L] CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION ORNL-2799 Reactors-Power ’I MOLTEN-SALT REACTOR PROGRAM 5 QUARTERLY PROGRESS REPORT FOR PERIOD ENDING JULY 31, 1959 OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION r\“~—“f‘\.‘? — e ORNL-2799 Reactors — Povwer TID-4500 (15th ed.) Contract No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT For Period Ending July 31, 1959 H. G. MacPherson, Project Coordinator DATE ISSUED 0CT1 1359 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the MARTIN MARIETTA ENERGY SYSTEMS LIBRARIE s woome memox conassion | IHNHINNEININ 3 4456 D351151 A MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT SUMMARY Part 1. Reactor Design Studies 1.1 DNuclear Calculations and Design Studies The nuclear performance of a two-region graphite-moderated molten- salt breeder reactor was studied. The cylindrical core was approximated by a 5-ft sphere for the purposes of caleculation; the reactor power was 125 Mw(th), 53 Mw(e); the fuel and blanket salts were each processed at the rate of 10 ft®/day; and the total fuel volume was L41.3 f£t®, The op- timum case was found to have an initial inventory of 42 kg of U232 and a doubling time ranging from 10 to 26 years, according to the degree of pessimism with which the values of m for U®32 in the near epithermal range were selected. 1.2 Component Development and Testing Salt-lubricated hydrodynamic Journal bearings with two axial grooves 180° apart were tested to failure. The first of these bearings seized after 33 hr and 22 start-stops, running at 1200 rpm and 200 lbf loading. Examination and analysis of these bearings after test indicate they do not have dynamic stability equal to previously tested bearings having three equally spaced grooves, Three further tests were conducted with bearings containing three equally spaced helical grooves, These tests were terminated because of impending failure as the load was increased. Analysis indicated that end leakage from the helical grooves interfered with the inward flow of liquid required to maintain a hydrodynamic film and suggested that flow restrictors at the ends of the grooves might remedy the condition. A test incorporating a flow restriction showed this to improve the load-carrying capacity of the bearing by a factor of 3. Fabrication of a pump incorporating a three-axial-groove hydrodynamic bearing was completed. This pump is being assembled and checked prior to operation with molten salt. The modified Fulton-Sylphon bellows seal operating in a PK type of centrifugal pump has accumulated 14,830 hr. The MF type centrifugal pump has continued to operate and has logged 17,664 hr (more than two years) of continuous operation. The small frozen-lead-seal pump operating with 1200°F molten salt has logged 9400 hr of continuous, maintenance-free operation. Construction of the Remocte Maintenance Demonstration Facility was completed on schedule. This facility incorporates a one-third-scale mockup of a typical molten-salt reactor together with remote manipulators and re- mote viewing equipment, The system 1s in the process of being checked out prior to operation with molten salt; then the subsequent disassembly and reassembly by use of remote maintenance techniques will be conducted. The objective is to uncover and find solutions to the problems of remotely maintaining a reactor system after it has operated at power. One long-term corrosion test loop was terminated after a year's opera- tion, leaving 1k test loops in operation. Eleven of these are constructed of INOR-8 and three of Inconel. The terminated loop contained specimens of graphite in contact with the hottest salt in order to determine the com- patibility of unclad graphite with the salt. Examination is not yet com- pleted. The second in-pile loop test was terminated because of pump seizure after approximately 1000 hr operation. 1.3 Engineering Research The enthalpy and viscosity of a high~thorium-content blanket salt mix- ture have been experimentally determined. The viscosity was observed to decrease somevhat more rapidly with temperature than do the viscosities of a number of other mixtures with high percentages of ThF, or UF,. Over the temperature range investigated, the viscosities of the several mixtures vere the same to within 10%. Preliminary measurements of the enthalpy of this blanket salt indicate a heat capacity consistent with the value pre- dicted by a correlation based on experimental measurements with other mix- tures containing various amounts of NaF, LiF, BeFs, UF4, and ThF,. A re- determination of the viscosities of two low-UFs-content mixtures with the modified viscometer resulted in lower values and altered temperature de- pendence for both salts. Assembly of the system for the study of inter- facial film formation and heat transfer with BeFs-containing salts is con- tinuing. Two schemes for heating the test sections are being considered. 1.4 Instruments and Controls The Q-1717-54 level probe tested previously was examined metallurgi- cally and found to be in excellent condition. A new series of tests was started with two level elements made of Inconel being tested in fuel 130 at 1200°F. The spans of both probes have decreased throughout the tests, probably because of an increase in resistivity of the fuel. Part 2, Materials Studies 2.1 Metallurgy Examinations of eleven INOR-8 thermal-convection loops, which had operated with fused fluoride mixtures, were completed; nine of the loops had operated for one year, and each of the other two had operated for 1000 hr. Only one of the loops experienced more than a very light attack, in the form of surface pitting or shallow surface roughening in hot-leg sec- tions, Twelve Inconel loops were examined; nine had operated for one year and three for 1000 hr. All except one showed attack in the form of inter- granular voids to a depth ranging from 1 to 18 mils. An INOR-8 pump-loop test to determine the compatability of INOR-8, graphite, and a fluoride fuel was terminated after one year. The loop wall showed light surface roughening and pitting to less than % mil; there was no evidence of car- burization; the fuel salt showed the expected chromium pickup but was un- affected by impurities in the graphite. A 6000-hr test and a 5000-hr test of INOR-8 exposed at 1300°F to a system containing graphite and fuel 150 have been completed, and no evi- dence of carburization was observed in the INOR-8 by metallographic, mechanical, and chemical examinations. Inconel and INOR-8 specimens ex- posed to graphite-sodium systems for 4000 hr at 1L00°F were carburized to a depth of 30 to 40 mils and showed a loss in ductility both at room tem- perature and at 1250°F. vi Fuel 30 (NaF-ZrFs-UF4, 50-46-4 mole %) and fuel 130 (LiF-BeFp-UFa4, 62-37-1 mole %) penetrated to 58% or more of the accessible void volume . in AGOT graphite specimens when 150 psig pressure was applied in a 100-hr exposure at 1300°F (704°C). Initial gettering and flushing tests to remove oxide contaminants from graphite specimens with fuel 130 in a 20-hr exposure at 1300°F (70L°C) appeared to reduce the amount of uranium oxide precipitation normally observed in tests with nongettered graphite. The program for evaluating the mechanical properties of INOR-8 has been reviewed, and the significant results obtained from this program are surmarized with respect to the variations observed in testing various heats of material. It has been found that two heats, SP-16 and SP-19, exhibit very similar mechanical properties, while two high-carbon heats, 8 M-1 and 1327, are stronger in tensile tests but have creep properties similar to those of SP-16 and SP-19. Initial studies to determine the claddability of prefabricated porous nickel for use as a leak-detecting component in a triplex heat exchanger tube have indicated that the material can be bonded to Inconel with reten- tion of core porosity. An attempt has been made to fabricate a 9-ft triplex tube by utilizing six 18-in. lengths of prefabricated porous nickel tubing as the core, Although evaluation of this tube is incomplete, radiographic inspection revealed slight separations between two of the five core-to-core joints. It has not yet been determined at which step during fabrication these flaws developed. An INOR-8 weld-metal composition containing 2% niobium has exhibited good ductility at 1500°F. A large heat of this composition is being pre- pared for further inert-arc studies and for metallic-arc studies. The ductility at 1500°F is quite sensitive to small variations in weld-metal chemistry, and careful chemical control should be observed. 2.2 Chemistry and Radiation Effects Solid-solution formation, resulting from the interchangeability of ThF, and UF,, plays an important role in the freezing behavior of breeder fuels, as demonstrated in the Na¥F-ThF,-UF, phase diagram. The existence vii of a quasi-binary system between NaF.BeF, and ThF, furnishes important clues to some of the controlling phase relationships in the NaF-BeFs-ThF, ternary system. Groundwork for a phase study of PuFs fuels is provided by cooling curves on a series of compositions from the NaF-PuF,; binary system. Freez- ing-point depressions for NaF, as influenced by the size and charge of solute cation, are being measured as an aid to theoretical interpretation of molten-fluoride behavior. Solubility measurements for Xe and COs yield information of both theoretical and applied interest; the temperature coefficient for CO- changes sign. Fission-product behavior, particularly as related to fuel reprocessing and recovery, has been studied in experiments on the solubility and pre- cipitation of rare-earth fluorides. Also, some additional investigations of the reactions of UF, with oxides are of interest in connection with re- processing and with exposure to oxide contamination during maintenance, Preliminary indications are that higher ThF, concentrations in breeder fuels diminish or perhaps eliminate the oxide sensitivity exhibited by LiF-BeFo-UF, melts. Solubility measurements on NiFp throw new light on the thermodynamic behavior of this corrosion product, but do not resolve the mystery of why the apparent activity coefficient of NiFp in fuels is so high. Practical aspects of corrosion behavior have been followed with periodic sampling of corrosion-~test loops, while systematic approaches to related problems have been advanced with the aid of varied techniques such as vapor-pressure studies and radiocactive tracers. In a study of the compatibility of graphite with fuel, no deletecrious or unfavorable effects were encountered with samples of an impervious graphite which were exposed to circulating fuel for one year in a corrosion- test loop. Other experiments with graphite, giving seemingly diverse and erratic results, can probably be correlated by taking into account the effects of oxide and of the pressure required for forcing a nonwetting liquid into the particular pore spectrum of a graphite sample. viii Work on the preparation of purified materials proceeded at an accel- erated pace. A thermal-convection loop was assembled and loaded for in-plle opera- tion. Two test assemblies of the capsule variety, containing fissioning fuel in INOR-8, have operated in the MIR for three months. 2.3 Tuel Processing Two modifications of the HF dissolution process for purifying the LiF-BeFs carrier salt are being investigated. One concerns the dissolu- tion of uranium along with the LiF and part of the BeFs, which are soluble in anhydrous HF. Addition of ClFs to HF oxidizes UF, to UFg, which has appreciable solubility in anhydrous HF. Dissolution of uranium and LiF together would permit recovery of all the valuable isotopically enriched material in a single step, which would result in considerable saving in equipment and operation, since the volatility step for uranium recovery from the fuel would be unnecessary. The other modification involves de- contamination of LiF from fission products that are soluble in HF, such as cesium, and therefore are not separated in the initial dissolution. Partial evaporation of a 90% HF solution of fuel salt yields a precipitate containing nearly all the lithium and a solution containing most of the cesium. CONTENTS S.[.WARY " EEEEEEENENENNENRE I I R R N R RN I N YR I N I IR I R B RN I BN B L B B L A 1.1. 1.3, 2.1. PART 1. REACTOR DESIGN STUDIES NUCLEAR CALCULATIONS AND DESIGN STUDIES s.oceveececeacssansnns Nuclear Performance of a Two-Region, Graphite-Moderated, MOlten—Salt Breeder Reactor O * & 8 8 0 & 5 0 8PS ST 0 B 00 E S EE N TSN COMPONENT DEVELOPMENT AND TESTING ceceveseservcasoacsessnens Molten-Salt-Lubricated Bearings for fuel PUumps .eeeceseceses Hydrodynamic Journal Bearings ..ecceceesssscessssssrsncns Hydrodynamic Thrust Bearings .c.eeiecesscscscesssssscssnns Test of Pump Equipped with One Molten-Salt-Lubricated Journal Bearing itecseescssonsesssccsscocsssscassssnssascs Bearing Mountings eeseeececsceescsscccssncesssosnssssssnacs Self-Welding of INOR=8 cveverecsssnsosscassscansonnsonsss Mechanical Seals fOr PUNDS teeseoercsccccosscssasvesnssssasas Pump Endurance Testing eeeecceceecetocasstcsssasssssnennnsses Remote Maintenance Demonstration Facility eseeescececssancas Frozen-Lead Pump Se€8l .tieeececrsasccccsarsssscssrsssssrsssances Mechanical Joints for the Reactor System .ceeeiecereccssencas Design, Construction, and Operation of Materials-Testing LOOPS ........I..III....l...Ill.....'..I.....l“.....ll... Forced-Circulation Corrosion LOOPS seecesssssccossssccss In_Pile LOOPS " EEEENENRENEREREN N IE TR BN BNEN BN S N B N I B R N N RN N R N IR N ENGINEERING RE SEA-RCH $ 6 9 4 & 9 4 5 ¢ 8 0 @ 0 8 0 e e BB e S s e e * 08 & & 0 00 Physical-Property Measurements .....c0ceceee.. cecsaneseveane Viscosity ..... Cesearceasseannens et esessesasesscsnsesane e Enthalpy and Heat Capaclily .eececertencercstsscnsonnsnnses Thermal EXPansioll seeseccsececsrsosessssssscsssrssssssscasss Heat-Transfer StUdies t.eieceessssssssrssasssssancsseanvsssssssss INSTR[H'ENTSANDCONTROLS 8 9 6 0 0 S8 0 B S B S O P BT S U PP A S S e RS S seDse MOlteIlflsalt—FU.El Level Indic&tors % 8 9 8 4 665 0PSSO S SRS PART 2. MATERIALS STUDIES mTALL[]‘RGY M EEEEEEEEEE e ar ar B R A B RN R R RN B I R B A B A B A A A e Dynamic Corrosion Studi€s ..eveeeereiserecarcracstsnscnasnns INOR"8 Themal—conveCtiOH LOOPS 9 ¥ 6 8 66 0 P PSS BSOS BB LSS iii Il’lCOl’lEl T].‘lemlal_convection LOOPS S 2 8 2 B PN P TP O N NPT ERESPRETETS INOR-8 FOI’CEd-CiI‘CUl&tiOI’l LOOPS * 9 50 00008000 ® e e 009000 General Corrosion STUudies seveiservecsrssssnesenss I Carburization Tests on Inconel and INOR-8 in Systems Containing Fuel 130 and Graphite ..vescevrecsescncasoss Sodium-Graphite Carburization TeSTS cesieccseersasaccassoee Penetration of Graphite by Moliten Fluorides cecseescasess Removal of Oxide Contaminants from Graphite ..ieesevecases Mechanical Properties Of INOR-8 tieeeeeneeeorennesesonssnsoas Materials ceesecesoesscasceces T e secasrtsensaerearetesnu s Tensile Properties teieeeecssesssosesssssssncssasesssscas Creep TestsS teeeeevncesernconsnns e treeerresarsearssaanas Relaxation TesStsS eeeieeeccosrsesssssassssccssssssnsnaas crrae Dimensional Instabllity seeeeserescccsssssscsssonssanssnns Fatigue STUudies seeseeessossassssonsssncssossssossssoocsnsa Carburizalion sieeseessssesovesssssssnssssassssacnnanssns . Materials Fabrication STudi€s teisesecnsarseasss ceissersaseas Triplex Heat Exchanger TubIng seeeeeevoesssccrscaserssssons Welding and Brazing STudiesS seeeseecaceasss veeanne ceesastnasans Welding Of INOR—B LR I T T I N T I I I S I B I S N I R R R R Y R R A A N L L R R ) CHEMISTRY AND RADITATION EFFECTS eieeeeenroccnnossassnsanansas Phase Equilibrium Studies ...ceieceaeessessosscsscssnsnacsas The System NaF-ThF -UF4 +eeeeteeiscescesosscssensonsasnse The System NaF-BeFo-ThF, veveeirtieetscearonsensassansnsss The System NaF-PuFg .euieieetesressseacsscesessonsensensas Cryoscopy 1in NaF ..eeveeesevecsas cesesseessesessuransso e Gas Solubilities in Molten Fluorides sisvecssssosssenssssocss Solubility of Xenon in LiF-BeFo seeeveverovsnsscscsscases Solubility of COz in NaF-BeFo ceveeenes e trcearssssassses Fission-Product Behavior seeeeseesoesesssesassosssssssssanss Solubility of Rare-Earth Fluorides ...eecesscsescsesssesas Precipitation of Rare-Earth Fluorides by Cooling ...ee... Reactions of UF, with Oxides in Molten Fluoride Solvents ... Reaction of UF, With BeD tiiieeeienreerennacsssnnsnsnnses Reaction of UF, with Strontium OxXide seeeerervcaserssonse Reaction of UF, with Structural-Metal Oxides .veeveveeens Reactions of UF4 With Alr ..ieeereesevsernsnssesacsoscnnas Chemistry of Corrosion ProcessSes tieereesvsrsaccerosecsosssns Solubility and Thermodynamic Properties of NiF, in LiF-BeF s tiveeeseesesessossssssossssssessssssssessnssnss Activities from Vapor Pressure Measurelments ceeeseseesess Tracer Analyses for Determining Efficacy of Reducing AFeNlS teverrerreennssorsssssoscssessesnssnsssssasascanss Analyses of Corrosion-Test LOOPS eseeescersarsessnccnsscee Stability of Zirconium Monochloride In Alr cievieseacasas 2.3, Xi Graphite compatibility Studies e 0 & 0 & & e 4 B A & S & P B S B RSB SN SRS Long-Term Loop Tests of Graphite with Circulating Fuel ® & & & & % & 0 B 2 PSS S S S e e DS eSS NN PO E AR S FEREYE Permeability of Graphite by Molten Fluorides .eeeeecscees