The Waste Isolation Pilot Plant: A Potential Solution for the Disposal of Transuranic Waste (1996)

As background for preparing its review of the actinide source term (AST), the AST subcommittee (Clark, Ewing, and Krauskopf) was supplied with abundant literature, including both in-house documents and articles from refereed journals that were considered by the Department of Energy (DOE) to be relevant to the Waste Isolation Pilot Plant (WIPP) program—many of which are listed at the end of Chapter 5. In addition, the subcommittee visited Los Alamos National Laboratory (LANL) in 1994 to view the proposed experiments with real and actinide-spiked waste (these are described as the Source Term Test Program [STTP]). In 1995 and early 1996, DOE scientists and contractors met with the AST subcommittee several times (March 6, April 27-28, and June 29, 1995; and February 14, 1996) to present the rationale for the AST program and some data from experiments completed in the past, and to describe experimental work in progress or still being planned (Novak, 1995a, b).

As late as April 1996, the subcommittee was still receiving revised versions of the proposed test plans, but no results of recent experiments or of the STTP program at LANL were available for review. The subcommittee was given a conceptual outline of the use of actinide source term data in the performance assessment (PA; Ruth Weiner, Sandia National Laboratories, personal communication to S. Clark, January 18, 1996); however, in the absence of experimental data explicitly used in the performance assessment, it was not possible to review this part of the program. Substantive review of AST results must await completion of at least a major part of the proposed work. To the extent that the actinide source term is an important parameter in the performance assessment and the determination of compliance, it will be necessary to review carefully the final data package, the selection and use of conceptual models, and the final incorporation of data and models into the performance assessment.

The remainder of this appendix discusses specific issues related to some of the tests planned as part of the DOE experimental program for the AST. The test plans may have changed after they were presented to the subcommittee and even before the experimental program was initiated; thus, these comments are applicable only to those portions of the test plans that have not changed.

The principal laboratories, principal investigators, and status of various components of the AST program are listed in Table E.1. This provides an indication of the materials that were reviewed by the AST subcommittee.

The application of Pitzer parameters (Pitzer, 1973, 1977) specifically to concentrated saline solutions has been described by Harvie et al. (1984) and by Felmy and Weare (1986). The methods of calculating solubilities suggested by these authors have been applied to the WIPP brines by DOE researchers. Particular advantages of using the Pitzer approach, rather than other means of correcting tabulated thermodynamic data to conditions of concentrated electrolytes, include the following:

1. The model is designed specifically for the geochemical conditions anticipated at WIPP and is well described in the peer-reviewed literature.
2. Data are available for many of the constituents in brine, making additional experiments unnecessary.
3. The model provides for complex interactions among more than just two species, reflecting the situation anticipated in highly saline media.

An additional general cautionary note is in order regarding the estimation of solubilities in concentrated solutions with the use of the Pitzer parameters. The thermodynamic formalism involved in such estimates is based on the assumption that equilibrium exists between material in solution and solids in contact with the solution—in other words, that all reactions leading to formation or dissolution of solids go rapidly to completion. There is no assurance that these assumptions are correct for the experiments or for in situ WIPP conditions. In fact, some of the experimental work has shown that concentrations in a solution, after a solid has precipitated, may change slowly over the course of months or a few years as the solid alters its form, for example, from an amorphous precipitate to a crystalline phase (Nitsche et al., 1994).

Because changes in the solid are generally in the direction of greater and greater stability (Ostwald's step rule), and thus lower and lower solubility, such changes commonly are thought to ensure concentrations of actinides lower than those first measured, and hence not to be important in estimating cautiously conservative values for actinide solubilities. As a general rule this is true; however, for solutions as concentrated, and with compositions as complex as WIPP brines, exceptions can readily be imagined.

Resolving issues such as this will require extending some of the solubility experiments well beyond their scheduled termination dates. These experiments can be conducted at the bench scale under carefully controlled conditions, so that the fundamental aspects of actinide solubility in brine can be determined. Such experiments may be needed for interpretation of the more complex experiments performed as part of the STTP at LANL. These longer-term experiments should not delay compliance of the repository if all else is favorable, but they represent one example of the kinds of research efforts that are recommended to continue after wastes are first brought to the repository.

