2.1 Bismuth Phosphate Process

The first microgram quantities of plutonium were produced [S6] in 1942 by irradiation of natural uranium with deuterons in the cyclotron of Washington University in St. Louis. This plutonium was separated at the Chicago Metallurgical Laboratory of the Manhattan Project by Seaborg and his collaborators, who employed the method of carrier precipitation frequently used by radiochemists to extract small amounts of radioactive material present at low concentration. As wartime urgency required that a plutonium separation plant be designed and built before macro quantities of plutonium could be available for process development, it was decided to use the same carrier precipitation process that had successfully produced the first small quantities of this element.

Seaborg and associates [LI] had found that tetravalent plutonium [Pu(IV)] could be coprecipitated from aqueous solution in good yield with insoluble bismuth phosphate BiP04, made by adding bismuth nitrate and sodium phosphate to an aqueous solution of plutonium nitrate. The bismuth phosphate process was developed at the Metallurgical Laboratory, demonstrated at the X-10 pilot plant at Oak Ridge National Laboratory in 1944, and put into operation for large-scale recovery of plutonium from irradiated fuel at Hanford in early 1945.

The bismuth phosphate process consisted of a number of steps in which plutonium is made alternatively soluble and insoluble. Fuel elements containing plutonium, uranium, and fission products were first dissolved in nitric acid. Plutonium was reduced to the tetravalent state by addition of sodium nitrite. Plutonium phosphate Pu3rv(P04)4 was coprecipitated with bismuth phosphate BiP04, by addition of bismuth nitrate and sodium phosphate. Coprecipitation of uranium was prevented by the presence of sufficient sulfate ion to form anionic U02(S04)22′. The BiP04 precipitate was redissolved in nitric acid and subjected to two decontamination cycles to purify the plutonium. In each cycle the plutonium was oxidized to the soluble hexavalent state by NaBi03 or other strong oxidant. Next bismuth phosphate was again precipitated, to remove fission products while hexavalent plutonium remained in solution. Then plutonium was reduced to the tetravalent state and again coprecipitated with bismuth phosphate.

After the third precipitation with bismuth phosphate, the plutonium was put through a similar cycle in which lanthanum fluoride LaF3 was used as carrier precipitate, to remove fission products not completely scavenged by bismuth phosphate in previous steps.

Despite the numerous steps, the overall recovery of plutonium exceeded 95 percent and the

^For some years most fuel will be stored much longer than 150 days, because of the large backlog of spent fuel awaiting reprocessing.

overall decontamination factor from fission products was 107. Serious disadvantages of the process were its batch operation, its inability to recover uranium, the large amount of process chemicals used, and the large volume of wastes.

Kuchler and associates [Кб, K7] of Farbwerke Hoechst have investigated the modifications necessary in the acid Thorex process to enable it to handle (1) the high concentration of fission products present in fuel with the burnups of up to 100,000 MWd/MT expected in fuel from the HTGR, AVR, and THTR, and (2) uranium concentrations of up to 20 percent in thorium, which may be used in these reactors when fissile uranium is diluted with 238 U to deter its use as a nuclear explosive. They found two difficulties with the acid Thorex process flow sheets previously used at Oak Ridge [B14] and Hanford [Jl]:

1. A second organic phase formed when the thorium concentration in first-stage solvent extraction feed was as high as 1.5 M.

2. Hydrolysis products of fission products precipitated when the feed was made acid-deficient.

To avoid these difficulties they reduced the thorium content of solvent extraction feed to 1.15 Af and developed a two-stage acid Thorex process. In this process thorium and uranium were coextracted from an acid feed to separate them from most of the fission products and then stripped back into the aqueous phase. By this means fission products were removed to such an extent that the Thorex process with acid-deficient feed could be used in the second stage without causing them to precipitate.

First stage. The flow sheet recommended by Kuchler et al. [K7] for the first stage of this two-stage process is shown in Fig. 10.22. Adjusted feed is 1.15 M in thorium and is assumed to contain from 4 to 20 percent as much uranium. The nitric acid content of feed is made from 0.7 to 1.1 M, depending on its uranium content. One volume of feed is extracted with 9.5 volumes of 30 v/o TBP in unit 1A, with eight extracting stages and eight scrubbing stages. One volume of 0.1 M HN03 is used for aqueous scrub, and 0.22 volume of 13 M HN03 is added to

the third extracting stage to complete extraction of thorium, as in the Hanford flow sheet Fig. 10.21. Uranium and thorium are returned to the aqueous phase by eight volumes of 0.01 Af HN03 in 16 stripping stages 1C. Aqueous product from 1C is concentrated and made 0.15 M acid-deficient in the evaporator and becomes partially decontaminated feed for the second stage.

Second stage. The second stage is shown in Fig. 10.23 with material quantities for the lower, 4 percent, uranium feed. In unit 2A, one volume of feed is extracted in eight stages with eight volumes of 30 v/o TBP and scrubbed in eight stages with one volume of 1 M HN03. The scrub contains 0.01 M H3PO4 to improve decontamination from protactinium and zirconium — niobium, as in the Hanford flow sheet Fig. 10.21. An additional scrub of 13 M HNO3 is added to the third extracting stage to complete recovery of thorium.

In unit 2B, thorium is returned to the aqueous phase by stripping in eight stages with 4.8

L_____ ^

Figure 10.23 Second stage of two-stage acid Thorex process for high-burnup fuel. (From К itchier et al. [K7].) volumes of 0.01 M HN03. Uranium is extracted from thorium product in eight stages by an additional 1.4 volumes of solvent.

In unit 2C uranium is returned to the aqueous phase by stripping in 16 stages with an additional 4.0 volumes of 0.01 M HN03.

Uranium product is further decontaminated by a third cycle of extraction with 5 v/o TBP in и-dodecane and stripping with 0.01 M HN03.

Process results. Decontamination factors observed by Kiichler et al. [K7] in processing 54,000 MWd/MT fuel with thorium/uranium ratio of 5.9, cooled 346 days, are listed in Table 10.19. Uranium losses were 0.012 percent to thorium product, 0.004 percent to solvent from 2C, and 0.0018 percent to solvent from third uranium cycle. Thorium loss was 0.025 percent to uranium product.

In these experiments, no mention was made of the disposition of the plutonium that will be present in fuel containing uranium irradiated to high burnup. This plutonium could either be

routed to high-level waste by adding ferrous sulfamate to the scrub solution for the second stage (as in Fig. 10.21) or could be made to accompany uranium into the third cycle. There, prior to extraction of uranium, plutonium could be reduced and made inextractable by addition of hydroxylamine.

The methods available for volume reduction are different for liquid and solid waste. For liquid waste evaporation, ion exchange, and flocculation are used; for solid waste incineration, baling and surface decontamination are the most common processes.

Evaporation. Evaporation is a process whereby a solution or a slurry is concentrated by vaporizing the solvent, normally water. Then a residue with a high solids content, usually a sludge, will be formed that is handled as the radioactive waste concentrate.

Evaporators coupled to efficient deentrainment devices provide capability for a high degree of separation for most radioactive materials. The inherently high operating cost of evaporation limits its application to those liquids that have a high concentration of dissolved solids and require high decontamination factors.

An evaporator consists basically of a device to transfer heat to the solution and a device to separate the vapor and the liquid phases. The principal parameters involved in evaporator design are heat transfer, vapor-liquid separation, and energy utilization. Common problems in radioactive waste evaporators are foaming, severe scaling, and corrosion. To resist corrosion, evaporators are usually constructed of stainless steel and operated at as low a temperature as is practical. Scale has to be removed periodically, either mechanically or chemically. Foaming can be avoided by foam-breaking devices inside the evapoarator or by antifoam agents.

The basic types of evaporators are pot evaporators and circulation-either natural or forced—evaporators. Figure 11.20 shows a natural-circulation evaporator. To improve the economy of the process, vapor compression may be employed. Vapor-compression evaporators make the latent heat of condensation available at a higher temperature to use the energy potentials of vapors by compressing it and combining it with fresh steam input.

The wiped-film evaporator is a special type of evaporator that permits evaporation to a much higher concentration of solids than do other evaporators. Liquid is fed into a heated cylinder that contains rotating blades or wipers to reduce the liquid to a film, thereby improving the heat-transfer efficiency. Wiped-film evaporators can also be operated as dryers. Other equipment that can be used for drying and calcining non-high-level waste is the same as for HLW, e. g., spray calciners, fluidized-bed calciners, and rotary kilns.

Ion-exchange. Ion exchange is a process whereby ions from an aqueous solution are bound to a solid adsorbent. Either the ion-exchanger itself, loaded with radioactive ions, will then be
handled as a waste concentrate, or it may be regenerated. In the latter case a liquid concentrate is obtained that has a volume greater than that of the ion exchanger but smaller than that of the original liquid waste. The decision as to which way will be more appropriate depends on the radioactivity concentration in the exhausted ion exchanger as well as on the price of the ion-exchanger material. In this respect, inexpensive inorganic ion-exchangers such as vermiculite are of some interest.

As the capacity of the ion-exchanger is equally exhausted by radioactive and inactive ions, this method is suitable only for waste solutions with a high radioactivity concentration relative to the total concentrations of dissolved solids.

According to the ionic nature of the radioactive contaminant, cation — or anion-exchangers have to be used, but usually the radioactive species in the waste are cations. The most efficient decontamination can be achieved by using a mixed-bed ion-exchanger as a final process stage. This is an intimate mixture of a cation-exchange resin in H+ form and an anion-exchange resin in OH" form in a 2:1 ratio. Its high decontamination effect is due to the favorable equilibrium of the reaction 2H+ + OH~ ^ 2H2 0. With mixed beds decontamination factors as high as 103 may be obtained. The product is fully demineralized water suitable as reactor coolant.

Flocculation. Flocculation is the least costly procedure to concentrate non-high-level waste. The principles are unspecific adsorption of radionuclides on a carrier, such as Fe203(aq) or calcium phosphate, or cocrystallization with a suitable crystalline precipitate, such as strontium with CaC03. The sludge has to be collected by settling or filtering and is handled as the radioactive waste concentrate. This technique, because of its rather poor decontamination effect, is suitable only for LLW. Usually, the concentrate has a high water content.

Volume reduction of solid waste. Concentration of burnable solid waste can be very effectively achieved by incineration. The ashes are handled as radioactive concentrate. This is a rather costly technique because of much effort spent for off-gas filtration and safe handling of the ashes. Figure 11.21 shows an example flow sheet of an incinerator.

A much simpler though less effective technique is baling of the waste under high pressure.

If bulky equipment, which is radioactive only because of surface contamination, is to be

discarded, the actual radioactive waste volume can be significantly reduced by complete decontamination. The techniques available include rinsing with acids or other suitable solvents, ultrasonic treatment, and sandblasting.

When the distribution equilibrium reaction involves hydrogen ions, changing the hydrogen ion concentration will have a strong effect on the distribution coefficient. An example of this is the extraction of metal complexes of acetyl acetone (HAa) and other weakly acid complexing agents by benzene. The equilibrium reaction for extraction of thorium by this reagent is

Th4+(o<7) + 4HAa(o) ^ Th(Aa)4(o) + 4H*(aq)

Hence, the thorium distribution coefficient should be

n _ [Th(Aa)4(o)] _ Kjh[HAa(o)]4 ,,

^Th " [Th*{aq)—————- [HW~ ( }

where ATh is the equilibrium constant for the above reaction. An inverse fourth-power dependence on hydrogen ion concentration is in fact observed for this distribution coefficient.

When an extractable cation, such as Zr4+, is readily hydrolyzed, reduction of hydrogen ion concentration will reduce the distribution coefficient by increasing the proportion of the element in the form of partially hydrolyzed, nonextractable ions such as ZrOJ+. This principle was used in the Redox process [B2, C7, C8] for the hexone extraction of plutonium from irradiated uranium, wherein the aqueous phase was made slightly acid-deficient with ammonium hydroxide, to reduce the extraction of zirconium and rare-earth fission products.

Uranyl nitrate has an unusual property, shared only by nitrates of a few other actinides, of being very soluble in a number of organic solvents. When such an organic solvent is immiscible with water, it can be used in a solvent extraction process to extract uranium from aqueous solutions and separate it from associated impurities. Such applications of solvent extraction are very important in extracting and purifying uranium from leach solution of uranium ores or from nitric acid solution of irradiated nuclear fuel. Examples of extractants that have been used for such separation processes are listed in Table 5.14.

The ability of diethyl ether to extract uranyl nitrate from aqueous solution has been known for a hundred years and was the method chosen by the Manhattan Project to purify the uranium used in the first nuclear chain reactors. This solvent has numerous disadvantages. It is very volatile, very flammable, and toxic, and it requires addition of sodium, aluminum, or calcium nitrate to the aqueous phase to enhance extractions. When solvent extraction was first applied to recovery of uranium and plutonium from irradiated fuel, other oxygenated solvents less volatile than diethyl ether that were first used were methyl isobutyl ketone, dibutyl

carbinol, and triglycoldicliloride. They had the disadvantages either of reacting with nitric acid or of requiring addition of solid salting agents. For solvent extraction of irradiated fuel, these have all been superseded by tributyl phosphate (TBP), in the Purex process described in Chap. 10. TBP has the advantage of being able to extract uranium efficiently from nitric acid solution without addition of solid nitrates. TBP is also used to purify natural uranium (Sec. 9.2 of this chapter).

The five solvents just discussed extract uranium in the form of neutral complexes of uranyl nitrate. With TBP the complex-forming equilibrium is

U02(N03)2 -6HsO(aq) + 2TBP(o) — U02(N03)2 -2TBP(o) + 6H20(aq)

where (aq) denotes the aqueous phase and (o) the organic.

The last two solvents in Table 5.14 are sometimes called liquid ion-exchangers because they react with water-soluble uranium-bearing ions to form organic-soluble compounds. Di(2- ethylhexyl) phosphoric acid is an example of a liquid cation exchanger, which acts through the equilibrium

О О

Ч, II

гССвНпО^РОНСо) + U022*(aq) * [(C8H170)2P0]2U02(o) + 2H+(aq)

Because of the long octyl group, both the acid and the uranyl salt are soluble in a hydrocarbon diluent and insoluble in water.

Trioctylamine is an example of a liquid anion-exchanger, which acts through the equi­librium

2[(C8H17)3NH]2S04(o) +U02(S04)34′(«?)- [C8H17NH]4U02(S04)3(o) + 2S04[16] [17] [18]-(«?)

Because of the long octyl group, both the amine and the uranyl sulfate complex are soluble in a hydrocarbon diluent and insoluble in water.

These liquid ion-exchangers have two advantages over TBP for extraction of uranium from leach liquors. Distribution coefficients are higher, so that uranium may be extracted at higher concentration from dilute leach liquors. The involvement of hydrogen or sulfate ion in the distribution equilibrium makes it possible to drive the reaction to favor either the organic or aqueous phase by adjusting the H2S04 concentration of the aqueous phase. Clegg and Foley [Cl], Merritt [М3], and Brown et al. [B8] describe many other long-chain amines and organophosphorus compounds that have been used to extract uranium from leach liquors. These may be used either for the hexavalent uranyl or the tetravalent uranous ion.

