CHAPTER 6 CHEMICAL PROCESSING* 6-1. INTRODUCTION One of the principal advantages of fluid fuel reactors is the possibility of continually processing the fuel and blanket material for the removal of fission products and other poisons and the recovery of fissionable material produced. Such continuous processing accomplishes several desirable objectives: (a) improvement of the neutron economy sufficiently that the reactor breeds more fissionable material than it consumes, (b) minimiza- tion of the hazards associated with the operation of the reactor by main- taining a low concentration of radioactive material in the fuel, and (c) im- provement of the life of equipment and stability of the fuel solution by removing deleterious fission and corrosion products. The performance and operability of a homogeneous reactor are considerably more dependent on the processing cycle than are those of a solid fuel reactor, although the objectives of processing are similar. The neutron poisoning in a homogeneous reactor from which fission product gases are removed continuously is largely due to rare earths [1], as shown in Fig. 6-1. In Fig. 6-1 the rare earths contributing to reactor poisoning are divided into two groups. The time-dependent rare earths are those of high yield and intermediate cross section, such as Nd43 and Nd'#% Prl4l and Pm!'47) which over a period of several months could accumulate in the reactor and result in a poisoning of about 209,. The constant rare-earth poison fraction is due primarily to Sm'#® and Sm!5! which have very large cross sections for neutron absorption but low yield, and therefore reach their equilibrium level in only a few days’ operation. Poisoning due to corrosion of the stainless-steel reactor system was cal- culated for a typical reactor containing 15,500 {t* of steel corroding at a rate of 1 mpy. It is assumed that only the nickel and manganese contribute to the poisoning, since iron and chromium will hydrolyze and precipitate and be removed from the reactor system; otherwise, corrosion product poisoning would be four times greater than indicated in Fig. 6-1. The control of rare earths and corrosion product elements is discussed in sub- sequent sections of this chapter. Removal of solids from the fuel solution also improves the performance of the reactor by diminishing the deposition of scale on heat-transfer surfaces and reducing the possibility of erosion of pump impellers, bearing surfaces, and valve seats. *By R. A. McNees, with contributions from W. E. Browning, W. D. Burch, R. E. Leuze, W. T. McDuffee, and 8. Peterson, Oak Ridge National Laboratory. 301 302 CHEMICAL PROCESSING [cHAP. 6 4.5 | ! A. Total Rare Ecrths 40 | B. Time Dependent Rare Earths ' C. Corrosion Products at 1 mpy D. Constant (98.2% Rare Earths 3.5 + 1.8% Cd) E. Sr8% 4 1c99 30} F. Alkali Metals ] ’ G. VIIB(13Y) R H. Nobie Metals e 2.5 (RR103) 2 z 0 30 60 90 120 150 Irradiation Time, days Fig. 6-1. Poison effect as a function of chemical group in core of two-region thermal breeder. D,O 2 To Reactor u, Th, Pa, Blanket Fission Products Slurry from DQ_O Blanket | Recavery Decay ! Storage 160 days) Overflow Returned to Reactor Core U and Fission Products F — eed from—b\ Z——Hyd roclone s / Fission Products Reactor Core Solvent v ‘ '{ Separation | Extraction - - " Pa, T B | Underflow Fission Products Containing Decay | U Insoluble | Storage | Fission u (120 days) To Reactor | Products To Reactor Core Core Fic. 6-2. Conceptual flow diagram for processing fuel and blanket material from a two-region reactor, The biological hazards associated with a homogencous reactor are due chiefly to the radioactive rare earths, alkaline earths, and iodine [2]. The importance, as a biological hazard, of any one of these groups or nu- clides within the group depends on assumptions made in describing ex- posure conditions; however, I'3! contributes a major fraction of the radia- tion hazards for any set of conditions. While the accumulation of hazardous materials such as rare earths and alkaline earths will be controlled by the processing methods to be described, less is known about the chemistry of 6-1] INTRODUCTION 303 Blanket 1.4 mUOQSO4 in D70 at 250°C [—PDQO 1 D50 = Dissolution Recovery [ | inHNO, - Solvent Extraction Partition Stripping Solvent === Lrgnium Fission Plutonium Uranium — Plutonium Product Waste FiG. 6-3. Conceptual flow diagram for processing blanket material from a two- region plutonium producer. iodine in the fuel systems and methods for removing it. Ixisting informa- tion on iodine processing is discussed in Section 6-3. Schematic flowshects for proposed processing schemes for two types of two-region aqueous homogencous reactors are shown in Figs. 6-2 and 6-3. In both cases, solids are removed by hydroclones and concentrated into a small volume of solution for further processing. The nature of such processing will be determined by the exact design and purpose of the reactor. Thus, for a two-region plutonium producer, the core and blanket, materials would have to be processed separately to avoid isotopic dilution, while for a thorium breeder, core and blanket material could be processed together. However, if an attractive method should be developed for leach- ing uranium and/or protactinium from a thorium-oxide slurry without seriously altering the physical properties of the slurry, the two materials could be processed separately. In a similar way, the relation between iodine control and fission product gas disposal is such that neither problem can be disassociated from the other. A specific, complete, and feasible chemical processing scheme cannot be proposed for any reactor without an intimate knowledge of all aspects of design and operation of the reactor. However, some of the basic chemical knowledge needed to evaluate various 304 CHEMICAL PROCESSING [crAP. 6 possible processing methods has been developed and is presented in the following sections. 6-2. CorE PRrROCESSING: SoLIDs REMOVAL 6-2.1 Introduction. Early in the study of the behavior of fisston and corrosion products in uranyl sulfate solutions at temperatures in the range 250 to 325°C, it was found that many of these elements had only a limited solubility under reactor conditions. Detailed studies of these elements were conducted and devices for separating solids from liquid at high temperature and pressure were constructed and evaluated. Based on this work, a pilot plant to test a processing concept based on solids separation at reactor temperature was installed as an adjunct to the HRE-2. These processing developments are discussed in this section. 6-2.2 Chemistry of insoluble fission and corrosion products. Of the nongaseous fission produets, the rare earths contribute the largest amount of neutron poison to a homogencous reactor after a short period of operation (Fig. 6-1). Therefore, a detailed study of the behavior of these elements TaBLE 6-1 SOLUBILITY OF LANTHANUM SULFATE IN 0.02 m U0:804,—0.005m HaSO4 as A FuNcTION OF SoLuTioON TEMPERATURE mg La2(S04)s/kg H20 Temperature, °C True Concentration required to solubility initiate precipitation 190 250 760 210 130 360 230 54 167 250 25 77 270 12 36 has been made. All the rare earths and yttrium showed a negative tem- perature coefficient of solubility in all the solutions studied and a strong tendency to supersaturate the solutions, as shown in Table 6-1. With the exception of prascodymium and neodymium, which are reversed, the solu- hility at a given temperature and uranyl sulfate concentration increased with increasing atomic number, with yttrium falling between neodymium 6-2] CORE PROCESSING: SOLIDS REMOVAL 305 and samarium, as shown in Table 6-2. Increasing the uranyl sulfate concentration increased the solubility of a given rare-earth sulfate, as shown in Table 6-3. TABLE (-2 SOLUBILITY OF VARIOUS RARE-IARTH SULFATES IN 0.02m U0s804—0.005m HoSO4 aT 280°C Solubility, Solubility, Salt mg/kg H20 Salt mg/kg H20 Las(804)3 10 Nd2(S04)5 110 Ce2(804)3 50 Y2(S04)3 240 PI‘2(804)3 170 SHlQ(SO4)3 420 TABLE 6-3 Errect oF URANYL SULFATE CONCENTRATION ON THE SOLUBILITY OF NEODYMIUM SULFATE AT VARIOUS TEMPERATURES Nd2(S04)3 solubility, mg/kg H20 U, g/kg H20 250°C 280°C 300°C 5.7 270 115 73 10.8 400 200 120 16.6 770 300 180 22 4 > 1000 o00 500 In a mixture of rare-earth sulfates the solubility of an individual rare earth 1s less than it would be if it were present alone. For example, the solubility of praseodymium sulfate at 280°C is 170 mg/keg H»0 with no other rare earths present, as compared with 12 mg,/ke H>0 in a solution made up with a rarc-earth mixture containing 6 praseodymiuin sulfate. Samples of the precipitating salts i1solated from solution at 280°C' have usually been the sulfates and contained no uranium. However, under special conditions a mixed sulfate salt of neodymium and uranium has been observed [3]. The alkaline earths, barium and strontiun, also show a negative tem- perature coeflicient, but not so strongly as do the rare earths; almost no effect can be seen when the temperature of precipitating solutions is in- 306 CHEMICAL PROCESSING [cHap. 6 creased from 250 to 300°C. At 295°C in 0.02 m U02504—0.005 m H2S04 solution, the solubility of barium sulfate is 7 mg/kg HoO and that of strontium sulfate is 21 mg/kg Ho0O. Both the alkaline and rare-earth sulfates show a strong tendency to precipitate on and adhere to steel surfaces hotter than the precipitating solutions, and this property can be used to isolate these solids from liquids at high temperatures. Other fission and corroston product elements hydrolyze extensively at 250 to 300°C and precipitate as oxides, leaving very low concentrations in solution. Iron(III) at 285°C has a solubility of 0.5 to 2 mg Fe/kg Ho0 and chromium(III), 2 to 5 mg/kg HoO. At 285°C less than 5 mg of zir- conium or niobium per kilogram of H20 remains in solution. For other elements of variable valence, such as technetium, the amount of the clement in solution is determined by the stable valence state under reactor conditions. In general, the higher valence states better resist hy- drolysis and remain in solution. Thus at 275°C in 0.02 m UO2S04 Te(VII) is reduced to Te(IV) if hydrogen is present, and only 12 mg/kg H20 re- mains in solution. However, a slurry of TcO» in the same solution but with oxygen present dissolves to give a solution at 275°C with a technetium concentration of more than 9 g/kg H»0. The same qualitative behavior is observed with ruthentum. Selenium and tellurium in the hexapositive state are much more soluble than when in the tetrapositive state [4]. A few eclements, e.g., cesium, rubidium, nickel, and manganese, intro- duced into the fuel solution by fission or by corrosion of the system, are very soluble under reactor conditions. Their removal and control are dis- cussed in Section 6-1. 6-2.3 Experimental study of hydroclone performance. It is evident from the preceding section that the amount of uranium withdrawn from the reactor diminishes if the collection, eoncentration, and 1solation of the insolubles can be effected at high temperature. One device capable of collecting and concentrating solids at high temperature is a solid-liquid cyclone separator called a “hydroclone,” or “clone.” A diagram of a hydro- clone is shown in Fig. 6-4. In operation, a solids-bearing stream of liquid is injected tangentially into the wide portion of a conical vessel. Solids concentrate in a downward-moving layer of liquid and are discharged from the bottom of the clone into the underflow receiver. Partially clarified liquid leaves from the top of the clone through a vortex finder. Use of the underflow recciver eliminates mechanical control of the discharge flow rate and, by proper choice of hydroclone dimensions, any desired ratio of overflow rate to underflow rate can be achieved. The driving foree for the system is provided by a mechanical pump. The factors influencing the design of an effective hydroclone for homo- geneous reactor processing use have been studied, and hydroclone designs 6-2] CORE PROCESSING: SOLIDS REMOVAL 307 t/Overflow ] ——-DO 1% Feed —-i Hydroclone | // Underfiow Port V5 3 £ e— Underflow Receiver Fig. 6-4. Schematic diagram of a hydroclone with associated underflow receiver. based on these studies have been tested in the laboratory and on various circulating loops [5]. All tests have shown conclusively that such hydro- clones can separate insoluble sulfates or hydrolyzed materials from liquid streams at 250 to 300°C. In the HRE-2 mockup loop a mixture of the sul- fates of iron, zirconium, and various rare earths, dissolved in uranyl- sulfate solution at room temperature, precipitated when injected into the loop solution at 250 to 300°C. The solids concentrated into the underflow receiver of a hydroclone contained 75% of the precipitated rare-earth sulfates. When the lanthanum-sulfate solubility in the loop solution was exceeded by 109, the concentration of rare earths in the underflow receiver was four to six times greater than in the rest of the loop system; some accumulation of rare earths was observed in the loop heater. A large fraction of the hydrolyzed iron and zirconium was collected in the gas separator portion of the loop. In the separator the centrifugal motion given to the liquid forced solids to the periphery of the pipe and allowed them to accumulate. Only about 109 of the solids formed in the loop was recovered by the hydroclone, and examination of the loop system dis- closed large quantities of solids settled in every horizontal run of pipe. 308 Reactor Cell Pump CHEMICAL PROCLESSING Underflow Pot Hydroclone Recombine: -Condenser 10 gal Separator Lsad e DZO Receiver g Decay Tank {2) Recombine: -Condenser HZO to Waste Separator 100 gal \ [cHAP. 6 Dissolver 7 gal @7\ Dissolver Solution DO Addition 67 »+ Fuel Addition Sampler o ] Carrier Transfer Tank Fie. 6-5. Schematic flow diagram for the HRE-2 chemical processing plant. TaBLE 64 Dmaensions oF HRE-2 HypROCLONES Dimension, in. Symbol Location . . _ 0.25-in. 0.40-1n, 0.56-1n. hydroclone hydroclone hydroclone D1 Maximum inside 0.25 0.40 0.56 diameter L Inside length 1.50 2.40 3.20 Dy Underflow port diameter 0.070 0.100 0.148 Do Overflow port diameter 0.053 0.100 0.140 Dy Feed port effec- tive diameter 0.051 0.118 0.159 2] CORE PROCESSING: SOLIDS REMOVAL 309 Samples taken from the loop after addition of preformed solids and without the hydroclone operating showed an exponential decrease in solids concen- tration with a half-time of 2.5 hr; with the hydroclone operating, the half- time was 1.2 hr. In the HRIE-2 chemical plant [5], operated with an aux- iliary loop to provide a slurry of preformed solids in uranyl sulfate solution as a feed for the plant, the half-times for solids disappearance and removal were 11 hr without the hydroclone and 1.5 hr with it. The efficiency of the hydroclone for separating the particular solids used in these experiments was about 109, With gross amounts of solids in the system, concentration factors have been as large as 1700. Correlation of these data with anticipated reactor chemical plant oper- ating conditions indicates that the HRE-2 chemical plant will hold the amount of solids in the fuel solution to between 10 and 100 ppm. If neces- zaryv. performance can be improved by increasing the flow through the chemical plant and by eliminating, wherever possible, long runs of hori- zontal pipe with low liquid veloeity and other stagnant areas which serve to accumulate solids. 6-2.4 HRE-2 chemical processing plant.* An experimental facility to test the solids-removal processing concept has been constructed in a cell adjacent to the HRI-2. A schematic flowsheet for this facility is shown in Fig. 6-5. A 0.75-gpm bypass stream from the reactor fuel system at 280°C and 1700 psi ig circulated through the high-pressure system, consisting of a heater to make up heat losses, a screen to protect the hydroclone from plugging, the hydroclone with underflow receiver, and a canned-rotor circulating pump to make up pressure losses across the system. The hydroclone is operated with an underflow receiver which is drained after each week of operation, at which time the processing plant is isolated from the reactor system. At the conclusion of each operating period 10 liters of the slurry in the underflow pot is removed and sampled. The heavy water is evaporated and recovered, and the solids are dissolved in sulfurie acid and sampled again. The solution is then transferred under pressure to one of two 100-gal decay storage tanks, Tollowing a three-month decay period, the solution 1= transferred to a shielded carrier outside the cell and then to an existing sulvent extraction plant at Oak Ridge National Laboratory for uranium decontamination and recovery. The sulfuric acid solution step is incor- porated n the HRIS-2 chemical plant to ensure obhtaining a satisfactory | — x — /4 Dr\OCe\ — 05 — gihca Gel-70 0.2 \at Gieves- AR —] Zeo-Dur MoleculS Micro Cel-A (1 1 1 [ [ [ | 0.1l 0 10 20 30 40 50 60 70 80 90 100 110 120 130 Krypton Pressure, mm Fic. 6-8. Adsorption of krypton on various adsorbents at 28°C. or further separated by conventional methods into an mert xenon fraction and a fraction containing Kr®5 and inert krypton. 6-3.2 Experimental study of adsorption of fission product gases. Evalu- atiou of vartous adsorber materials based on experimental measurements of the equilibrium adsorption of krypton or xenon under static conditions is in progress [7]. Results in the form of adsorption isotherms of various solid adsorber materials are presented in Fig, 6-8, A radioactive-tracer technique was developed to study the adsorption efficiency (holdup time) of small, dynamic, laboratory-scale adsorber systems [8]. This consists of sweeping a brief pulse of Kr8 through an ex- perimental adsorber system with a diluent gas such as oxygen or nitrogen 314 CHEMICAL PROCESSING [cHAP. 6 and monitoring the efluent gases for Kr®5 beta activity. The activity in the gas stream versus time after injection of the pulse of Kr®5 is recorded. A plot of the data gives an experimental elution curve, such as shown in Fig. 6-9, from which various properties of an adsorber material and ad- sorber system may be evaluated. Among the factors which influence the adsorption of fission product gases from a dynamic system are (1) adsorptive capacity of adsorber ma- terial, (2) temperature of adsorber material, (3) volume flow rate of gas stream, (4) adsorbed moisture content of adsorber material, (5) composi- tion and moisture content of gas stream, (6) geometry of adsorber system, and (7) particle size of adsorber material. The average time required for the fission product gas to pass through an adsorber system, fmax, 1s influ- enced by the first five of the above factors. The shape of the experimental elution curve is affected by the last two. The temperature of the adsorber material is of prime importance. The lower the temperature the greater will be the adsorption of the fission gases, and therefore longer holdup times per unit mass of adsorber material will result. The dependence of adsorptive capacity, k, on temperature as de- termined by holdup tests with some solid adsorber materials is shown in Table 6-5. TABLE 6-H ADsSoRPTIVE CaprAcITY OF VARIOUS MATERIALS AS A FUNCTION oF TEMPERATURE ce gas/g adsorbent*® Gas Diluent Adsorber 273°K 323°K 373°K Xe Oy Charcoal 4.7 x 103 4.0 x 102 80.0 Kr He Charcoal 1.8 x 10° 34 9.6 Kr Oz or N» Charcoal, 68 24 11.0 Kr O; Linde Molecular 23 9 4.5 Sieve SA Kr O Linde Molecular 11 5.7 3.5 Steve 10X *(Gas volume measured at temperatures indicated. 6-3] FISSION PRODUCT GAS DISPOSAL ald 1 | i Average Holdup Time (rmax) _l Kr85 Activity in Effluent Gas Stream —» Breakthrough Time (1} l I J Time After Injection of Kr85 — F1ac. 6-9. Experimental Kr83 elution curve. At a given temperature, the average holdup time, fyax, 18 inversely pro- portional to the volume flow rate of the gas stream. If the volume flow rate ix doubled, the holdup time will be decreased by a factor of two. All the solid adsorber materials adsorb moisture to some degree. Any adsorbed moisture reduces the active surface area available to the fission gases and thus reduces the average holdup time, The geometry of the adsorber system influences the relation between breakthrough time, f,, and average holdup time, ¢max, as shown in I'ig. 6-9. Ideally, for fission product gas disposal, a particular atom of fisslon gas should not emerge from the adsorber system prior to the time #,... Since this condition cannot be realized in practice, the difference between break- through and average holdup times should be made as small as possible. For a given mass of adsorber material a system composed of long, small- diameter pipes will have a small difference between ¢, and ., whereas a system composed of short, large-diameter pipes will not. The particle size of the adsorber material 1s important for ensuring n- timate contact between the active surface of the adsorber material and the fission gases. A system filled with large particles will allow some mole- 316 CHEMICAL PROCESSING [cHAP. 6 cules of fission gases to penctrate deeper into the system before contact is made with an active surface, while the pressure drop across a long trap filled with small particles may be excessive. Material between 8 and 14 mesh in size 1s satisfactory from both viewpoints. 6-3.3 Design of a fission product gas adsorber system. The design of an adsorber system will be determined partly by the final disposition of the effluent gas mixture. If ultimate disposal is to be to the atmosphere, the adsorber system should be designed to discharge only Kr®* plus inert krypton and xenon isotopes. If the effluent gases are to be contained and stored, the adsorber system may be designed to allow discharge of other radioactive krypton and xenon isotopes. In the following discussion it is assumed that final disposal of the effluent gas mixture will be to the at- mosphere. The following simple relation has been developed which is use- ful in finding the mass of adsorber material in such an adsorber system: fl’[ = ;,Fj’tmax, 14 where M = mass of adsorber material (grams), F = gas volume flow rate through adsorber system (cc/min), k= adsorptive capacity under dy- namic conditions (cc/g), and tnax = average holdup time (min). The operating characteristics of the reactor will dictate the composition and volume flow rate of the gas stream; fmax will be determined by the al- lowable concentration of radioactivity in the effluent gas; &k values for krypton and xenon must be determined experimentally under conditions simulating these in the full-scale adsorber system. It should be noted (Fig. 6-9) that a portion of the fission gas will emerge from the adsorber system at a time #, prior to the average holdup time, fmax. The design should ensure that radioactive gas emerging at time ¢ has decayed suffi- ciently that only insignificant amounts of activity other than Kr® will be discharged from the bed. The adsorber system should be operated at the lowest convenient temperature because of the dependence of adsorptive capacity on tempera- ture. Beta decay of the fission product gases while passing through the adsorber system will increase the temperature of the adsorber material and reduce the adsorptive capacity. Temperature control is especially eritical if the adsorber system uses a combustible adsorber material, such as activated charcoal, with oxygen as the diluent or sweep gas. 6-3.4 HRE-2 fission product gas adsorber system. The HRE-2 uses a fission product gas adsorber system containing Columbia G activated charcoal. Oxygen, contaminated with the fission produet gases, is removed 6-4] CORE PROCESSING: SOLUBLES 317 from the reactor, dried, and passed into this system, and the eflluent gases are dispersed into the atmosphere through a stack. The adsorber system contains two activated charcoal-filled units con- nected in parallel to the gas line from the reactor. Iiach unit consists of 40 ft of 4-in. pipe, 40 ft of 1-in. pipe, 40 ft of 2-in. pipe, and 60 ft of G-in. pipe connected in series. The entire system is contained in a water-filled pit, which is buried underground for gamma shielding purposes. lach unit is filled with approximately 520 Ib of Columbia G activated charcoal, 8 to 14 mesh. The heat due to beta decay of the short-lived krypton and xenon 1so- topes is diminished by an empty holdup volume composed of 160 {t of 3-1n. pipe between the reactor and the charcoal adsorber system. This pre- vents the temperature of the charcoal in the inlet sections of the adsorber system from exceeding 100°C. Iixecessive oxidation of the charcoal by the oxygen in the gas is further prevented by water-cooling the beds. Before the adsorber system was placed 1n service, its efficiency was tested under simulated operating conditions [Y]. A pulse of Kr®% (25 millicuries) was injected into each unit of the adsorber system and swept through with a measured flow of oxygen. In this way the krypton holdup time was determined to be 30 days at an oxygen flow rate of 250 e¢/min/ unit. Based on laboratory data from small adsorber systems, the holdup time for xenon is larger than that for krypton by a factor that varies from 30 to 7 over the temperature range of 20 to 100°C. Irom these data, it 1s estimated that the maximum temperature of the HRE-2 adsorber system will vary between 20 and 98°C after the reactor has been operating at 10 Mw power level long enough for the gas composition and charcoal temperature to have reached equilibrium through the entire length of the adsorber unit. The holdup performance of the adsorber system was cal- culated with corrections for the increased temperature expected from the fission gases. The calculated holdup time was found to be 23 days for krypton and 700 days for xenon; this would permit essentially no Xe!33 to escape from the trap. 6—4. CortE PROCESSING: SOLUBLES 6—4.1 Introduction. While the solids-removal scheme discussed i Sec- tion 6—1 will limit the amount of solids circulating through the reactor sys- tem, soluble elements will build up in the fuel solution. Nickel and man- ganese from the corrosion of stainless steel and fission-produced cesium will not precipitate from fuel solution under reactor conditions until con- centrations have been reached which would result in fuel instability and loss of uranium by coprecipitation. Loss of neutrons to these poisons would seriously decrease the probability of the reactor producing more § 318 CHEMICAL TPROCESSING [cHAP. 6 fuel than it consumes. Thercfore, a volume of fuel solution sufficient to process the core solution of the reactor at a desired rate for removal of soluble materials is discharged along with the inscluble materials concen- trated into the hydroelone underflow pot. This rate of removal of soluble materials depends on the naturc of other chemical processing being done and on the extentrof corrosion. For example, operation of an iodine re- moval plant (Section 6-5) reduces the buildup of cesium in the fuel to an insignificant value by removing .cesium precursors. 6—4.2 Solvent extraction. Processing of the core solution of a homo- geneous reactor by solvent extraction is the only method presently avail- able which has been thoroughly proved in practice. However, the amount of uranium to be processed daily is so small that operation of a solvent ex- traction plant just for core solution processing would be unduly expensive. Therefore, the core solution would normally he combined with blanket ma- terial from a thermal breeder reactor and be processed through a Thorex plant, but with a plutonium-producing reactor separate processing of core and blanket materials will be needed. These process schemes are discussed in detail in Sections 6-6 and 6-7. The uranium produect from either process would certainly be satisfactory for return to the reactor. Since solid fuel clement refabrication is not a problem with homogeneous reactors, decontamination factors of 10 to 100 from various nuclides are adequate and some simplification of present solvent extraction schemes may be possible. 6-4.3 Uranyl peroxide precipitation. A process for decontaminating the uranium for quick return to a reactor has been proposed as a means of reducing core processing costs. A conceptual flowsheet of this process, which depends on the insolubility of UO4 under controlled conditions for the desired separation from fission and corrosion products, is shown in Fig. 6-10. A prerequisite for use of this scheme is that losses due to the mmsoluble uranium contained in the solids concentrated in the hydroclone plant be small. Iowever, laboratory data obtained with synthetic solids simulating those expected from reactor operation indicate that the uranium content of the solids will be less than 19, by weight. Verification of the results will be sought during operation of the HRE-2. In the proposed method, the hydroclone system is periodically isolated from the reactor and allowed to cool to 100°C. The hydrolyzed solids re- main as such, but the rare-earth sulfate solids concentrated in the under- flow pot redissolve upon cooling. The contents of the underflow pot are discharged to a centrifuge where solids are separated from the uranium- containing solution and washed with D»0), the suspension being sent to a waste evaporator for recovery of Dz0. 6-5] CORE PROCESSING: IODINE 319 D0 D207 D20 D250, SU ‘\ fi* LW t SFC L SU | e uo, ‘ RE —— 100°C —— SFC—tm 0 C SFc—--l 3o°c ,......uo4—->1 100°C sy RE _ RE \ G IsFC ’ Llc—a -_— c O N w O > DQO RE D50 Recovery r——J SU = Soluble Uranium - SFC = Soluble Fission ~—H,O and Corrosion Products RE = Rare Earth Sulfates s IsU = Insoluble Uranium -5 FC o T Waste Storage IsFC = Insoluble Fission S FC and Corrosion Products RE Fig. 6-10. Schematic flow diagram for decontaminating uranium by uranyl peroxide precipitation. Uranium in the clarified solution is precipitated by the addition of either D202 or NaoOq2. By controlling pID and precipitation conditions, a fast settling precipitate can be obtained with less than 0.19 of the uranium remaining in solution. The U0y precipitate is centrifuged or filtered and washed with D20 and dissolved in 509 excess of 112804 at 80°C before being returned to the reactor. In laboratory studies uranium losses have been consistently less than 0.19, for this method and decontamination factors from rare earths greater than 10. Deecontamination factors from nickel and cesium have been 600 and 40, respectively. It is estimated that the product returned to the re- actor would contain about 20 ppm of sodium as the only contaminant in- troduced during processing. Although either the addition of D202 or use of D202 generated by radiation from the solution itself appears attractive, acid liberated by the precipitation of UO4 must be neutralized if uranium losses are to be minimized. Since the entire operation is done in a D20 system, no special precautions to avoid contaminating the reactor with ordinary water are needed. 6-5. Core Processing: lopine* 6-5.1 Introduction. The removal of iodine from the fuel solution of a homogeneous reactor is desirable from the standpoint of minimizing the biological hazard and neutron poisoning due to lodine and reducing the production of gaseous xenon and its associated problems. lodine will also *Contribution from S. Peterson, 320 CHEMICATL PROCESSING [crar. 6 poison platinum catalysts [10] used for radiolytic gas recombination in the reactor low-pressure system and may catalyze the corrosion of metals by the fuel solution. For this reason a considerable effort has been carried out at ORNI: and by Vitro [11] to mmvestigate the behavior of iodine in solution and to develop methods for its removal. In this regard, the iodine isotopes of primary interest are 8-day I'*! and 6.7-hr 1'35, 6-5.2 The chemistry of iodine in aqueous solutions. Iodine in aqueous solution at 25°C can exist in several oxidation states. The stable species are iodide ion, I7; elemental iodine, 12; iodate, 103 ; and periodate, 104~ or H5IO¢s. The last of these exists only under very strongly oxidizing con- ditions, and 1s immediately reduced under the conditions expected for a homogeneous reactor fuel. lodide ion can be formed from reduction of other states by metals, such as stainless steel, but in the presence of the oxygen necessary in a reactor system it is readily converted to elemental iodine; this conversion is especially rapid above 200°C. Thus the only states of concern in reactor fuel solutions are elemental iodine and iodate. Under the conditions found in a high-pressure fuel system the iodine is largely, if not all, in the elemental form. Volatility of todine. Since the volatile elemental state of iodine is pre- dominant under reactor conditions, the volatility of iodine from fuel solu- tion is the basis for proposed iodine-removal processes. The vapor-liquid distribution coeflicient [11] (ratio of mole fraction of iodine in vapor to that in liquid) for simulated fuel solution and for water at the temperatures expected for both the high-pressure and low-pressure systems of homo- geneous reactors 1s given in Table 66, TABLE 6-6 Varor-Liquip DisTriBUuTION OF lODINE Distribution coefficient, vapor/liquid Solution High pressure Low pressure (260-330°C) (100°C) Clean fuel solution 7.4 0.34 (0.02 m U02804—0.005 m HaS04— 0.005m CLISO4, 1-100 pp Is Fuel solution with mixed figsion and corrosion products 2.4 Water (pH 4 to 8, 1-13 ppm I2) 0.29 0.009 6-3] CORE PROCESSING: IODINE 321 TTTTTTTTI Tt T T T T Adsorber Bed Silvered Alundum | —— e — —— S I S G e \ P SRl SN SN SEnR SN Ejector Gas Separator —— (g Heater e | iquid] B - Sampler L lodine Spike —'6 Fic. 6-11. Vitro iodine test loop. Canned Rotor Pump A number of conclusions are evident from these data. Iodine 1s much more volatile from fuel solution than from water at either temperature, Fission and corrosion products appear to increase the volatility of 1odine from fuel solution at 100°C. Increasing the temperature from 100 to 200°C increases the volatility of 1odine relative to that of water. No systematic variation of 1odine volatility has been found with iodine concentration in the range 1 to 100 ppm or temperature in the range 260 to 330°C. The volatility of 1odine from simulated fuel solution has been verified by experiments in a high-pressure loop, shown schematically in Fig. 6-11 [11]. The circulating solution was contacted with oxygen in the ejector; the separated gas was stripped of iodine by passing through a bed of sil- vered alundum which was superheated to prevent steam condensation. Potassium iodide solution (containing a radioactive tracer, I'3!) was rapidly injected into the loop to give an 1odine concentration of 10 ppm. The lodine concentration decreased exponentially with time in the circu- lating solution. Table 6-7 gives the half-times for 1odine removal and the volatility distribution coeflicient, calculated from the removal rate and the flow rates, based on three experiments with clean fuel solution and two with added iron. Within the accuracy of flow rate measurement, the coef- 322 CHEMICAL PROCESSING [cHAP. 6 TABLE 6-7 lopiNE REmovaL FroM A HicH-PrESSURE Loor Todine Todine Tempera~ L \ i removal | distribution Solution ture, . .. % half-time, | coeflicient, min vapor/hquid ('lean fuel solution 230 13.0 7.6 (0.02 m U02304—0.005 m HeSO 34— 6.5 16.8 0.005 m CuS0y) 13.0 8.4 Fuel + 30 ppm Fe3* 220 11.0 10.9 Fuel + 300 ppm Fe?* 225 11.0 9.5 ficient agrees with the average value of 7.4 obtained in numerous static tests over the high-temperature range. Iron appears to have no effect. Oxidation state of 1odine at high temperatures and pressures. While 1odate ion is quite stable at room temperature, at elevated temperatures it de- composes according to the equilibrium reaction 4105~ 4+ 4HT === 21, 4 502 4 2H 0. The extent of this decomposition in uranyl-sulfate solutions above 200°C is not known with certainty, since all observations have been made on samples that have been withdrawn from the system, cooled, and reduced in pressure before analysis. Although the iodine in such samples is prin- cipally elemental, some iodate is always present, possibly because of re- versul of the iodate decomposition as the temperature drops in the sample line. Such measurements therefore give an upper limit to the iodate con- tent of the solution. If periodate is introduced into uranyl-sulfate solution. at elevated temperatures, it 18 reduced before a sample can be taken to detect its presence. Jodide similarly disappears if an overpressure of oxy- gen Is present, although iodide to the extent of 409, of the total 1odine has been found m the absence of added oxygen [11]. Methods that have been used for determining the iodine/iodate ratio in fuel solutions are (a) analysis of samples taken from an autoclave at 250°C at measured intervals after injection of iodine in various states [11], (b) analysis of samples taken from the liquid in liquid-vapor equilibrium studies at 260 to 330°C [11], (c¢) rapid sampling from static bombs at 250 to 300°C [12], and (d) continuous injection of iodate-containing fuel solution into the above described ejector loop at 220°C and determining 651 CORE PROCESSING: IODINE 323 oxidation states in samples withdrawn [11]. The iodine/iodate ratio in these samples has varied from slightly over 1 to about 70, with no apparent relation to variations in temperature, oxygen pressure, and total iodine concentration. The strongest indication of iodate instability was in the loop experi- ments, which gave the highest observed iodine/iodate ratio, even though iodine was continuously introduced into the flowing stream as iodate and removed by oxygen scrubbing as elemental iodine. The low iodate content of the samples from these experiments corresponded to a first- order iodate decomposition rate constant of 6.2 min~! Iodate con- tents averaging about 10% of the total iodine have been observed in 0.04 m U02804—0.005 m CuSO4—H2804 solution, rapidly sampled from a static bomb through an ice-cooled titanium sample line. The observed iodate content was unrelated to whether the free sulfuric acid concentra- tion was 0.02 or 0.03 m, whether the temperature was 250 or 300°C, and whether or not the solution was exposed to cobalt gamma radiation at an intensity of 1.7 watts/kg. Oxidation state of iodine at low temperatures. At 100°C the iodate de- composition and iodine oxidation are too slow for equilibrium to be es- tablished in reasonable periods of time. Thus both states can persist under similar conditions. In stainless-steel equipment both states are reduced to iodide, which is oxidized to iodine if oxygen or iodate is present {12]. In a radiation field the iodide is oxidized, iodine is oxidized if sufficient oxygen 1s present, and iodate is reduced [13]. At the start of irradiation, iodate is reduced, but in the presence of sufficient oxygen, iodine is later reoxidized to iodate, probably by radiation-produced hydrogen peroxide which aceumulates in the solution. Ifinally, a steady state is reached with a proportion of iodate to total iodine which is independent of total iodine con- centration from 1079 to 10~ % m and temperatures from 100 to 110°C, but strongly- dependent on uranium and acid concentrations and on the hydro- gen/oxygen ratio in the gas phase. When the temperature is increased to 120°C there is a marked decrease in iodate stability under all conditions of gas and solution composition. Experimental data on the effects of radiation intensity, temperature, and gas composition for the irradiation of a typical fuel solution containing 0.04 m T02504—0.01 m H2504—0.005 m CuSO4 are given in Ref. 13. The steady-state iodate percentages are also given in this reference. 6-5.3 Removal of iodine from aqueous homogeneous reactors. It is clear that under the operating conditions of a power reactor, iodine in the the fuel solution is mainly in the volatile elemental state. It can therefore be removed by sweeping it from the solution into a gas phase, stripping 324 CHEMICAL PROCESSING [cuaep. 6 it from the gas stream by trapping it in a solid absorber or by contacting the gas with a liquid. Numerous experiments have shown that silver supported on alundum is a very effective reagent for removing iodine from gas or vapor systems, although its efficiency is considerably reduced at temperatures below 150°C. Silver-plated Yorkmesh packing is very effective for removing iodine from vapor streams in the range 100 to 120°C. In one in-pile ex- periment [14] 909% of the fission-product iodine was concentrated in a silvered-alundum pellet suspended in the vapor above a uranyl-sulfate solution. This method of using a solid 1odine absorber, however, would present difhicult engineering problems, since xenon resulting from iodine decay would be expected to leave the absorber and return to the core unless the absorbers were isolated after short periods of use and remotely replaced. Iodine removal by gas stripping requires a continuous fuel letdown. In case this is not desirable, the vapor can be stripped of iodine in the high- pressure system by contacting with a small volume of liquid which is sub- sequently discharged. Liquids considered include water and aqueous solutions of alkali, sodium sulfite, or silver sulfate [11]. Although the so- lutions are much more effective iodine strippers than pure water, their use requires elaborate provision for preventing entrainment in the gas and sub- sequent contamination of the fuel solution. Thus most of the effort in design of iodine-removal systems is based on stripping by pure heavy water. One possible iodine-removal scheme uses Oz or Oz 4+ D2 stripping [15]. The iodine is serubbed from the fuel solution by the gas in one contactor and then stripped from the gas by heavy water in a second contactor. This water would then be let down to low pressure and stored for decay or proc- essed to remove 1odine. In most homogeneous reactors some of the fuel solution is evaporated to provide condensate for purge of the circulating pump and pressurizer. Since iodine is stripped from the fuel by this evaporation this operation can be used for iodine removal. This method, which is illustrated in Fig. 6-12, has been proposed for the HRE~3 [16]. Here a stream of the fuel solution is scrubbed with oxygen in the pressurizer. The steam is condensed and the oxygen recycled. The condensate is distilled to concentrate the 1odine into such a small volume that its letdown does not complicate reactor operation. Lodine removal in the HRE-2. lodine adsorption on the platinized alu- mina recombination catalyst, such as that used in the HRE-2, poisons the catalyst severely [10]. Although the catalyst can be restored by operation at 650°C, this would not be feasible in HRE-2 operation. A method for removing iodine from the gas stream by contact with alundum or York- mesh coated with silver was developed in the HRT mockup. Iodine was introduced into the system and vapor from the letdown stream and dump 6-5] CORE PROCESSING: IODINE 325 Oxygen Condenser 7.53 b/ min Condenser Pressurizer Holdup Tank To High Pressure D,OStorage 7.46 Ibfmin . I Core Solution 20 gal/min 536°F To Low Pressure To Reactor Core 596°F System 0,07 Ib/min Fia, 6-12. Todine removal system proposed for HRE-3. tank was passed through a silvered alundum bed and the recombiner, and then to a condenser. Condensate was returned to the high-pressure loop through a pressurizer and the circulating pump. After injection, the 1odine concentration of the high-pressure loop dropped from 1.8 mg/liter to 0.1 mg literin 2 hr. In similar experiments with silvered Yorkmesh, 1odine levels in the condensate and pressurizer were cven lower relative to the high-pressure loop. The Yorkmesh efficiency depended strongly on how densely 1t was packed. The iodine removal efliciencies calculated from these experiments and others are given in Table 6-8. In laboratory ex- periments with a I-in.-diameter bed which could not be tightly packed, Yorkmesh efliciencies were consistently poorer than those of silvered alundum. The ability of a bed of silver-plated Yorkmesh to remove iodine from the reactor system was apparently confirmed during the initial operating period of the HRE-2 [17]. Here the iodine activity in the reactor fuel appeared to be even lower than cxpected when iodine was removed at the same fractional rate as fuel solution was let down from the high- pressure system. Less than 3% of the iodine produced during 40 Mwh of operation was found in the fuel solution. Experience with the HRT mockup indicates that the iodine not in solution was held on the silvered bed. 326 CHEMICAL PROCESSING [cHAP. 6 TaBLE 6-8 Iopine REmovaL Erriciency orF SiLverep Beps 1N HRT Mockup Absorber Bed helght, Temp:arature, Efficiency, in. C Yo Silvered alundum 8 150 97.7 rings 8 120 81.0 5 110 64.0 Yorkmesh, 22 1b/ft3 10 120 97.0 Yorkmesh, 29 Ib/ft3 6 120 99.6 6—6. URANYL SULFATE BLANKET PRoCEsSING* 6-6.1 Introduction. The uranyl sulfate blanket solution of a plutonium producer is processed to remove plutonium and to control the neutron poisoning by corrosion and fission products. Although a modified Purex solvent extraction process can be used for plutonium removal, the method shown schematically in Fig. 6-3, based on the low solubility of plutonium in uranyl sulfate solution at 250°C, appears more attractive. A hydroclone similar to that used for reactor core processing is used to produce a con- centrated suspension of PuO2 along with solid corrosion and fission prod- ucts. The small volume of blanket solution carrying the plutonium is evaporated to recover the heavy water and the solids are dissolved in nitric acid. After storage to allow Np?3? to deecay, plutonium is decon- taminated by solvent extraction. 6—6.2 Plutonium chemistry in uranyl sulfate solution. The amount of plutonium remaining dissolved in 1.4 m U050, at 250°C is dependent on a number of variables, including solution acidity, plutonium valence, and initial plutonium concentration. Under properly controlled condi- tions, less than 3 mg/kg H20O has been obtained. Since plutonium is re- moved from solution by hydrolysis to PuQOg, solubilities are increased by increasing the acidity. Table 6-9 summarizes data on the solubility be- havior of plutonium for various acidities. *Contribution from R. E. Leuze. 6-6] URANYL SULFATE BLANKET PROCESSING 327 TABLE 6-9 SOLUBILITY OF TETRAVALENT PLUTONIUM IN 1.4 m UO2804 AT 250°C Excess sulfuric acid, Pu(IV) solubility, m mg/kg H20 3.7 17 39 68 105 oo = 0 D Plutonium behavior is difficult to predict because of its complex valence pattern. In the absence of irradiation, plutonium dissolved in 1.4 m U02804 under a stoichiometric mixture of hydrogen and oxygen at 250°C exists in the tetrapositive state. However, when dissolved chromium is present or when an overpressure of pure oxygen is used, part of the plutonium is oxi- dized to the hexapositive state. Experiments indicate that in the presence of Co% gamma irradiation [18], reducing conditions prevail even under an oxygen pressure and plutonium is held in the tetrapositive state. The valence behavior discussed here is somewhat in question, since actual valence measurements were made at room temperature immediately after cooling from 250°C. It is known that tetrapositive plutonium will dispro- portionate upon heating [19]. The disproportionation in a sulfate system is depressed by the sulfate complex formation with tetrapositive plu- tonium. These results indicate that plutomium in a reactor will be pre- dominantly in the tetrapositive state. When the plutonium concentration exceeds the solubility limit, plu- tonium will hydrolyze to form small particles of PuOs about 0.5 micron in diameter and in pyrex, quartz, or gold equipment forms a loose preci- pitate with negligible amounts adsorbed on the walls. However, if these solutions are contained in type-347 stainless steel, titanium, or Zirealoy, a large fraction of the PuOg adsorbs on and becomes incorporated within the oxide corrosion film. Attempts to saturate these metal surfaces with plutonium in small-scale laboratory experiments were unsuccessful even though plutonium adsorption was as much as 1 mg/em?. 6-6.3 Neptunium chemistry in uranyl sulfate solution. Neptunium dis- solved in 1.4 m UO2S804 at 250°C under air, stoichiometric mixture hy- drogen and oxygen, or oxygen is stable in an oxidized valence state, prob- 328 CHEMICAL PROCESSING [cHAP. 6 ably Np(V). The solubility is not known, but it is greater than 200 mg/kg H20. Since the equilibrium concentration is only about 50 mg/kg H0, for a 1.4 m UO2304 blanket with an average flux of 1.8 X 10'* neu- trons/(em?)(sec), all the neptunium should remain in solution in most reactor designs. 6-6.4 Plutonium behavior under simulated reactor conditions. Pluto- nium behavior in actual uranyl sulfate blanket systems has not been studied; however, small-scale static experiments with 100 ml of solution and circulating loop experiments with 12 liters of solution have been car- ried out in the absence of irradiation under conditions similar to those ex- pected in an actual reactor. In the static experiments, plutonium was added batchwise to 1.4 m U02804 at a rate of about 6 mg/kg HoO/day. The solution was heated overnight in a pyrex-lined autoclave at 250°C under 200 psi hydrogen and 100 ps1 oxygen. The solution was cooled to room temperature for analysis and for adding more plutonium. This was repeated until a total of 140 mg of plutonium per kilogram of water was added. Small disks of type-347 stainless steel were suspended 1n the solution throughout the experiment to determine the amount of plutonium adsorption. The behavior of plu- tonium for a stainless-steel surface area/solution volume ratio of 0.6 cm?2/ml is shown in Fig. 6-13. As the plutonium concentration was gradually in- creased to 45 mg/kg H20, essentially all the plutonium remained in solu- tion as Pu(VI). There was a small amount of adsorption, but no precipita- tion. During the next few additions the amount of plutonium in solution decreased rapidly to about 5 mg/kg Hz0. At the same time there was a rapid increase in plutonium adsorption and in the formation of a loose PuO, precipitate. All plutonium added after this was either adsorbed or precipitated. Other experiments were made with surface/volume ratios of 0.2 and 0.4 em?/ml. In all cases, the plutonium remaining in solution and the plu- tonium adsorption per square centimeter were essentially the same as that shown in Fig. 6-13. Thus, by decreasing the surface/volume ratio, it is possible to increase the amount of plutonium in the loose precipitate. For example, when the total plutonium addition was 130 mg/kg H20, 40% of the plutomium was as a loose precipitate for a surface/volume ratio of 0.6 em?/ml, 609% for a ratio of 0.4 em?/ml, and 689 for a ratio of 0.2 em?/ml. Plutonium behavior under dynamic conditions was studied by injecting dissolved plutonium sulfate and preformed PuQOg into a circulating stream of 12 liters of 1.4 m UO2S04 at 250°C under 350 psi oxygen. This solution was contained 1n a type—347 stainless steel loop equipped with a canned rotor pump, a hydroclone, metal adsorption coupon holders, and a small ti—t3] TRANYL SULFATE BLANKET PROCESSING 329 60 | l I | ! I ( \ 1.4 m U0,80, at 250°C 100 psiO4 4+ 200 psiH 2 2 9 Surface to Volume 0.6 cm “/ml p:- S o | \ \ In Solution, mg/kg HyO 3 W \ i ] \ i 0 | | 1 ! ] T T 1000 | ] 1 { | | { 100 |— Adsorbed, ng/cm? o [ 0.1 60 40 |— 20 In Precipitate, mg/kg H20 | | l l | | i 0 20 40 60 80 100 120 140 160 Total Plutonium Added, mg/kg H20 Fia. 6-13. Plutonium behavior in uranyl sulfate solution contained in type-347 stainless steel. pressure vessel that could be connected and removed while the loop was in operation. Plutonium was added and ecirculating-solution samples were taken through this vessel. Tetrapositive plutonium added to the circulating solution was completely oxidized to hexapositive i less than 5 min. When 45 mg/kg Ho0 of dissolved plutonium was added every 8 hr, the amount of plutonium circulating i solution increased to a maximum of about 150 mg/keg Ho0. As more plutonium was added, it was rapidly adsorbed on the loop walls. After the last addition of plutonium, the loop was operated at 250°C for several days. Twelve hours after the last addition the plutonium concentration had decreased to 100 mg/ kg H»>0, and about 40 hr later the amount of plutonium in solution had dropped to an ap- parent equilibrium value of 60 mg/kg H»0. Essentially all the plutonium removed from solution was adsorbed on equipment walls uniformly throughout the loop. Less than 0.19, of the plutonium was removed in 330 CHEMICAL PROCESSING [caAP. 6 the hydroclone underflow, and no precipitated solids were circulating. Even when 850 mg of plutonium as preformed PuQ» was injected into the loop, no circulating solids were detected 5 min later. Only 209, of this plu- tonium was removed by the hydroclone, 359 was adsorbed on the stainless steel, and the rest was distributed throughout the horizontal sections of the loop as loose solids. The hydroclone was effective for removing solids that reached it, but the loop walls and low veclocity in horizontal pipes were effective traps for PuO.. There are several differences in conditions between the loop runs and an actual reactor, the most important of which are probably the presence of radiation, the lower surface/volume ratio (0.4 compared with 0.8 cm?/ml for the loop), the slower rate of plutonium growth in the reactor (12 to 15 mg/kg H20/day) and the probability that a plutonium producer would have to be constructed of titanium and Zircaloy to contain the con- centrated uranyl-sulfate solution. Based on these laboratory results, how- ever, 1t appears that plutonium adsorption on metal walls may be a serious obstacle to processing for removal of precipitated PuQs. 6-6.5 Alternate process methods. Because of the problem of plutonium adsorption on metal walls, removal methods based on plutonium concen- trations well below the solubility limit have been considered. In a full- scale reactor plutonium will be formed at the rate of up to 12 to 15 mg/kg H20/day. In order to keep the plutonium concentration below 3 mg/kg H20, the entire blanket solution must be processed at least four to five times a day. By adding 0.4 m excess H2S04 (see Table 6-9), the plutonium solubility is increased to greater than 100 mg/kg H.O and the blanket processing rate can be decreased to once every 3 or 4 days. Slightly longer processing cycles can be used if part of the plutonium is removed as neptunium before it decays. Of the various alternate processes considered, ion exchange and ad- sorption methods show the most promise. Dowex—-50 resin, a strongly acidic sulfonic acid resin loaded with UO2" 1, completely removed tetra- positive plutonium from 1.4 m UO2504 containing 20 mg of plutonium per liter [20]. The resin capacity under these conditions, however, has not been determined. Because of the high radiation level it may not be feasible to use organic resins. Sorption of plutonium on inorganic materials shows some possibilities as a processing method {21]. Although rather low plu- tonium /adsorber ratios have been obtained, indications are that capacities will be significantly higher at higher plutonium concentrations. Special preparation of the adsorbers should alsc increase capacities. Attempts to coprecipitate plutonium with tri- or tetrapositive iodates, sulfates, oxalates, and arsenates were not successful, owing to the high solubilities of these materials in 1.4 m UO28504. 6-6] URANYL SULFATE BLANKET PROCESSING 331 r—=To Off Gas T -+ Irridiated Tharium Scrubber Additive Condenser Dissolvent — 7 } 1 Acid | e——Acid Catch Fractionator | 1 Dissolver > Waste e e —————— § | To Solvent | , | | Extraction . ] Reboiler te——Feed Adjustment Tank Recovered l f LACId Tank Fi1a. 6~14. Thorex process, feed preparation flowsheet. % - --—— First Cycle +———— Second Cycle — \-i ‘ I Aqueous Agueous Agueous Aqueous Aqueous Scrub Strip Scrub Strip Strip To Orgonic From Feed Preparation Solvent To Th Th Recovery First Cycle Th, U Extraction | Stripping Extraction Partitioning Stripping Reductant - Th iTh U Organic Organic | | Uranium x Extract- Scrub I Isolation Waste ont Sioc:cr:ge Evaporator Evaporator Il.'..-,. Th Product Waste F1c. 6-15. Thorex process, solvent extraction co-decontamination flowsheet. 332 CHEMICAL PROCESSING [cHAP. 6 -Third Cycle ——————————= e isolation —— - ——————— ! 1 Elutriant Elutriant Agueous - Scrub From Solvel_fl Effluent T Extraction to Waste ; Solvent Recovery Third Cycle Feed Preparation Organic Extractant U Sorption Column Extraction Column Th Removal Column Waste v or Storage To Storage Shigronen'r Fic. 6-16. Thorex process, uranium isolation and third cycle flowsheet. 6-7. TuoriuM OXIDE BLANKET PROCESSING 6-7.1 Introduction. At the present the only practical method available for processing irradiated thorium-oxide slurry is to convert the oxide to a natural water-thorium nitrate solution and treat by the Thorex process. Although this method is adequate, it is expensive unless one plant can be built to process thorium oxide from several full-scale power reactors. Therefore methods for ThO2 reprocessing which could be economically in- corporated into the design and operation of a single power station have been considered. Alternate methods that have been subjected to only brief scouting-type experimentation are discussed in Article 6-7.3. 6-7.2 Thorex process.* The Thorex process has been developed to sep- arate thorium, U233, fission product activities, and Pa?33; to recover the thorium and uranium as aqueous products suitable for further direct handling; and to recover isotopically pure U2 after decay-storage of the Pa233, The flowsheet includes two solvent-extraction cycles for thorium and three solvent-extraction cycles plus lon exchange for the uranium. Although only irradiated thorium metal has been processed, the process is expected to be satisfactory for recovery of thorium and uranium from homogeneous reactor tuels. The Thorex process may be divided into three parts: feed preparation, *(Contribution from W. T. McDuffee. 6-7] THORIUM OXIDE BLANKET PROCESSING 333 solvent extraction, and product concentration and purification. These three divisions are shown in Figs. 6-14, 6-15, and 6-16. In the feed preparation step, uranyl sulfate solution from the reactor core and thorium oxide from the blanket system, freed of D»0 and suspended in ordinary water, are fed into the dissolver tank. The dissolvent is 13 N nitric acid to which has been added catalytic amounts (0.04 N) of sodium fluoride. When short-cooled thorium is being processed, potassium iodide is added continuously to the dissolver to provide for isotopic dilution of the large amount of fission-produced 1'3! which is present. The dissolver solution is continuously sparged with air, and the volatilized iodine is re- moved from the off-gases in a caustic scrubber. The dissolver solution is transferred to the feed adjustment tank where aluminum nitrate is added, excess nitric acid recovered, and the resultant solution made slightly acid-deficient by evaporating until a temperature of 155°C is reached. During digestion in the feed adjustment tank any silica present is converted to a form that will not cause emulsion problems in pulse columng, and fission products generally are converted to forms less likely to be extracted by the solvent (4297 TBP in Amsco). In the solvent extraction step thorium and uranium are co-extracted in the first cycle; subsequent partitioning of thorium and uranium in the second cycle gives two decontaminating cycles to both products while using only five columns. For short-decayed thorium a reductant, sodium hydrogen sulfite, is continuously added to the feed streams of both cycles to decrease the effect of nitrite formed by irradiation. Without the sulfite addition, the nitrite formed by radiation decomposition of nitrates con- verts ruthenium to a solvent-extractable form. Aecid deficiency in the second cycle feed is achieved by adding dibasic aluminum nitrate (diban). The spent organic from the second cycle is recycled to the first cycle as the organic extractant. The spent solvent from the first cycle is processed through a solvent-recovery system and reused as the organic extractant in the second cycle. In the uranium product concentration and purification step (Fig. 6-16), uranium is isolated by ion exchange, using upflow sorption and downflow elution. In this way & concentrated uranium solution in 6 N HNO3 is ob- tained. This solution is stable enough for storage or is suitable as a feed for the third uranium extraction cycle. The third uranium cycle is a standard extraction-stripping solvent-extraction system using 15% TBP- Amsco as the organic extractant. Although installed as a part of the com- plete Thorex flowsheet, the third cycle may be used separately for re- processing long-stored uranium to free it of objectionable decay daughters of U232, When used as an integral part of the Thorex scheme, additional decontamination of the uranium is achieved and the nitrate product is well adapted for extended storage or future reprocessing. TaBLE 6-10 AVERAGE DECONTAMINATION FacTors rFor THorRIUM AND URANIUM PRODUCTS IN THE THOREX Piror PLANT Thorium irradiated to 3500 grams of mass—-233 per ton, two complete cycles for both uranium and thorium, one additional uranium cycle for material decayed only 30 days. Decontamination factors Gross Pa, Ru Zr-Nb Total rare earths I Thorium 400 days decayed 1 x 10° I x 104 4 x 103 3% 105 2% 108 — 30 days decayed 4 x 10* 7 X 10% 200 3 X 104 2 % 108 9 x 108 Uranium-233 400 days decayed 3 X 105 3 X 105 2 X 10° 8 x 10% 9 x 108 — 30 days decayed 5 x 107 5% 1010 4 x 108 7 X 1(® 3 X 108 3 x 107 ¥Ee DNISSTI0dd TVOIRAHD g "d¥HO] 6-7] THORIUM OXIDE BLANKET PROCESSING 335 For return to an aqueous homogeneous reactor the decontaminated uranium would probably be precipitated as the peroxide, washed free of nitrate, and then dissolved in D2SO4 and D20O. Product thorium would be converted to thorium oxide by methods described in Section 4-3. The adaptability of the Thorex flowsheet just described to processing thorium irradiated to contain larger amounts of U23% per ton and decayed a short time has been demonstrated in the Thorex Pilot Plant at Oak Ridge National Laboratory [22]. Fifteen hundred pounds of thorium irradiated to 3500 grams of U233 per ton and decayed 30 days was processed through two thorium cyecles and three uranium cycles. The decontamination fac- tors for various elements achieved with short-decayed material are com- pared in Table 6-10 with results obtained with longer-decayed material. While the decontamination factors obtained with the short-decayed ma- terial compare favorably with the factors for the long-decayed material, the initial activity in the short-decayed thorium was 1000 times greater than in the long-decayed. Therefore, while the thorium and uranium products did not meet tentative specifications after two complete cycles, the uranium product did meet those specifications after the third uranium cycle. Since the chemical operations necessary to convert these materials to forms suitable for use in a homogeneous reactor can be carried out re- motely, the products are satisfactory for return to a homogeneous reactor after two cycles. 6-7.3 Alternate processing method.* Attempts to leach protactinium and uranium produced in ThO» particles by neutron irradiation [23] in- dicate that both are rather uniformly distributed throughout the mass of the ThOs particle, and migration of such ions at temperatures up to 300°C is extremely slow. Since calculations show that the recoil energy of frag- ments from U233 fission is sufficiently large to eject most of them from a particle of ThO2 not larger than 10 microns in diameter, this offers the possibility of separating fission and corrosion products from a slurry of ThO. without destroying the oxide particles. Such a separation, however, depends on the ability to remove the elements that are subsequently ad- sorbed on the surface of the ThOs. Adsorption of various cations on ThO» and methods for their removal are discussed in the following paragraphs. Trace quantities of such nuclides as Zr® Nd'#7, Y°!, and Ru'®® when added to a slurry of ThO»s in water at 250°C are rapidly adsorbed on the oxide particles, leaving less than 10749 of the nuelides in solution. The tracer thus adsorbed cannot be eluted with hot dilute nitric or sulfuric acid. The adsorption of macroscopic amounts of uranium or neodymium on ThO; at 250°C is less for oxide fired to 1600°C than for 650°C-fired oxide, *Contribution from R. E. Leuze. 336 CHEMICAL PROCESSING [cHAP. 6 TABLE 6-1] ErrFect OoF CALCINATION TEMPERATURE ON URANIUM AND NEODYMIUM ADSORPTION ON THQs aT 250°C 0.5 g of ThO; slurried at 250°C in 10 ml of 0.005 m Nd(NO3)3 or 0.0bm U02804—0.05 m HzSO4. 1 * Calcination temperature, Adsorption, mg/g Th °C U Nd 650 3.3-4.4 7.4 850 1.9-2.4 6.1 1000 0.72-1.10 2.4 1100 0.08-0.19 0.5 1600 0.06-0.12 0.3 *Single numbers represent data from single experiments. In other cases the range for several experiments is given. TABLE 6-12 Use or PBO 10 DECREASE CATION ADSORPTION ON THO» 0.2 g of ThO; plus various amounts of PbO coslurried in 10 ml of solution at 250°C for 8 hr. . PhO/ThO, . Cation adsorbed on ThOs, Solids . Solution wt. ratio ppm ThO» 0.002 m UO2804 3100 PbO + ThO- 0.2 (0.002 m UO2804 220 ThOz 0.001m CE(N03)3 6200 PbO + ThO. 0.2 0.001 m Ce(NO3)s 10 ThO, (.01 m Nd tartrate 2700 PbO + ThO- 0.4 0.01 m Nd tartrate 10 6-71 THORIUM OXIDE BLANKET PROCESSING 337 as illustrated in Table 6-11. This change in amount of adsorption may be almost entirely due to decrease in surface area of ThO; with increased firing temperature. The surface area of 1600°C-fired ThO3 is only 1 m?/g ThO3,, while the 650°C-fired ThO» has a surface area of 35 m2/g ThO,. The cation adsorption on ThO;z can be decreased by coslurrying some other oxide with the ThO2. The added oxide must adsorb fission products much more strongly than ThO, and be easily separable from ThOz. The effectiveness of PbO in decreasing cation adsorption on ThO; is shown in Table 6-12. When PbO; was used, more than 999 of the cations added to the ThO2-PbO slurry was adsorbed on the PbOa. However, cations adsorbed on ThOgz were not transferred to PbO2 when it was added to slurry in which the cations were already adsorbed on the ThO, particles. Addition of dilute nitric acid to the ThO2-PbO coslurry completely dis- solved the PbO and the cations adsorbed on it without disturbing the ThOs-. In all cases, cations adsorbed on ThO: at 250°C are so tightly held that dilute nitric or sulfuric acid, even at boiling temperature, will not remove the adsorbed material. However, the adsorbed ions can be desorbed by refluxing the ThO» in suitable reagents under such conditions that only a small amount of 1600°C-fired ThOs is dissolved. Under the same treat- ment ThO» fired to only 650°C would be 90% dissolved. 338 CHEMICAL PROCESSING [cHAP. 6 REFERENCES 1. A. T. Gresky and E. D. Ar~NoLp, Poisoning of the Core of the Two-region Homogeneous Thermal Breeder: Study No. 2, USAEC Report ORNL CF-54-2-208, Oak Ridge National Laboratory, 1954. 2. E. D. Arnvouwp and A. T. Gresky, Relative Biological Hazards of Radiations Ezpected in Homogeneous Reactors TBR and HPR, USAEC Report ORNL-1982, Oak Ridge National Laboratory, 1955. 3. R. A. MoNErs and S. PETERsoN, in Homogeneous Reactor Project Quarterly Progress Report for the Period Ending July 81, 1955, USAEC Report ORNL-1943, Oak Ridge National Laboratory, 1955. (pp. 201-202) 4. D. E. Fercuson et al., in Homogeneous Reactor Project Quarterly Progress Report for the Period Ending Apr. 30, 1956, USAEC Report ORNL-2096, Oak Ridge National Laboratory, 1956. (p. 118) 5. P. A. Haas, Hydraulic Cyclones for Applicatton to Homogeneous Reactor Chemical Processing, USAEC Report ORNL-2301, Oak Ridge National Labora- tory, 1957. 6. W. D. BurcH et al., in Homogeneous Reactor Project Quarterly Progress Report for the Period Ending July 31, 1957, USAEC Report ORNL-2379, Oak Ridge National Laboratory, 1957. (p. 26) 7. R. A. McNEeEs ct al,, Oak Ridge National Laboratory, in Homogeneous Reactor Project Quarterly Progress Report, USAEC Reports ORNT-2379, 1957 (p. 137); ORNL-2432, 1957 (p. 147); ORNL-2493, 1958. 8. W. E. Browning and C. C. Bovta, Measurement and Analysis of Holdup of Gas Mixtures by Charcoal Adsorption Traps, USAEC Report ORNL-2116, Oak Ridge National Laboratory, 1956. 9. W. D. BurcH et al.,, Oak Ridge National Laboratory, in Homogeneous Reactor Project Quarterly Progress Report, USAEC Report ORNL-2432, 1957 (p. 23); ORNL-2493, 1958. 10. P. H. Harrey, HRT Mock-up Iodine Removal and Recombiner Tests, USAEC Report CF-58-1-138, Oak Ridge National Laboratory, 1958. 11. R. A. KxELER et al., Vitro Laboratories, 1957. Unpulblished. 12. S. PerERrsoN, unpublished experiments. 13. D. E. FErRGUsON et al,, Oak Ridge National Laboratory, in Homogeneous Reactor Project Quarterly Progress Report, USAEC Reports ORNL-2272, 1957 (pp. 133-135); ORNL-2331, 1957 (pp. 142-143, 148-149); ORNL-2379, 1957 (pp. 138-139). 14. R. A. McNEEs et al., in Homogeneous Reactor Project Quarterly Progress Report for the Period Ending Apr. 30, 1955, USAEC Report ORNL-1895, Oak Ridge National Laboratory, 1955. (pp. 175-176) 15. D. E. FErauson, Removal of Todine from Homogeneous Reactors, USAEC Report CF-56-2-81, Oak Ridge National Laboratory, 1956; Preliminary Design of an Todine Removal System for a 460-Mw Thorium Breeder Reactor, USAEC Report CF-56-7-12, Oak Ridge National Laboratory, 1956. 16. H. O. WEEREN, Preliminary Design of HRE-3 Iodine Removal System #1, USAEC Report CF-58-2-66 Oak Ridge National Laboratory, 1958. 17. S. PETERSON, Behavior of Iodine in the HRT, USAEC Report CF-58-3-75, Oak Ridge National Laboratorys 1958. REFERENCES 339 18. R. E. Leuze et al.,, in Homogeneous Reactor Project Quarterly Progress Report for the Period Ending Apr. 30, 1957, USAEC Report ORNL-2331, Oak Ridge National Laboratory, 1957. (p. 151) 19. R. E. CoxnNickg, in The Actinide Elements, National Nuclear Energy Series, Division 1V, Volume 14A. New York: McGraw-Hill Book Co., Inc., 1954, (pp. 221, 238-241) 20. 8. 8. Kigrsuis, Oak Ridge National Laboratory. Unpublished. 21. R. E. Levuze and 8. 8. Krrsuis, Oak Ridge National Laboratory, 1957. Unpublished. 22, W. T. McDvurrer and O. O. Yarsro, Oak Ridge National Laboratory, 1958. Unpublished. 23. D. E. Ferguson et al., in Homogeneous Reactor Project Quarterly Progress Report for the Period Ending July 31, 1955, USAEC Report ORNL-1943, Oak Ridge National Laboratory, 1955. (p. 221)