|
Radiation Effects and
Radioactive Waste Management |
|||
 |
 |
 |
 |
 |
 |
 |
 |
|
Radiation Effects and
Radioactive Waste Management |
|||
|
|
|
|
|
|
|
|
|
|
|
Radiation Effects in Nuclear Waste
Forms for High-Level Radioactive Waste Progress in Nuclear Energy An International Review Journal Vol.
29, No. 2, pp. 63-127 1995 R.C.
Ewing University
of New Mexico Albuquerque,
New Mexico 87131 W.J.
Weber Pacific
Northwest Laboratory, P. O. Box 999 Richland,
Washington 99352 F.W.
Clinard, Jr. Los
Alamos National Laboratory Los
Alamos, New Mexico 87545 High-level
nuclear waste (HLW) in the United States has three major sources: i.)
spent nuclear fuel from commercial and research nuclear reactors; ii.)
liquid waste produced at West Valley, New York, in the 1960s during the
reprocessing of commercial spent nuclear fuel; iii.) waste
generated by the nuclear weapons and naval propulsion programs. Much of
the defense waste is stored in tanks at Hanford and Savannah River sites
in Washington and South Carolina, respectively. The volumes are
substantial. At the Hanford site alone the inventory includes 11 million
m3 of fluids and 6,900 metric tons of nuclear materials, which includes
4,100 metric tons of uranium and 15 metric tons of Cs and Sr in
capsules. The total activity of the waste at the Hanford site is
estimated to be 446 million curies (Illman, 1993). For all of the
principal DOE sites (Hanford, Savannah River, Idaho Chemical Processing
Plant and the West Valley Demonstration Project) the combined total
volume of high-level waste (through 1988) is 385,000 m3, and the total
activity is 1,200,000,000 Curies (Office of Technology Assessment,
1991). Although the defense wastes account for 95 percent of the volume
of the current total United States HLW inventory, they account for only
9 percent of the total radioactivity. The majority of the radioactivity
is contained in commercial spent nuclear fuel (Weber, 1991a). Except in
the case of direct disposal of spent fuel which is the current U.S.
policy, these wastes must be treated and solidified prior to permanent
disposal. Finally, under the first and second Strategic Arms Reduction
Treaties (START I and II) and unilateral pledges made by the United
States and the former Soviet Union and Russia, many thousands of nuclear
weapons will be dismantled during the next decade. This will generate
over one hundred metric tons of plutonium and many hundreds of tons of
highly enriched uranium that will require disposal (National Academy of
Sciences Report, 1994). To
the extent that these radioactive wastes must be treated and solidified
prior to final disposition, the effect of radiation on waste form solids
is a critical concern. One of the most important applications of the
study of radiation effects in ceramics (vitreous and crystalline) during
the past decade has been the effects of radiation on nuclear waste
forms, e.g., borosilicate glasses, single and polyphase crystalline
ceramics, and the UO2 in spent nuclear fuel (Weber and Roberts, 1983;
Matzke, 1988a, 1988b; Ewing et al., 1988; Weber, 1991a). In the case of
spent fuel, the radiation dose due to the in-reactor neutron irradiation
is already substantial, and additional damage accumulation during
disposal is not anticipated to be significant. In contrast, the
post-disposal radiation damage to waste form glasses and crystalline
ceramics is significant. The cumulative
a-decay
doses per gram which are projected for the nuclear waste glasses to be
used for the West Valley Demonstration Project (WVDP) and the Savannah
River Plant (SRP) reach values of 1016
a-decay
events/g in 100 years (Weber, 1991a). Similarly, crystalline waste
forms, e.g. the titanate assemblage, Synroc, would reach values of
>1018 a-decay events/g in one
thousand years for a 20 weight percent waste loading. For this Synroc
composition, the total accumulated dose at the end of 10,000 years is
nearly equivalent to 0.5 displacements per atom. This dose is well
within the range for which important changes in physical and chemical
properties, most notably the transition from the crystalline-to-aperiodic
(amorphous or metamict) state can occur. Although
present performance assessments of radionuclide containment rely
primarily on the geologic repository, the physical and chemical
durability of the waste form can contribute greatly to successful
isolation of nuclear waste (Ewing, 1992a). Thus, the effect of radiation
on physical properties and chemical durability are of importance. The
changes in chemical and physical properties occur over relatively long
periods of storage, 104 to 106 years, and at temperatures which range
from 100 to 450°C depending on waste loading, age of the waste, depth
of burial, and the repository-specific geothermal gradient). Thus, one
challenge is to effectively simulate radiation damage at relatively low
fluxes over long periods of time to relatively high total doses. In this
regard, studies of naturally occurring U- and Th-bearing phases that are
structurally and chemically analogous to waste form phases have provided
important data and reference points for long-term effects (Ewing et al.,
1988). The
essential issues addressed and summarized in this review are the
following: 1.
What are the changes in physical properties (e.g., volume, fracture
toughness, and hardness), chemical properties (e.g., thermodynamic
stability and leach rate), and stored energy as a function of radiation
dose, temperature, and time? 2.
How can the changes in physical and chemical properties to relevant
radiation doses be simulated in accelerated experiments (for a waste
form age of 104 to 106 years)? We
begin with a brief history of the development of nuclear waste forms and
a summary of their types. 1.1 History The
early history of the development of nuclear waste forms is summarized by
Lutze (1988). These efforts, mainly in the 1950's, focused on the
incorporation of radioactive waste into glasses of widely varying
compositions. The first radioactive wastes were incorporated into a
nepheline syenite (a silica-deficient igneous rock) glass at Chalk
River, Canada. The processing temperature of the nepheline syenite glass
(1,350°C) was much higher than that of the now widely adopted
borosilicate glasses (1,100° to 1,150°). Major programs during the
1960's in the United States, Canada, Germany, France, Italy, Japan, the
United Kingdom, and the Soviet Union were directed at developing glass
compositions that had lower melting points in order to reduce loss of
radionuclides by volatization, incorporated waste without phase
separation, resulted in glasses which were durable, and for which
vitrification was a feasible large scale industrial process.
Borosilicate glass is now the waste glass composition of choice for most
nations. Industrial use of the vitrification process has been in
progress since 1978. During
the period from 1977 to 1982, there was a tremendous diversity in the
types of nuclear waste forms which were under development (Lutze and
Ewing, 1988). In the United States, much of this work ended with the
decision to use borosilicate glass as the waste form for U.S. defense
waste at Savannah River (Hench et al., 1984). Synroc, a titanate ceramic
waste form, was selected as the alternative U.S. waste form, but further
development in the United States ended with the absence of funded
programs. A major research and development program for the development
of Synroc continued in Australia, culminating in the construction of a
"cold" Synroc pilot-scale processing plant. Basic research on
the properties of Synroc were continued at the Australian National
University by the late Professor Ted Ringwood and his colleagues (see
Ringwood, 1985 and Ringwood et al., 1988) and at the Australian Nuclear
Science and Technology Organization, with collaborative work at the
Japanese Atomic Energy Research Institute and AERE Harwell in the United
Kingdom. Synroc is perhaps the most thoroughly studied ceramic
alternative to borosilicate glass. Investigations
into the properties and performance of other ceramic waste forms have
revived during the past ten years for application to special waste
stream compositions. At Lawrence Livermore National Laboratory, a Mixed
Waste Management Facility is being developed in order to demonstrate an
alternative to incineration (Oversby et al., 1994). The waste form under
study is a derivative of a Synroc composition originally developed for
the immobilization of reprocessed residues at Savannah River, and
typical phases include zirconolite, perovskite, spinel, nepheline and
rutile. Although the radioactivity is low, this does illustrate the
ubiquity of a rather limited set of phases, many of which are discussed
in this paper. At the Idaho National Engineering Laboratory, an
iron-enriched basalt waste form is under development, and the addition
of ZrO2 and TiO2 has produced zirconolite as an actinide host in a
silicate ceramic, i.e., a basalt (Reimann and Kong, 1994). At Argonne
National Laboratory glassy slags have been developed for mixed wastes
with high metal contents (Feng et al., 1994). The glassy slag may have
crystalline components which are as high as 80 volume percent in a
vitreous matrix. Spent fuel - a metal clad, ceramic oxide - has become
the major high level waste form under consideration in the U.S., as well
as in Canada, Sweden, and Switzerland, since these countries have
decided against reprocessing (see summary of recent work in Ewing,
1992b). 1.2 Principles of nuclide isolation in solid
waste forms Borosilicate
glass is, at present, the waste form of choice for most countries that
have defense waste or have reprocessed high-level waste (e.g., France,
Great Britain and the United States) and for most waste compositions.
The selection of borosilicate glass is based mainly on an anticipated
ease of processing (glass frit and the waste are mixed, melted at
relatively low temperatures, and poured into canisters), the fact that
the technology is well demonstrated for actual (radioactive) waste, and
finally the assumption that the glass as an aperiodic solid will easily
accommodate wide variations in waste stream compositions which are
extremely complex (20 to 30 component systems), as well as not be
susceptible to radiation-induced transformations (it is already
aperiodic). Ceramic waste forms, with the exception of Synroc, have
never been developed to a degree that the processing technology can be
evaluated or demonstrated with actual waste. The basis for comparison
and evaluation has been bench scale samples and usually samples that are
not radioactive. Despite the lack of data on waste forms that have
incorporated actual waste, the knowledge of radiation effects in ceramic
waste forms is much further advanced than that of radiation effects in
glass waste forms. In
contrast to glass waste forms in which the radionuclides are in
principle homogeneously distributed throughout the waste solid, ceramic
waste forms may incorporate radionuclides in two ways: (1)
Radionuclides may occupy specific atomic positions in the periodic
structures of constituent crystalline phases, that is as a dilute solid
solution. The coordination polyhedra in each phase impose specific size,
charge, and bonding constraints on the nuclides that can be incorporated
into the structure. This means that ideal waste form phases usually have
relatively complex structure types with a number of different
coordination polyhedra of various sizes and shapes and with multiple
substitutional schemes to allow for charge balance with radionuclide
substitutions. Extensive nuclide substitution can result in cation and
anion vacancies, interstitial defects, and changes in structure type.
The formation of polytypes and twinning on a fine scale is common. The
point defects can themselves become sites for the radionuclides. Except
in unusual situations (e.g., monazite, CeP04), the complexity of the
waste composition results in the formation of a polyphase assemblage
(e.g., Synroc consists of phases such as zirconolite, CaZrTi207;
perovskite, CaTi03; and "hollandite", BaAl2Ti6016), with
unequal partitioning of radionuclides between the phases. In Synroc,
actinides partition preferentially into the zirconolite and perovskite.
The polyphase assemblages are sensitive to waste stream compositions,
and minor phases form, including glass segregated along grain
boundaries. Ideally, all waste stream elements, radioactive and
non-radioactive, are important components in the phases formed. In some
rare cases, a single phase (e.g. monazite or sodium zirconium phosphate)
can incorporate nearly all of the radionuclides into a single structure. (2)
Radioactive phases, perhaps resulting from simply drying the waste
sludges, can be encapsulated in non-radioactive phases. The most common
approach has been to encapsulate individual grains of radioactive phases
in Ti02 or Al203, mainly because of their extremely low solubilities.
This approach requires major modifications to the waste stream
composition and special processing considerations to keep temperatures
low enough to avoid volatilization of radionuclides. A similar approach
may be taken with low temperature assemblages (e.g. mixing with
concrete), but in this case there is the possibility of reaction between
the encapsulating phase and the radioactive phases. Both
of the above types of waste forms are specifically formulated for the
incorporation or encapsulation of radionuclides. Spent
fuel - a metal-clad, U02 ceramic - results from the use of reactor fuel
that is designed without consideration of its waste form properties; but
in order to avoid reprocessing, spent fuel has become an important,
potential waste form. The properties of spent fuel as a waste form are
determined primarily by the irradiation history of the UO2 in the
reactor and the disposal conditions (e.g., nearly insoluble under
reducing conditions and easily corroded under oxidizing conditions).
Radionuclides are distributed through the fuel matrix as interstitial
defects, as exsolved/precipitated phases, along grain boundaries, or in
voids and cracks of the fuel-cladding gap. There is an extensive
literature of neutron damage effects in UO2 (Nakae, et al., 1979) and a-particle damage (Weber,
1981), as well as radiation damage in naturally occurring UO2+x (Janeczek
and Ewing, 1991). A summary of the work on spent fuel is beyond the
scope of this review; thus, radiation effects in UO2 as a nuclear waste
form will not be discussed. 1.3 Types of Nuclear Waste Forms The
following sections give brief summaries of the characteristics of the
different waste forms. A careful description of the characteristics of a
waste form is essential in anticipating potential radiation effects. As
an example, the waste loading ultimately determines the radiation dose
and the thermal history of the waste form (i.e., the decay of fission
products is responsible for the thermal pulse in the early history of
the waste form). The final damage state will be a function of the
balance between accumulated damage and the kinetics of annealing of the
individual phases. For polyphase waste forms, which includes both
glasses and ceramics, the partitioning of actinides and fission products
into separate phases can have dramatic differential effects, such as the
selective amorphization of phases, while others retain their
crystallinity. Anisotropic expansion of phases that experience different
radiation doses of different types of radiation may lead to
microfractures and disaggregation of the waste form. Despite the
complexity of crystalline, polyphase ceramics, certain structure types
are common to a wide variety of waste forms and waste stream
compositions. This summary focuses on the present knowledge of radiation
effects in the most common phases, e.g., pyrochlore, zirconolite,
hollandite, perovskite, zircon, apatite and monazite. Extended
discussions of the properties of the waste forms and additional details
on radiation effects can be found in Lutze and Ewing (1988). 1.3.1
Glass waste
forms can be of a wide variety of compositions, including silicate
glasses, borosilicate glasses, and phosphate glasses (summarized by
Lutze, 1988). In principle, radionuclides are evenly dispersed
throughout the glass, although noble metal precipitates and, in some
cases, crystalline oxide precipitates enriched in radionuclides may be
present. Waste loadings (defense and commercial) are typically in the
range of 10 to 30 weight percent. 1.3.2
Synroc
is
a titanium based, polyphase ceramic developed at the Australian National
University and the Australian Nuclear Science and Technology
Organization. The primary phases are: zirconolite (CaZrTi207),
hollandite-like phase (Ba1.2(Al,Ti)8016), perovskite (CaTi03), and
titanium oxides (e.g., TiO2). Other phases, such as pyrochlore
(A1-mB206(0,OH,F)1-n pH20; A = Ca, REE, U, Th, Pu, etc. ; B = Ti, Nb,
Ta) may be abundant depending on the waste stream composition. Minor
phases also include titanates and aluminates, as well as noble metals.
Processing conditions are sufficiently reducing to form Ru-rich and PdTe-rich
phases. The principal components can each accommodate a range of
radionuclides. Hollandite can retain fission products such as Cs, Rb, Ba;
zirconolite, U, Zr, Np, Pu; perovskite, Sr and transuranics such as Np
and Pu. The variable compositions of the waste are accepted over
relatively wide ranges by simple and complex atomic substitutions within
the structure of individual phases and by changing the relative
proportions of the principal phases. Certain elements (e.g., Na20
greater than 2 wt. % together with SiO2 and Al2O3) can lead to the
formation of more soluble silicate phases (e.g. nepheline, NaAlSi04).
Still, waste loadings up to 20 weight percent can be made without
deleterious effects on the properties of the waste form. 1.3.3
Tailored ceramics were
developed mainly at the Rockwell International Science Center (Harker,
1988). Another silicate and oxide based ceramic, supercalcine,
was developed at the Pennsylvania State University (McCarthy, 1977;
McCarthy et al., 1979). These crystalline, ceramic waste forms are also
polyphase assemblages, not different in principle from Synroc, but
perhaps broader in conceptual design with a wider variety of
"high-integrity mineral-like crystalline phases" (McCarthy et
al., 1979). The strategy again is to modify the waste stream composition
so as to form specific, stable, crystalline phase assemblages. A broader
range of phases (oxides, silicates and phosphates) are considered for
radionuclide immobilization. Actinides are incorporated into fluorite
(CaF2), zirconolite, pyrochlore, perovskite (CaTiO3), monazite (CePO4),
apatite (Ca5(PO4)3F), and zircon (ZrSiO4); Sr into magnetoplumbite
(Pb(Fe,Mn)12O19), perovskite and hollandite; Cs into nepheline,
perovskite, magnetoplumbite, hollandite; Tc into reduced metal alloys.
Residual glass with fission products can occur along grain boundaries.
Non-radioactive phases may be prominent, such as spinel (MgAl2O4),
corundum (Al2O3) or rutile (TiO2). As an example, a ceramic waste form
tailored for the Savannah River Defense waste consisted of
magnetoplumbite, spinel, nepheline, uraninite (UO2+x, and high oxides)
and corundum, and a ceramic waste form tailored for high-level Barnwell
waste consisted of pyrochlore, perovskite, monazite, fluorite and a Ru
alloy. In general, the silicate based assemblages are less stable than
the titanates in the presence of water. Waste loadings are comparable to
that in Synroc but the higher density of Synroc (4.35 gm/cc) vs. a
typical tailored ceramic (4.05 gm/cc) means that a slightly greater
amount of waste can be incorporated in Synroc for a weight-equivalent
waste loading. 1.3.4
The Ti02-ceramic matrix
waste form was developed at the Kernforschungszentrum Karlsruhe (Adehelm
et al., 1988) and encapsulates the waste in a rutile (Ti02) matrix.
Waste loadings of simulated waste oxides of up to 12 wt. % were attained
in the laboratory scale experiments. The final product has a low leach
rate as compared to borosilicate glass, because of the low solubility of
Ti02. 1.3.5
Glass ceramics have
been developed at the Hahn-Meitner-Institut in Berlin and at the
Whiteshell Nuclear Research Establishment in Canada (Hayward, 1988). The
Canadian glass ceramic consists of discrete crystals of sphene (the
proper mineral name is "titanite"), CaTiSi05, within a matrix
of aluminosilicate glass. The sphene structure is capable of accepting a
wide range of nuclides. Actinide elements, as well as fission products
such as Sr, can substitute for the Ca. Larger cations, such as Cs, do
not fit into the sphene structure and remain in the glass phase. The
aluminosilicate glass matrix that remains after sphene crystallization
is rather durable and can be produced in the glass-ceramic at lower
temperatures than those required to produce the glass alone. Leach rates
under comparable conditions are lower than for borosilicate glass. In
the Canadian repository (granitic), sphene, instead of perovskite, is
the thermodynamically stable phase. Optimal waste loadings are
approximately 15 wt. percent. A number of other glass ceramic
compositions have been studied in less detail. These include: 1) celsian
glass-ceramics which yield, as individual phases, celsian (BaAl2Si208),
perovskite (CaTi03), diopside (CaMgSi206), or eucryptite (LiAlSi206); 2)
fresnoite glass-ceramics which form fresnoite (Ba2TiSi208), priderite ((K,Ba)(Ti,Fe)8016),
pyrochlore (A1-mB206(0,OH,F)1-n pH20) and scheelite (CaW04); 3) basalt
glass-ceramics which form augite ((Ca,Mg,Fe)2Si206) and the spinel
trevorite (NiFe204); 4) iron-enriched basalt glass-ceramics which form
spinel (enriched in Fe304), plagioclase (CaAl2Si208), augite
((Ca,Mg,Fe)2Si206), fluorapatite (Ca5(P04)3F) and zircon (ZrSi04). Other
glass ceramic compositions have been studied mainly in the hope of
improving the leach resistance of the borosilicate glasses. Perhaps the
most important feature of the glass-ceramics is that the processing
technology is not very different from the demonstrated technology for
the borosilicate glass. The only difference is the required final heat
treatment, and depending on the specific glass-ceramic used (e.g. the
sphene glass-ceramic), there is a resulting increase in chemical
durability. 1.3.6
Monazite developed
at Oak Ridge National Laboratory is unique in that it consists of
essentially a single phase, monazite (CeP04) (Boatner and Sales, 1988).
