3 Research and Development in Support of Argonne National Laboratory 's Demonstration Project

METAL WASTE FORM TESTING AND PLANS

ANL has developed a metal waste form (MWF) test plan, a progress implementation test plan, and plans for qualification testing beyond June 1999.1 ANL plans a total of 856 tests in the MWF test plan.

The MWF test plan consists of attribute tests, characterization tests, accelerated tests, and service condition tests.2 Using electron microscopy, X-ray analysis, and neutron diffraction, the attribute tests are designed to provide information on properties of materials. Very good progress seems to have been achieved in identification of the various microstructures of SS-15 Zr-type materials. Noble-metal-rich precipitates have not been observed.

The characterization tests, 220 of which have been initiated, consist of immersion of the sample in sealed Teflon vessels at 90 °C in J-13 (simulated Yucca Mountain groundwater) and in deionized water.3 The 80 tests that have been terminated showed either no attack or only minor tarnish, and the test solutions have been submitted for elemental analysis. In a presentation to the committee on October 26, 1998,4 it was pointed out that some of the corrosion products might remain on the sample surface. A large number of the samples to be tested differ only very slightly in minor alloying elements.

The tests designed to accelerate corrosion rates involve immersion of the sample in deionized water in a titanium vessel at 200 °C for 28 days.5 In the six alloy compositions that have been tested, corrosion rates were very low and no correlation of elemental leaching with alloy composition was found.

Electrochemical corrosion testing is based on polarization resistance techniques [American Society for Testing and Materials G-59 (ASTM G-59)].6 At the committee's meeting in June 1998 at ANL-W, discussions included the observation that this method does not accelerate corrosion rates and is used to measure “instantaneous” corrosion rates. It was not clear whether the effect of the uncompensated solution resistance is eliminated by the use of an interrupter technique or other means in the present tests. If not, corrosion rates can be underestimated. Of the 360 tests planned, 105 were completed at the time of the committee's meeting in October 1998 at ANL-E. Corrosion rates of the MWF alloys

1  

D. Abraham, ANL-W, presentation to the committee, October 26, 1998, Argonne, IL.

2  

R. W. Benedict, H. F. McFarlane, J. P. Ackerman, R. K. Ahluwalia, L. L. Briggs, H. Garcia, E. C. Gay, K. M. Goff, S. G. Johnson, R. D. Mariani, S. McDeavitt, G. A. McLennan, C. Pereira, P. D. Roach, and B. R. Westphal, Spent Fuel Treatment Demonstration Interim Status Report, ANL-NT-74, Argonne National Laboratory, Argonne, IL, pp. 16-19.

3  

EBR-II Spent Fuel Treatment Program Monthly Report, June 1998, NT Technical Memorandum No. 84, Argonne National Laboratory, Argonne, IL, pp. 76-78.

4  

D. Abraham, ANL-W, presentation to the committee, October 26, 1998, Argonne, IL.

5  

EBR-II Spent Fuel Treatment Program Monthly Report, July 1998, NT Technical Memorandum No. 86, Argonne National Laboratory, Argonne, IL, pp. 110-114.

