Journal of MATERIALS RESEARCH Welcome Comments Help · 2015. 7. 20. · Materials Science Division, Argonne National Laboratory, Argonne, Illinois 60439 ... V. RESEARCH FACILITIES

http://journal

1946

HelpCommentsWelcomeJournal of

MATERIALS RESEARCH

Radiation effects in glasses used for immobilization of high-levelwaste and plutonium disposition

William J. WeberPacific Northwest National Laboratory, P.O. Box 999, M.S. K2-44,Richland, Washington 99352

Rodney C. Ewinga)

Department of Earth and Planetary Sciences, University of New Mexico,Albuquerque, New Mexico 87131

C. Austen AngellDepartment of Chemistry, Arizona State University, Tempe, Arizona 85287

George W. ArnoldConsultants International, 11729 S Highway 14, Tijeras, New Mexico 87059

Alastair N. CormackNYS College of Ceramics, Alfred University, Alfred, New York 14802

Jean Marc DelayeDTA/SRMP, Centre d’Etudes de Saclay, 91191 Gif-sur-Yvette Cedex, France

David L. GriscomNaval Research Laboratory, Washington DC 20375

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

Alexandra Navrotskyb)

Geological and Geophysical Sciences & Princeton Materials Institute, Princeton University,Princeton, New Jersey 08544

David L. PriceMaterials Science Division, Argonne National Laboratory, Argonne, Illinois 60439

A. Marshall StonehamDepartment of Physics and Astronomy, University College London,London WC1E 6BT, United Kingdom

Michael C. WeinbergDepartment of Material Science and Engineering, University of Arizona,Tucson, Arizona 85721

(Received 28 October 1996; accepted 17 April 1997)

This paper is a comprehensive review of the state-of-knowledge in thefield of radiation effects in glasses that are to be used for the immobilizationof high-level nuclear waste and plutonium disposition. The current statusand issues in the area of radiation damage processes, defect generation,microstructure development, theoretical methods and experimental methodsare reviewed. Questions of fundamental and technological interest that offeropportunities for research are identified.

c ials.

a)Present address: Nuclear Engineering and Radiological ScienUniversity of Michigan, Ann Arbor, Michigan 48109-2104.

s.cambridge.org Downloaded: 31 Oct 2012

J. Mater. Res., Vol. 12, No. 8, Aug 1997

es,b)Present address: Department of Chemical Engineering and MaterScience, University of California at Davis, Davis, California 95616

IP address: 144.82.107.169

1997 Materials Research Society

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

http://journals.cambridge.org

W. J. Weber et al.: Radiation effects in glasses used for immobilization of high-level waste

8

9

3

44

6

7

8

0

ndest

entrx.scallly

e-

aallof

myisonna,eof

ees.g-forl,eof

nd-enlFe,s-lserC.s

e in

TABLE OF CONTENTS

I. INTRODUCTION 1947A. High-level nuclear waste and plutonium 194B. Principles of radionuclide immobilization 1949

in glass

II. PRINCIPLES OF RADIATION EFFECTS 1950A. Radiation sources 1950B. Interaction of radiation with glass 1950

1. Ionization and electronic excitation 19512. Ballistic processes 19513. Transmutations and gas production 195

C. Irradiation techniques 19531. Actinide-incorporation 19532. Actinides in natural glasses 19533. Charged-particle irradiation 19544. Gamma irradiation 19545. Neutron irradiation 1954

III. STRUCTURE AND RADIATION EFFECTSIN GLASSES 1954

A. Structure of glasses 1955B. Silica glass 1956

1. Macroscopic changes 19562. Point defects 1956

C. Alkali silicate glasses 19581. Macroscopic changes 19582. Point defects 1958

D. Borosilicate glasses 19581. Macroscopic changes 19582. Point defects 1959

E. Simulated waste glasses 1951. Volume changes 19592. Stored energy 19603. Helium accumulation 19614. Microstructural changes 19615. Mechanical properties 19626. Enhanced diffusion 19627. Radionuclide release 1962

IV. FUNDAMENTAL SCIENTIFIC ISSUES 1963A. Glass structure and energetics 196B. Ionization versus ballistic effects 1964

1. Electron- and X-irradiations 19642. Ion-irradiations 1964

C. Relaxation processes and diffusion 1961. Relaxation processes 1962. Diffusion 1965

D. Temperature and dose rate effects 196E. Radiation-induced transformations 196

1. Phase separation 19662. Devitrification 19673. Amorphization of crystalline phases 196

F. Volume changes 1968G. Gas accumulation and bubble or void 196

formationH. Chemical durability 1969I. Theory and computer simulations 197

V. RESEARCH FACILITIES 1972A. Actinide research facilities 1972B. Gamma irradiation facilities 1972C. Ion-beam irradiation facilities 1972D. Electron-beam irradiation facilities 1973

http://journals.cambridge.org Downloaded: 31 Oct 2012

J. Mater. Res., Vol.

2

5

VI. RESEARCH NEEDS 1973A. Reference glass compositions 1973B. Structural and thermodynamic properties 1973C. Systematic irradiation studies 1973D. Theory and computer simulations 1974E. Facilities 1974F. Attract talented scientists 1974

ACKNOWLEDGMENTS 1974

REFERENCES 1975

I. INTRODUCTION

Stabilization and immobilization of high-level tankwaste, high-level sludge in fuel storage basins, aplutonium (Pu) residues/scraps are some of the largand most costly challenges facing the U.S. Departmof Energy (DOE) in its effort to close the nucleafuel cycle and cleanup the DOE weapons comple1

In addition, the immobilization and disposal of surpluweapons-grade plutonium creates growing technologidemands and raises political issues that are historicaand technically tied to DOE’s environmental managment and restoration activities.2,3 In most cases, thesehigh-level nuclear wastes (HLW) will be immobilizedin glass waste forms prior to permanent disposal ingeologic repository. The immobilization and disposof HLW and Pu in glass waste forms is a globaissue, as other countries also address the issuessurplus weapons Pu and continue to produce HLW frothe reprocessing of commercial nuclear fuels. A kechallenge in the permanent disposal of HLW and Puthe development of predictive models that are baseda sound scientific understanding of relevant phenomeincluding the effects of radiation, thus allowing thassessment of long-term performance. The durabilityHLW glasses has been studied in detail,4–6 but withlimited consideration of the effects of radiation from thdecay of the incorporated fission products and actinidSuch radiation effects could potentially impact the lonterm performance and environmental risk of glassesHLW and Pu immobilization and disposal. In generathere is insufficient scientific or engineering data in thcurrent literature to fully assess the actual impactradiation on glass waste form behavior.

In order to assess the current state of understaing, identify relevant scientific issues, and determindirections for future research in the area of radiatioeffects in glasses for the immobilization of high-levewaste and plutonium, a panel was convened in SantaNew Mexico, on February 25–29, 1996, under the aupices of the Department of Energy, Council on MateriaScience. The primary objective of this twelve membpanel, which was chaired by W. J. Weber and R.Ewing, was to identify the fundamental scientific issuethat need to be addressed and techniques availabl

IP address: 144.82.107.169

12, No. 8, Aug 1997 1947



orr

tieio

ntiior

sssc

in

dnaat

m

u

iaheyn

i

e

i

ds

yeped

edgtingthrs

e

fhendn-

s

of.s

oy

of

he,

e

.,ntndim

ofntiald

m

dd”n

r.tys

be

c-s

order to: (i) advance the understanding of radiatieffects in complex glasses, (ii) perform accelerated irdiation studies and computer simulations to doses cosponding to 104 to 106 years of storage, and (iii) providethe sound scientific base needed to develop predicstrategies and models for the performance of nuclwaste glasses. This report summarizes the deliberatand conclusions of the panel.

The effect of radiation on HLW glass is of speciaimportance because a major portion of the high-level dfense waste in the U.S. is destined for incorporation ina borosilicate glass waste form. The glass waste formthe first barrier to the release and dispersion of radioclides, and at the same time, the materials proper(e.g., leach rate, stored energy, and volume expansof glasses can be very sensitive to interactions with adiation field. The recent proposal to incorporate exceweapons plutonium in glass2,3,7 has only increased theneed to understand the effects of high-radiation doon the performance of a broad range of nuclear waglasses. The very high cost of vitrification facilitie[up to 4 billion dollars for the Defense Waste Proessing Facility (DWPF) at the Savannah River Site8],the long times required for construction of the facilitie[13 years and 12 years for the construction of DWPand the West Valley Demonstration Project (WVDP)New York, respectively9], and the extremely long timeperiods (104 to 106 years) required by disposal in ageologic repository, mandate a prudent need for a funmental, scientific understanding of relevant phenomesuch as the effects of radiation on the nuclear waste glTo begin, a brief review of the nature of the waste, chacteristics of the waste form glass, and the scale ofproblem are presented.

A. High-level nuclear waste and plutonium

The disposal of high-level radioactive waste fronuclear weapons production, naval propulsion programand the processing of commercial spent nuclear fuelstechnically advanced countries is a growing global issIn the United States, defense HLW was generated as asult of DOE reprocessing operations at the Hanford S(Washington), the Savannah River Site (South Carolinand the Idaho National Engineering Laboratory. TDOE also is responsible for the liquid waste producin the 1960s at a demonstration plant in West ValleNew York, for the reprocessing of commercial spenuclear fuel. Currently, about 100 million gallons oHLW under the stewardship of the U.S. DOE is stored243 underground tanks in Washington, South CarolinIdaho, and New York.1 These wastes are the legacy o50 years of research, production, and testing of nuclweapons at a cost of over 300 billion dollars (in 199dollars). The cleanup of the weapons complex is in


1948 J. Mater. Res., Vol. 1

na-re-

vearns

le-toisu-esn)

a-ss

este

-

sF

a-a,ss.r-he

s,ine.re-te),ed,t

fna,far5ts

early stages with an anticipated expenditure of hundreof billions of dollars in an effort that will continue wellinto the 21st century.

The HLW at the Savannah River and the West ValleSites are destined for immobilization in a borosilicatglass. Specific glass compositions have been develofor DWPF10 and WVDP11; however, glass compositionsfor the Hanford tank waste have not yet been selector developed. Vitrification plants have a long operatinhistory in France and have been operated or are operain England, Belgium, India, Japan, and Russia. Bothe DWPF and WVDP started production of canistefilled with HLW glass (approximately 3 m high and0.6 m diameter10) in 1996 and are scheduled to produc6000 canisters over 25 years12 and 300 canisters over2.5 years,11 respectively. At present, the majority othe high-level defense waste is stored in tanks at tHanford and Savannah River sites in Washington aSouth Carolina, respectively. The volumes are substatial. At the Hanford site alone the inventory include11 million m3 of fluids and 6900 metric tons of nu-clear materials, which includes 4100 metric tonsuranium and 15 metric tons of Cs and Sr in capsules13

The total activity of the waste at the Hanford site iestimated to be 446 million curies.13 For all of theprincipal DOE sites (Hanford, Savannah River, IdahNuclear Engineering Laboratory, and the West ValleDemonstration Project), the combined total volumehigh-level waste (through 1988) is 385,000 m3, andthe total activity is 1,200,000,000 curies.14 Althoughthe defense high-level wastes account for 95% of tvolume of the current total United States HLW inventorythey account for only 9% of the total radioactivity. Thmajority of the radioactivity inventory in the U.S. iscontained in commercial spent nuclear fuel. In the U.Sthe present policy is for the direct disposal of spenuclear fuel; however, in France, England, Japan, aRussia, the spent nuclear fuel is reprocessed to reclafissile material for power generation and/or productionnuclear weapons. This reprocessing generates substavolumes of high-level waste, most of which is destinefor vitrification.

There are also several tens of tons of plutoniuresidues and scraps in the DOE complex,1 primarily atthe Rocky Flats Plant in Colorado, the Hanford Site, anthe Savannah River Site. This material was “strandein the plutonium processing facilities due to the suddeshutdown of the facilities at the end of the Cold WaThe plutonium residues and scraps are in a wide varieof forms, including Pu dissolved in acid, rough pieceof metal, and nearly finished weapons parts.1,7 Similarquantities of plutonium residues and scraps are tofound in Russian weapons production facilities.

Under the first and second Strategic Arms Redution Treaties (START I and II) and unilateral pledge

IP address: 144.82.107.169

2, No. 8, Aug 1997



o

no

R

ges

r-h

n

a

ePe

i

us

ro

vn

e

t

d

es,tved

ofostes,stealasaseteds);alic

te20as

to

-es,

bleident.lly

stera-

corcalte

sat

-

a-

al,, ae

e

made by the United States and the former Soviet Uniand Russia, many thousands of nuclear weapons willdismantled during the next decade. This will generaover one hundred metric tons of plutonium and mahundreds of tons of highly enriched uranium, bothwhich will require disposition.2,3,15 Thus, the fissile ma-terials, whose separation and concentration generateddefense HLW, have become a new waste stream.cent studies by the National Academy of Sciences2,3 andothers16 have recommended consideration of vitrifyinweapons-grade plutonium in a mixture with high-levradioactive waste as one of two “promising” optionConsequently, earlier assumptions that the nuclear waglass would have a low actinide content and thus a locumulative exposure toa-decay radiation are no longevalid. Additionally, the durability of the waste form becomes a very important concern with the disposal of tfissile material,239Pu and its daughter235U (the half-lives of 239Pu and 235U are 2.413 104 and 7.043

108 years, respectively17), because any degradation othe waste form due to radiation effects may lead to icreased corrosion rates and subsequent nuclide transand reconcentration that could conceivably lead to cricality in a repository.18–20 Glass compositions currentlyunder development for immobilization of Pu at the Svannah River Site21 and Argonne National Laboratory22

are significantly different from the compositions thahave been investigated for immobilization of high-levtank wastes, and it is proposed that 5 to 20 wt. %be incorporated into these glass waste forms. For thnovel glass compositions, there are no data on thefects of radiation on glass structure or properties, anas discussed below, the lack of a fundamental scientunderstanding inhibits the extrapolation of the limiteexisting data to new glass compositions and the equalent higher radiation doses.

To the extent that the high-level defense or Pcontaining radioactive wastes must be treated andlidified prior to final disposition, the effect of radiationon the waste form solids must be an important conceOne of the more important applications of the studyradiation effects in ceramics (vitreous and crystallineduring the late 1970s and early 1980s, was the effectsradiation on borosilicate glasses and crystalline ceramfor HLW immobilization. Radiation effects in HLWforms and related materials have been the subjectseveral workshops23,24 and reviews.25–32 In general, mostof the studies on radiation effects in HLW glasses habeen engineering test studies that generated little, if asystematic understanding of radiation effects in HLWglasses. As a result, there has been little advancementhe understanding of radiation effects in HLW glassover the past decade.32 Based on some of the datain these reviews, the post-disposal radiation effectsglass waste forms could be significant. For example,


J. Mater. Res., Vol. 1

nbeteyf

thee-

l.stew

e

f-

portit-

-

tluse

ef-d,ficdiv-

-o-

n.f

),of

ics

of

ey,

t ins

tohe

cumulativea-decay dose that is projected for DWPF anWVDP nuclear waste glasses reaches a value of 7.63

1016 a-decay eventsyg in 500 years,30 while over thesame time period, glasses containing 10 to 20 wt. %239Puwill accumulate 3.6 to 7.23 1018 a-decay eventsyg,33,34

nearly 50 to 100 times higher. In both these casthe dose is well within the range for which importanchanges in physical and chemical properties are obserto occur in HLW glasses.

B. Principles of radionuclide immobilizationin glass

Borosilicate glass is, at present, the waste formchoice for most nuclear waste compositions and for mcountries (e.g., France, Great Britain, the United StatJapan, and China) that have high-level defense waor high-level waste from reprocessing of commercinuclear fuels. The selection of borosilicate glass wbased mainly on four assumptions: (i) an anticipated eof processing (glass frit and the waste are mixed, melat relatively low temperatures, and poured into canister(ii) that the technology is well demonstrated for actu(radioactive) waste; (iii) that the glass as an aperiodsolid will easily accommodate wide variations in wasstream compositions which are extremely complex (to 30 component systems); and finally, (iv) that glassan aperiodic solid will show only a minor responsethe effects of radiation, especially froma-decay events.Glass waste forms can be of a wide variety of compositions, including silicate glasses, borosilicate glassand phosphate glasses.35 In principle, radionuclides areevenly dispersed throughout the glass, although nometal precipitates and, in some cases, crystalline oxprecipitates enriched in radionuclides may be preseWaste loadings (defense and commercial) are typicain the range of 10 to 30 wt. %.

Assessments of the long-term performance of waforms is rare; present performance assessments ofdionuclide containment rely primarily on the geologibarriers (e.g., long travel times in hydrologic systemssorption on mineral surfaces). The physical and chemidurability of the waste form, however, can contribugreatly to the successful isolation of radionuclides;36

thus, the effects of radiation on physical propertieand chemical durability of waste forms are of greimportance. This was illustrated recently in a performance assessment of zircon for Pu disposal.34 The changesin chemical and physical properties occur over reltively long periods of storage, up to 106 years, and attemperatures that range from 100 to 300±C, dependingon waste loading, age of the waste, depth of buriand the repository-specific geothermal gradient. Thusmajor challenge is to effectively simulate high-dosradiation effects that will occur at relatively low-dos

IP address: 144.82.107.169

2, No. 8, Aug 1997 1949



e

c

e

ru

o

g

it

g

e

,e

ase

the

onsof

forr

tiche

es.

entnillt

rates over long periods of time at elevated storatemperatures.

II. PRINCIPLES OF RADIATION EFFECTS

A. Radiation sources

The principal sources of radiation in HLW arb-decay of the fission products (e.g.,137Cs and90Sr)and a-decay of the actinide elements (e.g., U, Np, PAm, and Cm), both of which lead to physical and chemcal changes in the waste form. Beta-decay produenergetic b-particles, very low energy recoil nucleiand g-rays; whereas, alpha-decay produces energa-particles (4.5 to 5.5 MeV), energetic recoil nuclei (7to 100 keV), and someg-rays. There also are minocontributions to the radiation field from spontaneofission events of the actinides and froma-neutron re-actions. The rates of spontaneous fission anda-neutronreactions are very low and do not significantly contributo the overall effects of radiation; consequently, radiatieffects from spontaneous fission are not consideredthis summary. In general,b-decay is the primary sourceof radiation during the first 500 years of storage, asoriginates from the shorter-lived fission products (e.the half-life of 137Cs is 30.2 years17 and the half-life of90Sr is 28.8 years17). The b-decay of fission products isresponsible for the high radioactivity, high self-heatinrates, and elevated temperatures early in the historywaste form storage. Due to the long half-lives of thactinides and their daughter products (e.g., the half-livof 237Np, 239Pu, and235U are 2.143 106, 2.413 104, and7.04 3 108 years,17 respectively),a-decay is generallydominant at longer times. Figure 1 shows the cumulatbeta and alpha decay events as a function of sage time that have been estimated for DWPF glass29,30

FIG. 1. Cumulative number of decay events per gram for DWPglass, a commercial (C-HLW) nuclear waste glass, and a glcontaining 10 wt. % plutonium.


1950 J. Mater. Res., Vol.

ge

u,i-es

,tic

0

s

tenin

it.,

gof

ees

veor-

Fass

FIG. 2. Cumulativea-decay dose as a function of plutonium loadinand waste storage time.

and non-U.S. commercial HLW glass.25 The cumulativeb-decay anda-decay events for the WVDP glass arsimilar to those for DWPF glass.29,30 Also shown inFig. 1 is the cumulative number ofa-decay events for awaste glass containing 10 wt. %239Pu,33 which reachessignificantly larger values relative to DWPF, WVDPor non-U.S. commercial HLW glasses. The cumulativa-decay events in glasses for239Pu immobilization reacha temporary plateau after 100,000 years, but increover much longer time periods due to the235U decayseries, as shown in Fig. 2 for several239Pu concentrationsin a waste glass.