This composition can be synthesized for the full range of lanthanide
orhophosphates. Depending on the lanthanide composition and the
temperature, at least three structure types are possible. A low
temperature hexagonal phase forms for compositions in the first half of
the lanthanide series (from La to Dy). The hexagonal phase is metastable
and will not form once the structure is transformed to the monoclinic
monazite structure (Beall et al., 1981; Mullica et al., 1985). The
heavier lanthanides (Er to Y) have a high temperature tetragonal
structure isostructural with zircon (ZrSi04) (Mullica et al., 1990).
Actinides are incorporated into the lanthanide site. Important divalent
fission products, Sr and Ba, are also accommodated in the monazite
structure in order to provide charge balance for tetravalent actinides.
As with many of the other ceramic waste forms, there is an important
improvement in chemical durability (factor of 20 in leach rate for
selected elements) as compared to that of borosilicate glass under
comparable conditions. In contrast to other waste form phases, the
lanthanide orthophosphates have a negative temperature coefficient of
solubility; thus, one may expect the chemical durability in the presence
of solutions to increase with increasing temperature. These compounds
also have a high thermal stability with melting points in excess of
2,000°C. Typical waste loadings for simulated U.S. defense waste are 20
wt. % (monazite waste form density is in the range of 4.0 to 5.0 gm/cc).
1.3.7
Cementitious waste forms are mainly considered for low-level
waste (Jiang, et al., 1993; Quillin et al., 1994). Although concrete has
been used for low-level waste, FUETAP (Formed Under Elevated Temperature
and Pressure) concrete was developed at Oak Ridge National Laboratory
and the Pennsylvania State University (McDaniel and Delzer, 1988) and is
unique in that it uses the inherent heat of the high-level waste (as
well as external heat sources) to accelerate curing to drive off up to
98% of the unbound water and form a hard, dense product of improved
(compared to normal concrete) physical properties. The identification
and characterization of the specific phases formed and the determination
of the fate of radionuclides in FUETAP have not been completed. Typical
waste loadings are on the order of 15 to 25 wt. percent (density is
approximately 2 gm/c3). Leach rates are comparable to those of
borosilicate glass. In addition to radiation damage of specific phases
in cementitious materials, there is the additional possibility of
radiolysis of cement-pore water and potential pressurization of waste
canisters due to an accumulation of the radiolytically produced gases.
Solid-state radiolysis of hydrated phases may also effect their
long-term stability. 1.3.8
Spent fuel (Johnson
and Shoesmith, 1988) was first considered as a possible waste form in
the early 1970's due to delays and cancellations of fuel reprocessing
programs. The continued, depressed prices for uranium and lower
projected nuclear power generating capacity have made the disposal of
spent fuel a likely possibility. Typical fuel consists of U02 pellets.
Stacks of the pellets are clad in Zircaloy-2 or -4. Individual pellets
are highly pure U02, with 95% of theoretical maximum density and with a
uniform grain size that is typically 2 to 4 mm.
The irradiated fuel is different chemically and structurally from the
unirradiated fuel. Actinides and lanthanides, as well as some fission
products, form solid solutions with the U02. In some cases new compounds
form. Cs and I react and form CsI, and some Cs is thought to form cesium
uranates or cesium molybdates. The elevated temperatures (up to 1700°C)
during irradiation promote grain growth with the resulting segregation
of fission products that are incompatible with the U02 structure.
Cracking and void formation are common and volatile species (Xe, Kr, Cs
and I) can accumulate. Noble metal inclusions (up to 10 mm) are also found. Despite
the apparent complexity of spent fuel as a waste form, its direct
disposal in a repository eliminates the need for reprocessing. The
chemical durability is difficult to assess because it is dependent on
the irradiation history and on the oxidation potential of the
repository. Under reducing conditions, U02 is very insoluble, but under
slightly oxidizing conditions matrix corrosion proceeds by oxidative
dissolution. Redox conditions may be influenced by a- and gamma- radiolysis of the
solution in contact with the spent fuel. Under the
"extended-dry" repository concept (e.g., at Yucca Mountain),
oxidation to U3O8, accompanied by a volume increase and cladding
rupture, is a concern. If air enters the waste packages at elevated
temperatures, the subsequent entry of water could result in relatively
high leach rates for the spent fuel. 1.3.9
"Novel" ceramic waste forms include consideration of
some additional, unusual phases. Most of this work has been completed at
the Pennsylvania State University (Ewing, 1988). One such phase is
sodium zirconium phosphate, NaZr2(P04)3 (NZP). The structure consists of
an open three dimensional framework with three distinctly different
atomic sites, which can accommodate a wide range of radionuclides. At
high simulated waste loadings (20 wt. %), an additional phase, monazite,
forms. NZP can also be mixed with concrete formulations to act as a host
for Cs, resulting in a CsZP. Its main advantages lie in its potential
for low-temperature processing and as a specific additive to concrete
formulations. A similar waste form ceramic is the low-temperature
hydroxylated ceramic. These consist of hydrated sheet silicates (clays)
and framework silicates (zeolites) that in most cases form at low
temperatures. Historically, clay phases were among the first to be
suggested for the immobilization of nuclear waste (Hatch, 1953). The
principal advantage of both of these low-temperature waste forms is that
they are the common, thermodynamically stable phases at the earth's near
surface, the environment of most repositories. 1.3.10
Multi-barrier waste forms are unique combinations of the previously
discussed waste forms and some special processing technologies.
Multi-barrier waste forms consist of a series of glass or ceramic
barriers. This is the "Russian doll" concept proposed by R.
Roy (1979). Phases are either encapsulated in other more durable phases
or the particles are coated with another phase. The coatings are applied
to enhance chemical durability, mechanical strength or thermal
stability. The coatings may be conventional ceramics (Al203, Ti02, or
Si02), carbon products (PyC, Cr7C3 or SiC), glass (borosilicate or
aluminosilicate) or metals (Ni, Si or Fe). A common difficulty is the
proper attachment of coatings to substrate particles without cracking or
flaking of the layer. Differential expansion of phases due to radiation
damage may exacerabate the cracking of the thin films. 1.4 Radiation Sources and Doses The
chief sources of radiation in high-level nuclear waste forms are the b-decay of fission products
(e.g., 137Cs and 90Sr) and the
a-decay
of actinide elements (e.g., U, Np, Pu, Am and Cm). There are minor
contributions to the damage process from spontaneous fission events of
the actinides, and
a-neutron reactions are
additional sources of fission fragments and neutrons, but the production
rates are low. Therefore, these latter events are not considered in this
paper. b-decay is the primary source
of radiation during the first 500 years of storage, as it originates
from the shorter-lived fission products (e.g., the half-life of 137Cs is
30.2 years and the half-life of 90Sr is 28.1 years). The b-decay
of fission products is responsible for heat generation and the elevated
temperatures early in the history of waste form storage. The
a-decay damage remains dominant
as a radiation damage source after approximately 1,000 years (e.g., the
half life of 239Pu is 2.411 x 104 years). The
b- and a-decay events can cause radiation damage through
three processes in the waste form: (1) elastic collisions between
nuclear particles (e.g.,
b-particles,
a-particles and
a-recoil nuclei) and the atoms
in the host matrix which cause atomic displacements and the creation of
isolated Frenkel defect pairs (combined interstitials and vacancies) or
intense collision cascades; (2) ionization effects associated with the b- particles and a-particles; (3) the
transmutation of radioactive parent nuclei into different elements. Of
these three effects, the most important are processes that cause atomic
displacements, as these are responsible for the atomic-scale
rearrangement of the structure and hence lead to the greatest changes in
physical and chemical properties. Each
a-decay event results in
approximately 1,500 atomic displacements, while a b-decay
event results in only 0.15 to 0.10 displacements. There is no compelling
evidence that ionization events result in important structural changes
in most oxide ceramics, (Weber et al., 1981, Weber et al., 1982; Weber
and Roberts, 1983; Weber et al., 1984); however, more recent work by
DeNatale and Howitt (1985, 1987) and summarized by Weber (1991a) has
shown that complex borosilicate glasses can decompose by an
ionization-driven radiolytic process that produces bubbles which contain
molecular oxygen. Although systematic studies are lacking, the bubble
formation, glass decomposition, and even changes in chemical durability
may correlate with the ionizing component of energy deposition. These
phenomena will be discussed in Section 4. The
cumulative doses for the waste forms can be substantial. Table 1
summarizes the anticipated cumulative doses for a,
b
and g radiations for waste glasses
that will be generated at the Savannah River Plant in South Carolina. If
one converts these values to atomic displacements per gram or absorbed
dose, the glass waste form with a 25 wt.% waste loading will reach
values in excess of 1020 atomic displacements/g ( = 0.0025 displacements
per atom, dpa) or 7 x 1010 rads in only 1,000 years (Figures 1 and 2).
Also included in Figures 1 and 2 are the cumulative doses generated
within non-U.S. commercial waste glasses, which are several orders of
magnitude higher than those for defense waste. Figure 3 shows the
accumulated a-decay
event dose in Synroc for 20 and 10 weight percent waste loadings. Note,
that for a 20 wt. percent waste loading, the accumulated dose after 104
years is 4 X 1018
a-decay
events/g ( = 0.32 dpa). This is well within the range of doses in which
one may expect amorphization of actinide-bearing phases. Phases such as
zircon can become aperiodic (or metamict) at doses as low as 0.3 dpa.
When solids are subjected to irradiation, any or all of three
responses are possible: heating, localized displacement of the
constituent ions, and disordering or other global rearrangements of
those ions. Further, the irradiating particles themselves can in some
cases have a significant effect on evolution of the damage
microstructure, through their deposition in the structure. In this
section we discuss how damage effects vary with type of radiation (which
in nuclear waste principally incompasses a,
a-recoil, b, and g radiation),
and how damage mechanisms can vary with type of waste form. 2.1.
Heating When
energy from any type of radiation is absorbed in a solid, an increase in
temperature will result. The magnitude of the temperature increase
depends on the rate of energy absorption, physical properties of the
solid (e.g., its specific heat), and the extent of thermal conductance
to the surroundings. In the case of nuclear waste, significant heating
is possible: it can be shown that commercial nuclear waste forms
containing typical waste loadings will, when placed in a repository,
generate enough self-heat from decay of fission products to result in an
initial storage temperature as high as 600°C. Even 100 years after
emplacement, the temperature may still be as high as 300°C. Temperature
increases of this magnitude can profoundly affect the response of
nuclear waste forms to self-irradiation damage. 2.2 a- and
a--recoil
irradiation
Self-damage
in nuclear waste forms results primarily from decay of
a-emitting actinide isotopes.
The a-particle (helium ion) has an
energy of 4.5 to 5.5 MeV; whereas, the recoil nucleus has an energy of
70 to100 keV; thus the
a-particle
carries 98% of the energy of the decay event. In estimating the
consequences of irradiation of a solid by light and heavy particles of
varying energies, it is important to estimate whether a given particle
will deposit its energy primarily by displacive (elastic) or by
ionization (inelastic) processes. This is determined by the relative
velocity of the bombarding particle and that of the orbital electrons of
the target ion. If the particle velocity is below that of the orbital
electrons, the likelihood of electronic excitation, is small and most of
the energy will be transferred to the nucleus of the ion. However, if
the particle velocity is higher than that of the orbital electrons,
electronic excitation will dominate. As a rough rule, inelastic
processes are important if the energy of the bombarding particle,
expressed in keV, is greater than its atomic weight. Thus an
a-particle
of atomic weight 4 and energy 5 MeV will predominately deposit its
energy by ionization, while an
a-recoil
ion of mass 240 and energy 100 keV will lose most of its energy in
elastic collisions. Reeve and Woolfrey (1980) have estimated that the a-particle accounts for only 6% of the displacive
energy from an
a-decay
event. This energy is deposited over a range of 10 to 20 microns, with
most of the damage occurring near the end of the path of the helium ion. In
the case of irradiation with a light but energetic ionized particle such
as a 5 MeV a-particle, several sequential
atomic collisions usually occur before the particle comes to rest. Each
of these events displaces a small number of ions, so that the overall
effect is a damage event spread widely throughout the solid in small
clusters. As a consequence, recombination of vacancies and interstitials
can easily occur and this process is enhanced by the prevalence of
nearby undamaged material. In
most solids the only structural consequence of electronic excitation is
heating. However, in those materials subject to damage by electrolysis,
breaking of atomic bonds and displacement of atoms can result. For
example, crystalline SiO2 can be transformed to the disordered or
metamict state by electrons whose energy is too low to directly displace
atoms (Hobbs and Pascucci, 1980). With respect to nuclear waste, it has
been observed that borosilicate waste glasses can decompose by
absorption of ionizing energy with the resultant production of molecular
oxygen (DeNatale and Howitt, 1987). The
more massive but lower energy (approximately 100 keV) a-recoil particle has been shown by Reeve and
Woolfrey (1980) to account for 94% of the total displacement energy
resulting from an a-decay event (Van Konynenburg
and Guinan (1983) calculated a value of 89% for Synroc phases). Such a
particle travels only a short distance in a solid, on the order of 10
nm. Thus the displacement event consists of a large number of disordered
atoms in a compact microvolume. In this case most disordered atoms are
surrounded by other disordered atoms, leading to an increased
probability of damage retention.
When an
a-particle stops in a solid, it
becomes a helium atom; if present in sufficient numbers and at high
enough temperatures to allow mobility, these gas atoms can result in the
formation of gas bubbles (Clinard et al., 1970). In some cases these
aggregates affect mechanical properties, especially if located at grain
boundaries. 2.3.
b-irradiation
In
nuclear waste, decay of fission products produces high-energy electrons
(b-particles). Because of their
low mass, these electrons usually induce only single displacement events
(Frenkel pairs); most of the energy of these particles is dissipated by
ionization effects, through Coulombic interactions. However, for those
waste forms in which radiolytic damage is significant, ionization
effects from b-particles could represent a
major source of damage. Additionally, b-decay
is the principal source of heating in waste forms during the first 500
years of waste storage, and therefore plays an important role in
determining the temperature (and thus damage response) of the waste
form. Calculations
have been completed by Weber et al. (1982) to determine the cumulative
number of non-radiolytic atomic displacements in commercial and defense
high-level waste that will result from the two most effective sources of
displacement damage, namely
a-decay
and b-decay . The results, shown in
Figures 1 and 2, demonstrate that a-decay
is dominant and that irradiation fluxes (and therefore heating effects)
are greatest early in the storage cycle. 2.4.
g-irradiation
Ionizing
radiation such as g-rays
can, through ionization processes, cause displacement damage or even
amorphization if the solid is subject to radiolysis. Another source of
damage related to g-irradiation is the energetic
electrons that result from interaction of high-energy electromagnetic
radiation with solids. However, the inefficiency of production of these
electrons and their inefficiency in producing displacement damage means
that this is not a significant source of radiation-induced changes in
nuclear waste materials. Total ionizing dose as a function of the age of
the waste form is illustrated in Figure 2. 3.
Radiation Simulation Techniques 3.1.
Actinide-doping of waste forms and component phases a-decay
events associated with the decay of actinides and their daughter
products are responsible for the atomic-scale structural damage that
will occur in radioactive waste forms. The a-decay
event consists of two, separate, simultaneous processes: (1) An
a-particle ( 5 MeV) with a
range of 10,000 to 20,000 nm dissipates most of its energy by
ionization; however, at low velocities near the end of its track, it
displaces one to two hundred atoms creating Frenkel defect pairs. (2)
The a-recoil
atom ( 0.1 MeV) with a range of 10 nm produces one to two thousand
atomic displacements creating "tracks" of disordered material.
These two damaged areas are separated by thousands of unit cell
distances and, as is illustrated in later sections of this review, will
have different effects on the atomic structure of the solid,
particularly for crystalline materials (see particularly the work by
Weber (1981) on UO2). The long-term effects of
a-decay events can be simulated
by incorporating highly-active, short-lived actinides, such as 238Pu
(half-life of 87.7 years) or 244Cm (half life of 18.1 years) into the
waste form phase in concentrations great enough that the total a-decay
dose reaches values of 1018 to 1019 a-decay
events/g in reasonable amounts of time (this may still require several
years). This is the accepted test procedure drafted by the International
Standards Organization and is the basis for the MCC-6 method for the
preparation and characterization of actinide-doped waste forms (Weber
and Turcotte, 1982). This test effectively simulates the simultaneous a-
and a-recoil
effects to doses that one expects over long periods of time (103 to 106
for waste forms and 108 years for natural phases), but at much
accelerated dose rates. However, as illustrated by the study of zircon
(Section 5.2.5), actinide-doping experiments, with a high dose rate, do
accurately simulate radiation effects at very much lower dose rates over
geologic time . This is also supported by the comparisons presented by
Van Konynenburg and Guinan (1983). 3.2 Actinides in minerals as natural analogues
(metamict state) Naturally
occurring phases (= minerals) contain 238U, 235U and 232Th and their
daughter products. Concentrations may be as high as 30 wt.% UO2 in
pyrochlores, which, depending on the age of the sample, may reach doses
of 1020 a-decay events/g. Other phases,
such as zircon, contain trace amounts of uranium (up to 5,000 ppm), but
with ages that are up to 109 years, the doses may reach 1019 a-decay events/g. In many cases, these doses are
sufficient to cause a radiation-induced periodic-to-aperiodic
transition, that is the metamict state. Metamict minerals are one class
of amorphous materials that have long been recognized to have resulted
from radiation damage (Ewing, 1994), and their properties have been
summarized by Pabst (1952), Ewing (1975) and Ewing et al. (1987).
Minerals that are isostructural with crystalline phases in ceramic waste
forms can therefore serve as natural analogues in the investigation of
radiation damage effects (Ewing and Haaker, 1980; Ewing et al., 1988).
