ORNL-4548 UC-80 — Reactor Technology Contract No. W-7405-eng-26 MOLTEN-SALT FIEACTOR PROGRAM SEMIANNUAL PROGRESS REPORT For Period Ending February 28, 1970 M. W. Rosenthal, Program Director R. B. Briggs, Associate Director P. R. Kasten Associate Dlrector -LEGAL NOTICE This report was prepared as an account of work sponsored by the United States Governmeni, Neither the United States nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contractors, subcontractors,.or their employees, makes any warranty, express or implied, or assumes any ‘legal liability or responsibility for the accuracy, com- pleteness or usefulness of any information, apparatus, product or process disclosed, or represents that its use wwld not infringe privately owned rights, AUGUST 1970 OAK RIDGE NATIONAL LABORATORY Ozk Ridge, Tennessee - operatedby UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION DISTRIBUTION OF THIS DOCUMENT IS UN »\ This report is one of a series of periodié repoi‘ts in which we describe‘ the progress of thé program. Other reports issued in this series are listed below. ORNL-3708 is especially useful because it gives a thorough re- - view of the design and construction and supporting development work for the MSRE. ~y ORNL-2474 Period Ending January 31, 1958 ORNL-2626 - Period Ending October 31, 1958 ORNL-2684 Period Ending January 31, 1959 ORNL-2723 Period Ending April 30, 1959 ORNL-2799 Period Ending July 31, 1959 ORNL-2890 Period Ending October 31, 1959 ORNL-2973 Periods Ending January 31 and April 30, 1960 ORNL-3014 Period Ending July 31, 1960 ORNL-3122 Period Ending February 28, 1961 ORNL-3215 Period Ending August 31, 1961 ORNL-3282 - Period Ending February 28, 1962 ORNL-3369 Period Ending August 31, 1962 ORNL-3419 Period Ending January 31, 1963 ORNL-3529 Period Ending July 31, 1963 ORNL-3626 Period Ending January 31, 1964 ORNL-3708 "Period Ending July 31, 1964 ORNL-3812 Period Ending February 28, 1965 ORNL-3872 Period Ending August 31, 1965 - ORNL-3936 Period Ending February 28, 1966 ORNL-4037 = Period Ending August 31, 1966 ORNL-4119 ~ Period Ending February 28, 1967 ORNL4191 Period Ending August 31, 1967 ORNL-4254 Period Ending February 29, 1968 - ORNL-4344 Period Ending August 31, 1968 ORNL-4396 Period Ending February 28, 1969 ORNL4449 Period Ending August 31, 1969 < Contents 1 2307 51804 1 (0 (N xi QUMM A RY .\ 0t ittttttiietianseeseeesseanseseausasasasssosnasiossoasssanssasscssssnsnans xiii PART 1. MOLTEN-SALT REACTOR EXPERIMENT 1. MSRE OPERATIONS ... iiittiiieetrenereinaesecosnersosssonasorsnsasssssecssasssannes 1 1.1 Chronological Account of Operatlons and Mamtenance .................................... 1 1.2 Operations Analysis ..........ccooeviniieeinnennn. ettt 5 12,1 Reactivity Balance ..........ccouiuiiiiiiineeneernientneeneioaereecrancannnns 5 122 XenonPoisoning .........ceeiiiiiniiiiiiiiiines i eeseceereeeenaanirenns 6 123 Dynamics Testing .. ..ccvvveveeeiiivennnnnnniennannesns e teeteeeaeeaaeaa, 7 1.24 Operational Diagnosis by Noise Analysns et eerarr et e 7 125 Tritium....oooivieneinnnaeneneennassnnes e eeseieaas meere e e 9 1.2.6 Fission Product Distribution in the Fuel System ...........coiivieniiaiininneana. 11 1.2.7 Graphite Samples Exposedin Fuel Pump ... iiiiin... 13 e ™ 1.2.8 Evaluation of Leakin Primary System . ........c.ciiiiiniaiiaanae, e 14 L 129 Salt Transfer tothe Overflow Tank ..........ccoiiiiiiiiinninernnneneennnenen 14 » 1.2.10 RadiationHeating ...........coovueeiiinnn... ettt 15 L 1.2.11 SystemHeat Balance ............c.ououtiiiiuinneenninteeeanncecacnnnooroannnns 15 1.2.12 Heat Transfer in Primary Heat Exchanger ..............oiiiiiienninainenn, 17 1.2.13 Thermal Cycle HIStOTY ... oiiveiiiiiiii it rinaniaet i iattersnnansoans 17 1.3 EQUIPIMENt ..ottt ittt iiiiniiiiineeeaseseesnnunenerenscaenataasenasnenns 18 1.3.1 SaltSamplers .......coovivriierrennnnan et aae et 18 1.32 ControlRodsandDrives ..........citiiiiieniieirnenenrninanasns Pael e .. 20 133 Off-GasSystems ..........ccivieieenennonnonennnans feereeeas etenencoananans 20 '1.34 Component Cooling System ......... P 22 1.3.5 Containment and Ventilation ..... e awen s e a et e M eeeeceeeraans 22 1.3.6 Heatersand Electrical System . ........veiiiiiiinaneniiieieenedonneionsnnsns 22 1.3.7 OilSystems for Salt Pumps .......coviiiiernuiniiinnaniecreerealineanecneannns 23 1.38 RadiatorandMain Blowers ...........coiiiieiiiiiierieineresaarosaccosasens 23 14 Remote Maintenance ............ eeieeneanais et teee et i 23 1.5 Final Shutdown and Standby Status .......... e e ... 24 1.6 Preparations for Inspection ..... e e feeeeceecaane 25 T 2. REACTOR ANALYSIS .. ..uvueneeennenenenennnns T PP 26 \j} ' 2.0 Introduction .......c.oeeeuuiareeneeeenesaocennnnonans et 26 ; * 2.2 Extensions of MSRE Core Physics Calculations to the Use of c XSDRN-CITATION Programis .. .. cvvuevruuervannerenanasetivassotsosocesssasennans 26 2.3 Evaluation of Long-Term Reactivity Behavior Durmg Operatlon with23%U ...l 30 iii iv 3. COMPONENT DEVELOPMENT ..... e, e 34 3.1 Freeze-Flange Thermal Cycle Test ........... s e P eea e 34 3.1.1 Facility Operation Problems ........ et ettt it e et e - 34 3.1.2 ‘InspectionoftheFlanges.‘.............' ................................ se... 34 32 PUMPS toiiineeiiiinneeeeannnnns R e, e 35 3.2.1 Mark2 Fuel Pump .......... e teeeaeeieer s R 35 322 Oil Pump Endurance Test .. ............... e eseeteresenesesiaeeaaeneeaan - 36 3.3 Development ofAnalytlcal Mode! for !35Xe Poisoningin the MSRE ........................ 37 4. INSTRUMENTS AND CONTROLS . .......... e e, e e 39 4.1 MSRE Operating Experience . ..... S e e, . 39 42 Control System Design . .. .. e e e ettt 40 4.3 MSRE On-Line COMPULET .. .vuvvueneneeneunenrnernrnensenenasnenenesnensnnes e 40 PART 2. MSER DESIGN AND DEVELOPMENT ~ S. DESIGN .. tteee ettt ete et e et et et et e et et et et e et ananaeea.. 41 5.1 Single-Fluid MSBR Design Study . ..... el S P el L4 5.2 MSBR Primary Salt Storage Tank . ........ . e e . 46 5.3 Startup and Shutdown Procedures .................... e e e, 46 ' 5.3.1 Startup Procedures ....... e etereiteretaeeraaaeeaas e ireeeceeeeas e 46 5.3.2 Normal Shutdown .................. s heeseassaaeraansnnes b iee i 48 54 Designs for Fn'st-Generatlon Molten-Salt PowerReactors ...........coiiieiininannnnnnnnn. 48 540 Gemeral .. ...ttt it it ettt ettt et s ... 48 542 Large MSRE-TypeReactor...........coiiiiiiiiiriininannnnnns. - 1 543 _SphericaIReactorwithGraph.iteBall'Bed_.....,.._.,................_....._. ...... . 49 5.5 Bayonet-Tube Heat Exchangers ....... e eeane P e et Ceeeean 54 - 5.6 Distribution of Tritiuminan MSBR ........... PP 6. REACTOR PHYSICS ......... PR T e 58 6.1 Physics Analysis of MSBR .........cccovvennnnnn.. e e s 58 6.1.1 Single-Fluid MSBR Reference Design ...........ciiiiriiiiirennnnrennrennnnnnns 58 6.1.2 Designs for First-Generation Molten-Salt Power Reactors G e e eer et ie e, 60 6.1.3 Reactivity Coefficients .................. ... . oLl e e 63 6.14 ControlRodWorth .................. st ierecenecasena e esenaseaaaa 64 6.2 MSR Experimental Physics ......... et erereseeareeaaaas [ A 65 6.2.1 Indication of Integrated Power by 2 23 SU Depletion ..... P v eenreieeereraaa 65 7. SYSTEMS AND COMPONENTS DEVELOPMENT .................... P el 6T 7.1 Molten-Salt Steam Generator .................. et teieeteeeea e, e 67 7.1.1 Steam Generator Industrial Program .......... rieeeeas P ereenan... 67 7.1.2 -Steam Generator Tube Test Stand ................ e I 1 7.2 Sodium Fluoroborate Test Loop ............. e e, .. 69 7.2.1 Salt Leak from Pressure Transmitter ...........cc00uvn.s SRR TR RN RS RN IR T, 69 722 GasSystemStudies ..........ccciiiiiiiiiiretnrienannaann e reeseeaas e 70 7.3 MSBRPumps...............................’ .......... O e 72 7.3.1 MSBE Salt Pump Procurement ................. eerrreeniaras eresarecesennae 72 m)..‘ ) - ~ - fl‘) 1) > 732 MSBESaltPump TestStand .............. e e 72 733 ALPHAPUMD ....iivitininineennensensesesessseseseneassssssssnsnconnnns 74 74 RemoteWelding ..................oooiiiiian... et e 74 8. MSBR INSTRUMENTATION AND CONTROLS ..........oooiviiiiiiiiiiiiin.., S 79 8.1 Control System Analysis ..........ciieiiiiiiiiiiiieenttscersasnssanns heeveeaaeaa. 79 9. HEAT AND MASS TRANSFER AND THERMOPHY SICAL PROPERTIES ......... . . 87 9.1 HeatTransfer ................ b e e et ee e e taaeesaiteeeaeseaaa et e 87 9.2 Thermophysical Properties ..........oi ittt irersonsonssorinessssnases ~. 89 9.3 Mass Transfer to Circulating Bubbles ........ e eseaeas eeeaeaae e eeerereaanaaas 90 _ PART 3. CHEMISTRY 10. CHEMISTRY OF THEMSRE ... ...0eeeeeeeeeninnnnnnnnnns e, e e .. 93 10.1 Corrosion of the MSRE Fuel Salt Circuit ... .cueevueevnnennnen.. e, 93 10.2 Relationship of the Distribution of ? *Nb in the MSRE to the Presence ’ ' of Uranium Trifluoride i intheFuel Salt . ...... ..ttt ittt e eiienennanan. 96 10.3 Power Output of the MSRE Based on the Isotopic Composmon of Plutonium ................. 98 104 Isotopic Composition of Uranium During 233U Operations ...........c.oovveenenvnnenenn. 100 - 10.5 Surface Tension and Wetting Behavior of the MSRE Fuel and Coolant Salts ................... 100 11. FISSIONPRODUCT BEHAVIOR .. ... .ottt ie it e e ieea e 104 11.1 Examination of the Fourth Set of Surveillance Specimens from the MSRE et reeieseesaeaaaaa 104 11.1.1 Radiochemical Analyses of the Graphite ..............vvv.t.. et iesesesaaeaas 104 11.1.2 Radiochemical Analyses of Fission Product Deposition on Metal ........ e 108 11.1.3 Fission Product Distributionin the MSRE ...... ... it niiiieririerannenns 110 11.2 Fission Product Distribution in MSRE Pump Bowl Samples....... ... 111 11.3 Investigations of the Behavior of Fission Product Niobium ............cooviiiiiiiin.., 118 114 Noble Metal Fission Product Chemistry . . e e e e 123 114.1 Introduction.........cceonnene teniesieinesitreinreinisteanens Cereeeresaen 123 1142 Synthesis ...........c 000, P . 11.4.3 Niobium and Molybdenum Fluoride Solutions in Molten Li, BeF4 e eeeereiaaaaa, 124 114.4 Mass Spectroscopy of Molybdenum and Ruthemum Fluorides .. ceeeeaean e 126 12. PROPERTIES OF THE ALKALI FLUOROBORATES e i e s iaaaann Cieeeaeeeaaaeas 133 12.1 Phase Equilibria in the Alkali Metal Fluoride—Metal Fluoroborate Binary Systems: The System RbF-RbBF, ......... ...t e 133 12.2 Solubility of Na;CrF; in Sodium Tetrafluoroborate Melts Ceeeeaeaan eeieressesainsenanens 134 12.3 Preparat;on of Pure Sodium Tetrafluoroborate ..............coiiiiiiiiiiiinn,, PR 134 124 Investlgatlon of Oxidation of Metals by Fluoroborate Coolant ...... 0cvvmeurvnnn N 135 12.5 Boron Trifluoride Pressure over Fluoroborate Coolant Salt Admixed _ ' with MSBR Fuel Salt ........... iedereeteneranens heesraeeenraensena eeeseraneen 136 12.6 Heat Content of Alkali Metal Fluoroborates ............ Cedesrienaeens et 138 13. PHYSICAL CHEMI STRY OF MOLTENSALTS ..o ovovinnerennnnn e B e 13.1 Equilibration of Rare -Earth-Containing Molten Fluondes with Vanous Solids .................. , 132 The Oxide Chemistry of Plutoniumin Molten Fluorides ....... ...t iiiiiiniinnennnns .. . 133 The CeF3-ThF, System .........coivieiinnannnnnn eeeae e eerenareanaens e 134 Equilibrium Phase Relationships in the System LiF-BeF,CeF3 ............... e . 13.5 An Investigation of Possible Polymorphic Transitions in Uranium Tetrafluoride ................ 13.6 Estimation of Activity Coefficients in LiF-BeF,-ThF, Melts e e PR -13.7 Estimation of Activity Coefficients of Alkaline Earths in Molten - Bismuth and as Chlorides or Fluorides .................... e eaeen M teeeresareaaes 13.8 Potentiometric Studies in Molten Fluorides . ........... e P 139 Electrical Conductivities and lonic Mobilities in the - Molten LiF-BeF, System ........couiiiiiiiriiiniernrennanacnnaesssnsons e 13.10 Electrical Conductivities of Proposed MSBR Fuel Compositions in ' : the LiF-BeF,-ThF,; System...... A 13.11 Determination of Liquidus Temperatures in the LiF -BeF, System from ' - EMF Measurements of Transference Cells ............. .ot et ecanoannnane 13.12 Coordination Effects on U(IV) Spectra in Fluonde Melts ...................... P 13.13 Hydrogen Behavior in Molten-Salt Reactor Systems .. ......... oo, .. 13.14 Crystal Structure of the Complex Fluoride (Na,Li); ThgFa;y «.vvevineneininiiniinnniinnn., . 14. CHEMISTRY OF MOLTEN-SALT REACTOR FUEL REPROCESSING TECHNOLOGY e .. 14.1 Distribution of Cerium, Europlum and Strontium Between anmuth ‘ : and Lithium Chloride .........ciuiiiriititeetinerscasscenssenasanssasssssnannsssse 142 Extraction of Cesium from Lithium Chlonde into Blsmuth by : Reduction with Lithium at 650°C ........ovverriiieirniaeaennn.. S 14.3 Removal of Chloride from Simulated MSBR Fuel Solvent by Reaction with Anhydrous Hydrogen Fluoride .......... et ieee e e 144 Calculated Behavior of Sodium, Rubidium, and Cesium in the | Molten-Salt Breeder Reactor Reprocessing Plant ......... e etteratieesaraes e .. 14.5 Reduction of Uranium and Protactinium with Titanium ........... e rebeseenns e 14.6 Bismuth-Gold Alloys as Extractants for Rare Earths and Thotium ........... s . 14.7 Bismuth-Platinum Solutions fqr the Extractlon of Fission Products from MSBR Salt + . v vttt it ceeseeeineassoseessnsasossnssasasaansenennnns 14.8 Extraction of Cerium and Thorium from L1F-BeF2 (66-34 mole %) into Bismuth 2t 600%C ..o v e s s ee e erncsoiesoeseaacanessanaesoeens e ' 15. DEVELOPMENT AND EVALUATION OF ANALYTICAL METHODS FOR MOLTEN-SALT REACTORS ........ e e 15.1 Determination of Oxidein MSRE Salt ..................... e et .. 152 Voitammet_ric Determination of {U*]/[ZU] Ratios in MSRE F_uel et teeeeeeeeria e, 7 15.3 Spectral Studies of MSRE Salts .............. e e s 154 Tritium in the Effluent Gases of the MSRE .............oooiiiniiininnne, s . 15.5 Reference Electrode Studies in Molten Fluorides . .............. o ettt e "l o i vii 15.6 Removal of Oxide fromNaBF; ...........ccvvuvevnnn. e, et 185 15.7 A Preliminary Study of Volatile AICla Complexes ........ccnvvuennn ettt aeeaeeare . 185 PART 4. MOLTEN-SALT IRRADIATION EXPERIMENTS - PART 5. MATERIALS DEVELOPMENT 16. MSRE SURVEILLANCEPROGRAM ................... P R 188 16.1 Propertiés of Hastelloy N Samples Removed fromthe MSRE .. ...........c..coveunrninennn.. 189 17. GRAPHITE STUDIES ...... e et T 200 17.1 - Fundamental Studies of Radiation Damage Mechanisms i in Graphlte e 200 17.2 Procurement of New Grades of Graphite . e et eeaeeeaia ettt 201 17.3 General Physical Property Measurements ................. e [P U 201 174 Graphite Fabrication .............. .. .ooiii.L, et e s easnssasaenssnsaensaanananan 203 174.1 Characterizationof Materials ............ciiiiiiuiuni it iiiininonrnennnns 203 174.2 Fabricationby Hot Pressing ............ ...t et e 203 17.4.3 Fabrication by IsostatlcPressmg Gt resreraea e ia s N . 204 1744 Conclusions ..........cooiiiiivinnerrnnnennnrnnnsnnnn. ettt 205 17.5 Measurement of the Thermal Conductivity of Graphite .. ... et seaerese et et 205 176 X-RayStudies ..........cciitiirnneanrncnsresscscanneaanans e e eaaaeeeeen 207 17.7 Electron Microscopy of Graphite ....... e e e ettt e ettt e e 209 17.8 Gas Impregnation of Graphite .............oeuoeneueneneennn. P e . 209 179 HFIR Irradiation Program . .......co ittt i ittt iiiieaaneanes . 213 17.10 Irradiation Behavior of Pyrolytic Carbons ..............covvuennn, et 215 17. 11 Calculation of Lifetime and Induced Stresses in MSBR Graplute Cores ...... b eeeeceraetaeans 218 18. HASTELLOY N ... ..uiutntintereaneaneaneenianeennaneaaneenn T 222 18.1 Aging of Modified Alloys ......oovvvrevennnenns. e, e e ... 222 18.2 Effect of Carbon and Titanium on Postirradiation Properties « . ...« «.vueeeeunneeeennseesnns 225 '18.3 Effects of Irradiation at 760°C on the Creep-Rupture Propertles : - ' - : | of Modified Hastelloy N ..............ociiiiiiiiiiiiiinn, Cirieiaereaeeaas e 229 184 Electron Microscopy of Modified Hastelloy N Alloys .‘ .................. e e ereeeeeaae 231 18.5 Weldability of Commercial Alloys ........ e e e e 238 18.6 CorroswnStudles.'.............j ..... P i erenraaa et rateaeecerarerereanas 240 18.6.1 FuelSalts ...... e benseaneiteaereeerietsnaiaaaseases eiecstesreeaeaans 240 18.6.2 Fertile-Fissile Salts ......... e earsaataaasr et N 242 18.6.3 BlanketSalts ........cc00iununnn. Crenieenanseaen eeiieaaans e e 242 18.64 Coolant Salts ...... i eeeeaieenaaaaas A ceees L . 242 18.6.5 Summary of Corrosion Studies .. ..... e e 247 187 Fluoroborate Purity Test ........... et aeeeen et e 247 18.8 Forced-ConvectlonLoopMSR-FCL- et easase b ean s et 247 - 18.8.1 Metallurgical Analysis ...........c..... eesens et eeeeeeereeeeaateeaaaaean 248 - 18.8.2 Forced-Convection Loop MSR-FCL-2 ....... veaiseeans i ieeaeaeenenaneaeeaaaas 251 189 Corrosion of Hastelloy NinSteam ....................... @t et et e 252 19. 20. 21. 22. 23. 24. 19.1 Fabrication Development.of Molybdenum Components ......... el e deedeeanaaaaes 253 19.2 Welding of Molybdenum ............... Crdeenetertatesacnasecenosoean P i T 19.3 Development of B1smuth-Re31stant Brazing Ffller Metals for Jommg Molybdenum .............. 255 194 Compatibility of Structural Materials with Bismuth ..ooovvrnrereenannnnnn. e erecaneens 256 19.5 Chemical Vapor Deposited Coatings . ......... e inene ediassreas e eeieraaceeas e 259 19.6 Scaling Resistance of Carbon Steels . ..... e S s e .. 261 SUPPORT FOR COMPONENTS DEVELOPMENT PROGRAM ..........civiiiiiinennnns aeaeees 265 | 201 Remote Welding Development .. ... .. ..orverenreneenenne. e e . 265 - 202 Support for Systems and Component Development ...... eeeaae . ... Gt ereestersenanane 265 20.3 Coated Bearing Specimens .......... eeeans eaes e e reaeeeaaa e 271 - PART 6. MOLTEN-SALT PROCESSING AND PREPARATION 7 FLOWSHEET ANALYSIS ....... e e e reeeeaeas 277 21.1 Rare Earth Removal Using the Metal Transfer PrOCESS .« .« vevsvneneneencnesoioeneeensenes 277 212 Protactinium Isolation Using Fluorination—Reductive EXtraction . ........ocecnzes e 279 21.3 MSBR Processing by Fluonnatlon—Reductlve Extractlon and the - Metal Transfer Process . ...... et el it e ieaaas 282 MEASUREMENT OF DISTRIBUTION COEFFICIENTS IN MOLTEN-SALT-MET AL SYSTEMS ....... 289 221 Metal Transfer Process STUGIES . .. .« v nvnueeereenensseneeioanaeacreesonesassensaeass 289 222 Extraction of Barium, Rare Earths and Thonum from Smgle-Flmd MSBR Fuels ..... PR 290 223 Solubility of Bismuth in Single-Fluid MSBR FUel Salts .. .......uuverrrrnnanereeunnnneenn. 292 224 Studies Involving BiFy ........... s . S SR eean.. 292 ENGINEERING DEVELOPMENT OF PROCESS OPERATIONS ......ovcvvnencnrnnns e 295 23.1 Reductive Extraction Engineering Studies ..........cciiiiiiiiiiiiiiin., s eeaes . 295 23.2 Design of a Processing Materials Test Stand and the First Molybdenum . Reductive Extraction Equipment ..............oviiiiieniinanan feeiesreaseeaeaa.s 298 233 Bismuth-Salt Interface Detector ... ...... B T e, 300 23_.4 Contactor Development: Pressure Drop, Holdup, and Floodmg in _ - Packed Columns .......ccvuiiiiiiianinnasaarannsstesinneronnnns eraeaae. e 302 23.5 Contactor Development: AxlalMlxmngackedColumns e e i, 304 23.6 AxmlMlxmngpenColumns..'......................‘ ...... e bereeeisenaareidireaens 307 23.7 Demonstration of the Metal Transfer Process for Removing Rare Earths ......... e e . 309 23.8 Frozen-Wall Fluorinator Development . ................. et ereeeanaaaanana, P 309 239 ElectrolytieCellDevelopment e ettt errene e e aev e srarrea ey 311 ~ 239.1 Static Cell Experiments ....... e e eeeteaeeatee e 312 239.2 Flow Electrolytic Cell Facility ............. Meeiesennne e eeiaeeea eeaeaae .. 316 DISTILLATION OF MSRE FUEL CARRIER SALT ........0uveeeeinunneeeeennnnnessennnnnnns 317 SUPPORT FOR CHEMICAL PROCESSING .. ......... e, RO feeiea.. 253 W ,_ £, i - o Y’ ) “ ix 25. SYSTEM CALCULATIONS ...........c..couiiennnnnns,. NP ... 324 25.1 MSBR Processing Plant Material and Energy Balance Calculations ........................... 324 25.2 Effect of Chemical Processing on the Nuclear Performance of an MSBR ...................... 325 26. CONTINUOUS SALT PURIFICATION SYSTEM ......... PP 329 27. DESIGN AND PREPARATION OF 22°PuF, CAPSULES FOR SMALL REFUELING ' 332 ADDITIONSTOTHEMSRE .. ...t i it ittt i et e e eaanas [ 2l ' ..‘ ") 4 'Introduction The objective of the Molten-Salt Reactor Program is the development of nuclear reactors which use fluid fuel§ that are solutions of fissile and fertile materials in suitable carrier salts. The program is an outgrowth of the effort begun over 20 years ago in the Aircraft Nuclear Propulsion program to make a molten-salt reactor power plant for aircraft. A molten-salt reactor — the Aircraft Reactor Experiment — was operated at ORNL in 1954 as part of the ANP program. ‘ Our major goal now is to achieve a thermal breeder reactor that will produce power at low cost while - simultaneously conserving and extending the nation’s fuel resources. Fuel for this type of reactor would be 233UF, dissolved in a salt that is a mixture of LiF and BeF,, but it could be started up with 225U or plutonium. The fertile material would be ThF, dis- solved in the same salt or in a separate blanket salt of similar composition. The technology being developed for the breeder is also applicable to high-performance converter reactors. A major program activity until recently was the operation of the Molten-Salt Reactor Experiment. This ~ reactor was built to test the types of fuels and materials that would be used in thermal breeder and converter reactors and to provide experience with the operation and maintenance of a molten-salt reactor. The MSRE operated at 1200°F and at atmospheric pressure and produced about 8 Mw of heat. The initial fuel con- tained 0.9 mole % UF,, 5 mole % ZrF,, 29 mole % BeF,, and 65 mole % "LiF, a mixture which has a melting point of 840°F. The uranium was about 33% 235U The fuel cxrculated through a reactor vessel and an external pump and heat exchange system. All this equipment was constructed of Hastelloy ‘N, a nickel- molybdenum-iron-chromium alloy with exceptional re- sistance to corrosion by molten fluorides and with high strength at high temperature. The reactor core con- tained an assembly of graphite moderator bars that were in direct contact with the fuel. The fuel salt does not wet graphite and therefore did not enter the pores. Heat produced in the reactor was transferred to a coolant salt in the primary heat exchanger, and the coolant salt was pumped through a radiator to d1ss1pate the heat to the atmosphere. Design of the MSRE started in the summer of 1960, and fabrication of equipment began early in 1962. The reactor was taken critical on June 1, 1965, operation at - low power began in January 1966, and full power was reached in May 1966. Sustained power operation was begun in December, when the MSRE was operated at full power for 30 days without interruption. In September 1967, a run was begun which continued for six months, until terminated on schedule in March 1968. Power operation during this run had to be ‘interrupted once when the reactor was taken to zero power to repair an electrical short in the sampler- enricher. - ' Completion of this six-month run brought to a close the first phase of MSRE operation, in which the objective was to demonstrate on a small scale the attractive features and technical feasibility of these systems for civilian power reactors. We believe this _objective has been achieved and that the MSRE has shown that molten-fluoride reactors can be operated at temperatures above 1200°F without corrosive attack on either the metal or graphite parts of the system, that the fuel is completely stable, that reactor equipment can operate satisfactorily "at these conditions, that xenon can be removed rapidly from molten salts, and that, when necessary, the radioactive eqmpment can be repaired or replaced. The second phase of MSRE operation began in August 1968, when a small facility in the MSRE building was used to remove the original uranium charge from the fuel salt by treatment with gaseous F,. - In six days of fluorination, 219 kg of uranium was removed from the molten salt and loaded onto ab- - sorbers filled with sodium fluoride pellets. The decon- Xi tamination and recovery of the uranium were very good. While the fuel was being processed, a charge of 233U was added to the original carrier salt, and in October 1968 the MSRE became the world’s first reactor to operate on 233U. The nuclear characteristics of the - xii MSRE with the 233U were close to the predictions, and, as expected, the reactor was quite stable. In September 1969 small amounts of PuF, were added to the fuel to obtain some experience with plutonium in a molten-salt reactor. The MSRE was shut down permanently December 12, 1969, so that the funds supporting its operation could be used elsewhere in the research and development program. Most - of the Molten-Salt Reactor Program is now being devoted to future moltensalt reactors. Con- ceptual design studies are being made of breeder reactors, and the program includes work on materials, on the chemistry of fuel and coolant salts, on proc- essing methods, and on components and systems development. | _ , Until two years ago most of our work on breeder reactors was aimed specifically at two-fluid systems in which graphite tubes would be used to separate uranium-bearing fuel salts from thorium-bearing fertile salts. We think attractive reactors of this type -can be developed, but several years of experience with a prototype reactor would be required to prove that graphite can serve as piping while exposed to high fast-neutron irradiations. As a consequence, a one-fluid breeder was a longsought goal. In late 1967 two developments established the feasi- bility of a one-fluid breeder. The first was demonstra- tion of the chemical steps in a process which uses liquid - bismuth to extract protactinium and uranium selec- tively from a salt that also contains thorium. The second was the recognition that a fertile blanket can be obtained with a salt that contains uranium and thorium ‘by reducing the graphite-to-fuel ratio in the outer part of the core. Our studies show that a onefluid, two-region breeder can be built that has fuel utilization characteristics approaching those of our two-fluid de- signs and probably better economics. Since the graphite serves only as moderator, the one-fluid reactor is more nearly a scaleup of the MSRE. ' These features caused us to change the emphasis of our breeder program from the two-fluid to the one-fluid - _ breeder. Most of our design and development effort is now directed to the one-fluid system. o ot i) ) Summary PART 1. MOLTEN-SALT REACTOR EXPERIMENT 1. MSRE Operations The power operation of the reactor that was in progress at the beginning of this report period was continued through October 1969. After a three-week shutdown, primarily to collect data on the deposition of fission products on metal surfaces in the system, . operation was resumed for a final run at power. Nuclear operation of the MSRE was terminated on December 12, 1969. Although plugging in the off-gas system continued to be an annoyance, no major equipment difficulties were encountered during the reactor opera- tion. ' . Experiments to study xenon behavwr in the reactor system were concluded early in the report period. Significant effects of circulating void fraction and cover gas solubility on xenon poisoning are still being evaluated. After completion of the xenon studies, the reactor was operated at high power for a month to facilitate the fission product deposition studies. This continued operation at power required the addition of more fissile material to the fuel. Six capsules of plutonium, as PuF,, were added through the sampler- enricher, demonstrating the suitability of plutomum as an MSR makeup fuel. Enrichments were also made with 233U _ The remote gamma-ray spectrometer was used exten- sively, with the reactor operating and during shut- downs, to study fission product behavior. The on-power studies provided additional information on nuclides with relatively short half-lives. Measurements were made on primary-system components, the reactor off- gas line, and samples exposed in the pump bowl. Extensive use was also made of the fuel sampler during the last two runs, both to expose materials in the pump bowl and to remove salt and cover gas samples. Subjects of particular interest were the effect of the oxidation- reduction state of the salt on fission product behavior and salt properties and fission product transport to the off-gas system. xiii Data were collected in the final power run in an effort to establish a tritium balance for the MSRE. Results indicated that 60 to 70% of the tritium produced in the fuel loop goes out through the fuel off-gas system. Other special measurements were made to improve the heat-balance power calibration of the system. After the final reactor shutdown, a leak developed in the drain-line piping which' released a small amount of activity into the secondary containment. The leak was stopped by freezing salt in the drain tank freeze valves; then the secondary containment was cleared by venting the activity (mostly xenon and krypton) through filters to the stack. With the fuel salt secured, the specimen array was removed from the core, and the reactor system was shut down completely. Detailed analysis of the reactor operation is continu- ing, and plans are being formulated for a final examina- tion of the system. 2. Reactor Analysis Reactor physics studies were continued, relating to the effects of revisions in nuclear data on the calcula- tion and interpretation of MSRE nuclear performance characteristics. This work was extended to include a comparison of results of using various computing techniques to perform the basic neutronics calculations. The study focused on the sensitivity of calculated multiplication factors and spectrum-averaged capture- to-absotption ratios for fissile materials to assumptions in the neutronics model and the computer programs used, Significant differences occurred in the calculated multiplication factors, which appeared to arise mainly from differences in calculated transport cross sections and their associated effects on MSRE neutron leakage. The capture-to-absorption ratios, however, were found to be much less sensitive to these effects and to depend mainly on the ‘underlying cross-section hbrarres for the fissile materials. An examination was also made .of the effects of revisions in nuclear .data on the interpretation of reactivity trends during the 225U operating history. All reactivity balance data logged at very low power over this period were modified to reflect the nuclear data ~ revisions . and to include certain carrections in .inter- Xiv pretation of the absolute magnitudes of several reac- tivity effects. The net result of all changes left the maximum variation in residual reactivity (reactivity not explicitly accounted for in the calculation) quite small (<0.1% &8k/k) during the entire operation with 23U, and within the region of validity of the reactivity balance model. 3. Component Development The freeze flange thermal cycle test was discontinued after 540 cycles. An inspection was made after cycle 470 which showed some changes in the cracklike pattern at the alignment stub in the main flange. However, as in all previous inspections, there were no indications in locations that would cause concern. The effects of flange cleaning procedures on the dye- penetrant indications are being examined. ' The mark 2 fuel pump has now operated 12,744 hr with molten salt. A new filter was installed in the off-gas line. Small salt particles are continuing to collect on the filter. Comparisons were made between the xenon poisoning observed in the MSRE and that predicted by a mathematical model developed in 1967 which ignored any effects of cover gas solubility. The model fits the observed behavior with argon cover gas much more closely than that with helium. The divergence with helium at low circulating void fractions is tentatively ascribed to effects of helium solubility in the salt. 4. Instruments and Centrols_ Most of the reactor instrumentation performed normally during the final periods of operation. How- ever, two unscheduled powef reductions were caused by dirty relay contacts in the load-scram circuits. Some changes were made in parts of the reactor containment system to provide prete_ction_for the tritium sampling - operations. The entire instrument system, including the . on-line digital computer, was placed in standby opera- “tion after the final reactor shutdown. PART 2. MSBR DESIGN AND DEVELOPMENT . 5. Design | The major effort of the design group during the past - period was the preparation of a comprehensive report on the single-fluid MSBR reference design. A first draft was circulated for comment. The only significant change made in the MSBR reference design was the addition of a 2500-ft> storage tank for the fuel salt. This tank would be used in the event that repairs are needed on the primary drain tank. The storage tank is located in the chemical processing cell and has a 1-Mw(th) heat removal system very similar to that used in the MSRE drain tank. Partialdoad operation and startup and shutdown procedures for the reference MSBR were given prelim- inary study, and it was concluded that these special conditions would not add undue operating complexity ot an unreasonable amount of equipment. Starting from the cold condition involves use of the cell electric heaters and circulation of helium inside the salt loops to preheat the system to about 1000°F prior to filling them with salt from the drain tanks. The steam system equipment is preheated by steam furnished from a supercritical-pressure, 1000°F auxiliary boiler fired by oil or gas. The steam generators are raised to essentially operating temperature before the coolant salt circula- tion is started. Generated steam is bypassed to the turbine condenser through a desuperheater until about 8% reactor load level is reached and the main turbines are brought up to speed. At about 20% load the control mode is essentially that for full-load operation with the reactor outlet temperature a function of the plant load. On loss of turbine load, steam would be let down through an extraction valve to the desuperheater and thence to the main turbine condenser. A portion of the steam would be used to continue operation of the boiler feed-pump turbine to maintain water circulation through the steam generators for removal of heat from the secondary salt. Another portion of the steam, if required, would be used to drive the standby power turbine-generator set to allow continued operation of the salt circulating pumps and as many as are needed of the main condensing water supply and feedwater pressure-booster pumps. In this hot standby condition, the plant would require relatlvely little time to again assume the system load. For normal shutdown the generator load is reduced until the system is essentially in the hot standby mode described above. Feedwater is supplied to as many of the steam generators as are required to remove the reactor afterheat. After about ten days the fuel salt is transferred to the primary drain tank and the steam system is allowed to cool. With completion- of the MSBR reference des1gn, preliminary studies could be made of molten-salt reactor types in which fuel conservation is given reduced emphasis and higher priority is given to o ) o) simplifying the reactor plant and its operation. Al- though the reactors would not have highest perform- ance as breeders, they would have high conversion ratios, and fuel cycle costs would be low enough to compete favorably with other converter reactors. They could use either 22*U or plutonium as the fissile fuel. Design and operation of these simplified molten-salt reactors would yield valuable experience directly appli- cable to the high-performance breeders to follow. The plants would be simplified in two major areas: (1) the Xv marily due to the (n, a) reactions associated with the lithium in the fuel salt. A large percentage of the production might reach the steam system in the MSBR reference design, but increasing the system purge gas - flow rates and the tube wall resistances and taking advantage of the chemical characteristics of the fluoro- borate secondary salt could sharply reduce the amount. Further studies are being made of the behavior of - tritium. treatment of the fuel salt would consist simply in . recovery of uranium by fluoride volatility and discard of the carrier salt after it became too contaminated for economical operation; (2) lower power densities in the reactor core would allow a graphite life equal to that of the rest of the plant, thus simplifying the reactor design, reducing the amount of equipment needed for core maintenance, and moderating the risks and nui- sances involved in disposal of a series of spent reactor - cores. One of these first-generation-type molten-salt reactors given particular study has a relatively large core (26 ft in diameter and 26 ft high) with graphite core elements and flow arrangements very similar to those used in the MSRE. The graphite is constrained in position by grids at the top and bottom, and a 2%-ftthick graphite reflector is provided for neutron economy and vessel wall protection. The core graphite would not normally require teplacement over the 30-year life of the plant. Another of the simplified types of molten=salt re- actors given brief study has a spherical reactor vessel containing 6-in.-diam graphite spheres as the moderator. Removal and replacement of the balls would probably be easier than handling the long graphite prisms used in other designs,but the concept has the disadvantage that the salt fraction in the packed bed is about 0.37, too high to be economical when the core is made large enough to obtain a 30-year life. _ The primary heat exchangers used in the MSBR reference design are of the shell-and-tube type, and the only practical maintenance method for repairing tube leaks would be replacement of the tube bundle. A brief study was made of a different heat exchanger concept employing bayonet-type tubes which could be repaired individually. The inventory of fuel salt held in the heat exchanger tubes would be increased by about 30%, but the ‘design may have merit, particularly in that the bayonet-tube construction would: greatly reduce the estimated amount of tritium reaching the steam system. A preliminary study was made of the distribution of tritium in 1000-Mw(e) MSBR power stations. About 2420 curies/day of tritium would be produced, pri- 6. Reactor Physics A new fuel cycle analysis for the reference single-fluid MSBR takes into account small modifications in system description, salt composition, and treatment of chem- ical processing that have occurred since the previous progress report. In addition, the quoted fuel cycle cost now includes the cost of replacing core graphite because of radiation damage effects. “Temperature coefficients of reactivity were calculated for the reference MSBR. The reactivity coefficient ‘associated with uniform changes in fuel salt tempera- ture is —3.5 X 107 &k/k per °C, while that associated with graphite temperature is +2.3 X 107 8§k/k per °C. The overall isothermal temperature coefficient of reac- tivity is calculated to be —1.2 X 10~5 8k/k per °C. A small amount of reactivity control can be achieved in an MSBR, without any loss of neutrons by capture in unproductive control poisons, by using graphite rods to displace fuel salt from circular passages in the graphite moderator. Because the effect of such rods, even if centrally located in the core, is expected to be quite small, and because self-shielding effects of neutron resonance cross sections are expected to be rather complex and important, reactivity worth calculations for such rods have been performed with the help of advanced Monte Carlo methods. The calculated reac- tivity effect of four such rods 4 in. in diameter grouped around- the core center line is calculated to be about 0.33% (roughly 2 dollars in reactivity), and the calcula- tions suggest that the effect of a smaller number of rods is proportional to their number. Fuel cycle calculations for several conceptual designs - of moltensalt converter reactors indicate that variations in design permitting 30-year operation without graphite - replacement, as well as variations in fuel cycle which permit - operation without continuous chemical proc- essing, may produce quite small increases in fuel cycle cost relative to the reference MSBR. A large converter reactor with a structure similar to that of the MSRE and requiring neither graphite replacement nor chemical processing (other than batch fluorination of the salt for uranium recovery) is estimated to have a fuel cycle cost ~0.1 mill/kwhr(e) greater than that of the reference MSBR. Such reactors can effectively utilize plutonium discharged from water-moderated reactors. Xvi -Calculations of the fuel conversion ratio were carried . out for different rates of chemical processing and for different fissile feed materials (333U, 235U, or PWR plutonium) for cases in which the conversion ratio is less than 1. The results -illustrate the importance of rapid removal of 233Pa as well as of fission products and indicate that plutonium discharged from light-water reactors should be a very satisfactory fuel for molten- salt converter reactors. " Detailed consideration of fuel depletion, 236U build- up, and time-dependent 24°Pa/?3°Pu ratios, as well as a careful reexamination of heat balances, now indicates that nominal full power of the MSRE was about 7.25 Mw(th), rather than 8 Mw as previously believed. | 7. Systems and Components Developir_)ent - The preparation of the conceptual system design description for the steam generator tube test stand is nearing completion. All sections have been written and are bemg combined into a rough d:aft for internal review, The 800-gpm NaBF4 test loop was operated for a total of 2954 hr during the report period to continue the investigation of the removal of salt and acid liquid ~ that is carried out of the pump bowl in the offgas stream. Four different designs of hot mist traps were tested. One type passed an excessive fraction of the solids. Another design removed the solids but plugged at the exit with a carbon-like deposit. A third removed 76% of the solids but also plugged at the discharge. The fourth design removed almost all solids and did not form a plug in the discharge. The acid liquid, thought to be a BF;-H, O compound and oil mixture, was removed - A proposal to provide the MSBE salt pumps was received from Westinghouse Electro Mechanical Divi- sion, Cheswick, Pennsylvania. The Bingham-Willamette Company and the Byron Jackson Pump Company declined to propose. The single proposal was reviewed, and a list of comments and questlons was submltted to Westinghouse. The Preliminary System Design Description (PSDD) for the salt pump test stand was completed and issued. Work on the pump test stand was suspended to be resumed at a later date. : The final fabrication of the ALPHA pump that is . being designed for small test loops is progressing. - Fabrication and assembly of the water test stand were completed. The remote welding program concentrated on para- metric studies of pipe end configurations and weld inserts to find the best joint design for use in remote welding. All the welding tests were performed with the orbital equipment, including a new and improved programmer-controller which provided very effective automatic self-regulatmg control of the welding var- 1abIes 8_. MSBR Instrumentation and Controls - Simulation studies of . the reference 1000-Mw(e) MSBR on an analog computer were continued. The basic plant components simulated were the reactor, primary heat exchanger, and steam generator. A lumped parametric model was used for the heat transfer system, and a two-delayed-neutron-group model was used for the nuclear kinetics. A provision for variable flow of the . primary salt, secondary salt, and steam, with the attendant variations in film heat transfer coefficients, - was included. The purpose of the load control system used in thlS .study was to maintain the temperature of the steam sufficiently by a cold trap operated at 32°F to prevent . fouling of the.off-gas pressure control valve. A salt leak developed in a weld joint in one of the loop pressure transmitters. The weld was defective when the trans- mitter was received from the vendor in 1956 and was repaired by back brazing. Over a period of at least eight years at a temperature of 1000°F, the braze material was leached out by the salt, openinga path through the pipe wall to atmosphere. This type of repair is no longer recommended for salt service. The fact that NaBF, was in the system was fortuitous, as any salt would have acted similarly at this temperature. No fire, explosion, or excessive corrosion occurred during the leak. delivered to the turbines at a design value of 1000°F during all steady-state conditions and within a narrow . band around this value during plant transients. It consisted of a steam temperature controller and a reactor outlet temperature controller similar to that used successfully in the MSRE. The investigation was concerned with the integrated plant response initiated by such perturbations as changes in load demand, loss of primary or secondary flow, and reactivity changes. The load demand transient results indicate a stable well-behaved -system. Normal load changes at a rate of 5%/min or less can probably be controlled by a system similar to that used on the MSRE with the addition of -y ¥ 9 the steam temperature controller. A change in load of 50%, from 100% to 50% of full load at 5%/min, produced a maximum steam temperature error of 2°F, The rate of change of salt temperatures was limited to 0.27°F/sec at the reactor outlet. Control reactivity of —0.06% 8k/k at a maximum rate of —107*% 8k/k per second was required. Transients initiated by positive and negatlve reactivity excursions of 0.15 to 0.20% &k/k were investigated. ‘The results indicate that certain reactivity transients may require additional control if undesirably low temperatures of the salts are to be avoided. For example, if an insertion of negative reactivity into the core reduces the reactor power, then the load must be ‘reduced at a rate sufficient to avoid overcooling the salts. For a positive step in reactivity of 0.15% 8k/k, the maximum rate of change of primary salt temperature was 50°F/sec occurring at the reactor outlet. In the secondary salt the maximum rate was 10°F/sec, and in the steam it was 2°F/sec. The loss of primary or secondary salt flow to a level of 10% of full flow was also investigated. The loss of salt flow in the primary or secondary salt loops decoupled the reactor system from the steam generating system. The reactor outlet temperature control system was able to control the reactor outlet temperature following the loss of primary or secondary flow with or without a subsequent reduction in load demand. If the load demand was not reduced, the control system maintained the reactor outlet temperature within 100°F of its design point of 1300°F. When a reduction in load demand to 20% followed 5 sec after the loss of flow, the controller brought the reactor outlet tempera- ture down in accordance with the accompanying reduction in its set point (1050°F at 20% load). The reactor inlet temperature, however, decreased well below the freezing point of the primary salt upon loss of the primary flow, due to the increased transit time of the salt in the primary heat exchanger, whether or not the load demand was reduced. Therefore, upon the loss of primary flow, steps must be taken to prevent a reduction in the reactor inlet temperature. Decreasing the secondary salt flow through the primary heat exchangers to transfer out less heat would probably be. the most effective way to accomplish this. The secondary salt temperatures also decreased upon loss of primary flow.' To prevent an undesirably low temperature of the cold leg, the load must be reduced sufficiently fast. Decreasing the secondary salt flow rate to control reactor inlet temperature, as discussed above, aggravates this situation, since the transit time of the secondary salt through the steam generator is increased, Xvii further lowering the secondary salt temperature. Upon loss of primary flow, then, the secondary salt flow rate must be decreased to prevent a low reactor inlet temperature, and the load must be reduced sufficiently fast to prevent low secondary salt cold leg tempera- tures. Upon loss of secondary salt flow to 10%, the reactor inlet temperature tended to increase and remain above 1050°F when the load demand was not reduced (i., constant reactor outlet temperature set point). When the load demand (and the reactor outlet temperature set point) was reduced, the inlet temperature remained above 960°F. Some additional control action may be required to maintain the inlet temperature above ‘1000°F upon loss of secondary flow. Loss of secondary salt flow rate produced undesirable decreases in the secondary salt cold leg temperatures. Therefore, as in the case of loss of primary flow, the load must be reduced at a rate sufficiently fast to prevent freezing of the secondary salt when loss of secondary salt flow rate occurs. - 9. Heat and Mass Transfer and Thermophysical Properties Heat Transfer. — An effect of heaf flux on the axial - temperature profile with a proposed MSBR melt (LiF- BeF, -ThF, -UF, ; 67.5-20-12-0.5 mole %) was observed in heat transfer tests in the low transitional flow range (Reynolds modulus approximately 4000). The effect - was manifested as a variation in the distance from the inlet of the heated test section to the position at which the local heat transfer coefficient became nearly inde- - pendent of length; this effect is believed to be related to the reduction in fluid viscosity near the wall due to heat transfer from the wall to the fluid. , An attempt to obtain heat transfer measurements in the low-Reynolds-modulus range with the test section oriented vertically for evaluation of possible natural ~convection effects was unsuccessful due to numerous salt’ leaks. A few preliminary runs were made with Reynolds moduli between 6000 and 12,000, where natural convection effects should not be significant. The heat transfer coefficients with upflow were about 12% lower than with downflow, but the difference can - be explained by errors in measurement of inlet and outlet fluid temperatures. Additional runs were made with the test section reoriented horizontally, prepar- atory to experiments with the more soluble gas helium replacing argon as cover gas. Thermophysical Properties. — The vanable-gap ther- mal conductivity apparatus was improved by the Xviii addition of a heat meter. The heat leaving the specimen can. now be determined by measurement of the axial temperature gradient in the heat meter and computa- tion of heat losses based on a two-dimensional heat- flow model. The modified apparatus will be used -to systematically study the thermal conductivity of fluo- ride salt mixtures containing LiF, BeF,, and ThF, over a wide range of compositions. From these measure- ~ments it is hoped to develop means for estimating the conductivities of the molten fluoride salt m:xtures having these same constituents. Mass Transfer to Circulating Bubbles. — Successful operation of the mass transfer apparatus enabled preliminary tests to be carried out for determination of the characteristics of the helium bubbles as a function of liquid flow rate, gas flow rate, generator probe position, and surfactant concentration. The bubble size | distribution was well described by a log-normal proba- bility function. It was also possible in the preliminary tests to establish the operating limits of the system with respect to the Reynolds and Schmidt moduli and the ~void fraction, and it was concluded that sufficient ranges of the primary variables were achieved to yield useful correlations of these variables in a fractional factorial experiment. It is significant that the expected linear variation in log .concentration with time was observed in experiments with water and with water containing a surfactant. The loop transit time, the interfacial area per unit volume (from analysis of the bubble photographs), and the mass transfer coefficient can be obtained from this linear variation. 'PART 3, CHEMISTRY 10 Chemistry of the MSRE “The cumulative results of MSRE fuel salt analyses demonstrated that at termination of reactor operations the total average corrosion sustained by the fuel circuit - extended to a depth of 0.46 mil. The major fraction of the indicated corrosion is attributed to airborne con- tamination introduced into the reactor during mainte- nance periods and when surveillance specimen arrays were removed and reinstalled in the core. Analyses of the circulating coolant salt indicated that during the four-year period of its use the coolant circuit was not corroded to any measurable extent. Correlation of the distribution of ®5Nb in the fuel system with the redox potential of the salt showed that when the concentra- . tion ratio [U3*)/[ZU] was poised at ~0.5%, disposition of °SNb toward solution in the salt or deposition within the reactor was at a null point. Comparisons of a series of 24°Puf?3%Pu isotopic ratios were made with “ ¢alculated values of the same ratio over a long operating period. From these, it was deduced that-the maximum power level of the MSRE was 7.41 + 0.05 Mw(th). Surface tension of simulated MSRE fuel and coolant mixtures was determined using the capillary depression method. Temperature coefficients, as determined by this method, were in excellent agreement with previous findings. The results permit interrelation of the labora- tory results with previous values of the MSRE fuel as determined in hot-cell experiments. ! 11. Fission Product Behavior Examination of the fourth set of surveillance spec- imens removed from the MSRE showed that omission of the use of flush salt before removal of the specimens did not significantly increase the amount of fission products deposited on or within the graphite moder- ator. One outstanding difference from previous results was the relatively small amount of °*Nb found; this was regarded to be a result of operating with the fuel in an oxidizing condition for extended periods. Investigations of the chemical behavior of fission products in the MSRE were continued using double- - walled sample capsules to minimize _cbntamin_ation. The _results suggest that the bulk of the noble metals remain accessible in the circulating loop, but with widely _ varying amounts in circulation at any particular time, ~that the proportional composition is relatively constant, indicating that the entire inventory is in substantial equilibrium with the new material being produced, and that deposits occur as an accumulation of finely divided, well-mixed material rather than as 2 p]ated layer. As compared with its distribution in the MSRE fuel circuit, *Nb was found to behave erratically in laboratory-scale experiments which sought to relate the distribution of nicbium between the fuel salt and the Hastelloy N metal surface to the redox potential of the fuel salt, Development of synthesis procedures for the noble metal fission product fluorides and oxyfluorides was continued. These materials were used in studies of the valence of njobium and molybdenum as fluorides and of their stability at low concentration in molten Li;BeF,, and in Raman, absorption, and spectropho- tometric studies of the pure compounds. 