Preparation of Purified Materials .....cecececcrnncncscocncas Purification, Transfer, and Service Operations ..,eeecees. Reactions of Stannous Fluoride .ceeesessesescescssscassnns RadiationEffectS S 9 0 Q8 P9 PSS 9P E ST e S e e TS e G RS e e INOR-8 Thermal-Convection Loop for Operation in the LITR o 8 8 & 2 9 ¢ ¢ 8 5 0 % d e S S S S S E e E SO E S PSSP H e PP Y PSP In-Pile Static Corrosion TesSTS ceeesesssscsssscsacsssssess F.UEL PROCESSING LI I R I O L I N B B SN N I R S R I A A I I A B R IR I B I A A A UFB Solubility in :H:E""C:LFS LI L I BE N 2N B R I R R U B I I B R Y B I Separation of LiF from CsF tuieeeeeeeersssescscscsscsasans PART 1. REACTOR DESIGN STUDIES 1.1 NUCLEAR CALCULATIONS AND DESIGN STUDIES Nuclear Performance of a Two-Region Graphite-Moderated Molten-Salt Breeder Reactor A typical conceptual arrangement is shown in Fig. 1.1.1. The main part of the core is fabricated from a single cylinder of graphite about 5 ft in diameter and 5 ft long. The end pieces are also prepared by machining solid pieces, and the three sections are held together by means of rods of INOR-8 or other suitable alloy. The fuel salt passes through the core and is circulated to an external heat exchanger. The blanket salt surrounding the core is maintained at a slightly higher pressure, and contains the fluorides of lithium, beryllium, and thorium in the same pro- portions as in the fuel salt. A proposed processing scheme is shown dia- grammatically in Fig. 1.1.2. In this scheme, fuel salt is withdrawn from the fuel circuit and passed through a tower wheére it is exposed to fluorine gas (the fluoride volatility process). Uranium is removed in the gas stream as the hexafluoride. The barren effluent salt, containing the nonvolatile fission products and Pa233, is sent to the blanket salt cir- cuit, where it is diluted by a factor of approximately 10, The blanket salt is treated with CeF5; to remove rare-earth fission products. A side stream from the blanket salt circuit is also treated with fluorine to re- cover the U=33 pred therein. This, together with the uranium recovered from the fuel salt, is combined with blanket salt that has been treated with CeFs to reconstitute fuel salt, By this scheme, rare-earth fission products are removed from the system, and other fission products (alkali metals, alkaline earths) and Pa=°3 are substantially diluted. Further, any leakage of blanket fluid into the fuel circuit can be compensated conveniently by increasing the rate of processing of the fuel stream with fluorine, since this operation returns barren salt to the blanket system, In a previous study, the fuel channels through the graphite core were 2.66 in, in diameter and vere arranged in an 8-in. triangular lattice.l IMSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 10. UNCLASSIFIED ORNL-LR-DWG 37258 FUEL BLANKET ? T, | re] ZZZIIZIZT! v | P 7 ra TITIIT %7 777 INOR - 8 FUEL * GRAPHITE 8L ANKET Fig. 1.1.1. Graphite-Moderated Two-Region Molten-5alt Thorium Breeder. UNCLASSIFIED ORNL—LR-DWG 40827 STEAM 10 = I | | TURBINE HEAT EXCHANGER BOILER - [] INTERMEDI ATE FUEL SALT SALT RECONSTITUTED - FROM FUEL SALT BLANKET CONDENSER : HE AT EXCHANGER BLANKET SALT —— CORE . 7 BLANKET SALT @_ PURIFIED SALT FUEL BARREN SALT BLANKET SALT SALT CONTAINING Pa CONTAINING U233, Pa, : AND FISSION AND FISSION PRODUCTS HF PRODUCTS , Cefy UFg UFg ‘ TREATMENT ™ REDUCTION FLUORINE TREATMENT H 1 1 RARE EARTH 2 EXCESS F FLUORIDES u2*3 10 SALES 2 Fig. 1.1.2. Processing Scheme for a Two-Region Graphite-Moderated Molten-Salt Breeder Reactor. The estimate of the performance was based on the assumption that the molten salt fuel would penetrate the solid graphite to a negligible ex- tent. In the present study, the consequences of fuel penetration were considered. Optimally, the fuel salt occuples about 10% of the core volume, If 1 vol % of the graphite should be penetrated by the fuel, approximately lO% of the fuel in the core would be incorporated into the graphite. Since the thermal neutron flux in the moderator is greater than in the fuel channels, somewhat more than 10% of the total fission heat would be released in the graphite, with the result that thermal stresses in excess of the maximum allowable (1000 psi) might be set up. Figure 1.1.3 shovs values of the estimated temperature rise in the graphite for a fuel pene- tration of 1 vol %, a fuel-channel volume of 10%, various fuel-channel radii, and various assumed heat release rates in the moderator. It was estimated that a temperature rise of LOO°F in the graphite could be toler- ated, and it is seen that the fuel-channel radius may not exceed 1/2 in. for a heat release rate of 15 Mw, which corresponds to a reactor powver of 125 Mw. The corresponding lattice spacing is 3.0 in. The nuclear performance of the reactor in equivalent spherical geometry was then studied by means of the multigroup Oracle programs Cornpone and Sorghum, Because of the smallness of the lattice parameters, the core was assumed, for the purposes of nuclear calculation, to consist of a homogeneous mixture of fuel salt and graphite. The basic core salt was 65 mole % LiF, 31 mole % BeFo, and 4 mole % ThF,. A core equivalent to a 5-ft sphere was selected for study, and a blanket 30 in. thick was specified. Although the processing scheme requires that the fuel and blanket salts have the same thorium concentration, the calculations were performed throughout using 1% mole % thorium in the blanket. The neutron leakage from the blanket can be matched at lower thorium concentrations by using a thicker blanket, but parasitic absorptions in lithium, beryllium, and fluorine in the blanket are somewhat underestimated at the higher con- centration. The error, hovever, 1s small. No correction for fuel inlet and outlet manifolds was made. The total reactor heat rate was set at UNCLASSIFIED 4 ORNL-LR-DWG 40617 10 | T RATE OF HEAT RELEASE IN MODERATOR —— / 15 Mw 5 ~~ / A 10 Mw o 2 e N - 10’ 7/ 7 s 77 7 W / / / = / / / A M & . / / ) d e 5 /[ / pd : Ry = = 10" f // 7 1/ / A4 / 10 0 1 2 3 FUEL-CHANNEL RADIUS (in.) Fig. 1.1.3. Moderator Temperature in a Two-Region Graphite-Moderated Molten-Salt Breeder Reactor. Core diameter, 5 ft; reactor heat, 125 Mw; fuel volume fraction in core, 0.10. 125 Mm(th), with a plant factor of 0.8. It was estimated that the various components of the external heat-removal circuit would entail the volumes of fuel listed in Table 1.1.1. Table 1.1.1. Fuel Volumes in Molten-Salt Breeder Component Fuel Volume (ft3) Core inlet 3.8 Piping 10.2 Pump bowl 1.5 Heat exchanger 16.8 Surge tank 2.2 Total 3L4.5 The power density in the external circuit is thus about L Mw/fts, which is moderately high performance but not soc high as specified for certain other molten-salt systems (e.g., the ART). The volume of fuel in the core is 8.2 ft3, bringing the total to L2.7 ft°. In the Scrghum program, provision is made for the treatment of the processing of only a single group of fission products. Accordingly, the performance of the scheme was assumed to be approximated by the complete removal of the fission products from the fuel-salt processing stream; the return of Pa®33 and fission products in the fuel-salt makeup stream was neglected, as was also the absorption of neutrons by fission products and Pa®33 in the blanket. Fluoride volatility processing rates for fuel and blanket systems were set at 10 fta/day each, which resulted in processing cycles of 4 and 40 days, respectively. The effect of varying the thorium concentration in the fuel salt is illustrated in Fig. 1l.1l.4 for a core having a fuel volume fraction of 0.10. It is seen that the doubling time reaches a rather broad minimum, near 13 years, between 1 and 4 mole % ThF,. Judging from the curves, the optimum thorium fluoride concentration is about 2.5 mole %. The corresponding UNCLASSIFIED ORNL-LR-DWG 40618 ThF, IN FUEL SALT {(mole %) 1 4 7 43 30 I I 150 — 3.5 \o 25 \ 125 ] 3.0 SPECIFIC POWER [ J 20 \ // 100 —4 25 \ o 2 m o 2 — [ ] > ~— 8 \ [ o =) > = g UEJ \ ./ / E 2'0 g). = 45 e / 75 2 z Z \e _~ ° W p ) DOUBLING TIME _ 1.5 = 3 o : s g — z Q 10 // \\ © S ./ ™ § —1 40 /%‘.AL INVENTORY /. 5 25 —1 0.5 0 0] — 0 0 10 20 30 40 50 (x10™%H RATIO OF CONCENTRATION OF THORIUM ATOMS TO CARBON ATOMS IN CORE, NTh/NC Fig. 1.1.4. Nuclear Performance of @ Two-Region Graphite-Moderated Molten-Salt Breeder Reactor. Core diameter, 5 ft; reactor heat, 125 Mw; reactor power, 52.5 Mw; plant factor, 0.8; total fuel volume, 41.1 ff3; fuel fraction in core, 0.10; fuel and blanket processing rates, 10 ff3/day. 10 U233 and the specific power (average) initial inventory is about 37 kg of is about 2.7 Mw/kg. However, parasitic losses to lithium, beryllium, and fluorine and leakage from the blanket become excessive when the thorium is too dilute. A salt having 4 mole % ThF, was selected as a reasonable compromise. The effect of varying fuel volume fraction at constant thorium con- centration is shown in Fig. 1.1.5. The lower curve 1s based on initial conditions (i.e., initial breeding ratio obtained in the clean core) and was computed by using the approximate equation 1.05 n X inventory (kg of U=32) Toreeding ratio - 1) X reactor heat (Mw) doubling time (years) = The upper curve was based on Sorghum lifetime calculations covering periods up to 20 years of operation. It 1s seen that estimates based on initial conditions are grossly misleading. Not only are the doubling times about half those computed in the lifetime studies, but the trend is downwvard with increasing volume fraction of the fuel in the core; whereas, when the accumulation of fission fragments and uranium isotopes is taken into account, the doubling time is a minimum when the volume fraction is about 0.125. An extract from the predicted performance of the optimum core is given in Table 1.1.2,where inventories, neutron balances, ete., for the initial state and after 20 years are given. It is seen that the initial inventory of U233 45 about 42 kg, giving an average specific power of about 2.5 Mw/kg. Over a period of 20 years, the increase in inventory required to override fission product poisons and uranium isotopes amounts to 18 kg. Also, in that period the excess production amounts to 78 kg of U233 giving, in terms of the initial inventory, an average doubling time of 11 years. This doubling time was calculated by using energy-dependent values of 1 for U%32 compiled by J. T. Roberts® of the Molten Salt Group in 27 1, Roberts and L. G. Alexander, Cross Sections for the Ocusol-A Program, ORNL CF-57-6-5 (June 1957). 1 UNCLASSIFIED ORNL-LR-DWG 40615 ~ 20-YEAR PERICD )\n e O™ 12 O v \\ Q [« 1] z Q w = AN k- \ g s} INITIAL CONDITIONS T QO a \ o e —C\ o 6 4 2 0 0 0.10 0.20 0.30 VOLUME FRACTION OF FUEL IN CORE Fig. 1.1.5. Doubling Time of a Two-Region Graphite-Moderated Molten-Salt Breeder Reactor, Core diameter, 5 ft; reactor heat, 125 Mw; reactor power, 53 Mw; plant factor, 0.8; 4 mele % ThF4 in fuel salt; fuel and blanket processing rates, 10 ft>/day. late 1956. Newer data are now available, and these have been analyzed by, among others, J. Chernick® and A. M. Perry.® A comparison of the three sets of n values 1s shown in Fig. 1.1.6. By using the Chernick and Perry values, the doubling time averaged over the first 20 years of operation was estimated to be 10 and 29 years, respectively. The nuclear performance corresponding to the three sets of n values is shown 1n Fig. 1.1.7. It should be noted that the excess production is negative during the early part of the cycle and that the time required to reproduce 3J. Chernick, Prepublication oral communication, June 1959. “*A. M., Perry et al., A Study of Graphite-Moderated Th-U=33 Breeder Systems, ORNL-2666 (1959). 12 Table 1.1.2. Nuclear Performance of a Two-Region Graphite-Moderated Molten-Salt Breeder Reoctor Core diameter, 5 ft Fuel processing rate, 90 times per year . Power, 125 Mw (th) Moderator, graphite Plant factor, 0.8 Blanket, 30 in. thick Fue!l volume, 42.7 f3 Fuel volume fraction in core, 0.125 Fuel geometry, 1-in. channels in 2.7-in. triangular lattice Before Operatian After 20 Yeors of Operation Inventory (kg) Absorption Ratio* Inventory (kg) Absorption Ratio* Fissionable isotopes U233 (fuel) 41.5 1.000 49.5 0.907 U233 (blanket) 4.2 u23s 6.0 0.093 Fertile Isotopes Th232 (fuel) 613 0.362 613 0.316 Th232 (blanket) 21,500 0.785 21,500 0.709 u234 3.3 0.096 Fuel carrier and moderator Be?, C 0.057 . 0.043 F, Li (99.999 % Li”) 0.030 0.022 Fission products 0.442 0.001 - Parasitic isatopes (U236, etc.) 8.9 0.014 Miscellaneous - PaZ33 (fuel) 1.41 0.010 Pa233 (blanket) 3.27 Core vessel, blanket salt, leakage 0.033 0.030 Cumulative power generation, Mwyr 0 2000 Neutron yield, 7 2,267 2.241 Total fuel inventory, kg 41.5 59.7 Cumulative net burnup, kg 0 -~ 959 Excess 233 production, kg 0 78 Regeneration ratio 1.147 1.109 *Neutrons absorbed per neutron absorbed in fuel. the initial inventory, for the first time, is 12, 11, and 33 years, re- spectively. Later, after the fission product concentration has stabilized, the excess production rate is higher. Doubling times based on this higher, "stable" rate are 10, 9, and 25 years, respectively. It 1s clear that the estimated performance is sensitive to the assumed values of n for U=33 . However, it seems probable that doubling - times of 25 years or less can be achieved in externally cooled molten- salt reactors. UNCLASSIFIED ORNL-LR-DWG 40620 2.6 ................... 2.5 2.4 I 2.3 : I -l I ...................................... 2.2 Eta 2.1 —— CHERNICK - RCBERTS — — PERRY 2.0 =5 20 19 18 47 6 15 14 13 12 N 10 9 8 7 ) 5 4 3 2 1 0 LETHARGY Fig. 1.1.6. Nuclear Data for U233 Used in Estimation of Nuclear Performance of Molten-Salt Breeder Reactors. UNCLASSIFIED ORNL-LR-DWG 40621 100 90 NET ACCUMULATIVE EXCESS u??? (CHERNiCy 80 _— 70 //// / o on _______.—-—-‘ g ~ TOTAL INVENTORY (CHERNICK AND ROBERTS) 7 / /%ACCUMULATIVE EXCESS U?33 (ROBERTS) \ o / NET ACCUMULATIVE EXCESS U233 (PERRY) [~ / . ] | -1 0 2 4 6 8 10 12 14 16 18 20 REACTOR OPERATING TIME (years} Fig. 1.1.7. Nuclear Characteristics of a Two-Region Graphite-Moderated Molten-Salt Breeder Reactor. Core diameter, 5 ft; reactor heat, 125 Mw; plant factor, 0.8; fuel volume fraction in core, 0.125; blanket width, 2.5 ft; ThF4 in fuel, 4 mole %; ThF4 in blanket, 13 mole %. 15 1.2 COMPONENT DEVELOPMENT AND TESTING Molten-Salt-Lubricated Bearings for Fuel Pumps Hydrodynamic Journal Bearings Investigations on journal bearings made of INOR-8 were continued throughout the quarter. Five tests were conducted in molten salt 130 (LiF-BeFo-UF,, 62-37-1 mole %), in which two types of bearing-groove con- figurations were investigated for application to a vertical-shaft centri- fugal fuel pump. Previous tests had been made with bearings having three equally spaced axial grooves. These bearings could carry acceptable radial loads in each of three known directions 120° apart, but their load-carrying ability was sharply reduced at other radial load angles. It was therefore decided to test bearings having two axial grooves 180° apart, as shown in Fig. 1.2.1, since this configuration should be capable of accepting reason- able loads over a wider range of angles. Tests 10 and 11 were conducted with a bearing containing two axial grooves. The same bearing and journal were used in each test except that the radial clearance measured at room temperature was increased from 0.005 in., for test 10 to 0.007 in. for test 1l. - In test 10, the bearing was inadvertently operated for 5 hr without load (at speeds up to 1800 rpm), and subsequently was operated for %% hr with a EOO—lbf load at 1200 rpm, during which time 22 start-stops were made. Termination resulted from seizure of the bearing. Postrun examina- tion of the test surfaces revealed a wear pattern that would be representa- tive of either vibration or film whirl. Test 11 covered a period of 6 hr, of which 2 1/2 hr was devoted to performance at various combinations of speed and load, and the remaining 3 1/2 hr was devoted to steady-state operating conditions of 200 lb_ and 1200 rpm. This test was also term- inated as a result of seizure.f The results of postrun examination were similar to those in test 10. The test bearing configuration with two axial grooves apparently does not have adequate dynamic stability. Three further tests were conducted with bearings containing three equally spaced helical grooves as shown in Fig. l.2.2. A1l the tests were 16 UNCLASSIFIED ORNL-LR-DWG 38966A Fig. 1.2.1. Bearing with Two Axial Grooves. UNCLASSIFIED - ORNL-LR-DWG 38968A Fig. 1.2,2. Bearing with Three Helical Grooves, 17 conducted at 1200°F and 1200 rpm, and the radial clearance in each test was 0.005 in.,, measured at room temperature. The groove cross-sectional dimensions were the same as those used with previous tests of axial grooves. In test 12, the load was applied in a direction such that the minimum film section should have crossed a groove at or near the mid-point of the bear- ing length. The bearing was subjected to various loads from 10 1b_, to 175 T 1b. for 26 hr. The test was terminated as a result of impending seilzure f when the load was increased from 150 1b_, to 175 1b Just prior to the T il impending seizure, the test had operated for 2 hr at 150 1b Postrun examination of the bearing surfaces revealed considerable wgar near the lower