ORNL-3872 UC-80 — Reactor Technology TID-4500 (46th ed.) " MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT FOR PERIOD ENDING AUGUST 31, 1965 OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION | for the U.S. ATOMIC ENERGY COMMISSION Printed in USA. Price $5.00. Available from the Clearinghouse for Federal Scientific and Technical Information, National Bureau of Standards, U.S. Department of Commerce, Springfield, Virginia 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 6wned_righfs; 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. Contract No. W-7405-eng-26 /‘ { . MOLTEN-SALT REACTOR PROGRAM R SEMIANNUAL PROGRESS REPORT For Period Ending August 31, 1965 R. B. Briggs, Program Director DECEMBER 1965 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION $ _L’{\; 7 "IN s C) 0 3/ 1ii Part 1. MSRE OPERATIONS AND CONSTRUCTION, ENGINEERING ANALYSIS, AND COMPONENT DEVELOPMENT 1. MSRE OPERATIONS..... s e ecesessecesececesene s ssce s e eeseo s e s 7 Chronological AcCoUnt..coeeeeeeess cecsescesssenns ceeeeesencens 7 Component and System Performance.....ccceeeececsscocsocccscsns 10 Control RodS.eeeeeeeoosns s s eesceccccceresns s s s seses s e s 10 Sampler-Enricher...ceceeeeeeces ceoscoescesesess s et e an e 11 Freeze Valve S .. eeeeerssooscesscososcossocosssosososossssscsossos 12 Freeze FlangeSeeeeeo oo 6t e ocecsecceevsesecessessssas s e e s s e s 13 Fuel and Coolant System Pressure Control.....cceeeeeecesss 13 Heaters and Insulation..eeeeeceeeccsccsoocess ceeseeseneene 15 Reliable Tnstrument Power SyStelm....eoceeessceosccosooosas 15 Thermal-Shield Cooling Waler..oseeeoeeeececocrssocescccnossoe 16 Fuel-Pump Overflow TanK...eeeeoeeoeceocssssssossesscccsnse 16 Analysis of Operation...cceeeeececesescscescsssoscccsossosccacscse 16 Basic Nuclear CharacteristiCS.eeeeececcsssooccscososccsesss 16 Dynamic TeStS.useeeeeoseeeeoeeeeososasasosssscssssossnosssss 24 Undissolved Gas in Fuel LOOPeeeceecsssssoosoosooocoossass coe 30 Salt Density and Inventory EXperience......ceeececesccesss 30 External NeUbron SOUILCE .. eeeessscesoccecooossossossoscsssssssse 31 Instrumentation and Controls Design and Installation.......... 32 CEIIET A L e e s 0o s ovoooooeoccsssossscsssesosasososssscssssssssessss 32 Design, Installation, and Checkout........ooeeveeencaeennn 32 Tnstrumentation and Controls System Performance........... 40 Revisions and ModificationsS.ceeeesceeesoccccosccosccsacsanss 43 DOCUMENLEEI0N . s s e e vesesoeecsesssscssssssssscssescaassccsses 46 2. COMPONENT DEVELOPMENT. . cocavcoscesossoascsasossssosssscocnssseocs 48 AT e T OE S e s e s s evoeosososoesasessesesseosssssssssonsescsssasssos 48 FrEEZE VOLV e e oo eseesoeosocessssssosassssssosessossosasconsss 48 Pipe Healer e iveeeeoseooonsocsocssassssassssscsasscses eeaes 48 Drain-Tank Heaber.. v e eeeeeesaosossssscssssossscccososssss 48 Checkout and Startup of ComponentS...eeeeeeeeeecescccoacasnocss 48 CONETOL ROAS . e s esecsosesesecsesscssessssssssscsscsscsscsnssse 48 Control Rod Drive Unibts..eeeeeeeesocosesoosssoscosccscncssss 49 Freeze ValveS..... cecoessssssess ceoeees ceecoseecsecssasesasene 51 Freeze Valves 204 and 206...eeeesseoesocccscssosssssossssnss 51 Freeze Valve 103, . ceeeecesoeeoooscsssosssocsssossssssssscascesaes 51 Freeze Valves 104, 105, and 106...cceeeesseccsccssancosons 53 Noble-Gas DynamicCS.eeeseeceosocscosscososss cecesereacacnos 55 Sampler-Fnricher..ceeeeeecescoscocsscsss ceseceensno e cesece 56 iv Removal Valve and S€a8l...ceecocssseossossoscsocassocsosnossss . 56 MAnIPULAEO e s e e veesrosossocososossssssssscsessocssasooscnsse 56 Access Porteeeeeesoseoes cecoeeaceos ceeeeseessssesss s e e e 57 Operational Valve...oeeeoeese ceceon e teceeseccescssssens e 57 Sampling CapsUle. e eeesseoeosooccsoesoseees ceeseeseceseene 58 Fuel-Processing SampPler.ceceessssscosscossossssssosocscassos 58 Off-Gas SySteMeoeesssseasssoseoosns Gt e e seeereseeseessasanns 58 Remobte MainbenanCe.eeeeeseoccsoossososssosesossoosssscssesesssscoseco 60 Pump Development.ecessseosoocoeocssesosssosscsocsscsssocassonsos 61 MSRE PUIDS e e e o sessosesesssscsososssossesssssososossossssss o1 Measurement of the Concentration of Undissolved Helium in Circulating Molten Sall..eeceeeecccoscessnscsoccacsnccne 62 Other Molten-Salt PUMPSeececscoceccccssossoosossssocsssocssss 65 Tnstrument Developmenteeeeeessesrsooossesesrscssssoossssessccscs 66 Ultrasonic, Single-Point Molten-Salt Level FProbe..... conee 66 High-Temperature NaK-Filled Differentlal-Pressure T aNSIMIi T e e e v sesveceseossccesosssssssoosossssscsssocssssss 70 Float-Type Molten-Salt Level Transmitter.....ccocceeeeecses 71 Conductivity-Type, Single-Point Molten-Salt Level Probe... 71 gingle-Point Temperature Alarm Switches.....coceeeeeeasees 72 Helium Control Valve Trim Replacement,..ecoeseesoesocoscocss 72 Thermocouple Development and Testinge.ceseeeseerecosseeses 73 MSRE REACTOR ANALYSTIS:eeoeeoeecsssosooscoscocns cessecees e e e 75 Theory of Period Measurements Made with the MSRE During Fuel Circulation..e.eeeeoseosoossssosescssoososssssssossssosssocesss 75 Part 2. MATERIALS STUDIES B/[ETA—LL.[JRGY....0..0...0..000.0......9.....0.0....‘..O’O...'0.0. 8]— Dynamic Corrosion StudieS..veeeeeseressecssooaesccsocscsescans 81 Corrosion Studies Using Lead as a Coolanf...eecevesoncssse 82 Revised Thermal Convection Loop DesSigleceoceecsoscocssonss 85 MSRE Maberials Surveillance TeStiNgeeeeececsecscesssssessssees O7 TNOR-8 Surveillance SpPeCimMeNnS.eceeseososscccscsssssssssssos 88 Graphite Surveillance SPECIMENS.ceeecoesssosossssosoocsons 88 Assembly of Survelllance SpPECIMENS.cseecesvessoocasacnscss 20 Fvaluation of Radiation-Damage Problems of Graphite for Advanced Molten-Salt ReaCtorS.eeeesescessscoossosssososscoscssce 93 Mechanical Properties of Irradiated INOR-8...ceeecevoscssccson 94 Mechanical Properties of Unirradiated INOR-8 Used in the Reactor VessSel..veeeeessoososososooososessssssososssos V. Postirradiation Stress-Rupture Properties of INOR-8....... 102 In-Reactor Creep TesTS..ieeereserncosrceeccancnnscsocccncen 104 RADIATTION CHEMISTRY ¢ v s evevescesosoncscsossssensossseassseassses 106 IntrOduCtiOIl.O..O..OOO..0.00‘0'0'..0.00.0..O...Q.OO’.O..O..... 106 Experimental Objectives and Design ConsiderationS..eeeeese 106 Experiment Design and Mockup Test Operation........ ceeeees 107 CHEMISTRY ¢ o s oecseoeesssosonocooncss ceceens ceecesensenen s e ees 111 Chemistry of the MSRE..eeseesss ceeev oo cessssses e cereserssen e 111 MSRE Fuel LOBA1NEeoeersoesossosososasossssssssocssssassassonsaes 111 MSRE Salt Chemistry During Precritical and Zero-Power Experiments.eeee.. c e s e s e e e e ees e s e e s et et es e et s e e e s ceees 112 Examination of Salts After the Zero-Power Experiment...... 117 Density of MSRE Fuel and Coolant SallsS.eieeeesescesaccenes 119 Chemistry of LiF-BeF, Systems.. ceereacns ceeecescacpencanes 123 Solubility of HF and DF in Molten LiF-Bel's (66-34 Mole B)eereensanss ceveeeessssesassseesss s ceeesas 123 VaPOr PreSSUlCessecessscescesocsasessscesss ceeeseseseens 123 The Quest for quuld Ligquid Tmmiscibility in LiF-BeF, Melts., 6t e e s e s cecesseese s s s e s s e s e ns ees 126 Removal of Iodlde from.LlF -Bek, MElts by HF-H, SPATrEING . ceeservocoses s s esseeresecesees ceseecessenees e s 127 Salt Compositions for Use in Advanced Reactor Systems........ . 133 Blanket Salt Mixtures for Molten-Salt Breeder Reactors.... 133 Coolants for the Molten-Salt Breeder Reactor...eeseeesss.. 135 Viscosity Of NaBF,eeveoececococcscossosoosssconsssocsocscss 136 Fuel and Blanket Materials for the Proposed MOSEL Reactor..... 137 Recovery of Protactinium from Fluoride Breeder Blanket MixXbUreSeseesoeess 6t oo s e e eeec e et s e e s e e st es s e s e e es s e ns 137 Introduction.seeeeeeos. ce e e e s s s e st e e s e e s et tee s s s e eeo 137 Facility DescriptioN..ceecee. cesessecsese st s ee s ee oo s s 137 Oxide Precipitation of Protactinium.......eceeceeeee cesess 140 Development and Evaluation of Methods for the Analy51s of MSRE Fuel..... creceseeessesscee s cecesecen cesessecenses eo. 140 Determination of Oxide in MSRE Fuel....... cesens ceeeseness 140 Electrochemical AN8lySiS..eseceeeesessesescsosssoscossasas 143 Spectrophotometric Studies of Molten-Salt Reactor Fuels... 145 Analysis of MSRE Blanket Gas. ceeeseseeresecsensnseenns 147 Development and Evaluation of Equlpment for Analy21ng Radioactive MSRE Fuel SampleS...... ceese s cereaes cecsesecesss 148 Sample Preparation..eeeecoececocesass ce s esesacesecesoseonees 148 Sample Analyses.. ceeereecenoaes cesecsveco s ceesescanee 148 Quality Control Program ........ cevenas ceeresetreseanas eee. 148 FUEL PROCESSING....... cessecssscevvsessreenes e veoecconon eesse 152 .- vii SUMMARY Part 1. MSRE Operations and Construction, Engineering Analysis, and Component Development 1. MSRE Operations The first run in which salt was circulated ended on March 4, after the planned prenuclear tests were completed. In this run flush salt was circulated for 1000 hr in the fuel loop and coolant salt for 1200 hr in the coolant loop; equipment performance was generally satisfactory. With the reactor shut down, the operators recelved advanced tralin- ing. They were then examined and, if they qualified, were certified as nuclear reactor operators. Meanwhile the system was prepared for low- power nuclear operation by installation of the nuclear instruments, the fuel sampler-enricher, and one layer of concrete blocks over the reactor cell. Fuel carrier salt (lacking enriched uranium) was charged in late April. As a final checkout and to provide base-line data, this salt was circulated for ten days before the addition of enriched uranium began on May 24. Criticality was first attained on June 1, with circulation stopped, control rods almost fully withdrawn, and a 2357 concentration very close to predictions. June was spent in adding more 235U} while one control rod was calibrated over its full travel, reactivity coeffi- cients were measured, and dynamics tests were conducted. The nuclear power was held to a few watts except during planned transients, in which it was allowed to rise to a few kilowatts. Nuclear characteristics were in good agreement with predicted values. Performance of the mechanical components of the system was gratifying. Only minor difficulties were encountered, interfering in no way with the experimental program. The fuel system was drained and flushed on July 4—5. Radiation levels were low enough to permit maintenance and installation work to begin immediately in all areas in preparation for high-power operation. The remainder of the period was spent in this work. Except for a small amount of work on the fuel-processing system sampler, completion of documentation, and a few minor additions and re- visions, the design, installation, and checkout of the MSRE Instrumen- tation and Controls System are now complete. Since our last semiannual report the design, installation, and checkout of all instrumentation and controls systems required for power operation of the reactor were com- pleted, the computer—data-logger system was installed and checked out, several revisions and modifications were made to improve performance or correct errors in design, and some safety instrumentation and associated control circuitry were added. Documentation of design and as-built changes viii is nearing completion. In general, performance of the instrumentation has been very good. Although some failures and malfunctions have occurred, once the causes of the failures were determined, they were easlly cor- rected, and no major changes in instrumentation or in design philosophy have been necessary. 2. Component Development A thermal cycling test was completed on the prototype of a freeze valve after 1800 cycles through the temperature range from 600 to 1200°F. No cracking or other physical damage was evident. Prototypes of the removable heater for 5-in. pipe and the drain-tank heater completed over 12,000 hr of satisfactory test operation. The modified control rods were operated in the reactor during the criticality test and performed satisfactorily. A satisfactory procedure for attaching the control rods to the drive unit by use of remote methods was demonstrated. The limit switch actuators, which operate at the limit of the con- trol rod stroke, were modified to improve the bearing surfaces. A pro- totype of the modified switch was operated through 14,760 switch opera- tions without difficulty. The buffer stroke of the shock absorber had changed by as much as 50% during the criticality test and had to be re- adjusted before reinstallation. Temperature alarm switches were installed in the lower end of each control rod drive to indicate gross changes in the cooling air flow. The wire cables for the electrical disconnects were lengthened to facilitate remote removal and installation. The cooling air flows to the freeze valves were increased to about 2/ sefm. Freeze valves Tor the coolant drain tank were modified to re- duce the thaw time under power failure conditions. This time is now about 13 min. We continued to experiment with the reactor drain valve in an effort to improve the procedures for operating the valve under the dual set of operating conditions. Also there was some difficulty in operating this valve when the reactor temperature was below about 1100°F. A 4if- ferential controller was installed in the cooling air stream, and the valve performed satisfactorily during the critical experiment. We also had some difficulty with the freeze valves in the distribution lines to the drain tanks, and methods were developed to ensure proper filling of the valves with salt before the valve was frozen. In addition, the problem of me- chanical failure and inadequate capacity for the heaters in these valves was solved by installing new heaters. Analysis of the experiment on the behavior of noble gas in the re- actor was continued. Preliminary results of one experimental run yielded five rate constants which control the transient behavior of 85Ky, A rea- sonable agreement was found between these constants and the equivalent physical constants measured elsewhere. 1X A total of 54 samples were isolated and 87 capsules of enriching salt were added by means of the sampler-enricher during the zero-power runs. The equipment was tested for the first time and operators were trained during the same period. Maintenance was sufficient to improve the performance of the sample- removal valve and the operational valve. The manipulator boot failed three times, but the causes have been eliminated by modifying the pro- tecting interlock system which controls the pressure and by altering a protrusion in the normal path of the manipulator. A sample capsule was retrieved from the top of the operational valve after it had been acci- dentally dropped. The design of the fuel-processing sampler was started. A holdup test using 85Kr as an indicator was made to check the per- formance of the MSRE charcoal beds. The beds performed as predicted. Difficulties with the operation of the valves in the fuel and coolant off-gas system were traced to an accumulation of glassy spheroids of the salt and a carbonaceous material which could be a normal residual of the manufacturing process. The filter element upstream of the valve is being changed from one with a 25-j pore size to one with a 1-K pore size, and no further trouble 1s expected. Design was started on an off-gas sample unit which is to incorporate an on-line indicator of the total contami- nants as well as provide the means of concentrating and collecting batch samples of the off-gas. Practice with the remote-maintenance tools and procedures was con- tinued. FPhotographs were taken of all the installed equipment in order to provide a final record of the as-installed condition. The spare rotary assembly for the fuel pump was operated for 2644 hr circulating salt. The radiation densitometer used to determine the concentration of undissolved gas in the circulating salt during this test was used again at the MSRE to make a similar measurement for the circulating fuel salt. Also, back-diffusion tests using 85Ky were made during the same test. The spare rotary assembly for the MSRE coolant pump was completed. Tests on the mockup of the MSRE lubrication system were completed. The priming problem with the standby lube pumps was resolved by modifying the jet pump in the oil return circuit to the reservoir. The PKP molten- salt pump was placed in operation at 1200°F for an endurance run. Fabrication and installation of an ultrasonic level probe system in the MSRE fuel storage tank were completed. The probe performed very well during the initial filling of the fuel storage tank but did not operate when the tank was drained. This malfunction was determined to have been caused by frequency drift of the excitation oscillator. Performance studies revealed other characteristics which could limit the usefulness of the instrument for long-term service under field conditions. Modifi- cations are being made or are under consideration which would eliminate the unwanted characteristics. The coolant-salt-system flow transmitter that falled in service at the MSRE was replaced by a spare transmitter. A new transmitter has been ordered for use as a spare. Tests are being performed on the defective transmitter to determine the cause of failure and to determine whether it can be repaired. Inspection of the core tube in a prototype ball-float-type molten- salt level transmitter showed that the buildup of vapor-deposited salts in critical areas has not been sufficient to affect the performance of the instrument. The fuel flush tank level probe was modified and repaired. Consid- erable difficulty was experienced with sulfur embrittlement of nickel wire in the replacement assembly. Performance of the repaired probe and of other probes installed in MSRE drain tanks was satisfactory during critical and low-power operations. Modifications of the temperature alarm switches to eliminate spurious set-point shifts were completed. 1t i1s not known at this time whether the modifications effectively eliminated the set-point shifts. Results of investigations indicate that the failure of four helium control valves in MSRE service was probably caused by misalignment and complete lack of lubrication rather than incompatibility of plug and seat materials. Calibration drift of eight thermocouples made of materials selected from MSRE stock remained within the limits previously reported. Ten MSRE prototype, surface-mounted thermocouples installed on the prototype pump test loop continued to perform satisfactorily throughout the test. Revisions were made in the MSRE coolant-salt radiator AT measurements system to eliminate long-term drifts and noise found to be present in the MORE installation. 3. MSRE Reactor Analysis As an aid in interpretation of the zero-power kinetics experiments with the MSRE, the theory of period measurements while the fuel is in circulation has been developed from the general reactor kinetic equa- tions. The resulting inhour-type equation was evaluated numerically by machine computation, and results are presented relating the reactivity and the asymptotic period measured during circulation. By means of this analysis, the measured and calculated reactivity differences between the noncirculating and circulating critical conditions were found to be in close agreement. X1 Part 2. Materials Studies 4. Metallurgy Thermal convection loops made of INOR-8 and type 304 stainless steel have circulated LiF-BeF,-ZrF,-UF,-ThF, (70-23-5-1-1 mole %) fuel for 29,688 and 18,312 hr respectively. A maximum attack of 0.002 in. was found on specimens removed from the type 304 stainless steel loop. Thermal convection loops were run to evaluate the compatibility of lead with Croloy 2—1/4 steel, low-alloy steel, type 410 stainless steel, and. Cb—1% Zr at 1100 to 1400°F meximum temperatures. The steel loops tended to plug in the cold regions and have general surface corrosion in the hot region. The Cb—1% Zr was found to have no measurable attack at 1400°F in 5280 hr. A new loop design that allows improved temperature control of the cold region was tested for use in coolant evaluation studilies. MSRE surveillance specimens were assembled and placed in the re- actor core, the control test rig, and the area adjacent to the reactor vessel. An assembly of graphite specimen, INOR-8 tensile bars, and flux monitor wires will be exposed to fluxes at various points of the reactor to anticipate and match the effects of radiation on the materials of the reactor core and vessel. The radiation-damage problems were evaluated for graphite in ad- vanced molten-salt reactors, considering growth rate, creep coefficient, flux gradient, and geometric restraint as important factors. The stress developed because of differential growth in an isotropic graphite should not exceed the fracture strength of the graphite and cause failures. Evi- dence exists that %raphite should have the ability to withstand damage to doses up to 4 X 10 2 Data are needed for greater dose levels. A study of INOR-8 specimens made from heats of material used in the MSRE reactor vessel indicates that (1) creep strength is comparable to those reported previously, (2) properties of the alloy are very sensitive to mechanical and thermal treatment, and (3) welding of air-melted heats without subsequent heat treatment causes large reductions in rupture life and ductility, whereas welding of vacuum-melted heats causes small effects on these properties. A microstructure study using the electron microprobe indicates that the large precipitates present are nickel-molybdenum inter- metallics with high silicon content. Several modified alloys are being studied that have potentially improved properties. Postirradiation stress-rupture properties of heat Ni-5065 and heat 2477 at 650°C were determined at doses varying from 5 X 1016 to 5 x 1020 nvt. These data show that ductilities at the higher levels vary from 1 to ~5%, the lower-boron-bearing heat 2477 having the higher ductility. otress-rupture life was reduced as dose level was raised. In-pile creep data were developed for two heats of material. Xi1 5. Radiation Chemistry The in-pile irradiation program is being changed from an MSRE-oriented program to one that will provide an understanding of both short-term and long-term effects of irradiation and fissioning on molten-salt reactor fuels and materials. Ixperimental objectives of the program are: (1) 200'w/bm3 fuel fission power, (2) meximum fission product production, and (3) long~-term in-pile operation (up to one year). The i1rradiation tests are to be conducted in beam hole HN-1 of the ORR with an autoclave (capsule type) experiment. Design features in- clude: (1) circulation of the salt by means of thermally induced flow, (2) sampling and replacement of fuel salt while operating in-pile, (3) cover gas sampling, and (4) keeping fuel molten at all times. oeveral prototype models of the in-pile molten-salt autoclave ex- periment have been constructed and operated in a mockup facility. Some 4000 hr of operation with a salt mixture similar to the MSRE fuel salt have been accumulated. Results of these mockup tests indicate that the presently designed autoclave is suitable for in-pile experiments with molten-salt fuel. 6. Chemistry A1l fluoride mixtures — coolant, flush, and fuel — for the operation of the MSRE were prepared and loaded into the reactor facility by the Re- actor Chemistry Division. Capsules of fuel concentrate, containing about 85 g of 2357 each, were provided for use in reaching criticality and for criticality maintenance during nuclear operation. Chemical analyses of the MSRE fuel were carried out during the pre- critical, zero-power, and postcritical stages for the purpose of estab- lishing analytical base lines for use in the full-power operating period. Chemical composition, contaminant levels, and isotopic analyses were ob- tained regularly on samples obtained daily throughout the zero-power ex- periments. Judging from the concentration of Cr<™, which is the primary corrosion product, essentially no corrosion occurred during the 1100-hr precritical and zero-power test period. From the standpoint of chemical evidence, the MORE salts were maintained in an excellent state of purity during all transfer, fill, and circulation operations. Unless dilution of the fuel by flush salt is postulated, uranium analyses for samples obtained in both precritical and zero-power experiments were about l% below book values. New experimental values for the densities of the fuel and the coolant agreed well with results on the weights and volumes of salts in the drain tanks at the MSRE site. The solubilities of HF and DF in the molten mixture LiF-BeF, (66-34 mole %) were measured over the range 500 to 700°C at pressures of 1 to 2 X111 atm. The solubilities were of the order of 2 X 10™% mole of HF per mole of melt, and DF solubilities were lower than HF solubilities by about 10%. Vapor pressures were measured for the LiF-BeF, system over the entire composition range. The vapor above liquid compositions containing 70% or more BelFp was virtually pure BeFy. Vapor pressures of importance in recovering the MSRE fuel by distillation were also determined. The fea81b111ty of removing 1357 from the fuel, as a way of reducing the amount of *3 Xe, was examined in greater detail. Half the iodide con- tent could be removed by using 388 cc of gaseous HF in an Ho-HF mixture per kilogram of melt. A Turther search for liquid-liquid immiscibility in the LiF-BeF, system at high Bel, concentrations was made; no immiscibility region was found. Considerations of phase behavior in ternary fluoride systems con- taining ThF, have led to the selection of suitable blanket systems for breeder reactors. A search for suitable coolants, however, continues. At present, interest is centered on the potentialities of fluorides and fluoborates, possibly in combination with Bs03. Reports in the Russian literature of a low-melting eutectic of NaF-NaBF, could not be confirmed. The new facility was designed for laboratory-scale studies of the removal of protactinium from fluoride breeder-blanket mixtures and for supporting research work. Glove boxes will permit use of the *21Pg isotope to give concentrations in the expected operating range of 50 to 100 ppm. Hot cells would be required for work with equivalent concen- trations of <° Pa but millicurie amounts of this gamma-active isotope will be mixed'w1th 231pg in order to minimize the need of alpha anal- yses., A preliminary experiment was performed to test the equipment and to confirm the previously reported precipitation of protactinium by addition of oxides. Protactinium at tracer concentration K1 ppb) was completely pre01p1tated'by addition of thorium oxide to molten LiF-BeF,-ThF, (73-2- 25 mole %), and treatment of the melt with a dry mixture of HF and H, redissolved the protactinium. A prototype apparatus was constructed and tested for the determina- tion of oxides in the MoRE fuel using the hydrofluorination principle. The water produced from the reaction of HF with oxides i1s measured auto- matically by means of an electrolytic moisture monitor. The entire ap- paratus 1is being assembled for insertion into a hot cell in order to an- alyze the fuel after the reactor has gone to power. otudies were continued on adapting electrochemical methods to in- line analysis of impurities in the fuel. When the simulated fuel is subjected to controlled-potential electrolysis, gas evolution, primarily X1V of COz, CO, and Oz, is observed at the indicating electrode. This indi- cates removal of oxide from the melt by electroreduction. This discovery holds promise for a possible in-line determination of oxide. Absorbance spectrophotometric studies are also under way which are designed to de- termine trivalent uranium and tetravalent uranium in the fuel by their characteristic absorbance peaks. A process gas chromatograph equipped with s helium breakdown-voltage detector is under construction for the continuous analysis of the helium cover gas 1n the reactor. A metal diaphragm sampling valve has been de- signed specially to withstand the temperature and radiation effects that nullify the use of conventional sampling valves. camples from the MSRE precritical and zero-power experiment were analyzed in the HRLAL hot cells. The results, using the specially de- veloped equipment and analytical methods, were satisfactory with the exception of those for uranium and beryllium. Statistical evaluation of the control data indicated a negative bias of ~0.8% for uranium and none for beryllium. 7. Fuel Processing Construction of the MSRE fuel-processing system was completed, the system was tested, and the flush salt was processed for oxide removal. Operation of the plant was generally satisfactory, and about 115 ppm of oxide was removed from the salt in reducing the concentration to about 50 ppm. INTRODUCTION The Molten-5Salt Reactor Program is concerned with research and de- velopment for nuclear reactors that use mobile fuels, which are solu- tions of fissile and fertile materials in suitable carrier salts. The program is an outgrowth of the ANP efforts to make a molten-salt reactor power plant for aircraft and is extending the technology originated there to the development of reactors for producing low-cost power for civilian uses. The major goal of the program is to develog a thermal breeder re- actor. Fuel for this type of reactor would be 33UF4 or 235UF4 dissolved in a salt of composition near ZLiF-Bel';. The blanket would be ThF, dis- solved in a carrier of simillar composition. The technology being devel- oped for the breeder is applicable to, and could be exploited sooner in, advanced converter reactors or in burners of fissionable uranium and plu- tonium that also use fluoride fuels. osolutions of uranium, plutonium, and. thorium salts in chloride and fluoride carrier salts offer attractive possibilities for mobile fuels for intermediate and fast breeder reactors. The fast reactors are of interest too but are not a significant part of the program. Our major effort is being applied to the development, construction, and operation of a Molten-Salt Reactor Experiment. The purpose of this Experiment is to test the types of fuels and materials that would be used in the thermal breeder and the converter reactors and to obtain several years of experience with the operation and maintenance of a small molten- salt power reactor. A successful experiment will demonstrate on a small scale the attractive features and the technical feasibility of these sys- tems for large civilian power reactors. The MSRE will operate at 1200°F and atmospheric pressure and will generate 10 Mw of heat. Initially, the fuel will contain 0.9 mole % UFy, 5 mole % ZrF,, 29.1 mole % BeF,, and 65 mole % LiF, and the uranium will contain about 30% 235U. The melting point will be 840°F. In later operation, highly enriched uranium will be used 1n lower concentration, and a fuel containing ThF, will also be tested. In each case the composition of the solvent can be adjusted to retain about the same liquidus temperature. The fuel will circulate through a reactor vessel and an external pump and heat exchange system. All this equipment is constructed of INOR-8,1 a new nickel-molybdenum-chromium alloy with exceptional re- sistance to corrosion by molten fluorides and with high strength at high temperature. The reactor core contains an assembly of graphite moderator bars that are in direct contact with the fuel. The graphite 1s a new material® of high density and small pore size. The fuel salt does not wet the graphite and therefore should not enter the pores, even at pressures well above the operating pressure. 1S0ld commercially as Hastelloy N and Inco No. 806. 