1. INTRODUCTION
To evaluate the postclosure performance of a potential monitored geologic repository at Yucca Mountain, a total system performance assessment (TSPA) will be conducted. A set of nine process model reports (PMRs), of which this document is one, is being developed to summarize the technical basis for each of the process models supporting the TSPA model. These reports cover the following areas:1.1 OBJECTIVE
The degradation and potential release of the radioactive waste placed into the potential Yucca Mountain repository are dependent on numerous features1, events2, and processes3 (FEPs) of the disposal system. Because of the different spatial and time scales and the large number of FEPs, they have been grouped into several modeling systems and components for modeling and discussion in reports. The overall objective of the Waste Form Degradation Process Modeling Report (Waste Form Degradation PMR) is to summarize the technical basis of the Waste Form Degradation model. The degradation of the waste eventually leads to mobilization of radionuclides and release into other components of the engineered barrier system (EBS) in the TSPA. Because a small fraction of the radioisotope in the waste can be transported through the EBS and the natural rock barrier, the source concentrations of radionuclides evaluated by the Waste Form Degradation Model can influence results of the TSPA analysis when the waste package (WP) is breached. This TSPA analysis, in turn, is an important basis of the Site Recommendation Consideration Report (SRCR) and, thus, designated as TSPA-SR in this report.1.2 PURPOSE AND SCOPE
Nineteen AMRs were written to document (1) various experimental results, (2) conceptual models and detailed numerical models, and (3) further abstraction of the experiments or numerical models and to define consistent model parameters pertinent to degradation of the waste form. The purpose and scope of the Waste Form Degradation PMR is to summarize the most important aspects of these 19 AMRs to provide a coherent view of the components of the Waste Form Degradation Model. More specifically, in order to satisfy several important requirements for the analysis for the TSPA-SR according to the DOE interim guidance (Dyer 1999) to the proposed U.S. Nuclear Regulatory Commission (NRC) regulation (proposed DOE 10 CFR 63.114 Interim Guidance [Dyer 1999]), Waste Form Degradation PMR purpose and scope are to briefly summarize:1.3 QUALITY ASSURANCE
The Quality Assurance (QA) program applies to this PMR and the AMRs that support it. The development of this analysis is conducted under activity evaluation Waste Form Analysis & Models (CRWMS M&O 2000b), which was prepared per AP-2.16Q, Activity Evaluation. The results of that evaluation were that the activity is subject to the Quality Assurance Requirements and Description (DOE 2000) requirements. This report has a formal planning document (CRWMS M&O 2000c). Development of this report did not involve the electronic management of data. The purpose of this ICN is to address DOE comments. This report was prepared in accordance with AP-3.11Q, Technical Reports and reviewed in accordance with AP-2.14Q, Review of Technical Products. The Waste Form Degradation PMR does not serve as a primary source for data, models, and codes described in the AMRs related to waste form degradation; this function is provided by the AMRs themselves. The AMRs supporting the Waste Form Degradation PMR were prepared in accordance with AP-3.10Q, Analyses and Models, including AP-3.15Q, Managing Technical Product Inputs, for data verification, AP-SIII.2Q, Qualification of Unqualified Data and the Documentation of Rationale for Accepted Data, for data qualification and AP-SI.1Q, Software Management, for management of software. The quality assurance documentation of the models, codes, and data is deferred to the AMRs cited in this document. The current quality status of data and software can be found in the Document Input Reference System (DIRS). This PMR includes the results from software codes and routines used in the supporting AMRs. However no software codes were used in the development of this PMR. The following statements are presented for information only. The only software code used in the AMRs was the EQ3/6 package, Version 7.2b, which was approved for QA work by LLNL and is identified as Computer Software Configuration Item (CSCI): UCRL-MA-110662 V 7.2b (CRWMS M&O 1998a) and its addendum (CRWMS M&O 1999c). The package was used both for chemical equilibrium calculations, and reaction path calculations (see Sections 3.2 and 3.7).1.4 RELATIONSHIP TO OTHER PROJECT REPORTS
The Waste Form Degradation PMR describes how information is used in the TSPA-SR analysis by summarizing the modeling components of the Waste Form Degradation Model. Although not directly coupled, the Waste Form Degradation Model, in turn, is related to the Waste Package Degradation Model and Engineered Barrier System Model and their corresponding PMRs: Waste Package Degradation PMR (CRWMS M&O 2000d) and Engineered Barrier System PMR (CRWMS M&O 2000y). The Waste Form Degradation PMR is also related to the Disruptive Events PMR (CRWMS M&O 2000an) since it requires waste inventory information. Eight summary/abstraction AMRs specifically support the Waste Form Degradation PMR (Figure 1.4-1). Other documents underlying these primary areas are mentioned within Chapter 3. Abstraction Level AMRs:1.5 OVERVIEW OF WASTE FORM DEGRADATION MODEL
The analysis for the TSPA-SR requires numerous models and analyses. Many of these models are components of larger models that, in turn, are models to system-level models. Rather than call all the parts "models," this report uses terms that identify the hierarchy of the models to help orient the reader. The system-level model, which uses total system simulation software, is referred to as the "TSPA-SR system-level model" or "TSPA-SR." The major parts of TSPA-SR are referred to as "models." The Waste Form Degradation Model is one such model. The models and analyses that are parts of the Waste Form Degradation Model are referred to as "components." (A possible alternative term is "submodel." However, not all the parts are models in the sense used by YMP but rather are analyses, so the term "components" was selected.) Although the Waste Form Degradation PMR has several broad purposes already listed, one important purpose is to describe the Waste Form Degradation Model in the TSPA-SR and its underlying components.1.5.1 Function of Waste Form Degradation Model
Within the analysis for the TSPA-SR analysis, the function of the Waste Form Degradation Model is to determine three outputs over time: (1) dissolved concentration, (2) reversible colloidal concentration, and (3) irreversible colloidal concentration of radionuclides (Figure 1.5-1). Numerous inputs are required for the Waste Form Degradation Model. Several inputs are intimately tied with the waste form degradation models such as initial cladding condition and thermodynamic data for radionuclides. They are summarized in this report. Other inputs are summarized in other PMRs. For example, the time dependent seepage flow into the WP and WP surface temperatures are summarized in the EBS PMR (CRWMS M&O 2000y). Corrosion rates of the inner stainless steel container and supports for waste inside the waste package are summarized in the Waste Package Degradation PMR (CRWMS M&O 2000d).1.5.2 Components of Waste Form Degradation Model
To determine the three output radioisotope concentrations, the Waste Form Degradation Model uses eight major modeling/analysis components to evaluate the degradation, deterioration, or alteration of the waste: (1) Radioisotope Inventory (described in CRWMS M&O 2000f), (2) In-Package Chemistry (described in CRWMS M&O 2000g), (3) Commercial Spent Nuclear Fuel (CSNF) Matrix Degradation (described in CRWMS M&O 2000h), (4) CSNF Cladding Degradation (described in CRWMS M&O 2000i), (5) DOE Spent Nuclear Fuel (DSNF) Degradation (described in CRWMS M&O 2000j), (6) High-Level Waste (HLW) Degradation (described in CRWMS M&O 2000k), (7) Radioisotope Dissolved Concentration (described in CRWMS M&O 2000l), and (8) Radioisotope Colloidal Concentration (described in CRWMS M&O 2000m) (Figure 1.4-1). The mathematical description of these eight components is one specific purpose of the Waste Form Degradation PMR. Another important purpose is to summarize the technical bases of the eight components. Generally, the primary technical bases or "foundations" of the eight components of the waste form degradation models are the CSNF, DSNF, and HLW degradation experiments described throughout the report.1.5.3 Major Models Communicating with Waste Form Degradation Models within TSPA-SR Analysis
For TSPA-SR analysis, the major model communication with the Waste Form Degradation Model is from the EBS abstraction models and the TSPA (Figure 1.5-2). The EBS abstraction models determine the water seepage quantity onto the waste form that is used by both the In-Package Chemistry Component and the Dissolved Radioisotope Concentration Components. In turn, the Waste Form Degradation Model determines dissolved and colloidal concentrations of radionuclides for the EBS. The Waste Form Degradation Model is also uncoupled from the Waste Package Degradation Model. Although conceptually, the Waste Package Degradation Model could calculate the rate iron oxides or other corrosion products are produced by the degradation of the WP and pass this information onto the Waste Form Degradation Model. Instead of this communication, the underlying in-package process model evaluated the resulting chemistry from a variety of corrosion rates and expressed this as uncertainty in the final in-package Chemistry Component.1.5.4 Linkage of Components in Waste Form Degradation Model
The components of the Waste Form Degradation Model are generally connected sequentially starting with the Radioisotope Inventory Component and ending with the Radioisotope Colloidal Concentration Component; however, some complications occur. First, the rates of degradation of the three general categories of waste forms modeled (CSNF, DSNF, and HLW) are not coupled, and so three different paths through the Waste Form Degradation Model are possible (Figure 1.5-3). Furthermore, because the in-package chemistry (primarily pH) is dependent upon the amount of CSNF exposed and the rate of alteration of the borosilicate glass of the HLW, some minor feedback does occur. Because the feedback is minor (i.e., secondary effect), the feedback lags by one timestep (i.e., the Waste Form Degradation Model does not iterate during the timestep).1.6 Principal and other factors considered
