CFNTRAL RESEARCH LIBRARY ORNL.-TM.3259 TR cy52 3 445k 0383070 9 ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT BREEDER REACTOR PROCESSING NO. 9 L. E. McNeese OAK RIDGE NATIONAL LABORATORY CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON If you wish someone else to see this document, send in name with document and the library will arrange a loan. 2, = OAK RIDGE NATIO LABORATORY OPERATED BY UNION CARBIDE CORPORATION = FOR THE U.S. ATOMIC ENERGY COMMISSION ORNL-TM-3259 Contract No. W-TL05-eng-26 CHEMICAL TECHNOLOGY DIVISION ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT BREEDER REACTOR PROCESSING NO. 9 L. E. McNeese DECEMBER 1972 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37830 operated by UNION CARBIDE CORPORATION for the U. S. ATOMIC ENERGY COMMISSION MARTIN MARIETTA ENERG T 3 445k 0383070 9 ii Reports previously issued in this series are as follows: ORNL-TM-3053 ORNL-TM-3137 ORNL-TM-3138 ORNL-TM-3139 ORNL-TM-3140 ORNL-TM-~3141 ORNL-TM-3257 ORNL-TM-3258 Period Period Period Period Period Period Period Period ending ending ending ending ending ending ending ending December 1968 March 1969 June 1969 September 1969 December 1969 March 1970 June 1970 September 1970 iii CONTENTS SUWARIES o » » . . . . ° . . . ° . . . . . . . . . . . . . . . . -Vii 1. 2. INTRODUCTION . @ > L n » . o o ° . . - . » . » . - . * L] L . ° FLOWSHEET ANALYSIS: ISOLATION OF PROTACTINIUM BY OXIDE PR.ECIPITATION » L . B . s . - . » o . ? o - . L * 8 L 2.1 TIsolation of Protactinium by Oxide Precipitation, and Recovery of Uranium Daughters by Fluorination. . . . . . . 2.2 1Isolation of Protactinium by Oxide Precipitation Without the Use of Fluorination for Recovering 233U Produced by Decay of 233Pa . v i e e e e e e e e e e e e e e e e e FLOWSHEET ANALYSIS: REFERENCE PROCESSING PLANT FLOWSHEET BASED ON FLUORINATION, REDUCTIVE EXTRACTION, AND THE METAL TRANSFER PROCESS ® - . o ° @ s . ® © . . . o > . - » L] ¢ . . . » » ; . " FLOWSHEET ANALYSIS: IMPORTANCE OF URANIUM INVENTORY IN AN MSBR PROCESSING PIJANT - o . o - - - . k] - e . > L b a * . » . s - . - FLOWSHEET ANALYSIS: REMOVAL OF RARE-EARTH FISSION PRODUCTS FROM LiCl IN THE METAL TRANSFER PROCESS . . « ¢« & & + o & o o FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON INDUCTION HEATING IN A CONTINUOUS FLUORINATOR SIMULATION. . « « « « o« « & 6.1 Experimental Procedure . . . « ¢ + o 5 s o s e v o4 s o4 s 6.2 Experimental Results . . s o o o o o o s o &+ s o o s o o PREDICTED CORROSION RATES IN CONTINUOUS FLUORINATORS EMPLOYING FROZEN—'WALL PROTECT ION . . . . . s . . . . o n ° . . ° ° . . » 7.1 Data on the Rate of Corrosion of Nickel in Gaseous Fluorine o . & . - o - . » - . - » 5 - - » - a a - » v o . 7.2 Predicted Corrosion RatesS. . « « o o o o s s s 6 s & s o PREDICTED PERFORMANCE QOF CONTINUOUS FLUORINATORS. . + « « o + & MEASUREMENT OF AXTAL DISPERSION COEFFICIENTS AND GAS HOLDUP IN OPEN BUBBLE COLUMNS ¢ &+ & « ¢ o o o o s 5 s o 3 s o o & o o« » 9.1 Previous Studies on Axial Dispersion . . « + + &+ « &+ + o+ 9.2 Equipment and Experimental Procedure . . . + « « « « « o 9.3 Experimental Data on Axial Dispersion. . . . . . + « . & 9.4 Experimental Data on Gas Holdup. . « « « « « + & s + & o o 9.5 C(Correlation of Data on Gas Holdup. . + « « + ¢ & « « + & 17 22 24 27 27 30 . 33 33 37 41 52 52 54 . 55 72 74 9.6 Compilation of Data on Axial Mixing. . . . . . . « « +. . 106 10. 11. 12. 15. 1h. iv CONTENTS (continued ) Page 9.7 Correlation of Data on Axial Dispersion. . . . . . . . . . 106 9.8 Discussion of Data on Axial Mixing and Gas Holdup. . 126 9.8.1 Flow Regimes and Effect of Gas Superficial Velocity . 126 9.8.2 Effect of Column Diameter . . . c e e e ... 12 9.8.3 Effect of Viscosity of the L1qu1d Phase c e v e v .. 1hs 9.8.4 Effect of Surface Tension of the Liquid Phase . . . . 146 9.8.5 Effect of Gas Inlet Orifice Diameter. . . . . 146 9.8.6 Effect of Number of Orifices in Gas Dlstrlbutor 148 9.9 Future WOork. « « + « ¢ + & v v v v v v v e e e e e 148 SEMICONT INUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL FACILITY. « « & « + & & o = o o v 4 o o 4 4« v v s o« v o o« « o llg 10.1 Run UTR-6 « + v v v v v v v vt v e e e e e e e e 150 10.2 Preparation for Run UTR-T; Installation of Molybdenum Draft Tube in Treatment Vessel. . . . + « « « & « + « . « « . . . 154 10.3 Preparation for Mass Transfer Run ZTR-1; Production of 9TZr by Irradiation of 90Zr. . . . . . . . . . . . . . ... 158 10.4 1Inspection of Salt Feed-and-Catch Tank, and Equipment Maintenance .« . « « « « ¢ 4 4 e v e e e e 4 e e e e e e 159 DEVELOPMENT OF THE METAL TRANSFER PROCESS: OPERATION OF EXPERI- MENT MTE=2. « « + v o o o o o o o o o o o o o o o o o v o o o« + . 167 11.1 Experimental Procedure. . . « . . « . . . 168 11.2 Mathematical Analysis of Transfer Rate. . . . . . . . . 169 11.35 Experimental Results. . « . « « « « + & o v & « & o « 17h 11.3.1 Rates of Transfer of Neodymium and Lanthanum. 174 DEVELOPMENT QF THE METAL TRANSFER PROCESS: DESIGN OF EXPERIMENT MIE=3 +© & & & & o o o o o o s + o s o o + o o e e e 4 e e e o« . . 196 12.1 Mathematical Analysis of Metal Transfer Experiment MTE-3. 197 12.2 Preliminary Design of Metal Transfer Experiment MITE-3 . 202 DEVELOPMENT OF MECHANICALLY AGITATED SALT-METAL CONTACTORS. . . . 205 13.1 Hydrodynamic Studies. . « . « « .« « .+ « ¢ . . . . . 205 15.2 Survey of Literature Relative to Mechanically Agitated, Nondispersing Salt-Metal Contactors « - « « « « « « « .+ . « 208 HYDRODYNAMICS OF PACKED-COLUMN OPERATION WITH HIGH-DENSITY FLUIDS 216 14.1 Equipment and Experimental Technique. . . . . . . . . . . . 217 15. 6. 17. CONTENTS (continued) ].)-".2 ResultS- . - . * . - . - . . - . . . . 14.3 Prediction of Flobding Rates and Dispersed-Phase Holdup in Packed Columns. e e e e . . . ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN MOLTEN SALTS AND LIQUID BISMUTH DURING COUNTERCURRENT FLOW IN PACKED COLUmS - ». . * . . - & ® - - . - - . 2 * L . * . - - * & 15.1 Literature Review. 15.2 Mathematical Analysis. 15.3 Calculational Procedure. . « « « « « + o o« o« o« o« STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS. 16.1 Removal of Oxide from Salt by Countercurrent Contact with an :[']_-_FI-I']'.2 GaS Stream- . . . . . » . * . . - . . . » . . . 16.2 Removal of Oxide from Column . . « « « + « v o« & + & « & 16.3 Iron Fluoride Reduction RUnms . « « « « o s+ o o o o o o 16.4 Calculated Values for the Mass Transfer Coefficient and the Reaction Rate Constant During the Reduction of Iron Fluoride ' REFERENCES . Page 219 233 238 238 240 2hé 251 252 254 256 256 261 vii SUMMARIES FLOWSHEET ANALYSIS: ISOLATION OF PROTACTINIUM BY OXIDE PRECIPITATION Two flowsheets that employ oxide precipitation for protactinium removal are described, and the effects of several operating parameters on the performance of the flowsheets have been investigated. In the first flowsheet, protactinium is selectively precipitated from MSBR fuel salt on a 3-day cycle. The resulting oxide and a small amount of fuel salt associated with it are hydrofluorinated in the presence of a secondary fluoride salt that is circulated through a fluorinator and a protactinium decay tank. A small fraction of the salt leaving the decay tank is returned to the primary reactor circuit to compensate for salt that is transferred to the decay tank along with the oxide. The uranium is removed from 10%Z of the fuel salt leaving the precipi- tator by fluorination or oxide precipitation, and rare earths are removed from the resulting salt by the metal transfer process. The purified salt leaving the metal transfer process is combined with the uranium removed earlier, and the resulting stream is returned to the reactor. A protactinium removal time of 5 days can be realized if 607 of the protactinium is separated from the salt in the precipitator, provided the fuel salt transfer rate to the decay tank is as low as 10 to 20 moles/day. For the same protactinium removal time, a pro- tactinium removal efficiency of 80% would be required in the precipitator if the fuel salt transfer rate to the decay tank were as large as 3000 moles/day. The uranium inventory in the decay tank would be negligible. In the second flowsheet, a fluorinator is not used for removal of uranium from the protactinium decay tank. Fuel salt is withdrawn from the reactor on a 3-day cycle and combined with a salt stream that is withdrawn from the protactinium decay tank. Part of the protactinium in the resulting salt stream is removed by precipitation, and the pre- cipitate and associated salt are hydrofluorinated in the presence of processed fuel carrier salt leaving the metal transfer process. The resulting salt stream then passes through a decay tank, from which it viii is fed to the protactinium precipitator in order to return the uranium to the reactor. Operation of the flowsheet is highly dependent on the fraction of the protactinium removed in the precipitator and on the amount of fuel salt that accompanies the oxide precipitate. FLOWSHEET ANALYSIS: REFERENCE PROCESSING PLANT FLOWSHEET BASED ON FLUORINATION, REDUCTIVE EXTRACTION, AND THE METAL TRANSFER PROCESS Operating conditions that will constitute the reference fluori- nation--reductive extraction--metal transfer flowsheet were selected, and additional calculations were performed to indicate the operating characteristics of the flowsheet. ZEssentially complete extraction of the protactinium is achieved with a 10-day processing cycle, a five- stage protactinium extractor, a lithium reductant addition rate of 200 equiv/day, and a uranium removal efficiency of 99%Z in the primary fluorinator. Rare earths are extracted from the fuel salt with removal times ranging from 16 to 50 days in a three-stage extractor. A three- stage extractor is also used for the selective transfer of the rare earths from the bismuth-plus-thorium phase and the extracted rare earths to a LiCl stream. The various waste salt streams produced by the pro- cessing system are combined into a single stream having the composition 76.3-12.3-9.8-0.64 mole 7 LiF—ThFa—BeFZ-ZrF4, 0.864 mole 7 trivalent rare-earth fluorides, and 0.114 mole % divalent rare-earth fluorides. The waste salt would be discarded from the processing system at the rate of 70 ft3 every 220 days. FLOWSHEET ANALYSIS: IMPORTANCE OF URANIUM INVENTORY IN AN MSBR PROCESSING PLANT The MSBR processing flowsheets considered to date have resulted in uranium inventories in the processing plant that are quite low, usually less than 17 of the inventory in the reactor. Since several potential processing flowsheets may result in uranium inventories as large as 10% of the reactor inventory, the importance of increases in this inventory was examined. It was found that increasing the ix processing plant inventory from O to 10% would increase the fuel cycle cost by only 0.03 mill/kWhr and would increase the system doubling time from 22 to 24.2 years. It was concluded that, while there are incentives for maintaining a low inventory, inventory values of 5 %o 10% would not rule out an otherwise attractive processing system. FLOWSHEET ANALYSIS: REMOVAL OF RARE-EARTH FISSION PRODUCTS FROM LiCl IN THE METAL TRANSFER PROCESS Calculations were made to determine the effect of varying the con- centration of lithium in the bismuth solution used for removing the trivalent rare earths from the LiCl in the metal transfer process; the reactor breeding ratio was found to decrease only slightly (from about 1.063 to about 1.060) as the lithium concentration in the bismuth was decreased from 5 at. % to approximately 1.67 at. 4. Calculations were also carried out which indicate that a single-stage extractor has essen- tially the same removal efficiency for the divalent rare earths in the reference flowsheet as a two-stage contactor; thus the use of a single- stage contactor was adopted. FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON INDUCTION HEATING IN A CONTINUOUS FLUORINATOR SIMULATION An experiment to demonstrate protection against corrosion by the use of a layer of frozen salt in a continuous fluorinator requires a corrosion-resistant heat source to be placed in the molten salt. High- frequency induction heating appears to be an acceptable heating method, and equipment has been installed for studying this method in a simu- lated fluorinator that uses a 31 wt Z HNO3 solution in place of molten salt. Experimental results on heat generation rates in the acid, in the pipe surrounding the acid column, and in the induction coil are presented for the first eight runs. PREDICTED CORROSION RATES IN CONTINUOUS FLUORINATORS EMPLOYING FROZEN-WALL PROTECTION Nickel is the preferred material of construction for fluorinators in MSBR processing plants since it exhibits greater resistance to attack by gaseous fluorine than other candidate materials. This resistance is due to the formation of a tightly adherent film of Nin, and it is proposed that a layer of frozen salt be used to prevent removal of the NiF, film via dissolution in the molten fluoride mixture that flows 2 through the fluorinator. However, it is expected that the NiF, film will be removed periodically as the result of deviations from ihe desired mode of operation, and an analysis was carried out for estimating the resulting corrosion rate under such conditions. It was found that, if the NiF2 film were destroyed 52 times per year, the average yearly corrosion rates at 450°C would be 2.9 mils and 0.97 mil for types 200 and 201 nickel respectively. It appears that either material will show satisfactory corrosion resistance if the NiF2 film is destroyed less frequently than once per week. PREDICTED PERFORMANCE OF CONTINUOUS FLUORINATORS Previous data on the extent of removal of uranium from a molten fluoride salt in a l-in.-diam, open-column fluorinator and recently obtained data on axial dispersion in open bubble columns were used to develop a mathematical model for predicting the performance of con- tinuous fluorinators having diameters ranging from 6 to 12 in. The results of the analysis are encouraging since they suggest that single fluorination vessels of moderate size will suffice for removing uranium from MSBR fuel salt prior to the isolation of protactinium. The ref- erence MSBR processing flowsheet requires fluorination of fuel salt at the rate of 170 ft3/day and a uranium removal efficiency of 99%; the present analysis indicates that an 8-in.-diam fluorinator having a height of 17.8 ft will meet these requirements. x1i MEASUREMENT OF AXTIAL DISPERSION COEFFICIENTS AND GAS HOLDUP IN OPEN BUBBLE COLUMNS Measurements of gas holdup and axial dispersion were made in open bubble columns having diameters of 1, 1.5, 2, 3, and 6 in. for a range of operating conditions. The effects of changes in the viscosity and surface tension of the liquid, the superficial gas velocity, the gas inlet-orifice size, and the number of gas inlets were determined. These data, as well as data obtained previously, were used to develop correlations for predicting gas holdup and axial dispersion in open- column, gas—liquid contactors such as continuous fluorinators in which a molten fluoride salt is countercurrently contacted with a gaseous mixture of fluorine and UF,. 6 SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL FACILITY We have continued to operate a facility in which semicontinuous reductive extraction experiments can be carried out in a mild-steel system. We are presently studying the mass transfer performance of an 0.82-in.-ID, 24-in.-long column packed with 1/4-in. molybdenum Raschig rings. Several experiments were carried out previously in which a salt stream containing UF4 was countercurrently contacted with bismuth containing reductant over a range of operating conditions. In order to measure mass transfer rates in the column under closely controlled conditions and under conditions where the controlling resistance is not in the salt phase (as was the case in previous exper- iments), preparations were begun for experiments in which the rate of exchange of zirconium isotopes will be measured between salt and bismuth phases otherwise at chemical equilibrium. Techniques for the production and charging of 97Zr (half-1life, 16.8 hr) to the salt were developed, and about 7 mCi of 97Zr was added to the salt in the feed tank. The first experiment using the 97Zr tracer was interrupted by a leak in the salt exit line from the feed tank. Because damage to the feed tank and Calrod heaters on the vessel made salvage of the xii tank impractical, a new vessel was fabricated and installed. Examin- ation of a specimen from the original vessel revealed that, although some graphitization of the steel had occurred, no evidence of embrittle- ment was present., DEVELOPMENT OF THE METAL TRANSFER PROCESS: OPERATION OF EXPERIMENT MTE-2 The second engineering experiment (MTE-2) for development of the metal transfer process was completed. This experiment was performed at 650°C in a 6-in.-diam carbon steel vessel that was divided into two compartments interconnected at the bottom by a pool of thorium-saturated molten bismuth. One compartment contained MSBR fuel carrier salt (72- 16-12 mole % LiF—Ber-ThF4) to which were added 7 mCi of 147Nd and sufficient LaF3 to produce a concentration of 0.3 mole %Z. The second compartment contained LiCl, a 35 at. % Li-Bi solution (in a cup), and a pump for circulating the LiCl through the cup at the rate of about 25 cmB/min. Gas~1ift sparge tubes were used to disperse droplets of bismuth in the salt phase and thereby improve contact of the phases. During a 3-month operating period, in which a total of 563 liters of LiCl was circulated through the cup containing the Li-Bi solution, more than 857 of the lanthanum and more than 507 of the neodymium were removed from the fluoride salt. No measurable accumulation of thorium in the Li-Bi solution (<10 ppm) was noted during this period. The observed values for the distribution coefficients for lanthanum, neo- dymium, thorium, and radium during the experiment were in general agreement with the expected values. From 70 to 100% of the quantities of the rare earths charged to the system could be accounted for through- out the experiment. A much greater decrease was observed in the con- centration of lithium in the Li-Bi solution than was expected; the reason for this discrepancy has not been determined. Eight days before the end of the experiment, 1 vol % of fuel carrier salt was added to the LiCl in order to study the effect of contamination of the LiCl with fluoride salt. All of the objectives of the experiment were achieved. xiii DEVELOPMENT OF THE METAL TRANSFER PROCESS: DESIGN OF EXPERIMENT MTE-3 Design of the third engineering experiment for development of the metal transfer process has been initiated. This experiment (MTE-3) will use salt and bismuth flow rates that are 1% of the estimated flow rates required for processing a 1000-MW(e) reactor. Mechanical agitators will be used for promoting mass transfer between the salt and metal phases in the experiment. A mathematical analysis was carried out in order to select approximate equipment sizes and to determine operating conditions for the system. The experiment will use about 35 liters of MSBR fuel carrier salt, 6 liters of Th-Bi solution, 6 liters of LiCl, and about 5 liters of Li-Bi solution having an initial lithium content of about 5 at. %Z. The salt-metal contactor will be a 10~in.-diam, two-compartmented vessel having a mechanical agitator in each compartment. DEVELOPMENT OF MECHANICALLY AGITATED SALT-METAL CONTACTORS A program was initiated for the development of mechanically agitated salt-metal contactors as an alternative to packed columns presently under consideration for MSBR processing systems. This type of contactor is of particular interest for the metal transfer process since designs can be envisioned in which the bismuth phase would be a near-isothermal, inter- nally recirculated, captive phase. It is believed that such designs will be less dependent on the technology for molybdenum fabrication than would a counterpart system based on packed columns. Preliminary tests on the hydrodynamics of mechanically agitated salt-metal contactors were carried out using mercury and water. Initially, tests were made using an agitator that was operated at the water-mercury interface in a manner designed to disperse the mercury in the water. However, results of these tests led us to conclude that the contactor should operate under conditions that minimize dispersion of the mercury. The Lewis contactor appears to have the greatest potential for achieving effective mass transfer rates with minimum dispersion of the phases. In this contactor, an agitator, located well away from the interface, is present in each phase. Each agitator is Xiv operated in a manner such that the phases are mixed as vigorously as possible without actually dispersing one in the other. Information in the literature on mass transfer rates in Lewis~type contactors was reviewed. It was concluded that the mass transfer rate correlation developed by Lewis may be applicable to salt-bismuth systems, and that adequate mass transfer rates for MSBR processing applications should be obtained. HYDRODYNAMICS OF PACKED-COLUMN OPERATION WITH HIGH-DENSITY FLUIDS Studies of the hydrodynamics of packed column operation were con- tinued, using fluids with high densities and a large density difference. Data were obtained in a 2-in.-diam, 24-in.-long column that was packed with 3/8-in. Teflon Raschig rings for determining the dependence of dispersed-phase holdup, pressure drop, and flooding on the viscosity of the continuous phase. An improved relationship was developed for predicting packed-column performance during the countercurrent flow of molten salt and bismuth. The effects of wetting of the packing by the metal phase on metal phase holdup, flooding, and pressure drop were also evaluated in a 2-in.-diam, 24-in.-long column packed with 3/8- in. copper Raschig rings that were wetted by the mercury. The inter- facial area between the aqueous and mercury phases was decreased sub- stantially when the packing was wetted, and the column throughput at flooding was about 40% greater than with nonwetted packing. ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN MOLTEN SALTS AND LIQUID BISMUTH DURING COUNTERCURRENT FLOW IN PACKED COLUMNS The transfer of materials between a molten salt and liquid bismuth results in a condition where the fluxes of the transferring ions are dependent on both concentration gradients and electric potential gra- dients. This greatly complicates the mass transfer process and makes the design of continuous reductive extraction columns difficult. A math- ematical analysis of mass transfer during reductive extraction processes XV was carried out to facilitate interpretation of results from present and proposed experiments in packed columns and as an aid in using these data for the design of larger reductive extraction systems. A calculational procedure was developed for solving the resulting relations with as many as ten transferring materials. Provision was made for calculating rates of mass transfer between solvent and electrolyte phases for a range of operating conditions. In future work, particular attention will be paid to the influence of the electric field on the rate of mass transfer and to the differences that result from the case where mass transfer rates are assumed to be dependent only on concentration gradients. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS Salt purification studies using 66-34 mole 7% LiF—BeF2 were terminated because of leaks that resulted in the loss of about half of the 1l4-liter salt charge. The composition of the remaining salt was adjusted to the approximate composition of the proposed MSBR fuel salt (72-16-12 mole % LiF—Ber-ThFA). The newly prepared salt was then countercurrently con- tacted with a H,--10% HF mixture in the packed column in order to remove oxide from the zalt. Although a considerable quantity of oxide was removed from the salt, a significant quantity still remained in the column. In the two flooding runs and one iron fluoride reduction run that were carfied out during this report period, the pressure drop across the column increased sufficiently to make operation of the system difficult., The column was then filled with molten salt, and an HF-H2 stream was allowed to contact the static salt charge for a period of 18 hr in order to remove the oxide from the column. After this opera- tion had been determined to be successful, eight additional iron flu- oride reduction runs were completed. Operation of the system was smooth in each case, and the pressure drop across the column remained low. How- ever, the results of iron analyses of the salt samples from the runs were inconsistent. This inconsistency was probably due to the low iron concentration in the system, although sample contamination was suspected in some cases. 1. TINTRODUCTION A molten-gsalt breeder reactor (MSBR) will be fueled with a molten fluoride mixture that will circulate through the blanket and core regions of the reactor and through the primary heat exchangers. We are developing processing methods for use in a close-coupled facility for removing fission products, corrosion products, and fissile materials from the molten fluoride mixture. Several operations associated with MSBR processing are under study. The remaining parts of this report discuss: 1. the description and analysis of a flowsheet for isolating protactinium from MSBR fuel salt by oxide precipitation, the description of the reference flowsheet for processing MSBR fuel salt by the fluorination--reductive extraction-- metal transfer process, an analysis of the importance of the uranium inventory in a processing plant, the results of calculations related to the removal of rare earths from molten LiCl in the metal transfer process, experiments conducted in a simulated continuous fluorina- tor for studying induction heating in molten salt, predictions of the rate of corrosion of the nickel vessel in continuous fluorinators employing frozen- wall corrosion protection, predictions of the extent of removal of uranium in con- tinuous fluorinators, measurement of axial dispersion coefficients and gas holdup in open bubble columns and the development of correlations for predicting these quantities, experiments conducted in a mild-steel reductive extraction facility, to increase our understanding 10. 11. 12. 1%. 14, 15. of the rate at which uranium is extracted from molten salt into bismuth in a packed column, operation of experiment MTE-2 for demonstrating the metal transfer process for the removal of rare earths from MSBR fuel carrier salt, design of experiment MTE-3 for studying operation of the metal transfer process using salt and bismuth flow rates that are 1% of those expected for processing a 1000-MW (e ) MSBR, development of mechanically agitated salt-metal con- tactors, studies of flooding, dispersed-phase holdup, and pressure drop during countercurrent flow of liquids having a large difference in densities in packed columns, analysis of multicomponent mass transfer between molten salts and liquid bismuth during countercurrent flow in packed columns, and studies of the purification of salt by continuous methods. This work was carried out in the Chemical Technology Division during the period October through December 1970. 2. FLOWSHEET ANALYSIS: ISOLATION OF PROTACTINIUM BY OXIDE PRECIPITATION M. J. Bell L. E. McNeese Ross, Bamberger, and Baesl have shown that protactinium can be pre- cipitated selectively as Pa205 from MSBR fuel salt by the addition of oxide to salt containing Pa>t, and that Pa4+ can be readily oxidized to Pa5+ by hydrofluorination. Mailen2 has measured the solubility of PaZO5 in MSBR fuel salt that is saturated with UO2 at temperatures between 550 and 650°C. Also, Bamberger and Baes3 have found that uranium oxide can be precipitated from protactinium-free fuel salt as a U02—Th02 solid solution in which the concentration of UO2 at equilibrium is dependent on the concentration of UF4 in the salt. Bell and McNeese4 have used the equilibrium data of Bamberger and Baes to calculate the performance of a countercurrent multistage uranium oxide precipitator and have found that greater than 997 of the uranium can be removed from fuel salt as a U02-Th02 solid solution that contains less than 107 ThO2 by using only a few equilibrium stages in which the salt and oxide are countercurrently contacted. These results indicate that oxide precipita- tion may be an attractive alternative process to fluorination-reductive extraction for isolating protactinium and removing uranium from the fuel salt of an MSBR. Two flowsheets that employ oxide precipitation are described in the remainder of this section, and the effects of several operating parameters on the performance of the flowsheets are discussed. 2.1 Isolation of Protactinium by Oxide Precipitation, and Recovery of Uranium Daughters by Fluorination Figure 1 presents a flowsheet and typical operating parameters for a process which employs oxide precipitation to isolate protactinium from MSBR fuel salt and fluorination to recover uranium produced by decay of the protactinium. Fuel salt is withdrawn from the reactor on a 3-day cycle, and protactinium is selectively removed by precipitation as Pa205. The precipitate and a small amount of salt associated with it are hydro- fluorinated in the presence of a secondary salt that is circulated through e 8.3x104 mole /day 275 mole/day U REACTOR 2250 MW (th) 1680 ft3 RECOMBINER X pof,=8.4x10° ORNL DWG 70-14086R1 URANIUM mdz-:d-” U METAL RARE REMOVAL |22 0] TRANSFER |—e EARTHS 999, SYSTEM Sr,Ba,Zr,U UFg ' 1 a 8.3x10° mole/day POT . S « REC'P',A OR1 5.2 molesday OXIDE dOA 60% p— — = o"" . e EFFICIENCY | 200 mole/day SALT n:}f XpaF;20xI0 0.66 mole/day o= XyF,=0.0033 UFy T Hz' HF 200 mole/day 0.4 mole/day PaF4 a o Pa DECAY TANK 3 395 moles PoF, & x [™] 10.0moles UF, g 0.19x10* moles = - 3 XyF,=56xI0 F2 0.19x10® mole/day ; Fig. 1. Flowsheet for Isolating Protactinium by Fluorination-Oxide Precipitation and Removing Rare Earths by the Metal Transfer Process. T a fluorinator and a protactinium decay tank. A small fraction of the salt leaving the decay tank is returned to the primary reactor circuit to compensate for salt accompanying the oxide precipitate. The main salt stream exiting from the precipitator vessel contains most of the fission products and uranium, plus 5 to 40%Z of the pro- tactinium in the salt leaving the reactor. Ten percent of this stream is processed to reccover a large fraction of the uranium, and rare earths are removed from the resulting salt by the metal transfer process. Puri- fied salt leaving the metal transfer process is combined with the recov- ered uranium and then returned to the reactor. Removal of the uranium can be accomplished either by fluorination or by oxide precipitation. A mathematical analysis of the protactinium isolation portion of the flowsheet was carried out using the nomenclature shown in Fig. 2. The following material balance relations can be written for protactinium: FO = (F + FS)*PEFF:CP1 s (1) CP2 = CP1* (1 ~ PEFF) , (2) (F + A VR):CPR = P + F-(CP2 . (3) A*(VR*CPR + VT-CPT) = P . (4) and FO + FS:CP2 = (FS + X+VT)*CPT . (5) where FO = flow rate of oxide leaving precipitator, moles/day, FS = flow rate of salt accompanying oxide leaving precipitator, moles/day, F = flow rate of salt leaving the reactor, moles/day, PEFF = protactinium removal efficiency in precipitator, r = 233pa decay constant, day ™1, VR = volume of salt in reactor, moles, VT = volume of salt in protactinium decay tank, moles, CP = concentration of protactinium in salt at point denoted by suffix (defined below), mole fraction. ORNL DWG 72-13525Ri F cu2 UFse Reactor cCP2 v S Precipitator Mydro- ° Pa Decay Efficiency - —7 ™ fluorinator }— g S | g Tank F Cul = PEFF Fo FPS o FPS Volume =VT CUR CPI F5 cu3 3 cu4 cP2 HF Fa FPS CUT CPT FS cuT CPT Fig. 2. Nomenclature Used in Mathematical Analysis of Flowsheet for Protactinium Isolation by Fluorination--Oxide Precipitation. The suffixes 1, 2, 3, 4, R, and T refer to the following locations in the flowsheet: 1, entering precipitator; 2, entering hydrofluorinator; 3, leaving hydrofluorinator; 4, entering protactinium decay tank; R, leaving reactor; and T, leaving Pa decay tank. In the analysis, the quantities F, PEFF, P, VR, VI, and A are known and the ratio FS/FQ is specified. The quantities FO, FS, CPl1, CP2, CPR, and CPT are to be determined. It should be noted that Eq. (4) implies that a negligible fraction of the total protactinium inventory is in the metal transfer system and assumes no removal of protactinium except by radioactive decay. However, as much as 5% of the protactinium inventory may be present in the metal transfer system in an actual processing plant. The equations to be solved are nonlinear because of the manner in which the unknown quantity FS enters the relations. The algorithm developed to solve Eqs. (1)-(5) involved assuming a value for FO, which fixes the value for FS and also reduces Eqs. (1)-(4) to a set of four linear algebraic equations that can be readily solved for the four unknown concentrations. Equation (5) is then used to improve the estimated value for FO, and the procedure is repeated until satisfactory con- vergence in the value of FO is obtained. After the concentrations of protactinium throughout the system and the quantities FO and FS have been determined, the concentrations of uranium throughout the system can be determined by the use of the following analogous set of material balance relations: (F + FS)*CU1l = F*CUR + FS-CUT ’ (6) FS.CUl + (FPS - FS)-CUT = FPS.CU3 s (7) Cu4 = (1 - H)-CU3 s (8) and FPS.CU4 + A-VT+CPT = FPS-CUT , (9) where CU = concentration of uranium in salt at point denoted by suffix defined above, mole fraction, H = uranium removal efficiency in fluorinator, FPS = salt flow rate from hydrofluorinator, moles/day. The quantities H and CUR are assumed to be fixed quantities. In making the analysis, it was assumed that the volumes of all vessels except the protactinium decay tank were negligible, and that the flow rates of the salt streams entering and leaving the fluorinator were equal. The effects of several parameters on the performance of the flow- sheet were calculated. As shown in Fig. 3, the protactinium removal time depends on the precipitator efficiency and the rate at which fuel salt is transferred to the protactinium decay tank along with the Pa205. A protactinium removal time of about 5 days can be realized if 60% of the protactinium is removed from the salt in the precipitator, provided the salt transfer rate to the protactinium decay tank is as low as 10 to 20 moles/day (a salt-to-oxide flow rate ratio of 2 to 4). If the salt- to-oxide flow rate ratio were as high as 600, a precipitator efficiency of about 80% would be required in order to obtain the same protactinium removal time. The uranium inventory in the decay tank depends on the efficiency of the fluorinator in the protactinium isolation loop and on the amount of fuel salt that is transferred to the protactinium decay tank along with the precipitate, as shown in Fig. 4. The uranium inven- tory in the decay tank will be only a small fraction of the uranium inventory in the reactor, and the associated inventory charge will be less than 0.001 mill/kWhr for a wide range of operating conditions. 