B. Interaction of radiation with glass

Beta and alpha decay affect waste glass throughinteractions of theb-particles,a-particles, recoil nuclei,and g-rays with the glass. These interactions fall inttwo broad categories: the transfer of energy to electro(ionization and electronic excitations) and the transferenergy to atomic nuclei, primarily byballistic processesinvolving elastic (billiard-ball-like) collisions. The totalenergy absorbed as a function of storage timeb-decay anda-decay processes is shown in Fig. 3 foDWPF glass,29,30non-U.S. commercial HLW glass,25 anda 10 wt. %239Pu glass.33 The partitioning of the energytransferred into electronic excitations and into elasnuclear collisions is an important process controlling teffects of radiation. Forb-particles andg-radiation, theenergy transfer is dominated by ionization processFor ions, such asa-particles and recoil nuclei, thispartitioning is governed by the relative velocity of thions and that of the orbital electrons of the constitueions of the glass. If the ion velocity is higher thathat of the orbital electrons, ionization processes wdominate. However, if the particle velocity is below tha

IP address: 144.82.107.169

12, No. 8, Aug 1997



ish

i

o

i

t

tg

s

sr

h

ag

nh

i

e

es

orin

ndto

llatreeanes

clee,ec-icgoftooentic

e-e-hes,

dnly

ghof

lyl,e

ic

seie

FIG. 3. Cumulative absorbed dose for DWPF, C-HLW, and 10 wt.plutonium glasses.

of the orbital electrons, the likelihood of ionizationsmall, and most of the energy will be transferred to tnucleus of the ion. A useful relation is that ionizatioprocesses dominate if the energy of the bombardparticle, expressed in keV, is greater than its atomweight, and elastic collision (Coulombic) interactionwith atomic nuclei are predominant if the energythe ion falls below this limiting approximation. Thusan a-particle with an atomic weight of 4 and energranging from 4.5 to 5.5 MeV will predominately deposits energy by ionization processes, while ana-recoil ionwith an atomic weight of 235 to 244 and energy 70100 keV will lose most of its energy in elastic collisionwith the nuclei of atoms in the glass. In additionthe transfer of energy, the particles emitted throuradioactive decay can themselves, in some cases, hasignificant effect on the nuclear waste glass as a reof their deposition and accommodation in the glastructure.

1. Ionization and electronic excitation

The amount of energy absorbed through interactioof the emitted particles with the atomic nuclei is onlyvery small fraction of the total energy absorbed; conquently, the results in Fig. 3 also approximate the enedeposited into ionization and electronic excitationsthe glasses. Energy loss byb-particles anda-particlesis predominately through Coulombic interactions. Tinteraction ofg-rays with matter is primarily throughthe photoelectric effect, Compton scattering, and pproduction. The high-rate of energy absorption throuionization and electronic excitation fromb-decay inHLW glasses and froma-decay in Pu glasses caresult in significant self-heating. The magnitude of ttemperature increase depends on the rate of energysorption, physical properties of the solid (e.g., its spec



%

enngicsf

,yt

osoh

ve ault

ss

nsae-gyin

e

irh

eab-

fic

heat), thermal conductivity, and heat transfer to thsurroundings. In the case of non-U.S. commercial HLWglass in a repository, self-heating fromb-decay of thefission products can result in initial storage temperaturgreater than 200±C, with temperatures falling below150 ±C several hundred years after emplacement. Fdefense HLW glass in the U.S., initial temperaturesthe Yucca Mountain repository may be as high 250±C,depending on the storage density of the waste glass aspent fuel packages, but temperatures will decreaseless than 100±C after several hundred years.37 Glassesincorporating high loading of weapons-grade Pu wiexperience similar temperatures in the repositoryYucca Mountain; however, the decrease in temperatumay be slower due to additional self-heating from thdecay of the long-lived Pu. Temperature in this range cprofoundly affect the response of nuclear waste glassto radiation effects.

In addition to self-heating, ionization and electroniexcitations produce a large number of electron-hopairs and can result in covalent and ionic bond rupturcharged defects, enhanced diffusion, and localized eltronic excitations. Bond ruptures and localized electronexcitations can lead to the formation of nonbridginoxygen defects, local decomposition (e.g., formationO2), and the permanent displacement of atoms duethe conversion of a localized electronic excitation intatomic motion (often referred to as radiolysis). Thradiolytic displacement of atoms can be a very efficieprocess in glasses, producing significantly more atomdisplacements than ballistic collisions.

2. Ballistic processes

Ballistic processes cause direct atomic displacments through elastic scattering collisions and are rsponsible for the atomic-scale rearrangement of tstructure. Beta particles, because of their low masdo not efficiently transfer energy to atomic nuclei aninduce well-separated single displacement events oat high energies. The recoil nuclei (b-recoils) producedin b-decay events are almost never energetic enouto be permanently displaced. Similarly, the emissiong-rays results in the recoil of nuclei (g-recoils), whichare also not sufficiently energetic to be permanentdisplaced for most decay products in HLW. In generab-decay of the fission products in HLW produces on thorder of 0.1 atomic displacements perb-decay event.The primary source of atomic displacements by ballistprocesses in HLW and Pu glasses are thea-particle anda-recoil nucleus produced in ana-decay event. Thea-particles emitted by actinide decay in HLW glassehave energies of 4.5 to 5.8 MeV, and the recoil nucla-recoils) have energies of 70 to 100 keV. In the casof Pu glasses, a 5.2 MeVa-particle and a 86 keV235U

IP address: 144.82.107.169

2, No. 8, Aug 1997 1951



h

y

n

h

neco

c

e

u

f

a

ns.s-e-eV

e

is00

pa)

ly

eic

s,-athe

oth

of

recoil are released by the decay of the239Pu (a-particlesanda-recoils of similar energies are released during tdecay of the235U series).

The a-particles dissipate most of their energy bionization processes over a range of 16 to 22mm, butundergo enough elastic collisions along their pathproduce several hundred well-separated atomic displaments. The largest number of displacements occurs nthe end of range of thea-particles. The partitioning ofenergy loss between ionization and elastic collisions athe distribution of ballistically produced displaced atomcan be determined by Monte Carlo calculations usingcomputer code, such asTRIM.38 The results of such acalculation usingTRIM-96 are illustrated in Fig. 4, wherethe positions of displaced oxygen ions and cations tresult from a 5.2 MeVa-particle in a borosilicate glasscomposition are shown on thex-y plane. (The wide useof the TRIM code is due to its many years of use by thsemiconductor industry where its predictions have beverified by many experimental determinations.)

The more massive but lower energya-recoil particleaccounts for most of the total number of displacemeproduced by ballistic processes in HLW and Pu glassThe a-recoil loses nearly all of its energy in elasticollisions over a very short range (30 to 40 nm), prducing highly localized damage (displacement cascaor thermal spike) with 1000 to 2000 displacements. Sua displacement cascade for a235U recoil in a borosilicateglass is illustrated in Fig. 5. The density of energdeposited into the glass by ana-recoil cascade is veryhigh (up to 1 eVyatom), occurs over a very short tim(,10212 s), and can lead to local “melting.”

In an a-decay event, thea-particle anda-recoilparticle are released in opposite directions and prod

FIG. 4. Distribution of displaced atoms in the collision cascade o5.2 MeV a-particle in borosilicate glass (calculated usingTRIM-96).



e

toce-ear

dsa

at

een

tss.

-deh

y

ce

a

FIG. 5. Distribution of displaced atoms in a collision cascade of86 keV U recoil in borosilicate glass (calculated usingTRIM-96).

distinct damage regions separated by several microBased on full cascade Monte Carlo calculations uing TRIM-96, the average number of atomic displacments generated in a borosilicate glass by the 5.2 Ma-particle and the 86 keV235U recoil released in thedecay of 239Pu are 380 and 1930, respectively. Thaverage number of displacements generated pera-decayevent by the decay of the actinides in HLW glassexpected to be similar and is estimated to be 23displacements pera-decay event. This is significantlymore than the 0.1 displacements generated perb-decay.The estimated number of displacements per atom (dgenerated byb-decay anda-decay in DWPF glass, anon-U.S. commercial HLW glass, and a 10 wt. %239Puglass are shown in Fig. 6. The results in Fig. 6 clearshow the dominance ofa-decay events (a-particles anda-recoils) in producing ballistic-type collision damagin glass waste forms (displacements from radiolytprocesses are not included).

3. Transmutations and gas production

In addition to the energy transferred to the glasb-decay anda-decay also lead to transmutation of radioactive parent nuclei into different chemical elements thmust be accommodated and may significantly impact tchemical properties of the glass.25,32 The principal sourceof transmutations in HLW areb-decay of the relativelyabundant fission products,137Cs and90Sr. Transmutationof these two elements is accompanied by changes in bionic radius and valence. Cs11 decays to Ba21 with adecrease in ionic radius of 20%, and Sr21 decays to Y31,which in turn decays to Zr41 with a final ionic radiusdecrease of 29%.39 In glasses for Pu immobilization,where the Pu concentration is high, the transmutation

IP address: 144.82.107.169

2, No. 8, Aug 1997



u

n

b

I

r

i

e

r

ac-h-tail

si-,

istes,

orestehodn

redrallnctsnilar

esvermeen

en-ltic

w-ceansesl

FIG. 6. Displacements per atom (dpa) from ballistic processes afunction of waste storage time for DWPF, C-HLW, and 10 wt. % pltonium glasses.

239Pu to235U (both of which are fissile) may also result ichanges in valence states and ionic radius, which coimpact long-term performance and safety [e.g., U61 asthe uranyl ion (UO2)21 is much more soluble than U41].

Helium atoms, which result from the capture of twelectrons bya-particles, are also produced and mustaccommodated in the glass interstitially, be trappedinternal defects, aggregate to form bubbles (or voidor be released at the glass surface.25,30 For HLW, mostof the He will be generated over long time periodnear ambient temperature, and the concentration shonot exceed 100 atomic parts per million (appm).the case of glass waste forms for Pu immobilizatioand disposal, He concentrations can become quite la450 appm and 1.5 atom% in 1000 and 100,000 yearespectively, for glasses containing 20 wt. %239Pu. TheseHe concentrations are significantly above the solubilfor He in borosilicate waste glasses, which is on the ordof 2 3 1022 atomsym3 (0.25 appm) at 400±C and 1.73

105 Pa (He).40,41 Consequently, all of the He generatemust be trapped at defects, collect into bubbles, orreleased at surfaces.

In addition to He production froma-decay, theevolution of molecular oxygen from a wide range oglasses irradiated with 20 keV electrons was confirmover 30 years ago using mass spectroscopy.42,43 Thetendency of the glasses to evolve O2 under ionizingirradiation conditions showed a strong dependenceglass composition, with fused silica showing the lowe



s a-

uld

oeat

s),

suldnnge:rs,

tyer

dbe

fed

onst

rate of O2 evolution. This radiolytic source of O2 canplay an important role in the evolution of microstructur(Secs. III and IV).

C. Irradiation techniques

Since the effects of radiation will accumulate ovevery long time periods, the effects of radiation in HLWglasses have been studied using a broad range ofcelerated irradiation techniques. These irradiation tecniques and procedures have been described in depreviously,25,32 but are briefly reviewed here.

1. Actinide-incorporation

The long-term effects ofa-decay events havebeen simulated in a number of waste glass compotions25–28,32,44 by incorporating short-lived actinidessuch as238Pu (half-life of 87.7 years17) or 244Cm (halflife of 18.1 years17) in sufficient concentrations (0.2to 3 wt. %) to achieve 1018 to 1019 a-decay eventsygin laboratory time periods of up to several years. Thtechnique effectively simulates, at accelerated dose rathe simultaneousa-particle anda-recoil effects that areexpected over long periods of time (103 to 106 years)for HLW waste glasses or over 100 to 1000 years fPu glasses. The incorporation of short-lived actinidinto waste forms in order to test the stability of wasforms toa-decay damage has become a standard met(ISO 6962) approved by the International Organizatiofor Standardization45 and is the basis for the DOEMaterials Characterization Center’s MCC-6 Method fothe Preparation and Characterization of Actinide-DopWaste Forms46 that has been used to prepare sevePu-doped HLW glasses.47 Phase separation and partiadivitrification can result in nonhomogeneous distributioof actinides in actual nuclear waste glasses. These effecan be readily simulated using actinide incorporatiotechniques and appropriate processing to produce simphase separation or crystallinity.

2. Actinides in natural glasses

Natural glasses do contain238U, 235U, and232Th andtheir daughter products. Unfortunately, natural glassdo not contain enough uranium and thorium to seras natural analogues for radiation effects in waste foglasses because of the generally low doses that have bobtained. In volcanic glasses, the uranium content gerally increases with the silica content; hence, basaglasses have uranium contents that range from 0.1mgygto a fewmgyg, and silicic glasses (obsidians), from a femgyg to about 20mgyg. Despite these low uranium concentrations, fission tracks occur in sufficient abundanto allow dating of natural glasses. There are more th400 successful age determinations on volcanic glaswhich utilize fission track dating. Tektites (terrestria

IP address: 144.82.107.169

2, No. 8, Aug 1997 1953



o

ac

mta

is

ear

k

la

ig

ss

atr

.u

os

og

la

o

f

dd

yder

icone

-

icis-of

-

a

le,de

-eat

ntu

e soa-r-

ge-e

-e

seesst

e-

impact glasses) have uranium contents on the orderfew mgyg, and more than 250 specimens have been dautilizing fission tracks. The fission tracks representextreme in radiation damage caused by massive parti(fission fragments) of high energy (hundreds of MeVwith the latent tracks being up to several tens of nin length. The fission fragment yield in nuclear wasglasses is very much lower than that of alpha-decevents, and thus this is not a major damage mechanStill, fission tracks offer an opportunity for the study otrack annealing in an aperiodic solid,48 and such resultsmay be relevant to the evaluation of the persistenof damage tracks in nuclear waste glasses. Therehowever, one instance in which a synthetic Th-dopborosilicate glass, 17 years in age, has been studied,enhanced chemical reactivity of the damage tracks pduced by thea-recoil nuclei was clearly demonstrated.49

This phenomenon, differential etching of damage tracis well documented in naturally occurring minerals anglasses.50

3. Charged-particle irradiation

Charged-particle irradiation using electrons, protona-particles, and heavy ions has been used to simuand study radiation effects in HLW glasses.32 These tech-niques generally employ particle accelerators and hdose rates; consequently, very significant doses canreached in short periods of time (e.g., minutes). Analyof the results from such irradiations is difficult becauthe damaged areas are thin surface layers of restriclateral extent. The high surface area can act as a sfor migrating defects, and electric field gradients mbe generated along the electron or ion paths. Implanlayers also can change the characteristics of the tamaterial as a result of compositional changes51 (e.g., Pbimplantation). Still,a-particle irradiation can be a goodsimulation of a-particle damage, and heavy-ion (e.gXe, Pb) irradiation could provide a reasonable simlation of a-recoil damage. In fact, the two techniquecould be used simultaneous to study the effects of ba-particles anda-recoils on glass. Electron irradiationhave generally been used to simulate and study thefects of ionization and electronic excitations fromb-particles andg-rays on glass. One disadvantagecharged-particle irradiations, particularly for simulatinthe damage from alpha-decay events, is that the daage is homogeneously distributed and does not simuheterogeneous irradiation effects that may result froa non-homogeneous distribution of actinides or fissiproducts.

4. Gamma irradiation

Gamma irradiation, utilizing60Co or 137Cs sources,also has been used to simulate the effects ofb-particlesandg-rays on glass.25,29–32The advantage of this type o



f atednles)

eym.

f

ceis,dnd

o-

s,d

s,te

hbeisetedinkyedget

,-

sth

ef-

f

m-te

mn

irradiation is thatg-rays are so penetrating that simulateHLW glasses can be irradiated in bulk and while sealein containers; furthermore, theg-rays provide a realisticsimulation since they interact with the glass primarilthrough ejected photoelectrons. Dose rates on the orof 2.5 3 106 Ryh are easily achieved.25 Most studieshave focused on the effects ofg-rays on HLW glass cor-rosion, where the most important effect is the radiolytdecomposition of species such as nitrogen and carbdioxide, which are present in air or dissolved in thwater, leading to the formation of nitric and carboxylicacids.25–32

5. Neutron irradiation

Irradiation with neutrons has also been used to simulate and study radiation effects in HLW glasses.25,32

Fast neutrons, which dissipate their energy by ballistprocesses, produce significant numbers of atomic dplacements but only moderately simulate the effectsthe ballistic interactions ofa-particles anda-recoils.25

This technique provides no simulation of helium buildup from the a-particles. Another approach52 that hasbeen used with borosilicate glasses is irradiation inthermal neutron flux to generate a 1.78 MeVa-particleas a result of the high10B(n, a)7Li cross section. Withthis technique, high rates of He formation are possibbut this does not simulate the dense collision cascaof the a-recoil particle. A third technique53 involvesfission of fissile nuclides (e.g.,235U) in the glass byirradiation in a thermal neutron flux; the fission event results in very high-energy fission fragments that producextensive regions of damage (i.e., fission tracks) thmay or may not simulate ballistic damage from thea-recoil particles. Such fission events provide an excellesimulation of spontaneous fission events in HLW and Pglasses; however, as mention above, these events arinfrequent that they are unimportant as damage mechnisms. Similar to charged-particle irradiation, fast and/othermal neutron irradiation produces nearly homogeneous damage that cannot simulate the nonhomoneous distributions of actinides or fission products duto phase separation or crystallinity.

III. STRUCTURE AND RADIATION EFFECTSIN GLASSES

Many glasses proposed for HLW and Pu immobilization and disposal (e.g., DWPF glass) comprisapproximately 50 wt. % SiO2. Therefore, a knowledgeof radiation effects in pure silica (i.e., amorphous formof SiO2) and simple binary silicate glasses can givuseful clues to the behaviors of HLW and Pu glassto be expected at high radiation doses. The earlieobservations of radiation effects in glasses were rported nearly 70 years ago by Curtiss54 and by Lind,55

IP address: 144.82.107.169

2, No. 8, Aug 1997



e

i

o

n

os

sif

e

e

rh

–

nn

iou

bor

he

s,

nen

esral

l.ral

a-

a,

gs

n ane

estednt

eestsn-y

les

notagensch

cas,

nto

ro-ns

ans-nt

lrk

who observed crazing and cracking in silica and Pyrirradiated with a-particles emitted by radium. Laterexperiments by Tuck56 showed that sucha-particledamage leads to increased leach rates for both a solime glass and Pyrex. Although there have been maobservations of irradiation-induced changes in propertof simple glasses (e.g., variations in refractive indedensity, and mechanical properties),57 the understandingof these macroscopic changes in structure is still nwell developed. On the other hand, the familiesradiation-induced point defects in pure silica and simpalkali silicate glasses are well understood from 40 yeaof study, principally by the technique of electron spiresonance (ESR).57,58

The current understanding of structure and radiatieffects in simple and simulated nuclear waste glasis briefly reviewed in this section. This review is nocomprehensive but rather reflects the topics discusby the panel in the deliberations designed to identimportant scientific issues pertinent to radiation effecin nuclear waste form glasses.

A. Structure of glasses

Some description of the structure of network glassis important for assessing the role of radiation effecin creating distinguishable structural rearrangements;other words, how does one identify and describe a defin an aperiodic solid. In the case of irradiated crystallinceramics, one may unambiguously define a displacatom, the point defect spectrum, and gross structurearrangement (e.g., the formation of dislocation loopmetal colloids, and voids59). Without a reference lat-tice, as is the case for an aperiodic (glassy) solid,is difficult to define an individual displaced atom othe gross structural rearrangements associated witdisplacement cascade. One useful approach is to mona characteristic structural parameter, such as the X–Obond angle (X network-forming elements Si, B, Al,P . . .) in oxide network glasses, that is independentcrystallinity. There is, for example, evidence that thSi–O–Si bond angle in vitreous SiO2 decreases bysome 11± after neutron and ion irradiation.60 Similarbond-angle changes have been observed in compoxides, such as (Ca, Th)ZrTi2O7 (i.e., zirconolite), thatexperience a radiation-induced crystalline-to-aperiodtransformation due toa-decay damage from the U andTh content.61 Spectroscopic measurements (EXAFS aXANES) show that the loss of long-range order izirconolite is due to minor variations (less than 5±)in the metal-O-metal bond angle. Also, the neareneighbor coordination sphere of the aperiodic stateessentially the same as that of the crystalline materdemonstrating the similarity of the structures in termsthe coordination polyhedra; the aperiodicity is the res



x

da-nyesx,

otf

lers

nestedyts

stsincteedrals,

it

aitor

X

ofe

lex

ic

d

stisal,flt

of variations in the linkages between the nearest neighcoordination polyhedra.

A more fundamental approach is to evaluate toverall structural topology of the glass,62 of which theX–O–X angle is a characteristic. In network glassethe topological entity that appears is thering, defined asa closed circuit along the edges of linked coordinatiopolytopes (e.g., [SiO4] tetrahedra in silicate glasses). Thcollection of lowest-order rings passing through a givesite (together with associated coordination polytopbelonging to them) represents a characteristic structuentity, much as a space group does for a crysta60

The ring complement can thus define the structutopology, and a change in the ring complement istopologically identifiable change in the network structure. For example, in neutron-irradiated vitreous silicboth Raman60 and neutron diffraction63,64 evidence pointto an increase in the complement of 3-membered rin(i.e., rings containing three linked SiO4 tetrahedra). Theassociation of two 3-membered rings per tetrahedron itetrahedral network has been shown to formally defia two-dimensional surface62; consequently, a significantincrease in the number of 3-membered rings impliformation of internal surfaces, such as would be expecto bound the voids or bubbles that are a persisteobservation32 in irradiated waste glasses.

