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.
2 Radiolysis and Nuclear Reactions This chapter provides technical details about current information available on the salts in the tanks. This assessment is important for any remediation strategy. The discussion that follows focuses on the radioactive species and nuclear reactions relevant to cleanup operations. The now solid 4650 kg of fuel salt in two of the drain tanks contains nearly all of the uranium, plutonium, and fission products, with a minor amount in the flush salt in the third tank. Ongoing radiolysis effects have had important chemical consequences on the contents of the drain tanks. Substantial amounts of both fluorine gas and uranium hexafluoride (UFO) have been produced, allowing vapor transport of more than 10 percent of the original 37.1 kg of uranium (see Table 2.~) from the fuel salt storage tanks. An unknown fraction of the uranium is in connecting piping and traps, with approximately 2.6 kg in the auxiliary charcoal bed (ACB) located at the end of the off-gas vent line (see Figure 1.5~. Salient features of the Peretz (1996c) report that are pertinent to this discussion are summarized here and cover radioactive source terms, radiation effects, and general comments, including discussion of radiation-induced liberation of fluorine (F2) and uranium hexafluoride gases, radiation `decomposition of solid uranium hexaBuoride, the effects of an excess of reducing species in the salt, and the difficulties associated with a simple remelting operation. RADIOACTIVE SOURCE TERMS When reactor operation was terminated in late 1969, the molten fuel salt was transferred from the reactor to two drain tanks and ~1000 kg (one metric ton) is approximately equal to 2205 pounds, slightly smaller than the English long ton (2240 pounds) and larger than the short ton (2000 pounds). 23
24 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE TABLE 2.1 Mass (kg) of Uranium and Plutonium in Fuel and Flush Salt Mass (kg) if No Losses Mass (kg) if 4.4 kg U Lost from Fuel Salta In In In Fuel Flush Fuel In Drain In Drain Isotope Salt Salt Total Salt Tank 1 Tank 2 Total U 37.1 0.5 37.6 32.7 17.4 15.3 232U b 233u 31.1 0.2 31.3 27.4 14.6 12.8 234u 2.8 0.02 2.8 2.4 1.3 1.1 235u 1.0 0.09 1.0 0.84 0.45 0.39 236U 0.04 0.001 0.04 0.03 0.02 0.02 238u 2.2 0.2 2.4 1.9 1.0 0.9 Total Pu 0.72 0.013 0.74 0.72 0.39 0.34 239pu 0.65 0.013 0.66 0.65 0.35 0.30 24opu 0.069 0.0006 0.070 0.069 0.037 0.032 Other Pu 0.003 0.0001 0.003 0.003 0.001 0.001 Abased on transport of 2.6 kg to the auxiliary charcoal bed and presence of 1.8 kg as UFO vapor in the off-gas piping and vessels. bThe 232U/233U ratio is assumed to be the same in both the fuel and the flush salts. This implied, as of 1995, a 232U content of 160 ppm (parts per million) of the total uranium content in the fuel salt and 75 ppm of the total uranium content in the flush salt (the 233U isotopic fraction in the flush salt is approximately half that in the fuel salt due to the higher concentrations of 235U and 238U in the flush salt). SOURCE: Modified from Peretz (1996c, Table 1.4~. allowed to coo! and solidify. During the intervening years, the short-lived radioactive fission products have