12. Properties of the Alkali FIuOro.lbOrates._ ' Investigafion of the RbF-RbBF, binary system was completed. The system was found to exhibit a single v "} ) ) eutectic, at 31.5 ‘mole % RbF, mp 442 + 2°C. By application of improved laboratory techniques, a ten- fold improvement in the purity of NaBF, was achieved. Coupons of chromium metal were found to react with contaminant moisture in NaF-NaBF, melts at moisture concentrations as low as ~100 ppm, to produce Na;CtFg¢ and hydrogen. Studies of MSBR . fuel— fluoroborate-coolant mixing reactions were initiated. Preliminary results indicated that the partial pressures of BF; resulting from such events may rise to ~150 psia. Determinations of the . high-temperature heat Coordination effects in the liquid state were demon- strated to affect the absorption spectra of tri- and tetravalent uranium in fluoride mixtures and to vary with composition and temperature. A survey of the literature was conducted to appralse the interrelation of transport processes and chemical behavior of hydrogen isotopes in MSR moderator-salt- . metal containment systems. The crystal structure of the complex fluoride com- pound (Na,Li);ThgF;, was elucidated by x-ray and “neutron diffraction. The structure was found to be content of NaBF,, KBF,, RbBF,, and CsBF; were completed. '13. Physical Chemistry of Molten Salts Continued efforts were devoted to attempts to develop ion exchangers for removal of lanthanide ions from molten fluoride mixtures. Direct measurements of the solubility of plutonium oxides in molten fluoride mixtures showed that such oxides are unlikely to be the saturating phases in MSBR fuels. Preliminary investi- gation of the CeF5-ThF, binary system disclosed that an intermediate compound (currently of undetermined stoichiometry) is formed from these components; it melts at 975°C. Solubility of CeF, in LiF-BeF, mixtures was found to be limited, with CeF3 precipi- tating from LiF-BeF, solvents with BeF, concentra- tions in the range 33 to 60 mole % as the primary phase at concentrations as low as 0.5 mole % CeF;. A recent report of dimorphism in crystalline UF, was tested experimentally and found to be erroneous. A correla- tion of activity coefficients for various ionic species in molten fluoride mixtures was devised as an extension of the polymer model for LiF-BeF, mixtures and was successfully applied in preliminary tests. Estimates of activity coefficients of the alkaline earths as chlorides or fluorides in molten bismuth were made possible through recent advances in molten-salt theory. In continuing investigation of nickel reference elec- trodes in molten Li, BeF,, successful use was made of sintered BeO as a container material for the reference atypical of those adopted by the other complex fluorides of the same stoichiometry. 14. Chemistry of Molten-Salt Reactor Fuel Reprocessing Technology ‘Studies of the distribution of Ce, Zr, and Sr between Bi and LiCl showed that these fission products will distribute ‘strongly into lithium chloride. Distribution coefficients for cesium in lithium chloride and bismuth solutions were measured at 650°C; results compared favorably with previous estimates. Initial attempts were made to effect the removal of chloride from LiF-BeF, - ThF,; melts by reaction with HF; the relative rate of removal with H,-HF sparge gas at 650 and 750°C was examined. An evaluation was made, based on thermodynamic data, of the probable chemical behavior of Na, Rb, and Cs in the MSBR reprocessing plant. Experiments were conducted to examine the potential application of metallic# titanium as a reductant for uranium and protactinium. The results were negative. The benefits of altering the composition of the bismuth alloy to include gold as a means of increasing the solubility of thorium in the alloy were examined and found to be of marginal value. Studies of the effect of platinum additions to the alloy were continued, in which it was found that electrode. Electrical conductivities were measured in LiF-BeF, melts ranging in composition from 34 to 70 mole % BeF,. The results afford an analysis of the zirconium platinide is formed as a nearly insoluble phase. Further studies were made of the extraction of cerium at low "and varying concentrations in the salt phase. The results obtained further substantiated the “free fluoride” effect on the activity of thorium in the variation of electrical conductivity with both compo- - sition and temperature. Specific conductances were measured for LiF-BeF,-ThF, melts at six compositions. Emf measurements of cells with transference were shown to have application as a means for precise determination of liquidus and solidus temperatures in multicomponent fluoride systems. salt phase. IS Development and Evaluatlon of Analytlcal Methods ~ for Molten-Salt Reactors Installation of a ‘hot-cell spectrophotometer and related equipment for studies of molten-salt reactor fuels was completed and tested with MSRE fuel. Analytical equipment was installed at the MSRE and used to measure the concentration of tritium in the various effluent gas streams from the operating reactor. A new type of reference electrode, employing single- crystal LaF;, was developed for use with molten fluorides. Continued efforts were devoted to the de- velopment of methods for assay and control of water, oxides, and tritium in fluoroborate coolant salts. PART 5. MATERIALS DEVELOPMENT 16. MSRE Surveillance Progrém HastelloyN samples were removed from the MSRE after exposure to the core for 22,533 hr at 650°C. The thermal fluence was 1.5 X 10?! neutrons/cm?®. These samples had a2 modified microstructure to a depth of about 4 mils, and this region cracked profusely when a - sample was deformed. The reactor had operated under more oxidizing conditions than normal, but the exact cause of the surface embrittlement has not been determined. Mechanical property tests on these samples revealed some further decrease in fracture strain from the previous samples irradiated to a fluence of 9.4 X 10%° neutrons/cm?. Some heats of modified Hastelloy N have been included in the surveillance program and show improved rupture life and fracture strain. '17. Graphite Studies With the essential completion of the sugvey of radiation behavior of commercially available graphites, the emphasis is turning increasingly to fabrication studies and a better understanding of the underlying phenomena. From a fundamental point of view, elec- tron bombardment of graphite single crystals in the electron microscope is proving to be an extremely useful tool. Quantitative results on interstitial cluster generation await better calibration of the microscope operating characteristics, but the displacement energy is being re.determined, and an upper limit of about 30 ev has been established. . Irradiation studies on bulk artificial graphites are continuing, with recent efforts on lamp- and furnace- black base materials and samples fabricated at ORNL. In the latter program, preliminary fabrications have been attempted using JOZ, air-blown, and Santa Maria graphite . powders as filler materials. The air-blown material looks most promising, at least from the standpoint of isotropy. XX . Samples of low-permeability graphites obtained by gaseous impregnation have shown a degradation in permeability at fluences of the order of 1 X 10?2 nvt (E > 50 kev). The undamaged samples possessed permeabilities of the order of 10™® cm?/sec (He STP), but this increased to ~1075 cm?®/sec after damage. These results are subject to difficulty in interpretation, and even the high values may still be good enough to exclude xenon from the graphite. The experiments are continuing. However, propylene-derived pyrocarbon coatings in a free state have been observed to be ‘dimensionally stable under irradiation to >2 X 10?2, and coated samples are being prepared for permeability studies. 18. Hastelloy N Alloys containing from O to 1% titanium were aged for 1500 hr at 760°C and tested at 650°C. Generally, the ductility and strength were increased. When alloys containing from 10 to 20% molybdenum were aged, the strength increased and the ductility decreased. An irradiation experiment at 550 C involving six alloys with titanium levels of 0.6, 0.9, and 1.2% each with carbon levels of 0.03 and 0.08% showed generally that the postirradiation creep properties at 650°C improved with increasing titanium and carbon levels. Postirradia-. tion expenments on alloys irradiated at 760°C and containing various amounts of Ti, Hf, Nb, and Si are incomplete, but the desired levels of these elements are being better defined. Silicon must be kept low in these alloys to obtain good properties, the precise maximum level likely being about 0.1%. This undesirable effect of silicon is associated with its tendency to form very coarse M¢C-type carbides. None of the modified alloys except those containing zirconium have given welding problems to date. : The corrosion rate of Hastelloy N in sodlurn fluoro- borate has been shown to be very dependent upon the water content of the salt, However, natural-circulation loops have continued to operate under conditions where steam was injected to cause a very high corrosion rate. A forced-convection loop constructed of Hastelloy N with sodium fluoroborate has operated over 6700 hr ~with an average corrosion rate of 1.2 mils/year and a rate during the last 2000 hr of operation of 0.7 mil/year. A process whereby BF3; and HF are bubbled through liquid fluoroborate was effective in removing water. Hastelloy N was exposed to steam at 3500 psi and 538°C for 2000 hr, and the average corrosion rate was about 0.1 mil/year. ' w) ¥ ] 19. Support for Chemical Processing Back extrusion offers several advantages as a tech- nique for fabricating containers of molybdenum. We have produced several 2.5-in.-diam capsules which have excellent surface finishes and flow patterns. Three tube-to-header molybdenum welds were made in a glove box by the gas tungsten-arc process, and no evidence of flaws was detected. The orbiting automatic gas tungsten-arc process is being evaluated as a tech- nique for field welding of molybdenum. We have shown that braze alloy composition and joint design affect the compatibility of braze alloys with bismuth. Less attack by bismuth was observed in a. . molybdenum lap joint brazed with an Fe-C-B alloy than in a T-oint brazed with the same alloy. The addition of molybdenum to Fe-C-B braze alloys also improved their corrosxon resistance, Molybdenum and TZM were unattacked by blsmuth after 3000 hr exposure in a quartz thermal convection loop test at 700°C with a AT of 95 = 5°C. Severe dissolution and mass transfer of niobium and Nb—1% Zr alloy was observed after only 115 hr under snmlar conditions. The adherence of tungsten coatings to conventional alloy substrates has been found to be related to substrate composmon rather than to differences in thermal expansion between tungste_n and the substrate. Bond adherence was evaluated by thermal cycling, spiral bend tests, and tensile tests. The effectiveness of several commercial coatmgs in preventing oxidation and scaling of carbon steel during thermal cycling from 650°C was evaluated. Croloys having from 1.1 to 8.7 wt % chromium were also thermally cycled from 650°C, and it was concluded that more than 5% chromium is requn'ed for improved oxidation behavior. 20. Support for Components Development Program = Welding parameters have been established_ that ensure good root passes in Hastelloy being welded by remote ~ welding equipment. Welds of high quality’ were ob- tained when the misalignment was as much as % ¢ in, A failed pressure measuring device from a sodium fluoro- . borate pump loop has been examined. The failure seemed due to the corrosion of a brazing alloy that had been used in repairing .a faulty weld. Four potential bearing materials have been plasma sprayed on Hastel- loy N by a commercial vendor. These coated samples have been thermal cycled and seem reasonably stable, The five failures that have been observed may be associated with oxides that are present. " PART 6. MOLTEN-SALT PROCESSING AND PREPARATION | 21. Flowsheet Analysis We have devised a new process known as the metal transfer process for removing rare-earth and alkaline- earth fission products from the fuel salt of a single-fluid MSBR. The process uses bismuth for transporting the rare earths from the fuel salt to an acceptor salt such as LiCl; the rare earths are then removed from the LiCl by contact with bismuth containing 0.05 to 0.50 mole fraction lithium. - The effective thonum—rare-earth separation factors for the various rare earths range from about 10* to about 10%, The new process does not require an electrolytic cell. This is an important advantage over the earlier reductive extraction process, which also had the disadvantage of separation factors near unity. A new process for isolating protactinium from a single-fluid MSBR has been developed. This process consists of fluorination of the salt to remove the uranium followed by reductive extraction to isolate the protactinium, Operation without an electrolytic cell is - possible since the cost of the reductant (lithium or thorium) is acceptably low. The process is highly efficient, quite stable with respect to variations in operating conditions, and requires a uranium removal efficiency in the fluorinator of less than 90%. We have adopted a flowsheet that incorporates the new processes for isolating protactinium and for re- moving the rare-earth fission products from single-fluid MSBR’s. Operating conditions which result in the same reactor performance as that obtained with the prevmus reference flowsheet are shown, . 22. Measurement of Distn'bution Coefficients in Molten-Salt—Metal Systems The distribution of thorium and rare earths between ‘molten LiCl and LiBr and liquid bismuth solutions is _being studied as part of the development of a metal ‘transfer process for removing rare-earth and other fission products from single-fluid MSBR fuels. The distribution coefficient data show that, from'a chemical 'wewpomt LiCl and LiBr will be equally good as acceptor salts in the process. Rare-earth—thorium sep- - aration factors of at least 10% are attainable with either salt. The distribution coefficients were not markedly affected by changes in temperature, Contamination of - the acceptor salt with a small amount of fluoride fuel salt significantly reduces the rare-earth—thonum sep- aration factor. The distribution of barium between LiF-BeF,-ThF, (72-16-12 mole %) and liquid bismuth was found to be nearly identical to that of europium. Preliminary studies showed that the solubility of metallic bismuth in LiF-BeF,-ThF, (72-16-12 mole %) was less than 5 ppm at 700°C. The solubility of BiF; was found to be ~ at least 4 mole % in both LiF-BeF, (66-34 mole %) and LiF-BeF, -ThF, (72-16-12 mole %) at 600°C. Graphite was the only material tested that was not rapidly attacked by BiF, dissolved in fluoride melts. 23. Engineering Development of Process Operations Minor modifications to the salt overflow from the extraction column had the desired effect of improving the flow control of both the bismuth and the salt streams. Four hydrodynamic experiments (runs 5—8) were made with the column. Run 5 was highly successful in that stable flow rates of about 80 ml/min were observed for each phase. Succeeding runs were increasingly plagued by the formation of porous iron plugs, which resulted from mass transfer of iron from ‘the carbon steel piping. The extraction column was replaced by a new column packed with ¥ -in. Raschig rings because of increased resistance to flow in the - original column. This increase in resistance was prob- ably caused by the deposition of iron within the interstices of the ¥,-in. solid cylindrical packing. Before the final hydrodynamic experiment was ‘made in the original column, zirconium metal was added to the bismuth for the purpose of inhibiting mass transfer -of iron in bismuth. . Experimental difficulties have been caused by using carbon steel as the material of construction for reduc- tive extraction systems handling bismuth. Significant progress in molybdenum fabrication techniques has led to the decision to build a small molybdenum reductive extraction system consisting of a packed extraction column with gaslift recirculation systems for pumping bismuth and salt through the column. A capital fund request has been approved for the project. The prelim- inary design is under way, and metallurgical and fabrication problems are being investigated. - A bismuth-salt interface detector is required for control of the interface location in salt-metal extraction - columns. Equipment has been assembled for testing eddy-current detectors for this application. A study of pressure drop, holdup, and flooding rates .in a 2-in-ID packed column was completed. Mercury and water were used to simulate bismuth and molten salt. Two different packing materials were studied: %,-in. solid cylindrical packing and %-in. Raschig rings. xxii The data obtained with the Y -in. packing agreed well with earlier data obtained with this packing in a 1-in.-diam column. , We are studying modifications of packed columns - which will decrease the effect of axial mixing and thereby improve column performance. The proposed . modifications. involve placing devices at various points along the column to reduce axial mixing across the column at those points. Two designs -for “backflow preventers” have been tested; these preventers resemble- inverted bubble caps. The experiments demonstrated that relatively simple devices are capable of reducing backmixing to an acceptable level (less than 15% of the net continuous-phase flow rate). - We have developed an empirical relation that predicts ~ the effect of axial mixing in both phases on the performance of a countercurrent contactor. A study of axial mixing in a 2-in.-diam open column using air and aqueous solutions of glycerol or butanol, ‘was made in support of continuous fluorinator develop- ment. Data were obtained on the effects of the viscosity and the surface tension of the continuous phase. Equipment has been fabricated for the study and demonstration of the metal transfer process for re- moving rare earths from single-fluid MSBR fuel salt. A quartz pump for circulating LiCl operated satisfactorily, - although the tests showed the need for improved methods for the initial purification of the LiCl in order to prevent attack of the quartz. ' Radio-frequency heating is being considered as a " method for generating heat in nonradioactive molten salt in studies of frozen-wall fluorinators. Approximate expressions for the rate of heat generation in molten salt and in an adjacent metal wall were derived; the - general validity of these expressions was substantiated in tests with a system in which sulfuric acid was used as a substitute for molten salt. Radio-frequency heating appears to be practical for nonradioactive experiments, although a number of problems with this approach- must still be solved. Several of the proposed flowsheets for processing -MSBR fuel salt require the use of electrolytlc cells, and the study and development of such cells is under way. To date, experimental work has been carried out in static cells. However, a facility that will allow cells to be tested under flow conditions at steady state is presently nearing completion, In tests with static cells, it was shown that a layer of ~ frozen salt can be maintained in the presence of high current densities if sufficient cooling is provided. Experiments with static cells were also carried out to identify the source of the black material that has %} ;Y " formed in the salt during all the runs made thus far. It was concluded that an ac component in the nominal dc power supply was not responsible for the formation of dark material and that the dark material does not form at cell temperatures of 675 to 680°C. It was also shown that a dark material was formed in the absence of bismuth, 24. Distillation of MSRE Fuel Carrier Salt Final analyses were obtained for the 11 condensate samples taken during the MSRE distillation experiment. The effective relative volatilities (with respect to LiF) of the major components and ®5ZrF, were essentially constant during the run and agreed with earlier meas- urements from equilibrium stills. However, the effective relative volatilities of the alkaline-earth and rare-earth fission products were substantially higher (one to two orders of magnitude) than values measured in equi- librium stills. Contamination of samples during analysis is suspected. 25, System Calculations Material and energy balance calculations for the reductive extraction process show that about 10 MW of xxiii - heat is generated in the chemical processing plant at equilibrium operation. This heat rate is about evenly divided between 233Pa decay and fission product decay. - A series of calculations was performed to investigate the effect of the removal of individual fission product elements on the performance of an MSBR. Changes in the fuel yield and fuel cycle cost are given as functions of element removal time for the most important fission product elements (Nd, Sm, Pm, and Zr). 26. Continuous Salt Purification System | Equipment is being installed to study the continuous purification of molten salt. The first purification step to be studied will be the hydrogen reduction of dissolved iron fluoride; the experiments will be carried out in a 7-ft-long, 1% -in.-diam nickel column packed with Y% -in. nickel Raschig rings. Installation of the equipment has been completed, and the system has been leak tested. 27. Design and Preparation of 23°PuF; Capsules for Small Refueling Additions to the MSRE Eight specially designed capsules were filled with 239PyF, powder for the MSRE. L }) 4l o) X Part 1. g e 23~-4y47212 Molten-Salt Reactor Experiment “P.-N, Haubenreich-< Nuclear operation of the MSRE was terminated on December 12, 1969, after all the original objectives of the reactor had been reached and many had been surpassed. (The heat production reached 13,172 equiva- lent full-power hours — over twice the 6000 EFPH originally set as the goal for demonstrating operability.) The reactor continued to run well, and as the termina- tion date, set by budget considerations, approached, experimentation was intensified, so that considerable 1. MSRE 1.1 CHRONOLOGICAL ACCOUNT OF OPERATIONS AND MAINTENANCE J.K.Franzreb T.L.Hudson - R.H. Guymon A.L Krakoviak P. H. Harley M. Richardson - At the beginning of the report period run 19 was in progress, with the reactor being operated to observe the effects of fuel pump speed and the type of cover gas (helium or argon) on entrained bubbles and !35Xe poisoning. These experiments, which had been going on. with fuel salt since August 16, were continued through September. As shown in Fig. 1.1, periods of operation at very low power (7.5 kw) to observe !35Xe decay alternated with periods at 70% of full power. (The power was limited to permit operation at reduced fuel circulation rates.) The depéndence of xenon poisoning on pump speed (amount of cover gas entrained in the loop) and the type of cover gas (widely - different solubilities) was generally as anticipated but did show the need for further refinements in the model used in the system analysis. - 'In the preceding run (run 18) there had been no fuel additions during the substantial fractional burnup of 233U required for the measurement of *23U cross- section ratios. As a result the excess reactivity had been allowed to reach a low level. Just after the start of fuel new information was developed during this report period. Part 1 of this report describes the analysis of the reactor behavior, equipment performance, and develop- ment work directly related to the MSRE. The studies of fuel chemistry in the MSRE, with emphasis on the behavior of fission products, are described in Part 3. Part 5 covers the information on reactor materials that was obtained from the MSRE. Operations salt operation in run 19, three capsules of 233U enriching salt were added to provide a wider operating margin. During September, six more capsules of fissile material were added to the fuel salt to compensate for the additional burnup and fission product accumulation expected through the end of the MSRE operation. All the additions were made in the normal manner, through the sampler-enricher with the reactor operating. For the six later additions, however, the fissile material was plutonium (94% 22°Pu). Special capsules with zir- conium windows! were used to expose powdered PuF; (about 30 g of plutonium per capsule) to the salt in the pump bowl. Dissolution of the PuF; was rather slow in the first addition, which was made with the pump at reduced speed, but satisfactory in the other five - additipns. Preparations had been made for remote gamma-ray . spectrometry of the fuel system? during operation and at the end of run 19. The purpose was to study fission product -distributions, particularly after the fuel was drained. To help make the results simpler to interpret, the reactor was operated steadily at full power for the 1MSR Program Semiann. Progr. Rept Aug. 31, 1969, ORNL-4449, pp. 245-46. 2MSR Program Semiann. Progr. Rept. Aug. 31, 1969 ORNL-4449, pp. 11-12. . ORNL=-DWG TO-3t76 l"/PM X SALT IN FUEL LOOP{ AUG SEPT t9€9 FUEL Frush ] NOV DEC Fig. 1.1. Qutline of MSRE Operations in Runs 19 and 20. flast 30 days of the run. During this period the sampler-enricher was used intensively to take samples of gas and salt, to expose various devices for measuring surface tension and fission product deposition, and to add beryllium. Apparatus that had been developed for measuring tritium in the gaseous effluents from the reactor was also put into service, and the first results were obtained. _ The scheduled shutdown began on November 2 with the initiation of a fuel drain with the reactor at full power. Scanning with the remote gamma-ray spec- trometer began within 2 hr after the fuel drain and continued around the clock for the next 17 days. Spectra were obtained repeatedly at 1-in. intervals over the primary heat exchanger and also on the fuel salt piping and the fuel off-gas ine. =~ . Plans had been made and preparations were under way to install a sampling device on the off-gas line near the fuel pump during the shutdown. This was to be used during a final two-week run in late December and early January. By November 18, however, the budget situation had become so stringent that ORNL manage- ment concluded that an earlier conclusion of MSRE operation was imperative. Therefore plans for the off-gas system work and replacement of the core specimen array were dropped, and the reactor was started up for a brief full-power run for more tritium determinations and some final sampling of the fuel salt. The reactor was taken critical for this run (run 20) on November 25. After several hours at 80% of full power to permit measurement of velocity and temperature profiles in the coolant stack at reduced air flow, the reactor was taken to full power and held there without any interruption for 16 days until the scheduled final shutdown on December 12. ' , The last run was a sustained flurry of sampling — 32 sampler-enricher operations in 16 days. These included 28 samples and exposures of various kinds, two beryllium additions, and two additions of 233U, (Ura- nium was added instead of plutonium to save a few hours on each addition.) At the same time tritium analyses were obtained on numerous samples of the radiator cooling air and the fuel off-gas. Also during the final run two measurements were made to help resolve the discrepancy in the indicated reactor power, (Salt system heat balances had been giving 8.0 Mw, while nuclide changes indicated 7.2 Mw as the maximum power.) Profiles of temperature and velocity were measured in the cooling air stack, and the gamma spectrometer was used in an attempt to measure ") a) coolant salt flow by decay of shortived activities between two points in the loop. Results (discussed in Sect. 1.2.11) more nearly agreed with the lower figute for maximum power. The reactor was shut down on December 12 by dropping the load, allowing the coolant salt to come up to temperature, then draining the fuel. Gamma-ray scanning of a spot on the primary heat exchanger started immediately after the power reduction. - Three hours after the fuel salt was drained, while the freeze valves to the fuel drain tanks were being frozen, the radioactivity in the cell atmosphere began to increase, indicating a very small leak from the primary containment. About 7 hr after the drain, the activity began to decrease after having driven the monitor on the recirculating cell atmosphere up to about 15 mr/hr. Pressurization of the fuel loop and the drain tanks had no apparent effect, and the activity continued to diminish. A sample of the cell atmosphere at this time indicated predominantly '33Xe. Since the leak ap- peared to have stopped, the flushing operation pro- ceeded. The flush salt was circulated for 17 hr to give a - thorough flushing and to permit three samples to determine its final condition. Before the flush salt was drained, the loop was pressurized to 20 psig for 2 hr without showing any sign of a leak. Suspicion was directed at the freeze valves in the fuel drain line because the leak was not evident while they were frozen. On December 15 the freeze valve, FV-105, leading to drain tank No. 2 was thawed, and the cell activity began to rise. Thus the leak appeared to be in FV-105 or its immediate vicinity. The rise continued after air was turned on to freeze the valve but leveled ~ off about 16 hr later. Although it appeared that the leak had stopped, two days later flush salt was transferred through the fill line to ensure that FV-105 and the adjacent line were full of salt. ' Examination of recorded temperatures on and near the freeze valves showed that the operation was apparently quite normal before and at the time of the leak. Radiographs of the welds at the freeze valve, taken during construction, were reexamined, and no defects were found. Thus the exact location and cause of the [eak could not be pinpointed. Consideration is being given to various methods of inspection that might be used later (fiscal year 1971). | Samples of the cell atmosphere showed about 20 to 25 curies of !32Xe, 16 mc of '3'I,and 20 mc of 321 in the cell. These amounts were low enough so that the cell could be safely vented to the stack. This was done, the cell membrane was cut, and on December 18 the experimental array in the core was rémoved. While the core access was being closed, a small amount of particulate contamination {(Ru, Te, Nb, Mo) was dis- ‘persed around the high bay of the reactor building. Mopping cleaned it up satisfactorily. Meanwhile the heat had been turned off the fuel drain tanks two days after the drain. Four days later, temperatures in the tanks reached the salt liquidus temperature. On January 5, 24 days after the drain, fuel temperatures were near the solidus, and the tank heaters were turned on at low settings to hold the salt in the range from 450 to 650°F. (This range was chosen to preclude the evolution of fluorine by radiolysis.) The system conditions for the interim period between operation and examination were planned, reviewed, and approved before Christmas, and on that day, for the first time in over five years, the reactor was left unattended. An article® describing the MSRE its operatmg his- tory, and the essence of the experience with it was prepared for the special issue of the American Nuclear Society journal that was devoted to the Molten-Salt Reactor Program. This article provides a convenient, concise picture of the MSRE experience through July 1969. Figure 1.2 (an extension of the graph in the article) outlines the MSRE operation ‘from the first experiments above a few kilowatts in January 1966 through the final shutdown in December 1969. When the fuel and flush salt began to freeze on- December 18, 1969, it was the first time that any salt had frozen in the fuel system (other than in the freeze - valves) since the first loading of molten salt into the fuel drain tanks on November 28, 1964, There had been no difficulty in keeping the salt molten throughout the intervening 44,364 hr. Other statistics as of the end of operation are given in Table 1.1. 3P. N. Haubenreich and J. R. Engel, “Experience with the Molten-Salt Reactor Experiment,” Nucl. Appl. Technol. 8, 118 (1970) Table 1.1. Final Operating Statistics of the MSRE 235y 233 Operation Operation Total Critical time, hr | 11,515 6140 17,655 Equivalent full-power hours 9,005 4167 = 13,172 Salt circulation time, hr ‘ ' : Fuel loop o 15,042 6746 21,788 Coo!ant loop 16,906 92170 26,076 i i t E { § i i P el OOES FLuse (] SALT IN . : FUEL LOOP POWER ‘ } OYNAMICS TESTS INVESTIGATE OFFGAS PLUGGING REPLACE VALVES AND FILTERS 1 RAISE POWER REPAIR SAMPLER ATTAIN FULL POWER CHECK CONTAINMENT FULL - POWER RUN -=— MAIN BLOWER FAILURE REPLACE MAIN BLOWER MELT SALT FROM GAS LINES REPLACE CORE SAMPLES . TEST CONTAINMENT RUN WITH ONE BLOWER 3 INSTALL SECOND BLOWER ROD OUT OFFGAS LINE CHECK CONTAINMENT 30-day RUN AT FULL POWER REPLACE AIR LINE DISCONNECTS SUSTAINED OPERATION " AT HIGH POWER REPLACE CORE SAMPLES TEST CONTAINMENT } REPAIR SAMPLER 0 2 4 6 8 10 POWER (Mw) SALT iN FUEL LCOP POWER 0 2.4 6 8 10 oL S POWER {(Mw) - FLuse [ ] Y Fig. 1.2, Outline of the Four Years of MSRE Power Operation. A Joe Py ! y— ORNL-DWG 69-T7293R2 | XENON STRIPPING EXPERIMENTS MAINTENANCE } INSPECTION AND .REPLACE CORE SAMPLES i TEST AND MOQDIFY + FLUORINE -DISPOSAL SYSTEM PROCESS FLUSH SALT PROCESS FUEL SALT LOAD URANIUM-233 REMOVE 1.0ADING DEVICE 233, ZERO-POWER PHYSICS EXPERIMENTS INVESTIGATE FUEL SALT BEHAVIOR CLEAR OFFGAS LINES REPAIR SAMPLER AND CONTROL ROD DRIVE 233 pYNAMICS TESTS INVESTIGATE GAS N FUEL LOOP HIGH-POWER OPERATION TO MEASURE 23U o /o REPLACE CORE SAMPLES REPAIR ROD DRIVES CLEAR OFFGAS LINES INVESTIGATE COVER GAS, XENON, AND FISSION PRODUCT BEHAVIOR ADD PLUTONIUM IRRADIATE ENCAPSULATED U ~ MAP F.P. DEPOSITION WITH GAMMA SPECTROMETER MEASURE TRITIUM, SAMPLE FUEL REMOVE CORE ARRAY PUT REACTOR IN STANDBY ") a) ) The MSRE operating staff was disbanded by January 1, 1970. Several engineers remained at the site, com- pleting the analysis of the MSRE experience and writing summary reports. MSRE maintenance specialists began planning for a limited amount of examination work to be carried out early in the next fiscal year. ' 1.2 OPERATIONS ANALYSIS Experiments conducted with the reactor during this period included those on xenon stripping, the effects of adding reducing agents to the salt, and the changes in the distribution of fission products. Meanwhile im- portant observations were made on the reactivity balance, system dynamics, the power level, and the amounts -of tritium leaving - the reactor in various streams. The chemistry experiments and much on the fission product behavior are discussed in Chaps. 10 and " 11. In Chap. 6 are discussed the measurements of uranium isotopic changes in MSRE fuel samples. The other items mentioned above are discussed in -this section. 1.2.1 Reactivity Balance ~'J.R.Engel i The zero-power reactivity balance results with 233U fuel are summarized in Fig. 1.3. Since there is still some uncertainty in the reactor power calibration, data are presented for two full-power bases: 8 Mw, derived from ‘system heat balances, and 7.25 Mw, derived from uranium isotopic-ratio changes in the 23°U operation. ‘The points with error flags were obtained when there was a substantial circulating void fraction in the loop, and the flags reflect only the uncertainty associated with the value of that parameter. These data also reflect an adjustment from previously reported results* to account for improved values for the samarium that remained in the salt from the 23°U operation. How- ever, other refinements in ‘nuclide concentrations, shown by the reevaluation of the 235U operation (see Sect. 2.3) to be needed, have not yet been incorpo- rated. These data fall into two groups, with no obvious trend within each group but with a distinct shift between 12,000 and 18,000 Mwhr. It appears likely that at least part of this shift was caused by the new experimental array that was installed in the core just before the startup for run 19 (ref. 5). The new array contained considerably more high-cross-section mate- rials than the earlier arrays of surveillance specimens, but no account was taken of this in the reactivity balance calculations. Detailed calculations are being made to evaluate the reactivity effect of changing the array. Other refinements in fission product yields and cross sections are also being made, but theu' effects are expected to be small. In spite of the apparent shift in the base line, the reactivity balance was used successfully to follow the xenon behavior (see Sect. 1.2.2) and the plutonium “MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 4-5. SMSR Program Semiann. Progr. Rept Aug. 31, 1969, ORNL-4449, pp. 14-15. ORNL-DWG 70-3175 0.4 ‘ : . BASIS ‘ ) o 1 1 1 e 8 Mw § % fl _' A 7.25 Mu 2 ' g = — . " ' uj . r { % =10 £ -o04 1 ofd > _ : ' wi . g ' | 5|5 { | <1 wi=Z fz g bl .. ~Q.2 ola T g =|E , | 3 uz-'q {{ | w . w ]‘ , @ "0.3 T o T -0.4 — - . : — . 0 4 8 12 16 20 24 28 32 (x10%) INTEGRATED POWER WITH 22%y FUEL (Mwh) Fig. 1.3. Zero-Power Reactivity Balances During Operaiion with 233U Fuel. additions, along with the routine monitoring of reactor behavior at power. The plutonium that was charged into the MSRE was added as pure PuF;, which had to dissolve in the salt (rather than melting, as did the LiF-UF, eutectic used in the uranium additions). In - addition, the openings in the plutonium capsules were sealed with - zirconium windows, which had to react with the salt before dissolution of the PuFj; could begin. ‘The combination of these factors required exposure of the capsules for 3 to 4 hr in the pump to complete an addition; the uranium enriching capsules emptied in 5 min or less. The first plutonium addition was made with the reactor at low power (10 kw) and the fuel pump running at 900 rpm to minimize reactivity effects from sources other than the plutonium itself. The reactivity - balance began to show the effect of the plutonium after the capsule had been in the pump bowl for about 10 min. The capsule was left in the pump bowl for 4 hr even though the reactivity balance showed no further detectable change was occurring after about 3 hr. ‘Subsequent examination revealed a 15.3-g salt residue in the upper part of the capsule that contained 5.5 of the 25.5 g of plutonium originally loaded into the capsule. Presumably there was simply insufficient agita- - tion with the pump running at the low speed to move salt through that region, The remaining five plutonium additions were made with the fuel pump running at full speed (1188 rpm), and essentially complete dissolution was obtained for the same exposure times; salt residues were between 04 and 0.75 g. Careful reactivity monitoring during each addition indicated that the dissolution was very orderly with no evidence of anomalous effects. _ .If we deduct the 5.5 g left in the first capsule, a total of 179 g of plutonium was added to the loop from the six capsules. This should have increased the plutonium concentration by 41 ppm with a positive reactivity effect of 0.443% 8kfk. The observed reactivity effect duplicated this value within the accuracy limits for short-term changes, +0.04% 8k/k. This represents £10% of the plutonium added or ~4 ppm in concentration. Analytical results of plutonium concentration measure- ments are presented in Sect. 10.3. 1.2.2 Xenon Poisoning J.R.Engel R, C, Steffy Observations of the xenon poisoning under various conditions were continued throughout the operation of the reactor. The previously reported® differences in the behavior of circulating bubbles with helium and argon - ORNL—-DWG 69— 10545R i, T © 5.5 Mw N o HELIUM 2 o i ] . s ARGON o8 ‘ o0 8 © ° - § osf- o = 2 ; S o o 04 1 z o o S 02° x = O : ; 0 0.4 0.2 03 0.4 05 06 0.7 CORE VOID FRACTION (%) Fig. 1.4. Obscrved Xenon Poisoning in MSRE. - cover gases had suggested that the Xenon stripping effect might also be different. Xenon poisoning was measured with both cover gases at a variety of fuel - pump speeds that gave core void fractions covering the range from zero to 0.7 vol %. (The correlation between fuel pump speed and core void- fraction had bee- established earlier at zero power.) The xenon measure- -ments were made at a reactor power level of 5.5 Mw (70% of full power) so that low pump speeds could be attained within the desired range of system tempera- tures. The results are shown in Fig. 1.4. The data with helium cover gas followed the same general pattern observed previously® at 7 Mw. However, with argon a marked difference was observed as the void fraction approached zero; instead of decreasing, the xenon poisoning continued to increase, reaching nearly 0.9% 8k/k at zero voids. - In the early predictions of the effect of circulating voids on xenon behavior,” the cover gas was treated as a totally insoluble material, and it was predicted that the poison fraction would decrease monotonically with increasing void fraction. Although the xenon poisoning observed with argon at zero voids did not quite reach the value that would be predicted by such a treatment (see Sect. 1.3) the relation to void fraction was qualitatively very similar to the prediction. Since the solubility of argon in molten salt is less than that of helium by a factor of 10, it appears possible that the deviations from the idealized treatment (minor for SMSR Program Semiann. Progr. Rept. Aug. 31, 1969, ' ORNL-4449, pp. 6-10, TMSR Program Semmnn. Progr. Rept Aug. 31, 1966, . ORNL-4037, pp. 14-16. o 1) argon but major for helium) are caused by solublhty effects. - The pronounced effects of core void fractlon on xenon poisoning that are evident in Fig. 1.4 cannot be described adequately by a simple analytical model of the fuel system. It had been shown that, for a soluble gas at low void fractions, the void fraction can vary - widely around the loop.® That such variations influence xenon behavior was illustrated by two measurements made with helium cover gas at fuel pump speeds of 600 and 900 rpm. In both cases the core void fraction was zero. However, 900 rpm is near the threshold for bubbles in the core, so voids were probably present in other parts of the loop. (The core is the only region in which small void fractions could be detected with reasonable confidence.) The xenon poisonings at 5.5 Mw for the 600- and 900-rpm pump speeds were 0.36 and 0.23% 8k/k respectively. Thus the out-of-core voids appear to have had some stripping effect. A computer program was prepared that includes the dynamic behavior of the cover gas bubbles and treats the various xenon transport pheriomena in some detail in a 23d-order system of equations, This program was designed to permit evaluation of both steady-state and transient xenon effects in an effort to find a set of system parameters that fit the observed effects. Results .are not yet available at this writing. The preceding discussion- deals with d1fferences ob- served over a relatively short period of time., However, significant long-term variations in xenon poisoning were also observed when the reactor operating conditions were nominally the same., Table 1.2 lists some typical . values observed at full power during the 223U opera- tion with full pump speed and helium cover gas; the - integrated power (based on 8 Mw full power) is shown for each point to provide a frame of reference. At least Table l 2. Full-Power Xenon Poisomng During 33y Operation Xenon Integrated Date "~ Poisoning © Power? (% 8kfk) . (Mwhr) 2/21/69 036 4,793 2/25/69 : 036 - 5,415 4/15/69 . 1 0.32 ' ' 12,460 " -4/18/69 0.40 12,880 - 10/6/69 0.51 . 24,810 10/20/69 , 0.39 27,700 11/29/69 044 30,850 12/11/69 .. 045 - 33,210 awith 233U fuel, based on 8 Mw full power. one similar change in xXenon poisoning was observed during the 23%U operation. Early in that operation (May 1966) the full-power xenon poisoning was 0.35% 8k/k; later values were in the range 0.26 to 0.28%, and near the end of the operation, it was 0.35% again. No correlation with other aspects of the operation has yet been found.. | 1.2.3 Dynamics Testing R. C. Steffy The only dynamics tests performed on the reactor during this report period were primarily for research aimed at improving frequency-response results by using test signals which had signal power only at certain predetermined frequencies. These tests were designed and analyzed by M. R, Buckner and T. W, Kerlin of the Nuclear Engineering Department of the University of Tennessee as part of a graduate studies program. The frequency-response results from these tests did verify - that the reactor was still responding in the expected manner to reactmty perturbations. 1.2.4 Operational Diagnosis by Noise Analysis R. C. Steffy ' During this report period we continued to assess the reactor’s operation by frequently analyzing the inherent fluctuations in the neutron flux and in the pump bowl pressure using detailed noise analysis techniques.® In -addition, the ondine noise monitors® which were installed just before the end of the last report period functioned well, giving a continuous indication of the average noise levels in the vicinity of 1 Hz. During run 19 the pressure noise in the pump bowl slowly increased as indicated by both the on-line monitor and the detailed analyses in the vicinity of 1 Hz. Since restrictions in the off-gas line had been known to cause increases like this,!© it was anticipated that the off-gas line might be slowly plugging. The detailed analyses of the neutron noise showed that in the vicinity of 1 Hz it also increased, but the on-line neutron noise monitor showed no increase, possibly because it was not as sensitive to changes at a particular frequency. During the first part of run 20, both the 8MSR Program Semiann. Progr. Rept. Féb. 29, 1 968, p. 32. SMSR Program Semizmn; Progr. Rept. Aug. 31, 1969, pp. 10, 11. loMSR Program Semiann, Progr. Rept. Aug. 31, 1 969 PP- 36, 37 pressure noise and neutron noise remained essentially constant at abouat the same values which they had attained at the end of run 19, The pressure noise monitor was observed to have several interesting characteristics during run 19 and the first part of run 20. Each time the sampler valves were opened, the pressure noise decreased by 10 to 20%, presumably because the effective gas volume was increased and this attenuated the magnitude of the ~pressure perturbations. Conversely, closing the over- flow-tank vent valve to burp the overflow tank (OFT) caused an increase in the pressure noise, as shown in Fig. 1.5, by reducing the effective pump bowl gas space. When an OFT burp was completed, the opening of the vent valve caused a sudden surge in pressure in the pump bowl that resulted in a large spike in the pressure _ noise, After an OFT burp on December 4 1969, it was observed that the indication from the on-ine pressure noise monitor did not return to its original value but remained about 15% higher than the original level. Then on December 6 the pressure noise level spontaneously increased by about 50%, as shown in Fig. 1.6. This increase coincided with a detectable increase in the restriction in the off-gas line near the pump bowl. From - this time until the end of run 20 the fuel pump pressure -increased slightly during burps of the OFT, whereas it had decreased when there was no restriction in the main off-gas line. Of course, this implies that a fraction of the gas flow into the pump bowl was bubbling out through the OFT. On December 8 the pressure noise increased by another 20% when the restriction in the pump bowl off-gas line increased to the point that essentially all the fuel pump off-gas began to flow through the OFT. ORNL-DWG 70-317¢ OFT VALVE OPENED WHEN BURP COMPLETED TIME (min}- - CLOSED OFT VENT VALVE . RELATIVE PRESSURE NOISE MAGNITUDE (arbitrary units) Fig. 1.5. Pressure 'Noise Monitor’s Trace During OFT Burp with the Off-Gas Line Not Restricted. There was also a sharp increase in OFT temperatures at this time because of increased fission product heating, and the character of the pressure noise during a burp. was changed significantly. The pressure noise trace for a typical burp after December 8 is shown in Fig. 1.7and may be compared with Fig. 1.5, which records a burp when -there was little restriction in the off-gas line. "ORNL-DWG 70-3!72 TIME (min) 5 8 & 8 o RELATIVE PRESSURE NOISE MAGNITUDE larbitrary units) Fig. 1.6. Pressure Noise Monitor’s Trace During Spontaneous Increase on December 6, 1969. . " ORNL-DWG 70-3173 60 z 45 E +— — w 30 OFT VALVE OPENED = WHEN BURP COMPLETED h . -l » o CLOSED OFT VENT VALVE RELATIVE . PRESSURE. NOISE MAGNITUDE ({arbitrary units) Fsg 1.7, Pressure Noise Mon:tors Trace During OFT Butp with the Off-Gas Lme Restncted. 4 3 Features to note about Fig. 1.7 include: (1) The pressure noise decreased when the OFT vent valve was closed. Closing the valve caused an increase in pressure in the OFT which stopped the gas from the fuel pump from bubbling into the OFT. The bubbling action - appears to be a strong noise source. (2) After the burp the noise level approximately returned to its original level, but the trace is seen to have a higher frequency of oscillation. After a burp the level in the OFT was quite low, and less differential pressure was required to initiate bubbling into the OFT. The pressure noise trace appears to indicate that at the lower differential pressure there was more bubbling action than for the higher differential pressure which existed with high OFT 1levels. It is possible that when the OFT level was high, differential pressure increased until a large bubble was released into the OFT and that there was no more bubbling until the pressure increased sufficiently for another large bubble. (3) The average noise level was slowly decreasing both before and after the burp, that is, between successive burps. This is thought to be a combined result of the increasing volume of the gas phase in the fuel pump bowl as more salt was transferred to the OFT and a decrease in the bubbling rate. The neutron noise, both from the on-line monitor and in the detailed analyses, remained essentially constant during all of run 20. 1.2.5 Tritium P. N. Haubenreich Tritium is produced as a fission product at a rate of about one atom per 10* fissions. In the MSRE this source amounted to about 0.1 curie per full-power day. Greater by a factor of several hundred was the production from neutron reactions with lithium. Both ~ thermal-neutron reactions in 6 Li and fast regctions (E> 2.8 Mev) in 7Li are significant. During this report period the calculated production rates for the several regions of the MSRE were reevaluated, and an intensive effort was made to measure the tritium content of the various effluent streams to detcrmme where the tntmrn was going. Calculated Production Rates. — There is lithium in the fuel salt, the coolant salt, the thermal insulation around the reactor vessel, and the corrosion inhibitor in the water that circulates through the thermal shield. The trace amount of lithium in the insulation is natural lithium (7.4% °Li), but the other lithium contains less than 0.01% ®Li. The greatest tritium production was in the fuel salt, where the neutron flux was by far the highest. The next greatest production was in the thermal insulation, with its high $Li fraction. Produc- tion rates in the coolant salt and treated water were relatively very small. The ratio of neutron flux in the fuel to reactor power can be calculated rather accurately; the same is true for the effective cross sections of ®Li and 7Li. The major uncertainties in the calculation of tritium production in the fuel are in the power measurement and in the ®Li ~ fraction. The production rates in Table 1.3 are based on full power being 7.25 Mw. The SLi fractions were obtained from assays of the LiF used to make up the fuel salt, the probable natural lithium content of the 'BeF, that was used, and the calculated burnup of the ‘high-crosssection ®Li. (The fractions used are appro- priate for the end of 22U operation and the end of 233y operation.) The production from ®Li was much “higher with 233U fuel because of the higher thermal- neutron flux resulting from the lower fissile concentra- tion. Since the fast flux changed only slightly between 235U and 233U operation, the production from 7Li is similar for both fuels. The coolant salt production was due almost entirely to exposure to neutrons from the fuel in the primary heat exchanger (including delayed neutrons and neu- trons from fissions within the heat exchanger). The calculated rate was lower with 233U because of the smaller delayed neutron fraction. The big uncertainty in the calculation of production from the-thermal insulation is the lithium content. Analysis of batches of insulation that were to be used in the MSRE showed 1000 ppm Li with an estimated accuracy of a factor of 2. Later analyses of material nominally the same gave around 10 ppm Li, The 3 curies/day is based on 1000 ppm Li. Table 1.3, Calculated Rates of Tritium Production in MSRE Production Rate Source A (curies/full-power day?) | " | ?*5UFue 2y Fue Fuel salt® ' , SLi o 20 35 Li | . 4 , 5 Ufission 01 0.1 - Total 4 40 Coolantsalt, total =~ 0.0002 0.0001 Insulation, total : - 3%3 - .3%3 Treated water, total 0,005 0,005 ‘Full power is taken to be 7.25 Mw. bCalculations were based on 0 005 1% 6L1 in, fuel salt lithium during 235y operation; 0 0048% ®Li dunng 33y operation. Observed Amounts, — Tritium was obser\}ed to build up in the treated water system, to occur in condensate 10 from the containment cell atmosphere, and to leave the reactor in the fuel salt off-gas, the coolant salt off-gas, and the air flowing across the coolant salt radiator. = The tritium in the 4000-gal treated water system gradually built up over the years to about 0.14 uc/ml. The changes with time were consistent W1th the calculated production rate of S mc/day. Moisture condensed from the containment cell atmos- phere generally contained around 1.3 mc of tritium per milliliter. The condensate was collected (with other liquid wastes) and sent in batches to the ORNL waste disposal system. Records show tritium accumulating in the MSRE waste tank at rates of 4 to 6 curies per full-power day over, several intervals of a few months each., _ The tritium sampling apparatus 'that was . developed for use on the gaseous effluent streams consisted of a heated bed of CuO followed by a refrigerated trap. The - first analyses were obtained with the reactor at full power near the end of run 19. The fuel off-gas sample ~ point was downstream of the charcoal bed, where the fission product concentrations were manageable. The first sample from here, with the CuO at 340°C; showed 9 curies of tritium per day. Raising the CuO tempera- ture to 800°C, so as to collect tritium in all hydro- carbons, showed 23 curies/day. On November 21, 19 days after the power had been shut down, tritium analyses of the fuel off-gas showed concentrations over half of what they had been at power. (The cover gas flow was less, so the tritium flow out of the off-gas system was down about a factor of 3.) After the reactor was takén back to power in run 20, the tritium in the fuel off-gas gradually came back up. Samples with the CuO at 800°C showed 11.6 curies/day six days after the power was raised and 15.0 curies/day after 16 days at - full power. The only sample taken from the coolant off-gas was near the end of run 19 and showed 0.6 curie/day ~ passing the sample point.. Tritium in the form. of moisture in the radiator cooling air was first measured early in October, using calcium chloride to trap moisture from the air in the - stack. Results of six samples ranged from 1 to 3 curies/day. The CuO apparatus could not be used effectively on the stack air to measure total tritium because of the extremely low concentration. In an . attempt of circumvent this problem, a 2-ft length of radiator tube was fitted with a jacket from which air ~ could be drawn to the tritium sampler. Sample results obtained in run 20, scaled up by the ratio of total - surface area to jacketed area, indicated only 0.2 t0 0.6 curies/day coming out of the radiator. Just before the’ final shutdown, four samples of stack air collected in large bulbs indicated a total tritium flow up the stack of - 3.3 to 4.6 curies/day. Attempts in run 20 to repeat the moisture measurements by the calcium chloride absorp- tion method gave widely scattered results (0 6 to 18 curies/day). \ , The containment cell exhaust was sampled with calcium chloride and also was analyzed with the CuO apparatus. The former indicated 10 mc/day and the latter 2.4 me/day. Comparison, — Uncertainties in the SLi content of the fuel salt and the reactor power level introduce 2 probable error of about 16% in the calculated produc- tion of tritium in the fuel salt (40 £ 6 curies/day). The observations on the off-gas downstream of the charcoal beds indicated that there was substantial holdup in the off-gas system but that the tritium effluent would gradually build up in a long run at full power to 2527 curies/day. This is between 60 and 70% of the calculated production rate in the fuel salt. ' Some of the tritium produeed in the fuel was certainly diffusing through the heat exchanger tubes’ into the coolant salt, The amount of tritium in the coolant salt off-gas, although relatively small, was more - than the calculated production in the coolant salt. More significant was the tritium that came out through the radiator tubes. It appears that around 10% of the tritium produced in the fuel salt left the reactor by this route. Roughly half of the tritium in the cooling air stack appeared to be in the form of moisture (based on comparison of calcium chloride collection with total tritium). The sleeve samples suggest that the tritium diffusion rate out of the tube inside the jacket was substantially less than the average for all the tubes exposed to the main air stream, ' The rate at which tritium appeared in the contain- ment cell (almost entirely in moisture condensed from the humid atmosphere) was about 8 to 13% of the calculated production rate in the fuel salt. This must be regarded only as an upper limit on the amount diffusing out of the salt systems, however, because production in the insulation around the reactor conceivably could have accounted for this much tritium. Although the sum of the observed effluent rates is 12 | to 25% less than the calculated total production 'r_ate, the discrepancy is hardly greater than the probable | error in measurement and calculations. ' 2 11 1.2.6 Fission Product Distribution in the Fuel System A.Houtzeel R.Blumberg F.F.Dyer Gamina-ray spectrometry studies of fission product distribution in the fuel system were continued during this period with the following progress: ' 1. Gamma-ray spectra were taken of the main reactor off-gas line (line 522) near the fuel pump with the reactor at different operating conditions. 2. Calibration of the spectrometer equipment with 110m Aq and 226 Ra sources was completed. 3. Several components of the MSRE system were scanned during the November shutdown period and . after the final reactor shutdown. 4. Work was continued on a computer program to automatically analyze the spectra. 5. Some preliminary analyses were made of the gamma-ray spectra taken during the June—July 1969 reactor shutdown. The equipment for this work was described previ- ously.!! Figure 1.8 is a photograph of the detector system in operation, Reactor Off-Gas Line, — The gamma-ray detector, together with the collimator body and insert, was set up over a hole in the reactor shield blocks. Through this hole it was possible to aim the detector directly at the main reactor off-gas line, with only the cell membrane interposed, at a location about 1 ft downstream of the fuel pump. The purpose of this experiment was to study any variation in the fission products contained in 1p5R Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 1112, 3133, Fig. 1.8. Remote Gamma-Ray Spectrometer on Portable Maintenance Shield, the off-gas line during different reactor conditions; xenon behavior was of primary interest. During this experiment the reactor power was varied from zero to 7.5 Mw; other variables were the fuel pump speed and the cover gas flow. Spectra were also taken when fuel salt was returned from the overflow tank to the pump bowl, during a beryllium exposure in the fuel system, - and when argon was used instead of helium for the cover gas. Altogether some 170 different gamma-ray 12 of their daughters, but it also shows clearly the presence of noble metals. Calibration. — An apprecmble effort was devoted to calibration of the equipment. The purpose was not only to determine the counting efficiency of the equipment in the existing geometry but also to establish the self-shielding effect of the MSRE heat exchanger along with the heater elements. We used two sources for this ~ calibration work, a small 226Ra source to determine spectra were recorded at this location. Another four spectra were taken from 2 sample bomb in the main reactor off-gas line about 45 to 50 min downstream from the fuel pump. ~ The main problem encountered was the very intense activity emitted by the reactor off-gas line. When the reactor was at power, the radiation beam was more than 1000 r/hr directly above the hole in the shielding; the neutron flux was also appreciable, By using the % g-in.- diam collimator, 2 in. of lithium-impregnated paraffin, and a Y-in. lead plate, we were able to keep the counting rate down and the detector-system dead time ~ within reasonable limits. It appeared that the then available computer program would not adequately analyze these data because there were too many photopeaks in the spectra (many of these peaks also overlapped, i.e. multiplets), so the detailed analysis was postponed. The results of a preliminary manual analysis are given in Table 1.4, This listing shows the expected noble gases and at least some - Table 14. Nuclides ldentified in & Spectrum Taken from the MSRE Off-Gas Line Reactor power, 5.5 Mw; September 1, 1969 Elements Chain 87 Kr . 88 K1, Ru 89 - Kr,Ru 90 Kr 95 Nb 99 Mo 103 Ru 106 Ru 129 Te (m) 131 Te, 1 132 Te, 1 133 Te{m) 135 Xe(m) 137 Xe, Cs 138 Xe,Cs 139 Xe, Cs , 140 Xe,Cs,Ba, La the detector sensitivity over a wide energy range and a 110m Ag source of approximately 25 curies to evaluate the geometry and shielding effects. The 6.4-in.-long by 0.5-in.-diam silver source was placed in all the different tube positions of the heat exchanger mockup. This way it was possible to determine the effect on the detector of activity in each of the different tube positions. The silver source was also used to determine the detector counting efficiency for the reactor off-gas line geom- etry. Shielding experiments were done for the different shielding materials used (lithium-impregnated paraffin, lead, copper, aluminum, cadmium, and a heater ele- _ ment). Since the source was made up of three short silver - ~ tubes that had been activated in the ORR, it was thought necessary to determine the activity along the ‘length of the source to establish an average source strength in relation to the detector efficiency for a given collimator geometry. Local variations of 13% from the average were found in the source. However, when the normal field of view at the source was considered, the effect of these variations on count rate at the detector was much smaller. Approximately ‘435 calibration and related spectra were taken, and the analysis is 95% complete. These calibration data will be used as input data for the automated computer analysis of the fission product spectra. ' - Scanning During Reactor Shutdowns. — The shut- down scanning program was started just prior to the reactor shutdown on November 2 and continued for almost three weeks in an around-the-clock operation. The reactor was drained from full power, and no flush salt was circulated. Gamma spectra were taken pri- . marily from the MSRE heat exchanger and main reactor off-gas line; other spectra were taken from the fuel pump bowl, the drain tank, and two fuel salt lines (lines - 101 and 102). Multiple spectra were taken to facilitate the identification of both short-lived and longer-ived species. This approach also tended to reduce statlstlcal " errors. 13 Table 1.5, Fission Product Residues in MSRE Heat Exchanger and Main Reactor Off-Gas Line " Four to five weeks after reactor shutdown 9SNB 103p . 106p 1370, Residue, curiesfin.? : Heat exchanger,av = 0.50 0.073 0.0038 522 off-gas line 0,73 2.13 0,17 0.48 (max) Residue, atoms/in,? o Heat exchanger,av 8.0 10'® 132 10'® o065 10! 522 off-gas line 117 107 388 10'7 283 10'7 2.42'? (max) Fission yield 64 29 0.39 ‘5.9 Half-life 35days - 40 days -1.01 years 30,1 years . Note: Other elements found in both the MSRE primary heat exchanger and reactor off-gas line were 95Mo, 129Te, 131l, ’321, l'")Ba, and *4%La, ~ In the first 72 hr, gamma spectra were taken from the main reactor. off-gas line, heat exchanger, and drain tank through the holes in the shielding blocks. By then the top shielding blocks had been removed, and the remote maintenance shield was used above the reactor cell. Especially during the first hours-after shutdown, extra shielding was necessary to keep the dead time of the detector system within reasonable limits. For example, we used 2 in. of lithium-impregnated paraffin and 1 in. of copper together with the ¥ ¢-in. collimator insert. Two weeks later, a 1-in. aluminum shield could be used with the ¥-in. collimator. , During the first three weeks, 235 spectra were taken from the heat exchanger, mostly along the longitudinal center line. Ninety spectra were taken along the first few feet of the main reactor off-gas line immediately downstream of the pump. Fourteen spectra were taken from the drain tank and a total of 35 from the main fuel lines (lines 101 and 102) and the fuel pump bowl. It is obvious that the drain tank, fuel lines, and pump bowl data can only furnish qualitative information, toi'y' (The program had originally been developed at the Lawrence Radiation Laboratory for use on CDC equip- ment.) This program was adapted by the ORNL . Mathematics Division to handle our data on ORNL. equipment and has been used to evaluate the calibration spectra. These results will be used with the same . program for a complete evaluation of the fission product spectra. Results. — None of the fission product spectra has . been completely analyzed at this writing, However, some of the data obtained from the off-gas line and the heat exchanger during the June—July 1969 shutdown were treated manually. The results, shown in Table 1.5, give some preliminary data for fission product residues - four to five weeks after reactor shutdown. For compari- since the calibration mockups did not include these geometries. Most of the spectra taken during and after the final MSRE shutdown were from the heat exchanger. The principal purpose was to study the effect of the flush salt on the fission product residues, - Computer Analysis. — It is clear that detailed analysis of the large number of gamma spectra described above can be performed effectively only with a digital computer. The job is further complicated by the large number of photopeaks in each spectrum. (Many spectra contain more than 200 peaks with several multiplets.) A program capable- of such analyses on IBM 360 com- son, the data taken in November indicate that the major activities after short cooling are !32I and ??Mo, isotopes that had completely decayed prior to the earlier measurements, 1 27 Graphite Samples chposéd in Fuel Pump C. H. Gabbard - The remote gamma spectrometer! 2 was set up on the sampler-enricher for about four days, and a series of gamma scans were taken on a special graphite sample, a - copper dummy capsule, and a standard 10-g salt sample. - These data were taken to study the deposition of ~ shortived fission products. The graphite specimen was contained in a special . capsule that would shield the graphite from the fuel puters was obtained from Argonne National Labora- 12)1sR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 31. . pump atmosphere during insertion and withdrawal through the gas phase but would allow the graphite to be exposed while the capsule was submerged in the salt. ' The samples were placed on a special fixture in the 3A 14 area of the sampler-enricher directly below the removal - tube. The graphite capsule could be opened on this fixture, exposing the bare graphite for. counting. The specimens were counted through a Plexiglas window . which maintained the containment of the 3A area while the removal valve was opened for counting. We were able to begin counting the specimens about - 50 min after the exposure was complete, so that any nuclides with halflives of 10 to 15 min or greater should be detectable. About 20 gamma spectra were recorded on magnetic tape, but no analysis of these data has been completed at this time. The graphite specimen was submitted for radiochemical analysis of some of the longer-lived nuclides, which will be used to convert the count-rate data to surface activity of the specimen. '1,2.8 Evaluation of Leak in Primary System R.H. Guymon P.N. Haubenreich | - The occurrence of the leak was clearly revealed by the cell air activity monitors. The general location of the leak was narrowed to the vicinity of freeze valve 105 by the response of these monitors to the freezing and thawing of the salt in this valve and adjacent piping. The leak could be effectively stopped by freezmg salt inside the pipe. Evidence for this is the lack of response of the cell air activity monitors to pressurization of the fuel system on December 12 and 13 after FV-105 was frozen. Leakage occurred again when FV-105 was thawed and the line was blown down on December 14, Both because of the contamination in the monitor and because the activity in the air was an unidentified mixture of fission products, the cell air monitor readings could not be used as an accurate measure of the total activity in the cell air. Self-consistent and apparently reliable information on this was obtained from gamma-spectrometric analysis of samples of the - containment atmosphere. The amount of activity that leaked into the cell was a tiny fraction of the total in the reactor. The 5.3-day 133%e in the cell amounted to 7 X .107° of the total inventory, The 8.0-day !3'I observed in the cell . atmosphere was less than 5 X 1077 of the reactor mventory - The amount of gas that leaked into the cell was estimated from the ! >3Xe. Assuming that the contents of the gas space above the salt during operation mixed uniformly with the gas entering the fuel loop as the salt drained, the xenon in the cell was equivalent to about 3 ft> of the mixture. The '*'I in the cell atmosphere was the equivalent of the amount in about 1 cm® of salt. 1.2.9 Salt Transfer to the Overflow Tank J.R. Engel During this final period of operation the transfer of salt from the pump bowl to the overflow tank followed the same qualitative pattern established earlier in the 2337J operation,! 3 with the rate varying as a function of the indicated level in the fuel pump. However, the level dependence was significantly stronger near the end of the operation. Figure 1.9 shows observed salt - transfer rates for two periods separated by about one When the cell air monitors did not decrease as expected after FV-105 was refrozen, it was suspected that the line was not full of salt, so some flush salt was put into ‘the line on top of the frozen fuel. Comparison of 133Xe concentrations in samples taken on December 15 and 16 showed, however, that the leak was nearly if not completely stopped even before this step. ~ The cell air monitors erred in the high direction at times because of radioactive contaminant that accumu- lated in the line at the monitors. This was apparent . from otherwise inexplicable drops in reading which sometimes occurred when the air flow by the monitors. was suddenly increased. The probable source was discovered when small amounts of water containing much radioactivity appeared at a nearby gas sample point. year. The earlier data were originally reported as a function of the level indicated by LE-593 (the shorter of the two bubblers in the pump bowl). However, that element suffered a substantial zero shift, so those results were reevaluated, and all the data are referred to the other bubbler (LE-596). Small zero shifts (1 to 2%) still leave some uncertainty in the relative positions of the two sets of data, but they do not affect the slope of the relations between overflow rate and indicated level. The estimated slopes differ by more than a factor of 2. All of this behavior is in marked contrast to that observed during the 235U operation, when the transfer rate was low and largely independent of the pump bowl level over the same indicated range. 13psR Program Semiann. Progr. Rept, Feb. 28, 1969, ORNL-4396, pp. 21—22. ORNL—-DWG 70-4807 44 - | { © OCT-NOV, 1968 ‘ ® OCT—DEC, 1969 12 & / o/ . / . 10 = ... % w / | o & 8 { 2 o et I° S LIFYA © & o Q, ul . 3 -6 3 oB‘/ o 5 |3/ ¢ < /% O w0 @ ‘,’/.u o L 4 s T .07 ° ': o O.I o & > % 0 o -0 o °, o o 0 - 45 : 50 55 60 65 70 INDICATED FUEL PUMP LEVEL, LR-596 (%) Fig. 1.9. Effect of Indicated Fuel Pump Level on Salt Transfer to the Overflow Tank During 233U Operation. 1.2.10 Radiation Heating C. H. Gabbard ‘The temperature differences between the reactor inlet and the lower head and between the inlet and the core support flange have been monitored as an indication of any sedimentation buildup within the reactor vessel. These temperature differences are shown.in Table 1.6 15 Table 1.6. Power-Dependent Temperature Differences Between Fuel Salt Entering and Points on the Reactor Vessel Temperature Difference (°FIMW) Run No. Date Core Support Flange Lower Head 6 4/66-5/66 1.90 1.39 7 1/67-5/67 1,93 1.35 12 6/67-8/67 1.98 : 1.40 14 9/67-2/68 2,03 1.28 17 1/69-4/69 231 1.54 18 4/69—-6/69 2,33 ‘1,57 19 9/69-11/69 241 1.54 © 20 11/69-12/69 2.58 - 1.53 is being more fully investigated.) In any event the indicated average temperature difference at full power was only 2.2°F higher than at the beginning of 233U power operationin run 17, ' 1.2.11 Systexfi Heat Balance C.H. Gabbard Efforts were continued in the attempts to resolve the discrepancy between the reactor power indicated by the heat balance and the power indicated by observed changes in the isotopic ratios of uranium and plutonium nuclides (Sects. 10.3 and 10.4). Two tests were con- ‘ducted which verified that the thermocouples reading the inlet and outlet salt temperatures at the radiator were satisfactorily accurate. The tests indicated that the thermocouple biasing procedure was applicable over the full operating temperature range of the coolant system and that the thermocouples were not affected by air leakage from the radiator enclosure. Additional efforts to resolve the power discrepancy included two air heat ~ balances, an attempt to measure the coolant salt flow for all periods of significant power operation. The change that accompanied the transition from 235U to 2337 fuel between runs 14 and 17 is clearly evident. This is attributed to the increased neutron leakage and fission heat generation in salt in peripheral regions ~ during 233U operation. Aside from this step, the temperatures in the lower head show no significant change with time. There does appear to have been a trend upward in the thermocouple readings adjacent to rate by radioactive decay of activation products in the “salt, and a recalculation of the head loss of the coolant system, Air Heat Balance. — The radxator air system prowdes an opportunity to make an independent measurement of the operating power of the reactor. Heat balances on ~ the air system which had been made in 1966 (ref. 14) the core support flange. However, the reactor vessel | thermocouples were not rebiased at the beginning of run 20, and the apparent increase in temperatures at the - core support flange during this run may be from thermocouple error rather than from other causes. (This indicated powers in rather good agreement with those calculated later from salt heat balances using the latest value for the coolant salt specific heat.!® The stack air outlet temperature for these early heat balances was 14ysR Progmm Semiann. Progr. Rept Aug. 31, 1 966, ORNL-4037, pp. 26-27. A5MSR Program Semiann. Progr. Rept. Aug. 31, 1968, ORNL-4344, pp. 24--26. measured at a single point near the stack wall, and the aitr flow measurement was made with a single-point device calibrated when the stack was cold. Since the air temperature could be nonuniform across the stack and since the -velocity distribution could also change at operating conditions, early in run 20 two sets of air heat balance data were taken which included tempera- ture and velocity profiles across two perpendicular diameters of the stack. One set of data was taken at nominal full reactor power, and the other was taken at the maximum power obtainable with one blower. The results of these two air-heat balances were 6.98 and 4.82 Mw, respectively, as compared with 7.96 and 6.31 Mw from salt heat balances calculated by the computer. The - corresponding values based on isotopic changes would be 7.25 and 5.75 Mw. Thus the new air heat balances did not agree exactly with either of the other indica- tions. No explanation has been found for the greater disagreement at the lower power level. The reactor power levels indicated by these air heat ‘balances are lower than indicated by the previous air heat balances (which were more in agreement with the salt heat balance) mainly because of lower indicated stack air velocities from a new calibration of the Pitot-Venturi air flowmeter., The Pitot-Venturi was originally installed permanently in the center of the stack. Since the air velocity was not uniform across the stack, the .calibration data provided with the Pitot- Venturi were not directly applicable. The relation between its reading and the average stack velocity had been determined by velocity traverses with a hot-wire anemometer at several flow rates with the stack cold. Because of some apparent discrepancies between the stack calibration and the manufacturer’s calibration, the Pitot-Venturi was recalibrated in connection with the run 20 measurements and was used to measure the - velocity profiles directly. The new calibration, which was in general agreement with the manufacturer’s data, was adopted. The new calibration gave flows about 10 - to 11% below the stack mhbratlon that had been used previously, Coolant Salt Flow Measurement by Decay of Activa- tion Products. — The remote gamma-spectrometry equipment, which was on hand for studying fission product distributions in the fuel system components and piping, was used in run 20 in an attempt to measure the coolant salt flow rate by the decay of ' *N and 2°F, (The 7.1-sec '*N and the 11-sec 2°F in the coolant salt are formed by neutron reactions with fluorine in the ~ primary heat exchanger.) Two holes were drilled in the high-bay floor to permit scanning the coolant salt 16 piping -at two locations in the coolant cell while the reactor was operating at full power. The two locations were separated by a total circulating salt volume of 20.9 " ft3, which would give decays of 1 and 1.5 halfives respectively for 2°F and ! °N at the design flow rate. The detector was located near the periphery of the reactor cell top shield blocks, where a relatively high - background of both neutrons and gamma radiation was unavoidable.!® Although lead bricks were stacked - around the detector to reduce the background as much as possible, a relatively large ‘% -in.-diam collimator was required to get a satisfactory count rate above back- ground. We expected to find activities of '®Nand 2°F and had hoped to find a longerlived activity of some" impurity in the coolant salt that could have been used to evaluate a geometry calibration factor between the two scan points. However, the only usable activity that was found was the 1.63-Mev peak from ?°F. Peaks were found at the proper energies for N, but they did not appear to be coming from the salt, since they were not significantly attenuated by about 6 in. of lead placed between the detector and the coolant piping. None of ~ the peaks other than the 1.63 Mev from *°F were significantly attenuated by this lead, and the peaks were believed to be from capture gammas that constitute part of the background radiation. Because the data were taken during the last few days of power operation of the MSRE, there was no time to refine the technique - after analysis of the data. The count-rate data were analyzed by the computer program mentioned in Sect. 1.2.6. The count rate, or 20F concentration, at the downstream counting station was adjusted to account for the salt density change in ‘passing through the radiator and for the differences in reactor power that existed when the two sets of data were taken. Corrections were also applied to the downstream count rate to account for mixing due to the bypass stream through the coolant pump tank and for the line 205 flow that bypassed the radiator volume. The residence time and coolant flow rate were then calculated from the relative count rates at the two counting stations. The coolant salt flow rate calculated this way was 610 gpm, well below the 850 gpm indicated by the Venturi flowmeter in the salt line. Although the coolant flow calculated from the 2°F decay is about 20 to 30% below what is believed to be the most probable value for the flow rate, this technique for measuring the flow appears to be feasible 16The background was high only by comparison with the detector efficiency and the radiation from the coolant lines. The total biological dose rate was less than 2 millirems/hr. c if proper precautions are taken in setting up the experiment. The large errors in this particular meas- ‘urement could be caused by differences in the counting geometry or background at the two scan points, errors in the system volume, or counting statistics. Coolant System Head Loss. — The only important uncertainty remaining in the salt heat balance is in the coolant salt flow rate, and the accuracy of the primary flow measurement device cannot be determined until the differential pressure cells and flow transmitter are tested next fiscal year. The flow rate indicated by the Venturi has been trusted, partly because it was consist- | ent with the design value predicted from the calculated head loss of the coolant system and the coolant pump performance. After the various measurements in run 20, the head loss of the coolant system was recalculated at the design flow rate to determine if this agreement was valid. Instead of the original design value of 78 ft, the new calculation gave a head loss at 850 gpm of 94 to 99 ft (allowing for a *15% uncertainty in the salt vis- cosity). A somewhat larger uncertainty band is prob- ably needed to account for a selection of the friction factor and other unknowns. The 99-ft head would give a predicted flow of about 800 gpm based on the performance of the coolant pump in water tests. The flow is not much less than the nominal flow, because the original design had allowed for a 10% greater head loss than the 78 ft calculated. Although the use of this lower predicted coolant flow rate would reduce the w Q Q 17 power discrepfincy, the final verdict on the power level should be deferred until the differential pressure cells on the Venturi are checked. 1.2.12 Heat Transfer in Primary Heat Exchanger C. H. Gabbard ' There have been no absolute measurements of the overall heat transfer coefficient of the main heat exchanger since March 1968, near the end of the 235U " power operation. However, the relative performance of the main heat exchanger is indicated by the heat - transfer index taken at full power and at full fuel pump speed. The heat transfer index is the ratio of reactor power to the temperature difference between the fuel outlet from the reactor and the coolant outlet from the radiator. Figure 1.10 shows the measured heat transfer coefficients and the heat transfer index since January '1967. These data indicate that the performance of the heat exchanger has remained constant since the begin- ning of power operation. ‘ 1.2.13 Thermal Cycle History C. H. Gabbard The final accumulated thermal cycle history for the various components sensitive to thermal damage is shown in Table 1.7. The larger number of power cycles on the fuel system was caused by the temperature ORNL-DWG 70-3174 ~ Q o o o o 7 X m AT TRANSFER COEFFIC IENT o o © H O o X o —' — I P = SFER INDEX d o Q N Q o - HEAT TRANSFER COEFFICIENT (Btu hr—! $1=2 of =) HEAT TRANSFER INDEX (Mw/°F) . 1967 Jivlalsiolnlplulrimlatmliululalsiolnlo 1968 _ 1969 Fig. 1.10. Observed Performance of MSRE Heat Exchanger. 18 . Table L.7. MSRE C_umulétive Thermal Cycle History Through Run 20 - Number of Equivalent Cycles ' Component : - Thaw , P Heé:):;'d Fll)umii"l:d Power 01(1):.;1&‘1 Thaw and ' Transfer Fuel system 13 55 101 Coolant system - - 11 18 .97 .Fuel pump : 16 51 ~ 101 711 Coolant pump - 12 19 .97 156 Freeze flanges 100, 101, 102 13 51 101 o Freeze flanges 200, 201 12 i8 97 - Penetrations 200, 201 12 18 97 Freeze valve . : 103 - 13 29 62 - 104 21 12 34 105 | 22 20 57 106 ' ‘ 23 34 44 107 15 14 22 108 . 16 - 17 28 109 ‘ 15 23 30 110 _ 8 4 10 111 _ 6 4 6 112 - ' 2 1 2. 204 12 15 42 13 41 206 - 12 coefficient of reactivity tests and other operations where the fuel system temperature was changed appre- ciably while the coolant system was drained. These thermal cycles were expressed as equivalent power cycles. . 1.3 EQUIPMENT 1.3.1 Salt Samplers A. 1. Krakoviak The sampler-enricher was used intensively to obtain a wide variety of samples during the three months of actual reactor operations in this report period. A total six PuF; and two 233UF,-LiF fuel additions. This brings the number of sampling cycles to a grand total of 745, of which 152 were fuel additions of either uranium or plutonium. A description and the number of each type of sample taken or addition made during “this report period are tabulated below. Some of the capsules were exposed as much as 12 hr in elther the - salt or cover gas in the pump bowl. ~ Freeze valve gas samples : 23 10-g salt samples for compositional analyses 13 Freeze valve salt samples » ' 11 . 50-g samples for oxide and other detérmmatlons "Solid nickel bars for tritium analyses "~ Niobium metal addition of 96 sampling operations were performed, including Evacuated capsules containing CuO, Pd, or Ni 50-g samples for U/U™ analyses PuF ;3 (powder) addition capsules Graphite and copper capsules for gamma scanmng Surface tension capsules (including two containing Be) Fission product plating capsules Hinged capsules for U /U"’F (spectrophotometncally) Addition capsules containing 23 UF4 *LiF » Beryllium additions Empty nickel cage Capsule contalmng electron microscope screens b b s b DWW W R AL N During the latter part of run 19, routine sampling was - suspended for approximately four days, and the sam- pler was adapted to accommodate a collimator and a germanium crystal detector atop the sampler where the carrier cask is normally positioned. A plastic plug in the ‘removal area permitted an unobstructed view into area 3A, where samples retrieved from the fuel pump were positioned. Gamma-ray spectrometry data were then collected on short-lived fuel fission products within 50 min of their removal from the circulating salt stream. -Although minor annoying problems were encountered with the sampler and although repairs caused a few days’ delay in the rather heavy sampling schedule, all of the planned samples and additions were accomplished. The main problems encountered concerned repair of the manipulator and -of the two flexible containment membranes (manipulator boots) between the manipu- lator and the main containment box (3A). As in the previous report period,! 7 the boots were replaced three times; one replacement was due to a small leak in the outer, or larger, boot, and two were due to ruptures in the inner boot (in contact with the manipulator). After the second failure, the boots were modified by length- ening the boot by 2 in. at the small end and increasing the boot thickness by 0.003 in. The negative ‘pressure support rings which keep the inner boot away from the manipulator rod were also eliminated on subsequent boot installations. During the early part of run 19, a small leak developed in one of the convolutions of the metal bellows which provides containment between the ma- nipulator finger mechanism (in area 3A) and the actuating rod (operating area). Although this bellows - was installed during the summer of 1968 after the 19 capsule retrieval work was completed,'® it had been - - decontaminated in an acid bath before and after the 1968 installation. The fuel processing sampler was cannibalized to make this repair, and sampling was continued. Contamination during repair work, although adequately controlled, was more of a problem because of the higher activity of the salt and the higher frequency of sampling with various types of ladles and capsules. About a month before the end of run 19 -Serious tangling of the cable, which had caused long delays on previous occasions,’® was averted because the prox- imity switch?® indicated something wrong on an otherwise apparently normal insertion. When the mag- " netic pickup did not actuate a light to indicate the passage of the latch as it should have, the insertion was stopped after 3 ft 10 in. of cable had been unreeled. The cable was rewound with no apparent difficulty until the position indicator on the reel indicated full a higher position than it had been in at the start. Somehow it had been lifted. Probably the capsule had lodged at the sampler tube entrance, a few inches below its starting point, and the stiff cable had been pushed on down the tube past it. Then as the cable was retrieved it must have snared the capsule and lifted it up to where it was found. Because the insertion was stopped before very much cable was paid out, no serious tangling had occurred. The capsule was hung again in its normal position, and sampling proceeded normally with no recurrence of this problem. During the latter part of run 20, the light bulb which illuminated area 3A failed. A temporary battery- powered light was improvised and inserted down the periscope channel, permitting sampling over the week- end until the light bulb was replaced on the following Monday. The double elastomer seals which contain the buffer gas at the various valves and the access port continued to show increased leakage, but it remained possible to provide a positive buffer zone. For example, the buffer pressure between the seals of the 1C access port decreased from ~50 psia in 1967 (ref. 21) to 37 psia at the end of run 19 and to ~32 psia at the end of run 20. (The seal leakage was into area 1C rather than into area 3A.) Although the radiation damage to the seal material _probably was becoming significant, the principal reason for the increased leakage from this buffer zone was no doubt the fact that the left-center Nu-vise clamp was loose and thus ineffective in compressing the seal during run 20. The increased leakage of buffer gas at the operational,2?:23 the maintenance,?* and (during run 19) the removal valve was all from the upper half of the . seal, This indicates that mechanical damage due to withdrawal. The isolation valves were then closed, and - the access port was opened. The cable had been fully rewound, but the capsule was not hanging straight down from the latch as would be normal. Instead it was lodged diagonally between the ledges of the access port, VSR Program Semiann. Frogr. Rept Aug. 31, 1969, ORNL4449, pp. 15-16. - 18pSR Program Semiann, Progr. Rept. Aug 31, 1968, ORNL-4344, p. 27, 19MSR Program Semiann. Progr. Rept. Aug. 31, 1967, ORNL4191, pp. 15, 32. 20)SR Program Semiann. Progr. Rept Feb, 29, 1968, ORNL-4254, p. 20, particulates falling on the valve was the probable cause of deterioration of the valve seals rather than rad:atlon damage. During this report period the coolant sampler was used to take eight coolant samples: four were salt samples for compositional analysis, and the other four were special exposures made in the study of tritium in both the gas and salt sections of the coolant pump. No operational difficulties were encountered. “MSR Program Semumn. Progr. Rept. Aug. 31, 1967, ORNL-4191, p. 32. 22y19R Program Semiann, Progr, Rept. Feb 28, 1966, ORNL-3936, p. 59. 23psR Program Semiann. Progr. Rept. Aug. 31, 1966, . ORNL4037, p. 72. 24MSR Program Semiann, Progr. Rept. Feb. 28, 1967, "ORNL-4119, p. 40. 1.3.2 Control Rods and Drives M. Richardson J. R.Engel All three control rod and drive assemblies performed - quite satisfactorily, with no difficulties, during this report period. Tests in late September, early October, and before the startup in November showed that scram times (times for release plus travel from full mthdrawal to lower limit) were all less than 0.81 sec. Some studies of the system used to extract control rod acceleration from records of position vs time during drop tests>® were performed to explore the sensitivity of the system to changes in acceleration. In these ‘studies an analog computer was used to generate a voltage signal in the same range as that produced by the position potentiometer on the actual rod drive. The time variation of this voltage was made to simulate rod 20 Such tests were performed at least monthly during the final periods of operation and at the start of each run. There was no further evidence of abnormal drag on any of the rods. 1.3.3 Off-GasSystems ~ _A.I Krakoviak Although chronic plugging recurred in both the fuel and coolant off-gas systems during this report period, the restrictions did not interrupt power operations. - The early part of run 19 was devoted to a study of ~ the effects of cover gas solubility on xenon stripping, position as a function of time for arbitrarily specified - values of acceleration as a function of rod position. The simulated signals were passed through the same data recording and processing system used for the actual signals so that a direct comparison could be made between real acceleration values and those produced by the processing program from position data. Com- parative results were good for long regions (more than 6 in.) of low acceleration (as was the case for the stuck control rod®® in June 1969). However, if a region of zero acceleration as short as 2 in. was present at a location more than 12 in. below the starting position of the rod, the effect on the final curve of acceleration vs position was nearly indistinguishable from other ran- dom variations, The failure of the program to reproduce such small aberrations was attributed to the filtering and data smoothing that were required to reduce random noise effects. Because of the difficulty in identifying short regions of low acceleration in a single rod drop from the fully withdrawn position, the rod testing procedure was modified to improve its fault detection capability. Accelerations were evaluated for rod drops from just which required several switchovers between helium and argon and finally back to helium as the supply cover gas. This -caused perturbations in system pressure control. The inlet gas flow is normally held constant, and the fuel system pressure is controlled by manually throttling the valve at the outlet of the charcoal beds (V-557B). The transit times (assuming slug flow) from the inlet meter to the charcoal bed inlets and to the outlet throttle valve are approximately 2 and 6 hr respectively. The varying gas flow rates (due to the different physical properties of the gases) at these three primary restrictions presented somewhat of a problem - of pressure control and pressure drop interpretation above the normal operating positions as well as from fully withdrawn. This would permit detection of any tight spots within the first few inches of travel during a drop. Since a very large restriction would be required to stop a rod that has accelerated normally for 6 in. or more, these measurements, coupled with the total drop times, gave adequate assurance of rod drop capability. ' 25MSR Program Semiann. Progr. Rept, Aug 31 1969, ~ ORNL-4449, pp. 16—-17. during the transition from one gas to another. No other operational difficulties were assoclated with the use of argon as a Cover gas. After approximately ten days of run 19 operatlons at the 5.5-Mw level, the pressure drop across the main charcoal bed increased from 3.0 psi with sections 1A and 1B in service to 4.6 psi with all three sections in service. The restrictions were partially cleared by lowering the water level in the charcoal bed pit and consecutively energizing the heaters-at the entrance region of each bed for periods of 8 hr each. The pressure drop across two beds in parallel was thereby reduced to 2.7 psi; however, plugging at the beds gradually increased over the next 12 days and reached a pressure drop of 4.9 psi. The restriction then remained constant for the next 20 days, after which it was necessary to valve in section 2B also. Five days later it was necessary to lower the water level in the charcoal bed pit and reheat the inlet section of each bed. At the end of each 8-r heating cycle, the heated bed was forward-blown with helium at 25 psig. This procedure - cleared the beds so that the pressure drop across beds 1A and 1B in parallel was lowered to 2 psi. No further problems were encountered with the main charcoal beds i in the subsequent Tun, Particle trap 1 had developed a restriction of 0.7 psi in May 1969, and V-522C was opened at that time to ~ put particle trap 2 in parallel service.?® Temperature measurements within the two traps then and during runs 19 and 20 indicated that particle trap 1 is still restricted and particle trap 2 carries essentially all of the off-gas flow. Approximately two months after the start of run 19, pressure drop measurements on the off-gas line seemed to indicate a restriction of 0.3 psi across - particle trap 2 also; however, cycling V-522C between its closed and open position cleared the restriction, thus indicating that the restriction' was in the valve rather than in the trap. Three days later the restriction reappeared (0.3 psi) and was cleared by the same method. No further problems were encountered with the traps or valve. Two additional attempts to clear the partial re- striction in the gas line entering fuel drain tank 1 (ref. 27) were made. Helium at 30 psig and then at 50 psig was directed from line 561 through HCV-573 and into FD-1. Although only marginal improvement was ob- tained, the gas flow through this line is adequate for any reactor drain situation. "Two weeks after the fuel fill of run 20, the restnctlon 21 directly below the hole that had been drilled in the top blocks for gamma spectrometry of the off-gas line. A tube leading from the discharge side of the filter would penetrate the top rim of the reactor cell and pass into the fuel sampler shield. A small tank that could be evacuated by the sampler vacuum pump would be used to pull measurable increments of gas through the filter pack, thus permitting determination of concentrations “of fission products in the off-gas at the pump exit. A in the fuel off-gas line at its exit from the pump bowl became detectable again for the first time since it was cleared in July 1969. In the next three days before reactor shutdown, the restriction increased from 1.3 to 2.2 psi as measured during the normal salt recovery operation from the overflow tank. The restriction was detected earlier by pressure noise data (Sect. 1.2.4); however, because of the proximity of the reactor shutdown date and the desirability of not interrupting the experiment in progress, it was decided not to use the heater which had been previously installed and sucoessful]y used to clear a snmlar restriction pre- v10usly Interest in better measurements of the fission products and tritium leaving the pump bowl led to the design and fabrication of two off-gas sampling devices, ‘which were to have been installed near the pump bowl after run 19. One device was a sampler to be used ‘during operation. A side outlet in the flange nearest the pump bowl would permit gas to be drawn through a connection into the tank was also to be provided for withdrawing batches of gas for tritium analysis, The other device, which was to be installed and recovered after operation, was a tubular specimen array to go in the off-gas line at the entrance of the holdup volume. Its purpose was to characterize the state of the fission products by measuring their distribution on the surfaces. The decision to bring MSRE operation to an early conclusion canceled the installation of these off-gas sampling devices. The sampling system?® in the vent house, which takes gas from the fuel off-gas line about 45 min downstream from the fuel pump, continued in service. In addition to trapping two samples of reactor gas onto the molecular sieve ‘during run 19, the offgas sampler was used to make ten checks of the amount of hydrocarbons in the off-gas stream. One of the samples trapped onto the molecular sieve was taken while argon was used as the - cover gas. This sample, however, was not very informa- tive, since argon as well as krypton and xenon were trapped on the sieve at —320°F. However, an attempt * was made to boil off the argon and thus concentrate ~-and retain the xenon as the sieve was allowed to heat up tO 32°Fo . . The recurring restrictions at the inlet valves to the off-gas sampler gradually increased during the hydro- carbon’ determination runs. Back-blowing the sampler valves with helium at 60 psig only partially restored the flow. After two additional hydrocarbon determinations, the restriction returned. No samples were taken with very efficient filter pack which would be located 26MS)‘(’ Program Semiann, Progr. Rept. Aug 31 1969, ORNL-4449, pp. 17-19, _ 2IMSR Program Semiann, Progr. Rept. Aug. 31, 1969, - ORNL-4449, pp, 17-19, rthls apparatus during the last run, The coolant off-gas system functioned satlsfactonly during this report period with the exception of the recurrent restriction at the sintered metal filter. The filter showed evidence of plugging on September 20, approximately six weeks after the coolant system was filled with salt and about two months after the filter. was replaced during the June—July shutdown. During 28)ISR Program Semiann, Progr Rept. Aug. 31, 1968, ORNL-4344,p. 30. the next month the restriction seemed to vary, as the before it eventually exceeded 10 psig. The signal from the coolant system pressure controller was then de- tached from PCV-528 and attached to HCV-536. Adequate pressure control was thus obtained (bypassing the filter) until the end of run 19, at which time the filter was replaced. No further coolant system off-gas problem was encountered during the subsequent run. Cursory visual examination of the filter used in the 22 ~gas pressure in the coolant pump bowl slowly mean- ~ dered back and forth several times between 5 and 9 psig 1.3.5 Containment and Ventilation | R. H. Guymon . The air inleakage into the reactor cell (at —2 psig) ranged from 15 to 30 scffday until the November. penultimate run showed no contaminant other than oil. No flow test was made on the filter after removal. The pressure drop of 5 to 9 psi that had developed across the filter while it was still in the system agrees rather well with the pressure theoretically required to enlarge or burst a film of oil from a 1-u-diam pore if one assumes that the bubble radius of a liquid film produced by a pore is the same size as the pore. The equilibrium equation for a spherical bubble is AP = - 20/r, where ¢ = surface tension = 16 dynes/cm for the oil in the coolant pump and r = bubble radius = pore radius = 0.5 X 10™* cm. This calculation indicates that any pore in the sintered metal filter smaller than 1 p, once plugged with condensed oil vapors, would remain plugged at pressure differentials of less than 9.3 psi. The slow pressure oscillation experienced could be at- tributed to a near i £ \L-.___.____. PLUTONIUM —240 - o -0 URANIUM — 236 - ul i . \ -0.2. \\ : ~0.3 NONSATURAT;N}\ . " FISSION PRODUCTS \ -0.4 : S » TN In this expression the parameter K is also the 23U concentration coefficient of reactivity at the minimum critical loading, that is, : e In our earlier analysis we had assigned a value of 0.223 for K, which represented the average coefficient for 235U variations made in the zero-power calibration experiments.® More recent analysis,'® however, indi- cates that a better fit to the data over the entire range of 235U concentration variations is obtained by use of Eq. (1) with K equal to 0.239. This correction is independent of the additional correction for delayed neutron effectiveness described above. The principal effects of the modifications in category 1, other than the 235U data changes, which are _separately treated, appear in a single term of the reactivity balance. This term includes effects of isotopic changes in the core which depend mainly on the time-integrated power, namely, the buildup of non- . ORNL—DWG 70—-6734 O 10 20 30 40 50 60 70 80 TIME—INTEGRATED POWER (10® Mwhr) Fig. 2.1. Reactivity Changes Due to Long-Term Isotopic Changes in MSRE During 235U Operation (Revised Calcula- tions). ‘ saturating or slowly saturating fission products and the changes in Li, 234U, 236U, 238y, 23%py, 24%Py, and 108 content (the last in the graphite). In our reference calculations the same assumptions were made con- cerning the degree of fission product removal from the system as were used in earlier studies,'! that is, that all of the noble gases and none of the noble metals were removed. The results of revised calculations of these component reactivity effects, together with their algebraic sum, which appears in the reactivity balance, are shown in Fig. 2.1. The most important change resulting from these revisions is in the 2*®Pu reactivity component, which increased by about 50% from earlier calculations. The magnitude of the negative component associated with buildup of nonsaturating fission products also increased by about 10%, due to the revision in the calculated epithermal-to-thermal flux ratxo in the reactor. The results of including all changes described under categories 1, 2, and 3 in the interpretation of the reactivity balance data are shown in Fig. 2.22. In this figure we have attempted to distinguish the data points based on measurements taken during separate power runs. Transfer and mixing effects during drain and flush operations between runs introduced variations in the composition of the salt which generally had larger uncertainties associated with them than the changes produced during the runs. The data shown in Fig. 2.2a indicate a small overall positive trend in the residual reactivity during the entire period of operation with 225U, (The residual reactivity is essentially equal to the “observed” reactivity minus the “predicted” reactivity.) One finds, however, that the average upward trend in reéctivity during this period is about 0.07% 8k/k, which is less than half the magnitude indicated in our earlier studies. The statis- tical spread in the data points logged over the -entire period prevents one from assigning a precise shape to the residual reactivity variation; however, the data in Fig. 2.22 taken during the longest uninterrupted power run (run 14) appear to be rising more steeply than the average trend during 235U operation. Thus the slope of the residual reactivity variation was probably increasing dunng this period. , It is useful to note that part of any long-term upward ' trend in reactivity might be explained if the actual amount of fission product removal was greater than assumed in the reference calculations (Fig. 2.1). For example, if complete removal of the noble metals Mo, Ru, and Te had occurred, the residual reactivity would have been reduced by an amount varying approximately RESIDUAL REACTIVITY (% 84/%) 32 ORNL-DWG 7O~ 6735 0.42 0.10 0.08 am n 0.06 »O 0.04 0.02 -0.04 -0.06 1 o & START RUN 9 (D) START RUN 10 (A) START RUN 8 (4) . START RUN 7 (&) {a) START RUN 13 (0) START RUN i4 (@) START RUN 12 (@) e e e e —— START RUN it (V) -0.10 0.06 0.04 0.02 kim | T ® -0.06 I -0.08 |- | | START RUN 11 (V) START RUN 7 (&) START RUN 8(#) T | sTaRT RUN 9(Q) 'START RUN 12 (m) START RUN 13 (0) START RUN 14 (@) o e v e e s g e e e i e e {5 -0.10 0 .{ START RUN 10(a) 20 - 30 o] TIME- INTEGRATED POWER (10° Mwhr) H © 50 60 70 Fig. 2.2. Long-Term Variations in Zeh—Power Residual Reactivity During 235y Operations. Maximum reactor power: (a) 8.0 Mw(th), (b) 7.25 Mw(th). linearly between zero at the start of operation and 0.04% Sk/[k at the end of 70,000 Mwhr of operation. Evidence exists that at least partial removal of these products did occur.!® : o 1SMSR Prograin Semiann. Progr. Rept. Aug. 31, 1967, ORNL4191, pp. 116-19. : Recently, some precise mass spectrometer measure- ments, made to determine the capture-to-absorption ratio for 235U in the MSRE spectrum,® have indicated - that the actual fission rate might be lower by about 10% than that indicated by the heat balance calibration. To determine what effect this difference might have on 33 the interpretation of the long-term reactivity trends, we recalculated the individual terms in the reactivity ‘balance assuming that the maximum reactor power was 7.25 Mw. (Those terms which are directly affected by the fission rate are the 2?°U depletion, the samarium poisoning, and the long-term isotopic changes described in Fig. 2.1.) These results are shown in Fig. 2.2b. The variations in residual reactivity remain quite small in magnitude and indicate a slight negative trend during the first part of the operation, again followed by a more gentle positive trend during run 14, Increased fission product removal could accentuate any downward trend by the amount indicated above. | It is quite likely that either mterpretatxon of the reactivity balance data described in Fig. 2.20 or b is a possible representation of uncertainties in the model or of other changes affecting nuclear operation with 235U which are not explicitly included in the calculations. The magnitudes of reactivity variations exhibited in - either case lie within the region in which the calculation model is judged to be valid, Hence the reactivity balance data alone cannot be used to support a spe01f1c choice for the fission rate. In this connection it is interesting to observe that - recent calculations by C. H. Gabbard'® relating to the effect of dimensional changes in the graphite during 2357 operation exhibit variations in reactivity similar to those of both Figs. 2.2z and b. These calculations were made in an attempt to bracket the expected reactivity variation due to deflections and density changes in graphite stringers caused by fast-neutron irradiation. In one limiting case these calculations indicated that a gradual increase in reactivity could occur, resulting in a net addition of about 0.06% 8k/k during 70,000 Mwhr of operation. An opposite limit, however, indicated that an initial decrease in reactivity "could occur, reaching a minimum of —0.06% 8k/k at about 50,000 Mwhr and then increasing again during . further exposure of the graphite. Precise mathematical modeling of the effects of the nonuniform dimensional changes on the core geometry is quite difficult, and the magnitudes obtained in these calculations are approxi- mate. However, the results of Gabbard’s studies lend further credence to the reactmty behavior exhibited in Fig. 2.2 . One other quite recent development should be mentioned ‘which could, in addition to the fission product removal discussed above, slightly modify the numerical values of residual reactivity shown in Fig. 16MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 5-6. 