Density is related to ring complement,65 andirradiation-induced changes in ring topology can bmanifested as density changes. The correlation suggthat the structures of both vitreous and irradiatioamorphized forms of silica are largely dominated b6-membered rings,63 a distribution generally confirmedby computer66 and molecular dynamics67 models. Theenergetics of the silica intertetrahedral bond angare fairly flat until 3-membered ring configuration(130.5±) are reached68; thus, a distribution of ringsizes is expected. Network structures, however, arenecessarily homogeneous, and it is possible to envisheterogeneous structures in which delimited regiodominated by 6-membered rings are linked to eaother by larger and smaller rings.69 Larger rings areresponsible for the higher density of the denser silipolymorphs65 because they can fold back on themselveand the densification of vitreous silica by irradiatiocould arise from opening up of 6-membered ringslarger rings in regional boundaries.

Radiolytic damage processes, such as those pposed in silica,70,71 result in a progressive accumulatioof broken network linkages. Simple bond-switching habeen demonstrated, at least heuristically, as a meof achieving topological disorder in monatomic tetrahedral networks72 and could operate as a subsequerearrangement mode in silica73 and other networks. Ra-diolytic decomposition of X–O–X linkages to a criticathreshold may result in the collapse of the netwo

IP address: 144.82.107.169

2, No. 8, Aug 1997 1955



r

o

gnidcrth

se

isuars

e

udd

e

suhs

s

s

e

n

s

e

at-edofnden

n-re

r-nd

n-keVehisn,s

toetes

eses

-

nf

,e

ioraln

betlythend

ed

ah

and its subsequent stochastic rebonding into altetopologies, just as the rigidity percolation threshold74

that is achieved in central-force networks by progresive bond removal leads to a sudden multiplicatiof floppy modes. Indeed, the maximum in detectabirradiation-induced oxygen vacancy (E0) centers in silicaoccurs at a dose corresponding to maximum ratedensity change,60 when structural collapse and rebondinwould likely re-form broken linkages. In dense collisiocascades, it is possible, as an alternative, to conswholesale restructuring as occurring according to logrowth rules appropriate to either higher- or lowetemperature crystalline polymorphs, depending oneffective temperature of the cascade.

Ease of structural rearrangement in irradiated glasis likely to correlate with a topological parameter likstructural freedom,75 which depends on coordinationpolytopes and their modes of connectivity, just as sugested for amorphization of crystalline ceramics.62 Theaddition of network-modifying cations (such as Na)generally considered to disrupt network connectivity, bthe coordination polytopes surrounding these cationsthemselves linked, to each other and to the netwoand their contribution to the structural constraint mube assessed. Some workers76 have suggested that thdistribution of these network modifying cations may binhomogeneous, in which case their influence on strtural stability may be less than if uniformly disperseA fundamental question in glasses, which are alreatopologically disordered, is whether radiation-inducere-arrangement results in a topologicallydifferent struc-ture (e.g., one with substantially different free energyIon-beam irradiation of lead pyrophosphate (Pb2P2O7)glasses does induce a change in topological structur77

Additionally, the preferential etching ofa-recoil tracksin glasses containing Th and U radionuclides49,50 sug-gests that there may be some difference, although denchange, strain, or segregation within the tracks coaccount for the preferential chemical attack without tnecessity to invoke topologically significant difference

B. Silica glass

1. Macroscopic changes

As noted,a-particle irradiation of silica glass leadto crazing and cracking.54,55 Primaket al.78 were the firstto report the compaction (densification) and increaserefractive index of silica glass when subjected to a faneutron fluence. The density of silica glass increasvery rapidly to a value 3% greater than the unirradiatglass at a neutron fluence of 53 1019ycm2; it thengradually decreases to a equilibrium density that2.7% greater than the unirradiated glass at a flueof ,4 3 1020ycm2.79 Crystalline quartz subjected tothe same fast-neutron fluence undergoes an expan



ed

s-nle

of

eral-e

es

g-

trek,t

ec-.y

d

).

.

itylde.

int-esd

isce

ion

of ,15%,79,80 and asymptotically approaches the samdensity as the irradiated silica glass.57 Both materialsprobably share a common network structure at this suration dose. Though amorphous, this radiation-inducstructure must differ substantially from the structuresmelt-quenched silica glasses. A similar compaction aincrease in the refractive index of silica glass has beobserved under ion irradiation,81–83 and compaction isalso reported under electron84 and g84–86 irradiations.Both ballistic and ionization energy losses lead to desification of silica glass; however, ballistic processes aabout a factor of 200 more efficient.81–83

Although silica glass normally compacts under iradiation, silica glasses impregnated with H expawhen irradiated withg-rays86 or with electrons.87,88 Thisso-called “impurity effect” has been associated with ioization processes because it can be produced by 18electrons,88 which cannot transfer sufficient energy to thnetwork atoms to cause direct atomic displacement. Teffect also is seen at low fluences under ion irradiatiobut is overwhelmed by ballistic collisional processewith continuing irradiation to higher fluences.89 Shelby86

has suggested that this observed expansion is dueinterstitial OH, incorporated into the silica structurduring processing, that inhibits occupation of such siby SiO4 tetrahedral groups after bond-breaking.

Along with the increase in density and refractivindex, the chemical etch rate of silica glass also increawith ion fluence.56,90 Although molecular oxygen isknown to evolve radiolytically from silica glass irradiated with 20 keV electrons,42 bubbles are not observedto form in silica glass under ion and electron irradiatioto high fluences.91 Irradiation-induced stored energy oup to 800 Jyg (0.2 eVyatom) have been reported92 forsilica glass.

In a review of radiation effects in silica glassDevine60 has shown that the concentration of thirradiation-induced oxygen-vacancy (E0) center saturatesbefore significant densification occurs; this behavis illustrated in Fig. 7. With respect to the structurmodifications, Devine60 suggests that the densificatio(2.5–3%) and the large bond-angle changes (,10±)require the formation of large voids, such as canattained by the structural change from a predominan6-membered ring structure to a 3-membered type;latter is typical of films deposited at high temperature aquenched (CVD deposition at 300±C).93 A Raman studyof ion-implanted silica94 showed the development of thD2 band at 605 cm21 that is due to the 3-memberering structure.

2. Point defects

The idealized structure of “defect-free” glassy silicis a “continuous random network” (CRN), wherein eac

IP address: 144.82.107.169

2, No. 8, Aug 1997



isa

r

n

e

e

z

s

nrep

i

of

es,

err

das

yt-

ledtti-ns,

forat

ce

n

e

isof

nral

ith

er-y.ese2

FIG. 7. Dependence of oxygen-vacancy center concentration and dsity on deposited nuclear (ballistic) energy in ion-irradiated amorphosilica (based on Devine60).

silicon is tetrahedrally bonded to 4 oxygens, and (pure silica) each oxygen is bonded to two siliconX-ray and neutron diffraction data for silica glasses cbe explained in terms of CRN-type models.95 A pointdefect in silica may comprise a group of perhaps 210 atoms present in a configuration qualitatively diffeent from any configurations found in the idealized CRNAn oxygen-centered portion of the ideal network cabe represented by the notation°°°°°° Si–O–Si°°°°°° , where thenotation “°°°°°° ” represents three bonds to other oxygein the glass network. Structures probed by ESR aconfined to those which are paramagnetic, i.e., thobinding an unpaired electron (denoted by “?” in thechemical notations used below).

A neutral oxygen vacancy in silica can be reprsented°°°°°° Si–Si°°°°°° . As implied by this notation, there islikely to be a chemical bond between the two siliconadjoining the vacancy site. The paramagnetic versioof the oxygen vacancy are calledE0 centers.96,97 Oneof the positively charged variants in amorphous SiO2,the E0g center,98 is represented°°°°°° Si ? Si°°°°°° , where theunpaired electron is depicted as residing on the lehand silicon, so that the positive charge formally residon the right-hand silicon. (This “asymmetric relaxationmodel99–101 was originally developed to explain the ESRcharacteristics of the analogousE0

1 center in crystallinequartz.) When silica glass (or even crystalline quartis exposed to a fast-neutron fluence of,1020ycm2 oran ionizing dose of,1010 Gy, the measured numberof E0 centers saturate at,1 per thousand silicons.102

The numbers of electrostatically neutral,nonparamag-netic oxygen vacancies induced by these irradiatioare larger, perhaps by an order of magnitude or moThough the nonparamagnetic oxygen deficiency centcannot be counted by ESR, they can be detected ocally as an absorption band peaking near 248 nm.103–106

Correlations have been established between the opt



en-us

n.n

to-.n

srese

-

sns

ft-s

”

)

se.rsti-

cal

bleaching of the 248-nm band and the appearanceE0 centers,104,106 consistent with the view that electronscan be optically detached from neutral oxygen vacanciyielding E0 centers:

°°°°°° Si–Si°°°°°° °! °°°°°° Si ? Si°°°°°° 1e2 . (1)

Oxygen vacancies in silica may be created eithby elastic collisions with energetic nuclear particles oradiolytically by decay of self-trapped excitons (bounelectron-hole pairs created by ionizing radiation, suchg-rays).107–110 A significant conclusion of careful ESRstudies by many workers109,111–113is that the displacedoxygen atoms move to interstitial positions, where thereadily dimerize to form molecular oxygen. Upon heaing to ,200 ±C, this interstitial O2 has been shown todiffuse through the network and react withE0 centersaccording to:

°°°°°° Si ? Si°°°°°° 1O2 °! °°°°°° Si–O–O? Si°°°°°° . (2)

The fragment°°°°°° Si–O–O? is a second fundamentadefect species in silica and is known as the bondperoxy radical.114 Any displaced oxygen that does noreact according to Eq. (2) either remains as interstial O2 molecules or else reacts with neutral oxygevacancies to form (nonparamagnetic) peroxy-linkage°°°°°° Si–O–O–Si°°°°°° :

°°°°°° Si–Si°°°°°° 1O2 °! °°°°°° Si–O–O–Si°°°°°° . (3)

Thus, there are at least 0.5 oxygen-oxygen bondsevery thousand silicon atoms, complementing twice thmany silicon-silicon bonds, at the fast neutron fluenfor saturation. This interstitial O2 is the likely sourceof the radiolytic O2 released under the 20 keV electroirradiations previously noted.42

A third elementary defect center in silica is thnonbridging-oxygen hole center (NBOHC),115 a config-uration represented by the notation°°°°°° Si–O?. Thoughthe NBOHC can have several possible origins, itmost easily understood as resulting from radiolysisa hydroxyl group:

°°°°°° Si–O–H°! °°°°°° Si–O? 1H0 . (4)

An example of a neutralE0 center, denotedE0b, results

when a radiolytic hydrogen atom [originating, e.g., ithe reaction represented by Eq. (4)] reacts with a neutoxygen vacancy98,116:

°°°°°° Si–Si°°°°°° 1H0 °! °°°°°° Si ? H–Si°°°°°° . (5)

Numerous species of defect centers associated wthe elements boron,117 germanium,118 and phosphorus119

have been well characterized by ESR studies of othwise pure silica doped with these elements individuallFor comparable doses, the defects associated with thsubstitutional impurities tend to outnumber by 1 to

IP address: 144.82.107.169

2, No. 8, Aug 1997 1957



a

u

d

p

e

-

r

c

f

f

led-

ltraled

be

edto

mi-ses

x-

esis

ionetheted

mssin

exende

no

ion

esseeds.a,d.

5

ess

orders of magnitude theE0 centers, peroxy radicals, andNBOHC’s observed in high-purity silica.

C. Alkali silicate glasses


Ion irradiation of Na2O–3SiO2 glass results in ex-pansion at low fluences and compaction at higher flences that scales with the energy lost by ionizatioprocesses.120 In addition, ion83,121 and electron122,123 ir-radiation of alkali silicate glasses induces alkali iomigration that modifies the glass composition in the irrdiated region. Such irradiation-induced alkali depletiocan bring about phase separation and low-temperatcrystallization, as observed in Li2O–2SiO2 glass.83 In-creased O2 evolution is observed in electron-irradiatesoda glasses relative to silica glass,42 and irradiation ofa 0.07Na2O–0.93SiO2 glass with 1.5 MeV electrons at200 ±C results in the formation of bubbles.124

2. Point defects

The defect centers induced in complex silicate (anborate) glasses involving alkali or alkaline-earth modfier oxides are created at rates which are higher by 1 toorders of magnitude than those observed in monocomnent SiO2 (or B2O3) glasses. This outcome is presumeto reflect the existence of many more electron and hotrapping sites in the complex glasses than are foundthe pure network glasses.

Radiation-induced defects in simple alkali silicatglasses include two types of nonbridging-oxygen hocenters, denoted HC1 and HC2.125–128 The former isanalogous to the NBOHC appearing in Eq. (4), if onpictures an alkali atom substituted for the hydrogen. HC2

is similar, except that it involves two nonbridging oxygens on the same silicon.126,128 These HC-type defectsare thought to be created by a process wherein a fhole first traps at a°°°°°° Si–O2Alk 1 site. The alkali ion(Alk 1) is then free to diffuse away to the site of atrapped electron, where it stabilizes the latter. BasedESR studies of alkali borate glasses,129 it is believedthat electrons and alkali ions trap sequentially at susites forming clusters that become increasingly stabletheir sizes increase. Although it has not been explicitdemonstrated, it seems possible that under sufficienhigh radiation doses, such alkali clusters may grow inmacroscopic colloids, analogous to processes knowntake place in NaCl.130

Some E0-type centers are observed in irradiatealkali silicate (and borate) glasses, although their numbers are relatively fewer that those of the other centementioned.127 TheseE0 centers are believed to be othe electrostatically neutral type, understandable by sustituting Alk0 for H0 in Eq. (5). Some bonded peroxyradicals have been revealed in the ESR spectra a



u-n

n-nre

di-

3o-

dlein

le

e

ee

on

haslytlytoto

d-

rs

b-

ter

the irradiated potassium silicate samples were anneaat ,270 ±C to diminish the intensities of the overlapping HC1 and HC2 signals.127 When these potassiumsilicate glass samples (g-irradiated to doses,106 Gy)were annealed to,310 ±C, the bonded peroxy radicacenters also disappeared. The remaining ESR specfeatures were unambiguously attributed to unbondsuperoxide (O22) and ozonide (O32) ions.127 The ob-served superoxide-ion ESR spectrum is believed tocharacteristic of O22 ions in a disordered alkali peroxidematrix58 (presumably K2O2 in the studied example).Here, then, is ESR evidence for a radiation-inducdecomposition of simple potassium silicate glasses inoxygen-rich peroxide phases on one hand and checally reduced phases on the other. The reduced phapresumably comprise SiOx (x , 2) or alkali metalcolloids, or both. There are well known tendencies131

for alkali peroxides to disproportionate into the monoide (Alk2O) and the superoxide (AlkO2), and for thesuperoxides in turn to disproportionate into the peroxidand O2 molecules. The appearance of ozonide ionsalso predicted in the course of these disproportionatsequences.131 Thus, this series of reactions could be thsource of the oxygen molecules suggested to because of radiation-induced bubbles observed in simulanuclear waste glasses.124,132–135

D. Borosilicate glasses


Shelby86 measured the effect ofg-irradiation on thedensity, refractive index, thermal expansion, and heliupermeability of several commercial borosilicate glacompositions. Compaction up to 1% was observedthe glasses after irradiation to 108 Gy. The refractiveindex increased with dose and the change in indwas linearly proportional to the change in density. Ththermal expansion coefficient decreased with dose awas linearly proportional to the change in density. Thhelium permeability was unaffected byg-irradiation to108 Gy.

In alkali borosilicate glasses, the high concentratioof alkalis generally “clogs up” the interstitial sites, sthat the initial effect of ion irradiation is to causeexpansion. The fluence regime over which this expansoccurs is proportional to the alkali content of the glass.120

Since the alkali content in the many borosilicate glassis known to be nonhomogeneously distributed (phaseparated),136 expansion and compaction are associatwith alkali-rich and alkali-depleted regions of the glasCompaction eventually dominates, as in fused silicafter the expansion in the alkali-rich phase is saturate

Utilizing cantilever-beam methodology and 0.to 4.0 MeV Xe ions, Snoecks and colleagues137,138

investigated the energy dependence of the lateral str

IP address: 144.82.107.169

2, No. 8, Aug 1997



e

i

n

e

d

to

s

ls

wro

j

t

ch-turetes;rectes.areare

n

an-ro-

rge

n

destent

or

gedith

ste

(volume change) in irradiated Na-borosilicate glassCompaction was observed at low energies (0.51.0 MeV), similar to the results of Arnold89,120,139;however, at higher energies (3.0 to 4.0 MeV), expansoccurred. These results suggest that between 1.0 Mand 3.0 MeV there may be a threshold in the electrostopping power for expansion in Na-borosilicate glassunder Xe irradiation. The electronic stopping powfor 2.0 MeV Xe ions is about 1.5 keVynm (based onTRIM-96) in these glasses, which is similar in magnituto the electronic stopping power threshold (2.3 keVynm)reported for anisotropic expansion in SiO2 thin filmsirradiated with Xe ions.140,141 Such a threshold may nobe relevant for nuclear waste glasses, since the electrstopping powers for thea-particles and recoil nucleiemitted by alpha decay will not exceed 0.5 keVynm.It was also observed in these Na-borosilicate glasse137

that compaction (up to 0.12%) occurred slowly wittime when the ion beam was turned off for both initialcompacting (1 MeV) and expanding (4 MeV) conditionThis was interpreted as a relaxation of the network dto annihilation of defects, possibly O-vacancies.

Significant evolution of O2 from commercialborosilicate glasses irradiated with 20 keV electrons hbeen measured.42,43 Oxygen bubble formation in highand low Na2O borosilicate glasses has been observeda result of irradiation with 1 MeV electrons135; the peakswelling temperature was 150±C in the high Na2O glassand 325±C in the low Na2O glass. Bubble formation inthese glasses was preceded by the migration of Na afrom the center of the irradiated region to the peripheLeach tests indicated higher leach rates for the regicontaining bubbles.

2. Point defects

Borosilicate glasses are expected to exhibit famlies of radiation-induced defect centers similar to thodescribed for silica and alkali silicate glasses, plusgroup of analogous defects which have been elucidain simple alkali borate glasses.58,142,143However, to dateno studies have been conducted which indicate whetsimple (or complex) borosilicate glasses are also subto radiation-induced disproportionation into peroxidesuperoxides, and ozonides.

E. Simulated waste glasses

The effects of radiation in nuclear waste glassare complex, and the fundamental understandingthe radiation damage processes and available datalimited. The high-radiation environment provided by tha-decay andb-decay of radionuclides in nuclear wastcan affect radionuclide release to the biosphere throuradiation-induced physical and chemical changes innuclear waste glass at both the atomic and macroscolevels. Several comprehensive reviews25–32 of radiation



s.to

oneVicesr

e

nic

hy.

ue

as

as

ayy.ns

i-se

ated

herects,

esofare

eeghhepic

effects in nuclear waste glasses provide excellent tenical assessments. There are no data on the naof intrinsic point defects or irradiation-induced poindefects in actual or simulated nuclear waste glasshowever, glasses for HLW and Pu immobilization aexpected to exhibit families of radiation-induced defecenters similar to those described for simple glassThe only data available on simulated waste glassesof measured macroscopic changes. These changesbriefly described below. More detailed information cabe found in the reviews25–32 and in the discussions ofspecific issues (Sec. IV).