decayed to insignificant levels. This has left strontium-90 (9~Sr; with a 28.5-year half-life) and cesium-137 (tics; with a 30-year half-life) and their short-lived daughters as the primary sources of fission product radioactivity (mostly beta and gamma radiation). According to Table 2.2, a total of 23,985 curies (Ci) exists in the fuel salt (in the inventory projected for December 1999),2 2A curie is a measure of radioactivity that is now defined as 3.7 x 10~° disintegrations per second; it originated historically as a representation of the radioactivity of one gram of radium (226Ra)
RADIOLYSIS AND NUCLEAR REACTIONS TABLE 2.2 Inventory of Fission and Activation Product Isotopesa Inventory Inventory Radioactivity in Fuel in Flush Isotope Half-Life (Ci) Salt (98.3%) Salt (1.7%) 90Sr 28.5 y 6,670 6,557 113 90Y 2.7 ~ 6,670 6,557 ~ I] 137Cs 30 y 5,600 5,505 95 137Ba 2.6 m 5,290 5,200 90 i5~Sm 90y 117 115 2 147Pm 2.62 y 13.4 13.2 0.2 is5Eu 4.96 y 4.4 4.3 0.1 i54Eu 8.8 y 3.1 3.0 0.1 is2Eu 13.3 y 1.1 1.1 0.0 99Tc 2.1 x lOsy 0.5 0.5 0.0 ~25Sb 2.73 y 0.3 0.3 0.0 93Zr 1.5 x 106 y 0.3 0.3 0.0 Total . . NOTE: d = day; m = minute; y = year. aBased on decay to December 1999. SOURCE: Peretz (1 996c, Table 1.6). 25 24,400 23,985 415 approximately 99 percent of which is contributed by the isobaric pairs 9 Sr/Y and ~37Cs/Ba.3 Uranium, plutonium, americium, and their decay products contribute substantially less radioactivity (1496 Ci, primarily alpha activity from the 232U and 233U decay daughters) in the fuel salt (see Table 2.3) but contribute a major radiation hazard (the 2.6-million electron volt [MeV] gamma from thallium-208) that affects worker safety. 3Y = Hum; Ba= barium.
26 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE TABLE 2.3 Inventory of Actinide Isotopes Based on Decay to December 1999. Radioactivity Isotope Half-Life (Ci)a 232 U and its daughter chain 232U 70 y 228Th 1.9 y 224Ra 3.66 d 220R~ 55.6 s 2~6po 150 ms 2~2pb 3.25 h 2~2Bi 1.01 h 2~2po 45 s 208Tl 3.05 m 233 U and its primary daughter chain 233u 1.59 x 105 y 229Th 7300 y 225Ra 14.8 d 22sAc 10 d 221Fr 217At 213Bi 213po 2o9pb Activity in Fuel Salt tci~b 27 4 kg 233 U 113 116 116 116 116 116 116 74.3 41.7 Activity in Flush Salt (Ci) b 0 2 kg 233 U 0.81 0.83 0.83 0.83 0.83 0.83 0.83 0.53 0.30 4.9 m 32ms 45.6 m 4ms 3.25 h 129 132 132 132 132 132 132 84.8 302 265 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.8 0.70 0.70 0.70 0.70 0.70 0.70 0.70 0.70 1.90 0.01 0.01 0.01 0.01 0.01 0.01 0.01 0.01 Other signifcant uranium and transuranicisotopes 2.5kg U 0.02 kg U 234u 2.45 x lOsy 17.4 15.2 0.11 98.2% Pu 1.8% Pu 23spu 87.7 y 0.89 0.87 0.02 239pu 24,110 y 41.7 40.9 0.75 24opu 6540 y 15.3 15.0 0.28 24~pu 14.4 y 212 208 3.82 241~ 433 y 23.2 22.8 0.42 Summary of actinide inventories Totalactinideactivity,Z<92 931 815 5.9 Total uranium activity 448 393 2.8 Total transuranic 293 288 5.3 activity Total activity 1672 1496 14.0 NOTE: d = day; h = hour; m = minute; ms = millisecond; s = second; y = year. aBased on the total inventory of 31.3 kg 233u and 2.8 kg 234u. 232u/233u ratio assumed to be constant. bThis assumes a migration of approximately 4 kg 233U from the salt. SOURCE: Modified from Peretz (1996c, Table 1.7).
