= B “\"_fl B P I--.——lfl.l"' ...... A L] — —_— — e 111 — OA i (e ¥ unie C 1B 05BBET0 3 g S NUCLEAR DIVISION U.S. ATOMIC ENERGY COMMISSION % / ORNL- TM - 3257 FEB 10 1972 DATE ISSUE: _ SR e ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT BREEDER REACTOR PROCESSING NO. 7 L. E. McNeese NOTICE This document conicins information of a preliminary nature ond was prepored primarily for internal use at the Oak Ridge MNational Loboratory. 1t is subject to revision or correction and therefore does not represent a tinal report. ORNL-TM-3257 Contract No. W-TL05-eng-26 ENGINEERING DEVELOPMENT STUDIES FOR MOLTEN-SALT BREEDER REACTOR PROCESSING NO. T L. E. McNeese FEBRUARY 1972 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37830 operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION ii Reports previously issued in this series are as follows: ORNL-4365 ORNI-L4366 ORNL-TM-3053 ORNL~-TM-3137 ORNL-TM-3138 ORNL~-TM-3139 ORNL-TM-31k0 ORNL-TM-31k1 Period Period Period Period Period Period Period Period ending ending ending ending ending ending ending ending July 1968 September 1968 December 1968 March 1969 June 1969 September 1969 December 1969 March 1970 iid SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN,A MILD-STEEL FACILITY © v v v v v 6 o o o o o o o o« o o o s o s o o o o o CONTENTS Page SUMMARIES ¢« v v v v v o v o o o o o o o o o o o o o o o v 1. INTRODUCTION « « & v o ¢ o o v o v v v o e e e e e e e 1 2. ANALYSIS OF THE FLUORINATION--REDUCTIVE EXTRACTION AND METAL TRANSFER FLOWSHEET . . o ¢« ¢« & ¢ v v v o v v 0 v v 4 o o v o v W 2 2.1 Distribution of Rare-Earth and Alkaline-Earth Elements Be- tween Molten Salt and Bismuth Containing Reductant 2 2.2 Isolation of Protactinium Using Fluorination~~Reductive Extraction « o v v v 6 6 e v e e b e e e e e e e e e e 3 2.3 Removal of Noble Metals with the Fluorination--Reductive Extraction Flowsheet . « . + ¢« ¢« v ¢« ¢ o o o o 10 2.4 Halogen Removal in the Uranium Removal System . . . . . 12 3. DEVELOPMENT OF A FROZEN~-WALL FLUORINATOR: DESIGN CALCULATIONS FOR INDUCTION HEATING OF A FROZEN-WALL FLUORINATOR . . . . 16 3.1 Effects of Wall Temperature, Current, Frequency, and Fluori- nator Diameter on the Thickness of the Frozen Film . . . . 16 3.2 Control of Frozen Film Thickness, and Approximate Dynamics Of Freezing .+ « o o o o o o s o o o o o o o o o o 23 3.3 Power Requirements for an Experimental Fluorinator . . 26 4. DEVELOPMENT OF THE METAL TRANSFER PROCESS .+ + v v o + . . 29 4.1 Equipment and Materials Used for Experiment MTE-1 . 29 4.2 Experimental Procedure .+ . « « « & o o + o o« o« o o o . 33 4.3 Experimental ResultsS .+ « o « & v « v o o & o o o« o o o o 35 4.4 Postoperational Equipment Examination . . . L5 4.5 Design and Testing of a Carbon-Steel Pump Having Molten- Bismuth Check Valves .« v « +v v v o o o o o o o o o o L6 5. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS L6 5.1 Batch Treatment of Salt for Oxide Removal . . . 48 5.2 Measured Flooding Rates During Countercurrent Flow of Molten Salt and Hydrogen or ATgOn . . « « « « o o« « o+« s Lo 6. 52 iv CONTENTS (Continued) 6.1 Preparation for Mass Transfer Experiments . . . 6.2 Mass Transfer Experiment UTR-1 . o &+ 4 & « o 6.3 Mass Transfer Experiment UTR-2 . . « . + . . . MEASUREMENT OF AXIAL DISPERSION COEFFICIENTS IN PACKED .1 Experimental Results .+ « ¢ o o« &« « ¢« ¢« o o« o & T.2 Comparison of Results with a Published Correlation REFERENCES . ¢ ¢ ¢ v ¢ v v v v v v v v v v o 0 o o SUMMARIES ANALYSIS OF THE FLUORINATION--REDUCTIVE EXTRACTION AND IIETAL TRANEFER FLOWSHEET Recently obtained data on the distribution of several rare earths between molten salt and bismuth containing reductant have been used “n additional calculations made to identify the important operating para- meters in the flowsheet and to determine the optimum operating condi- tions. The behavior of fission products more noble than uranium in the fluorination--reductive extraction process has also been considered, and the effects of these materials on the reactor breeding ratio have been calculated. Calculations were also carried out to determine the heat generation rates associated with the decay of halogen fission products that will be removed by fluorination. DEVELOPMENT OF A FROZEN-WALL FLUORINATOR: DESIGN CALCULATIONS FOR INDUCTION HEATING OF A FROZEN-WALL FLUORINATOR Calculations were made to show the effects of coil current, fre- quency, wall temperature, and fluorinator diameter on the thickness of the frozen salt film in a continuous fluorinator that employs high-fre- quency induction heating. An approximate analysis of the dynamics of frozen film formation was carried out, and methods for controlling the frozen film thickness were examined. Calculations were also carried out to estimate the power requirements for a 5-ft-long experimental fluorinator that employs rf heating. DEVELOPMENT OF THE METAL TRANSFER PROCESS The first engineering experiment (MTE-1) for studying the removal of rare earths from single-fluid MSBR fuel salt by the metal transfer process was completed during this reporting period. The main objective of the experiment was to demonstrate the selective removal of rare earths (La and Nd) from a fluoride salt mixture containing thorium fluoride. The experiment was performed at 660°C in a 6-in.-diam carbon-steel vessel, vi which contained two compartments interconnected at the bottom by a pool of molten bismuth that was saturated with thorium. One compartment con- tained fluoride salt to which 2 mCi of llwl\Td and a sufficient quantity of LaF3 to produce a concentration of 0.38 mole % had been added. The second compartment contained LiCl. The distribution coefficients for the rare earths between the fluo- ride salt and the thorium-saturated bismuth were relatively constant throughout the run and were in agreement with expected values. The dis- tribution coefficients for the rare earths between the LiCl and the thorium- saturated bismuth were higher than anticipated during the first part of the run but approached the expected values near the end of the run. Approximately 50% of the lanthanum and 25% of the neodymium originally present in the fluoride salt were removed during the run. The rates at which the rare earths were removed are in close agreement with expected re- moval rates; however, the rare earths did not collect in the lithium-bismuth solution (with which the LiCl was contacted) as expected. Instead, most of the rare earths were found in a 1/8-in.-thick layer of material located at the interface between the LiCl and the thorium-saturated bismuth. It is believed that the presence of oxide in the system may account for the ac- cumulation of the rare earths at this point. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS The system was charged with 28 kg of salt (66-3L4 mole % LiF-Bng), and ten flooding runs were carried out using hydrogen and argon. During these runs, salt flow rates of 50 to 400 cm3/min were used with argon and hydrogen flow rates of up to 7.5 and 30 liters/min, respectively. The temperature of the column was 700°C in each case. The pressure drop across the column increased linearly with increased gas flow rate; however, the salt flow rate had only a minor effect on pressure drop. The maximum flow rate possible with the present system is about 19% of the calculated flooding rate. vii SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL FACILITY Following routine H,-HF treatment of the bismuth and the salt in the system, the phases wire transferred to the respective feed tanks. Then 90 g of purified LiF—UFh eutectic salt was added to the salt phase to produce a UFh concentration of about 0.0003 mole fraction for the first mass transfer run (UTR-1). Hydrodynamic performance during the 140-min run was excellent, and ten pairs of bismuth and salt samples were taken. The column was operated at 62% and 76% of flooding (at a bismuth-to-salt volumetric flow rate ratio of unity); nevertheless, vir- tually none of the uranium was extracted from the salt due to an oper- ational difficulty that prevented reductant from being added to the bis- muth. Dissolution of thorium in the bismuth feed tank in preparation for the second mass transfer experiment proceeded slowly as the result of poor mixing in the tank. In run UTR-2, 95% of the uranium was extracted from the salt. The run was made with a 200% excess of reductant over the stoichiometric requirement and with bismuth and salt flow rates of 247 ml/min and 52 ml/min, respectively. These flow rates are equivalent to about 77% of flooding. This experiment represents the first known demonstration of the continuous extraction of uranium from molten salt into bismuth containing reductant. The results indicate that high uranium removal efficiencies can be obtained in a packed column having a reasonable length. MEASUREMENT OF AXTIAL DISPERSION COEFFICIENTS IN PACKED COLUMNS We have continued our measurements of axial dispersion in packed col- umns during the countercurrent flow of fluids having high densities and a high density difference. These experiments (which use mercury and water) were intended to simulate the conditions in packed columns through which bismuth and molten salt are in countercurrent flow. Results reported viii previously for a 2-in.-ID column packed with 3/8-in. Raschig rings are compared with data obtained during this reporting period for 1/4-in. Raschig rings, 1/4-in. solid cylinders, and 1/2-in. Raschig rings. In each case, the axial dispersion coefficient was independent of the dis- persed-phase (mercury) flow rate. Dispersion coefficients for 3/8- and 1/2-in. packing were also independent of the continuous-phase (water) flow rate; their values were 3.5 and 4.8 cm2/sec, respectively. Data for the 1/4-in. packing indicate that the dispersion coefficient is in- versely proportional to the continuocus-phase flow rate. The data ob- tained during this study are compared with a published correlation of axial dispersion coefficient data. The present data are found to be in good agreement with the published correlation; this is remarkable since the correlation was developed from data obtained with systems having density differences between one and two orders of magnitude less than the density difference of the mercury-water system. 1. INTRODUCTION A molten-salt 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 develop- ing 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 describe: (1) optimized conditions for operation with the combined flowsheet that utilizes both fluorination-- reductive extraction and the metal transfer process, and results of calcu- lations showing the effect of noble-metal removal time on reactor breeding performance and the heat generation rates associated with decay of the halogen fission products; (2) results of calculations that show the sen- sitivity of the frozen film thickness in a continuous fluorinator heated by high-frequency induction heating to coil current, frequency, wall temp- erature, and fluorinator diameter; (3) results of the first engineering experiment for demonstrating the metal transfer process for removal of rare-earth fission products from fluoride salt mixtures; (4) studies of the continuous purification of salt; (5) experiments made in a mild-steel reductive extraction facility to demonstrate the extraction of uranium from molten salt by countercurrent contact with bismuth containing reduc- tant; and (6) measurements of axial dispersion in packed columns during the countercurrent flow of mercury and water. This work was carried out in the Chemical Technology Division during the period April through June 1970. 