The hot-wire process was developed by Van Arkel and de Boer [V2], who used it to produce the first pure, massive specimens of many refractory metals, notably titanium, zirconium, hafnium, and thorium. An interesting account of early uses of this process is given in

Figure 7.14 Hot-wire reactor for zirconium produc­tion.

Van Arkel’s book Reine Metalle [VI]. This method was used by the Foote Mineral Company and the Westinghouse Electric Corporation to produce the first zirconium pure enough for nuclear reactors.

The apparatus used by Foote is shown in Fig. 7.14. It consists of an Inconel tube through which are led insulated tungsten leads capable of carrying a heavy electric current. Inside the tube, the leads are connected to a thin tungsten wire. The tube is charged with crude zirconium and evacuated, and a few grams of iodine are distilled into it. The tube is heated to a temperature at which iodine reacts with the zirconium and at which the iodide produced has a vapor pressure of several torn. The tungsten wire is heated electrically to a temperature high enough to dissociate the iodide, but below that at which the metal melts or has a substantial vapor pressure. Tetraiodide, formed from the crude metal, diffuses through the iodine vapor and deposits pure metal on the tungsten wire. As the latter increases in cross section, the electric current through it is increased to keep it above the dissociation temperature of the iodide. The run is concluded when the tungsten leads are carrying the maximum possible current.

In this way rods, or “crystal bars,” of compact, ductile zirconium or hafnium have been prepared. The usual crystal bar is 0.25 to 0.4 in in diameter in lengths up to 2 ft, but Westinghouse and Battelle Memorial Institute have produced zirconium bars as large as 1.7 in in diameter and 50 ft overall length [HI].

The hot-wire process eliminates oxygen, nitrogen, and carbon, the impurities most difficult to keep out of zirconium in other processes, but other metals that form volatile iodides are not removed completely. The main disadvantage of the process is its low capacity, the rate of production being limited by the rate of diffusion of iodide vapor to the small wire. Temperatures used for producing metals of the IVA group by the hot-wire process are listed in Table 7.13.

Eighty-five percent of the zirconium used in the first land-based prototype of a submarine reactor was made by the hot-wire process. In 1952, the hot-wire process for zirconium was superseded by the lower-cost Kroll process. However, the hot-wire process is still used to produce hafnium for control rods in U. S. naval reactors.

The phases of metallic neptunium, and their densities and transition temperatures, are listed in Table 9.12.

Metallic neptunium is prepared by reducing NpF4 with calcium. Neptunium yields of about 99 percent have been obtained from 100- to 400-g quantities of NpF4, with 30 percent excess calcium and with 0.25 to 035 mol of iodine booster per mole of NpF4. Metallic neptunium forms a protective oxide layer in air at room temperature, but it rapidly oxidizes at higher temperatures. It dissolves readily in HC1 and H2 S04 [K2].

Off-gases from decladding, voloxidation if practiced, and dissolution are passed through high-efficiency particulate filters, processed for radioiodine absorption and, in some plants, for krypton and xenon retention before discharge through the plant stack. Gases vented from downstream process equipment are also passed through high-efficiency particulate filters and radioiodine absorbers. This section describes briefly processes that have been developed for absorbing radioiodine and removing and packaging krypton and xenon. Retention of tritium and 14C may also be required in the future.

Radioiodine removal. Radioiodine removal is important because of its toxicity, the compara­tively high iodine content of fission products (0.69 w/o, Table 8.7), and the high fission yields at the mass numbers of the two principal radioiodines, 1.7 X 107 year 129I (1 percent) and 8.05-day 1311 (2.09 percent). Removal of radioiodine is complicated because of the numerous process streams in which iodine may appear and the variety of chemical forms it assumes. About 1 percent of the iodine is volatilized during decladding, some during voloxidation and a significant but incomplete amount during dissolution. If iodine is allowed to remain in the feed to solvent extraction, it reacts with solvent to form hard-to-remove compounds that eventually contaminate the entire system. It is thus important to remove as much of the iodine as possible before solvent contacting. Iodine may appear as I2, HI, HIO, or organic iodides in off-gases or aqueous or organic phases, or as HI03 in concentrated nitric acid solutions.

The preferred procedure for removing iodine is to route the gases from decladding and voloxidation to iodine absorbers and to distill iodine from dissolver solution before solvent extraction. Experiments at Oak Ridge showed that 95 percent of the iodine could be removed by distilling 2 percent of the volume from 4 M HN03 and 99 percent by distilling 20 percent [06]. Some of the remaining iodine is evolved with vent gases.