RADIOLYSIS AND NUCLEAR REACTIONS RADIATION EFFECTS AND GENERAL COMMENTS 27 Radiation-induced changes to salt chemistry from fission energy and fission events ceased at reactor shutdown, leaving radioactive decays (from alpha, beta, and gamma particle radiation sources) to continue to drive chemical effects. Such chemical effects result from ionization processes that occur within the salts as the radiation energy is absorbed in the solids. In a gross view, one can speak of the distances (ranges4) within which the energies of the alpha and beta particles are expended and in which the particles are stopped. in the fuel salt, which has a density of nearly 2.5 g/cm3 (Table 2.4), the range of an alpha particle is very short, nominally a few milligrams per square centimeter, whereas that of beta particles is longer and can be several grams per square centimeter. By contrast, gamma radiation is much more "penetrating," with intensity diminished primarily by the vessel wall thickness rather than by the salt mass. Consequently, the energy from gamma rays are not all be expended within the volume of the fuel salt. The gamma levels at the surface of the fuel salt tanks due to fission products are approximately 640 roentgens (R) per hours (Williams, 1995), largely due to the 0.66-MeV gamma rays from arcs. The existence of a high radiation field from the gamma radiation external to the fuel salt storage vessels makes it difficult for personnel to work on nearby equipment within the shielding except by remote- handling techniques or by brief operations designed to limit the episodic exposure to acceptable levels. Equipment can deteriorate over an extended period of time in the presence of radiation-induced chemically reactive gases such as fluorine and uranium hexafluoride. It would not be unusual for this to result in mechanical problems such as stuck valves, malfunctioning gauges (giving faulty readings), plugged lines (often at right-angle joints), and welds with pinholes. It is prudent to design for extensive equipment repair or replacement prior to any actual treatment of the radioactive salt. 4This range is often given as a weight per square centimeter. sHuman exposure to the radiation field at the external wall of the drain tanks would reach life-threatening levels in one to two hours, but such exposure is prevented because operations are performed outside the shielded drain tank cell.
28 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE TABLE 2.4 Inventory of Salt Stored in the Fuel and Flush Drain Tanks Volume Density Tank Mass (kg)a (mafia (~mL at 26°C) Fuel salt Fuel salt drain tank 1 2479 1.00 2.48 Fuel salt drain tank 2 2171 0.88 2.48 Total fuel salt in drain tanks 4650 1.~8 Flush salt Flush salt drain tank 4265 1.92 2.22 aMass and volume estimates that best correspond to process history. SOURCE: Peretz (1996c, Table 1.2). Radiation-Induced Liberation of Fluorine and Uranium Hexafluoride Gases As shown in Table 2.l, the stored fuel salt originally contained 37.6 kg of uranium isotopes (84 percent 233U, 7.5 percent 234U, 5.9 percent U. 2.6 percent U) and 0.74 kg of plutonium isotopes (90 percent Pu, 9.6 percent 240Pu, 0.4 percent 24~Pu, and other isotopes), which is important to radiolysis (and criticality) concerns. The radiolysis reaction is probably the most important chemical consequence (discussed further in Chapter 3) for storage of the fuel salt, causing the release of fluorine atoms and molecules and the formation of volatile UFO. Radiolysis led to pressurization of the piping system, migration of uranium into the pipes and ultimately to the ACE, and evidence that decomposition of the volatile UFO had deposited some solid plugs (probably uranium tetrafluoride [UF41) in the reactor vent pipes. Peretz (1996c, p. 