It is for me an honour and a privilege to have been asked to open the technical programme of the 1988 MRS conference with a talk on the important topic of research priorities in radioactive waste management. It is essential in this field that we continually review the relevance and applicability of our work; this is clearly recognized by the organisers of the MRS conferences, as witnessed also by the 1985 meeting where an entire session was devoted to overviews of research priorities, and by the 1982 meeting, where a very interesting panel discussion on the topic was led by Professor Tom Pigford.

Nuclear waste forms may be divided into two broad categories: (1) single phase glasses with minor crystalline components (e.g., borosilicate glasses) and (2) crystalline waste forms, either single phase (e.g., monazite) or polyphase (e.g., SYNROC). This paper reviews the materials properties data that are available for each of these two types of waste forms. The prinicipal data include: (1) physical, thermal and mechanical properties, (2) chemical durability; (3) radiation damage effects. Complete data are only available for borosilicate glasses and SYNROC; therefore, this comparison focuses on the performance assessment of borosilicate glass and SYNROC.

The comparative leaching behavior of a soda-lime glass and of a simulated nuclear glass has been investigated through H and Na depth profiles obtained with resonant nuclear reaction analysis. It is shown that glass surface hydration involves both H+/Na+ ion exchange and permeation of molecular water, the first process being dominant at low temperature and solution ionic strength and the second for high values of these two parameters. The shape of the H profile as well as the response of H retention to thermal treatment indicate that on the nuclear glass, most of the molecular water is more weakly bonded than on the soda-lime glass.

We have investigated the relationship between hydrated layer formation during aqueous corrosion of silicates and retention of heavy elements (Fe, REE, actinides). Our approach is based on the comparison of the dissolution behaviour of silicate glasses, silicate minerals implanted with increasing doses of lead ions (1×E+12 to 1×E+15 ions/cm2), sorption experiments on silica surfaces and direct precipitation of hydrosilicates. The characterization of reacted surfaces was performed by combining Rutherford backscattering spectrometry (RBS) for profiling heavy elements with Resonant Nuclear Reaction Analysis (RNRA) for hydrogen profilimetry. The accumulation of these elements does not necessarily imply a selective dissolution and can be explained by the “precipitation” of hydroxides or hydrosilicates.

Experimental basaltic glass dissolution in fresh water is compared with analyses made on subglacial hyaloclastites from Iceland. The dissolution is initially selective and remains selective if the solution is renewed, whereas it becomes apparently congruent in non-renewed conditions. The congruent dissolution is ascribed to a pH increase (up to 7.0–7.5) which is hampered in the former conditions. The palagonite hydrated layer on the Icelandic basaltic glasses is made up of amorphous to crystallized clay-like materials. The chemical composition of palagonite is close to that of the intergranular clayey material, thus, it is inferred that in most cases no significant chemical gradient exists in the solution between the reaction zone, namely the glass/palagonite interface, and the intergranular solution. We conclude that the dissolution of basaltic glass under subglacial conditions is controlled by thermodynamics and that kinetic constraints, such as the diffusion of species through the altered layers, do not play a major role.

Alteration (fresh water) products (palagonite and zeolites) of a 12, 000–28, 000 year old basaltic glass from Hanauma Bay, Hawaii, were examined by electron microprobe analysis, scanning electron microscopy, and X-ray diffraction analysis. The palagonite, 0 to 300 microns thick, is depleted in Si, Al, Ca, Na, and K and enriched in Ti and Fe. Ca-rich zeolites are associated with palagonite that is slightly more depleted in Ca; all samples contain zeolites and low Al-palagonite. The rate of palagonitization (0.002–0.005 volume percent/year) is less than the experimentally predicted rate for basalt glasses due to the episodic presence of the percolating groundwater. The Si/AI ratio of the zeolites decreases with increasing pH of the groundwater as it passes through the vertical section. Zeolite paragenesis is (first to last): analcime (NaAlSi2O 6·H2O), phillipsite [(1/2 Ca, Na, K)(Al3Si5O16).6H2O], chabazite (CaAl2SiO16.6H2O).

The long term geochemical consequences of basaltic glass dissolution in fresh water at 0°C have been calculated with the computer code DISSOL. The clay minerals were represented by an ideal solid solution model (CISSFIT) able to describe variations in chemical composition of a clay phase in response to variations of the solution chemistry. The predicted mineral phases were iron hydroxides followed by kaolinite, TOT clays, chabazite and cli-noptilolite. These results are in reasonably good agreement with experimental results and observations of altered subglacial hyaloclastites from Iceland. The formation of secondary products are mainly controlled by thermodynamic constraints. Kinetic effects, such as diffusion in the near glass surface are not important.