1. Volume changes

Radiation effects froma-decay in actinide-dopedsimulated nuclear waste glasses results in either expsion or compaction of the glass structure. The macscopic volume changes,DVyVO, generally follow anexponential dependence on dose,D, of the form:

DVyVO Af1 2 exps2BDdg , (6)

where A is the volume change at saturation, andB isthe amount of glass damaged per unit dose. For a larange of glasses studied in the United States25,29,30,32,44

and Europe,28,144 the volume changes normally reach aapparent plateau that is within the range of61.2% at adose of 1 to 23 1018 a-decaysyg (1 3 109 Gy). Thisbehavior is illustrated in Fig. 8 for several actinide-dopenuclear waste glasses studied at the Pacific NorthwNational Laboratory and the Savannah River Site. Recstudies of actinide-doped glasses in Japan145,146 showirradiation-induced expansions of similar magnitude fsimilar doses. In one of these studies,145 the volumeexpansion determined by density measurements ranfrom 0.4 to 0.6% and was in excellent agreement w

FIG. 8. Volume changes in several actinide-doped nuclear waglasses studied at the Pacific Northwest National Laboratory25 andthe Savannah River Site.44

IP address: 144.82.107.169

2, No. 8, Aug 1997 1959



ab

x

re

re

w

xbn

es

ds

ee

a

te

a

sl

e

d

at

ytionanorior

forof

iorrgy

as

n-68)-

as

s

s.

lass

the volume change (0.51%) determined from the meured size and concentration of irradiation-induced bubles. These bubbles may contain He from thea-decayevent or radiolytically produced oxygen. The volume epansion and compaction of neutron,147,148 g-ray,29,44,84,85

ion,89 and electron84 irradiated nuclear waste glasses aalso within the bounds observed for the actinide-dopglasses. Manaraet al.135,149 and Satoet al.85 have re-ported swelling up to 50% when bubbles formed duing electron irradiation. In the simulated waste glassthere is a general tendency for expansion in glasses whigh alkaliySi ratios and compaction in glasses with loalkaliySi ratios.32 Another point of view, offered byStoneham,150 is that the density changes in complewaste glasses are due to a radiation-enhanced equilition to the ideal glass density for the glass compositioThis, of course, assumes that bubbles do not form.

The effect of irradiation temperature on volumchange has been studied in only one simulated waglass (VG 98y3) doped with 238Pu144; in that study,the rate of expansion at 130±C was less than thatat 20±C. The thermal recovery of radiation-inducevolume changes in several238Pu-doped waste glassealso has been studied, and recovery rate constantsthe glasses have been determined.144 The rate constantfor the glass VG 98y3 was used in a simple model tocalculate the volume changes at 130±C; the calculatedvalues were much smaller than the measured valusuggesting that the model (based on first-order recovkinetics) may be oversimplified. No further work onthe temperature dependence of volume changes in wglasses has been reported.

Based on these limited data, it is impossibledetermine conclusively whether the volume changinduced bya-decay are due to ballistic or ionizationprocesses. Ion-irradiation studies of Arnold89,121 haveindicated that ionization processes are the dominmechanism for producing volume changes in borosilicanuclear waste glasses, and this issue will be discusin greater detail (Sec. IV). In the absence of bubbformation, the magnitudes of the above volume changare small; however, the data do not extend to the muhigher dose range that will be experienced by glassused for Pu disposition.

2. Stored energy

The interactions of radiation with nuclear wastglasses can result in the storage of latent energy asciated with the changes in structure and bonding this released as heat at elevated temperatures. Stuon actinide-doped simulated nuclear waste glassesthe Pacific Northwest National Laboratory,25,151 AERE-Harwell,152 and under a collaborative European Communities program153 indicate thata-decay can result



s--

-

ed

-s,ith

ra-.

te

for

s,ry

ste

os

ntteedeesches

so-atiesat

-

in a rapid increase in stored energy that saturatesvalues generally less than 150 Jyg (,0.03 eVyatom)at a dose of 0.1 to 0.33 1018 a-decays/g (108 Gy),as illustrated in Fig. 9. A comparison of the energand volume change results indicates that the saturain stored energy occurs at a much lower dose thfor the volume changes, as illustrated in Fig. 10 fone simulated waste glass composition. Similar behavhas been observed for silica glass (Fig. 7) andall the simulated waste glasses studied, regardlesswhether there is compaction or expansion. This behavindicates that the defects giving rise to the stored eneprecede the network rearrangements that manifestvolume changes.

The effect of irradiation temperature on stored eergy has also been studied in only one glass (PNL 72-doped with 244Cm; the stored energy in this glass decreased linearly with irradiation temperature and wcompletely absent at 350±C.151 Thermal recovery ofthe stored energy in several244Cm-doped glasses ha

FIG. 9. Stored energy in several Cm-doped simulated waste glasse25

FIG. 10. Stored energy and volume change in a simulated waste gas a function of cumulative dose.

IP address: 144.82.107.169

2, No. 8, Aug 1997



gl

etb

s-u,oon

e

eo

aa

t

a

e

it

ob-

ed

u-er,er

ofard-

onisc-nn-

eaisas

ond,tynotna

ra-,

ein

sly

ith

also been investigated, and the results show a sin(nearly linear) recovery stage that is similar for athe glasses, with full recovery occurring at 360±C.25

The similarity of the temperature dependence for storenergy accumulation and recovery suggests thataccumulation of stored energy (defects) is controlledthe defect recovery kinetics.

3. Helium accumulation

As noted earlier, He atoms will accumulate asresult of the capture of two electrons bya-particles, andthe He concentration will increase with cumulative dofrom a-decay events. The diffusivities of He in simulated waste glasses are one to two orders of magnitlower than in silica, alkali silicate, and Pyrex glasses41

and they have been reported to decrease by up tseveral orders of magnitude with increasing radiatidose, which suggests trapping of helium at radiatioinduced defects.40,52 Studies of He diffusion and releasin several simulated waste glasses40,52,152,153indicate thatmost of the He generated at ambient temperature waccumulate in the glass, and only a small fraction of Hwill be released. However, the fraction of He releasincreases with irradiation temperature, and the fractireleased at 170±C can range from 50 to 100%.152,153

Bubbles (presumed to be He) have been observedsimulated waste glasses containing 100 appm of Hintroduced by (n, a) reactions during neutron irradiationor by ion implantation, but only after annealing at temperatures between 600 and 700±C.52,154 When higher Heconcentrations (up to 104 appm) are introduced by(n, a)reactions, bubble formation has been observed at lowtemperatures (,100 to 230±C).148,154 Bubbles have alsobeen observed in a glass containing both238Pu and244Cmafter 8 3 1018 a-decaysyg at ambient temperature,145

and, as noted above, the integrated volume of the bubbcorrelated with the measured volume change. The exnature of the bubbles observed has not determined inof these studies, but are presumed to be He or HeyO2.

4. Microstructural changes

Phase separation has been induced in simulawaste glasses by electron irradiation.132,133,155Radiation-induced amorphization of crystalline phases in a244Cm-doped simulated waste glass has been observed,the volume expansion associated with this transformtion produced significant microcracking.25,156 As notedabove, a-decay145 and neutron irradiations to induce(n, a) reactions148,154 in waste glasses can result in thformation of bubbles.

In 1976, Hall and co-workers152 were the first toreport the formation of bubbles, assumed to contaO2, in simulated nuclear waste glasses irradiated welectrons in a high voltage electron microscope (HVEM



lel

dhey

a

e

de

an-

illedn

ine,

-

er

lesctny

ed

nda-

inh)

at 200± C to high doses (1010 to 1011 Gy) in experimentsto simulate the effects ofb-irradiation. Since then,electron-beam-induced bubble formation has beenserved in a wide variety of glasses.85,132–135,149,157–160Thepresence of a gas phase in the bubbles was confirmin the studies of Hallet al.152 and Heueret al.134 Ion-beam irradiation124,158,159andg-irradiation155,158–160alsohave been reported to produce bubbles in several simlated waste glasses, as illustrated in Fig. 11. Howevg-irradiation of one of these same glasses in anothstudy161,162 did not produce bubbles, even at an ordermagnitude higher dose. This has raised questions reging whether or not the bubble formation ing-irradiatedglasses reported by others155,158–160 may be an artifactof sample preparation or the electron-beam irradiatiduring examination using electron microscopy. In thregard, it is important to note that there is a charateristic difference in the bubble sizes and distributiofound in g-irradiated glasses as compared to electroirradiated glasses.155,158Gamma irradiation leads to moreuniform bubble sizes and distributions over a large ar(i.e., the whole sample or at least the area thattransparent to electrons in the microscope), as wellbubbles in very thin regions as observed by electrmicroscopy. Electron irradiation, on the other hanproduces a nonuniform distribution of bubble densiand bubble sizes only in a focused beam; bubbles canbe produced in very thin regions by electron irradiatioalone. Another point is that bubbles are observed ing-irradiated glass by electron microscopy at tempetures where bubbles are difficult, if not impossibleto form under intense electron irradiation.155,158 Con-sequently, if bubbles do not form directly under thg-irradiations, then the glass must be highly damagedsuch a way that the bubbles form nearly instantaneouwhen exposed to an unfocused electron beam.

FIG. 11. Bubbles formed in simulated DWPF glass irradiated wg-rays to 1.33 107 Gy (micrograph courtesy of J. P. Heuer and D. GHowitt, University of California–Davis).

IP address: 144.82.107.169

2, No. 8, Aug 1997 1961



e-ses

sesa

te

ndalnat

,rshwte

re

esforg-

fnc-of

on-rk

estsmin-ra-

up

f

tsate

Quantitative studies have shown that the kinetiof the radiolytic bubble formation process are constent with the motion of alkali metal cations and nooxygen.133,163The formation of oxygen bubbles is appaently brought about by radiolytic decomposition of thionic component of the glass, followed by the migratioof the cations into the glass and the local precipitatiof molecular oxygen.160 Irradiation studies158,159 utiliz-ing g-rays, ion beams, and electron beams at simitemperatures, indicate a significant dose-rate effectradiolytic bubble formation, as shown in Fig. 12. Threported bubble formation underg-irradiation occurs atmuch lower doses (106 to 107 Gy) than for either ionirradiation (107 to 1010 Gy) or electron irradiations (1010

to 1013 Gy). Under electron irradiation, bubble formatiooccurs over a broad temperature range with a maximin the rate of formation at 250±C.133,159Similar behavioris reported by Manaraet al.,135 who also observed thatthe peak for bubble formation occurs at a lower temperature (125 to 150±C) for high Na glasses (19 and25 wt. % Na2O) than the peak temperature (325±C) fora low Na glass (4 wt. % Na2O).

5. Mechanical properties

In a recent study of a simulated nuclear wasglass doped with238Pu and244Cm,164 the hardness andYoung’s modulus decreased exponentially with doswhile the fracture toughness increased exponentiawith dose. The maximum values of the relative changreported for the hardness, Young’s modulus, and fracttoughness in these glasses were225%, 230%, and145%, respectively. These changes in fracture toughnstrongly correlated to the formation of bubbles whicwere observed in these glasses.145 The bubbles, whichcan impede crack propagation, anneal with kinetics sim

FIG. 12. Dose and dose-rate dependence of radiolytic bubble formtion in simulated HLW glasses (adapted from DeNatale and Howitt158

and Heuer159).



csis-t

r-en

on

larfore

num

-

te

e,llyesure

essh

i-

a-

lar to the recovery of the fracture toughness. Similar dcreases in hardness and Young’s modulus and increain fracture toughness have been observed due toa-decaydamage in three different glasses containing244Cm.165–167

Irradiation-induced decreases in hardness and increain fracture toughness also have been reported innumber of other studies.168–170

6. Enhanced diffusion

Yamashita and Matzke171 investigated radiationenhancement of Na diffusion in a simulated wasglass (GP 98y12) under irradiation at relatively lowdose rates witha-particles emitted from244Cm and239Pu sources. (The244Cm irradiations simulate doserates for accelerated actinide-doping experiments athe 239Pu irradiations simulate dose rates for actuHLW glasses.) Although Na depletion is observed ithis glass under high dose rate electron irradiationroom temperature,172 these investigators found thatwithin experimental error, there was no evidence fosignificant radiation-enhanced Na diffusion in this glasunder the more realistic dose-rate conditions. Althougthe experimental errors are considerable at the lotemperature conditions of the study, the results indicathat the radiation-enhanced Na diffusion is no mothan a factor of 2 to 5 at temperatures below 150±C.

7. Radionuclide release

Radiation affects the release rate of radionuclidfrom waste glasses by increasing the surface arearadionuclide release (microfracturing) and by chaning the dissolution rate of the glass. The extent othe radiation-induced microfracturing will depend odifferential volume changes, development of microstrutures, and mechanical properties. The dissolution ratenuclear waste glasses may be affected by the radiatiinduced changes in chemistry, microstructure, netwobonding, and alonga-recoil tracks. As noted in sev-eral reviews,25–27,29–32,173almost all data concerning theeffects ofa-decay on dissolution (leach) rates of wastglasses were determined prior to 1982 in short-term tebased solely on weight loss. The data obtained frothese tests indicate no more than a factor of threecrease in the short-term dissolution rates as a result ofdiation effects froma-decay. Ion-irradiation149,174,175andneutron-irradiation176 studies of several HLW glasseshave shown irradiation-enhanced dissolution rates ofto a factor of four. Ion irradiation, however, is knownto introduce compositional variations,51 which affectthe interpretation of the results. Gamma-irradiation oseveral HLW glasses up to 109 Gy also leads to increasesin dissolution rates of up to a factor of four.44,177,178

It is now well established that the short-term tesbased on weight loss in these studies can underestim

IP address: 144.82.107.169

12, No. 8, Aug 1997



1ars

ob

ho

e

i

t

s

d

o

e

tci

c

a

ef-cals-

lexnrmir-ar

A)

asesr-ss

n,rsgtheser-i-

edves,

-edc-n

syth

ors inger

ndsmeay

-

sesy.nd

beinge

the radiation-induced increases in dissolution rate byfactor of three to four due to precipitation of alteratiophases on the surface of the glass29,179; consequently,based on the limited data and understanding availablehas been suggested by Weber29,30 that radiation-inducedchanges in HLW glass structure are not expectedincrease the leach rate by more than a factor ofprovided there are no radiation-induced phase septions or transformations. This assessment is suppoby a study49 of a seventeen-year-old borosilicate glacontaining 4 wt. % ThO2, where the Th daughter producrelease rate was increased 15 to 45% due to incgruent dissolution along the recoil tracks producedthe a-recoil nucleus. For thea-decay damaged glassthe initial 228Th leach rate was enhanced relative to tinitial 232Th leach rate due to the selective removalradiogenic228Th nuclides from228Ra recoil tracks (232Thdecays to228Ra by a-decay and the228Ra b-decays to228Th). Heating at 650±C prior to leaching eliminatesthe 228Thy232Th fractionation. Thus, it is evident that tha-recoil tracks in glass represent localized regionsseverely damaged, chemically reactive, material. Simlar phenomena have been well described in crystallmaterials.50,180,181 Another recent study,166 however, ofHLW glasses containing244Cm has shown that radiationdamage froma-decay does not significantly increase thbulk leach rate when measured by solution analysis,recommended by a previous workshop.23 The effects ofradiation-induced structural changes on glass dissolurange from insignificant to significant, depending oglass composition and what is actually measured. Aresult, this subject remains highly controversial andneed of systematic investigation.

IV. FUNDAMENTAL SCIENTIFIC ISSUES

Glass waste forms for the DWPF, WVDP, Hanforand weapons plutonium applications represent compmulticomponent systems. Though the compositiranges of borosilicate glass for DWPF and WVDP awell established, possible glass compositions for Haford HLW and weapons plutonium immobilization havnot yet been determined, and several competing glasscrystalline formulations are under consideration. Evfor DWPF, the waste vitrification campaign will proceefor over 20 years, the glass canisters must be stored ua repository is opened, and the repository will probabremain open for less than 100 years, at which timewill be sealed. Thus, there is time to identify and addrefundamental scientific issues, in particular those relato the effect of radiation on waste form properties, whimay have a bearing on cost, optimization of processtechnologies, the integrity of the HLW glass canisteuntil the repository is sealed, and long-term performan

In the following discussion, critical areas are oulined for which the panel determined that there is



an

, it

to0,ra-teds

tn-y

,ef

ofi-

ne

eas

ionn

ain

,lexn

ren-eandn

dntillyit

ssedhngrse.

t-n

absence of a systematic understanding of radiationfects on waste form glasses. The phenomenologisystematics require that the knowledge of simple glasradiation interactions be extended to glasses of compcompositions. The scientific basis for the extrapolatioof radiation effects data and models based on short-teexperimental studies (actinide-doping and ion-beamradiations) to the longer time periods relevant to nuclewaste disposal is reviewed.

A. Glass structure and energetics

Nuclear waste glasses are structurally complex.borosilicate glass contains network formers (Si, Al, Band network modifiers (alkalis and alkaline earths),well as large-size, high-valence-state cations (lanthanidand actinides) that occupy irregular sites of large coodination number and that generally are not good glaformers.182 The oxidation states of ions such as Fe, MCe, U, and Pu are variable. The ratio of network formeto network modifiers is such that the ratio of bridginto nonbridging oxygens varies. Boron can exist in botetrahedral and trigonal coordination. Structural studion simpler model glasses have produced a considable knowledge of short-range order and local coordnation, but few structural studies have been completon simulated or actual nuclear waste glasses themseleither before or after the accumulation of radiation damage. As noted above, there is a need for an improvmethod to describe glass structure and for more strutural data on all potential HLW glass compositions thacurrently exists.

What is clear is that such HLW melts and glasseare not very far from regions of glass-glass immiscibilitinherent to borosilicate glasses and to high field strengionic glasses.183 For La2O3 dissolved into potassium sili-cate glasses, Ellison and Navrotsky182 have argued thatthe enthalpy of mixing suggests extensive clusteringincipient phase separation. Radiation-induced changeglass energetics enhance heterogeneity and may trigphase separation.

Another possible effect of radiation damage in HLWglasses is mixing. Because atoms are displaced and bobroken, a more homogeneous structure may result. Soof the volume changes seen in actual waste glasses mreflect relaxation to a more stable structure.150 A studyof the glass transition in As2Se3 suggests that radiationdamage allows the glass to relax to a lower fictive temperature and an energetically more stable state.184 Thusit is not clear under what conditions radiation raisethe free energy of a glass and in what circumstancit allows the glass to find a state of lower free energThe possible kinetic enhancement of crystallization astable phase separation by such bond breaking mayanother manifestation of the latter case. Understandthe direction in which radiation-solid interactions driv

IP address: 144.82.107.169

2, No. 8, Aug 1997 1963



r

a

b

eo

aoett

o

s

y

iai

r-t

d

g

air

elts

ng,

nhe

nusei-s,nin

.

sm-ies

esghe

centn

mefy

ssnic

m-h

evy

ionses

nderustual

aer

waste glass energetics, toward chemical equilibriumaway from it, iscritical to understanding the long-termbehavior and durability.Another possibility is that, un-der steady state irradiation, the actual equilibrium stamay be altered, with different phases being stabilizeThe effects of radiation on glass energetics require cosiderable additional study. Because of the possible copeting mechanisms of defect production and recovethe effects of temperature are likely to be complex.

When either quartz or silica glass is neutron irradated, its energy increases. Quartz expands under irration, while silica glass compacts. Both crystalline anglassy silica appear to approach a similar (or perhaidentical) final density state or plateau characterized3% densification relative to silica glass and 18 kJymolexcess enthalpy relative to quartz.185 Spectroscopicallyand energetically, these irradiated glasses are differfrom fused silica and from pressure-densified silicacomparable density.185–188 Does damage by other formsof radiation follow similar trends? If radionuclides innuclear glasses are partitioned into different glass phasdoes the radiation (a-particles, a-recoils, b-particles,g-rays) result in similar or different changes in physicand chemical properties? Answers to these questi(and those posed below) will require a systematic undstanding of the structure, phase equilibria, and energeof both undamaged and radiation-damaged glassesrange in composition from simple to complex. Obtaining this understanding will require the application omodern characterization techniques to studies of bnonradioactive and radioactive glasses. These techniqinclude, but are not limited to, electron-spin resonancnuclear magnetic resonance, Raman spectroscopy,chrotron radiation spectroscopies, neutron and x-ray dfraction, analytical electron microscopy, and calorimetr

B. Ionization versus ballistic effects

1. Electron- and X-irradiations

E0-centers in fused silica glass are associated wESR (electron spin resonance)-active oxygen vaccies.189 An early attempt was made to measure the dplacement energy required to form the O-vacancyfused silica by monitoring the associated optical absotion (215 nm) as a function of a given flux of electrons at varying energies (0.5–2.0 MeV) normalizedthe absorption at 2.0 MeV.190 The concentration ofO-vacancies measured in this manner increased withcreasing electron energy and, indeed, correspondedmost entirely to the energy lost by the electrons throuionization processes as they traversed the silica saples. X-ray (50 kV) irradiations for doses matching thosof the electron-irradiated samples confirmed thO-vacancies could be produced by purely ionizingradiation and that the nonsaturating absorption show



or

ted.n-m-y,

i-di-dpsy

ntf

es,

lnsr-

icshat-fthuese,yn-if-.

thn-s-inp-

o

e-al-hm-et-ed

that they involved vacancy creation and not simply thfilling of existing vacancies by electrons. These resuwere later confirmed by Pfeffer191 in studies on thermaloxides that showed the effects of H in damage annealia factor that was noted in prior work.190 Still furtherconclusive evidence of the dominant role of ionizatioin producing actual displacements was provided by twork of Tsai and Griscom109 who showed that the de-cay of self-trapped excitons produced by a two-photoprocess (6.4 eV excimer laser photons) alone can cadisplacements of O atoms from the silica network. Simlar studies of HLW glasses that employ electrons, x-rayandg-rays are needed to verify whether or not ionizatioprocesses are an important mode of radiation damagethese complex HLW glasses, as has been suggested

2. Ion-irradiations

The dominance of ballistic (collisional) processeover electronic processes for producing ion-beam daage in fused silica glasses was established in studby EerNisse81 and by Presby and Brown.82 Both ofthese investigations estimated that ballistic processare about a factor of 200 more effective in producindamage than electronic processes in silica glass. Texperimental approach of EerNisse81 (i.e., measuring thecantilever-beam deflections produced by induced surfastress) has been exploited for determining the dominamechanism for producing defects under ion irradiatioin a variety of glass systems. The stress data (voluchanges) in fused silica irradiated with incident ions ovarying mass were found to scale with ballistic energdeposition.81,89 In simple alkali (Na, K) borosilicates,Pyrex (commercial borosilicate), and soda-lime gla(microscope slide), the stress values scale with electroenergy deposition.89,120,137,139For the alkali-borosilicatenuclear waste glasses, PNL 76-68 and DWPF, the daage from ion irradiation (He, Xe, Pb) also scales witenergy deposition in electronic processes.89 For a-decay,this means that the 5 MeVa-particle, which losesmost of its energy through ionization events, may bof greater importance rather than the 100 keV hearecoil in causing permanent radiation damage in HLWglass compositions. It should be noted that the saturatvalues for stress in the implanted nuclear waste glasoccur for deposited energies equivalent to 108 –109 Gy.Although ionization appears to be the dominant radiatiodamage process for borosilicate waste glasses unhigh-dose-rate electron and ion bombardment, this mbe established for the much lower dose rates that actglasses will experience froma and b decay.