2Grade CGB, produced by the Carbon Products Division of Union Carbide Corp. Heat produced in the reactor will be transferred to a coolant fuel 1n the heat exchanger, and the coolant salt will be pumped through a radiator to dissipate the heat to the atmosphere. A small facility is being installed in the MSRE building for occasionally processing the fuel by treatment with gaseous HF and F». Design of the MORE was begun early in the summer of 1960. Orders for special materials were placed in the spring of 1961. Major modifi- cations to Building 7503 at ORNL, in which the reactor is installed, were started in the fall of 1961 and were completed by January 1963. Fabrication of the reactor equipment was begun early in 1962. Some difficulties were experienced in obtaining materials and in making and installing the equipment, but the essential installations were completed so that prenuclear testing could begin in August of 1964. The prenuclear testing was completed with only minor difficulties in March of 1965. Some modifications were made before beginning the critical experiments in May, and the reactor was first critical on June 1, 1965. The zero-power ex- periments were completed early in July. Additional modifications, main- Tenance, and sealing and testing of the containment are required before the reactor begins to operate at appreciable power. This work should be completed in October, and the reactor should be at full power before the end of the year. Because the MoRE is of a new and advanced type, substantial research and development effort is provided in support of the design and construc- tion. Included are engineering development and testing of reactor com- ponents and systems, metallurgical development of materials, and studies of the chemistry of the salts and their compatibility with graphite and metals both in-pile and out-of-pile. Work is algso being done on methods for purifying the fuel salts and in preparing purified mixtures for the reactor and for the research and development studies. This report is one of a series of periodic reports in which we de- scribe briefly the progress of the program. ORNL-3708 is an especially useful report because it gives a thorough review of the design and con- struction and supporting development work for the MSRE. It also describes much of the general technology for molten-salt reactor systems. Other re- ports issued in this series are: ORNL-2474 Period Ending January 31, 1958 ORNL-2626 Period Ending October 31, 1958 ORNL-2684 Period Ending January 31, 1959 ORNL-2723 Period Ending April 30, 1959 ORNL-2799 Period Ending July 31, 1959 ORNL-28°90 Period Ending October 31, 1959 ORNL-2973 Periods Ending January 31 and April 30, 1960 ORNL-3014% Period Ending July 31, 1960 ORNL-3122 ORNL-3215 ORNL-3282 ORNL-3369 ORNL-3419 ORNL-3529 ORNL-3626 ORNL-3708 ORNL-3812 Period Ending Period Ending Period Ending Period Ending Period Ending Period Ending Period Ending Period Ending Period Ending February 28, 1961 August 31, 1961 February 28, 1962 Avgust 31, 1962 January 31, 1963 July 31, 1963 January 31, 1964 July 31, 1964 February 28, 1965 Part 1. MSRE OPERATIONS AND CONSTRUCTION, ENGINEERING ANATYSIS, AND COMPONENT DEVELOPMENT 1. MSRE OPERATIONS Chronological Account The principal accomplishments for the period from March through Au- gust 1965 were the preparation for, and completion of, the initial criti- cal and associated "zero-power' experiments. Initial criticality was achieved on June 1, and the experiments were concluded on July 3. The remainder of the period was used in preparing the system for operation at power. The initial, precritical operation of the reactor system (run PC-l), with flush salt in the fuel loop, was concluded on March 4, after the ex- periments on noble-gas behavior were completed. (The analysis of the re- sults of these experiments is discussed in Chap. 2, Component Development.) In this run, salt was circulated at high temperature for 1000 hr in the fuel loop and 1200 hr in the coolant loop. After the flush salt was drained from the fuel loop, it was transferred to the fuel storage tank, there to await processing to remove oxides. (The processing is described on page 152.) The next five weeks (until midaApril) were spent in advanced class- room training of the reactor operators and supervisors and the administra- tion of qualifying examinations. Before the nuclear experiments started, there were at least one engineer and one technician on each crew who had been qualified and certified. Others were certified later as they com- pleted individual oral examinations. While the operator training was in progress, final physical prepara- tions were made for zero-power nuclear operation. These included.: 1. 1installation and checkout of the fuel-salt sampler-enricher, 2. fTinal installation and checkout of the control rods and drives, 3. completion and checkout of the nuclear instrumentation and controls, 2 installation of a gamma-ray densitometer on the fuel-salt inlet line to the reactor, 5. installation of the lower layer of shield plugs on top of the reactor cell, 6. miscellaneous minor maintenance Jobs. The fuel "carrier" salt (a mixture of LiF, BeF,, and ZrF,) was charged into fuel drain tank NO.ZB(FD-Z), starting April 21. The con- tents of 35 shipping containers (4560 kg of salt) were melted and trans- ferred to the drain tank in six days. To this was added 236 kg of LiF- UF, eutectic containing 147 kg of 238y (depleted in 235U). The first operation with this barren salt was to obtain neutron counting rates with the salt at various levels in the core. Then, as a Tinal check on the operation of the equipment and to establish base lines for chemical analyses of the fuel salt, the carrier salt was circulated for ten days in run PC-2. Eighteen samples, taken through the newly in- stalled sampler-enricher, showed that the salt composition was as ex- pected. (See page 113.) This, coupled with satisfactory operation of all equipment, indicated that at last all was ready for the initial crit- ical experiment. The addition of #3°U was started on May 24, and initial criticality was achieved at 6:00 PM on June 1, 1965. The 255U’was added as the IiF- U, eutectic with highly enriched uranium (93%). The bulk of this ma- terial, containing 69 kg of 235U} was loaded in four charging operations to FD-2. After each addition the salt was transferred to the second drain tank (FD-1) and back again to ensure thorough mixing. The mixed salt was loaded into the reactor system after each charging operation, and count- rate data were taken at several salt levels in the core and with the re- actor vessel full. These data were compared with the barren-salt dats to monitor the neutron multiplication and to establish the size of the next addition. Extrapolation of inverse-count-rate plots with the re- actor vessel full showed that the loading after the fourth addition was within 0.8 kg 2327 of the critical loading when circulation was stopped and the control rods were withdrawn to their upper limits. The remainder of the #3°U was added directly to the circulating loop with enriching capsules. These were inserted into the fuel-pump bowl via the sampler- enricher to increase the loading 85 g at a time. Count rates were meas- ured after each capsule with the fuel pump off and the control rods with- drawn. The reactor became critical after the eighth capsule with the pump off, two rods fully withdrawn, and one poisoning 0.03 of its avail- able worth. After the initial critical condition was established, additional en- riching capsules were added to increase the uranium loading to the op- erating level. Enough excess reactivity was added in this way so that one control rod could be calibrated over its entire range of travel. The various zero-power experiments were performed during this phase of the operation. These included, in addition to the rod-calibration experi- ments, measurements of temperature coefficient of reactivity, uranium-concentration coefficient of reactivity, effects of fuel circulation on reactivity, effects of system overpressure on reactivity, bW dynamic characteristics. Use was made of the on-line digital computer for collecting data for some of these experiments even though the equipment was not completely checked out and in normal service. kxperiments specifically aimed at rod worth were stable period meas- urements and rod drop experiments. These were done with the fuel static and with it circulating. The results will give, as accurately as possible, the total and differential worths of the regulating rod (rod 1) over its entire travel with the other two rods fully withdrawn. In addition, worth values will be obtained for each of the three rods with the other two withdrawn and at intermediate positions. These will lead to evalu- ations of rod shadowing and "ganged" rod worth. After the initial critical experiment, another eight capsules were required before the reactor could be made critical at 1200°F with the fuel pump running (a consequence of the loss of delayed neutrons during circulation). Thereafter, we measured the critical rod position, with the pump running, after each capsule. At intervals of four capsules, we made period measurements with the pump running; then we turned it of'f, determined the new critical rod position, and made more period measure- ments. This continued until 87 capsules had been added. Three times during this experiment (after 30, 65, and 87 capsules), we observed rod drop effects. Period measurements were usually made in pairs. The rod on which the sensitivity was to be measured was adjusted to make the reactor Just critical at approximately 10 w; then it was pulled a prescribed distance and held there until the power increased about 2 decades. The rod was then quickly inserted to bring the power back to 10 w, and the measure- ment was repeated at a shorter stable period. Periods were generally in the ranges from 40 to 50 sec and from 70 to 120 sec. The available re- sults of these and the other zero-power experiments are discussed in the section Analysis of Operation. Most of the zero-power experimental program was carried out with the coolant system empty. However, some of the dynamic tests required cir- culation of the coolant salt. This loop was filled on June 20, and salt was circulated for 118 hr while the tests were in progress. The coolant loop was drained on July 1. The zero-power experiments were concluded, and the fuel loop was drained on July 4 after 764 hr of circulation in this run. The loop was then filled with flush salt, which was circulated 1.3 hr, sampled, and drained to prepare the system for maintenance. With the reactor shut down, the final preparations were started for operation at significant power. The principal jobs to be accomplished during this shutdown are: 1. modification of the coolant-radiator door assembly, 2. modification of the coolant-salt penetrations of the reactor contain- ment cell, 3. dnstallation of a new graphite-sample assembly in the reactor vessel, 4. removal and replacement of the fuel-pump rotary element for remote maintenance practice, closure and leak testing of the reactor containment, installation of stacked-block shielding. 10 Component and System Performance In general, the performance of the many mechanical components and auxiliary systems during operation was highly satisfactory. This is par- ticularly true in view of the fact that some of the items were being in- tegrated into the system operation for the first time. Some difficulties were encountered which caused temporary inconvenience, but no program de- lays resulted and no extensive modifications will be required to improve future performance. This section deals with the difficulties that were experienced, their actual and potential effects, and the changes which they incurred. In addition, some routine experience with selected com- ponents is discussed. Control Rods Two of the control rods and drives were installed during run PC-1 and were used in simulator training. Before PC-2 all three rods were installed and subjected to a test consisting of 100 cycles of full with- drawal and scram. The rods operated freely and never failed to scram, but occasionally the lower limit switches failled to clear properly as the rods were withdrawn. We found the cause to be galling in the cam actuator for the switch. After we installed Stellite bearing surfaces to remedy this problem, each rod was successfully raised and scrammed 30 times without any malfunction. (The lower limit switch on rod 2 at first stuck as before, and we found that a shim had been left out of the switch-actuator assembly. After the shim was replaced, there was no fur- ther trouble.) Operation continued throughout run 3 without trouble. Rod drop times were measured in the tests in PC-2 and in a geries of 40 scrams at the end of run 3. The results (Table 1.1) show that the drop times became slightly shorter and more uniform. This is con- sistent with development experience in breaking in new flexible rods. Table 1.1. Observed Drop Times of Control Rods . a Number of Timed Drops T eDrop Times éfseg) —— Rod 1 Rod 2 Rod 3 verag >tandar eviation Rod 1 Rod 2 Rod 3 Rod 1 Rod 2 Rod 3 Test Series Original 45 64 46 820 818 870 17 25 30 At end of 4l 4. 41 793 775 792 4.0 .1 3.7 run 3 aTim.e from actuation of the scram switch (with rods at 5l—in.'withdrawal) until actuation of lower limit switch (0 in.). Measurements of indicated rod positions with the lower end of the poison at the fiducial zero position were made after installation and at intervals during run 3. The data did not indicate any stretching of the rods. 11 Inspection of the control rods after run 3 showed that two were in very good condition, but there was severe damage at one point on the flex- ible extension tube of rod 3. (This extension, connecting the poison section to the drive assenmbly, consists of a flexible stainless steel tube covered with a braided stainless steel wire sheath.) A hole was worn in the sheath, and the tube convolutions were abraded at the point which was 1n contact with the lower roller when the rods were fully in- serted. The roller assembly was cut out of the thimble, and the roller was found to be worn and rough on one side, indicating that it had been stuck. Presumably, the roller was jammed when the housing was distorted, while the assembly was being welded in. The extension tube was replaced, and a new roller assembly will be installed. After run 3 the rod drives were also inspected. Modified limit switch actuators were installed, and the worm gears were replaced with new fully hardened, lapped gears which had a much smoother finish than the original gears. campler-Enricher The sampling of the circulating fuel-salt system during run PC-1 was done with a temporary samplerl which provided an inert atmosphere for the sample and prevented air from entering the fuel system. A total of 12 flush salt samples was taken with this sampler. The installation of the fuel-system sampler-enricher was completed during the shutdown period before run PC-2. Since that time, 53 samples were withdrawn, and 87 enriching capsules were added. Although several minor problems occurred during the reactor operation, none of these pre- vented the sampler from being operational. In some cases delays were in- curred during sampling or enriching, but the overall schedule was not significantly affected. ©GSpecific problems that were encountered are as follows: 1. Both the operational and the malntenance gate valves developed leaks through one of the two seats of each valve. 2. The removal valve leaked and required a greater-than-normal helium flow for buffering. 3. A solenoid valve on the removal-valve actuator failed. 4. The removal valve occasionally failed to close completely and had to be closed manually. 5. The access port periodically failed to close properly because of faulty operation of the clamps on the clamp actuators. 6. One sample capsule was accidentally dropped down the sample tube to the operational valve gate. 7. The drive motor stopped once during the removal of an empty capsule. The capsule was inserted about 12 in. and was then successfully with- drawn. The reason for this stoppage is unknown. 12 8. There were three boot failures on the manipulator, one of these in- volving both boots. 9. The manipulator arm and fingers were bent, causing some difficulty in gripping the latch cable and moving the manipulator arm. Most of the troubles occurred because this period was one of testing of equipment and training of operators. A mockup had been thoroughly tested, but the installation of the sampler-enricher on the reactor was completed just before the critical experiments were begun. The operators are now well trained, and some changes are being made to improve the re- liability of the equipment and the safety of the operation. The device 1s expected to perform satisfactorily during power operation. Freeze Valves The freeze valve problems of insufficient cooling air which were re- ported previously were corrected.?® In addition, the coolant system drain valves were modified to the same basic design as the fuel-system valves. The revised coolant valves now have sufficient heat capacity to thaw on loss of power. All the freeze valves except the system drain valves were relieved of the fast-thaw requirement by requiring that the valves to both fuel drain tanks be thawed while fuel is in the reactor. Previously one valve was to be normally frozen and was to thaw during an emergency drain. A proportional controller was added to the FV103 cooling-air supply after run PC-1 to maintain the freeze valve at a preselected temperature that would result in a suitable thaw time. This temperature and the valve temperature distribution were controlled satisfactorily at any steady fuel-system temperature level, but it was necessary to change the controller set points whenever the fuel-system temperature was changed. appreclably. The following thaw times were recorded for the drains after runs PC-2 and 3. Run Drain Thaw Time (min) PC-2 Carrier salt l6-l/2a 3 Fuel salt 10 3 Flush salt 18 ®System cooled to 1100°F. These thaw times compare with the 33-1/2 min which was required at the end of run PC-1. An incident occurred during run 3 which could have delayed an emer- gency drain of the fuel system had it been required. Electrical power to 13 the freeze-valve control modules was lost for a period of about 10 min be- cause the terminals of a temporary recorder were accidentally shorted. The loss of module power turned blast air onto the freeze valves. Normal op- eration of the valves was restored after module power was regained. Con- Ttrol circult revisions are being made to prevent the recurrence of this problem. Although the performance of the freeze valves was acceptable, it was still difficult to maintain the proper temperature profiles across the valves because one controller was used to supply heat to both shoulders of the wvalve. The heater control systems were revised after run 3 to provide separate controllers for each shoulder heater on the freeze valves that serve the fuel and coolant drain tanks and the fuel flush tank. Pro- portional controllers similar to that on FV1O03 were added to the cooling- alr supplies for the fuel and coolant drain-tank freeze valves (105, 106, 204, and 206). Freeze Flanges The five freeze flanges on the main fuel- and coolant-salt piping continued to perform satisfactorily. Leakage of the buffer (leak de- tector) gas was not excessive at any time. The buffer gas leakage rates measured at various times during the reactor operation are listed in Table 1.2. Table 1.2. Observed Leak Rates of Buffer Gas from Freeze Flanges Leakage Rate (stad cm3/sec) oystem freeze Initial Circulating Hot Drained System Cold System Cold After After Flange Heatup oalt After Run PC-1 Run 3 Run PC-1 X 1077 X 10™2 X 1073 X 1073 X 1077 100 2.0 0.57 Q.7 1.5 2.0 101 1.3 0.4 0.25 2 .20 1.2 102 0.5 0.3 0.30 2.33 1.0 200 1.0 0.21 0.41 1.48 0.3 201 0.6 0.22 0.21 1.23 1.8 The freeze flange leakages are normally monitored as a group,; leak- age of individual flanges would be measured if the group leakage were higher than normal. Fuel- and Coolant-System Pressure Control Difficulties were experienced in controlling the fuel- and coolant- system pressures within close limits because of accumulation of solids in the off-gas throttling valve used for fuel-system pressure control and in the filter just upstream of the coolant-system pressure-control valve. 14 During run PC-1 (Januvary-March 1965), when the coolant salt was circulated for 1200 hr, the coolant off-gas filter plugged and was re- placed twice. When the filters plugged, pressure was controlled at 5 * 2 psig by manual venting through a larger bypass valve. Inspection showed that the filter was covered with amorphous carbon containing traces of the constituents of the coolant salt and INOR-8. Before run PC-2 the filter was replaced with one having 35 times the surface area. Coolant salt was not circulated again until near the end of run 3 and then for only 118 hr. During this time the pressure control again became erratic, indicating obstruction of either the filter or the valve. Both were re- moved for inspection, and although there was no deposit on the filter, the valve was partially obstructed by a black, granular material. Rinsing with acetone restored the original flow characteristics of the valve, and 1t was reinstalled. The fuel-system pressure control became erratic near the end of run PC-1, and at the conclusion of the run the off-gas filter was removed. It was clean; so the pressure control valve was removed and found to be partially plugged. The obstruction was blown out with gas, and the valve was washed out with acetone. The valve then performed normally and was reinstalled. The acetone rinse was darkened and contained small (1—5 u) beads of a glassy substance. Fuel-system pressure control was satisfactory at the beginning of run PC-2, but within a week the valve began sticking again. This time it was replaced with one having a large CV (0.077 instead of 0.02). The original valve was cut open for inspection, and a black deposit was found partially covering the tapered stem. The deposit was about 20% amorphous carbon, and the remainder was the 1- to 5-p glassy beads, which proved to have the composition of the flush salt. The larger replacement valve in the fuel off-gas line gave adequate pressure control for the first four days of salt circulation in run 3. However, there were four occasions when it appeared to be sticking. When this happened, the pressure built up slowly to about 6 psig before the valve opened to drop it back to the normal 5 psig. For the next 20 days Tthe small pressure variations predominated, suggesting either that the valve was not functioning properly or that intermittent partial plugging was occurring. This condition cleared up abruptly, and during the last ten days of salt circulation the loop pressure remained completely stable. Since no corrective action had been taken and the valve was functioning properly when the system was shut down, no explanation for the earlier erratic. behavior could be established. The cause of the solids in the off-gas lines is not yet known. There is reason to believe that some carbon may have been introduced into the reactor with the salt, accumulated on the surface in the pump bowl, and carried into the off-gas line as a dust. O0il contamination of the salt system has also been suggested. The glassy salt beads in the fuel off- gas line are probably frozen droplets of mist caused by the stripper spray, but we do not know whether these were carried into the line con- Tinuously during operation or were swept out of the pump bowl by sudden venting. 15 Filters that are capable of removing 1l-u particles are to be installed in both the fuel and coolant systems. (The pore size of the original fil- ters was about 25 K.) Presumably, this will protect the valves from fur- ther accumulations. In any event, the coolant off-gas filter and pressure- control valve can be maintained directly after power operation; those on the fuel off-gas are designed for remote maintenance. Heaters and Insulation The two heater elements which failed during the initial precritical operation3 and their spares were replaced during the shutdown prior to run PC-2. In both these cases operation had been continued by using the installed spare elements. ©Six additional heater-element failures and an electrical ground at a disconnect were discovered during the checkout for run PC-2. 1In four of the heaters, the failure occurred at the junction of the lead-in wire and the heating wire inside the ceramic element. These elements and the electrical ground were repaired before startup, and the other two elements were left out of service for runs PC-2 and 3. During subsequent operation, one in-cell heater failure was noted, and one failure occurred at a heater power supply. (The latter was re- paired immediately.) 1t may be noted that a number of heater failures could occur without significantly affecting the operation of the reactor system. Unless a heater 1s in a particularly sensitive location, its failure may not be discovered until individual-element checks are made during a shutdown. Three more heater-element failures were found during the shutdown after run 3. In addition, a number of minor defects (grounds, improper resistances, damaged connectors) were found. All the defects, including those left from earlier operations, will be corrected before the next reactor startup. An important cause of heater-element failure has been separation of the lead-in from the heater wire. This is apparently due to a com- bination of a design weakness and excessive flexing during installation of the elements. The Jjoint was redesigned, and new elements which in- corporate the change are being installed where such failures occurred. Reliable Instrument Power System Motor-generator sets 1 and 4 and the 250-v battery system normally supply power to the fuel and coolant oil pumps and to the instrumentation system. MG-1 is an ac-to-dc set which supplies 250-v dc power to drive MG-4 and to charge the 250-v emergency batteries. MG-4 operates from either MG-1 or the 250-v batteries and normally supplies 120-v ac power to the instruments and to the fuel and coolant oil pumps. The initial operation of these MG sets was unreliable because of failures in electrical control components. The MG sets are used ones that were installed in the building for earlier experiments, and most of the failures resulted from aging of the control components. 16 The MG sets were repalred prior to run PC-2 and were placed back in service. Minor problems occurred, which were corrected by adjustments to the voltage controls. Both MG sets were placed in service under normal load at the begin- ning of run 3, and both performed satisfactorily for the duration of the run. MG-4 was shut down after the completion of run 3 because of the failure of an internal cooling blower. The blower motor was replaced to make the set operational. Thermal-Shield Cooling Water The thermal-shield water piping was modified to provide an adequate flow through the three removable segments. The three segments were con- nected in series, and a line was connected directly to the cooling water supply. Pressure regulators and pressure-relief valves were installed in the supply lines to both the main thermal shield and The removable slides to avold overpressuring the system. The above changes provided flow rates of 65.8 and 4.6 gpm to the main shield and to the removable segments respectively. An inspection of the thermal shield while the fuel system was heated indicated that the removable segments were adequately cooled by these modifications. Fuel-Pump Overflow Tank A continuous, but very slow, accumulation of salt in the fuel-pump overflow tank was observed throughout the operation of the fuel loop. In 1000 hr of circulation in runs PC-2 and 3, 39 kg of salt was collected with the pump-bowl level 3 in. below the overflow point. This salt was recovered Jjust before the run 3 shutdown. Accumulation at this rate will not significantly affect the operation of the reactor. Analysis of Operation Bagic Nuclear Characteristics The nuclear characteristics measured during the zero-power tests generally agreed quite well with the predicted values. Analysis of the data is still in progress, particularly with regard to control-rod worth and dynamic characteristics. The results that have been obtained so far are described below. Critical Concentration. Predicted and observed <3°U requirements for criticality are compared most logically on the basis of volumetric concentration. The required volumetric concentration is nearly invariant with regard to the fuel-salt density (unlike the mass concentration, which varies inversely with density) and depends not at all on system volume or total inventory. 17 The observed *3°U concentrations are on a welght basis, obtained from either inventory records or from chemical analyses. These weight concentrations must be converted to volumetric concentration by multi- plying by the fuel-salt density. The amounts of ?2°U and salt weighed into the system gave a *3°U weight fraction of 1.42% at the time of the initial criticality. The chemical analyses during the precritical op- eration and the zero-power experiments gave uranium concentrations which were 0.985 of the "book" concentrations. (Part of this discrepancy, about half we believe, is due to dilution of the fuel with flush salt left in freeze valves and drain-tank heels when the fuel salt was charged.) Applying this bias to the book concentration at criticality gives an "ana- lytical"™ #7°U weight fraction of 1.40%. We now believe that the density of the fuel salt at 1200°F is about 145.5 1b/ft3. This is the preliminary result of recent laboratory measurements of density, and it agrees with measurements made in the reactor using the two-point level probes in the drain tanks. DHRarlier measurements in the reactor, using the drain-tank welght indications and the volume of salt believed to have been trans- ferred into the fuel loop, gave 136.6 lb/ftB. Corrections must be applied because the initial critical conditions were not exactly the same as those assumed in the predictions. The core temperature was 1181°F instead of 1200°F, and the control rods were poi- soning 0.118% 8k/k instead of none. (Two rods were at meximum withdrawal, 51 in., and one was at 46.6 in.) The predicted 23°U concentration for criticality at the reference condition was 32.87 g/liter; corrected to the actual conditions, 1t is 33.06. This predicted value is compared with observed concentrations in Table 1.3. Table 1.5. Comparison of Critical 2327 Concentrations (1181°F, pump off, 0.118% dk/k rod poisoning) 235U 235U . Fuel Density . Concentration (lb/ft3) Concentration (wt %) (g/1iter) Predicted 33.06 Book 1.42 145.5 33.10 136.6 31.07 Analytical 1.40 145.5 32.60 136.6 30.60 1f, as we estimate, the true concentration was about halfway be- tween the book and the analytical and the density is about 145.5 lb/ft3, the actual concentration was extremely close to the prediction. Laboratory measurements now in progress should confirm the value of the fuel-salt density. The uncertainty between book and analytical con- centrations will be reduced as a result of the next startup, when analysis of the fuel after another flushing, draining, and refilling will help us evaluate the dilution effect. 18 Control Rod Worth. The only data relative to control rod worth that we have finished analyzing so far are the rod-bump, period measurements on rod 1 with the fuel static and the other two rods at their upper limits. These data were used to produce the curve of rod sensitivity as a function of position shown in Fig. 1.1. Because rod worth is affected by the 235y concentration in the core, it was necessary to apply theoretical correc- tions to the measured sensitivities to put them all on the basis of one concentration. The points in Fig. 1.1l were corrected to the initial crit- ical concentration, where the sensitivity is the highest. The correction factors which were applied increase linearly with 237 concentration to a maximum of 1.087 at the final concentration (the points between 1 and 2 in.). Had the points been corrected to the final concentration, the curve would have been lower by 8.7%. Figure 1.2 shows a curve of rod effect vs position at the initial concentration which is the integral of the differential-worth curve in Fig. 1.1. The curve for the final concentration is simply the first curve reduced by a factor of 1.087. The predicted worth of this rod at the initial critical concentration was 2.2%. We are working on the analysis of the period measurements with the fuel circulating. As will be discussed later, a new mathematical treat- ment of delayed-neutron effects in the MSRE has been developed; this will be used to relate period to reactivity. The sensitivities determined in this way should, of course, coincide with the results of the static-fuel measurements. Because of the difficulty of accurately treating the com- plex pattern of precursor distribution in the circulating fuel, the values obtained with the fuel static should be more reliable. ORNL- DWG 65-8033 0.07 : 006 1= T e T 0.05 | - . . 4 e \ o .7 ! N | - - \. . / . 003 | — ¢ e e Y, N ./ o O 2 N\ DIFFERENTIAL WORTH [(% 84/4)/in] N N | o = | | | \ \ \ | O 4 8 12 16 20 24 28 32 36 40 44 48 52 ROD POSITION (in.) O Fig. 1.1. Differential Worth of Control Rod No. 1, Adjusted to Initial Critical Loading. 19 ORNL-DWG 65-8034 e T T 2.0 — AR i | | \ \ 1.8 1 . | I I ’ o el . CRITICAL LOADING | (61.52 kg 235U IN LOOP) 2/ | e | | | | | | ! | | } | - | » | | ST FINAL LOADING N x > 1 0 & 235 ~ (67.98 kg U IN LOOP) T ? | - i o O = > T N = = }_. - @) < (N R (et 20 24 28 32 36 40 44 48 52 ROD POSITION (in.) Fig. 1.2. Integrated Worth of Control Rod No. 1. The rod-drop experiments are also being analyzed. They will give independent values of rod worth and may show the effect of 23°U concen- tration on rod worth. This latter effect was also the subject of multi- group calculations at the concentrations observed in the experiment. The correction factor used here is the result of these calculations. 2327 Concentration Coefficient of Reactivity. The 232U concentration coefficient of reactivity is §iven'by the ratio of the change in reactivity to the fractional change in 235U concentration (or circulating mass) as a result of a small addition. The effect of each capsule addition (after initial criticality) on the critical position of the control rod was de- termined with the pump running. The critical position with the pump off was measured after every fourth addition. Rod positions were converted to reactivity, using Fig. 1.2 and correcting for the concentration effect on rod worth. Results are shown in Fig. 1.3. The slope of a curve in Fig. 1.3 at any concentration multiplied by that concentration 1s the desired concentration coefficient of reactivity, (8k/k)/(8m/m). The coefficient obtained in this way is 0.226, independent of concentration. The coefficlient predicted from multigroup criticality searches about the minimum critical concentration was 0.248 (ref. 4). 20 ORNL-DWG 65-803¢ 2.4 e 2 e ——— o S e 37 08’ FUEL NOT CIRCULATING ¢ | _&*° 16 T . /;'-.? i;;.?ag e ] = e 2 ./ <" ~X- 1 2 - - ’;77’{“ ® R - - T E ./ ,o".. o ~" " FUEL CIRCULATING /, .‘.,oo 0.8 LA - - — /. ... //. ,o” 0.4 7 ./.‘.‘.; - S I . /: .... o8 ..3 O ./ W .'..’ 61 62 63 64 65 66 67 68 69 MASS OF 235U IN FUEL LOOP (kq) Fig. 1.3. Effect of *2°U Mass on oystem Reactivity. Reactivity Effect of Circulation. The reactivity effect of circula- tion, given by the difference of the two curves in Fig. 1.3, is —(0.212 = 0.004)% 6k/k. The effect of changes in delayed-neutron precursor distri- butions with circulation had been predicted to be —0.30% 8k/k.