2. ANALYSIS OF THE FLUORINATION--REDUCTIVE EXTRACTION AND METAL TRANSFER FLOWSHEET M. J. Bell L. E. McNeese A flowsheet that uses fluorination--reductive extraction and the metal transfer process for removing protactinium and the rare earths from the fuel salt of a single-fluid MSBR has been described previously.l Calculations to identify the important operating parameters in this flow- sheet and to determine the optimum operating conditions have been continued using recently obtained data on the distribution of several rare earths between molten salt and bismuth containing reductant. The behavior of fission products more noble than uranium in the fluorination--reductive extraction process has also been considered, and the effects of these materials on the reactor breeding ratio have been determined by means of calculations. Calculations were also carried out to determine the heat generation rates associated with the decay of halogen fission products that will be removed by fluorination. These items are discussed in detail in the remainder of this section. 2.1 Distribution of Rare-Earth and Alkaline-Earth Elements Between Molten Salt and Bismuth Containing Reductant Ferris and co—workers2 have continued to measure the equilibrium distribution of fission product and actinide elements between molten salt and bismuth containing reductant. They have found that, at a given temperature, the distribution coefficient for element M, defined as _ mole fraction of M in bismuth phase M mole fraction of M in salt phase > can be expressed as * log DM = n log XLi + log KM . where XLi = the mole fraction of lithium in the bismuth phase, n = the valence of M in the salt phase, and log KM* = a constant. Plots of the log KM* values vs reciprocal absolute temperature are linear over the temperature range 625-750°C, as shown in Fig. 1. Thus, the temp- erature dependence of log KM* can be expressed as log KM* = A + B/T. Val- ues of the constants A and B used in the present flowsheet calculation are shown in Table 1 for several elements. These data indicate that the dis- tribution of the rare earths is relatively insensitive to temperature and that the distribution coefficients for a given element are about the same, regardless of whether lithium chloride or lithium bromide is used as the salt phase. 2.2 Isolation of Protactinium Using Fluorination--Reductive Extraction Calculations were made for selecting optimum operating conditions for the protactinium isolation system. Optimum conditions were tentatively as- sumed to be those resulting in the minimum partial fuel cycle cost. The partial fuel cycle cost includes the following components of the fuel cycle cost which are associated with the isolation of protactinium: (1) bismuth and uranium inventories in the protactinium decay tank, (2) the loss of bred uranium resulting from inefficient protactinium isolation, (3) the T cost of 'Li reductant required to extract uranium and protactinium from the fuel salt, and (4) the cost of BeF,, and ThF), which must be added to the system in order to maintain a constant fuel salt composition. An in- terest rate of 14% per annum was used to compute inventory charges, and the value of 233U was taken to be $l2/g. The following costs were used for chemicals: bismuth, $5/1b; ThFh, $6.50/1b; Bng, $7.50/1b; and 7Li metal, $55/1b. Values that were obtained for the partial fuel cycle cost include only those charges directly related to the isolation of protactinium and include no contribution either for fluorination of the fuel salt to remove ORNL DWG 70-10,997 TEMPERATURE (°C) 10 750 700 650 1 I l' | I 1 1 ll T I I 1 l] 1 I 1 1 9 — Nd 8| A La _ 7 o 6 |- _ «Z | i x o 5 - ° 4 e —— Sm 3 — + — —— + — v v v a 0 OO0 EU 2 lo— o - s ——— Ba ! = —— i 1 — r— — o 1 ] ll ] l 1 1 Il i l ] ] 11 | l 1 } 9.5 10.0 10.5 1.0 104/ T(°K) Fig. 1. Effect of Temperature on the Values of log K& Obtained for Several Elements Using LiCl as the Salt Phase. ‘ * Table 1. Temperature Dependence of log KM for Several Elements: * log K, = A+ B/T (°K) (Temperature range: 625 to T50°C) Std. Dev.* Salt Flement A B of log K, 1ic1 Ba®" ~0.6907 2,189 0.02 a3t ~2.6585 9,697 0.1 nas? ~3.3568 10,900 0.08 sm°* 0.7518 1,950 0.05 £’ ~0.158k 2,250 0.05 LiC1-LiF (97.55-2.L45 mole %) P ~1.2356 8,536 0.33 LiBr Ba®" -0.0733 1,333 0.02 Nas* 4.0k 4,297 0.1 uranium or for removal of fission products (notably zirconium) in the protactinium isolation system. The effect of the number of equilibrium stages in the extraction .columns above and below the protactinium decay tank on the partial fuel cycle cost is shown in Fig. 2. In the final selection of the number of stages for these columns, one must consider the expense associated with an increased number of stages. The decision to use two stages below and five stages above the protactinium decay tank was made because a larger number of stages results in only a small decrease in cost. For a reductant feed rate of 429 equiv/day and a thorium concentration in the bismuth entering the column equal to 90% of the thorium solubility at 640°C, the bismuth-to-salt volumetric flow rate ratio in the columns is 0.1L4. The required column diameter is 3 in. if the column is packed with 3/8-in. molybdenum Raschig rings. The effects of changes in the reductant addition rate and in the volume of the protactinium decay tank on the partial fuel cycle cost are shown in Fig. 3. The capital cost of the decay tank, a relatively expensive equipment item, will also influence the final choices for the tank volume and the reductant feed rate; however, this cost has not yvet been taken into consideration. Values of 161 ft3 for the decay tank volume and 429 equiv of reductant per day were selected as optimum. De- creasing the reductant feed rate from 429 equiv/day to L0O equiv/day re- duces the partial fuel cycle cost by 2% and increases the inventory charge on bismuth in the decay tank by about 5%. The effect of the operating temperature on the performance of the protactinium isolation system, as shown by changes in the partial fuel cycle cost, is given in Fig. L. A minimum partial fuel cycle cost of 0.0453 mill/kWhr is observed for the following conditions: a temperature of 640°C, a column having two stages below and five stages above the protactinium decay tank, a decay tank volume of 161 ft3, and a reductant addition rate of 429 equiv/day. These conditions, which have been chosen as the reference processing conditions, result in a protactinium removal time of 10.7 days and a uranium inventory of 12.7 kg (about 0.67% of the reactor inventory) in the protactinium ORNL DWG 70-10,990 0.052 — T T T T T T REDUCTANT ADDITION RATE =429 EQUIVALENTS/DAY 90% OF Th SOLUBILITY IN Bi TEMPERATURE =640°C 7 Pa DECAY TANK VOLUME=161 FT3 Pa PROCESSING CYCLE =10 DAYS — 0.050 0.048 STAGES IN 0.046 LOWER COLUMN | 0.044 | PARTIAL FUEL CYCLE COST (mill/kw hr) 0.042 — — 1 I l I 1 l 1 l | I 3 4 5 6 7 8 STAGES IN UPPER COLUMN Fig. 2. Partial Fuel Cycle Cost for the Protactinium Isolation Sys- tem as a Function of the Number of Stages in the Extractors Above and Below the Protactinium Decay Tank. ORNL DWG 70-10,991 0.052 l N] — 1 T 1 T T STAGES INLOWER COLUMN - 1 - 24 Li FEED RATE \ (MOLES/DAY) _ 0.050 — 457 — < 3 — — £ = 429 £ 0.048 |- — — 3 - 400 ] o 4 0.046 |- — O > o | — - w L 0.044 — = 0.0 STAGES IN UPPER COLUMN =5 < TEMPERATURE =640°C = . Th CONC IN Bi =90% OF Th - % SOLUBILITY AT 640°C a Pa PROCESSING CYCLE=10 DAYS 0.042 |— — 140 160 180 200 Pa DECAY TANK VOLUME (ft3) Fig. 3. Partial Fuel Cycle Cost for the Protactinium Isolation Sys- tem as a Function of the Protactinium Decay Tank Volume and the Reductant Addition Rate. ORNL DWG 70-10,992 0.050 T T [ T [ T r 1 l — STAGES IN UPPER COLUMN=5 - STAGES IN LOWER COLUMN=2 0.049 |— REDUCTANT ADDITION RATE =429 MOLES/DAY 13 DECAY TANK VOLUME=161 ft3 Pa PROCESSING CYCLE =10 DAYS £ 2 0.048 |— — > T | | = - o 8 0.047 — — 12 w O = w - 4 [ 7 - > g o o _, 0.046 |— — s o w S (1 T [ — o o o g E 0.045 |— — {1 < a 0.044 |— — ] l 1 l I I | I | l 10 620 630 640 650 660 TEMPERATURE (°C) Fig. 4. Protactinium Removal Time and Partial Fuel Cycle Cost for Protactinium Isolation System as a Function of Temperature. 10 decay tank. The components of the partial fuel cycle cost are as follows: bismuth inventory charge, 0.0097 mill/kWhr; uranium inventory charge, 0.003 mill/kWhr; loss in 233 0.0013 mill/kWhr; 7Li metal consumption, 0.0151 mill/kWhr; and BeF2 and ThF) addition, 0.0163 mill/kWhr. U due to inefficient protactinium isolation, 2.3 Removal of Noble Metals with the Fluorination-- Reductive Extraction Flowsheet Previous calculations of fission product inventories and poisoning in an MSBR have assumed that most of the noble metals (Se, Nb, Mo, Tc, Ru, Rh, Pd, Ag, Sb, and Te) have been removed from the fuel salt on a short (50-sec) cycle by being plated out on metal surfaces or by being transported to the off-gas system as a "smoke." As a result of these as- sumptions, the neutron poisoning by these materials was negligible; how- ever, the heat load on the off-gas system was increased by about 10 MW. A more conservative assumptibn with regard to both neutron poisoning and heat generation in the processing plant would be that a significant fraction of these materials will remain in the fuel salt and will be re- moved in the processing plant. Accordingly, we have made calculations to estimate the neutron poisoning caused by these materials in the event that they remain in the fuel salt. These calculations assumed a chemical processing system in which protactinium was removed by fluorination-- reductive extraction on a 10-day cycle and the rare earths were removed on a 25-day cycle by the metal transfer process. Many of the noble metals (i.e., Se, Nb, Mo, Te, Ru, Sb, and Te) form volatile fluorides during fluorination and can be separated with varying degrees of dif- ficulty from UF6 by sorption on materials such as NaF. The remaining materials (Ga, Ge, Rh, Pd, Ag, Cd, In, and Sn) are relatively soluble in bismuth and will be extracted into the bismuth with the same removal time as Pa. In these calculations, the removal time of the individual noble- metal elements was varied from 2.5 to 640 days, and the effect of this variation on the reactor performance was determined. Figure 5 shows the (% per annum) CHANGE IN FUEL YIELD ORNL-DWG-71-9492 1 1 I ! { 1 l of e i -0.025} A -0.050 - -0.075} - -0.100} _ 1 ] 1 ] ] | 1 Tc 2 4 10 20 40 102 200 400 REMOVAL TIME (days) Fig. 5. Effect of Removal Time on Fuel Yield for Noble Metals That Form Volatile Fluorides. 