Dissolution of monazite. The first step in all of the sulfuric acid processes is dissolution of monazite. The procedure recommended by workers at Iowa State [Bl] is as follows. Monazite ground to minus 65 mesh is digested with 93% sulfuric acid for 4 h at 210°C in a stirred reactor. The mass ratio of acid to sand, based on 100% H2S04, is 1.56. The temperature must be kept below 230°C to prevent formation of water-insoluble ThP207. The monazite is converted into a thick paste soluble in cold water. The reaction mass is cooled to 70 C and diluted with about 10 kg cold water/kg monazite. Most of the thorium, rare earths, and uranium go into solution, leaving a sludge of silica, rutile, zircon, and some unreacted monazite. Most of the solution is decanted from the silica sludge and unreacted monazite. The denser monazite is separated from the sludge and recycled. The sludge is filtered and washed to recover additional solution.

Radium in the monazite may be removed with the sludge by adding barium carbonate before decantation. This forms barium sulfate, which removes radium as insoluble radium sulfate.

This process produces a solution of thorium, rare earths, and uranium cations with sulfate and phosphate anions.

Thorium recovery processes. Because of the many elements in the solution, their chemical similarity, and the presence of phosphoric acid, separation of thorium from this acid solution has proved to be difficult. Wylie [W5] has reviewed the numerous separation processes that have been developed. Figure 6.5 shows the principal steps in seven of these processes and gives references for more details. Processes 4 and 6 appear to be the most economic when thorium, rare earths, and uranium all are to be recovered. Process 4, involving separation of thorium and rare earths from phosphate and uranium by precipitation with oxalic acid, is described next. Process 6, involving separation by solvent extraction with organic amines, is described in Sec. 8.6.

Precipitation with oxalic acid. Figure 6.6 shows the principal steps in the process for separating the sulfuric acid solution of monazite into a thorium concentrate, a rare earth concentrate, and a uranium concentrate developed at the Ames, Iowa, Laboratory of the U. S. Atomic Energy Commission [В1].

The solution of monazite in sulfuric acid containing about 50 to 60 g of thorium and rare earths per liter is diluted with about 4.5 volumes of water and brought to a pH of 1.5 by addition of NH4OH. Oxalate ion is added in the form of recycle sodium oxalate, plus sufficient oxalic acid in 10% aqueous solution to provide 110% of the oxalate ion needed to precipitate thorium and rare earth oxalates. The precipitate is filtered and washed with 1 % oxalic acid in 0.3 N nitric acid. A clean separation of uranium from rare earths plus thorium is claimed.

Because of the comparatively high cost of oxalic acid, economics requires recovery of oxalate ion. This is effected by digesting the thorium and rare-earth oxalates with a
stoichiometric equivalent of sodium hydroxide at 95°C for 1 h, to convert the precipitate to hydroxide, which is filtered and washed with hot water. Oxalate ion is recovered as sodium oxalate, of which 95% is recycled.

Uranium is recovered from the sulfate and phosphate filtrate by anion exchange (Chap. S).

Thorium and rare earths in the hydroxide precipitate are dissolved in nitric acid and separated by solvent extraction with TBP (Sec. 8.7).

The rate of radioactive disintegration, e. g., curies, is only a crude measure of the importance of individual fission products in irradiated fuel and in radioactive wastes. A more meaningful measure of potential biological hazard must also include the sensitivity of humans to inhalation
or ingestion of these radionuclides. For this purpose we use the radioactivity concentration limit C, which is the concentration of radioactivity (curies) of a given radionuclide in air or water such that an individual who obtains his or her total intake of air or water from this source will receive a radiation dose from this radionuclide at the rate of 0.5 rem/year.+ Values of the public-exposure radioactivity concentration limit C for selected radionuclides are listed in App. D. A more complete listing appears in the Federal Regulations 10 CFR 20 [F2]. Assuming that the biological hazard to an individual exposed to low levels of radiation is proportional to the accumulated radiation dose, then the potential biological hazard from inhalation or ingestion of a mixture of radionuclides is proportional to the toxicity index, defined as

V Mi

V cik

l

where X( = radioactive decay constant for nuclide і Nt = number of atoms of nuclide і

Сік = radioactivity concentration limit for nuclide і in medium к (i. e., air or water)

The toxicity index is the volume of air or water with which the mixture of radionuclides must be diluted so that breathing the air or drinking the water will result in accumulation of radiation dose at a rate no greater than 0.5 rem/year. However, the toxicity index still does not measure ultimate hazards and risk, because it does not take into account the mechanisms by which the radionuclides could be released to air or water and transported to humans.

The inhalation-toxicity indices of the fission products in the fuel discharged yearly from the 1000-MWe uranium-fueled LWR are shown in Fig. 8.3 as a function of storage time. Ingestion toxicity indices for the same fission products are shown in Fig. 8.4. If Fig. 8.4 is compared with the activity plot of Fig. 8.1, it is apparent that the relatively high toxicity, i. e., low C, of bone-seeking 90Sr makes this nuclide more important than any other fission product in terms of potential inhalation or ingestion toxicity during the first few hundred years after discharge from the reactor. Thereafter, the long-lived thyroid-seeking 1291 is potentially the most important of the fission products, even though only about 1 Ci of 129I is produced yearly in a 1000-MWe reactor.

1.2 Effects of Fuel-Cycle Alternatives on Fission Products in Irradiated Fuel

Because the nuclides 232 Th, 233U, 23SU, 23®U, 239Pu, and 241 Pu yield different amounts of individual fission products, different fuel cycles such as uranium fueling without recycle, uranium-plutonium fueling, and thorium-uranium fueling will result in different amounts of fission products in the discharge fuel. Calculated yearly production and composition of some of the principal fission products for some of the alternative fuel cycles described in Chap. 3 are listed in Table 8.3.

Plutonium nitrate solution from fuel reprocessing is to be converted either to plutonium dioxide for fuel fabrication or it may be converted to intermediate compounds suitable for reduction to plutonium metal. Direct thermal decomposition of the nitrate solution to plutonium dioxide is possible, but it requires very pure feed solutions. Industrial-scale operations usually begin with the precipitation of plutonium peroxide or plutonium oxalate, which result in further decontami­nation of impurities. Routes to the production of plutonium metal may involve hydrofluorination of plutonium dioxide, peroxide, or oxalate to form the anhydrous tetrafluoride for metaUothermic reduction. Alternatively, anhydrous plutonium halides may be formed by precipitation of PuF3 or CaF2 — PuF4 , either of which can be reduced directly to the metal, or PuF3 may be converted to PuF4 or PuF4-Pu02 for subsequent reduction. Preparation of PuCl3 from calcined Pu02 is another alternative.

H20’I40g * As contained Pu

The important industrial-scale conversion operations are described below, followed in Sec.

4.7 by the description of processes to produce plutonium metal.

Plutonium peroxide. Plutonium nitrate solutions containing plutonium of any oxidation state can be used as the feed solution for peroxide precipitation, because plutonium in aqueous solu­tions is converted to the tetravalent state by hydrogen peroxide. A flow sheet of the peroxide precipitation process is shown in Fig. 9.7 [Ml, C2]. A solution of 30 to 50 percent H2 02 is added to the plutonium nitrate-nitric acid solution slowly to promote precipitation of large and easily filtered hexagonal crystals. About three peroxide oxygen atoms are required per plutonium atom. A low temperature, about 30°C or less, is desirable to reduce peroxide decomposition. Peroxide precipitation equipment used at the U. S. Savannah River plant is refrigerated so that precipitation can occur at 15°C, followed by digestion at 6°C, to minimize decomposition of H202 by impuri­ties in the feed [Ml]. The filtered plutonium peroxide cake can be calcined at 150°C to form plutonium oxide, although a final temperature of 900°C is required to produce stoichiometric Pu02. Alternatively, the peroxide cake can be dried at 25 to 55°C and fluorinated with HF to form PuF4 , for subsequent reduction to the metal.

Peroxide precipitation achieves excellent decontamination from cationic impurities, because there are so few metals that form peroxide precipitates. However, special care must be taken in

the preparation and handling of solid and dissolved peroxides, which can decompose explosively, especially in the presence of iron and other impurities that catalyze decomposition.

Phitonium(IV) oxalate. The tetravalent plutonium oxalate can be precipitated from a nitric acid solution of Pu(IV) nitrate according to the flow sheet of Fig. 9.8 [H3]. Hydrogen peroxide is added for valence adjustment to Pu(IV), either before or during the addition of oxalic acid. Best precipitations occur at a temperature of 50 to 60°C and for time periods of oxalic acid addition in the region of 10 to 60 min. Higher temperatures and more rapid addition of oxalic acid result in finely divided and gummy precipitates. The nitric acid concentration must be adjusted so that the final oxalate slurry is in a solution between 1.5 and 4.5 M HN03. At lower acid concentrations coprecipitation of impurities is favored and the precipitate is finely divided, whereas at higher acid concentrations the solubility of Pu(IV) oxalate is unsuitably high and the precipitate is thixotropic. The oxalate cake is washed and then calcined at 300°C, followed by restructuring at 900°C to form stoichiometric Pu02. If the final product is to be plutonium metal, the oxalate can be fluorinated directly with HF and oxygen to form PuF4 for subsequent reduction.

The Pu(IV) oxalate process achieves decontamination factors of about 3 to 6 for zirconium — niobium, 12 for ruthenium, 60 for uranium, and 100 for aluminum-chromium-nickel. As com­pared with peroxide precipitation, the oxalate process achieves less decontamination from im­purities, but the solutions and solids are more stable and safer to handle. It is more suitable for processing solutions containing high concentrations of impurities that would catalyze peroxide decomposition.

Plutonium(III) oxalate. If the plutonium solution from fuel reprocessing is concentrated by sorption on a cation-exchange resin, the plutonium eluent will be a nitrate solution of stabilized trivalent plutonium. This solution may be a logical candidate for the relatively simple precipitation of Pu(III) oxalate. The Pu(III) oxalate Pu2(C204)3 -9H20 can be easily precipitated by adding oxalic acid, either as a solution or a solid. Precipitation conditions are not critical, and the Pu(III)

oxalate precipitate settles rapidly and is easy to filter. The oxalic acid can be added rapidly, with a digestion period of about a half hour.

If the Ри(Ш) nitrate is not stabilized against oxidation, hydriodic acid, hydroxylamine or ascorbic and sulfamic acid may be used as a reducing agent. Problems of handling and corrosion result with HI. If appreciable Pu(IV) were present, more stringent care would have to be taken to avoid unmanageable precipitates of Pu(IV) oxalate, as discussed above.

The Pu(III) oxalate is calcined to form the dioxide or hydrofluorinated with HF and oxygen to form PuF4 , as has been described above for the tetravalent oxalate.

When process solutions of Pu(III) are available, Pu(III) oxalate precipitation may be desirable if impurity levels are too high for the PuF3-precipitation process described later.

Pu02 from direct calcination of Pu(N03 )4. The precipitation steps of the above processes can be avoided by the direct calcination of the plutonium nitrate solution to Pu02. Calcination has been carried out at 350°C in a liquid-phase screw calciner. Half a mole of ammonium sulfate per mole of plutonium is added to the feed solution to increase the production of reactive Pu02. The calcination time and temperature must be low enough to minimize sintering, which would other­wise reduce the chemical reactivity of the oxide particles for subsequent conversion to a halide.

Direct calcination of Pu(N03)4 involves no chemical separations that could remove impuri­ties, so a highly pure plutonium nitrate feed solution is required. The plutonium dioxide product can be hydrofluorinated to PuF4, or it can be used as a feed for the formation of PuCl3. Direct calcination has received less industrial-scale application than the precipitation processes described above [C2].

Plutonium trifluoride. Plutonium trifluoride can be converted directly to plutonium metal, or it is an intermediate in the formation of PuF4 or PuF4 — Pu02 mixtures for thermochemical reduc­tion, as described in Sec. 4.8. The stabilized Pu(III) solution, produced by cation exchange in one of the Purex process options for fuel reprocessing, is a natural feed for the formation of plu­tonium trifluoride, as is shown in the flow sheet of Fig. 9.9 [03]. A typical eluent solution from cation exchange consists of 30 to 70 g plutonium/liter, 4 to 5 M nitric acid, 0.2 M sulfamic acid, and 03 M hydroxylamine nitrate. The sulfamic acid reacts rapidly with nitrous acid to reduce the rate of oxidation of Pu(III) to about 4 to 6 percent per day. Addition of ascorbic acid to the plutonium solution just before fluoride precipitation reduces Pu(IV) rapidly and completely to Pu(III).

Addition of 2.7 to 4 Mhydrofluoric acid results in the precipitation reaction

Pu(N03 )i(aq) + 3HF(o<?) — PuF3 (s) + 3HN03 (aq) (9.47)

with a solubility product for PuF3 of 2.4 ± 0.4 X 10"16. Easily filterable precipitates result from controlled rate of addition of the reagents and by maintaining a HN03/HF ratio of at least 4. In contrast to PuF4, the PuF3 precipitate is crystalline and contains no water of crystallization, so that it is easily dried to the anhydrous salt desirable for metallothermic reduction to plutonium metal. The filtered trifluoride cake is washed with 0.8 M HF and dried by ambient air. Anhydrous PuF3 is produced by further drying with warm air, followed by heating to 600°C in argon to remove remaining volatile impurities [B6].

The PuF3 process does not attain the degree of decontamination from cationic impurities that can be achieved in peroxide or oxalate precipitation, but it is acceptable when the plutonium nitrate feed contains no more than a few hundred parts of uranium and aluminum per million of plutonium. This process has been in routine use at the U. S. plant at Savannah River [B6,03].

Plutonium tetrafluoride. Precipitation of PuF4 by adding hydrofluoric acid to a Pu(TV) nitrate solution is impractical because the hydrated precipitate PuF4 -23H20 is amorphous and difficult to Alter, and it is difficult to dehydrate to the anhydrous material necessary for subsequent reduc-

Pu concentrate

0. 6 5//min

Balch size’ I to 2 kg Pu

tion to the metal. Vacuum dehydration of the precipitate at 200°C yields PuF4 — H2 0, and further heating results in the trifluoride [Cl]. When heated in a moist atmosphere above 300°C, PuF4 hydrolyzes to Pu02 [K2]. Therefore, the more difficult process of hydrofluorination of a solid is necessary to obtain anhydrous PuF4.

If the plutonium to be fluorinated is the plutonium peroxide cake, as in one of the processes used at the U. S. Savannah River plant, the air-dried cake is reacted with HF gas at 600°C. The reaction time is quite sensitive to sulfate containment in the oxalate cake, which interferes with fluorination and requires a longer time for reaction of the oxalate with HF. The interfering sulfate is that present due to a sulfuric acid wash of the cation-exchange resin prior to peroxide precipita­tion.

Alternatively, PuF4 can be formed from Pu(III) or Pu(IV) oxalate cake by drying the cake in air at 100 to 120°C and fluorinating in HF at 400 to 600°C, or the dried oxalate can be calcined at 130 to 300°C in air to form Pu02, which is then hydrofluorinated in HF to form PuF4. The hydrofluorination temperature must equal or exceed the calcination temperature, and the latter must be kept below 480°C to prevent formation of refractory oxide [Ml]. Similarly, PuF4 can be prepared by hydrofluorinating Pu02 or PuF3.

With the availability of anhydrous plutonium trifluoride, as discussed in the previous section, the equipment problems associated with the direct conversion of PuF3 to PuF4 with HF and oxygen can be avoided by roasting the trifluoride in oxygen to form a mixture ofPuF4 and Pu02, according to the reaction

4PuF3 + 02 -*• 3PuF4 + Pu02 (9.48)

The PuF4-Pu02 mixture is suitable for metallothermic reduction, as discussed in Sec. 4.8.

Air-dried PuF3 cake is roasted in an inert atmosphere at 150 to 200°C for 1 і to 3 h, and then in an oxygen atmosphere at 400 to 600° C. This is one of the processes that has been em­ployed at the U. S. plant at Savannah River [C2,03].

The CaF2‘PuF4 process. A process for forming plutonium tetrafluoride, without the attendant corrosion problems of dry hydrofluorination, involves the precipitation of the double salt CaF2’PuF4 from a solution of Pu(N03)4. By contrast to the PuF4*2.5H20 precipitate, the CaF2 -PuF4 precipitate is less soluble, is readily dried, and may be directly reduced to the metal.

A flow sheet for the precipitation of CaF2 — PuF4 is shown in Fig. 9.10 [C2]. A solution of plutonium and calcium nitrates in 4 to 5 M HN03 is added to HF solution of 6 M or less. The precipitate is washed with dilute HN03-HF solution and dried at 300°C in argon or nitrogen to form the anhydrous CaF2 — PuF4, which must be crushed to particles suitable for reduction to the metal.

The process is attractive in its simplicity. However, in the subsequent metallothermic reduc­tion the CaF2 diluent absorbs a portion of the heat of reaction otherwise needed for slag melting. Also, there is less decontamination from impurities than in the case of the other precipitation processes described earlier.

Plutonium trichloride. Although PuCl3 is more hygroscopic than the plutonium fluorides, and although it generates less heat of reaction in subsequent metallothermic reduction to the metal,

the production of PuCl3 is motivated by the reduced shielding requirements. Because of the relatively weak reaction of plutonium alphas with chlorine to produce neutrons, the neutron emission of PuCl3 is one-sixty-fourth that of PuF4 [C2]. Also, the slag from PuCl3 reduction melts at a much lower temperature than the fluoride slags (see Sec. 4.8).

Plutonium dioxide, prepared by direct calcination of the nitrate or calcination of the peroxide or oxalate precipitates, can be chlorinated to PuCl3 by HC1-H2, gaseous ССЦ, or phosgene (COClj), the latter resulting in the most rapid reaction.

Chlorination of nitrate-calcined oxide has been carried out in a fluidized bed at 500° C. Oxide from oxalate calcination has been chlorinated in a continuous screw calciner at 250 to 350°C. Because many impurities form volatile chlorides under these conditions, relatively good decon­tamination from impurities results. Consequently, this is a logical conversion step to follow the direct calcination of Pu(N03 )4.

It is essential that PuCl3 be handled only in a very dry atmosphere, otherwise hygroscopic moisture accumulation can result in excessive pressures during subsequent reduction to the metal.

Characterization of aqueous wastes. Reprocessing plants generate many aqueous waste streams, which differ widely with respect to their content of radioactivity, solids, and nitric acid. Radioactivity is characterized as low-level, intermediate (or mediumyievel, or highrlevel, but with no generally accepted quantitative definition for each category. The terms low, medium, or high activity are also used. Until around 1975 low-level liquid wastes were regarded as those that could be discharged directly to groundwaters or the ocean because after natural dilution their radionuclide concentrations were below the maximum permissible values for general population exposure. More recently, the requirement that the concentration and amount of radioactive effluents be made as low as practicable has led to a preference for discharging no liquid wastes to ground or surface waters and disposing of excess water by evaporation into plant off-gases. About the only universally accepted usage is characterization of the aqueous waste stream from the first, codecontamination cycle as high-level, or high-activity waste. This waste contains many curies per liter and must be cooled to prevent self-boiling.

Liquid wastes are sometimes characterized as low-salt or high-salt wastes. Low-salt wastes are those that can be greatly reduced in volume by evaporation without precipitation of solids. High-salt wastes are those that can be only moderately reduced in volume.

Low-acid wastes are those whose nitric acid content is too low to justify fractionating the distillate for acid recovery. If necessary to remove the little acid present, this is better done by neutralization or ion exchange. High-acid wastes are those whose nitric acid can advantageously be recovered by fractional distillation.

Steps in aqueous waste processing. Because of the great variety of aqueous waste streams and differences in process flow arrangements in different plants, there is no standard flow sheet for processing aqueous wastes from the Purex process. Figure 10.10 shows the principal steps in one possible scheme for concentrating the wastes and recovering water and nitric acid from them.

Low-level, low-acid, low-salt wastes are neutralized if necessary and concentrated in a simple flash or vapor-compression evaporator to produce low-level waste concentrates and water sufficiently decontaminated for return to process. With simple wire-mesh entrainment separators, decontamination factors of several thousand are easily obtained.

The intermediate-level waste concentrator handles the low-level waste concentrate, con­taminated aqueous solutions from solvent washing, and many other streams with appreciable solids content. With more exhaustive entrainment removal, as by partial reflux of condensate through a bubble-plate or sieve-plate column, water sufficiently pure for return to process can be produced. If concentrator bottoms are concentrated to the point of incipient crystallization, they are routed to waste storage. If still unsaturated, they are routed to the high-level waste concentrator.

The principal feed to the high-level waste concentrator is the high-level waste stream (HAW) from the codecontamination solvent extraction cycle. This typically contains about 2.5 mol HN03, 3 to 9 g fission products, and 400 to 1200 Сі/liter and generates heat at the rate of 2 to 6 W/liter. Additional feed may be intermediate-level waste concentrate and nitric acid evaporator bottoms. The high-level waste concentrator is usually operated at subatmospheric pressure and made of a corrosion-resistant material such as titanium, to extend life and minimize maintenance. Wastes are concentrated as far as possible without appreciable solids formation. If solids other than fission products are absent, a concentration of about 90 g fission products per liter can be obtained. Products are contaminated nitric acid overhead, slightly under 2.5 M, and evaporator bottoms, about 7 Af in HN03. Because evaporator bottoms self-heat at a rate up to l°C/min, the evaporator and the bottoms storage tanks must be provided with reliable cooling.

Previous Page

Figure 10.10 Steps in Purex waste processing and acid recovery.

Denitration of high-level wastes. To reduce corrosion during subsequent storage of concentrated high-level wastes, it is desirable to reduce their nitric acid content from around 7 to between 2 and 4 M, a value high enough to prevent major precipitation of hydrolyzed fission-product nitrates. To avoid loading the wastes with additional nonvolatile solids, nitric acid concentration can be reduced either by steam distillation or by reduction to gaseous nitrogen oxides by an organic reducing agent such as sucrose or formaldehyde.

Steam distillation is used in British Magnox plants. There, when the HAW stream has been evaporated to the desired solids content, water is substituted for radioactive feed for about 2 days and evaporation continued at constant volume. In this way the acidity of the concentrate is reduced to about 4 M.

Some of the reactions that occur with organic reducing agents are

With formaldehyde: HCHO + 4HN03 ->• C02 + 3H20 + 4N02

With sucrose: C12H22Ou + 24HN03 -» 12CO + 23H20 + 24N02

Denitration with formaldehyde was first studied at Harwell [H5, H6]. Above 80°C the reaction proceeds smoothly, and the acidity can be reduced to 1 M in from 1 to 2 h. At lower temperature, if unreacted formaldehyde is allowed to accumulate, the reaction may become uncontrollable.

Denitration with sucrose was studied at Hanford [В16]. If the temperature is above 85°C the reaction proceeds smoothly through many intermediate stages, after an incubation period of several minutes. As with formaldehyde, sugar should not be added at lower temperature.

Treatment of process water. Water recovered in processing aqueous wastes usually contains so little radioactivity that it can be recycled to the reprocessing plant, but it must be treated
further if it is to be discharged. Some water must be discharged because more water is fed to the plant in nitric acid and aqueous solutions than leaves it in aqueous wastes. This excess water is evaporated into gases and ventilating air discharged through the plant stack. Before excess water vapor is discharged in this way, it is treated to remove radioiodine and filtered to remove suspended solids.

1 INTRODUCTION

1.1 Definition

Radioactive waste is any waste material—gas, liquid, or solid—whose radioactivity exceeds certain limits. These limits have been established by governments or by local authorities, guided by the recommendations of the International Commission on Radiation Protection (ICRP). The ICRP recommendations define the maximum permissible concentration (MPC) for each individual radionuclide and for mixtures of radionuclides in water or air. The U. S. regulation defines such limiting concentration as the radioactivity concentration limit (C), which is the terminology used in this text. Values of C for selected actinides and long-lived fission products in water or air are given in App. D.

The intention of regulations limiting the release of radioactive material from nuclear installations is to keep the radioactivity concentration in ground and surface water or in air well below the levels recommended by the ICRP. The regulations may follow either one of two principles or may combine them:

Limitation of the total amount of radioactivity associated with a certain material that may be released over a given period of time

Limitation of the radioactivity concentration in the material to be released

As a consequence of these limitations, most of the radioactivity arising as waste from nuclear technology has to be isolated from the environment by some storage or final disposal technique. The first step toward this goal is usually a volume reduction, preparing the waste for interim storage as a liquid or solid. This is considered part of the waste-generating technology rather than of the waste management. Waste management is defined to include interim storage, final conditioning, and long-term storage or disposal.

9.2 Difficulties

Production of uranium metal sufficiently pure for use in nuclear reactors is difficult. Uranium forms very stable compounds with oxygen, nitrogen, and carbon, and it reduces the oxides of many common refractories. Methods that yield uranium metal at temperatures below its melting point result in a fine powder that oxidizes rapidly in air and is difficult to consolidate into massive metal. Uranium cannot be deposited electrolytically from aqueous solution. It is not practical to purify uranium by distillation because of its very high boiling point, 3900° C. Any nonvolatile impurities introduced into uranium during production will remain in it during subsequent operations and contaminate the final product.

9.3 Alternative Methods

Four methods that have been used to produce uranium metal are

1. Electrolysis of fused salts

2. Reduction of UOj

3. Reduction of UF4

4. Reduction of UCI4

Electrolysis. Electrolysis of KUF$ or UF4 dissolved in a molten mixture of 80 percent Cad2 and 20 percent NaG was the method used by the Westinghouse Electric Company to produce the first pure uranium metal for the Manhattan Project [S4]. Because electrolysis was carried out below the melting point of uranium metal, 1130°C, the crude metal contained salt that had to be leached with water and had to be remelted before acceptably pure metal could be obtained. By 1943 this method was superseded by the less costly reduction of UF4 by magnesium, to be described later.

Possible reductants. Elements that might be considered for reducing U02, UF4, or UQ4 to metallic uranium are hydrogen, sodium, magnesium, or calcium. Carbon is impractical because of formation of uranium carbide, and aluminum is undesirable because it forms an intermetallic compound with uranium. Sodium, magnesium, and calcium do not do this.

To show which combinations of uranium compound and reductant are thermodynamically favorable, the ftee-energy change in reducing U02, UF4, or UQ4 by hydrogen, sodium, calcium, or magnesium has been evaluated in Table 5.29. These data are for a temperature of

1500 К (1227°C), which is high enough for rapid reaction and is above the melting point of uranium, 1130°C, so that the uranium product would be consolidated rather than a fme powder. Part 1 of Table 5.29 lists the free energy of formation of 1 mol of each of the uranium compounds from its elements and the free energy of formation of the number of gram-moles of the oxide, fluoride, or chloride of the four possible reductants needed to produce 1 g-mol of uranium. Part 2 of Table 5.29 lists the free-energy change in reaction of each possible combination of uranium compound and reductant. For example, the free-energy change in the reaction

UF4 + 2Mg -*• U + 2MgF2

is —407,219 — (—351,544) =—55,675 cal/g-mol uranium (5.16)

For reduction of the uranium compound to be complete without requiring a large excess of reductant, the free-energy change at 1500 К should be more negative than 10,000 cal/g-mol. Table 5.29 shows that hydrogen is completely impractical and that the only feasible reductant for U02 is calcium. For UF4 or UC14, sodium, magnesium, or calcium meet the free-energy criterion.

Reduction of U02. Production of metallic uranium by reacting U02 with calcium metal is thermodynamically possible and was practiced in Germany in 1942 [SI]. However, the melting point of CaO is so high, 2615°C, and the heat of reaction is so small, 44 kcal/g-mol uranium, that it is impossible to melt the lime to make a clean separation between it and uranium metal. The result is that the uranium product is in the form of small particles and the recovery of clean uranium is usually no more than 35 to 40 percent.

Reduction of halides. Table 5.30 lists pertinent properties of substances that might take part in the reduction of the uranium halides UF4 or UC14. As this table shows, the heat available per mole of uranium produced is much higher than in the reduction of U02. In addition, the melting point of the halide by-product is much lower than that of CaO and is near that of uranium metal. Consequently, the reaction temperature can be raised enough to melt both the uranium metal and the halide by-product, so that a clean separation between metal and slag can be obtained.

Production of reactive metals by reduction of the tetrachlorides is the basis of the well-known Kroll process for titanium or zirconium (Chap. 7). Although metallothermic reduction of uranium tetrachloride is thermodynamically possible for uranium also, its use in practice is made difficult by the hygroscopic character of UC14. This salt picks up water from moist air, which would contaminate uranium metal with uranium oxide after reduction. Moreover, the low boiling point of UC14 (789°C) relative to the higher melting point of uranium (1130°C) means that the UC14 would have to be fed into the reaction zone as vapor, a complication avoided with UF4, whose boiling point is higher (1457°C). For these reasons, UF4 is generally used.

Of possible reductants for UF4, sodium is less desirable than magnesium or calcium because, like UC14, its boiling point is far below that of uranium metal. Use of sodium would require either feed of sodium vapor from an external source or operation of the reactor at very high pressure. Choice of reductant for UF4 thus is essentially limited to calcium or magnesium. In practice magnesium is used for production of large batches of uranium, with calcium being used for smaller quantities, as when criticality considerations limit batch size. For example, at the French refinery at Malvesi [B5], calcium is used to produce uranium in small batches, under 100 kg, with magnesium used for larger batches. The same general practice is followed in the United States. Magnesium reduction has been used in the United States since 1943, when F. H. Spedding and co-workers at Iowa State College developed the process for the Manhattan Project [S4] to supersede fused-salt electrolysis.

The advantages of calcium include the following:

1. The reaction can be carried out at atmospheric pressure, because the melting point of calcium fluoride is below the boiling point of calcium metal.

2. The heat of reaction is sufficient to melt both uranium metal and CaF2 slag, with reactants initially at room temperature, so that preheating is unnecessary.

The advantages of magnesium include the following:

1. Magnesium costs much less than calcium, and only 60 percent as much mass of reductant is needed.

2. It is easier to obtain magnesium of the requisite purity than calcium, and magnesium does not pick up oxygen from air or moisture.

The problems in using magnesium are these:

1. The reaction must be carried out in a sealed reactor to contain the superatmospheric vapor

pressure of magnesium developed when the reactants are heated to the melting point of MgF3.

2. It is necessary to preheat the charge because the heat of reaction is too small to melt the reaction products when the reactants are initially at room temperature.

The synthetic isotopes 237U and 233U occur in the 4л + 1 radioactive series, of which 237Np is the longest-lived member. Figure 5.3 shows the nuclear reactions that occur successively as these nuclides decay into their nearly stable end product 209Bi. Table 5.4 gives the half-lives of these radioactive species and their principal decay radiations. Radiations from 229 Th and its short-lived daughters will be the most important contributors to the remaining toxicity of high-level wastes from irradiated reactor fuels containing 237 Np, after such wastes have been in storage for several hundred thousand years, when these daughters will be in secular equilibrium with 162,000-year 233 U.

Enriched or depleted uranium is usually produced in the form of UF6, but is used as metallic uranium or U02. This requires conversion of UF6 to UF4 or U02. UF6 is converted to UF4 by vapor-phase reduction with hydrogen. Because the heat of reaction is small, the mixture must be heated. In small reactors used for converting highly enriched uranium, heat is provided internally by reacting fluorine with hydrogen.

Three processes have been used for converting UF6 to U02. In one, UF6 is reduced to UF4, which is then hydrolyzed by steam,

UF4 + 2H20 -► U02 + 4HF

the reverse of the reaction used to make UF4. In a second process, UF6 is hydrolyzed to U02F2 by solution in water, after which ammonia is added to precipitate ammonium diuranate,

2U02F2 + 6NH4OH -+ 4NH4F + (NH4)2U207 + 3H20

The diuranate is then reduced to U02 with hydrogen at 820°C. In the third, AUC (ammonium uranyl carbonate) process, developed in West Germany by Nukem, streams of gaseous UF6, C02, and NH3 are fed batchwise into demineralized water, whereby (NH4)4U02(C03)3 is precipitated. The AUC is converted batchwise to U02 by contacting it with steam and hydrogen at 500° C in a fluidized bed, with recovery of C02 and NH3. Subsequently, steam at 650°C is supplied to the fluidized bed to reduce fluorine content to 50 to 60 ppm by pyrohydrolysis.

1.4 Valence States

The principal valence of zirconium and hafnium is +4. Halides of valence +2 and +3 have been prepared, but they are of less practical importance as they disproportionate when heated and react with water in aqueous solution.

3.1 Tritium from Neutron Activation

In addition to tritium produced by ternary fission, as shown in Table 8.1, tritium is also produced in reactors by neutron reactions with lithium, boron, and deuterium. Reactors can be designed to produce tritium by irradiating lithium targets with thermal neutrons, resulting in the (я, a) reaction:

|li + in-* 2He + ?H (8.47)

with a 2200 m/s cross section of 940 b. Lithium contaminants in reactor fuel, structure, or coolant will produce tritium by reaction (8.47). Also, the more predominant natural isotope 7 Li reacts with high-energy neutrons according to

Fast-neutron cross section [S3]

?Li+in->-?He-l-?H 55 mb (8.48)

2іі + £/і-*£л + $Не + ?Н 330 mb (8.49)

Although relatively little tritium is produced from natural lithium contaminant in thermal reactors by reactions (8.48) and (8.49), the 7Li source of tritium is also produced by the (n, a) reaction with boron used for reactivity control:

l? B+ l0n ■* ?Li + $He (8.50)

The cross section for reaction (8.50) is 3837 b for 2200 m/s neutrons. Boron also reacts with high-energy neutrons in reactors to produce tritium by the reactions:

Fast-neutron cross section [S3]

‘?В + іл-»-2$Не + ?Н 42 mb (8.51)

‘[В + іл-^Ве-ЧН 15 mb (8.52)

The cross section for reaction (8.51) can be interpreted as the spectrum-averaged value for neutrons of energy greater than 1 MeV. The threshold neutron energy for reaction (8.52) is 10.4 MeV. The flux of neutrons with energies above this threshold is negligible in fission reactors, so tritium production from reaction (8.52) is negligible.

Neutron absorption in deuterium in water coolant-moderator produces tritium by the (n, y) reaction

?H + in-4H (8.53)

for which the 2200 m/s cross section is 0.53 mb. This reaction is most important as a tritium source in reactors cooled and/or moderated by heavy water, but it is negligible in LWRs.

The activity (NX)T of tritium produced in a reactor can be estimated by assuming irradiation in a constant neutron flux for a period Tr and applying Eq. (2.27). For these tritium-producing reactions it is sometimes a good approximation to assume that the parent material is present in nearly constant amount during the irradiation period. The high (n, a) cross section for 10 В might suggest that this nuclide would decrease considerably in amount if exposed to the full reactor flux over a period of even 1 year, which is the typical time interval for reactivity adjustment between refueling intervals. However, in boiling-water reactors (BWRs), which use solid control absorbers for long-term reactivity control, the effect of the large thermal cross section of boron is to self-shield all but the surface of these absorbers from thermal neutrons, so that very little of the boron is actually consumed during a refueling interval or even during the period Tr of fuel irradiation. The boron cross section for fast neutrons is relatively small, so fast neutrons are not self-shielded and essentially homogenous exposure of all the boron to the average fast-neutron flux in the reactor can be assumed. In PWRs boron is dissolved in the coolant for long-term control of reactivity, with the boron concentration controlled by chemical means during the irradiation period between refueling intervals. Because this concentration change occurs over a time period short compared to the

half-life of tritium, and because the boron concentrations are repeated from one refueling cycle to another, a constant average concentration of boron in the coolant can be assumed for the purpose of estimating tritium production. Therefore, for those tritium sources in which the parent nuclide can be assumed to be of constant amount, Eq. (2.27) takes the form

(N)T = 2 NpMl-e-W) (8.54)

I

where Ni = number of atoms of species і producing tritium by neutron reactions о,- = cross section for species і to produce tritium T = radioactive decay constant for tritium TR = time of constant-flux irradiation

For an irradiation period Tr much smaller than the tritium half-life of 12.3 years, jTr •< 1, and Eq. (8.54) simplifies to

(NX)t — tTr ^ Л)а,-0 (8.55)

І

To illustrate, we shall consider a 1000-MWe PWR with the same core composition and power density as the reactor described in Chap. 3. The in-core inventory of water is approximately 13,400 kg. The tritium produced by 2H(«, y) during one calendar year in an average thermal-neutron flux of 3.5 X 1013 «/(cm2 ■ s) with an effective 2H(«, 7) cross section of 3.35 X IO-4 b is

mT = (їіт^) (°-8 yrXl-34x 107 g)(2X6-°^ff at0ms) (1.5 X 10-4 atoms 2H/atom H)

X (3.35 X.0- c^p. S X 10» (3,7 x,0.. 1-9 Cl

The actual irradiation time Tr is 0.8 years because of the assumed 0.8 capacity factor of the power plant.

Assuming an average dissolved boron concentration of 600 ppm in the coolant, the tritium produced from reaction (8.51) in an average fast-neutron flux of 7.2 X 1013 «/(cm2 — s) is similarly obtained by applying Eq. (8.55), resulting in an estimated yearly production of 360 Ci.

In a water-cooled reactor the coolant is processed continuously for control and removal of chemical and radioactive contaminants. In a PWR the lithium formed by (n, a) reactions in dissolved boron will add to whatever natural lithium is present as a contaminant and for corrosion control, but the continued processing will hold it at some steady concentration. For the purpose of this estimate we shall assume a concentration of 1.0 ppm of lithium in the coolant and will neglect the additional 7 Li produced by reaction (8.50). However, after the coolant lithium has been exposed to thermal neutrons for a few years it will become depleted in the 6 Li, because of the high absorption cross section of 6 Li. A typical isotopic composition of lithium in the coolant of a PWR is 99.9 percent 7Li [S2]. Applying Eq. (8.55) for tritium produced by 6Li (n, a) yields the yearly production of 34 Ci listed in Table 8.10. The yearly production of the tritium from 7 Li reactions is estimated at 4 Ci [S2].

The total yearly production of neutron-activation tritium in the PWR coolant is 400 Ci, as shown in Table 8.10. Another source of tritium in the coolant is fission-product tritium that diffuses through the fuel cladding and escapes through pin-hold penetrations through the cladding. Estimates of the amount of fission-product tritium reaching the coolant in LWRs with zircaloy fuel range from 0.2 to 1 percent of the total fission-production tritium produced within the fuel.

In the HTGR the principal nonfission sources of tritium are from lithium and boron contaminants in the graphite fuel elements. Typical contaminant concentrations assumed in the HTGR designs [HI] are

^ = 1.2 X 10’6 I = 1.36 X 10’4

At such low concentrations the lithium and boron are exposed homogenously to the neutron flux. Because of the large thermal-neutron cross sections for 6 Li and 10 B, these isotopes are depleted significantly during the typical fuel irradiation time of 4 years. Therefore, to calculate the tritium activity (A’A)j — in a fuel element after an irradiation time TR, we rewrite Eq. (2.100), recognizing that the chain-linking term here is фа instead of X. For the 6 Li reaction of Eq. (8.47),

(N)T = тМ°Ф?6- TR — e-xTrR) (8.56)

t — фО s

where iV° = initial number of atoms of 6 Li а і = (л, a) cross section for6 Li

For an effective 6 Li cross section of 294 b and an average thermal-neutron flux of 12 X 10I4n/(cm2 -s), the tritium in discharge fuel due to 6 Li(n, a) is calculated to be

(NX)T = 308 Сі/Mg of graphite

The tritium from fast-neutron reactions with 10 В is estimated to be about 0.6 Сі/Mg of graphite, and tritium from 7 Li and other sources is even less.

The fuel discharged yearly from the 1000-MWe HTGR of Fig. 3.33 contains 90.5 Mg of graphite [P3]. The yearly production of tritium from neutron activation of lithium impurities is then estimated to be

(308X90.5) = 27,900 Ci/year

Ibis compares with 9.59 X 103 Сі/year of fission-product tritium calculated to be present in the discharge fuel from a 1000-MWe HTGR [Gl].

Tritium is also produced in the HTGR helium coolant by neutron reactions with small amounts (1.7 X 10"5 percent) of 3He present in underground sources of natural helium:

ІНе + ія -*• jH + fH (8.57)

with a 2200 m/s cross section of 5327 b. For an inventory of natural helium of 618 kg in the core of a 1000-MWe HTGR [Bl], 3H is initially formed at the rate of about 8,020 Сі/year and is trapped by forming tritides with hot titanium in the coolant cleanup system. However, because of its large cross section, 3He is rapidly depleted by neutron absorption. It is replaced by fresh helium introduced to make up for coolant leakage. If a fraction /He of the coolant leaks from the coolant system per unit time, the steady-state concentration X*He of 3He within the reactor coolant can be calculated by

ЛЯе/н.*?Не = ЛГЙеЛГ»н.0а. н. + Лйе^не/не (8.58)

where Nue = total inventory of helium in the coolant system ІУне = total inventory of helium within the reactor core X? He = atom fraction of 3 He in natural helium (1.7 X 10" 7)

Solving for 2ГэНе, we obtain

He 1 +Лгне0азне/Л? Не/не

From HTGR design data, it is estimated [Bl ] that

^йе /не = 0.015/yr

For an effective &эНе = 2800 b, and for <p= 1.2 X 1014n/(cm2 ’s), we obtain

ЛГзНе = 2.63Х 10-9

The resulting steady-state rate of production of tritium in the coolant from 3He(n, p) is 124 Ci/year.

In the CANDU heavy-water reactor the dominant source of tritium is the deuterium activation reaction of Eq. (8.53). The data given in Prob. 3.3 for the Douglas Point Nuclear Power Station provide a basis for estimating the rate of production of tritium in the heavy-water moderator and coolant:

Electrical power = 203 MWe

Inventory of D20 coolant in reactor core = 2.82 X 106 g Average thermal-neutron flux in coolant = 6.10 X 1013n/(cm2 ‘s)

Inventory of D20 moderator in reactor core = 7.72 X 107 g Average thermal-neutron flux in moderator = 1.01 X 10I4n/(cm2,s)

Average 2H(n, y) cross section = 4.45 X 10"4 b

The rate of production of 3H in the moderator is then

cm2) Ci/yr

For a 1000-MWe CANDU power plant with the same reactor lattice and with the same ratio of DjO in-core inventory to uranium inventory as in the Douglas Point Reactor, the yearly production of tritium in the heavy water is then

(w)(2-60 x 1qS) = >-28 x 106 Сі/уг

Because of this large rate of tritium generation, it is necessary to operate a small isotope-separation unit to prevent the buildup of large concentrations of tritium in the heavy water. The losses of heavy water are kept small enough so that only a very small fraction of the tritium is released to the environment. The yearly release of tritium reported for the Douglas Point Station is typically about 4000 Сі/year, which is about 0.2 percent of the allowable release [Dl].

3.2 14C

14 C is an activation product of potential environmental importance in the nuclear fuel cycle because of its long half-life of 5730 years and because it easily appears in volatile form, such as C02. Most of the 14 C formed in reactors results from the (n, p) reaction with 14 N:

14 N + o’* ->■ 14C + J H (8.60)

The 14N, which constitutes 99.6 percent of natural nitrogen, is present as residual nitrogen impurity in oxide fuel of water reactors and fast-breeder reactors, as air dissolved in the coolant of water-cooled reactors, and as residual nitrogen in the graphite of HTGRs. The 14 N activation cross section for 2200 m/s neutrons is 1.85 b.

14C also results from the (n, a) reaction on 17O, which is present as 0.03 percent of natural oxygen, with a 2200 m/s cross section of 0.235 b:

ЧО+ ‘on -* 4C + ?He (8.61)

In graphite-moderated reactors another source of 14C is the (n, 7) reaction with 13C, which is present as 1.108 percent of the natural carbon in graphite:

4С + ІП-+4С + 87 (8.62)

However, the 2200 m/s cross section is only about 0.9 mb. Additional but less important reactions are

l? N+bn-*-4C + ?H (8.63)

with a 2200 m/s cross section of 2.4 Ж 10’7 b, and

10 + £и-*4С + !Не (g.64)

The activity (jVX)c of 14 C produced in a reactor can be estimated by assuming irradiation

in a constant-neutron flux for a period TR and applying Eq. (2.27). Because of the long half-life of 14C, the approximation XcTR < 1 leads, as in the case of Eq. (8.55), to

(7VX)c = Xc7’« 2 Ni0i<p (8.65)

І

where Л’,- = number of atoms of species і producing 14 C by neutron reactions о,- = cross section for species і to produce 14 C Xc = radioactive decay constant for 14 C

14 C produced in water coolant is important because of its possible environmental release at the reactor site. If 14 C forms carbon dioxide or a hydrocarbon such as CH4, and if no processes

are provided to recover the gaseous 14 C, the coolant-produced 14 C will be discharged along with the noncondensable gases removed by the main condenser air ejector in a BWR and through the gaseous waste disposal system for a PWR.

We consider here the production of 14C by reactions (8.60) and (8.61) in the reactor coolant, which requires estimates of the inventories of 17 О and dissolved nitrogen in the coolant within the reactor core. For the 1000-MWe PWR with an in-core water inventory of 13,400 kg, an effective 170(n, a) thermal cross section of 0.149 b, and an average thermal-neutron flux of 3.5 X 1013 n/(cm5,s), the 14C production from I70 is estimated to be

2.2 Ci/year.

To obtain the 14 C from dissolved nitrogen in the coolant, a dissolved nitrogen concentration of 1 ppm (by weight) is assumed, with an effective 14N(n, p) cross section of

1.17 b, resulting in a yearly production of 0.061 Ci of 14C. The total yearly production of 14C in the PWR coolant is then about 2.3 Ci/year, which is the source term for possible environmental release at the reactor site. A 1000-MWe BWR would contain about 33,000 kg of water in the core under operating conditions. Assuming the same values of neutron flux and cross sections, the yearly production of 5.6 Ci of 14C in the BWR coolant is estimated.

The 14C produced by 170(n, a) in U02 fuel, calculated as the yearly production per metric ton (Mg) of uranium originally in the makeup fuel, is again obtained by applying Eq. (8.65):

X [3.5 X 1013 (cm5• s)"1 ] ( inl0 ‘-C’——————— ;—- r) ((0.8)

3.7 X 1010 disintegrations/s/ 5730 yr / v

= 2.54 X 10-2 Ci/(yr-MgU)

For the 14N source in the fuel, it is assumed that the nitrogen impurity is present in U02 at a weight ratio of 25 ppm, although nitrogen contents from 1 to 100 ppm have been reported [Kl]. The yearly production per metric ton of uranium is

= 0.130 Ci/(yr-MgU)

The total amount of 14C produced yearly in the fuel is then 0.155 Сі/Mg of uranium.

To obtain the 14 C in the discharge fuel, we use the fuel life of 3 calendar years, as calculated in Chap. 3 for the reference PWR. Because there is negligible decay of the 14 C during this З-year period, the concentration in the discharge fuel is

3X 0.155 = 0.465 Ci/Mg

The quantity of 14C in the total fuel discharged yearly, which initially contained 27.2 Mg of uranium, is

0. 465 X 27.2 = 12.7 Ci/yr

In a PWR operating with plutonium recycle the thermal-neutron flux is lower than for uranium fueling because of the higher fission cross section for plutonium. As a result, less 14C is produced by thermal-neutron activation within the fuel, as shown in Table 8.11.

Fast-breeder oxide fuel is also assumed to contain 25 ppm of residual nitrogen [К1]. Typical average fast-spectrum cross sections are 0.135 mb for 170(и, у) and 14 mb for 14N(n, p)

Table 8.11 Volatile radionuclides in discharge fuel from neutron activation*

*1000-MWe reactors, 80% capacity factor:

*PWR, pressurized-water reactor; HTGR, high-temperature gas-cooled reactor; LMFBR, liquid — metal-cooled fast-breeder reactor. Data are calculated for 150 days after discharge for PWR and HTGR, 60 days after discharge for LMFBR.

within the reactor core [Cl], For an average fast-spectrum core flux [Cl] of

3.8 X 1015«/(cm2-s), and for the breeder parameters of Fig. 3.34, the estimated yearly production of 14C for a 1000-MWe fast breeder is estimated to be 3.3 Сі/year. Relatively little 14 C is produced in the blanket fuel because of the lower neutron flux there.

The fuel of the HTGR consists of uranium and thorium particles, as oxides and carbides, distributed through a graphite matrix. The important 14 C-producing reactions in this fuel are 14N(«, p) and 13 C(n, y). Residual nitrogen is assumed to be present in graphite at a weight ratio of 30 ppm [B4]. In the thermal-neutron energy spectrum of an HTGR the effective activation cross sections [B4] are 0.683 b for 14N and 3.3 X 1СГ4 b for 13C. For an average thermal-neutron flux of 1.2 X 1014 «/(cm2 — s) and a 4-year fuel life, the estimated concentra­tion of 14C in the discharged graphite fuel is calculated from Eq. (8.65), with the result:

The fuel discharged yearly from the 1000-MWe HTGR of Fig. 3.33 contains 7.95 Mg of heavy metal and 90.5 Mg of graphite. The yearly production of 14C by this reactor is then estimated to be

(1.33 X 10-3X90,500)= 120 Ci/yr

In another HTGR calculation 1 ppm of N2 in the graphite is assumed [HI], resulting in an estimated yearly production of 24 Сі/year for a 1000-MWe plant.

When HTGR fuel is reprocessed the graphite matrix is to be incinerated in oxygen, exposing the fuel particles for dissolution. The combustion gas, which contains the 14 C and all of the normal carbon from the graphite, is to be recovered to avoid release of14C to the environment.