1-11) and Williams et al. (1996) reported radiation-induced liberation of fluorine gas from simulated fuel salts. The release of F2 apparently is due to recombination of fluorine atoms liberated by radiation effects on the fluoride salts, with each radiation source (alpha, beta, and gamma) potentially responsible for this effect.6 6Radiolysis (i.e., energy deposition in the salt due to energetic alpha, beta, or gamma radiation) is the source of the energy needed to break chemical bonds, resulting in the release of fluorine and UFO. However, the particular radioactive species (e.g., alpha particles from actinides, beta particles from fission products, or gamma rays from gamma emitters such as those in the 232U decay chain) that caused the most radiolytic
RADIOLYSIS AND NUCLEAR REACTIONS 29 This was observed in 1964 in irradiation experiments using surrogate salt samples. At that time no uranium was found in the gas phase. With the aim of reversing this radiolysis and, in large measure, deterring the escape of fluorine from the bulk salt, the irradiated salt (Guymon, 1971) was "annealed" after reactor shutdown. Annealing was carried out by maintaining the salt for a week at 149°C, a temperature well below the melting point of the salts (Peretz, 1996c). Annealing of the bulk filet salt, carried out on an annual basis until December 1989, was unsuccessful in preventing the transport of F2 and UFO from the salt. This procedure was done presumably so that the fluorine released by radiolysis could recombine with the reduced species in the salt. Instead, the net effect over time was that fluorine (perhaps as fluorine radicals or F2) oxidized UF4 to UFO. Annealing was stopped because of concern that annual heating of the fuel salts may have aided the transport of some uranium from the fuel, which had been observed at that time, probably as volatile uranium hexafluoride. Presently, pressure sensors at different locations in the piping system indicate different overpressures, showing that the piping is plugged in several places. In recent years, gamma radiation was detected to confirm the migration of uranium within the piping system. The high-energy gamma from the thallium-208 ~ TI) daughter of U was used to trace the movement of UFO through the piping. In retrospect, the difference decomposition is uncertain. It can be argued that alpha particle decays are partially responsible for fluorine radiolysis. Compared to gamma radiation, alpha particles have a shorter range and a denser ionization track, considerations that are important because individual fluorine atoms must be produced in close proximity to combine to form F2. Another argument in favor of alpha (over gamma) radiolysis of UFO to lesser fluorides is that the uranium nucleus provides the alpha particle source, which is therefore always in close proximity to uranium fluoride compounds. These considerations are reflected in quantitative data. Trowbridge et al. (1995, Table 2, p. 19) summarize fluoride radiolysis, focusing solely on uranium fluorides and Molten Salt Reactor Experiment (MSRE) salts. They show G values (molecules of fluorine produced per 100 eV of absorbed radiation) as being highest for alpha radiation energy expended in solid UFO (G = 1.5~. For x-ray and gamma energies expended in solid fluorides to yield fluorine, the G values are much lower, of the order of 0.02 with large uncertainties. One experiment, cited without details, provided a G of zero for the effects of 238Pu alpha radiation on MSRE salts. However, since the observed yields of F2 and UFO were both extraordinarily greater than anticipated, other mechanisms are clearly at play; nevertheless, the effects of alpha radiation will have major consequences particularly on a condensed UFO phase.
