Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Appendix E Hazard Scoping of Major Actions for Remediation This appendix provides a more in-depth view of the detailed technical considerations for each of the process steps outlined in Chapter 4 in the development of a preferred approach. The execution of any remediation program involves many detailed selections in chemistry, engineering, and safety and monitoring provisions. A typical sequence (not necessarily optimal or comprehensive) of such selections is outlined here but is not intended to provide a definitive recommendation for any step in the process. Such decisions are best made by project personnel as new information is obtained. The suggested actions, some of which are already being planned or are under way at Oak Ridge National Laboratory (ORNL), include the following: Action 1: Removal (principally from the piping and the freeboard of the tanks) and analysis of the reactive gases f uorine IFS and uranium hexafI~oride LUFF from the vent lines of the salt drain tanks, now under pressure. The amount and composition of the gases can provide important but partial information about conditions of the stored fuel salts and some clues as to the condition of equipment (including possible leakage or exposure to water via a pressure of hydrogen fluoride tHF] and oxygen t023~. Action 2: Removal of UFO gas specifically from the piping and tankfreeboard. The gas should be collected and its amount and composition determined. These determinations will improve the estimates of the degree of reduction of the salt relative to the oxidation state while the E.1
E.2 AN EVALUATION OF DOE ALTERNATIVE FOR MSRE reactor was in operation. These data would aid a uranium mass balance calculation, and significant uranium recovery would reduce the criticality potential associated with the drain tank remediation alternatives. Action 3: Removal of nonvolatile uranium (if it existsJ from the piping by fluorination. If any aggregate amount remaining is small relative to criticality (e.g., less than 250 g), it may be left for the last step after the tank contents are processed unless the residue results in plugged lines. Final removal may be either by fluorination or by dissolution. If simpler measures (such as addition of any of the fluorinating agents of Appendix B as gases) fail, another possible approach for removing nonvolatile plugs in the lines is to introduce an aggressive oxidizer such as krypton difluoride (KrF2) in anhydrous HE, to carry KrF2 to the plugs in a liquid solution. This approach is not without difficulty it would require cooling of the lines below the 19°C boiling point of HE (or working at greater than atmospheric pressure), and in the presence of a base such as lithium fluoride (LiF), it would enhance the attack of KrF2 on nickel surfaces. Action4: Obtain information about the distribution and segregation of uranium in the salt, by one or both of the following methods: Option 4A. Gamma spectroscopy Option 4B. Collection of a salt sample Obtaining a solid sample of the fuel salt in the drain tank appears to be an operation with hazards that result from gaseous radon contamination, as discussed in Chapter 4 and Appendix C. The sampling method considered here by the panel is a simple core drilling operation, with the drill and contained sample removed from the drain tank to a hot cell for sample handling and analysis, conducted after the reactive gases F2 and UFO have been rem overt from the Coining ~nr1 the drain tank freeboard. r -or ---= Obtaining fully representative samples of the drain tank salt in its present solid condition is highly desirable but may not be possible without significant spreading of radioactive contamination. Any bulk
APPENDIXE HAZARD SCOPING FOR REMEDIATION E.3 sample provides some information on the extent of segregation present, some of which might be expected due to zone refining effects during initial solidification. The ideal goal of such characterization is to map the spatial distribution of uranium-233 (233U) and plutonium within the tanks, and their chemical states. If practical, several full-length core samples, at different radial distances from the center of the tanks, can provide a more precise measure of possible inhomogeneity in elemental composition and chemical state and, therefore, of the potential for segregation (and possible degree of precipitation) on melting. Because of solubility uncertainties, melting the salt may not by itself ensure homogeneity. Even though it may not be truly representative of the entire salt mass, one full-length core sample can provide significant data on the extent of segregation. Gamma spectroscopy and mapping are mentioned as actions that are less hazardous than core sampling. They can provide information on uranium distribution that will help inform a decision on remediation strategy. Action 5: Determination of the structural integrity of the file! and flush salt tanks and their potential for leakage during subsequent operations (see options 6A, 6B, and 6C). Several hazards to tank integrity must be evaluated. One hazard is that the tank may have experienced some degree of corrosion during operation or during exposure to fluorine (and possibly to nascent fluorine species and fluorine radicals formed by radiolysis) and to ionizing radiation over the extended storage period. Another hazard of tank failure is associated with the melting of the salt. This hazard includes possible clifferential thermal expansion effects, depending on the directionality and overall pattern of the melting process. These effects can be monitored during melting and the heat inputs arranged so as to minimize differential effects by relieving most volumetric effects on the free surface of the salt bed. A third hazard is that tanks may leak due to progressive corrosion during fluorination operations. A water leak could seriously complicate these problems. It is recommended that all available techniques for condition assessment be considered. Analyses of cell gas, tank gas, and any salt samples can provide bounding information on tank condition and possible general corrosion and on possible existing leaks. Ultrasonic