C. Relaxation processes and diffusion1. Relaxation processes

The subject of relaxation phenomena in glass iscomplex one, and the effects of radiation must furth

IP address: 144.82.107.169

2, No. 8, Aug 1997



ris

lltobo

e

hcdew

yedd

d

o

a

o

ep

cs

-m

eso

to

s

an

liellnnede

h

nc-

eionss

nt

dndr-hs

delnt

at

dsillin

re

is

ns,bewr

orghes.

toe

ti--nddagehis

complicate the process. At the glass transition tempeture, Tg, the fundamental enthalpy relaxation timeabout 100 s. AboveTg, radiation should have little ef-fect on the properties of a glass because of the shtime scale (,100 s) on which radiation damage is relaxed. NearTg, the stabilization of the glass structure wigreatly reduce the mobility of any species that are notdecoupled from the glass structure; this is presumarelated to the rapid decreasing probability of percolatipaths for such species as the glass structure compaThe overall structure of the glass becomes relativfixed at temperatures much belowTg; however, certainmodes of motion remain active. These are modes thave been split off (decoupled) from the principal strutural relaxation modes during cooling within the liquistate and continue to be active in the glass. The bstudied of such modes are those due to small, locharged ions such as Li1, Ag1, and (in fluoride glasses)F2, which can move via low activation energy pathwawithin the liquid and, at lower temperatures, within thglass structure. Most data come from conductivity stuies, though NMR and diffusivity have also been studie

Similar decoupling processes must occur for the mbility of network modifiers, especially the alkalis, anfor the gas atoms or molecules (i.e., He and O2) that aregenerated. The decoupled motion of these species shpersist to temperatures much lower thanTg. Understand-ing the motion of alkali ions and gas species in relationthe structural relaxation of the glass is especially lacing. Radiation may affect alkali mobility, as discussebelow, and this should be studied by diffusion, electricconductivity, and NMR measurements of original anirradiated glasses. The understanding of transport prerties in a glass under a strong radiation field shouldinvestigated using both computer simulations and expimental techniques. There are no studies of these tyof decoupling processes over long periods of time.

2. Diffusion

Both ionizing and displacive radiation can enhanthe diffusion of ionic and gaseous species in the glaThis is because “bottlenecks” in the glass structure cbe temporarily opened or reorganized either by locionization (modifying Coulombic barriers) or by displacement of blocking species—or simply by short-terlocal heating due to nearby radiation tracks. Radiatioinduced expansion of the glass structure may also lto enhanced diffusion of some species. In some glashowever, radiation can cause annealing that would nmally lead to decreased mobilities for small ions angaseous species. As2Se3, for instance, can be annealeda lower enthalpy state under neutron radiation belowTg

than could be obtained by normal thermal annealing.184

Which conditions will apply to the glass composition



a-

ort-

olyncts.ly

at-

st-

s

-.

o-

uld

tok-dl

dp-

ber-es

es.

anal

n-ades,r-

d

under consideration for radioactive waste disposal cbe determined only by experiment.

Marked modification of the near-surface alkaconcentrations in electron-irradiated glasses is wknown (e.g., it is a disturbing feature of Auger electrospectroscopy).192–194 For these effects, the changes ialkali compositional depth profiles appear to be governby ordinary and electric-field assisted diffusion where thelectric field is a function of the irradiation depth botfor electron- and proton-irradiated glasses.194,195 Thesurface depletion of alkali results in an accumulatioof alkali at a depth roughly corresponding to the eletron range.122,123 For MeV proton-irradiation, the alkalimovement is from the irradiated depth toward thsurface, resulting in an abnormal surface concentratof alkalis. The early stages of nuclear waste gladissolution involve ion exchange reactions (e.g., Na$

H3O), thus near surface increases in the alkali contemay affect the dissolution rate.

For heavy-ion irradiations, alkali ions are removefrom depths corresponding to the heavy-ion range, afor continued irradiation to high fluences, the entire neasurface region can be depleted of alkali ions to deptexceeding the ion range.121 These heavy-ion effectshave been interpreted using a phenomenological mobased on radiation-enhanced migration over the incideion range and preferential sputtering of the alkalithe surface.

Almost all the studies to date of radiation-enhancealkali diffusion have involved high dose rate irradiationwith electrons or ions. Actual nuclear waste glasses wbe exposed to much lower radiation dose rates. Onlythe study of Yamashita and Matzke171 has the radiation-enhanced diffusion of Na been studied under morealistic dose rates witha-particles emitted from244Cmand 239Pu sources. In this case, Na depletion, whichobserved in this glass under electron irradiation,172 isnot observed under the much lower dose-rate conditioand the enhancement of Na diffusion is estimated tono more than a factor of 2 to 5 at temperatures belo150 ±C. Although similar studies are needed for otheglass compositions, this work highlights the need fstudies at realistic dose rates combined with a thorouunderstanding of dose rate effects on various process

D. Temperature and dose rate effects

Accelerated irradiation methods must be usedachieve repository relevant doses in laboratory timframes. These methods include incorporation of acnides,g-irradiation, electron irradiation, and ion irradiation. At the high dose rates employed under electron aion irradiations, the effects of irradiation are observeat much higher temperatures because the high damrates overwhelm the thermal recovery processes. T

IP address: 144.82.107.169

2, No. 8, Aug 1997 1965



nniainoo

n

a

tirema

uWcnr

izs

utu

b

h

nble;d

bee-ionn-us

nse,he

lexestSoft

elycalses

naldrnd

he-edns,

nsch

ionnttingesce

nethe

ied, ifan

eep-edn,vetoL

che

effect is mitigated to some extent by the irradiatioassisted recovery processes that are also observed uelectron and ion irradiations. One of the key questionsIs there a dose rate below which the effects of dose rare negligible? A systematic, integrated understandof dose-rate effects is required for the extrapolationdata obtained using accelerated methods under high drates to the actual long-term, low dose rate behaviornuclear waste glasses.

A detailed study196 of radiation-induced amor-phization in zircon showed an excellent correlatiobetween heavy-ion (Kr and Xe) irradiations at 25±C anda-decay effects at ambient temperature in238Pu-containing synthetic zircon (6.5 years old) and UyTh-bearing natural zircons (570 million years old),difference in dose rates of over 1013. This correlation isobserved in the temperature regime where amorphizaof zircon does not exhibit significant temperatudependence, which suggests that dose-rate effectsnot be important at temperatures where the damis temperature independent (i.e., recovery processesnegligible). Such correlations have not yet been shownirradiation damage studies of glasses, and temperatindependent regimes for irradiation damage in HLglasses have not been identified. If radiation effein HLW glasses exhibit some temperature dependeat the anticipated storage temperatures, then dose-effects will generally be important.

E. Radiation-induced transformations

Possible radiation-induced phase transformationsclude phase separation, devitrification, and amorphition of crystalline phase impurities present in the glaor formed during cooling.

1. Phase separation

Phase separation (immiscibility) occurs over a limited temperature-composition domain. The curve (sface) that delimits the one phase (homogeneous) fromtwo phase region is termed the binodal. The temperatassociated with the highest binodal temperature is termthe consolute (or critical) temperature, and is denotedTc. AboveTc, the single phase liquid is the equilibriumstate for all compositions. IfTc occurs above the liquidustemperature (Tl), then the portion of the immiscibilityregion that occurs aboveTl is stable, and the equilibriumstate corresponds to the two liquid phase region. Tportion of the immiscibility domain that falls belowTl

is metastable, since the two liquids do not correspoto the equilibrium state. Hence, in regions of metastaimmiscibility, further phase transformations are possibStable immiscibility usually is produced quite rapidlythus, it is exceedingly difficult to avoid (even by rapiquenching methods).



-der

s:tegfse

of

on

aygeareinre-

tsceate

n-a-s

-r-hereedy

e

dle.

Phase separation is of importance since it candeleterious to the chemical durability of the glass. Dpending upon chemical composition, phase separatcan produce two types of morphologies: (i) an interconected structure, or (ii) a droplet structure in a continuomatrix.197,198 If immiscibility occurs and a structure oftype (i) is formed, or a structure of type (ii) results iwhich the droplets correspond to the more durable phathen phase separation will result in a degradation of tchemical durability of the glass.

Since many nuclear waste glasses are compsodium borosilicate (NBS) based glasses, it is of interto briefly review the phase separation behavior of NBglass. The NBS system exhibits a broad region (partially stable) immiscibility extending up to abou16 mole % soda and for nearly all SiO2 compositionsin excess of 20 mole %.199,200 Phase equilibrium studieshave indicated that typically one phase is an extremsilica poor phase which possesses inferior chemidurability, and the other is a chemically durable phaconsisting nearly of pure vitreous silica. Thus, adiscussed above, there exists a wide compositiorange for which NBS glass durability will be degradeby immiscibility. This undesirable feature is furtheexacerbated by the fact that most fission products teto partition into the less durable glassy phase.

Thus, one important scientific issue pertains to tinfluence of glass composition upon immiscibility behavior. Although considerable work has been performupon the influence of divalent and trivalent ions upophase separation in alkali borosilicate glass systemmuch less is known concerning the effects of catioof valence four or greater on phase separation in susystems. A second issue relates to the lack of informatregarding the influence of melting history upon the exteand nature of phase separation in these glasses. Meltime, temperature, redox conditions, and cooling ratare several specific factors which strongly influenimmiscibility behavior, especially if the compositioncontains cations which can occur in more than ovalence state (e.g., Fe, Pu, U). The valence states andeffects of irradiation on the valence states can be studby electron spin resonance techniques. For examplethe total iron in a glass sample should be known,ESR determination of the Fe31 content can be sufficientto determine the Fe31yFe21 redox ratio.201

An important consideration is the influence of thradiation field and the radiation products upon phase saration. There are scant data detailing radiation-inducchanges in composition or compositional redistributiobecause few of the requisite studies appear to habeen done. Radiolysis in the TEM has been observedinduce phase separation (possibly spinodally) in a PN76-68 waste glass into silica-rich and heavy-metal-ricomponents,132 and also to sharpen initially diffuse phas

IP address: 144.82.107.169

2, No. 8, Aug 1997



s

oiw

ri

yy-e

gulittivei

e

eth

nf

f

s

hmT

in-

itiomu

lob

mse

ong,st

na-

)-to

ne

e

nory

cedente.edr-mentis

ee

F-pt-

inesed-edretoineor-slys

nd-ant

boundaries in a K2O–Al2O3 –CaO– SiO2 glass derivedfrom earlier spinodal decomposition.202 Because phaseseparation is known to occur thermally in both theglasses, it is likely that radiation-enhanced diffusion simply accelerates the process. The potential for irradiatito induce phase separation more generally in otherwstable glass compositions is presently unknown; hoever, implantation of 2PbO–P2O5 glass with P1 ionshas been shown to result in phase separation of Pb-regions.203

Even if the as-prepared glass does not exhibit phaseparation, radiation effects could induce immiscibilitespecially if the glass composition lies in the vicinitof a region of immiscibility. Several possible conditions could produce this effect. First, collision cascadcould produce high energy states characterized byaltered number of nonbridging oxygens and a chanin average boron coordination number. This state cobe metastable with respect to a two-phase (nonequirium) state, and phase separation could ensue abeby radiation-enhanced diffusion. A second possibiliis that radiation-induced heating could cause selectvolatilization, thus altering the glass composition. If thas-prepared composition is near a region of immiscibity, then this composition shift could produce a glasof composition within the binodal. Furthermore, thelevated temperature would enhance the phase separakinetics. Finally, one must consider the influence of thradiation products. Since the radioactive decay procalters the glass composition, one must considerimmiscibility behavior of systems which contain newchemical components. The above effects must not obe examined individually, but also the possibility othese influences acting in a synergistic fashion mube studied.

The discussion given above suggests the needexperimental investigation of the thermodynamic ankinetic aspects of phase separation in unirradiated airradiated glasses. One should investigate three typeglass with compositions of increasing complexity. Thsimplest glass should be a binary glass composition texhibits phase separation, and whose thermodynaand phase separation features are well characterized.influence of radiation upon a composition that fallsthe vicinity, but outside, of the binodal should be investigated. The composition of intermediate complexshould be a 4 or 5 component NBS based compositwhich contains at least one cation which can assutwo (or more) valence states. For this glass, one shoinvestigate the influence of melting conditions and compositional excursions on immiscibility behavior, as weas the influence of radiation effects, especially ionizatieffects. The most complex glass composition shoulda multicomponent prototype waste glass compositioThis composition should be examined in a manner th



e-n

se-

ch

se,

sane

ldb-tedye

l-s

tionesse

ly

st

ordndof

eatiche

yneld-

lne

n.at

would be dependent upon the information gained frothe studies of glasses of intermediate complexity. Phaseparation can be monitored via the use of electrmicroscopy (TEM and SEM), selective phase etchinand small angle x-ray scattering (SAXS). For the moreliable experimental study, one should use a combition of these techniques.

2. Devitrification

The formation of crystalline phases (devitrificationduring cooling of the waste glass from the melt temperature is generally undesirable, as it often leadsdecreased chemical durability.204 Waste glass composi-tions are generally optimized to minimize the formatioof such crystals. The center line cooling rate of thglass is on the order of 10±Cyh until the tempera-ture reaches 200 to 250±C, at which time self-heatingfrom radionuclide decay slows the cooling rate. Thtemperature of the glass will generally fall to below150 ±C after several hundred years. Devitrification iwaste glass compositions is not observed in laboratstudies below 525±C204; consequently, devitrification isgenerally assumed not to occur once the glass is plain storage. Radiation-induced devitrification has not beobserved in any borosilicate glasses studied to daNear the glass transition temperature, radiation-inducdevitrification may be more probable as radiation inteaction processes provide additional energy to overcoany activation barrier. However, at repository-relevastorage temperatures, radiation-induced devitrificationless likely. The only exception to this would be thprecipitation of alkali metal crystals, such as Na, duto radiolytic decomposition and diffusion.

3. Amorphization of crystalline phases

High-level waste glasses being produced at DWPand WVDP will not contain a significant amount of secondary crystalline phases according to current acceance criteria. Such restrictions on secondary crystallphase formation could be relaxed for glasses propofor immobilizing HLW at Hanford and the Idaho National Engineering Laboratory or for glasses proposfor plutonium disposition. If such crystalline phases apresent and radionuclides partition preferentially inthese phases, then radiation effects in the crystallphases will be an issue. Alpha-decay induced amphization of such crystalline phases has been previouobserved25,156 and is accompanied by volume changethat can exceed 15%,32 which can lead to microcracking.There are a significant amount of data and understaing of amorphization in crystalline phases, as this isactive area of research,32,205 that can be used to predicthe effects of radiation on such phases.

IP address: 144.82.107.169

2, No. 8, Aug 1997 1967



g

m

eo

o

inc

o

elh

t

nands

oin

i

eo

u

h

n

n

e

ro-

ingdif-ects

ly

sofcts

om-heesto

nnd

ionin

es-er-

asies,hede

r-d

atebe

eb-

-

ren

d,erre,or.ingblea

ved

F. Volume changes

At low to intermediate dose levels, there is a sinificant amount of data on radiation-induced volumchanges in glasses ranging from fused silica to coplex borosilicate waste glasses, as discussed above.volume changes in bulk-irradiated specimens (actinidcontaining, g-irradiated, or neutron-irradiated) can bdetermined easily from density changes. For electror ion irradiations, the cantilever-beam technique81 hasbeen used very successfully. The radiation-induced vume changes are the result of the accumulationpoint defects, network rearrangements and rebondand the development of microstructural defects suas gas bubbles and voids. At relatively low dosethe accumulation of point defects in both simple ancomplex glasses, as observed by ESR, spectrosctechniques, and stored energy, reaches steady statefect concentrations without significant volume chang(Figs. 7 and 10). Following the saturation of the simpdefect concentrations, the compaction or expansion tis measured is generally within the range of61.2%.Whether compaction or expansion occurs at lowintermediate dose levels may be largely dependentcomposition, processing conditions, and whether orphase separation occurs. These changes are probthe results of network rearrangements and rebondias suggested by Devine.60 They also may be associatewith the radiation-enhanced equilibration of the glato its ideal density for the given composition.150 Fewdata are available on the effects of temperaturevolume changes. A more fundamental understandof the compaction or expansion that occurs at thedose levels is desirable in order to model and predbehavior at larger doses, such as expected for Pu dissition. Significantly larger compaction at higher dosesprobably not feasible, unless there is radiation-inducdevitrification. Volume expansion of the compactedexpanded structures, however, can result at higher dofrom the formation of gas bubbles or voids as a resof the He that is generated froma-decay and/or theradiolytically generated O2. More experimental researchis needed on the volume changes at higha-decay dosesas a function of temperature in glasses for Pu and otactinide disposition, and temperature dependent modare needed to describe the network rearrangementsformation of bubbles or voids.

G. Gas accumulation and bubble or voidformation

The oxygen vacancy centers, nonbridging oxygeand neutral alkali atoms, together with He atoms (geerated froma-decay) and possibly molecular oxygefrom disproportionation, comprise theobservablepri-mary defect spectrum of alkali silicate-based glass



-e

-Thee-

n

l-ofg,h

s,dpicde-s

eat

oonotblyg,

s

ng

sectpo-isdrseslt

erelsand

sn-

s.

Other point defects, which are not observable spectscopically, could be generated byb –g radiolysis andtransiently in the displacement cascades accompanya-recoil. These defects and gas atoms/molecules canfuse, interact, and aggregate to become trapped at defor form bubbles/voids. Existing models206 developed forbubble growth in metals under irradiation can probabbe modified to model the phenomena in glasses.

As noted above, radiation damage in alkali silicatemay produce clusters of alkali atoms and evolutionoxygen gas as endproducts of a complex series of defeinvolving peroxide and superoxide ions. In addition tpotential problems of gas bubble formation, such decoposition can create compositional heterogeneity in tglass. An important issue is whether similar processoccur in complex waste glasses and do they leadthe formation of bubbles or voids? The applicatioof electron spin resonance, Raman spectroscopy, aother advanced characterization techniques to radiatdamage studies of complex waste glasses will assistanswering these questions.

