MS LIBRA Iil HA'tnl]iTI"jiiIilvnsis'Timl Il ulmil 3 445k 03b132kL 7 CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION ! '{ LIBRARY LOAN COPY ‘ DO NOT TRANSFER TO ANOTHER PERSON If you wish Someone else to see this document, send in name with document ] and the librory will arrange a loan, ORNL-2626 Reactors=Power TID-4500 (14th ed.) Contract No. W-7405-eng-26 MOL TEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT For Period Ending October 31, 1958 H. G. MacPherson, Program Director DATE ISSUED SANG 1959 OAK RIDGE NATIONAL LABORATORY Ock Ridge, Tennessee operated by | UNION CARBIDE CORPORATION | for the JE 1111 i 3 4456 03kL32L 7 MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT SUMMARY PART 1. REACTOR DESIGN STUDIES 1.1. Conceptual Design Studies and Nuclear Calcvlations Further studies of the Interim Design Reactor have resulted in an improved fill-and-drain system and alternate ways of fransferring heat from the re- actor fuel to steam. The revised fill-and-drain system retains the previously described after-heat removal scheme, but the components have been re- designed to make remote repair and removal feasible. As now designed the entire heat removal system may be removed or inserted from overhead without dis- turbing the fuel circuitry. A simplified fuel-to-steam heat transfer system is being considered that eliminates intermediate heat transfer fluids. The direct fuel-to-steam heat trans- fer system used in conjunction with a Loeffler boiler cycle allows control of the reactor and load relationship with conventional steam cycle com- ponents. Preliminary cost studies based on a 640- Mw (heat) two-region homogeneous reactor were made for several heat transfer cycles, including the direct fuel-to-steam cycle, and it was found that the direct cycle gave the lowest first cost. A helium-cooled cycle gave the highest first cost, but optimization of the reactor system would lower the cost appreciably. Nuclear calculations relative to a 30-Mw experi- mental reactor with a 6-ft-dia core have indicated that the critical mass of the system at 1180°F would be 49 kg of U235 and the total inventory for a system volume of 150 f#2 would be 65 kg of u23s, After a year of operation, the critical mass would have increased to 61 kg, with a net integrated burnup of 14.4 kg. Calculations of the initial states of reactors were completed. Data were obtained for critical mass, inventory, regeneration ratio, etc. for core diameters ranging from 4 to 12 ft and thorium fluoride concentrations in the core of 2, 4, and 7 mole % for 600-Mw (thermal) reactors. Additional calculations covering a 20-year operating period of the 8-ft-dia core with 7 mole % thorium showed that, after five years, the reactor is self-sustaining. Calculations were also made of the fissions in the blanket of the Interim Design Reactor. To date, calculations covering only the first five years of U233 fyeled operation have been completed, but the fraction of fissions in the blanket appears to have stabilized at about 0.02. Methods for calculating the nuclear characteristics of graphite-moderated, hetero- geneous, molten-fluoride-salt reactors are being developed. Gamma-ray energy absorption coefficients were computed as a function of energy for 24 elements: hydrogen, carbon, beryllium, nitrogen, oxygen, magnesium, sodium, aluminum, silicon, sulfur, phosphorus, argon, potassium, iron, calcium, copper, molybdenum, iodine, tin, tungsten, platinum, lead, thallium, and uranium. The calculated values have been assembled on paper tape for use with the gamma-ray heating program Ghimsr being written for the Oracle. 1.2. Component Development and Testing Development tests of salt-lubricated pump bear- ings were continued. Satisfactory molten-salt hydrodynamic bearing films were established in two tests of an INOR-8 bearing and an INCR-8 journal. Slight rubbing marks found after extended service are being investigated. Equipment for test- ing hydrostatic bearings with stationary pockets was completed, and data on bearing loads and flows were obtained in preliminary tests with water. Air en- trainment encountered in preliminary dynamic tests of a rotating-pocket hydrostatic bearing was cor- rected, and the pressure distribution in the pockets is being investigated. A conventional aluminum bearing and an Inconel journal that had been lubricated with Dowtherm ““A’ (a eutectic mixture of diphenyl and diphenyl- oxide) during 3200 hr of operation in a pump that was circulating NaF-ZrF -UF , at 1200°F was ex- amined. The clearance between the bearing and journal had not changed and there were only slight indications of wear. The oil-lubricated centrifugal-pump rotary element assembly being operated in a gamma-irradiation facility at the MTR had accumulated a total of 5312 hr of operation and a gamma-ray dose rate to the lower seal region of 9.3 x 107 r by the end of the quarter. Samples of the bulk oil and the seal leakage oil show that the viscosity has increased and that the bromine number has increased. The increases have not been sufficient to affect oper- ation, however. The acidity number of the oil in- dicates that little oxidation of the oil has occurred. The test will be terminated when the dose to the lower seal region has reached 1070 . Materials required for a motor suitable for long- term operation at 1250°F in a radiation field are being investigated. Six coil assemblies that in- corporate electrical insulation developed for use at high temperatures were received from the Louis- Allis Company for testing. Suitable magnetic core materials and electrical conductors are being sought, Fuel pump design studies that were contfracted to the Allis-Chalmers Manufacturing Company and the Westinghouse Electric Company were completed. The results of both studies emphasized the need for the development of salt-lubricated bearings. Various layouts of centrifugal pumps, both sump type and submerged, were suggested. A small, submerged, centrifugal pump with a frozen-lead seal is being operated in evaluation tests. During continuous isothermal operation of this pump for 2500 hr at 1200°F, with @ molten salt as the pumped fluid, there has been no leakage of lead from the seal. A similar lead-sealed pump with a 3]/4-in.-dia shaft is being fabricated for test- ing. Freeze-flange and indented-seal-flange joints that had sealed successfully in high-temperature molten- salt lines were tested with sodium. Both joints operated successfully; there was no indication of sodium leakage. Two large freeze-flange joints were tested in a 4-in. line carrying a high-temper- ature molten salt, There was no salt [eakage, but gas leakage of the flanges was slightly in excess of the allowable 107 cm? of helium per second. Modifications are now being made to improve the gas tightness. Heater-and-insulation units designed for use on a 4-in. line were tested along with the large freeze flange. The thermal loss from two units was found to be about four times the heat loss from an equi- valent length of ''Hy-Temp’' pipe insulation 3 in. thick. Three commercially available expansion joints were tested and found to be unsatisfactory for use in molten salt or sodium lines. Since the bellows that failed in two of the tests were weakened by the attachment of thermocouples, similar units will be tested without thermocouples. Data were obtained from which a plot of the sub- limation temperature vs vapor pressure was pre- pared, and the information was used in the design of a thermal-convection loop for evaluating the heat . transfer performance of aluminum chioride gas. Exposure of Inconel and INOR-8 specimens to aluminum chloride vapor for 1000 hr showed that either material would be satisfactory for con- struction of the thermal-convection loop. Construction and operation of forced-circulation corrosion testiag loops was continued. Of the thirteen facilities with operating loops, six are of the new construction that gives maximum protection against freezing of the salt in the event of power failures, and the remaining seven loops have been improved so far as possible without changing the loop piping. Assembly of the first in-pile loop was almost completed, and a second loop is being assembled. 1.3. Engineering Research The enthalpies, heat capacities, and heats of fusion were determined for the mixtures NaF-BeF _- UF , (53-46-1 mole %), LiF-BeF -UF , (53-46-1 mole %); NaCl-CaCl , (49-51 mole %), LiCI-KCI (70-30 mole %), LiCI-KCl {60-40 mole %), and LiCl-KCI (50-50 mole %). Preliminary values were obtained for the surface tension of LiF-BeF -UF , (62-37-] mole %) over the temperature range of 460 to 750°C. Apparatus for measurements of the coefficient of thermal expansion, thermal conductivity, viscosity, and electrical conductivity of beryllium-containing fluoride salt mixtures are being assembled. The cause of a major discrepancy in heat balance in the study of the heat-transfer coefficients of LiF- BeF ,-UF, (53-46-1 mole %) flowing in a small- diameter Inconel tube is being investigated. A pump system for investigating such long-range effects as film formation by comparison of heat- transfer coefficients is being constructed. 1.4. Instrumentation and Controls The results of a series of tests of Inconel re- . sistance-type fuel-level-indicating elements have not been reproducible, as yet, but it has been established that fuel 130 has sufficient resistance - to provide a useful milliwatt output from the probe. Polarization and surface tension effects are being investigated as possible sources of the incon- sistencies. 1.5. Advanced Reactor Design Studies Graphitesmoderated molten-salt reactors fueled with low-enrichment material were studied, and the results of preliminary calculations indicated that initial enrichments as low as 1.25% U233 might be used in a circulating-fuel system. Highly enriched uranium would be added as makeup fuel, and such reactors could probably be operated to burnups as high as 60,000 Mwd/ton. A simplified natural-convection molten-salt re- actor for operation at 576 Mw (thermal) was studied to determine the approximate size of components and the fuel volume. Natural-convection power re- actors of this size are characterized by high fuel volumes and large numbers of heat exchanger tubes. Therefore it is expected that the initial cost of a natural-convection reactor would be higher than that for a forced-circulation system, and the large number of heat exchanger tubes casts some doubt as to whether the natural-convection system would actually be more reliable than the forced-circulation system. PART 2. MATERIALS STUDIES 2,1. Metallurgy Two Inconel and four INOR-8 thermal-convection loops, which circulated various fused-fluoride-salt mixtures, were examined. An Inconel loop which operated 1000 hr at 1350°F with the mixture LiF- BeF,-UF , (62-37-1 mole %) was attacked to a depth of 3 mils. A loop operated previously with this mixture at 1250°F showed approximately 1 mil of attack over a similar test period. Operation of an Inconel loop at 1050°F for 4360 hr with the mixture NaF-LiF-KF (11.5-46.5-42 mole %) resulted in the formation of subsurface voids to a depth of 2 mils. No attack was evident in any of the INOR-8 loops even after operation, in one case, for 3114 hr. Test results were also obtained for three Inconel forced-circulation loops operated with fluoride mixtures. A loop which circulated the mixture NoF-Zer-UF4 (57-42-1 mole %) and also con- tained a secondary sodium circuit showed maximum attack to a depth of 1.5 mils in the salt circuit and only slight surface roughening in the sodium cir- cuits. No cold-leg deposits were present in either circuit. Mixtures of LiF, BeF, and UF produced attacks of 5 and 8 mils respectively in two other loops operated for slightly more than 3000 hr at a maximum hot-leg temperature of 1300°F. INOR-8 samples removed from a forced-circulation loop that was shut down after 3370 hr of operation for repair of a leak at a heater lug were found to have no evidence of surface attack. Etching rates showed slight differences between as-received and as-tested specimens, Operation of the loop has been resumed. The mixture LiF-Ber--UF4 (53-46-1 mole %) is being circulated in this loop. The maximum walil temperature is 1300°F. The failure occurred at a weld of the tubing to a heating lug adapter, and it was found that the wrong weld material had been used, Routine inspection of the INOR-8 material now being received has revealed that the quality is as good as that of Inconel and of stainless steels made to ASTM standards. The rejection rate for fully in- spected welds is now about 10%, A sodium-graphite system was used to prepare Inconel and INOR-8 tensile test specimens for a study of the effect of carburization on the mechanical properties, The carburization at 1500°F caused in- creases in both the yield and tensile strengths of Inconel and decreased the ductility at room temper- ature. For the INOR-8 specimens, the yield strength was increased, the tensile strength was slightly decreased, and the ductility was greatly reduced. The changes were dependent on the amount of car- burization. In similar tests at 1250°F the carburized specimens again showed a trend to fower ductility, but some specimens were more ductile than control specimens. In comparative tests of the carburization of INOR-8 by fue! 130-graphite systems, 3-mil surface cuts of specimens exposed only to fuel analyzed 640 ppm carbon, while similar cuts of specimens exposed to fuel containing a graphite rod analyzed 940 ppm carbon. Mechanical property tests showed the same trends as those found for the specimens carburized in sodium-graphite systems. For a test of the penetration of graphite by fuel 130, a graphite crucible 4% in. long with a tapered hole 0.43 in. ID and 3‘/2 in. deep was loaded with the fuel and tested in a vacuum at 1300°F. Radio- graphs were made at 500-hr intervals to study the penetration of the graphite by the fuel. At the end of the first 500 hr test of a CCN graphite crucible, there had been no penetration of the graphite, but a disk of UO, had formed at the bottom of the tapered hole that was 0.1 in. thick. Subsequent analyses showed that it contained 28% of the vranium originally present in the fuel. Further tests with TSF in contact with fuel 130 have shown similar precipitation of UQ, but to a lesser extent. Since the precipitation was least in the TSF graphite, which is purer than the CCN grade, it is thought that the precipitation may be caused by impurities in the graphite. The precipitation that occurred in these tests was not observed in the carburization tests described above. Further it is in contrast to the lack of pre- cipitation in similar tests in which fuel 30 (NaF- ZrF ,-UF ,, 50-46-4 mole %) was used. Metallographic examinations of the crucible showed no attack and no penetration of salt into the pores. Corrosion tests of pure silver and silver-base brazing alloys in static fuel 130 showed silver and its alloys to have very limited corrosion resistance. On the other hand, gold-bearing alloys have shown excellent corrosion resistance in fuel 130. A time-temperature-stress relationship was formu- lated for the creep strength of INOR-8 to aid in predicting long-time creep strengths. Heatsto-heat variation studies indicated a 0,14%-carbon-content heat to be appreciably stronger than two other INOR-8 heats containing 0.02 and 0.06% carbon. Eleven alloys with compositions bracketing the maximum amount of each major alioying element of INOR-8 have been vacuum-induction melted and are presently being aged at temperatures in the range of 900 to 1800°F. These alloys will be tested in order to determine the effect of composition on structural stability. Specimens of commercially fabricated INOR-8 were aged for 5000 hr at test temperatures of from 1000 to 1400°F. These specimens exhibited no in- stabilities in tensile tests that would be detrimental during long-time service. Room- and elevated-temperature mechanical property tests of INOR-8 weld metal both in the as- welded and in the welded and aged conditions were made and the results were compared with similar data for Hastelloy B and Hastelloy W. Aging seriously reduces the room-temperature ductility of Hastelloy B and Hastelloy W weld metal, while the ductility of INOR-8 weld metal is improved. The influence of melting practice and carbide spheroid- ization on the elevated-temperature ductility of INOR-8 is being studied. The brazing of thick tube sheets is being studied. Promising results were obtained in a test braze of vi a ¥%-in,-OD, 0.065-in.-wall Inconel tube into a 5-in.- thick Inconel tube sheet with Coast Metals No. 52 alloy. The brazing alloy was preplaced in annular trepans in the tube sheet and fed to the joint through . three small feeder holes. Methods of joining molybdenum and nickel-base materials are also being studied for pump application. 2,2. Radiation Damage The electrically heated mockup of the INOR-8 in-pile thermal-convection loop was operated under simulated in-pile conditions. This experiment served as a performance test for the various com- ponents of the loop and indicated that modifications of the cooling=air injection system were required. The in-pile model of the loop is being assembled. Two fluoride-fuel-filled graphite capsules were examined that had been irradiated in the MTR for 1610 and 1492 hr at 1250°F. The graphite was un- damaged. 2.3. Chemistry Phase equilibrium studies of fluoride-salt systems containing UF , and/or ThF , were continued. Data are being obtained on the NaF-Ber-ThF4-UF4 system to provide a comparative basis for evaluation of lithium-base quaternary systems. Studies have revealed the cause of low concen- tration of uranium in 50-lb batches of LiF-BeF - UF4 (62-37-1 mole %) that were transferred from 250-1b batches. The first liquid to form when the - mixture is melted is the eutectic composition con- taining 3.1 wt % uranium rather than the nominal 6.5 wt % uranium in the 250-1b batch. Thus the composition of any partially molten batch must be intermediate between that of the lowest melting eutectic and the nominal composition of the liquid. Complete melting of the mixture is necessary to maintain composition homogene ity. Extensive gradient-quenching experiments were completed on the LiF-BeF ,-ThF, system. Asa result of these experiments several modifications were made in the preliminary phase diagram issued - previously. Three-dimensional models of this system and of the previously studied system LiF- ThF -UF , were constructed. The effects of divalent and trivalent fission- product ions on the solubility of PuF, in LiF-BeF, mixtures was studied, The effect of the trivalent ions was found to be sufficient to indicate that they should not be allowed to build up in the fuel, but the effect of the divalent ions was insignificant. Data were obtained that provide a basis for the use of trivalent ions in a scheme for reprocessing fused-salt mixtures containing plutonium. Solubility studies were made on several systems, including argon in LiF-BeF, (64-36 mole %); HF in LiF-BeF , mixtures containing 10 to 50 mole % Ber; Cer in Nc:F-LiF--BeF2 solvents of various compositions; CeF3, LoF3, and SmF3 in LiF-Ber- UF, (62.8-36.4-0.8 mole %); CeF, and LaF, simul- taneously in LiF-BeF -UF, (62.8-36.4-0.8 mole %); and CeF3 and SmF ; simultaneously in LiF-BeF ,- UF, (62.8-36.4-0.8 mole %). An elementary model was studied with which estimates can probably be made of the solubilities of the noble gases in media of high surface tension to within an order of mag- nitude, which is sufficiently accurate for reactor applications. Two methods of separating uranium from un- desirable fission products in fluoride-salt fuel solvents are being studied. Experiments are in progress in which a chromatographic type of separation of uranium and fission-product rare earths is effected by using beryllium oxide as the column packing material. In other experiments uranium precipitation upon the addition of water in the influent gas stream is being investigated. Tentative values were obtained for the activities and the activity coefficients of nickel metal in nickel-molybdenum alloys. The values were calcu- lated from measurements of the electromotive force of cells containing pure nickel and nickel-molybdenum alloy electrodes in a bath of NaC|-KCl eutectic con- taining 5 wt % NiCI2. Apparatus was set up for measuring the surface tensions of BeF , mixtures. Trial runs gave a value of 195 dynes/cm at 425 to 450°C for LiF-BeF ,-UF (53-46-1 mole %) and 196 dynes/cm at about 480°C for LiF-BeF2 (63-37 mole %). In the study of the chemical equilibria involved in the corrosion of structural metals, the effect of solvent composition on the equilibria f‘)Cer,?:"‘?CrF3 + Cr® and 3FeF2—\‘\__ 2FeF , + Fe® was investigated. The solvent LiF-NaF-ZrF was used because it permitted a continuous variation from a basic to an acidic melt without interference from precipitation. Increasing disproportionation ot CrF, was found with decreasing ZrF , content. Similar tests in solvents containing BeF ,, which is considered to be more basic than ZrF ,, showed that in an LiF-BeF, mixture containing 52 mole % BeF, the disproportionation of CrF, was roughly comparable to that in a solvent containing 35 mole % LrF . These results confirmed the supposition that BeF, mixtures are more basic than correspond- ing ZrF , mixtures and that the extent of dispro- portionation of (:rl:2 is greater in the more basic solvents. Experiments carried out with FeF, in the same solvents showed no disproportionation. Further measurements of the activity coefficients of CrF, dissolved in the molten mixture NaF-ZrF (53-47 mole %) were determined and equilibrium quotients were calculated. The results of the measurements of CrF_ and the corresponding in- vestigations of NiF, and Fer, along with the values for the activities of chromium in alloys, can now be used in the study of corrosion of Inconel and INOR-8 alloys. [t is thought that the rate of corrosion in these alloys is diffusion controlled. Calculations were made of the amounts of chromium metal which should be added to pre- equilibrate a salt prior to a test in an Inconel loop. The corrosion by a pre-equilibrated salt should de- pend only on the hot-to-cold-zone transfer mechanism, and the results should be indicative of long-term cofrosion rates in reactors, Studies of chromium diffusion in chromium-con- taining nickel-base alloys were continued. It is hoped that the results of these experiments will provide a means for predicting the void penetration distances to be expected for a given set of cor- rosion conditions. An experimental study was made of the corrosion of Inconel by aluminum chloride vapor in an all- metal system. The results indicated that the amount of attack of Inconel can be expected to be pro- portional to the pressure of the aluminum chloride vapor and that at 5 atm the corrosion after 300 hr would be about 0.7 mil. INOR-8, in which the activity of chromium is much lower than in Inconel, should prove quite resistant to aluminum chloride vapor. Other alloys selected on the basis of chemical and structural considerations are to be tested, The containment of aluminum chloride is of in- terest because the gas has unique properties which make it a heat transfer medium of potential utility vii at high temperatures. First, because of the large number of atoms per molecule, the heat capacity of the gas is high. Further, because at high temper- atures the major species is AICl, and at lower temperatures the major species is Al Cl, cooling the gas in a heat exchanger will remove, in addition to the heat obtained from the temperature change, the heat from the dissociation reaction, Estimates of the physical properties of the gas were made to provide a basis for the design of a thermal-con- vection loop in which to test the heat transfer characteristics and the corrosiveness of the gas. In tests of the penetration of graphite by molten salts, it was found that NaF-LiF-KF penetrated to greater extent than LiF-MgF ,, as indicated by weight gains. The NaF-LiF-KF mixture has a lower density and a lower viscosity than the LiF-MgF, mixture. In both cases there was even penetration to the center of the graphite rod. A new method of preparing chromous fluoride was explored in which the reaction is SnF, + Cr—CrF, + Sn The chromium-containing portion obtained in the viii experiment had the crystallographic properties of pure chromous fluoride. Improved methods for pre- paring vanadium trifluoride and ferrous fluoride were also developed. 2,4, Fuel Processing Sufficient laboratory work has been done to con- firm that fluorination of the fuel salts LiF-BeF ,- UF , or LiF-BeF -UF ,-ThF , results in good uranium recovery. Therefore the fluoride volatili- zation process, which was developed for hetero- geneous reactor fuel processing and was used successfully for recovery of the uranium from the fuel mixture (NaF-ZrF4-UF4) circulated in the Air- craft Reactor Experiment, appears to be applicable to processing of fuels of the type now being con- sidered for the molten-salt reactor. Developmental work has been initiated on the processing of the solvent salt so that it can be re- cycled. The recovery process is based on the solubility of LiF-BeF, in highly concentrated aqueous HF and the insolubility of rare earth and other polyvalent-element flueorides. CONTENTS SUMMARY .ottt ettt e e ets b e tee b e et e s ae st e b et beeae s o4eshb e bebe e st e sabete 4 ebmenbeentebe e seenseanserneseseennenneesenneene iii PART 1. REACTOR DESIGN STUDIES . 1.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS ....cccoeiiiiiiiinincniic 3 Interim Design Reactor ...ttt s 3 Fuel Fill-and-Drain Sy stem ... s 3 Fuel-to-Steam Heat Transfer ...ttt et s 3 Comparative Costs of Several Systems ... 4 Experimental Reactor Calaulations .......ccieciiiecienierei et s 6 NUCTEAr CalEUlHOoNS c.oo ottt e e et eee e easeae e st s e s eaneeeebe kb e e aas e bbb bt iasera e e s 6 Nuclear Characteristics of Homogeneous, Two-Region, Molten-Fluoride-Salt Reactors Fueled with UZ33 ettt es st b b st e 6 Blanket Fissions in the Reference Design Reactor ... 9 Comparison of Ocusol and Cornpone Caleulations ... 10 Heterogeneous ReaCIOrs ....o.coiviiieiiiiiiii ettt 11 Gamma-Ray Heating Calculations ... 11 1.2. COMPONENT DEVELOPMENT AND TESTING oottt 19 Fuel PUmp Development ...ttt bttt e bbbt 19 Development Tests of Salt-Lubricated Bearings......coocoveooiiiiiic 19 Development Tests of Conventional Bearings.......ccoiiniimiiiin i, 20 MEChANTCAI SEAIS ettt ettt et et esr e ettt er e ae s b rbs b s e ab s R e e e e et e e e et 21 Radiation-Resistant Electric Motors for Use at High Temperatures ..o 22 Design Studies of Fuel PUMPS .. et 22 . Frozen-Lead Pump Seal ..ot 23 Development of Techniques for Remote Maintenance of the Reactor System ... 25 Mechanical Joint Development ... i s s 25 i Evaluation of Expansion Joints for Molten-Salt Reactor Systems ..o, 32 Remote Maintenance Demonstration Facility....cociiiii e, 35 Heat Exchanger Development ... ..ot 37 ALCl, Thermal-Convection LOoop ... rineissiinis it e 37 Moflten-Solf Heat Transfer Coefficient Measurement.....o.coiiiriiiiiiiii e 38 Design, Construction, and Operation of Materials Testing Loops .o 39 Forced-Circulation LOOPS . oiiieieireeie ettt b e e e 39 [oPi 1@ LLOOP'S «eveeareuietiseesree st ceetsi e eeeeme bbb s h bbb bbb SRS 42 1.3. ENGINEERING RESEARGCH ...ooctocee ettt ettt ieteetec st b e st st en bbb st 44 Physical Property Measurements ... ..o i 44 - Enthalpy and Heat Capacity et st 44 S UPFGCE T NS TON otitvtereresssseeeeee e e eeeets et sasbasssaas samseanseeases b et b e s b e e e b e e R e s e s na e Ls e b4 e ab e e e e e be s sa e eas b s b e s ne b0 45 Apparatus Fabrication and Calibration ... 46 Molten-Salt Heat-Transfer SHUIES ..o ireierert et ettt s s eb e b et et s s s 46 1.4, INSTRUMENTATION AND CONTROLS oottt e 47 Resistance-Type Continuous Fuel Level Indicator ..o 47 1.5. ADVANCED REACTOR DESIGN STUDIES oottt et srae s ssaessssenssesmnesenesevesanssnnessnn 48 Low-Enrichment Graphite-Moderated Molten-Salt Reactors ......cccvemieiiviiniiniei i 48 A Natural-Convection Molten-Salt ReQCtor ....ciiiciii ittt et besses e s st s samae s s 48 2.1. 2.2, 2.3. MET ALLURGY ottt ettt ta et et e sttt e st s evtaeee s eteenbaeebeaasasseenssenseaseeseesseassebaeereensansbennnenten 53 Dynamic Corrosion StUAIies ..ttt et as b e ers e snrereesre et e snnanbes 53 Inconel Thermal-Convection Loop Tests ..ottt eer e e 53 INOR-8 Thermal-Convection Loop Tests ..t cere et e s re s sas et e annas 53 Inconel Forced-Circulation Loop T @SS oot cciicccctrievctes e ceres e veevsrsneesstessnneessreessneessnnnssnsres 53 Results of Examination of Samples Removed from INOR-8 Forced-Circulation 00D D354 T ettt st ets e et s b e ra e e e eatee e abeaearbbeareaenba e reneenteeabeaeanbebeas 56 MOtErIAl [NSPECTION eiiieiiiiieeicieie s et e et e rers e e s s nr e res e e s s ane e s e te e e s resaenbeseea ssnanaesareessernenerranteens 57 General Corrosion StUdI@s.......ooiiici it ce e s e e v e s e sbaea s er s e st e et e e e e e e teebsenraennrn 57 Carburization of Inconel and INOR-8 by Sodium-Graphite Systems ......ccccooeeiinvccrviinnnicineenne, 57 Carburization of INOR-8 by Fuel 130—Graphite Systems......coooiiiiiiiiiiiee e 59 UO, Precipitation in Fused Fluoride Salts in Contact with Graphite ........ccooniviiiiiinincn, 61 Precious-Metal-Base Brazing Alloys in Fuel 130 .o 64 Mechanical Properties of INOR-=8. ...t et et et eesser e s are s ne e 64 Influence of Composition on Properties of INOR-8 ..o 65 High-Temperature Stability of INOR-8......c..cveieiiicie e 67 Welding and Brazing Studies. ...ttt 68 INOR-8 Weldability ...ciciieieereierireieicee sttt s sim e e st bttt s b s ma s sbs s ba st 68 Brazing of Thick Tube Sheets for Heat Exchangers. ..., 71 Fabrication of Pump Components .......c.ueiiiier ittt sa bbb s 73 RADIATION DAMAGE ..ottt et e se st se e e st ree e e et e s e st es e et e e ete et e st e be e es s ese e s e e b e aabesn e e s n st a s s sanis 76 INOR-8 Thermal-Convection Loop for Operation in the LITR .. 76 Graphite Capsules Irradiated in the MTR ... e 76 CHEMI S T R Y oottt ettt e te sttt e b e eabeeteabs £ ert e aas e et e easear et e ane e sre s e s e e eamn s s ne st ene e e e st e st b e st b s sane b s nanass 78 Phase Equilibrium SHudies ..ooooiiiiiiiiiiiii s e s 78 Systems Containing UF4 and/or ThF4 ............................................................................................ 78 Solubility of PuF, in LiF-BeF, Mixtures ... s 80 FiSSiON-Product BehaVior ..ot ettt et e e ettt e eebe et s bbb s bbb er s e b ases pene 83 Solubility of Noble Gases in Molten Fluoride Mixtures ... 83 Solubility of HF in LiF-BeF, Mixtures ..., 85 Solubilities of Fission-Product Fluorides in Molten Alkali Fluoride—Beryllium FIUOEIE SOIVENES oot eti e et e et s ete b en st e rae e s e et e bt ereesaesas e st s s e s ebe ot e s bn bt e beere peas 87 A Simple Method for the Estimation of Noble Gas Solubilities in Molten FlUOTIAE MIXPUTES ooneeeeeeeiieieeiiieseei st e st e v e st e e e taeas e s e teesse sebe s eas s eatesaresamaan b e s reeabas s ssatsrnssrnnesanssnesns %0 Chemical Reactions of Oxides with Fluorides in LiF-BeF ..o 92 Chemistry of the Corrosion Process ... s 93 Activities in Metal Aoy s .o e 93 Surface Tensions of BeF , Mixtures ..., 94 2.4. Disproportionation of Chromous FIUoride..........uuviiiiiieiiieioiiie e e 94 Effect of Fuel Composition on the Equilibria 3CrF2\=‘ 2CrF, + Cr° and 3FeF2: 2FeF 5 4+ Fel s 95 Activity Coefficients of CrF, in NaF-ZrF 4o, 96 Equilibrium Amounts of CrF, in Molten Salts Containing UF; Which are in Contact With INConel ... ettt et eeene e 98 Chromium Diffusion in ATIoys ..ot et ettt s b eraas s b eas s e 99 Corrosion of Metals by Aluminum Chloride........occooviiiiiiiiiiiie et 100 Corrosion of Inconel by Aluminum Chloride Yapor in a Fused Silica Container ........ccoe...... 100 Corrosion of Inconel by Aluminum Chloride Vapor in an All-Metal System.......cccccvvveecivvvnnnnene, 100 Significance of Experimental Results ..o 101 Theoretical Considerations of Aluminum Chloride Vapor Corrosion .......ccccoveeeieerivicecniecrennnen, 101 Structural Metal Considerations ... s e s s e e e s e snerenes 102 Gaseous Aluminum Chloride as a Heat Exchange Medium .....ccooecviicriiniiiiiceeec e 103 Permeability of Graphite by Molten Fluoride Salts ..o 107 Preparation of Purified Materials........c.cciiiiiiiiiiiiiie st se e sresree s 107 Preparation of Pure Fluoride Compounds.......coceieiieriiiiiniiiccnnneneesrcnce e e e 107 Production-Scale Operations .......cciieciioriioiiiciiereiesctesse s ese s e se e s sar e s e ensseacsaessesennes 108 Experimental-Scale Operations ...ttt e et 108 FUEL PROCESSING ..ottt e stebs e ettt er et sb e et ets e b st enenb s esteseen e nae e e nseenesseenebeesns 110 Flowsheet for Fluoride Volatility and HF Dissolution Processing of Molten-Salt Reactor Fluids .ocoiviieriieeee e ettt e s aa s a e er et 110 Experimental Studies of Volatility Step ...ovvieiiiiiiineciec e 110 Solubilities of LiF-BeF, Salts in Aqueous HF ..oy 112 x1i Part 1 REACTOR DESIGN STUDIES 1.