Actinide Source Term
In the performance assessment (PA) of the repository described in Chapter 2, compliance with the Environmental Protection Agency (EPA) requirements is discussed relative to two scenarios, E1 and E1E2 (Figures 2.4 and 2.5), in which pressurized brine from below the Salado Formation flows into the repository and then either directly to the surface or through the Culebra Dolomite to the accessible environment. The amount of radioactive material that might move into the environment obviously will depend on the amounts of the five actinide elements (plutonium [Pu], americium [Am], neptunium [Np], uranium [U], and thorium [Th]) that the flowing brine can transport. This quantity is the actinide source term (AST), defined more specifically by scientists at Sandia National Laboratories as the cumulative amounts, concentrations in solution, and chemical nature of the radioactive materials that conceivably could be moved in solution or suspension from the Waste Isolation Pilot Plant (WIPP) repository into the environment.
Clearly, an estimate of the AST is an important part of any performance assessment. Although the AST program includes an evaluation of five actinides, the most important issue for periods greater than a thousand years is the fate of the more abundant, long-lived isotopes of plutonium (e.g., Pu-239 and Pu-240; see Table 1.1). The study of the other actinides is important because they provide a basis for the oxidation state model in which plutonium may assume a variety of oxidation states. As an example, the following elements may be used to model the behavior of plutonium: Am(III)1 for Pu(III), Th(IV) for Pu(IV), Np(V)O2+ for Pu(V), and U(VI)O22+ for Pu(VI) (Novak, 1995b).
Evaluation of the AST depends principally on the following:
- the total inventory of radioactive material (Table 1.1) and its decrease due to radioactive decay as a function of time;
- conditions within panels of the repository and along the route of transport (e.g., the volume of brine exposed to the waste and transported out of the repository); and
- the timing of potential release events.
The challenge in making a reasonable estimate of the AST is to gain an understanding of the coupling between the chemistry of the system and the boundary conditions and assumptions (e.g., redox potential, flow rate, gas generation, distribution of sorbing clays, presence of ligands) used in the performance assessment.
As noted in Chapter 2, performance assessments to date (e.g., that for 1992) confirm that if the repository, after completion, is left undisturbed by human activity, the presence of brine in contact with waste will not lead to migration and release of appreciable radioactivity beyond the WIPP boundary. The geologic barriers, combined with the planned engineered barriers discussed in Chapter 4, will be sufficient to prevent radionuclide escape. Thus, for the undisturbed case, the magnitude of the AST is of no consequence for assessing compliance.
If the repository is disturbed at some time in the future, however, it is possible that some radioactive material might escape. Therefore, estimates of possible amounts of released radioactivity are required for detailed assessment of the performance of the repository under various assumed conditions. In any such assessment, the magnitude of the AST is an important parameter.
In making a performance assessment of the scenarios described in Chapter 2, Department of Energy (DOE) scientists have undertaken an elaborate program of experimental work (Novak, 1995a) to ascertain the
possible range of actinide concentrations in the brine as it moves from the panels in the repository to the WIPP boundary. This chapter reviews DOE's proposed experimental program (Novak, 1995b; Papenguth and Behl, 1996a, b) on the AST.
Although the committee has been given a good overview of the work on the AST that has been completed or is still in progress (see Appendix E), much of the research on the source term is not yet complete, and some important results of the program will not be available for review in the immediate future (e.g., experimental studies of Pu speciation in brine). Consequently, this chapter should be viewed as a review of work in progress and cannot serve as an evaluation of the quality or applicability of the final experimental results.
Actinide Source Term Model
Brine that has been in contact with transuranic (TRU) waste (described in Chapter 1) likely will contain actinides as compounds or ions in solution and as colloidal particles. As the brine moves through rock units, the amount of actinide transported will change. Actinide concentrations in brine may be expected to decrease owing to precipitation of actinide-bearing phases or sorption on mineral surfaces; alternatively, actinide concentrations could increase in response to the formation of complexes in solution. To determine the amounts of actinides that might reach the compliance boundary, estimates of the following are required for each element of concern (particularly plutonium):
- the concentration in solution, as controlled by the solubilities of the actinide-bearing phases and the kinetics of the dissolution process;
- the amount that could form and be transported as colloidal particles; and
- the amount that could be removed from solution by sorption on the surfaces of dolomite or other minerals (e.g., the corrensite-lined fractures of the Culebra Dolomite) as the brine moves from the repository into the surrounding rock.