E.4 AN EVALUATION OF DOE ALTERNATIVE FOR MSRE testing, where feasible, could provide more direct evidence of general wall thinning. None of these methods is able to provide reliable evidence that there is not some localized pitting or cracking that may progress to leaks. However, if the measured rate of general corrosion is low and in accord with prior laboratory results, the likelihood of unexpected severe localized corrosion is reduced. The most likely failure mode is projected to be corrosion of the thinnest members, the cooling thimbles, and the more highly stressed region in and near the welds. This hazard can be eliminated by removing the bayonet coolers or by plugging thimble penetrations at the vessel wall. Corrosion rates might be reduced by using a lower-temperature fluorinating agent (such as bromine pentafluoride tBrFs]; see Appendix B) than fluorine. The relative corrosion rates for different time- temperature-reagent compositions for alternative fluorination processes should be tested and confirmed, and the process conditions optimized, by using small-scale laboratory tests. In risk terms, this hazard appears likely to remain seriously uncertain unless creative inspection methods are found or developed. This fact suggests that procedures having minimum reliance on tank integrity be given priority; these include the following: . FI?~orination at annealing temperatures without melting. Because some fluorine species have been able to diffuse out of the solid salt near room temperature, it is reasonable to test whether fluorine (or HF) will diffuse into the solid salt and react at useful rates at annealing temperatures lower than the melting point. · Controlled zone melting and simultaneous fluorination, musing the unmelted" portion of the salt bed as a "skull" or frozen crust for most of the melting and fluorination operation. With proper monitoring, this can be used to bound the possible extent of segregation and precipitation of reduced species. . Contingency plans for coping with leaks of reagent gas and UFO. · Contingency plans for stopping melting anal fluorination and reverting to removal of salt as a solid, if unavoidable. As mentioned in Chapter 4, solid removal by carbon dioxide (CO2) blasting is a contingency that it is preferable to avoid, because of the potential spreading of contamination of gaseous radioactive radon.
APPENDIX EHAZARD SCOPING FOR REMEDIATION E.5 Action 6: In-tank fluorination for extraction and collection of ~ U. At least three methods can be identified: . Option 6A. Preliminarily conduct hydrofluorination without melting to recover normal "oxidation" states of UF4 and PuF4 (uranium and plutonium tetrafluorides) (e.g., using HE or fluorine-helium mixtures at annealing temperatures), followed by options 6B or 6C to extract uranium. Option 6B. Zone melt the salt in place progressively and fluorinate the uranium content to UFO using an appropriate fluorination gas mixture at elevated temperature (e.g., fluorine at 500°C). · Option 6C. Zone melt the salt in place progressively and fluorinate the uranium content to UFO in place using alternative fluoridating agents (e.g., BrFs or KrF2; see Appendix B). . There are two noteworthy hazards to this operation. The first is failure of the vessel due to preexisting corrosion damage and the effect of accelerated corrosion during fluorination. Analysis of the hazard of vessel failure requires information about the present condition of the vessel and the projected corrosion rates for fluorinating conditions. The second hazard is the possibility of criticality due to possible segregation of fissionable material. The criticality hazard can be limited by using a low salt melting rate and by progressively fluoridating out the uranium content of the melt zone before further melting occurs. By operating in a quasi-batch size mode, it should be possible to monitor the uranium removed from the melted volume and avoid having a critical mass of uranium accumulating in the melt and subject to precipitation. This does not necessarily address the question of the behavior of possible insoluble plutonium species. The amount of plutonium appears small enough (650 g divided among the two drain tanks, as reported in Peretz, 1996c) that by itself it would be safely subcritical in any configuration in the absence of moderator. However, the case of incomplete dissolution and fluorination of uranium leaves open the possible scenario of a mixture of plutonium and uranium segregating in the drain tank vessel. Only if an effective fissile concentration of tens of grams per liter were achieved in a large enough, sphere-like configuration could criticality be possible. The detailed analyses under
E.6 ANEVALUATION OF DOE ALTERNATIVE FOR MSRE way at ORNE address the presence of plutonium along with the credible uranium and salt configurations. Regarding option 6A, even if recovery of the normal oxidation state is only partial, when combined with zone melting and monitoring it appears likely to eliminate the possibility of criticality caused by precipitation. Remaining uncertainties in the conditions required for this process can be determined by a series of laboratory melting tests with simulated "reduced" uranium in salt so as to permit measurement of diffusion and reaction rates. A continuous monitor of the degree of subcriticality of the system, such as sensitive neutron monitoring, would provide a measurement to support calculations and provide additional assurance of safety. The (a,n) reactions with fluorine and beryllium provide an internal neutron source that is augmented slightly by the subcritical neutron flux (ken approximately 0.85) from fission events. Significant increases in neutron flux might be readily observable far from the critical configuration. The sensitivity of such a measurement could be enhanced further, if necessary, by introducing a stronger external neutron source (e.g., a pulsed neutron source). If the processes of melting and fluorination to separate and recover the uranium and transfer it to another vessel are acceptable for selection on the basis of other criteria, means to ensure large margins of criticality safety appear feasible, subject mainly to further detailing of the process steps. To support this position, a series of conditions that the panel considers unlikely would all have to occur simultaneously without the exercise of detection or control measures. These conditions are 1. substantial segregation of the fissile species, on the order of a large fraction of the total inventory in the drain tanks; inventory; 2. attainment of high density of the fissile material; 3. absence of continuous tracking of the location of the 4. absence of monitoring of possible increased neutron multiplication; and 5. failure to take available corrective actions if trends toward any of these conditions were to occur.