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS H. G. MacPherson Reactor Projects Division INTERIM DESIGN REACTOR A study has been made of improvements and variations of the Interim Reactor Design described in ORNL-2634 and, briefly, in the previous quar- terly report.! Specifically, as described below, an improved fill-and-drain system has been de- signed, and alternate ways of transferring heat from the reactor fuel to steam have been con- sidered. Fuel Fill-and-Drain System G. D. Whitman Design studies of the fuel fill-and-drain system for the interim design reactor were described pre- viously, "2 and further study of possible configu- rations has resulted in a system that eliminates maintenance difficulties inherent in the initial design. The original concept of pipes and water walls resulted in a relatively low first cost sys- tem; however, the components were not readily accessible for repair. In addition the entire system was contained in a gastight thermally in- sulated volume that complicated primary entry for maintenance operations. The new concept retains the after-heat removal scheme, that is, radiant heat transfer to a boiling water system, but the components have been re- designed so that overhead accessibility makes remote repair and removal more practical. The fused salt is contained in a cylindrical tank into which a number of vertical bayonet tubes are in- serted. These tubes are capped at the bottom and welded into the top of the vessel, which forms a tube sheet. These tubes serve as the primary after-heat removal radiating surface and contribute substantial nuclear poison to the geometry. The water boiler, which consists of a number of mating double-pipe bayonet tubes projecting down- ward from a combination water-and-steam drum, rests on the top of the drain vessel. The boiler tubes fit inside the tubes in the fuel vessel, and 1H. G. MacPherson, Molten-Salt Reactor Program Status Report, ORNL-2634 (Nov. 12, 1958), and MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 3. 26, D. Whitman, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 15. the combination forms a radiant heat exchange system. This concept retains the double con- tingency protection ogainst fluid leakage from either system, Electric heaters and insulation are installed around the fuel vessel for preheating. The boiler tubes are flooded with water for after-heat removal, and the rate of heat removal may be controlled by the boiler ‘‘water rate’” or by dividing the boiler into sections which may be used in combinations to control the rate and locations of heat removal in the fuel vessel. The entire boiler system may be removed or inserted from overhead without disturbing the fuel circuitry. A failed fuel tube could plug at the tube sheet and only failure of the vessel walls would necessitate complete removal of the system. The drain system required for a 600-Mw (thermal) molten-salt reactor system was considered, and it was found that four vessels 5 ft in diameter and 12 ft high would be needed to contain the 600 ft3 of salt from the reactor. Each of the vessels would be capable of removing 1.8 Mw of after heat within the practical temperature limitations of the container materials. Criticality calculations were made at off-design conditions. The nuclear poison contribution of the removable boiler tubes was neglected and the bayonets projecting into the fuel tanks were con- sidered to be flooded with water. It was de- termined that approximately 60 lb of 1.5% boron steel would have to be placed in the system to give multiplication constants below 0.5 with the probable U239 concentrations in the fuel salt. Fuel-to-Steam Heat Transfer M. E. Lackey A direct fuel-to-steam heat transfer system has been studied for application in a molten-salt power reactor. The complete heat removal sys- tem consists of four circuits in parallel to remove the heat from the fuel and a single circuit to re- move the heat from the blanket. Each of the cir- cuits is separate and independent up to the point where the superheated steam paths join ahead of the turbine. Each circuit has a salt-to-steam heat MOLTEN-SALT REACTOR PROGRESS REPORT exchanger located in the reactor cell and an evap- orator and steam compressor located adjacent to the reactor cell. A simplified flow diagram of the heat transfer system is shown in Fig. 1.1.1. This system is entirely free of sodium and its associated prob- lems. The fuel-to-steam heat couple in conjunc- tion with the Loeffler boiler cycle provides a differential modulating thermal block between the fuel and the water which allows complete and conventional control of the reactor and load re- lationship. Comparative Costs of Several Systems G. D. Whitman Preliminary cost studies have been made for a number of heat transfer cycles for a molten-salt power reactor, and the capital costs for seven heat transfer systems are summarized in Table 1.1.1. In all cases the system was assumed to include a fixed-size, two-region, homogeneous reactor in which 640 Mw of heat was generated and then transferred to a reheat-regenerative steam cycle, The steam conditions were set at 1800 psia and 1000°F reheat. In all cases four parallel heat transfer loops and four steam generating systems were coupled to the fuel circuit, and a single heat transfer loop and steam generator were used in the blanket circuit. Reheat was supplied by the fuel system exclusively in all the systems. The reactor plant portion of the cost summary included the shielding, containment vessel, in- strumentation, remote maintenance equipment, auxiliary fluid systems, original fluid inventories, and the chemical plant equipment, in addition to the reactor and heat transfer circuitry. The con- ventional plant costs included the costs of land, structures, steam system (not including boilers, superheaters, or reheaters), turbine-generator, accessory electrical equipment, and miscellaneous power plant equipment. The general expense item listed in Table 1.1.1 was assigned to cover the design and indirect costs incurred during the construction and startup phases of operation. It was set at approximately 16.5% of the reactor and conventional plant sub- total. The heat transfer linkage used in each system is indicated, and in order to compare the costs more Table 1.1.1. Molten-Salt Recctor Costs Net electrical output: 260 Mw 0.80 Plant load factor: Interest rate on capital investment: 14% per annum Reactor with Reactor with Reference Interim Reactor with Natural- Design Reactor Design Reactor Loeffler System Loefter System Loetfler System Gas-Cooled Convection Reactor (Na-Steam) (Fuel-Steam) Reactar {Na-Na-Steam) (No-Steam) (No-Steam) {Steam-Steam (Steam-Steam {He at 300 psia) (Na-Na-Steam) {Na-Na-H,0) (Na-Na-H,0) (Na-5team Reheat} Reheat) Reheat) (Na-Na-H ,0) Reactor plant $20,232,000 $19,600,000 $17,432,000 $17,100,000 $15,688,000 $26,234,000 $24,500,000 Contingency (40%) 8,093,000 7,840,000 6,973,000 6,840,000 6,275,000 10,493,000 9,800,000 Subtotal 28,325,000 27,440,000 24,405,000 23,940,000 21,963,000 36,727,000 34,300,000 Conventional plant 29,350,000 29,350,000 31,200,000 31,850,000 32,400,000 29,350,000 28,250,000 Contingency (7,5%) 2,201,000 2,201,000 2,340,000 2,389,000 2,430,000 2,201,000 2,119,000 Subtotal 31,551,000 31,551,000 33,540,000 34,239,000 34,830,000 31,551,000 30,369,000 Reactor and conventional 59,876,000 58,991,000 57,945,000 58,179,000 56,793,000 68,278,000 64,669,000 plant subtotal General expense 9,950,000 9,800,000 9,625,000 9,650,000 9, 440,000 11,350,000 10,750,000 (including design) Total cost $69,826,000 $58,791,000 $67,570,000 $68, 629,000 $645, 233,000 $79,628,000 $75,419,000 Fuel volume, ft? 575 555 589 589 777 766 1765 Fixed cost contribution 5.37 529 5,20 529 5.09 6.13 5.80 ta power cast, mills/kwhr BLANKET @—» 1210°F 23 psia REACTOR BLANKET UNCLASSIFIED ORNL-L.LR-DWG 34504 408,100 Ib/hr Fig. 1.1.1. Flow Diogrom of Fuel-to-Steam Heat Transfer System For Interim Design Reactor. EMERGENCY RELIEF LEAK-DETECTION SYSTEM [} —100 psia FUEL FROM FUEL [ ' AND BLANKET TURBINE STOP RN ’ B HIGH -PRESSURE |_83 osia ATTEMPERATOR 408,1001b/hr - TURBINE . 1805 psia 260psia 78 psio— 333,000 1b/hr E‘ T iINTERMEDIATE- p Q22°F AND LOW-PRESSURE t < 240 psia TURBINES SUPERHEATER , REHEATER = | = % 1 COOLING : WATER 621°F 881°F RSETT’TC'NG CONDENSER 40 psia—1 — 1854 psia 1800 psia — ION 1,492,000 Ib/hr 1,083,900 Ib/hr DESUPERHEATER CONDENSATE PUMP \j N d Lt J 1050°F 13.5 cfs D ——- STEAM PUMP M ENTRAINMENT SEPARATOR - FEEDWATER ~ 17\ —- HEATERS IN SERIES [ [ BOILER } 1 P F— DEAERATOR (OF 3 PER FUEL CIRCUIT i BLANKET FUEL FEEDWATER .l PUMP 475°F FEEDWATER HEATERS IN SERIES 8561 ‘L€ ¥390L20 9NIGNI dOl¥3d MOL TEN-SALT REACTOR PROGRESS REPORT easily certain design features were fixed, as pre- viously described. More nearly optimized systems would probably result in lower costs; however, with the exception of the gas-cooled cycle, it was not evident that optimization studies would sig- nificantly alter the relative costs. The Loeffler direct fuel-to-steam system gave the lowest first cost. In this case the elimination of a sodium circuit resulted in a net saving even though the fuel-to-steam heat exchanger was more expensive than the heat exchangers in the sodium- coupled system. The helium-cooled cycle resulted in the highest first cost, This was due primarily to the relatively large amount of heat transfer surface required in the fuel-to-helium and helium-to-water-to-steam heat exchangers. This particular system would probably be the most sensitive to optimization studies. In particular, increasing the coolant gas pressure and using gases other than helium could reduce the size of the heat transfer equipment. Differences in fuel volume will also be reflected in fuel cycle costs, which are not included in the capital costs shown in Table 1.1.1. A first approximation, based on changes in fuel inventory charges alone, indicate that each additional 100 ft3 of fuel volume would add about 0.06 mill/kwhr to the cost. On the basis that each plant had the same net electrical output, the maximum deviation in the power cost contributed by the capital charge was found to be approximately 1 mill/kwhr. EXPERIMENTAL REACTOR CALCULATIONS J. W. Miller A 30-Mw experimental reactor is being con- sidered that would have a 6-ft-dia core enclosed in a ]/Z-in.-fhick INOR-8 pressure vessel. Sur- rounding the core there would be a 6-in.-thick layer of insulation, a 6-in.-thick air space, and five 2«in,-thick boron steel thermal shields cooled by 4-in.-thick layers of diphenyl. Nuclear calculations indicate that the critical mass at 1180°F would be 49 kg of U233, and the total inventory for a system volume of 150 43 would be 65 kg of U235, After a year of operation, the critical mass would have increased to 61 kg, with a net integrated burnup of 14.4 kg. Approximately 7% of the neutrons would be captured in the core vessel. This compares with 5% for the 600-Mw reference design reactor. Weighting these numbers by the reactor power levels, indicates that the heat-generation rate in the experimental reactor would be approximately /% of the heat-generation rate in the reference design reactor. ) Approximately 31% of the neutrons would escape from the core vessel, and these would ail be cap- tured by the time they reached the fourth boron steel thermal shield. NUCLEAR CALCULATIONS L. G. Alexander Nuclear Characteristics of Homogeneous, Two-Region, Molten-Fluoride-Salt Reactors Fueled with U233 The calculations entailed in the investigation of initial states of U233.fueled reactors were com- pleted, and it was discovered that serious dis- crepancies existed between results obtained from the Oracle program Cornpone and the UNIVAC program QOcusol, especially when the results were used as input for the Oracle burnout code, Sorghum. The difficulty was traced to fluctuating end-point fluxes generated by the Cornpone program, Ac- cordingly, the Cornpone cases were discarded, and the series of calculations was completed with the Ocusol program. Meanwhile, the defect in the Cornpone program was remedied by the author, W. E. Kinney, and the two programs are now in fair agreement, as discussed below. The results of the calculations are given in - Table 1.1.2, where critical mass, inventory, re- generation ratio, etc., are given for a series of cases covering core diameters ranging from 4 to 12 ft and thorium fluoride concentrations in the core of 2, 4, and 7 mole %. The regeneration ratio is plotted against the critical inventory for a 600-Mw (thermal) system in Fig. 1.1.2. The dotted line in Fig. 1.1.2 is the envelope of the curves shown and is the locus of optimum conversion for a given inventory. As may be seen, the 8-ft-dia cores are nearly optimum at all thorium concen- trations, and initial regeneration ratios of up to . 1.08 can be obtained with inventories ranging up to 1300 kg of U233, The subsequent performance of the 8-ft-dia core with 7 mole % thorium was studied by means of the Oracle code Sorghum. Extracts from the cal- culated results are given in Table 1.1.3, where inventories, absorption ratios, regeneration ratios, etc., are given for periods of operation of up to PERIOD ENDING OCTOBER 31, 1958 Table 1.1.2, Initial Nuclear Characteristics of Two-Region, Homogeneous, Molten-Fluoride-Salt Reactors Fueled with U233 Fuel salt: 37 mole % BeF, + 63 mole % LiF + UF,+ ThF, Blanket salt: 13 mole % ThF, + 16 mole % BeF, + 71 mole % LiF Total power: 600 Mw (heat) External fuel volume; 339 3 Case No, 51 52 53 54 55 56 57 58 59 60 61 62 63 64 85 Core diameter, ft 4 4 4 6 6 6 8 8 8 10 10 10 12 12 12 ThF , in fuel salt, mole % 2 4 7 2 4 7 2 4 7 2 4 7 2 4 7 U233 in fuel salt, mole % 0.619 0.856 1.247 0.236 0,450 0.762 0.152 0.316 0.603 0.121 0.262 0.528 0.101 0.222 0.477 U233 51om density, atoms/em> 19.8%10'7 27.4x 10'7 39.9% 10" 7.55x 10"7 14.4x10'7 24.4% 10" 4.88 10" 10.1x10"7 19.3x 10" 3.8 % 10" 8.39%1017 16.9x10'"7 3.24x10"7 7.39x 10" 1525% 10'7 Critical mass, kg of U233 72.0 100.5 146.5 94.2 177.8 301 143 299 566 221 481 970 320.5 733 1510 Critical inventory, kg of U233 810 1118 1630 374 7N 1204 325 677 1283 364 793 1600 441 1007 2078 Neutron absorption ratios* U233 (fissions) 0.871 0.874 0.881 0.864 0.868 0.876 0.867 0.865 0.873 0.87 0.864 0.871 0.873 0.864 0.870 U233 (n,) 0.129 0.126 0.119 0.136 0.132 0.124 0.133 0.135 0.127 0,130 0.136 0.129 0.127 0.136 0.130 Be-LiF in fuel salt 0.070 0,066 0,069 0.093 0,075 0.076 0.120 0.082 0.078 0.142 0.088 0.081 0.164 0.093 0.083 Core vessel Be-Li-F in blanket salt 0.129 0.111 0.095 0.101 0.080 0.062 0.086 0.059 0.043 0.062 0.044 0.031 0.049 0.032 0.020 Leakage Th in fuel salt 0.343 0.426 0.517 0.581 0.650 0.740 0.716 0.785 0.865 0.800 0.865 0.938 0.872 0.922 0.998 Th in blanket salt 0.653 0.600 0.538 0.403 0.382 0.330 0.264 0.254 0.213 0.189 0.180 0.146 0.115 0.130 0.092 Neutron yield, 7 2.20 2.20 2.22 2.18 2,19 2,21 2,19 2.18 2,20 2.19 2.18 2.20 2.20 2.18 2.19 Regeneration ratio 0.996 1.026 1.055 0.984 1.032 1.070 0.980 1.039 1.078 0.989 1.045 1.084 0.987 1,052 1.090 *Neutrons absorbed per neutron absorbed in U233. | 4 Fuel: U233 Core diameter: 8 ft ThF4 in core: 7 mole % Power: 600 Mw (thermal) Lood factor: 0.8 External volume: 339 503 Chemical processing rate: 1.66 ft3/day for core 1.97 ¥ /day for blanket Table 1.1.3. Long-Term Nuclear Performance of o Two-Region, Homogeneous, Spherical, Molten-Fluoride-Salt Reactor Initial State After 1 Yeor Afrer 2 Years After 5 Years After 10 Yeors Aftrer 20 Yeors Inventory Absorption Inventory Absorption Inventory Absorption Inventory Absorption Inventory Absorption laventory Absorption (kg) Ratio (ka) Ratio (kq) Ratio (kg) Ratio (kg) Ratie (kg) Ratio Core elements Th232 14,527.1 0.8652 14,527 0.8416 14,527 0.8343 14,527 0.8237 14,527 0.8086 14,527 0.7892 Pq?33 0 0 18.8 0,00692 18.43 0,00671 18.38 0.00642 18.11 0.00605 17.67 0,00562 Y233 1,283.28 1,0000 1,363 0.9998 1,388.3 0.9994 1,420.7 0,996 1,453.03 0.9876 1,500.02 0.9721 Fissions 0.8730 0.8740 0,8741 0.8721 0.8648 0.8519 n,y 0.1270 0.1258 0.1253 0.1245 0.1228 0.1202 234 0 0 27.32 0,00348 53,91 0,00672 129.3 0.015 242.05 0,0305 428.7 0,0472 y3s 0 0 0.353 1.73 « 104 1.32 6.37x107* 717 0.00339 23.17 0.01235 63.07 0.0279 Fissions 0 123« 10~* 4,53 x 10~* 0.00241 0,00880 0.0199 n,y 0 0.50 » 104 1.84 » 10-4 0,00098 0.00355 0,0080 Y23 0 0 0.00347 6.39 < 10~7 0.0252 4.50 = 10~* 0.343 5.89%10"% 2.206 0,000458 11.73 0.00178 Np237 0 0 274105 9.76x 10" 3.23410~* 113x10"7 0.00725 2.46 x 10-% 0.58 2.40 = 10~3 0,321 9.94x10°3 TEAL 0 0 412107 101100 9.44,10~% 225.107 565x10°* 1,30~ 1077 0,00922 2.863 x 10~¢ 0.104 2.711x10"% py23? 0 0 3.46 <107 1.98x10-'? 1.49%10°7 B37x10""" 2.29-10"% 1.25«10% 0.000732 5.952 = 10~7 0.0148 7,36 x 10~ Fissions 0 1.22% 1072 515« 10="" 0.77 » 108 3,687 » 10”7 4.58 x 10~¢ ",y 0 0.76 » 10~'2 3.22x 107" 0.48 » 10~® 2,255 < 10~7 2.78 x 10~8 Fission products 0 0 115,2 0,0165 158.2 0.0221 180,5 0,0244 181.5 0.0234 181.5 0.0221 Li 3,699.12 0.00522 36,991 0.00488 3,699 0,00478 3,699 0.00455 3,699.1 0.00449 3,699.1 0.,00429 Be 2,739.23 0.00907 2,739 0.00906 2,739 0.00906 2,739 0,00905 2,739.2 0.00903 2,739,2 0.00900 F 26,330.9 0.06433 26,331 0.0643 26,331 0.0642 26,331 0,0642 26,331 0.0640 26,331 0,0638 Blanket Th232 30,821.6 0,2132 30,822 0,2132 30,822 0.2131 30,822 0.2128 30,822 0.212 30,822 0.211 Pg23 0 4,796 4,79 4.78 4,771 4.752 y233 0 26.26 38.15 44.36 44.53 44,35 Other 0.0428 0.0428 0,0428 0.0428 0.0426 0,0423 Total inventory, kg of 1,283.28 1,389,6 1,427.8 1,472.2 1,520.73 1,607.5 fissionable isotopes Fission yield, 5 2.20 2.203 2.204 2.204 2,202 2.196 Fuel feed rate, ka/year 118 27.4 6.0 -0.1 -0.8 Net integrated fuel 0 11.46 2.5 -6.42 -75.87 -167.36 bumup, kg Regeneration ratio 1.079 1.051 1,047 1.046 1.045 1.042 LY0dI Y SS3IYO0AJ 4OLOVIY LTVS"NILTIOW UNCLASSIFIED ORNL-LR-DWG 34505 140 7 mole%. Th, 8 ’/T 10 4 mole% ThF, 4 4 / / 20412 4 | 2T 2 mole% ThE, ° | REGENERATION RATIO ?; ! " FUEL SALT: 37 mole%s BeF, + 63 mole7s LiF + , UF, + ThE, EXTERNAL FUELVOLUME: 339 fTS | ! | e ; 0.90 g §'mole ThE, IN FUEL SALT ' ‘ | NUMBERS ON CURVES ARE I 8 ‘ CORE DIAMETERS IN FEET I L H 1 1 2 ~ 4 6 8 10 12 44 16 18 20 (X109 TOTAL INVENTORY OF FISSIONABLE ISOTOPES {kg) 0 Fig. 1.1.2. Initial Regeneration Ratio in Two-Region, Homogeneous, Molten-Fluoride-Salt Reactors Fueled with U233. 20 years. The regeneration ratio, total inventory (including U233 in the blanket), annual feed rate, and net integrated burnup are plotted in Fig. 1.1.3. It may be seen that, although the regeneration ratio remains above 1.04 and therefore there is a net breeding gain, it is, nevertheless, necessary to add fresh U233 during the first five years to com- pensate for the ingrowth of fission products and nonfissionable isotopes. The annual additions amount to 118, 27.4, 10.8, 6.0, and 1.6 kg, re- spectively, but less than 1.0 kg per year is required after five years, and the reactor is then self-sustaining. By the end of the second year the net integrated burnup (total inventory less total fuel purchased) becomes negative. Blanket Fissions in the Reference Design Reactor The burnout code, Sorghum, used in this study does not take into account fissions occurring in blanket. In order to estimate these, a series of sequential Cornpone and Sorghum calculations was performed, in which the concentrations of the 16 elements in the reactor predicted by one-year Sorghum calculations were submitted periodically to a criticality test by the Cornpone program. The outputs, consisting of fraction of fissions in the PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED ORNL -LR-DWG 34506 1.08 o q < 106 = O = 1.04 Ll o 4 Z 102 (U] & £.00 1600 TOTAL INVENTORY [ - L INY I S w E _.——-'._—-—-——'_'-_-— 7= 1400 ! [%2] < 7 / % o =0 1200 0o e TOTAL POWER: 600 Mw ( HEAT) 1000 '— (0AD FACTOR: 0.8 EXTERNAL FUEL VOLUME: 339 1t3 CORE AND BLANKET PROCESSING RATE:ONCE PER YEAR 150 — CORE DIAMETER: 8 f1 [ O S O —~ 100 I T i 1 : g =1 AVERAGE FEED RATE, kg OF UZ33/year @ 50 a o 1 5 0 % T~ J -50 ~ m \ ; >\ © 100 |— NET INTEGRATED ~ . & BURNUP, kg — | S~ Y 450 ~ 200 0 2 4 6 8 {0 12 44 & 48 20 TIME OF OPERATION (years) Fig. 1.1.3. Nuclear Performance of Two-Region, Homogeneous, Molten-Fluoride-Salt Reactor Fueled with U233. blanket, together with space-averaged group fluxes and leakage probabilities, were then used as in- puts for further one-year Sorghum calculations. The neutron source in Sorghum was also reduced by the fraction of fissions occurring in the blanket. In addition to providing information about the fissions in the blanket, this series of calculations also provided a test of the usual method of operat- ing the code. The case selected for study was the Interim Design Reactor with a core 8 ft in diameter, a fuel salt consisting of 36.5 mole % BeF,, 62.5 mole % LiF, and 1 mole % ThF ,, to- gether with sufficient UF , to make the system critical (about 0.25 mole %). The case was run as Cornpone calculation 01045 and the resulting data were used as input for Sorghum calculation 01065. In the integration of the time-dependent difference equations, a time increment of 5 days MOLTEN-SALT REACTOR PROGRESS REPORT was used until the time variable corresponded to one year of operation of the reactor; at this time, the increment was increased to 30 days for the next 20 years of reactor time. This is customarily done to shorten the time required for the compu- tation and is justified by the fact that, after the first year, concentrations change siowly. A second series of calculations was performed as described above, except that the time incre- ment was held constant at 5 days. After one year of reactor operation, the concentrations predicted by the Sorghum code were found to render the reactor slightly subcritical (£ = 0.9982 on Corn- pone 02045), even though the U233 in the blanket was faken into account. The fraction of fissions in the blanket was estimated to be 0.013. This information, together with flux spectrum and leak- age probabilities, was used as input for further Sorghum calculations, etc., with the results shown in Fig. 1.1.4, where the fraction of blanket fissions is shown as a function of reactor operating time. Also shown is a comparison of two key parameters, [} g UNCLASSIFIED = ORNL—LR—DWG 34507 D o002 , : v W | l ! ! b= ! ‘ i O o ; | = ! S @ l — (51) Z L ! 08 [V o7 | | { | [ CORNPONE BASE, SECOND SERIES E [ ! ‘ [ /CORNPONE BASE, FIRST SERIES - R [ B [ { ey = =— 0CUSOL BASE | } | ! \ \ f 0.6 REGENERATION RATIO DIAMETER OF CORE: 8ft 0.5 1250 ThF, IN FUEL" {1 mole Do EXTERNAL FUEL VOLUME: 338 ft3 uLI)J TOTAL POWER: 600 Mw (THERMAL) % LOAD FACTOR: 0.8 > = ! O ! 55 ] _ Z L, 1000 l — & OCUSOL BASE _—F ; 2’ A== / —// 22 / S & / ] | A =0 CORNPONE BASE, FIRST SERIES S rso 1\l L bbb —t © ’ CORNPONE BASE, SECOND SERIES g ] | 500 ‘ Q 2 4 6 8 10 12 14 16 18 20 TIME {years) Fig. 1.1.4. Blanket Fissions in Interim Design Reactor. 10 critical inventory and regeneration ratio, com- puted in the two series of calculations. The cal- culations to date cover only the first five years of operation; however, the fraction of fissions in the blanket appears to have stabilized at about 0.02. A comparison between the two lower curves for critical inventory shows satisfactory agreement between the two methods of calculation, especially since the top points of the discontinuous curve represent the critical condition. Similarly, the low points of the discontinuous curve for regener- ation ratio agree well with the smooth curve. How- ever, it is not presently understood why the slopes of the discontinuous curves differ so much from the smooth curves, and it may be that the agree- ment of the points is fortuitous. Comparison of Ocusol and Cornpone Calculations As mentioned above, a mathematical defect in the Cornpone program was discovered, and hence all prior Cornpone calculations were discarded. After the error had been corrected, the program was used to recompute an Ocusol case. A com- parison of the results obtained from the two pro- grams with identical input conditions is shown in Table 1.1.4. it may be seen that, while there is good agreement between the regeneration ratios, there is a substantial difference between the critical inventories and that the Cornpone program predicts the lower values. The discrepancies have been resolved in favor of Cornpone, which embodies a fourth-order approx- imation to the difference equations, whereas Ocusol uses only second-order approximations that result in an under-estimate of the absorptions in the core. In order to achieve a neutron balance in Ocusol, the currents at the boundary of the core are adjusted arbitrarily and are not computed from the gradients of the flux. As a result, the leakages are overestimated. In Cornpone, the fourth-order solutions satisfy the neutron balances to less than one part in a thousand, and the boundary currents are correctly computed from the gradients of the flux. The difference in the behavior of this reactor as predicted by Sorghum calculations based on the Ocusol and on the Cornpone solutions are shown in Fig. 1.1.4. It may be seen that the differences in regeneration ratio are negligible, but the cal- culation based on Cornpone predicts a critical inventory about 50 kg lower than that based on Ocusol. PERIOD ENDING OCTOBER 31, 1958 Table 1.1,4, Comparison of Ocusol and Cornpone Calculations for the Interim Design Reactor Fuel: U235 Core diametar: 8 ft ThF4 in core: 1 mole % Power: 600 Mw (thermal) Load factor: 0.8 External fuel volume: 339 £ Ocusol Cornpone Inventory Absorption Inventory Absorption (kg) Ratio (kg) Ratio Core elements Th232 2,121 0.364 2,121 0.378 u233 616 565 Fissions 0.729 0.734 ny 0.271 0.266 y238 45.9 0.039 46.7 0.042 Li 3,969 0.033 3,969 0.037 Be 3,044 0.010 3,044 0.010 F 24,291 0.058 24,291 0.047 Blanket TH232 30,822 0.232 30,822 0.227 Other 0.063 0.071 Fission yield, 7 1.800 1.812 Regeneration ratio 0.635 0.647 Heterogeneous Reactors As mentioned in previous reports, graphite- moderated, heterogeneous, molten-fluoride-salt reactors show promise of achieving higher regener- ation ratios for a given investment in fuel than the homogeneous reactors. As a result, attention has been given to developing a reasonably reliable method of calculation. The multigroup Cornpone program for Oracle was modified by W. E. Kinney to provide a boundary condition of zero net cur- rent in each diffusion group. Thus it is now pos- sible to treat a heterogeneous unit cell in an infinite lattice and to compute group disadvantage factors which may then be applied to the group cross sections in a homogeneized version of the finite reactor. A program modifying the Cornpone program to apply these factors has been written, Gamma-Ray Heating Calculations Work on a program for computing the gamma-ray heating in the core vessel and blanket of homo- geneous, molten-salt reactors was resumed. Gamma-ray energy absorption coefficients as a function of energy for 24 elements were computed from the x-ray attenuation coefficients listed by Grodstein, 3 according to a procedure developed previously.® The defining equation is: Mo =tpe * [ppltpy + (T + [ duc o 3G. W. Grodstein, Natl. Bur. Standards (US) Circ. 583 (1957). 4L. G. Alexander, The Gamma Energy-Absorption Coefficient, ORNL CF-56-8-219 (August 8, 1956). 11 MOLTEN-SALT REACTOR PROGRESS REPORT where i, = gamma-ray energy absorption coefficient, Hpe = photoelectric absorption coefficient, .= Compton scattering coefficient, f. = fraction of energy of photon deposited locally in a pair-production collision, f_ = fraction of energy of photon deposited locally in a Compton collision, fo= fraction of energy of photon carried away from Compton collision by photons having energies less than E , the Heut-off”’ energy. In these calculations /., was taken to be equal to 1.0; that is, the annihilation radiation was in- cluded in the local heating effect. This is not strictly consistent with the use of a low-energy 5L. G. Alexander, The Integral Spectrum Method for Gamma-Heating Calculations in Nuclear Reactors, ORNL CF-56-11-82 (Nov. 4, 1956}. 12 cut-off limit, E ), of 0.1 Mev for the Compton process. The energy-absorption coefficients com- puted here will be used, however, in an ‘‘integral spectrum’’ calculation of the type described in ORNL CF-56-11-82 (ref 5) in which the coefficient is averaged over the photon spectrum which is not known with sufficient accuracy to warrant taking into account the annihilation radiation. For a cut-off energy of 0.1 Mev, /_ is equal to zero with negligible error for all energies down to and including 0.15 Mev. The energy-deposition coefficient for Compton scattering, f_, was taken from the values listed in ref 4, The calculated values are listed in Table 1.1.5, and they have been assembled on paper tape for use with the gamma-ray heating program Ghimsr being written for the Oracle. A closed, three- address subroutine has been written for the com- putation of Storm’s attenuation function for spheri- cally symmetric systems.® SM. L. Storm et al., Gamma-Ray Absorption Distribu- tions ..., KAPL-783 (July 24, 1952), PERIOD ENDING OCTOBER 31, 1958 Table 1.1.5. Gamma-Ray Energy-Absorption Coefficients, i, e . o - E., (Mev) barns /atom cm?/g E., (Mev) barns /atom cm?/g Hydrogen (1;2) Beryllium (4;9.02) 0.1 0.0620 0.0375 0.1 0.251 0.0167 0.2 0.0856 0.0514 0.2 0.342 0.0228 0.4 0.0982 0.0589 0.4 0.396 0.0264 0.6 0.0991 0.0595 0.6 0.396 0.0264 0.8 0.0970 0.0581 0.8 0.388 0.0258 1.0 0.0927 0.0556 1.0 0.372 0.0248 1.5 0.0850 0.0510 1.5 0.328 0.0219 2.0 0.0815 0.0589 2.0 0.328 0.0219 3.0 0.0669 0.0401 3.0 0.273 0.0182 4.0 0.0593 0.0356 4.0 0.247 0.0165 6.0 0.0491 0.02%4 6.0 0.213 0.0142 8.0 0.0425 0.0255 8.0 0.191 0.0128 10.0 0.0376 0.0226 10.0 0.177 0.0118 Carbon (6;12) Nitrogen (7;14) 0.1 0.391 0.0199 0.1 0.470 0.0202 0.2 0.512 0.0257 0.2 0.599 0.0258 0.4 0.589 0.0296 0.4 0.689 0.0296 0.6 0.594 0.0298 0.6 0.6%94 0.0298 0.8 0.582 0.0292 0.8 0.679 0.0292 1.0 0.557 0.0279 1.0 0.650 0.0280 1.5 0.512 0.0257 1.5 0.597 0.0257 2.0 0.493 0.0247 2.0 0.579 0.0249 3.0 0.416 0.0209 3.0 0.490 0.0211 4.0 0.381 0.0191 4.0 0.451 0.0194 6.0 0.334 0.0167 6.0 0.400 0.0172 8.0 0.306 0.0153 8.0 0.370 0.0159 10.0 0.390 0.0145 10.0 0.251 0.0151 13 MOLTEN-SALT REACTOR PROGRESS REPORT Table 1.1.5 (continued) 14 - . o . E., (Mev) barns /atom em?/g E., (Mev) barns/atom em?/g Oxygen (8;16) Sodium (11;23) 0.1 0.531 0.0200 0.1 1.04 0.0273 0.2 0.702 0.0264 0.2 1.031 0.0270 0.4 0.790 0.0297 0.4 1.119 0.0293 0.6 0.784 0.0294 0.6 1.098 0.0288 0.8 0.776 0.0292 0.8 1.067 0.8279 1.0 0.745 0.0280 1.0 1.021 0.0268 1.5 0.683 0.0257 1.5 0.940 0.0246 2.0 0.662 0.0249 2.0 0.895 0.0234 3.0 0.563 0.0212 3.0 0.791 0.0207 4.0 0.522 0.0196 4.0 0.745 0.0195 6.0 0.452 0.0170 6.0 0.658 0.0172 8.0 0.410 0.0154 8.0 0.608 0.0159 10.0 0.394 0.0148 10.0 0.599 0.0157 Magnesium (12;24.32) Aluminum (13;26.96) 0.1 1.280 0.0316 0.1 1.595 0.0356 0.2 1.085 0.0268 0.2 1.189 0.0265 0.4 1.185 0.0293 0.4 1.285 0.0287 0.6 1.188 0.0294 0.6 1.288 0.0287 0.8 1.162 0.0288 0.8 1.264 0.0282 1.0 IREN 0.0276 1.0 1.210 0.0270 1.5 1.025 0.0254 1.5 LN 0.0248 2.0 0.074 0.0248 2.0 1.088 0.0242 3.0 0.869 0.0215 3.0 1.031 0.0230 4.0 0.822 0.0203 4.0 0.901 0.0202 6.0 0.767 0.0190 6.0 0.848 0.0189 8.0 0.739 0.0182 8.0 0.822 0.0184 10.0 0.727 0.7180 10.0 0.714 0.0160 PERIOD ENDING OCTOBER 31, 1958 Table 1.1,5 (continued) = . e . E., (Mev) barns /atom em?/g E., (Mev) barns/atom em?/g Silicon (14;28.06) Phosphorus (15;31,02) 0.1 1.986 0.0426 0.1 2.49 0.0485 0.2 1.209 0.0259 0.2 1.45 0.0282 0.4 1.392 0.0300 0.4 1.49 0.0290 0.6 1.383 0.0297 0.6 1.48 0.0288 0.8 1.360 0.0292 0.8 1.45 0.0282 1.0 1.303 0.0280 1.0 1.39 0.0270 1.5 1.198 0.0257 1.5 1.28 0.0249 2.0 1.173 0.0252 2.0 1.26 0.0275 3.0 1.030 0.0221 3.0 1.11 0.0216 4.0 0.984 0.0211 4.0 1,068 0.0208 6.0 0.933 0.0200 6.0 1.017 0.0198 8.0 0.912 0.0196 8.0 (1.007) 10.0 0.906 0.0195 10.0 1.022 0.0195 Sulfur (16;32.06) Argon (18;39.91) 0.1 3.10 0.0593 0.1 4.7 0.0709 0.2 1.60 0.0300 0.2 1.95 0.0294 0.4 1.60 0.0300 0.4 1.82 0.0274 0.6 1.60 0.0300 0.6 1.80 0.0272 0.8 1.56 0.0296 0.8 1.75 0.0264 1.0 1.49 0.0280 1.0 1.67 0.0252 1.5 1.37 0.0257 1.5 1.55 0.0234 2.0 1.34 0.0252 2.0 1.52 0.0229 3.0 1.19 0.0224 3.0 1.36 0.0205 4.0 1.15 0.0211 4.0 1.32 0.0199 6.0 1.10 0.0206 6.0 1.25 0.0189 8.0 1.10 0.0206 8.0 1.29 0.0294 10.0 1.10 0.0206 10.0 1.30 0.0196 15 MOL TEN-SALT REACTOR PROGRESS REPORT Table 1.1.5 (continued) e 2 o 2 E., (Mev) barns/atom em?