ORNL-2890 UC-81 Reactors — Power TID-4500 (15th ed.) MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT FOR PERIOD ENDING OCTOBER 31, 1959 OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency Thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. DISCLAIMER Portions of this document may be illegible In electronic image products. Images are produced from the best available original document. Printed in USA. Prico&zs_. Available from the Office of Technical Services Department of Commerce Washington 25, D.C. LEGAL NOTICE This report was prepared as an account of Government sponsored work. Neither the United States, nor the Commission, nor any person acting on behalf of the Commission: A. Makes any warranty or representation, expressed or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of any information, apparatus, method, or process disclosed in this report. As used in the above, “person acting on behalf of the Commission'' includes any employee or contractor of the Commission, or employee of such contractor, to the extent that such employee or contractor of the Commission, or employee of such contractor prepares, disseminates, or provides access to, any information pursuant to his employment or contract with the Commission, or his employment with such contractor. ORNL-2890 Reactors — Power TID-4500 (15th ed.) Contract No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT For Period Ending October 31, 1959 H. G. MacPherson, Project Coordinator o DATE ISSUED MAR 8 1960 ) * OAK RIDGE NATIONAL LABORATORY o ‘ Oak Ridge, Tennessee A operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION THIS PAGE WAS INTENTIONALLY LEFT BLANK 1 f Na N ) MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT SUMMARY Part 1. Engineering and Component Development 1.1 Component Development and Testing Development work on molten-salt-lubricated bearings was continued. In the series of seven concluded tests the start-stop tolerance of journal bearings made of INOR-8 was investigated. In all but two cases, tests were carried to seizure or impending seizure. Time of operation ranged from 3 to 379 hr. The largest number of start-stop cycles per test was 260. Signs of metal-to-metal rubbing were evident on bearing surfaces; Serious ingassing was observed during bench tests of ‘the hydrodynamic thrust bearing. Remedial efforts are being continued, and further bench tests are proposed. Two pumps equipped with hydrodynamic bearings were bench tested and then tested in molten salt. Seizuré occurred in both after short operation in salt, and the failures were attributed to dry bearings caused by ex- cessive restrictions in the suction path of the bearings. | No evidence of self-welding was found between INOR-8 surfaces immersed in molten salt at 1200°F. Coefficients of dry friction for INOR-8 on INOR- 8 and for INOR-8 on stainless steel were found to be fairly constant in the 900 to 1800°F temperature range, a good average value being 0.25. Construction of the Remote Maintenance Demonstration Facility, in- cluding the viewing equipment, was completed. The layout is being modified and additional maintenance tools are being made as need arises. The loop was tested for vacuum-tightness. 'he remote viewing equipment was modified to improve its reliability and-picture quality. A set of four television cameras provides nearly 90% viewing coverage of the cell. All major com- ponents exéept the core vessel have been disconnected and reinstalled by remote means at least once. The loop will be run with molten salt after removal and installation of the core vessel have been successfully accom- plished. Eleven long-term corrosion loops completed one year of operating time *in this quarter. Tests on two INOR-8 loops and one Inconel loop were ter- minated for metallurgical examination; the results are not yet available. One loop was shut down temporarily due to pump failure and a subsequent 0il fire. The test will be continued after repairs and reloading. The second of three sample inserts was removed from loop 9354-4 after 10,000 hr of operation. Findings on this insert are described in Sec 2.1. Two new INOR-8 loops were started up with BULT-14 salt. These loops have double-walled inserts at the exit of the second resistance heater. Y The PKP centrifugal pump with Fulton Sylphon bellows-mounted seal has logged 17,878 hr of operation in NaK at 1200°F. Only negligible leakage was oObserved. The MF-type centrifugal pump is operating in a region of cavitation in salt 30, and service time of more than 20,000 hr has been recorded to date. The average upper- and lower-seal leakage rates amounted to 30 and 12 cc/day, respectively. Trouble-free operation of the small frozen-lead pump seal continued. Total service time is 12,000 hr. The large frozen-lead-seal test was dis- continued. An apparatus for testing graphite-metal seals is under construction. A seal in the annulus between a 5-in.-0D graphite tube and a larger, mating Inconel section will be attempted first. | Sectioning of the first two in-pile loops was completed, and metallo- graphic and analytical examinations are in progress. A test program was launched to explore the behavior of graphite under conditions similar to S those encountered in graphite-moderated molten-salt reactors. Bench tests have been completed, and the first experiment is scheduled for insertion in the MIR early in 1960. 1.2 Engineering Research The viscosity of the fuel mixture BULT-14 (LiF-BeF,-UF,-ThF,, 67-18.5- 0.5-14 mole %) was determined in the temperature range from 550 to 800°C. Comparison with previous results on other salt compositions containing a high percentage of ThF, shows a maximum deviation of 18% at the extremes of the temperature range. Enthalpy measurements with the BelT-15 mikture (LiF-BeF,-ThF,, 67-18-15 mole %) yielded a heat capacity varying from 0.343 Btu-1b~ 1. (°F)"! at 1000°F to 0.304 Btu-1b~1.(°F)~! at 1500°F. The appa- ratus for measuring the thermal expansion coefficient, P, of molten salt mixtures was tested by using HTS (NaNO,-NaNO;-KNO;, 40-7-53 wt %). The . v coefficient P was 3.63 x 107%/°C between 215 and 315°C; comparison with data obtained from density measurements shows good agreement. oy Preliminary results were obtained for heat transfer with BULT-14 flow- ing in heated Inconel and INOR-8 tubes. The system was operated continu- - ously for about five weeks without difficulty. Axial temperature profiles show the effect of the laminar-turbulent transition in the hydrodynamic boundary layer in the entrance section of the tube. The heat transfer ap- pears to be about half that which would be predicted from earlier studies with molten salts; this discrepancy may result from an uncertainty in the value of the thermal conductivity for BULT-14 used in the analysis of the data. A 23% difference between the results ih the Inconel and INOR-8 tubes cannot. be explained, at present. Part 2. Materials Studies 2.1 Metallurgy In an effort to join graphite to gréphite and graphite to metal for molten-salt reactor service, an alloy (48p Ti—48% Zr—% Be) was developed and had satisfactory wetting and flowing characteristics. However, because of its poor corrosion resistance to molten fluorides, other alloy systems are being investigated. Promising alloys in the Au-Ni-Ta system are under study, and allcys in the Ni-Nb, Ni-Mo, Ni-Nb-Mo, Ni-Nb-Ta, Pd-Ni-Ta, and P&-Ni-Mo systems are being prepared for testing. Preliminary work in electric-resistance brazing of graphite to graphite, using molybdenum as the brazing alloy, has also been undertaken. The second of three INOR-8 corrosion inserts was removed after 10,000 hr from an INOR-8 forced-convection loop. Weight-loss evaluations indi- cated the insert to have lost 2.0 mg/cm?, which corresponds to a wall thickness decrease of 0.09 mil, if uniform removal of the wall is assumed. Fxamination of three Inconel forced-convection loops, which circulated fluorides for one year, was completed. Two of the loops showed normal at- tack in the form of intergranular and general voids to a depth ranging from 1 to 14 mils. The third loop was heavily attacked to depths ranging from 18 to 38 mils. Salt in the latter loop showed appreciable oxide contami- nation. Fifteen grades of graphite were classified according to their resist- ance to permeation by molten fluorides at 150 psig at 1300°F in 100-hr tests. Grade GT-123 was found to have a bulk-volume permeation of less than 0.5%, the maximum specified for the reactor, despite the high test pressure. Bismuth has been used to fill the accessible voids in AGOT - graphite prior to exposing it to molten salts. Tests are being conducted to determlne to what extent the bismuth-permeated graphlte will pick up salts durlng standard permeation tests. . When held .in direct contact with graphite for 20 hr at 1300°F, LiF- BeF,-UF, (62-37-l,mole %) apparently gettered completely the oxide contami- nants in-thelgtapsitel”.flo precipitation was detected radiographically .in the_LiF-BngjThF4—UF4 (@7{18,5-14=O;5 mole %).after it had been in contact with-grephlte'fqr 500, hr at 1300°F in ‘a vacuum. - ' | - AlNOR-8 specimens showed no metallographic indications.that carburiza- tion had occurred after being in contact with TSF graphite at 1000 psi for - 700 hr in salt environment etiIBOO°Ffl Four braze materials showed good resistance to fuel 130 (LiF-BeF,-UF,, 62-37-1 mole %) during a 5000-hr corrosion test in a thermal-convectiefi loop with a 1300°F hot;leg temperature. Corrosion tests of 100-hr duration indicated that the usefulness of titanium-, zirconium-, or beryllium;cofi-». taining brasing alloys in fuel-130 would be limited in pufe nickel or INOR~ 8 containers at 1300°F. Heavy attack due to the transfer of the refractory material to the nickel or INOR-8 container walls was observed. 2.2 Chemistry and Radiation Effects Detailed description of" the ten'subsidiary binary and ternary systems 1ncorporated in the LiF-BeF,- UF4—ThF4 quaternary system is nearly completed. Among the questlons as yet incompletely- answered by phase-equilibria studies on breeder fuels is that of detailed crystalllzatlon paths for compositions of reactor interest. ‘ The 1dent1flcat10n of the'isomorphic‘compeunds NaF -BeF,- 3UF4 afid ‘NaF -BeF,-3ThF, is of espec1al academic interest since NaF- BeF2 3UF4 has been erroneously reported in the literature 'as NaF- 4UF¢,~““' The solubility of HF. in molten fluoridé” bree@e? fuels has been cor- related with solvent composition by'means‘of's’eonCentretien scale tased on the excess or def1c1ency, in equivalent per cent, of alkall fluorlde for complete coordination of higher-valence catlons 1n the solvent “that is, Na,BeF, and Na;ZrF- correspond to complete coordlnatlon-and representf the limit of complexing power of the cation. - | ' h Exchange reactions in molten fluorides, particularly those that might' be of interest in reprocessing sehemes, are being exploréd. Many of the i reactions potentially useful for removing rare earths are subject to in- terference by essential ingredients in the fuel. The diffusion of chromium, which is the rate-limiting process for long-term corrosion in molten fluoride breeders, is strongly dependent on grain size, and hence on annealing history. Under equivalent conditions there is no significant difference between the diffusion coefficients of chromium in INOR-8 and in Inconel; thus the steady-state corrosion rate is predicted to be higher by a factor of 15/7 in Inconel, since this is the ratio of chromium concentration in the two alloys. Except for one unexplained and palpably misfunctioning case, sampled corrosion test loops are demonstrating predictable and satisfactory be- havior in long-term tests. Exploratory studies on the hygroscopicity of fuels and on the possi- bility of drying at low temperatures without extensive hydrolysis have given additional qualitative evidence of the extreme sensitivity of the fuels to moist air. Evidence of formation of a graphite intercalation compound with MoFs was found, but this compound is not expected to be encountered in reactor operation. Several grades of extruded graphite, but none of the molded types, show less than 1 vol % permeétion by breeder fuels at 60 to 75 psig and 1300°F. About 740 kg of purified fuel mixtures was processed during the quar- ter. INOR-8 corrosion test capsules have been removed from the MIR after 5500 hr of exposure and a 75% uranium burnup. 2.3 Fuel Processing Modification of the hydrofluorination—volatilization process is being investigated for application to thorium-bearing fuel salts, a principal obJjective being the retention of the processed material in the fluoride form. The solubility of ThF, in hot aqueous ammonium fluoride solutions is as high as 4 to 5%, and this system is therefore being investigated further. The use of Cl¥;-HF mixtures does not appear feasible because of the suppression of LiF solubility and the correspondingly low UF, and ThF, solubilities. The NO,-HF system is also unattractive because of the low solubility of ThF,, although that of UF, is higher. vii THIS PAGE WAS INTENTIONALLY ' LEFT BLANK CONTENTS SUWMY".U...I'..l.......l..........Q....I...O.l.l............'.... l.l. 1020 2.1. PART 1. ENGINEERING AND COMPONENT DEVELOPMENT COWONENT DEVEIDPM}NT mDT‘ESTING"...II..Ill..l......-l'..... Molten-Salt-Lubricated Bearings for Fuel PumpSecccececcsecaas Hydrodynamic Journal BearingSe.eeesccecscscssscessssscscoses Hydrodynamic Thrust BearingS.cecceeescesccscssccosssessssses Test of Pump Equipped with One Molten-Salt-Lubricated Journal Bearing..ceccecescccssesssscsscasssscsssscoccsscs Self-Welding of INOR-8e¢vieeeeansssconnnsns Pessesescsssrnos Friction Tests On INOR-8eseeesosroosssssarsocssccsasocsassaes Mechanical Seals fOr PUmMPSeceecccrssoscssssscscssscssssssossssse MF Performance LOOPee:cescsessecssessssssssssscsssssscsostsasssss Frozen-Lead Pump S€a8l.ceseecrscssesscenssssnosscsscassssssscssons Remote Maintenance Demonstration Facilityeeeesesscscosccocssns Design, Construction, and Operation of Materials-Testing LOOPS....I'.....-.’l...'.'l.l.... ..... * & @ &9 & P88 e S8 8P &S s S8 o Forced-Circulation Corrosion LOODS.ssesceccsoscoscsscsccses Graphite—Metal Seal Tests..... ..... Q& & & & o & 60 ¢ ¢ ¢ 0 &0 20 e h e e R0 In-Pile LOOP.'...........'."............I."......'...'..... ENGINEERING RESEARCH: e coeceesoscosssoscsscsoscasssccsassssnosncs Physical-Property MeasurementS.ecsecssecsssosessscoscscsanssssns ViSCOSi Yy eeeoseeeiocosossssoscesasssscsssossvossscsocssssss Enthalpy and Heat CapacCityeceecessscocsscsssssscssosssacssne Thermal EXpanSiON.:cceeececcsssessscssssasscsosscsssancssossssss Heat-TranSfer Studies.l.'.l....-.... ..... * & 8 & O " 0" 8B E T RS PSS e PART 2. MATERIALS STUDIES M}TAIIIJ[]RGYOOO ........ L I B A e 8 & 0 ¢ e B ¢ 0 b e s e b bt R s s Graphite Brazing StudieS..scecscscesccsesessssccsoscssccsnnsnss DwmiC‘CorrOSion Studies-............OC......!......O.I..... Forced-convection Loops............l........!.Q.l..l.‘..l. General Corrosion StudieS.eeecccecsesoscsssccsscsctssssasrscsosnss Permeation of Graphite by Molten SaltSecececsececsscccsnsee Removal of Oxide Contaminants from Graphitee..cececececees Compatibility of INOR-8 and Graphite in Direct Contact.... Brazing-Material Evaluation in Fuel 130...ecseesccsecscscs iii UL W W O o wm 2.2. 2.3. CHEMISTRY AND RADIATION EFFECTSececcsccoccscacanes cetscsannas Phase Equilibrium Studies...........fll.'.......I...l..“..... TIle SyStem Ber"ThF4‘IJF4....--..---o-ooooouuoooo.oo.ocoooo l The System BeFo=Thl e ieteereseocesestsosnsesossososcsanasas The System NaF-BeFo=ThF, eeeeceecersoesssnsccssoroscscccnsas Molten-Fluoride Solvents..... Cesecertaratetats et es e The Compound NaF*BeFs *3UF jcessesceccoascssosssscssssssnnns Gas SOlubilities in Molten Fluorides... LI B B B B BN B B BN BN I RN B N IR R B N ) SOlubility Ofm‘in F‘llel Solvents‘..............l..'..l'.. FiSSion-PrOduct Behavior‘.!.........l..ll‘...llO'....l.ll.l..l Exchange Reactions in Molten FluorideS.iseescescescscscess Chemistry of Corrosion ProCesSsSeSiceecscscsssssessossnsvensosans Self-Diffusion Coefficients for Chromium in Nickel AllOySeesesssssssacsnsesnonoe Sesesssssressessresserasenne Chemical Analyses of Corrosion Test LOODPS:eescresccasscssns Solubility of CrF, in LiF-BeF, (62-38 Mole %)eceuvensennen Equilibrium Measurements on the Reduction of FeF, by Hs in LiF-BeF Equilibrium Measurements on the Reduction of CrF, and FeF, by H, in LiF-BeF, (62-38 Mole %)eeeeeacennsn Cheeean Hygroscopic Behavior and Dehydration of Breeder Fuels..... Graphite Compatibility StudieS.sececeacecssorssnecsnns . . Penetration of Graphite by Molten FluorideS.iieeseessascves Intercalation of Graphite with Molten SaltS.ececceecerecccse Diffusion of Rare-Farth Fission Products in Graphite€...... Preparation of Purified Materials.eisseesse teetrstsrasnrssens Technological Operations...... Gesesasescesesescasor s e Molten Ammonium Bifluoride as & Reagent.eieeseosesosasososass Radiation Effectseeeccsceccrsscscrcnccnncs cesescessesonenaas .. Corrosion Tests Under RadiatioNeseiecsesesascsccsscccssacs FUEL PROCESSTING: e e+ eseassossscssosccssscsosesssnssnsnssessassa . Thorium Fluoride Solubility in Aqueous NHLF Solutions........ Fuel-Component Solubility in ClF3-HFe.eeeeeisatnnnccoeaacnns Fuel-Component Solubility in NOs-HF..eieeoeosessccceanscosnns 49 49 49 49 49 49 50 52 52 54 54 55 55 59 61 62 63 65 65 66 67 67 67 68 69 69 70 70 71 71 PART 1. ENGINEERING AND COMPONENT DEVELOPMENT THIS PAGE WAS INTENTIONALLY LEFT BLANK 1.1 COMPONENT DEVELOPMENT AND TESTING Molten-Salt-Lubricated Bearings for Fuel Pumps Hydrodynamic Journal Bearings Investigations of journal bearings operating in molten salt 130 (LiF- BeF,-UF,, 62-37-1 mole %) were continued throughout the quarter. Four tests were performed to investigate start-stop ability of carburized INOR-8 journals operating at temperatures from 1200 to 1500°F in INOR-8 bearings having three axial grooves; three were conducted to investigate performance of the three-helical-groove INOR-8 bearing at L200°F. A test, No. 22, of the performance of Inconel for comparison with INOR-8 is in progress. All test pieces were fitted with a radial clearance of 0.005 in. (measured at room temperature). A summary of the test conditions is presented in Table 1.1.1. Test 15 was terminated when impending seizure became evident upon restarting the run following a 30-min shutdown for repairs in the off- gas line. The support pins were twisted and partially severed from the bearing sleeve. This condition, a result of poor attachment, possibly gave rise to the impending failure. Test 16, which was a repeat of test 15 with new parts, was terminated on schedule. Examination indicated that rubbing damage was somewhat less than that experienced during a similar test at 1200°F. Test 17 was terminated by seizure during start-stop. cycle 129. The test pieces exhibited galling and metal transfer, evidences of rubbing. Test 18 was conducted until seizure occurred between the bearing and journal. They did not meet dimensional specifications; the test should therefore be repeated. Test 19 was conducted with bearing grooves having a cross-sectional area less than that used in the other tests with helical-groove bearings. Flow passage was not restricted at the discharge end of the grooves. Seizure was impending when an attempt was made to place the bearing back in operation after the test had been stopped for a start-stop operation following a test at 300 1b Signs of rubbing were evident on the test £ pieces. Test 20 was performed with bearing grooves of the same cross section as that used in test 19, but with a flow restrictor over the discharge Table 1.1.1. Hydrodynamic-Journal-Bearing Tests; Summary of Test Conditions Operating Radial Operating Test Material Groove Speed Temperature load Time Start-Stop No. Bearing Tournal Configuration* (rpm) (°F) (lbf) (hr) Cycles 15 INOR-8 INOR-8, Axial 1200 1200 200 21 2 carburized ' 1300 24 16 INOR-8 INOR-8, Axial 1200 1300 200 123 260 carburized 17 INOR-8 INOR-8, Axial 1200 1400 200 94 - 129 carburized 18 INOR-8 INOR-8, Axial 1200 1500 200 3 1 carburized 19 INOR-8 INOR-8 Helical 1200 1200 10-300 136 1 20 INOR-8 INOR-8 Helical 1200 1200 10-150 22 1 21 INOR-8 INOR-8 Helical 1200 1200 500 379 106 600 1200 375 22 Inconel Inconel Axial 1200 1200 200 336 10 *¥A1l bearings contained three grooves. -2 end of the grooves. The flow restrictor was a thin disk having an inside diameter 0.010 in. larger than the inside diameter of the bearing. The test was terminated because of a slight roughness in the drive-motor power trace. Examination of the test pieces did not reveal any apparent cause for this problem. Test 21 was a continuation of test 20, except that the flow restric- tion on the bearing grooves was reduced by adding a 1/16- by l/l6-in. V- notch to the shim at each groove. The test was terminated by apparent seizure in the 600-rpm run with 375 lbf load. There was slight evidence of rubbing between the test surfaces. Test 22, now in progress, is being performed with an Inconel journal and a three-axial-groove Inconel bearing. Hydrodynamic Thrust Bearings Bench tests were performed on the thrust-bearing tester with the thrust bearing immersed in oil. Gross ingassing was found, and a number of modifications were investigated as possible corrective measures. Sev- eral of these were promising, but further bench testing with other modifi- cations is in order prior to hot tests in molten salt. Test of Pump Equipped with One Molten-Salt-Lubricated Journal Bearing Bench tests were conducted on the original pump design, shown in Fig. 1.1.1, to investigate the performance of the overhung motor, of the oil slinger in pumping lubricant to the upper bearing, and of the journal bearing immersed in oil. The overhung motor was found to exert a large rotating radial load. Investigations show, however, that the load can be reduced to tolerable levels if the electrical balance of the motor is improved. In the in- terest of speeding the investigation of a molten-salt bearing applied to a pump, the overhung motor was temporarily replaced by a conventional motor with a flexible shaft coupling (Fig. l.l.2), which will isolate the pump shaft from the effects of imbalance in the motor. The oil slinger had to be modified before satisfactory flow of lu- bricant to the upper bearing could be attained. The modifications con- sisted of slitting the slinger and increasing the diameter of the oil return ports. UNCLASSIFIED ORNL-LR-DWG 44831 SALT LEVEL — CANNED MOTOR & i3 Tl Yo, sy, | \Q\., % OIL SLINGER HELIUM COOLING BARRIER FACE-TYPE SEAL — SEAL LEAKAGE OUT B oz g 7 u\ Qi \A/\/A\/ /A/ /\/\\r\\\\\\\\\\\\\\ \\\\\\\\\\\\\\\\\\\\\\\\\A/\I\I\/\ll&l/&‘\\\\\\\\ __= __ I __; ; . i :z 7_‘_ l I LAk TTi \\ \///// QAN /I/’/ ,//\\\\\\\\\\\\\\\.\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\§\\ s \\\\\ § \\§ \\\\\ 722 \\\\\\\\\ 2227 »\\ \\\\\\\\\ 7Z77% \ .\\\\\\\.\\\\\\\\\ \\ ,\ ,! - “\\” N\ _ l\\\\\\\\\\\\&l_& T i I, \\ J V) AN NN NNRRNERNY :_— m. 7 Z ’ { & 7 TS0 5920717 \\\\\\\\\\\\\\\\\\\\\\\\.\\\\\\\//: Z A ST I N //52, 7 %. -.|_ - YR i + / o B2 z 2@ W 35 ¢ (] o oo = ._.._I._u s _rNu 383 5 gz ] 2 ag 77777777 RUuL 5 WB 29 5 @ i 8a w = z w = o = = £ Sg - 3 g x 59 = g = 2o 1 - HE o Fig. 1.1.1. Pump with Molten-Salt-Lubricated Journal Bearing and with an Overhung Motor. UNCLASSIFIED ORNL-LR-DWG 41765A MOTOR OIL SLINGER & // ROLL BALL BEARING OIL-LUBRICATED DOUBLE- FACE-TYPE SEAL =— SEAL LEAKAGE OUT HCLIUM PURGE IN —3 COOLING BARRIER \\\<\\\\\\ SALT-LUBRICATED HYDRO- T \\\\\\x\\\%\\% ‘ \\\\\\\\\\\\\-\.\\\\\ A0 L0 000 . DYNAMIC JOURLNAL BEARING ~— SALT LEVEL L \\§>\\\\V [ - J ?;;555 oo \\ 257577 ,//(1: 2 2 WA - ‘1 27;?2. NS [z E [ ; - &y AARAENONANAY S AN IMPELLER Pump with Molten-Salt-Lubricated Journal Bearing and with a Flexible Shaft Coupling to a Conventional Motor. Fig. ]Q]'2. Two Jjournal-bearing configurations were tested in oil at room tem- perature and in molten salt 130 (LiF-BeF,-UF,, 62-37-1 mole %) at 1200°F. The oil tests were performed with the pump assembled except for the im- peller; the complete assembly was operated in the tests with molten salt. The first bearing tested had three axial grooves and was mounted rigidly. The bearing operated successfully in oil under load, and the pump was subsequently placed in operation with molten salt. The pump was rotated for 10 min at 1000 rpm and another 10 min at 1200 rpm; at 1200 rpm the pump delivered a flow of 100 gpm. The operation was terminated as a re- sult of seizure of the pump rotary element, which is believed to have oc- curred in the molten-salt-lubricated journal bearing. The other bearing had three helical grooves with a flow restriction at the discharge (upper) end. The bearing was mounted flexibly on gim- bals. It operated satisfactorily in oil but subsequently failed by sei- zure in operation with molten salt. Study of the configuration near the bearing revealed that the im- peller hub probably presented too great a restriction to flow of salt into the bearing at the lower end; the bearing therefore pumped itself dry in the molten-salt tests. This condition was probably responsible for both the failures. During the oil tests this condition was not present, since they were made without the impeller present. The flow passage into the bearing has been increased by removal of material from the impeller hub and the lower end of the bearing sleeve. Preparations are being made for a third test. Self-Welding of INOR-8 One INOR-8 self-welding experiment was performed at 1200°F, on loose- fitting pins pressing against the sides of holes, a condition encountered with gimbal-mounted bearings. The test was performed in a bath of molten salt 107 for 1000 hr under a load of 200 lbf. Examination of the test pieces revealed no evidence of self-welding. A similar test will be per- formed at 1500°F. Friction Tests on INOR-8 Friction tests were performed at MIT! on INOR-8 rubbing against INOR-8 and on INOR-8 against stainless steel. Four tests were performed, each of i, Rabinowicz. which covered a temperature range of 70 to 1830°F. The relation between the coefficient of friction and temperature is shown in Fig. 1.1.3. Note thdt in each test the coefficient drops off over the range of 700 to 900°F and then, in general, remains at approximately 0.25 on up to 1830°F. Mechanical Seals for Pumps The modified Fulton Sylphon bellows-mounted seal,2 undergoing test in a PKP type of centrifugal pump, has accumulated an additional 3048 hr of operation since the previous report period, for a total of 17,878 hr. The pump continues to operate at constant conditions of 1200°F loop tem- perature, 2400-rpm shaft speed, and 1200-gpm NaK flow. The seal leakage in both the test seal and the Durametallic upper seal has been negligible. The pump was stopped once to check the drive-motor brushes. MF Performance Loop An MF type of centrifugal pump has continued in operation?® and has logged more than 20,000 hr of continuous operation. Since the previous report period the pump has continued operation in a region of cavitation at 2700 rpm, 645 gpm, and 2.5 psig pump-tank cover gas pressure in molten °MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 19. UNCLASSIFIED ORNL — LR—DWG 44488 1.0 ‘ | [ T | I | I | I INOR-8 ON INOR-8 IN ATMOSPHERE INOR-8 ON INOR-8 IN ARGON | —_ | o LOAD: 1750 g o LOAD: 1750 g SPEED: 4.6 cm/sec SPEED: 4.6 cm /sec 0.5 5 PO n==0 [0 d 40 w v I 1.0 T T T T T T T T T T w INOR-8 ON INOR-8 IN NITROGEN STAINLESS STEEL ON INOR-8 IN ATMOSPHERE : ot g LOAD: {750 g w LOAD: {750 ¢ S s MMN\ SPEED: 4.6 cm/sec o - \\ SPEED: 4.6 cm/sec \ ] . O e — %r\f 0 0 500 1000 1500 2000 O 500 1000 1500 2000 EMPERATURE (°F) Fig. 1.1.3. Coefficient of Friction vs Temperature for INOR-8 Against INOR-8 and for INOR-8 Against Stainless Steel. salt 30 at 1200°F. During the quarter the pump was stopped twice — once to replace the air filter on the motor-generator set and once (for 10 min) to replace the brushes in the d-c drive motor. The pump upper- and lower- seal leakage rates have averaged approximately 30 and 12 cc/day, respec- tively. Frozen-Lead Pump Seal The small frozen-lead pump seal being tested on a 3/l6—in.-dia shaft, as described previously,? has operated continuously since the start of its operation on June 13, 1958. The accumulated operating time, as of October 31, 1959, was 12,000 hr. There was slight leakage of lead during the first 100 hr of operation; there has been no further leakage. Testing of the large frozen-lead pump seal was discontinued. Remote Maintenance Demonstration Facility Construction of the equipment in the Remote Maintenance Demonstration Facility was completed during July. The entire loop of piping (made of 3-1/2- and 6-in. Inconel pipes) was stress-relieved at 1500°F for 2 hr. A check of torque settings on the bolts of the freeze flanges after the stress-relieving operation revealed that all suffered loss of tightness; in some instances the bolt torque was reduced to as little as 40% of the original setting. To assure constant bolt torque on the freeze flange Jjoints during thermal cycling, the use of disk springs was suggested. Three sets of Bellville-type springs are presently being tested for this applicatibn. A system of controlling the air flow to the cooling passages of the freeze flanges was designed, and construction drawings were completed. It is believed that such a flow control will eliminate the adverse effects of the differential thermal expansion between the Inconel flanges and copper seal rings and will improve the gas-tightness of the system during thermal cyclings. The loop was tested for vacuum-tightness with the flange bolts tight- ened to their original setting after stress-relieving. A pressure rise of 52 p was recorded in a 56-hr period with the loop isolated. The system was originally evacuated to 17 u. 3MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 23. 10 Television equipment for remote viewing of the cell has been in- stalled. The closed-circuit system consists of four remotely operated cameras. Two of these, equipped with three-lens turrets and used as a stereo pailr (Fig. 1.1.4), are mounted on a dolly and are free to travel the whole length of the cell on a rail mounted on top of the control room. The other two, with "auto zoom" lenses, are mounted in other locations of the cell. The complement of cameras allows nearly 90% coverage of the cell. All four cameras feed into a pair of monitors mounted on a stereo viewer above the General Mills manipulator console (Fig. 1.1.5). The picture quality and equipment reliability were not satisfactory during early operation; however, with the assistance of a factory repre- sentative, local maintenance personnel were able to correct most of the difficulties. Upon completion of suggested modifications, the frequency of maintenance on the viewing equipment was reduced, at times, from 1 in 24 hr to 1 in 200 hr of operation. Remotely viewed operation of the General Mills manipulator and over- head crane was started in September. Tasks successfully executed by re- mote means to date include removal and installation of the PK salt pump, UNCLASSIFIED PHOTO 35259 Fig. 1.1.4. Stereo Camera Pair for Viewing Operations in the Remote Maintenance Demon- stration Facility. 11 ; o ) UNCLASSIFIED . PHOTO 35261 | Wil s I L \ (g . iy iy, o ity g e, e = i, Fig. 1.1.5. Control Center for General Mills Manipulator Console, Closed-Circuit Tele- vision, and Overhead Crane; Remote Maintenance Demonstration Facility. its motor, the dump tank, the freeze valve, the electric heater and in- sulation unit, and the heat exchanger. Numerous modifications were incorporated in the loop, as the need arose, either to improve visibility through the cameras or to facilitate maintenance operations. Removal and installation of the heat exchanger, for example, were made practicable by providing guide pins and funnels at the mating flanges and by adding a screw-jack mechanism to close and back away the heat exchanger (Figs. 1.1.6 and 1.1.7). A variety of tools were fabricated and are in use in conjunction with the overhead crane and the manipulator. Remote removal of the core vessel is scheduled prior to running molten salt through the loop. Upon completion of the salt run the dis- assembly-assembly cycle will be repeated on all components. 12 § UNCL ASSIFIED PHOTO 35258 ; = 2 g ; Fig. 1.1.6. Guide Pins and Funnels for Alignment of Flanged Joints in the Remote Main- tenance Demonstration Facility. 13 UNCL ASSIFIED PHOTO 35260 Fig. 1.1.7. Auxiliary Screw Jack for Manipulating the Heat Exchanger in the Remote Main- tenance Demonstration Facility. Design, Construction, and Operation of Materials-Testing Loops Forced-Circulation Corrosion Loops The operation of long-term forced-circulation corrosion-testing loops was continued. Table 1.1.2 gives a summary of 16 loop operations during the period, with accumulated hours of operation at test conditions. Eleven of the loops have operated for a full year. (Four other loops with one year of operation had been terminated earlier.*) Operation of INOR-8 loops MSRP-8 and -9 and Inconel loop 9377-4 was terminated during this period for metallurgical examination. The two INOR-8 loops had operated without an incident; no dumping or freeze-up of the salt was experienced in the loops although both sustained at least five building-power failures. The LFB pump in the Inconel loop had been “MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684, p 61; Apr. 30, 1959, ORNL-2723, p 33; July 31, 1959, ORNL-2799, p 31. 14 Table 1.1.2. Molten-Salt Forced-Circulation Corrosion Loops; Operations Summary as of October 31, 1959 Operation at Sl Loop Material Cififig}fd Test Conditions Comments (hr)** 9354-3 INOR-8 84 14,538 Pump has been replaced; loop is awaiting a new batch of salt 9354-1 INOR-8 126 13,746 Normal operation 9354~4 INOR-8 130 10,756 Insert sample No. 2 removed at 10,000 hr; operation resumed MSRP-7 INOR-8 133 10,534 Tubing sample removed after 1 yr; operation resumed MSRP-6 INOR-8 134 10,31C Normal operation MSRP-9 INOR-8 134 9,687 Terminated Oct. 21, 1959 MSRP-8 INOR-8 124 9,633 Terminated Oct. 12, 1959 MSRP-10 INOR-8 135 9,553 Tubing sample removed after 1 yr; operation resumed 93774 Inconel 130 9,552 Terminated Sept. 14, 1959 MSRP-11 INOR-8 123 9,344 Normal operation MSRP-12 INOR-8 134 8,845 Normal operation 9377-5 Inconel 134 8,015 Normal operation 9377-6 Inconel 133 5,830 Normal operation MSRP-13 INOR-8 136 4,901 Normal operation MSRP-14 INOR-8 BULT-14 280 Began operation Oct. 14, 1959 MSRP-15 INOR-8 BULT-14 200 Began operation Oct. 23, 1959 *Compositions (mole %): 84 NaF-LiF-BeF, (27-37-38) 133 LiF-BeF,-ThF, (71-16-13) 123 NaF-BeF,-UF, (53-46-1) 134 LiF-BeF,-ThF,-UF, (62-36.5-1-0.5) 124 NaF-BeF,-ThF, (58-35-7) 135 NaF-BeF,-ThF,-UF, (53-45.5-1-0.5) 126 LiF-BeF,-UF, (53-46-1) 136 LiF-BeF,-UF, (70-10-20) 130 LiF-BeF,-UF, (62-37-1) BULT-14 LiF-BeF,-ThF,-UF, (67-18.5-14-0.5) **¥A11 tests are operating with a 1300°F high wall temperature and a AT of 200°F between high wall and low fluid temperatures, with one exception: MSRP-15 has a high wall temperature of 1400°F. replaced three times, which required dumping and cooling of the loop. Examination of the loops i1s in progress. INOR-8 loop 9354-3 was shut down in September when a pump failure resulted in an oil fire. The pump has been replaced, and operation of the loop will be resumed on receipt of a new batch of salt. INOR-8 loop 9354-4 was shut down in September, after 10,000 hr of operation, for the removal of the second of three sample inserts® con- tained in the hot leg. Results of the examination of this insert are given in Sec 2.1 of this report. Operation was resumed with the remaining (third) insert. INOR-8 loops MSRP-7 and -10 were shut down for the removal of tubing sections from the hot leg after 8760 hr (one year) of operation. The sec- tions were replaced and operation was resumed. The sections are now being examined by the Metallurgy Section. Two new INOR-8 loops, MSRP-14 and -15, were placed in operation with salt BULT-14 in October. Both loops contain molten-salt sampling devices and three Jjacketed insert sections. The double-walled inserts were placed at the exit of the second resistance-heater section as shown in Fig. 1.1.8. The close-fitting combination of Jjacket and insert was designed to give the same cross section and flow area as the matching loop tubing. Loop MSRP-14 is operating with a high wall temperature of 1300°F, and MSRP-15 with a high wall temperature of 1400°F. Graphite-Metal Seal Tests Several designs for graphite-to-metal frozen-salt seals have been proposed. In the first of the configurations to be tested, the frozen- salt seal will be formed in a narrow annulus between a 5-in.-0D graphite tube and a slightly larger Inconel tube. Such a seal could be used in a molten-salt breeder reactor as the seal between the fuel salt and breeder salt. A sectional view of the test-joint design is shown in Fig. o don S Manufacture of parts for the seal-test apparatus is nearly complete, and construction of the test loop has been started. It is expected that the first test of the seal will be made before the end of November 1959. 5MSR Quar. Prog. Rep. Apr. 30, 1959, ORNL-2723, p 33. 16 UNCLASSIFIED ORNL-LR-DWG 43719 /RESISTANCE— HEATER LUG o O O O ~=————— DIRECTION OF FLOW TOTAL LENGTH OF SECOND HEATER, 109 in. = = = =1 = 3§ \ %mu&a-mufio 3/g-in. SCHED-10 INOR-8 PIPE INSERT SECTIONS = =i 1) Fig. 1.1.8. Double-Walled Insert Sections Installed in INOR-8 Loops MSRP-14 and -15. In-Pile Loop Sectioning and metallographic examination of the first MSR in-pile loop are complete. Analysis of the fuel and of radioactive deposits in the gas passages of the loop is in progress. The second loop has been sectioned, and some preliminary metallographic examination made. The results will be issued after both loops have been examined. An irradiation program, in conjunction with out-of-pile tests, is under way to study the effects of radiation and fissioning on the prop- erties and fuel penetration of graphite under conditions simulating those of a graphite-core molten-salt reactor. Specimens of promising grades of low-permeability graphite will be submerged in enriched molten-salt fuel under pressure and irradiated in the MTR HB-3 beam hole at 1300°F. A bench test of the irradiation experiment has operated satisfactorily, and the first experiment is scheduled for insertion in the MIR early in 1960. 17 18 0 i 2 3 4 5 SCALE IN INCHES THERMOCOUPLE EXPECTED OPERATING TEMPERATURE: {300° F AIR-COOLING TUBES _ll EXPECTED OPERATING TEMPERATURE: ' 850°F i o, iyt ik INCONEL GRAPHITE GASKET—— ! I I : \\ \\\\\ HEAT REFLECTOR N UNCLASSIFIED ORNL-LR-DWG 44489 SALT INLET SALT OUTLET \ . TEMPERATURE CONTROL HEATER 3 THERMOCOUPLES D Z N R (o] o Z 0000 < » A& S ; INITIAL CLEARANCE : ‘Y4 in. 4 INITIAL CLEARANCE: %4 in. P (FROZEN-SALT SEAL) . : 5 5'%gin n |7~ CONAX SEAL ) I ! THERMOCOUPLES o | — ok ! : = 000 = ‘. 1§ l'li' N \\\%\\\\‘ Ml N\ ] : i i f N ")‘” “ NN MU\ TEMPERATURE CONTROL AN HEATER ‘ ": \ I STEEL POT A “Q\Q ! ml’r ; ) ’17 IRIN THERMAL INSULATION i Wl | R Al AR THERMOCOUPLE i "‘ \ IBERAN il st i\l* (I \\\\ 1 RO [ 1IN Al BN Vel l‘”'l lw” Fig. 1.1.9. Graphite-Metal Freeze-Joint Test Apparatus. 1.2 ENGINEERING RESEARCH Physical-Property Measurements Viscosity The viscosity of the fuel mixture BULT-14 (LiF-BeF,-UF,-ThF,, 67- 18.5-0.5-14 mole %) was obtained by using the skirted capillary efflux viscometer.! The data from three separate determinations using different viscometer cups are given in Fig. 1.2.1. Also plotted is the correlating line derived by a least-squares analysis of these data; in the temperature range from 550 to 800°C, this analysis yielded the equation L = 0.08126%422/T ) where p is the viscosity in centipoises and T is the absolute temperature (°K). The viscosity of the BULT-14 mixture is compared in Fig. 1.2.2 with earlier results'™ on salts 133 (LiF-BeF,-ThF,, 71-16-13 mole %), 134 1MSR Quar. Prog. Rep. Jan. 31, 1959, ORNL-2684, p 65. °MSR Quar. Prog. Rep. Apr. 30, 1959, ORNL-2723, p 39. 3MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 34. UNCLASSIFIED ORNL-LR-DWG 44490 f, TEMPERATURE (°C) 500 600 700 800 a ¢ T 1 | T e LIQUIDUS TEMPERATURE | g g \\‘\@ & 9 _ qa22 /1 = Q\A B e~ g 16 }—m—m—m ~ O S O CUP NO. 52 e S ——— @ CUP NO. 53 % : 8 ———— A CUP NO. 54 o) =Y '} \\\ - 6 & \\A 9‘\\\ a 0.0043 0.0042 0.0011 0.0010 0.0009 Yr [ex)™ ] Fig. 1.2.1. Viscosity-Temperature Relationship for the Fuel Mixture BULT-14 (LiF-BeF ,- UF ,-ThF ,, 67-18.5-0.5-14 Mole %). 19 UNCLASSIFIED ORNL-LR-DWG 44494 t, TEMPERATURE (°C) 450 500 600 700 800 T T T I I I | LIQUIDUS TEMPERATURES 134 133 I BULT-14 SALT NO COMPOSITION (mole %) 40 | BELT— 15 LiF [ BeFp | UFg | Thig Sa; 5 “ 133 71| 16 13 B No 13 134 62 |365| 05 | 10 ar S BULr\;’\Qr BELT—15 | 67 | 48 15 * 433 %\rs BULT—14 | 67 | 185 05 | 14 20 § o~ \\\ = : \ > 8 \\ BULT -144 2 s \\ ] = _— SALT NO. 133 ] i %\ \ ()] 4 BELT-15 — SALT NO. 134 2 0.0014 0.0013 0.0042 0.0011 0.0010 0.0009 Yr [(°K)~'] Fig. 1.2.2. A Comparison of the Viscosities of Four Thorium-Containing Molten Salt Mixtures. (LiF-BeF,-ThF,-UF,, 62-36.5-1-0.5 mole %), and BeLT-15 (LiF-BeF,-ThF,, 67-18-15 mole %). The viscosities of the three high-thorium-content salts deviate by a maximum of 18% (at the extremes of the temperature range). This large variation is rather surprising in view of the small composition change between BeLT-15 and BULT-14; the viscosities of these two mixtures will be redetermined in the near future. Results for the BeFj,-containing salts studied to date are summarized in Table 1.2.1. Enthalpy and Heat Capacity Analysis of the previously reported data* on the enthalpy of the BeLT-15 mixture was completed. For the solid (100-400°C), the enthalpy (in cal/g) can be expressed as H, — Hzo = 6.5 + 0.190t + (5.5 % 1077)t= , 4Tbid., p 38. 20 Table 1.2.1. Viscosities of BeF,-Containing Salts salt No. Composition (mole %) ggiiii;ii* IiF NeF BeF, UF, TuF, . = 123 53 46 1 .0234 5505 126 53 46 1 0104 6487 130 62 37 i .0580 4550 133 71 16 13 L0526 4838 134 62 36.5 0.5 3 L0680 4497 136 70 10 20 L0489 4847 BeLT-15 67 18 15 0311 5308 BULT-14 67 18.5 0.5 14 0.0812 4422 B/T_ | o *Constants appear in the equation, p = and for the liquid (550-800°C) as H - Hzo = =38.5 + 0.418t — (7.0 % 107 )J&> where t is the temperature (°C). The heat of fusion at 500°C was deter- mined to be 48 cal/g. Preliminary data on the enthalpy of the BULT-14 composition are given in Fig. 1.2.3; analysis is continuing. Thermal Expansion The apparatus for determining the thermal expansion of molten-salt mixtures is pictured in Figs. 1.2.4 and 1.2.5. This device obtains, as described in an earlier report,5 the thermal expansion of the liquid by measuring the difference in height of the fluid in the two legs of a U- tube when the legs are maintained at different temperatures. The density at any temperature, t, is expressed as Py = Poll = B(t = to)] SMSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 39. 21 UNCLASSIFIED ORNL-LR-DWG 44492 225 l LIQUIDUS L TEMPERATURE W 200 7 § o L \/( > 475 —g 8 Y s o 150 ) T / |K / * / . 125 ) / u / 4 L 5 100 S r o "} > 75 // o= z el Z 50 ™ 25 — 0 0 100 200 300 400 500 600 760 #, TEMPERATURE (°C) Fig. 1.2.3. Enthalpy-Temperature Relationship for the Fuel Mixture BULT-14 (LiF-BeF ,- UF ,-ThF ,, 67-18.5-0.5-14 Mole %). where p, is the density at some base temperature, tp. The volume coeffi- cient of expansion is thereby obtained, to a close approximation, as 5 d(x; — x5) s C + X3 daT . where x; is the height of the liquid in leg 1 and x,; 1s the height of liquid in leg 2, both measured above the arbitrary fixed point of the U- tube height, ¢ (note Fig. 1.2.6); T is the temperature difference (°C) between the two legs. In practice, it was found most convenient to vary the temperature of one leg only, the other being held at some fixed ref- erence temperature. Experimental data from an operational study of this apparatus using the salt mixture HTS (NaNO,-NaNO,;-KNO,;, 40-7-53 wt %) are shown in Fig. 1.2.6. From this figure, the slope, dX/dT (where X = x; — x,), was found to be 4.808 x 10~3 in./°C; B is then calculated to be 0.000363 (g/cm3)- (g/cm?®)~%-(°C)™* (22%) over the temperature range from 215 to 315°C. This can be compared with the value of B, 0.000378, obtained from density 22 UNCL ASSIFIED PHOTO 34940 LIQUID~-LEVEL MEASUREMENT \Q Fig. 1.2.4. Apparatus for Determining Thermal-Expansion Coefficient of Molten Salts. measurements by Vargaftik et 3&.6 Measurements with the BULT-14 mixture were initiated. Heat-Transfer Studies Preliminary measurements were made of the heat-transfer coefficients for the salt mixture BULT-14 (LiF-BeF,-UF,-ThF,, 67-18.5-0.5-14 mole %) flowing in heated Inconel and INOR-8 tubes. The apparatus, described in carlier reports,”’ is shown in Figs. 1.2.7 and 1.2.8. Initial operation of the system was terminated, after 24 hr of cir- culation at a mean temperature of 1180 to 1195°F, by an internal leak in the heat exchanger. Since it had been observed during this period that éN. B. Vargaftik, B. E. Neimark, and O. N. Oleshchuk," Physical Properties of a Fluid Heat-Carrier for High Temperatures," Izvest. Vsesoyuz. Teplotekh. Inst. im. Feliksa Dzerzhinskogo 21(9), 1-7 (1952). "MSR Quar. Prog. Rep. Apr. 30, 1959, ORNL-2723, p 41; MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 39. 23 UNCL ASSIFIED | PHOTO 34939 Fig. 1.2.5. Salt Container and Liquid-Level Probes for Thermal-Expansion-Coefficient Apparatus. 24 UNCLASSIFIED ORNL—LR—-DWG 44493 % = e 0.3 | 0.2 2X =4.808 %1072 in./°C / 1, =265°C (CONSTANT) < S o - ‘ 1 4 w W e o = 0 X, LEG LEG w > w e ® > | - =50 —40 =30 =20 =0 0 10 20 30 40 50 T=t—t,, TEMPERATURE DIFFERENCE (°C) Fig. 1.2.6. Experimental Data for Thermal-Expansion Coefficient of HTS (NaNO,-NaNO,- KNO,, 40-7-53 wt %). the heat loss from the piping alone was sufficient to maintain tempera- tures at the desired levels, the heat exchanger coolant lines were plugged; and circulation was resumed. At higher power levels, partial removal of the thermal insulation in the heat exchanger region satisfied the addi- tional cooling requirements. The system has now operated continuously fur sume [lve weeks without ditf't'iculty. During this period, At runs have been made at regular intervals; normal circulation takes place under isothermal conditions. Typical axial temperature profiles (outside tube surface) for both the Inconel and INOR-8 sections are shown in Fig. 1.2.9. The tube-wall temperatures have been adjusted to a consistent base by using data from a series of runs in which the power level was set so as to just maintain the salt passing through the test section at its inlet temperature. Under such essentially isothermal conditions, corrections for the tube-wall 25 UNCI ASSIFIED PHOTO 35004 L i S e e s e IS TEST Fig. 1.2.7. Continuous-Circulation System for Study of Heat-Transfer Coefficients with Molten Salts. 26 i UNCL ASSIFIED - ELECTRODE GUARD HEATER ; e i : (Al | THERMOCOUPLESQ&\ . P e g Fig. 1.2.8. Close-up of Inconel Test Section for Study of Molten-Salt Heat Transfer. thermocouples can be obtained by comparison with the inlet-fluid mixed- mean thermocouple, and the power consumed is a measure of the system heat loss at the operating temperature. Of interest is the unusual shape of the tube-wall temperature profiles of Fig. 1.2.9. Since the fluid enters the test section directly from the larger-diameter mixing chamber and heating occurs over the entire length of the test section, the entrance condition is complicated by the concurrent dewvelopment of the thermal and hydrodynamic boundary layers. The peaks in the temperature curves can then probably be explained in terms of a laminar-to-turbulent transition in the hydrodynamic boundary layer. Linke and Kunze,8 studying heat transfer in the entrancc region of tubes with the simultaneous develop- ment of the flow and temperature fields, found a similar dip in the heat- transfer coefficient (corresponding to a peak in the temperature) in the 8W. Linke and H. Kunze, Allgem. Warmetech. 4, 73-79 (1953); see also E. R. G. Eckert and R. M. Drake, Heat and Mass Transfer, p 213, McGraw- Hilly 1959, t, TEMPERATURE (°F) ', TEMPERATURE (°F) 28 1200 1180 1160 1140 1120 1280 n o) o N H o 1220 UNCLASSIFIED ORNL—LR— DWG 44494 x/d , DISTANCE ALONG TUBE /0 . ‘r_ / TUBE OUTSIDE SURFACE ,0 / / Af0ve|'-c|ll J) FLUID MEAN O (@) @ / \l . P—— TUBE OUTSIDE SURFACE ! Afover-cll FLUID MEAN ‘ — (6) 10 20 30 40 50 60 Fig. 1.2.9. Typical Axial Temperature Profiles as Obtained in the Study of the Heat- Transfer Characteristics of an LiF-BeF,-UF ,-ThF, (67-18.5-0.5-14 Mole %) Salt Mixture. (a) Run MA-2, Inconel section, Ngo= 7980; (b) run MA-13, INOR-8 section, Npe = 13,940. vicinity of x/d = 10. .Further, the location of this temperature peak should be related to the configuration of the tube entrance in that the transition should.occur farther downstream for a rounded entrance than for a sharp-edged entrance. This may explain the shift in the location of the peak between the curves a and b of Fig. 1.2.9; this same difference between the temperature patterns in the Inconel and INOR-8 sections has been observed in all runs made to date. Although a sharp-edged entrance was specified for both sections, welding of the thin-walled tubes into the heating lugs (mixing-chamber end plates) may have altered this situ- ation. Thus it is not possible at this time to define exactly the en- trance configuration. Far downstream (x/d > 40) the temperature curves assume the character observed in earlier studies with molten salts® under similar conditions of uniform heat input at the wall and established flow and tcmperature fields. The results obtained to date are summarized in Fig. 1.2.10 in terms of the heéat-transfer parameter,'NNu-NEg'4. The wall-to-fluid temperature difference used in the analysis of the data was taken in the region of fully developed conditions (as indicated in Fig. 1.2.9). The heat flux was based on the heat gained by the fluid in passing through the test section; heat balances (corrected for external heat loss) ranged from 0.9 to 1.2. The data are compared in Fig. 1.2.10 with the general heat- transfer correlation!® for normal fluids under moderate At conditions, ~ 0.8 0.4 Ny, = 0.023 N2 WY also shown are earlier results'l 12 with Flinak (LiF-NaF-KF, 11.5-42-46.5 mole %) and fuel mixtures 30 (NaF-ZrF,-UF;, 50-46-4 mole %) and 130 (LiF- BeF,-UF,, 62-37-1 mole %). The low results with the BULT-14 salt mixture may possibly be attributed to the uncertainty in the value of its thermal °H. W. Hoffman, Turbulent Forced Convection Heat Transfer in Circular Tubes Containing Molten Sodium Hydroxide, ORNL-1370 (Oct. 3, 1952). 10y. H. McAdams, Heat Transmission, 3d ed., p 219, McGraw-Hill, 1954. 114, W. Hoffman, Molten Salt Heat Transfer, ORNL CF-58-2-40 (Feb. 18, 1958) . 127, C. Amos, R. E. MacPherson, and R. L. Senn, Preliminary Report of Fused Salt Mixture No. 130 Heat Transfer Coefficient Test, ORNL CF- 58-4-23 (Apr. 2, 1958). 29 conductivity, which was assumed to be 1.3 Btu.-hr™t.£t7™1.(°F)™" in esti- mating the heat-transfer parameter. It is planned to determine experi- mentally the thermal conductivity of the BULT-14 mixture in the near future. The discrepancy observed between the Inconel and INOR-8 data (approximately 23% at NRe = 10,000) is not explainable at present. Future data, obtained at periodic intervals, will serve to establish whether there exists a progressive film formation for the BULT-l4 salt in contact with Inconel or INOR-8. UNCLASSIFIED ORNL-LR-DWG 44495 100 80 % 60 va Ny Vo0 =0.023 W 28 / 7 . 40 // ‘ 74 4// o, °* /// A . // ® ///‘ ¢ LA 20 p iy /4 ® « / /’ 7/ A i/ 3 7y ¢ | A Z 7/ a // , [1 ° L /Y [75] S 10 VA & / [/ 5 1/ w 8 / 7 I / . < /- O / / o 2 g / ~ S Z / 4 / ® INCONEL SECTION - FLUID TEMPERATURE = 1247 °F / A INOR-8 SECTION=-FLUID TEMPERATURE A FLINAK ~ HOFFMAN ‘ —— - ——FUEL MIXTURES 30 AND (30~ AMOS, MACPHERSON, SENN , , // ' ] 103 2 5 10* 2 3x10¢ _ REYNOLDS MODULUS, Mge ' Fig. 1.2.10. Experimental Results for Heat Transfer with BULT-14 Salt Mixture (LiF- BeF ,-UF ,-ThF ,, 67-18.5-0.5-14 Mole %). 30 PART 2. MATERTALS STUDIES THIS PAGE WAS INTENTIONALLY LEFT BLANK 2.1 METALLURGY Graphite Brazing Studies Graphite has been proposed for use as the moderator of a molten-salt reactor. This material could be placed in the core in simple shapes or as a complex structure such as a heat exchanger that would form a section of the primary loop system. The latter case would probably require graph- ite-to-graphite and graphite-to-metal joints that have structural integrity and, to an extent, leak-tightness. A program has consequently been initiated fo investigate various tech- niques for producing Jjoints with graphite and to evaluate their usefulness for molten-salt-reactor service. A preliminary literature survey revealed a notable scarcity of Jjoining techniques applicable to high-temperature- reactor technology. Procedures for wetting graphite are relatively well known, but experience in producing reliable joints in complex assemblies is lacking. One serious problem is related to the gross differences be- tween the thermal-expansion coefficient of graphite and those of common brazing alloys and structural materials. Metals having a strong tendency to form carbides wet graphite readily when they are molten. Therefore, alloyé containing these carbide-forming metals appear promising as brazing alloys for joining graphite to itself and to metals. Zirconium and titanium appear especially attractive, since their thermal-expansion coefficients are relatively low (see Table 2.1.1) and since their melting points can be reduced significantly by only small additions of other elements. Several brazing alloys containing zirconium and titanium as the major constituents have been developed for the Jjoining of refractory metals. Many of these alloys appear promising .for the joining of graphite, but one alloy in particular (48% Ti—48% Zr—% Be) exhibits excellent flowability. An inverted T-Jjoint made from type AGOT graphite and brazed with this alloy is shown in Fig. 2.1.1. A photomicrograph illustrating the excellent fil- leting and bonding observed in this joint is shown in Fig. 2.1.2. Figure 2.1.3 illustrates the good flowability obtained by brazing, with this al- loy, a low-permeability graphite pipe to a type AGOT graphite sheet. Brazed joints'made with this alloy appear to possess good strength and seem promising for a variety of high-temperature-reactor applications. 33 Table 2.1.1. Thermal-Expansion Coefficients of Various Materials Thermal -Expansion . Material - Coefficient Temperature Range [hin. +in.~t« (°F)~1] (OF) Graphite - 238 32~1800 Titanium AL Room Zirconium | - 2.éb | Room | Molyb&enum _ | 2.7b o S Room‘ Inconel _ 10.5° 32-1400 Tyfie 430 stainless steel 6.3b o 321500 INOR-8 8.3 - 32-1800 %The Industrial Graphite Engineering Handbook, National Carbon Com- pany. ‘ ' S bMetals Handbook. . _éEngineering Properties of Inconel and Inconel X, International Nickel Company . : ‘ dDesign Data for INOR-8 Alloy, Mgtallfirgy Divisiofi, ORNL. For molten-salt servicé, however, the uses of this particular alloy are limited by its poor resistance to corrosion in thisimedium (see "Brazing Material Evaluation in Fuel 130," this section). | o . This development of other brazing-alloy systems is therefore being undertaken in an effort to find more compatible materials. Utilizing niobium, tantalum, and molybdenum as the pfimary.cafbide-forming elements, a number of alloy systems are being studied in detail to determine their applicability-for joining gréphite. Alloys in the Au-Ni-Ta sysfiem have indicated promising wetting characteristics, and several alloy compositions from the gold-rich corner of the ternary diagramflhave'been selected for - further study. The relatively high‘thermal—expansion coefficients of gold- rich alloys, however, may 1imi£ their use for joining graphite. Other al- loy systems which will be investigated include Ni-Mo, Ni-Nb, Ni—Nb-Mo, Ni-Nb-Ta, Pd-Ni-Ta, and Pd-Ni-Mo. S Preliminary experiments have also been undertaken to determine the feasibility éf electricfrésistance bonding of gréphite to gfaphite, using molybdenum as the braze metal. 8o far, experiments_havé been confined to 34 UNCLASSIFIED Y-30131 Fig. .2.1.]. Graphite T-Joint Brazed with 48% Ti-48% Zr-4% Be. Excellent wetting and flowability are evident. graphite rods of l/4-in. diameter, or less, due to the high-amperage re- quirements of the process. The Jjoint is prepared by placing a thin disk of molybdenum between the ends of two graphite rods. Evaluation of speci- mens prepared to date indicates that good structural bonds can be made by this technique. Metallographic examination has shown, however, that care- ful control will be necessary to minimize porosity and inconsistent bond- ing. Dynamic-Corrosion Studies Forced-Convection Loops Examination of INOR-8 Hot-Leg Inserts from Loop 9354-4. — A second hot-leg insert was removed from INOR-8 forced-convection loop 9354-4 after 10,000 hr of loop operation. As previously discussed,1 three inserts were MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 59. 35 36 UNCLASSIFIED ¥-20624 Fig. 2.1.2. Photomicrograph of Graphite T-Joint. As-polished. 12X. UNCLASSIFIED 2 ' . Y-30986 Fig. 2.1.3. Graphite Pipe-to-Sheet Joint Brazed with 48% Ti—48% Zr-4% Be. installed at the end of the hot-leg section of this loop to provide data on the weight changes of INOR-8 in contact with a beryllium-base fluoride fuel mixture. The schedule specified removal of one insert after 5000 hr of exposure, another after 10,000 hr, and the third after 15,000 hr. Operation of the loop began in July 1958 under the conditions: Salt mixture LiF-BeF,-UF, (62-37-1 mole %) Maximum fuel-metal interface 1300°F temperature Minimum salt temperature 1100°F AT 200°F Reynolds No. 1600 Flow rate 2 gpm Examination of the second insert, although not yet complete, indicates that an average weight loss of 2.0 mg/cm® (£3%) occurred along its 4-in. length. If uniform removal of surface metal is assumed, this weight loss corresponds to a loss in wall thickness of 0.09 mil. No detectable loss in wall thickness was noted in comparing micrometer measurements of the inserts taken before and after test. The first of the inserts, which was removed from the loop after 5000 hr, showed a weight loss of 1.8 mg/cm2 (xt2%), or a loss in wall thickness of 0.08 mil. The similarity of these values to the values obtained after 10,000 hr indicates that no significant weight loss occurred after the first 5000 hr of loop operation. The errors listed with these values are based on estimated limits of precision of the instruments used to measure and weigh the insert. They do not include errors resulting from handling of the insert, which are difficult to estimate but could be significant considering the small weight. change. Metallographic examination of transverse and longitudinal sections of the 5000-hr insert revealed light surface roughening, as shown in Figs. 2.1.4 and 2.1.5. Along the exposed edge of the tubing there was evidence of a band between 1 and 2 mils thick containing precipitates and very small grains. The start of a very thin film was also evident along the surface. Examination of transverse and longitudinal sections of the 10,000-hr insert showed the surface film to have increased in thickness and the band of fine-grained material to have grown in size (Figs. 2.1.6 and g 37 UNCLASSIFIED T-16820 - 010 - ® i @ pa 01 < ° T AN e e ¢ « - . . Olz . . - ol3 ; . » \- & & .Nj' » o e . : o ReTEY 014 4 4 - L Lo 8 1.015 -~ = o B hd -Q - . ¥ 2 > Y : o ‘ 2 - ‘.‘ . . 0 " 7 'C ; L- m% ' i A o - . ™ ) - Fig. 2.1.4. Transverse Section of 5000-hr Insert Removed from INOR-8 Pump Loop 9354-4, Etchant: 3 parts HCI, 2 parts H,0, 1 part 10% chromic acid. UNCLASSIFIED T-16821 Fig. 2.1.5. Longitudinal Section of 5000-hr Insert Removed from INOR-8 Pump Loop 9354-4. Etchant: 3 parts HCI, 2 parts H,0, 1 part 10% chromic acid. 38 UNCLASSIFIED T-18200 Fig. 2.1.6. Transverse Section of 10,000-hr Insert Removed from INOR-8 Pump Loop 9354-4, Etchant: 3 parts HCI, 2 parts H,0, 1 part 10% chromic acid. UNCLASSIFIED T-18201 - / - = /‘ el N ——— ‘L/. { x L e y A f ik ~\\ - Pr— > N {.0l4 . oy ' 3 H T y ¢ e Why . - / 4 ] & e % 2 /,’ - i ) 5 Do frem s / - ‘ L. I« 4 ; | i Fig. 2.1.7. Longitudinal Section of 10,000-hr Insert Removed from INOR-8 Pump Loop 9354-4, Etchant: 3 parts HCI, 2 parts H,0, 1 part 10% chromic acid. As previously reported,1 a surface layer or film similar in appearance to that found on the 10,000-hr insert has been noticed on specimens removed from most of the long-time INOR-8 thermal- and forced-convection tests. Samples of this film are being studied, but no results are available. The band of precipitates and fine grains found along the surfaces of the in- serts has not been observed in other tests of INOR-8. At present it is not known whether these fine grains represent a second phase or only re- crystallization of the matrix near the inside surfaces of the insert. There is reason to suspect the latter possibility, since a reaming operation was performed on the inserts prior to service and undoubtedly induced some cold work along the inside surfaces. A review of the recrystallization properties of INOR-8 indicates that recrystallization of the material can be initiated at 1300°F following prior cold work.? Grain growth of the recrystallized material would be extremely slow at this temperature, which would help to explain the extremely fine crystallites apparent. Examination of Three Inconel Pump Loops. — Chemical and metallographic results have been obtained for three Inconel pump loops which completed one year's operation. Operating conditions for these loops are shown in ‘Pable 2.1.2. Metallographic examination of loop 9344-1, which circulated salt 123 for 8760 hr at a maximum temperature of 1300°F, showed heavy attack in the form of intergranular and general voids along both heater legs, as shown in Fig. 2.1.8. In the first heater leg, attack ranged from 18 to 31 mils in depth and in the second leg, from 18 to 38 mils.., The unheated bend 2J. Spruell, Recrystallization of INOR-8, ORNL CF-57-11-119 (Nov. 25, 1957) . Table 2.1.2. Operating Conditions for Three Inconel Forced-Convection Loops yieed G- Minimum Fluid Flow Salt Metal Interface AT Reynolds L No.* Temperature Tempefature (°F) hese No. (oF) ( F) (gpm) 9344-1 123 1300 1100 200 2 3250 9344,-2 12 1200 1100 100 2D 8200 9377-3 131 1300 1100 200 2 3400 %¥123: NaF-BeF,-UF, (53-46-1 mole %) 12: NaF-LiF-KF (11.5-46.5-42 mole %) 131: LiF-BeF,-UF, (60-36-4 mole %) 40 UNCLASSIFIED T-17816 INCHES r_ Ll -— SRy '...' .'. P - 0.02 5'. >3- ~ .fi . N " o D PR e .V. b;.‘ - L 3 ) el & N kS ...,\)' o - '-.\I > ~b N e ¥ - ‘.. - R, o .. EY “ *® Vone v i =d . e - -~ A} i -~ "Bk o e .. N, - - N S ’ - - " - s = . S s 9 '- A V‘ CERR T - L » "wa . %l lo.03 - # ] |0?x Fig. 2.1.8. Specimen Removed from End of Second Heater Leg of Inconel Pump Loop 9344-1, Showing Maximum Attack Found. Loop was operated for 8760 hr with salt mixture NaF-BeF - UF, (53-46-1 mole %). Etchant: modified aqua regia. connecting the two legs revealed only a few pits, not over 1 mil, and the cold-leg sections showed only slight surface roughening and surface pit- ting. No cold-leg deposits were observed visually or metallographically. Chemical analyses of the fused-salt mixture before and after circu- lation in this loop are shown in Table 2.1.3. Note that an extremely large increase in chromium concentration occurred during the test, along with a reduction in uranium concentration. Optical examination of the salt dis- closed a brown phase consisting of about equal amounts of UO, and NaF-BeF,. An x-ray examination of this phase confirmed the presence of UO,. Thus the salt appears to have undergone extensive extraneous oxidation either before or during the test. Metallographic examination of loop 9344-2, which operated for 8760 hr at a maximum temperature of 1200°F with salt 12, showed moderate attack in the form of intergranular and general voids along both heater legs. In 41 the first heater leg, attack ranged from 1 to 3 mils in depth and in the second leg, from 4 to 8 mils. The cold-leg sections and the unheated bend connecting the two hot legs revealed light intergranular voids to 1 mil. The coolant salt in this loop was replaced three times because of operating difficulties. A somewhat higher rate of attack occurred in loop 9377-3, which cir- culated salt 131 for 8760 hr at a maximum temperature of 1300°F. Exami- nation of the heater sections revealed attack in the form of heavy inter- granular and general voids, ranging from 2 to 8 mils in the first heater leg and from 6 to 14 mils in the second leg. The unheated bend connecting the two legs showed only light surface roughening except in one region where a few isolated voids to a maximum depth of 11 mils were found. The cold-leg sections revealed only a few pits, all less than 1 mil, Chemical analyses of salt 131 before and after circulation in loop 9377-3 are shown in Table 2.1.4. Except for the increase in chromium con- centration, these analyses show little difference. The present status of all forced-circulation loop tests now in prog- ress is given in Sec 1.1 of this report. Table 2.1.3. Analysis of Salt 123 Before and After Circulation in Pump Loop 9344-1 Major Constituents Minor Constituents Sample Taken (wt %) (ppm) U Be Ni cr Fe Before test 3.40 8.16 140 425 130 After test 258 8.25 175 4670 205 Table 2.1.4. Analysis of Salt 131 Before and After Circulation in Pump Loop 9377-3 Major Constituents Minor Constituents Sample Taken (wt %) (ppm) U Be Ni Cr Fe Before test 19.6 ¢ 4o e 145 345 200 After test 19.5 7.09 1S 860 275 42 General Corrosion Studies Permeation of Graphite by Molten Salts The permeation of various grades of graphite by molten fluoride salts is being investigated in tests® at 1300°F (704°C) for 100 hr at 150 psig, using the weight increase of the test specimen to calculate the volume of graphite that is permeated by the salt. Table 2.1.5, a summary of results, lists 15 grades of graphite in order of resistance to permeation. These grades were manufactured as low-permeability graphite, with the exception of CT-158, CCN, CEYB, R-0013, AGOT, and TSF. The permeation values for the pipe material are considered high, because no effort was made to seal the ends of the pipe specimens to prevent the exposure of the more contin- uous and accessible voids produced in the axial direction as a result of fabrication. Pipe specimens with sealed ends are being made to be tested for permeation, since values for them should be more applicable to reactor designs. Most of the testing was done with LiF-BeF,-ThF,-UF, (67-18.5-14-0.5 mole %); however, NaF-ZrF,-UF, (56-40-4 mole %) and LiF-BeF,-UF, (62-37-1 mole %) were also used. Under the test conditions the extent of permeation by the three salt mixtures was essentially the same; other test conditions will be investigated. ©No wetting of the graphite by the salts was ob- served. ' Salt permeation studies with additional grades of graphite, ihcluding studies to show the effect of exposure time and graphite shape, were begun. An effort 1s being made to preclude the fuel pickup by initially fill- ing accessible void spaces in the graphite with bismuth, thus allowing no space for the salt. Approximately 60% of the accessible* void spaces of ACOT graphite specimens were filled with bismuth when held at 1022°F for 4 hr at 125 psig. Tests were begun to determine the extent to which these impregnated specimens will take up salt in standard permeation tests. 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 May 1, 1958 to Dec. 31, 1958, ORNL CF-59-1-4, p 1-20. “The total accessible void volume is reported to be 21.7% of the bulk volume of AGOT graphite; W. P. Eatherly, M. Janes, and R. L. Mansfield, "Physical Properties of Graphite Materials for Special Nuclear Applica- tions," Second U.N. Intern. Conf. Peaceful Uses Atomic Energy, Geneva, 1958, paper A/Conf 15/P/708, p 4 (June 1958). 43 Table 2.1.5. Permeation of Varluub Grades of Graphlte ' by Molten Fluoride Salts All percentages are averages of three with the exception of. those with numerical superscripts, which denote the number of values averaged Percentage of Bulk Volume Gréphite Graphitg Bulk - of Graphite Permeated Grade : Density by Fluoride Salts (g/cc) Fuel 30 Fuel 130 BULT-14-.5U GT-123-82 1.92 0.2 R-4 1.93 1.6 R-0009 1.92 1.8 cEy-82° 1.87 1.9 R-0009-RG 1:90 2.5 cT-158° 1.77 3.5 186 1.8 | 4.61) 3.g6) ceN 1.92 AP / 4.2 CEYB 1.79 4.3 S=d 1.85 - 4.8 R-0013 1.87 6.0 ATJ-82-RG 1.83 6.4 CS-82-RG 1.80 | g.1 AGOT 1.68 4.8 148 4.2 TSF 1.67 13.2(2) 14.4 ®30: NeF- -7rF,-UF, (56-40-4 mole %) = . 130: LiF-BeF,-UF, (62-37-1.mole %) BULT-14-.5U: LiF-BeF,-ThF,-UF, (67-18.5- 14 0.5 mole %) bPipe, nominal dimensions, in.: 1-1/4 OD x 7/8 ID x 1/2 long. “Pipe, nominal dimensions, in.: 25/64 OD x 3/16 ID x 1-1/2 long. Removal of Oxide Contaminants from Graphite Previous tests have shown that exposing graphite to fuel 130 (LiF- BeF,-UF,, 62-37-1 mole %) at 1300°F (704°C) will result in the precipita- tion of UO,, and that this reaction is essentially complete within 5 hr when the fuel is in vapor-phase contact or ih direct contact with the graphite. This indicated that oxide contamination might be removed from the gréphite by vapor-phase or direct-contact gettering. Results that 44 indicated partial success in two such tests, with fuel 130 as the gettering agent, have been reported.’ A direct-gettering test, which has been completed, was apparently successful in removing oxide contaminants from the graphite. 1In the test a charge of fuel 130 was in direct contact with an AGOT graphite crucible for 20 hr at 1300°F (704°C) in a vacuum. The ratio of the bulk volume of the graphite to the volume of the salt was 27:1 (an arbitrary standard selected for such tests). Radiographic examinations at room temperature indicated that a normal quantity of UO, had precipitated. The gettering charge of fuel 130 was removed and replaced by a fresh charge of fuel 130, which was held an additional 20 hr at 1300°F. Radiographic examination showed no precipitation, indicating that the preceding gettering operation was successful. Since the cost of fuel 130 would make it an uneconomical gettering agent to use on a full-scale reactor, cheaper gettering agents are also being investigated., A hydrogen test is in progress. Tests similar to those described above are being made with LiF-BeF,- ThF,-UF, (67-18.5-14-0.5 mole %). Three such tests with AGOT graphite have continued for 500 hr, and no precipitation has been detected in the galt by radiographic examinations. Compatibility of INOR-8 and Graphite in Direct Contact The tendency for INOR-8 to carburize when in direct contact with graphite is being investigated as part of the over-all INOR-8—graphite— salt compatibility program. TSF-grade graphite and INOR-8 specimens were held in contact at 1300°F at 1000 psi. Mbdified stress-rupture apparatus wvas used, and the tests were held in salt environments to simulate reactor conditions. Metallographic examination of specimens held for 700 hr in NaF-KF-LiF-UF, (11.2-41-45.3-2.5 mole %) revealed no evidence of carburi- zation. Tests for 3400 hr in NaF-ZrF,-UF, (50-46-4 mole %) have been com- pleted, and the test specimens are now being examined. Brazing-Malerial Evaluation in Fuel 130 Since a barren-salt-to-fuel-salt heat exchanger is being considered for use in the Molten-Salt Reactor, the corrosion resistance of brazing SMSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 62. 45 alloys which may be useful in the fabrication of metal-to-metal joints in such a heat exchanger is being evaluated. The method used to subject brazing alloys to the corrosive environment in the hot leg of a thermal- convection loop has been described earlier,® and corrosion test results on a series of brazing alloys subjected to fuel 130 (LiF-BeF,-UF,, 62-37-1 mole %) for 1000 hr at a 1300°F hot-leg temperature in such a loop have been reported.’ The results of a 5000-hr test made under the same conditions are out- lined in Table 2.1.6. Copper, gold-nickel, and the Coast Metals alloys showed good corrosion resistance. A depleted region to a depth of 3 mils was observed along each of the Coast Metals alloy fillets after the test. Results of earlier corrosion tests on these alloys in NaF-ZrF,-base fuels and in liquid metals indicated that this depletion was due to the leaching of the minor constituents boron and silicon from the.alloy by the bath, and it appears that boron and silicon are also being leached by fuel 130.. This depletion has no detrimental effect on the alloy, since a nickel-rich, corrosion-resistant matrix is left. A 10,000-hr thermal-convection-loop test on Inconel and INOR-8 lap joints brazed with the five brazing ma- terials listed in Table 2.1.6 is in progress. Some refractory-metal-base alloys, being developed by the Welding and Brazing Group for possible application in joining graphite to graphite, were given a static corrosion test in fuel 130 for 100 hr at 1300°F. Nickel, because of its comparative inertness to fluorides, was used as the test- container material. Test results, listed in Table 2.1.7, showed large weight losses for each alloy. In order to determine the alloying element or elements responsible for the heavy attack, the following static corro- sion tests in fuel 130 were conducted at 1300°F: (1) titanium was tested in containers of titanium, nickel, and INOR-8, (2) zirconium was tested in containers of zirconium, nickel, and INOR-8, and (3) beryllium was tested in a nickel container. Large weight losses (Table 2.1.8) were ob- served for specimens tested in nickel or nickel-base containers, while small weight losses were found when specimen and container were of the same metal. Substituting an INOR-8 (containing 70% Ni) container for one ®MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 62. "Met. Ann. Prog. Rep. Sept. 1, 1959, ORNL-2839, p 166-67. 46 Ly Table 2.1.6. Results of Metallographic Examinations of Brézing Materials Tested in Thermal-Convection Loop Circulating Fuel 130 (LiF-BeF,-UF,, 62-37-1 Mole %) Test conditions: Hot-leg temperature, 1300°F Cold-leg temperature, 1100°F Time, 5000 hr Metallographic Results: Alloy Brazed to Alloy Composition (%) Tnconel INOR-8 Coast Metals No. 52 89 Ni-5 Si—4 B2 Fe No attack No attack Coast Metals No. 53 8l Ni—-8 Cr—4 B4 Si—3 Fe No attack No attack General Electric 70 Ni—20 Cr—10 si Heavy attack of 16 3-mil attack No. 81 " : mils Gold-nickel 82 Au—18 Ni No attack | No attack Copper 100 Cu No attack; diffusion No attack; diffusion voids 2 mils deep voids 2 mils deep on fillet ' Table 2.1.7. Results of Static Corrosion Tests on Refractory-Metal-Base Brazing Alloys in Fuel 130 in Nickel Containers Test conditions: 1300°F, 100 hr Alloy Composition (%) . Weight Change (%) 48 Ti—48 Zr—4 Be -4, 2 95 Ti—5 Be ' -6.5 95 Ti—5 Be -9.8 Table 2.1.8. Results of Static Corrosion Tests’on’Refractory Metals in Fuel 130 in Containers of Several Materials Test conditions: 1300°F, 100 hr Weight Change (%) When Tested in: Material Nickel INOR-8 Titanium Zirconium Ti ~12.1 . 276 ~0.71 7 -11.7 4.2 ~0.08 Be Excessive, portion of sample dissolved of pure nickel resulted in a weight-loss decrease on the titanium and zir- conium specimens. Metallographic examifiation and sPéctrographic results on the nickel capsules revealed surface layers containing titanium, zir- conium, or beryllium, corresponding to the major component of the brazing alloy or metal tested. Results of the tests indicate that the usefulness of titanium-, zirconium-, or beryllium-containing brazing alloys in fuel 130 would be limited in systems constructed of nickel or nickel-base al- loys. 48 Phase Equilibrium Studies The System BeF,-ThF,-UF, The most favorable fuels for breeding with molten fluorides are ob- tained from the LiF-BeF,-ThF,-UF, quaternary system. Of the ten binary and ternary systems which are incorporated as limiting cases in this qua- ternary system, bnly the system BeF,-ThF,-UF, has not been described pre- viously. Spot checks have shown that only two primary phases, those of BeF, and of the ThF,-UF, solid solution, are involved in the BeF,-ThF,- UF, phase diagram. As in the ThF,-UF, binary system, no minimum exists in the THF4-UF4 solid-solution liquidus surface in the ternary system. Isotherms for such of the liquidus surface in the ternary system can be - approximated by straight lines between the bounding binary systems BeF,- ThF, and BcF,-UF,. Thc System BeF,-ThF, Some of the information on previously reportedl preliminary phase diagrams for BeF,-ThF, was uncertain because BeF,-ThF, mixtures containing more than 75 mole % BeF, are too viscous to reach equilibrium, even in periods as long as three weeks. This difficulty is greatly alleviatéd by small additions of NaF or LiF, and the BeF,-ThF, binary diagram has now been established with the aid of short extrapolations from work on the ternary systems. Figure 2.2.1, showing a eutectic at 527°C and 2 mole % Th¥,, is regarded as a final diagram. The System NaF-BeF,-ThF, Continuing studies of the phase equilibria in the system NaF-BeF,- ThF, have established that this system involves 12 invariant equilibrium points. Of these, ten have been established to within 1 mole % and 5°C; these ten are listed in Table 2.2.1. Mollen-Fluoride Solvents A subsidiary interest, as far as breeder fuels are concerned, is at- tached to a recent search for alternate solvents for use in the fluoride- volatility dissolution process, in which a high solubility of ZrF, is a 1MSR Quar. Prog. Rep. Oct. 31, 1957, ORNL-2431, p 36. 49 UNCLASSIFIED ORNL-LR-DWG 29667R 1400 - L~ / 1000 ] / TEMPERATURE (°C) BeF, 10 20 30 40 50 60 70 80 90 ThR, ThF, (mole %) Fig. 2.2.1. The System BeF,-ThF,. requisite. Three mixtures, NaF-ZrF,, NaF-LiF, and NaF-LiF-ZrF,, have been used; of these, NaF-LiF has the largest capacity for ZrF,, but in this case liquidus temperatures approaching 600°C are encountered at 25 mole % ZrF, . Exploratory investigations of the phase equilibria in the system NaF- BeF,-ZrF, have shown that NaF-BeF, (69-31 mole %) affords a solvent for ZrF, which will accommodate as much as 52 mole % of ZrF, in solution at temperatures below 550°C and that no liquidus temperatures higher than 570°C will be encountered on the composition path between NaF-BeF, (69-31 mole %) and NaF-BeF,-ZrF, (33-15-52 mole %). The Compound NaF -BeF',-3UF, Studies by Barton et 2&'2 on the system NaF-UF, and more recent at- tempts in this laboratory to synthesize a compound NaF-4UF, by thermal- gradient quenches of a substance with the composition 20 mole % NaF-80 mole % UF, have failed to confirm the existence of a compound with this composition, which was reported by Eichelberger et gl.3 in their investi- gation of the system NaF-BeF,-UF,. Moreover, studies at this laboratory in the system NaF-BeF,-ThF, have shown that a primary-phase field exists corresponding to the "NaF-.4UF," primary-phase field in the system NaF-BeF,- UF, and that the primary phase in this case is NaF-BeF,-3ThF,. This com- pound is analogous in optical properties to the so-called "NaF-4UF," as ?C. J. Barton et al., J. Am. Ceram. Soc. 41(2), 63 (1958). 3J. F. Eichelberger et al., p 110 in Phase Diagrams of Nuclear Reactor Materials, R. E. Thoma, ed., ORNL-2548 (Nov. 6, 1959). 50 89 Table z.2.1. Invariant Equilibria in the System NaF-BeF,-ThF, Composition of Liquid (mole %) Invariant Type of Temperature Eaquilibrium Solid Phases Present NaF BeF, ThF (°c) < 2 9% - 2 527 ? ThF,, BeF,, and NaF.BeF,-3ThF, 43 55 2 . 365 ‘Eutectic NeF -BeF,, BeF,, and NaF-BeF,-3ThF, 49 49 2 - 375 7 NaF-BeF,, NeF-2ThF,, and | NaF - BeF, - 3ThF,, : 47 51 2 375 . DPeritectic NaF-BeF,, NaF:2ThF,, and | | | B-NaF -ThF, 56 42 2 320 Butectic NaF -BeF,, 2NaF-BeF,, and ' B-NaF -ThF,, 57 41 2 415 - Peritectic B,-2NaF:ThF,, 2NaF.BeF,, and B-NaF . ThF, . 72 22 6 510 Eutectic B,-2NaF:ThF,, 2NaF-BeF,, and NaF 76 11 A 13 540 Peritectic B,-2NaF.ThF,, B-4NaF.ThF,, and NaF 62 2 36 683 Peritectic B,-2NaF .ThF,, 3NaF.2ThF,, and B-NaF - ThF, 42 31 27 740 Peritectic NaF - 2ThF,, ThF,, and NaF - BeF, + 3ThF, reported by the Mound Laboratory. Since a thermal-gradient quench in the system NaF-BeF,-UF, at 35-55-10 mole % has produced a primary phase which is isostructural, by x-ray diffraction analysis, with the compound NaF - BeF,-3ThF,, it seems probable that the compound reported by the Mound Lab- oratory is actually NaF-BeF,-3UF,. Gas Solubilities in Molten Fluorides Solubility of HF in Fuel Solvents As previously repo:c"ced,l*’5 the solubility of HF has been measured over wide composition ranges in both BeF,- and ZrF,-base solvents. In each type of system, compositions rich in alkali fluorides showed marked increases in the solubility and in the heat of solution of HF. The tran- sition to higher values appeared to occur as the alkali fluoride content exceeded the stoichiometry for complete complexing of the Be'' as BeF, and the Zr“" as ZrF, , thereby providing "free" fluorides for forming HF Several schemes were examined in an attempt to correlate the solu- bility of HF with the composition of the solvent. One which works well is the use of a composition scale in terms of equivalent per cent free NaF as shown in Fig. 2.2.2. Compositions corresponding to Na,BeF, and Na,ZrF. are considered to contain 0% excess NaF. Other compositions con- tain an excess or a deficiency of NaF; equivalents are calculated on the basis of 2 for Na,BeF, and 3 for Na3ZrF, in obtaining equivalent per cent compositions according to this scheme. Figures 2.2.3, 2.2.4, and 2.2.5, applying to different temperatures, demonstrate the good correlation achieved for solubilities in NaF-ZrF, and NaF-BeF, compositions. Solubilities are expressed in terms of the Henry's law constant. The curves in the figures are logarithmic and give straight- line plots on semilog paper. The effects of both temperature and compo- sition are accommodated in the expression 8400 + 3263 RC 15.24 + 0.909 RC log K = RT - R ; “MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 85. MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-2474, p 93. 52 UNCLASSIFIED ORNL—LR—DWG 449497 100 . // . / NOaBqu NoyZrF; =] EQUIVALENT PER CENT FREE NaF o bl 4+ 0 20 40 60 80 100 NoF (mole %) Fig. 2.2.2. Correlation of Free NaF with Mole Fraction of NaF in Mixtures of NaF- BeF, and NaF-ZrF , (Based on NozBeF4 and Na,ZrF,). UNCLASSIFIED (x107%) ORNL — LR— DWG 44498 180 I | } [ 160 |— U NoF —BeF, / © NoF-ZrFy, A CALCULATED. 140 [— g @ EXTRAPOLATED FROM MEASURE— iy MENTS TAKEN AT HIGHER g — TEMPERATURE £120 of 5 100 (8] K N / T 80 5 3 (=] LE, 60 x . n 0 [ -~ - - B VR e ,/D/ No, BeF, 20 o= 257 ’C/o-/g/ No3 2rF, (o] -{100 -80 -—60 -4n =20 0 20 40 60 EQUIVALENT PER CENT FREE NoF " Fig. 2.2.3. Composition Dependence of Henry’s Law Constants for the Solubility of HF in NaF-BeF, and NaF-ZrF, Mixtures at 600°C. _g UNCLASSIFIED (x10 =) ORNI. — 1 R~ DWG 44499 140 120 0 NoF— BeF, TH00 [ O NaF-2ZrFg 5 & CALCULATED 8/ z 8 @ EXTRAPOLATED FROM MEASUREMENTS £ AT HIGHER TEMPERATURE 5 80 S O o I 60 / o k4 £ T 40 /{( 20 O/y/ Na,BeF, 2Befy 5._—0--"%/ No3ZrFy 0 -80 -60 —40 -20 0 20 40 EQUIVALENT PER CENT FREE NoF Fig. 2.2.4. Composition Dependence of Henry's Law Constants for the Solubility of HF in NaF-BeF2 and NaF-ZrF4 Mixtures at 700°C. UNCLASSIFIED (x10°%) ORNL— LR— DWG 4 4500 140 120 a} NoF—Ber =400 |- O NoF—ZrF, 3 &6 CALCULATED z 5 80 (8] ° ~ ('8 I 60 (=3 @ [} [ £ X 40 e oJ// Na,8eF, _o____o. N032rF7 0 -80 -60 -40 —20 0 20 40 EQUIVALENT PER CENT FREE NaF Fig. 2.2.5. Composition Dependence of Henry's Law Constants for the Solubility of HF in NaF-BeF, and NaF-ZrF, Mixtures at 800°C. 53 K = Henry's law constant, moles of HF/(liter of melt) (atm), Q Il concentration of effective free fluoride, equivalent per cent, H Il temperature, °K, R = gas constant. Calculated values'from this equation, denoted by triangles in Figs. 2.2.3 to 2.2.5, are within the experimental precision of the measurements. The arbitrary concentration scale used here can be related to a more conventional concentration concept by considering the system NaF-BeF, to be composed of the systems BeF,-Na,BeF, and Na,BeF,-NaF; then the postu- lated free-fluoride model corresponds to mole % BeF, in the former system and to equivalent per cent NaF in the latter system. Similar relations hold for the NaF-ZrF, system. Fission-Product Behavior Exchange Reactions in Molten Fluorides Experiments with exchange reactions for reprocessing molten-salt re- actor fuels have continued.® Important characteristics for potential solid exchange reactants are insolubility in molten fluorides and chemical in- ertness toward reactor fuel constituents. Studies of oxide reactions in molten fluorides have shown that the rare-earth oxides are sufficiently insoluble to qualify as possible ex- change agents. The exchange of Ce0O, with cerium ion in the solvent LiF-NaF (60-40 mole %) and the effect of subsequently added CeF; on the equilibrium Ce¥F5 (L) + CeO,(s) = Ce*0, + CeF, are shown in Fig. 2.2.6. Similarly, in the same solvent, the removal of lanthanum-tracer activity with Laz03 and the effect of subsequent additions of CeF3 are shown in Fig. 2.2.7. These experiments, while not carried out in realistic fuels, nevertheless illustrate the effectiveness of exchange proceéses in molten fluorides. The exchange of lanthanum with cerium is interesting as a possible method for replacing nuclear poisons, such as samarium, gadolinium, and europium, with more innocuous rare earths such as cerium. Unfortunately, as shown in Fig. 2.2.8, Ce0O, reacts readily with ®MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 8l. 54 UF, in the solvent LiF-NaF (60-40 mole %). Hence, reprocessing with rare- earth oxides would pfobably be limited to reactor fuels from which uranium had been removed by prior treatment. Figure 2.2.9 illustrates the unfavor- able effect of ThF, and BeF, on the removal of tracer-cerium activity from solution by exchange with CeO,. Experiments to investigate the exchange properties of carbides are also under way. 'The results of' two experiments to study the removal of hafnium activity with hafnium carbide are shown in Fig. 2.2.10. These re- sults illustrate the unfavorable effect of UF, on the exchange reaction in the solvent LiF-BeF, (63-37 mole %). Attempts to remove barium activity with Cr,05; and BeO from the solvent LiF-BeF,-ThF, (67-18-15 mole %) were not successful. Similarly an attempt to remove cerium activity from the same solvent with BeO was unsuccessful. The addition of A1203vto LiF-BeF,-ThF, (67-18-15 mole %) containing cerium activity caused the precipitation of ThF, without any detectable removal of cerium activity from solution, These‘experiments show that exchange reactions for reprocessing molten- salt reactor fuels are in many cases adversely affected by essential con- - stituents in the fuels. The presence of thorium and beryllium, for example, interferes with exchange reactions involving CeO, and HfC. Further searches for suitable exchange agents will include the rare-earth carbides. Chemistry of Corrosion Processes Self-Diffusion Coefficients for Chromium in Nickel Alloys A preponderance of the evidence accumulated over many years indicates that corrosion of nickel-chromium alloys by molten fluorides under aniso- thermal conditions should reach a steady-state rate controlled by diffusion of chromium in the metal in the colder part of the system. For this reason the value of the diffusion coefficient for chromium has been of paramount interest.” Data from several diffusion experiments, employing cr’l as a tracer, have been partially digested. The alloy specimens studied may be grouped as follows, according to the treatment they received prior to the experi- ments: Inconel annealed at 800°C for 8 to 12 hr, Inconel annealed at 7MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 99. 55 UNCLASSIFIED ORNL—LR—DWG 44501 100 £ 80 (&) £ 2 <{ 2 60 | ¥ 2 > a0 5 ) < o /o/ o A & 20 \ / o) :hf o] 4 8 12 16 20 24 28 ADDED COMPOUND {(g) Fig. 2.2.6. Exchange of Cerium Activity with CeO, and CeF, in LiF-NaF (60-40 Mole %) at 750°C. UNCLASSIFIED ORNL~LR-—DWG 44502 100 20 80 70 \ RADICACTIVITY REMAINING (%) ; e \/ 0 1 2 3 4 5 ADDED COMPOUND (g} 40 Fig. 2.2.7. Exchange of Tracer-Lanthanum Activity with La,0, and CeF, in LiF-NaF (60-40 Mole %) at 750°C. 56 UNCLASSIFIED ORNL—LR—DWG 44503 500 400 [~ Ce0, ADDED S £ 300 [=] 5 CERIUM (o] ® O u o | 2 200 w = 100 A o—e—¢ URANIUM . 0 0 5 10 i5 20 25 TIME (hr) Fig. 2.2.8. Reaction of UF, with CeO, in LiF-NaF (60-40 Mole %) ot 750°C. UNCLASSIFIED ORNL - LR~ DWG 44504 \, J’\ g g N N <1 = 60 = r>—- \ > = > 40 = [ 2 o A LiF—8eF, (63-37mole %) 3 20 O LiF—BeF, —Thf, (67=18--15 mole %) __| o ® LifF— NoF (60~ 40 mole %) N N I [¢] 2 4q 6 8 10 12 14 CeO, ADDED (g) Fig. 2.2.9. Removal of Tracer-Cerium Activity by Exchange with CeO, in Molten Fluoride Systems. UNCLASSIFIED ORNL-LR-DWG 44505 100 ®._0 \" 90 S :.'; (o) - g ® WITH URANIUM (7 wt%) § 80 O O WITHOUT URANIUM < . w o ). = > 2 4 o 70 g PN 2 \ < @ I: \ h\‘k 60 50 o) 5 10 5 20 25 TIME (hr) Fig. 2.2.10. Effect of UF, on the Removal of Hf'8" Activity with HfC (5 wt %) in LiF-BeF, (63-37 Mole %) at 600°C. 1150°C for 2 to 4 hr, and INOR-8 annealed at 800°C for 8 to 12 hr. The results given here are over-all coefficients obtained from a comparison between the total amount of Cr®' present in the metal and the amount in the contacting salt after a given exposure time. A description of the ex- rerimental procedures was given in previous reports.7'8 The data presently under discussion are not based on concentration profiles within the alloy. Companion experiments of the latter type will be discussed in a.subsequent report; the general trends discussed below are reflected by the available profile data. The results for Inconel specimens are presented as curves 5 and 7 on Fig. 2.2.11. Only two'experimental points for Inconel specimens are shown; however, curve 5 was taken from 19 individual points (34 experiments) at various temperatures, and curve 7 from 16 points. Curves 1 through 4, from 8MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p- 95. 57 UNCLASSIFIED ORNL~-LR—DOWG 44506 TEMPERATURE (°C) ' 1000 900 800 750 700 650 600 T “§$§§§§§EE:;: T 1 I — T N §%§E§Q§§\ ~ —1 : 2 \\:::;f§\ | | ~E\.Q\ S -3 B @ . Hy NG \9\ ° N . E - (§\®\ < Q 14— DI9.7%Cr, 4.3%Fe ,1.9%TI — ™) . U? @ 20.4 % Cr, 2.6% Ti SN d L . \ % @ 19.8 % Cr \ o ~ - @ PURE NICKEL \\H4 =< \ \ (® INCONEL ANNEALED AT 1150°C ‘ O\ \ —16 —— (® INOR-8 ANNEALED AT 800°C — < @ INCONEL ANNEALED AT 800°C . \\ 7 ~-0--DATA POINTS FOR INOR-8 N : ® "H" INCONEL SPECIMENS \ . \ . \ 18 | 1 L [ 1 | T 6 7 8 9 10 " (x10™% 1/7 (k") Fig; 2.2.11. Diffusion Coefficients qf Cr3' in Various Nickel-Base Alloys. a high-temperature investigation conducted on similar alloys by Russian % are shown for comparison. investigators, A It has been found that grain size, and thérefore the annealing history prior to the diffusion experiments, has a marked effect on the diffusion coefficients. For example, photomicrdgraphs'Of Inconel specimens Hy and H, revealed thatuthe average grain size was much greater ih H, than-in H,. The effect of annealing conditions is easily seen by comparing curves 5 and 7 of Fig. 2.2.11 and also the individual values for Hy and H,. An ad- ‘ditional anneéling effect is manifeét in the reproducibility of diffusion behavior. The data of curve 7 (specimens annealed at 800°C) tend to show a higher degree of scattering than those of curve 5.(specimens annealed at‘ll5060), which suggests that an 800°C annealing temperature fails to eliminate grain variations which might exist between specimens. °p. L. Gruzin and G. B. Fedorov, Doklady Akad. Nauk S.S.S.R. 105, 264 (1955). ' | 58 The over-all diffusion coefficients obtained for INOR-8 specimens are shown as points on curve 6 of Fig. 2.2.11. At temperatures greater than 700°C, the INOR-8 coefficients are of the same magnitude as the Inconel coefficients. This means that the rates of chromium diffusion in INOR-8 are approximately 7/15 of the rates in Inconel, at least at higher temper- atures. The ratio 7/15 corresponds to the ratio of the chromium concen- trations in the alloys. . At temperatures below 700°C, the INOR-8 coefficients appear to fall considerably below the Inconel coefficients. This sudden drop might afford a partial explanation of the encouragingly low molten-salt corrosion rates exhibited by INOR-8 within the 500 to 700°C temperature range. ' ' Another point of interest is that observed corrosion results fre- quently represent considerably faster rates than predicted onlthe basis of diffusion in the cold zone as the rate-controlling step if the diffu- sion coefficients from Fig. 2.2.11 are used. This suggests that there is frequently an extended period of uncompensated oxidation, largely extra- neous, which precedes the establishment of a steady-state regime. An ex- treme example of this is described in the next section. Chemical Analyses of Corrosion Test Loops Periodic sampling to follow the corrosion behavior in pumped loop systems has continued.® An INOR-8 loop, containing LiF-BeF,-ThF,~-UF, (62-36.5-1-0.5 mole %) has been operating for 9300 hr with a hot-zone tem- perature of 1300°F and a cold-zone temperature of 1100°F. As shown in Fig. 2.2.12, the chromium content reached a steady-state value of 500 ppm after about 1000 hr. Sampling was temporarily discontinued after 5000 hr in order to maintain a sufficient fuel volume in the loop; the scheduled test period for the loop is 15,000 hr. This loop appears to be following the predicted behavior. By contrast, an Inconel loop, operating under the same cdnditions, has shown a steady increase in chromium concentration in the fuel as il- lustrated in Fig. 2.2.13. After about 2000 hr some additional fuel was added to compensate for the depletion due to sample removals. A slight dilution of the chromium ion concentration was reflected in the analyses. At about 3000 hr a motor failure necessitated draining the loop into the 10MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p &8. 59 - ' ‘ UNCLASSIFIED : ORNL-LR-DWG 44507 - 800 600 ' ° 9 ° € o—¢© ° ¥ g ././ ® ".‘ a e ® "5 § 400 2 . o 3 & k X . U , 200 ‘ LOOP: MSRP-12 (INOR-8) : SALT : LiF-BeF,-ThF,~UF, ; (62-36.5-1-0.5 mole %) : 0 . 0 : 800 1600 2400 - 3200 4000 4800 5600 g TIME (hr} Fig. 2.2.12.. Chromium Concentration Due to Corrosion of INOR-8. : t Y,‘{:‘ v UNCLASSIFIED ORNL- LR-DWG 44508 5000 4000 / o | : { PUMP FAILURE / ' ' 3000 , — e // " NEW SALT ADDED CHROMIUM {ppm) LOOP: 9377-5 (INCONEL) SALT: LiF-'BeFZ—ThF‘;—UFQ (62 -36.5~1-0.5 mole %) 2000 | -/ - 3 * 1000 . 0 800 1600 2400- - 3200 4000 4800 5600 6400 7200 ) ) TIME (hr) Fig. 2.2.13. Corrosion of Inconel Due to an Unidentified Extraneous Source, 60 sump; this resulted in another dilution because, contrary to intended op- erating procedures, a }arge residue of uncirculated salt had remained in the sump since the oriéinal filling. As the chromium concentration has increased to 4730 ppm in 7000 hr, the uranium concentration has dropped from 2.8°to 1.5 wt %. This is the behavior predicted for continuing ex- posure to an extraneous oxidant, such as air, but a rigorous examination of the system did not disclose any evidence of a leak or other means of . exposure to air. In another loop, of INOR-8, a salt mixture with a higher concentration of uranium, LiF-BeF,-UF, (70-10-20 mole %), hasfibeen pumped under similar conditions for 5500 hr. The analyses suggest, as shown in Fig. 2.2.14, that the expected steady-state chromium éoncentration has not yet been reached; this is somewhat surprising. Three additional test loops, scheduled for periodic sampling, were _placed in operation during the past quarter. Solubility of CrF, in LiF-BeF, (62-38 Mole %) The solubility of CrF, in LiF-BeF, (62-38 mole %) was found to be greater than 7.0 mole % at 483°C. Uncombined CrF, was found to be pre- cipitated from the melt, containing 7.0 mole % CrF,, at 369 to 422°C at the phase boundary between CrF, and 2LiF-BeF, primary-phase fields. Besides indicating the extremely high concentration of CrF, required for deposition on heat exchanger surfaces, this information should lead UNCLASSIFIED ORNL-LR-DWG 44509 600 ® LOOP: MSRP - 13 (INOR-8) ’/,/’///////’ SALT : LiF -BeF, - UF, P 400 { 70-10 — 20 mole %) / € Q [ — - = /. = 3 ot g ° / ® () 0] 800 1600 2400 3200 ’ 4000 TIME (hr) Figb. 2.2.14. Corrosion of INOR-8 by Fuel Containing a High Concentration of UF,. 61 .to an experimental determination of the free energy of formation for pure CrF, at high temperatures (800°C) when combined with the equilibrium con- stants now‘being obtained for the reaction CrF, + H, = Cr® 4 2HF Equilibrium Measurements on the Reduction of CrF,-and FeF2 by H, in LiF- BeF, . (62-38 Mole %) High-temperature experimental equilibrium quotients,t?! Kg, for. the reduction of CrF, and FeF, dissolved in LiF-BeF, (62-38 mole %) are being obtained from the following reactions: <] CrFs + Hz%w0 + 2HF (o] FeF, +1H2飥L£éFe° + 2HF A comparison of recent experimental equilibrium quotiefits obtained for the above reactions with those for the same reactions in NaF-ZrF, (53- 47 mole %) is shown in Table 2.2.2. | Thé saturating phase for CrF, in LiF-BeF, (62-38 mole %) is pure CrF,; a complex compound precipitated on saturating NaF-ZrF, (53-47 mole %) with 11MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 85. Table 2.2.2. Effect of Solvent on the Equilibrium Quotients (Atmospheres/Mole Fraction), K, = MOPI‘?‘IF/XMFzPH2 K . Temperature x Compound (°c) . o In LiF-BeF, In NaF-ZrF, X 1074 X 1074 FeF, 700 ~4500° 6300 CrF, 750 1.8° 2.21 800 4.1° 6.77 ®ANP Quar. Prog. Rep. Mar. 10, 1956, ORNL-2061, p 84. P>.88 mole % FeF,. ©0.53 mole % CrF,. .62 T meniia T L Tiel. ol CrF, or FeF,. Previous studies have indicated that the saturating phase for FeF, in LiF-BeF, (62-38 mole %) is pure FeF,; this is still being veri- fied. The smaller equilibrium quotients in LiF-BeF, result from stronger complexing in a more basic solvent. Hygroscopic Behavior and Dehydration of Breeder Fuels It has been generally recognized that breeder fuels based on LiF-BeF, as solvent are very hygroscopic and prone to hydrolysis, but no specific experimental evaluation of the problem had been made. Some salts can be dried without extensive hydrolysis, and drying at low temperatures is fre- quently helpful.12 Trials were made to explore the prospects for removing water from breeder fuels at low temperatures. Two compositions, LiF-BeF,- UF, (62-37-1 mole %) and LiF-BeF,-ThF,-UF, (62-36.5-1-0.5 mole %), were prepared by melting with ammonium bifluoride. These preparations were then examined by x-ray diffraction and microscopic techniques and found to be typical samples. Only traces of UO, were present. A sample of each, ground to 100 mesh and weighing approximately 10 g,' was placed in a nickel crucible and exposed to water-saturated air at room temperature. The initial rate of weight increase was 2.0% per day for each composition. As exemplified in Fig. 2.2.15, for the sample with ThF,, after 40 days the total weight of the sample had increased by 23% and the rate of increase was 0.32% per day; for the sample without ThF, the corre- sponding values were 30% and 0.38% per day. These weight percentages are based on the original weight before exposure. The samples were then placed in an oven exposed to air at 133°C. Within 24 hr they had returned to the original weights. But on continuing the drying to obtain a constant weight, the samples continued to drop very slowly to values 0.3 and 0.4%, respec- tively, below the starting weight. Unfortunately, the salts at this state could not be distinguished from uncontaminated fuel by either x-ray dif- fraction or petrographic techniques. The samples were then melted in a vacuum dry box at 30 to 50 p, and a release of gas bubbles was observed. There was a further weight loss of 0.6 and 1%, respectively, for the samples with and without ThF,. At this stage the samples were predominantly green with many small black 12prgonne National Laboratory, Reactor Fuel Processing 2(4), 42 (1959). B 63 T K TR v UNCLASSIFIED ORNL-LR-DWG 43723 | I » SAMPLE PLACED IN 133°C OVEN\ / / TOTAL WEIGHT (g) 14 voe / SAMPLE FUSED IN VACUUM DRY BOX o | 0] 10 20 30 40 50 TIME (doys) Fig. 2.2.15. Hydration-Dehydration Cycle for LiF-BeF ,-ThF ,-UF ; (62-36.5-1-0.5 Mole %). areas. The green portion had a higher UO, content than the starting ma- terials; the black portiofis were predominantly UO,, and it was estimated that from 20 to 40% of the uranium had been converted to UO,. Evidently the uranium had served as an oxide scavenger for the melt. A second pair of samples was treated as above except that the'baking in éir'was eliminated. Each sample went directly from the humidifier to the vacuum dry box, where it was dried at 135°C for 2 hr at 30 to 50 u and then melted. Bubbling occurred as before, and the final samples were 1.5 and 1% below the startifig weight. The safiples fused after vacuum drying were essentially identical with the fused samples from the first experi- ment. It was concluded'that the weight losses on prolonged drying were at least in part due to the substitution of O (wt 16) for oF (wt 38) as a consequence of hydrolysis, and that a satisfactory reconditioning of fuel by a simple drying operation is unlikely. 64 Graphite Compatibility Studies Penetration of Graphite by Molten Fluorides Continuing studies!?® of the permeation of graphite by fuel mixtures at approximately 1 atm led to the following conclusions. | Early experiments, in which only slight permeation of graphites such as AGOT occurred, reflected the effect of surface oxide in the graphite as a source of plugging by the formation of insoluble UO,. Such results could be altered to give easy permeation by using fuels with a larger oxide solubility, or by previously removing the oxide by treatment with H,. When easy permeation was achieved, the relative composition of the fuel in the graphite was normal, in contrast to the earlier behavior which involved complications due to oxide precipitation. Long-term tests gave the same amount of permeation as short ones. There was good agreement between the permeation obtained by experiment, in the absence of oxide interference, and the permeation calculated on the basis of the physical interaction be- tween a nonwetting liquid and a porous medium. Tests at lower pressures have been discontinued in favor of more per- tinent studies at 60 psig, corresponding to reactor conditions. At these higher pressures the permeation is increased in the expected manner, and differences in behavior due to the presence of oxide are usually not ap- parent. For example, when dealing with a readily permeable graphite, such as AGOT, little effect is noted as a result of pretreatment with H,. Attention has recently centered on the amount of permeation to be ex- pected at 60 psig in various grades of graphite; typical results are shown in Table 2.2.3. In the usual procedure the graphite specimens were baked out under 40-p partial vacuum at 1300°F for 24 hr and then left in contact with the salt at 60 psig and 1300°F. A test of graphite coated with siliconized silicon carbide was acci- dentally invalidated when, as later realized, the coating was punctured during mounting for submersion in molten fuel. Consequently, ensuing large weight gains were not necessarily attributable to failure of the coating. From a metallurgical examination it appeared that the coating had been severely attacked, but no definite conclusions could be reached because no specimens in the original condition were available for comparison. 13MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 90. 65 Table 2.2.3. Permeability of Various Graphites Permeation conditions: temperature, 1300°F; pressure, 60 psig Weight Gains (%) Graphite LiF-BeF,-ThF, LiF-BeF,-ThF,-UF, (67-18-15 mole %) (67.0-18.5-14.0-0.5 mole %) AGOT 17-22 2023 TSF 19 R-0013 10 S-4 4.6—7.0 3.6 21-CT-158 -0.1 : -0.1 18-CT-158 -0.1 -0.1 GT-123 0.1 0.1 ATL-82 6.2 CEY* (RLM-24) Not completed 0.1-0.2 CEY** special 1.0 Not completed *¥Extruded tubing, 1-1/4-in. OD x 7/8-in. ID. *¥Extruded tubing, 1/2-in. OD. . Intercalation of Graphite with Molten Salts The possibility of intercalation compounds of graphite and molten fluorides is unlikely, in view of the wide experience in handling fuels in graphite with no evidence of chemical reaction. However, because of the prevalence of intercalation of chloride with graphite,14 it was of in- terest to explore the phenomenon. Of the 20 principal fission elements, niobium, technetium, and pos- sibly molybdenum fluorides seemed most likely to exhibit intercalation with graphite. Most of the other fission-element cations form stable fluorides which have been included in melts in graphite crucibles in the course of other investigations in the Reactor Chemistry Division. Also, the more noble metals would be expected to exist as the free element and therefore would not intercalate. While molybdenum is also likely to be uncombined, it could possibly form a low-valence fluoride. Fluorides of molybdefium with a valence lower than six have been studied very little. Niobium forms 14y, Rudorff, "Graphite Intercalation Compounds,” in Advances in In- organic Chemistry and Radiochemistry, H. J. Emelens and A. G. Sharpe, eds., p 223—266, vol. I, Academic Press, New York, 1959. 66 a stable pentafluoride which has not yet been evaluated in graphite. Very little is known of the chemistry of technetium, but it seems to resemble rhenium more closely than it does manganese. When treated with molten molybdenum pentafluoride, graphite swelled to form an intercalate. A graphite tube (3/8-in. OD, 1/4-in. ID) was sealed in a copper capéule with MoFs; and maintained at 100 to 150°C for 17 hr. The unreacted MoF, was distilled off by heating to 200°C. As a result of the treatment part of the graphite tube had crumbled, but the outside diameter of the intact portion had increased to 1/2 in. Exposure to water caused the tube to evolve heat and hydrogen fluoride in-a manner similar to the behavior of pure MoFs. Molybdenum hexafluoride also gave evidence of intercalation, and a similar phenomenon has been noted repeatedly with UFg. Such strong oxi- dizing agents as theése could not survive in a reactor, but would be reduced to a lower valence state by reaction with the containing walls. Dif'f'usion ot Rare-KEarth kission Products in Graphite The development of techniques for the study of diffusion of rare earths in graphite is in progress. Specimens of graphite are to be im- pregnated with breeder fuels containing a trace guantity of Eul52,154, The activity distribution in the graphite will be determined before and after exposure to clean fuel salt by microcore drilling of small graphite samples from the face of a section of the graphite rod. Tube furnaces, furnace controllers, and Inconel fuel containers for the leaching operation have been procured. Equipment for microcore drilling has been designed and fabricated and the technique tested on graphite samples containing Cs'37, This technique will be adapted for use in the hot cell to analyze graphite samples from the MIR graphite—fused-salt experiment. A simple glove box for contamination control is under construction, and a saw for sectioning the graphite rod is on order. Specimens of graphite for the first test were received, were machined to the desired dimensions, and are ready for impregnation. Preparation of Purified Materials Technological Operations In order to stock-pile sufficient quantities of fluoride salt mix- tures to supply the Molten Salt Reactor Program for the remainder of this 67 fiscal year, the production facility was activated for three weeks, and 600 kg of fluoride salt mixes was processed. Small-scale processing op- erations accounted for another 140 kg, making a total of 740 kg. The increased rate of transfer and service operations encountered during the previous quarter continued during the first month of this quar- ter; the rate then dropped. Considerable effort has been devoted to de- veloping a technique for filling recently designed in-pile capsules15 for testing graphite. The results obtained to date are promising but not com- pletely satisfactory. Molten Ammonium Bifluoride as a Reagent Reactions of molten ammonium bifluoride have been useful for prepar- ing a large number of metal fluorides and ammonium fluometallates.® Ad- ditional reactions of ammonium bifluoride have been explored. Ammonium dichromate dissolved in ammonium bifluoride at 125°C to form an orange melt which vigorously bubbled and rapidly became green at about 200°C. Ammonium hexafluochromate(III) had formed by the following reac- tion: 6NH,HF, + (NH,),Cry0, — 2(NH,)3CrF¢ + 7H,01 + N, T Chromic acid reacted similarly: 8NH,HF, + 2Cr0O; — 2(NH,);CrFg + 6H,01 + N, T + 4HF 1 Cerium trifluoride was prepared by the thermal decomposition of a complex, presumably ammonium hexafluocerate(III), made by reacting molten ammonium bifluoride with ceric oxide: 20NH, HF, + 6Ce0; 2—00-$12H20T + 4HF T + NoT + 6(NH,)5CeFg (NH, ) 3CeFg @%Cem + 3NH; 1 + 30F 1 Lithium hydroxide monohydrate, enriched in isotope 7, was converted into the anhydrous'fluoride by fusion with ammonium bifluoride and elimi- nating volatiles at about 350°C. 15For the ORNL-MIR-47 experiment. 16MSR Quar. Prog. Rep. Apr. 30, 1959, ORNL-2723, p 93—9%. 68 ‘e Radiation Effects Corrosion Tests Under Radialtion Two static-corrosion capsules, made of INOR-8 alloy, were irradiated in the Materials Testing Reactor. Fuel composition and test conditions are summarized in Table 2.2.4. | Failure became evident from the release of radioactivity into the off; gas from one of the capsules. The two capsules were installed as a unit; so it was necessary to terminate the test for both capsules. The'Chromel-Alumel thermocouples attached to the capsules showéd no sign of failure, and the capsules are being removed from the reactor by a process designed to preserve the thermocouples. The capsules will be examined in the hot cells at ORNL. ‘ Metallographic tests will be made on the metal, chemical analyses will be made on the fuel, and calibrations will be made on the_thermocouples. Table 2.2.4. Conditions for 5000-hr Corrosion Test Under Irradiation Composition of fuel _ Melting point of fuel Composition of INOR-8 Operating temperature Temperature control Weight of fuel per capsule Power density (initial) Thermal flux Total hours in reactor Total hours above 842°F Tétal hours below 842°F Total hours not operating (shutdowns, ete.) Integrated thermal flux Estimated uranium burnup Li7F-BeF,-U%?°F, (62-37-1 mole %) 842°F Mo, 17%; Cr, 8%b; Fe, 5%; Mn, 0.5%; ¢, 0.06%; bal Ni 1250°F Air-cooled 0.400 g 1.17 kw/cc 1-2 x 10** neutrons-cm™ 5508 3392 1883 .sec” S 233 2 x 10?1 nvt 75% 69 2.3 FUEL PROCESSING Emphasis is being placed on modification of the hydrofluorination— volatilization process for application to MSR salts containing TnF4. The proceés as. originally developed was satisfactory for the recovery of ura- nium from core and blanket salts in a two-region reactor and for the re- cycle of Li7F and BeF, used in the core fuel.! There was no provision in this process, however, for the handling of thorium, which is particularly desirable in core sali, where frequent processing is necessary. The hy- drofluorination—volatilization process has as an attractive feature the fact that the processed material remains in the flupride form and thus re- quires little further treatment before re-use in the reactor. Retention of this feature is one objective of the work. | The use of hot aqueous NH,F solution appears deflnltely prom1s1ng for processing ThF,-bearing fuels, and further ‘work in this direction seems | appropriate. The possibility of using C1F;-HF as a solvent for the simul- tanéous solution of UF, and LiF no longer appears attractive because an appreciable concentration of ClF; represses the LiF solubility. and because the sqlubilities of UF, and ThF, are unsatisfactory. The NOg'HF system appears attractive for the simultaneous dissolution of LiF, BeF,, and UF,, but it does not show sufficient ThF, solubility. Thorium Fluoride Solubility in.Aqueous NH,F Solutions An exploratory test indicated that ThF, was soluble to about 5 wt % in 50 wt % aqueous NH,F solution at the boiling point, which is high enough to be of interest for processing. Typical fuel salts contain about 60 wt % ThF,. Experimental data indicated increasing ThF, solubility with increas- ing temperature and increasing NH,F concentration, although some difficul- ties were encountered in sampling solutions at or near the bbiling point. . The ThF, solubility in 50% solution at 105°C was 49 mg/ml'in a sample pi- petted from a solution in which the solids had been allowed to settle, and was 43 mg/ml in a sample filtered through sintered nickel. The solubility dropped rapidly with decreasing temperature, to about 12 mg/ml at 60°C and 4.9 mg/ml at 25°C. 1In 43% NH,F solution at 25°C the solubility was only 1. G. MacPherson, Molten Salt Reactor Program Status Report, ORNL- 2634 (Nov. 12,.1958). 70 0.4 mg/ml; The solid formed when these NH,F solutions of ThF, were cooled or diluted was not ThF,, but rather a compound (or compounds) involving NH,F and ThF,. The compositions of solids formed under various conditions are being determined. The investigation is being extended to measurements of the solubili- ties of the other fuel components — LiF, BeF,, UF,, and rare earths — both “individually and collectively. The effect of other components on the ThF, solubility has not been measured. Fuel-Component Solubility in Cl¥F,4-HF Mixed ClF;-HF solvent appears unsatisfactory for a one-step process to dissolve both uranium and LiF. Further experiments confirmed the ob- servations reported previously2 that C1lF3 in HF causes a drastic reduction in the LiF and BeF,; solubilities with MSR fuel salt (LiF—Bng-ThF4-UF4, 62-36.5-1-0.5 mole %). The presence of significant amounts of ClF,; in HF decreased the solubility of LiF from ~100 to ~20 g/kg, and of BeF, from 3040 to < 5 g/kg. Uranium solubilities were generally about 0.25 g/kg; although significant, they are too low to dissolve all the uranium as- sociated with the dissolved LiF and BeF,. Rare-earth and thorium fluorides are very insoluble in these solutions; rare-earth solubilities were in the range 0.0001 to 0.0005 mole %, bhased on dissolved salt. The U(IV), which has a low solubility, is probably not converted under these conditions to the highly soluble U(VI). Operation of the ClF;-HF system at a higher temperature and pressure to obtain conversion appears impractical. Hence no further work in this direction is planned. Mel-Component Solubility in NO,-HF Scouting tests of the solubility of fuel salt (LiF-Bng—ThF4—UF4, 62-36.5-1-0.5 mole %, plus 0.05 mole % rare-earth fluorides and trace fis- sion products) in NO,-HF solutions were completed. The results (Table 2.3.1) show. that LiF and BeF, have reasonably good solubility if the NO, content is not much higher than 5 mole % (above which LiF solubility drops) or much lower than probably 2 mole %_(below which BeF, solubility drops). Uranium solubilities were generally about 1 g/liter, with some indication 2MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p %. 71 Table 2.3.1. Solubilities of Fuel Components and Fission Products in NO,-HF Solutions Solubility (mggg %) Major Components (g/kg) Rare Earths (mole %) LiF BeF, U Th Based on Ce Based on Pm 2.7 35 62 1.0 0.06 0.0013 0.0025 5.0 48 49 0.9 0.15 0.0006 0.001 10.0 19 36 1.4 0.04 0.001 0.0003 that values would be higher at NO, concentrations above 10 mole %; uranium solubilities are presumably high in the absence of LiF and BeF,. Thorium solubilities were generally low, and rare-earth solubilities were mostly about 0.001 to 0.002 mole %, based on dissolved salt (the most reliable value, from an experiment discussed below, was 50-fold less than this). With 2.7 mole % NO, (Table 2.3.1) the BeF, solubility was unexpectedly high, corresponding almost exactly with the NO, content on a mole basis. The LiF solubility was only about half as high as expected. The uranium solubilities account for about one-third of the uranium associated with the dissolved salt except with 10 mole % NO,, where it accounts for all of it. With 5 mole % NO, the total salt solubility is a little over 100 g/liter, which is a reasonable value for processing. In another experiment with approximately 2 mole % NO, in HF, the so- lution was decanted from the undissolved excess salt and both were analyzed separately. The LiF, BeFé, uranium, and thorium solubilities were 33, 52, 0.5, and 0.04 g/kg, respectively, in reasonable agreement with the values reported above. The NO,-HF solution dissolved 56% of the LiF, 79% of the BeF,, 5% of the uranium, and 62% of the cesium activity, but only 0.086% of the cerium and 0.05% of the promethium. The much larger samples per- mitted more accurate determinations of the rare-earth activities than in the experiments reported in Table 2.3.1. The total rare-earth solubilities were 0.000042 mole % calculated from the cerium determination and 0.000029 mole % calculated from the promethium determinations, based on dissolved LiF and BeF,. 72 The above results were obtained at room temperature. As NO, was added to HF, the boiling point rose rapidly to about 56°C at about 25 mole % NO, and then dropped to the boiling point of NO, (21°C) with about 60 mole % NO,, at which point a second phase appeared. The second phase appeared to be nearly pure NO,. Salt solubilities would probably be significantly highef‘at the boiling points, since salt solubilities in other HF systems have been observed to increase with temperature. 73 THIS PAGE WAS INTENTIONALLY LEFT BLANK ooe-aIo0owmdsw- * L] - * QNP PEHEEAHIGPEUEPUGI NG GOSN U S YEZQEQREAY E T ORNL-28&90 Reactors — Power TID-4500 (15th ed.) INTERNAL DISTRIBUTION Affel - 47. G. M. T. Kelley G. Alexander . | - 48. F. Kertesz S. Bettis : K 49, B. W. Kinyon S. Billington 50. M. L. Lackey F. Blankenship : : 51. J. A. lLane P. Blizara ' 52. R. S. Livingston L. Boch 53. H. G. MacPherson J. Borkowski 54, W. D. Manly F. Boudreau 55, E. R. Mann E. Boyd 56, L. A. Mann A. Bredig 57. W. B. McDonald J. Breeding 58, H. F. McDuffie B. Briggs 59, J. R, McNally E. Browning : €0, H. J: Mctz 0. Campbell _ 6l. R. P. Milford H. Carr . - 62, E. C. Miller I. Cathers 63. J. W. Miller E. Center (K-25) : | 64. K. Z. Morgan A. Charpie 65. J. P. Murray (Y-12) H. Coobs ‘ 66. M. L. Nelson L. Culler 67. G. J. Nessle H., DeVan y 68. W. R. Osborn A, Douglas : 69. P. Patriarca B. Emlet (K-25) ' 70. A. M. Perry K. Ergen : 71. D. Phillips Y. Estabrook ‘ 72. W. D. Reel E. Ferguson - 73. P. M. Reyling P. Fraas ' 74. J. T. Roberts A. Franco-Ferreira 75. M. T. Robinson H. Frye, dJr. 7. H. W. Davage R. Gull . 77, J. L. Bcott T. Gresky ' ‘ C 78. H. E. Seagren L. Gregg ' ' 79. E. D. Shipley R. Grimes o g80. M. J. Skinner Guth 4_ _ 8l. G. M. Slaughter S. Harrill 82. A. H. Snell R, Hill o 83. E. Storto E. Hoffman 8. C, D, Susano W. Hoffman - | 85. dJ. A. Swartout Hollaender : ' ~ 86. A. Taboada S. Houscecholder _ 87. E. H. Taylor H. Jordan o | 88. R. E. Thoma W. Keilholtz - _ » 89. D. B. Trauger P. Keim o 90. F. C. Vonderlage 75 M. Watson - 8. J. Zasler 91. G. .92. A. M, Weinberg ' 99-102, ORNL - Y-12 Technical Library, 93. M. E. Whatley : , Document Reference Section 94, J. C. White | 103-142., Laboratory Records Department 95. G. D. Whitman . 143. Laboratory Records, ORNL R.C. 9. G. C., Williams 144-145, Central Research Library 97. C. E. Winters EXTERNAL DISTRIBUTION 146. D. H. Groelsema, AEC, Washington 147. Division of Research and Development, AEC, ORO 148-742. Given distribution as shown in TID-4500 (15th ed.) under Reactors-Power category (75 copies - OTS) - 76