Thorium forms two nitrides, ThN and Th3N4. Th3N4 loses nitrogen at temperatures above 1500eC. At low pressures and temperatures above 2200 K, ThN also dissociates, into thorium and nitrogen. The Na pressure over solid ThN in equilibrium with liquid Th-N alloy is [II]

log10 p(atm) = 8.086 — 33,224 + 0.958 X 10*17Г5 2689 < T < 3063 К (6.5)

At a nitrogen partial pressure around 2 atm, ThN melts at 2820eC without dissociation [B4]. The density of ThN is 11.9 g/cm3.

Th3N4 is made by reacting thorium hydride with nitrogen at temperatures increasing from 200 to 900°C. ThN is made by pressing Th3N4 in vacuo at 1500°C, The nitrides react rapidly with water or moist air.

Нигё and Saint-James [H5] of the French Atomic Energy Commission have shown that TBP diluted with kerosene is a selective solvent for the fractional extraction of zirconium from hafnium. These workers recommend using an organic phase consisting of 60 v/o (volume percent) TBP and 40 percent refined kerosene, and an aqueous phase 3 N in nitric acid and

3.5 N in sodium nitrate, containing no more than 30 g zirconium/liter. Under these conditions, the distribution coefficient of zirconium is around 1.5, favoring the organic phase, and that of hafnium is only one-tenth as great, so that the separation factor is 10. Unlike thiocyanate extraction, zirconium concentrates in the organic phase with TBP.

This process was demonstrated in a pilot plant built by the French Atomic Energy Commission, which used the flow sheet shown in Fig. 7.8. The zirconium-extracting section consisted of six mixer-settler stages, and the hafnium-scrubbing section consisted of three stages. Each was 75 percent efficient. A single contact was used for the zirconium-stripping section. The plant produced 24 kg zirconium/day, containing less than 0.02 percent hafnium, from feed

Figure 7.8 Pilot plant of French Atomic Energy Commission for separation of zirconium from hafnium by solvent extraction with TBP. Solid line, aqueous; broken line, organic.

containing 2.4 percent, and made a hafnium concentrate containing 42 percent hafnium. Solvent losses by hydrolysis and solution in water were around 2 percent per cycle. Separation performance in this plant was analyzed in the examples given in Chap. 4.

Pilot-plant work on a similar process was conducted by Cox et al. [C3] at the Ames, Iowa, Laboratory of the U. S. Atomic Energy Commission. A 14-stage mixer-settler cascade was used, with 10 extracting and 4 scrubbing stages. Table 7.12 gives reported flow rates and compositions. This process differs from the one developed by Hure and Saint-James in using only HNO3 as salting agent, without NaN03. As HN03 is easier to recover and recycle (by distillation) than NaN03, HN03 alone is preferable in a commercial process. Reported separation factors ranged between 2.5 and 36. This pilot-plant work provided the design data for separation of hafnium from zirconium in the Columbia National Corporation plant [K2].

Earlier laboratory processes used to prepare metallic protactinium include vacuum decomposi­tion of the oxide by 35-keV electrons [G6] and thermal decomposition of the pentahalides [G2]. More recently, protactinium has been prepared by reducing the tetrafluoride by barium vapor [C5, S3, Zl], by calcium at 1250°C [М2] and by zinc-magnesium. The purest protactinium has been prepared by reduction in a barium-fluoride crucible at 1300°C [L2].

The two known crystalline phases of the metal, up to its melting point of 1575°C, are shown in Table 9.10 [K2].

U. S. plants. The principal U. S. reprocessing plants are listed in Table 10.3, together with their main process features. All use some form of the Purex process. In 1979, the only ones operating were the Savannah River and Idaho plants of the U. S. Department of Energy (DOE). The Hanford plant had been used primarily for recovery of plutonium and uranium from irradiated natural uranium, but was versatile and had been used, for example, for Thorex

Barnwell Nuclear Fuel Plant Barnwell, S. C.

Allied General Nuclear Services

Waiting for license

5

5

40,000

160

Shear-leach

Semicontinuous

Centrifugal + pulse columns 30

1

Electrolytic 2 cycles TBP

1 cycle TBP + silica gel Evaporate

Remote + direct [АЗ, B21, M10]

process runs (sec. 5.5). The F area at Savannah River is used primarily for irradiated natural uranium, but it, too, has been used for Thorex runs. The H area at Savannah River is a multi­purpose facility used for processing highly enriched uranium from production and test reactors.

The Idaho Chemical Processing Plant is a versatile, multipurpose facility used for recovering highly enriched uranium from a variety of fuels in naval propulsion, research, and test reactors. Materials processed [Al] include aluminum-alloyed, zirconium-alloyed, stainless steel-based, and graphite-based fuels. The West Valley plant, although designed primarily for low-enriched uranium fuel from power reactors, also processed plutonium-enriched and thorium-based fuels. It is the only U. S. plant to have reprocessed fuel from commercial nuclear power plants.

The Barnwell Nuclear Fuel Plant is the newest U. S. reprocessing plant. In 1979, it was nearly complete, but standing unused because of U. S. government policy unfavorable to reprocessing fuel from power reactors. Its main process features are to be described in Sec. 4.14 as an example of a modem Purex plant.

Overseas plants. Table 10.4 lists the reprocessing plants outside of the United States and the Soviet Union with capacities greater than 100 kg heavy metal per day and gives their principal process features. In addition to these plants, smaller plants have been operated in Italy, India, and, probably, other countries.

The Cogema plant at Marcoule, France, designed originally for natural uranium fuel from plutonium-production reactors, has also been used to reprocess natural uranium fuel from Magnox power reactors. Since 1976 the Cogema plant at La Hague, France, has been operating head-end facilities that enable it to handle slightly enriched fuel from water reactors. The capacity at La Hague is being increased, first by expansion of the plant to 800 t/year and, later, by construction of a larger plant [С5]. Figure 10.3 is an aerial view of the plant at La Hague.

The British Magnox reprocessing plant at Windscale, designed originally for natural uranium fuel from plutonium-production reactors, is being used to reprocess slightly enriched, low-bumup fuel from British gas-cooled power reactors. From 1970 to 1973 this plant also operated a Butex head-end facility that enabled it to process higher-burnup oxide fuel from LWRs. A second plant at Windscale, termed THORP, using the Purex process to treat oxide fuel, is planned [B17] as the British participation in United Reprocessors GmbH, a joint Anglo-French-German company created to coordinate commercial fuel reprocessing in Europe.

The 0.17 MT/day WAK Purex pilot plant at Karlsruhe, West Germany, has operated since 1971 [S3]. A joint venture of Kernforschungszentrum Karlsruhe (KFK) and Gesellschaft zur Wiederaufarbeitung von Kembrennstoffen mbH (GWK), this plant has provided operating experience to guide design of the full-scale Deutsche Gesellschaft fur Wiederaufbereitung von Kembrennstoffen (DKW) plant.

Additional European reprocessing experience was gained from the Eurochemic plant at Mol, Belgium [D1 ]. This joint undertaking of the Organization for Economic Cooperation and Development Nuclear Energy Agency operated a demonstration reprocessing plant from 1966 until the mid-1970s. This multipurpose plant could reprocess either 0.35 MT/day of slightly enriched uranium or 10 to 20 kg/day of 93 percent enriched MSU.

The 0.7 MT/day plant at Tokai-Mura is a prototype of a larger plant that Japan expects to build.

This section describes processes for recovering neptunium from irradiated uranium. Neptunium is an example of one of the numerous elements in irradiated fuel that could be recovered as by-products of extraction of uranium and plutonium in the Purex process.