Of the numerous iodine-removal methods discussed by Goode and Clinton [G9], the most significant are characterized below.

Absorption by aqueous NaOH removes HI and I2 but not organic iodides. No good procedure is available for disposal of spent solution.

The Iodox process under development by Oak Ridge National Laboratory uses absorption in boiling 21 to 23 A/HNO3 to convert iodine and its compounds to solid, nonvolatile I2Os.

An alternative process developed by Oak Ridge [08] uses boiling 8 to 14 M HNO3 containing 0.2 to 0.4 M Hg(N03)2 to absorb all forms of iodine as Hgl2. The absorber solution is evaporated from vermiculite, which retains the iodine in stable form suitable for storage. The Barnwell plant proposes [A3] use of a similar process.

Unglazed Berl saddles coated with silver nitrate and operated at 135°C were used at Hanford [Ml] to remove HI and I2 from dissolver off-gases. In 1958, an explosion occurred, which was attributed to an unstable compound of silver and ammonia formed when the reactor was periodically cleaned by washing with ammonium sulfite. After this was replaced by sodium thiosulfate, the reactor operated for 14 years without incident.

A more effective way of using silver is to impregnate with silver a zeolite catalyst of the type used in hydrocarbon processing. With moist air at 150°C all volatile iodine species are absorbed as stable silver iodide in a form suitable for packaging and permanent storage. Silver zeolites for iodine absorption have been developed at Idaho Nuclear [P3] and Karlsruhe, Germany [W7], Wilhelm et al. [W7] give data for fractional penetration of I2 and CH3I through an amorphous silicic acid zeolite impregnated with 0.06 to 0.08 g silver/g zeolite. More than 98 percent of the silver is available for reaction, permitting loadings of 0.1 g iodine/g zeolite. Fractional penetration of iodine is a function of many variables, as described in [W7]. Decontamination factors of from 102 to 104 have been reported. Long-term management of radioiodine as a radioactive waste is discussed in Chap. 11.

Krypton and xenon removal. The number of curies of krypton and xenon per megagram (metric ton) of spent fuel from pressurized-water, liquid-metal fast-breeder, and high — temperature gas-cooled reactors from Tables 8.7, 8.8, and 8.9 are listed in Table 10.5, together with the number of standard liters per megagram, assuming atomic weights of 85 and 133 for krypton and xenon.

Table 10.5 Curies and liters of krypton and xenon in spent fuel 150 days after discharge

The volume of gas is appreciable; 80 percent or more is xenon. Practically all of the radioactivity is due to 85 Kr. One year after discharge the xenon activity would be negligible. This xenon could be a significant commercial source.

Processes that have been studied for krypton-xenon removal are listed in Table 10.6 together with comments on the process from reference [M6]. All have achieved 99 percent krypton removal.

Room-temperature adsorption is used for off-gases from nuclear power plants to delay escape of krypton and xenon long enough for all radionuclides except 85 Kr to decay to innocuous levels. Retention of 85Kr would require very large bed volumes and a more complex system for bed regeneration. There is a fire hazard when treating reprocessing off-gases with charcoal, so that 02 and NO* must be removed from the feed.

In cryogenic adsorption, smaller bed volumes suffice, but the feed must be pretreated to remove condensibles. The fire hazard with charcoal remains and may be worse, because of the possibility of adsorption of ozone produced by radiolysis of oxygen.

Development of permselective membranes is only at the laboratory stage. For reprocessing

Table 10.6 Processes for removal of 85 Kr from reprocessing off-gas

Process Development status Comments

off-gases, disadvantages are the serious consequence of mechanical failure and deterioration from radiation and exposure to ozone and NO*.

The last two processes are the ones favored for reprocessing plants.

Cryogenic distillation has been extensively operated at Harwell [W8] and the Idaho Chemical Processing Plant [В9]. The principal concerns are (1) plugging of low-temperature equipment by condensed ice, solid C02 or Xe, or solid nitrogen oxides and (2) possible explosion from accumulation of solid hydrocarbons in the presence of condensed oxygen and ozone. To deal with them, feed gases must be pretreated for removal of impurities before condensing the krypton and xenon. At the Idaho plant [B9], N02 and C02 were removed from feed gas by scrubbing with sodium hydroxide solution. No attempt was made to package the C02 containing 14 C, but this could have been done by precipitation as CaC03 with lime. N20 was dissociated into N2 and 02 by passage over a rhodium catalyst at 650°C. The hydrocarbon content of feed gas was low enough that hazardous accumulation in low-temperature equipment was prevented by warm-up once per shift. More generally applicable practice would be to oxidize hydrocarbons by passing feed gas over copper oxide at 600°C.