There is no debate over whether or not bubblform under electron and ion irradiation; the only controversies are whether or not bubbles actually form undg-irradiation and whether or not the electron- and ionirradiation results are relevant. Surprisingly, there hbeen no research effort to resolve these controverswhich have lingered for nearly a decade, despite tfact that careful experimentation could probably leato resolution. There are three important points from thg-irradiation studies of bubble formation that are impotant to note. First, the minimum electron fluence (aroun5 3 1020 eym2) to image 5 nm bubbles in a TEMamounts to an ionization dose of about 107 Gy,207 orabout 1000 times less than the electron dose to nuclebubbles; thus, it appears that electrons can indeedused to monitor bubbles formed byg-irradiation, thoughcarefully controlled confirmatory experiments must bperformed. Second, the uniform density of smaller bubles produced in theg-irradiated glasses is significantlydifferent than the highly localized larger bubbles produced in a focused electron beam. It is possible thatg-irradiation causes the nucleation of small bubbles moeffectively, perhaps because of the longer irradiatiotime (five orders of magnitude smaller dose rate). Thirin the limited studies to date, bubble formation undg-irradiation generally occurs at temperatures whebubbles are difficult to form by electron irradiationwhich is consistent with expected dose-rate behaviConsequently, in the absence of scientific understandand based on data in the literature, it seems very probathat bubble formation could occur in HLW glasses asresult of ionization damage fromb –g radiation. Theissue of whether or not oxygen bubbles form in HLWglasses under actual storage conditions must be resol

IP address: 144.82.107.169

2, No. 8, Aug 1997



of

olc

o

enr

orrer

idu

s

dle

urxd

e

s

ioeic

ea

ge

esfor

nge

-an-ndn-teult

e-be

esesd

theraldre

ingpicon-

ion.db-ter

ion-meuldionatedgesiseresd,.g.,

enteedlds.y

ndrayerS).

le

scientifically; otherwise, in any performance assessmeone must assume some finite probability for bubbformation and potentially large volume changes, perhaas high as the 50% volume change85,135,149 observedunder electron irradiation. A detailed investigationglasses irradiated withg-rays over a broad range otemperatures is needed to resolve this issue.

Alkali, particularly Na, mobility and depletion hasbeen linked to bubble formation. Due to the high mbility of alkali, any chemical potential gradient wilcause some alkali depletion. Consequently, the effeof thermal, electric field, and stress gradients on buble formation should be investigated as a functiontemperature in an ionizing radiation field.

The relevance of the electron- and ion-beam resuis more difficult to address. The ion-beam results acertainly relevant toa-particles and suggest that tha-particles will generate radiolytic oxygen, in additioto helium, which under the appropriate temperatuand dose-rate conditions can lead to bubble or vformation. Similar studies need to be performed ovetemperature range that encompasses the actual stotemperatures. The electron-beam results do providmeans to evaluate whether or not bubbles can founder the TEM conditions used to examineg-irradiatedsamples. Electron irradiations also can be used to rapdetermine any temperature dependence; however, sstudies should include a careful investigations of dorate effects.

Bubbles must be distinguished from voids and voithat contain only a few gas molecules. Gas molecu(He and O2) accumulating interstitially soon generatphysically unrealistic strains, and it is more likely thaoxygen molecules from radiolytic decomposition accmulate in voids or cages effected by network structurearrangements or alkali loss. In alkali halides, for eample, excess halogen molecules are accommodatecation-anion vacancy pairs, whose aggregation in tugenerates voids filled with halogen.208 The differencein specific volumes of vacancy pairs and moleculrenders Cl2 bubbles in NaCl or KCl under very highpressure (probably as highly strained fluid inclusionbut iodine in KI in a strain-free state.209 Inclusions ofdifferent charge density are most efficiently imagedTEM by Fresnel (out-of-focus) contrast independentmatrix crystallinity, and strained inclusions in crystallinmatrices by Bragg (diffraction) contrast. In aperiodglasses, the latter contrast mode is not availableimage strain fields and thus distinguish empty or patially filled cavities from high-pressure bubbles. Thuse of small-angle neutron and x-ray scattering cbe used to measure microstructural changes, suchvoids and bubbles, and with careful experimental desiit may be possible to distinguish gas-filled bubblfrom voids.



nt,leps

f

-

tsb-f

ltsre

eidaage

am

lych

e-

ses

t-

al-in

rn

s

),

nf

tor-

nas

n,s

H. Chemical durability

Radiation can affect the release rate of radionuclidfrom waste glasses by increasing the surface arearadionuclide release (microfracturing) and by changithe dissolution rate of the glass. The extent of thradiation-induced microfracturing will depend on differential volume changes, microstructure, and mechical properties. Presently, there are insufficient data aunderstanding available to predict the extent of radiatioinduced microfracturing in HLW glasses, except to nothat it can be expected to occur, especially as a resof radiation-induced differential expansion of phasseparated regions or any crystalline phases that maypresent in the glass.

The actual dissolution rate of nuclear waste glassmay be affected by the radiation-induced changin chemistry, microstructure, network bonding, analong radiation-damage tracks. As noted above,changes in leach rates due to radiation-induced structuchanges range from insignificant to significant, anthe effects of radiation on radionuclide release ahighly controversial. This will be resolved only bydetailed studies of the changes in structure and leachstudies that combine elemental analysis and isotoanalysis of nuclides in solution. It is clearly important tunderstand the relationship between the radiatioinduced structural changes and enhanced dissolutThe actual radionuclide release rate will depenon geochemical conditions (e.g., presence or asence of silica-saturated solutions) and glass-canisinteractions,6 which need to be studied. Similarly, theeffects of elevated storage temperatures on the radiatenhanced dissolution cannot be predicted at this tiand should be investigated. Higher temperatures copromote recovery processes. However, bubble formatand phase separation could be enhanced at elevtemperatures, and the effects of these structural chanon dissolution rates need to be understood. Itequally important to perform studies at the much highcumulative doses that will be experienced by glassfor Pu and other actinide disposition. In this regarthe release rate of alpha-decay daughter products (e235U) can be expected to be higher than that of the parnuclide (239Pu) due to the damage track formed by threcoiling daughter. Although this has been observin several glass and crystalline systems, it shoube confirmed for nuclear waste glass compositionUnderstanding this behavior would benefit greatly bdetailed characterization of the local structure arouthe actinides using techniques such as extended x-absorption fine structure (EXAFS), site-selective lasspectroscopies, and anomalous x-ray scattering (AXIn general, AXS will be more powerful than EXAFSfor measurements of this type; it provides reliab

IP address: 144.82.107.169

2, No. 8, Aug 1997 1969



r

a

ie

octule

i

)nn

sfoa

besir.

nncrotn

aa

la

ra

anxr-e-ntat

itut.

ceasrs.

n-c-tsnttureok

g.,

gle-e.,

e-s-

an

eralasi-eas-al

u-

essd

edisntre

deliatiooedis

quantitative information, and probes the structuassociated with second and higher neighbors thatalmost impossible to obtain on disordered materiwith EXAFS.

I. Theory and computer simulations

Theory and computer simulation or modeling caprovide an explanation or interpretation of macroscopprocesses in terms of the underlying atomistic mechnisms. The noncrystalline nature of glass, coupled wthe inherent chemical complexity of the waste glasssuggests that theory and simulation are essentialcomplement the range of experimental techniques navailable. Only through modeling can the complex proesses be analyzed, and only through theory canaccelerated test procedures be understood well enoto serve as the basis for the extrapolation of modebehavior over long periods of time.

Atomic-level simulations of radiation effects inmetals, intermetallics, and semiconductors is an actarea of research.210 Quantum mechanical and empiricamodels of atomic bonding, energy minimizationmolecular dynamics (MD), and Monte Carlo (MCmethods can be used in atomic-level calculatioof radiation damage processes in crystalline (aamorphous) materials.211 Energy minimization is limitedto zero Kelvin but is widely used to study structurestable defect configurations, and energy minimadefect motion. Static defect properties as welldynamic processes on the order of picoseconds canmodeled by MD simulations. Molecular dynamics canused to study defect formation and migration energidamage mechanisms, and defect production procein cascades. Monte Carlo techniques, which predbehavior from a random sampling of initial states, auseful for calculating damage distributions (e.g., Figsand 5), thermodynamic equilibrium structures anproperties, phase separation, and long-range diffusio

As noted, atomic-level simulations of radiatioeffects have focused largely on metals, intermetalliand semiconductors, and the consistent conclusions fthese studies are that a liquid-like core is generaduring an intense collision cascade. This subsequerelaxes back to the crystalline structure, but withlarger than equilibrium number of defects, suchvacancies and interstitials. These tend to aggregforming vacancy loops and clusters of interstitials. Instudy of glassy metals, Mattilaet al.212 concluded thatthere is melting around the core of the cascade, ascrystalline metals, but this melted region does not reback to the original structural state, producing insteadpermanent structural change in these glassy metals.

Computer simulations of radiation effects in amophous silica and complex silicate glasses are gener



eis

ls

nica-ths,tow-heghd

vel,

sd

,rsbees,sescte4d.

s,m

edtlyaste,a

inxa

-lly

more complicated because of the presence of more thone atom type, mixied ionic-covalent bonding, complestructures, and stoichiometry effects. Effective inteatomic potentials for amorphous silica have been dveloped and used in MD simulations of displacemecascades213 and to calculate the structure, density, hecapacity, and compressibility.214 Molecular dynamics isespecially useful in modeling glass structures, sincedoes not require the precise structure to be given as inpMD simulations of the structure of lead silicate215 andalkali silicate216 glasses have been shown to reproduexperimentally measured bulk structural features, suchinteratomic distances and average coordination numbeWith the development of appropriate interatomic potetials, it should be possible to model the atomic struture and simulate both diffusion and radiation effecin more complex silicate glasses. For multicomponesilicate (or phosphate) glasses, the amorphous strucraises several problems. First, it may be necessary to loat a number of different structural arrangements (e.presence of more than one type of network215). Secondly,the amorphous systems have networks that require andependent forces, so simple potentials will not sufficThis is less of a problem for the network modifierswhich do not form part of the network itself. While MDsimulations of complex glasses is in an early stage of dvelopment, it has been shown that a number of ion tranport properties, such as alkali ion migration energies, cbe calculated in multi-component silicate glasses.217,218

Such calculations of network modifier motion rationalizthe compositional dependence in terms of the structufeatures of the simulated structures. In addition, it hbeen shown that MD simulations of multicomponent silcate glasses can be used to calculate a number of murable structural properties, including density, thermexpansion coefficient, and viscosity.217,218

Delaye and Ghaleb219 have modeled a simplified(5-component) waste glass using pair potentials modlated by 3-body terms for certain components (SiO2,B2O3, Al2O3). Simulations of displacement cascadevents have been performed in this model waste glaand provide information about ion displacements anconsequent structural rearrangement of the glass.220–222

Oxygen and sodium were the most frequently displacatoms in the simulations of cascade events in thsimplified waste glass. This observation is consistewith simulations of cascades in amorphous silica, wheoxygen atoms are predominately displaced.213 Theseinitial computer simulation studies suggest that a cascain a simplified borosilicate waste glass drives alkaatoms to the edge of the cascade and increases the rof trigonally to tetrahedrally bonded boron, thus alscreating more nonbridging oxygens and diminishing thnetwork connectivity. The calculations also highlightethe difficulty of defining a displacement when there

IP address: 144.82.107.169

2, No. 8, Aug 1997



ac

hsau

atee

i

c

v

ghv

t

re

aa

aresesi-thether re-here

ty.l-efestnitblec-

.ekeyha.

to.-s,ofofreiesr

rk

iesnd-s,

nc-,ses

ofp-

alses.der-

ec-m-ngete,

no reference crystalline lattice. Electronic excitations anot currently taken into account in these simulations.

Changes to the glass structure as a result of cascevents can be quantified from simulations, and sustructural changes should be investigated by spectscopic and diffuse scattering techniques, as well aschanges in physical properties. This is an example of tvery desirable interaction of modeling and experimenMolecular dynamics can be used to estimate directly tobservable consequences of radiation damage in glasIn some cases, other calculations may be needed,spectroscopic changes would need electronic structcalculations as well. It is therefore important to extenexisting simulation work on structure and diffusion inoxide glasses to the area of radiation effects andexpand on the use of MD simulations to study cascadin complex glasses, as in the work of Delaye anGhaleb.221,222

An important issue concerns the complexity of thwaste form glass systems. Here one is concerned wtwo types of structural detail: the large number odifferent species and the nature of the charge statethe key components. In particular, the actinides coccur in several different charge states, as canFe and Ti, which are invariably found in the waststream. The charge state issue is just one of the arin which the behavior of the electron is crucial. Acomplete understanding of the mechanisms associawith radiation effects must also include an explicconsideration of the behavior of electrons and holeThis means that one must move beyond purely classitechniques. For example, network ions and defect specmay change their charge state during the damage presses. Electronic excitation can also lead to damagedifferent molecular species through chemical processinvolving excited electrons, such as in the formatioof peroxy radicals or superoxide ions discussed aboIt is possible to speculate that in the center of tha-recoil cascade excited state dynamics will influencboth defect formation and ion transport, while the eneris carried away by both photons and electrons. Telectron displacements may be greater than ion moments, and this kind of charge redistribution may leato the buildup of electric fields at both microscopic anmesoscopic length scales, causing additional separaof charges, either of electron-hole pairs, or of positivand negative ions. The effects of this on the structurearrangements in a cascade event need to be considbecause the motion of charges is heavily influencedelectric fields.

The importance of electronic processes in radiatioeffects is obvious, but the nature of these processestheir influence over structural changes is far less cleGiven that the spatial extent of the electronic process



re

deh

ro-byhet.ees.ndre

d

toesd

eithfofn

he

as

tedts.al

iesoc-or

esne.eeyee-ddionealred

by

nndr.

es

may be substantial (electron mean free paths of 1 nmnot long, but access many sites), and that the host glaswill contain a variety of aliovalent ions (such as transtion metals from the waste stream and, of course,actinides), it is necessary to understand the nature ofcharge states of these species, since electron transfeactions also underlie chemical durability properties. Tobjective of the theory here is to understand the natuof the charge states and their role in structural stabili

As with any theoretical approach, the issue of “vaidation” is critical, if the techniques are to be predictivand significantly more than just a rationalization oempirical observations. The use of structure as one tis attractive for a number of reasons. Firstly, givethat atomic positions are a natural product of MD,can be used to predict the macroscopically measuraproperties, such as the neutron total correlation funtion, T srd, which is available from diffraction studiesSimilarly, EXAFS spectra may be calculated from thsimulations. In fact, the simulations offer a natural linbetween all the experimental probes of structure. Thoffer a route to develop a picture of the structure whicis consistent with all the available experimental datReverse Monte Carlo methods that have been appliedoxide glasses223,224 can be used for waste form glasses

Another approach useful in “validation” is thermodynamic. Many of the quantities of interest are energieand one must have confidence in the predictionsenergies from simulations. These are a by-productMD simulations, although other methods may be moaccurate once the geometry is defined. The energmight be the relative stability of different phases, owhether or not dissolution is exothermic. Some woin this area has been done for crystals,225 but not yet forglasses. In addition, defect and ion migration energare of considerable interest. For example, Cormack aCao217 have shown that the mixed alkali effect in multicomponent glasses is reproduced by MD simulationand changes in sodium ion activation energy as a fution of Al content also mirror experimental trendssuggesting that the structural geometry for these glasis basically correct.

Advancements in the simulation and modelingradiation effects in complex glasses will require the aplication of different methodologies.Ab initio methodscan be used to develop improved interatomic potentiand to investigate the electronic state of small structurMD simulations with improved potentials can be useto model the radiation damage state and defect engetics; however, it will be important to include newmethodolgies, as they are developed, that allow eltronic excitiations and changes in charge state. An iportant goal, however, is the modeling and simulatioof the effects of time and temperature on the damastate, defect interactions, diffusion, relative energy sta

IP address: 144.82.107.169

2, No. 8, Aug 1997 1971



stio

y

s

ts

n

uie

nm

u

artun

r

tioin

l

rees

nsofi-eweshetrondleftenre-

,a-ofd,

ala-rals ofre,ldss

ant,

n

ss.atsooilse

os-

hcer-

c-ntd-is

and evolution of microstructure (e.g., bubbles and voidThis will require Monte Carlo methods and theorecal models to investigate the effects of multiple radiactive decay (damage) events and defect diffusion aannealing on damage accumulation and microstructuevolution on longer time scales and macroscopic stems. For example, mathematical rate theory, similarthat used to describe microstructural evolution in crytalline materials,226,227 could be applied to the formationof bubbles in complex glasses. In addition, methods neto be developed to incorporate defects detected specscopically, incorporate potential gradients (e.g., strethermal, electric field), define a displacement, calculaproperty changes (e.g., stored energy), and allow copaction or expansion of the structure. An important oucome should be the development of predictive modelsradiation effects in complex glasses that can be combiwith atomistic theories/models of glass dissolution.

V. RESEARCH FACILITIES

Research facilities needed in order to investigathe effects of radiation on glasses for HLW or Pdisposal range include actinide research laboratorg-irradiation facilities, and modern electron- and ionirradiation facilities. For actinide research, the radiatioeffects 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 that pvide unique opportunities to systematically investigaradiation effects in nuclear waste glasses. The techniqrange from probes of the local atomic environmeto bulk characterization and make use of major usfacilities with accelerator-based photon sources, neutscattering facilities, and high-energy or high-resolutioelectron microscopes. These techniques and facilineed to be made available for the study of radiatieffects in radioactive materials, including those containg actinides.

A. Actinide research facilities

Modern laboratories that perform research on atinide containing materials are very limited. Many laboratories with capabilities to handle radioactive materiaincluding actinides, have been shutdown on the DOsites as a result of the transformation from weapons pduction to environmental remediation. The capabilitito handle and study actinides are decreasing each yIt is therefore disturbing that with the 33 metric ton



).--ndrals-to-

edro-s,tem-t-ofed

te

s,-nd-

butes.

-

reo-eest

eronnesn-

c--s,Eo-sar.

of plutonium residue and scraps, 50 to 100 metric toof weapons grade plutonium, and 650 metric tonsplutonium in spent commercial nuclear fuel to be stablized, stored, and sent to disposal that there are so flaboratories equipped with modern analytical capabilitito handle radioactive materials. Fortunately, many of tadvanced characterization techniques (e.g., ESR, neuscattering, Raman spectroscopy, and NMR) can hanencapsulated materials, but these instruments are olocated in laboratories where radioactive materials ano longer allowed. There is, unfortunately, no central multiuser facility for handling radioactive materialsparticularly actinides, with all the necessary characteriztion facilities; consequently, detailed characterizationspecimens often requires shipping from site to site anin some cases, re-encapsulation for each technique.

B. Gamma irradiation facilities

Several 60Co gamma irradiation facilities exist atDOE sites; these include the Pacific Northwest NationLaboratory, Sandia National Laboratories, Argonne Ntional Laboratory, and the Savannah River Site. Seveof these sites have been used in instrumented studieglass corrosion. Studies of the effects of temperatuthermal gradients, stress gradients, and electron fieare feasible.In situ optical and electrical measurementare also feasible. In many of these60Co facilities thegamma field is nonisotropic. More isotropic gammfields are possible using spent fuel, particularly the specylindrical fuel at HFIR, which is hollow along thecylindrical axis. Understanding ionization damage iHLW glasses fromb-particles andg-rays would benefitimmensely from controlled temperature studies andinsitu measurements of glass specimens ing-facilities.

C. Ion-beam irradiation facilities

There are numerous ion-beam irradiation facilitieat DOE sites, universities, and industry laboratorieSome include dual and triple beam capabilities thallow simultaneous irradiation with several species,as to simulate both the alpha particle and the recnucleus that are produced during alpha decay. Themultibeam facilities also allowin situ ion-beam charac-terization (Sandia National Laboratories and Los AlamNational Laboratory228) or electron-beam characterization (IVEM and HVEM facilities at Argonne NationalLaboratory229). The ion-beam analysis techniques, whicare primarily used to measure changes in near-surfachemical compositions, include Rutherford backscatteing (RBS), elastic recoil detection (ERD), nuclear reation analysis (NRA), and particle-induced x-ray emissio(PIXE). Light elements that are usually difficult to detecwith RBS are easily measured with ERD methods. In adition, microbeam capabilities would allow the analys

IP address: 144.82.107.169

2, No. 8, Aug 1997



t

af

hokgss

hi

h

cddcsnoie

cs

nmhd

ts,be

,.d

ts,ngltsin

dnsu-ins

blep totesdes

ires)ri-s,n-

ofiesillte

eveder

te,yts

nd-c-

esmyal-sedcnsnet-

of phase separated regions. These ion-beam irradiafacilities also could be equipped within situ opticalcharacterization. A cantilever-beam apparatus89 is avail-able (Sandia National Laboratories) for measurementsvolume changes induced by ion-beam irradiation. Theare also a number of ion-beam irradiation facilitiesuniversities (e.g., University of Michigan, University oIllinois, 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 higenergy beta particle damage. Conventional electrmicroscopes that operate in the range of 100 to 400can also be used to simulate beta particle damaUnfortunately, these facilities have extremely high dorates as compared with the damage rates in glascontaining HLW or actinides, producing in microsecondthe equivalent damage of 1000 years storage. Otuseful facilities may be the pulsed electron-beam facities at Argonne National Laboratory and BrookhaveNational Laboratory.In situ characterization capabilitiesinclude time-resolved EPR and fluorescence, and otcapabilities could be added.

