ORNL5143 ORNT.-51L3 Dist. Category UC-T64 Contract No. W-TL05-eng-26 CI-IEMICAL TECHNOLOGY DIVISION MEASUREMENT OF MASS TRANSFER COEFFICIENTS IN A MECHANTCALLY AGTITATED, NONDISPERSING CONTACTOR OPERATING WITH A MOLTEN MTXTURE OF LiF-BeFB-‘]th_ ANND MOLTEN BISMUTH . C. H. Brown, Jr, J. R. Hightower, Jr. J. A, Klein Date Published: November 1976 ,. OAK RIDGE NATTONAL LABORATCRY Oak Ridge, Tennessee 37830 operated by UNION CARBIDE CORPORATION for the : ENERGY ERESEARCH AND DEVELOPMENT ADMINISTRATTION REPRODUCED BY: m U.3. Department of Commerce National Technical Information Service Springfield, \firginia 22161 CONTENTS «ABSTRACT - . . [ ] - * . - - & .. - . - - - - - » * * 1. INTRODUCTION. . v ¢ & v o o 2 o « « = . - 2. EXPERIMENTAT. EQUIPMENT. . . « + o« + o « « o 2.1 Flow Diagram of Saglt-Bismuth System. . . . 2.2 Contactor Vessel . . . . . . . 0.3 Feed and Catch Tanks for Salt and Bismuth. 2.4 Treatment Vessel for Salt and Metal. . 2.5 5zZmPlers o v v v e e s 4 s e e e e . . 2.6 Freeze Valves and Lines. . . . . . . 2.7 Instrumentation and Control., . . . 2.8 Gas Purification and Supply Systems. . . . 3. EXPERIMENTAL PROCEDURES . & v « v & o . 3.1 Reductant Addition . . « « . . 3.2 Tracer Irradiation and Addition. 3.3 Run Procedld¥@. o v « ¢ o o+ o @ 3.L Trestment with Hydrogen-Hydrogen 3.5 Sample Preparation and Analysis. L, EXPERIMENTAL RESULTS. . « + « « & « . 5. INTERPRETATION OF RESULTS . . . . . . 6. CONCLUSIONS AND RECOMMENDATIONS . . . T. ACKNOWLEDGMENT. . - « . + « . « . 8., REFERENCES. . . + + o« v o o o « « « o APPENDIX A. v v v v « o o o o o o = « 4 APPENDIX B, . & ¢« v v « o« 4 =« o o o = « = APPENDIX C - . - - L] - - . " - - » - . - * APPENDIX D. & & ¢ v v v v o o o v o o o = APPENDTIX E. & v v 4 v v o« ¢« o o o o o o & =N o o M 10 12 1k ik 33 35 bk MBASUREMENT OF MASS TRANSFER COEFFICTENTS TN A MRECHANTCAITY - AGITATED, NONDISPERSING CONTACTOR OPRRATING WITH A MOLTEN MTXTURE OF IiF-BeF,-ThF, AND MOLTEN BISMUTH C. H. Brown, Jr. J. R. Hightower, dJr. J. A, Kiein ABSTRACT A mechanically agitated, nondispersing contactor in which molten fluoride salt and molten bismuth phases were contacted has been built and operated. The mass traansfer performance of the contactor was evalusted over a range of agitator speeds under conditions in which the major resist- ance to mass transfer was in the salt phase. The measured mass transfer rates were compared with rates predicted by literature correlations, The equipment necessary to contain the salt and bismuth at ~ A00°C is described along with the complete set of experimental data obtained during operation. 1. INTRODUCTION A molten-salt breeder reactor (MSBR) will be fueled with 2 molten fluoride mixture that will circulate continuously through the blanket and core regions of the reactor and through the primary'heat exchanger. Methods are being developed for use in a close-coupled processing facility for removing fission products, corrosion products, and fissile malerials from the molten fluoride mixture. The proposed MSBR processing plant is based on fluorinaticn to remove urénium, reductive extraction to remove protactinium, and the metal transfer process to remove the rare~earth fission products. The type of two-phase contactor being considered for the iatter two steps in the processing plant is a nondispersing, mechanically agitated contactor in which & molten-salt phase and molten-bismuth phase are contacted to effect the desired separation. A facility was instelled for measuring mass transfer rates across a salt-metal interface in a mechanically agitated, nondispersing contactor using a molten mixture of LiF-BeF,~ThF) (72-16-12 mole %) as the light phase, and molten bismuth as the heavy phase., Mass transfer rates for 237U and 2! 7r tracers were measured at nine different agitator speeds. The purpose of the experiments was 1o provide measurements of mass trans- fer coefficients in a fluoride salt-bismuth system with which existing correlations could be compared, and to provide data for developing new correlations for mass transfer coefficients which would allow large- scale contactors to be designed. Included in this report is a complete description of the experimental equipment, operating procedures, experimental data, and interpretation of the results. 2. EXPERIMENTAL EQUIPMENT Mass transfer rates between molten salt and bismuth in the mechani- cally agitated contactor were measured in steady-state experiments in which salt and bismuth streams flowed through the contactor. Concentra- tions of components which transferred between phases were measured in inlet and effluent streams. The equipment used to make these measure- ments consisted of the contactor véssel; feed and catch tanks for salt and bismuth; a vessel for purification of the salt and bismuth inventory: provisions for withdrawing samples of each phase from various locations; Treeze valves for salt and bismuth flow contrel; instrumentation for temperature, pressure, and gas flow measurement and control; and gas supplies and purification systems. A description of the equipment follows., 2.1 TFlow Diagram of Salt-Bismuth System A flow diagram of the system for flowing salt and bismuth streams through the contactor is shown in Fig. 1. Salt and bismuth were metered from the salt feed tank and the bismuth feed tank (vessels T-1 and T-3, respectively) by controlled pressurization of these tanks. The salt and *POTTBISUT J090B3U0D PalRLIIB ATTBOTUBYOSW I UG TM £3TTTORM UYSNOAUGMOTL YINWSTE-1TBY 9UT JO WeITeTp Mot T *STd TIASSIA LNIWLVEYL MNVL NOILLOZTTI02 MNVYL NOILLIOITI0D YOLOVLNOD 1AW ONV 1VS ANV (334 1vS dNV a33d VLAW .. | B ! f! | 5L “ mn#“?» |} LR | t I, Iy ! “ ” ““ i ] ! n | ! i -“ by b ] i ! oy R i i — — i i L v-nd g f1 e A | I i [ s i - 1 _ flll%dlnlill.w_ Y L- D | | = mnu,ul Qens — L-oT-go-mEE | | i S WAL 4 Hinwsig” H390YS 1Vs WP69-bL 9MA “INYO bismuth flowed to the contactor vessel; each phase entered below the surface of the contactor inventory of that phase and left the contactor through an effluent line at the salt-bismuth interface elevation. The interface thus was continuously renewed and mass-transfer inhibiting films were removed. The combined effluent stream was separated, and each stream flowed through a flowing stream sampler and then to the salt and bismuth catch tanks (vessels T-2 and T-4, respectively). The feed and catch tanks for each phase were concentric tanks to conserve space in the hood (see Sect. 2.3)., The salt and bismuth inventory could be sent to a graphite-lined treatment vessel (vessel T-5) for periodic treatment with H_-HF mixtures for removal of impurities and 2 adjustment of distribution coefficients. 2.2 Contactor Vessel A diagram of the contactor 1s shown in Fig. 2. The contactor was & 6-in. (152-mm)-diam low-carbon steel vessel conteining four l-in. (25-mm)-wide vertical baffles. The agitator consisted of two 2-7/8-in. (73-mm)-diam turbines with four 3/L-in. (19-mm)-wide straight blades. A 3/k-in. (19-rm)-diam overflow at the interface allowed the removal of interfacial films with the salt and metal effluent streams. Salt and bismuth were fed to the contactor below the surface of the respective phase. 2.3 Feed and Catch Tanks for Salt and Bismuth The duplex feéd and catch tanks for salt and bismuth were identical in construction. The feed tank, an inner cylinder of 8-in. sched 30 pipe, was designed to operate at pressures up to 50 psig (345 kPa) at 600°C, Both the inner feed tank and the outer catch tank had a capacity of about 20 liters of fluid; however, only about 15 liters of salt and 15 liters of bismuth were used. The top of each feed tank contained seven ports: (1) an inlet port (1/2-in. pipe with a fitting for 3/8-in. tubing), which did not extend into the tank; (2) an outlet line (1/2-in. pipe with a fitting for 3/8-in. tubing), which extended to within 1/2 in. (13 mm) of the ORNL DWG 76-584 AGITATOR DRIVE AND M SEAL ASSEMBLY BISMUTH |NLET>——9£::::§\ <« SALT INLET | AGITATOR - | | ]~ T SHAFT - ™1 [ SALT 1 / 'SURFACE\ / —~ TR - 1 U ; 2 h 14 —aSALT \ U; ( QUTLET N Lt | |Eh INT = — , ¢ — > BISMUTH T ) OUTLET FLAT-BLADED . 7 7 TURBINES ARy BAFFLES (4) Fig. 2. ©SOchematic diagram of mechanically agitated molten-salt— bismuth contactor. bottom of the tank; (3) a sparge and pressurization port (with a fitting for 3/8-in. tubing), which extended to within 1/2 in. (13 mm) of the bottom of the tank; (4) a 1/2-in. pipe (with a fitting for 3/8-im. tubing) used as a thermocouple well, which extended to within 1/2 in. (13 mm) of the bottom of the tank; (5) a 1/2-in. pipe with a fitting for a 1/2-in. ball valve and sampler and a fitting for 1/U-in, tubing below the valve; (6) a l-in. pipe with a 1-in. ball valve as an addition port; and (7) a 1/2-in. capped pipe as a spare port. Each catch tank had the same ports as the feed tanks except that no addition port was provided. The outer surfaces of the feed and catch tanks were flame sprayed with nickel aluminide to retard oxidation. 2.t Treatment Vessel for Salt and Metal The treatment vessel consisted of a 30LI, stainless steel pressure vessel that held a graphite crucible. The cylindrical portion of the pressure vessel was 26.5 in. (0.67 m) long [1/h-in. (6.L4-mm) wall thick- ness] by 18 in. (0.46 m) OD and with 18-in., (0.46-m)-OD by 1/lU-in. (6.4-mm)-thick dished heads on each end. It was designed to withstand HE-HF at 600°C at a pressure of 50 psig (345 kPa). The inner crucible, machined of graphite,® had an outer diameter of 16.75 in. (0.43 m) and was an overall 26.75 in. (0.68 m) high. The wall thickness tapered from 1.75 in. (44 mm) at the bottom to 0.75 in. (16 mm) at a point 16.75 in. (0.43 m) from the bottom, and was uniform from there to the top. The bottom of the crucible was 1.75 in. (44t mm) thick. The crucible had a 16.75-in. (0.43 m)-diam 1id, whose thickness varied from 1 in. (25 mm) at the rim tc 0,5 in. (13 mm) at the center. The graphite crucible rested on a support plate ingide the pressure vessel, and the 1lid was held loosely in position by three studs pro- jecting from inéide the top of the pressure vessel, The vessel had 13 nozzles, which are described in Table 1. *No. 8735, Speer Carbon Company, a Division of Air Reduction Company, St. Marys, Pennsylvania, Teble 1. Description of nozzles on treatment vessel Nozzlie Mo. Purposs Description i Bismrth charging 2-in. sched Lo pipe, flanged at the top o gccommodate o chute for loading bismuth. The greghite 1id below this nozzle hes & 1.625-in, - dism hole with & removable plug. 2 Bismuth sampling; salt 0.5=in. sched L0 pipe with bail valve and sampling; ges-phase szmpler. The 1id is fitted with & 1-in.-ID pressure connection graphite pipe into which the 0, 5-in, pipe slips, The graphite pipe extends through the gravhite 1id and into the crucible for a distence of 1 in, 3 Retwrning salt from the 0.5-in. sched 10 pipe nozzle conteining e galt rcceiver sleeved 0.375-in.-0D tube, PRelow the cerbon steel=-to-molybdenium transition, the 0.375-ixn.- 0D molybdenum tubing extends It in. belcw the graphite 1id. ly Returning bismuth from the Identical to nozzle No. 3. Dismuth receilver - 5 Transferring bismith to the 0.5=-in. sched L0 pipe nozzile containing a bismuth feed tank sleeved Q,370-in, 0D tube thet extends to within 0.5 in. of the bottom of the erucible. The tubing that extends into the crucible is made aof molybdenmum. 6 Transferring salt to the Similer to nozzle No. 93 set so that 15 liters galt feed tank of =21t can be transferred to the g2l feed tank, leaving 2 0.5-in. heel of salt on top of the bisruth. T Monitoring liquid levela Similsr to nozzle No. S. 8 Sparging with H,-HF gimilar to nozzle No. 5. 9 Adding salt Similay to nozzle No. 3. 10 Spare Similer to nozzle No. 3. 11 Thermocouple well 0.5-in. sched 40 pipe with fittings for 0.3735- in, 0D tubing. _ iz Meking miscellanecus addi- I1-in, sched LO pipe, with ball valve. tions, or vessel veniing . 13 Draining vessel C.5=in. sched 40 pipe extending from the hottom of the pressure vessel; this line is capped. | &'Acfis as a bubbler type of liquid-level monitor, 2,5 Samplers The treatment vessel and the feed and catch tanks were each pro- vided with a lfz—in. sched 40 pipe nozzle fitted with a ball valve and sample port. These tank sample ports held four sample capsules attached to capillary tubing that extended through a Teflon plug in the top. These capsules were lowered (while the system was under avgon pressure) through the ball valve into the tank below, and samples were drawn into the capsules by vacuum, In addition to the five sample ports on the vessels, there were twe Tlowing-stream sample ports that operated in a manner similar to that of the tank sample ports. These flowing-stream sample ports allowed seven samples to be taken from each of two flowing streams during operation. One sample port was located on the salt return line (between the con- tactor and the salt catch tank), and one was located on the metal return line (between the contactor and the metal cateh tank). The filtered sample capsules, which were used to take bismuth and salt samples, were made from 1/L-in. (6.L-mm)-diam stainless steel rod that was 3/h in. (19 mm) long. The sample capsules were fitted with a porous 347 stainless steel filter on one end and 1/16-in. (1.6-mm)-diam caplllary tubing on the other. Figure 3 shows a schematic diagram of a sample capsule and a typical tank sample port. 2.6 Freeze Valves and Lines Salt and metal flows through the facility were directed by four freeze valves in the transfer lines, located as indicated in Fig. 1. These valves were simply dips (in the carbon steel tubing) that were fitted with air cooling lines. Those freeze valves that had to be closed before any salt or metal could be transferred from the treatment vessel were equipped with small reservoirs (about 50 cmS) upstream and down- stream from the valve. The facility, which was of welded constructiofi, contained approximately 200 £t (A1 m) of salt and metal transfer lines (3/86~ and 1/2-in. pressure tubing). - PLASTIC TUBING fl iy b SAMPLERS 7 b TEFLON PLLEG VENT PURGE SAMPLE HOLDER BALL VALVE TOP GF VESSEL "'f/lfl]/ff’[’} FH I IT TP IITLT, -t et A Pl d 711 ;i et T Lt rnt Tt LT e ¢ it i L I : ! i L, TP T LTI Y e rrd T i Fig. 3. —=T0 ARGON AND VACUUHM SUPPLIES N = ORNL OWG 72-10408 ‘r,a-I/lfi-in. STAINLESS STEEL CAPILLARY TUBIMG, 40 in. LONG TYFICAL SAMPLER 3/4 in. 1 N ‘Q\\C\\Q:. 3/18-in. DRILL 1/4-in.-DIAN STAINLESS STEEL ROD FOROUS METAL FILTER, ¢0-p PORE SIZE, 347 STAINLESS STEEL Typical tank sample port and sample capsule. 10 2.7 Instrunentation and Control The principal objective of the instrumentation and control system was to provide closely reguleted flows of bismuth and molten salt to the contactor. The range of flow rates for both bismuth and molten salt was nominally L0 tc 500 ce/min, corresponding to experiment durations of about 5 to 0.5 hr. Pressures and liquid levels in the five vessels (treatment vessel and feed and catch tanks) of the facility were sensed by Foxboro differential-pressure transmitters, which sent signals to miniature pneumatic recorders or controllers. Liquid level was inferred from the pressure of the argon that was supplied to a dip-leg bubbler in each tank., Flow rates of bismuth and salt to the contactor were con- trolled by regulating the rate of change of liquid level in the two feed tanks., The feed and catch tanks, the treatment vessel, and the contactor were maintained at the desired temperatures by automatic controllers; transfer-line temperatures and temperatures of small components were controlled by manually regulating the appropriate voltage transformers that supplied power to Calrod tubular heaters. Figure 4 is a schematic diagram of the control system that regulated the flow of bismuth or salt to the coantactor. It was designed to circum- vent the flow-control problems that sometimes occur when gas pressure is used to maintain a constant flow of liguid from a heated feed tank. An adjustable ramp generator and an electric-to-pneumatic converter were used to linearly decrcase the set point of a controller that sensed liquid level in the feed tank, The level was controlled by controlling the flow rate of argon to the gas space of the feed tank. The result wes a uniformly decreasing liquid level and, hence, a uniform discharge rate of bismuth or salt from the tank. This control system was unaffected by small increases in back pressure, partial plugging of transfer lines, decreasing feed tank level, etc., or leakage of argon (a small argon bleed was provided to improve pressure control). Small gas pressure oscillations caused by temperature cycling was minimized by using time- proportioning controllers. Rates of transfer of salt and metal between the collection tanks and the treatment vessel were not required to be closely regulated; therefore, manual control of pressurization was used. ORNL DWG 7O-4357iRt FOXBORO ADJUSTABLE RAMP | _ | o verTer E/p |hrids RECORDER GENERATOR (E vst) CONTROLLER B 2% Al I - o o be A TR d P sl bl 4 5 / a e [Z4 £ Izl i ¥ " " CELL Y a ARGON BISMUTH QUT «@—— 50-500 ml/min. LEN TEMPERATURE RECORDER CONTROLLER (TIME PROPORTIONING) Fig. L. Schematic diagram of control system for metering bismuth from the pressurized feed tank, T-l. 12 Heating circuits were controlled manually for 11 transfer lines and the two flowlng stream samplers. On the transfer lines, the Calrods rated at 230 V were operated at 140 V or less, and provided up to 185 W per foot (600 W/m) of line, Typically, temperatures at three points on each line were recorded. The temperature of approximately 100 points was recorded for the system. 2.8 Gas Purification and Supply Systems Three gases were required for the experimental facllity: anhydrous hydrogen fluoride (HF), hydrogen, and argon. Because of the highly deleterious effect of small amcunts of oxygen or webter vapor, the nomi- nally pure bottled hydrogen and argon were further purified to remove traces of oxygen or water vapor. The anhydrous hydrogen flubride that was used only in the treatment vessel for hydrofluorination of the metal and the salt was given no additional purification. A schematic diagram for each of the three supply systems is shown in Fig. 5. Highly purified argon was used for all applications requiring an inert gas (e.g., pressurization of tanks for transferring bismuth and molten salt, dip-leg bubblers for liguid-level measurements, and purging of epparatus for sampling bismuth and galt). Cylinder argon with a minimm purity of 99.995% was first fed to & bed of molecular sieves (Fig. 5a), which reduced the water vapor content to about 2 ppm [-100°F (=73°C) dew point]. The argon then flowed through a bed of uranium metal turnings where the remaining oxygen and water vapor were removed. A porous stainless steel filter removed any uranium oxide dust that might have been carried from the uranium bed by the gas stream. The maximum argon flow rate, based on the capacity of the molecular sieve bed, was about 6 sefm (2.8 x 1073 std ma/sec). The hydrogen purification system was a commercially available deviceX that purified hydrogen by the selective diffusion of hydrogen across a *Serfass hydrogen purifier, product of Milton Roy Company, St. Petersburg, Florida. 13 ORNL OWG 70-280% /! {a) MOISTURE > MONITOR & X - 2 Cf’ ] MOLECULAR SIEVE BED | X URANIUM CHIF BED ~360 ¥ bl (WATER REMOVAL) A ! (OXYGEN REMOVAL) FILTER > ccth ~25°( ~E50°C [{max.) ARGON CYLINDERS {4} ARGON PURIFICATION SYSTEM ' IMPURITIES BLEED § {b) . T . i | SERFASS : ~45 e HYDROGEN : sefh PURIFIER (mex.) HYDROGEM CYLINDERS (2) HYDROGEN PURIFICATION SYSTEM {c) .~ Eib/hr | £ut t:‘:;'__:i . {max.) E—#HOT WATER SUPPLY L e PK-HUT AR BATH FOR CAFILLARIES HF TANK HOT WATER BATH PRESSURE RECORDER CONTROLLER HF SUPFLY SYSTEM Fig. 5. BSimplified diagram of the gas supply systems. 14 palladium alloy barrier. Impurities, along with a small flow of hydro- gen, were bled continuously from the upstream side of the barrier. The capacity of the unit wes 15 scfhi (1.2 x 107% std m3/sec). Controls for the purifier were self-contained. The anhydérous HF supply system utilized small capillaries for metering; a pneumatic controller maintained a specified pressure drop across a capillary by controlling the HF gas supply pressure. This was achieved by regulating the temperature of the water bath in which the HF supply tank was suspended (Fig. 5¢). Accidental overheating of the HF supply tank was prevented by a switch that released cold water into the bath if the temperature exceeded 60°C. The minimum flow range for the HF supply was nominally O to 0.25 1b of HF per hour (0 to 0.15 g/sec); the maximum range was O to 6 lb/hr (0 to 3.7 g/sec). 3. EXPERIMENTAL PROCEDURES In order to measure mass transfer rates in the eguipment previously described, 1t was necessary to perform several operations. The proper distridbution coefficient of the transferring materials was adjusted by adding reductant thorium and lithium to the bismuth. Tracers (237U‘and 972r) were prepared by irradiating 236U and 96Zr in the Oak Ridge Research Reactor (ORR) and these were added to the salt feed before each run. The salt and bismuth were fed {through the contactor. Samples were taken of salt and bismuth and were prepared for analysis. The salt and metal phases were treated periodically with mixtures of hydrogen and HF. Details of these procedures are described in this section. 3.1 Reductant Addition Pericdic ddjustment of the reductant inventory in the bismuth phase was necessary in crder to replenish reductant loss due to oxidation,. since even the high-purity argon which was used as a cover gas still contained a small amount of oxygen. The reductant inventory in the bismuth also required adjustment after H,-HF treatment of the salt and 2 bismuth. The method used for adjusting the reductant inventory was 15 electrolytic dissolution of beryllium ions in the salt phase while the .salt and bismuth were in contact in the treatment vessel. © A schematic diagram of the experimental apparatus used for electro- lytic beryllium addition is shown in Fig. 6. A 3/8-in. (8.2-mm) diam by 6~in, (152-mm) long berylliium rod was suspended in the treatment vessel and immersed in the salt phase. The beryllium rod was connected electri- cally to the positive terminal of a 12-V lead-acid storage battery vis wire and a stainless steel rod, which is insulated electrically from the treztment vessel. To couplete fihe circuit, the negative pole of the battery was connected to an ammeter, a variable resiétance, and finally to the treatment vessel. The bismuth phase served as the cathode in +this electrolysis. The overall reaction that tekes plaece in this electrolytic cell when current is pessed between the beryllium anode and bismuth cathode ig: 3 3 = o 5 Be” + U F_(salt) +—Be2+F2(salt) + U°(bismuth) . (1) 3( Thus, the electrolysis resulted in the oxidaticn of beryllium at the znode (the beryllium rod) and the reduction of uranium at the cathode (the bismuth surface). The reduced uranium dissolved in the bismuth and the concentrations of thorium and lithium dissolved in the bismuth adjusted to Sétisfy the equilibrium conditions that were reported previously.l 3.2 Tracer ITrradiation and Addition Mass transfer rates between the sali and bismuth phases were 23Ty or determined from the extent of transfer of and of Zr tracers that were dissolved in the salt phase prior to an experiment. The salt and bismuth were at chemical equilibrium with respect to the nonradiozctive uranium and zirconium, o7 The Zr tracer was prepared by irradiating a T.5-mg quantity of Q ’6Zr02 enclosed in a quartz ampul in the ORR for ~ 2L nhr. The 9TZrO2 was then transferred to a 0.75-in. (19-mm)-diam steel capsule after an *[9SSaA qUSWIBAIY YINWSIQq-1TBS 3Y3 ut eseyd 3TeS 9U3 0% WMTTTAISQ JO UOTAIPPE JIOJ OTHBWAYUDS °9 '8BT4 13SS3A INIWIVIYL HLNWNSIG-11VS aod WAWNITTAY3H Ad311vHd LNNSIE 171VS JOVHO1S daiov-avai 313NNV ¢ «§§¥T|l@|lw_ * dO1SIS3Y 318VIHVA I1I86-92 'OMA TINYO 17 3gs activity from the smpul) to 18-hr decay period (for decay of facilitate addition of the tracer to the salt phase. The capsule was then immersed in the salt phase in the salt feed tank while argon was sparged through the capsule to circulate galt through the capsuie and enhance mixing. ' | 236 Uranium-237 tracer was prepsred by irredisting v 1 mg U (as 236U308) encased in a quartz empul in the ORR for v T2 hr. As with the zirconium tracer, the uranium was then transferred to a steel =ddi- tion capsule and loaded into the salt phase in the salt feed tank. 3.3 Run Procedure A1l runs were performed using the same procedure. While the fluoride =salt and bismuth were in contact in the trestment vessel (T5), sufficient beryllium was added to the salt electrolytically to produce the desired distribution coefficient (D). Prior to a run, the salt and bismuth phases were separated by pressurizing the salt-bismuth treatment vessel and transferring salt and bismuth to their respective feed tanks. Approximately T mCi of ?Tzr-"Timo and 50 to 100 nCi of *>Tv,0g were alloved to diszolve in the salt phase sbout 2 hr prior to an experiment. Salt and bismuth streams were passed through the contactor vessel at the desired flow rates by controlled pressurization of the sgalt and bismuth feed tanks. The contactor was maintained at 520 to 600°C for gll runs. Both phases exited through é common effliuent line, separated, and returned to the salt and bismuth catch vessels. 3.4 Treatment with Hydrogen-Hydrogen Fluoride Mixture Periodic treatment of the salt and bismuth phases with HF-H2 mix- tures was necessary to remove oxides from the salt, and dissolved reductants and impurities (thorium, lithium, and iron) from the bismuth. The treatment procedure alsc served to adjust the equilibrium distri- bution of uranium and zirconium between the salt and bismuth phases. 18 The procedure followed was to sparge an HF—F2 mixture into the salt phase while both the salt and bismuth were in the treatment vessel (T5). In order to prevent excessive attack on containers and piping, the hydrogen fluoride concentration was kept below 30 mole % by dilution with hydrogen, although an attempt was made to keep the HF concentration as near 30% as possible to afford the maximum oxide removal rate. The total nominal flow rate was 30 scfh. The hydrogen fluoride flow rate was set by the pressure drop across a capillary, and the H, flow rate 2 was set by a rotameter which was calibrated with H From the treatment vegssel, the HF-H , stream passed through a sodium fiuoride bed (to remove HF) and then was exhausted to the atmosphere outside of Building 3592. The feed and off-gas from the treatment vessel were analyzed by diverting a small portion of the stream through a small agueous scrubber and a 0.05 ftB/revolution (1.4 x 10_3 m3/revolution) wet-test meter, which were connected in series. The concentration of HF in the gas stream was determined by passing the gas through 250 ml of a 0.k N NaOH solution in the scrubber. When 0.05 £65 (1.1 x 1073 m3) of H, had passed through the wet-test meter, the gas flow was stopped and the solution was removed for analysis. The HF concentration in the gas was determined by titrating small samples of the scrubbing solution with 0.1 N HC1. Utilization of the HF was calculated from the feed and discharge concentrations. 3.5 Sample Preparation and Analysis In an effort to avoid contamination of the samples obtained in each run with extraneous material, the sample capsules were cleaned of foreign material before analysis by the following procedure. Gross amounts of salt or bismuth were first removed with a file, the sample capsule was then polished with emery cloth, and, finally, the capsule was washed with acetone. The samples were analyzed by first counting the sample capsules for the activity of 51U (207.95 keV B~) and the activity of ° |Zr-> Nb (743.37 keV and 658.18 keV B, resvectively) after secular equilibrium was reached between QTZT and its daughter 97Nb. The material in the 19 237 sample capsules was then dissolved, and the activity of U was counted 97Zr~97Nb activity had decayed toc a very low level. This again after the was done to correct for self-absorption in the solid samples. Mass transfer rates were than calculated from the ratios of tracer concentrstions as discussed in Appendix B. Iy, EXPERIMENTAL RESULTS Nine runs were made in the experimental equipment to measure rates of mass transfer of 23TU and o1 Zr between selt and bismuth. In these runs, +the agitator speed was varied over the range of 68 to 24k rpm, and the operating temperature was held in the range of 590 to 600°C. Concentra- r Mgy and 237U were messured (as deseribed in Sect. 3.5) in the tions o salt input and both the salt and bismuth effluent streams. The counting dats obtained in a2ll runs are given in Appendix A. Using these concen- trations, three different equations were used to calculate the mass transfer coefficient between_the salt and biémnth in the contactor. The derived equations are given in Appendix B. The calculeted mass transfer coefficients for all the runs are summarized in Table 2. The .values given in Table 2 are the average of the values calculéted from the three equetions [Egs. (B-18)-(-20)] with the standard devietion. Values are given both for results based upon the uranium counting and for results based upon the zirconium counting. Run TSMC-1 was mainly a preliminary run designed to test the procedure, Salt and bismuth flows were approximately 200 ce/min, and the stirrer rate was 123 rpm. Unfortunately., the distribution coefficient (defined in Appendix B) was too low to effect any significant mass transfer, and mess transfer rates could not be determined accurately; thus, no results are shown for this run in Table 2. Operation of the equipment during run TSMC-2 was very smooth. The salt and bismuth flow rates were 228 and 197 ce/min, respectively. The ‘distribution coefficients were higher than for the previous run, but were lower than desired, resulting ian much uncertainty. Several determinations of the distribution coefficient of uranium D_ were made that ranged from ) C.9Lh to 3L, One determination was made of the distribution coefficient 20 .Afio\mo.l HV = PIJIISDISUBIY JI30BIY GOfipomhmd -—- gOT°0 + T2T'O 76°0 - I < he 19T 69T 6-0DHSL - 0T00*0 + 2200°0 G20 -- of < 0~ 79T 2sT §—OWST ——= 2T00°0 + LS00°0 o0 -- L6 < 89 OLT 2sT L~OWST TO°0 + 020°0 G00°0 + 6£0°0 719°0 2 gLt < 09T ¢eT 902 9=OHSL 6STO0 + €9TO0 £T00°0 + $600°'0 GE*0 e €n < LTA GLT 612 G—OWST, 20°0 + G&£0°0 200 + H®S0'O gL'0 e gL < S02 T 0LT H1—OWSL -—= €00°0 + 2T0°0 0$°0 - € < 29T €LT 991 € -OWST ¢G00'0 + £800°0 g2600°0 = 6500°0 LT°0 960 HE-46°0 12T L6T gee 2—OWSL IMTUODITZ wniue.ln pPSII8I SUBLYL MNQ Dm Aamhv {(utm/29) Anfla\uov uny uo poseq U0 paseg I90BIY 998X MOTJI Ujnmstg MOTJI 371BS Aomm\.fiov mum TUOTFOoOBILL I3JIATIG J040BIUOCD YLnUSIQ-4TeS 23 UT SqUallsINSeBSW J9JSUBLY S8BW JO sSyTnsad TeauswTJIodxy 21 of zirconium DZr-that indicated Dy.. was 0.96. Consequently, only a range of possible values for the overall mass transfer coefficient could be stated for the results based on uranium. A value for overgll mass transfer coefficient based on zirconium is given, but since the value for DZr is uncertain, there is more uncertainty in the mass transfer coefficient than is indicated by the reported standard deviation given in Table 2. A bismuth line leask occurred immediately preceeding run TSMC-3. During the resulting delay for repairs. the 9TZr decayed and only the 23TU tracer could be used. The remainder of the run went smoothly. Salt and bismuth flow rates were 166 and 173 cc/min, and the stirrer rate was 162 rpm. A high value for Dy (greater than 34) was maintained for this - run. Tn run TSMC-L, flow rates of 170 and 1L4 cc/min were set for the salt and bismuth flows, and a stirrer rate of 205 rpm was maintained. The distribution coefficients, which had been determined from samples taken before, after, and during the run, were greater than 172 for DU and greater than 24 for DZr' These values are sufficiently large so that Eqs. (B-18)-(-20) in Appendix B are valid. lLarge distribution coefficients cannot be determined precisely due to the inability to determine very smgll amounts of uranium in the sallt phase. No problems arose during this run. Runs TSMC-5 and -6 were performed without incident. The distribution coefficients were maintained at high lefiels for both runs. TSMC-5 had a stirrer rate of 12U rpm and salt and bismuth flows of 219 and 175 ce/min., TSMC-6 selt and bismuth flows were 206 and 185 cc/min, respectively, with a stirrer rate of 180 rpm. Prior to run TOMC-T, two leaks developed in the blsmuth transfer line from the bismuth feed tank to the contactor vessel. This transfer line was completely replaced slong with the associated Calrod and thermal insulaticn. The volumetric flow rates of salt and bismuth during the run were 152 cm3/min and 170 cmS/min, respectively. The stirrer rate was set at 68 rpm for the run. The uranium distribution coefficient was greater than 97. Seven sets of salt and bismuth flowing stream samples were taken from the contactor effluent during the run,. 22 Run TSMC-8 was performed with salt and bismuth flow rates of 152 cc/ min and 164 cc/min, respectively. The uranium distribution coefficient was maintained at a level (> 40) for this run which was greéter than the minimum desired value of 20. It was presumed that the agitator operated at 241 rpm, which is high enough to produce mild dispersion of the phases in the contactor and, therefore, a high measured mass transfer rate. However, results from this run indicated that very little (v 25%) of the 237U tracer was actually transferred from the salt to the bismuth. Inspection of the magnetically coupled, agitator drive assembly indicated that an accumulation of a highly viscous carbonaceous material between the upper carbon bearing and the agitator drive shaft had prevented proper rotation of the shaft. The drive assembly was cleaned of all foreign material, was reassembled, and was found to operate satisfactorily. The ninth tracer run, TSMC-Q, was performed as a repeat of the eighth run. Salt and bismuth flow rates were set at 169 cc/min and 164 ce/min, respectively. The agitator was operated at 2L4 rpm during this run. A high stirrer rate was maintained to determine the effects on the mass transfer rate of dispersal of one phase in the other, and to determine if large amounts of bismuth and salt are entrained in the other phase after settling in the contactor effluent line, Entrainment results from this run and another similar run are given in Appendix D. The uranium distribution coefficient was greater than 47 during this run. No systematic problems were encountered and the run was performed smoothly. 5. INTERPRETATION OF RESULTS In this section, the mass transfer coefficients measured in the salt- bismuth system are compared with typical mass transfer coefficlents measured in aqueous-organic systems at comparable agitator diameters and speeds and with mass transfer coefficients measured in a water-mercury system. The mass transfer coefficients measured in this study are also compared to predictions made by three mass transfer correlations taken from the literature that were developed from data measured in aqueous- organic systems. 23 EBTU given in Teble 2 &are prcobably T The mass transfer coefficients for more reliable than the results given for 237 U was greater than 80%, whereas that for 9TZr was Zr because, In all cases, the maeterial balance for consistently about 60%. Because of this, and also because more useful data points were obtained for 237U than for o Zr, the interpretation of results presented in this section is based mainly on the 237U measurements. The magss trangsfer coefficients for 237 U are compared in Fig. 7 with some typical mass transfer coefficients measured in aqueous-organic systems and with water-side mass transfer coefficients measured in a water-mercury 39'1*!' systemn. The figure shows that the mass transfer coefficients measured in the salt-bismuth system (curve A) are quite high compared to the agueous- organic and water-mercury results measured in cells of comparable size and at comparable agitator speeds (curves B through E). Except for curve E all the mass transfer coefficient data can be divided into two regimes: (1) at low agitator speeds the mass transfer coefficient is proportional to the agitator speed raised to a power'less than 1.0 (0.9 for the salt- bismuth results and 0.7 for results represented by curves.BAthrough E): -and (2) at higher agitztor speeds the mass transfer coefficient is pro- portional to the agitator speed raised to a power significantly grezter than 1.0 (1.5 for curve D, 1.95 for curves B and ¢, and 3.0 for curve A). Olander asnd Benedict2 suggest that the sudden change in expofient for their data (in the absence of phase dispersal) is caused by a laminar-turbulent transition at the interface. Observation has shown that phase dispersal did not occur at the break points of curves B, C, and D, but it was not possible to determine unequivocally when dispersal occurred in the salt- bismuth system since measurements of entrainment were Incconclusive cn 5,6 this point (see Appendix D). However, previous work with water-mercury and with agueous-organic systems Indicates that, for the agitator diameter used, dispersal of molten salt into bismuth should begin to cccur at an agitator speed of about 170 rpm. This entrainment would cause the apparent mass transfer coefficient to be greater than the mass Transfer coefficient that would have résulte& if phase dispersal had not occurred. The increase would be due to the increased arez for mass transfer. since the apparent mass transfer coefficient is based upon the area of The undisturbved inter- face. Since dispersal was expected, and because the dependence on agitator 2l speed is so different from the aqueous-organic data and the water-mercury data, it is concluded that dispersal of salt into bismuth occurred at an agitator speed between 160 and 180 rpm, even though entrainment measure- ments do net support this. 91 The mass transfer coefficients for Zr are shown in Fig. 8 compared to curve A from Fig. 7. In all but two cases (runs TSMC-2 and —5), the zirconium mass transfer coefficient was lower than the uranium mass trans- fer coefficient. This difference is probably related to the inability to correct for the self-absorption of the Ti3.37 keV B~ in the analysis of 97Zr in the solid bismuth samples. In run TSMC-2, all the resistance to mass transfer of uranium was in the salt phase, whereas there was sig- nificant resistance to mass transfer of zirconium in both phases. Never- theless, the overall mass transfer coefficients for uranium and zirconium were of comparable magnitude in this run. The mass transfer coefficients were compared with three literature correlations for mass transfer coefficient in stirred cells that were developed for aqueous-organic systems. The properties of the fluoride salt that were used to evaluate these correlations are given in Appendix D. The properties of bismuth at 600°C that were used are: s = Q.66 g,/cm3 and Np; = 1.0 x 1072 g/cm-sec. Lewis7 investigated mass transfer rates in mechanically agitated, nondispersing contactors, all of the same size, using several agueous- orgahic systems. He fitted his results with the following empirical equation: 60 k, 6 N, 1.65 = 6,76 + 10 Re. + Re,, — + 1, (1) V , 1 2 1N 1 1 where I\TL2 Re = Reynoclds number —E—E- ) = stirrer speed, rps , = gtirrer length, cm , = mass transfer coefficient, cm/sec , 25 ORNL DWG.Fe-582 ledl Ix 109 % 1G9 MASS TRANSFER COEFFICIENT (cm/sec) 1 { 1 1T 1T iT 237 U TRANSFER IN FLUORIDE SALT (B HEAVY PHASE) 73 mm AG!TATOR QUINONE TRANSFER IN WATER {(REF. 3,4) (Hg HEAVY PHASE) 89 mm AGITATOR QUINONE TRANSFER IN WATER (REF. 3,4) (Hg HEAVY PHASE)E4 mm AGITATOR WATER IN 20% TBP-70% n-HEXANE (REF 2) (WATER HEAVY FHASE) ?6 mm AGITATOR WATER IN ISOCBUTANCL (REF. 2) (WATER HEAVY FHASE) 76 mm AGITATOH llllll el i (x 107 . Fig. T. water-mercury coefficients. Comparigon of ! | 20 40 €6 80 100 200 400 600 AGITATOR SPEED {rpm) . 237 to run number in the salit-bismuth experimental facility. IOOO U mass transfer coefficients in fluoride salt with representative agueous-organic mass-transfer coefficients and Numbers in parentheses on top curve refer 26 ORNL DWG. 76-578 1x10"! = N w o o £ L Q0 e 01(4) w N o w w - 0 (8) 3 0(s) o L - » Ix10 2 i o(2) o L - wn L N I = L 2,"0'3 1 I 1 [ I i 1 1 1 0 20 40 60 80 100 200 400 600 800 1000 AGITATOR SPEED (rpm) Fig. 8. Effect of agitator speed on salt-side mass transfer coef- ficient of 9T%r in the salt-bismuth contactor. Numbers in parentheses refer to run number. 27 p = density, g/cm3 . n = viscosity, g/em sec , . = kinematic viscosity, n/p. cm?/sec, and 1.2 = phase being considered . For the case in Which,Nl = N2 and Ll = L2, the above equation can be reduced to the form: 60 k & L = 6.76 x 10” Re. | 1 + -2 +1 . (2) Ul 1 pl For Lewis' work, where the densities of the various phases varied from 0.8 to 1.2 g/cm3 but the stirrer length was kept constant, this correla- tion effectively uses only the Reynolds number of the phzse being con- sidered. The wranium mass transfer coefficients are compared to the Lewls correlation in Fig. 9. At agitator speeds bélOW’lTO rpm, the Lewis corre- lation overprediclts the mass transfer coefficient for uranium; it zalso shows a stronger dependence of mass transfer coefficient on the agitator speed than the data indicate., No dependence of molecular diffusivity is shown by fihis correlation. The omission of a term contasining molecular diffusivity from the correlation has been criticized in the recent liter- 8,9 9 sture. It has been shown theoretically and in recent experiments that the mass transfer coefficlent should be proportional to molecular diffusivity raised to a power near 1/2. Mayerslo developed a slightly more involved correlation of the following form: 171 i/2 (N2 Mo ~1/6 S = 0.1896 (Bel Reg) 0.6 + = (scl) . (3) 1 Ny 1 When both phases are stirred with identical paddles at the same speed, this equation reduces to: 28 ORNL DWG 76-583 1000 ! v - 00 — T’ - x_ — -y gl - - ~ o(3) . o(5) =T o7 - l L 1 1 1 Lt 1' 1 i 1 i L it ll 10 100 000 n .68 676 x 107 [ Re, + Re, (=2 )] i Fig. 9. Comparison of uranium mass transfer coefficients with the Lewis correlation. Numbers in parentheses refer to run number. 29 L - | - 60 kL n, 1.4 n, 2.h 6o 0.5 ——— = 0.1896 (Re;) | == 0.6 + —= (5c,) (L) 1 My Ny \P1 wnere S¢ = Schmidt number, n/p& . This correlation, which is based on data. covering a limited renge of den- sities (0.8 to 1.2 g/cm3), indicates that the viscosity and dénsity of each phase affect the mass transfer ccoefficient. A comparison of the uranium mass transfer coefficients with the Meyers correlation is shown in Fig. 10. At agitator speeds below iTO rpm, the Mayers correlation also overpredicts the mass transfer.coefficient, but the predicted dependence of mass transfer coefficient on agitator speed is more nearly in accord with the data than it is in the Lewis correlation, Note that the dependence on the Schmidt number (molecular diffusivity) is fairly weak. McManamey ™ correlated his data and the results obtained by Lewis by using the following expression, whilch i1s similar to that used by Lewis but includes the Schmidt number: 60 k N. Re _I — —= = 0.0861 (Rel)0'9 1+ -2 fi (se, ) 0.37 (5) 1 N & where S¢ = Schmidt number, n/p® , and ® = diffusion coefficient, cm?/sec. This equation can be reduced ta: 60 k | _ P o A L - 0.0861 (Re )O'9 1+ = (Sc )‘O'37 . ' (6) 1 1 Py . Vv for the case where Ll = L2 and Nl = Ng. Note that the numérical constant in Eqgs. (5) and (6) has the dimension of reciprocal centimeters. 30 ORNL DWG 76 - 580 1000 | 1 - _ r-— e~ ® (9) 100 4 - ®{4) ] n : ®(6) - —.l - 4 - i co - g ®(3) ® 10— (5) . . ®(7) J L_ — | o ' | ! 0 100 oums( e, e, )* (221" (064 387" (sc, 7™ Fig. 10. Comparison of uranium mass transfer coefficient with the Mayers correlation. Numbers in parentheses refer to run number. 31 The uranium mass transfer coefficients are compared with the.MbManamey correlation in Fig. 11. - This correlation shows good,agreementlwith the data at agitator speeds below 170 rpm. It must be pointed.out, however, that the Schmidt number for diffusion of uranium in molten fluoride selt was estimated by using correlations based upon materials with solution behavior that is gquite different from molten salt solutions. Hence, the very good agreement shown here should be considered somewhat coincidental. 6. CONCLUSIONS AND RECOMMENDATIONS The following conclusions and recommendations are based upon the experimental results and analysis presented in this report. (1) At comparable agitator speeds, salt-side mass transfer coeffi- cients for uranium are higher than water-side mass transfer k coefficients for quinone measured in water-mercury systems, and higher than mass transfer coefficients for other Coumponents in aqueous—-organic systems. Furthermore, the dependence of salt-side mass transfer coefficients on agitator spesd seems to be somewhat stronger than Tor the water-mercury system at low agitator speeds and with similsr diameter'agitators. (2) The change in slope of the mass transfer coefficient vs agitator speea curve at 170 rpm is probably caused by the onset of phase dispersal. The occurrence of dispersal at this speed is in reasonable agreement with data measured in water-mercury and aqueous—-organic systems. (3) There is & large increase in mass transfer rates with only Slighthhase dispersal. Tt may be possible Tc achieve mass transfer rates required in the MSER processing plant by operating under these cbnditidns without suffering bismuth entrainment in the salt. The data on bismuth entrainment 'presented in the Appendix suggest this; but further testing is required to confirm it. 32 ORNL DWG 76-579 1000 T T T T T T 7] T T T T T L1 11 i _J. 1 - ®(9) . 1 T £ _ x| wh - 10 == -1 - - - -1 - I L Llllllll 1 ) b Lk 1 | 10 100 0.0861( Sc, )37 (Re, )08 ( | +22 Bez ) ) Re, Fig. 11. Comparison of the uranium mass transfer coefficients with the McManamey correlation. Number in parentheses refers to run number. 33 (4) The mass transfer coefficients measured at agitetor speeds below 170 rpm.provide“data Which,should be compared with new correlations for stirred nondispersing contactors. The McManamey correlation correlated these data much beter than twd other litefature‘éorrelations; however, it should be used with extreme caution for scaleup and design if extrapolations to untested conditions are reguired. T. ACKNOWLEDGMENT The authors wish to acknowledge the corntribution of Mr. J. Beams, technician assigned to this project, for his diligenée and skill in operating the rather cantankerous experimental equipment used for this work. Appreciation is also due to My, Max Montgomery, pipefitter, for maintaining the equipment in working condition. 8. REFERENCES 1. L. M. Ferris et al., J. Inorg. Nuel. Chem. 32, 2019-35 (1970). D. R. Olander and M. Benedict, Nucl. Sci. Engr. 1k, 287-9L (1962). C. H. Brown, Jr.., in Engineering Development Studies for Molten-Selt Breeder Reactor Processing No. 2L, ORNL/TM-5339 (in preparstion). 4, J. Hernanz et al., Determination and Correlation of Mass Transfer Coefficients in a Stirred Cell, ORNL/MIT-22 (in preparation). 5. dJd. A, Klein and C. H. Brown, Jr.., unpublished dats. 6. H. 0. Weeren and L. E. McNeese in Engineering Development Studies for Molten-Salt Breeder Resctor Processing No. 10, ORNL/TM-3352 (December 1972), p- 52. T. J. B. Lewis, Chem. Eng. Sci. 3, 248-59 (195L4). 8. D. R. Olander, Chem. Eng. Seci. 18, 123-32 (1963). 9. W. J. McManemey et al., Chem. Eng. Sci. 28, 1061-69 (1973). 10. G. R. A. Mayers, Chem. Eng. Seci. 16, 69-T75 (1961). 11. W. J. McManamey, Chem. Eng. Sei. 15, 251-5k (1961). 3k l2. A. S. Foust et al., Principles of Unit Operations, p. 210, Wiley, New York, 1960. 13. C. R, Wilke, Chem. Eng. Progr. 45, 218-24 (1949). 14. R. B. Bird et al., Transport Phenomena, pp. 514-15, Wiley, New York, 1960, 15, &. Cantor in Molten-Salt Reactor Program Semisnn. Progr. Rep. for Aug. 31, 1969, ORNL-4LLk9, p, 1b5, 16. 5. Cantor in Pnysical Properties of Molten-Salt Reactor Fuel, Coolant, and Flush Salts, ORNL/TM-2316 (August 1968), p. 28. 17. S. Cantor in Molten-Salt Reactor Program Semiann. Progr. Rep. for Aug. 31, 1969, ORNL-4hhg, p. 1LT. 18. R. B. Lindauer, personal communication, Feb. 28, 1973. 35 APPENDIX A. Sample Analyses The counting data obtained during runs TSMC-2 through -9 are pre- sented in Tables AI-A8. Counting data are given for 237U (207.L5 xev R7) and ° 7z (743.37 keV B ) in the solid salt and bismuth samples, and for 2BTU after the samples were dissolved. All results are givén in terms of counts per minute per gram. 36 *aTdwes weaxis FUTMOTT = S 4L = G ‘Hl = €L =€ ‘8L =2 ‘TL = T ‘utrdtio aofdmes = ) pue ¢(378S = § ‘Uynwslq = g) 97dwes UT TBTISIBW = ¢ tTaqumu oTduBS = ¥ 8J9UM ‘H-g-Y¥ 07 FuTpuodsaiaoo spod B £q poqwuBTSep ST oTdmss sowmm 40T X G6°L gOT X w.m“M £0T X 656 G=5-gTL #0T X f6°e #OT ¥ ®E°9 LO0T ¥ ST S G-g-91T 70T X £€9°9 0T X 6°T > 0T X €6°T G-8-LTT 40T X Of"¢ /0T X 26" ¢ ,O0T X 9L°2 G-g-ST1 ¢OT X 98°T gOT X 16°c gOT X 662 H=g-%1T 70T ¥ £8°T 70T X 9L ¢ 70T X 96°T c—4-0TT gCT X 621 ¢OT X cg8°¢c ¢O0T X T2 h-S—€T1 x0T X T9'T 40T X 62°'€ yOT X g1°T 2—-g-60T 0T X €6°T gOT ¥ 1I6°6 gOT X HT°G £-g-21t 0T X G0'f x0T X RO°T ¢0T ¥ gT°K T-9-g0T 0T X 76T 0T ¥ 60°G gOT X 00°§ £-5-T1T ¢0T X Of"} 40T X QO0°T 0T X 60* 4 T-g-L0T I JI299J8 U2YB3 Sordweg 0T ¥ 49°T 0T X LG°¢ 0T X T6'e Sd~8-90T 70T X wl'2a 40T X T0°'S §OT X et Sd-g-66 gOT X ®9°'T 0T X Li*g gOT X fL°2 Sd-5-60T x0T X L6°T 40T X T6°¢ /0T X 98°T SJd-g-g6 ¢OT X 06°1 0T X lg-e gOT X 69°C SA-5—-40T ;0T X 76°T 40T X lg-¢ a0T X L A Sd-g-L6 - -= - S4-5-£0T x0T X 0T"¢ 40T X g2 % x0T X fc'c Sd-g-96 cOT X TL'T 0T X [G°2 ¢OT X "¢ Sd-g-c0t £0T X 00°¢ #0T X 082 70T X 16T Sd-g-66 OT X 25T 0T ¥ 0L°e c0T X Ix*c Sd-5S-T0T 70T X 69T #0T X 62°¢ {OT X o' T Sad-g-16 gOT X g°T 0T X 9¢°2 0T X gt°¢ Sd4-8-00T 40T X T0'¢ §OT X 61" € x0T X 08°T Sd-g-€£6 TUhd PULTAND UsyNB] So[duBg gOT X 6L°T ¢0T ¥ 60°2 gOT ¥ 96°1 g-g-cé GOT X 68T 0T X 6L°2 ¢OT X T0°¢ £=-5-16 SI90BIG JO UOTIFIPPRE J27JB 4nq ‘uUng 03 J0TJad usys] Sof g 0T X 9°9 > 0T X g9 > #OT X 9°6> £-5-06 _ _ _ 0T X 6°¢ > e0T X %°9 > g0T X m.m.w £-5-68 (0T X 9°9 > 0T X £°¢ > 0T X ¢°¢ > T-d-4%6 Z0T X P.N.W x0T X 8°T > gOT X 879 > S-s-lg 20T X mo.H.W 0T x m.HAW z0T X m.m.w T-g-£6 0T X G°'f7 > 40T X @°T > gOT X G°Q > 5—-5-9¢8 10T X LE"9 > 0T X g'c > 0T X €°€E > c-9-58 unI o3 JoTJd Ud¥e] SaTdweg (uTw-38/s3unod) (utw-3/sjunca) (uTw-%/s3unco) g9P02 (uTw-3/squnoco) (utw-g/squnoo) (uTwm-% /S3Unoo) g9POD I1Z,¢ 103 Nz ¥0F N;gz 207 oTdmeg IZ,6 I0F O,¢7 107 N;gz %03 oTdursg sTsfTBUB DPITOS STSATEBUR UOTINTOQ STSATBUB PLITOY STSATBUB PITOQ STSATBUB UCTINTOYG sTSATBUB PITOg Z=0WSI UnI woJgJ psulBlqo ®BIBP JFUTIUNOD *I-V STqEL 37 Teble A-2. Counting data ¢bteined from run TEMC=3 Solid =znalysis Solution analysis Solid anslysis Sclution enalysis Sample for 237y for 237y Sarmple for 237y for 2377 code? (counts/g~min) (counts/e-min) code® {ecunts/g-min} {counts/g-min) Semples taken prior Lo run 141-B-5 < 1.13 x 102 < 5.1 x 1083 1355 < 3.3 =x 108 <5.3 x 10" 142-B-5 < 1.6k x 102 < 5.5 x 108 ibkl-8-5 < 3.2 x 109 < 7.6 x 1o% 147-B-5 < §6.26 x 10} < 3.L x 108 145-5-3 < 3.k x 103 < 7.6 x 10“* 148-B-5 < 1.00 x 102 40T X 6°% > 40T X 0°8 ¢-5-t2e 40T X 29°¢ 0T X gE°2 40T X g2°¢ G-g-222 0T X 2'¢ > 0T X L' > LO0T X t°g g-g-tze 40T X ge e 0T X te'e 0T X 98" L S-g-T22g 40T X L9°} 0T X TL°¢ 0T X gt°¢ ~S=0c¢ 70T X g9°¢t g0T X LT°¢ 0T X L0°T c—d-9Te 70T X 88K 0T X 19°¢c 0T X 62°¢ f1-8-61¢ 40T X 69°€ ¢OT X %g°¢C ¢OT ¥ 80°T g-g-61e gOT ¥ 9t*¢ 90T X }5°T g0T X G2°T £-s-Q12 0T X T°T 0T ¥ L°L 0T X EL°T T-9-1Te 0T X 6£°¢ 0T X gH°T 0T X TE°T £-g-L1¢ ¢0T X 6T 0T x L°'g ¢0T X 00°¢ T-g~£T1¢ ungd J533J8B USYEBY SSTAWBS #OT X #4879 gOT X KT°¢ cOT X 26°¢ Sd-S-c1c 0T X 2T°L 0T X 2€°¢ 0T X LG°T Sd-g-502 50T X gh°8 0T X 0Tt 0T X £9°¢C SA-S-T12 ,0T X €7°9 0T X 9t ¢ ¢OT X £8°T Sd-g=w02 40T X 96°9 cOT X fE"E 0T X 9L°¢g Sd-5-0T 40T X 0°T> 0T X €°f > E-8-%6T 0T x L6 0T X L'yg> ¢0T X 6T°T > T-€-96T 0T X #°T > x0T X 9°t > 0T X b.:.w £-5-€61 20T X m.m.w g0T ¥ m.:.w ¢0T X 00°T > T-g-66T 0T X %79 > x0T X g°t > g0T X 0O°f > G-8S-06T 20T x m.m.w ¢0T X L°9 > 0T X €8°Q > G-g-261 0T X &°T > 40T X 9°¢ > ¢0T X 6°6 > G-S-68T 0T X #'f > 0T X T°6 > ¢0T X €0°T > G=g-T6T und 09 JOTJd Usye]} SoTdweg Anfls1m\mpzfioov {uTw-F/s3Uunod) fisflnTw\mpqzoou g9P00 AQHfllm\mpflfiouu hanTm\mvnfioov (uTw-3/squnov) goPO2 T, 0F N;¢z 103 fl,e7 I0J oTdureg 17,6 *0F 0,¢z 103 n £z I0J oTdureg STSATBUB PITOS ETafATBUR UOTINTOQ stsAiBUB PIIOY STSATEUB PITOS stsiTBUR UOTINTOYS stsATeus PITOS {=~OWSIL UNJI WOIJ PauT®3qO ®I8P JUTIIUNOD "£-Y 9TaBL 39 *oTdwss W8salS JUTMOTY = 8i SGL = § L = ¥ ‘€L = € %28l = g $TL = T ‘UTdiao ordums = J pus (3188 = g ‘uyjnusTq = ) STAdwes UT TBTLSIBUW = g ‘Jaqunu sTduss = y aaeym ‘p-g-y 07 Julpuodssaaoo spoo ® Aq pejvudrssp sT oTdwms Yowd, g0 ¥ S'E > - 0T % T'T > $=8-093 40T * 002 -- gOT ¥ HE"T S—€-gse gOT * 9'c > - 0T X 2°6 > G=5-642 yOT x Lf*g _ - gOT ¥ TL*T G~d-L52 gOT X LE'T - gOT ¥ OT'T {~-g-94¢2 40T X 08T - gOT X 00°T g-d-gsc gOT X 61°1 - gOT X 84°T f~5-562 40T X mm.fi.l - 40T X ow.b.l g-t-16e gOT * 6T'E - g0T X ge°t E=-g-152 20T X §°F > - 0T X mo.fi.w T-g-042 0T X 96'2 - g0T X TT°E £-g-€4e 20T X 0'ff > -— 0T X gE°T > T-g-6%2 “UTWE A99.J8 UaiB] SoTdusg 40T ¥ 9£°6 - 90T X 0%°G Sd=-5-ghe 40T X 26°¢ gOT X TE'S gOT X TH'T Sd-g-The gOT X 36°1 g0T X go°g 90T X 6L°T ' gd-g-Lhe 40T X #0°€E gOT X L9 0T ¥ L9°T SI-g-043 0T X 62'28 0T X 10'€ g0T X 96°T Sd=-5-9N2 4OT ¥ gL-2 0T X EQ'f gOT X SE'T Sd-g-6Eg gOT X £f'e g0T ¥ 0T'¢ 0T X gL'T Sd-g~5he 40T X 6g°2 0T X T19°% 0T X GE°T Su-4-RL 2 gOT X 19°'¢ gOT X 00°2 0T X 28°1 gd-5-fihe 40T X 46°2 gOT X 09°% gOT ¥ ES°T Sad-g-LEe gOT X 6L°T g0T ¥ 00'¢c g0T X #9°T SU-8-Ehe 40T X 29°¢ gOT X ET'H gOT X OE°T Sd~g~-9E2 0T X Li'g g0T X ET°¢ g0T X EQ°T Bd=G-2chg 40T X 4g°T 0T X 1g°¢ gOT x 70°T Si-g-GEC NI JUTINP Us¥e] SsTduwRg ¢OT X @gc’¢ g0T X gh°t g0T x LG*2 E=g-fEc gOT X 0T*2 90T X fif°t 90T X %5°'2 E-g-tte S490HI] JO UOTJIPPE 48748 I00 und 07 JOTJdd uaie] SoTdWegy 0T X #°T > - 0T X 0°9> £-5-0Ee 0T x 0TS -- 20T X T6°S > T-H-ZER 0T X 2°T > - g0T X 6% > £-5-622 0T x T°T > -- 20T X mm.:.w T-g~1te 20T X m.w.w - g0T X m.w.w ¢-g-92g 20T X m.mum - Z0T X 6f°g > S-g-gee 0T X 9°'T > - g0T X #*L > G-g~Scc 20T X w'T > - 20T X 19°'G > S-g-Le2 und 07 JI0TJIU Ul38q sajdwsg (utw-3/83UNco) (uTw=3/s3tmoo) (uTw-3/83UN00) 59P0OD (uTw-3/sq.unoo) (utw=-3 /89UM00) (utw-8/s3unoo} 5oDOD A7, 65 L0F gz I0F N7 70J ardureg a7, 4OF N,pg *OF ¢z 4OF a1duegy STSATBUB PITOQ sTsATBUE UOIINTOY sTsdTeUB PTTOQ gsTedTRUB PITOY gTsiTBUB UOTANTOQ STSATBUB PITOQ G-0OWSL Urnd WOIJ PAUTBIGO BBY BUTIUNOY =Y STUBL %0 puw mfivflmm = g ‘yinwstq = q) aTdums UT TBTJ93BW *o1dues weadds BUTMOTI = Sd 6L = ¢ {l = ‘€L =€ 2l =2 'TL = T ‘urdixo ordums = ) = g ‘aequmu sTdwes = Y 2J9UYsA ‘)-g-v 03 Jurpuodssaacc 9poos B £q parsBudisap sT oTdmss nodmd 0T X 0°6 > 0T X L'T> 0T X T°g> 6-S-%62 40T X Gi°% gOT X TE"E 0T X €€°T ¢-g-262 g0OT ¥ L'G > 0T X @°T > 0T X 2°6 > G-g=£62 40T X 9E°f 0T X TR'E 0T X 60°T ¢-g-T162 40T X 68°8 cOT X 0g8°¢ ¢OT X 06°€ t~-S-062 0T X 65°4 gOT X TT°¢ OT X ET°T c—d-98¢ «OT X 0%°8 0T X STy gOT X ®0'f h=S-69¢ #OT X £€°6 0T X ®E°E ¢OT X LO°T c-d~5Q2 0T X LG°¢E 0T X RE'T g0T X 0¢°T £-8-gg 0T X H°T > 0T X 9°6> €-5-t92 {0T X g6 40T ¥ Li°¢ 40T X 66°T T-4-992 g0T X O°T > K0T X ®°T > 0T X &°T > £—-8-£92 0T X 96°¢ 0T X B8L*S 40T X G6°T T-g~592 0T X €°T > x0T X £°'T > 0T X T°9 > ¢~5—-09¢ ¢0T X TG°C x0T X 6§°6 40T X LE°2 G-g-2c92 0T X g'¢c > 70T X £ T>> 0T X 2°9 > g=s-66¢ - 0T X 2£°6 40T X SL°T G-g-192 und 09 J0TJId Us¥Bq SsSTAWEBG (uTw-3/s3unoo) (uTw-3/s3unoo) (UTw-%/s3unod) »oPOO (uTw-3/squmoo) (uTw-8/squnod) {(UuTw-3/53UNo0) g°POD 17, IOF Nygz 20F Ng7 (0T X €°G > §-5-9vE ¢OT X BS°S c0T X G0°¥ S-B-VPE 0T X L°8 > g0T X 8'6 > S-S-GPE gOT X 5075 ¢OT % BE'V S-g-EbE g0T ¥ 6T°2 0T X ¥9°T b-5-2ZveE gOT X 98°¢ gOT X 66 Z-g-BEE —— - b-S-TPE gOT X BL'E gOT x £L°C C-d-LEE - -- £E~S5-0VE gOT X B8T"C 0T X L¥°1 T-8-9E¢ - - £-5-6L¢€ gOT x g2 ¢ g0T X 06°T T-g-5¢¢ unI J8jje usyel SaTdwes ¢OT % LT°S gOT x 2e°E Sd-H-LZE g0T X 06°€ g0T * €6°2 Sd-5-VEE gOT X 50°Ss gOT X ¥6°E SI-g=-9Z¢ g0T * 96°¢C g0T * ZB*C Sd-S-ELE gOT X 90°§ 0T X 9G°F Sd-g-6Z¢ g0T X L6°E 90T X* 68°C 5d-S-CEE 0T ¥ £0°S 0T X 86°E SI-H-vCE g0T * 92°¢ gOT X SL°2 S3-S-TEE gOT * 1I8°V 0T X €T°E 5J-g-E£C¢ gOT * ZE°E g0T X 68°2 S4-5-0t¢E gOT X EE°V 0T X 66°¢ Sd-g-E2¢ g0T X 68°¢C g0T X TT°C 8d-5-62¢ gOT * T6°¢ gOT X v9°C SI-g-TCE Und butanp Uede3 Sefdues _ g0T X T9°¢ g0T X G6°¢ £-5-02¢ ¢0T X 97°6 g0T X S6°¢ €-S-6TE SI190va] JO UOT4IPpPE A93JFF 3ng unit o3 I0Tad usyel saTdues g0T X S°6 M 0T X 9°L M £-S-9T¢t {O0T * LV'9 x0T X §9°6 T-2-8T¢E 20T X 8°T > g0T ¥ 979 > £-S~5TE 40T ¥ 6L°9 0T X S6°V T-9-LTE Q0T X &°T > 0T X T°5 > §-S~-CT¢E 0T X LL™9 40T X v9°V c-g-vTE ZO0T ¥ 8°T7 > 0T X T°% > G=-5-TTE 40T X 6£°9 40T X 06°S S-g-tTt una o3 I0Tad uslel Selawes (utw-5/s3UNnoo) (uTw-56/83UN0D) 29POD (uTw~-H/s3UNOD) (uTw-56/53UNO2) pdP02 Nyge *93 N,ez 103 a1duwes N pg 393 N, gz 393 arduwes stsAieur UOIINTOS sTsATeu® pPTITOS sTsATeUR UOTINTOS q sTsATeue pTITOS L—-OHSL una woxy peuteldo ejep buTjunod “9-¥ STqRl ho * oT7dures I = ‘€L = € ‘2L = ‘1L = T ‘ui@tao os7dus = 5 pur (3TES = § ‘UINWSIq = g) uresa3s FUTMOTI = Sd ST = § c tasqumu oTdures = ¥ SJ9UM ‘)-g-V 0% Jurpuodsaxioo apoo ® £q pelreUdTISSp sT aTdwes sommm cordures UT TBTJI27BW = g 0T x 2> SOT X 2°T > G-S-T6€ ¢OT ¥ 99°€ cOT X 6£°T G-g-68€ K0T X #°c > 40T X 0°T > 6-5-06¢ 0T X 6T°€ 0T X 22°T C-€—QgE 0T X GE°2 o0T X 08" T #=-5-19¢ #0T X L0 70T X 94T c—gd-£g¢ 0T X G1°2 0T X LL°T 1=5-9g¢ x0T X 2276 x0T X T6°T c—d-2g¢ g0T X gf*¢c 0T X 0T°¢ £-5-4Qt qf0T X gn°T ¢0OT X I8° 1 T-9-18¢ 0T X [9°2 g0T X g2°¢ £—S-fg¢ 20T X LG°T ¢OT X gL°% T-9-0Q¢ una Js1Je usayeq Sotdured 90T X ft'¢ g0T X £f°¢ mmlmlmwm cOT X 90°T 70T X 90°¢t SA-g-cLE 90T X QT*2 90T X 06°T Sd-S-g.LE +0T X 65°6 40T X 66°2 Sd-g-TLE 0T X 06°T gO0T X #6°T Sd~-S-LLE 0T X 69°Q x0T X 98°¢ Sd-g-0L¢E 0T X 6c*c 90T X g0°¢C Sd-S=-9.LE x0T X £0°8Q 40T ¥ £°2 SA-4-69¢ 0T X "¢ 0T X 9L'T SA-S~GLE 70T X 88°¢Q x0T X 99°¢ Sd-4-39¢t g0T X ¢0°'T 0T X 676 SA-S—HLE +0T X Tc°¢ 0T X 8T 6 Sd-g-L9¢ cOT X ¢L°¢E ¢OT X #0°¢ Sd~S—ELE x0T X 00°T ¢0T X xL°2 Sd—~g-99¢ unJg SUTJnp us¥Bl] SaTdueg 90T X 16°¢2 g0T X If°¢ £-5-69¢ 90T X 68°C 0T X ge* e €-8-19¢ JI90®JI] JO UOTZFTDPPE JI89IB 1ng UNJI 07 JOTJAA U3¥YE] SsSTAWERG x0T X w.fi”M ¢0T X w.:”M £-8-TOE ¢0T X 90° % 0T X G2°'T T-€-¢9o¢t x0T X 9°'T > ¢OT X w.j.w £-S-09¢ ¢OT X TS84 ¢0T X NN.H.I T-€-29¢ 0T X 6°Q > ¢0OT X g°f > G-g=LGE ¢OT X €€ 0T X #°T > G-g-65¢€ c0OT X g°t > G—8=94¢E ¢OT X QTN 0T X L°T > S-g—-gse unJg 0% JA0Tad Us¥eq SssTdweg (uTw-3/53UNnod) (uTw~-3/s3Unod) poPOD Aflflalw\mpqfloov (UTw-3 /83uMoD) gSPOD N,¢7 IO Ngg I0J aT1dureg 0;¢z 10 N,gz I0 aTdwreg sTsATBRUB UCTANTOS sTsATeU®B PITOQ sTsATBUR UOTANTOG STSATBUB DPITOQ Q-OWSL UNJd woJgj PauTejqo BIBP SuTjuno) *L-¥ STqEL 43 . *o1dwss wBaaANS FUTMOTE = Qi SGL = § SHi = ft SEL =€ ‘8L =2 ‘T = T ‘furdrao ordwes = J pus (4788 = g ‘YINUWETQ = f) “oTdurss UT TBTJA9BW = g (Jdaqunu ordums = Yy adeym )-g-y 01 Furpuodssdaon opoo v Lq pejvudrssp st ordwss sodad 40T X :.H“M g0T X :.m_M G—-g~-Lay gOT X G°g cOT X 6T"°T G-g~62h ZOT X G°T > 0T X T°L > G-8-9¢t cOT X 62°¢ gOT X g0°T S-g-w2t 0T X 69'1 gOT X #5°T h—8-£3h gOT X 9g°¢ gOT * EX°T - c-g-6Th gOT X 18°T gOT X #G*T h-g-gch gOT X G0°E gOT X LT T-g-8TH cOT X 69°8 cOT X 2T'. g-g-Teh ¢0T X T°E > Z0T X 626 T-a-LTH 0T X 2L°g cOT X 68°9 £-g-024 g0T X L€ > z0T x ¢6°g T-g-9THk unJg J99J8 uayw] ssTdweg gOT X 69°1 0T X 69°T Sd~-8=-5Th gOT X #6°4 ‘ 0T X 262 SI-g-80% gOT X g8°T 0T X %L°T Sd-g-HTH cOT X TL°G gOT X gg°'<¢ Sua~-g-1,0% 0T X 6L°T cO0T X g2'c S=-S—ETH gOT X 9E* 4 gOT X TE°E Sd-g-901 0T X 66°T cOT X 96T . Sd-8-cTh cOT X 62°% gOT X gl-g Sd=-g=-50t cOT X 00°¢2 cOT X #0°¢ Sd-S-TTH gOT X ¢€°9 0T X 96°2 S-g-t01 gOT X 9T'¢ gOT X 90'¢c SJd—-8-0TH gOT X KLy gOT X GT°E Sd-d-E01 0T ¥ 76°T gOT x LT'2 Si—g-601 gOT X 02°% 0T X 60°€ Su-g-g0t unJgd JUuTJUp UsyB] SBoTAWRY 0T X G2°€ - gOT X QT'¢ E-g-T0H g0T X 68°¢ 90T X LT'2 €-8~001% Jd90BI9. JO UOTETRPR J437JB Ing und 07 JO0TJId USYBT EITAWET wOT X &°'T M 0T X 679 M g-g-L6t 0T X 0°¢ M Z0T X £4°g T-4-66€ 40T X E°T > g0T X p.w.w £=-9-96¢ 0T X m.m.w 20T X :m.».l T-g-g6E x0T X €°T > g0T X 8'9 > G-8-E6E ¢0T X 9°¢ > 20T x 2S°'g > G—g~G6t xOT X £°T > 0T X g'9 > G-g-26¢t 0T X 2°¢ > 20T X g L > G-g~H6E UL 09 JOTJ4d Ua¥B] SaTuweg (uTw-2/83uUnoo) AanTm\mpnfioov 5oPOD (utw=-5/89unoo) Anflslm\mpmsoug goPOD Ngg 40F Ngz 40 aTdureg Ngg 40J [ypg 40F aTdwsyg sTsATBUB UOTINTOY STSATBUB PTTOQ grsATBUB UOTINTOL sTsATBUB PITOQ 6-0WSL UNJ WOIJ PauTelqo ©gep JuTqunoy 'g-y STUBI Ly APPENDIX B. Calculation of Mass Transfer Coeffilicients For a flow-through, continuously stirred contactor at steady-state conditions, a mass balance on the salt phase yields: = + - F.C, =FC +J, (B-1) where F, = flow rate of salt, cm3/sec, Cl = tracer concentration in salt inflow, units/cm3, 02 = tracer concentration in salt outflow, units/cmB, = rate of transfer of tracer across the interface, units/sec. Expressing the rate of transfer across the interface as the product of an overall mass transfer coefficient and a driving force times the area available for mass transfer yields:12 = - B-2 Jg=x/[c_-c /DA, (B-2) where = + - 1/Ks 1/ks 1/ka . {B-3) and KS = overall mass transfer coefficient based on salt phase, cm/sec, ks = individual mass transfer coefficient in salt phase, cnm/sec, km = individual mass transfer coefficient phase in metal, cm/sec, D = distribution cocefficient = ratio of concentration in metal phase to concentration in salt phase at eguilibrium, moles/cm3 3 2 moles/cen k5 C m A tracer concentration in metal outflow. units/cms, and . - 2 interfacial area, cm . Taking an overall mass balance results in: = ' ' =l C.F + FF, CF, +CF,, | (B-L) where 02 = {racer concentration in metal inflcw,-units/cm3, and F, = flow rate of metal, cmS/sec. Ir C, = 0, Eq. (B-L) can be rearranged to give the four following relations: = - ' (B-5) C,F, = CF +CF, , - Fy ¢, = c.+c |} (B-6) i Fy Cs = C_ - Cm 71 and . ' (B-7) ]_ . F F c = ¢ =-cl]. | (B-8) m o o Combining Egs. (B-1), (B-2), and (B-8) yields: K CAfF ) EcCA([F Y\ o Pl = PO P ECA-TEO\R )T T (5, ) (5-9) no which can be rearranged to give: KSA +-..-......-._. ¢ o F, + -5 (F/F,) - _ (5 10) s 1 K A_( ) - 8 Fl Fl * KSA'+ D 'fig Combining Eqs. (B-1), (B-2), and (B-T7) yields: L6 F2 KSCmA = - - —_—— B— FiCy FiC; - CF, + K CiA - (K C_A) . 5 , (B-11) which is rearranged to give: KSA C/Cy = F, (KA (B-12) Fo v (KA 5=+ 1 Combining Egs. (B-1), (B-2), and (B-6) yields K CmA F,C, +CF, = F.C +KCA-\—F ; (B-13) which is arranged to give: KSA c_/C = e (B-1h) m s KSA Fo # _5-) Rearranging Egqs. (B-10), (B-12), and (B-1k4) gives three expressions for the overall mass transfer coefficient in terms of the measured quantities Cl’ Cs’ Fl, F2, D, and A: CS LAt v 1 3 8 s\ af%\/f1\ aff1 e )*ole JI5 ) - olF, (3-15) 1 1/\ "2 o cll'l g ) ~ (B-16) 7 K= . | | (B-17) The above eguations can then be used to calculete mess transfer coefficients from experimental results (i.e., the ratio of tracer concentration in any two of the salt of bismuth flows). Within experimental error, the distribution coefficiefit D can be get as desired. To minimize effects of uncerteinties in the velue of D on the calculated value of the overall mass transfer coefficient, D should be made fairly large. For the values of concentrations and flow rates used in these experiments, the terms which contain D in Egs. (B-15), (B-16), and (B-17) are less than 5% of the values of the other terms for values of D greater than 20 and can be neglected with little error. By assuming that the terms that contain D can be neglected, Eqs. (B-15), (B-16), and (B-1T) reduce to K. = (B~18) kg = (B-19) F. C _ 2 m K., = 1= G (B-20) L8 Uncertainties in the distribution coefficient do not affect the accuracy of the overall mass transfer coefficient. However, as shown by Eq. (L), when D is very large, the overall mass transfer coefficient is essentially the individual salt-phase coefficient, since resistance to mass transfer in the metal phase is comparatively negligible. 49 APPENDIX C. Calculation of Diffusivity of UF. and Zth in Molten Salt 3 | The diffusivities of UF; and ZrF) in molten salt (72-16-12 mole %) LiFmBngnThFh were estimated from an empirical equation developed by Wilke,13 which was based on the Stokes-Einstein equation:l o () e E%B = T7.bL x 10 ——7:?;?:3—— s (c-1) A where 213 = diffusion coefficient of specie A in solvent B, cm?/sec, ?B = assoclation parameter for solvent B,which is equal to 1.0 for an unassocisted liquid, T = temperature, °K, Mz = molecular weight of solvent B, g/g-mole, %A = molar volume of solute A. cmB/g—mole, n = solubion viscosity, cP. This equation is good only for dilute solutions of nondissociating solutes; 1k for such solutions the error is within + 10% of the true value. Semple caleulation The diffusivity of UF3 mole %) at 600°C was calculsted as follows: and ZrFLL in molten LiF_BeFQ-ThFh (72-16-12 T = STSOK, n = 11.81 cp,’ M., = 63.16 g/g-mole, | %fiFh = U6.4 cnd/g-mole at 600°C.1° This value was assumed to apply to both UF3 and Zth, wsalt assumed = 1.0, _ 317 psalt = 3.35 g/cm”. 50 Substituting the above values in Eq. (C-1) results in: 1/2 & = T.4 x 10 = 4,45 x 10" em”/sec. UF3 or Zth ~ galt (11.81) (h6.h)0'6 see -2 8o = 1]2 _ (ll.8§ x 10 ° g/cm Zec)2 - 7920. P (3.35 g/em™) (L.L5 x 107 em™/sec) 51 APPENDTX D. Entrainment Studies One hydrodynamic test was performed to determine the smount of fluo- ride salt that might be dispersed and entfained.in the bismuth and the amount of blsmuth that might be digspersed and entrained in the fluoride salt effluent streams of the contactor at various agitator speeds. This hydrodynamic run was performed with salt and bismuth flcw rates of 150 cc/min and 1L0 ce/min, respectively. The agitator was operated at three different speeds during the run, 250 rpm, 310 rpm, afid 386 rpm. At 250 rpm gnd 310 rpm three sets of unfiltered salt and bismuth sampies from .the-contactor effluent streams were taken ot L-min intervals. Three sets of unfiltered effluent samples were also taken with the agitator operating at 386 rpm, but the samples were taken &t P-min intervals. The sample capsules were cleaned and the contents of each sample were removed as described in Sect. 3.5 of this report. The contents of each sample were inspected for evidence of gross entralnment of one fihase into the cther., No such evidence of entrainmeni'was found. Results of chefiical analysis of the bismuth and salt samples for beryllium.and,bismufih con- tent. respectively, are given in Teble D-1. The flowing~stream bismuth samples from runs TSMC-5, -6, and -9 were also analyzed for beryllium content, and the results of these analyses are given in Table D-2. Runs TSMC-5, -6, and -9 were performed with agi- tator speeds of 124 rpm, 180 rpm, and 24l rpm, respectively. The bismuth concentration meagured in the salt samples tzken during the hydrodynamic run (Table D-1) shows a general decrease with increasing stirrer speed, with very low values occurring at the highest stirrer speed. It also seems evident that the bismuth concentration in the salit phase may. have been a function of the run time. After the fourth sample, the bis- muth concentration in the salt samples reamined at a relatively_constant value of 50 + 11 ppm; this is quite different from the values reported for the first four samples,which ranged from 1800 ppm to 155 ppm. 52 e Svv 0S 9ty 98¢ e 7A"44 SLT seEd 98¢ i74°] evd OTT PEY 98¢ e vy - £eEv 0Tt €S 1vd 0T6 cev 0T¢ oLC o¥¥ S8 Ted 0T¢ esT 6ev S1c oev 0S¢ q0¢ 12184 QT 6Cv 0s¢ 008T LEY qTC 8cv 0S¢ (udd) aTes *ON (udd) yanuwsiq *ON (udx) ut ynwstyg oTdues 3Tes uT umtiTAiedg oTdures yaInwstyg poads I03e3THY unx oTweulpoxpdAy 9yl HutaInp usNel soTdues YInNWS I pue 3Tes 9yl JO STsATeur woxF sS3jtnssy *I-d ST9el 23 Table D-2. Results of analysis of flowing stream bismuth samples for presence of beryliium Run Agitator speed Bismuth flowing stream Beryllium in NO . (rpm) sample No. - bismuth (ppm) TSMC=5 124 235 8l 124 236 131 124 237 86 124 238 ig1 124 ' 23¢ o 162 124 240 : 111 124 241 : i32 TSMC-6 180 269 264 180 270 315 180 271 | 118 180 272 125 180 : 273 86 180 274 278 180 275 113 TSMC-9 242 402" | 85 ' 244 403 118 244 AQ4 41 244 405 | < 10 244 406 26 244 407 56 244 408 104 54 These results are significantly higher than those of Lindauer18 who saw less than 10 ppm of bismuth in fluoride salt that was in contact with bismuth in several different contacting devices. It is likely that sample contamination is a contributing factor to the high bismuth concentrations that were measured. Three possible sources of sample contamination have been reported:l9 (1) sample contamination during sampling by withdrawing the samples through a sample port that has been in contact with bismuth; (2) sample contamination during sample handling and in the analytical laboratory by the use of equipment that is used routinely for bismuth analyses; (3) sample contamination from a low-density bismuth material that may be floating on the salt surface. The beryllium concentration in the bismuth samples taken in the hydro- dynamic run (Table D-1) and in runs TSMC-5, -6, and -9 (Table D-2) show both high and low values with no discernible dependence on agitator speed. Based on previous experiments with water-mercury and organic-mercury systems, one would expect entrainment of the light phase into the heavy phase at an agitator speed of about 170 rpm. 55 APPENDIX E. Tocation of Originel Data A1l data and operating records for the salt-metal contactor studies are recorded in log book No. A-6886-G. Records for the Ffaeility up to the time at which the stirred interfece contactor was installed are con- tained in log books numbered: A-56U9-G, A~5965-G, A-6210-G, A-6LB2-G, and A-6T722-G, and are available from the authqr.