Minerals have the advantage of representing long-term radiation effects,
over hundreds of millions of years, which are the result of very low
dose rates. The dose rates of the actinide-doping experiments described
above and the dose rates experienced by minerals span a range of nine
orders of magnitude. Unfortunately,
there are no natural glasses which contain enough uranium and thorium
that they can serve as natural analogues for radiation effects in waste
form glasses. There is, however, one instance in which a Th-doped glass,
17 years in age, has been studied (Eyal and Ewing, 1993). 3.3 Neutron irradiations Neutron
irradiations can be utilized in three ways (Weber and Roberts, 1983).
Fast neutrons dissipate their energy by elastic collisions resulting in
numerous atomic displacements, and the samples become moderately
radioactive. This provides a fair simulation of a-recoil
damage, but there is no simulation of the effect of helium build-up due
to an absence of a-particles.
Additionally, the only phases in polyphase waste forms that will
experience a-recoil damage are those which
contain actinides. Other phases will experience only a-particle and b- and
g-irradiation.
Neutron irradiations damage all of the phases in the waste form, and
therefore, will not simulate the phase selective a-decay
damage experienced in actual waste forms. Neutron irradiations have been
used to simulate damage in Synroc phases (Reeve and Woolfrey, 1980;
Reeve et al., 1981). Using another approach (Malow and Andersen, 1979),
boron-containing glasses (e.g., the borosilicate glass) or boron-doped
waste forms can be irradiated in a thermal neutron flux, and because of
the high 10B(n, a)7Li cross section, an
a-particle (E = 1.78 MeV) is
generated. With this technique, high rates of He formation are possible,
but this does not simulate the
a-recoil
damage of the
a-decay event. Finally, thermal
neutron capture by selected nuclides (e.g., 235U) can lead to
spontaneous fission, which results in the formation of extensive zones
of atomic displacements, i.e., fission tracks (Antonini et al., 1979).
This is not a good simulation of
a-decay damage (Weber, 1981),
and the number of fission events in nuclear waste forms is so low as to
make the process an unimportant damage mechanism. 3.4.
Charged-particle irradiations Charged
particle irradiations using electrons (Hobbs and Pascucci, 1980),
protons, a-particles (Weber, 1981, 1982,
1985) or heavy ions (Weber et al., 1994) have been used to study
radiation damage effects. Significant doses can be reached in short
periods of time (e.g., minutes), but analysis of the results is
difficult because the damaged areas are thin surface layers of
restricted lateral extent. The high surface area can act as a sink for
migrating defects, and the dose required for amorphization may be
greater than that for bulk irradiations. Implanted layers also can
change the characteristics of the target material (e.g., an increase in
the apparent leach rate) as a result of compositional changes that are
enhanced in thin layers. Still, the
a-particle
irradiation is a good simulation of a-particle
damage, and heavy-ion implantation (Pb2+) provides a good simulation of
a-recoil damage (Headley et
al., 1982). In fact, the two techniques used together can be used to
unravel the relative contributions of
a-particle
and a-recoil damage to the changes
in the properties of the solid. All of these techniques have been used
to study a-decay damage effects in
nuclear waste glasses, nuclear waste ceramics and component phases. Most
recently, Weber et al. (1994) have completed a detailed study of the
radiation-induced crystalline-to-amorphous transition in zircon in which
they made a comparison between damage caused by a-decay
events (238Pu-doped synthetic zircon and U-bearing natural zircons) and
ion-beam irradiated zircon (0.8 MeV Ne+, 1.4 MeV Ar+, 1.5 MeV Kr+, 0.7
MeV Kr+, 1.5 MeV Xe+). The results demonstrated that the amorphization
dose under the high dose rates of the ion irradiation (10-4 to 10-3 dpa/s)
for heavy-ion irradiations is nearly identical to that of Pu-doped
zircon (3 x 10-9 dpa/s), which suggests that the amorphization process
is largely independent of the damage source or the dose rate for heavy
particles (a-recoil particles or heavy
ions) at constant temperature. This suggests that heavy ion beam
irradiations can be used to simulate
a-decay
event damage in individual phases in nuclear waste forms. Irradiations
with light ions (e.g., a-particles) can be used to
simulate a-decay in phases that do not
contain nuclides which decay by a-emission. 3.5 g-irradiations Gamma-irradiations
are easily completed on waste forms using 60Co or 137Cs sources. The
advantage of this type of irradiation is that g-rays
are so penetrating that waste forms can be irradiated in sealed
containers. Dose rates on the order of 2.5 X 106 R/h are easily
achieved. However, structural damage from g-irradiation
is minimal except in those waste forms that are subject to damage by
radiolysis (Hobbs and Pascussi, 1980). In terms of waste form corrosion,
the most important effect of g-radiation
is the radiolytic decomposition of species such as nitrogen and carbon
dioxide which are present in air or dissolved in the water, leading to
the formation of nitric and carboxylic acids. The oxidative dissolution
of UO2 can result from g-radiolysis of water in
contact with spent fuel (Sunder et al., 1992). 4.
Radiation Effects in Glass Waste Forms The
effects of radiation in nuclear waste glasses are complex, and the
fundamental understanding of the radiation damage processes and results
is limited. The high-radiation environment provided by the a-
and b-decay
of radionuclides in nuclear waste can affect radionuclide release to the
biosphere through radiation-induced physical and chemical changes in the
nuclear waste glass at both the atomic and macroscopic levels. Several
comprehensive reviews (Burns et al., 1982a and 1982b; Weber and Roberts,
1983; Day et al., 1985; Weber, 1988 and 1991a; Wronkiewicz, 1993) of
radiation effects in nuclear waste glasses provide excellent technical
assessments, and many unresolved scientific and engineering issues
regarding radiation effects were summarized by Matzke (1988a; 1988b)
following a 1987 workshop. 4.1.
Volume Changes Radiation
effects from a-decay
in actinide-doped simulated nuclear waste glasses can result in either
expansion or compaction of the glass structure. The macroscopic volume
changes, DV/Vo, are usually determined
by density measurements and generally follow an exponential dependence
on dose, D, of the form:
DV/Vo
= A [1 - exp(-BD)]
(1) where
A is the volume change at saturation and B is the amount of glass
damaged per unit dose. The volume changes normally saturate within the
range of +1.2% at a dose equivalent to about 1 x 1018 a-decays/g
(1 x 1011 rad), as illustrated in Figure 4 for several actinide-doped
nuclear waste glasses studied at the Pacific Northwest Laboratory in the
U.S. (Weber and Roberts, 1983; Weber, 1988). Similar behavior has been
observed in a collaborative study on actinide-doped simulated waste
glasses by the Centre de Marcoule, the Hahn-Meitner Institut, and the
Atomic Energy Research Establishment (Day et al., 1985), as shown in
Figure 5. Several proposed glass compositions for Savannah River Plant
waste have also been studied (Bibler and Kelley, 1978; Bibler, 1982;
Weber, 1988), and the volume expansions (Figure 6)are larger than for
other waste glasses studied to date. In a more recent study of an
actinide-doped waste glass by Inagaki et al. (1992), the volume
expansion determined by density measurements at a dose of approximately
5.5 x 1018 a-decays/g
ranged from 0.4 to 0.6%, which they found to be in excellent agreement
with the volume change (0.51%) determined from the measured size and
concentration of irradiation-induced bubbles. These results show a
strong correlation between bubble formation and macroscopic volume
change in nuclear waste glass and also support the hypothesis suggested
by Weber (1988 and 1991b) that the volume expansions observed in earlier
studies (Figures 4-6) may partly be due to the formation of bubbles.
Banba et al. (1994) have also observed a volume expansion of 0.75
percent after 1.5 x 1018 a-decay
events/g in an actinide-doped waste glass from the Tokai Reprocessing
Plant. The bubbles may be of trapped He from the a-decay
event or oxygen from radiolytic decomposition of the glass. The
increases in volume of neutron-irradiated nuclear waste glasses (Antonini
et al., 1980; Sato et al., 1988) are also within the bounds observed for
the actinide-doped glasses. Volume changes have not been measured in any
of the ion-irradiation studies. In the simulated waste glasses that have
been studied, there is a general tendency for swelling in glasses with
low alkali/silicon ratios and for compaction in glasses with high
alkali/silicon ratios (see shaded area of Figure 7). g-irradiation
studies in the U.S. on two simulated nuclear waste glasses have reported
only slight ( < 0.1%) volume changes (Bibler, 1982; Weber, 1988).
Shelby (1980), however, reported up to 1% compaction in commercial
borosilicate glasses after g-irradiation to 1010 rad,
which is comparable to the compaction observed in some actinide-doped
glasses (Figure 4). Sato et al. (1984) have reported swelling of 0.2% in
a simulated nuclear waste glass after g-irradiation,
and swelling of up to 50% (from oxygen bubble formation) after
electron-irradiation. Based on these results, it is impossible to
determine conclusively whether the volume changes observed in
a-decay studies are due to
atomic displacements or ionization damage. However, the ion-irradiation
studies of Arnold (1985, 1988) have strongly indicated that ionization
processes are the dominant mechanism for producing volume changes in
borosilicate waste glasses. Consequently, the volume changes in actual
waste glasses, if due to ionization damage, could saturate within 10,000
years for U.S. defense glasses and within several hundred years for
commercial waste glasses in other countries. While the magnitude of the
above volume changes do not pose any problems for the safe storage of
nuclear waste, they do indicate that changes in microstructure may
affect the integrity of nuclear waste glasses and radionuclide release
rates (which are proportional to the surface area in contact with
solutions). Futhermore, differential volume changes as a result of
temperature dependencies, which have not been investigated, and
compositional inhomogeneities may result in microcracking, residual
stresses, additional loss of integrity, and enhanced radionuclide
release. 4.2.
Stored Energy The
interactions of radiation with nuclear waste glasses can result in the
storage of latent energy associated with the changes in structure and
bonding that is released as heat at elevated temperatures. Studies on
actinide-doped simulated nuclear waste glasses at the Pacific Northwest
Laboratory (Roberts et al., 1976; Weber and Roberts, 1983), AERE-Harwell
(Hall et al., 1976), and under a collaborative European Communities
program (Malow et al., 1980) indicate that
a-decay can result in stored
energy that is generally less than 150 J/g and saturates at a dose of
0.1 to 0.3 x 1018
a-decays/g (1010 rad). This is
illustrated in Figure 8, where the total stored energy for several
simulated waste glass compositions has been determined using
differential scanning calorimetry (Weber and Roberts, 1983). In general,
the increase in stored energy with dose followed exponential behavior
similar to that shown by the changes in volume; however, the dose for
saturation of stored energy is considerably less than the dose required
for saturation of volume changes, which suggests different mechanisms
and associated defects. Primak and Roberts (1984) have postulated that
this stored energy is associated with the breaking of glass network
bonds and may be due to ionization damage from the a-particle.
If this is the case, stored energy from ionization damage associated
with b-decay could saturate in 30
years. Stored energy data on neutron-irradiated waste glasses (Roberts
et al., 1976; Antonini et al., 1980; Cousens and Myhra, 1983) are in
good agreement with the results in Figure 8. No data are available on
stored energy release in g- or electron-irradiated
glasses to evaluate the role of ionization in the absence of significant
displacive energy deposition. The
stored energy in actinide-doped simulated waste glasses is released over
a broad temperature range from 100 to 600°C (Roberts et al., 1976;
Weber and Roberts, 1983). Since the specific heat of the nuclear waste
glasses is on the order of 1 J/(g.K), the maximum temperature rise that
could occur from the rapid release of this stored energy is only 50 to
125 K. Temperature excursions of this magnitude should present no
serious problems. Furthermore, the results of Roberts et al. (1976)
indicate that the stored energy decreases almost linearly with
increasing storage temperature. Thermal recovery of the stored energy in
the PNL glasses (Figure 4) was investigated by 7-day isochronal anneals
(Weber and Roberts, 1983). The results show a single recovery stage that
is similar for all the waste glass compositons, with full recovery
occuring at 360°C (Figure 9). These results are similar to those of
Roberts et al. (1976) for the effects of storage temperature. 4.3.
Transmutations Nuclear
transmutation is the formation of a new chemical element as a result of
nuclear reactions or radioactive decay processes. The principal source
of transmutations in nuclear waste are b-decay
of the relatively abundant fission products, cesium and strontium.
Transmutation of these two elements is accompanied by changes in both
ionic radius and valence. Cesium (1+) decays to barium (2+) with a
decrease in ionic radius of 20%, and strontium (2+) decays to yttrium
(3+), which in turn decays to zirconium (4+) with a final ionic radius
decrease of 29%. In an accelerated study of cesium to barium
transmutation in five glasses and two ceramics, Gray (1982) found no
significant effect of the transmutations on microstructure, chemical
durability, or density. 4.4
Helium accumulation Helium,
which results from the capture of two electrons by
an a-particle,
must be accommodated interstitially or substitutionally, diffuse to
internal defects, or be released at the glass surface. Based on a number
of studies (Hall et al., 1976; Turcotte, 1976; Malow and Andresen, 1979;
Malow et al., 1980) in which helium diffusion and release in nuclear
waste glasses were measured, it can be concluded that the helium
generated will accumulate in the glass, and only a small fraction can be
expected to be released at ambient temperatures. Observed decreases in
diffusion coefficients with increasing dose (Turcotte, 1976; Malow and
Andresen, 1979) suggest trapping of helium at radiation-induced defects.
Low-level helium concentrations (up to 100 appm) introduced both by (n,a)
reactions under neutron irradiation and by accelerator implantation have
resulted in bubble formation in simulated nuclear waste glasses after
post-irradiation annealing at 600°C in one case (Dé et al., 1976a) and
750°C in another (Malow and Andresen, 1979). When higher helium
concentrations (up to 104 atomic parts per million, appm) are introduced
by (n,a)
reactions, bubble formation was observed at ambient irradiation
temperatures (<100°C) by Dé et al. (1976a) and at slightly elevated
irradiation temperatures (<230°C) by Sato et al. (1988). The exact
nature of the bubbles was not determined in either case, but they are
assumed to be helium or helium/oxygen bubbles. In a more recent study,
Inagaki et al. (1992) observed bubble formation in 238Pu- and 244Cm-
doped glasses as a result of a-decay. The change in glass
density determined by observed bubble size and abundance correlated well
with the measured change in density. 4.5.
Microstructural Changes The
principal radiation-induced changes in the microstructures of interest
for nuclear waste glasses are phase changes, phase separation,
microfracturing, and bubble formation. Thermally-induced devitrification
(Jantzen et al., 1984; Spilman et al., 1986) and phase separation (Tomozawa
et al., 1979) in HLW glasses are well known. Radiation-enhanced
devitrification of nuclear waste glasses has not been investigated to
date. Phase separation has been induced in simulated waste glasses by
electron-irradiation (DeNatale and Howitt, 1984 and 1985). Crystalline
phases that often form in waste glasses from thermal devitrification
during cooling and storage (additionally, crystalline phases form in the
glass during solidifcation from the melt) can undergo a
radiation-induced crystalline-to-amorphous transformation. In a study of
a Cm-doped waste glass containing crystallites of
Ca3(Gd,Cm)7(SiO4)5(PO4)O2 (apatite structure) and (Gd,Cm)2Ti2O7
(pyrochlore structure), the differential expansion (5 to 8%) associated
with the crystalline-to-amorphous transformation of these phases
resulted in significant microfracturing as a result of radiation-induced
differential volume changes (illustrated in Figure 10 of Weber et al.,
1979; Weber and Roberts, 1983). The behavior of these two particular
phases is discussed in detail in Sections 5.2.1 and 5.2.6. This
radiation-induced microcracking can significantly increase the surface
area for radionuclide release. As
noted above, helium implantation in waste glasses can result in the
formation of helium bubbles. Oxygen bubbles can also form and were first
observed by Hall et al. (1976) in simulated nuclear waste glasses
irradiated with electrons in a high voltage electron microscope at 200°C
to simulate the effects of b-irradiation. Since then,
irradiation-induced oxygen bubble formation has been observed by others
in a wide variety of glasses as a result of electron-beam irradiations
(Todd et al., 1960; Manara et al., 1982; Sato et al., 1983; DeNatale and
Howitt, 1984, 1985 and 1987; Arnold, 1985; DeNatale et al., 1986; Heuer,
1987; Heuer et al., 1986), ion-beam irradiations (DeNatale et al., 1986;
DeNatale and Howitt, 1987; Heuer, 1987), and g-
irradiations (DeNatale and Howitt, 1985 and 1987; Heuer, 1987; Howitt et
al., 1991). The presence of a gas phase in the bubbles has been
confirmed in the studies of Hall et al. (1976) and Heuer et al. (1987).
The formation of oxygen bubbles, which results from radiolytic
decomposition, has been correlated with the ionization component of
energy deposition and not the ballistic (atomic displacement) component
(DeNatale et al., 1986), in agreement with the results of Arnold (1985
and 1988). Consequently, ionization effects from
a-, b-,
and g-radiation
are more important to nuclear waste glass behavior than originally
assumed. Although high-resolution microstructural characterization was
not performed in the early studies of actinide-doped nuclear waste
glasses, Weber (1988 and 1991a) postulated that similar bubble formation
from a-particles
emitted in actinide-doped nuclear waste glasses may occur and could
account for the volume expansions observed in Figures 4 - 6. Recent
results by Inagaki et al. (1992) confirm that bubbles, probably of He,
with radii of 200 to 300 nm are formed as a result of
a-decay
in a 244Cm and 238Pu-doped simulated nuclear waste glass and are
primarily responsible for the measured volume expansion. 4.6.
Radiolytic Decomposition The
early emphasis of radiation effects studies in nuclear waste glasses on
displacement damage has led to some misconceptions concerning waste
glass stability. Although not originally anticipated, it is now evident
that complex borosilicate glasses, including those proposed for the
storage of nuclear wastes, can decompose by an ionization-driven
radiolytic process that produces molecular oxygen and may lead to oxygen
bubble formation as described above. The initial observation of oxygen
bubble formation by Hall et al. (1976) did not create much concern
because the electron dose required to produce the bubbles was quite
high, 1012 to 1013 rad, exceeding the dose equivalent to a million years
of storage. Nearly ten years later, DeNatale and Howitt (1985) reported
that gamma irradiation of simulated nuclear waste glasses at
significantly lower doses (109 rad) will result in bubble formation;
this observation was of significant relevance to nuclear waste glass
behavior. Subsequent g-irradiation
studies (DeNatale and Howitt, 1987; Heuer, 1987; Howitt et al., 1991)
have concluded that oxygen bubbles form (Figure 11) in a wide variety of
glasses at low doses. Furthermore, in another study (Tosten, 1984; Biber
et al., 1990) of one of the same glass compositions, g-irradaition-induced
decomposition or bubble formation was not observed, even at an order of
magnitude higher dose. These authors suggest that bubble formation in g-irradiated glasses which
were reported by others may be an artifact of sample preparation.
Howevber, quantitative studies have shown that the kinetics of the
formation process are consistent with the motion of alkali metal cations
and not oxygen (DeNatale and Howitt, 1984; Laval and Westmacott, 1980).