end and very slight wear near the upper end. Test 13 was conducted with the bearing and journal used in test 12 but with the surfaces cleaned up with hand tools, resulting in little or no change in the original 0.005-in. room-temperature clearance. The load was applied to the opposite side of the bearing from that in test 12. With this direction of loading, the minimum film section should have crossed the ends of two separate grooves located at opposite ends of the bearing. The test consisted in operation at various loads from 10 1b. to f 150 1b The test was operated for 2 hr at 150 1b_ load and was then stoppeg to investigate start-and-stop ability. Upin restarting, the bear- ing began to show evidence of impending seizure and was thus terminated. The total operating time was 23 hr. Postrun examination revealed that the failure was quite similar to that in test 12. The bearings having helical grooves that were used in tests 12 and 1% had inferior load-carrying capacity compared with bearings having three axial grooves. An analysis pointed out the possibility that the end leak- age through the helical grooves interfered with the flow of liquid required by the bearing hydrodynamic film. It was suggested that the addition of a flow restriction on the leakage end of the grooves would remedy the situation. In test 14, a restriction to flow at the discharge end of the grooves, consisting of a thin disk having an inside diameter 0.010 in. larger than the inside diameter of the bearing, was installed. Operation of this bear- ing covered a period of 215 hr, and loads up to 500 lbf were applied to 18 the bearing. During the period of operation, 260 start-stop operations were satisfactorily performed, and the test was terminated on schedule. Visual inspection of the surfaces showed signs of slight wear. Restricting the flow of molten salt lubricant from the grooves had the effect of trip- ling the load-carrying capacity of the bearing having three helical grooves at identical operating condition. This type of bearing should be able to accept loads in any radial direction. Further tests will be conducted in the next quarter with carburized INOR-8 journals operating in noncarburized INOR-8 bearings at temperatures above 1200°F. These investigations will be made in order to test the im- provement in start-stop ability of carburized Jjournals compared with non- carburized journals. Other tests are planned for bearings having three helical grooves but with smaller groove cross sections than were used in the bearings reported above. Hydrodynamic Thrust Bearings Assembly of the thrust bearing tester was completed, and one test was performed with fuel 134 [62 LiF, 36.5 BeFs, 1.0 ThF, 0.5 UF,a (mole )] at 1200°F. The runner was rotated at speeds up to 1400 rpm, and the load applied to the thrust plate was varied from O to 250 1bf. There was no evidence during the test, which operated for 30 min, that a hydro- dynamic film was developed. Postrun examination of the surfaces revealed evidence of gross metal-to-metal rubbing. The surfaces of the rotor and stator thrust plates may have distorted sufficiently at elevated tempera- ture, or the stator plate may have vibrated enough to prevent the estab- lishment of a hydrodynamic film. Bench tests with oll as the lubricant are planned during the next quarter prior to further tests with molten salt in order to establish a base point of bearing operation. Test of Pump Equipped with One Molten-Salt-Lubricated Journal Bearing The detailed design work was completed on the pump, and the fabrica- tion of parts is essentially complete. Assembly and checking of the pump are under way. Bench tests are planned for the operation of the upper ball bearing and the lower journal bearing in oil prior to pump operation 19 with molten salt. A bearing having three axial grooves has been selected for the initial molten-salt operation of the pump. The pump test stand was completed. Bearing Mountings A tester for investigating at room temperature the stiffness produced by various methods of mounting bearing shells to pump containers was de- signed and fabricated. Preliminary testing has begun, using diaphragms of various thicknesses. Self-Welding of INOR-8 Parts have been fabricated to investigate the self-welding of INOR-8 against INOR-8 immersed in molten salt. The experiments will be conducted during the next quarter in the stress-rupture test facility at the Metal- lurgy Division. Mechanical Seals for Pumps The modified Fulton-Sylphon bellows-mounted seal,1 undergoing en- durance test in a PKP type of centrifugal pump, has accumulated an addi- tional 2352 hr of operation since the previous report period, for a total of 14,830 hr. The pump continues to operate at constant conditions of 1200°F loop temperature, a shaft speed of 2500 rpm, and a NaK flow rate of 1200 gpm. The seal leakage in both the test seal and the Durametallic upper seal has been negligible, No pump stoppages occurred since the last report. Pump Endurance Testing An MF type of centrifugal pump has continued in operationl and has logged more than 17,66h hr (over two years) ot continuous operation. There has been no maintenance performed on the pump during this period. During the last 15,817 hr the pump has operated in a region of cavitation at 2700 rpm, 645 gpm, and 2.5 psig pump-tank cover gas pressure in molten IMSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 27. 20 salt 30 at 1200°F. During the quarter the pump was stopped three times. One stop was momentary and due to a power outage, one stop was of 5 min duration to replace filters in the motor-generator set, and the other was of 10 min duration to replace the brushes in the d-c drive motor. Remote Maintenance Demonstration Facility This facility has been constructed in Building 9201-3 in the ¥Y-12 area to study remote maintenance problems and to demonstrate the feasi- bility of devices and techniques for remote maintenance of a molten-salt reactor system. It incorporates, to approximately 1/5 scale, a mockup of a typical reactor system in its essential parts, and a complement of re- mote manipulating and viewing equipment. Molten salt will be circulated through the system at prototype temperature and pressure. It 1s proposed to demonstrate that any component can be removed and replaced and the original integrity of the system re-established. Main components of the facility consist of the following: 1. Reactor. — A dummy reactor functions as a support for the pump and serves to represent the space layout problems associated with the actual reactor. The dummy is made of steel framing and is covered with insulation to form the desired reactor contour (Fig. 1.2.3). 2. Reactor Circuit. — The reactor circuit is intended to represent the space layout and maintenance problems associated with several symmetri- cally arranged heat exchanger circuits which might be incorporated in a typical molten-salt reactor. The circuit includes a standard PK pump to circulate molten salt (1200°F) through the dummy reactor (Fig. 1.2.4), two dummy heat exchangers (Fig. 1.2.5), and two dummy secondary systems (Fig. 1.2.6). The heat exchanger and reactor connections have freeze- flange mechanical joints. The fill and drain system has been designed to transfer approximately 120 gal of fuel 30 by means of pressurized gas from the sump tank to the reactor circuit. The reactor circuit will be filled with hot fuel, and the fuel will be circulated for a short time before being dumped. The mechanical joints will be broken, a component renoved and replaced, and the joint remade to meet the required leak- tightness specifications. 21 o (- 2 5 g 2 3 PHOTO 34545 actor shown at left. of dummy re Demonstration Facility. Contour Fig. 1.2.3. Remote Maintenance 22 UNCLASSIFIED PHOTO 34544 Fig. 1.2.4. Remote Maintenance Demonstration Facility. PK pump on tep of dummy reactor. Fig. 1.2,5. Remote Maintenance Demonstration Facility, Showing Two Dummy Heat Exchangers. UNCL SSIFIED PNOYO 34547 | 4 TR TR o e TT PR T T - Y UNCLASSIFIED PHOTO 34546 Fig. 1.2.6. Remote Maintenance Demonstration Facility, Showing Two Dummy Secondary Systems. 25 Piping is 6-in.-IPS Inconel for the main header circuits and 5-1/2- in.-IPS Inconel for the secondary circuits. The reactor circuit piping is supported on specially designed adjustable spring hangers (Fig. 1.2.7) that permit all piping to expand or contract during operation without overstressing the pipe. All piping can be removed from above. Reactor piping connecting heat exchangers, sump tank, and reactor is connected by freeze flanges (Fig. 1.2.7) that permit remote disassembly by the manip- ulator. All reactor piping is preheated by standard clamshell and Calrod electric heaters, with the exception of one section which is preheated by a specially designed clamshell unit (Fig. 1.2.8) containing electric heaters and thermal insulation in one package. The units are removable and replaceable in short sections by the manipulator. Thermocouple and electric heater disconnects (Figs. 1.2.9 and 1.2.10) have been specially designed to permit remote handling by the manipulator. 3, Manipulateor. — A General Mills model 300 manipulator has been specially modified for use in this facility. This mechanical arm has a vertical travel of 20 ft and operates on a railway 40 ft long with a 20-ft span. In addition to the manipulator, a remotely operated overhead crane ig available with 5-ton- and 20-ton-capacity hoists. The manipulator will be used for making and breaking connections, carrying light loads up to 750 1b, and aiding the overhead crane in positioning heavy compcnents. . Remote Viewing Equipment. — Closed-circuit stereotelevision is used for all remote viewing. Two sets of General Precision remotely op- erated stereotelevision systems are being installed. One set, with a three-lens turret, will be on a remotely operated dolly that is located on the edge of the control room rcof., The second stereotelevision set has an auto-zoom lens and will be located on the manipulator. 5. Control Room. — An existing control room is being used for the demonstration facility. This control room houses the control console for the manipulator, remote control for the overhead crane, closed-circuit stereotelevision viewer, and all controls necessary to operate the re- actor system. UNCLASSIFIED PHOTO 34360 Fig. 1.2.7. Freeze Flange and Spring Pipe Support in Position. UNCLASSIFIED PHOTO 34432 LZ Fig. 1.2.8. Combination Electrical Preheater and Insulation Unit for Use on Pipe. UNCLASSIFIED PHOTO 34359 | . e | /flfg‘ e s = ul H ! | | INCHES 1 121 3 4 5 | 6 ) > ! O B e £ Il | | | i PR P R ' s Labotal ol Fig. 1.2.9. Remotely Operated Thermocouple Connector. Fig. 1.2.10. Remotely Operated Electrical Connection. 30 The design, procurement, and construction of this facility were essentially complete by the scheduled completion date of July 1, 1959, - with the exception of the installation of the two closed-circult stereo= television sets. Frozen-Lead Pump Seal The small frozen-lead pump seal being tested on a 3/16-in.-dia shaft, as described previously,2 has operated continuously with no leak- age since the start of its operation on June 13, 1958. The accumulated operating time as of July 10, 1959, was 9LOO hr. The large lead pump seal with a B-l/h-in.-dia rotating shaft has been operating for 5400 hr, since November 12, 1958. Leakage of solid lead from the seal is sporadic. The lead leakage would probably be elim- inated by a seal which would provide better coolant control and would in- corporate a resilient packing to decrease the annulus between the seal, and shaft. A detailed description of the equipment was presented pre- viously.3 Information was obtained on a new material called "Zirklor," a com- pound of zirconium and chlorine, which has potential use for high-tem- perature bearings and seals because of its self-lubricating and elastic properties. Tests made on finely powdered samples showed Zirklor to be subject to severe oxidation in air at 500°C. Apparently it will be neces- sary to protect the Zirklor with a coating if it 1s used in air, or to provide an inert atmosphere, especially for elevated temperatures. Mechanical Joints for the Reactor System Three sets of disk springs have been ordered, one set for each of the three sizes of freeze-flange joints in the Remote Maintenance Facility. These springs should prevent or reduce gas leakage, which may occur when - the flange Jjoints are thermal-cycled. ZMSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 23. MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-268L, p L3, 31 Design, Construction, and Operation of Materials-Testing Loops Forced-Circulation Corrosion Loops The operation of long-term forced-circulation corrosion-testing loops was continued. Table 1,2,1 gives a summary of opérations for the periocd, with the total hours of operation at test conditions. Fourteen loops are in operation at the time of this report, consisting of eleven made of INOR-8 and three of Inconel. One test loop, designated 955&—5,4 was terminated during this period after a year of operation at test conditions. The loop, shown in Fig. 2.1.5, contained graphite samples in the hot fluid stream, The loop, container, and graphite have been dissected. The results of the examina- tion will be reported, when completed, by the metallurgy and chemistry sections. One test loop, designated MSRP-13 in the operations summary, began operation during this period. This loop is fabricated of INOR-8 and is operating with salt 136 (LiF-BeFs-UF,, 70-10-20 mole %). Normal operation was experienced by all but two tests during this period. Test loop 9377-5 was temporarily drained of molten salt April L, 1959, after it experienced a complete failure of the pump motor. It was refilled and resumed operation following replacement of the motor. Test loop 9377-6 experienced a failure of an electronic power-con- trol unit and became plugged before auxiliary heat sources could respond, a result of the high freezing temperature of the salt circulated. During normal operation of this loop there is only a 30°F difference between the freezing temperature of 990°F and the lowest wall temperature -~ecorded on the loop. In-Pile Loops The second MSRP in-pile loop test operating in the MIR was terminated when the pump seilized after approximately 1000 hr of operation. Pump opera- tion had been smooth until intermittent speed perturbations appeared a few days before termination. The exact cause of the pump seizure cannot “MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 36, 32 Table 1.2.1. June 30 Summary of Molten-50lt Forced-Circulation Corrosion Loop Operations Composition Hours of Loop Loop Number of Operation Designation Material Circulated at Test Comments** Salt* Conditions 9354-3 INOR-8 84 12,745 Normal operation 9354-1 INOR-8 126 10,836 Normal operation 9354-5 INOR -8 130 8,938 Loop contains graphite samples in hot fluid stream; terminated May 20, 1959, after one year’s operation 9354-4 INOR-8 130 8,034 Normal operation MSRP-7 INOR-8 133 7,755 Normal operation 9377-4 Inconel 130 7,749 Normal operation MSRP-6 {NOR-8 134 7,364 Normal operation MSRP-8 INOR-8 124 7,139 Normal operation MSRP-9 INOR-8 134 6,973 Normal operation MSRP-10 INOR-8 135 6,802 Normal operation MSRP-11 INOR-8 123 6,399 Normal operation MSRP-12 INOR-8 134 5,920 Normal operation 9377-5 Inconel 134 5,097 Loop was dumped temporarily April 4, 1959, when pump drive motor failed 8377-6 inconel 133 2,926 Electronic failure of power control June 1, 1959, caused plugging with high melting point salt (990°F); the salt was frozen in the loop but was remelted and operation was resumed. MSRP-13 INOR-8 136 2,119 Began operation April 8, 1959 *Compositions: 84 NcF-LiF-BeF2 {27-35-38 mole %) 131 LiF-Ber-UF4 (60-36-4 mole %) 123 Nch-Ber--UF4 (53-46-1 mole %) 133 LiF-Ber-ThF4 (71-16-13 mole %) 124 NuF-Ber-ThF4 (58-35-7 mole %} 134 LiF-Bv;er,-ThF4 (62-36.5-1-0.5 mole %) 126 LiF-Ber-UF4 {53-46-1 mole %) 135 NoF-BeF2-ThF4-UF4 (53-45.5-1-0.5 mole %) 130 LiF-Ber-UF4 {62-37-1 mole %) 136 LiF-Ber-UF4 (70-10-20 mole %) “*MSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 33. 33 be established until the loop is examined, but accumulated products of radiation-damaged lubricating oil probably interfered with the rotating members, as in previous in-pile pumps. Loop operation was satisfactory in other respects, The average loop heat generation for 766 hr with the reactor at pover was 7.71 kw (57 w/cm®), based on total fuel inventory of the loop. The average loop maximum temperature during power operation was 1%25°F. Measured loop temperature difference averaged 125°F, giving a total effec~ tive loop temperature difference of 217°F when the calculated heat ex- changer salt film temperature drop is included. The loop is scheduled to be cut from the shield plug at the MIR dur- ing July for return to ORNL for detailed examination. The chemical and metallographic examinations of the fuel and con- tainer materials of the first loop are in progress. 