2.2. This concerns the assessment of the ®Li content in the fuel salt. Reexamination of all available evidence (Sect. 1.2.5 and ref. 17) now indicates that the most likely average isotopic assay of Li which went into the MSRE fuel salt preparation was between 0.005% and 0.006%, instead of 0.0074% as had been used in all previous reactor analysis calculations. In accordance with the calculations shown in Fig. 2.1, this would have “the effect of reducing the magnitude of the component associated with SLi burnout by the ratio of the initial ®Li concentrations. The net result of this change would be to add an amount of reactivity equivalent to this difference to the residual reactivity points shown in Fig. 2.2. Thus each point in Figs. 2.2¢ and b could be displaced upward by an amount varying approximately linearly between zero at the start of operation and 0.035 + 0.009% 8k/k at 70,000 Mwhr to correspond to this correction. However, the preceding conclusions regarding the validity of the reactivity balarice model and results would not be changed by this modification. Those sources of statistical variations and uncertain- ties in the reactivity balance data which can be identified include: (1) a component of about #0.01% 8kfk associated with random errors in reading rod positions and temperatures and with short-term varia- “tions in the amount of entrained gas circulating with the salt and (2) a component associated with uncertain- ties in fuel transfer during flush operations, increasing in magnitude from zero near the start of power operation to about +0.015% 8k/k near the end of operation. Any remaining statistical variations should at - present be regarded as inherent in the technique of recording and analyzing- the reactivity balance data rather than as evidence of some anomalous physical process in the reactor. : The data summarized in Fig. 2.2 represent measure- ments accumulated over nearly three years of operation of the MSRE. During that time 235U equivalent to ~125% in reactivity had been depleted in power operation; 235U equivalent to 0.72% 8k/k had been added to the salt; poisoning due to '*?Sm and !5!Sm ~ equivalent to 0.77% 8k/k had been formed; changes in isotopic content of other constituents of the salt had produced a net reactivity addition of 0.42% Sk/k. Our procedures for accounting for these effects indicate that the reactor behaved in a regular and predictable manner and that the reactivity balance is a valuable tool in performing nuclear operatlons analyms for molten-sa]t reactors 17p, N. Haubenreich, Tritium in the MSRE: Calculated Production Rates and Observed Amounts, -internal memo- randum (Feb. 4, 1970). 3 Component Development | Dcn]ap Scott 3.1 FREEZE-FLANGE THERMAL-CYCLE TEST - F.E.Lynch Thermal cyclmg of the freeze flange was continued through cycle 470 before it was shut down for a scheduled inspection and minor repairs. Operation of - the test was resumed at cycle 471 and continued through cycle 540, when thermal cycling of the flange was discontinued as scheduted. Visual inspection of the ‘flange exterior during and at the end of each cycle revealed no indication of thermal fatigue cracks. There were, however, the usual minor operation problems experienced during this period of operation. These operation problems and the results of the flange " inspection after cycle 470 are reported below. 3.1.1 Facility Operation Problems Downtime during both these periods of operation (cycles 401-470 and 471-540) was due to burned-out test section heaters. All minor operation problems, such as plugged vent lines and instrument troubles, were corrected during or at the end of the oscillation time. Burned-out heaters in both test section heater boxes resulted in all the downtime during the first period (cycles 401—470). Replacement of the test section heaters adjacent to the female flange resulted in the downtime during the last period of operation. 3.1.2 InSpectlon of the Flanges The flange clamps were: removed at the end of cycle 470 without any difficulty. The flanges were opened for inspection, and the oval ring gasket with stainless steel insert screen was removed for inspection. The average outer diameter of the frozen salt cake on the screen was 10 in. This was _approximately the same outer diameter as in all previous inspections except for cycle 103, when it was 11% in.- Inspection of the inner flange face and the bore of both flanges was again made with a fluorescent dye penetrant (Zyglo type ZL-22 penetrant with developer type ZP-9). As in previous inspections, the female flange face and bore were free of cracks. In the male flange the face and neck remained free of cracks or porosity indications, but as before there were indica- tions in the bore in the vicinity of the weld attaching the alignment stub to the face of the flange."! Figure 3.1 is a photograph showing the fluorescent indication of a crack at the base of the stub and a porosity band extending clockwise around the bore from the upper thermocouple. A dashed cracklike indication at a - position 1% in. from the end of the stub started at the upper bore thermocouple and extended circumferen- tially to approximately 15°. From this point around to 60° the cracklike pattern indication was continuous, with a dashed cracklike pattern continuing around to 120°, A porosity band extended 1 in. farther into the bore and continued around to 120° before necking down to %, in. width that continued around to 155°. Figure 3.2 shows a similar porosity band extended counterclockwise from the upper bore thermocouple around to 155°. From 15° counterclockwise around to 45° there was very little porosity indication. Comparison of these photographs with those taken of the same locations after cycle 400 (Figs. 3.1 and 3.2 of ref, 2) showed that there was less apparent porosity in the later inspection. This apparent anomaly was attrib- uted to differences in the cleaning procedure used to remove salt from the bore before it was inspected with the dye penetrant. (All trace of salt must be removed from the bore or it will absorb the dye and indicate porosity-type cracks.) The cleaning was more vigorous after cycle 470, when aluminum oxide cloth bands (180 'MSR Program Semiann. Progr. Rept, Aug. 31, 1968, ORNL-4344, pp. 33— 35. 2MSR Program Semiann. Progr. Rept. Aug 31, 1969 0RNL-4449 pp. 27-29. 35 PHOTO 77203 Flg 3.1. Photograph of Test Freeze Flange After 470 Cycles, Showing Fluorescent Dye-Penctrant Ihdication of a Crack on the Right Side of Bore. ' grit) were used on a small pneumatic rotary grinder to clean the bore. The ability to remove or obliterate the cracklike indication and the porosity indications could be evidence that these indications were not as deep as they had first appeared. Thus plans were made to make an effort in the final inspection to clean the surface by removing a very small amount of metal. The results will be photographed, and differences in the cracklike pattern noted. An effort will then be made to remove a portion of the cracklike pattern by grinding approxi- mately 0,002 in. from the bore of the male flange in an area where the depth of the cracklike pattern is greatest. 3.2 PUMPS - P.G.Smith _A.G.Grindell '3;2.1_ Mark 2 Fuel Pump The mark 2 fuel salt pump® was continued in opera- tion circulating - the molten :salt LiF-BeF;-ZrF,- ThF,-UF, (68.4-24.6-5.0-1.1-0.9 mole %). It has now operated for 12,744 hr at flows to 1350 gpm and temperatures between 1020 and 1325°F, The salt level 3mSR Program ‘Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 29. UPPER BORE THERMOCOUPLE PHOTO 77204 Fig. 3.2. Photogmph of Test Freeze Flange After 470 Cycles, Showmg Fluorescent Dye-Penetrant Indication of a Crack on the Left Side of Bore. was maintained at 5% in. above the normal level, and the pump was operated at constant conditions to investigate the endurance capabilities of the pump. During the first four months of this report period, partial plugging was experienced about twice weekly in the pump tank off-gas line at the filter or in the piping just upstream of the filter. Upon removal the filter was found to be nearly full of salt aerosols (15 p diameter or less) and was replaced with a new filter. During the remainder of the report period, the new filter did not plug although salt aerosols again began to build up in the filter. The shaft annulus (gas inlet to the pump tank) plugged slightly eight times during the report period. The plug was easily removed by stopping and sta'rting the pump three or four times or by allowing the salt temperature to increase to 1325°F for a penod of about 1 hr. , - Collection of oil leakage from the lower shaft seal has averaged less than 10 cc/day 322 011 Mp Endurance Test The oil pump endurance test* was contmued One week of operation was lost when the pump was stopped *MSR Program Semiann. Progr Rept. Aug. 31, I 969 ORNL-4449, p. 31. inadvertently by a craftsman. Operation was resumed, and the pump has now operated for 57,344 hr, circulating oil at 160°F and 60 gpm. - 3.3 DEVELOPMENT OF ANALYTICAL MODEL FOR 135 Xe POISONING IN THE MSRE “R.J.Kedl In 1967 an analytical model was developed to calculate !'35Xe reactivity effects in the MSRE. This model was based on conventional mass transfer phe- nomena, including effects of circulating bubbles of cover gas, which it treated as practically insoluble. This model predicted that the '3%Xe poisoning would decrease monotonically as the void fraction of bubbles increased. The data on xenon poisoning as a function of void fraction obtained in this report period (Sect. 1.2.2) showed the predicted kind of behavior when the cover gas was argon but not when it was helium. With helium cover gas, the Xxenon poisoning passed through a maximum at a core void fraction of about 0.25%. ‘The reason for the maximum in the experimental data with helium cover gas is thought to be associated with helium bubbles going into solution at the pump discharge and renucleating further around the fuel loop as the fluid static pressure is reduced. Apparently this renucleation is accompanied by gross removal of '35Xe from solution. Argon is an order of magnitude less soluble in fuel salt than helium; therefore one would expect the bubble dissolution and renucleation effects to be reduced and the argon data to be in better. agreement with the analytical model. To check this idea, 135Xe steady-state reactivities were recomputed with this model and the best available estimates of bubble parameters for comparison with the helium and argon data to see if any conclusions could be reached. ~ The bubble parameters which were used are shown in Table 3.1. The reasoning behind the choice is as 37 follows. It is generally thought that the circulating bubbles ‘in ‘the MSRE are very small, with estimates ranging from 0.001 to 0.010 in. The void fraction in the reactor was obtained by changing the pump speed. As the flow rate goes up, yielding a higher void fraction, one would expect that larger bubbles are carried under into the fuel loop. Table 3.1 then shows the bubble - diameter changing from 0 to 0.010 in. incrementally as the loop void fraction goes from O to 1%. An average bubble diameter is chosen for each increment of void fraction. In a gravity field of unity, the terminal rising velocity of a 0.001-in. bubble in the fuel salt is about 0.004 in./sec, and it is 0.4 in./sec for a 0.010-in. bubble. Obviously the 0.001-in. bubbles will travel “with the - salt”; that is, they will remain with the fuel as it is sprayed through the pump bowl and eventually reenters the loop. The bubble-stripping efficiency for these bubbles ought to be close to zero. The bubble-stripping efficiency will increase rapidly as the bubble size increases because in this range (Stokes range) the bubble rising velocity increases as the square of the bubble diameter. Table 3.1 also shows the bubble- stripping efficiency changing from 0 to 20% incre- mentally as the loop void fraction goes from 0 to 1%. An average bubble-stripping efficiency is chosen for each increment of void fraction. To illustrate, when the void fraction is in the range of 0 to 0.1%, the bubble diameter is 0.0005 in. and the bubble-stripping effi- ciency is 1%. When the void fraction is in the range of 0.1 to 0.2%, the first 0.1% is the previous bubbles, and that above 0.1% is composed of new bubbles of diameter 0.0015 in. and stripping efficiency 3%. The progression continues to the last range of 0.9 to 1.0% void fraction, which is composed of the previous nine ranges plus new bubbles of diameter 0.0095 in. and stripping efficiency 19%. Similar progressions with different sized efficiency increments were used to establish bubble parameters for other maximum values of stripping efficiency. ' Table 3.1. Bubble Parameters Used in ' *Xe Calculations Parameter Value Void percent 0 0.1 02 ... 0.9 1.0 " Bubble diameter (in.) 0 0.001 0.002 ... 0.009 0.010 ~ Average bubble diameter in 0.0005 0.0015 ... 0.0095 range indicated (in.) / » - ‘ . Bubble stripping efficiency (%) 0 2 4 18 20 Average bubble stripping efficiency 1 in range indicated (%) ORNL-DWG T0-6736 .2 T — Y o FUEL: 2%y 1.0 N POWER: 5.5 Mw OBSERVED VALUES N \ 5 \ \_ @ HELIUM COVER GAS \ o ARGON COVER GAS & 7 XENON-135 POISONING (% 34/4) O an \"\\\ ' : BUBBLE STRIPPING \u\ EFFICIENCY RANGE 4 - 0-20% o2t CALCULATED B o _ I | | 0-40% 0 0.2 0.4 0.6 0.8 1.0 CORE VOID FRACTION BUBBLES IN FUEL SALT (%) Fig. -3.3. Companson of Calculated and Observed 135xe Poisoning. The results of the calculation with these bubble parameters at maximum stripping efficiencies of 20 and ~ 40% are the curves in Fig. 3.3, A simplified calcula- tional method was used in which all core parameters were volume averaged into one lump. The reactivity was computed by first computing the poison fraction, multiplying it by a constant (0.656) to convert it to reactivity units, and then by a spatial correction factor (0.81) on the graphite contribution because its !35Xe concentration profile is dish-shaped due to burnup. 38 First consider the right side of the plot. The curves calculated with bubble-stripping efficiency ranges of O . to 20% and O to 40% straddle the data points, At higher void fractions one would expect bubble dissolution and renucleation effects to diminish so the original ana- lytical model (insoluble cover gas) should converge with - any new analysis where the solubility of the cover gas is considered. The good agreement between the analytical model and the data with “reasonable™ values of bubble parameters would seem to bear this out. Now consider the left side of the plot. The computed intercept with the ordinate could be brought down to . the argon data point (void fraction = 0%) by using a mass transfer coefficient to the graphite of half the originally estimated value. There is some indirect evidence from work with noble metal migration to ~ graphite that the mass transfer coefficient should be Iower. If we assume, however, that the mass transfer coefficient to the graphite is correct, then another observation can be made. This is that the points of intersection with the axis increase as the cover gas solubility decreases: they are about 0.25% &k/k for helium (soluble), about 0.85% 8k/k for argon (much less soluble), and about 1.12% 8k/k for the analytical model - (insoluble). This observation would lend support to the thesis that cover gas solubility is an important param- eter at low void fractions in the fuel loop. 4. Instruments and Controls 8. J. Ditto 4.1 MSRE OPERATING EXPERIENCE J.L. Redford | Daily testing of the 15 relays in the rodscram coincidence matrix was continued to the end of nuclear operation. There was no failure among these relays after the summer of 1968 (ref. 1). | There was no unscheduled control-rod scram during this report period. Table 4.1 lists the causes (ass.igned’ Y MSR Program Semiann. Progr. Rept. Aug. 31, 1 968 ORNL-' by the reactor operators) for the 37 unscheduled control rod scrams in four years of MSRE operation. None was caused by a process variable actually going out of limits. Nearly half of all the unscheduled scrams ‘and 70% of those attributed to instrumentation and control problems occurred in the first six months of the four-year period. - | During run 19 the load-scram circuit, which drops the radiator doors and stops the blower, was tripped twice while the reactor was at full power. Investigation showed that several relay contacts had developed unusually high resistance due to oxide films. Because of 4344, p. 43, the way the contacts were paralleled in the matrix, the . Table4.1. Summary of Unscheduled Scrams at MSRE with Fuel in the Core? . o " Operating Hours " Number of Unscheduled Rod Scrams Year Quarter - Fuelin R _ _ Human Power Core Critical Total Error Failures 1&C Other” 1966 1 672 62 4 2 0 1 1 ‘ 2 1293 1070 13 2 3 6 2 '3 554 413 2 0 2 0 0 4 1266 1221 3 1 1 1 0 1967 = - 1 1861 1852 2 1 0 1 0 - 2 1254 118 . - 2 1 1 0 0 3 . 1318 1292 . .1 0 1 0 0 4 2159 2144 2 0 1 1 0 1968 1 2048 2045 0 0 0 0 0 2 0 0 0 0 0 0 0 3 88 0 0 0 0 0 0 4 1000 735 1 1 0 0 0 1969 1 1850- 1800 2 0 o 0 2 | 2 ‘1385 1375 3 0 1 0 2 3 11076 1054 2 0 0 0 2 4 1203 1176 0 0 0 0 0 Total 19027 17425 37 8 10 10 9 %There is no record of any unscheduled scrams dunng 1965 when fuel was in the core for 1062 hr and the reactor was cnhcal (at 1 kw or less) for 230 hr. Mostly equipment faults, For example, five of the last six scrams due to “other” causes occurred when the speed of the variable-frequency generator being used temporarily to drive the fuel pump sagged below a prescribed limit. 39 e B e e T ey W s e o ets e o1 s " film was not bumned off each time the contact close(i, as in a normal application. 4 2 CONTROL SYSTEM DESIGN P.G. Hemdon | The fuel off-gas system was revised to mcorporate an " apparatus set up in the vent house to collect tritium from the gaseous effluent streams. Sample lines were connected from the apparatus to.the fuel and coolant salt off-gas lines, to the containment evacuation line, and to a sleeve on a tube in the coolant radiator (Sect. 1.2.5). Purge gas was supplied to this apparatus from high-pressure cylinders equipped with series-connected pressure regulators, rupture disks, and high-pressure alarms to assure that the reactor system would not be subjected to pressures in excess of 50 psig. One new safety block valve was connected to existing circuits, and all other lines were connected so as to take advantage of existing block valves without com- promising the integrity of the primary and secondary containment systems. Design of safety circuits for another system, a new off-gas sample line from line 522 near the pump bowl to the fuel sampler-enricher container, was completed, but the line was not installed before the final shutdown. Three small weldsealed valves were fitted with pneu- matic actuators for use in this system. The actuators were to be supplied by three-way solenoid valves which were added to the vacuum system circuits. Plans for setting up the instruments and controls to - adequately maintain and monitor reactor conditions during the shutdown period were also completed. 4.3 MSRE ON-LINE COMPUTER C. D, Martin, Jr. Operation and analysis of the MSRE was facilitated -by the use of a Bunker-Ramo 340 digital computer - connected direcily to sensors in the MSRE system. The ~ computer main frame included 12,288 28-bit words of core memory and 32,768 words of drum memory with a 7.5usec core cycle time. Peripheral equipment in- - cluded 350 analog inputs, 128 contact inputs, 36 analog outputs, 32 contact outputs, 2 magnetic tape units, an input keyboard, a paper tape punch, a paper tape reader, 4 output typewriters, and an x-y plotter. The system was installed in the summer of 1965 and was accepted on October 1, 1965 (before the beginning of MSRE power operation), Over the next 51 months, untii January 1, 1970, the computer system was available 95.35% of the time. During this report period the neutron noise analysis program was extensively modified to make it run as an ondine operator request. The previous version of the program required taking the computer offdine to acquire the data.and record it on magnetic tape. The data were then read back into the computer for processing ondine as a background operation with the reactor monitoring system running in the foreground. The new program eliminated the recording of data on magnetic tape. When this program is requested, the reactor monitoring system is preempted for a period of 10 min, during which time data are acquired and stored in the computer memory. Calculations are made as sufficient data are accumulated. The use of two data buffers permits the functions of data acquisition and processing to proceed concurrently. The reactor moni- toring system is automatically reinstated upon com- pletion of the noise analysis calculation, and the results of the calculation are plotted ondine as a background .. function producing the familiar graph of power spectral - density. The void fraction is calculated for the fuel salt using the results of the noise analysis calculation. This - value is printed on the control room typewriter. - ‘Following the reactor shutdown in December 1969, the ondine reactor monitoring system was revised. The new system exercises all units of the computer periodi- “cally except for the magnetic tape drives and three of the four typewriters. FOCAL is still available for engineering calculations using the input keyboard and console typewriter, and analog signals can be displayed on the digital display of the computer console. The company-owned magnetic drum memory unit was refurbished at the manufacturer’s plant and was reinstalled in the computer in January. While the drum was being reworked, system operation was maintained using a drum leased from the computer manufacturer. ¥ Part 2. MSBR Design and Develp'_pment - .S The purpose of the MSBR design and development activities is to prepare a reference design for a 1000-Mw(e) one-fluid MSBR plant; to design a molten- salt breeder experiment (MSBE), operation of which will provide the data and experience necessary to build large MSBR’s; and to develop the components and systems for the MSBE. ' Work on the reference design for the one-fluid MSBR ~ was begun in October 1967 and has taken most of the effort. Prior progress is reported in our semiannual reports for the periods ending in February and August 1968 and 1969. The studies have converged on a design, . and some of the more important details have been investigated. Writing is in progress on a topical report that will describe the plant and the results of the studies in considerable detail. Some studies are being made of ‘molten-salt power reactors that would have lower performance but would require less development and could be built sooner than the reference MSBR. _ With the general design of the reference MSBR reasonably well established, we have begun to look at the MSBE, Calculations indicate that a reactor with a power level of 100 to 200 Mw(th) can satisfy the requirements that have been proposed to date. A small effort is being spent on preliminary studies of reactor designs and on nuclear calculations. The development program is small, and the experi- mental work is limited to some of the most important problems. Work is being done on methods of dispersing bubbles of gas in and separating them from circulating liquids. Experiments are in progress to measure the coefficients for transfer of dissolved gas to bubbles in circulating liquids. These experiments are in support of the gaseous fission product removal system. Better values are being obtained for the thermal conductivities of salts for use in heat transfer calcula- tions. Experiments are in progress to confirm or improve on the relationships used to calculate heat transfer coefficients for molten fluoride salts. The sodium fluoroborate—sodium fluoride eutectic salt has, ‘because of its low melting point and low cost, been proposed for use in the intermediate coolant systems of large molten-salt reactors. Since this is a new salt to the MSR program, a forced convection loop is being operated in engineering tests with the salt. 5. Design E. S. Bettis 5.1 SINGLE-FLUID MSER DESIGN STUDY P. R. Kasten R.C. Robertson ~ E. S. Bettis W. K. Furlong The major effort of the MSBR design group was spent in writing a2 comprehensive report on the single-fluid 41 MSBR - reference design.' The material is essentially complete, and a first draft is being circulated for -comment. The updated characteristics of the plant are summarized in Table 5.1. | - 1p R. Kasten et al, ,' Single-Fluid Molten-Salt Breeder Reactor Design Study, ORNL-4541 (to be published). 42 Table 5.1, Principal Design Data for 10000 Mw(Electrical) MSBR Power Station General . Thermal capacity of reactor, Mw(th) . ) . 2250 Gross electrical generation, Mw(e) 1035 Net electrical output of plant, Mw(e) - - o 1000 Net overall thermal efficiency, % T , L . 444 Structures _ ' Reactor cell, ft : o "~ 72 diam X 42 high Confinement building, ft ' _ . 134 diam X 189 high Reactor o _ - Reactor vessel inside dlameter, ft S : 22 © Vessel height at center, ft? . , : _ - 20 _ Vessel wall thickness, in. o : 2 " Vessel head thickness, in. T 3 - Vessel design pressure, psi ‘ ' 78 - Core height, ft = o 13 Number of core elements . 1412 Length of zone I portion of core elements ft o Overall length of core elements, approxnmate, ft o 15 Distance across flats, zone I, ft : ' 14 Outside diameter of undermoderated region, zone II, ft i5 Overall height, zone I plus zone I, ft: = . 18 Radial distance between reflector and core zone I, in. . 2 Radial thickness of reflector, in. ‘ _ .30 - Average axial thickness of reflector, in. ~22 Volume fraction salt in zoneI ' - 0.13 Volume fraction salt in zone 11 , 0.37 Average core power density, kw/liter _ : 22.2 Maximum thermal neutron flux, neutrons cm 2 sec™ 7.9 X 1014 Maximum graphite damage flux (>>50 kev), neutrons em ™2 sec”! 32 X 1014 Graphite temperature at maximum neutron flux region, °F 1284 Graphite temperature at maximum graphite damage region, °F o 1307 - Estimated useful life of graphite, years 4 Total weight of graphite in reactor,Ib S - 650,000 Weight of removable core assembly, Ib 480,000 Maximum flow velocity in core, fps ’ . 8.5 Pressure drop due to salt flow in core, psi _ 18 Total salt volume, primary system, e ' 1720 Fissile fuel inventory, reactor plant and fuel proccssmg plant kg 1470 Thorium inventory, kg : 68,000 Breeding ratio ' 1.06 Yield, %/year 33 . Doubling time, compounded continuously, years 21 Primary heat exchangers (for each of four units) : _ Thermal capacity, Mw(th) ' 556.30 Tube-side conditions: _ _ ) Fluid : : Fuel salt Tube size, OD, in. , % ' Approximate total length, ft , 22.5b ~Number of tubes | ' ~ 5896° Inlet-outlet temperatures, F ' 1300-1050 Mass flow rate, Ib/hr - ‘ | 23.45 X 10%? ~ Volume of fuel salt in tubes, £t E ‘ 67.20 ' Pressure drop due to flow, psl ‘ 130% Total heat transfer surface, ft o 13,0092 Shell-side conditions: _ : : - Fluid . Coolant salt Shell ID, ft ‘ | 5.67° Central tube diameter, ft , ' 1.7 43 Table 5.1 (continued) Baffle spacing, ft Baffle cut, % Inlet-outlet temperatures, °F Mass flow rate, Ib/hr Pressure drop due to flow, psi Approximate overall heat transfer coefficient, Btu hr ! ft72 °F~1 : Fuel-salt circulating pumps (each of four units) Pump capacity, gpm Rated head, ft Speed, rpm Specific speed " Net positive suction head, ft ~ Impeller input power, hp Design temperature, °F Coolant-salt circulating pumps (each of four units) Pump capacity, gpm Rated head, ft Speed, rpm Specific speed Net positive suction head, ft Impeller input power, hp Design temperature, °F Primary salt drain tank Outside diameter, ft Overall height, ft Outside wall thickness, in. Bottom head thickness, in. Storage capacity, e - Design pressure, psig Design temperature, °F Number of coolant U-tubes Size of tubes, OD, in. Number of separate coolant circuits Coolant fluid o Composition, mole % : Volume of coolant inventory, e Under normal steady-state conditions: Maximum heat load, Mw(th)¢ Coolant circulation rate, gpm Coolant temperatures, in/out, °F . Maximum tank wall temperature, °F Maximum transient heat load, Mw(th) Primary salt storage tank ' Storage capacity, ft3 Heat removal capacity, Mw(th) Material of construction. ' Steam generator-superheaters (for each of 16 units) Thermal capacity, Mw(th) Tube-side conditions: Fluid Tube size, OD, in. Approximate length, ft Number of tubes Inlet-outlet temperatures, °F Mass flow rate, 1b/hr . 0.94° 40 850—1150 17.6 X 1052 1152 8510 16,000 150 890 2625 18 2200 1300 20,000 - 300 1190 2330 30 3100 1300 14 22 1 1Y% 2500 40 1300 1500 3 4 40 TLiF-BeF, 67-33 420 18 830 900-1050 ~1260 2500P 1.4P Stainless stqelb ‘ 120.7 Steam at 3600-3800 psia | 76.4% 3930 700—1000 630,000 44 Table 5.1 (continued) Total heat transfer surface, ft2 Pressure drop due to flow, psi Shell-side conditions: Fluid Shell ID, ft Baffle spacing, ft Inlet-outlet temperatures, °F Mass flow rate, Ib/hr Pressure drop due to flow psi Approx:mate overall heat transfer coeffic:ent Btuhr ! ft72 °F - Steam reheaters (for each of eight units) Thermal capacity, Mw(th) Tube-side conditions: Fluid Tube size, OD, in. Approximate length, ft - Number of tubes : Inlet-outlet temperatures, °F - Mass flow rate, 1b/hr Pressure drop due to flow, ps1 Total heat transfer surface, ft? Shell-side conditions: Fluid Shell ID, in. Baffle spacing, in Inlet-outlet temperature, °F Mass flow rate, Ib/hr Pressure drop due to flow, psi Approximate overall heat transfer coefficient, Btuhr™ ft 2 °F ! - Turbine-generator plant Number of turbine-generator units Turbine throttle conditions, psia/°F Turbine throttle mass flow rate, [b/hr Reheat steam to intermediate-pressure turbine, ps1a/° F Reheat steam mass flow rate, Ib/hr Condensing pressure, in. Hg abs. - Number of stages, regenerative feedwater heating Feedwater temperature leaving last regenerative heater, °F Feedwater temperature entering stéam generator, °F Boiler feed pump work, each of two, hp Booster feed pump work, each of two, hp Gross electrical generation, Mw(e) Net electrical output of plant, Mw(e) Net plant heat rate, Btu/kwhr . Net overall plant thermal efficiency, % Fuel processing system -Total inventory, fuel salt in primary system, ft> Processing rate, gpm Cycle time for salt inventory, days Heat generation in salt to chemical plant, kw/ft Design properties of materials Fuel salt: Components Composition, mole % . 8 39290 1540 Coolant salt 1.5 4 1150-850 3.82 X 102 610 490-53059 36.6 / b 3osb 4000 6501000 641,000 30 23820 Coolant salt 21.2b 9.8b 1150-850 1.16 X 10%% 60 3030 1 3500/1000 7.15 X 10° 540/1000 5.13 X 10° 1.5 551 700 19,700 6,200 1035 - 1000 7687 444 1720 . 298 3 56 ~ 7LiF-BeF,-ThF,-UF, 71.7-16-12-03 Steam at 550 psxa 45 Table 5.1 (continued) Molecular weight, approxunate Liquidus temgerature, Density, Ib/ft” at 1175°F Viscosity, Ib ft ™! hr™* at 1175°F Thermal conductivity, Btu hr ! ft™! °F ! at 1175°F Heat capacity, BtuIb~! °F ! Vapor pressure, mm Hg at 1150°F Coolant salt: Components Composition, mole % Molecular weight , Liquidus temgerature, Density, Ib/ft at 1000°F Viscosity, Ib ft~! hr™? at 1000 F Thermal conductmty, Btu hr 1 g1 oF -1 at 1000°F Heat capacity, Btu b °F! Vapor pressure, mm Hg at 1150°F Graphite: Density, Ib/ft® at 70°F Bending strength, psi Young’s modulus of elasticity, psi - Poisson’s ratio Thermal expansion per °F Thermal conductivity, Btu hr ™} ft 1 op-1 Electrical resistivity, ohm-cm X 10* Specific heat, Btu Ib ™! °F ™! at 600°F at 1200°F Hastelloy N: Composition, wt % Nickel Molybdenum Chromium Iron Manganese Silicon Boron Titanium Hafnium or niobium Cu,Co,P,S,C, W, Al Density, lblft Thermal conductmty, Btu hr f) -1 op-t Specific heat, Btu ! . ~ Thermal expansion, per F . Modulus of elasticity, psi - Electrical resistance, ugohm-cm Tensile strength, psi, approximate Maximum allowable design stress, psi Maximum allowable design stress, bolts, psi Melting temperature, approximate, °F 64 930 208 24b 0.71% 0.32 <0.1 NaBF4-NaF 92-8 104 725 117 3.4 0.23% 036 252 115 _ 4000—6000 1.7 X 10° 0.27 23X 10°° 35-42 8.9-9.9 0.33 042 Balance 12 7 0—-4 0.2-0.5 0.1 max. 0.001 max 0.5-1.0 - 0-2 03S At80°F At 1300°F. ~557 ~541 6.0 12.6 0.098 0.136 57%x10°° 95x107° 31x10° 25x10° 120.5 126.0 115,000 75,000 25,000 3,500 10,000 3,500 2,500 2,500 “Does not include upper extension cylinder. bIndicates change from or addition to data reported in Table 5.1 of ORNL-4449 “Due to decay of gases and noble metals only. 4Dye to varying heat transfer reglmes, termmal temperatures are not an indication of average driving force. 5.2 MSBR PRIMARY SALT STORAGE TANK E.S.Bettis H.L.Watts The only significant change in the single-fluid MSBR reference design since last reported® has been the addition of a primary salt storage tank. This tank would ‘be used to store the primary salt in event of the need to repair the primary drain tank or its heat removal system. . The storage tank is located in the chemical processing cell. After insertion of a spool piece in the transfer line, the salt would be moved to the storage tank by a jet pump located in the bottom of the primary drain tank. The spool-piece arrangement is used to provide assurance that the salt could not be transferred inadvertently. The salt is returned to the primary drain tank by a similar jet pump located in the bottom of the storage tank. Both jet pumps are actuated by a small auxiliary salt circulation pump. The 46 : dechyed before a transfer is made. The heat removal storage tank has only about 1 Mw of heat removal ;:apacity, since afterheat in the primary salt will be well 2MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, S540P 1000 F RH | REHEATER COOLANT SALT system will be similar to that used in the MSRE drain | tank, where water is boiled in double-walled U-tubes immersed in the salt. The storage tank is provided with electric heaters and thermal insulation to maintain the salt above its liquidus temperature. 5.3 STARTUP AND SHUTDOWN PROCEDURES E.C. Hise S.J.Ditto J. L. Anderson Preliminary studies were made of the startup and shutdown procedures for the MSBR reference plant to be sure that these special conditions, as well as those for partial load operation, would not add undue operating complexity or require an unreasonable amount of special equipment. A preliminary flowsheet developed for these special situations is shown in Fig, 5.1, 5.3.1 Startup Procedures Two startup procedures must be considered: (1) the cold startup, with all systems cold and empty, and (2) hot restart from a hot standby condition. As in any ORNL-OWG 70-6737 DEMINERALIZER 1 850 F CONDENSER —, I {1.5 in Hg ABS) HOT Fssévsnsn WELL EXTRACTION STEAM HEATERS E : FROM HPT [ RSE 13600 P 850 F et [T PUMP , REHEAT STEAM PREHEATERS 1050 P : - 950 ¥ T BF PUMP BTVY TURBINE BOILER sD BF| i pr———— 1 100 P BE}FDe:wgcnou - DEAERATOR EMPERATO T 860F : STANDBY - p—————————— o © POWER TURBINE DSH STEAM ‘ GENERATOR 3%'&.%?:&1_ DESUPER- DRYER VALVE : ] HEATER COOLANT SALT 1150 F 1100 P 556 F. EXTRACTION STEAM DURING - HEAT REJECTION START-UP AND STANDBY * THROTTLE VALVE _ ‘ _EXTRACTION STEAM 3300P 3590 P : -T£H - - 700 F z ? 895 F | L b} M COOLANT SALT 850 F PBP MIXER s6 BF BOOSTER PUMP STEAM GENERATOR . AUX BOILER {200,000 Lb/hr) HIGH PRESSURE FEEDWATER HEATERS 1P FEEDWATER HEATERS . IP BOQOSTER PUMP A-BFP AUX BFP MOTOR DRIVE | Fig, 5.1.- MSBR Steam Plant Startup and Shutdown System. thermal power station, the ability to hold the system in hot standby and to achieve quick starts from this condition is imperative to avoid excessive outage times for the plant. - Cold Start. — A normal startup from the cold-empty condition proceeds as follows: The primary and second- ary cell electric heaters are turned on, and the primary and secondary salt circulation pumps are started to circulate helium in the salt systems. When the tempera- ture of the secondary system reaches 850°F, the loops are filled with secondary salt from the heated drain tank, and circulation of salt is started. When the primary and secondary systems reach 1000°F, the primary system is filled from the primary-salt drain tank, and circulation of salt is commenced. Both salt systems confinue to circulate isothermally at 1000°F until power escalation is started. The primary and secondary salt flow rates are at the levels required for the lower end of the power range. The reactor is made critical at essentially zero power using standard flux control methods. This operation requires removal of safety rods and further addition of reactivity by insertion of graphite control rods under the surveillance of startup instrumentation and a flux - level control system. When the power reaches an appropriate level, which is still below the sensible power generating range, the automatic neutron flux level controller assumes control of the power. Concurrently with the salt systems being electrically heated, the steam system is being warmed and brought to operating conditions by means of an oil- or gas-fired auxiliary boiler. Deaeration and demineralization of the feedwater and warmup of piping, feedwater heaters, turbines, etc., proceed in a conventional manner with steam taken from this boiler. To avoid excessive thermal gradients in the steam generators, steam that is admitted must be nearly at full operating conditions of 3600 psia and 1000°F. As the auxiliary boiler is being raised to this pressure, steam from it is throttled through the boiler extraction valve BE, is passed -.through the . desuperheater DSH, and is used. for feedwater heating, for warming and rolling the boiler feed pump drive turbines BFP-T and for warming the . high-pressure feedwater heaters. When the auxiliary boiler reaches full pressure and temperature, circulation can be started through the steam generator. When the steam system is ready to begin to take on load, the set point of the flux controller is adjusted as 47 As the steam load is slowly increased, the reactor power is matched to the load, and salt temperatures are kept at the desired level by manipulating the flux set point,. At 2 to 10% power, temperature changes are slow, and such control should not be difficult. When the load - reaches 800,000 Ib/hr, or about 8 to 10% of full load, the reactor can be put in a temperature control mode instead of a flux control mode after matching the temperature set point with the existing outlet tempera- ture. The load should be held essentially constant until the system comes to equilibrium, and then the reactor outlet temperature set point should be adjusted to meet the requirements for subsequent load-following control. The boiler feedwater pressure booster pumps FBP can then be started to raise the steam generator inlet pressure to about 3800 psia, and the auxiliary boiler and its feedwater pump can be taken off the line. The - system is now self-supporting at about 8% load. Part of the steam generator output flows to the mixer M via the reheat steam preheater, and the remainder passes through the boiler extraction valve BE to drive the main boiler feed pumps, etc. The main turbines, which have previously been warmed as mentioned above, are now gradually brought up to speed and temperature, using steam from the hot standby equip- ment at first and then steam taken directly from the steam generators. | . The load is then increased to 20%, at which time the steam temperature controller is activated. At this power level the “normal” control system regulates steam temperature and reactor outlet temperature as func- tions of load. To prevent undesirable transients, the various system parameters and set points should be adjusted to the requirements of the existing power demand prior to closing the fully automatic loop. The precise conditions at which the various steps of the program are taken, as well as the rates of change of the different variables during the startup, must be determined by analyses of the complete plant. The extent of the present study is limited and is only - intended to establish feasibility, Given consistent and -acceptable steady-state (or quasi-steady-state) condi- tions for the startup operations, it appears that the plan ‘described can be implemented using already demon- required to maintain the desired salt temperatures as - the feedwater flow is increased. The feedwater tempera- ture to the steam generator is reduced by tempering the - ~ feed steam with 550°F water in the mixing chamber M. strated techniques of controlling the MSRE. “Hot Standby and Startup. — On reduction of the main turbine load and closure of the stop valve SV, steam will be immediately let down through the boiler extraction valve BE, through the desuperheater and heat rejection valve HRTV, and thence to the main turbine condenser. Except for extreme situations of sudden loss of turbine load, and possibly not then, the boiler pressure relief valves need not vent steam to the atmosphere. A portion of the steam from the steam generator can be used to drive the boiler feed pump turbine BFP-T 48 and to continue circulation of the feedwater to the steam generators for heat removal and rejection to the turbine condensers. Another portion of the steam will continue to drive the standby steam turbine-generator set to supply standby power (if not available from the grid through the station service transformer) to drive the salt circulation pumps, some of the main con- densing water supply pumps and hot well, pressure . booster, and other pumps required to maintain the feedwater system operative. - Afterheat from the reactor system will continue to be transferred to the steam system and maintain it at ~ operating temperature - for several hours, depending upon the burden of fission products in the system. As this heat source decays, the auxiliary boiler can be started if it is desired to maintain the system in the hot standby condition. The time required for restart from this mode, as described above, would be limited only by the acceptable rate of temperature rise in the main turbines, as in a conventional steam system. 5.3.2 Normal Shutdown A normal shutdown proceeds as follows: The system power is reduced under control of the operating circuits until\about 8% of full load power is reached. As outlined above, the hot standby condition is reached by gradually reducing the flow to the main turbines to zero and at the same time transferring the generated steam to the hot standby system through the boiler extraction 5.4 DESIGNS FOR FIRST-GENERATION MOLTEN-SALT POWER REACTORS E.S.Bettis H. L. Watts ~ 5.4.1 General The molten-salt reactor design effort to date has concentrated on developing a concept with the highest performance consistent with a feasible and practical design. As discussed in a previous report,? the criterion for performance was taken to be the combination of ‘breeding gain and salt inventory requirement that would result in- best utilization of the nation’s fissile fuel resources. However, if breeding capability is given reduced emphasis and higher priority is given to simplifying the plant and its operation, other designs of molten-salt reactor plants would result. Some of these would be more immediately realizable with less de- - velopment, since they would be similar to scaled-up valve BE and thence to the turbine condenser, as | required. If it is desired to stay in the hot standby condition, the auxiliary boiler can be started; if not, the main turbine can be allowed to cool, the rate being _ controlled by admitting some steam from the steam - dryer SD through the turbine seals and warmup system. Feedwater will continue to be supplied to as many of the steam generators as required (probably one or two) to remove reactor afterheat and to maintain the desired salt temperature profiles. After about ten days of afterheat removal (depending on the operating history of the reactor), the primary salt can be transferred to the drain tank, where the heat rejection system for that tank will continue to dissipate the heat. Cell heaters will - pick up the load to maintain the secondary salt in the molten condition until it is drained. With termination versions of the MSRE, which accumulated several years of successful operating experience. Although the simpli- fied plants would not have as high a performance as breeders (when processing methods are successfully developed), they would, as converters, have high con- version ratios, and the fuel cycle costs would be low enough to compete favorably with other converter reactors. Further, these molten-salt converter reactors would be able to use either 235U or Pu as fissile fuel makeup with about the same low fuel cost. Design and operation of these simplified plants would not represent a diversion from the main goal of achieving a molten-salt breeder reactor having good fuel conservation properties. They would, rather, be an intermediate step, yielding valuable experience directly applicable to the high-performance breeders to follow. Simplification of molten-salt reactor plants can be achieved in two important areas: (1) The complexity of the “processing of the primary salt can be greatly = reduced if the treatment consists simply in recovery of the uranium by fluoride volatility, a process which is “already relatively well developed. The carrier salt would be discarded after it became too contaminated with fission products for economical operation. (2) Use of lower power density within the reactor core would ‘provide a core graphite life equal to that of the rest of the plant. Elimination of the need for periodic graphite replacement would simplify the reactor design, greatly reduce the amount of special equipment needed for - core maintenance, and moderate the risks and nuisances of all steam generation, the steam system can be allowed to cool. - 3MSR Program Semiann, Progr. Rept Aug. 31 1967, ORNL-4191, p. 82. involved in disposal of a series of spent reactor core assemblies. To date, we have bnefly considered two single-fluid molten-salt reactor concepts for use in such power plants. The reactor designs are discussed here; the nuclear characteristics and fuel cycle performance are reported in Sect. 6.1.2. 5.4.2 Large MSRE-Type Reactor We began our investigation of these reactors by instructing the ROD computer code to optimize on the conservation coefficient that has been used previously in molten-salt reactor studies but to limit the peak flux " to a value such that the damage fluence to the graphite would not exceed 3.2 X 10%? neutrons/cm? in 30 years, This resulted in a reactor with a cylindrical core about 26 ft in diameter by 26 ft high. Although the core is much larger than the core of the reference design, replacement of the graphite would not be necessary, and the reactor vessel could be a cylinder with dished heads welded at top and bottom. As shown in Table 6.2, the reactor could be operated as a converter with only fluoride volatility processing of the 49 As shown in Fig. 5.3, the top grid is made up of l-in.-thick graphite plates. The top 4-in. length of each " prism is turned to a diameter of 2 in. to fit into holes in the plates. Each of the principal plates has 16 holes. At " the points where the corners of four 16-element groups ‘meet, a smaller *5-in-thick graphite plate with four fuel or as a breeder with complete fuel processing and with little difference in estimated fuel cycle cost. The increased core life is, however, obtained at the expense of an increase in inventory of fissionable material over the reference design and a reduction in fuel yield when . operated as a breeder. A conceptual layout of the reactor vessel is shown in Fig. 5.2. The similarity to the MSRE core is evident. The vertical graphite prisms constituting the core have a 4-in.-square cross section and have a central hole as well as milled flow passages in the four faces to provide channels for flow of salt upward through the core. The holes would be of different diameters, and the depth of holes is placed over the corner prism in each group to tie the four groups together. These smaller tie plates have standoff dowels to hold them away from the top reflector to provide a flow passage for the salt. The elements are positioned at the bottom of the core by graphite slabs arranged in egg-crate fashion, as shown in Fig. 5.4. The grid is arranged into squares, each accommodating 25 elements. By indexing the elements in groups of 25 at the bottom and in groups of 16 at the top, the position of the core graphite is stabilized, and gap distances cannot accumulate. The prisms are thus restrained in the radial direction but can move axially with temperature and radiation-induced dimen- sional changes. ' A 2%-ft-thick reflector is provided around the core for better neutron economy and to protect the vessel wall from radiation damage. The reflector is made up of large pieces of graphite which have salt flow passages to provide the necessary cooling. The graphite core ele- ments normally float in the fuel salt. When the reactor is empty of salt, the elements rest on the bottom reflector. A possible major shortcoming of this design is the 26-ft length of the core elements. Difficulty could be encountered with impregnation of the graphite surface - inside the central hole, and this would permit per- the passages on the faces would be varied, dependingon ‘the distance from the center of the core, to provide three different regions of salt-to-graphxte ratio as indicated in Table 6.2. : Each of the graphite prisms in the core must be meation of the graphite by '*°Xe and increased neutron losses. Also, elements of this length pose fabrication and handling problems. 5.4.3 Spherical Reactor with Graphite Ball Bed Procurement of graphite for a molten-salt reactor : would be greatly simplified if graphite spheres were constrained to a definite position with relation to other - elements in order to preserve the desired flow distri- bution. This is necessary because the temperature -coefficient of expansion of Hastelloy N is greater than that of graphite, which would tend to cause a tightly assembled unconstrained core to become loose when the system is brought up to temperature. Further, radiation-induced damage causes the graphite to shrink over certain ranges of neutron fluence. An indexing grid for the elements is therefore used at the top and bottom of the core. used as the moderator. We accordingly made a study of a spherical reactor vessel containing 6-in.-diam graphite balls. As may be noted in Table 6.2, however, the salt-to-graphite ratio in this type of core is about 0,37, a salt fraction too high to be economical when the core is - made large enough to obtain the low power density needed for long graphite life. Since this is true, the computer code was allowed to optimize fuel costs as a function of the reactor core diameter. On this basis a core diameter of 19 ft resulted, but the estimated core graphite life is only about 9 years. A conceptual drawing of this reactor type is shown in Fig. 5.5. 43 ft 6in. 264 Oin ORNL-DWG 70-4304 (3) « i PIPES O, (3) 18-in. PIPES 26ft Oin. - 31t Tin. Fig. 5.2 Conceptual Design — lOOO-Mw(e) MSRE-Type Motien-Sali_Reactor, Elevation. 51 ORNL-DWG 70-€6738 m D :0;0;0. e | &R H R w - ’ W w \\ Q ‘; o ) i w t a ' o @ E : | ? E i O s 1 . 3: 0;0;0; i > @ (%.Eb 3 in. _ ] TIE PLATE'./ \UPPER GRID AND ORIFICE PLATE Fig. 5.3. Top Plate Grid for MSRE-Type Reactor. _ ] (Y Y e a N N/ o/ ./ N4 ./ e } = ) N N L \_J Y, "/ /I L/ ) Y, \/ ) \/\/] <_/ .’ (\. S/ NG - 2 é/g | 5 1 e — " e U T o U A T tatieietelii ittt ittt dteiet et EZ { () £ rY (Y Z \,) \_) .’ ../ \_/ ................................... 2 1 : : . 1 - Fig. 54. Lower Grid for MSRE-Type Reactor. 52 ORNL DWG 70-4305 . TO PUMPS (3 TOTAL) RETURN (3 TOTAL) Fig. 5.5. Concepfnal Design of Spherical Molten-Salt Reactor with Ball-Bed Moderator. ‘While -the moderator graphite in this concept has to be changed probably twice during the life of the plant, it may be easier to design a maintenance system to handle the small spheres rather than the long prisms. One way of changing the spheres might employ a release gate above the salt-liquid level in the reactor which would utilize the buoyancy of the graphite to force out a sphere in an inversion of the method used in the German pebble-bed reactor. Batch removal could be effected by an unloading mechanism operating through a relatively small opening at the top of the vessel. This molten-salt reactor concept has not received extensive study but appears to have interesting possi- bilities for development into a practical reactor. 5.5 BAYONET-TUBE HEAT EXCHANGERS - C.E.Bettis - Molten-salt breeder reactors must be designed to minimize fuel salt volume in order to attain a good doubling time. The primary heat exchangers which have been designed to date for the molten-salt breeder concepts have been shell-and-tube types which use small tubes to carry the primary salt and are baffled to improve cross flow of secondary salt on the shell side. The only method considered practical for maintenance of such exchangers is to replace the tube bundle if a leak occurs. In order to make pluggmg of individual tubes possi- ble, a different primary heat exchanger concept has been devised which uses a central tube and an inner annulus for secondary salt and an outer annulus for primary salt. As shown in Fig. 5.6, each concentric tube assembly becomes a separate heat exchanger in which one can conceivably detect a leak and plug the assembly. ~ Since the maintenance possibilities are attractive, several exchangers of different sizes have been analyzed to see what size, shape, volume, etc., would be necessary to enable such an exchanger to be substituted for the primary heat exchanger of the MSBR reference design. A computer program was written to make the calculations. It determines the tube length and associ- 53 ated parameters for a given configuration, pressure -- drops, and number of tubes. The wall thickness of the tubes is adjusted to withstand the maximum pressures encountered. The log mean temperature difference allows for heat transfer to the coolant salt in the inner tube. Table 5.2 compares the reference design MSBR primary heat exchanger with the reentrant tube ex- changer designed for similar conditions. ORNL-DWG 70-6740 Fig. 5.6. Model Exchanger Unit, for Calculation of Bayonet Tube Heal S, 6 DISTRIBUTION OF TRITIUM IN AN MSBR R. B. Briggs A 1000-Mw(e) MSBR that operates at 2250 Mw(th) will produce tritium at a rate of 2420 cunes per full-power day. The sources of productlon are:* " R.B. Korsmeyer - 4H. T. Kerr and A. M. Perry, Tritium Production in MSBR’s, MSR-69-116 (Dec. 3, 1969). Ternary fission - 31 curies/day SLioya)H 1210 ' TLi(n,an)’H 1170 wF(rz,”O)BH ) ) 9 2420 The radiation emitted by tritium is low in energy (18.6-kev beta), and the hazard is small in comparison .with the hazard of most fission products. Tritium does, however, exchange with the normal hydrogen in hydrogenous compounds, and once it is released it becomes a part of the hydrogen of all living things about in proportion to its fraction of the hydrogen in water in the surroundings. In these preliminary calculations of the distribution in the MSBR reference design, we-assumed that tritium, shortly after birth, will be present as molecules of tritium gas (T, ) or tritium fluoride gas (TF) dissolved in the primary salt. The relative amounts of T, and TF will depend on the reducing power (the ratio of the concentration of UF; to that of UF, in the primary salt). The fraction of . tritium present as T, would increase with increasing UF3/UF,;. T, and TF in the fuel salt were assumed to diffuse to the helium bubbles that are provided to remove gaseous fission products and to be removed with the helium to the off-gas system. They were also assumed to diffuse to metal and graphite surfaces. At the metal surfaces the T, could dissociate into tritium atoms, dissolve in the metal, and diffuse through the metal walls. The TF could react with the metal wall to liberate tritium atoms, but we assumed that the reaction rate _of TF with the metal would be slow in comparison with the dissociation rate of T, and that none of the tritium in - TF would be available to diffuse through the walls. This assumption tends to give low results for the amount of tritium that would reach the steam system. The gas that diffuses to the graphite would reach a steady concentration in the pores that would be proportional to the concentration in the salt, or it would react with the graphite to form methane or other hydrocarbons that could not diffuse through metal and would be removed in the helium bubbles. Neglect of ~ these reactions tends to give high results for the amount of tritium that would reach the steam system.. _ The tritium that passes through the walls of the piping and reactor vessel would be contained by the reactor cell, and that which passes through the heat exchanger tubing would enter the secondary salt. In the secondary system, tr1t1um was assumed to be present only as T,; no reactlons Wlth the salt were consxdered For purposes of these calculations, we assumed that helium bubbles could be injected into the secondary salt and removed to an off-gas system as in the primary system. Some of the T, would enter the bubbles, and the remainder would diffuse to metal surfaces, dissoci- ate, and diffuse through the metal. In the secondary ~ system, tritium that -passes through the pipe walls would be contained in the secondary system cell. However, tritium that passes through the walls of the tubes in the steam generators, superheaters, and re- heaters would exchange with hydrogen in the steam. If most of the tritium were to reach the steam, it would give rise to major problems in operatmg the plant and disposing of the tritium. - The calculations were made by use of the usual mass transfer and diffusion relationships. For transport through the salt to metal, graphite, or bubble surfaces, Q= kA(C, — Cy), where, in consistent units, Q = net tra_nsbort rate, k= mass transfer coefficient, A =surface area‘, ) C,, = concentration of T, or TF in bulk salt, Cg = concentration of T, or TF in salt at surface. For diffusion through the metal walls, Q——(P"2 -p;*), where - P = penetration coefficient for the metal, t = thickness of the metal, p; = partial pressure of Tz at the mner surface of the ~ metal, p,, = partial pressure of T, at the outer surface of the metal. For all the calculations the concentration of tritium in * the salt at the surfaces was assumed to be in equilibrium with the concentration in the gas in the bubbles or on the metal surface, so p = HCg, where H is the solubility - coefficient (or Henry’s law constant) that relates the concentration of T, in the salt to its partial pressure. Calculations were run for a variety of conditions. The first were for conditions generally typical of the MSBR reference design. Those conditions were then modified - 55 Table 5.2, Companson of Reference Design MSBR Primary Heat Exchanget Basis for Hastelloy N properties with a Bayonet-Tube Exchanger MSBR Refcrencp Design Bayonet-Tube Primary Heat Exchanger® Heat Exchanger Rate of heat transfer per unit Mw 563 , 563 Btu/hr 1.923 x 10° 1.923 x 10° Tube-side conditions Hot fluid . Primary salt Primary salt ‘Entrance tcmperatgre, 'F 1300 1300 Exit temperature, F - 1050 1050 Entrance pressure, psi 180 180 Pressure drop across exchanger, psi 130 - 130.5 Mass flow rate, Ib/hr 23.74 X 10° 23.74 X 10° Shell-side conditions Cold fluid o Secondary salt Secondary salt Entrance temperature, F 850 850 Exit temperature, F 1150 1150 Exit pressure, psi 34 34 Pressure drop across exchanger, psi 115 115 Mass flow rate, Ib/hr 17.8 X 10° 17.8 X 10° Tube dimensions, in. 0.375 OD D; = 0.6849 ' 0.035 thick D5 =0.8049 D3 =0.9249 D4 = 1.0449 Ds =1.1969 D¢ = 1.2500 D-] =1,3900 Dg =1.4525 Tube length, ft 22,07 20.03 Shell ID, in. 66.2 85.1 Number of tubes 5543 2199 Pitch of tubes, in. 0.75 . L5775 " Total heat transfer area, ft2 12010.67- 14414 Basis for area calculation Outside of tubes Surface of D6. Overall heat transfer coefficient, 921 801 U, Btuhe™ ft72 °p 7! Volume of primary salt in tubes, ft 62.07 88.8 Basm for molten salt properties MSR Memo 68-135 MSR Memo 68-135 CF 64-6-43 CF 64-6-43 Data for tfie reference design heat exchanger do not exactly agree with data in Table 5.1 because the latter are based on more recent salt physical property data, ~ to include injection and removal of helium in the secondary salt system, increased rates of helium purge - in primary and secondary systems, and variation in UF,/UF,. Also calculated were the effects of adding hydrogen (supplied in the helium purge) to the primary salt at rates 10? to 10® times the tritium generation rate and of coating the heat exchanger and steam generator, superheater, and reheater tubing to increase the resistance (decrease the permeation coefficient) by factors of 10 to 1000. | ' Results of the calculations are summarized in Table 5.3. In the reference design, with shell-and-tube primary heat exchangers, a helium purge flow of 10 cfm in the ‘primary system, and little or no purge in the secondary system, about 69% of the tritium was found to pass into the steam system, 18% to diffuse through the - "metal walls into the cells, and 13% to be taken into the primary off-gas system in the helium bubbles. In- creasing the helium purge rate in the primary system to 100 cfm and providing a helium purge of 100 cfm in the secondary system reduced the amount of tritium reaching the steam to 28% of the production. The amount entering the cell was reduced in about the same proportion. The effect of adding hydrogen to the Table 5.3. Tabulated Results of the Computer Runs Helium Purge Tube-Wall Resistance it " Tritium Distribution (percentage of production) R tive to Ref: i : um ; . _ ate (cfm) {Relative to Reference Design) Generation Rate ] Primaty - Secondary Total Cells Steam Primaty Secondary Primary Heat SteamSystem (moleculeg/sec) _YSiom PW8® System Purge Primary Secondary System System | Syftem Exchanger Eqmp@et;t T, TF Purge 'System System . Reference Design 10 - 0,2 . | X1 29E17% 58 1.0 0.1: 129 8.7 . 94 69.3 Effect of Purge Rates 100 0.2 X1 X1 , X1 8.8 45.2 0.0 540 44 49 36.3 10 20 X1 X1 X1 722 18 6.2 212 107 8.1 60.3 10 50 X1 X1 X1 55 68 13.6 259 8.3 7.9 58.8 100 100 x1 X1 x1 83 440 111 634 43 3.9 28.2 | . Effect of Increasing Tube-Wall Resistance 100 100 X1 X10 x1 8.3 440 111 634 4.3 3.9 28.2 100 100 X1 X100 X1 8.7 451 14.0 618 4.6 438 22.5 100 100 x1 . X1000 X1 10.2 48.6 23.3 821 5.3 8.1 5.1 100 100 ~ X10 X10 X1 8.7 451 108 646 4.6 3.7 27.4 100 100 X100 X100 x1 1.6 520 10.0 736 6.0 3.4 17.4 100 . 100 X1000 %1000 X1 1.7 64.2 5.2 87.1 84 1.8 2.2 _ Effect of Adding Hydrogen to Primary Salt | 100 100 X1 X1 - x10? 159 61 194 414 36 61 485 100 100 X1 X1 x10* 326 09 268 603 08 24 36.2 100 100 X1 X1 x10% 782 01 170 953 0.1 0.3 44 100 100 ox1 X100 x10? 214 71 470 755 45 137 7.3 100 100 X1 X100 x10 390 1.0 556 956 09 3.7 0.8 100 100 X1 X100 X108 788 01 211 1000 0.1 0.3 0.0 100 100 X100 X100 x10? 602 1.8 1L8 838 87 3.8 3.5 100 - 100 X100 X100 x10* 949 15 24 98.8 LS5 0.5 0.2 100 100 X100 X100 x10% 1002 0.2 03 1007 0.1 0.0 0.0 10 0.2 X1 X1 10% 19.5 0.1 0.2 198 1.1 4.5 73.7 10 20 X1 X1 10% 178 - 01 155 334 11 38 . 61.4 10 20 X1 X100 10 3., 02 -s8.1 89.8 14 7.9 1.7 10 20 X100 X100 10* 914 03 3l 948 25 15 0.4 , Effect of Increasing UF3/UF, from 0.001 to 001 ' o 10 20 X1 X1 X1 60 07 1.0 13.7 89 9.2 67.9 100 100 X1 X1 10* 334 01 274 609 08 2.5 36.7 10 02 Xt - . X100 10* 749 00 - 25 774 22 16.2 3.3 100 100 X1 X100 - 10 385 01 556. 942 09 3.7 0.8 100 100 X100 X100 10* 945 0.1 24 970 LS 0.5 0.2 %Tritium generation rate is 2.9E17 molecules/sec. X1 indicates that no hydrogen is added to salt. X10% to X10° indicates that hydmgefl is added to salt at rates 10% to 10° times the Ty production rate. reference design was to increase the amount of T, reaching the steam system until the ratio H,/T, exceeded 10%. This is because addition of hydrogen results in an increase in the hydrogen concentration in the salt, a decrease in the ratio HF/H, in the salt,.and a 57 decrease in the fraction of tritium that is present as TF. . This makes a larger fraction of the tritium available as HT, a form that can dissociate and diffuse through the walls, Increasing UF3 /UF,4 had the same effect. Increasing the hydrogen addition rate to 10® times the rate of tritium production resulted in only 4% of the tritium reaching the steam system and 95% being removed by the purge systems. This amount of dilution of the tritium by hydrogen is probably excessive and would create problems of handling and storage. In the absence of added hydrogen the partial pressure of hydrogen (as tritium) in equilibrium with its con- centration in the primary salt is 107 to 10™* torr. At this. low pressure the salt films provide much more resistance to transport of tritium than do the metal walls. Increasing the resistance of the tube walls in the steam generating equipment and in the primary heat exchangers by factors up to 100 had no useful effect. With the tube wall resistances increased by a factor of 1000 and with maximum purge rates in primary and secondary systems, the transport to the steam system was reduced to less than 5% of the production. It is desirable to reduce the amount of tritium that enters the steam system to 1% or less of the production. The calculations indicated that this could be accom- plished by increasing the tube wall resistances by a factor of 100 and adding hydrogen to the primary ‘system at a rate near 10° times the production rate of the tritium. Under these conditions the UF3/UF; in the primary salt is unimportant. Gas purge is required in the reactor primary and secondary system, but the purge rates need not be as great as 100 cfm. Use of tungsten or' ceramic coatings on tubes in the steam generating equipment and in the primary heat exchangers and use of high purge gas flows in the primary and secondary salt systems can be expected to increase significantly the cost and complexity of an MSBR -The requlrement to add hydrogen to the primary system is an additional but less expensive complication. We believe that the calculations reported here are conservative, and factors that were not considered could considerably reduce the measures necessary to prevent excessive amourits of tritium from reaching the steam system. They include the following: 1. Some data indicate that the diffusion of hydrogen through Hastelloy N at 1200 to 1500°F deviates from the relationship Q/A =Mp'/? at pressures below about 10 torrs and is better described by an equation of the form Q/A =Mp‘72 [Np" /(1 + Np™)] where %, < | < 1. This relatlonslup implies a higher resistance - to d1ffus1on 'of hydrogen through the metal than was used in our calculations. . The sodium fluoroborate salt contains hydrogen in hydroxyfluoroboric acid or some other as yet undetermined form. Tritium can be expected to exchange with that hydrogen and to be removable by processing the salt. : . Data from the MSRE suggest that the exchange of tritium for hydrogen in small amounts of methane added to the helium purge gas could considerably increase the rate of removal of tritium by the purge gas in the primary system. 4. Preliminary calculations indicate that use of the bayonet-tube-type primary heat exchanger (see Sect. 5.5) in the MSBR reference design would reduce the amount of tritium reaching the steam system from 69% to about 9% of the production without any other provisions, This is because most of the tritium diffuses through the large surface area provided by the outer tubes that separate fuel salt from the gas - space in the exchangers. - We plan to study the effects of these factors, but new experimental data are required for some before much more progress can be made in the calculations. 6. Reactor Physics A.M, Penry 6.1 PHYSICS ANALYSIS OF MSBR '6.1.1 Single-Fluid MSBR Reference Design H. F. Bauman | The nuclear data for the single-fluid MSBR reference design have been reported in previous semiannual reports.!»> However, we have made a new calculation of the MSBR reference design in order to take advantage of some improved capabilities in the ROD code and to make use of the latest available data. The processing section of ROD was expanded so that it can now completely describe a processing scheme with up to ten processing steps. A cost (dependent on the throughput) may be assigned to any or all steps.In addition, the complete fuel cycle cost section of ROD is now operational. The new reference calculatxon (CC120) dxffers from the previous calculation (CC58) as follows: 1. It follows in detail the reference reductlve‘extraction processing scheme (see part B of Table 6.1). 2. It includes the fuel cycle costs based on the latest material costs® and a preliminary estimate of the - probable processing cost. 3. It includes a calculation of replacement cost for the core graphlte . 4. The reference 7 Li enrichment in fresh makeup salt is 99.995%, whereas in CC58 a value of 99.9988% had madvertently been selected. 5. The plutomum nuclides are mc!uded exphc:tly in the nuclear calculation, : 1MSR Program Semiann. Progr Rept Feb. 28, 1969, ORNL- 4396, p. 77, 2MSR Program Semiann. Progr. Rept. Aug. 31, 1 969, ORNL-4449, p. 59. 3Rr.C. Rbbertson, Material Requirements in a Growing MSBR Power Economy — Revised, MSR-69-36 (Apr. 24, 1969). 6.7 It uses the latest cross-section weighting emplbying the nuclide concentrations from CC58 and the latest est:mate of the average temperature The results of the new calculation are g:ven in Table' 6.1. The indicated performance is not much different ‘from the earlier calculation. The breeding ratio is about the same, the fissile inventory is about 2% greater, and the peak power density is 8% greater. MSBR Processing Rate, — The MSBR reference design . was calculated for fixed processing rates corresponding to a three-day protactinium removal cycle and a 50-day rare-earth removal cycle, The optimum processing rates cannot be determined until the processing steps and their costs are further developed, but we have studied = the effect of processing rate on the performance of the reference MSBR. The effect of processing rate on the conversion ratio is shown in-Fig. 61. The 3000-day protactinium cycle time curve is essentially the no-protactinium-removal case. Without protactinium removal the reference - MSBR is at best a break-even breeder. Three fissile fuels were examined as makeup for cases with the conversion ratio less than 1, These fuels were 233U, mixed plutonium representative of light-water-reactor dis- charge, and 93%-enriched 23%U. The 235U seems to be the least desirable fuel because of the buildup of the neutron poisons 226U and 237Np. It must be noted, however, that 237 Np in these calculations was assumed to be processed out of the primary fuel salt at the very slow rate of 6.3% per year, that is, on a 16-year cycle. A much more rapid removal of neptunium, which may be feasible, would make the 235 U feed look relatively - more attractive, This study po_mts up the importance of rapid (and by implication, inexpensive) processing to the breeding performance of the MSBR and indicates the great extent to which the performance reported for the - reference MSBR is a function of the selection of reference processing cycle times. 59 Table 6.1. Characteristics of the Single-Fluid MSBR Reference Design Reflector A. Description . Identification CC120 Salt fractions Power © Corezone 1 0.132 Mw(e) 1000 Core zone 2 0.37 Mw(th) 2250 Plena 0.85 Plant factor 0.8 Annulus 1.0 Dimensions, ft Reflector 0.01 Core zone 1 -Salt composition, mole % Height 13.0 UF, 0.232 Diameter 14.4 PuF; 0.0006 Region thicknesses ThF4 12 Axial BeF, 16 Core zone 2 0.75 LiF 72 Plenum 0.25 Reflector 2.0 Radial ' Core zone 2 1.25 Annulus 0.167 Reflector 25 B. Processing Processing Group Nuclides Units, Full Power Cycle Time 1. Rare earths Y,La,Ce,Pr,Nd,Pm,Sm,Gd Days 50 Eu Days 500 2. Noble metals Se,Nb,Mo,Tc,Ru,Rh,Pd,Ag,Sb,Te Seconds 20 -3, Semi-noble metals Zr,C4,InSn Days 200 4. Gases KrXe Seconds 20 5. Volatile fluorides Br,I Days 60 6. Discard Rb,51,Cs,Ba - Days 3435 7. Salt discard ThsLi,Be,F " Days 3435 8. Protactinium 23 Days 3 9. Higher nuclides 2?"7Np, 242py Years 16 ~ C. Performance - Conservation coefficient, [Mw(th)lkg)]z‘ 14.1 Breeding ratio 1,063 Yield, % per annum (at 0.8 plant factor) - 320 Inventory, fissile, kg 1504 Specific power, Mw(th)/kg 1.50 System doubling time, years ’ 22 Peak damage flux, £ > 50 kev, neutrons cm 2 sec ™! Lo Core zone 1 3.5 X 1014 'Reflector 3,7 X 103 Vessel 4.3 x 10'1 Power density, wiem® ' Average 22.2 Peak 70.4 Ratio 317 Fission power fractions by zone ‘Core zone 1 0.790 - Core zone 2 - 0.130 Annulus and plena 0,049 . 0,012 Table 6.1 60 (continued) D. Neutron Balance Absorptions Fissions 2321y 0.9779 0.0030 233p, 0.0016 S 233y 0.9152 0.8163 234y 0.0804 0.0004 235y - 0.0747 0.0609 236y 0.0085 237Np 0.0074 238py 0.0074 239p, 0.0073 0.0045 240p, 0.0027 ' 241p, 0.0027 0.0020 242p,, 0.0006 Li 0.0035 TLi 0.0157 %Be 0.0070 0.00459 19g 0.0201 Graphite 0.0513 Fission products 0.0202 Leakage 0.0244 ne 2.2285 E. FuelCycle Costs? In mills per kilowatt-hour Inventory Fissile 0.276 Salt - 0.045 Replacement Salt 0.040 Graphite 0.095 Processing 0.307 Fissile production credxt -0.088 Total - 0.674 @(n,2n) reaction. DAt 10% per year mventory charge on materials, 13,7% per year fixed charge rate on processmg plant, $13/g 233y, $11.2/¢g 235y, $12/kg ThO,, $120/kg "Li, $26/kg carrier salt (including "Li). ' 6.1.2 Designs for First-Generation Molten-Salt Power Reactors H. F. Bauman The successful operation of the MSRE suggests that a moltensalt reactor designed to exploit today’s tech- nology could help to meet the rapidly increasing demand for electric power in the United States. The reference single-fluid MSBR design goes beyond MSRE technology in two important areas; first, the core power density is higher so that the moderator graphite must be replaced at intervals because of radiation damage, and second, an ondine chemical processing plant is required to separate the 2?3Pa and fission products from the fuel stream. The extension of the technology into these - areas will require some further development, and carrying - this development forward to the point where, for example, a chemical process has been demonstrated on an engineering scale will take time. ORNL-DWG €9-12698A 1.08 . . 3days Pa CYCLE TIME o _ . o | - _ 1.05 — 10 \‘:k : o %\ % | g , \\X = 3000 : < Na € 00 S~ o A - @ 1:§A\ | & 233y W A \ % \‘\\“ . Pu o _ : 4 o — 0.95 \ e — 233, \A\ \“ Pu . , \Azasu 0.90 : 0 100 200 300 400 _ 500 600 RECYCLE TIME (days) - Fig. 6.1. Conversion Ratio of Single-Fluid MSBR Reactors as a Function of Processing Cycle Times and Feed Materials. | In a search for reactor designs that could, in the meantime, exploit the technology that we now have in hand, we have calculated the nuclear performance of several designs that sidestep one or both of these development areas. With respect to chemical processing, we considered either the case of an equilibrium fuel cycle with continuous processing for isolation of protactinium and removal of fission products or the case of a batch fuel cycle without removal of poisons except by infrequent discard and replacement of the fuel salt. In the latter case, the uranium in the discarded salt (but not the plutonium) is assumed to be recovered by the well-established fluoride volatility process and recycled to the reactor. Thus the higher isotopes of uranium, as well as *37Np, are allowed to reach therefore half the batch cycle time. The fission product nuclides in general lie between these two extremes, and we have arbitrarily taken two-thirds as the average ratio of the equivalent equilibrium cycle time to the batch cycle time in calculating the cost of discarding salt. It is " not necessary to know this ratio-accurately in order to - make a good estimate of the fuel cycle cost, because for batch discard cycles longer than two or three years the costs are not very sensitive to the cycle time. This is because cost penalties associated with decreasing con- version ratio, as the batch cycle is lengthened, are -approximately offset by reduced salt replacement cost. _ . Perhaps the most obvious design, and one of the most attractive, is the “big MSRE,” a single-fluid prismatic- - core reactor like the MSRE and large enough (that is, of equilibrium. The average performance of the reactor - during such a batch cycle is simulated by an equivalent equilibrium fuel cycle calculation (performed with the ROD code). The equivalence. is at best approximate, because saturation - effects associated with neutron ~ absorption are different for different nuclides. For ~ rapidly ‘saturating nuclides, the average concentration approaches the concentration at the end of the cycle, and the equivalent equilibrium cycle time approaches: the batch cycle time. For slowly saturating nuclides, which essentially build up linearly, the average con- centration is approximately half the end-of-cycle con- centration, and the equivalent equilibrium cycle time is sufficiently low power density) that the ‘moderator ‘graphite would not have to be replaced because of radiation damage in the anticipated life of the reactor. The performance of this type of reactor, with and without chemical processing, is given in columns 2 and 3 of Table 6.2. In both cases the core salt fractions were adjusted by zones to flatten the damage flux near the center of the core, giving a smaller core for the same - moderator life. The case with processing is a moderate- performance breeder, with a fuel-cycle cost not much “higher than calculated for the reference MSBR (column 1). It differs from the reference breeder mainly in having a larger core and a higher fissile inventory. For 62 - Table 6.2. First-Generation Molten-Salt Reactor Designs CC-167-9 CC-186, - o 88-12, S-9-12, Case CC-120, \opEqype MSRETYPE — pjped, Ball Bed, Reference Prismatic, T]Pnsmatnc, a “Throwaway, Throwaway, MSER Pennanenfl " Penn::ez’ Replaceable? Permanent A. Description o ' , oo Thotium concentration, mole % 12 12 8 _ 5 5 Process cycle time, days , 3/50 30/100% 2700°¢ - 1330° 2350¢ Core volume fraction salt , o _ B ‘ ~ Zonel - 0.132 0.151 0127 0.37 0.37 Zone 2 037 0.115 0.123 : Zone 3 ' - . 0140 0.116 Core diameter, ft . 17 25 27 19 (sphere) 31 (sphere) B. Performance _ _ : - Breeding ratio? | 1063 1.048 0.837 0.661 0.818 Consetvation coefficient, [Mw(th)/kg] > 14.1 44 ‘ Yield, % per annum’ 3.20 1.55 , - Inventory, fissile, kg ' 1504 2350 _ 1790 : 1594 _ 3172 Core life, years ‘ 3.5 - 30 30 8.7 28.6 Fuel salt volume, total, ft* - 1683 2918 2940 1996 6463 Power density, w/cm » : ‘ : L ' " Peak 70.4 10.1 8.9 33.2 10.3 Average 22.2 6.1 5.1 22.3 5.1 - Ratio 3.17 1.66 L75 149 2.02 Fuel cycle cost, mlllslkwhr ' : : _ Inventory: Fissile ©0.276 0.430 0.320 0.271 - 0.560 Salt ) 0.045 0.078 : 0.073 0.043 - 0.139. " Replacement® 0.135 0.068 0.082 0.108 0.175 - Processing/ | ' 0.307 0.209 - 0,127 0131 - 0.142 Fissile feed or credit - —0,088 -0.066 - 0.180 0.400 - 0.203 Total fuel cycle cost 068 0.72 - 0,78 0.95 1.22 %The terms “permanent” and “replaceable™ indicate that the graphite moderator is, or is not, intended to withstand radiation damage effects for the full design hfe of the reactor, 24 full-power years. The term “tlirowaway” refers to a salt-discard fuel cycle as explained in the text. bProtactinium removal cycle time/rare earth removal cycle time (contmuous) €Equivalent batch cycle time, assumed equal to 1.5 times computed continuous cycle time. _ dAssuming that the plutonium discarded with the fuel salt is not recovered. “Replacement includes thorium, carrier salt, and, for replaceable cores, graphite, fBased on scaling the reference MSBR processing costs with a 0,5-power scale factor. Since this scaling does not weight fixed costs sufficiently at very low processing rates, an allowance for fixed costs of 0.05 mill/kwhr was added for the 30/100 case, and 0.1 mill/kwhr for the throwaway cases. NOTES: The calculations were based on a l(}DO-Mw(e) reactor with a thermal power of 2250 Mw(th) and an (.8 plant factor. The fissile feed in converter cases was assumed to be mixed plutonium typical of light-water-reactor discharge with the following composition: 23%Py, 60%; 24°Pu, 24%; 24! Pu, 12%; 2?Py, 4%. the no-processing or “throwaway” case, it is assumed that the salt is discarded at an economically optimum interval. The throwaway case is a high-gain converter reactor rather than a breeder, but the fuel cycle cost is very attractive. It proved economically worth while in this case to lower the thorium concentration to 8%, reducing the fissile inventory at the expense of reducing the conversion ratio. Such a plant could be designed for the addition of processing at some ]ater time, thus upgrading it toa breeder. Another design approach‘ is to make the core of a random-packed bed of graphite spheres. Such a-pebble- bed core has several advantages. The manufacture of the small spheres of high-grade isotropic graphite is entirely within today’s technology. A pebble bed will accom- modate radiation-induced dimensional changes in the graphite, and, finally, if- necessary, small graphite spheres can be replaced more easily than large prismatic elements At The main disadvantage of the pebble-bed core is that the salt fraction is inherently about 0.37, whereas the optimum for the best nuclear performance lies in the ‘range 0,12 to 0.15, Even when the thorium concentra- tion in the salt is reduced as partial compensation for the high salt fraction, as in the cases shown in columns 4 and 5 of Table 6.2, the performance of the pebble-bed cores is well below that of similar prismatic cores. In case S-8-12, the core size was optimized for minimum fuel-cycle cost and resulted in a graphite life of about 9 years. In case S9-12, the core was made large enough so that the graphite would last the 30-year life of the reactor, but as a result the fissile inventory and the fuel cycle cost are particularly high. - Of the designs considered here, the “big MSRE” with throwaway processing (case CC-167-9) appears to best meet the criteria of low-cost power and little develop- ment beyond scaleup from MSRE technology. It could - be designed for the convenient future addition of a processing plant and would provide an excellent vehicle for the first fullscale test of mtegral molten-salt processing. 6.13 Reactivity.Coefficients O.L. Smi_th J. H. Carswell, Jr. A number of isothermal temperature coefficients of reactivity were calculated for the single-fluid MSBR, using the reference reactor geometry shown in a _previous progress report.* These calculations were performed with a detailed two-dimensional representa- tion of the reactor in-R-Z geometry, using the diffusion code CITATION® with nine neutron energy groups. Both forward and adjoint fluxes were calculated, and the effects of various changes in microscopic cross sections or in material densities were calculated by first-order perturbation theory. The cross sections themselves were obtained from a series of calculations, using the code XSDRN,® in which group-average cross sections were calculated for each major region of the reactor for each of three different temperatures (800, 900, and 1000°K) and for various combinations of material densities. In this way the effects of tempera- ture-dependent chang&s in microscopic cross sections 4MSR Program Semiann. Progr. Rept. Feb, 28, 1969, ORNL- 4396, p.71. ST, B. Fowler and D. R. Vondy, Nuclear Reactor Core Analysis Code: CITATION, ORNL-TM-2496 (July 1969). ®N. M. Greene and C, W, Craven, Jr., XSDRN: A Discrete Ordinates Spectral Averaging Code, ORNL-TM-2500 (July 1969). 63 can be calculated separately from those of tempera- ture-dependent changes in density. | " The calculated reactivity coefficients are summarized in Table 6.3. The. Doppler coefficient is primarily that of thorium. The graphite thermal base coefficient and the salt thermal base coefficient, that is, the effects of microscopic cross-section changes caused by chang- ing the temperatures of the graphite and the salt, respectively, are positive because of the competition between thermal captures in fuel, which decrease less rapidly than those of a 1fv absorber, and thermal captures in thorium, which decrease nearly as 1fv, with increasing temperature. The salt density compo- nent represents all effects of salt expansmn including the decreasing salt density. The graphite density component includes both chang- ing graphite density and displacement of graphite surfaces. In calculating the displacements it was . assumed that the graphite-vessel interface did not move, that is, that the vessel temperature did not change. For short-term reactivity effects, this is the most reasonable assumption, since inlet salt bathes the vessel’s inner face. These dimensional changes in the graphite without a concomitant expansion of the vessel produce a significant change in the thickness of the salt annulus between the core and the reflector. The reactivity effect - of this change is not readily calculated by perturbation theory and was therefore obtained by comparison of two conventional criticality calculations with different thicknesses of the salt annulus and with appropriately differing core density. In any case, it should be noted that the graphite density coefficient is a small and essentially negligible component. From Table 6.3 it is seen that the total core coefficient is negative. But more important, the total - salt coefficient, which is prompt and largely controls the fast transient response of the system, is a relatively - Table 6.3. Isothermal Temperature Coefficnents of Reactivity, Reference Single-Fluid MSBR Reactivity Coef! ficlent, Component : ; 1ok . : = —(per . _ % 3T (p °0 | X 10's Doppler o - —4.37 -Salt thermal base = . +0.27 ~ Salt density _ . +0.82 "Total salt ‘ ~3.22 Graphite thermal base =~ +2.47 Graphite density - - - - ~0.12 Total graphite : +2.35 ~ Total core ' ~-0.87 large negative coefficient and affords adequate reactor stability and controllability. The salt density coefficient is particularly important 64 with regard to bubbles in the core salt. It is expected that the salt will contain about 1% helium bubbles. Under certain circumstances the bubbles might expand or collapse without change in core temperature and “hence without inwvoking the total salt temperature coefficient. Since the salt density component is posi- tive, bubble expansion would produce a positive reactiv- ity effect. Using a salt expansion-coefficient § V/V =2.1 X 107%/°C, an increase in core bubble fraction from, say, 0.01 to 0.02 would yield a reactivity change of 8k/k = +0.00039. This is approximately one-fourth the worth of the delayed neutrons in the core. Analogously, complete collapse of 2 0.01 bubble fraction would yleld a reactivity change of 8k/k = —0.00039. Finally, the fuel concentration coefficient, (6k/k)/(6n/n), where n is atomic density, was calculated - to be 042 for *33U and 0.027 for ?33U. The large difference between these two numbers is primarily a result of the substantial difference in concentrations Gie., n;3 = 11 X nys), so that a given fractional increase in 235U concentration produces a far smaller reactivity effect than does the same fractional increase in 233U concentration. ~ 6.1.4 Control Rod Worth 'O.L.Smith G.W.Morrison W.R.Cobb Calculations were performed to determine the reactiv- ity worth of four graphite control rods situated in the . central part of the reference MSBR core. The control rod region, shown schematically in Fig. 6.2, consists of four prismatic graphite elements each ~6 in. on a side and with a 4-n.diam hole down the center. The elements are separated by thin salt passages. To account for the influence of the remainder of the core on the ORNL—DWG 70—-6741 TYPICAL CORE CELL jCONTROL CELL _X_ ________ _ 7 7 . 2 2 s —-|§--0.63cm__| 10.46 cm | » GRAPHITE lsaLT ]—— {6.24cm ——-I REFLECTED BOUNDARY CONDITION — — — ____————_..._.a.._.___'.....——-————'-‘ s s \s\\\ x\\\ _—————-—-.—.——-—-—-———— A — — — ——— — e m—— — = Ry=0.635cm R, =5.08cm an 6.2. Geometry Used in Monte Catlo Calculatlons to Pre- pare Cross Sectlons. _ - prisms, the other with graphite in the hdles. The resuits neutron spectrum of the contro! cells, the control rod region is surrounded by a ring of cells typical of the adjacent core, followed by a reflected boundary condi- tion. The calculatlons were performed in two stages. The first stage was to generate cross sections. Using the Monte Carlo code ESP, the regions shown in Fig. 6.2 were mocked up exactly, thereby explicitly taking into account all geometric and energy-dependent self-shield- ing effects. Two Monte Carlo calculations were per- formed, one with salt in the holes of the control cell of these calculations were two sets of homogeneous macroscopic nine-group cross sections for the control region, that is, one set for each configuration. Each set was based upon 5000 neutron histories, and the statistical uncertainties in the cross sections are believed to yield uncertainties in reactivity an order of magni- tude less than the computed rod worth. Standard transport theory cell calculations (using the XSDRN code) were used to generate cross sections for the remainder of the reactor. The second stage of computation was to calculate rod worth. The two sets of cross sections described above were used in the central regions of two CITATION diffusion-theory calculations in which the entire reactor ORNL—DWG 706742 0.004 ll, _ 0.003 -~ s @ w / 3 0.002 /1 : L > ‘ / - > ‘ '— 2 0.001 / & /7 o9 / 17 L o 1 2 3 4 ~NUMBER OF GRAPHITE RODS IN Fig. 6.3. Control-Rod Worth, was mocked up in R-Z geometry. The central regions were of the same sectional area as in the Monte Carlo calculations but circular in shape. The standard trans- port theory cross sections were used in the remainder of the reactor. In the two CITATION calculations the value of k.gr was computed using each of the two Monte Carlo crosssection sets in the central region while keeping other cross sections fixed. The change in reactivity in going from four graphite rods out to four 10ds in was 8k = k;, — Ko, ¢ = 10.0033. Since the change in reactivity was anticipated to be small, it was felt that the geometry of the control region should be represented in as much detail as possible to account for all self-shielding effects. Ideally the entire reactor would have been represented in the two Monte Carlo calculations, giving control rod worth directly. However, the machine computing time re- quired to get adequate statistics for such a large volume is prohibitively costly. Since the remainder of the reactor outside the region shown in Fig. 6.2 does not. change composition, it was deemed satisfactory to treat it with more conventional and cheaper computing methods. | | - Using the same technique as described above, the worth of one rod and the worth of three rods were computed. The results of the calculations are plotted in Fig. 6.3, together with estimates of the standard deviation of the results. There appear to be no significant interaction effects between control rods, and hence a straight line appears to be the best representa- tlon of the data. 65 6.2 MSR EXPERIMENTAL PHYSICS 6.2.1 Indication of Integrated Power by 235U Depletion G. L. Ragan The fractional depletion of 235U in the MSRE fuel was obtained as a by-product of 235U capture-to- absorption ratio measurements.” The fractional deple- tion between initial and final samples taken for that experiment was about 10% below that expected on the basis of certain parameters then in use: a fission energy release of 199.7 Mev/fission® and a power calibration corresponding to a nominal full power of 8.00 Mw. During the present report period, these two parame- ters have been undergoing careful reexamination. The heat balance power calibration is discussed in Sect. 1.2.11. Revised values of fission energy release for several fissile nuclides and for operation with 225U and 233U fuels are reported below. The appropriate value is then combined with the observed 235U depletion rate to derive an independent measurement of the nominal full power: 7.30 £ 0.10 Mw. A comparison is made with ~ a measurement based similarly on the buildup of 236U in the fuel:® 7.45 £ 0.18 Mw. A combined value of 7.34 1 0.09 Mw is suggested. . ‘Some recent values for the recoverable energy re- leased in the fission process are given in the second column of Table 64. The value for 233U is that of Walker,!'® and the others are averages of those of Walker and of James.!! From the total energy released, derived from a mass balance, they subtract the esti- mated average neutrino energy and the long-term (>3 years) fission product decay energy. The latter amounts to about 0.2 Mev/fission; I estimate that subtracting an additional 0.1 Meyv/fission would exclude decay energy released after about six months, and subtracting an- .other 0.1 Mev/fission would exclude decays after about two months. "MSR Program Semiann, Pragr Rept. Aug 31 1969 ORNL-4449, p, 70. * 8p, N. Haubenreich, J. R, Engel, B, E, Prince, and H, C. Claiborne, MSRE Design and Analysis Report, Part IlI: Nuclear Analysis, ORNL-TM-730, p. 182 (February 1964), ' 9MSR Program Semiann, Progr Rept Feb 29, 1968, ORNL- 4254, p. 10, mW H. Walker, ‘Mass Balance Estimate of the Energy Released per Fission in a Reactor, AECL-3109 (April 1968); - also, Addendum of August 1969. Uy, F, James, Energy Released in Fission, AEEW-M-863 May 1969). Table 6.4, Recoverable Energy Released . - per Fission in MSRE Basic Fission - Typical MSRE Operation, Nuclide gfiififif; _ :ncluding Captux::a Capture? With *>*UFuel With **U Fuel 3y 188409 199.7%09 = 2004%09 23y’ 1925105 - 203.2%05 203905 238y 193808 206.9%10 207.6 £ 1.0 239y, 1982108 © 2124109 2132109 Overall 203.4+05 2010209 - 13, _ bEach of the v — l excess neutrons releases an average of 7.49 "Aléo ekcludmg. neutrino energy aind loflg-term (>3 year) fission product decay energy, Values taken from refs. 10 and Mevlcapture W1th 5U fuel and 7.92 with ° 23 U fuel, The; third and'fourth colu_mns of Table 6.4 include the energy released when the v — 1 excess neutrons (ie., the neutron increase resulting from the fission event) are captured in the reactor materials. The capture energy released was averaged over the capturing materials present by using a neutron balance table!2 for the reactor loading being considered. This capture energy includes, in addition to the binding energy of the captured neutron, the energy released in any subsequent radioactive decays with halfdives less than three years. For each of the two fuels, the last line shows the overall reactor average, the energy release in each fissile nuclide being weighted by the fraction of the fissions occurring in that nudlide. Each average is for a typical fuel salt composition and varies but little (about 0.1 Mev/fission) during‘the operational history with either fuel. - In the course of the 35U capture-to-absorption ratio 66 effective full-power hour. The standard deviation indi- | cated (*1.9%) represents a slight increaseé over the . purely statistical curve-fitting error (+1.7%) reported, to allow for possible errors in the 7% adjustment for fuel drains and flushes. The adjustment error, assumed to be $0.7%, is incorporated in the manner of independent errors. Using the measured capture-to-absorption ratio listed above, one can convert the 235U buildup rate to a 335U depletion rate and proceed as before to infer a value for nominal full power: 745 + 0.18 Mw. | In summary, we have made two determinations of nominal full power that are basically experimental. The only dependence on calculated cross sections is through small corrections to the experimental data (see ref. 7); for this, we use calculated ratios of absorption cross sections (23¢U/%35U and 238U/?35U), and for these ratios a generous uncertainty assignment (£10%) has been included. The two power determinations proceed from entirely independent experimental data, but much of the subsequent data reduction involves common parameters. Hence the two power estimates are only partially independent. A summary of the results, with an estimated overall combined value, is given in Table - 6.5. experiment,” the 235U depletion rate was found to be 5.545 *+ 0.056 ppm per effective full-power hour. This corresponds to a nominal full power of 7.30 + 0.10 Mw on the basis of the measured capture-to-absorption ratio, the recoverable energy per 235U fission, and relevant system characteristics: 0.2006 * 0.0024 capture per absorption in 23°U, 203.2 + 0.5 Mev per fission of 235U, 69.65 +0.17 kg of 235U circulating, and 0.980 + 0.002 of power from 235 U. | The above result may be compared with those obtained using data from an earlier experiment. Steffy and Engel® observed a net 22°U buildup rate which may be expressed as 0.0797 £ 0.0015 g of 23€U per 12g_E. Prince, private communication, It is mterestmg to compare the above results with those of Thoma and Prince (Sect. 10.3). Those authors compare a series of experimental * °Pu/22°Pu isotopic ratios taken over a long operating period with calcu- lated values for the same ratio. In the calculations the operating power level is treated as a parameter, and that power level giving best agreement between experimental and calculated values is found to be 7.41 * 0.05 Mw. The quoted uncertainty is that of fitting the experi- mental data to the calculated curves and does not include any allowance for cross-section uncertainties or for uncertainty in the energy release per fission. The good agreement with the results given above may be taken as confirming the essential validity of the cross . _sections used and of the calculational methods. Further confirmation is given by Thoma in Sect. 104, where experimental and calculated values of the 23"UI"”U -isotopic ratios are compared Table 6.5, Determinations of Nominal Full Power | Nominal Full Method Power (Mw) 235y depletion 7.30£0.10 236y buildup 7.4510.18 7.34 £0.09 Combined value 7. Systems and Component's' Development Dunlap Scott 7.1 MOLTEN-SALT STEAM GENERATOR R. E. Helms A steam generator development program plan is being formulated to develop the technology required to produce and successfully operate a molten-salt-heated steam generator. This plan includes programs for the procurement of steam generators for the MSBE from industrial manufacturers and for the construction and operation of a test stand for evaluating steam generator concepts proposed for the MSBE. A discussion of these two programs follows. 7.1.1 Steam Generator Indfisfrial Program A request for proposal péckage is being prepared for use in obtaining proposals from industrial manufac- turers to prepare conceptual designs of steam generators for large MSBR’s and to design, develop, and fabricate ~ models of those steam generators for the MSBE. Rough drafts have been written of the proposed scope of work, a general description of the primary and secondary salt systems and the steam system of an MSBR, and the design criteria for a molten-salt generator. 7.1.2 Steam Generator Tube Test Stand ; o R.E.Helms = J. P, Sanders J.L.Crowley H.J.Metz E. J. Breeding The preparation of the conceptual system design description (CSDD) for the steam generator tube test stand (STTS)' is nearing completion. All sections of the CSDD have been written and are being combined into “the first rough draft for project review, Present concepts of the STTS employ a salt system capable of operation with salt temperatures up to \MSR Program Semiann, Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 73-74. Table 7.1. Test Section Parameters Startup, Operation, 0-20% 20-100% Heat input, Mw 0-0.6 0.6-3.0 Steam flow rate, Ib/hr 0-2185 2185--15,925 - Feedwater inlet, °F Steam at 240 550 or 700 ' to water at ' : 700 ‘ Steam outlet temperature, °F 950—1100 1000 Feedwater inlet pressure, psia 15-4000 3100—4000 - Steam outlet pressure, psia 15-3600 2400 or 3600 Salt flow rate, Ib/hr 18,960 18,960-94,800 Salt inlet temperature, °F 1000-1200 1150-1200 Salt outlet temperature, °F 1000-850 850-900 1200°F and a steam feedwater system capable of operation at subcritical and supercritical pressures with feedwater temperatures up to 700°F. The STTS will be used to test 3.0-Mw units containing full-length tubes in configurations representative of steam generator designs -for a molten-salt breeder reactor experiment. - We have examined startup and operating parameters for a 3.0-Mw molten-salt-heated steam generator test section for testing in the STTS. The ranges of startup and operating parameters are shown in Table 7.1. The salt system for the STTS includes a salt pump, - - direct resistance heater as the heat source, associated piping to the test section, a venturi meter to measure salt flow, a drain tank, and an associated freeze valve, Gas-fired furnaces® were investigated as a possible heat source. However, heating by passing electrical current through a section of the loop piping is preferred - because of the smaller salt inventory in the loop, the 67 “shorter loop circuit time, the faster response to load changes because of the smaller heat capacity, and better - apparent control of the system over a wide range of operating conditions (20—100%). 2MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL- 4396, p. 99. Sodium fluoroborate, with impurities in the ranges shown in Table 7.2, is the reference salt. The STTS may also be operated with fluoride salt mixtures that melt at temperatures as high as 850°F and as low as 640°F. (Sodium fluoroborate reference salt melts at 716°F.) A ~ higher-melting-point salt is ‘the MSRE coolant salt, LiF-BeF, (66-34 mole %). A lower-melting-point salt is NaF-LiF-BeF, (27-35-38 mole %). Both of these salts - have a higher viscosity than sodium fluoroborate. A steam feedwater system has been selected that will permit a broad investigation of the technology of salt-heated steam generators. A pressure letdown system provides a facility that can be used for demonstration of the coupling, startup, and operation of the salt and steam-feedwater system with the salt in the shell side Qf | the steam generator test section at an initial tempera-. ture of 1000°F. The pressure letdown system also permits the use of full-flow demineralization w1th conventional equipment.. In the operation of a high-pressure, once-through steam generating unit, high-purity feedwater is required. - The feedwater purity criteria established for the STTS are shown in Table 7.3. This requirement is also critical for the STTS because solids in the feedwater will either Table 7.2, Impurity Range for Operation of STTS with Sodium Fluoroborate ‘ Fe,ppm | 50100 Cr,ppm ' 50-250 Oxygen, ppm 300-1000 H,0, ppm 300-1000 deposit in the steam generating test unit, resulting in loss of steam generator efficiency, or will be carried through the steam generator unit and deposited in the expansion valve area. Failure to maintain high-purity - water in the STTS can result in faulty heat transfer data and costly outages for chemical cleaning or for repairing - equipment which has been damaged by erosion and corrosion, An externally regenerated ammonium hydrox:de . (AMMONEX) mixed-bed demineralizer system has been selected for the condensate purification system because of low operating and capital cost and because these beds do not deplete the ammonia which is added to the ~ condensate for pH control. Heat balances for three test sections wzth different feedwater-steam conditions at 3.0 Mw rated load were made to establish criteria for the steam-feedwater system components. The first system heat balance used - feedwater at 700°F and 3800 psia with outlet steam conditions of 1000°F and 3600 psia. These feedwater- steam conditions are typical of the ORNL MSBR reference design. The second system heat balance used feedwater at SSO°F and 4000 psia with outlet steam conditions of 1000°F and 3600 psia. These feedwater- steam conditions are typical of modern supercritical fossil-fired power plants. The third system heat balance used feedwater at 550°F and 3100 psia with outlet steam conditions of 1000°F and 2400 psia. These feedwater-steam conditions are typical of modern sub- critical fossil-fired power plants. The above feedwater- steam conditions are considered typical for use with the test sections that will be produced during the steam generator development program. Table 7.3, Feedwater Operation Requirements Cation - Maximum Limits (ppb) Sample Point - Condition pH - - (micromhos/cm) 0, N2H4 Si0, Fe - Cu Ni Na Condenser 1.0 8.8-9.3 .10 15 10 5 1 hot well ' . a - : - ’ Condensate 9.2-9.4 10 20 15 10 2 2 1 booster pump o o - ' , Deaerator 0 9,2-94 < 20 15 10 2 2 1 outlet ' ' _ ‘ ' Auxiliary : 0.3 9.2-94 , 100 psig. at 1200°F. tions were made for: ~ Fig. 7.1. Basically, the loop consists of two salt-to-air heat exchangers, a throttling valve, a flow restrictor, a venturi tube, the interconnecting piping, and the salt pump being tested. A salt storage tank is connected to the loop by a pipe containing a freeze valve, , During this period, preliminary designs and calcula- 2. Salt-to-air heat exchangers — 800 Btu/sec (total), salt temperature = 1050°F, air flow rate = 10,000 '~ cfm (total) at 150°F, and AP of 3.1 psi. 3. Temperature transients — could achieve a rate of temperature reduction of about 15°F/min or a rate of temperature rise of 25°F/min, . Pfimp characteristics — to investigate the probable head-flow relationships for the pumps to be tested. . Heat removal — comparative study of cooling the 1500-hp pump motor with plant water or an 74 air-cooled heat exchanger showed that the use of plant water was more economical. . Salt flow measurement instrumentation — to in- vestigate venturi tube charactenstlcs and select pressure transmitters. . Proof test of MSBE secondary pump in primary salt — to investigate pump relationships to provide the proof test at design speed, torque, temperature, and pressure. . Pressure profile in the salt piping — to investigate salt pressures at various locations in the loop-to - provide design criteria for the various components (throttling valve, venturi tube, pressure transmitter, piping). | . Stress analysis — to investigate stresses produced by pressure, weight, and thermal expansion using a piping flexibility analysis to determine that they are within allowable limits. The following tabulations were prepared: 1. Applicable specifiéations, standards, and other publi- cations. . Pipe line schedule — a detailed list of all loop, instrument, gas, air, and water lines. . Instrument tabulations — a detailed list of instru- ments to be used. electrical and mechanical equipment. . Instrument application diagrams. . Electrical schematic diagram. The Preliminary Systems Design Descriptions and the title I design of the facility were sent to the AEC for review and approval. Before this was completed, the facility was “deobligated” because of changed budget projections for the MSR program, and all work on the design was terminated. We plan to resubmit the facility for approval to design and build on a schedule that will be based on new schedules for the development of pumps for the MSBE. 7.3.3 ALPHA Pump The fabrication of the water-test model of the 5-to 30-gpm ALPHA pump® that is being developed for use - in small forced convection loops was nearly completed. . Equipment tabulation — a detailed list of valves and . The fabrication and assembly of the test stand were completed. Upon completion of the fabrication and assembly of the ALPHA pump, it will be installed in the test stand and the water testing will be initiated. ~ 7.4 REMOTE WELDING P.P.Holz C.M. Smith, Jr. The remote welding program'® has concentrated on parametric studies of pipe end configurations and weld inserts to find the best joint design for use in remote welding. All the welding tests were performed with the " orbital - equipment including a new and improved programmer-controller - which provided very effective automatic self-regulating control of the welding vari- | ables. A quantity of 6-in. sched 80 type 347 stainless steel pipe was available from surplus stock and was used for initial development work. Procedures were developed for welding with washer-shaped inserts and with com- mercially available Y-ring inserts.}* For some tests, the pipe wall at the joint was machined to simulate a flat washer “insert”; for others, weld metal was deposited on the pipe inner surface and then machined to simulate the washer “insert,” and for others, actual insert rings were tack-welded to one of the pipe joint ends. These joint configurations are illustrated in Fig. 7.2. The tests showed that the “buttered” insert, prepared by depositing weld metal around the pipe interior and then machining the deposit to a washer shape, is best for remote welding applications. Machin- ing the pipe walls to simulate a washer-shaped insert has " the disadvantage of thinning the pipe walls adjacent to the weld. The tack-welded inserts were too easily damaged to be suitable for remotely controlled installa- ~ tion, positioning, and alignment. The weld inserts do offer the most economical method of welding and are recommended for direct construction or shop welding. The washer-shaped insert gives best results for badly misaligned welds. It is a tribute to the orbital equip- - ment that high-quality welds were made every time, whichever type of joint design was used. Following the experimental joint welding tests with '6-in. pipe of type 347 stainless steel, we made addi- tional tests with type 304 stainless steel and Hastelloy N pipe. Hastelloy N pipe was available only in a small quantity of 5-in.-diam sched 40 pipe. Our prototype orbital carriage is designed to fit only 6- to 9-in.-diam loMSR Program Semiann, Progr. Rept Aug. 31, 1969, ORNL-4449, pp. 7982, , Pyweld: Ring Co., Inc., 7508 Kress Ave., Bell Gardens, Calif. 70° 2 //}“ . "'B ------- |&-| o~ - © 0| — . ~tel 9w o lflg 58—-'/84—8 Ol g . o ~lw ol 9| 0 | T w| o 213 2] = < -, o 3 zlel |8 3 o 1 21 2 z W © SIMULATED 6-in.-SCHED 40 PIPE n_.n TACKED "Y" INSERT 7 :7-50 - . \<(CKWELD\/ PIPE OD V"-BEVELLED PIPE JOINT DIMENSIONS ARE IN INCHES 75 ORNL-DWG 70-6744 _ "BUTTERED" WASHER INSERT N 70° Yetw PIPEOD PIPE ID PIPE OD ST 347 w="/¢ ST 304 w =332 HASTELLOY N w =3, TYPICAL PIPE JOINT TACKED WASHER INSERT ST 347 w=Y¢ ST 304 w=%32 HASTELLOY N w=33> TYPICAL PIPE JOINT Fig. 7.2. Insert-Type Joint Geometries for Orbital Pipe Welding. pipe. We therefore superimposed a 6-in.-OD sleeve over one side of the 5-in. pipe and extended the torch to reach the 5-in. pipe. Even under these extreme condi- tions, the equipment performed flawlessly. All the joint geometries worked as well for 304 stainless and Hastelloy N pipe welding as for welding type 347 - stainless steel, For type 304 stainless and Hastelloy N pipe, slight widening (approximately %, in.) of the lands of the weld joint groove was required to prevent capillary action from causing the weld metal puddle to wet the upper walls of the joint during the root fusion “weld. Metallographic analyses of the welds and discus- sions of their quality are covered in Sect. 20.1, In the tests with 6-in. pipe of types 304 and 347 stainless steel and 5-in. pipe of Hastelloy N, root pass fusion welds (without weld wire feed) were made with currents of 77 and 82 amp, at torch speeds of 2% to 3% in./min. Slightly higher current (85 to 92 amp) was necessary for the filler passes, which were generally made at the same torch speed and with feed rates of 10 to 15 in./min of 0.045-in.