The formation of oxygen bubbles is apparently brought about by
radiolytic decomposition of the ionic component of the glass, followed
by the migration of the cations into the glass and the local
precipitation of molecular oxygen (Howitt et al., 1991). The bubbles
observed in an actinide-doped simulated waste glass (Inagaki et al.,
1992) also may be produced by this mechanism, although the role of the
helium (a-particles) and the
composition of the bubbles (oxygen and/or helium) has not be
established. Irradiation
studies of simulated waste glasses carried out by DeNatale and Howitt
(1987) and Heuer (1987), utilizing g-radiation,
ion beams, and electron beams at similar temperatures, indicate a
significant dose-rate effect for glass decomposition and subsequent
bubble formation (Figure 12). Decomposition under g-irradiation
occurs at much lower doses than for either ion- or
electron-irradiations. The process for g-induced
decomposition under ambient conditions (<100°C) has an onset at
about 108 rad, saturates at 109 rad, and results in a uniform
distribution of bubbles (Figure 11) with an average diameter of 5 to 6
nm (Heuer, 1987; Weber, 1988 and 1991a). Under electron irradiation,
bubble formation occurs over a broad temperature range with a maximum in
the rate of formation at 250°C (DeNatale and Howitt, 1984; Heuer,
1987), as illustrated in Figure 13. The effects of temperature on size
and density have not been reported. The temperature dependence of
radiolytic decomposition and bubble formation under g-irradiation
has not been investigated (although DeNatale and Howitt (1985) reported
enhanced bubble formation at 50°C), but given the strong dose-rate
dependence in Figure 12, a shift in the temperature dependence from that
shown in Figure 13 is expected. The
implications of these results may be more serious than other irradiation
studies have suggested, as this radiolytic decomposition process, if
operable, can be completed in relatively short time periods (< 10
years) for actual nuclear waste glasses. In addition, the low-dose
character of this effect underscores the need for a better understanding
of the temperature dependence of this process, since the waste glasses
will be exposed to elevated temperatures and temperature gradients
during this time period. As noted above, two studies (Sato et al., 1984;
Inagaki et al., 1992) have correlated volume expansion with the size and
density of observed bubbles, which is consistent with previous
observation of bubble formation. Furthermore, this radiolytic process
apparently requires a high alkali content and the volume expansion
increases with alkali content (Figure 7). Concerns raised (Tosten, 1984;
Bibler et al., 1990) about whether radiolytic decomposition and bubble
formation actually occur under g-irradiation
emphasize the need for a more fundamental understanding of this process
in complex glasses. 4.7.
Mechanical Properties In
a recent study of a simulated nuclear waste glass doped with 238Pu and
244Cm, Inagaki et al. (1993) observed an exponential decrease in
hardness and Young's modulus with dose, while the fracture toughness
increased exponentially with dose. The maximum values of the relative
changes reported for the hardness, Young's modulus, and fracture
toughness in these glasses were -25%, -30%, and +45%, respectively. In
the same study, the changes in hardness and Young's modulus completely
recovered during annealing above 673 K with an activation energy of 0.28
eV; the recovery of the change in fracture toughness was initially more
rapid, with an activation energy of 0.21 eV, then the rate of recovery
slowed significantly, and complete recovery of the fracture toughness
was never observed. The changes in fracture toughness are strongly
correlated to the behavior of bubbles, which they observed previously in
these glasses (Inagaki et al., 1992). The bubbles, which can impede
crack propagation, anneal with kinetics similar to the recovery of the
fracture toughness. A
similar decrease in hardness (24%) was observed in an actinide-doped
glass by Bonniaud et al. (1979). In a study of a-irradiated
simulated waste glass (Weber and Matzke, 1987), the hardness decreased
15% and the fracture toughness increased 80% with dose (Figure 14). In
another study of
a-irradiated waste glass (Routbort
and Matzke, 1983), a 75% increase in fracture toughness was reported.
These large increases in fracture toughness may also be associated with
bubble formation; however, because the microstructural characterization
was not completed, the presence and role of bubbles cannot be confirmed.
g-irradiation of a simulated
nuclear waste glass to 6.6 x 108 rad indicates a slight decrease in
elastic modulus and a slight increase in hardness and fracture toughness
(Weber, 1988). In a study of a commercial borosilicate glass (Zdaniewski
et al., 1983), g-radiation did not appreciably
affect the strength or fracture toughness below 108 rad. The maximum
doses in these studies are low; consequently, the effect of g-radiation on mechanical properties is unknown
for g-doses equivalent to more than
one year of storage. 4.8.
Radionuclide Release Radiation
can affect the release rate of radionuclides from waste glasses by
increasing the surface area for radionuclide release (microfracturing)
and by changing the dissolution rate of the glass. The extent of the
radiation-induced microfracturing will depend on differential volume
changes, microstructure, and mechanical properties. Presently, there are
insufficient data and understanding currently available to predict the
extent of radiation-induced microfracturing in waste glasses, except to
note that it can be expected to occur, especially as a result of the
radiation-induced differential expansion of the crystalline phases in
the glass. The
dissolution rate of nuclear waste glasses may be affected by the
radiation-induced changes in chemistry, microstructure, and network
bonding. As noted in several reviews (Turcotte, 1981; Burns et al.,
1982a and 1982b; Weber and Roberts, 1983) and discussed in detail
recently (Weber, 1988 and 1991b), almost all data concerning the effects
of a-decay
on dissolution (leach) rates of waste glasses were determined prior to
1982 in short-term tests based solely on weight loss. The data obtained
from these tests (Figure 15) indicate no more than a factor of 3
increase in the short-term dissolution rates as a result of radiation
effects from a-decay. It is now well
established that short-term tests based on weight loss can underestimate
the radiation-induced increases in dissolution rate by a factor of 3 to
4 due to precipitation of alteration phases on the surface of the glass
(Westsik and Harvey, 1981; Weber et al., 1985a; Weber, 1988);
consequently, based on these limited data base, it has been suggested by
Weber (1991b) that radiation effects are not expected to increase the
nuclear waste glass dissolution rates by more than a factor of 10.
Although this may overestimate (or even underestimate) the actual
effects, it is the best assessment that can be drawn from the limited
data base and understanding that currently exists. This assessment is
supported by a recent study by Eyal and Ewing (1993) of a Th-doped
borosilicate glass, where the Th daughter product release rate was
increased 15 to 45% due to incongruent dissolution along the recoil
tracks produced by the
a-recoil nucleus. Ion-irradiation
(Manara et al., 1982; Dran et al., 1982; Primak, 1982) and
neutron-irradiation (Cousens and Myhra, 1983) studies of several HLW
glasses have shown irradiation-enhanced dissolution rates of up to a
factor of 4. Ion irradiation, however, is known to introduce
compositional variations (Arnold et al., 1982), which affect the
interpretation of the results. g-irradiation of several HLW glasses up to 1011
rad also leads to increases in dissolution rates of up to a factor of 4
(Grover, 1973; Bibler, 1982). The effect of radiolytic decomposition on
dissolution rates has not been investigated. In
summary, depending on the type of corrosion model used for waste glass
dissolution, the radiation-enhanced radionuclide release rates may be
expected to increase in proportion to the increase in surface area (due
to microfracturing). The actual release rate will depend on geochemical
conditions (e.g., presence or absence of silica-saturated solutions).
Additionally, in the absence of careful studies, a conservative factor
of approximately 10 is estimated for the enhanced dissolution rate due
to radiation-induced changes in the atomic structure of the glass (e.g.,
Eyal and Ewing, 1993). Any temperature dependence of the radiation
effects could further affect this enhancement. 5.
Radiation Effects in Crystalline Waste Forms 5.1.
Multiphase Ceramic Waste Forms
As alternatives to nuclear waste glasses, several multi-phase,
crystalline ceramic waste forms have been proposed that are tailored to
produce specific crystalline phases as hosts for the different
radionuclides. Generally, fission products (such as cesium and
strontium) are confined to one or more crystalline phases, and the
actinides (uranium, neptunium, plutonium, americium, and curium)
partition into other crystalline phases. Synroc and other related
titanate-based ceramic waste forms have received the most
attention(Ringwood et al., 1979 and 1981; Ringwood, 1982; Kesson et al.,
1983; Newkirk et al., 1982). There has been more limited study of
radiation effects in supercalcine, a silicate-based, tailored ceramic,
(McCarthy, 1977; McCarthy et al., 1979) and glass-ceramics (Dé et al.,
1976a and 1976b). There have been far fewer studies of radiation effects
in the bulk crystalline ceramic waste forms as compared to glass waste
forms. Details of the more extensive studies of radiation effects in
component phases are presented in Section 5.2. 5.1.1.
Synroc
Ringwood et al. (1980) have argued that Synroc is stable with
respect to a-decay
damage, based on an assessment of natural minerals of uranium and/or
thorium-bearing perovskite and zirconolite, which are the actinide host
phases of Synroc. Such data can provide only qualitative results, as
natural and actinide-doped zirconolites do become amorphous. The rate of
amorphization and subsequent changes in physical and chemical properties
may vary with modifications in composition. Therefore, quantitative
evaluation of Synroc as a waste form requires radiation-damage testing
of actual or simulated waste forms.
In the early 1980's, a limited amount of accelerated
irradiation-damage testing was performed on Synroc and its constituent
phases by Reeve et al. (1981) and Woolfrey et al. (1982) using fast
neutrons. Cold-pressed and sintered (cps) and hot pressed (hp) specimens
of Synroc B and Synroc C, as well as cps specimens of their constituent
minerals hollandite, perovskite, and zirconolite, were irradiated in the
fast-neutron flux of the HIFAR reactor at Lucas Heights, for different
lengths of time. (Synroc B does not incorporate simulated fission
products; Synroc C does.) The largest dose reported on the Synroc
specimens (10 wt.% waste loading) was 2.7 x 1026 n/m2 (E > 1 MeV),
which corresponds to 0.7 dpa or 8 x 1018
a-decays/g
and represents an equivalent Synroc storage time of 6.5 x 105 yr.
Volume, density, open porosity, and unit cell parameter changes were
measured. The volume expansion of Synroc specimens was generally larger
than that of the separately irradiated specimens of the constituent
phases. In addition, the results show that the volume expansion is much
larger for the Synroc-B (cps) than for the Synroc-B (hp). The volume
changes in Synroc-C are similar for both hp and cps specimens. The
macroscopic volume expansion results for Synroc-C(hp) are included in
Figure 16 (the neutron dose has been converted to an equivalent
a-decay dose). Based on these
data, volume expansion at saturation cannot be predicted. Accelerated
testing of Synroc-C and its constituent phases has been completed using
actinide-doping techniques at Harwell (Evans and Marples, 1985; Evans et
al., 1986). The radiation stability of hot-pressed Synroc specimens
doped with 2 and 5 wt.% 238Pu was investigated. X-ray diffraction
analysis indicated that the Pu-containing zirconolite and perovskite
phases were x-ray diffraction amorphous after a bulk average dose
corresponding to 2.8 x 1018 a-decay events/g (estimated to
be 0.37 dpa in the zirconolite and perovskite phases), which is
equivalent to a Synroc age of approximately 104 years for a 10 wt.%
waste loading. The macroscopic volume expansion (Figure 16) was
determined from the change in density (Weber, 1990). The swelling
exceeds 6% and shows only a slight indication of approaching a steady
state (saturation) value. The swelling of the Pu-doped Synroc-C occurs
at about twice the rate as the equivalent neutron-irradiated Synroc-C.
This may be due to the relatively simple assumptions made by Woolfrey et
al. (1982) in estimating the equivalent a-decay
dose to the measured neutron dose or may indicate that neutron
irradiation does not produce the same radiation damage effects. 5.1.2.
Titanate Ceramics
The Japanese have investigated a titanate ceramic consisting of
perovskite, zirconolite, hollandite, freudenbergite and loveringite for
encapsulation of sodium-rich high-level waste. This titanate ceramic was
doped with 0.69 wt% 244Cm, and the effects of
a-decay on density and
dissolution rate were investigated (Mitamura et al., 1990). The density
decreased linearly with dose. The density change was -1.0% after a dose
of 0.7 x 1018
a-decays/g, which was the
maximum dose reached. The release rates measured for Cm and soluble
non-radioactive elements (e.g., Na, Cs, Sr and Ca) showed a general
trend of increase with increasing dose. The maximum dose reached was
relatively low compared to those in other studies, corresponding to a
storage time of 5000 years.
Sandia National Laboratories investigated a titanate ceramic to
immobilize acidic high-level nuclear waste. This titanate ceramic is a
multi-phase assemblage that consists of rutile (TiO2), an amorphous
silicate phase, perovskite, zirconolite, hollandite, and a U-Zr-rich
phase believed to have a pyrochlore structure (Dosch et al., 1984).
Radiation effects from
a-decay
have been simulated by irradiating TEM specimens with Pb+ ions (Dosch et
al., 1985). Multiple energies (40 to 250 keV) were used to produce a
uniform damage profile. Irradiation at single energies of 240 to 250 keV
were also performed. Examination by transmission electron microscopy of
the irradiated specimens indicated considerable damage in all phases
after a dose equivalent to 1 x 1019 a-decay
events/cm3. All phases remained crystalline except a phase with a
chemistry typical of a pyrochlore, which became amorphous. 5.1.3.
Supercalcine
The only radiation damage study that has been performed on a
supercalcine formulation (SPC-2) was carried out at the Pacific
Northwest Laboratory (Rusin et al., 1979; Turcotte et al., 1982). The
SPC-2 formulation was doped with 2 wt% 244Cm. Stored energy, x-ray
diffraction analysis, and density measurements were made initially and
at 3-month intervals, reaching a total bulk dose of 1.2 x 1018
a-decay events/g. Initial
characterization showed the supercalcine consisted of three major
phases: fluorite, apatite, and a tetragonal phase. Subsequent analysis
suggested that curium predominately partitioned into the apatite and the
tetragonal phase.
Analysis by x-ray diffraction of the Cm-doped supercalcine, as a
function of dose, showed a gradual transformation of the apatite from
the crystalline to amorphous state, in agreement with the results of
Weber et al. (1979) on the amorphization of apatite crystallites in a
nuclear waste glass. An expansion in the unit cell of apatite was also
observed with increasing dose; the increase in unit cell volume
saturated at 2.5%. There was a slight increase in the intensity of the
diffraction maxima associated with the tetragonal phase, but changes in
the tetragonal unit cell could not be determined. The fluorite peak was
used as a standard, as this phase did not appear to contain Cm and was
unaffected by external radiation.
Stored energy and density measurements were also made as a
function of dose for the Cm-doped supercalcine. The stored energy
reached a maximum value of 42 J/g at a dose of 0.5 x 1018
a-decay
events/g and decreased slightly with further increases in dose. Energy
release was not complete at 600°C (the upper limit of the calorimeter
used); therefore, additional energy release may occur at higher
temperatures (> 600°C). The density decreased exponentially with
dose, resulting in a volume expansion of 1.4% at a dose of 1.2 x 1018 a-decay
events/g. The macroscopic swelling as a function of dose in SPC-2 is
shown in Figure 17. The volume expansion at saturation (1.4%) is
partially attributed to the volume expansion in the apatite as a result
of radiation-induced amorphization. 5.1.4.
Glass Ceramics
A glass ceramic is a fine-grained mixture of glass and ceramic
phase ideally derived from a homogeneous glass by heat-treating a glass
precursor at the temperature of maximum nucleation rate for the ceramic
phases(s), followed by a high-temperature treatment to yield a maximum
growth rate. The celsian glass ceramic, developed at the Hahn-Meitner
Institut and named for the predominant crystalline phase, is easy to
fabricate, is homogeneously crystallized, and contains a variety of
leach-resistant host phases for the radionuclides. Celsian glass ceramic
as a waste form has been studied for potential radiation effects both at
the Pacific Northwest Laboratory (Turcotte et al., 1982), in cooperation
with the Hahn-Meitner Institut, and as part of a collaborative research
program within the European Community (Malow et al., 1980). In the study
at PNL, the effects of radiation damage in 244Cm-doped celsian glass
ceramic were investigated. The density was observed to decrease
exponentially to saturation values at a dose of 0.9 x 1018
a-decay events/g. As shown in
Figure 17, the volume expansion at saturation was 0.5%. Stored energy
also increased exponentially with dose to a value of 80 J/g. Data from
the x-ray diffraction analysis revealed that a Cm-rich, rare-earth
titanate phase with the pyrochlore structure underwent a small volume
expansion with dose and eventually became x-ray diffraction amorphous at
the same saturation dose of 0.9 x 1018
a-decay events/g. In the study
of Malow et al. (1980), the celsian glass ceramic was doped with 2.5 wt%
238Pu. The volume also expanded exponentially with dose, as shown in
Figure 17. Saturation is predicted at a dose of 1.6 x 1018 a-decay
events/g, with a corresponding volume expansion of 0.5%, in good
agreement with the PNL results. The stored energy was measured to be 43 +5
J/g after a dose of 1.1x 1019
a-decay
events/g. This is approximately the value determined at PNL, but it is
not known whether saturation was reached in this material. Changes in
leach rates were measured for specimens stored at 23 and 170oC to a
cumulative dose of 1.1 x 1018
a-decay events/g. The leach
rate decreased by 4% in the specimen stored at 170oC. The fractional
helium release was also determined to be 3% after storage at 170o C to a
cumulative dose of 1.1 x 1018
a-decay
events/g. 5.2.
Crystalline Phases
There have been a number of studies of individual synthetic
phases and minerals that are structurally and/or chemically analogous to
phases found in multiphase ceramic waste forms. These studies contribute
significantly to the understanding of radiation effects in the
multiphase ceramic waste forms by providing a detailed understanding of
the behavior of each component phase. The effects of radiation on the
most important phases are summarized. Two of the phases to be discussed,
pyrochlore and apatite, have also been observed as crystalline phases in
nuclear waste glasses. 5.2.1.
Pyrochlores
Pyrochlore (Fd3m, Z = 8), VIIIA2VIB2IVX6IVY
is a derivative of the fluorite structure type (Chakoumakos 1984, 1986)
in which the A-site contains large cations
(Na, Ca, U, Th, Y and lanthanides) and the B-site consists of smaller,
higher valence cations (Nb,Ta, Ti, Zr, Fe3+). The essential feature of
the structure is sheets of corner-sharing, BX6 octahedra parallel to the
(111) plane which are arranged into three- and six-membered rings (the
hexagonal tungsten bronze structure). Actinides may be accommodated in
the A-site,
and charge balance is maintained by cation deficiencies in the
A-site and substitutions on
the B-site. Rare-earth titanates with the pyrochlore structure have been
observed as actinide-host phases in nuclear waste glasses (Weber et al.,
1979), in titanate ceramic waste forms (Dosch et al., 1984), and in a
glass ceramic waste form (Malow et al., 1980; Turcotte et al., 1982). In
addition, the rare-earth titanates with the pyrochlore structure are
chemically and structurally related to zirconolite, an important
actinide host phase in Synroc and titanate ceramic waste forms. In
Cm-doped nuclear waste glass (Weber and Roberts, 1983), crystallites of
the pyrochlore (Gd,Cm)2Ti2O7 transformed from a crystalline to an
amorphous state as a result of
a-decay
of the incorporated Cm. The volume expansion associated with the
amorphous transformation of this pyrochlore phase and an apatite phase
resulted in microfracturing of the simulated waste glass. Similarly, in
a study of a Cm-doped celsian glass ceramic (Turcotte et al., 1982), the
originally crystalline pyrochlore phase,
(Nd0.85Cm0.15)2(Ti1.65Zr0.35)O7, transformed to the amorphous state due
to self-radiation damage from a-decay.