The magnitude of the YMP and the complexities associated with both the natural and EBSs dictate that the YMP prioritize its activities and focus on the factors most important to performance, hereafter named the principal factors. The Repository Safety Strategy (RSS) (CRWMS M&O 2000a, Table 3-1) has identified seven principal factors and 18 other factors of lesser importance. The selection of the principal factors has been based on preliminary TSPA analyses and expert judgment, which identified these factors as likely to provide sufficient confidence for the safety case. The other factors were deemed to be less important to the safety case, nevertheless requiring representation in the current methodology (CRWMS M&O 2000a, Section 3.1). Table 1.6-1 lists the seven principal factors, the 18 other factors, and the PMRs that address each. The Waste Form Degradation PMR addresses the principal factor Solubility Limits of Dissolved Radionuclides. Solubility Limits of Dissolved Radionuclides is a principal factor because it describes the limitation to the mobilization of the relatively immobile radionuclides because of their limited solubilities in the water at Yucca Mountain. Some of these radionuclides present significant risk potential because of the long half-life and large dose conversion factor. In many cases, the solubilities of these radionuclides are so low they present no significant issue for the potential repository system. However, there are a few cases, notably neptunium, plutonium, and uranium isotopes, for which solubility limits could be very important to the postclosure safety case (CRWMS M&O 2000a, Section 3.2.4). In addition, the following six other factors are within the scope of the Waste Form Degradation PMR.1.7 Organization of Report
Chapter 2 discusses the evolution of the Waste Form Degradation Model from past TSPA analysis and the formal screening of FEPs to include in the current model. The major portion of this report, Chapter 3, discusses the inner workings and technical basis of the eight main components of the Waste Form Degradation Model used in the TSPA-SR analysis. The order of discussion for the eight components reflects the direction of data flow between components (Figure 1.5-3). Each section of Chapter 3 begins with an introduction to the component. This introduction includes a listing of the AMRs and calculations that provide the detailed technical basis for that component, and an overview of the conceptual model as appropriate. The description sub-section, which follows, includes summaries of (1) the descriptions of the analysis, models, and/or abstractions, (2) relevant data and uncertainties, (3) assumptions and basis, and (4) model results. The amount of detail provided varies with each component. For many components, the results are simply the models and parameters provided to the TSPA-SR and are, thus, included in the model/analysis abstraction description. In addition, many times, the relevant data and uncertainties are only briefly summarized, and the reader is referred to the supporting AMRs for detailed discussion. The Confidence/Limitations/Validation sub-section includes a discussion of model limitations, the justification and confidence, and validation that the component is appropriate for the intended use (i.e., appropriate as support for making a recommendation on the suitability of the Yucca Mountain site for a repository). Each component section ends with a sub-section, which summarizes alternative models and/or addresses known issues posed by groups that have project oversight, regulatory oversight, or stakeholder interfaces with the YMP. The last section of Chapter 3, Section 3.9, tabulates these issues for all eight components, and provides cross-references to the component sections. Chapter 4 of the report summarizes the relationship of the topics of this report with the NRC IRSRs, primarily the CLST KTIs. Chapter 5 is a brief summary of the report.2. EVOLUTION OF WASTE FORM DEGRADATION MODEL
The waste form degradation models have generally evolved over the years along with the overall TSPA model of the disposal system. Thus, this section first presents a short summary of the evolution of the TSPAs of the Yucca Mountain disposal system by the Yucca Mountain Project (YMP), followed by a summary of the evolution of various components in the Waste Form Degradation Model. More detail is presented along with each major section of the various components.2.1 GENERAL HISTORY OF ASSESSMENTS OF YUCCA MOUNTAIN PERFORMANCE
As summarized by Rechard (1999), simple analytic calculations to determine the relative importance of various phenomena postulated to occur at Yucca Mountain were conducted in 1984 (which identified 99Tc, 129I, and 237Np as important radionuclides for evaluating compliance) (Sinnock et al. 1984) and 1988 (performed in conjunction with the Site Characterization Plan (SCP) (DOE 1988, Section 8.3.5.13). The first large-scale analysis of water movement through the unsaturated zone occurred in 1990 (Prindle and Hopkins 1990). Shortly thereafter, a series of deterministic calculations using best estimates for model parameters were run by several organizations—Sandia National Laboratories, Pacific Northwest National Laboratory, and Los Alamos National Laboratory—to simulate the radioisotope transport in the unsaturated zone (Barnard and Dockery 1991). In 1992, the YMP completed the first probabilistic TSPA of the potential Yucca Mountain disposal system to evaluate releases to a 5-km boundary. Two different organizations conducted total system performance analysis: Sandia National Laboratories (TSPA-91-SNL) (Barnard et al. 1992) and Pacific Northwest National Laboratory (TSPA-91-PNNL) (Eslinger et al. 1993). For the first time, gaseous flow of 14C was included. In this first probabilistic assessment, the YMP was at a relatively early stage in conceptual model development; thus, parameter values and distributions were determined primarily by individual PA analysts. In 1993, a second iteration of the SNL TSPA (TSPA-93-SNL) (Wilson et al. 1994) was started that included an improved Waste Form Degradation Model (called source-term model) and a saturated zone model. The analysis also greatly expanded the data used for defining geochemical parameters. The formality increased as well in that distributions for many more parameters were developed and more often based on the consensus of several PA analysts, accompanied by input from site characterization scientists. In 1993, the newly assigned CRWMS M&O also conducted a TSPA (TSPA-93-Duke) (CRWMS M&O 1994) using the Repository Integration Program (RIP) modeling system intended to rapidly simulate the behavior of the disposal system to evaluate design systems. The system used a variety of techniques such as curve fits to previous results and selection of distributions for particular data to incorporate previous results. This simplified modeling style, called abstraction, had been originally proposed in the 1988 SCP (DOE 1988, Chapter 8, Section 8.3.5.13) as the culmination of sensitivity analysis on process models. The analyses using RIP were the only TSPAs performed by CRWMS M&O after 1993. The basic information on parameter distributions reported in TSPA-93-SNL (Wilson et al. 1994) was used for TSPA-93-Duke (CRWMS M&O 1994) and subsequent TSPAs in 1995, 1996, and 1997 (CRWMS M&O 1995; CRWMS M&O 1996a; CRWMS M&O 1996b; CRWMS M&O 1997a); although some differences did occur for radionuclide inventory, and other parameter values were sometimes changed for parametric sensitivity analysis. Besides TSPAs conducted specifically by CRWMS M&O for the Yucca Mountain Site Characterization Office of DOE (DOE/RW/YMSCO), TSPAs conducted specifically for the National Spent Nuclear Fuel Program (NSNFP) of the Office of Environmental Management of DOE (DOE/EM) by SNL in 1993, 1995, and 1998 examined the behavior of DSNF to test the viability of direct disposal of the waste in salt, granite, and tuff (DSNF-TSPA-93-SNL) (Rechard 1993), DSNF-TSPA-94-SNL (Rechard 1995), DSNF-TSPA-98-SNL) (DOE 1998a; Rechard 1993, 1995). The analysis included the effects of cladding since this was an important feature that distinguished ~250 types of DSNF and CSNF. The NSNFP also contracted with CRWMS M&O to conduct similar analysis in 1997 for DSNF, DSNF-TSPA-97-Duke (CRWMS M&O 1997a). The CRWMS M&O had done a similar sensitivity study the year before on the disposal of excess weapons and plutonium (CRWMS M&O 1996b). In 1998, NSNFP supported CRWMS M&O such that DSNF was included as a sensitivity study in the TSPA-VA (DOE 1998b) (viability assessment) technical basis document discussed below (CRWMS M&O 1998b). In 1997, Congress mandated in its energy appropriation bill that the YMP provide a viability assessment that (along with a preliminary design and costs estimates for constructing and operating that design) would include a TSPA describing the probable behavior of the potential Yucca Mountain disposal system (Energy and Water Development Appropriations Act 1997). A TSPA-VA was thus initiated and completed in November 1998 (DOE 1998b). For TSPA-VA, numerous changes and additions were made to the TSPA-95 (CRWMS M&O 1995) models, including the addition of more phenomena. Some of these changes included the influence of the Zircaloy cladding on CSNF, evaluation and inclusion of geochemistry changes near the waste package, colloid formation and transport, and a factor of 100 reduction in solubility of Np (DOE 1998b; CRWMS M&O 1997b; CRWMS M&O 1998b).2.2 EVOLUTION OF EACH MODEL COMPONENT
A brief history of the evolution of the waste form models and components is discussed.2.3 SCREENING OF FEATURES, EVENTS, AND PROCESSES
2.3.1 Screening Criteria
