Electrometallurgical Techniques for DOE Spent Fuel Treatment: Status Report on Argonne National Laboratory's R & D Activity Through Spring 1997 (1997)

Since the fall of 1997, ANL has achieved satisfactory progress in the preparation, characterization, and testing of development-scale ceramic and metal waste forms. The committee has concerns, however, regarding characterization of product homogeneity and scale-up of the manufacturing process for the ceramic waste form, as well as whether batch loading options may supplant ion-exchange column loading of radionuclides into zeolite or sodalite.

A key portion of the electrometallurgical technology is the creation of two novel waste forms. One is a metallic waste form composed predominantly of cladding metal. The second is a “ceramic” waste form composed of synthetic zeolite (or sodalite) loaded with fission products and trans-uranium elements and then encapsulated within a nonradioactive borosilicate glass-bonded matrix (also called glass-bonded zeolite, or GBZ) by hot isostatic pressing (HIP).

Scale-up, remote design, and “cold” operation of the column ion-exchange system for the loading of radionuclides into zeolite (or sodalite) are planned in parallel with the EBR-II demonstration. The committee learned, however, that no “hot” zeolite or sodalite will be generated by column ion-exchange. Part of the rationale for this approach is that the preparation of fully loaded “hot” ceramic waste forms must wait until the disposable fission products and transuranic ions have increased to higher levels in the process salt. With the quantities of fuel planned for the demonstration as reduced by the Environmental Assessment (100 driver and 25 blanket assemblies), the radionuclides will not increase to the level that requires use of the column ion-exchange until after the completion of the planned EBR-II demonstration.

The committee remains concerned about whether at the conclusion of the EBR-II demonstration (June 1999) there will be sufficient information available on ion-exchange loading of zeolite or sodalite under radioactive operating conditions and at relevant radionuclide concentrations to provide a basis for a decision to proceed with using electrometallurgical technology to treat spent fuel. In particular, it is worrisome that “hot” column ion-exchange tests will not be part of the EBR-II demonstration and that no fully loaded samples of the expected final ceramic waste form will be produced or characterized. DOE should consider whether the absence of such information will create difficulties in reaching a decision to proceed with this technology after the EBR-II demonstration is completed.

ANL plans to select a reference crystalline matrix for incorporation of radionuclides by October 1997. Alternative crystalline host phases being considered include the following:

1. Linde Type-A (LTA) zeolite alone, or LTA zeolite with chabazite (CHA) zeolite, or

2. Zeolite converted to sodalite (a felspathoid).

The conversion of the radionuclide-loaded zeolite to a denser form (e.g., sodalite) will require particular attention to the uniformity of the distribution of the salt between zeolite crystals. Because the denser phase contains less salt, it may be necessary to limit the highest salt loading in each individual zeolite crystal to no greater than the amount that can be contained in the individual sodalite crystals. Otherwise, portions of the salt (and radionuclides) can be expected to reside outside of the sodalite and, possibly, to be more easily leached from the final waste form. ANL should remain alert to issues regarding the scale-up of the manufacturing process for any of these waste forms, especially with regard to the conditions (time, temperature, mixing, etc.) required to achieve an adequate degree of uniformity of salt loading inside each of the crystalline host phases.

ANL expects to produce radioactive samples of its ceramic waste form by the end of the EBR-II demonstration period in 1999. It plans to demonstrate all of the key steps for ceramic waste form preparation, including creating radionuclide-loaded zeolite (or sodalite) and HIP of these crystalline phases with glass to manufacture the final waste form. The samples planned are as follows:

1. A Pu-238 loaded zeolite, to be tested to assess potential alpha-recoil damage;

2. A zeolite loaded from the “spent” processing salt of the EBR-II demonstration at ANL-W (the so-called “Throw-Away Option”); and

3. A zeolite batch ion-exchange equilibrated with salt in the electrorefiner, also to be done in parallel with the ANL-W demonstration (the so-called “Batch Equilibration Option”).

None of these samples closely matches the composition and radionuclide loadings to be expected in zeolites from actual column ion-exchange operations. Such samples are postulated, however, as providing preliminary and representative information regarding the performance of radionuclide-bearing zeolites that will be prepared in actual column ion-exchange operations after the EBR-II demonstration.