After purification the feed gas at Idaho was cooled to -160°C by passage through regenerators, precooled by outflowing cold gas, in which H20 and remaining traces of C02 and nitrogen oxides were condensed and removed. Finally, the purified feed gas was washed with liquid nitrogen to condense krypton and xenon, which were then concentrated by fractional distillation. The concentrate was separated periodically by batch distillation into an oxygen fraction, which was recycled to prevent loss of small amounts of accompanying krypton, and a krypton fraction and a xenon fraction, which were bottled separately for storage.

Absorption in halogenated solvents, such as refrigerant R-12, CF2C12, has been extensively studied at Brookhaven [S21], Harwell [T3], and Oak Ridge [M6, V4]. The process has several advantages. Fire or explosion hazards are minimal, and gas purification prior to absorption is not required. The process is flexible and does not use extremely low temperatures. Dis­advantages are operation at 8 to 10 bar pressure, a fairly complex flow sheet, and the need for an auxiliary system to separate krypton from xenon and C02.

Figure 10.7 shows one of the flow sheets for removing radioisotopes from reprocessing plant off-gases by refrigerated absorption tested by Oak Ridge National Laboratory [V4]. Contaminated feed gas, consisting of H20, C02, N20, Xe, Kr, Ar, N2, and 02, and possibly containing I2, CHjI, and N02 not previously removed, is compressed to 8 bar (100 psig) and cooled to —28°C in a cold trap. This removes most of the H20, N02, I2, and iodine compounds. The gas is then fed to a 5-m absorber-fractionator column refluxed with refrigerant R-12 at —28°C at the top and reboiled at 31°C at the bottom. Decontaminated Ar, N2, and 02 are taken off the top, and a solution of Kr, Xe, N20, and C02 in R-12 is taken off the bottom. This solution also contains traces of Ar, N2, 02, N02, and water, and iodine compounds if present. In the stripper, Kr, Xe, N20, and C02 diluted with small amounts of Ar, N2, and 02 are taken off as overhead product, together with some R-12. Solvent from the bottom of the stripper is distilled to separate it from small amounts of water and other less volatile impurities prior to recycle to the absorber-fractionator.

Overhead product from the stripper, although greatly reduced in volume from feed gas, requires further treatment (not shown) to separate and package C02 containing 14 C, Kr, and Xe. One possible sequence of operations would be

1. Absorb C02 for permanent storage on solid soda lime.

2. Remove R-12 for recycle with a selective molecular sieve.

3. Decompose N20 over a rhodium catalyst at 650°C.

4. Remove 02 with copper at 600° C.

5. Condense Xe, Kr, and some Ar in a cold trap refrigerated with liquid nitrogen.

6. Separate the condensate by low-temperature distillation into (a) an Ar-Kr fraction to be bottled for permanent storage as radioactive waste and (b) an Xe fraction.

Special campaigns for recovering neptunium from Purex solutions have been run at Oak Ridge [F4], Hanford [D3], Savannah River [P7], Windscale [Nl], and Marcoule [С6]. None of these sought complete recovery. A brief description will be given of the first three.

Oak Ridge [F4]. The solution obtained by dissolving irradiated, natural uranium in nitric acid was treated with 0.01 M NaN02 to convert most of the neptunium to extractable Np(VI) in 2 M HN03. In the first Purex cycle, 90 percent of the neptunium was extracted with the uranium and plutonium. Ferrous sulfamate used in the partitioning step reduced most of the neptunium to Np(V) and Np(IV), which followed Pu(III) into the aqueous phase, but some Np(IV) remained with the uranium. When the aqueous phase containing Pu(III) and most of the neptunium was reoxidized with HNO3 and NaN02 and extracted with TBP in the second cycle, from one-half to two-thirds of the neptunium was recovered with the plutonium, from which the neptunium was separated by anion exchange. The process produced 99.9 percent pure neptunium, but the recovery was incomplete and very sensitive to nitrite and nitric acid concentrations.

Hanford [D3]. Nitrite concentration in feed to the HA column of a standard Purex plant was adjusted to route most of the neptunium in irradiated natural uranium into the extract from the HS scrubbing column. Sufficient ferrous sulfamate was used in the partitioning column to reduce neptunium to Np(IV), which followed uranium. This neptunium was separated from uranium by fractional extraction with TBP in the second uranium cycle. The dilute neptunium product was recycled to HA column feed, to build up its concentration. Periodically, irradiated uranium feed was replaced by unirradiated uranium, which flushed plutonium and fission products from the system. The impure neptunium remaining was concentrated and purified by solvent extraction and ion exchange.