/g E., (Mev) barns /atom em?/g - ‘ Potassium (19;39.1) Calcium (20;40.07) . 0.1 5.8 0.0895 0.1 7.26 0.1092 0.2 2.14 0.0330 0.2 2.38 0.0358 0.4 1.94 0.0299 0.4 2.05 0.0309 0.6 1.90 0.0293 0.6 2.01 0.0325 0.8 1.85 0.0285 0.8 1.95 0.0294 1.0 1.76 0.0272 1.0 .85 0.0278 1.5 1.63 0.0251 1.5 1.72 0.0259 2.0 1.61 0.0248 2.0 1.69 0.0255 3.0 1.44 0.0222 3.0 1.52 0.0229 4.0 1.41 0.0218 4.0 1.50 0.0226 6.0 1.39 0.0214 6.0 1.49 0.0224 8.0 1.41 0.0218 8.0 1.50 0.0226 10.0 1.43 0.0220 10.0 1.53 0.0230 Iron (26;55.84) Copper (29;63.57) 0.1 20.7 0.223 0.1 32.5 0.308 . 0.2 4.45 0.0479 0.2 6.2 0.0287 0.4 2.84 0.0206 0.4 3.33 0.0316 0.6 2.67 0.0288 0.6 3.03 0.0287 0.8 2.57 0.0277 0.8 2.90 0.0275 1.0 2.44 0.0263 1.0 2.74 0.0260 1.5 2.24 0.0242 1.5 2.50 0.0237 2.0 2.23 0.0240 2.0 2.52 0.0239 - 3.0 2.07 0.0223 3.0 2.45 0.0232 4.0 2.08 0.0224 4.0 2.40 0.0228 . 6.0 2.15 0.0232 6.0 2.52 0.0239 8.0 2.22 0.0240 8.0 2.63 0.0250 10.0 2.40 0.0258 10.0 2.73 0.0259 16 PERIOD ENDING OCTOBER 31, 1958 Table 1.1.5 (continued) Photon Photon Energy fe Energy He E,, (Mev) barns/atom em?/g E., (Mev) barns /atom em?/g Molybdenum (42;96.0) Tin (50;118.7) 0.1 147 0.930 0.1 289 1.46 0.2 22.3 0.140 0.2 43.6 0.221 0.4 6.9 0.0433 0.4 10.5 0.0534 0.6 5.04 0.0316 0.6 6.90 0.0350 0.8 4.53 0.0284 0.8 5.86 0.0297 1.0 4.20 0.0264 1.0 5.29 0.0268 1.5 3.81 0.0239 1.5 4.71 0.0239 2.0 3.86 0.0242 2.0 4.77 0.0242 3.0 3.86 0.0242 3.0 4.79 0.0242 4.0 3.99 0.0250 4.0 5.12 0.0260 6.0 4.39 0.0275 6.0 5.77 0.0292 8.0 4.71 0.0296 8.0 6.27 0.0318 10.0 5.03 0.0315 10.0 6.74 0.0341 lodine (53;126.93) Tungsten (74;184.0) 0.1 363 1.72 0.1 1254 4N 0.2 54 0.266 0.2 192 0.630 0.4 12.4 0.0589 0.4 37.0 0.121 0.6 7.7 0.0365 0.6 18.8 0.0616 0.8 6.4 0.0303 0.8 13.1 0.0430 1.0 5.75 0.0272 1.0 1.2 0.0368 1.5 5.08 0.0240 1.5 9.5 0.0311 2.0 5.15 0.0244 2.0 8.5 0.0278 3.0 5.21 0.0247 3.0 8.75 0.0286 4.0 5.60 0.0265 4.0 9.51 0.0312 6.0 6.31 0.0299 6.0 10.86 0.0356 8.0 6.90 0.0327 8.0 11.82 0.0388 10.0 7.44 0.0352 10.0 13.0 0.0426 17 MOLTEN-SALT REACTOR PROGRESS REPORT Table 1,1,5 (continued) Photon Photon Energy Fe Energy Fe E., (Mev) barns/atom em?/g E., (Mev) barns/atom em?/g Platinum (78;195.2) Thallium (81;204.4) 0.1 1505 4.65 0.1 1715 5.04 0.2 232 0.716 0.2 268 0.789 0.4 44.8 0.139 0.4 51.5 0.152 0.6 21.6 0.0667 0.6 24.4 0.0719 0.8 15.2 0.0470 0.8 16.9 0.0497 1.0 12.1 0.0374 1.0 13.3 0.0392 1.5 9.5 0.0294 1.5 10.2 0.0301 2.0 9.3 0.0288 2.0 10.2 0.0301 3.0 9.6 0.0297 3.0 10.3 0.0304 4.0 10.4 0.0321 4.0 1.0 0.0324 6.0 10.8 0.0333 6.0 12.6 0.0371 8.0 12.9 0.0398 8.0 13.8 0.0406 10.0 14.2 0.0439 10.0 15.1 0.0445 Lead (82;207.20) Uranium (92;238) 0.1 1785 5.18 0.1 3801 0.962 0.2 282 0.819 0.2 533 1.35 0.4 53.7 0.156 0.4 82.2 0.208 0.6 25.4 0.0738 0.6 38.3 0.0970 0.8 17.4 0.0506 0.8 24.9 0.0630 1.0 13.8 0.0401 1.0 19.1 0.0483 1.5 10.5 0.0304 1.5 13.7 0.0346 2.0 10.4 0.0302 2.0 13.1 0.0331 3.0 10.4 0.0302 3.0 13.1 0.0331 4.0 1.4 0.0332 4.0 14.0 0.0354 6.0 12.9 0.0375 6.0 16.6 0.0419 8.0 13.2 0.0384 8.0 17.1 0.0432 10.0 15.5 0.0451 10.0 18.8 0.0475 18 PERIOD ENDING OCTOBER 31, 1958 1.2, COMPONENT DEVELOPMENT AND TESTING H. W. Savage W. B. McDonald Reactor Projects Division FUEL PUMP DEVELOPMENT W. F. Boudreau A. G. Grindell Development Tests of Salt-Lubricated Bearings P. G. Smith W. E. Thomas H. E. Gilkey Hydrodynamic Bearings. — The results of the first attempt ! to establish a molten salt (LiF-BeF ,-UF , 62-37-1 mole %) hydrodynamic bearing film have been assessed and three additional tests with the same combination of materials, an INCR-8 bearing and an INOR-8 journal, have been made. A review of the bearing installation procedure used for the first test revealed that the bearing had been placed into the test apparatus upside down. This error accounts for the difficulty experienced in maintaining a hydrodynamic film during bearing operation. Testing of a second bearing and journal of INCOR- 8 was terminated on schedule at the end of a 500-hr period of satisfactory operation. Examination of this bearing and journal revealed a very small amount of wear and a nearly unchanged surface finish. Every indication points to the establishment and maintenance of a hydrodynamic film during operation of this bearing and journal. Fifty-seven stops and starts were made during a 54-hr period of the test. The remainder of the test was con- ducted at constant conditions, that is, a shaft speed of 1200 rpm and an applied radial load of 200 lb. A report? covering the first and second tests was written. The same bearing and journal were then assembled for a third test to investigate the effect of repeated operation. The third test covered a period of 284 hr during which 37 stop-start tests were performed. The last 100 hr of operation were at constant con- ditions: shaft speed, 1200 rpm; applied radial load, 300 |b. Examination of the bearing and journal re- vealed slight rubbing marks at each end of the bearing. P, G. Smith, W. E. Thomas, and L. V. Wilson, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 20. 2p, G. Smith, Salt-Lubricated Hydrodynamic Journal Bearing Tests Nos. 1 and 2, ORNL CF-58-8-10 (Aug. 7, 1958). A fourth test with a new (third) bearing and journal of INOR-8 in molten salt at 1200°F is under way to trace the cause of rubbing marks to either stop and start tests or to steady-load operation. [t is planned to continue the testing of the second bearing and journal at a later date. Hydrostatic Bearings. — Fabrication of the hydro- static bearing tester® was completed and tests were run on two hydrostatic bearings with water as the lubricant. During the first test, in which a radial clearance of 0.003 in. was used, the bearing and journal experienced considerable wear. It was found that the supply pressures for the four hydrostatic pockets differed from each other, instead of being constant. The radial distribution of these pressures was very similar to the radial distribution of the volute pressures, which give rise to the load on the bearing. As a consequence, the bearing force tended to be least in the direction of the greatest bearing load. This difficulty was traced to four partitions which had been instalied in the manifold supplying the hydrostatic pockets. The partitions were re- moved for the second test, which was conducted with a second hydrostatic bearing having a radial clearance of 0.0075 in. For this test, additional instrumentation was installed to permit indication of the eccentricity of the journal. This instru- mentation consists of nozzles through which air flows radially inward toward the cylindrical shaft - surface; the back pressure on the nozzle is a function of the distance between the nozzle outlet and the cylindrical surface. Further refinements in this technique of measurement appear to be de- sirable. In the second test, data were obtained at shaft speeds from 600 to 3000 rpm at various pump cir- cuit resistances in an attempt to determine tne effect of varying the bearing load. Data on bear- ing loads and flows were obtained by measuring the pressure in each bearing pocket and the pressure differential across each orifice. The second bear- ing and journal experienced slight wear of the type that might be expected to occur during stopping and 3L. Y. Wilson, MSR Quar. Prog. Rep. June 30, 1958, ORNL.-2551, p 19. 19 MOLTEN-SALT REACTOR PROGRESS REPORT starting. At low speeds, during stopping and start- ing, no supply pressure is available to the hydro- static pockets; thus contact between the journal and bearing may be expected. Rotating-Pocket Hydrostatic Bearings. — Static tests were completed on the rotating-pocket hydro- static bearing,4 and dynamic tests are in progress. In the initial dynamic test it was found that there was entrainment of air in the process water. This problem has been solved by the use of an external circuit that carries flow from the upper side of the impe ller-bearing region to the impeller suction through a stilling well. Another problem that is being investigated is that of the pressure dis- tribution in the pockets. The pressure distribution obtained thus far in dynamic tests does not con- form to that found in the static tests and pre- dicted by the analytical studies. Bearing Mountings. -~ Two types of bearing mounts are being studied. One of these is a column type of sleeve mount designed at ORNL that takes into account the thermal expansion differences of the materials. Plans have been made to test this mount statically prior to applying it to an actual bearing in a molten-salt system. Fabrication of a mount for testing awaits a satisfactory solution to the problem of brazing molybdenum to Inconel and to INOR-8. The other type of mount being studied is a patented fixture supplied by Westinghouse. Tests of this fixture, designated ““Thermoflex,'’ are under way. Development Tests of Conventional Bearings D. L. Gray W. E. Thomas Organic-Liquid-L ubricated Bearings. — The high- temperature sump pump being used to conduct a demonstration of the behavior of Dowtherm **A’’ (eutectic mixture of diphenyl and diphenyloxide) as a bearing lubricant operated satisfactorily for a period of 3206 hr. The pump circulated fuel 30 (Na F-ZrF4-UF4, 50-46-4 mole %) at a temperature of 1200°F; the shaft speed was about 2600 rpm; the temperature of the Dowtherm ‘‘A’’ supplied to the bearings was maintained at 180°F, with approxi- mately a 5°F rise through the test bearings. The lubricant flow rate was 2 gpm. Upon completion of 3200 hr of operation, the pump was disassembled, and the test bearings and seals were inspected. The double-row angular- . v, Wilson, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 17. 20 contact ball bearings (MRC-5309), which were sub- merged in the lubricant during operation, showed no detectable wear or imperfections. The diametral clearance between the aluminum bearing and the Inconel journal was found to be essentially the same (0.0044 in.) as it was at the time of the previous 100-hr inspection. Seal nose heights were measured at the beginning and the end of a 2008-hr test period, and wear indications of 0.0016 in. on the upper seal and 0.0004 in. on the lower seal were noted. Lower seal leakage averaged 2 cm®/day during the operating period, but the upper seal leakage was essentially zero. The pump was reassembled and reinstalled in the test stand and operation at the above-mentioned steady-state conditions was resumed. After 6 hr of operation, that is, an over-all total of 3206 hr of operation, a building power interruption caused the lube oil pump to stop. Within about 30 sec, it was noticed that the drive equipment was heavily over- loaded and that the pump speed had been reduced to zero. Upon disassembly of the pump it was found that the pump shaft had seized in the aluminum bearing. It was further observed that the three oil- supply grooves in the bearing were not covered by the journal for approximately T/4 in. Thus a path existed during the entire test for direct escape of the pressure-fed lubricant to the region external to the load-carrying film of the bearing. However, it is believed that this bearing would have continued to operate for a longer period of time had the lube oil pump remained in operation. The demonstration was terminated after this incident and a final report is being written, Bearing and Seal Gamma Irradiation. — The centrifugal-pump rotary element assembly being operated without an impeller in a gamma-irradiation facility at the MTR had accumulated a total of 5312 hr of operation by the end of the quarter; during 4426 hr of the operating period, the assembly has been exposed to gamma irradiation. The accumu- lated gamma-ray dose to the lower seal region has reached approximately 9.3 x 10% r. The bulk oil has accumulated an approximate gamma-ray dose of 2.25 x 108 r. On June 12, 1958, a momentary power failure stopped shaft rotation and caused seal leakage of 53 ¢m? in the following 24 hr. During the succeed- ing 24 hr the leakage was 12 cm?, after which the leakage leveled off to the normal 6 to 7 cm®/day. On June 23, 1958, an attempt was made to ac- celerate the bulk oil radiation damage by reducing the lube oil dilution factor. A total of 5430 cm® of oil had been removed from the lube oil reservoir when the low lube oil flow alarm halted rotation of the shaft. Sufficient oil was then replaced to permit stable operation, but subsequently operation was stopped twice because of a low lube oil level. The leakage of the lower seal increased during this quarter from approximately 6 ml/day to approxi- mately 60 ml/day. It is believed that the stoppages may have contributed to the increased rate of leak- age. [t is planned to terminate the test when the lower seal region has received a dose of 10'% r. This should be achieved within three additional weeks, The test unit will then be returned to ORNL for examination. Table 1.2.1 indicates the effect of the radiation on samples of the lubricant, Gulfspin 34, which were taken at various times and tested by the ORNL Analytical Chemistry Division. The viscosity of both the bulk and the seal- leakage oil has increased, but is not yet deleterious to the operation of the test, The increase in the bromine number indicates that the gamma irradiation has displaced some hydrogen atoms from the hydro- carbon molecules of the lubricant. The acidity PERIOD ENDING OCTOBER 31, 1958 number indicates that the lubricant system has suffered very little oxidation during the test. Mechanical Seals D. L. Gray Labyrinth and Split-Purge Arrangement. — Load- deflection data were obtained for new seal springs and for the seal springs used in the NaK pump that was operated to test a labyrinth and split-purge arrangement, as described previously.® The data are being analyzed to determine the effect of the test conditions on the seal-face loading created by the springs. The load-deflection tests were made both with and without the elastomeric O-ring which seals between the stationary and floating members of the seal in order to determine the effect of the resistance offered by the O-ring to the axial move- ment of the flexibly mounted seal ring element of the seal assembly. The results of these tests and the operation of the pump are being correlated. Bellows-Mounted Seal. —~ The modified Fulton- Sylphon bellows-mounted seal being subjected to an endurance test in a NaK pump has accumulated a D, L. Gray, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 21. Table 1,2,1, Effect of Gamma Irradiation on Lubricating Properties of Gulfcrest 34 in a CentrifugalsPump Rotary Element Assembly Bulk Circulating Qil Accumulated Oil Leakage from Lower Seal Control Sample Taken 3-13-58 5.24-58 7-29-58 Sample 3.77.58 6-5-58 7-31-58 to 4-28-58 to to 3-27-58 5.21-58 8.8-58 Approximate gamma- 0 0.47 x10% 1.3x10% 1.82x108 ray dose (r) Viscosity (SUS at 90.2 89.4* 104.9* 118.5 108.0 112.6 140.2 168.2 25°C) Bromine No. {mg of Br 0.87 2.8 4.2 5.6 3.4 4,1 7.3 6.8 per 100 g of oil) Acidity (mg of KOH 0 0 0 0.013 0 0 0 * per g of oil) *These quantities were checked within 5.5% by an independent analysis performed by the Special Samples Laboratory of the Y-12 Technical Division. **Result not yet available. 21 MOLTEN-SALT REACTOR PROGRESS REPORT total of 8110 hr of operation. The pump has been operating at a temperature of 1200°F and a speed of 2500 rpm. The observed seal leakage has continued to decrease to the point that it has become difficult to measure and can be considered to be negligible. Two stoppages occurred: (1) a power failure caused the pump to be stopped for 5 min, and (2) the pump was stopped to replace the motor brushes. The test has been scheduled for termination at the end of 8760 hr, at which time the seal will be examined. Radiation-Resistant Electric Motors for Use at High Temperatures S. M. DeCamp The investigation of the materials that would be required for radiation-resistant motors for use at high temperatures has included studies of electrical steels and high-strength low-resistance electrical conductors, in addition to the studies of electrical insulation previously reported.® This work is directed toward the development of a motor suitable for long-term operation at 1250°F in a radiation field. Six coil assemblies that incorporate the electrical insulation system developed by the Louis-Allis Company for use at high temperatures were re- ceived early in September. These coil assemblies (called *‘motorettes’’) are arranged on a steel frame in such a fashion that the three separate aspects of the insulation problem normally encountered in an electric motor are properly simulated and can be separately measured. These aspects of the problem ' eoil-to- are usually referred to as *‘turn-to-turn, coil,’’ and “‘coil-to-ground’’ resistance. During the course of development work by the Louis-Allis Company it was found that a new wire product (Silotex-N) made by Anaconda Copper Company was superior to ceramic-coated wire at elevated temperatures. Silotex-N wire is a nickel- plated copper conductor covered with a glass serv- ing. The wire is also coated with a silicone varnish that is stable up to about 350°C. At this temperature the silicone is driven off, and SiO, is left as a residue. The glass serving appears to be a good turn-to-turn insulation at high temperatures based on tests run by the Louis-Allis Company. No tests have been run by ORNL to date. The coil-to-coil and coil-to-ground insulation supplied with the motorettes consists of mica that 65, M. DeCamp, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 22. 22 was coated and impregnated with a modified ceramic cement by the High Temperature Instrument Com- pany. Again, preliminary tests by the Louis-Allis Company at 1200°F indicate good results. After the coils are wound, they are placed in the “‘slots”’ of the motorette with the mica as a slot liner {(coil- to-ground insulation) and as a coil separator (coil- to-coil insulation). The whole assembly is then impregnated with a ceramic compounded by the Louis-Allis Company. All materials used in the motorettes appear to offer good insulating properties at a tamperature of 1200°F. Radiation resistance has not been checked by test, as yet, but this re- quirement was considered in the selection of ma- terials. Testing of the motorettes by ORNL will include assessing the effects of time at elevated temperatures, thermal cycling operation during nu- clear irradiation, and operation during nuclear irradi- ation at elevated temperatures. A preliminary investigation of magnetic core ma- terials for use at high temperatures indicates that little progress has been made in this field. More work on obtaining satisfactory magnetic steel is planned. Operation of an electric motor at 1250°F will require electrical conductors with good mechanical strength properties and low electrical resistivity. Work is presently under way to determine types of materials which might be considered for such an application. Design Studies of Fuel Pumps Engineering study contracts for the evaluation of various types of molten-fluoride-salt pumps and for the preparation of preliminary design studies were awarded to two companies late in the previous quarter, and work on both contracts (Allis-Chalmers Manufacturing Company and Westinghouse Electric Company) was completed during this quarter, The work by Allis-Chalmers tock the form of a feasibility study. Nineteen different types of pump and drive combinations were evaluated, including centrifugal pumps of the canned-motor type, similar pumps with remote drives, positive displacement pumps, jet pumps, and electromagnetic pumps. The study resulted in a recommendation that two types of centrifugal pumps be selected for further study: (1) a pump with a vertical shaft and a free gas-to- salt interface and a totally enclosed fan-cooled motor drive; and (2) a turbo-pump energized by a secondary fluid, possibly a barren-salt mixture that is compatible with the reactor salt and the container material. Three preliminary centrifugal pump layouts were selected by Westinghouse for inclusion in their report (a number of other, less applicable, layouts were shown to ORNL engineers during the course of the work). Two of the layouts included in the re- port showed pumps of the sump type (having a free liquid-to-gas interface in the body of the pump). Both of these layouts made use of a salt-lubricated lower bearing, while the upper bearing was lubri- cated by a low-melting-point salt in a container physically separated from the process salt. Both pumps made use of totally enclosed motors. The third Westinghouse layout showed a sub- merged pump completely filled with salt that made use of a motor designed for operation at high temperatures. Both the journal and the thrust bearings were lubricated by the molten salt. The motor windings were to be made with a specially formed mica insulation system being developed by Westinghouse for another reactor project. A sub- merged pump of this type would have advantages in a molten-salt reactor in that (1) the heating of the pump parts by fission gases would be eliminated and (2) the pump could be made considerably smaller than a sump type of pump. One of the primary purposes of these design studies was to determine which particular elements of a fuel pump should be selected for inclusion in the development program. In this respect, both vendors emphasized the need for salt-lubricated bearings. Frozen-Lead Pump Secl W. B. McDonald E. Storto J. L. Crowley A small, submerged, centrifugal pump with a frozen-lead seal was designed and fabricated and is being operated in evaluation tests. The pump volute and impeller were patterned after a standard Eastern Industries, Inc., Model D-11 oil pump and were fabricated of Inconel. The pump barrel, which is the container for the lead, contains the seal gland. The gland incorporates a long tapered PERIOD ENDING OCTOBER 31, 1958 annulus which narrows to a short annulus of con- stant diameter around which the cooling coil is wrapped (see Fig. 1,2,1). The pump shaft is a directly coupled extension of the motor shaft and depends upon the motor for support and guidance. There is a 0.015-in. diam- etral clearance between the pump shaft and the pump body at both the seal gland and the volute. The drive motor is a fractional horsepower (]/8) in- duction motor rated to draw 2.0 amp under a full load at 3460 rpm. During operation with the frozen-lead seal this motor utilizes 1.5 amp at 3580 rpm. The test loop in which the pump is operating con- sists of 3/3-in. sched-40 Inconel pipe with a small surge tank at the top. Heat is applied by Calred heaters controlled by Variacs, and temperatures are measured and recorded by thermocouples and a recorder. The pump, motor, and loop assembly is shown in Fig. 1.2.2 without the heaters and thermo- couples. In preparation for filling, the loop piping was heated to 1000°F. With the pump shaft rotating, 1100 g of pure lead was loaded into the pump barrel; the chamber was about two-thirds full. The remainder of the loop and the pump barrel were then filled with fused salt mixture 30 (NaF-ZrFA-UF“) and flow was started. |sothermal operation was established at 1200°F for the main loop, and the temperatures of the pump barrel were regulated to establish a lead-to-fused salt interface temperature of 1200°F and a seal-gland temperature of 820°F, Since the start of operation on June 13, 1958, the pump and seal have operated continuously for 2600 hr. During the first 100 hr of operation, a slight leakage of lead was observed. This was stopped by adjusting the coolant flow, and no leakage has occurred since. A closeup view of the seal gland while in operation is shown in Fig. 1.2.3. In order to further evaluate the possibility of a lead-sealed pump, a larger model is being pre- pared for testing. A 3]{‘-in.-d ia shaft, which will operate at 1200 rpm, will be used in a seal gland of similar configuration. The peripheral speed of the shaft will be 1020 fpm compared with 190 fpm for the smaller model. 23 MOLTEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED ORNL— LR—DWG 34508 / PUMPHEAD PUMP VOLUTE IMPELLER TANK BARREL N PUMP SHAFT 7 | | | 7 N / ZPUMP FOOT TANK FILL LINE 9 COOLING COIL LA U ! 2 1 I ; 1 r— O INCHES Fig. 1.2,1, Sectional View of Lead-Sealed Pump. 24 UNCLASSIFIED PHOTO 31626 Fig. 1.2.2, Frozen-Lead Sealed Pump and Test Loop Without Heaters and Thermocouples. PERIOD ENDING OCTOBER 31, 1958 Fig. 1.2.3. Frozen-Lead Seal Gland During Operation. DEVELOPMENT OF TECHNIQUES FOR REMOTE MAINTENANCE OF THE REACTOR SYSTEM W. B. McDonald Mechanical Joint Development A. S. Olson The necessity for the development of a reliable mechanical joint was explained previously,” and detailed descriptions of the three types of joints under consideration were presented. Screening tests of the three joints were conducted under thermal-cycling conditions in a high-temperature loop that circulated fuel 30 (NcF-Zer-UFA) at temperatures up to 1500°F. 87 During this quarter, the freeze-flange and in- dented-seal-flange joints were tested with sodium, since sodium may be used as the primary coolant for a molten-salt power reactor. The two joints are shown installed in the loop in Fig. 1.2.4. The testing procedure consisted of the following: (1) assemble joint and leak check (cold) on helium leak detector; leakage specification, =1 x 107 cm? of helium per second; (2) weld joint into loop and operate through 50 thermal cycles between 1100 and 1300°F; (3) drain loop and remove joint intact; A S, Olson, MSR Quar. Prog. Rep. Jan., 31, 1958, ORNL-2474, p 20. 8A. . Olson, MSR Quar. Prog, Rep. June 30, 1958, ORNL-2551, p 24. 9W. B. McDonald, E. Storto, A. S. Olson, Screening Tests of Mechanical Pipe Joints for a Fused Salt Re- actor System, ORNL CF-5 -8-33 (Aug. 13, 1958). 25 MOLTEN-SALT REACTOR PROGRESS REPORT CL AJQ 'EO flOTO 31627 Fig. 1.2.4. Freeze-Flange and Indented Seal-Flange Joints Installed in Loop for Testing with Molten Sodium. (4) leak check joint on leak detector; and (5) separate joint, remake joint, and leak check. The leakage rate for the freeze flange joint, which cgntained a seal ring made of annealed copper, was 2 x 10=8 cm3/sec before the test and 1 x 107 ecm3/sec after the test. The flangese were then separated and reassembled with a new seal ring of the same material, and the leakage rate dropped to 1.3 x 10~8 cm?/sec. The leakage rate for the indented-seal-flange joint, which contained a seal ring made of nickel- plated Armco iron, was 5 x 10~8 cm3/sec before the test and 3 x 10=8 cm®/sec after the test. The flanges were then separated and reassembled with a new seal ring of the same material, and the [eak- age rate dropped to 1 x 10~8 cm3/sec. Temperatures were measured during the test at the points indicated in Fig. 1.2.5. The molten 26 THERMOCQUPLE 3| ‘ EY Ll 1 L N T —4 = l [ SODIUM “u | FLOW 32 I L 33| | . J! - ] [ C3—fe — 4 FREEZE-FLANGE JOINT TESTNO.5 FLOW | o8 SODIUM ) ‘ [ / UNCLASSIFIED ORNL~LR-DWG 34509 THERMOCOUPLE 21 > Ry — Fe { ;J(\ ‘ 231 »] \ | ks /J \\ / | 24 o ( e A VK - - o~ ° ( | &% NDENTED-SEAL-FLANGE JOINT TEST NO. 3 Fig. 1.2.5. Diagram of Locations of Thermocouples on Mechanical Joints Tested in a Loop That Circulated Sodium. sodium was circulated at a flow rate of 2 gpm, and the average temperature cycle time for the 50-cycle test was 60 min. Representative temperatures measured during the test are listed in Table 1.2.2. Both joints operated successfully; there was no indication of sodium leakage. The separated joints are shown in Figs. 1.2.6 and 1.2.7 as they appeared after the test. The annealed copper seal ring used in the freeze- flange joint was identical to that used in the pre- vious tests with molten salt. The gasket used in the indented-seal-flange joint was made of annealed Armco iron that was nickel plated to prevent oxi- dation of the iron. This gasket material was also successfully tested with molten salt. A detailed description of the large freeze-flange foint for use in a 4-in.-dia line was presented previously.® Tests of two of these joints were con- ducted in a high-temperature loop that circulated PERIOD ENDING OCTOBER 31, 1958 fused-salt fuel 30 at temperatures up to 1300°F under thermal-cycling conditions. The flanges are shown installed in the loop, prior to testing, in Fig. 1.2.8, and the test loop is shown schematically in Fig. 1.2.9. A heater-insulation unit, designed for use on a 4-in.-dia line, was installed on the loop and tested along with the large freeze-flange joints. The heater-insulation unit is shown in Fig. 1.2.10. A spacer is shown in Fig. 1.2,11, Two of the heater- insulation units and one spacer, located between the units, are shown installed on the test loop in Fig. 1.2.8. It was found that the thermal loss from the two units was about four times the heat loss from an equivalent length of ‘‘Hy-temp'’ pipe in- sulation 3 in. thick on a 4-in. pipe. The cold leakage rates for the two large freeze- flange joints, before installation in the loop, were 6.7 x 10~ cm? of helium per second for flange No. Table 1,2,2, Temperatures Measured During Thermal Cycling of Flanged Joints in a Sodium-Filled Test Loop Thermocouple Minimum Cycle Maximum Cycle Nomber* Temperature Normber Temperature Nomber (°F) (°F) Freeze-Flange Joint 27 110 10-20-28-29 133 9 28 170 21 215 9-41 29 905 29-31-32-34 1095 1-2 30 1075 31-32 1295 8 31 1045 31-32-39 1255 Numerous 32 890 38 1085 2 33 170 20 212 8-9 34 108 10 130 8-9 Indented-Seal-Flange Joint 21 908 37 1025 35 22 920 4 1035 18-28-41 23 925 3.4 1050 35 24 965 Numerous 1110 35 25 1060 32-38 1055 Numerous 26 1078 21-33 1288 8-41 *See Fig. 1.2.5 for location of thermocouples. 27 MOLTEN-SALT REACTOR PROGRESS REPORT Fig. 1.2.6. Indented-Seal-Flange Joint After Removal from Molten Sodium Test Loop. Fig. 1.2.7. Freeze-Flange Joint After Removal from Molten Sodium Test Loop. 28 UNCLASSIFIED PHOTO 319R4 UNCLASSIFIED PHOTO 31985 PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED PHOTO 31950 Fig. 1.2.8. Large Freeze-Flange Joints and Heater-Insulation Units Installed in Test Loop. 1 and 1.6 x 10~8 cm?/sec for flange No. 2. During the loop test, temperatures were measured at the points indicated in Fig. 1.2.12. Molten salt was circulated at a flow rate of ~40 gpm. The salt temperature was cycled 50 times between 1100 and 1300°F; each cycle was of 2 hr duration. Repre- sentative temperatures measured during the test are listed in Table 1.2.3. There was no indication of salt leakage during the test, but gas leakage tests made on the flanges while they were installed in the loop indicated leakage rates in excess of the maximum allowable leakage rate of 1077 cm® of helium per second. The bolts on the special clamps around each joint were tightened from original loads of 180 ft-Ib each to approximately 400 ft-lb each without appreciable improvement of the leakage rates. The clamps were therefore removed from the flanges, and additional bracing was welded to them to permit higher loads. In addition, oxide was re- moved from the clamps, and they were given an oxidation-resistant coating. A special lubricant containing graphite and molybdenum disulfide in a resin base was also coated over the surfaces of the clamp and bolts to reduce friction between mating surfaces and thus provide a clamp of greater efficiency. Two seal rings are being fabricated from 2S aluminum. [t is believed that the use of these seal rings and a higher bolt loading will provide a gas-tight joint. The thermal-cycle testing will be repeated following these modifications. Both of the joints were separated, after removal of the clamps, and the inside surfaces of the flanges were examined for signs of cracking caused 29 MOLTEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED ORNL-LR~DWG 31380R , PRESSURE-MEASURING DEVICES g ! Ya=in PIPE poV Lo 4-in PIPE FREEZE FLANGES FILL LINE AND FREEZ\E VALVE ) Fig. 1.2.9. Freeze-Flange Mechanical Joint Development Test Loop. 30 PERIOD ENDING OCTOBER 31, UNCLASSIFIED PHOTO 31558 Fig. 1.2.10, Heater-Insulation Shown in Open Position. UNCLASSIFIED PHOTO 31557 Fig. 1.2.11. Spacer for Heater-Insulation Unit. 1958 31 MOL TEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED ORNL--LR—DWG 3451 () FLANGE NO.1 THERMOCQUPLES | ] FLANGE NO.2 THERMOCOUPLES DOUBLE LINES INDICATE WELLS Fig. 1.2,12, Diagram of Location of Thermocouples on Joints Tested in Mechanical Joint Development Test Loop. by thermal stress. No such stress cracks were ob- served in dye checks. The joints were easily disassembled in about 20 min. The frozen salt in each joint was firmly attached to the stainless steel screen that had been placed there for that purpose, as shown in Fig. 1.2,13. A flange is shown with the screen and salt seal removed in Fig. 1.2.14. Note that the flange face is sufficiently clean for reassembly of the joint. Evaluation of Expansion Joints for Molten-Salt Reactor Systems J. C. Amos Three commercially available expansion joints have been tested in the test facility shown in Fig. 1.2,15. The joints were cycled over their maximum allowable traverse at rated conditions of 1300°F and 75 psig in an environment of molten salt or Table 1.2.3. Temperatures Measured During Thermal Cycling of Freeze-Flange Joints in Mechanical Joint Development Test Loop Thermocouple Minimum Cycle Maximum Cyele . Temperature Nomber Temperature Number (°F) (°F) Flange No. 1 ] 440 17 550 3 2 827 17 1030 6-10 3 962 17-18 1140 10 4 965 17-18-29 1135 10 5 875 10-17-29 1030 10 6 353 5 435 30 7 265 17 305 3 8 247 17 288 11-32 Flange No, 2 9 458 28 598 32 10 882 28-29 1045 10 11 970 17-18 1140 Numerous 12 932 18 1098 26 13 898 18 1063 8 14 413 17 510 2-3-31-49 15 242 17-18-41 307 49 16 330 10-18 378 49 *See Fig. 1.2.12 for location of thermocouples. 32 PERIOD ENDING OCTOBER 31, 1958 ‘ UNCLASSIFIED PHOTO 32544 UNCLASSIFIED 32545 ’ S Fig. 1.2.14, Disassembled Freeze-Flange Joint After Removal of Frozen Salt Seal and Screen. 33 MOL TEN-SALT REACTOR PROGRESS REPORT Qo ~ oY '}. ; —.' PNEUMATIC OPERATOR i . L N\ _ VAPOR TRAP FURNACE Fig. 1.2.15. Expansion Joint Test Facility. sodium. The joints and the results of the tests are described in Table 1.2.4. The failure of the Zallea joints cannot be con- strued as a reflection of the integrity of the joints, since the bellows were evidently weakened by the attachment of thermocouples. The Cook Electric Company joint will be examined metallurgically in an effort to determine whether the failure was due to a weld defect or to excessive stresses at rated operating conditions. Three more expansion joints, shown in Fig. 1.2.16, are currently being installed in the test facility. Thermocouples will not be attached to these units. PERIOD ENDING OCTOBER 31, 1958 Remote Maintenance Demonstration Facility C. K. McGlothlan The contractor’s portion: of the field construction work for the remote maintenance demonstration facility (Fig. 1.2,17) in Experimental Engineering Building 9201-3 has been completed. This work consisted of fabricating and erecting the salt safety spill tray and enclosure and the railway for the manipulator. In addition, the contractor installed structures for the dummy reactor, piping, and dummy heat exchangers (Fig. 1.2.18). In Fig. 1.2.17 the fuel dump tank and two dummy heat exchangers are also shown in their approximate Table 1,2,4. Descriptions of Expansion Joints and Results of Tests Vendor Cook Electric Co. Zallea Bros. Zallea Bros. Material Stainless steel Inconel Stainless steel Test fluid Sodium Molten salt Sodium Maximum traverse of joint 2]/4 in. 25/8 in. 25/8 in. Number of cycles to failure 71 49 60 Location of failure Bellows seam weld Thermocouple attachment Thermocouple attachment to bellows to bellows COOK ELECTRIC CU TG s UNCLASSIFIED PHOTO 32544 FLEXCONICS CORP JIOINTS . JOINT Fig. 1.2.16. Expansion Joints To Be Tested in the Expansion-Joint Test Facility. 35 e e MOL TEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED PHOTO 32443 Fig. 1.2.17. Remote Maintenance Demonstration Facility. position. Although this facility contains a dummy reactor and two dummy heat exchangers, it will con- tain a loop with a pump for circulating molten salt at 1250°F through 3%-in. and 6-in. Inconel pipe. It will be demonstrated in this facility that any or all components of this quarter-scale simulated reactor system can be maintained remotely by a General Mills manipulator as viewed through a stereo-type closed-circuit television system. The procedures developed can be adapted for use on a full-size reactor system. The manipulator will operate on a 40-ft-long track with a 22-ft span. The vertical travel of the mani- pulator arm will be 20 ft. This manipulator is now being constructed at General Mills Company. The first of three shipments of equipment for the two closed-circuit television systems has been 36 made by General Precision Laboratory. The dummy reactor has been fabricated and is on the job site. Fabrication in a local job shop of the 3‘/2-in. and 6-in. loop piping subassemblies is in progress. Bids are being obtained for the fabrication of 20 sets of freeze flanges and the removable electric heater and insulation units for the 3%-in. and 6-in. piping. In addition, bids are being obtained to modify an existing molten-salt pump for use in this system. Approximately two-thirds of the engineering work for this job has been completed. The remaining portion of the engineering work consists mostly of electrical, instrumentation, and remote tooling de- sign. PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED PHOTO 12442 Fig. 1.2.18. Heat Exchanger, Piping and Reactor Support Structure of Remote Maintenance Demonstration Facility. HEAT EXCHANGER DEVELOPMENT J. C. Amos Al 2Cl6 Thermal-Convection Loop A thermal-convection loop has been designed for evaluating the heat transfer performance of aluminum chloride gas. This gas has been suggested as a possible low-pressure gas coolant which would eliminate many of the containment problems in- herent in high-pressure gas systems. Aluminum chloride gas is of interest for the molten-salt re- actor concept because of the almost complete dis- sociation reaction (Al ,Cl ,—=2AICl,) that occurs between 900 and 1200°F. This dissociation pro- cess is reversible and results in a large increase in the heat capacity of the gas over the temperature range presently considered for the molten-salt re- actor fuel (see Chap. 2.3, ‘‘Chemistry,”” for dis- cussion of ““Gaseous Aluminum Chloride as a Heat Exchange Medium'’). Operation of a thermal- convection loop should provide an initial qualitative evaluation of the effect of dissociation on the heat transfer characteristics of this gas. The small-scale apparatus shown schematically in Fig. 1.2.19 was assembled to obtain a subli- mation temperature vs pressure curve for aluminum chloride and to provide preliminary compatibility data for the gas in the container materials of principal interest. A plot of the temperature vs vapor pressure data is presented in Fig. 1.2.20. Inconel and INOR-8 corrosion samples installed in the test apparatus were exposed to aluminum chloride vapors for 1000 hr. Visual and metallo- graphic examination disclosed intergranular attack to a maximum depth of 2 mils at the high-temperature (1200°F) end of the Inconel specimen and grain- boundary penetration to a depth of 0.5 mil at the high-temperature end of the INOR-8 specimen. The 37 MOL TEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 34512 PILOT VALVE MOORE NULL- BALANCE PRESSURE- O O INSTRUMENT MEASURING DEVICE — AIR SUPPLY — : BELLOWS F M \PRESSURE GAGE . Pl CALROD HEATER—2 O HERMOCOUPLE ~—CLAMSHELL HEATER — — [—‘M_—\\J CREINTIONLNE o= SHEDULE 0 P />< w——CORROSION SPECIMENS THERM PLE << CRMOMIUPLE S % THERMOCOUPLE O (O—=——CALROD HEATER 7::- E— Ny __—>THERMOCOUPLE THEwocouHE\_,g 9; 4 —S0LID AICI4 CONTAINER Fig. 1,2.19. Diagram of Apparatus for Determining Aluminum Chloride Sublimation Temperature vs Pressure Data and Preliminary Compatibility Data. UNCLASSIFIED ORNL-LR-OWG 34513 130 110 | 100 PRESSURE ( psig) /La—OATA BY FRIEDEL AND CRAFTS, FROM _ C.A THOMAS, " ANHYDROUS ALUMINUM 7 CHLORIDE IN ORGANIC CHEMISTRY," REINHOLD J ~ 300 400 500 600 700 TEMPERATURE (°F) Fige 1,2.20. Aluminum Chloride Sublimation Tempera- ture vs Pressure Data Plot. 38 cold end (750°F) of both specimens revealed a light, metallic-appearing deposit of a maximum thickness of 2 mils on the INOR-8 specimen and 1 mil on the Inconel specimen. Spectrographic analysis showed this deposit to be predominantly aluminum. Representative photomicrographs of these specimens are shown in Figs. 1.2.21 through 1.2.24. These preliminary data indicate that either of the materials tested would be satisfactory for use in the thermal-convection loop and suggest that there may be no serious compatibility problem between aluminum chloride gas and INOR-8, the container material proposed for a molten-salt power reactor system. Molten-Salt Heat Transfer Coefficient Measurement Modifications to the molten-salt heat transfer co- efficient test facility required for a salt-to-salt heat transfer test are essentially complete. Data runs will be started early in the next quarter. A de- scription of this test facility and the results of a salt-to-liquid metal test were reported pre- viously. 1% 10, . Amos, R. E. MacPherson, and R. L. Senn, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 27. 1), C. Amas, R. E. MocPherson, oand R L. Senn, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 31, UNCLASSIFIED T-15799 o Fig. 1.2.21. Top of Inconel Specimen Exposed to Aluminum Chloride Gas at 1200°F for 1000 hr. Etchant: glyceria regia. 250X, I ) UNCLASSIFIED SAMPLE 1 - 3 ? T-15798 LT (& ] e z SAMPLE 2 & ok Fig. 1.2,22. Bottom of Inconel Specimen Exposed to Aluminum Chloride Gas at 750°F for 1000 hr. glyceria regia. 250X. Etchant: Fig. 1.2.23. Top of INOR-8 Specimen Exposed to Aluminum Chloride Gas at 1200°F for 1000 hr. Etchant: aqua regia. 250X, PERIOD ENDING OCTOBER 31, 1958 Fig. 1.2.24. Bottom of INOR-8 Specimen Exposed to Aluminum Chloride Gas at 750°F for 1000 hr. As polished. 250X, DESIGN, CONSTRUCTION, AND OPERATION OF MATERIALS TESTING LOOPS Forced-Circulation Loops J. L. Crowley Construction and operation of forced-circulation corrosion testing loops was continued. Operation of seven new loops was started during this period, and three loop tests were terminated. By the end of September, 13 loops were in operation and one loop was ready to be filled. Nine of the operating loops are constructed of INOR-8 and four of Inconel. A summary of operations as of September 30, 1958, is presented in Table 1.2.5. Of the three terminated loops, one was an In- conel loop (9377-1), one an INOR-8 loop (9354-2), and the other an Inconel bifluid loop (designated CPR). The decision was made to terminate loops 9377-1 and 9354-2 after the power failure on April 6 caused their failure.'? The CPR loop was termi- nated after completion of a year of operation to ]2J. L. Crowley, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 33. 39 oy Table 1.2.5. Forced«Circulation Loop Operations Summary as of September 30, 1958 Composition Aporoximate Approximate Maximum Minimum Maximum Hours of Loap Loop Material Number of pp Reynolds Wall Fluid Fluid Operation at . . . ) Flow Rate L Comments Designation and Size Circulated (gpm) Number at Temperature Temperature Temperature Conditions Fluid* 9 1100°F (°F) (°F) (°F) Given CPR Inconel, 3/fl in. sched 40, 122 1 5,000 1250 1095 1190 9148 Operation terminated to provide cperating 1 in. OD x 0,035 in. wall, Sodium 7 97.700 1085 1135 space for an additional loop % in. OD x 0,035 in. wall ' 9344.1 Inconel, ]/2 in. 0D, 123 2 3,250 1300 1105 1230 7930 Fluid dumped on July 3 for changing of 0.045 in. wall pump motor 9354-3 INOR-8 84 2.75 1200 1100 1145 6289 Operation stopped tempororily on June 16 - for removal of section of loop piping for Hot feg, /g In- sched 40 4,500 metollurgical examinotion; restarted Cold leg, ¥ in. OD, 5,400 with new batch of fluid after pump re- 8.045esi’n./3vclxrlll placed and emergency heaters added 9344-2 Inconel, ]/2 in. OD, 12 2.5 8,200 1200 1000 1090 5861 Fluid froze on July 2 when faulty relay 0.045 in. wall caused loss of pump drive; fluid partially froze again on Aug. 5 when a building power failure caused failure of the pump drive motor; operation re- sumed after both incidents 9377-3 Inconel, !/2 in. OD, 131 2 3,400 1300 1100 1240 4774 Fluid dumped on July 23 for repair of 0.045 in. wail leaking oil rotory wnion 9354-1 Inconel, I/2 in. OD, 126 25 2,000 1300 1100 1190 4308 Fluid dumped on June 20 for instal- 0.045 in. woll lation af emergency heaters; fluid dumped on July 29 when a lug adapter made of SP-16 material de- veloped a crack; adapter replaced and operation resumed 9377-1 Inconel, I/2 in. OD, 126 2 1,600 1300 1100 1175 3413 Operation terminated after rupture de- 0.045 in. wall scribed previously 9354.5 INOR-8, :"/8 in. OD, 130 1 2,200 1300 1095 1230 3383 Fiuid partially froze on Aug. 5 when 0.035 in. wall building power failure caused pump drive motor failure; loop contains graphite samples LY0d3 Y SSIHYI0¥d JyOLIVI Y LTVS*NILTIOW Table 1,2,5 (continued) Composition Approximate Approximate Maximum Minimum Maximum Hours of Loop Loop Material Number of PP Reynolds Wall Fluid Fluid Operation at - . . . Fiow Rate o Comments Designation and Size Circvlated (gpm) Number at Temperature Temperature Temperature Conditions Flyid* s 1100°F (°F) (°F) (°F) Given 9354-4 INOR-8 130 2.5 1300 1130 1200 1646 Operation started July 23; loop con- . toins three INOR-8 sample inserts Hot leg, 7% in. sched 40 3,000 Cold leg, % in. OD, 3,500 0.045 in. wall MSRP-7 INOR-8, I/2 in. OD, 133 Not Not 1300 1100 1210 1302 Operation started on Aug. 7 0.045 in. wall available available 9377-4 Inconel, ‘/2 in. OD, 130 1.75 2,600 1300 1100 1185 1241 First loop with cooler coil and con- 0.045 in. wall trols of new design; operation started July 24; new protective devices pre- vented freezing of flvid on two occasions of pump failures; pumps re- placed and operation resumed 9354-2 INOR-8, '/2 in. OD, 12 2 6,500 1200 1050 1140 1052 Operation terminated after rupture de- 0.045 in. wall scribed previously MSRP-6 INOR-8, ]/2 in, OD, 134 Not Not 1300 1100 1190 822 Pump bound on first startup attempt; 0.045 in. wall available available during preheating after pump change, loop tubing parted between cooler and pump before fluid was introduced; tubing being examined by Metallurgy Division for cause; loop repaired and operation storted on Aug. 27 MSRP-8 INOR.8, ‘/2 in. OD, 124 2.0 4,000 1300 1100 1210 605 Operotion started on Sept. 4 0.045 in. wall MSRP-9 INOR-8, 1/2 in. OD, 134 Not Not 1300 1m0 1225 471 Operation started on Sept. 10 0.045 in. wall available available MSRP-10 INOR-8, ‘/2 in. OD, 135 Not Not 1300 1100 1210 264 Operation started on Sept. 19 0.045 in. wall available available *Composition 12: NaF-LiF-KF (11,5-46,5-42 mole %) Composition 130: LiF-BeF ,-UF (62-37-1 mole %) Composition 84; Na F-LiF-BeF2 (27-35-38 mole %) Composition 131: LiF-BeF.‘,-UF4 (60-36-4 mole %) Composition 122: NaF-ZrF4-UF4 (57-42-1 mole %) Composition 133: LiF-BeF2-UF (71-16-13 mole %) Composition 123: NaF-BeF_.UF, (53-46-1 mole %) Composition 134: LiF-BeF,-ThF UF (62-36.5-1:0.5 mole %) Composition 126: LiF-BeF,-UF, {53-46-1 mole %) Composition 135;: Na F-BeF.‘,-ThF“-UF‘1 (53-45.5-1-0.5 mole %) 8% 8561 ‘I ¥3IFOLD0 ONIANT @Ooi¥d3d MOL TEN-SALT REACTOR PROGRESS REPORT provide space for an additional loop (see Chap. 2.1, *Metallurgy,’’ for results of metallurgical exami- nations). Other than minor occurrences involving individual loops, there were only two incidents during the quarter that affected all loops. The first was a building power failure on August 5 of less than 1 min duration that affected seven operating loops. At this time, while an extensive emergency elec- trical system was being installed, all loops were equipped so that momentary power failure would automatically place them on isothermal operation. Five of the loops did resume operation isothermally after the power was reapplied. On two loops, 0354-5 and 9344-2, the pump drive motors tripped out on overload when required to restart under the load of the magnetic clutch and the pump. These two loops were partially frozen when circulation was not maintained but were both successfully thawed and placed back into operation. The over- load breakers on all motors, although sized accord- ing to proper electrical code, were increased to carry a larger startup current. The other incident that affected all loops was a changehouse fire across the street from where the forced-circulation corrosion loops are located. The fire destroyed the poles supplying an alternate power source to the building and also threatened to spread to the operating area. All operating loops on that side of the building were placed on isothermal operation in preparation for dumping. It was not necessary to dump, however, and the loops were returned to normal operating conditions after the danger was past. Crders were placed for the fabrication of three INOR-8 loops and two Inconel loops. One INOR-8 loop will be used as a spare and two Inconel loops will replace two presently operating loops sched- uled to be terminated. Of the thirteen facilities with operating loops, six are of the new design, which gives maximum protection, as described previously. 2 The re- maining seven are in various stages of improved protection that could be provided without disturbing the loop piping. In-Pile Loops D. B. Trauger J. A, Conlin P. A. Gnadt Assembly of the first MSR in-pile loop will be completed by November 1, and the loop is sched- uled for insertion in the MTR December 1, 1958 42 (MTR cycle 113). A second loop is being as- sembled, and parts for a third loop are being fabri- cated. The second loop will provide additional data and serve as a backup loop in the event of a failure of the first loop. The instrumentation and controls required for loop operation are being pre- pared at the MTR. The schedule for operation of the second loop will be determined by the success of the first loop and the outcome of pump tests presently under way, as described previously.'® In these tests, two methods have been tested for starting the loop in the reactor. The first method consisted of filling the loop during assembly, freezing the salt, and then thawing it out after insertion in the reactor. It was demonstrated in four successive tests with the first prototype pump that progressive meltout from the pump to the nose coil could be achieved without damage to the loop. During meltout, how- ever, pump speed perturbations occurred, and several hours after the fourth meltout the pump seized. Investigation showed a large accumulation of salt in the intermediate region between the sump and bearing housing and salt in a granular form mixed with oil in the copper rotor region. Although x-ray examination indicated that salt had collected in the intermediate region without the concurrent meltout procedure, melting out of the loop was probably a contributing factor. A second method of starting the loop in the re- actor was found to be satisfactory in a second prototype pump test. The original filling system, which consisted of an integral fill tank that was heated and drained into the pump after installation in the reactor was modified by changing the in- ternal passages of the pump to accommodate the high surface tension of the salt. This system was incorporated in the first loop. The appearance of salt in the region between the sump and bearing housing in both the first and second prototype pump tests presents a further difficulty. The melting and thawing of the loop in the first test was thought to have been a con- tributory cause; however, the salt deposit appeared again in the second pump test but to a lesser degree. This salt transfer is not understood, but is be- lieved to be, in part, a function of the suspected high surface tension of the fluid. At present the second prototype pump, similar to the pump used ]3D. B. Trauger, J. A. Conlin, and P, A, Gnadt, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 34. PERIOD ENDING OCTOBER 31, 1958 in the first loop, is being tested to determine the the pump of the second loop. Since construction of effect of this phenomenon on long-term operation. the first ioop is beyond the point that would permit A third test of a modified pump is being set up in pump modification, the period of in-pile operation - an attempt to correct the situation. Modifications will be based on the duration of the pump life test indicated by these tests will be incorporated in presently in progress. 43 eSS MOL TEN-SALT REACTOR PROGRESS REPORT 1.3. ENGINEERING RESEARCH H. W. Hoffman Reactor Projects Division PHYSICAL PROPERTY MEASUREMENTS equation, W. D. Powers R. H. Nimmo Ho = Hygoo = a+ bT + T2 (cal/g) | Enthalpy and Heat Capacity where the temperature T is expressed in °C. The coefficients & and ¢ of Table 1.3.1 may be used to The enthalpies, heat capacities, and heats of | fusion were determined for six additional salt mix- obtain the heat capacity, “pt from the following i tures. Of these, two were fuels of the alkali metal- ~ €dquation: | berylllu.m sy.stem,' and the other four were chlorides ¢ = b+ 2T (cal /g-°C) of possible interest as secondary coolants. The p results are given in Table 1.3.1, in which the con- For comparison and completeness, previously de- stants «, b, c, are the coefficients in the enthalpy termined values of several related mixtures are Table 1.3.1. Enthalpy Equation Coefficients und Heots of Fusion of Several Fused Salt Mixtures Enthalpy Equation Heat of Fusion (cal/g) Temperature Coefficients* Salt Mixture Phase R o) Temperature ange (" C) H, - HS a b c (°C) x 10~ NaF-BeF,-UF Solid 100-300 ~7.8 +0.251 +11.83 (53-46-1 mole %) Liquid 400—800 ~5.7 +0.312 +10.87 365 23 LiF-BeF -UF, Solid 100-300 -10.9 +0.293 +19.33 - (53-46-1 mole %) Liquid 400-800 -12.8 +0.539 ~1.90 400 63 LiF-BeF -UF ,** Solid 100300 ~9.6 +0.293 +2.18 (62-37-1 mole %) Liquid 470-790 +24.2 +0.488 +2.52 430 122 . NaCl-CaCl, Solid 100~500 4.5 +0.158 +7.60 (49-51 mole %) Liquid 550_850 —18.4 +0.358 ~7.70 500 48 LiCl-CaCt,** Solid 100425 ~3.2 +0.167 +T11.31 (62-38 mole %) Liquid 520-895 +25.8 +0.289 496 61 LiCl-SrCl*+ Solid 100—495 3.6 +0.140 +4.25 (52-48 mole %) Liquid 550850 -39.8 +0.356 ~8.16 525 43 LiCl-BaCl,** Solid 100—500 _4.5 +0.146 +2.50 (70-30 mole %) Liquid 550850 —6.6 +0.278 ~3.93 31 45 LiCl-KCl Liquid 450—800 +15.3 +0.361 ~3.12 (70-30 mole %) i LiCl-KCl Liquid 400—800 +21.2 +0.315 —0.18 (60-40 mole %) . LiCl-KCl Liquid 500800 +38.5 +0.239 +4.62 (50-50 mole %) *Heaot copacity may be evoluated from table os cp = b+ 2cT; enthalpy given as HT - H300C =a+bT+ C‘T2. 44 **Previously reported results repeated for completeness. repeated in Table 1.3.1. Measurements of the solid state have not been completed for the three LiCl-KCl mixtures studied in the liquid state. The results of initial measurements of the mix- ture LiF-BeF ,-UF, (53-46-1 mole %) were some- what erratic. Specifically, the enthalpy values were observed to be a function of the heating- cooling cycle history of the sample. Thus, for samples heated to 450°C and suddenly quenched in the calorimeter, an average enthalpy of 225 cal/g was obtained. Data for four samples measured at 500°C resulted in an average enthalpy of 220 cal/g, a value less than that for the original measurements at 450°C. Since BeF , and mixtures containing large percentages of BeF, are known to form glasses, it was suspected that the anomalous be- havior was associated with this phenomenon. Ac- cordingly, the samples were remeasured with slower cooling in the calorimeter (100 min to equilibrium in comparison with the normal 20 to 30 min). The slower cooling rate was achieved by inserting an insulating liner within the cavity of the copper block. Under these conditions, the enthalpy values were temperature-consistent and of higher magni- tude. Since the observed enthalpy is the energy difference of the salt between the two given tem- perature levels, the higher values indicate a lower final energy state and hence a more stable form. Evidence obtained by varying the heating rate and pattern in attaining the initial sample temperature suggests that the upper energy state was identical for all the samples. The results presented in Table 1.3.1 for BeF ,-containing mixtures are based on the slow-cooling data. The enthalpies and heat capacities of the LiCl- KCI mixtures are compared in Table 1.3.2 on a gram-atom basis with results obtained by Douglas and Dever' at the National Bureau of Standards for the eutectic mixture LiCl-KCl (59-41 mole %). It is of interest to note the change in the tempera- ture dependence of the heat capacity with decreas- ing LiCl content in the mixture. Between 500 and 800°C the heat capacity of the 70-30 mixture de- creases by 5.7%, that of the 60-40 mixture by only 0.3%, and that of the 59-41 NBS mixture by 3.5%, while that of the 50-50 mixture increases by 9.7%. Additional data will be required for verification of this indicated effect of composition. ]T. B. Douglas and J. L. Dever, Nationa!l Bureau of Standards Report 2303, Feb. 20, 1953. PERIOD ENDING OCTOBER 31, 1958 Surface Tension Preliminary determinations of the surface tension of the mixture LiF-BeF ,-UF, (62-37-1 mole %) were made with the use of the previously described maximum-bubble-pressure method.? The results, presented in Fig. 1.3.1, can be represented by the equation, o (dyne/cm) = 235,5 - 0.09 T (°C) , to within 5% over the temperature range of 460 to 2W. D. Powers, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 38. Table 1.3.2. Enthalpies and Heat Capacities of LiCl-K Cl Mixtures of Various Compositions LiCl-KC! Composition (Mole %) Temperoture (°C) 70-30 60-40 50-50 59-41* Enthalpy (cal/g+atom) Hp — Hagec Hy - Hgoe 500 4890 4920 4950 5160 600 5740 5790 5800 6030 700 6580 6650 6670 6890 800 7390 7510 7570 7730 Heat Capacity (cal/g-otom+°C) 500 8.58 8.65 8.33 8.76 600 8.42 8.64 8.60 8.63 700 8.26 8.63 8.87 8.53 800 8.09 8.62 9.14 8.45 *Results of Douglas and Dever at NBS (ref. 1). UNCLASSIFIED 500 ORNL-LR -DWG 34514 —_ 1 . —=—Jo ] I | g | e S g ; : : A R — £ I . \ | | = {170 ) PR S — —— e ,l e | > | TUBE NO. TIP 1D {in.) : | 2 o 4 00700 3 : & a2 0.0520 | B H 100 — ' S S P O — " ° 3 00400 | <1 4 . L ‘ [ ‘ fid : ! A 50 L | i i i 450 500 550 600 650 700 750 800 TEMPERATURE (°C) Fig. 1.3.1. Surface Tension of LiF-Ber-UF4 Mix- ture (62-37-1 mole %). 45 MOLTEN-SALT REACTOR PROGRESS REPORT 750°C. Three tubes that differed in inside diam- eter were used. |t is felt that a major source of the observed scatter of the data is the uncertainty in the tube-tip inside diameter as a result of pro- gressive plugging during a test series. Future studies will include measurement of other salts of known surface tension and noncorrosive nature in order to isolate the plugging effect and to estab- lish accuracy limits for this technique. Apparatus Fabrication and Calibration The apparatus? for direct measurements of the coefficients of thermal expansion of beryllium salt mixtures has been fabricated and is being as- sembled. The equipment? for determining the thermal conductivity of the beryllium salts is being fabricated. The thermal gradient, and hence the heat flux, at the ends of the cylinders is to be de- termined from temperature measurements made at a large number of positions within the cylinders. Since it was not found to be feasible to locate the thermocouples coaxially, interpretation of the ther- mal profile under conditions of heat exchange at the cylinder walls becomes difficult. [t is planned to analyze statistically the experimental tempera- ture data from the initial measurements so as to ascertain the thermocouple locations that yield maximum heat flux precision. Efflux-cup viscometers are being calibrated that have larger volumes than the viscometers used previously.? It is believed that effects such as ““surface-humping’’ will be sufficiently reduced to allow more precise measurement of the viscosity of the beryllium salts., Apparatus for the determina- tion of the electrical conductivity of the beryllium salts is being assembled. 46 MOLTEN-SALT HEAT-TRANSFER STUDIES F. E. Lynch Heat balance difficulties have been experienced in the evaluation of experimental data on the heat- transfer coefficient of LiF-BeF2-UF4 (53-46-1 mole %) flowing through an electrical-resistance- heated smalil-diameter Inconel tube. Specifically, the measured electrical heat input has been ob- served to be only one-half the convective heat gained by the salt, as calculated from temperature A check made with recalibrated instruments and with additional instru- and flow rate measurements. ments indicated that the power measurements were not the cause of the discrepancy. New, precali- brated thermocouples have been installed, and the weigh-beam flow-measuring device has been re- checked. A second series of runs will be initiated shortly under conditions that closely approximate those of the earlier experiments. The molten-salt pumping system is being as- sembled. Since it is the purpose of this study to determine whether effects such as surface film formation occur after long exposure of the metal to the beryllium salt, it is sufficient to obtain rela- It is now planned, however, to include an experimental turbine-type flowmeter in the loop and also to de- tive values of the heat-transfer coefficient. termine the flow from the pump calibration, so that if the flow values obtained by these two methods consistently agree, an estimate of the heat-transfer coefficient can be made. The system will contain an INOR-8 test section, if available, in addition to the Inconel test section. The two test sections will be located in adjacent legs of the loop, and a three-electrode heating scheme will be used. It will also be necessary to include a heat exchanger to control the system temperature. PERIOD ENDING OCTOBER 31, 1958 1.4. INSTRUMENTATION AND CONTROLS H. J. Metz Instrumentation and Controls Division RESISTANCE-TYPE CONTINUOUS FUEL has sufficient resistance to provide a useful milli- LEVEL INDICATOR volt output from the probe, the data for a given set R. F. Hyland of conditions could not be reproduced. One pos- sible source of these inconsistencies is polariza- tion, a phenomenon commonly encountered in con- ductivity measurements. This and possible surface tension effects will be investigated in future tests. A series of tests of the Inconel resistance-type fuel-level-indicating elements described previ- ously] was completed, and typical data are sum- marized in Table 1.4.1. The results of these tests 'R. F. Hyland, MSR Quar. Prog. Rep. June 30, 1958, were somewhat disappointing. Although fuel 130 ORNL-2551, p 48. Table 1.4.1. Summary of Data from Tests of Resistance-Type Fuel Level Probes Probe supply voltage (regulated): 3.4 v ac Probe supply current (constant): 1.0 amp Frequency: 60 cps Readout device: Hewlett-Packard Model 400H VTVM Fuel 130: LiF-BeF,-UF, (62-37-1 mole %) Test Level Range Probe Output Zero to Maximum Run Date Temperature (in. of fuel at Zero Level Level Probe Qutput Probe Qutput NO- ]958 (OF) ]30) (mv) Range (mv) Spcn (mv) 1 7/29 1000 0-6 15.7 15.4-12.0 3.4 2 9/2 1000 0-6 15.7 14.9--8.0 6.9 1 8/5 1100 0-6 15.7 15.2-11.5 3.7 2 8/22 1100 0-6 15.7 14.2-6.0 8.2 1 8/11 1200 0-6 15.7 15.5-12.2 3.3 2 8/22 1200 0-6 15.7 14.6-7.0 7.6 1 8/20 1300 0~-6 15.7 14.8-7.4 7.4 2 9/9 1300 0-6 15.7 14.7=7.1 7.6 47 MOLTEN-SALT REACTOR PROGRESS REPORT 1.5. ADVANCED REACTOR DESIGN STUDIES H. G. MacPherson Reactor Projects Division LOW-ENRICHMENT GRAPHITE-MODERATED MOLTEN-SALT REACTORS H. G. MacPherson A rough survey of the nuclear characteristics of graphite-moderated reactors utilizing an initial com- plement of low-enrichment uranium fuel has been made.! The pertinent characteristics of a number of such reactors are presented in Table 1.5,1. |t may be seen that reactors could be constructed that would operate with fuel enriched initially with as little as 1,25% U235; initial conversion ratios as high as 0.8 could be obtained with a fuel en- richment of less than 2%. Highly enriched uranium would be added as make-up fuel, and such reactors could probably be operated te burnups as high as 60,000 Mwd/ton. The total fuel cycle cost would be approximately 1.3 mills/kwhr, of which 1,0 mill would be the cost of the enriched U233 that would be burned up. H. G. MacPherson, Survey of Lou-Enrichment Molten- Salt Reactors, ORNL CF-58-10-60 (Oct. 17, 1958). A NATURAL-CONVECTION MOLTEN-SALT REACTOR A study was made of a large natural-convection power reactor to determine the approximate size of the components and the fuel volume for comparison with other molten-salt reactors in which the fuel is circulated by a pump.2 A reactor with a power rating of 576 Mw (thermal) was chosen for study so that the heat generated in the core would be com- parable to that generated in the core of the interim molten-salt reactor. The rather simple configuration of reactor, heat exchanger, and piping shown in Fig. 1.5.1 was used in the study. All calculations were done on the basis of one heat exchanger, one riser, and one downcomer, Since the frictional losses in the piping are determined primarily by the expansion and contraction losses and are insensitive to wall 2J. Zasler, 576 Mwt Natural Convection Molten Salt Reactor Study, ORNL CF-58-8-85 (Aug. 25, 1958). Table 1.5.1. Choracteristics of Low-Enrichment Grophite-Moderated Molten-Salt Reactors Uranium Volume of Uranium Critical Mass Initial Case Yolume Fraction Multiplication g ;100 Fuelin Core in Core of U235 Conversion of Fuel in Core Constaont (%) (ft3) (kq) (kg) Ratio 1 0.05 1.05 1.30 395 22,100 298 0.546 2 1.10 1.45 143 8,000 16 0.492 3 0.075 1.05 1.25 427 23,900 298 0.635 4 1.10 1.39 167 9,350 130 0.600 5 0.10 1.05 1.275 445 24,900 318 0.707 6 1.10 1.46 179 10,000 146 0.668 7 0.15 1.05 1.525 474 26,600 405 0.796 8 1.10 1.80 197 11,000 198 0.780 9 0.20 1.05 2.24 520 29,100 652 0.865 10 1.10 2.88 206 11,540 332 0.865 1 0.25 1.05 4.36 575 32,200 1400 0.900 12 1.10 7.05 240 13,430 952 0.865 48 friction, the single riser can be replaced by a num- ber of risers having the same height and total cross sections, Likewise, the heat exchanger can be replaced by a number of heat exchangers having the same length of tubes and total number of tubes. The principal results of these calculations-are given in Fig. 1.5.2. The diameter of the riser needed to provide natural convection, the length and number of heat exchanger tubes, and the total fuel volume are given as functions of riser height PERIOD ENDING OCTOBER 31, 1958 and the temperature drop of the fuel in going through the heat exchanger. Natural-convection power re- actors of this size are characterized by high fuel volumes and large numbers of heat exchanger tubes. Therefore it is expected that the initial cost of a natural-convection reactor would be higher than that for a forced-circulation system, and the larger number of heat exchanger tubes cast some doubt as to whether the natural-convection system would actually be more reliable than the forced-circula- tion system. UNCLASSIFIED ORNL-LR-DWG 32561 L —= HEAT EXCHANGER . /S-ft DIA REACTOR / Fig. 1.5.1. Molten-Salt Reactor. Schematic Diagram of a Natural-Convection 49 NUMBER OF HEAT VOLUME OF FUEL LENGTH OF HEAT RISER 50 UNCLASSIFIED ORNL-LR-DWG 32565 (x 10%) /‘)'-\'7—7= H‘SOF 22 / } o = _— AT = 200°F o g e J— // AT = 250°F 16 (X 403) ‘ Q \%\ AT =475°%F R 40 < ‘< A7 = 260°F ; s S 30— AT = 225% = "k\_‘x : o7 =225 = S e § AT = 250°F J | \ —— L 20 [ AT = 250°F E 15 // ; / /AT = 225°%F b;i P | ———"TAT = 200°F a _ o g 10 ///i’// AT =1T5F T g '/_/ // / 5 6 : '_'-l—n_____ cr 5 ——‘.——_“_; R = 4 e —— AT = 200°F — ) AT = 225°%F AT = 25C°F 3 i 10 15 20 25 30 35 40 45 RISER HEIGHT (ft) Fig. 1.5.2. Design Characteristics of a 576 Mw (thermal) Natural-Convection Molten-Salt Reactor. Part 2 MATERIALS STUDIES 2.1. METALLURGY W. D. Manly D. A. Douglas A. Taboada Metallurgy Division DYNAMIC CORROSION STUDIES J. H. DeVan J. R. DiStefano R. C. Schulze The three-phase corrosion testing program, previ- ously described, for which Inconel and INOR-8 thermal-convection and forced-circulation loops are utilized, was continued, During the quarter, cor- rosion studies were completed of two Inconel and four INOR-8 thermal-convection loops. Examina- tions were also completed of three Inconel forced- circulation loops, which were the first loops oper- ated in the third phase of the corrosion testing program. Samples removed from an INOR-8 forced- circulation loop that was temporarily shut down for repair of a leak were also examined. A number of new forced-circulation loop tests were started; the present status of all such loop tests now in oper- ation is given in Chap. 1.2 of this report. 1), H. DeVan, J. R. DiStefano, and R. S. Crouse, MSR Quar. Prog. Rep. Oct. 31, 1957, ORNL-2431, p 23, UNCLASSIFIED T-15734 L0l0 \Q\ / .01 ] 013 Ql4 o R .018 : | sl 5] _ | |1z o~ » L ’ o~ Fig. 2.1.1. Hot-Leg Sample from Inconel Loop 1222 Which Circulated Salt 130 for 1000 hr ot 1350°F. Inconel Thermal«Convection Loop Tests One of the two Inconel thermal-convection loops examined was operated as a supplement to the first phase of corrosion testing of Inconel in contact with fused-fluoride-salt mixtures. Loop 1222 circu- lated salt mixture 130 (LiF-BeF2-UF4, 62-37-1 mole %) for 1000 hr at a maximum hot-leg tempera- ture of 1350°F, and it was attacked to a depth of 3 mils in the form of intergranular voids (Fig. 2.1.1). Previously,? loop 1178 had circulated the same salt mixture for 1000 hr at 1250°F and was attacked to a depth of 1 mil. Comparison of the results from the two loop tests emphasizes the effect of operating temperature on depth of corrosion in Inconel sys- tems. The 100°F increase in the maximum operating temperature caused an increase in the depth of at- tack from 1 to 3 mils in the 1000-hr period. The second Inconel thermal-convection loop (1207) that was examined was the first that had operated for an extended period of time. It circulated salt mixture 12 (NaF-LiF-KF, 11.5-46.5-42 mole %) for 4360 hr at o maximum temperature of 1050°F. The attack in loop 1207 was in the form of voids to a depth of 2 mils. INOR«8 Thermal«Convection Loop Tests The four INOR-8 thermal-convection loop tests which were completed concluded the first phase of the corrosion testing program for INOR-8, The op- erating conditions and chemical analyses of the salts circulated in the loops are presented in Table 2.1.1. None of the loops showed measurable attack, even though one loop (1213) operated for 3114 hr, A hot-leg section of loop 1203 after exposure to salt mixture 125 for 1000 hr at 1250°F is shown in Fig. 2.1.2. Comparable samples from the other INOR-8 loops were similar in appearance. Inconel Forced«Circulation Loop Tests The operating conditions and results of three Inconel forced-circulation loop tests are presented in Table 2.1.2, These three loops were the first 2), H. DeVan, J. R. DiStefano, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 57. 53 MOL TEN-SALT REACTOR PROGRESS REPORT Table 2.1.1. Test Operoting Conditions and Chemical Analyses of Salts Circulated in INOR-8 Thermal-Convection Loops Maximum loop temperature: 1250°F Operatin Major Constituents Minor Constituents Loop Salt e Salt Composition (wt %) (ppm) No. No. Time (mole %) Sample Taken (hr) ) Be Th Ni Cr Fe 1203 125 1000 53 NaF =46 Bel:2 -0.5 Before test 1,73 8.59 2.82 150 72 222 UF,~0.5 ThF, After test 1.82 10,5 2.90 68 235 195 1204 131 1000 60 LiF-36 Be F2-4 UF“ Before test 20.4 6.97 128 30 275 After test 20,6 7.15 <10 79 195 1213 128 3114 71 LiF=29 ’l'l'\F:4 Before test* 62 20 25 115 1221 128 1000 71 LiF=29 ThFA Before test 62.3 35 37 103 After test 61.3 28 70 245 *Results of after-test analyses not yet available. operated in phase three of the corrosion testing UNCTLSgg'sTEDI program. The loop designated CPR was the first forced-circulation loop operated under molten-salt reactor program conditions. |ts design provided for the operation of separate sodium and salt circuits through a U-bend heat exchanger. The loop com- ; .007 pleted 9148 hr of operation with salt mixture 122 ™ . (NoF-ZrF4-UF4, 57-42-1 mole %) at a maximum ’ temperature of 1260°F and with the sodium at a . 009 maximum temperature of 1120°F. Operation was terminated because of a pump failure.* Metal- 010 lographic examination revealed moderate surface ; j - roughening and slight intergranular attack in the . e heated leg of the molten-salt circuit. The ‘ . . maximum intergranular penetration was less than ; ' ' ' — 1.5 mils, as shown in Fig. 2.1.3. In the sodium - 014 circuit, attack was limited to light su.face rough- ening. No evidences of cold-leg deposits were 25 018 found in either the salt or sodium circuits. - o The second and third loops described in Table ) a 2 N 2.1.2 also operated for extended time periods. Fig. 2.1.2. Hot-Leg Sample from INOR-8 Loop 1203 3). L. Crowley, MSR Quar. Prog. Rep. June 30, 1958, 3 Which Circulated Salt 125 for 1000 hr at 1250°F, ORNL-2551, p 33, After circulating salt mixture 126 (LiF-BeF ,- UF4, 53-46-1 mole %) for 3073 hr at a maximum hot- leg temperature of 1300°F, loop 9377-1 showed moderate to heavy surface roughening and heavy intergranular void formation to a depth of 5 mils, as shown in Fig. 2.1.4. The third loop, 9377-2, UNCLASSIFIED T-15185 Fig. 2.1.3. Hot-Leg Sample from Salt Circuit of Inconel Forced-Circulation Loop CPR Which Circulated Salt 122 for 9148 hr at 1300°F. PERIOD ENDING OCTOBER 31, 1958 revealed moderate intergranular void formation io a maximum depth of 8 mils and an average depth of 4 mils, as shown in Fig. 2.1.5. The loop op- erated 3064 hr and circulated the salt mixture 130 (LiF-Ber-UF4, 62-37-1 mole %). UNCLASSIFIED T-15011 o - 8 Fig. 2.1.4. Hot-Leg Sample from Inconel Forced- Circulation Loop 9377-1 Which Circulated Salt 126 for 3073 hr ot 1300°F. Table 2.1.2. Operating Conditions ond Results of Inconel Forced-Circulation Loop Tests Maximum salt-to-metal interface temperature: 1300°F Minimum bulk salt temperature: 1100°F Temperature difference ocross loop: 200°F* Salt flow rate: 2 gpm Loop Fluid Composition Operating Time Reynolds Results No., (mole %) (hr) Number s CPR (primary circuit) 57 NaF-42 Zer—] UFA 9148 5,000 Intergranular penetration to a depth of <1.5 mils CPR (secondary circuit) Sedium 9148 97,700 Light surface roughening; no cold-leg deposits 9377-1 53 LiF =46 Ber—l UF4 3073 1,600 Voids to a depth of 5 mils 9377-2 62 LiF=37 BeF2—] UF, 3064 3,000 Intergranular voids to a maxi- mum depth of 8mils; average depth, 4 mils *The respective temperature differences across the fused-salt and sodium circuits of the loop were 100°F, 55 MOL TEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED T-14663 \ "" 2w " ol I/' 2 010 2 F /' - Ol - » S '~ / 258 b 2 0i2 C v - 013 014 . s ‘ L e 2184 >~ ¥ 2 x . .‘ f - - ; -' "8"‘ e o Fig. 2.1.5. Hot-Leg Sample from Inconel Forced- Circulation Loop 9377-2 Which Circulated Salt 130 for 3064 hr at 1300°F. Results of Examination of Samples Removed from INOR+8 Forceds«Circulation Loop 93541 On July 29, 1958, after successful completion of 3370 hr of operation, loop 9354-1 was temporarily shut down to permit the repair of a small leak in a lug adapter near the hottest section in the loop. The cause of the leak was traced to a probable de- fect in the material used in fabricating the loop (see subsequent section of this chapter on ‘“Welding and Brazing Studies’’). The loop was constructed of ]/é-in.-OD, 0.045-in.-wall INOR-8 tubing and it circulated salt mixture 126. The maximum loop wall temperature was 1300°F and the minimum fluid temperature was 1180°F. The salt flow rate was 2.5 gpm and the Reynolds number was 2000. The damaged section, which was about 1 ft long and which is shown in Fig.2.1.6, was removed from the loop, and 2-in. samples were cut from the tubing on both sides of the lug adapter. Metallographic examination of these samples did not show evidence of surface attack. However, the electrolytic etch used to prepare the samples for metallographic ex- amination produced a greater darkening of the grain UNCLASSIFIED T-15512 Fig. 2.1.6. Damaged Section of INOR-8 Forced-Circulation Loop 9354-1 Showing Location of Leak, boundaries near the exposed surface than in the interior, as shown in Fig. 2.1.7. This effect indi- cates a faster etching rate of the boundaries near the surface and is presumably due to a slight dif- ference in the chemical composition of the region near the surface. As may be seen in Fig. 2.1.8, this surface effect is not apparent in as-received samples of the tubing used for the loop. The loop was repaired by replacing the damaged section, and it has again been placed in operation. Material Inspection G. M. Tolson J. H. DeVan Routine inspection of the INOR-8 material now being received has revealed that the quality is as good as that of Inconel and of stainless steels made to ASTM standards. The high rejection rate previously found for INOR-8 welded and drawn tub- ing is felt to have been due to poor welding quality of the heat of INOR-8 from which the tubing was made. The rejection of seamless INOR-8 tubing, with defects greater than 5% of the wall thickness used as criteria for rejection, is running well below 10%. UNCLASSIFIED | [ T-15580 | | w I — O = Fig. 2.1.7. Somple Taken from Damaged Section of INOR-8 Forced-Circulation Loop 9354-1 Showing the Effect of Electrolytic Etch on the Grain Boundaries Near the Exposed Surface. PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED | [, | T-15578 v Fig. 2.1.8. Sample of As-Received Tubing Used for Loop 9354-1 Showing Absence of Effect of the Electro- lytic Etch on the Grain Boundaries. Fully inspected welds made on INOR-8 material show a rejection rate of about 10%, which compares favorably with the rejection rate of welds made on Inconel and stainless steels. Evaluation tests were made of thermocouple spot welds to INOR-8 tubing. Although nondestructive testing did not reveal any rejectable spot welds, metallographic examinations showed microcracks in some spot welds. GENERAL CORROSION STUDIES E. E. Hoffman W. H. Cook D. H. Jansen Carburization of Inconel and INOR-8 by SodiumsGraphite Systems D. H. Jansen The sodium-graphite system, which is an effec- tive carburizing medium for nickel-base alloys,* was used to carburize Inconel and INOR-8 samples for additional tests to determine whether the me- chanical properties of these materials are detri- mentally affected by various degrees of carburi- zation. The carburizing system and the test pro- 4E. E. Hoffman, W. H., Cook, and D. H. Jansen, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 54. 57 MOLTEN-SALT REACTOR PROGRESS REPORT cedures used were described previously.® Results of the tests completed thus far are presented in Table 2.1.3. As may be seen, the data indicate that the carburization treatment increases both the yield and tensile strengths of Inconel and de- creases the ductility at room temperature, Similar room-temperature mechanical property data for the carburized INOR-8 specimens showed that the yield strength was increased, the tensile strength was slightly decreased, and the ductility was greatly SE. E. Hoffman, W. H. Cook, and D. H. Jansen, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 59. reduced. The change in each mechanical property, especially the ductility, was dependent on the amount of carburization that had occurred. There was close agreement of all duplicate and triplicate mechanical test values, that is, 2 x 102 psi vari- ation for strength values and £1.5% for elongation values. Inconel and INOR-8 specimens that were exposed in the sodium-graphite system for 40 hr at 1600°F are shown in Fig. 2.1.9. Heavy carburi- zation to adepth of 2mils was found on the Inconel, and even heavier carburization was found on the INOR-8. Creep tests have also been conducted in argon at 1200°F on the Inconel and INOR-8 specimens that Table 2.1.3. Results of Mechanical Tests in Room-Temperature Air of Inconel and INOR-8 Specimens Carburized in Sodium-Graphite Systems Mechanical Properties Exposure Exposure . Specimen Yield Tensile Elongation in 2-in. Temperature Period (°F) (hr) Treatment Strength Strength Gage (psi) (psi) (%) Inconel x 103 x 103 1200 40 Control? 241 80.8 40,1 Carburized 27.2 84.2 38.0 1400 40% Control 23.9 81.3 38.8 Carburized 27.0 83.9 36.8 400° Control 24.3 81.7 37.9 Carburized 30.9 92.9 27.0 1600 40€¢ Control 20.9 75.9 47.3 Carburized 28.5 91.5 25,7 INOR-8 1200 40 Control? 53.3 120.3 43.6 Carburized 53.1 119.6 43.8 1400 40% Control 53.7 122.0 42,6 Carburized 53.9 120.5 36.9 400° Control 53.2 120.7 43.0 Carburized 57.4 103.2 8.8 1600 40¢ Control 50.5 116.3 50.0 Carburized 52.8 98.5 7.8 %Control specimens exposed to argon at the conditions given, bTesf values given for these specimens are averages of two tests, c . . Test values given for these specimens are averages of three tests, 58 ¢ ~ . “INCONEL Etchant: copper regia. 500X. were carburized for 40 hr at 1600°F in the sodium- graphite system, and the results are presented in Table 2.1.4. The results of tensile tests of similar specimens and control specimens in air at 1250°F are presented in Table 2.1.5. In general, the carbu- rized specimens show lower ductility thon that of the control specimens. Exceptions to this trend were found, as indicated in Table 2.1.5, for some of the carburized INOR-8 material and for a heavily carburized Inconel specimen that had been exposed for 40 hr at 1600°F in the sodium-graphite system; the carburized pieces were more ductile than the control specimens. These results were verified by duplicate tests, and therefore the results of the duplicate tests are also included in Table 2,1.5. Carburization of INOR-8 by Fuel-130-Graphite Systems A comprehensive test in which INOR-8 (nominal composition: 70% Ni-16% Mo-7% Cr-5% Fe-2% other alloying additions) specimens were exposed to (1) fuel 130 (LiF-Ber-UFa, 62-37-1 mole %) containing a graphite rod, (2) fuel 130 without added graphite, and (3) argon, all in the same test container, at a temperature of 1300°F for 2000 hr, PERIOD ENDING OCTOBER 31, 1958 lUNCLASSIFIED! Y-26201 f:fi , «m_; ( 4 Fig. 2.1.9. Inconel and INOR-8 Specimens After Exposure to Sodium-Graphite System for 40 hr at 1600°F. b INOR-8 has been completed. The 40-mil-thick tensile-test specimens used were then subjected to tensile tests in order to determine whether the various treatments had affected the mechanical properties. Table 2.1.4, Results of Creep Tests of INOR-8 and Inconel Control Specimens and Specimens Corburized in a Sodium-Graphite System for 40 hr at 1600° F 1250°F argon 18,500 psi Test temperature: Test environment: Test stress: Elongation in 2-in. Time to Rupture Specimen Goge ) (%) INOR-8 Control 1734 23.8 Carburized 4031* 9.0* Inconel Control 38 21.2 Carburized 160 47.5 *Still in test, 59 MOLTEN-SALT REACTOR PROGRESS REPORT Table 2.1.5. Results of Mechanical Tests in 1250°F Air of Inconel and INOR-8 Specimens Carburized in Sodium-Grophite Systems Mechanical Properties Exposure Exposure . Specimen Yield Tensile Elongation in 2«in. Temperature Period (°F) (hr) Treatment Strength Strength Gage (psi) (psi) (%) Inconel x 103 x 103 1200 40 Control* 50.3 38.5 Carburized 48,6 33.5 1400 40 Control * 46.7 36.5 Carburized 49.9 23.0 400 Control* 16.4 45.2 26.5 Carburized 22.0 50.7 21.0 1600 40 Control* 11.3 42.4 16.5 1.4 18.5 Carburized 20.5 53,9 24.8 51.1 28.5 INOR-8 1200 40 Control* 70.2 17.0 Carburized 74.4 19.0 1400 40 Control* 74.6 18.5 Carburized 76.8 19.5 400 Control* 36.8 70.3 16.5 74.5 20.5 Carburized 42.4 82,9 18.0 85.2 22,0 1600 40 Control* 33.8 68.9 18.5 70.6 6.5 Carburized 40.5 82.8 13.0 *Control specimens exposed to argon ot the conditions indicated, The specimens were also examined metallographi- system contained 980 ppm carbon. These data and cally, and, as shown in Fig. 2.1.10, no carburiza- data obtained from carbon analyses of the fuel are tion was found on the surface of the INOR-8 speci- given in Table 2.1.6. men that was exposed to the fuel-graphite system. Room-temperature mechanical tests showed the Chemical analyses and mechanical tests, how- INOR-8 specimen exposed to the fuel-graphite sys- ever, indicated that carburization did occur. Carbon tem to have higher yield and tensile strengths and analyses of 3-mil surface cuts showed that the lower elongation than the INOR-8 specimen that INOR-8 specimen exposed to fuel 130 withoutadded was exposed only to argon and given the same heat graphite contained 640 ppm carbon, while a similar treatment. The same trends were apparent in me- cut of an INOR-8 piece exposed to the fuel-graphite chanical tests at 1250°F. Data obtained in the 60 PERIOD ENDING OCTOBER 31, 1958 ? - \J \ 3 < Q\ - \ B ¢ \ ¥ ° . I - - - e—ea- '.'- ' TN F - 2 il | i - 1 .002 ’ ' \ ] . - / . »’ -} . | . ' t ’ 003 5 » SRt R w - . e / i /. o [ 5 - - P @ 2 " -] - = . 'O Y. (0) e UNCLASSIFIED | © (b )" - . UNCLASSIFIED Y-26647 ’ | 004 " 2 Y-26646 Fig. 2.1.10, INOR-8 Specimens After 2000-hr Exposures at 1300°F to (a) Fuel 130 Containing a Graphite Rod and (b) to Fuel 130 Without Added Graphite, Etchant: glyceria regia. 1000X, Reduced 9%. Table 2,1.6. Carbon Analyses of INOR-8 Specimens and Fuel 130 Mixtures Used in Tests of the Carburization of INOR-8 in a Fuel-Graphite System at 1300° F for 2000 hr Specimen Analyzed Carbon Content (ppm) Before Test After Test Fuel 130 exposed to graphite and INOR-8 220 190 Fuel 130 exposed only to INOR-8 180 150 INOR-8 exposed to graphite and fuel 130 160-170" 980 INOR-8 exposed only to fuel 160-170* 640 *Carbon analysis of as-received INOR-8 from heat SP-16. room-temperature and high-temperature mechanical tests are presented in Table 2.1.7. As in the me- chanical tests of specimens carburized in a sodium- graphite system, the decrease in elongation is an indication that carburization occurred, The fact that the tensile strength of the INOR-8 specimen exposed to fuel 130 was greater than that of the specimen heat treated in argon can be explained by the decrease in carbon concentration of the fuel exposed to the INOR-8. The INOR-8 in this case probably acted as a sump and picked up carbon from the fuel, This would account for the decrease in the carbon content of the fuel and the increase in mechanical strength of the INOR-8 specimen ex- posed to it. UO, Precipitation in Fused Fluoride Salts in Contact with Graphite W. H, Cook A test of the compatibility of fuel 130 and graph- ite at 1300°F in @ vacuum (<0.1 i) was terminated at the end of a 500-hr period because of UO, pre- cipitation from the fuel. In preparation for the test, 61 MOLTEN-SALT REACTOR PROGRESS REPORT Table 2.1.7. Results of Mechanical Tests of INOR-8 Specimens Exposed to Fuel 130 With and Without Graphite and to Argon at 1300°F for 2000 hr Yield Strength (psi) Specimen Treatment Tensile Strength Elongation in 2-in, Gage (psi) (%) Room-Temperatute Tests* X 103 51.5 51.5 Exposed to fuel 130 and graphite 40.0 39.9 Exposed only to fuel 130 39.0 31.3 Exposed to argon Tests at 1250°F Exposed to fuel 130 and graphite *x Exposed only to fuel 130 b Exposed to argon ** % 103 116.8 47.0 116.4 46.5 107.3 56.0 106.5 56.0 104.2 55.0 103.4 56.0 73.0 19.5 66.5 30.0 66.6 33.8 *Values obtained in each of duplicate tests are listed. **Yield strengths were not obtained because the results might have been obscured by a notch effect of the extensometer, a machined CCN graphite crucible, nominally, 2 in. oD, 0.43 in. ID, 31/2 in. deep, and 4]/2 in. long, with an average bulk density of 1.90 g/cm 3, was de- gassed at 2370°F for 5 hr in a vacuum of 1 to 3 p, cooled to room temperature, and blanketed in argon. The crucible was exposed to room atmosphere for approximately 3 min in order to weigh it. During all other preparations for the test, the crucible was either sealed under an argon atmosphere or was under a vacuum of <0,1 . Sufficient fuel 130, 13.70 g, to fill the crucible to a depth of 3 in. at 1300°F was cast into the graphite. The loaded crucible was tested in an inert-arc welded Inconel container. Radiographs were used to monitor the location of the fuel in the graphite so that the test could be terminated if the fuel threatened to penetrate the crucible walls. Such a radiograph revealed at the end of 500 hr a Q.1-in.-high disk in the bottom of the graphite crucible that was relatively opaque to the x-rays {see Fig. 2.1.11). The remainder of the fuel was less opaque to the x-rays than it had been at the beginning of the test. Subsequent examina- tion of the disk revealed that it contained U02; the 62 disk contained ~40 wt % U as compared with 7.03 wt % U in the as-received fuel. The uranium con- tent in the disk amounted to 28% of the total amount of uranium originally in the fuel 130 charge; stated in another way, the uranium concentration of the fuel above the disk had decreased to 59% of its original value. Petrographic examinations of the fuel used for this test did not reveal oxide contamination. The fuel remaining in the batch from which the test quantity was taken is being analyzed for oxygen. This same source of fuel was used in a 2000-hr static test at 1300°F, described in the preceding section, in which the fuel was exposed only to INOR-8. A radiographic examination of the fuel at the end of that test did not reveal any precipitation of urantum. Different grades of graphite and different fuels are being used in further tests of the compatibility of fuels and graphite. Radiographic examinations of systems now being tested, in which TSF graphite (average bulk density, 1.67 g/cm3) is used to con- tain the fuel 130, have also revealed uranium pre- cipitation, but the quantities precipitated are less in TSF graphite than in CCN graphite. Radiographic PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED PHOTO 45742 B ———VACUUM (<04 p) — & (@) 3 2. 2 = = _J = VU] i g1 N 180,000 STRENGTH AS-WELDED AGED AT {200°F 166,000 =1 FOR 200 hr ' = 7] AGED AT 1200°F = FOR 500 hr ' AGED AT 1500°F 140,000 FOR 200 hr AGED AT 1500°F FOR 500 hr 120,000 s (o T 5 100,000 p=4 Ll o = w % 80,000 wn) = L - 60,000 |- 40,000 |- 20,000 EN x ’ e HASTELLOY B HASTELLOY W INOR-8 INCONEL Fig. 2.1.19. Room-Temperature Mechenical Properties of All-Weld-Metal Specimens. PERIOD ENDING OCTOBER 31, 1958 ductility upon aging, particularly at 1200°F (duc- tility reduced from 24 to 3% after aging for 500 hr). Hastelloy W weld metal also exhibits a marked re- duction in room-temperature ductility after aging. INOR-8, on the other hand, maintains its good duc- tility, Mechanical property studies of aged specimens were also conducted at the aging temperature in an effort to more completely determine the behavior of these materials at elevated temperatures. The data obtained at 1200°F are shown in Fig. 2.1.20, and the data obtained at 1500°F are shown in Fig. 2.1.21. Aging at 1200°F definitely reduces the ductility of Hastelloy B and Hastelloy W at 1200°F, whereas the ductility of INOR-8 is slightly im- proved. At 1500°F, a significant ductility increase UNCLASSIFIED ORNL —LR-DOWG 34523 100 80 J— R [ —_— — . p— . — _ J— - J— DUCTILITY ELONGATION IN tin. (%) o o T 160,000 STRENGTH AS-WELDED AGED AT 4200°F FOR 200 hr . AGED AT {200°F FOR 500 br 140,000 |~ 120,000 - L] 100,000 r 80,000 r 60,000 — TENSILE STRENGTH {psi) 40,000 |- 20,000 [ - HASTELLOY W | HASTELLOY B INOR-8 [INCONEL Fig. 2.1.20. Mechanical Properties of All-Weld-Metal Specimens at 1200°F, 69 MOLTEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED ORNL-LR-0WG 34524 .. 100 8 s 80 o oDucTILITY _ 2 60 = = e 40 |— — —— AR — <1 2 Z 20 p— —— —] - BN o ‘ 80,000 AS-WELDED AGED AT 1500°F % 600 ] FOR 200 hr 2 60,000 AGED AT 1500°F he FOR 500 hr &) S =4 SN & o x 40,000 f— W wl ] & 20,000 | —F [ s ?:: N HASTELLOY B HASTELLOY W INCR-8 [INCONEL Fig. 2.1.21. Mechanical Properties of All-Weld-Metal Specimens at 1500°F. is evident for INOR-8, probably as a result of partial spheroidization of the eutectic carbides. The influence of carbide spheroidization is being studied further, and high-temperature stabilization treatments of weld metal are under investigation. Mechanical property studies of INOR-8 welded joints, both in the as-welded and aged conditions, are also being conducted, Filler Metal Modification Studies. ~ Since the ductility exhibited by INOR-8 weld metal at tem- peratures of 1300°F and higher is marginal, a pro- gram has been initiated to determine the factors that influence ductility and to develop procedures for improving it. Three heats of INOR-8weld metal have all shown approximately the same ductility, that is, 10% or less in a 1-in. gage length in the 1300 to 1500°F temperature range. Hastelloy W, however, has a minimum elongation of 20% in this range. Personnel of the International Nickel Company have indicated that they think the trouble probably results from trace elements in the 0,009 to 0.0001 wt % range, The use of proper addition agents during the melting of the ingot and the proper timing of the additions are extremely important in removing or neutralizing these trace elements. Accordingly, 70 melts containing recommended quantities of manga- ese, silicon, aluminum, titanium, magnesium, and boron have been made. The ingots are being fabri- cated into filler wire for the preparation of all-weld- metal specimens for room- and elevated-temperature testing. Heats of filler metal have also been made that contain additions of a commercially available aluminum-titanium-nickel master alloy. One experiment was conducted which verifies the premise that minor modifications in filler metal composition can result in major changes in elevated- temperature mechanical properties. A special ingot of INOR-8 filler metal was prepared in which most of the nickel and all the chromium were added as Inconel inert-arc welding wire. This Inconel wire, designated Inco No. 62 (efongation at 1500°F of 70%), also contains 2% columbium to prevent hot cracking in the weld deposit., The mechanical properties of this modified INOR-8 weld metal in the 1200 to 1500°F temperature range were superior to those exhibited by conventional INOR-8, as is shown in Table 2.1.12. As was previously mentioned, the improvement in ductility of INOR-8 weld metal upon aging appeared to be associated with the spheroidization of eutectic carbides. Therefore special heats of INOR-8 having no carbon and 0.03% carbon have been prepared and will be fabricated into filler wire as a means of de- termining the welding characteristics and mechani- cal properties of lower-carbon-content weld deposits. Results of Examination of INOR«8 Forceds«Circus lation Loop Weld Failure, — An INOR-8 heater lug section of forced-circulation loop 9354-1 failed during service. The component and the location of the failure are shown in Fig. 2.1.22. Metallographic examinations of the failed area and adjacent areas revealed profuse cracks in the INOR-8 adapter near the lug-to-adapter weld, with the cracks resembling those frequently observed in the heat-affected zones of the crack-sensitive heat SP-16 material.!! Chemical analyses of the adapter material re- vealed it to be a high-boron, low-carbon heat of INOR-8 and confirmed the suspicion that heat SP-16 material had been inadvertently substituted for the recommended heats of INOR-8. It is thought that the cracks initiated during welding and propa- gated to failure during service. Welding of INOR-8 to Other Metals. — An evalu- ation of Inco-Weld *A’’ wire was made to determine ”P. Patriarca and G. M. Slaughter, MSR Quar. Prog. Rep. Oct. 31, 1957, ORNL-2431, p 18. PERIOD ENDING OCTOBER 31, 1958 Toble 2,1.12, Comparison of Mechanical Properties of INOR-8 and Modified INOR-8 in As-Welded Condition Test Temperature | \ \ | ' Tensile Strength Elongation in 1-in, i (OF) Fl“ef Meful (Psi) (%) Room INOR-8 116,000 38 - Modified INOR-8 113,000 44 1200 INOR-8 73,000 18 Modified INOR-8 77,000 30 1300 INOR-8 64,000 10 Modified INOR-8 72,000 22 1500 INOR-8 53,000 6 | Modified INOR-8 59,000 15 UNCLASSIFIED Y-26847 Fig. 2.1.22. INOR-8 Heater Lug Section of Forced-Circulation Loop 9354-1 That Failed During Service. Arrow indicates point of failure. its suitability as a filler material for weldments with INOR-8. The study'? indicated that the as- welded ductility of the metal in short-time tests is satisfactory up to 1300°F but becomes marginal at : 1500°F. Significant impairment of the room- and elevated-temperature ductility was found after aging at 1200°F for 500 hr. 125, M. Slaughter, Mechanical Property Studies on INCO-Weld **A’* Wire Filler Metal, ORNL CF-58-10-30 (Nov, 5, 1958). Brazing of Thick Tube Sheets for Heat Exchangers G. M. Slaughter The construction of heat exchangers for molten- salt power reactor applications poses a variety of problems. For example, it is probable that the tem- perature and pressure conditions in some of the heat exchangers will necessitate the use of relatively thick tube sheets (thicknesses of 4 to 6 in. or greater). Present practice and fabrication proce- dures for constructing high-integrity tube-to-tube 71 MOLTEN-SALT REACTOR PROGRESS REPORT sheet joints in sections of this magnitude may not Inconel tube sheet with Coast Metals No. 52 alloy. be adequate for a molten-salt environment. The un- Brazing was performed at 1920°F with a 1 hr hold avoidable notch obtained in the conventional welded at temperature. A 600°F per hr rate of temperature tube-to-header joint is undesirable in several ways. rise was used in this experiment, and the sample - First, it may act as a starting point for crack propa- was furnace-cooled to room temperature. The weld ‘ gation in the welds during cyclic service. Also, the side of the tube sheet was also trepanned to mini- | notch may act as a crevice and thereby increase the mize stresses during welding, As may be seen in " possibility of stress-corrosion cracking. It is evi- Fig. 2.1.24, a photograph of the brazed specimen, dent that the technique of back-brazing of the tube- excellent filleting was obtained. to-header joints would remedy these conditions by eliminating the notch. Brazing would also improve heat transfer and decrease the possibility of a leak through a defective tube-to-header weld, since the leak-path would be significantly increased. Although considerable experience exists inbrozing small diameter tubes into relatively thin tube sheets, little is known of the problems associated with adapting these techniques to heavy sections. Ac- cordingly, a preliminary experiment has been con- ducted to determine the feasibility of back-brazing joints in thick tube sheets. A brazing procedure was utilized which had proved useful in fabricating simulated components containing ]]/Q-in.-thick tube UNCLASSIFIED Y-26383 | sheets. As may be seen in Fig. 2.1.23, the brazing alloy is preplaced in annular trepans in the tube sheet and is fed to the joint through three small feeder holes. Melting of the brazing alloy therefore does not take place until the joint has attained the tem- perature where wetting can occur, In an effort to extend this technique, a speci- men was prepared that consisted of a 5/B-in.-OD, 0.065-in.-wall Inconel tube brazed into a 5-in.-thick ‘ UNCLASSIFIED ORNL-LR-DWG 27066R _—~WELD SIDE — R e > TUBE SHEET E | E» I ', | : 3 HOLES, 0.062-in DIA, f } } AT 120 DEG—>__. . 1/ A ZN%) i PREPLACED BRAZING ALLOY” [ = TUBING ’ 7 Fig. 2,1.23, Alloy Preplacement Procedure for Fig. 2.1.24. Brazed Tube-to-Header Specimen Showing Brazing Thick Tube Sheets. Excellent Filleting. 72 Flow of the alloy along the joint was evaluated from metallographic examination of numerous sec- tions of the sample. A photomicrograph of the fillet on the braze side of the tube sheet is shown in Fig. 2.1.25, while the joint on the weld side of the tube sheet is shown in Fig. 2.1.26. On the basis of these observations it can be concluded that flow was probably continuous along the entire 5 in. of the joint and around the entire 360 deg of the cir- cumference. Also no cracking of the brazing alloy was noted. In view of the promising results of this experi- ment, a more comprehensive evaluation program is planned. Two 6-in.-dia, 5-in.-thick tube sheets containing five tubes are presently being machined to determine the influence of section size on braze- metal preplacement procedure, |t is probable that these experiments may suggest modifications in joint design and brazing procedure. In addition, other factors must also be studied if these experi- ments show promise. For example, the minimum rate-of-rise to brazing temperature to achieve satis- factory flow during brazing of heavy tube sheets will be affected by the choice of materials and the brazing atmosphere. This rate should be determined for the base material and brazing alloy compositions Fig. 2.1.25. Brazed Side of Tube Sheet Showing Fillet. PERIOD ENDING OCTOBER 31, 1958 of interest, Further, the influence of time and tem- perature on the diffusion of microconstituents from the brazing alloy into the tube walls and to a lesser extent into the tube sheet must be determined. Since it is reasonable to expect that the base metal me- chanical properties will be affected, the nature of this effect must be known if back-brazing is to be utilized, Fabrication of Pump Components G. M. Slaughter The problems associated with joining a mo- lybdenum extension to an Inconel shaft for a pump application are being investigated. The widely different coefficients of thermal expansion (that is, molybdenum, 2.7 x 10~¢ in./in./°F; Inconel, 6.4 x 10~% in./in./°F) present a condition of very high stresses during service. One method of at- tachment proposed by the Pump Development Group was to utilize molybdenum fingers brazed to the Inconel shaft around the shaft periphery. The amount of stress built up in each brazed lap joint would thus be influenced markedly by the size of the molybdenum fingers. In order to test this proposed method, several mo- lybdenum-to-Inconel brazed joints were fabricated UNCLASSIFIED Y-26634 vid w4 . K 4 LA R et g 4 B I ey P BT s e S Wy e P Etchant: glyceria regia, 20X. 73 MOLTEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED Y-26633 Fig. 2.1.26. Weld Side of Tube Sheet Showing Complete Flow to the Root of the Weld. Etchant: glycerio regia. 20X. and are being examined metallographically for cracking after brazing and after thermal cycling. The samples are ]8 mole % BeF, into solid solution and 3LiF+«ThF can take as much as 11 mole % BeF,. Moreover, the introduction of BeF, into the structure of 3LiF+«ThF, apparently causes enough distortion of the lattice for some 7LiF-6ThF4 also to be taken into solid so- [ution. Thus a roughly triangular area of homo- geneous solid solution exists with one apex located on the LiF-ThF, boundary at the composition 3LiF«ThF, and the other two apices at about the compositions 67 mole % LiF-11 mole % BeF ,-22 mole % ThF, and 60 mole % LiF-13 mole % BeF,— 27 mole % ThF ,. Detailed quenching tests have established that there is practically no solid so- lution of 7LiF«6ThF , in 3LiF«ThF in the binary system LiF-ThF4. The nature of the solid solutions in the ternary system is not known, but obviously such solid solutions are not of the substitutional type. With the increases of 7LiF:6ThF , in 3LiF+ThF, in the ternary system, the refractive indices of 3LiF-ThF4 are raised somewhat and the birefringence is lowered. The lines of the x-ray powder-diffraction pattern also shift with increasing solid solution. The compound 2LiF:BeF, appar- ently exists only as the pure compound in the ternary system, and there is no evidence that it enters into solid solution with any of the LiF-ThF, compounds. Further gradient-quenching studies have required the following modifications in the preliminary dia- gram of the LiF-Ber-ThF4 system. (1) The primary phase 7LiF-6ThF4 does not have a common boundary curve with the 2LiFThe construction of these models was a part of the assignment to C. S, Johnson, MIT summer participant, 1958, éC. J. Barton and R. A, Strehlow, MSR Quar, Prog. Rep. June 30, 1958, ORNL-2551, p 84. 80 T. Ward et al.,” in which 5B uF vhs (solubility) of F’uF3 in the solvent at a specified temperature, as shown by the top line in Fig. 2.3.6 (labeled *‘PuF; only™’), and Nqua(ss) is the mole fraction of PuF, in the solute mixture. Two of the four data points obtained with a 1:1 molar ratio of CeF:3 to PUF3 added to the LiF-BeF2 (63-37 mole %) mixture agree quite well with the theoretical curve, while four of the five points obtained with a 5:1 molar ratio of CeF ; to PuF, lie reasonably close to the theoretical curve. The reason for the divergence of the results in the other three experiments is not clear. Similar plots of CeF, solubility values based on preliminary analytical data show even poorer agree- ment with theoretical curves obtained with the use of data reported by W. T. Ward® for the solubility of Cel:3 in LiF-BeF2 (62.4-37.6 mole %). Analytical methods for determining cerium in the presence of plutonium in mixtures of this type are being studied and it is hoped that a better correlation between theoretical and experimental values will be obtained with later analytical results. It seems reasonably safe to assume that Eq. 1 holds well enough for mixtures of this type. Data in Fig. 2.3.6 show that the solubility of PuF, in LiF-BeF, (63-37 mole %) can be redyced from 0.15 to 0.026 mole % at 500°C by dissolving 0.75 mole % CeF , in the mixture at a higher temper- ature before cooling to 500°C. The calculated con- centration of CeF ; remaining in solution at 500°C is 0.14 mole %. By going to a lower temperature (the liquidus temperature for the solvent is about 458°C) or a higher cerium-to-plutonium ratio, the recovery of plutonium from used fuel could undoubt- edly be improved, but it would be difficult to recover more than 90% of the plutonium by a single operation of this type. By reheating the filtrate and adding more CeF, to the mixture, followed by a second cooling and filtration operation, further improvement in the efficiency of plutonium recovery could be cbtained. The data in Fig. 2.3.6 seem therefore to provide the basis for a suitable reprocessing scheme for fused-salt mixtures containing plutonium. Concen- tration of the plutonium by precipitation would is the mole fraction "W, T.Ward et al., Solubility Relations Among Some Fission Product Fluorides in NaF'-ZrF'4- UF'4 (50-46-4 mole %), ORNL-2421 (Jan. 15, 1958). 8w, 1. Ward, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 90. PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED PHOTO 32550 ThF, 1111°C + LIQUID LiF-4ThF4—LiF-4UF4 SOLID SOLUTION + LIQUID PR Y S P LiF-2ThF4 + LIQUIDE 3 LiF-ThF, + LIQUID — e ! g PR .6 ThF,—7LiF-6UF, ‘ SOLID SOLUTION + LIQUID . | + 400 TUF-6ug, u'+7[jf.‘u“ .5! 7[5‘-‘"; §' = A : 300 - + = S UF-4ug, UF-4UE, + ug, N ~_ Laanabo bl Vbl s, cul i __,_‘.J Fig. 2.3.4. Model of the System LiF-ThFa-UF“. 81 MOLTEN-SALT REACTOR PROGRESS REPORT LiF; 7LiF-6ThF, 845°C + LIQUID B 3 LiF ThF, . + LIQUID 70 600 < N g = g4 ThFg + LIQUID LiF-4ThF, + LIQUID 3\ s < LiF-2ThF, B + LIQUID § Fig. 2.3.5. Model of the System LiF-Ber-Th FA' 82 ThFg, 1111°C UNCLASSIFIED PHOTO 32551 UNCLASSIFIED ORNL—-LR-DWG 329414 TEMPERATURE {°C) 650 600 550 500 | I [ I ! T 4:4/\ - EXPERIMENTAL POINTS N— 0 —— 1] PuFz ONLY IN SALT . %: SOLUBILITY OF PuFy IN LiF -BeF, (63-37 mole o) 0.05 [— e - ® PuFy + ThF, IN SALT \\ & PuFy + BaFp IN SALT 4 4 1:1 MOLAR RATIO OF CeF3 TO \\ PuF3 IN SALT 0.02 I'g 5:4 MOLAR RATIO OF Cefy TO PuF3 IN SALT } | 0.0 ‘ W : ‘ ‘ 100 105 M0 M5 120 125 130 135 10%/7 (°K) Fig. 2.3.6. Effect of Other Fluorides on Solubility of PuF, in LiF-BeF, (63-37 mole %). simplify subsequent operations required to purify the plutonium, possibly by fluorination. The effect of trivalent fission products on the solubility of PuF, is large enough so that they could not be allowed to build up to high concentrations, and some of them probably should be removed anyway to reduce nuclear poisoning. Comparison of the PuF; solubility data for mix- tures containing ThF, and BaF, with the values ob- tained in the absence of additives, as shown in Fig. 2.3.6, seems to indicate that ThF, may reduce the solubility of PuF,; slightly at the lower end of the temperature range investigated, while the solu- bility in BaF ,-containing mixtures is about the same as that in the ThF ,-containing solvent at about 500°C and significantly less at 650°C. In neither case is there sufficient data to be con- clusive and additional experiments are planned, but the very odd behavior of the solubility curve for 3aF ,-containing mixtures may indicate that iso- therms in the psuedo-ternary system BaF ,-PuF ,- PERIOD ENDING OCTOBER 31, 1958 solvent are not linear, although they apparently are linear for the system CeF ;-PuF ;-solvent and for a number of similar systems investigated by Ward et al.” Phase behavior in the system BaF -PuF, has apparently not been investigated, but it is very likely similar to that in the system B 0 Q 7] v 1»/500"(: / 4 0 0 0.5 10 1.5 20 25 PRESSURE (atm) Fig. 2.3.7. Solubility of Argon in Molten LiF-BeF, (64-36 mole %) at Yarious Temperatures. tension of the solvent is presented in a following section of this chapter that is titled ‘‘A Simple Method for the Estimation of Noble Gas Solubilities in Molten Fluoride Mixtures.”’ Comparisons of the solubilities of helium'? and of argon in LiF-BeF, (64-36 mole %) with their solu- bilities in other solvents previously studied have indicated a number of similarities and one differ- ence. The similarities include (1) conformity to Henry’s law, (2) increase of solubility with in- creasing temperature, (3) decrease of solubility with increasing atomic weight of gas, and (4) increase in the enthalpy of solution with increasing atomic weight of gas. The difference is the value of the entropy of solution in the LiF-BeF, mixture. While all the entropies of solution for helium, neon, argon, and xenon in Nc:F-ZrF‘1 (53-47 mole %) were quite small, about —1 eu, the entropy values for helium and argon are —3.73 and —3.95, respectively, in the LiF—BeF2 (64-36 mole %) mixture. The significance of this difference is not immediately apparent, but the difference seems to be related to the nature of the solvent. PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED ORNL—LR—DWG 34533 TEMPERATURE (°C) -9 (X'BOO ) 8OO 700 600 500 ] I I T e ,<1,7 - 5 4\7}\ —~ARGON IN NgF-KF~LiF (REF1{3) | AH=12,400 cal /mole AS= —0.10eu | - ’E ‘ . .2 | ; 2 AN ] O E ™, ‘ ! - 20 e ———. ——— 8 | | £ \/ _~ARGON IN LiF -Bef, B 10 \ _ AH= 8850 ccl/moie _ n ! Oy £ -+ - — AS=—395eu - ) | A\ ‘ T 3 _ I R oS . L : \____ I = [ L o 5 i _ 5 - G E —— — —_t I i —— ] e i E NN X \ 2L L « 2.09% \ | | | 8 9 10 1 12 13 14 15 10%/T (°K) Fig. 2.3.8. Solubility of Argon in LiF-BeF, (64-36 mole %) and in NaF-KF-LiF (11.5-42-46.5 mole %). Solubility of HF in LiF-BeF , Mixtures J. H. Shaffer The solubility of HF in LiF-BeF, mixtures was determined as a function of temperature, pressure, and solvent composition. The composition was varied from approximately 10 to 50 mole % BeF,, at pressures from O to 3 atm, and at temperatures from 500 to 950°C, depending on the liquidus temper- atures of the mixtures.'* Within the precision of the measurements, the solubility of HF was found to follow Henry’s law over the pressure range studied. Accordingly, at a given temperature and solvent composition, the solubility of HF can be measured at several saturating pressures and the results can MComposi're diagram compiled by R. E. Thoma, ORNL, from D. M. Roy, R, Roy, and E. F. Osborn, J. Am. Ceram. Soc. 37, 300 (1954); A. V. Novoselova, Y. P. Simanov, and E. |, Yarembash, Zbur. Fiz. Khim. 26, 1244 (1952). 85 MOLTEN-SALT REACTOR PROGRESS REPORT be averaged as a Henry’s law constant character- istic of that temperature and composition. The average constants obtained experimentally are pre- sented in Fig. 2.3.9. For each point, at least three constants were obtained that agreed with each other to within £5%. The averages of the 800 and 900°C data at 31 mole % BeF2 are believed to be low, possibly because of consistently reproducible errors in the temperature measurement as a result of a de- fective thermocouple. The values obtained at this composition will be rechecked before proceeding -6 UNCLASSIFIED (Xx10 ) ORNL—LR-DWG 34534 30 | l ———EXTRAPOLATED o o 20 - a —— (1] — e =] e § 15 e B ““‘- 6OO°C E " ——— '\ 70 S 10 ke Ooc ‘xot [=] K S — ! \ N £ . l-._____\ \ 500 T~ [ » ° & . \\\C \.L\ (TR 3 E -\OOOC\D\ 5 » \ v % 5 \. E \ x 4 3 G 10 20 30 40 5C LiF BeF, (mole %) Fig. 2.3.9. Composition Dependence of Henry's Law Constants for the Solubility of HF in LiF-BeF2 Mixtures. with a corresponding investigation of another sol- vent system. For purposes of comparison, Henry’s law con- stants expressed at rounded values of composition are more useful, and therefore values taken from a large-scale plot of the data are summarized in Table 2.3.2. The enthalpies and entropies of solution of HF as functions of composition of the solvent are listed in Table 2.3.3. It may be seen that in LiF-BeF, solvents, the solubility of HF does not show a very strong com- position dependence, such as it did in NaF-ZrF, mixtures, as indicated previously. 'S The weaker 15, H. Shaffer, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 93. Table 2.3,.3. Enthalpies and Entropies of Solution of HF in LiF-BeF2 Mixtures Solvent Enthalpy Entropy Composition of Solution of Solution* (mole % LiF) (cal/mole) (ev) 100 -4700 -4.4 90 -4900 -4.8 80 -5100 -5.2 70 -5400 -5.8 60 -5400 -6.2 ' 50 —-5400 —6.6 *Entropy of solution calculated at equal concentra- tions in the gas and liquid phases at 800°C, Table 2,3.2, Henry’s Law Constants for the Solubility of HF in LiF-BeF, Mixtures Solvent K (moles of HF per t:m3 of melt per atmosphere) Composition (mole % LiF) At 600°C At 700°C At 800°C At 900°C x 1078 x 1078 x 1078 x 1070 100 18.2* 13.7+ 1.0+ 9.0 90 17.3%* 130+~ 10.3+* 8.4 80 16,0+ 1.8 9.1 7.6 70 13.9 10.1 7.7 6.3 60 11.2 8.2 6.4 5.2 50 8.6 6.6 5.2 4.2 *Graphical extrapolation to 0 mole % BeF2 (Fig. 2.3.9). **Extrapolated from measurements at higher temperatures, 86 composition dependence may be a result of much less stable acid fluorides in the LiF system'® than in the NaF system. However, this effect should be more precisely demonstrated when measurements using NaF-BeF, mixtures are carried out. Solubilities of Fission-Product Fluorides in Molten Alkali Fluoride—Beryllium Fluoride Sol vents W. T. Ward Solubility of CeF in NaF-LiF-BeF,. — The solu- bility of CeF, in NaF-LiF-BeF, solvents of various compositions was determined over the temperature range 400 to 750°C. The mixtures investigated con- sisted of the NaF-LiF eutectic (40-60 mole %) to which BeF, was added in amounts that varied from 0 to 56 mole %. The results of the experiments are shown in Fig. 2.3.10, in which the solubility of wH. J. Emeleus in Fluorine Chemistry, Yol. 1, p 25, J. H. Simons (ed.), Academic Press Inc,, N. Y. (1950). UNCLASSIFIED ORNL-LR-DWG 34535 © | B — — —_— I__ — 4 I I | s LAY GREATER THAN 4.5 \ \ | moie 7, AT 650°C \ 2 e \ - \ - 750°C o / '. L ] 1 \ o_Ye_700°C \ A\ ® » -~ 08 ANEAN l yd : N ovore a2 %’ 0.6 \\ \ / E . E 04 ¥. 6000% P_: ~— — c s = o 0.2 o . w© i \e” )/ [ ] 0.4 / 0.08 .// 0.06 // 0.04 e 0.02 0 10 20 30 40 50 60 t00% NoF - LiF Bef, IN SOLVENT {mole T} EUTECTIC Fig. 2.3.10. Solubility of CeF3 in Nt:F-LiF-BeF2 Mixtures. PERIOD ENDING OCTOBER 31,1958 CeF, (mole %) is plotted against solvent compo- sition at several temperatures. A comparison of the solubility of CeF , in this solvent with the solu- bilities of CeF, in the LiF-BeF, and NaF-BeF, binary systems is presented in Fig. 2.3.11 for the single temperature of 600°C. UNCLASSIFIED 2 ORNL-LR-DWG 34536 Al \\ o & . 1 [ VO 2 \ 2 = 08 : WY /4 = — \ - P 4 z \ a4 5 06 NN LFBeR o % o \\ b b E\/ g —e O L N 7 [0 04 © NuFuF~BeF2__,\€\_1/ \G'z\NcF—BeFZ 0.2 o 10 20 30 40 50 60 BeF, IN SOLVENT (mole Ta) Fig. 2.3.11, Comparison of CeF, Solubilities in Three Solvent Systems at 600°C. Solubility of CeF, LaF,, and SmF, in LiF-BeF ,- UF4. ~ The solubilities of CeF,, LoFa, and SmF3 in LiF-BeF ,-UF, (62.8-36.4-0.8 mole %)} were de- termined over the temperature range 500 to 750°C. The data are given in Table 2.3.4, and, for com- parison, are plotted against the reciprocal of the temperature in Fig. 2.3.12, As may be seen, the solubility of CeF, in this solvent is a little higher than in LiF-BeF, (63-37 mole %) over the full tem- perature range investigated. Simultaneous Solubilities of CeF, and LaF, in LiF.BeF -UF ,. — The solubilities of CeF; and LaF, in the presence of each other in LiF-BeF ,- UF4 (62.8-36.4-0.8 mole %) were determined with the use of two radioactive tracers in the manner de- scribed previously.!” The values obtained are listed in Table 2.3.5 and plotted in Fig. 2.3.13, in which the sums of the solubilities may be seen to be intermediate between the individual solubilities. The lines represent the solubilities of CeF, and 7W. T. Ward et al., Solubility Relations Among Some Fissiop Product Fluorides in NaF'-ZrF4 mole %), ORNL-2421 {(Jan. 15, 1958). -UF ; (50-46-4 MOL TEN-SALT REACTOR PROGRESS REPORT Table 2.3.4. Solubilities of CeF3, LuF3, and SmF3 in LiF-Beer-UF4 (62.8-36.4-0.8 mole %) Temperature CeF3 in Filtrate Temperature La F3 in Filtrate Temperature SmF3 in Filtrate (°C) Wt % Mole % (°C) Wt % Mole % (°C) Wt % Mole % 764 >9,1* >1.8* 741 8.65 1.71 757 >10.22*% >1.93* 661 5.1 0.97 676 >10.27* >1.94* 652 4,5 0.85 655 4.19 0.79 578 2.35 0.44 560 1.73 0.32 587 5.35 0.97 499 1.04 0.19 480 0.79 0.15 495 2.20 0.39 *Solution not saturated, UNCLASSIFIED UNCLASSIFIED ORNL-LR-DWG 34537 . ORNL-LR-DWG 34538 2 \ * CALCULATED TOTAL RARE—-EARTH FLUQRIDE PRESENT 0\ . e 182 mole % Cefs, 0.78 mole % LaF3 2 2\ & 1.85 mole % Cefy, 2.27 mole % LaFy _ \ C 1.83 mole % Cefs;3.30 mole % LaFy SmF o o ___.___E ] — \ s e i % AR * £ E \L:N N\ & = 0.8 *\ N E w AN \ E g L £ 06 NN ; E \N\eer, -\ z I E L w = ) \\\ 5 £ 04 = [} \ . N\ z = % w 0 LoF3 W CeF, ONLY w < & : ] o > Y = \ '_ \‘\ I I I N LY ot 9 {e) 11 12 13 14 15 (oK ] 9 10 11 12 13 14 104/ T{°K) 10Y/7 (o) Fig. 2.3.12, Solubilities of CeF,, LaF,, and SmF, in LiF-Ber-UF4 (62.8-36.4-0.8 mole %). LaF ; individually (same curves as those shown in Fig. 2.3.12), while the points represent the sums of the solubilities obtained when both are present in the proportions indicated. It may be observed from the data given in Table 2.3.5 that the solubility of a rare earth fluoride is decreased by the addition of another rare earth fluo- ride. This is believed to be the result of the for- mation of a solid solution by the rare earth fluorides 88 Fig., 2.3.13, Comparison of Solubilities of CeF, and LoF3 Individually and When Both are Present in LiF- Be FZ-U F4 (62.8'36.4‘0.8 mo'e %)o involved. |t was shown previously that the equi- librium quotient of the reaction LaF, (d) + CeF, (ss) &= LaF (ss) + CeF, (d) is given by the expression SO CeF3 PERIOD ENDING OCTOBER 31, 1958 where S° is the mole fraction (solubility) of a rare lated constants. These results are also illustrated earth fluoride in the solvent at the specified tem- in Fig. 2.3.14. perature and in the absence of a second rare earth fluoride. A comparison of the calculated and the Simultaneous Solubilities of CeF; and SmF in experimentally determined equilibrium quotients is LiF-BeF ,-UF,. — One experiment was run in which given in Table 2.3.6. It may be seen that, regard- both CeF; and SmF, were added to LiF-BeF,-UF , less of rather large composition changes, the equi- (62.8-36.4-0.8 mole %). There was evidence that librium quotients, or partition coefficients, remain this particular batch of SmF; was slower in dis- fairly constant and in fair agreement with the calcu- solving than usual, and the individual solubilities Toble 2.3.5. Solubilities of CeF, and L0F3 Separately and in the Presence of Each Other in LiF-Ber--UF4 (62.8-36.4-0.8 Mole %) Constituents of System Rare Earth Flucride in Filtrate (mole %) Temperature (mole %) €c) CeF3 L0F3 Solvent CeF3 L<:|F3 Total 180 0 98.20 700 1.36 0 1.36 600 0.55 0 0.55 500 0.193 0 0.193 1.82 0.78 97.40 700 0.890 0.387 1.28 600 0.375 0.141 0.52 500 0.148 0.043 0.191 1.85 2.27 95.88 700 0.588 0.525 1.1 600 0.271 0.195 0.47 500 0.120 0.067 0.187 1.83 3.30 94.87 700 0.486 0.775 0.126 600 0.220 0.296 0.52 500 0.093 0.102 0.195 0 2.20 97.80 700 0 1.19 1.19 600 0 0.475 0.48 500 0 0.177 0.177 Table 2.3.6. Solid-Solvent Extraction Coefficients for LaFg-Cng in LiF--Ber-UF4 (62.8-36.4-0.8 Mole %) Experimental KN* Temperature Calculated Ky (OC) © /5° For 1.82 Mole % CeF3 + For 1.85 Mole % CeF3 + For 1.83 Mole % CeF3 + CeFgy = LaF, 0.78 Mole % LaF , 2.27 Mole % LaF 3.30 Mole % LaF, 700 1.14 0.97 1.55 1.18 600 1.15 1.18 1.82 1.38 500 1.09 1.52 2.28 1.68 NLaFS(ss) NCeF3(d) CeF 4(s2) NLaFB(d) 89 MOLTEN-SALT REACTOR PROGRESS REPORT SOLVENT UNCLASSIFIED ORNL-LR—DWG 34539 & 3 / o . é AN\ &m s Nooe N OVER-ALL 3 COMPOSITION \\ / / \\ / \ / \\ ///' \ ) ’ A . 4q / S ovER-ALL S\ / \\ COMPOSrTlON _ \ / / \\ // \ . OVER-ALL COMPOSITION “ Fig. 2.3.14. Portion of Psvedo-Ternary System Ce F3~ La F3-So|venf Where the Solvent is LiF-Ber-UF4 (62.8'36.4'0.8 mo'e %)- obtained were considered to be unreliable. How- ever, the sums of the solubilities were intermediate between the solubility curves obtained when only CeF3 and only SmF3 were present, as shown in Fig. 2.3.15. A Simple Method for the Estimation of Noble Gas Solubilities in Molten Fluoride Mixtures M. Blander G. M, Watson The solubilities of noble gases in liquids may be interpreted with the use of a model similar to that described by Uhlig. 817 Although the model is elementary, it yields an interesting correlation with experimental data. If the gas does not interact with the liquid in which it is dissolved, the free-energy change upon solution is related to the ‘‘surface’” energy of the hole created by the gas molecule or atom. The solubility of such a gas in a continuous fluid medium can be shown to be given by the re- lation A d -18.08r%y . /RT mic K T —— = @ ©c C g 8p 4, Emmett, private communication. "9H. H. Uhlig, J. Phys, Chem. 41, 1215 (1937). 90 UNCLASSIFIED ORNL- LR-DWG 34540 3 [ ] 2 g . SmF5 ONLY B \ [} E 1 = N\ N\ 5 o8 N\ AN o X H N\ N\ z \ N\ = 2 06 \ . \ (& ] T @} o = oa \ z N 2 L \CeF3ONLY Ll o < . o | 2002 : N ° \ CALCULATED TOTAL RARE-EARTH — FLUORIDE PRESENT . 2.83 mole % C,eF3 2.22 mole % SmFs 0.4 9 10 1 12 13 14 10% 7 (°K) Fig. 2.3.15. Solubilities of CeF3 and SmF3 Indi- vidually and When Both are Present in Li |:-Be|:2-U|:4 (62.8‘36-4‘0-8 mole %)l where K _is the equilibrium ratio of the concen- trations in the liquid (Cd) and in the gas phase (Cg), 7 is the radius of the gas atom (in ‘Angstroms), Y .ic iS the microscopic surface tension of the sol- vent (in ergs/cmz), R is the gas constant (in cal/moles®K), and the number 18.08 is the resultant of the factors (6.02 x 1023) (477 x 10~ 1¢) (2.39 x 10‘8), which are used in the calculation of the sur- face free energy (in cal/mole of a mole of spherical holes of radius r (A} in a continuous liquid of sur- face free energy y (e (ergs/cm?). Although a real liquid cannot be considered a continuous fluid, it is interesting to speculate that it behaves as if it were one and that y face tension Y., The area of the holes is at least as large as the surface area of the spherical gas molecules, and because of thermal motions the hole would be ex- pected to be larger than the gas atom. The radii of ;. is related to the macroscopic sur- PERIOD ENDING OCTOBER 31, 1958 the noble gas atoms in the solid will be a lower Atomic_Radius limit of the radius of the hole, and the solubility (A) calculated on this basis would be expected to be ‘ . high. Values of K_ are listed in Table 2.3.7 that He lium 1.22 were calculated for the noble gases by assuming Neon 1.52 thaty . =y_._and by using the radii of the noble Argon 1.93 - gas atoms in the solid. The numerical values used Xenon 2,18 for the surface tensions of the solvents are given in Table 2.3.8, and the radii of the noble gases, as given by Goldschmidt, 20 are listed below: The comparison in Table 2.3.7 of the calculated and measured values of K_ for solutions of noble gases in NaF-KF-LiF and in NaF-ZrF, shows good agreement in view of the elementary nature of the 20y, M. Goldschmidt, Geochemistry, Oxford University mo.dt'al. -I_-he n-mgnn‘ude and the variation of the sol- Press (1954). ubility with size of the gas molecule are correctly Table 2.3.7. Comparison of Calculated and Observed Values of Henry’s Law Constants, K= Cd/cg’ for Noble Gas Solubilities in Molten Fluoride Mixtures Sol c Temperature Kc Ratio of Experimental to olvent as (°C) Experimental Calculated Calculated K. Values x 1073 x 1073 NaF-ZrF, Helium 600 15.5 137 0.11 700 23.3 188 0.12 800 37.0 243 0.15 Neon 600 8.09 45,9 0.18 } 700 14.7 74.9 0.20 800 21.7 112 0.20 Argon 600 3.62 7.33 0.49 . 700 6.44 16.0 0.40 800 10.6 30.2 0.35 Xenon 600 1.39 1.77 0.79 700 2,84 4.84 0.59 800 5.56 10.7 0.52 NoF-KF-LiF He lium 600 8.09 (28.3)* (0.29) 700 14.0 (46.8) {0.30) 800 20.3 (70.7) (0.29) Neon 600 3.12 (3.94) (0.79) 700 6.00 (8.63) (0.70) 3 800 9.84 (16.4) (0.60) Argon 600 0.645 (0,146) (4.4) 700 1.43 (0.509) (2.8) ) 800 2.99 (1.41) (2.1) Xenon 600 (0.011) 700 {0.057) 800 (0.212) *Values enclosed in parentheses are based on estimated surface tensions. MOLTEN-SALT REACTOR PROGRESS REPORT Surface Tensions of Molten Fluoride Table 2.3.8. Mixtures Surface Tension (dynes/cm) Temperoture (°C) NaF-KF-LiF* NaF-Zr F4** (11.5-42-46.5 mole %) (53-47 mole %) 600 230 128 700 220 120 800 210 112 *Surface tension values estimated from extrapolated values of surface tensions of pure components [F. M. Jager, |. fir anorg. chemie 101 (1917)] assuming addi- tive surface tension. Surface tension measurements of this mixture have not as yet been made. **F, W, Miles and G. M, Watson, Oak Ridge National Laboratory, unpublished work. predicted. This order of solubility has not often been experimentally observed, since most liquids have surface tensions so low that other solution effects predominate. The difference between the calculated and experimental values is in the di- rection expected if the estimated size of the hole is too small (except for argon in NaF-KF-LiF). The last column of Table 2.3.7 shows a trend in the ratio of the experimental to the calculated values of the Henry’s law constants. An approxi- mate calculation indicates that this trend can be explained if the polarization of the noble gas atom by the highly ionic salt medium is taken into ac- count. It may be concluded from these calculations that estimates can probably be made of the solubilities of the noble gases in media of high surface tension to within an order of magnitude. This accuracy is satisfactory for reactor applications. Chemical Reactions of Oxides with Fluorides in LiF-BeF2 J. H. Shaffer An effective means for separating uranium from undesirable fission products is required in order to develop a suitable reprocessing scheme for fused- salt nuclear reactor fuels. Experiments described previously?! demonstrated that uranium can be sepa- rated from rare earth fluorides, such as CeF3, in the 21 J, H. Shaffer, MSR Quar. Prog. Rep. Jan, 31, 1958, ORNL-2474, p 99. 92 simple binary solvent KF-LiF. In these experiments uranium was preferentially precipitated as UO, by the addition of calcium oxide to the fused salt mix- ture. In the presence of BeF,, uranium probably can be separated from cerium and cerium oxides, but apparently oxides of cerium and beryllium will coprecipitate after the precipitation of uranium is essentially complete. In practical applications the addition of calcium oxide to precipitate uranium is objectionable. Con- tamination of the fuel with calcium fluoride might induce serious complications, such as increasing the melt liguidus temperature. However, two alter- nate methods of uranium separation are being in- vestigated. First, beryllium oxide may be substi- tuted for calcium oxide to precipitate uranium and the rare earths. The primary advantage of this sub- stitution is that no foreign constituent is introduced into fuels containing beryllium. The results of pre- liminary experiments with the use of beryllium oxide have been reported.2?2 A second method would em- ploy a precipitant which would form a volatile product. Reagents of this class might be water, boron oxide, and silicon dioxide. Precipitations with BeO. — A chromatographic type of separation of uranium and fission product rare earths might be effected by using beryllium oxide as the column packing material. Separations of this type would yield a purified solvent as the column effluent and would provide a means of con- centrating the uranium in the column for subsequent recovery. Pertinent information for this process in- cludes knowledge of reaction rates, optimum particle size of BeO, column recovery techniques, etc. The rate of reaction of UF , with a onefold excess of BeO as a function of the particle size is illus- trated in Fig. 2.3.16. Experimental results indicate that the reaction of UI:4 with solid BeO is surface controlled. Further experiments are in progress to determine additional characteristics of the ex- traction process. |t is planned, for the next series of experiments, to use actual columns packed with BeO and to force the liquid LiF-Ber-UF4 mixture to flow through the columns to determine the amount of uranium extracted per pass. Formation of Yolatile Products. — Results of ex- perimental precipitations of uranium from UF | in LiF-BeF2 (63-37 mole %) at 600°C by the addition of water to the influent gas stream are presented in 22§ H. Shaffer, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 90. UNCLASSIFIED ORNL-LR-DWG 34544 | t)o BeQ PARTICLE SIZE \ ® 35 TO 60 MESH g0 fe) A 20 TO 35 MESH \. A\ 0 12 TO 20 MESH o] \ 0\3\ [ ] 2 \ \ \‘ )'\_\ 70 \ s X < 5 < 60 w z \ A D \ z . Z 50 \ s . \ i o A : \ S 40 N 2 N g \‘\ 30 \ [ ) 20 - *J N 0 0 20 40 60 80 100 120 TIME (min) Fig. 2.3.16. Rote of Reaction of BeO (100% Excess) With UF4 in LiF-BeF2 (63-37 mole %) at 600°C. Fig. 2.3.17. As may be seen, the stoichiometry in- dicates the formation of U02. This assumption has been confirmed by petrographic examination. Further experiments are in progress to investigate the precipitation with water of cerium, as well as the selective precipitation of uranium in the pres- ence of cerium, in various fused salt solvents. CHEMISTRY OF THE CORROSION PROCESS Activities in Metal Alloys S. Langer Studies of the activity of nickel in nickel-molyb- denum alloys by measurement of the electromotive force of cells of the type NiCl2 (5 wt %} in i | Ni (metal) NaCl-KCl eutectic Ni-Mo Alloy PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED ORNL-LR-DWG 34542 [ /3/ URANIUM ME TAL FOUND (w1 %) (5] 1 THEORETICAL END-POINT FOR UF, e g 0 100 200 300 400 500 HF EVOLVED {meq) Fig. 2.3.17. Precipitation of UO2 from UF, in Li F- BeF2 (63-37 mole %) With Hzo- have been continued. The disagreement between the potentials of identical pure nickel electrodes which was experienced in the pusm‘23 has been greatly reduced by changing to an uncompartmented cell. The potentials of the nickel electrodes now agree to within £1.0 mv and frequently to better than £0.5 mv. A condensation of the results obtained to date is given in Table 2.3.9. Cell potentials measured in the temperature range 680 to 820°C were plotted as a function of temperature, and a straight line was drawn that best fitted the data points. The data are not yet sufficiently precise to warrant more refined methods. The potential of a given alloy at 750°C was read from the graph, and the activities and ac- tivity coefficients listed in Table 2.3.9 were cal- culated from the potential data. The alloys used as electrodes were prepared by annealing for 16 hr at 1200°C and quenching to room temperature; thus the electrodes may not rep- resent equilibrium samples even after remaining at 23g, Langer, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 111. 93 MOLTEN-SALT REACTOR PROGRESS REPORT Table 2.3.9. Tentative Values of Activities and Activity Coefficients of Nickel Metal in Nickel-Molybdenum Alloys Nickel Potential Activi Activity Atom Fraction, at 750°C, civity, Coefficient, a.,. XNi E (mv) Ni Ni 0.937 1.95 0.96 1.02 0.904 2.95 0.94 1.04 0.800 3.45 0.92 1.16 temperature in a cell for a week. In future experi- ments, alloys which have been annealed at 750 to 850°C for at least 6 weeks will be used. Surface Tensions of BeF2 Mixtures B. J. Sturm Experiences in pumping LiF-BeF ,-UF , melts led to interest in the surface tensions of BeF, mix- tures, and preliminary measurements have been made. For these measurements a Cenco-DuNouy interfacial tensiometer?4 was used to measure the force required to pull a platinum ring out of the melt. Since the measurements were made with the liquid surface exposed to air, the results are ap- proximate. 24p oduct of the Cenco Scientific Company. A trial measurement of LiF-Ber-UF4 (53-46-1 mole %), with poor temperature control, gave a sur- face tension of 195 dynes/cm at 425 to 450°C. Measurements of a melt consisting of LiF and BeF (63-37 mole %) gave 196 dynes/em at about 480°C. The surface tension of the liquid can affect the drainage behavior, and the level of a liquid in a capillary is directly proportional to the surface tension and inversely proportional to the density. Therefore a comparison of the rise in a 2-mm-1D capillary is indicative of the validity of the results. Such a comparison is presented in Table 2.3.10 for two melts and for water. Disproportionation of Chromous Fluoride B. J. Sturm Chromous fluoride in solution with fused alkali fluorides disproportionates into trivalent chromium and chromium metal (see the following section). In an attempt to demonstrate this reaction in concen- trated solutions, and also to prepare compounds of NaF and CrF,, nickel capsules containing mixtures of sodium fluoride and chromous fluoride {(NaF-CrF 50-50 and 67-33 mole %) were heated to 1000°C. For capsules maintained at the elevated temperature for 2 hr and cooled slowly, x-ray diffraction showed sodium hexafluochromate (IHI) to be the principal phase. Capsules heated for only 2 min and quenched on a copper plate contained only a trace of sodium hexafluochromate (l1), the principal phase being one regarded as sodium trifluochromate (i1). The rapid cooling and quenching did not allow time for disproportionation; the product was N(:lF-CrF2 instead of 3N0F-CrF3. ! Table 2.3.10. Comparison of Surface Tensions and Approximate Rise in a 2-mm-ID Capillary of Two Fused Salts and Water Temperature Surface Tension Density Capillary Rise (°C) (dynes/cm) (g/cma) (em) NaF-ZrF ,-UF, 630 1329 3.34% 0.8 (50-46-4 mole %) LiF-BeF, 480 196 1.92° 2.1 (63-37 mole %) Water 25 72¢ 1 1.5 94 . 1. Cohen, W. D. Powers, and N. D, Greene, private communication. W. D. Powers, private communication, September 1958, “N. A. Lange (ed.), Handbook of Chemistry, 5th ed. (1944), p 1577. Effect of Fuel Composition on the Equilibria 3CrF, &= 2CrF, + Cr° and 3FeF2\—-——-——_)2FeF3 + Fe° J. D. Redman A study of the chemical equilibria involved in the corrosion of structural metals by fluoride fuels has continued, with attention being given to the effect of solvent composition on the equilibria 3CrF2#2CrF3 + Cr° and 3FeF2 —4\—- 2FeF2 + Fe® . Chemical analyses were made of equilibrated samples contained in nickel filtration apparatus. The samples were obtained by filtration after an equilibration period of 5 hr at temperatures chosen between 525 and 800°C. Chromium Fluorides. — The extent to which the reaction 3CrF2 #QCrF3 +Cr° proceeds to the right was found to decrease with increasing ZrF , concentration in the solvent LiF- NaF-ZrF ,, as shown in Fig. 2.3.18. The solvent LiF-Nch-ZrF4 was useful because it permitted a continuous variation from a basic to an acidic melt without interference from precipitation. The data of Fig. 2.3.18 are based on duplicate experiments in PERIOD ENDING OCTOBER 31, 1958 which excess chromium metal was equilibrated with 2 wt % Cr*** in the solvents described in Table 2.3.11. For this series of solvent compositions, in which the mole ratio of NaF to LiF was held con- stant at 40:60, the effect of Zr*" in competing with UNCLASSIFIED CGRNL.-LR-DWG 34543 [o2d O o (@) " (&) o Z o o y T~ o { PERCENTAGE OF DISSOLVED GHROMIUM PRESENT AS cr ¥+ 0 10 20 30 40 50 60 Zrf, IN SOLVENT (male To) Fig. 2.3.18, Effect of Solvent Composition on the Reaction 3Cr++‘=‘- 2Cr+++ + Cr° When 2 wt % Cr+++ is Equilibrated with Excess Cr® in the Solvent LiF-NaF (60-40 mole %) + ZrF , at 800°C. Table 2.3.11. The 3CrF2;—_-—'\2CrF3 + Cr® Equilibrium in LiF-NaF-ZrF, ot 800°C (2wt % ottt Equilibrated with Excess Chromium Metal) ZrF ; added (mole %) to LiF-NaF 0 15 25 35 47 55 (60-40 mole %) Cr found by analysis (wt %) in duplicate experiments cettr ottt 2.59 2.76 2.80 2.79 2.76 2.83 2.62 2.70 2.78 2.77 2.75 2.86 cett 1.38 2.10 2.30 2.30 2.52 2.82 1.25 2.21 2.27 2.34 2.55 2.50 Calculated Concentrations (wt %) Average Cr' ' 1.32 2.16 2.28 2.32 2.54 2.66 Average Cr'tt 1.29 0.66 0.50 0.46 0.22 0.18 K, = (cha)z/(chz)3 gx 10-1 4x10~2 2x10°2 1x1072 3x107¥ 2x7107° 95 MOLTEN-SALT REACTOR PROGRESS REPORT Cr*** for fluoride ions is apparent from the decreas- ing stability of Cr**7, Similar experiments at lower temperatures and lower dilutions of chromium ion consistently showed the same trend. Because of minor quantitative dis- crepancies, which appeared presumably as a result of analytical difficulties, the numerical results are mainly of qualitative interest and have been omitted from this report. The same is true of numerous trials involving the addition of CrF,, with no chro- mium metal or Cr*** initially present, to check the reaction 3CrF, — 2CrF, + Cr° Increasing dispro- portionation (more CrF,) with decreasing ZrF, content was quite evident. Since the chromium metal was deposited on the nickel walls at an ac- tivity of less than unity, the ratios of CrF, to CrF, were higher when only CrF, was added than for the equilibrium values in the presence of excess chro- mium metal. Tests were also made in solvents containing BeF,, which is considered to be more basic, or less acidic, than ZrF ,, since it is a better fluoride donor for furnishing fluorides for complexing. An appre- ciable amount of disproportionation, CrF3/CrF 0.15 for 2 wt % Cr at 800°C, was found in LiF-BeF, (48-52 mole %). This is roughly comparable to the amount of disproportionation in the 35 mole % ZrF solvent. In another BeF , mixture (LiF- BeF, ThF 67-23-10) the effect of ThF (acidic) was more fhan counterbalanced by the |ncrecsed LiF (basic) con- tent to give a net increase in the complexing of CrF,. Here the equilibrium Cr|:3/CrF2 ratio was about 0.4 at 800°C with 2 wt % Cr; thus the solvent containing ThF , has about the same complexing tendency as the 15 mole % ZrF, solvent. These results confirm the supposition thcf BeF, mixtures are more basic than corresponding ZrF, mnxtures and that the extent of dlsproporflonohon is greater in the more basic solvents. Numerous auxiliary trials were made to ensure that solubility limits were not exceeded and to explore the effect of lower dilutions and temperatures; the results were in good qualitative agreement with the foregoing conclu- sions. Quantitative discrepancies, such as a lack of material balance, were found that appeared to be due to experimental difficulties, particularly with respect to analyses. Ferrous Fluorides. — Experiments similar to those with chromium fluorides were also carried out with FeF, and FeF3 in the same solvents. No dispro- portionation of FeF, was detected in any of the sol- vent compositions at either 600 or 800°C. In the 96 course of these experiments it was noted that the solubility of FeF, in LiF-NaF- ZrF, (27-18-55 mole %) is 4.4 wt % Fe at 800°C ond 0.3 wt % at 600°C; in LiF- NaF-ZrF4 (32-21-47 mole %) the solubility is 0.4 wt % at 600°C. For comparison the solubility of FeF in NaF- ZrF, (53-47 mole %) at 600°C is 0.4 w'r %, as defermlned by electrometric measurements, 2 and 0.26 wt %, as determined by an earlier filtration experlmenf Activity Coefficients of CrF, in NuF-ZrF4 C. M. Blood The activity coefficients of CrF, dissolved in the molten mixture NaF- ZrF (53-47 mole %) at 850°C were determined and reporfed previously. 27 The re- sults of experimental measurements at 750°C are summarized in Table 2.3.12. The tabulation gives the experimentally determined partial pressures of HF and mole fractions of CrF,, together with the resulting equilibrium quotients. In each case the partial pressure of hydrogen was determined by sub- tracting the partial pressure of HF from the meas- ured total pressure, which was kept essentially constant and very close to 1 atm. The equilibrium quotients are also shown graphically as functions of the mole fraction of CrF, in Fig. 2.3.19. Examination of the results indicates that, within the experimental precision, the results are inde- pendent of the mole fraction of CrF, over the range from 1 to 10 mole %. If the extrapolation of the values of K, is valid to infinite dilution, as shown in Fig. 2.3.19, the activity coefficient of Crr:2 is unity, over the concentration range studied, with respect to the standard state of reference having unit activity coefficient at infinite dilution. Com- parison of the extrapolated equilibrium quotient with the equilibrium constants calculated from tabulated values of thermodynamic properties?® for the re- actions: CrF2 (s) + H, (g) #Cr (s) + 25 L. E. Topol, private communication. 6 . . . J. D. Redman, private communication. e M Blood, MSR Quar, Prog. Rep. Jan. 31, 1958, ORNL-2474, p 105. 28 L. Brewer et al,, Natl. Nuclear Energy Ser. Div, |V, 198 (1950). PERIOD ENDING OCTOBER 31, 1958 Table 2.3.12, Equilibrium Quotients at 750°C of the Reaction Cl'F2 (d) + H2 (g) =— Cr {s) + 2HF (g) in Nt:lF-ZI'F4 (53-47 mole %) Pyr XCer _ Pur xCrF2 - {atm) {mole fraction) * {atm) {mole fraction) * x 1073 x 1072 x 10~4 x 1073 x 102 x 10~4 4,23** 9.14 1.97 2,63*%** 3.42 2.03 4.97*** 9.18 2.71 2.51** 3.38 1.87 4,38*%** 9.23 2.09 2,76%** 3.44 2.22 5,12%*x 9.23 2,85 2,58*** 3.40 1.96 4,78*** 9.25 2.48 2,55%** 3.44 1.90 4,57*** 9.22 2.28 2,87*** 3.44 2.40 4,31** 9.25 2.02 2,66*"* 3.46 2.05 5.33* % 9.42 3.03 2,71+ 3.44 2,14 4,27** 9.44 1.94 2,70%** 3.50 2.09 4,61** 9.44 2,26 2.60** 3.52 1.93 4,96*** 2.50 2.60 3.02*%* 3.59 2,55 4.84*** 9.56 2,46 2,93*** 3.53 2.44 4,41%* 9.48 2,06 2,69*** 3.55 2,04 4.39%* 9.56 2.03 1,31%** 0.953 1.80 5.56*** 9.63 3.23 1.26** 0.943 1.69 4.72%* 9.71 2.31 1.62%* 1.07 2.47 4.86** 9.60 2.47 1.59** 1.09 2,33 5.56%** 9.88 3.15 1.53** 1.05 2,24 2,18** 3.07 1.30 1.54** 1.05 2.27 2,05** 3.07 1.37 1.57** 1.07 2,32 2.03%** 3.09 1.34 1.60** 0.988 2.60 2,18** 3.13 1.52 1.70** 1.03 2.82 1.91** 3.13 1.17 1.55** 1.05 2.30 2,15** 3.13 1.48 T.49* 1.07 2.09 2.08** 3.09 1.40 1.61*** 1.09 2.39 1.98** 3.13 1.25 1.84*** 1.10 3.07 2,09%* 3.13 1.40 1,79*** 1.09 2.96 1.99** 3.09 1.28 1,69** 1.14 2.51 2,28*** 3,13 1.66 1.86%** 1.07 3.26 2.06** 3.13 1.36 L71** 1.10 2,65 2,00** 3.1 1.29 1,89%** 1.16 3.08 1,93** 3.13 1.1¢9 2,06*** 1.22 3.49 1.91** 3.15 1.16 2, 12%** 1.26 3.58 2.05** 3.11 1.35 1.92** 1.26 2.93 2,31*** 3.13 1.71 2, 11 *** 1.28 3.49 2,84**%* 3.32 2.43 2,07*** 1.32 3.26 2.51** 3.42 1.85 2,16** 1.39 3.35 2,28** 3.38 1.54 Av 2,21 10,52 *K, = PE-I F/XCrF Py where X is mole fraction and P is pressure in atmospheres, **Determined under reducing conditions. ***Determined under oxidizing conditions. 97 MOLTEN-SALT REACTOR PROGRESS REPORT e DETERMINED UNDER * -3 UNCLASSIFIED (:é% ) ORNL-LR—DWG 34544 . { i [ _ -4 o DETERMINED UNDER K= (224 £ O-f’f’ x 10 REDUCING CONDITIONS 0.80 K, {s) = 4.2 x10 -4 K,(1) = 8.6 x 10 OXIDIZING CONDITIONS 0.60 Y(s) = 0.526 B, ) . o y(1) = 0.257 0.40 —1- "‘c?'o _ e ¢ o I A, A 020 f=z====s Vi it @ffi; G e T 0 0 1.0 20 3.0 40 5.0 6.0 7.0 8.0 3.0 (x1079) CrFy IN LIQUID (mote fraction) Fig. 2.3.19. Equilibrium Quotients at 750°C of the Reaction CrF2 (d) + H2 (9) ==Cr (s) + 2HF (g) in Na F-ZrF4 (53-47 mole %). and CrF, (1) + H, (g) ==Cr (s) + + 2HF (g), K_(I) = 8.6 x 1074, yields values of activity coefficients of CrF, of 0.53 and 0.26 with respect to the solid and liquid standard states respectively at 750°C. The corre- sponding values at 850°C (ref 27) are 0.28 and 0.18. The results of this investigation and of the corre- sponding investigations of NiF, and FeF, and of the activities of chromium in alloys?® can now be applied to the study of corrosion of Inconel and INOR alloys, where it is believed that the rate of corrosion is diffusion controlled. Experiments might be carried out where corrosion would be in- duced by the passage of HF-H, mixtures through salts in contact with the alloys. By controlling composition of the gas mixture it may be possible to attack the chromium without significant effects on iron or nickel corrosion. Equilibrium Amounts of CrF, in Molten Salts Containing UF ; Which Are in Contact with Inconel R. J. Sheil R. B. Evans The results of two previous invesm‘igc:tionsw'30 have been combined to calculate the equilibrium amounts of CrF, in two molten salts containing UF in contact with an infinitely thick Inconel container for an infinite period of time. The calculated values 29M. B. Panish, R. F. Newton, W. R, Grimes, and ¥, F. Blankenship, J. Pbhys., Chem., 62, 980 (1958). 30|, G. Overholser and J. D. Redman, private communicatjon, 98 are presented in Table 2.3.13 and in Fig. 2.3.20 as functions of temperature. The following assumptions were made for the cal- culations: 1. All the CrF, present at equilibrium resulted from the reaction Cr° + 2UF ;== 2UF, + GrF, 2. The activity coefficient for chromium in In- conel is independent of temperature and salt com- ponents. The value utilized was 0.563.2° 3. The average atom fraction of chromium in In- conel is 0.1688. 4. The salt contained no UF; or CrF,, initially. 5. The mean temperature which defined the equi- librium values corresponds to the balance-point3! temperature for polythermal flow systems. 31 R. B. Evans, MSR Quar. Prog. Rep. . 31, 1938 ORNL-2474, p 107. € 8. Rep- Jan 2% Table 2.3.13. Equilibrium Amounts of Chromium in UFA-Containing Salts in Inconel Equilibrium Cr (ppm) Temperature (OF) NflF-ZrF4'UF4 (50-46-4 mole %) LiF-Ber-UF4 (47.8-51.7-1.5 mole %) 1000 717 242 1100 785 345 . 1200 852 467 1300 920 600 1400 985 734 UNCLASSIFIED ORNL~-LR-DWG 34545 1400 1000 — o) _am® __ 900 a6 5 Z 800 L Q 700 = a L7 o ((\0\e [ a] =z 1/\' g 600 g B 1 s} & = axe 2 2 500 57 i NS (&) 400 - / // iiiii 200 1000 1100 1200 1300 1400 MEAN SALT TEMPERATURE (°F) Fig. 2.3.20, Calculated Equilibrium Amount of Chromium Present in UFA-Confuining Salt in Inconel, The chromium concentrations shown also repre- sent the amounts of chromium metal which should be added to pre-equilibrate a salt prior to loop tests. Under this condition, the corrosion rates will de- pend only on the hot-to-cold-zone transfer mechao- nism. Such results will be highly indicative of the long-term corrosion rates to be expected in molten- salt reactors. Chromium Diffusion in Alloys R. B. Evans R. J. Sheil W. M. Johnson Studies of chromium diffusion in chromium-con- taining nickel-base alloys were continued along the lines described previously.32 New equipment is being developed for the depletion-method experi- ments that are based on the rate of loss of Cr37 (as Cr H*) from a noncorrosive salt contained in the alloy being studied. The pots used in previous tests were costly, and the experimental procedures for their use were complicated. They were designed to assure a constant ratio of the salt-exposed area of the container to the salt volume (A4/V). Capsules 325, B. Evons and R. J. Sheil, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 95. PERIOD ENDING OCTOBER 31, 1958 are being designed to replace the pots, and equip- ment for testing several capsules simultaneously is being developed. The constant-potential method was utilized further to prepare specimens for the study of the distri- bution of Cr®! within the alloy. In these investi- gations, measurements are made of the gain of Cr5! by a metal specimen exposed to molten salt con- taining a dissolved amount of chromium 51 (as Cr***) that remains constant with time. It has been found that the results obtained by this method are extremely sensitive to the manner in which the specimens are prepared prior to the experiment. The diffusion coefficients resulting from initial experiments, which were conducted with specimens which had been polished but not hydrogen fired, were one to two orders of magnitude higher than those obtained with specimens that had been pol- ished and then hydrogen fired at 1100 to 1200°C for 3 hr. This information is of importance with respect to the use of tracers in corrosive media, since a large portion of the dynamic corrosion studies at ORNL are conducted in nonhydrogen-fired con- tainers. The over-all diffusion coefficients obtained to date, including data obtained by Gruzin and Fedorov3?3 and by Price et al., 4 are presented as log D vs 1/T plots in Fig. 2.3.21. Little signifi- cance is attached to the low values for the hydro- gen-fired specimens tested at 640 and 675°C, since the radioactive count of the Cr>! was virtually at background level for these measurements. The coefficients indicated at 750, 700, and 640°C are based on the results of two or more determi- nations at each temperature. The data from which they were obtained are plotted on Fig. 2.3.22. The straight lines indicate that, with the exception of a small discrepancy around zero time, the square-root time law utilized for the calculations is valid. Equipment for sectioning specimens by electro- polishing techniques is now being developed jointly with members of the Dynamic Corrosion Group of the ORNL Metallurgy Division. The technique thus de- veloped will be used for tracer studies of chromium diffusion phenomena in thermal-convection loops. It is felt that a comparison of diffusion coefficients 33p, L, Gruzin and G, B, Fedorov, Doklady Akad, Nauk. S.5.5.R. 105, 264 (1955). 34R. B. Price et al., A Tracer Study of the Transport of Chromium in Fluoride Fuel Systems, BMI-1194 (June 18, 1957). MOLTEN-SALT REACTOR PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 343546 TEMPERATURE (°C) 1000 900 800 . 700 600 ace IINTO NICIKEL (GRUZIN AND FEDoéov, REF 33) | & CAPSULE EXPERIMENTS (PRICE &/ o/, REF 34) _____| ® POT WITH CONSTANT A/V AND FILLED WITH NGF —KF=LiF (11.5-46.5-42 mole %) -9 [ 0 CONSTANT POTENTIAL EXPERIMENTS WITH UNFIRED | SPECIMENS _10 |® CONSTANT POTENTIAL EXPERIMENTS WITH HYDROGEN FIRED SPECIMENS -1 ”“‘—\ ~ log,, D (WHERE D IS IN cm%sec) ’l ? / ‘I ¢ ! F—r 7.5 8C 85 90 95 100 105 #1.0 15 104/7 (°K) Fig. 2.3.21. Current Results of Chromium Diffusion Experiments. obtained from salt behavior with those obtained from_ distribution curves is essential to a complete under- standing of the diffusional processes which take place within the metal. It is hoped that the results of these experiments will provide a means for pre- dicting the void penetration distances to be ex- pected for a given set of corrosion conditions. CORROSION OF METALS BY ALUMINUM CHLORIDE R. E. Moore Corrosion of Inconel by Aluminum Chloride Vapor in a Fused Silica Container Purified aluminum chloride vapor in a container of fused silica at 730°C was found to have a very de- structive effect on Inconel. The results of the ex- periment showed that the aluminum chloride had reacted with the silica to produce silicon tetra- chloride, which attacked the metal. In essence, this experiment was a corrosion test of Inconel ex- posed to silicon tetrachloride rather than to alu- minum chloride. Aluminum chloride should be much 100 UNCLASSIFIED ORNL~LR-DWG 34547 (x10‘3) [ / | 100 // 90 : 80 A + . 1(0 | l s/ /! 70 o 17 / & y A S 60 O & 50 o 40 Q/’ O / ACTIVITY IN SALT PER UNIT WEIGHT o £ 30 1P Q}?/ l\% 1 N — ] | | 20 / cg\)q ’ v . / & [ 10 ) 1 ={762 ppm ’ + / /.MELI\SUREMENTS ot 6a0°C,Lert ] 0 /b 2 2! 0 1 2 3 4 5 6 ?| 8 SQUARE ROOT OF EXPOSURE TIME (hr'2) RATIO OF Cr3' ACTIVITY IN METAL PER UNIT AREA TO Cr5! Fig. 2,3.22. Time Dependence of Chromium Diffusion Data Obtained in Constant Potential Experiments. more stable toward Inconel on the basis of free energies of formation. The three polished Inconel strips that were tested lost about 1% in weight in 500 hr and were covered with a dull coating, which was identified by x-ray diffraction examination as mainly Ni Si. Large amounts of chromous chloride and ferrous chloride, identified by x-ray and spec- trographic examinations, were found in the aluminum chloride in the part of the apparatus that constituted the reservoir. This aluminum chloride reservoir had been held at about 180°C during the test in order to maintain a pressure of aluminum chloride of about 1 atm. A small amount of nickel was also found by analysis to be present in the aluminum chloride in the reservoir. A white coating on the fused silica was identified by x-ray diffraction as an aluminum silicate. Corrosion of Inconel by Aluminum Chloride Yapor in an All-Metal System A further corrosion experiment was carried out with an all-metal system in order to avoid the for- mation of SiCl,. Purified aluminum chloride gas was maintained in contact with the walls of an In- conel tube for 300 hr at 720°C. A copper metal reservoir containing aluminum chloride, which was attached to the Inconel tube, was held at about 150°C to maintain a pressure of aluminum chloride of about 125 mm Hg during the test. Metallographic examination®3 of the metal showed the complete absence of subsurface voids. A photo- micrograph of the test specimen (Fig. 2.3.23) shows very large grains 1o a depth of about 10 mils in com- parison with the untreated fine-grained Inconel sample (Fig. 2.3.24). The grain growth, as well as the irregular nature of the surface, is probably a re- sult of the heat treatment or hydrogen firing rather than the corrosive attack. Chemical analysis of the contents of the reservoir showed 3.0 mg chromium, 0.2 mg iron, and 0.5 mg nickel. As expected, chro- mium was attacked to a much greater extent than iron or nickel. The chromium in the reservoir repre- sents attack on the Inconel of the equivalent of 0.024 mil over the 44-cm? surface area exposed to attack at 720°C. Considering the worst case and assuming that the reaction does not become dif- fusion controlled after the initial attack on the sur- face, the amount of attack would be proportional to 35E xamination performed by R. S. Crouse of the ORNL Metallurgy Division. UNCLASSIFIED T-15846 Fig. 2.3.23. Tube Wall of Inconel Test Specimen Which was Exposed to Aluminum Chloride Vapor for 300 hr at 720°C. 250X, PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED T-15845 / ‘- ~i, oy - = X {2 ¥ ;.- \ S e .Q',/ e ? N 1 <(i /) = X a . 3 " 7 [ " .' . ?/ 09 U I. ‘ { e " D ‘C? Y r = s (/v.‘ = L R 4 1T ] g . . ,‘ ’r - ’\ K 4 - 'y, " k’., P 2 :«2 (“\ ) t \ s — < 0l o _F‘ i - Are A= 1 ¢ g o % ) R et ol 4 L.ol2 : (.25 L e o 2 “® \ — ‘/. L‘/ ( S . S p = . y s t A -v__\ \ o3 T Q:'A // "I'- - | -‘ A : v.‘f_,— — \ A RS : {_ s ' ek < - 014 3 C . ' ~ . .| i . v 3 VS < ¥ ~ . v ’ “/‘ L (/ jf‘ ols & L [ <=5 N2 {\\ N &, .;fi’u P B A g ~. - N \ o, okl L9 Jo L% 5 EAU N N 18 b M . L " < - " - .T 2 O 4T AR TR At S Fig. 2.3.24. Tube Wall of Untreated Fine-Grained Inconel Test Specimen. 250X. the pressure of aluminum chloride and at 5 atm the corrosion after 300 hr would be about 0.7 mil. Significance of Experimental Results The existence of a reservoir of aluminum chloride at a temperature of 150 to 200°C in a corrosion test is unfavorable because the relatively cold reservoir allows condensation of NiCl, and FeCl,,. In future tests the container will be maintained at 650°C on one end and 450°C on the other end to duplicate actual conditions in a reactor. Also, the aluminum chloride pressure will be maintained at about 5 atm. Even though it is expected that CrCl, will condense at 450°C from alloys containing chromium, the slight attack on Inconel found in the all-metal test described above is very encouraging. INOR-8, in which the activity of chromium is much lower than in Inconel, should prove quite resistant to aluminum chloride vapor. The corrosion of a number of metals, including nickel, steel, stainless steel, INOR-8, and Inconel, is under study. Theoretical Considerations of Aluminum Chloride Vapor Corrosion Theoretical calculations of M. Blander and R. F. Newton (see following section) suggest that alu- minum chloride vapor would be an excellent coolant 101 MOLTEN-SALT REACTOR PROGRESS REPORT for a power reactor. Its primary advantoge over other gas coolants is the large amount of heat which could be transferred by the reaction AlLCl, ==2AICI, in the vapor phase. An important consideration, however, in the practical use of aluminum chloride is a choice of a suitable container material so as to minimize corrosion. The reactions of importance in corrosion of metals by aluminum chloride vapor are (1) 2AICI, + 3M° ==3MCl, + 2AI° (2) AICI, + 2A1° &==3AICI (3) AICH, + M° ==AICl + MCI, In these equations M represents any component of a structural metal. Equation 3 was obtained by adding Egs. 1 and 2. For the case of a reactor operating at 650°C with the temperature of the heat exchanger at 450°C, equilibrium constants may be calculated for the equations given above for various components from values of free energies of formation given by Glassner.38 The principal uncertainty in such cal- culations is the extent of absorption of aluminum metal into the surface of the container as a solid solution. Equation 3 should be the important reaction for most components of structural metals because the pressures of MCl, calculated from this equation are higher than those calculated from Eq. 1. The equi- librium pressure of AICI in AlCl3 at 5 atm and 650°C in contact with aluminum metal is about 0.28 mm Hg. The equilibrium pressure of AICI calcu- lated from Eq. 3 for Fe, Cr, and Ni under the same conditions is very much less than 0.28 mm Hg. Hence, disproportionation of AICI should occur only by the process of absorption of aluminum metal into the surface of the container. When this occurs there should be a reduction in pressures of MCI, action with the aluminum in the container. The equilibrium pressures of AICI and MCI, at 450°C are less than those at 650°C. Thus, transfer of metal from the reactor core to the heat exchanger should take place. by re- 364, Glassner, A Survey of the Free Energies of For- mation of the Fluorides, Chlorides, and Oxides of the Elements to 2500°K, ANL-5107 {Aug. 1953). 102 Another important factor is the vapor pressure of MCl, at 450°C. If the vapor pressure is lower than the equilibrium pressure from Eq. 3, MCI, should deposit in the heat exchanger. Therefore, suitable vapor should have the (1) a low oxidation-re- metals for containing AICI, following characteristics: duction equilibrium pressure for the metal chloride at 650°C and (2) a vapor pressure for the metal chloride at 450°C that is higher than the equilibrium pressure at 650°C. Structural Metal Considerations Considerations of various possible components of structural metals has led to the following generali- zations: Chromium. — A high chromium content may be un- desirable because the equilibrium pressure of CrCl at 650°C over an alloy such as Inconel or INOR-8 is much higher than the vapor pressure of CrCl, at 450°C. The deposition of CrCl, on the cooler walls of the heat exchanger might interfere with heat transfer. Iron. - The equilibrium pressure of FeCl, is somewhat less than that of CrClz, but the vapor pressure’’ of FeCl, at 450°C is higher than its equilibrium pressure at 650°C. lron would therefore be expected to deposit in the heat exchanger as metal (Eq. 3) and not seriously interfere with heat transfer. Nickel. — The equilibrium pressure of NiCl, 650°C is less than that of FeCl,. The vapor pres- sure of NiCl, at 450°C is probably high enough to preclude deposifion of NiCl, in the heat exchanger. Manganese. — Manganese appears to be an unde- sirable component because of a very high equilib- rium pressure of MnCl . Molybdenum. — The equilibrium pressures of mo- lybdenum chlorides are very low. Molybdenum should be a satisfactory component of a container metal. These generalizations are based solely on equi- librium considerations and have no relation to rates of corrosion. After initial attack on the surface of an alloy the rate will probably be diffusion con- trolled, and it could be very slow. 37C Beusman, Activities in the KCl- FeCl and LiCl- FeCl Systems, ORNL-2323 (May 15, ]957) GASEOUS ALUMINUM CHLORIDE AS A HEAT EXCHANGE MEDIUM M. Blander R. F. Newton Aluminum chloride has faverable and unique prop- erties which make it, as a gas, a heat transfer medium of potential utility at high temperatures. Gases ordinarily transfer heat because of the lower- ing of their heat content upon cooling from a temper- ature T, to a temperature T, in a heat exchanger according to the equation where H, — H , is the heat content increment and Cp is the heat capacity of the gas at constant pressure. Aluminum chloride exists as a dimer Al,Cl, at low temperatures and as a monomer AICI, at high tem- peratures. Because of the large number of atoms per molecule the heat capacity of the gas is high. Yalues of the heat capacities, of AI2CI6 and of AICI; were estimated theoretically by statistical mechanical methods, 38 from the infrared vibrational frequencies measured or estimated by Klemperer.3? 38J. E. Mayer and M. G. Mayer, Statistical Mechanics, Wiley (1940), p 440. 3w, Klemperer, J. Chem., Phys. 24, 353 (1956). o PERIOD ENDING OCTOBER 31, 1958 The values obtained were 42 cal/moles®C for AlCl, and 19 cal/moles°C for AICl;. For ease of calculation the approximation was made that the heat capacity of AlCl; was half that of A|2C|6, which was equivalent to assuming that the heat ca- pacity of AICl is 21 cal/moles°C. Table 2.3.14 presents comparisons of the heat transferred per mole of Al ,Cl,, CO,, and He for various temper- ature intervals. On this basis alone, it may be seen from Table 2.3.14 that the heat transfer capacity of AlL,Cl, is high. A further mode of heat transfer is possible, however, since at high enough temper- atures the major gaseous species is AlCl; and at lower temperatures the major species is Al Cl,. Cooling the gas in a heat exchanger will remove, in addition to the heat obtained from the temperature change, the heat from the reaction 2AICI, == AlCI, The heat content as a result of the dissociation re- action was calculated by using a value of 29.6 kcal for the heat change for the reaction and 34.6 cal/°C for the entropy change,*® and the data are presented in Table 2.3.15. The heat changes per mole of Al,Cl, from the reaction are given in column 5 of 404, Shepp ond S. H, Baver, J, Am, Chem, Soc. 76, 265 (1954). Table 2.3.14. Heat Transfer Data for AI2C|6' C02, and He for Several Temperature Intervals 1 2 3 4 5 6 7 Heat Change Temperature Heat from Sum of Heat Changes, Adjusted Heat Change, Interval (keal/mole) Association of Columns 2 + 5 Column 6 Divided by (°C) Al,Cl, CO, He AICI4 (kcal/mole) (kcal/mole) (1 + ¥)* (kcal/mole) 4501000 12.8 3.6 1.5 8.8 21.6 18.8 450-1300 19.8 5.7 2.3 25.1 4.9 31.7 750—1300 12.8 3.8 1.5 23.7 36.5 26.0 750-1000 5.8 1.7 0.7 7.4 13.2 1n.2 10002000 23.3 7.3 2.8 20.6 43.9 26.6 1500-2000 11.7 3.8 1.4 1.4 13.1 6.6 1000-1500 11.3 3.6 1.4 19.2 30.9 19.0 12001500 7.0 2.2 0.8 6.7 13.7 7.5 *In this expression, (1 + X} is the number of moles of gas molecules at the average temperature of the interval stated. MOLTEN-SALT REACTOR PROGRESS REPORT Toble 2.3.15. Dissociation of AI2CI6 and Its Contribution to Heat Content Temperature Temperature Fraction of Gas Dissociation Heat Content Due to Dissociation (°F) (k) at One Atmosphere (kcal per formula weight of AI2CI6) 450 505 0.001 0.03 750 672 0.05 1.48 1000 811 0.30 §.88 1200 922 0.725 21.46 1300 978 0.85 25.16 1500 1089 0.95 28.12 2000 1366 0.997 29.5 Table 2.3.14 for the stated temperature intervals. In column 6 of Table 2.3.14 are listed the sums of the heat changes resulting from cooling and the heat contribution from the dimerization. Since the number of moles of gas molecules in a forced-circulation loop is greater than that calculated if it is assumed that the gas is all Al Cl, the values of the heat changes must be adjusted to obtain the heat change of the gus per mole of gas molecules. A crude ad- justment can be made by dividing the heat change of column 6 by the number of moles of gas molecules at the average temperature of the interval stated. These values are listed in column 7 of Table 2.3.14. The large values of the heat transferred by this cycle of association and dissociation further en- hance the possible utility of aluminum chloride as a heat transfer medium. No measurement of viscosity or thermal condue- tivity of AICI, or A|2CI6 appears to have been made. Values for these properties are needed in design of thermal-convection loops for the equilib- rium mixture of these gases. Accordingly values for these properties have been estimated for the gases at 500 and 1000°K and at 1 atm pressure. These approximations are not claimed to be accurate and are presented only to serve as rough working figures for engineering calculations. The calculations were performed as suggested by Hirschfelder, Curtiss, and Byrd,“ and it must be 41 J. 0. Hirschfelder, C, F. Curtiss, and R. B. Byrd, Molecular Theory of Gases and Liquids, Wiley (1954). 104 emphasized that the theory used may be applied with real accuracy only to the very simplest of gases. The crudest model of a gas is an assembly of monatomic, noninteracting hard spheres. Hirsch- felder et al. show the equations for viscosity, 7y, and thermal conductivity, A, where the subscript HS indicates the hard sphere approximation, to be: VMT (]) T]HS = 2.6693 x ]0—5 —2 ' g/cm.sec‘ o T/M (2) A, =1.9891x 1074 = 52 15R ——7, cal/em®Cisec, 4 M where . -« O o = molecular diameter in A, T = temperature in °K, M = gram molecular weight, R = gas constant. The molecular diameter may be estimated crudely for AI2C|6 from electron diffraction data;4? struc- tural estimates have been published for AlC[a. The crude approximations which result from these calcu- lations are presented in Table 2.3.16. Application of the hard-sphere model to poly- atomic molecules yields values for thermal conduc- tivity which are too small. The transfer of energy 42\ | R, Maxwell, J. Optical Soc. 30, 374 (1940). PERIOD ENDING OCTOBER 31, 1958 Table 2.3.16. Viscosity and Thermal Conductivity of Al2C|6 and AlCi3 Calculated According to Hard-Sphere Approximation —7 Viscosity Thermal Conductivity (gz (g/cm.sec) (cal/cm:°C.sec) ) At 500°K At 1000°K At 500°K At 1000°K x 108 x 10~8 x 10~ x 1076 AI2CI6 65 150 212 4.2 5.9 AI('.:I3 40 172 243 . 9.3 13.6 to or from vibrational and rotational degrees of free- T*, defined by dom (which are obviously missing in the hard-sphere LT case) permits greater heat transfer per collision for (6) T* =, £ polyatomic molecules. The anharmonic vibrations of AICl, and Al,Cl, and the relatively high temper- atures for the system of interest make it likely that all the modes of vibration, rotation, and translation are near equilibrium at all times. |f such equilib- rium is attained, the Eucken correlation 4 ¢, 3 (3) Ap=Ays | T 5o where C_ is heat capacity at constant volume, R is the gas constant, and A is the corrected thermal conductivity, is valid. Since C_ is about 40 for A12C|6 and about 17 for A|C|3, the AE/AHS values are about 6 and 2.9 for the dimer and monomer, re- spectively. For Al,Cl the values for A are 24,3 x 10-¢ and 34.3 x 107¢ cal/em«°C sec at 500 and 1000°K, respectively, and the corresponding values for AICI3 are 27.1 x 10~% and 38.1 x 105, respec- tively. The hard-sphere model must be further corrected, however, for effects of energetic interactions of the gas molecules (van der Waals forces, etc). These corrections take the form @ MTHs n= Q ’ (5) A e - where Q, for the Lennard-Jones potential, can be calculated from first principles for potential func- tions of sufficient simplicity. Hirschfelder et al. tabulate values of £ as a function of the parameter where & is the gas constant per molecule and € is the depth of the Lennard-Jones potential welli: m wae(7) -] Unfortunately there is no way to estimate € for AI2CI6 or AICFB. However, Hirschfelder et al. list values of €/k for a variety of substances. The values for halogen-containing substances range from 324 for HI to 1550 for SnCl . If these limits for /k are used, the values for T* and Q shown in Table 2.3.17 result. These values indicate that 7 and A are smaller than 5, ¢ and A, by a factor less than 3. A value of 2 for Q was, accordingly, chosen ar- bitrarily, and the resulting reasonable estimates of viscosity and thermal conductivity for dimer and monomer of aluminum chloride are presented in Table 2.3.18. Table 2,3.17. Probable Limits for Lennard-Jones Potential Function Temperature € & °K) /k T Q 500 324 1.5 1.3 1550 0.32 2.7 1000 324 3.09 1.0 1550 0.65 2.0 105 MOL TEN-SALT REACTOR PROGRESS REPORT Toble 2.3,18. Estimated Values for Viscosity and Thermal Conductivity of A|2CI6 and AlCI3 Thermal Conductivity {cal/em:°Cssec) Viscosity {(g/cmesec) At 500°K At 1000°K At 500°K At 1000°K the usual manner?® by assuming a constant compo- sition for a change of temperature, AH is the heat of dimerization, D , 5 is the interdiffusion coefficient of the dimer A and monomer B, P is the gas pres- are the weight fractions of dimer and monomer, R and R“ are the gas constants in the different units required, and T is the temperature. sure, W, and W % 106 % 10~ % 10-8 % 106 The interdiffusion coeffi4cien'r D ,gP may be esti- mated from the equation?’ A|2C16 75 106 12,2 17.2 ; 0.0026280 / T3[(M , + M )/2¥ ,Mp] AICl 86 122 13.6 19.1 B 3 (9 D, zP = UiB Q- Substances which dimerize have, at temperatures at which there is an appreciable fraction of both monomer and dimer, a higher effective thermal con- ductivity than a gas whose composition is invariant with temperature.43~4% This increased thermal conductivity is a result of the change of the equi- librium constant for the association with temper- ature. Theoretical calculations?? for this effect yielded the equation 8 N AH2 D apP WoWg ® e =" g2 R'T 2 where A _ is the effective thermal conductivity, A is the "“trozen’’ thermal conductivity calculated in 43k, P. Coffin and C. O’Neal, Jr., NACA Technical Note 4209 (1957). 44 ). N. Butler and R. S. Brokaw, J. Chem. Phys. 26, 1636 (1957). 45}, 0. Hirschfelder, J. Chem. Phys. 26, 274 (1957). where M, and M are the molecular weights of the dimer and monomer, O,p i8S the sum of the collision radii for dimer and monomer, where 0, = ]/2(0A + ch), and £ is a correction parameter similar to that described in Egs. 4 and 5 above and which can be calculated for simple potential functions. It does not differ greatly from € (ref 48). The values of the parameters chosen for the calculations described below were: AH =29.6 kcal, M; = 133.4, 0, = 8.06A, o, = 6.32A, o, = 7.19A, and Q" = 2. Some of the calculated values of D, P and DABPAHz/ RR’T3 for several temperatures are given in Table 2.3.19. Values of the weight fractions of the two gaseous species can be calculated from the equilibrium con- stant for the dimerization. These were calculated, 46Hirschfe|der, Curtiss, and Byrd, Op. cit., p 534. 47 . 16id., p 539, 8 1bid., p 534. Table 2.3.19. Calculated Effective Thermal Conductivities of Aluminum Chloride ()\e) and Some Parameters Needed in the Calculations 2 Thermal Conductiviti .deg- Temperature DABP DABP AH WB for WB for N " f Aermc:ACfon c:vmes (CUVCH; d:g sec) W ok 2 —————— p-4.4at P =10 ot . — A for — A, for veroge*® or or (k) [(cm?/sec)-atm] RR'T? o omm P= 4.t{atm Pe= 10 atm A/ P =2.4 atm P =e'|0 atm x 10-3 x 10-3 x 10~6 x 10-% x 10-6 x 10~¢ x 10—% 500 21.4 0.92 5x10"4 3.2x10°4 <0.1 <0.1 12 12 12 600 28.1 0.70 6x10"% 39x10-3 2 1 13 15 14 700 35.4 0.56 3.5x10"2 23 x10"? 10 6 14 24 20 800 40.3 0.42 0.13 0.087 24 15 16 40 31 900 51.6 0.38 0.35 0.24 43 35 17 60 52 1000 60.5 0.33 0.65 0.49 38 4) 18 56 59 1100 69.3 0.28 0.86 0.74 17 27 20 37 47 *A/ is a weighed average of the calculated values for AICl, and Al,Cl . 106 as before, by using a value of 29.6 keal for the heat of dimerization and 34.6 cal/°C for the entropy of dimerization in the equation (10) AF°=AH -~ TAS°=-RT In K The calculated weight fractions of monomer are listed in columns 4 and 5 of Table 2.3.19 for total pressures of 4.4 and 10 atm, respectively. Columns 6 and 7 list the contribution to the thermal con- ductivity from the dimerization process at each of the two pressures, 4.4 and 10 atm. These values are to be added to the ‘‘frozen’’ thermal conduc- tivity given in column 8. The calculated values of A_are given in the last two columns of the table. It is to be noted that the largest values of the thermal conductivity are for compositions which exhibit the largest change of composition with temperature, and hence at temperatures at which there is the fastest release of the ‘‘chemi- cal’’ heat of association. Such results are a con- sequence of the theory and lead to the result that within the range that the ‘‘chemical’’ heat is available, the average effective heat conductivity is about 80% of the maximum value. Thus the most effective temperature range for the utilization of AICl, as a heat transfer medium is limited to temperatures for which the composition of the vapor still has an appreciable fraction of both monomer and dimer (that is, at least 5% of either). The effective thermal conductivities and effective heat capacities are not very different from the “‘frozen’’ values outside this range. These calculations contain many approxima- tions. Although an attempt was made to be con- servative in these estimates, they should be used only as tentative working figures until experi- mental measurements can be made. PERMEABILITY OF GRAPHITE BY MOLTEN FLUORIDE SALTS G. J. Nessle J. E. Eorgan Tests of the penetration of graphite by molten salts were continued. Graphite rod specimens }4, ]/2, and 1 in. in diameter, prepared as de- scribed previously,? were outgassed and then immersed in NaF-KF-LiF (11.5-46.5-42 mole %) at 1700°F for a specific length of time. The weight gains of the graphite in this medium are 49(5. J. Nessle and J. E, Eorgan, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 99. PERIOD ENDING OCTOBER 31, 1958 compared in Table 2.3.20 with similar data for graphite tested in LiF-MgF , (67.5-32.5 mole %). 1 The NaF-LiF-KF mixture has a lower density and | a lower viscosity than the LiF-MgF , mixture. It may be seen from the data in Table 2.3.20 that NaF-LiF<-KF penetrated the graphite to a greater extent than did the LiF-MgF ,, mixture. Chemical analyses of machined cuttings from the ]/2- and l-in. rods exposed to NaF-LiF-KF indi- cated uniform penetration of the salt to the center of the rods, as was found to be the case with the LiF-MgF , mixture. Table 2,3.20. Weight Gains of Graphite Exposed to Fused Salts Average Net Weight Gain of Rod Graphite Rod (%) Diameter (in.) In NaF-LiF-KF in LiF-Mg F2 (11.5-46.5-42 mole %) (67.5-32.5 mole %) ) 9.1 A . 6.8 A 9.8 7.5 1 1.6 8.4 PREPARATION OF PURIFIED MATERIALS J. P. Blakely G. J. Nessle F. F. Blankenship Preparation of Pure Fluoride Compounds B. J. Sturm Further attention was given to the problem of preparing anhydrous chromous fluoride for use in research related to corrosion problems with fused fluorides. Chromous fluoride resulting from the hydrogen reduction of chromic fluoride is con- taminated with chromium metal, while that made from ammonium hexafluochromate (l11) usuaily contains chromium nitride. A difficult recrystalli- zation is necessary to purify these products., The product of the reaction of stannous fluoride with anhydrous chromic chloride is contaminated with chloride. 59 A different method of preparing chromous fluo- ride was therefore explored. Stannous fluoride was 30g, J. Sturm, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 101. 107 MOLTEN-SALT REACTOR PROGRESS REPORT heated with granulated chromium metal to over and/or ThF , component, under an HF atmosphere, 1100°C and the meolten material was allowed to and treating the barren melt, at 1500°F, first with stand at this temperature for a few hours, The H. for 2 hr and then with HF for 12 hr. The melt products were chromous fluoride and tin metal in is then cooled to below freezing (about 400°F), well-defined layers that were easily separated. and the UF , and/or ThF is added. The melt is The reaction is then reheated to 1500°F and treated with HF for an additional 4 hr. The melt is then stripped with SnF, + Cr — CrF, + Sn 2 2 H, gas until chemical analyses of periodic ‘‘thief"’ The chromium-containing portion had the crystal- samples indicate the batch to be acceptable. ; . . 51 lographic properties of pure chromous fluoride. The 9 batches processed included the six dif- The chemical analysis of the material was in good ) ferent compositions listed in Table 2.3.21. The agreement with the theoretical analysis. Micro- analyses of the final batches are given in Table scopic observation showed the product to be free 2.3.99 of opaque material, and thus the absence of free Batch 1149, for which the analysis showed 1140 metal is indicated. In cooperation with the ORNL Isotopes Division, vanadium trifluoride was prepared by thermal de- ppm of Fe, is being rechecked and if necessary will be reprocessed or replaced before use. Copper-lined stainless steel reaction vessels and copper thermocouple wells and dip legs were used during this operation of the production composition at about 500°C of ammonium hexa- fluovanadate (l11) made by fusing ammonium bi- fluoride with vanadium trioxide.”? This method _ . facility, because the relatively high sulfur content is more convenient than the usual hydrofluorina- of the beryllium fluoride had previously caused tion procedures. frequent failures of nickel thermocouple wells and Fusion of ammonium bifluoride with hydrated ferric fluoride formed ammonium hexafluoferrate (111} of the same crystallographic properties as those reported for the aqueous preparation.53 The dip legs. The copper dip leg was replaced with a nickel dip leg when a batch was completed and was to be transferred to a storage vessel. No _ e failures of thermocouple wells or dip legs occurred crystals were isotropic with a refractive index of 1.444. Thermal decomposition at about 500°C in a helium atmosphere yielded ferrous fluoride according to the equation C. R. Croft J. Truitt during this operation, Experimental-Scale Operations 6(NH ). FeF ,—> 6FeF. + 16NH .t + 24HF* + N_? Eleven small batches totaling approximately 4/3 6 2 3 2 . 88 kg were processed for use in small-scale cor- Production-Scale Operations rosion tests and physical property studies. Some of the small-scale units are being modified to J. E. Eorgan permit their use in a thorough investigation of the The production facility was activated for approx- graphite permeability problem. imately 5 weeks of the quarter to process 9 batches ; . . . Table 2.3.21. Beryllium-Containing Mixtures of various beryllium compositions to be used in Prepared for Experimentation component tests. Since the purification process used for ZrF , mixtures is not adequate for oxide Constituents (mole %) removal from these compositions, the procedure Composition was modified and extended to allow longer hydro- No. UF, ThF, BeF, LiF NaF fluorination time. In brief, the procedure consists of melting the blended salts, without the UF 123 1 46 53 124 7 35 58 STy, Insley et al., Optical Properties and X-Ray Dif- fraction Data for Some Inorganic Fluoride and Chloride 130 1 37 62 Compounds, ORNL-2192 (Oct, 23, 1956). 52B. J. Sturm and C, W, Sheridan, Preparation of Va- nadium Trifluoride by the Thermal Decomposition of Ammonium Hexafluovanadate (111), ORNL CF-58-5-95. 33Data on Chemicals for Ceramic Use, Bulletin of the 135 0.5 1 45.5 53 National Research Council, No, 118, p 6, June 1949, 133 13 16 71 134 0.5 1 36.5 62 108 PERIOD ENDING OCTOBER 31, 1958 Table 2.3.22, Analyses of Production Facility Batches of Beryllium-Containing Mixtures Batch Composition Major Constituents (wt %) Minor Impurities (ppm) ’ No. No. U Th Be F Li Na Ni Cr Fe S g 1142 130 5.02 9.22 73.4 12.1 40 220 70 2 1143 134 3.14 5.27 9.34 71.3 1.2 65 250 335 2 1144 134 3.24 5.86 8.82 70.5 1.2 100 150 95 6 1145 133 46.5 1.76 43.9 7.76 45 65 190 2 1146 134 3.50 5.88 70 210 150 2 1147 124 26.8 4.80 47.0 21.6 65 15 65 2 1148 135 2.47 3.76 58.4 27.3 20 60 205 35 1149 123 3.61 8.83 61.2 27.0 100 65 1140 6 109 MOL TEN-SALT REACTOR PROGRESS REPORT 2,4, FUEL PROCESSING M. R. Bennett G. L Chemical processing of the fuel and blanket salts of a molten-salt reactor will be required in order to remove the neutron poisons and fransuranic elements and to prepare the fissionable materials and the solvents for recycling. The fluoride volatilization process, which was developed for heterogeneous re- actor fuel processing and was used successfully for recovery of the uranium from the fuel mixture (NcF-ZrF4-UF4) circulated in the Aircraft Reactor Experiment (ARE), appears to be applicable to processing fuels of the type now being considered for the molten-salt reactor.! Sufficient laboratory work has been done to confirm that fluorination of the fuel salts LiF-BeF -UF , or Li|=-Bef:2—UF4-'|'|'1F4 results in good recovery of the uranium. Developmental work has also been initiated on the processing of the solvent salt to prepare it for recycling. The salt-recovery process is based on the solubility of LiF-BeF, in highly concentrated aqueous HF (70-100%) c:nd the insolubility of rare- earth and other polyvalent-element fluorides, 2 which results in the recovered salt being decontaminated from the important rare-earth neutron poisons. FLOWSHEET FOR FLUORIDE VOLATILITY AND HF DISSOLUTION PROCESSING OF MOLTEN-SALT REACTOR FLUIDS A tentative flowsheet for application of the fluoride volatility and HF dissolution processes to molten- salt reactor fluids has been prepared (Fig. 2.4.1). In this system the uranium is separated from the salt as UF before HF dissolution of the salt, although the reverse system also seems to be feasible. The LiF-BeF, salt is then dissolved in concenfrated HF (>90% HF) for separation from the rare-earth neutron poisons. The salt is reformed in a flash evaporation step, from which it proceeds to a final makeup and purification step. This last step would perhaps in- volve the H. and HF treatment now believed to be necessary for all salt recycled to a reactor. 1G. 1. Cathers, W. H. Carr, R. B, Lindaver, R. P. Milford, and M. E. Whatiey, Recovery of Uranium from Highly Irradiated Reactor Fuel by a Fused Salt-Fluoride Volatility Process, UN-535. 25, W. Jache and G. W. Cady, J. Phys. Chem. 56, 1106-1109 (1952). 110 D. O. Campbell Cathers Chemical Technology Division Some of the specific features of the process are to be investigated. For example, a high solubility of UF, the voloflllzcn‘lon step, except for final recovery. in aqueous HF would eliminate the need for This possibility appears to be unlikely, however, in view of the low solubility of many tri- and quadri- valent elements. The solubilities of corrosion- product fluorides are to be determined; a method for recovering thorium from the waste salt is needed; and a study is to be made of the behavior of plu- tonium, neptunium, and protactinium in the process steps. Experience in the loborufory and in ARE volatility pilot plant operations? has shown that the recycled UF is highly decontaminated from plu- tonium. EXPERIMENTAL STUDIES OF VOLATILITY STEP A series of small-scale fluorinations was carried out with a 48 mole % LiF-52 mole % BeF , eutectic mixture containing 0.8 mole % UF ,. (MSR fuel will contain from 0.25 to 1.0 mole % UF ,, depending on the operating time.) The eutectic salt was used in- - stead of the fuel salt in order to permit operation at lower temperatures, Fluorinations at 450, 500, and 550°C indicated that the rate of uranium removal in- » creased with the temperature (Table 2.4,1). It is not known whether the increased fluorination time ap- pdrently required for quantitative uranium recovery at the lower temperature was compensated for by a decreased corrosion rate. The thorium-containing blanket salt cannot be pro- cessed at so low a temperature as that used to process the fuel salt. The uranium concenfration in the blanket, however, is very low; it has been esti- mated that with continuous processing at the rate of one blanket volume per year, the blanket salt (LiF-BeF -ThF 71-16-13 mole %) will contain * opproxmafely 0. 004 mole % UF after one year and 0.014 mole % UF , after 20 years.? Fluorinations of 3Communicmion from W. H. Carr at ORNL Volatility Pilot Plant, 4Moiten-Solt Reactor Program Status Report, ORNL CF 58-5-3, p 40. PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED ORNL-LR-DWG 28749A i —»-UF iF - 6 LiF-Bef, SALT | FLUORINATOR U,Th,FP FROM ~ 450°C REACTOR SALT HF, H,0 VAPORS SOLVENT Th,FP - 7 CONDENSER Y DISSOLVER > 30 % HF,< 10% H,0 32°C ~10% SALT SALY SOLUTION SOLUTION SOLID FLASH IN HF =H20 LIQUID EVAPORATOR +Th,FP SOLIDS SEPARATOR 100-400°C CONTINUOUS BATCH TH, FP SOLIDS MOLTEN SOME SALT AND SALT SOLVENT HF,H M2 U, ThF, HF H,0 ! WASTE EVAPORATCR AND FUEL MAKEUP PURIFICATION Th RECOVERY LiF - BeF,-Thf,-UF, TO REACTOR Fig. 2.4.1. Tentative Flowsheet for Fluoride Volatility and HF Dissolution Processing of Molten-Salt Reactor Fuel. two such mixtures at 600°C for 90 min gave uranium concentrations in the salt of 0.0001 to 0.0002 wt %, which were the lowest uranium concentrations ever observed in laboratory fluorinations. Over 90% of the uranium was removed in 15 min. It is concluded therefore that fluorination of uranium from the blanket salt offers no problem. The behavior of protactinium in the blanket salt during fluorination is of interest, although the pro- tactinium is not lost, in any case, since the salt is returned to the reactor. A LiF-Ber-Th F4 (71-16- 13 mole %) mixture containing sufficient irradiated thorium to give a Pa?33 concentration of 5.5 x 10~° g per gram of salt was fluorinated for 150 min at 600°C; no measurable decrease in protactinium activity in the salt was observed. Protactinium volatilization in the process seems to be unlikely, but the protactinium concentration in the blanket of the reference design reactor is ~ 104 g per gram of salt, which is 2 x 104 times larger than the quantity used in the experiments. Further tests will there- fore be required for a definite conclusion regarding volatility. 111 MOLTEN-SALT REACTOR PROGRESS REPORT Table 2.4,1. Effect of Fluorination Temperature on the Fluorination of Uranium from LiF-BeF2 (48-52 Mole %)* Fluerination Uranium in Salt (wt %) Time After Treatment (hr) At 450°C At 500°C At 550°C 0 3.39*+ 5.10 4.91 0.5 1.96 0.20 0.55 1.0 0.39 0.17 0.20 1.5 0.21 0.12 0.06 2.5 0.32 0.11 0.05 *No induction period before uranium eveolution. **5 wt % added; some of the uranium probably pre- cipitated as oxide. SOLUBILITIES OF LiF-BeF2 SALTS IN AQUEOUS HF Initial measurements of the solubilities of LiF and BeF., separately, in aqueous HF solutions indi- cated that both materials are soluble to an appre- ciable extent in solutions containing 70 to 90 wt % HF (Tables 2.4.2 and 2.4.3). In general, LiF is more soluble than BeF, at higher temperatures, but the effect of temperature on BeF, solubility was not definitely established. In fact, it is thought that some of these preliminary values for beryllium solubility are too low. The LiF solubility decreases rapidly as water is added to anhydrous HF, and the BeF, solubility increases from near zero; the solu- bilities are roughly the same in 80 wt % HF, being about 25 to 30 g/liter. The BeF. was quite glassy in appearance and was slow in dissolving. The solubility values reported for LiF in Table 2.4.2, except those at 12°C, were obtained after refluxing HF over the salt for 3 hr. No further measurements were made with LiF or BeF2 alone, since the solu- bilities of the two components together were of primary importance. The BeF, solubility was measured in nominal 70, 80, and 90 wt % HF containing a known amount of LiF that was below the solubility limit. The results of these measurements (Table 2.4.4) show a marked increase in BeF, solubility with LiF present, as compared with the results given in Table 2.4.3. Values obtained for the solubilities of LiF and BeF, in solutions saturated with both salts are 112 Table 2,4,2, Solubility of LiF in Aqueous HF Solutions LiF in Solution {mg/g of solution) Temperature °C) HF Concentration in Selvent (wt %) 73.6 82.3 90.8 96.2 100 12 10.7 31.7 58.4 74.8 88.2 62* 20,6 47* 41.0 37 62.8 80.4 32.5* 91.4 *Reflux temperature. Table 2.4.3. Solubility of BeF2 in Aqueous HF Solution BeF, in Solution (mg/g of solution) Temperature - - °C) HF Concentration in Solvent {wt %) 72.8 78.3 90.8 95.2 100 12 45.8 26.3 9.2 2.8 0.0012 Table 2,4.4. Solubility of Ber in Aqueous HF Containing LiF BeF2 in Solution (mg/g of solution) Temperature LiF Added P , €O mafg et solveny e %) 68.6 79.5 90.0 12 7.0 68.8 12 15.0 42.8 15.6 ~60 7.0 65.8 —60 15.0 8.2 16.5 presented in Table 2.4.5. Comparison with the previous results indicates again that LiF strongly increases the solubility of BeF ,; a plot of the BeF, solubilities (Fig. 2.4.2) clearly demonstrates this. The presence of BeF2 appears to increase the LiF solubility, but to a smaller extent; this effect may Table 2.4.5. Solubility of LiF and BeF2 in Aqueous HF Saturated with Both Salts {10 T | Salt in Solution {mg/g of solution) 100 ; 'é';: / Temperature HF Concentration in Solvent (wt %) i / (°C) 68.6 79.5 90.0 90 ] LiF BeF, LiF BeF, LiF BeF2 Q SOLI-_lSBILITY_7 80 N 14 12 12,9 824 29.9 53,8 64.2 48,2 \ . SATURATED -60 8.5 768 22.6 54.2 46.0 58.2 N be more apparent in more dilute HF where the LiF solubility is lower and the BeF, Because of the increase in BeF2 solubility in the higher HF concentrations when LiF is present, the 90 to 100 wt % HF range was studied further. [t was expected that the more nearly anhydrous system would present fewer complications resulting from hydrolysis. The solubilities obtained in 90 to 100 wt % HF saturated with both LiF and BeF, are pre- sented in Table 2.4.6. Measurements were made at the reflux temperatures indicated, at 12°C, and at ~60°C; in some cases, the measurements at ~60°C were probably taken without the system being at equilibrium. The BeF, solubility is appreciable even in anhydrous HF in the presence of a large concentration of LiF, in marked contrast to the solubility in the absence of LiF. Thus solubilities solubility is higher. PERIOD ENDING OCTOBER 31, 1958 UNCLASSIFIED ORNL—-LR—-DWG 34548 AN 60 \ / 50 40 SALT SOLUBILITY (mg /g of solvent} 30 i X \\ 20 /l/ | 1 l NO Li 65 70 75 80 85 90 95 HF CONCENTRATION IN SOLVENT {wt %) Fig. 2.4.2. Effect of LiF on BeF2 Solubility in HF- H20 Solutions at 12°C. Table 24,6, Solubility of LiF and BeF2 in Solvents Containing 89,5 to 100 wt % HF Salt in Solution (mg/g of solution) Temperature HF Concentration in Solvent {wt %) (°C) LiF 95 98 LiF 60 50 69 78 27 60 87 66 98 46 105 49 *Reflux tem perature. 113 MOLTEN-SALT REACTOR PROGRESS REPORT high enough for a workable processing system are available even in anhydrous HF, provided the LiF concentration can be maintained at the saturation value. The effect of temperature on the solubilities is complicated by the interaction between the two con- stituents, LiF is less soluble at lower temperatures. However, the change in solubility of BeF, with temperature is less obvious, with an apparent in- crease in solubility at =60°C compared with 12°C, especially at the higher HF concentrations. The fuel salt will contain 48 wt % LLiF and 52 wt % BeF ,, sothe highest fuel solubilities should be ob- tainable with the HF concentration for which the LiF and BeF2 solubilities are about the same, that is, about 90 wt % HF. In order to check this ob- servation, solubility measurements were made in 80, 90, 95, and 98 wt % HF solutions for an LiF- BeF2 (48-52 wt %, 63-37 mole %) salt containing about 0.1 mole % ZrF4, 0.2 mole % mixed rare-earth fluorides (Lindsay Code 370), and trace fission products, The salt was crushed; it was not ground to a powder. Average particles were flakes about 10 mils thick and 50 to 100 mils across. The so- lutions were sampled 15 and 60 min after salt addition to permit an estimate of the rate of disso- lution. The results (Table 2.4.7) indicate that an appreciable concentration is reached rapidly, but that the solutions are not saturated, especially with respect to BeF,, in less than several days. These results also demonstrate that the fused salt be- haves similarly to the two components added in- dividually. The aqueous HF solutions of the salt were fil- tered through sintered nickel and radiochemical analyses were made in order to determine the fission-product solubilities. The results, pre- sented in Table 2.4.8, show that rare earths, as represented by cerium, were relatively insoluble in the HF solution and were therefore effectively separated from the salt. The solubility decreased as the HF concentration was increased. The total rare earth {TRE) and trivalent rare earth {except cerium) analyses do not show this separation be- cause of the presence of the yttrium daughter of strontium. No rare earth activity other than cerium was detected in the HF solutions. The slight apparent decontamination from strontium is not understood; strontium is expected to be fairly soluble in these solutions. Cesium is known to be soluble, but the solubilities of other neutron poisons have not yet been investigated. Thus, the rare earths, as represented by cerium, are removed from LiF-BeF, salts by dissolution of the salt in an aqueous HF solution; and strontium and cesium are not. The rare-earth solubility in HF saturated with LiF and BeF, increases from about 10~4 mole % in 98 wt % HF to 0.003 mole % in 80 wt % HF, based on the amount of LiF + BeF, dissolved; the reactor fuel will contain rare earths at a concentration of about 0.05 mole %. Table 2.4.7. Solubility in Aqueous HF Solution of LiF and BeF, from Salt Mixture Containing 63 Mole % LiF and 37 Mole % BeF, Solt in Solution {mg/g of solution) HF Concentration in Solvent {wt %) Time Temperature (°C) 79.5 89.5 95 98 LiF BeF, LiF BeF, LiF BeF, LiF BeF, 15 min 12 28 33 50 44 40 17 28 4 1 hr 29 35 51 50 43 17 22 3 20 hr 31 58 68 74 4 days 32 68* 79 82~ 63 28 70* 22 5 hr -60 21 65 74 102 70 36 71* 22 20 he 12 34 100 76 92 88* 53 60* 22 *Essentially all the component indicated had dissolved and therefore the solubility may be higher than the value given. Insufficient salt was added to some of the solutions to ensure an excess of both components. 114 PERIOD ENDING OCTOBER 31, 1958 Table 2.4.8. Fission-Product Solubilities in Aqueous HF Solutions Activity in Original Salt and in Selution (counts/min/g of salt) P Fission .. Product* Original HF Concentration in Solvent (wt %) Salt** 79.5 89.5 95 98 x 104 x 10* x 104 x 10 x 104 Gross 3 745 230 200 225 250 Gross vy 225 293 213 244 282 Cs ¥ 174 251 196 223 258 Sr 8 104 73 67 72 75 TRE 8 650 97 92 94 101 Ce 510 7.9 3.6 1.3 0.28 Trivalent g*** 105 73 66 74 81.5 *Zr and Nb precipitated from molten salt before this experiment and were not present in significant concentration, **LiF-BeF2 (63-37 mole %) + ™~ 0.2 mole % rare-earth fluoride + trace fission products between 1 and 2 years old. ***This activity later identified as Y90, daughter of Srqo; activity of rare-earth elements 59-71 not detectable in HF solutions. 115 XN AN TAC-TWINEFOQOEPP>PIOMECPE-MPOC-ICETME-DO0OEIOEAIMZIOEO>»PMMOME . Alexander Bettis Billington . Blankenship . Blizard . Boch . Borkowski , Boudreau . Boyd . Bredig Breeding . Briggs . Browning . Campbell . Carr . |. Cathers . Center (K-25) . Charpie . Coobs . Culler . DeVan . Emlet (K-25) Ergen . Estabrook . Ferguson . Fraas . Franco-Ferreira Frye, Jr. . Gall . Gresky . Gregg . Grimes Guth S. Harrill W. Hoffman Hollaender A4 nIr>rom» OO . Savolainen Scott . Seagren . Shipley . Skinner . Snell . Storto A. Swartout . Taboada . Taylor . Thoma . Trauger . Vonderlage . Watson . Weinberg . Whatley . Whitman . Williams . Winters J. Zasler ORNL - Y-12 Technical Library, Document Reference Section Laboratory Records Department Laboratory Records, ORNL R.C. Central Research Library OEEEFPONOAMP>PE-MP>PIMIS->IZS-VOP>VEOQT-A-MATE-FICME T-omC=x44% mOooOmTTNOomMmI 117 EXTERNAL DISTRIBUTION 139. F. C. Moesel, AEC, Washington 140. Division of Research and Development, AEC, ORO 141.728. Given distribution as shown in TID-4500 (14th ed.) under Reactors—Power category (75 copies — OTS) 118