Radiation effects in crystalline ceramics for the immobilization of high-level nuclear waste and...

1434

HelpCommentsWelcomeJournal of

MATERIALS RESEARCH

Radiation effects in crystalline ceramics for the immobilizationof high-level nuclear waste and plutoniumW. J. WeberPacific Northwest National Laboratory, P.O. Box 999, M.S. K2-44, Richland, Washington, 99352

R. C. EwingNuclear Engineering and Radiological Sciences, The University of Michigan,Ann Arbor, Michigan 48109

C. R. A. CatlowThe Royal Institution, 21 Albemarle Street, London W1X 4BS, United Kingdom

T. Diaz de la RubiaLawrence Livermore National Laboratory, P.O. Box 808, Livermore, California 94550

L. W. HobbsDepartment of Materials Science and Engineering, Massachusetts Institute of Technology,Cambridge, Massachusetts 02139

C. KinoshitaDepartment of Nuclear Engineering, Kyushu University, Fukoka 812, Japan

Hj. MatzkeEuropean Commission, Joint Research Center, Institute for Transuranium Elements, Postfach 2340,76125 Karlsruhe, Germany

A. T. MottaDepartment of Nuclear Engineering, Pennsylvania State University, University Park, Pennsylvania 16802

M. NastasiLos Alamos National Laboratory, P.O. Box 1663, Los Alamos, New Mexico 87545

E. K. H. SaljeDepartment of Earth Sciences, University of Cambridge, Cambridge CB2 3EQ, United Kingdom

E. R. VanceMaterials Division, ANSTO, Menai NSW 2234, Australia

S. J. ZinkleOak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, Tennessee 37831

(Received 25 November 1997; accepted 17 February 1998)

This review provides a comprehensive evaluation of the state-of-knowledge of radiationeffects in crystalline ceramics that may be used for the immobilization of high-levelnuclear waste and plutonium. The current understanding of radiation damage processes,defect generation, microstructure development, theoretical methods, and experimentalmethods are reviewed. Fundamental scientific and technological issues that offeropportunities for research are identified. The most important issue is the need for anunderstanding of the radiation-induced structural changes at the atomic, microscopic, andmacroscopic levels, and the effect of these changes on the release rates of radionuclidesduring corrosion.

J. Mater. Res., Vol. 13, No. 6, Jun 1998 1998 Materials Research Society

http://www.mrs.org/publications/jmr/comments.html

W. J. Weber et al.: Radiation effects in crystalline ceramics for the immobilization of high-level nuclear waste

V

A

R

I

tcai r11aia ur to-nTr sRl

g oft rvt edbna ,FR arf -e f

TABLE OF CONTENTS

I. INTRODUCTION 1435

II. RADIONUCLIDE IMMOBILIZATION INCRYSTALLINE PHASES 1437

III. PRINCIPLES OF RADIATION EFFECTS 1438A. Radiation sources 1438B. Interaction of radiation with solids 1438

1. Ionization and electronic excitation 14392. Ballistic processes 14393. Transmutations and gas production 1440

C. Irradiation techniques 14411. Actinide-incorporation 14412. Actinides in natural minerals 14413. Charged-particle irradiation 14414. Gamma irradiation 14415. Neutron irradiation 1442

D. Damage production 1442E. Radiation damage kinetics 1444

IV. RADIATION EFFECTS IN MULTIPHASECERAMIC WASTE FORMS 1444A. Synroc 1445B. Synroc-related ceramics 1445C. Supercalcine 1446D. Glass-ceramics 1446E. Scientific issues for multiphase waste 1446

forms

V. RADIATION EFFECTS IN CRYSTALLINEPHASES 1447A. Phases and structures 1447

1. Actinide-bearing phases 14472. Cs and Sr-bearing phases 1449

B. Defects and defect interactions 14501. Role in damage accumulation 14512. Enhanced diffusion 14513. Defect kinetics 1452

C. Volume changes 14521. Unit-cell volume 14532. Macroscopic swelling 14533. Temperature dependence 1454

D. Stored energy 1455E. Amorphization 1456

1. Structure of the amorphous state 14562. Susceptibility to amorphization 14573. Amorphization processes 14584. Amorphization kinetics 14605. Recrystallization 1463

F. Helium accumulation, trapping, and 1464bubble formation

G. Mechanical properties 1464H. Thermal properties 1465I. Chemical durability 1465

1. Summary of available data 14652. Mechanisms and needs 1466

VI. APPLICATIONS OF COMPUTERSIMULATION METHODS 1466A. Structure and defects 1467

1. Modeling techniques 14672. Modeling achievements 14683. Challenges for ceramics modeling 14684. Modeling opportunities and priorities 1468

J. Mater. Res., Vol. 13,

B. Primary damage production 14691. Threshold displacement energy 14692. Displacement cascade simulations 14703. Damage evolution and amorphization 1470

C. Modeling amorphous structures 1471

VII. IRRADIATION FACILITIES 1472A. Actinide research facilities 1472B. Gamma irradiation facilities 1472C. Ion-beam irradiation facilities 1473D. Electron-beam irradiation facilities 1473

III. CHARACTERIZATION TECHNIQUESAND FACILITIES 1473A. Electron microscopy 1473B. Diffraction 1474C. X-ray absorption spectroscopies 1475D. Phonon spectroscopies 1475E. Hyperfine techniques 1475F. Liquid-phase chromatography 1475

IX. RESEARCH NEEDS ANDOPPORTUNITIES 1475A. Research 1476B. Facilities 1477C. Attract talented scientists 1478

CKNOWLEDGMENTS 1478

EFERENCES 1478

. INTRODUCTION

During the next decades, one of the largest and mosostly challenges facing the world is the stabilizationnd immobilization of nuclear high-level waste (HLW)

n a solid form. For example, at the present time, ove00 million gallons of HLW with a total activity of.2 3 109 curies are stored in 243 underground tankst U.S. Department of Energy (DOE) sites in Wash-

ngton, South Carolina, Idaho, and New York.1–4 Inddition, there are special waste streams, such as Pesidues/scraps (tens of tons) and excess weapons pluium (50 metric tons), throughout the DOE complex.1,5–7

he immobilization and disposition of fissile materialsequires special considerations. Other countries, such aussia, face similar problems of equal magnitude, a

egacy of their defense programs.8,9

In the United States of America (USA), the HLWenerated by the defense program and large fractions

he associated low-level wastes are currently destined foitrification (i.e., immobilization in glass). The charac-eristics of these defense wastes have been summarizy Croff.4 In the USA, commercially generated spentuclear fuel is currently intended for direct disposal ingeologic repository without reprocessing. In contrast

rance, England, Belgium, Japan, China, India, andussia currently reprocess the commercial spent nucle

uel and reclaim the fissile material. Reprocessing genrates substantial volumes of high-level waste, most o

No. 6, Jun 1998 1435


y

-

which is presently destined for vitrification in borosili-cate glass, the generally accepted first-generation wasform.10,11 Consequently, two waste forms are presentlyenvisioned for most nuclear waste:spent nuclear fueland borosilicate glass. However, neither of these wasteforms was designed nor selected on the basis of thehigh chemical and physical durability. Rather, thesewaste forms are single components of a multibarriesystem that relies mainly on geologic isolation to prevenradionuclides from reaching the biosphere. As an example, the UO2 in spent nuclear fuel is not stable underoxidizing conditions.12,13 Likewise, borosilicate glassesare metastable and will inevitably corrode when incontact with water or humid air.14 Both spent nuclear fueland current borosilicate HLW glasses may be adequawaste forms within the multibarrier context of geologicisolation, but second-generation highly durable wastforms may become necessary to immobilize the presentdiverse and complex nuclear waste streams existinworldwide.

Due to the emergence of increasingly diverse sourceof nuclear wastes (e.g., separated fission products, pltonium residues/scraps, weapons-usable plutonium aenriched uranium, other high-actinide wastes, and higburn-up commercial nuclear wastes), there is reneweinterest in the development of alternatives to borosilicatglass waste forms in France,15,16 Russia,17 Japan,18 andthe USA.19,20 In general, many of the alternative wastematrices under consideration are complex but highlstable crystalline ceramics or glass-ceramics, in whicthe radionuclides are incorporated into the structureof crystalline phases. For example, waste forms havnot yet been selected for the majority (by volume) ofthe radioactive wastes within the DOE complex, andalternatives to borosilicate glasses are being considerfor many of these nuclear wastes. The recent Recoof Decision21 and Strategic Plan22 for the storage anddisposition of weapons-usable fissile materials clearlidentified ceramics as a candidate form for plutoniumdisposition, and a recent assessment23 recommendedceramics as the preferred immobilization technologyThere are also other “special” waste streams within thDOE complex, such as the 15 metric tons of CsCl anSrF2 stored in capsules at the Hanford site,4 where stabi-lization in alternative matrices is under consideration. Ihas also become increasingly apparent that vitrificatioof the large volumes of chemically complex wastes aHanford (over 200,000 m3 of HLW and sludge) mayrequire pretreatment. Any pretreatment of these complenuclear wastes may result in additional and chemicallunique waste streams that may be difficult to vitrify (e.g.due to the high phosphorus or zirconium contents) omay be more suitably stabilized in crystalline ceramics(e.g., the stabilization of Cs in silicotitanates). For somewaste streams, glass-ceramics may be used, where

1436 J. Mater. Res., Vol. 1

te

ir

rt-

te

elyg

su-ndhd

e

yhse

edrd

y

.ed

tnt

xy,r

a

substantial proportion of the radionuclides, particularlythe actinides, are incorporated in crystalline phases.24–26

This renewed interest11 in the design and use ofalternative waste forms highlights several specific andhighly desirable advantages to the use of more durablewaste forms, such as crystalline ceramics, as a primarybarrier to radionuclide release. These include:

(i) Initially, the radionuclides are isolated within thewaste-form matrix, and the only part of the repositorythat is radioactive is the waste form. The successfulperformance of the waste form results in near-fieldcontainment. This is much preferred to geologic iso-lation, which essentially relies on long travel times,dilution and dispersion, and sorption on mineral surfaces.These geologic processes implicitly imply the releaseand movement of radionuclides.

(ii) The chemistry and physics of the corrosion andalteration of a durable waste form, with the subsequentlimited release or retention of radionuclides over somerange of conditions, are inherently more simply modeledand extrapolated over time than the use of coupledhydrologic, geochemical, and geophysical models of themovement of radionuclides through the far-field of ageologic repository. That is, the extrapolation of thecorrosion behavior of material over long periods restson a firmer scientific foundation than the extrapolatedbehavior of, as an example, hydrologic systems that aresite specific and highly dependent on assumed boundarconditions (e.g., climate and recharge).

(iii) Naturally occurring phases, mineral “ana-logues,” provide fundamental data for the “confirmation”of extrapolated or interpolated behavior of the wasteforms over long periods of time. This approach holdsgreat potential for the confirmation of performanceassessments related to radiation effects27 and corrosionmechanisms28 in the near-field.

All of these features are characteristic of durablecrystalline ceramics that are potential candidates as nuclear waste forms. Fortunately, there is already a longhistory of research and development on crystalline ce-ramics as nuclear waste forms.11,29 Further research anddevelopment has been initiated in both the USA20 andRussia30 in efforts to develop immobilization matricesfor the disposition of excess weapons plutonium and inEurope31–35 to develop nonfertile fuels (to burn up excessactinides) that are also suitable waste-form phases. Inall of these applications,one of the critical concernshas been to evaluate the effects of radiation on thesecrystalline phases,because the effects of radiation canimpact the performance of waste forms containing HLWand Pu. The most prominent effect in crystalline wastephases is the radiation-induced transformation from acrystalline to an amorphous state, mainly as a result ofthe self-radiation damage associated with thea-decayof incorporated actinides.36,37 This transformation results

3, No. 6, Jun 1998


Ee

elaiir

rtr

i)

ndea

hu

n

oetelasnicn

fiie,i

itnn

-,tt

,

e

ss-

s

m-r,d

,-

,rl

d

f

o

in significant changes in the structure and propertiesthe waste form, as discussed in detail below (Sec. V.Although there is already a substantial body of rsearch available on radiation effects in crystalline phas(simple oxides, semiconductors, intermetallics, and mals), waste-form phases are, in general, more compin their structures and compositions, because thesephases that must be able to incorporate a rather wvariety of chemically diverse radionuclides into thestructures at the atomic scale.

These “complex” ceramics include materials that agenerally strongly bonded (mixed ionic and covalenrefractory, and frequently good insulators. They adistinguished from simple, compact ceramics (e.g., Mgand UO2) by atomic-scale features that include: (open network structures that are best characterizeda consideration of the shape, size, and connectivitythe coordination polyhedra; (ii) generally complex compositions that characteristically lead to multiple catiosites and lower symmetry; (iii) directional bonding; an(iv) bond-type variations, from bond-to-bond, within thstructures. Additionally, some of these structures mcontain structural (OH) and molecular (H2O) water. Theresponse of these materials to radiation is compleas discussed in detail below (Sec. V), and is strongdependent not only on the total dose but also on trate of damage (radionuclide decay) and the temperatof the waste form, both of which decrease over lontime periods.

In order to assess the current state of understaing, identify relevant scientific issues, and determindirections for future research in the area of radiatioeffects in ceramic phases relevant to the immobilizatiof high-level waste and plutonium, a panel was convenon January 13–17, 1997, under the auspices ofDepartment of Energy, Council on Materials SciencThe primary objective of this twelve-member panewhich was chaired by W. J. Weber and R. C. Ewing, wto identify the fundamental scientific issues that mube addressed in order to (i) advance the understaing of radiation effects in relevant crystalline ceramoxides, (ii) perform accelerated irradiation studies acomputer simulations to doses corresponding to 104 to106 years of storage, and (iii) provide the sound scientibase needed in order to develop predictive strategand models for the performance of crystalline phasused for the immobilization and disposition of HLWPu residue/scraps, weapons-useable Pu, and other hactinide or high-fission-product waste streams. Indeliberations, this panel drew heavily on the recereport of a similar DOE panel on radiation effects iglass nuclear waste forms38 and on previous workshopreports39,40 and reviews11,37,41–44related to this topic. Thisreview summarizes the deliberations and conclusionsthe panel.

J. Mater. Res., Vol. 1

of).-est-exre

de

e),eO

byof-

y

x,lyere

g

d-ennd

he.

,std-

d

cess

gh-st

of

II. RADIONUCLIDE IMMOBILIZATIONIN CRYSTALLINE PHASES

In contrast to glass waste forms in which the radio-nuclides are, in principle, homogeneously distributedthroughout the waste solid, ceramic or vitreous-ceramicwaste forms may incorporate radionuclides in two ways:

(i) Radionuclides may occupy specific atomic posi-tions in the periodic structures of constituent crystallinephases, that is, as a dilute solid solution. The coordination polyhedra in each phase impose specific sizecharge, and bonding constraints on the radionuclides thacan be incorporated into the structure. This means thaideal waste-form phases usually have relatively complexstructure types with a number of different coordina-tion polyhedra of various sizes and shapes and withmultiple substitutional schemes to allow for charge bal-ance with radionuclide substitutions. Extensive nuclidesubstitution can result in cation and anion vacanciesinterstitial defects, and changes in structure type. Theformation of polytypes and twinning on a fine scaleis common. The point defects can themselves becomsites for the radionuclides. Generally, the complexityof the high-level waste composition usually results inthe formation of a polyphase assemblage (e.g., vitreouceramics contain multiple ceramic phases in a glasmatrix, and Synroc consists of phases such as zirconolite, CaZrTi2O7; perovskite, CaTiO3; and “hollandite,”BaAl2Ti6O16), with unequal partitioning of radionuclidesbetween the phases. In vitreous ceramics, the actinidepreferentially partition to rare-earth or zirconium-basedphases; while in Synroc, the actinides partition prefer-entially into the zirconolite and perovskite. In general,the polyphase assemblages are sensitive to waste-streacompositions and waste loadings, which affect the variations and abundance of the constituent phases. Howeveif certain elements are present that are not incorporateinto existing phases, minor phases will form, includ-ing glass segregated along grain boundaries. Ideallyall waste stream elements, both radioactive and nonradioactive, are important components, or at least insolid solution, in the phases formed. In some casesa single phase (e.g., zirconolite, monazite, apatite, osodium zirconium phosphate) can incorporate nearly alof the radionuclides into a single structure, especially ifthe radionuclides have been partitioned into chemicallysimilar groups, such as actinides.

(ii) Radioactive phases, perhaps resulting fromsimply drying the waste sludges, can be encapsulatewithin nonradioactive phases. The most commonapproach has been to encapsulate individual grains oradioactive phases in a matrix of TiO2 or Al2O3, mainlybecause of their extremely low solubilities.45 This ap-proach requires major modifications to the waste-streamcomposition and special processing considerations t

3, No. 6, Jun 1998 1437


i

n

era

e

c

d

o

e

s

itueters.

meorte.ldan

gs

alrterme2e.

ls

eofs)

ly)toisn.

sas

OEof

keep temperatures low enough to avoid volatilization oradionuclides and any reaction of the radioactive phaswith the matrix phase. A similar approach may be takewith low-temperature assemblages (e.g., mixing witconcrete), but again, there is the possibility of reactiobetween the encapsulating phase and the radioactphases.

Both of the above types of waste forms are specifically formulated for the incorporation or encapsulatioof radionuclides.

Spent fuel (a metal-clad, UO2 ceramic) results fromthe use of reactor fuel that has traditionally been dsigned without consideration of its waste form propeties, but in order to avoid reprocessing, spent fuel hbecome an important waste form. The properties of spefuel as a waste form are determined primarily by thirradiation and thermal history of the UO2 in the reactorand the disposal conditions (e.g., nearly insoluble undreducing conditions and easily corroded under oxidizinconditions). Radionuclides are distributed throughout thfuel matrix in solution, as exsolved/precipitated phasealong grain boundaries, or in voids, cracks, and the fuecladding gap. There are considerable data on the effeof neutron irradiation,46–49a-particle irradiation,50–52andion-beam irradiation53,54 on UO2, as well as radiationdamage in naturally occurring UO21x .55 Radiation effectsin UO2 as a nuclear waste form are not reviewehere; however, radiation damage froma-decay events atrepository temperatures, as opposed to reactor operattemperatures, could further affect the microstructurethe spent fuel.

III. PRINCIPLES OF RADIATION EFFECTS

Many of the principles of radiation effects in wasteforms for HLW and Pu disposition have been discussein great detail previously.37–44 Key principles that arespecific to crystalline ceramic phases are highlightehere, and some specific scientific issues are identifiand discussed.

A. Radiation sources

The principal sources of radiation in HLW areb-decay of the fission products (e.g.,137Cs and90Sr) anda-decay of the actinide elements (e.g., U, Np, Pu, Amand Cm). Beta-decay produces energeticb-particles,very low energy recoil nuclei, andg-rays, whereas,a-decay produces energetica-particles (4.5 to5.5 MeV), energetic recoil nuclei (70 to 100 keV),and someg-rays. There are also minor contributionsto the radiation field from spontaneous fission of somof the actinides and from (a, n) reactions. The rates ofspontaneous fission and (a, n) reactions are very lowand do not significantly contribute to the overall effectof radiation. In general,b-decay is the primary source


fesnhnve

-

--snt

erges,l-ts

ingf

d

dd

,

e

of radiation during the first 500 years of storage, asoriginates from the shorter-lived fission products. Dto the long half-lives of the actinides and their daughproducts,a-decay is generally dominant at longer timeFigure 1 shows the cumulativeb-decay anda-decaydose (events per gram) as a function of storage tithat might be expected for multiphase ceramicsglass-ceramics containing DOE high-level tank wasCeramics containing non-USA commercial HLW wouexperience doses of at least a factor of 10 higher thfor the DOE tank wastes.38 Also shown in Fig. 1 isthe cumulativea-decay dose for a ceramic containin10 wt. %239Pu, which reaches significantly larger valuerelative to DOE tank wastes or non-USA commerciHLW. The cumulativea-decay dose in ceramics fo239Pu immobilization reaches a temporary plateau af100,000 years, but increases over much longer tiperiods due to the235U decay series, as shown in Fig.for several239Pu concentrations in a ceramic host phas

B. Interaction of radiation with solids

Beta and alpha decay affect crystalline materiathrough the interactions of theb-particles,a-particles,recoil nuclei, andg-rays with the ceramic phases. Thesinteractions fall into two broad categories: the transferenergy to electrons (ionization and electronic excitationand the transfer of energy to atomic nuclei, primariby ballistic processes involving elastic (billiard-ball-likecollisions. The partitioning of the energy transferred inelectronic excitations and into elastic nuclear collisionsan important process controlling the effects of radiatioFor b-particles andg-radiation, the energy transfer idominated by ionization processes. For ions, such

FIG. 1. Cumulative number ofb-decay anda-decay events pergram for a multiphase ceramics or glass-ceramics containing Dhigh-level tank waste. Also shown is the cumulative numbera-decay events per gram for a ceramic containing 10 wt. %239Pu.

3, No. 6, Jun 1998


,

n

t

,

FIG. 2. Cumulativea-decay dose in ceramics as a function of239Puloading and waste storage time.

a-particles and recoil nuclei, interactions involve ioniztion processes and elastic collisions. Ana-particle willpredominantly deposit its energy by ionization processwhile an a-recoil ion will lose most of its energyin elastic collisions with the nuclei of atoms in thsolid. In addition to the transfer of energy, the particlemitted through radioactive decay can themselvessome cases, have a significant chemical effect onnuclear waste material as a result of their deposition aaccommodation in the structure.

1. Ionization and electronic excitation

The bulk average energy absorbed by ionization aelectronic excitations in ceramic waste forms for DOtank wastes and for Pu disposition is similar to that fglass waste forms,37,38 and the relative absorbed dosefor different types of nuclear wastes are illustratedFig. 3. For non-USA commercial HLW applications, thabsorbed doses will be at least a factor of 20 higher ththose indicated for DOE tank wastes (Fig. 3). In mulphase waste forms, partitioning of radionuclides indifferent phases will affect the local absorbed energy. Eergy loss byb-particles anda-particles is predominantlythrough Coulombic interactions. The interactiong-rays with matter is primarily through the photoelectreffect, Compton scattering, and electron-positron pproduction. The high rate of energy absorption throuionization and electronic excitation fromb-decay inHLW ceramics, and to a lesser extent froma-decayin Pu ceramics, can result in self-heating. In additito self-heating, ionization and electronic excitatioproduce a large number of electron-hole pairs thcan result in covalent and ionic bond rupture, chargdefects, enhanced self-ion and defect diffusion, localiz

J. Mater. Res., Vol.

a-

es,

ees

inthend

ndEorsinean

ti-ton-

oficairgh

onsateded

FIG. 3. Cumulative absorbed dose in ceramics containing DOEhigh-level tank waste or containing 10 wt. %239Pu.

electronic excitations, and, in some ceramics, permanendefects from radiolysis.

2. Ballistic processes

Ballistic processes cause direct atomic displace-ments through elastic scattering collisions and are re-sponsible for the atomic-scale rearrangement of thestructure. Beta particles, because of their low mass,do not efficiently transfer kinetic energy to atomic nu-clei and only induce well-separated single displacementevents at high energies (. several hundred keV). Therecoil nuclei (b-recoils) produced inb-decay eventsare almost never energetic enough to be permanentlydisplaced. Similarly, the emission ofg-rays results inthe recoil of nuclei (g-recoils), which also are gener-ally not sufficiently energetic to be permanently dis-placed. Likewise, the electrons emitted by the interactionof g-rays with the structure are generally not suffi-ciently energetic to produce displaced atoms. In generalb-decay of the fission products in HLW produces on theorder of 0.1 atomic displacements perb-decay event.41

The primary sources of atomic displacements by ballisticprocesses in ceramics for HLW and Pu disposition arethe a-particle and a-recoil nucleus produced in ana-decay event. Thea-particles emitted by actinide de-cay in HLW ceramics have energies of 4.5 to 5.8 MeV,and the recoil nuclei (a-recoils) have energies of 70 to100 keV. In the case of Pu ceramics, a 5.2 MeVa-particle and an 86 keV235U recoil are released by thedecay of the239Pu (a-particles anda-recoils of similarenergies are released during the decay of the235U series).

The a-particles dissipate most of their energy byionization processes over a range of 16 to 22mm, butundergo enough elastic collisions along their path to

13, No. 6, Jun 1998 1439


nt

sa

-

c

0t

tne

eay

nn

e

e

b

p

t

icf

nd

-res

ns

ius

fu

ing-

osteb-W,suldhic

n--

er

produce several hundred isolated atomic displacemeThe largest number of displacements occurs nearend of the a-particle range. Both the partitioning ofthe energy loss bya-particles between ionization andelastic collisions and the distribution of displaced atomif the threshold displacement energies are known, cbe determined by computational approaches basedthe binary collision approximation (BCA). These computational approaches generally involve using the BCcomputer codes,TRIM56 and MARLOWE,57,58 which arealso used for charged-particle irradiations.

The more massive but lower energya-recoil particleaccounts for most of the total number of displacemenproduced by ballistic processes in HLW and Pu ceramiThe a-recoil loses nearly all of its energy in elasticcollisions over a very short range (30 to 40 nm), producing a highly localized displacement cascade of 10to 2000 displacements. The density of energy deposiinto the crystal structure by ana-recoil cascade is veryhigh (up to 1 eVyatom) and occurs over a very shortime (,10212 s). The distribution of displacements in aa-recoil cascade can also be calculated by BCA codbut the binary collision approximation is generally noas quantitative for such cascades, as discussed be(Sec. III. D), and does not account for any simultanous recombination events that may occur. Moleculdynamic simulations (Sec. VI. B) may provide the onlmeans of calculating the primary damage state in sucascades.

In an a-decay event, thea-particle anda-recoilparticle are released in opposite directions and produdistinct damage regions separated by several microBased on full cascade Monte Carlo calculations usiTRIM-96 (assuming a threshold displacement energy25 eV), the average number of atomic displacemengenerated in a ceramic, such as zircon, by the 5.2 Ma-particle and the 86 keV235U recoil released in thedecay of 239Pu are 220 and 1180, respectively. Thaverage number of displacements generated pera-decayevent by the decay of the actinides in HLW ceramics is expected to be similar and is estimated to1400 displacements pera-decay event. This is signifi-cantly more than the 0.1 displacements generatedb-decay event (see above). The relative numbersdisplacements per atom (dpa) generated byb-decay anda-decay in typical ceramics containing DOE tank wasor 10 wt. % 239Pu are shown in Fig. 4. The results inFig. 4 clearly show the dominance ofa-decay events(a-particles anda-recoils) in producing ballistic-typecollision damage in ceramic waste forms.

3. Transmutations and gas production

In addition to the energy transferred to the ceramb-decay anda-decay also lead to transmutation o


ts.he

,non

A

tss.

-0

ed

s,tlow-r

ch

ces.goftsV

-e

erof

e

,

FIG. 4. Relative number of displacements generated byb-decay anda-decay events in typical ceramics containing DOE tank waste aby a-decay events in typical ceramics containing 10 wt. %239Pu.

radioactive parent nuclei into different chemical elements that must be accommodated into the structuand may significantly impact the chemical propertieof the ceramic. The principal source of transmutatioin HLW is b-decay of the relatively abundant fissionproducts,137Cs and 90Sr. Transmutation of these twoelements is accompanied by changes in both ionic radand valence. Cs11 decays to Ba21 with a decrease inionic radius of 20%, and Sr21 decays to Y31, which inturn decays to Zr41 with a final ionic radius decrease o29%.59 In ceramics for Pu immobilization, where the Pconcentration is high, the transmutation of239Pu to235U(both of which are fissile) may also result in changesvalence state and ionic radius, which could impact lonterm performance and safety [e.g., U61 as the uranyl ion(UO2)21 in solid phases is much more soluble than U41].

Helium atoms, which result from the capture of twelectrons bya-particles, are also produced and mu(i) be accommodated in the ceramic interstitially, (ii) btrapped at internal defects, (iii) aggregate to form bubles, or (iv) be released at the ceramic surface. For HLmost of the He will be generated over long time periodnear ambient temperature, and the concentration shonot exceed 100 atomic parts per million (appm), whicshould be easily accommodated. In the case of ceramwaste forms for Pu immobilization and disposal, He cocentrations can become quite high and difficult to accommodate within the structure, exceeding 2 atom % aft100,000 years for ceramics containing 20 wt. %239Pu.

3, No. 6, Jun 1998


-;ot

nht

o

s

e

n-orteseedinkyedrial

r

yd

d

le

ci-e)ald.

seersto

thseedr

ies.

ivemfor-

lk

s.

-le

C. Irradiation techniques

The effects of radiation from the decay of radionuclides will accumulate over very long time periodsconsequently, a broad range of accelerated irradiatitechniques must be utilized to study radiation effecin ceramics for HLW and Pu disposition. The relativedamage rates in natural minerals, HLW forms, ceramics for Pu disposition, and for accelerated irradiatiomethods are shown in Table I. These irradiation tecniques and procedures have been described in depreviously,37,38,41 and are briefly reviewed here.

1. Actinide-incorporation

The long-term effects ofa-decay events can besimulated by incorporating short-lived actinides, such a238Pu (half-life of 87.7 years60) or 244Cm (half life of18.1 years60) in sufficient concentrations (0.2 to 3 wt. %)to achieve 1018 to 3 3 1019 a-decay eventsyg in labora-tory time periods of up to several years. This techniquhas been utilized in studies of several HLW ceramicand single-phase ceramics, as discussed below, aeffectively simulates, at accelerated dose rates, the simtaneousa-particle anda-recoil effects that are expectedover long periods of time in ceramic phases for HLWor Pu immobilization.

2. Actinides in natural minerals

Unlike natural glasses, which do not contain enouguranium and thorium (,100 ppm) to serve as naturalanalogues for radiation effects in waste form glassenatural minerals can contain large concentrationsU and Th impurities (up to 30 wt. %) and, thereforecan serve as natural analogues27 for radiation effectsin analogous waste form structures, as discussedmore detail below. For example, the enhanced chemicreactivity (differential etching) ofa-recoil tracks is welldocumented in naturally occurring minerals.61

3. Charged-particle irradiation

Charged-particle irradiation using electrons, protona-particles, and heavy ions can be used to simulaand study radiation effects in ceramic phase relevantHLW and Pu disposition, particularly over a wide rangof temperatures. These irradiation techniques genera

TABLE I. Relative damage rates.