The unit cell volume was estimated to expand 0.8% before amorphization
was complete. Dosch et al. (1985) reported that the pyrochlore phase in
a titanate ceramic waste transformed to an amorphous state as a result
irradiation with Pb+ ions. More specific details of the amorphization
process could not be obtained from the study of these multi-phase waste
forms.
The effect of
a-decay on the pyrochlore phase
Gd2Ti2O7 doped with 244Cm has been extensively investigated (Weber et
al., 1985b and 1986). The decrease in diffracted x-ray intensity with
increasing dose is shown in Figure 18 for Cm-doped Gd2Ti2O7 (Weber et
al., 1986) and for minerals of the pyrochlore group (Lumpkin and Ewing,
1988). The progression of the amorphization process in Cm-doped Gd2Ti2O7
was also followed by transmission electron microscopy and selected area
electron diffraction (Figure 19). At a low dose of 0.6 x 1018 a-decays/g
the material exhibits a strongly crystalline electron diffraction
pattern and clearly resolvable individual defects and fission tracks. At
increasing dose levels, the density of radiation-induced defects
increases to the point that individual defect clusters are no longer
readily distinguished and a radial ring of diffuse intensity associated
with the presence of amorphous material is observed in the electron
diffraction pattern. At higher doses, the material is nearly fully
amorphous with some residual crystallinity evident in the electron
diffraction pattern. At a dose of 3.1 x 1018 a-decay events/g, the material became fully
amorphous with no evidence of any residual crystallinity. The
macroscopic swelling, DVm/Vo,
of Cm-doped Gd2Ti2O7 is shown in Figure 20 as a function of dose, D, and
follows an exponential relationship given by the expression:
DVm/Vo
= Am [1 - exp(-BmD)]
(2) where
Am is the swelling at saturation, and Bm is the amount of material
damaged per a-decay event. The values of Am
and Bm have been determined to be 5.1% and 6.8 x 10-19g, respectively
(Weber et al., 1985b and 1986). Wald
and Weber (1984) studied the radiation-induced changes in dissolution
kinetics of Cm-doped Gd2Ti2O7 by testing fully-damaged (amorphized
specimens and a second set of specimens that had been fully
recrystallized to the original structure by a 12 hour anneal at 1,100°C).
Specimens were immersed in distilled, deionized water at 90°C for 14
days. Weight loss measurements indicated an increase in dissolution rate
of a factor of 2.5 as a result of the radiation-induced swelling and
amorphization. More accurate radiochemical analysis indicated a
radiation-induced increase of a factor of 20 to 50 in the dissolution
rates of the non-network (TiO6 octahedra form the network) Cm and Pu.
The concentration of Ti in solution was below the detection limit of the
inductively coupled plasma emission spectroscopy system, suggesting some
dissolution resistance of the titania network. Analysis for non-network
Gd was not performed. These results suggest that the dissolution may
occur incongruently, selectively leaching the non-network ions. Similar
behavior is observed for zirconolite (Wald and Weber, 1984). The
increase in dissolution rate due to amorphization is probably due to the
increase in the size of the network tunnels as the structure disorders
and expands.
The fracture toughness in Cm-doped Gd2Ti2O7 increased with
cumulative dose to a broad maximum and then decreased slightly (Figure
21). This is similar to the behavior of zirconolite (Clinard et al.,
1985a; Weber et al., 1986) and apatite (Weber and Matzke, 1986b). The
increase in fracture toughness is attributed to the composite nature of
the microstructure. At low to intermediate doses, the microstructure
consists of amorphous tracks in a crystalline matrix, and this composite
microstructure can inhibit crack propagation and increase the fracture
toughness. As the amorphous phase becomes the dominant matrix at high
doses with remnant crystallites, the fracture toughness decreases
slightly as some of the internal stresses are relieved. This is
supported by the observations of zirconolite that suggest a relaxation
of disorder at high doses (Foltyn et al., 1985) and the analysis of
strain accumulation in natural pyrochlores (Lumpkin and Ewing, 1988).
Isochronal (12 h) annealing of fully amorphous Cm-doped Gd2Ti2O7
shows a linear recovery of density with temperature up to the
temperature where recrystallization begins (700°C). Recrystallization
results in a sharp recovery of the density with a peak in the recovery
rate at 750°C. Full recovery of the density and recrystallization of
the original pyrochlore structure are essentially complete at 850°C.
Thermal recovery studies of natural minerals of the pyrochlore group
indicate that recrystallization is an exothermic reaction that peaks in
the temperature range from 650 to 700°C and releases 120 to 200 J/g of
stored energy (Lumpkin et al., 1986). The temperature range is in
reasonable agreement with the isochronal recovery behavior of Cm-doped
Gd2Ti2O7, and the stored energy released is similar in magnitude to that
released during recrystallization of zirconolite (Weber et al., 1986;
Foltyn et al., 1985) and apatite (Weber, 1983). Systematic
studies (Lumpkin and Ewing, 1985; Lumpkin and Ewing 1986; Lumpkin et
al., 1986; Lumpkin and Ewing, 1988; Lumpkin et al., 1988) have been
completed of a-decay event damage of
natural, isometric pyrochlore structure types. Minerals of the
pyrochlore group are among the common rare element accessory minerals
occurring in a wide range of igneous rocks: carbonatites, nepheline
syenite pegmatites and granitic pegmatites. Actinides may be
accommodated in the A-site,
and in natural pyrochlores the substitution of uranium can be extensive
(up to 30 wt. %). Charge balance is maintained by cation deficiencies in
the A-site
and substitutions on the B-site (typically, Nb, Ta and Ti). Natural
pyrochlores are classified on the basis of their major B-site cations:
microlite (Ta>Nb, Ti< 33 mole percent); pyrochlore (Nb >
Ta, Ti < 33 mole percent); betafite (Ti > 33 mole percent)
(Hogarth, 1977). Most of the detailed work on radiation damage effects
has been completed on microlite, in which the compositions may reach
values as high as 30 wt % UO2 and 5 wt.% ThO2. Depending on the age of
the sample, the total calculated dpa may be as high as 80. Careful
analysis of samples from the same deposit, e.g., the Harding Pegmatite
in New Mexico which is 1300 million years old, allows one to study
radiation damage over a range of dpa (0.2 to 49) as function of the
uranium concentration (0.05 to 8.6 wt. % UO2). This provides a good
analogue for long term radiation damage effects in an actinide-bearing,
crystalline waste form. In the remainder of this section, we summarize
the results of such a study in which the periodic-to-aperiodic
transition was characterized by x-ray diffraction (XRD) analysis, high
resolution transmission electron microscopy (HRTEM) and extended x-ray
absorption spectroscopy (EXAFS) and x-ray absorption near edge
spectroscopy (XANES) as a function of increasing
a-decay event dose. Figure
22 illustrates the change in the x-ray diffraction pattern of microlites
as a function of increasing
a-decay
event dose. With increasing dose (#147, dpa = 0.2; #153, dpa = 49), the
intensity of the diffraction maxima decreases and the peaks broaden, but
not as quickly as in the actinide-doped material (Figure 18). The
positions of the diffraction maxima do not shift to lower values of 2Q (as in the case of zircon
when the unit cell volume increases) because, although unit cell
expansion with increasing dose and the accumulation of defects is
commonly observed, the increasing
a-decay
dose in natural pyrochlore is the result of the substitution of U for Ca
and Na. The contraction of the unit cell due to the chemical
substitution evidently compensates for the expansion of the unit cell
due to radiation damage effects. Volume expansion has been well
documented in pyrochlores (Wald and Offerman, 1982) in which the
composition remains constant. Debye-Scherrer patterns of microlite
single crystals show, with increasing dose, loss of Ka1 and Ka2 peak splitting (0.5 - 1.0 x 1019
a-decay events/g), loss of back
reflections (1 - 5 X 1019
a-decay
events/g), and progressive loss of low angle peaks until none are
discernible (5 - 15 x 1019
a-decay
events/g). The Bragg diffraction maxima remain essentially symmetric
throughout the transition from the crystalline to the metamict state. A
detailed analysis (Karioris et al., 1982) of the decrease in Bragg
intensity can be used to estimate the amount of material damaged by each a-recoil
(Lumpkin and Ewing, 1988). The calculated spherical diameter of an
a-recoil track was estimated to
be 4.6 nm, in good agreement with the value of 5.4 nm estimated by
Clinard et al. (1985b) for Pu-doped zirconolite. The
x-ray line broadening was analyzed using the technique of Williamson and
Hall (1953), which provides information on crystallite size and strain.
The maximum strain reached is 0.003, halfway through the
crystalline-to-metamict transition. Average crystallite dimensions
decrease from approximately 450 nm to less than 15 nm just prior to the
microlite becoming x-ray diffraction amorphous. This suggests that the
early stage of damage accumulation is characterized by unit cell
expansion and strain caused by isolated defects (predominantly due to
a-particle damage). During the
second stage, strain is reduced as the overlap of aperiodic regions (a-recoil tracks) increases.
With the increase in the volume of the aperiodic domains, there is a
reduction in the size of the crystallites that remain in the still
crystalline regions. The slight rotation of these crystallites can
contribute to the reduction in strain. The
results of the high resolution transmission electron microscopy confirm
the interpretation of the x-ray diffraction data (Figure 23). High
resolution TEM images of highly crystalline samples recorded in the
[110] orientation show variations in contrast resulting from the B2X6
framework and channels parallel to [110]. A typical image is shown in
Figures 23-a and 23-b.. Most of the images (e.g., 23.c - 23.g) were
taken in the (111) orientation, exhibiting 0.6 nm lattice fringes which
probably result from the hexagonal tungsten bronze layer stacking
parallel to (111). Some images also display 0.3 nm (222) lattice
fringes. Effects of increasing U-content on the microstructure are
illustrated by the series of HRTEM images in Figure 23. The U-contents
have been converted to a corrected dose assuming the the mean life of an
a-recoil track is on the order
of 100 million years. The U-content and equivalent corrected dose (a-decay events/mg) for each
image are given in the caption of Figure 23. Three stages of damage are
evident: 1.) At an early stage of damage (< 1018 a-decay
events/g), lattice fringes are essentially continuous, but small
variations in diffraction contrast suggest the presence of isolated
defects. 2.) At an intermediate stage of damage (1018 - 1019 a-decay
events/g), lattice-fringe free areas indicate the increasing presence of
aperiodic domains of
a-recoil
damage tracks. As the dose increases the proportion of aperiodic domains
increases, and "islands" of crystallinity decrease in size as
the tracks increase in number and begin to overlap. 3.) At the highest
doses (> 1019 a-decay
events/g), no lattice fringes are apparent and the material is
electron-diffraction amorphous. Electron diffraction maxima are absent,
having been replaced by a diffuse, x-ray scattering halo. Once
long range periodicity is lost, XRD and HRTEM techniques provide little
insight into the results of further damage; however, spectroscopic
techniques provide a powerful means of examining changes in the first
and second coordination sphere geometries of the aperiodic material. In
a series of studies, Greegor and colleagues (1984a, 1984b, 1985a, 1985b,
1986, 1987) applied EXAFS and XANES analysis to partially and
fully-metamict members of AxByOz oxides (e.g., AB2O6: euxenite,
polycrase, priorite and blomstrandine; A2B2O7: zirconolite; A1-2B2O6Y0-1
nH2O; pyrochlore, microlite and betafite). Based on these results, which
are summarized in Ewing et al. (1988), a good model of the structure of
the metamict state is one in which there is: (1) a slight distortion of
the nearest-neighbor coordination polyhedra; (2) a decrease in the
coordination number and bond length in the primary coordination
polyhedra; (3) a slight increase in the mean second nearest neighbor
distances; (4) a loss of second nearest neighbor periodicity, probably
due to the rotation of coordination polyhedra across shared corners and
shared-edges. The structure of Ta2O5, with its large cell volume and
complex structure, is intermediate between the highly symmetric periodic
pyrochlore structure and the aperiodic, metamict state; thus serving as
a model for the aperiodic atomic arrangement of fully-damaged pyrochlore
(Greegor et al., 1987).
Lumpkin and Ewing (1988) have noted that if suites of pyrochlores
(of different ages) are analyzed from single localities, the calculated
dose for the crystalline-to-metamict transition increases with the age
of the geologic deposit (Figure 24). This is clear evidence for the
annealing of a-recoil
damage over geologic time. The relationship between critical
amorphization dose and geologic age was modelled assuming
"fading" of the
a-recoil "tracks" in
a fashion similar to that used to describe fission fragment track
fading. Critical parameters are the
a-decay
event dose, the age of the sample, the thermal history, and the mean
life of an a-recoil track. The variation
of dose as a function of age for natural pyrochlores can then be
corrected by the consideration of recoil track annealing. Annealing
under ambient conditions over long periods results in a higher apparent
dose being required for amorphization. The actual dose vs. time curves
(Figure 24) can be modelled by an exponential equation which considers
the annealing to be described by the half-life of a recoil-track, ta,
in the material (Lumpkin and Ewing, 1988). Figure 25 illustrates such a
correction when the mean life of a track, ta
, is assumed to be 100 million years. A similar analysis of data for
zircon and zirconolite gave mean "track" lives of 400 million
years and 700 million years, respectively. This is in contrast to the
mean lives determined by Eyal and co-workers using a differential
etching technique to determine the mean life of damaged areas in
monazite, UO2, metamict betafite, thorite, and metamict samarskite (Eyal
and Kaufman, 1982; Eyal and Fleischer, 1985; Eyal et al., 1985; Eyal et
al., 1987; Lumpkin et al., 1988). For phases which retain their
crystallinity despite
high
a-decay event doses (> 10
dpa), the mean life of an
a-recoil
track was on the order of 104 years. Thus, the natural samples provide
clear evidence for annealing over geologic time under ambient
conditions. The final state, crystalline or metamict, of the mineral
depends on the mean life of the
a-recoil track, a material
dependent property. One
of the important issues with any waste form is the amount of stored
energy associated with the radiation damage. For natural pyrochlores
thermal recrystallization effects (measurement of the heat of
recrystallization and identification of phases formed) have been
determined on metamict members of the pyrochlore group which have
received a-decay
event doses of up to 4 x 1020
a-decay
events/g. The heats of recrystallization range from 125 to 210 J/g. The
temperature of the exotherm of recrystallization varies between 400 and
700°C; the exact temperature is determined by the sample composition
(i.e. 400-600°C for pyrochlores and microlites; 650-700°C for
betafites). Release of energy decreases as a function of the
crystallinity (estimated on the basis of the intensity of X-ray
diffraction maxima), with the fully-metamict samples approaching 20 J/g.
Lower measured values (40-125 J/g) are the result of alteration of the
pyrochlores. Other metamict, complex oxides with stoichiometries of ABO4
and AB2O6 have lower heats of recrystallization (40-85 J/g) and are
easily distinguished from pyrochlore group minerals. The
recrystallization energies for the pyrochlore group minerals are
consistently higher than those measured in synthetic, Pu-doped
zirconolite (50 J/g at doses of 3 x 1016 a-decay events/mg = 15 x 1026 a-decay events/m3), synthetic Cm-doped zirconolite
(127 J/g) at doses of 5 x 1018
a-decay
events/m3, and natural zirconolite (40-50 J/g at doses of 2 x 1016
a-decay events/mg = 1.0 x 1026
a-decay events/m3). Activation
energies of recrystallization range between values of 0.29 to 0.97 eV,
less than those measured for Pu-doped zirconolites (Foltyn et al., 1985;
Weber et al., 1986). 5.2.2.
Zirconolite
The waste form phase zirconolite (monoclinic CaZrTi2O7) is one of
the three principal phases of Synroc, and zirconolite is one of the most
extensively studied phases, as it is the primary actinide host.
Monoclinic zirconolite is a fluorite-derivative structure closely
related to pyrochlore. The (001) layers of the corner-sharing
arrangement of TiO6 octahedra are similar to the (111) layers in
pyrochlore and the (001) layers in the hexagonal tungsten bronzes. The
A-site
cations (Ca, Zr and actinides) are "sandwiched" between the
(001) sheets of B-site cations (corner-sharing Ti06 octahedra). There
can be several polytypes depending on the stacking arrangements of the
sheets. In this section, we discuss the damage response as investigated
by study of synthetic compositions doped with the short half-life
actinides, 238Pu (87.7 y,
a-decay to 234U) and 244Cm
(18.1 y, a-decay to 240Pu), natural
zirconolites damaged by
a-decay
of uranium and thorium, and finally ion-beam irradiated zirconolite. 5.2.2.1.
238Pu-substituted zirconolite A
zirconolite (CaZrTi2O7) composition can be synthesized with the
substitution of PuO2 for ZrO2. X-ray absorption spectroscopy
demonstrated that the 238Pu was incorporated into the
a-site
(Greegor et al., 1988). The effect of this chemical substitution is to
convert the zirconolite composition to an isometric, pyrochlore
derivative structure (Clinard et al., 1984). A major change in CaPuTi2O7
with increasing dose is the large decrease in density (Figure 26).
Storage at ambient temperature (approximately 350 K, due to
self-heating) resulted in significant swelling, up to a saturation value
of approximately 5.5 vol. % in material that had reached an a-decay
dose of 2-3 x 1025 a-decay
events/m3 (for 200 days of storage time). Storage at 575 K resulted in
less swelling, which required a longer time (300 days) to reach
saturation. When this material was held at 875 K, swelling was minimal.
Transmission electron microscopy conducted after various storage
times and temperatures (Clinard et al., 1984) showed that with
sufficient a-decay
damage, Pu-zirconolite at 350 and 575 K converts from the crystalline to
the aperiodic, metamict state. Shortly after the test material was
fabricated, examination by transmission electron microscopy of the the
Pu-zirconolite at ambient temperated showed a typical crystalline
diffraction pattern (Figure 27). At intermediate states of damage, the
microstructure consisted of a mixture of crystalline and aperiodic
domains (Figure 28). The observed microstructural heterogeneities are
consistent with a random distribution of spherical damage tracks 5 nm in
diameter, a microstructure that can be predicted from the shape of the
corresponding swelling curve (Clinard, 1986). At high damage levels,
electron diffraction patterns (Figure 29) show full conversion from
crystalline materials with sharp diffraction maxima to an inner
correlation ring characteristic of the nearest-neighbor interatomic
spacing in an aperiodic structure.