103 1T 12 effect of the removal time for important noble metal elements which form volatile fluorides on the fuel yield of an MSBR. The elements that have the most effect on neutron poisoning are Tc, Ru, and Mo; Se, Sb, Nb, and Te have virtually no effect. The curves in Fig. 5 are relatively flat for removal times shorter than about 100 days, indicating that the re- moval efficiency for an individual element may be as low as 10% without seriously impairing reactor performance. TFigure 6 shows the effect of removal time on the fuel yield for elements with nonvolatile fluorides. Here, a relatively large decrease in fuel yield is associated with long removal times for Rh, while the elements Cd, Pd, and Ag have only a small effect. Negligible effects were observed for Ga, Ge, In, and Sn. In these calculations, a 10-day removal time for all noble metals was taken as the reference condition. For this condition, the total neutron poi- soning (¥) for the noble metals is 0.0010 absorption per fissile absorp- tion. The principal isotopes contributing to this poisoning are: lOSRh Y = 0.00085; 113 Cd, ¥ = 0.00005; 99Tc, ¥ = 0.00003; and lOBRh, ¥ = 0.00002. % If the noble metals are removed on a 10-day cycle, their combined thermal power will be 0.98 MW. The heat load on the fluorinator off-gas system will depend on the fractions of noble-metal fluorides that are collected in this system and their residence time. The important heat sources among noble metals having volatile fluorides are given in Table 2. The maximum amount of heat that could be produced in the UF6 separa- tion system, assuming that these isotopes were collected with 100% ef- ficiency and retained indefinitely would be 0.95 MW. Similar data are shown in Table 3 for the noble-metal fission products whose fluorides are not volatile. These isotopes will add a maximum of 30 kW of heat to the bismuth stream in the protactinium isolation system. 2.4 Halogen Removal in the Uranium Removal System In the fluorination--reductive extraction flowsheet for removal of uranium and isolation of protactinium from MSBR fuel salt, the halogen fission products will be removed as volatile fluorides and will enter the UF6 separation system with a 10-day removal time. These materials CHANGE IN FUEL YIELD (% per annum) ORNL-DWG-71-9493 -0.025 I -0.050 I —-0.075 T — 0.100 - = e - e REMOVAL TIME (days) Fig. 6. Effect of Removal Time on Fuel Yield for Do Not Form Volatile Fluorides. 200 400 Noble Metals That €T 1k Table 2. Thermal Power of Isotopes of Noble-Metal Elements That Have Volatile Fluorides Removal time, 10 days Thermal Power Isotope (kW) Half-Life 1320e L56 78 h o 139 67 n 103g4 106 Lo 4 lO6Ru 45,1 1.0 y 129y 4k .6 3L 4 2211 Lh.o 35 d 1eTgy 27.2 3.94d 13Llmy 18.7 1.24 l3hTe 14.3 Lo m 1265, 12.3 12.5 4 129y, 10.7 L.5 n 133ty 9.4 50 m M 4.5 72 m 125y, 3.5 2.7y 13lgy 3.0 25 m 12T, 3.0 105 4 1285 2.8 9.6 h 105k, 2.2 4.43 h I 1.7 6.0 h 130 Sb 1.5 39 m 15 Table 3. Thermal Power of Isotopes of Noble-Metal Elements That Form Nonvolatile Fluorides Removal cycle, 10 days Thermal Power Isotope (kW) Half-Life 126Sn 20.2 '\JlO5 y 125an 5.2 9.62 4 129ms 4.9 1h 128Sn 2.4 62 m 127Sn 2.k 2.1 h 1058n 1.8 36 h g 0.56 7.5 a H2pg 0.48 21.0 h 504 0.47 2.3 a 109p4 0.18 13.5 h 106m,, 0.1k 2.2 h Homeg 0.09 43 4 H3)g 0.08 5.3 h T04 0.07 2.5 h H2pg 0.07 3.2 h 107 Rh 0.03 21.7 m 16 will contain a number of isotopes which could contribute as much as 0.3 MW to the heat load on the fluorinator——UF6 collection system. The principal heat sources in this group are given in Table 4. With the ex- 131I, all of the important materials are isotopes of iodine ception of with half-lives of less than 1 day. The maximum rate at which iodine and bromine would be collected in the uranium removal system is 1 g/day. 3. DEVELOPMENT OF A FROZEN-WALL FLUORINATOR: DESIGN CALCULATIONS FOR INDUCTION HEATING OF A FROZEN-WALL FLUORINATOR J. R. Hightower, Jr. C. P. Tung We are continuing to study rf induction heating of molten salt as a method for providing a corrosion-free heat source for an experimental continuous fluorinator in which a film of salt is frozen on the walls to protect against corrosion. A previous study3 showed that induction heating may be suitable for batch fluorinators, and we have made calcu- tions and experimentsh which indicate that induction heating can also be used with a continuous fluorinator. This section summarizes results of calculations that show the ef- fects of coil current, frequency, wall temperature, and fluorinator diameter on the thickness of the frozen salt film in a continuous fluori- nator employing high-frequency induction heating. Methods for control- ling the thickness of the frozen film are also discussed. Since the ef- ficiency of heating the salt cannot be reliably calculated, an experiment that uses an aqueous electrolyte as a substitute for molten salt will be carried out in order to measure heating efficiency in equipment similar to the fluorinator. 3.1 Effects of Wall Temperature, Current, Frequency, and Fluorinator Diameter on the Thickness of the Frozen Film In the proposed fluorinator configuration (designated previouslyh as configuration I), the induction coils are embedded in the frozen salt film 17 Table 4. Thermal Power of Isotopes of Halogens Removed on a 10-day Cycle Thermal Power Isotope | (kW) Half-Life 131; 117 8.05 4 133; 93.2 21 n 1357 66.3 6.7 h 132; 12.8 2.3 h 134, 11.2 53 m 8k 1.8 32 m 83Br 0.58 2k h 8Ty 0.37 5.5 s 88Br 0.1k 16 s 1361 0.12 83 s 85y 0.11 3.0 m 86Br 0.10 54 s 823r 0.06 35.7T h 1371 0.06 2k s 138, 0.06 6.3 s 89Br 0.0k L.5 s 1301 0.02 12.5 h 18 near the fluorinator vessel wall. In treating the system mathematically, it was assumed that the molten zone would behave as a solid cylindrical charge in an induction coil and that the effect of bubbles in the molten salt would be negligible. The heat generation rate in an infinitely long cylindrical charge placed inside an infinitely long coil is given by: , (1) 2_2 ber a ber! ( )+-be1 (;- bei! _2mI (g_ g P ber —; + | bei where -1 = heat generated in molten salt, W'm ~, conductivity of molten salt, Q-l'nfl, = coil spacing, turns/m, H B3 &8/ O i = rms coil current, A, radius of molten salt zone, m, Y Il 1l/2 p = (2nfgu) / , m, f = frequency, Hz, U = permeability of salt, assumed to be 4 x lO;YWb.A—l.m'l ber, ber', bei, and bei' are bessel functions. At steady state, the heat generated in the molten zone will be trans- ferred by conduction through the frozen film that surrounds the molten core of the fluorinator. The equation relating the rate of heat flow to the sys- tem dimensions and properties is: 2flk(Ti - TC) where heat transferred through the salt film, W'm_l, thermal conductivity of the frozen salt, W-m_l-OC_l, ~ O I ] 19 t = thickness of the frozen salt film between the molten zone and the inside of the induction coil, m, ry = inside radius of the induction coil, m, i = liquidus temperature of the salt, °C, o temperature at the induction coil; here we assume that this temperature can be set when, actually, the fluorinator wall temperature is the quantity set. Combining Egs. (1) and (2) results in an expression that defines the steady-state frozen film thickness in terms of system properties and the operating variables. Equation (1) contains bessel functions, and it is somewhat cumbersome to use; a more useful approximate equation, valid for a/p < 1.4, is: 2.2 . ol _2m’I !0.06077 C%) 3.988: . (3) g P i T i i The salt phase, which is assumed to have the composition 68-20-12 mole % LiF-BeF,-ThF) , has the following properties: 1.5k Q_l'cm—l'at L80°C (ref. 5), 0.0159 Weem Leoct at 4L55°C (ref. 6). g k It was assumed that the induction heating generator operated at 400 kHz. Combining Egs. (2) and (3) and using the above values yields the following expression for the frozen film thickness: D 1 3.988 _ > _AT In { 5= | (D} - 2t) = 6.376 x 10”7 5= , (k) 1 n I where = the inside diameter of the coil, in. (the equation is wvalid s ! for D, <5 in.), t = the frozen film thickness, in., AT Ti - Tc’ °C, 20 coil spacing, turns/m, s I rms coil current, A. The effect of temperature difference across the frozen film for a range of coil currents was determined from Eq. (4), using an assumed coil spacing of 78.7 turns per meter (2 turns/in.)and a 5-in.-ID coil. The results are shown in Fig. 7. For a constant coil current, the steady- state thickness of the frozen film thickness increases with increasing temperature difference across the film, as would be expected. The sensi- tivity of the film thickness to changes in temperature difference also increases as the temperature difference increases. Finally, the temp- erature difference can become so large that a condition is reached in which the heat generated in the molten zone is not sufficient to balance the rate of heat loss from the system and the fluorinator will freeze completely. The thickness of the film at this critical temperature dif- ference is dependent only on the diameter of the fluorinator vessel and the relationship between the heat generated per unit length of fluorinator and the diameter of the molten zone. To illustrate this, assume that the heat generation rate is given by the relation b P =a(D, - 2t), (5) where P = heat generation rate, W-cm—l, a,b = constants, and D, and t are as defined for Eq. (L4). If P in Eq. (5) is equated to Q in Eq. (2) (rewritten in terms of diameter rather than radius), the following general equation is obtained: o Dl b AT = —— —_— - 2nk in Dl - 2t 6.35 As shown in Fig. 7, the maximum stable film thickness with a coil current of 11 A is obtained with a temperature difference of 62°C. Calculations were made with an assumed temperature difference of 82°C (20°C greater than could be allowed for steady-state operation while maintaining a frozen film). Other conditions assumed in the calculation were: a coil spacing of T78.7 turns per meter, and coil currents of O A and 11 A. The results of the calculations are shown in Fig. 8, which shows that, with no heat generation in the molten zone, the time required to freeze the fluorinator completely will be greater than 1.5 hr. With the heat gen- eration resulting from an 11-A coil current, the time required for com- plete freezing will be about 2.9 hr. Equation (12) shows that the freezing process can be reversed by decreasing the temperature difference. We conclude, on the basis of this analysis, that the frozen film thickness in the fluorinator could be controlled, without severe problems, by adjust- ing the temperature of the wall of the fluorinator. Reliable control of the frozen film thickness could be accomplished most effectively by using a direct measurement of the film thickness in a feedback loop to adjust the wall temperature. We are attempting to devise a convenient method for measuring the thickness of the frozen film. 3.3 Power Requirements for an Experimental Fluorinator One of the purposes for making the previous analysis was to estimate the size of the high-frequency generator that would be required for a 5-ft- long experimental fluorinator. In the fluorinator considered, the power 27 supplied by the generator would be dissipated as heat in the molten salt in the fluorinator, in the induction coil, and in the metal walls of the fluorinator. Estimates of the amount of heat generated in the molten salt core and in the induction coil can be made from relationships found in standard textbooks on induction heating (e.g., ref. 8). However, no relationships exist in the literature for estimating the heat generated in a metal cylinder surrounding a cylindrical coil. For our calculations, we have assumed that heat would be generated in the fluorinator vessel at the same rate as in a vessel which has an outside diameter equal to the inside diameter of the fluorinator and which 1s placed inside a coil having the same spacing and carrying the same current as the coil in the fluorina- tor. We have assumed that the fluorinator vessel and the induction coil will be made from nickel since nickel exhibits excellent resistance to fluorine when molten salt is not present and since its specific electrical resistivity is lower than that of other alloys which might be used (pro- viding the coil temperature can be kept above the Curie transition temp- erature of nickel, which is 358°C). A low specific resistivity for the coil material results in a low heat generation rate in the coil. Sufficient power can be generated in the molten salt to maintain a salt film 0.3 in. thick inside the coil with a temperature difference of 51°C between the liquid-solid interface and the induction coil when the following conditions are used: a 5-in.-ID coil plaéed inside a fluorina- tor vessel made from 6-in. sched 4O pipe, a coil spacing of 2 turns per inch, and a coil current of 11 A at a frequency of 400 kHz. The estimated heat generation rates in the molten salt, the induction coil, and the fluorinator vessel wall are 1661, 133, and 42.7 W/ft, respectively. A total heat generation rate of at least 9180 W would be generated in the case of a 5-ft-long fluorinator. The reactance of a 5-in.-ID by 5-ft-long coil is very large, and a power factor of 0.15 was estimated from standard expressions for this type of coil. With a power factor this low, a generator that has a 28 reactive power of 61,200 V-A is required. In order to obtain a coil cur- rent of 11 A, more than 5500 V would have to be impressed across the coil. The required voltage can be reduced by dividing the coil into a number of shorter coils that have that same diameter and spacing and are connected in parallel, with adjacent coils wound in opposite directions. If the coil is divided into 10 sections having 12 turns per section, the es- timated power factor would be 0.11 and the voltage that must be impressed across each coil would be about 760 V (which is still undesirably high). If the coil were divided into 20 sections having 6 turns per section, the power factor would be 0.093, and the required voltage would be about 450 V. Although voltages of the magnitude mentioned above would be potential- ly hazardous and would complicate the removal of molten salt samples from the system during operation, the sampling hazard is decreased by the fact that the induction coil must be electrically insulated from the fluorinator vessel. Use of a higher frequency may result in lower required voltage values; we will investigate this possibility later. Even if high voltages are required, we do not expect to encounter any serious problems in provid- ing the necessary electrical insulation. In estimating the power requirements for the generator, we have as- sumed that the coil and pipe are infinitely long and that the voids pro- duced by bubbling fluorine through the molten salt will not affect the heat generation rate; also, we have essentially assumed a value for the heat generation rate in the metal wall, although it is probably not accurate. The effect of these assumptions will probably be to decrease the efficiency of heating the molten salt in the fluorinator to a value lower than that which we have estimated. With a 5-ft-long coil, the magnetic field strength is weaker than that predicted by equations for an infinitely long coil; therefore, the heat generation rate in the salt would be lower than the calculated value. The presence of bubbles will probably cause the heat generation rate to be lower than the rate that would be ob- tained with no bubbles present. The bubbles would probably reduce the ef- fective conductivity of the salt, and hence the heat generation rate, since the heat generation rate in the salt is directly proportional to the con- ductivity of the salt. 29 The importance of the effects mentioned above cannot be easily de- termined by calculations; hence, we plan to carry out experiments using an aqueous electrolyte solution (probably HNO3) as a substitute for molten salt. We conclude that induction heating appears to be a reasonable means for providing a corrosion-free heat source in a molten salt fluorinator and that the operational problems which have been examined appear to be tractable. L. DEVELOPMENT OF THE METAL TRANSFER PROCESS E. L. Youngblood L. E. McNeese E. L. Nicholson W. F. Schaffer, Jr. An improved rare-earth removal process has been devised, based on the observation that rare earths distribute selectively into molten lithium chloride from bismuth solutions containing rare earths and thorium. Work that will demonstrate all phases of the improved process, which is known as the metal transfer process, is under way. The first engineering experi- ment (MIE-1) for studying the removal of rare earths from single-fluid MSER fuel salt by this process was completed during the current reporting period. The main objective of this experiment was to demonstrate the selective re- moval of rare earths (La and Nd) from a fluoride salt mixture containing thorium fluoride. L.1 Equipment and Materials Used for Experiment MTE-1 The experiment was performed in a 6-in.-ID carbon-steel vessel (shown 9 in Fig. 9), which has been described previously. The vessel contained two compartments interconnected at the bottom by a pool of molten bismuth that was saturated with thorium. One compartment contained MSBR fuel carrier salt (72-16-12 mole % LiF—BeFZ—ThFh) to which 2 mCi of lh?Nd and a suf- ficient quantity of LaF3 to produce a lanthanum concentration of 0.38 mole % had been added. The other compartment contained lithium chloride, a cup that contained a lithium-bismuth solution, and a pump for circulating the lithium chloride through the cup. The pump was constructed of quartz and 30 ORNL DWG 70-4504 ™ L ] //—QUARTZ CARBON-STEEL— L PUMP PARTITION 6-in. CARBON- »" STEEL PIPE 24 in. \ N LiCl 72-16-12 MOLE Yo~ D FUEL CARRIER SALT Li- Bi Th-Bi 1 Fig. 9. Carbon-Steel Vessel for Use in the Metal Transfer Experiment. 31 9 used sapphire balls as check valves. Lithium chloride was forced in and out of the pump body by varying the argon pressure inside the pump chamber. Electrical probes were used to detect the liquid level in the pump chamber and to actuate solenoid valves in order to change the argon pressure during a given pump cycle. Use of this type of pump permitted ac- curate control of the flow of lithium chloride through the lithiumbismuth container. In order to obtain mixing in the main bismuth pool, about 10% of the metal volume was forced to flow back and forth every 7 min through a 1/2- in. slot below the partition between the fluoride and chloride compart- ments. This flow was effected by reducing the pressure of the argon cover gas in the fluoride compartment relative to that in the chloride compart- ment. Argon sparge tubes were placed in each side of the vessel as well as in the lithium-bismuth cup in order to promote contact between the salt and metal phases. Lumps of thorium metal (a total of 501 g) measuring approximately 0.5 in. on a side were placed in the bottom of the vessel in order to ensure that the bismuth phase was saturated with thorium. The amount of thorium added in this manner was about 20 times the amount that could dissolve in the bismuth. The exterior of the carbon-steel vessel used for the experiment was spray-coated with nickel aluminide to prevent air oxidation of the carbon steel. The quantities of materials used in the experiment are shown in Table 5. Most of the materials were purified to remove oxides and other impuri- ties before they were introduced into the system. The lithium chloride was purified by contact with bismuth saturated with thorium at 650°C. The carbon-steel vessel and the bismuth used in the experiment were first treated separately with hydrogen at 650°C for about 12 hr for oxide re- moval. The bismuth was then added to the carbon~steel vessel and was sub- sequently treated with hydrogen at 650°C for an additional 10 hr. Purified MSBR fuel carrier salt (72-16-12 mole % LiF—BeFE—ThFh) was obtained from the Reactor Chemistry Division. The argon used for cover gas and for op- eration of the lithium chloride pump was purified by passage through a molecular sieve dryer and a bed of uranium turnings operated at 600°C. Table 5. Materials Used in Metal Transfer Experiment MTE-1 Volume (cm3) g-moles Fluoride salt 709 36.6 72.0-15.5-12.1-0.4 mole % LiF-BeFpo-ThF)-LaF3; 2 mCi of 14TNdFs Bismuth saturated with thorium 797 36.7 (0.35 at. % Th, 0.24 at. % Li) LiCl 10Lk2 36.6 Li-Bi (35 at. % Li) 192 11.1 2t 33 The first quartz pump used in the experiment failed to operate properly; when we attempted to operate the pump to begin the experiment, we found that the check valves were stuck. Therefore, the system was cooled to 300°C and the pump was replaced with a second quartz pump of a similar design. The latter pump operated satisfactorily throughout the experiment. L.2 Experimental Procedure The sequence of operations carried out during the experiment can be described as follows. The lithium chloride was pumped through the lithium- bismuth container for 3 hr at the flow rate of about 25 cmB/min. Pumping was then stopped, and the system was allowed to approach equilibrium during a L-hr period. At the end of this period, filtered samples of the salt and bismuth phases were taken. As shown in Fig. 10, this sequence was repeated for 11 cycles. Three additional cycles were carried out in which the pumping period was increased to 6 hr and the equilibration pe- riod was omitted. During the pumping and equilibration periods, a por- tion of the main bismuth pool was forced back and forth between the flu- oride and the chloride compartments at a rate equivalent to 10% of the metal volume every 7 min. The experiment was carried out at 660°C. During the 5-day operating period, the pump was operated 50.5 hr and 81.2 liters of lithium chloride was circulated through the lithium-bis- muth container. The filtered samples taken during the experiment were prepared for analysis by cutting off the filter section with tubing cutters and cleaning the external surfaces with emery cloth. The lw&d activity in each sample was determined by direct counting of the 0.53- MeV gamma rays emitted by the sample. The lanthanum concentration in each sample was determined by neutron activation. ORNL DWG 70-6256 PUMPING TIME PER CYCLE = 3 hr TOTAL PUMPING TIME =50.5 hr EQUILIBRATION TIME PER CYCLE =4 hr TOTAL VOLUME PUMPED = 81.2 LITERS SAMPLE TIME PER CYCLE =~I hr NUMBER OF SAMPLES TAKEN OF EACH PHASE =15 SAMPLING - - - - - - - - - - - - - - - HEAT UP (s oo b e b ey by by by 0 10 20 30 40 50 60 70 80 90 100 {10 120 130 TIME— (hr) Fig. 10. Operating Schedule for Metal Transfer Experiment MTE-1. we 35 4.3 Experimental Results During the experiment the rare earths should have distributed between the salt and metal phases in a manner depending on the concentration of reductant in the metal phase. The relative amount of rare earth in a salt and metal phase at equilibrium can be expressed by the distribution coef- ficient, which is defined as: _ mmole fraction of M in the bismuth phase M mole fraction of M in the salt phase At equilibrium, a portion of the rare earths originally present in the fluoride salt would have been extracted into the bismuth phase, from which it should have distributed into the lithium chloride. The lithium concen- tration in the lithium-bismuth solution (35 at. % lithium) was sufficiently high that, at equilibrium, essentially all of the neodymium and lanthanum would be removed from lithium chloride in contact with this metal phase. Therefore, continued circulation of the lithium chloride through the lithium-bismuth cup should have gradually removed the neodymium and lan- thanum from the fluoride salt and deposited these materials in the lithium- bismuth solution. 