DOE analyses (Sandia National Laboratories, 1992, 1995; DOE, 1995a) consistently have identified these three quantities as important to the long-term performance of WIPP. The same three quantities also have been identified as important by EPA and the Environmental Evaluation Group (EEG; Neill et al., 1996).
Estimates of these quantities can be approached theoretically with models using tabulated chemical data and basic chemical principles, and empirically through either carefully controlled laboratory experiments or measurements under actual field conditions. Each approach has advantages, and all have weaknesses. Although well-controlled experiments using ideal or model systems may provide a fundamental understanding of the chemical processes, conditions of the experiments may be simplified to the point that their direct applicability to the larger-scale, more complex conditions in a waste panel is not clear. Field-scale experiments (e.g., sorbing tracer tests) or experiments with actual waste (e.g., Source Term Test Program [STTP] experiments at the Los Alamos National Laboratory [LANL]) may be difficult to interpret, but they do provide qualitative information on potential chemical interactions and bounding estimates of actinide concentrations in brine. Confidence in the models used in performance assessment will rest inevitably on consistency in the results and the general applicability of the models to all results—experimental, theoretical, and field.
Three principal factors currently being investigated by DOE that determine actinide transport in brine—solubility, colloids, and retardation by sorption—are examined below. Only limited attention is being paid to other factors (e.g., kinetics and redox effects on the reactions) that may also finally influence the transport of actinides to the accessible environment.
Estimating actinide solubilities from published data is particularly difficult because the electrolyte concentrations in brines are very high, whereas most existing tabulated data apply to dilute solutions. This means that the thermodynamic formalism related to solubility must incorporate the use of correction factors. Of the numerous such factors that have been proposed, the generally accepted approach has proven to be the parameters developed by Pitzer (Pitzer, 1991). These are the correction factors used by DOE scientists
(see Appendix E for a more detailed discussion of Pitzer parameters).
Difficulties in determining actinide solubilities are compounded by the many gaps in published tables of thermodynamic data. DOE scientists are responsible for filling such gaps through their experimental work. A further complication is the fact that the solubilities of a number of actinide compounds are influenced by the formation of complexes, either with inorganic ions such as Cl- and CO32- or with C-H-O compounds from the organic materials present in much TRU waste. Add the difficulty of knowing or determining the relative proportions of the oxidation states of an actinide element in a given solution, and the complicated nature of the task that DOE researchers have set for themselves becomes apparent.2
Based on a review of the AST program, the committee has identified areas of concern that are described in Appendix E. Although the experimental approach chosen to estimate actinide solubilities is acceptable, it will be defensible only when supported by extensive experimental data, which were unavailable at the time of preparation of this report. When compared to published experimental data, current modeling efforts are not capable of conservatively estimating actinide solubilities. For example, Novak and Roberts (1995) found that their model for Np(V) predicted solubilities as much as an order of magnitude less than those determined experimentally. Thus, completing the experimental program in order to generate Pitzer parameters for all oxidation states of the actinides is essential for obtaining credible estimates of dissolved actinide concentrations.
Additionally, the estimate of actinide concentrations in a brine will depend on the solubilities of actinide-bearing phases. In most of the experimental programs, too little attention is being paid to identification of the solubility-limiting phases. These phases may be unusual in composition, and thus difficult to identify, because of the complex composition of the brine and the waste inventory. Many of the thermodynamic data required by geochemical codes to model actinide concentrations in solution are either absent or contradictory. A key factor in the final result will be the "fairly subjective" choice of the solubility-limiting phase (McKinley and Savage, 1993).