VI. RESEARCH NEEDS

There are many scientific issues to be resolvedorder to advance the understanding of radiation effein complex glasses and to provide for predictive moeling of nuclear waste glass behavior over long perioof time. In this section, recommendations for researpriorities are summarized from the previous discusions. The primary objective of these recommendatiois to develop fundamental knowledge and modelsradiation-glass interactions at the atomic, microscopand macroscopic levels in order to provide for thevaluation and performance assessment of glass waforms used for the immobilization and disposal of highlevel waste and weapons-grade Pu.

A. Reference glass compositions

Because so many different techniques and displines must contribute to progress in this field, it iimportant that radiation effects studies, as well as othstudies (e.g., dissolution and thermal stability), be peformed on well-characterized reference glasses. Itrecommended that a few specific glass compositio(simple to complex) be prepared, characterized for coposition and homogeneity, and made available to tscientific community. As a minimum, a three-tiereapproach to glass complexity should be followed: (i)simple system (an alkali silicate), (ii) a model system



ion

ofret

aty-nVe.ees

serl-n

er

ints-sh-sf

c,

ste-

i-

err-iss-e

a

of intermediate complexity (a four or five componenglass), and (iii) a simulated waste glass (DWPF glasPu glass, or others). Other glass systems that shouldstudied include: (i) Na-B-silicate, (ii) Na-B-Fe-silicate(iii) Na-B-(Ln,Ac)-silicate, and (iv) lead-iron phosphateThere should be a major attempt to coordinate ancommunicate the results of all studies (radiation effecdissolution, thermal stability) on these glasses amothose scientists active in the nuclear waste field. Resushould be published in the refereed literature and notthe gray literature of laboratory reports.

B. Structural and thermodynamic properties

Much of the data on structure, phase equilibria, anthermodynamic parameters for the glass compositiocurrently under consideration are missing, poorly docmented, or not readily accessible as they are foundinternal DOE or laboratory reports. Systematic studieof the structure and energetics of stable and metastaphases in these glass systems are an essential steunderstanding how the system approaches or deviafrom equilibrium under radiation. Such structural anthermodynamic studies must include both model glassand glasses containing actinides, which may requplacing appropriate instrumentation (e.g., calorimeterin radiation zones or gloveboxes. Structural charactezation should include determination of intrinsic defectshort and intermediate range order, and the local evironment of important radionuclides. Measurementsheat capacities, glass transition temperatures, enthalpof rapid relaxation processes, and total stored energy wbe required in order to determine the relative energy staof the irradiated glasses.

C. Systematic irradiation studies

Systematic studies of irradiation effects in referencglasses and simulated waste glasses, including archiglasses containing Pu or Cm, must be performed ovthe widest range of conditions (i.e., dose, dose raand temperature). This is required in order to identifthe fundamental processes underlying radiation effecin these glasses and to develop a scientific understaing of the balance between radiation damage prodution, microstructure evolution, and recovery processas functions of temperature and dose rate. Only frosuch understanding will it be possible to predict with andegree of certainty the effects of radiation under acturepository or disposition conditions. Simple engineeringtype tests at expected temperatures, as might be propofor some acceptance criteria, provide limited scientifiunderstanding and unreliable performance predictiowhen the fundamental damage processes, damage kiics, and dose-rate effects are unknown.

IP address: 144.82.107.169

2, No. 8, Aug 1997 1973



ttte

nd

c

t

m

c

tu

e

i

i

a

inextheies

ofons,n).edg.,nt,ndtondsillls

ic

dus

rme)

esants-atnslet

sbleralndthd

noff

el

sy

The systematic studies must include measuremeof the energetics of the radiation damage processesthe relative energy state of the damaged glasses bto understand the effects of radiation on properties areactivity and to provide a means of validation focomputer simulations and model predictions. These stematic studies should focus on the following: (1) undestanding the relative effects of ionization and balliscollision processes on defect generation and radiaeffects; (2) resolving the controversy regarding radioly(oxygen) bubble formation in HLW glasses; (3) dveloping an understanding of radiation-induced buble formation, radiation-induced phase transformatioradiation-enhanced diffusion, and the role of thermstress, and electric field gradients on these processefunctions of dose rate and temperature; (4) understandhelium accumulation, trapping, bubble formation, arelease in glass compositions for Pu (or other actinidisposition; (5) since accelerated irradiation techniqumust be employed in these studies of radiation effeit is imperative that a thorough understanding of dosrate effects be developed and that high-dose-rate teniques (e.g., ion and electron irradiations) be validathrough comparisons with data on actinide-containinggamma-irradiated glasses and with theoretical modand computer simulations.

D. Theory and computer simulations

Theoretical modeling and computer simulation tecniques are needed to provide: (i) atomistic interpretatioof experimental results; (ii) relationships between coposition, atomic structure, and volume changes; (iii)understanding of ionization and ballistic-collision proesses; (iv) atomistic details of the primary damage sta(v) defect structures and energetics; (vi) an understaing of dose-rate effects and the limitations of accelerairradiation techniques; (vii) models of defect/gas diffsion and interactions; and (viii) models of microstructuevolution (e.g., bubble formation). In addition, computsimulations should be used to ascertain and interpthe short-range and intermediate-range order in glas(e.g., local bond coordination and ring-size distributioncalculate spectral features (e.g., x-ray and laser sptroscopies), and predict observable effects of radiatdamage in glasses for experimental confirmation. Avancements in this area will require the use of differemethodologies:ab initio methods to develop improvedinteratomic potentials and investigate the electronic stof small structures; molecular dynamic simulations wimproved potentials to model the radiation damage staand Monte Carlo methods and theoretical modelsinvestigate the effects of multiple radioactive dec(damage) events and defect diffusion on damage acmulation and microstructure evolution on longer timscales and macroscopic systems.



ntsandothndrys-r-icionic-b-n,

al,s asingde)ests,e-ch-edorels

h-ns-

an-te;

nd-ed-

rerretses),ec-ond-nt

atethte;toycu-e

Electronic excitation processes are importantthe development of radiation effects in these complglasses. Classical methodologies do not account forelectronic processes; consequently, new methodologneed to be developed that allow the transferenergy to electronic excitations, the ability of ions tchange charge states, and the formation of excitoperoxy radicals, and molecular species (e.g., oxygeIn addition, methods to incorporate defects detectspectroscopically, incorporate potential gradients (e.stress, thermal, electric field), define a displacemecalculate property changes (e.g., stored energy), aallow compaction or expansion of the structure needbe developed. In all cases, validation of the models asimulation methods is critical. Finally, predictive modelof long-term nuclear waste glass performance wrequire combining atomistic and macroscopic modeof radiation effects in complex glasses with atomisttheories and models of glass dissolution.

E. Facilities

Existing facilities for radiological work need to bemaintained and a central multiuser facility to handle ancharacterize radioactive materials is needed. The varioirradiation facilities (electrons, ions, x-rays, andg-rays)currently available are adequate, but resources to perfoin situ and controlled environment (e.g., temperaturstudies need to be available.

F. Attract talented scientists

The panel expressed concern that this field donot attract graduate students, and much of the relevknowledge and experiences will disappear with the pasage of time. With the complex needs for remediationthe DOE sites, and the need for disposition of weapoPu, there is a need for a relatively small but stabpool of expertise and facilities. It is imperative to attracand involve bright new minds with new ideas to thifield. Increased opportunities need to be made availato support graduate students and provide post doctoexperience at universities, the national laboratories, aother DOE sites. A long-term research program, wia continuity of purpose, is essential to attracting anretaining high caliber scientists.

ACKNOWLEDGMENTS

This panel was sponsored by the Council oMaterials Science of the United States DepartmentEnergy, Office of Basic Energy Sciences, Division oMaterials Sciences (DMSyBES) under Contract DE-FG02-96ER45439 at the University of Illinois. The Panwould like to thank Dr. Iran Thomas, DMSyBES Direc-tor, and Dr. Robert Gottschall, Metallurgy and CeramicBranch Chief, for their support. They also gratefull

IP address: 144.82.107.169

12, No. 8, Aug 1997



fe,m

n

cen

edr

n

u

m,

l

i,

rky

ra

ry

.

e

in

.

in

.

gen

,

nd

nd-

.

s

in

i,

t,

n

a,

acknowledge the support and encouragement of Prosor C. Peter Flynn (University of Illinois–Urbana)Chairman of the Council on Materials Science, whoprovided liaison with DMSyBES and actively partici-pated in the Panel Meeting in Santa Fe. The Panel thaProfessor David Clarke (University of California–Santa Barbara) and Dr. Michael Knotek (PacifiNorthwest National Laboratory), members of thCouncil on Materials Science, and Dr. Yok Che(DOE, DMSyBES) for their active participation in thePanel Meeting. The Panel also gratefully acknowledgDr. John Bates (Argonne National Laboratory) anDr. Ned Bibler (Savannah River Technology Centefor the background tutorials presented to the panel.

REFERENCES

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

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

3. Management and Disposition of Excess Weapons PlutoniuReactor-Related Options(National Academy Press, WashingtonDC, 1995).

4. G. G. Wicks, A. R. Lodding, and M. A. Molecke, MRS Bull.XVIII (9), 33 (1993).

5. C. M. Jantzen, J. Am. Ceram. Soc.75 (9), 2433 (1992).6. B. Grambow, MRS Bull.XIX (12), 20 (1994).7. Storage and Disposition of Weapons-Usable Fissile Materia

Draft Programmatic Environmental Impact Statement,DOEyEIS-0229-D (U.S. Department of Energy, Washington, DC, 1996).

8. Defense Waste Processing Facility - Cost, Schedule, and TechnIssues,GAOyRCED 92-183 (U.S. General Accounting OfficeWashington, DC, 1992).

9. W. Lutze, presented at the National Research Council Woshop onGlass as a Waste Form and Vitrification TechnologWashington, DC, May 13–15, 1996 (unpublished).

10. Cunnane and J. M. Allison, inScientific Basis for Nuclear WasteManagement XVII,edited by A. Barkatt and R. A. Van Konynen-burg (Mater. Res. Soc. Symp. Proc.333, Pittsburgh, PA, 1994),p. 3.

11. Waste Form Compliance Plan for the West Valley Demonsttion Project High-Level Waste Form,WVDP-185 (West ValleyDemonstration Project, West Valley, NY, 1995).

12. D. B. Amerine, presented at the National Research Council Woshop onGlass as a Waste Form and Vitrification TechnologWashington, DC, May 13–15, 1996 (unpublished).

13. D. L. Illman, Chemical & Engineering News71 (25), 9 (1993).14. Office of Technology Assessment,Complex Cleanup: The En-

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

15. Disposal of Weapon Plutonium,edited by E. R. Merz and C. E.Walter (Kluwer Academic Publishers, The Netherlands, 1996)

16. A. Makhijani and A. Makhijani,Fissile Materials in a Glass,Darkly (IEER Press, Takoma Park, MD, 1995).

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

18. G. Taubes, Science268, 1836 (1995).19. C. D. Bowman and F. Venneri, Sci. Glob. Secur.5, 279 (1996).20. R. A. Van Konynenburg, Sci. Glob. Secur.5, 303 (1996).



s-

ks

s

)

,

m

–

s

cal

-,

-

k-,

y

21. N. E. Bibler, W. G. Ramsey, T. F. Meaker, and J. M. Pareizs,Scientific Basis for Nuclear Waste Management XIX,edited byW. M. Murphy and D. A. Knecht (Mater. Res. Soc. Symp. Proc412, Pittsburgh, PA, 1996), p. 65.

22. J. K. Bates, A. J. G. Ellison, J. W. Emery, and J. C. Hoh,Scientific Basis for Nuclear Waste Management XIX,edited byW. M. Murphy and D. A. Knecht (Mater. Res. Soc. Symp. Proc412, Pittsburgh, PA, 1996), p. 57.

23. Hj. Matzke, Special issue for a workshop on Radiation DamaEffects in Nuclear Waste Materials, Nucl. Instrum. Methods iPhys. Res. B32, 453 (1988).

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

25. W. J. Weber and F. P. Roberts, Nucl. Technol.60, 178 (1983).26. W. G. Burns, A. E. Hughes, J. A. C. Marples, R. S. Nelson, a

A. M. Stoneham, J. Nucl. Mater.107, 245 (1982).27. W. G. Burns, A. E. Hughes, J. A. C. Marples, R. S. Nelson, a

A. M. Stoneham, inScientific Basis for Nuclear Waste Management V,edited by W. Lutze (Mater. Res. Soc. Symp. Proc.11,Pittsburgh, PA, 1982), p. 339.

28. D. H. Day, A. E. Hughes, J. W. Leake, J. A. C. Marples, G. PMarsh, J. Rae, and B. O. Wade, Rep. Prog. Phys.48, 101 (1985).

29. W. J. Weber, Nucl. Instrum. Methods in Phys. Res. B32, 471(1988).

30. W. J. Weber, J. Minerals, Metals and Materials Sociery43 (7),35 (1991).

31. D. J. Wronkiewicz,Effects of Radionuclide Decay on Waste GlasBehavior –A Critical Review,ANL-93y45 (Argonne National Lab-oratory, Argonne, IL, 1993).

32. R. C. Ewing, W. J. Weber, and F. W. Clinard, Jr., ProgressNuclear Energy29 (2), 63 (1995).

33. W. J. Weber and R. C. Ewing, inScientific Basis for NuclearWaste Management XVIII,edited by T. Murakami and R. C. Ewing(Mater. Res. Soc. Symp. Proc.353, Pittsburgh, PA, 1995), p. 1389.

34. 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.

35. Radioactive Waste Forms for the Future,edited by W. Lutze andR. C. Ewing (North-Holland, Amsterdam, 1988).

36. R. C. Ewing, J. Nucl. Mater.190, vii (1992).37. T. A. Buscheck, J. J. Nitao, and L. D. Ramspott, inScientific Basis

for Nuclear Waste Management XIX,edited by W. M. Murphy andD. A. Knecht (Mater. Res. Soc. Symp. Proc.412, Pittsburgh, PA,1996), p. 715.

38. J. F. Ziegler, J. P. Biersack, and U. Littmark,The Stopping andRange of Ions in Solids(Pergamon, New York, 1985).

39. W. J. Gray, Nature (London)296, 547 (1982).40. R. P. Turcotte, Radiation Effects in Solidified High-Level

Wastes –Part 2, Helium Behavior,BNWL-2051 (Pacific North-west National Laboratory, Richland, WA, 1976).

41. S. Sato, H. Furuya, K. Morikawa, M. Sugisaki, and Y. InagakJ. Nucl. Sci. and Technol.27, 343 (1990).

42. B. J. Todd, J. L. Lineweaver, and J. T. Kerr, J. Appl. Phys.31,51 (1960).

43. J. L. Lineweaver, J. Appl. Phys.34, 1786 (1963).44. N. E. Bibler, inScientific Basis for Nuclear Waste Managemen

edited by S. V. Topp (Mater. Res. Soc. Symp. Proc.6, North-Holland, New York, 1982), p. 681.

45. Standard Method for Testing the Long Term Alpha IrradiatioStability of Solidified High-Level Radioactive Waste Forms,ISO6962 (International Organization for Standardization, GenevSwitzerland, 1982).

46. Nuclear Waste Materials Handbook, Test Methods,DOEyTIC-11400 (Pacific Northwest Laboratory, Richland, WA, 1981).

IP address: 144.82.107.169

2, No. 8, Aug 1997 1975



c

n,

,

,g

d

n

r,

.

.

s

,

k

e

ki,

cl.

ntr.

k,

o,

o.

nd

,

5

),

rty,

s

),

ks,

r.,

47. W. J. Weber, J. W. Wald, and G. L. McVay, J. Am. Ceram. So68 (9), C-253 (1985).

48. G. Wagner and P. van den Haute,Fission-Track Dating(KluwerAcademic Publishers, Dordrecht, The Netherlands, 1992).

49. Y. Eyal and R. C. Ewing, inProceedings of the InternationalConference on Nuclear Waste Management and EnvironmeRemediation, Vol. 1,edited by D. Alexandre, R. Baker, R. Kohoutand J. Marek (ASME Press, New York, 1993), p. 191.

50. R. L. Fleischer, Science,207, 979 (1980).51. G. W. Arnold, C. J. M. Northrup, and N. E. Bibler, inScientific

Basis for Nuclear Waste Management-V,edited by W. Lutze(Mater. Res. Soc. Symp. Proc.11, Pittsburgh, PA, 1982), p. 357.

52. G. Malow and H. Andresen, inScientific Basis for Nuclear WasteManagement, Vol. 1,edited by G. J. McCarthy (Plenum PressNew York, 1979), p.109.

53. M. Antonini, F. Lanza, and A. Manara, inCeramics in NuclearWaste Management,edited by T. D. Chikalla and J. E. MendelCONF-790420 (National Technical Information Service, Sprinfield, VA, 1979), p. 289.

54. L. F. Curtiss, Nature (London)120, 406 (1927).55. S. C. Lind, Science68, 643 (1928).56. D. G. Tuck, Int. J. Appl. Radiat. Isotopes15, 49 (1964).57. E. Lell, N. J. Kreidl, and J. R. Hensler, inProgress in Ceramic

Science, Vol. 4,edited by J. Burke (Pergamon Press, OxforNew York, 1966), p. 1.

58. D. L. Griscom, inGlass: Science and Technology Vol. 4B,editedby D. R. Uhlmann and N. J. Kreidl (Academic Press, Bosto1990), p. 151.

59. F. W. Clinard, Jr. and L. W. Hobbs, inPhysics of Radiation Effectsin Crystals,edited by R. A. Johnson and A. N. Orlov (ElsevieAmsterdam, 1986), p. 387.

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

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

62. L. W. Hobbs, J. Non-Cryst. Solids192 & 193, 79 (1995).63. A. C. Wright, B. Bachra, T. M. Brunier, R. N. Sinclair, L. F

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

64. L. W. Hobbs, Nucl. Instrum. Methods in Phys. Res. B91, 30–42(1994).

65. C. S. Marians and L. W. Hobbs, J. Non-Cryst. Solids119, 269(1990).

66. L. F. Gladden, J. Non-Cryst. Solids119, 318 (1990).67. P. Vashishta, A. Nakano, R. K. Kalia, and I. Ebbsjo, J. Non-Cry

Solids 182, 59 (1995).68. A. G. Revesz and G. V. Gibbs, inThe Physics of MOS Insulators

edited by G. Lucovsky, S. T. Pantelides, and F. L. Galeen(Pergamon, New York, 1980), p. 92.

69. C. S. Marians and J. K. Burdett, J. Non-Cryst. Solids124, 1(1990).

70. E. J. Friebele, D. L. Griscom, M. Stapelbroek, and R. A. WeePhys. Rev. Lett.42, 1346 (1979).

71. L. W. Hobbs and M. R. Pascucci, J. Phys.41 (C6), 237 (1980).72. F. Wooten and D. Weaire, J. Non-Cryst. Solids64, 325 (1984).73. U. Jain, A. C. Powell, and L. W. Hobbs, inDefects in Materi-

als, edited by P. D. Bristowe, J. E. Epperson, J. E. Griffith, anZ. Lillental-Weber (Mater. Res. Soc. Symp. Proc.209, Pittsburgh,PA, 1991), p. 201.

74. D. J. Jacobs and M. F. Thorpe, Phys. Rev. E52, 3682 (1996).75. P. K. Gupta, J. Am. Ceram. Soc.76, 1088 (1993).76. G. N. Greaves, J. Non-Cryst. Solids71, 203 (1985).77. B. C. Sales, J. O. Ramey, J. C. McCallum, and L. A. Boatn

J. Non-Cryst. Solids126, 179 (1990).78. W. Primak, L. H. Fuchs, and P. Day, Phys. Rev.92, 1064 (1953).



.

tal

-

,

,

t.

er

s,

d

r,

79. W. Primak, Phys. Rev.110, 1240 (1958).80. M. Whittels and F. A. Sherrill, Phys. Rev.93, 1117 (1954).81. E. P. EerNisse, J. Appl. Phys.45, 167 (1974).82. H. M. Presby and W. L. Brown, Appl. Phys. Lett.24, 511 (1974).83. G. W. Arnold, Radiat. Eff.65, 17 (1982).84. S. Sato, H. Furuya, Y. Inagaki, T. Kozaka, and M. Sugisa

J. Nucl. Sci. Technol.24, 920 (1987).85. S. Sato, H. Furuya, K. Asakura, K. Ohta, and T. Tamai, Nu

Instrum. Methods in Phys. Res. B1, 534 (1984).86. J. E. Shelby, J. Appl. Phys.51, 2561 (1980).87. W. Primak and R. Kampwirth, J. Appl. Phys.39, 6010 (1968).88. C. B. Norris and E. P. EerNisse, J. Appl. Phys.45, 3876 (1974).89. G. W. Arnold, inScientific Basis for Nuclear Waste Manageme

VIII, edited by C. M. Jantzen, J. A. Stone, and R. C. Ewing (MateRes. Soc. Symp. Proc.44, Pittsburgh, PA, 1985), p. 617.