Savannah River [Р7]. At the Savannah River Purex plant, neptunium in irradiated natural uranium was recovered by the alternative method of forcing most of it into the aqueous waste stream HAW from the first extraction cycle and then recovering it from waste directly by anion exchange. Neptunium in the first extraction step was converted mostly to the inextractable pentavalent state by adding sufficient nitrite to the next-to-the-last mixer-settler stage of the HA section to make the solvent 0.007 M in HNOj.

A simple way to maintain interfacial area and dispersion in a vertical gravity-flow column, and to reduce axial mixing, is to fill the column with loose packing to provide tortuous flow paths. Typical packing consists of ceramic rings or saddle shapes, dumped in random arrangement. The

^ Figure 4.29 The spray column of Elgin [Е1]. (From Treybal [T2], by

liquid permission.)

packing reduces the available space for liquid flow and also introduces frictional drag, so the liquid throughout per unit of cross-sectional area is less than for spray columns. Neutron poisons can be incorporated into the packing to increase the criticality safe diameter.

Spray column contactors were used in the first large-scale solvent extraction plants at Hanford, Washington, for recovering plutonium from irradiated natural uranium and in the first chemical processing plant at Idaho for recovering enriched 235 U [L2].

Although packed columns are simple and have no moving parts, their large space requirements have resulted in the replacement of packed columns by pulsed columns, or by other more compact contactors, in more recent installations for reprocessing irradiated reactor fuel.

As received by the uranium refinery, uranium ore concentrates now usually consist of uranium oxide or sodium, magnesium, or ammonium diuranate. These concentrates still contain appreciable amounts of elements other than uranium and some of uranium’s radioactive decay products present in the original uranium ore, such as radium and radon.

The first step in the conventional process for refining uranium is dissolution in nitric acid. When the concentrates have been produced by chemical leaching and are in the form of diuranates, dissolution proceeds rapidly and leaves little solid residue. When the concentrates have been separated mechanically and are in the form of the original uranium mineral, dissolution may require more concentrated acid, higher temperatures, longer times, and addition of oxidants such as Mn02. Also, filtration to remove undissolved residues is usually required. In either case, dissolution produces an aqueous solution of uranyl nitrate hexahydrate U02(N03)2’бНгО, containing excess nitric acid and variable amounts of nitrates of metallic impurities present in the concentrates.

The next step in purification is separation of uranyl nitrate from the other metallic impurities in the dissolver solution by solvent extraction. Practically all uranium refineries now use as solvent tributyl phosphate (TBP) dissolved in an inert hydrocarbon diluent. The first U. S. refinery used diethyl ether as solvent and later refineries have used methyl isobutyl ketone or organic amines, but practically all have now adopted TBP. It is nonvolatile, chemically stable, selective for uranium, and has a uranium distribution coefficient greater than unity when the aqueous phase contains nitric acid or inorganic nitrates.

Although uranium refineries use widely different types of solvent extraction contactors, their basic process flow sheets are similar, along the lines of Fig. 5.22, which illustrates the

specific process developed by the Mallinckrodt Chemical Company for the Weldon Springs refinery. A similar flow sheet is used in the Kerr-McGee plant.

Uranium ore concentrates are digested with hot 40% nitric acid. The resulting mixture is about 1 N in nitric acid and contains about 400 g uranium/liter and some suspended solids. The aqueous mixture is fed to a series of pumper-decanter mixer-settlers, where the uranyl nitrate is extracted by countercurrent flow of 30 v/o TBP in normal hexane. The flow ratio of organic to aqueous is about 13:1. Uranium concentration in the organic extract leaving the first stage is about 95 g uranium/liter and in the aqueous raffinate leaving the last stage is under 0.1 g uranium/liter. The raffinate is neutralized with lime. It contains most of the radioactive impurities in the ore concentrates, principally 230Th and 226 Ra.

In the scrubbing section all nonuranium metallic impurities and some uranium are removed from the organic phase by counterflowing dilute nitric acid, which is returned to the extracting section. In the stripping section purified uranium in the organic phase leaving the scrubbing

Makeup Nitric

HN03 acid

Recycle

HN03

Hj

Anhydrous HF

Figure 5.21 Steps in conventional uranium refining processes.

section is transferred to an aqueous phase by back-extraction with 0.01 normal nitric acid. Pulse columns are used for the scrubbing and stripping sections.

A portion of the aqueous stream leaving the stripping section is withdrawn, washed with hexane to remove dissolved and entrained TBP, and leaves the TBP-removal column as product uranyl nitrate solution (UNH).