228 The concentrations of lanthanum, neodymium, and Ra, a decay product of 232‘I‘h, were determined in the salt and bismuth phases periodically throughout the experiment. These data are summarized in Table 6. Values for the distribution coefficients thus obtained were relatively constant during the run. As chown in Figs. 11 and 12, the average values for the distribution coefficients between the fluoride salt and bismuth for lan- thanum and neodymium were 0.044 and 0.073, respectively. The distribution coefficient values are in good agreement with values obtained in previous studies.lo Distribution coefficient wvalues for lanthanum and neodymium between the bismuth and lithium chloride are shown in Figs. 13 and 1k. These data show considerable scatter during the first two-thirds of the experiment because of the low concentrations of rare earths in the lithium chloride. The neodymium concentration in the lithium chloride was too low to be determined accurately, and the lanthanum concentration was lower than Table 6. and Bismuth Phases During Metal Transfer Experiment MTE-1 Concentration of Rare Earths, Lithium, and Thorium in the Salt 36 Voi;me Fluoride Salt Bismuth-Thorium Lithium Chloride Lithium-Bismuth | Pumping LiCl 1 Tyge 14Tyqa 14Tyqa i Time Pumped La (dis min-lg-1 228g,b La (dis min—lg-1 228p b Li Th La (dis min-lg=1) 228p4b La 228Rqa Li Th (hr) (liters) (mg/g) x 10-6) (counts min-lg=1) (mg/g) x 10-6) (counts min~ (ppm) (wt %) (mg/g) x 10- (counts min—1g=1) (mg/g) (counts min=lg—1) (wt %) (ppm 0 0 8.21 1.56 1429 0.0k9 0.36 10 92 0.hlL 0.013 c 1h1k 0.0003 6.6 1.63 3.4 3 5.4 7.51 1.55 0.099 0.37 93 0.37 0.017 c 0.021 1.56 L.2 6 10.4 1.61 1183 0.37 10 1403 453 9 15.1 T.22 1.k49 0.098 0.41 78 0.38 0.022 c 0.031 1.58 3.0 12 20.0 1.50 LT 0.37 9 2044 477 15 25.0 6.58 1.L46 0.068 0.36 270 0.k4k2 0.013 c 0.029 1.63 22.5 | 18 29.9 1.36 Thu 0.29 T 27L8 Lu2 21 3k4.9 5.87 1.41 0.090 0.33 L20 0.L45 0.021 c 0.027 1.5k 18.4 2l 39.7 1.17 51k 0.25 5 Th 2980 481 | 2T L .0 5.h2 1.30 0.979 0.26 80 0.43 0.162 c 0.028 1.54 10.0 30 48.8 1.16 443 0.060 0.22 3 78 2.85 3030 475 | 33 54,2 L.L4s5 1.11 0.228 0.18 80 0.28 0.323 3.23 0.024 1.60 L.s! 39 64.0 3.89 1.0k 340 0.22 3 1.61 2650 L3k L5 65.5 3.72 0.99 456 0.2L4 2 76 1.23 2900 h§3 50.5 81.2 3.92 0.94 0.058 0.18 120 0.32 0.131 2.46 0.11%L 1.53 7.0 ytions per minute per gram, corrected for decay. ~ minute per gram. ted. 37 ORNL DWG 70-6269RI 0.07 | | | | | | | | | 8 | » | | 0.064 CALCULATED VALUE 0.590 0.i67 |~ 0.06 |— — 0.05 e o —-1 ¢ o o 0.04 +— — ® o —|’ 0.03 |— — W a . 0.02 - —_ % ’ 0.0 |- — 0 | | | | | | | | | | | | | | 0 | 2 3 4 5 6 7 8 9 10 it 12 i3 14 15 SAMPLE NUMBER Fig. 11. Distribution Coefficient for Lanthanum Between Fluoride Salt and Bismuth Saturated with Thorium During Metal Transfer Experiment MTE-1. 38 ORNL DWG 70-6257 RI 0.10 0.09 ¢ 0.08 | e ¢ ° 0.07 — , * ° ° o 0.064 CALCULATED VALUE ® 0.06 — 0.05 — © f 0.04 - 003 I~ 0.02 — 00! |~ oooL—L L 1 1 1 1 0 141 I 2 3 4 5 6 7 8 9 0 1 2 13 14 |5 SAMPLE NUMBER Fig. 12. Distribution Coefficient for Neodymium Between Fluoride Salt and Bismuth Saturated with Thorium During Metal Transfer Experiment MTE-1. ORNL DWG 70-6275RI S0 l l T & 7 T I T 20 - o - 1 |5 p— (8] o 10| 5k o [0.9 CALCULATED VALUE o 1 1 1 i 1 1 | | 0 48 o 0 12 24 36 50.5 PUMPING TIME ( hours) Fig. 13. Distribution Coefficient for Lanthanum Between LiCl and Bis- muth Saturated with Thorium During Metal Transfer Experiment MIE-1. 6¢ ko ORNL DWG 70-6258RI ‘(/YI)CALCULATED VALUE Dc-nd ol 1Lt 10 4 5 6 7 8 9 10 1 2 13 14 I5 SAMPLE NUMBER Fig. 14. Distribution Coefficient for Neodymium Between LiCl and Bis- muth Saturated with Thorium During Metal Transfer Experiment MTE-1. b1 expected. These results probably indicate that the contact between the bismuth and lithium chloride phases was not sufficient to maintain equi- librium during this period. An argon sparge tube was inserted into the lithium chloride about midway through the run, and during the last tuird of the experiment the average distribution coefficients for lanthanum (2.8) and neodymium (5) between the bismuth and lithium chloride phases were reasonably close to expected Valuesll (0.9 for La and 7.0 for N4d). The distribution of radium was also followed during the experiment. Most of the radium was introduced with the fluoride salt; a small amount was introduced with the thorium metal that dissolved in the bismuth phase. During the experiment, the radium slowly transferred from the fluoride salt to the lithium chloride and lithium-bismuth phases, as shown in Fig. 15. At the end of the experiment, 72% of the radium was in the lithium chloride, 15% was in the fluoride salt, 12% was in the lithium-bismuth solution, and less than 1% was in the bismuth-thorium solution. Approximately 50% of the lanthanum and 25% of the neodymium (after 1hT correcting for Nd decay) originally present in the fluoride salt were removed during the experiment. Figure 16 shows the decrease in the lan- thanum and neodymium concentrations in the fluoride salt as a function of pumping time and run time. The rates at which the rare earths were removed are in agreement with the expected removal rates. However, the lanthanum and neodymium removed from the fluoride salt did not collect in the lithium-bismuth as expected. No neodymium was detected in the lithium-bismuth solution during the run, and only a small fraction of the lanthanum (<1%) was collected in the lithium-bismuth. The concentra- tion of lanthanum in the lithium-bismuth solution during the experiment is shown in Fig. 1T7. At the conclusion of the experiment most of the rare earths that had been removed from the fluoride salt were found in a layer of material located at the interface between the LiCl and the thorium-saturated bismuth. Reaction of oxide impurities in the system with the rare earths is thought to have caused the rare earths to deposit at the lithium chloride--bismuth-thorium interface rather than in the lithium-bismuth solution. Details are given in the following section. PERCENT OF TOTAL RADIUM 100 @ o N o ¥ O 20 ORNL DWG 71-7856 x__LiCl . X/ l urated with Thorium, and LiCl During Metal Transfer Experiment MTE-1. X/ \ .\ .\ e FLUORIDE SALT o Li-Bi T ° - o 0 o o 12 24 36 48 50.5 PUMPING TIME ( hours) Fig. 15. Distribution of Radium Between Fluoride Salt, Bismuth Sat- cft 43 ORNL DWG 70-11,002 R1 I T ! F o = o ? ] 0.9 08 0.7 0.6 0.5 @ NEODYMIUM (AFTER CORRECTING FOR DECAY) FRACTION OF RARE EARTHS REMAINING IN FLUORIDE SALT 0.4} X LANTHANUM - 03 1 1 L 1 e 1 1 1 1 1 I 1 1 L 1 1 0o 12 24 36 48 50.5 PUMPING TIME (hr) L 1 1 1 I 0 327 65.9 95.0 108.0 RUN TIME (hr) Fig. 16. Rate of Removal of Neodymium and Lanthanum from the Fluoride Salt in Metal Transfer Experiment MTE-1. LANTHANUM CONCENTRATION {mg /gram) ORNL DWG 70-627|R!I 0.12 , 0.114 0.0 |- 0.08 0.06 - 0.04 |- 0.021 e i 0.0 . 1 1 ] | 1 1 ] ] ' o 12 24 36 48 50.5 PUMPING TIME({ hours) Fig. 17. Concentration of Lanthanum in the Lithium-Bismuth Solution During Metal Transfer Experiment MTE-1. L5 4.4 Postoperational Equipment Examination At the end of the experiment, the carbon-steel vessel was cut apart for inspection. Most of the La and Nd that had been removed from the fluoride salt during the experiment was found in a 1/8-in.-thick layer of black material located at the interface between the lithium chloride and the bismuth that was saturated with thorium. Chemical analysis showed the black layer to contain 3.6 wt % La, 7.7 wt % Th, 12 wt % Li, 0.95 wt % Si, 0.54 wt % Fe, 59 wt % Cl, 4.4 wt % 0, with the remaining material being bismuth. Thorium oxide and lithium chloride were the only compounds that could be positively identified by X-ray diffraction. Thorium oxide was expected to be present at the point where the black material was found, as the result of reaction of thorium with oxide impurities in the system. The chemical form of the rare earths in the black material was not deter- minded; however, it is believed that the presence of oxide in the system is responsible for the accumulation of rare earths at the interface between the lithium chloride and the thorium-saturated bismuth rather than in the lithium-bismuth solution. Oxide impurities were probably introduced into the system in two ways: (1) by deterioration of the quartz pumps, and (2) by exposure of the salt to air during installation of the second quartz pump. Other areas in the system were also checked for accumulation of rare earths. A layer of material approximately 1/8 in. thick was noted at the bottom of the bismuth that was saturated with thorium. The appearance of this layer was different from that of the remainder of the metal phase. The material was pyrophoric, and sparks and yellow smoke could be produced by scratching its surface. It contained 20 wt % thorium and is assumed to be composed of a mixture of thorium bismuthide particles and bismuth. Al- though the Nd and La concentrations in this layer were about ten times the concentrations of the rare earths in filtered samples of the bismuth-thorium solution taken during the run, the actual rare-earth inventory amounted to only about 3% of the total rare earths in the system. Lumps of undissolved thorium metal could also be seen in the thorium-bismuth phase, which in- dicates that sufficient thorium was present to maintain the thorium con- L6 centration in the bismuth at its solubility during the run. The carbon- : 1k steel dip tubes and portions of the quartz pump were examined for 7Nd 147 activity. No significant amounts of Nd were found on these components or in the lithium chloride that had condensed in the upper portion of the vessel. The carbon-steel vessel and other steel components showed no evidence of corrosion. The quartz pump was also in reasonably good condition. The surfaces of the guartz that had been in contact with the lithium chloride were white in appearance, and the quartz was somewhat weaker than initially as the result of devitrification. 4.5 Design and Testing of a Carbon-Steel Pump Having Molten-Bismuth Check Valves Because of the difficulties encountered with the operation of quartz pumps for circulating the lithium chloride in metal transfer experiments, a carbon-steel pump having molten-bismuth check valves was designed and tested. The pump, shown schematically in Fig. 18, was operated with lithium chloride at a flow rate of about 25 cm3/min at 650°C for about 2 weeks. The pump operated satisfactorily and appears to be suitable for use in future metal transfer process experiments. 