Throughout the 20 years of analysis of the potential Yucca Mountain disposal system, various FEPs have been identified and their influence on the disposal system evaluated. The FEPs have been identified by a variety of methods, such as hypothesis by scientists and engineers working on the project or through review of their work. The variety of project participants and reviewers have helped ensure that a wide variety of FEPs has been considered. These hypotheses and reviews have been the impetus for the changes and evolution of the models discussed above. An initial set of FEPs was created for TSPA-SR analysis by combining three general lists of FEPs: 82 from various YMP workshops in 1998 and 1999, 292 FEPs gathered from YMP literature and site studies, and 1,412 FEPs from a draft report of the Nuclear Energy Agency (NEA) of the Organization for Economic Cooperation and Development (OECD) for a total of 1,786 FEPs. The latter list is a compilation of FEPs from seven geologic repository programs in other countries and, thus, the most complete attempt internationally to develop a comprehensive list of FEPs relevant to radioactive waste disposal. To ensure the list was comprehensive, all potentially relevant FEPs identified have been included; however, this list has led to considerable redundancy. Consequently, late in 1998, the FEPs were classified as either primary or secondary. The 310 primary FEPs are those FEPs for which detailed screening arguments are developed. The remaining 1,476 secondary FEPs are either completely redundant or can be reasonably aggregated and mapped into a single primary FEP through modification of the description the primary FEP. To develop screening arguments, the primary FEPs have been assigned to applicable PMRs. However, since a FEP can effect many facets of the disposal system, a FEP may be assigned to several PMRs. For example, some FEPs that affect the degradation of waste form also influence the waste package degradation and the evolution of the Engineered Barrier System and, thus, have been assigned to all three corresponding PMRs. The 87 primary waste form FEPs assigned to the Waste Form Degradation PMR have been evaluated by subject-matter experts and discussed in three different AMRs: 57 primary waste form FEPs are discussed in Miscellaneous Waste-Form FEPs (CRWMS M&O 2000n, Table 1), 15 are discussed in Waste Form Colloid-Associated Concentrations Limits: Abstraction and Summary (CRWMS M&O 2000m, Attachments I through XV), and 16 are discussed in Clad Degradation—FEPs Screening Arguments (CRWMS M&O 2000s). One FEP (Mutation) appears in two AMRs (CRWMS M&O 2000n, CRWMS M&O 2000m). The decision to include or exclude a FEP related to the waste form was based on two criteria in the NRC’s proposed rule 10 CFR 63.114 (Dyer 1999)4.2.3.2 Screening Decisions
Most of the waste form FEPs that have been excluded from further consideration were excluded based on low consequence. Several waste form FEPs have been excluded based on the FEP not being credible for the waste characteristics and repository design at Yucca Mountain. For the discussion herein, the 87 FEPs have been grouped and tabulated according to the eight major modeling components of the waste degradation model (Figure 2.3-1). Each of the eight tables provides the YMP FEP number, the short title, the decision as to whether to include or exclude the FEP, the PMRs where the FEP was discussed, the AMR that provides the full argument for the screening decision related to the waste form, and a brief synopsis of the screening argument. The synopsis only pertains to the Waste Form Degradation PMR. The reasoning to include or exclude a FEP in relation to other PMRs is not given but can be found in those other PMRs or the underlying AMR. Several waste form FEPs have been excluded because the FEP is not credible for the characteristics of the waste proposed for disposal at Yucca Mountain (i.e., first criterion, T1). The succinct description for this reasoning is "Excluded based on low probability (credibility)." For example, cellulosic material is not present in CSNF, DSNF, or HLW that would be disposed of at Yucca Mountain; furthermore, organic material will be excluded based on current waste acceptance criteria. Therefore, FEP 2.1.02.10.00 in Table 2.3-2 that discusses this feature of the waste is excluded based on a low probability (credibility) argument. No FEPs were excluded based on quantitatively evaluating the probability of a FEP (i.e., criteria T2 and T3). Rather, most of the excluded FEPs related to the waste form were excluded based on the fourth criterion (T4), "no significant change in the expected annual dose." FEPs excluded based on this criterion are succinctly described as "Excluded based on low consequence." For example, generation of H2 is credible in an anoxic environment when steel corrodes; yet, it is unlikely that much will be produced since the drifts at Yucca Mountain will be mostly oxic. Only in small pores and crevices would an anoxic environment be found in the waste. Because evaluating a probability for this phenomenon is very difficult, the low consequences of this phenomenon were used to exclude this FEP. H2, if produced, is very mobile and kinetically unreactive at low temperatures, so it is expected to leave the drift before it reacts with any of the emplaced materials. In addition, oxidation of hydronium ion to H2 instead of sulfur to sulfate results in less pH suppression. The latter reason is a minor aspect of the argument to exclude the process, but in other FEPs it can be a major aspect of the argument because bounding estimates are often used in the TSPA-SR. That is, the argument for exclusion is often that the exclusion of the FEP provides a bound on the expected annual dose (i.e., inclusion of the FEP would only decrease the expected annual dose). For this case, the succinct description of the FEP reasoning used in the summary tables is "Excluded based on low consequence (conservative bound)." For example, a few radioisotopes, such as Np, may be incorporated into the structure of phases of other minerals (primarily uranium minerals) that form during degradation of the waste. For chemically similar elements such as Np and U, the effective solubility of Np as a minor constituent of a uranium mineral is less than the solubility of the analogues pure Np mineral. However, this FEP, "secondary phase effects on dissolved radionuclide concentrations at the waste form" in Table 2.3-7, was conservatively excluded in TSPA-SR. The degree of conservatism is not quantified since the collection of data to justify the model would require either too much time or too many resources for the surmised potential benefit. The exclusion of FEPs that have a potential beneficial consequence is a conservatism of the FEPs screening process. This conservatism is in addition to the many conservative approaches used to include FEPs as more fully discussed in Chapter 3.2.3.2.1 FEPs Related to the Radioisotope Inventory Component
Five FEPs relate to the Radioisotope Inventory Component (Table 2.3-1). The waste inventory was reconsidered and 27 important radionuclides identified. The radioisotopic gas inventory was omitted because the screening analysis showed that gases were not significant contributors to dose. However, one radioisotopic gas, 14C, is included as a dissolved species. Not all isotopes of uranium and plutonium are included in the 27 important radionuclides, and so the proper proportion ("dilution") of the included uranium and plutonium isotopes is evaluated for the Waste Form Degradation Model when evaluating solubility. For the TSPA-SR analysis, radioactive decay (and ingrowth6) is modeled, and the heterogeneity of the inventory in the waste form is included using three waste forms (CSNF, DSNF, and HLW).2.3.2.2 FEPs Related to the In-Package Chemistry Component
Twenty-five FEPs relate to the In-Package Chemistry Component (Table 2.3-2). Prior to TSPA-SR analysis, uncertainty in the in-package chemistry was evaluated through scientific judgment; but objectively demonstrating that the uncertainty from these 25 FEPs was included was difficult; an in-package chemistry component was developed for TSPA-SR analysis to explicitly include many of these FEPs. Furthermore, a few FEPs dealt with the coupling of the in-package chemistry with other model components; these FEPs can now be included in the current TSPA-SR analysis.2.3.2.3 FEPs Related to the CSNF Matrix Degradation Component
Seven FEPs relate to the CSNF Matrix Degradation Component (Table 2.3-3). The process of CSNF dissolution and radioisotope release has been included in TSPA analysis for a number of years through an regression equation of experimental results on both irradiated and unirradiated fuel. The use of samples of irradiated fuel implies that the potential enhanced dissolution from radiolysis, radiation damage, and alpha recoil has indirectly been included in the dissolution rates. The small differences in results and theoretical arguments, however, show that these latter effects are actually insignificant and can be excluded from consideration. While an unimportant distinction for the CSNF since experimental results are available that include these effects, enhanced dissolution from radiolysis, radiation damage, and alpha recoil are excluded from DSNF and HLW degradation as discussed below. The process of magma interacting with the waste has formally been included this year in the analysis for the TSPA-SR and is described in the Disruptive Events PMR.2.3.2.4 FEPs Related to the CSNF Cladding Degradation Component
Twenty FEPs relate to the CSNF Cladding Degradation Component (Table 2.3-4). These 20 FEPs were developed based on comments received on early TSPAs (e.g., DSNF-TSPA-93-SNL [Rechard 1993] and DSNF-TSPA-94-SNL [Rechard 1995] and the TSPA-VA [DOE 1998b]). Several perforation mechanisms have been included in the CSNF Cladding Degradation Component: initial perforation in reactor before receipt of the waste, creep perforation, stress corrosion cracking, mechanical failure due to severe earthquakes, and localized corrosion. The influence of temperature on creep and stress corrosion cracking are also included. Localized crevice corrosion, and diffusion-controlled cavity growth, have been excluded based on FEP arguments. Based on arguments in an AMR dealing solely with this topic, hydride perforation mechanisms such as hydride embrittlement and DHC have also been excluded. Complete failure of the cladding after perforation is modeled through unzipping of the cladding in an aqueous environment7. Unzipping in a very hot, dry environment where the UO2 is rapidly oxidized to U3O8 is excluded based on low consequence since the required high temperature conditions do not exist when WP failure typically occurs.2.3.2.5 FEPs Related to the DSNF Degradation Component
Three of the FEPs related to CSNF Degradation Component (Section 2.3.2.3) also apply here (radiolysis, alpha recoil, and radiation damage). These three radiation effects are included by conservatively bounding the degradation of rate of experiments on irradiated N-Reactor fuel. Six additional FEPs specifically related to the DSNF Degradation Component are listed here (Table 2.3-5). Four of these FEPs relate to the potentially increased reactivity of some types of the DSNF. However, the additional reactivity is of low consequence and so these four FEPs are screened out. One FEP relates to the degradation of cladding for DSNF, but is conservatively excluded by not including cladding performance for DSNF except for the naval SNF, whose analysis will account for the performance of the cladding.2.3.2.6 FEPs Related to the HLW Degradation Component
