(3¢ [T 3 4456 03b13LL 3 ORNL-2684 Reactors-Power o 4 = OJ«/BS/ MOLTEN-SALT REACTOR PROJECT " QUARTERLY PROGRESS REPORT FOR PERIOD ENDING JANUARY 31, 1959 CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION .} LIBRARY LOAN COPY ' DO NOT TRANSFER TO ANOTHER PERSON B If you wish someone else to see this : document, send in name with document and the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY operated by i UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION ORNL-2684 Reactors — Power TID-4500 (14th ed.) Contract No. W-T405-eng-26 MOLTEN-SALT REACTOR PROJECT QUARTERLY PROGRESS REPORT For Period Ending January 31, 1959 H. G. MacPherson, Project (oordinator DATE ISSUED MAR 4 1059 OAX RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the U.5. ATOMIC ENERGY COMMISSION : Lignames I 3 Y456 03b13kY 3 - 1iii - MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT SUMMARY PART 1, REACTOR DESIGN STUDIES l.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS A design study is being made of a 30-Mw, one-region, experimental, molten~salt reactor. A power plant arrangement has been developed that is based on fabricability and maintenance considerations. Designs have been prepared of a fuel sampling and enriching mechanism, system pre- heating devices, and off-gas system valves, A preliminary design and costeestimate study has been completed on a 315-Mw (e) graphite-moderated moltenesalt power reactor. In this reactor the molten-salt fuel flows through holes in the cylindrical, unclad, graphite moderator, which is contaified in an INOR-8 vesgsel. The low-enrichment (1.30% U-3°, initially) fuel mixture is LiF~BeF,, - UFM (70+10-20 mole %). Nuclear calculations relative to the experimental reactor have been completed in which the reactivity effects of various modifications were evaluated. It was estimated that the experimental reactor would have an over-all temperature coefficient of reactivity of -6.75 x 10"5 (6x/k) /°F., The effect of neutron and gamma-ray activity in the secondary heat exchanger of the interim design reactor of using the salt mixture LiF -BeF,, (63-37 mole %) as the coolant, in place of sodium, has been studied. It was found that the gamma-ray dose would be only one-half that expected from sodium, and that no appreciable neutron activity would be present. = v = The nuclear effects of processing molten-gsalt fuels to remove the rare-earth fission products by ion exchange with CeF3 have been investi- gated. A comparison with performance of the system utilizing the fluoride-voletility processing method showed that the ion=exchange method gives 50% lower average critical inventory and a 30% lower cumulative net burnup. Modifications have been made to the Oracle programs Cornpone and Sorghum to permit evaluations of heterogeneous reactors;, and Oracle program GHIMSR has been written to expedite heating caiculations. Follow- ing test calculations with the modified programs, initial and long-=term nuclear performance calculations were made for two - low-enrichment, one- region, graphite-moderated, unreflected, molten-salt reactors. The results indicate the effect of adding thorium to the fuel mixture. The fuel used in one calculation was LiF=BeF2=-UFh (70-10-20 mole %), as described above, and in the other the fuel wasg LiF=-BeF2-=‘I‘hFh (67-16-13 mole %) plus sufficient U235Fh=-U238.Fh to make the system critical. For a cumulative power generation of T200 Mwy, the critical inventory with- out thorium was 4386 kg and, with thorium, was 1416 kg. The cumulative net burnup without thofium was 1400 kg, and, with thorium, was 1083 kg. Calculations of the performance of a thorium-containing graphite- moderated and =reflected system were initiated. 1.2, COMPONENT DEVELOPMENT AND TESTING Investigations have been continued of salt-lubricated bearings for molten-salt pumps. Bearing and journal parts fabricated of INOR-8 are being subjected to numerous tests to determine optimum design and operating conditions. Equipment for testing salt-lubricated thrust bearings was designed and construction is nearly complete. A PK type of centrifugal pump is being prepared for service tests of molten-sslt lubricated bearings. The favorable results being obtained with thesge -V - bearings have resulted in the postponement of tests of the more complicated hydrostatic type of bearing. The modified Fulton-Sylphon bellows-mounted seal being subjected to an endurance test in a PK-P type of centrifugal pump has continued to seal satisfactorily for more than 10,000 hr of operation. Operation of an MF type of centrifugal pump with fuel 30 as the circulated fluid has continued to be satisfactory through more than 13,500 hr, with the past 11,500 hr of operation being under cavitation damage conditions. A small frozen-lead pump seal on a 3/16-in.-dia shaft has operated since the first 100 hr of almost 5000 hr of operation with no lead leakage. Test operation of a similar seal on a 3 1/4-in.-dia shaft has been initiated. Techniques for remote meintenance of the reactor system have been investigated further. Specifically, improvements were made in the freeze-flange Jjoints being developed for remote connection of molten- salt and sodium lines. Construction work continued on the remote-mainte- nance demonstration facility. Equipment for salt-to-salt heat transfer tests has been completed. Data are being obtained with which to determine heat transfer coefficients. A heat exchanger has been designed with which to evaluate triplex tubing for use in moiten-salt or liquid metal-to-steam superheater appli- cations. A small heat exchanger test stand will be modified for testing of the heat exchanger. Further tests of commercial expansion joints in molten-salt lines have demonstrated conclusively that such joints are fiot suitable. Typical failures -indicated, however, a high-stress point that could be eliminated by redesign. Operation of forced-circulation corrosion-testing loops has been continued, and most of the 15 loops have been modified to prevent freeze- ups of the salts being circulated. - V] = Operation of the first in-pile loop was started at the MIR on December 3, 1958 and had continued satisfactorily through 860 hr, when it was shut down because of the leakage of fission gases. The radiation release has been traced to a partially plugged pump-sump purge=-outlet line whicfi caused flssion gases to back-diffuse up the pump-sump purge- inlet line to a point outside the cubicle, where they escaped through a leak. The assembly of an improved in-pile loop is approximately 60% complete. 1l.3. ENGINEERING RESEARCH Modifications in the design of the capillary efflux viscometers and in the method of calibration have led to more precise data on the viscosities of the salt mixtures LiF-BeF,-UF, (62-37-1 mole %) and LiF-BeFE-U'Fh-ThFh (62-36.5-0.5-1 mole %). At the lower temperatures the viscosities are significantly lower than those previously reported. The solid-phase enthalpies and heats of fusion have been obtained for three LiCl-KCl mixtures. Fabrication, assembly, and initial calibrations are continuing on components for the forced-circulation loop designed to ascertain the presence and effect of interfacial thermal resistances in systemsg cir- culating BeFe-containing salts. Heat-transfer coefficient measurements will be made as a means of studying surface film formation. PART 2. MATERIALS STUDIES 2.1. METALLURGY Examinations of three Inconel thermsl-convection loops that had operated for long periods of time and one Inconel loop that had operated - Vvii = for 1000 hr have been completed. Anomalous results were obtained that are largely attributable to impurities in the circulated salt mixtures. No forced-circulation loop tests were completed during the quarter. Mechanical property tests have been made on Inconel and INOR-8 speci=- mens that were carburized by exposure for LOO0 hr at 1200°F in g system containing sodium and graphite. Carburization was found to decrease the tensile strength and elongation of INORe8. On the other hand, carburi- zation increased the tensile and yield strengths of Inconel and decreased the elongation. The results of chemical analyses of the specimens were used to establish relationships of carbon content vs distance from specimen surface for various time intervals. Similar specimens exposed for 400 hr to sodium containing graphite showed no evidence of carburization. Specimens of Inconel and INOR-8 were also exposed at l300°F to fuel 130 containing graphite for 2000 and 4000 hr. The Inconel specimens were attacked and showed reductions in mechanical strength. The INOR-8 specimens showed no carburization and only slight attack after 4000 hr. Further tests have been made in the study of the compatibility of graphite and molten-salt mixtures. Uranium oxide precipitation has occurred in all tests with fuel 130. The quantity of uranium precipitated appears to be an indirect function of the purity of the graphite. Attempts are being made to eliminate the precipitation by using purer graphite. Tests are also under way to investigate the penetration of graphite by fuel mixtures as a function of time, temperature, and pressure. The strength properties of INOR-8 at high temperatures are being investigated in tensile, creep, relaxation, and fatigue tests. Prelimi- nary data indicate that the creep properties in molten salts at 1200°F are the same as those in air. Tests of notch sensitivity have indicated that INOR-8 is notch strengthened at room-temperature and at l500°F. Rotating-beam fatigue studles are being conducted under subcontract at Battelle Memorial Institute. - viii - Fabrication and welding and brazing studies are being made of triplex- tube heat exchangers. The tubing consists of two concentric tubes with a porous-metal-filled annulus. Means are being sought for obtaining good conductivity across the annulus and adequate permeability of helium through the porous metal for leak detection. Prefabricated porous materials are being obtained from commercial sources for evaluation. Studies of composites of INOR-8 and type 316 stainless steel have shown the alloys to be compatible at temperatures up to 1800°F. Such composites therefore appear to be promising for use in fused salt-to- sodium heat exchangers. Various means for deoxidizing and purifying ingots of INOR-8 weld metal have been studied as means for improving the ductility of INOR=8 weldments. The ductility at lSOOOF has been increased from an average of 7% to 13 to 15%, and further improvements are thought to be possible. No increase in high-temperature ductility was obtained by decreasing the carbon content. A niobium-containing coated electrode, designated Inco Weld "A" electrode, was investigated for use in joining dissimilar metals. The studies showed that where the metallic-arc welding process could be used, such weld deposits would be satisfactory for joining dissimilar metals for high-temperature service. Molybdenum~to-Inconel joints were brazed with several materials in connection with the fabrication of a pump-shaft extension. All joints showed a tendency to crack because of stresses built up by the different thermal expansion properties of the base materials. 2.2. CHEMISTRY AND RADIATION DAMAGE Phase studies are being conducted to determine whether the NaF- BeFQ-UFh system has any advantages over the LiF-Ber-ThFh-UFh system. The need for liquidus temperatures of 55000 or lower in nuclear reactor o X = systems will restrict the ThFh concentration to the range 10 to 15 mole %. The substitution of UFh for part of the ThFh would be expected to lower the liquidus temperature. Plutonium trifluoride has been found to be sfifficiently soluble in LiF--BeF2 and NaF-BeF2 melts to form a fuel mixture for a high-temperature plutonium-burning reactor. An apparatus was developed in which l=-g samples can be used for phase-relationship studies. The system KF-LiF-BeF2 coolant. Such mixtures would have lower gamma activity than similar NaeF-base mixtures after irradiation. The mixture LiCl-RbCl is also being examined as a possible coolant. Experimental studies of the com- patibility of the chloride mixture with fuel mixtures have shown that is being investigated as a possible reactor no uranium compound would deposit as the primary phase at reactor temper- atures. In studies of figsion-product behavior, it was found that additions of UF, had no effect on the solubility of CeF3 in LiF-BeF, (62-38 mole %). In esnother experiment it was demonstrated that the addition of CeF, to remove SmF_ from LiF-BeFE-UFh mixtures was effective even 3 3 when‘the SmF3 was present in trace amounts. The behavior of oxides in molten fluorides is being studied as part of an effort to explore chemical reactions which can be adapted to £he reprocessing of molten- fluoride-galt reactor fuels. In the study of the chemistry of the corrosion process, the activity coefficients of the fluorides of structural metals in dilute gsolutions of molten fluorides have been measured. The data are useful in understanding and predicting the corrosion reactions which take place in systems in which molten fluorides are in contact with alloys of these metals. Techniques have been developed for determining the self-diffusion coefficients of chromium in chromium-nickel alloys with Cr51 as a radiotracer. The results of recent experiments have shown that the o X = grain structure has a marked influence on the gelf-diffusion rates of chromium in Inconel. Grain size appears to be a controlling factor. Hydrogen firing and annealing in helium had the same effects on the over-all diffusion rate. Salt samples taken from operating forced-circulation corrosion= testing loops with decreasing frequency were analyzed. Samples taken from an INOR-8 loop during the first 1000 hr showed a slow but steady increase of chromium from about 420 to 530 ppm. Samples from an Inconel loop showed an increase in chromium content from about 350 to 450 in 500 hr, The chromium content of neither salt has reached a steady- state value, Vapor pressurescfi'CsF-Bng are being measured to obtain values on which to base estimates of the vapor pressures of LiF-BeF2 mixtures., Activities in the LiF-BeF2 system cannot be obtained from vapor-pressure measurements because of complex association in the vapor phase. The vapor pressure of liquid UFh wag measured between 4 and 180 mm Hg (1030 to 1300°C). Further studies have been made of gaseous aluminum chloride as a heat transfer medium. Calculations indicate that very low pumping power would be required and that it would be useful as a turbine working fluid. Experimental investigations of the compatibility of aluminum chloride and structural metals have shown very little attack. Reactor-grade graphite impregnated with LiF=-MgF2 has been tested in LiF-BeF,-UF, at 1250°F for 800 hr and examined for uranium pene- tration. Complete penetration of beryllium ard uranium was found in decreasing amounts with distance from the surface. A series of tests was run to determine the effect of thermal cycling on salt stability. It was found to be important in the handling of beryllium~based fuels to ensure complete melting of a batch before any portion is transferred to another container or test rig. Thermal cycling under static conditions must be avoided. - X1 - The in-pile thermal convection loop that was being readied for insertion into the LITR was found to have a deposit at the bottom of the lower bend. Examination of the fuel from which it was filled revealed the presence of UO2 on solid pieces. A second loop is being assembled. Methods have been developed for synthesizing simple fluorides by reactions with stannous fluoride. The following fluorides have been MoF,, ALF,, FeFp, VF, and UF,. prepared: CrFE, 7 ) 3 2.3. FUEL PROCESSING The solubilities of fluorides of neptunium, corrosion products, uranium, and thorium were investigated to evaluate their behavior in the proposed EF dissolution process for recovering LiF-BeF2 salt. On the basis of this work, the recovered LiF-BeF, salt after processing 2 by the HF dissolution method would contain, at most, 0.001 mole % neptunium. The solubility data indicate that the fluorides of neptunium and possibly plutonium behave similarly to the rare earth fluorides. Separation of the recovered LiF--BeF2 from corrosion~product fluorides would probably be less effective. However, the solubility of any particular contaminant fluoride is usually less in the presence of others. - Xjiidi - CONTENTS SUWARY P B0 0O C P OB EOCE S BN AEADSSOPERVRIOCO0OCG0CSONCAIBS S LONGCOO0D0CHEOSASD000E0 PART 1. REACTOR DESIGN STUDIES 1.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS ssecaecescoces Design of a 30-Mw, One Region, Experimental Molten-Salt | ReacCtor seseeeecsescecnoesccncsococoescssosencooonooosoesceccoas Reactor System ceeececcccccscsscccossncanscsecoecsccososcoose Power Plant Arrangement coccoccococcccseecsssccscoococcoscsosssaoc Fuel Sampling and Enriching Mechanism ccccccccccaccoscecean System Preheating Device c.cccocecoccecoccocsccoecnoccsscascsn Off-Gas System Valves .seecsccsssoesosssscescssccosscscssocs Preliminary Design of a 315-Mw (e) Graphite-Moderated Molten- Salt PO’WGI‘ Reagtc‘?:o--o-tooooooooouoaoooooooooooooooooooooo.io Nuclear Calculations .esseescoccoscscscesacecescsecococoosceenssse Reactivity Effects in Experimental Molten-Salt Reactor c... Neutron and Gamma Ray Activity in the Secondary Heat - Exchanger of the Interim Design Reactor cccsecaecscsceoos Effect of Ion-Exchange Processing of Rare-Earth Fission Products on Performance of Interim Design Reactor cccesos Multigroup Oracle Programs for Heterogeneous Reactor - COmpu‘bationS P00 800000 GEOPE0EER00CO08 000000000 S0D0000C0O00D Oracle Program GHIMSR for Gamma-Heating Calculations ccceoe Initial and Long-Term Nuclear Performance Without Fuel Processing of Low-Enrichment, One~Region, Graphite- Moderated, Unreflected, Molten-Salt Reactor cececccscccss Initial and Long-Term Nuclear Performance Without Fuel Processing of One=-Region, Graphite-Moderated, Unreflected, Th232-Conversion, Molten-Salt Reactor seece. Initial Nuclear Performesnce of One-ge on, Graphite- Moderated, Graphite~Reflected, Th23<-Conversion, Molten- Salt Reactor .eescococcescessscsccoecosaosoecscocoososoanos 102. COMPONENT DEVELOPBEN.II.I AND lI’ESTING .UOUGOOOOO000000{)‘000.000000000 Salt-Lubricated Bearings for Fuel PumpsS coccccococcocceccsccccecocs Hydrodynamic Journal Bearings coceccceccosccocccsoccasscasncs Hydrodynamic Thrust Bearings ccccccoccococscscoscccosccescoccooo Test Pump Equipped with a Salt-Lubricated Journal Bearing sccococeescesccccococecesscccooonsooen0000006000000000 Hydrostatic Bearings ccoceesesscesccocoeccccccssccoesooccnnos Bearing Mountings ccccocccoscoccceeessccococscocneascosccocncococaa iii (WS 30 31 1.3. - Xiv - Conventional Bearings for Fuel Pumps ...... cecsecoerssosscesasa Mechanical Seals for Fuel PUmps ....eeceoess ceesssesansracnanas Labyrinth and Split-Purge Arrangement .......oveeees. ceasos Bellows-Mounted Seal ..ceceevn. cesossesseseassrannnns ceces Pump Endurance TeSting .voevoererseereoonccasacsesoosaconcsccasse ,e Frozen-Lead Pump S@8L . .icciooecooaesneeeasoosoonesasnacsonensss Techniques for Remote Maintenance of the Reactor System ....... Mechanical .Joint Development ..... s ereeccsanssecannns ceoen Remote Maintenance Demonstration Facility ...... cesecannns Molten-Salt Heat-Transfer-Coefficient Measurement .....eeeese.. Triplex-Tubing Heat Exchanger Development ......... sevessencacs Evaluation of Expansion Joints for Molten-Salt Reactor Systems .eiviveennnvace cscscessessarse s cesacasretssessesnoa Design, Construction, and Operation of Materials Testing Loops ceveunn cocetrmrers seseceans i reesasesansensee teocesnuae Forced-Circulation LOOPS tevevieceeeecnnsas ceenssossnsenesn In-Pile LOOPS sevsneerssteroonnnnannsnnnnasanans et enscans ENGINEERING RESEARCH v evennootoronnonensnsooecssssnnsonnessses Physical Property Measurements ........... Sesesenssserasasassns ViSCOSItY cvvieninniiieiineeetoonoeennssonnnsesonaaneesnes Enthalpy and Heat Capacity ceceeeeeenennens s easssasannsas Molten~Salt Heat Transfer StUdi€sS ceececcococsocscscoecooooososseos PART 2. MATERIALS STUDIES METALL[JRGY .....0“0&.00.00.00000..00000000000.0900.000GQQOOO.-O Dynamic Corrosion Studi€s cccoccccossesococccoecssstoscoscsscsa Thermal-Convection Loop TeStS cocoococcscosossosoesscccoosss Forced=-Circulation Loop TeS8t8 scccsesccescocososcescoooocaoo General Corrosion Studies sesececococoosocecnoooscnncoonoeossssas Carburization of Inconel and INOR-8 in Systems Containing Sodium and Graphite cceseecoccceccocceccosesososoeccosescsss Carburization of Inconel and INOR-8 in Systems Containing Fuel 130 and Graphite ccceoccoccscosscocecocccccoccosocsnso Uranium Precipitation From Fused Fluoride Salts in Contact with Graphite cccecececaccccaoscosccsoccoesncoao Fused Fluoride Salt Penetration Into Graphite coeoocccoccecoe Mechanical Properties of INOR=8 cccoococoocescassccooosscoocossos Materials Fabrication Studies cecococcccsccccocoscococessessccsse Triplex Tubing for Heat EXChangers scecceocccoccocoeececcsos High-Temperature Stability of INOR=8 .ceo0co0cco0cscecoccocoss Commercial Production of INOR=8 coocccococcocoscnsecosnoass Evaluations of Composites of INOR-8 and Type 316 Stainless Steel cccoooccvecooecescnscsasassnonsonnoossoo 2.24; 2.3. Welding and Brazing Studies cocoocosocococccccsescccscocescossesssa Heat Exchanger Fabrication coccococcocococococceccosccocococossocoses Welding Of INOR=8 ceeseecscscoccconocccoscossooscoononnsso Welding of Dissimilar Metals eecocococcococococococovesssococoscssco Pump Component Fabrication cecoceoccococcosococoecescocoscsoanscnos CHEN[ISTR.Y AND RADIATION DAMA-GE 0900 00S#000DO0O00CO0LO00D0OB00000OCQ0D0CO0OESE Phase Equilibrium Studies cecocesscceconcccccscosssocssccscossss Systems Containing UF) and/or ThF) sccccosssccossscsosssos The System LlF"PUF * 8006 CE0GE0D008000000008000000R0EDCOESD The System KF'-LiF- F2 ©08000008€000000000000000060000000DCOCO Compatibility of Fuel with Chloride Coolants ceoceococccecses Fisgion=Product Behavior «ocsoceocoeccssesscsscooecococscsoscnoccses Effect of UF) on Solubility of CeFy in LiF-BeF Solvents . Removal of Traces of SmF3 by the Agdition of CeF3 to LlF-BeF “UF 200009 SO0 0OSAO00DO0RSSICOVGO0BO00SEOARSEUOOOOD Chemical Reactions of Oxides with Fluorides cccccscocooccses Chemistry of the CorrOSion Process occecesecocooocscosesccecscons Activity Coefficients of CrF, in NaF—Zth cacsccoosesensoo Chromjum Diffusion in A110YS scsasvcscn seassoenas coasns Sampling of Operating LOOPS ceccscsnceossccsssecssessscsas - Effect of Fuel Composition on Corrosion Equilibria ....... Vapor Pressures of Molten Salts .ccoecoe soosocos ceosveessnansan BeFp MiXtures .scceccccssocessccsoesessoccscosesoasoesonns UEF'), soooscasooccsocsescosscoocnsoooasscnosonooescooonoooossosss Aluminum Chloride Vapor as a Heat Transfer Medium and Turbine Working Fluid :scccocscccesccscscassacnosasoscosocensnscossccsao Estimation of Thermodynamic Properties cccceccceccccsccses Corrosion of Nickel by AlCl. ccocscocecocoscacasecesosnons Permeability of Graphite by Moltef Fluoride Mixtures .cceccecsoes Effects of Thermal Cycling on Salt Stability ccsceocecccoeccacaance Radiation Damage Studies .ccocccesoesccoesaccsescascocessssssaesse INOR-8 Thermal-Convection Loop for Operation in the LITR .. In-Pile Static Corrosion TesSt .ecccsescsseccsssscsccscssesse Preparation of Purified Materials ccoesccoesessaocssccconcsssessas Fluorides of Chromium coesecsesescecscsscosoesascoscscnnescace Synthesis of Simple Fluorides by Reactions with Stannous Fluoride cccocosescecccesnsessscncsscosocesccsccocsossssos Experimental-Scale Purification Operations .cocseccosccososo Transfer and Service Operations cceeeeccococscoccocosccncecse FIEL PROCESSING 00..00.OQDOOQOAUQUOOOOOICOCO.0...0.0000.....9.0. Solubility of Neptunium in Aqueous HF Solutions sesecccsscsccss- Solubility of Corrosion-Product Fluorides in HF Solvent ccceces Solubilities of Uranium and Thorium Tetrafluoride in HF SOlvent 0 000 & 6 & 0000 0SS0 S S S 000 ES 8 E S SO0 DS SO 000 S8S S O0¢E BO0S SO0 OC & 6D A Part 1. Reactor Design Studies - 3 = 1l.1. CONCEPTUAL DESIGN STUDIES AND NUCLEAR CALCULATIONS DESIGN OF A 30-Mw, ONE-REGION, EXPERIMENTAL, MOLTEN-SALT REACTOR Conceptual designs of an experimental moltem-salt reactor and power plant arrangement have been prepared. In these designs, consideration has been given to fabricability and to the elimimation of awkward design situations. An attempt has been made to take into account all features of a finished power plant. Improvements in plant layouts and equipment design will, of course, result from further study. Reactor System. The proposed reactor consists of a 6-ft-dia spherical core, a heat exchanger above the core, and a sump-type fuel pump offset from the main vertical axis and located slightly above the heat exchanger, as shown in Fig. 1l.1.1l. Liquid fuel leaving the pump discharge is directed to the top of the heat exchanger, where it passes downward. Flow through the heat exchanger is parallel and countercurrent, with the fuel outside helical tubes containing an inert molten-salt heat transfer fluid. The fuel leaves the heat exchanger through a pipe that communicates with the top of the spherical core, passes downward through the center of the sphere, and reverses its direction of flow at the bottom. It leaves the core through an annulus at the top. This annulus feeds the fuel directly to the sump bowl of the fuel pump. A small portion (about 10%) of the total fuel flow is bypassed through a gas-stripping system. Directly above the heat exchanger, at the end of the pump-discharge diffuser, there is a plate perforated with small nozzles which spray the bypassed fuel above the plate and cause i1t to be intermixed with helium flowing to an off-gas system. This helium originates as a purge gas in the seal region of the fuel pump. The reactor fuel system may be filled with or drained of fuel through a dip line that communicates with two valves in series and a storage tank COOLANT COOLANT OUT UNCLASSIFIED ORNL-LR-DWG 35663 GAS BLOWER K -oe " HEAT —EXCHANGER a -3 GAS HEATERS . ¢ COOLER - . et TR .-.-- = A N i “i ‘ S . . .l | 4 |} NN ' . : “ . '_A a’ ... DEGASSING CHAMBER i ! . R AN . . a ) N * " ‘ .. ‘a B . . . ‘J‘ S * [} . . . o Py ey i “H A . i PURIRE pe it s " s R i -3 -, B . . § = IS ] a . OFF GAS . _.° . - .. * D+ A . a ’ . D N - . o - 2 . : :A.n 2 * ) s s o . a A s a ' a’ . .'. 'A 7 s 2 . R o « . - Fs . N :A . - - b LAMINATED ot .o . SHlELDiNG—\.' ' . . 4 : o & T ) ‘ < o . ) €, 4 re . a o a a a . B :- ' . a sl Fa) o " 4 . * - a - . 'y A 3 s ° - aa . - ‘, a‘. ‘\‘: a 4 ' AN 4 )' a : = = o \ ——— - - ry “ a A; . & a a "' i a4 A ' o P - * = N - - . - Fig. 1.1.1. Conceptual Design of a One-Region, Experimental, Molten-Salt Reactor. (not shown). A fuel enriching and sampling system, which is described in a following section, is appended to the pump bowl. There is an insulated gas passage surrounding the core vessel which constitutes one leg of a heating or cooling loop. This gas loop is provided to maintain the fuel above its melting point and below a fixed maximum temperature during all off-design conditions. The remainder of the gas loop consists of a blower, a heater, a cooler, and two dampers, which operate to direct the flow through either the heater or the cooler. The heater is described in a following section. The reactor system is surrounded by a double=walled steel cell. The space between the two walls is monitored for leaks. Space coolers within the cell keep the cell walls below lSOOF. The design of the reactor system provides for semiremote maintenance. Those components which have the highest probability of requiring maintenance or replacement are located so that replacement involves a minimum of remote operations. Power Plant Arrangement. Preliminary layouts of the power plant are presented in Figs. 1l.1.2 and 1.1.3. It may be noted that the reactor is at a higher elevation than the main condenser in the steam plant. This arrangement avoids any possibility of flood water entering the reactor cell. Obviously this design feature will raise the building cost, and it may be moreheconomical to lower the reactor in relation to the condenser and at the same time raise the whole plant to above the level: of the cooling-water source. In this case the price to be paid would be extra pumping costs. A detailed study of a particular site would be required to establish the proper relations. In contrast to most steam power plant designs, the main crane in the turbine-generator room also provides services to many of the other ma jor components of the plant, such as the evaporators, feedwater heaters, and deaerator. A schematic flow diagram of the power plant is presented in Fig. 1.1l.k. UNCLASSIFIED ORNL-LR-DWG 35664 14 - COOLANT PUMP REMOVABLE SHIELDING —.. . SUPERHEATER | H FUEL PUMP | }7“——':" T h ‘l Tflgfl‘) . by BLOWER ~ o | | %z»OFF*GAS ) ;LT HOLDUP TANKS REAGTOR GAS HEATER — {GOOLER BEHIND) ~ TURBINE-GENERATOR L T ] FEEDWATER HEATERS +===— Fig. 1.1.2. Elevation Drawing of Power Plant for 30-Mw, Experimental, Molten-Salt Reactor. FUEL DRAIN TANKS (UNDER) AN M N ; | - | ACCESS TO REACTOR FUEL GAS HEATER, COOLER AND BLOWER —— ——|_ | SAMPLER, ENRICHER - S CELL COOLER - DRAIN TANK COOLER ] B I ofey HEATER. I {UNDER) DRAIN VALVES * (UNDER) -REACTOR CENTER (UNDER) CONTROL ROOM SUPERHEATER - A 20 % Maccess To ACCESS TO EQUlPMETI/// REMOVAL TUNNEL \l “~CHARGOAL TRAPS {UNDER) COOLANT; \»J PUMP - | I/ EVAPORATORS S - d :i DEAERATOR ™ ;)f il . UNCLASSIFIED CRNL-LR-DWG 35665 L«— TURBINE-GENERATOR I Fig. 1.1.3. Plon Drawing of Power Plant for 30-Mw, Experimental, Molten-Salt Reactor. FUEL PUMP I 1210°F REACTOR 96,3001b 1000°F 1450 psi e COOLANT PUMP 125°F L |- PRIMARY HEAT EXCHANGER / "\_Fl / ATTEMPERATOR UNCLASSIFIED ORNL-LR-DWG 35666 - TUBINE-GENERATOR —D SUPERHEATER —# (COOLANT-STEAM) OO _F r - 4 1075°F 3.33cfs 1,460,000 Ib/hr 349,0001b 610°F 1470 psi STEAM PUMP 593°F 228,700 1b EVAPORATOR 97,260 Ib 450°F/ / Y FEEDWATER HEATERS / 1% BLOWDOWN 260 1b FEEDWATER PUMP/ \\ j - y Y [ CONDENSER } 1% MAKE UP 960 Ib / Y ¥ namVAVAVA ] — 1 - / . \CONDENSATE PUMP FEEDWATER HEATER DEARERATOR Fig. 1.1.4, Steam System Flow Diagram for 30-Mw, Experimental, Molten-Salt Reactor. -8- - 9 = Fuel Sampling and Enriching Mechanism. The fuel sampling and enriching mechanism, shown schematically in Figs. 1l.1.5 and 1.1.6, consists of a sampling-pot elevator; a horizontal conveyor; a sample- depositing elevator; an enriching-capsule elevator; four, identical, motorized worm-gear drives; several enriching-capsule magazines; several shielded semple carriers; an enricher-magazine press; a sample-carrier press; a solenoid-operated stop; a canned remotely controlled gate valve; a number of canned remotely controlled ball valves; a supply of sampling capsules; and a supply of enriching capsules. All parts of the system are enclosed in vacuum-tight piping that is insulated and shielded where required. Every operation is remotely controlled and electrically interlocked so that accidental improper sequences of operation cannot damage the equipment. The sampling-pot elevator carries capsules from the horizontal conveyor down into the sampling pot of the reacter, either to obtain samples or to add enriching capsules. It is actuated by a motor-driven bronze worm-gear. Between the horizontal conveyer and the reactor are a ball valve and gate valve. On the down stroke of the elevator plunger, these must be opened in turn as the plunger approaches them. On the up stroke, they must be closed immediately behind it. The capsules are propelled along the horizontal conveyor by a transfer plate suspended from a track. There are breaks in the track where the three elevators cross it. These breaks are filled by the plungers of the elevators when they are in the up position. The transfer plate has a hole in it through which the elevator plungers operate %o obtain or deposit capsules. There is a solenoid-operated stop provided to center the transfer plate hole over the sample-depositing elevator. End-~of -travel stops center the hole over the other two elevators. A bolt-on blind flange is provided for access at the reactor end of the conveyor, and a bolt-on flange matching that of the drive unit is welded to the other end. -10- UNCLASSIFIED ORNL-LR-DWG 35667 MCTORIZED WORM-GEAR DRIVES MCTCRIZED WORM-GEAR DRIVES : SAMPLING-POT ELEVATOR ki ENRIGHING-CAPSULE / ! ELEVATOR—_ | SAMPLE-DEPOSITING ELEVATOR —\ HORIZONTAL CONVEYOR SAMPLING CAPSULES ON GONVEYOR CANNED, REMOTELY CON- TROLLED BALL VALVES SHIELDED SAMPLE N 45 CARRIER S ENRICHER-MAGAZINE PRESS ENRIGHING-CAPSULE X GANNED, REMOTELY CON- | , MAGAZINE TROLLED BALL VALVE SAMPLE-CARRIER PRESS — GANNED, REMOTELY CONTROLLED P GATE VALVE REACTOR CELL TCP SHIELD SAMPLER ACCESS AREA Fig. 1.1.5. Fuel Sampling end Enriching Mechanism, UNCLASSIFIED ORNL-LR-DWG 35668 \/Kgl {T\_/\/\ Nl ’ ?'b I\si - é 7N Nz N7 NS A » N4 \\\ G i @ Z \\ // NN R RN IR Y (g) (b) AN AT IR / A\ SRR Fig. 1.1.6. (a) Sampling-Pot Elevator Showing Plunger Under a Sampling Can. (b) Sample-Depositing Elevator Showing Tongs. - 12 - At the bottom of the sample-depositing elevator plunger is a pair . of remotely controlled tongs that grip and release capsules as desired. In their up position, the tTongs bridge the gap in the horizontal conveyor track. The tongs cannot open until the capsule is within the shielded sample carrier. At the bottom of the elevator, the shielded sample carrier rests upon a hydraulic press which forces it against a seal in the elevator pipe flange to eliminate the necessity of bolting it. The enriching capsule elevator is similar to the sample-depositing elevator. System Preheating Device, Two designs of the heater bank for heating the gas used to preheat the reactor system are being considered. Tubular heaters with an Inconel sheath are included in one design and . flat ceramic heaters with an Inconel enclosure are utilized in the other design. The main features of the units being studied are listed in - Table 1.l.l. The chief advantages of the fiat ceramic heaters are: (1) the vattage per unit surface area is less, (2) any number of heater dimensions can be obtained, (3) the support for the ceramic plate is provided by the enclosure, (4) fins can be added to the enclosure to provide additional heat transfer surface if required, and the (5) vertical air flow may simplify duct work. A diagram of a heater unit with two, flat, ceramic heaters is shown in Fig. 1.1.7. A group of these units would provide a heater bank with a capacity of 6 to 50 kw in a 2-ft by 2-ft duct. Off-Gas System Valves. Frozen-bismuth valves are being considered for use in the off-gas system, because (1) conventional metal-to-metal seated valves are not consigtently tight in the contaimment of fission- - product gases, (2) valves employing elastomer closures are of limited life because radiation destroys the elastomer, and (3) devices such as bellows to replace stem glands tie the life of the valve to that of the fatigue limit of the bellows. The use of frozen bismuth in a trap to create a perfectly tight valve both processwise and environment- Table 1l.l.l. - 13 - Features of Tubular end Flat Ceramic Heaters Variable Tubular Heater Flat Ceramic Rating, w/in.2 Rating per heater, w Voltage, v Dimensions Shape Surface fins Sheath temperature, ° Mounting Air flow Power source Voltage control 4O 750-3500 230 18- to 8h=in. heated length; 0.315 in. OD Hairpin loop Not available for this temperature Up to 1650 Vertical, tie wires or supports may be required which will tend to cause hot spots | Horizontal A group of T.5«kva, 0 to 270-v Power- stats ganged and operated by a common shaft Manual or automatic by an air-operated motor Up to 20 500-3500 230 Thickness, ~ 3/ in.; width, up to 12=in.; length, up to 18 in, Flat plate Enclosure can be made with fins Up to 1650 Vertical support for ceramic plate is provided by the enclosure Vertical, leads at end of heater; horizontal, leads at top of heater Same as for tubular heaters Same -14- UNCLASSIFIED ORNL- LR—DWG 35669 HEATER LEADS METAL CONTAINER HEATERS Fig. 1.1.7. Heater Unit With Two, Flat, Ceramic Heaters. - 15 - wise has been demonstrated.l Valves of this type have been considered - in the past only for limited operation {once open and once closed), but two schemes for repeat action valves, shown in Figs. 1.1.8 and 1.1.9, are now being studied. The valve shown in Fig. 1.1.8 is gravity and temperature operated. In the valve open position the bismuth in the lower trap is in the frozen state., Likewise, in the valve closed position, all the bismuth shown is in the frozen state. Heating zones 2 and 3 are operated together only when the valve is to be opened; the bismuth is melted for a sufficient time to allow complete drainage of the trident-shaped upper trap. Heating zones 1A or 1B are operated in conjunction with heating zone 2 when it is desired to close the valve. In this operation bismuth flows ) from one of the charge containers to the trident section and is subsequently frozen in place. There is enough heat capacity in the lower trap to prevent melting during the closure operation. New charges of bismuth may be placed in either charge container whenever the valve is in the closed position. If the valve is to have a limited application, enough charge containers are built in place to co@er the number of operations expected, and no attempt is made to add charges after the original installation. A gas-pressure and temperature-operated valve is shown in Fig. 1l.1l.9. Although it is simpler in construction than the gravity-operated device, it may-be argued that it is not foolproof. Strict control of the maximum operating gas pressure while the sealant is in the molten state would be highly necessary in this device. lW. H. Kelley, Fuel Sampling Bismuth Valve Test, ORNL CF=57=7=36 (July 24, 1957). -16- UNCLASSIFIED CRNL—LR—DWG 35670 HEATING ZONES 1A AND 18 SEALANT CHARGE CONTAINERS HEATING ZONE 2 GAS IN GAS OUT VENT c = o w =< o i (&) = T GA SSSSSSS GGGGGG e A ; S 5 b _ = AAAAAAAAAAAA - 18 = PRELIMINARY DESIGN OF A 315-Mw (e) GRAPHITE-MODERATED MOLTEN=-SALT POWER REACTOR ' A preliminary studyg’3 was made of a one-region molten-salt power reactor in which the fuel is in contact with the unclad graphite moderator. The molten-salt fuel flows through holes in the graphite, which is contained in a cylindrical INOR-8 vessel. The graphite core is 12.25 ft in diameter and 12.25 ft high, with 3,6-in.-dia holes on 8=in. centers. The fuel, LiF=-BeF,-UF) (70-10=20 mole %), contains low-enrichment (1.30%) uranium, and it occupies 16% of the core volume. A diagram of the reactor system is shown in Fig. 1.1.10. The choice of the power level for this design study was arbitrary, since the core is capable of operation at 1500 Mw (th) without: exceeding safe power densities. An electrical generator of 333-Mw (e) capacity was chosen, which would give a station output of 315 Mw (e). The reactor power rating would thus be 760 Mw (th). A plan view of the reactor plant layout is presented in Fig. 1.l.12; and an elevation view is shown in Fig. 1.1.12. The reactor and the primary heat exchangers are contained in a large rectangular reactor cell, which is sealed to provide double containment for any leakage of fission gases. The rectangular configuration of the plant permits the grouping of similar equipment with a minimum of floor space and piping. The plant includes, in addition to the reactor and heat exchanger systems 2H° G. MacPherscn et al., A Preliminary Study of a Graphite Mcderated Mplten Salt Power Reactor, ORNL CF=59-1-26 {(Jan. 13, 1959). 3¢, E. Guthrie, Fuel Cycle Costs in a Graphite-Mcderated Siightly Enriched Fused Salt Reactor, ORNL CF-=59-1-13 {(Jan. G, 1959). -19- UNCLASSIFIED ORNL~ LR-DWG 35088A SEAL WELD - FUEL PUMP . SALT CONTAINER - — GAS HEATING SHELL GRAPHITE CORE FUEL PASSAGE (TYPICAL) Fig. 1.1.10, Diagram of Reactor System. UNCL ASSIFIED ORNL-LR-DWG 350864 COOLANT SALT PUMP REHEATER COOLANT SALT CELL | / EVAPORATORS PRE-HEATING LOOP ( / EQUIPMENT 1 T ~ REMOVAL HATCH y 4 S \ WINDOWS | |& [):) 2\ - N B FUEL ENRICHERA & ~ | T=—STEAM PUMP fi\; 3 REHEAT LINE CELL COOLINGA - —s= TURBINE - GENERATOR \ 3 HIGH PRESSURE LINE / & kl_ ACCESS LOCK —] L IA: :JJ FUEL PUMPS (4) - { ]T HEAT EXCHANGERS (4} i BD COOLANT SALT LINES N T~ AN N FUEL FILL-AND-DRAIN TANKS O O ) SUPERHEATERS Fig. 1.1.11. Plan View of 760-Mw (th) Graphite-Moderated Molten-Salt Power Reactor Plant, -OZ- UNCLASSIFIED ORNL-LR-OWG 35087A . : COCLANT PUMP R 1§t SUPERHEATER STEAM PUMP J .I.Z- FUEL PUMP —] REACTOR M HEAT EXCHANGER FEET )'/\ O . e 4 1 / : 1 EVAPCRATORS Fig. 1.1.12. Elevation View of 760-Mw (th) Graphite-Moderated Molten-5alt Power Reactor Plant. - 22 - and the electrical generation systems, the control room and the fill-and- drain tanks for the reactor fluids. The plant characteristics are listed below: Fuel Fuel carrier Neutron energy Active core Fuel equivalent diameter Moderator Thermal shield Primary coolant Power Electric (net) Heat Regeneration ratio Clean (initial) Estimated capital costs Refueling cgele at full power Shielding Control Plant efficiency Fuel conditions, pump discharge Steam -system (Loeffler) Temperature Preséure Secondary loop fluid (coolant salt) 1.30% U235Fh (initially) LiF-BeF,-UF ) (70-10-20 mole %) Near thermal 1k £t 12.25-Ft-dia, 12.25-ft-high graphite cylinder with 3.6- in.-dia holes on 8=in. centers 12 in, of iron Fuel solution circulating at 35,470 gpm 315 Mw T60 Mw 0.79 $79,250,000 or $252/kw Semicontinuous Concrete room wall, 9 £t thick Temperature and fuel concentration 41.5% 1225°F at 105 psia 1000°F with 1000°F reheat 2000 psia LiF-BeF, (65-35 mole %) Structural materials Fuel circuit INOR-8 Secondary loop INOR=8 Steam generator Croloy steel (2.5% Cr, 1% Mo) Steam superheater and reheater INOR-8 Temperature coefficient, (Ak/k)/°F Negative Specific power 1770 kw/kg Power density 117 kw/liter Fnel inventory Initial (clean) 700 kg of e Critical mass, clean 178 kg of y°32 Burnup Unlimited The off-gas system provides for the continuous removal of fission- product gases and serves several purposes. The safety of the system in the event of a fuel spill is considerably enhanced if the radicactive gas concentration in the fuel is reduced by stripping the gas as it is formed. Further, the nuclear stability of the reactor under changes of power level is improved by keeping the high-cross-section Xe135 continuously at a low level. Finally, many of the fission-product poisons are, in their decay chains, either noble gases for a period of time or end their decay chains as stable noble gases, and therefore the buildup of poisons is considerably reduced by gas removal. Details of the nuclear calculations pertinent to this design study are presented irn a subsequent section of this chapter. NUCLEAR CALCULATIONS Reactivity Effects in Experimental Molten-Salt Reactor. Various reactivity effects in the experimental reactor described above were calculated. The first calculation was an estimate of the reactivity effect of draining the first layer of diphenyl coolant from the thermal -2l - shield. This draining resulted in a loss in reactivity, Sk/k, of 1.8%%. Another reactivity calculation was performed to estimate the effect of inserting 100 g of U235 in the form of UFh at the center of the reactor. This resulted in a gain in reactivity, ¢5kafi of 0.39%. The Cornpone program was used for these reactivity calculations. A temperature coefficient was also calculated for the experimental reactor by using the procedure described in the Molten-Salt Reactor Program Status Reportoh The over-all temperature coefficient of reactivity was estimated to be =6.75 x 1077 ( Sk/k;/°F, Neutron and Gemma-Ray Activity in the Secondary Heat Exchanger of the Interim Design Reactor. The present specifications for the Interim Design Reactor call for the use of sodium as the heat transfer fiuid in the first intermediate heat exchange loop that links the reactor to the steam system. There are several disadvantages in the use of sodium, among which are its flammability, incompatibility with the fuel salt (uranium is reduced and may precipitate), ard a 15
J @ Z g 700 3 5 g u IS L4 \ = 4 600 2 E oo \ E O \ ON-EXCHANGE PROCESSING, Jp— = 500 —\tffls/day — o o 400 1600 J i -5 / —= Y 3 4200 [ FLUORIDE-VOLATILITY y - o5 PROCESSING, 1.66 f1~/day - > o 4—’ E < 800 i 4 % -+~ TION-EXCHANGE PROCESSING, g = - 3 = % 400 :’/ 18 ft /de Q @ ’/ 0 0 4 8 12 16 20 TIME OF OPERATION (year) Fig. 1.1.13, Long-Term Performonce of Spherical, Homogeneous, Two-Region, Molten- Fluoride-Salt Reactor Fueled With U235. - 29 - mixed in proportions corresponding to those used in the unit cell, The diameter of the equivalent spherical core and the thicknesses of core vessel, blanket, reflector, etec., were specified, as usual. A subprogram was added by means of which the cross sections of the fuel and moderator materials were then multiplied groupwise by the appropriate disadvantage factors. From that point the calculation proceeds in the usual manner. The modified program was tested by computlng the multlplicatlon constant for the ORNL Graphite Reactor. Cross sections for U23 metal were estimated by taking into account resonance saturation and Doppler broadening. Neutron energies ranged down to 0.079 ev, the thermal energy (11800F) in the molten-salt reactors. Cross sections for aluminum were also added. With the use of specifications given by Ramsey and Cagle,7 k was estimated to be 0.967 for the reactor at 1180°F. This value appears to be reasonably correct. Further calculations toc permit éstimation of the multiplication at the operating temperature (3500F) are planned. Cross sections for U238 in a mixture containing 70 mole % LiF, 10 mole % BeFE, and 20 mole % UFh were computed and added to the cross-section library. With the use of the new code and the previously calculated cross sections for various ThFh melts, several uranium and thorium cycle reactors were computed, as described in subsequent sections of this chapter., In order to study the long-term behavior of heterogeneous reactors, the Oracle program Sorghum was modifiéd. Provisien for absorption of neutrons by graphite was made, and group disadvantage factors derived from the modified Cornpone program were applied separately to fuel and graphite. The fission product Smlhg was treated separately from the TM. E. Ramsey and C. D. Cagle, Research Program and Operating Experience on ORNL Reactors, Geneva Conference Paper 450 11958). - 30 - other figsion products. For evaluating plutonium converters, the 240 241 u , Pu 7, existing program was modified to compute ingrowth of P and ° Puahz, in place of Pa233, U233, and U23h. The modified program works well with bare and blanketed reactors, but the critical calculation was found to be unstable in the case of graphite-reflected reactors. In the reactor model embodied in Sorghum, the net leakage in each group is assumed to be proportional to the space-averaged flux in the corresponding group. This is a reasonably good approximation in bare and blanketed reactors in which the sloving down in the second region is small, and the net leakage in any given group is nearly independent of the leakage in groups of higher energy. In graphite-reflected reactors, however, this approximation is not good. Two cases of graphite-moderated reactors have been studied for periods of up to ten years by means of the modified program described in subsequent sections of this chapter. Oracle Program GHIMSR for Gamma-Heating Calculations. A code was written to expedite gamma-heating calculations, in which an integral- spectrum method is used to calculate the heating in a spherical reactor system. Gamma-energy absorption coefficients for 12 groups are stored and used to obtain average absorption coefficients for the reactor and core vessel, The necessary inputs for the calculations include the radius of the reactor, the core and core vessel compositions, and gamma source distributions in space and energy. The time of computation for one source is approximately 3 min. The program was checked against hand computations and was found to be performing satisfactorily. Initial and Long-Term Nuclear Performance Without Fuel Processing of Low-Enriéhmenfi, One-Region, Graphite-Moderated, Unreflected, Molten- Salt Reactor. The modified Cornpone program described above was used to estimate the initial nuclear performance of the graphite-moderated, molten-salt reactor described above. As stated, the reactor has a ..31.. spherical core 14 ft in diameter enclosed in an INOR-8 pressure vessel 1 3/L4 in. thick surrounded by an iron thermal shield 12 in. thick. Fuel channels 3.6 in. in diameter penetrate the core and are arranged in a square array measuring 8 in. between centers. The fuel solution is LiF-BeF,-UF, (70-10-20 mole %), and the calculation indicated that a fuel enrichment of 1.3 mole % is required to make the system critical. The slowing=-down, leakage, and disadvantage parameters obtained from the Cornpone calculation were used in a Sorghum calculation to study the long-term nuclear performance of the system without fuel processing. An extract from the results is given in Table 1l.l.3, and key parameters are graphed in Fig. 1.1l.1lk. The accumulation of fission products progressively poisons the reactor. The regeneration ratio, initially 0.79, falls steadily to 0.5 after 7200 megawatt years (Mwy) of cumulative power generation. The critical inventory falls, at first, because of the ingrowth of plutonium isotopes, but it then rises steadily from an initial value of about 1000 kg to about 4500 kg. The cumulative net burnup amounts to about 1500 kg in 7200 Mwy; the enrichment of the fuel added varies, but averages 60% during the first five years. The cost figures given in ref 2 were calculated on the basis of a one-group nuclear calcuhation3 in which progressive hardening of the neutron spectrum was ignored. Revised cost estimates based on the multigroup calculations described here are expected to show a fuel cost of about 2.5 mills/kwhr. Initial and Long-Term Nuclear Performance Without Fuel Processing of One-Region, Graphite-Moderated, Unreflected Th 232 -Conversion, Molten- Salt Reactor. The nuclear performance of the reactor described in the preceding section fueled with the mixture LiF-BeF,-ThF, (67-16=13 mole %), together with sufficient U235F U238Fu to make the system critical, was studied. The initial critlcal inventory was found to be 829 kg, and the initial regeneration ratio was 0.79. The long-term performance Table 1.1,3. Initial and Long=Term Nuclear Performance Without Fuel Processing of Low=-Enrichment, One-Region, Graphite-Moderated, Unreflected, Molten-Salt Reactor Core diameter: 14 ft Power: 760 Mw (th) Fuel volume: 900 i3 Plant factor: 0.8 After Cumulative Power Initial State Generation of 480 Mwy After Cumulative Power Generation of 2400 Mwy After Cumulative Power Generation of 4800 Mwy After Cumulative Power Generation of 7200 Mwy Neutron Inventory ' Neutron Neutron Neutron Neutron Absorption Inventory . Inventory . Inventory ) Inventory . {kg) .. . Absorption i Absorption . Absorption Absorption Ratio (kg) Ratio* (ko) Ratio* (kg) Ratio* (ko) Ratio* Fissionable isotopes u23s 788 1.000 652 0.578 1,145 0.486 2,275 0.561 3,715 0.630 py239 106 0.409 248 0.426 384 0.346 528 0.290 pu24l 4 0.013 63 0.088 12 0.093 143 0.080 Fertile isotopes y23s 55,150 0.790 54,992 0.672 54,443 0.533 53,870 0.455 53,386 0.403 py240 16 0.056 55 0.140 62 0.122 66 0.102 Fuel carrier Li/ 5,760 0.052 5,760 0.036 5,760 0.016 5,760 0.009 5,760 0.006 g9 37,900 0.026 37,900 0.022 37,900 0.018 37,900 0.017 37,900 0.016 Moderator Be? 1,060 1 0.001 1,060 0.001 1,057 0.001 1,057 0.000 1,057 0.000 cl? 0.051 0.036 0.017 0.011 0.008 Fission products sm14? 0.029 0,010 0.078 0.009 0.192 0.009 0.415 0.00% Other 781 0.026 908 0.085 1,815 0.125 2,723 0.150 Parasitic isotopes U236 27 0.002 109 0.009 228 0.016 379 0.024 Np237 5 0,001 18 0.004 37 0.006 py242 5 0.002 n3 0.041 235 0.071 314 0.078 Core vessel and leakage 0.158 0.133 0.103 0.089 0.080 Neutron yield, 7 2.078 1.996 1,973 1.928 1.882 Total fuel inventory, kg 788 762 1457 277 4386 Cumulative net burnup, kg 58 347 817 1400 Net fuel requirement, kg of U235 788 820 1804 3588 5786 Regeneration ratio 0.790 0.728 0.673 0.577 0.505 *Neutrons absorbed per neutron absorbed by fissionable isotopes. ce _33. UNCLASSIFIED ORNL-LR-DWG 35673 0.8 = 2 o7 - ~N O b o = w 0.6 ™~ G} w 5 \ FUEL SALT: 1_iF~-BeF2—UF4 {(70-10-20 mole %) 0.5 L CORE DIAMETER: 12.25 ft FUEL-TO~-MODERATOR VOLUME TOTAL FUEL VOLUME: 900 ft° TOTAL POWER: 760 Mw {th) PLANT FACTOR: 0.8 4000 // CRITICAL INVENTORY Fd 2000 CUMULATIVE / NET BURNUP 0 2000 4000 6000 8000 TIME OF OPERATION (Mwyr) CRITICAL INVENTORY AND BURNUP (kg of fissionable isotopes) Fig. 1.1.14. Long-Term Nuclear Performance Without Processing of a One«Region, Unreflected, Heterogeneous, Graphite-Moderated, Molten-Fluoride-Salt Reactor Fueled With 1.30% Enriched Uranium, - 3L - of the system was studied by means of the Oracle program Sorghum. An extract from the calculations is given in Table 1.l1.4. For a cumulative power generation of 7200 Mwy, the critical inventory rose to 1416 kg, and the cumulative net burnup amounted to 1083 kg. This performance is substantially better than that exhibited by the corresponding low-enrichment system, which required an inventory of 4386 kg and had a cumulative net burnup of 1400 kg. The disparity between the two systems is expected to increase with time. Initial Nuclear Performance of One-Region, Graphite-Moderated, Graphite-Reflected, Th?32-Conversion, Molten-8alt Reactor. The initial nuclear characteristics of a heterogeneous, graphite-moderated and -reflected, one-region, molten-salt reactor were studied. The reactor considered in the calculations consists of a cylindrical core, 15 ft in diameter and 15 ft high; surrounded by a 2.5-ft-thick graphite reflector contained in a l-in.-thick INOR-8 pressure shell. The core is penetrated by cylindrical fuel passages arranged on an 8-in. triangular lattice paraliel to the core axis. The resulting unit cells are hexagonal and 15 ft long. This system has not been optimized, and a much better design might conceivably be evolved. The initial nuclear characteristics of the system are compared in Tables 1.1.5 and 1.1.6 with the characteristics of the interim design reactor. The modified Oracle program Cornpone was used to calculate k and group disadvantage factors for the fuel and graphite. These results were then used for complete reactor calculations in spherical geometry of a core having the same volume as the actual cylindrical core. Mean "homogeneized" densities (atoms/cm3) were used for each element. It may be seen that the regensration ratio of the heterogeneous reactor is more favorable than that of the interim design reactor. The initial critical inventory of U2359 on the other hand, is 60% higher than in the interim design reactor. It is estimated that a power level of 760 Mw (th) can be attained in the heterogeneous system compared with 600 Mw (th) in the homogeneous reactor. 35 Table 1.1.4. Initial and Long=-Term Nuclear Performance Without Fuel Processing of One-Region, Graophite-Moderated, Unreflected, Thzaz-Conversion, Molten-Salt Reactor Fuel volume: 900 ft°3 Plant factor: 0.8 Core diameter: 14 ft Power: 760 Mw (th) After Cumulative Power Generation of 7200 Mwy After Cumulative Powaer Generatian of 4800 Mwy After Cumulative Power Generation of 2200 Mwy After Cumulative Power Initial nitial State Generation of 480 Mwy Neutron Inventory Absoroti Invent Neutron : , Neutron | Neutron Neutron (kg) . r-p 1:t:m vekn ory Absorption nventory Absorprion nventory AI::SOrpfion lnvenfory Absorpfion atio (kg) Ratio* (ke) Rotio* tkg) Ratie* (kg) Ratio* Fissioncble isctopes y233 129 0.175 424 0.508 541 0.529 593 0.470 y23s 829 1.000 702 0.819 473 0.480 570 0.457 817 0.513 pu239 1 0.006 3 0.012 4 0.014 7 0.017 Fertile isotopes Th432 38,438 0.783 38,438 0.767 38,438 0.700 38,437 0.597 38,437 0.511 Y234 4 0.001 42 0.009 97 0.016 143 0,019 Y238 64 0.008 78 0,008 108 0.013 141 0.016 187 0.020 Fuel carrier Li’ 6,328 0.054 6,328 0.052 6,328 0.045 6,325 0.035 5,328 0.027 - g1? 37,570 0.025 37,570 0.025 37,570 0.024 37,555 0.022 37,570 0.021 Moderator Be? 1,840 0.001 1,840 0.001 1,840 0.001 1,840 0.001 1,840 0.001 cl2 0.051 0.049 0.043 0.034 0.027 Fission products 181 0.034 908 0.154 1,815 0,263 2,723 0.336 Parasitic isotopes y23s 3 0.004 10 0.013 174 0.020 240 0.026 Np237 7 0.004 19 0.008 32 0.012 Miscellaneous Pa233 18 0.009 16 0.008 14 0.006 12 0.004 Core vessel and 0,149 0,148 0.141 0.128 0.116 leakage Neutron yield, 1 2.071 2.099 2.155 2.146 2,120 Total fuel inventory, kg 829 832 900 1,115 1,417 Cumulative net burnup, kg 69 287 638 1,083 Net fuel requirement, kg 829 ¢01 1,187 1,753 2,500 of U235 * Regeneration ratio 0.791 0.768 0.714 0.623 0.546 *Neutrons absorbed per neutron absorbed by fissionable isotopes, Table 1.1.5. - 36 - Initial Nuclear Performance of One-Region, Graphite=-Moderated Graphite-Refiected, Th232-Conversion, Molten-Salt Reactor Compared With That of the Two-Region, Homogeneous Interim Design Reactor 760-Mw (th) Th>3°- Conversion Reactor 600-Mw (ti) Interim Design Reactor Core Blanket Thorium content 13 mole % 1 mole % 13 mole % Core size 15 x 15 ft 8 ft sphere 2=in.=thick cylinder layer Fuel channel diameter 5 in. Fuel volumetric fraction in core 0.3543 All Rlanket Total fuel volume 1480 £t3 607 £t Volume of fuel 3 3 3 in core 9o ft 268 ft 913 ft Volume of fuel 3 3 external to core 540 ft 339 ft Volume of graphite 5340 ft3 Weight of graphite 623,000 1b Thermal fissions 2% 11% Thermal absorptions 5% Neutron yield, YZ 2.073 1.80 Regeneration ratio 0.79 0.63 Table 1l.1l.6. Initial-State Inventory and Neutron Absorptions 760-Mw (th) Tho3° - 600-Mw (th) Interim Design Reactor Conversion Reactor Core Blanket Inventory Neutron Inventory Neutron Inventory Neutron (kg) Absorption (kg) Absorption (kg) Absorption Ratio¥* Ratio¥* Ratio¥ Figsionable isotope ye3? 970 1.000 604 1.000 Fertile isotopes U238 81 0.010 45.3 0.039 THhe 32 41,000 0.782 2,100 0.36h 30, 500 0.228 Fuel carrier Li7 5,250 0.038 3,920 5,030 Fld 4, 000 0.031 24,000 25,100 Moderator Be9 3,890 0,050 3,008 0.102 1, 460 0.011 cte 283,000 Core vessel and leakage 0.162 0.052 ¥* Neutrons absorbed per neutron absorbed by fissionable isotopes. - 38 - l.2. COMPONENT DEVELOFPMENT AND TESTING SALT-LUBRICATED BEARINGS FOR FUEL PUMPS Hydrodynsmic Journal Bearings. The fourth test of a salt-lubricated hydrodynamic bearing was completed and terminated on schedule after 1000 hr of operation. The test was performed with salt No. 130 (LiF-BeFE-UFh, 62-37-1 mole %) under steady-state conditiens: temperature, 1200°F; Journal speed, 1200 rpm; bearing radial. load, 200 1lb. Postrun examination revealed that slight rubbing had occurred between the bearing and journal. Since this test bearing was started and stopped only a very few times as compared with the bearing which was tested previously and started and stopped 87 times (rather than 94 times, as reported erroneously in the previous reportl) during a total operating time of 784 hr, it appears that the rubbing marks may not be attributed entirely to starting and stopping operations. Wear on the journal bemering used in the fourth test appears to be slightly less than on the Jjournal bearing which was sub jected to 87 starts and stops. A fifth test of 360 hr duration, for which new bearing and journal parts fabricated of INOR-8 were used, was also completed. Salt No. 130 was used as the lubricant and the system was maintained at 1200°F. The Journal speed and bearing radial load were varied over wide ranges (600 to 2700 rpm end O to 500 1b, respectively). Thirty combinationg of speed and load were applied for periode of 2 hr each. There were no difficulties observed during these tests. Postrun examination of the bearing revealed that slight rubbing occurred between the bearing and journal. 1P. G. Smith, W, E. Thomas, and H. E. Gilkey, MSR Quar. Prog. Rep. Octo 31’ 1958, ORNL-2626, P 1.90 - 3"1.. A sixth test with the bearing and journal used in the fifth test is under way to investigate bearing performance at steady-state conditions of journal speed and bearing radial load (1200 rpm and 200 1b) and at temperatures varying in 5OOF intervals from 1200 to 15OOOF. Bearing performance is being observed for 4 hr at each temperature level. All these tests have been performed with a radial clearance of 0.005 in. between the bearing and journal surfaces, as measured at room tempera- ture. The continuing analytical studies and surveys of recent literature indicate that it may be possible to improve the performance of this bearing by increasing the clearance to 0.007 in. Hydrodynamic Thrust Bearings. Construction of a thrust-bearing tester is approximately 80% complete. The tester shown schematically in Pig. 1.2.1, utilizes a PK type of centrifugal pump (without impeller) that was modified in the impeller region to permit installation of the test thrust bearing. The tester includes provisions for varying bearing load, journal speed, and test temperature. The thrust runner is mounted on the lower end of the pump shaft and the stationary member of the thrust bearing is supported by a load actuator. The load actuator consists of two concentric bellows with common end flanges. Pressure applied internally to the bellows assembly resultis in an axial force {design maximum, 500 1b at 1200°F) applied to the thrust bearing. The first bearing to be investigated will be of the step type. Test Pump Equipped with a Salt-Lubricated Journal Bearing. A PK type of centrifugal pump is being modified to include a molten-salt- lubricated journal bearing near the impeller, an upper bearing (both radial and thrust) lubricated either with oil or grease, and an overhung drive motor. The molten-salt-lubricated bearing will be tested through the full scope of loadings and clearance problems existing in a pump operating at high temperatures. - 40- UNCLASSIFIED ORNL-LR-DWG 35674 IEMARAERS HIS NN oL INmmmefi | | EQUALIZER [0 SN L PRESSURIZING LINE T !‘I T LOAD ACTUATOR N Wi / T N it THRUST-BEARING STATOR THRUST-BEARING RUNNER Fig. 1.2.1. Diagram of Thrust-Bearing Tester, ")"'l" Hydrostatic Bearings. The previously described2 water tests of hydrostatic bearings were completed and a report describing the tests and giving the results is being prepared.3 The final test, for which the bearing radial clearance was 0.005 in., was made at various speeds with various system flow resistances. Postrun examination revealed that slight rubbing occurred between the bearing and journal. Further work with this type of bearing (and also with rotating-pocket hydrostatic bearings) has been deferred because of the favorable results being obtained with the less complicated hydrodynamic type of journal bearings and the increased emphasis on thruste-bearing testing. Bearing Mountings. Investigations of bearing mounts to accommodate dimensional changes resulting from differences in coefficients of expansion of the bearing and mounting materials have been de-emphasized because of the favorable results being obtained with the use of INOR-8 for Jjournal, sleeve, and mount. Promising techniques have not yet been found.; The "Thermoflex" journal bearing mounts designed by Westinghouse Electric Manufacturing Company are not yet available in the sizes of interest. CONVENTIONAL BEARINGS FOR FUEL PUMPS Tests with Dowtherm "A" as the pump lubricant were completed, as stated previously,h and a final report5 on the tests and results has been issued. 2P. G, Smith, W. E. Thomas, and H. E. Gilkey, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL=2626, p 19. 3P. G. Smith and H. E. Gilkey, Hydrostatic Journal Bearing Water Tests « Modified PK~-A Pump, ORNL CF~59~2-1 (to be issued). hD. L. Gray, W. E. Thomas, MSR Quar. Prog. Rep. Oct, 31, 1958, ORNL'2626, p 20. 5w. E. Thomas, Dowtherm "A" Journal Bearing Test, ORNL CF=58-10=40 (Nov. 6, 1958). - 43 - MECHANICAL SEALS FOR FUEL PUMPS Labyrinth and Split-Purge Arrangement. The previously described results of tests of a PK type of pump with a labyrinth seal and split- purge gas arrangement were correlated and a test report was issued. Bellows«Mounfied Seal., The modified Fulton-Sylphon bellows-mounted seal7 being subjected to an endurance test in a PK~P type of centrifugal pump has accumulated an additional 2520 hr of operation since the previous report period, for a total of 10,630 hr. The pump has been operating at a temperature of 1200°F, a shaft speed of 2500 rpm, and a NaK flow rate of 1200 gpm. The test=seal leakage continued to be negligible throughout the peridd. Three pump stoppages occurred; one that was caused by a power failure and two that were for replacing brushes in the drive motor. PUMP ENDURANCE TESTING An MF type of centrifugal pump has been circulating fuel 30 (NaF'- ZrF) -UF, , 50=46=l mole %) in continuous operation for 13,500 hr since June 26, 1957, without maintenance. For the past 11,500 hr the pump has been operated under cavitation damage conditions. The fuel 30 flow rate is 645 gpm; the shaft speed is 2700 rpm; the pump tank cover gas is at a pressure of 2,5 psig; and the temperature of the circulating salt mixture is 1200°F, A recently pu‘blish.ed8 operational resumé,presents an account of various pump starts and stops and changes in operating conditions. During this quarter the pump was stopped three times. One stop was - Gb L. Gray, Test of PK Pump Split Purge Gas Labyrinth Seal, ORNL CF-59-1-5 (Jan. 17, 1959). 7b. L. Gray, MSR Quar. Prog. Rep., Oct. 31, 1958, ORNL-2626, p 21. 8(}° G. Smith, Resumé’MFnj Pump Test Operation in MFF Loop, ORNL CF-58-11-38 (Nov. 17, 1958). | ...],’_3- momentary because of a power outage. It was stopped the second time for 5 min to replace the tachometer generator on the drive motor. The third stop was of 8 min duration to replace drive-motor brushes. FROZEN-LEAD PUMP SEAL The small frozen-lead pump seal being tested on a 3/16-in.-dia shaft, as described previously,9 has operated continuously since it started operstion on June 13, 1958. As of January 1, 1959 the accumulated operating time was L4L84LO hr. Following slight leakage of lead during the first 100 hr of operation, there has been no furthef leakage. A large frozen-lead seal on a 3 l/h-in.-dia rotating shaft has been placed in operation. The frozen-lead seal retains molten lead at a temperature of approximately lOOOOF in a tank. The test equipment is shown in Figs. 1.2.2 and 1.2.3. The three separate cooling coils used to maintain the lead below its melting point in the bottom of the sealing gland are shown in Fig. 1.2.4. Thermocouples, located between the cooling coils, may also be seen in Fig. 1.2.k. The ghaft is mounted in two self-aligning, pillow=-block bearings attached to the support frame. A 10-hp Louis Allis motor and magnetic clutch are conmected to the bottom end of the shaft through an adjustable coupling. As may be seen in Fig. 1l.2.2, the sealing gland incorporates a long, tapered amnulus, which narrows to provide a clearance between the shaftl and the sealing gland of a few thousandths of an inch. The upper end of the shaft rotates in the lead tank. An inert atmosphere of argon at atmospheric pressure or higher can be provided over the surface of the molten lead in the tank. Either water or air can be circulated through 9. B. McDonald, E. Storto, and J. L. Crowley, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 23. -44 - UNCLASSIFIED ORNL—LR— DWG 35675 Te i LEAD TANK l RN l SNy S \ % Tc9 } 3 N N Tc3 % T 10 — 3Y,-in-DIA SHAFT AR [ ) im fi\\\ § th z:l Te Tc2 TeS Tc8 Tc 4 Te7 Tct7 Tce ‘l = | ) \ UPPER BEARING - COOLING COILS ON SEALING GLAND LOWER BEARING C_— SUPPORT P E T_\l LATE &/ U — ] L COUPLING TO MOTOR Fig. 1.2.2. Sketch of Frozen-Lead-Seal Test Equipment Showing Thermocouple Locations. UNCLASSIFIED PHOTO 32824 =28 L n-Lead-Seal Test Equipment. .45. Fig. 1.2.3. Froze -46- " UNCLASSIFIED | PHOTO 32826 Fig. 1.2.4. Sealing Gland, Shaft, Cooling Coils and Thermocouples for Frozen-Lead Seal Test. ..)_l_T.- the cooling coils attached to the outside of the sealing gland. The flow in each of the three cooling coils can be controlled separately. The supply pressures of watef and alr are maintained at constant levels. In preparation for filling with lead, the equipment was heated to operating temperature. With the shaft rotating at 100 rpm, the tank was filled with 120 1b of lead at 1000°F. At the present time, the shaft is being run at 1300 rpm. The corresponding peripheral speed of the shaft is 1100 fpm, which is about 10 times the peripheral speed of the 3/16-in.- dia shaft operating in the small frozen-lead seal test. Water at a flow rate of approximately 2 gpm is being used as the coolant for the seal gland. Smell amounts of lead have leaked sporadically from the sealing glafld. Representative operating temperatures for the sealing gland and tank at the thermocouple locations indicated in Fig. l.2.2 are listed below: Thermocouple Temperature Thermocouple Temperature Number (OF) Number (°r ) 1 880 T 590 2 875 8 810 3 935 9 925 L 590 10 985 5 815 il 860 6 275 17 290 TECHNJQUES FOR REMOTE MAINTENANCE OF THE REACTOR SYSTEM Mechanical Joint Development. Detailed descriptions of the three types of mechanical joints being considered for use in a remotely maintainabie reactor system were presented previously,lo‘as well as the results of lQA. S. Olson, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-247k, p 20. - L8 - screening tests of the three joints, which were conducted under thermal cycling conditions in a high-temperature loop that circulated fuel 30 (NaFnZ_thaUFh, 50=46-4 mole %) at temperatures up o 1_500° Rl Additional tests were made in a high-temperature loop that circulated sodium at temperatures up to 13OO°F°13 Tests were continued during the quarter on the two, previously described,ll large freeze-flange mechanical joints for use in a L=in, pipe line. The tests were conducted in a high-temperature loop that circulated fuel 30 at temperatures up to 1300°F under thermal cycling conditions. | 13 the freeze=flange clamps Subsequent to the first thermal cycle, were removed, and additional braces were welded onto the clamps to permit greater clamping loads. The clamps were cleaned and given an oxidation- resistant coating with the use of the "Black Magic" process. The coated gsurfaces of the clamps and the bolts were then covered with a special lubricant. Seal rings of annealed 28 aluminum were installed in each Joint, and the seal was effected with a bolt loading of 200 ft=-1b per bolt. The cold leakage rates for these two large freeze-flange jointé before preheating of the loop were 6 x 10='11 cm3 of helium per seEond or less. The leakage rate after preheating of the loop to 13OOOF was 2 X 10“"8 cm3 of helium per second or less. During the loop test, temperatures were measured at the points indicated in Fig. 1l.2.5. The molten salt was circulated at a flow rate of approximately 4O gpm. The 1, S. Olson, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-24Th, p 20. 12W. B, McDonald, E. Storto, A. S. Olson, Screening Tests of Mechanical Pipe Jointg for a Fused Salt Reactor System, ORNL CF=58=-8-33 (Aug. 13, 1950}. 13A. S. Olson, MSR Quar. Prog. Rep. Oct., 31, 1958, ORNL=-2626, p 25, -49. UNCLASSIFIED ORNL—LR—-DWG 3454 (O FLANGE NO.1 THERMOCOUPLES [ | FLANGE NO.2 THERMOCOUPLES DOUBLE LINES INDICATE WELLS Fig. 1.2.5. Diagram of Large Freeze-Flange Mechanical Joint Showing Location of Thermocouples. salt temperature was cycled 12times between 1100 and l300°F; each cycle was of 2 hr duration. Representative temperatures measured during the test are listed in Table 1l.2.1. Table 1.2.1. Temperatures Measured During Thermal Cycling of Freeze~-Flange Mechanical Joints in Development Test No. 2 Thermocouple Minimum Cycle Maximum - Cycle Number Temperature Number Temperature Number (°F) (°F) Freeze«Flange Joint No. 1 1 450 8 540 2=3=1] 2 875 3-4-5-8-11~12 1025 9 3 980 3el=5=11=12 1150 2=23=526=T=3=10 L 980 3elje5-11-12 1145 9 5 880 = 3-4-5-809-11-12 1020 9 6 420 8 495 2-3 7 300 lei=5=829 335 3 8 245 8 290 a3 Freeze~Flange Jolint No. 2 9 510 h=5 585 2 10 900 -3 1055 9 11 995 3 1175 9 12 960 3=5-11-12 1135 9 13 910 3 1080 9 1Y 465 8«3 550 T=10 15 315 5=8 350 2-4=10=11 16 300 4=8 330 1=2«3=b-T=10-11 N There was no indication of salt leakage during the test, but gas leakage tests after the high-temperature salt test yielded leakage rates 3 in excess of the maximum alloweble leakage rate of 10_7 cm” of helium per second. After the clamp bolts on both flanges were tightened to again provide a bolt loading of 200 ft-1b per bolt, the flanges were leak tested a fourth time. The leakage rates were 9.4 x 10-801113 of helium per second or less and were thus again below the specified maximum allowable rate, Both joints were easily disassembled in about 15 min. The frozen salt in each Jjoint was caked to the stainless steel insert screen, as shown in Fig. 1.2.6. One fiange is shown with the screen and salt seal removed in Fié. 1.2.7. The flange faces were sufficiently clean for reassembly of the Joint. Leakage rates were again measured after reassembly of the joints with the same aluminum sealing rings. The leakage rate for joint No.l was in excess of lO-7 cm3 of helium per second. The leakage rate for joint No., 2 was 6 x 10"'8 em> of helium per second. Remote Maintenance Demonstration Facility. Construction of the remote maintenance demonstration facility iz expected to be essentially complete by Juily 1. The General Mills remotely operated manipulator for this facility will be delivered in February. MOLTEN-SALT FEAT-TRANSFER.COEFFICIENT MEASUREMENT The modifications reguired for salt-to-salt heat transfer tests in the test facility for molten-salt heat-transfer-coefficient measurements were completed, and 52 data runs were made. The heat transfer data -52. UNCLASSIFIED PHOTO 45806 Fig. 1.2.6. Disassembled Large Freeze-Flange Joint With Screen and Sealing Ring in Place. 2535 UNCL ASSIFIED PHOTO 45807 Fig. 1.2.7. Disassembled Large Freeze-Flange Joint After Removal of Screen. -5)_|_- obtained are currently being analyzed. A description of this test facility and the results of a salt-to-liquid metal test were presented previously;lh’15 TRIPLEX-TUBING HEAT EXCHANGER DEVELOPMENT A seven-tube test heat exchanger was designed (Fig. 1.2.8) with which to evaluate the triplex heat exchanger tubing being developed by the Metallurgy Division for molten-salt or liquid metal=-to-steam superheater applications. The triplex tubing consists of inner and outer tubes with sintered metal powder or wire mesh in the annulus between them. This tubing was designed to provide positive separation of fluids- witfi minimum thermal resistance between the inner and outer tubes and to provide a means of leak detection in that the pressure of a gas occupying the voids in the sintered metal in the annulus could be monitored. Details of the tube development program are included in Chap. 2.1 of this report. The design of a small heat exchanger test stand has been revised to accommodate this heat exchanger for heat trénsfer and reliability e#aluation tests. EVALUATION OF EXPANSION JOINTS FOR MOLTEN-SALT REACTOR SYSTEMS A total of six commercially available expansion joints has been evaluated in the expansion joint test facility. A description of the th, C. Amos, R. E. MacPherson, and R. L. Senn, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-24Th, p 27. 15J. C. Amos, R. E, MacPherson, and R. L. Senn, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 31. ‘ MOLTEN SALT [P S— (B mam] UNCLASSIFIED ORNL-LR-DWG 35676 SPACERS i . /g-in. RODS —— AN ‘ OUTSIDE TUBE : 1-in. GD, 0.065-in. WALL “__J INSIDE TUBE: -in.OD, 0.065-in. WALL ANNULUS: SINTERED NICKEL POWDER FINAL DRAWN TUBE: 0.875-in. OD TS HEAT EXCHANGER SHELL 3Y5-in. SCHED-40 PIPE -gg- Lo idml moLTeN | g ) SALT _~GAS HEADER i - e —- - - - “HELIUM GAS CONNECTIONS FOR LEAK-DETECTION SYSTEM — - = = — — — 83— — - = = ] ~ ~— NaK HEADER Fig. 1.2.8. Triplex Tubing Heat Exchanger. -56- test facility and results of the first three tests were presented previously.16 The results of the last three tests are presented in Table l.2.2. The points of failure are shown in Fig. 1.2.9. Table 1.2.2. Descriptions of Expansion Joints and Results of Tests Vendor Cook Electric Co. Flexonics Corp. Flexonics Corp. Material Inconel Inconel Stainless steel Test fluid Molten salt Molten salt Sodium Rated maximum traverse 2 1/k in, 2 5/8 in. 2 5/8 in. Test traverse 2 1/4 in. 2 5/8 in. 2 in. Number of cycles to failure 34 29 31 All the units were cycled over their maximum allowable traverse at rated temperature and pressure of-l3OOOF and 75 psig, respectively, except the Flexonics Corp. stainless steel unit, which was cycled over approximately 75% of its rated maximum traverse. This reduction in the cycle did not appreciably increase the life of the unit. The early failure of all the units indicates that this type of expansion joint design would not be satisfactory for molten-salt reactor application at 13OOOF and 75 psig without design modifications and further development testing. The typical failure, shown in Fig. 1.2.10, indicates a high- stress point that could be eliminated by redesign. 163. C. Amos, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL~-2626, p 32. UNCLASSIFIED PHOTO 32923 FAI LUR s -/~ COOK ELECTRIC CO. INCONEL FLEXONICS CORP STAINLESS STEEL FLEXONICS CORP INCONEL Fig. 1.2.9. Bellows Removed from Expansion Joints Tested in a Molten Salt or Sodium at 1300°F Showing Location of Fallure UNCL ASSIFIED PHOTO 32921 Fig. 1.2.10. Cracked Area of Flexonics Corp. Inconel Expansion Joint Bellows That Was Tested in a Molten Salt ot 1300°F. —59- DESIGN, CONSTRUCTION, AND OPERATION OF MATERIALS TESTING LOOPS Forced-Circulation Loops. The operation of forced-circulation corrosion-testing loops was continued. Three new loops were started and one loop was terminated, on schedule, during the quarter. By the end of October, the 15 forced-circulation loop test stands were operating at full capacity. Eleven of the facilities occupying five column bays of Bldg. 9201-3 may be seen in Fig. 1.2.11. Thé remaining four loops are located to the right and across the aisle from the area shown, Two of the three new loops started during the quarter were constructed of INOR-8 and the other was constructed of Inconel. One INOR-8 loop and one Inconel loop (designated MSRP-12 and 9377~5, respectively, in the operations summary, Table 1.2.3) were equipped with a molten-salt sampling device,lT which makes possible the removal of samples during operation. The results of analyses of the 57 samples that have been removed from these two loops are presented in Chap. 2.2 of this report. The loop terminated was the Inconel loop listed as 93hhi~l in Table 1.2,3, which had operated a full year with a temperature difference across the loop. The facility from which this loop was removed was modified according to the new design18 for improved operation before loop 9377~5 was installed. Of the 15 facilities in ofieration at the present time, nine are of the new design and the remaining six are in various stages of improvement. In November the entire facility, with 1k operating loops, survived a mamentary power failure without a freeze-up of the salts. All loops regained flow as soon as power was restored. Those loops which had been Y5, 1. Crowley, A Sampling Device for Molten-Salt Systems, ORNL-2688 (to be issued). 18J. L. Crowley, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 36. Fig. 1.2.11. View of Operating Area for Forced-Circulation Corrosion-Testing Loops. -09- 1 . v v '] * Table 1,2,3. Forced-Circulation Loop Operations Summary as of December 31, 1958 Composition A ] A . Maximum Temperature Hours of Loop L M ol and S; Number of plprox;Qmute Ppmx';:me Wall Difference Operation at o Designation oop Material and Size Circulated Flow Rate Reynolds Temperature Across Loop Conditions Special Features Comments Fluid® (gpm} Number (°F) (°F) Given 9344-1 Inconel, ]/2 in, OD, 123 2 3250 1300 200 8800 Loop terminated Nov, 5, 1958, 0.045 in. wall after one year of operation 9354-3 INOR-8 84 2,75 1200 100 8374 Contains a Hastelloy B Loop fluid dumped temporarily Hot leg,“';/B in, sched 40, 4500 pump on Oct. 25, 1958, when small Cold leg, ]/2 in. OD, 5400 oil fire broke out near pump 0.045 in, wall 9344.2 Inconel, 1/2 in. OD, 0.045 12 2.5 8200 1200 200 8067 Normal operation in, wall 9377-3 Inconel, ‘/2 in, OD, 0.045 131 2 3400 1300 200 6980 Normal operation in. wall 9354-1 INOR-8, ]/2 in, OD, 0.045 126 2.5 2000 1300 200 6503 Normal operation in, wall 9354-5 {NOR-8, 3/8 in. OD, 0,035 130 1 2200 1300 200 5590 Contains box of graphite Normal cperation in, wall rods in hot fluid sire-cu'nh 9354-4 INOR-8 130 2.5 1300 200 3852 Contains three machined Normal operation Hot leg,s/a in. sched 40, 3000 sample inserts in re- Cold leg, ‘/2 in, OD, 3500 sistance-heated 0.045 in, watl section® MSRP-7 INOR-8, ]/2 in. OD, 0.045 133 Not available 1300 190 3418 This loop and all others Loop fluid dumped Oct. 13, in. wall listed below contain 1958, when pump stopped redesigned safety after loss of power to both features® magnetic clutchs 9377-4 Inconel, 1/2 in. OD, 0.045 130 1.75 2600 1300 200 3408 LFB pump rotary element, in. wall changed Dec. 9, 1958, be- cause of bad bearing MSRP-6 INOR-8, ]/2 in. OD, 0.045 134 Not available 1300 200 3028 Normal operation in, wall MSRP.8 INOR-8, 1/2 in, OD, 0.045 124 2.0 4000 1300 200 280¢% Normal operation in. wall L9 Table 1.2.3 {continued) Composition A ) A ) Maximum Temperoture Hours of Loop L M ‘ol and Si Number of :Iprommote pproxu;r;ote Wall Ditference Operotion at ) Designotion oop Materiol and Size Circulated ow Rote ReYn: s Temperature Across Loop Conditions Special Features Comments Fluid® (gpm) Number (°F) (°F) Given MSRP-9 INOR-8, ]/2 in, OD, 0.045 134 Not available 1300 190 2636 Normal operation in., woll MSRP-10 INOR-8, l/2 in. OD, 0,045 135 Not available 1300 200 2469 Normal operation in, wall MSRP-11 INOR-8, ‘/2 in. OD, 0.045 123 2.0 3200 1300 190 2128 Started Oct, 3, 1958; normal in. wall operation MSRP-12 INOR-8, ]/2 in. OD, 0.045 134 Not ovailable 1300 200 1521 Contains o molten-solt Storted Oct. 29, 1958; normal in. wall sampling device operation 9377-5 Inconel, 1/2 in. OD, 0,045 134 Mot available 1300 190 873 Contains a molten-salt Started Nov. 25, 1958; narmal in. wall sampling device operation z9 “Composition 12: NaF-KF-LiF (11.5-42-46.5 mole %) Composition 84: NaF-LiF-Be F2 (27-35-38 mole %) Composition 123: NaF-BeF -UF‘1 {53-46-1 mole %) Composition 124: NaF-BeF —ThF4 (58-35-7 mole %) Composition 126: LiF-BeF --UF‘1 (53-46-1 mole %) Composition 130: LiF-BeF ‘UFA (62-37-1 mole %) Composition 131: LiF-BeF -UF4 (60+36-4 mole %) Composition 133: LiF-BeF -ThF4 (71-16-13 mole %) Composition 134: LiF-BeF 'ThFd'UFd (62-36.5-1-0.5 mole %) Composition 135: NaF-Ber-ThFd-UF4 (53-45.5-1-0.5 mole %} I:’J. L. Crowley, MSR Quar, Prog, Rep. Jan. 31, 1958, ORNL-2474, p 31. 8] NN R €J. L. Crowley, MSR Quar, Prog. Rep. June 30, 1958, ORNL-2551, p 36, Fig. 1.2.16. - 63 - provided with the necessary equipment assumed isothermal operation automatically, and steps were taken to prevent freeze-ups of the remaining semiprotected loops. All loops were returned to test conditions with no adverse effects and with a minimum of lost time. The preventive maintenance program on the corrosion-testing facilities requires that the loops be operated isothermally when it is necessary to repair or replace worn or defective equipment. Such events are listed in Table 1.2.3 under comments only if it was necessary to drain the operating fluid into the system sump or to change the operating fluid when a system was opened. In-Pile Loops. Operation of the first in-pile loop was started in the MTR December 3, 1958. The loop operated satisfactorily at design conditions until Januvary 8, 1959 (nearly two MIR cycles) when a high radiation‘field accompanied by fission-gas release developed in the vicinity of the cubicle containing the loop. Loop operation was terminated. The radiation release has been traced to a partially plugged pump-sump purge-outlet line which caused fission gases to back- diffuse up the pump-sump purge-inlet line to a point outside the cubicle, where they escaped ithrough a leak. The loop had accumuldted a total of 860 hr of operation, of which approximately 700 hr were with the reactor at power. The operation of the second prototype in-pile pump was terminated after 2500 hr of satisfactory operation at the design conditions of the first loop. Examination showed the pump to be in the as-installed condition, except for an accumulation of salt in the intermediate region between the pump sump and bearing=housing region. The accumulation had been obgerved by X-ray examinmation during operation of the pump, but it was considered to be of insufficient quantity to cause difficulty. The third prototype in-pile pump, which is still being tested, has accumnulated over 2000 hr of satisfagtory operation. This pump is identical to that installed in the second in-pile loop and incorporates means to - 64 . eliminate salt accumulation in the intermediate region. X~ray examinations of the operating pump have verified the effectiveness of the modification. The assembly of the second in~pile loop is approximately 60% complete. This loop will incorporate modifications to the purge system which should prevent a recurrence of the difficulty encountered with the first loop. - 65 . 1.3. ENGINEERING RESEARCH PHYSICAL PROPERTY MEASUREMENTS Viscogity. Viscometers of the gravity-flow capillary-efflux type; which have been employed almost routinely in determinations of the viscosities of molten-salt mixturesyl have recently given unexpectedly irregular results2 for beryllium-containing fluoride mixtures. Efforts to increase the experimental precision of the data have led to modifications both in the efflux-cup design and in the cup calibration technique. The design of the modified efflux-cup viscometer is illustrated in Fig. 1.3.1. It differs from previous designs in two significant respects: (1) the length of the cup was increased to provide either 120 or 240% more fluid volume, and (2) a serrated "skirt" was added to the lower outside portion of the cup. The first of these changes allowed longer flow times and diminished the effect of initial surface "humping," and the gecond change assured that drainage along the outside surface did not interfere with the discharge from the capillary tube. The new cups were calibrated with a NaNOE-NaN03-KNO (LO=T7=53 wt %) mixture of known v1scosity.3 In the temperature range 150 to 400° C, this salt has a kinematic viscosity that is in the same range as the kinematic viscosities of the fluoride salts at 500 %o 800°¢C. Comparison of the NaNO.-NaNO_-KNO_ calibration with room-temperature 2 3 3 calibrations with glycerine-water solutions of similar kinematic viscogities lS, I. Cohen and T. N. Jones, Viscosity Measurements on Molten Fluoride Mixtures, ORNL-2278, (June 28, 1957}. W. D. Powers, MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 38. 3W. E. Kirst, W. M. Nagle, and J. B. Castner, Trans. Am. Inst. Chem. Engrs. 36, 371 (1940). 266 . UNCLASSIFIED ORNL-LR-DWG 35677 CUP, 1.0 in.0D, 0.881 in. ID % OR 2% in. LONG. T i M, i [ ppates WEEP HOLE — | "SKIRT" EIGHT EQUALLY CAPILLARY TUBE, 0.039 OR SPACED POINTS 0.046 in. ID, 1.0 in. LONG Fig. 1.3.1. Modifled Efflux Cup Viscometer. - 67 - indicated a significant temperature dependence. The room-temperature calibration leads to high and inconsistent results for the fluoride salts. New viscosity values were obtained for the mixtures LiF-BeFE-UFh (62-37-1 mole %) and LiF-BeFe--UFrThFh (62-36.5-0.5-1 mole %} with the modified and recalibrated efflux cups, and comparison of the new values with the data obtained previously confirmed the calibration effects discugsed above. Over the temperature range of 500 to SOOOC, the new viscosity values for the two mixtures may be closely correlated by the equation Ho=h eB/T , where‘/I is the viscosity in centipoise, T is the temperature in OK, and A and B are experimentally determined constants. The values obtained for A and B for LiF-BeF,-UF) (62-37-1 mole %) are A = 0.0580, B = L4550, and for LiF-BeF,-UF) -ThF) (62-36.5-0.5-1 mole %) are A = 0.0680, B = L49T. The new viscosity values for the mixture LiF-BeF,-UF) (62-37-1 mole %) are presented in Fig. 1.3.2, along with the values obtained previously for comparison. Enthalpy and Heat Capacity. Measurements of the enthalpies in the solid state of three mixtures of lithium chloride and potassium chloride were completed. The results are given in Table 1l.3.1, along with the previously reported values for these mixtures in the liquid phase. MOLTEN-SALT HEAT TRANSFER STUDIES The "continuous-operation" apparatus designed for studying surface film formation in molten-salt systems by heat transfer coefficient measurements is indicated schematically in Fig. 1.3.3. The mixing chambers nave been redesigned so that the thermocouples can be placed along the centerline of a chamber to obtain better fluid mixed-mean temperatures. H, VISCOSITY (centipoise) N AN ' ’ \ o) | j — A | Q C, 20 \\0 ; \ 41 =0.0580 e%5%/7 ‘\A\/o/- @] . o NOTE: IN THE VISCOSITY EQUATION, a L 7 1S IN °K 10 | o & . \"; Py N (] e .J l \ . %e 6 —— SYMBOL TYPE OF CUP CALIBRATION MEDIUM —| & o A \tr o @ ® e A WITH SKIRT NaNO, - NaNOz = KNO5 D . 2N A WITH SKIRT NoNO, - NaNOz ~KNO, - - P WITH SKIRT NGNO,~NaNO5 ~KNO5 o o WITHOUT SKIRT GLYCERINE - WATER SOLUTION o~ ° WITHOUT SKIRT 'GLYCERINE - WATER SOLUTION _ 2 R - 1 400 450 500 550 600 650 700 750 800 -68- UNCLASSIFIED ORNL-LR-DWG 35678 TEMPERATURE (°C) Fig. 1.3.2. Viscosity of Molten Salt Mixture LiF-BeF,-UF , (62-37-1 mole %). B50O Table 1.3.1. Enthalpy Equation Coefficients and Heats of Fusion of Several LiCl=-KC1l Mixtures Salt Mixture Phase Temperature Enthalpy Equation Coefficients* Heat of Fusion (cal/g) Range (°C) a b c Temperature H - Hg - (c) x 1077 LiCl-KC1l Solid 100-350 -6.8 +0.212 +6.35 L43 T0 (70-30 molie %) Liquid 450-800 +15.3 +0.361 -3.12 LiCl-KCl Solid 100-350 -7.1 +0.204 +6.26 370 61. (60-40 mole %) Liquid 400-800 +21.2 +0.315 -0,18 LiCl-KC1 Solid 100-350 -5.9 +0.192 +5.40 450 6l (50=50 mole %) Liquid 500800 +38.5 +0.239 +1 .62 * The enthalpy is given as HT - H3O°C = a + bT+—cT2; the heat capacity is then evaluated as cp = b + 2cT. -69- .70- UNCLASSIFIED ORNL-LR-DWG 35679 FLOWMETER ELECTRODES : \ HEAT SINK TEST SECTIONS MIXING CHAMBER (LOCATED IN PIPE PUMP AT EACH END OF TEST SECTION) Fig. 1.3.3. System for Heat Transfer Coefficient Measurements of BeF,-Containing Molten Salts. - 71 = Further, the thermally indeterminate regions between the test-section electrodes and the mixing chambers have been eliminated by incdrporating the electrodes as the end plates of the mixing chambers, The components for this system are being fabricated and assembled. | Initial calibration of the turbine-type flowmeter to be used in the apparatus has been completed at room temperature with water as the éalibrating medium. No deterioration of the calibration with time was observed. The essential components of the apparatus are being mocked up with the use of Tygon-coupled sections in order to determine both the pump speed vs flow rate characterigtics and the over-all system pressure Arop. Part 2. Materials Studies - 75 « 2.1. METALLUGRY DYNAMIC CORROSION STUDIES Thermal-Convection Loop Tests. Examinations of two Inconel thermal- convection loops that operated with fused fluoride mixtures for one year at 13500F vere completed during the quarter. One of the loops (1181) which circulated a blanket salt of the composition LiF-ThFh (71-29 mole %), showed attack in the form of subsurface voids to a depth of 6.5 mils along hot-leg sections. No attack or deposits were observed in cold-leg sections. The other loop (1184) which circulated the fuel mixture NaF--Zth-UFh (55.3=40.7=4 mole %), was similarly attacked to a depth of 7 mils and showed no attack or deposits in cold-leg sections. A third Inconel loop (1223) was examined after completing 1000 hr of operation at 1350°F with the fuel mixture LiF-BeF,-UF . Attack in thig loop occurred as light intergranular void formation to a maximum depth of 1 mil. The fuel mixture circulated in this loop was obtained from the same batch as a mixture circulated in loop 1222, described previously.l However, prior to operatiocn in loop 1223, an analysis of the fuel was made to detect sulfur contamination. In the course of this analysis, the fuel, while molten, was purged extensively with HF and hydrogen gas. No measurable sulfur contamination was found, and, in general, the impurity concentrations appeared to be unaffected by this treatment. However, the as-received fuel caused heavy attack to a depth of 3 mils in loop 1222 under conditions otherwise identical to those of loop 1223. An effect of the purging treatment on the corrosiveness of 1 P 53. J. H. DeVan et al., MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, - 76 - the fuel is therefore evident, but the mechanism through which this effect occurred has not been identified. Operation of an Inconel thermal-convection loop with the salt mixture NaF-LiF-KF (1l.5-46.5=42 mole %) was terminated after 4673 hr in order to provide additional test space for salts of currently greater program interest. The maximum operating temperature of the loop was 1250°F. Examination of this loop (121k4) revealed extensive attack along hot-leg sections to a depth of 13 mils, but the cold-leg sections showed no evidence of mass transfer. Chemical analysis of the salt after the test showed a significant increase in the chromium content compared with that found by analysis prior to the test. The occurrence of such gross attack in the absence of observable mass transfer strongly ifidicated that the fuel became contaminated either prior to or subsequent to introduction into the loop. The high chromium concentration was unexpected because there was no UFh in the salt. A subsequent check of the salt by petrographic and x-ray analyses revealed the presence of approximately 15% KF°2H20. The presence of this constituent suggests that water, a highly oxidizing impurity in tpg fluoride systems, was introduced into the salt at some point befafe or during the test and was responsible for the extensive attack. Forced-Circulation Loop Tests. The status of the forced-circulation loops now in operation with fluoride salts is presented in Chap, 1.2; no tests were completed during the quarter. GENERAL CORROSION STUDIES Carburization of Inconel and INOR-8 in Systems Containing Scdium and Graphite. A system containing sodium and graphite, which forms an —77- effective carburizing medium for nickel-base alloys,2 was used to carburize Inconel and INOR-8 by exposure for 4OOO hr at 1200°F. The test procedures and apparatus were described previouslyo3 Room- temperature mechanical tests show that the tensile strength and elongation of the carburized INOR-8 decreased (~6.5% and 50%, respectively) in comparison with the tensile strength and elongation of the control specimens. The control specimens were subjected to an argon atmosphere for the same time and at the same temperature that the carburized specimens were exposed to the sodium-graphite system. The tensile and yield strengths of the carburized Inconel specimens increased (~24% and 9%, respectively), while the elongation decreased, in comparison with the same properties of Inconel control specimens. Carburization was found metallographically to a depth of about 7 mils on both metals. Chemical analyses of the carburized specimens of both TInconel and INOR-8 showed decreasging carbon conteht with depth as indicated in Table 2.l.1l. The values obtained from the chemical analyses were used to prepare the curvesh shown in Fig. 2.1.1. From these curves, a carbon content vs distance relationship for certain times was determined, as indicated in Fige. 2.1.2, 2E. E. Hofffian, W. H. Cook, and D, H. Jansen, MSR Quar. Prog, Rep, Jano 311 1958, 0RNL"’2’-|'7,+) D 5""0 3E. E. Hoffman, W. H. Cook, and D. H. Jansen, MSR Quar. Prog. Rep. June 30, 1958, ORNL=-2551, p 59. o “w. H. Cook and D. H. Jansen, A Preliminary Summary of Studies of INOR-8, Inconel, Graphite, and Fluoride Systems for the MSRP for the Period from May 1, 1958 to Dec. 31, 1958, ORNL CF=59-l=h (to be published).. - 78 - Table 2.,1.1. Carbon Content of Millings Taken from Inconel and %NOR-B After Exposure to Sodium-Graphite System for LOOO hr at 1200 F Carbon Content (wt %) Material . At 0 =3 mils At 3-6 mils At 6~9 mils At 9=12 mils Inconel 0.21 0.12 0,055 0.030 INOR=8 0.27 0.10 0.048 0.026 * Surface of specimen. A second carburization test was completed for which the test conditions were identical to those described above, except that the duration of the exposure was LOO hr, rather than L4000 hr. No carburization .of the specimens could be detected metallographically after this test. Mechanical property tests showed no reductions in elongation of the Inconel or INOR-8 specimens. Carburization of Inconel and INOR-8 in Systems Containing Fuel 130 and Graphite. Two tests at l300°F in which Inconel'and INOR-8 were exposed to fuel 130 (LiF-BeFE-UFh, 62-37~1 mole %) in systems containing graphite were terminated after 2000 and LOOO hr, respectively. The Inconel specimens were attaéked to depths of 3 and 7 mils, respectively, in the 2000- and 400O~hr tests, and there were reductions in mechanical strength as a result of the attack,h The INOR-8 specimens showed no carburization in either case. Surface roughening, which was found only on the specimens from the L400O-hr test, was to a depth of less than 0.5 mil. No significant changes in mechanical properties were foufid for the INOR-8 specimens from either the 2000- or LOOO=hr tests in argon or in fuel 130 exposed to graphite. The mechanical test results gave indications that INOR~8 was not carburized under the conditions of this test.h -79- UNCLASSIFIED ORNL-LR-DWG 35449R 4.0 T [ I 1 ‘ ALLOY O (cm?/sec) C, (wi%) | INCONEL 7X10™%2 0.26 os | | _INOR-8 3.5%40° "% 0.40 _ i ‘ . i | | | | by — 7 AN G o6 a ‘7’0 9 - - — b NS g‘ :_Z‘ —i + X 7, ~ o} B 2 AN - 3] R < CIESAY? o V e (g AS -~ O ® m (@] O¢ A q_ o4 2\ A RN\S 2, e oloz\5 ¢ - 5y, Pl O\Z @N\'p '9(‘\ i “‘l AR AN 2\e P\ LN\& \ o\~ | 0.2 PO — NS N ; -\é.. Tl«,_ \ \ ¢ 10 20 30 40 50 DISTANCE FROM INNER SURFACE {mils} C=WEIGHT %, CARBON AT POINT WITHIN TUBE WALL C;.=WE|GHT %> CARBON AT SOLID—-LIQUID INTERFACE (INIT1AL TUBE SURFACE) Co= INITIAL WEIGHT 7. CARBON IN ALLOY O =DIFFUSION CONSTANT Fig. 2.1.1. Approximate Carbon Distribution in Walls of 50-mil Inconel and INOR-8 Tubing Exposed for Various Times to Liquid Sodium and Graphite at 1300°F on Inner Surface and Argon on Outer Surface. UNCLASSIFIED ORNL-LR-DWG 25120R 040 | W | ‘ | | 030 } F | T— 88 3 . Z 020 —— - - }— — & oy S ey@Or i ; 1 J’eors <9 ‘ €arg . | | 0 10 20 30 40 50 DISTANCE FROM INNER SURFACE (mils) Fig. 2.1.2. Amount of Carbon (wt %) at Points Within a 50-mil Thick INOR-8 Tube Wall After Yarious Exposure Times Surface and Argon on Outer Surface. at 1300°F to Liquid Sodium-Graphite on Inner - 8 - Uranium Precipitation From Fused Fluoride Salts in Contact with Graphite, Tests were continued in the study of the compatibility of p, graphite and molten-salt fuel mixtures. As described previously, the test systems are radiographed periodically to check for UO2 precipitation from the fuel mixture and penetration of the fuel into the graphite. Uranium oxide precipitation has been found in all teste in which the fuel mixture 130 has been used in contact with graphite at 1300 h in these tests the ratio of the calculated projected surface area of the graphite to the volume of the fuel was 3.6:1. The fraction of the uranium that precipitates from fuel 130 appears to be finite and eomplete within a 5-hr exposure to graphite at 13000F under a vacuum of less than Oal/u. The basis for the conclusion that the precipitation is finite is that no detectable changes in the initial quantities of precipitate have'been seen radiographically in any of the test systems, even after éxposures as long as 3000 hr. The quantity of uranium precipitated appears to be an indirect function of the purity of the graphite; that is, the greatest quantities of precipitated uranium are found in the test systems containing the most impure graphite. The impurities in the graphite are traces of' metals, and since the adsorptive qualities of graphite are increased and degassing is made more difficult by metallic contaminants, it is believed that the oxygen source for the uranium precipit;tion was adsorbed to a greater extent on the less pure graphiteo The graphites used in the test systems in which uranium precipitation has been found were degassed at 2350 F for 5 hr under a vacuum of less than h/l, flooded with argon, and then exposed to air for 3 min while they . H. Cook, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 61. -8l - were being weighed. Apparently the brief exposure to air was sufficient for the graphite to pick up enough contamination to cause the uranium precipitation. In contrast to the results being obtained for fuel 130, no segregation or precipitation has been found radiographically when fuel 30 (NeF-2rF)-UF), 50-L6-4 mole %) is in contact with graphite at 1300°F under a vacuum of less than 0.14 for 3000 hr or longer. In the tests under way,.steps are being taken to determine definitely whethef graphite is directly or indirectly responsible for uraniup precipitation from fuel 130 under these,con@itions. Since, on the basis of the preliminary.results,,graphite does appear to be responsible, attempts are being made to determine whether the precipiiation can be eliminated by using purer graphite and/or by purging the graphite with gas or liquid prior to exposure to fuel 130. ' Fused Floride Salt Penetration Into Graphite. Fuel 30 penetration into graphite pore spaces has been erratic in tests of 1000 hr or less at 13OOOF under a vacuum of €0.1 Pl The gi'aphite specimens have 2, b The penetration is probably shown no penetratibn to heavy penetration. due to contamination of the graphite; reaction products are being sought to verify this supposition. The effect of pressure on the penetration of- fuel 30 into graphite pore spaces at 13OOOF is being investigated. The pressure being used in a test now in progress is 150 psia. In examinations just started, weight-change measurements of TSF-grade graphite specimens that had been exposed for 2000 and 4000 hr to fuel 130 at 13OOOF under a vacuum of approximately O.;}( have indicated that fuel 130 did not penetrate the graphite during the 2000-hr exposure. A weight gein resulting from the LOOO-hr exposure is being investigated to determine whether there was penetration or whether reaction products were formed. In two tests of 100-hr duration at 1300°F, fuel 30 under a.pressfire of 150 psia has not penetrated evacuated (<112/4) graphite.h - 82 - MECHANICAL PROPERTIES OF "INOR-8 Fabricability and corrosion resistance to fused salts are perhaps the most important requiremehts for a structural material for the molten-gsalt reactor, but the high-temperature'strength of the material is also important from a design standpoint. The INOR-8 alloy presently being considered as the structural material for the molten-salt reactor system was developed principally to possess corrosion resistance; it | is presently being studied from the standpoint of its strength properties. Tests are under way in which tensile strength, creep, relaxation, and fatigue effects are being evaluated. This tésting program has two objectives: first, to obtain reliable data suitable for use in design engineering, and,‘second, To explore many of the factors, other than stress and temperature, fihich may affect the strength properties. Tensile and creep data obtained in air were presqnted previously6’7 which indicated that the strength of INOR-8 approaches thatrof the strongest stainless steels. Tests to determine the creep strength of INOR-8 in molten salts are under fiay. Times to 1% strain as a function of temperature and stress in short-time tests were presented previously,7 and long-time tests are in progress at 1100 and 1200°F . Specimens stressed at 25,000 psi at 1100°F have mnot yet reached 1% creep strain. even after 6000 hr. The results obtained by ORNL in molten salts were compared with results obtained by the Haynes Stellite Company in air in the previous report,7 but differences in the heat treatment of the specimens prevented any conclusions from being drawn with respect to the relative effects 6b A. DOuglas, MSR Quar. Prog. Rep. June 30, 1958, ORNL=-2551, p 6k4. 7]E{ W. Swindeman and D. A. Douglas, MSR Quar. Prog. Rep. Oct. 31, 958 ORNL—2626, p 64. - 8% - of the environments. Subsequently, tests in air were.initiated at ORNL on material identical to that tested in molten salts. The data thus obtained in air are compared with the moltenesalt data in Table 2.1l.2. Table 2.1.2. Comparison of Creep Data for INOR-8 Sheet Specimens Tested in Fuel Salt 107 and in Air Temperature Stress FEnvironment Time (hr) to Specified Strain Time to (°F) (psi) 0.5% 1.0% 2.0% 5.0% Rupture _ (hr) 1800 3,000 Salt 4.5 17 48 115 130 Air 8 17 35 75 1700 5,000 Salt 4.5 12.5 31 110 250 Air 5.4 12.5 25 58 90 1500 8, 000 Salt 19 52 125 350 886 Air 29 49 86 175 260 1250 20,000 “Salt* 110 265 680 1600 2400 Air 180 320 720 1500 2180 - . Interpolated data. On the basis of a 1% creep strain, the data indicate that environment does nqt‘appear to seriously affect the creep properties. At 1500, 1700, and lBOOoF, however, the times to greater strains and to rupture are much longer in molten salts than in air. At 12500F and a stress of 20,000 psi, the molten salt and air data agree fairly we%l at all strain levels. A correlation of data from tests on sheet and rod specimens in air is presented in Table 2.l.3. Rod specimens appear to be the strongér'afi 20,000 psi, and the sheet specimens are slightly stronger at 10;000 psi. - 8k - Table 2.1.3. Comparison of Creep Data for INOR-8 Sheet and Rod Specimens Tested in Air at 1250°F Stress Specimen Time (hr) to Specified Strain Time to (psi) 0.1% 0.2% 0.5% 1.06 2.0% 5.0% Rupture - (hr) 20,000 Sheet 30 80 190 350 730 1700 1786 Sheet 30 6L 180 320 720 1500 2177 Rod 20 125 300 510 920 1800 3535 Rod 15 100 300 510 920 2200 15,000 Sheet 190 350 750 1400 Sheet 120 330 780 1500 Rod 130 300 40 1500 12,000 Sheet 600 931 2275 Rod 320 TLO 2082 10,000 Sheet 1650 2600 4270 Rod 760 1600 3596 Tensile data available for INOR-8 show that the ductility decreases rapidly with increasing temperature between 1000 and lSOOOFo For example, elongations ranging from 15 to 30% have been found at lhOOOF; whereas, at 1000°F, the ductility may be as high as 60%. Since low ductilities are sometimes associated with notch senéitivity, several tensile tests were conducted in order to establish the notched-tosunnotched strength ratio. A ratio greater than unity is considered to indicate notch strengthening, while a ratio less than unity signifies notch sensitivity. Results from these tests are presented in Table 2.1.4. On the basis of - 85 - Table 2.1.4. Effect of Notches on the Tensile Strength and Ductility of INOR-8 Tensile Reduction Notch Temperature Geometry Strength in Area Strength (psi) (%) Ratio Room Unnotched 116,000 48 1.08 Notched 125,000 20 1500°F Unnotched 148, 500 15 1.37 Notched 66, 500 8 (0.0045 =4 LYY Byt O Xy SNIEERSy V- L TR ¥ AR PRA . ™ I P { ) «\‘ e 1 [ coe/ iy . 2 = N\ " 5 oF . N 003 ‘. 3 P SN . .+ P £nd 010 e R 2 > - L= o« o 2 fa ) LA Tt Y o P \‘ g fore | - ~ LB o3 S " o AR oL foue - J‘\-’ Lol & \ . - LILE ' R T o Los| ¢ - » - SO 4 . R D At 3 Jd P ~af.- < - P or g ¥ ' e . - - . abin® - . L 3 = yu e & "““\ b &'n\':' \ S - - Fig. 2.1.5. Carbide Spheroldization Occurring In INOR-8 Weld Metal Aged for 500 hr at 1500°F. 200X. Etchant: copper regia. 1000X Fig. 2.1.6. Carbide Precipitation Along Grain Boundaries of As-Welded INOR-8 Weld Metal. 1000X. Etchant: copper regia. Table 2.1.6. Effect of Melting Practice on Mechanical Properties of INOR- 8 Weld Metal in AsAWelded Condition Tensile Strength Elongation Heat No. - Welding Practice (psi) (% in 1 in. eage) ’ At Room At At At Room At At Temperature 1200°F 1500°F Temperature 1200°F 1500°F MP-1 Inconel weld wire used to obtain 113,000 77,000 59,000 Ll 30 15 50% of the nickel, 100% of the chromium, and 90% of the Fe; 1% Cb wag also obtained from the weld-wire addition MP -2 Master alloy obtained from 108, 000 69,000 57,000 43 22 15 Electro Metallurgical Company used to add 2% Al + Ti MP-3 0.5% Mn, 0.10% S8i, 0.2% Al, 0.2% 107,000 67,000 48,000 46 25 13 Ti, 0.025% Mg, and 0.005% B added as recommended by International Nickel Company MP-4 Same as MP-3 except boron omitted Tests in progress MP-5 100% of the nickel added as nickel weld wire Tests in progress "g6." - o4 - As may be seen in Table 2.1.6, all the methods described above for deoxidizing and purifying the ingot have shown promising results. The ductility at 15000F has been increasged from an average value of 7% to 13 to lB%. Other modificafions in melting practice will be invesgtigated in an effort to obtain a ductility at 1500°F of 20%. In order to determine the influence of carbon content on the duc- tility, vacuum-induction melts containing 0.00+%, 0.03%, and 0.06% carbon were made, fabricated into weld wire, and deposited as weld metal. The results of the all-weld-metal tensile tests of these speci- mens are summarized in Table 2.1.7. The data show that no increase in Table 2.1.7. Influence of Carbon Content on As-Welded Mechanical Properties of INOR-8 Weld Metal Carbon Tengile Strength Elongation Heat Content (psi) (% in 1 in. gage) No. (%) At Room At At At Room At At 0 0 Temperature 1200°F 1500°F Temperature 1200°F 1500 F 0.06 117,000 72,000 53,000 37 17 7 MP-7 0.003 112,000 70,000 52,000 35 16 7 MP=6 0.00+ 76,000 47,000 40,000 9 9 7 high-temperature ductility was attained by decreasing the carbon content of the INOR-8 weld deposit. Also, when no carbon addition to the ingot was made, the ductility values at room temperature and at 1200°F were seriously diminished. These poor ductilities are assumed to result from a preponderance of grain-boundary films (probably oxides). These films were of such a magnitude that grain-boundary fissures were numer- ous. The properties of lower-carbon-content melts should also be investigated when melted with improved practices of the type discussed in the preceding paragraphs. Welding of Dissimilar Metals. The problem of selecting & suitable filler material for welding dissimilar metals has often been an obstacle in the fabrication of various components. A dissimilar metal weld is composed not only of filler metal, but substantial quantities of both of the base materials being joined. The filler metal should thus have the following properties: (1) high tolerance for dilution by elements such as iron, nickel, chromium, and copper without forming brittle or crack-sensitive alloys, (2) a moderate coefficient of thermal expansion, (3) ability to withstand high service temperatures over long periods of time without suffering from harmful effects such as sigma-phase formation and carbon migration, (4) good oxidation and corrosion resistance, and (5) good gstrength and ductility over the temperature ranges of interest. The International Nickel Company has developed such a filler wire for inert-arc welding of dissimilar metals, and the mechanical properties have been studied and reported.9 The high titanium content of the wire makes it highly susceptible to age hardening at the molten-salt reactor service temperature of lQOOOF, and its application may, therefore, be limited. A niobium-containing coated electrode, designated Inco Weld "A" electrode, was also developed for the metallic-arc welding of dissimilar metals, and the mechanical properties were determined in both the as- welded and as-aged conditions. The results of this study are shown in Table 2,1.8. As would be expected from the composition,; the metallic- arc weld deposits are not subject-to appreciable aging. The tensile strength at room and elevated temperatures is not appreciably altered, and the ductility is not decreased. It thus appears that this electrode might be useful for Jjoining dissimilar metals for high-temperature 9%. M. Slaughter, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p T0. - 96 - Table 2.1.8. Mechanical Properties of Inco-Weld "A" Electrode in As-Welded and As-Aged Conditions Test Temperature Tensile Strength Elongation (°F) (psi) (9 in 1 in. gage) Room 93, LoO L1 1200 68, 500 28 1300 61,700 19 1500 37,000 32 Room (after aging 500 hr at 1200°F) 97, 400 38 1200 (after aging 500 hr at 1200°F) 65,000 W2 Room (after aging 500 hr at 1500°F) 92,600 40 1500 {after aging 500 hr at “1500°F) 32, 300 49 service where the metallic-arc process is permissible. Pump Component Fabrication. A metallographic study of numerous molybdenum~to-Inconel brazed Jjoints has been conducted in connection with the fabrication of a pump-shaft extension,9 Molybdenum sheets 1/8 in. thick and of varying sizes were brazed to 1/2-in.-thick Inconel plates with Coast Metals No. 52 alloy, copper, and an 82 wt % Au-18 wt % Ni alloy. Examinations of the samples were conducted visually and metallographically after brazing and after thermally cycling 15 times from 1200°F. It was found that extreme stresses are built up in the molyhbdenum- to~-Inconel brazed Jjoints from differential thermal expansion (molybdenum, 2.9 x 1070 in./in./°F; Inconel, 8.7 x 1070 in./in./°F). The stresses may cause severe cracking alpng the center of the brazed joint, along ...97.. the brazing alloy-molybdenum interface, or in the molybdenum base material. As would be expected, the cracking tendencies are decreased as the area of contact decreases. Joints brazed with copper exhibited no cracks after brazing, but: numerous transverse and longitudinal cracks in the Jjoint were noted on all specimens which were cycled 15 times from 1200°F td room temper- ature. It appears that the copper lacks adequate strength and ductility at the elevated temperatures to accommodate the stresses bullt up in thermal cycling. The stronger,lless ductile alloy, Au-Ni, was characterized by longitudinal cracks along the Jjoint. In many cases, the molybdenum base metal near the joint was also cracked. Joints brazed with Coast Metals No. 52 alloy, a strong, brittle material, showed varied properties. In most samples, only slight cracking was observed after thermal cycling, while severe cracking was noted in one sample. The fillets were subject to cracking in most cases. Since Coast Metals No. 52 alloy appeared to be the most promising of the three brazing alloys tested for this application, a prototype pump shaft and extension were constructed. The conditions of the test unit will be limited, however, since some of the molybdenum fingers were cracked during machining. 2.2. CHEMISTRY AND RADIATION DAMAGE PHASE EQUILIBRIUM STUDIES Systems Containing UF) and/or ThF). Phase studies are being conducted to deFermine whetper the N’aF-BeFE-ThFh-UFh system has any advantages over the LiF-BeFe-ThFh-UFh system as a fuel for a graphite-moderated one-region reactor. Investigations of the phase relationships in the system NaF- BeFE-ThF,4 have shown that they are grossly similar to those derived for the system NaF-BeFE---UFh by the Mound Laboratory,l The choice of NaFaBngm ThFh mixtures for use in nuclear reactor systems is limited to compositions having liquidus temperatures of 550°C or lower, and thus the ThFh concens tration is restricted to the range 10 to 15 mole %. All observations of the effect of UFh substitution for ThFh in solid solutions (LiFeThFhm UFEQLiF-Ber-ThFh-UFh, NaF-ThFhJUFh) have indicated that UF)4 substi- tution causes the liquidus temperature of the resultant solid solution to be lowered. It would appear, therefore, that the liquid temperatures encountered in the system Na.F-BeFE--ThFl+ represent maximum temperatures to be encountered for the liquidus as UFh is substituted for part of the ThF, in NaF-BeF -ThFh mixtures. L 2 The System LiF—PuF3. Chemical analyses of filtrates from saturated solutions of PuF3 in LiF—BeF2 and N‘aF-BeF2 atures and compositions have shown that Put o melts at a variety of temper- 3 is sufficiently soluble in such melts to form a fuel mixture for a high-temperature plutonium- burning reactor.2’3 A sensitive apparatus capable of detecting thermal 1C. J. Barton et al., MSR Quar. Prog. Rep. Oct. 31, 1957, ORNL-2431, p 26, Fig. 2,22, 2Co J. Barton, W. R, Grimes, and R. A. Strehlow, Solubility and Stability of PuF3 in Fused Alkali Fluoride-Beryllium Fluoride Mixtures, ORNL-2530 (June 11, 1958). 3C. J. Barton and R. A. Strehlow, MSR Quar. Prog. Rep. Oct, 31, 1958, ORNL-2626, p 80. ..99.. effects on cooling curves from l-g samples has been developed to study phase relationships in these and similar systems. Thermal analyses of LiF-PuF3 mixtures containing 5 to 38 mole % PuF3 show a single eutectic at 19.5 mole % PuF., which melts at 743 + 2°C. Examinations of slowly- cooled mixtures by means of a polarizing microscope mounted in a glove box aided in the location of the eutectic composition and showed no evidence of compounds or solid solutions in this system. The System KF-LiF-BeFe. The possibility of using low-melting point mixtures in the system KF-LiF-BeF2 ag coolants for nuclear reactors was studied. S8Such mixtures would be attractive because of the lower gamma activity they would have in comparison with similar NaF-base mixtures after irradiation in a reactor. Examinations of 12 melts from thermal analysis experiments by the use of optical and x-ray techniques showed that the system KF-LiF- BeF, contains two termary compounds which have the probable formulas 2 KF+LiF*BeF. and KF.LiF-2BeF.. Phases present in the cooled melts indicate tiat the system isedivided into the following assignable compatibility triangles: KF-LiF-3KF*BeF,, 3KF*BeF,-LiF-2KF-BeF,, 2KF'BeF2-LiF-KF'LiF‘BeFE, KF-LiF-BeFecLiF-aLiF-BeFe, and KF-LiF-BeFQ— 2LiF-BeF2~KF-LiFe2BeF2v The first two of the triangles will each contain a eutectic. Neither of these eutectics will have a melting temperature lower than h7500u Mixtures displaying liquidus temperatures as low as 450°C will probably contain as little as 37 mole % BeF,. Mixtures having liquidus temperatures of 100°C or lower will probably not be found that will contain less than 40 mole % BeF,. Compatibility of Fuel with Chloride Coolants. Some consideration has been given to use of a mixture of LiCl and RbCl (58.3 mole % LiCl) as an unreactive coolant for a molten fluoride reactor. Some of the consequences of a leak between the fuel (69.5 mole % LiF, 29;5 mole % BeF,, 1 mole % UFh) and this coolant mixture have been examined 2, experimentally. - 100 - The vapor pressure at 900 to 988°C of an equimolar mixture of fuel and coolant is intermediate between the vapor pressures expected for the ummixed fluids. The material distilled from the mixture consists of fluorides and chlorides of Li, Rb, and Be and about 0.2 at. % U, It is obvious that no compounds of low boiling point result. Mixtures of coolant and fuel contained so little uranium that examinations of quenched specimens did not demonstrate whether a uranium compound was the primary phase. Quenching of mixtures of this coolant with an LiF-BeFe-UFh (67-28-5 mole %) fuel mixture containing more uranium showed that Li2Bth is the primary phase from pure fuel to below 50 mole % fuel in the mixture. The liquidus temperatures dropped from 440°C for the pure fuel to below 360°C for a mixture with 25 mole % fuel. ‘It is obvious that no uranium compound would deposit as the primary phase at reactor temperatures. FISSION-PRODUCT BEHAVIOR Effect of UFh on Solubility of CeF3 in LiF-BeF2 Solvents. The solubility of CeF3 was determined in LiF-BeF2 (62-38 mole %) containing no UFh' Two additions of UFh were then made to this solvent to give mixtures containing 1.2 and 1.9 mole % UFh’ respectively, and the solubility of CeF, was determined in each mixture at several tempera- 3 tures. No effect of UFh on the solubility of CelF., was observed. 3 Removal of Traces of SmF. by the Addition of CeF 3 3 An experiment was completed to determine whether the method used to f i - F - 20 bl 6. hid 5 from LiF-BeF,, UFh (62.8-36.4 0.8 mole %) would be satisfactory for the removal of trace amounts. remove relatively large amounts of SmF It was also desired to determine whether the concentration of SmF 3 remaining in the liquid could be calculated a priori from a knowledge of the individual solubilities of SmF3 and of CeF3 and a material - 101 - balance of the system. The method of calculation was based on relation- ships described ];;reviously.)4 The results are summarized in Table 2.2.1. Table 2.2.1. Removal of Traces of SmF3 from LiF-BeF -UFh 2 (62.8-36.4-0.8 mole %) by the Addition of CeF3 SmF3 added: 578 ppm as Sm SmF,_ in Filtrate (as ppm Sm) CeF3 Added Temperature 3 (vt %) (Oq) Calculated Observed 2.1 487 342 394 10.1 736 480 480 10.1 587 218 288 10.1 492 95 190 The tabulated results show that the method is effective in removing traces of SmF., from this liquid. Agreement between the 3 calculated and observed concentrations is considered good in view of the extensive extrapolations involved when dealing with trace quantities. Chemical Reactions of Oxides With Fluorides. The behavior of oxides in molten fluorides is being studied as part of an effort to explore chemical reactions which can be adapted to the reprocessing hW. T. Ward, R. A. Strehlow, W, R. Grimes, and G. M. Watson, Solubility Relations Among Some Fission Product Fluorides in NaF-ZrF)- UFy (50-56-5 mole %), ORNL-2L2L (Jan. 15, 19587. - 102 - of molten-fluoride-salt reactor fuels. Attempts to precipitate U0, from solutions of UF) (6.5 wt % U) in LiF-BeF,, (63-37 mole %) by treatment with excess BeO have shown that the UFh content of the melt decreases rapidly at first but that the precipitation rate becomes very slow after about 30 min; less than 40% of the UFh precipitates in 3 hr. It is likely that the BeO becomes coated with U02 and that the reaction is effectively prevented. The surface area of the BeO should, accordingly, prove to be an important variable. Sparging such an incompletely reacted system with HF in an attempt to redissolve the precipitated UO, appeared to accelerate 2 the precipitation of U0 In a typical experiment the UFh content of the melt was reducedgfrom 6.5% to 4.4% U by 3 hr of contact with excess BeO at 600°C and was further reduced to 2.9% U by 3 hr of sparging with HF at this temperature. A possible explanation for this behavior is that the HF is capable of exposing additional surface area of the BeO pellets previously coated with insoluble UO2 and thereby accelerating the reaction. An alternate and equally likely explanation is that HF serves as an intermediate reactant which reacts with BeO to produce H20 and bring about precipitation of UO2° CHEMISTRY OF THE CORROSION PROCESS Activity Coefficients of CrF2 in NaF~Zth° The activity coef- ficlents of the fluorides of structural metals such as chromium, nickel, and iron dissolved in dilute solutions of molten fluorides are of interest to the understanding and prediction of corrosion reactions taking place in systems in which molten fluorides are in contact with alloys of these metals. The activity coefficients of FeEe and NiF2 dissolved in molten NaF-Zth (53-L47 mole %) were determined - 103 - > and reported previously.” The activity coefficients of CrF, in the seme solvent were determined at 850°C (ref 6) and at 750°C %ref ), and the activity coefficients determined at 800°C are reported here. The study of the activity coefficienfis of CrF2 in this solvent, as originally planned, was concluded with these measurements. The average equilibrium quotients and activity coefficients for CrF, in NaF-ZrF) (53-47 mole %) at 850, 800, and 750°C are summarized in Table 2.2.2. The equilibrium constants listed for reaction 2 in Table 2.2.2 represent simply the arithmetic averages of the experimentally determined equilibrium guotients. Since no significant effect was noted #ith changes of CrF2 concentration, thg arithmetic aversges may be consldered as equilibrium constants obtain€d by extrapolation to ififinite dilution. Chromium;Diffusion in Alloys. For the ultimate understanding of the corrosion mechanism and prediction of corrosion rates occurring in systems containing molten fluoride solutions in contact with chromium~bearing alloys, it is necessary to determine the various factors that affect the diffusion rates of chromium in the alloys. As a first step of a systematic study of the fundamental corrosion processes, technidues have been developed aqd describeds to determine self-diffusion coefficients of chromium in chromium-nickel alloys with Cr51 As was reported previously,8 the magnitudes of the diffusion ag a radiotracer. coefficients obtained from experiments conducted with specimens which had been polished but not hydrogen fired were one to two orders 5C° M. Blood, W. R. Grimes, and G. M. Watson, Activity Coefficients of Ferrous Fluoride and of Nickel Fluoride in Molten Sodium Fluoride— Zirconium Fluoride Solutions, paper No. 75 Division of Physical and Inorganic Chemistry, 132nd Meeting of the American Chemical Society, New York, Sept. 8=12, 1957. _6C° M. Blood, MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-24Thk, p 105.. 7. M. Blood, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 96. 8R° B. Evans, R. J. Sheil, and W. M. Johnson, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 99. ' - 104 - Table 2.2.2. Equilibrium Constants for the Reduction of CrF2 by Hydrogen Gas and Activity Coefficients of . - CrF, in NaF=Zth (53=47 mole %) Reactions:® (1) CrF, (1) + H, (g)=cCr (s) + 20F (g) (2) CrF, (d) + H, (8)=>Cr (s) + 28F (g) Heat of fusion of CrFE:b 5500 cal/mole Melting point of Cr'F2:b 1375°K . Activity Equilibrium Constant, Kfi Coefficients Temperature - m (°c) Reaction 1 Reaction 2 y (1) y (d) 850 8.0 x 1073 (1.45 + 0.17) x 1073 0,18 1.0 ' 800 2.8 x 1073 (6.77 + 1.12) x 1074 0.2l 1.0 750 8.6 x 107" (2.21 + 0.52) x 1074 0.26 1.0 ) a'Su.bscrip‘t.s s, & 1, and d refer to solid, gaseous, supercooled liquid and dissolved states, respectively. bL. Brewer et al., Natl. Nuclear Energy Ser. Div, IV (1950). Ccalculated equilibrium constants obtained from thermodynamic properties listed by Brewer et al. (footnote b). of magnitude higher than those obtained with specimens which had been polished and then hydrogen fired at 1100 to 1200°C for 2 hr. Photo- micrographs of the fired and unfired specimens showed that considerabie grain=growth had occurred in the fired specimens. The differences in the alloy structures of the annealed and unannealed specimens may be seen in Fig. 2.2.1. . From the experiments during this quarter with specimens annealed in helium, diffusion coefficients were obtained that were very similar to those obtained with hydrogen-fired specimens. The numerical Fig. 2.2.1. 2 hr ot 1100 to -105- UNCLASSIFIED T-16052 UNCLASSIFIED T-15972 e ‘L ois oA R g g R _,’ x 778 (1Led \ |/ |r@ Y X < Py Effect of Annealing on Grain Size of Inconel. (a) Annealed for 1200°C. (b) Unannealed. magnitudes of the over-all diffusion coefficients obtained by contacting the specimens with NaF-ZrF (53-L47 mole %) at 675°C were 1.3 x 10 =15 -15 cm /sec for the cm /sec for the annealed spe01men and 63 x 10 unannesled specimen. The results of these experiments have shown that the grain structure has a marked influence on the self-=diffusion rates of chromium in Inconel. Grain size appears to be a controlling factor. Hydrogen firing and annealing in helium had the samé effects on the over=all 4iffusion rate. Sampling of Operating Loops. Two forced=circulation loops for corrosion testing of materials in circulating fluoride salts were equipped during the quarter with devices which provide for frequent sampling of the circulating salt during loop operation. One loop equipped with a sampling device was fabricated from INOR-8 and the other from Inconel. Both loops were charged with a salt composed of LJ'_F‘=-BeF2-ThFh=UFh (62-36,5-1-G.5 mole %). Salt samples are being teken from these loops, while they are operating, with gradually decreasing frequency. The samples taken from the INOR-8 loop during the first 1000 hr of operation showed a slow but steady increase of chromium in the salt from about 420 to 530 ppm. For the Inconel loop, the increase in the chromium content of the salt was somewhat more rapid, with the chromium content in the salt having increased from about 350 to about L50 ppm in 500 hr of operation. Operation of these loops is continuing. In neither loop has the chromium content of the salt reached a steady-state value. Effect of Frel Composition on Corrosion Equilibria., Equilie- bration studies of the irdividual reactions involved in the corrosion of structural metals by molten fluorides have been continued. Most of the recent experiments have been designed to refine previous results by obtaining better material balances and to corroborate pre- - JCT .- viously described9 equilibrium behavior. Attention has also been given to some instances of anomalous results. The repeated experiments have shown the anomalous results to be fallacious, but the puzzle of their origin has not been com- pletely solved. There are preliminary indications of an effect of the pretreatment of pure Fe on the extent of the reaction O_A 2UFh + Fe ir——.Fee + 2UF3 in typical fuel mixtures, which might mean that thermodynamic equilibrium concentrations for this reaction are not yet definitely known. VAPOR PRESSURES OF MOLTEN SALTS BeF2 Mixtures. An apparatus for obtaining vapor pressures by weighing the salt saturating a known volume of transpired inert gas is being used for measurements of the CsF-BeF, system. The objective of the experiments is to determine the thermodynamic activities as a function of composition; a combination of total vapor pressure and transpiration results is necessary in order to ascertain the molecular species in the vapor. The LiF«-BeF2 systen, because of complex association in the vapor phase, is not amenable to a determination of activities from vapor pressures, but the results for the CsF-BeF2 good estimates for LiF--BeF2 mixtures. UFh’ The vapor pressure of liquid UFh was measured between b and 180 mu Hg (1030 to 1300°C), and the following relationship system are expected to provide a basis for 9J. D, Redman, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL=2626, P 9. -..108. - was obtained: 202.44 £ 0,15 82,280 + 220 = 19 2.303 R ’ 2.303 RT R log p (mm Hg) = - log T + where T is in %K. At the normal boiling point, 1729°K, the heat of vaporization is L49.4 kcal/mole. A re-evaluation of available literature data for the sublimation pressure of the solid led to the re}ationship 16,500 X 450 T log p (m Hg) = - + 13,294 + 0,420 and to a heat of fusion estimate of approximately 20 kcal/moleo ALUMINUM CHLORIDE VAPOR AS A HEAT TRANSFER MFEDIUM AND TURBINE WORKING FLUID Estimation of Thermodynamic Properties. Gaseous aluminum chlor;de under a pressure of 1 atm consists almost entirely of A12016 at h50 ¥, whereas at 1200 OF the fraction dissociated is O. 7. The resulting increases in specific heat and thermal conductivity make this gas attractive as a working fluid for a turbine. In the pre- vious report,10 results of calculations of dissociation equilibria, effective specific heat, thermal conductivity, and viscosity of the gas at various temperatures and pressures were presented. These calculations were extended considerably during the quarter and are being incorporated in a topical report.ll The results of a study of lQM Blander and R. F. Newton, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 103. ;M Blander et al., Aluminum Chloride as a Thermodynamic Working Fluid and Heat Transfer Medium, ORNL-2677 (to be issued). - JUJ - the equilibria of possible mechanisms for the corrosion of metals by gaseous AlCl3 are also presented in the reporto11 The velocity of sound (co) in the gaseous medium, which would be an important parame- ter in turbine design, was calculated from the expression12 2 . /X o ""’/(ap)s ? where v is the specific volume of the gas in cm3/g and p is the pressure of the gas. The term (3v/5 p)S was evaluated at constant entropy (S) from standard thermodynamic relationships and the equation of state for an ideal associating gas. The calculations indicate that aluminum chloride is probably an attractive gaseous heat transfer medium and that it would require very low pumping power. As a working fluid for a turbine, the rela= tively large difference Qf the specific volume of the gas at the high and low temperatures of the cycle, as compared with a working fluid that does not associate, minimizes the effect of inefficiencies of the pump and turbine system. Corrosion of Nickel by.AlCl3° the compatibility of aluminum chloride and structural metals were Experimental investigations of continued. Aluminum chloride gas was held at a pressure of about 1 atm for 850 hr in a capsule of metallic nickel whose ends were maintained at temperatures of 670 and hSOOC, respectively. A very small metallic deposit, which was identified spectrographically as nickel, was found on the capsule wall in the 450°C zone. It is likely that this deposit was formed by the temperature-sensitive reversible reaction 12J. O. Hirschfelder, C. F. Curtiss, R. B. Bird, Molecular Theory of Gases and Liquids, Wiley (1954). -- 110 - AlCl3 + Ni ==~ AlCl + N:LC]..2 9 which was described previously.13 The quantity of material deposited could account for very little attack in the hot zone. Photomicrographs of the capsule showed general and intergranular attack in the 67000 zone, with the maximum penetration of the intergranular attack being to a depth of about 1 mil. No subsurface voids were observed. Further examination of the capsule after completion of the experi=- ment indicated that a cold spbt was present on the capsule near the 67000 zone during the run. It is possible that condensation of Nicl2 on this cold spot could have resulted in increased corrosion of the wall. PERMEABILITY OF GRAFHITE BY MOLTEN FLUORIDE MIXTURES Previous reports in this serieslu have described tests in which reactor-grade graphite was impregnated by exposing out-gassed specimens to molten LiF-MgF, or NaF-LiF-KF mixtures at 1700°F for several hours under inert gas at a pressure of about 20 psig. In order to ascertain whether such treatment protects the graphite from penetration by reactor fuel mixtures, 0.5~ and l.0-in.-dia specimens impregnated with the salt mixture Lin-MgF2 (67.5-32.5 mole %; mp, 1350°F) were exposed at 1250°F for 800 hr to a typical fuel mixture, LiF-BeF,~UF) (62-37-1 mole %). - No evidence of gains or losses in weight of the graphite specimens resulted from exposure to the fuel mixture. However, chemical analysés df successive 1/32-in. layers machined from the specimens indicate that penetration of the fuel mixture had occurred. The beryllifim concen= 13R. E. Moore, MSR Quar. Prog. Rep. Oct. 31, 1058, ORNL-2626, p 102. 1L G. J. Nessle and J. E. Eorgan, MSR Quar. Prog. Rep. June 30, 1958, ORNL=-2551, p 99; MSR Quar. Prog. Rep. Uct. 3L, 1950, URNE-QEEE, P fU?. -.111 - tration of the successive layers decreased from 5000 ppm in the outer layer tq 2000 ppm in the center of the rod; the uranium concentration similarly decreased from 3500 to 1000 ppm. These figures indicate that nearly two=-thirds of the LiFuMgF2 in the outer layer of the specimen was replaced by the fuel. The ratio of beryllium to uranium was sub=- stantially lower in the center of the graphite than in the-ffiel mixtures, but it is evident that impregnating graphite with LiFnMgFg will not satisfactorily prevent the subsequent penetration by LiF-sBeF2-UFh mixtures. EFFECTS OF THERMAL CYCLING ON SALT STABILITY A geries of tests was run with sealed nickel capsules to determine the effects on beryllium-containing salt mixtures of freezing and remelting, without agitation. The mixtures LiF-BeF,-UF) (62-37=1 mole %), LiF-BeFQ-ThFh (71-16-13 mole %), and LiF-BeF,-ThF) -UF, (62+36.5=1~0.5 mole %) were tested. In one set of experiments, the salts were held at‘12509F'under gtatic conditions for 25, 100, and 500 hr. In another set, the salts were thermally cycled five times from 400 to 850, 1000, 1200, and 1400°F, respectively. The maximum temperature of the cycle was held for 30 min before cooling was started. All tests were run in duplicate. Analyses of sections from the salts held at 1250°F for varying periods of time showed no detectable change in composition. The salts which were subjected to thermal cycling showed, however, definite increases in uranium and/or thorium content in the bottom sections of the capsules, with a corresponding decrease in lithium and beryllium content. The degree of concentration of‘UFh or ThFLl in the bottom sections was greater at the lower maximum cycle temperatures. The samples taken to the higher temperatures were apparently mixed by convection. It is evident that when beryilium-based fuels are handled care must be taken to ensure compiete melting of the batch before any portion - 110 of it is transferred to another container or test rig. Thermal cycling of these salts under static conditions must be avoided. RADIATION DAMAGE STUDIES INOR-8 Thermal-Convection Loop for Operation in the LITR. The ine pile thermal-convection loop comstructed for testing fused-salt fuels in an INOR-8 System in the LITR was operated in preliminary out=of-pile tests, and satisfactory circulation was obtained. The loop was filled with solid fuel, which was melted in Place., Removal of gas bubbles from the fuel was found to be a problem but was accomplished by pressure changes brought about by alternately refrigerating and warming the charcoal-adsorber bed of the loop. During radiographic examinations of the loop to determine the extent of gas removal, a dense deposit was noted at the bottom of the lower bend, which is presumably U0, . "The amount of UO2 observed would generate an intoierable amount of heat at that point if the loop were operated in-pile. Examination of the batch of fuel from which this loop filling was obtained revealed UO2 deposits on the surfaces of some pieces. Analyses indicate that the fuel circulating in the loop is free of U02° A second in-pile loop iz being assembled that will be filled with selected pieces of Uogafree fuel. In-Pile Static Corrosion Test. Two INOR-8 capsules that had been filled with LiF-BeF,-UF, (62-37-1 mole %) and shipped to the MIR for 2 irradiation were found upon arrival to have suffered mechanical damage to the thermocouples that prevented their insertion., These capsules are being repaired. PREPARATION OF PURIFIED MATERIALS Fluorides of Chromium. Chromous fluoride of good purity results -.113 - 15 from the reaction of molten stannous fluoride with chromium metal, The container material needed for the reaction is "graphitite," an expensive, low-porosity grade of graphite; stannous fluoride melts leak through containers of ordinary graphite. If cadmium fluoride or bismuth trifluoride is substituted for the stannous fluoride in the reaction, comparable yields (over 90%) of chromous fluoride are obtained in crucibles of ordinary graphite. Molten cupric fluoride and lead fluoride also react with chromium metal to form chromous fluoride, but the yields are lower, probably as a result of loss of material by leak- age through graphite. Cadmium metal volatilizes from the charge when CdF2 is used as the oxidant. When the other oxidants are used, the salt and metal separate into well-defined layers. Batches of less than 4O g of chromous fluoride are conveniently prepared by maintaining chromic fluoride at 1100°C in a platinum or graphite container for 4 hr. The yield of chromous fluoride is con- sistent with the equation 3CfiF3———> 2Cr.F2 + CrF5 ° Attempts to prepare large batches of CrF2 by the disproportionation of CrF3 at 1100°C for a limited time, however, yield a product whose composition is intermediate between CrF2 and CrF3° This intermediate fluoride has also resulted from (1) partial reduction of CrF3 by E, and (2) partial oxidation of CrF, by HF. It can apparently be prepared in nearly pure form by treatment of CrF2 with Sn‘F2 or by reaction of Cr, CrClQ, or CrCl3 with an excess of Sano The y=llowish green crystals of this compound are biaxial negative and show a distinctive X-ray pattern by which the compound can be identified readily. Chemical analysis shows the F/Cr atom ratio to be 2.4. The compound may plausibly 5. J. Sturm, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626;, p 9. - 11h-- be represented as chromous hexafluochromate, that is, CrII3(CrIIIF6)20 The reaction, presumably, is 5CrF +-SnF2——€> Cr3(CfiF6)2 + Sn 2 The oxidation does not procezd further, even if a 400% excess of SnF, is provided. Synthesis of Simple Fluorides by Reactions with Stannous Fluoride. Stannous fluoride is also useful for synthesizing fluorides of other metals in a low valence state. Molten stannous fluoride was reacted with manganese, zinc, aluminum, iron, vanadium, and uranium to yield, s ZnF,, AlF.,, FeF,, VF_, and Uic The equation for regspectively, MnF 09 3 59 3 a typical reaction is Mn + San———>-MnF2 + Sn . The manganous and zinc fluorides prepared by this procedure were pure, since the molten salt and the tin resulting in the reaction formed well-defined layers which were easily separated. The reaction of the uranium metal with stannous fluoride was so vigorous that the molten product was spattered about. Experimental-Scale Purification Operations. Small-scale proceszing of various molten fluoride salts for use in corrosion tests, physical property studies, and small=gcale component testing continued in the experimental facilities located in Building 9201-3 and Building 9928. A total of 265 kg of mixtures without beryllium and 73 kg of beryllium- containing compositions was procesged during the guarter, Trangfer and Service Operations. Service cperations were about b, normal during the guarter, with approximately 98 kg of liquid metals and 1123 kg of fluoride salt mixtures being dispensed. Approximately - 112 operations wers involved in the normal procegses of filling, draining, and sampling various test equipment, - 115 - 2.3. FUEL PROCESSING A process for the recovery of the fuel and blanket salts of a molten- salt reactor is being developed that is based on the volatilization of uranium as UF6 and on the appreciable solubility of LiFnBeF2 salts in nearly anhydrous HF. Previously reported work indicated this procedure to be generally applicable to molten-salt reactor systems employing LiF-BeFe-UFh or LiF--BeFe-UFh-ThFh salts. The process is particularly promising in the degree of separation of the recovered LiF-=BeF2 salt from important rare-earth neutron poisons as a result of the general insolubility of polyvalent-element fluorides in anhydrous HF. A tentative process flowsheet was presented previously.l Studies of the behavior of neptunium, uranium, thorium, and corrosion=product fluorides have been made in further development work on the HF salt-diszolution process. The results appear to be consistent with the proposed process flowsheet in that the recovered LiF--BeF2 salt would effectively be decontaminated from neptunium. Much further investi- gation of some of the variables involved will be necessary, however, before specific flowsheets for specific reactor designs can be prepared. SOLUBILITY OF NEPTUNIUM IN AQUEQUS HF SOLUTIONS The rate of neptunium buildup in a U235 burner reactor increases with time as a result of buildup of the parent U236 isotope. After 20 years, the neptunium concentration will be approximately 0.015 mole % if there is continuous removal on a once-per=year cycle. For such a reactor, the neptunium would contribute about 40% as much poisoning as lM R. Bennett, D. O, Campbell, and G. I. Cathers, MSR Quar. Prog. Rep, Oct. 31, 1958, ORNL-2626, p 111, Fig. 2.4.1. - 116 contributed by all fission fragments at that 'bime.2 The solubility of neptunium in 80 to 100% HF solvents saturated with LiF and BeF2 was found experimentally to be sufficiently low to permit its removal along with the rare earths. In the experiments, small quantities of neptunium (agueous nitrate solution) were added incrementally to HF splutions saturated with LiF and BeFe. The concen- tration of nitrate added was approximately 0.02% of the fluoride salt concentration. The observed solubilities of neptunium are reported in Table 2.3.1 as milligrams of neptunium per gram of solution and as Table 2.3.1 Solubility of Neptunium (IV) in LiF-BeF Saturated Aqueous HF e Approximate Solubility in Solvent Np Solubility HF Concentration - (mg/g of solution) Relative to Salt In Solvent (wt %) TiF Béféfii Np Solubility (mole %) 80 29 111 0,026 0.0031 90 64 70 0.011 0.0012 94 96 60 0.0086 0.00072 100 112 40 0.0029 0,000214 mole % of neptunium relative to the mole % of dissolved LiF-BeF. salt. The neptunium was determined by alpha counting and pulse-height 2 analyses in order to distinguish neptunium from the plutonium present as an impurity. The plutonium incidentally, appeared to be carried 2Molten-Salt Reactor Program Status Report, ORNL-2634 (Nov. 12, 1958). - 117 - with the neptunium to a considerable extent; the ratio of the amount of plutonium in solution to the amount undissolved was within a factor of 2 of the ratio for neptunium although the plutonium concentration was 500 times smaller. | Additions of iron and nickel metal to the HF solutions resulted in a significant reduction in the gross alpha activity, probably as the result of reduction to neptunium (III) and to plutonium (III). The reduction has not yet been verified by neptunium pulse-height analyses. In the trivalent state, behavior somewhat similar to that of the rare earths might be expected; the observed solubilities are in the same range as those reported previously for rare earths. It is éxpected that the rare earths, neptunium, plutonium, and, possibly, uranium will behave as a single group and will therefore exhibit lower solubilities when present together than the values reported here for the separate components. SOLUBILITY OF CORROSION-PRODUCT FLUORIDES IN HF SOLVENT Measurements were made of the solubilities of corrosion=-product fluorides (iron, chromium, and nickel) in HF solutions saturated with LiF. Chromium fluoride was found to be relatively soluble, with the solubilities being about 8, 12, and 18 mg of chromium per gram of solution in 100, 95, and 90 wt % HF solvents, respectively. Iron end nickel fluorides were found to be less soluble. The iron fluoride solubility values varied from 0.08 to 2 mg of iron per gram of solution; the average deviation of the values, neglecting extreme values, was 0.1l i 0.10. The nickel fluoride solubility varied from 0.0l to 1.3 mg of nickel per gram of solution, the average being 0.15 i.0°06° The regults were so inconsistent that no trend with HF concentration could be established. When iron, nickel, and chromium fluorides were presént together, the solubilities of all appeared to be somewhat lower. < 118 - Similar measurements in 90 and 100 wt % HF solvents with no dissolved salt gave values in the same range, but in the 80 wt % HF solvent the iron and nickel solubilities appeared to be sgignificantly higher. SOLUBILITIES OF URANIUM AND THORIUM TETRAFLUORIDE IN HF SOLVENT Measurements of the solubilities of thorium and uranium fluorides in the HF solvents indicated that these fluorides are relatively insolu- ble. All ThFh determinations gave less than 0.03 mg of thorium per gram of solution, which is the limit of detection. Better analytical methods are available for UFh’ and the solubility values were, in general, in the range 0.005 to 0.010 mg of uranium per gram of solution. In the presence of fresh iron or nickel metal, somewhat lower solubilities (0.002 to 0.005) were observed, indicating, perhaps, a higher solubility for higper oxidation states. | » WO O Ovan =W o = L] » GBES R. Alexander Bettis Billington Blankenship Blizard Boch Borkowski Boudreau Boyd Bredig Breeding Briggs Browning Campbell Carr Cathers Center (K=25) Charpie Coobs Culler DeVan Emlet (K-25) Ergen Estabrook Ferguson Fraas Franco-Ferreira Frye, Jr. Gall , Gresky Gregg Grimes Guth Se R. W. Harrill Hill Hoffman Hollaender Se H. We P. T. Householder Jordan Keilholtz Keim Kelley Kertesz - 119 - INTERNAL DISTRIBUTION L6, L. 48, L9, 50. 51. 52. 53. 5k, 55. 56. 5T 58. 59. 60. 61. 62. 63. 64, 65. 66. 67. 68. 69. 70. 71. 2. B. M. Je R. H. W. E. L. W. Je H. R. E. Je K. Je M. G. W. P. A. D. P. Je M. H. A. Jo H. W. E. A. S. G. D. R. A, BQ R. Je P. C. W. Z. P. L. Je R. ORNL-268L4 Reactors-Power TID-4500 (14 th ed.) Kinyon Lackey Lane Livingston MacPherson Manly Mann Mann McDonald McNally Metz Milford Miller Miller Morgan Murrey (Y-12) Nelson Nessle Osborn Patriarca M. Perry Phillips M. T. T. We. W. L. E. D. Je H. Reyling Roberts Robinson Savage Savolainen Scott Seagren Shipley Skinner Snell storto A. Swartout Taboada. H. E. B. Ce M. M. E. D. Taylor Thoma. Trauger Vonderlage Watson Weinberg whatley Whitman - 120 - 89. G. C. Williams 0. C. E. Winters 91. J. Zasler 92-95. ORNL = Y-12 Technical Library, Document Reference Section 96-135. Laboratory Records Department 136. Laboratory Records, ORNL R.C. 1537-139. Central Research Library EXTERNAL DISTRIBUTION 140. F. C. Moesel, AEC, Washington 141. Division of Research and Development, AEC, ORO 142-729. Given distribution as shown in TID-4500 (14th ed.) under Reactors- Power category (75 copies - OTS) Reports previously issued in this series are as follows: ORNL~-2378 Period Ending September 1, 1957 ORNL~2431 Period Ending October 31, 1957 ORNL~247k Period Ending Januvery 31, 1958 ORNL~2551 Period Ending Junz 30, 1958 CRNL~2626 Period Ending October 31, 1958