All TBP is washed with an aqueous solution of sodium carbonate in a spray column to remove any hydrolyzed TBP and impurities that might accumulate in the TBP if it were not cleaned in this way. Sodium hydroxide is added to the aqueous sodium carbonate stream leaving the spray column to precipitate any uranium that might have been carried to this point. This impure uranium is recycled to the dissolver.

Variants of this basic process are used in other plants. For example, the Comurhex plant at Malvesi [B5] filters the output from the dissolver, uses pulse columns in the extracting section, and dilutes TBP with n-dodecane instead of л-hexane.

The principal uses of ThF4 are as intermediate in the production of thorium metal or, potentially, as a compound in the fuel mixture of the molten-salt breeder reactor. For both applications anhydrous, oxide-free ThF4 is required.

Such ThF4 cannot be made by precipitation from aqueous solution, as the precipitate contains water that, during heating or evaporation, hydrolyzes some ThF4 to Th02 or ThOF2. Instead, ThF4 is produced by gas-phase hydrofluorination of Th02 with anhydrous HF:

Th02 + 4HF ThF4 + 2H20

This reaction is exothermic and proceeds rapidly at 566°C, but the equilibrium gas mixture contains some unreacted HF (Prob. 6.3). At lower temperatures nearly complete utilization of HF can be obtained, but the reaction is slow. The process was developed at Iowa State College

Table 6.21 Principal processes for producing metallic thorium

Electrolysis of fused salts Electrolysis of KThFs in NaCl Electrolysis of ThF4 in NaCl/KCl Electrolysis of ThC^ in NaCl/KCl Reduction with Reactive Metals Reduction of Th02 with Ca Reduction of ThCU with Mg Reduction of ThF4 with Ca Thermal Dissociation of Thl4

and used industrially for the U. S. AEC by the National Lead Company at Femald, Ohio. A summary of the process, described in detail by Cuthbert [C6], pp. 152-154, follows.

Equipment consists of four externally heated, screw-fed, horizontal reactors positioned vertically one above another. The reactors are made of 309 Nb stainless steel, and the screw of Inconel and Illium R. Solids flow through the four reactors in series. In the first reactor, counterflowing air at 650 to 675°C removes residual H20 and C02. Anhydrous HF vapor enters the fourth reactor4 and flows counter to the solids through the fourth reactor held at 566°C, the third at 370°C, and the second at 260°C. In this way, Th02 can be converted completely to ThF4, the highly exothermic reaction can be controlled and most of the HF can be reacted. However, the process was usually operated to produce 70 w/o aqueous hydrofluoric acid, which was sold as a by-product.

10.2 Production of ТЪСЦ

Several of the processes for producing thorium metal start with anhydrous ThCL,. As with ThF4, anhydrous ThCL, cannot be prepared from aqueous solution, but must be made by gas-phase chlorination. A process used in England [B3] involved chlorinating a mixture of Th02 and carbon at or above 600°C:

Th02 + 2C + 2C12 ->• ThCL, + 2CO

The chloride must be purified by distillation to free it from unreacted solids and from impurities in the carbon. This is difficult because of the hygroscopicity of ThCl4 and its high boiling point, 942°C. An alternative process [C6] reacts thorium oxalate with an excess of carbon tetrachloride and a small amount of chlorine as catalyst,

Th(C204)2 + ССЦ -* ThCl4 + 2CO + 3C02

batchwise in a vertical graphite reactor at 600°C. This uses more expensive materials but produces pure solid ThCl4 in a single step.

The principal actinides involved in using thorium-uranium fuel are shown in the actinide chains of Fig. 8.11. The important reactions are the fission of 233U and 235U and the absorption of neutrons in 232 Th to form 233 U.

The relatively long 27.0-day half-life of 233Pa, the precursor of 233U, affects the time that irradiated fuel must be stored prior to reprocessing. If the discharged fuel is stored only for 150

Figure 8.11 Actinide chains in thorium fuel.

days, as is frequently specified for sufficient decay of 1311, some of the 233 Pa will remain during reprocessing. Protactinium is one of the more difficult elements to separate from uranium, and the high radioactivity of protactinium will contribute to the problem of decontaminating the uranium product after it is separated from the fission products and thorium. Also, if protactinium is not recovered, the loss of undecayed 233Pa will represent some loss in the production of 233U for recycle.