DOE must, therefore, consider the question of whether this approach will provide adequate information to proceed beyond the June 1999 demonstration. Implementation of the “Throw-Away

Option” or “Batch Equilibration Option” as an alternative to column ion-exchange for the preparation of the final ceramic waste form would concern the committee. In particular, these options significantly increase the volumes of radioactive waste forms produced and should be carefully evaluated.

Small-scale samples of the metal waste form have been fabricated by melting the cladding residues at 1600°C in an inert atmosphere environment and subsequently casting them into ingots. Noble-metal fission products, cladding, and alloying zirconium constituents from the original EBR-II fuel are expected to reside in this metal waste form. Tests to date show satisfactory corrosion, metallurgical, and mechanical properties.⁶ Production of “hot” metal waste forms from the demonstration products is scheduled to commence in April 1998 in order to evaluate scale-up issues.

Because of the extremely low corrosion rate observed for the metal waste form (similar to rates for stainless steel and Zircaloy), no formal waste form qualification testing procedures are being developed or applied. While this assumption of waste form acceptability does not appear to be unreasonable, the committee believes that by June 1999, ANL ought to apply its corrosion rate data to the same conceptual water contact and release models that are being developed for the ceramic waste form.

ANL is now implementing a Qualification Testing (QT) program for its proposed ceramic waste form, with less emphasis on qualification testing of its proposed metal waste form. The committee is awaiting ANL's detailed project plan to assess the adequacy and scheduling of QT activities. ANL has indicated that such QT activities are likely to include conceptual model development to identify important performance characteristics of the waste form and evaluation of characteristics under simulated repository disposal conditions.

ANL has recognized the importance of developing a conceptual model to assess how its proposed ceramic waste form will behave under long-term conditions in a deep geologic repository. ANL's preliminary conceptual model for reaction of the ceramic waste form with repository groundwater and subsequent transport of released radionuclides is divided into three steps:

• Step I: contact of the ceramic waste form with water and/or water vapor;

• Step II: primary release of radionuclides from the ceramic waste form; and

• Step III: subsequent migration of radionuclides, either into secondary solid phases or as mobile aqueous species (or colloids) undergoing diffusive or advective-diffusive transport.

With respect to Step I, one issue to be resolved is whether the encompassing, nonradioactive glass binder will provide physical isolation of the radionuclide-loaded zeolite/sodalite, limiting the available area of the ceramic contacted by water. Demonstrating physical protection for long periods of time may be difficult and should be based on consideration of the full range of thermal, mechanical, hydrological, and chemical conditions of planned geologic repositories.

In contrast to its physical effects, the presence of the glass binder will undoubtedly have an impact on the chemical reactions occurring between water and the radionuclide-bearing zeolite/sodalite. For example, glass dissolution may be a source of solvated cations that could expedite the subsequent ion-exchange release of radionuclides from the ceramic waste form. Furthermore, water-glass reactions typically lead to increases in pH and alkaline conditions, which could increase the dissolution rate of zeolitic materials.

ANL does not yet have enough QT data on its proposed ceramic waste forms to unambiguously distinguish which release mechanism (Step II)—ion-exchange or dissolution—controls the primary release of radionuclides. Ion-exchange is now acknowledged as one likely mechanism, supported by comparison of the release behavior of nonradioactive ceramic waste forms by ANL using dilute brine contacting waters. ANL plans an extensive set of both long-term and short-term tests (see following section) to attempt to evaluate which primary release mechanism dominates at what time period.

Step III relates to the impact of transport conditions of a repository on the long-term rate of migration of radionuclides released from the waste packages. Information on the expected flow and transport conditions from DOE's high-level waste (HLW) repository program is needed so that reasonable test conditions can be simulated in planned long-term QT of the ceramic waste form.

ANL has listed a broad range of planned testing techniques that will form the initial phase of QT for its ceramic waste form. ANL noted that many of the same staff and much of the same analytical equipment necessary to support QT must be shared with parallel activities such as waste form preparation and characterization.

In view of ANL's limited human and facility resources, it would be prudent for ANL to involve external technical experts to review the need for a full set of tests to support QT. There may be needless overlap and redundancy among these proposed tests. Furthermore, alternative unsaturated