Material Damage rate (dpays)

HLW forms 10216210211

10 wt. % Pu waste form 10211

Natural minerals ,10217

Actinide doping 1021021028

Neutron irradiation 102721026

Ion-beam irradiation 102521022


ns

-

-ail

s

esndul-

h

s,f

,

inal

,teto

lly

employ particle accelerators and high dose rates; cosequently, very significant doses are reached in shperiods of time (e.g., minutes). Interpretation of thresults from such experiments can be difficult becauthe irradiated areas are thin surface layers of restrictlateral extent. The high surface area can act as a sfor migrating defects, and electric field gradients mabe generated along the electron or ion paths. Implantions also can change the properties of the target mateas a result of compositional changes. Still,a-particleirradiation, using either sealed actinidea-sources (e.g.,238Pu) or particle accelerators, is an effective tool founderstandinga-particle effects; similarly, heavy-ion(e.g., Xe, Pb) irradiation is an effective method to studa-recoil effects. In fact, the two techniques can be usesimultaneously to study the effects of botha-particlesanda-recoils on ceramics. Similarly, electron irradiationcan be used to study the effects of ionization anelectronic excitations fromb-particles andg-rays onceramics. One minor disadvantage of charged-particirradiation, particularly for simulating the damage froma-decay events over a large area in a multiphase spemen, is that it cannot selectively irradiate (i.e., damagonly the actinide host phases as would occur in an actumultiphase waste form; instead all phases are irradiateThis is not a disadvantage in studies of single phamaterials, and homogeneously damaged surface laycan be produced using multiple energy beams. In ordercompare the results of charged-particle irradiations withose from actinide-decay studies, the comparable doin displacements per atom (dpa) can be calculated, bason the binary collision approximation, using computecodes, such asTRIM56 and MARLOWE,57,58 and known orassumed values for the displacement threshold energ

4. Gamma irradiation

Gamma irradiation, utilizing60Co or 137Cs sources,has been used to simulate the effects ofb-particlesand g-rays on glass.37,41,42 Very little of this workhas been done on ceramics because of the relatinsensitivity of many oxides to permanent damage froionization processes. However, some host phasesfission products could be susceptible to ionization damage. The advantage of this technique is thatg-rays areso penetrating that samples can be irradiated in buand while sealed in containers; furthermore, theg-raysprovide a realistic simulation since they interact withthe ceramic primarily through ejected photoelectronDose rates on the order of2.5 3 104 Gyyh are easilyachieved.41

5. Neutron irradiation

Irradiation with neutrons can also be used to simulate and study radiation effects in multiphase and sing

3, No. 6, Jun 1998 1441


t

gi

-

e

o

ttint

ttn

t

n

o

tr-c-

s.yon

o

g-

r

c-

nt

--

nt

t

phase waste ceramics. The main advantage of irradiawith neutrons, which dissipate their energy by ballisprocesses, is that significant numbers of atomic displaments can be produced in bulk specimens, makineasier to measure physical property changes assocwith ballistic damage. Neutron irradiation, however, onmoderately simulates the ballistic damage froma-par-ticles anda-recoils.41 In addition, fast-neutron irradiation of ceramics provides only limited simulation of Hbuildup froma-decay through the generation ofa-par-ticles by (n, a) reactions, which is different than thcase for borosilicate glasses where the B(n, a)Li reactioncan generate significant He.38,41 Another technique62

involves fission of fissile nuclides (e.g.,235U) in theceramic by irradiation in a fast and/or thermal neutrflux; the fission event results in very high-energy fissifragments that produce extensive regions of damage (fission tracks) that may or may not simulate ballisdamage from thea-recoil particles. Such fission evenprovide an excellent simulation of spontaneous fissevents in HLW and Pu ceramics; however, as mentioabove, these events are so infrequent in actual HLWthey are unimportant as damage mechanisms. Similacharged-particle irradiation, neutron irradiation generadamages all phases in a multiphase ceramic to sometent, although fission-track damage may be localizedphases containing fissile nuclides; consequently, neuirradiation does not provide an accurate simulation ofheterogeneous damage that may occur due to actipartitioning.

D. Damage production

The production of radiation damage effects inwide-range of ceramics has been studied for many yeand the current state of knowledge is summarizedseveral reviews.37,40,63–69 One of the most importanfundamental parameters affecting radiation damagea material is the threshold displacement energy,Ed,which is the minimum kinetic energy necessary to dplace an atom from its normal site. As noted aboand discussed below, this parameter is essentialquantifying the number of displaced atoms producin irradiated materials using computational approachand provides primary damage information for modeliradiation effects. Since ceramics generally consistmultiple sublattices,Ed must be separately measure(or calculated) for each sublattice. Table II summarizthe recommended values of the threshold displacemenergies for several ceramics, as determined fromrecent evaluation of the literature.69 Typical values rangefrom 20 to 60 eV. There are no known measurementsthe displacement energies in the ceramic phases thaunder consideration as matrices for radionuclides. Ufortunately, there are vanishingly few electron irradiati

1442 J. Mater. Res., Vol.

tingicce-

itatedly

e

onn

i.e.,icsonedhatr tollyex-to

ronheide

aars,in

in

is-vefor

edes,gofdesent

a

oft aren-n

facilities still available in the USA where the thresholddisplacement energies of materials can be determined.

The recommended method for measuringEd is toirradiate the sample with electrons while monitoring theinduced-defect concentration as a function of electronenergy to determine the minimum electron energy awhich defects are produced. Preferred techniques fomonitoring the defect concentration are optical spectroscopy or electron paramagnetic resonance (EPR) spetroscopy, since they are capable of uniquely monitoringthe behavior of specific defects, such as anion vacancieHowever, in some cases, the EPR or optical signal manot be detectable unless displacement damage occursboth sublattices.70,71 Other techniques (e.g., high voltageelectron microscopy) have been successfully used tdetermineEd .69 Measurements are typically performedat cryogenic temperatures to minimize defect recoveryprocesses. The relationship betweenEd and the threshold(minimum) electron energy,Ee, is given by72,73

Ed 2EesEe 1 2mec2dyMc2, (1)

whereme is the electron mass,c is the velocity of light,and M is the mass of the displaced ion. Some studieshave estimatedEd by performing electron irradiationsat energies above the threshold value and comparinthe measured point defect concentration with the concentration predicted by theory; this method typicallyoverestimatesEd.69 Measurements on Al2O3 and MgOindicate thatEd is approximately constant over the tem-perature range from 78 to 400 K.74,75 The temperaturedependence ofEd in ceramics relevant for HLW or Pudisposition is unknown and needs to be investigated.

Computer simulation techniques, as discussed inSec. VI, can be used to confirm experimental values foEd or to calculate values ofEd when experimental valuesare unavailable, as is the case for most of the ceramiphases of interest for nuclear waste applications. Recently, values ofEd calculated by computer simulationtechniques have been shown to be in good agreemewith many of the oxide values76 and SiC values77 inTable II. In the case of zircon (ZrSiO4), computer simu-lation results,76 which are given in Table III, provide theonly available estimates forEd values.

Given the threshold displacement energies, theproduction of displacements in these polyatomic materials by ballistic processes can be determined by computational approaches based on the binary collisionapproximation (BCA). As already mentioned, thesecomputational approaches generally involve usingcomputer codes, such asTRIM56 andMARLOWE.57,58 TRIM

assumes a structureless medium (i.e., it takes no accouof crystallinity) and uses a Monte Carlo approach todetermine the scattering angle and energy transfer tharesults from each binary elastic collision.MARLOWE is

13, No. 6, Jun 1998


e

ts

ndmn

oa

su

hrrni

uoa

c

lk

ici-oforte

toofghsis

etitearetti-ry

rn

toar

nts,

ts

tofor

ted

tedldly

o

TABLE II. Recommended threshold displacement energies for seral simple ceramics.69(* indicates less reliable data.)

Material Threshold displacement energy

Al 2O3 Ed sAl d 20 eV, EdsOd 50 eVMgO EdsMgd 55 eV, EdsOd 55 eVCaO EdsOd 50 eVMgAl 2O4 EdsOd 60 eVZnO EdsZnd 50 eV,p EdsOd 55 eVBeO EdsBed 25 eV,p EdsOd 70 eVp

UO2 EdsUd 40 eV,p EdsOd 20 eVSiC EdsSid 40 eV,p EdsCd 20 eVGraphite EdsCd 30 eVDiamond EdsCd 40 eV

TABLE III. Minimum threshold displacement energies,Ed , calcu-lated for zircon.76

Specific ion Ed (eV)

Zr 79Si 23O 47

a BCA code that does take into account crystalliniby assigning all atoms to well-defined initial positionin a crystal lattice. The net number of displacemencan also be determined directly by numerical solutioto the integro-differential equations of Parkin anCoulter,78 which also assume a structureless mediuIf the same scattering cross sections and electrostopping powers are used, the results ofTRIM simulationsare in agreement with numerical solutions to thintegro-differential equations.79 The binary collisionapproximation, however, is a high-energy approximatithat is appropriate for high-energy collisions and mbe reasonable for light ions, such asa-particles. Atlow energies, where the trajectories of recoils are neasily described by discrete collisions, such as in aa-recoil cascade, the binary collision approximation is leuseful; it also fails to take into account the simultaneorecombination events that can occur. Molecular dynamsimulations, as discussed in Sec. VI. B, may provide tonly means of calculating the nature of the primadamage state in such cascades. Nonetheless, in oto semiquantitatively compare the results of irradiatiowith ion beams of different masses and energies weach other and with the results of self-radiation duea-decay, the BCA codes often must be used becamolecular dynamic simulations cannot be performedall the materials of interest due to the lack of adequinteratomic potentials.

Radiolysis can also result in defect production asconsequence of localized electronic excitations produby ionization.63,65,66,68 This mechanism is important insome ceramic systems, such as alkali halides, a


v-

y

tss

.ic

e

ny

ot

ssice

ydersthtosente

aed

a-

line earth fluorides, and silica,63,65,68 but is generallyconcluded to be insignificant in many other cerammaterials. Under the irradiation environments anticpated for nuclear waste materials, the productiondisplaced atoms from radiolysis is expected to be minin most crystalline ceramics of interest. Some silicaand phosphate systems, however, may be sensitivethis mechanism. Table IV summarizes the sensitivitycandidate nuclear waste ceramics to radiolysis at hidose rates. A wide range of chemical compositionis possible for most of these structure types, and thmay affect their sensitivity to radiolysis. This can bimportant for some structure types, such as the apastructure, where both silicate and phosphate systemsof interest. The sensitivity of potential fission-produchost phases (e.g., apatite, monazite, NZP, and silicotanate structures) to radiolysis under expected repositoconditions is unknown.

Obviously, threshold displacement energies foelastic collisions cannot be measured by electroirradiation methods in materials that are sensitiveradiolysis because the ratio of electronic-to-nuclestopping powers for electrons with energies,1 MeV isgreater than 104 and increases with decreasing electroenergy. Therefore, the radiolytic production of defecin much higher than that due to elastic collisionsmaking any determination ofEd impossible. Estimatesof Ed in radiolysis-sensitive materials may be besobtained by theoretical or semi-empirical comparisonwith other ceramics. AdditionalEd data on a widerange of ceramic materials are needed in orderallow accurate estimates of displacement energiesradiolysis-sensitive materials.

Due to the low average knock-on energies associawith a-particle and b-particle interactions with thehost ceramic, most of the knock-on energies associawith these processes will be close to the threshoenergies for atomic displacements. It is, therefore, like

TABLE IV. Sensitivity of candidate nuclear waste ceramics tradiolysis.

Ceramic phases Radiolysis sensitive

Actinide hosts:zircon nozirconolite nosilicate apatite nofluoroapatite yesmonazite yesNZP unknownZrO2 no

Fission product hosts:CsCl, SrF2 yes

Hydroxylated ceramics:clays, zeolites, etc. yes

Silicotitanates unknown

3, No. 6, Jun 1998 1443


m

eocni

t

i

o

t

b

i

,

atnfedhee

n

exnenet,

d

-ntse)

s

lls-

latg,e

se

te

ofis

that nonstoichiometric fractions of displacement daage will be produced on the different sublattices, dto differences inEd and atomic mass. Although thetotal number of displacements associated witha-particlecollisions is less than 20% of the nearly stoichiometrdisplacements calculated for thea-recoil cascade, theseisolated displacements have higher survival rates (lin-cascade recombination), and it is possible that thdisplacements may play a major role in the evolutiof microstructure in irradiated nuclear waste ceramiFurther work is needed to determine if there is asignificant effect associated with nonstoichiometric dplacements and the role of thea-particle displacementson damage accumulation processes.

E. Radiation damage kinetics

Because equilibrium thermodynamics are not stricapplicable under irradiation conditions, the accumulatiof point defects, evolution of microstructures, and fomation of metastable phases (e.g., amorphization) unirradiation are controlled by the nucleation and kineproperties of the system.80–83 The kinetics of radiationdamage accumulation are controlled by the competitbetween the rate of damage production and the rof various simultaneous recovery processes (e.g., clopair recombination, defect migration, defect aggregatioepitaxial recrystallization). The kinetics of simultaneourecovery processes can be expected to vary dependinthe irradiating ion mass and energy, the bulk temperatof the sample, and the melting temperature or cohesenergy of the irradiated material. Experiments in cerammaterials have shown that light ions are less efficiewith increasing temperature, at irradiation-induced amphization than heavy ions.84–89 These observations areconsistent with metal irradiation studies that show thlight ions will have a larger fraction of the initiallyproduced Frenkel defects remaining at the end ofcascade event as compared with the fraction remainin cascades produced by heavier ions.90 In ceramics,ionization-induced diffusion can also affect recovekinetics.67,69,91,92

Accelerated irradiation methods (Sec. III. C) mustused to achieve dose levels, in laboratory time framthat are relevant to long-term repository performancConsequently, a detailed understanding of the depeence of radiation-damage accumulation on time, teperature, and damage rate is important to predictthe long-term behavior of ceramic waste forms undexpected geologic repository conditions. For exampbecause the high damage rates overwhelm the simuneous thermal recovery processes, the effects of elecand ion irradiation generally occur at much higher temperatures than the effects from self-radiation damagesignificantly lower damage rates) in actual nuclear wa


-ue

ic

esssens.y

s-

tlyonr-deric

onatese-n,s

g onureiveic

nt,r-

at

heing

ry

ees,e.nd-m-ngerle,lta-tron

-(at

ste

ceramics, in ceramics doped with short-lived actinidesor in mineral analogues. This effect is mitigated to someextent by the irradiation-assisted recovery processes thare also observed under high-dose electron and ioirradiation. A systematic, integrated understanding odose-rate effects is necessary to extrapolate data obtainusing accelerated methods, under high dose rates, to tlong-term low-dose-rate behavior of actual nuclear wastceramics. This dose-rate difference ranges from 106 to1014 (Table I). Comparisons to natural mineral data caninvolve even larger dose-rate differences. Some data odose-rate effects will be presented below (Sec. V. E).

IV. RADIATION EFFECTS IN MULTIPHASECERAMIC WASTE FORMS

Waste forms for the immobilization of HLW atDOE sites in the USA, HLW worldwide, plutonium, andother special nuclear waste streams represent complmulticomponent systems, and the clean-up campaigwill proceed for several decades. Thus, there is timto identify and address fundamental scientific issues, iparticular those related to the effect of radiation on wastform properties, which may have a bearing on the cosoptimization of processing technologies, the integrity ofthe waste forms during interim storage/operations, anlong-term performance.

Multiphase ceramic waste forms are tailored to produce specific crystalline phases as hosts for the differeradionuclides. Generally, fission products (such as Cand Sr) are confined to one or more glass or crystallinphases, while the actinides (U, Np, Pu, Am, and Cmgenerally partition into other crystalline phases. Synrocand other related titanate-based ceramic waste formhave received the most attention.93–97 There have beenmore limited studies of supercalcine,98,99 a silicate-basedtailored ceramic, and of glass-ceramics,11,100–102 whichare prepared by controlled crystallization of suitableglasses. In general, the cumulative radiation effects in athe component phases will contribute to the actual effectof radiation on the performance of these and other potential multiphase waste forms. However, due to differentiaeffects, such as radiation-induced volume changes thcan lead to differential stresses and even microcrackinradiation effects in multiphase waste forms may be morthan the sum of the effects in individual phases, whichare described in more detail below (Sec. V). The studieof radiation effects in several multiphase ceramic wastforms are summarized below.

A. Synroc

Synroc is a dense, multiphase titanate-based wasform designed for the immobilization of HLW, withphases based on mineral analogues. The attractionthis multiphase ceramic is that the phase assemblage

13, No. 6, Jun 1998


i

s

n

n

e

egs

a

w

ii

l

o

u

ss

.

-

ion

-n

ea

e,.

at

0,ordil-eso

t-ic

insensitive to the HLW/precursor ratio; in most cases, thratios of the phases vary as the waste loading changfrom 0 to 35 wt. %. The Synroc phases, in approxmately equal abundance, are zirconolite (CaZrTi2O7),which incorporates rare earths (RE) and actinides (Anperovskite (CaTiO3), which incorporates RE, An, andSr, hollandite (BaAl2Ti6O16), which incorporates Cs,Rb, and Ba, and rutile (TiO2). In some formulations,pyrochlore, A2B2O7, which is structurally closely relatedto zirconolite, may become an important actinide hophase. Some fission and corrosion products (Fe, NCr, Pd, Rh, Ru, Tc, Mo, etc.) will form metal alloys.In general, the composition of the precursor can breadily adjusted for different types of HLW. Possiblenear-term applications of Synroc are immobilization oTc (technetium) and other elements separated duricleanup of U.S. DOE tank waste liquids, Synroc/glascomposites aimed at HLW sludges from U.S. DOEtank wastes, and zirconolite/pyrochlore-rich ceramics foimmobilization of Pu-rich materials. There has beeinterest in France and more particularly in Russia sinc1994 in producing Synroc and Synroc/glass compositby induction melting.

Ringwoodet al.103 have argued that Synroc is stablewith respect toa-decay damage, based on an assessmof natural minerals of uranium and/or thorium-bearinperovskite and zirconolite, which are the actinide hophases of Syntoc. Such data, however, can provide onqualitative results, as both zirconolite and perovskitare susceptible to irradiation-induced amorphization.37

In the early 1980s, a limited amount of accelerateradiation-damage testing was performed on sinterecoarse-grained Synroc and its constituent phasesirradiation with fast neutrons to fluences of up to2.7 3

1026 nym2 (E . 1 MeV), which is equivalent in damageto about8 3 1018 a-decaysyg.104–106Microcracking wasobserved in the neutron-irradiated samples and cotributed to the large volume expansions (up to 8.5%) thwere observed. Accelerated testing of238Pu-doped Syn-roc (and its constituent phases) has also been compleon hot-pressed materials.107,108The macroscopic swellingin the Pu-doped Synroc increases with dose, as shoin Fig. 5, and exceeds 6%, with only a slight indicationof approaching a steady-state (saturation) value. Studof ion-irradiated Synroc indicate significant increasesthe leach rate at high damage levels.109,110

For the multiphase Synroc ceramic, the hollanditphase will be subject only toa-particle andbyg irra-diation, while the zirconolite and perovskite phases wibe subject toa-decay damage processes, involving bota-particles anda-recoils, as well asbyg irradiation.Since byg irradiation damage is generally assumed tbe negligible in these phases, studies of radiation effecin the major constituent phases of Synroc (i.e., hollandite, perovskite, and zirconolite) have been carried o


ees-

),

ti,

e

fg

s

r

es

nt

tly

e

dd,by

n-t

ted

n

esn

e

lh

ts-t

FIG. 5. Macroscopic swelling in the Pu-doped Synroc,107 Cm-dopedsupercalcine,123 and Cm-doped123 and Pu-doped126 glass-ceramics.

using actinide-doping,a-particle irradiation, neutron-irradiation, and ion-beam irradiation techniques. Detailfrom the studies of these specific crystalline phaseare reviewed elsewhere37 and are briefly described inthe following section on relevant crystalline phasesThe essential conclusions are thata-particles induce alarge unit-cell volume expansion (up to 2.5%) and adisplacive transformation from the tetragonal to a lowersymmetry monoclinic structure in hollandite.111 In ad-dition, the zirconolite107,112–116and perovskite107 phaseswill undergo a-decay-induced amorphization with vol-ume changes in the range of 6 to 8%.

B. Synroc-related ceramics

Japanese researchers have investigated self-radiateffects in a244Cm-doped Synroc-related titanate ceramicconsisting of perovskite, zirconolite, hollandite, freudenbergite, and loveringite that is intended for encapsulatioof sodium-rich high-level waste.117,118 The density ofthis material decreased linearly with dose to a valucorresponding to a volume expansion of 1.0% afterdose of0.7 3 1018 a-decaysyg; above this dose, the rateof density change increased to a new constant valuand this phenomenon is attributed to microcrackingThe maximum volume expansion approaches 4%1.8 3 1018 a-decaysyg, with no indication of saturation.The leach rates for Na and Cs increased by a factor of 1while the leach rates for Sr and Ca increased by a factof 100.118 These increases were attributed to increasesurface area from microcracking and decreased durabity due to radiation damage. A Synroc-related titanatceramic containing sodium-free simulated waste has albeen studied by doping with 0.91 wt. %244Cm.119 Boththe perovskite and zirconolite phases exhibited unicell expansions on the order of 2.6%. The macroscop

3, No. 6, Jun 1998 1445


lu

iz

n

h

rr

ea

re

ts

s

aT

so

emtn

ts,

l

eo-%a

se

ys

red

d,l

s

e

de

b-

s

he

il

t-aln.

volume expansion increased linearly with dose to a vaof 1.7% after a dose of1.2 3 1018 a-decaysyg.

Researchers at Sandia National Laboratories120 haveinvestigated a multiphase titanate ceramic to immobilacidic high-level nuclear waste. This titanate waste cramic consisted of rutile (TiO2), an amorphous silicatephase, perovskite, zirconolite, hollandite, and a U-Zrich phase that was believed to have the pyrochlostructure. Radiation effects froma-decay in this titanateceramic were simulated by irradiating TEM specimewith Pb1 ions.121 Multiple energies (40 to 250 keV) wereused to produce a uniform damage profile. Irradiationsingle energies of 240 to 250 keV were also performeExamination by transmission electron microscopy of tirradiated specimens indicated considerable damageall phases after a dose equivalent to2 3 1018 a-decayeventsyg. All phases remained crystalline, except thphase with a chemistry typical of a pyrochlore, andbecame amorphous.

C. Supercalcine

Radiation effects in this multiphase waste form westudied at the Pacific Northwest National Laboratoin the late 1970s by incorporating244Cm in a super-calcine formulation consisting of three major phaswith the fluorite structure, the apatite structure, andunresolved tetragonal structure type.122,123 Subsequentanalysis suggested that the Cm predominantly partitioninto the apatite phase and the tetragonal phase. X-diffraction indicated a gradual transformation of thapatite from a crystalline to an amorphous state,agreement with the results124 on the amorphization ofsimilar apatite crystals in a devitrified nuclear wasglass. There also was a slight increase in the intenof the diffraction maxima associated with the tetragonphase. The stored energy reached a maximum vaof 42 Jyg at a dose of0.5 3 1018 a-decay eventsygand decreased slightly with further increases in doThe energy release was not complete at 600±C (theupper limit of the calorimeter used); therefore, additionenergy release may occur at higher temperatures.density decreased exponentially with dose, resultinga volume expansion of 1.4% at a dose of1.2 3 1018

a-decay eventsyg (Fig. 5).

D. Glass-ceramics

A glass-ceramic is a fine-grained mixture of glaand ceramic phase ideally derived from a homogeneglass through a heat-treatment at the temperature of mimum nucleation rate for the ceramic phases, followby a high-temperature treatment to yield a maximugrowth rate. The celsian glass-ceramic, developed atHahn-Meitner Institut and named for the predominacrystalline phase, is easy to fabricate, is homogeneou


e

ee-

r-re

s

atd.ein

eit

ey

sn

eday

in

eityallue

e.

lhein

susax-d

het

sly

crystallized, and contains a variety of leach-resistanhost phases for the radionuclides. In the late 1970radiation effects in a244Cm-doped celsian glass-ceramicwere investigated at the Pacific Northwest NationaLaboratory.123 At the same time, similar studies werebeing performed on both a244Cm-doped celsian glass-ceramic125 and a 238Pu-doped celsian glass-ceramic126

within the European Community. In these studies, thdensity decreased exponentially with dose, and the assciated volume change was projected to saturate at 0.5(Fig. 5). The stored energy increased with dose tosaturation value of 80 Jyg,123 and the fracture toughnessincreased by 25%.125 Data from the x-ray diffractionanalysis revealed that a Cm-rich, rare-earth titanate phawith the pyrochlore structure underwent a small vol-ume expansion with dose and eventually became x-radiffraction amorphous. The fractional helium release wadetermined to be 3% in samples stored at 170±C to acumulative dose of 1.13 1018 a-decay eventsyg.

In the study of a devitrified Cm-doped wasteglass,124 the Cm partitioned into two crystallinephases: Ca3(Gd, Cm)7(SiO4)5(PO4)O2 (apatite structure)and (Gd, Cm)2Ti2O7 (pyrochlore structure). The volumeexpansion of the glass saturated at 1%, and the measustored energy was 90 Jyg. Both of these crystallinephases were observed to undergo a radiation-inducecrystalline-to-amorphous transformation. The differentiaexpansion (estimated to be 5 to 8%37) associated with thecrystalline-to-amorphous transformation of these phaseresulted in significant microfracturing (Fig. 6), whichcan greatly increase the surface area for radionuclidrelease.

Vance et al.127 used 3 MeV Ar1 ions to study ra-diation effects in a sphene glass-ceramic. They reportethat complete amorphization of the crystalline sphenphase occurred at an ion fluence equivalent to 73

1018 a-decaysyg. No significant enhancement in theleaching of the irradiated sphene glass-ceramic was oserved. A sphene glass-ceramic containing 2 wt. %238Puwas reportedly prepared to studya-decay effects128; noresults have been reported to date.

E. Scientific issues for multiphase waste forms

Radiation effects in multiphase ceramic waste formconcern both behavior of the individual crystallinephases and behavior due to the multiphase nature twaste form. Key issues regarding radiation effects inindividual crystalline phases are discussed in detain the following section, such as the nature of thepoint defects produced by radiation, the displacemenenergies for ions on different sublattices, and radiationinduced amorphization and volume changes. For actumultiphase ceramic waste forms, the behavior at graiboundaries and interfaces will become important

3, No. 6, Jun 1998


FIG. 6. Effect of a-decay in crystalline phases on microstructure in a partially devitrified glass: (a)6 3 1015 a-decaysyg and (b)2.4 3 1017 a-decaysyg.124

dyiahifne

icccioayao

a

tahsrsioi

issnin

gesr

ite

ey

nceZr

dr

tei,,

ies

ve

Consequently, studies should (i) measure and mothe microfracturing that occurs in polyphase and polcrystalline ceramics due to radiation-induced differentand anisotropic volume expansions, (ii) determine tinfluence of grain size on the tendency of such ceramto microcrack, and (iii) determine the influence ointergranular or interfacial stresses on properties abehavior. Clearly, additional characterization techniqumust be employed in investigations ofa-decay effects inthese nuclear waste ceramics. Point defects in manythese ceramics can be investigated by EPR and optspectroscopies. In principle, changes in the valenstates of actinides, rare-earth, and transition elementsbe addressed by Auger, x-ray photoelectron/absorptspectroscopy, electron energy loss spectroscopy,(possibly) diffuse UV-near IR reflection spectroscopNearly all the ceramics under consideration for nuclewaste immobilization are oxide-based, and neutrdiffraction enjoys advantages over x-ray diffraction fothe study of radiation-induced oxygen defects.

V. RADIATION EFFECTS INCRYSTALLINE PHASES

There have been a number of studies of individusynthetic phases and minerals that are structurally andchemically analogous to the phases found in mulphase ceramic waste forms or to proposed single phwaste forms. These studies contribute significantly to tunderstanding of radiation effects in the multiphaceramic waste forms by providing a detailed undestanding of the behavior of each component phaThe current state of understanding regarding radiateffects and radiation damage processes in crystall


el-lecs

ds

ofaleann

nd.rn

r

l/ori-seee-e.n

ne

phases proposed as host matrices for radionuclidesbriefly summarized below and critical scientific issueare identified and discussed. More detailed informatioand data are available in the original references anda recent review.37

A. Phases and structures

1. Actinide-bearing phases

The most prominent of the potential actinide-bearinphases are listed in Table V. In these structures, actinidare generally accommodated by substitution for Zr othe rare-earth (RE) elements. Phases with the fluorstructure (PuO2 and UO2) or modified fluorite structure(ZrO2) are also important. The low solubility of PuO2

in many materials can result in the formation of PuO2

in both glass and ceramic waste forms. Uranium dioxid(UO2) is the main constituent of spent fuel and maalso form in U-rich wastes. Zirconia (ZrO2) is a verystable material under a wide range of oxidation/reductioenvironments and may be very radiation resistant. SinPu and other actinides can be accommodated on thesite, ZrO2 is a candidate material for Pu disposition anfor use as an inert matrix for Pu burn-up in a reactor oaccelerator-based neutron source.31,32

Pyrochlore (A2B2O7) is a derivative of the fluoritestructure type129,130 in which the A-site contains largecations (Na, Ca, U, Th, Y, and lanthanides) and the B-siconsists of smaller, higher valence cations (Nb, Ta, TZr, Fe31). Actinides may be accommodated in the A-siteand charge balance is maintained by cation deficiencin the A-site and substitutions on the A- or B-sites.131

Rare-earth titanates with the pyrochlore structure ha

3, No. 6, Jun 1998 1447


at

for

yrct

-d-easinof

l

ificss-

eas

on

rth

si-

ilyal

ilyalth

-ge

for

-

ble

TABLE V. Potential actinide-bearing phases.

Structure type Composition

Oxidesfluorite PuO2; UO2

pyrochlore A2B2X6Y; RE2Ti2O7; CaPuTi2O7

zirconolite CaZrTi2O7

perovskite CaTiO3zirconia, ceria ZrO2; CeO2

Silicateszircon ZrSiO4

apatite Ca42xRE61x(SiO4)62y (PO4)yO2

sphene (titanite) CaTi(SiO4)OPhosphates

monazite CePO4apatite Ca42xRE61x(SiO4)62y (PO4)yO2

NZP NaZr2(PO4)3

been observed as actinide-host phases in some nucwaste glasses,124 in titanate ceramic waste forms,120 andin a glass-ceramic waste form.123,126 The synthesis ofPu2Ti2O7 and several other Pu-containing titanates withe pyrochlore structure has also been reported.132 Asnoted above, zirconolite (monoclinic CaZrTi2O7) is afluorite-derivative structure closely related to pyrochloand is the primary actinide-host phase in Synroc. Raation effects in Pu- and Cm-containing pyrochlores azirconolite have been studied.36,107,112–116,133In addition,studies of metamictization (i.e., amorphization inducegeologically bya-decay of U and Th impurities) in min-eral analogues for pyrochlore134–137and zirconolite138–141

provide data on radiation effects over geologic time peods. There have also been ion-beam142,143and neutron144

irradiation studies of pyrochlores and zirconolite. Iaddition to extensive studies of radiation effectsnaturally occurring pyrochlores, systematic studies habeen completed of the alteration patterns of the naturaoccurring compositional end-members of this isometphase: microlites,145 pyrochlores,146 and betafites.137 Ithas been shown that the increased degree of metamtization (i.e., highera-decay event doses) increases thsusceptibility of these phases to alteration under natuconditions.

Perovskite, CaTiO3, is a mineral that may assumewide range of compositions as stable solid solutions ais a major constituent phase in Synroc and other titanceramic waste forms. Perovskite is orthorhombic aconsists of a 3-dimensional network of corner-shariTiO6 octahedra with Ca occupying the large void spabetween the octahedra (the corner-sharing octaheare located on the eight corners of a slightly distortcube). The actinides, rare-earths, and other large catireadily replace Ca in the structure, while smaller-sizcations replace titanium. Natural perovskites containiU and Th impurities that have received doses ofto 2.6 3 1018 a-decaysyg show a high degree of


lear

th

redi-nd

d

ri-

ninvelly

ric

ic-eral

andatendngcedra

edonsedngup

atomic periodicity, no evidence of amorphization, andunit-cell volume expansion of 1.8% due to defecaccumulation.147 However, perovskites do amorphizeunder ion-beam irradiation.148–152

Zircon (ZrSiO4) forms as one of several crystallinephases in glass-ceramic waste forms.24–26,153,154BecausePu can readily substitute for Zr in this structure,155,156

zircon has been proposed as a durable ceramic phasethe immobilization and disposition of Pu in the USA19,20

and high-actinide wastes in Russia.157,158 Additionally,zircon is a prominent actinide-bearing phase formed bcrystallization in the core-melt at the Chernobyl NucleaPower Plant,157 has been observed as a corrosion produon HLW glass,159 and is a phase extensively used inUyPb radiometric age-dating, as it may contain uranium in concentrations up to 20,000 ppm. Holland anGottfried160 investigated radiation effects and metamictization in natural zircons in an effort to use the degreof damage as a measure of the zircon’s age. This wthe first study to quantitatively correlate the decreasedensity, decrease in refractive index, and expansionthe unit cell parameters with increasinga-decay dose.The effects ofa-decay in both Pu-containing and naturazircons,84,155,156,161–163as well as ion-beam irradiationeffects in zircon,84 have been extensively studied andmodeled. These studies have resulted in a model (specto zircon) for describing the kinetics of radiation effectover geologic time periods in zircon containing weapongrade Pu or other actinides.164,165

Several rare-earth silicate-phosphates with thapatite structure have been observed or proposedactinide-host phases in HLW glass.124 supercalcine,98,99

glass-ceramics,100,101,166 and cement.167 Apatite phaseshave also been observed as recrystallized alteratiproducts on the leached surfaces of simulated HLWglasses,168,169 indicating their inherent stability relativeto HLW glasses. These apatite phases are rare-easilicate-phosphate isomorphs of natural apatite, Ca10-(PO4)6(F, OH)2, which is the most abundant of thephosphate minerals, and generally have the compotion: Ca42xRE61x(SiO4)62y(PO4)yO2 (where RE La,Ce, Pr, Nd, Pm, Sm, Eu, and Gd). The actinides readsubstitute for the rare-earth elements in this hexagoncrystal structure, and fission products are also readincorporated. Some apatites from the Oklo naturreactor site in the Republic of Gabon have retained boa significant235U enrichment (due to the decay of239Puincorporated during crystallization) and a high fissionproduct concentration in their structures, despite an aof nearly 2 billion years.170 Not surprisingly, the apatitestructure has been proposed as a potential host phasePu and high-actinide wastes.20,170–172Natural apatites cancontain U and Th, but are generally found in highly crystalline states173; however, partial metamictization hasbeen observed in natural apatites containing apprecia

13, No. 6, Jun 1998


i

u

t

s,

e

he

r-ly

e,

r

s

c

e-

al

g

rare earths, silica, and Th.174 Radiation effects inapatite phases have been extensively studied by Cdoping162,163,175–178and ion irradiation.85–87,179,180Thesestudies have resulted in another model, specificapatites, describing the kinetics of radiation effecover geologic time periods in apatite phases containweapons-grade Pu or other actinides.164

Sphene, CaTi(SiO4)O (the proper mineral name is ti-tanite), is a prominent phase in the sphene-glass-ceramSphene is an actinide-host (the structure will not accomodate Cs but will accommodate minor concentratioof Sr); thus radiation damage due toa-decay of actinidesis an issue of primary concern. In irradiation studieemployingg-radiation and 200 keV electrons to simulatirradiation effects from fission product decay elsewhein the waste form,181 sphene has been shown to brelatively resistant to ionization damage. In studiesnatural sphenes containing U and Th,182,183 the meta-mict state is reached after a cumulative dose of 53

1018 a-decaysyg, which is similar to that observed inion-irradiation studies.127

Monazite has been proposed as a single-phaseramic to incorporate a wide variety of nuclear wasteparticularly those rich in actinides.184,185Monazite is alsoa potential host phase for excess weapons Pu,20 but itsrelatively low thermal conductivity may preclude its usas a potential inert matrix to incinerate actinides.186 Themineral monazite is a mixed lanthanide orthophosphaLnPO4 (Ln La, Ce, Nd, Gd etc.), that often containsignificant amounts of Th and U (up to 27 wt. % combined). As a result, many natural monazites have besubjected to significanta-decay doses over geologictime (up to two billion years of age). In spite of the largradiation doses received by monazite minerals, theygenerally found in a highly crystalline state.88,165,187,188

The apparent resistance of natural monazites to radiatiinduced amorphization is an important factor in thproposed use of monazite for the immobilization onuclear wastes and the disposition of excess weapPu. Monatize is, however, readily amorphized under iobeam irradiation,88,89,189but it recrystallizes at relativelylow temperatures.190 The irradiation-induced swelling inmonazite may be on the order of 20%.186

Crystalline phosphates of the NaZr2(PO4)3 (NZP)family continue to be of interest as alternative wasforms for HLW and Pu disposition because the uniqNZP structure can incorporate a complex varietycations, including fission products and actinides.191–194

The NZP structure is a three-dimensional networkcorner-sharing ZrO6 octahedra and PO4 tetrahedra inwhich Cs and Sr can be accommodated in the intersticavities occupied by Na. The rare-earth elements aactinides can substitute for Zr, and other radionuclidcan substitute for either Na, Zr, or P. Thus, the NZstructure can accommodate nearly all the ions pres


m-

totsng

ic.m-ns

sereeof

ce-s,

e

te,s-en

eare

on-ef

onsn-

tee

of

of

ialndesPent

in HLW, as well as Pu. In a study of NaPu2(PO4)3

prepared with either239Pu or an isotopic Pu mixture thatwas predominantly238Pu, the samples with238Pu becameamorphous after a dose of 9.33 1018 a-decaysyg.195

Irradiation of NZP compounds withg-rays up to dosesof 3 to 5 3 108 Gy to simulatebyg damage does notproduce significant structural changes.195,196

2. Cs and Sr-bearing phases

Barium hollandite, BaAl2Ti6O16, is a major phase inSynroc that is intended to accommodate fission productsuch as Cs and Sr.93–97The structure of barium hollanditeconsists of a framework of TiO6 octahedra that arelinked (edge-sharing and corner-sharing) to form squarchannels parallel to thec-axis. These channels canaccommodate a wide variety of large cations, includingfission products. Although barium hollandite will notcontain actinides, it will experience irradiation froma-particles emitted in adjacent actinide-containingphases. In barium hollandite, thea-particles cause aunit-cell expansion and a displacive transformation fromthe tetragonal structure to a lower symmetry monoclinicstructure.111

Several phases have been proposed as hosts for tselective removal and solidification of fission productsfrom defense waste streams presently held in large stoage tanks. The removal of the fission products can greatsimplify the vitrification process for the remainder ofthe waste. These phases include the mineral analcimNaAlSi2O6 ? 2H2O, which forms a solid-solution serieswith pollucite, (Cs, Na)2 (Al 2Si4)O12 ? H2O, and a Cs-bearing sodium zirconium phosphate, NZP.192 Recently,much work has also focused on the crystalline silicoti-tanates (CST) that are a class of “zeolite-like” frameworkstructures consisting of titanium in octahedral (6-fold)and silicon in tetrahedral (4-fold) coordination withoxygens or hydroxyls. The framework structures havelarge “zeolite-like” columns or voids (.0.5 nm) thatare generally occupied by alkali elements and moleculawater; hydroxyl groups substitute for oxygen in order tomaintain charge balance. The resulting stoichiometriecan be quite complex, and in principle a wide varietyof structure types with voids of a variety of sizes canbe synthesized.197 Despite the apparent complexity ofthese structures, each framework with its characteristivoids is uniquely described by its molecular units andthe characteristics of the chains that incorporate thesunits. One of the potentially useful aspects of the crystalline silicotitanates is the very selective ion exchangecapacities of these structure types. Recently, severcrystalline silicotitanates with high affinities for Cs1 andSr21 in the presence of Na1 have been developed198–200;these materials have important potential for separatinCs1 from the Na-rich solutions in the Hanford Tank

3, No. 6, Jun 1998 1449


y

s

t

nze

n

sah

nC

e

nbnaler

r

dloa

noa

ani

d

e

,n

-

d

sal

u-

d

g

s

l

,

t

Wastes.201 The cumulative ionization dose from Cs decain such materials can be quite large, reaching 109 Gy in30 years.

Continuing radiation damage from solid-statebyg

radiolysis may be occurring in the 15 metric tons oradioactive Cs and Sr that are currently stored as Cand SrF2 in capsules at the Hanford site.2 Solid-stateradiolysis can lead to the formation of dislocation loopand metal colloids in alkali halides and alkaline-earfluorides.68 Studies202–205 have shown that metal colloidformation in NaCl, CaF2, SrF2, and BaF2 is sensitive toboth temperature and dose rate.

In addition to the potential for metal colloidformation, radiolysis may induce other defects anmicrostructural changes in phases specific for Cs aSr immobilization. For example, radiolysis in quart(SiO2) can lead to amorphization and large volumchanges.206 In alkali silicate, alkali-borosilicate, andHLW glasses, radiolysis can result in the formatioof point defects, free alkali, alkali peroxides, moleculaoxygen, and bubbles.38 The accumulation of such defectmay induce volume changes, produce stored energy,affect other properties. In addition, transmutations of tCs and Sr can affect phase stability due to changesionic radii and charge. Consequently, the radiation atransmutation stability of any phases proposed forand Sr immobilization should be investigated.

B. Defects and defect interactions

Almost nothing is known about the nature of thdefects and their interactions in the ceramic phasesinterest for nuclear waste immobilization. The migratioenthalpies of interstitials and vacancies, the recomnation volume for Frenkel pairs, and the displacemethreshold energies for cations and anions are importto understanding radiation effects in these materiaThe stable configurations of defects and defect clustshould also be known in order to describe microstructuevolution under irradiation. Although little is knownabout specific ceramic waste phases, there is a labody of literature on oxides, such as MgO, Al2O3,and MgAl2O4 that suggests that both stoichiometric annonstoichiometric interstitial clusters can form, depening on irradiation conditions. In addition, it is welknown that structural vacancies can suppress the fmation of defects and defect clusters in such materias in TiC12x

207,208 and MgO.209 The incorporation ofradionuclides could very well affect the concentratioof structural vacancies on both the cation and anisublattices and ultimately the accumulation of radition damage. Experiments have shown that damarecovery kinetics are well correlated with certain mterials properties. Research on the radiation resistaof magnesia spinel suggests that crystalline ceram


fCl

sh

dd

r

ndeinds

of

i-tnts.rsal

ge

d-

r-ls

n-

ge-cecs

will be more radiation resistant when the concentrationof equilibrium structural vacancies is high, barriers tointerstitial-structural vacancy recombination are low, thedefect energies of compositional disorder are small, anthe critical nucleus for the formation of stable interstitialloops is large.210–212

A number of scientific issues regarding defects anddefect interactions need to be resolved. These includ(i) the nature of the atomic defects produced by irradi-ation; (ii) fundamental properties of the atomic defectssuch as the atomic configuration and the recombinatiovolume of a Frenkel pair for both cations and anions, thedisplacement threshold energies, and the migration energies of point defects; (iii) nucleation and growth proc-esses of defect clusters in different radiation fields, interms of the fundamental properties of the materials; an(iv) the effects of deposition energy density, electronicexcitation, and/or low energy knock-ons on the structureand the stability of microstructures. Many of these issueregarding defects can be investigated by EPR and opticspectroscopies of electron-irradiated single crystals.

1. Role in damage accumulation

Natural minerals provide clear evidence for defect-recovery processes affecting radiation damage accumlation over geologic time under ambient conditions. Therecovery of point defects over 570 million years innatural zircons suppresses unit-cell expansions155–160,161

and has been shown to correlate with the decreaserate of amorphization (i.e., a phase transformation) innatural zircons relative to Pu-doped zircon.155 In naturalpyrochlores, the critical dose for complete amorphizationincreases with the geologic age,213 clear evidence forthe recovery ofa-recoil damage over geologic time.This recovery behavior has been modeled by assumin“fading” of the a-recoil “tracks” similar to that usedto describe fission track fading. From this model, themean life of a-recoil tracks at ambient temperaturein pyrochlore has been estimated to be 100 millionyears.213 Similar modeling of data for zirconolite andzircon yields mean lives for thea-recoil tracks of700 million and 400 million years, respectively.213 Suchstudies to determine meana-recoil track lives, however,are of limited value in predicting behavior at elevatedtemperatures without details of the kinetics of the variousrecovery processes (i.e., characteristic activation energieand frequency factors).

The standard approach to modeling microstructuraevolution under irradiation is to find the defect concen-trations from a mathematical rate theory approach.214,215

Self-radiation from radionuclide decay creates vacanciesinterstitials, and He atoms that can migrate throughthe structure and either recombine or be absorbed adefect sinks, thereby providing the driving force for

3, No. 6, Jun 1998


d

,

s

e

a

es

d

e

ad

e

t

-

d

in

s

-

processes such as defect clustering, bubble formatioand amorphization. This theory, however, was developefor use with isotropic materials, notably metals, anincludes several approximations that are of consequenwhen modeling microstructural evolution in generaland amorphization, in particular, in complex ceramicsSeveral issues of concern for application to compleceramics are as follows:

(i) There are more types of point defects (vacancieinterstitials and possibly antisite defects in each sublattice) in complex ceramics, including variable chargstates; hence, it is necessary to solve a much greanumber of coupled equations, since these defects rewith each other.

(ii) The assumption of a homogeneous mediumis clearly not valid when dealing with both multiplecations, which may be ordered, and more than onsublattice, as the production, migration, annihilationand defect reaction rates are all orientation and chemicspecies dependent.

(iii) Depending on their migration mode, differentdefects can have different annealing efficiencies.

(iv) Defect migration kinetics could also be affectedby ionization, as in the Bourgoin–Corbett mechanism.91

(v) As damage accumulates and disorder progressthe defect migration and formation energies (and posibly migration modes) may change. At the limit, theperiodic structure itself may disappear.

It is, thus, a formidable task to solve the general ratequations for ceramic structures. Furthermore, the resuobtained for one structure will not necessarily be valifor others because of the different topological propertieof the structures.

2. Enhanced diffusion

The dynamics of chemical diffusion are sensitive tothe structural state of a material.216 As is well knownfrom the study of ionic diffusion in aluminosilicatesof feldspar composition (e.g., CaAl2Si2O8, NaAlSi3O8),the diffusion coefficients of cations in the glass statare orders of magnitude faster than in the crystallinform.217 Thus, radiation-induced amorphization incrystalline phases can be expected to result in similincreases in ionic diffusion. Indeed, both theory anexperiments have demonstrated that diffusion cabe significantly influenced in ionic solids by bothmicrostructure and particle irradiation. Cation diffusionin both UO2 and (U, Pu)O2 is enhanced during reactorirradiation.218 In Al 2O3, the thermal diffusion of Fehas an activation energy of 3.09 eV for the crystallinstate219 and an activation energy of only 1.77 eV forthe amorphous state; furthermore, the diffusion oFe in amorphous Al2O3 is greatly enhanced underion irradiation, with the activation energy decreasing


n,d

ce

.x

,-

terct

e,al

s,-

elts

s

e

r

n

f

to 0.69 eV.220 Similar results have been reported forirradiation-enhanced diffusion of an Fe2O3 markerinto crystalline Al2O3 relative to that of amorphousAl 2O3.221–223

Several studies performed on irradiated ceramicshave concluded that ionization-induced diffusion cansignificantly affect microstructural evolution.67,69,91 Thesensitivity of different materials to ionization-induceddiffusion can vary considerably; unfortunately, littlequantitative information is available on fundamentalproperties, such as ionization-induced changes in poindefect migration energies. Recent work67,69,92,224 hasshown that dislocation loop formation is suppressed inhigh-ionization environments for several ion-irradiatedoxide ceramics and that the electronic-to-nuclearstopping power ratio is an important parameter. Consequently, ceramics that exhibit good radiation resistanceduring light-ion or neutron irradiation (high electronic-to-nuclear stopping power ratio) may not be radiationresistant to thea-recoil cascade (low electronic-to-nuclear stopping power ratio). Such behavior hasbeen observed in a silicate apatite that exhibited goodresistance to amorphization undera-particle irradiation,but readily amorphized due to244Cm decay.175,179

In a similar silicate apatite, as illustrated in Fig. 7,simultaneous irradiation with 300 keV electrons during1.5 MeV Xe1 irradiation is observed to decrease therate of amorphization as compared to irradiation with1.5 MeV Xe1 alone225; this effect is due to enhancedrecovery processes associated with ionization-inducediffusion or subthreshold collision events. Additionalexperimental studies are needed on relevant ceramicsorder to determine their sensitivity to both the irradiationspectrum (ion mass and energy) and simultaneouionization. Such studies will allow the assessmentof the relative importance of ionizing radiation on

FIG. 7. Decreased rate of amorphization due to simultaneous irradiation with 300 keV electrons.225

, No. 6, Jun 1998 1451


g

h

a

z

i

tdl

r

n

i-ofn

lts.nin

8

r

ve,

esse

ble

dl

trts

-

both amorphization and microstructural evolution awill also provide valuable information on the relativefficiency of displacement damage produced by liand heavy ions. Finally, although ionization-inducdiffusion generally improves radiation resistanceenhancing point defect recombination, it is also possthat the enhanced diffusion might lead to segregationsolute atoms and possible phase decomposition. Wthe exception of colloid formation,68,202–204 there hasbeen relatively little attention given to radiation-inducsolute segregation in ceramic materials.

3. Defect kinetics

The kinetics of defect accumulation and microstrutural evolution under irradiation are controlled by tdamage rate and rate of recovery processes asated with point defect migration. With the exceptioof some of the fluorite structures, such as UO2

51–54 andPuO2,52,226studies of point defect annealing kinetics hanot been carried out on relevant phases of interestCa2Nd8(SiO4)6O2 irradiated with 3 MeV Ar1 ions,179

the damaged crystalline state exhibited an exotherdefect recovery stage at 350±C with an activation energyof 1.3 6 0.1 eV; however, the nature of the defectrecovery process was not determined. Activation engies for defect migration in other relevant materialsunavailable. It is important to determine the migratipaths in order to assess the effect of defect migraon recover processes. Simple defect kinetics (interstand vacancy recombination/migration kinetics) caninvestigated by annealing studies of materials containsignificant defect concentrations introduced by irradtion, as, for example, in previous studies of UO2,51,52

PuO2,52 CeO2,52 SrF2,227 and BaF2.227 The effects ofdefect kinetics on the amorphization process candetermined by annealing studies of partially amorphimaterials.

C. Volume changes

Radiation-induced volume changes in crystallphases can result from the accumulation of podefects in the crystalline structure, solid-state phtransformations (e.g., amorphization), and the evoluof microstructural defects (e.g., gas bubbles, voids,locations, and microcracks). The macroscopic swelis the overall sum of these effects. In general,main concern with volume changes, particularly ladifferential changes, is that it may lead to high internstresses that could cause microcracking, segregationincreased dissolution rates. In addition, the large voluchanges associated with amorphization significaaffect atomic bonding, local coordination, and tpathways for ion exchange, all of which can impathe release rates of radionuclides.


ndeht

edbyible

ofith

ed

c-e

soci-n

ve. In

mic

orer-re

ontionitialbeingia-

beed

neint

aseionis-

ingthegeal

, andmetly

hect

1. Unit-cell volume

In many crystalline ceramics, barriers to recombnation and aggregation can lead to the accumulationsignificant point defect concentrations. The accumulatioof these point defects in the crystalline structure resuin an increase in unit-cell volume with cumulative doseThe magnitude of the unit-cell volume expansion ofteis dependent on the irradiation species due to changesthe primary recoil-energy spectra, as illustrated in Fig.for polycrystalline Ca2Nd8(SiO4)6O2 under irradiationconditions (a-particles and 3 MeV Ar1) where amor-phization does not occur179 and where it does (244Cm-decay).175 The higher displacement doses required foexpansion under 3 MeV Ar1 irradiation illustrates thedecrease of the in-cascade survival of defects relatito a-particle irradiation. For the case of Cm-decaythe in-cascade amorphization process results in excinterstitials (ejected from the cascade) in the crystallinmatrix relative toa-particle irradiation. Similar recoilspectra effects have also been observed for the stastructures UO2,50,52 PuO2,52 and BaAl2Ti6O16.111 Ingeneral, the unit-cell volume expansion,DVucyV0,follows the characteristic exponential behavior predicteby models50,228for the accumulation of isolated structuradefects and is given by the expression:

DVucyV0 Aucf1 2 exps2BucDdg , (2)

whereAuc is the relative unit-cell volume expansion asaturation,Buc is the rate constant (per unit dose) fothe simultaneous recombination of structural defecduring irradiation, andD is the dose. Unit-cell expan-sions due toa-particle irradiation have been determined for a number of ceramics, including barium

FIG. 8. Unit-cell volume expansion in polycrystalline Ca2Nd8-(SiO4)6O2 due to244Cm-decay,175 where amorphization occurs, andirradiated with a-particles (emitted from a238PuO2 source) and3 MeV Ar1 ions,179 where amorphization does not occur.

13, No. 6, Jun 1998


,

st-

ea

i

o

e

ic

hollandite,111 a Synroc phase subject only toa-par-ticle (and byg) irradiation. As already noted abovea-particle irradiation of barium hollandite results ina large unit-cell expansion (up to 2.5%) and a diplacive transformation from the tetragonal structurea lower symmetry monoclinic structure. Unit-cell expansions due toa-decay have also been determinefor actinide dioxides,228,229 actinide-doped ceramics37

(e.g., zircon,155,156 apatite,162,163,175pyrochlore,112,116 andzirconolite115,116), and natural minerals.27,37,160 Some ofthese unit-cell expansions due toa-decay are sum-marized in Fig. 9. The unit-cell volume expansion pa-decay event for Pu-doped zircon is much larger thfor natural zircon (Fig. 9), which indicates that annealinof point defects in natural zircon occurs under geologconditions, as previously suggested.155,156 Holland andGottfried160 also noted evidence for defect annealingnatural zircons. In general, the unit-cell expansionisotropic in cubic structures; while in noncubic structures, the unit-cell expansion is anisotropic and has beshown to be consistent with topological constraintsthe changes in the polyhedral connectivity within thstructure.179

2. Macroscopic swelling

The macroscopic swelling is usually determineexperimentally by the relative change in macroscopdensity, Dryr0. If the density changes are small, thswelling can be reasonably approximated as2Dryr0.However, in cases where the changes in density and cresponding swelling are large, as can occur in ceramthe actual relationship betweenDVmyV0 andDryr0 thatis given by the expression

DVmyV0 2sDryr0dys1 1 Dryr0d (3)

FIG. 9. Unit-cell volume expansions due toa-decay in Ca2-(Nd, Cm)8(SiO4)6O2,162,163,175 (Zr, Pu)SiO4,155,156 natural zircon,160

and PuO2.229


-o

d

rn

gic

nis-enne

dic

or-s,

should be used.155,163

The available data on macroscopic swelling due toa-decay in ceramics are extensive37 and include resultsfrom PuyCm-containing ceramics for doses up to3 3

1019 a-decaysyg and from mineral analogues for muchhigher doses (up to4 3 1020 a-decaysyg). The macro-scopic swelling in these ceramics and minerals increaseswith dose and at saturation ranges from about 1% tonearly 20%, as summarized in Fig. 10. In general, stablestructures, such as the actinide dioxides, exhibit onlyabout 1% swelling, while larger values of swelling areassociated with radiation-induced amorphization. Satu-ration swelling values reported for some of the ceramicphases of interest are summarized in Table VI, which in-cludes the value for neutron-irradiated quartz determinedfrom reported density changes.230,231These large volumeexpansions raise important questions: (i) what is thephysical basis for the large volume change in ceramics(e.g., quartz and zircon) due to amorphization, and (ii)can the magnitude of the volume change be predictedor modeled, particularly as a function of temperature.Answers will require detailed structural analysis of theamorphous states of these ceramics and application ofcomputer simulation methods, as discussed in Secs. V. Eand VI. C, respectively.

If amorphization does not occur and there are nosignificant contributions to macroscopic swelling fromextended defects, such as dislocation loops/networks,voids, bubbles, or microcracks, then the macroscopicswelling is dominated by the unit-cell expansion inducedby the accumulation of point defects. In many of therelevant phases of interest, however, amorphization doesoccur, and the formation of helium bubbles at very high

FIG. 10. Summary of extensive data37 on macroscopic swelling be-havior in PuyCm-containing ceramics and natural minerals.

3, No. 6, Jun 1998 1453


-o

e

r

n

y

d

trd-

-

yed

a-

TABLE VI. Saturation macroscopic swelling observed in severceramic phases.

Phase Swelling References

SiO2 (quartz) 17.5% 230, 231ZrSiO4 (natural) 18.4 155, 160, 163(Zr, Pu)SiO4 16.6 155, 163Ca2Nd8(SiO4)6O2 9.4 163CaPuTi2O7 5.5 112, 113Ca(Zr, Pu, Cm)Ti2O7 6.0–7.1 115, 116(Gd, Cm)2Ti2O7 5.1–6.5 116, 133PuO2 0.97 52, 228, 229

doses is possible. Under these conditions, the unit-cexpansions, amorphous regions, and extended defectcontribute to macroscopic swelling. Due to the compoite nature of the radiation-induced microstructure, thtotal macroscopic swelling,DVmyV0, in these materialscan be expressed as:

DVmyV0 fcDVucyV0 1 faDVayV0 1 Fex , (4)

wherefc is the mass fraction of crystalline phase,fa isthe mass fraction of amorphous phase,DVucyV0 is theunit-cell volume change of the crystalline phase,DVayV0

is the volume change associated with the amorphizstate, andFex is the volume fraction of extended microstructures (e.g., bubbles, voids, microcracks). In mstudies ofa-decay effects,Fex is negligible, and Eq. (4)reduces to that proposed by Weber.163 As noted above,the unit-cell volume changes,DVucyV0, have been meas-ured for a number of relevant materials. In most casit can be inferred163 that the volume change associatewith amorphization,DVayV0, is constant for a givencomposition and structure (at any given temperatuand is equal to the saturation value of the macroscoswelling for the amorphized structure (Table VI).

If Fex is negligible and the amorphous fractiofa, is known, thenfc 1 2 fa, and the total macro-scopic swelling, DVmyV0, can be calculated, as habeen done for apatite, Pu-zircon, and natural zircon.163

The calculated total macroscopic swelling is shownFig. 11 for Pu-zircon, along with the experimental dafor macroscopic swelling and the calculated crystallinfc DVucyV0, and amorphous,fa DVayV0, components.The calculated macroscopic swelling, based on Eq. (provides an excellent fit to the experimental data, athe dominant contribution of the amorphization proceto macroscopic swelling at high doses is clearly evide

3. Temperature dependence

Of the candidate ceramic waste materials, the teperature dependence of volume changes accompana-decay has been measured only in238Pu-containingCaPuTi2O7 (pyrochlore structure).113,114 In this material,storage at ambient temperature (approximately 350


al

ells alls-e

ed

st

s,d

e)pic

s

intae,

4),ndssnt.

m-ing

K,

FIG. 11. Experimental data for macroscopic swelling in Pu-zircon,along with the calculated crystalline and amorphous components anthe total swelling based on Eq. (4).84,163

due to self-heating) resulted in macroscopic swellingthat saturated at a value of approximately 5.5 vol %after a cumulative dose of,5 3 1018 a-decaysyg (for200 days of storage time). Storage of this material a575 K resulted in less swelling and required a longetime (300 days) to reach saturation due to the increaserate of simultaneous damage recovery. When this material was held at 875 K, swelling was minimal. Themacroscopic saturation swelling as a function of temperature in this material is summarized in Fig. 12.

D. Stored energy

The defects and structural changes introduced bradiation increase the energy of materials above thminimum energy configuration. Since these defects an

FIG. 12. Temperature dependence of macroscopic swelling at saturtion in CaPuTi2O7.113,114

3, No. 6, Jun 1998


g.eh

ei

iai

n

ertnm

g

cs

ofde-

sdII.ofurer

ither,ted

ofose

y

es

structural changes are metastable, radiation-damamaterials tend to react to reduce their energy contentthe amount of stored energy is large and rapidly releassignificant temperature excursions may occur, as in tcase of neutron-irradiated graphite232 where the storedenergy can exceed 2400 Jyg and cause temperatureincreases of 1000±C. Consequently, an important issufor waste forms is the amount of energy storedthe structure due to radiation damage. Temperatuexcursions due to the release of stored energy caffect the physical and chemical properties of ceramwaste phases. In addition, the stored energy providesexcellent measure of the relative damage accumulatand stability of the radiation-damaged state. In generboth the total stored energy due to accumulated radiatdamage and the distribution of the stored energy releaas a function of temperature are of interest. Measuristored energy release is also useful in studying damarecovery, since prominent recovery processes produpeaks in the stored energy release spectrum.

Sources of stored energy in radiation-damagceramics are: (1) point defects in the crystal structu(2) the significant atomic disorder associated wiamorphization, and (3) the high strains from formatioof amorphous domains. The maximum contribution frostrain to the total stored energy has been estimatedthe case of CaPuTi2O7 to be 18%.233 Although onlylimited data113,234,236exist on the accumulation of storedenergy as a function of dose in the ceramic phasesinterest, it is generally observed that the stored enerincreases with dose until the fully amorphous statereached, as shown in Fig. 13 for natural zircons.236

The accumulated stored energy in zircon occurs mufaster than the rate of amorphization, which suggea significant contribution to the stored energy fromdefect accumulation in the crystalline structure. Th

FIG. 13. Accumulation of stored energy in natural zircon.236


edIfd,e

nreanicanonl,

onseggece

de,h

in

ofy

is

hts

e

total stored energy released during recrystallizationthe amorphous state has been measured for (i) actinidoped apatite176 and zirconolite/pyrochlore116,234,235

phases; (ii) natural metamict zircon,236 zirconolite,237 andpyrochlores238; and (iii) ion-beam amorphized phasesuch as monazite.239 The total stored energy measurefor each of these materials is summarized in Table VThe stored energy released during recrystallizationthe amorphous state occurs over a narrow temperatrange in these materials, as shown in Fig. 14 foCa2(Nd, Cm)8(SiO4)6O2.176 The release of stored energydue to defect recovery in Ca2Nd8(SiO4)6O2 occursin a lower temperature peak.179 In most cases, thelargest contribution to stored energy is associated wthe atomic disorder of the amorphous state; howevcomparable energy may be stored by the accumulapoint defects in the crystalline structure.

In a study of CaPuTi2O7, differential scanningcalorimetry measurements showed that the amountstored energy released increased with increasing dand peaked at 100 Jyg after a dose of5.5 3 1018

a-decaysyg.234 Beyond this dose, the stored energdecreased exponentially to a value of 50 Jyg after a

TABLE VII. Total stored energy released in several ceramic phasduring recrystallization of amorphous state.

Material Stored energy (Jyg) Reference

Ca2(Nd, Cm)8(SiO4)6O2 130 176Ca(Zr, Cm)Ti2O7 127 116CaPuTi2O7 100 234, 235CePO4 (ion-irradiated) 31 239Natural ZrSiO4 318 236Natural CaZrTi2O7 48 237Natural pyrochlores 125–210 238

FIG. 14. Stored energy release in amorphized Ca2(Nd, Cm)8-(SiO4)6O2 during heating at 20±Cymin.176

3, No. 6, Jun 1998 1455


na

es

V

e

tt

h

ls

ce

e

ooe

g

ru

e

o

a

s

i-

o-

ic

.

s-

S

s;

lts

dose of3.6 3 1019 a-decaysyg. These data suggest thacontinued self-radiation damage in this material beyothe dose for complete amorphization produces a damstate of lower energy. Based on these results, Coghand Clinard240 estimated that the local stored energproduced in a single (nonoverlapped)a-recoil cascade ison the order of 200 Jyg in CaPuTi2O7. A peak in internalstrain at intermediate doses is a common observationirradiated ceramics. This effect may explain the lowstored energy (48 Jyg) measured in natural zirconolitedamaged to high doses (2 3 1019 a-decaysyg).237

E. Amorphization

Self-radiation damage froma-decay events andion-irradiation damage results in a crystalline-toamorphous transformation in most of the crystallinceramics of interest as actinide-host phases (TableOnly the fluorite-related structures (ZrO2, PuO2 andUO2) appear to be not susceptible to radiation-inducamorphization (at least for temperatures above 20for ZrO2 and 5 K for UO2); however, they may besusceptible to polygonization.241 Although susceptibleto ion-irradiation-induced amorphization,86,88 naturalmonazite88,165,187,188 and apatite173,174 minerals aregenerally found in highly crystalline states despiexperiencing higha-decay event doses, which suggesthat for these two minerals the rate of damagrecovery may exceed the damage-production ratethe geologic environment. This is consistent witthe low critical temperatures for ion-beam-induceamorphization86,88,89,189and low temperatures for thermarecrystallization180,190,239 measured for these phaseSilicate apatites, zircon, zirconolite, pyrochlores, anperovskite have all been shown to be susceptiblea-decay induced amorphization. Although the exanature of the amorphization process is not well definfor most of these phases, amorphization appearsoccur heterogeneously in these phases by sevpossible mechanisms: (i) directly in thea-recoildisplacement cascade, (ii) from the local accumulatiof high defect concentrations due to the overlapa-recoil cascades, or (iii) by composite or collectivphenomena involving more than one process. Thradiation-induced amorphization leads to decreasesatomic density and changes in local and long-ranstructure, which can have a profound effect on thperformance and properties of the materials. In geneamorphization in these ceramic phases is more difficwith increasing temperature and occurs only belowcritical temperature; the amorphized state, once formis stable under further irradiation. In this section, thcurrent state of knowledge regarding amorphizatiof ceramic phases, structure of the amorphous staamorphization processes, and amorphization kinetics


tdge

lany

inr

-e).

dK

esein

d

.dtotdtoral

nf

isineeal,ltad,ente,re

reviewed, and fundamental scientific issues regardingamorphization are identified.

1. Structure of the amorphous state

Radiation-induced amorphization of crystalline com-pounds leads to aperiodic structures with large changein volume. In the geological literature, such aperiodicstructures, which form in (U, Th)-containing minerals asa result ofa-decay, are referred to as metamict.242–244

These radiation-induced amorphous states lack both orentational and long-range translational correlations andare, therefore, more properly referred to astopologicallydisordered; however, in this paper, the simpler termamorphous is used synonymously with topologicallydisordered.

Although various techniques have been used toelicit structural information, a coherent overall descrip-tion of such amorphous structures, integrating the in-dividual pieces of information available, has yet toemerge. Consequently, the large volume changes assciated with amorphization cannot be explained quantita-tively. It is possible, however, to obtain information onstructure from a variety of techniques, including (i) high-resolution electron microscopy and the allied techniqueof spatially resolved energy-loss spectroscopy; (ii) x-rayabsorption spectroscopies (XAS); (iii) x-ray, electron,and neutron diffraction; (iv) measurement of changesin macroscopic parameters such as density and elastmoduli; (v) vibrational spectroscopies, such as FTIR andRaman methods; and (vi) liquid-phase chromatographyIt is important to stress, however, that for each phasethere may be nouniqueamorphous structure; the finalatomic structure will be dependent on initial structureand chemistry, temperature, and irradiation parameter(e.g., dose rate and ion mass/energy). Computer simulation techniques, as described in Sec. VI. C, can bepowerful tools in modeling the amorphous structures andinterpreting experimental structure data.

There have been few studies of the structure of theamorphous state in these ceramic phases. Based on XAstudies of partially and fully metamict minerals,27,245

fully metamict zirconolite,237,246,247and fully amorphizedCaPuTi2O7,248 a reasonable model of the structure of theamorphous state is one that entails37 (i) a slight distortionof the nearest-neighbor coordination polyhedra; (ii) adecrease in the coordination number and bond lengthin the primary coordination polyhedra; (iii) a slightincrease in the mean second nearest neighbor distanceand (iv) a loss of second-nearest-neighbor periodicity,probably due to the rotation of coordination polyhedraacross shared corners and shared-edges. Recent resuindicate that in metamict zirconolite Ti41 is predomi-nantly fivefold-coordinated,247 as compared to mainly6-coordinated in highly crystalline zirconolite. It has also

3, No. 6, Jun 1998


i

t

ib

e

aleoeon

e

,

o

hle

e

u

h-

e

e

eennos

re

r-he

eis

ly-e

nhe

ti-sgesep-an

ol

ic

p-non-s,es,

nre

o-u-

.g.,s)

llytonce

been reported183 that amorphization in natural sphenesaccompanied by the reduction of Fe31 to Fe21. Valencechanges of this type can have significant effects onsolubility of radionuclides in the aperiodic structure.

2. Susceptibility of amorphization

Several models64,252,249,250 describe amorphizationsusceptibility in ceramics and may be useful for comparing different ceramics. As already noted, howeveall the ceramic nuclear waste phases of interest, wthe exception of the fluorite structures, are susceptito irradiation-induced amorphization. Most models osusceptibility to amorphization ignore kinetics and artherefore, only applicable at very low temperaturewhere point defects are immobile; also some modelsbest suited for electron irradiation conditions (i.e., simppoint defect generation without complications associatwith cascade quenching). Some of the principlesmaterial susceptibility to amorphization are discussbelow. The more important aspects of amorphizatiprocesses and kinetics, which control the rate atemperature range of amorphization, are discussedthe following sections.

Different crystalline structures exhibit a range osusceptibility to amorphization. The variation in susceptibility can be explained, at least in broad outlinby structural connectivity, which is a measure of thtopological freedom251,252 available to each structureand the corresponding redundancy of structural costraints that must be destroyed upon amorphizatioStructures such as MgO, comprising edge-sharingtahedra, are seriously overconstrained and difficultamorphize, although they may polygonize.253 Corner-connected tetrahedral structures, however, like SiO2,silicates, and phosphates are only marginally constrainand amorphize easily. One pertinent observation is tcomplex oxide structures that amorphize easily are amore difficult to recrystallize and hence exhibit a highcritical temperature for amorphization.254 Another obser-vation is the apparent anomaly of SiC, which amorphizalmost as readily as SiO2 (two [SiO4] tetrahedra pervertex) yet is considerably more highly connected (fo[SiC4] tetrahedra connected at each vertex) than evSi3N4 (three [SiN4] tetrahedra per vertex), which isalmost unamorphizable. One explanation is that antisdisorder, which is more possible with SiC than witSi3N4, is required for amorphization of SiC; randomization of site populations rendersb –SiC topologicallyequivalent to Si, which has the same structural fredom as SiO2 and amorphizes with equivalent easThere is evidence for the existence of antisite defectsSiC255 and for radiation-induced antisite disorder frommolecular dynamics simulations,256,257 which suggestthat antisite defects may be the most common def


s

he

-r,thlef,sreedfdndin

f-,

e

n-n.c-to

edatsor

s

ren

ite

e-.in

ct

produced in the cascade; such antisite disorder has bimplicated in electron irradiation-induced amorphizatioof SiC.258–261 Short-range chemical disorder may alsplay a role in amorphization of ionic crystalline phasewith otherwise ordered multiple cations.143,262,263In ad-dition, compositional changes within a given structutype can affect the susceptibility to amorphization.189

Amorphization of ceramic phases induced by enegetic heavy-ion irradiation can also be compared to tprocess of glass formation.264,265 The main mechanismof radiation-induced amorphization is proposed to bfast quenching of a “molten” cascade. The cascadevisualized as a thermal spike, in which atoms are highmobile. The susceptibility of a phase to irradiationinduced amorphization is then inversely related to thcrystallization tendency during the cooling of the moltecascade. In an irradiation study of the phases in tMgO–Al2O3 –SiO2 system,266 the activation energy ofannealing during irradiation was correlated to the acvation energy of viscous flow. Both activation energieshowed the same trend in response to systematic chanin chemical composition. The viscosity of the melt at thliquidus temperature is inversely related to the suscetibility to amorphization. These correlations suggestparallel between ion irradiation-induced amorphizatioand the glass formation process.266–268

In order to describe the tendency of a melt tcrystallize, Wanget al.269 have proposed an empiricaparameter, S, which allows an evaluation of thesusceptibility of a melt to glass formation and a ceramto irradiation-induced amorphization. The calculatedSparameter gives a good relative ranking of the suscetibility to radiation-induced amorphization of phases ithe MgO–Al2O3 –SiO2 system. This parameter has alsbeen successfully applied to the description of radiatioinduced amorphization in other complex ceramicsuch as monazite-structure types, zircon-structure typperovskites, calcium aluminates, and micas.

3. Amorphization processes

Four different types of amorphous accumulatiobehavior have been observed in materials that asusceptible to amorphization: (i) progressive (homgeneous) amorphization due to point defect accumlation (e.g., quartz,206 coesite,270 silicon carbide,271,272

intermetallics273,274), (ii) interface-controlled amor-phization (e.g., intermetallics275 and ceramics276), (iii)cascade-overlap (heterogeneous) amorphization (ezircon84,155,162,163), and (iv) in-cascade (heterogeneouamorphization (e.g., apatite162,163). These differentamorphization processes are illustrated schematicain Fig. 15. Curve “A” represents amorphization thaoccurs nearly spontaneously after a critical concentratiof defects has accumulated. The linear dependen

3, No. 6, Jun 1998 1457


cd

y

i

meool

l

h

el

,

,

,

FIG. 15. Amorphous fraction as a function of dose for several proesses: (A) defect accumulation, (B) interface control, (C) cascaoverlap, (D) direct impact, and (E) direct impact, with cascade twicas large as in (D).

of the amorphous fraction on dose illustrated bcurve “B” indicates an interface-controlled processas in amorphization induced by ballistic mixing inintermetallics.275,277 When amorphization occurs bycascade overlap, then a curve of type “C” is obtained.278

The curvature near the end of the irradiation timecaused by the decreasing probability of cascade impaon the remanent crystalline regions,279 and the curvatureat the beginning is caused by the necessity to creadamaged zones so that cascade superposition becomore probable. The curves “D” and “E” represent thcases of direct impact amorphization in a cascade fdifferent cascade sizes. Therefore, if the crystallineamorphous fraction can be determined experimentaas a function of cumulative dose for different types oirradiation (e.g., by careful measure of the intensity odiffracted electrons or x-rays), then it may be possibto discern the amorphization mechanism.

For amorphization directly in a displacement cascade, the rapid quenching of a melt-like region may leato the amorphous state. For cascade overlap and defaccumulation processes, a critical damage level maynecessary to trigger amorphization, heterogeneouslyhomogeneously. If the difference in free energy betweethe crystal and the amorphous state is associated witcritical level of damage, then the change in free energdetermines phase stability,273,280,281 and amorphizationoccurs when

DGirr . DGca , (5)

where DGca is the difference in free energy betweenthe crystalline and amorphous phases andDGirr is thedifference in free energy (local or global) between thirradiated and unirradiated solid due to all the possib


-e

e

,

sct

tees

rr

lyffe

-dectbeorna

y

e

mechanisms of energy storage in the solid (e.g., pointdefects, chemical disorder, defect clusters). Another cri-terion for the critical damage level is the modifiedLindemann criterion,282 which states that amorphizationoccurs when

kx2l . kx2lcrit , (6)

where kx2l is the mean square static displacement ofthe atoms in the solid from their equilibrium latticepositions andkx2lcrit is the critical mean square displace-ment at melting. According to this model, amorphizationcan be seen as solid-state melting, in which a highdefect concentration softens the lattice and reduces themelting temperature. The advantage of this approach isthat it reduces the different damage contributions (e.g.,point defects, chemical disorder) to a single parame-ter. Molecular dynamics simulations of intermetallicsshow that amorphization always occurs at a fixed valueof kx2l, whether the origin of the increase in mean-square displacements is point defects, antisite defectsor thermal motion.275 This type of behavior has beenobserved experimentally in coesite.270 For amorphizationprocesses where a critical damage level must be reachedthe accumulation of damage and defect concentrationscan be modeled by rate theory,214,215 as discussed above(Sec. V. B. 1).

Homogeneous amorphization is an importantprocess for intermetallics and some ceramics; howeverin nearly all the ceramics of interest as actinide hostphases, amorphization due toa-decay and heavy-ionirradiation is known to occur heterogeneously.Thishas been confirmed in apatite, zircon, pyrochlore,and zirconolite structure types by electron microscopyand diffraction studies that characterized the amor-phization process as a function of increasing dose inactinide-containing ceramics of interest,36,113–116,133,178

in their mineral analogues,84,140,161 and under heavy-ion irradiation.84–87,254 The heterogeneous nature of theamorphization process in these phases is illustrated inthe high-resolution electron microscopy (HREM) imagesin Fig. 16 for ion-irradiated apatite.86 In general, theamorphization process due toa-decay in these materialscan be described as follows: (i) at low dose levels(,1018 a-decaysyg), the materials generally exhibit astrongly crystalline electron diffraction pattern, clearlyresolvable individual damaged regions (observableby their strain contrast), and nearly perfect latticeimages under higher resolution; (ii) at increasing doselevels (.1 3 1018 a-decaysyg), the density of damagedregions increases to the point that individual regions areno longer readily distinguished, a radial ring of diffuseintensity associated with the presence of amorphousmaterial is observed in the electron diffraction pattern,and amorphous domains that exhibit mottled contrastare clearly observable in high-resolution images; and

3, No. 6, Jun 1998


a.
FIG. 16. HREM images of Ca2La8(SiO4)6O2 irradiated with 1.5 MeV Kr1 ions: (a) unirradiated, (b) 0.03 dpa, (c) 0.06 dpa, and (d) 0.24 dp86
y.

ds

/

d

o

tt

sh-d,g,s.

(iii) at higher doses (on the order of 5 to10 3 1018

a-decaysyg, depending on the phase), the materiabecomes fully amorphous with no evidence of anresidual crystallinity in the electron diffraction patterns

Although amorphization occurs predominantly byheterogeneous processes in these ceramic phases,dominant process (direct-in-cascade or overlap) has nbeen identified, except in two cases. In244Cm-dopedCa2(Nd)8(SiO4)6O2,162,163 it has been shown that amor-phization primarily occurs directly in the cascades othe recoil nuclei emitted duringa-decay of Cm. For thismaterial, the increase in amorphous fraction,fa, withdose,D, generally follows the direct impact model278 foramorphization that is given by the following expression

fa 1 2 exps2BaDd , (7)

whereBa is the amount of amorphous material producepera-recoil andD is the dose. The change in amorphoufraction with dose for Ca2Nd8(SiO4)6O2 is shown inFig. 17. At the higher doses,fa for Ca2Nd8(SiO4)6O2

deviates somewhat from the behavior predicted by thdirect impact model, suggesting that additional mechanisms (e.g., defect accumulation processes, cascaoverlap processes, strain-induced amorphization, andinstability of crystalline “islands” below a critical size)may be contributing in a collective manner to the overall amorphization process. Amorphization in Pu-dopeand natural zircons has been shown84,155,161–163 to beconsistent with the double cascade-overlap model famorphization,278 which is given by the expression:

fa 1 2 fs1 1 BdD 1 Bd2D2y2d exps2BdDdg , (8)


l

theot

f

:

e-deor

-

r

where Bd is the amount of material damaged pera-recoil. The amorphous fractions for Pu-doped andnatural zircons (Fig. 17) exhibit very similar behaviorsas functions of cumulative displacement dose, excepfor an offset at low dose in natural zircons due to defecrecovery processes over geologic time.155,163

The above results and models represent idealizationof the amorphization process as caused by single mecanisms, without any consideration for temperature anannealing effects. Under actual irradiation conditionsthere may be a mixture of different processes causinamorphization, such as a combination of direct-impactcascade-overlap, and defect accumulation processe

FIG. 17. Amorphous fraction as a function of dose Ca2(Nd, Cm)8-(SiO4)6O2, (Zr, Pu)SiO4, and natural zircon.163

, No. 6, Jun 1998 1459


st

li

e

snsf

ie

nieaciar

e

c

o

a

nt

n

a

t

n

oin

re

dn

ly

ral

on

There are also the influences of temperature aannealing, which can affect the amorphization procesand relative kinetics in a complex manner. Consequenthe dominant amorphization process may change wtemperature. The effects of temperature and anneaon amorphization processes are discussed in the nsection.

4. Amorphization kinetics

During irradiation, the production of radiationdamage is mitigated by simultaneous damage recovprocesses, and the rate of damage evolution (e.g., amphization) will depend on the relative magnitude of thetwo processes under the irradiation conditions at agiven time. For irradiation-induced amorphization, therecovery processes may be associated with point demigration and annealing in the crystalline state, defeand ion migration in the amorphous state, or epitaxrecrystallization at the crystalline/amorphous boundariObviously, if the rate of amorphization is less than thrate of the recovery processes, amorphization canproceed. Since most available data on amorphizatkinetics in relevant ceramics have been obtainusing greatly accelerated ion-irradiation techniques,important goal of understanding amorphization kinetifor nuclear waste ceramics is to be able to predwhether or not amorphization will occur under the acturadiation conditions (damage rate and temperatuexpected in a geologic repository, and if amorphizatiooccurs, what the rate of amorphization will be undthe repository conditions. Such understanding cangained only through detailed studies of the dependenof amorphization on time, temperature, dose rate, aion mass/energy.

a. Summary of available data.There have been onlylimited studies on the kinetics of amorphization in ceramic nuclear waste phases. As noted above, amphization in natural zircons has been observed to ocat slightly higher doses than those measured for Pcontaining and ion-irradiated zircons.84 This behaviorhas been determined to be due to defect recovery presses over geologic time.155,163Similar recovery behaviorover geologic time periods is postulated for naturzirconolite, since the dose required for complete amophization in natural zirconolite is six times higher thathat measured for both actinide-doped and ion-irradiazirconolites.283

The detailed temperature dependence ofa-decay-induced amorphization has not been studied in amaterials; however, there is some limited informatioavailable from a study of the temperature dependenof a-decay damage in CaPuTi2O7.113,114 In this study,the critical dose for complete amorphization is appaently slightly higher at 575 K than at 350 K, and at


ndesly,ithngext

ryor-eyeectctals.eot

ondn

sctl

e)nr

becend

-or-uru-

c-

lr-

ed

ynce

r-

temperature of 875 K, no loss of crystallinity is evideneven at1 3 1019 a-decaysyg. Consequently, at the doserate in this material (1029 dpays), the critical temperatureabove which amorphization does not occur is betwee575 and 875 K.

More recently, ion irradiation has been used tstudy the temperature dependence of amorphizationa number of relevant crystalline phases,84–89 as illus-trated in Fig. 18 for several phases and in Fig. 19 fodifferent compositions within the same zircon structurtype.88,164,189 A detailed study84 of radiation-inducedamorphization in zircon at room temperature showethat the amorphization dose for heavy-ion irradiatio(1024 to 1023 dpays) is nearly identical to that froma-decay in238Pu-doped zircon (3 3 1029 dpays). Thus,at 300 K, the amorphization process in zircon is near

FIG. 18. The temperature dependence of amorphization in sevepotential actinide host phases under 1.5 MeV Kr1 irradiation.84–86,89

FIG. 19. The temperature dependence of amorphization in the zircstructure type.88,164,189

3, No. 6, Jun 1998


eoc

f

as

noe

e

ua

so

t

nv-

-e,s

lethela-e

tdes

theinate

ese

sle

)]

seel

onrylserr

as

rseron-icn,ithionese

tedndis-sed

independent of the damage source and damage rate,high-resolution electron microscopy results84 confirmthat the damage ingrowth processes for ion-irradiatand natural zircons are similar. Since amorphizatiof zircon does not exhibit a significant dependenon temperature (i.e., mobile defects) at 300 K,84 theseresults are consistent with rate theory models that predno dose-rate dependence at temperatures where deare immobile. In apatite,85 the critical dose at 300 Kfor amorphization from heavy ions (5 3 1024 dpays) isslightly higher (40%) than that measured due to244Cmdecay (3 3 1029 dpays); this behavior, however, maybe due to the effects of irradiation spectrum rather thdose-rate effects. At elevated temperatures, the effectdose rate in both zircon and apatite are expected tomore pronounced.

There have been several studies of the influenceion species on amorphization in ceramics.85,271,272,284–286

In general, heavy ions increase the temperature raover which amorphization is possible relative telectron (or very light ion) irradiation. In the casof Ca2La8(SiO4)6O2

85 (Fig. 20) and SiC,271,272,284 thedifference in critical temperature due to particlmass/energy can amount to several hundred±C. Thismay be due to differences in the distributions antypes of defects formed and an ionization effect dto differences in the ratios of the electronic to nuclestopping powers. As noted above (Fig. 6), simultaneoirradiation of Ca2La8(SiO4)6O2 with 300 keV electronsduring 1.5 MeV Xe1 irradiation significantly suppressethe rate of amorphization as compared to irradiatiwith Xe1 alone.225

b. Existing models.Several models have been useto describe the temperature dependence of ion-beainduced amorphization in ceramics and semicondu

FIG. 20. Temperature dependence of amorphization in ion-irradiaCa2La8(SiO4)6O2.85


and

dne

ictects

nof

be

of

ge

der

us

n

dm-c-

ed

tors.84,85,287–289In general, the effects of temperature oamorphization can be described by simultaneous recoery processes with activation energies,Ea, associatedwith defect migration/recombination in the crystal structure, defect and ion diffusion in the amorphous stator epitaxial recrystallization at the crystalline/amorphouinterface. In many of the existing models, rather simprelationships are derived between temperature anddose required for complete amorphization. These retionships are generally given by an expression of thform:

lnf1 2 sD0yDd1/mg C 2 EaynkT , (9)

where D0 is the dose for complete amorphization a0 K, C is a constant dependent on ion flux and cascasize, and bothm and n are model dependent constant[Morehead and Crowder287 derived m 2 (or 3) andn 2, Dennis and Hale288 derivedm 2 and n 1,and Weber84,85 derived m 1 and n 1]. Althoughthese models have been adequate in describingobserved temperature dependence of amorphizationboth semiconductors and ceramics, it is recognized ththe kinetic processes controlling amorphization may bmore complex than the single activated process that thmodels represent.

Recently, the approach of single defect kinetichas been used to modify the direct impact and douboverlap models for amorphization [Eqs. (7) and (8of apatite and zircon, respectively.164 Using activationenergies and other parameters from the literature, themodified kinetic expressions have been used to modthe time, temperature, and dose rate of amorphizatiin both apatite and zircon under expected repositoconditions. Although these are relatively simple modethat certainly require further refinement (if data are evavailable), they predict that zircon will amorphize undeconditions typical of a near-surface repository, suchYucca Mountain, while apatite will remain crystallinedue to the kinetics of the recovery processes.

In the field of irradiation-induced amorphization inintermetallic compounds, where amorphization occuhomogeneously by the accumulation of defects, othmodels have been used to describe amorphizatikinetics.273,281,290 Although such homogeneous amorphization is not generally observed in the ceramphases of interest for nuclear waste immobilizatiothese intermetallic approaches may be applicable, wsome modifications, to the heterogeneous amorphizatprocesses that have been observed in most of thceramics. For example, using rate theory,214,215 theconcentration of accumulated defects can be calculaas a function of dose rate, time, and temperature, athe corresponding change in free energy (or static dplacement) determined. Such an approach could be u

3, No. 6, Jun 1998 1461


o

a

ieoe

o

s

o

r

hg

sfd

-s,s

)

gtol

t

s-e

-r

nge

ess

-

to model the amorphization kinetics. In order to appsuch models, however, it will be necessary to determexperimentally the critical valuesDGca or kx2lcrit for therelevant ceramics. It will also be necessary to generreliable experimental data on amorphization kineticsfunctions of different ion species and dose rates.

c. Modeling considerations.Comprehensive modelsof amorphization kinetics are clearly needed and shotake into account damage rate, PKA spectrum,cascade amorphization, and ionization-induced diffusiThe standard approach to modeling evolutionary chanin microstructure controlled by defect kinetics, suchsome amorphization processes, is to use rate theory.creation of vacancies and interstitials that can migrthrough the lattice and either recombine or be absorbedefect sinks, including amorphous domains, providesdriving force controlling the kinetics of amorphizationAs noted, the application of rate theory to ceramis in itself a formidable task, and the results obtainfor one structure type will not necessarily be valid fothers because of the different topological propertiIt is nevertheless necessary to solve the equationspredict amorphization kinetics, as has been donea few cases in intermetallics273,281,290 and anisotropicmaterials.291 The use of kinetic Monte Carlo techniquealso has potential applications in this area; howevit will be necessary to obtain reliable values of thdefect properties from theory, computer simulations,experimental methods.

Other approaches may be required in the caseheterogeneous amorphization. In the case of amorphtion directly in thea-recoil displacement cascade, threcovery processes may be associated with epitarecrystallization at the crystalline-amorphous interfaof the amorphous domains and less dependent onmigration enthalpies of interstitials and vacancies. Fsuch epitaxial recovery processes, there will in genebe no unique activation energy, as it will dependthe size of the amorphous domain. Since heterogeneamorphization is apparently the dominant mechaniinduced bya-decay in nuclear waste ceramics, therea great need for additional studies in this area.

Ionization-induced diffusion67,69,91,92 might inhibitamorphization for both the homogeneous (point dfect accumulation) and heterogeneous (in-cascade) casince both defect and recrystallization kinetics woube enhanced. The effect of ionization-induced diffusiwould be expected to be more pronounced for the homgeneous amorphization process, since large changethe interstitial migration enthalpies may occur (includinpossibly athermal interstitial migration). Data for seveoxides189,225 suggest that ionization-induced diffusiocan retard or inhibit amorphization under high-dosrate electron and ion irradiation conditions where telectronic to nuclear stopping power ratios are lar


lyine

ateas

uldin-n.

gesasThete

d atthe.cs

drs.to

for

ser,eor

ofiza-exialcetheorralnousmis

e-ses,ldno-

s ingalne-ee.

However, ionization-induced diffusion may be less ef-fective in retarding amorphization froma-recoil nucleiunder repository conditions, where both the electronicto nuclear stopping power ratios and the dose rateare significantly smaller. Consequently, the effects odose rate on ionization-induced diffusion processes ankinetics must be investigated in order to interpret ion-irradiation data and predict behavior under repositoryconditions.

Many crystalline ceramics exist in a narrow rangeof stoichiometry, since any departure from exact stoi-chiometry has to be accommodated by interstitials, vacancies, or antisite defects at great energy cost. Thuit is important to understand the influence of departurefrom stoichiometry (caused, for example, by the recoil-induced transfer of atoms from the crystalline phase intothe glassy phase in glass-ceramics or by transmutationon the amorphization susceptibility. In addition, the pres-ence of extended defects, such as dislocations, stackinfaults and antiphase boundaries, have been shownfacilitate amorphization in some materials. Preferentiagrain-boundary amorphization under electron irradiationhas been observed in quartz292 and recently observedin spinel and coesite.293 It is particularly importantto understand the amorphization susceptibility in grainboundaries, as leaching often occurs preferentially athose sites.

5. Recrystallization

a. Epitaxial recrystallization.Solid-state epitaxial re-crystallization at the crystalline/amorphous interface inmany of these ceramic phases could control the kineticof the amorphization process. Annealing of ion-beamamorphized layers, such as has been studied in thperovskite phases, CaTiO3

148,149 and SrTiO3,150,151 canprovide activation energies for solid-state epitaxial re-crystallization at an effectively semi-infinite interface.Similar studies have not been performed for other relevant phases, but such studies would provide an uppelimit to the activation energy for annealing discreteamorphous zones in each phase. Fission track annealiin natural minerals and reactor-irradiated phases or thannealing of tracks produced by high-energy (GeV)heavy ions also can provide information on solid-stateepitaxial recrystallization at finite surfaces if the tracksare amorphous. Unfortunately, in many of the studiesof relevant phases, the nature of the track damag(i.e., whether disordered-crystalline or amorphous) hanot been determined. In zircon, such heavy-ion trackhave been shown to be amorphous,294 and annealingstudies295–297 of such tracks in zircon yield activationenergies for solid-state epitaxial recrystallization. It hasbeen shown that fission track annealing in zircon exhibits anisotropy, and the activation energies for track

13, No. 6, Jun 1998


1

io

e

ar

a

re

t

n

zl

e

-

s

,s

ld

s

t

s

annealing in the (001), (011), and (100) planes are 2.2.87, and 3.60 eV, respectively.297 Heavy-ion track an-nealing in fluoroapatite is reported to have an activatenergy of 0.76 0.1 eV296,297; however, the nature of thedamage track (crystalline or amorphous) has not breported.

b. Recrystallization of fully amorphous state.Recrys-tallization of the fully amorphous state in ceramics isexothermic reaction that results in the release of stoenergy. Isochronal annealing studies of fully amorphospecimens of (Gd, Cm)2Ti2O7,116 Ca(Zr, Cm)Ti2O7,116

Ca2(Nd, Cm)8(SiO4)6O2,163,176(Zr, Pu)SiO4,155,156,163andnatural zircon298 have been performed. The isochron(12 h) recovery results for several materials are sumarized in Fig. 21. For both (Gd, Cm)2Ti2O7 (Fig. 21)and Ca(Zr, Cm)Ti2O7 (not shown), there is a linearecovery of density with temperature prior to singlstage recrystallization,116 which suggests the amorphoustate for these materials has a range of densitiesvary with temperature, consistent with the observatioof Clinard and co-workers.113 There is little recoveryin density prior to the onset of recrystallization iCa2(Nd, Cm)8(SiO4)6O2, (Zr, Pu)SiO4, and natural zircon(not shown), suggesting a narrower range of densitiesthe amorphous state in these two structures. AmorphiCa2(Nd, Cm)8(SiO4)6O2 recrystallizes to the originaapatite structure in a single recovery stage,163,176 whilerecrystallization of amorphized (Zr, Pu)SiO4 occurs intwo stages involving the initial crystallization of sompseudo-cubic ZrO2 nuclei prior to full recrystallizationback to the original zircon structure.155,156,163Recrystal-lization of amorphized Ca2(Nd, Cm)8(SiO4)6O2, whichhas been studied in some detail,176 releases 130 Jyg ofstored energy and occurs with an activation energy3.1 6 0.2 eV. Although rigorous determinations of th

FIG. 21. Recovery of density during recrystallization of Ca2(Nd,Cm)8(SiO4)6O2,163,176 (Gd, Cm)2Ti2O7,116 and (Zr, Pu)SiO4.155,156,163


5,

n

en

nedus

lm-

-shatns

fored

ofe

recrystallization kinetics are not available for otherrelevant ceramics, the activation energies for recrystallization in zirconolite/pyrochlore,116,235,238 zircon,155,156

and monazite239 structures have been estimated.

F. Helium accumulation, trapping,and bubble formation

Helium has a limited solubility in most ceramics. Atincreased concentrations, it may form bubbles, which cancause swelling and affect many of the physical propertiesof the solid. The consequences of high gas concentrationare well known for nuclear fuels and the heavy fissiongases Kr and Xe. Although the effects for the smallerHe atoms may be less pronounced, the effects of Heaccumulation need to be studied and understood. Fewdata exist on the behavior of He in potential actinide-host phases. Consequently, studies of He solubilitydiffusion, trapping, and release in these ceramic phaseare needed. In addition, the evolution of He bubbles,including the temperature and dose dependence, shoube investigated. Such studies of bubble formation willrequire careful characterization by transmission electronmicroscopy and small angle x-ray scattering, such as habeen done recently for a nuclear waste glass.299

In order to understand the potential effects of Heaccumulation, it is instructive to consider the mecha-nisms of a damage effect of great importance for UO2

fuel, which is the grain subdivision or polygonizationof high burnup UO2 fuel (the so-called rim-effect).Recent work310 on ion-implanted UO2 indicates thatthis effect may be gas-driven. The threshold conditionsfor polygonization in UO2 correspond to about 1 wt. %Xe, a displacement dose of 2000 dpa, and,9 cm3

Xe(NTP)ycm3 fuel. Most experiments with actinide-containing ceramics have not exceeded 1 dpa or 0.1 cm3

Heycm3 ceramic. In the case of Pu-containing ceramicwaste forms, however, the He concentration after longstorage times will be much larger (a factor of 100) thanthe threshold Xe concentration for polygonization.

In some natural minerals, most notably natural UO2

(uraninite) or (U, Th)O2 (thorianite) with ages rangingfrom 200 to 500 million years, high damage levels(100 to 200 dpa) and high He concentrations (on theorder of 0.5 wt. % He), corresponding to some 300 cm3

He(NTP)ycm3 mineral, are found. Similar He concentra-tions will only be reached in waste matrices with highactinide loading over long time periods that correspondto many half-lives. Transmission electron microscopystudies of such minerals with high damage levels andhigh He concentrations show evidence for polygoniza-tion, and these minerals “explode” when heated to abou800 ±C.300 This phenomenon is accompanied by a burstrelease of He and extensive bubble formation, whichstrongly supports the need to study the consequence

3, No. 6, Jun 1998 1463


H

r

o

i

hess

ofes,a

erese

s,thehe

,al

-

f

s

thents

he.iong

one

ndis

instt ofsoraliore

s

n

tor

of He buildup in waste ceramics. Bubbles that aassumed to contain He have been observed in high-dmetamict minerals301; however, little is known on theeffect of the He bubbles on materials properties. Itpossible that the crystalline-to-amorphous phase chacould sweep the He out of the matrix. Furthermore,diffusion may be faster in the amorphous state, caing more He release during storage; this could reduthe overall He concentration and associated effectsthe matrix but could possibly cause canister failuValuable exploratory investigations could be completusing natural analogue phases containing high U andconcentrations (i.e., high damage levels).

Several studies140,237 have reported the presencenear-spherical microvoids or bubbles ranging in diamter from 100 to 400 nm in highly damaged naturzirconolites. These bubbles have been attributed toaccumulation of He from thea-decay of U and Thimpurities, and their daughter products. Bubbles haalso been observed in metamict Nb–Ta–Ti oxides,301 aswell as in annealed uraninite, UO21x .302 There are alsoextensive data and models303,304of fission gas bubble for-mation in UO2 that may be relevant to bubble formatiofrom He accumulation in nuclear waste ceramics.

G. Mechanical properties

Ion implantation is widely used to change the mchanical behavior of ceramic surfaces. It is thus nsurprising that similar radiation-induced changes in mchanical properties are observed in actinide-host phadue toa-decay. The effects ofa-decay induced amor-phization on hardness, elastic modulus, and fracttoughness have been investigated as a function of d(up to 5 3 1018 a-decaysyg) in several ceramic phasecontaining238Pu or 244Cm and in some natural mineraanalogues. Additional data have been obtained frstudies utilizing heavy-ion irradiation.

The hardness and elastic moduli of Cm-dopGd2Ti2O7 and CaZrTi2O7 have been shown to decreassteadily with increasinga-decay dose116; a similardecrease in hardness with increasinga-decay dose hasbeen reported in CaPuTi2O7.305 Decreases in hardnesand elastic moduli with increasing dose have also beobserved in natural zircons306–308and in synthetic zirconirradiated with 540 keV Pb ions.309 In general, thedecreases in hardness range from 25 to 50%, anddecreases in elastic moduli are in the range of 5070%. The softening of these materials with increasdose is due to the radiation-induced amorphization tresults in a lower density structure.

The fracture toughness in Cm-doped Gd2Ti2O7,116

CaZrTi2O7,116 and Ca2Nd8(SiO4)6O2,178 as well as inCaPuTi2O7,305 has been shown to increase by up 100with increasinga-decay dose to a broad maximum an


reose

isngee

us-cein

e.edTh

fe-

althe

ve

n

e-ote-ses

ureoseslom

ede

sen

theto

nghat

%d

then decrease slightly as the materials approach tfully amorphous state. The increase in fracture toughneis attributed to the composite (two-phase) naturethe microstructure, which consists of mixed crystallinand amorphous domains. At low to intermediate dosethe microstructure consists of amorphous tracks incrystalline matrix, and this composite microstructurcan inhibit crack propagation and increase the fractutoughness. At high doses where the amorphous phabecomes the dominant matrix with remnant crystallitethe fracture toughness decreases slightly as some ofinternal stresses are relieved. This is supported by tobserved decrease in stored energy in CaPuTi2O7 athigh doses,234 which suggests a relaxation of disorderand by the analysis of strain accumulation in naturpyrochlores.213

Additional helpful information has recently becomeavailable from indentation tests in the high burnup, socalled rim zone of LWR UO2 fuel.310 In this zone,the burnup is very high due to a high production o239Pu by resonance neutron capture in238U, and a grain-subdivision process (polygonization) occurs. A Vickerindentation study along radial directions of the UO2

fuel showed a decrease in hardnessH (by ,30%) andan increase in fracture toughnessKIc (by ,100%) thatfollow the radial burnup profile. It is worth noting thatthe concentration of actinides, and hence thea-decayrate, increases with the increasing burnup as doesrare-gas (Xe and He) content. Since these measuremeare made 2–3 years after removal from the reactor, tcontribution of thea-decay damage may be importantTherefore, these results can be taken as a first indicatof the behavior that might be expected in a Pu-containinphase that has been stored for a long time, has undergpolygonization, and contains helium bubbles.

The net result of the decreases in hardness aelastic modulus and increases in fracture toughnessa reduced brittleness and enhanced resistance agacrack propagation in these ceramic phases as a resula-decay damage. The extensive experience withmany materials of different structures and the similanature of the radiation-induced changes in mechanicproperties strongly suggest that the expected behavfor other potential waste ceramic phases will not bsignificantly different.

H. Thermal properties

The thermal conductivity of actinide compounds iknown to decrease significantly due toa-decay damage,mainly by phonon scattering at point defects. As aexample, the thermal conductivity of239PuO2, whichdoes not become amorphous, decreases by a facof 3 at 60±C within one year due to point defectaccumulation.311 Depending on actinide loading and

13, No. 6, Jun 1998


la

y

oo

ire

o)

aie

b

a

l

g

zo

d

t

t

le

ng

cebellyd

terave

asion

isr-

nt

egense

aruch

nde is

agehis

or

sered

nd

erere,

ofcanhen

rm,achue

dimensions of the ceramic waste form, the originalexisting temperature gradient will increase and msignificantly affect microstructural evolution and durability. Thermal diffusivity (laser flash) and heat capacitmeasurements on both as-produced and self-damaactinide waste forms are, therefore, needed.

I. Chemical durability

The principal mechanism for the release of radinuclides from the emplaced waste forms is by corrosiof the waste forms in the presence of water. There habeen surprisingly few systematic studies designedevaluate the change in leach rate of a crystalline ceramas a function of radiation damage, whether due to ioization or ballistic processes. Proper studies will requcarefully controlled irradiations, over a range of dosand temperatures, combined with systematic studiesboth the corrosion rates and mechanisms as functiof temperature, solution pH, oxidation potential (Ehand flow rate. No satisfactory experiments have ybeen completed; however, increases in leach rates hbeen noted for actinide-doped phases, naturally occurrphases containing U and Th, and in some ion-irradiatceramics, as discussed below. In general, the increain leach rates range from a factor of 10 to 100. It mustnoted that the leach rates of the amorphous (damagstate in the ceramics studied to date are still lower ththose of glass waste forms.

1. Summary of available data

The radiation-induced changes in the leach ratesCm-doped Gd2Ti2O7, CaZrTi2O7, and Ca2Nd8(SiO4)6O2

have been determined from leach tests of both fudamaged (amorphized) specimens and a second sespecimens that had been fully recrystallized to the orinal structure by appropriate annealing.177 The results ofthese studies indicate that radiation-induced amorphition increased the leach rate by a factor of 20–50 fGd2Ti2O7, and by a factor of 10 for both CaZrTi2O7 andCa2Nd8(SiO4)6O2. The leaching of Gd2Ti2O7 appearedto occur incongruently, with the non-TiO6 network ionsbeing selectively leached.

Because of the important implications in determining the UyPb systematics of zircons for geologic agdating, there are data on the increased loss of nuclias a function of increasing degree of damage.312–316 In astudy of a suite of natural zircons with varying contenof uranium and thorium, Ewinget al.313 found thatover a dose range of 1017 to 1019 a-decay eventsyg,there was a one order of magnitude increase inleach rate. For completely metamict zircons (dose>1019

a-decay eventsyg), the increase in leach rate was neartwo orders of magnitude. These results are consistwith the observation that loss of nuclides and increas


yy-

ged

-nvetoic

n-esofns,et

vengd

sese

ed)n

of

lyt ofi-

a-r

-ees

s

he

ynt

ed

solubility are, at least in part, related to the increasilevel of radiation damage in zircon.312,314,316 As earlyas the 1970s, it was recognized that the discordanin UyPb age determinations of zircon crystals couldreduced by mechanically (i.e., abrasion) or chemica(i.e., etching with hydrofluoric acid) removing the altere(metamict) areas of the zircons.317,318These altered areasoften have higher U concentrations and occur in the ouzones of single crystals; thus, these altered zones hexperienced morea-decay event damage.319,320

Results by Saleset al.321 suggest that the leachrate of a synthetic monazite, LaPO4, containing simu-lated waste remains low even after the material hbeen transformed to an amorphous state by irradiatwith 250 keV Bi1 ions. However, work by Eyal andKaufmann322 on natural monazite indicates that therea preferential dissolution of the radionuclide daughteproducts,234U, 230Th, and228Th, by factors of between1.1 and 10 relative to the structurally incorporated pareisotopes 238U and 232Th, respectively. This isotopicfractionation is attributed to radiation damage in thtracks of the recoiling daughter nuclei emitted durina-decay of the parent isotopes. While there have besome concerns regarding the interpretation of theresults,185 the increases in dissolution rates are similto those observed in other actinide-host phases, sas naturally occurring thorianite (ThO2) and uraninite(UO2).323 Leaching experiments of these phases ameasurements of the Th and U isotopes show that thera strongly enhanced release of short-lived228Th relativeto 232Th, but only slight enhancement of long-lived234Uand 230Th relative to238U and 232Th. These results canbe interpreted as due to natural annealing of the damcascade over periods of tens of thousands of years. Tchemical etching technique is a powerful method fprobing the thermal stability ofa-recoil damage tracksin a wide variety of crystalline materials, as well aglasses.324,325 These results are consistent with the ovan order of magnitude increase in leach rate observin heavy-ion irradiated UO2.326 Increased leaching of upto a factor of 10 also has been reported in sphene asphene-ceramics irradiated with 280 keV Bi21 ions.327

2. Mechanisms and needs

Depending on the type of solid, there are a numbof different corrosion mechanisms, and in fact, thmechanism may change as a function of temperatuEh, pH, solution composition, and flow rate. Eachthese parameters, particularly temperature and pH,affect the leach rate by many orders of magnitude. Tmicrostructure of a solid can also affect the corrosiorate. Thus, radiation damage to a ceramic waste foas the results above demonstrate, can affect the lerate as a result of (i) an increase in surface area d

3, No. 6, Jun 1998 1465


oinain

h

unvns

e

inu

eoibnanaer

n

oocsnou

eie

-

-

-iau

y

n

nh

,

,

,

r-

-

d

,

d

to microfractures created because of differentialanisotropic radiation-induced volume expansions, (the reduced thermodynamic stability of the radiatioinduced aperiodic domains, (iii) the higher chemicreactivity due to strain fields around amorphous doma(e.g., an a-recoil track), (iv) the higher chemicalreactivity due to decreased cation coordination in tamorphous state, (v) an increase in ionic diffusivitiein the damage state that provides a convenient medifor ion exchange reactions that initiate corrosion, a(vi) solute segregation that leads to enhanced remoof radionuclides. In addition, actinides released duriaqueous corrosion of glass and ceramic waste forms ually have a relatively low solubility and precipitate in thleached gel layer as actinide-bearing phases.159,168,169,328

However, radiation damage to these actinide-bearprecipitates may lead to the formation of amorphocolloids that are mobile, thus, enhancing actinide tranport and increasing the concentrations of radionuclidin solution. All of these possibilities are expected tcontribute to an increase in the release (and posstransport) of radionuclides depending on the nature aextent of the radiation damage. The microstructurchanges in the radiation-damaged solid will depemost critically on the cumulative dose and the thermhistory of the waste form over extended periods of tim

The radiation-induced changes in microstructuevolve over long time periods and are of criticaimportance to the long-term radionuclide release duricorrosion. Thus, the evaluation of radiation effects ochemical durability requires a substantial databasethe development of microstructure as a functionradiation dose in host phases for both fission produand actinides. These data must be understood in termdamage-accumulation models that can then be combiwith models that describe the alteration and corrosiof the waste form. The models of both phenomena mbe extrapolated over long periods (104 to 106 years),but at least in the case of radiation effects, the modof damage accumulation can be confirmed by studof naturally occurring phases of similar structure ancomposition that have accumulated damage for timperiods up to 109 years.

VI. APPLICATIONS OF COMPUTERSIMULATION METHODS

Atomic-level simulation of radiation effects in metals, intermetallics, and semiconductors is an areaactive research.329 Quantum mechanical and empirical models of atomic bonding, energy minimizationmolecular dynamics (MD), and Monte Carlo (MD) methods can be used in atomic-level calculations of radtion damage processes in crystalline (and amorphomaterials.330 Energy minimization is widely used to


ri)-ls

esmdalgu-

gss-s

ledldl.

elgnnftsof

ednst

lss

de

of

,

-s)

study structures, stable defect configurations, and energminima for defect motion. Static defect properties as wellas dynamic processes on the order of picoseconds cabe modeled by MD simulations. Molecular dynamicscan be used to study defect formation and migrationenergies, damage mechanisms, and defect productioprocesses in cascades. Monte Carlo techniques, whicpredict behavior from a random sampling of initialstates, are useful for calculating damage distributionsthermodynamic equilibrium structures and properties,and long-range diffusion.

The prospect of using computer simulations to cal-culate irradiation-induced defect production and amor-phization in ceramic waste forms, and to follow theevolution of the defect structures, has been a beckoningyet elusive goal. However, progress achieved in com-putational physics for developing reliable yet tractableinteratomic potentials, coupled with vast improvementsin computational power, have created the possibilityof computing microstructure evolution during prolongedirradiation. The aim of such simulation work on radiationeffects in ceramics should be to develop an effectivemodeling technology for predicting radiation-inducedstructural and chemical stability and the long-term per-formance of nuclear waste form ceramics.The ultimateresult of this work should be a physically based modelof the effect of radiation damage on defect accumula-tion, phase transformations, and the evolution of themicrostructure over geological time scales.Develop-ment of a predictive simulation and modeling hier-archy for radiation effects in nuclear waste forms wouldenable (i) detailed understanding of defect energetics(ii) atomic-scale understanding of the mechanisms anddose dependence of phase transformations (e.g., amophization), (iii) accurate interpretation of experimentalresults in terms of underlying atomistic mechanisms, and(iv) evaluation of behavior of new waste forms underirradiation as needed.

A. Structure and defects

Owing to the pivotal role played by defects in in-fluencing the physical and chemical properties of solids,there is a long history of the characterization of impurityand defect states in ceramics using theoretical and computational techniques. The earlier achievements in thisfield have been described in detail by Stoneham,331 whilethe more recent developments have been summarizeby others.332–334 Considerable success has been enjoyedin many applications, especially for closed-shell de-fects in strongly ionic materials; indeed, in many casesquantitative agreement with experiment can now beachieved, and the computational methods have evolveinto straightforward and generally applicable tools thatmay be used routinely together with experiments in char-

3, No. 6, Jun 1998


st

d

ct

rbot

lo

n

d

t

a

sts.s

y

s,s

er

-

a-sge.-t-

nd

-ndn

ll

r-

s

d

sgr

is

nt

acterizing defect and impurity states. In other cases, mnotably open-shell defects in covalent materials, thestill remain substantial problems in achieving a higlevel of quantitative accuracy, although computationtechniques with present methodologies can still maa substantial contribution to characterizing energies athe structures of defects and impurities.

With the constant growth in computer power, thrange and accuracy of modeling techniques continueexpand. Computer simulation methods have a substial and growing role in examining the structures anstability of complex inorganic materials, including solisolutions, in modeling defect formation and migratioprocesses, in investigating radiolysis processes, andmodeling surface and interfacial phenomena, includisurface segregation.

1. Modeling techniques

Defects may be modeled either by embedding tdefect (or defect aggregate) in an infinite perfect lattior by employing an infinite periodic array of defecembedded in a supercell. The latter approach hasadvantage of being able to employ the wide rangecomputational solid state tools available for the peodic infinite solid; however, these methods mustadapted for charged defects. Moreover, the methautomatically include interactions between defects indifferent supercells, which may require the use of largcells, hence increasing the computational expense ofcalculation. The use of isolated embedded defects avothese problems, but these methods inevitably invodescribing the regions close to the defect using meththat are different from those employed for more distaregions of the structure; interfacing the different regioinvariably presents problems, and indeed, the embeddproblems remain amongst the most enduring and chlenging problems in computational solid-state physics

Within the context of the strategies outlined abovthe whole range of current computational solid-statools—quantum mechanical (QM) methods emploing ab initio Hartree–Fock, local density functiona(LDF), and semi-empirical methodologies—may bimplemented. Methods based on interatomic potentimay also be employed, such as those based on the wiand successfully used approach of Mott and Littleton335

that involves embedding an atomistically modeleregion containing several hundred ions surroundeda quasi-continuum description of the remaining latticWith the growth of the applicability of explicit QMmethods, the core of the region (including the defeand one or two shells of immediate neighbors)described using such techniques; although as nothere are difficulties in interfacing this core (QM) regiowith the region described more approximately by


ostrehalkend

eto

an-d

nin

ng

hee

stheofi-eds

heertheidsveds

nts

ingal-.e,tey-lealsely

dbye.

ctised,nn

interatomic potential. Nonetheless, explicit QM methodare essential for accurate modeling of open shell defec

One key aspect of all defect modeling approacheis the necessity of including full and explicitrelaxa-tion of the lattice around the defect. Omission of thiscrucial effect will inevitably lead to overestimation ofdefect formation energies and may result in qualitativelincorrect results. The majority of defect calculationshave used static lattice, energy minimization techniquewhich may be extended by lattice dynamical techniqueto include entropy calculations.332 There also have beenfruitful applications of molecular dynamics methodsin modeling highly mobile defects and Monte Carlomethods in treating heavily disordered systems. Furthdiscussion can be found elsewhere.332–334

Although it is possible to use computational techniques to model the structures of amorphous solids,332

such as those produced by radiation-induced amorphiztion, it must be emphasized that modeling of impuritieand defects in these states represents a major challenThe extension of modeling techniques to include surfaces and surface defects is, however, relatively straighforward, as reviewed elsewhere334 and the referencescited therein.

2. Modeling achievements

Computer modeling is now a well-established tool inceramic science. Crystal structures can be accurately aincreasingly predictively modeled. Moreover, modelingof the properties of crystalline ceramics, including elastic, dielectric, and lattice dynamical properties, has beeincreasingly successful. Defect modeling has enjoyenotable success in the last 20 years. Accurate formatioenergies may be routinely calculated for closed shedefects in ionic and semi-ionic oxides.332–334 Similarsuccess is enjoyed in the calculation of migration enegies. A good illustration is provided by the recent workof Cherry et al.336 who successfully modeled vacancymigration mechanisms in perovskite structured oxideand predicted the A/B cation radius ratio required togive optimal oxygen ion mobility.

Other notable achievements that are relevant anpossibly applicable to studies of radiation effects inceramic waste phases are:

(i) Defect interactions and clusteringhave long beena successful field for modeling studies; such studiehave made a major contribution to the understandinof the complex structures of heavily defective doped ononstoichiometric solids.337

(ii) Defect processes and reactionsin solids may beinvestigated, as in the detailed studies on the radiolysof NaCl338 and CaF2.339 The energetics of formation ofsolid state solutions may be calculated, as in relevastudies of fission product solution in UO2

340 or following

3, No. 6, Jun 1998 1467


rici

ant

oe

m

r

en

dite

rce

b

ot

r-r-

-byy-l.h

ony

hed-se

r-le,ic/e

of,rs

te.reesedo

tom-nd

u-tioneornd.

ereughnfndndnrip-nle,

ro-or

int

the recent work of Grimeset al.341 who have calculatedthe solution energies of a variety of dopants in Y2O3.

(iii) Modeling of surfaces and surface defectsisa long-standing field of growing importance.334,342 Asnoted, modeling has made a particularly useful contbution to understanding the key processes of impurand defect segregation. Modeling of molecule-surfainteractions, which is of major importance in catalysand sensor studies, is developing rapidly.

3. Challenges for ceramics modeling

Despite these successes, there remain problemand difficult areas for computer simulation efforts iceramic science. Those of greatest relevance to radiaeffects concern the following:

(i) Modeling of the energies (and free energies)phase transitions in ceramics, where the widely usshell model potentials often give inaccurate and somtimes qualitatively incorrect results.

(ii) Modeling of open shell defects and of interstitiadefects in dense structures. Models based on interatopotential methods are intrinsically unsuitable for thformer, while in the latter, it is difficult to obtain aninteratomic potential of the necessary quality.

(iii) Modeling of reaction mechanisms involvingbond-breaking processes in solids and on surfaces.

(iv) Predictive modeling of the structures of amophous materials (for comparison to data from x-raspectroscopies or neutron scattering).

All of these challenging problems are consideresolvable given the appropriate methodological devopment and access to the appropriate computatioresources.

4. Modeling opportunities and priorities

With the ongoing developments in computer harware and software, there is now a real opportunto develop detailed models of the structure and defchemistry of both crystalline and amorphous ceramicThe following areas should be given high priority:

(i) Development of detailed models for the structuand energies of pure and doped forms of all relevantramic phases (including zircon, zirconolite, pyrochlorzirconia, monazite, apatite, and perovskite). The comptational predictions should be validated and refinedcomparison with experimental structural data.

(ii) Calculation of formation, migration, andinteraction energies (and free energies) of all relevadefects including metal and oxygen interstitials anvacancies and their aggregates as well as defect-dopcomplexes. A principal objective of this work will beto provide the data necessary for implementation inkinetic Monte Carlo schemes of the damage evolutiprocess. In particular, it is strongly recommended tha


i-tye

s

tic

ion

fd

e-

lic

e

-y

dl-al

-ycts.

ee-,u-y

ntdant

tona

pilot study be made on zirconia in view of the consideable existing knowledge of defect structures and of inteatomic potentials in this system.

(iii) Development of improved models of the structure of amorphous (and amorphized) ceramics bothrefinement of conventional melt/quench molecular dnamics (MD) techniques and by coupling of topologicamodeling with MD and energy minimization methodsAgain, the modeling should be closely coupled witimproved experimental data on amorphous systems.

(iv) Modeling of segregation processes. Segregatiof impurities to surfaces or grain boundaries is a keaspect of the materials chemistry of ceramics, and teffects may be enhanced in irradiated materials. Moeling of both the energetics and the kinetics of theprocesses is needed.

(v) Modeling of solution processes at ceramic sufaces. Such studies, which are ambitious but feasibwould require the development of models of the ceramwater interface, of the reaction of water with the surfacof the material, and of the solution mechanismssurface ions. The work could build on, for exampleearlier studies343 of the mechanisms of reactions of watewith silicate solids and eventually expand to reactionwith calculated structures of the radiation-damage sta

All these developments are feasible; however, thewill be substantial requirements related to methodologiand hardware. Methodologies are needed for improvinteratomic potentials (where there is a need to gbeyond the traditional shell model approach) andfurther develop the procedures for embedded quantumechanical cluster techniques. Further refinement aautomation of procedures for vibrational entropy calclations are also needed. The use of the new generaof massively parallel processing (MPP) platforms will bessential, in particular for large scale MD calculations fgenerating improved models of amorphous systems afor certain electronic structure calculations on defects

B. Primary damage production

A complete description of the initial damage stat(i.e., the source term for all subsequent microstructuand property changes) can best be ascertained throthe use of computer simulation methods in conjunctiowith experimental methods. In addition, the evolution othe damage and its effect on structural properties aperformance can be evaluated by the use of kinetic aforce-bias Monte Carlo (MC) methods. The combinatioof these techniques enables a complete atomistic desction of irradiation damage and microstructure evolutioover all the relevant length and time scales. For exampradiation-induced amorphization can occur either hetegeneously in the cores of the displacement cascadeshomogeneously as the result of the accumulation of po

3, No. 6, Jun 1998


l

nso

s

n

.

o

set-ithyreent

-

y

us

ss)

.lhessor

od,as

ets

s,e

f

.e

rctnt

-

s-

defects and small defect clusters. The actual mechanisfor amorphization are complex, and experimental resuare difficult to interpret because little is known about thdefect kinetics that may control the processes. MD aMC simulations have the advantage that the importaparameters controlling amorphization, such as damazone stability and defect migration and recombinatiocan be studied in isolation; therefore, their relativimportance can be evaluated.

Model potential molecular dynamics, by virtue oits simplicity and the appropriateness of the lengscale (e.g., a cube of2 3 105 nm3 contains 107 atoms),is an extremely powerful tool to obtain atomic-scainformation and physical insight into the mechanismsirradiation-induced displacements. In a MD simulatiothe phase-space trajectories of a collection of degreefreedom (atoms) in a box are obtained from integratiof the classical equations of motion.344 The forces areobtained from the interatomic potential between thatoms in the system.

To investigate radiation effects in oxide ceramicthe interactions between atoms can be described withteratomic potentials that have both long-range Coulominteraction terms and short-range interactions betwepairs of ions. These types of potentials have been uextensively345,346 to model the behavior of ionic solidsand minerals in molecular dynamics simulations ahave been shown to be successful in modeling tstructure and thermodynamic properties of silicates.347

In fact, according to Wright and Catlow348 there is goodevidence that these potentials, which perform well on tcalculations of structure and thermodynamic propertiealso work well in defect simulations. The parameterssuch potentials for many binary and ternary oxides habeen derived from first principles by Bushet al.,345 andthe parameters for Zr–O, Zr–Si, Si–O, and O–O pain various valence states are available in the literature349

These parameters can be used in initial calculationsthe primary damage state in phases such as zirco(ZrO2) and zircon (ZrSiO4). The results of first principlecalculations can be used to optimize and validate texisting empirical interatomic potentials for zirconia anzircon. First principles calculations can also be useddevelop parametrizations for the interatomic potentiarequired for pyrochlore and its monoclinic derivativezirconolite. These new potentials can in turn be usedinvestigate the radiation stability of these ceramics.

1. Threshold displacement energy

As noted in Sec. III. D, there are only very limiteddata on the threshold displacement energies for ionsthe different sublattices of ceramics.69 Such data arecrucial for estimating the irradiation dose (numberdisplacements per atom, or dpa) to achieve a cert


msltsendntgen,e

fth

eof,ofn

e

s,in-b

ened

dhe

hes,

inve

irs

ofnia

hedtols,to

on

fain

damage state, such as amorphization, under a givenof irradiation conditions and also for correlating accelerated test data, based on ion-beam experiments, wresults due to longer-term damage from actinide decaor other radiation sources (e.g., neutrons). Moleculadynamics simulations can be used to explore the valuof the threshold energy for stable defect production, thlength of replacement collision sequences along differecrystallographic directions, and the spontaneous recombination volumes of various Frenkel pair configurationsin complex crystalline ceramics. The threshold energis defined as the minimum recoil energy required toform a stable (i.e., separated beyond its spontaneorecombination volume) Frenkel defect and is typicallyin the range of 30 to 50 eV. Such defects are formed athe result of replacement collision sequences (RCS’along low index crystallographic directions that werefirst predicted by Vineyard and co-workers350 on thebasis of MD simulations of low energy recoils in copperRCS’s are the critical mechanism whereby an interstitiacan be efficiently separated from its own vacancy. SucMD simulations have not yet been applied to any of thceramic phases of interest (Table V), but their usefulneand application to ceramics have been demonstrated fSiC.77 Static energy minimization techniques can alsobe used to calculate the minimum energy barrier tdisplacing an ion to a stable or metastable site anthus, to estimate the threshold displacement energy,has been recently done for several oxides.76 However,such static lattice calculations must be used with somcaution, since the neglect of quantum chemical effeccould lead to errors.

Knowledge of the threshold displacement energieEd, is necessary in order to accurately describe thproduction of nearly isolated point defects bya-particleson the different sublattices. The rather large number oisolated defects produced bya-particles can have a sig-nificant effect on the kinetics of microstructure evolutionValues forEd are also needed in order to estimate thdamage rates fora-recoils, charged particles, and neu-trons through the modified Kinchin–Pease expression351

or through use of the binary collision codes56–58that werediscussed above (Sec. III). Therefore, knowledge ofEd

for both the cation and anion sublattices is critical foaccurate predictions of damage production and defeaccumulation during prolonged exposure to irradiatioand for comparison of data obtained using differenirradiation techniques. Significant differences inEd forcations and anions can result in nonstoichiometric damage production.

2. Displacement cascade simulations

Ultimately, the study of radiation effects in nuclearwaste form ceramics requires that the displacement ca

3, No. 6, Jun 1998 1469


cy

a

ela

y-m

h

h

p

hdue

d

n

me

ze

gs

)de

ct

til-t,

lyin

ede-

-iali-

cades generated along the path of the energetic partiproduced bya-decay be well understood, particularlthat of thea-recoil. The nature of energetic cascadesmaterials has been the subject of intense study for myears.352 However, the small volume (,4 3 103 nm3)and short lifetime (,10211 s) of the cascades maktheir investigation by experimental means very difficuThe primary damage state that survives the cascevent (i.e., the source term for subsequent diffusion amicrostructure evolution) can only be inferred indirectl

Atomic-level computer simulation by MD is a valuable technique for investigating the mechanism of daage production in displacement cascades in cerambecause the lifetimes and volumes of the cascades, ware too small for direct experimental study, are withthe realm of current computational power, such asavailable on large parallel processing platforms. Tresult of a typical MD simulation is illustrated in Fig. 22which shows the primary damage state resulting froa 10 keV Si1 cascade inb –SiC.256 In this figure, thedistribution of interstitials and vacancies is shown 9.4after the initiation of the Si primary knock-on atom(PKA), and the energy deposited in the crystal by tSi PKA has been dissipated by the damped perioboundaries. The most striking features of this resare that the number of surviving C defects is thrtimes larger than the number of surviving Si defecand no well-defined amorphous regions are producedrectly in the cascade. The application of MD techniquto studies of displacement cascades in oxide ceramhas been primarily limited by the lack of fundinghowever, the availability of suitable potentials is alsan issue that must be addressed. Recently, empirpotentials of the Born–Mayer–Huggins type have beused with some success in MD simulations of cascain a SiO2–B2O3–Na2O–Al2O3 –ZrO2 simulated wasteglass353,354 and in ZrSiO4.355

3. Damage evolution and amorphization

Although MD simulations can be used to studthe structure and the initial evolution of the radiatiodamaged state, the computational time becomes prohtive beyond the first few nanoseconds of the simulatioeven for state-of-the-art scaleable parallel machinThis is an important consideration when studying mcrostructural evolution and amorphization mechanisduring prolonged irradiation at elevated temperaturOne of the critical issues is the competition between tirradiation-induced damage production and amorphition and the tendency of the system to equilibrium drivby the lower free energy of the crystalline state. Anoted in Sec. V. E, amorphization occurs either heteroneously by the accumulation of discrete damage zone


les

inny

t.dend.

-icsich

inise

,m

s

eiclte

tsdi-

esics;oicalenes

y-ibi-n,

es.i-ss.

hea-nse-or

FIG. 22. Interstitials (black) and vacancies (gray) of Si (large circlesand C (small circles) at the end of a 10 keV Si displacement cascain SiC.256 The cell size shown is 8.72 nm3 8.72 nm3 13.08 nm,which is half the size of the simulation cell that was used.

homogeneously by the accumulation of a critical defeconcentration. It is extremely difficult to predict themechanism of amorphization for a given system withoua detailed knowledge of the damage morphology, stabity, and defect kinetics. From a simulation standpointhe difficulty of simulating microstructural evolutionand amorphization arises as a result of the extremedisparate time scales between damage productionthe cascade (10211 s), the evolution of the damagezones (1029 to 1023 s), defect diffusion between cascadeoverlaps at a given dose rate (1023 to 102 s), andmicrostructure evolution over geological time scales (107

to 1013 s).In order to bridge the time scales between th

primary damage state creation and the subsequentfect and microstructure evolution, kinetic Monte Carlo(KMC) methods must be used. The input for the diffusion source term can be taken from the model-potentMD simulations of the displacement cascade. The phys

3, No. 6, Jun 1998


n

io

ehit

esn

ie

Dsa

i

s

r

nio

t

o

nu

:n

s

a

esns

,

de

r-c-s

e

lt

the

aisiness-lly

for

es

ged

ae

lteity)r

h

-o-t-

e-

th.

cal input parameters needed for the KMC simulatiomust be obtained from a combination of experiments acalculations based on transferable interatomic potentand/or more accurate electronic-structure-based meth

A hybrid model-potential MD/KMC simulation canbe performed by generating the displacement cascadmodel-potential MD and passing the information on tresulting defect distribution to the KMC code. In thtype of simulation, the vacancies and interstitials creain a displacement cascade are given random jumpsa rate derived from their diffusivities at temperaturthat allow for vacancy-interstitial recombination, clutering of like defects, re-emission from the clusters, atrapping and de-trapping of interstitials at native trapAnnihilation occurs at internal sinks with a specified sinefficiency, and periodic boundary conditions are applin the lateral directions. Diffusion proceeds until tharrival of a new cascade from the model-potential Mas determined by the irradiation dose rate. Succescascades are generated and annealed until the irradidose or the geological time scale are reached. An imptant application of such an MD/KMC approach is to thstudy of dose rate effects during damage accumulatas has been done in Si.356

In addition to point defects and small clusterthe cascade damage simulated by the model-potenMD code may contain large regions of highly disodered amorphous-like material, which are not eastreatable by MC simulators currently in use. For threason, extensive MD calculations of the stability aannealing kinetics of second-phase amorphous regin crystalline ceramic matrices should be performeThese studies can provide a tabulation of a setrecrystallization-kinetics data for amorphous pockets acrystal-amorphous interface regions. These data canbe part of the input data set for the KMC simulatoThus, the KMC simulator can treat the kineticsrecrystallization in a phenomenological way, albeit bason the atomistic results of the MD simulations.

The KMC codes require the following informatio(obtained from a combination of experiments, MD simlations, and first principles calculations) to simulate dfect diffusion in crystalline ceramics during irradiation

(1) Point defect (cation and anion vacancy ainterstitial) formation energies and diffusivities.

(2) Point defect-impurity interaction energy.(3) Point defect cluster binding energies.(4) Extended defect nucleation kinetics and the ro

of impurities.(5) Bias effect of point defect clusters on point defe

diffusion.(6) Bias effect of sink microstructure (dislocation

stacking faults, etc.) on point defect diffusion.(7) Information on any possible vacancy-interstiti

recombination barrier.


sndalsds.

bye

sed(at)-d

s.kd

e,

ivetionor-eon,

,tial-ilyisdns

d.ofndhenr.f

ed

-e-

d

le

ct

,

l

C. Modeling amorphous structures

Two approaches to modeling amorphous structurhave been reported: molecular dynamics simulatioand topological modeling. Starting with suitablepotentials, amorphous models of SiO2 and Si3N4 havebeen generated by molecular dynamics simulations,357

with a full set of atom coordinates available. Similarlycollision cascades in SiC,256,257 Ge and Si,358 inZrSiO4,355 and in a SiO2–B2O3 –Na2O–Al2O3–ZrO2

simulated waste glass353 have been modeled. Fullatomic coordinates can, in principle, be determineat time intervals during the evolution or change of thamorphous state within a cascade.

A second technique has involved hand- or computebuilding of models using algorithms based on reproduing experimentally observed bond-angle distributions, ain a recent model of vitreous silica.359 A somewhat dif-ferent approach is to investigate with modeling the rangof topological possibilities for amorphous structures.This approach was initially carried out using hand-buimodels of network silica,360,361 and characterization ofthe assemblages was carried out using the concept oflocal cluster,which is the set of all tetrahedra belongingto the set of undecomposable rings passing throughgiven tetrahedron in the structure. The local clusteran analogue in amorphous structures to the unit cellcrystals: an element of local structure that characterizthe topological properties of the whole amorphous asemblage. The local cluster assignments can be equaapplied to crystalline structures as has been donesilica polymorphs,360–362as well as for SiC and Si3N4.362

An efficient computer-based building approach that uslocal rules based construction362 can quickly constructcrystalline structures with a compact set of local buildinrules or can generate amorphous structures with modifirules. In this way, it is possible to construct fullyconnected distinguishable models for amorphized silicwith topological properties close to those of crystallinquartz or crystalline cristobalite or tridymite and withcredible bond-angle distributions. Using this topologicaapproach, it has been shown that it is possible to simulacascade disorder/damage by destroying the connectivwithin a designated region of a crystal (e.g., in quartzand regrowing with a different set of rules (e.g., those fothe higher-temperature polymorph cristobalite), whicyields an amorphous structure.362

Approaches to developing models of the radiationinduced amorphous structure in ceramics used for immbilization of radionuclides should include: (i) refinemenof conventional melt/quench MD techniques, (ii) coupling of topological modeling with MD and energy min-imization methods, and (iii) determining the dependencof the amorphous structure (and its density) on temperature. Such modeling should be closely coupled wiimproved experimental data on amorphous structures

3, No. 6, Jun 1998 1471


h

rno

a

e

lltoi

u

a

g

i

q

a-

let)g

ds

to-ss

l

-

tons,ild

em

s

atygyese

edrntyte

y

d

VII. IRRADIATION FACILITIES

Irradiation facilities are needed to investigate teffects of radiation on ceramics for HLW or Pu disposal, and these include actinide research laboratog-irradiation facilities, and modern electron- and ioirradiation facilities. For actinide research, the radiatieffects accumulate continuously over long times, ain situ measurements under controlled conditions (teperature, thermal gradient, stress, etc.) are possibleperhaps less necessary than for accelerated techniqIn situ measurements are possible in manyg-irradiationfacilities, ion-beam irradiation facilities, and electronbeam or electron microscopy facilities.

Many modern research techniques and facilitiesavailable to the materials science community and pvide unique opportunities to systematically investigairradiation effects in ceramics. The techniques ranfrom probes of the local atomic structure to bulk chaacterization and make use of major user facilities, suas accelerator-based photon sources, neutron scattfacilities, and high-energy or high-resolution electromicroscopes. These techniques and facilities need tomade available for the study of radiation effects inra-dioactive materials, including those containing actinide

A. Actinide research facilities

Modern laboratory facilities to perform research oactinide-containing materials are limited. Many facities with the capability to handle radioactive materiaincluding actinides, have been shut down at DOE sibecause of the change in the DOE mission from weapproduction to environmental remediation. The capabties to handle and study actinides are decreasing eyear. There are very few laboratories equipped wmodern analytical capabilities to handle radioactive mterials. Many of the advanced characterization techniq(e.g., ESR, neutron scattering, Raman spectroscopy,NMR) can handle encapsulated radioactive materibut these instruments are, unfortunately, often locain laboratories where radioactive materials are no lonallowed. There is, regrettably, no central multi-usfacility in the USA for handling radioactive materialsparticularly actinides, with all the necessary charactezation facilities. Consequently, detailed characterizatof specimens often requires shipping from site to sand, in some cases, re-encapsulation for each techni

B. Gamma irradiation facilities

Several intense60Co gamma sources and irradiatiofacilities exist at DOE and other government sitethese include the Pacific Northwest National LaboratoSandia National Laboratories, Argonne National Laoratory, Savannah River Site, and the Naval ReseaLaboratory. Even using the most intense60Co sources,


e-ies,-n

ndm-butues.

-

rero-teger-chringnbe

s.

ni-s,esns

li-achitha-es

andls,

teder

er,ri-oniteue.

ns;ry,b-rch

irradiation experiments may take years.In situ opticaland electrical measurements in these facilities are fesible. In many of these60Co facilities, the gamma fieldis nonisotropic. More isotropic gamma fields are possibusing spent nuclear fuel, particularly the cylindrical spenfuel elements at HFIR (Oak Ridge National Laboratorythat are hollow along the cylindrical axis. Understandinionization effects fromb-particles andg-rays in ceramicphases relevant to fission-product immobilization woulbenefit immensely from controlled temperature studieand in situ measurements utilizing these facilities.

C. Ion-beam irradiation facilities

There are numerous ion-beam irradiation facilities agovernment laboratories, universities, and industry labratories. Some include dual and triple beam capabilitiethat allow simultaneous irradiation with several speciein order to simulate both thea-particle and thea-recoilnucleus that are produced duringa-decay. These multi-beam facilities also allowin situ ion-beam characteriza-tion (Sandia National Laboratories, Los Alamos NationaLaboratory,363 and Oak Ridge National Laboratory) orelectron-beam characterization (IVEM and HVEM facilities at Argonne National Laboratory364). The ion-beam analysis techniques, which are primarily usedmeasure changes in near-surface chemical compositioinclude Rutherford backscattering (RBS), elastic recodetection (ERD), nuclear reaction analysis (NRA), anparticle-induced x-ray emission (PIXE).365 Light ele-ments that are usually difficult to detect with RBS areasily measured with ERD methods. These ion-beairradiation facilities also could be equipped within situoptical characterization. Ion-beam irradiation facilitieat universities include the University of Michigan, Uni-versity of Illinois, and Alabama A&M University.

D. Electron-beam irradiation facilities

High-voltage electron microscopes, such as thoseArgonne National Laboratory and Lawrence BerkeleNational Laboratory, can be used to simulate high-enerb-particle damage. Conventional electron microscopthat operate in the range of 100 to 400 kV can also bused to simulateb-particle damage. Unfortunately, thesefacilities have extremely high dose rates as comparwith the damage rates in ceramics containing HLW oactinides, producing in seconds the damage equivaleto 1000 years of storage. Other useful facilities mabe the electron accelerator facilities at Oklahoma StaUniversity and Wright State University or the pulsedelectron-beam facilities at Argonne National Laboratorand Brookhaven National Laboratory.In situ charac-terization capabilities include time-resolved EPR anfluorescence, but other capabilities could be added.

13, No. 6, Jun 1998


cectp

nnle

nh

n

neonleV

oe

ce

n

vd

t

ns

nsc-ere-

sic

g

hic

s

-

ith

gar-

aser

t-seet

ci-

ntres

us

n

e-ssaly

arerd

esfbys.then

There is a continuing and urgent need for an eletron-accelerator facility with the capability to measurin situ the threshold displacement energies in ceramiThis is needed not only for ceramics for nuclear wasapplications, but also for ceramics for fusion reactor aplications, space applications, advanced electronics, aion-beam processing/modification. High-voltage electromicroscope (HVEM) facilities, such as the HVEMfacility at Argonne National Laboratory, have beeutilized to indirectly determine threshold displacemeenergies through the measurement of the threshold etron energy required to produce an observable chanin microstructure (e.g., amorphization or dislocatioloop formation). As noted above (Sec. III. D), sucan approach often overestimatesEd, and for complexoxides, such as those for nuclear waste applicatiocan generally only determineEd for one sublattice.Such microscopes are also limited in energy, and cantransfer sufficient energy to displace heavy ion specisuch as Zr and the rare-earths. Although some electraccelerator facilities exist at several universities, nois currently utilized or equipped to measure threshodisplacement energies. In general, a facility is needthat has variable electron energies (0.1 to 2.0 Meand an end station capable ofin situ measurementsduring irradiation at liquid helium temperature. Ideallya facility capable of providing pulsed electron beamis desirable, as it would allow direct measurementsthe defects formed, during the pulse, by time-resolvspectroscopy techniques between pulses, as has bpreviously demonstrated for CaO,75 MgO,75,366 andAl 2O3.75,367

VIII. CHARACTERIZATION TECHNIQUESAND FACILITIES

Detailed characterization of the defects, microstrutural evolution, and structure of the amorphized statproduced by irradiation is necessary to advance the uderstanding of radiation effects in ceramics for HLW anPu immobilization. Two fundamental characterizatioobjectives are (i) the need for state-of-the-art facilitieto characterize defects and structures in radioactiprimarily actinide-containing, materials; and (ii) the neeto characterize defects and microstructures producedcharged-particle irradiation in small volumes, often asfunction of depth. The techniques that are highlighted blow should not be deemed as inclusive; other techniqumust and should be utilized.

A. Electron microscopy

Transmission electron microscopes are currenavailable with lateral spatial resolution under 0.15 nmwhich encompasses interatomic bond distances of oxidand phosphate compounds of interest as nuclear wa


-

s.e-ndn

tc-

ge

s,

ots,n-edd)

,sfdeen

-sn-d

se,

byae-es

ly,esste

forms. Such capabilities can be applied to specimein situ during ion-beam irradiation using new facilitiesat Argonne National Laboratory, and to cross sectioof ion-beam irradiated specimens. High-resolution eletron microscopy (HREM) has and will continue to maksignificant contributions to the understanding of structuand structural defects in nominally crystalline materials.368 However, conventional HREM is of limited usefor directly determining atomic structure in amorphoumaterials, even though a little information about atomcorrelations does survive in such images.369,370 Never-theless, HREM can be a useful technique for trackinthe progress and heterogeneity of amorphization,84,206

provided it is recognized that in projections througmixed crystalline and amorphized material, periodfeatures will dominate the image.371,372

The projection problem for amorphous materialmay be partially offset by limiting the local volume inwhich atomic correlations are probed by controlled reduction of the spatial coherence employed in imaging.373

Variable spatial coherence, which can be achieved whollow-cone illumination in conventional TEM and withannular dark-field detector configurations in scanninTEM (STEM), has been shown to be sensitive tohigher order than pair-correlations in imaging of amophous specimens.374 Information obtained this way aboutpair-pair correlations in amorphous Ge specimens hrevealed evidence for greater intermediate-range ord(IRO) that is consistent with continuous random nework models of the amorphous state. Hence, the uof variable-coherence HREM imaging methods has thpotential to provide at least some information abouamorphized structures and with greater spatial speficity than diffraction methods, which is important forion-irradiated specimens. The greatest utility at preseappears to be comparisons of real and model structuor comparisons of ion-amorphized structures witha-decay-induced (or other irradiation-induced) amorphostructures.

An advantage of an electron beam is that it careadily be focused to small areas (,0.5 nm diameterin some cases) and becomes a highly spatially rsolved spectroscopic probe. The electron energy lospectra contain information about the bonding and locenvironment,375 which can be used to establish densitand local density variations.376 The densities of the smallvolumes amorphized by ion beams or electron beamsotherwise difficult to establish. In addition, the behavioof the energy loss spectrum immediately before anafter the higher-energy characteristic ionization edgcontains information about the local environment ospecific excited atom types, analogous to that suppliedEXAFS spectra but from substantially smaller volumeThe method has been used to distinguish betweentrigonal and tetrahedral coordination of boron by oxyge

3, No. 6, Jun 1998 1473


i

e

a

o

n

e

m

ic

t

)f adge

fngor-

asndos-nd

nts)ndve,ral

e

e-ons

lenn-ralcts

dem

ndto

silyiththe

ticles

alsistasn-

andr-

in borate minerals377 and to establish the oxygen coordnation around aluminum at an Al2O3 grain boundary.378

Similar determinations should be possible in irradiatnuclear waste ceramics.

B. Diffraction

X-ray, neutron, and electron diffraction techniquehave all been used to explore the atomic structuof crystalline and amorphous materials. In crystallinceramics, pair correlations are strong, leading to shBragg diffraction peaks that depend on scattering froseparate sublattices. As an example, studies of iinduced amorphization of spinel262,263have indicated thatdisorder on the cation sublattice may be a precursoroverall amorphization.

In amorphous materials, the diffracted intensities cyield radial density functions (RDF’s) that are relateto the radial distribution and pair correlation functionthrough the local average density. In most cases, at lesome of the peaks in the RDF’s can be correlatedspecific short-range correlations (e.g., the oxygen codination shell about Si or P in silicates or phosphateand the integrated peak area can yield the coordition number. Unfortunately, because pair correlatiodo not efficiently probe higher-order correlations, thRDF’s are not enormously sensitive to differencesintermediate range order. A prominent feature of madiffraction patterns, however, is a first sharp diffractiopeak379 (FSDP), which represents the signature of somcharacteristic intermediate-range structural feature, suas the topologically dominant ring structure in silicnetworks.380 A significant experimental difficulty in ap-plying diffraction techniques to irradiation-amorphizespecimens is the small volumes rendered amorphouscharged-particle beams. Energy-filtered electron diffration (EFED) provides a means to obtain structural dafrom amorphized volumes significantly smaller than rquired for x-ray or neutron diffraction, as in studies oamorphized silicas381 and phosphates.382,383

The characterization of radiation damage in ceraics requires the simultaneous observation of crystalliand amorphous material and the analysis of theterfacial behavior of these two structural states. Tcrystalline material will generally have a periodic atomarrangement with a substantial amount of point defeand other extended defect aggregates present. Thdefect aggregations will also have an approximateperiodic atomic arrangement. In order to observe tvarious structural features (periodic, quasiperiodic aaperiodic regions in the material) with statistical significance, advanced x-ray diffraction (XRD) techniquemust be employed. Recently, a novel seven-circle x-rdiffractometer has been developed and optimized forsimultaneous observation of the diffraction signals fro


-

d

sreerpmn-

to

andsastto

or-s),a-

nseinnynech

a

dby

c-ta-f

-nein-hectsesely

hend-sayhem

crystalline regions (i.e., the Bragg diffraction maximaand the diffuse scattering from less-ordered regions osample.384,385 This unique diffractometer has been usein ongoing characterization studies of radiation damaand annealing behavior in natural zircons.386 Such acapability in the USA would be useful in studies oa-decay-induced amorphization in actinide-containiphases, and a diffractometer of this type could be incporated in a target chamber forin situ analysis duringion irradiation.

C. X-ray absorption spectroscopies

X-ray absorption spectroscopies (XAS), suchextended x-ray absorption fine structure (EXAFS) ax-ray absorption near-edge structure (XANES) spectrcopies, can be utilized to evaluate the first and secocoordination geometries (nearest neighbor environmearound selected elements, including the actinides arare earths, in nuclear waste ceramics. As noted abothese XAS techniques have been applied to structustudies of both Pu-containing ceramics248 and naturalminerals.27,237,245–247It is even possible to investigate thlocal structure of the energetically inserteda-recoilnuclei, as has been done recently for the234U daugh-ter product in238Pu-containing zircon.387 At the presenttime, these techniques can be applied to actinidcontaining compounds at the Stanford SynchrotrRadiation Laboratory. With the much higher intensitieof the new synchrotron light facilities, it may be possibto probe small volumes of materials, which is aadvantage for radioactive materials and critical for ioirradiated materials. Understanding the local structuenvironments of the actinides and their daughter produ(a-recoil particles) in ceramic phases can proviatomic-level information on changes in bonding frothe a-recoil cascade.

D. Phonon spectroscopies

Measurements of phonon spectra by infrared aRaman spectroscopy have infrequently been appliedirradiation damage studies. These techniques are eaadapted for application to radioactive materials and, wnew micro-beam techniques, can be used to probesmall damaged volumes produced under charged-parirradiation. The characterization of vibrational modewould provide valuable information on local structurchanges, and the results would complement and asin the interpretation of other structural data, suchfrom XAS. For example, Raman spectra of neutroirradiated vitreous silica376 show additional vibrationalmodes characteristic of threefold rings in the silicnetwork topology, in agreement with neutron diffractioresults.388 In addition, Raman spectra of ion-irradiateGd2Ti2O7 indicate the appearance of a normally fo

3, No. 6, Jun 1998


stngof

tellnot

a-naprchetalic,e

di-edl

-yf

meals.nsofingd-

nar-

tyttoin

or

se,e,stedn-n

es.eens

ses

bidden Raman mode due to disorder on the catsublattice.143

E. Hyperfine techniques

Hyperfine techniques that use nuclear probesstudy point defects in crystalline solids can be usin studies of some ceramic phase for nuclear waimmobilization. These techniques include: perturbed agular correlation spectroscopy (PACS), nuclear magneresonance (NMR), and M¨ossbauer spectroscopy (MSBoth NMR and MS are limited to phases containinone of several nuclei as major constituents (e.g., FeSn for Mossbauer). The PACS technique, which hbeen used to study defects and phase transitionsceramics,389 involves substituting trace concentrationsradioactive probe atoms, such as181Hf or 111In, into thecrystal structure. One limitation is, therefore, the neto fabricate ceramics with radioactive probes or hathe probes activated in a nuclear reactor. Since mawaste ceramic phases contain Zr and rare-earths,181Hfis an ideal candidate to study these phases. In fact, PAhas been applied, in nonradiation damage studiesseveral structures of interest, such as perovskite,390–392

pyrochlore,393 and zirconia.394 Using PACS, it is possibleto obtain information on defect migration and bindinenergies. In addition, PACS exhibits no line-broadeniwith increasing temperature, which makes it a sensittool to obtain data on defect kinetics.

F. Liquid-phase chromatography

The degree of polymerization of network structurcan be monitored by dissolution followed by solutiochromatography in which polymerized units (e.gcorner-sharing [SiO4] tetrahedra in silicates, cornersharing [PO4] tetrahedra in phosphates) retainedsolution can be separated by mass. Phosphatesmore amenable to this technique than silicates ahave been shown to polymerize significantly upoamorphization.395,396

IX. RESEARCH NEEDS AND OPPORTUNITIES

Both single-phase and multiphase crystalline ceraics can be used as durable hosts for the immobilizationradionuclides. Radiation damage (ionizing and ballistfrom incorporated radionuclides affects the physical achemical properties of the ceramics and the release rof the radionuclides during corrosion.

Because of the many applications for ceramicsradiation environments and the development of iobeam processing technologies for ceramics, there alreexists a substantial database and some understandinradiation-solid interactions in ceramics. However, thpresent understanding of radiation effects in compceramics is insufficient in detail and scope to pred


ion

toedsten-tic

).gor

asin

of

edveny

CS, to

gngive

esn.,-inarendn

m-of

ic)ndates

inn-adyg ofe

lexict

the effects of radiation on the various radionuclide-hophases under expected repository conditions over lotime periods. The cumulative macroscopic responsesdifferential radiation effects (e.g., swelling) in differenphases within multiphase waste forms are even less wunderstood. In many cases, the scientific issues arelimited to only ceramics for the immobilization of HLWand Pu; these issues are also critical for other applictions, such as fusion technology, ion-beam modificatioof ceramics, and ion-beam processing of wide-band-gsemiconductor devices. Recommendations for reseapriorities are summarized below. The primary objectivof these recommendations is to develop a fundamenknowledge and models of radiation effects at the atommicroscopic, and macroscopic levels in order to providfor the evaluation and performance assessment of invidual crystalline phases and multiphase ceramics usfor the immobilization of HLW, Pu, and other specianuclear waste streams.

A. Research

The development of predictive models of radiation effects in nuclear waste forms under repositorconditions will require a fundamental understanding othe effects of dose, dose-rate, temperature, and tion the radiation-induced, atomic-scale microstructurevolution in relevant ceramic phases and waste formConsequently, relevant structure types and compositioshould be systematically studied over the widest rangeconditions (e.g., dose, dose rate, and temperature) usa variety of radiation sources and techniques. Such stuies will require careful microstructural characterizatioduring the damage accumulation process, and such chacterization may best be performed utilizing a varieof techniques involving collaborations between differeninvestigators and institutions. Scientific issues relatedradiation effects in these solids have been discussedthis review. Recommendations and considerations ffuture research include the following:

(i) The most prominent of the actinide-host phaseamong the promising structure types include: pyrochlorzirconolite, zircon, apatite, perovskite, titanite, monazitzirconia, and NZP (sodium-zirconium-phosphate). Moof these structures are susceptible to radiation-inducamorphization, and for some of these structures, a cosiderable amount of preliminary work has already beecompleted on synthetic phases and mineral analoguAlthough the identified structure types should be thfocus of initial studies, other phases may yet assumimportance depending on the waste stream compositioand the development of new waste forms.

(ii) Investigations of potential actinide-host phasemust include studies of self-radiation damage in phascontaining short-lived actinides (e.g.,238Pu and244Cm).

13, No. 6, Jun 1998 1475


ebeg

si

na

o

inu

n

g

e

e

rte

nebfahcs

o

a

hi

nts

ded tollys

etalectmayts,-ta,udy

stof

n

usx

a-ntd

ey

r-s

ctus

dl,d

rei-

e

f

lpice.ted

Since such studies may last for many years (up to tyears or more), carefully planned experiments need toinitiated immediately on the most promising structurtypes. The limitations often imposed on the handlinof actinide-containing materials may severely restricthe number of analytical techniques that can be employed. Thus, studies of self-radiation damage in thephases must be complemented by investigations of meral analogues and ion-irradiated ceramics, for whichwider range of irradiation conditions can be studied ishorter time periods using a greater variety of analytictechniques.

(iii) Because phases used as hosts for fission proucts (e.g., Cs and Sr) will be subjected to high ionizingradiation fields for several hundred years, studiespotential fission-product phases must focus on the efects of ionizing radiation using electron-irradiation techniques org-irradiation facilities. Many of these phasesare hydroxylated ceramics (clays, zeolites, and concrephases), and principal radiation effects include radiolytgas formation, bubble formation, and changes in catioexchange or sorption capacities. Other host phases, sas CsCl and SrF2, are susceptible to the evolution ofradiolytic gas and the formation of metal colloids. Inaddition, some of these phases will not only be potetial waste forms, but are also important in processintechnologies because of their selective ion exchancapacities for Cs, Sr, and, in some cases, actinides.

(iv) The predictive capacity of contemporarycomputer simulation techniques and greatly expandcomputational capabilities (e.g., massively parallel computers) should be exploited to provide (a) atomic leveunderstanding, (b) calculations of fundamental materiaparameters and the primary damage state, (c) modof crystalline and amorphous structures, and (d) defechemistry of the crystalline ceramic phases of inteest. The results should be confirmed or relatedexperimental data. Such modeling efforts will requirthe development of improved interatomic potentialsAttention should be paid to modeling the sites and the eergetics associated with the substitution of radionuclidinto the proposed structures. Detailed models shoulddeveloped of the migration mechanisms and energieshost lattice ions, radionuclides, point defects, and gatoms. Calculations of defect energetics, migration patways, and stable defect configurations using static latti(energy minimization) methods will be essential for thieffort, but quantum chemical effects must be considerein the application of these static methods. The processof defect and cascade formation occur over such shtime scales (,10211 s) that they can only be studiedby computer simulation methods, such as moleculdynamics. Kinetic Monte Carlo techniques can be useto model diffusion processes and defect interactions. Textension of already existing models may be useful


ne

t-e

n-a

l

d--ff--

tec-ch

-ge

d-llsls

ct-o

.-se

ors-e

desrt

rde

n

describing microstructure evolution (e.g., amorphizatioand bubbles) and the role of radiation-induced defecand structural changes on dissolution and radionuclirelease. These methods and models should be useexplore and interpret experimental results, specificato correlate results from different irradiation technique(e.g., electron, neutron, ion, anda-decay).

(v) The interpretation of experimental results and thdevelopment of predictive models require a fundamenunderstanding and data on structure, properties, defenergetics, and radiation-damage processes. Thesebe determined directly by experimental measuremenindirectly by modeling experimental data, or by theoretical methods and computer simulations. The daproperties, and mechanisms that require systematic stinclude:

(a) Structural studies of potential radionuclide-hophases and the specific site occupanciesimpurities (e.g., incorporated radionuclides) ithese complex structure types.

(b) The threshold displacement energies for varioions on different sublattices in these compleceramics.

(c) The radiation-induced stable defect configurtions and primary damage state, the differedefect migration energies and pathways, anthe kinetics of defect annealing, including theffects of ionization and subthreshold energtransfers.

(d) The mechanisms of radiation-induced amophization (e.g., defect accumulation versuin-cascade amorphization).

(e) The mechanisms and kinetics of defect-defeinteractions, defect annealing, and amorphodomain annealing.

(f) The structure and evolution of radiation-inducemicrostructures, including detailed structurastudies of mixed periodic/aperiodic domainstheir interfaces, the fully amorphous state, anthe nature of the large density changes.

(g) The effects of dose, dose-rate and temperatuon defect accumulation, amorphization, and mcrostructure evolution.

(h) Helium accumulation, trapping, and bubblformation.

(i) Self-ion diffusion, impurity diffusion, and ther-mal conductivity.

( j) Radiation-induced diffusion and segregation oimpurities and major element components.

(k) The effects of radiation-induced differentiachanges (e.g., volume changes) on macroscoresponses (e.g., microfracturing) in multiphasceramics as a function of dose and grain size

(vi) In all these studies, the results from differenirradiation techniques and conditions must be correlat

3, No. 6, Jun 1998


so

ite

toc

il

aeeenn

uydds

t

ine

ot

yinflsoi

f

,

e-t

-

t

,

-

to some standard basis of dose, such as the ballidamage dose (i.e., dpa) or, in the case of ionizatidamage, the absorbed ionization dose (Gy). In addition,these correlations and models of damage accumulatand annealing should be validated by comparisoncomputer simulations and data from mineral analoguthat have experienced high radiation doses due to U aTh decay over long times (.100 million years).

(vii) Most importantly, corrosion (dissolution) ex-periments or test methods must be developedthat arespecifically designed to relate the effects of radiation ostructure to changes in the dissolution rate or releasof radionuclides.In general, the standard tests usedinvestigate the chemical durability of waste forms are nappropriate or sensitive enough to elucidate the effeof radiation damage on the corrosion process.

B. Facilities

The necessity for handling radioactive materials wrequire new facilities or access to existing nonactivfacilities. In some cases, researchers are deniedcess to facilities even though samples are small, safencapsulated, and of low activity. Facilities must bavailable for the preparation of actinide-doped sampland their subsequent study. The limitations on handliradioactive materials severely restrict the number of aalytical techniques that can be used to study the damaaccumulation and annealing processes. Minimum anlytical requirements include determination of physicaproperties (e.g., density, hardness, and thermal condtivity), scanning electron microscopy, autoradiographx-ray diffraction, transmission electron microscopy, anelectron microprobe analysis. A facility for handling ancharacterizing radioactive materials should find wide uthroughout the DOE complex.

Recent advances in x-ray diffraction techniques384,385

can be used in conjunction with ion-beam irradiationsperform radiation damage studiesin situ over a range oftemperatures. Although some electron accelerator facties exist at several universities, there is a continuing aurgent need for an electron accelerator facility with thcapability to measurein situ the displacement energiesin ceramics. These measurements will require variabelectron energies (0.1 to 2.0 MeV) and an end staticapable ofin situ measurements during irradiation aliquid helium temperature. Finally, there are numerousolid-state characterization techniques that have notbeen applied to the investigation of radiation effectscrystalline ceramics of interest for immobilization ohigh-level nuclear waste and plutonium (e.g., doubnuclear resonance techniques: ENDOR, ODMR). Invetigators are urged to consider potential applicationstheir techniques to the scientific issues outlined in threview.


ticn

onos

nd

neotts

lec-ly

sg-

gea-lc-,

e

o

li-d

len

set

e-f

s

C. Attract talented scientists

This field does not, in general, readily attract or sup-port young scientists, and much of the relevant knowl-edge and experiences will disappear with the passage otime. With the long-time scales envisioned for immobi-lization of HLW, Pu, and other nuclear waste streamsthere is a need for a relatively small but stable poolof expertise and continued training of new scientists inthis area. It is important to attract and involve brightnew minds with new ideas to this field. Increased oppor-tunities need to be made available to support graduatstudents and provide postdoctoral experience at universities, the national laboratories, and other governmensites. A long-term research program, with a continuityof purpose, is essential to attracting and retaining highcaliber scientists.

ACKNOWLEDGMENTS

This panel was sponsored by the Council on Materi-als Science of the U.S. Department of Energy, Office ofBasic Energy Sciences, Division of Materials Sciences(DMS/BES) under Contract DE-FG02-96ER45439 atthe University of Illinois. The Panel thanks Dr. IranThomas, DMS/BES Director, and Dr. Robert Gottschall,DMS/BES Team Leader, for their support. They alsogratefully acknowledge the support and encouragemenof Professor C. Peter Flynn (University of Illinois–Urbana), Chairman of the Council on Materials Sciencewho provided liaison with DMS/BES. The Panel thanksDr. Michael Knotek (Council on Materials Science)and Dr. Yok Chen (DOE, DMS/BES) for their activeparticipation in the Panel Meeting. The Panel also gratefully acknowledges the assistance of Susan G. McGuire(Pacific Northwest National Laboratory) during the panelmeeting and in the preparation of this manuscript.

REFERENCES

1. Closing the Circle on the Splitting of the Atom(U.S. Departmentof Energy, Office of Environmental Management, Washington,DC, 1995).

2. D. L. Illman, Chem. Eng. News71 (25), 9 (1993).3. Office of Technology Assessment,Complex Cleanup: The Envi-

ronmental Legacy of Nuclear Weapons Production,OTA-O-484(Congress of the United States, Washington, DC, 1991).

4. A. G. Croff, in Glass as a Waste Form and Vitrification Technol-ogy: Summary of an International Workshop(National AcademyPress, Washington, DC, 1996), p. D.1.

5. Management and Disposition of Excess Weapons Plutonium(National Academy Press, Washington, DC, 1994).

6. Management and Disposition of Excess Weapons Plutonium–Reactor-Related Options(National Academy Press, Washington,DC, 1995).

7. Storage and Disposition of Weapons-Usable Fissile MaterialsDraft Programmatic Environmental Impact Statement,DOE/EIS-0229-D (U.S. Department of Energy, Washington, DC, 1996).

8. F. von Hippel, Phys. Today47 (6), 26 (1995).

3, No. 6, Jun 1998 1477


i

i

d

.

l,

i,

-

s

.

,

in

a,

a

.

.

.

.

r.

9. D. J. Bradley, C. W. Frank, and Y. Mikerin, Phys. Today49 (4),40 (1996).

10. Glass as a Waste Form and Vitrification Technology: Summaof an International Workshop(National Academy Press, Wash-ington, DC, 1996).

11. I. W. Donald, B. L. Metcalfe, and R. N. J. Taylor, J. Mater. Sc32, 5851 (1997).

12. W. M. Murphy and R. T. Pabalan,Review of empirical thermo-dynamic data for uranyl silicate minerals and experimental planReport 95-014 (Center for Nuclear Waste Regulatory AnalysSan Antonio, TX, 1995).

13. D. Langmuir, Geochim. Cosmochim. Acta42, 547 (1978).14. B. Grambow, Mater. Res. Soc. Bull.XIX (12), 20 (1994).15. Rapport d’evaluation no. 1 de la Commission Nationale

d’Evaluation relative aux recherches sur la gestion des déchetsradioactifs instituee par la loi 91-1381 du 30 décembre 1991,juin 1995 (CNE, Paris, 1995).

16. Rapport d’evaluation no. 2 de la Commission Nationaled’Evaluation relative aux recherches sur la gestion des déchetsradioactifs instituee par la loi 91-1381 du 30 décembre 1991,juin 1996 (CNE, Paris, 1996).

17. A. S. Aloy, B. Ya. Galkin, B. S. Kuznetsov, R. I. Lyubtsev, anV. M. Esimantovskii, inWaste Management ’89,edited by R. G.Post (Arizona Board of Regents, Tucson, AZ, 1989), Vol. 1p. 677.

18. H. Yokoi, Y. Arita, T. Matsui, H. Ohno, and K. Kobayashi, JNucl. Mater. 238, 163 (1996).

19. R. C. Ewing, W. Lutze, and W. J. Weber, J. Mater. Res.10, 243(1995).

20. R. C. Ewing, W. J. Weber, and W. Lutze, inDisposal of WeaponsPlutonium, edited by E. R. Merz and C. E. Walter (KluwerAcademic Publishers, The Netherlands, 1996), p. 65.

21. Record of Decision for the Storage and Disposition of WeaponUsable Fissile Materials Final Programmatic EnvironmentaImpact Statement(U.S. Department of Energy, Washington, DCJanuary 14, 1997).

22. Strategic Plan, Storage and Disposition of Weapons-Usable Fsile Materials (U.S. Department of Energy, Washington, DCAugust, 1997).

23. S. G. Cochran, W. H. Dunlop, T. A. Edmunds, L. M. MacLeanand T. H. Gould, Fissile Material Disposition Program FinaImmobilization Form Assessment and Recommendation, UCRID-128705 (Lawrence Livermore National Laboratory, Livermore, CA, October 3, 1997).

24. X. Feng, J. E. Surma, C. G. Whitworth, R. C. Eschenbach, aG. L. Leatherman, inGlass as a Waste Form and VitrificationTechnology: Summary of an International Workshop(NationalAcademy Press, Washington, DC, 1996), p. E.68.

25. D. A. Knecht, T. P. O’Holleran, K. Vinjamuri, S. V. Raman,and B. A. Staples, inGlass as a Waste Form and VitrificationTechnology: Summary of an International Workshop(NationalAcademy Press, Washington, DC, 1996), p. E.78.

26. S. V. Raman, R. Bopp, T. A. Batcheller, and Q. Yan, inScientificBasis for Nuclear Waste Management XIX,edited by W. M.Murphy and D. A. Knecht (Mater. Res. Soc. Symp. Proc.412,Pittsburgh, PA, 1996), p. 113.

27. R. C. Ewing, B. C. Chakoumakos, G. R. Lumpkin, T. MurakamR. B. Greegor, and F. W. Lytle, Nucl. Instrum. Methods in PhyRes. B32, 487 (1988).

28. R. C. Ewing, inScientific Basis for Nuclear Waste ManagemenXVI, edited by C. G. Interrante and R. T. Pabalan (Mater. ReSoc. Symp. Proc.294, Pittsburgh, PA, 1993), p. 559.

29. W. Lutze and R. C. Ewing, inRadioactive Waste Forms for theFuture, edited by W. Lutze and R. C. Ewing (North-Holland,Amsterdam, 1988), p. 699.


ry

.

,s,

,

s-

s-

,lL-

nd

i,.

ts.

30. E. Anderson, B. Burakov, B. Galkin, V. Starchenko, and VVasilev, in Plutonium Stabilization and Immobilization Work-shop,CONF-951259 (U.S. Department of Energy, WashingtonDC, 1996), p. 269.

31. C. Degueldre, U. Kasemeyer, F. Botta, and G. Ledergerber,Scientific Basis for Nuclear Waste Management XIX,edited byW. M. Murphy and D. A. Knecht (Mater. Res. Soc. Symp. Proc.412, Pittsburgh, PA, 1996), p. 15.

32. C. Degueldre, P. Heimgartner, G. Ledergerber, N. SasajimK. Hojou, T. Muromura, L. Wang, W. Gong, and R. Ewing,in Microstructure Evolution During Irradiation,edited by I. M.Robertson, G. S. Was, L. W. Hobbs, and T. Diaz de la Rubi(Mater. Res. Soc. Symp. Proc.439, Pittsburgh, PA, 1997),p. 625.

33. P. G. Boczar, M. J. N. Gagnon, P. S. W. Chan, R. J. Ellis, R. AVerrall, A. R. Dastur, Canadian Nucl. Soc. Bull.18 (1), 2 (1997).

34. V. M. Oversby, C. C. McPheeters, C. Degueldre, and J. MParatte, J. Nucl. Mater.245, 17 (1997).

35. Hj. Matzke, J. Magill, and J. van Geel, inProc. 21st AnnualSymp. of the Uranium Institute,London, Sept. 4–6, 1996.

36. W. J. Weber and Hj. Matzke, Radiat. Eff.98, 93 (1986).37. R. C. Ewing, W. J. Weber, and F. W. Clinard, Jr., Prog. Nucl

Energy 29 (2), 63 (1995).38. W. J. Weber, R. C. Ewing, C. A. Angell, G. W. Arnold, A. N.

Cormack, J. M. Delaye, D. L. Griscom, L. W. Hobbs, A. Navrot-sky, D. L. Price, A. M. Stoneham, and M. C. Weinberg, J. MaterRes. 12, 1946 (1997).

39. Hj. Matzke, Nucl. Instrum. Methods in Phys. Res. B32, 453(1988).

40. W. J. Weber, L. K. Mansur, F. W. Clinard, Jr., and D. M. Parkin,J. Nucl. Mater.184, 1 (1991).

41. W. J. Weber and F. P. Roberts, Nucl. Technol.60, 178 (1983).42. W. J. Weber, J. Minerals, Metals and Materials Society43 (7),

35 (1991).43. W. J. Weber and R. C. Ewing, inScientific Basis for Nuclear

Waste Management XVIII,edited by T. Murakami and R. C.Ewing (Mater. Res. Soc. Symp. Proc.353, Pittsburgh, PA, 1995),p. 1389.

44. W. J. Weber, R. C. Ewing, and W. Lutze, inScientific Basis forNuclear Waste Management XIX,edited by W. M. Murphy andD. A. Knecht (Mater. Res. Soc. Symp. Proc.412, Pittsburgh,PA, 1996), p. 25.

45. C. Adelhelm, C. Bauer, S. Gehlert, and G. Ondracek, inRa-dioactive Waste Forms for the Future,edited by W. Lutze andR. C. Ewing (North-Holland, Amsterdam, 1988), p. 393.

46. A. D. Whapham and B. E. Sheldon, Philos. Mag.12, 1179(1965).

47. N. Nakae, A. Harada, T. Kirihara, and S. Nasu, J. Nucl. Mate71, 314 (1978).

48. N. Nakae, T. Kirihara, and S. Nasu, J. Nucl. Mater.74, 1 (1978).49. N. Nakae, Y. Iwata, and T. Kirihara, J. Nucl. Mater.80, 314

(1979).50. W. J. Weber, J. Nucl. Mater.98, 206 (1981).51. W. J. Weber, J. Nucl. Mater.114, 213 (1983).52. W. J. Weber, Radiat. Eff.83, 145 (1984).53. Hj. Matzke, O. Meyer, and A. Turos, Radiat. Eff.119–121, 885

(1991).54. A. Turos, Hj. Matzke, M. Wielunski, and L. Nowicki, Nucl.

Instrum. Methods in Phys. Res. B80/81, 1259 (1993).55. J. Janeczek and R. C. Ewing, J. Nucl. Mater.185, 66 (1991).56. J. F. Ziegler, J. P. Biersack, and U. Littmark,The Stopping and

Range of Ions in Solids(Pergamon, New York, 1985).57. M. T. Robinson and I. M. Torrens, Phys. Rev. B9, 5008 (1974).58. M. T. Robinson, Phys. Rev. B27, 5347 (1983); J. Appl. Phys.

54, 2650 (1983).59. W. J. Gray, Nature (London)296, 547 (1982).

3, No. 6, Jun 1998


y

-

,

z

c

.

.

.

)

d

,

.

.d,

e,

l

,

,

.

,

r.

,

,

,d

.

),

tr.

L.

s,

.

d

60. E. Browne, J. M. Dairiki, and R. E. Doebler,Table of Isotopes,edited by C. M. Lederer and V. S. Sherley, 7th ed. (John Wileand Sons, New York, 1978).

61. R. L. Fleischer, Science207, 979 (1980).62. M. Antonini, F. Lanza, and A. Manara, inCeramics in Nuclear

Waste Management,edited by T. D. Chikalla and J. E. Mendel,CONF-790420 (National Technical Information Service, Springfield, VA, 1979), p. 289.

63. L. W. Hobbs, J. Am. Ceram. Soc.62, 267 (1979).64. H. Matzke, Radiat. Eff.64, 3 (1982).65. N. Itoh and K. Tanimura, Radiat. Eff.98, 269 (1986).66. L. W. Hobbs, F. W. Clinard, Jr., S. J. Zinkle, and R. C. Ewing

J. Nucl. Mater.216, 291 (1994).67. S. J. Zinkle, inMicrostructure Evolution During Irradiation,

edited by I. M. Robertson, G. S. Was, L. W. Hobbs, and T. Diade la Rubia (Mater. Res. Soc. Symp. Proc.439, Pittsburgh, PA,1997), p. 667.

68. F. W. Clinard, Jr. and L. W. Hobbs, inPhysics of RadiationEffects in Crystals,edited by R. A. Johnson and A. N. Orlov(North-Holland, New York, 1986), p. 387.

69. S. J. Zinkle and C. Kinoshita, J. Nucl. Mater.251, 200 (1997).70. D. R. Locker and J. M. Meese, IEEE Trans. Nucl. Sci.19, 237

(1972).71. P. Agnew, Philos. Mag. A65, 355 (1992).72. F. Seitz and J. S. Koehler, Solid State Phys.2, 305 (1956).73. B. T. Kelly, Irradiation Damage to Solids(Pergamon Press, New

York, 1966).74. G. P. Pells, Radiat. Eff.64, 71 (1982).75. K. J. Caulfield, R. Cooper, and J. F. Boas, J. Am. Ceram. So

78, 1054 (1995).76. R. E. Williford, R. Devanathan, and W. J. Weber, Nucl. Instrum

Methods in Phys. Res. B141 (1998, in press).77. R. Devanathan, T. Diaz de la Rubia, and W. J. Weber, J. Nuc

Mater. 253 (1998, in press).78. D. M. Parkin and C. A. Coulter, J. Nucl. Mater.101, 261 (1981);

J. Nucl. Mater.103–104, 1315 (1981); J. Nucl. Mater.117, 340(1983).

79. W. J. Weber, R. E. Williford, and K. E. Sickafus, J. Nucl. Mater.244, 205 (1997).

80. M. Nastasi and J. W. Mayer, Mater. Sci. Rep.6, 1 (1991).81. M. Nastasi, J. W. Mayer, and J. K. Hirvonen,Ion-Solid Inter-

actions: Fundamentals and Applications(Cambridge UniversityPress, Cambridge, 1996).

82. K. C. Russell, Prog. Mater. Sci.28, 229 (1985).83. G. Martin, Phys. Rev. B30, 1424 (1984).84. W. J. Weber, R. C. Ewing, and L. M. Wang, J. Mater. Res.9,

688 (1994).85. W. J. Weber and L. M. Wang, Nucl. Instrum. Methods in Phys

Res. B91, 63 (1994).86. L. M. Wang, M. Cameron, W. J. Weber, K. D. Crowley, and R. C

Ewing, inHydroxyapatite and Related Materials,edited by P. W.Brown and B. Constants (CRC Press, Boca Raton, FL, 1994p. 243.

87. L. M. Wang and W. J. Weber, inMicrostructure of Irradiated Ma-terials, edited by I. M. Robertson,, L. E. Rehn, S. J. Zinkle, andW. J. Phythian (Mater. Res. Soc. Symp. Proc.373, Pittsburgh,PA, 1995), p. 389.

88. A. Meldrum, Ph.D. Thesis, The University of New Mexico(1997).

89. A. Meldrum, L. M. Wang, R. C. Ewing, Nucl. Instrum. Methodsin Phys. Res. B116, 220 (1996).

90. R. S. Averback, R. Benedek, and K. L. Merkle, Phys. Rev. B18,4156 (1978).

91. J. C. Bourgoin and J. W. Corbett, Radiat. Eff.36, 157 (1978).92. S. J. Zinkle, Nucl. Instrum. Methods in Phys. Res. B91, 234

(1994).


.

l.

,

93. A. E. Ringwood, S. E. Kesson, N. G. Ware, W. O. Hibberson, anA. Major, Geochemical J.13, 141 (1979).

94. A. E. Ringwood, V. M. Oversby, S. E. Kesson, W. SinclairN. Ware, W. Hibberson, and A. Major, Nucl. Chem. WasteManagement2, 287 (1981).

95. A. E. Ringwood, Am. Scientist70, 201 (1982).96. S. E. Kesson, W. J. Sinclair, and A. E. Ringwood, Nucl. Chem

Waste Management4, 259 (1983).97. H. W. Newkirk, C. L. Hoenig, F. L. Ryerson, J. D. Tewhey, G. S

Smith, C. S. Rossington, A. J. Brackmann, and A. E. RingwooCeram. Bull.61, 559 (1982).

98. G. J. McCarthy, Nucl. Technol.32, 92 (1977).99. G. J. McCarthy, J. G. Pepin, D. E. Pfoertsch, and D. R. Clark

in Ceramics in Nuclear Waste Management,edited by T. D.Chikalla and J. E. Mendel, CONF-790420 (National TechnicaInformation Service, Springfield, VA, 1979), p. 315.

100. A. K. De, B. Luckscheiter, W. Lutze, G. Malow, and E.Schiewer, Ceram. Bull.55, 500 (1976).

101. A. K. De, B. Luckscheiter, W. Lutze, G. Malow, E. Schiewerand S. Tymochowicz, inManagement of Radioactive Wastes fromthe Nuclear Fuel Cycle, Vol. II,IAEA-SM-207 (InternationalAtomic Energy Agency, Vienna, 1976), p. 63.

102. P. J. Hayward, inRadioactive Waste Forms for the Future,editedby W. Lutze and R. C. Ewing (North-Holland, Amsterdam1988), p. 427.

103. A. E. Ringwood, V. M. Oversby, and W. Sinclair, inScientificBasis for Nuclear Waste Management, Vol. 2,edited by C. J. M.Northrup, Jr. (Plenum Press, New York, 1980), p. 273.

104. K. D. Reeve and J. L. Woolfrey, J. Aust. Ceram. Soc.16, 10(1980).

105. K. D. Reeve, D. M. Levins, E. J. Ramm, J. L. Woolfrey, W. JBuykx, R. K. Ryan, and J. F. Chapman, inWaste Management’81, edited by R. G. Post (Arizona Board of Rgents, TucsonAZ, 1981), p. 249.

106. J. L. Woolfrey, K. D. Reeve, and D. J. Cassidy, J. Nucl. Mate108 & 109, 739 (1982).

107. K. A. Boult, J. T. Dalton, J. P. Evans, A. R. Hall, A. J. InnsJ. A. C. Marples, and E. L. Paige,The Preparation of Fully ActiveSynroc and Its Radiation Stability–Final Report, October 1988AERE-R-13318 (Harwell Laboratory, Harwell, UK, 1988).

108. E. Vernaz, A. Loida, G. Malow, J. A. C. Marples, and Hj. Matzkein Proc. 3rd EC Conf. on Radioactive Waste Management anDisposal,edited by L. Cecille (Elsevier, London, 1991), p. 302

109. Hj. Matzke, inIon Beam Modification of Insulators,edited by P.Mazzoldi and G. Arnold (Elsevier Science, Amsterdam, 1997p. 501.

110. A. G. Solomah and Hj. Matzke, inScientific Basis for NuclearWaste Management XII,edited by W. Lutze and R. C. Ewing(Mater. Res. Soc. Symp. Proc.127, Pittsburgh, PA, 1989),p. 241.

111. W. J. Weber, inScientific Basis for Nuclear Waste ManagemenVII, edited by C. M. Jantzen, J. A. Stone, and R. C. Ewing (MateRes. Soc. Symp. Proc.44, Pittsburgh, PA, 1985), p. 671.

112. F. W. Clinard, Jr., L. W. Hobbs, C. C. Land, D. E. Peterson, D.Rohr, and R. B. Roof, J. Nucl. Mater.105, 248 (1982).

113. F. W. Clinard, Jr., Am. Ceram. Soc. Bull.65 (8), 1181 (1986).114. F. W. Clinard, Jr., D. E. Peterson, D. L. Rohr, and L. W. Hobb

J. Nucl. Mater.126, 245 (1984).115. F. W. Clinard, Jr., D. L. Rohr, and R. B. Roof, Nucl. Instrum

Methods in Phys. Res. B1, 581 (1984).116. W. J. Weber, J. W. Wald, and Hj. Matzke, J. Nucl. Mater.138,

196 (1986).117. H. Mitzmura, S. Matsumoto, T. Miyazaki, T. J. White, K.

Nukaga, Y. Togashi, T. Sagawa, S. Tashiro, D. M. Levins, anA. Kikuchi, J. Am. Ceram. Soc.73, 3433 (1990).

, No. 6, Jun 1998 1479


R

a

a

.

.

,

d

.

r

,s

TX

118. H. Mitamura, S. Matsumoto, K. P. Hart, T. Miyazaki, E.Vance, Y. Tamura, Y. Togashi, and T. J. White, J. Am. CeraSoc. 75, 392 (1992).

119. H. Mitamura, S. Matsumoto, M. W. A. Stewart, T. Tsuboi, MHashimoto, E. R. Vance, K. P. Hart, Y. Togashi, H. KanazawC. J. Ball, and T. J. White, J. Am. Ceram. Soc.77, 2255 (1994).

120. R. G. Dosch, T. J. Headley, and P. Hlava, J. Am. Ceram. S67, 354 (1984).

121. R. G. Dosch, C. J. Northrup, and T. J. Headley, J. Am. CerSoc. 68, 330 (1985).

122. J. M. Rusin, W. J. Gray, and J. W. Wald,Multibarrier WasteForms, Part II; Characterization and Evaluation,PNL-2668-2(Pacific Northwest Laboratory, Richland, WA, 1979).

123. R. P. Turcotte, J. W. Wald, F. P. Roberts, J. M. Rusin, andLutze, J. Am. Ceram. Soc.65, 589 (1982).

124. W. J. Weber, R. P. Turcotte, L. R. Bunnell, F. P. Roberts,J. H. Westsik, inCeramics in Nuclear Waste Management,editedby T. D. Chikalla and J. E. Mendel, CONF-790420 (NationTechnical Information Service, Springfield, VA, 1979), p. 294

125. J. L. Routbort, P. Offermann, and Hj. Matzke, inScientific Basisfor Nuclear Waste Management VI,edited by D. G. Brookins(Mater. Res. Soc. Symp. Proc.15, Pittsburgh, PA, 1983), p. 415

126. G. Malow, J. A. C. Marples, and C. Sombret, inRadioactiveWaste Management and Disposal,edited by R. Simon and SOrlowski (Harwood Academic Publishers, Chur, Switzerlan1980), p. 341.

127. E. R. Vance, F. G. Karioris, L. Cartz, and M. S. Wong,Nuclear Waste Management,edited by G. G. Wicks, and W. A.Ross (Advances in Ceramics8, American Ceramic SocietyWesterville, OH, 1984), p. 62.

128. P. J. Hayward, inRadioactive Waste Forms for the Future,editedby W. Lutze and R. C. Ewing (North-Holland, Amsterdam1988), p. 457.

129. B. C. Chakoumakos, J. Solid State Chem.53, 120 (1984).130. B. C. Chakoumakos, inMcGraw-Hill Yearbook of Science an

Technology 1987,edited by S. P. Parker (McGraw-Hill, NewYork, 1986), p. 393.

131. B. C. Chakoumakos and R. C. Ewing, inScientific Basis forNuclear Waste Management VIII,edited by C. M. Jantzen, J. AStone, and R. C. Ewing (Mater. Res. Soc. Symp. Proc.44,Pittsburgh, PA, 1985), p. 641.

132. S. S. Shoup, C. E. Bamberger, and R. G. Haire, J. Am. CeSoc. 79, 1489 (1996).

133. W. J. Weber, J. W. Wald, and Hj. Matzke, Mater. Lett.3, 173(1985).

134. G. R. Lumpkin and R. C. Ewing, inScientific Basis for NuclearWaste Management VIII,edited by C. M. Jantzen, J. A. Stoneand R. C. Ewing (Mater. Res. Soc. Symp. Proc.44, Pittsburgh,PA, 1985), p. 647.

135. G. R. Lumpkin and R. C. Ewing, inMicrobeam Analysis–1986edited by A. D. Romig, Jr., and W. F. Chambers (San FranciPress, San Francisco, 1986), p. 145.

136. G. R. Lumpkin, E. M. Foltyn, and R. C. Ewing, J. Nucl. Mate139, 113 (1986).

137. G. R. Lumpkin and R. C. Ewing, Am. Mineral.81, 1237 (1997).138. R. C. Ewing and T. J. Headley, J. Nucl. Mater.119, 102 (1983).139. G. R. Lumpkin, K. L. Smith, and R. Gier´e, Micron28, 57 (1997).140. R. C. Ewing and T. J. Headley, J. Nucl. Mater.119, 102 (1983).141. G. R. Lumpkin, K. L. Smith, M. G. Blackford, R. Giere, and C.

Williams, in Scientific Basis for Nuclear Waste Management Xedited by I. G. McKinley and C. McCombie (Mater. Res. SoSymp. Proc.506, Warrendale, PA 1998), p. 215.

142. W. J. Weber, N. J. Hess, and G. D. Maupin, Nucl. InstruMethods in Phys. Res. B65, 102 (1992).

143. W. J. Weber and N. J. Hess, Nucl. Instrum. Methods in PhRes. B80/81, 1245 (1993).


.m.

.a,

oc.

m.

W.

nd

al.

d,

in

,

am.

,

co

r.

.I,

c.

m.

ys.

144. G. R. Lumpkin, K. L. Smith, and R. Blake, inScientific Basis forNuclear Waste Management XIX,edited by W. M. Murphy andD. A. Knecht (Mater. Res. Soc. Symp. Proc.412, Pittsburgh,PA, 1996), p. 329.

145. G. R. Lumpkin and R. C. Ewing, Am. Mineral.77, 179 (1992).146. G. R. Lumpkin and R. C. Ewing, Am. Mineral.80, 732 (1995).147. W. Sinclair and A. E. Ringwood, Geochem. J.15, 229 (1981).148. J. Rankin, L. W. Hobbs, L. A. Boatner, and C. W. White, Nucl.

Instrum. Methods in Phys. Res. B32, 28 (1988).149. J. Rankin, J. C. McCallum, and L. A. Boatner, J. Appl. Phys.78,

1519 (1995).150. J. Rankin, J. C. McCallum, and L. A. Boatner, J. Mater. Res.7,

717 (1992).151. J. Rankin, B. W. Sheldon, and L. A. Boatner, J. Mater. Res.9,

3113 (1994).152. A. Meldrum, L. A. Boatner, and R. C. Ewing, inPhase Trans-

formations and Systems Driven Far From Equilibrium,edited byE. Ma, P. Bellon, M. Atzmon, and R. Trivedi (Mater. Res. Soc.Symp. Proc.481, Warrendale, PA, 1998), in press.

153. A. B. Harker and J. F. Flintoff, J. Am. Ceram. Soc.68, 159(1985).

154. A. B. Harker and J. F. Flintoff, J. Am. Ceram. Soc.73, 1901(1990).

155. W. J. Weber, J. Mater. Res.5, 2687 (1990).156. W. J. Weber, Radiat. Eff.115, 341 (1991).157. E. B. Anderson, B. E. Burakov, and E. M. Pazukhin, Radiochim.

Acta 60, 149 (1993).158. B. E. Burokov, E. B. Anderson, B. Ya. Galkin, V. A. Starchenko,

and V. G. Vasiliev, inDisposal of Weapons Plutonium,edited byE. R. Merz and C. E. Walter (Kluwer Academic Publishers, TheNetherlands, 1996), p. 85.

159. A. Abdelouas, J. L. Crovisier, W. Lutze, B. Grambow, J. C. Dran,and R. Muller, J. Nucl. Mater.240, 100 (1997).

160. H. D. Holland and D. Gottfried, Acta Crystallogr.8, 291 (1955).161. T. Murakami, B. C. Chakoumakos, R. C. Ewing, G. R. Lumpkin,

and W. J. Weber, Am. Mineral.76, 1510 (1991).162. W. J. Weber, Nucl. Instrum. Metods in Phys. Res. B65, 88

(1992).163. W. J. Weber, J. Amer. Ceram. Soc.76, 1729 (1993).164. W. J. Weber, R. C. Ewing, and A. Meldrum, J. Nucl. Mater.250,

147 (1997).165. A. Meldrum, L. A. Boatner, W. J. Weber, and R. C. Ewing,

Geochim. Cosmochim. Acta (unpublished).166. D. J. Wronkiewicz, S. F. Wolf, and T. S. DiSanto, inScientific

Basis for Nuclear Waste Management XIX,edited by W. M.Murphy and D. A. Knecht (Mater. Res. Soc. Symp. Proc.412,Pittsburgh, PA, 1996), p. 345.

167. C. M. Jantzen and F. P. Glasser, Ceram. Bull.58, 459 (1979).168. W. L. Ebert, J. K. Bates, and W. L. Bourcier, Waste Mgmt.11,

205 (1991).169. W. L. Gong, R. C. Ewing, L. M. Wang, E. Vernaz, J. K. Bates,

and W. L. Ebert, inScientific Basis for Nuclear Waste Manage-ment XIX,edited by W. M. Murphy and D. A. Knecht (Mater.Res. Soc. Symp. Proc.412, Pittsburgh, PA, 1996), p. 197.

170. R. Bros, J. Carpena, V. Sere, and A. Beltritti, Radiochim. Acta74, 277 (1996).

171. F. Audubert, J. Carpena, J. L. Lacout, and F. Tetard, Solid StateIonics 95, 113 (1997).

172. L. Boyer, J. Carpena, and J. L. Lacout, Solid State Ionics95,121 (1997).

173. D. McConnell,Apatite (Springer-Verlag, New York, 1973).174. M. L. Lindberg and B. Ingram, U.S. Geol. Surv., Prof. Paper

501-B (1964), p. 64.175. W. J. Weber, J. Am. Ceram. Soc.65 (11), 544 (1982).176. W. J. Weber, Radiat. Eff.77, 295 (1983).

13, No. 6, Jun 1998


t

e

r

e

-

s

g

A

u

r

.

,

.

i,

.

,

,

177. J. W. Wald and W. J. Weber, inNuclear Waste Managemenedited by G. G. Wicks and W. A. Ross (Advances in Ceram8, American Ceramic Society, Westerville, OH, 1984), p. 71.

178. W. J. Weber and Hj. Matzke, Mater. Lett.5, 9 (1986).179. W. J. Weber, R. K. Eby, and R. C. Ewing, J. Mater. Res.6, 1334

(1991).180. A. Meldrum, L. M. Wang, and R. C. Ewing, Am. Mineral.82,

858 (1997).181. E. R. Vance, C. D. Cann, and P. G. Richardson, Radiat. Eff.98,

71 (1986).182. E. R. Vance and J. B. Metson, Phys. Chem. Minerals12, 255

(1985).183. F. C. Hawthorne, L. A. Groat, M. Raudsepp, N. A. Ball, M. K

mata, F. D. Spike, R. Gaba, N. M. Halden, G. R. Lumpkin, R.Ewing, R. B. Greegor, F. W. Lytle, T. S. Ercit, G. R. RossmaF. J. Wicks, R. A. Ramik, B. L. Sherriff, M. E. Fleet, and CMcCammon, Am. Mineral.76, 370 (1991).

184. L. A. Boatner, G. W. Beall, M. M. Abraham. C. B. Finch, P. GHuray, and M. Rappaz, inScientific Basis for Nuclear WastManagement,edited by C. J. M. Northrup, Jr. (Plenum PresNew York, 1980), Vol. 2, p. 289.

185. L. A. Boatner and B. C. Sales, inRadioactive Waste Forms fothe Future,edited by W. Lutze and R. C. Ewing (North-HollandAmsterdam, 1988), p. 495.

186. K. Bakker, H. Hein, R. J. M. Konings, R. R. van der Laan, HMatzke, and P. van Vlaanderen, J. Nucl. Mater.252, 228 (1998).

187. R. C. Ewing, Am. Mineral.60, 728 (1975).188. R. C. Ewing and R. F. Haaker, Nucl. Chem. Waste Managem

1, 51 (1980).189. A. Meldrum, L. A. Boatner, and R. C. Ewing, Phys. Rev. B56,

13805 (1997).190. A. Meldrum, L. A. Boatner, and R. C. Ewing, J. Mater. Res.12,

1816 (1997).191. R. Roy, E. R. Vance, and J. Alamo, Mater. Res. Bull.17, 585

(1982).192. B. E. Scheetz and R. Roy, inRadioactive Waste Forms for th

Future, edited by W. Lutze and R. C. Ewing (North-HollandAmsterdam, 1988), p. 596.

193. H. T. Hawkins, B. E. Scheetz, and G. D. Guthrie, Jr., inScientificBasis for Nuclear Waste Management XX,edited by W. J. Grayand I. R. Triay (Mater. Res. Soc. Symp. Proc.465, Pittsburgh,PA, 1997), p. 387.

194. V. N. Zyryanov and E. R. Vance, inScientific Basis for NuclearWaste Management XX,edited by W. J. Gray and I. R. Triay(Mater. Res. Soc. Symp. Proc.465, Pittsburgh, PA, 1997),p. 409.

195. A. I. Orlova, Y. F. Volkov, R. F. Melkaya, L. Y. Masterova, I. AKulikov, and V. A. Alferov, Radiochemistry36, 322 (1994).

196. A. I. Kryukova, I. A. Kulikov, G. Y. Artem’eva, O. V. Pechenevskaya, and V. A. Alferov, Radiokhimiya34 (6), 82 (1992).

197. D. M. Chapman and L. Lawrence Roe, Zeolites10, 730 (1990).198. R. G. Dosch, N. E. Brown, H. P. Stephens, and R. G. Antho

in Waste Management ’93,edited by R. G. Post and M. E. Wack(Arizona Board of Regents, Tucson, AZ, 1993), p. 1751.

199. R. G. Anthony, C. V. Phillip, and R. G. Dosch, Waste Manament 13, 503 (1993).

200. M. L. Balmer, Q. Huang, W. Wong-Ng, R. S. Roth, andSantoro, J. Solid State Chem.130, 97 (1997).

201. E. A. Klavetter, N. E. Brown, D. Trudell, R. G. Anthony, D. Gand C. Thibaud-Erkey, inWaste Management ’94,edited by R. G.Post and M. E. Wacks (Arizona Board of Regents, Tucson, A1994), p. 709.

202. W. J. Weber, G. J. Exarhos, and L. M. Wang, inMicrostructureof Irradiated Materials,edited by I. M. Robertson, L. E. RehnS. J. Zinkle, and W. J. Phythian (Mater. Res. Soc. Symp. P373, Pittsburgh, PA, 1995), p. 311.


,ics

i-C.n,.

.

s,

,

j.

ent

,

.

ny,

e-

.

,

Z,

,oc.

203. L. W. Hobbs, J. de Phys.37, C7-3 (1976).204. A. E. Hughes, Radiat. Eff.74, 57 (1983).205. A. E. Hughes, Radiat. Eff.97, 161 (1986).206. M. R. Pascucci, J. L. Hutchison, and L. W. Hobbs, Radiat. Eff

74, 219 (1983).207. H. Matsui, M. W. Guinan, T. Iida, and M. Iseki, J. Nucl. Mater.

133/134, 718 (1985).208. C. Kinoshita and K. Nakai, Jpn. J. Appl. Phys. Series2, 105

(1989).209. C. Kinoshita, T. Sonoda, and A. Manabe, Philos. Mag. A (1998

in press).210. C. Kinoshita, K. Fukumoto, K. Fukuda, F. A. Gardner, and G. W.

Hollenberg, J. Nucl. Mater.219, 143 (1995).211. F. W. Clinard, Jr., G. F. Hurley, and L. W. Hobbs, J. Nucl. Mater.

108/109, 655 (1982).212. K. E. Sickafus, A. C. Larson, N. Yu, M. Nastasi, G. W. Hollen-

berg, F. A. Garner, and R. C. Bradt, J. Nucl. Mater.219, 128(1995).

213. G. R. Lumpkin and R. C. Ewing, Phys. Chem. Minerals16, 2(1988).

214. A. D. Brailsford and R. Bullough, J. Nucl. Mater.44, 121 (1972).215. U. Jain and A. B. Lidiard, Philos. Mag.35, 245 (1977).216. P. G. Shewmon,Diffusion in Solids(McGraw-Hill, New York,

1963).217. R. Freer, Contrib. Mineral. Petrol.76, 440 (1981).218. Hj. Matzke, Radiat. Eff.75, 317 (1983).219. B. Lesage, A. M. Huntz, and G. Petot-Ervas, Radiat. Eff.75,

283 (1983).220. E. A. Cooper and M. Nastasi, J. Amer. Ceram. Soc.77, 268

(1994).221. E. A. Cooper and M. Nastasi, Appl. Phys. Lett.64 (22), 2958

(1994).222. E. A. Cooper, M. Nastasi, and K. E. Sickafus, inBeam-Solid

Interactions: Fundamentals and Applications,edited by M. A.Nastasi, L. R. Harriott, N. Herbots, and R. S. Averback (Mater.Res. Soc. Symp. Proc.279, Pittsburgh, PA, 1993), p. 457.

223. E. A. Cooper and M. Nastasi, Nucl. Instrum. Methods in PhysRes. B91, 558 (1994).

224. S. J. Zinkle, J. Nucl. Mater.219, 113 (1995).225. R. Devanathan, K. E. Sickafus, W. J. Weber, and M. Nastas

J. Nucl. Mater253 (1998, in press).226. R. P. Turcotte and T. D. Chikalla, Radiat. Eff.19, 99 (1973).227. W. J. Weber, Radiat. Eff.70, 217 (1983).228. W. J. Nellis, Inorg. Nucl. Chem. Letters13, 393 (1977).229. T. D. Chikalla and R. P. Turcotte, Radiat. Eff.19, 93 (1973).230. M. Wittels and F. A. Sherrill, Phys. Rev.93, 1117 (1954).231. W. Primak, Phys. Rev.110, 1240 (1958).232. W. K. Woods, L. F. Bupp, and J. F. Fletcher, inProc. Int. Conf.

on Peaceful Uses of Atomic Energy(United Nations, New York,1956), Vol. 7, p. 455.

233. P. G. Klemens, F. W. Clinard, Jr., and R. J. Livak, J. Appl. Phys62, 2062 (1987).

234. E. M. Foltyn, F. W. Clinard, Jr., J. Rankin, and D. E. PetersonJ. Nucl. Mater.136, 97 (1985).

235. D. E. Peterson and F. W. Clinard, Jr., inScientific Basis forNuclear Waste Management VI,edited by D. G. Brookins (Mater.Res. Soc. Symp. Proc.15, Pittsburgh, PA, 1983), p. 465.

236. S. Ellsworth, A. Navrotsky, and R. C. Ewing, Phys. Chem.Minerals 21, 140 (1994).

237. G. R. Lumpkin, R. C. Ewing, B. C. Chakoumakos, R. B. GreegorF. W. Lytle, E. M. Foltyn, F. W. Clinard, Jr., L. A. Boatner, andM. M. Abraham, J. Mater. Res.1, 564 (1986).

238. G. R. Lumpkin, E. M. Foltyn, and R. C. Ewing, J. Nucl. Mater.139, 113 (1986).

239. T. C. Ehlert, K. Appaji Gowda, F. G. Karioris, and L. Cartz,Radiat. Eff. 70, 173 (1983).

13, No. 6, Jun 1998 1481


a

i

.

r

t

n

s

s

.

i-

.

,

)

240. W. A. Coghlan and F. W. Clinard, Jr., J. Less-Common Met140, 255 (1988).

241. Hj. Matzke and M. Kinoshita, J. Nucl. Mater.247, 108 (1997).242. W. C. Brøgger, “Amorf,” inSalmonsens store illustrede Konver

sationslexikon1, 742 (1893).243. M. V. Stackelberg and E. Rottenbach, Z. Kristallogr.102, 173

(1939).244. M. V. Stackelberg and E. Rottenbach, Z. Kristallogr.102, 207

(1939).245. R. B. Greegor, F. W. Lytle, B. C. Chakoumakos, G. R. Lumpk

R. C. Ewing, C. L. Spiro, and J. Wong, inScientific Basis forNuclear Waste Management X,edited by J. K. Bates and W. B.Seefeldt (Mater. Res. Soc. Symp. Proc.84, Pittsburgh, PA, 1987),p. 645.

246. F. Farges, R. C. Ewing, and G. E. Brown, Jr., J. Mater. Res8,1983 (1993).

247. F. Farges, Am. Mineral.82, 44 (1997).248. R. B. Greegor, F. W. Lytle, R. J. Livak, and F. W. Clinard, J

J. Nucl. Mater.152, 270 (1988).249. H. M. Naguib and R. Kelly, Radiat. Eff.25, 1 (1975).250. L. W. Hobbs, J. Non-Cryst. Solids182, 27 (1995).251. L. W. Hobbs, Nucl. Instrum. Methods in Phys. Res. B91, 30

(1994).252. L. W. Hobbs, A. N. Sreeram, B. Berger and C. E. Jesurum, Nu

Instrum. Methods in Phys. Res. B116, 17 (1996).253. Hj. Matzke and J. L. Whitton, Canadian J. Phys.44, 995 (1966).254. R. C. Ewing, L. M. Wang, and W. J. Weber, inMicrostructure of

Irradiated Materials,edited by I. M. Robertson, L. E. Rehn, S. JZinkle, and W. J. Phythian (Mater. Res. Soc. Symp. Proc.373,Pittsburgh, PA, 1995), p. 347.

255. Hj. Matzke, Defects and Diffusion Forum83, 111 (1992).256. R. Devanathan, W. J. Weber, and T. Diaz de la Rubia, Nu

Instrum. Methods in Phys. Res. B141 (1998, in press).257. T. Diaz de la Rubia, M.-J. Caturla, and M. Tobin, inMicrostruc-

ture of Irradiated Materials,edited by I. M. Robertson, L. E.Rehn, S. J. Zinkle, and W. J. Phythian (Mater. Res. Soc. SymProc. 373, Pittsburgh, PA, 1995), p. 555.

258. H. Inui, H. Mori, and H. Fujita, Philos. Mag. B61, 107 (1990).259. A. Matsunaga, C. Kinoshita, K. Nakai, and Y. Tomokiyo

J. Nucl. Mater.179–181, 457 (1991).260. H. Inui, H. Mori, A. Suzuki, and H. Fujita, Philos. Mag. B65,

1 (1992).261. H. Inui, H. Mori, and T. Sakata, Philos. Mag. B66, 737 (1992).262. N. Yu, R. Devanathan, K. E. Sickafus, and M. Nastasi, J. Ma

Res. 12, 1766 (1997).263. N. Yu, K. E. Sickafus, and M. Nastasi, Philos. Mag.70, 235

(1994).264. S. X. Wang, Ph.D. Dissertation, The University of New Mexic

(1997).265. L. M. Wang, S. X. Wang, W. L. Gong, and R. C. Ewing, i

Microstructure Evolution During Irradiation,edited by I. M.Robertson, G. S. Was, L. W. Hobbs, and T. Diaz de la Rub(Mater. Res. Soc. Symp. Proc.439, Pittsburgh, PA, 1997),p. 583.

266. S. X. Wang, L. M. Wang, and R. C. Ewing, inMicrostructureEvolution During Irradiation,edited by I. M. Robertson, G. S.Was, L. W. Hobbs, and T. Diaz de la Rubia (Mater. Res. SoSymp. Proc.439, Pittsburgh, PA, 1997), p. 619.

267. S. X. Wang, L. M. Wang, and R. C. Ewing, Nucl. InstrumMethods in Phys. Res. B127/128, 186 (1997).

268. S. X. Wang, L. M. Wang, R. C. Ewing, and R. H. DoremuJ. Appl. Phys.81, 587 (1997).

269. S. X. Wang, L. M. Wang, R. C. Ewing, and R. H. DoremuJ. Non-Cryst. Solid (unpublished).

270. W. L. Gong, L. M. Wang, R. C. Ewing, and J. Zhang, Phys. ReB 54, 3800 (1996).


ls

-

n,

.,

cl.

.

cl.

p.

,

er.

o

ia

c.

.

,

,

v.

271. W. J. Weber and L. M. Wang, Nucl. Instrum. Methods in Phys.Res. B 106, 298 (1995).

272. W. J. Weber, L. M. Wang, and N. Yu, Nucl. Instrum. Methods inPhys. Res. B116, 322 (1996).

273. A. T. Motta and D. R. Olander, Acta Met.38 (11), 2175 (1990).274. J. A. Faldowski, A. T. Motta, L. M. Howe, and P. R. Okamoto,

J. Appl. Phys.80, 729 (1996).275. A. T. Motta, J. Nucl. Mater.244, 227 (1997).276. L. M. Wang, W. L. Gong, R. C. Ewing, and W. J. Weber, in

Thermodynamics and Kinetics of Phase Transformations,editedby J. S. Im, B. Park, A. L. Greer, and G. B. Stephenson (Mater.Res. Soc. Symp. Proc.398, Pittsburgh, PA, 1996), p. 233.

277. A. T. Motta and C. Lemaignan, J. Nucl. Mater.195, 277 (1992).278. J. F. Gibbons, Proc. IEEE60, 1062 (1972).279. J. L. Brimhall, H. E. Kissinger, and A. R. Pelton, Radiat. Eff.70,

241 (1985).280. A. T. Motta, D. R. Olander and A. J. Machiels, inEffects of

Radiation on Materials,edited by N. H. Packan, R. E. Stoller, andA. S. Kumar, ASTM STP 1046 (American Society for Testingand Materials, Philadelphia, PA, 1989), p. 457.

281. E. P. Simonen, Nucl. Instrum. Methods in Phys. Res. B16, 198(1986).

282. N. Q. Lam and P. R. Okamoto, Mater. Res. Soc. Bull.XIX (7),41 (1994).

283. R. C. Ewing and L. M. Wang, Nucl. Instrum. Methods in Phys.Res. B65, 319 (1992).

284. S. J. Zinkle and L. L. Snead, Nucl. Instrum. Methods in Phys.Res. B116, 92 (1996).

285. C. J. McHargue, P. S. Sklad, and C. W. White, Nucl. Instrum.Methods in Phys. Res. B46, 79 (1990).

286. C. J. McHargue, G. C. Farlow, C. W. White, J. M. Williams, B. R.Appleton, and H. Naramoto, Mater. Sci. Eng.69, 123 (1985).

287. F. F. Morehead and B. L. Crowder, Radiat. Eff.6, 27 (1970).288. J. R. Dennis and E. B. Hale, J. Appl. Phys.49 (3), 1119 (1978).289. K. A. Jackson, J. Mater. Res.3, 1218 (1988).290. D. F. Pedraza and L. K. Mansur, Nucl. Instrum. Methods in Phys

Res. B16, 203 (1986).291. C. H. Woo, J. Nucl. Mater.159, 237 (1988).292. M. R. Pascucci, Ph.D. Dissertation, Case Western Reserve Un

versity (1983).293. L. M. Wang, W. L. Gong, R. C. Ewing, and W. J. Weber, in

Thermodynamics and Kinetics of Phase Transformations,editedby J. S. Im, B. Park, A. L. Greer, and G. B. Stephenson (Mater.Res. Soc. Symp. Proc.398, Pittsburgh, PA, 1996), p. 233.

294. L. A. Bursill and G. Braunshausen, Philos. Mag. A62, 395(1990).

295. R. L. Fleischer, P. B. Price, and R. M. Walker, J. Geophys. Res69, 4885 (1964).

296. A. S. Sandhu, L. Singh, R. C. Ramola, S. Singh, and H. S. VirkNucl. Instrum. Methods in Phys. Res. B46, 122 (1990).

297. H. S. Virk, Radiat. Eff.133, 87 (1995).298. E. R. Vance and B. W. Anderson, Mineralogical Mag.38, 605

(1972).299. R. Evron, Y. Cohen, O. Regev, Y. Eyal, Hj. Matzke, and K.

Tinschert, Proc. Nucl. Soc. Israel18, VIII-2 (1994).300. Hj. Matzke, Y. Eyal, and A. van Veene, private communication.301. T. J. Headley, R. C. Ewing, and R. F. Haaker, Nature (London

293, 449 (1981).302. J. Janeczek and R. C. Ewing, J. Nucl. Mater.185, 66 (1991).303. V. F. Chkuaseli and Hj. Matzke, J. Nucl. Mater.201, 92 (1993).304. V. F. Chkuaseli and Hj. Matzke, J. Nucl. Mater.223, 61 (1995).305. F. W. Clinard, Jr., D. S. Tucker, G. F. Hurley, C. D. Kise, and J.

Rankin, inScientific Basis for Nuclear Waste Management VIII,edited by C. M. Jantzen, J. A. Stone, and R. C. Ewing (Mater.Res. Soc. Symp. Proc.44, Pittsburgh, PA, 1985), p. 664.

306. H. Ozkan, J. Appl. Phys.47, 4772 (1976).

3, No. 6, Jun 1998


A

.

s

s

.

307. B. C. Chakoumakos, W. C. Oliver, G. R. Lumpkin, and R.Ewing, Radiat. Eff.118, 393 (1991).

308. B. C. Chakoumakos, T. Murakami, G. R. Lumpkin, and R.Ewing, Science236, 1556 (1987).

309. W. C. Oliver, J. C. McCallum, B. C. Chakoumakos, and L.Boatner, Radiat. Eff.132, 131 (1994).

310. Hj. Matzke and J. Spino, J. Nucl. Mater.248, 170 (1997).311. H. E. Schmidt, J. Richter, Hj. Matzke, and J. van Geel,

Thermal Conductivity 22(Technomic Publ. Co., Lancaster, PA1994), p. 920.

312. R. T. Pidgeon, J. R. O’Neil, and L. T. Silver, Science154, 1538(1986).

313. R. C. Ewing, R. F. Haaker, and W. Lutze, inScientific Basis forNuclear Waste Management V,edited by W. Lutze (Mater. ResSoc. Symp. Proc.11, Pittsburgh, PA, 1982), p. 389.

314. M. P. Tole, Geochim. Cosmochim. Acta49, 453 (1985).315. K. Suzuki, Geochem. J.21, 173 (1987).316. A. K. Sinha, D. M. Wayne, and D. A. Hewitt, Geochim. Co

mochim. Acta56, 3551 (1992).317. T. E. Krogh and G. L. Davis, Carnegie Institution of Washingt

Yearbook73, 560 (1973).318. T. E. Krogh, Geochim. Cosmochim. Acta46, 637 (1982).319. T. E. Krogh, EOS, Trans. Amer. Geophysical Union56 (6), 472

(1975).320. T. E. Krogh and G. L. Davis, Carnegie Institution of Washingt

Yearbook74, 619 (1975).321. B. C. Sales, C. W. White, and L. A. Boatner, Nucl. Chem. Wa

Management4, 281 (1983).322. Y. Eyal and A. Kaufmann, Nucl. Technol.58, 77 (1982).323. Y. Eyal and R. L. Fleischer, Geochim. Cosmochim. Acta49,

1155 (1985).324. G. R. Lumpkin, Y. Eyal, and R. C. Ewing, J. Mater. Res.3, 357

(1988).325. Y. Eyal and R. C. Ewing, inProc. Int. Conf. on Nuclear Waste

Management and Environmental Remediation, Vol. 1,edited byD. Alexandre, R. Baker, R. Kohout, and J. Marek (ASME PreNew York, 1993), p. 191.

326. Hj. Matzke, J. Nucl. Mater.190, 101 (1992).327. D. V. Stevanovic, D. A. Thompson, and E. R. Vance, J. Nu

Mater. 161, 169 (1989).328. T. Murakami, T. Banba, M. J. Jercinovic, and R. C. Ewing,

Scientific Basis for Nuclear Waste Management XII,edited byW. Lutze and R. C. Ewing (Mater. Res. Soc. Symp. Proc.127,Pittsburgh, PA, 1989), p. 65.

329. Computer Simulation of Radiation Effects in Solids,edited byT. Diaz de la Rubia, G. H. Gilmer, and M-J. Caturla (NorthHolland, Amsterdam, 1995).

330. J. B. Adams, A. Rockett, J. Kieffer, W. Xu, M. Nomura, K. AKiliam, D. F. Richards, and R. Ramprasad, J. Nucl. Mater.216,265 (1994).

331. A. M. Stoneham,Theory of Defects in Solids(Oxford UniversityPress, 1975).

332. J. H. Harding, Rep. Prog. Phys.53, 1403 (1990).333. C. R. A. Catlow, Ann. Rev. Mater. Sci.16, 517 (1986).334. C. R. A. Catlow, R. G. Bell, and J. D. Gale, J. Mater. Chem4

(6), 781 (1994).335. N. F. Mott and M. J. Littleton, Faraday Trans.34, 389 (1938).336. M. Cherry, M. S. Islam, and C. R. A. Catlow, J. Solid Sta

Chem.118, 125 (1995).337. C. R. A. Catlow, inNonstoichiometric Oxides,edited by O. T.

Sorensen (Academic Press, New York, 1981), p. 61.338. C. R. A. Catlow, K. M. Diller, and L. W. Hobbs, Philos. Mag. A

42 (2), 123 (1980).339. C. R. A. Catlow, J. Phys. C. Condensed Matter12 (6), 969

(1979).


C.

C.

.

in,

-

on

on

ste

s,

cl.

in

-

.

te

340. R. W. Grimes and C. R. A. Catlow, Phil. Trans. Roy. Soc. Lond.A 341, 10,389 (1992).

341. R. W. Grimeset al., private communication.342. D. H. Gay and A. L. Rohl, J. Chem. Soc., Faraday Trans.5, 925

(1995).343. K. V. Wright and C. R. A. Catlow, J. Phys. Chem. Minerals23,

38 (1996).344. M. P. Allen and D. J. Tildesley,Computer Simulation of Liquids

(Oxford University Press, Oxford, 1987).345. T. S. Bush, J. D. Gale, C. R. A. Catlow, and P. D. Battle, J. Mater.

Chem. 4, 831 (1994).346. K. V. Damodaran, V. S. Nagarayan, and K. J. Rao, J. Non-Cryst.

Solids 124, 233 (1990).347. G. D. Price, S. C. Parker, and M. Leslie, Min. Mag.51, 157

(1987).348. K. V. Wright and C. R. A. Catlow, J. Phys. Chem. Minerals20,

515 (1994).349. T. F. Soules, J. Non-Cryst. Solids49, 29 (1982).350. J. B. Gibson, A. N. Goland, M. Milgram, and G. H. Vineyard,

Phys. Rev.120, 1229 (1960).351. M. J. Norgett, M. T. Robinson, and I. M. Torrens, Nucl. Eng.

Design 33, 50 (1975).352. D. N. Seidman, R. S. Averback, and R. Benedek, Phys. Status

Solidi 144, 85 (1987).353. J. M. Delaye and D. Ghaleb, J. Nucl. Mater.244, 22 (1997).354. J. M. Delaye and D. Ghaleb, inScientific Basis for Nuclear Waste

Management XX,edited by W. J. Gray and I. R. Triay (Mater.Res. Soc. Symp. Proc.465, Pittsburgh, PA, 1997), p. 633.

355. J.-P. Crocombette and D. Ghaleb, inScientific Basis for NuclearWaste Management XXI,edited by I. G. McKinley and C. Mc-Combie (Mater. Res. Soc. Symp. Proc.506, Warrendale, PA,1998), p. 101.

356. M.-J. Carturla and T. Diaz de la Rubia, inMicrostructureEvolution during Irradition, edited by I. M. Robertson, G. S.Was, L. W. Hobbs, and T. Diaz de la Rubia (Mater. Res. Soc.Symp. Proc.439, Pittsburgh, PA, 1997), p. 125.

357. P. Vashishta, A. Nakano, R. K. Kalia and I. Ebbsj¨o, J. Non-Cryst.Solids 182, 59 (1995).

358. T. Diaz de la Rubia, M.-J. Caturla, and M. Tobin, inMicrostruc-ture of Irradiated Materials,edited by I. M. Robertson, L. E.Rehn, S. J. Zinkle, and W. J. Phythian (Mater. Res. Soc. Symp.Proc. 373, Pittsburgh, PA, 1995), p. 555.

359. L. F. Gladden, J. Non-Cryst. Solids119, 318 (1990).360. C. A. Marians and L. W. Hobbs, J. Non-Cryst. Solids119, 269

(1990); 124, 242 (1990).361. L. W. Hobbs, J. Non-Cryst. Solids192 & 193, 79 (1995).362. C. E. Jesurum, V. Pulim, B. Berger, and L. W. Hobbs, in

Microstructure Evolution During Irradiation,edited by I. M.Robertson, G. S. Was, L. W. Hobbs, and T. Diaz de la Rubia(Mater. Res. Soc. Symp. Proc.439, Pittsburgh, PA, 1997),p. 633.

363. N. Yu, M. Nastasi, T. E. Levine, J. R. Tesmer, M. G. Hollander,C. R. Evans, and C. J. Maggiore, Nucl. Instrum. Methods in Phys.Res. B99, 566 (1995).

364. C. W. Allen, L. L. Funk, E. A. Ryan, and A. Taylor, Nucl.Instrum. Methods in Phys. Res. B40/41, 553 (1989).

365. Hj. Matzke, Surf. Interface Anal.22, 472 (1994).366. K. J. Caulfield, R. Cooper, and J. F. Boas, J. Chem. Phys.92

(11), 6441 (1990).367. K. J. Caulfield, R. Cooper, and J. F. Boas, J. Nucl. Mater.184,

150 (1991).368. L. M. Wang, Nucl. Instrum. Methods in Phys. Res. B (1998) in

press.369. O. L. Krivanek, P. H. Gaskell, and A. Howie, Nature (London)

262, 454 (1976).

13, No. 6, Jun 1998 1483


.

.

..

.

J.

c.

370. G. Y. Fan and J. M. Cowley, Ultramicroscopy17, 345 (1985);24, 49 (1988).

371. M. L. Miller and R. C. Ewing, Ultramicroscopy48, 203 (1993).372. L. C. Qin and L. W. Hobbs, inMicrostructure of Irradiated

Materials,edited by I. M. Robertson, L. E. Rehn, S. J. Zinkle, andW. J. Phythian (Mater. Res. Soc. Symp. Proc.373, Pittsburgh,PA, 1995), p. 329.

373. D. E. Jesson and S. J. Pennycook, Proc. R. Soc. London. SA 441, 261 (1993); M. M. J. Treacy and J. M. Gibson, Ultrami-croscopy52, 31 (1993).

374. M. M. J. Treacy and J. M. Gibson, Acta Crystallogr. A52, 212(1996).

375. T.-C. Lee and J. Silcox, inMicrostructure of Irradiated Mate-rials, edited by I. M. Robertson, L. E. Rehn, S. J. Zinkle, andW. J. Phythian (Mater. Res. Soc. Symp. Proc.373, Pittsburgh,PA, 1995), p. 335.

376. R. A. B. Devine, Nucl. Instrum. Methods in Phys. Res. B91,378 (1994).

377. R. Brydson, H. Sauer, and W. Engel, inTransmission ElectronEnergy-Loss Spectroscopy in Materials Science(TMS, Warren-dale, PA, 1982), p. 131.

378. J. Bruley, Microsc. Micro. Anal. & Microstructure4, 23 (1993).379. S. R. Elliott, Nature (London)354, 445 (1991).380. L. C. Qin and L. W. Hobbs, inMicrostructure of Irradiated

Materials,edited by I. M. Robertson, L. E. Rehn, S. J. Zinkle, andW. J. Phythian (Mater. Res. Soc. Symp. Proc.373, Pittsburgh,PA, 1995), p. 341.

381. L. C. Qin and L. W. Hobbs, J. Non-Cryst. Solids192 & 193,456 (1995).

382. A. N. Sreeram and L. W. Hobbs, inBeam-Solid Interactions:Fundamentals and Applications,edited by M. Nastasi, L. R.


er.

Harriott, N. Herbots, and R. S. Averback (Mater. Res. Soc. SympProc. 279, Pittsburgh, PA, 1993), p. 559.

383. A. N. Sreeram, L. W. Hobbs, N. Bordes, and R. C. Ewing, NuclInstrum. Methods in Phys. Res. B116, 126 (1996).

384. E. K. H. Salje, Phase Transitions55, 37 (1995).385. K. R. Locherer, S. A. Hayward, P. J. Hirst, J. Chrosch, M

Yeadon, J. S. Abell, and E. K. H. Salje, Phil. Trans. Royal SocLondon 354, 2815 (1996).

386. E. K. H. Salje, private communication.387. N. J. Hess, W. J. Weber, and S. D. Conradson, J. Alloys &

Compounds (1998, in press).388. A. C. Wright, B. Bachra, T. M. Brunier, R. N. Sinclair, L. F.

Gladden, and R. L. Portsmouth, J. Non-Cryst. Solids150, 69(1992).

389. G. L. Catchen, Mater. Res. Soc. Bull.XX (7), 37 (1995).390. G. L. Catchen, W. E. Evenson, and D. Allred, Phys. Rev. B54,

3679 (1996).391. G. L. Catchen, T. M. Rearick, and D. G. Schlom, Phys. Rev. B

49, 318 (1994).392. T. M. Rearick, G. L. Catchen, and J. M. Adams, Phys. Rev. B

48, 224 (1993).393. G. L. Catchen and T. M. Rearick, Phys. Rev. B52, 9890 (1995).394. J. A. Gardner, H. Jaeger, H. T. Su, W. H. Warnes, and J. C

Haygarth, Physica B150, 223 (1988).395. B. C. Sales, J. O. Ramey, J. C. McCallum, and L. A. Boatner,

Non-Cryst. Solids126, 179 (1990).396. A. N. Sreeram and L. W. Hobbs, inCrystallization and Related

Phenomena in Amorphous Materials,edited by M. Libera, T. E.Haynes, P. Cebe, and J. E. Dickinson, Jr. (Mater. Res. SoSymp. Proc.321, Pittsburgh, PA, 1994), p. 25.

, No. 6, Jun 1998

Radiation effects in crystalline ceramics for the immobilization of high-level nuclear waste and...

Documents

Transcript of Radiation effects in crystalline ceramics for the immobilization of high-level nuclear waste and...