Another problem of the thorium fuel cycle results from the radioactivity of 72-year 232 U, and its daughters. 232U is formed by (n, 2n) reaction with 232Th according to

232 Th 231 Th 231 Pa M2Pa __J1————— > 232 и (8.13)

25.5 h 1.31 days v ‘

and by

233 jj n’2n, 232 у

Also, many thorium ores as well as thorium, which is obtained as a by-product of uranium mining, contain traces of 230Th, a radionuclide in the decay chain of 238U. Neutron absorption in 230Th also results in the formation of 232 U:

Although significant alpha activity results from 232 U in the 233 U to be recovered and recycled, more of a problem results from the 232U daughters. The 232U decay daughter is 1.91-year 228Th, a radionuclide that is also formed by the radioactive decay of 232Th. As shown in Table 6.3, the decay daughters of 228Th are all short-lived, so they reach secular equilibrium with 228Th after a delay time of only a few days. The decays of 212Bi and 208T1 are accompanied by very energetic and penetrating gammas, so gamma shielding is required when fabricating fuel from recycled uranium containing 232U.

Although chemical reprocessing yields essentially pure uranium, storage after separation and time elapsed in shipping to fabrication allow the buildup of 228Th and its decay daughters. Consequently, the gamma activity in separated uranium containing 232 U increases continuously with storage time, until it reaches a maximum at about 10 years after separation. Once uranium has been separated from thorium, there is considerable incentive to complete the uranium purification and fuel fabrication quickly to avoid the increasing gamma radiation due to the buildup of 228Th. Hydrogenous shielding is also necessary because of the high-energy neutrons from alpha decay in recycled uranium. The alphas from the decay of 233U, 232U, and 228Th interact with light elements such as oxygen and carbon to form neutrons, so the neutron activity also increases with storage time.

The 228Th and 234Th appearing with irradiated thorium fuel results in appreciable radioactivity in the separated thorium. Consequently, as discussed in Sec. 2.9, it may not be practicable to recycle the recovered thorium until it has been stored for about 5 to 20 years.

When 235 U is used as fissile makeup in the thorium cycle, as in the reference high-temperature gas-cooled reactor (HTGR) fuel cycle, the high bumup and uranium recycle result in considerable production of “’’Np, according to the reactions shown in Fig. 8.11. The 237Np then forms a relatively large activity of 238Pu. These plutonium activities are important because of the problems of decontaminating uranium from plutonium when reprocessing the uranium. Also, even though fissile plutonium is formed by neutron absorption in the 238 U
accompanying the highly enriched 23SU makeup, the high activities of 238Pu may discourage the utilization of the fuel value of plutonium in the discharge fuel.

Relatively little 239Pu, 240 Pu, 241 Pu, americium and curium are formed in the irradiation of thorium-uranium fuel with 235 U fissile makeup. However, when plutonium is used as fissile makeup for a thorium fuel cycle, considerable quantities of americium and curium are formed. As discussed in Sec. 2.4, these are the radionuclides that are the greatest contributors to radioactivity and ingestion toxicity after about 600 years of waste isolation, when the fission products have decayed.

Material quantities and activities of the actinides calculated [HI, P3] in the cooled discharge fuel from the uranium-thorium-fueled HTGR (cf. Fig. 3.33) are listed in Table 8.6. The natural thorium is assumed to contain 100 ppm 230Th, so the quantities of 228Th and 232U in the discharge fuel are greater than would occur for thorium consisting of pure 232 Th. The strongest actinide beta source is 233Pa, which contributes 7.58 X 106 Сі/year after 150 days of cooling. In the uranium, which is to be recovered and fabricated into recycle fuel, the main contributors to alpha activity are 232 U and 233 U. Both are important as potential environmental contaminants, but the activity of the 232U daughters, which grow into separated uranium prior to fabrication, dictate the requirements for semiremote and remote fabrication. By comparison with the data in Table 8.5, the total alpha activity of 5.16 X 103 Сі/year in the uranium to be fabricated as recycle HTGR fuel is much less than the 1.70 X 10s Сі/year of alpha activity in the plutonium to be fabricated for recycle in a 1000-MWe LWR.

The total alpha activity in the plutonium in the HTGR discharge fuel is within 20 percent of the total alpha activity in plutonium from the uranium-fueled LWR (Table 8.4). In both cases the plutonium alpha activity is dominated by 238 Pu. However, the HTGR plutonium consists of 66 percent 238Pu, and the high alpha activity, the high heat generation rate, and the low fissile content mitigate against the recycle of HTGR plutonium.

Because of the relatively small amount of high-mass plutonium nuclides produced in uranium-thorium fueling, the amounts of americium and curium produced are about two orders of magnitude less than in a uranium-fueled reactor with plutonium recycle.