Four FEPs specifically relate to the HLW Degradation Component (Table 2.3-6). One primary FEP and its underlying secondary FEPs deal with the enhanced degradation that may occur if extensive phase separation or glass recrystallization occur in the encapsulating glass; however, production controls will prevent significant initial phase separation. Furthermore, temperatures in the repository are low relative to the melting point of the glass such that extensive recrystallization will not occur in the glass. The HLW Degradation Component explicitly includes the effect of temperature on HLW glass degradation. Another FEP deals with the proper reactive surface area to use when evaluating the degradation rate of the glass. The HLW Degradation Component bound the reactive surface area by multiplying the geometric surface area by a very conservative, but constant cracking factor of 20.2.3.2.7 FEPs Related to the Dissolved Radioisotope Concentration Component
Many processes influence the dissolved concentration of radionuclides in the Dissolved Radioisotope Concentration Component. However, most of those processes have been discussed separately in the In-Package Chemistry Component since these processes also affect other model components. Here four FEPs are listed that have a more direct connection with the dissolved radionuclide concentration (Table 2.3-7). First, solubility control based on pure phases has been and continues to be included in the TSPA-SR analysis. Although a few radionuclides such as Np may be incorporated into the structure of phases of other minerals (primarily uranium) that form during degradation of the waste such that the solubility of these other minerals control the concentration of the radioisotope, these effects have been conservatively excluded. Complexation of radionuclides by organic ligands has been excluded because organic material is generally prohibited from the repository and the incidental amounts of organic material that may inadvertently occur cannot alter the solubility of significant amounts of radionuclides.2.3.2.8 FEPs Related to the Colloidal Radioisotope Concentration Component
Sixteen FEPs relate to the Colloidal Radioisotope Concentration Component (Table 2.3-8). Three types of colloids were included: those already present in the groundwater, those formed during corrosion of the WP, and those formed during degradation of the waste. Condensed Pu polymer colloids were excluded since they have not been observed in fuel or glass dissolution tests. Biological activity was screened out, and so microbial colloids were also excluded. Transport and filtration of colloids inside the WP were conservatively excluded.2.3.3 Relationship of Model Components, FEPs, and Factors
The model components listed in Section 2.2 are closely aligned with the principal and other factors that affect repository performance. Table 2.3-9 shows the correspondence. Note that FEPs 2.1.02.09.00 and 2.1.09.02.00 are related to two different factors.2.4 Influence of design changes on Waste form degradation model
The design of the repository continues to evolve in preparation for the future license application. In general, these design changes can directly influence the screening arguments for FEPs and their inclusion in or exclusion from the waste form degradation model. The elimination of backfill is an important design change that has been made after the development of the current Waste Form Degradation Model. The primary effect of the elimination of backfill is the decrease of peak temperatures inside the WP, which is beneficial. For example, the temperature decrease (1) reduces the chance for creep rupture and stress corrosion cracking of CSNF cladding; (2) reduces the degradation rates of the CSNF and HLW matrices; (3) improves the applicability of the current data for in-package chemistry; and (4) decreases the solubility of uranium. However, design change to eliminate backfill does not affect the Waste Form Degradation Model because temperature is explicitly included as a model variable. That is, the model is an explicit function of the surface temperature of the waste package (Figure 1.5-1), so any changes in the surface temperature of the waste package because of the design change are automatically included.3. MODELS AND ABSTRACTIONS
3.1 RADIONUCLIDE INVENTORY COMPONENT
The function of the radioisotope inventory abstraction component is to estimate the inventory of those radionuclides most important to human dose. The inventory abstraction component is input for the waste form degradation models and is developed from a series of steps that starts with radioisotope inventories of various spent nuclear fuel assemblies and HLW then estimates the radioisotope inventory when packaged in disposal containers. Three important aspects of the radionuclide inventory are (1) selecting the most important radionuclides for human dose out of the few hundred found within the waste, (2) obtaining the radioisotope inventories of various wastes, and (3) grouping the fuels into the waste packages selected for modeling in the TSPA-SR analysis. Using radionuclide activities for CSNF assemblies, DSNF canisters, and HLW canisters, the radionuclide inventory component provides an estimate for activities in containers destined for disposal in the potential Yucca Mountain repository as shown in Figure 3.1-1. The inventory abstraction is described in the Inventory Abstraction AMR (CRWMS M&O 2000f) and the eleven supporting calculations (Figure 3.1-2). Three sources were used for inventory data: the commercial utilities for CSNF (CRWMS M&O 1999d), the DOE NSNFP for DSNF (DOE 1999b), and the Yucca Mountain EIS program for the HLW, mixed oxide fuel (MOX), and plutonium ceramic (DOE 1999a). From this radioisotope inventory, the most important radionuclides for human dose were evaluated. Arrival scenarios were developed for CSNF, and the inventory was assigned to WP configurations. Average and bounding inventories were developed for each package configuration recommended for the repository. Then, the package-specific radionuclide activities were combined using the number of WPs in each group as a weighting factor to get the radionuclide activities in each allocation category.3.1.1 Description
3.1.1.1 Radionuclides Important for Total System Performance Assessment
The relative importance of individual radionuclides for human inhalation and ingestion doses was evaluated for several waste types, time frames, and release scenarios. In this evaluation, the effects of inventory abundance, radionuclide longevity, element solubility, and element transport affinity were considered. To address inventory abundance, eight waste types were examined9. To address radionuclide longevity, the fuels were evaluated between 100 and 1 million years after repository closure. The elements were evaluated in two solubility groups, the relatively soluble and the relatively insoluble (Am, Cm, Zr, Th, Nb, Pa, and Sn) and three transport affinity groups: (1) highly sorbing, (2) moderately sorbing, and (3) slightly to non-sorbing. The isotopes within each group were compared to one another for relative importance. Three release scenarios were considered: nominal case, human intrusion, and direct volcanic release. Two time frames were considered: 100 to 10,000 years and 100 to 1 million years. The set of important isotopes was different for each scenario and time frame. For the first 10,000 years, 99Tc and 129I are normally the primary contributors to dose (DOE 1998c, Figure 4-29). For a direct release from a disruptive event scenario out to 10,000 years, 90Sr, 137Cs, 227Ac, 229Th, 231Pa, 232U, 233U, 234U, 238Pu, 239Pu, 240Pu, 241Am, 243Am were screened in. These are the isotopes that contribute most to the dose when release is not mitigated by either solubility or transport. For a nominal release and human intrusion scenario out to 10,000 years 14C, 99Tc, 129I, 227Ac, 229Th, 232U, 233U, 234U, 236U, 238U, 237Np, 238Pu, 239Pu, 240Pu, 241Am, 243Am were screened in. By modeling the plutonium isotopes (238Pu, 239Pu, 240Pu), the americium isotopes (241Am, 243Am), 229Th, and 227Ac, doses that could result from colloidal transport of radioactive material to the biosphere will be adequately represented. By modeling 14C, 99Tc, 129I, the uranium isotopes (232U, 233U, 234U, 236U, 238U), and 237Np, doses that could result from transport of solutes, either by fracture flow or matrix diffusion, will be adequately represented. In addition, 63Ni, 90Sr and 137Cs are screened in for a human intrusion scenario because the event could occur as early as 100 years after repository closure. For the time period from 10,000 to 1 million years, 210Pb, 226Ra, 230Th, and 242Pu are screened in for all of the scenarios, and 231Pa is screened in for the human intrusion and nominal release scenarios. The result is 24 isotopes recommended for modeling in the TSPA-SR analysis and TSPA-FEIS. The 24 isotopes are shown in Table 3.1-1, along with the isotope selection from previous performance assessments. For TSPA 1993 (Wilson et al. 1994) and 1995 (CRWMS M&O 1995), all radionuclides contributing to 99.99 percent of the potential dose at any time from 1,000 to 1 million years were retained for the inventories. For TSPA-SR analysis, all radionuclides contributing to 95 percent of the potential dose at any time from 100 to 1 million years were retained for the inventories. Thus, radionuclides that contribute to the last five percent of the dose estimate were eliminated from consideration in the TSPA-SR analysis. For the TSPA-VA (CRWMS M&O 1998b), a smaller number of representative radionuclides were chosen. A radionuclide was included in the TSPA-VA (DOE 1998b) if it had: (1) a high solubility, (2) a low sorption affinity, (3) a significant inventory10, (4) a high dose conversion factor, and (5) a long half-life. In addition, some radionuclides were included based on previous experience (1993 TSPA calculations [Wilson et al. 1994], 1995 TSPA calculations [CRWMS M&O 1995], and scoping calculations) and the experience of other organizations (NRC performance assessments [Wescott et al. 1995]). For the Iterative Performance Assessment (IPA) (Wescott et al. 1995), the NRC screened radionuclides to include only the major contributors to cumulative release and dose. A radionuclide was retained in the inventory if, in preliminary calculations, it contributed more than one percent of the EPA cumulative release limit for that radionuclide11. The screening analysis also checked the maximum dose to a farm family to see if any of the radionuclides that might have been screened out on the basis of cumulative release should have been kept on the basis of dose. The differences between the isotope selection in these PAs are primarily due to: (1) the changing regulations (dose - cumulative release - groundwater protection; 10,000 years - 1 million years, groundwater - human intrusion - volcanic releases), (2) inventory data relied upon, and (3) the screening techniques.3.1.1.2 Commercial Spent Nuclear Fuel Inventory