An experimentally altered (1–14 days, deionized water, 90°C), synthetic borosilicate glass and naturally altered (104 years, fresh water, 0–100°C), basalt glass were characterized by analytical electron microscopy to evaluate the development of alteration layers. Layers formed on borosilicate glass are initially amorphous. With increasing age, nontronite, chlorite, septechlorite, and/or stilpnomelane appear to precipitate and grow within the amorphous matrix and on the surface of the layer. Crystals are enriched in Fe, Co, and Ni, and depleted in Si relative to the surrounding amorphous matrix. Alteration layers on basalt glass are texturally and mineralogically similar to layers formed on borosilicate glass, but the degree of crystallinity is less.

A considerable range in alteration rates of basaltic glasses in various environments has been reported in previous studies. However, these studies paid only cursory attention to the environmental conditions under which the glass had been altered. In this study, the alteration of basaltic glasses was investigated and the environmental conditions and the alteration rate were discussed. Two sample ages were represented: 280 years and 2800 years. Basaltic glasses and their alteration layers were analyzed by electron probe microanalyzer (EPMA) and the thickness of the alteration layers were measured by scanning electron microscope (SEM). The ground water collected near the sampling point of Zunazawa Scoria and the pore water of both samples were analyzed. The alteration temperature and flow rate of water are estimated to be about 13t and 0.2 1/cm2/y respectively on the basis of meteorological data. The alteration layers of young aged basaltic glasses in freshwater conditions are similar to those of leached borosilicate glasses. The alteration rates of these basaltic glasses are estimated to be several um/1000y. The elemental concentrations in the ground water can be roughly explained as the result of leaching of the glasses.

The kinetic results of corrosion experiments of glass R7T7 in brine 2 (high Mg and Ca concentrations) and brine 3 (high Na, low Ca concentrations) in this study are similar to those reported for the corrosion in brine 1 (high Mg concentration, no Ca) in [1]. Further support is lent to the idea that corrosion can be described, by a transient process that depends on the Si concentration in solution (prior to arriving at steady state Si concentration) and another one where this dependence vanishes or becomes small (after constant Si concentration is reached). This is analogous to what is known about corrosion of glass R7T7 and of other glasses in deionized water [4].

Borosilicate glasses are proposed as waste form for the final disposal of High Level Waste (HLW) in salt formations by the Federal Republic of Germany. To check the safety of this waste form the corrosion of the non-radioactive borosilicate glass R7T7 in salt brines has been studied in collaboration with the Hahn-Meitner-Institut, Berlin. Temperatures up to 190°C and S/V ratios (surface area to solution volume ratio) between 3.3 and 1000 m were applied. S/V was varied by the addition of glass powder. The results presented here concern the characterization of the surface composition of the corroded glass (corroded at 190°C in a brine containing high amounts of MgCl-, called Q brine in this paper) after removing the reaction product layer. The corrosion conditions were chosen to obtain samples typical of long term (silica saturation) corrosion. The surface composition was analyzed quantitatively by X-ray photoelectron spectroscopy (XPS) by using external standard glasses with a composition similar to R7T7. Additionally scanning electron microscopy (SEM) was employed.

R7T7 glass behavior was investigated in contact with eleven different materials including smectites, bentonites, illites, granite and sand. The experiments were conducted at 90°C under atmospheric pressure with a 4 cm−1 SA/V ratio using two grams of material in double distilled water under static leaching conditions without renewal of the leachate. For each material a control test was conducted under the same conditions but without the glass specimen, as well as a series of tests on glass alone, without the environmental material. Kinetics studies lasting up to 364 days were performed on four selected materials.

The test results (mass loss, ICP analysis of the leachates, SEM and TEM observations) showed relatively slight glass alteration in contact with sand, granite and one bentonite (i.e. the same order of magnitude as glass alone in deionized water), but greater alteration of glass in contact with all the other clays.

The significant differences in glass behavior in contact with environmental materials can be accounted for by the behavior of the latter in solution: activated bentonite releases very large amounts of silicon and sodium into solution while smectite seems to remove silicon from solution.

R7T7 glass samples were tested to determine their sensitivity to variations in the chemical composition and in industrial scale operating parameters. Variations investigated included the composition of the feed solution and the glass frit, the frit/glass ratio and the glass melting temperature. The iniportant properties of the resulting glasses were measured. Permissible variation ranges defined on the basis of the results obtained ensure that the glass properties remain acceptable compared with the reference glass.

The corrosion experiments of crushed, radioactive waste glass at Studsvik within phase V of the JSS Project are described. The experiments were performed with high S/V ratios (1100 and 4000 m−1) which resulted in silica saturation already after the shortest reaction time studied (91 d).Thee long term leach rate of the soluble elements is about 10−3 g.m d−1, both in the presence and absence of bentonite. In the presence of bentonite + magnetite the leach rate is higher, about 0.007 g.m1.d−2. The release rate to the solution of Cs is one order of magnitude lower and that of Pu-238 + Am-241 three orders of magnitude lower than the release rate of the soluble elements Mo and B. Most of the Pu-238 and Am-241 fraction, found in the solution, is present as colloids.