Microhardness indentation studies indicated that as damage
accumulated the zirconolite became softer (Figure 33), yet fracture
toughness increased (Figure 34) (Clinard et al., 1985a). The first of
these changes is consistent with conversion to the metamict state, while
the latter was interpreted as resulting from the toughening effect of
the two-phase microstructure (i.e., mixed periodic and aperiodic
domains).
When the Pu-zirconolite which had been damaged at ambient
temperature was heated to 875 K, the result was densification up to a
value of 4 vol.% over a period of 400 days. Transmission electron
microscopy confirmed that this recovery corresponded to
recrystallization from the metamict state (Clinard et al., 1985b).
Differential scanning calorimetry measurements (Foltyn et al., 1985)
showed that as expected, recovery of metamict material to the
crystalline state was an exothermic process. The amount of energy
released initially increased with damage dose, but then peaked and
subsequently decreased to about half the maximum value (Figure 30). This
suggests that redamage of previously-damaged material results in a
reordering of the damage state to one of lower energy.
Modeling calculations (Figure 31) were completed to determine the
fraction of material that was undamaged, damaged by a single-track, and
damaged by multiple overlap of tracks as a function of irradiation dose
(Coghlan and Clinard, 1988). Comparison of Figure 31 with Figure 26
shows that swelling occurs under a constantly-changing damage condition,
not just with respect to the proportion of damaged and undamaged
material, but also with respect to the extent of redamaged material.
A comparison of stored energy and its corresponding damage state
at different damage levels is shown in Table 2 (Coghlan and Clinard,
1988). The greatest amount of stored energy, 100 J/g, was measured in
material that contained the largest fraction of single-track damaged
material. At higher doses stored energy decreased (Table 2), although
both the amount of redamaged material and the total amount damaged have
increased. From these results, Coghlan and Clinard (1988) determined
that the stored energy for the single track damaged condition is on the
order of 200 J/g. Probable sources of the energy which is released on
annealing of irradiation-damaged CaPuTi2O7 are: (1) the atomic disorder
that is induced, and (2) the accompanying strain that results from
formation of aperiodic domains. An estimate was made of the maximum
contribution from strains, 18%, (Klemens et al., 1987); therefore, by
difference, the decrease in stored energy can be primarily attributed to
atomic-scale rearrangement of the aperiodic structure. The width of the
inner correlation ring observed by TEM is another indication of the
extent of disorder. Measurements of CaPuTi2O7 (Clinard, 1986) showed
that this parameter varied with history of the test sample (Figure 32).
The greatest width of the inner correlation ring (diffuse scattering
haloe) resulted from material that was held for 460 days at 350K. The
extent of disorder decreased with irradiation; thus, it is not
surprising that holding the material for 957 days caused a narrowing or
sharpening of the correlation ring. Further sharpening was observed when
the material was stored at 575 K for 430 days, or reheated to 875 K for
a few days after being self-irradiated to the fully-redamaged aperiodic
structure (625 days at 350 K). Thus, measurements of the inner
correlation ring are consistent with the understanding of damage and
redamage processes obtained by other techniques.
EXAFS/XANES studies completed on 238Pu-substituted zirconolite
held at 350 K for at least 500 days helped to identify details of the
damage state (Greegor et al., 1988). For comparison, measurements were
made on essentially undamaged material (that is, substituted with 239Pu
with a half-life of 24,100 years as compared to that of 238Pu of 87.7
years). In the heavily-damaged material, the Pu/O nearest-neighbor bond
length contracted by 0.003 nm, while expansion was observed for the more
distant coordinating atoms, corresponding to a 6 to 7% overall volume
expansion. This observation is close to the measured volume swelling of
5.5 % (Fig 26). 5.2.2.2.
244Cm-doped zirconolite
Self-damage studies of zirconolite have also been carried out by
doping with 244Cm, an
a-active isotope with a
half-life of 18.1 y. Weber et al. (1986) doped CaZrTi2O7 with 3 wt.%
244Cm and assessed changes in swelling, microstructure, hardness,
fracture toughness, leachability, and recrystallization behavior. The
Cm-doped zirconolite retained the monoclinic structure; whereas, the
238Pu-doped zirconolite was isometric (although up to concentrations of
5 mole % PuO2, the monoclinic, zirconolite structure was retained (Clinard
et al., 1984). More importantly, the composition of the Cm-doped
zirconolite was not greatly changed by the substitution; whereas, the Pu-substituted
structure was significantly changed by replacement of ZrO2 by PuO2.
Swelling data at 350 K were obtained to near-saturation, and by
use of an exponential relationship, a saturation swelling value of 6.0
vol.% was calculated for a dose of 2 x 1025
a-decays events/m3. These
results are in good agreement with those for the Pu-substituted
material, which exhibited saturation swelling of 5.5 vol.% at nearly the
same damage level (Clinard et al., 1984). TEM images and diffraction
patterns confirmed that Cm-doped zirconolite amorphized by the
accumulation of damage tracks produced by the 240Pu recoil ions that are
emitted during a-decay
of 244Cm. A smaller amount of additional damage resulted from
spontaneous fission of 244Cm, which introduced a low density of fission
tracks.
Changes in hardness and fracture toughness as a function of
increasing a-decay
event dose are shown in Figs. 33 and 34 combined with the results for Pu-doped
zirconolite. Both materials showed a decrease in hardness and an
increase in fracture toughness; the former change is expected as a
result of the radiation-induced transformation from a periodic to an
aperiodic state, while the latter may result from the heterogeneous
microstructure (Clinard et al., 1985a).
Differential thermal analysis of the Cm-doped zirconolite after
self-irradiation to 2.5 x 1025
a-decays events/m3 showed a
strong exothermic reaction during recrystallization, with a sharp
recovery stage of 114 J/g at 680-730° C. These values of stored energy
and recovery temperature agree closely with those obtained from Pu-substituted
zirconolite, although the latter material showed a halving of stored
energy at higher radiation doses as a result of damage beyond the dose
required for amorphization (Foltyn et al., 1985). Measurements on the
Cm-doped material showed that the temperature of the energy release peak
increased with heating rate; this behavior allowed the calculation of an
activation energy of recrystallation of 5.8 eV.
The dissolution (leaching) characteristics of the Cm-doped
zirconolite were determined (Weber et al, 1986) by testing fully-damaged
(amorphized) specimens and a second set of specimens that had been fully
recrystallized to the original structure by a 12 h anneal at 1,100°C.
Samples were immersed in distilled, deionized water at 90°C for 14 d.
The results, shown in Table 3, indicate a notable increase in leach rate
with amorphization; especially significant is the factor of 10 increase
for the a-decay product Pu. The high
final pH is believed to limit the solubility of the Cm to the value
shown. The low concentrations of Ti as compared to that of Ca suggests
that the dissolution is incongruent, with the Ti ions (which are part of
the TiO6 octahedra comprising the network structure) being more
resistant to leaching than the interlayer cations (e.g., Pu and Ca). The
mechanism for the enhanced leaching of the radiation-induced amorphous
state may be associated with the broken bonds and the increased free
volume (a more open framework) of the amorphous state. 5.2.2.3
Ion beam-irradiated zirconolite Radiation
damage effects in zirconolite have also been investigated by bombardment
with energetic heavy ions (Weber et al., 1992). TEM of samples
irradiated with 3 MeV Ar+ ions showed damage buildup similar to that
observed in material damaged by
a-decay of 238Pu or 244Cm,
namely, accumulation of small defects (alpha-recoil tracks) to the point
where the material became globally disordered. Recent studies (White et
al., 1994) of the temperature dependence of amorphization in zirconolite
have shown an increase in amorphization dose with temperature, in
general agreement with the results of Clinard et al. ()1984) for
238Pu-substituted zirconolite. Natural
zirconolites have also been used in ion-beam irradiation simulations of
a-decay event-induced
amorphization (Ewing and Wang, 1992). Recrystallized samples (1,130°C
for 8 hours in air) were irradiated with 1.5 MeV Kr+ ions using the HVEM-Tandem
Facility at Argonne National laboratory. The Kr+ dose rate used during
the irradiation was 3.4 X 1011 ions/cm2s, which is a damage rate 2 x
1012 times higher than that which has occurred in the natural
zirconolite due to decay of uranium. In-situ transmission electron
microscopy was completed during the ion irradiation. After a Kr+ dose of
4 X 1014 ions/cm2( = 0.3 dpa), the zirconolite grains were amorphized.
The dpa level for ion beam-induced amorphization is six times lower than
that calculated for the natural a-decay
induced process. Some of the a-decay
event damage may have recovered by point defect migration over geologic
periods of time. It is interesting to note that inclusions of thorianite,
ThO2, in the zirconolite did not become amorphous at equivalent doses.
This is consistent with the already mentioned apparent radiation
resistance of UO2. The resistance of thorianite to amorphization
compared to zirconolite is consistent with a model of damage
accumulation in insulators in which the critical dose for amorphization
is inversely proportional to the topologic and chemical complexity of
the phase (Wang et al., 1991). Thoriainite is isometric with only one
unique cation site. Zirconolite is monoclinic with three unique cation
sites. Such criteria may be useful in evaluating the response of waste
form ceramics to
a-decay
event damage. Dose-dependence
of damage in zirconolite has been studied by Headley et al. (1982) using
Pb+ ions of 40 to 240 keV at ambient temperature. Damage proceeded in
four stages: isolated tracks, track overlap, domination by aperiodic
domains, and finally complete amorphization. Results were compared with
those for natural zirconolite self-irradiated to various damage levels
by a-decay over geologic times.
Both the nature and extent of damage were similar, regardless of the
large difference in damage rates. Later, Clinard et al. (1984) extended
the comparison to include 238Pu-substituted zirconolite and showed that
damage at the intermediate rate characteristic of zirconolite was
similar to that observed in natural and Pb+ ion-bombarded material. 5.2.2.4.
Natural zirconolite Although
zirconolite occurrences are rare, selected specimens have been the
subject of extensive studies (Ewing et al., 1982a; Ewing and Headley,
1983; Ewing, 1983; Oversby and Ringwood, 1981; Sinclair and Ringwood,
1981; Sinclair and Eggleton, 1982). Samples from Sri Lanka (550 million
years old) typically can have concentrations of UO2 of 2 weight percent
and of ThO2 of 20 weight percent, thus reaching doses of 1016 a-decay
events/mg (=1026
a-decay
events/m3). This is a dose equivalent to nearly two displacements per
atom. X-ray
diffraction analysis of fully-metamict zirconolite shows no diffraction
maxima confirming that zirconolite is X-ray diffraction amorphous.
Because of the paucity of samples which span a range of a-decay
event doses, it is not possible to follow the change of diffraction
maxima intensity and position with increasing a-decay
event dose. A zirconolite annealed at 1130°C did give a diffraction
pattern with sharp peaks which could be refined on the basis of a
monoclinic unit cell (Lumpkin et al., 1986b). Referring to the work of
Wald and Offerman (1982) on a 244Cm-doped zirconolite, the expansion of
the unit cell was anisotropic with the greatest expansion perpendicular
to the sheets of TiO6 octahedra (Figure 35) reaching a maximum value of
2.1% volume expansion at a dose of 1023
a-decay events/m3. High
resolution transmission electron microscopy showed no evidence for
crystallinity (Lumpkin et al., 1986b), and in approximately ten percent
of the grains examined, there were subspherical microvoids ranging in
diameter from 100 to 400 nm (Ewing and Headley, 1983). The microvoids
are attributed to He-gas accumulation which is a result of the
radioactive decay. Annealed samples (1,100°C) recrystallized producing
crystals up to 2,000 nm in length. The crystallites were often twinned
at the unit cell scale as a result of different stacking sequences of
the TiO6 sheets, thus resulting in laminar intergrowths of the various
zirconolite polytypes (Lumpkin et al., 1986b). The same texture of
fine-scale twinning has been observed in synthetic zirconolite in Synroc
C (White et al., 1984) and synthetic zirconolite of
Ca0.6Sm0.4ZrNb0.4Mg0.4Ti1.2O7 composition (White, 1984). The defects
along the twin boundaries may become sites at which actinide elements
are incorporated. In some instances (Lumpkin et al., 1986b), there was
evidence for the formation of crystals (5-100 nm) with d -spacings
similar to those of the fluorite structure type. Lumpkin et al.(1986b)
suggested that metamict zirconolite may recrystallize initially with a
disordered, fluorite-type structure. Continued heating to higher
temperatures (1,000° to 1,200°C) appears to favor the highly twinned,
monoclinic zirconolite structure. Differential thermal analysis (30° to
1,200°C) showed exothermic recrystallization at 780°C with a heat of
recrystallization of 50 J/g for a zirconolite with a total dose of >
1026 a-decay events/m3. This is
similar to the value obtained by Foltyn et al. (1985) for synthetic
CaPuTi2O7 at nearly the same
a-decay
event dose of 1026
a-decay
events/m3 after 1,200 days. Although the measured energy release for the
natural zirconolites and Pu-doped zirconolites of equivalent dose are
nearly identical, both of these values are distinctly lower than the
value (84 J/g) measured for Pu-doped zirconolite that had experienced a
smaller dose (approximately 1025
a-decay events/m3) (Foltyn,
1985) or for Cm-doped zirconolite (126 J/g) and (Gd,Cm)2Ti2O7 at doses
of 2.0 X 1025
a-decay events/m3 (Weber et
al., 1985a; Weber et al., 1986). Foltyn et al. (1985) have suggested
that the decrease in released energy with increasing dose is due to a
redamaging process than results in a reordering of atoms in the metamict
state. Detailed
EXAFS and XANES studies have been completed on fully-aperiodic
zirconolite (Lumpkin et al., 1986b; Greegor et al., 1984a; Farges et
al., 1993). These studies, in many cases on the same sample from Sri
Lanka (#111.35), were on the Ca, Th, U, Zr and Ti EXAFS spectra, that is
all of the major cations of zirconolite. Principal conclusions included:
1.) The loss of long range atomic periodicity occurs by changes in the
metal-oxygen-metal (M-O-M) angles across mainly edge-sharing
coordination polyhedra. The tilting of nearest-neighbor coordination
polyhedra occurs without significant changes in the first and second
neighbor distances. Thus, the second-neighbor coordination geometries
for Ca, Ti, and Zr may not be very different in metamict and crystalline
zirconolite, despite the loss of long-range periodicity. This is
illustrated in Figure 36 which shows that a slight change (5°) in the
mean Zr-O-M angle correlates with a relatively large variation (0.01 nm)
in the Zr-M distance. Thus, variations in the mean M-O-M angle do not
have to be large in order to have a significant effect on long-range
periodicity and the medium range order in this radiation-damaged phase.
2.) Although there is some evidence for reduction in coordination number
and an associated decrease in bond-length with increasing
a-decay
event dose (Greegor et al., 1984), the short range (nearest-neighbor
oxygen) environments of Zr4+, Th4+, Ca2+ and Ti4+are little changed by
increasing radiation damage. 3.) The structural environments around Zr4+
and Th4+ in metamict zirconolite (7-coordinated and 8-coordinated,
respectively) are clearly different from those measured in silicate and
borosilicate glasses in which they are commonly 6-coordinated with
oxygens. This suggests that there are fundamental structural differences
in cation environments between radiation-induced aperiodic (metamict)
phases and glasses quenched from melts (Sales and Boatner, 1990). 5.2.3.
Perovskite
Perovskite, CaTiO3, is a mineral that may assume a wide range of
compositions as stable solid solutions. Perovskite is also a highly
refractory phase that has been stable in near surface environments for
billions of years. Thus, CaTiO3 is an excellent candidate for
immobilization of radioactive waste elements, particularly actinides,
and is a major constituent phase in Synroc and other titanate ceramic
waste forms. Perovskite is orthorhombic and consists of a 3-dimensional
network of corner-sharing TiO6 octahedra with Ca occupying the large
void space between the octahedra (the corner-sharing octahedra are
located on the eight corners of a slightly distorted cube). The
actinides, rare earths, and other large cations readily replace Ca in
the structure, while smaller-sized cations replace titanium.
There are limited data on radiation effects in perovskite.
Sinclair and Ringwood (1981) have studied natural perovskites of varying
age and uranium and thorium contents. Natural perovskite samples that
have received a dose of up to 2.6 x 1018
a-decay events/g (equivalent to
20,000 years storage for Synroc) showed a high degree of atomic
periodicity with no evidence of amorphization. The unit-cell volume
expansion at this dose was 1.8% due to the accumulation of simple
structural defects. Loparite is an isometric perovskite that contains
large amounts of Ce and Nb and is also highly crystalline. The unit-cell
volume expansion is 1.5 to 1.7% after a dose of 1 to 2 x 1018
a-decay
events/g (Ringwood et al., 1980; Ringwood and Sinclair, 1981).
Cold-pressed and sintered specimens of synthetic CaTiO3 have been
neutron irradiated in the fast-neutron flux of the HIFAR reactor at
Lucas Heights, for different lengths of time (Reeve et al., 1981;
Woolfrey et al., 1982). The highest dose reported was 2.5 x 1026 n/m2 (E
> 1 MeV), which represents an equivalent Synroc storage time of 5.8 x
105 yr and corresponds to 0.7 dpa or 8 x 1018 a-decay events/g. The specimens exhibited
macroscopic volume expansion, and a corresponding decrease in density
with increasing dose. The volume expansion reached a value of 6% at the
maximum dose achieved, but showed no indication of approaching a
saturation value (Figure 37). X-ray diffraction analysis indicated sharp
diffraction maxima, suggesting strong resistance to amorphization, but
shifts in specific peaks were not analyzed.
While the above studies indicate that the perovskite structure is
fairly resistant to radiation-induced amorphization, studies by Karioris
et al. (1982) indicate that CaTiO3 readily amorphizes under 3 MeV Ar+
ion irradiation. Similarly, the perovskite-type compound CmAlO3 has been
reported to transform from a crystalline to a fully amorphous state
after a dose of 1.4 x 1018 a-decay
events/g, which was achieved in 7.5 days from the decay of 244Cm
(Mosley, 1971). The crystal structure CmAlO3 was fully recovered after
annealing at 1000°C for 30 minutes. Questions regarding the radiation
stability of perovskite are being addressed in a new study of a-decay
damage in Cm-doped perovskite. Preliminary results indicate a linear
increase in volume with dose (0.75 % at 5 x 1017 a-decay
events/g), as well as an increased leach rate (Mitamura et al., 1994). 5.2.4.
Barium Hollandite
Barium hollandite, BaAl2Ti6O16, is a major phase in several
different Synroc formulations (Ringwood et al., 1979 and 1981; Ringwood,
1982; Newkirk et al., 1982). The structure of barium hollandite consists
of a framework of TiO6 octahedra that are linked (edge-sharing and
corner-sharing) to form square channels parallel to the c-axis.