%»© Another ~0.2% dk/k was expected because of entrained bubbles of helium in the cir- culating salt. As will be discussed later, the evidence shows that there were practically no circulating gas bubbles except for a brief period when the fuel level in the pump bowl was lowered. Therefore the gas effect attending circulation was practically nil. The difference between the predicted and observed delayed-neutron losses was apparently due to inadequate accounting for delayed neutrons emitted just outside the graphite region of the core in the upper and lower heads. A more realistic model has been developed to account for these effects. The program computes precursor distributions under both steady-state and transient conditions, taking into account mixing, ve- locities, volume fractions, and flux distributions in each of the prin- cipal regions. The welghted contributions of the delayed neutrons from each group are computed, taking into account the initial energies of the neutrons and the nuclear importance of each region. The result is s "circulating-fuel inhour equation" for the MSRE whose uses will include the analysis of the circulating-fuel period measurements and, as a spe- cial case, the steady-state effect of circulation. Application of this equation to the steady-state condition gave a reactivity effect due to circulation of —0.22% 8k/k. Temperature Coefficients of Reactivity. We measured the effect of temperature on reactivity by adjusting the electric heaters to change the system temperature slowly (about l5°F/hr) while we observed the critical 2l position of the regulating rod. This experiment gave the overall tempera- ture coefficient, that is, the sum of the fuel and graphite coefficients. We alsc attempted to separate the fuel (rapid) and graphite (Sluggish) coefficients by an experiment in which the coolant system was used to in- crease the fuel-salt temperature rather abruptly. Figure 1.4 shows results of the three experiments involving slow changes in temperature. The first experiment, with 68 kg of %3°U in cir- culation, gave a line whose slope ranges from —(6.6 to 8.3) x 10™° (°F)~L. At 70 kg the experiment gave a straight line with a slope of —7.24 X 107° (°F)~t. In the last experiment, at 72 kg, the slope of the curve above sbout 1140°F is —7.3 X 1072 (°F)~t. A value of =7 X 10™° (°F)~1 nad been predicted: —3.3 X 1072 (°F)~! for the fuel and —3.7 X 10~° (°F)~! for the graphite. The fuel-salt coefficient of thermal expansion used in this calculation was obtained by an empirical correlation of density and com- position of salts other than our fuel salt. Recent measurements of fuel- salt density gave a higher thermal-expansion coefficient, leading to a calculated fuel temperature coefficilent of reactivity of —5.6 X 10™° (°F)"1. The new value for overall temperature coefficient is —9.6 X 107° (°F)~t. The observed coefficient is in better agreement with the earlier prediction (on which the safety analysis of the MSRE was based) . The density measurements are belng reviewed and will be checked by an independent method of density determination. The experiment at 72 kg 232U shows a lower slope below about 1140°F. We do not believe that the temperature coefficient is lower in this range; we believe that another phenomenon became significant during this part of the experiment. This phenomenon was the appearance of an increasing amount ORNL-DWG 65-8032R 16 .- . b L [ - i,,,,, fi,,:’ip - ] e 67.86 kg 23°U IN LOOP ,.¢° 1.4 - - - e e . /o‘ CD/O $o b — A £ 10 |- -~ —t - N / ..J 3 - al S 0.8 — — e b e et o ] 235 ° N 71.71kg U IN LOOP e o 7 &7 06 | 4 — e — - - -~ -~ ° o® 235 oa | T weasskg 2PU N Loop ./ o/ " 0.2 e —_— - - e B e 0 1000 1050 1100 1150 1200 1250 FUEL TEMPERATURE (°F) Fig. 1.4. Effect of Fuel Temperature on Reactivity. 22 of helium bubbles in the circulating salt as the temperature was lowered. The evidence for this is discussed in the next section. The effect, so far as the temperature experiment is concerned, was that the bubbles tended to reduce the amount of fuel salt in the core, compensating to some extent for the increase in density of the salt itself as the temperature was low- ered. Thus the slope of the lower part of the curve cannot be interpreted as a temperature coefficient of reactivity in the usual sense. The hot-slug transient was done by stopping the fuel pump, raising the temperature of the circulating coolant salt and the stagnant fuel in the heat exchanger, and then restarting the fuel pump to pass the hotter fuel salt through the core. The output of a thermocouple in the reactor- vessel outlet, logged digitally at 1/4-sec intervals, showed a brief in- crease of 5 to 6°F as the hot salt first passed. It then leveled at about 3.5°F for a few loop transit times before decreasing gradually. The noise in the analog-to-digital conversion (£1°F) limited the accuracy of the measurement, but by taking an average of 50 points during the level period after mixing and before the graphite temperature had time to change sig- nificantly, a value was obtained for the step in fuel temperature. Re- activity change was obtained from the change in rod position, corrected for the decrease due to circulation, and ascribed to the fuel temperature increase. The result was a change of —=(4.9 + 2.3) x 10~° (°F)~Lt. Pre- dicted values of the fuel temperature coefficient lie in this range. We propose to repeat this test with the thermocouple signals biased and am- plified to reduce the effect of noise in the analog-to-digital conversion. More precise results can be expected in this case. Effect of Pressure on Reactivity. We performed three tests to ex- plore the effect on reactivity of changing system overpressure. Theo- retical considerations had indicated that for slow changes a very small, possibly negative, pressure coefficient of reactivity could be expected, but for rapilid changes the coefficient would be positive. The existence of any pressure coefficient was based on the assumption that undissolved helium would be entrained in the circulating fuel. In each of the three tests the loop overpressure was slowly increased from the normal 5 psig to 10 to 15 psig and then quickly relieved, through a bypass valve, to a drain tank that had been previously vented to atmospheric pressure. The first two tests were carried out at normal system temperature with the normal operating level of salt in the fuel-pump bowl. No change in control rod position was required to maintain criticality, and no significant change in pump-bowl level was observed during either of the tests. These indicated that the pressure coefficient was negligibly small and that essentially no helium bubbles were circulating with the salt. IFurther evidence of the lack of circulating voids was obtained from a gamma-ray densitometer on the reactor inlet line; this instrument showed no change in mean salt density during the tests. The third test was performed at an abnormally low pump-bowl level, which was obtained by lowering the operating temperature to 1050°F,. Fig- ure 1.5 shows the pressure transient and the responses of the regulating control rod, densitometer, and fuel-pump level during the rapid pressure 23 ORNL-DWG 65-8074R FUEL PUMP LEVEL (%) OVERPRESSURE (psig) DENSITOMETER CONTROL ROD POSITION (in.) O 4 8 12 16 20 24 28 TIME (min) Fig. 1.5. Conditions During Ra@id Pressure Release While Cir- culating Helium Bubbles. release. Independent evaluations of the void fraction from these three parameters all gave values between 2 and 3% by volume. The frequency re- sponse characteristics of the effects of pressure on reactivity were cal- culated from the pressure and rod motion and are shown in Fig. 1.6, along with the predicted high-frequency response for a void fraction of 1.2%. Extrapolation to the observed curve gives a void fraction of 2 to 2-1/2%. The low- and high-frequency pressure coefficients were +0.0003 and +0.014 (% 6k/k)/@si, respectively, for this particular condition. 24 ORNL-DWG 65—-8904A 1073 ® FREQUENCY VALUE o PREDICTED FOR 1.2 % VOID MN 1076 0.0001 0.004 0.01 0.1 1 10 w, FREQUENCY (radians/min) Fig. 1.6. Reactivity-Pressure Transfer Function with 2 to 3% Void Volume in Fuel. Calculated from pressure release experiment using Samulons method with 0.2-min sampling interval. Dynamic Tests We performed a variety of dynamic tests during the operation at zero power. These tests were the start of an extensive program to evaluate experimentally the inherent nuclear stability of the MSRE at all power levels. The reactor system has been analyzed on a theoretical basis, and the tests are designed not only to characterize the present system but also to evaluate the techniques and mathematical models used in the theoretical analysis. Preliminary results from the analysis of the zero- power tests are presented below. Frequency Response Measurements. A series of tests was run to de- termine the frequency response of neutron level to reactivity perturba- tions. These experiments included pulse tests, pseudorandom binary re- activity-perturbation tests, and measurements of the inherent noise in the flux signal. Tests were run with the fuel pump on and with it off. Noise measurements were also made during the special run with low pump- bowl level, when there were entrained bubbles in the core. The frequency response is a convenient measure of the dynamic char- acteristics of a reactor system. Classically, the frequency response is obtained by disturbing the reactor with a sinusoidal reactivity per- turbation and observing the resulting sinusoidal neutron-level variations. The magnitude ratio i1s defined as the ratio of the amplitude of the output sinusold to that of the input sinusoid. The phase angle is defined as the phase difference between the output sinusoid and the input sinusoid. Other procedures, such as those described in this report, can be used to yleld the same results as the classical method with less experimental effort. 25 The zero-power frequency response tests serve to check the theoret- ical zero-power frequency response predictions, but they do not furnish direct information on the stability of the power-producing reactor. The zero-power tests, however, do serve as an indirect, partial check on the at-power predictions because the dynamic behavior at power is simply the zero-power case with the addition of feedback from the system. Thus, ver- ification of the zero-power kinetics predictions lends some support to the predictions regarding power operation. In the pulse tests a control rod was withdrawn 1/2 in., held there for 3-1/2 or 7 sec, and then returned to its original position. The rod was placed so that this rod motion caused a change in kgope of about 0.03%. The rod-position signal and the flux signal were recorded digitally at 0.25- sec 1ntervals using the MSRE on-line digital computer (referred to in Figs. 1.13 and 1.14 as "logger"). The frequency response Was obtained by numer- ical Fourier transformation of the input and output signals. The pseudorandom binary tests consisted of a gpecially selected series of positive and negative pulses. The series contained 19 bits each with a duration of 7 sec and was repeated several times so that the initial transients could die out. The rod motion about the average position cor- responded to a Mkgrp of about *£0.015%. The rod-position signal and the flux-level signal were recorded digitally at 0.25-sec intervals using the MSRE computer. The frequency response was obtained by two different methods. The direct method used a digital filtering technique to obtain the spectral density of the input and the cross spectral density of the input and the output. The frequency response of the system is the ratio of the cross to input power spectral densities. The indirect method in- volved calculation of correlation functions and subsequent numerical Fou- rier transformation. Both methods gave essentially the same results. The nolse measurements and analyses used direct analog filtering of an analog tape record of the inherent noise in the flux-level signal. The tests required 1 hr of data recording to give statistically accurate re- sults. These tests were hampered by the unfavorable location of the de- tector and the resulting low flux-signal level at low power QYlO'w). Preliminary results of the frequency response measurements are shown in Figs. 1.7 through 1.11, along with theoretical predictions for the runs with the fuel pump on and with the pump off. As shown on the legend in the figures, several different procedures were used to obtain thesgse re- sults, but this represents only a partial analysis of the available data. All the experimental points except for the noise-analysis results are in absolute units (fractional change in neutron flux per change in keff)' The noise-analysis results are based on the assumption of a white-noige input and contain an unobtainable proportionality constant. Thus, the noise-analysis data were arbitrarily multiplied by the factor required to give agreement with theoretical results at a frequency of 9 radians/ sec. Figure 1.12 shows the noise-analysis results for operation at a re- duced pump-bowl level, which caused increasgsed bubble entrainment in the 20 fuel salt. Comparison of the nolise spectrum obtained with bubbles cir- culating (Fig, 1.12) and the noise spectrum obtained with no bubbles cir- culating (Fig. 1.7) indicates that the bubbles increased the amplitude of the power spectral density significantly in the 1- to lO—radian/sec region. Previous experiments with the MSRE-core hydraulic mockup indi- cated that random, hydraulically induced pressure fluctuations in the core would probably cause a significant modulation of the core void frac- tion, thus causing reactivity fluctuations. Hence, additional flux noise in this frequency range was expected, although it was not possible to pre- dict the "shape' or characteristics of this spectrum. ORNL—DWG 65—8905A 104 SN/, Sk/ kg ® BINARY TEST — INDIRECT ANALYSIS A BINARY TEST — DIRECT ANALYSIS o NOISE ANALYSIS 10 0.004 0.01 0.1 4 10 100 w, FREQUENCY (rodicns/sec) Fig. 1.7. Frequency Response — Magnitude Ratio of (SN/NO)/ (6k/ko) with Fuel Circulating. Results of binary and noise tests. ORNL-DWG 65-8906A 104 ® TEST NO.1 © TEST NO. 2 103 ‘SAUM@ S/ kg 102 10" 0.001 0.0 0.1 1 10 100 w, FREQUENCY (radians/sec) Fig. 1.8. Frequency Response — Magnitude Ratio of (BN/NO)/ (8k/ko) with Fuel Circulating. Results of two 7-sec-pulse tests. 27 The information presented in Figs. 1.4 through 1.8 was obtained from the data by stralghtforward analysis, using basic techniques. The pulse test data were filtered prior to analysis, but the binary test data were used in unmodified form. A more sophisticated analysis, including smooth- ing and trend removal, is being made, but little change in the results is expected. The noise test results will be corrected to remove the in- fluence of the tape-recorder characteristics, which may be significant at frequencies below 6 radians/sec. ORNL—DWG 65—-8907A 10 o 7—sec PULSE, TEST NO.1 A 7—sec PULSE,TEST NO.2 o BINARY-INDIRECT ANALYSIS -10 s BINARY-DIRECT ANALYSIS PHASE (deg) —-80 —90 —-100 0.001 0.0/ 01 1 10 w, FREQUENCY (radians/sec) Fig. 1.9. Frequency Response — Phase of (5N/NO)/(6k/ko) with Fuel Circulating. Results of pulse and binary tests. ORNL—DWG 65—8908A O 7-sec PULSE, TEST NO.1 A 7—sec PULSE, TEST NO.?2 a 3-1/p—sec PULSE, TEST NO.3 ® 2—'/>5—sec PULSE, TEST NO.4 O NOISE ANALYSIS 0.001 0.01 0.1 1 10 100 w, FREQUENCY (rodions/sec) Fig. 1.10. Frequency Response — Magnitude Ratio of (5N/NO)/ (6k/ko) with Fuel btatic. Results of pulse and noise tests. 28 ORNL-DWG 65-8909A 20 -20 -40 PHASE (deg) e 7sec PULSE, TEST NO. 1 60 4 Tsec PULSE, TEST NO.?2 5 3% sec PULSE, TEST NO.3 _80 ° 3% sec PULSE, TEST NO. 4 -100 0.001 0.01 0.1 1.0 10 w, FREQUENCY (radians /sec) Fig. 1.11. Frequency Response — Phase of (6N/NO)/(SK/RO) with Fuel Static. Results of pulse tests. ORNL-DWG 65-8910A 10° T T T T i T T T T ] R S SN/ Ny | THE FREQUENCY RESPONSE S_k/T IS PROPORTIONAL 5 _ TO A/PSD IF THE DISTURBANCE IN 84 IS WHITE NOISE | 51—y —— - ® . _ | O N g - [ | 5 _ [ ] ® 9 o N ° 102 4 2 5 10 20 50 100 w, FREQUENCY (radians/sec) Fig. 1.12. Frequency Response — Square Root of Power Spectral Density (PSD) of Flux oignal — Fuel Circulating with Bubbles. Re- sults of noise analysis; normalization same as for noise data in Fig. 1.7. The experience with these "zero-power" dynamics tests has led o plans for more tests with some improvements in data recording. Transient Flow-Rate Tests. The purposes of the transient flow-rate tests were: (1) to obtain startup and coastdown characteristics for fuel- and coolant-pump speeds and for coolant-salt flowrate; (2) to infer fuel- salt flow-rate transients from the results of (l); and (3) to determine the transient effects of flow changes on reactivity and void fraction. 29 Figures 1.13 and 1.14 show the fuel-pump speed, coolant-pump speed, and coolant-salt flow rate vs time for pump startup and coastdown. Data were taken with the computer and with a Sanborn oscillograph. The output of a differential-pressure cell across the coolant-salt Venturi was re- corded directly on the oscillograph, and the square root of that signal was taken as flow rate. The lag in the response of the computer flow signal is due to the response characteristics of the emf-to-current con- verter and the square-root converter between the differential-pressure transmitter and the computer input. PERCENT OF FULL SCALE ] ORNL—DWG 65—8911A 120 100 —o-a= 5 O—O0—O0— FUEL e qp—‘—.—.-fi 8—0—-0‘ 8-g—9-8 e §==o PUMP /COOLANT ’O,o-"’)/),ozo’( W SPEED PUMP o®’ O/O/ z[l 80 e 1.0 SPEED !‘v’., | @] PYd 3 7/ / ¥ / — o > /° a4 L ® [ 60 f—f /- 5 s T - /g> ./ />'COOLANT FLOW RATE m © e / E—:) 40 / o ( i O ) / / / ’ / 20 —e45 / o LOGGER /° /" ® OSCILLOGRAPH ¢ / | \ ’ 0 §—o—G%- 0 1 2 3 4 5 6 7 8 9 10 TIME (sec) Fig. 1.13. Pump Speed and Flow Startup Transients. ORNL-DWG 65-8942A 100 ¢mgz— J } . NN | | | | \ COOLANT SALT FLOW RATE o LOGGER k\\ %\:E\\\ e OSCILLOGRAPH COOLANT PUMP SPEED ;;;fif D O 0 i 1 O g ‘ 3 L - O 2 4 o 8 10 12 14 16 18 20 TIME (sec) Fig. 1.14. Pump Speed and Flow Coastdown Transients. 30 It was hoped that the coclant-pump speed and coolant flow rate would coast down in unison so that the fuel flow-rate coastdown could be inferred directly from the fuel-pump speed coastdown curve. This was not the case, however. Other methods of analysis will be attempted later. Reactivity effects of fuel flow-rate transients were measured by letting the flux servo controller hold the reactor critical during the transients; the reactivity added by the rod is then equal (and opposite) to the reactivity change due to the flow perturbations. The data for the pump startup were taken on the computer but were inadvertently erased; the reactivity transient for the pump coastdown was recorded. Since there were no voids in the fuel loop during normal operation, this transient 1s due entirely to flow effects on delayed neutrons. A further analysis of this curve will be made to try to determine the flow coastdown tran- sient and to check on the model used to represent the circulating delayed- neutron precursors. Conclusions from Dynamic Tests. The two most significant conclusions to be obtained from the dynamic tests are: (1) the information obtained gives no indication of the existence of inherent characteristics that might lead to operating difficulties in the low-power runs, and (2) the selected tests were, on the whole, quite satisfactory. These tests gave results which show good agreement with theoretical predictions, giving increased confidence in the theoretical model and in the predictions for stable power operation. ©Since the zero-power tests of this type are al- ways more difficult than power-level tests, very good results are expected from later tests. Undissolved Gas in Fuel Loop Although the nuclear experiments showed that there is no undissolved gas clrculating in the fuel loop when the salt is at the normal level in the fuel-pump bowl, the behavior of the pump-bowl level on stopping and starting the pump suggests the presence of a small compressible volume somewhere in the loop. Under normal conditions the level change corre- sponds to 0.4 ft° of gas uniformly distributed around the loop or 0.23 ft> located at the point of maximum pressure. The only known trapped gas is about 0.14 ft2 in the annuli at the reactor access nozzles. At- tempts to resolve this anomaly have been inconclusive. oalt Density and Inventory Experience The weigh cells on all the salt drain tanks were calibrated with lead weights shortly after the equipment was installed and before the Tanks and connected piping were heated. Additional data were obtained with the tanks hot during various salt-charging and transfer operations. These data served both to recalibrate the weighing systems and to give a measure of the density of the salts at operating temperature. Through- out the operation, salt inventories were computed from weigh-cell readings, using scale factors and tare corrections obtained from the calibration tests. 31 Cold and hot calibration of the coolant drain-tank weigh cells gave scale factors differing by less than 0.5%. Fuel drain tank 2 (FD-2) was calibrated hot twice, with flush salt and with fuel carrier salt; scale factors were within 0.2% of each other but were about 4% higher than the original, cold calibration. The reason for this discrepancy has not been established. The coolant-salt density was measured in the reactor by three dif- ferent methods; values ranged from 121.3 to 122.3 lb/ft3 at 1200°F, with an average of 121.9 1b/ft>. When flush salt, which is identical with coolant salt, was charged into ¥FD-2, the amount of salt added between two level probes indicated a density of 124.5 1b/ft2. A density of 120.9 lb/ft3 was computed from the pressure required to 1ift salt from the drain tank to the fuel loop. Weigh-cell indications of flush-galt den- sity, using the "hot" calibration factor, ranged from 123.2 to 131.5 1b/ft> at 1200°F; the "cold" calibration factor would have given 118.4 to 126.3 lb/ft3. The data on coolant- and flush-salt densities thus tend to support the "cold" calibration factor for FD-2. The density of the fuel carrier salt (65 LiF—30 BeFo—5 ZrF,) was measured as the salt was being charged to FD-2. This measured density, computed from externally measured weights and the volume between the level probes in FD-2, was 140.6 1b/ft> at 1200°F. Addition of all the uranium added through run 3 would be expected to increase the density by about 5.3 1b/ft>. Four measurements were made after the uranium was added, using- the weigh cells and the level probes. Densities based on the "hot" calibration of the wei§h cells ranged from 149.9 to 152.2 1b/ft>, with an average of 151.0 1b/ft> at 1200°F. With the "cold" scale factor the av- erage was 145.1 1b/ft°, very close to the expected density. oalt densities were computed on several occasions from the change in welgh-cell readings as the fuel loop was filled. In every case the com- puted density was less than given by other means, suggesting that a full loop volume may not have been transferred. The temperature coefficients of density for the salts were computed from the change in salt level with loop temperature. Measured values of (Ap/p)/bm were: for the coolant salt, —1.06 X 1074 (°F)—1L (average of three measurements); for the flush salt, —1.15 X 10—% (°F)“l; and for the fuel salt, —1.09 X 107% and —1.15 X 107% (°F)~! (two measurements). The bulk of the inventory data accumulated to date is on the flush salt, because more transfer and fill-and-drain operations have been done with this salt. Calculated inventories (using "ot" scale factors) have ranged from 1.7% below to 2.6% above the nominal or "book" inventory for no ascribable reason. Ixternal Neutron Source The MSRE fuel contains an internal neutron source (from the inter- action of 224U alphas with beryllium and fluorine) that provides about 4 X 10° neutrons/gec in the core when the reactor vessel is full. This 32 source fulfills all the safety requirements imposed on a reactor neutron source. However, it is highly desirable to monitor closely the operation of filling the reactor to ensure an orderly sequence of operations. This requires an external neutron source that will produce significant counting rates on the nuclear instruments before the filling operation is started. The geometry of the MSRE is such that an extremely strong neutron source is needed to satisfy this requirement. (The source tube is in the thermal shield on the opposite side of the reactor from the neutron detectors.) The Am-Cm-Be source that was procured was expected to produce about 10 counts/sec on the fission chambers with the reactor vessel empty of salt. This would have given the source a useful life of one year, with- out regeneration, for monitoring reactor fills. (A minimum of 2 counts/ sec has been specified as a criterion for starting a fill.) However, the source, as delivered, produced only 2.5 counts/sec,'which made 1t adequate for the initial operation but inadequate for future fills. oince an even stronger source is impractical, because of handling dif- ficulties and cost, it is expected that some instrument changes will be made to permit adequate monitoring of the early stages of reactor fills. Instrumentation and Controls Design and Installation General Ixcept for a small amount of work on the Fuel Processing oystem sampler, completion of documentation, and a few minor additions and re- visions, the design, installation, and checkout of the MSRE Instrumenta- tion and Controls System are now complete. During the period since the last report7 the design, installation, and checkout of all instrumenta- tion and controls systems required for power operation of the reactor were completed, the data-logging and processing system, referred to as the data-logger—computer system, was installed and checked out, several re- visions and modifications were made to improve performance or correct errors in design, and some safety instrumentation and associated control circultry were added. Documentation of design and as-built changes is nearing completion. In general, performance of the instrumentation has been very good. Some failures and malfunctions have occurred; however, once the causes of the failures were determined, these troubles were easlly corrected, and no major changes in instrumentation or in design philosophy have been necessary. Design, Installation, and Checkout Vapor-Condensing System. Design and procurement of instrumentation for the vapor-condensing system were completed. Instrumentation was pro- vided to monitor the water level, pressure, and temperature in the tanks. Commercial devices were used for all applications in this system; how- ever, since the vapor-condensing system is part of the secondary contain- ment, some special design and development were required to maintain the integrity of the containment. Water level is monitored by a four-position 33 float-type switch. DPressure is monitored by a local gage and pressure switch. Temperature is monitored by thermocouples. Abnormal conditions produce alarms at the vapor-condensing system and in the main control room. A dip tube was installed for use in continuous measurement of water level during shutdown. Installation of this system is in progress and will be complete before the start of power operation. Cell Air Oxygen Analyzer. Design, procurement, and preliminary checkout of a cell air oxygen analyzer system were completed. This sys- tem, which will be used to measure the amount of air leakage into-the nitrogen-filled secondary containment system, is required to detect a change of 0.02% (by volume) of oxygen over the range of 4.9 to 4.1%. The instrument supplied is capable of measuring oxygen content with an accuracy of 1% of full scale. Two ranges are provided (0-10% and 0-25%). When the O~25% range is selected, the oxygen content of the cell air can be moni- tored over the range from normal air (approximately 22%) to, and below, the normal 5% operating level. The requested sensitivity (0.0Z%) is ob- tainable over the full range of the instrument when a potentiometer-type device is used to read the signal. A continuous sample of the cell air will be obtained by connecting the analyzer across a restriction in a component cooling air system line. The component cooling air system ob- tains its air from the reactor cell and is part of secondary containment. oince the construction of the analyzer would not satisfy the requirements of the containment system, and since a possibility existed that the sampled alr could be contaminated, a safety-grade block-valve system was installed to maintain the integrity of the containment and to protect the operator. Component parts of the analyzer system were obtained from commercial sources, and the system was assembled and checked out at ORNL. Figure 1.15 shows the analyzer cabinet. Figure 1.16 shows the block- valve header assembly. Installation of this system in the MSRE vent house 1s nearing completion. Final checkout and operational testing will be per- formed before the start of power operation of the reactor. Fuel-Processing oystem. Fabrication, installation, and checkout of instrumentation and controls for the fuel-processing system were completed. Design efforts on this system during the past report period involved the completion of design, fabrication, and calibration of special flow ele- ments used in measurement of Fp, HF, and He gas flow and revision of in- strumentation required to improve performance and to satisfy safety re- quirements. Controlled Test Rig for Surveillance Program. The instrumentation and controls design for the surveillance program controlled test rig8 wa.s completed. The rig will be used to expose samples to conditions which, except for nuclear radiation, approximate the conditions experienced by similar samples installed in the MSRE core. Since it is desired that the out-of-pile samples have the same temperature history as the in-pile samples, and since 1t 1s not practical to measure the temperature of the in-pile samples directly, on-line computer control of the test rig heaters 3 PHOTO 72554 Fig. 1.15. Reactor Cell Oxygen Analyzer Cabinet — Front View. 35 PHOTO 72556A HAND SOLENOID HAND SOLENOID SHUT—-OFF BLOCK SHUT-OFF BLOCK VALVE == VALVES = VALVE VALVES »om ] LEAK TEST LEAK TEST ; LEAK TEST /| LEAK TEST BLOCK = BLOCK =~ BLOCK / BLOCK OUTLET Fig. 1.16. Reactor Cell Oxygen Analyzer — Block Valve Header Assembly. will be used. ©OSet points, error signals, and control functions (propor— tional, reset, and derivative actions) will be developed in the MSRE data- logger—computer. oet points (simulated sample temperatures) will be computed using ex- isting temperature and power input information to the computer and reactor temperature profile equations. Error signals will be obtained by computing the difference between the computed set polnts and computer inputs obtained from measured temperatures in the test rig. The 10- to 50-ma control signal supplied by the computer (using the existing analog output capa- bilities of the data—logger—computer) will be transduced to a pneumatic signal which will drive the Variacs which control the test rig tempera- tures. Three zones of heat control are required on the test rig; therefore, three control channels are used. To prevent complete loss of control in the event of computer fallure and to permit completely manual operation of the test rig, the computer control was superimposed on a manually ad- justable base heat control, and the amount of control allotted to the computer was limited to that required to vary the test rig over the nor- mal range of reactor sample temperature variations. Due to the inherent flexibility and capability of the MSRE data- logger—computer, the only difference in equipment requirements (other than the computer) between the computer-controlled system and a strictly manually controlled system was three pneumatically operated Variacs and three current-to-pneumatic converters. Fabrication of the control panels has been completed, and installa- tion of equipment in the MSRE high-bay area is in progress. Closed-Circult Television for Remote-Maintenance Operations. Design of the television system installation and design and fabrication of the console assembly were completed. Figure 1.17 shows the console. Although 36 PHOTO 72443 Fig. 1.17. Remote Maintenance Television Control Console. 37 PHOTO 72445 Fig. 1.18. Radiation-Resistant Camera and Accessories for Remote Maintenance Television System. only two monitors are used, three camera systems are installed. A video switching system permits the operator to display the signal from any of the three cameras on either or both monitors. The "Joy Stick" controls mounted on the front of the console table enable the operator to control pan, tilt, focus, and zoom operations on two of the three cameras with wrist and finger motion. Other, less frequently used controls and ad- Justments are located on the sloping panel in front of the operator. opace was provided on the table for addition of crane controls. The de- sign of this system was strongly influenced by the limited space avail- able in the maintenance control room. Figure 1.18 shows one of the ra- diation-resistant cameras with a radiation-resistant zoom lens and a pan and tilt unit. The complete system was tested after assembly to demon- strate performance and reliability. A number of minor failures occurred during the first few weeks of operation; however, since these troubles were corrected, the performance and reliability have been excellent. Fuel Sampler-Enricher. Installation and checkout of instrumentation and controls for the fuel sampler-enricher system were completed. 38 Nuclear Instrumentation. Installation and checkout of the nuclear instrumentation were completed. Data System. Fabrication of the data-logger—computer was completed at the vendor's plant in early February and released to the programmers for program loading and checkout. Starting about the middle of February and continuing through March and April, the system was checked out at the manufacturer's plant by both ORNL and the wvendor!'s (Bunker-Ramo) personnel The checkout consisted in debugging and running all the system programs written by ORNL and Bunker-Ramo. The programs and system operation were verified as meeting the preliminary acceptance requirements except for the analog input system, which could not be effectively checked because of the lack of live input signals, and for the computer power failure circuitry, which was not installed. * The system was shipped to ORNL and arrived gbout May 1. The instal- lation of the system, including the connection of the input signals, was completed in three weeks. All necessary building modifications and in- stallations of power distribution wiring, signal input transducers, and signal input wiring were completed prior to delivery of the system. The installed system is shown in Figs. 1.19 and 1.20. In the front row of PHOTO 80019 Fig. 1.19. MSRE Computer—Data-Logger System Console, Typewriters, and Tape Units. 39 Fig. 1.20. Computer, Typewriters, and Tape Units of the MSRE Computer—Data-Logger System. cabinets the first three contain the computer, and the last contains the power supplies. Five cabinets are located in a row behind the cabinets shown in Fig. 1.20; they contain the input relays, amplifiers, analog-to- digital converter, and the associated control circultry. After the installation was complete, the programs were loaded and The system was operated with live input signals to test hardware, pro- grams, and input signals. These tests were continued from the middle of May through most of June. Numerous hardware and some program dif- ficulties were corrected. Also, some difficulties with the input signals were found and corrected. Most of the system problems during this period were caused by the analog input equipment, which includes the input relays, amplifiers, analog-to-digital converter, and the associated control cir- cultry. These problems seemed to result from inadequate design and test- ing of the input system, which is unique and which was built special for the MORE installation. On June 24, after many corrections and modifications of the hard- ware, the system was initially accepted with the provision that it op- erate continuously for one week without an error, excluding errors caused 40 by typewriters or magnetic tapes. The system ran for six days before fail- ing. From then until the middle of August the test was attempted many times but usuvally failed after a day or two. Many problems were found in the hardware, usually the analog input equipment. Attempts to run the test continued until August 16, when the system was shut down at the ven- dor's request to install extensive modifications intended to improve re- liability. These modifications are scheduled for completion about Septem- ber 1. During the reactor critical experiments in June and July, the data logger and computer were used to collect and process some of the data from the reactor dynamics, rod drop, and pressure coefficient of reac- tivity experiments. The data system was not completely effective during these experiments due to hardware failures and poor signals from the re- actor instruments. However, some benefit was obtained from its use. MSRE Training Simulators. Two reactor kinetics simulators were de- veloped for the purpose of training the MSRE operators in nuclear start- up and power level operation procedures. (The startup, or zero power, simulator was run in February 1965; the power level simulator will be run in September 1965.) Both were designed to be tied in to the MSRE reactor control and instrumentation system. The "on-site" simulator has several advantages over one set up in a computing facility: (1) the operators get used to the actual instrumentation and controls system; and (2) much of the actual hardware, such as control rods, is used rather than simu- lated. The main disadvantage of the on-site simulation is that less com- puting equipment is gvailable, so the simulation cannot be as accurate as with the off-site simulator. The startup simulator used the control rod position signals as in- puts, and provided outputs of log count rate, period, log power, and linear power. The reactor's period interlocks, flux control system, and linear flux range selector were also operational. The power level simulator will include the effects of flux on tem- peratures. Radiator door position and cooling ailr pressure drop signals, as well as control rod positions, will be used as inputs, and the usual nuclear information plus key system temperature outputs will be read out on the reactor instrumentation. The reactor's servo controller and ra- diator load control systems will be used. Both simulators are set up on general-purpose, portable EAI TR-10 analog computers. Instrumentation and Controls System Performance Performance of the MoRE Instrumentation and Controls System has con- tinued to be very good. As systems become operational and as operating time was accumulated, additional minor troubles with instrumentation oc- curred; however, once the causes were determined, the troubles were easily corrected and no major changes in instrumentation or in design philosophy have been necessary. 41 Weigh Systems. In addition to the calibration drift previously re- ported,? difficulty was experienced with the multiposition pneumatic se- lector switches, and a failure occurred in one of the weigh cells. The difficulty with the selector switches, which are used to select signals from any one of the ten weigh cells for precision readout on mercury ma- nometers, was determined to be mechanical in nature and was eliminated by a redesign of the switching mechanism. The weigh cell failure was determined to be due to pitting of the baffle and nozzle in the cell. This pitting was apparently caused by amalgamation of mercury with the plating on the baffle and nozzle. How the mercury got into the welgh cells has not been determined; however, it is presently believed to have come from the manometers and to have been precipitated on the baffle by expansion cooling of the air leaving the nozzle. Appreciable quantities of mercury were also found in the tare pressure regulators on the con- trol panel; however, no mercury was found in the interconnecting tubing or in other portions of the system. Several methods of preventing this problem from reoccurring in the future (including replacement of the ma- nometers with precision gages) are under consideration. Temperature Scanner. The scanners continued to operate satisfacto- rily until March, when four differential amplifiers failed. This failure was caused by overheating due to failure of the fans in the amplifier cabinet. The amplifiers were completely rebuilt, since many of the com- ponents were burned up. The two fans were repaired, and the system was restarted. However, the ambient temperature in the cabinet was still around 140°F, partly due to the high ambient temperature in the heater panel control area. To remedy this, a flexible hose from the control room was brought down to the input of the amplifier cabinet so that cool alr from the control room was pulled through the amplifiers by the two cabinet exhaust fans. This seemed to improve scanner operation consid- erably, and no further amplifier failures have occurred. A better ar- rangement of this cooling system is nearing completion. The thermocouple scanner reference voltage supply for the two ra- diator scanners was installed and checked out. Before reactor power operation begins, all the mercury switches will be cleaned and checked and the complete scanner system checked and cali- brated. An additional 17-in. oscilloscope is being modified for use as a spare. Drain-Tank Level Probe Power Supply. Failures due to overheating also occurred in the drain-tank level probe power supply. In this case the damage was limited to vacuum tube failures. This problem was elimi- nated by reducing the load on the supply and by providing better ventila- tion. It should be noted that the level probe power supplies and the scanner amplifiers are vacuum tube equipment. No failures due to over- heating have occurred in any of the solid-state equipment used in the MSRE . 42 Thermocouples. Performance of the 1033 thermocouples in the MSRE system continues to be excellent. There were no known thermocouple failures during critical and low-power operations. An apparent sensi- Tivity of several thermocouples, located inside the containment, to cell ambient temperature was determined to have been caused by a double re- versal of the Chromel and Alumel extension wire leads at the disconnects and at the out-of-cell Junction box. Manually Operated Helium Throttling Valves. Considerable difficulty was experienced by the MoORE operators in setting the bellows-sealed hand valves, which control helium flow to the fuel and coolant pump bowl level systems, to maintain a constant flow of 366 cm’®/min. The valves were dis- mantled during the precritical shutdown and found to have been contami- nated with oil and metal particles. It was also found that the trim had "on-off" instead of throttling characteristics. The valves were cleaned, and new trim with throttling characteristics was fabricated and installed in the valves. Performance of these valves was satisfactory during crit- ical and low-power operation. The source of the contamination has not been determined. Component Cooling Air System Control Valves. As previously reported,lo cooling air flow to the freeze valves was found to be inadequate during precritical operations. Air flow to the control rods and reactor access nozzle was also found to be inadequate, and flow to the pump bowl shroud was fTound to be excessive. Measurements made during the precritical shutdown showed that the sizing of all of the freeze valve cooling air control valves wasg adequate. In most cases the low flow was due to restrictions in the lines or freeze valve shrouds. The low flow to the coolant-salt system freeze valve was found to be due to improper adjustment of the cooling air control valves. The sizing of the cooling air control valves for the fuel drain freeze valve (FV103) was found to be adequate to supply the required blast air flow; however, the control during the "hold" mode of operation was very poor because the air flow needed was so low that the valve was forced to control on the seat. This problem was successfully corrected by in- stalling specially designed trim in the valve. The low flow to the reactor access nozzle and control rods was found to be due, in part, to the equal-percentage characteristics of the valve together with insufficient actuator stroke. This problem was corrected by reshaping the valve trim to obtain the full rated capacity of the wvalve with the available operator stroke. Reduced area trim was installed in the fuel pump bowl cooling air control valve to obtain satisfactory control at the new (reduced) air flow requirement. oome trouble was experienced with hysteresis in the specially de- veloped valve actuator motion multipliers.ll This problem was eliminated by increasing some clearances in the multiplier assenbly and by adding 43 a floating coupling between the valve stem and the multiplier to provide greater freedom of motion. Helium Control Valve. Except for some trouble with plugging of the fuel and coolant system letdown valves, performance of the helium con- trol valves has been satisfactory. There have been no failures of these valves due to galling since the start of precritical operations. The cause of plugging in the letdown valves was determined to be deposition of particles of carbon and salt on the valve trim. The sizes of these valves were increased to reduce the effect of the deposits. Revisions and Modifications Nuclear Safety Instrumentation. The nuclear safety system was modi- fied by the addition of period safety amplifiers in each of the three channels; see Fig. 1.21. The monitor and test unit was redesigned to include in-service testing of these. A second modification to the safety system lowers the flux scram level by a factor of 1000 when fuel salt is not being circulated by the pump. A measurement of the three-phase current to the pump provides the input information to accomplish this. Initial criticality tests disclosed that the neutron flux attenua- tion in the instrument penetration departed from the ideal exponential curve. A flat, or nearly flat, region in the attenuation curves prej- udiced operation of the wide range counting instrumentation. Addi- tional shielding is being designed for the penetration to obtain the desired characteristics. Control and Alarm Circuits. ©Safety-grade circuits were added to lower the flux scram level when fuel salt is not circulating, to pre- vent the occurrence of excessively low pressures in the containment which might damage the reactor containment vessel, and to block the sample lines that comnect the reactor cell oxygen analyzer to the sec- ondary contalnment system in the event of high reactor cell pressure or air activity. The center heaters and the high (1250°F) center temperature inter- locks and alarm were removed from all freeze valve circuits. The siphon break and permissive-to-thaw circults were revised to make clear the causes for alarms. On freeze valves 103, 104, 105, 106, 204, and 206, the "Freeze Valve Frozen' logic circulitry was changed from one-of-two to two-out-of-three to prevent false operation of control interlocks when normal control actions occurred. A number of revisions were made 1n control-grade circuits for the purpose of correction minor errors in the design, improving performance of the system, or providing additional alarms. Ixamples of these changes are: removal of drain-tank pressure interlocks from the prefill mode cir- cuits, installation of additional jumpers on the jumper board, revision 4ty "WBJISBI(Q 00T WBIDG POY TOJIJUO) HMSW °TZ°T "ITd ¢S (¢401no2) [ 5 , INIWIT3 ¢ ON XIHLVW AV 13 ft— 35S —— ¢ STINNYHO 30N30IONI0D < €S ALIIVS =—P— g gy | \ >60,000 cycles and remain in good condition. Freeze Valves oceveral problems were found in the operation of the freeze valves during the zero-power experiments.2 Changes were made in the electrical and alr supply to the valves to provide more margin for operation during abnormal conditions. The results of the changes are described below. Cooling air flow rates to the freeze valves have been increased by reducing the line pressure drop. Maximum air flows (scfm) to the valves with an 8-psi air supply are now: FV204, 22.5; FV206, 25.0; FV104, 26.6; Fv105, 27.7; and FV106, 24.8. Freeze Valves 204 and 206 The coolant-salt freeze valves 204 and 206 have been modified by re- moving the center heater and enclosing the 2-in. valve plate in a metal shroud to contain the cooling air. These valves are now similar to all the other valves in the system with the exception of FV103. The heat ca- pacity in the furnace area surrounding the valves was increased by instal- ling a ceramic liner around the valve heaters. Freeze valve 204 now melts in 13 min on power failure and 206 melts in 12 min. The control of the temperature distribution along these valves will be improved by installa- tion of separate control circuits to the valve shoulder heaters. These valves formerly used a single Variac to control the heat on both sides of the frozen wvalve. Independent control of the shoulder heaters should alleviate the temperature gradient shown in Fig. 2.2. A differential flow controller is being added to the cooling air supply system to each of these valves and will be operated from a single selected thermocouple. The function of the controller is to maintain the valve temperatures within the module control temperature boundaries by automatically varying the cooling ailr flow to maintain a constant temper- ature at the valve. The purpose of this change is to reduce the amount of temperature cycling, the frequency of the alarms that occur when the module set points are exceeded, and the attention required by the reactor operators. Freeze Valve 103 The operation of FV103 is complicated by a dual set of operating conditions: (1) After the reactor system is filled with salt, FV103 is 52 ORNL—-DWG 65—11795 58 e C Y T v X T e N\ A4 B4 B4 A4 B b 1200 l TEMPERATURE (°F) X\ / X X% /X 800 < \ X X/ 13-min POWER t2-min POWER FAILURE MELT FAILURE MELT 400 \ X V/ | 1 FV-204 FV-204 : 1 || A4 fiB 2B 3B B4 5B B4 38 2B 1B A4 TE NUMBER Fig. 2.2. Freeze Valves 204 and 206 Temperature Distribution — Hold Freeze Condition. frozen, and salt is retained in the 103 line for some period. At this time there is salt on both sides of the valve. (2) After determining that the salt level in the pump bowl is correct, the 103 line is vented via line 519. Freeze valve 103 then has hot salt on the reactor side of the valve and an empty pipe on the drain-tank side. During condition 1, the temperatures are symmetrical across the frozen valve. During condi- tion 2, a temperature gradient of ~400°F exists between the reactor side and the drain-tank side of the wvalwve, probably due to the difference in the heat conductance of the full and the empty line. The module controls were set to control over this wide range. The freeze time for FV103 is 15 to 30 min if the salt is steady in The system. Thaw time average has been 11 min with the reactor tempera- ture at 1150 to 1200°F. A differential controller was installed on FV103 prior to the criti- cality tests and operated smoothly throughout the run. The controller 53 ORNL-DWG 65-11796 3A_ 38, 2A 28 1A 1B N\ T/ | ( . . Nul g/ 0C - ~ REACTOR U ] A\ A\ e )0 \ \ B A1B FV-103 1400 1000 o”/////j\\ 4””’t:::::::::::; < — %’L ‘/)]( \ X/ & 800 X \\\~ g \\ /47 X CONTROLLER ON FV103-{B-—960°F SET o O CONTROLLER ON FV103-3B-540°F SET _| 600 S Nx_ / L}E I\X 400 i \[ VALVE IN HOLD FREEZE CONDITION 200 | 0 B AlB 38 2B 1B R27-30 AVG TE NUMBER Fig. 2.3. Temperature Distribution on FV103 with Controller. was operated initially from the drain-tank-side thermocouple (FVlOB-BB), and 1t was found that this couple was relatively insensitive to changes in reactor temperature. The control was then shifted to the thermocouple on the reactor side of the wvalve (FVlOB-lB) and remained there for the balance of the run. Over the range of 1150 to 1200°F no changes in con- troller set point were necessary. The reactor temperature was lowered to 1050 to 1100°F at one stage of the operation, and it was necessary to lower the control point by 50°F. Operation of the controller from the valve center temperature (FV103-2B) will be tried next to see if an op- timum mode of operation can be found. Figure 2.3 shows how the tempera- ture distribution across FV103 changes with the location of the control thermocouple when the valve is in the hold-freeze condition. Freeze Valves 104, 105, and 106 Operation of these wvalves has been difficult for a number of reasons: (l) pressure differences in the drain system which affect the salt dis- tribution in the valve tree (FV105 and 106), (2) poor heat control and resultant temperature distribution across the valves, (3) inadequate heat at specific locations, (4) loss of heater elements for mechanical reasons, and (5) difficulties with the temperature control modules. 54 Methods are being‘developed with operational experience which ensure adequate Tilling of the valves with salt. Separate shoulder heater controls were installed and tested on freewze valves 105 and 106 which relieve the temperature distribution difficul- ties. Power to the shoulder heaters was increased from 15 w/in.2 to 30 w/in.®. The effect of these changes is shown in Fig. 2.4. The loss of some of the heaters from mechanical difficulties was due to improper installation; warping of the heater boxes to which these heat- ing elements are attached, with resultant breaking of the ceramic elements; and abuse of the heater element lead wires during installation. New heat- ing elements were installed in all the freeze-valve heating units. These ORNL—-DWG 65-11797 LINE 103 FUEL DRAIN TANK _— FUEL DRAIN TANK 7 J/KM1Q13 1200 I ]] ] ¢ 1000 5 3 / | 7 {/ ) _ / < 800 3 w C | X o D — | < Q: b a |1 A = 600 |i - {l O BEFORE MODIFICATION | ‘| X AFTER MODIFICATION |l 400 , 200 l ! 1 2 3?5 6 7 8 910, 12 13 14 ¢ — - Fig. 2.4. Effect of Freeze-Valve Heater and Control Modifications on Temperature Distributions for FV105 and FV106. 55 elements were installed to fit loosely to permit some freedom of movement. The lead wire size was reduced from No. 12 to No. 1l4. The method of at- taching the lead wire to the element wire was modified to prevent twisting and breaking the leads at the welded joint. This was done by adding a short length of wire at 90° to and extending outward from the welded joint and embedding the cross and joint in the ceramic. 1t was noted that in a number of units with the straight connection, the weld had broken; but the lead and resistance wires maintained conti- nulty due to the twisting together of these wires prior to welding. How- ever, breakage at this weld point can create an area of high resistance and possible burnout; so the change in attachment method was made. The module difficulties are being corrected by the instrument group. These were multiple alarm points for a single set point and drifting off set points in some instances. Differential cooling air controllers are being installed on freegze valves 105 and 106. Noble-Gas Dynamics Analysis of the experiment® on the noble-gas dynamics was continued. The experiment was designed to evaluate the constants needed to describe the xenon migration in the MSRE and consisted in adding 85Ky to the gas space of the fuel pump bowl for an interval of time and then observing the rate of removal during an interval of stripping. Run 3, which consisted of an 11-1/2—day addition phase and a 6-day stripping phase, yielded five exponentials. The half-lives associated with these exponentials and their physical interpretation are given in Table 2.1. Exponentials 3 and 4 are still subJject to interpretation. Although listed as two graphite Table 2.1. Results of ®9Kr Experiment, Run 3 Fxponential Half-Life Process Rate Constant Rate Constant *P (hr) Involved Value 1 0.119 Purging pump bowl Purging 74% efficiency 2 1.04 Stripping salt Stripping 129 efficiency 3 4e52 Probably stripping Mass transfer 2.0 ft/hr graphite coefficient 4 15.5 Probably stripping Mass transfer 0.59 ft/hr graphite coefficient 5 198 Stripping bulk Mass transfer 0.046 ft/nr graphite coefficient 56 regions, they may actually be one graphite region in which krypton diffu- sion in the graphite must be considered. The estimated stripping effi- ciency, from work done at the University of Tennessee with a CO,-water system, is about 17%. The estimated mass transfer coefficient for the bulk graphite, assuming turbulent flow, is about 0.1 ft/hr. It is rea- sonable that the measured value is lower than this because the flow is not turbulent through the entire channel, as was assumed in the estimate, but only through a part of it. scampler-Enricher Installation of the sampler-enricher system for the fuel pump was completed just before the start of the precritical run (PC-2). Therefore, shakedown testing and training of operators were done concurrently with sampling and enriching of the fuel loop. The system was used to isolate 54 samples and add 87 capsules of enriching salt during PC-2 and the zero- power run (run 3). Twenty operators were trained in the use of the sampler during this period. Little or no contamination of the work area occurred during operations or maintenance. A maximum radiation field of 5 mr/hr was measured at the outside of the unshielded transport container which contained samples isolated either during an extended 10-w operation or immediately after a high power peak. During this initial operating period some maintenance was required, and some changes were indicated which were delayed until the completion of run 3. BSeveral components were removed from the system for maintenance and were decontaminated easily by washing. The mechanical troubles encountered during this initial period are discussed below. After the indicated changes are completed and after the Operators gain additional experience, less maintenance should be needed. Removal Valve and Seal The transport container and removal tool assembly did not slide through the removal valve and seal unit freely. Examination revealed that the valve and seal were not aligned properly with area 3A (see Fig. 2.5). After clearances were increased, no further binding occurred. During the entire test the ball of the removal valve failed to seal properly even though both the ball and the seals were replaced. There- fore, the entire valve assembly will be replaced with a modified design which should improve the sealing characteristics and access for future maintenance work. During the run, one of the three-way solenoid valveg which controls the air supply to the removal valve operator failed and was replaced. Manipulator The boots used to seal the manipulator arm to area 3A were replaced three times. On one occasion area 3A was inadvertently evacuated without 57 ORNL-DWG 63-5848R REMOVAL VALVE AND SHAFT SEAL PERISCOPE LIGHT N N /] L xrr? CASTLE JOINT (SHIELDED WITH DEPLETED URANIUM) CAPSULE DRIVE UNIT— e LATCH\W F ACCESS PORT— & N | i B ! N | ] MANIPULATOR %\AREA 3A (SECONDARY CONTAINMENT ) SAMPLE TRANSPORT CONTAINER | EAD SHIELDING "D — ——. .7 o 2 2 ) AREA 1C (PRIMARY CONTAINMENT) — SAMPLE CAPSULE ] OPERATIONAL AND MAINTENANCE VALVES SPRING CLAMP \ DISCONNECT | ——=AREA 2B (SECONDARY CONTAINMENT) TRANSFER TUBE (PRIMARY CONTAINMENT )~/ LATCH STOP —+ CRITICAL CLOSURES REQUIRING A BUFFERED SEAL MIST SHIELD ’ O ! 2 CAPSULE GLIDE FEET Fig. 2.5. ©Schematic Representation of Sampler-Enricher Dry Box. the manipulator cover, and the resultant pressure gradient across the boot ruptured it. Modifications to the interlock system are planned to prevent recurrence of this type of accident. On another occasion, while using the manipulator to release a sticking access port operator, the boot was snagged on the bottom piece of the transport container. The height of this bottom pilece was reduced, and the access port operators were read- Justed. The third failure resulted from pinching the boot between the manipulator arm and the housing. Access Port On several occasions one or two of the six access port operators failed to open when they should have. The manipulator was required to release the sticking operator. The tension on the operators was read- Justed, and the gas-supply tubing to the pistons was realigned to relieve & restraining stress. To increase the ease of emergency operation, a knob was added tc the pin which must be moved with the manipulator to manually release the operator. Operational Valve The leak rate of buffer gas through the seals on both the operational and maintenance valves exceeded 5 cm® of helium per minute. In both cases 58 the upper gate seal had the greater leak rate. The operational valve was removed and examined. A thin black ring, which was easily removed, had formed. at the upper sealing surface of the valve gate. When the valve stem and gate were lubricated, the valve sealed almost completely. When the lubrication was then removed from the gate, the leak rate of buffer gas (helium) increased to about 2 cm®/min through the upper gate seal and remained at zero through the lower seal. The maintenance valve will be removed and cleaned, and the stem will be lubricated. A small quantity of salt spheres (estimated at <1 g) had collected between the seats of the gate valve. The analysis of the uranium in the spheres showed the concentration to be less than enriching salt and more than fuel salt. Therefore, the spheres must have come from the used en- riching capsules as they were retrieved from the pump bowl. A possible explanation is that the bottom hole in the enriching capsule was not drilled completely through the nickel wall; a small droplet of salt could bridge the hole and then could be dislodged from the capsule during han- dling with the manipulator while removing it from area 1C. campling Capsule On one occasion a sampling capsule was knocked into area 1C before the latch key was completely engaged in the latch, and the capsule as- sembly dropped down to the operational valve. The capsule was retrieved by removing the manipulator assembly from area 3A, opening the access port, and retrieving with a hook on the end of a stiff wire. Fuel-Processing Sampler Design of the sampling equipment and instrumentation for the fuel- processing system has been started. The sampler-enricher mockup will be modified for this use., Off-Gas System Charcoal Beds. A gas-retention test was made to check the perform- ance of the MSRE charcoal beds. With the bed temperature at 85°F and helium flowing at a constant rate, a pulse of 85K was injected at the bed inlet. The time required for passage of the krypton through the bed was determined by monitoring the effluent gas with a G-M tube. The holdup time for krypton at the design flow rate of 4.2 liters of helium per minute was 5~l/2 days. Adsorption data published by Ackley and. Browning4 indicate the equivalent xenon holdup time to be a minimum of 88 days. By adjusting the holdup times downward to allow for tempera- ture effects, an estimate was made of the atmospheric concentrations of krypton and xenon which will be produced by the charcoal bed effluent during 10-Mw operation. The estimate was based on the following assump- tions: 59 Minimum holdup time (t) — 4-3/4 days krypton and 75 days xenon Minimum stack flow — 7 X 10° cm?/sec (15,000 cfm) Atmospheric dilution factor — 1500 (ref. 5) Maximum atmospheric concentration will be: (Ri) (g ) (e™MY) (108) C; = = 2.6 X 10712 Rjhie ("7 x 10°) (1500) (3.7 x 1010) =N\t J where I Ry flow rate at pump bowl outlet, atoms/sec, Ci atmospheric concentration, pc/cm?, For the indicated holdup times, 8°Kr, 121MXe, and 133Xe are the only iso- topes which yileld concentrations of significance. The concentrations for These are given in Table 2.2 together with the maximum permissible con- centration (MPC). Table 2.2. Atmospheric Concentrations of Radiocactive Gases from the MSRE Off-Gas System MPC (uc/cmB) -7\1't 3 * Ry Ay © Cy (he/en?) 168-hr Week® 85Ky 9.4 %X 1014 2,14 x 10°° 1 5.2 x 1072 3 x 1076 13y, 9,2 x 1012 6.7 x 1007 1.3 x 1072 1.4 x 107° 4 X 1070 133x%e 2.0 x 1016 1.5 x107%® 6.0 x 1075 4.7 x 107° 3 x 10°6 SNBS Handbook 69. Solids Entrainment. Difficulties were encountered with operation of the back-pressure control valves (PCV 522 and 528) in the fuel and coolant off-gas systems during the precritical and critical test periods. The trouble was due, at least in part, to an accumulation of solids at the valve seats. Visual examination indicated that the solids were a mixture of glassy spheroids, about 1 p in diameter, and carbon or a car- bonaceous material. Optical measurements indicated that the chemical composition of the spheroids was the same as that of the circulating salts, and it is assumed that they are due to carryover of the salt mist in the pump-bowl gas spaces. The carbon is assumed to be a residual con- taminant from the manufacturing process. The problem is being approached in two ways: (1) a study is being made to determine whether the solids 60 entrainment can be reduced or eliminated, and (2) filter elements immedi- ately upstream of the wvalves are being changed from a 25-u pore size to l-u size. cample Unit. Design work was started on a sample unit for the re- actor off-gas stream. The unit will include: 1. @a thermal conductivity cell for on-line measurement of total contami- nants, & gas chromatograph unit for measurement of gaseous constituents, a refrigerated molecular sieve trap for batchwise collection of gas- eous constituents; the trap contents will be transferred to a shielded sample bottle and then to a hot cell for analysis. Supply to the sample unit will be a small side stream (<100 cmB/min) taken from the off-gas line immediately downstream of PCV 522. Initial design effort is being concentrated on areas which must be completed prior to operation of the reactor at power (e.g., piping connections to existing equipment and shielding revisions). Remote Maintenance The previously established programs for ensuring the maintainability of the MSRE were continued. Broadly stated, these programs consist of monitoring new designs and changes, inspecting the installation of the reactor equipment, and trying out the maintenance procedures where the possibility of future trouble could be determined. One control rod and its drive were removed using the work shield. This Jjob required handling five 3/8 -in. socket head bolts with long-han- dled tools. Vision and maneuverability were hampered because of close clearances and because the bolts are not quite accessible from directly above. Previously developed lights and viewing techniques were not effec- tive for all the bolts. However, a viewer with a collimated light source mounted above the work shield was used with success. Because it takes longer to set up this type of light source, it will be used only on those bolts where the internal lights are not good enough. In addition, align- ment guides were added to both the adapter flange and the housing to help in the installation and removal of the drive housing. All three space coolers and many of the freeze-valve tree replaceable heaters were removed from their installed positions for alterations. This provided the opportunity to check out several remote-maintenance require- ments. The space coolers were balanced for handling, flange guides were added, and the effectiveness of blowing out the lines was observed. All the space coolers were observed to lose water when the flanges were broken, indicating the need for improving the procedures for blowing out the lines. The replaceability of some of the drain cell heaters has been adversely affected by thermal distortions. This has the effect of producing a tight fit between adjacent heaters. The proposed solution is to alter the heater units if they have to be removed for maintenance. 61 All the maintenance procedures are affected by revisions, new in- stallations, and unanticipated requirements; so a continuing effort is being made to keep track of changes. Toward this end, a program of pho- tographing installed equipment and marking identifying numbers on thermo- couple boxes, heaters, and electrical leads is under way. Revised freeze-~flange clamp operators were assembled and tried for fit. Long-handled tools were designed and built for the installation of the revised graphite samples. Preparations are being made for handling the pump rotary element and the graphite samples, and for inspecting the core through the 2-1/2—in. access flange. Procedures were worked out for handling the newly designed control rod shielding. Preliminary procedures were prepared for maintaining some of the elements of the sampler-enricher. Pump Development MSRE Pumps Molten-Salt Pump Operation in the Prototype Pump Test Facility. The spare rotary assembly for the fuel pump® circulated the salt LiF-BeF,- ZrF,-UF,; (66.4-27.3=4.7-0.9-0.7 mole %) for 2644 hr at 1200 to 1400°F. During this time, measurements were made to determine the concentration of undissolved helium in the circulating salt and the effectiveness of the down-the-shaft helium purge against the upward diffusion of 85K+ to the catch basin for the lower shaft seal. Operation of the pump was terminated when strident noises and an un- wanted increase in power to the pump drive motor were noticed. Inspection revealed that three vanes, each approximately 2 X 24 X 1/8 in., had be- come detached from the flow-straightener section in the salt loop. One of the vanes was carried by the circulating salt to the impeller inlet, where 1t became lodged and rubbed against the impeller. The only apparent damage to the pump was the removal of metal from the inlet portion of the impeller vanes. The failure of the vanes apparently resulted from metal fatigue caused by vibration induced by the flowing salt. This unit will be reassembled with a new impeller, shaft bearings, and seals and given a cold shakedown test to prepare it for future reactor service. The spare rotary assembly for the coolant pump6 was prepared for re- actor service and will be held in standby. The prototype pump test loop is being modified. The Venturi flow- meter is being relocated upstream of the orifice flow restricter, and a new flow-straightener section is being installed upstream of the Venturi. Lubrication Systems. Tests were conducted in the transparent mockup of the lubrication reservoir to determine the source of gas entrainment in the circulating lubricant that was noted during shakedown tests of the MSRE lubrication systems.7 The entraimment made it difficult to prime a standby pump during emergency operation. Sources of entrainment were 62 found to be the action of the jet pump used to scavenge oil from the bear- ing housing and the agitation of the oil caused by the rotation of the shaft and bearings. The scavenging Jjet pump was modified to reduce the entrainment of gas, and the standby pump primed satisfactorily. A modi- fied jet pump will be installed in the lubrication systems for the fuel and coolant pumps at the MSRE and tested prior to power operation of the reactor. The lubrication pump, which is circulating oil in an endurance® test, has operated without incident for 18,000 hr. The thermal cycling tests (1000 cycles, 1 hr running and 1 hr off) were completed. The endurance test is continuing at 160°F oil temperature and 70 gpm. Back Diffusion of Fission Gases in the Pump Shaft Annulus. Addi- tional tests were made to investigate the diffusion of $2Kr in the shaft annulus against a flow of helium purge gas. In these tests the radial clearance in the shaft annulus was 0.005 in. compared. to 0.0025 in. used in the previous tests.’ The following data were obtained with the 0.005- in. annulus: 85Kr Concentration Shaft Purge Cat;firBisln (curies/cm?) Dilution (1iters/day) (literi/da ) Factor Y/ In Pump Tank In Catch Basin 212 173 2.66 X 107¢ <1.5 x 10710 >317 700 232 164 5.50 X 1007¢ <1.5 x 10719 >36,700 72 105 be2e X 1076 8.17 x 10~? 520 It was difficult to measure the dilution factor (ratio of concentration of fission gas in the pump tank to that in the catch basin) except for very low purge flow rates of little interest to the project. The diffi- culty stems from our inability to measure the concentration of fission gas below 10710 curie/cm®. An additional factor was the relatively low concentration of fission gas permitted in the pump tank because of hazard to personnel, the quantity of the gas available, and the problems of han- dling it. However, the data do indicate that purge flow rates in excess of 1000 liters/day should protect leakage 01l in the catch basin from polymerization by fission gas. Measurement of the Concentration of Undissolved Helium in Circulating Molten Salt Circulating helium in the MSRE fuel will reduce the fuel density and, therefore, the reactivity of the reactor. Measurements of both the void fraction and the void-fraction reactivity coefficient were required. 63 A radiation densitometerl® (ORNL Dwg. 433-7.0 R 41482) was used to obtain a measure of the helium void fraction. The void-fraction coeffi- cient was determined from transient measurements with the densitometer in conjunction with other reactor parameters. The detector signal current was fed to a suppression circult that indicated only variations in de- tector output. The circuilt output was recorded on Visicorder tape. The densitometer sensed only mass variations; these variations were inter- preted in terms of void fractions and temperature effects in the following presentation. Investigations were made on two different experimental installations. The first was the prototype pump test loop. Here, the densitometer was evaluated, and a standardization technique was developed. The second in- stallation was on the fuel inlet line to the MSRE reactor vessel. Prototype Pump Test Loop. Helium concentration measurements were made with a radiation densitometer at two pump bowl levels, upper and normal. The salt was circulated through a 6-in.-diam system at 1615 gpm; an additional 85 gpm flowed through the pump bowl spray ring. The densitometer axis was horizontally positioned normal to and in- tersecting the pipe axis. The scintillation detector was maintained at 70°F by water cooling. The detector voltage was maintained at 1700 v, and the voltage divider-network drain was 9 ma. The measurement technique involved the following. First, after a steady-state salt flow and temperature condition was established, refer- ence measurements were obtained. A Tused-silica beam filter was used as a mass sensitivity standard. The mass sensitivity was checked frequently throughout all the measurements. ©Second, measurements of void behavior were obtained during salt-flow stoppage and subsequent reestablishment. The third technique involved varying the salt temperature. Flow and no- flow measurements were obtained at several temperatures. The two Visicorder traces shown in Fig. 2.6 give a measure of helium concentration in the loop at two pump bowl levels. The traces are plotted as 1T normalized to the flow density in order to illustrate the transient behavior of both the helium purge and buildup when the pump is first turned off and then turned on. Trace I relates to the normal pump bowl fuel level, which is 3-3/8 in. above the center line of the volute (ORNL Dwg. F-RD-9830). The circulating void fraction was approximately 4.6 vol %. Trace II was obtained when the fuel level was 4-7/16 in. above the center line of the volute; the fuel covered the spray ring discharge ports. The void fraction was approximately 1.7 vol %. The volume sensi- tivity of the densitometer was 0.64 vol % per division in each case. The volume sensitivity and the validity of using the fused-silica beam-filter standard are illustrated in Fig. 2.7. The densitometer volume sensitivity is calculated from the following: 100 M S >S5 s C 64 where Mg = mass of standard, DS = recorder divisions deflection due to standard, ME = salt mass subtended by the collimator. The numerical value of S is controlled by adjustment of the electronic system gain. The density at temperature T is obtained from: Mr S S pT=V;-+Pr1—O—O-(Dt-Dr)=Pr[l+-l—o—o-(Dt—-Dr)} , where Mr = salt mass subtended by the collimater at the reference tempera- ture (in this case, 1200°F), Pr = salt density at reference temperature, Dr = deflection at reference temperature, Dt = deflection at temperature in question, Vé = volume subtended by the collimator. The measured densities are in very close agreement with those obtained by independent measurements.tt ORNL-DWG 65-11798 SILICA BEAM PUMP ON FILTER I 4.6% DENSITY INCREASING I NORMAL PUMP BOWL LEVEL L L O II UPPER PUMP BOWL LEVEL II .7% DEFLECTION (div) SALT TEMPERATURE : 1185°F VOLUME SENSITIVITY: 0.64% /div O 20 40 o0 80 100 120 140 160 180 200 TIME (sec) Pig. 2.6. Transient Characteristics of the Helium Concentration in the Prototype Pump Test Loop. 65 ORNL-DWG 65—-11799 e g N — 3 C—2.286 g/cm P 16 o =2.29 g/cm3 SENSITIVITY = 0.275 VOL.%/div M p. = CALCULATED DENSITY 14 ~\\\\\ FM==MEASURED DENSITY N 2 < \; 259 g/cm? /%=2 \ N DEFLECTION (div) ® \ p.=2.203 g/cm3 _| 4 ¢ _ 3 \%2.20 g/cmx 2 \\ 0 \\\j 1100 1150 1200 1250 1300 1350 1400 TEMPERATURE (°F) Fig. 2.7. Temperature Characteristics of Loop Salt — No Flow. Molten-5alt Reactor Fuel Circuit. The densitometer was horizontally mounted on the reactor input fuel line; it was operated in the same fashion as on the prototype pump test loop. The fuel line was a 5-in.-diam pipe instead of a 6-in. one. The system flow was 1150 gpm, with a 50 gpm spray ring flow, The first measurements were obtained from carrier salt only. The measurement sensitivity was 0.076 vol %. The data indicated that there was no measurable void fraction at 1190°F. The flow was stopped and then started again numerous times, and the quiescent period was varied from 5 to 10 min. The only density changes observed were those caused by temper- ature changes (approximately 0.0095 mass % per °F). The second set of measurements were obtained after fuel was added to the carrier salt. At 1200°F there was no measurable void fraction, either with pump excursions or with 15 1b overpressure. Overpressure measure- ments made at 1050°F indicated a void structure. The low-temperature densitometer data, in conjunction with other measurements, were used in determining some of the nuclear parameters discussed previously in the section Analysis of Operation. Other Molten-Salt Pumps PKP Fuel Pump High-Temperature Endurance Test., This pump12 was oper- ated for 384 hr circulating the salt LiF-BeF,-ThF,-UF, (65-30-4-1 mole %) 66 at 1200°F and 1550 rpm. The test was halted temporarily due to failure of the lower bearing in the drive motor, apparently due to insufficient lubrication. The bearing was replaced, and the pump has since been oper- ated at 1200°F, 1650 rpm, and 800 rpm for 800 hr. MK-2 Fuel Pump. Water testing 1%°13 with the pump tank models (full scale and 4-in. section) was concluded. There was no noticeable reduction of gas content attalned with the various baffle arrangements tested in these models. The salt pump tank design12 is near completion, and the internal baffling and flow passages will duplicate that used on the full- scale pump tank model. Instrument Development Ultrasonic, Single-Point Molten-Salt Level Probe Fabrication and installation of an ultrasonic level probe system for the MSRE fuel storage tank (developed by Aeroprojects, Inc., under con- tract to the AEC with ORNL assistance) were completed prior to the start of MSRE chemical-processing system operations in May. Figure 2.8 shows the probe assembly before installation in the fuel storage tank. Figure 2.9 shows the force-insensitive mount assembly that permits ultrasonic energy 1o be transmitted through the tank wall. Note the rugged all- welded construction of this device. A final seal weld was made at the periphery of the flanged section when the probe was installed in the fuel storage tank. A Tunctional diagram of the system is shown in Fig. 2.10. Except for the frequency of operation, the principle of operation of the MSRE probe system is the same as that of the prototype probe system pre- viously described.t® A lower (25-kc) frequency was required on the MSRE probe because of the heavier construction required to provide adequate allowance for the high corrosion rates expected during operation of the chemical-processing system. Performance of the probe was satisfactory during initial operation of the chemical-processing system; however, some difficulties were experi- enced during subsequent operations. The probe operated very well when the tank was filled but did not operate when the tank was drained. A check of the instrument made at this time revealed that the oscillator frequency had drifted 40 cps from the original setting. Correction of PHOTO 72098 Fig. 2.8. MSRE Fuel Storage Tank Ultrasonic Level Probe Assembly. o7 PHOTO 72099 Fig. 2.9. MSRE Fuel Storage Tank Ultrasonic Level Probe, Force- Insensitive Mount, and Excitation Rod. 68 ORNL-DWG 65-11800 2] e FUEL STORAGE CELL SHIELDING = CONCRETE WAL L iaf"flméfi PIEZO ELECTRIC : L CRYSTALS LEVEL LIGHTS MAGNETO- STRICTIVE TRANSDUCER DIFFERENTIAL AMPLIFIER ) \\\\*ELECTRONKIPACKAGE 7 % FORCE INSENSITIVE MOUNTX ‘. ) % // 17 g 1V # el / 07 v g 7 EXCITATION ROD % 4 f 7 7 7 ] // % L /////// *PATENT NO. 2891180 AEROPROJECTS INC. Fig. 2.10. MSRE Fuel Storage Tank Ultrasonic Level Indicator oystem. the frequency restored the instrument to an operative condition. Further checks revealed that the frequency wvaried as much as 300 cps over a period of a few days. Since the probe is basically a sharply tuned (high-Q) resonant system, small shifts in oscillator frequency from the resonant point caused the probe to become inoperative. The problem of frequency drift was further complicated by the pres- ence of a number of resonant peaks within the range of oscillator fre- quency adjustment (some of which were not responsive to level changes) 69 and by the difficulty of checking instrument performance in the field withoug interfering with operations (salt level must be varied to check probe performance). The difficulties experienced with the MSRE probe showed that some improvements were needed if the device was to be useful for long-term operation under field conditions. To gain a better understanding of the problems involved, studies were made of the frequency response and per- formance characteristics of the probe gystem. Because of the need to minimize interference with MSRE operations, these studies were made using the prototype probe system installed in the MSRP level test facility. Results of tests performed showed that a number of resonant peaks existed (see Fig. 2.11) and that, while several of the peaks were level sensitive, only one was usable. When the molten-salt level was ralsed, one peak disappeared when the salt touched the sensing plate. This peak was not the one with the highest amplitude. Other peaks disappeared as the salt level rogse and covered more of the excitation rod. Some peaks were not affected by level. The MSRE probe will be checked the next time that salt is transferred to the fuel storage tank, using information gained and procedures devel- oped from the studies discussed above. An energy-absorbing (Q-reducing) slug will be installed in the excitation rod to broaden the bandwidth of the resonant peaks and to reduce the effects of frequency drifts. The oscillator will be stabilized by component changes and/or circuit revi- sions. Installation of a notch filter in the amplifier to eliminate un- wanted resonant peaks from the signal is being considered. As previously 10 [ ORNL-DWG 65-11801 | | —— MOLTEN SALT BELOW SENSING PLATE H -——— MOLTEN SALT TOUCHING SENSING PLATE | --------- SENSING PLATE 0.4 in BELOW SURFACE fl 8 f--- —t—— OF MOLTEN SALT —f ... . 7 - = O —— e e — e — | | ] et i T et — oo 2 g 0 0 0§ LEVEL DETECTOR OUTPUT (mv dc) el oo 08B i — | o S e —— . . ) , {\ o4 d A . .//J \\ : : II ‘\ r o . \ . » A e o\e 1, . ":.z\\ . fl°0 \\ o .. ~~ A ----- O — Lfim“b—- | e & 3 49,000 50,000 51,000 52,000 ULTRASONIC GENERATOR FREQUENCY (cps) Fig. 2.11. Ultrasonic Level Detector Resonance Peaks. 70 reported,l4 replacement of the oscillator with one that would automatically adjust to desired frequency is also being considered. The manufacturer (Aeroprojects, Inc.) is studying the effects that dimensional changes of the sensing plate will have on the resonant fre- quency. This information will be of wvalue in predicting the effect of corrosion on probe performance and might be of some value in determining the rate and extent of corrosion during chemical-processing system opera- tions. The work reported above was done with the assistance and cooperation of the probe manufacturer (Aeroprojects, Inc.). High-Temperature NaK-Filled Differential-Pressure Transmitter The coolant-salt system flow transmitter that failed in service at the MSREL® was replaced with a spare transmitter after further field tests and observations failed to reveal the reason for the shifts in zero and span calibration. An attempt to return the defective transmitter to the manufacturer for repair was abandoned because the instrument was contaminated with beryllium and the manufacturer did not have the facilities for handling contaminated materials. The possibility of decontaminating the instrument at ORNL was discounted when it was determined that the cost of decontami- nation without further damage to the instrument would probably exceed the cost of a new instrument. A new instrument has been ordered for use as a spare. Two high- temperature, INOR-8, diaphragm seal assemblies, fabricated and inspected on the original purchase order, will be used in the fabrication of the replacement instrument. The defective tTransmitter is being tested at ORNL, under controlled conditions, to determine the cause of the zero and span shifts and to de- termine whether repair of the transmitter, at ORNL, is feasible. Results of these tests to date are inconclusive; however, it has been determined that the shifts are predominantly temperature-induced zero shifts which are mechanical in nature. A 10°F change in ambient temperature around the body of the transmitter caused the output to vary approximately 50% of full scale. ©Static pressure changes applied equally to both high and low inputs caused no appreciable shifts in output. The possibility that the shifts originated in the strain-gage transducers or associated elec- tronic circuitry was eliminated by disconnecting the strain gage and ob- serving drift under various load and temperature conditions. The results of the controlled tests confirm the conclusions reached from field test and observationl® but have added very little to our under- standing of the causes of the malfunction. An attempt will be made to eliminate the shifts by refilling the transmitter body with silicone oil. If this attempt is not successful, the transmitter will be disassembled and visually inspected. 71 Float-Type Molten-5Salt Level Transmitter During a recent shutdown of the MSRE prototype pump test loop, the developmental ball-float-type level transmitterl® was dismantled and in- spected to determine whether any buildup of vapor-deposited salts was occurring which would affect future performance of the instrument. The transformer was removed, and the core chamber was cut off at the top of the float chamber. No significant deposits were found. A thin deposit of salt was found on the part of the core tube that was normally in the float chamber, but the amount was not enough to restrict motion of the core. A sample of this material was taken for analysis. Performance of the ball-float-type transmitter installed on the MSRE coolant-salt pump continues to be satisfactory. The errors in calibra- tion previously reported still exist; however, the information required To correct the calibration has been obtained, and corrections will be made when the system is filled. Design of the ball-float transmitter installation in the MK-2 MSRE fuel-circulating pump is in progress. All information necessary for this installation has been transmitted to the pump designers. Due to consid- erations of piping layout and pump configuration, it will not be possible to design this installation to permit remote removal and replacement of the differential transformer as originally planned. However, since no failures have occurred on any of the transformers on prototype or field test installations, it is expected that this compromise will not affect the overall reliability of the transmitter. Conductivity-Type, Single-Point Molten-Salt Level Probe Modification and repair of the fuel flush tank level probe excitation cable, which failed in service due to oxidation and embrittlement,17 was completed during the precritical shutdown in March. Repairs were accom- plished by replacing the copper-clad, mineral-insulated copper wire exci- tation and signal cables and portions of the probe head assembly with a stainless-sheathed, ceramic-beaded nickel-wire cable assembly. Consider- able difficulty was experienced with breakage of the nickel wire in this assembly during fabrication and installation. Investigations showed that Tthe breakage was due to embrittlement produced by sulfur which originated in potting materials (Thermostix and Ceramicite 200) used to seal the cable against leakage of gas from the probe head. Although the amount of sulfur found in these materials was very small and the materials had been successfully used with other material (such as copper, stainless steel, Inconel, Chromel, and Alumel), the quantities of sulfur present were ap- parently sufficient to produce severe embrittlement of the nickel wire. Since the potted seals were being installed only as a precautionary measure and were not necessary for either contaimment or performance of the in- strument, the seals were eliminated. Performance of the repaired probe and of other probes installed in the MSRE was satisfactory during subsequent critical and low-power opera- tions of the reactor. 72 Inspection of the cable assemblies on probes installed in other fuel and coolant system drain tanks showed that no damage had occurred in those installations and that no further probe modifications were needed. oingle-Point Temperature Alarm Switches Modification of the temperature alarm switch modules to eliminate the spurious set-point shifts experienced during precritical operationsl8 was completed. Printed circuit-board contacts were gold plated to reduce contact resistance, the trim pots used for hysteresis adjustment were re- placed with fixed resistors, and resistor values in modules having ambig- uous (dual) set points were changed to restore the proper bias levels. It is not known at this time whether these modifications have completely eliminated the drift problems. The necessity of resetting many of the modules during critical and low-power operation and of minimizing inter- ference with reactor operations prevented the accumulation of the data required to determine whether further shifts had occurred. There was some evidence of set-point shifts in two of the freeze- valve control modules; however, since the module could not be checked during operation, it was not established whether the shifts were due to module malfunction or poor connections in the thermocouple circuits. Checks made during the prepower shutdown showed that additional modules had developed dual set points; however, there was no evidence of recur- rence of dual set points in modules which had previously been modified to eliminate this problem. The occurrence of dual set points is presently believed to be due to component aging and is expected to diminish as com- ponents stabilize. Module set points are being rechecked, using improved procedures, as settings of various channels become firm. Associated thermocouple in- put circuits are being checked to determine whether poor or intermittent connections exist which could cause set-point shifts. Observation of the performance of the temperature alarm switches will continue during power operation of the reactor. Helium Control Valve Trim Replacement Results of investigation of the cause of failure of four helium con- trol valves during MSRE precritical 0perationsl5 indicate that the severe galling between the 17-4 PH plug and Stellite 6 seat was probably caused by misalignment and the complete lack of lubricant rather than by incom- patibility of the plug and seat materials. Inspection of the defective valves revealed no serious misaligmment in the body, bonnet, or stem as- semblies. The original trim was damaged too badly to determine whether misalignments had existed; however, inspection of replacement trim re- vealed that (in some cases) the hole in the seat was not perpendicular to the seat within the tolerance required to prevent excessive side forces. oeveral alternative trim material combinations were tested to determine whether they would be less susceptible to galling under the conditions of extreme cleanliness (no lubricant) and dry helium atmosphere. The com- binations tested were Stellite to Stellite, chrome-plated 17-4 PH to 73 Stellite, and 440C to Stellite. None could be cycled more than 100 times without evidence of incipient damage even when perfect alignment was main- tained. However, the 440C to Stellite and the 17-4 PH to Stellite combi- nations withstood 5000 additional cycles without evidence of further dam- age after the plug was lubricated with a minute amount of light-grade machine oil., There was some evidence that the properties of the 440C to Stellite combination may be superior to those of 17-4 PH to Stellite. Thermocouple Development and Testing Drift Tests. Calibration drift of eight metal-sheathed, mineral- insulated Chromel-Alumel thermocouples fabricated from material randomly selected from MSRE stock has remained within the limits previously re- ported.lg Figure 2.12 shows the drift of these thermocouples since the start of the test. ORNL-DWG 65-8089A /fi‘\~f I ‘/f:_f AN 8 Tol I, %’ =1 /i/f\'# fi \}fi\i\f”? o — | — ‘_—“‘—'—_—_—__'_-_—"—"— ERROR (°F) 3/8 % OF 1250°F TEST TEMPERATURE : 1250°F > TEST ATMOSPHERE : AIR O 50 100 150 200 250 300 350 400 450 500 550 600 620 700 TEST TIME (days) Fig. 2.12. Average Drift of Eight MSRE-Type Thermocouples. Thermocouples on the Prototype Pump Test Loop. Observation of the performance of ten MSRE prototype surface-mounted thermocouples installed on the prototype pump test loop was terminated in July, when the section of pipe on which the thermocouples were installed was removed from the loop. Performance of these thermocouples continued to be satisfactory to the end of the test. Coolant-5alt Radiator AT Thermocouples. Investigation of the effects of mismatch of thermocouple and thermocouple lead-wire materials on the accuracy of MSRE coolant-salt radiator AT measurement?® was continued. Tests performed at the MSRE showed that the effects observed in the lab- oratorygo were present in the MSRE thermocouple lead-wire installation and that excessive noise was present on the signal. The existing lead wire was replaced with a continuous run of higher-quality shielded lead wire., Tests performed after the shielded lead wire was installed indi- cated that the long-term drift previously observed had been eliminated T4 but that excessive intermittent noise was still present. The thermocouple lead wire was then insulated to eliminate ground loops. Further tests are being performed to determine whether the noise has been eliminated. References 1. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 29. 2. Ibid., p. 8. 3. Ibid., p. 12. 4s R. Do Ackley and W. E. Browning, Jr., Equilibrium Adsorption of Kryp- ton and Xenon on Activated Carbon on Linde Molecular Sieves, ORNL- CF-61-2-32 (Feb. 14, 1961). 5. MSRE Design and Operations Report. Part V, Reactor Safety and Anal- ysis Report, ORNL-TM-732, p. 269, 6. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 50, 7' Ibido) po 500 8. Ibid., p. 51. 9. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, pp. 155— 60. 10. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 51— 52. 11l. MSR Program Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, pp. 123— 24 12. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 52. 13. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 166, 14. MSR Program Semiann. Progr. Rept. Feb., 28, 1965, ORNL-3812, pp. 38— 40, 15. Ibid., p. 48. 16. Ibid., p. 4l. i7. Ibid., pp. 4l—42. 18. Ibid., pp. 46—47. 19. Ibid., p. 43. 20. TIbid., p. 44. 75 3. MSRE REACTOR ANALYSIS Theory of Period Measurements Made with the MSRE During Fuel Circulation Zero-power kinetics experiments performed with the MSRE while the fuel circulation is stopped can be analyzed by means of the conventional inhour equation, which relates the measured asymptotic time dependence of the neutron flux to the reactivity.l When the fuel is circulating, however, the conventional analysis becomes inadequate due to the trans- port of delayed neutron precursors by fluld motion and subsequent emis- sion of the delayed neutrons in regions where they either do not contribute at all to the chain reaction (i.e., in the external loop) or contribube with substantially different weight than the prompt fission neutrons (snift in the spatial distribution of emission within the reactor core). These effects have long been recognized, and various approximations have been used for their representation in reactor kinetics analysis.2 The particular effect of the shift in the spatial distribution of delayed neutron pre- cursors relative to the prompt fission neutrons has been the subject of recent studies by'Wolfe3 and Haubenreich.# Wolfe employs a perturbation approach and obtains an inhour equation for a slab reactor, through which fuel circulates in the direction of axial variation of the neutron flux. In an independent approach, Haubenreich considers an explicit analytical model representing the MSRE in the circulating, just-critical condition and, by means of a modal analysis, obtains effective values of delayed neutron fractions for this condition. In the latter analysis, a bare- cylinder homogenized approximation was used to represent the MSRE core, with boundaries corresponding physically to the channeled region of the actual core. We have extended the preceding analyses to include the contribution of delayed neutrons emitted while the fuel is in the upper and lower ple- nums and to include the case in which the flux 1s changing according to a stable asymptotic period. The result of this analysis is an inhour-type equation, relating the measured period to the static reactivity of an equivalent state in which the fuel is not circulating. The static re- activity, p_, is defined by the relation” vV — v . C ps vV J in which VvV is the physical energy-averaged number of neutrons emitted per fission and V., 1s the fictitious value for which the reactor would be "virtually'" critical with the fuel stationary. Since the static re- activity is the quantity normally obtained in reactor calculation pro- grams, it 1s most convenient to relate this quantity directly to the asymptotic period. This analysis is summarized below. Only the essen- tlial steps are indicated in obtaining the inhour relation. Specific approximations sultable for MSRE analysis, together with preliminary nu- merical results, are given following the discussion of the general prob- lem. Further details of the analysis will be included in a subsequent report. 76 When the flux and precursor densities are behaving in time as ewt Tthe general reactor equations, written to include the transport of de- layed neutron precursors by fluid motion in the axial direction, are: J 6 — - = -1 Do + (1 BT)fPP¢ + L A f vrwe (1) . diCi 1 d o B;Pe — AC, — 5= (vci) = uC, i=1,2, oo, 6 . (2) The symbols ¢ and C represent the neutron flux and delayed precursor densities, which in general are functions of position, energy, and angle variables. The operators D and P, assumed time independent, represent neutron destruction (leakage, absorption, energy transfer by scattering) and fission production processes, respectively. Their explicit repre- sentation depends on the model used in analysis. In Egs. (1) anda (2), 1 — Bp is the fraction of all neutrons from fission which are prompt, and PB; and %i are the fractional production and decay rate for the ith precursor group. The quantities fp and f5 are energy spectrum operators which multiply the total volumetric production rates of prompt and de- layed neutrons to obtain the net production of neutrons of a specific energy. The symbols v and V represent the neutron velocity and the fluid velocity, respectively. As defined above, the static reactivity is the algebraically largest eigenvalue of the equation: ~Do_ + (1 - ps)fP¢S =0, (3) where {5;\ T = (1 —-fiT)fp + fii B.f . - (4) In order to relate the reactivity, Pgs» to the stable period, wrl} use 1is made of the static adjoint flux, ¢§, the solution of the adjoint equation corresponding to Eq. (3), as a welghting function in the integration of Bgs. (1) and (2). The purpose of this procedure is to convert the re- actor equations from a form which involves only local reactor properties to a relation which utilizes global, or integral, quantities. This anal- ysis 1s similar in principle to that for stationary fuel reactor systems, details of which have been presented in several sources (see, e.g., ref, 5). By forming the inner product of of with Eq. (1), that is, by multi- plying Eq. (1) by ¢& and integrating over the position, energy, and angle variables, one obtains the relation 6 6 + (o1, vo) .Z; Bi(¢g, £4;P¢) —-.Z; A (o, £4:C;) 0 = — . 1=1 ~ 1=1 5 (5) ® (of, TPo) (¢, TPo) 77 where the symbol (x, y) denotes the inner product of the two functions x and y. In the special case when w = 0, this equation gives the static reactivity difference between the stationary-fuel and circulating-fuel critical states. In the general case, the dependence of the precursor densities on w is given implicitly by Eg. (2). Several approximations sultable for MoORE analysis have been made in performing the integrations of Egs. (2) anda (5): (a) The shape of the asymptotic flux distribution, ¢, is assumed to be sufficiently well approximated by the static flux, o5, corresponding to the stationary fuel state. This is essentially the perturbation approxi- mation. (b) The correction for the difference in the energy spectra for emission of prompt and delayed neutrons appearing in Eqg. (5) can be cal- culated approximately as a separate step. This is done by reducing the age for the ith group from that of the prompt neutrons, and using the relative nonleakage probability factors appropriate to a bare reactor. This is equivalent to modifying the values of the static fraction, B . The net correction for the differing "energy effectiveness" of the de- layed neutrons is small compared with the spatial transport effects under consideration. (c) The principal difference in the spatial distributions of prompt and delayed neutron emission 1s in the direction of fuel salt flow. The veloclty profile was assumed to be flat in the radial direction across the core. Also, for this initial study, radial averaging of the neutron %r§duction rates was neglected in calculating the inner products in Eq. 5). ORNL-DWG 6512048 0.8 0.7 - / . // os / Y/ v 2 05 / - /| C / = 0.4 ~ /| 2 yd Ll L * 03 7 S T —“‘/ 0.2 =T O 10?5 2 1021072 o 5 1072 2 5 10! 2 5 100 2 5 10! w, FREQUENCY (sec™!) Fig. 3.1. Static Reactivity Difference Between otationary-Fuel Critical Condition and Circulating-Fuel Transient Condition vs Stable Inverse Period Observed During Circulation. 78 Equation (2) was integrated numerically, using a three-region approxi- mation for the MSRE core, representing the channeled region and the upper and lower plenums. The fluid axial velocities in each region corresponded to the average fuel residence times for that region. The axial distribu- Ttlions of the static neutron flux and adjoint flux corresponded to the case of all three control rods fully withdrawn from the reactor core. Lffective fuel residence times in the core regions and the external loop were ob- tained from ref. 6. The numerical results obtained from the analysis are summarized in Fig. 3.1, in which the static reactivity is plotted as a function of the inverse asymptotic period, w. Only the range of values of w important to experimental measurements of the stable period were con- sidered in this initial study. The calculated reactivity difference be- tween the static-fuel and circulating-fuel critical conditions (0.22%, corresponding to w = 0 in Fig. 3.1) agrees favorably with the value of 0.21%, measured by excess addition of 23°U during the zero-power experi- ments. This analysis is now being applied to rod calibration measure- ments made during fuel circulation, and the results will be included in a future report. References 1. 5. Glasstone and M. C. Edlund, The Elements of Nuclear Reactor Theory, chap. X, Van Nostrand, New York, 1952. 2. dJ. A. Fleck, Jr., Theory of Low Power Kinetics of Circulating Fuel Reactors with Several Groups of Delayed Neutrons, BNL-334 (April 1955). 3. B. Wolfe, Nucl. Sci. Eng. 13, 80-90 (1962). 4. P. N. Haubenreich, Prediction of Effective Yields of Delayed Neutrons in the MSRE, ORNL-TM-380 (October 1962). 5. T. Gozani, "The Concept of Reactivity and Its Application to Kinetic Measurements, " NUcleonik!g(Z), 55 (1963). 6. R. B. Lindauer, internal memorandum (June 1964) , p. 8. Part 2. MATERIALS STUDIES 81 4o METALLURGY Dynamic Corrosion Studies A test program is in progress to study the compatibility of struc- tural materials with fuels and coolants of interest to the Molten-Salt Program. Thermal convection loops described previouslyl are used as the standard test in this program. Presently, two long-time tests are in operation with fuel salts at conditions reported in Table 4.1. Table 4.1. Operating Conditions for Thermal Convection Loops Containing LiF-BeF-ZrF 4-UF 4-ThF, (70-23-5-1-1 Mole %) Loop 1255 Loop 1258 Loop material INOR-8 Type 304 SS Maximum salt temperature 1300°F 1250°F AT 160°F 180°F Operating time as of 29,688 hr 18,312 hr Auvg. 31, 1965 Insert specimens in INOR-8-2% Cb Type 304 SS hot leg The INOR-8 loop contains insert specimens made of INOR-8 modified with 2% Cb to improve weldability and mechanical properties. In addition to studying the corrosion resistance of this modified alloy, the loop serves to demonstrate the long-time compatibility of INOR-8 with an MSRE- type fuel salt. The type 304 stainless steel loop is being operated to investigate the potential of this lower-cost alloy for molten-salt applications. This material 1s being reconsidered at present in view of the lower tempera- tures of interest for coolants and the improvements in salt processing that result in lower corrosion rates. Specilmens were removed from the stainless steel loop after 15,000 hr of operation and examined metallographically. A maximum attack of 0,002 in. was observed that was generally intergranular in nature (see Fig. 4.1). This loop is continuing to operate. 82 .035 INCHES N 100X [ \ Fig. 4.1. Appearance of Specimen Removed from Region of Maximum Temperature (1250°F) in Thermal Convection Loop 1258 Made of Type 304 otainless Steel. The loop had operated for 15,000 hr at this time. Corrosion Studies Using Lead as a Coolant Six uninhibited thermal convection loops were started at ORNI to ex- plore the compatibility of structural materials with lead at conditions expected for molten-salt reactor coolants. The operating conditions of these loops are summarized in Table 4.2, along with results of a 2-1/4 Croloy steel loop run at Brookhaven National Laboratory2 in which lead was circulated at 1022°F for several years and in which no measurable cor- rosion was Observed. Two of the ORNL loops (one of type 410 stainless steel and of 2-1/4 Croloy steel) had Cb—1% alloy liners in the surge tanks within which MSRE fuel salt floated on the lead surface, as shown in Fig. 4.2. These tests also contained graphite specimens in the surge tank which were exposed to the salt, to the lead-salt interface, and to the lead. A hot-leg tem- perature of 1200°F was maintained in the loops with a 300°F AT. Two other loops that contained no salt, Cb—1% Zr, or graphite, were operated at 1100°F with a 200°F AT. One of these loops was constructed of 2-1/% Croloy steel, the other of low-carbon steel. A fifth loop was operated to test the compatibility of columbium alloys with lead. This loop was made from Cb—1% Zr clad 446 stainless steel and contained no samples or inhibitors. Of the two loops that operated at 1200°F and contained salt, the 2-1/4 Croloy steel plugged after 288 hr, and the type 410 stainless steel plugged after 1346 hr. The plugs were made up of dendritic crystals, 83 Table 4.2. Operating Conditions of Thermal Convection Loops Circulating Lead Maximum AT Operating Loop Material Temperature ° Time Comments o (°F) (°F) (hr) Croloy 2-1/4% 1022 227 27,765 25 ppm Mg, Zr inhibitor added Croloy 2-1/4 1210 300 266 Graphite, fuel salt, Cb—l% Zr placed in surge tank ATST type 410 SS 1215 300 1,346 Graphite, fuel salt, Cb—1% Zr placed in surge tank Croloy 2-1/4 1100 200 5,156 Low-Carbon Steel 1100 200 5,064 Co—1% Zr (clad with SS) 1400 400 5,280 Croloy 2—1/4b 1200 230 1,848 Zr specimen in hot leg aBrookhaven National Laboratory loop. bPrototype test of new loop design. which were determined to be iron and chromium by x-ray diffraction and chemical analysis. The maximum depth of attack in the hot leg of the 410 stainless steel was 0.002 in. and on the 2-1/4 Croloy loop was 0.001 in. Typical attack is shown in Fig. 4.3. No corrosion was observed on Cb—1% Zr liners., The 2—1/4‘ Croloy steel loop that operated with a hot leg at 1100°F showed signs of plugging after 648 hr of operation. A radiographic ex- amination of the dumped loop at this time indicated several areas where lead had wet the metal, indicating selective attack. The lead from the loop contained crystals of iron and chromium. The loop was refilled with new lead and operated for a total of 5156 hr. Metallographic examination of the loop revealed large amounts of dendritic crystals in the cold leg and also revealed hot-leg attack to a depth that varied from 0.004 to 0.008 in. The crystals in the cold leg were similar in composition to the original alloy. The carbon-steel loop was constructed of a larger-diameter pipe and operated for 5000 hr before shutdown. However, prior to being shut down, it did show some signs of restricted flow. Metallographic examination showed a maximum attack of 0,010 in., similar in nature to that found in the 2-1/4 loops. The Cb—1% Zr loop operated for 5280 hr with a hot-leg temperature of 1400°F and a 300°F AT. Metallographic examination showed no hot-leg attack or crystals in the cold leg. Results on the tests made to date indicate that uninhibited lead will cause excessive corrosion in low-alloy steel and stainless steel systems at temperatures above 1100°F., However, Cb—1% Zr appears to be a promising container. Inhibitors should be investigated in steel systems in view of the BNL results. 84 | . PHOTO 74545A VENT LINE . . PROBE LINE TO DETERMINE ~ LIQUID LEVEL FILL LINE GAS LINE STOP TO HOLD LINER IN PLACE Cb—4 Zr LINER LEAD SALT INTERFACE TOP OF HOT LEG Fig. 4.2. Section Through the Surge Tank Used on the Type 410 octainless Steel Loop. Notice that the salt is floating on the lead and that the only metal which contacts the salt is the Cb—1% Zr liner. The graphite specimens were suspended from the small wire running into the salt and cannot be seen in the picture. 85 Ton 0.007 IN HES > 500X Fig. 4.3. Metallographic Section of a 2-1/4 Croloy Loop Run in Lead at 1200°F for 266 hr. Unetched. Revised Thermal Convection Loop Design The thermal convection loop was redesigned for studying coolants and now includes the following features: 1. removable hot-leg samples (for weight gain data and metallographic analysis), 2. 7portable thermocouple control in the hot leg to plot temperature pro- files, 3. surge tank with Cb—1% Zr liner and graphite to allow the presence of fuel salt, 4., smaller size to allow radiographing the complete loop at once, 5. sampling devices for both lead and salt, 6. a method by which the lead may be drained and refilled without af- fecting the salt so that the hot leg can be radiographed to find se- lective attack and so that mass transfer may be removed by gravity separation, 7. better control of the cold-leg temperature. 86 Figure 4.4 shows the revised thermal convection loop design. In order to obtain better control over the cold-leg temperature, a heater was placed on the cold leg upstream from the coldest part of the loop. This heater 1s controlled by the cold-leg thermocouple. A shut- down device to eliminate burnout was also incorporated which turns the power off when the temperature of the cold leg is equal to the temperature of the hot leg. With this design, the degree of plugging may be deter- mined by the power consumption of the cold-leg heater. A prototype loop of the new design was run at conditions listed in Table 4.2 to test the new features of the loop, the temperature distribu- tion, and the sample removal and cleaning techniques. The cold-leg tem- perature control worked well, as did the sample-removal and dumping fea- tures. The samples were cleaned by amalgamating the lead with mercury ORNL—DWG 65—-4744 /,-in. CONNAX FITTING Hrkp FILL FILL : AND SAMPLE~_ “AND SAMPLE EVACUATION LINE WCORCHESTER {-in. BALL VALVE SURGE TANK HOT LEG CONTROL THERMOCOUPLE GAS V4-in. BUTT WELD SWAGELOCK 34 in. OD x.0.089 in. WALL THICKNESS SPECIMENS SUSPENDED FROM END 4-in, - o CLAMSHELL -in. HEATER—— SCHED-40 PIPE | 4Y/p-in. LONG 6-in. CLAMSHELL : HEATERS { {“_—— T T —_——_-_-_—.]I i C—C— a3 e )~ e ] Lo e \J = E - GRAPHITE 71| 1 COLDLEG -1 Zr REMOVABLE ' CONTROL LINER SAMPLES . THERMOCOUPLE i ////) SLOTS IN GRAPHITE e ENSURE LEAKAGE SALT 4-in. | LT ‘ R CLAMSHELL HEATER . . LEAD SLOTS IN DUMP TANK—~] Cb-1 Zr LINER ENSURE LEAKAGE LOOP PIPING THERMAL CONVECTION LOOP SECTION THROUGH SURGE TANK USED FOR LEAD CORROSION STUDIES OF THERMAL CONVECTION LOOP USED FOR LEAD-SALT STUDIES Fig. 4.4. Sketch of Prototype Loops Used to Study Lead-Salt Low- Alloy Steel Systems. 87 and then dissolving the amalgam with concentrated nitric acid. A re- movable sample made of zirconium was included in the test and did not change weight. The loop plugged after 1848 hr of operation and showed a maximum attack of 4.5 mils. This high rate of attack indicates that the zirconium metal specimen did not inhibit corrosion in the loop. MSRE Materials Surveillance Testing The MSRE surveillance specimens® of INOR-8 and grade CGB graphite have been assembled and placed in the reactor and control test systems. These specimens will be used to make periodic surveys of the effects of the reactor operations on the materials from which the reactor and moder- ator were constructed. Three stringers of graphite specimens, each with a pair of INOR-8 Tensile specimen assemblies and flux monitors, were placed in the central position of the MSRKE core and extend axially the full height. Matching sets of graphite and INOR-8 specimens were installed in the control test rig, which will subject these control specimens to molten fuel salt under approximately the same temperature profile and major temperature fluctua- tions experienced by the reactor specimens.4 Three other sets of INOR-8 Tensile specimens and flux monitors without graphite specimens have been installed outside the reactor vessel adjacent to the reactor flow dis- tributor. The specimens in the reactor core will be subjected to a range of temperatures and neutron fluxes (shown in Fig. 4.5) that bracket the tem- peratures and fluxes of the reactor core materials. Thus results from ORNL-DWG 65-12049 //\QST FLUX 1220 (x10'3) o T - \ GRAPHITE TEMPERATURE o >\ (APPROX) n —— T — . ~ | - // | 1210 5 / L | ——==—FUEL . P _ TEMPERATURE _ - 7 ~ L 3 P 2 6 — g 1200 W £ _ _ 5 x v ~ < 5 - > 5 L 4 -~ . \ oo & z / // // ~ U§J o & s - \ = 5 7~ | 2 //,‘ ' THERMAL FLUX— 3 1180 - 0 170 0 10 20 30 40 50 60 70 80 DISTANCE ABOVE BOTTOM OF HORIZONTAL GRAPHITE GRID (in.) Fig. 4.5. Temperatures and Neutron Fluxes Along Graphite-Sample Assembly. Control rods at upper limits. 88 these specimens will anticipate the condition of the reactor vessel as well as match the condition of both the thimbles and the reactor vessel. The specimens adjacent to the reactor vessel will be exposed to neu- tron fluxes that more nearly approach those at the reactor vessel wall. Also, the sgpecimens will not be exposed to the salt environment, facili- tating any comparisons of test results with data developed in the Irradi- ation Test Program. INOR-8 Surveillance Specimens The INOR-8 specimen assembly consists of 27 tensile bars approxi- mately 2 in. long and with a 0.125-in. gage diameter welded end to end. The first and the 27th specimen are machined from weldments. The lengths of the sections between tensile bars have been adjusted to hold the graph- ite specimens firmly in the final assembly. The specimens in the core and control test were made from heats Ni- 5085 and Ni-5081 material. Ni-5085 is the heat from which most of the reactor vessel shell is made; Ni-5081 is one of the original heats used to develop much of the mechanical properties data for INOR-8. Both heats have been investigated in the INOR-8 radiation damage program. The weld specimen was made with heat Ni-5055 weld rod and heat Ni-5085 base metal. The specimen assembly rods outside the reactor vessel and the flux monitors are 82 in. long and parallel the vertical axis of the reactor. These specimens are similar in shape to the reactor core specimens but are made from heats Ni-5085 and Ni-5065. Ni-5065 is the heat used to make the reactor top and bottom heads and is also being tested in the radia- tion-damage program. The analyses of the INOR-8 specimens will include: (1) metallo- graphic examination for structural changes, corrosion effects, and pos- sible layer formations; (2) tensile and creep properties, with emphasis on creep ductility; (3) a general check for material integrity and di- mensional changes; and (4) chemical analyses for compositional changes and fission product deposition. Graphite Survelllance Specimens The graphite specimens were machined perpendicular to and parallel with the extrusion (grain) directions of the graphite bars listed in Table 4.3. The stringer bars are the vertical bars with grooved fuel channels that constitute about 98% of the moderator volume. The lattice bars are the horizontal criss-crossed bars that support the stringers. The lattice bars were fabricated with a slightly higher permeability than the stringer bars in order to secure maximum structural integrity. 89 Table 4.3. Typical Data on Grade CGB Graphite Used to Fabricate MSRE Surveillance Specimens Electrical Resistivity bulk (microhm-cm) Bar No. Lot No. Bar Type Density - - (g/ch) Parallel Perpendicular 635 8 Stringer 1.853 640 1210 1229 11 Stringer 1.848 690 1400 1559 13A Lattice 1.868 635 1564 13A Lattice 1.881 630 “The orientation is with respect to the extrusion, the grain, or ag direction. There appears to be somewhat more turbostratic graphite in some of the stringer bars. Since this less-graphitic material might show more irradiation-damage effects than the more crystalline graphite, specimens containing more than the normal quantity of turbostratic graphite were included with the surveillance specimens. Specimens from bars Nos. 635 and 1229 in Table 4.3, respectively, represent the stringer bars and the slightly-less-graphitic graphite. The higher electrical resistivity of bar No., 1229 reflects the latter. Typical graphite specimens used in a set of surveillance specimens are shown in Fig. 4.6. All these can be utilized in the (1) metallo- graphic examination for structural changes and material deposition; (2) radiographic and autoradiographic examination for salt permeation and possible wetting effects; (3) dimensional checks for shrinkage effects; and. (4) measurement of physical properties such as electrical conduc- Tivity, thermal conductivity, and Hall coefficient. Chemical analyses for salt and fission product deposition will be made on the larger pieces, A and B in Fig. 4.6. These have cross sections of 0.470 X 0.660 in. The lower specimen in Fig. 4.6A is from the lattice bar and will be located at the bottom of the moderator core. The specimen at the top of Fig. 4.6A and. the specimen B in Fig. 4.6, respectively, are located at the top and center of the moderator core and are from stringer bars. The specimens shown in Fig. 4.6C are essentially multiples of a basic 0.11 X 0,47 X 2,25 in. chape, the smallest pieces. These are to be used to make flexural strength and modulus of elasticity measurements. The small specimens can be tested directly, while the long specimens will have to be cut to the proper lengths. The objective was to preclude machining of irradiated graphite in hot-cell facilities. Such machining would be difficult and could modify some of the property changes in the graphite. These specimens, as shown in Fig. 4.6C, when placed together create a cross-section dimension that matches those of the larger pieces discussed above, 20 Y-64822 C A B Fig. 4.6. Typical Graphite Shapes Used in a Stringer of Surveil- lance Specimens. Assembly of Surveillance Specimens A subassembly of graphite and INOR-8 surveillance specimens bound. with INOR-8 straps is shown in Fig. 4.7. For the reactor, three sets were bound together and placed into a perforated tube as shown in Fig. 4.8. The specimens were sealed in the container by the ball lock assembly (Fig. 4.8b) and placed into the reactor (Fig. 4.8c). Similarly, the matching control specimens were assembled in three sets; however, instead of being bound together and placed into a common container, each set was placed into a separate, sealed perforated con- taingr as shown in Fig. 4.9. These were placed in the controlled test rig. The flux monitors included in the assemblies placed in the core and just outside the reactor vessel are 0.020-in.-diam wires of pure iron, pure nickel, and a type 302 stainless steel. The type 302 stainless wire 91 PHOTO 80761 Fig. 4.7. Three Stringers (S, R, and L) of CGB Graphite and INOR-8 Surveillance Specimens. PHOTO 81671 SPACER AND GUIDE PIN UPPER GUIDE ———s= BASKET LOCK\ ASSEMBLY INOR-8 ROD OF TENSILE SPECIMENS GRAPHITE (CGB) SPECIMENS BINDING STRAP BASKE T—_ FLUX MONITORS TUBE BASKET AN \ BALL LOCK ASSEMBLY REACTOR ¢ CONTROL ROD GUIDE TUBE TYPICAL GUIDE BAR FUEL CHANNEL I REACTOR ¢ o \ Oc 0.400 in. () SURVEILLANCE SPECIMENS 2in. TYPICAL (c) R=REMOVABLE STRINGER Fig. 4.8. 1INOR-8 and Grade CGB Graphite Surveillance Specimens and Container Basket. (a) opecimens partly inserted into the container. (b) Container and its lock assemblies. (c) Location of surveillance specimens in the MGRE. was selected for 1its uniform cobalt content. The selection of materials for flux monitors was limited by the relatively high temperature, 1225°F, of the reactor. R PHOTO 81672 Y-66055 LIFTING TOOL ~ “ SPECIMENS T~ BASKET (a) Fig. 4.9. One Stringer of Control Specimens of INOR-8 and Grade CGB Graphite for the MSRE Surveillance Specimens Shown with Its 73-1/2- in.-long Container. (a) Bottom view. (b) Top view. The flux monitor wires installed outside the reactor vessel are bare, while those placed near the central part of the moderator are protected from exposure to the salt. These have been sealed into an evacuated, 1/8-in.-0D by 0.20-in.-wall INOR-8 tube by tungsten—inert-gas arc welding. At an appropriate time determined by the radiation-damage program, examinations as described above will be made on a set of specimens from each of the three test sites: (1) the central part of the reactor moder- ator, (2) the controlled test rig (the control test for those from the reactor), and (3) outside but adjacent to the reactor vessel. A set of specimens constitutes one-third of the total at each test site. As these are removed, sets of specimens of like or of advanced materials will be placed into the empty positions. The sampling and replacing of specimens will be sequential for each one-third set at each test site. 93 kEvaluation of Radiation-Damage Problems of Graphite for Advanced Molten-Salt Reactors Recent progress in developing and testing nuclear grades of graphite provides a basis for specifying the properties and estimating the life of a graphite for advanced molten-salt reactors such as the MSBR. There are, of course, several uncertainties which limit the ability to state categorically that any graphite will withstand the MSBR environment. The graphite in the regions of highest flux in an MSBR will be irradiated to doses of 2 to 5 X 10%% nv (>0.18 Mev) per year, and there is no evidence to demonstrate that any graphite can withstand massive doses of >10%2 nvt and still retain its integrity. Isotropic graphite has demonstrated the greatest potential for accommodating this condition. The tubular, thin- walled design that is proposed for the MSBR is one of the better config- urations for reducing stresses resulting from differential growth of the graphite. The tubular shape also is easier to fabricate and to test non- destructively to ensure maximum integrity. The pertinent properties of the graphite grade that could be devel- oped for the MSBR can be projected from those of available grades with some degree of certainty. The main questions to be answered relate to the growth and strength of graphite under irradiation. Values for growth and strength can be estimated fairly well to dose levels where experi- mental evidence exists; however, it would be presumptuous to extrapolate these growth and strength values beyond dose levels of 3 to 4 X 102 nvt. The magnitude of the differential growth problems depends on the growth rate, creep coefficient, flux gradient, and geometric restraint in the particular structural component. With a tube 4 in. in outside diameter, a flux drop of approximately 10% could be expected across a l/Z-in. wall. The growth rate for an isotropic graphite at 700°C is con- servatively estimated to be about half®’? that for grade AGOT or about 2.4 %X 107%% (in./in.) (nvt)™L. The creep coefficient® should be about that for grades AGOT and CGB or about 7 X 107%° (in./in.) (psi)~! (nvt)~ 1. The geometric restraint for a tubular configuration of this size 1is very close to 1/2. The problem is solved by obtaining the maximum strain rate caused by the differential growth, which is: s (é?i) (6)(R) = #1.2 X 10725 (in./in.) (avt)~! , ¢ where ¢ = mechanical strain rate, 4%12 flux drop, 0.10, G = growth rate, —2.4 X 10724 (in./in.) (nvt )™, R = restraint factor, *1/2. 9 This strain rate is maximum on the inside surface and is minimum on the outside surface as in a tube with a thermal gradient induced by heating from the outside. The maximum stress is then obtained from the creep co- efficient. —25 - L2 X 10 = 1.7 x 10° psi , 0.7 x 10~<8 Mo where o 1s maximum fiber stress, and K is creep coefficient. The bend fracture stress of isotropic graphite is expected to be >5000 psi; so the stresses induced by differential shrinkage should not cause the tubes to fail, at least to doses of 4 X 10°% nvt. Also, there is some evidence”? that the shrinkage rate of isotropic graphite diminishes after 1 to 2 X 102 nvt. In effect, this will reduce the stress level and introduce more conservatism into the extrapolation to higher dose levels. The ability of the graphite to absorb the creep strain regardless of the stress intensity has been demonstrated. ’? 8 Although a strain lim- itation to fracture has not been demonstrated, one probably exists. This limitation, however, is greater than the 2% tensile strain observed by Perks and Simmons.,’ Thus, to obtain a 2% maximum fiber strain, it would require a 1.6 X 10°2 nvt dose if the shrinkage rate remains constant. Existing data show, then, that irradiation effects should not produce failures in one year in the graphite in the highest flux regions of a large MSBR. The data can be used to infer a life of several years. The actual life of the graphite will, however, remain uncertain until its ability to withstand irradiation damage beyond 1023 nvt has been demon- strated. Mechanical Properties of Irradiated INOR-8 Mechanical Properties of Unirradiated INOR-8 Used in the Reactor Vessel A program has been initiated to evaluate the properties of the heats of INOR-8 used in constructing the MSEE reactor vessel. Although mechan- ical property tests were conducted previously on several MSRE heats,g the radiation effects were not superimposed. The present studies include further mechanical property tests on the heats of material comprising the reactor vessel in both the wrought and welded conditions to evaluate their properties under service conditions. The obJjectives of this study are to estimate the safe operating life of the MSRE and to determine whether any reasonable steps can be taken to increase this 1life, oignificant experiments and conclusions in this program include the following: 95 1. The creep strengths of the wvarious heats of material appear to be comparable with those reported previously.9 However, the creep-rupture data appear to be more accurately represented by an equation of the form e’ = AtB than the more conventional form of o = AtB. Hence, the extrapo- lated values are altered slightly. This is illustrated by the data shown in Figs. 4.10 and 4.11 for heat Ni-5065. 2. The properties of the alloy at 650°C are very sensitive to me- chanical and thermal treatments. The data in Table 4.4 show that, in general, both annealing at temperatures below 1600°F and cold working im- prove the rupture life and ductility. If working and annealing are con- tinued sufficiently to cause recrystallization, the strength is reduced. Annealing at 2300°F caused a significant reduction in rupture life. 3. Welding the air-melted heats of material caused a large reduction in rupture life and ductility. This is also illustrated by the data in Table 4.4. The weld between the top head and the flow distribution ring of the MSRE, which was not stress relieved, was reproduced as nearly as possible. ©Since this weld was found to have very poor properties in the as-welded condition, the rate of recovery of strength and ductility was measured at several temperatures. The results of these tests are given in Figs. 4.12 and 4.13. Properties similar to those of the base metal can be obtained by stress relieving for 8 hr at 1600°F or for 50 to 100 hr at 1400°F, 4o Welding of a vacuum-melted heat of INOR-8 was found to produce very small changes in properties. This is illustrated in Fig. 4.14. 5. Microprobe and autoradiographic studies have shown that the large precipitates in INOR-8 are intermetallics rather than carbides. The structure is typified by the microstructure shown in Fig. 4.15. When the alloy is heated to about 2500°F, the microstructure reverts to that shown in Fig. 4.16. A grain-boundary lamellar phase results, with some local- ized melting. The transition from one phase to another is illustrated by the photomicrograph of a fusion line shown in Fig. 4.17. Table 4.5 pre- sents the results of microprobe analyses on both types of precipitate. The precipitates are basically nickel-molybdenum intermetallic phases with large amounts of silicon also present. The spherical phase is ap- proximately Ni-Mo (8), and the lamellar phase is Thought to be NisMo (7), since iron and chromium are both known to suppress B-phase formation in nickel-molybdenum alloys. The autoradiograph shown in Fig. 4.18 was made on a heat of INOR-8 containing the isotope 140, The light band shows where the emulsion has been scraped from the surface to reveal the la- mellar phase., The film is present on the dark area, and the darkening is produced by beta-ray emissions from the 140, The lamellar precipitate is actually depleted in carbon rather than enriched. 6. Several experimental alloys are being studied. The amount of molybdenum is being reduced to about 12 wt % to produce a solid solution, and the silicon content is being reduced to reduce the "hot-short" char- acteristics of the alloy. Various alloy additions, such as zirconium, titanium, and niobium, are being investigated as a means of improving the resistance of the alloy to neutron irradiation. 96 ORNL-DWG 65-10806 100 80 60 40 20 STRESS (1000 psi) 1 10 100 1000 10,000 100,000 RUPTURE LIFE (hr) Fig. 4.10. Stress-Rupture Properties of Hastelloy N (Heat 5065) at 650°C. ORNL-DWG 65-10805R S T A e AS RECEIVED o/ 70 ® 60 50 S 40 J STRESS (1000 psi) 30 \\\ \ N 1 10 100 1000 10,000 100,000 RUPTURE LIFE (hr) Fig. 4.11. Creep-Rupture Properties of Hastelloy N (Heat 5065) at 650°C. 97 Table 4.4. Influence of Cold Working and Heat Treatment on the Creep Properties of Hastelloy N (Heat 5065) o = 40,000 psi, T = 650°C Heat Trestment Rupt%ii)Life Elon%%gion iieiggzi?%) As received 312.3 16.6 15.6 2 hr at 871°C 502.8 48.4 40.8 8 hr at 871°C 652.3 54.7 30.3 200 hr at 760°C 566.3 46 .9 35.3 Cold worked (C.W.) 0% 179.7 9.4 8.1 C.W. 10%, 2 hr at 871°C 729.1 57.6 31.7 C.W. 10%, 8 hr at 871°C 373.5 37.5 21.8 C.W. 20% 316.1 28.1 8.5 C.W. 20%, 2 hr at 871°C 414 .3 29.7 34.1 C.W. 20%, 8 hr at 871°C 156.3 25.0 24.0 1 hr at 1260°C 34.0 A 13.2 As welded 18.7 1 0.77 . ORNL-DWG 65-10797 O 2 < Ll o 2 < = =Z 1 o i_ O D 8 0.5 o = 40,000 ps o 7 =650°C 871°C 0.2 760°C 650°C O.1 0.1 0.2 0.5 1 2 5 10 20 50 100 200 500 {000 TIME AT TEMPERATURE (hr) Fig. 4.12. Influence of Stress Relieving on the Rupture Ductility of Weld 1. (hr) RUPTURE LIFE 98 ORNL-DWG 65-10798 1000 40,000 ps 650°C 500 871°C 760°C 650°C 200 100 50 20 10 ' 041 02 0.5 1 > 5 10 20 50 100 200 500 1000 2000 TIME AT TEMPERATURE (hr) Fig. 4.13. Influence of Stress Relieving on the Rupture ILife of Weld 1. ORNL-DWG 65-10808 70 N | 2477 BASE METAL 50 40 STRESS (1000 ps1) DN 30 . 10 1 10 100 1000 10,000 RUPTURE LIFE (hr) Fig. 4.14. Influence of Welding on the Creep-Rupture Properties of Hastelloy N (Heat 2477) at 650°C. 99 | 0 | o " o S3IHONI o ° XOOl : 1 o o L o »* A % Kossorcsenaroe: e « o - e e w i X\x&\\a .w,« ) %\ !.Sez( ékfi\%«w’ W - ARSI ; N g “ . » 0 * ,\&w{ o ) » rs&, . T - J.\.t. »< m. y o e s l&,\\} > - . | i,.{z.,\‘\,,,v.. . . . § ~ : e s * \p\ 1 - \w\ . & j: ~~ o~ 0 B . < Y M . ! ,\« l» fi . 'X\ o..'«‘ Wrflw@ e w ° ..H» * * - ’ . /?((J..Mz . S * 9u o‘ s o g& \.Q % W &, Z . n}, f» &V A&o’-l & P ) .&M Ya s a‘ Ginflufl %;LA& t%.w« * °° e » ¢ ¢A, flw« 2 AT , M \M -/ \« \\ x - » \\«\ \ \N - \\‘»\. - N\ * P \\\ - ;- / o EEy e QWWP&%¥PA%W? OO»«/«}»‘ © #* Typical Microstructure of As-Received INOR-8, Showing Fig. 4.15. Large Intermetallic Prec ingers. tates in Stri ipi 100 e Fig. 4.16. Boundary Lamellar Phase and Localized Melting. Y-49700 \ é:""‘:fl. . INCHES 0.02 0.03 {0.05 Structure of INOR-8 Heated to 2500°F, Showing Grain 101 Table 4.5. Microprobe Analysis of Hastelloy N (Heat 5075) Element Bulk opherical Precipitates Lamellar Precipitates Composition Matrix Precipitate Matrix Precipitate Ni Bal 71.0 30.0 71.8 59.5 Mo 15.95 11.5 49 . 4 11.5 20.4 Cr 6.87 6.8 4.3 6.8 6.8 F'e 3.84 4.0 0.7 3.7 245D Sl 0.62 0.3 24 0.3 1.0 Mn 0.50 0.6 0.01-0.10 0.6 0.6 Mo _ Mo _ T 1.65 T 0.34 Mo\ _ Mo\ _ Mo\ _ (=) com (=) -om (=) -1 Fig. 4.17. Normal Structure to Structure with Lamellar Phase. T 0.007 INCHES 1> 500X [on [ Photomicrograph of INOR-8, Showing Transition from 102 l Tro 0.005 INCHE S 750X T [ Fig. 4.18. Autoradiograph of INOR-8, Showing Carbon Depletion in Lamellar Phase. Postirradiation Stress-Rupture Properties of INOR-8 The bulk of the data available on the effects of irradiation on INOR-8 are postirradiation tensile data.lo In order to better assess the lifetime of reactor components that were designed on the basis of un- irradiated creep data, a program was started to ascertain the in-reactor creep behavior and the postirradiation creep behavior of INOR-8. The dose dependence of the ductility and rupture life of two heats of INOR-8 (Ni- 5065 and 2477) are being determined at several stress levels. Heat Ni-5065 is a reactor vessel heat, and heat 2477 is a low-boron heat. These uni- axial creep tests at 650°C show the influence of irradiation on creep rate, time to rupture, and ductility. The postirradiation stress-rupture properties of heat Ni-5065 and heat 2477 at 650°C are given in Tables 4.6 and 4.7 for various thermal- neutron doses. These data show that ductilities for heat Ni-5065 at the higher dose levels are less than 1% and for the low-boron heat are ~5%. With both heats the time to rupture for a given stress level decreased with increasing dose. An attempt is being made to correlate these data To predict the rupture time for stresses below 10,000 psi and to determine 103 a correlation between in-reactor and postirradiation tests. With these correlations one would be able to predict allowable stress limits and rupture lives for reactor structural application. Table 4.6. Stress-Rupture Times and Ductilities® for INOR-8 (Teat Ni-5065) Tested at 650°C ot Times (hr) and Ductilities (%) for Specified (pgifis Thermal-Neutron Dose (nvt) 0 5% 106 1.3 x10% 9 x 108 5 x 102 5 x 102° X 107 52,57 52 21.2 11.9 5.1 0.43 0.9 (10) (2.6) (2.0) (3.4) (0.85) (0.64) 39,80 ~270 424 .7 147.0 80.2 23.5 5.3 (9.2) (2.8) (2.4) (0.2) (0.3) 32.35 ~800 686.7 647 .0 232 5.2 8.0 (6.7) (6.3) (2.5) (1.8) (0.42) 21.5% ~2500 Test in 1603 521.8 312 progress (2.7) (1.5) (0.79) a C e . Ductilities are in parentheses. Table 4.7. Stress-Rupture Times and Ductilities”™ of Irradiated INOR-8 (Heat 2477) Tested at 650°C Times (hr) and Ductilities (%) for Specified otress Thermal-Neutron Dose (nvt) (psi) 5 x 10%° 1.3 X 10%8 9 x 10%8 5 x 10%° 5 x 1020 X 103 52.6 70.6 57 .4 30.4 21.1 19.9 (2.8) (3.0) (4.6) (4.3) (4.7) 32 .4 1070 1525.7 1425 849.6 515 (37) (6.3) (5.6) (5.1) (3.4) a'Ductill_:'Lt:'Les are in parentheses. 104 In-Reactor Creep Tests ceveral tests have been run in which the specimens were irradiated and stressed simultaneously. The testing technique was described previ- ously.!! The data obtained on heat Ni-5065 at 650°C and a nominal thermal flux of 6 X 10'° nv are compared with unirradiated data in Fig. 4.19. A similar comparison is made in Fig. 4.20 for another experiment on heat Ni- 5085. In both experiments the specimen at the lower stress did not fail, and its behavior suggests that some experimental problem may exist. The reduction in rupture life is greater for heat 5085 than for heat 5065. This behavior seems consistent with the fact that the boron content of heat 5085 is about double that of heat 5065. ORNL-DWG 65-12050 70 6O\\\\\\\\ N ~\\ \ 50 \UNIRRADIATED 2 40 NG O \ Q N O N = \\ n \\ wn g 30 ® \ n ° \\\ 20 ? o—» \\ N N ® IN-REACTOR DATA FOR SPECIMENS EXPOSED TO 650°C AND TO A THERMAL FLUX OF 6X41013 nv {0 0 10 20 50 100 200 500 1000 2000 5000 10,000 RUPTURE LIFE (hr) Fig. 4.19. Influence of Irradiation on Creep-Rupture Properties of INOR-8 (Heat 5065). 105 ORNL-DWG 65-12051 60 50 \\ N N \\ \\ \\ 40 \\ n —~ UNIRRADIATED 2 \\k 3 o ™ = 30 S o \ & o = N wn \\ 20 ° ™ o— N 10 | ® SPECIMENS EXPOSED TO THERMAL FLUX OF 6X10'3 nv AND 650°C IN REACTOR 0 10 20 50 100 200 500 1000 2000 5000 10,000 RUPTURE LIFE (hr) Fig. 4.20. Influence of Irradiation on Creep-Rupture Properties of INOR-8 (Heat 5085). References G. M. Adamson, Jr., et al., Interim Report on Corrosion by Zirconium Base Fluorides, ORNL-2338 (Jan. 3, 1961). A. J. Romano, C. J. Klamart, and D. H. Gurmski, The Investigation of Container Materials for Bi and Pb Alloys, Part 1, BNL-811 (T-313) (July 1963). MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 83. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 84—86. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 84—86. J. W. Helm, Radiation Effects in Graphite at High Temperature, Bat- telle Northwest, BNSA-67. A. J. Perks and J. H. W. Simmons, 'Dimensional Changes and Irradia- tion Creep of Graphite at Very High Doses," paper No. 188, Seventh Biannual Conference on Carbon, June 21-25, 1965, held at Case In- stitute of Technology, Cleveland, Ohio. C. R. Kennedy, "Irradiation Creep of Graphite," Metals and Ceramics Ann. Progr. Rept. June 30, 1965, ORNL-3870. J. T. Venard, Tensile and Creep Properties of INOR-8& for the MSRE, ORNL-TM-1017 (February 1965). 10. W. R. Martin and J. R. Weir, Nucl. Appl. 1, 160 (April 1965). 11. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 82-83. 106 5. RADIATION CHEMISTRY Introduction A number of irradiation experiments with molten-salt reactor fuel and construction materials have been carried out in irradiation facili- ties in the Materials Testlng Reactor (MTR). Results from these experi- ments have been reported and indicate that operation of the MSRE should not be adversely affected by irradiation effects on fuel salt and ma- terials. However, a continuing irradiation program is needed to pro- vide supporting information for an understanding of both short-term and long-term effects of irradiation and fissioning on fuels and materials, especially for reactors that will follow the MSRE. Design and development of an in-pile molten-salt experiment have been in progress during the past year. A description of the proposed in-pile experlment and preliminary results of mockup tests have been reported else- where?® but are briefly summarized here along with progress made through this report period. Experimental ObJjectives and Design Considerations oome objectives of the proposed irradiation (in—pile) program with an MSR fuel salt are tabulated below: Irradiation objectives: ZOO'w/me fuel fission power Maximum fission product production One year of in-pile operation Information objectives: Graphite—INOR-8-salt compatibility Salt (fuel) stability Fission product chemistry To meet these objectives, the following experimental requirements were established: oalt circulation through hot and cold regions oalt sampling during irradiation Cover gas sampling Ability to drain fuel salt valt addition to replace that removed in sampling Keeping fuel molten at all times The above objectives and requirements indicate the need for daily attention to the experiment operation and data, rapid analysis of gas and fuel samples, and possible concessions in reactor operating sched- ules. These requirements and the availability of hot-cell and analytical facilities especially equipped to handle radicactive fuel dictate that such experiments be carried out in the Oak Ridge Research Reactor (ORR). 107 Therefore, the experimental equipment is designed to make use of horizon- tal beam hole HN-1 in the ORR. Previous neutron flux mapping of beam hole HN-1 indicates that a thermal flux of ~5 X 10%*3 would be avallable. It is estimated that a thermal flux of 5 X 10%3 will produce a fission power density of l9O'w/cm3 in a fuel salt containing 0.6 mole % 235U and a den- sity of 320 w/em® in a 1 mole % 23°U salt. Experiment Design and Mockup Test Operation In order to avoid costly, complex pump loop experiments, an auto- clave (capsule type) experiment with thermally induced flow was designed. Initial tests with prototype autoclaves showed that flow rates of 2—10 cmB/min could be achieved in an autoclave of this type. These flow rates are considered adequate to study transport and deposition of fission and corrosion products and to provide mixing for fuel sampling and enrichment. Figure 5.1 is a schematic diagram of the autoclave assembly, and a photo- graph of the partially assembled experiment is shown in Fig. 5.2. ORNL-DWG 65-12052 SALT RESERVOIR PRESSURE MONITOR LINE SALT FLOW PASSAGES GRAPHITE CORE THERMOCOUPLE WELL S SAMPLE LINE SALT LEVEL THERMOCOUPLE . harlier models of the autoclave experiment had a horizontal line connecting the "cold leg" with the main autoclave. However, the design incorporating this horizontal return line was changed to the triangular configuration shown in Fig. 5.1, after test operation indicated flow block- age by deposition of bubbles of cover gas in the return line. Gas depo- sition and flow blockage were observed after ~16 hr of operation with helium cover gas and ~135 hr with argon cover gas. Flow invariably re- covered after evacuation and repressurization of the capsule, but con- tinuous salt flow in the experiment 1s desirable, so the horizontal return line was eliminated to facilitate gravity release of the bubbles. The bubble formation appears to be consistent with a mechanism based on mass transport due to the temperature dependence of gas solubllity, but ap- parently is under diffusion control in the return line. Successful elimi- nation of the loss of flow by removal of the horizontal return line was evidenced by the fact that the redesigned test model (Fig. 5.1) has op- erated for 550 hr under helium cover gas with a continuous flow rate of 913 cm®/min. Calrod heaters and cooling coils, in which air or air-water mixtures are used as coolant, provide temperature control and maintain the thermal gradients necessary to induce flow. Heaters and cooling coils are em- bedded in a graphite Jjacket around the autoclave body. Tubes of capillary dimensions connect the vapor space of the reservoir tank with remotely located pressure monitoring and gas sampling equipment. The salt sample line (1/8-in. OD X 0.075-in. ID) is ~12 ft long and is traced with Calrod heaters. The sample line is connected to a sample station at the reactor shield face, where samples of molten salt can be routed either to a dump tank for storage or a sample capsule for subsequent chemical analysis. The sample station also contains necessary tanks, valves, and lines so that salt can be added to the autoclave experiment during in-pile opera- tion. To date some 4000 hr of operation has been accumulated with several prototype models of the in-pile molten-salt experiment. Test operation has been conducted in a mockup facility which duplicates beam hole HN-1 of the ORR. Test operation is at a nominal salt temperature of ~1200°F with a salt mixture whose composition is LiF-BeF,-ZrF,-UF, (65-29-5-1 mole %, liquidus temperature ~840°F). Salt circulation rates of ~L10 cm /min are obtained by maintaining a median temperature gradient of 50-100°C between the main autoclave and the '"cold leg." The flow rate is monitored by heat balance measurements around the "cold leg." The adequacy of the salt sampling and addition system was demonstrated by removing 15 samples (10 cm® each) of the salt while operating at high 110 temperature. After each sample removal, an equivalent amount of salt was added in order to maintain the original salt inventory. A test was made of the ability of the autoclave heater-cooler unit to remove the predicted 12 kw of fission and gamma heat (8 kw of fission heat) during in-pile operation. Some 9 kw of heat, the maximum that could be generated in the test equipment, was removed with the autoclave operating at 1000°F; the heat removal capacity is accordingly considered adequate for in-pile operation, taking into account the higher operating temperature of 1200°F. Also, the cooler is capable of a larger through- put of air-water coolant than was used in the above tests. Results of mockup tests indicate that the presently designed auto- clave is suitable for use in a fuel salt irradiation program with the obJjectives previously outlined. Additional design and development work to provide necessary facilities in beam hole HN-1 of the ORR is under way. Included in these facilities are a beam hole liner, a beam hole shield plug, salt and gas sampling equipment, revisions to an existing instrument panel, and shielded carriers needed to remove the experiment after in-plle operation. References 1. Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp. 36—45 (April 1965). 2. Ibid., pp. 45-48. 111 6., CHEMISTRY Chemistry of the MSRE MSRE Fuel Loading The initial fuel loading of the MSRE required approximately 75 ft> of fused fluorides having the composition 7LiF-BeF2-ZrF4-UF4 (65.0-29.1- 5.0-0,9 mole %); fissionable <3°U comprised about one-third of the uranium inventory. To provide for an orderly approach to critical operation of the reactor and to facilitate fuel preparation, the fuel was produced as Three component mixtures. The enriched fuel concentrate mixture, in which all ?3°U was combined with ‘LiF as UF, 93% enriched in 23°U to form the binary eutectic mixture (27 mole % UF,), was prepared in six small batches (15 kg 2357 each) for nuclear safety and for planned incremental additions to the reactor fuel system. The balance of the uranium required for the fuel was provided as a chemically identical mixture with UF, depleted of 235U, The third component mixture, the barren fuel solvent, consisted of the remaining constituents of the reactor fuel and had the chemical composition of ‘LiF-BeFy-ZrF, (64.7-30.1-5.2 mole %). The preparation of these mixtures from simple fluoride salts has been previously re- ported.*t Reactor fueling operations began on April 20, 1965, with the loading of the barren fuel solvent and the depleted fuel concentrate mixture. Approximately 10,050 1b of solvent from 35 batch containers and 520 1b of depleted fuel concentrate from two batch containers were added directly into the fuel drain tank. Batch containers were heated above the liquidus temperature of the salt mixture in auxiliary furnaces and connected by a small-diameter Inconel tube to the drain tank. The connecting tube ex- tended to the bottom of the batch container so that the salt mixture could be transferred as a liquild by controlling the differential gas pressure in the two containers. All fill operations were accomplished in a routine manner without causing detectable beryllium contamination to the reactor facilities., The addition of the enriched fuel concentrate mixture to the MSRE to within 1 kg #2°U of criticality was accomplished during the latter part of May 1965. This operation was coordinated in accordance with planned zero-power experiments of the reactor system.2 The first major addition of enriched fuel concentrate consisted of the transfer of about 44.17 kg 232U from three containers directly into the fuel drain tank. Three sub- sequent additions of 235U to the reactor drain tank increased its 23°U inventory to 59.35, 64.42, and finally 68.76 kg. The transfer of less than batch-size quantities of 437U was made by inserting the salt transfer line to a predetermined depth in the batch container. All fluoride mixtures — coolant, flush, and fuel — required for the initial operation of the MSRE were prepared and loaded into the reactor facility by the Reactor Chemistry Division. Enough fuel-enriching cap- 112 sules, each of which contained about 85 g 227U each, to bring the MSRE to its critical operation and to maintain nuclear operation of the reactor during its scheduled tests were also provided. MSRE Salt Chemistry During Precritical and Zero-Power Experiments Chemical analyses were conducted during the MSRE precritical and zero-power experiments for the purpose of establishing analytical base lines for use in the full-power operating period. Results of these tests provided not only the desired base lines, they also indicated that, with respect to preventing contamination of the fuel salt, current reactor op- erating procedures have afforded excellent protection of the fuel salt. Fuel samples were removed daily from the MSRE pump bowl throughout the zero-power experiments. Chemical composition, contaminant levels, and isotoplic analyses were determined on a regular basis. The salient con- clusion derived from the results of these analyses i1s that no corrosion was sustained by the fuel containment circuitry during the approximately 1100-hr precritical and zero-power test period. At the end of the zero- power experiments, the fuel salt was drained from the fuel circuit and replaced briefly by the flush salt. The amount of uranium subsequently detected in the flush salt (0.0195 wt %) is sufficiently small to indicate that no appreciable holdup of fuel salt occurred in the drain operation. The MSRE fuel was constituted within the reactor in two steps. The first was addition of "LiF-?28UF, to the "LiF-BeF,-7ZrF, (64.75-30.09-5.16 mole %) carrier salt. The mixture produced in this manner was circulated through the fuel circuit for some 250 hr during the PC-2 precritical test. The reactor fuel for the zero-power experiments was produced subsequently by adding small increments of LiF-235UF4 into this carrier salt in the drain tanks and finally into the pump bowl. The final composition of the salt was controlled by the amount of 235UF4 required for criticality to be sustained with one control rod completely inserted. It was anticipated that the composition of the fuel at this point would be 7LiF—BeF2—ZrF4-UF4 (65.0-29.17-5.0-0.83 mole %); the composition calculated from the weights of the carrier and enriching salts added to the reactor was 7LiF-BeF2- ZrF4-238UF4—235UF4 (64.85-29.28-5.04-0.55-0,28 mole %). The composition of the fuel salt was changed steadily throughout the zero-power experi- ments as capsules of the enriching salt were added to the pump bowl. Com- positional analysis during this period served, therefore, to permit eval- uation of the fuel composition dynamically rather than as a statistical base for reference during the power run. Final composition established analytically on the basis of the last four samples was 7LiF-BeF2-ZrF4— 238yF,-23°UF, (62.31-31.68-5.19-0.54-0.28 mole %). Complete results of all analyses performed with MSRE salts during the precritical and zero-power experiments were reported earlier.? The data are summarized in Table 6.1. Uranium Assay. During the initial fill operations, ‘LiF-BeF, flush salt was admitted to all parts of the fuel system. On draining the re- actor, flush salt remained in the freeze valves as well as residue in the drain tanks. At this point, some 80 1b of salt was unaccounted for by *quawtaadxe T-0d JO PpUS 4V *quawTIadXs T=Dd JO SUTUUTISQ 4V " (%'3eP qBT-30U) quel "H D *(®B3EP QBT-PTOO) UUSNeBA *Jd "M * SUSWTOads JNOJ UO Pasey BABP fAQUSUTISIXD JoMOd-0J9Z JO PUS 1V 113 0L6 ¥ GOTT GLOT oCE7 O 6L ¥ 8% e ¢c LT ol i TN 8 F LE LE 61l cc oG 13 )O/, Gk \io \Q\<_\ 6\\@ N \.\" ST T L l ) ] 1 B . 0.5 N S na - i . \\\."* | @=44.7 kg mole < " Fig. 6.7. Removal of Todide L 02 e ' S N by HF Sparging. S .\\ 8T Of | — S NG e e ] < N c 005 — [1719: ~0.005 mole kg=! |~ 8'~ * Hy, FLOW: ~90 ml min~" A T < — Pyg: ~0.047 atm - ° MELT: 275 g 2LiF —BeF, ~ TEMPERATURE: 487°C 0.02 +— o 0.04 1 0 5 10 15 20 (x1073) MOLES OF HF PASSED iodide added to the salt initially. Runs with 1307 revealed that this poor material balance definitely was not caused by retention of iodide in the reaction vessel or on the walls of the off-gas line. Rather, these tests indicated that a fraction of the HI passing through the aqueous scrubbers was adsorbed on particulate matter in the gas stream and in this form was not caught by the NaOH solution. While this effect will be investigated further, it is felt that it will not appreciably alter the present estimates of Q (Figs. 6.8 and 6.9). For 2LiF-Befs, the variation of @ with temperature is given by the fol- lowing expressions: log @ = —1.094 + 2.079(10%/1) . (7) The initial results on the effect of the BeF, mole fraction on Q indicate a maximum in Q somewhere in the range XBng = 0.33 to 0.50, From Eq. (6), the number of moles of HF which must be passed per kilogram of melt in order to remove half the iodide present in the melt is given simply by 0.693/Q. For example, at 500°C, Q = 40 kg/mole, and hence, half the iodide present will be removed by passage of 0.0173 mole of HF (388 std cm3) per kilogram of melt. In order to achieve a removal half-time of 1 hr, the HF flow rate thus must be 0.0173 mole/hr per kilo- gram of melt. In a reactor, 1357 could be removed by passing continuously a frac- tion F of Tthe fuel per hour through a gas-liquid contactor. If This con- tactor removed a fraction f of the iodide present in the side stream, then the half-time in hours for removal of iodide from the entire system would be ty/2 = 0.693/F(f) . (8) 131 Fig. 6.8. Variation with Tem- perature of Equilibrium Quotient for ORNL—-DWG 65-—9874 TEMPERATURE (°C) 650 550 600 500 1.314 1.35 T [ lodide Removal from Z2LiF-BeFs,. 2 AH=-95 keal < S 1.07 .1 145 119 1.23 .27 1000/ 7 (CK) ORNL-DWG 65-12056 80 /. o 70 LN v N\ s N\ d s \\ s s AN 60 ~ . s — s N\ T v N g _~ \ o o N\ @ 50 N o £ \ S \\\ \ 40 N\ \\ ® S 30 N, 20 0.35 0.40 0.45 0.50 MOLE FRACTION BeF, Fig. 6.9. Variation with LiF-BeFy Melt Composition of Equilibrium Quotient for Iodide Removal at 482°C. 132 If a countercurrent gas-liquid contactor equivalent to n ideal stages were usg%, the fraction of lodide extracted from the side stream would be given by 0.9) 1.1 7 1.2 5 1.5 Z, 2 3 Hence, the removal half-time is determined primarily by the side-stream flow rate F. Under these conditions the molegs of HF per hour required to achieve the desired removal half-time is given by moles HF 0,693 an e (l*>wT, hr T1/2 WS where Wp 1s the total weight of the fuel in kilograms. Since the product (QHF/WS§Q must be greater than unity, the minimum amount of HF required would be moles HF 0.693 W T hr T > hr tl/g In future tests attempts will be made to improve the efficiency of HI trapping. Preliminary results of tests with 1307 indicate that the trapping efficliency is greatly increased by prefiltering the gas stream, In addition, the effect of melt composition on Q will be investigated further. 133 calt Compositions for Use in Advanced Reactor Systems Blanket Salt Mixtures for Molten-Salt Breeder Reactors The blanket salt moest recently proposed for use in molten-salt breeder reactors:hsen17LiF-BeF2—ThF4 mixture (71-2-27 mole %) whose liquidus tem- perature is 1040°F (560°C). Other compositions of ‘LiF-BeF,-ThF, may be attractive as blanket salts if the advantages of lower liquidus tempera- tures are not offset by the attendant reduction of thorium concentration. (The concentrations of BeF, are not sufficient to affect the viscosity adversely.) Mixtures of 7LiF-BeF2—ThF4 whose compositions lie along the even-reaction boundary curve L = 3LiF«ThF,(ss) + 7LiF-6ThF, appear to qualifly as the best blanket salts from phase-behavior considerations. The ThF, concentration of these mixtures varies from 6.5 to 29 mole % as LiF concentrations change from 63 to 71 mole %. Liquidus temperatures for this range of compositions vary from 448°C (838°F) to 568°C (1044°F). In the molten state these mixtures contain thorium concentrations ranging from 850 to 2868 g/liter at approximately 600°C. These data are summa- rized in Figs. 6.10 and 6.11 and in Table 6.6. Optimization of thorium ORNL-LR-DWG 37420AR2 Theg, 1444 0 MO TEMPERATURE IN °C COMPOQSITION IN mole % 1050 LiF-4ThE, LiF-2ThF, 1000 LiF-2ThE; 950 P 897 2L|F-BeF2 900 7LiF-6ThE, P62 , 850 P 597 £ 568 \ 3LIF-ThE 4., 750 E 565 65 6. 20 06 %?%bQ’C)Qb 20 S 0 00 S50 £ 526 o0/ \L 3 LiF BeF, 845 2LiF- BeF; 500T450 400 400 450 500 548 P 465 £ 370 Fig. 6.10. Phase Diagram of the System LiF-BeF,-ThF,. ORNL-DWG 65—-12057 (°F) (°C) 1100 593 3400 1050 566 < 3000 LIQUIDUS TEMPERATURE‘/////, /////////// o o & 1000 538 , 2600 3 D / O 5 - o < LJ -+ ¢ a Q 2 / 2 W 950 510 2200 ~ / o //////////// Z N O T - 900 482 , 1800 = /// o O = O O / THORIUM (g/li‘rer AT ~600°C) s 850 454 ——1 1400 S / / = O I // ] // 1000 600 5 10 15 20 25 30 ThF, (mole %) Fig. 6.11. Thorium Content of LiF-BeF,-ThF', Compositions on the Even-Reaction Boundary Curve L = 3LiF-ThF, (ss) + 7LiF .6ThE, . Table 6.6. MSBR Blanket Compositions Thorium Liquidus Temperature Composition Concentration (g/liter) °C °F ThF, LiF BeF, 863 450 842 % 63 30 1207 469 876 10 66 24 1828 500 932 16 69 15 2419 550 1022 22.5 71 6.5 2868 568 1054 29 71 135 concentration and liquidus temperatures may be made with the aid of Fig. 6.11; if 2000 g/liter of thorium is adequate to achieve good breeding gain, the composition containing 17.5 mole % ThF, may be used, and the liquidus temperature of this mixture is 516°C (960°F). Coolants for the Molten-Salt Breeder Reactor The steam circuit of the Bull Run Plant of the TVA, the type of plant to which a proposed 1000-Mw(electrical) MSBR would be coupled, operates with temperature extremes of 700 and 1125°F (371 and 607°C). An essential requirement for the development of the reactor is the availability of a cheap, chemically stable secondary-coolant fluid which in the liquid state has acceptable nuclear, chemical, and physical properties. No salt mix- ture has as yet been discovered which meets all the criteria imposed. Preliminary examination of the NaF-NaBF, system, which according to Russian reports®®’?® forms a eutectic mixture melting at 304°C, disclosed that the eutectic formed from the pure fluorides actually melts at 373 to 375°C, but that the oxide contaminant B,0; markedly depresses the liquidus temperatures. Although high solubility of oxide ion in the coolant is chemically advantageous, By05 exhibits a strong tendency to form glasses in the molten state, an effect which has already been noted in the vis- cosity of NaP-NaBF,-B,O; liquids we have examined in the laboratory. Little information is available from the literature concerning the prop- erties of these and other liquid mixtures of fluorides and fluoborates. Evaluation of thelr potential as secondary coolants for the MSBR will accordingly require considerable additional investigation. In another set of experiments, phase transition temperatures of two samples, one consisting of pure NaBF,, the second of a mixture of 39.1 mole % NaBF, and 60.9 mole % NaF, were determined from an examination of cooling curves. The samples were encapsulated under vacuum in INOR-S8. Temperatures were read with an NBS-calibrated Pt-Rh thermocouple in an INOR-8 thermowell which extended about 1 in. into the melt. Cooling curves for the NaBF, sample were obtained at cooling rates of from 1.0°C/min down to 0.2°C/min. Every run exhibited supercooling, ranging from 5 to 15°C, which made it difficult to obtain a very precise melting temperature. FIFrom these curves, the best estimate for the melting point of NaBF, is 396 * 2°C. Cooling curves for the 39.1 mole % NaBF;~60.9 mole % NaF mixture were obtained at cooling rates from 0.3°C/min to 1.4°C/min. Relatively little supercooling was encountered at the only phase transition temperature, 375 £ 1°C, observed for this mixture. The sample was cycled through the temperature interval 255 to 465°C. Again, the 304°C eutectic reported in the Russian literature?? was not observed. The capsule containing the NaBF,-NaF mixture was cut open after com- pleting the cooling curves and was examined; there was no evidence of corrosive attack. 136 Viscosity of NaBF, In a glove box containing very low concentrations of oxygen and mois- ture, the viscosity of NaBF, was measured with a Brookfield viscosimeter and found to be 7 * 2 centipoises at 466°C and 14 * 3 centipoises at 436°C. The precision is poor because the apparatus was primarily designed to measure high viscosities. No mixtures of NaF and NaBF, were measured, but such melts would most likely have been less viscous than pure NaBF,; the poor precision occur- ring at Tthese lower viscosities discouraged us from attempting further measurements. ORNL-DWG 64-1993 ThE, 114 PRIMARY PHASE AREAS: _ LiF TEMPERATURES IN °C NaF i} COMPOSITIONS IN mole % TNaF - 2ThF, . +H-+++++ INDICATES SOLID SOLUTION 4NaF - ThF, 2NaF - ThE, 1050 2LiF-NaF - 2ThF, ~NaF - ThE, LiF-4ThFg—g—- — SOLID SOLUTION 3LIF- ThF, 1000 7LiF- 6ThE, LiF-2Thg N\ W /\ LiF- 4ThF, LiF- 2ThF,~ 950 —NaF - 2ThF, NaF - 2ThF, 3NaF-2ThF, A 900 A ThE P-897—/— [~ P-83 4 SGlIGICISIOMUIWICIGICIS) 7LiF- 6ThFg AN 3NaF - 2ThE, 712 "TN—£-690 \~7NaF - 2ThF, N>P-645 4NaF - Th, 8507 o, //Kjf/// 950 7 NaF 845 E-652 995 Fig. 6.12. The System LiF-NaF-ThF,. 137 Fuel and Blanket Materials for the Proposed MOSEL Reactor A proposal for investigation of the feasibility for developing an intermediate-energy, molten-fluoride-salt reactor with the fuel salt cooled by direct contact with liquid lead was described recently.28 In preliminary discussions, Kernforschungsanlage Julich and ORNI staff mem- bers have arrived at initial choices for fuel and blanket compositions. Inspection of the ternary phase diagrams containing thorium fluoride and mixtures of lithium fluoride, sodium fluoride, beryllium fluoride, or lead fluoride shows that at concentrations above 10 mole % only the LiF-NaF- ThF, system (Fig. 6.12) affords mixtures in which the liquidus is as low as 500°C. Accordingly, the initial choice of the fuel composition is approximately LiF-NaF-ThF,-UF, (44-32-20-4 mole %). The blanket-salt composition would be LiF-NaF-ThF, (44-32-24 mole %); alternatively, the composition LiF-ThF, (73-27 mole %) may be used. Composition of the salt mixtures used ultimately in the experimental program will be based on the results of further calculations to be performed in the next several months. Recovery of Protactinium from Fluoride Breeder Blanket Mixtures Introduction Demonstration of a process for extracting protactinium from a molten- fluoride breeder blanket 1s an important part of studies directed toward establishing the feasibility of molten-fluoride thermal breeder reactors. An earlier investigation29 showed that protactinium could be removed by oxide precipitation from one molten mixture, LiF-BeF,-ThF, (67-18-15 mole %), that is suitable for use as a breeder blanket. Other mixtures are currently of interest, and it is also desirable to develop other, more efficient, removal processes. The glove box used in the earlier investigation, when 231pg ywas present in the fluoride mixture in addition to ?33Pa, is no longer available. The use of ?31Pa permits attainment of expected operating concentration levels (50 to 100 ppm) without the ex- cessive gamma activity associated with milligram quantities of 233pg, This report contains a brief description of a new glove box facility suit- able for laboratory-scale studies of methods of recovering protactinium from fluoride mixtures and results of a preliminary experiment on precip- i1tation of protactinium at the tracer level, Facility Description The High-Alpha Molten-Salt (HAMS) ILeboratory is located in room 127, Building 4501, an area formerly occupied by a hot cell. A view of the laboratory, Fig. 6.13, shows the seven interconnected glove boxes that are presently installed in the laboratory. The glove box at the right end of the train is connected to a hood which exhausts into the hot cell off- gas system. Air is pulled into the glove boxes through a rectangular- shaped filter on the left side of the top of each box and exhausts through 138 Fig. 6.13. View of High-Alpha Molten-Salt Laboratory. a pair of 8 X 8 in. absolute filters mounted in the back of the box and through another enclosed absolute filter before reaching the glass-fiber- reinforced header pipe which discharges into the plant hot off-gas system. The pressure in the header line is controlled automatically to —0.5 in. Ho0 in the glove boxes, and the loss of a glove from two boxes in the glove train results in a negligible increase in pressure in the other boxes. The room can be maintained at a pressure of —0.5 in. H20 with respect to the adjoining areas by removal of air through the glove boxes and hood. The large filters on the wall back of the glove boxes form the alr intake of a recirculating air-conditioning system. Most of the molten-salt work will be performed in the stainless steel box shown in Fig. 6.14. The manifold and gages mounted on the back wall of the box control application of vacuum or admission of helium, hydrogen, and anhydrous HEF to a flanged nickel pot in a well below the floor level of the glove box.20 The well, made of 4-in.-ID stainless steel pipe, is heated by a 5-in. tube furnace supported by a jack below the box. The Pyrovane controller at the top of the panel to the right of the box con- trols the furnace temperature indicated by a thermocouple adjacent to the heating coil, while the Brown recorder shows the temperature at the junc- tion of an Inconel-sheathed thermocouple immersed in molten salt about 1/2 in., from the bottom of the nickel liner inside the pot. Other boxes in the train contain an analytical balance (Mettler, type H-l5), a Zeiss polarizing microscope, and a quenching furnace for supporting research. 139 PHOTO 80772 Fig. 6.14. Stainless Steel Glove Box for High-Temperature Studies with Molten Salts Containing 231Pa. 140 Oxide Precipitation of Protactinium The precipitation of protactinium from LiF-BeF,-ThF, (73-2-25 mole %) was studied at tracer level (~1 ppb) using “32Pa in purified salt to test the equipment and operating procedure. Filtered samples of fused salt were removed from the melt at 620 or 630°C by inserting a l-l/Z—in. length of 3/8-in. copper tubing with a sintered-copper filter welded in one end and 18 in. of l/8—in. nickel tubing brazed to the other end. Weighed amounts of ThO, were added to the molten mixture through the same fitting that permitted introduction of the filter stick, and mixing was accomplished by bubbling helium through the melt for 1/2 to 1 hr. After precipitation of protactinium was complete, the mixture was allowed to cool to room temperature under an atmosphere of helium. It was later remelted and treated with a mixture of H, and HF (10 to 1 by volume) to demonstrate that the precipitated protactinium could be returned to so- lution. The results of this experiment confirmed the conclusion of the earlier study that protactinium can be readily precipitated from a molten fldoride mixture by addition of thorium oxide and that the precipitate can be returned to solution by treatment with HF. It also appeared that precipitation occurred due to inadvertent addition of water. Quantitative conclusions are not Jjustified because insufficient time was allowed in the early stages of the experiment for equilibrium to be obtained, and because a large part of the mixture was found to have solidified on the liner about 2 in. above the surface of the melt. A large vertical gra- dient in the furnace well is necessitated by the requirement of a cool glove box floor. The operating temperature was only 60 to 70° above the freezing point of tTthe mixture GV560°C), and gas bubbling through the melt apparently carried portions of it to a point on the liner that was cool enough to permit freezing to occur. Operating experience obtained in this experiment indicated that no special problems should be encountered with similar experiments in the glove boxes with higher-specific-activity materials such as %31pa., A speclal furnace and other equipment modifications to minimize the thermal gradient problem are under consideration. Development and Iivaluation of Methods for the Analysis of MSRE Fuel Determination of Oxide in MSRE Fuel On the basis of continued study of methods of oxide determination in MSRE salts, the hydrofluorination method was selected as most adapt- able to immediate hot-cell operation. While the method is less sensgitive than the inert—gas-fusion3l or KBrF432 methods, the limits of oxide de- tection can be lowered by increasing the sample size. 141 d33 The hydrofluorination metho is based on the reaction 0%~ + 20F(g) — Hy0(g) + 27 which occurs when a molten salt sample 1s purged with an Ho-HEF gas mixture. Either the water evolved or the HF consumed can serve as a measure of oxide; however, the water offers the more facile measure- ment . Preliminary tests have shown that the oxide is readily removed from LisBeF, salts at 500°C with 0.02 atm of HF in the purge gas. Adding zirconium to the melt causes a shift in equilibrium of the re- action, resulting in an incomplete recovery of the oxide at these re- action conditions. By increasing the melt temperature to 700°C and enriching the HF in the hydrofluorinating gas to 0.15 atm, the oxide is removed from a 100-g melt in about 4 hr. The application of this method to the analysis of radioactive samples requires the development of (1) a sampling technique which minimizes atmospheric contamination, (2) the incorporation of a water- measurement technique which is convenient for hot-cell operations, and (3) the fabrication of a compact apparatus to conserve space in the hot cells. It was necessary to adapt the sampling technigques from methods already developed to obtain samples for wet analyses. By using a copper enricher ladle, a 50-g sample which can be transported in the existing transport container is obtained. Atmospheric expo- sures will be minimized by remelting and hydrofluorinating the entire sample 1in the sampling ladle. The ladle will be sealed in & nickel- Monel hydrofluorinator, with a dellvery tube spring loaded against the surface of the salt. Prior to melting, the apparatus will be purged with hydrofluorinating gas mixture to remove water on the inner surface of the hydrofluorinator and, hopefully, any contamination on the exposed surface of the salt. When the sample melts, the delivery tube will be driven by spring action to the bottom of the ladle for efficient purging of the melt. In all preliminary tests, the water in the effluent gas has been measured by Karl Fischer titration. While the Karl Fischer reagent has been shown to be remarkably stable to radiation, the titration would be difficult to perform in the hot cell. An electrolytic moisture moni- tor is ideally sulted to remote measurements but is subject to interference and damage from HF. A sodium fluoride column operated at about 90°C removes HF from the effluent gas without significant holdup of water. A schematic flow diagram of the apparatus for hot-cell installa- tion is shown in Fig. 6.15. The apparatus has been designed and is now being fabricated during component testing. A modular design has been selected to facilitate any changes found necessary during com- ponent testing and to permit necessary repairs in the hot cell. Except for the hydrofluorination furnace, all hot-cell components are contained within a 16-in.-square compartment. 142 ORNL-DWG 65—8855A TO SODA LIME TRAP | MOISTURE / MONITOR | 5| Pb SHIELD SPLITTER NN % P I @D 1 HF | LHe BACK FLUSH __——NaF —HF TRAP PURIFIER \ CELL WALL N\ HYDROFLUORINATOR Fig. ©.15. ©Schematic Flow Diagram of the Hydrofluorination Ap- paratus for the Determination of Oxide. Improved results have been obtained with a components test fa- cility. The apparatus, which has nickel gas lines maintained at a temperature above 100°C, is designed to minimize void space and dead legs and to improve melt purging efficiency. While an apparatus for filling ladles and evaluating the effects of atmospheric contamination was being assembled, test analyses were run on standard additions of ZrOz and U0y to a 50-g fuel melt. In each analysis the purge gas flared at 150 cm”/min. When the salt temperature was held at 700°C with the HF' concentration =0.1 atm, the oxide was evolved in 1 hr. Recovery data obtained by the Karl Fischer titration are shown below. Sample Oxide Added. (ppm) Oxide Recovered P Based on 50-g Sample (ppm) UO2 425 455 355 345 415 390 ZTOQ 405 435 520 490 805 790 425 435 415 420 325 320 143 oeparate tests of a Beckman electrolytic moisture monitor have yielded 98 * 2% recovery of injected water. These tests also indi- cated that the cell efficiency is easily decreased by overloading; therefore, the capillary splitter shown in Fig. 6.15 is required. A moisture monitor and splitter have been installed on the component test facility, and recoveries of water from ZrO; addition are being measured. Electrochemical Analysis Work is continuing toward the goal of adapting electroanalytical methods to the in-line analysis of impurities (e.g., corrosion products) and other electroactive species in the molten MSRE fuel.3%: 3% A new phenomenon, which suggests a potential method for the coulometric de- termination of oxide, has been observed in both the MSRE fuel and solvent salts. Current-voltage curves recorded in the positive di- rection have always shown an anodic peak at the pyrolytic-graphite indicator electrode (PGE) at about +1.0 v vs the platinum quasi-reference electrode. It has been concliuded from the shape of the curve that the electrode was belng passivated. Recently, when the PGE was observed visually while running current-voltage curves under vacuum conditions, 1t was noted that gas bubbles were coming from the end of the sheathed PGE at the potential of the anodic wave. Gas formation was also ob- served at this potential when a platinum wire was used as an indicator electrode. The gas lines servicing the electrolytic cells were redesigned to incorporate a liquid-nitrogen molecular-sieve trap to collect the gases which are evolved during electrolysis of the melt. The results of the gas chromatographic analyses of the evolved gases are shown in Table 6.7. These preliminary results indicate that the anodic wave which has been observed consistently in these fluoride melts is due to the oxi- dation of oxide (0%~ — 1/2 Os + 2e). It is believed that the CO and COp result from the reaction of the oxygen with the graphite cell prior to being pumped from the electrolytic cell. Table 6.7. Composition of Gases Obtained from the Electrolysis of LiF-Bels-Zrl, and LiF-BeF, Values in mole percent 0, CO CO, CH, H Anode LiF-BeF,-Z2rF, 2.3 3.4 83.5 8.1 3.0 Pyrolytic graphite 0.05 5.2 95.6 Graphite cell walls LiF-BeF» 6.6 16.2 75.6 0.7 <0.3 Platinum 144 The presence of Hp and CHy, may be a result of evolution of Hy at the platinum counterelectrode followed by reaction with the graphite cell to form CHy,. This is supported by the fact that neither Ho nor CH, was found to be present in the gases evolved from the second elec- trolysis experiment in Table 6.7. In this case, due to a high current density, zirconium was plated on the counterelectrode and probably re- acted with any Hp to form zirconium hydride. Further support is ob- tained from the fact that if the potential of the cathodic 1limit of LiF-Bel's-Zrt, is exceeded (reduction of zirconium) and the scan is re- versed, gas evolution from the electrode is observed only at that in- stant when all the zirconium has been stripped from the electrode. The current efficiency for the reduction of oxide in the first run of Table 6.7 was about 85%; however, a low current density was used. Higher current densities were used in the second and third runs, but the current efficiencies dropped to about 40%. In order to make feasible the controlled-potential coulometric analysis of oxide in these melts, it will be necessary to determine the proper conditions for theoretical current efficiencies. To de- crease the time of analysis, a better means of stirring the melt under vacuum is also required. It is planned to study these problems through further electrolysis experiments. Electrochemical studies were also made to identify a small re- duction wave which occurs in the MoRE fuel solvent salt. This wave, which 1s observed at both platinum and pyrolytic-graphite indicator electrodes at approximately —1.2 v vs the platinum quasi-reference electrode, 1s now believed to be due, in part, to the reduction of s small amount of hydroxide ion impurity in the melt. As previously pointed ou.t,35 however, the overall electrode reactions appear to be more complex than first envisioned. After making standard additions of anhydrous lithium hydroxide to the melt, it was observed that the overall limiting current increased and then slowly decreased over a period of several days to a reason- ably constant value. Further tests revealed that appreciable quantities of O do not remain as such in the melt bul instead react with the fluoride ion (OH™ + F~ — HF + 0°7) to form oxide and HF. This was confirmed by adding LiOH and pumping the off-gas from the electrolysis cell into a cold trap that contained a known amount of standard sodium hydroxide and then determining the amount of sodium hydroxide neutral- ized at 24-hr intervals. In one test where 220 mg of LiOH was added to the molten LiF-BeF, (43 ml), approximately 80% appeared to have reacted with the melt to evolve HI' over a period of eight days. Current-voltage curves recorded with the melt under reduced pres- sure were not significantly different from those recorded with the melt in a helium atmosphere; however, a few gas bubbles could be seen evolv- ing from the indicator electrode at a potential of approximately —1.2 v. This appears to be visual evidence that the wave is in part due to the reduction of hydroxide impurity (O + e — 0% + 1/2 Ho) and that hy- drogen is being discharged at the cathode. 145 opectrophotometric Studies of Molten-Salt Reactor Fuels The possible use of spectrophotometry is being investigated as an analytical tool for the in-line determination of U(III) at the parts- per-million level and estimation of U(IV) in molten-fluoride-salt re- actors. In past studies, experimental techniques have been developed for the spectrophotometric study of molten fluoride salts, which are corrosive to most known window materials. Spectra of both U(III)3° and U(IV)37 in fluoride salts of interest have been obtained. Tri- valent uranium in LiF-BeFp-type melts exhibits a very intense absorp- tion peak at_ 360 mp with a molar absorptivity of approximately 500 liters mole™t em™t. Tetravalent uranium in LiF-Bel's-type melts is almost transparent in the region of 360 mi, with a molar absorptivity of less than 2 liters mole™t cm™t. At 1000 mu, U(IV) exhibits a peak with a molar absorptivity of 15 liters mole ™t cm™t. Both U(IV) and U(III) are essentially transparent in the region of 700 to 710 mp. It is therefore feasible to determine U(III) down to a level of ca. 300 Ppm in the presence of up to 1 mole 9 of U(IV) in molten LiF-EEFg—ZrF4. Likewise, if the concentration of U(III) does not exceed ca. 1000 ppm it would be possible to at least estimate the concentration of U(IV) in this molten salt by determining the absorbance of the solution at 1000 mp. For both determinations the optical response of the solvent could be measured at 710 mp. The results of other spectral studies show that the corrosion products, which could be dissolved in the salt, will not interfere with the proposed determination. An idealized drawing of the spectrophotometric facility is shown in Fig. 6.16. A modified cylindrical captive-liquid cell3® would be used to contain the molten salt fuel in a windowless cell for the spectral determination. The cell would be filled either by continuocusly dripping the fuel into it or by activating the freeze valve periodically to flush the cell with fuel salt. In either case, the volume of the fuel in the cell will rapidly, in less than 6 sec, attain an equilibrium volume and a relatively constant optical path length. Excess liquids in either case will flow out of the bottom of the cell, collect on the cone- shaped bottom, and drip into the return circuit. With liquids that wet the modified captive-liquid cell, the optical path length is very re- producible, 0.064 cm * 2%, with either method of filling. With liquids that do not wet the cell, the path length is less reproducible, but tech- niques are being developed which will improve this situation. It is ex- pected that the radiation level of the volume of fuel to be held in the cell, ca. 100 pl, will not be sufficiently high to cause defect colora- tion of the windows if the Al,03 is pure. The model 14H Cary recording spectrophotometer is optically and electronically designed so that any interference caused by thermal emission from heated samples i1s essentially eliminated. This optical design will also eliminate interference from Cerenkov radiation. In the optical design of a spectrophotometer for in-line use with a nu- clear reactor, it is necessary to remove the electronic components from the highly radioactive region. OSeveral advantages can occur as a re- sult of the necessary lengthening of the optical path. The light beam 146 ORNL-DWG 65—-3975A FREEZE VALVE R Ya-in- diam CELL\\\\\\\\Q\ Al,05 WINDOW FOCUSED LIGHT BEAM {/g-in—diam APERTURE He — MATERIAL: INOR-8 Fig. 6.16. Molten-5alt Reactor In-ILine Spectrophotometric Fa- cility. of the instrument can be concentrated and imaged specifically for use with the captive-liquid cell; presently, ca. 90% of the light which is incident on a captive-liquid cell is lost because of necessary masking. oecond, the extension of the light beam by the use of mirrors will re- move any possible problems caused by high-energy gamma rays entering the spectrophotometer; these rays will, of course, be lost in reflec- tion. In cooperation with the vendor of the Cary spectrophotometer, a proposed extended optical design is being considered. A study of the analytical usefulness of reflectance spectra for the identification and estimation of U(III) and other ionic species 147 of interest in solidified powdered fluoride salts has been initiated. This technique should be of value in the analysis of irradiated solid fluoride salts and as an aid in identification of reduced species if significant reducing power 1is ever found in solid MSRE salts. In screening tests, excellent reflectance spectra have been ob- tained for salts such as UF,, PrFi, NdF3, and halides of structural metals. These spectra were obtained with a model 14 Cary recording spectrophotometer equipped with a model 1411 diffuse-reflectance ac- cessory. A sealable reflectance cell with a quartz window has been designed and fabricated so that the reflectance spectra of hygroscopic samples can be obtained without exposure of the samples to air; the cell would also serve to confine toxic or radioactive dusts in order to prevent contamination. With this cell it should be possible to examine the reflectance spectra of solidified LiF-BeFs-type salts which contain Uk, and/or UF3. This work is continuing. Analysis of MSRE Blanket Gas otudies have been continued on the helium breakdown-voltage de- tector for determining permanent gases in helium. The application of this type of detector to the analysis of helium blanket gas which will contain radiocactive fission gases seems promising. Tests with x-ray and gamma-~-ray sources and with the off-gas of the MSRE capsule test No. ORNL 47-6 indicated that radioactive gases would not be detrimental to the detector operation. The detector has been tested in a Beckman model 520 process chro- matograph for an evaluation of its operation on a continuous analysis basis. On a series of 35 samples of a standard gas mixture at the 10- ppm level, taken over a 5-hr period, the relative standard deviation of the peak heights was l%. This indicates the possibility of analysis at sub-ppm levels. The sensitivity of the detector for water in helium is estimated to be 1 ppb. A major problem in water determinations at the ppm and sub-ppm levels is the transmittal of the gas sample from the source to the detector. All lines and valves must be heated to 250-300°C to pre- vent excessive time delays in detecting concentration changes. At the present time there 1s no chromatographic sampling valve which will op- erate satisfactorily at these temperatures. A design has been conceived utilizing a pneumatically actuated six-way metal diaphragm sampling valve. Because of sealing problems in the first models of this valve, a redesigned version is being fabricated. | oince all components of this valve are metallic, problems of ra- diation damage are virtually eliminated. This valve will also be ideally sulted for application to gas analyses in a proposed in-line hydrofluori- nation method for the determination of oxide content and reducing power in the fuel of future molten-salt reactors. 148 Development and Evaluation of Equipment for Analyzing Radioactive MSRE Fuel Samples The equipment39 necessary to analyze radioactive MSRE salt samples for Be, Zr, U, Fe, Cr, Ni, ¥, and Mo was installed in the hot cells of Building 2026 by April 1965. The reliability of it was checked out by the Analytical Chemistry Division personnel responsible for its use. oeveral minor modifications were necessary; however, it was performing properly at the conclusion of the checkout period. oample Preparation The salt samples delivered to Building 2026 during the precritical and zero-power experiments were crushed, welghed, and dissolved in the hot cells. Two portions of each pulverized salt were weighed and dis- solved. The preparation of the salt samples for analysis proved to be satisfactory after a few modifications were made to the equipment and to the method of dissolution. Sample Analyses The solutions prepared from the salt samples were analyzed for U, Cr, Zr, Li, Be, Fe, Ni, and Mo. The crushed salt was analyzed for fluoride. The results obtained were generally satisfactory except for uranium and'berylliuma4o The uranium determinations exhibited a nega- tive bias of approximately l% from book values. The bias was evident during both the precritical and zero-power experiments. The beryllium results exhibited a positive bias of approximately 2% from book values. Both methods are being investigated to determine if the biases actually exist and, if so, what steps are necessary to correct them. Quality Control Program A quality control program was initiated prior to precritical sam- pling. Control samples of synthetic solutions similar to dissolved non- radioactive fuel-salt samples were analyzed along with the fuel-salt samples. The program was established to verify the true percent standard deviation in the individual methods employed in analyzing the MSRE fuel- salt samples. The accumulated results on the synthetic solutions were insufficient to determine the true percent standard deviations. How- ever, the estimated values for U, Be, Zr, Cr, Fe, Ni, and Mo were 1.0, 5.0, 5.0, 15, 15, 15, and 15%, respectively, based on the data to Au- gust 23. 10. 11. 12. 13. 14. 149 References MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 146. P. N. Hauvbenreich, Preliminary Report on Results of MSRE Zero- Power Experiments (internal memorandum). R. E. Thoma, 'MSRE Salt Chemistry During Precritical and Zero-Power Experiments, ' MSR-65-40 (Aug . 2 1965) (internal use only). R. E. Thoma, 'Chemical Analysis of MSRE Flush and Coolant Salts in Pren?clear Test Period,"” MSR-65-19 (March 19, 1965) (internal use only/. R. B. Lindauver, Preoperational Testing of the MSRE Fuel Reprocess- ing Facility and Flush Salt Treatment No. 1 (internal memorandum). Resgults are summarized in memo K-1-2079 from J. G. Million to R. E. Thoma, Sept. 1, 1965. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, pp. 23841 C. M. Blood et al., "Activities of Some Transition Metal Fluorides in Molten Fluoride Mixtures,' pp. 108-10 in Proceedings of the In- ternational Conference on Coordination Chemistry, /th, Stockholm and Uppsala, Jan. 25-29, 1962, Butterworths, London, 1963. S. I. Cohen, W. D. Powers, and N. D. Greene, A Physical Property Sumary for ANP Fluoride Mixtures, ORNL-2150 (Aug. 23, 1956, de- classified Nov. 24, 1959). S. I. Cohen and T. N. Jones, A Summary of Density Measurements on Molten Fluoride Mixtures and a Correlation for Predicting Densities of Fluoride Mixtures, ORNL-1702 (July 19, 1954, declassified Nov. 2, 1961). B. C. Blanke et al., Density and Viscosity of Fused Mixtures of Tithium, Beryllium, and Uranium Fluorides, MIM-1086 (December 1956, issued March 23, 1959). MSR Program Semiann. Progr. Rept. Feb. 28, 1961, ORNL-3122, pp. 118, 122-25. P. B. Bien, S. Cantor, and F. F. Blankenship, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1961, ORNL-3127, pp. 24—25. S. Cantor, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1962, ORNL-3262, pp. 3841. 15. 16. 17. 18. 19. 20. 21. 22 . 23, 2 150 Ixperimental work by P. B. Field, summer participant with Reactor Chemistry Division, 1965, Assistant Professor of Chemistry, Vir- ginia Polytechnic Institute. D. J. Rose and M. Clark, Jr., Plasmas and Controlled Fusion, p. 296, MIT Press, Cambridge, Mass., 1961. J. H. Shaffer, W. R. Grimes, and G. M. Watson, J. Phys. Chem. 03, 1999 (1959). = 5. T. Benton, R. L. farrar, Jr., and R. M. McGill, Preparation of Anhydrous Deuterium Fluoride by Direct Combination of the Elements, K-1585 (Jan. 29, 1964). W. H. Rodebush and A. L. Dixon, Phys. Rev. 26, 851 (1925). A. Buchler and J. L. Stauffer, IAEA Symposium on Thermodynamics, July 1965, paper SM-66/26. A. L. Mathews and C. F. Baes, Oxlide Chemistry and Thermodynamics of Molten Lithium Fluoride—Beryllium Fluoride by Equilibration with Gaseous Water—Hydrogen Fluoride Mixtures, ORNL-TM-1129, p. 104 (May 7, 1965). R. E. Thoma, ed., Phase Diagrams of Nuclear Reactor Materials, ORNL-2548, p.33 (Nov. 6, 1959). MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 137. Equation (9) was obtained by use of the following relation which applies to a countercurrent extraction system of n stages if the distribution coefficient D is constant: fraction extracted = 1 — { fiiifl 1 ; (br)™ " =1 where r 1s the ratio of the flow rates of the two phases. In the present case, and P__/RT HI : D = —=—— = QP __/RT kg/liter [I...] I_IEw/ g/ J r = liters of gas per kilogram of melt = Vg/'wS 5 QP .V o g br = WRT Ay /W 25. 20. 27 . 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. Transl.) 3, 279 (1958). 151 V. G. Selivanov and V. V. Stender, Zh. Neorgan. Khim. (1958). hw 5 a7 Tbid., 4, 934 (1959). V. G. Selivanov and V. V. Stender, Russ. J. Inorg. Chem. (English W. F. Schilling, Proposed MOSEL Reactor Program at KFA Jiilich (internal memorandum). J. H. Shaffer et al., Nucl. Sci. Eng. 18, 177 (1964). J. H. Shaffer, Reactor Chemistry Division, designed the manifold system and supervised its construction. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 160. G. Goldberg, A. 5. Meyer, Jr., and J. C. White, Anal. Chem. 32, 314 (1960). MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 162. MSR Program Semiann. Progr. Rept. Jan. 31, 1964, ORNL-3626, p. 151. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 164. J. P. Young, Anal. Chem. Div. Ann. Progr. Rept. Nov. 15, 1964, ORNL-3750, p. 6. J. P. Young, Anal. Chem. Div. Ann. Progr. Rept. Dec. 31, 1961, ORNL-3243, p. 30. J. P. Young, Anal. Chem. 36, 390 (1964). F. K. Heacker et al., MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 155-60. | R. E. Thoma, MSRE Salt Chemistry During Precritical and Zero-Power Experiments (Aug. 2, 1965) (internal use only). 152 7. FUEL PROCESSING The MSRE fuel processing system has been described in a previous report.l Construction was completed, the system was leak tested, and the tanks were calibrated. A test run using dry nitrogen with a small metered flow of vaporized water was made to test and calibrate the water metering equipment, that is, the cold trap and siphon pot for measuring the total volume of HyO and HF leaving the salt and the water analyzer developed at ORGDP . ? Operations were satisfactory, and preparations were then made to process the flush salt for oxide removal. The MSRE flush salt (66 mole % LiF, 34 mole % BeF,) had been cir- culated in the reactor fuel system for 1000 hr to remove oxide film and shake down the reactor system. The salt was tranferred, by gas pressure, to the fuel storage tank in the fuel processing system and sparged with a mixture of Hp and HF. Oxide removal was followed mainly by means of the water analyzer. Difficulty was experienced with the HF flow control causing occasional overloading of the cold trap and erratic siphon pot discharges. After 32 hr of operation, the analyzer indicated a removal of 115 ppm of oxide, and the removal rate had decreased by a factor of 15 to <1 ppm of oxide per hour. By this time the operating pressure in the system had increased from 2.2 to 6 psig due to an incorrectly in- stalled charcoal trap, and the increased pressure had caused partial tranfer of the KOH scrub solution to the liquid waste tank. Sparging with Hp and HF was therefore terminated. After the system was purged with helium at a low rate, the defective trap was removed. The salt was next sparged with Hsy, and 135 liters of HF was evolved as measured by an HF monitor in the off-gas stream. This indicated that 24 ppm of oxide as Be(OH)g was left in the salt at the end of Hy-HF sparg- ing. From equilibrium quotients at the salt temperature of 1200°F and the peak HF concentration of 0.0l atm, an [OH™] to [0°~] ratio of 0.8 was calculated. The total oxide remaining in the salt after processing was therefore 54 ppm. Although the Hp:HF ratio was closer to 5:1 than the desired 10:1 ratio, there was no discernible increase in the chromium content of the salt. A more detailed description of these operations has been reported.3 References 1. MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 20L. 2. W. 5. Pappas, Continuous Moisture Analyzer Monitors Hydrogen Fluoride Off-Gases During Oxide Removal at MSRE, K-L-6059A (June 29, 1965). 3. R. B. Lindauer, internal memorandum (July 1965). 153 OAK RIDGE NATIONAL LABORATORY JUNE 1, 1965 R. B. BRIGGS, DIRECTOR P. R. KASTEN, DEPUTY DIRECTOR* RD W. B. McDONALD, ASST. DIRECTOR** RD D MOLTEN-SALT REACTOR PROGRAM MSBR STUDIES ACepomArPOoEO I M REACTOR DESIGN BETTIS BAUMAN BETTIS* BRAATZ* CARTER** CRISTY* GRINDELL* PICKEL* ROBERTSON SCOTT* E. SPALLER* H. WESTSIK* V. WILSON* J. YOST W. TERRY C. JONES** W. KRICK* DEePprr-mMm ¥ MSRE OPERATIONS COMPONENT DEVELOPMENT INSTRUMENTATION AND CONTROLS R. L. MOORE j&C J. R. TALLACKSCN j&C A. H. ANDERSON 1&C S. J. BALL* 1&C G. H. BURGER 1&C D. G. DAVIS [1&C E. N. FRAY 1&C P. G. HERNDON 1&C J. W. KREWSON 1&C J. L. REDFORD** 1&C T. M. CATE 1&C B. J. JONES* 1&C P. E. SMITH 1&C C. E. STEVENSON 1&C W. WEIS |&C NUCLEAR ANALYSIS B. E. PRINCE R T. W. KERLIN* R FUEL PROCESSING DEVELOPMENT M. E. WHATLEY* CT R. E. BLANCO* CT H. E. GOELLER* CcT W. L. CARTER** CT G. I. CATHERS cT J. R. HIGHTOWER CT R. W. HORTON CT R. B. LINDAUER CT L. E. McNEESE CT C. E. SCHILLING CT C. D. SCOTT~ CT J. BEAMS CT C. J. SHIPMAN CT W. G. SISSON* CT mIm=E oMM rEm = 0= REACTOR CHEMISTRY GRIMES* BLANKENSHIP* McDUFFIE* BOHLMANN~* MSRE ON-SITE CHEMISTRY THOMA** S. KIRSLIS . D. NEUMANN o em IRRADIATION PROGRAM . COMPERE KELLY SAVAGE R. HART W. MYERS RC RC RC RC RC RC RC RC RC RC RC RC PHYSICAL AND INORGANIC CHEMISTRY COENTTAEEP METALLURGY TABOADA~ H. COOK™** W. DAVIS* G. DONNELLY* G. GILLILAND* INOUYE™ R. KENNEDY* R. MARTIN* M. TOLSON* R. WEIR™ J. L. GRIFFITH* E. J. LAWRENCE* M&C M&C M&C M&C M&C M&C M&C M&C M&C M&C M&C M&C * PAUL R. KASTEN* R P N. HAUBENREICH 2 Dl. zzloETWTAK E W. B. McDONALD** R . CAPITAL COST COORDINATION NUCL EAR AND MECHANICAL ANALYSIS FLOW MODELS R R. C. OLSON” R C. H. GABBARD** R J. R. ENGEL R R. J. KEDL* R R M. L. MYERS R C. H. GABBARD** R GE&C J. J. GEIST*** CE&C A. HOUTZEEL*** cT H. B. PIPER R GE&C J. A WATTS R R E G &(é METALLURGY OPERATION : O ADA Jec REMOTE MAINTENANCE GE&C A-TAB R. H. GUYMON R 2 W. H. COOK* M&C J. H. CROWLEY R R. BLUMBERG R 2 C. R. KENNEDY* M&C P. H. HARLEY R CHEMISTRY J.R. SHUGART R R G. M. TOLSON* M&C V.D.HOLT R 0 T. L. HUDSON R R. E. THOMA** RC R A. |. KRAKOVIAK R R H. R. PAYNE R H. C. ROLLER R R. STEFFY R CHEMISTRY W. $ E\;mfi:{z E COMPONENTS AND LOOPS L ANALYTICAL CHEMISTRY W. H. DUCKWORTH R J. D. EMCH R MATERIALS A. N. SMITH R J. C. WHITE* AC E. H. GUINN R R. B. GALL AHER R T.G. HILL R W. H. COOK™ M&C M. RICHARDSON R REACTOR CHEMISTRY C. C. HURTT R R. E. CARNES R J. C. JORDAN R C. A. GIFFORD R W. R. GRIMES* RC LA H. F. McDUFFIE* RC L. E. PENTON R C. JONES** R W. E. RAMSEY R C. J. BARTON* RC R. SMITH, JR. R F. F. BLANKENSHIP* RC J. L. STEPP R S. CANTOR* RC J. L. UNDERWOOD R S. S. KIRSLIS RC s R WEST 5 CHEMICAL PROCESSING J. H. SHAFFER* RC R ps J. E. WOLFE R R. B. LINDAUER** R PUMP DESIGN AND DEVELOPMENT AC ANALYTICAL CHEMISTRY DIVISION MAINTENANCE | . CT CHEMICAL TECHNOLOGY DIVISION ’;' g ;‘:I'?HDELL g , B. H. WEBSTER** R - G. D DIRECTOR’S DIVISION N E GILLEN® R L. V. WILSON R GE&C GENERAL ENGINEERING AND CONSTRUCTION DIVISION L. P. PUGH** R 18C INSTRUMENTATION AND CONTROLS DIVISION M&C METALS AND CERAMICS DIVISION P&E PLANT AND EQUIPMENT DIVISION R REACTOR DIVISION RC REACTOR CHEMISTRY DIVISION DESIGN LIAISON * PART TIME ON MSRP ** DUAL CAPACITY C. K. McGLOTHLAN** R #x* EURATOM EXCHANGE SCIENTIST C. F. BAES, JR. RC J. H. BURNS RC S. CANTOR RC J. H. SHAFFER RC R. E. THOMA** RC G. BRUNTON RC F. A. DOSS RC H. A. FRIEDMAN RC G. M. HEBERT RC K. A. ROMBERGER RC D. R. SEARS RC H. H. STONE RC W. T. WARD RC W. K. R, FINNELL RC B. F. HITCH RC W. P. TEICHERT RC ANALYTICAL CHEMISTRY L. T. CORBIN* AC J. C. WHITE™ AC R. F. APPLE AC J. M. DALE AC F. K. HEACKER AC C. E. LAMB* AC w. L. MADDOX* AC H e O NWH O 17-26. 27 . 28. 29. 30. 31. 32. 33. 34, 35. 36. 37. 38. 39. 40, 41 . 42 . 43. 45, 46 . 47 . 48. 49, 50. 51. 52. 53. 54. oo~ o0ut W H = GgrEHgOOPHA W HEH OO EAE® TN EsQgrHdgEudgsoEge R ESgE S STEHZEOOQOEHPPROEHH OEHON QHQ DO d@m@HdRdoEHepQEHETQHHPEPHEHOPPHEPES QOETW Adamson . Alexander Baes Beall . Bettis . Billington Blankenship . Blizard lumberg Bauman . Boch Bohlmann Borkowskil Boyd . Bredig . Breeding . Briggs Bruce . Burger antor . Cardwell Carter Compere Conlin Cook Corbin . Cristy Crowley . Culler . Davis . Davis . DeVan . Donnelly Douglas . Dunwoody . fIngel Iipler hrgen Fraas Frye, Jr. Gabbard Gall . Gallaher . Gilliland . Grimes 155 INTERNAL DISTRIBUTION 55. 56. 5. 58. 59. 60. ol. 02 . 63. 64 . 65. 66. o'7. 68. 69. 70. 71. 72-81. 82 . 83. 84 . 85. 86. 87. 88. 89. 20. 9. 2. 93. X, 95. 9. 7. 98. 99. 100. 101, 102. 103. 104. 105. 106. 107. 108-110. ORNL-3872 UC-80 — Reactor Technology TID-4500 (46th ed.) 2Rz aoHdomnssssm=EaddHdoagrrHaoR =2 0gdg = >=$=Wj Ho=ooxHE o238 n@mDEDnae GO HOQONPHE Y rOREHdOEOQHOWEepEHEHS HS Wg ey 0H Grindell Guymon Harley Harrill Haubenreich Herndon Hibbs (Y-12) . Hill . Hise . Hoffman . Holt Holz . Hollaender . ©. Householder . L. Hudson . Inouye . Jordan . Kasten . Kedl . Kelley . Kelly Kennedy . Kerlin . Krakoviak Krewson . Lamb . Iarson (K-25) . Lincoln . Lindauver . Livingston . Lundin . MacPherson . Martin . McDonald . McDuffie . McGlothlan . Miller . Mills Mixon . Moore Morgan Moyers Nelson . Osborn . Parker 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134, 135. 136. 137. 138. 200-207. 208. 209, 210. 211. 212. 213. 214, 215. 216. 217-554, C)fijk1ziw>fij>=C>E:E1tqfiitfl>=Eizizizififlziti?iF:21C7E1w1fi 156 . F. Parsly 139. C. D. Susano . Patriarca 140. A. Taboada . R. Payne 141. J. R. Tallackson . Phillips 142, E. H. Taylor B. Pike 143. R. E. Thoma B. Piper 144. G. M. Tolson E. Prince 145. D. B. Trauger L. Redford 146. R. W. Tucker Richardson 147. W. C. Ulrich C. Robertson 148. D. C. Watkin C. Roller 149. G. M. Watson W. Rosenthal 150, B. H. Webster C. Savage 151. A. M. Weinberg W. Savage 152. dJ. R. Weir W. Savolainen 153. J. H. Westsik Scott 154. G. C. Williams E. Seagren 155. J. C. White H. Shaffer 156. L. V. Wilson D. Shipley 157. K. J. Yost J. Skinner 158, G. J. Young M. Slauvghter 159. Bioclogy Library N. Smith 160-161. Reactor Division Library G. Smith 162-166. ORNL Y-12 Technical Library, H. Snell Document Reference Section F. Spencer 167-169. Central Research Library opiewak 170-204. Laboratory Records Department oteffy 205. Laboratory Records, ORNL RC E. Stevenson FXTERNATL DISTRIBUTION . Cope, AEC, ORO . Deering, AEC, ORO . Garrison, AEC, Washington . McNamee, Manager, Research Administration, UCC, New York . Shaw, AEC, Washington . E. Slnclalr, AEC, Washington . L. Smalley, AEC, ORO . A. Swartout, AEC, Washington . J. Whitman, AEC, Washington Division of Research and Development, AEC, ORO Given distribution as shown in TID-4500 (46th ed.) under Reactor Technology category (75 copies — CFSTI) Szibd*lj Sy EHE00 QO