34 1.3 ENGINEERING RESEARCH Physical-Property Measurements Viscosity The viscosity of a high-thorium-content blanket salt mixture, termed BeLT-15 and having the composition LiF-BeFo-Thly (67-18-15 mole %), has 1 Qver been determined by using the "skirted" capillary efflux viscometer. the temperature range from 510 to 800°C, the results are described by the equation L = 0.0311 ¢2308/T where g 1s the viscosity in centipoises and T is the temperature (°K). The data are compared in Fig. 1l.3.1 with previous measurements® on salt mixtures 133 (LiF-BeFo-ThF,, 71-16-13 mole %), 134 (LiF-BeFo-ThF,-UF., 62-36,5-1-0.5 mole %), and 136 (LiF-BeFs-UF., 70-~10-20 mole %). The viscosity of the BeLT-15 mixture decreases somewhat more sharply with temperature than do the viscosities of the comparison salts. To estab- lish the effect of thorium content on the viscosities of BelT compositions, it is planned to study a number of additional ThF,-containing mixtures. The modified (skirted) efflux viscometers were used to redetermine the viscosities of mixtures 123 {NaF-BeFo-UF,, 53-46-1 mole %) and 126 (LiF-BeFo-UF,, 53-L46-1 mole %).2® The current data are compared with the earlier measurements in Fig. 1l.3.23 also shown are the results with com- position 1350 (LiF~BeF2-UF4, 62-37-1 mole %). It is to be noted that the new data indicate lower viscosities for both mixtures. With salt 126 the difference between the two sets of data increases with Increasing temperature, while with composition 123 this difference decreases with temperature. Between the temperatures of 500 and 900°C the viscosities IMSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684, p 65. “MSR Quar. Prog. Rep. April 30, 1959, ORNL-272%, p 39. SMSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 38. TEMPERATURE ({°C) 35 UNCLASSIFIED ORNL-LR-DWG 40622 450 500 600 700 800 100 r | T T | | [ [ | [ 8C 60 LIQUIDUS TEMPERATURES (°C) 40 | | ! | bt s i N \\ - 133 ~. \\ BelT — 15 20 \ "\.‘\\ . SALT NO. \\"‘ 3 134 \_ 3 136 = @ S > 10 = S g 8 2 -3 {6 S e SALT NO. COMPOSITION (mole 7o) =3 4 Q LiF BeF, The, UR, 134 62 36.5 { 05 136 70 10 - 20 BeLT-15 | &7 18 15 - ! 13 - N 133 7 6 4 } 0.0014 00013 0.0042 0.001 4 0.0009 Figo 1.3.1- 17 ek ] Viscosity-Temperature Relations for BeL. T-15 and Related BeF, Salt Mixtures. L, VISCOSITY (centipoises) 36 TEMPERATURE (°C) 600 UNCLASSIFIED ORNL-LR-DWG 40623 100 [ 80 | | —— SALT NO. COMPOSITION {mole %) | TYPE OF VISCOMETER ] 0.0009 LiIF NaF BeF, UF, |SKIRTED PLAIN | | >l 426 53 - 46 | s A S T S N 123 - 53 46 | o . T 130 62 - 37 | |=——-— NOTTESTED ; | | | | L ¥ | | | ] | 0.0014 0.0013 0.0012 0.001 0.0010 1/7 [k ™'] Fig. 1.3.2. Viscosity-Temperature Relation as Determined in Skirted and in Plain Viscometer Cups. 37 of salts 123 and 126 are represented to within #5% by the equations: salt 123 m 0.0234 e5505/III 2 salt 126: u = 0,010k e6u87/T b4 where the viscosity is given in centipoises when the temperature is ex- pressed in absolute units (°K). Enthalpy and Heat Capacity It has been the practice to report the results of experimentally de- termined enthalpy-temperature relations for fused salt mixtures in the form of a three-constant least-squares-evaluated equation, HT - H50 = a + bt + ct The heat capacity equation derived therefrom, %% = cp =Db + 2ct , retains a temperature dependence. Data obtained both in this Laboratory and elsevhere have indlicated that for some salt mixtures the heat capacity increases with temperature and for others, decreases with increasing tem- perature. Douglas and co-workers? have observed that for liquid metals and fused salts the heat capacity decreases slowly with temperature in the region near the solidification temperature, may pass through a minimum, and then may increase at the higher temperatures. On this basis, a general prediction of the heat capacities of mixtures of various compositions is not readily obtainable. However, the value of such a correlation in engineering analysis 1s obvious. 8ince the curvature of the enthalpy- temperature equation is, in most cases, small, it appears feasible to attempt a straight-line (two-constant) correlation,® yielding the heat 4T, B. Douglas et al., Trans. Am. Soc, Mech., Engrs. 79, 23 (1957). >The subscript 1 (e.b., Cp,1 ) is used to designate the heat capacity obtained from the straight-line correlatlon of the enthalpy data. 38 capacity as a constant over the temperature range investigated. Tae heat capacity of the liquid can then be related®»® to such easily calculable - properties of the mixture as the average molecular weight and the average nunber of atoms per molecule; thus = \B c = A M s Dy fi where'fi, the average molecular weight, 1s defined as =), MT and fi, the average number of atoms per molecule in the mixture, is =), NT, - In these expressions, Mi is the molecular weight of component 1, Ni is the number of atoms of component i, and fi is the mole fraction of com- ponent 1. The constants A and B are empirically determined. In accord with this discussion, the results of measurements with six BeFs-containing mixtures are given in Table 1.3.1 and are plotted in Fig. 1.3.3 as a function of the ratio fi/fi'(the average atomic weight). From this, A and B have been evaluated; the resulting correlation is — \0.785 . =5 <_M_> Pl 'fi- This equation represents the available data to within #5% and may be used within these limits to predict the heat capacities of mixtures containing various amounts of NaF, LiF, BeFo,, UF,, and ThF,. Preliminary data on the enthalpy of the Bel.T-15 mixture are presented in Fig. 1.3.4. The ratic M/N for the mixture is calculated to be 27.h; then, from Fig. 1.3.3, the heat capacity 1s predicted to be 0.285 cal- g"l-(°C)'l. The correlation of Fig. 1.3.4 yields a value approximately - 5 to 10% above thic estimate. Study of the Bel.T-15 salt is continuing. 4. D. Powers and G. C. Blalock, Enthalpies and Heat Capacities of Solid and Molten Fluoride Mixtures, ORNL-1956 (Jan. i1, 1956). 39 Table 1.%5.1. Heat Capacities and Average Atomic Weights of Mixtures Containing BeFo Composition (mole %) Average Heat Capacity Salt . o Atomic calrgLl+(°C)~1 ** 1iF NaF BeF, ThF, UF, Veight & 123 53 L6 1.0 15.99 0.442 126 53 L6 1.0 135.09 0.517 130 62 37 1.0 12.84 0.520 134 62 36.5 1.0 0.5 13.24 0.493 133 71 16 13 25.87 0.312 136 70 10 20 31.72 0.245 Thermal Expansion A furnace and an inert-atmosphere enclosure have been designed for the thermal expansion apparatus previously described.” An essential feature of the furnace design is the inclusion of five independent heat- ing circuits so as to effect fine control of the temperatures in the two legs of the U-tube, Heat Transfer Studies Assembly of the apparatus for determining surface film formation in molten-salt systems by heat transfer coefficient measurements is pro- ceeding., All electrical and piping facilities have been installed; sub- assembly welding of the loop has been completed. Modular instrument and control units are being installed in the new laboratory area. As previously described,8 the loop will contain two test sections (see Fig. 1.3.5). It is planned to electrically parallel these two test units, with the pump sides of the sections being maintained at ground "MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 39. 8MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p L6, 40 UNCLASSIFIED ORNL-LR-DWG 40624 1.0 0.8 0.6 SALT NO. — 130 t ®) 126 &) 434 e - @ 123 IU‘ \ o 0.4 - T (=] \ il 5 = 1 %[J @133 o < \ Q = L 136 @ T & 02 N 3,821 /V\ DT (47)0.785 N cA 10 20 30 40 50 60 AVG MOLECULAR WEIGHT AVG NUMBER OF ATOMS M N Fig. 1.3.3. Relation Between Heat Capacity and Average Atomic Weight for Mixtures Containing BeF ,. 4] UNCLASSIFIED ORNL-LR-DWG 40625 300 250 ’,’, — //,. G ///‘ 5 L 200 8 X ¢/ | E ./ ‘ z e % 150 6 tl") [ tl\ 3 100 ¢ 50 300 400 500 600 700 800 900 TEMPERATURE (°C) Fig. 1.3.4. Enthalpy-Temperature Relation for Mixture Bel T-15, potential, The heat-gcink sides of the units will then be electrically isolated from ground. If the resistances of the two test sections are not identical (as is the case for an Inconel and INOR-8 pairing), a by- pass current will exist in the heat-sink leg of the loop. While it is anticipated that this leakage will be small, its magnitude must be known in order to effect a heat balance for the test units. Since a current transformer capable of enduring the temperatures in the vicinity of the loop is not available, the voltage drop across this leg and its resistance will be used to calculate the current. The effect of such a heat-sink leg current on the flowmeter 1s unknown; if it should be appreciable, it may become necessary to modify the heating circuits. A possible scheme would provide independent controls for the two units. Again, the pump- side electrodes would be maintained at ground potential; the I®R heating of the sections would then be adjusted to provide zero voltage difference between the sink-side electrodes. The magnitude and effect of heat generation in the Inconel electrodes are being evaluated. 42 UNCLASSIFIED ORNL-LR-DWG 35679 FLOWMETER R £ \ HEAT SINK ELECTRODES > sk ) RISR \% &4 TEST “ SECTIONS MIXING CHAMBER (LOCATED IN PIPE PUMP AT EACH END OF TEST SECTION) Fig. 1.3.5. System for Heat Transfer Coefficient Measurements on Ber-Conmining Molten Salts. 43 1.4 INSTRUMENTS AND CONTROLS Molten-Salt-Fuel Level Indicators Metallurgical examination of the Q-1717-54 level probe tested pre- viously® has been completed. Corrosion was negligible; the probe was in excellent condition, both mechanically and electrically. A second series of tests is in progress in which two Q-1717-54 level elements made of Inconel are being tested in fuel 130 (LiF-BeFs-UFy, 62-37-1 mole %) at 1200°F. Initially, as in previous tests, both ele- rnents provided an output of approximately 8.5 mv over a level range of 0—6 in. of fuel. TFor approximately three days this output was a function of level, with a sensitivity and reproducibility of +3 5%, On the fourth day the span of both probes gradually decreased. This decrease has con- tinued at a reasonably constant rate until, at present, the span of one of the probes has decreased to 2.0 mv after two months of continuocus operation. When the span of the second probe reached 3.6 mv after 37 days of continuous operation, the fuel was dumped and the vessel was refilled with fresh fuel in order to obtain evidence that the probe span decrease was due to an increase in the resistivity of the fuel with time. The fresh batch of fuel provided a probe span of 9.0 mv approximately one day after the filling operation was completed. Further evidence to support the theory that a fuel resistivity change had occurred was obtained by agitating the fuel in one vessel and measuring the probe output before and after agitation. The span, alfter agitation, had increased approximately 3.0%. IMSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p U47. PART 2 . MATERTALS STUDIES 47 2.1 METALLURGY Dynamic Corrosion Studies INOR-8 Thermal-Convection Loops Eleven INOR-8 thermal-convection loops were examined during the quarter; the results are summarized in Table 2,1.1, Nine of the loops had operated for one year; each of the other two had operated for 1000 hr. All the cne-year tests except for lcop 1208 revealed only a very light attack, in the form of surface pitting, along hot-leg sectilons. In loop 1208, which circulated salt 12 (NaF-KF-LiF, 11,5-L2-46.5 mole %) at a hot-leg temperature of 1250°F, subsurface voids were alsc apparent in hot-leg sections to depths of up to 1 mil. A typical hot-leg section from this loop is shown in Fig. 2.1.1. Chemical analyses of salt samples removed from hot- and cold-leg sections of loop 1208 showed a significant increase in nickel over the amount reported to be in the salt before test. Increases in the amounts of iron and chromium were somewhat less, although analyses of salt samples from the trap area, which is colder than the circulating salt in the loop, indicated the presence of 2KF+:NaF«.CrlF'n. All salt samples also contained traces of KF.2H-0, which similarly had been ncted in salt samples removed from Inconel loop 1214, reported previously.t A salt mixture from the same batch as that used in loop 1208 was circulated in loop 1214 for L4670 hr at 1250°F and resulted in extensive attack along hot-leg sections of the Inconel to a depth of 13 mils. The abnormally high attack in loops 1208 and 1214 was very likely associated with the presence of KF:2H-0, since its presence shows that water, a highly oxidizing impurity, entered the salt during its preparation or testing. The most extensive attack found among the other loops that had op- erated for a year occurred in loop 1212, which circulated salt 125 (NaF-BeFo-ThF4-UF,, 53-46-0.5-0.5 mole %) at a maximum temperature of 1250°F. Hot-leg sections from this loop contained surface pits to 1MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684, p 76. Table 2,1.1. Results of Metallographic Examinations of INOR-8 Thermal-Convection Loops Metallographic Results Loop Tegt Maximum Fluid- Salt No. P?rlOd fletal Interface No.* Hot-Leg Cold-Leg hr) Temperature (°F) Appearance Appearance 1219 8760 1350 122 Light surface roughening No attack 1200 8760 1250 130 Light to moderate surface Light surface roughening with a few pits, roughening all less than 1 mil deep 1196 8760 1350 130 Light surface roughening No attack 1206 8760 1350 126 Light surface pitting No attack 1212 8760 1250 125 Moderate pits <1 mil in depth No attack 1215 8760 1350 125 Moderate surface roughening Light surface roughening 1209 8760 1350 128 Heavy surface roughening No attack 1190 8760 1250 127 Light to moderate surface No attack roughening with a few pits, ‘ all less than 1 mil deep 1208 8760 1250 12 Moderate surface roughening No attack and pits up to 1 mil deep; a few subsurface voids 1 mil deep 1241 1000 1250 136 No attack No attack 1242 1000 1350 136 Moderate surface roughening No attack *100: NaF-ZrF,-UF, (57-42-1 mole %) 128: LiF-ThF, (71-29 mole %) 1%30: LiF-BeFo-UF, (62-37-1 mole % 127: LiF-BeFo-ThF, (58-35-7 mole %) 126: LiF-BeFo-UF, (53-46-1 mole % 12: NaF-LiF-KF (11.5-46.5-42 mole %) 125 NaF-BeFo-UF4~ThF, (53-46-0.5-0.5 mole %) 136: LiF-BeFs-UF, (70-10-20 mole %) 49 UNCLASSIFIED T-17347 Lo 3 Fig. 2.1.1. Specimen Taken from Hot Leg of INOR-8 Thermal-Convection Loop 1208 at Point of Maxi- mum Loop Temperature (1250°F). Loop was operated for one year with salt mixture NaF-KF-LiF (11.5- 42-46.5 mole %). Etchant: 3 parts HCI, 2 parts H,0, 1 part 10% chromic acid. 50 approximately 5/h mil in depth, as shown in Fig. 2.1.2. Loops 1200, 1190, and 1208, which similarly operated at 1250°F, also showed surface pits but to smaller depths. However, only shallow surface roughening was evi- dent in loops operated at a maximum temperature of 1350°F (1219, 1196, 1206, 1215, 1209), Figure 2,1.3 illustrates the appearance of loop 1215, which contained salt 125 at 1550°F, and is typical of the loops that op- erated in this temperature range. It will be noted in both Figs. 2.1.2 and 2.1.3% that a very thin con- tinuous film has formed along the surfaces exposed to fuel mixtures. oimilar films were observed in the examinations of loops 1196 and 12063 the composition of these films is not yet known. Cold-leg surfaces of all the above-mentioned INOR-8 loops showed no evidence of fluoride attack, although a surface film (similar to the films described above) was evident on some of the samples, Analyses of salt mixtures circulated in these loops (except loop 1208) showed no significant increase in corrosion products during test. In the 1000-hr INOR-8 tests, salt 136 (LiF-BeFp-UF,, 70-10-20 mole %) was circulated at maximum hot-leg temperatures of 1250°F (loop 1241) and 1350°F (loop 1242). TLoop 1241 showed no evidence of attack in hot- or cold-leg regions., Loop 1242 showed evidence of moderate surface roughen- ing in sections of the hot leg but no attack in sections of the cold leg. Inconel Thermal-Convection Loops Examinations were completed of nine Inconel thermal-convection loops that operated with fluoride mixtures for one year and of three that op- erated for 1000 hr., Test conditions and metallographic results are re- ported in Table 2.1.2. Except for loop 1202, all tests which operated for a year at a maximum temperature of 1250°F revealed intergranular void formation to a depth of 5 to 7 mils in sections of the hot legs. The surfaces of the cold-leg regions revealed only heavy surface roughening. No evidence of mass-transfer deposits was found. Loop 1202, which cir- culated salt 125 (NaF-BeFp-UF,-ThF,, 53-46-0.5-0.5 mole %) at a maximum hot-leg temperature of 1250°F, was terminated just short of its scheduled operating time of one year because of a leak which developed at the bottom 51 UNCLASSIFIED T-17133 Fig. 2.1.2. Specimen Taken from Hot Leg of INOR-8 Thermal-Convection Loop 1212 at Point of Maxi- mum Loop Temperature (1250°F). Loop was operated for one year with salt mixture NaF-BoFZ-UF‘-ThF‘ (53-46-0.5-0.5 mole %). Etchant: 3 parts HCI, 2 parts H2O, 1 part 10% chromic acid. UNCLASSIFIED T-17242 Fig. 2.1.3. Photomicrograph Showing Appearance of Hot Leg of INOR-8 Thermal-Convection Loop 1215. Loop was operated for one year at 1350°F with fuel mixture NaF-BeF ,-ThF ;UF , (53-46-0.5-0.5 mole %). Etchant: 3 parts HCI, 2 parts H20, 1 part 10% chromic acid. 52 Table 2.1.2, Results of Metallographic Examinations of Inconel Thermal-Convection Loops Maximum - Test Fluid-Metal . Metallographic Results Loop Period Interface Salt No. (hr) Temperature No.* Hot-Leg Appearance Celd-Leg Appearance °F) 1217 8760 1350 122 Light surface pitting 1 mil; moderate inter- Heavy surface roughening granular voids up to 7 mils deep 1218 8760 1350 123 Moderate surface roughening and pitting to @ Maderate surface raughening depth of 1 mil; heavy intergronular voids to and pitting less than 1 mil a depth of 18 mils deep 1180 8760 1350 130 Light surface roughening ond pitting to a Light surface roughening with depth of 1 mil; heavy intergronular voids grain-boundary penetrotions to a depth of 14 mils to a depth of 1 mil 1198 8760 1250 126 Light surface roughening and pitting to a No ottack depth of 1 mil; moderote to heavy inter- granular voids to a depth of 7 mils 1199 8760 1350 126 Light surface roughening and pits to a Very light surface roughening depth of 1 mil; heavy intergranular voids to a depth of 15 mils 1201 B6B4 1250 126 Light surface roughening and pits to a depth of 2 mils; heavy intergranulor voids to a depth of 7 mils - 1202 8308 1250 125 Moderate surface roughening; heavy inter- Na attack granular voids to a depth of .13 mils 1211 8760 1350 125 Moderote surface roughening with pitting No attack to a depth of 1 mil; intergranular voids to a depth of 15 mils 1210 8760 1250 128 Moderate surface roughening and pits to No attack a depth of 1 mil; light to moderate inter- granular subsurface voids to a depth of 5 mils 1234 1000 1250 133 Light surface roughening and pits to a Light surface roughening and depth of 1 mil; few intergranular voids penetrations less than 1 mil to a depth of 1 mil deep 1232 1000 1250 134 Light surface roughening and pits to a Light surfece roughening with depth of 1 mil; moderate intergranular grain-baundary penetrations voids to a depth of 3 mils to a depth of 1 mil 1243 1000 1250 135 No attack No attack *Compositions: 122 123 130 126 125 NoF-Zer—UF4 (57-42-1 mole %) chF-Ber-Uf:4 {53-46+1 mole %) LiF-Ber-UF4 (62-37-1 mole %) LiF-Ber-UF4 (53-46-1 mole %) Nc':F-BeI:2-U|:4-T|'1F4 (53-46-0.5-0.5 mole %) 128 LiF-ThFl4 (71-29 mole %) Nc:F-Ber-Uf=4-T'r1F4 (53-45.5-0.5-1 male %) 133 LiF—Ber-ThF4 (71-16-13 mole %) 134 LiF-Ber-UF4-ThF4 (62-36.5-0.5-1 mole %) 135 53 of the cold leg. Examination of the hot-leg regions revealed heavy inter- granular subsurface void formation to a depth of 13 mils, as shown in Fig. 2.1.4. The cold-leg surfaces showed no evidence of attack or mass transfer. As shown in Table 2,1.3% chemical analysis of the after-test salt showed considerably more chromium than did the before-test salt. Table 2.1.3. Analyses of Salt Mixture 125 Before and After Circulation in Loop 1202 for 8308 hr Major Constituents Minor Constituents Sample Taken (vt %) (ppm) U Th Be Ni Cr Fe Before test 1.56 3.07 8.22 L5 105 220 After test 2.16 3.07 10.70 25 2905 290 Examination of the loops which operated for one year at a maximum hot-leg temperature of 1350°F revealed moderate to heavy subsurface wvoid attack ranging in depth from 7 to 18 mils. All three 1000-hr tests were carried out at a maximum hot-leg tem- perature of 1250°F. Under these conditions, circulation of salt 133 (LiF-BeFs-ThF,, 71-16-13 mole %) in loop 1234 resulted in only slight attack (<1 mil), occurring as subsurface void formation. Attack was somewhat more severe in loop 12%2, which circulated salt 134 (LiF-Bng— UF 4-ThF,, 62-36.5-0.5-1 mole %); void formation occurred to a depth of 3 mils. Loop 1243, which circulated salt 135 (NaF-BeFs-UF,-ThF,, 53-45,5-0,5-1 mole %), showed no evidence of attack in either hot- or cold-leg regions. As reported previously,® examination of Inconel loop 1182, which circulated salt 126 (LiF-BeF2-UF,, 53-46-1 mole %) for one year at a ®MSR Quar. Prog. Rep. Apr. 30, 1959, ORNL-2723, p 5kh. 54 Fig. 2.1.4, Specimen Taken from Hot Leg of Inconel Thermal-Convection Loop 1202 at Point of Maxi- mum Loop Temperature (1250°F). Loop was operated for 8308 hr with salt mixture NeF-BoFZ-UF‘-ThF‘ (53-46-0.5-0.5 mole %). Etchant: modified aqua regia. 55 maximum hot-leg temperature of 1350°F, revealed the presence of a large quantity of metallic particles in the trap area of the loop. Chemical analysis of these particles indicated them to be composed primarily of chromium. These deposits apparently formed as a result of mass transfer occurring across the very steep temperature drop which exists in the trap region of the loop. INOR-8 Forced-Circulation Loops An INOR-8 pump loop, 935L-5, which was designed to determine the compatibility of INOR-8 and graphite in a fluoride fuel medium, was terminated after successfully completing one year of operation. Graphite samples for this experiment consisted of thirty-two l/2—in.-dia rods and eighteen 5/16—in.-dia rods, each 11 in. long, stacked in a horizontal array. The graphite was a low-permeability type, National Carbon grade GI'-123, The rods were contained in a rectangular box constructed of INOR-8 and located at the outlet of the second heater of a standard INOR-8 forced-circulation loop, as shown in Fig. 2.,1l.5., Space was main- tained between each of the rods and between the rods and the sides of the box by means of 0.035-in.-dia wire spacers wound around the l/2—in.-dia graphite rods. Operating conditions for the experiment were: Salt composition (salt 130) LiF-BeF>-UF, (62-37-1 mole %) Maximum fuel-metal interface 1300°F temperature Temperature of salt in contact 1250°F with graphite Minimum fluid temperature 1100°F The test was designed to evaluate the graphite for such effects as weight changes, fluoride penetration, and mechanical distortion. Also of interest were the effects of graphite on the INOR-8 container material, particularly with respect to carburization and influences on the corrosion process. UNCLASSIFIED ORNL-—-LR- DWG 39365 Ti THROUGH T14 PT ¢ THROUGH PT 14 SELECTOR SWITCH NO.§ @ T154 THROUGH Ti5M —® ELEC PT.15 . 16 POINT TEMPERATURE SELECTOR SWITCH NO.2 RECORDER TI6A THROUGH TI6D —®7 PT 16 GRAPHITE CONTAINER DETAIL "aA" SAMPLE LEG FREEZE VALVE DRAIN TANK Fig. 2.1.5. Isometric Sketch of INOR-8 Forced-Circulation Loop 9354-5. 9S 57 Metallographic examination of a section removed from the loop wall at a point of maximum temperature revealed light surface roughening and pitting to less than l/h mil., A layer approximately 1/5 mil thick was also evident in this sample, as shown in Fig. 2.1.6. A section removed from the exit end of the graphite container also showed light surface roughening and a similar but much thinner surface layer, as shown in Fig. 2.1.7. There was no evidence of carburization found in any of the sections examined., Identification of surface layers on the sections is being attempted using x-ray and electron diffraction technigques. Chemical analyses of the salt before and after the test are shown in Table 2.1.4, Except for an expected increase in chromium concentra- tion, the analyses showed little difference. Examinations of the after- test salt were also made under the petrographic microscope and x-ray diffraction unit to detect the presence of oxides. These analyses showed . the salt to be apparently unaffected by impurities in the graphite, although by analytical methods the presence of 3400 ppm of oxides was found. Table 2.1.4k. Analysis of Salt 130 Before and After Circulation in Pump Loop 9354-5 Major Constituents Minor Constituents Sample Taken (vt %) (PPm) U Be Ni Cr Fe Before test 8.37 L.87 5 135 235 After test Loop 9.77 k.97 25 550 330 Sump 3.55 4.59 85 585 L60 58 UNCL ASSIFIED T-17467 Fig. 2.1.6. Specimen Taken from Hot Leg of INOR-8 Forced-Circulation Loop 9354-5 at Point of Maxi- mum Loop Temperature (1300°F). Loop was operated for 8938 hr with salt mixture LiF-BeF ,-UF , (62-37-1 mole %). Etchant: 3 parts HCI, 2 parts H20, 1 part 10% chromic acid. 250X. UNCLASSIFIED| T-17468 Fig. 2.1.7. Specimen Taken from Top of Box Containing Graphite Rods Located at End of Second Heater Leg in INOR-8 Forced-Circulation Loop 9354-5. Loop was operated for 8938 hr with salt mixture LiF-Bch-UF‘ (62-37-1 mole %). Etchant: 3 parts HCI, 2 parts HZO' 1 part 10% chromic acid. 250X. 59 The first of three INOR-8 corrosion inserts® has been removed from INOR-8 forced-circulation loop 9354-4, circulating salt 130 (LiF-BeFo-UF,, 62-37-1 mole %) at a maximum fuel-metal interface temperature of 1300°F. As previously discussed, these inserts are intended tc yield a measure of corrosion in terms of weight changes which take place at the point of maximum loop wall temperature. The insert removed had been located at the end of the second heater leg for 5000 hr. There was a small weight loss along the 4-in. length of the specimen. This loss, which is an average value for the L-in., length, amounted to approximately 1.7 * 0.03L mg/cm2 and corresponds to a loss of 0.08 + 0.0016 mil of the tube wall if uniform removal of the material is assumed. The errors listed with these values are based on estimated limits of precision of the instruments used to measure and weigh the insert. They do not in- clude errors resulting from handling of the insert, which could be signifi- cant, considering the small weight change, but are difficult to estimate. For this reason, care should be exercised in extrapolating these weight losses beyond the conditions under which they were obtained. Additional weight measurements to be determined for two other inserts still in test will afford a sounder basis Tor such extrapolations. The present status of all forced-circulation loop tests now in pro- gress 1s given in Sec 1.2 of this report. General Corrosion Studies Carburization Tests on Inconel and INOR-8 in Systems Containing Fuel 130 and Graphite A test in which Inconel and INOR-8 specimens were exposed to fuel 150 (LiF—Bng—UF4, 62-37-1 mole %) at 1300°F in a system containing graphite was terminated after 6000 hr operation. Similar tests for 2000 and L4000 hr have been reported.? No evidence of carburization in the INOR-8 tested was shown by metallography or by tensile tests of both test and control specimens 3MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-247k, p 31. “MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684, p 78. 60 (control specimens were subjected to argon under the same conditions of time and temperature). Mechanical-test results on INOR-8 and Inconel . specimens from this system are shown in Table 2.1.5. Chemical and metal- lographic examinations of Inconel specimens are being made. Table 2.1.5. Results of Tensile Tests on Inconel and INOR-8 Exposed to Argon and Fuel 130 plus Graphite for 6000 hr at 1300°F Control Specimens Test Specimens Material Tensile Elongation Tensile Elongation Strength (4 in 2 in.) Strength (% in 2 in.) (psi) / ' (psi) : ' Room-Temperature Tests Inconel 83,400 36.5 76,000 3%.8 . INOR-8 124,700 39.0 124,200 36.0 1250°F Tests Inconel L6, k00 29.5 39,400 18.0 INOR-8 72,800 16.5 T4, 500 16.0 A second carburization test, similar to the one described above but containing only INOR-8 specimens, was terminated after 5000 hr operation at 1300°F. Specimens in separate compartments were exposed to (1) fuel 130 plus graphite, (2) fuel 130 only, and (3) argon. No evidence of car- burization was observed metallographically in the specimens exposed to fuel 130 with graphite, and results of mechanical tests and carbon analysis of these specimens were not significantly different from results for pieces exposed to fuel 130 only or to argon, again indicating no carburi- zation, Sodium-Graphite Carburization Tests The last of a series of seven carburization tests with Inconel and INOR-8 exposed to a sodium-graphite system was terminated after 4000 hr 61 operation at 1LOO°F. A description of the carburization system has been presented earlier.® Both materials were heavily carburized (30-40 mils in depth), which severely affected their room-temperature tensile properties. Elongation values at room temperature dropped from M5% in 2 in. for control specimens (exposed to argon) to 6% in 2 in., for the test specimens, The difference in elongation values at 1250°F showed less spread; elongation was 19 and 22% in 2 in. for the control and test specimens, respectively. Penetration of Graphite by Molten Fluorides Studies concerning the penetration of graphite by molten fluorides were continued by using procedures described previously,6 except that the 2%72°F (1300°C) graphite degassing operation was eliminated because such a treatment would be impractical in reactor systems. Graphite, AGOT grade, was penetrated by fuel 30 (NaF—ZrF4—UF4, 50-46-4 mole %) and by fuel 130 (LiF-BeF2-UF,, 62-37-1 mole %) in 100-hr exposure tests at 1300°F (706°C) with the systems pressurized to 150 psig. Table 2.1.6 gives the amount of salt entering the graphite in each test. The AGOT graphite, a relatively permeable grade, has been used in an effort to establish the characteristics of salt penetration into graphite. Low-permeability grades of the type expected to be used for reactor appli- cation are being acquired for testing. The difference in the gquantity of fuel 130 penetration in tests 1 and 2 (Table 2.1.6) was possibly due to pressure variations in test 2, resulting from an argon leak that occurred during the test and lasted for approximately 30 hr. In most of the future penetration studies, graphite of low perme- ability will be used, in order to determine: (1) the quantity and con- figuration of molten flucride penetrations into these graphites as func- tions of pressure, time, and temperature; (2) the mechanisms causing penetration of voilid spaces of graphites by the molten fluorides; and °MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 59. ®MSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 5k. 62 Table 2.1.6. Penetration of AGOT-Grade Graphite by Fuels 30 and 130 Test conditions: Graphite specimens: 1-7/8 in. OD x 1/4 in., ID x 2 in. long Temperature: 1300°F (704°C) Test period: 100 hr Pressure: 150 psig Specimen Calculated Accessible Test No. Fuel No. Weight Void Veolume Filled* Gain (%) (%) 1 130 19.1 75.3 2 130 14.3 58.4 > 50 29.0 63,4 *The total accessible void volume is reported to be 21.7% of the bulk volume of AGOT graphite; W. P. Eatherly, M. Janes, and R. L. Mansfield, "Physical Properties of Graphite Materials for Special Nuclear Applications,"” Second U.N. Intern. Conf. Peaceful Uses Atomic Energy, Geneva, 1958, paper A/Conf 15/P/708, p 4 (June 1958). (3) whether there is a reactor-grade graphite that is impermeable to the molten fluorides. The graphites that are penetrated with salts will be thermal-cycled to determine whether volume changes of the entrapped salt will cause physical damage to the graphite. Removal of Oxide Contaminants from Graphite Previous tests have shown that exposing graphite to fuel 130 will result in a precipitation of UOCs. This precipitation reaction, which is complete in 5 hr at 1300°F, has occurred in tests where the graphite was not in direct contact with the salt as well as in tests where the salt was contained in a graphite crucible, ™ These results suggested that 7MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 62. 8MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684%, p 80. ®MSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 58. 63 1t might be feasible to remove oxide contaminants from the graphite by gettering or flushing the graphite with a suitable salt, such as 130. A gettering test has been completed in which an empty crucible of AGOT graphite was heated for 20 hr at 1300°F in a capsule which con- tained fuel 130 in a compartment separated from the graphite. Uncon- taminated fuel 130 was then placed in the graphite crucible and the capsule was again heated to 1300°F for 20 hr. Radiographic examination after each heating operation showed that the UOs precipitated in the gettered fuel sample was reduced to approximately one-third of that normally found in ungettered samples. In a test to evaluate the ability to flush oxide contaminants from graphite, a crucible of AGOT graphitel® was exposed to fuel 130 for 20 hr at 1300°F (704°C) in a vacuum atmosphere. As expected, radiographic examinations indicated a uranium precipitation from the fuel. The ex- posed fuel and precipitate were then removed from the graphite crucible and replaced with a fresh charge of fuel 130, and the system was again heated for 20 hr at 1300°F., A uranium precipitate approximately one- third of that normally expected with this graphite was observed in radio- graphs of the test. Mechanical Properties of INOR-8 The investigation of the high-temperature mechanical properties of INOR-8 includes tests of properties such as short-time tensile strength, creep, relaxation, and fatigue. Most of these studies have been in the temperature range from 1000 to 1500°F in air and in molten-salt environ- ments. The effects of shape, compositional variations, aging, and car- burization have been surveyed. Much of this work has been completed and is reported in past quarterly reports,** 13 4 summary report containing 19The machining dust was removed in a sonic bath of ethyl alcohol, and the crucible was dried for 15 hr at 200°C. *MSR Quar, Prog. Rep. June 30, 1958, ORNL-2551, p 6k, L°MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 6h. T3MSR Quar. Prog. Rep. Dec. 30, 1958, ORNL-2684, p 82, 64 the detailed results of this program is now being compiled.1* 1In this quarterly report a general review of the entire program is presented. In attempting to evaluate the mechanical properties of INOR-8 in relation to those of other high-temperature alloys, it is interesting to consider that most of the heats tested were experimental. Thus the composition and melting practice did not conform to the specifications now established. However, in spite of these variations the properties vere fairly consistent, especially with regard to creep behavior. Thus significant variations in properties would not be expected in future heats of the alloy. Although the program represents a fairly comprehensive study of the mechanical properties, there are several important areas which have not received sufficient attention. These include strain fatigue in the fused- salt enviromment, multiaxial-stress-rupture tests on wrought and weld metal, and radiation-damage studies. Materials Five heats of INOR-8 have been studied: §P-16, SP-19, 8 M-1, 1327, and 1566. These heats have carbon contents of 0.02, 0.06, 0.1k, 0.1k, and 0.08%, respectively. Heat SP-16 is unusually high in boron (0.02-0.0%%) and exhibits brittleness when welded. Heats 38 M-1 and 1327 have higher carbon contents than that set forth in the maximum specification. Heat 1%27 is an arc-cast vacuum-remelted heat with a composition nearly identi- cal with that of 8 M-1. Tensile Properties Tensile tests on sheet material indicate that INOR-8 has good strength between 1000 and 1500°F, The 0.2%-offset yield strength exhibits very little temperature dependence and falls in a band between 25,000 and 40,000 psi in this temperature range. Heats 8 M~1 and 1327 appear to be the strongest materials, and data for these heats lie near the top of the band. Heats SP-16 and SP-19 fall near the lower limit. 14R. W. Swindeman and D. A. Douglas, The Mechanical Properties of INOR-8, ORNL-2780 (to be published). 65 Tensile tests on rod specimens agree, in general, with sheet data except for heat SP-19, which shows considerable scatter. Yield strengths as low as 21,000 psi have been observed between 1000 and 1500°F for SP-19 rods., The tensile ductility also varies from heat to heat. Maximum and minimum tensile elongations are constant up to 1000°F, but rapidly de- crease between 1000 and 1500°F. Sheet specimens of heats 8 M-1 and 1327 exnibit minimum elongation. Minimum values are near 55% up to 1000°F and drop to lO% at 1500°F. Heat SP-16 is the most ductile, with elonga- tions as high as 67% for rod specimens and 55% for sheet specimens. Heat SP-19 falls between 8 M-1 and SP-16, and its ductility at 1500°F appears to be improved by aging at 1650°F., Aging studies'® on SP-19 and 8 M-1 reveal no pronounced aging effects for temperatures up to 1400°F, however, Creep Tests Long-time creep data on INOR-8 sheet specimens have been obtained at 1100, 1200, and 1300°F in molten salt. Several tests at 1100 and 1200°F are still in progress, after more than 10,000 hr, without reaching 1% strain. Tests which have been run at 1250°F in air indicate that the creep strength of INOR-8 is comparable with that of type 316 stainless steel. Comparison of tests in air and salt shows that there is no pro- nounced weakening effect due to the static salt environment. All heats, except 1566, exhibit similar creep properties when tested in salt. Heat 1566 is weaker at 1200 and 1300°F than any other heat. Relaxation Tests Relaxation data are available from 1200 to 16C0°F. At 1200°F the relaxation rate 1s appreciable, For example, the maximum stress supported after 2000 hr at 1200°F is 4000 psi. Dimensional Instability Relaxation and low-stress creep tests have indicated that some dimensional instability may occur in certain heats of INOR-8; dilatometry 5MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 67. 66 has supported this conclusion. The change in the linear dimension, however, may be either positive or negative depending on the test. In any case, the change is so small as to be negligible in most instances. Fatigue Studies Fatigue studies have been conducted on SP-19 by the Battelle Memorial Institute under a subcontract. Three temperatures, 1100, 1300, and 1500°F, were employed. Two frequencies, 100 and 3000 cps, were used. The results show that when the grain size is fine the alloy has good fatigue strength at elevated temperatures. No frequency effect was observed at 1100°F, but at 1500°F a pronounced difference occurs. The high-frequency tests endure a greater number of cycles before failure. it was discovered that there was a marked difference in the grain size between two billets of heat SP-19 used in preparing the fatigue speci- mens. As expected, the fine-grain material exhibited superior fatigue resistance. Carburization In comparing the effect of the nominal carbon level on the tensile properties, it has been noted that heats 1327 and 8 M-1, which are high in carbon, exhibit only slightly higher tensile strengths than the low- carbon heats. The yield strength is significantly increased and the tensile ductility is lowered by the increased carbon level, Studies of the tensile properties of carburized specimens (sP-19) possessing a high-carbon case also indicate that the yield strength is increased while the ductility is reduced by carburization. Creep tests on carburized and homogenized specimens have been per- formed at temperatures between 1300 and 1800°F. These tests show that carburization increases the creep strength. The degree of strengthening increases with the test temperature. A comparison of the creep properties of the high-carbon heats, 8 M-1 and 1327, with SP-16 and SP-19, however, reveals that no significant strengthening or weakening occurs below 1300°F. 67 Materials Fabrication Studies Triplex Heat Exchanger Tubing Work is continuing on the development of techniques for cladding prefabricated porous nickel tubes on the outer and inner surfaces with Tnconel or INOR-8 to provide triplex heat exchanger tubing with a leak- detecting component. The initial attempt to clad with Inconel was carried out with a sample piece of grade F porous nickel (0.0008-in. mean pore opening) received from Micro Metallic Corporation. The method of com- ponent assembly and the cold-drawing schedule have been reported.l® After drawing, the triplex assembly was sintered in vacuum for 2 hr at 2150°F to promote bonding between the porous core and the cladding tubes. Sub- sequent metallographic examination of the sample showed the interface be- tween cladding and core to be clean, with evidence of bonding and reten- tion of porosity in the core. Figures 2.1.8 and 2.1.9 show representative sections of interfaces in the assembly. Recently 25 £t each of grades E and G porous nickel tubing (0.015- and 0.00k-in. mean pore opening, respectively), 5/8-in. ID x 1/16-in. wall thickness, were received from Micro Metallic Corporation for addi- tional cladding studies. The porous tubing was formed from powder- metallurgy sheet stock. An experiment is in progress to determine whether a 9-ft triplex tube can be fabricated. Six 18-in. lengths of grade E porous nickel tubing were placed end to end over a 5/8-in.—OD Inconel tube. Figure 2.1.10 shows an enlarged view of a porous tube in place. A 0.913-in.-0D Inconel tube was slipped over the porous nickel, and the assembled triplex was reduced about 13% in area by cold draving. Two 1-in. samples were cut from the tube and sintered Tor 1 hr at 2150°F and 2200°F to determine whether this would be sufficient time to bond the cladding to the core. Metallography showed that both samples were bonded; therefore the remainder of the 9-ft triplex was evacuated and sinteregd at 2200°F by drawing through the hot zone of the furnace at the rate of 18 in./hr. *6MSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 67. 68 UNCLASSIFIED Y-29736 . A Fig. 2.1.8. Triplex Tube Cladding Experiment. Interface between outer Inconel cladding and pre- fabricated porous nickel core (0.0008-in. mean pore opening). Assembly was cold drawn and sintered in vacuum for 2 hr at 2150°F. Transverse section as polished. Fig. 2.1.9. Triplex Tube Cladding Experiment. Interface between inn Y-29733 ; & S INCHES L T 100X er Inconel cladding and prefabri- cated porous nickel core (0.0008-in. mean pore opening). Assembly was cold drawn and sintered in vacuum for 2 hr at 2150°F. Transverse section as polished. 70 UNCLASSIFIED Y-30275 Fig. 2.1.10, Prefabricated Porous Nickel Tube Surrounding a z-ln.-OD Inconel Tube. The outer Inconel tube is not shown. 71 After sintering, the triplex was radiographed, and two of the five core-to-core Jjoints were found to be separated to a maximum of 5/8 in. It 1s not known whether the core separations occurred during assembly, the pointing operation (swaging), or drawing. Future experiments will be designed to determine the cause and remedy of this difficulty. Addi- tional studies on the 9-ft triplex will include evaluation of bond quality, permeability of the core, and ability of the tube to be bent intoc a U- configuration of approximately 26-in. radius. Welding and Brazing Studies Welding of INOR-8 A summary report covering the welding characteristics and weld-metal mechanical properties of Hastelloy B, Hastelloy W, and INOR-8 has been com.pleted.l7 Studies are continuing in an effort to improve the high-temperature ductility of INOR-8 weld metal. Small additions of aluminum, titanium, manganese, silicon, boron, and magnesium to the vacuum-melted basic charge of INOR-8 resulted in a slight improvement of the weld-metal ductility at 1500°F.*® A more substantial improvement has been accomplished, however, with a 2% niobiun addition to the basic charge followed by deoxidation and purification with aluminum, etc., as described above. A reduction in the carbon content from 0.06 to 0.05% also resulted in a marked improvement in ductility with only a minor reduction of the tensile strength. The results of mechanical-property tests on several heats of INOR-8 weld metal are summarized in Table 2.1.7. In view of the highly promising characteristics of the alloy containing 2% nicbium, approximately 90 1b of this composition has been vacuum cast for fabrication into inert-arc weld wire for subsequent experimentation. A portion of this wire will also be submitted to various vendors for the application of electrode coatings suitable for metallic-arc studies. *7G. M. Slaughter, P. Patriarca, and R. E. Clausing, Welding of Nickel-Molybdenum Alloys, ORNL-2760 (July 1959). L8MSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p T70-72. 72 Table 2.1.7. Mechanical-Property Studies on INOR-8 Weld Metal Basic melt charge: 72% Ni—~16% Mo—4% Fe—8% Cr Additions to Elongation (% in 1 in.) Basic Charge At Heat No. (wt %) Other Additions Roor At At 1200° ° Carbon Other Temp. 0O°F 1500°F Westing- 0.08 36 18 5 house M-5 Haynes SP-19 0.06 38 18 10 ORNL MP-3 0.06 None Al,Ti,Mn,Si,B,Mg L6 25 173 ORNL MP-9 0.06 2 Nbo Al,7i,Mn,Si,B,Mg 2 2L 31 ORNL MP-12 0.03 None Al,Ti,Mn,Si,B,Mg 52 2L 20 ORNL MP-16 0.06 None None Being Determined The influences of small variations in the amount of silicon, titanium, and magnesium additions upon the ductility at 1500°F were also investigated, but the results to date are inconclusive. The ductility at 1500°F, however, appears to be very sensitive to small variations in weld-metal chemistry. The chemical factors should therefore be carefully controlled, and it is highly desirable that each particular heat be evaluated tc determine its suitability for high-temperature service. It also appears that the welding procedure should be controlled to minimize dilution of the base metal. 73 2.2 CHEMISTRY AND RADIATION EFFECTS Phase Equilibrium Studies The System Nal-ThF,-UF, A preliminary phase diagram (Fig. 2.2.1) has been obtained for the system NaF-ThF,-UF,. The close similarity between Th** (radius 1.05 K) and U*t (radius 1.10 E) results in extensive solid-solution formation, which is of some practical advantage since it decreases the extent of segregation on cooling; also, from the standpoint of crystallography and phase behavior, there are many surprising complexities, several of which are not yet resolved. For example, in spite of dissimilar stoichi- ometry, the compounds *NaF.UF, and 4NaF:ThF, form solid solutions, 3NaFs (U, Th)F,-4NaF-(U,Th)F,, between both the high- and low-temperature modi- fications of these compounds. These so0lid solutions appear to exhibit a continuous primary-phase field along the 3NaF.UF4-4NaF+ThF, composition line. The binary compounds 3NalF.UF, and MNaF-ThF4 have no equilibrium stability at low temperatures, nor do the ternary solid sclutions. Other examples of unusual behavicr are the following. Phase relationships along the line between 2NaF.ThF, and 2NaF.UF, involve an immiscibility gap between compounds of the same formula but of different crystal structure, as shown in Fig. 2.2.2. The compounds 3NaF.2ThF, and 5NaF.3UF, form face-centered-cubic crystals with unit cell edges differing by less than 0.5 A. The two compounds form solid solutions, but, on the basis of preliminary evidence, do not appear to have a continuous primary-phase field in the system NakF-ThF4-UF4. Both compounds are unstable at low temperatures, and the solid solutions between them are not preserved in slowly cooled melts. A large primary-phase field of the sclid sclution 7NaF-6(U,Th)F4 occurs, although the compound TNaF.6ThF, does not appear as an equilibrium Phase in the system NaF-ThF,. Although the compounds NaF.2ThF, and NaF-2UF, are not isostructural, some solid solution does occur along the NaF.2ThF,-NaF.2UF, composition 74 UNCLASSIFIED ORNL-LR- DWG 39407 'I'hF;l M 7y (2) COMPOSITION IN mole %o TEMPERATURE IN °C FP=PERITECTIC F=EUTECTIC -NC!F'ZThEl % OJ\O £ 834 / NaF-The, 7NaF- 6Th§ P 730 £ 705 .Q 3NaF- ZThF;l z £ 690 © %, 2ZNaF-ThE 5 4 A (9] Ks) 00 / o, Og ok & OO S < Y &) X /oq C N1 618 3NGF-U:4//3648// F673 TNoF-6UE, “£680 “NoF-2UF, 1035 £ 6297 2NaF-UF,” ONaF-3UF, Q0 QO O ™ Fig. 2.2.1, The System Nch-ThF4-UF4 (Preliminary Diagram). TEMPERATURE (°C) UNCLASSIFIED ORNL-LR-DWG 34986 R~ //82—2N0anf4+lJoum 700 ~ 8- 2NaF - U(Th)F,, - f SOLID SOLUTION AND LIQUID | 5NaF - 3UF — 7 . 675 ‘§;i:\\\\:;\\f~a.§‘\\~ ANDLJOwD\\\\ \ k\; 650 ;Ss \ 5NaF - 3UF \\ +§- 2NoFlKT ANDlJoum 625 \ B, - 2NaF - Th{U)F,, \\\\\ ,‘\\\\\\ 600 — +8-2NaF - U(Th)F, Sad SOLW)SOLUTDNS \\ _ , 8 - 2NaF - U(Th)F, SOLID SOLUTION 575 \\\ 550 SOLID SOLUTION 525 — \ 500 2NaF - ThF, 5 10 15 20 25 UF4 (mole %) Fig. 2.2.2. The Section 2Na F-ThF4-2Na F-UFA- 30 2NaoF-UF, 76 line. The compound NaF.2UF,, subsolidus in the binary system, does not appear as a primary phase in the ternary system. The System NaF-BeFo-ThF, Recent results from thermal analysis and thermal-gradient quenching experiments have shown that in the system Nal-BeFs-Thi'y the compounds NaF.BeFs and ThF, form a quasi-binary system, as illustrated in Fig. 2.2.3. The system contains a ternary compound, NaF-.BeF5+3ThF4. The occurrence of the primary-phase field of this compound has been reported previously.l The compound NaFe.BeFs+3ThF, may be sald to melt semicongruently in the ternary system NaF-BeF--ThF,, although i1t melts Incongruently in the quasi-binary system NaF-.BeFo-ThF,; this emphasizes the point that the maximum temperature on the boundary line between the primary-phase fields of ThF, and NaF:BeFs-3ThF, occurs at the peritectic on the quasi-binary, that is, at NaF-BeFo-ThF, (38-38-24 mole %) and T730°C. The eutectic in the system NaF+BeFo-ThF, has the composition NaF-BeFs-ThF, (49-49-2 mole %), mp 374°C. This eutectic invariant point is also the maximum temperature on the boundary line between the primary-phase fields of NaF+BeFs and NaF.BeFs+3ThF, in the ternary system NaF-BeFo-ThF,. The compound NaF:«BeFs+«3ThF,, which has not been described in the literature, has now been defined with respect to optical properties and x-ray diffraction pattern. The System NalF-Pulg Investigations on the NaF-PuF5 system are part of a study of phase relationships in the ternary systems NaF-BeFs-PulF's and LiF-BeFo-Pul's, which have been found to dissolve sufficient PuFg at 550 to 650°C for use in reactors.® Preliminary thermal analysis data obtained with NaF-PuF5; mixtures containing 5 to 30 mole % Pul';, supplemented by micro- scopic examination of the slowly cooled melts, indicate the existence of a eutectic between NaF and a binary compound presumed to be NaPuf',. The eutectic composition melts at 727 % 5°C, the lowest ligquidus temperature in the binary, and contains between 20 and 25 mole % PuFs,. IMSR Quar. Prog. Rep. April 30, 1959, ORNL-2723%, p 79. ZMSR Quar. Prog. Rep. April 30, 1959, ORNL52723%, p 80. TEMPERATURE (°C) 77 UNCLASSIFIED CRNL-LR-DWG 39406 1200 1100 Vi LIQUID 1000 / 300 / LIQUID + ThF, 500 / 700 / 600 L £ '_ M Liquic + L ThFs + NaF-BeF, 3 ThF, s NaF-BeF, 3 ThF, o 500 2 400 300 NoF-Bef, + NaF-BeF,-3ThF, 200 NoF-BeF, 20 40 60 80 ThF, ThF4 (mole %) Fig. 2.2.3. The System Na F°BeF2-Th Fy 78 Cryoscopy in NaF As an aid in correlating the thermcdynamic properties of molten fluorides with the constitution of the fuels, cryoscopic measurements have been carried out on 120-g samples of NaF solutlions contained in a freezing point apparatus made of nickel and having a diameter of 2 in, Freezing points were found with an estimated accuracy of +0.6°C from cooling curves obtained with a Pt—Pt-Rh thermocouple inserted in a well extending 2 in. below the surface of the liquid. Temperature and cool- ing rates were regulated manually by means of an autotransformer in con- junction with a S5-in.-dia pot furnace and a nickel-block thermostat. The melts were protected from the atmosphere and stirred with argon. The effect of alkaline earth cations, shown in Table 2.2.1, reflects the increasing deviation from ideality found for smaller solute lon size. Most simply interpreted, fluoride ions are more strongly complexed the smaller the cation. Table 2.2.1. Cryoscopic Measurements in NaF (mp 995°C) Freezing Point Depression Calculated from Ideal Solutions Freezing Point Depressions (°C) Caused by: Concentration of Alkaline- Earth Solute (mole %) (°C) BeFo MgFso CaF> SrFo BaFo 2.0 8.0 8.9 8.1 8.2 8.6 8.5 5.0 20.2 24 .4 22.2 22.0 20.3% 20.5 10.0 L0.9 55.8 L9.8 hs.3 Wz 4 42,8 15.0 62.2 102.7 90.5 71.9 69.2 66.1 20.0 84,2 162.0 127.5 101.0 91.6 89.0 Gas Solubilities in Molten Fluorides Solubility of ¥Xenon in LiF-BeFo The solubility of xenon in LiF-BeFy (6L-36 mole %) has been deter- mined over a pressure range of 0 to 2 atm at 500 to 800°C. Plots of these 79 data as a function of pressure indicate that Henry's law is followed and that the solubility constants are 1.55 X 107°, 2.%3 x 1072, 3.33 x 1079, 5.05 X 1072, and 8.63 x 107 moles of Xe per cubic centimeter of melt per atmosphere at 500, 600, 700, and 800°C, respectively. A heat of solution of 12,000 cal/mole was calculated from these constants. Solubility of CO- in NaF-BeFs The relatively low melting point cof NaF-BeFs (57—&5 mole %) per- mitted measurements of the solubility of CO- at temperatures as low as 415°C; these were of interest in confirming the previously reported® minimum observed in the plot of the logarithm of the Henry's law con- stant vs the reciprocal of the absolute temperature. Saturating pres- sures from O to 2 atm of CO- at 415°C gave rise to a Henry's law con- stant of (10.36 % 0.45) x 1078 moles of CO- per cubic centimeter of melt per atmosphere, compared with 7 x 10-® at 600°C. Thus, from L15 to 600°C, CO2 exhibits the usual negative temperature coefficient of solubility, as does HF in fluoride solvents; at temperatures greater than 600°C, CO» has the less familiar positive temperature coefficient encountered with the noble gases. Fission-Product Behavior Sclubility of Rare-Earth Fluorides The solubility data for CeFs in LiF-BeFo-UF, (73-6-21 mole %) and LiF-BeFo-ThF,-UF, (72-15.5-11.5-1 mole %) in Table 2.2.2, obtained from radioactive tracer measurements on filtered samples of saturated liquid, should serve as reasonable estimates of the total solubility in mole fraction of any combination of rare-earth fluorides in these fuel com- positions. It has been previously demonstrated that the rare-earth cations are approximately equivalent from the standpoint of solu.bility.4 SMSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 85. *W. T. Ward et al., Solubility Relations Among Rare Earth Fluorides in Selected Molten Fluoride Solvents, ORNL-2749 (in press). 80 Table 2.2.2. Solubility of CeFs in LiF-BeFo-UF, (73-6-21 Mole %) and LiF-BeFo-ThF,-UF, (72-15.5-11.5-1 Mole %) . CeF5 Solubility in CeFs Solubility in Temp. (°C) LiF-BeFo-Ul'y LiF-BeFo-ThF 4-UF 4 Weight % Mole % Weight % Mole % 700 7.8 3.6 9.0 3.1h 600 3.9 1.8 L.32 1.48 500 1.91 0.86 2.05 0.68 On a molar basis, solubilities of CeFs in solvents having high con- centrations of thorium and/or uranium are approximately three to four times higher than those previously found in the LiF-BeFp (63-37 mole %) . solvent, An effect in this direction was expected because the higher LiF concentration (basic constituent) is conducive to stronger complexing . and enhanced solubility; however, a partial neutralization of such an effect, due to the substitution of ThF, and UF, (acidic) for BeFp (acidic), was not as large as expected, indicating that ThF, and UF, are not much stronger acids than BelFs. Precipitation of Rare-~Earth Fluorides by Cooling Exploratory experiments directed toward fuel reprocessing have been made on the feasibility of removing rare-earth fluorides from a fuel by means of a '"cold finger." Two devices were tested, one -using hydrogen gas as a coolant, the other using water. Seven of nine tests gave useful results, but the apparatus was not amenable to precise control. On the average, 0.16 g of salt containing 25 wt % CeFs was frozen and removed from a comparatively large amount of LiF-BeFo (63-37 mole % with 4 wt % CeFs) during test periods which did not exceed 2 hr. Varying the type of coolant for the same area of the . cooling surface seemed to have no effect. However, the transfer of heat from the flange top to the cold finger had a marked influence on the 81 efficiency of the process. Attempts will be made to insulate the finger from the flange for subsequent tests. Another type of experiment, also related to reprocessing schemes to remove rare-earth fluorides, is under way; this is the measurement of the kinetics of the exchange Ce*¥F5(d) + LaFs(ss) = LaF5(d) + CexFs(ss) where d and ss denote dissolved and solid solution. Reactions of UF, with Oxides in Molten Flucoride Solvents The reactions of various oxides with components of a fused-salt reactor fuel are being investigated to determine the feasibility of fuel reprocessing by oxide precipitation reactions and to develop an under- standing of the behavior of these oxides with respect to reactor opera- tion. The reactions of water vapor and beryllium oxide with UF, in molten fluorides are of current interest. The behavior of wvariocus structural-metal oxides as well as fission-product oxides 1s also of interest to reactor fuel technology. Reaction of UF4 with BeO The precipitation of U0z by reaction of BeO with UF, in fused salt solvents has been considered as a reprocessing scheme, The process has been studied by an extraction column method,® and it has been shown that the reaction is controlled by the surface area of the BeO particles.® Three experiments were conducted to study the efficiency of this reaction by contacting BeO particles (>%25 mesh) with about 8 wt % of UF. in the solvents LiF-BeF» (63-37 mole %), LiF-KF (50-50 mole %), and LiF-NaF (60-40 mole %). With a 50% excess of BeQ, the reaction was about 96% complete within 30 min in each case. These reactions are illustrated in Fig. 2.2.4., The uranium concentration at the completion of these experi- ments was 1520 ppm in LiF-BeFp, 355 ppm in LiF-KF, and 190 ppm in LiF-NaF, SMSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 83. SMSR Quar. Prog. Rep. Oct. 31, 1058, ORNL-2626, p 92. URAN|UM REMAlN\NG (/o) URAN[UM REMA|N|NG (0/0) URANIUM REMAINING (%) 100 n o 100 w @] 100 n (@] 82 UNCLASSIFIED ORNL-LR-DWG 40627 | 6.2 \ LiF - BeF, (63-37mole %) ‘ | \ | 5.5 | : LiF ~KF (50-50 mole %) ! ! : | i i . { : 6.5 I i 1 i 11 ! LiF-NaF {60-40mole %} 50 100 150 200 BeO ADDED (%%} Fig. 2.2,4, Reaction of BeO (>325 Mesh) with UF4 in Molten Fluorides. URANIUM IN LIQUID {wt %) URANIUM IN LIQUID {wt %) URANIUM IN LIQUID {(wt %) 83 In a separate experiment, the addition of a fivefold excess of BeQ (>325 mesh) to a mixture containing CeFs (2 wt %) in LiF-BeFs (63-37 mole %) failed to precipitate a significant amount of cerium, as deter- mined by radiochemical analysis of the filtrate. These experiments are further demonstrations that beryllium oxide might be effective in separating uranium from the rare-earth fluorigde fission products in a fused-salt reactor fuel. Reaction of UF, with Strontium Oxide A systematic investigation of the reactions of various metal oxides with UF, and CeFsz in mixtures of LiF-BeFs (63-37 mole %) and LiF-NaF (60-40 mole %) is being continued. The reaction of strontium oxide with UF, and CeFs in LiF-NaF (60-40 mole %) at 750°C is illustrated in Fig. 2.2.5. These results, showing sharply selective metathetical end points, are consistent with those obtained for the reactions of Ca0 and BaO with UF, and CeFs in this same solvent. Reaction of UF, with Structural-Metal Oxides? The reactions with the oxides of various structural metals used in equipment containing fused fluoride mixtures are of interest in connection with maintenance problems and other opportunities for oxide contamination. In three experiments, NiO, Fep03, and CrpOs were added independently to a mixture containing approximately 1 mole % UF, in LiF-NaF (60-40 mole %) at 750°C. The results are illustrated in Fig. 2.2.6. Equilibrium con- ditions were probably not attained, and further experiments will be re- quired to explain the differences in reactivity. Reactions of UF, with Air An experiment designed to study the reaction of air with a mixture of LiF-BeFs-ThF,-UF, (71-15-13-1 mole %) was accomplished by bubbling dried. alr through the fused mixture in a nickel container at 600°C. Samples of the mixture and of the effluent gas were taken periodically during the passage of 600 liters of air through the molten salt. An apparent "Experiments performed by R. St. James. METAL FOUND (wt %) 84 UNCLASSIFIED CRNL-LR-DWG 40628 URANIUM A Afl L CERIUM \ \ bt A \ X END POINT — FOR Ce F3 THEORETICAL END POINT FOR UF, \ ’ .\__H 100 200 300 400 500 SrO ADDED (megq) Fig. 2.2.5. Reaction of StO with UF , and CeF in LiF-NaF (60.40 Mole %) at 750°C, 85 saturation of the melt with nickel (approximately 3000 ppm) and a con- stant though small consumption of oxygen were noted. Examinations of filtered and unfiltered salt samples by x-ray diffraction and petro- graphy indicated no apparent alterations of the melt by the imposed con- ditions. Further experiments are planned to check the preliminary in- dications that air reacted with the wall rather than the melt, producing NiO, and that the reaction of NiO with the melt was too incomplete to give a detectable precipitation of U0s. Chemistry of Corrosion Processes Solubility and Thermodynamic Properties of NiF, in LiF-BeFs The solubility of NiFp in LiF-BeF- (62-38 mole %) is of considerable interest since it has been established that pure NiF,, rather than a com- pound or solid solution, is the saturating phase. This means that the measured equilibrium quotients for the reaction NiFs(d) + Ho =2 2HF + Ni which provide a value for the free energy of formation of NiFo in solution, can be extended to the saturated solution to yield the free energy of for- mation of so0lid NiFp and thus furnish an experimental check of the litera- ture value. The evidence that pure NiFs is the saturating phase was obtained from petrographic and x-ray examination, as well as from the fact that no difference in saturation solubility was obtained regardless of the amount of NiF, added. The recently measured solubilities (precision i3%), supplanting pre- viously reported lower values® for NiFp in LiF-BeFs (61-3%9 mole %), are given in Table 2.2,3, The slope of the straight line obtained from plotting log NNiFE Vs 1/T yields a heat of solution for NiFz of 9900 * 600 cal/mole. ®MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 93. 86 UNCLASSIFIED ORNL-LR-DWG 40629 100 % [ J . . o \ f | S \\\\ — A\ [©] > 60 z \\\ | q | E | \‘ i o | \ | % 40 Fe203 z < o ] 20 0 50 400 150 200 250 OXIDE ADDED (%) Fig. 2.2.6. Reactions of Structural Metal Oxides with UF, in LiF-NaF (60-40 Mole %) at 750°C. Table 2.2.3. Solubilities of NiFs, in LiF-BeFs (62-38 Mole %) Concentration Temperature (°c) ppmn Nitt Mole Fraction 87 It is interesting to note that if a previously reported® value for P2 K_N= HE = 2 x 10% NNiF2 PH2 in 2 x 1072 mole fraction NiFo solution at 600°C is assumed to hold for the saturated solution, the standard free energy of formation of solid NiFz is -123 kecal/mole, compared with a published value of -125 keal/mole. The calculations leading to this result imply that the activity coefficient of NiFp, based on pure supercooled liquid as standard state, is about 40O0. Such a marked positive deviation from Raoult's law ideality is difficult to reconcile with previous experience, which indicated that no marked posi- tive deviations are encountered in solutions of this type. The discrepancy, if the limitation of positive deviation is a real one, might be rectified by a higher value for Ky, a higher solubility, or a lower standard free energy of formétion of HF. A higher value of Ky is the more likely ex- planation; otherwise Ni++ in fuel solvents is highly exceptional in its behavior. The value of Ky in the saturated solution is difficult to measure by current techniques because of high HF and low Hs content of the equilibrium gas. Activities from Vapor Pressure Measurements A program to study the activity of uranium tetrafluoride in binary systems with an alkali fluoride as a function of temperature and composi- tion has been renewed. The first investigations will consist in vapor pressure measurements to establish the region of applicability of the technique. The limiting difficulty is association in the vapor phase, Cryoscopic measurements will be used to supplement the vapor pressure data. Preliminary vapor pressure measurements on LiF-UF, at 80 and 90 mole % UF, showed small negative deviations from Raoult's law, as expected. Assuming the vapor phase to be pure UF,, the activity coefficients of UF, are in the range 0.84 to 0.87. °MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 93. 88 Tracer Analyses for Determining Efficacy of Reducing Agents Previous work directed toward improving the reliability of analytical procedures related to corrosion studies!® demonstrated that 7r° would com- pletely remove labeled FeFp dissolved in an NaF-ZrF, (53-47 mole %) mix- ture. This type of study was extended to the reduction of the labeled FeFs- by Ho and Cro at 600°C. With Hp, the Fe content was reduced from 950 to approximately 10 ppm in three days, and to approximately 2 ppm in six days. With Cro, 690 ppm of Fe in solution was reduced below the limit of detection within 6 hr. Another experiment was performed to demonstrate the removal of CrFs by Zro at 600°C and to check the wet analysis data for low Cr contents; a solution containing 560 ppm of cr't was treated with zr°. Wet analyses confirmed the elimination of CrFs. However, for all the experiments, re- sults of wet analyses for the Fe content of the reduced solvents unaccount- ably ranged from 135 to 155 ppm, indicating that an unidentified source of contamination is still interfering with the analyses, Analyses of Corrosion-Test Loops Periodic sampling of the melts in three forced-circulation corrosion- test loops has continued. The chromium concentration in loop MSRP-12 (INOR—B), which reached a steady-state level of 550 ppm after 1200 hr of operation, is still maintaining that level after 4000 hr of operation. The salt charge is LiF-BeFo-ThF,-UF, (62.0-36.5-1.0-0.5 mole %). Loop 9377-5 (Inconel), which has shown an increase in chromium con- centration well beyond values possible without extraneous oxidation, may be approaching saturation. At 3100 hr of operation, the chromium content of the melt had risen to 3050 ppm. Unfortunately, a pump failure at this point caused the loop to be shut down, and the salt was drained into the sump. Since the sump had not been drained after the initial loading of the loop, considerable dilution of the loop salt resulted., Samples of the melt in the lcop were taken immediately after the pump was repaired and the loop started. The chromium concentration had dropped to 1800 ppm, *OMSR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 87. 89 indicating a dilution of about 30% in the shutdown mixing. After re- starting, the chromium content in the melt climbed to 2350 ppm in 72 hr. At 4000 hr of operation, the chromium had again reached the 3000-ppm level., The salt charge in this loop is the same as for loop MSRP-12. Loop MSRP-13 (INOR-8) has been started recently, using LiF-BeFo-UF, (70-10-20 mole %), After 900 hr of operation, the chromium content in the melt had risen from 160 to 210 ppm, which is lower than expected in view of the high UF, concentration. Stability of Zirconium Monochloride in Air A new cormercial product known as "Zirklor," stated to be a graphite- like product having the formula ZrCl, has been recommended as a lubricant and elastic packing material for use at temperatures up to 1000°C.11 There was reason to suspect that the product would be unstable in air, though the commercial literature stresses the marked thermel stability cf the material. An evaluation for MSR application revealed that samples gained weight in air at room temperature, showed evidence of severe attack on heating to 325°C, and reacted completely with air according to the eguation LZrCl + 302 — 3ZrOs + ZrCl, in 15 hr at 800°C. It was concluded that the material was unsuitable for uses which involve exposure to oxidizing agents. Graphite Compatibility Studies Long-Term Loop Tests of Graphite with Circulating Fuel In a study of the feasibility of using graphite as a moderator in direct contact with fluoride fuels, 56 samples of an impervious graphite were immersed in a flowing stream of LiF-BeFp-UF, (62-37-1 mole %) in a pumped loop (INOR-8) for one year at 1300°F. The fuel did not wet the graphite, and no penetration by the fuel could be detected by weight change. *'R. 8. Dean, Ind. Labs. 10(4), 45-L48(1959). 90 The graphite, after calipering and weighing, had been degassed in the loop under vacuum for 24 hr at 1100°F and repressurized with argon before charging with salt. A flow rate of 1.1 gpm was maintained, causing an effective head due to pumping of 10 psi. An additiocnal pressure of 3 psig, due to the helium cover pressure, resulted in a total pressure of 1% psig on the graphite. Dimensional changes for the thirteen l/2-in.— dia rods for which data were obtained corresponded to an average loss of less than 0.5 mil in diameter; this is close to the probable error of the measurements. Otherwise, there was no evidence of erosion. Weight losses, which varied from negligible to 0.05% and averaged 0.02%, could possibly be attributed to desorption of residual gases from the graphite; no statistically significant differences were noted between the thirteen l/2-in.—dia rods and eight 3/16—in.-dia rods for which weight data were available. Chemical analyses of the graphite are not yet finished, but the indications are that traces of uranium, of the order of a few parts per million, will be found. This may represent vapor phase deposition as a consequence of reaction with residual oxides in the graphite, and thus could presumably be avoided by pretreating the graphite with Ho at 600°C. The chromium content of the salt increased from 135 to 550 ppm during the year's operation. Permeability of Graphite by Molten Fluorides An initial objective was to saturate graphite with a high-melting inert salt, LiF-MgFs (67.5-32.5 mole %), in order to ascertain whether further penetration of the graphite by typical reactor fuels containing uranium was prevented or decreased.,t? Test rods, l/h, 1/2, and 1 in. in diameter by 3 in. long, were machined from AGOT graphite. The rods were mounted on a rack in a nickel can and degassed at 900°C for 24 hr in a 100-p vacuum. While still under vacuum, the rods were covered with the LiF-MgF, melt at 900°C, and then a helium pressure of 15 psig was applied., The temperature 12MsR Quar. Prog. Rep. April 30, 1959, ORNL-2723, p 88. 91 and pressure vere held at these levels for 48 hr, after which the pres- sure vas released and the salt was pumped away from the graphite rods. Welght gains of 12.5%, equivalent to filling about half the void volume, were obtained. To determine the distribution of the salt in the graphite rods, machine cuttings 1/52 in. deep were taken from the 1/2- and 1-in.-dia rods and analyzed for lithium and magnesium., The 1/L-in. rods were submitted for analysis without machining. Within analytical error, the salt had penetrated uniformly to the center of the rod with- out change in composition or concentration. Micrometer measurements of the rods before and after treatment showed no dimensional change. Since the LiF-MgF, mixture melts at 742°C and the normal operating temperatures of a reactor might be as low as about 675°C, it was of in- terest to socak samples of AGOT graphite rods, prepermeated with LiF-MgFo, in a fuel salt such as LiF-BeFo-UF, (62-37-1 mole %) at 675°C under 1 psig of helium for 1000 hr. The resulting weight changes were small (£0.2%), and since they were both positive and negative, little signifi- cance could be attached tc them. Micrometer measurements of the rods again showed nc dimensional change. Successive cuttings (1/32 in. deep) were machined from the 1/2— and l-in.-dia rods and submitted for analysis to determine the amount and depth of uranium and beryllium penetration. The 1/4-in.-dia rods were submitted for uranium and beryllium analysis without machining. Table 2.2.4 shows typical results obtained on the machined samples. It was obvious from the results that the uranium and beryllium were penetrating well, and that, in general, the beryllium appeared to diffuse into the prepermeated graphite faster than the uranium. There is also an indica- tion that the uranium concentration goes through a minimum, both abso- lutely and relative to the beryllium, between the outer and central por- tions of the rods. Additional experiments confirmed this general behavior for other types of graphite, such as TSF and code 82, and for slightly different fuel compositions. The minima in the uranium concentrations were most pronounced for the code 82 graphites, and the uranium concentrations in 92 Table 2.2.4. 1000-Hour Penetration of Fuel into AGOT Graphite Rods Preimpregnated with LiF-MgFso Rod Rod Cutting U Be Ratio, No., Diameter (in.) No.* (ppm) (ppm) U/Bex* 2 1/k 3870 2160 0.558 L 1/2 1 3550 4800 0.740 2 2460 Liks 0.553 3 1940 3545 0.547 L 1560 3105 0.502 5 1875 3255 0.576 6 1365 2710 0.50% 7 1070 2535 0.422 8 1080 2285 0.472 9 1210 2425 0.499 | 10 9355 2000 0.467 Center 1265 2160 0.586 ' 9 1 1 3600 5125 0.702 2 2045 L8360 0.606 3 2555 4615 0.553 L 2210 4190 0.527 5 1995 025 0.496 6 1530 2345 0.652 7 1685 2580 0.653 8 1695 2730 0.621 9 1645 2525 0.651 10 1410 2L55 0.57h 11 1130 2225 0.508 12 1135 2040 0.556 Center 1835 2485 0.738 *1/%2-in. cuts. **U/Be ratio in the fuel, 0.767, based on chemical analysis of the original salt batch, nominally LiF-BeFo-UF, (62-37-1 mole %): U, 6.84 wt %; Be, 8.92 wt %. 93 the center were so high as to suggest localized deposits of UOs. The code 82 graphites resisted penetration by the LiF-MgFs, mixture and also contalned relatively little fuel. For comparison with results on samples which were preimpregnated with LiF-MgFo, short-term experiments were run by using the techniques described above for impregnation, but substituting reactor fuels for the LiF-MgF», mixture. Slight modifications, such as in the degassing temperature, were made in the procedure to simulate probable reactor operating temperatures., The fuels permeated ordinary graphites to an extent corresponding to about lO% of the void volume of the graphite in test periods of 6 to 48 hr, and minima in the uranium concentration as a function of distance were usually found. Some of the samples were special graphites obtained from National Carbon: CB8-312-0 was untreated graphite; CS-312-2 was specially treated . to make CS-312-0 impervious; and R-0018 was a special low-permeability graphite. At the operating pressures involved, the impervious graphites resisted penetration, showing weight losses rather than gains. Graphite R-001% was a special high-density graphite which had received noc previous treatment to make it Impervious; it behaved about like AGOT with respect to permeation. Tests involving 1000-hr exposures are now in progress. Preliminary indications are that fuels containing 1% to 15 mole % T™hF, will penetrate graphite more readily, in response to pressure, than fuels con- taining UF, without ThF4. This result seems to be related to a tendency for UF4 to form UOp precipitates which partially plug the graphite; thorium tetrafluoride seems to make the oxide more soluble. Preparation of Purified Materials Purification, Transfer, and Service Operations The purification processing of molten fluorides of various composi- tions increased sharply during the quarter to approximately 120 kg of ¢ mixtures containing no beryllium fluoride and 510 kg of beryllium-con- taining mixtures. Transfer and service operations increased in a similar fashion. 94 Reactions of Stannous Fluoride Stannous fluoride, a useful oxidant for preparing simple fluorides from metals,m!l4 has also been found useful for preparing complex fluo- rides. Chromium metal reacted with a molten mixture of stannous fluoride and lithium fluoride to yield lithium hexafluochromate(III): 2Cr + %SnFa + 6LiF — 2LisCrFs + 3Sn The corresponding treatment of vanadium metal gave a product with an x-ray diffraction pattern slightly shifted from that of lithium hexa- fluochromate(III); so, presumably, it was lithium hexafluovanadate(III). Molten stannous fluoride reduced cupric fluoride, bismuth trifluo- ride, and antimony trifluoride to the free metals. The other product of these reactions, presumably stannic fluoride, appeared not to have melted at any temperature up to 500°C. These reductions are surprising, because published?> thermodynamic estimates give considerably unfavorable energy differences for their occurrence. Radiation Effects INOR-8 Thermal-Convection Loop for Operation in the LITR Assembly of the in-pile thermal-convection loop for testing fused salt fuel in INOR-8 tubing at the LITR was completed, and the loaded loop was installed in its outer can. This loop, which contains LiF-BeFp-UF4 (62-37-1 mole %), will be operated at a later date in the LITR. In-Pile Static Corrosion Tests Two fused-salt capsules have been in the MTR for three months. This is the longest test pericd for a fused-salt in-pile experiment to date. The two capsules are made of INOR-8 and contain Li’F-BeFo-U35F, (62-37-1 mole %). The capsules are operating at 1250°F and had an initial power 13MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 107. 14MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684, p 112, 154, Glassner, The Thermochemical Properties of the Oxides, Fluo- rides, and Chlorides to 2500°K, ANL-5750 (1957). 95 density of 1200 w/cc. They have accumulated approximately 1700 hr of operating time. It is estimated, based on flux and duration of the exposure, that approximately 50% of the U35 has been consumed by fission at this time. Irradiation of these capsules will be continued until the uranium in the fused salt no longer holds the capsule at operating tem- perature. After irradiation, the capsule will be subjected to metallo- graphic examination to determine the extent of corrosion. A total of 47 fused-salt capsules have been placed in the MIR, giving a combined operating time of 24,97L hr or 2.85 years; the average irradia- tion time was 531 hr. Most of these capsules were Inconel, contained zir- conium-base fluoride fuel, and were operated at 1500 or 1600°F., The amount of corrosion observed by metallography was considered negligible and was The same as that observed for control experiments operated without radia- tion. Four capsules contained beryllium fluoride in the fuel and were held at 1250°F for an average exposure of 1600 hr; two of these are still being irradiated. 96 2.3 FUEL PROCESSING Work has continued on the HF dissolution process for recovery of the LiF-BeFs carrier salt following uranium recovery by the fluoride volatility process. It is based on dissolution of LiF and Befp in about 90% aqueous HF, in which the major nuclear poisons are insoluble, followed by evaporation of the HF and water to reconstitute the carrier salt., An alternative process is dissolution in anhydrous HF, which gives better decontamination and LiF recovery but results in the loss of most of the BeF-. A flowsheet based on the 90% HF dissolution pro- cess has been presented previously.l The possibility of dissolving uranium along with the carrier salt, or at least the LiF, is under investigation. Simultaneous dissolution of the uranium and lithium fluoride would permit elimination of the volatility step for the fuel salt, resulting in & one-step process with a considerable saving in plant complexity, equipment, and operation., Experiments are also being carried out on the partial decontamination of reprocessed LiF from fission products that are soluble in HF. De- contamination of LiF from soluble fission products, such as cesium, will be required on a much slower schedule than rare-earth removal since the soluble fission products have smaller cross sections. Unless re- moved, however, they may eventually build up in the fuel salt until they are significant poisons. UFg Solubility in HF-ClFa Preliminary experiments indicate that the addition of ClFs to anhydrous HF oxidizes UF4 to UFg, which should be somewhat soluble in the HF solution of fused salts. The solubility of UFg in anhydrous HF at the bolling point2 is over 100 g per kilogram of HF and it increases *E. G. MacPherson, Molten Salt Reactor Program Status Report, ORNL-263k (Nov, 12, 1958). 2G. P. Rutledge, W. Davis, Jr., and R. L. Jarry, J. Phys. Chem. 57, 541 (1953). 97 steadily as ClFs 1s added,® but the large amount of LiF and BeFs in the solution of fuel salt would be expected to decrease this consider- ably. 1In laboratory experiments, the solubility of uranium from fuel salt (LiF-BeFp-ThF4-UF,, 62-36.5-1.0-0.5 mole %) in anhydrous HF-ClFs sclutions has generally been in the range 2 to 3 g/kg, and occasionally twice as high. However, in these same experiments very little of the LiF and BeF- was found in the solution. In order to check the effect of ClFs on LiF and BeF, sclubilities in HF, an experiment was carried out in which ClF,; was added to the extent of 25 wvol % to anhydrous HF saturated with LiF and BeFso. No precipitation occurred, and analysis of the final solution showed 100 g of LiF and 15 g of BeFs per kilo- gram. The discrepancy between the low results in the earlier experi- ments and the results in the later one 1s not understood, but it may be a result of the experimental apparatus. Although good solubility data have not yet been obtained for the ClF5~HF solvent system, the results indicate sufficient promise to war- rant further investigation. A flowsheet based on this system would differ from that presented previcusly! in the following respects: the fuel fluorination operation would be eliminated; the dissclver would cperate with anhydrous HF containing ClFq; all the LiF and uranium would be dissolved and recovered in the evaporator; but most of the BeFs would go with the waste, and new Bel, would have to be added to the evaporator or purification eguipment to compensate for this loss. In order to eval- uate this system properly, better solubility data must be cbtained for the fuel salt components, and the effect of ClF5 on fission product solubilities must be investigated. It should also-be mentioned that other solvent systems involving anhydrous HF as one component may dis- solve uranium from the fuel salt, and tests of some of these are planned. Separation of LiF from CsF A lsboratory experiment indicated that CsF can be separated from LiF by dissolution of the fuel salt in 90% HF followed by partial instead ®G. P. Rutledge and W. Davis, Jr., J. Phys. Chem. 63, 166 (1959). 98 of complete evaporation. A 90% HF solution containing 80 g of LiF, 48 g of BeFs, and 2 g of CsF per kilogram, together with tracer Cs'®7, was evaporated to about one-fourth its volume; a crystalline solid formed during the evaporation. The solution remaining was h2.5% HF, and it contained 65% of the cesium activity and only 2% of the LiF. Evaporation of 90% HF yields a decreased volume of more dilute HF, because the vapor is greatly enriched in HF, until the HF-Ho0O azeotrope is reached. The LiF solubility decreases rapidly as the HF concentra- tion is decreased,? and most of the LiF precipitates from the solution as it is evaporated. The major heavy fission product soluble in 90% HF is cesium, which would build up in the fuel until eventually it would become a significant poison. However, CsF is soluble in both water and HF, so it might be expected to remain in solution as the HF concentra- tion decreased during the evaporation. The experiment confirms that nearly all the LiF is precipitated and that most of the CsF remains in solution, as does the BelFs. 4MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 113. 99 ORNL-2799 Reactors — Power TID-4500 (1lhth ed.) INTERNAT. DISTRIBUTION 1. R. G. Affel L7. M. T. Kelley 2. L. G. Alexander 48. F. Kertesz 3. E. 8. Bettis 49. B. W. Kinyon L. D. S. Billington 50. M. E. Lackey 5. F. F. Blankenship 51. J. A. Lane 6. E. P. Blizard 52. R. 8. Livingston 7. A. L. Boch 5%. H. G. MacPherson 8. C. J. Borkowski sk, W. D. Manly 9. W. F. Boudreau 55. E. R. Mann 10. G. B. Boyd 56. L. A. Mann 11. M. A. Bredig 57« W. B. McDonald 12. E. J. Breeding 58. H. F. McDuffie 13. R. B. Briggs 59. J. R. McNally 14, W. E. Browning 60. H. J. Metz 15. D. 0, Campbell 61. R. P. Milford 16. W. H. Carr 62, L. C. Miller 17. G. I. Cathers 63. J. W. Miller 18. C. E. Center (K-25) 64, K. Z. Morgan 19. R. A. Charpie 65. J. P. Murray (v-12) 20. J. H. Coobs 66. M. L. Nelson 2l. F. L. Culler 67. G. J. Nessle P2, J. H. DeVan 68. W. R. Osborn 23. D. A. Douglas 69. P. Patriarca 2L, L. B. Emlet (K-25) 70. A. M. Perry 25. W. K. Ergen 7l. D. Phillips 26. J. Y. Estabrook 72. W. D. Reel 27. D. E. Ferguson 5. P, M, Reyling 28, A. P. Fraas 74, J. T. Roberts 29. E. A. Franco-Ferreira 75. M. T. Robinson %0. J. H. Frye, Jr. 76. H. W. Savage 31. W. R. Gall 77. J. L. Scott 32, A. T. Gresky 78. H. E. Seagren 33, J. L. Gregg 79. B®. D. Shipley 34-36. VW, R. Grimes 80. M. J. Skinner 37. E. Guth 81, A. H. Snell 38. C. 8. Harrill 2. E. Storto 39, M. R. Hill 83. C. D. Susano 40. E. E. Hoffman 84, J. A. Swartout 41. H. W. Hoffman 85. A. Taboada L2, A. Hollaender 86. E. H. Taylor 4z, A. S. Householder 87. R. E. Thoma 44, W. H. Jordan 88. D. B. Trauger 45. G. W. Keilholtz 89. F. C. VonderLage L6, C. P. Keim 90. G. M. Vatson 100 9l. A. M. Weinberg 92. M. E. Whatley 9%. J. C. White o, G. D. Whitman 95. G. C. Williams 96. C. E. Winters 7. J. Zasler 98-101. ORNL - Y-12 Technical Library, Document Reference Section 102-141. Laboratory Records Department 142. Laboratory Records, ORNL R.C. 143-144, Central Research Library EXTERNAL DISTRIBUTION 145. D. H. Groelsema, AEC, Washington 146. Division of Research and Development, AEC, ORO 147-723, Given distribution as shown in TID-4500 (15th ed.) under Reactors-Power category (75 copies - OTS)