In aqueous solution americium exists in the four oxidation states Ат(Ш), Am(IV), Am(V), and Am(VI). In the absence of complexing agents trivalent, pentavalent, and hexavalent americium exist as Am3*, Am02+, and Am022+, usually in hydrated form. In aqueous solution tetravalent americium rapidly disproportionates, except in concentrated fluoride and phosphate solutions.

Trivalent americium is the most stable state in solution. As shown by the oxidation-reduction potentials of Table 9.6, the higher oxidation states of americium are strong oxidizing agents, so they exist only in solutions that contain no oxidizable species.

Trivalent americium forms relatively unstable complexes with Cl” and N03” and more stable complexes with the thiocyanate ion CNS”. These americium complexes are more stable than those of the corresponding lanthanide compounds, so that americium can be separated from trivalent lanthanides by anion exchange with concentrated solutions of LiCl, IiN03, or NH4CNS. Tri­valent americium can be extracted with TBP from a concentrated nitrate solution. It can also be extracted with TBP from a molten LiN03 — KN03 eutectic at 150°C, with much higher distribution coefficients than in extraction from aqueous solutions. Americium is more readily extracted by this process than is trivalent curium [K2].

Although TBP and the hydrocarbon diluent are comparatively stable compounds, they slowly react in Purex systems with formation of degradation products that impair separation performance. The principal deleterious reactions are reaction with radioiodine, hydrolysis, and radiolysis. These will be discussed in turn.

Reaction with radioiodine. Any iodine left in dissolver solutions slowly reacts with TBP and diluent to form iodine compounds that cannot be removed by subsequent alkaline washing. Thus, it is important to remove as much iodine as possible from the dissolver solution before solvent extraction and to use low-inventory contactors in the first, HA, extraction step.

Hydrolysis of TBP. Hydrolysis of TBP occurs stepwise via dibutyl and monobutyl phosphoric acid ([C4H9O] 2P02H and C4H9OPQ3H2) and leads eventually to phosphoric acid. Dibutyl phosphoric acid is the most abundant degradation product. Its rate of formation is influenced by temperature, the nitric acid concentration, the uranium content, and the presence of a diluent, beside the radiation dose. The acidic nature of these hydrolysis products allows in principle cleanup by an alkaline wash.

The effect of TBP degradation products, particularly of dibutyl phosphoric acid, is the formation of strong complexes with uranium(VI), plutonium(IV), zirconium, and niobium. The sequence of complexing strength is Zr > Pu(IV) > U(VI) > Nb.

The uranium and plutonium complexes are strong enough to remain in the organic phase during stripping. In reprocessing LWR fuel, uranium is mainly affected because of its great

excess. As a consequence, uranium is lost to the waste in the solvent wash. In LMFBR fuel, plutonium is taken up by the degradation products to a significant extent. It can be removed only by an alkaline wash with fluoride addition.

The other detrimental effect of TBP degradation is its complexing of zirconium. This increases the zirconium distribution coefficient and consequently decreases the decontamination coefficient. Moreover, solvent residual radioactivity is increased because of incomplete zirconium reextraction. Another and even more troublesome consequence of zirconium complexing is the formation of precipitates known as crud. This is a severe problem, particularly in mixer-settlers, and has led to a preference for pulsed columns or centrifugal contactors in the first extraction cycle when high-burnup fuel is to be processed.

Table 10.15 shows the effect of temperature and organic-phase nitric acid concentration on the rate of formation of dibutyl phosphate in 30 v/o TBP, as reported by Siddall [SI5].

With the relative volumes of aqueous and organic phases usually present in Purex systems, the amount of DBP formed in the aqueous phase is much smaller than in the organic phase because of the low aqueous solubility of TBP. The practical consequence of these rates is that if the solvent is washed with water to remove HN03 and with aqueous sodium carbonate to remove DBP after less than 15 min contact with process solutions at temperatures under 70°C, the concentration of DBP in process contactors can be held so low that solvent separation performance is not degraded. The DBP concentration after 15 min at 70°C is approximately

Nitration and oxidation. Nitric acid does not react appreciably with TBP at temperatures up to 70°C. At sufficiently high temperatures, however, nitration and oxidation take place. In two instances reaction of TBP-hydrocarbon mixtures with hot, concentrated solutions of nitric acid and uranyl nitrate led to destructive explosions. At Savannah River in 1953 [Cl 1], an evaporator was destroyed while concentrating a solution of nitric acid and uranyl nitrate that contained TBP and a kerosene diluent. At Oak Ridge in 1959 [A8], an explosion occurred in a radiochemical plant evaporator that was concentrating a nitric acid solution of plutonium nitrate possibly contaminated by TBP, diluent, and their radiation degradation products.