These channels can accommodate a wide variety of large cations,
including fission products, such as Cs. Although hollandite will not
contain actinides, it will experience irradiation from a-particles
emitted in adjacent actinide-containing phases . The cumulative
a-particle fluence incident on
barium hollandite as a function of waste storage time is shown in Figure
38.
Weber (1985c) investigated the effect of
a-particle
irradiation on the structure of barium hollandite utilizing a-particles emitted isotropically from an
effectively semi-infinite 238PuO2 source. The energy spectra of the
a-particles emitted from this
source closely approximate that expected to be incident on the barium
hollandite phase in Synroc. Irradiation with the a-particles
resulted in a large expansion of the unit-cell volume that saturates at
2 to 2.5% and also induces a displacive transformation from the
tetragonal structure to a lower symmetry monoclinic structure. The
anisotropic unit-cell expansion is primarily along the
a- and b-axes and causes
a small increase of the angle b
from 90° in the tetragonal structure to 91° in the deformed monoclinic
structure. A similar tetragonal-to-monoclinic transformation has been
observed as a result of heavy-ion and electron irradiations (Barry et
al., 1983). The anisotropy of the unit-cell expansion caused an increase
in the size of the channels along the c-axis, which could
significantly affect the ability of the barium hollandite structure to
retain channel-site atoms, such as Cs, in an aqueous environment. The
implied increase in dissolution rates of non-network atoms as a result
of radiation-induced volume expansions has been previously described for
both the pyrochlore and zirconolite structures (Wald and Weber, 1984)
which also consist of a TiO6 octahedra network.
Comparisons of the
a-particle irradiation results
of Weber (1985) with the fast-neutron irradiation results of Woolfrey et
al. (1982) indicate a significant irradiation source dependence. Both
studies were completed using specimens from the same batch preparation
of hollandite. The volume changes under
a-particle
irradiation occur at an order of magnitude lower dose, calculated as
displacements per atom, than the volume changes under fast-neutron
irradiation (Figure 39). This is significant since an order of magnitude
in dose corresponds to two or three orders of magnitude in time. The
a-particle irradiation
primarily produces isolated interstitials and vacancies along the path
of the a-particles that give rise to a
large unit-cell expansion; whereas, the fast-neutron irradiation,
according to Ball and Woolfrey (1983) produces planar interstitial
clusters and isolated vacancies, with only a small increase in unit-cell
volume.
While there was no evidence for amorphization in the
a-irradiated
barium hollandite, heavy-ion irradiation of barium hollandite has been
reported to cause amorphization (Headley et al., 1982; Karioris et al.,
1982). These results clearly demonstrate that the proper selection of
irradiation source is important if an accurate simulation of waste form
behavior is desired. 5.2.5.
Zircon Zircon
(ZrSiO4, Ir1/amd, Z = 4) occurs in silica-rich tailored ceramics,
and its structure type is the high-temperature form of the heavy rare
earth phosphates (the low-temperature form has the monazite structure
type). Pu can substitute for Zr, and radiation damage studies
have been completed on 238Pu-doped zircon up to concentrations of 8.1
mole percent 238Pu (Weber, 1990, 1991b). Additionally, zircon is a phase
extensively used in U/Pb radiometric age-dating, as it may contain
uranium in concentrations up to 5,000 ppm. Early workers (Holland and
Gottfried, 1955) investigated radiation damage effects in an effort to
use the degree of damage as a measure of the age of the zircon. Theirs
was the first study to quantitatively correlate the decrease in density,
decrease in refractive index, and expansion of the unit cell parameters
with increasing
a-decay
event dose. Based on their study of zircon, Holland and Gottfried
proposed the first model of defect accumulation and damage ingrowth for
a complex ceramic. This early history of the study of radiation damage
effects in metamict zircons is summarized by Ewing (1994). There
are data on radiation effects in natural zircons (irradiation time: 570
million years (Murakami et al., 1986, 1991, Woodhead et al., 1991a;
Woodhead et al., 1991b)), Pu-doped zircon (irradiation time: 6.5 years (Exarhos,
1984; Weber, 1990, 1991b)) and most recently ion-beam irradiated zircon
(irradiation time: 30 minutes (Babsail et al., 1991; Wang et al., 1993;
Weber et al., 1994). The damage rate for the 238Pu-doped zircon is 6 x
1010 Bq/g, and the damage rate for natural zircons is <103 Bq/g. This
extraordinary data set with a dose rate range of >108, allows one to
investigate the efficacy of using accelerated techniques
(actinide-doping and ion-beam irradiations) to simulate the long-term
effects of a-decay
event damage. The
study of natural zircons culminated with the very detailed analysis
(Murakami et al., 1991; Woodhead et al., 1991a) of damage in-growth in
suites of zircons from Sri Lanka (570 million years old) which had
experienced variable
a-decay event doses due to
variable U and Th contents (0.06 x 1015 to 6.8 x 1015 a-decay
events/mg). In addition to the decrease in density and refractive
indices with a-decay event dose as
determined by Holland and Gottfried in 1955, later workers applied
detailed x-ray diffraction analysis (Murakami et al., 1986, 1991),
electron microprobe analysis (Murakami et al., 1991; Chakoumakos et al.,
1987), high resolution transmission electron microscopy (Murakami et
al., 1991; McLaren et al., 1994), infrared spectroscopy (Woodhead et
al., 1991b), and x-ray absorption spectroscopy (Farges and Calas, 1991;
Farges, 1994) to the study of this radiation-induced transition to the
aperiodic (metamict) state. Most recently, the decreases in hardness
(40%) and elastic moduli (60%) over the dose range of 0.15 to 0.65 dpa
were determined in zoned zircons using a mechanical properties
microprobe (Chakoumakos et al., 1991), and increases in stored energy
have been measured as a function of increasing
a-decay
event dose (Ellsworth et al., 1994). The enthalpy of annealing at room
temperature varies sigmoidally as a function of radiation dose and
reaches a saturation plateau at doses of 5 x 1015
a-decay
events/mg. The large magnitude of the enthalpy of annealing plateau, -59
kJ/mol, suggests that the damage to the structure is pervasive on the
scale of nanometers. Based
on x-ray diffraction analysis and high resolution transmission electron
microscopy, Murakami et al. (1991) described three stages (Figure 40) of
damage in-growth. Stage-I (< 3 x 1018 a-decay
events/g) is characterized by sharp Bragg diffraction maxima with a
minor contribution from the diffuse-scattering component. Electron
diffraction patterns were sharp. Damage is dominated by the accumulation
of isolated point defects which cause unit-cell expansion and distortion
that account for most of the decrease in density. These defects may
partially anneal over geologic time. Stage-II (3 to 8 x 1018
a-decay
events/g) is characterized by significant decreases in the intensity of
the Bragg diffraction maxima which become asymmetric due to increased
contributions from the diffuse-scattering component. High-resolution
transmission electron microscopy revealed that the microstructure
consisted of distorted crystalline regions and amorphous
"tracks" caused by
a-recoil nuclei. With
increasing a-decay dose, damaged
crystalline regions are converted into aperiodic regions, but with no
further significant expansion of the unit cell in the remaining
crystalline regions. Stage-III (>8 x 1018
a-decay
events/g) consists of material which is entirely aperiodic as far as can
be determined by X-ray or electron diffraction. Figure 40 is a schematic
representation of the changes in x-ray diffraction and electron
diffraction pattern that occur at each of the three stages. These stages
of damage are illustrated in HRTEM micrographs of the range of 0.0025 up
to 0.50 dpa (Figure 41). Finally,
there was no HRTEM or XRD evidence for the formation of Zr02 or Si02 as
final products during the last stage of metamictization: however, recent
work based on microcalorimetry (Ellsworth et al., 1994), X-ray
absorption spectroscopy (Farges, 1994) and SIMS and electron microscopy
(McLaren et al., 1994) has postulated the presence of zirconium- and
silicon-rich domains which are less than 10 Å in size. These domains
have been identified at higher temperatures by neutron diffraction
experiments (Mursic et al., 1992). Based
on modeled density changes, aperiodic regions continued to experience a
change in structure as they were redamaged. During Stage-II of the
process, the modeled density of aperiodic regions changed from 4.5 g/cm3
to 4.1 g/cm3. The amorphization process is consistent with a model for
the multiple overlap of displacement cascades (Carter and Webb, 1979;
Webb and Carter, 1979, 1981), suggesting amorphization occurs as a
result of defect accumulation rather than directly within a single
displacement cascade. Figure 42 compares the fraction of amorphous
material (estimated from density and optical data) for Pu-doped and
natural zircons (solid and dashed lines, respectively) to the calculated
values for a double overlap model (i.e., the defect density for
amorphization requires the overlap of three cascades as discussed
below). Note the incubation dose (0.05 dpa) required for the natural
zircon which reflects annealing of the isolated point defects.
Comparison of the results for natural zircon with those for Pu-doped
zircon (Weber, 1991b) showed that dose-rate variations (even as great as
a factor of 108) had no substantial effect on the damage accumulation
process. Unit
cell parameters and volume increased (Figure 43) and density decreased
more for the Pu-doped zircon than for natural zircon in the early stages
of damage accumulation (< 3 X 1018 a-decay
events/g), suggesting that annealing of point defects in the early
stages of the damage accumulation process occurs in natural zircon under
ambient conditions. Evidence for annealing under ambient conditions was
noted by Holland and Gottfried (1955) and the mean lifetime of an
a-recoil track was estimated to
be 400 million years by Lumpkin and Ewing (1988). The annealing accounts
for the distinct sigmoidal shape of the damage curves for natural zircon
and the apparent incubation period before the onset of amorphization
(see Figure 42). Thus, radiation damage from a-decay
events in zircon results in the simultaneous accumulation of simple
structural defects and amorphous (or aperiodic) regions that exhibit
loss of long-range order (Chakoumakos et al., 1987; Murakami et al.,
1991; Farges and Callas, 1991). The amorphous regions eventually overlap
to produce a completely amorphous state that may be structurally unique.
The local amorphization process is believed to result from the
spontaneous local collapse of the crystal structure and disappearance of
long-range order due to a local high defect concentration; amorphization
is triggered when the free energy of the high-defect region becomes
equal to the free energy of the amorphous state. The local defect
concentrations necessary to trigger amorphization can occur directly in
the displacement cascade (particle track) of a single heavy particle
(recoil nucleus in a-decay)
or from the overlapping cascades of individual light and heavy particles
(a-particle and recoil
nucleus). Several
models of both direct and multiple-overlap processes have been developed
to describe amorphization in ion-irradiated semiconductors (Carter and
Webb, 1979; Webb and Carter, 1979). These models have been used to gain
some insight into the process for amorphization in zircon due to
internal a-decay events (Weber et al.,
1994). As shown previously (Weber, 1990; Weber, 1993) amorphization in
Pu-doped and natural zircons is consistent with amorphization occurring
by the catastrophic collapse of local structure through an increase in
local defect concentration by the double overlap of displacement
cascades. The fraction of amorphous phase, fa, is given by the
expression
fa = 1 - [(1 + RD + R2D2/2) exp(-RD)]
(3) for
the double-overlap model (Carter and Webb, 1979; Webb and Carter, 1979),
where R (with dpa as the unit of dose, D) is the average ratio of total
atoms to displaced atoms within the cascades (tracks) produced by the
a-decay event particles and D (a-decay events per unit volume
or mass). The amorphous fractions have been determined previously for Pu-doped
zircon from macroscopic swelling and unit-cell expansion data (Weber,
1990) and for natural zircons by Holland and Gottfried (1955) from both
refractive indices and density data. The results and model fits are
shown in Figure 42. The amorphous fractions for Pu-doped and natural
zircons exhibit very similar behaviors as functions of cumulative
displacement dose, except for an offset at low dpa due to the incubation
dose in natural zircon that is caused by the point-defect annealing over
geologic time as discussed above. In the case of Pu-doped zircon,
nonlinear regression analysis yields a value of R = 9.51 dpa-1. An
incubation dose, Do, must be assumed in the case of natural zircon for
the double-overlap model to provide a reasonable fit to the results of
Holland and Gottfried (1955). Nonlinear regression analysis yields
values of R = 11.67 dpa-1 and Do = 0.05 dpa for natural zircon. The
value of R is in good agreement with the value determined for Pu-doped
zircon, and the incubation dose is consistent with the incubation dose
determined from both the XRD intensity data and the onset for expansion
along the a-axis. Physically, the double-overlap model
suggests that some limiting local defect concentration that corresponds
to the overlap of three displacement cascades is required to trigger
amorphization; this is consistent with the dependence of amorphous
fraction on defect concentration reported earlier, with the observed
effect of defect annealing, and with the observations of some
intermetallic compounds. If
it is assumed that the crystalline fraction contains both isolated
defects and disordered regions (recoil tracks) that give rise to the
measured unit-cell expansion, then, due to the composite nature of the
radiation-induced microstructure the total macroscopic swelling, ∆Vm
/Vo, of zircon can be expressed as:
∆Vm /Vo = fc ∆Vuc /Vo + fa ∆Va /Vo
(4) where
fc is the mass fraction of crystalline phase, fa is the mass fraction of
amorphous phase, ∆Vuc /Vo is the average unit-cell volume change
of the crystalline phase, and ∆Va/Vo is the relative volume change
associated with the amorphous phase (Weber, 1993). The unit-cell volume
change, ∆Vuc /Vo, is given by Eq. (2) with the fit parameters
determined directly from experimental measurements. It can be inferred
that the volume change associated with amorphization, ∆Va /Vo, is
constant and equal to the value of the macroscopic swelling at
saturation, (∆Va /Vo = 16.6% for the Pu-doped zircon and ∆Va
/Vo = 18.4% for natural zircons). Consequently, as fa is known and fc =
1 - fa, the total macroscopic swelling, ∆Vm /Vo, can be calculated
using Eq. (4) and is shown in Figure 44, along with the experimental
data on total macroscopic swelling and the calculated crystalline, fc
∆Vuc /Vo, and the amorphous, fa ∆Va /Vo, components. The
calculated macroscopic swelling, based on Eq. (4), provides an excellent
fit to the experimental data. The dominance of the unit-cell expansion
at low doses is clearly evident for both the Pu-doped and natural
zircons.
Most
recently, particle irradiations (2 MeV He+ (Babsail et al, 1991); 0.8
MeV Ne+, 1.5 MeV Ar+, 1.5 MeV Kr+, 0.7 MeV Kr+, 1.5 MeV Xe+ (Wang et
al., 1993) have been used to induce amorphization in zircon, and the
results compared (the time scales in this comparison range from 0.5
hours to 570 million years!) to the results of
a-decay
event damage (Weber et al., 1994). High resolution transmission electron
microscopy confirms that the damage ingrowth processes are similar for
natural and ion-beam irradiated zircons (Figure 45). The defects are
produced in varying concentrations along each particle track, depending
on mass and energy of the particle. Figure 46 shows the dependence of
amorphization dose in zircon at room temperature with the average energy
transferred to recoils per ion. These results show that the
amorphization dose for the heavier ions (Kr+ and Xe+ dose rates of
approximately 10-4 to 10-3 dpa/s) is nearly identical to that determined
from the 92 keV U+ recoil emitted during
a-decay in 238Pu-doped zircon
with a dose rate of 3 X 10-9 dpa/s. Thus, the amorphization process is
nearly independent of the damage source (a-decay events vs. heavy-ion
beam irradiations) and is consistent with models based on the multiple
overlap of particle tracks, indicating amorphization occurs as a result
of a critical defect concentration. Using
in situ TEM to determine the critical amorphization dose of zircon
irradiated by 1.5 MeV Kr+ as a function of temperature provides insight
the competing roles of defect recovery (annealing) and damage
accumulation in a waste form ceramic (Figure 47). The dose for complete
amorphization in zircon increases with temperature in two stages. This
increase in amorphization dose with temperature is believed to be due to
a decrease in the average cascade size (intracascade defect density) or
shrinkage of the amorphous regions (in overlapped cascades) as a result
of the thermal annealing of irradiation-induced defects. The first stage
(Stage I), which occurs below 300 K, may be associated with the
intracascade recombination of closely-spaced defects. As the temperature
increases above 300 K, the Stage I annealing process in zircon is
complete (all Stage I defects have recombined), and the dose for
complete amorphization from the remaining (unannealed) defect structure
within the cascade is constant, independent of temperature, from 300 to
473 K. Above 473 K, the dose for complete amorphization again increases
with temperature in a second stage (Stage II), which may be due to
additional intracascade recombination from the thermal migration of
remaining point defects or irradiation-enhanced epitaxial
recrystallization (point defect annihilation at the
crystalline-amorphous interface). Because
both the evolution of the amorphous state and the dose for complete
amorphization in the 1.5 MeV Kr+ irradiated zircon at room temperature
are similar to that observed above for the Pu-doped and natural zircons,
it is assumed that amorphization under heavy-ion irradiation is also due
to the multiple overlap of cascades. The effects of temperature on
amorphization can then be modeled by kinetic processes with activation
energies, Ea, that decrease the average cascade size (or average defect
density) and, consequently, the rate at which the fraction of material
damaged directly by the cascades, accumulates at constant flux. Weber et
al. (1994) have determined the activation energy, Ea, associated with
each annealing stage. The activation energy for Stage I defect annealing
is 0.02 ± 0.01 eV, and the activation energy determined for Stage II
defect annealing is 0.31 ± 0.03 eV. In addition to the activation
energies, the critical temperature, Tc , above which complete
amorphization does not occur due to the dominance of Stage II defect
annealing over damage production, can be determined. The critical
temperature, Tc , for zircon under these irradiation conditions is
estimated to be 1100 K. The critical temperature for complete
amorphization in zircon for other ions and energies may be different, as
has been observed for Ca2La8(SiO4)6O2 (Weber and Wang, 1994). The low
value of the activation energy for Stage I suggests that this annealing
stage in zircon is associated with an intracascade, close-pair
recombination process at low temperatures that is active up to 300 K
where annealing is apparently complete and above which it is assumed
that all close pairs recombine instantaneously. This process, which is
insufficient to prevent complete amorphization, reduces the number of
defects that survive each cascade at temperatures above 300 K. The
activation energy for Stage II annealing in zircon is similar to
activation energies reported in other materials for an
irradiation-enhanced epitaxial recrystallization process associated with
irradiation-enhanced defect mobility; however, because the amorphization
process in zircon is apparently due to the accumulation of a critical
defect concentration from cascade overlap, the Stage II annealing could
also be due to additional defect recombination and annealing from the
mobility of one or more defects above 473 K. Although the exact nature
of the second-stage annealing process is not yet known, it competes with
the irradiation-induced amorphization process and dominates above Tc,
thus preventing complete amorphization. The activation energy for the
second annealing stage, 0.31 eV, is considerably less than the
activation energy, 3.6 eV, for thermal track annealing (epitaxial
thermal recrystallization) in zircon (Sandhu et al., 1990) or the
activation energies, 5.1 and 6.6 eV, for bulk thermal recrystallization
of fully amorphous zircon (Weber et al., 1994). This type of thermal
annealing analysis is essential in the study of crystalline ceramic
waste forms in order to determine their final damage state as a function
of the thermal history of the waste form. Because
of the important implications in determing the U/Pb systematics of
zircons for geologic age dating, there are data on the increased loss of
nuclides as a function of increasing degree of damage (Pidgeon et al.,
1966; Ewing et al., 1982b; Tole, 1985; Suzuki, 1987; Sinha et al.,
1992), as well as efforts to relate these data to nuclear waste
containment (Gentry et al., 1982; Ludwig et al., 1984, Sinha et al.,
1991). In a study of a suite of natural zircons with varying contents of
uranium and thorium, Ewing et al. (1982b) found that over a dose range
of 1019 to 1021 a-decay
events/g, there was an increase of one order of magnitude in the weight
percent of zircon that is dissolved and an increase in the leach rate
from 2.9 X 10-8 to 2.3 X 10-7 g/cm2 day. For completely metamict zircons
(dose > 1022 a-decay
events/g) the leach rate was as high as 1.8 X 10-6 g/cm2 day. These
results are consistent with the observation that loss of nuclides or
increased solubility is, at least in part, related to the increasing
level of radiation damage (Pidgeon et al., 1966; Tole, 1985; Sinha et
al., 1992). This
unusually complete data set for damage effects in zircon provides a
basis for confirming the use of accelerated damage processes
(actinide-doping of synthetic phases and heavy-particle irradiations) to
study the long term behavior of ceramics in a radiation field. Similar
studies should be possible for members of the pyrochlore group as these
minerals occur abundantly in nature and can be synthesized with a wide
variety of actinide-bearing compositions (Chakoumakos and Ewing, 1985). 5.2.6.