6  

EBR-II Spent Fuel Treatment Program Monthly Report July 1998, NT Technical Memorandum No. 86, Argonne National Laboratory, Argonne, IL, pp. 114-115.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 3 Research and Development in Support of Argonne National Laboratory 's Demonstration Project METAL WASTE FORM TESTING AND PLANS ANL has developed a metal waste form (MWF) test plan, a progress implementation test plan, and plans for qualification testing beyond June 1999.1 ANL plans a total of 856 tests in the MWF test plan. The MWF test plan consists of attribute tests, characterization tests, accelerated tests, and service condition tests.2 Using electron microscopy, X-ray analysis, and neutron diffraction, the attribute tests are designed to provide information on properties of materials. Very good progress seems to have been achieved in identification of the various microstructures of SS-15 Zr-type materials. Noble-metal-rich precipitates have not been observed. The characterization tests, 220 of which have been initiated, consist of immersion of the sample in sealed Teflon vessels at 90 °C in J-13 (simulated Yucca Mountain groundwater) and in deionized water.3 The 80 tests that have been terminated showed either no attack or only minor tarnish, and the test solutions have been submitted for elemental analysis. In a presentation to the committee on October 26, 1998,4 it was pointed out that some of the corrosion products might remain on the sample surface. A large number of the samples to be tested differ only very slightly in minor alloying elements. The tests designed to accelerate corrosion rates involve immersion of the sample in deionized water in a titanium vessel at 200 °C for 28 days.5 In the six alloy compositions that have been tested, corrosion rates were very low and no correlation of elemental leaching with alloy composition was found. Electrochemical corrosion testing is based on polarization resistance techniques [American Society for Testing and Materials G-59 (ASTM G-59)].6 At the committee's meeting in June 1998 at ANL-W, discussions included the observation that this method does not accelerate corrosion rates and is used to measure “instantaneous” corrosion rates. It was not clear whether the effect of the uncompensated solution resistance is eliminated by the use of an interrupter technique or other means in the present tests. If not, corrosion rates can be underestimated. Of the 360 tests planned, 105 were completed at the time of the committee's meeting in October 1998 at ANL-E. Corrosion rates of the MWF alloys 1   D. Abraham, ANL-W, presentation to the committee, October 26, 1998, Argonne, IL. 2   R. W. Benedict, H. F. McFarlane, J. P. Ackerman, R. K. Ahluwalia, L. L. Briggs, H. Garcia, E. C. Gay, K. M. Goff, S. G. Johnson, R. D. Mariani, S. McDeavitt, G. A. McLennan, C. Pereira, P. D. Roach, and B. R. Westphal, Spent Fuel Treatment Demonstration Interim Status Report, ANL-NT-74, Argonne National Laboratory, Argonne, IL, pp. 16-19. 3   EBR-II Spent Fuel Treatment Program Monthly Report, June 1998, NT Technical Memorandum No. 84, Argonne National Laboratory, Argonne, IL, pp. 76-78. 4   D. Abraham, ANL-W, presentation to the committee, October 26, 1998, Argonne, IL. 5   EBR-II Spent Fuel Treatment Program Monthly Report, July 1998, NT Technical Memorandum No. 86, Argonne National Laboratory, Argonne, IL, pp. 110-114. 6   EBR-II Spent Fuel Treatment Program Monthly Report July 1998, NT Technical Memorandum No. 86, Argonne National Laboratory, Argonne, IL, pp. 114-115.

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 in J-13 and in solutions at pH 2, 4, and 10 were low and similar to those of SS-316 and Alloy C-22. Corrosion rate data for MWF materials were also compared to those for copper and mild steel. These results are not surprising, considering that the solutions tested did not contain chloride ions that could have initiated localized corrosion. The list of tests proposed to be performed after June 1999 had not been finalized completely. Electrochemical tests are to be performed at elevated temperatures in order to assess the effect of increased temperature on corrosion rates. Given that electrochemical processes proceed at higher rates at higher temperatures, it might be better to concentrate on a few key samples, expose them at higher temperatures, and obtain electrochemical and surface analysis data. Tests were also to be conducted in chloride solutions at concentrations up to 10,000 ppm. These are credible conditions that might be encountered in a repository. In this case it might be better to carry out pitting scans according to ASTM G-61, which allows a comparison of different alloys' relative susceptibility to pitting. Data on corrosion rate can also be obtained from these measurements. Finally, a study of corrosion mechanisms was proposed by ANL. Metal Waste Form Release Rate Work was presented to the committee at its meeting at ANL-E on the development of a radioisotope release rate model for the stainless steel MWF.7 Because the MWF behaves much like stainless steel, the modeling approach is based on the results of ANL's current uniform corrosion test program, the use of known stainless steel degradation mechanisms as a basis of MWF release modeling, and the use of MWF corrosion rate data to adjust the models empirically. Factors considered important for radioisotope release include MWF metallurgy, degradation mechanisms, and environmental conditions. A number of potential degradation mechanisms were discussed, including uniform corrosion, localized corrosion (crevice, pitting), microbially influenced corrosion (MIC), selective leaching, intergranular attack, and galvanic corrosion. 8 (Models will also be established for the chemistry of the repository water.) Waste canister modeling is concerned with the corrosive attack of the carbon steel and C-22 shells. Corrosion testing has used concentrated J-13 well water. In discussions between representatives of ANL and the committee, it was suggested that it is highly unlikely that any corrosive attack of C-22 would be observed under the mild conditions used in the corrosion testing. Corrosion models for long-term waste canisters will be based on empirical multivariate regressions. Key variables considered in ANL's corrosion testing included temperature, pH, and chloride content. For stainless steel, additional variables such as H2O2, HCO3-, and NO3- concentrations will be taken into account. The experimental data to be used for the development of models for uniform corrosion were those discussed in ANL's presentation at the committee's meeting at ANL-E in October 1998. It was mentioned that additional tests were scheduled to reduce model uncertainties, although it was not clear what types of tests would be used. In 7   M. C. Petri, ANL-E, presentation to the committee, Argonne, IL, October 26, 1998. 8   In discussions between representatives of ANL and the committee at the committee's meeting at ANL-W in June 1998, a member of the committee suggested that it was unlikely that MIC would play an important role and that obtaining meaningful MIC data would probably be quite difficult.