Commercial nuclear power plants use a variety of fuels and fuel configurations in their reactor cores to generate power. The predominant fuel is enriched uranium dioxide, but a plutonium/uranium mix is also planned for use in commercial reactors to dispose of excess plutonium from the government inventory. Fuel pellets are packed into fuel rods (which vary in size depending on the application), and fuel rods (clad in Zircaloy or stainless steel) are bundled into assemblies. The number of fuel rods per assembly and the number of assemblies in a reactor core vary depending on the reactor design. Once a reactor fuel has been irradiated to the extent that it can no longer effectively sustain a chain reaction, it is removed from the reactor and becomes spent nuclear fuel. Its isotopic composition at that point depends on the initial enrichment of the fuel, the reactor configuration (pellet size, fuel rod size, assembly layout, and other parameters), and the extent of irradiation (called burnup). Once removed from the core, the fuel is placed in storage and it ages. Almost 230,000 CSNF assemblies will need to be disposed, and each assembly, depending on the reactor configuration, initial fuel enrichment, burnup, and the age of the waste (time in storage), will have a unique isotopic composition. The radionuclide inventory abstraction collects this information and synthesizes it into a form that can be modeled in TSPA-SR analysis. A 1995 data submittal from the commercial utilities provided the basic information from which the TSPA-SR analysis inventory for CSNF was developed. In 1995, the utilities supplied historical information about reactor assembly discharges up through December 1995, and they provided five-cycle forecasts for assembly discharges from their reactors. With this information, a design basis waste stream was developed (CRWMS M&O 1999d), and forecasts for assembly discharges over the lifetime of each commercial power reactor were developed. For the base case repository design of 70,000 MTHM, three alternative schedules were developed for moving the CSNF assemblies out of storage and shipping the assemblies to Yucca Mountain. The schedules, called arrival scenarios, include the year of receipt for each shipment, the number of assemblies in each shipment, the type of fuel in each shipment (pressurized water reactor (PWR) versus boiling water reactor (BWR), and the enrichment/burnup characteristics of the fuel. Radionuclide activities for each assembly in the waste stream were estimated, and the WP configuration that could accommodate the assembly based on its potential criticality level was determined. The result was a grouping of the 230,000 CSNF assemblies into five proposed WP configurations. For each group, average and bounding assembly radionuclide activities were calculated.3.1.1.3 U.S. Department of Energy Spent Nuclear Fuel and High-Level Waste
For the most part, DSNF (with the exception of the naval SNF) and HLW are planned to be disposed together in codisposal waste packages. Therefore, they are discussed together here. The DSNF consists of more than 250 distinct types of spent fuel, and much like CSNF, radionuclide inventories for defense fuels will vary widely depending on the history of the fuel. The NSNFP grouped the fuels into 11 groups (DOE 1999b). DOE fuels will be packaged in three types of canisters before they are shipped to Yucca Mountain; short, long, and multi-container overpack (MCO). Similarly, the naval fuels will be packaged in two types of canisters: naval long and naval short (Dirkmaat 1997, Appendix F, Attachment, p. 1). For analysis, the fuels are grouped by canister type and average and bounding per canister radionuclide inventories were calculated. The HLW in storage at DOE sites is the result of reprocessing SNFs (some CSNF and some DSNF). The proposed technology for immobilization of HLW is vitrification in a borosilicate glass. The vitrified waste will be placed in one of two canister types (long and short). A small amount of HLW glass has been produced at the West Valley Demonstration Project (WVDP) in New York. Production of HLW has started at the Savannah River Site (SRS) and HLW will eventually be produced and stored at two other sites—the Hanford Reservation (HR) and the Idaho National Engineering and Environmental Laboratory (INEEL). Because the fuels reprocessed at each of these sites differ, the radionuclide inventory of the HLW and resultant glass product will vary slightly among the sites. The data source for the HLW source terms, which include radionuclide inventory, decay heat, and radiation sources due to gamma rays and neutrons, is the response to the EIS data call as described in DOE (1999a). The sites included in the data call are the HR, SRS, INEEL, and WVDP. HLW fuels will be packaged in three types of canisters before they are shipped to Yucca Mountain: short, long, and Pu disposition. The information from the data call was used to calculate average and bounding per canister inventories. The average radionuclide activity from DSNF for one of these WPs is the number of canisters times the average per canister radionuclide activity calculated. The average radionuclide activity from HLW for one of these WPs is the number of canisters times the average per canister radionuclide activity calculated. The radionuclide activity for the DSNF allocation group is an average, weighted by the number of packages in each group, over the six WP configurations. Similarly, the radionuclide activity for the HLW allocation category is an average, weighted by the number of packages in each group. The naval SNF, by agreement, is conservatively treated separately using average CSNF inventory and degradation characteristics.3.1.1.4 Radionuclide Masses in Allocation Categories for Total System Performance Assessment for Site Recommendation
The waste types, allocations, and waste packages for commercial spent nuclear fuel, DOE spent nuclear fuel, high-level radioactive waste, and plutonium disposition waste are shown in Figure 3.1-1. In the TSPA model, over 250 types of DOE spent nuclear fuels and HLWs are represented as being packaged in the ten types of canisters listed in Table 3.1-2. These canisters and the CSNF assemblies are, in turn, represented as being emplaced in ten types of waste packages, which are also listed in Table 3.1-3. The waste packages and canisters combine to give a total of thirteen waste package configurations. The waste package configurations and the number of packages with each configuration for TSPA modeling purposes are listed in Table 3.1-3. These waste package numbers were specified by management edict (Stroupe 2000), and are somewhat different from the "Proposed Action" numbers used in the DEIS (DOE 1999a) or the "truncated case" in recent input transmittals (CRWMS M&O 2000aj and 2000ae). The management specified numbers are "not to exceed" numbers and are rounded up from the truncated case, in order to convey flexibility in the design. The truncated case has more HLW canisters than the DEIS, in order to use the "ideal waste emplacement scenario" for codisposal. Neither the management case nor the truncated case may be emplaced without a redefinition of "HLW MTHM equivalency" or the elimination of the 70,000 MTHM limit specified in the Nuclear Waste Policy Act of 1982. Using the historical definition of 0.5 MTHM per HLW canister, these cases contain more than 4,667 MTHM HLW equivalent. Table 3.1-3b compares some of the canister and assembly numbers for the various scenarios. There will be 7860 CSNF packages and 3590 codisposal packages in the TSPA-SR analysis model. The 300 naval packages were grouped with the codisposal packages in the first version of CRWMS M&O 2000f, but were regrouped with CSNF packages in the subsequent input transmittal (CRWMS M&O 2000ak). Naval spent fuel is expected to perform very well within the repository, and it is better represented by commercial packages than by codisposal packages in the TSPA. Accordingly, the naval fuel inventory was removed from the high level waste and DOE spent nuclear fuel averages, and the averages were recalculated for the input transmittal (CRWMS M&O 2000ak). The impact analysis of this change is expected to show that the old values were conservative, and the total TSPA values are not significantly affected. The radionuclide inventories in grams for the selected isotopes are shown in Table 3.1-4. Three radionuclides appear in Table 3.1-4 that were not listed in Table 3.1-1. 228Ra, 232Th, and 235U were not identified as important contributors to dose for the direct release scenario, the human intrusion scenario, or the nominal release scenario. However, 228Ra and 232Th will be modeled in TSPA for the groundwater protection scenario and therefore were included in the inventory abstraction. 235U will be modeled in the TSPA because it is a parent to 231Pa. To get accurate estimates of the dose from 231Pa, TSPA must track the transport of 235U.3.1.2 Uncertainties, Limitations, and Conservatisms
The screening analysis is conservative because all fuel types, bounding and average, all time periods, and all transport groups were examined. Any changes that might be expected in the wastes that may be disposed at Yucca Mountain will not change the radionuclides that were screened in for modeling in TSPA. The representative inventories were derived from unqualified projections of future waste streams. The actual waste streams will be known only at the time of actual repository loading. The projected waste streams could differ from the actual waste streams in their fuel burnups, fuel ages, fuel enrichments, and utility efficiencies. However, changes that might be expected in the waste stream will produce only minimal (less than 20 percent) changes in the radionuclide activities in the fuels. Given this understanding, the values chosen for initial inventories in CSNF and codisposal waste packages are reasonable representations of the inventory that may be disposed at Yucca Mountain (CRWMS M&O 2000f).3.1.3 Other Views
