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 85
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report APPENDIX B Meeting Summary Meeting of the Committee on Electrometallurgical Techniques for DOE Spent Fuel Treatment Argonne National Laboratory–West Idaho National Engineering and Environmental Laboratory July 21-22, 1999 JULY 21, 1999—OPEN SESSION—AGENDA 8:15 a.m. Demonstration Status and Plans R.W. Benedict 9:00 a.m. Driver Processing Results R.D. Mariani 9:30 a.m. Cathode Processing and Casting Results B.R. Westphal 10:00 a.m. Break 10:15 a.m. Electrorefiner Throughput Studies J.L. Willit 10:45 a.m. Blanket Processing Results S.R. Sherman 11:15 a.m. Uranium Disposition Options H.F. McFarlane 11:45 a.m. Lunch 12:30 p.m. Ceramic Waste Process and Materials Studies S.McDeavitt 1:00 p.m. Ceramic Waste Demonstration Processing Results K.M. Goff 1:30 p.m. Ceramic Waste Qualification Testing L.R. Morss 2:00 p.m. Ceramic Waste Uranium/Plutonium Behavior Studies S.G. Johnson/L.R. Morss 2:30 p.m. Ceramic Waste Product Consistency Testing T.P. O’Holleran 3:00 p.m. Break 3:15 p.m. Metal Waste Qualification Testing D. Abraham 3:45 p.m. Metal Waste Product Testing D.D. Keiser 4:15 p.m. Metal Waste Release Modeling M.C. Petri 4:45 p.m. Repository Performance Modeling E.E. Morris 5:15 p.m. Adjourn
OCR for page 86
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report SUMMARY OF PRESENTATIONS Gregory Choppin, committee chair, opened the open session with an introduction of the committee members. Robert W. Benedict, ANL, spoke on the EBR-II spent fuel demonstration status and plans. The EBR-II spent fuel treatment flow sheet was reviewed, demonstrating the separation of uranium, and the ceramic and metal waste forms individually. An overview of repeatability results for electrorefining was given. The specific success criterion relating to this issues states, “Freeze process modifications and operating parameters while demonstrating a continuous throughput of 16 kg of driver uranium for three months.” The electrorefiner repeatability demonstration began on November 14, 1998, and ended on January 22, 1999, for a total of 61 working days. The average treatment rate was approximately 24 kg per month. The Mark-V electrorefiner is used to treat blanket fuel. The success criterion relating to blanket fuel treatment requires a throughput rate of 150 kg per month sustained for 1 month. At the time of the meeting (July 21, 1999), consecutive treatment had started with three ports on the electrorefiner running in parallel, with hopes of getting the fourth port running. Significant achievements include the following: the latest run conditions allow 190 to 240 g of uranium per hour as an average production rate, four ports are operational, five blanket assemblies have been treated, and control software allows unattended operation. Operations with irradiated fuel include both the cathode processor and the casting furnace. The cathode processor has treated 40 driver batches, 6 blanket batches, and 8 cladding hull batches. The casting furnace has treated 40 driver batches, 6 blanket batches, and 7 metal waste batches. Significant accomplishments in the treatment process include the following: driver treatment has processed 100 driver assemblies, 8 assemblies were treated in 1 month, 1110 kg of low-enriched uranium were cast, the cathode processor batch size increased from 12 to 19 kg, and the casting furnace batch size increased from 36 to 54 kg. Blanket treatment has processed 5 of 25 blanket assemblies, the Mark-V electrorefiner has run 11 batches of irradiated blankets, the cathode processor has consolidated up to 42 kg of product in a batch, 125 kg of blanket product have been cast, and the blanket element chopper is operational. Significant accomplishments in waste activities include the following. The stainless steel-zirconium alloy continues as the metal waste form. The test matrix for qualification testing has been established. Three of three full batches of irradiated cladding hulls have been cast. Spiked and cold samples castings are complete, and waste qualification testing has started. Glass-bonded sodalite is the ceramic waste form. Initial uranium and plutonium studies are available. Nonradioactive demonstration-scale equipment testing is complete. Equipment has been installed in the Hot Fuel Examination Facility. Laboratory-scale samples containing plutonium for accelerated alpha decay tests have been fabricated. At the time of the meeting (July 21, 1999), 4 of the 10 demonstration-scale cans had been processed. A number of reports have been or will be produced by ANL relating to the demonstration project. Overall demonstration reports include Spent Fuel Demonstration Final Report, Production Operations for the Electrometallurgical Treatment of Sodium-Bonded Spent Nuclear Fuel, Analysis of Spent Fuel Treatment Demonstration Operations, and Uranium Disposition Options. Treatment operation reports can be divided into three groups: overall treatment reports, driver treatment reports, and blanket treatment reports. The overall report will be the Development of Cathode Processor and Casting Furnace Operating Conditions. Treatment operation reports include Process Description for Driver Fuel Treatment Operations and Development of the Electrorefining Process for Driver Fuel. Blanket treatment reports include Process Description for Blanket Treatment Operations and Development of the Electrorefining Process for Blanket Fuel. Waste operation and qualification reports are divided into three groups: overall waste reports, ceramic waste reports, and metal waste reports. The overall reports include Waste Form Qualification Strategy, Waste Form Acceptance Product Specifications, Waste Compliance Plan, and Waste Form Degradation and Repository Performance Modeling. The ceramic waste reports include Ceramic Waste Form Process Qualification Plan and the Ceramic Waste Form Handbook. The metal waste reports include Metal Waste Form Process Qualification Plan and the Metal Waste Form Handbook. The Environmental Impact Statement covers sodium-bonded fuel treatment. The EBR-II spent fuel treatment
OCR for page 87
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report demonstration was limited. It includes 100 EBR-II driver assemblies containing 410 kg of highly enriched uranium and 25 EBR-II blanket assemblies containing 1,200 kg of depleted uranium. Robert D. Mariani, ANL-W, spoke on driver electrorefining results. One hundred driver assemblies have been treated in the Mark-IV electrorefiner within three years. Before the repeatability demonstration, 72 assemblies were used in tests. During the repeatability demonstration, 12 assemblies were used. Twelve more assemblies were electrorefined with repeatability conditions. The remaining four assemblies were used in additional tests. The driver electrorefining test parameters, as detailed in ANL-NT-112, include anodic conditions; electrode configurations; the electrorefining sequence; cell current/cell voltage profiles; the uranium source; electrode, electrolyte, and cadmium agitation; and the Mark-IV electrorefiner temperature. Repeatability conditions for the Mark-IV electrorefiner, as described in ANL-NT-111,1 include an anode voltage of <0.4 V when using fuel dissolution baskets, a 3,200 to 3,450 Amp-hour (A-h) per anode assembly, a dual anode/serial cathode electrode configuration, the electrorefining sequence, stepped current profiles, a 5 rpm electrode speed, a 20 rpm cadmium mixer speed, and a temperature of 500 °C. The dual anode/serial cathode configuration combines two anode assemblies (four driver assemblies) and two cathode assemblies. The electrorefining sequence is direct transport number 1, direct transport number 2, followed by dissolution to the cadmium pool and deposition from the cadmium pool. There is a stepped current profile in direct transport number 1, with an A-h range from 0-480 A-h to 480-1,680 A-h to 2,700-3,500 A-h. A summary of the repeatability demonstration for the Mark-IV electrorefiner shows that fixed electrorefining process conditions were used and that over the 61 working days of the demonstration (November 14, 1998, to January 22, 1999) the average treatment rate was approximately 24 kg per month, including an average electrorefining rate of 0.06 to 0.12 kg/h. In summary, 100 driver assemblies were treated in 3 years, a continuous throughput of 16 kg of driver fuel per month over a 3-month period with fixed process conditions was exceeded (48 kg of driver fuel was treated in 61 working days), and material balances for uranium, plutonium, neptunium, sodium, and fission products were consistent with experimental error. Moreover, the uranium and plutonium balances complied with DOE Nuclear Material Control and Accountability requirements (DOE order 5633.3B). Brian R. Westphal, ANL-W, presented information on cathode processing and casting results. Over the course of the demonstration, the cathode processor treated 40 driver batches, 6 blanket batches, and 8 cladding hull batches. The casting furnace during the demonstration treated 40 driver batches, 6 blanket batches, and 7 metal waste batches. Typical cathode processor/casting batch quantities were as follows: (1) for the driver, dendrites: 18 to 19 kg, salt (cadmium): 3 to 4 kg, DU: ~35 kg, uranium product: ~50 kg, and (2) for the blanket: blanket material: 40 to 41 kg, salt: 6 to 7 kg, uranium product: ~33 kg. Cathode processor/casting operating conditions for the driver and blanket include the following: for the cathode processor, the maximum crucible temperature is 1200 °C, the operating pressure is 0.1 torr, and the salt distillation step takes place over 1 h at 1100 °C. During casting, the maximum crucible temperature is 1300 °C, the operating pressure is at 900 torr until cast, and there is one stir cycle. In summary, the driver demonstration was completed and the success criteria for these steps were met. One hundred driver assemblies were processed, 56 kg of uranium were processed during the repeatability step, condensate was returned during repeatability, operating conditions for the cathode processor/casting were specified, and three metal waste batches were cast. The blanket demonstration was initiated at the time of the meeting (July 21, 1999). James L. Willit, ANL-W, spoke about electrorefiner throughput studies. The Mark-IV throughput goals for the demonstration phase were to electrorefine 150 kg of uranium in 1 month. Both glove box and hot cell tests 1 R.D. Mariani, D. Vaden, B.R. Westphal, D.V. Laug, S.S. Cunningham, S.X. Li., T.A. Johnson, J.R. Krsul, and M.J. Lambregts, Process Description for Driver Fuel Treatment Operations, NT Technical Memorandum No. 111, Argonne National Laboratory, Argonne, IL, 1999.
OCR for page 88
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report demonstrate that this rate can be achieved. For treatment of the inventory of blanket fuel, the goal was to electrorefine 450 kg of uranium per month. The required average production rate has been demonstrated in glove box tests. A process cycle was developed for sustained operation of the Mark-IV anode-cathode module (ACM). This process consists of loading, a stripping step, a washing step, a deposition step, another washing step, and then return to the stripping step, unless the product collector is full. The Mark-V ACM stripping step removes the dense layer of uranium that builds up on the cathode tubes and stalls anode rotation. Reversing polarity electrotransports the dense layer of uranium from the cathode back to the surface of the anode baskets. The stripping step is terminated when all of the dense layer has been removed, as evidenced by a sharp increase in voltage. Periodically, low-current stripping is needed to fully clean the cathode tubes. The washing step dislodges electrodeposited uranium trapped in the electrotransported zone, especially on the anode surfaces. Higher rotation speeds and multiple changes in direction of rotation improve the efficiency of this step. Implementation of this step decreases the frequency of stalls. The deposition step’s purpose is to electrotransport uranium from the fuel in the anode baskets to the cathode surface. The deposition step shows two voltage plateaus. There is a lower voltage plateau due to electrotransport of uranium from the basket surface back to the cathode. There is also a higher voltage plateau due to electrotransport of uranium from clad fuel segments inside the anode baskets. Electrodeposited uranium is scraped off the cathode surface and falls down into the product collector. For throughput measurements in the Mark-V ACM, the key parameter is the average production rate, measured in grams of uranium per hour per ACM. The components of the average production rate include the net ampere hours per kilogram (338 A-h/kg being theoretical), which reflects the efficiency of electrotransport of uranium from the chopped fuel pins to the cathode tubes, and the shorting and U4+/U3+ parasitic reaction decrease in coulombic efficiency and the increase in net A-h/kg. A second component is the stripping A-h/total A-h ratio. Minimizing this value has a huge impact on throughput. Additional components of the average production rate include the frequency and duration of low-current strips and the duration of the washing step. A number of parameters affect the average production rate. For the deposition and stripping steps, these include the current and voltage, the anode rotation speed, and the cutoff voltage (resistance). For the washing step, parameters include the duration of the washing step, the rotation rate during the washing step, and changes in the direction of rotation. Other parameters include the cutoff voltage (resistance) and frequency of the low-current strips, and holdup and shorting. Holdup and shorting always decrease throughput by decreasing coulombic efficiency (>338 A-h/kg uranium) and/or by increasing the stripping ampere hours. Product holdup has been found on the surfaces of the anode baskets, on isolated hard patches on the cathode tube, and in the region of the cathode supports. Cathode holdup must be removed by periodic low-current stripping. Anode holdup must be removed physically by hand or by aggressive washing. Optimization of deposition parameters can be achieved in a number of ways. Higher deposition current favors higher throughput, where 600 A is the practical upper limit. Exceeding 200 A-h as the length of the deposition step results in a greater frequency of stalls. A higher voltage cutoff allows longer operation at higher currents. The present cutoff is 0.75 V, which avoids appreciable corrosion of the anode baskets. Lower rotation speed favors higher throughput but can increase the frequency of stalls. A range of 20 to 60 rpm has been examined, and additional optimization steps are in progress. Optimization of stripping parameters can also be achieved in a number of ways. Lower rotation speed favors higher throughput but can increase the frequency of stalls. A range of 20 to 60 rpm has been examined, and additional optimization tests are in progress. A higher stripping current favors higher throughput. This decreases the length of the stripping cycle. A current of 600 A is the practical upper limit. Optimization of washing step parameters was also accomplished in a number of ways. A shorter washing step increases the average production rate but can increase the frequency of stalls. The rate typically ranges from 1 to 6 min. A higher rotation rate decreases the frequency of stalls. The optimal rate was found to be 60 rpm. Multiple
OCR for page 89
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report rotational direction changes were found to decrease the frequency of stalls. A forward-reverse-forward sequence was found to work better than a constant direction of rotation. Other parameters were looked at as well. For product collector harvesting, removal by an inverted bake-out is slow but leaves little residue in the product collector. Removal by grinding is fast but leaves residues in the product collector. A higher concentration of uranium in the salt was found to increase the limiting current density. Lower concentrations favor a finer deposit morphology. A number of comparisons were made between glove box and hot cell Mark-V ACM tests. Changes in operation parameters in general have the same qualitative effect in hot cell and glove box tests. Hot cell ACM tests show a somewhat lower average production rate (g/h/ACM). This is due to a higher stripping A-h/total A-h ratio in the hot cell tests. Holdup, shorting, and differences in adhesion of material to the cathode surface are the most likely sources for this difference. A strategy was adopted for increased throughput. The deposition current was increased from 500 to 600 A. To decrease holdup and shorting, the stripping A-h/total A-h ratio was decreased, as was the A-h/kg rate. Finally, to improve handling operations, the uranium was ground out of the product collection reservoir rather than baked out. Steven Sherman, ANL-W, presented information on blanket processing results in the hot cell. The setup of the Mark-V electrorefiner for blanket fuel is as follows. The fuel is depleted uranium at a rate of ~0.2 percent burnup. A 25-in. electrorefiner design of similar configuration has also been developed. The Mark-V contains four ports, each of which measures 10 in. and each of which operates independently. Each port has its own power supply, with a current of up to 600 A per port. The experimental goals of the Mark-V electrorefiner were to maximize the production rate, to achieve simultaneous operation of the ports, to achieve a production rate of 150 kg of uranium per month, to realize reliable operation, and to perform unattended operation while passing current. Significant achievements for the Mark-V electrorefiner were as follows. At the time of the meeting (July 21, 1999), the latest run conditions allowed 190 to 240 g of uranium/h/ACM as the average production rate, with 60 percent equipment utilization. There were four operational ports. Simultaneous operation of three ports was achieved, and routine operation of two ports was possible. More than five blanket assemblies were treated. Control software allowed unattended operation. Two product collector harvesting methods were developed: a bake-out oven, which gave a gravity-assisted product dump at 500 (C, and a product collector harvesting tool, a rotating multibladed tool for grinding out product at room temperature. For Mark-V electrorefiner operation, the anode basket was loaded with 9 or 10 assemblies. Each assembly has 19 elements, and there are two anode baskets per assembly. Three product collectors are needed per anode basket. The operation is cyclic—each cycle is composed of deposition, stripping, and wash steps. The deposition step (per cycle) consists of a controlled current, initially 500 A for 200 A-h or until the 0.75-V cutoff is reached. If the voltage cutoff is reached before 200 A-h, the current is reduced and the step is continued. The rotation rate is 20 rpm, and the direction is forward. For the stripping step (per cycle), the current is 600 A, 300 A, and 165 A, each operating to a –0.75-V cutoff. The rotation rate is 10 rpm, and the direction is forward. For the wash step (there are two per cycle) the rotation rate is 60 rpm for 1 to 2 minutes in the forward direction. The end point of a batch is reached when the voltage is 0.75 at 80 A during deposition. Materials/coatings tests were performed for intermediate containers. Alternative materials looked at included quartz, Pyrex, and aluminum oxide. A number of coatings on mild or stainless steel crucibles were also tested. Plasma-coated materials included zirconium oxide, zirconium oxide with a nickel-aluminum bond, tantalum, and tungsten. Titanium nitride was tested using chemical vapor deposition. Finally, boron nitride was tested by painting. For each of these materials, the Mark-V product was placed into crucibles and heated to 500 °C for 2 to 3 hours. The crucibles were then removed from the furnace and cooled, and turned upside down to test for product adhesion. In all cases, the product failed to release at room temperature but released at 500 °C. Test results and metallography examinations suggest no chemical interaction between uranium metal and the crucibles. Tests with titanium nitride (conductive coating) were in progress at the time of the meeting (July 21, 1999).
OCR for page 90
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report At the time of the meeting (July 21, 1999), future work included the completion of the demonstration of 150 kg/month using established conditions. Run cycles were adjusted as needed to increase process reliability and production rates. Future equipment modifications to the ACM and support equipment were to be tried to increase process reliability and production rates. Testing of alternative materials and coatings and studies of the electrorefining characteristics of uranium were also to be continued. Harold McFarlane, ANL, spoke about uranium disposition options. Disposition of uranium is constrained by the quantities and characteristics of uranium produced by treatment of sodium-bonded spent fuel. Options for disposition of uranium are constrained and defined by overarching DOE plans. These are discussed in the DOE publications Disposition of Surplus Highly Enriched Uranium, Strategies for the Long-Term Management and Use of Depleted Uranium Hexafluoride, and Storage and Disposition of Weapons-usable Fissile Materials. Treatment of the remaining 3.2 MTHM of EBR-II driver fuel results in 10.7 MT of low-enriched uranium (19 percent 235U). Fifty-five MT of depleted uranium (DU) would be recovered from treatment by EMT of the remaining blanket fuel in the DOE sodium-bonded SNF inventory. Approximately 1.6 MT of uranium ends up in the waste forms (0.89 MT in the metal waste form and 0.67 MT in the ceramic waste form). Available options for uranium disposition include commercial use, continued storage, disposition as low-level waste (LLW), disposal as transuranic (TRU) waste, or other DOE use. Excess highly enriched uranium (HEU), according to the Record of Decision for the Disposition of Highly Enriched Uranium Final Environmental Impact Statement (August 1996), will be dispositioned by one of two means: either conversion to low-enriched uranium light water reactor (LWR) fuel or disposal as LLW after blending down to 0.9 percent 235U. The HEU in DOE spent nuclear fuel (SNF) is considered part of the excess HEU inventory. To process HEU into LWR fuel for disposal as commercial reactor fuel, the blended product must be able to meet LWR fuel specifications. Standard fuel specifications are defined by ASTM. Specifications for off-spec fuel have also been developed for reactors operated by the Tennessee Valley Authority. For disposal as LLW, the HEU needs to be blended to 0.9 percent 235U. The key specification for disposal as LLW is that the TRU activity must be less than 100 nCi/g. After blending to 0.9 percent 235U, the LEU product would contain approximately 40 nCi/gram of TRU activity. Some fraction of the U.S. excess plutonium will be disposed of by can-in-canister. Current plans are to add plutonium-bearing LEU to the ceramic. Downblending will take place to 5 percent enrichment. Ceramic pucks will be fabricated in a remote glove box line. The feasibility of this option for the disposition of uranium is dependent on reducing the radiation levels of the metal ingots. At present LEU cannot be directly converted to LWR fuel. Options are being explored to reduce key contaminants. Downblending processes are being considered that would improve the quality of the final product. Continued storage is feasible. More than 250 MTHM is being stored at ANL-W. Disposal as LLW is also feasible. Integration into the plutonium disposition program may be feasible. Consideration is also being given to disposal of depleted uranium as low-level or transuranic waste. The TRU concentration in the DU ingots may be >100 nCi/g. ANL will determine whether it can lower this concentration to allow its classification as low-level waste. From a review of the waste acceptance criteria for the Waste Isolation Pilot Plant (WIPP), disposal as TRU waste appears to be feasible. According to the Final Programmatic Environmental Impact Statement for Alternative Strategies for the Long-Term Management and Use of Depleted Uranium Hexafluoride (DOE/EIS-0269), the preferred alternative for management of DU is continued storage after conversion to oxide or metal. As part of this program, DOE will also support the development of markets for the oxide and metal DU products. The DU produced from blanket treatment is a small fraction of that in the DOE complex (about 55 MT of 600,000 MT). Storage of 55 MT of recovered DU would be consistent with the programmatic Environmental Impact Statement. The material will already be a metal, which is the form most desirable for commercial application. Depleted uranium disposition as an LLW is questionable without additional treatment. Disposition as TRU
OCR for page 91
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report metal waste may be feasible, with conversion to oxide as a fallback position. Storage and commercial uses are feasible. Options for uranium disposition are constrained by DOE programmatic decisions based on environmental impact analyses. Within those constraints, there are viable options for both the LEU and DU from spent fuel treatment. Sean McDeavitt, ANL, presented information on ceramic waste process and material studies. Salt-borne waste in the molten lithium chloride-potassium chloride electrolyte include transuranic isotopes, active fission products (e.g., cesium, strontium, barium, and iodine), and rare earth fission products (e.g., lanthium, cerium, and neodymium). Salt-borne wastes in the ceramic waste form include salt incorporated in the aluminosilicate (zeolite 4A) lattice. Zeolite transforms to sodalite during processing. The waste form is a glass-bonded sodalite composite. Long-term thermal exposure of the ceramic waste form (CWF) was performed to assess the durability of the waste form. Five hot isostatic pressing (HIP) cans with reference ceramic waste composition were tested. Post-test examination was performed by leach testing, X-ray diffraction, microscopy, and bulk density determination. In summary, there were no major effects at 500 °C. When looking at uranium and plutonium, the principal question is whether reactive salt components will react with the zeolite structure. There is concern about uranium chloride and plutonium chloride (UCl3 and PuCl3), and the rare earth chlorides. Understanding continues to improve. UCl3 is known to react with water in the zeolite. PuCl3 reaction with water is implied. Dried zeolite contains more than enough water to react with all the UCl3 and PuCl3 in the spent electrorefiner salt. Uranium does not appear to react with the lattice. Plutonium behavior is still being investigated. Test methods for uranium include differential scanning calorimetry (DSC), which is the standard analysis tool. Post-test X-ray diffraction (XRD) is used to identify reaction products. DSC with simultaneous X-ray diffraction is a custom apparatus used at the Advanced Photon Source. X-ray diffraction is performed in situ during the DSC experiment. Evolved gas analysis (EGA) is also used. In this test the sample is heated in a vacuum chamber connected to a mass spectrometer. The evolved gas species are then identified. In a high-uranium-content, heavy-salt mixture (i.e., lithium chloride-potassium chloride-UCl3 with ~52 wt % uranium and 30 weight percent salt + 70 wt % zeolite), enough UCl3 is present to consume the water in dried zeolite and still react with the zeolite lattice. DSC revealed the nature of the reactions. A small exothermic reaction begins at ~180 °C, followed by a strong exothermic reaction at ~350 °C. The reaction product contains uranium dioxide (UO2). Solid phases in the reaction were tracked by DSC/XRD. It was found that the zeolite lattice contraction begins at ~180 °C and that UO2 formation begins at ~350 °C. Above 350 °C, the zeolite peak intensity decreases. Similar changes are also observed for lithium chloride-potassium chloride with and without UCl3. Lattice reaction products are notably absent. EGA confirmed the reaction with water by detecting hydrogen chloride and hydrogen vapor evolution at ~350 °C. These data indicate that uranium lattice reaction is not a problem. At high uranium concentrations, the zeolite lattice is not attacked by the salt. This implies that at low concentrations, the salt will not attack the zeolite lattice during waste processing. In addition, the salt cannot get past the water. These data also indicate that reaction of UCl3 with water is a confirmed fact, and that the waste form will contain oxidized uranium particles. The impact of these particles on repository performance is being addressed through qualification testing. There are a number of test methods to analyze plutonium in the CWF. X-ray diffraction has been used for blended powders contacted with plutonium-bearing salt. X-ray diffraction has also been used for hot, uniaxially pressed (HUP) ceramic waste form samples in powder form. X-ray absorption fine-structure (XAFS) spectroscopy has been used for HUP waste form samples in bulk and powder form. Electron microscopy has been used for 238Pu-bearing HUP ceramic waste form samples. A review of the data on plutonium in the CWF gave the following information. X-ray diffraction revealed that the zeolite lattice is unaffected during small-scale blending. It has also found that plutonium dioxide (PuO2) is formed during small-scale blending. XAFS revealed information about the local plutonium environment. Plutonium was found to be present as PuO2, not as PuCl3. The PuO2 particle size was estimated to be at least 20 to 40 Å. This
OCR for page 92
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report indicates that the PuO2 could not be inside the 12 to 16 Å zeolite cage. Electron microscopy revealed direct evidence of PuO2. PuO2 particles were observed near the sodalite-glass boundaries. Fine particles were observed by tunneling electron microscopy outside the sodalite. Uranium and plutonium matrix samples were prepared to address questions about the CWF. The waste forms were evaluated using representative processing methods and compositions. Samples were prepared with reference fission product loading. The samples were subjected to heating blending and HIP. Two samples were prepared that contained plutonium- and uranium- bearing salts (1.5 mole percent PuCl3 and 0.5 mole percent UCl3, and 0.5 mole percent PuCl3 and 1.5 mole percent UCl3). Two samples were prepared with water in the zeolite (0.12 weight percent water and 3.5 weight percent water). Finally, four small HIP cans were prepared for each salt/zeolite combination. Analysis of the blended materials and salt precursors indicated low free chloride content after blending (0.03 to 0.3 percent). Also, low plutonium release was observed in the free chloride test solution (parts per thousand levels). X-ray diffraction of the blended material was under way at the time of the meeting (July 21, 1999). X-ray diffraction of the salt precursors was pending at the time of the meeting (July 21, 1999). Electron microscopy details were deferred to a later presentation. A post-HIP test matrix was under way. The tests will evaluate the behavior of actinide-bearing phases to assess their impact on repository performance. Pressureless consolidation has been looked at as a potential alternative to HIP. Preparation steps between pressureless consolidation and HIP are nearly identical. Complex HIP steps, including HIP can welding, powder compaction, evacuation, sealing, and welding, are eliminated. In addition, no high-pressure equipment is used in the hot cell with pressureless consolidation. The process is simple, fast, and compact. The blended powder is poured into settlers and the top surface is leveled. Loaded settlers are passed through a tunnel kiln at 850 °C for ~4 hours. The waste form is then removed and the settler is reused. ANL is approaching the decision point for pressureless consolidation. The product is similar to the HIP product, without the can. The product has a uniform external appearance, and corrosion tests show no penalty in waste form leach behavior. The process also appears to be very easy to scale. Samples up to 1 kg have been prepared, and at the time of the meeting (July 21, 1999), lab scale-up to 5 kg was under way. The materials used are all available commercially. The graphite settler material is reusable. Finally, for the same cycle time, throughput may be more than 15 times higher than for HIP. K.M. Goff, ANL-E, spoke on ceramic waste demonstration processing results. A number of nonirradiated CWF tests were summarized. The drying cycle for a 30 to 40 kg batch has been developed and implemented at an outside vendor. Thirty batches of material have been processed. Material from salt processing (crusher-mill/classifier) has been well characterized to obtain appropriately sized material for mixing. Twenty-two V-mixer experiments were performed, including one 80 kg test. The ability to produce salt-loaded zeolite with low free chloride has been demonstrated repeatedly. Ninety-four HIP cans have been produced, and the reliability of the HIP, the HIP can, the bake out/evacuation system, and the weld have all been demonstrated. It has been verified that the glass-zeolite ratio is maintained throughout all process steps. The durability of the waste form containing surrogates has also been well demonstrated. Operating parameters for zeolite drying were as follows. The operating temperature was 550 °C, with a heating rate of ≤3 °C per minute. Time at temperature was seven hours, the degree of vacuum was ≤100 torr, the moisture content was 0.3 (≤1 weight percent), and rehydration was ≥18 weight percent. For the salt crusher and mill/classifier, the mill speed was 1,500 rpm, the classifier speed was 350 rpm, the feed frequency was 7.0 Hz, and secondary gas readings were ΔP 0.8 in. For the V-mixer, the operating temperature was 505 °C, time to temperature was ~2 hours (with no limits for heater ramp rates), and the time at temperature was 15 hours. The rotation rate was 17 rpm, free chloride measured ≤ 0.5 weight percent, and salt content was 3.8 Cl ions per unit cell. For demonstration runs using HIP, loading was as follows. Twenty-one bellows cans were loaded with 1,600 g of material. ANSTO cans were loaded with 950 g of material. The witness tubes were loaded with 24 g of material.The HIP cans themselves were made of 304L stainless steel. The compactor operating time was between 0 and 10 minutes.
OCR for page 93
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report Parameters for HIP evacuation bake-out and crimp and weld were as follows. For evacuation/bake-out the vacuum was at <100 mtorr, furnace time at temperature was ≥6 h, and the furnace temperature was 500 °C. For crimp and weld, the magnitude of the crimp was 12 tons, and the type of weld was tungsten inert gas (TIG). Eleven nonirradiated HIP cans were processed through the HIP in the Hot Fuel Examination Facility (HFEF). At the time of the meeting (July 21, 1999), processing of irradiated materials was ongoing. Operating parameters for the hot isostatic press are as follows: the hold temperature is 850 °C, the maximum pressure is 14,500 psi, the temperature ramp is 5 °C per minute, the pressurization rate is 240 psi/min on average, the depressurization rate is 85 psi/min on average, the hold time is 60 min, the shelf time is 45 min, and the shelf temperature is 750 °C. Demonstration-scale irradiated samples were produced. Irradiated salt was removed from the electrorefiner and transferred to the HFEF. Mill/classifier tests were performed to confirm the behavior of the electrorefiner salt. A batch of 100-driver salt was successfully processed through the V-mixer. At the time of the meeting (July 21, 1999), irradiated samples and witness tubes were being processed through the HIP. In the HIP scale-up tests, no density variations were detected from 27 samples taken throughout the can. X-ray diffraction was performed on 13 samples, and the diffraction patterns were the same as seen in the demonstration products. Initial performance results were the same as those obtained at the demonstration scale. An 18-in. can was processed in June 1999. This diameter is essentially production scale. Hold time was approximately 10 h. The can held 60 kg of material. An 18-in. ANSTO can was also processed. This can held almost 100 kg of material. Samples were produced for accelerated alpha decay studies. Two batches of material containing 238Pu were produced, and characterization is ongoing. The salt contained 15 mole percent plutonium and surrogates for fission products. The HUP sample contains more than 2 weight percent plutonium. Results of the alpha decay studies show no significant changes since last reporting. The plutonium normalized release rate was 1 × 10-4 g/(m2 day). There were no changes in density. There were also no changes in the unit cell volume of sodalite. L.R. Morss, ANL-E, spoke about CWF qualification testing. The CWF is ~75 percent sodalite, ~25 percent glass, with small amounts of UO2, PuO2, oxychlorides, halite, and nepheline. Qualification needs that are addressed in the testing program cover many areas. Repository performance assessment needs require the characterization of matrix corrosion behavior and radionuclide release, and also require a mechanistic model with model parameters. Waste specification needs require the identification of phases containing radionuclides, and methodology must be provided to monitor product consistency. For process qualification, a methodology must be provided and a database established for process operating parameters. Scoping tests identify corrosion modes and provide model parameter values. Solution exchange tests demonstrated that the primary release mechanism is matrix dissolution, not leaching of occluded material. Product consistency tests characterize the dissolution behavior under concentrated solution conditions. The MCC-1 static leach tests characterize dissolution behavior under dilute solution conditions and provide model parameter values. The pH buffer tests measure pH dependence of the dissolution rate. Accessible free salt measurements measure the amount of soluble salt (mostly sodium chloride). The 7-day product consistency test is specified in the waste acceptance systems requirements document (WASRD) for defense high-level waste (DHLW) glasses.2 Crushed material in demineralized water at 90 °C is used. The high surface area of material/volume of solution (S/V) ratio is representative of anticipated Yucca Mountain conditions. ANL has shown that the performances of the reference CWF and DHLW glasses are similar in this test. Longer-term tests are used to study corrosion behavior. The normalized elemental mass loss is predicted by the following equation: 2 American Society for Testing and Materials, Standard Test Methods for Determining Chemical Durability of Nuclear Waste Glasses: The Product Consistency Test (PCT) Standard C1285-94, Annual Book of ASTM Standards, Vol. 12.01, West Conshohocken, PA.
OCR for page 94
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report where C(i) is the measured solution concentration, S/V is the surface area of material divided by the volume of solution, and f(i) is the mass fraction of the element in the material. This equation allows direct comparison of the release of different elements from different materials and at different S/V ratios. The MCC-1 test was developed in the early 1980s to evaluate the relative chemical durabilities of simulated and radioactive monolithic waste forms at low S/V ratios. Monolithic samples in demineralized water at 90 °C are used. The low S/V ratio is not representative of anticipated Yucca Mountain conditions. The low S/V ratio allows the MCC-1 test to be useful for determining the initial (forward) rate and is thus useful for providing modeling parameters. MCC-1 tests found that the response of reference CWF was similar to that of DHLW glasses. In the scoping tests, the dissolution behaviors of sodalite, the glass binder, free salts, and oxides/oxychlorides were determined. The dissolution of free salts becomes controlled by the dissolution of sodalite and glass binder. The dissolution behaviors of sodalite and glass binder phases can be described with the glass dissolution model in the total systems performance-viability assessment. Model parameter values for DHLW glass provide the upper bound to parameter values for the sodalite and glass binder phases. For the corrosion model, a rate expression was developed for aluminosilicate minerals:3 The rate depends on orthosilicic acid concentration, but not on time. The value for kf depends on temperature and pH. The value for [H4SiO4]sat depends on temperature. The equation has been modified for application to DHLW glasses:4 The rate expression is: where R is the alteration rate of the waste form (g/(m2d), k0 is the intrinsic rate (g/(m2d), η is pH dependence, Ea is temperature dependence (kJ/mole), [H4SiO4] is the concentration of H4SiO4 in solution (mg/l), [H4SiO4]sat is silica saturation concentration (mg/l), and klong is the residual long-term rate (empirical term). Tests showed that [H4SiO4]sat is lower for sodalite, the glass binder, and the CWF than for HLW glasses. Implementation of the CWF model in the repository integration program uses radionuclide inventories and the specific surface area of the CWF. The HLW glass rate parameters are currently used by ANL to model CWF behavior. Experimentally determined rate data for the CWF will be included in future models. Scoping tests have identified the primary corrosion mode as matrix dissolution. They have also characterized short-term and long-term corrosion. The scoping tests have also shown that the CWF corrosion follows the model for aluminosilicate minerals and HLW glasses. For the model parameters, pH dependence of the forward dissolution rate is now being determined for sodalite and glass. Temperature dependence will be determined. The saturation concentration of H4SiO4 is being determined. A simplified model has been implemented in the Repository Integration Program (RIP) program. At the time of the meeting (July 21, 1999), work in progress included tests to study release behavior of PuO2from plutonium-loaded CWF. Tests to determine the temperature coefficient of the forward rate were under way. Tests to determine if the aluminum concentration in solution affects the dissolution rate had begun. Tests were also initiated under repository-relevant conditions (drip tests, vapor hydration tests). 3 P. Aagaard and H.C. Helgeson, Am. J. Sci. 282, 1982, pp. 237-285. 4 B. Grambow, Mat. Res. Soc. Symp. Proc. 44, 1985, pp. 15-24.
OCR for page 95
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report S.G. Johnson, ANL-W, spoke about ceramic waste uranium/plutonium behavior studies. The approach is to prepare the CWF with uranium, plutonium, and fission products by blending and HIP. Two uranium/plutonium ratios are used, as are two water contents in the zeolite. The microstructures of these materials are then analyzed, corrosion tests are carried out, and the disposition of uranium and plutonium released from the CWF is determined. The main objective of this work is to determine uranium/plutonium release as the CWF matrix corrodes. This includes determining uranium and plutonium released into solution, uranium- and plutonium-containing colloids in solution, if any, and alteration phases containing uranium and/or plutonium. The purpose of the corrosion test matrix is to achieve accelerated corrosion conditions. Here also the goal is to determine release of uranium and plutonium into solution. Also, altered solids remaining with the CWF are characterized. Plutonium colloids are also characterized in solution by sequential filtration, dynamic light scattering, and transmission electron microscopy (TEM). A product consistency test (PCT) at 120 °C in demineralized water at S/V ratios of 1,000 to 2,000 m–1 was also performed. Corrosion tests were also carried out at reference conditions. This was to verify that accelerated conditions do not change the release mechanism. PCT at 90 °C was performed in demineralized water at S/V 2,000 m–1. The approach to repository relevant conditions was to perform drip tests with cores from the CWF. Physical characterization of the CWF entailed determination of the coordination and distribution of uranium and plutonium. This included structural identification of uranium/plutonium-containing phases, location of uranium/plutonium-containing phases in the waste form, and distribution of phases within HIP cans (radially and axially). Phases were also identified in the precursors (zeolite and salt-loaded zeolite), glass-bonded sodalite (CWF), solid reaction products after corrosion tests, and colloids present in solution after corrosion tests. Physical characterization tests looked at plutonium before, during, and after the corrosion tests. Before the corrosion tests, plutonium was looked at in the CWF phases. This was done by X-ray powder diffractions, X-ray absorption fine structure spectroscopy (XAFS), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). During the corrosion tests, plutonium was released into solution. This was observed by sequential filtration to estimate the size and composition of colloidal particles. Dynamic light scattering determined the size distribution of colloidal particles. TEM looked at colloidal particles wicked onto the grid. Plutonium was also looked at in the alteration products of corrosion tests. SEM and TEM were used on the surfaces of reacted solids after the corrosion tests. Initial observations from the analyses are as follows. Plutonium and uranium are found in inclusions with rare earths in the form of solid solutions of oxides, silicates, or a mixture of these two. The plutonium/uranium oxide phase is small in size, typically ~10 nm, while the silicate phase is larger, typically 50 to 1,000 nm. in size. Most actinide inclusions are in the glassy regions near sodalite grains. More actinide inclusions are present in the sodalite granules in the waste forms produced with the dry zeolite than in those produced with the wet zeolite. The uranium/plutonium ratio does not affect the overall microstructure of the waste form. SEM analysis observations include the following. Plutonium, uranium, and rare earths are present as oxide solid solutions and silicate solid solutions. Actinide-bearing crystals accumulate mostly in the glassy regions. Small amounts are observed in the sodalite regions (more in the dry samples than in the wet samples). The uranium/plutonium ratio does not affect the microstructure of the ceramic waste form. TEM analysis showed that uranium, plutonium, and rare earths are present in silicate phases (50 to 1,000 nm size). In addition, (uranium, plutonium, rare earth) oxides are present in solid solution (~10 nm in size). XRD observations show that the XRD pattern is similar to the reference CWF pattern. The plutonium/uranium oxide phase behaves as a solid solution. No significant variation was observed in the sodalite crystalline pattern between samples with low and high moisture content. The minor halite phase reflection is roughly twice as intense in high moisture samples as in low moisture samples. Tests and analyses to study the release of uranium and plutonium from the CWF as colloids and dissolved species had been initiated and were in progress at the time of the meeting (July 21, 1999). The morphology and chemical composition of the uranium/plutonium containing phases were also being determined at the time of the meeting.
OCR for page 96
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report Thomas P. O’Holleran, ANL-W, spoke on ceramic waste product consistency testing. The goal of product consistency testing of the CWF is to ensure that the waste form production process is well controlled. A well-controlled process will create waste form products that meet ANL’s waste acceptance product specifications and can be subsequently disposed in a geologic repository. A suite of tests are to be applied to make sure that the waste form is consistent from run to run. A number of tests and methods were applied to the ceramic waste form for product consistency. These included density, X-ray diffraction, microscopy, immersion tests, and toxic characteristic leaching procedure (TCLP). Density was examined because the degree of consolidation in the HIP cans is approximately twice as great in the pre-HIP stage as in the post-HIP stage. This consolidation was important and served to re-check that the HIP went through the appropriate temperature-pressure (T-P) cycle. X-ray diffraction gave XRD patterns that revealed whether the salt-loaded zeolite 4A had completely converted to sodalite during the processing in the HIP. Conversion to sodalite was important to waste form durability and double-checked that the HIP had gone through the appropriate T-P cycle. X-ray also served as a check for the formation of any anomalous phases during production that could alter the release characteristics or other important properties. Microstructure was an important indicator of the interaction of the feed materials with the environment and each other during the processing and the appropriate T-P cycle in the HIP. Scanning electron microscopy (SEM) gave information on grain size, relative porosity, and semiquantitative elemental composition of various phases. Transmission electron microscopy (TEM) gave information similar to SEM but on a much smaller scale. In addition, it gave electron diffraction for the determination of the crystalline structure of phases. Immersion tests were used to assess the consistency of the product via the interaction with water at 90 °C. The release of pertinent elements from all significant phases present in the waste form were monitored. These included matrix elements silicon, aluminum, and boron; alkali/alkaline earth elements lithium, sodium, potassium, rubidium, cesium, strontium, and barium; chloride and iodide anions; and the actinides uranium, neptunium, and plutonium. The toxicity characteristic leaching procedure (TCLP) was used because Yucca Mountain will not be licensed as a RCRA facility. The TCLP test to assess the potential escape of hazardous material from the sample. The hazardous character of the waste must be determined. The tests/methods were put in place to monitor the CWF product to ensure its consistency, demonstrating a well-controlled process. These tests/methods are sensitive to density changes, phase composition, microstructure, release rate of important elements and possible release of hazardous constituents. Daniel P. Abraham, ANL-E, spoke about metal waste qualification testing. The metal waste form test plan is based on an ASTM C 1174. At the time of the meeting (July 21, 1999), 99 percent of phase 1 testing (for the demonstration) was complete; the results are in the Metal Waste Form Handbook.5 Phase 2 testing, which extends beyond the demonstration, had just begun at the time of the meeting (July 21, 1999). The focus was on experiments. This was to generate input for the total systems performance assessment (TSPA) analysis. These experiments are to specify the composition range for process qualification. The TSPA analysis requires input on corrosion behavior. This includes data feeds into corrosion models. These models will be incorporated into RIP performance assessment software. The analysis also requires input on physical properties and phase stability. The process qualification input requires information on composition-property relationships. A range of compositions was studied: 0-20 weight percent zirconium, 0-5 weight percent noble metals (with up to 2 weight percent technetium), and 0-11 weight percent actinides (uranium, plutonium, and neptunium). Alloy properties considered were microstructure, mechanical and thermophysical properties, and corrosion behavior. 5 D.P. Abraham, S.M. McDeavitt, D.D. Keiser, Jr., S.G. Johnson, M.L. Adamic, S.A. Barker, T.D. DiSanto, S.M. Frank, J.R. Krsul, M. Noy, J.W. Richardson, Jr., and B.R. Westphal, Metal Waste Form Handbook, NT Technical Memorandum No. 121, Argonne National Laboratory, Argonne, IL, 1999.
OCR for page 97
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report Microstructural studies were performed on the MWF to identify phases and to study noble metal and actinide distribution. There were 102 tests performed, including SEM, TEM, and neutron diffraction. All tests had been completed at the time of the meeting (July 21, 1999). No noble metal-rich precipitates were observed. Actinides are present only in the intermetallic phases. It was also found that the as-cast microstructure is stable up to ~1100 °C. A number of tests were performed on the physical properties of the MWF. For mechanical properties, tensile, compression, and impact tests were performed on the MWF. There were 45 tests using specimens from eight ingots. For thermophysical properties, density, thermal conductivity, specific heat, and the thermal expansion coefficient of the MWF were all determined. All tests had been completed. Immersion tests provide information on the selective leaching of elements. Static immersion tests (90 °C and 200 °C, up to 1 year) were performed on 176 samples. All are complete. Forty-two pulsed-flow immersion tests (90 °C) were performed. At the time of the meeting (July 21, 1999), all were in progress. Results show good fission product and actinide retention for the MWF alloys. Immersion tests at 90 °C show minor surface tarnish and negligible weight change. Fission product release at one year for technetium was 0.05 g/m2. The maximum rate was ~ 6 × 10–4 g/m2 per day. Losses of palladium, ruthenium, rhodium, and niobium were 0.01 g/m2. Other elemental releases after one year include 0.2 g/m2 for iron, chromium, nickel, manganese, and molybdenum. For uranium, the release was 0.2 g/m2. The maximum rate for uranium was ~ 5 × 10–3 g/m2 per day. Electrochemical tests provided a relative measure of the corrosion rates for various compositions tested. Three hundred and sixty tests were performed, and all were completed. Vapor hydration tests provide information on the corrosion behavior in superheated steam (200 °C). There were 44 tests. At the time of the meeting (July 21, 1999), 36 were complete. Electrochemical corrosion testing parameters included using samples containing a range of zirconium, fission product, and actinide contents. The pH of the solution was varied (pH = 2, 4, 9, and 10). MWF corrosion rates were similar for all alloys (even those with uranium and technetium). Rates were comparable for those for SS316 and Alloy C22 and were two orders of magnitude smaller than for mild steel. Similar results were seen for tests in pH = 4, 9, and 10 solutions. Rates for pH = 2 were higher. Vapor hydration corrosion tests were performed in steam at 200 °C and 100 percent humidity. Test periods were for up to six months. The oxide layer thickness for the MWF samples were ≤1 μm for both 56- and 182- day tests. The adherent oxide layer appears to retard MWF corrosion. TEM and AES studies were performed to identify oxide(s) and to determine the corrosion mechanism. For galvanic corrosion, electrochemical tests provide information on the galvanic interaction between waste forms and candidate container materials. The galvanic current is measured between the metallic samples and Alloy C22. Samples are short-circuited through a computer-controlled potentiostat, which acts as a zero-resistance ammeter. There were 10 tests, and all were complete. The steady-state current value for mild steel in J-13 (pH = 9) was 4 μA. At pH = 2, the value was 40 μA. For SS-15Zr in J-13 and pH = 2, the value was <1 μA. Galvanic interaction between the MWF and C22 was found to be insignificant. At the time of the meeting (July 21, 1999), the MWF test plan was 99 percent complete. Alloy microstructures are well understood. The alloys show favorable mechanical and thermophysical properties and good corrosion resistance. Dennis D. Keiser, Jr., ANL-E, spoke about metal waste product testing. Three primary issues surround the MWF. Have appropriate casting conditions been determined? What characterization techniques are employed for the MWF product? What are typical results from these techniques? MWF casting results in the fuel conditioning facility (FCF) are as follows. Three MWF ingots in the FCF that can accommodate an amount of cladding hulls corresponding to that of two driver fuel assemblies have been cast. A fourth MWF ingot with blanket cladding hulls substituted for driver hulls has been cast. The bulk of EBR-II material left to be processed will consist of blanket elements. The conclusion is that appropriate casting conditions have been determined and implemented for both driver and blanket hulls. For characterization of irradiated MWF ingots, the sampling strategy involved core drilling, where samples are representative of the final product, and many were obtained to establish the homogeneity of the final product.
OCR for page 98
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report Injection casting was used to obtain a sample to correlate bulk chemistry results with those obtained from core drilling. Analysis methods included microscopy, bulk chemistry, immersion tests, and toxicity characteristic leaching procedure (TCLP). Observations from chemical analyses show that elemental analyses of core drilled samples compares well with those for injection-cast samples. The variability of the chemical analysis associated with the use of core-drilled samples has been established. The noble metal fission products technetium, ruthenium, and palladium are present in the MWF at tenths of a weight percent. Observations from microstructural analyses show the expected amounts of alloy constituents present in the MWF alloys (i.e., stainless steel components, zirconium, uranium, and noble metal fission products). The constituents are partitioned between typical alloy phases. The major phases are ferrite and Zr(Fe, Cr, Ni)2+x. The minor phases are austenite, Zr6Fe23, and uranium phases containing technetium and selenium. Noble metals and uranium favor specific phases. Technetium favors iron solid solution phases; ruthenium, palladium, and uranium favor intermetallic phases. The conclusion is that the results show good agreement with those for surrogate MWF alloys. The purpose of immersion testing is to show the correlation between actual MWF samples and doped MWF samples. CFMW06 and 07 monolithic samples were tested in the following manner: core-drilled samples 0.8 cm in diameter × 1.3 to 2.3 cm thick were employed. The samples were cleaned ultrasonically several times in 200-proof absolute ethanol. Deionized water and J-13 well water were used as test solutions. The surface area/volume ratio equaled 50 m–1. Stainless steel containment vessels were used as blanks. The vessels were heated to 90 °C for 2 weeks. Leachates were analyzed and acid stripping of the vessels was performed and analyzed. Immersion testing results for hot ingots were comparable to results obtained for cold samples. The purpose of TLCP testing was to satisfy the waste acceptance system requirements document, which requires that a waste producer determine if the waste for disposal has hazardous characteristics. The method used was to core-drill samples from MWF ingots CFMW06, 07, and 08 and subject them to TCLP. TCLP tests for release of cadmium, chromium, lead, arsenic, selenium, silver, barium, and mercury utilized contact with mild or buffered acid. The solution was agitated at 18 rpm for 18 h. A solid:liquid ratio of 100 g to 2 liters was used. Solid waste that was tested had to pass through a 9.5-mm mesh. The conclusion of these tests was that the core samples from the three MWF ingots pass the TCLP release limits for the eight toxic metals. The answer to whether appropriate casting conditions had been determined was affirmative. The three MWF ingots were cast in the FCF that incorporate an amount of cladding hulls corresponding to two driver assemblies, with the third MWF ingot having been cast using blanket hulls. The characterization techniques employed for the metal waste form product were bulk chemistry, SEM, an immersion test, and the TCLP analysis. The immersion test shows a durable waste form that compares well to cold surrogate alloys. The TCLP analysis indicates that the MWF is not a characteristic waste. The bulk chemistry confirms the anticipated composition. The microscopy identifies the structure and the disposition of the radionuclides to be immobilized. Mark C. Petri, ANL-E, gave information about metal waste release modeling. The purpose of this modeling is to develop a radioisotope-release-rate model for the stainless-steel-based MWF. This in turn is used as an input module for the RIP, the performance assessment code. Another purpose was to assess experimental needs to support modeling beyond June 1999. The MWF modeling approach uses known stainless steel (SS) degradation modes. These include uniform aqueous corrosion and crevice corrosion. A comparison is made between empirical correlations for SS corrosion and MWF data. Assumptions are made about the extent of crevice corrosion across the MWF surface. It is assumed that the corrosion rate gives the radioisotope release rate. Results of the modeling effort for SS uniform aqueous corrosion rate data show that limited data are available for repository conditions. There is significant data scatter. The MWF and SS have similar corrosion rates. The data also show a correlation between electrochemical results and immersion results. The SS crevice corrosion rate data show that data are virtually nonexistent for repository conditions. One electrochemical study indicates a [Cl–] dependence. An empirical corrosion rate correlation has been developed based on SS electrochemical data. The MWF is
OCR for page 99
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report assumed to follow the same correlation. The MWF degradation model has been incorporated into the RIP performance assessment code. Additional MWF corrosion experiments are planned to support modeling. Regression of 316 SS electrochemical uniform aqueous corrosion data shows two key variables: pH and the product pH•T. For 316 SS, the chloride content has no significant effect on uniform corrosion. For 304L SS, passivation breakdown is observed beyond 100 mg/l Cl–. Results for MWF uniform aqueous corrosion data were reported. In simulated J-13 water at room temperature, pH = 4 gave a value of 88 mg/l Cl–, while at pH = 2, the value was 443 mg/l Cl–. MWF corrosion rates were independent of composition within the specified composition range (zirconium, technetium, uranium, etc.). The MWF has corrosion rates similar to 316 SS. No corrosion rate jump was observed with Cl– levels up to 443 mg/l, unlike 304L SS. For 304L SS crevice corrosion, crevice corrosion rates are difficult to measure; very few data are available. There are no data on pH or temperature dependence. A previous 304L SS electrochemical study showed a logarithmic relationship between crevice corrosion rate and chloride and fluoride content (up to 800 mg/l). These are relative (not absolute) corrosion rates. MWF corrosion rate modeling assumes that SS relationships hold for the MWF. Another assumption is a conservative 1:4 ratio between the active crevice area (anode) and the cathodic area. This ratio is the largest seen for SS. Immediate crevice corrosion initiation is also assumed. It is assumed that the crevice provides no cathodic protection, i.e., all regions undergo corrosion at either a uniform rate or at the crevice rate. The net MWF corrosion rate model follows the equation where CR = the net corrosion rate over the MWF surface in g/m2/yr, pH = the pH of the water in contact with the MWF (2 < pH < 10), T = the temperature of the MWF surface in °C (20 °C < T < 95 °C), and [Cl–] = the chloride (and fluoride) ion concentration of the water in mg/l ([Cl–] < 443 mg/l). When looking at the net MWF corrosion rate model vs. chloride concentration, a noticeable increase in the predicted MWF corrosion rate beyond 200 mg/l Cl– was observed. This was attributed to crevice corrosion. E.E. Morris, ANL-E, spoke about repository performance modeling. The purpose of the Argonne performance assessment work is to evaluate the performance of ANL waste forms under repository conditions. This provided feedback on the effect of waste form dissolution parameters on performance in the repository. The performance assessment used a simplified version of the TSPA-VA RIP model. The simplified model permits calculations to be performed 10 times faster than with the TSPA-VA RIP model. It also facilitates parameteric studies. The model was constructed for ANL by Golder Associates Inc., the developer of RIP. Calculations show that the ANL waste form performance is comparable to the TSPA-VA waste form performance. The simplified model used the same engineered barrier system model as the TSPA-VA model. Unsaturated and saturated zone radionuclide transport is represented by simpler models. This produces essentially the same dose rates at the 20-km well as does the TSPA-VA model. The expected-value time histories for dose rate hold, as do complimentary cumulative distribution functions (CCDF) for the peak dose rate. Results for the 39- and 9-isotope model results were compared. The 39 radionuclides used in TSPA-95 are adequate to represent the ANL waste forms. The 9-isotope simplified model is needed because the 39-isotope model is too slow. The 39- to 9-isotope comparisons with no ANL waste are used to verify the understanding of how the actinide inventory adjustments were made for the TSPA-VA. The 39- to 9-isotope comparisons with only ANL waste are used to verify that actinide inventory adjustments are satisfactory for the ANL waste forms. The 9-isotope total dose rate time history agrees well with the 39-isotope result. Several radionuclides that contribute a few percent to the total dose rate in the 39-isotope model are not included in the 9-isotope model. The general agreement within 15 or 20 percent between the total dose rate in the two models was considered satisfactory.
OCR for page 100
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report Conclusions of the parametric study were that the dominance of commercial SNF and the importance of its cladding credit, are evident for the TSPA-VA base case repository loading. In the absence of cladding credit, a waste matrix with a dissolution rate comparable to or slower than that for DHLW glass can significantly reduce the peak dose. The RIP implementation is similar to that for DHLW glass. The model is currently the same as DHLW. The rate data are those from DHLW. The surface area per gram is 1.8 × DHLW and is constant throughout the RIP calculation. The radionuclide inventory is that of ANL ceramic waste. ANL experimental data will be included in future models. The impact of waste form dissolution rates was assessed by comparing time histories of the normalized cumulative release from the engineered barrier system. The impact on the repository was assessed by comparing complimentary cumulative distribution functions for the peak dose rate at the 20-km well. In summary, a RIP modeling capability for ANL waste forms has been developed and demonstrated. The repository performance assessment effort is on schedule to meet the data call for Yucca Mountain project site recommendation/license. Stephen G. Johnson, ANL-W, presented information on the waste qualification strategy at ANL. The goals were to comply with all pertinent legislation on production and ultimate disposal of the waste forms. Another goal was to provide comprehensive plans and adequate records to prove that the process is well behaved operationally and creates a consistent product that can be dispositioned in a geologic repository. A further goal is to consult with the cognizant branches of DOE about the spent fuel treatment process under development at ANL and the waste forms that will be produced. Pertinent legislation on this topic includes Nuclear Regulatory Commission legislation 10CFR60. The NRC regulations in Part 60 of title 10 of the code of federal regulations provide that containment of radionuclides shall be substantially complete for a period that shall be no less than 300 years nor more than 1,000 years, unless otherwise permitted by the NRC. Any release of radionuclides after the containment period shall be a gradual release and limited to certain small fractional amounts based on the calculated inventory present at 1,000 years after closure. This stipulates that data on the waste forms must be gathered and integrated into the performance assessment of the geologic repository. Another pertinent law is Environmental Protection Agency (EPA) legislation 40CFR191,6 which provides that cumulative releases of radionuclides from the disposal system for 10,000 years after disposal shall have a likelihood of less than 1 chance in 10 of exceeding the values stated for each radionuclide in the regulation. This also stipulates that data regarding waste form performance must be gathered and integrated into the performance assessment of the geologic repository. There are a variety of data needs of the geologic repository for high-level wastes (HLW). An Environmental Impact Analysis is required to determine the quantity and type of material acceptable and to determine repository design. Performance assessments, including a viability assessment, site suitability, site recommendation, and license application are also needed. Two documents are related to waste acceptance criteria needs, the WASRD, and the mined geologic disposal system waste acceptance criteria (MGDSWAC). The geologic repository is responsible for the total system performance assessment (TSPA). The objective of the waste disposal system is to contain and isolate the radioactive constituents so that the dose impact to humans is attenuated to a relatively benign level. Water contact is limited, the waste package is long lived, the release rate of radioactive constituents from breached waste packages is low, and a concentration gradient is set during transport of the radioactive constituents that are released from the waste packages. As for how to provide the data required for the repository performance assessment, the National Technology Transfer and Advancement Act (Section 12, 1996) stipulates, “All federal agencies now must use technical standards developed or adopted by voluntary consensus standards bodies (SDOs) when carrying out procurements and policy objectives…. Federal agencies and departments are now expected to work with SDOs and to have their employees participate in the SDO’s development of technical standards.” 6 Yucca Mountain is exempted from 40CFR191.
OCR for page 101
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report To provide the data required by the repository performance assessment, ASTM C1174 is used to provide guidance in the selection of the types of tests and methods to apply. These include attribute tests, characterization tests, accelerated tests, service condition tests, analog tests, and confirmation tests. ASTM C1174 provided the basis for developing the test matrices for the CWF and the MWF. These test matrices are extensive and will provide the data necessary to develop a model for assessing the performance of the waste forms that result from the spent fuel treatment process. The WASRD stipulates chemical composition, radionuclide inventory, phase stability and integrity, and product consistency as requirements. The waste acceptance qualification specification (WAPS), the waste compliance plan (WCP), and the waste qualification report (WQR) are all required documents. The WAPS sets out specifications for the process, e.g., canister dimensions, radionuclide content, and hazardous characteristics. The WCP states specifically, in detail, how each specification in the WAPS will be demonstrated and how that compliance will be documented. WQR compiles the results from the waste form/process testing and analysis to demonstrate the ability of the producer to be in compliance. Test matrices are organized to fulfill the broad data needs for waste qualification. Data needs include the performance assessment, the waste product specification, process qualification, and background tests. Preliminary discussions between ANL and the DOE Office of Environmental Management, Office of Civilian Radioactive Waste Management, and Office of Nuclear Energy have taken place. Topics discussed included product specifications, compliance plans, and classification of electrometallurgical technology waste forms. Waste qualification activities that were to be completed by the end of the demonstration include the test matrices, except for long-term tests and radioactive sample tests. Methods will be established for applying both waste form models to the repository time frame. Sensitivity performance studies will be performed using bounding degradation models. For both waste forms, degradation modeling refinements based on experimental data will be incorporated into the simplified TSPA model. Other activities to be completed include the preliminary waste acceptance product specifications, which will be submitted to DOE for comment. Process parameters will be identified for full-scale operations qualifications Future waste test matrices for license application were to be established by June 1999. Post-demonstration-phase waste qualification activities include the completion of waste form test matrices that support repository license application and are included in the WQR. The WQR was to be written. Waste process qualification tests with full-scale equipment will be performed. JULY 22, 1999—CLOSED SESSION The committee met in closed session on July 22, 1999 at the Shilo Inn in Idaho Falls, ID. Discussions consisted of the previous day’s presentations by ANL personnel and preliminary discussion of the structure and writing assignments for the committee’s final report.
OCR for page 102
Electrometallurgical Techniques for DOE Spent Fuel Treatment: Final Report This page intentionally left blank.
Representative terms from entire chapter: