ifi"lW#Mfi%flfififlfifiiififlfiflfififillfi“fl 3 4456 0362LL0 & ARTIN MMARIETTA RCY SYSTENMS i 1 i i ORNL-3936 Contract No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT For Period Ending February 28, 1966 R. B. Briggs, Program Director JUNE 1966 OAK RIDGE NATIONAT, LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the Sfimms e W il 3 4u5kL 0382610 8 £ iii CONTENTS SUMMARY 4 v s eeeansesnsosasasssassasscsoesasnsasassasnasssssnnnassanecss VIiL INTRODUCTION e s v voansssansnsas cenesaen e creeaaeen eeenes ceeenn .. 1 Part 1. MSRE OPERATIONS AND CONSTRUCTION, ENGINEERING ANALYSTS, AND COMPONENT DEVELOPMENT 1. MSRE OPERATIONS...viesennecanse checrsacscectoateasesrnansceanes 7 Chronological Account...coevesrnanes e nsedmseetcesaneaaasnan cons 7 Analysis of Experiments.. e eacerrerterenoecaronnnnnns s 10 Reachivity BalanCe.iseesreseesssssssnsoeastoansasossassssnass 10 Power Calibration........... eseearesrevesoseesssebaaacas s 12 Flux MeasurelertS s ceeaeeseacosnsssssssossossasnnescannssass 12 MSRE Dynamic Tests.c.cocinennenssnn Ceeeneeean cenaan veeansaenses 13 Description of Dynamic Testt.reer i et eiiierrenonnne 13 Frequency Response TeshbS.iviieeirriesrssssscnssssnans cerans 15 Implementation of Pseudorandom Binary Tests....eceesas.. .o 16 Analysis Procedures...... cenreceaerneonas cCeeraenecans sosnn 17 Results..... et sassastieasat e s et ennenseen e oe eo e 18 Temperature Response Tesh.c.eeeeeiniraittssorsesssssosnassnes 22 Systems Performance..eeecoacasas teoenemassnas ceeaans ternasseanns 23 Off-(a8 SySbelleeeessseeennsnaronssasooranasnanssennsonsonss 23 Salt-Pump Oll SysbemS..cesereessosssnnesssssarssenanssssasana 27 Treated Cooling-Water System......oeee. Creersesesasnassenss 28 Secondary Containment....... . ... tecmesensasaenas cennns veeas 29 Sh1elding.esecasesanseassaosonssssasosansasassoronnaanaenss O Component Performance.coeeeeeeesss. Ceeeaneseereesesentaanann oo 34 RAALAE O e e e svesvensencuonsonnssoasscnssossansssasnosaasnses Cther Components........ tasssesesarssiesestnasrasnoannss 36 Inspection of the Fuel Pump e eteeesacenean Creceanaanas .. 38 Heat Treatment of Reactor Vessel...eeeeerescencsocarsssanns 39 Stress Analysis of Reactor Piping and Nozzles...eeeernecann 40 Instrumentation and Conbrols.ee s e nrvacsassaonsas besesa s 41 General.sevesveesanecns feasseseseseesesoaarsannans cesans .o 4L Safety Instrumentation....... e eresnasseavreeennroennasoes AL Wide-Range Counting ChannelsS...ceeeseesoveoescanccosononsan 42 Nuclear Instrument PenetratioN......ceceseeses Creneeneeneas A2 BF3 Confidence Instrumentation..veessessoeecanenccaasaseses 44 Personnel Monitoring Systelm...cveieieesteensciececsanns creens 45 Control Instrumentation..s.ceeereessersesscesssosassnansasns 4O Operating Experience — Process and Nuclear Instruments..... 47 Data Sysbem..seiieensonaearoas e Cerceenanaraoans vesna. A9 MSRE Training SimulatorS..ceseececrsessssosesssscsacnnn casen 50 Documentation..... caseesaansan ceessesaaesessseesotarrneroas 51 iv 2. COMPONENT DEVELOPMENT ... .eeeeeeenanoosas ceeatesaanens e ceen 53 Freeze ValveS.s.eusas Ceeesacens creesecaea teneacannrease ceeteaa 53 Control RodS.eeevasen. thecesreassneres e na e cerieasans e 53 Coutrol Rod Drive Unlts ceranas Ceeseaseas s chearseneas 54 Radiator DoorS.ieesssessensranas ceraees feeet et et e 54 Radiator Heater Electrical Insulation Failure.......... carerree 57 S L e - e e st s v ettt st et ansossostsesesosarensoscnansocnses 58 Coolant Dalt Sampler...coeeeeenss Ceeesereresearsara e ceseenas 59 Fxamination of Components from the MORE Off-Gas System......... 60 Capillary Flow Restrictbor FE 52l.. it eieeinererescranennns 60 Check Valve CV 533............ Ceeeesareerann Ceretsasseseans 60 Charcoal Bed Inlet Valve HV 62l.. ¢ iiiiiinennrencncanrons . 60 Line 522 Pressure Control Valve PCV 522. ...t iiinnennas 62 Line Filber 522. it iiiiuiiteeestnsensessatsenssassanaananas O Flow Test on the MSRE Filter from Line 522 ........... ceenan 65 Fuel Processing System Sampler...... C i tiieeretees s cee e . 65 Of T ~Ga8 CaND T sttt eeroerossoanosseacnsssasosssnsasssnsssssnss 67 Xenon Migration in the MSRE......cvveineveen.. Ceeteraneceanaas . 69 Remote MaintbenanCe .o reeerisorscsssrssssnanns ceeaceenaas cheean 70 Practice Before Operation..vee e eeieersteceeceraasanscanans 72 Maintenance of Radioactive Systems..iieiiieierescens Pr e 72 Pump Development. oo eseeeeersitotoesosnssassssssscanasnnasssenas T MSRE PUMPS e e v vevencensns ceeeenne C e easerereiteeetateranans T4 Other Molten-Salt PUmpPS . vesesseetecenssscencsnanaan ceeeaas 75 Instrument Development..oeeeeeeieierveseoassnsasssososas cereans 77 Ultrasonic Single-Point Molten-Salt ITevel Probe..... ceaean .77 High-Temperature NaK-Filled Differential-Pressure Transmitter.....cieevevvenn. trectravescrsacerassssssanens 77 Float-Type Molten-Salt Level Transmlttef C et et i etn e 78 Conductivity~Llype Single-Point Molten-balt Ievel Probe..... 78 Single-Point Temperature Alarm SwitchesS..viieeereereeeennnn. 78 Helium Control Valve Trim Replacement........... tereeraanaa 79 Thermocouple Development and Testing......... ceenes ceceiens 79 Temperature SCanner....... Gttt eseassecaer et cerenes ceeea 80 3. MSRE REACTOR ANALYSTS..ieeeseaoenaes sesecsnraeenaa seersescensss 82 Least-5Squares Formula for Control Rod Reactivity...ieveeeoensons 82 Spatial Distribution of '3°Xe Poisoning in MSRE Grephite....... 87 Part 2. MATERIALS STUDIES 4o METATILURGY . iveveeennens cirerareas ceeceanaan e itesesieaenar e 95 Dynamic Corrosion Studies......... teeeenrasens b aseecessassenans 95 MSEE Material Surveillance TestsS..ieesreitensoseens ceseaaeraease 96 Reactor Survelllance SpeCimeNnS...eescietsessaessaasearscsnns 96 Survelllance Control SpeCllelS .. oeseessossassssscasasssasss 99 Hot-Cell Metallographic Examination of Hagstelloy N from Ixperiment MIR-47-6 for Evidence of Nitriding........ cerrraan 100 Ut Posgtirradiation Metallographic Examination of Capsules 1—4 from Experiment ORNL MIR-47-6.ceirrernnareennnaan e 101 Development of Craphite-to-Metal Joints....... caranaaen vessanae 101 Tests of Graphite-Molybdenum Brazed Joint for Contalning Molten Salts Under Pressure..cvvrseesenssass feseaenas e censanss 104 New Grades of Graphite . veeiescoerssoctssensacsssassososssnnas . 107 Evaluation of the Effects of Irradiation on Graphite.......... . 108 " EBffects of Irradiation on Hastelloy N.ivesriooneeesesassoneanass 111 Weld Studies on Hastelloy W....... teecnecaes chearesaeans sereasae 115 MY T RY 4 s v sesonoassssonssasnsasanssassassosnssaass heereear e 122 Chemistry of the MORE..veeiereeorensacensesoronsoncssnssssesess L22 Apalyses of Flush, Fuel, and Coolant Selts..... e reas. 122 Fxamination of Materialsg Trom the MSRE Off-Cas System...... 124 Uranium~Bearing Crystals in Frozen Fuel....... teesenaesones 128 Physical Chemistry of Fluoride Melts........eovennns e . 128 Vapor Pressure of Fluoride Melbs...veeiieieeernonneasnan, oo 128 Methods for Predicting Density, Specific Heat, and Thermal Conductivity in Molten Flucrides.....oevee.. eeees 130 Oxide Solubilities in MSRE Flush Salt, Fuel Salt, and Their MIxtures........ e N B X Ceparations in Molten Fluorldes....ccevvearsvcncnse cereanna ceees 136 Fvaporative-Distillation Studies on Molten-Salt Fuel Components...osne. ceenaen Ceesreenanes Ceresosceeaunenan eas 136 Effective Activity Coefficilents by Evaporative Disbillatlion.siseseeesrtoccoeesnsresocassconsssnacaans e 139 Fxtraction of Rare Earths from Molten Fluorides into Molten Mebals. eeeeeeeeacoosncrssnnonsonnsnse eeson cesesaes 141 Removal of Protactinium......... P 75 Radiation Chemistry........... feieesatsesacs e O 57 Tn~-Pile Molten-Salt Irradiation Experiment........c.cocena.. 152 Development and Evaluation of Methods for the Analysis of the MSRE Fuel..ovvvenvnnannas cersanan feerana cevons cireas. 154 Determination of Oxide in MSRE FUel...ceeerenionnarrccocnen 154 Voltammetric Determination of Ionic ITron and Nickel in Molten MSRE FUEL.s.uecovrarncossssosasansssscsosnsansnness 1OZ Development and Evaluation of Equipment and Procedures for Analyzing Radicactive MSRE Salt Samples..... et ear s eee 165 Sample Analyses....ceveaen Cheeseaarsesresaesaenas cerarnenanseass 167 Quality Control Program........ e e sseanssecannans Ceeeraraaaan 168 MOLTEN-SALT BREEDER REACTOR DESIGN STUDIES. .. .eeeeevensaenaaass 172 MSBR Plant Desigi.vsevecsasscssorascsesessssesoscsarsassason eaee 172 FloWSheet oo venreresannnannnas Ceeeeeseeteeeeeesaneonansanass 172 Reactor Desigheeeseeeeoceonnass cheraraeennas ceteensasaanaes 174 FTuel Processing...ceeeessesesss faareesesaaans serseas eenes L77 Heat Exchange and Steam SystemsS.....veessescesesesseraeeasas 181 Capital Cost Bstimates......ocevnnnenn.. O < ¥4 Reactor Power Plant..... rteeseastas et eaeanesnnan cessaeeas 182 Fuel Recycle Plant.......... g < 2 vi Nuclear Performance and Fuel Cycle AnalyseS....vveeeenen. veess. 186 Analysis ProcedireS. . voiersesssanens ceaeans ceeareenes ceeeen 186 Basic Assumpbions..eieeeevieicenean. Gt et rerrtesasesstesens ... 186 Nuclear Design Aralysis...... ceeriaeseanaaaes Cerecesaan ce.. 189 Power Cost and Fuel Utilization Characteristics......ccvcciaens 191 MOLTEN-SATY REACTOR TROCESSING STUDIES. ... iieirevrsesesrosonenns 193 Semicontinuous Distillation..c et ieieneisacssssassosssnansans 194 Fuel Reconstibubion..coieiivr ittt eneoncnearonnnns Crreeaean 199 Continuous Fluorination of a Molten Salt.......coveveennn Ceeaaee 202 Chromium Fluoride Trapping..cvevrieeceesss tettesseaasanenan ceees 202 Degign and Evaluabtion SHuly...ovveeireniiiiniiiiiianoriinanannss 203 Description of Fuel Process.....ceevens Cereeiraeteaes Ceeenaenue 204 Description of Ferbtile Process..ieeeiieeriinenenneens Ceeeaes 208 Waste Treatment........cveinenne.. et eesanenseensnsaneeas 208 OFf-Gas Treatmenl .. e e cnrernsceonserosssnasnsesons Ceraesae 209 Surmary of Capital and Operating Costs............. Ceaaaae 209 Processing CosSt.iernanienrnnncnans e et e cesa e 210 SUMMARY Part 1. MSRE Operations and Construction, Fngineering Analysis, and Component Development 1. MSRE Operations Preparations for power operation were completed, and the MSRE was operated at nuclear powers up to 1 Mw before the system was snhut down to replace a space-cooler motor and to relleve plugging problems in the off- gags system. The power preparabtions included some system modifications shown to e required by operating experlence and by continuing development and analysig work. Remote maintenance techniques were tried and evaluated, some specilal tests were performed, the operators were trained and guali- fied for power operation, and the secondary containment was sealed and shown to have an acceptably low leak rate. The nuelear performance of the system at 211 powers up Lo 1 Mw was highly satisfactory. Reproducible reactivity behavior and a lack of gignificant cross contamination hetween the fuel and flush salis were demonstrated. Dynamics tests at power showed that the reactor has a slightly wider margin of stability than had been predicted from calcu~ lations. Preliminary results indicate that xenon poisoning may be lower than was anticipated. : The performance of most of the egquipment was satisfactory, but substantial operational difficulty was caused by plugging of very small copenings in off-gas system components by organic material. This problem was extensively investigated after the shutdown from 1 Mw. Other, less gserious problems ineluded the {reak failure of an electric motor inside the secondary containment, activation of the corrosion inhibitor in the treated cooling water, air entrainment in the cooling water, and ex- cegsive radiation levels in a Tew remote areas. Solutions have been developed for all the problems except the off-gas plugging, which 1s still under study. Formal design of the instrumentation and controls systems for the MSRE was completed. Additions and modifications are now velng mads as needed to provide additional protection, improve performance, or provide more Information for the operators. The addition of a low-level BF3 counting channel with control func- tions, the addition of cadmium shielding in the neutron instrument pene- tration, and changing reactor period Interlock trip points were required to obtain satisfactory performance of the nuclear instrumentation system. The remaining changes were mostly of secondary importance. GSome sporadic viii difficulties were experienced with individual hardware items such as air and helium valves and electronic switches. Most of the work done on the instrument system can be characterized as debugging the original instal- lation. The data~logger—computer was put into operation in conjunction with the reactor. Although the performance has not been up to expectations, it is proving useful to the operation of the reactor experiment. A power level simulator was asscmbled, and it operated satisfactorily for the training of operators. 2. Component Development The "fast thaw" requirement was eliminated in all freeze valves ex- cept for those which control the emergency drain of the reactor and of the coolant system. The operation of all the valves which might contain sufficient radiocactivity in the salt to produce radiolytic fluorine at low temperature are now operated above 400°F, which is sbove the threshold Tfor flucrine release. The braided wire sheath cover for the convoluted hose of control rod No. 3 was found to be severely torn about 2 f{ below its upper end. The cause was traced to a Jammed roller in the upper bend of the control rod thimble. The roller was replaced, and the upper rod sheath was repaired. There has been no further difficulty after several months of operation. Control rod drive unit Wo. 3 was replaced because of a shift in the remote position indication and because of a tendency for the lower Iimit switch to stick. The shift of the indicated position was eliminated by removing the excess slack 1in the chain, which had allowed the chain to slip over the sprocket. The sticking lower limit switch is being cured by replacing the return spring with a stronger one. Modifications were made {0 the radiator door guide tracks and lock mechanisms to allow for thermal distortion, found after the initial oper- ation of the radiator doors at temperature. Alterations were made o the limit switch system to prevent a damaging overtravel of the door in the upper end of the travel. A "loss-of-tension" device was designed which will stop the radiator door drive unit should the door support cables show any slack as the door 15 being lowered. This arrangement is intended to prevent damage to the support cable if the radiator door Jams, as well as to indicate a mal- function. Failure of the insulation on the electrical leads to the radiator heaters was traced to excessive heat leakage into the areca immediately above the radiator. Changes were made to reduce the temperature in this area, and electrical insulation with a higher temperature rating was in- stalled. ix Several changes were made to the sampler-enricher to improve the operation and safety of this system. Among these were the changes made to the interlock circuit, which require that additional barriers be present during certain critical operations, thereby assuring double con- tainment at all times. Forty samples were taken during runs 4 and 5, nine of which were large 50-g samples for oxygen analysis. These larger samples caused some difficulty until the capsule design was altered slightly to make it hang straighter. One of the operational valves developed a 20-cc/min helium leak across one of the two sealing surfaces of the gate. Since this is one of two valves in the line and the leak 1s clean buifer gas, the valve was not replaced. During the same period ten samples were removed from the coolant system, two of which were the larger 50-g samples. The first sample taken after an extended shutdown had a black film on it, which was ldenti- fied as decomposed oil. Although there was oll in the general arca the exact source was not established. No films were found on subsequent samples. The design and installation of the fuel processing system sampler is proceeding. A system is being designed to permit analysis of the reactor off-gas gtream. It will contain: 1. a thermal conductivity cell for on-line indication of the gross con- taminant level, 2. a chromatograph for quantitative determination of contaminant, 3. & refrigerated molecular sieve trap for isolation of a concentrated sample for transfer to a hot laboratory for isotoplc asnalysis. Estimates of the '3%Xe poison fraction for the MSRE were computed 35 a function of several parameters. At 10 Mw the resulls indicate that the poison fraction is 1.6%. It was found that the mass transfer coef- ficient from the salt to the graphite is controlling the transfer and that the properties of the graphite are not important. The remote maintenance group gained more experience with the reactor components during the period pricr to power operation. Among these were removing and replacing the pump rotary element and replacing the graphite sampler assembly. After a short period of power operation several oper- ations were performed, using remote maintensnce technigues, on & mildiy radiocactive gysten. The MSRE pump test Tacility was modified, and the prototype pump was operated for periods of 165 and 166 hr at 1200°F to provide shake- down of the spare fuel pump impeller and the spare coolant pump drive motor. The spare rotary element for the fuel pump was modified to pro- vide positive sealing against oil leakage Trom the shaft lower seal catch bhasin into the system past the outside of the shield plug. The drive motor containment vessel was redesigned, and the new design will be used for the fifth drive motor vessel. Modified ejectors were installed on tne lubrication systems for the MSRE salt pumps, and the lubrication pump endurance test was continued. The ME=2 fuel pump tank design was completed and is being reviewed. The PK-P molten-salt pump continues on endurance operation and has operated for 22,622 hr. The pump containing the molten-salt bearing was placed in operation, but the bearing seized after 1 hr of operation. Efforts to improve the stabllity of the ultrasonic level probe in- stalled in the MSRE fuel storage tank were continued without success. Testing of a NaK~filled differential pressure transmitter which failed in service at the MSRE was continued. Performance of the instru- ment was improved by refilling with silicone o0il but is stilli not satis- factory. Performance of the ball-float-type transmitter iunstalled at the MSRE continues to be satisfactory. Some difficulties were experienced with a similar (prototype) transmitter on the MSRE pump test loop; how- ever, these troubles were anticipated and corrected in the design of the MSRE model. Performance of the conductivity~type level probes installed in the MSRE drain tanks continues to he acceptable. Observation of the performance of 110 single-point temperature- alarm switches is continuing. Data obtained Lo date are insufficient to determine whether set-point drift in these switches is excessive. Testing of alternate trim meterial combinations for the helium con- trol valves was terminated. Some additional valve failures have occurred. Results of final checks indicate that errors in the coclant-salt- radiator differential temperature signal, produced by thermocouple and lead-wire mismatch, have been eliminated. Drift testing of selected MSRE-type thermocouples was concluded. The Cinal temperature equivalent drift values were between +4.7 and +6.4°F, Performance of the MSRE temperature scanning system continues to be satisfactory. Calibration drift appears to have been eliminated, and reliability is much better than had been expected. xi 3. M3HEE Reasctor Analysis For the purpcse of on-line computation of control rod reactivity with the TRW-340 data logger, a mathematical formula was fitted to the rod-worth vs position curves obtained from calibration experiments. The form of the expression used was obtained by applying a periturbation technique to evaluate the integral expression for the rod reactivity. A linear least-squares curve-fitting procedure was then used to evaluate the unknown coefficients in the resulting Tunctional expression. Close agreement between calculated and experimental curves was cblalned for those configurations of shim and regulating rods of interest in monitors- ing the contrel rod reactivity during operation. Theoretical calculations were made to estimate the inlfluence ol the overall spatial distribution of 135%e gbsorbed in rores near the grapnite surfaces in the reactor core. The purpose was to determine spatial cor- rection factors for use in the on-line calculation of *2%Xe reactivity with the TRW-340. Basged on an approximate model of the reactor core, thege calculations indicated that the equilibrium 1353 reactivity at 10 Mw is reduced by a factor of about 0.76 relative to the value obtained from a "point" calculation. In addition, this correction was found to depend on the time history of the power level. Results of calculations are presented for step changes in power level, increasing to and de- creasing from 10 Mw, Part 2. Materials Studies 4o Metallurgy Thermal convection loops made of Hastelloy N and type 304 stainless steel have circulated molten fuel salt for 33,000 and 22,000 hr, re- spectively, without incident. A Cb—1% Zr loop circulating lead at 1400°F with a 400°F AT was found to produce columbiwm crystals by mass transfer. Specimens of Hastelloy N and grade CGB graphite showed no detectable changes as a result of 1100 hr exposure to molten flucride salts in the MSRE core during the precritical, initial critical, and assoclated zerc- power experiments. The: reactor control specimen rig, which will establish base-line data by exposing graphite and Hastelloy N survelllance specimens t0 ap- proximately the operating conditions of the MBRE except for radiation, has been loaded with salt and is veing calibrated with the computer that monitors the MSRE. Metallographic examination of capsules from in-pile experiment MR 47-6 showed no evidence of nitriding of the Hastelloy N. Ho apparent X111 change in wall thickness or evidence of attack was observed, although an unexplained change in the etching characteristics of the grain boundaries at the surface was noted. Development of methods of Joining graphite to metal has included: (1) the design of a transition joint to reduce shear stresses arising from thermal expansion differences and (2) screening tests on potential brazing alloys. A small pipe of grade CGB grapnite brazed to molybdenum satisfactorily contained molten fluoride salts at 700°C under pressures of 50, 100, and 150 psig for periods of 100, 100, and 500 hr respectively. This is the first of a serles of tests of graphite-to-metal Jjolnts to determine if such Jjoints are corrosion resistant and mechanically adequate for the re- quirements of molten-salt breeder reactors. A few samples of needle~coOke graphite and isotropic graphite have been obtained and are being evaluated to determine their suitability for use in molten-salt breeder reactors. The radiation-damage problems were evaluated for graphite in advanced molten~-galt reactors, considering growth rate, creep coefficient, flux gradient, and geometric restraint as important factors. The stress de- veloped by differential growth in an isotropic graphite should not be allowed to exceed the fracture strength of the graphite and thus cause failures. The estimated life of graphite is at least five years before failure from inability to absorb creep deformation. The major uncertainty scems to be the ability of grapnite to sustaln dosecs of 2 X 1023 nvt with- out loss of integrity. Creep~-rupture life of Hastelloy N was found to be less affected by ir- radiation as the stress levels are lowered. The effects of irradiation temperature on the postirradiation creep life of air-melted heats are un- certain. Vacuun-melted heats show a large dependency on irradiation tem- perature. Pretest heat treatment can improve the ductility of irradiated specimens. The creep~rupbure properties of structural material in the M5RE appear ©o be better than originally predicted on the basis of linear extrapolation of data for stress vs log of rupture time. Experimental welds have been made {0 study methods of improving the weldability of Hastelloy N. 5. Chemistry Three innovations have been introduced in the chemical analysis of MSEE salts: a new end point for uranium titrations, a new method for determining structural-metal ions, and a new method for oxide analyses. Together they have given increased assurance that fuel conforms to the inventory composition and that the chemical purity of the salt has been maintained. Examination of deposits believed to have been responsible for the plugging of offwgas lines in the MSRE revealed the presence of oil and polymer products presumed to have formed from oil. A negligivle amount of salt was found. The formula for the uranium-bearing crystals in the frozen fuel has been found to be I4iF-UF, rather than 7[iF-0UF,; as formerly supposed. In a study of the physical chemistry of fluoride melts, vapor-pres- sure measurcments have been made for three compositions in the ILiF-Bels gystem. Because of vapor-phase association, the apparent volatility of LiF increases with decreasing concentration of IiF in the melt. Methods have been developed for predicting density, specific heat, and thermal conductivity in molten fluorides. The solubility of oxide in MSRE~-related fluoride mells has been re- evaluated with improved experiments. When increasing amounts off ZrF,, as present in the MSRE fuel, are added to flush salts, the capaecity for oxide first decreases, then Increages. Interest in reprocessing methods for MSBR fluorides has led (0 coOn- tinuing studies of distillation and of chemical reduction as a means of separating rare earths from fuels or protactinium from blankets. The composition that yields MSRE barren solvent as dilstillate has been found; this product distills, leaving the rare earths behind. Alloys of bismuth containing a small amount of lithium have proved very effective for re- ducing and extracting rare earths inte a liquid-metal phase. Thorium has been found effective as a reducing agent for removing protactinium from tlanket melts. Protactinium can also be removed on ZrOs. The in-pile molten-salt loop experiment and assoclated auwxiliary equipment are being fabricated and assembled so that modifications to beam hole HN-1 in the ORR and installation of equipment can begin in April. The precision and accuracy of the hydrofluorination method for de- termining oxide in MSRE salts were established. The method was applied to the analysls of nonradicactive samples taken during the startup of the M3R; the results were in reascnable agreement with those obtained by the KBrr, method. The hydrofluorination apparatus for the determination of oxide in radiocactive samples was fabricated and tested and is now being installed in a hot cell. Tonic iron and nickel were determined voltammetrically on a sample of molten fuel withdrawn from the MSR. These measurenments indicate that the major fraction of iron and of nickel in the fuel 1s present in an un~ionized state, presumably as finely divided metal. Also, a well- defined voltammetric wave for the reduction U(IV) — U(IIL) was observed. xXiv Efforts were continued on the development and evaluation of equip- ment and procedures for analyzing radicactive MSRE salt samples. The coulometric uranium procedure was modified to eliminate a negative bias. Both flush- and fuel-salt samples were analyzed Tor U, Zr, Cr, Be, ¥, Fe, Ni, and Mo. The analyses were routinely performed in the hot cells of the High-Radlation-Level Analytical Iaboratory. The quality control program was continued during the past period. The results obtalned on synthetic solutions established more reallstic limits of error for the methods employed. 6. Molten-Salt Breeder Reactor Design Studies Design and evaluation studies were made of tThermal molten-salt breeder reactors (MSBR) in order to assess their economic and nuclear potential and to identify important design and develcpment problems. The MSBR reference design concept is a two-region, two-fluid system with fuel salt separated from the blanket salt by graphite tubes. The energy produced in the reactor fluid is transferred to a secondary coolant-salt circuit which couples the reactor to a supercritical steam cycle. On- site fluoride volatility processing is employed, which leads to low unit processing costs and economic reactor operation as a thermal breeder. The resulting power cost is estimated to be 2.7 mills/kwhr for investor- owned utilities; the associated fuel cycle cost is 0.45 mill/kwhr (electrical); the specific fissile inventory is 0.8 kg/Mw (electrical); and the fuel doubling time is 21 years. Development of a protactinium removal scheme for the blanket region of the MSBR could lead to power costs of 2.6 mills/kwhr (electrical), a fuel cycle cost of 0.33 mill/kvwhr (electrical), a specific fissile inventory of 0.7 kg/Mw (electrical), and a fuel doubling time of 13 years. 7. Molten~Salt Reactor Processing Studies A close-coupled facility for processing the fTuel and Tertile streams wi.ll be an integral part of a molten~salt breecder reactor system. The Tuel salt will be processed on a 40-day cycle. The uranium will be re- moved from the carrier salt and fission products by fluorination, and the carrier salt will be recovered from the fission products by a semicon- tinuoug vacuum distillation. Relative volatilities between lithium and the rare carths have been measured to be 0,01 to 0.04 at 200 to 1050°C. The reconstitution of the Tuel salt, by combining the purified carrier salt with the purified UFg, can be done by direct absorption of the Ulg in fuel salt which already contains some UF, and subsequent reduction of the intermediate uranium fluoride to UF, with hydrogen. Experimental tests showed rapid and complete absorption. The primary proolems in con- tinuous Tluorination of the fuel salt to remove the uranium are corrosion XV and getting adequate mass transfer and countercurrent flow fto assure good recovery. Corrosion can probably be eliminated by the use of a layer of frozen salt on the wall of the vessel., BExperimental work with a small countercurrent continuous fluorinator gave recoveries of 290 to 96% of the uranium. Fluorination during the processing of the fuel from the MSRE produces volatile chromium flucorides. These can be effectively trapped, with negligible uranium losses, by use of sodium fluoride beds. A pre- liminary design study has been made on a conceptual processing plant in- corporating the above concepts. Among the problems which this study illuminated were the complications from the handling of high-heat- generating materials. The fixed capital cost for the conceptual plant was $5.3 million; the salt inventory cost was $0.196 million, and the direct operating cost was $787,790.00 per year. INTRCDUCTION The Molten-Salt Reactor Program is concerned with research and de- velopment for nuclear reactors that use moblle fuels, which are solu- tions of fissile and fertile materials in sultable carrier salts. The program is an outgrowth of the ANP efforts to make a molten-salt reactor power plant Tor aircraft and is extending the technology originated there to the development of reactors for producing low-cost power for civilian UELS. The major goal of the program is to develop a thermal breeder re- actor. Fuel for this type ol reactor would be 233UF4 or 2'35UF4 dissolved in a szalt of composition near 2LiF-BeFs. The blanket would be Thly dis- solved in a carrier of similar composition. The technology being devel- opad Tor the breeder is applicable to, and could be explolited sooner in, advanced converter reactors or in burners of fissiocnable uranium and plu- tonium that also use fluoride fuels. Bolutions of uranium, plutoniurm, and thorium szalts in chloride and fluoride carrler salts offer attractive possibilities for mcobile fuels for intermediate and fast breeder reactors. The fast reactors are of interest Loo bubt are not a significant part of the progran. Our major effort is being applied to the development, construction, and operation of a Molten-Salt Reactor Bxperiment. 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 obtaln several years of experience with the operation and maintenance of a small molten- salt power reactor. A successful experiment will demonstrate on a small cscale the attractive teatures and the technical feasibility of these sys- tems for large civilian power reactors. The MSRE operates at 1200°F and at atmospheric vrecsure and willl generate 10 Mw of heat. Initlially, the fuel conbains 0.9 mole % UF., 5 mole % ZrF,;, 29.1 mole % BeF,, and 65 mole % LiF, and the uranium is about 30% °°°U. The melting point is 840°F, In later operation, we expect to use highly enriched uranium in the lowzr concenbration typilcal of the fuel for the core of a breeder. In each case, the composition of the solvent can be adjusted to retaln about the same liguidus temperature. The fuel cirvculates through a reactor vessel and an external pump and heat-exchange system. All this egquipment is constructed of Hastelloy N,* a new nickel-molybdenum-chromium alloy with exceptional resistance to corrosion by molten fluorides and with high strength at high tempera- ture. The reactor core conbains an assembly of graphite modzsrator bars that are in direct contact with the fuel. The graphite is a new material? of high density and small pore size. The fuel salt does not wet the sraphite and therefore should not enter the pores, even at pressures well above the operating pressure. LA so s0ld commercially as Inco No. 806. “Grade CGB, produced by Carbon Products Division of Union Carbide Corpe. Heat produced in the reactor is transferred to a coclant salt in the heat exchanger, and the coolant salt is pumped through a radiator to dissipate the heat to the atmosphere. A small facility is installed in the MS5KEE building for occasionally processing the fuel by treatment with gaseous HF and Fs. Degign of the MSRE 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 obltaining materials aand 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 experiments were completed carly in July. Additiocnal modifica- tions, maintenance, and sealing and testing of The containment were required before the reactor began to operate at appreciable power. This work was completed in December, and the power experiments were begun in January 1966. The reactor had been operated for a short time at 1 Mw at the time of this report. lIurther increases in power were delayed by difficulties with the off-gas system. 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 alsc being done on methods for purifying the fuel salls and 1in preparing purified mixtures for the reactor and for the research and development studies. Some studies are being made of the large power breeder reactors for which this technology is being developed. This report is one of a series of periodiec 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 MSKE. It also de- scribes much of the general technology for molten-salt reactor systems. Other reports issued in this series are: ORNT,-2474 Period Ending January 31, 1958 ORNL-2626 Period Ending October 31, 1958 ORNI-2684 Period Ending January 31, 1959 ORNL-2723 Period Ending April 30, 1959 ORNT,~2799 Period Ending July 31, 1959 ORNL,-2890 Period Ending October 31, 1959 ORNL~2973 ORNL-3014 ORNL-3122 ORNL-3215 ORNL-3282 ORNL-3369 ORNL-3419 ORNI-3529 ORNL-3626 ORNL-3708 ORNL-3812 ORNL~3872 Periods Ending January 31 and April 30, 1960 Period Period Period Period Period Period Period Period Period Period Period Ending Ending Ending Ending Ending Ending Ending Inding Ending Ending Ending July 31, 1960 February 28, 1961 August 31, 1961 February 28, 1962 August 31, 1962 January 31, 1963 July 31, 1963 January 31, 1964 July 31, 1964 February 38, 1965 Auvgust 31, 1965 Part 1. MSRE OFPERATIONS AND CONSTRUCTION, ENGINEERING ANALYSIS, AND COMPONENT DEVELOPMENT 1. MSRE OFPERATIONS Chronological Account Preparations for operation at high power were completed, and the ex- perimental program was resumed. The power ascension was interrupted at 1 Mw, however, by partial or complete plugging at several points in the fuel off-gas system. The plugging materials were identified as organics, probably the products of oil decomposition. Figure 1.1 outlines the major activities in the period covered by this report. A brief account follows; details are given in later sec- tions. Two of the larger modification jobs scheduled before power opera- tion, the coolant line anchor sleeves and the installation of new ra- diator doors, were completed in August. Late that month, the assembly of graphite and Hastelloy N surveillance SpéClmenS,'Whlch had been in the core from the beginning of salt operation, was removed. While the reactor vessel was open, inspection revealed that pleces were broken from the horizontal graphite bar supporting the sample array. The pleces were recovered for examination, and a new sample assembly, degsigned for exposure at high power and suspended from above, was installed. The fuel pump rotary element was removed in a final rehearsal of remote maintenance and to permit inspection of the pump intermals. It was reinstalled after inspection showed the pump to be in very good con- dition. Tests had shown that heats of Hastelloy N used in the reactor vessel had poor high-temperature rupture life and ductility in the as-welded condition. The vessel closure weld had not been heat treated, so it was heated to 1400°F for 100 hr, using the installed heaters, to improve these properties. At the conclusion of the heat treatment, the reactor cell was sealed for the first time. The drain cell had already been sealed, and some of ORNL--DWG 66 -3546R FUEL-PUMP } { INSPECTION CONTAINMENT TESTING FUEL CIRCUIL ATlON VAT VR T T PR A 22T Ao o A 2 o it - HEAT TREATING PP FLUSH FOOLANT CIRCU{ ATION REACTOR VESSEL STREssgs & CETEEZR oo IR0 CORE ] V2T RADIATOR — 4 \ OFF GAS SYSTEM %AIVIP FS WIRING T T A """ DOQFS_'MECHANSMS EZITE I E’_}?‘;'A&] ................... Q] OPERATOR TRAININF '-2 m ......... AAAAAAA -_SEPT OCT fi DEC JAN FEB the closure devices on containment penetrations had been taested. Now test- ing of the containment provided by the reactor cell, drain cell, and vapor- condensing system became the primary effort. After preliminary tests at pressures up to 5 psig, the program was interrupted on October 21 to in- stall Masonite sheets between the cell mewbranes and upper blocks to modify the access opening over the core. Testing at 20 psig disclosed many small leaks at penetrations, which were repaired (many'while the reactor cell was cpen for ten days for straln-gage measurements of pip- ing stresses). After the repairs, leakage rates were measured at 10, 20, and 30 psig. Ixtrapolation to 39 psig (the peak pressure in the maximum credible accident) gave a leak rate of O.4%/day, compared to l.O%/day assumed in the safeby analysis. Leakage rates at —2 psig (the normal operating pressure) were measured to serve as a reference during subsequent operation. This program was completed on December 5. Stiresses in the reactor cell piping and vessels were caleulated in detalil to permit evaluation of the service life of the Hastelloy N parts under irradiation. Some adjustments of supports were made to minimize stresses, after which the calculated stresses were acceplably low ex- cept at the heat exchanger nozzles, where a complicated geometiry made calculations unreliable. The strain-gage measurements in early Novem- ber showed tolerable stresses at this point also. While the contaimment testing and strain-gage measurements were under way, the operators and supervisors underwent further training with emphasis on power operation. Classroom lectures were followed by practice on a simulator which included the actual controls, instrumen- tation, control rods, and radiator doors.t Examinations and certifica- tion of qualified operators followed. Early tests with the radiator hol and the exercises during simula- tor practice showed that the operaticn of the radiator doors was unre- liable. TFour weeks in November and December were spent in modifying and adjusting the door rollers, tracks, seals, and limit devices before tests showed they would operate reliably hot or cold. Air leakage from the radiator enclosure when the main blowers were operated proved to be excessive, both from the standpoint of coolant- cell ventilation capacity and because of excessive heating outside of the enclosure. Hoods were installed, into which the radiator doors re- tracted, and the sheet-metal enclosure was generally tightened and modi- fied before leakage became acceptable. (Even after the lmprovements it was necessary to supplement the cell exhaust with ducting to one annulus blower to attaln a negative pressure in the coolant cell.) When the main radiator blowers were operated with the radiator hot, ailr leaking from the top of the enclosure overheated electrical insula- tion in that area. It was necessary to Install ceramic insulation on leads on top of the radiator and revoute the leads to cooler locations. Ducting was also installed to redirect cooling air flow across the top of the enclosure where tne door hoods had blocked the original flow pat- terns. The radiator work lasted from early Novemper to mid-January, delay- ing the filling of the coolant system and the start of power operation. As socon as conbaloment testing was Tinished, the insbruments, con- trols, and equipment were given the checkoubs required prior to startup. The fuel system was then heated, and flush salt was clrculated for three days., oSamples of the flush salt taken at this time were analyzed for oxides by an improved, more reliable method. Results averaged lessg than 100 ppm, well below tolerable levels, (Evidently the measures taken o avoid oxygen contamination while the reactor vessel and fuel pump were open were effective.) Fuel salt was charged into the loop, and the reactor was taken crit- ical on Decenmber 20. Between then and January 18, when the ccolant loop was tilled, nuclear experiments were restricted to powers below 25 kw, Even so, useful measurements were made on {lux distributions in the thermal shield and beside the reactor vessel, control rod shadowing el- fects, and zero-power kinetics of the nuclear system. At the same time, numerous fuel-salt samples were analyzed, showing uranlum in excellent agreenent with expectations, low oxide concentration, and practically no corrosion. With the coolant system in operation, the power was raised to 100, 500, and then 1000 kw as heat was extracted at the radiator. Dynamics tests and heat balances were conducted at each power. On January 23, while the power was at 500 kw, the fuel pressure control valve (or its filter) showed signs of plugging, but the situvation cleared up in a Tew hours. There was also evidence of an abnormal restriction in the egqual- izing line between the fuel pump and the drain tanks. The next day the reactor was taken to 1L Mw, and a few hours later signs of intermittent plugging in the fuel off-gas line again appeared. Lt was established then that the equalizing line was completely plugged. When 1t was also discovered that the auxiliary vent line was almost completely plugged, the reactor was taken subcritical. Lfforts to blow oub the plugs in the equalizing line and the auwxiliary vent line failed, and the fuel and coolant loops were drained to shub down the reactor. (The syslens were kept hot, with helium circulating.) The check valve in the vent line to the auwxiliary charcoal bed was removed, and the line was recomnected. Gas then freely passed through the lines. The fuel drain cell was opened, and the flow-restricting capillary in the equalizing line was removed. This line then proved to be clear. The check valve and capillary were taken to the High-Radiation- Level BExamination Laboratory (HRLEL), where they were observed to contain small aocunts of intensely radiocactive material. They were not plugged, however. Meanwhile the plugging of the fuel pressure conbrol valve (or filter) recurred. Blowing backward throusgh the line did not clear the obstruc- tion, so the assembly of valve and filter was removed. A new valve was installed, and the assembly was replaced to see 1f it were open. (This was simpler than flow testing the contaminated parts outside the sys- tem.) At first the pressure drop was acceptable, but over a few hours, excegsive pressure drop again developed. After the assembly was removed, the new valve was proved to be c¢lear, so the filter was taken to the HRLFEL for examinaticn. The old pressure control valve was alsc cut open and examined in the HRLEL after flow tests showed its C was a factor of 10 3 below ils original value. There was some solid material in the pores cof the Tilter, but the pressure drop across the Tilter was not as high as 1t had apparently been while installed. Inspection showed that the valve trim was coated with oll and varnish-like material. The total amount of material in the valves and filter was quite small. Therefore we decided to resume opevations, using a Tllter with larger pores (50 p instead of 1 w) and controlling pressure with a hand valve (CV = 1.0) already in the line instead of the pressure control valve (CV = 0.07). (The pressure control valve was eliminated.) During the operation at significant power, activation of potassium in the cooling water corrosion inhibitor had been more than expected. While the reactor was shubt down, the original corrosion inhibitor was replaced with i nitrite and borate, which would give only a small amount of activity (britiom). Also while the reacltor was down, pressure drops across the charcoal beds were measured. 'There had been intermitient indications of excessive pressure drops during the L-Mw operation and just afterward, bub when measured before the next startup they were normal. Operators and ex- tension handles on the charcoal-bed hand valves had been repaired during this interval. (Ice, rust, and binding had prevented easy operation, and setscrews in two of the extensions had sheared.) Operation was resumed, and fuel was circulated for 44 hr before the power was raised to 1 Mw. Lt was necessary to gradually open the hand valve being used for pressure contbrol aboub l/4 turn during this period in order to hold the fuel pressure in the range of 4 to 6 psig. The implication was that a small amount (perhaps 0.01 in.) of material had built up on the valve trim. Within 3 hr after going to L Mw, the ap- parent accumulation of material accelerated. The valve was adjusted to maintain control, but a short time after makling an unusuvally large ad- Justment, the differential pressure across the charcoal bed and valves suddenly increased, indicaling hlockage in those lines. Jugt at this point, the motor on the space cooler in the fuel drain cell failed. Rising temperatures in this cell then required that the Tuel e drained and gecured s¢ the cell could be opened. The remainder of the period was spent in replacing the space cooler motor and Investigating the trouble in the off~gas system. Analysis of lxperiments Reactivity Balance The reactivity balance involves calculation of the effects of changes in all known variables affecting reactivity. The net effect of changes between any two times that the reactor is critical should be gzero if the calculations are acceuwrate., Deviations from zero indicate eifher error in the measurements and calculations or some unforseen effect. 11 The reference conditions Tor the reactivity balance were established in run 3 (the zero-power experiments which ended July 4, 1965). Between then and the startup in December, one variable changed significantly — the 227U concentration in the fuel. Because of uranium additions to the circulating fuel, at the end of run 3 the 233 concentration in the loop was apbout lO% higher than that ian the sgalt remsining in the drain tank. Thus the drain and mixing at the end of run 3 caused a substantial de- crease in 227U concentration, but one that could be caleulated. On the other hand, the fuel loop was flushed al the end of run 3 and the be- ginning of run 4, resulting in mixing into the fuel an amount of flush salt which was probably small but was not directly measurable. Before criticality in run 4 the 2357 conceuntration was calculated, assuming no dilution with flush salt. The change in concentration from rux 3 was converted intc reactivity change, using the concentration coefficlent of reactivity that had been cobserved in the zero~power experiments.2 The resultant reactivity was converted to a predicted control-rod con- figuration for criticality at 1200°F and zero power, using the control- rod polsoning representation based on the rod calibrations. (See "MSRE Reactor Analysis,” this report, for a discussion of the egquation for rod poisoning.) The predicted critical position of the regulating rod (rod 1) with the two shim rode withdrawn 34 in. was 24.7 in., and the observed critical position was 23.3 in. This difference, equivalent to only 0.08% Bk/k, may he due to several factors, lncluding (1) differences in temper- ature measurement, (2) inventory errors at the end of run 3, and (3) errors in the rod-poisoning calcwlabion. The most important conclusion to be drawn from this is that there was no significant dilution of the fuel salt by the flush-salt operation. Critical operation of the reactor during runs 4 and 5 provided an opportunity for checking out some parts of the reactivily-balance cal- culstion. 'The computer-executed reactivity balance was nct used be- cause (1) the accurate representation of control~rod poisoning had nob been incorporated into the program, (2) the calculation of xenon poison- ing was not fully developed, and (3) there were indications of other systematic errors in the program that had anot been resolved. However, several manual calculations were made at very low powers where the fig- sion product poisoning and fuel~burnup bterms are insignificant. Some calculations were also made at powers up to 1 Mw put without the xenon and. samarium terms. : Tn general, the zero-power reactivity balances showed a variation of +0.02% 8k/k. This corresponds to a variation of *0.3 in. in the critical position of the regulating rod or +3°F in system temgerature and probably represents the limit of accuracy of thig portion of the calewlation. Calculations after up to 24 hr of operaticn at 1 Mw showed no mea- surable change from the zero-power reactivity valances. Neglect of the fission product terms should have resulted in a change of about -0.2% &k/k if the anticipated xenon behavior had oceurred. This is much larger than the scatter in the measurements, so one can conclude that the xenon poigoning was actually much less than predicted. 12 The difference belween indicated and actual position of control rod 3, which developed after the start of run 4 (see p. 54), was reflected ag a shift of about 0.05% 6k/k in the zero-power reactivily balance. This demonstrated the abllity of the reactivity balance to show up small changes, at leastc under simple conditions, that is, no Tission product poisoning and low power. Power Calibration One cbjective of the early power tests was to correlate the thermal power of the reactor and the oubputs of the various neutron sensing ele- ments. The principal requirement of such a calibration is an independent method for measuring the thermal or fission power of the reactor. Vari- ous methods were used in the apprceach to 1 Mw with reasonably good agree- mexnt . The first approximation, used throuvghout run 3, was based on cal- culatiocns of the inherent (alphamn) source in the fuel and subecritical multiplication. "The primary purpose of this calibration was to permit positioning of the safety chambers so that they would provide a control- rod scrawm at a few kilowatis and still not interferce with the planned experiments at lower powers. Absolute accuracy was not required since none of the experiments lnvolved operation at powers where thermal ef- Tects were gpparent. A thermal-power calibration was made shortly after the start of run 4 at a nominal 25 kw. At this time, only the fuel loop wasg full and circulating; the thermal power was evaluated frowm the rate of tenpera- ture rige and the kncwn neat capacity of the loop. Thisg provided a basis for subsequent operation at higher powers and, incidentally, showed that the initial source-power calibration had been in error by only 8%. A gimilar test was performed al higher power with both the fusl and coolant loops full and circulating. However, the second test was di- vided in two parts. First, the nuclear power was set at about 75 kw, and 75 kw of electrical-heat input to the loop was turned off. These operations did not lead to a steady temperature, but the change in tem- perature slope witnh time before and after the increase to 75 kw was used to correct the thermal power. The nuclear power was then increased by 50 kw, to 125 kw, and the increase in the rate of temperature rise was measured. The thermal-power calibration from these two steps was within 5% of the 25-kw calibration. several overall heat balances were calculated with the nominal nu- clear power at 1 Mw. The results in run 4 varied between 0.94 and 1.18 M, with an average value of 1.0l Mw. In run 5, the steady-state heat balances at an indicated power of 1 Mw gave 1,19 = 0.02 Myw. The reagons for this apparent discrepancy have not been established. Filux Measurements Flux measurements were made in the neutron source tube to determine 1T operation at power would regenerate the external neutron source (Am- 13 Cm-Be) to a significant degree. The re%eneration would take place as a result of thermal-neutron captures 1n e lAm‘(Tl/g = 462 years), which would produce the more active 2420 (Tl/g = 162.5 days). The measure- ments were made using cadmium-covered and bare copper folls to determine the thermal-neutron flux at the permanent source position, an elevation of 830 £t & in. The thermal flux at that point was found to be 3.8 X 107 neutrons cm? sec”* with the reactor at 1.02 kw. This extrapolates to 3.7 X 10*1 neutrons cm™® sec™t at a power of 10 Mw. Using conservative assumptions, it was calculated that the source strength will be kept at an acceptable level for at least 1-1/2 yvears after power operation has begun. Flux measurements were also made in the reactor furnace (™~ in. from the reactor vessel wall) in the location later occupied by metallurgical surveillance specimens. The measurements were made using gold and copper foils. The foils were sealed in a stainless steel envelope from which alr had been purged to protect the copper from oxidation during the irra- diation at 1200°F. Fluxes were measured along a vertical line 80 in. long. The peak flux (at approximately the core midplane), extrapolated to a power of 10 Mw, was 7 X 1012 neutrons cm™? gec”t. Fast and thermal contribubions to this flux were 4.8 X 10Y® neutrons cm™? sec™ and 2.3 X 10%2 neutrons cm™? sec™t respectively. MSRE Dynamic Tests Description of Dynamic Tests A number of dynamic tests were performed during operaltion at zero power and subseguently at power levels up to 1 Mw. The purpose of these tests was to investigate the inherent stability of the sgystem, as well a5 1o evaluate the mathematical models and the paramseters that were used to predict its dynamic characteristics. The inherent stability of the system (i.e., without automatic con- trol) was the subject of an exhaustive theoretical analysisB in which 1T was concluded that the reactor would be stable at all powers and that its stability characteristics would improve as the power level increased. Teste made on the reactor operating without servo conitrol at power levels of 75 kw, 465 kw, and 1 Mw confirmed these predictions, as shown in Fig. 1.2, which shows the response of the power to arbltrary perturbations. The improvement in the anatural damping characteristics and the decrease in the period of oscillation with increasing power level are clearly shown. Figure 1.3 shows a comparison of the experimental and theoratical veriods of oscillation. A subject of considerable interest during the MoRE development was whether or not the reactor power would "wallow" when operating without servo control. Wallowing would occur if the system were unstable or il random perturbations in reactivity or ccoling load induced lightly damped 14 power oscillations. At both 75 and 465 kw some wallowing was observed, but the power level perturbations were less than ifi%. In voth cases neither of the main radiator blowers was on, so the bulk of the cooling load was due to natural convection and radiation, and random fluctuations in load were secen. When one radiator main blower was Lturned on for the ORNL-DWG 66~-47861 7O o P —e - POWER 22 75 kw ‘ \ 1 ‘ £ 60 ff —F | o x5 | | - i“‘mfi \J 5> , —~T J’»i:fi 5 ¢ T ‘ o = | i = 1 g e ) e . = i ,,,,,,,,,,,, | B P ‘ I = | 5 o 2 - — . R Ll oo | g0 | | - ‘ | e T = 20 b - e — b e P i . _.l_ . 2 POWER = 465 kw \ i | | i ! i ! 0 ‘ L i : — ; . 80 — 71— e ey - — T T e e 2 POWER = t Mw | ; o 5 70 b I . _T‘_ - C e e L - Y L /‘HMA‘ -------------------------- s ‘ i D | i £z 60t ] g b — & . | i ! — 5 - e — — e a1 o 400 800 1200 1600 2000 2400 2800 3200 3600 TIME (sec) 1..2. TInherent Response of MSRE at Three Low Power Levels. OQRNL-DWG 66—-4782 100 ooy 4 - e 50 - SobSNdx e e e = IS Z 20 = a X—EXPERIMENTAL = 10 X THEORETICAL ..l 1l (6] - . o beeed 4y vy ™ Lyt — —k ‘:_ n >0 5”//\/ r 2 8)'(/:( SAMPLING INTERVALS: INDIRECT ANALYSIS 1 sec . DIRECT ANALYSIS -0.25 sec 103} Q.004 0.002 Q.005 0.01 0.02 0.05 o 0.2 0.5 1.0 FREQUENCY {rodinns/seq) Fig. 1.12. Freguency Response - Magnitude Ratio of (bN/No)/ ok/ko) for Ng = 1.0 My, 27 ORNL-DWG 66-4772 TR T fi—'—'r_'“%—*—’“—v—‘ . R o ‘| - CASE SRB-{Mw—*H . 54 - bit SEQUENCE — INDIRECT ANALYSIS e ~ 5{4—bit SEQUENCE — DIRECT ANALYSIS a . CASE SRB-4Mw-12 20 : 127-bit SEQUENCE — DIRECT ANALYSIS a @ ; ; | SAMPLING INTERVALS : INDIRECT ANALYS! 2 30 \— s e ’ l_4seCL S w L SAMPLING INTERVALS: DIRECT ANALYSIS w20 b — S — < : ; —0.25 sec &40 e — oo - - - 5 ' _—~THEDRETICAL — L } e R S0 e e o b — i ! | o ‘ —20 o — I — e e ——— .< — o TT \‘ e .- - I . ! Vo ‘ } -3 e b T © S firfifiéfif | “?: A0 e | — a«fififiq; g%& e e e P Tl M el C.001 0.002 0.005 0.0t 0.02 0.05 0.4 0.2 0.5 1.0 FREQUENCY (radians/sec) Fig. 1.13. Frequency Response - Phase of (6N/N0)/(6k/ko) for Ng = 1.0 Mw. parameters appear as a quantity No@f/(MCp)f. It was found that excellent agreement helween theoretical and experimental results may be obtained at all frequencies and all three power levels 1if, in Lhe theoretical calcu-~ Jations, the increased effective core volume 1s used and the guantity Noaf/(MCp)f is inecreased by a factor of 1.5 to 2. The plausibility of such changes 15 now belng examined. Temperature Response Test In order to get a more accurate estimate of the fuel temperature coefficient of reactivity, the hot-slug transient test as reported earlier’ was modified and rerun in January 1966. In this test the re- actor power level was controlled by the flux servo at 100 w with the fuel loop circulating and the cooclant loop stagnant. After heating the stagnant coolant salt about 20°F hotter than the fuel sali, the coolant pump was started, producing a2 hot slug of fuel in the heat exchanger which was subsequently introduced intc the core. 1In this test, then, the change in reactivity was due entirely to the change in temperature, and the rod-induced reactivity required to keep the reactor critical was approximately equal and opposite to that due to the Tuel and graphite temperature change. ©Since the graphite responds slowly to changes in fuel temperature, the immediate effects on reactivity can be attributed to the fuel temperature coefficient alone. Hurther corrections have yet 23 ORNL-DWG 66-4773 9-REGION CORE MODEL bl el + ® EXPERIMENTAL RESULTS FROM A 1 MOT-SLUG TEST USING SAMULON'S \ ‘ 02 | METHOD; SAMPLING INTERVAL = A L 0.25 sec ‘ 0.4 ’ 0.001 0.002 Q.005 0.0 0.02 0.05 A 0.2 0.5 1.0 FREQUENCY ({radians /sec) Fig. 1.14. TFrequency Response — Magnitude Ratio of Reactor Outlet Temperature Response to Inlet Temperature Perturbations. to e applied to the results of the test, but a preliminary estimate of ~4.7 % 10° (°F)1 has been obtained. 'This compares Tavorably with the predicted value of —4.84 x 1077 (°F7)~! used in the theoretical calcu- lations. The transfer function of reactor outlet temperature respconse to changes in reactor inlet lLemperature was also calculated from this test, and the plot of magnitude ratio vs frequency is shown in Fig. 1l.l4. DNote that the experimental curve indicates much greater attenuation over most of the freguency range shown. This is probably due to higher-than-ex- pected rateg of heat transfer to the graphite, since experimental mixing studies’® indicated that there is very little attenuation due to mixing in the frequency range below O.1 radian/sec. Reasons for this discrepancy are being studied. oystems Performance Off-Gas System Some difficulty with plugging and partial restrictions in the MSRE off-gas systems has been encountered at various times in the operation of the reactor systems. In the past these obstructions have developed at in-line Tilters and pressure-control valves with extremely small flow passages. During early operations, the plugging of filters was attributed to having filters with inadeduate Tlow areas to accommocdate the material 1o be removed from the gas gLireams. (For example, the original filter in the coolant off-gas line, 528, was a tubular element only 1 in. long by 1/2 in. in diameter.) The plugging of valves appeared Lo be a re- sult of filters which were too coarse to remove all the particles that were capable of causing plugs. During shutdown which followed the zero-pewer experiments (zun 3), larger filters, capable of removing pvarticles greater than 1 p in di- ameter, were installed ahead of the pressure~control valves in both the 24 fuel and coolant off-gas lines. The fuel-system pressure-~control valve (Pcv 522), which had exhibited some ervatic behavior in run 3, was tested and found to have the same pressure-drop characteristics that it had when first installed. This valve was reinstalled for the next operation of the reactor. No further difficulties were encountered with the coolant of f-gas system, bul several problems developed in the fuel system. (See Fig. 1.15 for a simplified flowsheet of the fuel off-gas system.) The performance of the fuel off-gas system was satisfaclory during the suberitical and low-power (up to 25 k) operation of the reactor in December 1965 and January 1966. The first indication of difficulty in this run (run 4) occurred on Janvary 23 with the reactor power at 500 kw. The total integrated power accumulated up to this time was about 3 Mwhr. At that time the system pressure began Lo rise slowly, indicating either a restriction in the vieinity of PCV 522 or a malfunction of the valve. (A routine check of system pressure interlocks on the previous day had indicated normal behavior of all components.) shortly after the start of this pressure transient, the response of the drain-tank pressures indicated an abnecrmal restriction at a capillary flow restrictor in line 521, the fuel-loop-to-drain-tank pressure equalizing line. The excess pressure in the fuel system was relieved by venting gas through HCV 533 to the auxiliary charcoal bed until the restriction at PCV 522 was cleared by operating tne valve several Gtimes through its full stroke. These opera- ORNL-DWG €£6-2422, FE 524 N B e Cv 528 V 524B FROM COOLANT Eo- b SYSTEM AND 80TH OIL SYSTEMS GAS HOLDUP AMD COOLER Pd| e 170 liters 556 T PCY — 3 o2 V5228 VOLUME (ki ¥ HOLDUP AN 185 fters v 5188 % SAMPLE STATION A . HOV 533 e D—ba—T VAIEE v 5188 VAIBCH[ Ty 51ap v 622 CV533 (T . V51852 ~ vsace FE 521 (T e V 51883 518C Y o ¥ V518G HCY HCY HC IS & mem ve20 545 546 544 r : V 5384 YV 561 AUXILIARY CHARCOAL BED IACB) V5624 FD-1 FROM SAMPLER ENRICHER Fig. 1.15. BSimplified Flowsheet of MSRE Fuel Off-Gas Sysbem. 25 tiong slso revealed at least a partial resgtriction in the line Lo the auxiliary charceoal bed, apparently at HCV 533, The reactor was made sub- critical after 4-1/2 hr operation at 500 kw. Satisfactory pressure control was maintalned for sbout 24 hr, but the plugging at HCV 522 recurred shortly after the reactor power was raiged to 1 Mw, Again, the operation of this valve fhrougn its full gtroke was effective in reliceving at least part of the restriction. However, oun this occasion, evidence of restrictions at the charcoal-bed inlets began to appear. These restrictions were bypassed by putting the two spare charccal-bed sections (24 and 2B) in service and cloging the inlet valves to the two that were restricted (1A and 1B). Within sbout 6 hr the inlets of beds 28 and 2B also plugged. When the pre- vicusly plugged beds were checked, at least cne was found to be clear and off-gas flow was reestablished through bed 1B. The manipwlation of the charcoal-bed inlet valves revealed substantial mechanical difficul- tles with the manual operators. These were caused by misalignment of extension handles and corrosion due to weabher exposure and resulted in two of the valves becoming inoperative. The combination of difficulbies in the off-gas system rezulted in a reactor shutdown to correct the conditions. The restrictions in the equalizing line (521) and the suxiliary vent line (533) were relieved. by removing the capillary (FE 521) and a check valve (CV 533) respec- tively. The valve and filfer assembly, and a second valve that had been tried vriefly, were removed from line 522. 'The original valve and the filter were gubsequently subjected to detailed examinations which are deseribed in "Component Development,” this report. Pressure-drop meas- urements showed a higher pressure drop by a factor of 3 Lo 4 in the valve and a factor of 10 in the filter. However, the observed restric~ tions, particularly in the case of the Tilter, were not high enough to account for the observed pressure behavior in the operabing system. While the reactor was shut down, several measurements were made of the pressure drop across individual sections of the charcoal beds. Measurements made shortly after the shutdown showed essentially normal pressure drops for three of the four sectionsg. The pressure drop across the fourth section (1B) was about 60% higher than thal scross the others. (These meagurements include the pressure drops across the inlet and oub- let valves of each bed section.) Although nothing was done to correct this condition, measurements made Jjust before the next startup (run 5) showed that all four sections had essentially identical, normal pressure drops. Mechanical repairs had been made Lo the operating mechanisms of the valves, but these did not affect the flow characteristics. Measure- ments were also made of the pressure drop across the auxiliiary charcoal bed; no abnormalities were observed there. The preparations for the next startup included installation of a large, relatively coarse (50-u) filter and a large, open hand valve at the PCV 522 location. This arrangement eliminated the swall, easily plugged passages associated with PCV 3522, and 1t appeared that the filter would still remove any particles large enocugh to plug valve 5228, which was bto be used for system pressure ccentrol. (Satisfactory manual pres- sure control using valve 522B had been demonstrated during the shutdown. ) 26 In the next operation of the fuel loop (run 5), fuel salt was cir- culated for 44 hr with the reactor suberitical or at very low power (100 w) vefore the power was raised. The evidence is not conclusive, but there was some indicatlion ol an dncrease in flow resistance at wvalve 5228 during this time. ©Shortly after the power was raised to 1 Mw, the increase in flow resistance at 522B accelerated sharply. Freqguent adjustment of the valve was required to maintaln reasonable pressure control. The restriction at the charcoal beds which occurred after one such adjustment developed quite rapidly and resulted in almost com- plete plugging of the two beds that were in service (1A and 1B). After the reactor drain, which was reguirved because of the failure of the space-cooler motor, it was debermined thal the auxiliary charcoal bed was also restricted. This bed comes into service automaltically during a drain as a backup to rellieve any excess gas pressure. The helium in the fuel system was then released through the two standby charcoal beds. Detailed investigatlions after the shutdown showed that the charcoal- bed restrictions were probably at the inlet valves. Valve 621, the in- let valve to bed 1B, was removed for examination in the HRIEL. The re- sults of these examinatlions are described in the section entitled "Com- ponent Development.' The pressure drop across bed 1B with valve 621 in place was very much higher than normal. With the valve oul, the pres- sure drop was nmucen lower bub still higher than for a normal, unrestricted ted. An excess of helium was then forced through the bed for a few min- utes, and the pressure drop suddenly decreased. BSubsequent pressure-drop measurements under controlled conditions gave values which were normal for an unrestricted bed. everal gas samples were taken from the fuel loop during this shut- down 1n an effort to identify some of the plugging constituents. For several days after the drain, all helium flow through the fuel system was stopped, and the hellium in the loop was circulated at ~1100°F. The helium in the off-gas holdup volume (170 liters) was stagnant at about 122°F. Three successive pairs of samples were then trapped from the discharge of the holdup volume; each pair of sample traps consisted of a pipe coil at liquid-nitrogen temperature followed by a U-tube filled with molecular sieve material, also at liguid-nitrogen temperature. ‘The first pair of samples was isolated from the gas that had been in the hold- up volume by displacing it through the traps with helium from the fuel loop. The helium which entered the holdup volume during this displace- ment was presumed to be representative of the material circulating in the loop. A second displacement through anobther pair of traps was used to isolate material from that gas. For the third pair of samples, the re- actor cell temperature was allowed Lo rise to heat the off-gas holdup volume to 173°F before the gas was passed through the traps. The pur- pose of this was to see 1T any additional material was released from the holdup volume by the increase in temperature. Analysis of the samples has not yet been completed because of the high activity associated with thiem. bince particulate matter originating al the fuel pump was regarded as one possible cause of plugging in the off-gas system, the main fuel off-zas line (522) was opened at the fuel puwp for viswal inspection. cmall amounts of gray-~green powdery material were found between the faces of the flanges thal were opened, but no deposits were seen in the part of the holdup volume that could ve inspected. btamples of the solid ma- terial and swabs of the holdup volume were removed Tor analysis. Additional work to be performed before the next startup includes: 1. a blowout of the holdup volume Lo remove any loose material, 2. design, fabrication, and installatlion of a new fillbter-control-valve assembly at PCV 522, and 3. replacenent of several valves 1in the system with valves having larger trim, Salt-Pump 01l Systems Botrained gas in the circulating olil tends to accumulate in the volute of the standby oll pump. From the beginning of MSRE operation, it was necessary Lo prime the gtandby pump frequently o keep it ready Tor operation. (Priming wags done routlinely once a shift.) Observations in a development loop showed that entrainment and frobthing in the res- ervoilr were due to the action of a Jjel pump used to scavenge oll from the bearing housing (uging shield plug oil flow to drive the Jet) and agl- tation of oll by the shaft and.bearings.6 In Beptember 1965, the jets on the fuel and coolant pumps were replaced with less powerful jJets to decrease entralnment. After this modificaticn, it was possgible to re- duce the frequency of priming the standby pump to once a week in the fuel o0il system; on the coolant lube oll system it was necessary Lo con- tinuve as before. After the Jets were changed 1t was found that changes in Tuel~pump bearing~-oll flow or reservoir pressure caused changes in reservoir level, apparently because of holdup of oil in the salt-pump motor cavity. OILL from the bearing housing can enter the mobor cavity eltner by leakage past the upper seal or through a passage comecting the motor cavity to the area beltween the two shaft bearings. Apparently an excess of flow to the bearing housing causesz oil to be forced up Into the mobor cavity. Although the breasther line affords a return to the olil reservoir, it has bveen found that once the oil collection beglins in the motor cavity 1T msually conbinues until elther the oil zystem is shut down completely or the bearing Tlow is reduced to below normal. Apn increase in oll {low to the bearing housing from the normal 4 gpm to 5 gpm causes a level de- crease in the oll reservoir of over 2 gal. (Much more than this could collect without reaching the motor, but normally action is taken to stop collection at less than a gallon.) Leakage pasl the lower seal drains to the oil catch tank, which is equipped with a level indicator and which is perlodically drained to the waste oll recelver. After the installation and calibration of a siphon tube in the fuel-pump oil catch tank in Novenber, oll was not added to bring the level indicator on scale. The fuel-pump lower-seal leak rate during the months of December and January is therefore nobt known pre- cigely. However, when oll was deliberately added to the oil cabtceh tank 28 in February, it took 900 cc to briang the level instrument on scale. Tae seal leakage evidently had been very low, because this volume is very near that calculated to fill the empty pipe from the oil catceh Lank to the drain valve at the waste oll receiver. After {the oil catch tank level indicator was brought on scale, the indicasted leak rate was approximately 2 cc/day. Phe coolant-salt-pump Llower-seal leakasge measured during Januvary and most of February was zero. Near the end of February a leak rate of ap- proximately 4 cc/day began. The levels in tne oll reservoirs are recorded roubinely once a shift. As was seen above, the amounts of oil held up in the motor cavities vary with operating conditions, and these cause variations in the olil reservoirs. DBut there has been no reason why this variable holdup should show any long-term trend. Therefore, the reservolr levels recorded over a period of weeks should reveal changes in oil inventories. Figure 1.16 is a plot of the fuel-pump o0il reservoir levels from December 1965 through bFebruary 1966, when the oil system was in continuous opera- tion. Lt is apparent that leakage was less than can be measured by this ethod. The coclant-pump oil reservolr level showed similar behavior except that the fluctuations were smaller (prdbably reflecting less motor- cavity holdup). Treated Cooling-Water System The treated-water system 1s a closed systenm removing heat from the thermal shield and other items in the reactor and drain cells. Problems ORNL - DWG 66 -3524 o Y —— . | ; | | 1. a5 | e { liter ] U% = 149 liter) _ : e s B 5 @% 2 o9 ; %o sp ©PHEPDE ® U>J @@@ b 80 e ——— - s 0000 000F 1 i ® e 232 epper o396 ersne | @ s 9 |@ ‘ ® , o s ¥ ® @ ose & 23920035950 2L O o e | ® ’ |S e | ? [™~DRAIN e | R , | MOTOR | j | cavTy ! = ] sol --------- Ao b ‘ , e | 10 20 30, 10 20 30 10 20 l | DEC | JAN | FEB | Fig. 1.l6. Fuel-Pump Oil~Reservoir Ievel in Runs 4 and 5. 29 encountered during this report periocd were neutron activation of the cor- rosion inhibitor and interruptions of flow through the thermsal shield slides. Activation. During power operation in run 4, induced activity in the treated water rose to an unexpectedly high level. Extrapolation of observed radiation levels indicated that at 10 Mw, gamma radiation near the heat exchanger in the diesel room would be about 400 mr/hr; gimilar levels would exist in the water room. The activity proved to be 12.4-hr 42K produced in the corrosion inhivitor (2000 ppn of a 75% potassium ni- trite, 25% potagsium tetraborate mixtuce). The level indicated that the average lux in the system.(about 80% of which is included in the thermal shield) was equivalent to sbout 7 X 10%9 thermal neutrons cm™? sec™t at 10 M. A survey of possible cation replacements for potassium led to the choice of Lithium, highly enriched in the 71i isotope to minimize tritium production. An adequate amouvnt of lithium nitrite solution was prepared cormercially by ion exchange from potassium nitrite and lithium hydroxide. After the 4000-gal treated-water system was diluted with demineralized water to reduce the potassiwm from 800 to 3 ppm, the desired inhivitor concentration was attained by adding i nitrite, boric acid, and 14 hydroxide. When the reactor was next operated at power, in run 5, objecticnable activation again occurred, this time due to 1.0 ppm of sodium which had evidently come in with the demineralized water from the ORNL facility. Condensate wag produced at the MSRE with less than 0.1 ppm of scdium, and this was used to dilubte the sodium in the treated~water system to 0.3 ppm. The concentrations of sodium and 511 are now low enough so that shielding and zero-leakage containment of the water system will not be necessary. Flow Interruptions. On several cccasions flow through the thermal shield slides stopped. Each time it was found that when the water supply line was disconnected and water was allowed to flow backward through the slides, considerable amounts of alr came oub with the water. The con- clusion was that air could accumulate in the longest slide until the nead exceeded the pressure available to cause flow. (This should be im- posgible if the piping in the slide were installed as designed.) Vent- ing the inlet line restored the flow, and provislons were made Lo fa- cilitate venting. The source of the air was evidently a vorbtex in the surge tank which occurred when the level was low. After the flow through the surge tank was drastically reduced (it had been about 30 gpm) there was no recurrence of the flow interruptions. Secondary Contalinment The secondary contalnment, that is, the envelope surrounding the reactor and drain-tank cells, was tested and put iuto service for the first time during this report periocd. The testing procedure followed generally the MSRE Operating Procedures, Sect. 4E.7 This procedure in- 30 cludes testing of individual closure devices on lines penetralbing the of thne overall leakage rate. All lines carrying nhelium, air, or water into the cells are eguipped elther with check valves or block valves actuated by radiation monitors. All of these valves were tested, in place wherever possible; obherwise they were removed and bench-tested. "Yest pressures were 20 psig or more, and leakage rates were reguired Lo meet conservalive criteria. This work wag done concurrently with maintenance before the cells were closed., Leaks which occurred in the cooling-water system were mostly in the cneck valves. These were the resuwlt of Toreign parcicles, apparently washed out of the system and trapped between the sliding and moving parts of the valves. Two hard-seated valves in the treated-waler system had to be replaced even thiough the scats were lapped in an effort to get them to seal. One was a check valve with a swinging check, in g wabter line between the surge tank and condensate tank, and the other was z spring- loaded hand valve on top of the surge tank, which has to contain air. They were replaced with soft-seated valves. The leakage rates on water- contalning valves were determined by collecting the leakage, and a flow- meter wag uvsed to measure air leakage through valves on tup of the surge tank. The majority of the valves in the helium system of bobth primary and secondary containment had satisfactory leak rates. Those which had ex- cegssive leakage were check valves and were found to have damaged O-rings and/or foreign particles, usually metal chips from machining, in them. AfTter cleaning and lnstalling new O-rings, they were satisfactory. Many of the check valves had to be removed from the system to pressurize them. After reinstalling, the lines were pressurized and the fittings were checked with a helium leak detector. Leakage rates througn the valves vwere determined by a flowmeter or by displacement of water in a calibrated tube. The inshrument-air-line block valves with Tew exceptions were found to be satigfactory; however, nunerous tube fitlings were found to be leaking by checking with leak-detector solution when the lines were pressurized to check the valves. Several quick disconnects on alr lines inside the reactor cell and drain-tank cecll were found to be leaking when checked with leak-detector soluticon and were repaired. These leaks do not constilitute a leak in secondary contalnment, since each line has a block valve coubside the cell; bull a leak here doss affect the cell leak rate when air pressure 1s on the line to operate the valve. The butterfly valves in the 30-in. line used for ventilating the cells during maintenance operations were {irst checked by pressurizing between them. The leakage measured by a flowmeter was excesslve, and the valves had to be removed from the system to determine the cause. There was considerable dirt on the rubber seats, and one was cub; these seats were cleaned and repaired. We found that the motor drive units would slip on their mounting plates by a small amount, thus causing a slight error in the indicated position of the valve. Dowelg were in- stalled in the mounting plates to prevenlt this. Small leaks were also 31 found around the pins which fasten the butterfly to the operating shaft; these leaks were repalired with epoxy resin., The line from the thermal- shield rupture disks to the vapor-condensing systems has numerous threaded joints that leaked badly when pressurized with nitrogen. Iach Joint was broken, the threads were coated with epoxy resin, and the joint was re- made. All joints were leaktlight when rechecked with leak-detector solu- tion. When closing the cells, all membrane welds were dye checked. After conpletion of the membrane welding, alternate top blocks were installed, and the cells were pressurized to 1 psig to leak check all membrane welds with leak-detector solution. DNo leaks were found in the welds. The re- actor access cover plate, which has a double O-ring seal, was found to e tight by pressurizing between the O-rings. With the cells at 1 psig, all penetrations, pipe Joints, tube fit- tings, and mineral-insulated (ML) electrical cable seals subjected to this pressure were checked with leak-detector solution. Numerous leaks were found ia tube £ittings and ML cable seals, and the leak rate was about 4500 ft3/day, indicating a major leak. Many of the small leaks were shopped by simply tightening the threaded parts of the seal. How- ever, the leak rate was still aboub 4500 £t3/day. The cells were then pressurized to 5 pslg, and leak hunting con- tinued. Three large leaks were located: one in the sleeve of line 522, the off-gas line from the pump bowl to the carbon beds, under the benb house floor, another in an instrument air line to valve HCV 523, and anobher from the vapor-condensing system to the steam domes and oubt to the north equipment service arca through a line that was temporarily open. All penetrations, tube fittings, and ML cable seals were again checked with leak-detector solution. Many ML cable seals which had not leaked at 1 psig were found to be leaking, and some of bthose which had been tightened and sealed at 1 psig now leaked. The hock gege being used to monitor the leakage rate along with conventional pressure gages did not work properly, and the trouble was found to be the resull of extremely swall leaks in the tube fittings on the reference volume side of the system. This and temperature changes in the coolant drain-tank cell and special equipment room caused asppre- ciable difficulty in determining the leakage rate. The piping to the hock gage was evenbually replaced with virtually all welded tubing. Temporary closures were put on the special equipment room, and the door entering the coolant drain-tank cell was kept closed as much a8 possible. These measures considerably improved the reliabllity of the data. The MI cable seals as a group accounted for a large percentage of the remaining leaks. Many were sealed by tightening, but many of them required tightening more than once. As a result, several of the gland nubs split and soldering was required. All large leaks were sealed or greatly reduced, and many of the small leaks were stopped. To stop these and other small leaks which may not have been located, all ML cable seals were coated with epoxy resin at the seals oubtside of the cells. Teflon tape, used extensively on ML cable seals, other threaded pipe, and tube fittings, did not perform gatisfactorily in providing a gas seal., 32 After stopping or reducing the leaks in the ML cable seals and in- strument air lines, testing was begun at 10, 20, and 30 psig. At these hignher pressures, large leaks were located in bobh component coolant pump dome flanges by leak-detector solution, and onc of them was audible. To correct these leaks, the width of the gaskel was reduced to increase the loading pressure on the gasket. All penetrations, ML cable seals, tube fittings, and external parts of valves which are part of the secondary containment were checked wilh leak-detector solution at each of the above pressures. In each of these tests, the containment system was pressurized to a specific pressure. Then all gas additions and controlled exhausts were stopped, and the leak rate was measured by the change in pressure corrected for changes in temperature. The calculated leakage rates are presented in Fig. 1.17. ORNL-DWG 66-4753 BOO [ e e — 400 S 300 - — — r— ] ~ / D o 2 ud |_ z A _ SECTIGN THROUGH S REACTOR SHIZLD ™. NUCLEAR INSTRUMENT | . 7. PENETRATION . A \ C by \,\ 10—2 \ Iy ’\1 // > - ; !; A w ; \ ‘ = ' . i = | \ i . GUIDE TUBE = ; .\l " NO. 6 2 5 = Y oo O i T\ - e m e o ‘ - Y i REACTOR i | \‘ . A CONTAINMENT Tt . | \ Y MIDPLANE |+ Bia /i CELL 3 i c OF CORE 4 L & ! ‘ ) \\ i 158 |’ = ; ; \‘ ‘ . e oS CURVE "B" COUNT \ : } :‘I‘ o RATE WITH CADYIUM ‘. ‘ }eH ! SHIELD \ : L , . { ¢ e | .) REACTOR o 101 | CURVE "A"COUNT RATE - \ VESSEL ; {10} ) L OWITHOUT CADMIUM ot et e SHIELD | ! i | : !) 2 1 P | Lo CADMIUM SH:ELDING ~ {5} (4) ol . o SUBASSEMBLY — TYPICAL N e s o 20 40 60 80 IN GUIDE TLUBES ~THIS LENGTH 100% . x, DISTANCE WITHORAWN (in.) 8 AND 9 CADMIUM WRAP ~— T . 10f1-Qin 1011-0in SECTIOM A-A o Y LOCATION OF T L GUIDE TUBES Eh 1B R A 1O “GUTER SHELL REMOVED FOR CLARITY--SHADED WEDGES ARE CADMIUM —UNSHADED WEDGES ARE ALUMINUM Fig. 1.20. Guide Tube Shield in the MSRE Instrument Penetration. ever-increasing shielding from side-cntry neutrons as the counber is with- drawn. ‘The bases of these long narrow wedges occupy the full periphery of the shield tube at its upper end and thereby provide a smooth tran- sition to the 100% wraparound cadmium sleeve which Tollows. This shield- ing, Fig. 1.20, was an eminently successful answer to the problem; curve B, normalized count rate vs distance, wag oblained after this shield was installed in guide tube 6. Bl Confidence Instrumentation Because of very unfavorable geometry the strongest practicable neu- tron source would not produce 2 COuntb/SPC Trom the fission counters in the wide-range counting channels until the core vessel was approximately nalf full of fuel salt; neither would it produce 2 counts/sec with flush salt in the core at any level. This i1s the minimum count rate required to obtain the permissive "confidence" interlock which allows filling the core vessel and withdrawing the rods. Therefore a counting channel using a sensitive BF3 counter was added to establish "confidence" when the core vessel 1s less than half filled with fuel salt. With the revised system, a Tuel-sall f1ll may begin when the reactor vessel 1s empty if BFs count 45 ORNL-DOWG 862446 ELECTRONIC RESEARCH ASSOCIATES 2.5/10UC CHAMBER | pECRDE POWER SCALER SUPPLY Q-1743-C1 BF; COUNTER REUTERS-STOKES RSN-28A PULSE | AMPLIFIER 3t ] RCI3-10-59 RC13-10-57 LOG COUNT |_ = * = L RATE METER e e e RCA3440-58 7 [ TRIP COMP] TRIP COMP LCR>10cps LCR=10% cps L CAL USE|+ conso e e o - PRI e e RCI3-10-56 | IO BF CONFIDENCE CIRCUIT Fig. 1.21. Block Diagram of BF3 Counting Channel. rate "confidence" is established but cannot be continued after the re- actor vessel is half full unless "confidence’ is established with either of the two wide-range counting channels., When f1lling the reactor vessel with flush salt, rods may be withdrawn and the £ill allowed to proceed at any level if BF; count rate "confidence” is established and if the drain- tank selector switch is in the fuel flush tank (FFT) position. Selection of the ¥FY position bypasses the half-Tull weight interlock and requires administrative approval. Figure 1.21 is a block dlagram of this BFa count- ing channel. Personnel Monitoring oSystem The reactor building radiation and contamination warning system was revised to correct some deficlencies and to Improve its effective- ness as indicated by tests and experlence during reachtor power operation. An alarm relay was installed in the personnel and stack monibtor alarm system circultry to provide an ammunciation upon loss of the 24-v de power supply. Monitron RE-701Z2, located in the south end of the high bay, was moved east about 20 £t to monitor Tor possible radiation escaping from the nuclear instrument penetration. CAM RE-T700L was moved closer to the sampler-enricher to monitor its operation. Several of the ¢-2091 beta-gamma monitors were relocated to provide move protection, One unit was installed in the instrument shop, one wiit was removed from the health physics office and installed in the hall between Buildings 7503 and 7509, and. the unit in the vent house was relocated to reduce its background. response from lines in The vent house. Two sdditional alr horns were added to increase the area covered by the horn evacuation signal. One norn was installed on the southeast corner of Building 7509 and the other in the diesel houge. Four additicnal beacon alarm lights were installed in areas where the horn might not be heard. A light was installed in the vent house, coolling~water equipment room, bthe switch house, and the Plant 46 and Eguipment shop building. A Q-2277 rate meter and Q-2101 alpha probe will be installed in the hot change house. Control Instrumentation Containment. To maintain the integrity of the reactor secondary containment, solenoid block valves with safety-grade wiring were In- stalled in the fuel-dralin-tank steam dome drain plping. These valves are interlocked to close when reactor cell alr activity or pressure 1is above limits. 'The interlocks override a manuval switch used for normal operation of the steam dome drain system. Actuating signals were ob- tained from existing circuits. Safety-grade control circuits were installed to operate four weld- sealed solenoid valves serving the fuel off-gas sample system. The valves arec speeially designed and constructed Tor use on MSRE contain- ment systems and are identical to those previously purchased for use in the fuel-pump-bowl bubbler level system. The procurement of four addi- tlonal valves from a commercial source is nearing couwpleblion. Fuel Sampler-Enricher. The fuel sampler-enricher safety-grade con- trol circuits were revised to increasce the relisbilitiy of the two con- talmnent barriers between the primary system~(fue1 pump bowl) and the operator. Originally, the circults were designed so that the sample access port could be opened if either the operational valve or the maintenance valve was closed. With Uhis arrangement a condition existed wherein a single failure could permit the access port to be opened when both The operatiocnal valve and the maintenance valve were open. 1f this had cccurred when the manipulator cover was removed, the manipulstor boot could have become the only containment barrier belween the operator and the fuel pump bowl. ©Since the boot could he ruptured by system pressures in excess of 10 psig, this condition was considered to be hazardous. The circuits are now designed so that both the maintenance valve and the operational valve must be closed before the access port is opened. Also, the removal valve and access port must be closed and the manipulator cover must be in place before either the coperational or the maintenance valve can be opened. The positlon of the cover is detected by a newly installed vacuum pressure switch. Additional protection for the manipu- lator oot has also been provided by a new circuit that prevents develop- ment of excessive differential pressure across the hoob during sampler evacuation operation. This protection is accomplished by closing a valve in an exhaust lioe to the vacuum pump when the differential pressure across the boot exceeds 30 in. (water colimm). Fuel and Coolant Pumps. To satisfy established operating criteris, the fuel and coolant salt circulating purp control. elircuits were revised as fTollows: 1. In both the fuel and coolant salt civculabing pump circuits, the existliong pump bowl level swilech actuvation valve was changed to re- duce the level at which the pump is stopped, and one new switch was installed to prevent pump startup until normal Ti11 level is reached. Since the salt level drops 8 To 12% after pump startup, the single 47 level switch system previously used did not leave enough operating margin to prevent normal level [luctuations from stopping the pump and shutting down the reactor. 2. To prevent unnecessary shutdowns, the control which indicates that #V 103 is frozen is now sealed oubt afier the fuel pump starts. The only purpose of this interlock is to prevent pump startup until the freeze valve is closed. 3. Jumpers were added avround the coolant salt system helium off-gas ra- diation contacts in the fuel pump circuit to prevent the pump from stopping each time a’ cireult test is conducted. The radiation monitor ig a safety-grade device, and its primary function is to initiate an emergency drain. Both channels must be tested routinely. Fuel Processing System. A mass flow-rate meter was installed to indicate the flow of HF to the fuel storage tank during fuel processing operations. The purpose of this second flow measurement is to provide an independent check on the existing orifice meter because the HY flow rate is important in calculating the amount of oxide removed from the fuel salt. To prevent possible diffusion of Hp into the HF gas suoply cylinder, interlocks were installed to close both the Hp and HF gas supply station valves when the main gas supply valve to the fuel storage tank is closed,. Operating Experience — Process and Nuclear Instruments Control System — Relays. Little or no trouble has been experienced. Virtually all design changes to the relay control gear have been made to meet new requirements developed by operating experience, not to correct malfunctions of this equipment. ‘ Valves. The difficulty experienced with the pressure control valve, PCV 522A, in the pump bowl off-gas system ls a part of the larger general problem of off-gas contamination:® by carryover of solids and vapors which are deposited in the off-gas lines. Valve-selection criteria for the off- gas system did not include considering nongaseous forelgn matter in the off~gas stream. Two pressure control valves, PCV 500J and PCV 510A1, in the main in- let helium line gave difficulty from galling. These valves are being re- worked. Pressure Transducers. One straln gage unit in the sampler-enricher suddenly shifted calibration, bubt during operation returned to its original state. The most likely explanation is moisture which, after geliing into the device, was baked out during service. Thermocouples. Thermocouple performance has been excellent. Only one in-cell thermocouple was lost during the six-month period covered by this report. The plastic insulation on the radiator thermocouple lead wires suffered from local temperaturesl7 which were substantlially in ox- cecs of those anticipated. Remedial measures such as insulating indi- vidual wires with ceramic beads, directed flows of cooling air, aand in- 48 sulating the high-temperature regions from heat sources have reduced, hut not eliminated, the problen. ILigquid-Level Bubblers. The helium bhubbler instrumentation used for fuel salt level instrumentation experienced a mechanical faillure in two of the differential pressure-sensing instruments. The Tailure was a leaky connection caused by weak welduments fabricated by the vendor and used to attach autoclave Titltings Lo the pressure inlet tubing. With the assistance of the Metals and Ceramics Division, the weldment was re- designed and rewelded. No further difficulties have been expericnced. Weight Tostrumentation on Salt Tanks. The system has not been en- tirely satisfactory. The basic input Instrumentation (weigh cells, mount- ing, etc.) has functioned satisfactorily, but the readout has given trouble. Manometer readout is accomplished by seclecting a particular weigh cell channel wiblh pneumatic selector valves. The valves are composed of a stacked array of individual valves operated by cams on the operating handle shaft. The valves leak; proposals to eliminate the problen are being con- sidered. Nuclear Safety Instrumentation. The electronic instruments have given little trouble. The solid-state modular instruments, hitherto un- tried In an operating installation, have needed very little service, and no ma jor problems have developed with use. Channel 3 of the safety system produced an abnormal numver of spurious trips. A large number of {aese are believed to have originated In faullty, vibratilon-sensitive relays in a commercial electronic swibceh which provides the high-temperature trip signal in the channel. Vibration isolation and substitution of a dif- ferent relay are being considered as possible antidotes. Another source of occasional false trips is believed to have originated with a chabter- ing of the relays which change the sensitivity of the flux amplifiers in the safety circuits. This has been corrected. Very generally the source of spuriocus trips has been difficult to trace, since they are random and usvally appear Lo be unrelated to events clsewhere in the reactor sys- tem. As electrical noise-producing components elsewhere in the system are eliminated, the number of false {rips is expected Lo be reduced. Wide-Range Counting Channels. Moisbure penetration has been ex-~ perienced with the wide-range counting channel "snake"'® assenbly. An improved waterproof Jacket is expected Lo cure tnis affliction. A faulty vernistat and an overloaded geary reducer in the drive unit of the wide-~ range counting channel required replacement. Linear Power Channels and Servo Controller. 'The compensated ion chambers use a small electric mobor to change compeunsation. One motor drive has given some trouble and has been responsible for the maintenance required by these chambers. The servo rod controller has been used for automatic control in both the Tlux and tewmperature modes. kExcepting the provlem associated with the wide~range counting channel, the servo's per- formance was very satisfactory. Flectrical Power System. Substitution of the new 50-kva solid-~state converter for the existing 25~kva motor generator is expected to reduce or eliminate mwany of the problems slemming from instability, noise, aod poor voltage regulation. o~ O False alarms from the monitron in the esst service tunnel were trace to electrical anclise from the sampler~enricher vacuum pumps. An electrical nolse Tilter is being designed to remedy this. A faulty solid-state switeh in the sampler-enricher was responsible for nolse 1n the cubput of the Sorensen regulator. Data bystem The installation of extensive modifications®® to the data-~-logger— computer was completed on August 31, 1965, and testing began immediately. ceveral Talled compubter components were found and replaced, and two loose connections in the analog input system were found and repaired. Additional deglgn changes and adjustments were required as a2 result of the modifica- tlons, and, with the exception of the digital filter-integrator, which re- guired complete redesign, these changes and testing were completed on Sep- tember 16, The seven-day acceptance test was restarted on Sepbember 16, 1965, and completed September 23, 1965, The remainder of Septepber was spent correcting miscellaneous hardware problems, principally locse cir- cult card connections. An alr-conditioning failure caused the computer room ambient temperature to rise to 85°F, and, simultaneously, the ac supply voltage Tell to 103 v. These two conditions, although simultaneous, were not coupled. I[o these cilrvcumstances it was impossible 4o keep the systenm on line for more than a few hours at a tiwme. It was concluded that suc- cessful operation of the camputer regquires that ambient Temperature be held below 85°F and that ac supply voltage be maintained between 105 and 120 v. catisfactory operation was restored, and the system was accephed on October 1, 1965, provided that Bunker-Ramo Corporation (L) supplies and installs a digital filter~integrator for the analog input signals, and (2) provides and installs, at OBNL's option, cirecuitry which prevents damage to the core memory by restart transients subsequent to a power logs and which provides a contbrolled sequence automatic restart when power is restored. October was spent checking and calibrating inout instrumentation and in troubleshooting hardware fallures, principally in the analog in- put area of the system. Instrument calibrations were completed in No- vember, as were substantlal modifications to lmprove reliability sched- uled by Bunker-Ramo. The digital filter-integrator and restart circuitry were alsce installed during this shutdown. In the period of December 1965 through February 1966, the logger- computer achieved operating status. In addition to the roubtine and periodic collection of operating data, 1t was used to obhain transient and Trequency responses and to determine fusl temperature coefficients of reactivity. It was programmed to operate the control rod for the pseudorandom bilnary tests reporbed in "MSRE Dynamic Tests," this report. The log of operating data, which did not seem significant during opera- ticn, became extremely useful during analysis of the off-gas problem. During this period the various compubing and logging programs were being modified in accordance with the reguirements developed during reactor op- eration, 50 84 Figure 1.22 charts, on a weekly basis, the "in service” or "on" time as a percent of total time. The shaded arcas on the figure are scheduled shutdowns and cannot be charged against the system as malfunctions. Duriung the period immediately following sysbem acceptance, performance, as noted above, was disappointingly low and did not approach the specified require- ments. 'The system has shown a slow but steady overall improvement 1n op- erating reliability as debugging progressed. IFor example, during February the logger-computer "in service” time was 99.7%. MSRE Training Simulators The power level tralining similatort? was operated successfully in October 1965 as part of the operator training program. It was set up on two general-purpose, portable BAT-TR-10 analog computers and tied in with the reactor instrumentation system. No special hardware was required. The control of the simulator was effected cntirely from the operator's console, where actual rod moticn, radiator door positioning, and blower CRNL-DWG 66-2622 'INTEGRATED ,/ "ON" TIME 100 - / ) . | L—j . - _T\"!‘ - *_&H-a.iw}-o—w— -m-».,a,__g..-;—‘l‘w *- ”“’\ ] \,,—m" i e ; 95 98 20 ‘ S | 801 E 3 "ON" TIME 4 \ SCHEDULED SHUTDOWNS § FOR MAINTENANCE AND 70 \ MODIFICATIONS 4 j {] SCHEDULED SHUTDOWNS i Y FOR MAINTENANCE OF €0 ‘ PERIPHERAL EQUIPMENT ;\; - w504 | = 4 £ ] N 40 30 20 10 OCT A NOV. 5. DEC.3. JAN.7 FEB.4. MAR.4. Pig. 1.22. Availability Record of MSRE Data System, October 1965 to February 1966, Inclusive. 51 menipulation were used as inpubs Lo the simulator. Readouts of the siwmu- lated linear and log power, key system temperatures, and healt power were provided by the reactor instrumentation. The simdator was used Lo check out the reactor flux and lcad conbrol systems and the procedures used to gwiteh from flux to temperature servocontrol. Documentation Except for some revisions and additions to instrument speciflication sheet and preparation of a design report, documentation of the MSRE in- strumentation and controls system design is complete. During the past report period, instrumect application and switch tabulations were com- pleted and issued, and design drawings were revised to incorporate as- built revisions and recent additions to and revisions of the system. References 1. S. J. Ball, Simulators for Tralning Molten-sSalt Reactor Experiment Operators, ORNL-TM-1445 (in preparation). 2. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, ». 19. 3. 8. J. Ball and T. W. Kerlin, Stability Analysis of the Molten Halt Reactor Experiment, ORNI~-TM-1070 (December 1965). 4 R. L. T. Hampton, Simulation 4(3), 179-90 (1965). 5. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNI-3872, p. 22. 6. MSR Program Semimnn. Progr. Repb. Aug. 31, 1965, ORNL-3872, o0p. 61—62. 7. R. H. Guymon, MSRE Design and Operations Repord, Part VIIL, Operating Procedures, vol. 1, ORNL-TM~-908 (December 1965). 8. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL~3872, pp. 118-19. 2. MSR Program Semiana. Progr. Bept. Aug. 31, 1965, CRNL-3872, vp. 94—948, 10, MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 43. 11. S. E. Beall et al., MSRE Design and Operations Report, Part V, Re- actor Safety Analysis Report, ORNL-TM-732, Fig. 2.27, p. 98, pp. 100-10L1. 12. Ipid., p. 104. 13. 1. 16. 17. 18. 19. 52 lpid., pp. 108-12. Ibid., p. 100. Thid., p. 98, Fig. 2.27. R. B. Briggs, Status of the Problem of Plugging in the Off-Cas Lincs in the MSRE, MSR-66-3 (Mar. 1, 1966) (internal use only). See Chap. 2 of this report. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 38. Ibid., p. 40, 53 2. COMPONENT DEVELOPMENT The efforts of the development group were devoted Lo assisting in the prepower operation and testing of the reactor. Several changes in the equipment and procedures were mades, and these are described velow. Freeze Valves The specifications Tor all but three of the Treeze valves were sim- plified by eliminating the "rapid" thaw requirvement. This requirement is needed in FV103, which controls the emergency drain of the reactor, and in FV204 and FV206, which control the emergency drain of the coolant gystem, but had been included in the remaining valves only as an opera- tional convenience. Since the maintenance of the proper temperature dis- tribution needed to ensure a rapid thaw was more difficult than was con- sistent with good operaling practice, the rapid thaw requircment was eliminated, and these valves are now operated elther in the thawed or desp~-frozen condition. Thne valves which might contain suffiecient radio- activity in the salt to produce radiolytic fluorine at low temperatures are maintained above 400°F at all times, Experience with the in~pile ex- periments and other tests have shown That essentlally wo fluorine is re- leased avbove this temperature. The addition of the modulating air-flow controllers on FVI03, FV204, and FV206 and the separation of the heater control circuits on FV204 and 206 have resulted in greatly improved operation of the freeze valves. Control Rods Control rod units 1 and 2 have operated without difficulty since the initial installation at the reactor. Examination of rod 3 at the end of the criticality tests revealed that the braided wire gheath had besn torn 2l a point 28 in. below the tow block. The inner convoluted hose was worn bub not completely through the wall. Cause of the damage appeared to be the upper roller in the control rod thimble, which had jammed and would not rotate. The roller had a galled flat area on one side. The thimble roller and upper control rod hose were replaced. Recentl examina~- Lion indicated no further dAifficulty after zeveral months of operation at, temperature. Rod drop times for 51 in. fall remain at 0,72 sec to 0.8 gec for all three rods. Overall rod lengths remained within 0.10 in. of the installed lengthg. 54 Control Rod Drive Units Thermal switches were installed in the lower end of the drive housing and are gcet to alarm when the temperature rises above 200°H, 'The air flow through the drive wunlt cases is abhout 1.4 scfm, which appears to be ade- quate to maintain the temperature within the housings at less than 200°F, 1t was necessary to replace the No. 3 drive unit because the position indicators revealed an 0.8-in. deviation from the original zero set point. The deviation was caused by a partially restricted spring actlion of the preload spring operator of the drlve chain. This allowed enough slack in the drive chain to permit the chain links to slip over the sprocket gear teeth. The unit continued to operate without difficully with the 0.8-1n. crror until replaced. The lower limlt switch of unit No. 3 showed an intermittent tendency to stick when the control rod was dropped from zbove 24 in. The switch could be released by fully withdrawing the coutrol rod. We found that the shock absorber stroke was 25% greater than normal when the control rod was dropped from 51 in. There is evidence that the inertia of the switch actuating arm was sulficient, when the control rod was dropped from 51 in., to overcome the force of the actuator recovery spring to a point where the lower end of the actuating rod could strike the lower flange of the drive unit, The actualing rod had been slightly bent duve to striking the flange, cauging it to bind in the guide bushing. A strooger recovery spring will be installed to prevent the overtravel. figure 2.1 is a photograph of the control rod drive units in position on the reactor. The shielding and access hatch have been removed, showing the air and electrical disconnects in the drive unit cover. When a unit is removed,the gmall hatch on the drive unit cover is removed (note unit No. 2 in Fig. 2.1), which permits access to the cortrol rod tow block Tor releasing the control rod from the drive unit, and also access to the lower drive unit flange Tor releasing the drive unit from the thimble flange. This method of removal has been utilized for control rod mainte- nance since the initial installation at the reactor. Radiator Doors Thermal warping of the radiation doors and door cgeasls permitted air to leak through the radiator enclosure at an excessive rabe, and modifi- cations had to be made to reduce the leakage. The following modifications were made (Fig. 2.2 shows the face of the radiator with the inlet door raised after repairs and modifications): 1. The door trip locks. The Jocks were modified and adjusted so that, when the door wag Tully closed, the trip locks Torced the metal door, by a rCorward camming actiocon, against the soft seal on the face of the radiator enclosure. The forward force 1s exerted by movement of the door in the downward direction into the rolling cam locks located on the radiator frame. 23 CONTROL-ROD ACCESS | H REMOVED dREMOTE MAINTENANCE PICKUP BAIL GRAPHITE SAMPLING ACCESS HATCH FOR 'N;TAL*’-‘«TION ,E DRIVE—-UNIT AND CONTROL-ROD REMOVAL < CONTROL-ROD~COOLING AIR-SUPPLY DISCONNECT ELECTRICAL DISCONNECT DRIVE - UNIT HOUSING .~ COOLING-AIR-SUPPLY DISCONNECT L % \-\\ " DRIVE UNIT NO. Fig. 2.1. Control Rod Drive Units in Operation Position at MSRE. : = PHOTO 52447 , RADIATOR DOOR (N RAISED POSITION CAM PULLOVER BRACKET A LOWER -LIMIT HOCK-ABSORBER ARM i Fig. 2.2. Inlet Side of Radiator Door in Raised Position. 56 Door guide tracks. The door moved too close to the radiator seals when it was raised or lowered. Warpage of the door caused some of the seal elements to hang on the door, and they were torn lcose. The position of the door track was moved to provide 1 in. of clearance between the seals and the face of the door when the door was released from the fully closed position. Cam follower bracket. The doors tended to drag and jam in the guide tracks. Cams had been installed in the tracks to force the doors away from the seals when they were not fully closed. The cam follower brackets are mounted on the doors, and contain rollers which ride on the track cams. These rollers were damaged by the wedging action of the trip locks against the door rollers and tracks when the doors were dropped into the closed position. The cams were changed to provide clearances of 1/8 in. between the rollers and cams when the doors were closed. A short movement of a door in the upward direction brings the roller into contact with the cam, forcing the door away from the soft seal. End rollers were installed on the cam follower brackets to prevent side motion from jamming the doors against the tracks. Reliable operation of the limit switches is important in reducing the number of ways in which the doors can malfunction, and the original switches were not very reliable. New heavy-duty switches and actu- ators were installed to obtain more positive action. An additional upper limit safety switch was added plus a mechanical "hard stop" above the added upper limit switch. In the event of complete switch failure the door strikes the "hard" stop and an overload switch stops the drive motor before the door seals can be damaged. The door position indicators are synchro-driven devices located on the drive shafts. The doors are lifted by steel cables which are wound and unwound on chain-driven sheaves on the drive shafts. If the doors jam while being lowered, the sheaves continue to turn and give incorrect indications of the true door positions. Continued ac- tion of the drive unit, after the doors have stopped moving, causes the 1ifting cables to unwind from the sheaves and attempt to rewind in the reverse direction. The cables usually Jjump out of the grooves in the sheaves and become snarled and kinked. A "loss-of-tension" device has been designed but not installed which will stop any action of the drive unit should the cables become slack. It is simply a switch which will be actuated by movement of the cable from i1ts normal tightened position. 7. When the clearance between the door and door seals was changed to 1 in. with the door open (see No. 2 above) it was necessary to provide additional clearance between the door and the radiator enclosure hood. The hoods on both inlet and outlet doors were shifted 1-1/2 in. to provide this clearance. B Radiator Heater Electrical Insulation Failure The electrical leads from the radiator heaters extended vertically up through the radiator enclosure and terminated in Jjunction boxes on the floor of the area between the radiator door hoods. These leads are insulated with ceramic beads. The electrical leads extending from the Junction boxes to the supply and control circuits were plastic-insulated wires which were rated for 140°F. Heat leaks up through the radiator enclosure along the wire bundles overheated the junction boxes, and the plastic insulation melted. The heat leakage into this area was reduced by welding the original sheet-metal enclosure wherever possible, by patching some openings, and by reinsulating. The Jjunction boxes were removed from the high-temperature area between the hoods to a cooler position on the east wall of the pent- house. Electrical wire extensions between the relocated Junction boxes and the beaded heater lead wires were insulated with 194°F thermoplastic and asbestos composition. Figure 2.3 shows the present arrangement of the heater wiring between the hoods over the radiator doors. Increased -+ . PHOTO 82264 ‘. CROSSOVER DUCT i OUTLET DOOR SHROUD ;i 7 TOP OF RADIATOR (8-in. INSULATION) e ey e o= Fig. 2.3. Area Between Radiator Door Shrouds and Top of Radiator Enclosure. 58 circulation of cooling air in that area was provided by rerouting some of the existing exhaust ducts. Cool air is directed downward through the drive mechanism into the heater lead area; exhaust ducts remove the heated air from the flcor between the hoods. The south exhaust duct can be seen in Fig. 2.3. There are two additional exhaust ducts on the north end of the ares; these are 10-in.-diam flexible tubes. Also a duct was installed between the door hoods so the hoods will be cooled when the main blowers are running. The average ambilent temperature in the area between the doors after the above modifications was approx 100°F with the blowers off. Tempera- ture within the beaded electrical lead bundle penetrations at the top of the radiator was 600°F with the blowers off. These temperatures are sat- isfactory, and the temperatures also were satisfactory with the blowers omn. Sampler-knricher During the shutdown period following run 3, the low-power experiment, several modifications were made to the sampler-enricher to increase the efficiency of operation. Most of these were discussed in the previous semignnual report.l The major modifications made at that time included: (1) replacing the removal valve with a modified version, (2) adding a knob to one linkage pin of each access port operator to aid in opening it with the manipulator if it should fail to function properly, (3) put- ting steel rings in the convolutions of the manipulator inner boot to hold it free of the manipulator arm, and (4) adding a pressure switch to prevent an excessive pressure gradient across the manipulator boot. The manipulator arm which was bent during run 3 was replaced. The hand was straightened, and a 1/4 X 1/4 in. projection was welded to the bottom of each finger to aid in handling the capsule cable. Clearances in the Castle Jjoint were increased to reduce the force needed to operate the manipulator. Additional changes have been made to improve the reliability of op- eration. The brass keys used to attach the capsules to the latch were replaced with nickel-plated mild steel keys. In case a capsule is dyopped it can now be retrieved with a magnet. Two changes were made in the interlock circuit. A pressure switch installed on the manipulator cover requires a negative pressure in the cover before the operational or the maintenance valve can be opened. Also, both the operational and the maintenance valves must be closed be- fore either the access port or the removal valve can be opened. These changes were made to ensure that there is double containment at all times. Installaticn of lead shielding around the sampler-enricher has been completed. Ten inches of lead or the equivalent was installed around the main components and 4 in. around the off-gas system. While the re- actor was operating at 1 Mw, the radiation level around the equipment was < 1 mr/hr. During runs 4 and 5, 40 samples were taken, 9 of which were 50-g samples for oxygen analysis In place of the usual 10-g samples. Two of the large capsules failed to trap a sample. The first assembly was not long enough to submerge the £ill opening of the capsule in salt in the pump bowl. All other assemblies were made longer. The second fallure to trap a sample probably resulted from the capsule catching on the latch stop at the top of the pump and not entering the pump bowl. The capsule has been modified slightly to make it hang straighter and is therefore less apt to hang. While withdrawing one of the 10-g capsules, the position indicator stopped, indicating that the capsule had hung or the cable had jammed. After repeated attempts, the cable was withdrawn completely. The cable was examined to determine the cause of trouble. Apparently the capsule or latech had hung on the gate of the maintenance or coperational valve causing the cable to coll up. The cable also backed up into the drive unit box and caught ian the motor gears. The cable was straightened, the open limit switches on the operational and maintenance valves were reset to open the valves wider, and the cutside diameter of the latch was re- duced. No similar trouble has been experienced. The equilibrium buffer pressure Lo the operational valve dropped from 55 to 35 psia during one sampling operation. This represents a leak of aboul 20 cc He/min through the seal on the wpper gate of the valve. A particle of salt or metal apparently dropped on the gate while the capsule was being removed or attached to the latch and lodged be- tween the gate and the seat. Repeated efforts to blow the particle from the sealing surface have falled. The leak rate has not increased with continued use. The leak of heliun from the buffer system is malnly an inconvenlence, as 1t slowly pressurizes the volume above the valve, and the gas must be vented to the auxiliary charcoal ved about once a day. Besides this seal, three more sealing surfaces between the pump bowl and the sample access area provide adequate safety; therefore the valve will not be replaced at this time. The sarple capsule used while the reactor was operating at 1 My left conslderable contamination in the transport contalner. The open end of the bottom cup read 40 r/hr near ccontact after the capsule was removed. Replaceable liners have been Tabricated to reduce the contamination left in the transport container. Coolant Salf Sampler During runs 4 and 5, ten samples were isclated from the cooclant punp bowl, two of which were 50-g samples. During one sampling the latch failed to slide freely through the transfer tube. No reason for the difficulty could be found, but the outside diameter of the latch was reduced to lessen the possibility of future trouble. During the shutdown period following run 3 a capsule was left hang- ing on the latch. A dry helium atmosphere was maintained in the sampler. o0 When the capsule was retrieved from the pump bowl in taking the first sample of run 4 (CPH4—1), a black film covered the salt and adhered to the outside of the capsule. The fl1lm was identified as decomposed oil. An- other sample taken 4 hr later was clean with no evidence of black £ilm. The sampler was opened and examined for oil contamination. A small amount was found on the top flange of the glove box near the elastomer gasket. No evidence of oil was found around the drive unit or capsule. he source of the oil was not located, bub subscqueni samples have been free of the film. Ixamination of Components from the MSRE Off-Gas System The difficulty with plugging and partial restriction in the MSRE of f~gas system 1s described in the section on systems performance. bSev- eral of the components which had given trouble during the early operation at 1/2 and 1 Mw were removed to the HRLEL for examination. Photographs were taken through periscopes, and many samples of the foreign material found in the components were removed and transferred to the analytical chemistry hot cells for analysis. The results of these analyses are re- ported in the section on chemistry. A summary of the visual examinations made in HDRLEL is given below. Capillary IFlow Restrictor FE 521 The flow restrictor consists of a short length of 0.08-in.-ID tubing welded into the line at ong end and coiled to fit inside a 3-~in.-ID con- tainer. The other end of the line 1s left free. The container was cut open, and the entrance and exit ends of the restrictor tublng were ex- amined. The entrance region wag clean, and only =a small deposit was found on the container wall near the end of the discharge region. This deposit is shown in Fig. 2.4. The restrictor was not completely plugged when examined in the hot cell. Cheeck Valve CV 533 The check valve is a small spring-loaded poppet-type valve installed in the gas line to prevent a backflow into the fuel pump gas space during the venting operation on the drain tanks. Some foreign material was found on examination of the poppet, butl this was nol sufficient {c stop the gas flow. The valve poppet is shown in Fig. 2.5, and for the purpose of the examination is shown in the open position. The soft O-ring which makes the seal can also be seen In its normal position ou the poppet. The other material on the cone of the poppet is an amber varnish-like material, which was identified as a hydrocarbon. Charcoal Bed Inlet Valve HV 627 The valve examined was one of four which control the gas fiow into each of four parallel charcoal sections which make up the main charcoal Bl bed. The valve had given indications of almost complete plugging before removal from the system. The valve was cut apart with a saw in the hot cell, and the valve stem is shown in Fig. 2.6. The small metallic chips shown near the large end of the cone are from the sawing operation. The entire cone was shiny as though wet with an oillike material. There was some white amorphous powder on the tapered section of the valve trim. Fig. 2.4. Deposit in Flow Restrictor FE 521. Fig. 2.5. Check Valve CV 533. 62 Fig. 2.6. Charcoal Bed Inlet Valve Stem HV 621. The powder seemed to be adhering fairly strongly to the metal surface, although there were some discontinuities. A similar material was found in the valve body. The general appearance of this area indicated that the powder had moved up and down with the motion of the valve stem and was therefore not dislodged by the small stroke of the valve. Samples of the white material were removed for analysis. Iine 522 Pressure Control Valve PCV 522 PCV 522 controls the gas overpressure on the reactor system by vary- ing the gas letdown (off-gas flow). The flow passage at the valve seat was quite small (about 0.010 in.) in order to control the relatively low gas flow (4.2 liters/min). The valve was mounted with the seat above the stem. Flow was down through the seal (reverse of normal). The valve stem was covered with an amber oillike coating, and there was an accumulation of fluid in the recess formed by the bellows-to-stem adapter piece as shown in Fig. 2.7. The tapered flat (flow area) on the stem had small accumulations of solid material. The valve body was coated with similar material and had a semisolid mass of material on the sur- face near the seat port as shown in Fig. 2.8. Samples of these materials were removed for analyses. 63 [ R27812 Fig. 2.7. PCV 522 Valve Stem. R27813 Fig. 2.B. POV 522 Valve Body. 64 Line Filter 522 Filter 522 is a l-in.-diam by 15-in.-long cylindrical type 316 stain- less steel sintered metal element in a l-l/2—in. (iron pipe size) pipe housing. Element thickness 1is l/l6 in. Flow 1is from the outside in. The filter area is 0.34 ftz, and the removal rating is 98% particles > 0.7 w. The pressure drop (clean) at 4.2 liters/min helium (normal MSRE flow) is 2 in. Hz0. The filter was installed immediately upstream of the reactor gas pressure control valve (PCV 522) to protect the very fine valve trim. The filter assembly was removed from line 522 on February 8, 1966. Initial visual examination was made at HRLEL on February 9, 1966. The upper one-third of the element was largely steel gray in appearance. As shown in Fig. 2.9, the steel gray also appeared in the lower portion at the seam weld and in a few isolated spots. The major portion of the lower two-thirds of the element had a frosty white appearance. At no place was there any evidence of a measurable buildup, or cake, on the filter surface. Visual examination was continued on February 10, 1966, at which time it was noted that the frosty white areas had become darker in ap- pearance, tending toward the steel gray. Inspection of the interior of the filter housing showed an accumulation of foreign material in the bottom as shown in Fig. 2.10. ©Samples were scraped from the side of the element. The filter was then reassembled, and a flow pressure drop test (see below) was run on February 11, 1966. The filter was once again dis- assembled, and the element was sectioned into short pieces. The foreign material in the bottom of the housing was removed readily by rinsing with carbon disulfide. R27776 Fig. 2.9. Frosty Deposits on the Surface of Filter 522. 65 R27814 Fig. 2.10. Deposit in the Bottom of Filter 522 Housing. Flow Test on the MSRE Filter from Line 522 The filter and filter housing which had been removed from line 522 on February 8, 1966, was tested in the hot cell of the HRLEL as indicated in the flow diagram given in Fig. 2.11l, and flow tests were conducted on February 11, 1966. The gas was discharged into the gas disposal filter system of the hot cell. By taking readings of the filter pressure gage through the hot cell window for various helium flow rates the relative plugging of the filter was determined. The data obtained are shown in Fig. 2.12. The graph also shows the results of a helium flow test conducted on a duplicate but clean filter. The data indicate that for a given pressure difference the 522 filter was passing only 5% of the corresponding flow of the clean filter. How- ever, extrapolation of the data to the normal MSRE off-gas flow indicate that, even though 95% plugged, the contribution of the filter (0.075 psi) to the total system pressure drop (5 psi) was negligible. Fuel Processing System Sampler The mockup that was used to develop the sampler-enricher for the MSRE is being modified and will be installed as the sampler for the fuel 66 ORNL-DWG 86-4754 DISCHARGE UPSTREAM — PRESSURE : GAGE T HOT~CELL WALL = gl il iy b === FILTER ASSEMBLY ?ggfifi&gRE P NO. 522 LINE 522 THROTTLE / VALVE SINTERED METAL PREFILTER HOT CELL NN tFLOWMETER HELIUM SUPPLY Fig. 2.11. Test Arrangement at Building 3525 Hot Cell. ZFlow vs pressure-drop test. Filter No. 522 removed from MSRE on Feb. 8, 1966, ORNL-DWG 66-4755 HOT-CELL TEST 2-11-66 ON FILTER 522 REMOVED FROM MSRE LINE 522 ON 2-8-66 .’§ = HELIUM o G 0 w z NORMAL MSRE 9 OFF-GAS FLOW 9 R o & od BENCH TEST BY ORNL ON CLEAN TYPE-522 FILTER 0.0t 100 10' 10° 10° 10 HELIUM FLOW RATE (liters/min/ft%) Fig. 2.12. Flow vs Pressure Drop. MSRE filter No. 522. Filter element: Pall Trinity type G. 316 ss sintered metal. Area: 0.34 f£t2. N ~J processing system., Design of the modifications to the mechanical equip- ment is finighed except for the ghielding. Modifications to the equip- nent to upgrade it for 15-psig service are nearly complete. Redesign of the instrumentation is complete. This work involved some revisions Lo panel and field instrumentation, some changes in control circuitry, preparation of installation and interconnection drawings re- quired to install the system at the MSRE site, and revisions of nomen- clature and numbering of instruments and circuits to conform with the MSRE practices. Functionally, the ingtrumentation and control of thie chemical proc- ess sampler is similar to that provided on the fuel sampler-enricher system; however, the detail design of the chemical prccess system sampler ig simpler. This reductlion in complexity resulted from reduced require- ments Tor containment, which, in turn, resulted from the lower radiation levels present in the chemical processing system. The reduced contain- ment requirements permitted the use of conventlional (sifigle tracked) control circuitry instead of the redundant (dual tracked) "safety grade circultry used in the fuel sampler-enricher. Also, it was possible, in scme cases, Lo use commercial-grade components instead of special weld- sealed components. In other cases the reduction in requirements for con- tainment and/or redundancy elimlinated the need for components. 1 O0ff-Gas Sampler A system is being designed to permit analysis of the reactor off- gas stream. The arrangement is shown schemstically on Fig. 2.13. A side stream of 100 cc/min is taken from the reactor off-gas stream at a point downstream of the in-cell volume holdup. The sample stream passes to the sample unit inlet manifold, whence it may be routed to one of three alternate paths: 1. a thermal conductivity cell for on-line determination of gross con- taminant level, 2. a chromatograph cell for batch-wise and gquantitative determination of contaminants, 3. a refrigerated molecular sieve trap for isocolation of a concentrated sample, which will be transferred to a hot cell for analysis. The sample equipment will be housed in a containment box located in the pipe trench south of the vent house. A recirculating air system coupled with a radiation detector is provided to permit rapid detection of system leaks. In addition to the off-gas sampler, an abtempt will be made to study the nature and character of off-gas contaminants by inserting various gample coupons into the pump-bowl vapor space through the sampler-enricher line. Instrumentation and controls design for the off-gas sampler is in progress and nearing completion. Instrumentation is being provided for 68 ORNL-DWGC 56-4756 CONTAINMENT {5¢f, O psig 30 cfm - T T T T T T T T T T ‘]_ . VENT "TO STACK JTHERMAL—CONDUCTIVITY o CELL AIR SYSTEM | [ | | | _ - ! CHROMATOGRAPH t) : —— {2 CELL B 1 | T | | i TO AUXILIARY ! i e b — CHARCOAL BED NOTE: DELAY TIME ‘——ee ——i BETWEEN PUMP BOWL AND @ wfl RECIRCULATING SAMPLE UNIT APPROX | | | | | | ) Fomin L — A X REFRIGERATED SAMPLE TRANSFER | MOLECUILAR-SIEVE TRAP SOTTLE ENRICHER SAMPLE-ISOLATION | SAMPLER BOMBS e Ll \ 3 . L 100 em™/ min 3psig . 4200 cmd/min volume | L. | o | | voume i — CAD:(AJ woon BE“E;‘ HOLDUP HOLDUP "L_i_Lm 1?5 ---------- o ' STACK FUEL PUMP LINE 522 BOWL REACTOR OFF-GAS Fig. 2.13. Schematic Diagram — MSRE Off~-Gas Sampler. on-line chromatographic and conductivity analysis; for measurements of flows, pressures, and temperatures required for proper operablon of and interpretation of data from the chromatograph and conductivity analyzers; for control of temperature of a molecular sieve trap and of the level of a liquid-nitrogen bath in which the molecular sieve is immersed; for de- tection and annunciation of undesirable operating conditions; and to pre- vent the occurrence of harzardous conditiocns. Since the sampler will be an integral part of the primary contain- ment during sampling cperations, and since some components of the sampler would not meet the requirements for primary containment system components, solenoid block valves will be installed in the inlet and outlet lines which connect the sampler to the reactor system. Two valves will be in- stalled in series in each line. These valves will automatically close and isolate the sawmpler from the reactor system in the event of high pres- sure in the reactor containment cell, high pressure in the fuel-pump bowl, or high air activity in the sampler enclosure. High reactor cell pressure is indicative of a rupture of the primary contalinment and the occurrence of the maximum credible accident. High fuel-pump-bowl pressure indicates that conditions exist that could result in a rupture of The sampler pri- mary containment. High sampler-air activity indicates that a rupture of the sampler primary containment has occurred. Closure of the block valves resulting from high sampler-air activity (and tne aCCcompanying alarm) will also provide protection to the sampler operator against the occurrence of high background radiation resulting from small leaks in the sampler. Sampler-air activity will be detected by two GM-tube-type (ORNL model Q- 1916) gamma monitors, which will monitor two separate and independent 69 alr samples collected from and rebturned to the sampler enclosure. The isolation block valves and assoclabed detecting instruments and control cireuitry were designed in accordance with the recommendations of the ORNL "Proposed Standard for the Design of Reliable Reactor Protective Systems.” At least two independent devices were used to supply input signals to the system and to effect the corrective blocking action. Cne-out-of-twe logic was used in the control circuitry, and separation and identification were maintained in the detail design of the wiring. Panel~-mounted instrumentation will occupy 5 lin £t of panel (6 £t nigh X 2 ft deep). These panels will be located in the vent house, south of the reactor building. The sampler and associated instrumentation will be located in a trench outside the vent house. ©OGince all major sampling operations will be carried out at the sampler, all readout of ianformation will be presented at the sampler panels; however, occurrence of an alarm condition at the sampler will actuate an annunciator in the main control room, and provisions are being made to permit some information to be trans- mitted to the Computer—-Data Logger. Also, a sample permissive switch will be located in the main control room. This switch, which is connected in the vlock valve circuits, will be used to prevent operation of the sampler without knowledge of the reactor operators. Most of the instrument components used in the off-gas sampler were salvaged from the ORNL MIR-47-6 experiment located at Idaho Falls or were on hand in ORNL Reactor Division stores. Reconditioning and callbration of this equlipment and procurement of additional components are nearing completion. Preparaticn of instrument applicatlion drawing is essentially complete. Panel design is approximately 75% complete. Interconnection wiring and piping design is under way, and fabrication of panels is also under way. Xenon Migration in the MoRE Based on results of the ©7Kr expérimenta and xenon stripper effi- ciency reported by the University of Tennessee, the steady-state 135%e poison fraction for the MSRE was computed. The following major assump- tions were made: 1. Wo bubbles circulating with the fuel salt. 2. Todine remains in solubtion, that is, it 1s not adsorbed on any sur- faces and is not volatilized in any form. 3. Xenon is not adsorbed on any surfaces. 4. The mass transfer of xenon across the salt-gas interface in the graph- ite pores will not be lnterfered with, for example, by the accumula- tion of sludge or fission products at this interface. 70 To compute the 135%e poison fraction, thne 135%c concentration in the salt must first be calculated. This 1s accomplished by the following rate bal- ance: 135%c generation rate = 12°Xe decay rate in salt g S + 135%e burnup rate in salt + 135%e stripping rate (1.) + 135%e migration rate Lo graphite, where 135%e migration rate to graphite = 135%e burnup rate in graphite + 127%e decay rate in graphite.(2) At a given power level the left gide of Eq. (1) is a constant, and the right side is a fuaction of the '?7Xe concentration in the salt. After the concentration is computed the 135%e poison fraction may then be computed. The results of This calculatlion at eguilibrium are shown in ['igs. 2.14a, b, and ¢, as a function of the variable indicated when all other variables remain constant. The circle on the curve indicates the ex- pected or design value. For this calculation the core was divided into 72 reglons, and average [luxes and adjoint fluxes were used for each region. Figure 2.14c shows that the poison fraction is considerablg lower at low power levels. This is because a higher fraction of 135%e decays alt these power levels. One notable result of the calculation is that the xenon poison fraction did not change in value when the diffu- sivity of xenon in graphite ranged from 1077 to 5 X 1077 ftz/hr. This is because the mass transfer coefficient from salt to grapnite is con- trolling the transfer process. Therefore, 1f future reactors have low mass transfer coefficients as in the MSRE, it may not be necessary to purchase the more expensive low-void-fraction and low-diffusion-coefficient graphite, ifT r35%e poison fraction is the only consideration. The final measured *37Xe poison fraction may be different Irom these predicted values because of the agsumptions involved. As more information is gained about the reactor system and its chemistry, the equations will be adjusted to provide a better model for calculation of the migration in future systems. Remobe Maintenance The activities of the remote maintenance group were dictated largely by the condition of the reactor. During the last stages of constructlion and for as long as the cell was open, attention was given to trying the remote maintenance techniques on the installed eguipment and in check- ing to ensure and improve maintainablility. Once the cells were closed the effort was turned to recording cxperiences and planning for later work. Tlater, actual maintenance was performed in several areas. A de- scription of this work is given below. 71 ORNL -DWG 66 -4757 3 2 \fi\ \“\ 1 S = o (@) 5 0 L 2 0 5 10 15 20 25 1w STRIPPING EFFICIENCY (%) = Q o g -3 € > w 0 2 _.-——-‘—"—"_-—-_-—- .‘-_-'-,‘—--" /.’ q 8 = (£) S 0 5 0 0.02 0.04 0.06 0.08 0.10 g o MASS-TRANSFER GOEFFICIENT TO BULK GRAPHITE (ft/hr) = Q 0 o 3 D e w0y M (e} 1 10 100 REACTOR POWER (Mw) LOG SCALE Fig. 2.14. (2) Effect of Stripping Efficiency on 135%e Poison Fraction; (b) Effect of Masgs Transier Cogfficient on 13°Xe Poison Frac- tions; (c) Effect of Reactor Power on 135%e Poison Fraction. = Practice Before Operation Practice with reactor components was had in handling the pump ro- tary element and in changing the existing graphite sample assembly for the final uwnit. The sampling procedure included the handling and/or operation of (1) an inert atmogphere in the standpipe, (2) a special heater tool to thaw out a frozen salt annulus around the outside of the sample holddown assembly, (3) disconnects, valves, and flanges inside the standpipe, (4) stowing the holddown assembly, and (5) the sample it- self. In addition to these a limited visual inspection of the core was conducted using a 7/8—in.~diamASCOpe, A special fitting had to be in- stalled in the strainer basket above the core to acconmodate the revised sample assembly. The pump rotary element was lifted out of the cell for inspection using the lifting yoke and crane with in-cell direct guiding. During the 1ift, clearances of all the auxiliary eguipment were observed and points where damaging interferences could occur were noted. The unit was reinstalled by remote means using the view afforded through the port- able maintenance shield. The process of torgueing up the twenty-two 1- 1/2-in.-diam bolts in the pump's lower flange was started in the portable shield bul was finished with direct means to save time in the reassembly process. While set up in this area, representative electrical and ther- mocouple disconnects, pump-bowl heaters, and auxiliary Tlanges were han- dled with the long-handled tools. Procedures were revised, and in cases like the pump, graphite sam- pling, and the control rods were written up in minute detail. Tabula-~ tilons were prepared relating neaters, Lhermocouples, spare disconnects, and other eguipment with shield-block locations and elevations. Additional tools were designed, and some existing tools were revised. Maintenance of Radiocactive Systems In January, after a short period of power operation, it was necessary to use remote maintenance on a number of tasks. [Table 2.1 shows the op- erations which were done and the actual time and manpower used. Radiation levels were significantly lower than those which are anticipated after prolonged power operation. While this required only miniwal shielding, enough of the shielding was used for the operations with the control rod and. PCV 522 to provide experience in setting up and in using the tools through the portable shielding. Although PCV 522 and its filter assembly was quite a large radiocactive source, 100 r/hr at the cuter surface of the filter, no significant personnel exposure was encountered. 1In general the tools, technigues, and previously made preparabions proved more than adequate. In some cases it was necessary to fabricate special tools or revise existing ones, but these cases did nob hold up progress to any great extent. Good cooperation belween management, reactor operations, maintenance planning, health physics, and the craft foreces contributed toward efflicient and smoothly run operations. Table 2.1. Summary of Remote Mainternance Work at MSRE — January 27 Through February Perscanel Involved Llapsed Description . Tim Dat : Rlgger . : s Remote (hr? ’ and Millwright Pipe Fitter Maintenance OUperator enanc 1. Remove check valve from line 533 G 1 2 2 3-1/2 Jan. 27 (use Cam-loc on extensicn) 2. Remove FE 521 and repisce with a 0 o 4 2 2 Jan. 28 new unit 3. Control rod drive No. 3: remove 2 1 0 1 6 Jan., 29 shielding, disccunects, and cover plate Qontrol rod drive No. 3¢ remove 2 2 O 2 g8 Jan. 31 drive ard install new drive Control rod drive No. 3: hook up 2 2 0 2 2 Feb. 1 air and electrical disconnects Control rod drive No. 3! replace 2 2 0 1 2 Feb. 2 cover and shield 4. POV 522, unsuccessful abtempi in-cell 2 0 2 2 3 Feb. 3 replacemen 5. PCV 522, remove and put intce deconteami- 3 1 2 2 6 Feb, 5 nation cell PCV 522, revuiid and replace in systenm 3 1 2 6 Feb. PCV 522, remove from system 2 1 i 2-1/2 Feb. & PCVY 322, cut filter loose from valve, 2 2 1 2 Feb. put in can and carrier 9, PCV 522, attach flanges for fiow l{ests 0 0 i 1 2 Feb, @ and 10 10, PCV 322, install new assembly 2 G 1 1 1 Feb. 10 Total {hr) g9 48 4y 76.5 46 €L T4 Pump Develobment MSEE Punps Molten-Salt Pump Operation in the Prototype Pump Test Facility. The modifications® to the test loop were completed. The venturi flowmeter was relocated upstream of the orifice flow restriction, and a new flow- straightening section was installed near the pump discharge. 'The new flow straightener was modified o provide additional weld attachment of the blades inside tae pipe. The new arrangement of the system is shown in Fig. 2.15. "wo test runs were made with the prototype pump circulating the salt TiF-BeFa-Zr¥ ,-Tht 4-UF 4 (68.4-24.6-5.0-1.1-0.9 mole %) at 1200°F. The ORNL-DWG 66-2048 AlR § 4 " ] WELL -~ _____ {\\ _fl/ ” SALT PUMP AR COOLER VENTURI METER ,,,,, :Ih L . ' ST nql--m ) . e :::-"t-:-~Tf.i;f.‘.f,i_‘1j:_j_'_';:_:-_—‘k— o J i DISCHARGE = “SALT FLOW < FH;- .n.%fin_iFT\\\\\ PRESSURE STRAIGHTENER — THROAT INLET PRESSURF —~*j[ PRESSURE ). 41 _8-in. PIPE ) i i} _&-in. PIPE [ FREETE FLANGE J - e T \ R s L gifiifif mm J e SALT FLOW . T , T T T \ QRIFICE FLOW L:J RESTRICTER _____ L FREEZE VALVE \\ ’ MSRE FREEZE FLANGE Fig. 2.15. MSRE Salt Pump Hot Test Facility, Modified. 75 first run covered a period of 165 hr to provide hobt shakedown of the im- peliler (13-1in.-0D) for the fuel pump spare rotary element. The other run covered a period of 166 hr to provide shakedown of the drive motor for the spare coolant pump. The prototype pump ie being prepared for test with an 11-1/2-in. im- peller. Measurements will be made with the radiation densitometer to determine the undissolved. gas content of the circulating salt. OCther tests will be made in the pump tank off-gas circuit in connection with the plugging incidents experienced at the MORE. Pump Rotary Element Modificaticon. The spare rotary element for the fuel salt pump was modified to provide a positive seal against leakage of oil from the catch basin past the outslide of the shield plug and into the pump tank. Previcusly, a solid copper O-ring compressed between the bearing housing and shield plug was used to provide the seal. Incidents have occurred wherein oil leaked past the O-ring. Cross-sectional views of the part of the pump which ineludes the modifications are shown In IMig. 2.16. The larger section shows the relationships between the pump shaflt, the shaft lower seal, bearing housing, cabch basin, shield plug, and tThe pump tank. The exploded views indicate the nature of the modificatlon. Positive containment of the oil is obtained by providing a seal weld be- tween the bearing housing and the shield plug. The rotary element is being assembled for cold shakedown, and after satisfactory operation will be prepared for MERE use and stored. The spare rotary element for the coolant pump will receive the same modification. Tubrication System. Modified Jet ;pum.ps4 were ingtalled in the re- turn oil lines of the MSRE salt pump lubrication systems. The lubrication pump endurance test’ was continued, and the pump has now operated continuously for 22,622 hr circulating oil at 160°F and. 7C gpm. MK-2 Fuel Pump. The pump tank design was completed, and it is being reviewed. Drive Motors, A new design was conpleted for the containment vessel for the drive motors of the MSEE salt pumps. It provides single contain- ment for the electrical power penetrations, in contrast to the double con- tairmment provided in the original design. The resulting simplification should eliminate a serious fabrication problem previously experienced with weld-induced laminations 1In the vessel wall. Other Molten-Salt Pumps PK-P Fuel Punp High-Temperature Imdurance Test. Endurance wperation6 of this pump was continued, and it has now operated continuously for 5160 nr circulating the salt IiF-BeF,-Thi,4-UF, (65-30-4-1 mole %) at 1200°F, 800 gpm, and 1650 rpm. 70 Pump Containing a Molten-Salt-lubricated Journal Bearing. This pumpV'B’g was placed in operation circulating salt at 1200°F, but it ex- perienced a seizure after 1 hr of operatlion. IExamination revealed the seizure to be in the molten-~salt-lubricated bearing. As was the case with the test previous To this one,s two of the snap rings for the bear- ing sleeve gimbals support were lost and the other two were only loosely retained. Whether the bearing seizure or loss of the snap rings occurrcd first is a mool guestion. The design of the gimbals support is being ORNL-DWG 66-2049 SOLID COPPER O-RING BUNA N O-RING Ve LOWER SEAL //// CATCH BASIN 7 /// 7 7 - LOWER SEAL 7 — //// \ CATCH BASIN [N, 77 A d = IR SRR BEFORE MODIFICATION AFTER MODIFICATION SHAFT LOWER SEAL--- / - BEARING HOUSING SEAL OIL LEAKAGE OQUT T ,‘ g OIL LEAKAGE s PUMP TANK SHIELD PI.UG SHAFT § -7 Fig. 2.16. MSRE Salt Pump Rotary Llement, Modification. 7' modified to eliminate the use of the snap rings, which are intended for retaining the four fulcrum pins in the support. Instead, the pins will be retained by plugs tack-welded at thelir outer ends. Instrument Development Ultrasonic Single-Point Molten-8alt Level Probe As no fuel has been processed during the last report period, the ultrasonic level preobe installed in the MoRE fuel storage tank'® has not been used. Efforts to correct this instrument's deficiencies have been unsuc- cessful. The excitation osecillator which had proved so unstable was modified considerably (using information supplied by the manufacturer) in an attempt to eliminate its excessive frequency drift. These modi- fications made no improvement in the stability of the oscillator. As the design and success of other intended modifications depended upon the stability of this conmponent, none of them have been attempted. To cor- rect this situation, the ORNL Instrumentation and Controls Division Elec- tronic Design Group has been requested to investigate the electronic equip- ment asscciated with the ianstrument and make recommendations as to the most practical action to take. High~Temperature NaK-Filled Differential-Pressure Transmitter The coolant-gsalt system flow transmitter that failed in service at the MSREYL, 12,13 yag refilled with silicone oil and tested with the seals at room temperature. Prior to refilling the transmitter body, a vacuum punp wag connected to the instrument in such a manner that the pressure could be reduced on the process side of both seals and both sides of the silicone~filled body at the same time. Liguid braps were installed in the body evacuation lines to catech any oil that might be Torced out of the transmitter body by the expansion of trapped gas during evacuation. When the pressure wag reduced to 28 in. Hg vacuum (the lowest pressure attainable with the system at that time), oil was forced from the capil- laries: 8.2 ml from the low-pressure side of the transmitter, and 15.8 ml from the high-pressure side. This indicates that there was gas trapped in both sides of the transmitter body, and that the amounts trapped were unequal. The exact amount of gas trapped cannot be calculated from the amounts of oil caught in the contalners, because the capillaries are not connected to the two sides of the transmitter body in the same horizontal plane. After it was refilled, the temperature sensitivity of the instru- ment was 1 in. (water column) change in indicated outpul for each degree Fahrenheit change in ambient temperature. This sensitivity is greatly reduced from that observed prior to refilling but is still excessive. A1l the above testing was done with the seals and transmititer body at room temperature. The seals are now being heated to 1200°F, at which temperature the test will be continued. 78 The new NaK-~filled differential-pressure transmitter ordered as a sparc for the MSRE is scheduled for delivery in March. Float-TLype Molten-8alt Level Transmitter Performance of the ball-float-type transmitter installed on the MSRE coolant-galt pum_pl4 continues to be satisfactory. The necessary actions have been completed to correct the previously reported errors in calibra- tlon.t%Y% A these were calculated corrections, some additional future adjustment may be reguired. The ball-float level transmitter on the MsSRE pump test loop con- tinues to operate satisfactorily, although on one fill the float would not rise untlil the temperature of the heater on the bottom of the float chamber was increascd. bGither uanmelted salt or curvabure of the bottom inside the float chamber may now be assumed to be the cause of this stick- ing. A previous inspection of the core tube showed no deposits that would cause it to stick in the core chamber. The bottom inside curvature of the float chamber on the MSRE pump test loop fits The curved bottom of the float very well and may block the entrance of the molten salt into the chawper. This design error was noted prior to the fabrication of the iloat chamber for the MSRE and corrected on the assumption that this valving action might occur. The bottom of the float chamber was flat- tened so that the round-bottom float could not block the flow of the molten salt into the chamber. Bxcept for detailing of the differential transformer assembly, de- sign of the ball-float transmitter installation in the MK-2 MSRE fuel circulating pump has been completed. Cne of the prototype ball-float level transuitterst® installed on the level test loop completed four years of operation at temperature this month. Tt is still operating satisfactorily. The other was re- moved last year so the ultrascnic single-point level indicator could be installed. It was operating satisfactorily when removed. Conductivity-Type Single-Point Molten-5Salt Level Probe Performance of the conductivity-type level probes installed in the MSRE fuel, flush, and coolant drain tanks continues to be satisfactory. There have been no further failures of excitation and signal cables on these probes. Single-Point Temperature Alarm Switches Observation of the performance of 110 single-point temperature alarm switches installed at the MSRE has continued. As reported previously17 the switch modules were resel prior to power operation of the reactor. Data obtained from subseguenlt spot checks of module set points indicated 79 that a few modules had shifted excesgsively; however, these data are con- sidered to be inconclusive because records indicated that the modules may have been readjusted. Some additional cases of duvual set points were dis- covered and corrected. Routine checks and observations of module perfor- mance will be continued until sufficient data are obtained to permit an accurate evaluation of the reliability of thege devices. Helium Control Valve Trim Replacement Testing of alternate material combinations for helium control valve trim has been terminated. BSince tests had shown that all the alternate material combinations under consideration would gall when operated in dry helium without lubrication and that none would gall if a minute amount of lubricant was present,l7 the valves which had failed in MSRE service were refitted with spare trim using the original (L7-4 PH plug and Stel- lite No. 6) material combination. Close attention was given to alignment during reassembly of the valves, and all trim was given a light coat of machine oil before installation. All the repalred valves operated satis- factorily in shop tests. One of the repaired valves was installed in the MSRE main helium supply line and operated for several months before stick- ing. One additional heliuvm control valve failure (also due to sticking) has occurred. Both valves are being disassembled and will be refitted with spare (17-4 PH to Stellite No. 6) trim. Because the failures may be due to evaporation of the lubricant, an attempt will be made to find a less volatile oll or grease to use for trim lubricant. Previous at- tempts to use grapnite-base lubricants were not successful. Thermocouple Development and Testing Coolant Salt Radiator Differential Temperature Thermocouples. Noise in the coolant salt radiastor differential temperature thermocouple circult which was under imvestigationla was finally reduced to an acceptable level. A final check on the effects of thermocouple and lead-wire material mis- match was made by heating thelr disconnects both individually and simul-~- taneously to 150°F. No readable change in the output voltage was nobted. A geven-point calibration was run on the thermocouple pair between 1040 and 1250°F, which resulted in a constant error with the ocutlet thermo- couple reading 0.220 mv high with respect to the inlet thermocouple. The resistance of the thermocouple loop was determined to be high encugh to reguire recalibration of the receiving instrument to maintain reguired accuracy. Also, the zero of this instrument was offset to correct for the thermocouple loop error. Thermocouple Drift Tests. The checking of elght metal-sheathed mineral-insulated Chromel-~Alumel thermocouples fabricated from MSRE ma- terial for calibration drift at 1250°F was concluded.!® A1l thermocouples continued to show some drift to the end of a 26-month test period. The final temperature equivalent drift values were between +4.7 and +6.4°TF. 80 Temperature Scanner Performance of the temperature scanning system developed for use at the MSRE?? has continued to be satisfactory. The modifilcation previ- ously reporte&jzl together with improved calibration and maintenance pro- cedures, appears to have eliminated the calibration drift problems. Per- formance of the mercury switches has been excellent. We had expected that the swiltches would need freguent attention and that the mean 1life between routine cleaning or repair would be about 1000 hr. The switches have given very little trouble, and the mean life of the switches hag been much greater than 1000 hr. Since the start of operation of this system in September 1964, there have been no bearing or other mechanical failures of the five switches installed. IFour switches developed excessive noise during this period and required cleaning and replacement of the wercury. In one of these cases the switch failure was caused by a failure in the nitrogen purge gas supply system. All five switches were cleaned and reconditioned before the start of power operations as a routine precaulbionary measure. References 1. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL~-3872, pp- 56-58. 2. Ibid., p. 55 3. MSR Progrem Semiaonn. Progr. Rept. Aug. 31, 1905, ORNL-3872, p. 61. 4. Tbid., p. 61. 5, TIbid., p. 62. 6. TIpid., p. 65. 7. MSR Program Semiann. Progr. Rept. July 31, 1963, ORNL-3529, p. 54, 8. MSR Program Semiann. Progr. Rept. Jan. 31, 1964, ORNL-3626, p. 40. 9. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 52. 10. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 66. 11. Tpid., p. 70. 12. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, v. 48. k2 13. MSR Program Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p. 45, 14. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 7L1. 81 MSR Program Semiann. Progr. Rept. Feb. 28, 1965, MEBR Program Semiann. Progr. Reph. Feb. 28, 1962, MSR Program Semiann. Progr. Rept. Aug. 31, 1965, lbid., pp. 73, 74. Ibid., p. 73. MSR Program Semiann. Progr. Rept. Aug. 31, 1962, M3R Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL~3812, ORNL-~3282, ORNL-3872, ORWI-3269, ORNL-3812, P D. . 62 3. MSRE REACTOR ANATYSIS least-bquares Formula Tor Control Rod Reactivity One routine fTunction of the MSRE data logger will be the periodic calculation of the separate reactivity effects assoclated with reactor operation above some base line, or zero-reactivity condition. For this purpose, one of the guantities the computer must calculate is the nega- tive reactlvily corresponding to each configuration of the shim and regulating rods. By use of a least-squares curve-tfititing procedurc, we have obtained an analytical formula which closely approximates the in+ew gral curves determined from rod calibration experiments at zero power The Tunctional expression used for fitting the experimental curves was determined by anplying a perturbation technigue to the integral ex- pression for the rod reactivity, as outlined below. If a reference state is chosen to corregpond to zero rod ingertion, an expression for the static reactivity change when 2 single rod (or rod group) is inserted to a position Z in a core with effective height H and radius R 1is fflR ffiZ ¢§ BA @r dr dz Apg = 0 - 0 —— . o) fO fO Pof0r dr dz In this Tormula, @ is the column vector of group fluxes corresponding to the condition with the rods inserted, ®O is the row vector of adjoint fluxes for the reference state, 8A 1s the local perturbation in the neu- tron removal operator due to absorptions in the rods, and P is the neu- tron production operator. This expressiocon can also be applied if the shim and regulating rods are inserted to different positions. To simplify the analysis, we have assumed that the tips of the shim rods are at egual insertions, Z, which is always equal to or above the position of the regu- lating rod, Z;. We Turther specialilze the formula to two-~group diffusion theory and one dimension, corresponding to the directlon of rod Insertion. Equation (1) becomes Z1 % (1) . Lo ¥ (2) fO dpo d)zap ¢'2 Az +le ¢02 o5 a0 ¢2 dz 0 j; o1 (vEg, o + Vi, 0,) de where subscripts 1 and 2 designate the fast and thermal groups respec~ (1) a? gsorption cross section in the region O = 2 = 75, polsorned by the combined tively. The quantity 875 1s the effectlve Increment in thermal ab- 83 (2) az in Zp & 2 £ 7o, polceoned by the regulating rod alone. shim and regulating rod group, and BJ; is the corresponding increment [ &) 8) -.L’ p & In the usual approximation of perturbation theory, the flux distri~- butions ¢1(z) and ¢5(z) are replaced by the unperturbed fluxes ¢pi1(z) and ¢02(Z), corresponding to the reference state. IF the approximations nZ 92(z) ® ¢pp(z) = sin T (3) * * . 7 dp1r ~ Pgp ~ sin 3 (4) are introduced into (2) and the integrations performed, the usual formula for the reactivity-worth curve for a partially inserted control rod is obtained.® This result was tested by adjusting the parameters 8282 and ég) in order to fit the formula to the rod-worth curves obtalined from experiment. While reasonably good agreement was obtained in fitting the curve for the single (regulating) rod, poorer results were obtained when combinations of insertions of shim and regulating rods were considered. The functional expression obtained from standard Tirst-order perturbation theory was, therefore, Jjudged inadequate for representing the rod-re- activity curves. 02 We have found that a significant improvement in the fit can be made by assuming that the axial distribution of thermal flux, ¢, 1s perturbed according to the position of the shim-regulating rod bank (Z;) but is un- affected by the position of the regnulating rod (Z5). "The approximate formula for ¢, used for this analysis was: 0=z = Zy: 0s(z) = _—-—A----O; sinh oz + —— sin 2, (5) 2 2 D,BZ + D07 DEg, + DaZ 7hn = 2 = H: 95 (2) = B sinh Q(H — z) + — T sin =, (6) D. B2 + D.OF DB2 + Dor? 2 T 2 B wihere . (0) e - ...:a..:?m Do 7’ 62(;) O = O 4 i p Dy - (3)° T H ’ and D, and Z;g) are, respectively, the thermal diffusion coefficient and the macroscopic thermal cross section of the core in the absence of the control rods. The coefficients A and B depend on the strength of the 1) shim-regulating rod bank absorption, 52;2’, tors. They can ve determined by requiring continuity of Llux and current at the interface, Z;, between the "rodded” and "unrodded" regions. The final result of applying this analysis in Eq. (2) was together with gecmetric fac- _Co + C1 F1i(Zy) + Co Fo(Zy, Zp) + C3 Fa(2q) + Cy Fu(Zy, 72) Ap_ = 2 (7) ” Cs + Cg F1(2y) where 217 . 217 1 (Zl) = ""iLé‘;"* — 310 -——«Hl ) (8) I I ) ‘_; FolZa, Z2) = gflififififfiiéi# + sin 2151 m-sin42?;2 , (9) FB(ZT) = (COSh a7y sin %3:, — aj:i[-“j: sinh 071 cos Tlgl > T T ginh ol Zy) cos ZZA J , (10) x o . [ cosh O(H —~ Z;) sin T = - . | . . N [ 107 AVS = h & — e QL ol b4( 1, Zo) (\0051 71 sin 5 i sinh 73 cos T > X { cosh G — Z,) sin Moz, T sinn C(H -~ Zo) cos ) ] . (11) OH The parameters Co, Ci, ... , Cg were found to depend on the rod absorp- : (1) (2) tion strengths, 82a2 and 62&2 ters characterizing the reactor core in the absence of the rods. In addition, it was found that » together with other macroscopic parame- << 1, 0273 sH. (12) ng 7y (Zy) Cs This latter result is useful in modifying expression (7) Turther, in order that linear leasl-squares analysis could be used to determine the numerical values of the parameters. Thus, I [Cs +Cg Fr(Z2)]™t = 6%- (1.-£€?L ) . (13) By introducing this approximation in Eq. (7) and redefining the constants, it was found that Ao, (Za, Z2) =ag + 81F1 + apF, + asFy + agf, + asFf + 2gF1Fs + agfiFa + agfF, . (14) Bxpression (14) was fitted to the experimental rod calibration curves, using standard linear least-squares analysls to determine the coefficients ag, «.. , 8g. The position of the rods relative to the extrapolated zero of the unperturbed axial flux distribution (Z = 0) is Zi,2 = %o + X1,2 , (15) where Zg is the "zero point" insertion of the rods when withdrawn to their upper limit, relative to the extrapolated zero of the flux distribution, and X1,z are the measured rod insertions. Adequate approximations for the parameters characterizing the unperturbed flux distribution, Zo, &, and I, were obtained from earlier core physics studies. The results of this analysis are swmarized in Table 3.1 and Fig. 3.1. 1In Fig. 3.1 the magnitude of the rod reactivity is plotted as a function of rod position for various configurations of shim and regulating rods. The ordinate scale in thnis figure i1s arbitrarily normalized to zero measured reactivity when the regulating rod is fully inserted and the two shim rods are with- drawvn to 51 in. The leftmost curve represents the reactivity change when the regulating rod is moved with the shim rods fully withdrawn. The rightmost curve represents the case when the three rods are moved in a banked position with the tips of the rods at equal elevations. The re- maining curves represent the reactivity change when the regulating rod is moved with the shim rods held fixed at variocus intermediate positions. 86 The so0lid curves are the reactivity magnitude calculated from the least- The points shown as solid dots are sample points deter- mined from the experimental rod calibration curves. Over most of the range of rod movement the calculated and measured reactivity are in squares formula. agreement within about 0.02% 3k/k. Table 3.1. Near the extreme positions of the Numerical Values of Parameters in Ieast-Squares Formula for Control Rod Reactivity (Eq. 14) Parameter Value Parameter Value o, cmt 0.082 as ~1.944 x 1077 7o, Cil 24 .6 8, 5.891 x 1078 ag 2.125 g 2.865 x 1073 aj —0.3935 an —1.747 % 210™7 a, -0.3585 ag —4.358 x 108 55 ORNL-DWG 66-4758 , o 20 1~ e POINTS OBTAINED FROM ROD-CALIBRATION EXPERIMENTS .8 = = CALCULATED FROM LEAST-SQUARES : FORMULA T 16 - {ROD REACTIVITY NORMALIZED TO | < 7171 kg 235U IN LOOP) 82 * g ol T | - 5 S2 >— — 5 10 5 3 & 08 |- Lot i | | % oe-———L—‘fi - Lo | = 04 |— . . ® d/ / 0.2 ')fl//% L ] 0 4 8 12 16 20 24 28 32 36 40 44 48 52 POSITION OF ROD MOVED (in. WITHDRAWN ) Fig. 3.1. Comparison of Control Rod Reactivity from kxperimental Curves and from IlLeast-Squares Formula. g7 rods the error is somewhat larger. However, the least-squares formula should be adequate for automatic moniteoring of the control rod reactivity under most normal operating circumstances. Spatial Distribution of '2°Xe Poisoning in MSRE Graphite To supplement the experimental studies of the behavior of noble gas injected into the MSRE fuel,? theoretical calculations were made to de- termine the influence of the spatial distribution of 135%e sbsorbed within the graphite core. Tt is expected that the concentration of xenon In the salt will be relatively uniform throughout the volume of circulation. However, within the graphite pores, the 127Xe would tend to assume an overall spatial distribution governed by the burnout rate in the neutron flux. This distribution would be concave, with minimum concentration occurring near the position of maximum thermal Flux and maximum concen- tration near the boundaries of the reactor core. It is intended that the resactivity due to 135%e poisoning will be periodically calculated during operation by use of the TRW-340 data logger. TFrom a practical standpoint we are limited to the use of a relatively simple "point" kinetics model for on-line calculations. Therefore, any corrections for the spatial distribution of the 135%e poisoning must be predetermined from theoretical studies with a more elaborate model of the reactor core. If we consider a step change [rom one ?Qwer level to another, the correction for the spatial distrivution of 2%¥e within the graphlite region can be determined from X ___g . . ) Srapaze ¢ ) Fln G, 6] o) av, = AN NXE((I):L) t) f W(Po, P1, % , (16) o™ (r) o(r) av_ graphite where Pgp, P1 = initial and final power levels, t = time after power level is changed, @l(r) = thermal Tiux at position r relative Lo the center of the core after the power level 1s changed, @1 = thermal flux after the power level is changed, averaged over the graphite volums, 'fig [(P (,.) .-] = ] ...-,_:1_135- - trati s sraphl be % 1lr), tl = locally averaged Xe concentration in graphlite at position r and time €, 88 fig (61, t) = locally averaged 135%e concentration in graphite at time t, corresponding to a fictitious "flat” neutron flux distributiocn equal tTo the spatially averaged flux, = normalized distribution of thermal group importance (adjeint flux). ) = normalized distribution of thermal flux, ) As defined by Bq. (16), W is the factor by which the *3°Xe reactivity calenlated according to a "point" kinetics model must be multiplied to account for the spatlal distrivution of the poisoning. In this equation the "local average" '?°Xe concentration, N%e, is the radial average over the graphite volume associated with a single fuel channel. It is useful to compute this quantity before performing the integrations over the entire core graphite volume indicated in Eq. (16). Although nearly all the xenon in the graphite would be expected to be in the pores nearest the graphlite-salt interface, this local averaging procedure can be used because the radial variation of neutron flux across a single graphite stringer is negligible. We have used a one-dimensional model for the fuel channels in order to simplify the caleculations. The fuel and graphite widths corresponding to a single channel were chosen 50 that the ratios of mass transfer sur- face area to volumes were eqgual to those of the actual channel. 'The eguations governing the production and mass transfer for 135%e between the salt and graphite were: (17) . g (fifl - l ) , (18) yo o VL Xe Xe NG _ . SN o St T DXe ”axg - [GXe (=) + 7\Xe]NXe ? (19) 4 ¥1 By V1 J; Nie(x, 1) ax . (20) g9 In these egquations N%e(x, ) =0 Nxe(t) =0 NI,Xe Z NXe yo Xe Xe Yo Ji VC/VL il local concentration of *3%%e at position x within the graphite stringer, measured from the graphite-salt interface, atoms per cublic centimeter of graphite, local concentration of 135Xe, averaged over the graphite velume assoclated with & single fuel channel, average concentration of 1357 ang *35%e in the circu- lating fuel salt, atoms per cublc ceatimeter of liquid, concentration of '3°Xe in salt nearest the graphite interface, directly exposed to the graphite pores, atoms per cubic centimeter of liguid, fission density in core salt, fissions per cubic centi- meter of liguid per second, thermal neutron flux at position r, measured from the center of the graphite core, neutrons cm™? sec'l, fission yield of Y?°T and 137xe, radioactive decay constants for *3°I and 135Xe, sec"l, effective 1?7Xe removal rate due to external strippiog, sec” , 135Xe 2 thermal neutron sbsorption cross section of , cm®, diffusivity of xenon in graphite, square centimeters of graphite per second, coefficient for liquid film at the interface, cm/sec, mass transier graphite-salt half-width of single fuel channel, cm, half~width of graphite assocliated with a single fuel channel, cm, ratio of volume of salt within the graphite-moderated region to the total volume of circulating salt. Boundary conditions are required to solve the differential equations (18) and (19). fuel channel, At the outer boundary of the graphite associated with a single g - BNXG ~= | - X=y1 90 At the graphite-salt interface, 2 g N = BNXe(x = 0) , (22) J0 > B Xe _ =L a8 _ 0o\ Se = h { Moo BNXe(x =0)| , (23) x=0 where B = RT/HXee, R = universal gas constant, 82.07 cn® atm mole™ (°K)™%, T = temperature, °K, HX° = Henry's law coefficient, cm® atm mole“l, € = graphite porosity, cubic centimeters of void per cubic centi- meter of graphlite. AL time © = 0, equilibrium conditions corresponding to the initial power level (PO) were assumed. The initial concentrations are the solu- tions of Fgs. (17) through (23), with the time derivatives cgual to zero. For times following the change in power level, the differential equations were solved by means of laplace transforms. A sufficient approximation to the exact solutlion was obtained by use of the condition a2 Yo~ o 4 (24) B - Xe Physically, this condition implies that most of the resistance to mass transfer between salt and graphite 18 due to the fluid film. Results of krypton injection experiments appear to support this conclusion.?»? some typical numerical calculations based on the preceding model are given in Figs. 3.2 and 3.3. Numerical values Tor the effective mass transfer coefficient, stripping rate, and graphite porosity characteristics were obtained from ref. 3. Figure 3.2 shows the time variation of the correction factor, W, as the power level ascends to 10 Mw. In the limit- ing case of a step change from O to 10 Mw, W decreases monotonically to an equilibrium value of approximately 0.76. Curves are also glven for the case when the power level 1s increased in successive steps, each time 1.00 91 CRNL-OWG 66-4759 0.20 0.80 ///fll | 1 w Y3Bye POISONING SPATIAL CORRECTION FACTOR ( 5.0-=10.0 Mw Q=10 Mw 2550 Mw 070 O o @ Fig. 3.2. Time Dependence of the Spatial Correction for the Poisoning in MSRE Graphite: 16 20 24 28 TIME (hr) 32 36 40 Ascending Power Level. ORNL-DWG 66-4750Q 100 - 1 , {O—=0 Mw a e e g | et 5 ] 10,00 Mw ‘&) | L // ‘_'____.-—M 5 /,,ff”"“ 2.5-m10 Mw Eé 090 / e 1 © 5.0 =25 Mw I prommn % ] o ) 2 100~ 50 Mw z I . M £ 0.80 | e LT = R v o o Q = @£ 0 ¥ 0.70 0 4 8 12 16 20 24 78 32 36 40 TIME (hr) ¥ig, 3.3. Time Dependence of the Spatial Correction for the Poisoning in MSRE CGraphite: Descending Fower Level. 135Xe 135Xe 92 allowing equilibrium poisoning conditions to prevail before further in- creasing the power. The initial "dip" in the curve representing Lthe 5- to 10-Mw step is a consequence of the relative time lag between the in- crease 1n the burnocut rate in the graphite and the ilncrease in the pro- duction rate of *2°Xe in the salt. Figure 3.3 gives the analogous results for power levels descending from 10 Mw. Further studies will be made with this theoretical model in order to determine the sensitivity of the calculated corrections to the ¢f- Tective mass Lransfer coefficlent and stripoing rate, and to determine ....... 1. The Reactor Handbook, vol. IIT, Part A (Physics), ed. by H. Soodak, pp. 204—6, Interscience, New York, 1962. 2. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNI~3812, pp. 12-14. 3. R. J. Kedl, personal communication, dJanuary 19606. Part 2. MATERIALS STUDIES 95 4o METALIURGY Dynamic Corrosion Studies A test program is in progress to study the compatibility of struc-~ tural materials with fuels and coclants of interest to the Molten-Salt Reactor Program. Thermal convection loops described previouslyl’z are used as the standard test in this program. Circulation of lead in a Cb—1% Zr alloy thermal convection loop? was terminated after 5280 hr. This loop operated with a hot-leg tem- perature of 1400°F and a 400°F AD. Metallographic examination of the hot leg, shown in Fig. 4.1, revealed no evidence of attack; however, a small guantity of dendritic crystals was observed in the cold leg. Electron probe analysie indicated these crystals to be columbium, as shown in Fig. 4.2. Mass transfer of columbium in lead has not been re- ported previously. Two 2-1/4% Cr—1% Mo steel thermal convection loops were started with lead coolant to study the effect of magnesium additions on mass OO INCHES Pig. 4.1. Section Through the llot Leg of a Cb—1% Zr Loop Which Op- erated with Lead for Over 5000 hr at 1400°F, 96 Y-67208 LIGHT OPTICS COLUMBIUM L, X-RAY IMAGE Fig. 4.2. Results of Electron Probe Analysis Showing Presence of Columbium Crystals on the Cold Leg of a Cb~1% Zr Loop Which Operated for Over 5000 hr with a Hot-Leg Temperature of 1400°F and a 400° AT. 250X. Reduced 35%. transfer of primary elemenis. The loops will operate with a hot-leg temperature of 1100°F and a 200° AT. Magnesium was added to act as a deoxidizer and an inhibitor. An original plan to include titanium in the lcad was deferred because of difficulty in making a Ti~Pb alloy. Two loops containing molten fluorides continuecd to operate without incident. A type 304 stainless steel loop with removable specimens has operated for 22,000 hr, and a Hastelloy N loop containing specimens of Hastelloy N modified with 2% Cb has operated for 33,000 hr. MSRE Material Surveillance Tests Reactor Surveillance Specimens Specimens of Hastelloy N and grade CGB graphite were exposed for approximately 1100 hr to molten fluoride salts in the core of the MSRE during the precritical coperation and the 1nitlial critical and assoclated zero-power experiments. The purposes for these were: (1) to monitor materials during these preliminsry operations and (2) to be the mass eguivalent for the similar specimens that replace them for surveillance of the power experiments.3 These specimens showed no changes as a re- sult of exposure during these Tirst, mild experiments. 97 The graphite specimens were nominally 0.8-in.-diam by 10-in.-long rods. They were machined from MSKEE graphite, grade CGB, having a rela- tively high concentration of cracks, with the idea that this would magnify adverse changes that might occur in the relatively short, mild exposure. There was essentlally no salt on the graphite, and the machining marks appeared unaltered. As in the laboratory tests, radiographs showed that surface-connected cracks tended to be filled with salt, with no salt penetrating into the graphite. The volume of salt in the graphite aver-~ aged 0.06% of the bulk volume of the graphite, which is approximately one-tenth of the maximum permitted in the MORE design specifications. The relatively small dimensions of the specimens and the presgence of a more-than-typical quantity of cracks would tend to give apnormally high salt pickunp. The Hastelloy N specimens, in the form of tensile specimen rods, also drained free of salt. They had lost their bright, shiny, machined surface and had a bright, gray-white matte surface similar to that ob- tained in hydrogen firing of the metal. Fig. 4.3. Microstructure of Hastelloy N Tensile Specimen Exposed in MSRE During Zero-Power Testing of Reactor. Microstructure is ex- tremely fine grained. Black band approximately 0.0005 in. below surface is 0.0005 in. wide and is composed of grains whose boundaries have been enriched or depleted in some constituent. Etchant: aqua regia. 28 Two of the specimens were sectioned and were examined metallographi- cally. For comparative purposes, two control specimens of the same composition were also examined. Longitudinal and transverse sections of the shoulder and gage-length regions of the lensile specimens were ex- amined. Both the tested and control specimens exhibited extremely fine- grained microstructure (see Figs. 4.3 and 4.4). The grain poundaries in a 0.0005-in. band approximately 0.0005 in. below the surface of the tested specimens were darkened. 1t appears that the boundaries were perhaps either depleted or enriched in some unknown constituent. A section of a tested survelllance specimen has been submitted for analysis with the electron microprchbe analyzer; however, the analysis has not been completed yet. a4 A longitudinal view of the control specimen is shown in Fig. 4.5. The microstructure of the control specimen indicates that the surveil- lance specimens were in a severely worked conditicon prior to testing. During exposure in the reactor at approximately 1200°F, the alloy re- crystallized to the smaller grain size shown in Fig. 4.3. R-27463 T —3.007 IHCHES = 500X ™ Fig. 4.4. Microstructure of Control Hastelloy N Tensile Specimen. Although the microstructure is extremely fine grained, the specimen is coarser grained than the one removed from the MSRE. 99 Pig., 4.5, Longitudinal View of Control Hastelloy N Teusile Specimen. Microstructure indicates that specimen was in severely worked condition. Etchant: agua regia. Lurvelllance Control Specime The reactor core control speclimen rig4 wag charged with fluoride c salts in the latter part of December 1962, and 1t has beein operating satisfactorily. This unit containsg three sets of vahlt@ and Hastelloy N specimens that mabch the sets of reactor core specimens” in the MSRE. The function of this unit 1s to subject its zpecimens to approximately the temperature profile and the major temperature and pressure fluctua- tions of the reactor. These unirradiated specimens will be used to db- tain base-line data for those thaet are irradisted in the reactor. The reactor control specimen test unit will copy changes of the re- actor operation, as specified above, through directions relayed to it by the computer that monitors the MERE. Calibration of the unit with the computer is in progress. 100 Hot-Cell Metallographic Examination of Hastelloy N from Experiment MIR-47-6 for Evidence of Witriding When the MSRE 1s operating, the atmosphere in the reactor cell and the drain tank cell will be nitrogen containing about 3% oxygen. There has been some concern that the nitrogen, when ionized by radiation, will nitride and embrittie the Hastelloy N of the reactor components. Since the capsules in experiment ORNL MI'R-47-6 were ccooled with air and with alr and nitrogen mixtures during irradiation, we examined the bottoms of capsules 1 and 2 for evidence of nitriding of Hastelloy under irradiation. The capsules were contalned in a copper heater block so that only the tops and bottoms were exposed to nitrogen. Construction of the cap- sules and the operating conditions during lrradiation have been re- ported.6 Unfortunately, the outer surfaces of the boltoms were damaged during initial disassecmbly of the capsules at the MIR hot cells. Ma-~ chined grooves on the capsule bottoms were still visible, and the bottoms R-27480 o - z - & - e A . i o L i "’ £ / o ¥ p— S " ’ — - 2 { sR f > : ./'.’ : E ' # B Y e < - i ; (\ . - 0y A “ & ¢ # © .« . S - v - i . * * 3 7z~ % - . . I . . -~ . Lfié ». - L : : F P # = * . ' . I ¥ . ? e ! ¥ C f( € - e O . 5\ - . ] . & L . % « - o e - 0.18 Mev) reasonably soon. Samples of needle-coke graphite and isotropic graphite are being tested. A needle-coke graphite was used in the MGSRE to minimize irradiation contraction and associated stresses. Isotropic Y-69255 Fig. 4.12. Radiograph of Thin Sections Machined (a) Transversely and (b) Longitudinally from the Graphite-Molybdenum Brazed Joint Test Showing That No Salt Penetrated into the Graphite Pipe Walls. (salt would appear here as a white phase.) 108 graphite has been included in these studies because it may be superior in mechanical properties and more resistant to damage under the high ex- posures required in the MSBR. 8» 9 A part of this effort is to obtain graphite with the configuration required for current designs of molten-salt breeder reactors. This is being done because the shape and size can significantly affect the final properties that can be bullt into the graphite. Current designs require graphite pipe nominally 4 in. 0D X 3 in. ID X 8 to 12 ft long and 2-1/2 in. OD X 1—1/2 in. ID X & to 12 ft long. Secondary attention is given to the small-scale laboratory-produced materials. The prime requirements of the graphite for the initial procurement are that it have pores small enough to prevent the entry of molten salts and that it have a low gas permeability. The pore entrance diameters should be less than 0.5 p, and permeability to helium at 1 atm of pres- sure should approach 10™7 cmz/sec. The suppliers currently propose ma- terial with a permeability in the range of 107 to 1072 cm?/sec, with the higher values being favored. To date we have obtained graphite samples from the Carbon Products Division of the Union Carbide Corporation, Great Iakes Carbon Corporation, Poco Graphite, Inc., Speer Carbon Company, Stackpole Carbon Company, and the Y-12 Chemical Engineering Group of the Development Division. The material obtained from the Carbon Products Divislon is a needle- coke graphite, and the materials obtained from the others are isotropic graphite. Short pieces of pipe 4.7 in. OD X 3.5 in. ID and 3.6 in. 0D X 2.5 in. ID have been supplied by the Great Lakes Carbon Corporation and the Carbon Products Division respectively. The materials from the others were rods or blocks. To determine if these grades of graphite (and future grades) are potentially useful for an MSBR, we are routinely examining them for the entrance diameter spectrum of the accessible pores, permeability to helium gas, permeation by molten fluoride salts, microstructure, specific registance, flexural strength, and coefficients of thermal expansion. The current samples are in this stage of examination. Those that appear to have promise will be given additional tests for purity and crystallo- graphic development and will be included in the irradiation studies and graphite-metal Jjoint development mentioned previcusly. Evaluation of the Effects of Irradiation on Graphite Recent progress in the development of graphite for reactor use may be applied directly in estimating properties of a graphite designed for the MSBR. There are, of course, several unknowns which 1limit the ability to state categorically that any graphite will withstand the MSBR environ- ment. The major limitation is the lack of evidence to demonstrate that any graphite can sustain massive doses of 1023 nvt or greater and still 109 retain its integrity. The tubular thin-wall design in the MSBR is one of the better configurations for reducing the differential-growth problem to within the capavilities of the graphite. The tubular shape is also easy Lo fabricate and to test reliably and nondestructlvely in order to ensure maximum integrity. The properties of the most promising grade of graphite can be pro- Jected Trom available grades with a Tair degree of certalnty (see Table 4.2). "These estimations are based primarily on properties obtained from isotropic grades with densities required for the MSBR grade. The magnitude of the stress generated by differential growth can he fairly well approximated. The main uncertainty is in the flux gradient across the tube wall. Using the conservative estimation of 2.4 X 1024 in. in.™?* nvt™t as the growth rate and a 10% change in flux across the wall, a differential growth rate of 2.4 X 1072% in. in.™t nvt™t is ob- tained. The restraint is internal; therefore, only about half of the differential growth is restrained, so the effective strain rate is 1.2 X 10725 in. in.7t netT. The creep rate coefficient is 4 X 10727 in. in.”* psi“l nvt"l, and the stress 1is simply and directly calculated to be D5 = Loz X 10777 = 30 psi g o= < =z — 4 % 10727 Thig stiress could hardly be a cause for failure, and to obtain stresses in excess of 100 psi would require the use of grossly unreslistic ma- terial properties. Failure, however, could result from the inability of the grapnite to absort creep deformatlion even though the stress level is much less than the fracture stress. For lifetimes of 1.5 x 10?7, 3 x 10?3, and 6 x 10?2 nvt, the strain to be absorbed would be about 1.8, 3.6, and 7.2% respectively. This corresponds to 5-, 10-, and 20-year lifetimes with a dose rate of 9 x 1014 nv, the maximun fast flux in the present design of the MSBR. The consideration of a strain limit for failure is realistic; however, the strain limit for fracture has not been estab- lished. It has been demonstrated that graphite can absorb stralns in excess of 2% in 1072 nvt without loss of mechanical integrity. There 1s also some evidence that the growth rate will diminish after a 1022 nvt dose; thus, the graphite might not be forced to absorb the total quantity of strain calculated. Therefore it appears that failure by reaching a strain limit will require at least five years of service. The main uncertainty, as mentioned earlier, is simply the ability of the graphite to sustain the massive dose without loss of integrily. There 18 no experimental evidence beyond 2 X 1022 nvt on which to extrapolate the irradiation damage of graphite, so extrapolation of data to 1023 yould be pure conjecture. There are, on the other hand, several factors which force one to be optimistic about the ability of graphite to sustain the Tablie 4.2. Properties of Advanced Reactor Grades of CGraphite Grade Property 1-207-85% H-315-A% E-319% MSBR Density, g/em’ 1.80 .85 1.80 ~1.83 Bend strength, psi >6000 4500 >£000 >5000 Modulus of elasticity, psi 1.35-1.65 x 10° ~2 % 10° Thermal conductivity, 1622 21=2"7 23 24 Btu hr™t £t (°F)TH Thermal expansion, ¢ 5.4=6.,0 4.8-5.8 3.-4.4 0 R5.0 10°%/°¢ Electrical resistivity, 8.9 9.9 =9 omm-cm X 10% Isotropy factor, CTE}}/CTEL 2.11 1.20 1.18 N1.15 Creep coefficient at 700°C, N4 x 10727 in. in.”t psi"‘~L vt Permeability to helium, cu®/sec 8 x 1072 2 x 1072 <10-2 Dimensional instabiiity? at 700°C 1/2 to 1/3 of AGOT 50 p in diameter. On resuming operation at 1 Mw, restrictlions appeared to develop in three locations: at wvalve 522B, at the entries to charcocal beds 1A and 1B, and in the lines ahead of the auxiliary charcoal bed. Currently, & specimen has been obtained only from valve 621, which controls flow into charcoal bed 1B. Examinations and tests are still being performed Table 5.1, Summary of MSRE Flush and Fuel Salt Analyses, Full-Power Experiment Concentration {wt %) Concentration (ppm) Sample Circulation Number Period (ar) Li Be Zr v F > Fe Cr Ni Mo o” 0° Fiush Salt FP4-1 13 35 FP4s2 19 13.65 9.83 <0.0025 0.0210 80.52 104,02 110 <10 33 <15 56 FP4e3 25 46 FP4ed 27 13.55 9,35 <0.0025 0,0207 79,34 102.26 212 62 30 <15 74 FP4e5 29 72 P46 a2 13.50 9.96 <0.0025 0.0200 80.07 103.55 125 <10 <20 <15 180 FP4.7 40 13.65 9,46 <0.0025 0.0241 77.85 100.98 180 54 <20 <15 150 FP4-8 48 106 FP4.9 50 13.35 9.98 <0,0025 0.0221 75,80 99.15 210 " 57 <20 <15 142 FP4-10 64 13.55 9.49 <0,0025 0.0230 75,05 98,11 128 60 <20 <15 1300¢ Fue! Suf'r FP4-11 3 120 FP4-12 19 10,45 6.33 10,52 4.673 66.45 98.42 96 56 41 <15 144 FP4.13 24 105 FP4-14 27 10.20 6.41 10.77 4,634 64.67 96.68 79 41 69 <15 92 FP4-15 31 80 FP4.16° 48 FP4-17 53 65 FP4-18 64 10.25 6.68 11.24 4.651 67.44 100,26 144 a3 44 <15 85 FP4.19! 68 FP4.20° 75 FP4.21 99 10,47 6.40 10,82 4,671 66.79 99,15 121 60 52 <15 FP4.22 124 10,54 6.54 10.95 4,664 64,68 97,37 116 44 84 <15 FP4.23 148 10.27 5.74 10.85 4,642 65.06 97.57 99 46 35 <15 FP4-24 172 10.65 6,55 10.96 4,655 67,66 100,58 116 48 45 <15 FP4-25 159 10.60 6.37 11.41 4.646 65.44 98.47 26 35 42 <15 FP4-26 217 10.60 5,63 11.20 4,642 66,90 99,07 89 48 7 <15 FP4.27 245 10.55 6.53 11.07 4.618 67.68 100,45 227 50 41 <15 FP4-28 268 10.60 6.42 11.10 4.663 66,19 98,97 211 49 34 <15 FP4.29 314 10,70 6,71 11.54 4,654 69,75 103,35 111 39 31 <15 FP4-30 362 10,63 6,81 11.19 4,661 57.32 100.58 83 49 27 FP4.31 435 10,30 6.63 11.26 4,632 68,51 101,33 150 37 41 GCt Table 3.1 {continued) Concentration {(wi %) Concentration {(ppm) Sample Circulation Number Period (hr) i Be Zr ye r 2, Fe Cr Ni Mo o° 0° FP4.32 457 10,55 6.71 11,80 4,625 67.66 101,34 173 43 33 FP4-33 529 11,208 6.75 11.07 4,596 56.25 99,97 55 30 39 £P4-34 601 11.35% 6.4% 11,13 4,601 68.20 101.82 164 58 16 FP4-35 659 11.36% 6.68 10.86 4,721 69,35 103,93 74 54 <5 FP4-36 i1.25% 6.32 11,20 4.632 56,76 100.16 125 47 <5 FP4-37 11,302 5.33 11,08 4,622 69.35 102,68 189 53 80 FP4-38 10.65 6.54 11,46 4,608 67.25 100.56 311 51 25 FP4-39 10.60 6.33 11.54 4,519 68.33 101.42 78 51 40 Coolant Salt CP4=1 7 CP4=2 il 83 25 <5 13¢ CP4-3 59 13.78 8,91 75.70 99,39 21 41 37 1i0 185 CP4-4 67 46 53 24 60 CP4-5 180 14,20 8.87 <0.002 76.80 99.87 50 50 20 60 CP4-6 194 33 CP5-1 236 13.86 B.B5 <0.002 76.7 99,41 50 35 <10 150 CP52 i3.82 B.55 <0.002 70.6 99,67 59 35 <10 110 8values corrected to 33.241 wt % bHF-purge method. CKBrF4 method. dSample exposed to dry-box atmosphere for 48 hr, e gL No sample obiained. f . . For amperometric analysis, 35, 235y gErrc;m~.=ously nigh; attributed to failing batteries in automatic pipette. act Table 5.2. Summary of MSRE Fuel Composition Anslyses Component Nominal Book Zevo-Power, s FP4-11 to FP4-39 Experiment Tank (mole %) TiF 65.00 64,88 62.31 64,40 63.36 = 0.567 BeF, 29.17 29,26 31.68 29,68 30.65 * 0,533 ZrF, 5,00 5.04 18 5.11 5.15 £ 0.116 237.003yp, © 0.83 0.82 825 0.803 C.825 * 0.011 (wt %) Ti 10.95 10.93 10.25 10.94 10.51 £ 0.137 Be 6.32 6.34 .71 6,49 6.55 £ 0,161 77 10.97 11.06 11.11 11.32 11.14 £ 0.295 237,003 473 4 646 4,602 4611 4 642 + 0.028 “Based on four samples. Based on four samples obtained on completion "237.003y - 33,241 w4 235y, of the zero-power experiment. L1 128 with the available specimens, esults are sumarized in Table 5.3. On the basis of the results shown 1in Table 5.3, we would conclude tentatively that the restrictions in the off-gas line may be attributed to varnish-like organic material. Gulfspin-35 is used as the lubrical- ing oil for the rotary element. It is composed primarily of a mixture of long-chain aliphatic linear and branched hydrocarbons. Recent measurements show that its refractive index is 1.473, The refractive index of the heaviest 10% volume fraction obtained on vacuum distillation of the o0il was found to be approximately 1.50. The high refractive index of the varnish-like materials removed Trom the off-gas system suggests that radiation polymerization has produced the plugging phases. This inference appears to be strenglthened further by the observation that the varnish-like materials appeared to have limited or no solubility at room temperature in xylene, petroleum ether, carbon tetrachloride, or accltone. Uranium-Bearing Crystals in lFrozen Fuel X-ray diffraction studies of uranium-bearing salts from solidified MSRE fuel and/or concentrate have yielded information of chemical interest. A complete crystal structure analysis of "7IiF*6UF," has shown this tetragonal substance actually to be LiUFs and Lo have a basically different structure from the rhombohedral compounds which do have the 7:6 ratio. Compounds of the latier stoichirmetry do not appear in the present fuel mixtures. In LiUFs5, each U*" ion is coordinated by nine F ionms in the shape off a trigonal prism with each rectangular face bearing a pyramid. This polyhedron is similar to those in UpFs5, but different from those of g-coordinated U* in Ut¥ze It is not presently known whether this dif- ference is significant or Just a coincidence of packing of F~ ions. A determination of the structure of Li,UFg is partially completed: the orthorhombic crystals have unit-cell dimensions a = 2.96 A, b = 9.88 A, ¢ = 5,99 A, and the space group is Pama or InaZy. Four formula weights of Ti,UFg in a unilt cell correspond to a calculated density of 4.71 g/em?. The positions of U%' ions have been determined, but the other ions are yel to be located. Physical Chemistry of Ituoride Melto Vapor Pressure of Fluoride Melts Apparatus wag constructed to obtain vapor pressures by the carrier- gas method in order to determine (1) vapor composition in the LiF-BeFs system (to complement the manomelric pressure data already obtained for this system’), and (2) rare-carth vapor concentrations in equilibrium with liquid mixtures of importance to the molien-salt reactor distilla- tion procec (from these concentrations more accurate decontamination factors for the rare-~earth fission products will be obtained). The apparatus, shown schematically in Fig. 5.1, closely resembles that used by Sense ¢t g&.,g for studies of flucride nmelts. The reliability 129 Table 5.3. Results of Examinations of Specimens Removed from the MSRE O#f-Gas Lines Predominant Specimen Description Refractive Activity at Isctopes Spectrochemical and Qrigin Morphology Index Contact (from gamma Data (t/hr) scan) 1, Deposit from exit orifice of Isotropic particles, 1.520 Li, 99 ug; Be, capillary restrictor appearing as 124 lg; Zr, partly coalesced 100 jig amber globules 2. Scrapings from spool] piece Same as 1 1.540 adjacent to capillaty 3. Deposit from check Isctropic, amber, Be, 0.95 ug; Li, valve 533 varnishelike 2 pgy Zr, <0.5 particle, ™~ 50 x be 100 u 4, Scrapings from valve 522 Isotropic, amber, 1,540 2.5 No Zr, Nb, poppet varnish-like Ce malrix contain~ ing embedded isotropic(?} crystalline material of lower refractive index 5. Scrapings from valve 522 Same as 4 1.544 to 1.5 8981’, 140}33, seat 1.550 1407 4 6. Qil drops Ffrom 522 valve 1,500 No Zr, Nb, body Ce 7. Bcrapings from stainless Isotropic, faintly 1.524 to 14DBa, 14°La, steel filter element colored material, 1,526 msRu, more nearly 13703;. no scale-like than Ce, Zr, Nb glassy in ape~ pearance 8. Metallic scrappings from Granular opaque stainless steel filter particles; low element index, transparent, birefringent crystal- line material spalled off metal on microscope slide 9, Deposit from HV 621 valve Isotropic, faintly ~1.,526 132.p¢ stem colored material, varnish-like in ap- pearance with peb- bly surface 120 ORNL-DWG 6€5-13115 /—MOLECULAR SIEVE DRYING COLUMN / HEATED COPPER TURNINGS — PRESSURE GAGE (O—10in. Hz0) P A THERMOCOUPLE NO 3-_ N|CK$L STAINLESS STEEL SLEEVE ‘ A BO (5Yerin LONG), THERMOGOUPLE (5in LONGH =~~~ J/ NO. 2i , E,—f e g = e L*_% ARGON TO DRYING THERM(OCOUPLE NO1 L U e el ) SALT TEMPERATURE) - MOLTEN 3 J , COLUMN; THEN / SALT - ’ ; ON TO WET / LAVITE ) , ‘ P , TEST METER HIGH TEMPERATURE ALUMINA TUBE -~ NSULATOR / (36-in. LONG) ——T ,,,,, / MARSHALL FURNACE (16-in LONG)- ! i, "~ CONDENSER TUBE {CONDENSER HAS 0.03%-in, CPENING IN END ABOVE CENTER OF SAMPLE BOAT) Fig. 5.1. Transpiration Apparatus. of the apparatus and the efficacy of the washing procedures for remov- ing condensed vapor were tested with pure LiF. Satisfactory agreement with the reliable transpiration data obtained by Sense’ was attained. Three compositions of the Lib-BeF, system have thus far been investigated. The vapor-pressure data are summarized in Table 5.4, based on the assumption that the wvapor consists only of monomeric LiF and BeFp. It should be noted that the apparent partial pressure of LiF increesecs with decreasing concentration of LiF in the melt. This behavior is consistent with the expectation that the vapor species is predominantly a compound of LiF and Bels, such as TLisBeF, or LiBeF,, On the basis of the observed vapor composition, it appears that a liquid compogition of 88-12 mole 9% LiF-Bel's will provide a vapor composi- tion of 67-33 mole % TiF-BeF,. Hence 88-12 should be the correct composi- tion in the still pot for the MSR distillation process., This melt com- position is currently being measured to confirm that the vapor has the composition 67-33 mole % LiF-BelF,. The latter composition would serve as MSBR fuel solvent (this solvent will probably not contain ZrF.). Methods for Predicting Density, Specific Heat, and Thermal Conductivity in Molten Fluorides Densitx.lo Several years ago,t’ after the published data on density of molten fluoride had been examined, it was oroposed that the simple Tule of additivity of molar volumes might be very useful Tor estimating densities of fluoride melts., Since that time, the results of all additlional experimental investigations have been studied. 'The rule of additivity of molar volumes appeared to hold quite well except for one system; this was the NalF-UF, system™? where positive deviabions as great as 6% were obgerved. Thus it appears that, although there may be 131 Table 5.4. Vapor-Pressure Constants, Assuning that the Vapor Phase is Composed of LiF and BeF, log plmm) = A — —2 T(°K) Composition (mole %) Temperature Range p LiF p BeFs e (°c) LiF Bef, A B A B 20 10 8971052 10,370 13,330 10.431 13,890 85 15 8391036 10.43%7 13,330 9.483 12,270 75 25 895-1055 9.720 12,350 8.611 10,710 exceptions, the rule of additive molar volumes describes the experimental data on molten fluorides quite well and remalns the simplest, most accurate method for predicting densities of Tfluoride melts. To improve the method of estimation, a revised set of empirical molar volumes is given in Table 5.5. The origin of these values is dizcussed in ref., 10. For estimating a densilty expression of the form o =a +bt, (1) first solve for densities at two temperatures by using the squation Il - (.M. ) iZ—;:‘L s (2) pt T e e 1 \ v ()] i=1 where Ni and. Mi are the mole fraction and gram-formula weight of com- ponent 1, and Vi(t) is the molar volume of compornent i at temperature t. Substitute molar volumes from Table 5.5 at the twd different temperatures in order to obtain the two values of pt; since density is linear with temperature, substitution of the two pairs of values of p, and t in Eq. (1) provides the solution for the constants a and b. cpecific Heat. Examination®® of the heat-capacity wmeasurements of molten fluorides indicated that the heat capacity per gram-atom of melt is approximately & cal/°C. Hence an expression similar to that of Dulong t 132 Tahle 5.5. Empiricasl Molar Volumes of Fluorides Molar Volume (cc/mole) At 600°C AL 800°C LiF 13.46 14.19 NoF 19.08 20.20 KF 28.1 30.0 RbF 33.9 36.1 Cal 40.2 43,1 BeFs 23,6 XA Mg T 22,4 273, 3 CaFs, 27.5 28.3 ST, 30. 4 31.6 Bal, 35.8 37.3 A1T5 26.9 30,7 YT, 34.6 35.5 TaF, 37.7 38.7 CeT, 36. 3 37.6 Prl, 36.6 37.6 Smf'a 39,0 39. 7rF,, 47 50 ThI', L6.6 477 UF,, 45.5 46" and. Petit may be used to estimate specific heats of fluoride melts. 'The expression 1s n \ 1 g L (Nipi) c = ral e 3 (3) 1 , (NiMi) i=1 where ¢ is the specific heat in cal (°K) * g *, p, is the number of . . i . atoms in a molecule of component i, and Ni and M, are the mole fraction and the gram-formula weight, respectively, of cofiponent 1. 133 Sample calculation: For MSEE coolant, 66-34 mole % LiF-Def,, ) () ), (W) i 0.66(2) + 0.34(3) = 2.34 i 0.66(26) + 0.34(47) = 33.1 , L B(2.34) 4 s o)1 7L ¢ = —55== = 0.57 cal (°K) + g * . Thermal Conductivity. A new semltheoreticsal expression based on Bridgman's theoryl3 of energy transport in liguids has been developed; the derivation of the method is given in ref. 10. The expression is 13 k = -, (4) V2/3 where kX is the thermal conductivity, V i1s the molar volume, and f is the velocity of sound in the melt; all three variables should be in cgs units. To use Eq. (4), the molar volume and velocity of sound are necessary. Molar volume 1s easlly estimated by the rule of additivity as outlined above. The velocity of sound may be estimated from thermo- dynamic quantities, using the expression [(c_/c.) —1lc W2 = v P (5) aZrmd.I where Cp and CV are the molar heat capacities at constant pressure and volume, respectively, @ is the expansivity, T is the absolute ftempera- ture, CP/M is the specific heat [and should be expressed in ergs (°K) 1 g 1], The ratio Cp/CV varies between 1.2 and 1.35 for most fuzed salts (1.2 works quite well). The expanslvity is the negative of temperature coefTicient of density, divided by the density; that is, . ?.9) a = . ,(aT p By using the additivity rule, & is easily estimated. o - Oxide Sclubilities in MSRE Flush Salt, Fuel Salt, and Their Mixtures Bgtimates of the oxide tolerance of MSRE flush-salt—fuel-salt mixtures reported previouslyl4 wvere based upon transpiration meagsure- ments of the following equilibria: HaoO(g) + 2F (d) =207 (a) + 2ur(g) , . 2 271 . . 134 H,0(g) + BeFy(d) +=80(s) + 2HF(g) , = P2 [i . Uy = PHF/PHzO 5 (7) PH,0(g) + ZrF,(d) =%r0,(s) + 4EF(g) Q, = PfiF/(P§20[2r4+]} : (8) By combining the results for reasctions (6) and (7), BeO(s) + 2F(4) = Bely{d) + 0% (a) , L L (9) estimates were obtained for the dissolved-oxide concentration al BeO saturation of 2LiF-BeF, (approximately the composition of MSRE flush salt). For 2LiF-Bels melts containing more than 0.1 mole/kg of Zri, — vherein ZrO, becomes the least-soluble oxide — reactions (6) and (8) were compined in a similar way to obtalin estimates of the concentration oif dissolved oxide at Zr0, saturation: $ir02 (5) + 26 (a) w=2ur¥,(a) + 02 (a) , /2 1/2 _ 4t/ 2rn2 U0, = Qn/Qz' " = [zx ¥ 2[0% ] . (10) However, these estimates were of limited accuracy, owing to difficulties in the determination of QO by the transpiration method. In particular, this quotient was fTound too large ©o be measured with useful accuracy in the MSRE fuel composition, and hence it was necessary to estimate the oxide tolerance of the fuel by extrapolation from results for fuel salt diluted by flush salwv. During the past year, a more direct method for determining these oxide solubilities has been adopted. A measured volume of salt, which previously had been saturated by equilibration with excess solid Zr0O; or BeO, was passed upward through a sintered nickel filter into a heated reaction vessel (Fig. 5.2). There it was sparged with an H,-HF mixture to remove the dissclved oxide as H,0. The effluent HF-U,0-H, mixture was passed through a sodium fluoride column maintained at 90°C to remove the unreacted HF, and then 1t was passed through a Karl Fischer titrstion assembly where the water content was determined. Typlcally, 125-g samples of salt were fillered into the upper vessel and sparged with a gas flow of 150 cmB/min at a temperature of 500°C. The HF partial pressure was 0.07 atm or less. The time requirecd to remove egsentially all of the oxide increased as the ZrF, content of the mixtures was increased, hut generally did not exceed 4 hr. The blank was usually equivalent to less than 0.002 mole per kg of oxide in the sample, 135 The results obtained by this method for simulated flush-salt—fuel- salt mixtures (Fig. 5.3) are in reasonable agreement with previous estimates; % that is, as ZrF, (present in MSRE fuel) is added o 2LiF- BeF, (flush salt), the oxide tolerance at first falls, reaches a mini- mum, and then rises. This behavior reflects a monotonic increase in the solubility product QZrOg with ZrF,; concentration (the dashed lines shown in Fig. 5.3 Indicate the behavior which would be found if QZr02 remained constant). The temperature dependence of the oxide tolerance is indicated in Fig. 5.4 for three compositions: (1) the flush salt, (2) the approximate composition of minimum oxide tolerance, and (3) the Tuel salt. When the present directly measured values of QZr02 are comblned with previocusly determined* values of QZ in the expression QO - (QZrOQQZ)l/2 ? improved estimstes are obtalned for the equilibrium quotient Qg (Table 5.6), corresponding to reaction (6). This quotient should be useful in predicting the efficiency of oxide removal from MSRE fuel salt by HF sparging during fuel reprocessingl” and during oxide analysis,d® The solubility results obtained thus far for BeO in 2LiF-BeF, (Fig. .4) are somevhat above the previous values obtained by combination of B and Q0,14 but these new results showed considerable scatter, This 5 Q ORNL-0OWG €6-4780 PROBE FOR DEPTH MEASUREMENTS . Hp FROM CONSTANT-FLOow e Hy — CONTROL MaF TRAP ‘ ~WET-TEST HE-Hp ‘ METER MIXING ~~REACTION VESSEL o —~ 00015 -in, NICKEL FILTER BUBBLE-O- METER FURNAGE -—» NaOH SOLUTION KARL FISCHER ANHYDROUS HF SOLUTION IN CONSTANT — TEMPERATURE BATH EQUILIBRATION VESSEL Fig. 5.2. Apparatus for Determination of Solubilities by Hydro- Tluorination. 136 ORNL-DWG 66-4781 wt % FUEL SALT IN FLUSH SALT e > 19 20 50 100 P T TR T oo BeO SATURATION « + — & o 700 °C 002 Lo 0.04 0.005 OXIDE CONCENTRATION (moies /kg) OXI{DE CONCENTRATION {ppm) 0002 |~ - 0.001 ‘ - ; _ 0005 Q.01 002 005 Q4 0.2 0.5 1 2 ZrF, CONCENTRATION (moles/kg) Fig. 5.3. Oxide Concentration Required to Produce Precipitation of BeC or ZrO, from Simulated MSRE Flush-Sali—~Fuel-Salt Mixtures. is thought to be caused by the presence of finely divided BeO which was not Tiltered completely from all the samples. These measurements are being repeated with the use of longer equilibrium times, in The hope that this will improve the filterability of the BeC. The oxide tolerance of 2LiF-BeF, (flush salt), indicated by the lines in Figs. 5.3 and 5.4, represents our best estimate at present. Separations in Molten Fluorides vaporative-Distillation Studics on Molten-Salt Fuel Components Vacuum distillation separation of molten-calt fuel or fuel components from the rarc-ecarth fission products is an attractive method of decreas- ing neutron losses by capture. To design process equipment for this task, both the mass rate of distillation and the relative volatility of the rare earths must be known for the particular salt system used. Completed experiments concern the process demonstration planned on MSKEE fuel, but are directly applicable to any proposed thermal MSBR fuel. For MSRE fuel, removal of the uranium by fluorination is proposed. The fuel solvent and remaining fission products would then be fed at the distillation rate to a2 vacuum still charged with LiF-Bel,-ZrF, at that composition which will yield the fuel solvent as distilled product; the rare-carth fission products would concentrate in the still. The regidue would be discarded or procegsed, when necessitated by heat from the Tission products or concentration of rare earths in the product. 137 ORNL~DWG 66-47872 TEMPERATURE (°C) 800 530 500 0.05 0.02 t- FLUSH SALT 0.005 S— - - B — 15wt%FU;?Z;}\\ """ IN FLUSH SALT 100 OXIDE CONCENTRATION {moles /kg) OXIDE CONCENTRATION {ppm) 0.002 | 0.001 - 1.0 14 1.2 1.3 1OOO/T ) Fig. 5.4. Temperature Dependence of Oxide Tolerance: (1) in Flush Salt (BeO-Saturated), (2) in Flush-Salt—Fuel-Salt Mixture of Minimum Oxide Tolerance, and (3) in Fuel Sals. Table 5.6. Summary of Equilibrium Quotients for Reactions (6), (&), and (10) in MSRE Fuel Salt . 2= Tenperature [0<7] QZTOQ Q, QO (°c) (moles;kg) (moles3/kg?) (atm® kg mole™t) (atm mole kg—+) 500 0.012 1.8 % 10 % 7.1 x 10 ? 3.6 x 10 2 600 0.024 7.3 x 10 % 3.95 x 10 * 1.7 % 10 2 700 0.041 2.1 % 10 3 150 5.6 x 10 2 138 A 10-ml graphite cylinder, containing ~17 g of salt with a free surface (when molten) of 1 om®, was used for the still pot. This cylinder fitted into a "[I' shaped Hastelloy N tube heated electrically to a given temperature as measured by four thermocouples in the graphite cylinder. When the desired temperature was reached, distillation was initiated by evacuating the assembly. Vaporized sall then passed up and over the top of the "O" with a small portion (attributed to thermal reflux) collecting at the base of the graphite cylinder. The product salt condensed in a cooler (~450°C) collecting cup in the opposite leg. Distillation was stopped by helium addition after preselected time periods. Mass-rate data for ‘LiF weve determined first, and then MSEE solvent was added to the 'LiF and distilled in aliquot portions from it. Bach repetition brought the BeFp and ZrF, concentrations in the pot closer to those which would yield fuel solvent (LiF-BeFs-ZrF,, 65-30-5 mole %) as product. After the equilibrium concentration was approached, 2200 ppm of neodymivm (as NAFs) was added to the still bottom, and several distillate samples were taken. Then the neodymium concentration was raised to 22,000 ppm, and the sequence was repeated. For both "TiF and the nominal eguilibrium composition, the effect of temperature on mass rate was determined. The data of Fig. 5.5 show the effect of temperature and mass rate; single experiments show considerable scatter, so inclusive bands are chown. The slope of the bands is consistent with the heat of vaporiza- tion of the components and strongly suggests that ebullition does not occur. Solvent surface heat flux ig <1/100 of the available heat to the graphite cylinder; it is doubtful that the distillation is heat limited. The fact that the observed rate for ‘LiF is only ~10% of -2 ORNL-DWG GG-966 (g/cmzv sec) ._,_ - = DISTILLATION RAT s 100°C 1000=C 72 73 74 715 76 7.7 7.8 7.9 fig. 5.5. Observed Distillation Rate vs lTemperature. 139 Theoretical is unexplained, but it has been observed in many distilla- tions using several experimental configurations. The solvent for any proposed MSBR should distill at rates above that shown for ‘LiF with the posslble exception of Th¥F,-bearing systems., The effect of ThF, is being studied. Bffectivenesse of separation from rare earths was determined by activation danalysis for neodymlwia in the product, the still bottom, and the refluxed salt deposited around the base of the graphite. The data are shown in Table 5.7. The equilibrium composition in the still pol undoubtedly changes with temperature due to the change in activity coefficients of the components. The compobjtiom found at nominal equilibrium after several solvent additions of 5 to 7% by weight and afier distillations at 1030°C wag LiF-BeF,-ZrF, (85,4-10.7 3.9) in the presence of 22,000 ppm of neodymium as {luoride. mifective Activity Coefficients by Evaporative Distillation Evaporation into a vacuum from a quiescent wmolten-salt mixture of low vapor pressure should not permit vapor-liquid equilibrium at the surtace of the melt; transport from the surface should be controlled by the evaporation rates of the individual components, The amount vaporized is a function of The equilibrium vapor pressure, molecular mass, temperature, and surface area; according to Langmuir,l7 _ grams of 7 T W, = 5—-—-~—-—-2 - KP, \fMZ/'I‘ , (11) cme -gec where P, 15 vapor pressure of component 4, M is molecular weight of component Z, T is absolute temperature, and K is a constant dependent on units. In a multicomponent system at equilibrium, P, = 7ZNZP7 s (12) W, o= w = Ky _N_ Q/ /T (13) Z cme -sec Z 7 IT we use a fixed mechanical configuration and, for simplicity, define the activity coefficient of the major component (in this cas LiF) as unity, the activity coefficient For anmy other constituent mdy be determined by using the ratio W Ky, M, ) NM, /T 1 gy KON, P7 MLlr*/ T 140 Table 5.7. Councentration of Neodymium in Fractions from Vacuum Distillations - e Nd Concentration Fraction (ppm) 2200 ppm Added 22,000 ppm Added Still bottom 2570 22,000 Product 21 6OOb ReTlux 133 34 aMeaa values from several determinations; data show little scatter exceplt for reflux specimens. bThese product samples alsc showed Ce and La, which undoubtedly represent contamination during grinding and handling of specimens; Nd analysis, therefore, may well be toc high, altered to N_o.. W e — y _ L1k . 7 . LiF . JI\JLLD /M . (15) 2y W pC 1z 7 LiF z Proceeding with this technique, the data shown in Fig. 5.6 have been taken from LiF-BePFs-ZrF, melts. TFor comparison purposes, selected data from other sources are included,t820 Although internal consistency exists, the values are dependent on vapor-pressure data for the pure components. Calculations were made using the values (PO) for the pure components shown in Table 5.8, The leagt-precise measurements are from MSRE-colvent distillations; thesc are included to substantiate the surmise that when ZrF, is included in the melt, it effectively removes LiF from the solvent. The 1000°C LiF-Bel’; line has been extended to 100 mole % IiF, since the initial LiF-BeFo-ZrF, experiment contained <0.35 mole % ZrF,, a guantity so small that the system can be assumed to be Li¥-BeFs. On the other hand, as 7ZrF, builds up in the mixed system (to approximately 3 mole %), en- hancement of The BeF,; activity dis noted; this is consistent withh re- sults from MSERE solvent (LiF-BeFs-ZrF,, 65-30-5 mole % initially). The temperatures reported are those of the bulk melt, and 1t is rec- cgnized that the surface temperature may be significantly lower, causing an indeterminate {(for the present) error. Both surface-tempersa- ture effects and the pure LiF-ZrF, system are being studicd. It is interesting to note that the assumption of unit activity coefficient for TiF does not lead to incompatibility with the data of others. 141 Txtraction of Rare Farths from Molten Fluorides into Mcolten Metals Studies of the removel of rare earths from a molten-fluoride solvent and their dissolution in molten metals have been oriented toward the development of a liquid-liguid extraction process for re- moving rare-earth fission products from molten-salt reactor fuels. Fluoride mixtures obtained by dissolving a selected rare earth iato LiF-BelFy (66-34 mole %) have been usged to simulate the fuel solvent of the reference-degign MSBR. In fuel reprocessing schemes proposed for the reactor, uranium will be removed prior to the removal of non- O ORNL-DWG 66-267 Y COEFFICIENT i ACTIVIT 1000°C e LiF - BeF,, BAES, 700°C 200%¢ Lik- ZrF,, SENSE ef af. —rr—2 Lif - Bel, + Zr¥, , BAES — = =~DERIVED FROM MSRE SOLVENT DISTILLATIONS 900 ~1050°C LiF - Bef, , BUCHLER, 600°C LLiF - BeF,, 1000°C LiF«8eF,, 900°C LiF - BeF, » ZrFy , 1000°C 5 70 75 a0 g5 30 95 100 LiF IN MELT (mole Ys) Fig. 5.6. Effective Activity Coefficients Caleculated from BEvapo- rative-Distillation Data for MSBER Scolvent Compositions. 142 a,b,c Table 5.8. Vapor Pressures for FPure Fluorides Vapor Pressure Compound () 1000°¢ 200°C LiF 0. 47 0.072 Bels 65 12.0 LrE, 2700 780 “Handbook of Chemistry and Physics, 44th ed., p. 2438, Chemical Rubber Publishing Co., Cleveland, Ohlo, 19562. b B. Porter and E. A. Browa, J. Am. Ceram. Soc. 43, 49 (1962). C : . 8. Cantor, personsal communication. volatile [ission products. When this simualated fuel mixture is contacted with a molten bismuth-lithium mixture, rare earths are reduced to their elemental valence states and are dissolved in the molten-metal phase. Ixperimental studies conducted thus Tar have examined the distribution of rare earths between the two liquid phases as functions of the lithium concentration in the metal phase. BSubsequent experiments will elaborate on these findings and will examine the back extraction of rare earths from the liquid metal into a second salt mixture. The removal of rare earths from the reactor fuel, followed by thelr concen- tration in a solid, disposable salt mixture will constitubte the re- procegsing method. Fluoride starting materials were prepared in nickel equipment by adding sufficient rare-earth {luoride to approximately 2 kg of the gol- vent, LiF-BeF, (66-34 mole %), to attain a rare-carth concentration of about 10 % m.f. in the salt phase. When possible, a selected radio- isotope of the rare carth was added for analytical purpcses. The mixture wag treated with an HF-H; mixture at 600°C to remove oxide impurities and at 700°C with H, alone to reduce concentrations of structural-metal difluorides in the fluoride melt. The bismuth was purified by sparging the 2.35-kg batch with Hp at 600°C to remove oxide impurities. The bismuth purification was carried out in the experimental extraction vessel of type 304L stainless steel lined with low-carbon steel. Follow- ing the materials-preparation phase of the experiment, the fluoride mix- ture was transferred as a liquid to the extraction vesgsel. ILithium metal was added directly to the metal phase without prior contact with the salt phase; this was done through a loading port that extended to near the bottom of the extraction vessel, TLithium, Tor incremental additions to the experiment, was freshly cut and weighed under mineral 0il, affixed to a small-diameter steel rod, rinsed in benzene, and dried in the flowing inert atmosphere of the loading port prior to its 143 insertion and dissolution in the molten bismuth. Filtered samples of each phase were taken under assumed equilibrium conditions (approxi- mately 2-hr periods) after each addition of lithiwn. Radiochemical analyses of each phase for rare-ecarth gamma sctivity and spectrographic analyses of the metal phase for rare-earth and lithium concentrations provided data for calculating the distribution of the rare earth in the system and its dependence on the lithium concentration of the metal phase. In experiments conducted thus far, the distributions of lanthanum, cerium, neodymium, samarium, and europium in the extraction system have been examined separately, but under essentially comparative conditions. A summary of thege results, illustrated in Fig. 5.7, shows that a mixture of bismuth containing 0.02 m.f. of lithium metal sufficed for removing egsentially all cerium, lanthanum, and neodymium and substantial guantities of samarium and curopium from the barren fuel solvent. 1In all of the experiments, rare earths that were reduced from solution in the salt phase were found as dissolved components of the metal phase. Bach incremental addition of lithium to the system resulted in a near-linear increase in the concentration of lithium found in the metal phase, However, this dissolved quantity accounted for only 25 to 50% of the amount added to the system. The results of a blank extraction experiment in which rare earths were omitted from the system could not be distinguished from those obtajined when rare earths were present. Fach incremental addition of lithium to an extraction system was an approximate threefold excess over the amount required for the stoichio- metric reduction of all of the rare earth contained in the experiment. Wi SAMARIUM 29 CERIUM FITa00—— o LANTHANUM —| N MEQDYMIUM = FUROPIUM L5 z\z |z 300 v — : 4 L S g | < [ =T wilfur Ll i 05 water electrolysis cell, and the soda-lime trep for cleaning the effluent gas stream. These components are shown in Fig. 5.19, where the cover plate has been removed. The modular construction was chosen in hopes that any maintenance necessary can be performed remotely. If not, any component can be removed manually in 2 to 5 min if the hot cell is sufficiently decontaminated. Figure 5.20 shows the disassembled hydrofluorinator. The copper sample ladle is contained in the nickel liner, which fits into the bottom of the hydrofluorinator. The baffles on the spring-loaded sparge tube fit inside the liner and confine the molten salt to the liner during the bubbling operation of the analysis. This arrangement allows the solid sample, after analysis, to be removed from the bottom of the hydrofluorinator and then to be disconnected at the Swagelock fitting NaF TRAP (UNDER VALVE PLATE) - CAPILLARY SPLITTER COMPARTMENT Fig. 5.19. Top View of Valve Compartment of Hot-Cell Apparatus. 161 PHOTO 82187 Wi AMOLVHOBY T YN [ ) P P | 3 OlLLYN Disassembled Hydrofluorinator. Hir, 5.2, 162 and discarded along with the ladle, liner, and bottom of the sparge tube. This design affords considerable saving in cost per analysis, because it permits the repeated use of the complete hydrofluorinator. Figure 5.21 shows the Hp-HF mixer section, which is located in the access area behind the hot cell. The HF pressure is regulated by con- trolling the temperature of the HF tank shown at the left. The flow rates of the Hr and HF are controlled by nickel and platinum capillaries respectively. This apparatus also contains safety interlocks and limit- ing capillaries to prevent release of HF and excessive pressures inside the hot cell. Figure 5.22 shows the master control panel, which is located in front of the hot cell. This panel contains the moisture recorder (not shown), helium and hydrogen pressure controls, HF control switch, hydro- fluorinator coupler switch, temperature recorder, furnace power controls, and power control for the coupler-line heater. After they were fabricated, the apparatus and auxiliary equipment were assembled in the hot-cell mockup, and all mechanical operations were performed successfully with Master Slave manipulators. The equip- ment was then transferred to the laboratory and was assembled on the bench top for trial analyses of samples. Individual components were tested exhaustively and, when necessary, were modified or replaced to obtain maximum dependability and to minimize hot-cell maintenance. A final check of the splitter and cell was made by injecting a known quantity of water into the flow system; quantitative recovery of the water resulted. The entire system was then checked out by analyz- ing fuel-salt samples. Satisfactory results were obtained. The equipment is now being transferred to the hot cell and installed therein. Voltammetric Determination of Ionic Iron and Nickel in Mclten MSRE Fuel By controlled-potential voltammetry, Fe2+ and Ni2+ were determined in a 100-g sample of MSRE fuel. The sample was withdrawn from the re- actor in enricher ladles and was transferred under an inert atmosphere to the graphite crucible of an electrochemical cell assembly for remelt- ing and analysis. The cell assembly and electrodes developed for electrochemical studies of molten-fluoride salts are described else- where 29721 The iron and nickel in the molten fuel are suspected to be in the metallic state, as well as in the form of soluble ionic species. However, only in the divalent oxidation state are these metals electro- reducible in the melt and can thus give voltammetric reduction waves. By voltammetry and by the standard-addition technique, the concentration of Fe?” was determined to be ~10 ppm. The concentration of Ni2™ was be- low the limit of detection by voltammetry (<1 ppm). Average total concentrations of iron and nickel, determined by conventiocnal methods, are about 125 and 45 ppm respectively. Thus it appears that most of the iron and nickel in the fuel is present in the metalliic state, probably as finely divided particles. 163 PHOTO 82182 Fig. 5.21. Hp-HF Mixer Section of Hydrofluorinator Apparatus. 164 PHOTO 82189 Fig. 5.22. Main Control Panel of Hydrofluorination Apparatus. 165 Chromium concentration, determined conventionally, is ~50 ppm. Interference from uranium prevents the voltammetric determination of chromium in the molten salt. A well-defined wave was oObserved that corresponds to the reduction U(Iv) —U(II1). Possibly, this wave can be used to continuously monitor uranium in the system; more work is contemplated in this area. Development and Evaluation of Equipment and Procedures for Analyzing Radioactive MSRE Salt Samples As stated earlier,32 statistical evaluation of the control data indicated that a negative bias of approximately 0.8% existed in the coulometric uranium results. The uranium procedure consisted primarily of six steps. . Determination of blank. . Preparation of sample. Prereduction of sample at +0.125 v vs S.C.E. Reduction of uranium and copper at —0.325 v vs S.C.E. Oxidation of copper at +0.125 v ve S.C.E. . Calculation. O W In steps 1, 3, 4, and 5, the titration was allowed to proceed until the cell current had decreased to 5 Ha. The calculation was performed using the following formula: [(A=B) —C — (bt/E)IDE = micrograms of uranium per test solution |, where A = readout voltage for uranium and copper reduction in miliivolts, B = blank of electrolyte in millivolts, C = readout voltage for copper oxidation in millivolts, D = 1.233 x 10 ? ug of uranium per microcoulomb, E = coulometer calibration constant for 10-microequivalent range in microcoulombs per millivolt, b = background current in microamperes (5 na), t = time of uranium and copper reduction in seconds. In an attempt to eliminate the bias, steps 1, 4, and 6 were modified. The determination of the blank was eliminated. The cutoff point for the uranium and copper reduction was changed from 5 pa to 50 pa. An X-Y recorder (Fig. 5.23) was attached to the coulometer to monitor the titrations. When the readout voltage and cell current were plotted on the X and Y scales, respectively, a straight line with a negative slope was obtained after the recorder pen had reached a maximum Y value. By extrapolation of the curve (Fig. 5.24), it was possible to obtain the readout wvoltage corresponding to a cell cur- rent of O amp. Due to the error introduced by this technique, a poten- tiometer was connected to the integrator circuit of the coulometer in order to obtain more precise readout voltages. This necessitated terminating the titrations at a specific end point. A cell current of 166 PHOTO 82047 Fig. 5.23. High-Sensitivity Coulometric Uranium Apparatus. 50 pa was found empirically to be the most practical point of termina- tion. The coulometer was automatically turned off when the ammeter needle indicated a cell current of 50 pa. At this point, the titration was approximately 99% complete. This portion of the readout voltage was taken from the potentiometer; the remaining readout voltage was obtained by extrapolating from 50 to O pa an expanded cell current vs readout voltage curve (Fig. 5.25). The two readings were combined to obtain the total readout voltage for the uranium and copper reduction, Based on more than 200 titration curves, the readout voltage represented by the extrapolated portion of the curve was approximately 5 mv; there- fore, the readout voltages used for the uranium and copper reductions were oObtained by adding 5 mv to the potentiometer readings. Due to the above changes, the formula for calculating the uranium present in the test solution then became [(A — B) + CIDE = micrograms of uranium per test solution , where = readout voltage for uranium and copper reduction in millivolts, readout voltage for copper oxidation in millivolts, = 5 mv (value obtained by extrapolation of the cell current vs readout voltage curve from 50 to O ua}, = 1.233 x 10 ° pg of uranium per microcoulomb, = coulometer calibration constant for 10-microequivalent range in microcoulombs per millivolt. Qe I = | 167 The uranium present in the test solution could be calculated, since the voltages read out on the potentiometer were proportional to the coulombs required in the reduction of uranium and copper and the oxidation of copper. Sample Analyses From December 16, 1965, through January 26, 1966, six flush-salt and twenty-two fuel-salt samples were received in the High-Radiation- Level Analytical Laboratory. All 28 samples were analyzed for U, Zr, ORNL~-DWG €66-4784 S 5 g '\ '_ = Lt o S S 4 ] - Lt Q 3 A\ 2 , \ 0 0 01 0.2 0.3 0.4 05 0.6 VOLTAGE Fig. 5.24. Cell Current vs Readout Voltage Curve. 168 ORNL-DWG 66-4785 1.6 © £ 12 |_ = w [1 g o 3 o 08 | W (0] OA; \ EXTRAPQLATED VOLTAGE POTENTIOMETER VOLTAGE o | o 0.1 0.2 0.3 0.4 0.5 VOLTAGE Fig. 5.25. A Portion of the Cell Current vs Readout Voltage Curve. Cr, Be, F, Fe, and Ni. All six flush-salt samples and the first twelve fuel-salt samples received were analyzed for molybdenum. Quality Control Program The quality control program initiated prior to precritical sampling was continued during the past period. Synthetic soluticns similar to dissolved nonradioactive fuel-salt samples were analyzed along with each flush-salt and fuel-salt sample. Due to the relatively smali number of samples analyzed to date, the accumulated control data is insufficient tc calculate the true percent standard deviations of the methods. The values shown in Teble 5.13 were obtained during the fourth quarter of 1965. The 2S5 and average values shown were obtained by four different groups of shift personnel. Molybdenum values are not shown in the table since it was not added to the synthetic solution. The nickel values indicate that a positive bias exists in the method. The values given for the coulometric uranium procedure indicate that the negative bias was eliminated. 169 Table 5.13. Control Program, Fourth Quarter of 1965 Average Determination Shift Num?er ?f Added Value F%und 25 Percentage Determinations (ug/ml) : (hg/ml) Coulometric, uranium A 19 656.6 658 1.11 B 40 656.6 656 1.55 C 30 656.6 659 0.57 D 34 656.6 658 0.29 Amperometric, zirconium A 26 1117 1122 7,27 B O 1117 1077 13.53 C 14 1117 1119 5.17 D 0 Amperometric, chromium A 22 11.8 12.202 9.15 B 16 11.8 12.278 11..06 C 8 11.8 12.520 16.84% D 5 11.8 12.270 14.38 Colorimetric, iron A 18 6.36 6.286 2.30 B 10 6.36 6.355 10.91 C 16 6.30 6.302 5.51 D 3 6.36 6.253 2.40 Colorimetric, nickel A 16 6.08 6.713 3.92 B 10 6.08 6.748 6.24 C 18 6.08 6.669 14.32 D 0 Neutron acti- vation, beryllium A 3 847 853 1.08 B 0 C 2 847 849 1.83 D 4 847 856 1.48 18. 19. 20. 21. 22. 23. 2. 170 References MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 321. See "Development and Evaluation of Equipment and Procedures for Analyzing Radiocactive MSRE Salt Samples,” this report. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 117. See "Voltammetric Determination of ITonic Iron and Nickel in Molten MSRE Fuel," this report. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 140. See "Oxide Solubilities in MSRE Flush Salt, Fuel Salt, and Their Mixtures,” this report. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp. 123-25. K. A. Sense et al., J. Phys. Chem. 58, 223 (1954). K. A. Sense and R. W. Stone, J. Phys. Chem. 62, 1411 (1958). S. Cantor, Reactor Chem. Div. Ann. Progr. Rept. Dec. 31, 1965, ORNL-3913, pp. 2729, 5. Cantor, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1962, ORNL-3262, pp. 38-41. E. A, Brown and B. Porter, U.S. Bur. Mines Rept. Invest. 6500 (1964). P. W. Bridgman, Proc. Am. Acad. Arts Sci. 59, 162 (1923). MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, pp. 129— 38, MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 152. Ibid., pp. 140-43, T. Langmuir, Phys. Rev. 2 (Ser. 2), 329 (1913) (and subsequent papers). Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp. 59-62. K. A. Sense and R. W. Stone, J. Phys. Chem. 62, 96 (1958). A. Buchler and J. L. Stauffer, Symposium on Thermodynamice with Emphasis on Nuclear Materials and Atomic Transport in Solids, Vienna, July 22-27, 1965, paper SM-66-26, p. 15. J. H. Shaffer et al., Nuel. Sci. Eng. 18, 177 (1964). Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1961, ORNL-3127, p. 8. Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, p. 56. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp. 137 41 . 25. 27 28' 29, 30. 31. 32. 171 MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 109, Fig. 5.2. MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 140. MSR Program Semiann. Progr. Rept. Feb. 28, 1965, ORNL-3812, p. 162. G. Goldberg, A. 8. Meyer, Jr., and J. C. White, Anal. Chem. 32, 314 (1960). D. L. Manning, J. Electroanal. Chem. 6, 227 (1963). D. L. Manning and G. Mamantov, J. Electrosnal. Chem. 7, 102 (1964). D. L. Manning, J. Electroanal. Chem. 7, 302 (1964). MSR Program Semlann. Progr. Rept. Aug. 31, 1965, ORNL-3872, p. 148. 172 6. MOLTEN-SALT BREEDER REACTOR DESIGN STUDIES Design and evaluation studies have been made of thermal molten-salt breeder reactors (MSBR) in order to assess their economic and nuclear po- tential and to identify the ilmportant design and development problems. The reference reachor design presented here contains design provlems re- lated to molten-sallt reactors in general. The MSBR reference design concept i1s a two-region, two-fluld system, with fuel sall separated from the blanket salt by graphite tubes. The fuel salt consists of uranium fluoride dissolved in a mixture of lithium- beryllium fluorides, while the blanket salt is a thorium-lithium fluoride containing 27 mole % thorium fluoride. The energy generated in the re- actor fluid is transferred to a secondary coolant-salt circuit, which couples the reactor Lo a supercritical steam cycle. On-site fluoride volatllity processing 1ls employed, leading to low unit processing costs and economic operation as a thermal breeder reactor. MSBR Plant Design Flowsheet Figure 6.1 gives the flowsheet of the 1000-Mw (electrical) MSPR power plant. Fuel flows through the reactor al a rate of about 44,000 gpm (velocity of about 15 fps), entering the core at 1000°F and leaving at 1300°F. The primary fuel circuit has four loops, each loop having a pump and a primary heat exchanger. Racn of these pumps has & capacity of sbout 11,000 gpm. The four blanket-salt pumps and heat exchangers, although smaller, are similar to corresponding components in the fuel system. The blanket salt enters the reactor vessel at 1150°F and leaves at 1250°F. The blanket-salt pumps have a capacity of about 2000 gpm. Four 14,000-gpm coolant pumps circulate the sodium fluoroborate coolant salt, which enters the shell side of the primary heat exchanger at 850°F and leaves at 1112°F. After leaving the primary heat exchanger, the coolant salt is further heated to 1125°F on the shell side of the blanket~salt heat exchangers. The coolant then circulates througn the shell side of 16 once-through superheaters (4 superheaters per pump). In addition, four 2000-gpm pumps circulate a portion of the coolant through eight reheaters. The steam system flowsheet is essentially that of the new TVA Bull Run plant, with modifications to increase the rating to 1000 Mw (elecm trical) and to preheat the working fluid to 700°F prior to entering the heat-exchanger—superheater unit. A supercritical power conversion sys- tem 1is used, which is appropriate for molten-sall application and takes advantage of the high~strength structural alloy employed. Use of a super- critical fluid system results in an overall plant thermal efficiency of about 45%. ORNL-DWG 66-4786 FUEL _ NET OUTRUT 1000 Mwe . BLANKET ~—-— GROSS GENERATION 1034.9 Mwe A ——s ——— COOLANT —--— BF BOOSTER PUMPS 9.2 Mwe \ | STEAM ————- STATION AUXILIARIES 25.7 Mwe |‘ (58,5 10.067* 7152 % ; H,0 o REACTOR HEAT iNPUT 2225 Mwt jlocoer T T | * 07 ib/hr ; - 550 | S . , NET HEAT RATE 7601 Btu/ kwhr | 3309 I [ | | 1518.5h -540p P psia ) e 570p~650°F 551.7°F NET EFFICIENCY 49% | o e _ 220 ! o h Btu/b REACTOR VESSEL (1) I T - : - g '_P i | 1000%F —3— FREEZE 2225 Mw: : 3600p, REHEAT ' —ry I | | VALVE ! ; 00051 STEAM ‘ | | | : | razan | PREHEATERS(S)‘ | 00p_ | o ; i | 100.5 Mwt | 2587 | \ | | 350Cp I BLANKET SALT FUEL SALT | | | B65°F I PUMPS (4) PUMPS (4} ! ) | 1307.8h ! 2000 gpm |, L, 14,000 gom ++_ J . a d___ o F e (NOM) EACK (Now) EAcH | \ | ) ! -4 1P TURBINE 1-527.2 NMwe}— t q72e” i 57.352"% | | ; GROSS : 17.3 fts/sec\ \ | I ; ; : . PREHEATER COOLANT | : BOiLER-SUPERHEATER i ! 1180° F P SALT PUMPS (4) : ; ‘ : — | 0°F L = o o i , CODLANT SALT PUMPS (4) T Em —————- e ¥ § ; 14,000 gpm (NOM) EACH N LP e L fi Y 1250°F o | {NOM)EACH i ; : ! TURBINE | 7‘52;'(7)5?& ‘ IS [T e - 850°F ; i ' S K LT . Vi | ™ j 7 850°F 1 Ig SLANKET SALT HEAT 1125°F | ‘ L E CONDENSER AND i1 1| EXCHANGERS (4) : L - 1125°F | 1769.2h -3800p FEEDWATER SYSTEMS ’: T 1 Mt (oL ) P > —— - | 200° 7 | SEE STEAM SYSTEM o ‘. o ; : : FLOWSHEET) L ) 1 S BCILER- : ) FUEL SALT HEAT ; REHEATERS | — 850°F SUPEREREATERS ! i ] | EXCHANSERS (4} ) N (8) - : i (i6! : U 3500 p - 548.6°F : 2114 Mwi (TOTAL)- T . 293.5 Mwi LA 1934.5 Mwt N 546.3h I 7 — . —_— . ——— _— + T 1/ 3475p-695°F - 766.4 h BLANKET SALT i DRAIN TANKS (2] ! BOILER FEEDWATER i PRESSURE BOOSTER FUSL SALT COOLANT SALT PLUWPS (2} { DRAIN TANKS {4) Fig. 6.1. DRAIN TANKS i4) Flowsheet of MSBR Power Plant. 20,000 gpm (NOM) EACH 4.6 Mwe EACH el 174 Reactor Design Figure 6.2 shows a plan view of the MBBR cell arrangement. The reactor cell is surrounded by four shielded cells containing the super- heaters and reheater units; these cells can be individuvally isolated for maintenance, The processing cell, located adjacent to the reactor, is divided into a high-level and a low-level activity area. Figure 6.3 shows an elevation view of the reactor and indicates the position of egquipment in the various cells. Figure 6.4, a plan view of the reactor cell, shows the location of the reactor, punps, and fuecl and blanket heat exchangers. Figure 6.5 is an elevation of the reactor cell. The Hastelloy N reactor vessel has a side wall thickness of aboutb 1-1/4 in. and a head thickness of about 2-1/4 in.; 1t is designed to operate at 1200°F and 150 psi. The plenum chambers, with 1./4-in.-thick walls, communicate with the external heat cxchangers by concenbric inlet-outlet piping. The inner pipe has slip joints to accomumodate thermal expansion. Bypass flow through these slip Jjoints is about 1% of the total flow. As indicated in Fig. 6.5, the heat exchangers arc suspended from the top of the cell and are located below tThe reactor. kach fuel pump has a free fluid surface and a storage volume which permit rapid drainage of fuel fluid from the core upon loss of flow. In addition, the fuel salt can be drained to the dump tanks when the reactor is shut down for an ex- tended time. The entire reactor cell is kept at high temperature, while cold "fingers" and thermal insulation surround structural supporl members and all special equipment which must he kept at relatively low tempera- tures. The control-rod drives are located above the core, and the con- trol rods are inserted into the central region of the core. ORNL-DWG 66-785 | REHEAT STEAM *W N FEED H,0- T T 30 H.P. STEAM T He sTEAM--—1 | FEED Hp0- - i T I.P. STEAM -~ WASTE GAS CELL ~FUEL HEAT ¢ : EXCHANGER ! 4 — =z =D COOLANT SALT PUMPS- { \ iy 20 —==\ ) ; 5 5 [ e 2 { 5 i£; fi” = . | £ R l: 1fi," ,,,,, - r—m Lo REACTOR ____b_mgfl_p. ¢ BLANKET HEAT EXCHANGER’ STORAGE v 26 30 FUEL PROGESSING CONTROL AREA ALL DIMENSIONS ARE IN FEET DECONTAMINATION i.....—_. o e e e ,180 [, — - U Fig. 6.2. Plan View of MOSBR Cell Arrangement. 175 ORNL-DWG G&6-733 COOLANT FUEL CIRCULATING CONTROL ROD DRIVE /" saLT PUMP ! eumies BLANKET CIRCULATING PUMP - CONTROL ROOM LEVEL — 58 ft ANALYTICAL LAB LEVEI—" BLLANKET HEAT EXCHANGER (F?EHEATERS Fig. 6.3. EBElevabtion View of MSBR Reactor Arrangement. ORNL.-DWG 66-4787 STEAM GENERATOR CELL BLANKET HEAT EXCHANGER - — BLANKET REGICN | _—— CORE REGION e T DL CONTROL ¢f ROD # DRIVES [ A ’ii 9 c‘COOLANT SALT TO SUPERHEATERS AND REHEATERS ‘ ; S EAT EXCHANGER -COOLANT SALT SUPPORT STRUCTURE RETURN LINE PRIMARY HEAT EXCHANGER 36-ft-diom REACTOR & CELL A Fig. 6.4. Plan View of MSBR Reactor Cell Showing Location of Re~ actor Equipment. 176 CHNL.-DWG 86-794 FUEL PUMP MOTOR \ » CONTROL ROD DRIVES /BLANKET PUMP MOTOR i / 3 / - CONSTANT SUPPORT HANGERS . - A0 -ft-digm CCRE — REACTOR VESSEL — FUEL SALT DISTRIBUTION PLENUMS FUEL DUMP TANK ‘ WITH COOLING I3 COILS FOR AFTER HEAT REMOVAL — 1. 42 1 — T~ BLANKET HEAT EXCHANGER PRIMARY HEAT J EXCHANGER-unfigTJW -5 in, - — 19 ft \ REACTOR CELL HEATERS Fig. 6.5. Elevation of Reactor Cell. The reactor vessel, about 14 £t in diameter by about 15 ft high, contains a 10-ft-dlam core assembly composed of reentry-type graphite fuel cells. The graphite tubes are attached to the two plenum chambers at the bottom of the reactor with graphite-to-metal transition sleeves. Fuel from the entrance plenum flows up fuel passages in the ouber region of the Tuel cell and down through a single central passage to the exit plenum. The fuel flows from the exit plenum to the heat exchangers, then to the pump, and back to the reactor. A l«l/2-ft—thick molten-salt blanket plus a 1/7~ft—thick grapnite reflector surround the core. The blanket salt also permeates the interstices of the core lattice so that fertile material flows through the core without mixing with the fisgile fuel salt. The MSBR requires structural integrity of the graphite fuel cell. In order to reduce the effect of radiation damage, the fuel cells have bee 177 made small to reduce the fast-flux gradient across the graphite wall. Also, the cells are anchored only at one end to permit axial movement. The core volume has been made large in order to reduce the flux level in the core. In addition, the reactor i1s designed to permit replacement of the entire graphite core by remote means if required. Figure 6.6 shows a cross section of a fuel cell. Fuel fiuid flows upward through the small passages and dowmward through the large central passage. The outside diameter of a fuel cell tube is 3.5 in.; there are 534 of these tubes spaced on a 4.8-in. trisngular pitch. The tube assem- blies are surrounded by hexagonal blocks of moderator graphite with blanket salt filling the interstices. The nominal core composition is 75% graphite, 18% fuel salt, and 7% blanket salt by volume. A gummary of parameter values chosen for the MSBR design is given in Table 6.1. Fuel Processing The primary objectives of fuel processing are to purify and recycle fissile and carrier components and to minimize fissile inventory while holding losses to a low value. The [luoride volatility—vacuum distilla- tion process fulfills these obJjectives through simple operations. The core fuel is conveniently processed by fluocride volatility and vacuum distillation. DBlanket processing is accomplished by fluoride volatility alone, and the processing cycle time is short enough to main- tain a very low concentration of [issile material, The effluent UFg is absorbed by fuel salt and reduced to UF, by treatment with hydrogen to reconstitute a fuel-galt mixture of the desired composition. Molten-salt reactors are inherently sulted to the design of process- ing facilities integral with the reactor plant; these facilities require only a small amount of cell space adjacent to the reactor cell. Because all services and equipment avallable to the reactor are available to the rrocessing plant and because shipping and storage charges are eliminated, integral processing facilities permit significant savings in capital and operating costs. Also, the processing plant inventory of fissile material is greatly reduced, resulting in low fTuel inventory charges and improved fuel utilization characteristics for the reactor. The principal steps in core and blanket stream processing of the MSER are shown in Fig. 6.7. A small side stream of each fluid is con- tinuously withdrawn from the Tuel and blanket circulating loops and is circulated through the processing system. After processing, the decon- taminated fluids are returned to the reactor at some convenient point — for example, via the fuel and fertile stream storage tanks. Fuel inventories retained in the processing plant are estimated to be aboulb lO% of the reactor system inventory for core processing and less than l% for blanket processing. 178 ORNL-DWG 66-4788 FUEL PASSAGE (DOWN) 3Y%-in. OD FUEL TUBE MODERATOR (GRAPHITE ) FUEL PASSAGE (UP) —-—MODERATOR HQLD DOWN NUT (GRAPHITE) REACTOR P:_;’- . 6: r J,//T‘ SPACER L L ,_L/~f—_—' “““ “METAL TO GRAPHITE SLIP-JOINT i%jjr’ fifi/e#”‘""' METAL TO GRAPHITE BRAZED JOINT [ E.u H =TT -BRAZED JOINT 221z2) \ FUEL INLET PLENUM ~-w FUEL QUTLET PLENUM = 77 ] Fig. 6.6. Cross Section of a Fuel Cell. S SORBERS / e ~ COLD e FR‘AP ey ORNL-DWG 65-6194 3 IOO 400 NaF /MgF, /F2 NGF/;/ // MoF‘/ /—/70" // g/ 7 /44 {_/4? EXCESS (LS C’F\'ODJCTION /‘ 4 UF, + < VOLATILE FP MAKE UP LiF/Bes,/ThE, < % FERTILE / MAKE up/ / CONTINUOUS . FLUORIDE VOI_MI_.TY / %501;// / - LiF/BeF, / Tni, /FP [~ DiSCARD FOR FP REMOVAL s ;o | FERTILE STREAM RECYCLE ] Fig. 6.7, MSBR Core UR RECYCLE TO REACTOR y SORBEDS 70 //’ //// o /// N e a VL MygF, S =-70°C o 2 Mgra 100-200°C /E// v WASTE STORAGE NoF/MgF,/ o UF, + VOLATILE FP MAKE U LiF/BeF, v o o POLDJP//J CONTINUOUS |/ i |/ / FOR 7t FLUORIDE // VAZUUM / DISTILLATE | ", U= UR /// SPENT . :D DC\;AY//VOLATI._ATY 7 // DisT |=_;_ATfO'\|/ ~ LJi-/BeFZ/// /t\(’ED\JC~I ION / FlLTP\AT]ON 71 ~B50°C /// 1000°c/ ~500 ’// 550- 600°C///// U+, =UF, jflmmhg //// S / o) i’ / A h & LiF + RARE —H, EARTH FP - REDUCED METALS Cr, Fe, N LiF/Befy/UFq RECYCLE and Blaunket Stream Processing Scheme. 641 180 Table 6.1.. Parameter Values of MSBR Design Power, Mw Thermal 2220 Blectrical 1000 Thermal efficiency 0.45 Plant factor 0.80 Dimensions, ft Core height 12.5 Core diametler 10.0 Blanket thickness Radial 1.5 Axial 2.0 Reflector thickness 0.25 Volumes, £t> Core 982 Blanket 1120 Volume fractions Core Fuel salt 0.169 Fertile salt 0.0735 Moderator Q.7575 Blanket Fertile salt 1.0 Salt volumes, £t° Fuel Core 166 Blanket 26 Plenums 147 Heat exchanger and piping 345 Processing 33 Total 717 Fertile Core 72 Blanket 1121 Heat exchanger and piping 100 Storage (protactinium.decay) 2066 Processing 24 Total 3383 Salt compositions, mole % Fuel LiF 63.6 Bel'o 36.2 UF, (fissile) 0.22 Fertile LiF 71.0 Bel's 2.0 Thi'y, 27.0 Ul (fissile) 0.0005 181 Tmne6J,hmmflmmd) Core atom ratios Th/U 41,7 c/u 5800 Fissile inventory, kg 769 Fertile inventory, thousands of kilograms 260 Processing by fluoride volatility Fuel Fertile Stream stream Cycle time, days 47 23 Rate, ©t2/day 14.5 14 Unit processing cost, $/ft3 183 6.85 Heat Exchange and Steam Systems The structural material for all components contacted by molten salt in the fuel, blanket, and coolant systems, inecluding the reactor vessel, pumps, heat exchangers, and piping and storage tanks, is Hastelloy N. The primary heat exchangers are of the tube-~and-shell type. Hach shell contains two concentric tube bundles connected in series and at- tached to fixed tube sheets. The fuel salt flows downward in the outer section of tubes, enters a plenum at the bottom of the exchanger, and then flows upward to the pump through the center section of tubes. hn- tering at the top, the coolant salt flows on the baffled shell side of the exchanger down the central core, uander the barrier that separates the two sections, and up the outer annular section. Since a large temperature difference exists in the two tube sec- tions, the tube sheets at the bottom of the exchanger are not attached to the shell. The design permits differential tube growth between the two sections without creating troublesome stress problems. To accom~ plish this, the tube sheets are connected at the bottom of the exchanger by a bellows-type joint. This arrangement, essentially a floaling plenum, permits enough relative motion between the central and outer tube sheets to compensate for differential tube growth withoubt creating intolerable stresses in the joint, the tubes, or the pump. The blanket heat exchangers increase the temperature of the coolant eaving the primary core heat exchangers. BSince the coolant-salt temper- ature rise through the blanket exchangers is small and the flow rate is relatively high, the exchangers are designed for a single shell-gide pass for the coolant salt, although two-pass flow is retained for the blanket salt in the tubes. Straight tubes with two tube sheets are used. The superheater is a U-tube U-shell exchanger using disk and donut baffles with varying spacing. It is a long, slender exchanger having 182 relatively large baffle spacing. The baffle spacing is established by the shell-gide pressure drop and by the tewmperature gradient across the tube wall and is greatest in the central portion of the exchanger, where the temperature difference between the flulds is high. The gsupercritical fluid enters the tuve side of the superheater at 700°F and 3800 psi and leaves at 1000°F and 3600 psi. The reheaters transfer energy from the coolant salt to the working Tfluid before its use 1n the intermediate-pressure turbine. A shell-tube exchanger is used, producing steam at 1000°F and 540 psi. Since the freezing temperature of {he secondary salt coolant is about 700°F, a high working-fluid inlet temperature is reguired. Preheaters, along with prime fluid, are used in raising the tempersture of the work- ing fluid entering the superheaters. Prime fluid goes through a preheater exchanger and leaves at a pressure of 3550 psi and a temperature of about g70°F. It is then injected into the feedwater in a mixing tee, producing fluid at 700°F and 3500 psi. The pressure is then increased to about 3800 psi by a pressurizer (feedwater pump) before the fiuid enters the super- heater. Capital Cost Hstimates Reactor Power Plant Preliminary estimates of the capital cost of a 1000-Mw (electrical) MSBR power station indicabe a direct construction cost of about $80.4 miliion. After applying the indirect cost factors used in the advanced converter evaluation,l an esbimated total plant cost of $113.6 million is obtained. A summary of plant costs is given in Table 6.2. The con- ceptual design was not sufficiently detailed to permit a completely re- 1iable estimate; however, the design and estimates were studied thoroughly enough to make meaningiul comparisons with previous converter-reactor plant cost studies. The relatively low capital cost estimabe obtained results from the small physical size of the MSBR and the simple control require- ments. The results of the study encourage the belief that the cost of an MGBR power station will be as low as for stations utilizing other re- actor concepls. The operating and maintenance costs of the MSBR were not estimated. Baged on the ground rules used in ref. 1, these costs would be about 0.3 mill/kvhr (electrical). uel Recyecle Plant The capital costs associated with fuel recycle equipment were ob- tained by itemizing and costing the major process eguipment required and by estimating the costs of site, bulldings, instrumentation, waste dis- posal, and building services associated with fuel recycle. 183 Table 6.2. Preliminary Cost-Estimate Summarya for a 1000-Mw (Blectrical) MSBR Power Station Federal Power Commission Account (thousandgozgsdollars) 20 Land and land rightsb 360 21 Structures and improvementis 211 Ground improvements 866 212 Buildings and structgres .1 Reactor building 4,181 .2 Turbine building, auxiliary building, 2,832 and feedwalter heater space .3 Offices, shops, and laboratories 1,160 4 Waste disposal building 150 .5 BStack 76 .6 Warehouse 40 .7 Miscellaneous 30 Subtotal account 212 8,469 Total account 21 9,335 22 Reactor plant equipment 221 Reactor equipment .1 Reactor vessel 1,610 .2 Control rods 250 .3 Shielding and containment 1,477 4 Heating-cooling systems and 1,200 vapor-suppression system .5 Moderator and reflector 1,089 .6 Reactor plant crane 265 Subtotal account 221 5,801 222 Heat transfer systems .1 Reactor coolant system 6,732 .2 Intermediate cooling system 1,947 .3 Steam generator and reheatersg 92,853 .4 Coolant supply and treatment™ 300 .5 Coolant salt inventory 354 Subtotal account 222 19,186 223 Nuclear fuel handling and storage 1,700 (drain tanks) 225 Radioactive waste treatment and disposal 450 (off—gas system) 226 Instrumentation and controls 4,500 227 Feedwater supply and treatment 4,051 228 Steam, condensate, and FW piping 4,069 229 Other reactor plant eguipment 52000e (remote maintenance) Total account 22 bty , B4T 184 Tsble 6.2 (continued) 23 Turbine-generator units 231 Turbine-generator units 19,174 232 Cilrculating water system 1,243 233 Condensers and auxiliaries 1,690 234 Central lube oil system 80 235 Turbine plant instrumentation 25 236 Turbine plant piping 220t 237 Auxiliary equipment for generator 66 238 Other turbine plant equipment 25 Total account 23 22,523 24 Accessory electrical 241 Switchgear, main and station service 550 242 Bwitchboards 128 243 Station service transformers 169 244 Auxiliary generatox 50 245 Distributed items 2,000 Total account 24 2,897 25 Miscellaneous 800 Total direct construction cost?® 80,402 Total indirect costs 33,181 Total plant cost 113,583 YEstimates are based on 1966 costs, assuming an established molten- salt nuclear power plant industry. b . . . ; Land costs are not included in total direct construction costs. °MSER containment cost is ineluded in account 221.3. Yssumed as $300,000 on the basis of MSRE experience. “The ample MGBR allowance for remote mainbtenance may be too high, and some of the included replacement equipment allowances could more logically be classified as operating expenses rather than first capital costs. fBased_ on Bull Run plant cost of $160,000 plus ~37% for uncertainties. ®Does not include account 20, land costs. This is included in the in- direct cosis. Table 0.3 summarizes the direct construction costs, the indirect costs, and total costs associated with the integrated processing facility having approximately the required capacity. The operating and maintenance costs for the fuel recycle facility inelude labor, labor overhead, chemicals, utilities, and maintenance materials. The tobal annual cost for the capacity considered here (15 62 of fuel salt per day and 105 712 of fertile salt per day) is esti-~ mated to be $721,230, which is equivalent to about 0.1 will/kwhr (elec- trical).® A breakdown of these charges is given in Teble 6.4. 185 Table 6.3. Summary of Processing-FPlant Costs 1000-Myw {Electrical) MSBR Processing-FPlant Expendituresg Costs Installed process equipment $ 853,760 Structures and improvements 556,770 Waste storage 387,970 Process piping 155,800 Process instrumentation 272,100 Electrical auxiliaries 84,300 Sampling commections 20,000 Service and ubtility piping 128,060 Insulation 50,510 Radiation monitoring Total direct costs Construction overhead (30% of direct costs) Total construction cost Engineering and inspection (25% of total construction cost) Ssubtotal plant cost Conbingency (25% of subtotal plant cost) Total plant cost 100,000 $2,609,270 782,780 3,392,050 848,010 4,240,060 1,060,020 $5,300,080 Table 6.4. Summary of Operating and Malntenance Charges for Fuel Recycle in a 1000-Mw (BElectrical) MSBR Operation and Maintenance Annual Charges Expenditures Direct labor $222,000 Labor overhead 177,600 Chemicals 14,640 Waste containers 28,270 Utilities 80, 300 Maintenance materials Site 2,500 Services and utilities 35,880 Process equipment 160, 040 Total annual charges $721,230 186 Nuclear Performance and Fuel Cycle Analyses The fuel cycle cost and the fuel yield are closely related, yet independent in the sense that two nuclear designs can have similar costs but significantly different yields. The objective of the nuclear design calculations was primarily to find the conditions that gave the lowest fuel cycle cost, and then, without appreciably increasing this cost, the highest fTuel yield. Analysis Procedures Calculation Method. The caleculations were performed with OPIIMERC, a combinaticon of an optimization code with the MERC multigroup, diffusion, equilibrium reactor code. The program MERC? calculates the nuclear per- Tormance, the equilibrium concentrations of the variocus nuclides, in- cluding fission products, and the fuel cycle cost for a given set of con- ditions., OPTIMERC permits wp to 20 reactor parameters to be varied, within limits, in order to determine an optimum, by the method of steepest ascent. The designs were opbimized essentially for minimum fuel cycle cost, with lesser welght given to maximizing the annual fuel yield. Typical param-~ eters varied were the reactor dimensions, blanket thickness, fracticns of Tuel and fertile salts in the core, and fuel and Tertile stream proc- esging rates. beveral equations were included in the code for approximating cer- tain capital and operating costs that vary with the design parameters (e.g,, capital cost of the reactor vessel, which varies with the reactor dimensions). These costs were automatically added to the fuel cycle cost in the optimization routine so that the optimization search would take into account all known economic factors. However, only the fuel cycle cost itself 1s reported in the results. Modified GAM-1-[HERMOS cross-section libraries were used to compute the broad group cross sections for these calculations. It was assumed that all nuclides in the reactor system are at their equilibrium con- centrations. To check this assumption, a typical reactor design was examined o determine the operating time required for the various uranium isotopes to approach their equilibrium concentrations from a startup with 235y, It was found that 222U and 227U were within 95% of their equilibrium concentrations in less than two years. Uranium-234 was withia 95% of equi- librium after elght years, while 2367 was within 80% after ten yecars. oince the breeding performance depends mainly on the ratio of 233y 4o 2357 in the fuel, the equilibrium calculation appears to be a good rep-~ resentation of the lifetime performance of these reactors, even for start- Up on 2337, Basic Assumptions Economic. The basic economic assuuptions employed in the calcula- tions are given in Table 6.5. The values of the fissile isotopes were taken from the current AEC price schedule. 187 The procesgssing coghs are based on those given in the seclion entitled "Capital Cost Estimates" and are included in the fuel cycle costs. The capital and operating costs were estimated separately for each stream as a function of plant throughput, based on the volume of salt processed. The total processing cost is assumed to be a function of the throughput to some fractional power called the scale Tactor. Processing. The processing scheme is that indicated in Fig. 6.7. A Tissile material loss of O.l% per pass through processing was assumed. In addition to the basic processing scheme employed, results were also obtained for the case where protactinium can be removed directly from the blanket stream. The improvement in performance under these circumstances is a measure of the incenbive to develop protactinium re- moval ability. Fission Product Behavior. The disposition of the various fission products was assumed as shown in Table 6.6. The behavior of 135%e and other fission gases has a significant influence on nuclear performance. A gas stripping system is provided to remove these gases from the fuel salt. However, part of the xenon could diffuse into the moderator graph- ite. In the calculations reported here, an 135%¢ poison fraction of 0,005 was assunmed. Corrosion Product Behavior. The control of corrosion products in molten-salt fuels does not appear to be a signiflcant problem, and the effect of corrosion products was neglected in the nuclear calculations. The processing method considered here can conbrol corrosion product buildup in the fuel. Table 6.5. Basic Economic Assumptions Reactor power, Mw (electrical) 1000 Thermal efficiency, % 45 Load factor .80 Cost assumptions Value of 222U and 22°Pa, $/g 14 Value of *3°U, $/g 12 Value of thorium, $/kg 12 Value of carrier salt, $/kg 26 Capital charge, annual rave, % Plant 12 Nondepreciating capital, in- 10 cluding fissile inventory Processing cost, dollars per cubic foot of salt Fuel {at 10 ft3/aa§) 228 Blanket (at 100 f£t2/day) 8.47 Processing cost scale factor 0.4 (exponent ) 188 Table 6.6. Disposition of Fission Products in MSBR Reactor and Processing Systems Flements present as gases; assumed to be partly absorbed by graphite and partly removed by gas stripping (1/2% polsoning assumed) Elements which plate out on metal surfaces; as- sumed to be removed instantaneousl. 5 Elements which form volatile fluorides; assumed to be removed in the fluoride volatility process Elements winich form stable fluorides less vol- atile than LiF; assumed to be separated by vacuum distillation Elements which are not separated from the car- rier salt; assumed to be removed only by salt discard Kr, Xe Ru, Bh, Pd, Ag, In e, Br, Nb, Mo, Tec, Te, T sr, Y, Ba, la, Ce, Pr, Na, Pm, Snm, Eu, G4, Tb Rb, Cd, Sn, Cs, Zr Table 6.7. MSBR Performance Fuel yield, %/year Breeding ratic Fissile losses in processing, atoms/fissile absorption Neutron production per fissile absorption, ne Specilic inventory, kilograms of fissile material per Mw (electrical) Specific power, Mw (thermal) per kilogram of fissile material Power density, core average, kw/liter Gross In fuel salt Neubtron flux, core average, 10*% neutrons e ? sec™t Thermal fast Fast, over 100 kev Thermal flux factors, core, peak/mean Radial Axial Fraction of fissions in fuel stream Fraction of fissions in thermal neutron group Mean n of 233y Mean n of 235y 4 .86 1.049 0.0057 2.221 0.769 2.89 80 473 6.7 12.1 3.1 2.22 1.37 0.987 0.806 2.221 1.958 ed 189 Nuclear Design Analysis The important parameters describing the MSBR design are given in Table 6.,1. Many of the parameters were basically fixed by the ground rules for the evaluation or by the engineering design. These include the thermal efficiency, plant factor, capital charge rate, maximum fuel velocity, size of fuel tubes, processing costs and fissile loss rate, and the out-of-core fuel inventory. The parameters which were optimized by OPTIMERC were the reactor dimensions, the power density, the core composition, including the C/fi and Th/U ratics, and the processing rates. Nuclear Performance. The results of the caleculations for the MSBR design are given in Table 6.7, and the neutron balance is given in Table 6.8, The basic design has the inherent advantage of no neutron losses Table 6.8. MSBR Neubron Balance Neutrons per Fissile Absorption Material Absorbed Absorbed Produced Total by Fission 232, 0.9710 0.0025 0.0059 233p, 0.0079 233y 0.9119 0.8020 2.0233 234y 0.0936 0.0004 0.0010 2335y 0.0881 0.0708 0.1721 2361; 0.0115 0.0001 0.0001 237y 0.0014 238y 0.0009 Carrier salt (except 611) 0.0623 0.0185 614 0.0030 Graphite 0.0300 135ye 0.0050 149gm 0.0069 151qy 0.0018 Other fission products 0.0196 Delayed neutrons lost® 0.0050 Leakageb 0.0012 o L Total 2.2209 0.8828 2.2209 aDelayed neutrons emitlted ocutside the core. bleakage, including neutrons absorbed in the reflector. 190 to structural materials cother than the moderator. Except for some un- avoidable loss of delayed neutrons in the external fuel circult, there is almost zero neutron leakage from the reactor because of the thick blanket. The neubtron losses o Tission proeoducts are minimized by the availability of rapid and inexpenslve integrated processing. Fuel Cycle Cost. The components of the fuel cycle cost for the MoBR are given in Table ©6.92. The main components are the fissile inventory and processing costs. The inventory costs are rather rigid for a given reactor design, since they are largely determined by the assumed external fuel volume. The processing costs are, of course, a function of the processing cycle times, one of the chiefl parameters optimized in this study . SBR Performance with Protactinium Removal Scheme. The ability to remove protactinium directly from the blanket of the MSER has a marked effect on fuel yield and fuel cycle cost. This isg due primarily to the marked decrease in protactinium necutron absorptions when protactinium is removed from the blaonket region. A simple and inexpensive scheme for the removal. of protactinium from the blanket would give the MSBR the perfor- rmance indicated under MSBR (Pa) in Table 6.10; for comparison, the results without protactinium removal are also given in the table. Table 6.9. Fuel Cycle Cost for MSBR Costs (mills/kwhr) Fuel Fertile Total Grand Stream Stream obe Total Inventory Fissile” 0.1180 0.0324 0.1504 Fertile (0.0000 0.0459 0.0459 Salt 0.0146 0.0580 0.0726 Total 0.2690 Replacement Fertile 0.0000 0.0185 0.0185 Salt 0.0565 0.0217 0.0782 Total 0.0967 Processing 0.1102 0.0411 0.1513 Total 0.1513 Produetion credit 0.0718 Net fuel cycle cost 0.4452 a'I:acluding 233Pa, 233U, and 237U, 191 Power Cost and Fuel Utilization Characteristics Based on the above, the power cost, specific fissile inventory, and fuel doubling time Tor the MSBR and MSBER (Pa) are summarized in Table 6.11. Table 6.11 illustrates the economic advantage of MSBR's as nuclear power plants. Also, the fuel utilization characteristics as measured by the product of the specific inventory and the square of the doubling time? are excellent. On this basis the MSBR is comparable to a fast breeder with a specific inventory of 3 kg/Mw (electrical) and s doubling time of 10.5 years, while the MSBR (Pa) is comparable to the same fast breeder with a doubling time of 6 years. Table 6.10, Comparison of MSER Performance With and Without Protactinium Removal MSBR, Without MSBR (Pa), Protactinium with Protactinium Process Removal Fuel yield, %/year 4 .86 7.95 Breeding ratio 1.049 1.071 Fuel cycle cost, mills/kwhr 0.45 0.33 Specific inventory, kg/Mi 0.769 0.681 (electrical Specific power, Mw (thermal) /kg 2.89 3.26 Neutron production per fissile 2.221 2.227 absorption, mne Volume fractions, core Fuel 0.169 0.169 Fertile 0.0745 0.0735 Moderator G.7565 0.7575 oalt volumes, 3 Fuel Core 166 166 External 547 551 Total 713 717 Fertile Total 3383 1317 Core atom ratios Th /U 39.7 41.7 ¢/u 5440 5800 192 Table 6.11. Power Cost and Fuel Utilization Characteristics of the MSPR and the MSBR (Pa) Cost [mills/kwhr (electrical)] MSBR MSER (Pa) Capital cost™ 1.95 1.95 Operating and maintenance costb 0.30 0.30 Fuel cycle cost® 0.45 0.33 Total power costh 2.70 2.58 Specific fissile inventory, Q.77 0.68 kg/Mw (electrical) Fuel doubling time, years 20.6 12.6 a’.]_2% fixed charge rate, 80% load factor, 1000-Mr (electrical) plant. b \ . . - Nominal value used in advanced converter evaluation (see ref. l). C. . . | . . . . Costs of on-site integrated processing plant are included in this value. References 1. M. W. Rosenthal et al., A Comparative Evaluation of Advanced Con- verters, ORNL-3686 (January 1965). 2. C. D, Scott and W. L. Carter, Preliminary Design Study of a Con- tinuous Fluorination—Vacuum Distillation System for Regenerating Puel and Fertile Streams in a Molten Salt Breeder Reactor, ORNL- 3791 (January 1966). 3. T. W. Kerlin, Jr., et al., The MERC-1 Equilibrium Code, ORNL-TM-847 (Apr. 22, 1964). 4. P. R. Kasten, "Nuclear Fuel Utilization and Eccnomic Incentives,” paper presented at the American Nuclear Society Meeting, Nov. 1518, 1965, Washington, D.C. 193 7. MOLTEN-SALT REACTOR PROCESSING STUDIES A close-coupled facility for processing the fuel and fertile streams of a molten-salt breeder reactor (MSBR) will be an integral part of the reactor system. Studies are in progress for obtaining data relevant to the engineering design of such a processing facility. The processing plant will operate on a side stream withdrawn from the fuel stream, which circulates through the reactor core and the primary heat exchanger. For a 1000-Mw (electrical) MSBR, approximately 14.1 ft? of salt per day will be processed, which will result in a fuel-salt cycle time of approxi- mately 40 days.t The probable method for fuel-stream and fertile-stream processing is shown in Fig. 7.1. The salt will first be contacted with F, for re- moval of uranium as volatile UFg. Purified UFg will be obtained from the fluorinator off-gas (consisting of UFg, excess Fp, and volatile fission product fluorides) by use of NaF sorption. A semicontinuous vacuum distillation will then be carried out on the remaining salt for the re- moval of the rare earths, barium, stroontium, and yttrium. These fission products will be removed from the still in a salt volume equivalent to 0.5% of the stream. A small fraction of salt may also have to be dis- carded at some stage in the process for removal of fission products such CRN|-DWG 65-1801R24A Fp RECYCLE e Eongs LD MgF, AP SORBER 8! #tday NoF SORBER 27 -day CYCLE 100-400¢C BLANKET SALT SPENT NaF +MgF, l"’ NaF + MgF, Ufg F, LiF - BeF,- UF, 15 f1%/day WASTE . FERTILE SALT LiF - BeF,~UF, Lif MAKE-UP -—— UFe 046 #t3duy ‘ LiF+BsF, PRODUCT U= Z?jdg/day WASTE = 0059 f1%/day TLi=078 kg May FP's Fig. 7.1. MSBER Fuel and Fertile Stream Processing. 194 as zirconium, rubidium, and cesium. The barren salt, the purified U¥g, and the makeup salt will then be recombined. This step involves re- duction of UFg to UF,, mixing of these streams, and sparging the result- ant material with an H,-HF stream. Finally, the salt mixture will be filtered before return to the reactor. Semicontinuous Distillation The present concept of the distillation step in the MEBR processing plant will use a continuous feed stream and vapor removal; however, there will be a buildup of less volatile fission products (FP) in a static pool of liguid in the still, with perlodic discard.’ Fission products will be allowed to build up in the still liquid until the heat generation rate becomes excesslive, or until the liquid FP concentration becomes too large for useful decontsmination. One measure of the decontamination achieved in the distillation process is the relative volatility of the nonveolatile fission products as compared with the carrier salt. The relative volatility of a non- volatile component A compared with a more volatile component B in & mix- ture of A and B is defined as Va/%, 0 = @ 5%, v where QhB = relative volatility of A compared with B, = vapor-phase mole fraction, x = liguid-phase mole fraction. To achieve good decontamination from the less volatile P, the relative volatility must be small. For systems in which the FP concentration is small, the relative volatility can be approximated by o 3 which is the Henry's law constant, HA’ used in the expression Yo = g% - (3) Thus, determination of the Henry's law constant or relative volatiility for each of the nonvolatile FP will be sufficient for determining the size and operating conditions for the distillation step. The importance of relative volatility in determining the operating characteristics of a distillation system is shown by the following calcu- lation. Consider a material balance of an FP in the proposed distillation 195 process. The amount of FP fed into the still per unit time (I ) must equal the FP leaving the still per unit time (Dy), plus the rate of change of FP in the still liquid a(vx)/at: a{Vx) (4) where x = mole fraction of FP in liquid, Xo = injet mole fraction of FP, = megs feed rate in moles/unit time, = vaporization rate in moles/unit time, mass of 1liquid holdup in moles, = mole fraction of FP in vapor, ¢ — THERMOCOUPLE 101160 E§§§§1 \ Y in. A et — 5/g—in. NICKEL TUBE L R I B _E§§§§\ / 'lg 7 1-in. NICKEL TUBE ; ; E 5 |/4i|"l. RYAERE I C < b 7 S THERMOWELLS / 4 N2 , 2 ‘ ” i 9 7 in. ; | / , 4 / ~—THERMOWELL [ - 1 1] 2 ] / AN \\\WNSULAHON Fig. 7.3. BEquilibrium Still with Cold Finger. 199 Table 7.1. Relative Volatilities of Rare-FKarth luorides in ILithium Fluoride Rare~Farth Tiguid Mole Average Relative Volatillities Fluoride Fraction 900°C 950°C 1000°¢C 1050°C Cely 0.0067 0.133 0.167 0.208 DmE 5 0.01 0.033 0.009 NdF 3 0.01 0.025 0.016 PrF 3 0,001 0.038 0.020 0.014 ks 0,001 0.041 0.037 0.028 0.012 CeF3 0.01 0.043 0.033 0.018 Lafs 0.001 0.035 0.024 LaF3 0.01 0.051 0.027 0.011 0.008 Fuel Reconstitution A necessary step in the processing of the MSBR fuel is the recombl- nation of the purified uranium hexafluoride with the purified carrier salt, which includes reduction of UFg, the product of the [luorination step, to UF,. The usual method for reducing Uy to UFy uses excess hydrogen in an Hp-Fp flame which produces hydrogen flucride ag a by- product. The resulting UF, powder is collected at the base of a tall reaction vessel. Although this operation has veen reduced to rouvine production, it appears undesirable for radiochemical application because of the inherent solids handling problem. An alternative 1ls the reduction of Uf'g to UF, in a molten salt, involving only gases and liquids. When UF; ig contacted with a molten flvoride salt containing UF,, it is absorbed with reaction to form intermediate fluorides of uranium such as UFs. These intermediate fluorides can then be reduced to UF, by contacting the salt with hydrogen. Initial but definitive tests have sghown this alternative to be quite feasible. A tower which might serve well to conduct this gequence of reactions is shown in Fig. 7.4. Questions of feasibility are raised concerning the eguilibriunm distribution of the various species of uranium fluorides and the rate at which the reactions proceed. It is believed that the addition of Ulg to a molten salt containing UM, results in the formation of dissolved FTluorides of uranium with a valence intermediate between 4+ and 6+. This pehavior 1s indicated by the fact that quantities of F, sufficlent for the Formatlon of UFs can be absorbed by molten salt containing UF, with- cut the evcolution of UFg. Similar behavior 1s also noted in reactions between UF, and UFg in the absence of molten salt to yield intermediate 200 fluorides such as UgF3,. It could be expected that the homogeneous re~ action rate would be very rapid and that the absorption rate would proba- bly depend upcn diffusion to and from the interface. Previous data on the reduction of uranium fluorides intermediate between UF, and Uy in molten salts do not exist; however, rate data may be inferred from the reduction of UF% with hydrogen in molten mixtures of IiF and Bel', per- formed by Long. He observed thal the ratio of the concentrations of hydrogen and HF in gas bubbles rising through the molten salt reached equilibrium in only a few inches. His data also indicate only l% re- duction of UF, to Ul; by a gas stream containing 1% HF in hydrogen at pressures of 1 atm at 600°C. The experimental equipment consisted of a reaction vessel in which molten salt containing UF, could be contacted with a metered stream of UFg, HF, Hp, or N, and Na¥ traps to collect UFg and/or HF in the off-gas (Fig. 7.5). Two NaF traps were provided downstream of the vessel; one ORML~DWG 65— 4794RA H, + HF SALT + UF, s g FUELL SALT I “"fi'-'-"-—HE 5 MAKE~ UP 2 SALT o ; UFe BARRENSALT”“J“”““J——"—££;2:}-—_——} PUMP Fig. 7.4. Continuous Reductilon of Ulfg by Hs in a Molten Salt. ORNL-DWG 65— 30984 Ho No UFG vNZ - OFF - GAS 4 —-in. NICKEL PIPE NaF BEDS REDUGCTION VESSEL Fig. 7.5. Equipment Used in Reduction of U¥g to UF,; 1n a Molten Salt. 201 trap was used only for trapping UFg from the vessel off-gas during UFg addition to the molten salt, and the other trap was used for all other HF or UFg absorption. The reduction vessel was constructed from 4-in.- diam sched-40 nickel pipe and was 26 in. long. A 3/8-in. nickel inlet line was located in the center of the vessel and terminated 1/4 in. from the bottom of the vessel. A 3/4—in. fitting on the top flange allowed the insertion of a cold, 3/8-in. nickel rod, which was used for sampling the salt. A 3/8-in. off-gas line was connected to the top flange. The vessel was heated by two Nichrome-wire resistance furnaces. Three experiments were carried out at 600°C in which UFg was intro- duced at the rate of 1.5 g/min at a polnt 12 in. below the surface of a molten IiF-ZrF, mixture containing ~0.5 mole % UF,. The initial salt charge consisted of 5320 g of ZrF,, 863 g of LiF, and 61.8 g of UFy (0.197 g-mole of UF,) and had a melting point of approximabely 510°C. Complete absorption of the UFg was observed during each of the tests, which resulted in the sgbsorption of a total of 147 g of UFg during a period of 98 wmin. During a typical run, the salt charge from the previous run was heated to 600°C and sparged with Np for 15 min at the rate of 100 cm?/min (51P), after which a salt sample was taken. The salt was then sparged with HF at the rate of 0.5 1b/hr for 1 hr and with N, for 15 min, after which a second salt sample was taken. Uranium hexafluoride was then bubbled into the salt at a rate of 1.5 g/min for a specified length of time, with the vessel off-gas passing through an Nal bed used exclusively during this period. The salt was then sparged with Np for 12 min and sampled. The salt was then sparged with Hp at the rate of 95 cmB/min (8TP) for 30 min and sampled, after which the Hp sparge was continued for an additional 30 min. Two questions related to the experimental work are of primary in- terest. These are (1) the fraction of UFg which was absorbed by the molten salt and (2) the valence of the uranium in the resulting mixture. It was concluded that, within the accuracy of the experimental data, complete absorption of the UFg by the molten salt had occurred. TNo uranium was found on the NaF trap used during the UFg addition period. The concentration of Uy in the salt sample taken after UpF; addition was below the limlt of detection of 0.05 wt %. Reductlon of the uranium to UF, probably occurred during the addition of UFg by the reaction of the intermediate {luorides with nickel from the vessel wall. From these tests it appears that Ul could be rapldly absorbed by molten fluoride salt containing about 1 wt % Uf, at 600°C and thal the intermediate fluoride formed could be reduced with hydrogen to UF,. Sub- sequent studies are recommended to provide more guantitative data for engineering design; however, this form of recombination of UFg with the purified carrier salt will be indicated on all subsequent [lowsheets. 202 Continuous Fluorination of a Molten Salt Since the presence of uranium in the distillation step to separate the carrier galt from the Tission products would cause unnecessary compli- cations, it is removed continuously in a prior fluorination step. Pre- vious experience with the removal of uranium from molten salt by fluori- nation includes the operation of the Molten-Salt Fluoride Volatility Pilot Plant at ORNL.% In this facility, batch fluorinations completely volatilized the uranium as UFg, which allowed its subsequent purification and recovery by absorption and cold trapping. Observed corrosion in this facillity was severe but acceptable in a batch process of this sort. How- ever, it would be intolerable in a continuous unit, or in any unit with enough capacity to handle the processing stream for an MSBR. A possible solution to the corrosion problem is the cperation of the fluorination vessel, presently envisioned as a tower, with a layer of frozen salt on the vessel wall. FExperience with this type of system was obtained with batch fluorinations made at Argonne National Laboratory5 in support of the molten-salt fluoride volatility process. Successful tesis were also made at ORNL using ohmic heating to provide the internal heat generation, where it was found that a gas flow could be maintained through an unheated line which entered the Tluorinator vessel at a point below the molten- salt surface when a frozen salt layer was present on the fluorinator wall. 1In application to the MSBR fluorinator, internal heat generation will be provided by the fission product decay heat. Experimental studies of continuous fluorination of molten salt are being made in a l-Iin.~diam nickel column with a salt depth of 48 in. MNo provision is being made in the preseanl experimental work for corrosion protection by a frozen layer of salt (Fig. 7.6). Fluorination tests in which 15 cmB/min of molten salt (NaF-IiF-ZrF,) containing 0.5 wt % UFg was contacted countercurrently with 70 cm®/min of F, {STP) at 600°C showed removal of uranium from the salt at 96 to 92.4% efficiency during a l-hr period of continuons operation. Material balances were compli- cated by the inevitable corrosion of the nickel vessel. Complete removal of uranium from the salt with no corrosion would yield, for the above corn~ ditions, a UFg concentration of 17.56 mole % in the off-gas. Observed con- centrations ranged as high as 35 mole % UF. These results indicate that subsequent development can be expected to produce acceptab.le recoveries of uraniwn by continuous fluorination. Chromivm Fluoride Trapping At the conclusion of tests on the MSRE, uranium will be recovered from the fuel salt as Ulg by sparging the salt with F,. Fluorides of chromium will be present in the fuel salt as a result of corrosion of reactor plping and of equipment used for hydrofluorination or fluorination of the salt. A potential problem associated with the recovery of the uranium is the presence of volatile fluorides of chromiuvm {(Cr¥, and CrFs) 203 ORN(. - DWG 65—-9354A INFRARED ANALYZER FOR UFg ANALYSIS SALT SAMPLING METERED L VESSEL SODA LIME TRAP {.5=in. diam MaF TRAP 7 kg e 1.25 kg METERED Ny FOR SALT DISPLACEMENT 1t 1 ’ ’ b oFF-cas SALT RECEIVER SALT FEED TANK 14 liters 14 liters NICKEL FLUQRINATOR 1—in. diam 72-in iong Fig. 7.6. DEguipment for Removal of Uranium from Molten Salt by Con- tinuous Fluorination. in the fluorinator off-gas; these fluorides cannot only contaminate the U product but also render equipment inoperative by deposition in lines, valves, etc. A study has been completed which will permit the design of a trapping system for removing these fluorides from the oiff-gas, which will also contain Ufg and Fso. Ixperiments were carried out in which 1 1iter/min of Fs (STP) was sparged through a molten NaF-1iF-ZrF, mixture at 650°C which contained 0.5 to 4 wt % Crf3. The resulbing off-gas contalning fluorine and vola- tile fluorides of chromium then passed through beds of pelleted Nal' at 400°C for removal of chromium fluorides. In some tests, a UFg flow of 100 em?/min was added to the Fi. It can be concluded that (1) fixed beds of NaF at 400°C are effective in removing fluorides of chromium from a gas stream which also contains UFe and Fo; (2) pelleted NaF having a surface area of 0.074 w?/g and a void fraction of 0.277 is superior to material having a surface area of 1 m?/g and a void fraction of 0.45, and has an effective capacity of about 20 g of chromium per 100 g of NaF; (3) uranium losses to the 400°C NaF bed of less than 0.01% are achievable when working with a gas stream that containg 0.4 mole of CrFs per mole of UFg dn Fp. Deslgn and Evaluation Study A preliminary design study has been made of a conceptual processing plant to treat irradiated fuel and fertile streams from the 1000-Mw (electrical) MSBR described in the section "Molten-Salt Breeder Reactor 204 Design Studies"” of this report. The study evaluated the engineering Teasibility and costs for a plant that operated continucusly as an inte- gral part of the reactor system, belng located in two cells adjacent to the reactor cell. The plant was designed to treat 15 ftB/day of fuel salt and 105 ft3/day of fertile salt. The fuel salt was an Lil-Bel's (69-31 mole %) mixture containing the fissionable 272UF,; fertile salt was a 71-29 mole % mixture of Ii¥~Thf,. The lithium component of each stream was enriched to about 99.995 at. % 7Ii. The processing cycle was selected to give the optimum combination of fuel cycle cost and breeding gain. Description of Fuel Process The primary obJective of the fuel process is to recover uranlium and carrier salls sufficiently decontaminated from fission and corrosion products so that the reactor has an attractive breeding potential. The recoverad materials are recycled to the reactor, and the fission products are discarded. Only four major operations are required to accomplish this for the fuel stream: fluorination, sorption of UFg, vacvum distil- lation, and salt reconstitution. These operations are shown schematically in Fig. 7.1. As it enters the processing cell, fuel salt is only a few seconds removed from the Tission zone and is extremely radiocactive. The strean is delayed for about 36 hr before fluorination to allow the heat genera- tion rate to decrease to a point that temperature control in the fluori- nator is made easier. The curve in Fig. 7.7 shows the gross heat gener- ation rate of the fuel salt. The molten salt flows into the top of =a column anc is contacted by a countercurrent stream of fluorine, which strips out the uranium according to the reaction 500~550°C Uy, + Fp —————=> Ul'g . figsion products Ru, Tc, Nb, Cs, Mo, and Te are also volatilized and ac- company the Ulg. The system, consisting of molten LiF-BeF,-UF,, fission products, and clemental fluorine, is extremely corrosive to the walls of the fluorinator, requiring clever design if a significant lifetime is to be obtained. It is proposed to Jacket the fluorinator with a coolant that will maintain a 0.5- to 0.75-in.-thick layer of frozen sallt on the inner surface of the column to shield the wall from the molten salt.” A schematic diagram of the fluorinator is shown in Fig. 7.8. The gas stream leaving the fluorinator passes through a sorption system composed of temperature-controlled beds of Nal and Mgl, pellets. The first section of the Nab bed is held at about 400°C and sorbs most of the rission products; the secound seection of the bed at about 100°C e O R GRNL —~DOWG 65 - 9394 RA | HEAT GENERATION RATE {8tu/ hr-ft>) 10 100 MINUTES o lsgr bt il ] i 10 i 1iC 20 | HOURS l YEARS 100 | (RN L rre Foi Lot L L Lo oL 1072 1072 o+ 10° 10 102 to” i? TIME AFTER DISCHARGE FROM REAGCTOR (days) Fig. 7.7. Fission Product Decay Heat in MSBR Fuel Stream for a 1000- My (Electrical) Reactor. sorbs technetium, part of the molybdenum, and UFg, and allows the re- maining fission products to pass. Upon heating from 100 to 400°C, the second section of the sorber releases molybdenum, technetium, and UFg, which passes through Mgls for retention of technetium while allowing UFg to pass. Uranium hexafluoride is frozen in cold traps and retained for recycele to the reactor. Uranium-free salt flows from the fluorinator into a continucus : distillation unit, which is operated at about 1 mm Hg pressure and 1000°C. Under these conditions, it is possible to distill IiF and Bel¥, from the bulk of the fissicn products.6 Rare-earth fission products are much less volatile than lithium or beryllium fluoride, allowing a good separation to be achieved. Zirconium fluoride, however, is sufficlently wvolatile that this fission product will contaminate the Lif-Bel's product. In this study the vacuuwn s5till was a 2.5-ft-diam by 4-ft-high vessel containing a bank of cooling tubes over most of its height. A condensing surface at the top condensed and collected the overhead product. To initiate the operation the interior of the still is charged with 4 £t of molten ILiF; the still is evacuated and brought to temperature, and salt from the fluorinator is allowed to flow into the pool of molten LiF. Temperalure 1s controlled so that liguid is vaporized at the same rate at which 1t enters the still. There is no bottom discharge, so the still volume remains constant. Accordingly, the concentration of fission ORNL—-DWG 65—-3037TRA f}sz,UFG, F.P. TO SORPTION SYSTEM DE-ENTRAINMENT SECTION COOLANT FUEL FROZEN WALL OF SALT — 1/2 TO 3/4 in. THICK GRAVITY %\‘“‘FUEL SALT DRAIN NaK COOLANT::;b( = & Fa Fig. 7.8. Continuous #luorinator with Frozen Salt Wall for Cor- rosion Protection. 207 products steadily increases in the 4 ft? of LiF. After about 67 days® operation the accumulated heat generation rate (see Fig. 7.9) has become so great that the heat removal capablility of the cooling system is reached; the still contents are then drained to waste storage, and the operation 1s repeated. Heat 1s removed by forced circulation of NaK on the shell side of the tubes. The concentration factor for rare~earth fission products 1ln the gtil1l is about 250. The fraction of the process stream, which 1s almost entirely “LiF, discarded at this point is slightly lesas than 0.4%. Be- cause of the volatility of Zrf., an additional discard of the distiliate ig required to purge this fission product. As much as a 5% throwaway might be necessary in this type of operation. At the time of this szstudy, data were not available to assess the effect of increasing fission product concentration in the stlll on relative volatilities. Conse~ guently, the overall decontamination factor (DF') of the distillate cannot be predicted accurately, but it is believed that a DF of at least 100 can ve attained. ORNL~DWG €65 ~ 1821 R2A 1 O6 _________ e L L LA . HEAT GENERATION RATE (Btu/hr) 0 AT 1o’ 10? 10* 255 days TIME AFTER DISGHARGE FROM REACTOR (days) Fig. 7.9. Healt Generation Rate in the LiF Pool Resulting from Fission Product Accumilation in the Still. 208 The final step in fuel processing is reconstitution to make a suitable feed for the reactor. The LiF-BeF, distillate 1s admitted to a reduction column containing molten (~600°C) IiF-BeF,-UF, that is approximately the correct fuel composition. Concurrently, gaseous UFg from the cold traps is introduced near the bottom of the column, and hydrogen gas is admitlted at a point a little farther up the column. The UFg absorovs in the molten salt to form an intermediate Tluoride of nranium such as UFs, which reacts with H; according to the reaction UF's +~%H2-——-——-——9’ UF, + 1F . Makeup UFg from the blanket process and makeup IiF and BeF; are added at this point. The reconstituted fuel is sent to the reactor core to com- plete the fuel processing cycle. Degcription of Tertile Process The fertile stream process consists only of continuous fluorination and UFg purification by sorption. The operation is analogous Lo the corresponding fuel stream operatiocn but at a higher volumetric rate. The cycle time of the fertile stream is purposely kept short (20 to 25 days) to keep a low uraniuvm concentration in the blanket, thereby keeping the fission rate low. The low fission rate cnsures a low fission product accumuiation rate so that it is unnecessary to remove them on the same cycle as uranium. In fact, a 30-year discard cycle of the barren fertile stream is a sufficient purge rate for Tisslon products. Excess UlFg over thal required to refuel the core 1is sold. Waste Treatment Four waste streams requiring storage leave the processing facility: (1) aqueous waste from the KOH scrubber, (2) NaF and MgF, sorbent from the Ul'g purification system, (3) molten-salt residue from the distillation unit, and (4) molten salt from the fertile-stream discard. The aquecus waste comes from vent-gas scrubbing and is small in volume; it was as- sumed that this stream could be combined with reactor agueocus wastes for storage. The Uwo molten-salt wastes are stored in underground tanks, and the pelletized sorbents are stored in cylindrical containers in an under- ground vault. Forced draft cooling is provided for these three storage areas. This study includes a charge for 30-year interim storage of the molten-salt wastes and for 5-year interim storage of the solld waste. Perpetual storage beyond these times was not considered. 209 Off-Gas Treatment Most of the off-gas from the process comes from the continuous fluorinators. Although fluorine is recycled, a small amount is bled of f to purge gaseous Tission products. The off-gas is scrubbed with an agueous caustic solution, filtered, and discharged 0 the atmosphere. cummary of Capital and Operating Costs The desglign study included an estimation of capital and operating costs for the integrated processing plant. Space requirements and costs were estimated Tor a typical layout (Fig. 7.10) adjacent to the reactor system. FBach item of major equipment was designed to the extent that a reascnably accurate estimate of its cost could be made; the costs of auxiliary items, such az utilitles, piping, instrumentation, electrical connections, insulation, and sampling, were estimated by applying appro- priate factors to process equipment costs. Direct operating costs were estimated for labor and supervision, e consumed chemicals, utilitlies, and maintenance materials. These costs and capital costs are summarized in Table 7.2. ORNL — DW5 85 ~41804RA % "SORBERS Naf - Mgf, WASTE Lfi?@??@YlfiNfi \ TWASTE ' RECEIVER FLUORINATOR ------ HEAT EXCHANGER REACTOR POWER GENERATION SYSTEM REACTOR 69 ft \ SFERTILE MAKE-UP FLUORINATOR — SORBERS —COLD TRAPS Tk A T L e I ek Fig. 7.10. Reactor Integrated Processing Plant FPreliminary Layoutb. 210 Processing Cost The costs summarized in Table 7.2 contribute about 0.2 mill/kwhr to the fuel cycle cost when the fixed charges are amortized at lE%/year and the plant factor is taken at 80%. The amortization charge includes 10%/year for depreciztion, 1%/year for taxes, and 1%/year for insurance. Table 7.2. Cost of an Integrated Processing Plaant for a 1000-Mw (Electrical) Molten-Salt Breeder Reactor Fixed Capital Costs ($) Building space 1,130,200 Process equipment 1,734,200 Interim waste storage 785,100 Services and utilities 1,648,300 Total 5,301,500 Taventory Costs® (§) Fuel salt carrier 89,460 Fertile salt £9,200 NaX coolant 40,000 Total 198,660 Direct Operating Costs ($/year) Supervision and labor 399,600 Chemicals 70,390 Waste containers 28,270 Utilities 80,300 Maintenance materials 209,230 Total 787,790 a . . . Ixcludes fissile material. 211 References C. D. Scott and W. L. Carter, Preliminary Design Study of a Con~ tinuous Fluorination—Vacuum Distillation System for Regenerating Fuel and Fertile Streams in a Molten Salt Breeder Reactor, ORNI-3791 (January 1966). M. J. Kelly, ORNL, unpublished data (May 21, 1965). G. Long, Stebility of UFs (in preparation). R. P. Milford et al., Ind. Eng. Chem. 53, 357 (1961). R. W. Kessie et al., Process Vessel Design for Frozen-Wall Contain- ment of Fused Salt, ANI-6377 (1961). M. J. Kelly, Removal of Rare Earth Flssion Products from Molten Saltb Reactor Puels by Distillation, paper presented at 1lth annual meeting of American Nuclear Society, Gatlinburg, Tenn. (June 21=24, 1965). 213 OAK RIDGE NATIONAL LABORATORY MOLTEN-SALT REACTOR PROGRAM FEBRUARY i, 1966 R. B. BRIGGS, DIRECTOR D P. R. KASTEN,* DEPUTY DIRECTOR R W. B. MCDONALD,** ASST. IRECTOR R MSBR STUDIES MSRE OPERATIONS COMPONENT AND 5YSTEMS DEVELOPMERT J8C DESIGN AND DEYELOPMENT REACTOR CHEMISTRY METALLURGY PAUL R. KASTEN* R ‘ i W, R. GRIMES* RC A, TABOADA" M&.C - R.L. M W, 3. M DONALD: ¢ R P. N, HAUBENREICH R D. SCOTT R L. MOORE 18C Bl ANKENSHIP® RC £ G BOHLMANN® RC D. A. CANIOCO™ MaC H F. McDUEFIE* ac W, H. CO0K** MAC . R, T el " D. D. RAY™ 18C M. RICHARDSON R 0. R. SEARS RC R. C. OLSON* R 5. 5. KIRSLIS RC LGUHILL s T R. E. CARNES R L e ~ccper SLC HURTT R V. R.SILER” 1&C - e H. H. STONE RC M. L. MYERS™ R J.H. SHAFFER RC s C. A. GIEFGRD R R ~ . R. E. THOMA® " RC J. C. JORDAN R R. #ARD™ 1&C - A . W. T. WARD RC . C. ROBERTSON R i1 MILLER R C. JONES R W. K. R. FINNELL RE L. E. PENTON R 8. F.HITCH < W. E. RAMSEY R W. P, TEICHERT C R. SMITH, JIR. R J. L. STEPP R FUEL FROCESSIMG DEVELOPMENT J. L. UNDERWOOD R I 5. R, WEST R ! CHEMISTRY T M. E. WHATLEY? cT ) E.WOLFE R - i ! R.E. THOMA*! RC ! W, L. CARTER™ cT ; J. R HIGHTOWER cT AC AMALYTICAL CHEMISTRY DIVISION ! R. B, LINDAUER ™ 7 CT CHEMICAL TECHNOLOGY DIVISION PUMP DESIGN AND DEVELOPMENT L. £. McNEESE cT r . T L C. £ SCHLLING cT D D|RECTORS» DIVISICN A. G. GRINDELL " R . 0. 5C0TT CT E ANALYTICAL CHEMISTRY GZ&C GENERAL ENGINEERING AND CONSTRUCTION DIVISION DESIGH LIAISON o o omith N S BEAMS it i o 1&C IKSTRUMENTATION AND CONTROLS DIVISION ] L. V. WILSON * R €. J. SHIPMAN cT Ijg wmii-l‘N :g MRC METALS AND CERAMICS DIVISIDN €. K. McGLOTHLAN R CHEMICAL PROCESSING : W. G. SISSON: T o - R. F. APPLE AC R REACTOR DIVISION R. 8. LINDAUER"" R J. M. DALE AC RC REACTOR CHEMISTRY DIVISION £. K. HEACKER AC " PART TIME ON MSRP C. E. LAMB* AC v DUAL CAPACIT | W, L.MADDOX" AC v Y i D.L.MANNING AC i MAINTENANCE B. H. WEBSTER R MATERIALS A. F. CILLEN R fl L. P. PUGH R W, H. COOK** M&C 215 ORNL-393%6 UC-80 — Reactor Technology INTERNAL DISTRIBUTION 1. G. M. Adamson 57. A. G. Grindell 2. L. G. Alexander 58. R. H. Guymon 3. C. F. Baes 59. P. H. Harley 4. J. M. Baker 60. C. S. Harrill 5. 8. E. Beall 6l. P. N. Haubenreich 6. E. S. Bettis 62. P. G. Herndon 7. D. 8. Billington 63. R. F. Hivbs (Y-12) 8. F. F. Blankenship 64. M. R. Hill 9. R. Blumberg 65. E. C. Hise 10. H. F. Bauman 66. H. W. Hoffman 11. A. L. Boch 67. V. D. Holt 12. E. G. Bohlmann 68. P. P. Holz 13. C. J. Borkowski 69. A. Hollaender 14. G. E. Boyd 70. A. 5. Householder 15. M. A. Bredig 71. T. L. Hudson 16. E. J. Breeding 72. H. Inouye 17-26. R. B. Briggs 73. W. H. Jordan 27. F. R. Bruce 74-83. P. R. Kasten 28. G. H. Burger 84. R. J. Kedl 29. 8. Cantor 85. M. T. Kelley 30. D. W. Cardwell 86. M. J. Kelly 31. W. L. Carter g7. C. R. Kennedy 32. E. L. Compere 88. T. W. Kerlin 33. J. A. Conlin 89. H. T. Kerr 34. W. H. Cook 90. A. I. Kraskoviak 35. L. T. Corbin 91. J. W. Krewson 36. G. A. Cristy 92. C. E. Lamb 37. J. L. Crowley 93. C. E. Larson 38. F. L. Culler 9%. T. A. Lincoln 39. J. M. Dale 95. R. B. Lindauer 40. D. G. Davis 96. R. S. Livingston 41, W. W. Davis 97. M. I. TLundin 42. J. H. DeVan 9¢. H. G. MacPherson 43, R. G. Donnelly 99. F. C. Maienschein 44. D. A. Douglas 100. W. R. Martin 45, N. BE. Dunwoody 101. H. E. McCoy 46. J. R. Engel 102. W. B. McDonald 47. E. P. Epler 103. H., F. McDhuffie 48. W. K. Ergen 104. C. K. McGlothlan 49. D. E. Ferguson 105. E. C. Miller 50. A. P. Fraas 106. C. A. Mills 51. J. H ¥rye, dJr. 107. W. R. Mixon 52. C. H. Gabbard 108. R. L. Moore 53. W. R. Gall 109. K. Z. Morgan 54. R. B. Gallaher 110. J. C. Moyers 55. R. G. Gilliland 111. M. L. Nelson 56. W, R. Grimes 112. W. R. Osborn 113-114. 115. 11e. 117. 118. 119. 120. 121. 122. 123. 124 125. 126. 127. 128. 129. 130. 131. 132. 1373, 134. 135. 136. 137. 138. 139. 140. 141. 209-210. 211. 212. 213. 214 215. 216 217. 218. 219-549. R. B. Parker L. F. Parsly Patriarca R. Payne Phillips B. Pike B. Piper E. Prince L. Redford Richardson C. Robertson C. Roller W. Rosenthal C. Savage W. Savolainen Scott Seagren Shaffer Shipley Skinner Slaughter Smith Smith Snell Spencer Splcwak Steffy E. Stevenson QUHZedrRHGEHUPPOA R R R G EENE T DT @jflzqi?fg:citfifiim . F. Cope, . B. Deering, Z?UOU 142. 143. 144 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163-164. 165-169. 170-172. 173-207. 208. nEEHmxdg - - - N H YR Y Li$=EiC2C12 WG W o Gy t+<:c:czhzmimiz m=E Qo G J . Susano . Tallackson . Taylor Thoma Tolson Trauger . Tucker Ulrich . Watkin Watson . Webster Weinberg . Weir . Wegtsik Whatley Williams . White . Wilson Yost Young Biology Library Reactor Division Library ORNL — Y-12 Technical T.ibrary Document Reference Section Central Research Library TLaboratory Records Department Laboratory Records, ORNL R.C. EXTERNAL DISTRIBUI'ION Atomic Energy Commission, ORO Atomic Energy Commission, ORO G. Garrison, Atomic FEnergy Commission, Washington Shaw, Atomic Energy Commission, Washington F E. Slnclalr Atomic Energy Commission, Washington W. L. Smalley, Atomic Energy Commission, ORO J. A. Swartout, 270 Park Avenue, New York 17, New York M. J. Whitman, Atomic Energy Commission, Washington Research and Development Division, AEC, ORO Given distribution as shown in TID-4500 under Reactor Technology category (75 copies — CFSTI)