30 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE between the simulated test (showing a benefit from annealing) and the real salt system (in which annealing may have aided migration of uranium away from the drain tanks) was probably the existence of a cold chemical trap in the Molten Salt Reactor Experiment (MSRE) system that the simulated test did not contain. An alternative explanation is that tests on the simulated salt using reactor radiation did not adequately mode! the enhanced fluorine generation from alpha (over gamma) radiolysis. In March 1994, a gas sample was taken from an off-gas system that was connected to the vapor space existing over the solid fee} salts in the drain tanks. The analysis (Peretz, 1996c, Table 1.1) showed that fluorine gas was presents at a pressure of 350 mm of mercury. Furthermore, the gas sample had the relatively high UFO gas pressure of 69 mm at 21°C, which is significant because it is only a few millimeters lower than the vapor pressure of UFO gas in equilibrium with its solid phase at that temperature. This may represent a major concern because, where UFO gas is in true equilibrium with its solid phase, the pressure would not provide a measure of the amount of uranium present. Thus, other methods of measurement would be required to establish the amount and location of UFO as a solid and of any of the three lower fluorides of uranium (i.e., UFs, UF4, and UF3) that do not contribute importantly to the vapor pressure. Pressure-volume relationships (Peretz, 1996c, p.l-12) established that a minim am of I.8 kg of uranium exists as vapor in the vessels and piping. Further radiation and thermal measurements established that 2.6 kg of uranium had been transported from the fuel salts and was deposited in the upper 1-foot section of the ACE located outside the main reactor building. Both fluorine and uranium hexafluoride react with charcoal, forming CFX and nonvolatile uranium fluorides. The charcoal provided a "sink" (by greatly lowering both the F2 ant! the UFO pressures) to collect the uranium as UF4 or UF3 (uranium trifluoride). The ACB currently is isolated from any source of additional UFO. Planning and process design are in progress at Oak Ridge National Laboratory (ORNL) to remove the charcoal fraction containing the uranium deposit. Relief of the different ove~pressures observed in the vent piping (currently greater than the one 7As reported to the panel by Rushton et al. (1996b), the pressure over the drain tank fuel salts had increased to about two atmospheres, as measured on October 3, 1996.
RADIOLYSIS AND NUCLEAR REACTIONS 31 atmosphere pressure found in 1994) is being addressed by work in progress at ORNL. Once solid UFO is deposited on surfaces throughout the apparatus, there is no driving force for its migration, provided all surfaces are at the same temperature. Temperature differences (currently several Celsius degrees) between the interior of the salts and the upper tank shell that caps the vapor space could provide a mechanism for buildup of solid UFO. Even though only small temperature differences result from energy deposited in the salt by radiation, time favors transport. Radiation Decomposition of Solid Uranium Hexafluoride The reverse of radiation-induced formation of UFO from the fuel salts also can be a problem. In deposits of solid UFO, decomposition by its alpha radiation from the uranium must be considered. Whereas this effect is negligible in the case of depleted uranium and minimal in the case of 235U, it is enhanced greatly in the current situation where 233U and 232U (with its decay chain of six alpha-emitting daughters) are involved (see Figure 2.1~. Although 232U is present to only 160 parts per million (ppm) in the uranium, radiation from 232U (and its daughters) exceeds that from 233U (see Table 2.3), which is the major uranium isotope (84 weight percent). The data in Peretz (1996c) indicate about ~ Ci of alpha activity per mole of uranium. The neutrons produced by (a,n) reactions on fluorine and beryllium can be measured to show uranium deposition. Perhaps of more importance from a process aspect, a deposit of solid UFO that has migrated from the salt and has aged sufficiently for the 232U daughter chain to be reestablished, would have released in its lattice the energy from 3 x 10~ alpha particles per second per mole. If 6 MeV is used as the average alpha energy, in 100 days about 25 eV per molecule of UFO will have been deposited. This energy far exceeds the chemical bond strengths of UFO. Trowbridge et al. (1995) cite a value of 1.5 molecules of F2 produced for Here again, the alternative explanation of alpha radiolysis from 232U daughters can be invoked to explain transport. Reduction of a solid UFO deposit to lower fluoride salts (UFs, UF4, and UF3), if it were to lower the vapor pressure, would provide a driving force for transport and a mechanism to build deposits that cause eventual plugging.
32 U232 OR ~' U232 99% - a 1% - SF Th228 1 00% - a Ra224 1 00% - a Rn220 1 00% - a Po216 100% - a Pb212 1 00% - p- Bi212 OR ~' Bi212 64.07% - p~ 35.93% Po212 100% -a AN EVALUATION OF DOE ALTERNATIVES FOR MSRE U233 OR ~' U233 99.99% - a 1 e-003% - SF Th229 100% - a Ra225 1 00% - p- Ac225 100% - a Fr221 100% -a At217 OR 99% - a Bi213 OR ~' Bi213 97.84% - ,B- 2.16% - a Ti208 100% - p~ Pb208 Pb209 1 00% - p- Bi209 At217 1 % - p- Rn217 100% - a Po213 Ti209 Po213 100% - a 100% - ,8~ 100% - a Pb209 1 00% - p- Bi209 Pb209 1 00°~ p- Bi209 FIGURE 2.1 Decay chains for 232U and 233U. NOTE: SF = spontaneous fission. During reactor operations, 233U(n,2n) reactions produced small quantities of 232U, which decays by successive alpha and beta particle emissions to 208T1, an emitter of a penetrating 2.6-MeV gamma ray. SOURCE: Peretz (1996c, Figure 1.12).
RADIOLYSIS AND NUCLEAR REACTIONS 33 100 eV deposited in solid UFO. Thus, in spite of inefficiencies of bond breaking and the occurrence of back-reactions, substantial radiolytic decomposition of the deposited UFO will be observed in aged deposits. This decomposition results in deposits of nonvolatile uranium fluorides (e.g., UF4 and UF3) that cannot be removed readily from unheated pipes and vessel surfaces by use of room-temperature fluorine in an inert purge gas. in piping or equipment where heating is difficult or impossible, the use of fluorinating agents such as krypton difluoride (KrF2), bromine pentafluoride (BrFs), or chlorine trifluoride (CIF3) (see Appendix B) to refluorinate the lower fluoride salts may prove useful, if the reaction with fluorine at room temperature does not produce UFO. if oxygen has been introduced through minor leaks back- streaming from the vent line, or through oxide coatings or adsorbed moisture, urany! difluoride (UO2F2) may be formed, which- along with uranium in reduced oxidation statescan be converted to UFO upon refluorination, as with KrF2 or BrFs at room temperature (see Chapter 3 and Appendix B). Long-Term Effects of Leaving Plutonium in the Salt After Uranium Removal Plutonium migration in the MSRE system has not received significant attention because the formation and migration of quantities of UFO dominate nuclear criticality safety implications. Prior to migration from the salt, the initial uranium concentration was about 7100 ppm, far exceeding the plutonium concentration of 155 ppm. Indeed, competition leading to the formation of UFO would have suppressed formation of transient PuF6, which is known to be reduced by the lower fluorides of uranium. In anticipation of uranium removal, it is prudent to anticipate effects of radiation on the plutonium remaining in the fluoride salts. Radiation effects are known to disturb the normal PuF6 ~ PuF4 + F2 equilibrium with two opposing results. One is that radiation is known to decompose PuF6. Conversely, radiation effects on F2 gas in contact with PuF4 at ambient temperature will lead to PuF6 in concentrations exceeding those calculated from the equilibrium constant (Katz and Seaborg, 1986, and extensive references cited therein).
34 AN EVALUATION OF DOE ALTERNATIVES FOR MERE It is likely that radiation effects on the plutonium remaining in the fluoride fuel salts will lead eventually to formation of some PuF6 and its migration from the stored salt. Criticality will not occur because of the limited amount of plutonium in any one container. The getters provided for fluorine also will react with any PuF6, which would be an alpha particle hazard (and not a criticality hazard) at low plutonium concentrations. Cautious handling of the fluorine getter material on subsequent opening of the interim storage vessels is warranted. EXCESS OF REDUCING SPECIES IN THE SALT AND HAZARDS OF SIMPLE REMELTING The loss of F2 and UFO has left a reservoir of reducing agents distributed throughout the mass of fuel salt that can have some chemical effects on salt remelt. Whether these reducing agents are isolated metal atoms as described in Peretz (1996c, p. I-30) or electrons located in fluoride lattice vacancies is less important than their potential effects. The current estimate (Rushton et al., 1996a) is that 290 equivalents of excess reductant have been produced in the fuel salt. In reacting with 135 moles of total uranium in the two fuel drain tanks, all of the uranium initially present as UF4 would be reduced to the less soluble UF3, precipitation of which is unlikely at a uranium concentration of approximately 0.12 mole percent.9 Further reaction of the 155 remaining unsatisfied equivalents of chemical reducing agents could yield 12 kg of uranium metal. Melting of the salts without eliminating this excess of reducing equivalents could result in formation of some uranium metal, although it is not clear that uranium metal would be formed in preference to zirconium metal or possibly the zirconium fluoride ZrF3. 9This number can be calculated two ways. One way uses 4650 kg of salt and 39.5 as the quasi molecular weight of the mixture, derived from the mole percent of the constituents, to calculate 118,000 moles of salt and 142 moles of uranium. The ratio of these mole numbers is 0.0012, or 0.12 mole percent. The second calculation uses data from Hollenbach and Hopper (1994, Table 1) to deduce 118,400 moles of salt, of which 233UF4, at 309 g/mole, would constitute 150 moles. The ratio of these mole numbers is 0.0013, or approximately 0.13 mole percent. This uranium concentration is a factor often higher than the value quoted in Liebenthal et al. (19944.