Apatite
Several rare-earth silicates with the apatite structure have been
observed or proposed as actinide host phases in a nuclear waste glass
(Weber et al., 1979), supercalcine (McCarthy, 1977; McCarthy et al.,
1979), a glass ceramic waste form (Dé et al., 1976a and 1976b), and a
cement waste form (Jantzen and Glasser, 1979). The apatites in nuclear
waste solids are rare-earth silicate isomorphs of natural apatite,
Ca10(PO4)6(F,OH)2, and generally have the composition:
Ca4-xRe6+x(SiO4)6-y(PO4)yO2 (where RE = La, Ce, Pr, Nd, Pm, Sm, Eu, and
Gd). The actinides readily substitute for the rare earths in this
hexagonal crystal structure when it is present in the waste form.
As noted above in the study of a Cm-doped nuclear waste glass
(Weber et al., 1979; Weber and Roberts, 1983), crystallites of
Ca3(Gd,Cm)7(SiO4)5(PO4)O2 transformed from a crystalline to an amorphous
state as a result of a-decay
of the incorporated Cm. The volume expansion associated with the
amorphous transformation resulted in microfracturing of the simulated
waste glass. Similarly, in a study of Cm-doped supercalcine (Rusin et
al., 1979), the originally crystalline apatite phase,
Ca2(Re,Cm)8(SiO4)6O2, transformed to an amorphous state due to
a-decay.
The unit cell of this phase was estimated to expand 2.5% before the
amorphization process was complete. More specific details of the
amorphization process in these phases, structural data, and the effects
of amorphization on the integrity of these apatite phases could not be
obtained from the study of these multiphase waste forms. Several natural
apatite minerals with significant amounts of rare earths and silica are
known to have an appreciable thoria content. These minerals are reported
to be partially or fully metamict (amorphous) due to self-radiation
damage from the alpha decay of the Th (Lindberg and Ingram, 1964; Haaker
and Ewing, 1981), but detailed structural data on the effects of
radiation in these minerals have not been reported. Studies of fission
track dating and annealing (Koul, 1979) and ion-beam-induced
amorphization (Wang et al., 1994) in naturally occuring apatites have
also been reported.
In order to understand the effects of self-radiation damage from a-decay
on apatite phases in nuclear waste forms, an extensive study of
single-phase, 244Cm-doped Ca2(Nd)8(SiO4)6O2 has been carried out (Weber,
1982, 1983 and 1993; Wald and Weber, 1984; Weber and Matzke, 1986a and
1986b). Radiation damage from
a-decay in this material
results in the simultaneous accumulation of simple structural defects
(i.e., vacancies and interstitials on both the cation and anion
sublattices) and amorphous regions that exhibit loss of long-range
order. The amorphous regions eventually overlap to produce a completely
amorphous state (Figure 48). Consequently, as the radiation damage
process proceeds, the fraction of crystalline material, fc, decreases
and the fraction of amorphous material, fa, increases. For this
material, Weber (1993) has shown that amorphization primarily occurs
directly in the cascades of the recoil nuclei emitted during
a-decay of Cm and that the
increase in amorphous fraction with dose, D, is given by the expression:
fa = 1 - exp(-BaD)
(5) where
Ba, the mass of amorphous material produced per decay event, has been
determined to be 4.42 x 10-19g. The change in amorphous fraction with
dose for Cm-doped Ca2Nd8(SiO4)6O2 is shown in Figure 49. The crystalline
fraction, fc = 1 - fa, is consistent with the observed decrease in
diffracted x-ray intensity (Weber, 1993).
The accumulation of structural defects in the crystalline
fraction of material results in an increase in unit-cell volume with
cumulative dose. The unit-cell volume expansion, DVuc/Vo, which is shown in
Figure 50, follows the characteristic exponential behavior predicted by
models (Nellis, 1977; Weber et al., 1991) for the accumulation of
isolated structural defects and given by the expression:
DVuc
= Auc [1 - exp(-BucD)]
(6) where
Auc is the relative unit-cell volume expansion at saturation and Buc is
the rate constant (per unit dose) for the simultaneous recombination of
structural defects during irradiation. The values for Auc and Buc have
been determined to be 4.95% and 5.08 x 10-19g
(Weber, 1993). The unit-cell expansion is anisotropic with the expansion
of the unit cell along the
a-axis
about twice that along the c-axis. This anisotropic behavior is
similar to that observed in ion-irradiated Ca2Nd8(SiO4)6O2 and shown to
be consistent with topological constraints on the changes in the
polyhedral connectivity within this structure (Weber et al., 1991).
The overall macroscopic swelling, DVm/Vo, which has
contributions from the crystalline and amorphous fractions of material,
is given by equation 4 (Weber, 1993). The overall macroscopic swelling
determined experimentally from changes in immersion density is shown in
Figure 51 along with the predicted contributions from the crystalline
and amorphous fractions. The solid curve is the sum of the crystalline
and amorphous components (Eq. 4) and provides an excellent fit to the
data. The nearly equal contribution of the unit-cell expansion and
amorphization process to the total macroscopic swelling at low doses is
clearly shown as is the dominance of the the amorphization process at
high doses.
The normalized dissolution rate of Cm-doped Ca2Nd8(SiO4)6O2 at 90°C
in deionized water increased by about a factor of ten as a result of
radiation-induced swelling and amorphization (Weber, 1983; Wald and
Weber, 1984). The dissolution appears to occur congruently in this
apatite phase. The mechanism for the enhanced dissolution of the
amorphous phase is most likely associated with the high degree of
disorder, broken bonds, and more open structure.
The fracture toughness in Cm-doped Ca2Nd8(SiO4)6O2 increased with
cumulative dose to a broad maximum and then decreases slightly (Figure
52). This is similar to the behavior discussed above in pyrochlores and
zirconolite (Clinard et al., 1985; Weber et al., 1986a). The increase in
fracture toughness is attributed to the composite nature of the
microstructure. At low to intermediate doses, the microstructure
consists of amorphous tracks in a crystalline matrix, and this inhibits
crack propagation and increases the fracture toughness. As the amorphous
phase becomes the dominant matrix at high doses with remanent
crystallites, the fracture toughness decreases slightly.
The activation energy for recrystallization from the fully
amorphous state in Cm-doped Ca2Nd8(SiO4)6O2 has been rigorously
determined by Weber (1983) to be 3.1 +0.2 eV. Recrystallization
of the amorphous state occurs in a single exothermic recovery stage that
releases 130 J/g of stored energy. Full recovery of the density and
original apatite crystal structure occurs during this single-stage
recrystallization process.
The temperature dependence of amorphization in single crystals of
the apatite phase, Ca2La8(SiO4)6O2, has been been determined in recent
ion-irradiation studies by Weber and Wang (1993 and 1994). The dose for
complete amorphization in Ca2La8(SiO4)6O2 is independent of
recoil-energy spectra at low temperatures (< 15 K) and increases with
temperature in two stages for both 0.8 MeV Ne+ and 1.5 MeV
Kr+irradiations (Figure 53). The low-temperature stage (below 200 K) has
an activation energy of 0.01 +0.003 eV and may be associated with
close-pair recombination. The high-temperature stage (above 200 K) has
an activation energy of 0.13 +0.02 eV and may be associated with
irradiation-enhanced defect mobility. The critical temperature above
which complete amophization does not occur increases from 360 K for Ne+
to 710 K for Kr+. The amorphization dose below room temperature, 0.3 to
0.5 dpa for the heavier Kr+ irradiations is similar to the dose, 0.4 dpa,
determined for full amorphization of Cm-doped Ca2Nd8(SiO4)6O2 at room
temperature. The dose for complete amorphization increases more rapidly
for the lighter Ne+ irradiations because of the larger fraction of
recoverable defects and smaller cascades that are produced by the
low-energy recoils. 5.2.7.
Monazite Monazite
is unique, as it has been proposed as a single-phase ceramic to
incorportate a wide variety of nuclear wastes, particularly those rich
in actinides (Boatner et al., 1980; Boatner and Sales, 1988). The
mineral monazite is a mixed lanthanide orthophosphate, LnPO4 (Ln = La,
Ce, Nd, Gd etc.), that often contains significant amounts of Th and U
(up to 27 wt. % combined). As a result, many natural monazites have been
subjected to significant
a-decay event damage over
geologic time (some samples are up to two billion years of age). In
spite of the large radiation doses received by natural monazites, Ewing
(1975) and Ewing and Haaker (1980) have noted that monazite is always
found in the crystalline state. The apparent resistance of natural
monazites to radiation-induced amorphization was an important factor in
the initial proposal of monazite ceramics as potential candidates for
the immobilization of nuclear wastes. Karioris
et al. (1981) have established that natural monazite (Ce,La,Y)PO4 can be
readily transformed to an amorphous state by irradiation with 3 MeV Ar+
ions to moderate doses. Gowda (1982) determined that under 3 MeV Ar+ ion
irradiation the damage cross-section for amorphization was 0.54 +
0.10 nm2 for natural monazite and 0.56 + 0.08 nm2 for synthetic
monazite, CePO4. In natural monazite that was irradiated to a partially
amorphous state (diffraction intensities were reduced by one-third),
complete recovery of the fully crystalline state occurred after
annealing at 300°C for 20 hours (Karioris et al., 1981). The apparent
stability of natural monazite is attributed to this relatively low
temperature of recovery of the radiation damage. Differential scanning
calorimetry (DSC) measurements of natural (Gowda, 1982) and synthetic (Gowda,
1982; Ehlert et al., 1983) monazites irradiated with 3 MeV Ar+ ions have
been performed. In natural monazite, the stored energy released during
recrystallization was 33 J/g, the peak in the release rate occurred at
450°C, and the activation energy associated with the recovery process
was estimated to be 2.77 eV. For synthetic monazite, CePO4, the stored
energy release was 30.5 + 1.3 J/g, the temperature for maximum
release was slightly lower at 370°C, and the activation energy was
estimated to be 2.7 + 0.3 eV. Results
by Sales et al. (1983) suggest that the leach rate of the host matrix of
a synthetic monazite, LaPO4, containing simulated waste remains low even
after the material has been transformed to an amorphous state by
irradiation with 250 keV Bi+ ions. However, work by Eyal and Kaufman
(1982) on natural monazite indicate that there is a preferential
dissolution of the radionuclide daughter-products, 234U, 230Th, and
228Th, by factors of between 1.1 and 10 relative to the structurally
incorporated parent isotopes 238U and 232Th. This isotopic fractionation
is attributed to radiation damage in the tracks of the recoil nuclei
emitted during a-decay
of the parent isotopes. While there have been some concerns regarding
these results (Boatner and Sales, 1988), the increases in dissolution
rates are similar to those observed in other actinide-host phases. 5.2.8.
Sphene ( = titanite) Sphene,
CaTi(SiO4)O (the proper mineral name is titanite) is a prominent phase
in the sphene-glass ceramic. Sphene is an actinide-host (the structure
will not accommodate Cs; but will accommodate minor concentrations of Sr),
thus damage due to a-decay
of actinides is an issue of concern. Although
there have been studies using heavy particle irradiations, 3 MeV Ar+ and
280 keV Bi2+ (Vance et al., 1984; Stevanovic et al., 1989) and studies
using g-radiation, 200 keV electrons
(Vance et al., 1986), most work has focused on natural sphenes which
contain U and Th (Vance and Metson, 1985; Hawthorne et al., 1991). Up to
doses of 2 x 1027 e/m2 there was no evidence of structural damage. The
Ar+ irradiation indicated that a fluence of 7 x 1018 a-decay
events/g was required to produce complete metamictization. This is in
good agreement with the value of 5 x 1018 a-decay
events/g required to reach the metamict state in natural sphenes.
Leaching experiments on damaged sphenes indicated that the increase in
dissolution rate was no more than a factor of five; however, leaching
experiments on Bi2+ implanted sphene and sphene-ceramics to doses of 6 x
1015 ions/cm2 resulted in an order of magnitude increase in the leach
rate (Stevanovic et al., 1989). The
study of natural sphenes by Hawthorne et al. (1991) was unusual in that
a wide variety of techniques (electron microprobe analysis, powder x-ray
diffraction, single-crystal x-ray diffraction, infrared spectroscopy, Mössbauer
spectroscopy, magic-angle spinning nuclear magnetic resonance
spectroscopy, high resolution transmission electron microscopy, extended
x-ray absorption spectroscopy and x-ray absorption near edge
spectroscopy) were used to monitor structural changes with increasing a-decay
event dose. The results substantiated the model for damage ingrowth
outlined by Murakami et al. (1991) and illustrated in Figure 40. The
work did suggest that the damage process is accompanied by the reduction
of Fe3+ to Fe2+, which resides mainly in the aperiodic domains. Valence
changes of this type can have significant affects on the solubility of
certain actinides (e.g, uranium and plutonium) particularly when they
are partitioned into the aperiodic domains. The
issues surrounding the permanent disposal of nuclear waste are both
political and technological. The delays in winning public acceptance for
current nuclear waste policy and technology threaten shut-down of
current commercial nuclear power plants and hamper acceptance of new
nuclear power reactors, both of which could play a vital role in the
U.S. strategy to minimize human impact on global processes, such as CO2
emissions. Thus, the successful demonstration that nuclear waste can be
disposed of safely is an important step in the development of nuclear
power. One of the critical barriers to radionuclide migration to the
biosphere is the nuclear waste form; however, the physical and chemical
properties of the waste form may change dramatically as a result of
irradiation damage. Additionally, radiation-solid interactions might
also affect important geochemical processes that are relied upon for
nuclide retardation (e.g. sorption of radionuclides on clays, radiolysis
of transporting fluids, radiolytic oxidation and corrosion of UO2, and
radiation damage in secondary phases which by their solubilities control
radionuclide concentrations in the ground water). Unfortunately, due to
the state of understanding and limited research in this area (Weber et
al., 1991b), the nature, dimensions, and implications of these radiation
effects remain elusive, particularly with regard to how these effects
impact regulatory criteria for radionuclide release. More
specifically, certain topics require immediate attention: 1.
The effect of ionizing radiation on borosilicate waste glasses,
particularly the possibility and impact of radiolytic decomposition and
bubble formation, are neither well understood or under current
investigation. Studies are needed to determine the temperature and dose
rate dependence of these phenomena and their impact on radionuclide
release rates. A thorough understanding of the synergistic effects of
ionizing and displacive radiation on waste glass behavior is needed.
This should include the effects of radiolytic decomposition and
microstructural evolution on increased surface area and leachability,
which determine the release rate of radionuclides under unsaturated
conditions. Also, there are only limited (and in most cases
questionable) data on the effect of displacive events (e.g., a-recoil
damage) on the leachability of borosilicate waste glasses. 2.
Computer simulation of displacement processes. The final state of
a radiation-damaged glass or ceramic is a result of the
temperature-dependent balance between damage and annealing kinetics. The
initial damage stage can only be ascertained by models of collision
cascade formation ( these events occur in periods of 10-10 to 10-12
seconds). Very little is known about the process of collision cascade
formation in glasses and ceramics, as compared to the state of knowledge
for metals. As an example, we have only very limited data on the
displacement energies of cations in complex, multi-atom ceramics which
are typical of nuclear waste form assemblages. 3.
Radiation-induced periodic-to-aperiodic transformations
(amorphization). Little is known of the mechanisms of
crystalline-to-amorphous transformations, particularly in ceramics.
There has been no systematic evaluation of the role of structure-type,
bond-type, bond-strength and structural topology on this phenomenon.
This knowledge is necessary for the efficient evaluation of crystalline
ceramic waste phases and their response to different types and doses of
radiation. This mechanistic understanding of the amorphization process
is particularly important in extrapolating the long-term behavior of
ceramics and glasses in a radiation field. 4.
Structure of the aperiodic state. Preliminary research has
demonstrated that there are many possible structures and topologies in
aperiodic materials. These structural variations may reflect the mode of
amorphization (e.g., radiation-induced, pressure-induced or quenched
from a melt), and the aperiodic structure may change with increasing
level of damage. These changes can have significant effects on the
behavior (e.g., leach rate, devitrification and mechanical properties)
of aperiodic phases, such as the borosilicate glass waste form. Acknowledgments The
authors gratefully acknowledge the support that they have received from
the Office of Basis Energy Sciences of the U.S. Department of Energy (WFC:
contract #W-7405-ENG-36; RCE: grant #DE-FG03-93ER45498; WJW contract
#DE-AC06-76RLO 1830). The BES support of basic, interdisciplinary
research on radiation effects in minerals and ceramics has contributed
greatly to the present understanding of radiation damage effects in
nuclear waste forms. F.W. Clinard, Jr. expresses his appreciation for
the contributions of his co-workers in the Materials Science and
Technology Division at Los Alamos National Laboratory; W.J. Weber
acknowledges the contributions of his co-workers at the Pacific
Northwest Laborary; R.C. Ewing thanks his colleagues and students for
their collaboration in the field of radiation damage effects in minerals
and ceramics and the good times at El Patio. All errors of omission and
interpretations of the data in the literature remain those of the
authors. Finally, R.C. Ewing thanks his colleagues in the Department of
Earth Sciences of Aarhus University for their support during his visit
in 1994 during which most of the writing of this review was completed.
|