Technical Report NTB 91-10

In several national radioactive waste management programmes, it is the intention to isolate Iow- and intermediate-level wastes (L/ILW) in cementitious matrices or in repositories backfilled with cementitious materials. In addition, some countries intend to dispose of long-lived, intermediate-level radioactive waste (so-called TRU) in a cement-dominated environment in a part of the high-level waste (HLW) repository. Cement and concrete are chosen predominantly for their favourable engineering properties, although the low solubilities of most radionuclides of concern in the hyperalkaline cement pore waters plus the high sorption characteristics of cementitious backfill are also important contributions to the safety case.

Models of the interaction of the host formation groundwater with cement/concrete suggest that KOH and NaOH leachates will be produced initially, followed by a longer period of Ca(OH)2 buffered leachates. Such solutions will interact with the repository host formation and may well alter the rock properties, especially in the vicinity of the repository. This may result in local alteration of the original retardation qualities for which the repository site was originally chosen (including affecting the form of water-conducting fractures or the nature of sorbing minerals) and is therefore of considerable interest from a safety assessment viewpoint.

To date, some of these processes have been studied in the laboratory but, as always, the results remain somewhat ambiguous due to the necessarily short timescales over which the experiments are conducted. This is compounded by the fact that, to achieve results on a reasonable timescale, most experiments have to be conducted at unrealistically high temperatures, thus making extrapolation of the results to repository relevant conditions even more problematic than usual.

An additional method of assessing models of repository behaviour does, however, exist: the study of appropriate natural systems which, in some way, are analogous to the processes or systems present in a repository. This approach (usually termed a natural analogue study) has the advantage of being able to study the processes of interest in action over more realistic timescales and, where the analogue site is carefully chosen, under more realistic conditions than can be produced in the laboratory. Although natural analogue studies now have a well established role in the safety assessment of radioactive waste repositories, it is only rarely that analogues have been examined in terms of a cementitious repository. One notable exception to date was the Nagra and UK Nirex joint funded Oman project, where geochemical thermodynamic codes were tested and the behaviour of microbes in a hyperalkaline system was assessed. The results of this study were encouraging enough to persuade Nagra and UK Nirex, this time in association with Ontario Hydro, of the value of continuing such natural analogue work; this led to the present study at Maqarin, in northwest Jordan.

This site differs from that in the Oman ophiolite in that the hyperalkaline groundwater chemistry is controlled by naturally occurring cement minerals and gels, thus making the analogy with a cementitious repository even better. In addition, the site contains relatively high levels of U and other elements of interest to the safety assessment of a radioactive waste repository.

Nagra, Ontario Hydro and UK Nirex jointly funded a study of a range of processes of relevance to the performance of a cementitious repository, including:

• The interaction of hyperalkaline groundwaters with an unaltered host formation; this was the first study of its kind which focussed on a process directly analogous to the potential reaction of hyperalkaline leachates from a cementitious repository with the repository host rock
• The evaluation of the solubility and speciation of trace elements in hyperalkaline waters in order to test databases of thermodynamic geochemical codes, especially for those elements of interest to performance assessment (including U, Th, Ra, Pb, Se, Ni and Sn) 
• The assessment of the significance of microbiological activity in hyperalkaline systems, thus establishing the likely impact of such activity in the near-field of a cementitious repository
• The measurement of colloid populations in hyperalkaline groundwaters as a guide to the role of colloids in and around the near-field of a cementitious repository, this being the first such study of its kind
• The evaluation of grout carbonation reactions of relevance to ¹⁴C attenuation in cement engineered barriers of a HLW (spent fuel) repository.

In the case of the interaction of hyperalkaline groundwaters with the host formation, only preliminary data are presented due to the necessity to first fully characterise the cementitious source material in some detail. Nevertheless, initial data presented here and in an associated report (MILODOWSKI et aI., 1992) indicate clear interaction of the hyperalkaline groundwater with the host rock. The full implications of the results are not yet clear, but a further study is currently underway on additional material from Maqarin.

The thermodynamic database testing exercise proved to be very successful, with several clear database inconsistencies being identified. Further testing was also carried out for Ni and Sn, following re-assessment of the databases, via comparison with well controlled laboratory experiments. In general, all of the databases behaved in a conservative manner (i.e. overestimating solubilities) even though the mineral phases included in the databases were a poor representation of reality. This, in itself, is reassuring from the safety assessment viewpoint, but perhaps more important is the fact that, in most cases, the modellers were able to explain precisely why the differences between the code predictions and reality occurred, thus further increasing confidence in the databases used. One significant failure was the case of U, where the various databases disagree significantly on the U speciation and solubility and none predict the behaviour of U in the natural system particularly well.

The microbiological study indicates that nutrient availability, rather than the high pH conditions, appears to control microbial activity. This has important implications for those repository designs which cannot be specifically shown to have low microbial activity levels due to limited availability of nutrients.

Degradation of cement may enhance colloid populations in and around the near-field of a cementitious repository, although the potential of cement to produce colloids has actually been little studied. The data produced here are thus of great importance as they represent the only known results for colloids in aged hyperalkaline groundwaters. The data indicate low colloid populations in comparison with other sedimentary groundwaters, in spite of the fact that some types of colloids should be stable under hyperalkaline conditions and that the relatively open fractures at the Maqarin site would presumably offer little chance of significant colloid filtration.

Although little is known of the chemistry of the colloids, it is at least certain that there is little U associated with them. This is contrary to previous predictions (observations at Oman) of colloidal U under hyperalkaline conditions.

It should be noted, however, that the colloid populations measured here may be low due to precipitation during sample storage and that the data should therefore be treated as preliminary until the work can be repeated with an in-situ colloid sampling rig.

The grout carbonation study was carried out as a trial effort to evaluate the potential of studying natural systems to complement laboratory-based experiments. The results indicated that carbonation has been extensive in some areas at Maqarin and that the site could therefore be studied as an analogue to repository carbonation reactions and ¹⁴C retardation.

Taken overall, the Maqarin natural analogue project has produced valuable (and, in some cases, unique) data on hyperalkaline leachate/repository host rock interaction, microbial and colloid populations in hyperalkaline groundwaters and on the applicability of current databases for thermodynamic geochemical codes used in assessing radionuclide releases from a cementitious repository. In some cases, the results of the study have already been included in repository safety assessment calculations and, as the project co-funders move towards constructing their own cementitious repositories, will probably be used even more so.