RADIOL YSIS AND NUCLEAR REA CTIONS 35 The conclusion that there may be significant amounts of metallic uranium present in the salt (Peretz, 1996c, p. 1-30) is based on the assumption that all reducing potential available as a by-product from fluorine generation results in preferential uranium reduction. The panel does not concur with the use of such an extreme assumption. The fact that zirconium fluoride reduces before uranium fluoride is well established and is one of the reasons that zirconium fluoride is present in the melt. The melt composition is almost certainly different than at reactor discharge and may contain metallic zirconium, but even if single atoms of uranium metal were formed locally in the solid phase, there does not appear to be any mechanism that could cause this material to agglomerate into a dense metal phase. Surrogate samples of fuel salt were irradiated (in 1963) in the Materials Test Reactor (MTR) in Idaho to a fluoride removal of 2.1 percent (Peretz, 1996c, p. 3-4~. This value is to be compared with the current (minimum) estimate of 0.13 percent fluoride removal from the fuel salts. Only about half of the salt was recovered on remelt of the irradiated sample, leaving a high-melting residue that could not be removed by draining. Recent tests (Peretz, 1996c, p. 3-6; Williams et al., 1996) of irradiated surrogate fuel salts indicate similar behavior because insoluble, metallic-appearing precipitates are observed on remelt. These tests involve the melting behavior of surrogate salts when irradiated to a fluorine deficiency of 0.l percent, more closely resembling the 0.13 percent fluorine loss estimated for MSRE fuel salt. In these experiments (Williams et al., 1996, p. 12), the gray-black irradiated salt melted to form an opaque pool. With these results, there is currently no . , ~ ~ ~ experimental knowledge of the rate at which chemical treatment (e.g., fluorination) will restore the melted irradiated salt to the homogeneous condition of the melted unirradiated salt. As of this writing, further information is desirable on the composition of the nonmeiting precipitate. In summary, nuclear decays of radioactive constituents of the salt have resulted in radiolytic formation of F2 and UFO gases, which have migrated out of the solid salt. This migration implies a loss of fluorine and formation of associated reducing agents in the salt. These conditions could lead to both chemical and physical problems accompanying a simple remelt, since the phases that result from melting of a reduced salt mass could cause line plugging and prevent the transfer of molten salt. Based on considerable past experience in this area, it would be desirable
36 AN EVALUATION OF DOE ALTERNATIVES FOR MSRE to obtain a homogeneous melt that is likely to be handled successfully by routine, we11-established procedures. The goal would be to restore the salt to its previous process condition, in which melting was straightforward and the desired oxidation states in the melted salts were maintained readily. Cautious approaches to attaining this condition were suggested (Peretz, 1996c) and are discussed in greater detail in Chapters 4 through 6, after further elaboration of salt chemistry in Chapter 3.