Alternative approaches for developing the radionuclide inventory have been explored in previous TSPAs (Wilson et al. 1994; CRWMS M&O 1995; CRWMS M&O 1998b) and in the DEIS (DOE 1999a). In previous TSPAs, (Wilson et al. 1994; CRWMS M&O 1995; CRWMS M&O 1998b), radionuclide activities for CSNF were developed by assuming an average set of fuel characteristics (enrichment, burnup, and age of the waste) and using the DOE’s Characteristics Database to determine radionuclide activities for a fuel having the specified characteristics. Radionuclide activities for DSNF were developed by assuming that the defense fuels can be grouped into a few representative groups (based on their chemical characteristics) and calculating inventories for a representative fuel from each group. Radionuclide inventories for HLW were developed by assuming that an average of glass waste from the Savannah River Vitrification Plant and processed waste from Hanford, INEEL, and the West Valley Facility can represent HLW. The TSPA-SR analysis screening method used a qualified process to ensure traceability. It looked at both bounding and average fuels, DSNF and HLW as well as CSNF, time periods from 100 to 1 million years, and both inhalation and ingestion doses. By using two solubility groups and three different transport groups and comparing the radionuclides within the groups, better discrimination was achieved. For example, if all isotopes were lumped together, plutonium would dominate the dose. But this is true only if Pu is readily transported to the accessible environment. The full TSPA-SR analysis and its colloid models will predict the likelihood of significant Pu transport. The screening could not rely on TSPA output and, thus, did not rely on any assumptions about solubility or transport affinity. The TSPA-SR inventory analysis is more detailed and flexible than previous analyses and is tied to the waste stream. Changes in WP configuration or waste stream are more easily reflected in the per package inventory of representative WPs. An example of such a change is seen in the recent regrouping of naval spent nuclear fuel.3.2 In-package Chemistry Component
The function of the in-package chemistry model component is to estimate the fluid chemistry inside the WP over time after the initial breach of the disposal container. This chemistry is then used by the several other model components (see Figure 1.5-3) since the rate of degradation of the matrix of waste, the resulting dissolved concentration of radionuclides, the stability of any colloids, and degradation of cladding are all dependent on the chemistry of fluids within the WP. The rate of degradation of the waste matrix and inner stainless steel container, in turn, influences the fluid chemistry and so there is a coupling between all the chemically interacting components of the system. The In-Package Chemistry Abstraction AMR (CRWMS M&O 2000g) is the primary document describing the regression equations used for evaluating the in-package chemistry. This document, in effect, abstracts the process models that are developed in Summary of In-Package Chemistry for Waste Forms (CRWMS M&O 2000o). The In-Package Chemistry Abstraction (CRWMS M&O 2000g) also relies on a few of the FEP screening arguments in Miscellaneous Waste-Form FEPs (CRWMS M&O 2000n). While abstraction of process models is numerically convenient, it is technically defensible only if the abstracted model conservatively bounds the predictions of process-level models. In this case, dissolved radionuclide levels predicted with the abstracted model must be equal to, or greater than, those predicted by the underlying process-level models. The most reliable way to assure technical reliability of the abstraction is to force it to be mechanistically identical to the process-level model. This was done for the solubilities of the majority of the primary radionuclides by building abstractions of in-package chemistry and solubility expressions on equilibrium speciation output of EQ3NR.3.2.1 Description
The fluid chemistry inside the package (in-package chemistry) is dependent upon the initial chemistry of the water entering the breached package, the volume of water flowing through the package, the amount of water remaining within the package, and the amount of time that inflows into the waste package occur. The inflows were assumed to have the composition of J-13 groundwater. Early breach of a WP would almost certainly entail chemical interaction under substantially higher temperatures. Under such a scenario, evaporative concentration of reacting fluids would be expected to result in in-package fluids that diverge substantially from the compositions calculated under the 25°C, zero-evaporation limiting case considered here (see below). Various breach scenarios can be envisioned for the container, but as explained later, the current model assumes a fully flooded container (bathtub scenario) for the volume of water present in the container at all times. In addition, the in-package chemistry is dependent upon the degradation rates of the contents of the package. Two representative WPs were modeled, a CSNF package and a DSNF/HLW codisposal package (Figure 3.2-1). In both cases, there is an inner stainless steel disposal container, but the basket materials and waste forms are different and influence the fluid chemistry at least for short time periods. Direct use of a complex equilibrium code within the TSPA-SR analysis calculation was not practical; rather, a limited number of simulations were run with the complex equilibrium code, for a variety of input conditions and degradation rates of the contents. Regression equations were then developed for use directly in the TSPA-SR analysis calculations as explained below (Figure 3.2-2).3.2.1.1 In-Drift Fluids
The calculations for revision 00 of the Summary of In-Package Chemistry for Waste Forms AMR (CRWMS M&O 2000o) were performed in parallel with the testing and analysis of the effects of the near-field environment (NFE) on the water that may contact and enter the waste package. This parallel work is summarized in the NFE (CRWMS M&O 2000z), EBS (CRWMS M&O 2000y), and waste package degradation (CRWMS M&O 2000d) process model reports. However, for these first calculations, the J-13 well water composition was used for water entering the waste package (Harrar et al. 1990, Section 11). J-13 well water is quite dilute and its composition is not expected to significantly affect the in-package chemistry. On-going work will verify this assumption, or provide better information. J-13 well water is used as a surrogate for the groundwaters passing through the repository and possibly into breached WPs several thousand years after the repository has closed. The most abundant dissolved constituents in J-13 well water include Na and C, along with Si, Ca, K, F, Cl, N, and Mg. Dissolved Na, K, and Ca came from dissolution of feldspars into down-flowing meteoric water. Si was contributed by dissolution of feldspars and Si polymorphs. C and N came from equilibrium with soil and atmospheric gases encountered in transit. J-13 well water has undergone cation exchange with zeolitic rocks (causing increased Na/Ca) compared to UZ pore waters at the repository horizon. This is important for understanding the overall chemical budget of repository waters. Because Na and Ca play non-specific roles in the overall degradation process, ion exchange is less important to predicting what occurs once J-13 comes into contact with WPs. J-13 well water has a pH of ~8 and is thought to be close to equilibrium with carbon dioxide at levels slightly higher than that present in the atmosphere. The slightly higher than neutral pH of the solution is a direct result of alkalinity production that occurs when primary phase silicates are weathered. Waste package alteration scenarios are less sensitive to the exact composition of the water entering the package than they are to dissolution rates of a number of waste package components because: (1) The rates of WP component dissolution are not overly sensitive to minor variations in these parameters and (2) WP components rapidly come to dominate the dissolved phase after onset of reaction. Calculations have been started for a new version of the Summary of In-Package Chemistry for Waste Forms analyses and models report to test the sensitivity of the in-package chemistry to incoming water composition. It is expected that the incoming water will consist of the water found in the unsaturated zone subsequently modified by evaporation and/or condensation and fluid-rock interaction. Accordingly, sensitivity calculations are being done using a range of waters including the compositions in Table 3.2-1 as well as 100-fold evaporated J-13 water. Table 3.2-1 provides the major element composition of J-13 well water, Drift-Scale Test (DST) water, and the output of THC modeling. DST compositions represent an average of waters collected from Alcove 5 near the DST and were used as input compositions for modeling of thermal-hydrologic-chemical coupling in the NFE. DST water is thought to have moved appreciably through fractures and interacted with the rock matrix. The modeling output for waters expected to seep into the drifts is shown for two time periods, the transitional cool-down period (1,000-2,000 years) and the extended cool-down period of 2,000-100,000 years. By using these water compositions as well as 100-fold evaporated J-13 well water, the sensitivity calculations are covering a broad range of chemistries and should give a good indication of the sensitivity of the resulting in-package chemistry on the input water composition. The in-package chemical reaction calculations used input fluid flows of 1.5 to 150 L/yr. The temperature of the simulations was set at 25°C to represent the conditions that will occur several thousand years after WP emplacement, when the original thermal pulse has passed and temperatures have returned to near ambient levels. Waste form degradation may occur at temperatures up to 100°C. It is assumed that the process can be modeled adequately with the 25°C thermodynamic database. The calculations represent what occurs at times > 10,000 years, after the thermal pulse has passed and package temperatures are at, or below, 100°C. The justification for using 25°C thermodynamic data to model processes that might occur at somewhat higher temperatures is that many of the input thermodynamic parameters are not strongly sensitive to temperature over the range of 25 to 100°C, hence the broad scale features of the output fluid compositions are deemed independent of temperature. Dissolution rates of WP internal components would be greater at temperatures greater than 25°C. Typically, dissolution rates of minerals roughly double with every 10°C increase in temperature. In the absence of high temperature dissolution rates, this is probably a reasonable approximation for the behavior of metallic and oxide WP internal components as well. The net effect of temperatures greater than 25°C would therefore be a more