-diam wire. The torch was oscillated approximately %, ¢ 'in. back and forth during the root passes; oscillation width was increased for 76 Table 7.5. Experimental Welds Made with the Orbital Equipment August 20, 1969, Through December 31, 1969 - g N No.of Root No. of Joint Description” - Pass Joints Fill Passes 6-in. 347 ss5 butt joint without inserts 3 2 - 6-in, 347 ss joint type A .3 3 6-in, 347 ss joint type C 5 18 6-in. 347 ss joint type D 6 4 6-in. 347 ss joint type D, l/16 in. 3 4 mismatch - 6-in. 347 ss joint, loose Y insert 7 13 6-in, 347 ss joint type D, loose 2 5 washer _ _ §-in. 304L ss joint type D 4 14 5-in. INOR-8 joint type D 5 S 5-in. 304L ss joint type E 2 211 5-in. INOR-8 joint type E - _i _1_2 Total (through end of 1969) 41 89 2Type A joint washer is '/16 in. deeper than the l/s-in.-'.v,qum:e washer. Type C joint has a Y. o-in. land one side of the f-in.-square washer and a ¥%,-in. land on the other. Type D ~ joint has % g-in. lands and a Yg-in.-square washer. Type E joint subsequent fill passes. Current pulsing in the relatively low frequency ranges presently available within our programmer circuitry appeared to offer no improve- ment, and therefore current pulsing was omitted from the initial experiments, Table 7.5 shows the pipe joint welds that were actually made with different pipe sizes and materials and with various types of washer inserts. The washer insert for the type A joint is % ¢ in. deeper than the Yg-in.-square washer, while the type C joint washer is off-center in the groove by %, in. Joint types D and E both use Y%-in.-square washers, but D has ¥ ¢-in. lands adjacent to the washer and E has lands 3%, in. w1de The D and E joints gave the best results. Tlustrations in Sect. 20:1 show the good quality and full penetration of welds on 6-in. type 347 stainless © steel, even when the pipe ends mismatch by as much as ' 6 in., which is the full thickness of the machined land protrusions at the pipe joint ends. Being able to make a good weld with this much mismatch will reduce the alighment problems in remote maintenance. The pipe in any nuclear reactor system will be under stress from thermal effects, from gravity, and from other factors relating to installation and operation of the system. One must expect springback of the ends when a pipe is cut from maintenance. The springback we have experienced in the experimental reactors and test loops has often been as much as several inches. We know, therefore, that we must be prepared to realign has 3/32-in. lands and a Vg-in.-square washer. replacement pipe assemblies and hold them in position with not more than ¥ ¢ in. misalignment. - To maintain pipe ends in alignment for final joint preparation and welding, we are presently thinking of using a split sleeve which would be cradled over and clamped to the pipe. The split sleeve includes split roller bearings and a bearing-mounted gear drive at one end to rotate the sleeve about the pipe. Our orbital machinery would then be clamped -over this sleeve, and the programmed carriage rotation would be transferred from the carriage rollers to the sleeve drive. The sleeve could be utilized to collect chips during cutting to keep them from entering the reactor system; it would also provide a properly indexed and concentric platform for mounting internal pipe cleaning gear. The sleeve can serve as an alignment and restraining device to guide the pipe ends into proper position during reassembly. It offers “glove-box™ welding capability in that the sleeve forms a secondary containment over the pipe. A gas purge stream can be used to seal off the cutter and torch access slot. Other sleeve-type devices of different design but with similar advantages are also bemg studned “During the reporting period we were able to borrow additional Air Force orbital equipment for 1- to 3-in. pipe and for 12- to 16-in. pipe. Minor alterations were made to the large Air Force carriage so that it would accommodate the ORNL-built cutter and weld-head inserts and the modified programmer-controller cir- cuitry, An extra Air Force welding head was modified to incorporate some of the ORNL design changes and thus to provide an interchangeable spare. We also made major improvements to our original ORNL remote welding head to provide additional torch oscillation capability in the short-width ranges. We provided a single-unit disconnect for all torch lines, using the commercial Air Reduction Sales Company HP20A water-cooled TIG torch cable set, The wire feeder spool was provided with more precise bearing support, and the torch position adjustments were modified to in- crease the horizontal and vertical travel capability, Most important, however, was the modification to in- corporate a miniature Hayden dc motor atop the weld head to drive a ball screw assembly for vertical positioning of the torch. Figure 7.3 shows the motor addition, 77 Preliminary weld trials of the root passes on 12-in, sched 10 type 304 stainless steel pipe had indicated severe problems in maintaining proper arc length control to provide the steady arc voltage needed for uniform weld metal puddling and weld bead control. When the pipe was very much out of round, the mechanical spring-loaded torch positioning device could not adjust well enough to provide acceptable control. The incorporation of an automatically controlled drive . motor to move the torch up and down and thus stabilize the welding arc voltage has given vastly improved control. The arc voltage signal that was already available in the programmer is used to actuate the new drive motor and to regulate the torch-to-work spacing. . ' T — The original control system used the arc-voltage signal to control the welding wire feed rates. Close torch-to- work spacings lowered the arc voltage and thus called Fig. 7.3. Orbital Welder Welding 12-in. Pipe. The new torch drive motor is centered on the weld head insert, for a decrease in the wire feed rate; distant torch-to- work spacings gave higher arc voltages and stimulated an increase in wire feed rate. The system worked well within limited arc voltage changes but was not satis- factory for responding to large changes in arc voltage. Our newly incorporated control system uses the same arc voltage signal and the original wire feed control system within the range where its control is satis- 78 factory, but actuates the new drive motor when a predetermined and preset upper or lower arc voltage setting is reached. The motor then drives the torchto a: torch-to-work spacing that gives the designated control point arc voltage. Here the wire-feed-rate control again takes over until its control range limits are reached once more. The actuation range settings for the new torch drive control system are variable to permit numerous weld operations. under different welding conditions. Although “the old arc voltage control system was repair cdpability It is now possiblé to let the torch seek the vicinity of the pipe, start the arc, and automatlcally return to a preset arc voltage. c Having a torch which adjusts itself to maintain the proper arc gap for best welding will make it easier to use the remote welding system without the orbital carriage to work on components other than pipes. A variety of welding head suspension and guided move- ment systems can be used for remote welding much - more readily now that self-adjusting features are avall- able for regulating the arc gap. We also obtained, on loan from the Air Force, automated Orbit-Arc equipment for fusion TIG welding of thin-wall tubing in sizes from % to 1% in. in diameter. This equipment was developed by North - American Rockwell Corporation for operation with the satisfactory within its range limits, it works even better in conjunction with the new drive motor control. We have found that we can use the arc voltage to set and ‘control the torch-to-work spacing as well as to maintain this spacing within limits that offer uniformity in welding. The new system is also very useful for root - - pass fusion weldments without wire feed, as uniformity in torch spacing assures even and full penetratlon with proper bead shape. The new arc voltage control system also mcorporates automatic “touchdown weld start” and improved weld orbital system programmer which we already have. Orbit-Arc tube welding equipment is now commercially marketed by the Merrick Engineering Corporation of Nashville, Tennessce, under license from North American Rockwell. In its present form Orbit-Arc equipment cannot be mounted on tubing by remote means, though design changes could permit remote .operations. Preliminary results from weld tests to check out the equipment on 1-in. stainless tubing look very promising. The equipment will next be tried on ‘molybdenum tubing and T-111 tantalum alloy tubing. 8. MSBR Ins_trumentation and Controls S. J. Ditto 8.1 CONTROL SYSTEM ANALYSIS Simulation studies of the reference 1000-Mwi{e) MSBR on an analog computer were continued. These - simulation studies are an extension of those reported earlier.'*? The basic plant components simulated were the reactor, primary heat exchanger, and steam genera- tor., The lumped-parameter model used for the heat transfer system included ten spatial lumps in the primary heat exchanger and in the steam generator and nine spatial lumps 'in the reactor core. Two-delayed- neutron-group circulating-fuel kinetics equations were used. A provision for variable flow of the primary salt, secondary salt, and steam, with the attendant variations in film heat transfer coefficients, was included. The investigation was concerned with the integrated plant response; it was not concerned with a safety analysis of the system, although several of the transients introduced would be of an-abnormal nature (e.g., loss J. L Andérson W. H. Sides, Jr. of flow). It was an initial probe into the response of the system initiated by such perturbations as changes in- load demand, loss of primary or secondary flow, and reactivity changes. A complete report on the methods and results of these studies is in preparation,? So that the model would have the maximum dynanuc range, the system differential equations were not linearized, and, as a result, the available quantity of. equipment required the model to be severely limited spatially to minimize the number of equations. In addition the pressure in the water side of the steam generator, as well as in the rest of the plant, and the physical properties of the salts and water were taken to be time invariant, The temperature of the feedwater to the steam generators was also held constant. YMSR Program Semiann. Progr. Rept. Feb, 28, 1969, ORNL- 4396, p. 113, 2w. H. Sides, MSBR Control Studies, ORNL-TM-24S9 (June - 2,1969). 3W. H. Sides, Control Studies of a 1000 Mw(e) MSBR, ORNL-TM-2927 (1o be issued). The plant control system investigated was one which controlled the reactor outlet temperature as a function - of plant load. Steady-state calculations showed that, by specifying the steady-state primary salt flow rate, the reactor outlet temperature, and the feedwater and steam temperatures as a function of load, the remaining steady-state system temperatures and flow rates can be determined. The primary salt flow rate was held constant at its 100% design point value, and the feedwater and steam temperatures were held constant’ at 700 and 1000°F respectively. The reactor outlet temperature was varied with the plant load (Fig. 8.1). The resulting variations of the reactor inlet temperature and the secondary salt hot and cold leg temperatures and flow rates for the analog simulation model are also shown in Fig. 8.1. The reactor outlet temperature was varied as a linear function of the plant load between 1300 and 1125°F for loads between 100 and 50% respectively. For loads below 50% the reactor outlet temperature was varied linearly from 1125 to 1000°F at no load. The break- point at 50% load was requlred to maintain the reactor inlet temperature above 1000°F at low loads. Figure 8.1 shows that the steady-state secondary salt flow rate decreased with a decreasing load at a rate roughly proportional to load. The secondary salt AT between the hot and cold legs was, therefore, approximately constant. Hence the cold leg temperature was required to decrease from its design point value of 850°F at full load to below 725°F at loads below about 30%. It - dropped below its minimum acceptable value of 800°F ~ at approximately 70% load. Steady-state calculations 79 “for this model indicate that, by decreasing the reactor outlet temperature more rapidly with decreasing load in the range near 100% load, the secondary salt cold leg temperature decreased less rapidly with load and lowered the power level at which it crossed the 800°F minimum. Since it may be undesirable to decrease the reactor outlet temperature more. rapidly with decreasing load than is shown in Fig. 8.1, other methods may be 80 'ORNL-DWG T70-6745 1400 1300 | PRIMARY SALT REACTOR OUTLET TE wPERAle-/ { & 1200 : / 1.0 % 3 / ._____——— r 2 1100 08 2 & |- SECONDARY SALT HOT V7 3 & LEG TEMPERATURE_ ] —] & 1w ] ‘/ - - 1000 PRIMARY SALT REACTOR 06 }._1' INLET TEMPERATURE ' 2 o T SECONDARY SALT FLOW x > 900 — | | o.4§ 800 ™ f 0.2 % 1 | SECONDARY SALT COLD . " LEG TEMPERATURE 700 : ' L 1 L ‘ 0 o X 02 03 04 O05 06 - 07 08 .09 1.0 LOAD DEMAND Fig. 8.1. Steady-State Temperatures and Flows as Functions of Load. required to maintain the steady-state cold leg tempera- ture above its 800°F minimum at the lower power levels. Such methods are: (1) increasing the steam temperature above its 1000°F design point as the load decreases, with subsequent attemperation of the steam with injected feedwater; (2) increasing the feedwater temperature above its 700°F design point as the load decreases; and (3) reducing the number of steam generators in use as the load decreases. If valves are considered for use in the salt systems, other methods ‘may prove feasible as well. Further “investigations of steady-state temperatures and flows should be carried out, including studies of off-design conditions in the steam generator. Insuffi- cient machine time was available to adjust the present analog model to include a variable steam or feedwater temperature with load, and insufficient equipment was available to include more than one steam generator. ‘The objective of the load control system used in this . study was to maintain the temperature of the steam delivered to the turbines at a design value of 1000°F To accomplish plant load control in this simulation, an external plant load demand signal was used as input to the plant control system (Fig. 8.2). The steam flow rate was made to follow the demand w1th a 5-sec time constant. Steam . temperature control was accomplished by varying the secondary salt flow rate. This method was chosen because of the relatively tight coupling which existed between steam temperature and secondary salt flow rate. The measured steam temperature was com- pared with its set point of 1000°F, and any error caused the secondary salt flow rate to change in the appropriate direction at a rate proportional to the error if the error was 2°F or less. If the error was greater than 2°F, the rate of change of the secondary salt flow rate “was lifited to .its rate of change for a 2°F error, which was approximately 11%j/min. To accomplish reactor putlet temperature control, an external plant load demand signial was used to obtain a " reactor outlet temperature set point. The outlet temper- during all steady-state conditions and within a narrow band around this value during plant transients. The control system used in this simulation is shown in Fig. 8.2. It consisted of a steam temperature controller and a reactor outlet temperature controller similar to that used successfully in the MSRE.* ~ %7, R. Tallackson, MSRE Design and Operations Report, Part ITA: Nuclear and Process Instmmentanon, ORNL-TM-729 (February 1968). ature set point vs load demand was the same as that for the steady-state reactor outlet temperature vs load in Fig. 8.1. The measured value of the reactor inlet temperature was subtracted from the outlet tempera- ture set point, and since the primary salt flow rate was constant, a reactor (heat) power set point was generated by multiplying this AT by a proportionality constant. The reactor power set point was a function of inlet temperature during a transient and thus a function of dynamic load. The measured value of reactor power (from neutron flux) was compared with the reactor ORNL-DWG 70-6746 PpEMAND ROD DRIVE SERVO fe PRIMARY : HEAT Fo ) EXCHANGER ri + . & a7y (1000°F) 75t , : STEAM s GENERATOR REACTOR | 7 _ _r:_D Fig. 8.2. Plant and Control System Simulation Model. ~ power set point, and any error was fed to the control rod servo for appropriate reactivity adjustment, Under normal conditions the control ,rod servo added or removed reactivity at a rate proportional to the reactor power error if the error was 1% or less. If the error was greater than 1%, the addition or removal rate was limited to the rate for a 1% error, which was about 0.01% 8k/k per second. The maximum magnitude of reactivity which the simulation allowed was 1% 8k/k. During a reduction in load demand the control system responded as follows: A reduction in load demand reduced the steam flow rate and the reactor outlet temperature set point. The decreasing steam flow transferred less heat out of the steam generator, and the steam temperature rose. The resulting steam tempera- ture error decreased the secondary salt flow rate to transfer less heat into the steam generator. Simultane- ously, the decrease in the reactor outlet temperature set point- decreased the power demand set point. The resulting power demand error caused the control rod servo to add negative reactivity to bring the reactor . power down, The system reached the new steady state ~ when the heat transfer system was in equilibrium; the steam temperature was 1000°F, and the reactor outlet . temperature was at its set point. Various load demand transients were investigated, including changes in demand from 100% load to 90% and from 100 to 50% at 10%/minute and from 100 to 50% at 5%/min. The results of these studies are summarized in Table 8.1. Listed in the table are the Table 8.1. Maximum Magnitude and Rate of Change of System Temperature, Flow Rate, and Reactivity During Change in Load Demand 100 to 90% 100 to 50% 100 to 50% Load Demand Change at 10%/min at 10%/min at 5%/min " Reactor outlet Temperature, °F —35 -175 175 Rate of change, °F/sec —0.55 -0.56 -0.27 Reactor inlet ‘ Temperature, °F —10 -50 -50 Rate of change, °F/sec —0.10 -0.15 -0,09 Secondary salt hotleg Temperature, °F -8.8 - ~60 ~60 Rate of change, °Ff/sec -0.10 —~0.22 - -0.17 ~ Secondary salt cold leg - | o Temperature, °F -18 —80 —80 Rate of change, °F/sec —0.31 -0.36 -0.18 . Steam o Temperature, °F . 7 10 2 " Rate of change, °F/sec —0.19 ~-1.0 <0.1 Secondary salt flow rate 7 Magnitude, % ~14 -8 —56 Rate of change, %/min 11 -11 -9 Control reactivity Magnitude, % &k/k -0.013 -0.06 -0.06 Rate of change, %fsec 0.0002 —0.0002 —0.0001 values of the maximum magnitude of the deviation from the initial steady state of a system variable and the maximum rate of change of that variable. The values listed are the maxima encountered at any time during a transient; they are not necessarily initial rates of change or differences in steady-state magnitudes. The results of 2 load demand change from 100% to 50% at 5%/min are shown in' Fig. 8.3, The steam temperature was controlled to within 2°F of its design point, and the reactor outlet temperature closely followed its set point. The load demand transient results indicate a stable 82 well-behaved system. Normal load changes at a rate of 5%fmin or less can probably be controlled by a system similar to that used on the MSRE with the addition of the steam temperature controller. " QRNL-DWG 70-6747 - T T {00 1 ! LOAD DEMAND ——-I— 100% TO 50% AT 5%,/ min % 50 _ - 1 | 100 REACTOR POWER l | | | 100 SECONDARY SALT FLOW RATE — 7% 50 P 0 0425 ROD REACTIVITY % 34/k © -0.42% 1050 : TEMPERATURE °F 4000 950 1550 REACTOR QUTLET TEMPERATURE °F 4300 - 1050 1150 REACTOR INLET TEMPERATURE *F 1050 950 . . o 200 400 600 800 000 1200 1400 * TIME (sec) Fig, 8.3. Load Demand Change from 100 to 50% at 5%/min. Transients initiated by positive and negative reactivity excursions were investigated. The excursions included a negative step in reactivity of —0.2% 8k/k and positive steps of 0.15% 8k/k with and without the control rod servo operative. The results of these studies are shown in Table 8.2. The positive reactivity excursions were begun with the plant at an initial power level of 25%in order to obtain the maximum positive power excursion for this simulation, The maximum allowable power was 160% of design power. The results indicate that certain reactivity transients may require additional control if undesirably low temperatures of the salts are to be avoided. For example, if an insertion of negative reactivity in the core reduces the reactor power, then the load must be reduced at a rate sufficient to avoid overcooling the salts. — Table 8.2, Maximurm Magnitude and Rate of Change of System Temperature, Flow Rate, and Reactivity Resulting from Step Changes in Reactivity Reactivity Step +0.15% + fi(g;ll;?% from 25% -0.2% Power Power Level - Level with No Control . Reactivity Reéctorr outlet o | Temperature, “F -100 100 592 Rate of change, °F/sec —36 50 63 Reactor inlet .. _ Temperature, °F —-40 56 580 Rate of change, °Ffsec —6.9 14 19 Secondary salt hot leg _ . Temperature, °F -24 65 - >350 Rate of change, °Flsec -1.1 9.7 13 Secondary salt cold leg Temperature, °F —4 -15 . —40 Rate of change, °F/sec 7.1 048 0.67 Steam N N Temperature, °F -32 . 28 195 Rate of change, °F/sec 3.4 2.2 . 5.0 . Secondary salt flow rate _ Magnitude, % 10 -6.5 12 Rate of change, %/min - 11 11 C11 Control reactivity Magnitude, % &k/k 022 028 0 Rate of change, %/sec 0.01 001 - 0 Figure 8.4 shows the results of a positive step in reactivity of 0.15% from an initial power of 25%. The reactor power increased rapidly to about 144% while the control rod added negative reactivity at its maxi- 83 mum rate. The sudden increase in the reactor power caused a rapid increase in the reactor outlet tempera- ture. An increase in the reactor inlet temperature from its initial value followed. When the inlet temperature returned to 1040°F, the reactor power had decreased to 8.5%. Since the reactor outlet temperature set point was constant during this transient at 1063°F, the onm.- DWG 70— 6748 300 ~ 200 % REACTOR POWER - 100 o 100 CONDARY SALT FLOW RATE % 50 - 0.5 : INPUT REACTIVITY -0.5 1 N 1 0.5 T T ML CIONTROL REACTIVITY % -0.5 | 1050 1\ STEAM TEMPERATURE. °F 1000 950 1300 T T T T 1 REACTOR OUTLET TEMPERATURE °F 1050 800 1100 T T T T T REACTOR INLET TEMPERATURE — °F 1000 900 1200 1100 1000 900 200 800 700 €00 SECONDARY T TEMPERATURE - LEG °F SECONDARY SALT COLD - LEG TEMPERATURE oF 50 100 450 200 250 300 TIME AFTER CHANGE BEGUN (sec) . 0 Fig. 8.4. Input Reactivity Step of 0.15% from 25% Power Level. : ' reactor power set point at this time was 9%, and thus the control system began to add positive reactivity to the system to increase the reactor power. As the inlet temperature approached 1000°F, the power set point approached the initial level of 25%. The temporary increase in the reactor inlet temperature beginning at approximately 70 sec was due to the decrease in the secondary salt flow rate, which was attempting to control the delayed response in the steam temperature. The increasing reactor temperatures produced an in- crease in the steam temperature, which was delayed by about 65 sec because of the transit time of the secondary salt between the heat exchangers at the initial 22% flow rate. The steam temperature rose to about 1028°F before the decreasing secondary salt flow rate returned it to 1000°F, The relatively long second- ary salt loop transit time reduced the capability of the secondary salt flow rate to control the steam tempera- ture, and several oscillations were allowed to occur before the system returned to normal steady-state conditions at 25% power level. The total excess energy added to the system by the reactor power “pulse” from the initial power rise to the point at which the power first returned to the 25% level was approximately 13,000 Mwsec. Several transient cases were studied involving primary and secondary-salt flow rates. Some results of these studies are summarized in Table 8.3. The simultaneous coastdown of all four primary pumps to an arbitrary - minimum flow of 10% was investigated. It was assumed that some device such as a battery-powered pony motor on the primary pumps would maintain some minimum pumping capacity in the primary loop upon loss of power to the main primary pump motors. The primary salt flow rate was thus reduced in all parts of the primary loop to 10% of full flow at a rate of 10%fsec. The results of this transient are shown in Fig. 8.5, The proportionality constant between desired reactor AT and reactor power set point was .decreased with the - primary salt flow rate, which produced the reactor power error signal. Negative reactivity was thus intro- duced at the maximum rate, and the reactor power - decreased in about 25 sec to about 12%. The maximum amount of control reactivity required was about —0.21% &k/k. The reactor outlet temperature rose at first to about 1400°F in 15 sec, then decreased to about 1340°F, The inlet temperature fell below 1000°F in 15 sec, The loss of primary flow while maintaining full heat extraction from the. steam generators caused the secondary salt hot leg temperature to fall sharply. The cold leg temperature also decreased. The decreasing secondary salt temperatures caused a severe reduction ORNL-DWG 70-6749 - REACTOR POWER - PRIMARY SALT FLOW RATE %% ) ‘ 100 % : | SECONDARY SALT FLOW RATE 0 Lt -1 - 0.25 . - CONTROL ROD REACTIVITY % 3k/k 0 : -0.25 1550 °F $300 : REACTOR OUTLET TEMPERATURE 1050 : : 1300 : - . REACTOR INLET TEMPERATURE °F 1050 800 - 0O 30 60 80. 120 150 180 210 240 270 300 330 TIME (sec) Fig. 8.5. Loss of Primary Flow to 10% at 10%/sec. 84 in steam temperature. The secondary salt flow rate increased to its limit of 110% in an attempt to maintain the steam temperature at 1000°, but with little success. Due to the assumptions concerning the variations in - steam properties made in formulating the model of the steam generator used in this simulation, the useful range of the steam generator model was greatly limited. The model, therefore, simulated only small variations in steam temperature near 1000°F. Loss of primary or secondary flow to 10%, however, effectively decoupled the reactor from the steam system, and large-magnitude changes in the steam generator had a greatly reduced effect on the reactor system. Only the du'ectlon of ‘these changes was important, A second case involved the same loss of primary flow as described in the first case but with a reduction in load demand from 100 to 20% at a rate of 20%/sec. This rate was determined by the assumed maximum rate at which the turbine steam interceptor valves could close. It was assumed that an auxiliary heat rejection system would be capable of disposing of 20% of the full plant power. The reduction of load was initiated 5 sec after the initiation of the primary pump coastdown in order to simulate some delay time for the system to sense and evaluate the incident. The proportionality constant between the desired reactor AT and reactor power set point again was reduced with the reduction in Table 8.3, Maximum Magnitude and Réte of Change of System Temperature, Flow Rate, and Reactivity During Flow Transients Flow Rate Change - Loss of Loss of Primary ' | Loss of Seéofigss Ofl.’"low Primary Flow and » Secondary an dai.yda d a il a Flm\a'.r Load Reduction Flow Reduction® b Reactor outlet Temperature, F - 100 - -250 -30 -320 Rate of change, °F/sec 13 13 -4.4 17 Reactor inlet c Temperature, °F ~200 ~220 210 135 Rate of change, °F/sec -8.3 -8.8 20 20 Secondéry salt flow rate Magnitude, % 10 10 -9 - - =90 Rate of change, %/min 11 11 —-600 -600 Control reactivity ; N Magnitude, % sk/k- -0.21 . —0.46. ~0.063 -1.0 Rate of changc, %|sec -0.01 - ~0.01 ~0.01 -0.01 : "Flow rate decreased to 10% at a rate of 10%/ sec, bLoad demand reduced to 20% at 20%/sec initiated § sec after uutlatlon of flow reduction. . 85 flow rate. The reactor power reached 12% in about 20 sec. The reactor outlet temperature again rose to about 1400°F, then decreased to about 1200°F at 60 sec, and continued to decrease at a rate of about 0.3°F/sec. The reactor inlet temperature transient was much like that in the previous case, as were the transients in the secondary salt hot and cold leg temperatures. | The steam temperature initially rose in this transient since the fast load reduction dominated the response in .the steam generator when it occurred 5 sec after the primary flow coastdown. However, this did not prevent large sudden decreases in the secondary salt tempera- tures. Some additional corrective action may be re- quired to prevent such decreases in temperature, such as a reduction in secondary salt flow rate when primary flow is lost. ‘ ~ The results of these pnmary flow transients indicate a need for further investigation of the conditions existing in the secondary salt loops and steam generators - following a loss of primary flow transient. Attention must be paid to resulting magnitudes and rates of change of temperature in this part of the system. The model of the steam generator used in this simulation was not adequate for such studies due to the approxi- mations made. The result of the simultaneous reduction of the secondary salt flow rate in all four secondary loops to a level of 10% of full flow (the assumed level of auxiliary pumping power) at a rate of 10%/sec is shown in Fig. 8.6. The load demand was maintained constant at 100%. As in the case of loss of all primary flow to 10%, the loss of secondary flow decoupled the reactor from the steam system. For the case of constant load demand, the reactor inlet temperature initially rose about 200°F in about 60 sec. Since the load demand remained at 100%, the reactor outlet temperature set point remained at 1300°F. The rising inlet temperature thus decreased the reactor power set point, and negative reactivity was added to reduce the reactor power. The outlet temperature control system maintained the outlet temperature at 1300°F with a maximum varia- tion of 30°F. The reduction in secondary salt flow rate wuh, constant load demand on the steam generators caused an increase in the difference between the secondary salt hot and cold leg temperatures. The hot leg temperature increased, and the cold leg temperature decreased. The cold leg temperature approached the freezing point. ~ With the loss of secondary salt flow, there was no steam temperature control. When the load demand was decreased rapidly (20%/sec) to 20% starting 5 sec after the start of the ORNL-DWG T0-6750 REACTOR POWER "~ % DARY SALT FLOW. RATE Yo 0425 : ROD REACTIVITY % Bk/k 0 ~0.425 £550 . : QUTLET TEMPERATURE °F 1300 1050 1300 *F 1050 REACTOR INLET TEMPERATURE 800 O 30 60 90 120 150 180 210 240 270 300 330 TIME (sec) Fig. 8.6. Loss of Secondary Flow to 10% at 10%/sec. loss of secondary salt flow, the initial parts of the transients in system temperatures were the ~same. However, when the load demand was decreased, the reactor outlet temperature set point was decreased to 1050°F for 20% load. Therefore, the reactor outlet temperature controller began to bring the outlet tem- perature down to 1050°F. The initial rise in reactor inlet temperature caused a decrease in the reactor power set point as before, and negative control reac- tivity was inserted to bring the power down. The power decreased to about 10% in-30 sec. The secondary salt hot leg temperature initially tended to rise and the cold leg to fall as before, but now the decreasing reactor outlet temperature decreased the hot leg temperature after its initial increase. The cold leg temperature again ‘approached its freezing point. The loss of salt flow in the primary or secondary salt loops decoupled the reactor system from the steam generating system. The reactor outlet temperature control system was able to control the reactor outlet temperature following the loss of primary or secondary flow with or without a subsequent reduction in load demand. If the load demand was not reduced, the control system maintained the reactor outlet tempera- ture within 100°F of its design point of 1300°F. When a reduction in load demand followed the loss of flow, the controller brought the reactor outlet temperature down in accordance with the accompanying reduction in its set point (1050°F at 20% load). The reactor inlet temperature, however, decreased well below the freez- ~ing point of the primary salt upon loss of the primary flow, due to the increased transit time of the salt in the primary heat exchanger, whether or not the load demand was reduced 5 sec after loss of flow. Therefore, upon the loss of primary flow, steps must be taken to prevent a reduction in the reactor inlet temperature. Decreasing the secondary salt flow through the primary 86 must be decreased to prevent a. low reactor inlet ~temperature, and the load must be reduced éufficiently fast to prevent low secondary salt cold leg tempera- tures. Upon loss of secondary salt flow to 10%, the reactor inlet temperature tended to increase and remain above 1050°F when the load demand was not reduced (ie., - constant outlet temperature set point). When the load heat exchangers to transfer out less heat would proba- bly be the most effective way to accomplish this, The secondary salt temperatures also decreased upon loss of primary flow. To prevent an undesirably low temperature of the cold leg, the load must be reduced sufficiently fast. Decreasing the secondary salt flow rate to control reactor inlet temperature, as discussed above, aggravates this situation, since the transit time of the secondary salt through the steam generator is increased, further lowering the secondary salt temperature, Upon loss of primary flow, then, the secondary salt flow rate demand (and outlet temperature set- point) was re- duced, the inlet temperature remained above 960°F. Some additional control action may be required to maintain the inlet temperature above 1000 F upon loss of secondary flow. i Loss of secondary salt flow rate produced undesirable ~ decreases in the secondary salt cold leg temperatures. Therefore, as in the case of loss of prifnary flow, the- load must be reduced at a rate sufficiently fast to prevent freezing of the secondary salt when loss of secondary salt flow rate occurs. : 9. Heat and Mass T_ra'nsfer and 'Ihetmoph)fsical Properties H.W.Hoffman I.1J.Keyes, Jr. - 9.1 HEAT TRANSFER diameters long) for values of the Reynolds modulus as ' ' high as 5000. LW COOke - Heat Flux Effects — In Fig. 9.1 local heat transfer In the previous semiannual report, ! results of heat coeffic1ents relative to the heat transfer coefficient at transfer experiments employing a proposed MSBR fuel the test-section outlet are compared for heat fluxes of ‘salt (LiF-BeF,-ThF4-UF,; 67.520-120.5 mole %) 0.74 X 10° and 255 X 10° Btu hr™' ft™ at flowing in a horizontal tube 0.25 in. OD, 0.035 in. wall, = comparable Reynolds moduli in the low-transitional 245 in. long, were summarized; correlations were range. Whereas at the low heat flux (upper curve) the presented for values of the Reynolds modulus N less normalized heat transfer coefficient approaches a nearly than 1000 and greater than about 3500. It was pointed constant-value of unity (indicative of developed flow) out that, in the upperdaminar and lower-transitional about 14 in. from the test section inlet, at the high heat - flow ranges (1000 < Ngp, < 3500 approximately), flux (lower curve) the normalized coefficient does not irregular axial temperature profiles precluded analysis approach unity until very close to the outlet. These of the data to obtain a valid heat transfer coefficient for results suggest that the effect of heating the flowing salt developed flow. We also observed that the magnitude of is to retard the development of fluid-dynamical equi- the heat flux affected the temperature profile signifi- librium through the effect of temperature on fluid cantly with the result that, at sufficiently high flux, viscosity. It is known from hydrodynamic stability turbulent flow did not develop in the test section (130 theory, for example, that heat transfer from a solid ' ‘ : interface to a fluid whose viscosity decreases with : increasing - temperature has the effect of delaying 'MSR - Program Semiann. p,og,. Rept. Aug. 31, 1969, tramsition. This is thought to be the applicable mecha- ORNL-4449, pp, 8589, _ , nism in these experiments. It would also be expected ORNI-DWG 70-1238A HEAT FLUX=0.74x{0% Btu/hr ft2 NUMBER = 3560 AT FLUX=2.55x40% Btu/hr $12 NUMBER=3760 LOCAL HEAT TRANSFER COEFFICIENT 0 2 4 6 8 10 12 14 16 18 20 22 249 " DISTANCE"FROM TEST .SECTION INLET (in.} Fig. 9.1. Variation in Ratio of Local to Outlet Heat Transfer Coefficlent w:th Distance from the Inlet for Two Values of Heat Flux (LiF-BeF;-ThF,-UF,; 67.5-20-12-0.5 Mole %). 87 A that the effect of heat flux would be greatest for fluids 88 having high Prandtl moduli because the thermal bound- ary layer is relatively thinner, in respect to the hydrodynamic boundary layer, than for low Np_ fluids. The Prandtl modulus for the MSBR fuel salt is about 13 at the conditions for which the results depicted in Fig. 9.1 were obtained, probably high enough to show a significant effect of heat flux. Additional experiments will be carried out with the objective of establishing quantitatively the influence of heat flux in the transi- tional flow range, . Natural Convection Effects. — The posmblhty that ‘natural convection may have influenced the heat transfer results at low values of the Reynolds modulus in a horizontal tube led to an attempt to obtain coefficients for upflow and downflow in a vertically oriented test section. Accordingly, the test element . employed in the horizontal flow studies was repo- sitioned vertically; it was planned to use the test section in this orientation for studies of the effect of buoyancy forces and later also for gas dispersion studies. Eight preoperational check-out runs were made before 10 5 < o . P ® HORIZONTAL FLOW Z 2 A VERTICAL DOWNFLOW @ A VERTICAL UPFLOW x 2 T w 0 f wn < 2 - 5 - < w I 2 5 e Wor)”® ()™ O 2 10° - 102 2 5 10° 2 Nge + REYNOLDS MODULUS numerous salt leaks necessitated shut down of the loop. These leaks appeared to be associated with thermal stresses arising from repeated freezing and melting of the salt and were likely intensified by piping rigidity necessitated by the vertical orientation. : ‘ Results of the eight heat transfer runs (four in upflow and four in downflow) are compared with the earlier results for horizontal flow in Fig. 9.2. While the measured heat transfer function Ny o/ Ve )1 B3 (f1g)014 in downflow agrees well with that for horizontal flow, the function for upflow is ‘about 12% lower. In the range of Reynolds modulus over which the vertical flow results were obtained (6000 < Ng, < 10,000), well-established criteria _indicate that no significant effect of natural convection should have been observed. The observed difference between the upflow and downflow measurements is believed to result from errors associated with fluid temperature measurement. Discrepancies in the heat balances.(ratio of sensible heat gained by the fluid plus heat losses to heat generated electrically) were such as ‘to account for- the differences in the heat transfer ORNL-DWG 70—67514 5 w0 2 5 ©0° Fig. 9.2.. Comparisons of Previous Heat Transfer Measurements for Horizontal Flow with the Results for Vertical Flow for a Proposed MSBR Fuel Salt (LiF-BeF,-ThF4-UF, ; 67.5-20-12-0.5 Mole %). The curves are published empirical correlations. function. The sensible heat gained by the fluid was used directly in determining the Nusselt modulus Ny, for all l'uns. . . ' . Cover-Gas Effects. — Thus far, all tests have been run with argon cover gas. To determine whether 2 more soluble gas might have an effect upon. the heat transfer coefficient, measurement will be made using helium as cover gas. Helium is an order of magnitude more soluble in molten salts than argon, and hence more gas evolution would be expected from inlet to outlet. The test section has been reoriented horizontally, and 40 new runs have been made with argon to establish that the system is functioning properly and that the results are consistent with those obtained earlier using the same test section in the horizontal orientation. These runs have covered a Reynolds modulus range from 8000 to 12,000 and a heat flux range from 100,000 to 200,000 Btu hr™* ft™2 at an average fluid temperature of 1330°F and corresponding Prandtl modulus of about 6.0. The results agree on the average with the earlier results to within 3%. | | o Upon completion of studies with helium as cover gas, we plan to investigate the effect of injecting helium bubbles into the flowing salt directly upstream from the test section. A water-flow mockup is being used to evaluate possible injection techniques. 9.2 THERMOPHYSICAL PROPERTIES J. W, Cooke ' The variablegap apparatus used to measure the thermal conductivity of molten salts?> has been im- proved by the addition of a heat meter to permit comparison of the heat flowing out of the specimen with the heat flowing into the specimen. The latter heat flow is obtained from the current flow and voltage drop through the main specimen heater. In Fig. 9.3 the heat - meter is seen to consist of a 1.25-n.-diam by 4-in.dong bar of type 347 stainless steel surrounded by two concentric cylinders of the same material and welded to a water-cooled stainless steel heat sink. A %,4n. gap ‘between the cylinders reduces radial heat flow. The heat flow out of the specimen is determined from the measured temperature gradient and the known conductivity of the stainless steel bar. A movable thermocouple probe in a 0,070-in.-diam hole along the center line of the heat meter is used to measure the axial ‘temperature gradient. Thermocouple conduction 2MSR Program Semiann, Progr. Rept. Aug 31, 1968, ORNL-4344, p, 100. ‘container walls is eliminated.? ORNL-OWG 64-8427RAR DIAL INDICATOR VARIABLE-GAP ADJUSTMENT EECOLL TIPS DI EI T A BARALLRLRRAR AR RERRR GRS / -GUARD HEATER CALRQOD HEATERS ] 2% ol s VARIABLE GAP HEAT METER L Ll SINK COOLER S & 7 AN e THERMOCOUPLES Al Ak onddlode Al O AR Rt L L N N s e L AN N S | Fig. 9.3. Schematic _Crosé Section of Therma! Conductivity &u. error is minimized by using a special probe with a 0.5-in. extension beyond the junction having the same conductivity as the probe itself. ' - Errors associated with radial heat flow in the heat ~ meter will be reduced by (1) adjusting the furnace heat - . rate to reduce the radial temperature gradient that is measured by a thermocouple probe in the adjacent concentric cylinder and (2) calculating the radial heat losses by means of a computer analysis of the axial ‘temperature gradient along the center line in terms of a two-dimensional heat-flow model and the known ‘properties of type 347 stainless steel. The new appa- ratus will particularly improve the accuracy of determi- “nation of thermal conductivities of low-conductivity specimens. By using the heat meter rather than .the input ‘electrical measurement to determine heat flow through the specimen, much of the uncertainty arising from shorting of heat around the specimen by the SMSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 89-91. The modified apparatus will be used to study conduc- tivities in the LiF-BeF,-ThF, system. Measurements will begin with purified BeF, (in the solid and liquid states) and end with purified LiF. Six binary composi- tions (Li-Be) and several ternary mixes (Li-Be-Th) will be examined. From these measurements we hope to develop means for estimating the conductmty of other molten fluoride salt mixtures of these constituents with the same certainty but with less effort and cost than ‘would be required for detailed experimental measurements. ‘ | S 9.3 MASS TRANSFER TO CIRCULATING BUBBLES T S Kress The ‘apparatus was operated successfully after some modification necessitated by results of shakedown tests, and the instrumentation was calibrated. Preliminary tests were made to determine the characteristics of the helium bubbles produced by the bubble generator,* including effects of liquid flow rate, gas flow rate, and addition of a surface-active agent. Satisfactory bubbles were produced within the desired diameter range, but independent control of bubble size was not obtained. The size depended strongly on liquid flow rate, moder- ately on probe position, and weakly on void fraction. 4MSR Program Semmnn. Progr Rept. Aug. 31, 1969, 0RNL-4449 p.93. 9 Sample photographs of the bubbles produced are shown ~ in Fig. 94. A typical photograph was analyzed manu- ally to establish the bubble-diameter distribution shown in Fig. 9.5. These data, replotted on a log-normal _ probability graph, are presented in Fig. 9.6, which indicates that a log-normal function does indeed satis- factorily describe the distribution. ‘The preliminary tests also served to estabhsh the operating limits of the system. It was determined, for example, that experiments can be carried out within the following ranges of the primary variables: ) Variable : S From To Reynolds modulus 8X 10° 10° Schmidt modulus 400 4000 Void fraction, Qg/Q 7% 0.1 " 0.5 Bubble dlameter in, 0.01 0.05 Other variables and ranges of interest include test section diameters (2 and 1.5 in.), test section orienta- tion (horizontal and vertical), and surfactant concen- tration (zero or “high”). Since the magnitudes of the prime variables are not independent of the magnitudes of the other vanables, a fractional factonal expenment ‘ will suffice, Experiments have been completed with and without addition of a surfactant (butyl alcohol) for a Schmidt modulus of 400 (water with no glycerol added) over the desired ranges of Reynolds modulus, void fraction, and PHOTO 99443 Fig. 9.4. Typical Sami:le of Bubbles Produced by the Bubble Generator IMustrating Effect of Liquid Flow on Bubble Size. Probe -fully retracted to produce maximum bubble size. Fluid is 70% water plus 30% glycerol (a) Liquid flow, 40 gpm; void fractlon, 0.5%. () anuxd flow, 75 gpm, void fraction, 0.2%. 91 ORNL-DWG 70-67s2 bubble diameter. The results have not been analyzed; 140 however, in each experiment the anticipated linear lfi ' transient curve of log concentration vs time appears to have been obtained, as shown in Fig. 9.7. A curve f20 — . similar to Fig. 9.7 should be obtained for each | ' combination of the system variables. The slopes of the transient curves are related to the e 9 d & O L o = D . ® 100 SMOO‘TH CURVEI FITTED TO DATA - test-section-average mass transfer coefficient and to the ’é MEASURED HISTOGRAM | loop transit time 7; for run i The overall loop transit @ time is seen to be the sum of the transit times for each 3 80 - section 7, i % - _ - & 60 = ‘:': Tik ~ ELkAk/Qi— VL/Q;‘, @ ] in which © ) . § 40 \ - 7; = loop transit time for ith run, & ': \ : ' Tj = transit time through section k for ith run, § 20 |-f _ L;, = effective length of section k, fi 1 . -\ / - Ay, = effective cross-sectional area of section k, . : . - o ! Q; = volumetric flow rate of the /th run, and 2> — ‘l‘a’ 0 0.01 0.02 003 0.04 005 V; =the total loop volume assuming there are no | d,BUBBLE DIAMETER (in.) regions of zero (or little) flow. Fig. 9.5. Typical Bubble Diameter Distribution Produced by The loop volume was measured during the pre- Bubble Generator. Probe partially withdrawn, water flow =100 liminary tests and .found to-be 2.52 ft*, with an gpm, ratio of volumetric gas to liquid flows, @ /Q; =0.5%. estimated 0.1 ft* being “low-flow™ volume. PERCENT LESS THAN ¢ 99,9 99.8 99.5 99 98 95 90 80 70 60 50 40 30 20 10 5 2 { 05 0.2 | 04 103 2 5 92 0-2 d, DIAMETER (in.) ORNL—DWG 70— 6753 10! Fig. 9.6. Replot of Data from Fig. 9.5 Showing (l:omp'axison with a Log-Normal Distribution. Fig. 9.7. Typical Experimental Curve of Log-Concentration Ratio vs Time. Water flow = 75 gpm, void fraction of bubbles = 0.4%. o : 0.9 < = 0.8 o - 207 o o Q o o > OXYGEN CONCENTRATION/OXYGEN CONCENT o w o n " ORNL—DWG 70—6754 N N N N N\ . N \\ N N\ \I 7, TIME {min) | Part 3. Ch‘emis/try The chemical research and development efforts de- scribed below provide extensive support to the Molten- Salt Reactor Experiment (MSRE) and to the develop- ment of advanced molten-salt reactor systems. : A substantial fraction of these efforts was devoted to investigations of the chemistry of the MSRE fuel salt and off-gas streams and the transport, distribution, and chemistry of fission products in these streams, Studies of the relation of redox potential and distribution of ~ fission products within the fuel containment system - have continued. Investigations of fission product be- havior have been continued with specimens removed from the MSRE fuel circuit, with the MSRE off-gas sampler-analyzer, with “synthetic” fuel mixtures, and by investigation of the chemistry of molybdenum, niobium, and ruthenium in molten fluoride mixtures. A broad program of fundamental investigation into the physical chemistry of molten salt systems was maintained; from it are derived the basic data for reactor and chemical reprocessing design. Within the scope of these efforts is included research in solution thermodynamics and phase equilibria, crystal chem- istry, electrochemistry, spectroscopy, transport proc- - esses, and theoretical aspects of molten-salt chemistry. Effective separation of rare-earth fission products from fluoride salt streams which contain. thorium fluoride is the keystone to development of semicon- | tinuous reprocessing fiethods of single-fluid molten-salt ‘reactors. Efforts to develop chemical separations proc- esses for this application continue to emphasize meth- ods which employ selective reduction and extraction into molten bismuth containing elther lxthmm or thorium as the reducing agent. As a consequence of recent development of a fluoride-metal-chloride transfer process and laboratory-scale demonstration of its effi- cacy for separation of the lanthanide fluorides from MSBR fuel, part of the present effort has been devoted to quantitative verification of the chemical equilibria of importance to newly evolving modifications of the reductive extraction process. The principal emphasis of analytical chemical de- velopment programs has been placed on methods for use in semiautomated operational control of molten-salt breeder reactors, for example, the development of indine analytical methods for the analysis of MSR fuels, for reprocessing streams, and for gas streams. These methods include electrochemical and spectrophoto- metric means for determination of the concentration of U?* and other ionic species in fueéls and coolants, and adaptation of small online computers to electro- analytical methods. Parallel efforts have been devoted ‘to the development of analytical methods related to assay and control of the concentration of water, oxides, " and tritium in fluoroborate coolants. 10. Chemistry of the MSRE 10.1 CORROSION OF THE MSRE FUEL SALT CIRCUIT .~ R.E. Thoma Completion of the Molten-Salt Reactor Experiment has demonstrated that as a container material for resistant than has been estimated préviously. As evi- denced by changes in the chromium concentration of * ‘the fuel salt, the average cumulative corrosion sustained molten fluorides, Hastelloy N is even more corrosion . 93 by the MSRE fuel circuit since nuclear operations were initiated in 1965 extends to a depth of 0,46 mil. This value derives from revised rates based on analyses of the chromium content of Hastelloy N heats used in the ‘fabrication’ of the fuel circuit components, for which the average is 7.25 wt % rather than the nominal value of 6 wt % used in previous estimates. From chemical 94 evidence accumulated in operation of the MSRE with 233Y fuel, the slight attack which has occurred seems to be attributable to airborne contaminants introduced into the reactor during periods when the reactor vessel : was opened for maintenance. A summary of the results of chemical analyses of fuel salt samples (Fig. 10.1) shows that the concentration of chromium in these samples increased during the initial stages of runs 4, 8, 12, 15, and 19. During the first period of power operation, when a significant tempera- ture differential was imposed on the circuit for the first time (run 4), some corrosion was anticipated as the Cr°® + 2UF, = CrF, + 2UF; equilibrium reaction! adjusted to the temperature profile of the circuit. Under these conditions the increase in the concentration of chro- mium in the fuel salt resulted from the establishment of the equilibrium reaction and was not a signal of the presence of oxidizing contaminants. In each of the “ beginning periods of runs 8, 12,15, and 19, however, the increasing concentration of chromium in the fuel salt was not anticipated. In retrospect, we find that in each of these instances a common set of circumstances Cw. R, Grimes, Chemical Research and .Developmevnt for Molten-Salt Breeder Reactors, ORNL-TM-1853, pp. 40—45 (June 1967)., 150 140 130 - 120 10 100 90 80 " CHROMIUM (ppm) 60 50 ‘ 40 - 30 RUN FLUSH . ; 9 1 s ! ' 1 existed: the reactor core vessel was opened, and test arrays positioned within the graphite moderator lattice were exchanged. Although reasonable measures were adopted to minimize the possibility that airborne contaminants might be introduced into the system during these periods, it seems, nonetheless, that signifi- cant amounts of oxidants were then introduced into the open vessel, The increase in the concentration of chromium in the fuel salt well after run 14 began (see Fig. 10.1) seems to be inconsistent with the premise that external contami- "nants were the principal cause of corrosion. It may be recalled, however, that in the period preceding run 14 part -of the graphite and metal specimens in the core were removed and replaced. It seems quite possible, ~ therefore, that the residual concentration of reductant which was generated within the fuel salt during run 12 ~ was sufficient to offset the combined oxidizing effects of whatever contamination was incurred during shut- down and that characteristic of the fission reaction only through the early part of run 14, and that the subsequent rise in chromium concentration represents the normal compensating shift in the equihbnum corrosion reaction, The inference that moist air was the corrosion- inducing contaminant calls into question the efficacy of the flush salt. As an agent for removal of adsorbed moisture, molten LiF-BeF, flush salt is extremely ORNL-DWG 70-2164 DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASOND 1969 1966 1967 1968 Fig. 10.1. Corrosion of the MSRE Fuel Circuit in >>*U and >33 U Power Operations. effective, as demonstrated in numerous laboratory experiments, and should have scavenged moisture from ~ all ‘exposed surfaces. If an oxidizing contaminant or contaminants were capable of diffusing within the graphite, the probability of its removal by brief circulation of flush salt might be slight; instead it might be released into the salt gradually-after the moderator was heated to high temperatures. Thus oxygen, rather than water, seems.to be the most likely cause of the observed corrosion. This conclusion is reinforced by the fact that the scale found on the nickel cages which were used to expose Be® to the salt during run 15 was 95 comprised preponderantly of iron, whereas on other . occasions the principal structural metal in such scales was chromium. Results of chemical analyses showed that the prior fuel reprocessing treatment was effective in reducing the concentration of chromium in the salt * from 133 to 34 ppm, and iron from 174 to 110 ppm, The effectiveness of the reprocessing operations in reducing Cr** precludes the likelihood that significant amounts of Fe?