In the repository environment, many processes can lead to the formation of colloidal suspensions (Laaksoharju et al., 1995), some of which may include actinide elements. In fact, the transport of actinides by colloids greatly can exceed that anticipated from models based solely on the solubility of actinide-bearing phases (Kim, 1994). Recent results (Reed et al., 1994) have demonstrated the important potential role of colloids in retaining plutonium in suspension, with subsequent transport potentially being greater than expected (Ibaraki and Sudicky, 1995).
When a solid actinide compound is precipitated from solution, it may form suspended colloidal particles rather than a crystalline solid, or actinide ions may attach themselves to colloidal particles formed in other ways, either as inorganic particles or as complex compounds derived from organic constituents of the waste. Some of the varieties of bacteria that grow in the waste solution may themselves be particles of colloidal size and may contain sorbed actinides.
Colloid particles (usually in the size range of 10-9 to 10-5 m) generally have an electric charge, either positive or negative, that helps keep the particles apart and prevents their flocculation as a precipitate. In concentrated electrolyte solutions such as WIPP brines, the charge on particles is neutralized readily and the colloids are generally unstable. This is especially true for colloids of inorganic material (either colloids consisting entirely of actinide molecules or inorganic
compounds with sorbed actinides) and less so for colloids in which the particles consist largely of organic material. The former are called "hard-sphere" or hydrophobic colloids; the latter, "soft-sphere" or hydrophilic colloids. In general, hard-sphere colloids readily flocculate in a strong electrolyte and, hence, are ineffective in mobilizing actinides in the WIPP environment. In contrast, some soft-sphere colloids (including colloidal-size bacteria) remain stable even in concentrated solutions and so can serve as effective transporters of actinides. It is emphasized that colloid stability is a complex function of many other variables, such as acidity and alkalinity, the nature of sorbed ions, and the kinds of organic matter and electrolytes present.
Bacteria may be the most effective vehicle for transport of actinides in the Culebra Dolomite. Their total numbers, sizes, and morphologies in indigenous Culebra brines will be determined, as well as the increase in their numbers if nutrients are added from organic material in the waste. Bacteria can influence the behavior of actinides in several ways:
- Their metabolic activity will tend to keep conditions slightly acidic, thus helping to determine the oxidation state of actinide elements and so affecting their solubilities.
- By sorbing ions from solution, bacteria obviously can aid in the transport of actinides.
- The well-known tendency of microorganisms to adhere to surfaces (Marshal, 1976) may serve to retard actinide transport.
- If significant quantities of gas are produced as a result of microbial metabolism, this gas also may influence the movement of actinides.
In addition to bacterial activity, any corrosion process will influence the pH and/or the gas content of the brine and thereby influence solubilities and subsequent transport of actinides.
The possibility of substantial transport of actinides in colloidal form has been recognized only fairly recently (Kim, 1994) and has spurred much effort by DOE scientists to investigate the complexities of colloid behavior in strong electrolyte solutions such as WIPP brines (Papenguth and Behl, 1996a). From the proposed experimental program, DOE scientists expect to learn whether there is a serious threat of active transport of actinides by colloidal particles and what measures might be taken to modify conditions in the WIPP repository to reduce this threat.
The DOE study began with a comprehensive review of recent literature on the formation and stability of colloids in electrolyte solutions and now includes detailed plans for experiments designed to answer the many remaining questions about this mode of actinide transport (see Appendix E). This effort is commended by the committee, with the expectation that these results will provide much of the new information needed to gain a deeper understanding of colloids as part of the actinide source term. The experimental work has started and is scheduled for completion soon (see Table E.1), but no results have been presented to the committee.
If, as postulated, the dissolved and colloidal actinides that make up most of the source term move with the brine into the accessible environment through the Culebra Dolomite, their transport will be retarded to some extent by sorption on the surfaces of the dolomite and minor amounts of clay minerals (e.g., corrensite) contained in this member. Thus, the amount of radioactive material that might actually reach the accessible environment depends on the effectiveness of this sorption. Accordingly, the amount of sorption is a critical factor in any attempt to estimate whether or not the WIPP project complies with EPA regulations.