Because of these accidents laboratory studies were made at Hanford [Wl] and Savannah River [Cl 1, N5] to determine the conditions under which nitric acid solutions possibly containing TBP could be safely evaporated. Wagner [Wl] reported that a “red oil” formed by extended refluxing of a concentrated aqueous solution of uranyl nitrate, nitric acid, and TBP decomposed autocatalytically when heated to 150°C. Nichols [N5] found that a mixture of 10.5 M HN03 and TBP enters into a runaway reaction when heated rapidly to 130°C, but not at 125°C. Protective measures recommended to prevent future explosive reactions were to (1) minimize the amount of TBP added to the evaporator; (2) permit TBP to steam distill during evaporation; (3) hold the temperature below 130°C until all the TBP has been distilled.

Table 10.16 Effect of radiolytic energy density on decontamination in Purex process

Wh/liter Effect

0. 1 Process performance unimpaired

1. Noticeable but not too serious effects

10 Catastrophic loss in decontamination

Radiolysis. Radiation degrades both TBP and hydrocarbon diluent in Purex systems, with formation of molecular fragments, polymers, and nitration products. The main product, however, is the same as from hydrolysis, namely, DBP. The yield of DBP in radiolysis of TBP varies somewhat with the diluent used, water content, type of radiation, and dose rate. Baumgartner and Ochsenfeld [B6] cite production of 20 to 30 mg DBP/liter in 30-min exposure of 30 v/o TBP to 0.2 Wh/liter of radiation in mixer-settlers processing fuel cooled 240 days after 33,000 MWd/MT burnup. Because the density of DBP is 1065 g/liter, the volume percent DBP was

(100) f° °2 ),03>) = 0.0019 to 0.0028 v/o (10.8)

1065 /

Hence the radiation exposure that would produce the same amount of DBP as the 0.0005 v/o produced by hydrolysis for 15 min at 70°C, which has negligible effect on decontamination, is

In addition to DBP, ionizing radiation produces in TBP-hydrocarbon mixtures long-chain acid phosphate esters, nitrohydrocarbons, and nitrate esters that also complex uranium, plutonium, and zirconium, and that cannot be removed by simple alkaline washing. These must eventually be removed either by purging a fraction of the solvent or treating it with strong oxidants [B8].

Siddall [SI 5] summarizes the effects of increasing radiation exposure on decontamination in the first Purex extraction contactor as shown in Table 10.16. The power density, or dose rate, also has an effect on solvent performance. Baumgartner [B5] cited experiments in which

1.2 Wh/liter, delivered to 20 v/o TBP in one pass through the HA and HS contactors, reduced the zirconium decontamination factor from 1000 to 10.

The principal causes of radiolysis in a Purex plant are beta and gamma radiation in the first extracting unit (HA) and alpha radiation from plutonium in the plutonium purification units. Accurate calculation of radiation absorption by solvent is difficult, because it depends on details of the dispersion of aqueous and organic phases and contactor geometry. Blake [Bll] has given equations for estimating solvent radiation absorption when these details are known.

An upper bound for the exposure in the HA unit may be obtained by treating the solvent as uniformly dispersed as small droplets in the aqueous phase, assumed as containing all of the radiation sources. Then the radiation exposure is

R = vD[f + 6(1 ~f)]t (10.10)

where R = exposure of organic phase in watt-hours per liter

t = residence time, hours, of organic phase in contact with aqueous и = volume fraction of aqueous phase D = power density in aqueous phase, watts per liter f = fraction of radiation as beta radiation

1 — / = fraction of radiation as gamma radiation

в = fraction of gamma radiation absorbed in contactor This equation will be applied to the HA contactor of the Barnwell plant. From Tables 10.7 and 10.8, the activity of the aqueous waste HAW stream is

The watts per curie in this stream may be obtained approximately from Table 8.7 as

Hence D = (1251 Ci/liter) (0.00469 W/Ci) = 5.9 W/liter (10.13)

Blake [Bll] gives approximate values for/(0.65) and в (0.4). The volume fraction of aqueous phase v is lower with organic phase continuous than with aqueous. It will be assumed that the HA contactor will be run with organic phase continuous, with и = 0.25. Then

R = (0.25)(5.9 W/liter)[0.65 + (0.4)(0.35)] t = 1.17fh Wh/liter (10.14)

To keep organic exposure below 0.1 (Wh)/liter, the organic residence time should be below

0. 1/1.17 h, or 5 min. Thus, short-residence-time contactors, like the centrifugal contactor specified for Barnwell, are desirable for this primary decontamination service.