90. A. Hiraiwa, H. Usui, and K. Yagi, Appl. Phys. Lett.54, 1106(1989).

91. M. Antonini, A. Manara, and S. Buckley, Radiat. Eff.65, 55(1982).

92. M. Antonini, A. Manara, and P. Lensi, inPhysics of SiO2 andits Interfaces,edited by S. T. Pantiledes (Pergamon, New Yor1978), p. 316.

93. R. A. B. Devine and M. Marchand, Appl. Phys. Lett.63, 619(1993).

94. K. Fukumi, A. Chayahara, M. Satou, M. Hayakawa, M. Hangyand S. Nakashima, Jpn. J. Appl. Phys.29, 905 (1990).

95. A. C. Wright, J. Non-Cryst. Solids179, 84 (1994).96. R. A. Weeks, J. Appl. Phys.27, 1376 (1956).97. R. A. Weeks, Phys. Rev.130, 570 (1963).98. D. L. Griscom, Nucl. Instrum. Methods in Phys. Res. B1, 481

(1984).99. F. J. Feigl, W. B. Fowler, and K. L. Yip, Solid State Commun.14,

225 (1974).100. K. W. Yip and W. B. Fowler, Phys. Rev. B11, 2327 (1975).101. K. C. Snyder and W. B. Fowler, Phys. Rev. B48, 13 238 (1993).102. R. A. Weeks, Proc. VII Intl. Cong. on Glass, Brussels, Paper N

41 (1965).103. R. Tohmon, H. Mizuno, Y. Ohki, K. Sasagane, K. Nagasawa, a

Y. Hama, Phys. Rev. B39, 1337 (1989).104. H. Imai, K. Arai, H. Hosono, Y. Abe, T. Arai, and H. Imagawa

Phys. Rev. B44, 4812 (1991).105. D. L. Griscom, J. Ceram. Soc. Jpn.99, 923 (1991).106. A. V. Amossov and A. O. Rybaltovsky, J. Non-Cryst. Solids 7

(1994).107. D. L. Griscom, inProc. Thirty-Third Frequency Control Sym-

posium (Electronics Industries Assn., Washington, DC, 1979p. 109.

108. W. Hayes, M. J. Kane, O. Salimen, R. L. Wood, and S. P. DoheJ. Phys. C: Solid State17, 2943 (1984).

109. T. E. Tsai and D. L. Griscom, Phys. Rev. Lett.67, 2517(1991).

110. N. Itoh, T. Shimizu-Iyayama, and T. Fujita, J. Non-Cryst. Solid179, 194 (1994).

111. A. H. Edwards and W. B. Fowler, Phys. Rev. B26, 6649 (1982).112. R. L. Pfeffer, inThe Physics and Technology of Amorphous SiO2,

edited by R. A. B. Devine (Plenum Publishing, New York, 1988p. 181.

113. R. L. Pfeffer, inThe Physics and Chemistry of SiO2 and the Si-SiO2 Interface, edited by C. R. Helms and B. E. Deal (PlenumPress, New York, 1988), p. 169.

114. E. J. Friebele, D. L. Griscom, M. Stapelbroek, and R. A. WeePhys. Rev. Lett.42, 1346 (1979).

115. M. Stapelbroek, D. L. Griscom, E. J. Friebele, and G. H. Sigel, JJ. Non-Cryst. Solids32, 313 (1979).

116. D. L. Griscom, J. Non-Cryst. Solids68, 301 (1984).

IP address: 144.82.107.169

2, No. 8, Aug 1997



y

o

y

c

i

incle,

i41.r,

nd

al

an-

ing,

m.

9).

2

,

.

ic

G.

al

,

nt,

117. D. L. Griscom, G. H. Sigel, Jr., and R. J. Ginther, J. Appl. Ph47, 960 (1976).

118. E. J. Friebele and D. L. Griscom, inDefects in Glasses,edited byF. L. Galeener, D. L. Griscom, and M. J. Weber (Mater. Res. SSymp. Proc.61, Pittsburgh, PA, 1986), p. 319.

119. D. L. Griscom, E. J. Friebele, K. J. Long, and J. W. FleminJ. Appl. Phys.54, 3743 (1983).

120. G. W. Arnold, Radiat. Eff.98, 55 (1986).121. G. W. Arnold, Nucl. Instrum. Methods in Phys. Res. B1, 516

(1984).122. G. Battaglin, G. Della Mea, G. De Marchi, P. Mazzoldi, an

O. Puglisi, J. Non-Cryst. Solids50, 119 (1982).123. G. Battaglin, G. Della Mea, G. De Marchi, P. Mazzoldi, an

O. Puglisi, Radiat. Eff.64, 99 (1982).124. J. F. DeNatale, D. G. Howitt, and G. W. Arnold, Radiat. Eff.98,

63 (1986).125. J. W. H. Schreurs, J. Chem. Phys.47, 818 (1967).126. D. L. Griscom, J. Non-Cryst. Solids64, 229 (1984).127. R. Cases and D. L. Griscom, Nucl. Instrum. Methods in Ph

Res. B 1, 503 (1984).128. R. Cases and D. L. Griscom, inProc. XVI Int’l Cong. Glass, Vol. 3,

Glass Structure(1992), p. 53.129. D. L. Griscom, J. Non-Cryst. Solids6, 275 (1971).130. P. W. Levy, J. M. Loman, and J. A. Kierstead, Nucl. Instrum

Methods in Phys. Res. B1, 549 (1984).131. N-G. Vannerberg, inProgress in Inorganic Chemistry 4(1962),

p. 125.132. J. F. DeNatale and D. G. Howitt, Am. Ceram. Soc. Bull.61, 582

(1982).133. J. F. DeNatale and D. G. Howitt, Nucl. Instrum. Methods in Ph

Res. B 1, 489 (1984).134. J. P. Heuer, H. W. Chan, D. G. Howitt, and J. F. DeNatale,

Nuclear Waste Management II,edited by D. E. Clark, W. B. White,and J. J. Machiels (Advances in Ceramics20, The AmericanCeramic Society, Westerville, OH, 1986), p. 175.

135. A. Manara, M. Antonini, P. Camagni, and P. N. Gibson, NuInstrum. Methods in Phys. Res. B1, 475 (1984).

136. P. B. Adams and D. L. Evans, Mater. Res. Sci.12, 525 (1978).137. E. Snoeks, T. Weber, A. Cacciato, and A. Polman, unpublish138. E. Snoeks, Ph.D. Thesis, Utrecht University (1995).139. G. W. Arnold, Nucl. Instrum. Methods in Phys. Res. B32, 504

(1988).140. S. Klaumünzer, Radiat. Eff.110, 78 (1989).141. A. Benyagoub, S. Löffler, M. Rammansee, and S. Klaumünzer,

Nucl. Instrum. Methods in Phys. Res. B65, 228 (1992).142. D. L. Griscom, inExperimental Techniques of Glass Scienc

edited by O. El-Bayoumi and C. J. Simmons (The AmericaCeramic Society, Westerville, OH, 1993), p. 161.

143. D. L. Griscom, J. Chem. Phys.55, 1113 (1971).144. J. A. C. Marples, Nucl. Instrum. Methods in Phys. Res. B32, 480

(1988).145. Y. Inagaki, H. Furuya, K. Idemitsu, T. Banba, S. Matsumoto, a

S. Muraoka, inScientific Basis for Nuclear Waste ManagemeXV, edited by C. G. Sombret (Mater. Res. Soc. Symp. Proc.257,Pittsburgh, PA, 1992), p. 199.

146. T. Banba, S. Matsumoto, S. Muraoka, K. Yamada, M. SaH. Ishikawa, and N. Sasaki, inScientific Basis for NuclearWaste Management XVIII,edited by T. Murakami and R. C.Ewing (Mater. Res. Soc. Symp. Proc.353, Pittsburgh, PA, 1995),p. 1397.

147. M. Antonini, P. Camagni, F. Lanza, and A. Manara, inScien-tific Basis for Nuclear Waste Management-2,edited by C. J. M.Northrup, Jr. (Plenum Press, New York, 1980), p. 127.

148. S. Sato, H. Furuya, T. Kozaka, Y. Inagaki, and T. Tamai, J. NuMater. 152, 265 (1988).


J. Mater. Res., Vol.

s.

c.

g,

d

d

ys.

.

s.

in

l.

ed.

e,n

ndnt

to,

cl.

149. A. Manara, P. N. Gibson, and M. Antonini, inScientific Basis forNuclear Waste Management-V,edited by W. Lutze (Mater. Res.Soc. Symp. Proc.11, North-Holland, New York, 1982), p. 349.

150. A. M. Stoneham, Nucl. Instrum. Methods in Phys. Res. B91, 1(1994).

151. F. P. Roberts, G. H. Jenks, and C. D. Bopp,Radiation Effects inSolidified High-Level Wastes –Part 1, Stored Energy,BNWL-1944(Pacific Northwest National Laboratory, Richland, WA, 1976).

152. A. R. Hall, J. T. Dalton, B. Hudson, and J. A. C. Marples,Management of Radioactive Wastes from the Nuclear Fuel CyVol. II, IAEA-SM-207 (International Atomic Energy Agency,Vienna, 1976), p. 3.

153. G. Malow, J. A. C. Marples, and C. Sombret, inRadioactive WasteManagement and Disposal,edited by R. Simon and S. Orlowsk(Harwood Academic Publishers, Chur, Switzerland, 1980), p. 3

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

155. J. F. DeNatale and D. G. Howitt, Radiat. Eff.91, 89 (1985).156. W. J. Weber, R. P. Turcotte, L. R. Bunnell, F. P. Roberts, a

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.

157. S. Sato, K. Asakura, and H. Furuya, Nucl. Chem. Waste Magement4, 147 (1983).

158. J. F. DeNatale and D. G. Howitt, Am. Ceram. Soc. Bull.66, 1393(1987).

159. J. P. Heuer, M. S. Thesis, Department of Mechanical EngineerUniversity of California–Davis 1987.

160. D. G. Howitt, H. W. Chan, J. F. DeNatale, and J. P. Heuer, J. ACeram. Soc.74, 1145 (1991).

161. M. H. Tosten,TEM examination of irradiated waste glass,WSRC-RP 89-584 (Westinghouse Savannah River Co., Aiken, SC, 198

162. N. E. Bibler, M. H. Tosten, and D. C. Beam, inHigh-Level Radio-active Waste Management, Vol. 2(American Nuclear Society,La Grange Park, IL, 1990), p. 1103.

163. J. Y. Laval and K. H. Westmacott, inElectron Microscopy andAnalysis, 1979,edited by T. Mulvey, Conference Series, No. 5(Institute of Physics, London, 1980), p. 295.

164. Y. Inagaki, H. Furuya, Y. Ono, K. Idemitsu, T. BanbaS. Matsumoto, and S. Muraoka, inScientific Basis for NuclearWaste Management XVI,edited by C. G. Interrante and R. TPabalan (Mater. Res. Soc. Symp. Proc.294, Pittsburgh, PA,1993), p. 191.

165. Hj. Matzke and E. Toscano, Europ. Appl. Res. Rep.7 (9), 1403(1990).

166. Hj. Matzke and E. Vernaz, J. Nucl. Mater.201, 295 (1993).167. Hj. Matzke and J. van Geel, inDisposal of Weapons Pluto-

nium, edited by E. R. Merz and C. E. Walter (Kluwer AcademPublishers, The Netherlands, 1996), p. 93.

168. R. A. Bonniaud, N. R. Jacquet Francillon, R. L. Laude, and C.Sombret, in Ceramics in Nuclear Waste Management,editedby T. D. Chikalla and J. E. Mendel, CONF-790420 (NationTechnical Information Service, Springfield, VA, 1979), p. 57.

169. J. L. Routbort and Hj. Matzke, Mater. Sci. and Eng.58, 229(1983).

170. W. J. Weber and Hj. Matzke, Europ. Appl. Res. Rep.7 (7), 1221(1987).

171. M. Yamashita and Hj. Matzke, inModifications Induced byIrradiation in Glasses,edited by P. Mazzoldi (Elsevier ScienceAmsterdam, 1992), p. 53.

172. I. L. F. Ray and Hj. Matzke, unpublished results.173. R. P. Turcotte, Radioactive Waste Management2, 169 (1981).174. J. C. Dran, Y. Langevin, M. Maurette, J. C. Petit, and B. Vasse

in Scientific Basis for Nuclear Waste Management,edited by S. V.

IP address: 144.82.107.169

12, No. 8, Aug 1997 1977



o

c

c

d

,

c.,

te

g,

er.,

olspy,

h

.

.

lids

ds

cts.

d

r,ys.

.

Topp (Mater. Res. Soc. Symp. Proc.6, North-Holland, New York,1982), p. 651.

175. W. Primak, Nucl. Sci. Engr.80, 689 (1982).176. D. R. Cousens and S. Myhra, J. Non-Cryst. Solids54, 345 (1983).177. J. R. Grover and B. E. Chidley, Reactor Science and Technol

(J. Nucl. Energy Parts AyB) 16, 405 (1962).178. J.R. Grover, in Management of Radioactive Wastes from F

Reprocessing (Organisation for Economic Co-Operation and Dvelopment, Paris, 1973), p. 593.

179. J. H. Westsik and C. O. Harvey,High-Temperature Leaching of aSimulated High-Level Waste Glass,PNL-3172 (Pacific NorthwestLaboratory, Richland, WA, 1981).

180. Y. Eyal and R. L. Fleischer, Geochimica et Cosmochimica A49, 1155 (1985).

181. D. R. Olander and Y. Eyal, Geochimica et Cosmochimca A54, 1889 (1990).

182. A. J. G. Ellison and A. Navrotsky, J. Am. Ceram. Soc.75, 1430(1992).

183. A. J. G. Ellison and P. C. Hess, J. Non-Cryst. Solids127, 247(1991).

184. R. Calemezuk and E. Bonjour, J. Non-Cryst. Solids43, 427(1981).

185. A. Navrotsky, Transaction ACA27, 1 (1991).186. S. Susman, K. J. Volin, R. C. Liebermann, G. Gwanmesia, a

Y. Wang, Phys. Chem. Glasses31, 144 (1990).187. R. J. Hemley, K. K. Mao, P. M. Bell, and B. O. Mysen, Phys. Re

Lett. 57, 747 (1986).188. R. H. Stolen, J. T. Krause, and C. R. Kurkijan, Discuss. Fara

Soc. 50, 103 (1970).189. D. L. Griscom, SPIE541, 38 (1985).190. G. W. Arnold and W. D. Compton, Phys. Rev.116, 802 (1959).191. R. L. Pfeffer, J. Appl. Phys.57, 5176 (1985).192. G. W. Arnold and P. Mazzoldi, inIon Beam Modification of

Insulators, edited by P. Mazzoldi and G. W. Arnold (ElsevierAmsterdam, 1987), Chap. 5.

193. P. Mazzoldi and A. Miotello, Radiat. Eff.98, 205 (1986).194. A. Miotello and P. Mazzoldi, J. Phys. C: Solid State Phys.15,

5615 (1982).195. A. Miotello, Phys. Lett.103A, 279 (1984).196. W. J. Weber, R. C. Ewing, and L. M. Wang, J. Mater. Res.9, 688

(1994).197. P. F. James, J. Mater. Sci.10, 1802 (1975).198. Phase Separation in Glass,edited by O. V. Mazurin and E. A.

PoraiKoshits (North-Holland, Amsterdam, 1984).199. P. Taylor, Atomic Energy of Canada Limited Report, AECL

10173, August, 1980.200. W. Haller, D. H. Blackburn, F. E. Wagstaff, and R. J. Charle

J. Am. Ceram. Soc.53, 34 (1970).201. R. F. Fudali, M. D. Dayar, D. L. Griscom, and H. D. Schreibe

Geochim. Cosmochim. Acta51, 2749 (1987).202. Y-M. Chiang and W. D. Kingery, J. Am. Ceram. Soc.66, C171

(1983).203. A. N. Sreeram and L. W. Hobbs, inBeam-Solid Interac-

tions: Fundamentals and Applications,edited by M. Nastasi,



gy

uele-

ta

ta

nd

v.

ay

-

s,

r,

L. R. Harriott, N. Herbots, and R. S. Averback (Mater. Res. SoSymp. Proc.279, Pittsburgh, PA, 1993), p. 559; A. N. SreeramSc.D. Thesis, MIT (1995).

204. D. B. Spilman, L. L. Hench, and D. E. Clark, Nucl. Chem. WasManagement6, 107 (1986).

205. L. W. Hobbs, F. W. Clinard, Jr., S. J. Zinkle, and R. C. EwinJ. Nucl. Mater.216, 291 (1994).

206. J. Yu, W. F. Sommer, and J. N. Bradbury, J. Nucl. Mat141–143, 65 (1986); J. Yu, W. F. Sommer, J. N. BradburyW. V. Green, and M. Victoria, J. Nucl. Mater.227, 266 (1996).

207. L. W. Hobbs, inQuantitative Electron Microscopy,edited by J. N.Chapman and A. J. Craven (Scottish Universities Summer Schoin Physics Publ., Edinburgh, 1984), p. 399; Scanning MicroscoSuppl. 4, 171 (1990).

208. C. R. A. Catlow, K. M. Diller, and L. W. Hobbs, Philos. Mag.42,123 (1980).

209. L. W. Hobbs, J. Phys. 37 (C7), 3 (1976).210. Computer Simulation of Radiation Effects in Solids,edited by

T. Diaz de la Rubia, G.H. Gilmer, and M-J. Caturla (NortHolland, Amsterdam, 1995).

211. 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).

212. T. Mattila, R. M. Nieminen, and M. Dzugutov, Nucl. InstrumMethods in Phys. Res. B102, 119 (1995).

213. N. V. Doan, Philos. Mag.A49, 683 (1984).214. R. G. Della Valle and H. C. Andersen, J. Chem. Phys.97, 2682

(1992).215. G. Cormier, T. Peres, and J. A. Capobianco, J. Non-Cryst. So

195, 125, (1996).216. B. Vessal, A. C. Wright, and A. C. Hannon, J. Non-Cryst. Soli

196, 233 (1996).217. A. N. Cormack and Y. Cao, Mol. Eng. (1996), in press.218. J. M. Delaye and D. Ghaleb, J. Non-Cryst. Solids195, 239 (1996).219. J. M. Delaye and D. Ghaleb, Mater. Sci. Eng.B37, 232 (1996).220. J. M. Delaye, presented at this panel meeting (unpublished).221. J. M. Delaye and D. Ghaleb, J. Nucl. Mater. (in press).222. J. M. Delaye and D. Ghaleb, inProc. of the Third Interna-

tional Conference on Computer Simulations of Radiation Effein Solids,University of Surrey, Guildford, UK, July 22–26, 1996

223. J. D. Wicks and R. L McGreevy, J. Non-Cryst. Solids192yyy193,23 (1995).

224. J. D. Wicks, L. Börjesson, G. Bushnell-Wye, W. S. Howells, anR. L. McGreevy, Phys. Rev. Lett.74, 726 (1995).

225. J-G. Park and A. N. Cormack, Philos. Mag. B73, 21 (1996).226. A. D. Brailsford and R. Bullough, J. Nucl. Mater.44, 121 (1972).227. U. Jain and A. B. Lidiard, Philos. Mag.35, 245 (1977).228. N. Yu, M. Nastasi, T. E. Levine, J. R. Tesmer, M. G. Hollande

C. R. Evans, and C. J. Maggiore, Nucl. Instrum. Methods in PhRes. B 99, 566 (1995).

229. C. W. Allen, L. L. Funk, E. A. Ryan, and A. Taylor, Nucl. InstrumMethods in Phys. Res. B40yyy41, 553 (1989).

IP address: 144.82.107.169

2, No. 8, Aug 1997


Documents

Journal of MATERIALS RESEARCH Welcome Comments Help · 2015. 7. 20. · Materials Science Division, Argonne National Laboratory, Argonne, Illinois 60439 ... V. RESEARCH FACILITIES