Fully radioactive waste glass of HLLW generated in Tokai Reprocessing Plant was produced by using a small Joule-heated ceramic melter at the Chemical Processing Facility (CPF), Tokai Works, PNC, Japan. Static leach tests using the glass were carried out in doubly distilled water for a period of up to three months at 30°C and 90 °C. Gamma spectrometry was performed to measure 134Cs, 137Cs and 60Co in the leachates. After the chemical separation of Cs isotopes in the leachates, it was newly possible to measure 125Sb, 144Ce, 154Eu and 155Eu. Alpha spectrometry was performed to measure 239Pu+240Pu, 241Am and 244Cm in the leachates. Adsorption of the actinides on the wall of the leach vessel was recognized in all cases. Leaching behavior of the measured radionuclides were evaluated by considering mass balance of the radionuclides which were leached.

The corrosion behaviour of Magnox Waste (MW) glass in deionized water has been investigated under the aegis of the Japanese-Swedish-Swiss (JSS) project. The studies carried out at PSI and BNFL included static leaching and corrosion layer characterization. The results were used by HMI to derive the modelling parameters for the MW glass with the PHREEQE/GLASSOL codes. The silica concentration was found to strongly influence the corrosion rate of the glass. The MW glass selectively retains Cs due to the formation of a Cs-Mg-Al-Si-O-rich layer on its surface.

Using doped glasses, the concentrations in aqueous solutions of Tc, Np, Pu and Am have been measured, over periods of up to 12 months, in the presence of various possible components of a future repository for vitrified high-level waste. The experiments were carried out under the reducing conditions anticipated in a repository. One possible repository design would incorporate a 20cm thick iron ‘overpack’ round the glass cylinders and a backfill of Ordinary Portland Cement (OPC) and Pulverised Fuel Ash (PFA). With this combination, the concentrations of Tc, Np, Pu and Am in the ‘repository’, after an 0.45um filtration, were respectively 7×10−9M, 7×10−10M, 2×10−12M and 2×10−12M i.e 0.1, 4, 1 and 45 times the ‘Limiting Concentration’ which, if present in drinking water, would lead to an annual dose of lmSv to someone drinking 2 litres per day.

A variety of surface sensitive techniques has been used to assess surface layers formed on SRL 165 waste glass specimens from the Stripa burial study and from a laboratory study designed to simulate conditions of the Stripa environment. The laboratory study included static and slow flow tests carried out for periods up to 6 and 24 months, respectively. Comparisons of leached layers formed in the two studies are based upon secondary ion mass spectrometry (SIMS), Fourier transform infrared reflection spectroscopy (FTIRRS), and scanning electron microscopy-electron microprobe (SEM-EMP) analysis.

Results demonstrate that surface layers having similar trends in elemental profiles were developed in both lab and field tests. It was also found that both sets of samples showed similar changes in FTIRR spectra. One significant difference between the lab and field samples was that surface alteration, as indicated by changes in FTIRR spectra and leached layer thickness (from SIMS), occurred more rapidly in the lab tests.

Licencing procedures for the disposal of vitrified HLW require that there should be no unforeseen changes in the product properties prior to and after emplacement of the waste. Such changes could occur upon deliberate or accidental self-heating of the glass up to or even beyond the glass transformation temperature, Tg, over extended periods of time. Hence, an evaluation of the consequences of excessively high temperatures in the waste form is necessary.

Boulos et al. [1], derived a mathematical expression predicting the minimum time needed for a glassy system to experience fractional crystallization, not exceeding a few volume percent at temperatures lower than the glass transition temperature, T. The range of applicability of this expression is extended to include crystallization processes governed by either homogeneous or heterogeneous nucleation and by diffusion controlled one-, two-, and three-dimensional crystal growth. This study demonstrates that the Boulos et al. expression is fairly independent of the nucleation, growth habit and dimensionality of the crystallization process. Moreover, the inherent theoretical as well as experimental limitations of the technique are reviewed and discussed.

The calculation for evaluating this minimum time limit has been applied for an alkali ferrisilicate alpha-waste reference glass. It is found that, for this system, a barely observable crystallized volume fraction (less than 5 % vol) will be reached at 25°C only after storage for more than 10 years. This time period is certainly beyond the times to be considered for evaluations of geological disposal. In addition some preliminary calculations of the minimum time limit for two types of borosilicate high level waste glass forms (HLW) are presented. These indicate safe time limits of the order of 10–100 years at 450°C, suggesting that crystallization effects cannot be ignored if the assumptions made in the calculations are valid.