* were delivered to the fuel circuit just prior to 233U operation. The reduction of Fe?* to Fe? to 51.31 equivalents, or 410.5 gof 0", and corresponds to a cumulative exposure to approxlmately 50 ft3 of air. During 233U operations the amount of oxygen entering the salt might have been expected to increase the concentration of oxide by 83 ppm, in excess of the sensitivity limits for the analytical method and well above the concentration observed. It must be recalled however, that early in 235U power operations the concentration of 0?7, as measured experimentally, declined from 120 to 60 ppm, suggesting that under power operation 0% is partially removed from the salt as a volatile species. It may be concluded, therefore, that the corrosion observed in the MSRE is likely to have been caused as described above but that the mechanism has, as yet, not been demonstrated un- equivocally. The rationale proposed above has several implications concerning the behavior of the MSRE during 233U operations. During the first 16 hr in which fuel salt was - circulated at the beginning of run 15, the salt did not by Be? suggests rather that the residual iron delivered with the purified carrier salt was oxidized after the ‘beginning of fuel circulation and that the oxidizing contaminant was contained in the closed fuel circuit. As increasing amounts of Be® were added to the salt mixture, the ratio of metallic iron to chromium found on the nickel metal cages was reduced until a normal balance was reestablished and corrosion ceased (see Table 10.1). o transfer to the overflow tank and behaved as though it contained a negligibly small bubble fraction. There- after, Be® was introduced, and the bubble fraction began to increase; with further exposures of the salt to - Be® the fraction varied erratically.? Certainly the Be® reduced the surface tension and thereby allowed easier transport of gas from graphite to salt, followed prob- ably by oxidation of the metallic iron impurity, which acts as an oxidant to the circuit walls. Corrosion would continue until the oxidants were consumed. The model ~ of corrosion proposed here has a relation to the changes Appraisal of the premise that maintenance operations might possibly permit the ingress of a reasonable in bubble fraction. The corrosion data suggest that with - respect to its physical and chemical properties, the fuel quantity of oxygen requires the following considera- tions. The cumulative amount of oxidation introduced . into the fuel salt (during both 235U and 233U operations) based on the increases of chromium at the beginning of runs 8, 12, 15, and 19 and the apparent losses of UF; at the ends of runs 7, 18, and 19 amounts Table 10.1. Relative Fractions of Fe® and Cr° Reduced _ from MSRE Fuel Salt inRun 15 : Equivalénts ~ Corrosion | ‘-FeOICto Sa;lx:)ple of qu . Rate _ on Nickel t Added (mils/year) Cage? FP157 224 088 113 FP 15-30 ‘ 4.09. 0.54 ) 19,5 _FP 1562 _ 617 - 035 ' 0.61 did not achieve a reference state until the beginning of run 17, That its bubble fraction then was greater than observed in 23°U operations probably was related principally to its lower density. The equilibrium constant for the reactlon Cr® + 2UF4 - = 2UF; + CrF, varies with temperature so that under nonisothermal conditions there is a tendency for - chromium to mass transfer from hot to cold zones in - the Hastelloy N container alloy. Such transfer is ‘diffusion controlled and under the relatively small temperature differentials of the fuel circuit is regarded as negligible in the MSRE. This tentative conclusion has already been substantiated by examinations of the * metal surveillance specimens which have been removed from the reactor core on previous occasions, which - showed no detectable depletion of Cr to within 10 u of A verage Fe/Cr in carrier salt was 2,24 at the inception of run 15. - o 2 - ) P. ‘N. Haubenreich, MSR Program Semiann. Progr. Rept. * Feb. 28, 1969 ORNL-4396, p. 3. the surface.? It is anticipated that examination of the heat exchanger in planned postoperational tests will confirm that mass transfer in the MSRE has been mconsequennally low 10.2 RELATIONSHIP OF THE DISTRIBUTION OF °3Nb IN THE MSRE TO THE PRESENCE , OF URANIUM TRIFLUORIDE IN THE FUEL SALT R.E. Thoma smnally during the period when the MSRE was oper- ated with 235U fuel. Their primary purpose was to offset the oxidizing effects anticipated to result from the fission reaction. Within this period the [U>*]/[ZU] ‘concentration ratio was estimated to have varied within the range 0.1 to 1.54%, as shown in Fig. 10.2. Such - variation effected no significant changes on either Minor adjustments in the concentration of uranium trifluoride in the MSRE fuel salt were made occa- 3R.E. Thoma, MSR Program Semiann. Progr. Rept. Feb, 29, 1968, ORNL-4254, p. 88. corrosion rate or fission product behavior in the fuel salt within the reactor. The reason for this derives from the fact that the 235U fuel was a highly buffered system in comparison with the 233U fuel used later, since the total amount of uranium in the 23°U fuel . exceeded that contained in the 233U fuel by sixfold. In contrast, operation of the MSRE with 233U fuel showed pronounced changes in the corrosion rates and ORNL-DWG 69-14557 3.0 28 b 2.6 : U3+/SU NOMINAL CONCENTRATION = —— U3%/3U NOMINAL CONCENTRATION, ASSUMING CREDIT FOR CORROSION AT BEGINNING OF RUN NO. {7 2.4 & 95\p EXPERIMENTAL _ ma DISPOSITION OF 25Nb 2.2 : 2.0 1.8 € e § 1.4 ‘ N\ 2 - |'\\- 1 \ \\\\\ - '\\ \\\ \\ : \ 1.2 o : N N N . | \\\ | - \\\\ \\ ‘ ' \ 10 N . N \‘ ‘ 50 0.8 _ \\ | ] | N 9 : — N~ . 40: ‘ \\ : . A a 3 0.6 N 303 Nm\\\fix\\\\\\\%M\\\\\ NN 2 0.4 : A~ ‘ A 208 1 \ 1/l 1 ' 0.2 \ ' : 10 0 [ s J = s V ‘ ‘ 0 0 2 4 6 8 10 12 14 s 18 20 22 24 26 28 30 O _ - - Mwhr(x103) \ ) RUN 17 RUN 18 RUNY 19 'RUN 20 Fig. 10.2. Et‘t‘ect of U3I £U on Distribution of >5Nb in the MSRE Fuel Salt Based on assumptlon of 0.762 equivalents ox1dat10n per gram-atom uranium fissmn and maximum power of the MSRE = 7.40 Mw(th). . fission product chemistry as the concentration of UF; was altered. Of considerable interest was the appearance of ®°Nb- in the fuel salt, noted for the first time in initial operations with ?23U fuel. This observation signaled the potential application of the disposition of 95Nb as an indine redox mdlcator for molten-salt reactors.! During August and September 1968 the 233‘U fuel charge was constituted from ?LiF-233UF, and "LiF- BeF,-ZrF4 carrier sait that had previously contained . 235,238YF, . During the initial periods of circulation, analysis of the newly prepated 7LiF-BeF;-ZtF,- 233UF, fuel salt showed that the fuel .circuit was undergoing corrosion. At that time the entire inventory of 25Nb appeared in the fuel salt and persisted there until 6.54 equivalents of Be® had been added. Within this period the concentration of chromium in the salt rose from 35 to 65 ppm, indicating the removal of 5.89 equivalents of chromium from the circuit walls. Thus, when °*Nb disappeared from the fuel salt after the final addition of beryllium (sample FP 15-62) a total of 12.43 equivalents were involved in the reduction of Fe?' and establishment of the Ct® + 2UF, = 2UF, + CrF, equilibrium, Samples of salt obtained during the brief period of the subsequent run (No. 16) as well as at the beginning of 233U power operations in run 17 (FP 164 and FP 17-2) showed the presence of 52 and 29% 25Nb inventory, respectively. In Fig. 10.2, nominal values of [U3+]I[EU] are shown for runs 17 to 20. These operations comprise nearly the total power operation of the MSRE with 233y fuel. The concentrations of UF; shown in Fig. 10.2 are based on the assumption that 0.76 equivalent of oxidation results from the fission of one atomic weight of uranium,? and that the maximum power achieved by the MSRE was 725 Mw(th). The equilib- rium constant for the corrosion equilibrium Cr° + 2UF, = 2UF, + CrF, reaction at 650°C, assuming an activity for Cr° in the Hastelloy to be 0.03,is 1.271 X 107 ppm. Thus, in a regime such as that which prevailed during the initial stages of run 19, the rate at which Cr° is leached from. the Hastelloy N circuit gradually. decreases as the Cr?' concentration of the fuel salt increases, During the initial period of run 19 the Cr** concentration of the circula_ting fuel. salt rose from 72 R, E. Thoma, MSR Program Semiann, Progr Rept Feb, 28 : 1969, ORNL-4396, p. 130. c, F. Baes, “The Chenustty and Thermodynarmcs of Molten-Salt Reactor Fuels,” in Reprocessing of Nuclear Fuels {Nuclear Metallurgy, vol. 15), P. Chiotti, ed., USAEC, 1969. 97 to 100 ppm. At that point the equilibrium concentra- - tion of [U**]/[ZU] in the fuel salt anticipated from free energy and activity data is ~0.5%. : ' ‘The disposition of ®*Nb in the 223U fuel dunng the initial. period- ‘of run 19 indicates that when [U3*]/[ZU] is less than ~0.5%, Nb becomes oxidized and enters the salt, possibly as Nb*" or Nb*, Then, as the corrosion reaction Cr® + 2UF, = 2UF; + CiF, proceeds to equilibrium, the [U3*]/{ZU] concentration ratio increases, and at a' [U**]/[ZU] wvalue of 0.5%, 95Nb precipitates from the fuel salt. Two levels of nominal U*/ZU concentration are shown in Fig. 10.2 for Nos. 17 and 18, the higher values based on the assumption that corrosion of the fuel circuit during the early stages of run 17 may have accounted for a fraction of [U*]/[ZU]. The extent to which this reaction might have contributed to the total concentra- tion of UF; in the fuel at the beginning of run 17 is obscure, because the MSRE was operated at full power at the inception of run 17. Power operation deposits - the noble metal fission products on the surface of the Hastelloy N and should cause the activity of Cr° at the alloy surface to be effectively reduced. The beginning period of run 19 is not analogous, for not until the corrosion equilibrium was established was the reactor operated at full power for sustained periods. Disposition of ?*Nb is indicated by the data points in Fig. 10.2. It is evident that when the [U**]/[ZU] of the 233U MSRE fuel was poised at ~0.5%, disposition of ®>Nb toward solution in the salt or deposition within the reactor was at a null point. If chemical transport of °*Nb in the MSRE proceeded as Nb”* (in salt) = Nb® (on metal or graphite) and if changes in the oxidation number of niobium were caused by the reaction Nb”* + nUF; = Nb® + nUF,, then UNp© (NU'F..'“U F Thus, at any given ratio of UF, to UF;, Nb should have behaved identically in either 235U or 233U fuel because for low concentration the activity coefficients for U* and U* should not have differed appreciably. Although there were. occasions during 23° U operations when the [U**]/[ZU] fraction was less than 0.5% and this inference might have been tested, this period preceded- that in which techniques for obtaining representative samples of the salt had been developed successfully. The presence of Nb was not observed, but - as shown in Fig. 10,2, freeze valve samples were obtained mostly after the [U*]/[ZU] concentration had been increased to values =0.5%. As noted below, it seems more likely that in its reduced form niobium is stabilized by reactions which produce other species than or in addition to niobium metal. Recent labo- ratory experiments (Sect. 11.3) suggest as well that reduction of niobium n+ from salts does not result in the deposition of metallic niobium. Preliminary results of laboratory expenments con- 98 productlon and would possibly serve as an accurate indicator of the power produced. About 600 g of plutonium was produced dunng ~ power operations with 235U fuel. Thereafter, addi- ducted by C. F. Weaver (see Sect. 11.4) indicate that - under mildly oxidative conditions niobium assumes an " oxidation number of ~3.6. During run 15, when the oxidation potential of the fuel salt was sufficiently high to permit Fe?* to exist in the salt in significant concentrations, nearly all of the ®5Nb inventory of the fuel salt was in solution, whereas in subsequent opera- tions when the oxidation potential was less, no more than ~50% of the ®SNb was found in the salt. This behavior would be explained if, when niobium is deposited on the moderator graphite, it reacts to form " niobium carbide and if, under the various redox regimes ‘which have been established in the MSRE during runs 17 to 20, niobium carbide has not been removed from the moderator graphite. The prevalence of niobium as -the carbide is compatible with the experimental obser- vations noted by Cuneo and Robertson,® who found that the concentration of ?SNb at all profiles in three different types of graphite was greater than would have been anticipated if after deposition the motope remained as the metallic species. 10.3 POWER OUTPUT OF THE MSRE BASED ON THE ISOTOPIC COMPOSITION OF PLUTONTUM R.E. Thoma B. E. Prince ~ In the MSRE, fluid fuel was circulated at rates which were sufficiently rapid with respect to changes in the isotopic composition of the fissile species that the salt samples removed from the pump bowl were representa- tive of the. circulating stream. This characteristic of moltensalt reactors makes it possible to use the results of isotopic analyses for a variety of purposes. One potential application, that of appraising the cumulative power generated by the MSRE at various periods, became apparent with the initiation of 233U opera- tions, for with 222U fuel the isotopic composition of the plutonium inventory (produced partly by that generated in 235U operations as well as from that added later) would change significantly during power 3D. R. Cuneo and H. E. Robertsén, MSR Program Semiann. - Progr. Rept. Aug 31, 1968, ORNL-4344, p. 141, tional plutonium was introduced into the fuel salt as a contaminant of 7 LiF-233UF, enriching salt and later to replenish the fissile inventory of the MSRE during 233y power operations. In the interim since a pre- liminary attempt was made to use isotopic dilution methods to compute the power output of the MSRE,} calculated values for the average reaction cross sections for a number of the cations in the MSRE fuel salt were revised using post-1965 data.? These values showed that the plutonium production rate during 23°U operations was greater -than previously anticipated and, cor- respondingly, that the relative changes in 24%Pu and '239py during 233U operations were also slightly different. Estimates of the variation of 23°Pu and 240Py during recent power operations require that the quantity and composition of the initial plutonijum inventory be . known accurately. As noted previously,! attempts to determine the concentration of plutonium in the fuel salt from gross alpha count measurements were not very satisfactory . because of the high specific activity of 238py, An improved estimate of the plutonium in- ventory of the system was made from extrapolations of the observed changes in 23°Pu and 2*°Pu in the beginning stages of power operation with 233U fuel. - The initial 24°Pu/?3?Pu concentration ratio was com- puted to be 0.0453, with the plutonium of the reactor at that point as 568 g, ~2% more than estimated from previous analyses. Current estimates of inventory values have been computed for this revised starting inventory. The values obtaining at the time the samples were taken were based on estimated average values for the rates of change of 23°Pu and 2‘"’Pu in the penod between samples. . Samples of the MSRE fuel salt were submitted routinely for determination of the isotopic composition of the contained fissile species. Comparisons of the results of plutonium assays with nominal values which should result from operations at various power levels from ~7 to 8 Mw were made. Within this range, best agreement between calculated and experimental values was obtained for 2 maximum power output of ~7.40 Mw(th). A comparison of calculated and observed IR, E. Thoma, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 98. iB. E. Prince, MSR Program Semiann, Progr. Rept. Aug. 31, - 1969, ORNL-4449, p. 22, | ( . * " . ¢ n ¢ * * * * ( © Table 10.2. Isotopic Composition of Plutonium in the MSRE Fuel Salt at a Power Generation Rate of 7.40 Mw(th) . . o Fuel Circuit | X Isotopic Compositioh Isotopic Composition o . - uel Circuit Inventory . * Sample Pl:\lvl:r \whed A23%y A240p, (Calculated) (2)° (Calculated) (Analytical) -No, . (g per 1000 Mwhr) (gper 1000 Mwhy) 35— .0 . _Wt% Pu/ZPu 28400 239 Wt % Pu/ZPu 2404 1239 ; : Hour§ : _ o Pus TPu ZPu 239p, 240p, Pu/"""Pu 239p, 240p, - Pu/"""Pu "Rl 0 0o 541.5 © 24.53 568.4 9526 4.32 0.0453 -FP179 - 148 537 -3.025 - +1,178 539.8 25.16 5674 95.14 443 . 0.0466 95.26 4.35 0.0457 "FP17-18 466 2,227 - =2.996 +1,144 . 5348 27.09 5643 9476 4.80 0.0507 94,28 5.16 0.0547 ‘FP17-19 ' 542 3,656 —-2949 - +1,180 530.6 28.78 561.8 9444 5.12 0.0542 9448 5.00 0.0508 "FP17-20 = 697 4,492 =2.924 +1.105 . 528.1 29.70 560.3 9426 5.30 0.0562 94.20 5.25 0.0557 FP17-23 = .920 5,862 -2.894 +1.086 5242 31,18 557.8 9397 595 0.0595 FP17-27 . 1047 7,131 - 2871 +1.072 520.5 - 32,54 5555 93,70 5.86 0.0625 93.58 5.80 0.0620 FP17-28 1145 17,946 -2.854 +1.063 518.2 3342 5540 9353 6.03 0.0645 94.30 5.97 0.0639 FP17-30. = 1290 '83827 . —2.845 +1.054 515.7 3435 551.5 9351 6.23 0.0666 93.16 6.18 0.0663 "Run17F 1536 10,245 ~2.824 +1,039 5117~ 3582 5499 93.04 651 0.0700 | FP 181 1536 10,245 R . i " 514.1 3491 5514 9322 6.33 - 0.0679 " FP 18-1 i536 10,245 -2.875 - +1.075 514.1 3491 5514 9322 6.33 0.0679 92,92 6.36 0.0684 FP 18-5 1562 11,231 . -2.860 . +1.066 513.5 3511 551.1 93,18 6.37 0.0684 92.65 6.61 0.0713 FP 18-10 1851 12,3713 -2.843 +1.055 507.6 -~ 3732 5473 9273 6.82 0.0735 9238 6.84 0.0740 FP 18-13 1976 13,873 - =2826. +1.038 505.0 3826 5457 9254 101 0.0757 192,16 7.04 0.0764 FP 18-22 1976 13,873 -2.826 . +1.038 505.0 38.26 < 545.7 9254 1.01 0.0757 . 91.80 17.36 0.0802 FP 18-27 2306 15,523 =2.799 . +1.017 498.3 40.69 5414 92,03 7.52 0.0817 9163 7.49 0.0817 FP 1843 2461 17,281 -2.773 +0.999 495.2 4181 5395 91.80 7.75 0.0844 91.48 17.63 0.0834 ‘Run 18-F 2544 - 18,143 —-2,758 +0.989 493.5 4241 5384 9167 1.87 0.0859 Run 191 2544 18,143 ' o B 495.2 4181 5394 9179 1175 0.0844 , FP 19-17 2625 18,739 -2.770 - +0.993 493,5 4239 5384 91.67 17.87 0.0858 " 91.22 7.84 0.0860 FP 19-18 2642 19,093 -2.763 +0.989 493.2 42,51 5382 9165 190 0.0862 91.19 7.87 0.0863 'FP 19-21 2724 19,449 ~2.755 +0.985 491.6 43.10 537.1 9152 8.02 0.0877 91.02 8.03 0.0882 "FP19-22 2791 19,693 -2.752 +0.985 490.2 " 43,58 536.3 9142 8.13 0.0889 90.90 8.11. 0.0892 FP 19-24 2791 19,693 -2,752 +0.985 - 490.2 43.58 536.3 91.42 8.13 0.0889 89.88 8.99 0.1000 FP-19-256 2791 19,693 - -2.760 +0.983. 541.1 46,74 5906 9162 791 0.0864 : FP 19-27 2791 19,693 -2,760 +0.983 541.1 46.74 590.6 91.62 791 0.0864 91.01 8.03 0.0882 "FP 19-30 2818 19,693 -2.760 +0.983 540.5 4693 5899 91.63 17.96 0.0868 90.89 8.13 0.0895 FP 19-314 2818 19,693 -2,750 +0.983 - 6626 54,51 719.6 92.08 7.58 0.0823 . FP 19-35. 2964 20,961 -3.820 +1.369 658.6 5596 717.1 91.85 - 7.80 0.0850 91,35 1.1 0.0851 FP 1943 . 3102 21,989 --3.810 +1.357 6548 57.32 7145 91.64 8.02 0.0875- 91.18 7.90 0.0866 FP 19-53 3294 23,185 =3.795 +1.339 649.5 59,18 7111 91.33 8.32 0.0911 90.88 8.25 0.0908 FP 1963 3561 24,631 -3.775 +1.339 642.2 6177 7064 9091 8.74 0.0962 90.49 8.50- 0.0939 FP 19-74 3693 26,078 -3.750 -+1.316. 639.3 62,77 7045 89.09 891 0.0982 90.15 8.78 0.0974 Run 19-F 3774 27,069 -3.720 +1,302 637.1 63.54 703.1 90.62 9.04 0.0997 Run20d 3774 27,069 | | - 6258 6181 690.0 90.69 896 0.0988 FP 206 3820 27,236 -3.670 +1.309 624.6 62,25 689.3 90.61 . 9.03 0.0996 89.89 8.99 0.1000 FP 20-31 4159 28,294 -3.660 +1.293 6156 - 6543 6834 90.07 9.57 0.1062 89.36 9.43 ‘0.1055 Run 20-F 4159 28,294 ~3.660 +1,293 615.6 6543 6834 90.07 9.57 0.1062 8 Average for period between samples, b A ssumes 92% fuel charge in circulation, 66 100 ORNL-OWG 70-6755 oM 0.10 g / - 8 y_l 0.09 S / s s 5 0.08 a . _ | // 9 < ® & 0,07 — | | , - .’.\ | . 0.06 // : NOMINAL VALUES OF 240/23%y AT 7.40 Mw (th) . ~ . . ‘/ ) ' 0.05 / . - - : - RUN -LRUN 17———‘-——RUN 18——-1-—_—RUN,49——|-¥—| i ] ! 1 L ] L o . 1000 2000 equivalent full-power hours 3000 4000 " Fig. 10.3. [240/239py] in the MSRE Fuel Salt Circuit During 233U Operations. values for the isotopic composition of plutonium which should result from a2 maximum power output of 7.40 Mw is shown in Table 10.2 and Fig. 10.3. Agreement tests indicate that the standard deviation between _ calculated and observed values is +0.63% and that the average positive bias in the experimental data is 0.093%. On this basis the maximum power output was 7.41 + 0.05 Mw(th). The precision of this value seems to be adequate for related analyses of reactor operations. It is considered unlikely that further refinements in cross-section data for the plutonium isotopes will require any substantial changes in the calculated inventories used for this comparison. However, for the purpose of estimating “uncertainty” in the combined cross-section data and neutronic model used to cal- culate reaction rates, an equivalent of 2% in the power output is judged conservative. Any further improve- _ ments in the data, model, or computing approximations used in the analysis which have significant influence on these results will be reported at a later date, 10.4 ISOTOPIC COMPOSITION OF URANIUM DURING 233U OPERATIONS R.E. Thoma For operation of the MSRE with 2 33y f_uél, estima- tion of the power output of the MSRE from measured changes in isotopic composition of the fissile material is achieved with considerably greater precision from analyses of plutonium than from uranium. This arises from the fact that the rate of change in the relative fraction of the most abundant isotopes for plutonium, 239py and 24°Py, is some four times that for the uranium pair, 233U and 234 U. Analyses of the isotopic composition of uranium in the fuel circuit during 233U operations- were employed, therefore, primarily to determine whether they afforded approximate con- firmation of the power estimate as inferred from plutonium data (Sect. 10.3). Calculations of the iso- topic composition changes of the uranium in the MSRE fuel circuit which should have accompanied operation of the reactor at a maximum power output of 7.41 Mw(th) were made and compared with the results of mass spectrometric analyses. The results of this com- parison are shown in Table 10.3 and in Fig. 10.4; they indicate that the changes observed in the isotopic composition of the uranium were in excellent agree- ment with those of plutonium. ' 10.5 SURFACE TENSION AND WETTING BEHAVIOR OF THE MSRE FUEL AND COOLANT SALTS , H. W. Kohn .We have measured‘fhe surface tension of molten fuel salt and of Li, BeF, in the temperature range from the 101 Table 10.3. Isotopic Composition of Uranium in the MSRE Fuel Salt Circuit? U/2U (wt %) Sample b 234y77/,233 No. EFPH 233 234y 235y; 236y 238y o) Run 171 0 84.687 6.948 2.477 0.0808 5.807 0.08204 FP17-18 466 84.590 - 7.011 2.489 0.084 5.828 0.08288 / 84.690 6.990 2.470 0.084 5.771 0.08253 FP 17-24 920 84.489 7.073 . 2.501 0.087 5.849 0.08371 | | 84.382 7.058 2.487 0.089 5.986 0.08364 FP 17-32 1338 84440 7.131 2.510 0.090 5.867 0.08445 - . 84.445 7.128 2,487 0.087 . 5843 0.08440 Run 17-F 1536 84.363 7.152 2.513 0.091 5.875 0.08477 Run181 1536 84.393 7136 2511 0.091 5.870 0.08455 Run 18-2 1536 84199 7.138 2.507 . 0.089 6.067 0.08477 FP 18-4 1563 84.385 7.141 2.511 0.091 5.871 0.08462 84,249 7.158 2.527 0091 5975 0.08496 FO 18-10 1852 84.326 7.180 2.518 0.092 5.883 0.08514 - 84.269 7.178 2.507 0.091 - 5.955 0.08517 FP 18-13 1976 84.298 7.199 2521 0.092 5.890 0.08539 84.060 7.203 2.517 0.087 6133 0.08568 FP 18-22 2221 84.241 7.232 2,529 0.095 5.902 0.08584 84167 7.208 2517 0.098 6.016 0.08563 FP 1843 2461 84.189 7.265 2.534 0.097 5.912 0.08629 84.041 17.232 2.537 0.093 - 6.097 0.08605 Run 18-F - 2544 84.169 7.279 2.536 0.098 5.916 0.08648 Run 191 2544 84,185 7.267 2,534 - 0.097 5.913 0.08632 FP 19 10-12¢ 2544 84.224 7.268 252 0,097 5.882 0.08629 - FP 19-35 2964 84377 1.349 2,543 0.100 5919 0.08709 ‘ 83.987 7.338 2,537 0.099 6.036 0.08737 FP 1943 3102 84.103 7.348 2.539 0,101 5,909 0.08726 - ' . 83994 7328 2.533 0.099 6047 0.08724 FP 19-53 3294 84.060 7.375 2.546 0102 5917 0.08773 | 83912 7.358 2.537 0101 - 6.092 - 0.08768 FP 19-63 3561 84.000 7.413 2.553 0,104 5.931 0.08825 o '83.927 7408 2542 0102 6.021 - 0.08826 - FP19-74 3693 83.971 7.430 2555 0.105 5937 0.08848 S ' 83.801 7418 2569 - 0102 6.102 0.08851 Run19-F 3774 83.953 7442 . 2,557 0105 5.942 0.08864 Run20d 3774 83973 7.427 2.555 0.105 5.938 v 0.08844 FP 20-3€ - 3774 83996 7426 2.549 0105 5923 0.08840 FP206 = 3820 83986 . 7.433 2.550 0.105 5925 . 0.08850 | - : 83.614 7.435 2.577 0.104 6271 = 008892 FP 20-31 4159 83911 7482 2.558 0.108 5.940 " 0.08916 83.742 o 7484 2.577 - 0,105 6.092 0.08936 + 2Upright type indicates values computed on the basis that the maximum power generated by the MSRE was 7.41 Mw(th). The following rates (furnished by B. E. Prince) were used: 233U: —4.643 X 1072 g/Mwhr; 234U: +3,6325 X 103 g/Mwhr; ?35U: +9.5596 X 105 g/Mwhr; 236U: +3.0725 X 10~ g/Mwhr;233U: —2,90 X 10~ g/Mwhr. Results of mass spectrometric analyses are listed in iralicized type. : : - ' bEquivalent full-power hours. - ©Fuel addition. ' : - 102 ORNL-DWG 70-6756 0.090 . ® / : I /"/."‘/ 0,088 —— o | A"/ g: 0.086 o / . = o] < o n o "/ . 0.084 ,,/’<\‘ : _ : L NOMINAL VALUES OF 234/233; AT 7.41 Mw (1h) / - . 0.082 T |———Ru~ 7 —|——RUN 18 ———[——RUN 19 R;g 0.080 L I : 0 | aooo _ aooo equivalent full=power hours 3ooo oq . Fig. 10.4. 234/233y jn the MSRE Fuel Circuit During 233U Operations. freezing point to 709°C. We used the capillary depres- sion method, not only because of its simplicity but also to check the results reported previously.! This method is not as precise as some others, but it is accurate within the error limitations. The equipment was similar to that described previously except for the following changes: Only three capillaries, % in., %, in., and % in., in diameter, were used. These were drilled in a carbon plunger which fitted closely into a carbon crucible. The whole assembly was in a low-attenuation furnace so that they could be x radiographed. The holes and the rather irregularly shaped salt well were placed so that a maximum amount of salt ¢ould be used without having any of the configurations casting a shadow on the other. An extra blind hole in the carbon plunger contained a calibrating wire to allow us to measure magnification and to compensate for film shrinkage (see Fig. 10.5). , An absolute value of the surface tension determmed this way depends on an accurate measure of the contact angle A commonly used formula for this is: 2ycos 0 =(h +R[3)R(0;—0,)8, where v = surfdce tension, . R -=radius of capillary, 15, S. Kirslis, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p, 109, = density of liquid, | 0,= density of vapor, g= accéleration of gravity, @ = contact angle, | h = height of capillary depressmn or rise. This assumes that the surface is hexmsphencal Further refinements (Sugden, Schroedinger)? are out- side the limits of error here and are not even applicable _since they have been worked out for capillary rise phenomena. As the construct of Fig. 10.6 shows, the ratio of the radius of the capillary to the radius of curvature of the meniscus is equal to the cosine of the contact angle, provided the meniscus is hemispherical. The radii can be measured (#5%) from the x-ray shadowgraph. In this work the cosine of the contact | angle so determined is 0.875, which does not agree very well with values previously obtamed in sessile drop measurements (0.77). We were able to measure capillary depression and contact angle on fuel salt, giving the followmg empirical equation: v =(237 i‘l‘o-)-,'" (0.08 + 0.01)¢ dynes/cm where ¢ is in degrees centigrade. For Li,BeF, the shadowgraph was too faint to allow an accurate 2Sugden, Schroedinger, Surface and Colloid Science, ed by E. Matijevié, J. F. Padday, chap. Il, Wiley, New York, 1969, 103 5 _ _ PHOTO 99444 ORNL - DWG 70- 6757 Fig. 10.6. Construct Illustrating Relationship of Radii to Contact'Angle._ measure of the contact angle, but using the same value obtained with fuel salt we get Y= (245t 5) — (0.12 + 0.01)¢ dynes/cm. We found that allowing the salt to freeze in the capillaries would give results about 4 to 6% low. Reducing the fuel salt melt with metallic zirconium and with metallic lithium (each dosage calculated to reduce 10% of the uranium present) failed to change meas- urably either the contact angle or the capillary depres- ‘sion. - Fig. 10.5. Shadowgraph of a Capillary Depression Molten-Salt - Surface Tension Experiment. 11. Fission Product Behavior 11.1 EXAMINATION OF THE FOURTH SET OF SURVEILLANCE SPECIMENS FROM THE MSRE F.F.Blankenship E.L.Compere S.S. Kirslis 11.1.1 Radiochemical Anatyses of the Graphite Details regarding samples, sampling procedures, and examination have been given previously,! along with results on short-lived fission products. All of the data on graphite are summarized in Table 11.1 for salt- soluble products and in Tab]e 11.2 for metallic fission products. ' Some of the salt-soluble products have rare-gas precursors and enter the graphite by diffusion through the gas in the graphite pores. The resulting deposition profiles are like those in Fig. 11.1. Products with longer-lived precursors, such as 51-day 3°Sr and 29-year 137Ce, have flatter profiles because the krypton or xenon penetrates deeper before decaying. Correspond- ingly, the ‘slopes for 58.8-day °'Y and 12.8-day '*°Ba are steeper because of the shorter half-lives of the gaseous precursors. The half-lives of the precursors are listed in Fig. 11.1. Sometimes, even for the same element, a sharp turnup is found at the surface and sometimes not; this is not understood. To a first approximation, because the fuel does not wet graphite, salt-soluble fission products were not expected in the graphite, and in general the amounts actually found there were trivial. As expected, the highest amounts were found for those species entering the graphite as gases. The amount of 3°Sr is about twice the amount found previously. The presence of other salt-soluble species in the graphite could probably be attributed to the existence of a gas-borne mist of salt that has been found to be associated with the radioactive fuel. This mist was presumed to have entered the graphite pores. If this were the mechanism, then the percents of inventory of 95Zr and '*%Ce in the graphite should be the same, 1F. F. Blankenship, E. L. Compere, and S. S. Kirslis, MSRK Program Semiann. Progr, Rept. Aug 31, 1969, ORNL-4449, p, 104, Unfortunately, there is too much scatter in the data to warrant a firm conclusion on this point. The effect of 76,000 Mwhr of exposure as compared with 20,000 was especially pronounced for the 29-year 137Cs and is explained by the long half-life. - Since flush salt was not used on these samples, in contrast with past practice, it was of interest to note whether there was an increase in fission products in the graphite. Inspection of Table 11.1 shows that, in general, for this class of products, there was not a large increase and indeed the change was in the opposite direction for **®Ba and ***Ce. 12 : . ORNL-DWG T0-1345 V-47, IMPREGNATED CGB 2 : 2.34x160% dpm/em? : 3.2 min %%r— %9 10“ 5 > V-47, IMPREGNATED CGB ) 3.23240% dpm /cm? 10'© 10 sec ¥'kr -y e 5 € o a7 2 | V-47, IMPREGNATED CGB 10° 2.29x10'° dpm/em2 16 sec *Oxe—+66 sec #40¢cs —- 14984 | M1, ORDINARY CGB 1.29x107 dpm/em? 2 | 23 sec '1—»3.9 min Bxe 137, 10" — : 0 0 .20 - 30 - 40 S0 60 70 ' DISTANCE FROM SURFACE (mils) Fig. 11.1. Concenttauon Profiles for Fmon Products I-Iavmg Inert Gas Precursors. 104 Table 11.1. Deposition of Salt-Soluble Fission Products on Graphite from the MSRE Core Graphite Graphite - Exposure Distance 895y . iy 95@- . 131, 137cq ll'olg 1840, Sample type® Mhr - from ° Dis. Percent Dis. Percent ° Dis, Percent Dis. Percent Dis. Percent Dis. Percent Dis, Percent No. and yP bottom , min™ of min-1 of min-1 of min-1 of min~1 . min” . min=1 of Face ‘ core(in)’ cm-2 c cn~2 . cm'za em2 cn™2_ ° em~2 o em™2 x 1010 Inventory® x 10° Inventory® x 10% Inventory® x 108 Inventory® x 107 Inventory® x 10? Inventory® x 107 Inventory® V=47 wide . CGBI 20,000 575 - 2,31 17.3 3.23 2.67 ‘07.959 0.076 3.77 0.47 _4.87 0.894 2,29 1.49 3.50 0.057 "P-129 wide CGB 20,000 54~7/8 2.47 26.5 4,18 3.46 1.67 " 0.13 6.34 0.78 1.33 0.244 2.70 1.75 P-129 narrow _ ‘ 3.58 26,8 3.65 3.02 1.23 0.098 5.16 0.64 0.97 . 0.178 2,40 1,56 K=3 wide Pocol 20,000 50% - 2,91 21.8 5.59 4,63 0.583 0.047 2.36 0.29 - 1.36 0.250 3.1 2.41 K=3 narrow ‘ : - .3.07 23.0 6.10 5.06 3.96 0.32 8.23 1.02 1.04 0.190 4,37 2,84 YY-22 narrow CGB i0,000 ©32% 2,50 18.7 5.05 4.18 2,81 0.22 '1.90 0.23 1.14 0.208 3.65 2.37 8,82 0,144 YY=-22 wide ‘ 3.17 23.8 5.54 4.59 7.07 0.56 5.65 0.70 4,35 . 0,799 5.01 3.25 11,0 0,180 V=43 wide CGBI 20,000 &3 3.08 23.1 2,61 @ 2,16 2.81 0.22 2.28 0.28 3.50 0,642 3,32 2,15 7.49 '0.123 V=43 narrow ) 5.98 ' 44.8 0.426 0.35 1.99 0.16 2,64 0.33 4.58 0.841 5,56 3.61 0.62 - 0.010 M1 wide CGB 20,000 57 2,78 20.8 3.09 2.56 1.12 0.089 5.43 0.67 1.29 0.236 3.03 1.97 5.00 0.082 M 1 narrow . . . 2.56 19,1 9.09 7.53_ 1.30 0.104 4,55 0.56 - - 3.26 - 2,12 7.54 0.123 X=-23 wide CGB 20,000 32% 4.8 . 36,1 5.668 -';.69 6.71 0.54 6.58 0.81 3.48 0.638 5,61 3.65 18.60 0.305 X-23 narrow ) ‘ 4.87 36.5 5.88 4.87 5.11 - 0,41 5.73 0.71 2,97 0,545 5.65 3.67 14.30 0.234 K-1 wide CCB 20,000 B 8.59 - 64.5 8.70 7.21 20.5 1.64 2.64 0.33 7.78 1.43 8.48 5.50 N-1 narrow ‘ i 5.60 42.0 4,78 3.96 15.1 1.20 5.10 0.63 2,82 0.517 5.09 3.30 Average for 20,000-Mwhr samples 3.89 _ 4.90 4,86 4,56 2,96 4.27 8.54 NL~-7 wide CGB 76,000 5% 2.3 17.6 2,65 2,19 6.94 0.55 T 4,04 0.50 7.15 1.312 2.51 1.63 5.83 0.095 N1~7 narrow ) ) - 3.23 24,2 3.71. 3.07 1.83 0.15 7.86 0.97 10.8 1,975 -2.97 - 1.92 8.21 0.134 X-10 wide €GB 76,000 32% 3.68 27.6 4,17 3.46 5.14 0.41 10.5 1.29 14,90 2,738 4.01 2,60 26.0 0.426 X~10 narrow ’ L 6.20 46.5 7.08 5.86 5.26 0.42 11.1 1.37 14,90 2,739 6.62 4,22 25.8 0.423 Mi~3 wide CGB 76,000 &Y 4,25 31.8 4,96 4.11 2,43 0.19 1.73 0.21 28,90 5.302 3.15 2.04 15.10 0.247 ML~3 narrow . 6.02 45.1 2.67 2,21 2.99 0.24 11.9 1.47 31,50 5,779 6.10 3.95 22,50 0.368 Average for 76,000 Mvhr samples 4,29 - . s21 4,10 7.85 18,02 4,23 17.24 Average for all samples . 4.00 4.71 4.64 5.50 7.48 4.26 12.02 Previous average (Survey 3) o 2.25 - 4,83 2,42 - 7.26 18.2 Control ccad : o 1.19x10° 7.17x10% 1.87x10% 3.50x10% 1.68x10° 8CGRI 1s impregnated CGB; Pocol is impregnated Poco. Ppistance from bottom of the core to bottom of the sample. ®The amount on 1 ca? expressed as percént of the calculated amount in 1 gram of fuel if it escaped only by decay. dPhced in hot cell for sampling with previous exposure to salt or radiation. SOT ' ¢ Table 11.2. Deposition of Metallic Fission Products on Graphite from the MSRE Core > Graphite Dif::me ‘ ?S\b Mo 103p, 106, 1 1ag 125g, 1290, 132, ‘Sample Craphite Exposre o % Dis Percent Dis Percent Dis Percent Dis Percent Dis Percent Dis Percent Dis Percent Dis Percent Number Type' (Mwh) L~ min7? of min~? of min™! of min~} " of min“! of min~t of min~? of min~} of and Face (in)?® em™2 Inventory® cm™ Inventory® cm™ Inventory® em™ Inventory® om™2 Inventory® em™2 Inventory® cm™ Inventory® cm™? Inventory® " % 10° x 1010 x 10° X 10* x 10% x 10 x10° X 1010 v4Twide CGBI 20000 $7% 27 30 120 963 L7 3.94 0.661 1.91 475 .2 345 118 6.35 3.39 0.903 7,84 P-129 wide CGB ' 20,000 54% 5.87 66 LB2 1464 180 40 Li4 3,28 6.04 90.5 4,40 1.51 8,02 4.28 101 8.77 P-129 narrow 3.30 3.1 0.746 600 0958 2,13 0,792 2,29 5.54 83.0 211 0.72 5.38 2.87 0.947 8.22 K-3 wide Pocol 20,000 50% 0.69 0.7 0136 149 0255 0.57 0.198 0.57 1.19 17,7 0.53 0.18 1.54 402 0.703 6.10 K-3 narrow 9,27 10.4 1.72 1379 298 6.63 2,25 6.49 16.20 242.1 6.53 224 10.80 5.78 157 13.63 YY-22 narow OGB 20000 32Y% 22.3 250 129 10.38 176 391 1.22 3.53 558 83,5 1.60 0.55 5,07 2.70 102 8.86 YY-22 wide : 190 2.3 127 1020 252 5.61 2,23 6.43 6.79 101.7 6.90 2.36 14.20 7.56 0.779 6.76 V43 wide CGBI 20,000 &Y 1.72 8.6 0.769 618 158 351 1.32 3.81 7.99 119.7 5.91 2.03 4.1 2.54 0.557 4.84 V-43 narrow oo L 126 141 089% 718 1.84 4.1 1.51 437 6.46 96.8 5.52 1.89 4.41 2.38 0.565 4,90 M1 wide CGB 20000 57% 4.84 54 0836 671 06717 15 0319 092 6.41 96.0 491 168 8.42 449 1.07 9.27 M-1 natrow .67 64 174 1394 . 276 6.1 1.85 5.33 - 8.31 1240 7.27 249 - 10,70 57 156 13.5 X-23 wide - CGB . -~ 20000 32% 184 206 157 1263 373 8.3 3713 10.8 9.51 1420 1L0 375 9.29 4,96 0.91 79 X-23 namrow 279 31,3 119 9.56 353 7.8 2.84 8.21 - 12.50 187.0 138 . 472 729 - 3.89 118 10.2 M-1 wide CGB - 20,000 4% 1 154 0989 7.94 248 5.5 175 5.05 11.90 177.0 699 - 24 10,10 537 N-1 patrow : 8.09 9.7 087 699 0267 0.5 157 4.54 11.70 175.0 13 3.88 9.67 5.16 1.36 11.78 Average for 20,000-Mwhr samples 10,804 1.1394 1.927 ' 1559 8.057 6148 7.738 1.010 NL-7 wide CGB 76000 57% 4.82 54 06R 508 122 2.72 173 4.98 6.06 90.85 - 25.0 8.58 19,30 10,30 145 1259 . . NL-7 wide o : 6.12 69 0936 752 135 2.0 169 4.83 705~ 10568 20.3 6.96 18.20 X7 2.00 1735 X-10 wide CGB 76,000 32% 450 50.5 206 16.54 362 8.05 5.05 14.6 . 9.82 147.1 904 310 - 2370 12.66 2.60 22.57 X-10 natrow - ) SR X 578 203 1633 4.68 10.4 6.86 198 1040 1559 83.1 . 285 . 3350 11.90 3.04 26.37 ML-3 wide COGB 76,000 4% 24,3 273 0.652 523 162 3.6 2.85 8.22 8.81 13L9 527 ~ 1.80 11.80 6.31 1.42 1230 ML-namow 12,9 144 0.568 456 163 3.63 2.57 7.42 13.20 1917 36.1 1.24 22,30 1193 218 18.92 Average for 76,000-Mwhr samples 24,123 1146 2.353 3.458 9,223 51,267 21.467 2115 ' Average for all ssmples 14.610 L1141 2.049 2,101 © 8.390 19.039 11.660 1.341 Previous average (survey 3) 111.6 4,11 5.98 5.18 - 10,0 1.64 Control samples? 5.68 X107 2.07 X 107 145 X107 000032 839X10° 0000242 370X 10° 005546 839Xx10° 0000287 326X 10° 00174 210X10° 0.01820 4CGBI is impregnated CGB; Pocol is impregnated Poco, 5pistance from bottom of the core to bottom of the sample. ®The amount on 1 cm? expressed as percent of the calculated amount in 1 g of fuel if it escaped only by decay. . 4Placed in hot cell for sampling with previous exposure to salt or radiation, C . ‘ C . n 2 n e . * a 4 * » " ' %01 Some small fraction of these products was deposited in the graphite by fission recoil. This would tend to show up as high values for the samples at 34", in. from the bottom of the core as compared with samples from the bottom and top. In general such an effect was not consistently recognizable for this class of products. The percent of *3'I tended to run higher than that of 95Zr or' 1*4Ce, for examiple, because some of the *3!1 entered the graphite as a tellurium precursor. The results on metallic fission products given in Table 11.2 show that the lack of use of flush salt did not result in an increase in activity found on the graphite. The percents, based on calculated - inventories, were higher, however, than for the salt-soluble products. Also, higher values were frequently found on graphite | samples near the middle of the core, where the flux was highest. This tendency had sometimes been encoun- tered previously.? The effect of longer exposure time was noticeable for 35-day ?SNb, 367-day '°SRu, 986-day '2°Sb, 34-day '*°Te, and 3.25-day '32Te, but not for the other metals. This effect was qualita- tively in accord with our expectations for ruthenium and antimony, but not for the shorter-lived products such as niobium and the telluriums, Also, the telluriums showed atypical profiles. Figure 11.2 illustrates a typical profile. There seemed to be a high-capacity, slow process that accounted for deposition near the surface where the profile is steep. High capacity in this sense means accessibility to a large population of fission product atoms. This process could have been solid or volume diffusion through the graphite or intercrystal- line diffusion. Fission recoil undoubtedly caused some of the deposition in the first 30 p. Farther into the graphite, the position and slope of the typical profile correspond to a low-capacity, fast process. This could be diffusion through the gas or on the surface of the pores in the graphite. Throughout the study of fission- product behavior in the MSRE, a surprising amount of metallic product activity has_been found in the gas space in contact with fuel. There is no reason that gas in the pores of the graphfle should be - dlfferent In some instances the metals appear to be a smoke ‘ 107 14 ORNL-DWG 70-1344 ~7, ORDINARY CGB 1.73x10°% dpm /em? 76, 000 Mwhr V-47, IMPREGNATED CGB 6.64x107 dpm/em? 20, 000 Mwhr 20 30 40 50 60 70 DISTANCE FROM SURFACE {mils) 10 Fig. 11.2. Effect of Exposure on Deposition of 367-day 106Ru on Graphite. are born an atom at a time and fail to nucleate (homogeneously) in the bulk liquid. Reaching a solid surface, they nucleate heterogeneously; reaching a gas surface, they pass into the vapor phase at a tremendous supersaturation. In the gas phase, nucleation is easier because rates of diffusion are much greater. If the mist ~ particles are efficient in nucleating - metals, the two viewpoints may not be as much at variance as they ~ appeared at first glance. But it should be remembered - that at the tracer level ?5Nb passes into the gas phase independent of the salt mist. Some of the most recent data indicate that the metals can ride the salt mist (cf. Sect. 11.2) and that they are somehow highly concen- trated in the mist as compared with the concentration in the fuel. According to the first viewpoint, the metals 28, S. Kirslis and F. F. Blankenship, MSR Program Semiann, Progr. Rept. Aug. 31, 1968, ORNL-4344, p, 125, with almost no accompanying salt mist. , Figures 11.3 and 11.4 show the profiles for the telluriums. The steeply descending portion of the profile near the surface is greatly foreshortened, and the . inner portion is steeper than for typical species. The bend between the two slopes is much sharper than for typical species. This suggests that the fast, low-capacity mode of migration is accentuated exceptlonally in the case of tellurium, 108 12 ‘ ORNL—DWG 70-1346 NL-7, ORDINARY CGB 1.93x10° dpm fcm? 76,000 Mwhr dpm/g V- 47, IMPREGNATED CGB 6.35x10% dpm/cm? 20,000 Mwhr o 10 20 30 40 50 60 70 ' - DISTANCE FROM SURFACE {mils) Fig. 11.3. Effect of Exposure on Deposition of 34-day !2%Te’ in Graphite. One outstanding difference from previous results was the relatively small amount of ?5Nb; it was down by a factor of 7 or more. This was thought to be a result of operating with .the fuel in an oxidizing condition, that is, with low UF;, so that the stable state was not the metal but a niobium cation. 11.12 Radiochemical Analyses of Fission Product Deposition on Metal The first two times that the surveillance assemblies were removed from the MSRE, metal samples were obtained by cutting the perforated cylindrical Hastelloy N basket that held the assembly. The next two times Y-in. tubing that had held dosimeter wires was cut up ‘to provide samples. When the tubing was first. used, lower values were obtained for the deposition of fission products. After the fourth assembly was removed, the basket was no longer of use, and could again be cut up to see if the difference in deposition was due to the deference in samples. ' o' : o ORNL-DWG 70-~4347 dpm /g X-10, ORDINARY CG8 2.60x40' dpm fem? 76,000 Mwhr v-47, IMPREGNATED CGB 9.09x10% dpm/em2 20, 000 Mwhr 0 10 20 30 40 50 60 TO DISTANCE FROM CENTER (mils) ‘ Fig. 11.4. Effect of Exposure on Deposition of 3. 4-day 132Tc: : in Graphite, ~ Because the basket samples were exposed for 50,000 hr, as contrasted with 20,000 for the tubing samples, ~ only the short-lived isotopes are suitable for compari- son. Among the isotopes listed in Tables 11.3 and 11.4, this rules out 367-day ! °®Ru, 986-day ' 2*Sb, 284-day 144Ce, and 10,958-day '37Cs, which would be high because of the longer exposure. For the remaining isotopes, the basket samples are indeed higher, but by widely varying factors. This variation was not under- stood and seemed greater than could be ascribed to experimental errors. The tubing samples were taken at seven equally spaced locations from the top to the bottom of the core. Our notion was that turbulence might influence the deposition and that basket samples experienced the greatest turbulence. An increased depo- sition in turbulent regions had been possibly indicated by the gamma spectroscopy of the MSRE fuel circuit. Except for '3'1, the agreement between the new averages and the previous ones, as shown in Table 11.3, was deemed fair, Thus a consistent effect of not using 109 Table 11.3. Deposition of Salt-Sotuble Fission Products on Hastelloy N from the MSRE Core a The amount on 1 ex® expressed as percent of the calculated amount that would be in 1 g of fuel if it escaped only by decay. b Average from Survey 3. ¢ Basket samples B and T were ta.ken from the bot'bon and top of the perforated cylindrical basket that held the sample assembly in the core. B9y 3y Bz 131, el 1"°na Wk, Distance Dis., Percent ) Dis:l Percent ‘Dis._l Percent Dis 5 Percent Dia.l Percent Dis I Percent l):ls;l Pércent of 8auple of of & of mip of "1”-2 ot 1 of i of : Frog Bottn = g (8] T8 (») =8 ) %9 (&) =7 (s) °.B (a) 7 (a) {in) x 10 Inverntory x 10 Inventory x )0~ Ioventory X Inventary X 10 Inventory x 10" Inventory x 10 Inventory (L5 (16.2) (05.7) (15.6) (1.29) {5.91) (0.47) (8.30) (10.20) {21.8) ' (k.c1) (25-5) (1.66) (1r.3) (o.28) 133 6.02 0.45 5.08 O.b2 3.810 0.30 bo1 5.03 .T92 0.315 6.8% - 0.4s 1.7 0.19 - 21.17 6.01 0.b5 6.52 0.54 L38 0.35 6.08 T.50 836 0.15 6.20 0.0 ) 0.20 30.0 6.09 0.k6 6.23 0.52 458 0.37 5.8 T.21 1.00 0.18 10.5 0.69 12.8 0.21 39.0 6.08 0.46 L.53 0.38 3.53 0.28 5.04 6.22 .B93 ©0.16 6.08 0.k 9.13 0.15 531 b.92 0.37 1.3% 0.11 L2 0.35 5.33 6.5T .521 0,10 6.39 o0.k2 5.22 0.09 60.5 2.20 0.1T .69 0.06 .39 0.0 3.65 h.50 360 0.07 9.78 0.64 2.72 0.05 Average 5.22 406 5.2k 5.00 .5 7.63 8.91 Pl‘efi”(b) . 4.8 7.31 1T 1,56 .61 5.40 13.%4 Basket - . Sample Bt} 15.3 1.15 6.1 0.53 5.45 0.43 8.8 10.90 13.3 2.43 T.92 0.52 8.98 0.15 Barket (e) ’ . Sample T 7.57T 0.5T7 6.88 0.57 2,71 0.22 11.4 14,00 18.5 3.40 6.50 0.4l 2.57 0.04 Bagket Average 1.k 6.65 4.08 6.75 15.90 T.11 . 5.78 {a} The amount or 1 cma expreszsed as peroent of the calculated amount that would be in 1 g of fuel if it escaped only by decay. {b) Average from Burvey 3. {c) Basket samples B and T were taken from the bottom and top of the perforated cylindrical basket that held the sample lss-bly in the core. f&) This semple, 1.5 inches from the bottam, gave high results for salt soluble isotopes; . the regults from this sample were not included in the average for fear that they represented coptaminatiom, Sbmgely it iz only for apec:;eu with rare gas precursors that the values are exceptionally high. Table 11.4, Deposition of Metallic Fission Products on Hastelloy N from the MSRE Core Py o 10300 1065, ] gy 129y, 132, Diatance Dis. 1 Percent Dis. 1 Percent Pis. 1 Percent Dis. 2 Percent Dis. N Percent Dis. Percent Dis. 1 Percent D""';]_ Percent f Sampl in™ min=. win’ in n . min> , min®™ m !‘:'om ;':I'Ih)tzm :,,-2 oot 2 of cm .ot of . cmn'laa of em g of en=2 of cm'am of of core (in) x 1010 Inventory® x 1011 Taventory® x 1010 ynventory® x x 108 Toventory* x 108 tmventory® x 10° Imvemtor® % 10% entoryt x 101 Imventory®_ L5 .6T5 8 .89 69.9 1.26 8 180 52 . T8 108 2.7 B8 2.06 110 5.7 16 11.33 1.59 1B 1.8 T .6 i 9.86 29 509 T 352 121 112 - 60 8.06 O 21.17 2.5& ’ 28 2.38 191 1.62 56 23.7 68 1 6.18° 93 612 . 23 b.73 253 8.57 ™ 30.0 1.85 21 2.01. - 181 918 20 8. 3 by €3 672 23 .03 5 8.96 T8 39.0 1.29 - m . 1.8 . g 631 b 8.95 26 212 32 1.37 47 .905 48 8.9 T - 53.1 L2k 123 % 19 a7 9.12 & 3.84 58 Bz 564 30 3.56 ST 6.5 LI 13 .2 763 17 731 . 2 Lok 16 2.58 88 815 AT 9.20 8o Average 1.47 1.58 96 13.11 b3 1.67 1,61 8.93 Previoua ’ ‘ . - Averagel 8.59 2.13 1.06 9.40 0.30 - .51 T.29 Basket ‘ ) - . . ) Semple BS 9.T3 109 ] 2.98 167 1.91 i3 . sh,5 . 157 36.% . 543 1.5 393 3.61 195 Basket o ’ . 3ample T° 2.18 2k.b - L2 1% 117 39 ba.7 11 - 26.4 395 1.6 kel 2.47 B2 Basket : . . . Average 5.96 1.75 -~ 1.84 51.60 3135 12,55 3.04 110 flush salt was not found for salt-soluble products. For the salt-soluble products the tendency for deposition to - be highest where the flux was highest is apparent, and, except for !311, the amounts found were low enough to be of little importance. The '?'I count was about twice the amount that could be attributed to the tellurium precursor, based on the assumption that the total inventory of tellurium deposited on metal and remained in place during decay. With respect to the metallic fission product results in Table 11.4, a decrease in the amount of °*Nb was encountered for the same reason that was mentioned in connection with **Nb in the graphite. N In general, greater deposition is found toward the -bottom of the core than toward the top. Last time, the profile for metals was usually more uniform, but -occasional decreases toward the top were noted. Among the five cases suitable for comparing the effect of not using flush salt, ®*Mo was lower on the unflushed sample, '°*Ru and '3?Te were virtually unchanged, and ''*Ag and !2°Te were higher on the unflushed samples. 11.1.3 Fission Product Distribution in the MSRE It is not possible, from measured deposition in the center of the core, to draw any firm conclusions about deposition throughout the circuit. However, to gain some inkling about what the numbers mean, an exercise of some interest is to ask what would be the deposition throughout the reactor if the méasured samples were assumed to be’ representative. Such a hypothetical distribution is presented in Table 11.5; it is based on the fourth set of surveillance samples. The most objectionable feature of the table is the large unaccounted balance for the metallic products. The amount that can be accounted for by escape to the gas phase is now believed to be of the order of a few percent. The trouble probably lies mainly in the fact that figures for deposition of metals obtained from the Table 11.5. Distribution of Fission Products in the MSRE on June 1, 1969 i Percent of Inventory? to - sorope InSalt? On Graphite OnMetald Balance _ Salt Soluble 7 89gr 80 16 0.08 —4 oly 130 2 | 0.07 +32 %7r 100 0.2 0.05 0 131y - 60 : 0.3 0.1 -40 1370 80 07 003 -19 140, 110 _ 1.0 0.1 o+l 14406 120 01 0.03 +20 , Mgtals . °Nb Variable 7 3 ‘ 99Mo 1 4 26 -79 103Ru 6 2 4 -88 106pu 3 3 8 -86 1 ,p 7 5 16 72 1255p 7 3 12 ~78 1294, 1 3 18 -78 1327 1 5 16 - -78 %Based on calculated amount in reactor (4.336 X 108 g of fuel) and a full power of 7.25 Mw. Escape by decay only. bTypical figures from data of Compere and Bohlmann. “Based on 2 X 10® cm?® of graphite in reactor, "dBased on 7.9 X 10° cm? of Hastelloy N surface in fuel circuit. metal are very small, and, since the deposition is by fission recoil, vanishingly small amounts should be on the metal outside the core. - ' Except for °Nb, which forms a very stable carbide, - more of the metallic products are on the metal than in graphite, even though the area of the metal is smaller. Exposures for this and all the other surveillance assemblies' :*~> have been based on several different values for the full power of the reactor. Using 7.25 Mw core samples are probably lower than what is actually deposited on the heat exchanger. Other features of the table are more believable. The salt-soluble fission products are undoubtedly mainly in the salt; the spread in the numbers showing this reflects . correctly the reproducibility of the measurements involved. Also, there are only a few percent of metals in . the salt. The amounts of salt seekers in the graphite are small; the larger amounts are from species with gaseous precursors. The amounts of salt-soluble species on the as the full-power value, the figure 7800 Mwhr used for exposure at the time of removal of the first assembly becomes 7610 Mwhr. Similarly, 32,000 Mwhr for the second removal becomes 32,700, and 64,000 Mwhr for the third becomes 65,600 Mwhr; 76,000 Mwhr ex- posure on graphite that was in the reactor for three _ 38, 8. Kirslis and F. F. Blankenship, MSR Program Semiann. Progr. Rept. Aug. 31, 1968, ORNL-4344, p. 115, *S. 8. Kirslis and F. F. Blankenship, MSR Program Semiann. Progr Rept. Aug. 31, 1967, ORNL-4191, p. 121. SW. R, Grimes and R. E. Thoma, MSR Program Semiann, Progr. Rept. Feb, 28, 1967, ORNL-4119, p. 124, surveillance cycles becomes 76,900 Mwhr. For the ~exposure during the fourth cycle, 20,000 Mwhr be- comes 18,500 Mwhr, Before this semiannual, reactor fission product inventories were frequently based on 8 Mw full-power operation, which accounts for slight differences from earlier figures for percents of inven- tory. Such differences are thhm the experimental error. 11.2 FISSION PRODUCT DISTRIBUTION IN MSRE PUMP BOWL SAMPLES E.L.Compere E.G. Bbhlmami MSRE run 19 extended from August 11 to November 2, 1969, including over four days of flush salt opera- tion, During this period we caught (for fission product analysis) 2 flush salt and 13 fuel salt samples and 23 gas samples (including 1 postdrain sample and 3 controls which were sealed at the upper end of the nozzle regxon) 2 of the gas sample capsules had fine metal filters over the nozzles. Run 20 extended from November 26 to December 12, 1969. Two salt samples and four gas samples (with nozzle filters) were obtained during this run; one of these was taken after the fuel was drained, and one was reserved for recovery of cesium isotopes by J. J. Pinajian of the Isotopes Division. Data from the samples from runs 19 and 20 are presented below. Sample Capsules. — Sample capsules were of the double-wall type! to minimize contamination, a par- ticularly important consideration for gas samples. A sketch of the capsule is given in Fig. 11.5, The salt sample capsule volume was 7. 8 cm?; for gas - samples, volumes of 15.0 and 13.8 cm® were more frequently used. Sampling Procedure. — The capsules for salt samples were lowered into the pump bowl spray baffle so that . the nozzle was as far as possible below the salt surface. " It remained in salt for 60 min, more than five times as 111 long as bench tests showed was required to soften the solder holding the plug. The capsule was then lifted 18 in. which brought it into the sampler entry tube, and was held there for 20 min before proceeding with removal. During the entire sampling interval, a purge’ flow of helium amounting to about 575 std cm®/min ~ was maintained down the sample transfer tube. lE L. Compere and E G. Bohlmann, Fission Product Behavior in the MSRE During TM-2753, ORNL-DWG 70-€758 @] CUT FOR SAMPLE REMOVAL == L-——NICKEL CUTER TUBE —n: > VACUUM COPPER INNER CAPSULE | —BALL RETAINER - = BALL WITH SOFT SOLDER SEAL '[/—COPPER NOZZLE TUBE J Fig. 11.5. Double-Wall Freeze Valve Capsule. ABRADE FOR SAMPLE REMOVAL For gas samples in runs 19 and 20 the sample capsule was normally lowered to a depth that brought the nozzle to about 2 in. above the liquid surface. Usually, a period of 20 to 40 min was allowed in the pump bowl. The purge flow of helium down the sample ‘transfer tube was held to ~75 std ¢m®/min for all of the gas samples but the first four, After transfer to the High Radiation Level Analytical Laboratory, the sample tube was examined and - weighed, and the nozzle tip was abraded until the inner capsule was free. For samples with a filter over the nozzle, the filter was first cut off with a tubing cutter. Then, after the inner capsule was free, the upper part of the outer nickel tube was cut off using a tubing cutter and the inner capsule dumped into a new dissolver. For gas samples in run 19, after the inner capsule was caught in a plastic bag, the nozzle part of the capsule . was sawed off and analyzed separately. 