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 discussions between representatives from ANL and the committee, it was pointed out that so far most corrosion tests have shown very little if any corrosion damage of MWF alloys. As a result of the lack of observed corrosion damage, the tests that were being performed at the time of the committee's October 1998 meeting may not yield the information needed by ANL for its planned MWF release rate modeling. MWF samples analyzed in a variety of tests so far all indicate little or no corrosion of the MWF. These include samples from immersion testing in sealed Teflon vessels at 90 °C in simulated J-13 and deionized water for up to 3 years, samples immersed in deionized water in a titanium vessel at 200 °C for 28 days, samples that were exposed to water vapor for 56 days at 200 °C, and samples from electrochemical corrosion testing. Each of these tests was performed, and continued to be performed, under a variety of mild experimental conditions, and ANL reported that up to the time of the committee's October 1998 meeting, no sample had shown more than slight tarnish. FINDINGS Since no corrosive attack was observed in the immersion tests conducted thus far, it is unlikely that tests under mild conditions will yield any additional information. The lack of information from these tests under mild conditions has been a general observation. As a result, it is questionable if the large number of MWF tests (856) is really necessary. The proposed study of crevice corrosion requires careful design of an artificial crevice with consideration of the proposed application of the alloys used in the MWF. ANL's observation that the MWF behaves like stainless steel is critical in that the overall MWF test program is modeled after tests used for stainless steel. MWF tests to date have been conducted under mild conditions, and little attack has been observed. RECOMMENDATIONS Surface analysis by X-ray photoelectron spectroscopy (XPS) or Auger electron spectroscopy (AES) should be performed for selected samples drawn from the characterization tests. The committee notes that ANL-E has one of the leading experts in this area. It is recommended that only a few of these samples be fully characterized. ANL needs to refocus near-term testing by reducing product consistency testing under relatively mild conditions and instead emphasizing product performance testing under more stringent conditions that may reveal significant corrosion effects and address success criterion 2, goal 2 (Appendix B).

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 ANL should evaluate how the MWF performance model will be used and whether ongoing MWF testing will provide information needed for developing the performance model. CERAMIC WASTE FORM TESTING AND DEVELOPMENT Considerable progress has been made in the demonstration-scale ceramic waste form (CWF) operations.9 The reference ceramic waste form for the demonstration is glass bonded sodalite formed by hot isostatic pressing (HIP).10 Sodalite was the thermolysis product formed during HIP of the salt-loaded zeolite 4A. Key accomplishments included production of 61 demonstration-scale HIP cans, production of salt-loaded zeolite within process specifications using the V-mixer, completion of a mill/classifier process test matrix, use of a crusher for pre-mill salt processing, and routine production of dried zeolite at demonstration scale. Results from performance testing indicated that the demonstration-scale samples are as durable as laboratory-scale samples. Scale-up problems encountered in heating of the V-mixer, and particle size mismatch between the salt and zeolite, 11 have been resolved successfully. At the time of the June 1998 meeting at ANL-W, the HIP equipment was just being installed into the Hot Fuel Examination Facility. Hot Isostatic Pressing Hot isostatic pressing technology arose from pioneering work conducted in the 1960s at Battelle's Columbus laboratory. After decades of improvements, the technology is now widely used to (1) achieve theoretical densification of powder compacts, including metal, ceramic, and composite systems, and (2) heal shrinkage porosity that is commonly encountered in precision-cast parts. The utility of the technology is indicated by its adoption (as sinter-HIP) by the hard-metal industry to produce drill bits and machine tools, and by the gas-turbine engine industry to produce pore-free castings and fully consolidated preforms for hot forging. However, despite these successes, concerns remain about the safety of HIP technology. For example, an incident described in the technical press involved unexpected failures of the pressure wall and explosive decompression of the system.12 As a result, much effort has been devoted to establishing protocols that enable essentially low-risk operations. Nevertheless, an occasional failure 9   R. W. Benedict, presentation to the committee, June 25, 1998, Idaho Falls, ID, and October 26, 1998, Chicago, IL. 10   EBR-II Spent Fuel Treatment Program Monthly Report, August 1998, NT Technical Memorandum No. 91, Argonne National Laboratory, Argonne, IL, pp. 90-94. 11   From ANL Monthly Highlights of the Electrometallurgical Treatment Program (Appendix C of this eport): “Tests were initiated using the reference, as-procured, granulated zeolite from an outside vendor (UOP) in the demonstration-scale equipment. Problems were encountered with segregation of salt and zeolite in the blending operation in the V-mixer. The reason for the segregation was determined to be the different particle sizes of the salt and zeolite. Modification to the salt grinding operation are being implemented to remedy this problem.” 12   International Journal of Powder Metallurgy, APMI International, October 1998, p. 12.

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 of an operational unit is to be expected. Thus, in any practical use of HIP technology, it is prudent to balance the risks against the benefits. ANL's research staff have conducted such a risk assessment and arrived at a protocol that ANL finds acceptable, despite the fact that the operation eventually will be conducted in a hot cell.13 The HIP treatment is the last step in ANL's procedure for the production of ceramic waste forms. In a typical case, the radionuclides are incorporated within a matrix of zeolite 4A (70 wt %) that is first granulated with a clay binder phase (5 wt %), and then blended with glass 5714 (25 wt %). The blended mixture is finally canned and subjected to HIP at ~14,000 psi/600 °C to realize a fully dense glass-bonded ceramic in which the high-level waste is packaged in a form suitable for interim storage at ANL-W's radioactive scrap and waste facility. ANL-W has entered into a cooperative effort with the Australian Nuclear Science and Technology Organisation. The joint project will provide the first demonstration of radioactive waste processes that use the HIP process in a remotely controlled hot cell. The committee reconsidered the HIP approach in light of recent ANL-E research15 on hot uniaxial pressing (HUP), and pressureless consolidation as an alternative to HIP. In the new HUP approach, the amount of glass 57 is increased to about 50 wt %, which enables densification to be accomplished by hot pressing. ANL-W stated that although the amount of solid waste form for disposal is necessarily increased, this new approach is an attractive alternative to HIP. ANL-W's stated reasons for this view included the following: The compression density realized by pressureless consolidation is comparable to that of HIP; HUP technology removes uncertainties about the safety of the final consolidation step in the processing of the ceramic waste form; Preliminary data on leaching for material generated by pressureless consolidation is comparable to that for HIP-processed material; It appears that there are no impediments to scale-up of HUP technology; The additional glass matrix phase in HUP-processed material provides greater assurance that the radionuclides are more effectively contained in the zeolite matrix phase (more correctly sodalite, which is the transformation product of salt-loaded zeolite during hot pressing); The higher glass matrix content in HUP-processed material also provides a more effective barrier against sodalite's exposure to ground water; The greatly simplified procedure for HUP processing should markedly increase throughput of the ceramic waste form; and The powder preparation procedure for HUP is identical to that for HIP. 13   Morman, J. A. USQ Safety Assessment: FCF Electrorefiner Waste Salt Processing in HFEF, Argonne National Laboratory, Idaho Falls, ID, 1998. 14   Glass 57 is borosilicate glass that has a specific composition having approximately 66.5% SiO2, 19.1% B2O3, 6.8% Al2O3, 7.1% Na2O and 0.5% K2O. 15   John P. Ackerman, ANL-E, presentation to the committee, June 25, 1998, Idaho Falls, ID, and October 26, 1998, Chicago, IL.

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 Against these advantages, the committee notes that with HUP technology the amount of sodalite, and therefore radionuclides, per volume of as-produced waste form would be reduced. Thus, about twice as much waste form would have to be processed in order to dispose of the same amount of fissionable material. Given that the total amount of EBR-II spent fuel to be processed into the ceramic waste form is not large, this may be considered an acceptable penalty. It appears that there is sufficient interim storage capacity at ANL-W's facility to handle the quantity of waste form produced. Subsequent to the committee's meetings at ANL-W and ANL-E in June and October of 1998, ANL confirmed that HIP is the baseline reference consolidation process to be used in the demonstration project. The ceramic waste form process specifications will be developed for the HIP process only. The HUP process is not being considered as an alternative to the HIP process. ANL intends to evaluate pressureless sintering for post-demonstration use in early 1999. RECOMMENDATION Assuming that an increased amount of glass in the waste form is acceptable, then in addition to HUP, conventional cold pressing and sintering should also be considered as a viable processing option. It may turn out that pressureless sintering offers technical advantages over HUP, especially in view of the difficulties associated with conducting ceramic waste form operations in a hot cell using remote handling methods, not to mention the particularly onerous task associated with repair and maintenance of equipment. Additional Ceramic Waste Form Development Issues The issue of reaction between excess UCl3 and the initial zeolite-A host remains unresolved. For ANL's batch-loading procedures, which form the basis of the current EMT demonstration project, the water that remains on “dried” zeolite seems to react with excess UCl3 to protect the zeolite structure. However, in column-loading procedures that were originally envisioned as the preferred method of loading actinides and non-noble fission products into the zeolite, the presence of excess water is not possible. As a result, it appears that the problem of creating a scaled-up process for loading of radionuclides onto the zeolite will not be resolved by the end of the demonstration project in 1999. ANL also noted a related concern regarding the potential for reaction of PuCl3 with the zeolite. Tests to date, over a wide range of plutonium loadings, show no indication of such a reaction. However, the combined effects of uranium, plutonium, and other radionuclides at levels expected for final zeolite loading have not been investigated as part of the demonstration project. The committee is concerned about the distribution of the waste radionuclides in the CWF. To date, the characterization tests in the waste form plan provided to the committee (e.g., transition microscopy and SEM-element mapping to investigate the distribution of elements directly) have not been carried out. The committee notes some of the ionic size considerations that may affect the retention of the radionuclides in the

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 zeolite 4A and in its sodalite transformation product. The pore size of the sodalite structure has an effective diameter of 2.2 Å. Some of the radionuclides have ionic diameters larger than 2.2 Å (Cs+1, d = 3.34 Å; Rb+1, d = 3.04 Å; Cl-1, d = 3.62 Å; and I-1, d = 4.40 Å). Thus these ions would not be present in the sodalite cages contained in the original zeolite 4A structure. The mechanism, now unknown, of transformation of salt-loaded zeolite 4A to sodalite may affect whether or not radionuclides with diameters larger than 2.2 Å will be retained in the sodalite transformation product. RECOMMENDATION The committee believes that characterization of the ceramic waste form should be accelerated in order to determine the mechanism of transformation of salt-loaded zeolite 4A to sodalite. Demonstration-Scale CWF Processes In presentations to the committee, representatives from ANL reported a problem, attributable to grain-size differences in the zeolite and the salt, in achieving desired, reproducible radionuclide-loading of zeolite in the batch process. Although some progress had been made, attempts to optimize the process are continuing, and an empirical process model based on important processing variables is being developed by ANL's program on EMT. The potential impacts of alpha-decay and recoil damage to the CWF are being examined by a number of selective doping tests involving 239Pu and 238Pu. No results are yet available. No mention was made by ANL representatives regarding tests on possible transmutation effects on waste form stability, such as short-lived, univalent 137Cs decay to stable, divalent 137Ba. Local charge balance may be an important aspect of the long-term structural stability of zeolite and sodalite host matrices for radionuclides. ELECTROMETALLURGICAL TREATMENT OF OXIDE AND ALUMINUM ALLOY FUELS Oxide Fuels ANL is pursuing work on electrometallurgical treatment of oxide-containing spent DOE fuels,16 which cannot be electrorefined directly but must first be converted to metals.17 Lithium metal is used as the reducing agent in molten LiCl to effect this conversion, and the Li2O that is produced dissolves in the molten salt. The salt and lithium metal are recycled by electrochemical deposition of lithium at a cathode and evolution of oxygen at an anode. 16   C. C. McPheeters, ANL-W, presentation to the committee, October 26, 1998, Chicago, IL. 17   Electrometallurgical Techniques for DOE Spent Fuel Treatment: Spring 1998 Status Report on Argonne National Laboratory's R&D Activity, National Academy Press, Washington, D.C., 1998, p. 44.

OCR for page 15
ELECTROMETALLURGICAL TECHNIQUES FOR DOE SPENT FUEL TREATMENT: STATUS REPORT ON ARGONNE NATIONAL LABORATORY'S R&D ACTIVITY AS OF FALL 1998 Laboratory-scale work has been conducted to investigate the limits for reduction of PuO2, elucidate the kinetics of UO2 reduction, develop an electrowinning cathode for handling metallic lithium, and evaluate possible anode materials. Available thermodynamic information on both Pu2O3 and PuCl3 introduces uncertainties about the extent of reduction of Pu2O3 and the behavior of the resulting PuCl3 in the electrorefiner. Several anode materials have been tested. Platinum performs well and has low overpotential but is expensive and reacts with both metallic lithium and gaseous chlorine. Iron oxide (Fe3O4) is cheap and leads to an already existing corrosion product but presents thermal shock and fabrication difficulties. Doped tin oxide (SnO2) is cheap and is commercially available, but results in a high cell overpotential. The desired anode reaction is production of gaseous oxygen from Li2O. Use of a high cell overpotential can result in production of gaseous chlorine, which would react with the anode material. Although several anode materials have been tested, including platinum, iron oxide, and doped tin oxide, the committee believes that further progress will be enhanced through analysis of published material in this area. Work is in progress to study oxide fuel reduction kinetics, optimum fuel basket design for oxide fuel reduction and electrorefining, and development of methods for handling metallic lithium. Based on the six engineering-scale tests conducted to date, the committee agrees that lithium electrowinning has been successfully demonstrated using a cathode consisting of stainless steel screen wrapped on a stainless steel rod, that progress has been made on understanding factors important in oxide fuel reduction kinetics, and that the reduction step can be interfaced successfully with the electrorefining step without carryover of metallic lithium or Li2O. Aluminum Alloy Spent Fuels The feasibility of electrometallurgical treatment of aluminum alloy spent fuels has been demonstrated in laboratory-scale experiments. 18 The key step is electrorefining of the aluminum, which represents about 90% of the spent fuel volume and which can potentially be discarded as low-level waste. ANL has developed a flow sheet in which initial separation of the aluminum as a metal waste is followed by separation of metallic uranium from fission products and transuranic elements. An engineering-scale aluminum electrorefiner has been installed for further testing. Although the laboratory-scale work is promising, significant development problems remain to be resolved before the process can be adapted to engineering practice. Many of the difficulties being encountered with performance of the anode-cathode module are apt to be encountered with aluminum alloy fuels. 18   C. C. McPheeters, ANL-W, presentation to the committee, October 26, 1998, Chicago, IL.