rapid release of WP components into solution than what is reported here. Note that the existing calculations tend to err in this direction already because they use very high rates as input. Higher temperatures would also affect the boundary conditions of the calculation by decreasing the solubilities of oxygen and carbon dioxide in the fluids reacting with the WP. A more important impact of temperature on in-package chemistry would almost certainly be the evaporative concentration of inflowing and reacting solutions, which could conceivably change the concentrations of a number of important chemical species by orders of magnitude. Evaporation of J-13 water before interaction with WP components would result in a high ionic strength, alkaline solution. For example, 200-fold open-system concentration of J-13 water produces a pH 10, 1 mol/L ionic strength solution, which has 200 times as much chloride and total carbonate as J-13. Evaporation of J-13 type water is expected to lead to the formation of calcite, chalcedony, fluorite, and possibly sepiolite and other less abundant minerals. The important consequences of evaporation for fluid influx chemistry are increases in chloride, pH, and bicarbonate. Increased chloride impacts WP internal component corrosion. Glass and mineral dissolution rates tend to double for every unit increase in pH above 8. Many radionuclides are more soluble in the presence of bicarbonate. The sensitivity calculations will provide an indication of how important evaporation may be on in-package chemistry. The partial pressure of carbon dioxide and oxygen were set to, respectively, 10 -3.0 and 10-0.7 bar (DOE 1998c, Figure 3-36). The latter represents equilibrium with oxygen at atmospheric levels. Reaction with rock bolts (and steel supports) was neglected because the interaction of the fluid with waste package components would greatly outweigh any transient contact of the fluid with the rock bolts and the steel ground support. The total surface area of rock bolts likely to be encountered by seepage before contact with waste package components is small. It is likewise difficult to envision substantial interaction with the support material before seepage into the waste package.3.2.1.2 Water Contact Scenarios
As described in the Summary of In-Package Chemistry for Waste Forms Process Model CRWMS M&O (2000o), a simple degradation scenario for the WP entails breaching the WP (presumably through stress corrosion cracking and/or mechanical failure); filling of the WP void volume with seepage fluids from the drift; reaction within the WP; and releasing contaminated fluids at the same rate at which fresh fluids entered. This simple breach scenario was adopted and each breached WP was modeled as a continually stirred, fixed volume vessel. Specifically, the calculations used the solid-centered, flow-through mode of EQ3/6; in this mode, an increment of fresh fluid solution is added to the WP system and a like amount of solution is removed at each time step. The void volumes considered were a constant 4,507 L for the CSNF WP and a constant 5,811 L for the DSNF WP. The assumption of a continually stirred vessel is a simple and common approach to estimating the nature and extent of chemical reaction. The basis of the assumption is that either the fluid is continually stirred, or equivalently, the residence time of the fluids in the vessel is sufficient for diffusion to eliminate chemical gradients. In general, the latter basis is valid because of the geologic time scale of seepage and chemical reaction involved. Further justification is reasoned as follows. If wetting of WP components were complete, but void filling was not, the reduced fluid/solid ratio would tend to maximize the amount of dissolved solids in the effluent. On the other hand, the overall chemical reaction would be less for fluids that found themselves in "dead-end" channels, or if portions of the WP were physically inaccessible to fresh fluids. Furthermore, fluids that found "fast paths" that short-circuited interaction with WP solids would bear a much fainter signature of chemical reaction. Evaporation of fluids by residual heat could increase the salt content (i.e., ionic strength), and, thereby, increase some radioisotope solubilities, and possibly increase the rate of WF or WP degradation. The increased salt content, however, is not expected to significantly alter the overall reaction path (CRWMS M&O 2000o, Section 7) and was examined in a FEP on the use of J-13 well water (Section 2.3.2.2). This breach scenario provides an upper bound for in-package radionuclide transport and release. It does not include any time delay for water that enters the package; it does not include any retardation of radionuclides into immobile secondary phases by coprecipitation or sorption; it does not allow for radionuclides to be held-up in stagnant areas of the WP, and it maximizes contact between oxidizing fluids and WF and WP materials. The analyses using the in-package chemistry results also used bounding assumptions or large uncertainty (see Section 3.3, CSNF Matrix Degradation; Section 3.4, CSNF Cladding Degradation, Section 3.6, HLW Degradation, and Section 3.7, Dissolved Radioisotope Concentration). In particular, the possible effects of localized chemistry are included in the CSNF cladding degradation model.3.2.1.3 Waste Package Contents
In-drift solutions seeping into a breached WP would encounter a number of kinetically reactive solids whose reaction rates are only known within orders of magnitude (Table 3.2-2). Fluids intruding into WPs containing CSNF would encounter UO2 fuel wrapped in Zircaloy cladding, Al alloy, 316 stainless steel (with and without neutron absorbers, such as boron or GdPO4), and A516 low carbon steel. A range of degradation rates was used for each. Dissolution of the fuel, and release of radionuclides, occurs only after degradation of some of the cladding. General corrosion of cladding is likely to be insignificant under the geochemical conditions expected inside reacting WPs. Uncertainty in the initial failures, localized corrosion, mechanical damage, and other degradation mechanisms are addressed in the cladding degradation AMRs and were not available when the in-package calculations were started, so a range of clad damage and CSNF fuel exposure was investigated (100 percent, 20 percent, and 1 percent). The WP configuration used for the codisposal package calculations was that of the Fast Flux Test Facility (FFTF) DSNF with five HLW glass logs.3.2.1.4 EQ3/6 Reaction Path Modeling
As described in the Summary of In-Package Chemistry for Waste Forms AMR (CRWMS M&O 2000o) simulations of WP alteration by ambient groundwater were done using the qualified reaction path code EQ3/6 (CRWMS M&O 1998a; CRWMS M&O 1999c), which titrates masses of WP components at a rate determined by input reaction rates and fluid influxes into the breached WP. The software package, EQ3/6, Version 7.2b, was approved for QA work by LLNL and is identified as Computer Software Configuration Item (CSCI): UCRL-MA-110662 V 7.2b. EQ3/6 periodically assesses the chemical equilibrium state of the solution and removes newly saturated, secondary mineral or gas phases from the fluid. In addition to kinetic inputs (e.g., rates, compositions, and masses of reactants) EQ3/6 relies on a thermodynamic database describing the chemical stability of minerals, aqueous species, and gases. When the reacting solution becomes saturated with respect to solids or gasses, EQ3/6 converts dissolved components into the respective phase and then allows the latter to act as a reservoir of the respective components for use at subsequent times. In this way, EQ3/6 tracks the elemental composition of the reacting fluid for the duration of the reaction path calculation while at the same time providing estimates of the nature and masses of secondary phases that are predicted to form. The code, however, does not provide for the kinetic inhibition for phase formation, which is, therefore, the responsibility of the analyst. Generally, the assessment of kinetic inhibition must be made on the basis of there being an absence of the particular phase in low-temperature (~25°C) environments. For example, dolomite (CaMg(CO3)2) is predicted to form in a number of runs due to the accumulation of Ca, Mg, and alkalinity upon reaction with WP components. Dolomite, though thermodynamically favored to grow in a number of low temperature natural solutions such as seawater, apparently must overcome severe kinetic obstacles to actually form and is actually observed growing only in highly evaporated brines. These kinetic considerations are then the basis for suppressing formation of dolomite in the reaction path calculations. Similar arguments are used to suppress the formation of a number of oxide and silicate minerals that are typically observed to form only under high temperature conditions.3.2.1.5 Results of Waste Package and Waste Form Alteration Modeling
Table 3.2-3 summarizes the range of in-package fluid compositions predicted to occur for both the CSNF and codisposal packages (see Section 3.2.1.6 for further discussion of this chemistry and Section 3.2.2 for limitations). CSNF WPs–The majority of WPs in the repository will contain spent fuel rods from commercial reactors. The primary reactions that occur upon breach of CSNF WPs and subsequent alteration by intruding fluids are: (1) a rapid decrease in pH caused by dissolution of low-carbon steel and (2) a subsequent rise in pH caused by CSNF dissolution and oxidation of Al alloy (Figure 3.2-3). The latter rise is more pronounced under conditions of high clad failure. By the same token, the lowest predicted pHs are those calculated for the least amount of clad failure. Feedback between the pH and clad corrosion rate was not included within the calculations, as the pH is not expected to go low enough to cause significant increases in clad failure rates (see Section 3.4.2.4). Under the assumed fixed CO2 conditions of the simulations, carbonate levels are relatively high at high pH and low at low pH. Because both U and Pu form soluble complexes with carbonate, the dissolved levels of each tend to increase at high pH and decrease at low pH. However, the lowest pHs occur in the early stages of the reaction paths before appreciable U or Pu has been dissolved from the CSNF matrix. Minerals predicted to form during alteration of CSNF WPs include: copper minerals, silicates (clays, uranium silicates), oxides (metal oxides, actinide oxides), carbonates, phosphates, sulfates, and fluorides. Although the inputs to the calculation are normalized to a liter of fluid, and a liter volume is assumed to persist, the large volumes of lower density corrosion products (mainly hematite and UO3·2H2O) suggest that alteration products might actually fill the void space and seal off portions of the WPs to further fluid influx. This is further discussed in Section 3.4. Codisposal WPs–In all codisposal runs (Figure 3.2-4), the input pH of ~8.1 prevails initially. Then the system pH progresses towards a minimum, primarily due to the oxidation of the A516 carbon steel. The high specific rate and relatively high surface area of the latter mean that its dissolution tends to dominate the whole reaction path, at least as long as it remains. Oxidation of sulfur in the steel to sulfate is the primary proton-producing reaction. After the steel is exhausted, dissolution of base cation-containing HLW glass leads to increased pHs. pHs as high as 10 and ionic strengths as high as ~5.8 mol/L were predicted for codisposal WPs under conditions of high glass dissolution and low flow rates. Minerals predicted to form during alteration of codisposal WPs typically include the same minerals formed during CSNF alteration, though borates are also predicted to form, and the uranium silicates are more important.3.2.1.6 Abstraction of In-Package Chemistry