5. STUDY OF THE PURIFICATION OF SALT BY CONTINUOUS METHODS R. B. Lindauer L. E. McNeese To date, the molten salt required for development work, as well as for the MSRE, has been purified from harmful contaminants (mainly sulfur, oxygen, and iron fluoride) by batch processes. It is believed that the labor costs associated with salt purification can be reduced considerably by the use of a continuous process for the most time-consuming operation, which is the hydrogen reduction of iron fluoride. Equipment12 has been installed in which molten salt and gas can be countercurrently contacted in a packed column to obtain data for design of a full-scale continuous salt purification facility. ORNL DWG 70 -898I TO VENT AND Hmm‘r”ARGON SUPPLY TEFLON PLUG - ELECTRODES 172" CARBON STEEL FOR LIQUID TUBING 28" LONG LEVEL MESUREMENT I/4"CARBON STEEL s TUBING DISCHARGE Ty = MOLTEN BISMUTH > | I . [ x4 INTAKE Fig. 18. Carbon-Steel Pump with Melten-Bismuth Check Valves for Metal Transfer Experiments. L8 During this reporting period, the system was charged with 28 kg of LiF-BeF,, (66-34 mole %), and ten flooding runs were carried out using hydrogen and argon. After the first flooding run with hydrogen had been completed, the feed line to the column became plugged. Since oxide in the salt was believed to be causing the restriction, the salt was treated with a H2—HF mixture. Oxide exceeding the amount that is soluble in the salt was removed, as indicated by analysis of the off-gas stream for water; however, the plugging recurred after salt was again fed through the column. The difficulty was apparently caused by plugging of the 1/16-~ in.-diam holes at the bottom of the inlet salt distribution ring. Since no restriction was found in the 1/4-in.-0D feed line, a 3/16-in.-diam hole was drilled through the inner wall of the distributor ring to provide an alternate salt inlet and the feed line was replaced with a 3/8-in.-diam line. After this modification, satisfactory salt flow to the column was obtained. Details of the operations carried out for the removal of oxide from the salt and of the flooding runs are given in the remainder of this section. 5.1 Batch Treatment of Salt for Oxide Removal After the first run, most of the salt in the system (13.3 liters) was treated with a L% HF—H2 mixture in the feed tank for 9 hr. At the beginning of the treatment period, the salt was heated to 650°C to dissolve as much of the oxide as possible. As oxide was removed from the salt, the temperature of the salt was slowly decreased to a final value of 535°C in order to increase the solubility of HF in the salt and thereby increase the rate of reaction between HF and the oxide. The off-gas stream from the oxide removal operation was continuously analyzed for water by use of a electrolytic hygrometer, which was preceded 13 A total of 580 ppm of oxide was removed; by a NaF bed for removing HF. this is slightly more than twice the solubility of BeO in the salt at 650°C. The average HF utilization was 30%. L9 5.2 Measured Flooding Rates During Countercurrent Flow of Molten Salt and Hydrogen or Argon During the flooding runs, salt flow rates of approximately 50 to Loo cmB/min were used with argon and hydrogen flow rates of up to 7.5 and about 30 liters/min, respectively, as shown in Table 7. The temp- erature of the column was T00°C in each case. The pressure drop across the column increased linearly with increases in gas flow rate as shown in Fig. 19, which summarizes the best data from the ten runs. At the flow rates studied thus far, the salt flow rate had only a minor effect on pressure drop. In the figure, points obtained with the higher salt flow rates usually lie near the upper side of the band, although some data scatter was observed. Five of the ten runs were terminated when the salt-gas interface was depressed below the seal loop in the exit line in the column by the combined pressure drop in the off-gas line and in the column. The maximum flow rate possible with the present system is about 19% of the calculated flooding rate, although smooth operation is observed only up to about 12% of the calculated flooding rate. At the higher flow rates, operation was erratic and the pressure drop across the column fluctuated over a range of 2 to 3 in. of salt. The variation in pressure drop was possibly the result of poor salt distribution at the top of the column. One run was carried out in which the system was op- erated with a high salt-gas flow ratio (using an argon flow rate of about 1 liter/min). The salt flow rate was increased stepwise to 280 cc/min; then, after 15 min at this condition, the salt flow rate was increased to 400 cm3/min for 5 min. No indication of flooding was observed. The pres- sure drop across the column was about 30 in. H20. These observations indicate that a column of this type should have a much higher capacity for an operation, such as removal of oxide from salt by treatment with H -HF, in which the HF utilization is high and low gas-to-salt flow ratios 2 can be used. In the final flooding run, stable operation was obtained with a salt flow rate of lOOcmB/minand hydrogen flow rates up to 20 liters/min. This . 2+ nydrogen flow rate is sufficient to reduce 400 ppm of iron (as Fe” ) from Table 7. Summary of Countercurrent Runs Made with LiF—BeF2 Maximum Percent of Average Salt Gas Flow Duration of Calculated Maximum Run Flow Rate Rate Gas Contacting Flooding Salt Holdup No. (cm3/min) (1iters/min) Used (min) Rate (%) 1 53 30.5 H, 115 12 13 2 173 T.5 Ar L5 23 >15 3 140 6.3 Ar 130 16 15 L 120 23.5 H, 45 15 2k 5 90 25.5 H, 35 1L 15 6 282% ol Ar Lo 6 18 T 175 2k.0 H, 58 19 21 8 183 6.2 Ar 60 19 24 9 122 6.8 Ar 80 17 11 10 104 2k.3 H, 95 15 18 04 ®Rate was increased to 400 cm3/min toward the end of the run without flooding the column. 51 ORNL DWG 70-11,003 1och T T T T 1 ¥ T T I 1 1 ¥ T ¥ ¥ ¥ T - — r— — - - ° N X £ a 1o - a o ] > HYDROGEN ) = o 3 _ -l o o . O i COLUMN DATA ) HEIGHT = 81-in. . DIAMETER =1.38-in. . PACKING =0.25-in. RASCHIG RINGS 1 1 1 1 1 1 1 1 1 l 1 1 4 1 1 1 1 1 1 10 100 GAS FLOW RATE (liters/min) Fig. 19. Column Pressure Drop During Countercurrent Flow of Molten Salt (66-34 mole % LiF-BeFE) and Either Argon or Hydrogen. o2 a 100 cm3/min salt stream, assuming that the gas and salt reach equilibrium at T700°C; however, the mass transfer coefficient is probably not sufficiently high to allow equilibrium to be reached with the present column height (81 in. ). 6. SEMICONTINUOUS REDUCTIVE EXTRACTION EXPERIMENTS IN A MILD-STEEL FACILITY B. A. Hannaford C. W. Kee L. E. MclNeese Following routine HE—HF treatment of the bismuth and salt in the sys- tem, the phases were transferred to their respective feed tanks. Then 90 g of purified LiF—UFh eutectic salt was added to the salt phase to produce a UFh concentration of about 0.0003 mole fraction for the first mass transfer run (UTR-1). Hydrodynamic performance during the 140-min run was excellent, and ten pairs of bismuth and salt samples were taken. The column was op- erated at 63% and 82% of flooding (at a bismuth-to-salt volumetric flow rate ratio of unity); nevertheless, virtually none of the uranium was ex- tracted from the salt due to an operational difficulty that prevented re- ductant from being added to the bismuth. Dissolution of thorium in the bismuth feed tank in preparation for the second mass transfer experiment proceeded slowly as the result of poor mix- ing in the tank. In run UTR-2, 95% of the uranium was extracted from the salt. The run was made with a 200% excess of reductant over the stoichio- metric requirement and with bismuth and salt flow rates of 247 ml/min and 52 ml/min, respectively. These flow rates are equivalent to about 83% of flooding. 6.1 Preparation for Mass Transfer Experiments In preparation for the first mass transfer experiment with uranium, the salt and bismuth were treated with H2-HF to remove traces of oxides. The procedure was essentially the same as described ear]_ier,:LLL except that the gas flow rate (30% HF-H.) was only 15 scfh. The utilization of 2 53 HF decreased from an initial value of about 30% to 5 to 10% during most of the 15-hr treatment period. At the end of the period, it decreased rapidly to about 1%. Following periods of hydrogen sparging to reduce iron fluoride, and argon sparging to remove HF, a 400-g charge of thorium metal was added to the treatment vessel in order to transfer zirconium from the salt phase to the bismuth phase and to supply the necessary re- ductant in the bismuth for the first mass transfer experiment. About 24 hr after the thorium metal had been added to the treatment vessel, the salt and bismuth phases were sampled and transferred to the respec- tive feed tanks. A 270-g batch of LiF—UFh eutectic salt was prepared by placing 51.6 g of LiF and 231 g of UFh in a small graphite crucible in which the materials were melted and sparged with a HF-H, mixture for about 6 hr at 650°C. The 2 H,.-HF gas flow rate was too low during this period to obtain meaningful gis samples without upsetting the off-gas pressure, and in turn, the gas flow rate and the composition of the gas. Following the treatment period, the eutectic mixture was solidified and divided into two parts. A 90-g portion was added to the salt feed tank in preparation for the first mass transfer experiment (UTR-1). 6.2 Mass Transfer Experiment UTR-1 The first uranium mass transfer run (UTR-1) was very successful from the standpoint of hydrodynamics and demonstrated that simultaneous samples of the bismuth and salt leaving the extraction column can be taken easily. However, the run failed to provide mass transfer data for reasons described below. During the run, a uniform flow rate of each phase was maintained at 100 ml/min (about 63% of flooding) for 90 min, during which seven sets of simul- taneous samples of the salt and bismuth streams leaving the column were taken. Three additional sets of samples were taken during the 25-min period in which the salt and bismuth flow rates were increased to 130 ml/min (about 82% of flooding). The flow rates were determined from the changes in the 5k bismuth and salt volumes in the feed tanks during the run, as shown in Fig. 20. Analyses of the flowing bismuth samples for uranium and thorium showed little, if any, of these materials (<2 ppm of U and <4 ppm of 1h). A review of the procedure followed during the earlier addition of thorium metal to the treatment vessel showed that most of the thorium had failed to enter the crucible containing the bismuth and salt because a cube of thorium had blocked the hole in the graphite crucible cover, causing the thorium cubes that were subsequently added to come to rest on the crucible cover . Uranium analyses of salt from the salt feed tank and from the salt receiver vessel were identical (1200 ppm), which indicates that there was no transfer of uranium from the salt. This confirmed analytical results for the bismuth samples, which also showed that no transfer of uranium to the bismuth phase occurred during the experiment; however, the uranium concentrations reported for the set of nine flowing salt samples averaged 1092 ppm; each value in this set was less than the expected value of 1200 ppm. A possible explanation for the low uranium concentrations in the flow- ing salt samples was considered and discarded. If the salt contained as much as 3 vol % entrained bismuth, the effect would be a decrease in the uranium concentration in the sample to about the average value observed since the bismuth contained no uranium. However, the analysis of random salt samples indicated bismuth concentrations of less than 80 ppm, or less than approximately 0.003 vol %. Similarly, there was no inverse correlation of reported sample weights with indicated uranium concentra- tions. The entrainment detector (located between the salt outlet from the column and the salt sampler) also provided evidence that bismuth entrainment in salt leaving the column was less than (probably much less than) 1.5 vol %, the entrainment rate that would be required in order for the bismuth level in the detector to reach the lower probe. The bis- muth entrainment measurements were made while the column was operating at 82% of flooding; the salt in the drain line below the entrainment detector was frozen during this period. 55 ORNL-DWG-71-949I 20 T T 1 1 T T 1 1 v 1 1 ¥ 0 1 1 + + + + + + + + + + SAMPLE PAIRS ’BL BISMUTH ] —e—— SALT 16f 4 14| = 7 S [+ = 2} - ) = Z O < & o aL o o ] o St 4r— - 2r- - 1 L 1 A 1 1 L b 1 1 1 1 1 1 \l 40 50 60 70 80 90 00 0 120 130 140 150 160 170 180 190 200 TIME (min) Fig. 20. Volumes of Bismuth and Salt Remaining in Feed Tanks vs Run Time for Run UTR-1. The bismuth and salt flow rates were approximately equal and had the values of 100 ml/min for the period 60 to 150 min and 130 ml/min for the period 150 to 175 min. 56 Following experiment UTR-1, the salt and bismuth were returned to their respective feed tanks in preparation for experiment UTR-2. 6.3 Mass Transfer Experiment UTR-2 Because of the difficulty encountered in adding thorium metal to the treatment vessel prior to run UTR-1, the thorium reductant for ex- periment UTR-2 was added directly to the bismuth feed tank. The 192.5-g charge dissolved very slowly in the bismuth, possibly because of inadequate agitation of the metal phase by the argon sparge. Periodic samples of the metal phase (taken with a standard fritted sampler and with a graphite ladle) showed that the thorium concentration was increasing at the rate of only about 2 ppm/hr. After 200 hr at 540°C and 120 hr at 630°C, a filtered sample and a ladle sample indicated thorium concentrations of 750 ppm and 1100 ppm, respectively. These values should be compared with the thorium solubility in bismuth at 540°C (i.e., 880 ppm) and a material balance value of 1165 ppm, assuming complete dissolution of the thorium. The balance of the LiF-UF) eutectic (180 g) was added to the salt feed tank to bring the uranium concentration to about 3000 ppm. Because of uncertainty in the value for the thorium concentration in the bismuth, a bismuth-to-salt flow rate ratio of 5 was used for experiment UTR-2. The first attempt to begin run UTR-2 was halted by the failure of bis- muth to transfer from the bismuth feed tank. The site of the obstruction was apparently the open end of the feed tank dip tube since the line was cleared by back-flowing argon through the line. After bismuth and salt flow to the column had been initiated, the bismuth flow rate was increased to about five times the salt flow rate in order to achieve the desired run conditions. The flow rates of both bismuth and salt were steady over a period of about 4O min, as shown in Fig. 21. During this period, seven pairs of samples were taken of bismuth and salt streams leaving the column; the bismuth and salt flow rates were 247 and 52 ml/min, respectively, cor- responding to operation at about 83% of the column flooding rate. The small variations in the flow rates during this period are indicated by the close approximation to straight lines shown by the liquid-level curves in Fig. 21. o7 ORNL -DWG-71-9490 20 T 1 T T T T 1 1 1 1 1 1 T + + + + + + + SAMPLE PAIRS TAKEN 181 ———— BISMUTH —-— SALT FEED REMAINING (liters) 1 1 1 i 1 1 ) 1 1 1 1 1 1 40 45 50 55 60 65 70 75 80 85 90 95 100 105 o TIME (min) Fig. 21. Volumes of Bismuth and Salt Remaining in Feed Tanks vs Run Time for Run UTR-2. Flow rates inferred from slopes; bismuth, 247 ml/min; salt, 52 ml/min. Fraction of flooding, 83%. 58 In addition to samples of the salt and bismuth streams leaving the column, samples were also taken from bismuth and salt feed and receiver tanks. The resulting data are tabulated in Table 8. The fraction of uranium extracted, based on the data from the salt samples, was 95%. The scatter in the uranium analyses of the salt and bismuth streams leaving the column was larger than expected. A uranium material balance based on the average uranium concentration in salt leaving the column would require an average uranium concentration of about 230 ppm (as com- pared with the observed value of 174 ppm) in bismuth leaving the column. 7irconium concentrations in the bismuth samples were within about 25% of the expected values. ZExcellent agreement was observed between thorium and lithium concentrations in the bismuth stream leaving the column and in the bismuth receiver vessel. The thorium concentration in the bismuth fed to the column is believed to have been 659 ppm, as indicated by the amount of dissolved metals in the bismuth after the experiment. The quantity of reductant present in the bismuth during the experiment (thorium and lithium) amounted to a stoichiometric excess of about 200% over that required to reduce the uranium and zirconium in the feed salt. This experiment is important because it represents the first known demonstration of the continuous extraction of uranium from molten salt into bismuth containing reductant. The results indicate that high uranium removal efficiencies can be obtained in a packed column having a reasonable length. Additional experiments will be carried out in order to determine more exactly the mass transfer performance of packed column extractors over a range of operating conditions. 7. MEASUREMENT OF AXTAL DISPERSION COEFFICIENTS IN PACKED COLUMNS J. S. Watson L. E. McNeese Axial dispersion in the continuous (salt) phase can reduce the per- formance of packed column contactors that are proposed for the MSBR fuel processing system. Effects of axial dispersion will be most severe where Table 8. Summary of Mass Transfer Data Obtained in Run UTR-2 Bismuth flow rate, 247 ml/min; salt flow rate, 52 ml/min Fraction of uranium transferred = 399%666;29 = 0.05. U balance, flowing stream samples = U entering Bi _ 1.53 millimoles/min _ U leaving salt 2.10 millimoles/min 0.73. Bismuth Phase U Cone. in ’“A ’ TOTal Origin Salt Dissolved of Phase U Zr Th Li Metals Sample (ppm) (ppm) (meq/liter) (ppm) (meg/liter) (ppm) (meq/liter) (ppm) (meq/liter) (meq/liter) Feed tanks 3000 21 3.k "0 "0 11108 109 20 27.6 1k40° 6170 Flowing stream 1 4o6d 2L 36.3 22 9.3 330 55 L7 65 166 Flowing stream 2 1kl 223 36.3 2k 10 390 6L.9 39 54 165 Flowing stream 3 11k 138 22.4 22 9.3 320 53.3 45 62 147 Flowing stream U 137 140 22.7 21 8.9 300 50 33 46 128 Flowing stream 5 163 197 31.9 18 T.6 270 45 33 46 131 Flowing stream 6 161 1ko 22.7 25 11 420 70 34 L7 151 Flowing stream T 181 155 25.1 16 6.8 1ko 23.3 28 39 9k Flowing stream ave. 150 17k 28.2 21 8.9 310 51.6 37 51 1ko Receiver tanks 146 73712 - -~ — 308 51.3 3k W7 aQ,uestionable value. bCalculated as the difference: 144 meq/liter minus the sum of the concentrations of U, Zr, and Li. “Most probable value, based on average for flowing bismuth samples. dQuestionable value; not included in calculated average value. 65 60 high flow rate ratios are required. An experimental program is in progress in which axial dispersion coefficients in packed columns are measured under conditions similar to those in the proposed reductive extraction processes; in this program, mercury and water are used to simulate bismuth and molten salt. The measured axial dispersion coefficients will be used to estimate column performance in the proposed system. Devices for reducing axial dis- persion in packed columns are also under development. The equipment and experimental technique used for these studies have been described previously.15 The technique consists of injecting a tracer solution at a constant rate near the top of a packed column in which water and mercury are in countercurrent flow. The tracer tends to diffuse up- stream in the continuous (water) phase as a result of axial dispersion caused by the downflowing mercury droplets. The system is allowed to reach steady state, and then the concentration profile of the tracer in the continuous phase is determined. As indicated previously, a semilogarithmic plot of relative tracer concentration vs distance down the column produces a straight line having a slope that is inversely proportional to the axial dispersion coefficient. During this reporting period, measurements were carried out in a 2-in.-ID column having a length of approximately L ft. The column was packed with 1/4-in.-diam solid cylinders or Raschig rings or 1/2-in.-diam Raschig rings. The packing, which was made of either poly- ethylene or Teflon, is not wet by either mercury or water. T.1 Experimental Results The results of six runs made with 1/U-in. Raschig rings are shown in Figs. 22 through 27. Data for five runs with 1/L-in. solid cylindrical packing are shown in Figs. 28 through 32, and data obtained during 13 runs with 1/2-in.-diam Raschig rings are shown in Figs. 33 through 45. The data from all these runs are summarized in Tables 9-11, and are shown along with previous data from seven runs with 3/8-in.-diam Raschig ring packing in Fig. 4L6. In each case, the axial dispersion coefficient was independent of the dispersed-phase (mercury) flow rate. Dispersion coefficients for 3/8- and 1/2-in. packing were also independent of the continuous-phase (water) flow rate; their values were 3.5 and 4.8 cm3/sec, respectively. Data for 61 ORNL DWG 71-860l NORMALIZED TRACER CONCENTRATION (C/Cca“bm,e) 0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 22. Variation of Normalized Tracer Concentration with Distance Along a 2-in.-diam Column Packed with 1/k-in. Raschig Ring Packing During Run 1. 62 ORNL DWG 71-9474 2.0 cal ibrafe) NORMALIZED TRACER CONCENTRATION (C/C 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 23. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/4-in. Raschig Ring Packing During Run 2. cal ibrafe) NORMALIZED TRACER CONCENTRATION (C/C 63 ORNL DWG 71-947| 2.0 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.1 0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 24. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/b4-in. Raschig Ring Packing During Run 3. 6k ORNL DWG 71-8605 2.0 cal ibrate) NORMALIZED TRACER CONCENTRATION (C/C S 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 25. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/4-in. Raschig Ring Packing During Run L. 65 ORNL DWG 71-8595 R2 2.0 : cali brafé) NORMALIZED- TRACER CONCENTRATION (C/C °"o 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 26. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/b4-in. Raschig Ring Packing During Run 5. 66 ORNL DWG 71-9472 10.0 8.0 6.0 4.0 N o cali brcfe) o - o O o o 0.4 0.2 NORMALIZED TRACER CONCENTRATION (C/C 0.l 0.08 0.06 0.04 0.02 0.0l ‘ 0 5 I0 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 27. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/b-in. Raschig Ring Packing During Run 6. 67 ORNL DWG 71-8594 2.0 o © 0 0oo0 O ~N o V0 o o NORMALIZED TRACER CONCENTRATION (G/C__ ., ) o o N H 0"0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 28. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/4-in. Solid Cylindrical Packing During Run 1. 68 ORNL DWG 71-8598 2.0 cal ibrate) NORMALIZED TRACER CONCENTRATION (C/C "0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 29. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/hk-in. Solid Cylindrical Packing During Run 2. 69 ORNL DWG 71 —948l| 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 O.1 0.09 0.08 0.07 0.06 0.05 0.04 NORMALIZED TRACER CONCENTRATION (C/C__jip are) 0.03 0.02 0.0l o 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 30. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/4-in. Solid Cylindrical Packing During Run 3. 10 ORNL DWG 71-8599 calibrcfe> NORMALIZED TRACER CONCENTRATION (C/C 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 31. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/b-in. Solid Cylindrical Packing During Run L. T1 ORNL DWG 71-8600 calibrate) NORMALIZED TRACER CONCENTRATION (C/C 0.1 o 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 32. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/bk-in. Solid Cylindrical Packing During Run 5. T2 ORNL DWG 71-9480 cal ibrclfe) NORMALIZED TRACER CONCENTRATION (C/C "0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 33. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 1. T3 ORNL DWG 71-8608 2.0 o © O ooo o O g 0o o s o o NORMALIZED TRACER CONCENTRATION 2.0 ORNL DWG 71-8609 NORMALIZED TRACER CONCFNTRATION (C/C_ . . ) o 5 10 IS 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 36. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run L. 76 ORNL DWG 71-8597 2.0 calibrate) NORMALIZED TRACER CONCENTRATION (C/C S 10 15 20 25 30 38 DISTANCE FROM WATER EXIT (in.) Fig. 37. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 5. 1T ORNL DWG 71—-8607 2.0 cali brate) NORMALIZED TRACER CONCENTRATION (C/C "0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 38. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 6. 78 ORNL DWG 71-8596 NORMALIZED TRACER CONCENTRATION (C/Cca“bmte) 0 8 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 39. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 7. 9 ORNL DWG 71-9477 1.0 cal ibrafe) NORMALIZED TRACER CONCENTRATION (C/C 0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 40. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 8. 80 ORNL DWG 71-8604 2.0 ) calibrate NORMALIZED TRACER CONCENTRATION (C/C 0 S 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. b1. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 9. 81 ORNL DWG 71-8602 cal ibrate) NORMALIZED TRACER CONCENTRATION (C/C 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 42. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 10. 82 ORNL DWG 71-9475 cal ibrate) NORMALIZED TRACER CONCENTRATION (C/C 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 43. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 11. 83 ORNL DWG 71-9473 0.9 0.8 0.7 o o0 cal ibrate) o o 0.4 0.3 NORMALIZED TRACER CONCENTRATION (C/C 0.2 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. LL. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 12. 8l ORNL DWG 71-8603 O 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 NORMALIZED TRACER CONCENTRATION (C/C_ ., ) 0.l 0 5 10 15 20 25 30 35 DISTANCE FROM WATER EXIT (in.) Fig. 45. Variation of Normalized Tracer Concentration with Distance Along 2-in.-diam Column Packed with 1/2-in. Raschig Ring Packing During Run 13. 85 Table 9. OSummary of Axial Dispersion Measurements with 1/h=in. Solid Cylinders in a 2-in.-diam Column Run (mlyi%n) (mzfifign) (cm27sec) 1 350 23 1.41 2 . 350 L6 1.18 3 600 69 0.463 L 820 23 2.08 5 820 L6 0.669 Table 10. Summary of Axial Dispersion Measurements with 1/4-in.-diam Raschig Rings in a 2-in.-diam Column Run (mlyfifin) (m¥§%gn) (cmg7sec) 1 350 22.8 2.05 2 350 L7 1.33 3 350 68.4 0.918 L 820 22.8 2.16 5 820 L5.6 1.88 6 820 64.8 1.03 86 Table 11. Summary of Axial Dispersion Measurements with 1/2-in.-diam Raschig Rings in a 2-in.-diam Column Run (ml/min) (ml/min) (cm?/sec) 1 380 L6 6.48 2 380 68 L.78 3 655 23 3.25 L 655 L6 L.59 5 655 68 L.52 6 655 91 L.66 7 655 137 L.68 8 1060 L6 6.23 9 1060 L6 6.65 10 1060 91 L.56 11 1060 137 3.53 12 1060 137 L.38 13 1060 137 5.80 87 ORNL DWG 71-9545 0 ° T I T I ] 8- _ 7 A _ 6| A - 1/2 in. RASCHIG RINGS T R ——p——D— | . A 3/78in. RASCHIG RINGS A A —- 3 - [&] & N\ o . - < W © W W = _ S 09 : 0.8} — C_) O.7r— —t % 0.6[ — 3 05 — O 04l _ A 1/2in. RASCHIG RINGS 0.3k O /4 in. RASCHIG RINGS | O 174 in. SOLID CYLINDERS 0.2} - 0.1 | 1 L ] ] 10 20 30 50 70 100 200 WATER FLOW RATE (mi/min) Fig. 46. Variation of Dispersion Coefficient with Water Flow Rate for Several Packing Materials in a 2-in.-diam Column. 88 the 1/4-in. packing indicate that the dispersion coefficient is inversely proportional to the continuous-phase flow rate. Differences between the dispersion coefficients for the two 1/4-in.-diam packing materials are probably related to the difference in the packing void fractions of these materials. The data appear to portray two types of behavior, one for 1/k4- in. packing and one for larger packing. T.2 Comparison of Results with a Published Correlation Experimental data on axial dispersion in packed columns have been ob- tained by Vermeulen, Moon, Hennico, and Miyauchi,l6 who reported a correla- tion for predicting axial dispersion. Their study involved several fluids which had a range of physical properties and a number of packing materials; however, as the authors pointed out, the study did not include a sufficiently wide range of the difference in fluid densities to evaluate the effect of this variable. The data obtained during the present study afford a test of the Vermeulen correlation for fluids that have a large density difference. The present data are compared with the correlation of Vermeulen, Moon, Hennico, and Miyauchi in Fig. 47, where the dimensionless quantity chdp/De is plotted against the dimensionless quantity (Lp\)/dec)l/2 Vd/vc' The agree- ment of the present experimental data with this correlation is remarkable since the correlation was developed from data obtained with systems having density differences between one and two orders of magnitude less than that of the mercury-water system. The scatter of the present data around the curve is not significantly greater than that of the earlier data. Closer examination of the data reveals an apparent difference between the correlation and the present data with regard to the dependence of dis- persion coefficient on the flow rates of the two phases. However, it was not possible to cover a sufficiently wide range of mercury and water flow rates to confirm the suspected dependence. ORNL DWG 71- 123 RI 0.1~ ; T T T 1T 71 T T T T T T T 17117 T T T T 17 T T TTT] 8 |- ] B VERMEULEN, MOON ,HENNICO, AND MIYAUCH! CORRELATION — 6 — - q I~ — - . _ - ® ) L - 2 — B A o a B A — © LY ol 0.0l |- - 8 +— — e — 4 e — | & 1" scLip CYLINDERS ¢ =0.874 e =0314 . 2 e & RASCHIG RINGS ¥ =0.4 ¢ = 0.68 — " A | A 3 RASCHIG RINGS ¢ =0.4 ¢ = 0.66 | © 3 RASCHIG RINGS §=0.4 ¢ = 0.7I ol 2 4 6 8 10 2 4 6 8 10 2 4 6 8100 (¥ z Vd dec Vc Fig. 47. Comparison of Mercury-Water Data Obtained in This Study With the Vermeulen, Moon, Hennico, and Miyauchi Correlation. 68 10. 11. 12. 13. 1k, 15. 16. 90 8. REFERENCES M. J. Bell and L. E. McNeese, "MSBR Fuel Processing Using the Flu- orination--Reductive Extraction and Metal Transfer Flowsheet," Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (in press). L. M. Ferris, '""Measurement of Distribution Coefficients in Molten Salt--Metal Systems,'" MSR Program Semiann. Progr. Rept. Aug. 31, 1970, ORNL-L622, p. 20k. R. W. Kessie et al., Process Design for Frozen-Wall Containment of Fused Salt, ANL-6377 (August 1961). J. R. Hightower, Jr., C. P. Tung, and L. E. McNeese, "Use of Radio- Frequency Induction Heating for Frozen-Wall Fluorinator Development Studies," Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (in press). S. Cantor (Ed.), Physical Properties of Molten Salt Reactor Fuel, Coolant, and Flush Salts, ORNL-TM-2316 (August 1968), p. 1kL. J. W. Cook, MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL- 4396, p. 122. S. Cantor (Ed.), Op. Cit., p. 25. P. G. Simpson, Induction Heating, McGraw-Hill, New York, 1960, p. 1ul. E. L. Youngblood et al., "Metal Transfer Process Development," Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (in press). MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL-4396, p. 28k, MSR Program Semiann. Progr. Rept. Feb. 28, 1970, ORNL-4548, p. 291. R. B. Lindauer and L. E. McNeese, "Continuous Salt Purification Studies," Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 6, ORNL-TM-3141 (in press). W. S. Pappas, Anal. Chem. 38, 615 (1966). L. E. McNeese, Engineering Development Studies for Molten-Salt Breeder Reactor Processing No. 3, ORNL-3138, p. 3bi. Ibid., p. 72. T. Vermeulen, J. S. Moon, A. Hennico, and T. Miyauchi, Chem. Eng. Progr. 62, 96 (1966). » O o1 Ol W o 7T 78. 9. 80. 81l. 82. 83. 8. 85- 91 INTERNAL DISTRIBUTION C. F. Baes ho., J. H. Pashley (K-25) H. F. Bauman 43, A. M. Perry S. E. Beall L4-45, M. W. Rosenthal M. J. Bell 46. A. D. Ryon M. R. Bennett LT7. W. F. Schaffer, Jr. R. E. Blanco 48, Dunlap Scott F. F. Blankenship 49. J. H. Shaffer G. E. Boyd 50. M. J. Skinner R. B. Briggs 5. F. J. Smith R. E. Brooksbank 52. D. D. Sood K. B. Brown 53. Martha Stewart W. L. Carter 54. 0. K. Tallent H. D. Cochran, Jr. 55. R. E. Thoma F. L. Culler 56. D. B. Trauger J. R. Distefano 57T. W. E. Unger W. P. Eatherly 58. C. D. Watson D. E. Ferguson 5. J. S. Watson L. M. Ferris 60. A. M. Weinberg J. H. Frye 61. J. R. Weir W. R. Grimes 62. M. E. Whatley A. G. Grindell 63. J. C. White P. A. Haas 6L, W. M. Woods B. A. Hannaford 65. R. G. Wymer J. R. Hightower, Jr. 66. E. L. Youngblood C. W. Kee 67-68. Central Research Library R. B. Lindauer 69-70. Document Reference Section H. E. McCoy 71-T3. Laboratory Records L. E. McNeese T4. Laboratory Records, RC D. M. Moulton 5. Y-12 Document Reference Section J. P. Nichols 76. ORNL Patent Office E. L. Nicholson EXTERNAL DISTRIBUTION J. A. Accairri, Continental 0il Co., Ponca City, Oklahoma 74601 R. M. Bushong, UCC, Carbon Products Division, 12900 Snow Road, Parma, Ohio L4130 D. F. Cope, Atomic Energy Commission, RDT Site Office (ORNL) C. B. Deering, Black & Veach, P. O. Box 8405, Kansas City, Missouri 64114 A. R. DeGrazia, USAEC, RDT, Washington, D.C. 20545 Delonde R. deBoisblanc, Ebasco Services, Inc., 2 Rector Street, New York, N.Y. 10006 D. Elias, RDT, USAEC, Washington, D.C. 20545 Norton Haberman, RDT, USAEC, Washington, D.C. 20545 T. R. Johnson, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, Illinois 60L39 86. 87-88. 89. 90. 91. 92. 93. 9k. 95. 9. 97. 98-99. 92 EXTERNAL DISTRIBUTION (Continued) Kermit Laughon, Atomic Energy Commission, RDT Site Office (ORNL) T. W. McIntosh, Atomic Energy Commission, Washington, D.C. 20545 E. H. Okrent, Jersey Nuclear Co., Bellevue, Washington 98004 R. D. Pierce, Argonne National Laboratory, 9700 S. Cass Avenue, Argonne, Illinois 60439 J. Roth, Combustion Engineering Inc., Prospect Hill Road, Windsor, Connecticut 06095 M. Shaw, Atomic Energy Commission, 'Jashington, D.C. 205u45 N. Srinivasan, Head, Fuel Reprocessing Division, Bhabha Atomic Research Center, Trombay, Bombay T4, India C. L. Storrs, Combustion Engineering Inc., Prospect Hill Road, Windsor, Connecticut 06095 B. L. Tarmy, Esso Research and Engr. Co., P. 0. Box 101, Florham Park, N.J. 07932 J. R. Trinko, Ebasco Services, Inc., 2 Rector Street, New York, N.Y. 10006 Laboratory and University Division, ORO Division of Technical Information Extension, ORO, AEC