2,2 Isolation of Protactinium by Oxide Precipitation Without the Use of Fluorination for Recovering 233y Produced by Decay of 233p, As shown in Fig. 5, the isolation of protactinium by oxide precipi- tation can also be carried out without the use of a fluorinator for 233P recovering uranium that is produced by the decay of the a. In this flowsheat, fuel salt is withdrawn from the reactor on a 3-day cycle and combined with a salt stream that is withdrawn from the protactinium decay tank. Part of the protactinium in the resulting salt stream is removed by precipitation as PaZOS' The P3205 precipitate is hydrofluorinated in the presence of salt exiting from the metal transfer system, and the resulting salt stream then passes through a decay tank, where part of 233 233U. the Pa decays to The salt stream leaving the protactinium Pa REMOVAL TIME(days) ORNL DWG 70-14098 IS I I T T 1 T Pa CYCLE TIME:3 DAYS FLUORINATOR EFFICIENCY:95°%% RATIO OF SALT FLOW RATE TO OXIDE FLOW RATE 2 - - 600 9 - 400 200 40 6 - 0 3 i 1 L | 1 L 1 < 20 30 40 50 60 70 80 90 100 Pa PRECIPITATOR EFFICIENCY (%) Fig. 3. Effects of Protactinium Precipitator Efficiency and Rate of Transfer of Salt with Precipitate on Protactinium Removal Time for Fluorination--Oxide Precipitation Flowsheet. FISSILE U INVENTORY IN Pa DECAY TANK (% OF REACTOR FISSILE INVENTORY) ORNL DWG 70-14099R! T T T ! | | 0.0008 0.26 |- a 90% PRECIPITATOR EFFICIENCY - 0.24 | DAY SALT RESIDENCE TIME 0.0007 X poF, <0.0022 REACTOR FISSILE INVENTORY=1460 Kg 10% INVENTORY CHARGE 0.22 |- FLUORINATOR EFFICIENCY — 0.0006 0.20 0.18 ~-10.0005 0.16 1 | | [ i I 1 0 100 200 300 400 500 600 700 800 RATIO OF SALT TRANSFER RATE TO OXIDE FLOW RATE (MOLES SALT /MOLE OXIDE) Fig. 4. Effects of Rate of Transfer of Salt-plus-Precipitate and Fluorinator Efficiency on Uranium Inventory in Protactinium Decay Tank for Fluorination--Oxide Precipitation Flowsheet. hr) INVENTORY CHARGE (mill/ kW 01 56 tt3/day RECONSTITUTION ‘273 mole/day U 5500 mole/day SALT (4 t3) ORNL DWG 70-14087R1 RARE —a EARTHS Sr, Bo, Zr, U Pa DECAY TANK 150 ft3 U inventory=11.4 kg = -8 Xy, *250x10 X PqF, *0.0020 2.75 REACTOR 60f3/d URANIUM METAL 2250 MW(th | 60f7/day | RemovAL |-mole/day L) o ANSFER 1680 ft3 99% EFFY SYSTEM ] -6 XpgF,=8.4x107¢ XPaF,20x10 7 s ¢13/day XyF, =0.0033 4 szOs Pa 90 mole/day HYDRO ~ PRECIPITATOR ey 6% EFFY FUEL SALT FLUORINATOR 96% 3600 mole/day Hz-HF 60 ft3/day Fig. 5. Flowsheet for Isolating Protactinium by Oxide Precipitation and Removing Rare Earths by the Metal Transfer Process. 11 12 decay tank is combined with the salt entering the precipitator in order 233 to return U to the reactor without including a large quantity of 255Pa. A mathematical analysis was carried out for the protactinium iso- lation portion of the flowsheet, using the nomenclature shown in Fig. 6. The concentrations of protactinium at various points in the flow- sheet are defined by the following material balance relations: (FPS + F):CPl = F-CPR + FPS-CPT , (10) FPS-CP3 = FO + FS-CP2 , (11) P = A-(VR‘CPR + VI-CPT) , (12) FPS*CP3 = (FPS + A-VT)-CPT , (13) CP2 = (1 — PEFF)-CPl , (14) and (FPS + F):CPl = FO + (FPS + F)-CP2 . (15) The quantities F, FPS, VR, VI, P, and A are known; the ratio FS/FO is specified; and the values of CPl, CP2, CP3, CPR, CPT, FO, and FS are to be determined. The equations were linearized by assuming a value for FO, and Eqs. (10)-(14) were solved for the five unknown concentrations. Equation (15) was then used to improve the estimate of FO. As in the analysis for the earlier flowsheet, Eq. (12) implies that the protac- tinium inventory in the metal transfer system is a negligible fraction of the protactinium inventory in the total system. Changes in the flow rates of the salt streams through the precipitator and the hydro- fluorinator as the result of precipitation and dissolution of oxide were neglected. The concentrations of uranium in the system are defined by the following material balance relations: (FPS + F):CUL = F-CUR + FPS-CUT , (16) FPS-CU3 = FS-CUl + (FPS — FS)-CUL , (17) ORNL DWG 72-13524 < 1 Uranium Removal and F Metal Transfer| Systems FPS-FS cu4 CP4=0 Reactor cuz cpP2 — wl Precipitator Hydro - Pa Decay Efficiency __F:_o_.. fluorinator Tank = PEFF — » > FPS Volume sVT FS cu3 cuz f cP3 cP2 HF CPT CuUuT Fig. 6. Nomenclature Used in Mathematical Analysis of Flowsheet for Isolating Protactinium Without the Use of Fluorination for Recovering 2330 Produced by Decay of 255Pa. 14 cut = (1 —H)-CUL , (18) and FPS:CU3 + A*VI'CPT = FPS-CUT , (19) where H = uranium removal efficiency in the step prior to the metal transfer system, and the other quantities are as defined previously. The mathematical model describing operation of the protactinium isolation system was used to investigate the effects of several operating parameters on the performance of the flowsheet. The variation of protac- tinium removal time with changes in the precipitator efficiency and the fraction of the salt from the metal transfer system which is fed to the protactinium isolation system are shown in Fig. 7 for a protactinium processing cycle time of 3 days, a rare-earth processing cycle time of 50 days, a decay tank volume of 150 fti, and a salt-to-oxide molar ratio of 40 in the stream leaving the precipitator. Since the rate of flow of salt between the decay tank and the precipitator must be relatively large in order to limit the uranium inventory in the decay tank, high precipitator efficiencies are required with this flowsheet. A precipi- tator efficiency of about 96% would be necessary in order to obtain a protactinium removal time of 5 days. The uranium inventory in the pro- tactinium decay tank is relatively sensitive to the amount of salt accompanying the oxide in the stream leaving the precipitator, as shown in Fig. 8. For a salt-to-oxide flow rate ratio of 50 (corresponding to a salt flow rate of ~ 5000 moles/day), the uranium inventory will be about 1% of the reactor fissile inventory and the associated inventory charge will be about 0.003 mill/kWhr. However, if separation of salt from the oxide is difficult and salt-to-oxide flow rate ratios of the order of several hundred are required (corresponding to a salt flow rate of about 5 x 1Obr moles/day), the uranium inventory in the protactinium decay tank will be about 5% of the reactor fissile inventory and the associated inventory charge will be about 0©.015 mill/kWhr. Pao REMOVAL TIME (days) ORNL DWG 70-14097 50 I T T I [ T | % OF SALT FROM METAL : TRANSFER SYSTEM TO Pa CY . o CYCLE TIME:3DAYS o TANK RARE EARTH CYCLE TIME:30DAYS 40 |~ RATIO OF SALT TRANSFER RATE B TO OXIDE FLOW RATE:40 100% 90% 5% \o% _ . N 20 - — 10+ -T } ] ] i | e | ] 20 30 40 50 60 70 80 90 100 Pa PRECIPITATOR EFFICIENCY (%) Fig. 7. Effects of Protactinium Precipitator Efficiency and the Fraction of Salt Fed from the Metal Transfer System to the Protactinium Removal System on Protactinium Removal Time. ®22U Inventory in Pa Decay Tank (% of Reactor Fissile Inventory) ORNL DWG 72-13528 T T T T I T Po Processing Cycle: 3 days ' 0.020 70 Rore Eorth Processing Cycle: 30 days r Pa Precipitator Efficiency : 99 % Reactor Fissile inventory : 6000 moles 60 Inventory Charge: [0 percent per annum ' , 3 Decay Tank Volume: 150 ft 40015 Percent of Salt from Metal 50 Transfer System to Decay Tank —~0.010 —~ 0.005 0 1 1 ] ] ] 1 1 0 o 50 100 150 200 250 300 350 400 Fig. 8. Ratio of Salt Tronsfer Rate to Oxide Flow Rote (Moles Salt / Mole Pa,0,) Transfer System to Decay Tank. Effects on Uranium Inventory of Salt-—to—Pa2 in Stream Leaving Precipitator and of the Percent of Salt Q0. Molar Ratio gent from Metal Inventory Charge (mill /kWhr) 21 17 5. FLOWSHEET ANALYSIS: REFERENCE PROCESSING PLANT FLOWSHEET BASED ON FLUORINATION, REDUCTIVE EXTRACTION, AND THE METAL TRANSFER PROCESS L. E. McNeese Previously, we described an improved method for removing rare earths from the fuel salt of a single-fluid MSBR and presented calculated results 5,6 on the system performance for a range of operating conditions. More T recently we described an improved method for removing protactinium from fuel salt. This method is based on the use of fluorination for removing uranium and reductive extraction for removing protactinium. The isoclated 253Pa is held for decay to 253 U in a secondary molten fluoride salt. We also devised a method for combining the various wastes streams produced during the isolation of protactinium and during the removal of rare T earths into a single stream. During this report period, we combined these three processing methods into a single flowsheet and adopted a set of operating conditions that constitute the reference fluorination-- reductive extraction--metal transfer flowsheet. The reference flowsheet is shown in Fig. 9. For a 1000-MW(e ) MSBR,_ fuel salt (71.7-16-12-0.33 mole % LiF-BeF —ThFh_UFM) is removed from the reactor at the rate of 0.88 gpm, which reiresents a 10-day processing cycle. The salt passes through a delay vessel, which results in an average decay period of 30 min, before being contacted with fluorine at the rate of 19.3 liters (STP)/min in a continuous fluorinator (employing frozen-wall corrosion protection) to remove 99% of the uranium from the salt stream. The fluorine feed rate is equivalent to 150% of that required for conversion of the UFH to UF6. The fluoride salt stream leaving the fluorinator is contacted with a 28.3-liter/min hydrogen stream in order to remove dissolved fluorine from the salt and to reduce the valence of the residual uranium from 5+ to 4+. The salt stream is then contacted countercurrently with a 0.072-gpm bismuth stream containing 0.011 equiv of reductant per gram-mole of bismuth in a salt-metal contac- tor (which is equivalent to five theoretical stages) in oraer to extract the protactinium and uranium from the salt. The bismuth stream leaving 18 ORNL DWG 72-13529 3 Fiuorinator Salt 4+ Fission Products To Waste Be F, HeHF Th F, b1 ] : Salt | UF, UF, J ) Purification Reduction Absorber -: i ! . Hy : | k l . UFs-F, Mo e m e - o =5 :. ______ . 1 | | Reactor| 1 I | 1 ! | Hy-HF 1 . 1 "___l Li | ! : _y__ - Soomesny [ D.lfly Fluorinator parger r -@ ; é Vessel ' LiCl : | 1 [ H,-HFl : : : : F' H. T ' l —{ E E : Hydro- Pa P fiuerinator UFe=Fa Hg-HF Decay Podeeemee Loo® | Tank T 1 T ! ! I _J HF | Fluorinator—» Sparger : ! I 1 1 | ? T I i F, H ! ! * ' | HF-H.I ' ! | t e ! . Waste Hydro - 1 \ ®) Holdup fluorinater 1 I I T I s ? t.a ! HF 1 I 1 1 I | ! | | | ' b Fig. 9. Reference Flowsheet for Processing MSBR Fuel Salt via Fluo- rination, Reductive Extraction, and Metal Transfer. 19 the salt-metal contactor is subsequently contacted with an 11.8-1iter/ min HF stream in the presence of a 0.78-gpm secondary fluoride salt stream in order to transfer the extracted materials (protactinium, uranium, and zirconium) to the secondary salt. After leaving the hydrofluorinator, the secondary salt passes through a fluorinator, where 90% of the uranium is removed as UF6 by contact with a 1.56-liter/ min fluorine stream. (The assumed value for the fluorine utilization during this step is 1T7%.) The secondary salt is then contacted with a 2.8-liter/min hydrogen stream in order to remove dissolved fluorine and to reduce the valence of the residual uranium from 5+ to L+. Finally, the salt passes through the protactinium decay tank, where 2 53Pa present in the reactor and processing plant is held 253, most of the for decay to . The 0.78-gpm secondary salt stream that is fed to the hydrofluorinator is taken from the protactinium decay tank. The fuel carrier salt stream leaving the protactinium extraction column is essentially free of uranium, protactinium, and zirconium; however, its rare earth concentration is about the same as that in the reactor. This stream is fed countercurrent to a 12.%-gpm stream of bismuth containing lithium and therium (0.0l11 equiv of reductant per mole of bismuth) in a three-stage contactor in order to extract frac- tions of the rare earths from the salt. The bismuth stream leaving the extractor contains the rare earths, thorium, and lithium; this stream is countercurrently contacted with a 33.%-gpm LiCl stream in a three-stage salt-metal contactor. Because of highly favorable dis- tribution ratios, the rare earths, along with a negligible quantity of thorium, transfer to the LiCl. The trivalent rare earths are removed from the LiCl in a single-stage contactor by contact with an 8.1-gpm recirculating bismuth stream that contains 5 at. % lithium. The net flow rate of the Li-Bi solution through the contactor is 21.5 liters/day. Two percent of the LiCl stream exiting from the trivalent rare-earth contactor is fed to a two-stage salt-metal contactor, where the divalent rare earths are removed by contact with a 0.67-gpm recir- culating bismuth stream that contains 50 at. % lithium. The net flow of the Li-Bi solution through the contactor is 2.2 liters/day. All of the extraction operations are carried out at 640°C. 20 The bismuth stream exiting from the divalent rare-earth contactor is fed to the hydrofluorinator below the protactinium extractor in order to add lithium to the protactinium decay tank at the rate of about 50 moles/ day. With this addition, the composition of the salt in the decay tank is 79.6-17.7-2.12 mole % LiF -ThF) -Z1F), 0.367 mole % divalent rare-earth fluorides, and 0.247 mole % of trivalent rare-earth fluorides. The salt has a liquidus temperature of about 600°C; at this temperature, the rare- earth fluoride concentration is well within solubility limits. Salt is removed from the protactinium decay tank at the average rate of (0.094 fti/day to eliminate the fluorides of lithium, thorium, zirconium, and the rare earths that accumulate in the secondary fluoride salt. About 0.12% of the fuel carrier salt leaving the rare-earth extractor is dis- carded as a means of removing LiF that is added durine the extraction of protactinium and the rare earths. The discards of fuel carrier salt and secondary fluoride salt from the protactinium decay tank are made on a 220-day cycle. During each cycle, 20.7 ft5 of the secondary fluoride salt and 44.2 ft3 of the processed fuel carrier salt are transferred to the waste-salt holdup tank. The Li-Bi stream leaving the trivalent rare- earth stripper is hydrofluorinated in the presence of the resulting salt mixture during the 220-day period that is allowed for decay of the 255Pa in the waste salt. The salt in the waste holdup tank would be fluorinated 253 ' in order to recover the tion of the waste salt would be 76.3-12.%-9.8-0.64 mole % LiF -ThF, -BeF .- U before the salt is discarded. The composi- ZrF , 0.864 mole % trivalent rare-earth fluorides, and 0.114 mole % divalent rare-earth fluorides. Although the liquidus temperature of the salt is near 550°C, the salt temperature would have to be maintained at about 600°C to prevent precipitation of trivalent rare-earth fluorides. Waste salt would be discarded from the processing system at the rate of 70 ft5 every 220 days. The 0.072-gpm bismuth stream leaving the hydrofluorinator below the protactinium extractor is combined with reductant (200 moles of 7Li per day ), and the resulting stream is effectively returned to the protactinium extractor. The stream is actually combined with the 12.3%- gpm bismuth stream that circulates through the rare-earth removal system, 21 and an equal quantity of material is removed and fed to the protactinium extractor; this mode of operation allows for the removal of materials that would tend to accumulate in the otherwise captive bismuth phase in the rare-earth removal system. A small quantity of bismuth (2.2 liters/ 7 day ) is combined with sufficient 'Li to produce a stream containing 50 at. % lithium for return to the divalent rare-earth extractor. The processed fuel carrier salt remaining after discard of 0.12% of the salt stream exiting from the rare-earth extractor is combined with sufficient quantities of BeF and ThF)_L (47.8 and 47 moles/day, respec- tively) to make up for the discard of these materials in the waste salt and burnup of thorium in the reactor. The resulting fuel carrier salt is then fed to the fuel reconstitution step, which is carried out in a vessel having two compartments. To the first compartment are fed the 63-37 mole % UF6—F2 stream from the fluorinators (at the flow rate of 20.9 liters/min), the processed fuel carrier salt (at the rate of 0.88 gpm), and a 1.7-gpm fuel salt stream that is recycled from the second compartment. The rate at which fuel salt is recycled from the second compartment is such that the quantity of UFM in the recycled fuel salt is sufficient to give an average uranium fluoride valence of 4.5 (an equimolar mixture of UFu and UF5)~in the resulting 2.58-gpm salt stream leaving the first compartment. The 2.58-gpm salt stream leaving the first compartment is contacted with a 12.8-liter/min hydrogen stream in the second compartment in order to reduce the UF5 to UFM' A hydrogen utilization of 50% is assumed during this operation. Fuel salt is with- drawn from the second compartment for return to the reactor at the rate of 0.88 gpm, and the remaining salt is recycled to the first compartment. Before being returned to the reactor, the fuel salt is contacted with a 12.8-liter/min hydrogen stream to effect reduction of 1% of the UFM to UFB- The salt is also contacted with nickel wool as a means of removing traces of bismuth from the salt before its return to the reactor. The protactinium removal time obtained with the reference flowsheet is 10 days; the rare-earth removal times range from 16 to 50 days. The calculated value for the breeding ratio is about 1.065. 22 L. FLOWSHEET ANALYSIS: IMPORTANCE OF URANIUM INVENTORY IN AN MSBR PROCESSING PLANT M. J. Bell L. E. McNeese The MSBR processing flowsheets considered in the past have uni- formly resulted in very low uranium inventories in the processing plant, that is, inventories below 1% of the uranium inventory in the reactor in most cases. Since several potential processing systems might result in uranium inventories as large as 5 to 10% of the reactor inventory, we have investigated the importance of increases in the uranium inventory. The major effects of an increased uranium inventory are: (1) an increase in inventory charges on fissile material, and (2) an increase in the reactor doubling time. The variation of each of these quantities with processing plant uranium inventory is shown 233 in Fig. 10. The fissile inventory (which includes the Pa in the processing plant ) was assumed to be 1504 kg for the reference reactor Z 253y was taken to be $14/g, and and processing plant, the value of the capital charge rate was assumed to be 10% per year. The calcu- lated system doubling time for the limiting case of a zero uranium inventory in the processing plant is 22 years. It is seen that a processing plant uranium inventory of 5% of the system fissile inven- tory would increase the fuel cycle cost by 0.015 mill/kWhr and would increase the system doubling time from 22 to 23.1 years. A uranium inventory of 10% of the system fissile inventory would result in a fuel cycle cost increase of 0.03 mill/kWhr and an increase in doubling time from 22 to 24.2 years. Thus, while there is incentive for main- taining a low uranium inventory in the processing plant, it does not appear that a uranium inventory as high as 5 to 10% of the system fissile inventory would rule out an otherwise attractive processing system. INCREMENTAL INVENTORY CHARGE (mill/kWhr) ORNL DWG 71-2860 0'025 T ]’ T l T r v ' v I v ' ¥ l T 0.020 |- 0.015 — ——— 0.010 SYSTEM FISSILE INVENTORY ... 1504 kg VALUE OF 233y ... $14/9 0.005 - CAPITAL CHARGE RATE........... 10% /year o 1 4 ] L ] ) ] . 1 3 ] g ] 3 0 f 2 3 4 5 6 7 % OF SYSTEM FISSILE INVENTORY Fig. 10. Effects of 255U Inventory in Processing Plant on System Doubling Time and Inventory Charges. 25 24 23 22 21 20 DOUBLING TIME (years) ¢c 2L 5. FLOWSHEET ANALYSIS: REMOVAL OF RARE-EARTH FISSION PRODUCTS FROM LiCl IN THE METAL TRANSFER PROCESS M. J. Bell L. E. McNeese A flowsheet has been described previously8’9 for removing rare- earth fission products from MSBR fuel salt using the metal transfer process, which employs contact of LiCl with Li-Bi solutions for removing the rare earths and other fission products. 1In the reference flowsheet, the trivalent rare earths are removed by contacting LiCl at the rate of 33.4 gpm in a single equilibrium stage with an 8.1- gpm recirculating bismuth stream having a lithium concentration of S5 at. %. Bismuth containing extracted rare earths is withdrawn at the rate of 5.7 gal/day, and an equivalent amount of Li-Bi solution is added. Early data indicated that mutual solubility problems might be encountered between thorium and trivalent rare earths in bismuth having a lithium concentration as high as 5 at. %. Although this has been found not to be the case,10 we have made calculations to determine the effect that varying the lithium concentration in the lithium-bismuth alloy has on the thorium concentration in the metal and on the reactor performance. The results are presented in Fig. 11, which shows the effect of increasing the flow rate of the Li-Bi withdrawal stream while holding the amount of added reductant constant. The thorium concentration in the metal is reduced from 420 ppm at the reference withdrawal rate of 1000 moles/day to 1LO ppm at the discard rate of 3000 moles/day. The effect on reactor performance is slight; the breeding ratio decreases from about 1.063 to about 1.060. It would be possible to compensate, in part, for this loss in breeding ratio by operating the trivalent rare-earth stripper as a once-through batch contactor having more than one stage. In the reference flowsheet, the divalent rare-earth fission prod- ucts are removed from the LiCl by contacting 2% of the LiCl leaving the trivalent rare-earth stripper with a 50 at. % Li-Bi solution at the flow rate of 0.56 gal/day in a two-stage contactor. To accommodate MOLE FRACTION Th in Li-Bi SOLUTION (x 10%) 400 ORNL DWG 7i-2859R1 MOLE FRACTION Li IN METAL 0.05 0.033 0.025 0.020 0.0167 Li-Bi FLOW RATE (moles/day) Fig. 11. Effect of Lithium-Bismuth Discard Rate on tration in Discard Stream and on MSBR Performance. T T T T T 0.064 50 MOLES/DAY Li METAL IN Bi o 640°C 10063 — —— e - 0.062 p - -1 0.061 - -1 0.060 " L . A A A \ 0.059 1000 2000 3000 Thorium Concen- BREEDING GAIN (BREEDING RATIO-1.0) Gc 26 the high heat generation rates expected in the resulting Li-Bi solution, it is advantageous to operate the contactor as a continuous column through which the Li-Bi solution is recycled at a relatively high flow rate. A large fraction of the fission product decay heat could be dissipated by placing a decay tank in the recycle stream. However, this system would result in a contactor having only a single equilibrium stage. A calculation was made of the system performance for a flowsheet in which the divalent rare-earth contactor consisted of only one equilib- rium stage. Removal times for the divalent rare earths were increased only slightly over those obtained with a two-stage column; the resulting decrease in the breeding gain of the reactor was negligible. Therefore, the mode of operation described above has been adopted as part of the reference MSBR processing flowsheet. 27 6. FROZEN-WALL FLUORINATOR DEVELOPMENT: EXPERIMENTS ON INDUCTION HEATING IN A CONTINUOUS FLUORINATOR SIMULATION J. R. Hightower, Jr. C. P. Tung An experiment to demonstrate protection against corrosion by the use of a layer of frozem salt in a continuous fluorinator requires the use of a corrosion-resistant heat source in the molten salt. High- frequency induction heating has been proposed as the source of heat, and the estimated performance11 of a frozen-wall fluorinator having an induc- tion coil embedded in the frozen salt near the fluorinator wall has indicated that this may be an acceptable heating method. However, there are uncertainties associated with the effect of bubbles that will be present in the molten salt on the heat generation rate and in the amount of heat that will be generated in the metal walls of the fluorinator. Equipment has been installedl2 for studying heat generation in a simu- lated frozen-wall fluorinator by induction methods. In the simulation, a 31 wt Z HNO3 molten salts is used to simulate molten salt in the fluorinator vessel. solution with electrical properties similar to those of Results of the first eight experiments with the fluorinator simulation are described in the remainder of this section. 6.1 Experimental Procedure A simplified flow diagram for the continuous fluorinator simulation is shown in Fig. 12; the system is shown in greater detail in Fig. 13. The acid drain tank was filled with 28.4 kg of 31 wt % HNO3 that was used during the initial operating period and also during the first eight runms that were carried out. In each run, the acid recirculation system was filled with acid by pressurizing the drain tank to about 6 psig with valve V-7 open (see Fig. 13). When the liquid level in the column (as indicated by the level in the sight glass) reached the level of the top of the jacketed pipe, valve V-7 was closed and the pressure in the drain tank was reduced to 1 atm. The acid recirculation pump and the cooling water were turned on, and all flow rates were adjusted to the desired ORNL DWG 70-8964 RI 28 AIR COOLING WATER PUMP HEAT EXCHANGER COCGLING WATER JACKETED PIPE SN S S e N AAIRIA RN R SAINSEN Il \N\\\NN\\NNN\\\\N\\N\NNNN\NNNN\\\NN\\\N\HN\KNNN\\\NN\\\\\\\NI\\ bDbDDDDDDDbDDDDDDDD\ = / U U U U000 0D vy Oy Y g O U 0O U ~~~~~~~~~~~~~~~~~~~ i S e 0 2 AP W N 3 2 mul‘ flflflflflflflflflflflflflflflflflflflfl LTt T LTt T C TS TTTS ns ) 25 a3 o o NITRIC ACID 9 INDUCTION COIL Simplified Flow Diagram of a Fluorinator Simulation in ig. 12. Which Heat Is Generated by Induction Heating. COOLING WATER 29 ORNL DWG 70-14719 OFF—-GAS P | s FI-4 V-4 GLASS COLUMN | HEAT INDUCTION EXCHANGER COOLING WATER JACKETED < Y T PIPE INDUCTION § P COIL < h LS 9 10 « X Fl-2 Y o COOLING WATER TO JACKETED PIPE ..n.-J.n-A-y—-- - o DRAIN TANK Fig. 13. Flow Diagram of Fluorinator Simulation and Nitric Acid Recirculation System. 350 values. The radio frequency (rf) generator was then started, and the current to the induction coil was adjusted to a value between 100 and 160 A (rms value). The flow rates and the coil current were maintained at constant levels by manual adjustment until the temperatures around the recirculation system became steady. Steady-state temperatures were usually achieved within 1 hr of operation with constant values for the flow rates and coil current. After steady-state conditions had been established, the run was completed by recording the operating conditions. The rf power and the acid recirculation pump were then turned off, and the acid was drained from the recirculation system by opening valve V-7, 6.2 Experimental Results Eight runs have been carried out with the continuous fluorinator simulation to determine heat generation rates in the acid, in the pipe surrounding the acid column, and in the induction coil. Heat generation rates were calculated from the steady~state increase in the temperature of the acid as it passed through the column and in the temperature of the cooling water as it passed through the coil and the jacketed pipe. The experimental results for the first eight runs with the system are summarized in Table 1. During these runs, the induction coil consisted of 17 smaller coils arranged with alternate coils wound in opposite directions. Each coil section consisted of about 6.5 turns of 1/4-in.- diam Monel tubing that had been wound on a 5.6-in.-diam mandrel. The total length of the 1l7-section coil was 5 ft. The frequency of the rf current was 412 kHz during each of the runs. The heat generation rates in acid in the 5-ft-long column section, in the pipe representing the fluorinator wall, and in the induction coil were proportional to the square of the total coil current as expected. The ratios of the heat generation rates (in watts) to the square of the coil current (rms value, in amperes) were designated as effective resist- ances that will be used subsequently in circuit calculations for design of the rf generator. The average effective resistance for the acid was 0.0161 & at average acid temperatures of 25 to 29°C and 0.0177 Q for average acid temperatures of 46 to 51°C. The difference between these 2] Table 1. Summary of Results from Experiments on Induction Heating in a Simulated Continuous Fluorinator Total Average Run Coil Acid Heat Generation Rate Equivalent Resistance No. Current Temp. (W) () (CFS-) (A) (°C) Acid Pipe Coil Acid Pipe Coil 1 24.1 242 140 2 130 24.6 409 167 0.0242 0.0099 3 150 26.5 316 279 1442 0.0140 0.0124 0.0641 4 100 25.6 141 116 0.0141 0.0116 150 28,1 378 263 0.0168 0.0117 150 29.1 383 309 0.0170 0.0137 5 160 26.7 448 308 0.0175 0.0120 6 150 46.5 393 282 0.0175 0.0125 7 150 28.9 377 273 1356 0.0167 0.0121 0.0585 120 26.6 235 178 870 0.0163 0.0126 0.060 8 150 51.4 402 275 1442 0.0178 0.0122 0.0641 values may be within experimental error. The effective resistance of the acid should have increased by about 257Z as the temperature was increased from the lower to the higher value; however, the observed increase was only about 10%. The smaller variation in effective resis- tance with changes in temperature may be due to the use of short coil sections since the effective resistance of the acid should be proportional to the specific conductance for an infinitely long coil. The average effective resistance for the stainless steel pipe surrounding the acid was 0.0123 Q; the average effective resistance of the induction coil was 0.0617 Q. 52 The fraction of the total heat generation that occurred in the acid can be found by dividing the resistance of the acid by the sum of the resistances of the acid, pipe, and coil. About 197 of the total heat generatjon occurred in the acid during these runs. A 5-ft-long fluori- nator having a 4.7-in.-diam molten zone would require generation of about 8.3 kW of heat in the molten salt.11 If an efficiency of 197 were obtained for heating the molten salt, a 43.7-kW generator would be required. The calculated values for the resistances of the acid and coil under the experimental conditions are 0.135 and 0.017 & respectively. The dis- crepancies between the calculated and measured values probably result from the use of a coil composed of a number of short sections and the use of equations that were derived for infinitely long coils. We are presently evaluating alternative coil designs with a view toward improving the heating efficiency. 55 T. PREDICTED CORROSION RATES IN CONTINUOUS FLUORINATORS EMPLOYING FROZEN-WALL PROTECTION J. R. Hightower, Jr. Nickel is the preferred material of construction for fluorinators in MSBR processing plants since it exhibits greater resistance to attack by gaseous fluorine than other candidate materials. This resistance is due to the formation of a tightly adherent film of NiF,_ (a corrosion product ) through which additional fluorine must diffusi in order to cause further corrosion of the nickel. The rate of attack is greatly reduced, although not to zero, once the protective film of NiF2 is formed on the exposed nickel surface. 1In the proposed continuous fluorinators, a'layer of frozen salt will be formed on the fluorinator wall in order to prevent the protective NiF,. film from being dissolved 2 by the molten salt. 1In the present analysis, no credit is taken for resistance to diffusion of fluorine which may be afforded by the layer of frozen salt. Since the protective NiF, film will likely be destroyed 2 several times during operation of a fluorinator, an analysis has been carried out for estimating the corrosion rate under conditions such that the NiF_ film is destroyed periodically. 2 -1 Data on the Rate of Corrosion of Nickel in Gaseous Fluorine Data relative to the corrosion of Ni-200 and Ni-201 in gaseous fluorine at a pressure of 1 atm were collected from the literature. A summary of this information is given in Tables 2 and 3. It has been 15 shown - that the reaction of high-purity nickel with fluorine initially follows a parabolic rate law at temperatures of 300 to 600°C but that, after a period of time, the reaction rate decreases and follows a third- or higher-order rate law. The literature-derived corrosion-rate data were used to calculate rate constants for the reaction of fluorine with nickel in the temperature range of 360 to TO0°C under the assumption that the reaction rate follows a parabolic rate law at all times. The assumption of a parabolic rate law correctly represents the initial reaction rate but should yield high estimates of the corrosion rate for long periods of exposure. 3L Table 2. Calculated Parabolic Reaction Rate Constants for Ni-200 Exposed to Gaseous Fluorine Depth of Exposure Calculated - Parabolic Rate Constant Temp. Penetration Time 1/2 a °c) (mils) (hr) (mils/hr ) Reference 362 0.0048 120 0.000438 [14]1 (5) 400 0.00681 33.3 0.00118 [13} 400 0.00701 50 0.000991 [15] 475 0.0187 240 0.00242P [16] 475 0.0375 480 0.00342P [16] 475 0.0750 960 0.00484b [16] 500 0.0132 33.3 0.00229 [13] 500 0.0144 5 0.00643 [17] (27) 500 1.97 240 0.1272 [17] (25) 500 0.0754 50 0.01066 [171 (23) 500 0.0206 150 0.00168 [17] (13) 500 0.043 150 0.00352 [17] (13) 500 0.0243 9 0.0081 [14] 535 0.181 120 0.0165 [17] (18) 550 0.00299 6 0.00122 (141 () 550 0.192 120 0.0175 [14] (5) 600 0.0616 33.3 0.01067 [13] 600 0.00684 5 0.00306 [171 (27) 600 0.0657 96 0.00671 (17} (20) 600 0.315 96 0.0322 [171 (20) 600 0.00404 8 0.00143 [(17) (24) 600 0.0931 8 0.0329 [17] (24) 600 0.753 50 0.1065 (171 (23) 600 0.00887 13 0.00246 [18] 600 0.0736 28 0.0139 {18] 600 0.0417 28 0.00788 [18] 600 0.0287 77 0.00327 (18] 600 0.0171 77 0.00195 rig] 600 0.134 93 0.0139 [18] 600 0.0744 93 0.00772 [18] 600 0.0416 124 0.00374 (18] 600 0.0307 124 0.00276 [18] 600 0.162 132 0.0141 (18] 600 0.0936 132 0.00815 [18] 600 0.0690 195 0.00494 [18] 600 0.0628 195 0.0045 [18] 600 0.104 243 0.00666 (18] 700 0.103 8 0.0364 (17] (24) 700 0.0274 8 0.00969 [17]1 (24 700 3.28 240 0.212 [171 (25 700 0.727 5 0.325 (171 (@27 a . . : The numbers shown in brackets are primary references; the numbers shown in parentheses are reference numbers in the primary references. b Measured in F2—N2 (50-50%) mixture; rate constant corrected to 100% Fz. 35 Table 3. Calculated Parabolic Corrosion Rate Constants for Ni-201 Exposed to Gaseous Fluorine Calculated Parabolic Depth of Exposure Temp. Penetration Time Rate Coniignt (°c) (mils) (hr) (mils/hr™/ <) Reference” . 380 0.00085 5 0.000380 [14] (5) 500 0.00959 5 0.00429 [17] (27) 500 0.00899 6 0.00367 [14] 550 0.0115 5 0.00514 [14] (5) 600 0.00259 5 0.00116 [17]1 (5) 600 0.0471 28 0.00891 [18] 600 0.0283 28 0.00534 [18] 600 0.0936 7T 0.01067 [18] 600 0.235 77 0.0268 [18] 600 0.109 93 0.0113 [18] 600 0.111 93 0.0115 [18] 600 0.142 132 0.012k4 [18] 600 0.160 132 0.0139 [18] 700 0.0595 5 0.0266 [(17] (27) a . . The numbers shown in brackets are primary references; the numbers shown in parentheses are rjference numbers inm-the primary references. of nickel exposed to gaseous fluorine is described by the following relat where For the assumption of a parabolic rate law, the extent of corrosion ion: 0. it 1l W 1 d =kVt 3 depth of nickel attacked by F2, mils, (20) time of exposure of nickel metal to fluorine after a zero film thickness, hr, parabolic rate constant, mil hr 1/2_ 36 The rate constant values calculated from the literature data are shown in Tables 2 and 3, along with information on the length of exposure of the nickel specimens to fluorine, the extent of attack on the specimens, and the references from which data were obtained. It was assumed that the temperature dependence of the parabolic rate constants would be of the Arrhenius type; thus the calculated values were fitted to the following equation: In k = B + A/T , (21) where T = temperature, °K, A,B = constants, with the criterion that the best fit occurred when the quantity ]2 Z[ln k — 1n kobs was minimized. This criterion places more importance on reaction rate constant values resulting from low corrosion rates than does the usual criterion that the quantity )2 E:(k - kobs have a minimum value. This procedure was used since it was believed that experimental errors were likely to yield corrosion rates that were too high rather than rates that were too low. The resulting equations for the variation of the parabolic rate constants for Ni-200 and Ni-201 are as follows: In k = 0.3773 — 3961/T , (22) I and In k = 4.308% — 7836/T , (23) 57 where -1 parabolic rate constant, mil hr /2, o 1t T = temperature, °K. The variation of these constants with temperature is shown in Figs. 14 and 15. 7.2 Predicted Corrosion Rates If it is assumed that the protective NiF2 film is removed n times per year at equal time intervals, the extent of corrosion experienced during a l-year period is given by the expression: d = n kV/8760/n (2k) where d I corrosion rate, mils/year, n = number of times NiF, film is destroyed annually. 2 Figure 16 shows the variation of the average corrosion rate with fre- quency of destruction of the NiF, fium at 450°C, the approximate wall temperature that will be used inea frozen-wall fluorinator. If the NiF2 film were destroyed 52 times annually, the average corrosion rates at this temperature would be 2.9 mils/year and 0.97 mil/year for Ni-200 and Ni-201 respectively. It appears that Ni-201 is more resistant to corrosion by fluorine than Ni-200. However, either of these materials will show satisfactory corrosion resistance if the NiF2 film is destroyed less frequently than once per week. 33 ORNL DWG 72-353| Temperature (°C ) 100 €50 00 5850 800 480 400 380 e T T Ty Ty T T T T r [y T T T T I - - - L . - . - * o L . i : — a ' ] —~ w 0.0 - - < ~ EF 2 ] - - - £ - : 1 2 [~ . j - [~ E S . : b * —ef c . o . © . . o 0001 1 o - < - - x ~ . o - - = L . - o a » - 2 9 L - a 0000t | Variation of Rate Constant with Temperature for Corrosion fi E of Ni-200 in Fluorine 3 1 ] 1. 1 I 1 0.0000i 1.00 1o .20 .30 1.40 1.80 .60 170 1000 / T°K Fig. 1l4. Corrosion Rate Constant for Ni-200 as a Function of Temperature. 59 ORNL DWG 72-13530 TEMPERATURE (°C) 100 0 s00 $80 so0 480 400 380 1o ‘ - TrT LI DL L | T 17T 1 F T T 1 | [ 3 - vl 0.1 L ~ " i - o - -1 0.0i - — -~ N - - = - ~ | - E — x o = o 0.00I - “1 ° C - - - .y c - i o N - o [ i ° b - e o b -1 @ O - -— ) L O o.0001 | 7 o - 3 a — -~ - w— | - - e - — = — ! | | ] l 1 000001, 56 110 120 1.80 i.40 150 .60 i.70 1000 / T°K Fig. 15. Corrosion Rate Constant for Ni-201 as a Function of Temperature. ORNL DWG 72-13527 o , (mils /year) Lo g% Corrosion Rate o Avera o w [ T T TTTTI i P T T TTTT F0% T TTT] W — o1 | 1111 ] 1 ] 111 I { [] 10 30 I 80 100 n, times NiF, film is destroyed annually Fig. 16. Corrosion Rates of Ni-200 and Ni-201 at 450°C. Effect of Frequency of Destruction of NiF. Film on the 2 L1 8. PREDICTED PERFORMANCE OF CONTINUOUS FLUORINATORS L. E. McNeese J. S. Watson T. O. Rogers Most of the flowsheetslg_22 considered to date for processing MSBR fuel salt require fluorination of molten salt for removal of uranium at one or more points., These applications include: (1) removal of trace quantities of uranium from relatively small salt streams prior to discard, (2) removal of uranium from a captive salt volume in which 233Pa is accumulated and held for decay to 233U > (3) removal of most of the uranium from relatively large fuel salt streams prior to isolation of protactinium and removal of rare earths, and (4) nearly quantitative removal of uranium from a salt stream containing 233Pa in order to produce isotopically pure 233U. Not all of these applications require continuous fluorinators; in fact, the use of batch fluorinators results in definite advantages in certain cases. However, as the quantities of salt and uranium to be handled increase, the use of continuous fluorinators becomes mandatory in order to avoid undesirably large inventory charges on uranium and molten salt as well as the detrimental increase in reactor doubling time that is associated with an increased fissile inventory. We previously estimated23 the performance of continuous fluorinators by assuming that the rate of removal of uranium from the salt is first order with respect to the concentration of uranium in the salt. If the transfer of uranium in the salt by axial dispersion and by convection is taken into account, the concentration of uranium in the salt is defined by the following relation: 2 D-é~% - v-%% - dx kC =0 (25) Lo where D = axial dispersion coefficient, cmzlsec, C = concentration of uranium in salt, moles/cm3, X = position in column measured from top of column, cm, V = superficial salt velocity, cm/sec, k = reaction rate constant, sec—l. The terms in Eq. (25) represent the transfer of uranium in the salt by axial dispersion, the transfer of uranium in the salt by convection, and the removal of uranium from the salt by reaction with fluorine respec- tively. The assumption of the first-order reaction does not imply a particular rate-limiting reaction mechanism; however, it is consistent with the assumption that the rate-limiting step is diffusion of uranium in the salt to the gas-liquid interface. In this case, the first-order expression would imply that the concentration of uranium in the salt at the interface is negligible in comparison with the uranium concentration in the salt at points a short distance from the interface. The boundary conditions chosen for use with Eq. (25) assume that the diffusive flux across the fluorinator boundaries is negligible: at X = 0 (top of fluorinator), dc _ v X = -5 [Cfeed = Co4l ’ (26) X=0+ and at X = L (bottom of fluorinator), dc =0 . (27) dX " xar, where Cfeed = concentration of uranium in salt fed to the fluorinator, CO+ = concentration of uranium in salt at top of the fluorinator. Note that C0+ is not equal to Cfeed since there is a discontinuity in uranium concentration in the salt at the top of the fluorinator where the salt enters. ¥ Solution of Eq. (25) with the stated boundary conditions yields the following expression for the ratio of the uranium concentration in salt leaving the fluorinator to the concentration in the feed salt: feed 1 J__LE__ML + 1/2 E[‘V 1;)"' + n = 1/2] ng__fl ]/2 "é[l/g +,/1;)—l + 'r]] e — — e s 1 + 4 1 J1+ by J1 + by (28) where c(L) concentration of uranium in salt leaving fluorinator, n__EB ===, Vv VL € =7, L = length of fluorinator, cm. Application of Eq. (28) to the design and evaluation of continuous fluorinators requires values for the rate constant k and the axial dispersion coefficient D. When we made the earlier estimates of fluorinator performance,go only limited data were available for the axial dispersion coefficient; these data resulted from studies with air and water in 1.5-, 2-, and %-in.-ID columns. At that time it was assumed that the axial dispersion coefficient was represented by the following relation: D =5.22,/G , (29) where . . . . 2 D = axial dispersion coefficient, cm /sec, G = gas flow rate at top of fluorinator, cm3/sec. The rate constant, k, was evaluated from experimental data obtained with a l-in.-diam open-column, continuous fluorinator.20 In correcting the data for the effect of axial dispersion, results obtained with the 1l.5- in.-diam column were used and no correction was made for the differences in the physical properties of molten salt and water. Since that time, additional data on axial dispersion in open bubble columns have been Ll obtained in 1-, 1.5-, 2-, 3-, and 6-in.-diam columns using widely varying gas flow rates and aqueous solutions having a range of physical properties. One method for correlating the data yields the following relations: for low gas flow rates (bubble flow), N n Npe = 18.0 Ny Npr Mgy ’ (30) and at high gas flow rates (slug flow), 0.4 0.11 -0.38 NPe = 0.46 NRe NAr NSu s (31) where N, =2 = peclet numb pe = p = Peclet number, _ pdV _ NRe . Reynolds number, d3 2 N, =%L2 8 - Archimedes number , Ar 2 Y . NSu = Q%E_z Suratman number, u d = column diameter, V = superficial gas velocity, p = density of liquid, p = viscosity of liquid, o = surface tension of liquid, g = acceleration of gravity, n = number of gas inlets in disperser. The transition from bubble to slug flow occurs at the point represented by the following relation: 1.14 -0.635 n0.099 _ 4 . Ne_ = 4.81 x 107" N, _ Ng. . (32) These relations for the axial dispersion coefficient differ somewhat from those developed most recently (see Sect. 9); however, the axial dispersion coefficient values predicted by the two sets of relations are in good b5 agreement. It is not believed that the use of Egs. (30)-(32) introduces significant error in the calculated performance data for continuous flu- orinators given later in this section. The reaction rate constant, k, was reevaluated from data from a 1- in.-diam continuous fluorinator operated at 525°C with an inlet uranium concentration of 0.35 mole %23 by using the mathematical model represented by Egqs. (25)-(27) and the data on axial dispersion represented by Egs. (30)-(32). The results, summarized in Table 4, show no trend with salt or fluorine flow rate; that is, the values for k are seen to be reasonably constant. Table 4., Summary of Data for Evaluation of Fluorination Reaction Rate Constant from Data Obtained at 525°C in a l-in.~diam Continuous Fluorinator F . . 2 Salt Super?lClal c(L) Flow Rate D k Velocity C T3 2 _1 (cm/sec) feed (cm™/sec) (cm™/sec) (sec . ) 0.0625 0.0257 6.8 17.6 0.00805 0.0445 0.0096 5.0 14.6 0.01033 0.0225 0.00457 3.82 10.6 0.00886 Avg 0.00908 The performance of large open-column, continuous fluorinators (6, 8, 10, and 12 in. in diameter) was estimated from Eq. (28) using the pre- viously discussed estimate of k and the correlations for predicting the axial dispersion coefficient, D. The required fluorinator heights are shown in Figs. 17-20 for fractional uranium removal values of 0.9, 0.95, 0.99, and 0.999. The uranium concentration in the inlet salt was assumed to be 0.0033 mole fraction in each case, and the fluorine flow rate was assumed to be 150% of the stoichiometric requirement. These results are encouraging since they suggest that single fluorination vessels of moderate size will suffice for removing uranium from MSBR fuel salt prior to the ORNL OWG 72-13526 100 1 1T T TTTTT T T T T IT1T1T1 T T T T T 1T 70 Uronium Removael Efficiency 90% = sot- Inlet Uranium Concentration 0.0033 mole froction ~ - Fluorine Feed Rote 150 % of stoichiometric = wii — 10.0}p- 3 7-0"" — > S0 — * — < Fluorinator =4 - . - ° Diometer z - 6 in, — Y o e 10} - - - — 3 o ] LW osp ~ o.sh— L o.1 ] L1ttt | I I 1131t 1 L1111 i0 30 850 100 300 500 1000 3000 5000 10,000 Salt Flow Rate (ft’/day) Fig. 17. Variation of Calculated Fluorinator Height with Salt Flow Rate and Fluorinator Diameter for a Uranium Removal Efficiency of 90%. FLUORINATOR HEIGHT {ft) by URNL DWG 71-10788 IO_ { | IIIIII[ I T lTlIll] | 1 F v 1171 : Uranium Remcval Efficiency 95 % : - Inlet Uranium Concentration 0.0033 mole fraction — - Fluorine Feed Rate 150 % of Stoichiometric — - mamy 10— ] : Fluorinator - | Diameter _ I i Lt o1yl 1 L1 1 111 10 100 1000 KPOC SALT FLOW RATE(#7Yday) Fig. 18. Variation of Calculated Fluorinator Height with Salt Flow Rate and Fluorinator Diameter for a Uranium Removal Efficiency of 95%. FLUORINATOR HEIGHT (ft) ORNL DWG 7i-10793 100, T T T T T T 1 7T [ | T T T T T lTl T T T 1 11 Uranium Removal Efficiency 99% ; 70 Inlet Uranium Concentration 0.0033 mole fraction 7 — Fluorine Feed Rote 150% of stoichiometric N 50+ T 30 - FLUORI IO_— _j F - ' _ | . 5 ] 3+ O ] ' 1 i L i L .1 11 l 1 1 i 1 1411 I —1_ 1 1 r i 1 4.1 10 100 1000 10000 SALT FLOW RATE (ft3/day) Fig. 19. Variation of Calculated Fluorinator Height with Salt Flow Rate and Fluorinator Diameter for a Uranium Removal Efficiency of 99%. 81 FLUORINATOR HEIGHT (ft) ORNL DWG 72-13509R]| 00 T T T T T T T T T T T T T T T T T T 1] 80| ' — 60 — — 50+ _ FLUORINATOR 30 DIAMETER ] k ng. 8 in. 10in. 20 12 in, - 10 — sl ] - — 6 — 8 URANIUM REMOVAL EFFICIENCY - 999% — INLET URANIUM CONCENTRATION - 0.0033 mole fraction — FLUORINE FEED RATE -150 % OF STOICHIOMETRIC “ 3 ] 2+ _ I 1 1 Lo | i Lo o1l 1 1 I L1 11t 10 100 1000 10000 Fig. Rate and SALT FLOW RATE (ft3day) 20. Variation of Calculated Fluorinator Height with Salt Flow Fluorinator Diameter for a Uranium Removal Efficiency of 99.9%. t 50 isolation of protactinium by reductive extraction. The reference flow- sheet for isolating protactinium by fluorination--reductive extraction2 requires fluorination of fuel salt at the rate of 170 ft3/day, which is equivalent to a 10-day processing cycle. A 6-in.-diam fluorination having a height of 10.2 ft will be required for a uranium removal efficiency of 95%Z; an 8-in.-diam fluorinator having a height of 17.8 ft will be required for a uranium removal efficiency of 99Z%. Fluorinators having a high uranium removal efficiency are required in the production of high-purity 233U because incomplete removal of uranium from a salt stream containing 233Pa would result in contamination of the 233U with other uranium isotopes. Therefore, fluorination of salt streams having flow rates of 550 to 1700 ft3/day with uranium removal efficiencies as high as 99.9% may be required. As shown in Fig. 20, a column diameter of 10 in. and heights of 42.5 to 60 ft would be required if a single, open-column, continuous fluorinator were used. In this case, the fluorinator would be divided into several open—-column fluorinators operating in series. If two columns were used, the required heights of each column would be less than half the height required for a single column since there would be no axial dispersion across the fluorinator inlets and outlets. The required uranium removal efficiency for each column would be 96.8%; and, as shown in Fig. 21, column heights of 17 to 28.3 ft would be required for a 10-in.-diam fluorinator. The use of three columns, each with a 90Z uranium removal efficiency, would reduce the total column height even further. Column heights of 7.8 to 17.2 ft would be required for a 1l0-in.-diam fluorinator in this case. FLUORINATOR HEIGHT 100 80 60 50 t+ 40 30 20 w S0 @ ~N ORNL DWG 72-13508 T T ¥ 1 — FLUORINATOR DIAMETER " 6in. 8 in. " 10in. - 12 in. ‘\ vl 1 I U rrrvy 1 L L L f T T v 0 URANIUM REMOVAL EFFICIENCY-96.8% INLET URANIUM CONCENTRATION-0.0033 MOLE FRACTION FLUORINE FEED RATE-150% OF STOICHIOMETRIC L i1l 1 { 1t 11 el 1 L1 11 - 1 1 L 1111 1 10 Fig. 21. Variation of Calculated Fluorinator Height with Salt Flow Rate and Fluorinator D 100 1000 SALT FLOW RATE ( ft3/day) {ameter for a Uranium Removal Efficiency of 96.8%. 16 52 9. MEASUREMENT OF AXIAL DISPERSION COEFFICIENTS AND GAS HOLDUP IN OPEN BUBBLE COLUMNS J. §. Watson L. E. McNeese Axial dispersion is important in the design and performance of continuous fluorinators to be used in processing MSBR fuel salt. Since molten salt saturated with fluorine is corrosive, the fluorinators will be simple, open vessels having a protective layer of frozen salt on all exposed metal surfaces. In such systems the rising gas bubbles may cause appreciable axial dispersion throughout the salt. For the past few years, we have been involved in a program for measuring axial dis- persion resulting from the flow of air through liquids in open bubble columns. The objectives of this program are to evaluate the effect of axial dispersion on fluorinator performance and to account for this effect in the design of fluorinators. 9.1 Previous Studies on Axial Dispersion Initial studies on axial dispersion in open bubble columns were carried out by Bautista and McNeese,25 who studied axial dispersion during the countercurrent flow of air and water in a 2-in.-ID, 72-in.- long column. Two regions of operation were observed. The first of these consisted of a "bubble flow" region at low gas flow rates in which the air moved up the column as individual bubbles and coalescence was minimal. The second consisted of a "slug flow" region at higher gas flow rates in which the air coalesced rapidly into bubbles having diameters equal to the column diameter. A plot of the logarithm of the dispersion coefficient vs the logarithm of the gas flow rate was linear in both regions. However, the slope of the line representing data in the slug flow region was higher than that for data in the bubble flow region. The transition between the two regions was well defined. The same column and equipment were used by A. M. Sheikh and J. D. 26 Dearth, of the MIT Practice School, for investigating the effects of the viscosity and surface tension of the liquid. The dispersion coef- ficient was found to decrease in the bubble flow region as the viscosity 53 of the liquid was increased from 1 cP to 15 cP by the addition of glycerol to the water; little effect was noted in the slug flow region. An increase in the dispersion coefficient was observed as the surface tension of the liquid was decreased by the addition of n-butanol to the water. 21 of The equipment was also used by A. A. Jeje and C. R. Bozzuto, the MIT Practice School, who investigated the effects of gas inlet diameter and column diameter on axial dispersion and obtained data on gas holdup in bubble columns. In the slug flow region, the dispersion coefficient appeared to be proportional to the square root of the volumetric gas flow rate, but was independent of column diameter. 1In the bubble flow region, the dispersion coefficient was dependent only on the volumetric gas flow rate in the case of columns having diameters of 2 in. or larger. Dispersion coefficient data obtained with a 1.5- in.~diam column deviated from this condition. At low gas flow rates, the gas holdup was linearly dependent on the superficial gas velocity but independent of column diameter. At superficial velocities above the transition from bubble to slug flow, the gas holdup data for the various column diameters diverged; the holdup was greatest for the smallest column diameter. All of the dispersion coefficient data obtained by the above investigators resulted from measurements of the steady-state axial distribution of a cupric nitrate tracer that was continuously injected into the bottom of the column near the water exit. The studies indi- cated that the axial dispersion coefficient is independent of both axial position in the column and water superficial velocity in the range of interest. The steady-state experimental technique had two principal disadvantages: (1) the measurements were time-consuming since about 2 hr was required for the column to attain steady state; and (2) at high gas flow rates, air was entrained with water withdrawn from the column for determination of the tracer concentration, and the resulting error in the dispersion coefficient data was unacceptably high. 1In order to circumvent these problems, Bautista28 developed a transient technique for obtaining data on axial dispersion. In this 5L technique there was no net flow of water through the column; however, data obtained with the steady-state technique indicated that the water flow rate did not affect the axial dispersion coefficient at the water flow rates of interest. A small amount of electrolyte tracer (KCl) was quickly injected into the top of the column and the concentration of the tracer was measured continuously at a point near the bottom of the column by use of a conductivity probe. The resulting data were in agreement with earlier data obtained with the steady-state technique; on the other hand, data obtained with the transient technique showed minimal scatter even at high gas flow rates and could be obtained in less than 10% of the time required for the steady-state technique. During this report period, additional studies were carried out using the transient technique in order to determine the effects of changes in column diameter, gas inlet design, and physical properties of the liquid phase on axial dispersion and gas holdup. 9.2 Equipment and Experimental Procedure The equipment and experimental procedure used in the present studies have been described previously-28 The equipment consisted of an open bubble column, a means for injecting KCl tracer solution at the top of the column, a conductivity probe located at an intermediate axial point along the column for determining the KCl concentration in the aqueous solution at the point, an electronics system and a recorder for recording the output from the conductivity probe, an air supply and metering sys- tem for feeding air at a known flow rate to a gas disperser located in the base of the column, and a manometer for obtaining data on gas holdup in the column. Eight-foot-long Plexiglas columns with inside diameters of 1.0, 1.5, 2, 3, and 6 in. were used. Gas distributor plates having different numbers and sizes of orifice openings were installed at the bottom of the column in order to determine the effect of gas inlet design on axial dispersion and gas holdup. The air flow rate was measured at the top of the column. A soap bubble buret was employed for flow rates below 15 cm5/sec; a wet-test meter was used for higher flow rates. The solutions used in the study consisted of demineralized water or mixtures of demineralized water and glycerin or 55 n-butanol. The aqueous solutions were prepared in a tank and pumped to the column. Demineralized water could be introduced at the top of the column to facilitate cleaning of the column between runs. After the column had been filled with liquid having the desired physical properties, a sufficient volume of 2.k N KCl tracer solution (5 to 15 cmj, depending on column size) was added to the liquid in the column in order to obtain a recordable reading from the conductivity probe. The air flow rate was then adjusted to the desired value, and a second volume of tracer was quickly injected at the top of the column. Subsequently, the response of the conductivity probe was recorded until the tracer was uniformly dispersed throughout the column. The height of the gas-liquid mixture in the column and the height of the liquid with no gas flow were measured. Samples of the liquid were then taken for surface tension and viscosity measurements. The viscosity of the liquid was determined with a Ubbelohde viscometer, while surface tension measurements were made using the capillary rise method. Visual observa- tions of the gas and liquid in the column were made during the course of the experiments. 9.3 Experimental Data on Axial Dispersion Experimental data on axial dispersion in open bubble columns were obtained during this report period in a series of four separate studies. The first study, made by a group of students at the University of Tennessee, was carried out with a l-in.-ID, 8-ft-long column. The gas disperser at the bottom of the column consisted of a single inlet having an inside diameter of 4.3 mm. The studies were carried out with demin- eralized water, and the superficial gas velocity was varied from 0.156 to 76.6 cm/sec. The data obtained during this study are summarized in Table 5. The axial dispersion coefficient values obtained in the l-in.- diam column fall below the values obtained previously at the same super- ficial gas velocity in columns having diameters of 1.5 and 2 in. The second study was carried out by J. C. Bronfenbrenner, L. J. Marquez, and J. F. Mayer, of the MIT Practice School, who determined the effects on axial dispersion caused by changes in column diameter, 56 Table 5. Summary of Data on Axial Dispersion Obtained in a 1.0-in.-ID Open Bubble Column Containing Demineralized Water at 25°C 3 Tracer injection volume: ~ 5 cm Relative probe positiona: 0.825 Gas inlet: one orifice, 4.3 mm in diameter Gas Flow Superficial Dispersion Run Rateb Gas Velocity Coefficient No (em/sec ) (cm/sec) (em®/sec) 1 ok.6 18.7 82.3 2 387.9 T6.6 319.0 3 387.9 76.6 261.8 L 231.1 65 .k 255.2 5 378.4 Th. T 182.3 6 283.8 56.0 319.0 T 189.2 37.3 170.2 8 ok.6 18.7 63.8 9 141.9 28.0 170.2 10 236.5 46.1 232.0 11 52.0 10.3 52.0 12 QL. 6 18.7 92.8 13 146.6 28.9 128.3 14 118.2 23.% 102.6 15 70.9 14.0 61.1 16 165.6 22.7 146.6 17 189.2 37.3 213%.8 18 236.5 Lo.7 185.5 19 321.7 63.5 285.1 20 227.0 4.8 181.1 21 5.91 1.166 20.8 22 5.51 1.088 18.3 23 5.12 1.011 17.1 24 4. 73 0.933 22.5 25 4.33 0.855 17.1 26 3.94 0.778 15.5 27 3.55 0.700 14.1 28 0.79 0.156 5.84 29 6.33 1.25 14.8 30 5.17 1.02 2.6 51 3-95 0.78 19.7 32 2.37 0.468 1L.8 3% 0.79 0.156 7.04 3h 6.7k 1.33 22.h 35 5.12 1.01 17.8 36 3.56 0.702 1k.5 37 1.98 0.390 10.7 38 0.79 0.156 6.42 a . . Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. Measured under conditions at top of column. 5T viscosity of the liquid phase, and superficial gas velocity. Thle columns used in this study consisted of 1.5-, 2-, and 6-in.-ID Lucite tubes, each having a length of 8 ft. A conductivity cell was inserted in the columns at a height of 70.5 cm from the bottom of the column. The liquid in the column consisted of mixtures of distilled water and glycerin in which the glycerir concentrations were O, 25, and 65 wt %. The physical properties of these solutions are summarized in Table 6. Air entered the column through a single orifice at the bottom of the column; the orifice ID was 0.04 in. for the two smaller columns and 0.4 in. for the 6-in.-diam column. The superficial gas velocity was varied from 0.26 to 40 cm/sec in the smaller columns and 0.27 to 9.5 cm/sec in the 6-in.-diam column. Data on axial dispersion obtained during the second study are summarized in Tables 7-9. The results obtained with water in a 1.5-in.-diam column are in good agreement with those obtained previously;28 the axial disper- sion coefficient shows little change as the viscosity of the liquid is increased from 0.9 cP to 1.8 cP. Similarly, there is little difference in the axial dispersion coefficient values obtained in a 2-in.-~diam column with a liquid having a viscosity of 0.9 cP and those obtained with a liquid having a viscosity of 1.8 cP. Dispersion coefficient values obtained with a liquid having a viscosity of 12.1 c¢P are about 50 to 70% of those obtained with liquids having viscosities of 0.9 and 1.8 ¢cP. Essentially no difference was observed in the axial disper- sion coefficient values obtained in a 6-in.-diam column for liquids having viscosities of 0.9, 1.8, and 12.1 cP. Table 6. Physical Properties of Water-Glycerin Solutions Used During Second Study of Axial Dispersion and Gas Holdup in Bubble Columns Glycerin Surface Concentration Viscosity Density Tension (wt %) (cP) (g/cm>) (dynes/cm) O 0.89 0.997 T3 25 1.8 1.05 T2 65 12.1 1.16 67.9 58 Table 7. Summary of Data on Axial Dispersion Obtained in a 1.5-in.-ID Column During Second Study Relative probe position®: 0.69 Gas inlet: one orifice, 1.0 mm in diameter Gas Flow Superficial Glycerin Axial Dispersion Run Rate Gas Velocity Conc. Coefficient No. (em>/sec) (cm/sec ) (wt %) (cm®/sec ) 1 187.0 16.4 0 138.5 2 97.8 8.58 0 85.6 3 140.2 12.3 0 114.8 L 112.8 9.89 0 93.L4 5 19.6 1.72 0 31.6 6 4L0%.6 5.4 0 4h82.2 7 7.87 0.69 0 20.8 8 403.6 35,4 0 380.5 9 3.1 2.99 0 39.0 10 58.2 5.11 0 65.0 11 10.4 0.912 0 24.8 12 255.4 22.4 0 177.0 13 118.6 10.4 0 149.0 1k 191.5 16.8 0 136.0 15 155.0 13.6 0 95.2 16 118.6 10.4 25 88.9 17 18.9 1.66 25 2.9 18 372.8 32.7 25 813.9 19 566.6 Lo.7 25 655.9 20 372.8 32.7 25 155.1 21 46.3 4.06 25 Li.6 22 118.6 10.4 25 88.9 2% 2L9. 7 21.9 25 115.3% 2L 149.4 13.1 25 131.4 25 118.6 10.4 25 56.2 26 75.5 6.62 25 €2.2 a Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. b Measured under conditions at top of column. 29 Table 8. Summary of Data on Axial Dispersion Obtained in a 2.0-in.-ID Column During Second Study a Relative probe position”: 0.69 Gas inlet: one orifice, 1.0 mm in diameter Gas Flow Superficial Glycerin Axial Dispersion Run Rate Gas Velocity Conc. Coefficient No. (em?/sec ) (cm/sec ) (wt %) (cm®/sec ) 1 81.1 4.0 0 51.0 2 h11.4 20.3 0 206.0 3 482.4 23.8 0 172.7 L 291.9 4.k 0 144.9 RS 11.4 0.56 0 22.3 6 78.6 3.88 0 58.7 7 120.8 5.96 0 67.5 8 117.2 5.78 25 72.2 9 1rr.3 8.75 25 97.8 10 210.8 10.4 25 119.7 11 377.0 18.6 25 178.3 12 553.3 27.3 25 L27.0 13 56.5 2.79 25 45.3 14 81.3 L.01 25 52.1 15 10.7 0.53 25 26.5 16 : 8.31 0.41 25 26.5 17 15.2 0.75 25 31.2 18 2L .5 1.21 25 34.5 19 L. 8 2.21 25 48.2 20 18.4 0.91 65 27.6 21 9.93 0.49 65 21.h 22 6.28 0.31 65 18.9 23 25.1 1.24 65 27.8 24 L7.6 2.35 65 zh.,1 25 60.8 3.0 65 27.2 26 83.3 4.11 65 L. L 27 116.9 5.77 65 51.0 28 184.4 9.1 65 75-3 29 208.8 10.3 65 79.2 30 326.3% 16.1 65 105.3 31 504.7 29 65 1hl.2 32 758.0 37.4 65 213.6 #Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. bMeasured under conditions at top of column. 60 Table 9. Summary of Data on Axial Dispersion Obtained in a 6.0-in.-ID Column During Second Study Relative probe position?: 0.69 Gas inlet: one orifice, 10 mm in diameter Gas Flow Superficial Glycerin Axial Dispersion Run Rate Gas Velocity Conc. Coefficient No. (cmj/sec) (cm/sec) (wt %) (cm?/sec ) 1 1530 8.39 0 207.0 2 561.8 2.08 0 178.8 3 145.9 0.8 0 152.1 L 9k.8 0.52 0 156.6 5 ok .8 0.52 0 160.7 6 271.8 1.49 0 183.2 T 394.0 2.16 0 161.3 8 1299 T7.12 0 370.8 9 113.1 0.62 0 173.4 10 698.6 3.83 0 156.9 11 195.2 1.07 0 123.1 12 60.2 0.33% 0 146.0 13 286.4 1.57 0 146.2 14 1665 9.1% 0 280.6 15 899.3 4.93 0 245.8 16 479.8 2.63 0 141.0 17 1372 7.52 25 231.5 18 1757 9.63 25 265.9 19 L1757 9.63 25 303.0 20 hh.3 2.6 25 169.5 21 923.0 5.06 25 262.0 22 1572 8.62 25 229.3 23 58.4 0.32 25 124.7 24 372.1 2.0h 25 187.7 25 217.1 1.19 25 180.5 26 5436 2.98 25 184.0 27 707.8 3.88 25 210.8 28 5h3.6 2.98 25 181.1 29 1094 6.0 25 215.9 30 220.7 1.21 65 152.8 31 295.5 1.62 65 148.6 32 Lg.2 0.27 65 161.8 35 T727.8 5-99 65 181.1 3k 521.7 2.86 65 179.6 35 361.2 1.98 65 145.6 36 1096 6.01 65 215.6 61 Table 9. (continued) Relative probe position®: 0.69 Gas inlet: one orifice, 10 mm in diameter Gas Flow Superficial Glycerin Axial Dispersion Run Rate Gas Velocity Conc. Coefficient No. (cm?/sec ) (cm/sec) (wt %) (cm?/sec ) 37 1532 8.h 65 o5k .2 38 93.0 0.51 65 157.2 39 136.8 0.75 65 172.6 a Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. b Measured under conditions at top of column. The third study carried out during this period was made by A. K. Padia, G. T. Marion, and R. H. McCue, of the MIT Practice School, who studied the effects of changes in the number and size of gas inlet orifices, column diameter, superficial air velocity, and viscosity and surface tension of the liquid phase on the axial dispersion coef- ficient and gas holdup. The ranges of the independent parameters that were varied in this study are summarized in Table 10. Column diameters of 1.5, 2, and 3 in. were used with both single and multiple orifices ranging in size from O.4 to 6.4 mm. The viscosity of the liquid phase was varied from 0.9 to 11.% c¢cP, and the surface tension of the liquid was varied from 27 to 7O dynes/cm. The superficial gas velocity was varied from 0.0318 to 20 cm/sec in 12 to 17 increments for each value of colurn diameter, gas distributor design, and property of the liquid phase. Data obtained during the third study are summarized in Tables 62 Table 10. Ranges of Parameters During Third Study of Axial Dispersion in Open Bubble Columns Parameter Values Used Column Diameter, in. 1.5, 2, 3 Number of Orifices in Gas Inlet 1, 5, 19, 37 Gas Inlet Orifice Diameter, mm 0.4, 1, 2, 4, 6.4 Surface Tension of Liquid, dynes/cm 27, 45, TO Viscosity of Liquid, cP 0.9, 2.05, 10.7, 11.3 Superficial Gas Velocity, cm/sec 0.0318 to 20 11-19. The variation of the axial dispersion coefficient with changes in the superficial gas velocity and orifice diameter for a 2-in.-ID column for which the gas distributor consisted of five orifices is in general agreement with that obtained previously for a 2-in.-diam column operated with a single gas inlet. The variation in the dispersion coefficient in the slug flow region with changes in the diameter of the gas inlet orifices is not believed to be significant. However, the differences observed in the bubble flow region are probably meaningful. Data of the same type, obtained with a gas disperser consisting of 37 orifices, show even less deviation from previous values obtained with a single gas inlet. For a column diameter of 2 in. and gas distributor orifice diameters of 1 mm, only slight differences in the dispersion coefficient are noted in the slug flow region; however, in the bubble flow region, a progressive increase in axial dispersion coefficient is observed as the number of orifices is increased. The variation of the axial dispersion coefficient with changes in the viscosity of the liquid phase and the superficial gas velocity for column diameters of 1.5 and 63 Table 11. Summary of Data on Axial Dispersion Obtained in a 1.5-in.-ID Column Containing Water During Third Study 5 s s - ‘ Tracer injection volume: /~ 5 cm , P - Relative probe position : 0.79 Gas inlet: one orifice, 1 mm ID Gas Flow Superficial Axial Dispersion Run Ratel Gas Velocity Coefficient No. (cm5/sec) (cm/sec ) (cm®/sec ) 1 11.9 1.0k45 25.8 2 2.42 0.212 12.9 3 5.14 0.451 18.9 L 8.20 0.719 2L ) 5 15.2 1.33 32.9 6 19.2 1.68 31.6 7 25.2 2.21 39.5 8 Ly, 0 3.86 50.0 9 Th.6 6.54 69.4 10 101.8 8.93 85.9 11 2Lh2.8 2l.3 242.0 ®Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. b ‘o Measured under conditions at top of column. 6L Table 12. Summary of Data on Axial Dispersion in a l1.5-in.~ID Column Containing Aqueous Isopropanol During Third Study 3 Tracer injection volume: ~ 5 cm Relative probe position®: 0.79 Surface tension of liquid: L45.3% dynes/cm Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Axial Dispersion Run Rate Gas Velocity Coefficient No. (em’/sec) (cm/sec ) (cm®/sec ) 1 11.9 1.045 20.9 2 8.09 0.710 23%.3% 3 6.10 0.535 20.5 L 0.87 0.252 17.3 5 1.06 0.0930 14.6 6 16.6 1.456 23.1 7 17.7 1.55 26.7 8 29.h 2.58 33.7 9 49.1 L.31 Lo.6 10 66.0 5.79 50.0 11 BL4.9 T .45 58.0 %Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. Measured under conditions at top of column. 65 Table 15. Summary of Data on Axial Dispersion in a 1.5-in.-ID Column Containing Aqueous Isobutanol During Third Study 3 Tracer injection volume: -~ 5 cm Relative probe positiona: 0.79 Surface tension of liquid: 27.3 dynes/cm Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Axial Dispersion Run RateP Gas Velocity Coefficient No. (cm’/sec ) (cm/sec ) (cm?/sec ) 1 11.6 1.02 21.9 2 1.52 0.133 15.0 3 0.804 0.0705 11.9 L 2.4k 0.214 15.8 5 5.63 0.LokL 20.6 6 8.70 0.763% 22.2 7 15.6 1.37 22.3 8 23.6 2.07 27.9 9 40.6 3.56 36.2 10 73.0 6.4 51.5 11 98.0 8.6 65.7 fRatio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. bMeasured under conditions at top of column. 66 Table 14. Summary of Data on Axial Dispersion Obtained in a 1.5-in.-ID Column Containing a Water-Glycerin Mixture During Third Study 3 Tracer injection volume: ~ 5 cm s .4 Relative probe position : 0.79 Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Glycerin | Axial Dispersion Run Rate Gas Velocity Conc. Coefficient No. (em?/sec ) (cm/sec ) (wt %) (cm?/sec) 1 0.727 0.0638 25 T.65 2 2.13 0.187 25 12.5 3 3.80 0.333% 25 24.6 4 6.02 0.528 25 26.4 5 10.2 0.895 25 28.4 6 h.2 1.25 25 36.4 T 18.5 1.625 25 36.4 8 29.8 2.61 25 k3.5 9 65.9 5.78 25 60.6 10 92.0 8.07 25 81.0 11 2.4k 0.21%4 65 10.6 12 1.01 0.0888 65 7.45 13 3.57 0.313 65 12.3 14 5.13 0.45 65 15.2 15 T.42 0.651 65 21.3 16 11.4 1.0 65 21.8 17 15.7 1.38 65 27.% 18 19.8 L.74 65 27.7 19 33.1 2.9 65 3.7 20 66.0 5.79 65 55.6 #Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-~liquid mixture. b Measured under conditions at top of column. 67 Table 15. Summary of Data on Axial Dispersion Obtained in a 2.0-in.-ID Column Containing Water During Third Study 5 Tracer injection volume: ~ 5 cm a Relative probe position : 0.79 Gas inlet: one orifice, 1 mm ID Gas Flow Superficial Axial Dispersion Run RateP Gas Velocity Coefficient No. (cmd/sec) (cm/sec ) (cm?/sec ) 1 0.896 0.04kL2 19.6 2 0.255 0.0126 28.1 3 L. 2k 0.209 28.9 L 5.98 0.295 31.1 5 8.21 0.405 37.6 6 10.4 0.515 40.5 T 12.5 0.618 L3.4 8 16.2 0.798 45.5 9 19.2 0.95 L8.2 10 230.8 1.52 52.7 11 59.2 2.92 60.8 12 92.8 L.58 T2.6 13 145.9 T.2 81.5 4Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. bMeasured under conditions at top of column. Table lt. Summary of Data on Axial Dispersion Obtained in Number of orifices: Tracer injection volume: Relative probe position?: a -in.-ID Column Containing Water During Third Study ~ L cm-’ Runs 1-12 = 0.76T Runs 135-68 — 0.730 Gas Flow Superficial Gas Inlet Axial Dispersion Gas Flow Superficial Gas Inlet Axial Dispersion Run Rate Gas Velocity Orifice ID Coefficient Run Rateb Gas Velocity Orifice 1D Coefficient No. (cm”/sec) (cm/sec) (mm ) (cm/sec ) No. (cm’/sec) (cm/sec ) (mm ) (cm Table 20. Summary of Data on Axial Dispersion Obtained in a 3-in.-ID Column Having a 1l-in.-ID Side Inlet Joined to the Column at an Angle of 45° with Respect to the Column Axis 5 Tracer injection volume: ~ 5 cm -Relative probe positiona: 0.79 Gas Flow Superficial Dispersion Run RateP Gas Velocity Coefficient No. (em?/sec ) (cm/sec) (cm?/sec ) 1 852.8 18.7 111.9 2 706.8 15.5 159.2 3 579.2 12.7 137.2 L L83.4 10.6 132.6 5 367.6 8.06 114.9 6 277 T 6.09 91.1 T 118.8 b1k 67.3 8 98.0 2.15 4.3 9 192.0 .21 T0.7 10 162.3 3.56 Th .k 11 141.8 3,11 65.0 12 124.0 2.72 68. 4 13 oL.4 2.07 53T 14 89.4 1.96 58.0 15 72.0 1.58 49.0 16 67.9 1.49 59.1 17 4O.4 0.887 56.8 18 20.0 0.438 h7.3 19 26.2 0.57h 67.8 20 18.0 0.395 37.6 21 13.5 0.296 39.9 22 9.62 0.211 L7.1 03 5.88 0.129 31.8 ol 2.79 0.0612 28.1 25 97.1 2.13 60.7 26 141.8 3.11 65.% 27 9l.2 2.00 54.9 28 21.9 0.481 42.0 29 218.9 4.80 T1.8 30 5.65 0.124 29.5 %Ratio of distance of probe from surface of gas-liquid mixture to total height of gas-liquid mixture. bMeasured under conditions at top of column. Th Table 21. Ranges of Parameters During Studies of Gas Holdup in Open Bubble Columns Parameter Values Used Column Diameter, in. 1.0, 1.5, 2, 3, 6 Number of Orifices in Gas Inlet 1, 5, 19, 37 Gas Inlet Orifice Diameter, mm 0.4, 1, 2, 4, 6.4 Surface Tension of Liquid, 27, 45, 70, 73 dynes/cm Viscosity of Liquid, cP 0.9, 1.8, 2.05, 10.7, 11.3, 12.1 Gas Superficial Velocity, cm/sec 0.013 to 49.7 9.5 (Correlation of Data on Gas Holdup The data on gas holdup obtained in the second and third studies of this report period and in studies by Bautista28 were correlated by the method of least squares. These data, a total of 349 holdup determina- tions, cover column diameters from 1.5 to 6 in. and include a range of values for the superficial gas velocity and physical properties of the liquid. The results obtained for the l-in.-diam column in the first study were not used in developing the correlation since these values are believed to be of a lower quality than the remaining ones. It was found that gas holdup could be represented by the relation Vv h - 8 — (33) 1.468 Vg + 0.407IY gd where h = fraction of column volume occupied by gas, V_ = superficial gas velocity, cm/sec, g = acceleration of gravity, cm/secg, d = column diameter, cm. 5 Table 22. Summary of Data on Gas Holdup Obtained in a 1.0-in.-ID Open Bubble Column Containing Demineralized Water at 25°C Gas inlet: one orifice, 4.3 mm in diameter Gas Flow Superficial Bubble Rise Run Rate? Gas Velocity Gas Velocity No. (em?/sec ) (cm/sec ) Holdup (cm/sec ) 1 gl . 6 18.7 0.492 38.0 2 387.9 76.6 0.719 106.5 3 387.9 76.6 0.719 106.5 L 331.1 65 .4 0.707 92.5 5 378.4 Th. 7 0.719 103.9 6 28%.8 56.0 0.713 78.5 T 189.2 27.3 0.641 58.2 8 9.6 18.7 0.513 36.5 9 141.9 28.0 0.6 he. T 10 236.5 6.1 0.686 67.2 11 52.0 10.3 0.309 33,3 12 9Lk.6 18.7 0,450 h1.6 13 146.6 28.9 0.529 54.6 1k 118.2 23.3% 0.492 WL 15 70.9 1L.0 0.372 37,6 16 165.6 32.7 0.568 57.6 L7 189.2 37.3 0.575 64.9 18 236.5 L6.7 0.587 79.6 19 321.7 63.5 0.642 98.9 20 227.0 Ly .8 0.584 6.7 21 5.91 1.166 0.0658 7.7 22 5.51 1.088 0.0598 8.2 23 5.12 1.011 0.0523 19.3 2k .73 0.933 0.0516 18.1 o5 .23 0.855 0.0486 17.6 26 3.9k 0.778 0.0412 18.9 27 3.55 0.700 0.0387 18.1 28 0.79 0.156 0.0098 15.9 29 6.33 1.25 0.0645 19.4 30 5.17 1.02 0.0387 26.4 31 3-95 0.78 0.0451 17.% 32 2.37 0.468 0.0254 18.4 33 0.79 0.156 0.00645 2h.2 3L 6.7h 1.33 0.068 19.6 35 5.12 1.01 0.056 18.0 36 3.56 0.702 0.0L5 15.6 37 1.98 0.390 0.023 17.0 38 0.79 0.156 0.015 10.L ®Measured under conditions at top of column. T6 Table 23. Summary of Data on Gas Holdup Obtained in a 1.5-in.-ID Column During Second Study Gas Flow Superficial Glycerin Bubble Run Rated Gas Velocity Concentration Gas Rise Velocity No. (em3/sec) (cm/sec) (wt %) Holdup (cm/sec) 1 187.0 16.4 0 0.328 50.0 2 97.8 8.58 0 0.211 40.7 3 140,2 12.3 0 0.272 45.2 4 112.8 9.89 0 0.233 42.4 5 19.6 1.72 0 0.052 33.1 6 403.6 35.4 0 0.516 68.6 7 7.87 0.69 0 0.022 31.4 8 403.6 35.4 0 0.523 67.7 9 34,1 2.99 0 0.09 33.2 10 58.2 5.11 0 0.14 36.5 11 10.4 0.912 0 0.03 30.4 12 255.4 22.4 0 0.394 56.9 13 118.6 10.4 0 0.242 43.0 14 191.5 16.8 0 0.333 50.5 15 155.0 13.6 0 0.291 46.7 16 118.6 10.4 25 0.307 33.9 17 18.9 1.66 25 0.056 29.6 18 372.8 32.7 25 0.598 54.7 19 566.6 49.7 25 0.675 73.6 20 372.8 32.7 25 0.598 54.7 21 46.3 4.06 25 0.149 27.2 22 118.6 10.4 25 0.307 33.9 23 249.7 21.9 25 0.514 42.6 24 149.4 13.1 25 0.281 46.6 25 118.6 10.4 25 0.262 39.7 26 75.5 6.62 25 0.161 41.1 *Measured under conditions at top of column. 1T Table 24. Summary of Data on Gas Holdup Obtained in a 2.0-in.~-ID Column During Second Study Gas Flow Superficial Glycerin Bubble Rise Run Rate? Gas Velocity Conc. Gas Velocity No. (cmB/sec) (cm/sec) (wt %) Holdup (cm/sec) 1 81.1 4.0 0 0.108 37.0 2 h11.4 20.3 0 0.359 56.5 3 482.4 23.8 0 0.378 63.0 L 291.9 4.k 0 0.239 60.3% 5 11.4 0.56 0 0.012 hé.T 6 78.6 3.88 0 0.098 39.6 T 120.8 5.96 0 0.136 43,8 8 117.2 5.78 25 0.159 36.4 9 1773 8.75 25 0.207 ho.3 10 210.8 10.4 25 0.233 L. 6 11 377.0 18.6 25 0.35 5%.1 12 553.3% 27-3 25 0.466 58.6 13 56.5 2.79 25 0.075 37.2 14 81.% 4.01 25 0.106 37.8 15 10.7 0.53% 25 0.009 58.9 16 8.31 0.41 25 0.009 L5.6 17 15.2 0.75 25 0.02 37.5 18 ol .5 1.21 25 0.035 3.6 19 .8 2.21 25 0.063 35.1 20 18.4 0.91 65 0.021 43,3 21 9.93 0.49 65 0.012 40.8 22 6.28 0.31 65 0.009 3L .4 23 25.1 1.24 65 0.035 5.0 2l L7.6 2.35 65 0.054 43,5 25 60.8 3.0 65 0.073 h1.1 26 83.3 4.11 65 0.095 43,3 27 116.9 5.7T 65 0.124 h6.5 28 184 .4 9.1 65 0.18 50.6 29 208.8 10.3 65 0.199 51.8 30 226.3 16.1 65 0.254 63.L 31 50L.7 2.9 65 0.325 76.6 32 758.0 37 L 65 0.k 03.5 8Measured under conditions at top of column. T8 Table 25. Summary of Data on Gas Holdup Obtained in a 6.0-in.-ID Column During Second Study Gas Flow Superficial Glycerin Bubble Run Rated Gas Velocity Concentration Gas Rise Velocity No. (cm3/sec) (cm/sec) (wt 7%) Holdup (em/sec) 1 1530 8.39 0 0.146 57.5 2 561.8 3.08 0 0.071 43.4 3 145.9 0.8 0 0.021 38.1 4 94.8 0.52 0 0.016 32.5 5 94.8 0.52 0 0.016 32.5 6 271.8 1.49 0 0.036 41.4 7 394.0 2.16 0 0.054 40.0 8 1299 7.12 0 0.128 55.6 9 113.1 0.62 0 0.016 38.8 10 698.6 3.83 0 0.081 47 .3 11 195.2 1.07 0 0.028 38.2 12 60.2 0.33 0 0.009 36.7 13 286.4 1.57 0 0.034 46.2 14 1665 9.13 0 0.156 58.5 15 899.3 4.93 0 0.089 55.4 16 479.8 2.63 25 0.056 47.0 17 1372 7.52 25 0.16 47.0 18 1757 9.63 25 0.217 44,4 19 1757 9.63 25 0.217 44,4 20 474.3 2.6 25 0.076 34.2 21 923.0 5.06 25 0.126 40.2 22 1572 8.62 25 0.196 44,0 23 58.4 0.32 25 0.008 40.0 24 372.1 2.04 25 0.06 34.0 25 217.1 1.19 25 0.032 37.2 26 543.6 2.98 25 0.074 40.3 27 707.8 3.88 25 0.09 43.1 28 543.6 2.98 25 0.074 40.3 29 1094 6.0 25 0.118 50.8 30 220.7 1.21 65 0.028 43.2 31 295.5 1.62 65 0.038 42.6 32 727.8 3.99 65 : 0.081 49.3 33 521.7 2.86 65 0.066 43.3 34 361.2 1.98 65 0.052 38.1 35 1096 6.01 65 0.105 57.2 36 1532 8.4 65 0.13 64.6 37 93.0 0.51 65 0.013 39.2 38 136.8 0.75 65 0.017 44,1 #Measured under conditions at top of column. 9 Table 26. Summary of Data on Gas Holdup Obtained in a 1.5-in.-ID Column Containing Water During Third Study Gas inlet: one orifice, 1 mm ID Gas Flow Superficial Bubble Rise Run Rate? Gas Velocity Gas Velocity No. (em?/sec) (cm/sec) Holdup (cm/sec ) 1 2.42 0.212 0.00862 oh.6 2 8.20 0.719 0.0308 23.3 3 15.2 1.33 0.0424 31.4 L 19.2 1.68 0.0531 31.6 5 25.2 2.21 0.0675 32.7 6 WL, o 3.86 0.121 31.9 T Th.6 6.5k 0.183 35.7 8 101.8 8.93 0.250 357 9 2h2.8 21.3 0.427 49.9 a Measured under conditions at top of column. Table 27. Summary of Data on Gas Holdup in a 1.5-in.-ID Column Containing Aqueous Isopropanol During Third Study Surface tension of liquid: L45.3 dynes/cm Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Bubble Rise Run Rate? Gas Velocity Gas Velocity No (cm?/sec ) (em/sec) Holdup fem/sec) 1 11.9 1.045 0.034L 30,4 2 8.09 0.710 0.0247 28.7 3 6.10 0.535 0.019 28.2 4 2.87 0.252 0.00976 25.8 5 1.06 0.0930 0.00423 22.0 6 16.6 1.456 0.048T7 29.9 7 17.7 1.55 0.0567 27.3 8 29.4 2.58 0.089 29.0 9 9.1 h.31 0.149 28.9 10 66.0 5.79 0.18% 21.6 11 84.9 745 0.167 .6 8Measured under conditions at top of column. 80 Table 28. Summary of Data on Gas Holdup in a 1.5-in.-ID Column Containing Aqueous Isobutanol During Third Study Surface tension of liquid: 27.3% dynes/cm Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Bubble Rise Run Rate?® Gas Velocity Gas Velocity No. (emd/sec) (cm/sec) Holdup (cm/sec) 1 11.6 1.02 0.0328 31.1 2 1.52 0.133 0.00463 28.7 3 0.80L 0.0705 0.00278 25.4 b4 2.44 0.214 0.0088 24.3 5 5.6% 0.494 0.017k 28.4 6 8.70 . 0.76% 0.0276 27.6 7 15.6 1.37 0.052 26.3 8 23.6 2.07 0.081 25.6 9 40.6 3.56 0.141 25.2 10 73.0 6.4 0.247 25.9 11 98.0 8.6 0.38 22.6 a Measured under conditions at top of column. Table 29. Summary of Data on Gas Holdup Obtained in a 1.5-in.-ID Column Containing a Water-Glycerin Mixture During Third Study Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Glycerin Bubble Rise Run Rate? Gas Velocity Conc. Gas Velocity No. (em3/sec) (cm/sec) (wt %) Holdup (cm/sec) 1 2.13 0.187 25 0.00793 23.6 2 3.80 0.333% 25 0.0126 26.4 3 6.02 0.528 25 0.0194 27.2 L 10.2 0.895 25 0.0298 30.0 5 4.2 1.25 25 0.0387 20,3 6 18.5 1.625 25 0.052 31.3 7 29.8 2.61 25 0.0732 35,7 g 65.9 5.78 25 0.171 33,8 3 92.0 8.07 25 0.20k 39.6 1o 3.57 0.313 65 0.0101 31.0 11 5.13 0.45 65 0.016 28.1 12 7.4p 0.651 65 0.0216 30.1 13 11.4 1.0 65 0.0335 29.9 1L 15.7 1.%8 65 0.0L435 31.7 15 19.8 1.7k 65 0.0554 31.4 16 33.1 2.9 65 0.0925 31.4 17 66.0 5.79 65 0.227 25.5 a Lo Measured under conditions at top of column. 81 Table 30. Summary of Data on Gas Holdup Obtained in a 2.0-in.-ID Column Containing Water During Third Study Gas inlet: one orifice, 1 mm ID Gas Flow Superficial Bubble Rise Run Rate? Gas Velocity Gas Velocity No. (em?/sec) (em/sec) Holdup (cm/sec) 1 0.896 0.04L2 0.00225 19.6 2 2.55 04126 0.00L5 28.0 3 h.2k 0.209 0.00625 33,0 L 5.98 0.295 0.00936 31.5 5 8.21 0.405 0.012 33.8 6 10.4 0.515 0.0146 35.3 7 12.5 0.618 0.0186 33,0 8 16.2 0.798 0.0216 36.9 9 19.2 0.95 0.0252 7.7 10 30.8 1.52 0.0L437 34.8 11 59.2 2.92 0.07 h1.7 12 92.8 L.58 0.107 h2.8 13 145.9 7.2 0.152 hy.L ®Measured under conditions at top of column. Table 1. Summary of Data on Gas Holdup Obtained in a Z-in.-ID Column Containing Water During Third Study Number of orifices: Gas Flow superficial pubple 1 G Fl Superticial Bubble Rate 2 Gas Gas Inlet Rise a;ategw Gas Gas Inlet Rise Run . Velocity Orifice ID Gas Velocity Run Velocity Orifice ID Gas Velocity No. {em /sec) (cm/sec ) (mm ) Holdup (cm/sec) No. (cm5/sec) (cm/sec ) (mm ) Holdup (cm/sec) 1 .80 0. %65 0.4 0.0142 27.1 35 h.76 0.235 2.0 0.0129 18.2 ;2 0.926 0. 0Ly 0.k 0.00283 16.1 36 10.2 0.503 2.0 0.0237 21.2 5 2.50 0.123y O.h4 0.00k(3 26.1 37 20.3 1.0 2.0 0.0302 33.1 L L.9o 0,02 O.4 0.00988 2h .y 38 5.13% 0.253 4.0 0.0075 33,7 g 10.9 0.54%7¢ Ok 0.0186 28.9 39 10.3 0.509 L.o 0.0219 25.2 6 1,.2 0.k 0.k 0.0276 7.1 ~ Lo 31.2 1.54 4.0 0.0376 41.0 o 19.4 0. 96 0.h 0.03544 27.9 41 8h.5 L1t 4.0 0.098 4o.6 8 344 1.7 0.4 0.0632 26.9 42 1.90 0.0938 L.o 0.0033%5 28.0 9 65.3 3.22 Ouh 0.101 31.9 43 10.4 0.512 4.0 0.0159 32.2 10 97.1 ] 0.4 0.127 7.7 LYy Ls.8 2.26 L.0o 0.0713 39.7 11 141.9 ‘.00 0.4 0.157 4h.6 4s 111.5 5.5 4.0 0.125 k.o 12 171.5 846 0.k 0.176 48.1 L6 450 2.2 L.o 0.339 65.5 15 13.9 0.688 1.0 0.0205 33.6 47 282 13.9 4.0 0.24k7 56.3 1h 1.5 0.0767 1.0 0.00264 29.1 43 1y. f 0.9 4.0 0.026% 56.9 1. 3.20 0.158 1.0 0.0048¢ 32.6 49 2.45 c.121 4.0 0.00418 25.9 1 L 0.26 1.0 0.00615 42.3 50 26.3 1.3 4.0 0.0346 37-6 17 T.0% O. 547 1.0 0.00835 L1.6 51 14.6 0.72 4.0 0.0188 38.% 18 3,30 0459 1.0 0.011hL 40.3 52 7.56 0.373 L.o 0.0368 10.1 19 11.9 0.589 1.0 0.014 Lol 53 1.04 0.0-12 2.0 0.00176 29.1 20 1§.6 0.92 1.0 0.025% 36.4 54 2.53 0.125 2.0 0.00Uk 28.4 21 50.8 1.52 1.0 0.0405 37-5 55 L.09 0.202 2.0 0.0057 35.4 22 59. & 2.95 1.0 0.0738 40.0 56 6.16 0.304 2.0 0.0105 28.9 2% - 86.5 .oy 1.0 0.10 k2.7 57 8.21 0.405 2.0 0.018 22.5 24 113.3 5.59 1.0 0.124 4.1 58 10.6 0.525 2.0 0.0219 24.0 25 147.8 T.29 1.0 0.127 57.4 59 13.2 0.65 2.0 0.0175 37.1 26 280 15.8 1.0 0.228 £0.5 €0 15.4 0.76 2.0 0.0145 52.4 27 88.5 Lh.35 2.0 0. 107 LO.T 61 18.7 0.923 2.0 0.024T 37-4 28 5.37 0.265 2.0 0.011% 23.5 62 30.9 1.525 2.0 0.0k413 26.9 29 16.2 0.80 2.0 0.0543 23.% 63 63.8 3.15 2.0 0.0745 L2.3 50 50.7 2.4 2.0 0. 061 L1.0 o 89.2 i 2.0 0.0982 4h.8 31 185.4 9.1% 2.0 0.18 50.8 65 117.8 5.81 2.0 0.127 Ls.7 32 458 21.6 2.0 0. 529 65.7 66 1hh .5 7.13 2.0 0.15% hA. 0 53 2.2% 0.11 20 0.0052 21.2 67 330 16.5% 2.0 0.26 €2.7 34 k.ol U198 2.0 0.0082 2h.1 e a Measured under conditions at top of column. 83 Table 32. Summary of Data on Gas Holdup Obtained in a 2.0-in.-~ID Column During Third Study Gas inlet: 19 orifices, 1 mm ID Gas Flow Superficial Bubble Run Rate? Gas Velocity Gas Rise Velocity No. (cm3/seC) (cm/sec) Holdup (cm/sec) 1 1.02 | 0.0505 0.00219 23.1 2 2,31 0.114 0.00392 29.1 3 4.22 0.208 0.0061 34.1 4 6.10 0.301 0.0087 34.6 5 8.19 0.404 0.0117 34,5 6 10.8 0.535 0.0156 34.3 7 12.8 0.632 0.0173 36.5 8 14.7 0.726 0.021 34.6 9 18.6 0.919 0.0253 36.3 10 30.0 1.48 0.0353 41.9 11 61.2 3.02 0.0748 40.4 12 85.9 4.24 0.11 38.5 13 115.7 5.71 0.133 42,9 14 149.6 7.38 0.158 46.7 15 280 13.8 0.249 55.4 dMeasured under conditions at top of column. 8l Table 33. Summary of Data on Gas Holdup Obtained in a 2.0-in.-ID Column Containing Water During Third Study Gas inlet: 37 orifices Gas Flow Superficial Gas Inlet ' Bubble Run Rate? Gas Velocity Orifice ID Gas Rise Velocity No. (cm3/sec) (cm/sec) (mm) Holdup (cm/sec) 1 1.06 0.0525 1.0 0.00307 17.1 2 2.41 0.119 1.0 0.00438 27.2 3 3.46 0.171 1.0 0.00788 21.7 4 5.86 0.289 1.0 0.00918 31.5 5 7.96 0.393 1.0 0.0118 33.3 6 10.7 0.529 1.0 0.0144 36.7 7 13.9 0.687 1.0 0.0195 35.2 8 15.8 0.78 1.0 0.0282 27.7 9 19.2 0.945 1.0 0.0288 32.8 10 30.8 1.52 1.0 0.0474 32.1 11 60.2 2.97 1.0 0.0895 33.2 12 88.4 4.36 1.0 0.097 44.9 13 113.5 5.6 1.0 0.144 38.9 14 149.0 7.35 1.0 0.174 42.2 15 259 12.8 1.0 0.233 54.9 16 503 24.8 1.0 0.325 76.3 17 746 36.8 1.0 0.395 93.2 18 2.27 0.112 2.0 0.0039 28.7 19 7.62 0.376 2.0 0.0112 33.6 20 13.6 0.67 2.0 0.0185 36.2 21 1.42 0.0703 2.0 0.00175 40.2 22 5.31 0.262 2.0 0.00825 31.8 23 9.69 0.478 2.0 0.0148 32.3 24 17.7 0.875 2.0 0.024 36.5 25 27.8 1.37 2.0 0.0372 36.8 26 44.8 2.21 2.0 0.0592 37.3 27 101 5.0 2.0 0.12 41.7 28 72.6 3.58 2.0 0.0895 40.0 29 3.69 0.182 2.0 0.0065 28.0 30 148 7.28 2.0 0.16 45.5 31 590 29.12 2.0 0.385 75.6 32 14.9 0.737 4.0 0.0218 33.8 33 1.25 0.0618 4.0 0.0026 23.8 34 3.00 0.148 4.0 0.00488 30.3 35 5.09 0.251 4.0 0.00574 43.7 36 6.71 0.331 4.0 0.00795 41.6 37 8.17 0.403 4.0 0.015 26.9 38 11.8 0.585 4.0 0.0185 31.6 39 20.5 1.01 4.0 0.0282 35.8 40 30.2 1.49 4.0 0.0424 35.1 41 53.9 2.66 4.0 0.071 37.5 42 73.4 3.62 4.0 0.0876 41.3 43 94.4 4.66 4.0 0.109 42.8 44 118 5.8 4.0 0.131 44.3 45 150 7.43 4.0 0.155 47.9 46 537 26.5 4.0 0.34 77.9 a . Measured under conditions at top of celumn. 85 Table 34. Summary of Data on Gas Holdup Obtained in a 3.0-in.-ID Column Containing a Water—Glycerin Mixture During Third Study Gas inlet: one orifice, 0.638 cm ID Gas Flow Superficial Glycerin Bubble Run Ratead Gas Velocity Concentration Gas Rise Velocity No. (cm3/sec) (cm/sec) (wt 72) Holdup (cm/sec) 1 36.9 0.81 25 0.0252 32.1 2 1.60 0.0352 25 0.000928 37.9 3 3.50 0.0768 25 0.0045 17.1 4 5.56 0.122 25 0.004 30.5 5 8.94 0.196 25 0.0056 35.0 6 13.0 0.285 25 0.01 28.5 7 19.4 0.425 25 0.0107 39.7 8 25.2 0.553 25 0.0162 34,1 9 46.7 1.025 25 0.0295 34.7 10 84.8 1.86 25 0.0492 37.8 11 130.0 2.85 25 0.0765 37.3 12 258 5.66 25 0.113 50.1 13 377 8.26 25 0.171 48.3 14 1.47 0.0323 25 0.00141 22.9 15 3.51 0.0769 65 0.00188 40.9 16 4.92 0.108 65 0.00327 33.0 17 7.07 0.155 65 0.00468 33.1 18 10.6 0.232 65 0.00658 35.3 19 15.4 0.338 65 0.00889 38.0 20 19.3 0.424 65 0.0117 36.2 21 30.8 0.675 65 0.0199 33.9 22 76.2 1.67 65 0.043 38.8 23 57.0 1.25 65 0.0316 39.6 24 104 2.28 65 0.0628 36.3 25 135 2,96 65 0.0701 42.2 26 245 5.37 65 0.114 47.1 aMeasured under conditions at top of column. 86 A comparison of the experimental data with values predicted by the corre- lation is shown in Fig. 22. Equation (33) predicts that gas holdup should be essentially proportional to superficial gas velocity at low gas flow rates and that the dependence should decrease as the gas flow rate is increased. A limiting value for gas holdup of about 0.68 is predicted. The following expression for the bubble rise velocity, Vg/h, can be obtained from Eq. (33): % \Ved = = 1.468 vg + 0.4071 gdc . (34) This relation predicts that the bubble rise velocity should be essen- tially constant at low gas flow rates and should depend on the square root of the column diameter. The bubble rise velocity should increase at a rate that is essentially proportional to the superficial gas velocity at high gas flow rates and should show little dependence on the column diameter. The physical properties of the liquid are of negligible importance in determining either gas holdup or bubble rise velocity. Comparisons of the predicted and experimentally determined values for gas holdup and bubble rise velocity are shown in Figs. 23-40. The measured and the predicted values are seen to be in good agreement. Most of the predicted values lie within 15% of the measured values. Figures 23-32 show the variations of gas holdup and bubble rise velocity resulting from changes in the viscosity of the liquid phase and the superficial gas velocity for columns having diameters of 1.0, 1.5, 2, %, and 6 in. In the bubble flow regime, the bubble rise velocity is essentially constant at about 25 to 40 cm/sec, and the gas holdup increases almost linearly as the superficial gas velocity is increased. Increases in the column diameter result in small increases in the bubble rise velocity and small decreases in the gas holdup. In the slug flow regime, the bubble rise velocity increases with increases in the super- ficial gas velocity and the gas holdup is not linearly dependent on the superficial gas velocity. 87 COMMON LOGARITHM OF MEASURED (Vg /h) e | ] | ] | 1.4 15 1.6 0.7 1.8 19 2.0 COMMON LOGARITHM OF PREDICTED (Vg/h) Fig. 22. Comparison of Experimentally-Determined and Calculated Data on Gas Holdup in Open Bubble Columns Having Diameters Ranging from 1.5 to 6 in. 83 ORNL DWG 72-11742 1T T T 11 T o. T T TTT] GAS HOLDUP T T T 1771717171 T | PREDICTED VALUES Lit1i1al 1 0.0l 5 : 0.00I i 1113111l 1 L3 1 13119l 1 L 1.1 1111 o.1 1.0 10 100 SUPERFICIAL GAS VELOCITY (cm /sec) Fig. 25. Variation of Gas Holdup with Changes in Superficial Gas Velocity in a 1.0-in.-ID Bubble Column. BUBBLE RISE VELOCITY (cm/sec) ORNL DWG 72-11743 140 1 T T 1T T T TT] T T T 120 - 2 pts. v 100 - — 80 - — 60 - 40 |- = PREDICTED VALUES o O 20 o oy —d o o © 860 © 0 L 1 1111 1 1 i P 11 1al 1 1 11 1 111 0.1 1.0 10 100 SUPERFICIAL GAS VELOCITY(cm /sec) Fig. 24. Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity in a 1.0-in.-ID Bubble Column. €3 90 ORNL DOWG 72-13507 LERLLIL » I GAS HOLDUP T T VrrrTg f T rrreg | T 1T 7T 7T — — - PREDICTED VALUES L 11l ) 1 1 VISCOSITY (¢P) - 0.89 O 208 A 1.3 Ll 1 L1 1]l L 11l 1 L 1 L 111l ] 1 1 0.001 0 Fig. Velocity 25. 1.0 10 SUPERFICIAL GAS VELOCITY (cm/sec) 100 Variation of Gas Holdup with Changes in Superficial Gas and Viscosity of Liquid in a 1.5-in.-ID Bubble Column. BUBBLE RISE VELOCITY (ecm/sec) ORNL DWG 72-13506 140 T T T T T T 17717 1 T T T UTTg I 1 v rrrT VISCOSITY (cP) 120 e 0.89 - e 205 a L3 IOO = -t 80 | 4 PREDICTED VALUES 60 - ey \Zp't - \fi 40 | , 1 ® .. . A & e %M ® ® —— A e 20 e - 0 1 1 1 1 11 11 1 i 1 1 1 1 2l 1 1 1 A 14 i 1 0.1 I 10 100 SUPERFICIAL GAS VELOCITY (cm /sec) Fig. 26. Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Viscosity of Liquid in a 1.5-in.-ID Bubble Column. 92 ORNL DWG 72-13519 |.0_ T T T !riiI] 1 T T T T 1117 T T T rriT)] e A -y = . m - [ /‘3 4 0 - . EP - PREDICTED VALUES fi/ B . 0.1 : = l — : - o 3 - o T VISCOSITY - g (cP) © . 0.89 7 a 2.05 0.9 D .3 — 0.001 1 1 1411l L 11t 2114l 1 1 11 1 13} 0.1 1O 10.0 100.0 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 27. Variation of Gas Holdup with Changes in Superficial Gas Velocity and Viscosity of Liquid in a 2.0-in.-ID Bubble Column. BUBBLE RISE VELOCITY (cm/sec) ORNL DWG 72-13518 140 T 7 T T T T, T T T T T T T T T Y - VISCOSITY (cP) 120 |- * 0.89 4 a 1.8 - @ 2.1 - 100 | i . | o 80 |- i e 40 o 4 e ? e * o a. * D ° ’ * PREDICTED VALUES 20} | | : 0 L1 | | 1 11 1a1 11 1 11 L1 1 11 B 1 1 L1 11 l 4.3 111 0.04 0.05 007 ol 02 05 Q7 |10 20 130 50 70 10 20 40 SUPERFICIAL GAS VELOCITY (cm /sec) Flg 28. Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Viscosity of the Liquid in a 2.0-in.-ID Bubble Column. ¢6 GAS HOLDUP 9L ORNL DWG 72-1352I| 1.0 T T 0T YT T T Illill[ T Y T Uiy T Y T T rrT7T [ VISCOSITY (¢P) 3 - e 0.89 - A 2.05 - - g 10.7 - 0.4} - - - — PREDICTED VALUES 1 0.0l |- — n ] - - a 0.00! i £ 1t il 1t 4 vl i 1t e et by 11 0.01 0.1 1.0 10.0 100.0 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 29. Variation of Gas Holdup with Changes in Superficial Gas Velocity and Viscosity of Liquid in a 3.0-in.-ID Bubble Column. BUBBLE RISE VELOCITY (¢m/sec) 140 130 120 11O 90 80 70 60 50 40 30 20 95 ORNL DWG 72-13517 s VISCOSITY (cP) ® a o T rrIrty 1 v rTrTrrirng ! 1 TV rrrry v T 1T 7T 7177177 0.89 2.05 10.7 PREDICTED VALUES ~ 05 1111l 1 L Lt 11l i L4 431911 L L.l 1 1111 Fig. 30. 0.l 1.0 10.0 100.0 SUPERFICIAL GAS VELOCITY (cm/sec) Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Viscosity of Liquid in a 3.0-in.-ID Bubble Column. 96 ORNL DWG 72-13516 1.0 T T T T T T7TT] T T T T rTTITg 1 — T T I - 3 0.1 - - a - ] = g VISCOSITY (cP) - T — . 0.89 - o A 1.8 - - B 12.1 - PREDICTED VALUES 0.0l . 0.00!| i L4 111l 1 11 1133l 1 1 1 1 1111 0.1 1.0 10.0 100.0 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 51. Variation of Gas Holdup with Changes in Superficial Gas Velocity and Viscosity of Liquid in a 6.0-in.-ID Bubble Column. (cm/sec) BUBBLE RISE VELOCITY 70 50 40 20 ORNL DWG. 72-13515 a i PREDICTED VALUES / - * ° VISCOSITY (cp) . 0.89 - A 1.8 a 12.1 1 i ) i 1 1 1 1 0.1 0.2 0.5 1.0 2 5 10 20 50 100 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. %2. Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Viscosity of Liquid in a 6.0-in.-ID Bubble Column. L6 GAS HOLDUP g8 ORNL DWG 72-13522 |.0:‘ 1 1 FTETITTg I T Flllll] ; ! ! VITTETT I T T TTT1 :SURFACE TENSION ] - (dynes/cm) [ . 7200 A _ ) » L 453 - & 273 0.1 - 4 I PREDICTED VALUES i 0.0l .: 0.00| 1 1 ¢ttt 1 Lt 1istal 1 1 4 11l 1 L L L 111y 0.0l 0.1 | 10 |00 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 33. Variation of Gas Holdup with Changes in Superficial Gas Velocity and Surface Tension of Liquid in a 1.5-in.-ID Bubble Column. BUBBLE RISE VELOCITY (cm/sec) ORNL DWG 72-13514 80 T T 1] T T T T 7 T T 1] T T T 1T 171t [ I T T SURFACE TENSION (dynes/cm) . T2 701- O 453 . A 723 o* 60+ — 50 — a0 - o PREDICTED o * VALUES 30 A 20 — A D A A A a . Q A 20} 4 jol—1 11 Ll 1 | Lt 1t ] i L1 11l ] i | 0.05 0.1 0.2 0.5 1.0 2.0 50 10 20 50 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 34. Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Surface Tension of Liquid in a 1.5-in.-ID Bubble Column. 66 GAS HOLDUP 100 ORNL DWG 72-13523RI |0 T ]|]||||] T ]]][1[[| T I‘Illllll T r 17T . 7T71T07 C ORIFICE DIAMETER - t {mm) . N + 0.4 i » a |0 , i O 20 PREDICTED VALUES R0 - © 40 g° Ol | - r_ . 0.0l - . . i 0.001 i L 111l 1 i 4 1ok 1 1 1 i L4 1.1 0.01 0.1 t.0 10.0 100.0 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 35. Variation of Gas Holdup with Changes in Superficial Gas Velocity and Diameter of Orifices in Gas Distributor in a 2.0-in.-ID Bubble Column Filled with Water. The gas distributor consisted of five orifices of the diameter indicated. BUBBLE RISE VELOCITY (cm/sec) 101 ORNL DWG. 72-13513 80 g ! T T T T T ORIFICE DIAMETER (mm) . 0.4 70} A |.0 4 ® 2.0 + 4.0 60 - a 50+ o A A . 40} 4 a + "a - u+ » + Q ‘. A VALUES f ‘\‘ . ® ® . s ‘g a s s 20} - = _ » 10 ] 1 i L 1 i l L 0.04 0.08 0.2 04 08 20 40 8.0 20 40 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. %6. Variation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Diameter of Crifices in Gas Distributor in a 2.0- in.-ID Bubble Column Filled with Water. The gas distributor consisted of five orifices of the diameter indicated. 102 ORNL DWG 72-135I12R!I 1.0 T T T TTTTT T T. T 7T TTT] T T T T T TT71] — - |- 0.l i : — PREDICTED VALUES ] o i - D — - Q - [~ - o I - wn g — - o 0.0l = — 0.00| 1 1 L1 1 1 1 11| i 1 L1 0.05 0.1 03 0.5 1.0 3 5 10 30 50 SUPERFICIAL GAS VELOCITY (cm/ sec) Fig. 5(. Variation of Gas Holdup with Superficial Gas Velocity in a 2.0-in.-ID Bubble Column Filled with Water. The gas distributor con- sisted of 19 orifices, each having a diameter of 1.0 mm. BUBBLE RISE VELOCITY (cm/sec) ORNL DWG 72-1351I TOr 60 - 50 - 40 |- 30 | PREDICTED VALUES i L 1 1 1 L 1 | 0.1 0.2 0.5 | 2 5 10 20 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 38. Variation of Bubble Rise Velocity with Superficial Gas Velocity in a 2.0-in.-ID Bubble Column Filled with Water. The gas distributor consisted of 19 orifices, each having a diameter of 1.0 min « 50 ¢01 10k | ORNL DWG 73-Ti7 0 T T T TTTTT T ™ T T TTTTT T =T T T 0.7 e — — 1 0.5 C.3- PREDICTED Ol VALUES 0.07 L1 111t 0.05 ] 0.03 0.02 ORIFICE DIAMETER (mm) . 1.0 A 2.0 B 4.0 GAS HOLDUP 0.0l 0.0C7 0.005 11 1 111 i 0.0C3 0002 0.001 1t 111 g1 1 L L1111 ] 1 1 (htr] 0.05 ol 02 03 05 1.0 20 30 50 10 20 30 50 SUPERFICIAL GAS VELOCITY (cm/sec) Fig. 39. Variation of Gas Holdup with Changes in Superficial Gas Velocity and Diameter of Orifices in Gas Distributor in a 2.0-in.-ID Bubble Column. The gas distributor consisted of 37 orifices of the diameter indicated. BUBBLE RISE VELOCITY {(cm /sec) ORNL DWG 72-135I10 140 T T T T T T T T . ORIFICE DIAMETER 120 | (mm) _ e |.0 A 20 100 - g 4.0 80 PREDICTED VALUES 50T o I 1 l 1 ] 1 1 ] 0.05 0.1 0.2 0.5 I 2 5 10 20 50 SUPERFICIAL GAS VELOCITY (cm/ sec) Fig. 40O. vVariation of Bubble Rise Velocity with Changes in Super- ficial Gas Velocity and Diameter of Orifices in Gas Distributor in a 5.0-in.-ID Bubble Column. The gas distributor consisted of 37 orifices having the indicated diameter. 106 The variations of gas holdup and bubble rise veloéity with changes in the surface tension of the liquid and the superficial gas velocity are shown in Figs. 33 and 34 for a column diameter of 1.5 in. Decreases in the surface tension tend to increase the gas holdup, to decrease the bubble rise velocity, and to cause the transition from bubble flow to slug flow to occur at higher superficial gas velocities; on the other hand, the importance of changes in these parameters is very slight. Data on gas holdup and bubble rise velocity obtained in a 2-in.-diam column with multiple gas inlets having diameters of 0.4, 1.0, 2.0, and 4.0 mm are shown in Figs. 35-40. Within the accuracy of the data, there is no dependence of gas holdup or bubble rise velocity on the number or diameter of the orifices in the gas distributor except in the case of the smallest orifice diameter studied (O.4 mm). The bubble rise velocity observed with the O.4-mm-diam orifice is appreciably lower, and the gas holdup is higher, than for the larger orifice diameters. 9.6 Compilation of Data on Axial Mixing A compilation was made of the axial dispersion data that have been accumulated in experiments made thus far. This was done in preparation for developing a general correlation for predicting axial dispersion in open bubble columns. The information, which consists of about 420 measurements of the axial dispersion coefficient in columns with diameters ranging from 1.0 to 6.0 in., was divided into 34 data sets (as shown in Table 35). The results obtained in the individual runs with column diameters of 1.0, 1.5, 2.0, 3.0, and 6.0 in. are summarized in Tables 36, 37, 38, 39, and 40, respectively. 9.7 Correlation of Data on Axial Dispersion The axial dispersion data were analyzed using dimensional analysis and were fitted by the method of least squares to the resulting power- law expression. The power-law expression is of the form 107 Table 33. Description of Data Sets for Data on Axial Dispersion in Open Bubble Columns Data Column Orifice Number Fluid Surface Number Set Diameter Diameter of Viscosity Tension Density of Data No. (in.?} {cm) Orifices (cP) {(dynes/cm) {g/cm3) Points Reference 1 1.5 0.432 1 0.894 73.0 0.997 15 a 2 2.0 0.432 1 0.894 73.0 0.997 7 a 3 6.0 0.432 1 0.894% 73.0 0.997 16 a 4 2.0 0.1 1 0.894 73.0 0.997 13 b 5 2.0 0.04 5 0.89% 73.0 0.997 12 b b 2.0 0.1 5 0.89% 713.0 0.997 14 b 7 2.0 0.2 5 0.894 73.0 0.997 26 b 8 2.0 0.4 5 0.894 73.0 0.997 16 b 9 2.0 0.1 19 0.894 73.0 0.997 15 b 10 2.0 0.1 37 0.894 73.0 0.997 17 b 11 2.0 0.2 37 0.894 73.0 0.997 14 b 12 2.0 0.4 37 0.894 73.0 0.997 15 b 13 1.5 0.432 1 0.894 73.0 0.997 11 b 14 1.5 0.432 1 1.07 45,3 0.995 11 b 15 1.5 0.432 1 1.09 27.3 0.996 11 b 16 1.5 0.432 1 2.05 71.5 1.04 10 b 17 1.5 0.432 1 11.3 70.0 1.15 10 b 18 3.0 0.432 1 2.05 71.5 1.04 14 b 19 3.0 0.432 1 10.7 70.0 1.15 13 b 20 2.0 0.432 1 1.8 71.8 1.05 12 a 21 2.0 0.432 1 12.1 67.9 1.16 13 a 22 1.5 0.432 1 1.8 71.8 1.05 11 a 23 h.0 0.432 1 1.8 71.8 1.05 13 a 24 6.0 0.432 1 12.1 67.9 1.16 10 a 25 2.0 0.102 1 0.894 37.6 0.997 6 23 26 2.0 0.102 1 0.89% 53.2 0.997 4 23 27 2.0 0.102 1 0.894 73.0 0.997 5 23 '8 2.0 0.102 1 15.0 65.2 1.15 5 23 29 2.0 (G.102 1 0.89% 67.4 0.997 1 23 30 2.0 0.102 1 0.894 73.0 0.997 21 22 31 1.5 0.432 1 0.894 73.0 0.997 7 25 52 2.0 0.432 1 0.89% 73.0 0.997 13 75 33 3.0 0.432 1 0.894 73.0 0.997 9 25 1L 1.0 0.432 1 0.8%4 73.0 0.997 38 C dhata obtained during second study. I Pata obtained during third study. "bata obtained during first study. Table 36. Experimentally Determined Values® for the Axial Dispersion Coefficient in a 1.0-in.~-ID Open Bubble Column Filled with Water Daita set No. 34 Superficial Axial Dispersion Superficial Axial Dispersion Run Gas Velocity Coefficient Run Gas Velocity Coefficient No (em/sec) (cm?/sec) No. (cm/sec) (cm?/sec) 1 18.7 82.3 20 44.8 181.1 2 76.6 319.0 21 1.166 20.8 3 76.6 261.8 22 1.088 18.3 4 65.4 255.2 23 1.011 17.1 5 74.7 182.3 24 0.933 22.5 6 56.0 319.0 25 0.855 17.1 7 37.3 170.2 26 0.778 15.5 8 18.7 63.8 27 0.700 14.1 9 28.0 170.2 28 0.156 5.84 10 46.1 232.0 29 1.25 14.8 11 10.3 52.0 30 1.02 24.6 12 18.7 92.8 31 0.78 19.7 13 28.9 128.3 32 0.468 14.8 14 23.3 102.6 33 0.156 7.04 15 14.0 61.1 34 1.33 22.4 16 32,7 146.6 35 1.01 | 17.8 17 37.3 213.8 36 0.702 14.5 18 46.7 185.5 37 0.390 10.7 19 63.5 285.1 38 0.156 6.42 a Data obtained during first study. 801 Table 37. Experimentally Determined Values for the Axial Dispersion Coefficient in a 1.5-in.-ID Open Bubble Column Gas Properties of Liquid Axial Data Superficial Disperser Surface Dispersion Set Run Gas Velocity Diameter? Density Viscosity Tension Coefficient No. No. (cm/sec) (cm) (g/cm3) (cP) (dynes/cm) (cm?/sec) Ref. 1 1 16.4 0.635 0.997 0.894 73.0 138.5 b 1 2 8.58 0.635 0.997 0.894 73.0 85.6 b 1 3 12.3 0.635 0.997 0.894 73.0 114.8 b 1 4 9.89 0.635 0.997 0.894 73.0 93.4 b 1 5 1.72 0.635 0.997 0.894 73.0 31.6 b 1 6 35.4 0.635 0.997 0.894 73.0 482.2 b 1 7 0.69 0.635 0.997 0.894 73.0 20.8 b 1 3 35.4 0.635 0.997 0.894 73.0 380.5 b 1 9 2.99 0.635 0.997 0.894 73.0 39.0 b 1 10 5.11 0.635 0.997 0.894 73.0 ' 65.0 b 1 11 0.912 0.635 0.997 0.894 73.0 24,8 b 1 12 22.4 0.635 0.997 0.89% 73.0 177.0 b 1 13 10.4 0.635 0.997 0.894 73.0 149.0 b 1 14 16.8 0.635 0.997 0.894 73.0 136.0 b 1 15 13.6 0.635 0.997 0.894 73.0 95.2 b 13 1 1.045 0.635 0.997 0.89 72.0 25.8 c 13 2 0.212 0.635 0.997 0.89 72.0 12.9 c 13 3 0.451 0.635 0.997 0.89 72.0 18.9 c 13 4 0.719 0.635 0.997 0.89 72.0 24.4 c 13 5 1.33 0.635 0.997 0.89 72.0 32.9 c 13 6 1.68 0.635 0.997 0.89 72.0 31.6 c 13 7 2.21 0.635 0.997 0.89 72.0 39.5 c 13 8 3.86 0.635 0.997 0.89 72.0 50.0 ¢ 13 9 6.54 0.635 0.997 0.89 72.0 69.4 c 13 10 8.93 0.635 0.997 0.89 72.0 85.9 c 601 Table 37 (continued) Gas Properties of Liquid Axial Data Superficial Disperser Surface Dispersion Set Run Gas Velocity Diameter? Density Viscosity Tension Coefficient No. No. (cm/sec) (cm) (g/cm3) (cP) (dynes/cm) (cmzlsec) Ref. 13 11 21.3 0.635 0.997 0.89 72.0 242 c 14 1 1.045 0.635 0.995 1.07 45.3 20.9 c 14 2 0.710 0.635 0.995 1.07 45.3 23.3 c 14 3 0.535 0.635 0.995 1.07 45.3 20.5 c 14 4 0.252 0.635 0.995 1.07 45.3 17.3 c 14 5 0.0930 0.635 0.995 1.07 45.3 14.6 c 14 6 1.456 0.635 0.995 1.07 45.3 23.1 c 14 7 1.55 0.635 0.995 1.07 45.3 26.7 c 14 8 2.58 0.635 0.995 1.07 45.3 33.7 c 14 9 4,31 0.635 0.995 1.07 45.3 46.6 c 14 10 5.79 0.635 0.995 1.07 45.3 50.0 c 14 11 7.45 0.635 0.995 1.07 45.3 58.0 c 15 1 1.02 0.635 0.996 1.09 27.3 21.9 c 15 2 0.133 0.635 0.996 1.09 27.3 15.0 c 15 3 0.0705 0.635 0.996 1.09 27.3 11.9 c 15 4 0.214 0.635 0.996 1.09 27.3 15.8 c 15 5 0.494 0.635 0.996 1.09 27.3 20.6 c 15 6 0.763 0.635 0.996 1.09 27.3 22,2 c 15 7 1.37 0.635 0.996 1.09 27.3 22.3 c 15 8 2.07 0.635 0.996 1.09 27.3 27.9 c 15 9 3.5 0.635 0.996 1.09 27.3 36.2 c 15 10 6.4 0.635 0.996 1.09 27.3 51.5 c 15 11 8.6 0.635 0.996 1.09 27.3 65.7 c 16 1 0.0638 0.635 1.04 2.05 71.5 7.65 c 16 2 0.187 0.635 1.04 2.05 71.5 12.5 c 16 3 0.333 0.635 1.04 2.05 71.5 24,6 c 11 Table 37 (continued) Gas Properties of Liquid Axial Data Superficial Disperser Surface Dispersion Set Run Gas Velocity Diameter? Density Viscosity Tension Coefficient No. No. (cm/sec) (cm) (g/cm3) (cP) (dynes/cm) (cm2/sec) Ref. 16 4 0.528 0.635 1.04 2,05 71.5 26.4 c 16 5 0.895 0.635 1.04 2.05 71.5 28.4 c 16 6 1.25 0.635 1.04 2.05 71.5 36.4 c 16 7 1.625 0.635 1.04 2.05 71.5 36.4 c 16 8 2.61 0.635 1.04 2.05 71.5 43.5 c 16 9 5.78 0.635 1.04 2.05 71.5 60.6 c 16 10 8.07 0.635 1.04 2.05 71.5 81.0 C 17 1 0.214 0.635 1.15 11.3 70.0 10.6 c 17 2 0.0888 0.635 1.15 11.3 70.0 7.45 c 17 3 0.313 0.635 1.15 11.3 70.0 12.3 c 17 4 0.45 0.635 1.15 11.3 70.0 15.2 c 17 5 0.651 0.635 1.15 11.3 70.0 21.3 c 17 6 1.0 0.635 1.15 11.3 70.0 21.8 c 17 7 1.38 0.635 1.15 11.3 70.0 27.3 C 17 8 1.74 0.635 1.15 11.3 70.0 27.7 c 17 9 2.9 0.635 1.15 11.3 70.0 34.7 c 17 10 5.79 0.635 1.15 11.3 70.0 55.6 c 22 1 10.4 0.635 1.05 1.8 71.8 88.9 b 22 2 1.66 0.635 1.05 1.8 71.8 24.9 b 22 3 32,7 0.635 1.05 1.8 71.8 814 b 22 4 49.7 0.635 1.05 1.8 71.8 656 b 22 5 - 32.7 0.635 1.05 1.8 71.8 155 b 22 6 4.06 0.635 1.05 1.8 71.8 44,6 b 22 7 10.4 0.635 1.05 1.8 71.8 88.9 b 22 8 21.9 0.635 1.05 1.8 71.8 115 b 22 9 13.1 0.635 1.05 1.8 71.8 131 b TTT Table 37 (continued) Gas Properties of Liquid Axial Data Superficial Disperser Surface Dispersion Set Run Gas Velocity Diameterd Density Viscosity Tension Coefficient No. No. (em/sec) (cm) (g/cm3) (cP) (dynes/cm) (em2/sec) Ref. 22 10 10.4 0.635 1.05 1.8 71.8 56.2 b 22 11 6.62 0.635 1.05 1.8 71.8 62.2 b 31 1 0.631 0.102 0.997 0.89 73.0 18.3 25 31 2 7.88 0.102 0.997 0.89 73.0 74.2 25 31 3 3.11 0.102 0.997 0.89 73.0 39.4 25 31 4 25.3 0.102 0.997 0.89 73.0 194 25 31 5 20.7 0.102 0.997 0.89 73.0 154 25 31 6 13.3 0.102 0.997 0.89 73.0 100.3 25 31 7 11.2 0.102 0.997 0.89 73.0 97.6 25 cll %cas disperser consisted of a single orifice having the indicated inside diameter. bData obtained during second study. “Data obtained during third study. Table 38. Experimentally Determined Values for the Axial Dispersion Coefficient in a 2.0-in.-ID Open Bubble Column Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec) Inlets (cm) (g/cm?) (cP) (dynes/cm) (cme/sec) Ref. 2 1 4.0 1 0.432 0.997 0.894 73.0 51.0 a 2 2 20.3% 1 0.432 0.997 0.894 7%.0 206.0 a 2 3 2%.8 1 0.432 0.997 0.894 73%.0 172.7 a 2 h 4.4 1 0.432 0.997 0.894 73.0 14h.9 a 2 5 0.56 1 0.432 0.997 0.894 73.0 22.3 a 2 6 3.88 1 0.432 0.997 0.894 73.0 58.7 a 2 7 5.96 1 0.432 0.997 0.894 73.0 67.5 a L 1 0.0442 1 0.1 0.997 0.894 72.0 19.6 b L 2 0.126 1 0.1 0.997 0.894 72.0 28.1 b L 3 0.209 1 0.1 0.997 0.894 72.0 28.9 b b L 0.295 1 0.1 0.997 0.894 72.0 31.1 b L 5 0.405 1 0.1 0.997 0.894 72.0 37.6 b b 6 0.515 1 0.1 0.997 0.894 72.0 40.5 b L T 0.618 1 0.1 0.997 0.894 72.0 43,4 b L 8 0.798 1 0.1 0.997 0.894 72.0 45.5 b L 9 0.95 1 0.1 0.997 0.894 72.0 418.2 b b 10 1.52 1 0.1 0.997 0.894 72.0 52.7 b b 11 2.92 1 0.1 0.997 0.894 72.0 60.8 b b 12 4.58 1 0.1 0.997 0.894 72.0 72.6 b L 13 7.2 L 0.1 0.997 0.894 72.0 81.5 b 5 1 0.385 5 0.04 0.997 0.894 7%.0 40.6 b 5 2 0.045T 5 0.0k 0.997 0.894 7%.0 15.4 b 5 3 0.124 5 0.04 0.997 0.894 73.0 16.4 b 5 L 0.242 5 0.0h 0.997 0.894 73.0 22.8 b 5 5 0.537 5 0.0k 0.997 0.894 73.0 33,1 b ¢11 Table 38 (continued) Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec) Inlets (cm) (g/cm”) (cP) (dynes/cm ) (cme/sec) Ref. 5 6 0.749 5 0.0k 0.997 0.894 73.0 35.4 b 5 7 0.96 5 0.0k 0.997 0.894 73.0 36.4 b 5 8 1.7 5 0.0k 0.997 0.894 73.0 41.5 b 5 9 3.22 5 0.04 0.997 0.894 7%.0 57.8 b 5 10 L. 79 5 0.0k 0.997 0.894 73.0 60.8 b 5 11 7.00 5 0.0k 0.997 0.894 7%.0 72.6 b 5 12 8.46 5 0.04 0.997 0.894 7%.0 83%.6 b 6 1 0.688 5 0.1 0.997 0.894 7%.0 37.2 b 6 2 0.0767 5 0.1 0.997 0.894 73%.0 31.7 b 6 3 0.158 5 0.1 0.997 0.894 73.0 31.5 b 6 L 0.26 5 0.1 0.997 0.894 73.0 4.8 b 6 5 0.347 5 0.1 0.997 0.894 73.0 47.0 b 6 6 0.459 5 0.1 0.997 0.894 73.0 37.9 b 6 7 0.589 5 0.1 0.997 0.894 73.0 35,5 b 6 8 0.92 5 0.1 0.997 0.804 73.0 39.0 b 6 9 1.52 5 0.1 0.997 0.894 73.0 55.1 b 6 10 2.95 5 0.1 0.997 0.894 73.0 57.5 b 6 11 h.o7 5 0.1 0.997 0.894 7%.0 63.5 b 6 12 5.59 5 0.1 0.997 0.894 73.0 92.4 b 6 13 7.29 5 0.1 0.997 0.894 73.0 111.1 b 6 14 13.8 5 0.1 0.997 0.89%4 73.0 137.0 b T 1 h.35 5 0.2 0.997 0.894 75.0 71.6 b T 2 0.265 5 0.2 0.997 0.894 73.0 35,8 b 7 3 0.80 5 0.2 0.997 0.894 73.0 23,8 b 7 b 2.5 5 0.2 0.997 0.894 73.0 5Tk b W1l Table 38 (continued) Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No No. (cm/sec) Inlets (cm) (g/cm ) (cP) (dynes/cm ) (cm®/sec ) Ref. T 5 9.15 5 0.2 0.997 0.894 73.0 139.0 b T 6 21.6 5 0.2 0.997 0.894 73.0 200.0 b 7 T 0.11 5 0.2 0.997 0.894 7%.0 29.8 b T 8 0.198 5 0.2 0.997 0.894 73.0 30.5 b T 9 0.235 > 0.2 0.997 0.894 73.0 35.8 b T 10 0.50% 5 0.2 0.997 0.894 73.0 30.0 b 7 11 1.0 5 0.2 C.997 0.894 73.0 Lh.o b 8 1 0.253 5 0.4 0.997 0.894 73.0 27.1 b 8 2 0.509 5 O.L 0.997 0.89k 73.0 41.2 b 8 3 1.54 5 0.4 0.997 0.894 73.0 41.8 b 8 L h.17 5 0.4 0.997 0.89L4 73.0 65.5 b 8 5 0.0938 5 0.k 0.997 0.894 73.0 30.8 b 8 6 0.512 5 0.k 0.997 0.894 7%.0 38.3 b 8 7 2.26 5 0.4 0.997 0.894 73.0 54.0 b 8 8 5.5 5 0.4 0.997 0.894 73.0 69.5 b 8 9 22.2 5 0.4 0.997 0.894 73.0 326 b 8 10 13.9 5 0.4 0.997 0.894 73.0 129 b 8 11 0.97 5 0.4 0.997 0.894 73.0 33,7 b 8 12 0.121 5 O.k 0.997 0.894 73.0 26.4 b 8 13 1.3 5 O.L 0.997 0.894 3.0 41.8 b 8 14 0.72 5 0.4 0.997 0.894 73.0 36.8 b 8 15 0.373 5 0.L 0.997 0.894 73.0 32,5 b 8 16 0.076 5 0.4 0.997 0.89L 73.0 25.5 b T 12 0.0512 5 0.2 0.997 0.894 73.0 20.8 b T 13 0.125 5 0.2 0.997 0.804 73.0 41.0 b T 14 0.202 5 0.2 0.997 0.894 7%.0 L45.6 b S11 Table 38 (continued) Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec ) Inlets (cm) (g/cmB) (cP) (dynes/cm) (cme/sec) Ref. T 15 0.30kL 5 0.2 0.997 0.894 73.0 37.0 b 7 16 0.405 5 0.2 0.997 0.894 73.0 36.5 b T 17 0.525 5 0.2 0.997 0.894 3.0 ho.2 b 7 18 0.65 5 0.2 0.997 0.894 73.0 56.0 b T 19 0.76 5 0.2 0.997 0.89% 73.0 €2.7 b 7 20 0.923 5 0.2 0.997 0.894 73.0 52.2 b 7 21 1.525 5 0.2 0.997 0.894 73.0 46.8 b T 22 3,15 5 0.2 0.997 0.894 73.0 59.5 b T 23 L. 5 0.2 0.997 0.894 73.0 69.5 b T 24 5.81 5 0.2 0.997 0.894 73.0 75.2 b 7 25 T.13 5 0.2 0.997 0.894 73.0 96.5 b 7 26 16.3 5 0.2 0.997 0.894 73.0 87.2 b 9 1 0.0505 19 0.1 0.997 0.89%4 73.0 26.4 b 9 2 0.114 19 0.1 0.997 0.894 73.0 36.7 b 9 3 0.208 19 0.1 0.997 0.894 73.0 38.5 b 9 L 0.301 19 0.1 0.997 0.894 73.0 39.7 b 9 5 0.404 19 0.1 0.997 0.894 73.0 40.1 b 9 6 0.535 19 0.1 0.997 0.894 T73.0 4hh.5 b 9 T 0.632 19 0.1 0.997 0.894 73.0 46.5 b 9 8 0.726 19 0.1 0.997 0.894 73.0 L7.0 b 9 9 0.919 19 0.1 0.997 0.894 73.0 4.9 b 9 10 1.48 19 0.1 0.997 0.89L4 73.0 42.8 b 9 11 3,02 19 0.1 0.997 0.894 73.0 s5h.L b 9 12 4,24 19 0.1 0.997 0.894 73.0 62.5 b 9 13 5.71 19 0.1 0.997 0.894 3.0 80.4 b 9 14 T.38 19 0.1 0.997 0.894 73.0 124 b 9 15 13.8 19 0.1 0.997 0.894 73.0 155 b o1t Table 38 (continued) Superficial Gas Dispersor Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec) Inlets (cm) (g/cmj) (cP) (dynes/cm) (cmg/sec) Ref. 10 1 0.0525 37 0.1 0.997 0.89k 73.0 27.3 b 10 2 0.119 37 0.1 0.997 0.894 73.0 33.3 b 10 3 0.171 37 0.1 0.997 0.89k 73.0 Lo. L b 10 L 0.289 37 0.1 0.997 0.894 73.0 37.8 b 10 5 0.393 37 0.1 0.997 0.894 73.0 33,7 b 10 6 0.529 37 0.1 0.997 0.894 73.0 3h.2 b 10 T 0.687 37 0.1 0.997 0.894 73.0 45.0 b 10 8 0.78 37 0.1 0.997 0.894 73.0 48.0 b 10 9 0.945 37 0.1 0.997 0.894 73.0 48.0 b 10 10 1.52 37 0.1 0.997 0.894 73%.0 L. L b 10 11 2.97 37 0.1 0.997 0.894 73.0 58.2 b 10 12 4.36 37 0.1 0.997 0. 894 7%.0 121 b 10 13 5.6 37 0.1 0.997 0.894 73%.0 95.0 b 10 14 T.35 37 0.1 0.997 0.894 73.0 101 b 10 15 12.8 37 0.1 0.997 0.894 73.0 136 b 10 16 24 .6 37 0.1 0.997 0.894 73.0 214 b 10 17 36.8 37 0.1 0.997 0.894 73.0 228 b 11 1 0.112 37 0.2 0.997 0.894 73.0 36.7 b 11 2 0.376 37 0.2 0.997 0.894 73.0 34,7 b 11 3 0.67 37 0.2 0.997 0.894 73.0 39.3 b 11 L 0.0703% 37 0.2 0.997 0.894 73.0 20.5 b 11 5 0.262 37 0.2 0.997 0.89L 73%.0 35.5 b 11 6 0.478 37 0.2 0.997 0.894 73.0 33,2 b 11 T 0.875 37 0.2 0.997 0. 89k 75-0 38.0 b 11 8 1.37 37 0.2 0.997 0.894 7%.0 hg.5 b 11 9 2.21 37 0.2 0.997 0.894 73.0 61.0 b 11 10 5.0 37 0.2 0.997 0.894 7%.0 91.8 b L11 Table 38 (continued) . 20 Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec ) Inlets (em) (g/cm’) (cP) (dynes/cm) (cm?/sec) Ref. 11 11 3.58 37 0.2 0.997 0.89k 73.0 80.0 b 11 12 0.182 3T 0.2 0.997 0.894 73.0 33,9 b 11 13 7.28 37 0.2 0.997 0.894 73.0 100 b 11 14 29.1 37 0.2 0.997 0. 89k 73.0 269 b 12 1 0.737 37 0.4 0.997 0.894 73.0 33,4 b 12 2 0.0618 37 0.4 0.997 0.894 73.0 28.2 b 12 3 0.148 37 0.4 0.997 0.89L 73.0 33,1 b 12 L 0.251 37 0.k 0.997 0.894 73.0 31.9 b 12 5 0.331 37 0.4 0.997 0.894 73.0 30.0 b 12 6 0.403 37 0.4 0.997 0.894 73.0 33,7 b 12 T 0.585 37 0.4 0.997 0.894 73.0 32.4 b 12 8 1.01 37 0.4 0.997 0.89L 73.0 51.3 b 12 9 1.49 37 0.4 0.997 0.894 73.0 50.0 b 12 10 2.66 37 0.4 0.997 0.894 73.0 59.7 b 12 11 3.62 37 0.4 0.997 0.894 73.0 61.0 b 12 12 L. 66 37 0.4 0.997 0.894 73.0 72.2 b 12 13 5.8 37 0.4 0.997 0.89k 73%.0 79.5 b 12 14 T.43 37 0.4 0.997 0.894 73.0 77.5 b 12 15 26.5 37 0.4 0.997 0.894 73.0 164 b 20 1 5.78 1 0.432 1.04T 1.8 71.8 T72.2 a 2 8.75 1 0.432 1.047 1.8 71.8 97.8 a 20 3 10.4 1 0.432 1.04T 1.8 71.8 120 a 20 4 18.6 1 0.432 1.047 1.8 71.8 178 a 20 5 27.3% 1 0.432 1.0LT 1.8 71.8 Lot a 20 6 2.79 1 0.432 1.0LT 1.8 71.8 45.3 a 20 7 4.01 1 0.432 1.047T 1.8 71.8 52.1 a 811 Table 38 (continued) Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec) Inlets (cm) (g/cm) (cP) (dynes/cm) (cm?/sec) Ref. 20 8 0.53 1 0.432 1.047 1.8 71.8 26.5 a 20 9 0.41 1 0.432 1.047 1.8 71.8 26.5 a 20 10 0.75 1 0.432 1.047 1.8 71.8 31.2 a 20 11 1.21 1 0.432 1.0LT 1.8 71.8 3L.5 a 20 12 2.21 1 0.432 1.04T 1.8 71.8 L8.2 a 21 1 0.91 1 0.432 1.163 12.1 679 27.6 a 21 2 0.49 1 0.432 1.163 12.1 67.9 21 .4 a 21 3 0.31 1 0.432 1.163 12.1 67-9 18.9 a 21 L 1.24 1 0.432 1.163 12.1 67-9 27.8 a 21 5 2.35 1 0.432 1.163 12.1 67.9 3h.1 a 21 6 3.0 1 0.432 1.163 12.1 67-9 37.2 a 21 7 h.11 1 0.432 1.163 12.1 67.9 i L a 21 8 5.77 1 0.432 1.163 12.1 67.9 51.0 a 21 9 9.1 L 0.43%2 1.163 12.1 67-9 75.3 a 21 10 10.3 1 0.432 1.163 12.1 67-9 79.2 a 21 11 16.1 1 0.432 1.163 12.1 67.9 105 a 21 12 2L.9 1 0.432 1.163 12.1 67-9 144 a 21 13 37.h 1 0.432 1.163 12.1 67.9 21k a 30 1 0.405 1 0.102 0.997 0.894 73.0 29.6 22 30 2 0.493 1 0.102 0.997 0.894 73.0 27.% 22 30 3 0.7h4 1 0.102 0.997 0.894 73.0 29 22 30 L 1.50 1 0.102 0.997 0.89h 73.0 33,3 22 30 5 2.55 1 0.102 0.997 0.894 7%.0 39.1 22 30 6 3.6L 1 0.102 0.997 0.894 73.0 52.0 22 30 7 0.256 1 0.102 0.997 0.894 73.0 26.9 22 30 8 0. 404 1 0.102 0.997 0.894 7%.0 28.4 22 30 9 0.405 1 0.102 0.997 0.894 73.0 28.6 22 611 Table 38 (continued) Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec) Inlets (cm) (g/cm”) (cP) (dynes/cm) (cm?/sec) Ref. 30 10 O. LT 1 0.102 0.997 0.89k 73.0 26.4 22 30 11 0.582 1 0.102 0.997 0.894 73.0 31.4 22 30 12 0.671 1 0.102 0.997 0.894 73.0 28.9 22 30 13 0.784 1 0.102 0.997 0.894 73.0 29.6 o2 30 14 0.918 1 0.102 0.997 0.894 73.0 28.9 22 30 15 1.159 1 0.102 0.997 0.894 73.0 32.1 22 30 16 1.50 1 0.102 0.997 0.89k 73.0 32.6 22 30 17 1.55 1 0.102 0.997 0.894 73.0 35.6 22 30 18 2.26 1 0.102 0.997 0.894 73.0 35 22 30 19 2.91 1 0.102 0.997 C.894 73.0 43,6 22 30 20 3.69 1 0.102 0.997 0.894 73.0 L8 22 30 21 5.28 1 0.102 0.997 0.894 73.0 68.7 22 27 1 0.432 1 0.102 0.997 0.894 73.0 28.5 23 o7 2 3.31 1 0.102 0.997 0.894 T3.0 Lo.T 23 27 3 3.3] 1 0.102 0.997 0.894 73.0 61.7 23 27 L 18.7 1 0.102 0.997 0.894 73%.0 6T.7 23 27 5 10.9 1 0.102 0.997 0.894 73.0 76.8 23 28 1 0.432 1 0.102 1.15 15 65.2 19.1 23 28 2 3.3] 1 0.102 1.15 15 65.2 Lg.7 23 28 3 2.28 1 0.102 1.15 15 65.2 ho.T 23 28 Y 1.77 1 0.102 1.15 15 65.2 28.2 23 28 5 1.25 1 0.102 1.15 15 65.2 23.6 2% 25 1 0.432 1 0.102 0.997 0.894 37.6 38.8 23 25 2 1.25 1 0.102 0.997 0.894 37.6 Ly, 6 23 25 3 2.28 1 0.102 0.997 0.894 37.6 45.6 23 25 L 3.31 1 0.102 0.997 0.89k 37.6 57.1 23 0cl Table 38 (continued) Superficial Gas Disperser Design Properties of Liquid Axial Data Gas Number Inlet el Surface Dispersion Set Run Velocity of Diameter Density Viscosity Tension Coefficient No. No. (cm/sec) Inlets (cm) (g/cm”) (cP) (dynes/cm) (cm?/sec) Ref. 25 5 0.201 1 0.102 0.997 0.894 37.6 37.9 23 25 6 . L8 1 0.102 0.997 0.894 37.6 89.1 23 26 1 0.201 1 0.102 0.997 0.894 5%.2 37.9 23 26 2 3.31 1 0.102 0.997 0.890L 53,2 55.0 23 26 3 0.432 1 0.102 0.997 0.894 53.2 ho.2 23 26 L L. 48 1 0.102 0.997 0.89k 5%.2 9.k 23 29 1 1.25 1 0.102 0.997 0.89k 647 39.0 23 30 1 10.2 1 0.432 0.997 0.894 73.0 113 25 32 2 15.0 1 0.432 0.997 0.894 73.0 148 25 32 3 18.5 1 0.432 0.997 0.894 73.0 139 25 32 L 23.3 1 0.432 0.997 0.894 73.0 176 25 32 5 21,0 1 0.432 0.997 0.894 7%.0 145 25 32 6 2k L 1 0.432 0.997 0.894 7%.0 282 25 32 T 25.4 1 0.432 0.997 0.894 3.0 192 25 32 8 37,7 1 0.432 0.997 0.894 73.0 265 25 32 9 29.8 1 0.432 0.997 0.894 73.0 258 25 32 10 0.384 1 0.432 0.997 0.894 7%.0 21.4 25 32 11 2.83 1 0.432 0.997 0.894 7%.0 46.0 25 22 12 1.59 1 0.432 0.997 0.894 73.0 33,1 25 22 13 0-380 1 0.432 0.997 0.894 7%.0 31.0 25 %Data obtained during second study. bData obtained during third study. el Tabte 39. lxperimentally Determined Values for the Axia! Dispersion Coefficient in a 3.0-in.-ID Open Bubble Column Gas disperser : one orifice, 0.432 cm ID Properties of Liquid Axial Data Superficial Surface Dispersion Set Run Gas Velocity Density Viscosity Tension Coefficient No. No. (em/sec) (g/cm3) (cP) (dynes/cm) {(cm b - 200 Vg = 0.5 cm/sec 00 - 30 80 - T 60} — S50 - AXIAL DISPERSION COEFFICIENT (cmsec) 40} . 30 - 10 1 1 1 ] 1 1 0 I.O 2.0 3.0 4.0 5.0 6.0 COLUMN DIAMETER {in)) Fig. 56. Variation of Axial Dispersion Coefficient with Changes in Column Diameter and Superficial Gas Velocity. 145 Although the transition from bubble to slug flow in the present studies did not occur at a single value of the superficial gas velocity, certain qualitative conclusions can be inferred. As the column diameter is increased, the superficial gas velocity at which slug flow is obtained increases rapidly. Slug flow was obtained in the l.5-, 2-, and 3-in. columns at superficial gas velocities of 0.005, 1.14, and 5.54 cm/sec, respectively, for bubble columns filled with water; how- ever, slug flow was not obtained in the 6-in.-diam column at velocities as high as 9.5 cm/sec. 9.8.3 Effect of Viscosity of the Liquid Phase The variation of the axial dispersion coefficient with changes in the viscosity of the liquid and the superficial gas velocity is shown in Figs. 4k, 46, 47, 53, and 54 for columns having diameters of 1.5, 2, 3, and 6 in. 1In each case, no observable change was noted in the axial dispersion coefficient as the viscosity of the liquid was increased from 0.9 cP to about 2 cP. A further increase in viscosity to about 11 ¢P resulted in a decrease of about 20% in the axial dis- persion coefficient for the smaller columns (1.5 and 2 in. in diameter), while the effect was less pronounced for the 3-in.-diam column. How- ever, increasing the viscosity of the liquid by as much as a factor of 12 produced no observable change in the dispersion coefficient for the 6-in.-diam column. Although the axial dispersion coefficient values for the more viscous solutions were lower, the observed dependence of the axial dispersion coefficient on the superficial gas velocity was the same as that observed for columns filled with water. These obser- vations are in agreement with the correlations shown in Eqgs. (37) and (38), which predict that the dispersion coefficient depends on the viscosity of the liquid phase to the -0.057 and -0.072 powers for bubble and slug flow respectively. The variation of gas holddp and bubble rise velocity with changes in the viscosity of the liquid is shown in Figs. 25-32 for 1.5-, 2-, 3-, and 6-in.-diam columns. An increase in the viscosity of the liquid results in a slight increase in the gas holdup for the smaller-diameter 146 columns. However, the gas holdup values for the larger columns appear to be unaffected by changes in viscosity. The relation used for corre- lating the holdup data [Eq. (33)] shows no dependence of holdup on viscosity of the liquid phase. 9.8.4 Effect of Surface Tension of the Liquid Phase The effects of changes in the surface tension of the liquid on the axial dispersion coefficient and the gas holdup in a 1.5-in.-diam column are shown in Figs. 45 and 33 respectively. A decrease in the surface tension of the liquid from about 72 dynes/cm to about 27 dynes/cm resulted in a 25% decrease in the dispersion coefficient during slug flow; how- ever, essentially no change was observed during bubble flow. A slight increase in the gas holdup was noted as the surface tension was decreased for a given value of the superficial gas velocity. The correlations for the axial dispersion data indicate no dependence of dispersion coeffi- cient on surface tension during bubble flow and a 0.43-power dependence during slug flow. The correlation for gas holdup indicates that holdup is not affected by changes in surface tension. Decreases in the sur- face tension of the liquid delayed the change from transition flow to slug flow; that is, the change in flow regime occurred at progressively higher values of the superficial gas velocity as the surface tension of the liquid was decreased. 9.8.5 Effect of Gas Inlet Orifice Diameter The variation of the axial dispersion coefficient with a change in the gas inlet diameter is shown in Fig. 49 for a 2-in.-diam column filled with water. The gas distributor consisted of a single orifice having a diameter of either 0.1 or 0.432 cm. Only a minor effect at low superficial gas velocities was noted. The effect of the diameter of the gas inlet orifice on the axial dispersion coefficient in a 2-in.-diam column employing a gas distribu- tor consisting of five orifices is shown in Fig. 50. The gas inlet ori- fice diameter was varied from O.4 to 4 mm, which corresponds to a range of 13 to 127 for the column diameter/orifice diameter ratio. For gas 1h7 distributors containing multiple orifices, the axial dispersion coef- ficient values in the transition flow regime exhibit an anomalous behavior, which is probably due to complex bubble-bubble interaction effects. A similar behavior was observed by two other 3’.nvestJ'.gatc'rs,56’57 who considered it to be a hindered mixing effect caused by bubble coalescence. In the slug flow regime, the axial dispersion coefficient values show no dependence on gas inlet orifice diameter. This is to be expected since the large bubbles or slugs formed by coalescence of smaller bubbles are independent of the initial bubble population. 1In the bubble flow regime, however, the smallest orifice diameter (0.4 mm) produced axial dispersion coefficient values that were significantly lower than those produced by the other, larger orifices. The axial dis- persion coefficient values for the larger orifices were essentially the same (within the accuracy of the data). The gas holdup values for the 1.0-, 2.0-, and L4.0-mm-diam orifices were identical, while the values obtained by using a O.L-mm-diam orifice were higher. The higher values imply a lower bubble rise velocity. Thus, except in the case of very small orifice diameters, the dispersion coefficient is independent of orifice diamster,.as is assumed by the dispersion coefficient corre- lations. Data were also obtained with gas distributor plates containing 37 orifices. No effect on axial dispersion coefficient was observed (within the accuracy of the data) when the orifice diameter was varied from 1 to 4 mm, as shown in Fig. 51. Reith58 found no effect ef orifice diameter on axial dispersion during slug flow; during bubble flow, the axial dispersion appeared to be marginally affected, although no quanti- tative dependence was reported. For column-to-orifice diameter ratios less than 80, Ohki35 observed a very small influence of orifice diameter on axial dispersion. However, for higher values of this ratio, a marked decrease in the dispersion coefficient was observed. 148 9.8.6 Effect of Number of Orifices in Gas Distributor The effect of the number of orifices in the gas distributor on the axial dispersion coefficient is shown in Fig. 52 for a 2-in.-diam bubble column filled with water. The orifice diameter was 1 mm in each case. In the bubble flow regime, the axial dispersion coefficient increased as the number of orifices in the gas distributor was increased from 1 to 37. 1In the transition flow regime, no clear dependence on number of orifices is apparent. In the slug flow regime, the axial dispersion coefficient values are essentially independent of the number of orifices in the gas distributor. This is to be expected since, for a given super- ficial gas velocity, the number of orifices will only alter the initial size and number of bubbles, and the effects of these quantities are eliminated by coalescence and slug formation. The correlations of the dispersion coefficient data indicate that the dispersion coefficient is weakly dependent (0.098 power ) on the number of orifices in the gas distributor during bubble flow and independent of the number of orifices during slug flow. 9.9 Future Work It is believed that sufficient experimental data on gas holdup and axial dispersion have been obtained; hence no further experimental work in this area is planned. In the future, efforts will be made to develop relations that correlate the data on gas holdup satisfactorily and also extrapolate well to known limiting cases, such as the rate of rise of individual bubbles in vessels having very large diameters. Attempts will also be made to correlate the dispersion data obtained during slug flow without including gas holdup dependence since such a correlation would probably be much simpler. 1k9 10. SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL FACILITY B. A. Hannaford C. W, Kee L. E. McNeese We have continued operation of a facility in which semicontinuous reductive extraction experiments can be carried out in a mild-steel 39 system. Initial work with the facility was directed toward obtaining data on the hydrodynamics of the countercurrent flow of molten salt and bismuth in a 0.82-in.-ID, 24-in.-long column packed with 1/L-in. molybdenum Raschig rings. We have been able to show that flooding data obtained with this column are in agreement with predictions from a correlationbrO based on studies of the countercurrent flow of mercury and aqueous solutions in packed columns. We have carried out several experiments for determining the mass transfer performance of the packed column in which a salt stream containing UFu was countercurrently con- tacted with bismuth containing reductant over a range of operating conditions. During the second uranium mass transfer run (UTR-2), 95% of the uranium was extracted from the salt during a 40-min period in which the bismuth and salt were countercurrently contacted at flow rates of 247 and 52 cmj/min respectively.hl Two additional uranium mass transfer runs (UTR-3 and -4) were carried out under conditions such that the uranium extraction factor remained high throughout the column. The fraction of uranium extracted from the salt phase increased from 0.63 to 0.91 as the metal-to-salt flow rate ratio was varied from 0.75 to 2.05. It was found that the rate at which uranium transferred to the bismuth was controlled by the diffusive resistance in the salt film, and that the extraction data could be correlated in terms of the height of an overall transfer unit based on the salt phase. The HTU values ranged from O.77 to 2.1 ft as the bismuth-to-salt flow rate ratio was decreased from 2.05 to 0.75.h2 In order to measure mass transfer rates in the column under more closely controlled conditions and under condi- tions where the controlling resistance is not necessarily in the salt phase, preparations were begun for experiments in which the rate of exchange of zirconium isotopes will be measured between salt and bismuth 150 phases otherwise at chemical equilibrium. In preparation for these experiments, reductant was removed from most of the bismuth by hydro- fluorination of the salt and bismuth with a 70-30 mole % HE—HF mixture. After hydrofluorination, the salt and bismuth were transferred through the system (during run UTR-5) in order to obtain uniform concentrations in the salt and bismuth phases throughout the system. There was a sur- prisingly large variation (155%) in the reported concentrations of b3 uranium in salt samples removed from the column effluent during the run. 10.1 Run UTR-6 The objectives of run UTR-6 were: (1) to obtain additional informa- tion relative to the variation in the reported uranium concentrations in samples taken from the salt stream leaving the extraction column, because of the large variation (+35%) in these values during run UTR-5; and (2) to obtain additional hydrodynamic data on the countercurrent flow of salt and bismuth in the packed column by operating under conditions where flooding was predicted by the correlation based on countercurrent flow of mercury and water in packed columns. Following run UTR-5, the salt and bismuth were returned to the treatment vessel, from which the salt and bismuth phases were transferred to their respective feed tanks. After the transfers, the bismuth feed tank contained about 17 liters of bismuth in which the thorium concentration was about 6 ppm. The salt feed tank contained about 15 liters of salt having a uranium concentra- tion (as UFM) of about 3100 ppm. Essentially no extraction of uranium from the salt during the run would be expected because of the negligible concentration of reductant in the bismuth. As shown in Fig. 57 the run was initiated by starting a salt flow through the column; then, a few minutes later, a bismuth flow was begun. Constant salt and bismuth flow rates of 125 and 117 cmi/min, respectively, were attained after a short time, and seven pairs of bismuth and salt samples were withdrawn from the salt and bismuth streams leaving the column. Subsequently, the flows of salt and bismuth to the column were stopped, and the freeze valve below the specific gravity pot was closed by cooling the line to a temperature below the salt liquidus temperature. The run was then REMAINING () FEED ORNL DWG.73-48 20 IvvivllTIllllrr'trIITTTTlIlIt!TT1rlllirrvlvtrllrillllltr1ll"'llillrv1111l1I1I111IVII[vr1ll1llv'1l1|||lt1|llluvlllllrvtwvl||1I111l|1]11]1||1|lrl]_ + + + + + o+ o+ + + + + + + 4+ SAMPLE PAIRS TAKEN: I 2 3 4 5 6 7 8 9 10 11 12 13 14 3 I8 H7 cm3 l BISMUTH /min BISMUTH ————— SALT S ~ S 14 s 12 209 cm3 0 ~__ l BISMUTH/min I25cm:5 8 e SALT 7/ min '] 6 4 2 O 0 10 20 30 40 50 60 70 80 90 100 1o (20 130 140 TIME (min) Fig. 57. Volume of Bismuth and Salt Remaining in Feed Tanks vs Run Time, Run UTR-6. Volumetric flow rate (ml/min) for each indicated inter- val was inferred from the slope. L unlluluunnllll lllll|lluu_1un|1ulluu_l_uu IlllllIllllllllluAlllllllllllll_l_lllll 233 o O 16T 152 resumed with salt and bismuth flow rates of 150 and 209 cmB/min. This combination of salt and bismuth flow rates was selected in order to further test the prediction of flooding by the correlation based on countercurrent flow of mercury and aqueous solutions in packed columns (see Sect. 1k). Seven additional salt samples were taken during the remainder of the run for bismuth analyses. Shortly after the salt and bismuth flows were resumed, the apparent dispersed-phase holdup (as indicated by column pressure drop) stabilized at 30%. The salt flow rate was then increased slightly (to 151 cm3/min). This caused the apparent dispersed-phase holdup to increase until it reached a value of about 60%, at which point the specific gravity pot began to fill with bismuth (which clearly indicated that the column was flooded ). The conditions under which the column flooded agree reasonably well with the values predicted from the correlation since the ratio of the observed column throughput to the predicted throughput at flooding was 1.13. Results of the analyses of the salt and bismuth samples taken during the first part of the run are shown in Table 4l. Little signif- icance is attached to the analyses of the bismuth phase; the indicated increase in the concentration of reduced metals is probably the result of slight contamination of the samples with salt or may simply represent analytical errors. The uranium concentration in the salt samples should have remained constant at a value of 3100 ppm; nevertheless, the data show considerable scatter. The average of the uranium concentrations in the flowing salt samples is only 2714 ppm; however, the uranium content of the salt sample removed from the receiver tank is in agree- ment with the expected value. The variation of uranium concentrations in samples removed from the salt stream during this run is similar to that observed during run UTR-5; on the other hand, a much smaller variation in reported uranium concentrations was observed during an b1 earlier, similar run (UTR-1). Bismuth concentrations in the salt samples taken during the second half of run UTR-6 are given in Table L42. These concentrations are relatively low, ranging from 0.8 to 15 ppm. It should be noted that 153 Table 41. Summary of Results for Samples Taken During Run UTR-6 Materials in Uranium Conc. Sample Source Bismuth Samples (ppm) in Salt Li Th U (ppm) Feed tanks 0.052 <2 <1 3100 0.42 6 1 3200 Flowing stream 1 0,060 9 2.45 2700 2 0.046 15 <1 3200 3 0.046 5 <1 2400 4 0.055 14 <1 3100 5 0.042 4 <1 2300 6 0.035 6 f this contacting period, no circu- lation c¢f the LiCl was allowed to occur. The relative concentration of the transferring material in the fluoride salt after a period in which the LiCl was circulated is given by Bq. (45), and the relative concen- tration of the transferring material after a period during which no circulation occurred is given by the relation: s | (At ), (46) XFS,p s where XFS,S = concentration of the transferring material in the fluoride salt at time ts’ mole fraction, tS = length of time during which no LiCl circulation occurs, 5ec. Combination of Eqs. (45) and (46) yields the following relation for the concentration of the transferring material in the fluoride salt after one cycle of operation in which the LiCl is circulated for part of the time period and no circulation occurs during the remainder of the period: xFS =225 - exp [—K(t + t )] exp (-Kt ), (47) Xs, i > P b 2 h where DF chf - K = ¢ DF Mpg + MppPp * Mgg D, Similarly, the relative concentration of the transferable material in the fluoride salt after n cycles in which both pumping and equilibra- tion occur is given by the following expression: 173 XFS sn [ = ] - - = exp | -A z: (t . + t .) exp ('Kl E: t -) » (AB) Xps, 1 j=1 83 PJ j=1 PJ where xFS,sn = concentration of the transferable material in the fluoride salt after n cycles in which LiCl circulation and equili-~ bration occur, mole fraction, tsj = length of time during which no LiCl is circulated during cycle j, sec, tpj = length of time during which LiCl is circulated during cycle j, sec. It should be noted, however, that the two time summations in Eq. (48) have the following meanings: n t_ = .Z—Zl(tsj tr), (49) _]__. ) t = t ., (50) pn j=1 Pl where tn = length of time between the beginning of the experiment and the end of the nth cycle, sec, tpn = length of time that LiCl circulation has occurred between the beginning of the experiment and the end of the nth cycle, sec. Substitution of Eqs. (49) and (50) in Eq. (48) yields the following rela- tion for the relative concentration of the transferable material in the fluoride salt at the end of the nth cycle: XFS n —22°20 -~ exp (-A tn) exp (-Kt_ ) . (51) Xps, 1 pn This relation indicates that a plot of the quantity §—§L§E exp (htn) Vs FS,1i the length of time that the LiCl has been circulated should yield a straight line having a slope of -K. The quantity f can be calculated after the value of K has been determined from a plot of this type. 17k 11.3 Experimental Results 11.3.1 Rates of Transfer of Neodymium and Lanthanum During the experiment, the lanthanum and 147Nd that were added to the fluoride salt initially behaved as expected. The rates of accumu- lation of these two rare earths in the Li-Bi solution during the experi- ment are summarized in Table 43 and shown in Fig. 64. There was essentially no accumulation of either of the rare earths in the solution prior to operation of a gas-lift sparge tube in the cup (about 50 liters of LiCl was circulated through the Li-Bi compartment before operation of the sparge tube was initiated). After 400 liters of LiCl had been circulated (about two-thirds through the run), approximately 50% of the lanthanum and approximately 30Z of the neodymium originally in fluoride salt were found to be in the Li-Bi solution. During the last one~third of the run the rare earths continued to accumulate ih the Li-Bi solution, but the rate of accumulation could not be determined accurately because of a leak that developed in the lithium-bismuth cup. This leak allowed about 30% of the solution to flow into the area between the cup and the holder. The extent of removal of lanthanum and neodymium from the fluoride salt is summarized in Tables 44 and 45 and shown in Figs. 65 and 66. As seen in the figures, more than 857 of the lanthanum and more than 50% of the neodymium had been removed from the fluoride salt at the end of the experiment. After about 50 liters of LiCl had been circulated through the cup containing the Li-Bi solution, the concentrations of lanthanum and neodymium decreased in the manner suggested by Eq. (51). There is some uncertainty in the f wvalues that correspond to the lines shown in the figures because of uncertainty in the values for the distribution coefficients, as will be discussed later. The concentration of thorium in the Li-Bi solution in which the rare earths were accumulated remained below the limit of detection throughout the experiment (< 10 ppm). Data on the distribution of lanthanum and neodymium between the fluoride salt, the thorium-saturated bismuth, and the LiCl are summarized in Tables 46 and 47. The distribution coefficient data were generally Table 43. Rate of Accumulation of La and Nd in the Li-Bi Solution During Metal Transfer Experiment MTE-2 Volumé of Volume of Pumping LiCl La Nd Pumping LiCl La Nd Time Circulated Concentration Concentration® Time Circulated Concentration Concentratiop® (hr) (liters) (mg/g) (dpm/g) x 107 {hr) (liters) (mg/g) (dpm/g) x 107 0 0 0.0 0.0 108.2 173.8 - 0.149 6 9.6 0.005 0.0 124.2 195.3 4.46 0.180 12 19.6 0.023 0.001 137.6 211.2 4.91 0.200 18 25.1 0.058 -0.003 150.1 226.5 4.80 0.199 24 39.0 0.092 - 156.1 233.4 5.42 0.214 27 44.0 - 0.003 169.1 250.7 6.05 0.202 30 48.8 0.11 0.004 185.3 273.9 5.97 0.225 36 59.0 0.82 0.037 212.6 319.6 6.25 - 44.5 73.2 1.40 0.050 219.2 330.5 6.75 - 47.5 77.9 1.66 0.067 250.5 384.3 7.17 0.290 53.5 87.6 2.26 0.083 273.5 421.0 - 0.345 59.5 97.1 2.70 0.089 302.7 471.2 - 0.568 65.5 107.0 2.67 0.106 309.7 482.8 10.53 0.584 94.2 153.2 3.74 0.131 343.9 538.8 10.31 - 101.3 164.0 4.27 0.129 358.1 563.2 11.17 - a . . . Corrected for radiocactive decay during the run. QLT PERCENT OF TOTAL IN Li-Bi 80 70 60 4 o & W o 20 ORNL DWG TI-TOR1 METAL TRANSFER EXPERIMENT-MTE-2 -~ L. XV i 7 i XV /008 ] Y | ! T X—~LANTHANUM ! ! ' ' ' [ ' O-NEODYMIUM X alT X" x o O ooooooo o ~0 ©Oo _ L x/x&flflp ] *° Lx 00X 1 I 1 l 1 l i I 1 i00 200 300 400 500 600 VOLUME OF LiCi PUMPED (LITERS) Fig. 64. Rates of Accumulation of Lanthanum and Neodymium in the Li-Bi Solution During Metal Transfer Experiment MTE-2. Table 44. Concentration of Lanthanum in Fluoride Salt During Metal Transfer Experiment MTE-2 Pumping Volume of LiCl La Concentration Pumping Volume of LiCl La Concentration Time Circulated in Fluoride Salt Time Circulated in Fluoride Salt (hr) (liters) (mg/g) (hr) (liters) (mg/g) 0 0 6.54 124.2 -195.3 4.14 6 9.6 6.15 137.6 211.2 3.70 12 19.6 5.78 150.1 226.5 3.93 18 29.1 6.26 156.1 233.4 2.96 27 44.0 6.33 169.1 250.7 3.40 30 48.8 6.20 185.3 273.9 3.09 36 59.0 6.72 212.6 319.6 2.04 44 .5 73.2 5.54 219.2 330.5 1.94 47.5 77.9 5.63 250.5 384.3 1.42 53.5 87.6 5.75 273.5 421.0 1.38 59.5 97.1 5.65 309.7 482.8 1.17 65.5 107.0 5.30 343.9 538.8 1.10 94,2 153.2 4.31 358.1 563.2 0.86 101.3 164.0 4.84 LLT Table 45. Concentration of Neodymium in Fluoride Salt During Metal Transfer Experiment MTE-2 Pumping Volume of LiCl Nd Concentration? Pumping Volume of LiCl Nd Concentration? Time Circulated in Fluoride Salt Time Circulated in Fluoride Salt (hr) (liters) (dpm/g ) x 10T (hr) (liters) (dpm/g) x 107 0 0 0.508 80.0 130.4 0.422 3 h.7 0.545 g4.2 153.2 0.390 6 9.6 0.L87 101.3 164.0 0.387 9 4.7 0.531 108.2 173.8 0.383 12 19.6 0.474 114.9 183.5 0.358 15 2h.5 0.473 124.2 195.3 0.397 18 29.1 0.458 130.7 202.3 0.329 21 3%.9 0.458 137.6 211.2 0.313 27 Lh.o 0.492 143.5 218.2 0.323 30 L8.8 0.463 150.1 226.5 0.347 33 5h.1 0.450 156.1 233.4 0.289 36 53.0 0.517 162.6 241.2 0.310 hh.s 73.2 0.493 169.1 250.7 0.268 LT.5 T7-9 0.LT77 175.9 258.5 0.300 50.5 82.7 0.471 185.3 27%.9 0.315 53.5 87.6 0.478 250.5 384.3 0.278 56.5 92.3 0.458 27%.5 Lh21.0 0.251 59.5 97.1 0.441 302.7 Lh71.2 0.114 62.5 102.1 0.469 309.7 482.8 0.138 65.5 107.0 0.43L QLT a Corrected for radioactive decay during the run. 179 ORNL. DWG Ti-63R1 METAL TRANSFER EXPERIMENT-MTE-2 1 I | I [ 1 [ T T T T T - — o o z =z —4 < = o - @ =z - =2 z ! e - - = o - & z © - - Q < x W X ¥ 0.1 1 l i [ L l L 1 1 ] 1 0 100 200 300 400 500 600 VOLUME OF LiCl PUMPED (LITERS) Fig. 65. Variation of Lanthanum Inventory in the Fluoride Salt Surge Tank with Time During Metal Transfer Experiment MTE-Z2. 1.0 0.9 o8 0.7 0.6 0.5 0.4 03 FRACTION OF NEODYMIUM REMAINING (CORRECTED FOR DECAY) Are T [ T [ T I T I T l o T $ ;:/‘ o1 7 180 ORNL DWG 7i-64R1 METAL TRANSFER EXPERIMENT-MTE-2 1 I L l 1 I 1 I 1 0.2 o 100 200 300 400 500 VOLUME OF LiCl PUMPED (LITERS) Fig. 66. Variation of the Neodymium Inventory in the Fluoride Salt Surge Tank with Time During Metal Transfer Experiment MTE-2. Table 46. 181 Variation of Lanthanum Distribution Coefficient with Time During Metal Transfer Experiment MTE-2 Volume of Distribution Distribution Pumping LiCl Coefficient at Coefficient at Time Circulated Fluoride Salt-- LiCl--Bi~Th (hr) (1iters) Bi-Th Interface Interface 0 0 0.045 4,54 6 9.6 0.046 2.78 12 19.6 0.042 5.97 18 29.1 0.043 2.04 27 44,0 0.052 - 30 48.8 0.039 3.67 36 59.0 0.034 4.49 44,5 73.2 0.044 9.28 47 .5 77.9 0.038 2,18 53.5 87.6 0.039 1.84 59.5 97.1 0.023 1.17 65.5 107.0 0.034 0.93 94.2 153.2 0.056 2.07 124.2 195.3 0.055 4.17 124.2 195.3 0.078 3.51 137.6 211.2 0.045 1.88 150.1 226.5 0.045 2,00 156.1 233.4 0.055 2.45 169.1 250.7 0.055 0.93 185.3 273.9 0.054 - 212.6 319.6 0.071 7.54 219.2 330.5 0.083 - 250.5 384.3 0.058 3.14 273.5 421.0 0.056 2.69 309.7 482.8 0.046 2.24 343.9 538.8 0.036 2.32 358.1 563.2 0.048 2,27 358.1 563.2 0.051 3.56 Table 47. 182 Variation of Neodymium Distribution Coefficient with Time During Metal Transfer Experiment MTE-2 Volume of Distribution Distribution Pumping LiCl Coefficient at Coefficient at Time Circulated Fluoride Salt-—- LiCl1--Bi-Th (hr) (liters) Bi-Th Interface Interface 0 0 0.066 2.88 3 4,7 0.055 6.98 6 9.6 0.064 4,94 9 14.7 0.062 4.56 12 19.6 0.062 4.54 15 24.5 0.064 4.83 18 29.1 0.067 4,12 21 33.9 0.078 5.10 27 44,0 0.051 3.02 30 48.8 0.062 3.39 33 54.1 0.063 2.60 36 59.0 0.063 3.72 44.5 73.2 0.052 6.00 47.5 77.9 0.058 5.73 50.5 82.7 0.056 10.59 53.5 87.6 0.059 6.50 56.5 92.3 0.058 4.85 59.5 97.1 0.057 4,17 62.5 102.1 0.054 4,89 65.5 107.0 0.058 3.49 80.0 130.4 0.069 5.96 94.2 153.2 0.063 - 101.3 164.0 0.060 - 108.2 173.8 0.071 4.52 114.9 183.5 0.071 4.42 124.2 195.3 0.063 5.76 130.7 202.3 0.083 —— 137.6 211.2 0.093 — 143.5 218.2 0.085 - 150.1 226.5 0.048 - 156.1 233.4 0.076 —— 162.1 241.2 0.082 — 169.1 250.7 0.092 - 175.9 258.5 0.070 - 185.3 273.9 0.058 —— 185.3 273.9 0.069 - 250.5 384.3 0.052 —_ 250.5 384.3 0.038 - 183 obtained from samples taken after circulation of the LiCl had been stopped and the phases had been allowed to approach chemical equilibrium for four or more hours. The variation of the coefficient for the distribution of lanthanum between the fluoride salt and the Th-Bi phase during the experi- ment is shown in Fig. 67. Although there is some variation in the distri- bution coefficient values, the average value (0.05) is relatively close to the value of 0.06 that was calculated from previously reported equilib- rium relations. The variation of the coefficient for the distribution of lanthanum between the LiCl and the thorium-saturated bismuth during the run is shown in Fig. 68. The average value (3.1) is somewhat higher than the value predicted by equilibrium relations reported earlier (i.e., 0.9). Data on the coefficient for the distribution of neodymium between the fluoride salt and the thorium-saturated bismuth during the run are shown in Fig. 69. The average value for the distribution coefficient (0.06) is in good agreement with the expected value (0.062). The variation of the coefficient for the distribution of neodymium between the LiCl and the thorium-saturated bismuth during the experiment is shown in Fig. 70. The average value for the distribution coefficient (4.8) is slightly higher than the expected value of 3.5. The uncertainties in the distri- bution coefficient values for lanthanum and neodymium during the experi- ment result in some uncertainties of the values for f, the fractions of the lanthanum or neodymium that are removed from the LiCl during its passage through the cup containing the Li-Bi solution. The values of f for lanthanum and neodymium are 0.27 and 0.21, respectively, based on the average values for the distribution coefficients, and 0.08 and 0.16 based on the predicted values for the distribution coefficients. The variations of the neodymium and lanthanum inventories in the system during the experiment are shown in Fig. 71. From 70 to 100Z of each of the rare earths initially charged to the system could be accounted for, based on filtered samples. This indicates that essentially all of the rare earths remained in solution during the experiment. 13k ORNL DWG Ti-61R2 METAL TRANSFER EXPERIMENT-MTE-2 0.'0 v —r ¥ "— ¥ l T [ L I ¥ 0.09} _ - . - - 0.08 br ° "‘ 2 W 0.07} * - O t " 4 w 006 ° . ] ® AVERAGE=0. o Z 0.05 ERAGE=0.05 - - ° o0 * - 2 °® o ® ® - " o o ) 5 o003} - | . ] 0.02 |- - - ; 0.0l - . o . | L 1 ) ] . 1 L ] . o 100 200 300 400 500 600 VOLUME LiC! PUMPED (LITERS) Fig. 67. Variation of the Coefficient for the Distribution of Lanthanum Between the Fluoride Salt and the Thorium-Saturated Bismuth During Metal Transfer Experiment MIE-2. 185 ORNL DWG T7I-59R1 METAL TRANSFER EXPERIMENT-MTE-2 Io-o T '[ LB I T ' v I L I T 90 - - 80 . 7.0 — 60e - ® ° AVERAGE=3. ® - DISTRIBUTION COEFFICIENT o O = L o 3 l 2 l 3 l i l 3 | 2 0 100 200 300 400 500 600 VOLUME OF LiC! PUMPED (LITERS) Fig. 68. Variation of the Coefficient for the Distribution of Lanthanum Between the LiCl and the Thorium-Saturated Bismuth During Metal Transfer Experiment MTE-Z. 1866 ORNL DWG 7i-62R! METAL TRANSFER EXPERIMENT-MTE-2 0.'0 I = T [ r T ™ T T ® 0.09 . ) ° ® 008 * _ ; . - Z 007 « °° ® - o AVERAGE=0.06 u °% ' w006 o - Q . o ° 0.05 B S - o J 5 3 004 . . £ - v — @ 0.03 _ { 002 | i (lOI[ i, O L 1 . ] L | L i — 1 ) 0 00 200 300 400 500 600 VOLUME OF LiCl PUMPED (LITERS) Fig. 69. Variation of the Coefficient for the Distribution of Neodymium Between the Fluoride Salt and the Thorium-Saturated Bismuth During Metal Transfer Experiment MTE-Z2. DISTRIBUTION COEFFICIENT 1.0 — 100 |- 9.0 | 80| 70 |e 60 |- 50 4.0 |- 187 ORNL DWG T7I-60R1 METAL TRANSFER EXPERIMENT-MTE-2 ] ¥ ] T ® o @ AVERAGE=4.8 1 A 1 Neodynium Between the LiCl and the Thorium-Saturated Bismuth During Fig. T0O. 100 200 VOLUME OF LiCl PUMPED (LITERS) Variation of the Coefficient for the Distribution of Metal Transfer Experiment MTE-2. 300 188 ORNL DWG 7I-7IRI 120 T T - T T T o. X 1I00» ,5!:: . - "d‘.‘ Xe o o .x . X X .o. > X . ® < X 1 e .x ° X X 80 .ox. .! . o . x _ * % . @ X X ® 2 > 60 — a O '_ 4 s Z 40 e NEODYMIUM _ F x LANTHANUM 20 .l —y 0O i ] ] l 1 1 0 100 200 300 400 500 600 700 VOLUME OF LiClI PUMPED (LITERS) Fig. Tl. Variation of the Inventories of Neodymium and Lanthanum During Metal Transfer Experiment MTE-2. 189 Decrease in Concentration of Lithium in the Li-Bi Solution. — During the experiment, the concentration of lithium in the Li-Bi solution decreased from an initial value of 1.8 wt %7 to 0.5 wt Z (0.35 to 0.13 mole fraction), as summarized in Table 48 and shown in Fig. 72. A decrease of about 0.2 wt Z in lithium concentration was expected as the result of extraction of the rare earths into the Li-Bi solution; however, the observed decrease was about six times this value. The reason for the decrease has not been determined at this time; it may have resulted from a slight solubility of lithium or a lithium-bismuth intermetallic in the LiCl or from transfer of lithium ions through the LiCl due to the difference in emf between the two bismuth phases in the system. The decrease in lithium concentration appears to have occurred only while LiCl was being circulated through the Li-Bi cup since no decrease was noted during long periods in which no circulation occurred (e.g., weekends). Distribution of Radium Between the Salt and Metal Phases. — Radium was present in the system as a decay product of thorium, which was a con- stituent of the fluoride salt and of the bismuth phase (thorium-saturated). Data on the distribution of radium in the system during the experiment are summarized in Table 49 and shown in Fig. 73. At the end of the 76-hr equilibration period before circulation of the LiCl was begun, about 607% of the radium had transferred to the portion of the LiCl that was in con- tact with the thorium-saturated bismuth; the remaining 407 was present in the fluoride salt. No radium was present in the Li-Bi solution prior to circulation of the LiCl, and the radium concentration in the thorium- saturated bismuth remained below the limit of detection throughout the run. The radium slowly transferred into the Li-Bi solution as the LiCl was circulated through the Li-Bi cup. When operation of the gas-lift sparge tube in the Li-Bi cup was initiated (after about 50 liters of LiCl had been circulated), the concentration of 228Ra in the Li-Bi solution increased abruptly and the concentration of radium in the LiCl decreased simultaneously. After a short period of time, the concentrations of radium throughout the system appeared to have reached equilibrium values, with about 40% of the radium in the Li-Bi solution, 407 in the LiCl, and 20% in the fluoride salt. During the remainder of the experiment, the radium slowly transferred out of the Li-Bi solution as the concentration of lithium in the solution decreased. Table 48. Variation of Lithium Concentration in the Li-Bi Solution During Metal Transfer Experiment MTE-2 Pumping Volume of LiCl Lithium Conc. Pumping Volume of LiCl Lithium Conc. Time Circulated in Li-Bi Time Circulated in Li-Bi (hr) (liters) (wt %) (hr) (liters ) (wt %) 0 0 1.80 137.6 211.2 1.29 6 9.6 1.94 150.1 226.5 1.20 12 19.6 L.76 156.1 233 .4 1.23 18 29.1 1.76 169.1 250.7 1.13 24 39.0 1.81 185.3 273.9 1.08 30 48.8 1.70 212.6 319.6 1.03 i .5 73.2 1.72 219.2 330.5 1.03 47.5 77.9 1.62 250.5 38L4.3 0.95 53-5 87.6 1.52 273.5 421.0 0.89 59.5 97.1 1.63 309.7 482.8 0.83 65.5 107.0 1.56 343.9 538.8 0.80 94 .2 15%.2 1.45 358.1 563%.2 0.76 124.2 195.3 1.35 433.6 690.2 0.5 06T (wt %) LITHIUM CONCENTRATION IN Li-Bi 0.5 ORNL DWG 71-65R2 I I T 1 T | TST I ! I 1 | 100 200 300 400 500 600 700 VOLUME OF LiCl PUMPED (liters) Fig. 72. Variation of Lithium Concentration in the Li-Bi Solution with Time During Metal Transfer Experiment MIE-Z. Tabie A47). Variation of the Distribution of Radium with Time During Metal Transfer Experiment MIE-2 Volume of Radium Cone . Radiumn Radium Volume of Radium Conc. Radium Radium Pumping LicCl in Fluoride Conc. Conc . Pumping LiCl in Fluoride Conc - Conc. Time Circulated Salt in LiC1 in Li-Bi Time Circulated Salt in LicCl in Li-Bi (hr) (liters) (cpm/g )® (cpm/g )® (cpm/g )? (hr) (liters ) (cpm/g )? (cpm/g )? (cpm/g )? 0 0 557 1565 0 143 .5 218.2 166 1028 901 3 .7 ugy 1500 103 150.1 £226.5 152 1107 96U 6 §.6 hs1 1410 216 156.1 233,14 136 1103 885 9 4.7 L57 -- 284 162.6 2hk1.2 154 1010 856 12 19.6 koo 1511 L6 175.9 258.5 97 1241 997 15 2o 348 1511 384 185.3 273%.9 115 -- 907 18 29.1 Loy 1540 508 250.5 384.,3 182.3 898 771 21 3%.9 549 1506 51% 250.5 28,3 -- 1372P -- 24 39,0 -- 1361 621 250.5 384 .3 187.4 1846 97T 27 Lh.0 265 1310 617 250.5 z8l.3 - 1309P -- 30 L8.8 318 1508 -- 273.5 421.0 63 1356 733 33 54.1 316 1571 692 273.5 L21.0 52 1177 700 36 59.0 275 1029 1100 302.7 471.2 -- 1103 71 %9 65.9 -- 907 1057 309.7 L82.8 -- 1176 628 bh.s 73,2 278 950 1235 %09.7 482.8 -- 1135P -- hT.5 7.9 96 966 1261 34%.9 538.8 -- 1296 684 50.5 2.7 197 1588 1184 343.9 538.8 -- 1362P -- 53.5 87.6 188 824 121% 343.9 538.8 -- LL86 665 56.5 2.3 200 916 115k 343.0 538.8 -- 1h19P -- 59.5 971 184 701 1151 358.1 56%.2 -- 1311 640 62.5 102.1 226 o7 1117 358.1 563.2 -- 1676P - 65.5 107-0 op 850 1160 358.1 563.2 -- 1463 696 80.0 130. 4 183 10%2 1286 358.1 563%.2 -- 15700 -- k.5 153.2 15% 101y 1197 591.1 619.1 -- 1367 51t 108.2 17%.8 e 106k 108L 351.1 615.1 -- 1318P -- 114.9 183. 5 5 1186 1134 413.3 656.0 -- 1549°¢ Ls7 124.2 195.% g2 1220 1097 h13.% 656.0 -- 1739¢ -- 130.7 202. % 80 1117 112% §1%.5% 656.0 -- 1424€ -- 137.6 211.2 -~ 1069 1029 hio.2 665.2 -- 1378¢ 498 261 a Counts per minute per gram. b Samples taken from Li-Bi alloy container. “After 1 vol % fluoride salt was added to LiCl. PERCENT OF TOTAL RADIUM 100 ORNL DWG 71-!'72RlI 90 |- 80 |- 1 I ! A Li-Bi o LicCl X FLUORIDE SALT Negligible amount of Ra in Bi-Th Sy o X nl\x X X X X x X % oL | X i 1 1 0 100 200 300 400 VOLUME OF LiCl PUMPED (liters) Fig. 73. Variation of the Distribution of Radium with Time in the Salt and Metal Phases in Metal Transfer Experiment MTE-Z2. 500 €61 194 Addition of Fluoride Salt to the LiCl. — During operation of the metal transfer process, it is possible that small quantities of fluoride salt will be transferred to the LiCl by entrainment of salt in the bismuth that contacts these salt streams. It has been determined previously that the thorium distribution ratio between the LiCl and the bismuth is quite sensitive to the concentration of fluoride in the LiCl, and hence that the thorium--rare-earth separation factor in the process decreases markedly as the concentration of fluoride in the LiCl is increased. Throughout the experiment, very little fluoride was transferred to the LiCl, as indicated in Table 50; a final fluoride concentration of 1174 ppm was noted. Eight days before the end of the experiment, 1 vol % fluoride salt (72-16-12 mole % LiFwBer—ThFa) was added to the LiCl in order to simulate entrainment of fluoride salt in the circulating bismuth. The concentrations of beryllium, thorium, and fluoride in the LiCl were deter- mined periodically after the addition of the fuel carrier salt, as shown in Table 51. During the 93-hr period in which the LiCl was not circulated through the Li-Bi container, the beryllium concentration remained constant at the initial value of 490 ppm and the thorium concentration decreased, as expected, from the initial value of 9480 ppm to a value of 644 ppm because of transfer of thorium into the Th-Bi solution. When circulation of the LiCl was resumed, the beryllium concentration in the LiCl began to decrease, probably because of reduction of Bez+ by the Li-Bi solution. After 27 hr of LiCl cifculation, a2 beryllium concentration of 135 was observed in the LiCl. The thorium concentration in the LiCl was 171 ppm at this time. 195 Table 50. Variation of the Fluoride Concentration in the LiCl and the Chloride Concentration in the Fluoride Salt During Metal Transfer Experiment MTE-2 Volume of Chloride Fluoride Pumping LiCl Concentration Concentration Time Circulated in Fluoride Salt in LiCl (hr) (liters) (ppm) (ppm) 0 0 - 77 9 14.7 175 530 56.5 92.3 121 - 101.3 164.0 - 517 273.5 421.0 392 205 413.3 656.0 — 1174 Table 51. Variation of Beryllium, Thorium, and Fluoride Concentrations in the LiCl After Addition of 1 vol 7 Fuel Carrier Salt Time After Pumping Addition Time Fluoride Beryllium Thorium (hr) (hr) (wt %) (ppm). (ppm) 0 0 0.982 4902 94802 19.8 0 1.49 490 7200 45.4 0 1.08 500 1100 69.7 0 1.81 320 —— 93 0 2.03 490 644 115.9 5.9 2.34 470 - 163.9 20.3 1.07 120 1400 188.8 27 .4 0.91 - 350 b b b 188.8 27 .4 1.7 135 171 aCalculated values based on the amount of fluoride salt added. bSalt taken from vessel after experiment was concluded. 196 12. DEVELOPMENT OF THE METAL TRANSFER PROCESS: DESIGN OF EXPERIMENT MTE-3 L. E. McNeese E. L. Nicholson W. F. Schaffer, Jr. E. L. Youngblood H. 0. Weeren It has been found that rare earths distribute selectively into molten LiCl from bismuth solutions containing rare earths and thorium, and an improved rare-earth removal process based on this observation has been devised. We are currently engaged in experiments designed to demonstrate all phases of the improved rare-earth removal method, which is known as the metal transfer process. In a previous engineering 46 experiment (MTE-1), we studied the removal of rare earths from single- fluid MSBR fuel salt by this process. During the experiment, approxi- mately 50Z of the lanthanum and 257 of the neodymium originally present in fluoride salt were removed at about the predicted rate. Surprisingly, however, the lanthanum and neodymium that were removed from the fluoride salt did not accumulate in the Li-Bi solution used for removing these materials from the LiCl. Reaction of impurities in the system with the rare earths is believed to have caused this unexpected behavior. A second engineering experiment (MTE-2) has been designed and put into operation (see Sect. 1l). The design of the third engineering experi- ment is currently under way. The third experiment (MTE-3) will use salt and bismuth flow rates that are about 1% of the estimated flow rates required for processing a 1000-MW(e) reactor. In the two previous experiments, the salt and bismuth phases were only slightly agitated, resulting in a low rate of transfer of rare earths from the fuel carrier salt to the Li-Bi solution. In experiment MITE-3, the salt and metal phases will be mechanically agitated in order to increase the rate of transfer of materials between the phases. In the remainder of this section, results from a mathematical analysis carried out for estimating the performance of the system are presented and the experiment is briefly described. 197 12.1 Mathematical Analysis of Metal Transfer Experiment MTE-3 A mathematical analysis of experiment MTE-3 was performed in order to determine the approximate operating conditions for the system and to aid in setting values for parameters such as flow rates, solution volumes, etc. In considering the conceptual design for the experiment, it was concluded that the salt-metal contactor should be of the two-compartmented, stirred-interface type. As pointed out in the following section of this report (Sect. 13), such a contactor is of interest because, with it, adequate mass transfer rates can apparently be achieved without dispersal of either the salt or bismuth phases. Contact of the salt and metal streams without dispersion of the phases should considerably diminish the problem of entrainment of bismuth in the processed fuel carrier salt and the subsequent transfer of bismuth to the reactor, which is con- structed of a nickel-base alloy that is subject to damage by metallic bismuth. The bismuth in a salt-metal contactor of the type being considered would be a near-isothermal, internally circulated phase, which is a desirable condition. Also, it is believed that a processing system employing this type of contactor may be more easily fabricated from molybdenum and graphite than one employing packed columns. The major equipment items considered in the mathematical analysis are shown in Fig. 74. Fuel carrier salt (72-16-12 mole Z% LiFmBer-ThF4) con- taining rare-earth fluorides would be circulated between the fluoride salt surge tank and the side of the salt-metal contactor containing fluoride salt. Lithium fluoride would be circulated between the other side of the contactor and a vessel containing a Li-Bi solution having a lithium concentration of about 5 at. %Z. During operation of the system, the rare earths would be extracted from the fluoride salt and would accumulate in the Li-Bi solution. In carrying out the mathematical analysis, it was assumed that the salt and bismuth phases in the salt-metal contactor remain at equilibrium at all times. Although this condition will clearly not be met during operation of the experiment, such an assumption will allow a useful representation of the salt-metal contactor. The concentrations of a rare earth in the salt and bismuth phases in the salt-metal contactor at time t are then related by the following expressions: 198 ORNL DWG. 73-44 — F.X o Fe ) X¢ 7 FLUORIDE —->—%» FLUORIDE e o LiCl AND < A I e . o RARE o EARTHS Li-Bi Th-Bi FLUORIDE SALT SALT - METAL RARE -EARTH SURGE TANK CONTACTOR ACCUMUL ATION VESSEL Fig. 74. Equipment Representation and Nomenclature Used for Mathe- matical Analysis of Experiment MTE-3. 199 D, =— ’ (52) where D, = rare-earth distribution ratio between fluoride salt and bismuth containing reductant at time t, X,, = equilibrium concentration of rare earth in bismuth in con- tactor at time t, mole fraction, X, = equilibrium concentration of rare earth in fluoride salt in contactor at time t, mole fraction, and D =— ’ (53) where D = rare-earth distribution ratio between LiCl and bismuth containing reductant at time t, Xx = equilibrium concentration of rare earth in LiCl in con- tactor at time t, mole fraction. Combination of Eqs. (52) and (53) yields the following expression: o c Xp = oo X, , (54) = which relates the concentration of rare earth in the fluoride salt to that in the LiCl in the contactor at time t. The rate at which the rare earth is transferred through the contactor is then given by the expression rate = Fc X, (1 — £) . (55) where Fc = flow rate of LiCl between the contactor and the vessel con- taining the Li-Bi solution, moles/sec, f = fraction of rare earth removed from LiCl during its pas- sage through the vessel containing the Li-Bi solution. 200 The rate of transfer is also equal to the instantaneous rate at which rare earths are removed from the flunride salt surge tank, which is given by the expression rate = F(x —-xF) , (56) where F = flow rate of fluoride salt between the contactor and the fluoride salt surge tank, moles/sec, X = concentration of rare earth in fluoride salt surge tank at time t, mole fraction. If it is assumed that the rate of accumulation of rare earth in the salt and metal phases in the salt-metal contactor is negligible, Egs. (55) and (56) can be equated. Substituting Eq. (54) in the resulting relation and solving for Xpo the concentration of rare earth in the fluoride salt in the contactor at time t, yields the following expression: (57) X F F D ¢c F 1 + T D c 11— £ A material balance on the rare earth in the fluoride salt surge tank yields the following relation: dx _ _ -V ac F(x XF) ’ (58) where V = volume of fluoride salt in the surge tank, moles. Combination of Eqs. (57) and (58) yields the following differential equation: F D F X =5 -0 t dx _ F c x v F_D_ dt (59) . l+"F—-—D-"(l—f) 0 o Cc 201 where X, = initial concentration of rare-earth fluoride in fluoride salt surge tank, mole fraction. This equation can be integrated between the indicated limits to yield the following expression for the concentration of rare-earth fluoride in the fluoride salt surge tank at time t: t —=e (60) 0 where F D ¢ F , £ £ -1 _F F D Vv FC DF 1+ fr-fiz'(l — f) If it assumed that the flow rates for the fluoride salt and the LiCl are, in each case, 1% of the expected flow rates required for removing rare earths from a 1000-MW(e) MSBR (see Sect. 3) and if the concentration of reductant in the bismuth in the salt-metal contactor corresponds to a thorium concentration equal to 907 of the solubility of thorium in bismuth at the operating temperature of 640°C, the following values are obtained for the parameters in Eq. (60) for the transfer of neodymium: DF = 0.0564 , Dc = 3.5 , and 2251 F If it assumed that 50% of the neodymium is extracted from the LiCl as it passes through the vessel containing the Li-Bi solution, the expression for the quantity k can be written as: « = 0.168 % : (61) If it is desired to reduce the concentration of neodymium in the fluoride salt surge tank by 20%Z during a 24-~hr operating period, the quantity « must have the value 0.223, which then sets the value for the quantity V/F 202 at 0.753 day. The resulting volume of the fluoride salt in the surge tank is 36.1 liters. Table 52 summarizes the resulting design parameters which are dependent on the following assumptions: (1) the salt and metal phases in the salt-metal contactor remain at equilibrium at all times, (2) salt and metal flow rates equal to 1% of the expected rates required for processing a 1000-MW(e) MSBR are used, (3) neodymium is the transferring rare earth, and (4) the concentration of neodymium fluoride in the fluoride salt surge tank is to be reduced by 20% during a 24-hr operating period. Table 52. Calculated Values of the Design Parameters for Experiment MTE-3 Quantity ' Value DF 0.0564 Dc 3.5 3 33 cm” /min . 1.25 liters/min v 36.1 liters 12.2 Preliminary Design of Metal Transfer Experiment MTE-3 The planned experiment, shown schematically in Fig. 75, will use mechanical agitators to promote efficient contact of the salt and metal phases. Fuel carrier salt containing rare-earth fluorides will be circu- lated between one side of the salt-metal contactor and a fluoride salt reservoir. Lithium chloride containing rare-earth chlorides will be cir- culated between the other side of the salt-metal contactor and a rare- earth stripper, where the rare earths will be extracted into a Li-Bi solution. The experiment will use approximately 35 liters of MSBR fuel carrier salt, 6 liters of Th-Bi solution, 6 liters of LiCl, and about 5 liters of Li-Bi solution having an initial lithium content of about 5 at. Z. ORNL -DWG-71-147-RI AGITATORS VENT LEVEL ELECTRODES LEVEL /ELECTRODES FLUORIDE =~ SALT PUMP VENT i 1 ] 1 ] L J L ] L J ARGON SUPPLY — ARGON SUPPLY 33 Cm3/ min I.25 — Zmi 3 — \\§/ BETWE _—PA RTngN CELL HALVES Bssmum_ GAP FQR P BISMUTH FLOW Fig. 76. Mechanically Agitated Salt-Metal Contactor Proposed for Use with Metal Transfer Experiment MTE-3. 208 Investigation showed that use of a 3-in.-diam, four-bladed paddle having blades canted at a 45° angle did not result in entrainment of water until an agitator speed of 400 rpm was reached. It is believed that the flow patterns in the water and mercury that are created by the use of canted blades inhibit the formation of a vortex around the agita- tor shaft and thereby reduce the extent of dispersion of water in the mercury phase. 13.2 Survey of Literature Relative to Mechanically Agitated, Nondispersing Salt-Metal Contactors Various investigations have been carried out for determining the rate of mass transfer between aqueous and organic phases in mechanically agitated contactors in which the phases are not dispersed. Lewisa9 has investigated mass transfer rates in six solvent-water systems and two nonaqueous systems. The solvents used were aniline, ethyl acetate, ethyl acetoacetate, ethyl formate, furfural, and isobutanol; formic acid--benzene and aniline-cyclohexane were the nonaqueous systems. A sketch of the apparatus used, with reported dimensions, is given in Fig. 77. The speed and direction of rotation of the agitators were controlled individually; most of the data were taken with the agitators contrarotating, but it was stated that the relative direction of rotation made little difference. Lewis correlated his data by the expression: 1.65 60 k 1 6 - 2 y = 6.76 x 10 Re, + Re2 EI + 1, (62) 1 where k1 = individual mass transfer coefficient, cm/sec, v, = kinematic viscosity (u/p), Cmg/sec, | Re = ndg/v, dimensionless, . = viscosity, poises, o = density, g/cmj, d = agitator diameter, cm, and n = agitator speed, rps. 209 ORNL DWG 73-36 LEWIS 300 mA | n 4in GORDON STIRRER BARS 0.48cm in DIAMETER 8 SHERWOOD 15.2¢cm | 12 STIRRER BARS 3in. LONG OLANDER —t - 3l:m 6.2c¢cm — - 12 6in I |_GRID PROCHAZKA . PADDLE DIAMETER 4.5cm BULICKA '¢™ =t = L1 10.7em Fig. 77. Schematic Diagrams of Equipment Used By Various Investi- gators for Measuring Mass Transfer Coefficients in Stirred Interface Contactors. 210 The subscripts 1 and 2 refer to phases 1 and 2, respectively. In a sub- sequent paper, Lewis50 reported results for the rate of transfer of a third component between the two phases of several solvent-water systems. He stated that the relation given by Eq. (62) also correlated the data on the rate of transfer of the third component. 51 McManamey” correlated Lewis' results and his own data (obtained in a cell similar to that used by Lewis ) by the expression: 60 k 1 = 0.1 cm_1 v -0.3% )0_9 vy D 1 where D = diffusivity of the transferable material in the phase indicated, cmg/sec. 52 Mayers correlated Lewis' results and his own data (obtained in a cell similar to that used by Lewis ) by the expression: k,d M) 1-9 o) -2k 1/2(\;1 5/6 51—: 0.00316 J;} (0.6 + “—1_) (Re1 Reg) 5; . (6h4) 53 Gordon and Sherwood determined individual mass transfer coeffi- cients in the system isobutanol-water with and without a third solute. A sketch of their apparatus is given in Fig. T77. The agitators in each phase were mounted on a common shaft. No correlation was suggested. L . Olander5 determined mass transfer coefficient values in several solvent-water systems (many of which were identical to those used by Lewis) in the apparatus shown in Fig. T7. The agitators for each phase were mounted on a common shaft. The suggested correlation was: 1 )o.lm 0.67 1 = 0.046 (%) , (65) where w = agitator speed, radians/sec. 211 55 Prochazka and Bulicka determined mass transfer coefficient values in water--ethyl acetate and water-isobutanol systems. A sketch of their apparatus is shown in Fig. 77. The agitators were controlled separately, and the agitator blades were canted at an angle of 45°. The suggested correlation for their mass transfer coefficient data was: e 1T (Re, ) — . (66) P 1 + (—L) Po The magnitudes of the individual mass transfer coefficients obtained =~ o ¥ — < g o | by most of the investigators were approximately the same -- from about 0.007 cm/sec for ethyl acetate--water systems to about 0.001 cm/sec for isobutanol -water systems. The values for the individual mass transfer coefficients reported by Prochazka and Bulicka, however, were two orders of magnitude greater -- from about 1.1 cm/sec for ethyl acetate --water systems to about 0.1 cm/sec for isobutanol-water systems. It was later found that a decimal point has been misplaced in their literature arti- 56 cle. After the appropriate correction had been made, the reported results were found to be consistent with the data of other investiga- tors. Although the magnitudes of the individual mass transfer coeffi- cients obtained by different investigators were about the same, the reported dependence of the mass transfer coefficient on the agitator speed varied widely. Part of this difference is apparently due to the different relations used to correlate the data, and part is due to a different degree of interfacial turbulence created in the various pieces of equipment at a given agitator speed. The scatter in the reported data is shown in Figs. 78-80. In Fig. 78 the data from all the investi- gators are plotted in the manner suggested by Lewis; in Fig. T9 the data are plotted according to the correlation suggested by QOlander; and in Fig. 80 the data are plotted as suggested by Prochazka and Bulicka. ORNL DWG 73-67 1000 T T T TTTT] T T T 7T77H T T T T 7773 700 » 500 B 300 _ zoor- o Et. Ac.-Water >PROCHAZKA and BULICKA ;o4 1 4141 1 1 1 Et. Ac.- OLANDER 2 30 LEWIS .a:lo:. '_ \ 1l 20 Water - Isob TE 3 — PROCHAZKA - 85 and - — BULICKA - 3 » 2l _ Water -1sob. ' + Bi-LiCl ] ~L . _ LiCI-Bi \\ Bi-F ; F-8i \\\’ l ’/ 05 | 1ot J Lot || o1 1000 3000 5000 10000 30000 50000 100000 300000 1000000 [} Rcl + Rea r""l Fig. 78. Correlation of Data on Mass Transfer Coefficients in Stirred- Interface Contactors in the Manner Suggested by Lewis. The operating con- ditions that will be obtained at the various interfaces in metal transfer experiment MTE-3 with an agitator speed of 3 rps are shown on the abscissa. The notation F-Bi denotes transfer from fluoride salt to a bismuth phase; LiCl-Bi, transfer from molten LiCl to a bismuth phase; Bi-LiCl, transfer from a bismuth phase to molten LiCl; and Bi-F, transfer from a bismuth phase to fluoride salt. )0.44 213 ORNL DWG 73-66 100 T ! T T T 1 1 1 1 1 T 50 30 20r a LEWIS: Reg fural \ 10}~ PROCHAZKA and BULICKA = 4750 2L a 5 =2370 ™ OLANDER 3r 8 8 2k ° Bi-F | L L a1 3 | | L 11 I 1 100 300 500 1000 3000 5000 10000 30000 50000 w/v Fig. 79. Correlation of Data on Mass Transfer Coefficients in Stirred Interface Contactors in the Manner Suggested by Olander. The operating conditions that will be obtained at the various interfaces in metal transfer experiment MTE-3 with an agitator speed of 3 rps are shown on the abscissa. The notation F-Bi denotes transfer from fluo- ride salt to a bismuth phase; LiCl-Bi, transfer from molten LiCl to a bismutl phase; Bi-LiCl, transfer from a bismuth phase to molten LiCl; and Bi-F, transfer from a bismuth phase to fluoride salt. 21k ORNL DWG 73-65 100 1 ¥ A ] | 1 T T I I T T T T I 701 -fi 50~ -J 3CH Et. Ac.- Water — zor — Et. Ac.-Water &~ LEWIS, Water-Furfural 7 . 5 //. - \Waur- [sob. L . _ 2_ — Water - Isob. L _ F Salt-Bi LiCl- B Bi-LiCl Bi-F 100 300 500 1000 3000 5000 10000 30000 50000 w 2§ Fig. 80. Correlation of Data on Mass Transfer Coefficients in Stirred Interface Contactors in the Manner Suggested by Prochazka and Bulicka. interfaces in metal transfer experiment MTE-3 with an agitator speed of 3 rps are shown on the abscissa. The notation F-Bi denotes transfer from fluoride salt to a bismuth phase; LiCl-Bi, transfer from molten LiCl to a bismuth phase; Bi-LiCl, transfer from a bismuth phase to molten LiCl; and Bi-F, transfer from a bismuth phase to fluoride salt. The operating conditions that will be obtained at the various 215 Examination of the three methods for correlating the data shows that the Olander correlation has serious shortcomings. This correla- tion assumes that the individual mass transfer coefficient in one phase is independent of the level of turbulence in the second phase. How- ever, data of Lewis and data of Prochazka and Bulicka show that the individual mass transfer coefficient in one phase is definitely depen- dent on the level of turbulence in the other phase. The Prochazka- Bulicka relationship does not seem to adequately correlate data from different systems; for instance, the data for water-isobutanol and isobutanol-water do not fall on the same line. Of the three correla- tions, the Lewis correlation seems to represent the data best. The operating conditions that will be obtained with an agitator speed of % rps for the phases to be used in experiment MTE-3 are indi- cated along the abscissas of the figures for the three correlations. The values for the Reynolds-number grouping lie outside the range of the data obtained by Lewis but inside the range of the data obtained by Prochazka and Bulicka. The Prochazka-Bulicka correlation indicates a probable value for the individual mass transfer coefficient in the fluoride salt of about 0.0015 cm/sec; extrapolation of the Lewis corre- lation indicates a probable value of 0.05 cm/sec. Estimates of the individual mass transfer coefficients from the MTE-2 experiment indicated a coefficient of about 0.001 cm/sec under conditions where practically no agitation oécurred. The value obtained from the Lewis correlation seems much more likely to reflect operating conditions that will be used in experiment MTE-3. 216 14. HYDRODYNAMICS OF PACKED-COLUMN OPERATION WITH HIGH-DENSITY FLUIDS J. S. Watson L. E. McNeese The hydrodynamics of packed-column operation with fluids having high densities and a large density difference is being studied in order to evaluate and to design countercurrent contactors for use in MSBR processing systems based on reductive extraction. Mercury and water are being used to simulate bismuth and molten salt in these studies. Earlier experiments57 carried out with 1/8- and 1/L-in. solid cylinders and with 3/16- and 1/4-in. Raschig rings in a 1~in.-ID column demon- strated that a transition in mode of flow of the dispersed phase occurs between the packing sizes of 3/16 and 1/4 in. This transition appeared to be a function of the packing size only and was not related to packing shape (solid cylinders or Raschig rings ). With the larger packing the mercury was dispersed into small droplets, which produced a large inter- facial area. With the smaller packing, the mercury flowed down the column in continuous channels and resulted in a much smaller interfacial area. A large interfacial area (and hence large-sized packing) is desired in order to obtain high mass transfer rates between the salt and metal phases. A 2-in.-diam column was installed in the experimental system in order to carry out studies with packing of larger sizes, and measurements were made of flooding rates, dispersed-phase holdup, and pressure drop Wéth 1/4-in. solid cylinders and with 3/8- and 1/2-in. > Raschig rings. The results of these studies have shown that the dis- persed-phase holdup and the column throughputs at flooding can be corre- 58 lated on the basis of a constant superficial slip velocity in the following manner: Vc Vd Tx "% " Vs> (67) 1/2 1/2 1/2 Ve, Vo =V o (68) 217 where Vc = superficial velocity of continuous phase, ft/hr, Vy = superficial velocity of dispersed phase, ft/hr, v, = superficial slip velocity, ft/hr, and X = dispersed-phase holdup. The subscript f denotes conditions at flooding. These relations were previously58 extended to cover dispersed-phase holdup and throughput at flooding with salt-bismuth systems by assuming that, for a given packing size, the slip velocity was (1) independent of the viscosity of the continuous phase, (2) proportional to the difference in the densities of the phases, and (3) proportional to the packing void fraction. Although the resulting relations predicted flooding rates that were in excellent agreement with flooding rates measured with molten salt and bismuth, it was realized that the agreement did not constitute veri- fication of the assumed effects of the continuous-phase viscosity and the difference in the densities of the phases. During this reporting period, data showing the dependence of slip velocity on continuous-phase viscosity were obtained by a group of students from the MIT Practice School.59 A 2-in.-diam, 24-in.-long column packed with 3/8-in. Teflon Raschig rings, which were not wetted by either the continuous phase or the dispersed phase, was used in the study. Results of this study and the development of an improved rela- tion for predicting packed column performance during the countercurrent flow of molten salt and bismuth are discussed later in this section. Studies for evaluating the effect of wetting of the packing by the nominally dispersed phase were also carried out. 14.1 Equipment and Experimental Technique The experimental facility used in the present study has been 58 described previously. However, the system (shown in Fig. 81) was modified for the present studies so that water or water-glycerin solu- tions could be circulated through the column at constant temperature by installation of a heat exchanger and aqueous-phase surge tank. During the study in which the effect of the viscosity of the continuous 218 ORNL DWG 72-3243 PACKED DISENGAGING. 7, SECTION ///- ' BALL VALVE NN GLASS COLUMN 2%-in. LONG x t~in. L.D. 24-in. LCNG x 2-in. 1.D. i BALL. VALVE| 1 I 1 | P ! : . | — | =T J e s WATER PUMP MERCURY PUMP (CENTRIFUGAL) (POSITIVE DISPLACEMENT) Fig. 81. Flow Diagram of Equipment Used for Determining Dispersed- Phase Holdup, Flooding Rate, and Pressure Drop in a Packed Column During the Countercurrent Flow of Mercury and Aqueous Solutions. 219 phase on slip velocity was determined, the column consisted of a 2-in.- ID, 24-in.-long glass tube that was packed with 5/8-in. polyethylene Raschig rings having a void fraction of 0.66. During subsequent tests for determining the effect of wetting of the packing by the nominally dispersed phase, the column was packed with 3/8-in. copper Raschig rings (void fraction, 0.81) that had been treated with nitric acid in order to promote wetting of the packing by the mercury. A 6-in.-long section above the column was packed with O.5-in. Raschig rings in order to dis- tribute the mercury uniformly over the column cross section. Ball valves having internal diameters equal to that of the column were pro- vided above and below the column for use during measurements of dispersed- phase holdup. The holdup measurements were effected by simultaneously closing both ball valves and subsequently draining the mercury into a graduated cylinder. It was found that approximately 35 ml of mercury remained in the column during the draining operation. The mercury flow rate was measured during the studies by closing the lower ball valve and determining the time required for filling a 10-in.-long section of the column which had been calibrated. During the flooding measurements, the water flow rate was increased incrementally with a fixed mercury flow rate until flooding was observed. Steady-state conditions were normally reached within 30 min after each adjustment of the aqueous-phase flow rate. The column was considered to be flooded when it became impossible to maintain fixed flow rates without a constantly increasing pressure drop. With the part}cular equipment used in this experiment, the continuous—-phase flow rate and the pressure drop across the column fluctuated widely under "flooding" conditions. Mercury could be seen accumulating in the column until the water flow was essentially blocked. Mercury would then drain from the column, and the water flow rate would return to the desired value. This behavior was repeated in a cyclic manner. 14.2 Results Data were obtained on dispersed-pfiase holdup, column pressure drop, and flooding, using viscosities of 1, 7.5, and 15 c¢P for the continuous phase. Tables 5%-55 and Figs. 82-84 present data on holdup and pressure 220 Table 53. Experimentally Determined Values for Dispersed-Phase Holdup and Pressure Drop During Countercurrent Flow of Mercury and Water Through a 2-in.-diam Column Packed with 5/8-in. Teflon Raschig Rings Continuous-phase viscosity, 1 cP Superficial Velocity (ft/hr) Dispersed- Pressure Dispersed Continuous Phase Drop a Phase Phase Holdup (mm Hg-mm H,0) 0 36.3 -- ~0 0 50.3 -~ 1 0 65.54 -- 1 0 1hk -- 2 0 206 -- 3 0 272 -- L 0 340 -- 6 71 0 0.0506 18 71 67 0.0678 18 71 1hbk 0.0826 16 71 205.7 0.0838 19 71 271 0.0666 24 71 339 0.0801 5 145 0 0.1%6 35 145 14l 0.183 26 145 205.7 0.160 39 145 271 0.17h 50 145 339 0.183 52 188 0 0.183 L9 188 14k 0.206 39 188 206 0.217 52 188 271 0.2Lh7 79 188 339 0.271 102 Millimeters of liquid having a density of 12.6 g/cma. drop for a range of values of the continuous-phase superficial velocity. The points at which flooding was observed are indicated in the figures. Table 56 and Fig. 85 give values of holdup and pressure drop obtained during operation with a continuous-phase viscosity of 1 cP for the case in which the packing was wet by the dispersed phase. Data on flooding are presented in Figs. 86-89, where the square root of the dispersed-phase Table 54. 221 Experimentally Determined Values for Dispersed-Phase Holdup and Pressure Drop During Countercurrent Flow of Mercury and a Water-Glycerin Solution Through a 2-in.-diam Column Packed with 3/8-in. Continuous-phase viscosity, T.5 cP Teflon Raschig Rings Superficial Velocity {ft/hr) Dispersed- Pressure Dispersed Continuous Phase Drop Phase Phase Holdup (mm Hg-mm HQO)a 0 13.9 - 1 0 2l .1 -- 1.5 0 33.9 -- 2 0 81.1 = 3 0 136.2 -- L 0 198.4 -- 57 0 259.5 -- - 8-13 0 303 -- 11-14 0 373 ~- 15-18 0 431.5 -- 22-25 0 L84 - 31 0 568 -- 3T 67 0 0.0716 28 67 81.1 0.0900 33 67 136.2 0.0998 4o 67 198.4 0.0839 48 67 172 0.0962 L3 67 218 0.122 61 67 259 0.131 70 67 28k 0.128 60 107 0 0.154 26 107 81.1 0.154 L 107 3%.9 0.148 27 107 136.2 0.168 33.5 107 198.4 0.143 63 107 160 0.138 50 107 150 0.127 38 107 230 0.181 67 12k 0 0.154 40 12k 33.9 0.17h 61 124 81.1 0.181 50 124 136.2 0.192 65 124 172 0.211 72 Millimeters of liquid having a density of 12.6 g/cm3. 222 Table 55. Experimentally Determined Values for Dispersed-Phase Holdup and Pressure Drop During Countercurrent Flow of Mercury and a Water-Glycerin Solution Through a 2-in.-diam Column Packed with 5/8-in. Teflon Raschig Rings Continuous-phase viscosity, 15 cP Superficial Velocity (ft/hr) Dispersed- Pressure Dispersed Continuous Phase Drop Phase Phase Holdup (mm Hg~-mm H20)a 0 22.4 -- ~0 0 51.2 -- 1 0 116.5 -- 9 0 167 -- 12 0 216.2 -- 15 0 270 -- 19 0 330 -- 2l G 397 -- 28 71 0 0.107 5 71 100 0.109 9.5 71 200 0.141 12 71 220 0.17h 62 s 0 0.131 0 85 22.4 0.157 -- 85 51.2 0.143 -- 85 110.5 0.158 -- 85 167 0.181 -- 85 196 0.220 -- 85 228 0.25% -- 126 0 0.204 31 126 22,4 0.195 3% 126 51.2 0.197 36 126 110.5 0.217 57 ®Millimeters of liquid having a density of 12.6 g/cm5. 223 ORNL DWG 73-47 ' { T [ I I W - L 0.3 L { o’ -A { & - ® .‘. e 1 AT 2 _ A 8 O 02 ‘ .......................... . A.-’/—' " = x© wXx Q 0 0.1} a - o T""‘"""‘"‘ ______ -._ —————————— ._ ----- o 0 —+ : t t t : += Dispersed Phase Velocity . 8o} {ft/br) A 7 a 8 - 0 o _———_—g-- T eof . ! o —— 145 E | s W E ; Ao 188 o 8 o & E ______ —— —— T T 50 100 150 200 250 300 350 CONTINUQUS PHASE VELOCITY (ft/hr) Fig. 82. Variation of Dispersed-Phase Holdup and Column Pressure Drop During the Countercurrent Flow of Mercury and Water in a 2-in.-ID 2h-in.-long Column Packed with 5/8-in. Teflon Raschig Rings. 2 ool ORNL DWG. 73 -35RI 5 _DJ T T B v 0 ¥ L O T tJ O 3 = = 2 = FLOCDED IS 1 ruonoes > ] oOT w n @ u—l — a v o 0 FLOODED 80| | FLOODED FLOTDED ~ / 7/ 60 - 1 3o DISPERSED PHASE 5 N VELOCITY, Vg4 o T | (£1/hr) W E 40 e o T T -—--8-——- 67 A o ——e—— 107 Ll X 20 .‘ ....... |24 = x e - a —p - E e o - = v L 1 1 1 0 50 100 150 200 250 300 350 CONTINUOUS PHASE VELOCITY, V¢ (ft/hr) Fiz. 83. Variation of Dispersed-Phase Holdup and Column Pressure Drop During the Countercurrent Flow of Mercury and a Water-Glycerin Solution in a 2-in.-ID, 24-in.-long Column Packed with 5/8—in. Teflon Raschig Rings. The viscosity of the continuous (aqueous ) phase was 7.5 cP. ORNL DWG. 73-37 a 3 . : ‘ i | ] J - % > FLOODED _ - FLOODED g Q ‘ < o T e FLOODED I~ 02*.‘..‘. | 5 . o uJI e n o B xoe OIF """ | W~ a » 0 0 } 4 : : % : DISPERSED PHASE 3 VELOCITY, Vg — Qo g (ft/nr) xrT FLOODED _ _ _ e o e FLOODED E 60} b & - 2. * 'l e—— 85 o I ST I ol 126 nxT | Ll 40 ', - @x e ' G.E | ODED e Pl - - | | 0 50 100 150 200 250 300 350 CONTINUOUS PHASE VELOCITY, Vg (ft/hr) Fig. 8L4. variation of Dispersed-Phase Holdup and Column Pressure Drop During the Countercurrent Flow of Mercury and a Water-Glycerin Solution in a 2-in.-ID, 24-in.-long Column Packed with 3/8-in. Teflon Raschig Rings. The viscosity of the continuous (aqueous ) phase was 15 cP. Table 56. Experimentally Determined Values for Dispersed-Phase Holdup and Pressure Drop During Countercurrent Flow of Mercury and Water Through a 2-in.-diam Column Packed with 3/8-in. Copper Raschig Rings Continuous-phase viscosity, 1 cP Superficial Velocity (ft/hr) Dispersed~-Phase Pressure Drop Dispersed Phase Continuous Phase Holdup (mm Hg-mm HEO) 113 0 0.108 15 113 Lhk 0.106 L6 113 272 0.113 4O 113 406 0.119 14 315 0 0.201 33 315 RN 0.204 36 315 206 0.194 -- 315 339 0.215 41-50 315 Lo6 0.270 75 315 L0 0.308 103 315 380 0.233 58 Millimeters of liquid having a density of 12.6 g/cmj_ 9¢e 227 ORNL DWG 73-I ! I T T 1 T T 1 a 03 / o | 3 & a ---------------------- - _ § 0.1k - W = 0 ] 1 1 | L 1 I I 1 T T 1 i I | I 100~ Dispersed Phase Velocity = (ft/hr) e enee 13 QS 315 o N o« I w B % ' e 40} a - @~ . E tF————————em et e === [ e & E L . | 0 1 ] { 1 1 i 1 1 0 50 0O 150 200 250 300 350 400 450 CONTINUOUS PHASE VELOCITY (ft/hr) Fig. 85. Variation of Mercury Holdup and Column Pressure Drop During the Countercurrent Flow of Mercury and Water in a 2-in.-ID, 2L-in.-long Column Packed with 3/8-in. Copper Raschig Rings. The packing was wet by the mercury. 172 I/72 (DISPERSED PHASE VELOCITY, V;) (ft/hr) ORNL DWG 73-2RI 20 T T I T ! T 18 ® — Not Flooded 7] o - CALCULATED FLOODING LINE for 72 172 IV2 'ZF ¢ ¢ o Ve +VYy . 33.6 (1t/hn) B IO- T ® ® ® o . ¢ n 4}~ - 2 - 0 5 10 1S 20 25 30 35 Fig. 86. /2 I/ (CONTINUOUS PHASE VELOCITY, V) (ft/hw) 2 Flooding Data Obtained During the Countercurrent Flow of Mercury and Water Through a 2-in.-ID, 24-in.-long Column Packed with %/8-in. Teflon Raschig Rings. gece /72 (ft/hr) i172 (DISPERSED PHASE VELOCITY,Vq) ORNL DWG. 73-42 RI 14 T T T T T 2 CALCULATED FLOODING LINE for ° ° o o + Vc|/z+ V‘ln + 29.1 (tt/n)”? ol ® e oo o o + - sl ® e o o0 O - 6} - 4} - ® NOT FLOODED - + FLOODED o) i 1 L L 1 ] o) 5 |10 15 20 25 30 35 / (CONTINUOUS PHASE VELOCITY ., V)2 (ft/nn)V? Fig. 87. Flooding Data Obtained During the Countercurrent Flow of Mercury and a Glycerin-Water Solution Through a 2-in.-ID, 24-in.-long Column Packed with %/8-in. Teflon Raschig Rings. The viscosity of the continuous (aqueous ) phase was 7.5 cP. 6cc ORNL DWG. 73-43 /2 172 (CONTINUOUS PHASE VELOCITY, Vo) (ft/hn) Fig. 88. Flooding Data Obtained During the Countercurrent Flow of Mercury and a Glycerin-Water Solution Through a 2-in.-ID, 24-in.-long Column Packed with 3/8-in. Teflon Raschig Rings. The viscosity of the continuous (aqueous ) phase was 15 cP. ~N = £ = N:: 14 r { T Y I T > o I12F CALCULATED FLOODING LINE for - =) . . o 4+ 4+ 4 2 12 172 > Ve +Vy4 =25.95 (ft/hr) ~ 10} . O r . ° ® o o S et ¢ w ] > w 6 - 2 Y - i ® NOT FLOODED 8 2 + FLOODED - v T o 1 1 1 L | 1 % 0 5 10 15 20 25 30 35 a oeEe I/72 (ft/hr) 172 (DISPERSED PHASE VELOCITY, V4 ) ORNL DWG. 73-38 20 1 T 1 I L | CALCULATED FLOODING LINE p . . e o060 72 1/2 172 for Vo +Vq = 39.75 (ft/hr) TES ® NOT FLOODED 5k - + FLOODED 0 | i 1 1 1 l ] 0 5 10 15 20 25 30 35 40 /2 172 (CONTINUOUS PHASE VELOCITY,V(;)I (ft/hr) Fig. 89. Flooding Data Obtained During the Countercurrent Flow of .~ Mercury and Water in a 2-in.-ID, 24-in.-long Column Packed with 3/8-in. Copper Raschig Rings. The packing was wet by the mercury. 232 superficial velocity is plotted vs the square root of the continuous- phase superficial velocity. The conditions for which the system was operated in a nonflooded state, as well as conditions for which flooding was observed, are indicated. The results obtained during this study were similar to those observed previously for mercury and water. The data on dispersed-phase holdup involving nonwetted packing could be correlated in terms of a constant superficial slip velocity, as shown in Table 57. The relative standard deviations for the slip velocities were about #10%; no dependence of slip velocity on the flow rate of either phase was noted. The variation of slip velocity with the continuous-phase viscosity is shown in Fig. 90, which indicates that the superficial slip velocity is proportional to the -0.167 power of the continuous-phase viscosity. As expected, the dependence is not large; however, neglect of this effect was significant in the earlier extrapolation of data from a mercury-water system to a salt-bismuth system since the continuous-phase viscosity differs by a factor of 12. The effect of wetting of the packing by the metal phase on dispersed- phase holdup, flooding, and pressure drop was also evaluated by packing the column with 3/8-in. copper Raschig rings that had been etched with nitric acid (to promote wetting of the packing by the mercury). Com- plete wetting of the packing was obtained. Although the mercury was saturated with copper, the solubility of copper in mercury is quite low and no iwportant changes in other physical properties should occur. After a few hours of operation, solids were observed at the water-mercury interface below the column; these solids were believed to be copper oxide that was formed as the result of the reaction of dissolved copper with oxygen in the water. Periodic additions of nitric acid to the aqueous phase quickly removed the solids. The interfacial area between the aqueous and mercury phases was decreased substantially when the packing was wetted by the mercury, as noted by comparing Figs. 91 and 92, which show operation with nonwetted and wetted packing respectively. No dispersion of the mercury was observed with the wetted packing, and the interfacial area was essentially equal to the packing surface area. The superficial slip velocity (and hence the flooding rates) was considerably greater with wetted packing 233 Table 57. Variation of Superficial Slip Velocity with Continuous-Phase Viscosity and Wetting of the Packing for a Column Packed with 5/8-in. Raschig Rings . . Slip Continuous- Packing . Standard Type Phase Void Ve13c1ty, Deviation, D P?riént of Viscosity Fraction, S g EVl7 ton, Packing (cP) € (ft/hr) (ft/hr) Vg Nonwetted 1 0.66 1129 +105 9.3 Nonwetted 7.5 0.66 8Le +88 +10.4 Nonwetted 15 0.66 674 +49 +7.3 Wetted 1 0.82 1583 +268 +16.9 than with nonwetted packing, as shown in Table 57. It was not clear whether the data on metal-phase holdup with wetted packing could be correlated on the basis of a constant superficial slip velocity. Only nine metal holdup measurements were made, and a quantitative analysis of the relation between mercury holdup and the water and mercury super- ficial velocities was not possible. 14.3 Prediction of Flooding Rates and Dispersed- Phase Holdup in Packed Columns After the dependence of slip velocity on the continuous-phase vis- cosity had been determined, it was possible to reevaluate the dependence of slip velocity on the difference of the densities of the two phases by using the two data points afforded'by the meréury—water data and salt- bismuth data. The salt-bismuth data indicated a superficial slifi velocity of 388 ft/hr with a continuous-phase viscosity of 12.1 cP, a density difference of 6.28 g/cm;, and a packing void fraction of 0.84. It was assumed that the dependence of slip velocity on the difference in den- sities is a power-type dependence, and the resulting power value was 0.5. This result is interesting because it is the same as the dependence of SLIP VELOCITY x10~2 (ft/hr) ORNL DWG 71-2867 20 T l O 1 @ I SLOPE=—(VISCOSITY EXPONENT) =—b2 0.167 )] | H ' N i I | | | 1 1 11 11 | | | 1 1 1 111 0. 0.2 04 06 08 10 2 4 6 8 10 20 CONTINUOUS PHASE VISCOSITY (cP) Fig. 90. Variation of Superficial Slip Velocity with Continuous- Phase Viscosity in a 2-in.-diam Column Packed with 5/8-in. Teflon Raschig Rings. hee PHOTO 0009-71 R s LELS Fig. 91. Countercurrent Flow of Mercury and Water in a 2-in.-diam Column Packed with 3/8-in. Teflon Raschig Rings. The packing is not wetted by the metal phase, and the interfacial area is much greater than the surface area of the packing. Fig. 92. Countercurrent Flow of Mercury and Water in a 2-in.-diam Column Packed with 5/8—in. Copper Raschig Rings. The packing is wetted by the metal phase, and the interfacial area is essentially equal to the surface area of the packing. 237 drop terminal velocity on the difference in densities for conditions under which inertial forces predominéte and the drag coefficient is essentially constant. The final relation for predicting the variation of superficial slip velocity with packing void fraction, the difference in the densities of the phases, and the continuous-phase viscosity is, then: -0.167 0.5 Vs = Vs Hg-H_O eE ' L_ih—) (25_49—__- ’ (69) PP “Ret/ P10 Hg -H,,0 where VS = superficial slip velocity, v = slip velocity for mercury-water for the packing size being s,Hg-HQO used, e = packing void fraction, _ . . I . d _ €Ref void fraction for packing for which Vs,Hg-H 0 was deter mined, 2 i = viscosity of continuous phase, My o = viscosity of water at 20°C, 2 Mo = difference in the densities of the phases, and fipHg-H o= difference in the densities of mercury and water at 20°C. 2 Slip velocity values calculated from Eq. (69) can be used with Eqs. (67) and (68) for determining both the dispersed-phase holdup and throughputs of the continuous and dispersed phases at flooding. 238 15. ANALYSIS OF MULTICOMPONENT MASS TRANSFER BETWEEN MOLTEN SALTS AND LIQUID BISMUTH DURING COUNTERCURRENT FLOW IN PACKED COLUMNS C. P. Tung J. S. Watson Reductive extraction, an important operation in the removal of prot- actinium and rare earths from MSBR fuel salt, involves the exchange of metal ions in the salt phase with neutral (reduced) metal atoms in the bismuth phase. Since no net electric current flows between the salt and metal phases, the rate at which metal ions are reduced must equal the rate at which metal atoms are oxidized (taking into consideration differ- ences in the charges of the ions involved). 1In the bismuth phase, the fluxes of the transferring atoms are dependent only on concentration gradients. However, in the salt phase, electric potential gradients are generated near the salt-metal interface as the result of differences in the mobilities and/or charges of the various diffusing ions. This results in a condition where the fluxes of the transferring ions are dependent on both concentration gradients and electric potential gradients. These effects greatly complicate the mass transfer process and make the design of continuous (differential) reductive extraction columns difficult. We have begun a mathematical analysis of mass transfer during reductive extraction processes to facilitate interpretation of the results from present and proposed experiments in packed columns, and as an aid in using these data for the design of larger reductive extraction systems. The results from this study will also be applicable to several other sol- vent extraction operations that involve exchange reactions between uncharged species in a solvent and ions in an electrolyte. The extrac- tion of metals from aqueous solutions by use of "liquid ion exchange" solvents such as amines is a typical example of such an operation. 15.1 Literature Review It has been recognized for some time60 that the diffusion of ions in a liquid is inherently different from the diffusion of uncharged species. The interdiffusion of charged species as a result of concentration grad- ients produces an electric potential gradient, which, in turn, alters the ionic fluxes and prevents further deviation from electroneutrality. An electrical potential gradient in an electrolyte will produce a flux of 259 ion i, which depends on the concentration of the ion in the electrolyte and on the valence and mobility of the ion, as indicated by the relation = d Ji uiZiCi grad ¢ , where Ji =\f1ux of ion i, u, = mobility of ion i in electrolyte, ;= valence of ion i, ;= concentration of ion i, and grad ¢ = electrical potential gradient. The mobility of an ion is related to its diffusion coefficient in the electrolyte by the Nernst-Einstein equation as follows: D.F a = —= i RT ’ where Di = diffusion coefficient of ion i in the electrolyte, F = Faraday constant, R = gas constant, and T = absolute temperature. If a concentration gradient is also present in the electrolyte, an addi- tional flux, which is described by Fick's First Law, is produced; the flux resulting from this effect is given by the relation Jl = -Dl grad Ci The net flux of ion i will be the sum of these contributions, which is knowvn as the Nernst-Planck equation: Z,C F ii RT = - d . Ji Di(grad Ci + grad ¢) Several studies related to mass transfer during reductive extraction operations have been reported in the literature. The effects of electric fields on the rate of diffusion of ions were considered by Schldgl and Helfferich.6O Copeland, Henderson, and Marchello61 derived an analytical solution for the rate of transfer of ions through the liquid film sur- rounding solid ion exchange resin beads for the case in which the 240 exchanging ions have the same charge. Turner and Snowdon62 extended these results to cover the case in which the exchanging ions have different valences. Kataoka, Sato, Nishiki, and Ueyama65 studied systems involving two different coions (e.g., ions having electric charges opposite in sign to that of the exchanging ions). This study will attempt to extend the previous studies in three important ways: 1. Countercurrent liquid-liquid columns will be treated rather than portions of fixed beds. 2. Diffusion resistances may exist in both phases. Previous studies have treated either resistance to diffusion in solid ion exchange resins or resistance to diffusion through a liquid film surrounding the resin. The case for which significant resistance occurs in each phase is considerably more complex than either of these. 3. Any number of transferring ions may be considered. The results will not be limited to the case in which only one ion exchanges for another ion. The method that will be developed will be general, and any number of ions can be treated. However, provisions are being made for only as many as ten ions in computer programs that are being developed. 15.2 Mathematical Analysis The model selected for the mathematical analysis in the present study is based on a model similar to the Whitman two-film model; that is, an effectively stagnant film of liquid is assumed to exist on both sides of the interface, and all concentration gradients and all electric potential gradients are assumed to lie within these two films. The phases are assumed to be in chemical equilibrium at the interface. The rates of dif- fusion of species in the solvent film are controlled by Fick's First Law since the transferring materials have no electric charge in the solvent (bismuth). Thus, at a point within a packed column, the flux of compo- nent i across the solvent film is given by the relation 1= _—c.), (70) wvhere i flux of component i across the solvent film, si diffusion coefficient of component i in solvent phase, 241 68 = thickness of the solvent film, SiB = concentration of component i in the bulk solvent, and = concentration of component i in the solvent phase at the interface. sil The rate of transfer of ions in the electrolyte film, however, involves electrical transference as well as diffusion, and both concentration and electrical potential gradients are important. For this case, the flux is given by the Nernst-Planck equation: Zz.C .F _ g L diei ] Tei Dei[gra Cei RT 8¥ad ¢ (71) where Z. = valence (electric charge) of component i, = electric potential in the electrolyte phase, = diffusion coefficient of transferring component, Faraday constant, = concentration of transferring component, m 0o ™M g e - i = gas constant, and T = absolute temperature. The subscript i refers to component i, and the subscript e denotes the electrolyte phase. Since the films are thin, it is assumed that the rate of accumulation of materials in the films is negligible in comparison with the flux of the materials through the films. Thus, the flux of component i is con- stant across the films so that Jei - Jsi ’ (12) The fluxes, however, will vary with axial position in the column; this effect will be considered subsequently. There is no net transfer of electric charge between phases (e.g., no net electric current ); this is expressed by the relation E T 7 = E T 7 =0 . {5 . . = - . . = ( ) Also, the accumulation of electric charge at any point can be neglected, 2o so that at all points in the electrolyte film Zi:cei 2 = 4 & o (T4) where tlie subscript Y refers to the pontransferring ion. If the trans- ferring ions have a positive charge (as is the case with reductive extrac- tion), at least one other ion in the electrolyte phase must have an opposite (negative) charge. (Ions that are charged oppositely to the transferring ions are referred to as coions.) If the transferring ions are negstively charged, the coions would be positively charged. The rate of transfer of coions between the phases is assumed to be negligible. It will be convenient to develop two general expressions from Eqs. (7t) and (74 ) which will be used subsequently. Multiplication of Eq. (71) by the quantity ZiDe' and summation of the resulting expression over the i i transferring ions yields J. z Z.°C . F el 1 1 el Z - Zzi grad C_, Z = grad ¢ . (75) 1 el 1 1 Taking the gradient of Eq. (T4) yields E;Zi grad C . = — Z, grad C, . (76) i For most aqueous solutions, the concentrations of all ions will be small compared to the concentration of water molecules; hence it is assumed that no solubility limits are reached at any point in the elec- trolyte phase. For this case, the coion concentration will vary across the electrolyte film, but there will be no net transfer of coions. This condition is described by the relation JY = 0 = -DeY grad CeY + —RT grad o] , (T?a) where the subscript Y refers to the coion. Thus, the concentration pro- file for coion Y is independent of the diffusion coefficient of the coion. Kataoka, Sato, and Ueyama have solved Eqs. (71), (73), (74), and (77a) analytically for the case where there are two different cations 2L3 (transferring ions).65 Combination of Eqs. (76) and (77a) yields the relation 2 E: ZY CY F d = ——— a z, grad C_. = grad ¢ , (78) and substitution of Eq. (78) into Eq. (75) yields the final relation which describes the electric potential gradient in the electrolyte film. The resulting relation is J . Z z: el 1 F i Dei Ef-grad 0= = — - z: - 5 - . (79a) ZY Y T i ei For the case in which there are more than two transferring ioms, Eq. (71) must be integrated numerically for each ion. If the electrolyte phase is a molten salt, the concentration of transferring ions in the electrolyte film relative to the total concen- tration of ions can be high, and the volume of each ion should be taken into account. For molten salt systems of immediate interest, this effect can be treated approximately by a simple technique In the case of MSBR fuel salt, the electrolyte phase will consist of a mixture of the fluo- rides of Th, Li, Be, U, Pa, Zr, and rare earths. In reductive extraction operations involving molten LiCl, the electrolyte phase will consist of a mixture of the chlorides of these materials (except for beryllium) and of the rare-earth and alkaline-earth fission products. 