The predicted compositions of in-package fluids are of particular importance for estimating releases of radionuclides to the biosphere. The Summary of In-Package Chemistry for Waste Forms (CRWMS M&O 2000o, Section 6) provides reaction-path modeling assuming steady-state CO2 and O2 concentrations in the drift and breached WPs. The analysis indicates the broad range of effluent compositions that might be expected from a breached WP and provides a basis from which to establish abstractions describing radioisotope solubilities. The analysis of calculated in-package fluid compositions possesses four broad features. Carbonate alkalinity increases with pH due to the assumed fixed partial pressure of CO2. System Eh decreases with pH due to the assumed constant partial pressure of O2. Low early pHs are only seen with low cladding failures, and alkalinities typically correlate with high ionic strengths associated with high glass dissolution and low flow rates. The EQ3/6 simulations calculate the fluid chemistries that are operative over long periods of time (> 1,000 years). The simulations over this time period are typically different than those that prevail over shorter periods of time. Recognition of this change in process is important for estimating long-term behavior. Typically reaction times of less than 1,000 years result in relatively minor degradation of WP components; hence, the concentrations of many radionuclides are often well below their respective solubility limits. In effect their concentrations depend directly on the dissolution rates of WP components. At much greater time spans, many, but not all, radionuclides have reached saturation with at least one mineral phase. Once this has occurred, dissolved levels of the particular radionuclides will depend on the thermodynamics of secondary phase stability and much less directly upon the dissolution rates of the WP components. The abstraction of in-package processes is based upon a series of multiple linear regression analyses of the output from the EQ3/6 simulations as presented in the In-Package Chemistry Abstraction (CRWMS M&O 2000g, Attachment I). The multiple linear regression analyses take explicit account of the change in process control with time and treats results for times less than 1,000 years after waste package breach separately from those for later times. The specific relationships between pH, Eh, clad failure, and component degradation are illustrated in Figure 3.2-2. The predictions from the multiple regression analyses, and the output from the EQ3/6 simulations are presented in the abstraction AMR (CRWMS M&O 2000g) and summarized here. Because it is difficult to track the failures of individual packages in TSPA, the in-package chemistry abstraction was implemented in an approximate way for a group of packages. Let t be the current time and let tfail be the average time of failure for all the breached waste packages. The short-time abstraction would apply for t – tfail < 1000 years, and the long-time abstraction would apply for t – tfail > 1000 years. The pH abstraction follows two lines of reasoning based on the WP type (CSNF or codisposal ), and the difference in kinetic rate laws between the two. The rate law for CSNF is proportional to the hydrogen ion activity, i.e., proportional to pH, such that at low pH the dissolution rate increases (CRWMS M&O 2000o, Section 6.1.1). For HLW the rate law is "U" shaped with the minimum at pH 7 and the rate increasing above and below pH 7 (CRWMS M&O 2000o, Section 6.1.1). Therefore, conservative assumptions for one waste form may not be conservative for the other. In the case of the CSNF (and naval SNF), for example, assuming the lowest observed pH for the first 1000 years after breach, and the averaged pH thereafter would be the most conservative, while still honoring the pH–time history. However, for the codisposal package using the lowest observed pH would be conservative for the first 1000 years after breach but not for later times, when use of the maximum pH is conservative. The difference in the rate laws between HLW and CSNF and the difference in their pH–time profiles for the two waste forms predicates the use of different assumptions in the abstractions. For the CSNF, at times less than 1,000 years after waste package breach, minimum pH values for each flow/cladding/corrosion scenario were regressed to produce two abstractions of pH; one for low WP corrosion rates and the other for high WP corrosion rates, where the axes were x = cladding coverage, y = water flow (Q), and z = pH. This process was repeated for times greater than 1,000 years after breach, where the average pH for the entire modeled duration (0 to 10,000 years after breach) was used to calculate the average pH. The average pH for the entire duration is conservative because it includes the low pH values at early times, which tends to lower the average value compared to averaging over the late time information (1,000 to 10,000 years after breach). The resulting four data matrices were used to generate four response surfaces describing pH at less than and greater than 1,000 years after breach. These response surfaces were provided for TSPA to use in pH calculations. Table 3.2-4 provides the equations of the planes that bound the pH parameter space for CSNF, and Figures 3.2-5 and 3.2-6 show these planes. Each plane represents either the low or high corrosion rate data for the WP materials, 316 stainless steel, aluminum alloy, A516 carbon steel, and borated stainless steel. An uncertainty range of ±1 pH unit was recommended for TSPA. Table 3.2-5 and Figures 3.2-7 and 3.2-8 show the equations describing minimum and maximum pH parameter space for codisposal WPs. The balance of the acid production from the Alloy 516 carbon steel and base production from the glass determine the pH. At early times (< 1,000 years after breach), acid production dominates, but as the steel is consumed, base production overcomes the acid production. In most cases, this is not calculated to occur until considerably after 10,000 years after breach. To obtain conservative glass degradation rates, the maximum pH in the 100,000 year simulations was used. Due to the simplicity of the mathematical expressions used to calculate pH, it is possible, by increasing the water flow Q, to calculate unrealistic pH values. Therefore, it was necessary to set limits for the maximum allowable Q. Since the calculated pH from all of the expressions in Tables 3.2-4 and 3.2-5 are directly proportional to water flow through the WP and as the water flow increases the potential for reaction of the through-flowing water with the WP/waste form materials decreases, the pH of the water exiting the WP should approach that of J-13 well water. Therefore, above some maximum Q-value, the water exiting the WP should have the same pH as the water entering the WP. This critical value of Q was calculated by solving the expressions in Tables Tables 3.2-4 and 3.2-5 for Q and using the input, J-13 well water value for pH and assuming linear behavior. For CSNF at time less than 1,000 years after breach, the pH expressions (Table 3.2-4) are valid within the range of 0 < Q < 5.3 (m3/yr). Likewise, when Q
5.3 m3/yr, then the pH is constant at 8.1 (±1). For time greater than 1,000 years after breach, the pH expression (Table 3.2-4) is valid for 0 < Q < 0.24 (m3/yr), and when Q 0.24 m3/yr the pH is constant at 8.1 (±1). Similarly for codisposal waste packages, the pH expressions (Table 3.2-5) are valid within the range of 0 < Q < 1.1 (m3/yr) at time less than 1,000 years after breach and 0 < Q < 0.36 (m3/yr) for time greater than 1,000 years after breach. Above these values, the pH is constant at 8.1 (±1).
Total dissolved carbonate affects the solubilities of a number of actinide phases as well as the degradation rate of CSNF. The total dissolved carbonate is abstracted as a function of pH by assuming a partial pressure of atmospheric CO2 = 10-3 atm and equilibrium between the carbonate, bicarbonate, and carbonic acid in solution:
10-3 to 2.922 10-3 mol/L) and glass rate/corrosion rate scenario (beta distribution from 2.537 10-3 to 3.479 10-3 mol/L). At times greater than 1,000 years after breach, the average value of ionic strength was used from each scenario to get the ionic strength range used in the abstraction (beta distributions from 2.827 10-3 to 3.943 10-1 mol/L for CSNF and from 7.86 10-3 to 1.353 mol/L for codisposal). This is a conservative approach compared to using the maximum value because low ionic strength is conducive to colloid formation, and using the minimum value would also create the worst case, which is not reflective of the actual in-package conditions.
Fluoride was considered in the cladding degradation model. Fluoride was only pertinent for the CSNF and naval SNF because there was no cladding protection included in the codisposal package modeling. The CSNF calculations indicate that the fluoride concentration is controlled by fluorapatite. The abstracted distribution was normal between 1.148 10-4 and 9.114 10-4 mol/kg with a standard deviation of 2.2 10-4 mol/ kg. Chloride levels remain constant at 2.0
10-4 mol/kg and therefore are too low to influence the models that rely on bulk chemistry.
3.2.2 Confidence/Limitations/Validation
This work represents the first attempt by the DOE to model the chemical interactions that occur within the WP for use within the TSPA calculations. Only limited sensitivity runs have been completed to study the importance of issues such as: