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.
Fluoride Salt Chemistry, Partitioning, and System Corrosion To provide a technical evaluation of the options identified for the disposition of stored fluoride and flush salts, the chemistry of these salts needs to be understood. The effects of radiolysis and the processing steps required for remediation of these salts, as well as the corrosive interactions of the salts with the piping, storage tanks, and other components, also lead to important chemical changes in the salts. CHEMISTRY RELEVANT TO THE PRESENT STATUS As shown in Table 3.1, the composition of the flush salt in mole percent is 65.9 lithium fluoride (LiF), 33.9 beryllium fluoride (BeF2), and 0.~8 zirconium fluoride (ZrF4~. The fuel salt has the mole percent composition 64.5 LiF, 30.4 BeF2, 4.9 ZrF4, and 0.12 uranium tetrafluoride (AFT. The LiF and BeF2 components were chosen to be thermodynamically more stable than UF4, with a liquidus temperature for the flush salt of about 460°C. The 4.9 mole percent ZrF4 in the fuel salt reduces the liquidus temperature slightly but otherwise serves two functions: 1. For the composition of the fuel salt, depending on the temperature, zirconium may be reduced slightly more readily than uranium (tending to minimize the formation of metallic uranium). 2. The ZrF4 component reacts more readily with oxygen than UF4 and can serve as a getter to scavenge any oxygen impurity. An unusual and favorable property of these salts is that their solidification shrinkage for the given compositions is only about 2 percent. 37
38 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE TABLE 3. ~ Composition of the Fuel and Flush Salts Fuel Salt Flush Salt Current salt composition (mol YE LiF 64.5 65.9 BeF2 30.4 33.9 ZrF4 4.9 0.18 Mass off ssile elements in salt big) Uranium <33.2a 0.5 Plutonium 0.724 0.013 Current distribution of uranium and plutonium isotopes (wt tY) 232u 160 ppm 233u 83.92 234u 7.48 235u 236u 23su 239pu 24opu Other Pu 2.56 0.104 5.94 90.1 9.52 0.35 75 ppm 39.4 3.6 17.4 0.245 39.4 94.7 4.8 0.50 aAssumes that at least 1.8 kg of uranium has migrated to the omegas system and at least 2.6 kg of uranium is loaded onto the auxiliary charcoal bed. SOURCE: Peretz (1996c, Table 1.3). The fuel salt solids can be viewed as a matrix of closely packed fluoride ions with the much smaller Be2+ and Li+ ions occupying interstices. At their low concentrations, Zr4+ and the larger U4+ cation contribute little to spatial requirements (Zachariasen, 1948~. Indeed, on the basis of fluoride volume alone, the fuel salt density is estimated (Robert Penneman, unpublished), using the data of Zachariasen (1948), as 2.52 g/cm, close to the 2.48-g/cm3 density reported (Peretz, 1996c). Chemical Consequences of Radiolysis As a consequence of the radiolysis reaction, some cations of the stored crystalline salt may be reduced to metal atoms. Peretz (1996c) suggests that any of the metal atoms lithium, beryllium, uranium, or
CHEMISTRY, PARTITIONING, AND SYSTEM CORROSION 39 zirconium could be formed; this is uncertain because the reduction potentials for the metals in fused salt at 500°C are not relevant in the crystalline solid phases. Perhaps the identity of the metal atoms is not important, since they can be considered simply as electronic point defects in the crystalline salt phases, not metal particles. Another important consequence of the radiolysis reaction, besides the loss of gaseous fluorine and uranium hexafluoride (UFO) by migration to the piping system, is that the highly reduced multiphase salt has physical and chemical properties, such as an apparent reduction in the solubility of various fluorides in the melt, that differ from the properties of the original, single-phase fused salt. To better inform remediation strategy, it is important to gain additional information on the (as yet) poorly defined chemical properties of the remelted salts. PARTITIONING OF URANIUM FROM THE SALT As noted in Chapter 2, sampling of the gas phase near the end of the off-gas vent line indicated high concentrations of F2 essentially saturated with UFO vapor at the current temperature of the vent system (~21°C). This gaseous product results from radiolysis reactions in the solidified salt phase, and it would be prudent to assume that UFO has been distributed to, and condensed in, all regions of the system, including the freeboard volume in the upper regions of the drain tanks as well as areas of restricted flow in the off-gas vent system preceding the auxiliary charcoal bed (ACB). Due to the significant amount of alpha decay arising from the 232U daughters that grow into the chain with a 1.9-year half-life, radiolysis effects make it probable that a significant fraction of the initial UFO may now be present as non-volatile lower uranium fluorides. It is possible that back-migration of oxygen or water vapor from the ACE into the vent line has occurred. This could result in the precipitation of urany! difluoride (UO2F2) as one of the species currently restricting flow in the vent line and in the production of hydrogen fluoride (HF) gas. Today the ACB is isolated from the vent line. The staff at Oak Ridge National Laboratory (ORNL) proposes to clear the off-gas line by initially purging with an inert gas (helium or
40 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE argon), followed by evacuation to volatilize any UFO residing in the off- gas system; this approach shouicI remove all UFO, but it would not remove lower fluorides that may be present in the system. Several chemical options can be considered for the removal of nonvolatile uranium residues remaining after the initial pumping. Considerable fluorination technology that has been developed over the last 20 years at other Department of Energy (DOE) sites may be applicable to the Molten Salt Reactor Experiment (MSRE) remediation effort. These options may be of use both for clearing the piping and traps in the off-gas lines and for recovering 233U from the remelted MSRE salt fractions by volatility processing methods. Concern exists about whether the solidified salt can be remelted to form a homogeneous liquid since significant fluorine has been lost from the original melt (Chapter 2~. To address this issue, Chapter 4 describes a recommendation to attempt to refluorinate the salt components during a heating phase short of actual melting by addition of HE gas in helium (He) at a slightly positive pressure. It is expected that HF-He would oxidize all uranium or plutonium compounds to the IV oxidation state at 400°C (just below the melting temperature) if adequate permeability exists in the solidified salt. Additionally, refluorination can be continued during progressive melting with a rock-melting, laser, or cairod apparatus to ensure that significant quantities of fissile material do not accumulate as a critical deposit. Hydrogen fluoride will not oxidize uranium or plutonium to the hexafluorides UFO or PuF6, and molecular fluorine at temperatures less than 300°C may not be effective in producing UFO or PuF6.2 In order to generate UFO from UF4, especially at the temperature existing in the off-gas piping, alternative fluorination agents should be considered to boost the reaction rate. These would include atomic As is well known to MSRE project personnel (Rushton et al., 1996a,b), any such central heat source to induce localized melting need not be powerful enough to supply all the requisite heat energy to the salts. The resistance heaters on the external tank wall could be used to elevate the salt temperature. 2Fluorine reacts with UF4 to produce UFO in argon-neon mixtures at temperatures as low as 12 K. Large-scale, efficient production of UFO may require higher temperatures or ultraviolet irradiation (Margrave et al., 1976, 19773.
CHEMISTRY, PARTITIONING, AND SYSTEM CORROSION 41 fluorine, bromine pentafluoride (BrFs), chlorine trifluoride (ClF3), fluorine dioxide (O2F2), dioxygen monofluoride (O2F), and krypton difluoride (KrF2). All of these have significant advantages and disadvantages, as discussed in Appendix B. PLUTONIUM PARTITIONING BY FLUORINATION Peretz (1996c, p. 3-24) referred to laboratory and pilot plant studies involving fluorination of surrogate salt containing both uranium and plutonium. The results showed that very little plutonium was removed, even after most of the uranium was volatilized as UF6.3 Furthermore, the high-temperature fluorination yield of PuF6 was low, and long fluorination times caused increased corrosion. These results are consistent with the understanding that, because PuF6 is a powerful fluoridating agent, little PuF6 would be expected to be volatilized from the molten salt phase until all other volatile fluoride gases (including all the UFO) are volatilized and driven out of the molten salt solvent. Experience at Los Alamos (Mills, 1996) suggests that volatilization of such small amounts of PuF6 (at the level of ~ 55 parts per million [ppm] in the existing salt) is difficult, especially if plutonium exists in the salt phase as a double salt with one of the major components (e.g., LiF). Ambient-temperature fluorination using the reagent O2F2 is also unlikely to be successful in removing plutonium from the bulk salt (see Appendix B). Based on these considerations, fluorination does not appear to be usable for removing plutonium from the salt. The depleted salt will retain the plutonium (and fission products). 3A 45-hour fluorination of 2000 liters of MSRE molten salt removed 216 kg of uranium and left a uranium residue of 26 parts per million (ppm; Peretz, 1996c). No uranium was volatilized during the first 7~/: hours while conversion of lower oxidation state uranium to uranium pentafluoride (UFs) occurred. Subsequently, UFO was volatilized at about 6-7 kg per hour. During the last 6 hours, fluorine utilization dropped to nearly zero, and essentially no plutonium was removed (the final plutonium concentration was 1 10 ppm, compared to an initial value of 1 12 ppm).
42 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE NONFLUORINATION OPTION FOR PLUTONIUM SEPARATION If plutonium separation becomes essential, the following work at Los Alamos might be the basis for development: Los Alamos investigators (Ioel Williams, Los Alamos National Laboratory, personal communication with Melvin Coops) have found that it is feasible to strip small amounts of plutonium from molten salt electrorefining residues (a eutectic salt of sodium chloride and potassium chloride [NaCI-KCI]) by bubbling oxygen through the molten salt to precipitate plutonium (and americium) as the insoluble dioxide. The salt is then evaporated for recycle, leaving the oxide residues as a waste product. Since oxygen gas has a low solubility in the molten salt, addition of sodium carbonate to the molten salt has been found at Los Alamos to be a suitable way to precipitate the plutonium-americium oxides. If the MSRE molten salt is sufficiently fluid to be either filtered or centrifuged to separate a precipitate of zirconium-plutonium oxide, the method utilized at Los Alamos may be an applicable technique for isolation of the small amount of plutonium present in the MSRE drain salt. Addition of lithium carbonate (Li2CO3) to the salt after uranium removal is complete may be a simple and effective method of adding oxygen to precipitate plutonium oxide (Pu02), with zirconium oxide (ZrO2) also likely. SYSTEM CORROSION ISSUES In considering the recommendation for in situ melting of the stored salts (first the flush salt and then, if successful, the fuel salt) and pressurization of the tanks to push the molten salt through the existing (cleaned and heated) valves and pipes, it must be ascertained that these Hastelloy N components have not suffered severe corrosion damage during their use and storage periods. Hastelloy N was a good choice of material initially, because the high-nickel alloy is least susceptible to fluorination attack at high temperature (Lad, 1990~. Although fluorine corrosion at high temperature is accelerated significantly by the presence
CHEMISTRY, PARTITIONING, AND SYSTEM CORROSION 43 of water vapor (HF is formed), in this case water vapor would also have been removed from the gas phase by the formation of ZrO2 and UO2F2. Any significant corrosion of the alloy at ambient (or warm, 200°C) temperature would require the presence of an aqueous (or water vapor) phase. There is no evidence that system internals have been exposed to moisture via any leak in the system. ORNI, corrosion researcher James Keiser inspected one valve that had been plugged and cut out of the system and found no evidence of observable corrosion attack. As one possible means of observing on a gross scale the condition of the Hastelloy and the salt inside the storage tanks, an optical fiber inspection system could be considered. A cored salt sample taken from the center of a storage tank could not be expected to contain the chromium, iron, and other solutes that would indicate tank corrosion because any corrosion products (particularly nonvolatile compounds or those that complex with fluoride salts) would be localized to the immediate surface of the alloy. Radiation-Induced Corrosion Questions The 30-year-old drain tanks have been experiencing a radiation field and radiation-produced compounds, and one can ask what effect these compounds have on the integrity of the Hastelloy N walls. As noted previously, there is about one atmosphere of excess pressure over the fuel drain tanks. This is likely to be largely F2 combined with the saturation pressure of UFO (and HF-O2 [molecular oxygen] if there has been any leakage of moist air). Chapter 2 addresses the effects of alpha self- radiation on decomposing, solid UFO. However, radiation effects on the F2 gas could cause some reversal of UFO decomposition. Indeed, F2 plus radiation, especially in the presence of 02, iS an aggressive reaction mixture. Even at ambient temperature, at which F2 itself will not react, enhanced chemical attack of the fuel tank wall might take place in a radiation field, liberating molybdenum or chromium hexafluorides (MoF6 or CrF6) as volatile gases and possibly corroding the nickel by forming Li2NiF6 The F2 and UFO gases are more reactive and corrosive than the solid salt, and it is unknown at the present time whether these gaseous
44 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE; species are in contact with the interior drain tank wall. Because the fuel salt shrinks approximately 2 percent on solidifying, a gas space may exist between the solid salt surface and the interior tank walls. Alternatively, a coating or film of solid salt may be deposited, protecting the tank wall from reactive gas attack. Some laboratory-scale corrosion tests could be performed, but these would take time and would be of only limited value, since it would be difficult for such tests to duplicate stresses or other local conditions such as pitting. The ideas4 presented below are offered for consideration by ORAL staff to weigh the merits of obtaining this information against the time and efforts required for meaningful results. Corrosion tests would be conducted over irradiation times short compared with the 27- year radiation exposure of the fuel tank walls, even though the integrated dose could be similar. For example, samples of metal of composition equivalent to the tank and thimble tubes could be prepared and sealed in contact with a variety of gas environments, such as F2, F2+ 02, and 50-50 mixtures of UF4 and UFO. Then these samples could be irradiated with gamma rays (for example, from the High Flux Isotope Reactor tHFTR] spent fuel elements) or with a source of alpha particles to simulate the history of the drain tank walls in the presence of radiation fields. Over time, the samples could be tested for metal corrosion and for formation of volatile UFO. To provide perspective on the severity of these corrosion issues, the fact that the tanks and associated piping continue to hold gases at a pressure of more than one atmosphere after more than 25 years suggests that there are no leaks in the metal confinement at this time. 4Many techniques, such as standard procedures in nondestructive testing, are not practical due to the intensity of the radiation field at the tank wall (see Chapter 2), necessitating the use of remote operations to access the drain tanks.