1In both cases, only cations will exchange between the bismuth and salt phases. As shown in Tables 58 and 59, the equivalent volumes of most of the salts in each individual group are approximately the same (12.1 CmB/equiv for fluorides and 27.1 cmB/equiv for chlorides ). It has also been found that the equiv- alent volumes are linearly additive for fluoride mixtures containing the elements of interest.6u In view of these factors, the concentration of fluoride (or chloride) ion will be assumed to be constant throughout the electrolyte film. Thus CY = constant = l/D , (TTb) 2Ll Table 58. Empirical Molar and Equivalent Volumes of Fluorides at 600°Ca Molar Volume Equivalent Volume Material (cm3/mole) (cm3/equiv) LiF 13.46 13.46 BeF2 23.6 11.8 ThF, 46.6 11.65 UF4 45.5 11.38 ZrF4 47 11.75 YF3 34.6 11.53 LaF3 37.7 12.57 CeF3 36.3 12.1 PrF3 36.6 12.2 SmF3 39.0 13.0 SrF2 30.4 15.2 BaF2 35.8 17.9 %pata taken from ref. 64. Table 59. Molar and Equivalent Volumes for Chlorides at 650°C Molar Volume Equivalent Volume Material (cm3/mole) (cm3/equiv) LiCl 28.6 28.6 SrCl2 55.5 27.8 BaCl2 61.5 30.7 YCl3 76.7 25.6 L3013 72.7 24.3 CeCl3 72.5 24.2 ThC1 112.6 28.2 4 U5 where v = equivalent volume of salt, cmi/equiv. This assumption will not be universally valid for all molten salt systems, but it is believed to be reasonable for cation exchange between molten salt and bismuth since the anions in this instance will usually be much larger than the cations. Equations (75), (76), and (77b) may be combined to obtain an expression that describes the electric potential! gradient for the case where the electrolyte is a molten salt. The resulting relation is L %f grad ¢ = — EET———E;-——— . (79b) i In summary, the relations which define the rate at which components transfer between an aqueous and an organic phase are as follows: D ., si Joi = B (Csr'.l CsiB) ’ (70) Zi Cei F _ - Ll grad o Jos Dei[grad C,. ™+ o grad o) , (71) = e Joi = Jdg1 7 (72) E: Jei %4 = E: Tsi zl ’ (73) i i and J . Z, z: ei i F D . RT grad ¢ = — 5 (79a) i el D Z," Cy + Zl: z.“c_, Similarly, the relations which define the rate at which components transfer between a molten salt and bismuth phase consist of Eqs. (70), (T71), (72), (7%),anc the relation 246 J . Z 2: el 1 %f grad ¢ = «-ni-———Egi——— . (79b) z C . i el Equilibrium relations for the distribution of solutes between salt- metal or aqueous-organic phases can take many forms, and in principle, almost any type could be used with the calculational procedure being con- sidered. A mass action type of equilibrium relation was chosen, which can be expressed as N i Cei Csr Zr C, (E::) = Q. , (80) where Q1 is the equilibrium constant and the subscript r refers to a reference component. Ferris and co-workers65 have measured the equilib- rium distributions of several materials between molten salts and bismuth. Their studies have shown that the activity coefficients for the species in both phases are essentially independent of composition; hence, the equilibrium quotients Qi are essentially constant. Equation (80) is believed to be appropriate for several aqueous systems also. 15.3 Calculational Procedure The rates of transfer of materials between two liquid phases at a given axial position in a packed column are calculated by a trial-and- error integration of Eq. (71) across the electrolyte film. When only two transferring ions are present, the integration can be made analyti- cally in a manner similar to that shown by Kataoka, Sato, and Ueyama.65 When more than two transferring ions are present, a numerical procedure such as the one discussed below is recommended. The bulk concentrations of the transferring materials in both phases, the mass transfer coef- ficient (Di/fis) for each component in both phases, and the equilibrium constants [defined by Eq. (80)] are required for the calculations. The transfer rate for each ion and the concentration of each ion in the electrolyte at the interface are computed by the following procedure: 2L A value is assumed for the flux of each transferring ion through the electrolyte film. The electrolyte film is divided into several thin incre- ments. Equations (79b) and (71) are used to calculate the electric potential gradient and concentration gradi- ents of all components in the increment adjacent to the region where the bulk concentrations occur. The concentrations of the transferring materials at the boundary between the first and second increments of the electrolyte film are calculated using the calculated values for the concentration gradient. Steps 2 and 3 are repeated until values for the concen- trations of the transferring materials are established throughout the electrolyte film. A value is assumed for the concentration of the reference component (which can be any one of the transferring ions ) in the solvent phase at the interface. Equation (80) and the concentrations of the transferring mate- rials in the electrolyte phase at the interface (from step 4) are used to calculate concentrations of all other transferring materials in the solvent phase at the inter- face. The rate of transfer of each ion across the solvent film is calculated using Eq. (70). The sum of the products of the transfer rates and the valences of the components should be zero, as shown in Eq. (73). 1In general, this will not be the case for the first value assumed for the concentration of the reference component in the solvent phase at the interface. The next value for the concen- tration of the reference component is obtained by using Newton's method with Egs. (73) and (80), as follows. For the expression = - 1 £ 2; Z; Doy (CsiB CsiI) ’ (81) f should have the value of zero when the proper value of C is found. The derivative of f with respect to C srl srL is: ., C - £' = 9. 2.0 D, it . (82) i si Zr CSrI The next value to be assumed for CS is then given by r the relation I f? (Cer)n+1 - (Cer)n T F (83) 248 7. Steps 5 and 6 are repeated until the concentration of the reference component in the solvent phase at the interface is kgown within the specified convergence criterion (10°° has been used thus far in this study). 8. The calculated values for tne flux of each transferring component across the solvent film are compared with the assumed values for the respective fluxes across the electrolyte film. If the values are not equal, a new set of values is assumed for the fluxes of the trans- ferring materials through the electrolyte film. The new value for the flux of compoaent i is taken to be the arithmetic average of the previously assumed value and the flux of component i across the solvent f£ilm that was calculated using this value. Steps 1 through 7 are repeated until the difference between the values for each material is less than a specified quantity (1% of the average of the flux values has been used thus far in this study ). The calculational procedure ou:lined above was found to corverge rapidly; however, it is not necessarily the optimum procedure for solving Egs. (70)-(80). After a procedure had been developed for calculating the concentra- tions and fluxes for the transferring components in the electrolyte and solvent phases at a given axial location in a packed column, it was necessary to develop a calculationai procedure that could be used through- out the column. The column was divided axially into a number of incre- ments; the concentration of each component was assumed to be constant throughout a given increment. It was assumed that the electrolyte and solvent phases were in constant volumetric flow throughout the column and that no dispersion occurred in either phase. The interfacial area per unit column volume and the values for the film thickness in the solvent and electrolyte phases were assumed to be constant throughout the column. Material balances for each phase in a given column incre- ment then resulted in the following relations: d Csi(z) Sa —y " —'Vg' Ji(Z), (84) d ¢ (z) . S —5— =~ 1.(2), (85) 249 where Csi(Z) = concentration of component i in the solvent at height Z in the column, C ,(Z) = concentration of component i in the electrolyte at ei height Z in the column, VS = volumetric flow rate of solvent, Ve = volumetric flow rate of electrolyte, S = cross—-sectional area of column, and a = interfacial area per unit column volume. In deriving these relations, it was assumed that transfer of material from the solvent phase to the electrolyte phase resulted in a positive value for the flux. These relations can be expressed in finite dif- ference form as follows: n-1 At 1 - = — =J,.(jAZ 86 c (0 —C_, (L 7_n 1J(J ) , (86) j=0 n-1 A c ,(0)— c .(L) = L2y (342) (87) j=0 where At = Sal., total interfacial area in column, j = column increment number, (j = 0, ...n), n = number of increments in column = L/AZ, AZ = height of column increment, and Jij = flux of component i in increment j. These relations were used in the following manner for calculating concen- trations and fluxes for the transferring materials throughout a column: (1) Values were assumed for the concentration of each component in the solvent phase in the increment from which the solvent phase exits the column. Equations (70)-(79) were used for 250 calculating the fluxes of the transferring materials between the solvent and electrolyte phases in the increment. (2) Equations (86) and (87) were used for calculating the con- centrations of the transferring materials in the solvent and electrolyte phases in the next increment of the column. The fluxes of the transferring components were again calculated in this increment, using Egqs. (74)-(79). This procedure was repeated until concentrations and fluxes for each trans- ferring material had been calculated in each increment of the column. (3) The calculated values for the concentrations of the trans- ferring components in the solvent phase entering the column were compared with values that were specified by the operating conditions. Usually, differences between the calculated and specified values were observed. When the difference was greater than desired, a new set of values was assuned for the concentrations of the transferring materials in the solvent stream leaving the column. Each new value was cal- culated by summing the values for the concentrations of component i in the exit solvent stream with one-half of the difference between the specified concentration of component i in the inlet solvent stream and the calculated value for the concentration of component i in the inlet solvent stream from the last iteration. This procedure was repeated until the calculated and the specified values for the con- centrations of transferring materials in the inlet solvent stream were in agreement. In the future, work will be carried out for calculating rates of mass transfer between solvent and electrolyte phases for a range of operating conditions. Particular attention will be paid to the influ- ence of the electric field on the rate of mass transfer, and to the differences that result from the case where mass transfer rates are assumed to be dependent only on concentration gradients. 251 16. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS R. B. Lindauer L. E. McNeese We have previously described equipment for studying the purification of salt by continuous methods.66 Initial work with this system was directed at the measurement of the flooding rates in a 1l.25~in.-diam, 7-ft-long column packed with 1/4-in. nickel Raschig rings. Flooding data were obtained during the countercurrent flow of molten salt (66-34 mole 7% LiF—Ber) and hydrogen or argon.67 The objective of the present work is to study the continuous reduction of iron fluoride in molten salt by countercurrent contact of the salt with hydrogen in a packed column. We reported previously68 on the first two iron fluoride reduc- tion runs (runs 1 and 2), which were carried out at a temperature of 700°C. Reasonable values for the mass transfer coefficient for the transfer of iron fluoride from the bulk salt to the gas-salt interface were obtained. Operation of the column during these runs was erratic, and the pressure drop across it increased to twice the initial wvalue. The increased restriction was believed to have resulted from precipi~ tation of BeO on the column packing as the result of an accumulation of oxide in the system. During this report period, salt purification studies using 66-34 mole 7% LiF—BeF2 were terminated because of leaks that resulted in the loss of about half of the l4-liter salt charge. The composition of the remaining salt was adjusted to the approximate composition of the proposed MSBR fuel salt (72-16-12 mole % LiF—Ber—ThFA) by addition of sufficient quantities of salt having the composition 72.6-27.4 mole %Z LiF-ThF, and LiF powder to produce 17 liters of salt having the com- 4 position 72.0-14.4-13.6 mole 7% LiF—Ber—ThFA. The newly prepared LiF- Ber—ThF4 salt was then countercurrently contacted with a Hz——IOZ HF mixture in the column in order to remove oxide from the salt. Although considerable oxide was removed, the pressure drop across the packed column was reduced only slightly. This indicated that a significant quantity of oxide still remained in the column. Two flooding runs and 252 one iron fluoride reduction run were then made. During these runs, the pressure drop across the column increased to the point where operation of the system became difficult. The packed column was then filled with molten salt, and an HF—H2 stream was allowed to contact the static salt charge for a period of 18 hr in order to remove the oxide from the column. This operation was successful in reducing the pressure drop across the column to approximately the value observed after the first two iron fluoride reduction runs. Eight additional iron fluoride reduction runs were subsequently completed. During these runs, operation of the system was smooth, and there was no increase in pressure drop. The results obtained by analyzing salt samples from the runs for iron were incon- sistent, probably because of the low iron concentration in the system although sample contamination was suspected in some cases. These oper- ations are described in greater detail in the remainder of this section. 16.1 Removal of Oxide from Salt by Countercurrent Contact with an HF--H2 Gas Stream We have found that routine measurement of the pressure drop across the packed column is a useful means for detecting the buildup of mate- rials such as metallic iron or insoluble oxides in the column. As shown in Fig. 93, the pressure drop across the column with an argon flow rate of 5 liters/min increased from 5.6 in. H20 to 12.0 in. HZO during previously reported68 experiments with the LiF—BeF2 salt (flooding runs No. 3-10). Since less than 6 g of iron would have been reduced during this time, the restriction is believed to have been due mainly to an accumulation of insoluble Be0O on the column packing. The oxide in the salt could have originated from several sources. Although the gas purification and supply equipment for the hydrogen and argon was used throughout all the experiments, the purification traps were not regenerated before the experimental work was initiated. Instrumentation for monitoring the moisture content of the gases was installed after 1 month of operation, and an oxygen analyzer was installed 3 months after operation of the system began. Typical water and oxygen contents of the unpurified gases were 15 ppm and 2 ppm, respectively, and are much *UT!t "¢6 814 /sxa:n ¢ 3o 93eyg Mmo1d uolay ue yiim doig sanssaid uwnion 0l61=-31vQ COLUMN PRESSURE DROP, in, of H20 - - n ) o H O 3] o w o w S o o I | | 1 1 1 g o AFTER FLOODING RUN NO, 3 o N AFTER FLOODING RUN NO. IO o @ AFTER REDUCTION RUN NO. 2 ® Q 0 AFTER COUNTERCURRENT HZ-HF TREATMENT w § AFTER REDUCTION RUN NO. 3 o o < AFTER 3-DAY COLUMN CLEANOUT WITH H2-HF n i AFTER REDUCTION RUN NO. 4 3 ® AFTER REDUCTION RUN NO. 5 = ® AFTER REDUCTION RUN NO. € o § ® AFTER REDUCTION RUN NO. 10 ® < o AFTER REDUCTION RUN NO. || S i 1 1 i | ! Sh—-cL O9MQd TINMO £Ge 25k too low to.account for the quantity of oxide (several hundred ppm) that had accumulated in the salt during the six months of operation. Some contamination undoubtedly occurred during repair of plugged vent lines and possibly as the result of insufficient purge rates when the system was vented to the atmosphere between runs. Recent practice has been to close the system vent valve between experiments, while maintaining a sufficient purge rate to keep the system at a few psi of positive pressure (a small amount of gas leakage occurs through the vent valve). In order to remove the oxide that had accumulated in the salt, the salt was contacted with a 10-90 mole 7% HF-—H2 gas stream for 165 min at average salt and gas flow rates of 103 cm3/min and 5 liters/min respec- tively. Some difficulty was experienced in maintaining a constant gas sample flow rate to the analyzer used for determining the concentration of HZO in the gas stream leaving the column; however, during a period of steady flow near the end of the run, the analyzer indicated that 100 ppm of oxide was being removed from the salt as it passed through the column. The HF utilization at this point was 15Z. Treatment of the salt with the HF—H2 from 12.0 in. H,O0 to 10.3 in. H 2 2 min, which indicated that the column still contained an appreciable stream reduced the pressure drop across the column O with an argon flow rate of 5 liters/ quantity of oxide. 16.2 Removal of Oxide from Column Following the countercurrent contact of the salt with an HZ—HF gas stream, two flooding runs and one iron fluoride reduction run (No. 3) were carried out. During the second flooding run, it was necessary to heat the molten salt (downstream from the column) to a much higher temperature (650°C on the bottom of the filter housing) than usual in order to maintain the flow of salt through the filter. This could have been due to the presence of oxide in the salt since the solubility of BeO in salt increases as the temperature is increased. During the reduction run, the pressure drop across the column increased, as had been observed in the two previous runs made with salt having the 255 composition 66-34 mole % LiF—Ber. However, a test with argon after reduction run No. 3 showed an even higher column pressure drop (see Fig. 93) than had been observed during the run. An HF-—H2 gas stream was then passed through a static charge of molten salt in the column in order to provide more opportunity for the dissolution and hydro- fluorination of insoluble oxide. The column was filled with salt to the top of the packing (about 1-1/2 liters) by observing the hydro- static pressure at the bottom of the column. For the first 6.5 hr, the column was held at 650°C and a 29-71 mole ¥ HF--H2 gas mixture was passed through the column at the rate of 8.3 liters/min. The concen- tration of HF in the stream was higher than planned because of an error in the size of the capillary used for measuring the HF flow rate (2.4 liters/min). Only about 1.5% of the HF was utilized at this HF flow rate, and oxide was removed from the column at the rate of 0.024 mole/hr. Passage of the I-IF—H2 stream through the column was continued for 6.5 hr, after which it was terminated at the end of the day shift. When the treatment was resumed the following day, the column was cooled to 590°C in order to increase the concentration of HF in the salt, and the HF flow rate was reduced to 750 cm3/min. The oxide removal rate (as indicated by the water analyzer in the off-gas stream) was about the same as had been observed previously, and the HF utilization increased to about 5%Z. This compares favorably with the value obtained during countercurrent flow of salt and gas through the column (i.e., 15%). The treatment was resumed on the third day with a column temperature of 700°C to facilitate the dissolution of any remaining oxide. The output from the water analyzer decreased to a value below the reference value, which indicated that the NaF trap in the sample stream had become saturated and was allowing HF to reach the sample electrode. Since the rate of oxide removal could no longer be measured, the oxide removal operation was terminated after the salt had been contacted with the HF—H2 gas stream for a total of 18 hr. The remainder of the batch of salt was then passed countercurrent to a 10-90 mole Z HF—H2 gas stream in the column. On completion of these operations, the pressure drop across the column had been reduced to 14.8 in. H20° After two additional iron 256 fluoride reduction runs (runs No. 4 and 5), the column pressure drop had decreased to 9 to 10 in. H20 (probably from dissolution of insoluble oxide remaining on the packing) and remained in this range for the rest of the report period. 16.3 Iron Fluoride Reduction Runs During this report period, nine iron fluoride reduction runs were carried out with salt having the composition 72.0-14.4-13.6 mole % LiF- Ber—ThF4. Table 60 summarizes the data from these runs, along with the data from the two previous reduction runs (Nos. 1 and 2) made with salt having the composition 66-34 mole % LiF-BeF,. During the nine current runs, the concentration of iron fluoride in the salt was decreased from 220 ppm to 70 ppm or less. Operation of the equipment was satisfactory following run 3 (after removal of oxide from the column); however, analyses of the resulting salt samples for iron produced inconsistent data. The reported data for four of the last eight runs showed an increase in the iron fluoride content of the salt. Two of the reported values exceeded the total iron concentration believed to be possible (276 ppm). These unusually high analyses may have been caused by con- tamination of the samples during their removal from the nickel samplers or by the diminished accuracy of the analytical method as the iron con- centration is decreased. 16.4 Calculated Values for the Mass Transfer Coefficient and the Reaction Rate Constant During the Reduction of Iron Fluoride The rate of reduction of FeF, by reaction with hydrogen can be affected by a number of factors, 2One of the objectives of the present work is to determine the rate-controlling steps involved in the reduction of FeF2 in a packe§8column, and to evaluate the associated rate constants. We have previously = carried out mathematical analyses for two limiting cases: 257 Table 60. Data From Iron Fluoride Reduction Runs Column Temperature, 700°C Gas Flow Analyses of Rate Salt Flow Filtered Samples Percent of Run (std. liters/min) Rated (ppm_iron) Batch No. Date H2 Ar (cm3/min) Feed Product Contacted 1 7/27 20.0 -— 100 425 307 68.5 2 7128 13.5 -- 100 307 228 27.6 3 10/22 16.6 -— 100 220 158 75.3 4 11/3 14.6 ’ - 210 158 373 84.7 5 11/4 18.2 - 161 373 137 68.8 6 11/5 4.5 4.5 106 137 110 81.8 7 11/6 3.9 3.9 142 110 70 81.8 8 11/9 24,0 - 103 70 75 86.5 9A 11/13 4.5 4.5 93 b c 75 55 79.4 9B 11/13 3.5 3.5 136 d b 10 11/17 14.1 -— 105 69 77 81.8 11 11/19 3.1 4.5 117 207 339 81.2 104° #The first two runs used LiF-BeF salt; the remaining runs used LiF—BeFZ—ThF4 salt. 2 bAverage of flowing-stream samples. “A value of 100% is used in calculating kpa since samples are of flowing-stream tvpe. dCalculated from flowing-stream samples from previous run. 258 (1) the case in which the rate of reduction is limited by the rate of transfer of Fer to the gas-salt interface from the bulk salt, and (2) the case in which the rate of reduction is limited by the rate of reaction between H2 and FeF2 dissolved in the salt at the gas-salt interface. The associated rate constants can be evaluated by the relations given below. For the case in which the rate of reduction is limited by the rate of transfer of FeF2 from the bulk salt to the gas-salt interface, the product of the mass transfer coefficient k2 and the interfacial area a (which is not known) is given as: X k,a = XH In Xi , (88) o where kfl = mass transfer coefficient for the transfer of FeFs from the bulk salt to the salt-gas interface, moles/sec*cm?, a = gas-salt interfacial area per unit column volume, cm2/cm3, A = cross-sectional area of the column, cmz, L = salt flow rate, moles/sec, H = column height, cm, Xi = concentration of FeF2 in salt fed to column, mole fraction, XO = concentration of FeF2 in salt leaving column, mole fraction. This relation is valid only under conditions such that the concentration of FeF2 in salt that would be in equilibrium with the HF--H2 mixture adjacent to the salt being considered (X*) is small in comparison with the concentration of FeF2 in the bulk salt in the region being considered. For the case in which the rate of reduction is limited by the rate of reaction between FeF2 in molten salt and H2 in the gas at the gas-salt interface, the rate constant is given by the relation L Xi k = ————— 1ln — , (89) s KH HA Py Xo 259 where -1 reaction rate constant, sec ~, = i Henry's law constant for hydrogen in salt, moles/cm3-atm, ol partial pressure of hydrogen in gas, atm, 0 N [ and the other quantities are as defined above. Values for the mass transfer coefficient and the reaction rate con- stant were calculated for the current runs for which meaningful results could be obtained. These values, along with values from the two runs reported previously, are summarized in Table 61. Since the value for X* is negligible compared with the FeF, con- centration in the bulk of the salt for the runs considered, thezexPression used for calculating kza should be valid. Also, since the partial pressure of hydrogen (sz) is nearly constant throughout the column, the expression given above for calculating kS should be valid. Although the inlet hydrogen partial pressure was reduced to 0.5 atm in the last three runs for which rate constants are shown, the variation in the cal- culated constants is not sufficiently large to determine whether the rate of iron fluoride reduction is countrolled by the rate of transfer of FeF, to the gas-salt interface or by the rate of reaction between 2 FeF2 and hydrogen at the interface. This question will be resolved after additional data are obtained. Table 61. Calculated Values for the Mass Transfer Coefficient and the Reaction Rate Constant fer Runs R-1 and R-2 Mass Transfer Controlled Reaction Rate Controlled FeF, Concentration kya 3 X* Py Run (mole fraction x 10 ) (moles/sec-cm (mole fraction (se2-1 2 (atm) . No. Xq X, x 109) x 10%) x 10%4) Inlet Exit Equilibrium’ R-1 5.13 3.70 2.4 0.033 2.4 1.0 0.9987 0.9850 R-2 3.70 2.75 5.4 0.198 5.4 1.0 0.9967 0.9959 R-3 2.65 1.90 1.9 0.035 1.9 1.0 0.9993 0.9974 R-6 1.65 1.33 1.2 0.042 2.5 0.5 0.4995 0.4969 R-7 1.33 0.84 2.9 0.170 5.7 0.5 0.4989 0.4975 R~9A 0.90 .66 1.8 0.032 3.7 0.5 0.4995 0.4978 Cace a Calculated from the composition of the exit gas. bCalculated from the FeF2 composition in the salt leaving the column. 10. 11. 12. 13. 4. 15. 261 17. REFERENCES R. G. Ross, C. E. Bamberger, and C. F. Baes, Jr., MSR Program Semiann. Progr. Rept. Aug. 31, 1970, ORNL-4622, pp. 92-95. J. C. Mailen, MSR Program Semiann. Progr. Rept. Feb. 28, 1971, ORNL- L6T6, pp. 245-48. C. E. Bamberger and C. F. Baes, Jr., J. Nucl. Mater. 35, 177 (1970). M. J. Bell and L. E. McNeese, MSR Program Semiann. Progr. Rept. Aug. 31, 1970, ORNL-4622, pp. 202-8. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 5, ORNL-TM-3140 (October 1971), pp. 2-15. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (December 1971), pp. 4-13. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 3-2L. L. E. McNeese, MSR Program Semiann. Progr. Rept. Feb. 28, 1970, ORNL- 4548, pp. 282-88. M. J. Bell and L. E. McNeese, MSR Program Semiann. Progr. Rept. Aug. 31, 1970, ORNL-4622, pp. 199-202. F. J. Smith and C. T. Thompson, Ibid., pp. 207-8. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. 16-29. L. E. McNeese, Enginéefing Development Studies for Molten-Salt Breeder Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 2h4-30. R. L. Jarry, W. H. Gunther, and J. Fischer, The Mechanism and Kinetics of the Reaction Between Nickel and Fluorine, ANL-6684 (August 1963 ). W. H. Gunther and M. J. Steindler, Laboratory Investigations in Support of Fluid-Bed Fluoride Volatility Processes. Part XIV. The Corrosion of Nickel and Nickel Alloys by Fluorine, Uranium Hexa- fluoride, and Selected Volatile Fission Product Fluorides at 500°C, ANL-7241 (December 1966 ). Chemical Engineering Division Summary Report, July-September, 1958, ANL-5924 (1958). 16. 17. 18. 19. 20. 21. 22. 23, 2kh. o5, 26. 27. 28. 29. 30. 31, 32, 262 Chemical Engineering Div. Semiann. Rept., January-June, 1964, ANL- 6900 (August 1964 ). P. D. Miller and W. E. Berry, A Survey of Corrosion in the Fluidized- Bed Volatility Process, BMI-X-362 (November 1965 ). F. T. Miles et al., Progress Report of the Reactor Science and Engi- neering Department, BNL-176 (March 1952), pp. 17-18. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 2, ORNL- 22-43. MSR Program Semiann. Progr. Rept. Feb. TM-3137 (February 1971), pp. 29, 1968, ORNL-4254, pp. 260-63. H. F. Bauman, ORNL, personal communication, August 1970. L. E. McNeese, Engineering Development Reactor Processing No. 5, ORNL-TM-3140 L. E. McNeese, Engineering Development Reactor Processing No. 6, ORNL-TM-3141 L. E. McNeese, Engineering Development Reactor Processing No. 8, ORNL-TM-3258 L. E. McNeese, Engineering Development Reactor Processing No. 4, ORNL-TM-3139 L. E. McNeese, Engineering Development Reactor Processing No. 5, ORNL-TM-3140 L. E. McNeese, Engineering Development Reactor Processing No. 6, ORNL-TM-31L41 L. E. McNeese, Engineering Development Reactor Processing No. 8, ORNL-TM-3258 Studies for Molten-Salt Breeder (October 1971), pp. 15-22. Studies for Molten-Salt Breeder (December 1971), pp. 23-39. Studies for Molten-Salt Breeder (May 1972), pp- 3-21. Studies for Molten-Salt Breeder (August 1971), pp. 38-83. Studies for Molten-Salt Breeder (October 1971), pp. 22-30. Studies for Molten-Salt Breeder (December 1971), pp. 13-23. Studies for Molten-Salt Breeder (May 1972), pp. 31-58. M. E. Whatley et al., Unit Operations Section Quarterly Progress Report, April-June 1968, ORNL-L365 (February 1970), pp. 33-45. V. G. Levich, Physiochemical Hydrodynamics, pp. 47-52, translated by Scripta Technica, Inc., Prentice-Hall, H. P. Meissner, Process and Systems in Englewood Cliffs, N. J., 1962. Industrial Chemistry, Chap. 9, Prentice-Hall, Englewood Cliffs, N. J., 1971. G. D. Towell, C. P. Strand, and G. H. Ackerman, "Mixing and Mass Transfer in Large Diameter Bubble Columns," A.I.Ch.E. J., Chem. Eng. Symp. Ser., No. 10, p. 97 (1965). 35 3. 35 36. 57 58. 29- L0, h1. Lo, Lz, L. L5. L46. b L8. 4g. 263 Y. Aoyama, 0. Katsunori, K. Koide, and H. Kubota, "Liquid Mixing in Concurrent Bubble Columns,"” J. Chem. Eng. Japan 1(2), 158 (1968). K. Imafuku, T. Wang, K. Koide, and H. Kubosta, "The Behavior of Sus- pended Solid Particles in the Bubble Column," J. Chem. Eng. Japan 1 (2), 153 (1968). Y. Ohki and H. Inoue, "Longitudinal Mixing of the Liquid Phase in Bubble Columns," Chem. Eng. Sci. 25, 1 (1970). K. B. Bischoff and J. B. Phillips, "Longitudinal Mixing in Orifice Plate Gas-Liquid Reactors," Ind. Eng. Chem., Process Design Develop. 5(4), 416 (1966). S. H. Eissa, M. M. El-Halwagi, and M. A. Saleh, "Axial and Radial Mixing in a Cocurrent Bubble Column,” Ind. Eng. Chem., Process Design Develop. 10(1), 31 (1971). T. Reith, S. Reuken, and B. A. Israel, "Gas Holdup and Axial Mixing in the Fluid Phase of Bubble Columns," Chem. Eng. Sci. 23, 619 (1968 ). L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 1, ORNL-TM-3053% (November 1970}, pp. 1l-14. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (December 1971), pp. 73-T9. L. E. McNeesé, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. T, ORNL-TM—525f_zFebruary 1972), pp. 52-58. L. E. McNeese, Engineering Development Studies for Molten-~Salt Breeder Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 64-89. D. T. Goldman and J. R. Roesser, "Chart of the Nuclides," 9th ed., 1966. N. E. Holden and F. W. Walker, "Chart of the Nuclides," 10th ed., 1968. H. E. McCoy, personal communication to L. E. McNeese (February 1971). L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. 29-L6. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 58-63. Ibid., p. 61. J. B. Lewis, Chem. Eng. Sci. 3, 248-59 (1954 ). J. B. Lewis, Chem. Eng. Sci. 3, 260-78 (1954). 51. 55 Sh. 55 56. o~ 58. 29 60. 61. 62. 63. 6L 65. 66. 67 . 68. 264 W. J. McManamey, Chem. Eng. Sci. 15, 251-54 (1961). G. R. A. Mayers, Chem. Eng. Sci. 16, 16-75 (1966). K. F. Gordon and T. K. Sherwood, Chem. Eng. Progr.,Symp. Ser. 50(10), 15-23 (195h4). D. R. Olander, Nucl. Sci. Eng. 1k, 287-94 (1962). J. Prochazka and J. Bulicka, Session VA: Transport and Interfacial Phenomena, Inst. Chem. Proc. Fund. Prague, Czechoslovakia (April 1971). Jaroslav Prochazka, personal communication to H. O. Weeren (Mar. 2k, 1972). J. S. Watson and L. E. McNeese, Unit Operations Section Quarterly Progress Report, July-September 1968, ORNL-4366 (April 1970), pp. 57-98. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 5, ORNL-TM-3140 (October 1971), pp. 102-3. P. A. Carroad and A. C. Meyerey Flooding Characteristics of a Counter- current Mercury-Water Packed Column, ORNL-MIT-115 (Nov. 18, 1970). R. Schldgl and F. Helfferich, J. Chem. Phys. 26, 5 (1957). J. P. Copeland, C. L. Henderson, and J. M. Marchello, A.I.Ch.E. J. 13, bho (1967). J. C. R. Turner and C. B. Snowdon, Chem. Eng. Sci. 23, 1009 (1968 ). T. Kataoka, M. Sato, T. Nishiki, and K. Ueyama, Chem. Eng. Japan 33, 59 (1966). S. Cantor, Reactor Chem. Div. Ann. Progr. Rept., Dec. 31, 1965, ORNL- 5913, pp. 27-29. L. M. Ferris, J. C. Mailen, J. J. Lawrance, F. J. Smith, and E. D. Nogueira, J. Inorg. Nucl. Chem. 32, 2019-35 (1970). L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (December 197l), pp. 59-72. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 7, ORNL-TM-3257 (February 1972), pp. L46-52. L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 8, ORNL-TM-3258 (May 1972), pp. 97-106. oo~V MLWND 63. 64. 65. 66. 67. 68. 69. 70. 71. 72-73. 74. 265 INTERNAL DISTRIBUTION S. E. Beall 38-39. M. W. Rosenthal M. R. Bennett 40. A. D. Ryon R. E. Blanco 41, W, F. Schaffer, Jr. R. E. Brooksbank 42. J. H. Shaffer K. B. Brown 43, M. J. Skinner W. L. Carter 44, Martha Stewart H. D. Cochran, Jr. 45, 0. K. Tallent F. L. Culler 46. D. B. Trauger D. E. Ferguson 47. W. E. Unger L. M. Ferris 48. C. D. Watson W. R. Grimes 49, J. S. Watson P. A. Haas 50. A. M. Weinberg B. A. Hannaford 51. W. M. Woods J. R. Hightower, Jr. 52. R. G. Wymer C. W. Kee 53. E. L. Youngblood J. A. Klein 54-55. Central Research Library R. B. Lindauer 56. Document Reference Section H. G. MacPherson 57-59. Laboratory Records L. E, McNeese 60. Laboratory Records, RC J. P. Nichols 61. Y-12 Document Reference Section E. L. Nicholson 62. ORNL Patent Office J. H. Pashley (K-25) EXTERNAL DISTRIBUTION J. A. Accairri, Continental 0il Co., Ponca City, Oklahoma 74601 M. J. Bell, Effluent Treatment Systems Branch, USAEC, Washington, D. C. 20545 Mail Station Beth. 008 Delonde R. DeBoisblanc, Ebasco Services, Inc., 2 Rector Street, New York, New York 10006 Norton Haberman, USAEC, RDT, Washington, D. C. 20545 T. R. Johnson, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, Illinois 60439 R. D. Pierce, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, Illinois 60439 J. Roth, Combustion Engineering Inc., Prospect Hill Road, Windsor, Connecticut 06095 D, D. Sood, Radiochemistry Division, Bhabha Atomic Research Center, Bombay 85, India N. Srinivasan, Head, Fuel Reprocessing Division, Bhabha Atomic Research Center, Trombay, Bombay 74, India TIC, Oak Ridge, Tennessee Research and Technical Support Division, AEC, ORO