The following are specific concerns about details of the planned experimental work on actinide solubility:

• Most aspects of the experimental program and assumptions in the performance assessment are based on thermodynamic models; however, reaction kinetics may be important in determining the actinide concentrations measured in experiments or modeled in the performance assessment.
• One question is the suitability of the specific-ion interaction theory (SIT), activity coefficient formalism, which has been used in the test plan for the Alternative +6 Actinide Model (Grenthe and Wanner, 1992). This approach for estimating concentrations of actinides in the +VI oxidation state is not scientifically defensible, because SIT theory is designed specifically to extrapolate to zero ionic strength; it is not applicable at ionic strengths higher than 3.5 molal; and it is typically used for applications involving a single dominant electrolyte, rather than the complex brines anticipated at WIPP.
• Another question regards the use of experiments planned in the STTP as direct challenges to the submodel of solubilities. The STTP involves measuring actinide concentrations in solutions formed by allowing waste containers to stand for long periods in contact with brines having compositions similar to those at WIPP; in other words, determining concentrations under laboratory conditions set up to duplicate, as nearly as possible, the conditions expected in the WIPP repository. Such experiments will certainly provide useful qualitative data on actinide concentrations, and this program is applauded. However, the assumption that experimental conditions can be made similar in all respects to repository conditions over time seems tenuous at best. Experiments that challenge the thermodynamic model will require not only careful solution analyses as a function of extended periods of time (more than a year), but the identification of the actinide-bearing solids that are in equilibrium with the solution. This will be difficult, because many of the solids are poorly characterized and may have fairly complex compositions.
• Although the oxidation state model (the assumption that the chemistry of a given oxidation state is similar for all of the actinides) is an appropriate beginning to a difficult problem, deviations from the oxidation state analogy are well known in natural and experimental systems. Substantial experimental verification will be needed to establish the limits of this analogy.

Although actinide concentrations in WIPP brines are expected to be low, the possibility of transport via colloids must be considered one means by which actinides may reach the accessible environment (Kim, 1994). There are two issues: (1) whether colloids will form in a concentrated brine, and (2) whether these colloids can transport significant quantities of actinides effectively.

In designing their research effort, Papenguth and Behl (1996a, b) distinguish four kinds of colloidal particles:

1. mineral fragments, hard-sphere or hydrophobic particles that may contain actinides or have sorbed actinides on their surfaces;
2. actinide intrinsic colloids, particles consisting of actinide macromolecules that may mature on standing into mineral–fragment type colloids;
3. humic substances, hydrophilic or soft-sphere particles that readily sorb actinide ions from solution; and
4. bacteria, which may themselves be colloidal-sized particles and may contain actinides as part of their structure or sorbed on their surfaces.

All types of colloidal particles are relatively unstable in strong electrolyte solutions, and some of the experiments are designed to show what concentration of electrolyte (''critical coagulation concentration," or c.c.c.) is required to produce destabilization and precipitation.

The procedure is simple: small amounts of a simulated Salado-Castile brine are added to samples of colloid until coagulation is noted. Mineral fragments and mature actinide intrinsic colloids are expected to be the most sensitive to electrolytes (i.e., to have the lowest c.c.c.), whereas humic substances and bacteria will not precipitate as readily. For these latter two types, DOE plans more elaborate experiments to explore the effects of pH and the addition of specific ions and organic solutes.

The simple laboratory experiments on colloids will be supplemented with transport experiments designed to show the extent to which various kinds of colloids may be retarded, by precipitation or sorption, if the solution containing them travels through the dolomite of the Culebra Dolomite. Solutions carrying the colloids will be made up to simulate brine concentrations expected in the Culebra, and samples of the solutions will flow through both crushed dolomite and intact-core columnar dolomite. Transport experiments with bacteria as colloidal particles will be especially important, because microorganisms may serve either to increase radionuclide transport by sorbing abundant actinide ions or to decrease transport because the organisms carrying sorbed actinides are themselves strongly sorbed on rock surfaces (Marshal, 1976).

Again, to the extent that colloid formation and transport remain important issues in the performance assessment, the proposed experimental programs may well extend beyond their planned end dates. Long-term experiments on colloid stability and transport may be required even into the operational phase of the repository.

During transport, actinide concentrations in brine may be lowered by sorption onto rock and mineral surfaces. Such phenomena can be studied by batch experiments to determine K_d values, column experiments with actual rock units from the relevant geologic horizons, or field-scale tracer tests. The limitations of laboratory-scale tests are discussed briefly in Appendix F as part of the discussion of regional hydrology. The inherent weakness of the concept of a retardation factor determined in laboratory experiments, combined with highly uncertain and variable values for the retardation, requires careful analysis of the laboratory data and some field-scale verification. The proposed suite of DOE experiments has been planned carefully and should give a good indication of the effectiveness of sorption in controlling actinide concentrations; however, no results of the experiments completed to date were available.

Concerning the column experiments, the sorbing materials to be used are cylinders of Culebra Dolomite, 10-50 cm long, cut out of the bed with a drill bit and mounted vertically in metal containers. Various

samples of dolomite will be chosen to ensure adequate representation of different permeabilities, different grain sizes, and different proportions of clay mineral impurities. It is emphasized that the column experiments (plus, hopefully, the field experiments) should be carried on long enough and with sufficient variation in conditions to ensure an adequate replica of the natural situation. In particular, the amount of sorption provided by the corrensite coating on parts of the Culebra dolomite requires careful study, a point on which the State of New Mexico Environmental Evaluation Group (EEG) has expressed considerable skepticism regarding the DOE approach.