APPENDIXE HAZARD SCOPING FOR REMEDIATION E.7 Given the normal disciplines of carrying out the intended types of operations with intensive planning and training and with the usual multiple levels of internal and external review and monitoring, the pane! believes that the risk of criticality can be kept acceptably small. Specifically, it should be readily feasible to maintain less than ~ percent chance each for conditions 3, 4, and 5 above. Condition 2 may be physically impossible and is subject to experimental verification. The combined chance of the occurrence of criticality can be estimated reasonably to be substantially below the target levels generally considered satisfactory by nuclear safety authorities worldwide. This preliminary estimate is assumed to be subject to documentation and verification by a Safety Analysis Report, or its equivalent, normally required for all operations involving the processing or transport of Missile materials. Accordingly, the pane! believes that criticality poses negligible risk to the planned operations. The hazards of uranium-depleted salt removal appear to be substantially smaller than the hazards of "loaded" (i.e., fissile material- bearing) salt removal discussed in the next action. Action 7: Processing office! andflush salt external to the present drain tanks. Here two options have been identified: . Option 7A. Melt and transfer the salt to a separate vessel for fluorination of uranium to UFO, preferably using option 6C or well- monitored zone melting to prevent undetected segregation, and subsequent recovery of uranium. The following procedural comments are offered: I. Conduct a series of trial runs with melting of simulated reduced-state salt to establish possible segregation behavior. 2. Consider a possible option for further evaluation: run a full- scate pilot melting test with the flush tank salt, preceded by an annealing treatment with a gas atmosphere to simulate the partially reduced conditions of the fuel salt. Option 7B. Excavate the salt as a solid from the drain tanks (e.g., by CO2 blasting) and transfer to a new vessel for
E.8 AN EVALUATION OF DOE ALTERNATIVE FOR MERE 1. conventional fluorination of the uranium to UFO with options similar to options 6A, 6B, and 6C above; 2. electrorefining according to methods proposed by Argonne National Laboratory; or 3. no further treatment operations beyond stabilization for storage. The following principal hazards in option 7B are noted: 1. It may be impractical to effect complete removal of the salt by CO2 blasting due to internal tank structures or to impractically low rates of salt removal. 2. Carbon dioxide could react with F2 to form carbony! fluoride (COF2) or even carbon tetrafluoride (CF4) in an exothermic reaction. 3. Containment of radioactive species (in particular, dispersal of radon, as discussed in Chapter 4 and Appendix C) would seem to be aggravated by a technique such as CO2 blasting. 4. The large volumes of gas to be handled could be subject to leaks or mechanical blowouts and subsequent material releases. Secondary containment may be required to limit this hazard. 5. This procedure may not entirely eliminate the hazards of segregation. in fact, segregation may be caused by density differences of solid particulates. 6. Dry runs with simulated salts can establish optimum removal conditions and likely rates of removal. 7. A full-scale test can be done by using the flush tank and salt, avoiding criticality hazards, and minimizing corrosion and leakage hazards. However, the flush tank may not be a good surrogate for the physical removal test because of the absence of thimbles. The mitigation measures already taken are recognized as useful, but the pane} believes that preventive measures are also needed because an unexplained event, even if well mitigated, can raise concerns of institutional credibility that could render it difficult for the Department of Energy to proceed with fuel and flush salt remediation on an orderly schedule and budget.
APPENDIX E HAZARD SCOPING FOR REMEDIATION E.9 Action S: Interim storage of the separated 233 U. Transport of UFO to the existing on-site 233U storage facility without further chemical conversion is a convenient option with well- def~ned and experienced protocols and generally well-controlled hazards. Radiation decomposition would be expected if the uranium were left as a fluoride. This could be avoided by conversion to the uranium oxide (UPON), a standard operation. The option of converting UFO to U3O~ for storage is also a practical one with experienced conversion and transport techniques, generally well-def~ned procedures and equipment, and well-controlled hazards. Stabilization of the salt residues after fluorination by chemical Bettering is a less practiced operation but with less significant hazards.