As with solubility, the study of retardation by sorption is made difficult by the high concentration of dissolved material in the brine. For dilute solutions, determining the amount of material sorbed is straightforward and experimentally simple. It is commonly expressed in terms of Kds, or distribution coefficients, which are merely the ratios of amounts sorbed to amounts left in solution. Complexities arise with concentrated solutions, however, because (1) many substances are competing for places on any given sorption site, and (2) the nature of the sorbing substance is influenced by its speciation and the chemical conditions (e.g., acidity and alkalinity) of the solution (see discussion in Appendix E). Additionally, batch Kds measured in the laboratory rarely have been
used successfully to predict sorption on a larger field scale.
Because simple Kds are of little use as predictors, amounts of sorption are best determined by direct experiments in which samples of natural or synthetic brine are allowed to flow through samples of the intended sorbent material, and actinide concentrations are measured before and after the passage (see Appendix E). Such experiments have 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 have been presented to the committee. The experimental results will have to be confirmed by results from actual field experiments. Throughout the history of the WIPP project, such studies ("sorbing tracer tests") have been planned, delayed, and canceled. The status of such tests at present is unclear. It will be difficult for DOE to take credit for sorption in the Culebra (or any other rock unit) in the absence of well-planned and completed field tests.
As noted above, the ambitious schedule DOE has set for completion of experimental work for the AST (1996) is unrealistic in that final, definitive values for some of the numbers being sought cannot possibly be obtained and evaluated critically in so short a time. Enough data may be generated to provide a reasonable basis for the proposed compliance application, but important details about the source term and its appropriate use in performance assessment will almost certainly require additional work. Strictly accurate final values are not essential for showing compliance with EPA regulations, but are nevertheless numbers for which DOE will find many uses later. The committee recommends that selected research programs be extended into the operational phase of the repository.
Summary And Discussion
The actinide source term can be envisioned as the entire mass of actinides destined for disposal at WIPP, partly dissolved and partly in colloidal suspension in a brine similar in composition to Salado Formation or Castile Formation brines (or a mixture of the two). That the entire mass would ever be dissolved or suspended is hardly likely, but a minor fraction could plausibly be mobilized at some time or times in the future. Determining how large this fraction might be requires knowledge of the solubility of each of the actinide elements in brine and its behavior as a colloid, plus data on the sorption of ions and colloidal particles onto mineral surfaces with which the brine may come in contact. Data of this sort, both from the literature and from experimental work, have been and are being accumulated in impressive amounts by DOE scientists and contractors.
Data on the source term, as noted previously, will have no importance if the repository remains undisturbed by human activity for 10,000 years because no appreciable radioactivity will likely escape to the surface. However, for any of the scenarios considered in Chapter 2 describing the results of possible human disturbance of the repository, knowledge about the source term is essential for estimating possible radioactive releases from various assessments of repository performance and, hence, for making a decision on the application for a certificate of compliance to open the WIPP repository.
The ongoing experiments in DOE laboratories and DOE-sponsored universities, if carried to completion, would begin to provide the necessary information. The present schedule will not allow all the work to be finished and evaluated before the Fall of 1996. Nonetheless, enough preliminary data will be obtained to make possible a reasoned judgment of whether WIPP meets the compliance criteria for the disturbed case.
Overall, the scientific program outlined by DOE for study of the source term is adequate, provided that the program is carried to completion. Because the program at this time consists largely of work planned or in progress, it has not been possible to critically review experimental results or to judge whether these results are used appropriately in the PA analysis. Much of the AST work has not been subjected to the type of peer review that is part of publication in scientific journals, and the committee recommends that such review be sought as the work progresses. Interaction between principal investigators in the AST program and modelers in performance assessment has been
disappointingly limited, and the committee recommends that this situation be corrected.
In the opinion of the committee, as soon as substantial results are available from the experiments—provided, of course, that the results show the expected low solubilities, which will permit calculation of low actinide releases based on some of the less improbable variants of the borehole scenario—then DOE will have the information on actinide solubilities needed to prepare the certification of compliance.
Experiments should continue for as long as is needed to obtain reliable values for actinide solubilities, colloidal transport, and nuclide retardation by sorption, a time during which the repository will be in full operation. The new data obtained will be important at the WIPP site as a basis for changes in operational procedures (e.g., use of backfill, improved seals, or reduced waste loading) and, more generally, for providing the fundamental understanding needed to predict actinide behavior in other situations. Actinide chemistry is important, for example, in the disposal of other kinds of radioactive waste and in the nationwide environmental remediation program that DOE has undertaken. Because the experimental work will provide basic data needed for a variety of purposes, the committee recommends that DOE continue this work beyond the scheduled completion dates in mid-1996.
There appears to be no reason that the further required work on the actinide source term cannot be accomplished simultaneously with the disposal of waste in the repository. This additional work will improve the scientific basis of the performance assessment and increase public confidence in the ability of WIPP to isolate TRU waste.
Conclusions And Recommendations
The committee offers the following conclusions regarding the actinide source term:
- For the undisturbed case, the magnitude of the actinide source term is of no consequence to the demonstration that WIPP can isolate waste from the biosphere effectively for many thousands of years.
- In the disturbed case, if it is necessary to rely on estimates of actinide concentrations in the brine, colloidal transport, or nuclide retardation by sorption, a well-documented, reviewed, scientific and technical basis for these estimates does not yet exist. With this in mind, the following observations and recommendations are made.
With the exception of specific scientific issues raised in this chapter (and detailed in Appendix E), the scientific program outlined by DOE is adequate if the proposed program is carried to completion.
Plutonium is the principal radionuclide of interest in any release scenario. Although some aspects of its geochemical behavior can be estimated by studies of other actinides using the oxidation state model, ideally, data for each oxidation state should be followed by experiments with plutonium.
Because the AST program at this time consists mainly of proposed work, the committee has not had the opportunity to review experimental results or to evaluate whether these results were used appropriately in the performance assessment analysis. At present, much of the AST work in support of the WIPP project has not been subjected to the type of peer review that is part of publication in scientific journals. This review process is the essence of quality assurance, and the committee recommends that scientific results of the AST program continue to be published in peer-reviewed journals.
The existing schedule (see Table E.1) will not allow for the timely (i.e., by the Fall of 1996) completion of the proposed experimental work, proper analysis of the results, or publication in peer-reviewed scientific journals. In addition to the STTP, long-term scientifically based experiments are crucial in evaluating the long-term extrapolations that are required by PA analysis. These experiments are not part of the present experimental program; however, they should be an essential aspect of the iterative PA process, which will be required even during the operational phase of the repository. The needs of the AST program are not unique to the WIPP project. DOE inevitably will need a strong program in basic actinide chemistry as part of its nationwide environmental remediation effort. Prudent and efficient planning could consolidate the needs of the WIPP program within a broader program of basic research in actinide chemistry.
The interaction between principal investigators in the AST program and modelers in performance assessment has been disappointingly limited. Thus, it still is not clear how the AST data (experimental or theoretical) will be used in the performance assessment. The PA analysis has not yet provided the type of guidance to the experimental program that is required in order to determine when the work is, in fact, complete.
Concerning the use of performance assessment, there is still no clear distinction between the uncertainty associated with the selection of conceptual models in the AST and the uncertainty that results from the experimental data. This distinction is essential in evaluating the impact of the AST on PA analysis and conclusions.
The use of experimentally determined batch KdS alone is not sufficient to justify the assumption of retardation of actinides during transport through the Culebra Dolomite. Field confirmation and/or a more fundamental understanding of sorption processes are needed.
Based on the current assessment of the WIPP site, the committee believes that the required work on the actinide source term (actinide concentrations in brine, colloid formation and transport, and nuclide retardation) can be performed simultaneously with the operation of the facility.