quasipar-ticle peaks broaden and diminish to nearzero.The center of the Brillouin zone islocated near the 18˚ angle. One can seethat there is no 5f intensity whatsoeverat angles larger than about 23˚. This constitutes a classic case of f-d



T = 12 K TN = 200 K

(b) hυ = 60 eV

(a) hυ = 40 eV

1214

16

18

20

16

18

20

35.9535.90

35.8535.80

35.75

55.9555.90

55.85

Energy (eV)

Energy (eV)

55.8055.75

12

14

Angle (θ)

Angle (θ)

Figure 15. Angle-Resolved PES Data for Antiferromagnetic USb 2This is essentially E vs k dispersion, with k represented by the angle (1 degree =

0.07 Å–1 for 60-eV data and 0.06 Å –1 for 40-eV data). Only the first 200 meV within the

Fermi energy are shown. The 40-eV data in (a) are primarily sensitive to 6d emission,

whereas the 60-eV data in (b) represent primarily 5f emission. Note that both 6d and

5f bands show dispersion and that the 5f intensity disperses below the Fermi energy.

The center of the Brillouin zone is near 18˚.

4 John Sarrao of Los Alamos grew this single crystal.

hybridization. Similar-quality singlecrystals of plutonium compounds haverecently been grown with flux growthtechniques.5 Data such as those in Figure 15 could therefore be obtainedfor plutonium compounds with a Scienta energy and angle analyzer. It should then be possible to view theband structure of these plutonium com-pounds directly.

Conclusions

The large photon fluxes now avail-able from bright, low-emittanceundulator beamlines have made it pos-sible to perform some photoemissionexperiments on plutonium at synchro-tron sources. This development willenhance our knowledge of the electron-ic structure of this material. We havealso developed the LPLS as a viablesource for conducting photoemissionexperiments within a conventional laboratory environment. With intensitiesrivaling those of second-generation synchrotrons, the LPLS is particularlyuseful for those experiments on hazardous materials that benefit from a highly controlled and containedenvironment. A combination of PES tools will be used to tackle the least studied, yet the most interest-ing, series in condensed matterphysics—the transuranics.

Results on α- and δ-plutonium sug-gest that the electronic structure ischaracterized by an intense, sharp fea-ture of mixed 5f and 6d character nearthe Fermi energy, as well as a manifoldof broad 5f features centered at a bind-ing energy of about –1 eV. The sharpfeature near the Fermi energy is muchbroader in α- than in δ-plutonium, con-sistent with stronger hybridization. Theintense, narrow 5f peak in δ-plutoniumis reminiscent of the sharp features incerium compounds referred to as aKondo resonance, but this feature

appears to be reproduced in a GGAband calculation that constrains some ofthe 5f electrons to be localized. The useof periodic models based on the Ander-son Hamiltonian may be eventuallynecessary for modeling the electronicstructure of δ-plutonium, but the con-strained GGA result for δ-plutoniumprovides a reasonable starting point.The similarity between α- and δ-plutonium spectra raises suspicionabout a surface reconstruction in α-plutonium to a δ-like surface, partic-ularly since the sharp feature inα-plutonium is not reproduced in aband calculation. However, measure-ments with varying escape depths do not support this notion. Future measurements on single crystals willopen a whole new field of study forplutonium materials. �

Further Reading

Arko, A. J., J. J. Joyce, and L. Morales 1999. J. Alloys and Comp. 286: 14.

Arko, A. J., J. J. Joyce, L. E. Cox, L. Morales, J. Sarrao, J. L. Smith, et al. 1998. J. Alloys and Comp. 826: 271.

Blank, H., and R. Linder, eds. 1976. Plutonium 1975 and Other Actinides. 1976. North Holland: American Elsevier.

Fano, U. 1961. Phys. Rev. 124: 1886.

Freeman, A. J., and J. B. Darby, eds. 1974. The Actinides: Electronic Structure and Related Properties. Vols. I and II. New York: Academic Press.

Gunnarsson, O., and K. Schonhammer. 1987. In Handbook of the Physics and Chemistry of Rare Earths, Vol. 10, p. 103. Edited by K. A. Gschneidner, Jr., L. Eyring, and S. Huffner. Amsterdam: Elsevier Publishers.

Lee, P. A., T. M. Rice, J. W. Serene, L. J. Sham, J. W. Wilkins. 1986. Commun. Cond. MatterPhys.12: 99.

Lynch, D. W., and J. H. Weaver. 1987. In Handbook of the Physics and Chemistry of Rare Earths, Vol. 10. Edited by K. A. Gschneidner, Jr., L. Eyring, and S. Huffner. Amsterdam: Elsevier Publishers.

Miner, W. N, ed. 1970. Plutonium 1970 and Other Actinides. Nuclear Metallurgy 17, Vols. I and II. New York: Metallurgical Society and American Mining, Metallurgical, and Petroleum Engineers.

Shirley, D. A. 1978. In Topics in AppliedPhysics: Photoemission in Solids, Vol. 26, p. 192. Edited by M. Cardona and L. Ley. New York: Springer-Verlag.

Yeh, J. J., and I. Lindau 1985. At. Data Nucl. Data Tables32: 1.



5 In April 2000, John Sarrao of Los Alamos hasgrown single crystals of the compounds PuIn3,PuGa3, and PuSn3.



Aloysius J. Arko received his Ph.D. in condensed matter physics from NorthwesternUniversity. He is currently a staff member at the

Los Alamos NationalLaboratory. His primaryresearch interest hasbeen photoelectron spec-troscopy in narrow-bandmaterials. Arko is a Fellow of the AmericanPhysical Society and ofLos Alamos NationalLaboratory. In recogni-tion of his contributions

to the Laboratory, Arko received the Los AlamosFellows Prize and the Distinguished PerformanceAward. His recent interests are increasingly directed toward the study of plutonium and itscompounds with the unique photoemission capabilities of the laser-plasma light source.

Jeffrey H. Terry received his B.S. in chemistryfrom The University of Chicago and his Ph.D. inchemical physics from Stanford University. Afterjoining the Laboratory as a postdoctoral fellow,Jeff focussed on the application of vacuum ultra-violet synchrotrons to the study of plutoniummetal and compounds. Currently a staff memberat Los Alamos, Jeff hasbeen working on expandingactinide research at thethird-generation synchro-tron radiation facilities inthe United States. His research interests includethe surface chemical behavior of actinide solids,mineral surface interactionswith molecular actinides inenvironmental situations, and the electronic struc-ture of actinide materials.

Roland K. Schulzereceived his B.S. in chemistry from the University of Minnesota, Duluth, and his Ph.D. in chemistry from the University of Minnesota, Minneapolis. Rolandjoined the Los Alamos Laboratory as a postdoc-toral fellow in the Analytical Chemistry Group,where he conducted surface science research anddeveloped and built a time-of-flight secondaryion mass spectrometer for surface and interfacestudies. Since becoming a technical staff mem-ber, Roland has been with the MaterialsCharacterization Team in the Chemical and Metallurgical Research(CMR) Facility, first inthe Materials Scienceand currently in the Nu-clear Materials Division.In the past four years, hehas built up an analyticalsurface-science facilityin the CMR to study radiological materialsand perform fundamentalmeasurements on actinide solids. Roland’s research interests include the surface chemicalbehavior of actinide solids, mineral surface inter-actions with molecular actinides in environmentalsituations, and the electronic structure of actinidematerials.

Luis A. Morales received his undergraduate degree in chemistry from Arkansas State Univer-sity and his Ph.D. from the University of Kansas,where he was in theHigh-Temperature/Materials Science group.After a postdoctoralstint at Los Alamos National Laboratory,Luis became a staffmember in the Nuclear Materials andTechnology group. Hisresearch interests coverareas of solid state and solid-gas phase chem-istry. In order to understand structure-propertyrelationships, Luis applies aspects of thermody-namics, kinetics, and crystallography toactinide-containing chemical systems.

John J. Joycereceived a B.A. in Physics Summa Cum Laudefrom St. Mary’s Universityand a Ph.D. in Materials Science from the University of Wisconsin.He has conducted photoemission research,the results of which havebeen published in over80 articles and presentedin 50 invited and con-tributed talks. Johnjoined the Laboratory in 1990. His current research interests includeelectronic properties ofrare earths and actinides with photoemission research conducted at the Synchrotron RadiationCenter and the Laser Plasma Light Source(LPLS). He is a member of the Advisory Com-mittee for the Synchrotron Radiation Center andone of the principal investigators on the LPLSproject. For his work on the LPLS, John receivedthe Los Alamos Distinguished PerformanceAward and the Los Alamos Achievement Award.

Photoelectron Spectroscopy of α- and δ-Plutonium · Photoelectron Spectroscopy of α- and...

Documents

Transcript of Photoelectron Spectroscopy of α- and δ-Plutonium · Photoelectron Spectroscopy of α- and...