3U Operation ORNL- The outer nickel tube was dissolved for analysis, sometimes after one or more prior acid leaches. The inner capsule and nozzle were dissolved in 2 N HN03 with suitable traps. Analyses, — Analysis for salt constituent elements included lithium and beryllium by spectrographic 1. salt-seeking elements: fuel was used. However, methods, uranium (and zirconium on salt samples) by wet chemical methods, and 233U by neutron activa- tion—delayed neutron counting techniques. Radiochemical analyses for a number of isotopes representing several classes of fission products were obtained. These included: 85Zr, *41Ce, 14%Ce, and 147Nd. 2. daughters of rare gases: ®°Sr, 'Y, !'37Cs, and 140p,. | 3. noble metals: **Nb (from °%Zr, *°Mo, l"3Ru, 106Ry, and ! 1! Ag; 4. tellurium-related: 12%8b, 2°Te, '*?Te, and '3'I. Data for these 1sotopes were corrected to the time of sample removal; ®$Nb was first corrected for ingrowth from ®*Zr in the sample after removal. 112 necessary, and none was made, for fuel burnup or the addition of plutonium fuel. (The yields of 1°¢Ru and a few other isotopes would be somewhat increased by the added plutonium.) In order to obtain a sultably continuous and detailed . ~ power history of the MSRE, during 233U operation the ' Data pertaining to salt were generally expressed as amount (disintegrations per minute, micrograms etc.) per gram of salt, and data for gas samples could be expressed as amount per cubic centimeter of sample volume. In this way the values could be compared with appropriate inventory values or production rates calcu- lated from the associated MSRE operations. Inventory. — The inventory value for a given isotope is the total amount of the isotope that existed in the circulating fuel at a given time as a result of the power history (and drains) of the MSRE divided by the amount of fuel salt circulated. It accounts for all production and decay in the fuel since the first power operation with 235U in January 1966 and assumes no losses. However, the inventory of niobium was arbi- trarily set at zero at the end of the chemical reduction operations that were part of the August 1968 fuel reprocessing; subsequent sample data established the different levels of operating power were recovered daily from fission chamber charts which were calibrated from time to time by heat balance calculations. Since no appreciable or long-term deviation developed between fission chamber and heat balance values, no adjustment for such deviation was made. However, the inventory values are proportional to the nominal full power of the MSRE, and adjustments changing full power to 7.25. Mw from 8 Mw have been made where suitable. In comparing observed isotope data with inventory, it should be recalled that the part of operating power history reflected in the inventory of a given isotope depends on the decay rate of the isotope in question; the inventory for an isotope such as 3.25-day !32Te reflects only the most recent several days, while that of 30-year '37Cs is determined by the entire history since startup. If an observed dwmtegratxon rate per gram (dlS min~ g™!) is divided by inventory (dis min ~* g™"), the ratio is dimensionless and indicates what fraction of the average possible amount is present. If the mass of sample is not available, dividing observed values by inventory values provides an indica- tion of the corresponding number of grams of “inven- tory” salt. This is most useful with salt-seeking isotopes (and constituent elements). - Gas samples require a different basis of comparison. Assuming that no accumulation occurs, the maximum ‘steady number of atoms that could be transported in a validity of this assumption. Inventory per gram is thus the amount of an isotope or element that should be " found in fuel salt if no losses occurred. - Most isotopes were treated in this calculation using a single-decay model, ignoring all precursors, without introducing appreciable error. For the daughter isotopes in the ?!'Sr'Y, couples, the two-element parent-daughter decay model ‘was used.? Fission yields of 23sU were used for operation with that fuel and were changed appropriately when 233U no adjustment appeared - 2E, L. Compere and E. G. Bohlmann, MSR Program Semthfin Progr. Rept, Feb, 28, 1969, ORNL-4396, pp. 138—40. given volume Qf gas would be given by MSRE fission rate times fission yield of chain divided by the MSRE pump bowl purge flow rate. This can be multiplied by 9SZr9$Nb and 147Nd 147Pm | the isotope decay constant to obtain production in disintegrations per minute per cubic centimeter. When divided into an observed amount, if a valid sample was taken, the result would indicate what fraction of production was gas-borne in the pump bowl gas. In the case of rare-gas fission products and their daughters, it is possible to estimate the maximum amount that might be transferred as gas into the pump bowl. Comparison with observed values can help estab- lish whether valid samples of pump bowl gas were taken. A ' mathematical model perrmtted calculatlon of steady-state values for a number of pertinent isotopes, This model took into account the flow around the circuit, the production and decay of rare-gas precursor and rare gas, the liquid and gas flows into the pump bowl, and decay of the rare gas in the pump bowl. Slug flow was assumed in all regions, except the pump bowl gas was regarded as being completely mixed. Complete - stripping of all rare gas from the liquid entering the pump bowl was assumed. A calculation involving about 90 iterative cycles gave steady values, Typical results are shown in Table 11.6. It may be seen that for the 89 chain, only 96% of the 89Gr has a 3%Kr precursor, and of this, under typical conditions, 61% of the original chain decays before it enters the pump bowl and 35% of the original chain enters the pump bowl. The holdup in the pump bowl permits 21% of the original chain to decay to ®*?Rb atoms, which doubtless leave the gas phase as soon as a surface is encountered, whether salt, mist, or metal. The average 2 Kr content of the well-mixed pump bowl gas and that of the effluent off-gas are the same, amounting in this case to 14% of the original chain, Thus, except for salt mist or deposited or entrained ®?Rb or 27 Sr atoms produced by 3?Kr decay in the pump bowl, the highest value we could expect for ®?Sr in the pump . bowl gas is 14% of production; !37Cs is similar. The isotopes *!Y and !*°Ba, which have rare-gas precursors of shorter half-life and also about 40% of the chain yield produced by direct yield of rare-gas daughters, can have only 0.07-0.16% of production in this way; and 141Ce is essentially negligible as a gas-borne spécies. Salt Samples. — Thirteen salt samples were obtained with the new double-capsule-type sampler during MSRE runs 19 and 20.! Results of the radiochemical and - compositional analyses carried out on these samples are - summarized in Table 11.7. The data are presented in Table 11.6. Percentage of Given Chain Involving " Rare Gas Which Decays in Various Regions? Chain 89 137 91 140 141 Isotope counted Sr. Cs Y Ba Ce " Rare-gas precursor 4 23 half-life, sec ' Rare-gas half-life, sec 191 234 10 16 2 Percent decay: ' ' - As rare-gas precursor 4(4) 2 41 40 177 In salt . 61(74) 58 S8 57 23 In pump bowl 21(15) 22 18 28 0.06 Inexit gas . ‘147) 18 0.07 0.16 0.0004 aValues correspond to 1210°F, 1200 gpm loop flow, 65 gpm spray and fountain flow to pump bowl, 60% salt level in pump bowl, 3.2 std liters/min He flow, 5 psig. Values in parentheses for 89 chain illustrate 616 gpm, 2.3 std liters/min He flow, 33 gpm flow to pump bowl. - terms of concentration, mg/g and dis min™ g™, and the ratio of the amount found to the calculated inventory. The salt constituents inventories were based on the recorded additions to the system, and the radioactive isotopes inventories on the 222U yields and conventional equations involving power history and decay; full power was taken to be 7.25 Mw. Since the data sometimes scatter considerably, the medians and +25 percentile ranges are given at the bottom of the table. Evaluations of these and other data are continu- ing, but the following observations can be made at this - time. o : Salt Constituents. — In order to minimize contamina- tion the inner copper capsule containing the salt sample was dumped directly from the opened outer capsule into a new glass dissolver as soon as it was free of the outer capsule, with no contact with manipulators or ‘other cell equipment. This meant that the solution analyzed contained substantial amounts of copper ion, interference from which probably accounts for the relatively low values obtained for the uranium concen- tration by coulometric titration. Except for the 233U by activation analysis our data here scatter rather more than the results from ladle samples; this is, at least in part, attributable to the presence of the copper ion and a less precise determination of the sample weight: different balances were used outside of and in hot cell, .sample weight is the difference between two relatively large numbers, material on the outer capsule would count as sample weight. Such considerations, however, do not satisfactorily explain the considerable scatter and biases shown here for constituents and fission products. The analyses were run in duplicate and redone when reasonable (+10%) checks were not obtained. Probably the explanation is that the results ~ represent the state of the art when dealing with large numbers of extremely radioactive samples of various kinds and origins under hot-cell conditions. This state- ment is in no way meant to impugn the knowledgeable, interested, and cooperative efforts of the Analytical - Chemistry Division personnel who carried out the analyses. Salt-Seeking Isotopes. — The ratios to mventory for these species scatter about 1.0, as they should. How- ever, only the ?5Zr data are reasonably consistent and give the right level for the median of the middle 50 percentile group of ratios to inventory. Quite aberrant values were also obtained even for this isotope for samples 19-9 and 19-76. Noble Metal Isotopes, — The so-called noble metal isotopes behaved unusually in runs 19 and 20. During earlier operations these isotopes, with the exception of 114 ®>Nb as described previously,” were found in salt samples at concentrations equivalent to fractions of 1% of inventory.® In runs 19 and 20, however, the noble - metal concentrations in salt samples are often in the range 10 to 90% of inventory. Two points should be emphasized regardmg the noble metal group, as shown in Fig. 11.6. The observed activity values for the various isotopes in a sample remdin in.a reasonably constant ratio to each other from sample to sample when expressed in terms of inventory, and the values of these ratios to inventory . very seldom exceed unity in spite of the fact that the total amount present varies by two orders of magnitude from sample to sample. 3S. S. Kirslis and F. F. Blankenship, MSR Program Semiann, . Progr. Rept. Aug. 31, 1968, p. 136, Table 11.4. ' 103 'PERCENT OF INVENTORY IN SALT SAMPLES 1072 2 4 12 19 44 45 46 9 24 36 - SAMPLE NUMBER To the extent that the different. isotopes are in constant proportion from sample to sample; it is indicated that the samples carry material from a well-mixed pool of these substances. Since these rela- tionships hold for isotopes with half-lives ranging between 2.7 days for ®®Mo and 367 days for ' °°Ru, it would seem that the noble metals in the sample reflect a long period of power history, rather than bemg just from most recent production. The secc'md point argues that the samples reflect what is borne by salt rather than what is picked up as a result of some localization in the pump bowl or sample . region; else values should exceed unity rather readily. It would seem to be more than a conceivable coincidence to have such a preponderance of the values | for percent inventory found to be in 2 realistic range. The consistently reasonable -results for the several fission products, in spite of the wide variations in the ORNL-DWG 70-5112 9 42 44 41 55 57 58 59 76 1 19 Fig. 11.6. | Noble Metal Isotopes in Salt Samples (as Percentage of Calculated Inventory). 115 hailf-lives, ‘also sufiport the a'tg'ilment.‘that the s'amples ’ are “real.” A somewhat disturbing point in this regard, however, is the wide fluctuation in concentration observed in samples taken within a short span of time. Thus samples 57 through 59 were all taken over a period of 6% hr, and the noble metal concentrations vary between one and two orders of magnitude. No - satisfactory correlation of noble metal concentration and any operating parameter is apparent thus far. These - results suggest the following about the noble metals in the MSRE, 1. The bulk of the noble metals remain accessible in the circulating loop, but with widely varying amounts in circulation at any particular time. - 2. In spite of this wide variation in the total amount found in a particular sample, the proportional composition is relatively constant, indicating that the entire inventory is in substantial equlhbnum with the new material being produced. 3. The mobility of the pool of noble metal material suggests that deposits occur as an accumulation of finely divided, well-mixed matenal rather than as a “plate.” ‘ Examinations of the data are continuing. Of most . recent interest is the use of the ruthenium and tellurium isotope pairs to study the history and behavior of this material. Preliminary use of this tool, however, has revealed the necessity for refining our inventory calcu-. lations and rechecking obvious discrepancies in the analytical results. Niobium-95. — The analyses indicate that appreciable . percentages of the *SNb were dissolved in the salt in samples 199, 19-24, and 20-1. The above observations - on the noble metals in general raise a question about such .a conclusion being justified in all cases, however. Thus, in samples 19-24 and 20-1 comparable per- centages of all the noble metals were present in the salt. As a consequence the ®SNb present may have been associated with the other noble metals rather than being in solution. In view of this possibility, a review of the niobjum data will have to be carried out. A complica- tion in this regard is the fact that the *SNb is the daughter of ®%Zr, which is present in large quantities in sffifibing ‘of the *°Kr and '*"Xe pre'curso“rs‘ in the pump bowl. The general scatter of the results for these isotopes, however, suggests that this may be coinci- dental. Gas Samples. — Data for gas samples from runs 19 and 20 are shown in Table 11.8. In this table, the total amount (as disintegrations per minute) for a given isotope is shown for capsule and nozzle regions. The ratio of each value to salt inventory is also shown, indicating in micrograms the amount of inventory salt that would correspond to the observed activity., Examination of wvalues for salt-seeking isotopes implies that significant quantities of salt are found in both nozzle and capsule regions. The values for the various salt-seeking isotopes in a given region of a sample are similar, which is consistent with an entrain- ment of salt mist. The amounts of such mist in each region of each sample were estimated by averaging the microgram values indicated by appropriate salt-seeking isotopes and constituent elements. Comparison with other isotopes in the given region of the sample indicated quite generally that the gas samples were considerably enriched, relative to salt, in daughters of rare gases, noble metals, and tellurium-iodine isotopes, all of which are missing to some extent from subsurface salt samples. - However, gas data must be examined in terms of the amounts of material per unit volume of gas, both including and excluding any contribution of salt mist. - Such data should then be compared with MSRE production rate - per unit- purge gas flow rate to determine the fraction of production indicated to be transported into the off-gas system, This was done by taking the amounts of each nuclide found in the nozzle and inner chamber, dividing by capsule volume, and comparing with production per unit volume of off-gas. The range and average of such gross values are shown in appropriate columns of Table 11.9. It is also of interest to see what substances, if any, were to any extent transported independently of salt . mist, possibly as gases. To do this we have noted that salt samples, necessitating large corrections in the *>Nb results to arrive at the actual amount present at sample time. As can be seen in the table, this often leads to negative values, Isotopes with Noble-Gas Precursors — The slightly low median ratios to inventory for °Sr and !37Cs are in surprising agreement with the expected (10 to 20%) the volume of the nozzle tube was negligible. We thereby assumed that the nozzle deposits were largely mist and assign the composition of the nozzle region to capsule mist substance. The amount of a given isotope ~ assigned to the capsule mist is consequently determined by the amount in the nozzle region of the sample and’ by the relative weights of “salt” in capsule and nozzle regions, Such salt weights were averages of values for this region of the sample, as indicated by appropriate salt-seeking isotopes and constituent elements. i Table 11.7. Fission Product and Constituent Afialyses from MSRE Salt Samples — Runs 19 and 20 Weight of Salt Constituent Elements® Salt-Seeking Isotopes® Noble Metal Isotopes 'Teliurium-lodiné Isotopes Isotopes with Noble Gas Precursors Sample D.ate Operation? SamPle - 233ye ud Li Be Zr 9S4, 1410, 1440, ‘1”Nd 9%Nb Mo - luAg 103p, - 106p, 1324, 1299, 131 89g, 1375, 9y 140p, No. (Mwhr) Obtained 605 709 461 198 605 4.8 0.02 18 0.24 44 072 29 586 658 557 54 o ® 65. 33. 284, 11.1 35. 2.79 1.5 39.6 367. 3.25 34, 8.05 52. 10958. 58.8 128 . . ‘ : - : I 19-9 8/18/69 18,405 10.20 6.5E+00 4.1E+00 7.7E+01 5.2E+01 64E+01 3.7E+10 2.0E+10 4.5E+10 7.5E+07 2.6E+10 8.5E+06 <3.3E+06 3.1E+10 2.8E+09 1.8E+10 1.8E+09 _ 0978 0512 0.67 0.78 0.557 0.797 0.662 1.01 '0.203 0.378 0.00087 <0.00113 . 0.773 0569 0434 0936 1924 9/10/69 20,169 1.79 _7.1E+00' 5.3E+00 9.5E+01 7.2E+01 1.3E+02 5.2E+10 4.8E+10 5.0E+10 3.0E+09 2.2E+10 1.8E+10 23E+08 3.8E+09 1.2E+09 1.8E+10 2.2E+09 2.5E+10 5.0E+10 4.7E+09 7.6E+10 1.2E+10 1.06 0652 0.826 1.07 1.13 0.948 0.825 1.1 0.131 0.364 0.243 0.69 0.255 0.461 0.261 0375 0657 0978 0935 1.51 0.209 19-36 9/29/69 21,594 14.02 6.3E+00 64E+00 8.7E+01 6.1E+01 12E+02 5.7E+10 5.9E+10 5.SE+10 2.7E+10 —5.8E+09 1.SE+09 4.2E+07 4.7E+(7 4 4E+08 1.2E+07 2.5E+10 5.2E+10 6.2E+09 6.0E+10 7.1E+10 ' 0.94 0.796 0.75 0916 1. 0.947 0.843 1.18 1.04 -0.0951 0.0211 0.127 0.00266 . 0.00678 0.00169 0.636 0.905 1.22 1.1 1.03 1942 10/3/69 22,104 2.17 7.3E+00 3.7E+00 9.3E+01 7.1E+01 1.5E+02 6.5E+10 7.3E+10 S5.8E+10 4.1E+10 —4.1E+09 4.8E+10 9.0E+07 3.1E+09 4.8FE+08 9.6E+09 7.2E+08 3.5E+10 5.9E+10 7.7E+09 7.2E+10 B.6E+10 1.09 0462 0.808 1.06 1.3 1.03 0.928 1.24 -1.34 -0.0675 0475 0.223 0.163 0.175 0.108 0.0917 0.72 0948 1.5 1.24 1.08 1944 10/6/69 22,610 498 6.9E+00 4.9E+00 9.TE+01 6.8E+01 1.6E+02 7.5E+10 7.8E+10 S5.8E+10 4.7E+10 —3.5E+09 S5.8E+10 3.2E+08 1.2E+09 0.1E+07 2.3E+09 5.7E+10 5.7E+10 4.6E+08 8.2E+10 1.0E+11 ) 1.04 0611 0.842 1.02 1.36 1.12 0.903 1.21 1.3 -0.0581 0.455 0658 0.0547 0.0259 0.02 0983 0.86 0.0898 1.23 1.09 1947 10/7/69 22,828 2.11 7.2E+00. 5.2E+00 1.1E+02 6.1E+01 1.3E+02 7.5E+10 8.6E+10 6.1E+10 5.1E+10 1.8E+09 1.3E+11 5.9E+)7 4.1E+09 29FE+08 8.5E+09 9.0E+08 7.4E+09 6.JE+10 4.5E+09 8.7E+10 1.1E+11 1.08 0649 0964 0918 1.12 1.09 0.949 1.27 '1.33 0.029 0915 0.112 0.186 0.105 0.0702 0.0982 0.119 0981 0.878 1.37 '1.08 19-55 10/14/69 24,144 429 7.6E+00 6.8E+00 1.2E+03 6.8E+01 1.3E+02' 8.3E+10 1.1E+11 6.7E+10 7.2E+10 1.1E+11 1.2E+06 6.6E+09 7.2E+08 3.8E+10 6.8E+10 4.0E+09 1.5E+11 ' . t.14 0.848 10.7 1.02 1.15 1.05 0.996 134 148 0.691 0.00178 0.0464 0.065 049 0.849 = 0.76 1.23 1957 10/17/69 24,616 3.35 7.1E+00 7A4E+00 14E+02 6.8E+01 1.6E+02 84E+10 1.1E+11 6.5E+10 6.8E+10 3.5E+09 1.3E+11 5.8E+07 9.SE+09 6.0E+08 1.2E+10 1.8E+09 4.8E+10 6.9E+10 5.0E+10 9.5E+10 1.6E+11 . , 1.07 0.918 1.2 1.01 1.39 1.03 0.958 .29 135 0.0557 0.833 0.0856 0.341 0.211 0.088 0.149 0595 0.824 9.36 1.24 1.21 19-58 - 10/17/69 24,629 12.38 6.8E+00 5.8E+00 9.9E+01 6.5E+01 1.3E+02 8.3E+10 1.0E+11 6.5E+10 7.5E+10 —24E+09 3.1E+09 1.2E+08 1.2E+07 2.3E+08 5.6E+10 6.7E+10 S5.1E+0% 99E+10 1.5E+11 1.02 0.716 0.86 0.969 1.1 1.02 0.898 1.3 1.5 -0.0371 0.0208 0.00448 0.00409 0.00167 0.706 - 0.805 0.953 1.29 1,14 19-59 10/17/69 24,629 8.50 6.7E+00 5.9E+00 1.1E+02 6.6E+01 1.3E+02 B8.1E+10 1.0E+11 5.9E+10 6.4E+10 —9.5E+08 3.0E+10 L8E+07 1.1E+09 84E+07 1.0E+09 14E+08 13E+10 6.0E+10 4.7E+09 1.9E+11 1.6E+11 0999 0.736 - 0927 0985 1.13 0.99 0.872 1.17 128 -0.0149 0.206 0.0277 0.0399 0.0292 0.0075) 0.0118 0.0163 0.712 0.879 246 1.21 19-76 10/30/69 26,866 1348 7.3E+00 6.4E+00 1.1E+02 6.2E+01 7.6E+01 7.9E+10 1.3E+11 6.8E+10 7.3E+10 2.3E+09 2.3E+09 8.0E+07 4.7E+07 3.8E+10 B.0E+10 S5.9E+09 7.6E+10 1.7E+ii . 1.09 0.791 0939 0934 0655 0.821 0.934 1.28 : 1.29 0.033 0.0156 0.00238 0.00034 0.444 0.807 1.09 0.839. 1.15 20-1 11/26/69 - 27473 3.11 T.1E+00 7.4E+00 1.0E+02 6.9E+01 1.1E+02 7.3E+10 6.1E+10 6.2E+10 1.5E+10 4.7E+10 S$.2E+06 7.5E+09 7.8E+09 6.4E+09 9.5E+09 4.8E+10 4.5E+09 9.4E+10 4.8E+10 " 1.07 0922 0.871 1.03 0.957 .01 0.743 1.24 i 0.205 3.69 0.0544 0.353 0.277 .0.604 0.748 0.69 0.827 143 1.18 20-19 12/5/69 28,934 548 ‘ 7.3E4+00 8.8E+01 6.SE+01 1.7E+02 7.6E+10 B.0E+10 6.0E+10 Z 4 4E+09 11E+11 17E+09 3.5E+0% 29E+08 2.6E+10 24E+10 4.8E+10 4.9E+09 8.8E+10 0907 0.759 0978 1.47 0.936 0.765 1.17 0.0595 0.862 3.7 0.135 0.1 0.228 0449 0.587 0.875 1.01 Ratio to inventory : ' First quartile 1.01 0639 0796 093 0989 0944 0.810 117 110 -0.0476 0.114 0.0544 0.00357 0.0284 0.00714 00251 0446 0.758 0.810 1.13 1.02 Median 1.06 0.736 0.86 = 0985 1.13 1.01 0.898 1.24 1.30 0031 0465 0.119 0.0949 0.105 0.0583 00917 0615 0.824 0.879 1.29 1.09 Third quartile 1.08 0.863 0,945 1.02 . 131 1.03 0.938 1.28 1.35 0.132 0.847 0658 0.22 0.227 0.168 0.136 0.713 0916 1.12 141 1.18 “Cumulative Mwhr operation on 233y; full power = 7.25 Mw, bThe first value represents concentration (mg/g) and the second is the ratio to calculated inventory. - CBy activation analysis, dCoulometric titration. i €The first value represents concentration (dis min -1 g'l) and the second is the ratio to calculated inventory. Full power assumed, 7.25 Mw; data represent: activity in total sample region, dis/min; ratio to inventbry salt, 1 Table 11.8. Fission Product and Constituent Analyses for MSRE Gas Samples from Runs 19 and 20 Operating Conditions Isotopes with Noble Gas Precursors Noble Metal Isotopes Antimony-Tellurium-lodine Isotopes Sample Dat Mwhe " oM Pnrge PSIG Sample 895‘ 137C‘ 91Y MOB. 95Nb 99Mo l“AS !OSR“ ) IOGRu lszb 132.“ 129Tc - W . - . . No. ate v (liters/min) Volume (c¢) Capsule Nozzle Capsule Nozzle Capsule -Nozzle Capsule Nijzzle Capsule Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzle 19-13 - 821 18406 ©0.007 - 1176 83HE 5.2 78 33E+07 1.0E+07 2.8E+06 3.0E+06 9.6E+05 3.4E+06 1.1E+08 6.1E+07 7.4E+07 1.9E+07 19EH07 5.7E+06 3.2E+06 5.0E+06 5.6E+06 94E+06 J . 840 260 570 620 24 85 1,600 910 8000 2,100 7,300 2,200 45,000 71,000 1,800 3,000 J 19-14 . 8/21 18406 0007 1176 82HE 54 15.0 3.9E+07 4.6E+07 34E+07 1.8E+07 1.7E+06 1.2E+07 14E+06 5.3E+08 1.1E+08 5.3E+06 4.0E+06 1.7E+08 23E+07 4.8E+07 5.8E+06 14E407 1.9E+07 1.8E+07 2.1E+07 ‘ 1,000 1,200 7,000 3,800 43 290 860 7,800 1,700 61,000 46,000 18,000 2,500 18,000 2,200 190,000 270,000 5,800 6,800 1915 8/21 18406 0.007 = 1176 83HE . 52 78 1.8E+07 1.9E+08 7.4E+06 8.2E+06 2.8E+06 2.1E+07 1.3E+09 5.9E+08 1.4E+07 5.9E+07 2.0E+08 1.8E+07 S5.7E+07 S.SE+07 1.9E+06 3.3E+07 4.5E+06 S5.7E+07 . © 470 4900 1,500 1,700 70 540 19,000 8,800 160,000 6,400 21,000 6,800 22,000 230,000 25,000 440,000 1,500 19,000 19-16 8/21 18406 0.007 1176 83HE 5.1 15.0 9.8E+06 1.1E+08 3.3E+06 2.1E+07 7.SE+05 6.0E+06 2.6E+0S 2.3E+08 2.5E+08 1.8E+06 1.1E+07 3.0E+07 6.0E+07 B.8E+06 1.3E+07 1.2E+06 2.9E+06 S3E+07 4.8E+06 4.1E+07 2 250 2,800 680 4,400 19 150 160 3400 3,700 20,000 120,000 ‘ 3,200 6,500 3,300 5,000 5,000 38,000 630,000 1,600 14,000 1919 9/4 19,536 5. 1110 6.0AR 5.3 18 2.1E+08 1.8E+07 L.7E+08 6.4E+07 3.7E+06 6.1E+06 6.1E+06 8.3E+07 2.2E+07 7.0E+08 S.SE+08 4.4E+06 14E+06 5.7E+07 1.9E+07 4.1E+06 2.3E+06 1.6E+0S . 1.8E+08 6.SE+08 1.5E+07 7.SE+06 1 4,500 380 35,000 13,000 & 82 130 140 1,300 350 9,800 8,000 17,000 - 5600 4400 1,500 1,500 850 630 2,800 10,000 3,100 1,500 1920 9/4 1953 5. 1110 S9AR 5.5 150 ' 19E+07 LSE+07 34E+07 2.1E+07 3.6E+06 6.3E+06 1.9E+06 2.5E406 2.3E+07 7.2E+08 1.0E+09 2.8E+06 2.6E+06 4.9E+07 24E+07 8.1E+06 2.0E+06 . 2.8E+08 1.2E+09 64E+06 1.5E+07 4 . ‘ 400 330 6,800 4,200 79 140 44 41 370 9,800 14,000 11,000 10,000 3,800 1,800 3,000 730 4400 19,000 1,300 - 3,000 1923 9/10 20,169 0009 1176 . 6.0AR 5.1 150 43E+07 2.7E+07 2.0E+08 8.3E+06 1.1E+07 4.0E+07 9.1E+07 9.5E407 3.3E+07 2.1E+09 3.8E+08 2.5E+06 5.2E+05 8.7E+08 2.5E+07 7.9E+07 2.3E+06 7.5E+08 1.9E+08 2.8E+07 6.0E+06 J ‘ 830 520 39,000 1,700 230 790 1,500 1,500 530 27,000 4,700 7400 1500 57,000 1,600 29,000 850 10,000 2,700 4,800 1,000 1928 9/23 21,240 . 1199 - 6.1HE 5.0 15.0 1.3E409 2.SE+07. 1.2E+07 3.2E+06' 6.6E+06 5.0E+06 3.9E+07 3.6E+08 2.5E+08 1.6E+10 1.5E+10 1.7E+07 2.0E+07 6.0E+08 4.3E+08 29E+07 2.0E+07 3.2E+09 6.8E+09 9.0E+07 1.3E+08 9 oo = 23,000 440 2,300 640 120 91 550 5900 4,200 190,000 180,000 48,000 53,000 34,000 25000 11,000 7,400 43,000 90,000 13,000 19,000 1929C 9/23 21,259 5. 1199 6.1HE. 5.0 29E+07 2.8E+07 3.5E+06 2.1E+06 1.2E+06 5.6E+06 3.6E+06 1.8E+07 2.8E4+07 14E+09 1.1E+09 2.3E+06 3.5E+06 8.9E+07 4.9E+07 SAE+06 3.6E+06 2.9E+08 4.0E+09 1.8E+07 64E+07 J ‘ , - 500 480 690 410 22 100 50 300 460 17,000 13,000 6200 9,600 5,100 2800 2,000 1,300 3,900 53,000 2,500 9,000 1937 9/30 21,692 0.009 617 64HE 45 15.0 53E+07 6.2E+07 6.4E+07 2.6E+07 4.7E+06 23E+06 5.2E+06 3.7E+08 2.5E+08 S6E+09 3.2E+09 53E+06 5.5E+06 3.6E+08 2.3E+08 24E+07 2.9E+07 1.6E+09 6.3E+08 1.2E+08 7.2E+07 1 900 1,100 13,000 5,200 85 41 75 6,200 4,100 77,000 45,000 - 16,000 16,000 20,000 13,000 8,800 11,000 24,000 - 9,600 16,000 9,900 1938 10/1 21,773 . 616 SSHE 5.9 150 - 3.6E+08 2.8E+06 3.9E+06 6.4E+0S 4.1E+0S 14E+05 7.1E+06 28E+06 34E+06 26E+08 7.1E+07 1.1E+06 3.7E+06 3.8E+07 6.8E+D6 24E+06 4.9E+0S g 14E+08 2.6E+08 7.6E+06 7.6E+06 J ‘ 6,100 47 770 120 7 3 100 47 56 3,300 920 3,200 11,000 2,100 380 900 180 2,100 3,800 1,000 1,000 1941 10/3 22,069 63 1186 © 64HE 40 15.0 6.8E+07 2.6E+07 2.2E+07 3.5E+06 2.8E+06 2.5E+06 1.1E+07 24E+08 2.5E+08 9.7E+09 7.5E+08 9.0E+06 1.7E+06 4.SE+07 7.5E+07 3.1E+07 S$.8E+06" 4.5E+09 2.2E+09 14E+08 4.9E+07 § : . . 1,100 420 4,300 680 49 43 150 4,000 4,100 99,000 7,600 23,000 4200 - 2400 3,900 11,000 2,100. 52,000 26,000 17,000 - 6,300 1946 10/7 22,795 1.2 1187 83HE 52 15.0 2.1E+09 14E+08 2.3E+07 7.0E+06 3.2E+07 3.8E+07 14E+08 2.2E+08 1.5E+08 1.0E+10 3.7E+09 .3.0E+07 4.5E+07 6.0E+08 2.7E+08 3.2E+07 1.6E+07 4.5E+09 3.3E+10 1.4E+08 5.6E+08 1 31,000 2,100 4,500 1,400 510 600 1,400 3,600 - 2400 77,000 28,000 58,000 87,000 27,000 12,000 12,000 5900 38,000 270,000 15,000 61,000 19-54 10/14 24,107 7.2 - 1196 83HE 5.1 15.0 3.3E+07 7.8E+06 3.2E+06 S5.2E+06 1.9E+06 5.7E+05 2.2E+06 3.3E407 93E+06 3.1E+08 1.1E+08 1.0E+06 23E+05 7.2E+07 1.1E+07 8.3E+06 8.9E+06 44E+07 1.0E+07 1.1E4+07 2.5E+06 1 410 97 610 980 26 8 18 §30 150 2,000 680 1,600 360 2,700 440 2900 3,100 310 71 1,000 230 19-56 10/15 24,307 0.009 1196 82HE 5.7 15.0 4.6E+06 1.2E+07 S.1E+06 14E+06 14E+06 3.1E+06 3.3E+06 1.2E+07 44E+07 1.1E+09 1.6E+09 7.2E+05 8.1E+05 6.2E+07 4.SE+07 39E+06 2.7E+06 1.SE+08 1.1E+09 6.1E+06 3.5E+07 87 150 960 270 18 42 26 180 710 7,100 10,000 1,100 1,200 2,300 1,700 1,400 930 1,100 8,100 540 3,100 1962 10/22 25441 7.2 1197 82HE. 53 15.0 3.6E+08 2.1E+07 6.7E+07 3.7E+06 2.2E+06 4.7E+06 1.8E+07 3.9E+07 1.2E+07 15E+09 8.0E+08 1.6E+06 7.1E+05 7.9E+07 24E+07 4.8E+08 1.3E+06 3.1E+08 23E+09 1.2E+07 4.2E+07 . : 4,000 230 13,000 690 27 57 130 590 190 10,000 5400 - 2400 1,000 2600 - 800 1,700 440 2,300 17,000 %0 3,300 1964C 10/22 .25,540 7.2 1199 83HE 5.1 3.0E+06 3.0E+07 8.8E+06 1.3E+07 6.7E+04 14E+06 1.8E+05 8.3E+06 4.8E+06 1.0E+08 S.6E+08 8.6E+05 2.7E+06 1.0E+07 1.3E+07 8.2E+05 7.4E+0S 5.9E+07 5.9E+09 2.5E+06 9.8E+07 s 33 340 - 1,600 2,500 1 17 1 130 72 690 3,800 1,200 3,900 ' 330 430 280 250 430 © 43,000 200 7,600 { 1965 10/23 25646 7.2 1196 8OHE 55 15.0 7.1E+08 1.6E+07 1.9E+07 4.1E+06 9.6E+06 6.5E+05 6.9E+07 4.3E+06 5.SE+06 2.6E+08 3.SE+08 1.5E+05 10E+06 1.0E+07 7.1E+06 9.4E+05 5.3E+06 2.8E+08 34E+09 6.JE+06 6.7E+07 1 . 7,800 180 3,500 750 120 8 500 65 84 1,800 2400 220 1,400 330 230 320 1,800 © 2,000 25,000 520 5,100 ‘ 1970 10/28 26480 7.2 1199 83HE © 5.2 7.8 5.1E+09 1.SE+09 7.4E+08 9.7E+07 8.2E+09 2.3E+09 1.0E+10 1.1IE+10-3.5E+08 1.3E+10 1.2E+10 3.5E+07 6.6E+08 3.2E+08 33E+07 1.5E+07 53E+09 1.1E+10 2.6E+08 4.2E+08 3 52,000 15,000 140,000 18,000 - 93,000 26,000 70,000 160,000 -5,100 90,000 83,000 43000 20,000 9,700 11,000 4,900 39,000 ‘80,000 19,000 31,000 ‘ 1973C 10/29 26,689 7.2 1196 7.8HE 6.0 54E+07 2.6E+07 7.5E+06 1.9E+06 2.0E+06 24E+06 1.5E+07 2,SE+07 L1E+07 20E+09 1.2E+09 3.9E+06 34E+06 9.2E+07 3.5E+07 5.9E+06 2.0E+06 1.0E+09 54E+09 2.5E+07 9.7E+07 1 o 550 270 1,400 350 22 27 100 370 170 13,000 8,000 5500 4,700 2,800 1,000 2,000 670 7,500 40,000 1,800 6,900 \ 1977 10/31 27,051 1.2 1196 SOHE 58 15.0 1.6E+09 2.5E+07 14E+07 7.5E+05 4.2E+06 1.5E+06 7.1E+07 7.9E+07 8.6E+05 19E+09 5.9E+07 1.3E+06 2.0E+08 2.9E+07 1.5E+07 2.7E+06 3.1E+08 1.0E+07 1.8E+07 4.6E+0S 1 16,000 250 2,500 140 46 17 470 1,100 12 13,000 390 1,800 5,900 850 4,900 910 2,200 75 1,200 32 1978 10/31 27,060 72 1187 8OHE 58 15.0 1.8E409 2.2E+07 14E+07 1.6E+06 3.7E+06 5.8E+06 7.6E+07 46E+07 1.2E+06 7.6E+08 7.0E+08 6.7E+05 6.7E+07 2.3E+07 5.2E+06 1.SE+06 2.2E+08 3.3E+08 1.2E+07 9.2E+06 6 18,000 220 2,600 300 40 63 510 660 17 5100 4,700 920 2,000 690 1,800 510 1,600 2,400 840 640 1979 11/2 27411 00 1199 88HE 40 15.0 1.0E+07 5.3E+06 2.8E+06 7.1E+05 54E+05 23E+05 1.1E+06 1.8E4+07 4.2E+06 3.7E+08 14E+08 8.0E+05 1.1E+06 3.9E+07 1.SE+07 2.9E+06 1.2E+06 1.0E+08 4.9E+08 8.2E+06 24E+07 2 ) 97 52 510 130 6 2 7 250 58 2600 1,000 1,100 1,600 1,100 450 . 980 390 800 3,800 560 1,600 209 12/t 28,286 1.2 1199 80HE 58 13.8 1.9E+09 1.1E+07 9.6E+07 1.6E+07 8.8E+06 1.2E+07 4.1E+06 5 | ) 25,000 1,900 1,400 230 370 4,200 420 2012 122 28,483 1.2 1199 8IHE 5.7 13.8 8.9E+08 7.4E+06 1.4E+06 3.7E+07 8.0E+07 1.7E+09 3.4E+07 1.6E+07 $.3E+06 2.1E+08 7.7E406 - 3 . . 11,000 1,300 100 490 1,100 14,000 93,000 3,100 1,900 2,100 780 2027 12/10 29,838 7.2 1201 . 88HE 4.1 13.8 1.0E+09 6.0E+06 3,1E+08 6.0E+07 9.3E+07 2.2E+09 8.4E+06 8.7E+07 4.4E+06 7.8E+09 3.0E+08 3 12,000 1,100 3,800 560 1,200 15,000 15,000 3,000 1,500 60,000 25,000 2032 12/12 30,175 0.0 0 88HE 40 15.0 2.1E+07 1.4E407 5.3E+06 7.0E+06 7.3E+07 8.9E+08 2.7E+05 6.0E+07 3.9E+06 1.3E+08 1.2E+07 2 : 230 2,500 64 62 980 6,200 450 2,000 1,300 1,000 1,000 116 g Salt-Seeking Isotopes Salt Constituent Elements 131 . 957, 1410, 'f‘Ce 1474 . 133y U L | Be fapsule Nozzle - Capsule Nozzle Capsule ~ Nozzle Capsule Nozzle Capsule * Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzle Capsule Nozzie | 3E+06 1.1IE+06 2.2E+06 4.5E+05 1,2E+06 1.0E+06 4.0E+06 9.1E+04 0.001 0.009 1.73 0.012 14,000 25 49 15 42 23 9 290 79 1,300 210,000 100 L4E+06 1.7JE+06 2.6E+07 2.0E+07 9.SE+06 1.1E+07 2.3E+07 4.6E+07 0.003 0.003 0016 0052 0.11 0.13 0.041 16,000 19,000 570 450 330 - 390 520 1,000 - 510 410 - 2,000. 450 950 1,900 610 6E+05 2.8E+06 . 2.1E+06 L6E+07 1.0E+06 1.2E+07 26E+06 2.5E+07. 33E+05 0.001 0.004 0.05 0.006 0.003 11,000 32,000 46 350 36 410 58 550 ‘ 1,100 - 97 670 430 94 37 L2E+06 5.3E+06 1.1E+06 7.6E+07 2.3E+06 5.0E+07 1.0E+07 1.1E+08 1.7E+0S 16E+06 0.001 0.016 0.015 10.22 0024 - 0.12 25,000 60,000 4 1,700 82 - 1,700 230 2,500 560 5,300 110 2400 1,900 1,900 360 1,800 |8E+08 9.9E+07 2.7E+06 9.2E+05 1.1E+06 - ‘4.1E+06° 3.0E+04 3.1E+05 2.3E+06 0.001 0.001 0.006 0.005 0.002 6,000 3,300 ‘52 - 18 24 90 1 18 130 120 82 740 75 30 |0E+08 4.9E+08 4.3E+06 B.0E+05 1.4E+06 3.6E+05 3.9E+06 S.3E+06 © 2.1E+06 0,001 - 0001 ' 0.008 0.001 0.024 - 0.003 13,000 16,000 84 15 - 30 7 -85 120 120 110 87 . 990 120 . 360 45 [2E+08 1.3E+09 4.8E+06 2.7E+07 2.1E+07 24E+07 4.0E+07° 29E+07 3.5E+06 14E+07 0.004 0.003 0.014. . 0.006 0.052 0.017 0.02 5,700 33,000 86 480 360 410 870 620 150 620 620 450 1,700 740 450 250 300 L9E+08 8.2E+08 3.7E+05 1.7E+06 1.7E+05 3.1E+06 8.0E+05 2.7E+06 3.2E+06 0.0 0.0 ‘ 0045 0016 0.002 23,000 19,000 6 28 3 44 17 59 120 36 58 390 240 30 [0E+07 1.2E+09 L.SE+05 2.8E+06 53E+04 4.1E+06 1.5E+06 5.8E+06 2.2E406 0.001 0.001 0004 0.005 0.003 450 27,000 3 7 47 1 58 . 32 120 , 81 84 140 . 35 15 - 45 [SE+09 1.2E+09 6.0E+06 1.8E+07 2.7E+06 1.8E+05 4.6E+06 9.0E+06 1.2E+06 0.001 0002 Co0012- 37,000 29,000 98 300 - 37 3 ~100 190 46 210 - 330 180 SE+08 6.6F+07 8.5E+04 29E+05 2.8E+(4 1.6E+04 -2.0E+05 1.0E+05 ' 0.0 0.0 0.014 0.026 6,800 1,600 ' 1 5 G 0 4 2 34 6 1,700 390 6E+08 4.7E+08 2.2E+06 3.0E+06. 8.4E+05 1.9E+06 1.1E+06 2.8E+06 29E+05 1.1E+06 0.001 0.0 002 0.001 14,000 10,000 35 47 11 25 24 60 10 37 100 49 ‘ . . 300 15 4E+09 1.2E+09 1.2E+07 1.7E+07 1.8E+07 3.SE+07 1.6E+07 1.7E+07 B8.3E+06 1.3E+07 0,002 0.003 0.054 0005 0035 0.034 0012 0.014 22,000 20,000 170 250 200 390 340 360 220 340 330 490 6,600 620 300 290 180 210 6E+07 2.2E+06 1.5E+06 5.8E+05 1.8E+06 1.5E+04 2.0E+06 2.1E+06 9.6E+05 8.9E+04 -0.0 0.0 o 0.007 0.002 210 28 19 7 16 0 41 43 20 2 .10 33 100 30 1E+07 7.3E+07 6.2E+05- 1.9E+06 2.0E+06 3.7E+06 1.SE+06 2.2E+06 1.2E+06 2.3E+06 0.0 0.0 0.033 0.009 1770 930 8 24 18 32 30 44 26 46 30 45 ) 490 130 3E+07 1.9E+08 84E+05 5.2E+05 L.1E+06 2.7E+06 6.9E+05 1.6E+06 74E+05 2.2E+06 0.0 0.0 0.1 0.019 0015 0.001 880 2,300 10 6 9 22 13 31 14 41 54 24 870 160 220 15 1E+06 4.8E+08 1.6E+05 1.6E+06 4.9E+04 2.1E+06 14E+05 1.1E+06 1.5SE+06 0.0 0.0 0.09 0.03 0.016 - 0.002 37 5,800 2 .18 0 16 3 22 - 28 60 21 780 250 240 30 4E+08 1.7E+08. 9.9E+04 1.7E+05 64E+04 14E+)5 9.1E+04 1.1E+05 ' 1,700 2,100 1 2 L 1 3 , 0E+08 29E+08 4.7E+09 1.8E+(9 6.9E+09 2.2E+09 3.7E+09 1.1E+09 1.7E+09 . 004 012 3,500 3400 50,000 19,000 50,000 - 16,000 70,000 21,000 30,000 50,000 15,000 N 6E+08 4.7E+08 7.6E+05 1.BE+06 6.3E+05 2.3E+06 1.0E+06 1.3JE+06 1.8E+06 1.3E+07 0.002 0.001 147 . 0.004 0018 0.005 0.003 0.001 ' 1,900 5,500 8 19 5 16 19 24 33 220 - 250 75 180,000 500 160 , 43 . 45 15 9E+08 3.3E+07 6.8E+05 3.1E+0S 6.2E+05- 3.2E+05 6.3E+05 '2.SE+04 L.SE+04 0.0 0.0 : 0.02 - 2,200 30 . 7 3 4 2 - 1n 0 0 45 30 170 2E+07 4.1E+07 14E+07 3.1E+06 S5.3E+H)6 3.3E+06 4.2E+06 24E+06 2.7E+06 0001 00 0024 0.004 0.00% 710 470 150 kY 37. 23 79 44 48 120 60 210 - 35 15 1E+08 3.3E+08 6.5E+05 2.7E+03 6.7E+0S 2.6E+05 6.1E+05 3.5E+0S 0.0 0.003 0.033 0.018 0.02 2,500 3,900 7 3 5 2. 1 -7 24 380 290 . 160 300 9E+06 1.5E+06 6.5E+05 3.7E+06. - ' 150 19 7 74 8E+07 4.2E+06 1.2E+06 4.7E+06 0.001 0.013 870 54 12 _ 93 75 1,600 . 4E+08 6.6E+06 3.5E+06 3.9E+06 0.001 0.01 0.011 5,200 ' 75 30 ' 74 100 87 160 SE+08 4_.7E+06 - 4.9E+06 3.7E+06 0.001 046 0.065 3,600 53 41 70 100, 57,000 560 -— . e . - Table 11.9. Gas-Borne Percentage of MSRE Production Rate Double-wall capsules, runs 19 and 20 (sampled during power operation) b . s s Isoto Gross? Net Stripping pe " ‘ ; Number Range Median Mean Number Range Median ~ Mean (calcd) Isotopes with Gaseous Precursors %8s 13 0.3-17 5.2 651 1 0.06-15 3 57%1.2 14 137¢s 11 . 6-98 22 336 9 -16-91. 23 2516 18 oy 13 0.005-3 0.08 0.36%0.17 11 -0.11-0,08 0003 0.006+0.010 0.07 1408a 13 0.005-0.4 0.08 0.10£0.02 11 -0.004-0.18 ~ 0027 0.056+0.013 0.16 _ ) . Satt-Seeking Isotofies ' . %zr . 13 0.002-0.3 0.04 0057+0014 11 . -0.007-005 0.006 0.012 *0.004 141a 13 0.002-0.2 0.009 0.025£0.011 10 —0,03-0.009 -0.0003 —0.003 +0.003 144¢cs 13 0.01-1.7 022 0.32+0.09 11 ~0.12-0.42 0.007 0.05+0.03 147Ng 9 0.0001-0.1 0,012 0.021 +0,007 9 ~0.01-0.01 —0.001 0.002 £0.002 - “Noble” Metal Isotopes Nb 13 0.07-7 0.7 1.90.5 11 ~0.2-36 04 0.9 +0.2 Mo 13 0.16-16 1.0 2709 11 ~0,6-7.3 0.3 1.5 £0.5 111xg 13 0.2-20 1.5 40+%1.1 11 -0.9-4.1 0.3 0.7+0.3 193u 13 0.31-20 1.8 43%1.2 11 0.05-10 1.1 23107 1%%Ru 13 3.8-67 13 - 22%4 11 -2-36 6 13 ‘ Tellurium-lodine Isotopes 1291¢ 13 0.3-27 1.8 5115 11 —0.11—4 01 . -1%08 13274 13 0.03-23 1.0 3.5%1.2 11 ~12-2 ~0.4 ~1+0.8 131 13 0.04—6 0.8 1.6 0.4 11 —0.1-2 0.2 0.5+0.1 9Gross includes capsule plus nozzle isotopes, bNet includes capsule isotopes only, less proportional quantity of material of nozzle composition, for the given sample, A net gas-borne value for the capsule was then calculated by subtracting the capsule mist contribution. from the observed value and, as usual, dividing by capsule volume. This net value could then be ratioed to the isotope production rate per unit purge flow rate as before. The range and average of such values are given in Table 11.9. It is evident from this table that #?Sr and 13"Cs - which would presumably exist as 3°Kr and '37Xe at the time of sampling, are present in net amounts of the proper magnitude if valid gas samples are assumed. Since salt mist of nozzle composition has been sub- tracted, it appears legitimate to accept the samples in general as being acceptable samples of the gas regxon within the pump bowl spray baffle. . Insofar as the material in the shielded region repre- sents the pump bowl gas in general and the gas entering the off-gas line in particular, the data of Table 11,9 also indicate the following: The gross amounts of “noble metal” isotopes carried by the off-gas are in general a low few percent of production (though ' °®Ru appears higher, in part because the high yield from 23° Pu in the fuel was not taken into account). Much but not all of this appears to be associated with salt mist, but some carries through into the -net value, indicating additional noble metal transport independent of salt mist. Further, the tellurium isotopes also are found only to the extent of a low few percent; however, essentially all . ‘of these appear to follow the salt mist, so that net quantities tend to approach zero. It then appears likely that noble metal isotopes and tellurium isotopes could be separately transported to some extent. These data are ratioed to production rate and imply steady-state conditions and short holdup periods in any region; the simple behavior model thus assumed appears stralghtforwatd reasonable, and adequately conserva- tive. However, our recent studies of activity ratios of isotope pairs, such as '®3Ru/'°®Ru and ! *?Te/! 2°Te, indicate that the material found in gas capsules may not be from recent production but from a mix of it with a presently indeterminate but nevertheless substantial portion of the inventory from a considerable span of the previous power history. This is consistent with the rapid equilibration of recent production and inventory -indicated by the salt sample noble metal analyses. To apply such a model will require further definition of 118 For a. period of six weeks, experiments were con- ducted with the same salt charge, and 27 salt sam- ples were removed from the melt. The first three ~ contained anomalously high concentrations of nio- partition of losses between off-gas and other ‘possible “sinks,” a matter under current consideration. Present indications are that the fraction of noble metals lost to off-gas will not be greater than shown in Table 11.9. Correlation of the gas sample data with possibly relevant operating conditions is not apparent (other than reactor power, gas flow rate, pump bowl pressure, and temperature, which are incorporated into the value of production rate per unit purge flow rate). In these runs, samples 19-1 through 19-29 and 20-1 showed an appreciable fraction of the **Nb inventory in salt samples; consequently, the fuel salt was less reducing during this interval. However, 5Nb quantities in gas samples gave no evident response to this condition of the salt. ' 11.3 INVESTIGATIONS OF THE BEHAVIOR OF FISSION PRODUCT NIOBIUM H. W. Kohn Laboratory investigations of the chemical behavior of the fission products ?>Nb and ?5Zr were con- tinued. The principal emphasis of the current efforts was directed toward preliminary studies of the rela- tionship of the redox potential of fluoride melts to the distribution of these isotopes between the melts and metal containers. All experiments were con- ducted at 650°C. The following experimental procedures were employed: 2.15 kg of simulated fuel salt (Li,BeF, containing 5 mole % ZrF, and bium, indicating that the initial zirconium-niobium separation was not complete. All subsequent samples were deficient in niobium, indicating behavior unlike that observed in the MSRE. In the laboratory exper- iments, niobium was not quickly dissolved and trans- ferred to the melt by introduction of HF as an ‘oxidant, nor removed from the melt by metallic total " inventory lithium as a reductant. Instead, it behaved erratically under varying conditions of redox potential, as is indicated by the data in Table 11.10. Niobium on the gas probes was low, usually less than 1% of that formed during a sampling period and consequently a much smaller fraction of the in ‘the container vessel. The ®5Nb/?5Zr ratio in the gas probes was always ob- served to be greater than the ratio in the salt, usually a factor of 10 or so, and was higher for oxidized melts than for reduced melts. Samples from reduced melts exhibited increases in the ?5Nb/?5Zr ratio with increasing gas flow rates. The amount of salt in the probes, as judged from the zirconium - activity, was too great to be accounted for by vaporization. We believe it to be carried as fine droplets. The electron microscope slides suspended above the melt were found to be encrusted with salt, The size of the salt particles from the melt ~ when it was in an oxidized condition was about 0.1 0.3 mole % UF,) was tagged with 6.5 mc of °°Zr which had been freshly separated from its ?3Nb daughter. The salt was contained in a nickel vessel, and the apparatus was arranged so that gas samples could be collected either by chemical probes, de- scribed previously, or on electron microscope slides which were positioned just above a gas bubbler tube. The amounts of niobium and zirconium contained in the gas sampling tubes and also on immersion speci- mens were determined by radiochemical analysis. A nickel—nickel oxide electrode was also used to moni- ‘tor the redox potential of the melt. Gas flow rate, redox potential, niobium content, bubble movement, and character of the immersion probes were the principal variables studied. ‘ p. They appeared to have either rhombohedral out- lines or, in addition, adhering fragments which seem- ingly lacked morphological symmetry (Fig. 11.7a). The salt from reduced melts, however, exhibited dendritic or needle growth patterns and larger par- ticle sizes (Fig. 11.7b). It should be noted that the two photographs do not represent identical magnifi- cations. Electron diffraction analyses of the com- posite particles - indicated that in addition to the fluoride salt, nickel metal and carbon particles were present. : S Loglog plots of zirconium and of niobium deposi- tion in and on the probes vs flow rate were made in order to estimate the order and rate of deposition. Zirconium deposition, not surprisingly, was - always found to be about first order (0.7 to 0.9). Niobium deposition rates, though highly erratic, appear to have a higher order rate as high as 1.4 to 4, most probably 1.5. The actual numbers are not firm enough to postulate mechanisms because of the inac- curacy of the data and the small number of experi- 119 Table 11.10. Distribution of Niobium in Salt Samples ' . At Sampling Time , uHu* ' Date — State of Melt Treatment - Remarks Nb/Zr Observed Nb/Zr Calculated Observed/Calculated - if Any - 8/7 0.110 - 0.020 5.50 Oxidized HF+H,; 1 hr 8/ 0.110 - ‘0,020 - 5.50 Oxidized - HF+H, 4 hr 8/7 - 0.094 0.020 . 470 Oxidized HF + H; 4 hr ‘ 8/11. 0.076 0.097 0.784 Oxidized 5%) made the melt wetting toward these metals. These results are consistent with previous observations.? The behavior described above seems to have quali- tative relevance to the migration of niobium in the MSRE fuel, where at slightly above and below approximately 0.5% [U*]/[ZU] major differences in ?*Nb disposition within the fuel circuit were 'ob- served. (Sect. 102, this report). The results of the current study tend to mdlcate that wetting behavior may play a significant role. in controlling the distribution of woids and fission. products in molten-salt reactor fuel systems, but do not, at the present stage of development, provnde' posmve mdlcatlons of the mechamsms 1p, 3. Kreyger, S. S. Kirslis, and F., F, Blankenship, Reactor Chem, Div. Ann. Progr. Rept. Jan, 31, 1964, ORNL-3591, pp. 38—42; 1. P, Young, K. A. Romberger, and J, Braunstein, MSR Program Semiann. Progr. Rept. Feb, 28, 1969, ORNL-4396, p. 205. _ Table 11.12, Comparison of Observed Amount with Calculated Inventory After Correction of Observed ~ Count to the Time of Isolation Ratio of Amount Found - to Calculated Inventory per Gram of Salt " Nickel Cage . NbFoil - Foil Leachings 89gy 0.759 0.00191 0.0883 95Zs . 0.0155 - 0.000066 - 0.104 1408, " 0.0274 0.000466 0.128 147Nd ©0.0236 0.000234 0.156 ?5Nb 395 . 10,0246 0.064 Mo - 49.8 . 0.683 ‘ 915 1115 20.8 ~0.621 4,08 1327¢ 220 0.273 0.514 131 2.32 0.00711 0.296 114 NOBLE METAL FISSION PRODUCT CHEMISTRY 11.4.1 Introduction | C F. Weaver The unpredlctable behavior of the noble metal fis- sion products (Nb, Mo, Tc, Ru, and Te) in the MSRE -caused us to initiate studies of their high- temperature chemistry some time ago. Emphasis was given to molybdenum fluoride chemistry because the combination of fission yield and cross section for the molybdenum isotopes poses a potential problem in the breeding efficiency of an MSBR. To a lesser extent the behavior of niobium and ruthenium fluo- rides was investigated. The fluorides of technetium and tellurium were given attention only in our litera- ture surveys. Previously reported studies of molybde- num fluoride chemistry, along with the reasons for initiating this work, have been reviewed.'”? Recent evidence that niobium metal may be oxidized and .enter fuel salt when U**/U* ratios are quite low suggested that this fission product might - provide a means of monitoring the redox potential of such a system. Consequently, increased attention e, F. Weaver et al., Reactor Chem. Div. Ann. Progr Rept. Dec, 31, 1968, ORNL-4400, pp. 33-41, 2C. F. Weaver et al., MSR Program Semiann. 'Progr Rept. Feb. 28, 1969, ORNL-4396, pp. 157—-62. : 3C. F. Weaver et al., MSR Program Semiann, Progr Rept Aug. 31, 1969, ORNL-4449, pp. 113 21 - 124 has recently been given to niobium fluoride chem- istry. The following subjects were discussed in earlier reports: _potentiomet'ric measurements,* niobium pentafluoride volatility,’” synthesis of NbFs,® thesis of NbF,,’ -spectroscopxc investigations,®* mass spectrometric studies,” -and ?°*Nb tracer experiments.® . - ‘ 1142 Synthesis C. F Weaver H. A. Fnedman J. S. Gill Several gram batches of MoFs, MoF,, and MoF; were synthesized according to the procedure reported in ref. 3. The materials are used in the kinetic and mass spectrometric studies described below as well as in Raman®:'° and absorption!!-12 “spectroscopy studies. A 385-g batch of NiF; was synthesized by hydrofluorinating NiCl,-4H,0, using a procedure developed by B. J. Sturm.!3® Nickel fluoride is 2 mild fluorinating agent commonly used by several groups in the Reactor Chemistry Division to oxidize Nb metal. OQur current efforts are directed toward the synthesis of the lower fluorides of nio- bium by reduction of NbFs!* and from dispropor- tionation reactions of the lower fluorides. 11.4.3 Niobium and Molybdenum Fluoride ‘Solutions in Molten Li; BeF, C. F. Weavér H. A. Frieflman A series of experiments was initiated to investigate the reaction Nb+xHF#Nbe+32-c-H2 in molten Li, BeF, with respect to the valence of niobium in equilibrium with niobium metal, the partial pressures of H, and HF as a function of Nb** .concentration, and the stability of the NbF, solutions in a neutral (He) atmosphere. The niobium ‘was rapidly fluorinated by the direct addition of HF. After the Nb concentration reached 300 ppm the system was held under flowing He (3.3 liters/hr) for six days. No loss of niobium occurred. Hydro- fluorination was resumed and continued until the concentration of niobium rose to 1500 ppm. This solution was held for 32 days under He without loss of niobium. Sweeping the system with 1 atm of H, (5.5 liters/hr) produced an initial burst of HF. This syn- , was caused by thle' reduction of the fluorides of impurities more noble than niobium and was fol- lowed by a very slow evolution of HF. These results - indicate that approximately 1.5 X 10™ atm of HF exists at equilibrium with 1 atm of H, and 1500 ppm of Nb** in its lowest valence state in molten Li,BeF, at 500°C, and imply as well that under these conditions niobium is only slightly more noble than Cr. This behavior of Nb contrasts sharply with that of Mo under similar conditions. In a period of 32 days under He, 1500 ppm of Mo** would have been reduced by disproportionation to 1000 ppm. With 1 atm of H,, Mo® would be reduced as fast as the H, entered the system. Based entirely on - experimental information, the order of increasing ‘nobility with respect to fluorination of metals in molten Li,BeF, is Cr, Nb, Fe; Ni, Mo, Cu, and Ru. Information in the literature suggests that Tc is approximately equivalent to Cu in this respect. ' ‘A stronger reductant, Li metal, was used to reduce the dissolved niobium and, hence, determine its valence. The stoichiometry of the reaction indicated that niobium was present as Nb 3.6 + 0.1. The absence of concentration dependence suggests a cluster compound, Nb;F,;. We are now repeating this experiment at a higher initial concentration of Nb3-¢* to improve the accuracy of the valence deter- mination, to provide a more stringent test of con- centration dependence, and to determine the solubil- ity limit. At present the concentration is 2300 ppm and has shown no variation over a period of 15 44, R. Nichols, Jr., K. A. Romberger, and C. F. Baes, Reactor Chem. Div. Ann. Progr Rept. Dec. 31, 1966, ORNL-4076 p. 26. SC. F. Baes, Ir., Reactor Chem. Div, Ann. Progr. Rept Dec 31, 1966, ORNL-4076, p. 50. L. M. Toth, H. A. Friedman, and C. F. Wmver, MSR Program Semzann. Progr Rept. Feb. 29, 1968, ORNL-4254, p. 1317, : L. M. Toth and G. P, Smith, Rezctor Chem, Div. Ann. Progr. Rept. Dec 31, 1967, ORNL-4229, p. 64, , 8H. W, Kohn, MSR Program Semiann, Progr. Rept. Aug. 31, - 1969, ORNL-4449, p. 112, 9J B. Bates, Chemistry Division. 104, 5. Quist and L. M. Toth, Reactor Chemistry Division. llL M. Toth, G. D. Brunton, andG P, Smith, Inorg. Chem, 8, 269497 (1969). 12J P. Young, Analytical Chemistry Division. 13g . Sturm, ANP Quart. Progr. Rept. June 30, 1957, ORNL-2340, p. 164. 14F. P. Gortsema and R, Didchenko, Inorg. Chem, 4, 18286 (1965). -days with a helium flow rate of 3 liters/hr. These - observations should not be extrapolated to other - temperatures since there is evidence in the litera- 125 ture!® that both the wvalence and stability of the niobium fluorides are strongly tempet?qture‘ depend- ent. ' - The relative stability of this niobium fluoride and the carbides, Nb,C and NbC, which can occur in MSR operations may now be discussed. The standard ~ free energy changes (or equilibrium constants) for the following reactions are not available: Nb(s) + XUF, = NbF,(d) + xUFs(d) AFy, K; (1) NbC(s) + xUF.(d) = NbF,(d) | | + xUFs(d) + C(s) AF;, Kz () However, subtraction yields | Nb + C = NbC AF; (3) so that 'AFl - AFz = -—RT]H KI/Kg = AF3 ’ . which has been reported!®+!7 to be —32 kcal/mole at 773°K. Further, experiments in progress in which excess Nb metal was hydrofluorinated into molten - Li; BeF, at 500°C (773°K) have given by Li titration a tentative value of x = 3.6 + 0.1 which has shown no apparent concentration dependence and suggests a cluster compound, Nb3 F, ;. Hence, if one assumes that the activity coefficients for uranium cancel: o (f_)/zi) ~1 - 7_“ 1 7“ 2 - ‘and defines UF3/UF4 T, then o & =fl * K, (’2) 15g, Fairbrother, The Chemistry of Niobium and Tantalum, pp. 121, 142, Elsevier Publishing Co., New York, 1967. under conditions such that both reactions produce the same activity of NbF; . ) "1 ’3 .6 AF; = ~RTIn (——-) , r 3.6 _32=—R(173)In (fi) , _ ” Thus if 7, of 0.5 X 1072 will oxidize Nb metal,’® then r, of 1.6 X 1075 will oxidize NbC as well as the less stable Nb,C.'6-!7 This value of r is insufficient to oxidize Mo metal® but may oxidize Fe, which is more noble than Nb but less noble than Mo. The stability studies of Mo®*" in molten Li,BeF, at 500°C were continued. It was previously reported that a tenfold increase in the surface area of the copper container did not affect the rate or the apparent order of removal of Mo* from solution over a 1400-hr period, This experiment was continued to 4500 hr, giving the results shown in Fig. 11.8 and yielding a half-order rate constant, defined as K = ¢t 1(C}/? — C1/2), of 6.42 X 1072 ppm™1/2 hr™? by least-squares analysis. It is now planned to study the temperature dependence of the disproportionation kinetics and the valence of molybdenum in this solvent. , ~ An experiment was initiated to determine the effect of graphite on the removal of molybdenum from solution at 500°C. The preliminary results suggest that graphite has no significant effect on the kinetics. An atteihpt to introduce Mo** into molten Li, BeF, - at 500°C by direct hydrofiuorination of the metal was’ : barely successful. After 18 hr of hydrofluorination, using undiluted HF at 1 atm for the last 8 hr, only 92 ppm of molybdenum 1 was found in the filtered samples. Again the behavior of molybdenum and njobium may be contrasted. Niobium hydrofluorinates into molten - Li, BeF, with ease, molybdenum with difficulty. On the other hand, the reduction of the molybdenum ~ solutions with hydrogen occurs very rapldly, while 16p, R, Stull and G. C. Sinke, Thermodynamic Properties of the Elements, p. 67, American Chemical Society, 1956, ‘ 17Fdmund K. Storms, The Refractory Carbides, p. 72, _ Academic Press, New York, 1967; J, F, Elliot and M. Gleiser, Thermochemistry of Steelmaking, vol. I, p. 142, Addison- - Wesley Publishing Co., Inc., Reading, Mass., 1960. reductlon of niobium solutions is extremely low., 18R E. Thoma, personal communication, Dec. 23, 1969. 126 ORNL-OWG 70-6759 30 N o A Pe 3 / 500°C, 12 liters/hr He FLOW, 1800 cm? OF Cu SURFACE o ] /- CONCENTRATION (ppm) s N 0 1000 2000 3000 TIME (hr) Fig. 11.8. Removal of Mo®" from Molten Li, BeF,. : 11.4.4 Mass Spectroscopy of Molybdenum ‘and Ruthenium Fluorides - C.F.Weaver J.D.Redman Studies of the vapors associated with the evaporation, decomposition, and reactions of the fluorides and oxyfluorides of niobium, molybdenum, and ruthenium ‘have continued. In addition, further efforts to refine the pressure measurements were made. Recently we reported® that single-crystal 7LiF and Ag were used as pressure standards to check the sensitivity of the time-of-flight mass spectrometer before and after an experiment. These results improved the reliability of our. pressure estimates from an order of magnitude to within a factor of 2. The most severe problem with respect to further improvement in ac- curacy is that the corrosive noble metal fluorides cause a deterioration in machine sensitivity during an experi- | ment and that this decline is unlikely to be linear in time. The feasibility of using an internal standard to continuously monitor the machine sensitivity during an experiment was investigated. Krypton was selected because of its - availability, absence as a2 common impurity, nearness in mass to materials currently under investigation, and chemical inertness. The results were highly encouraging, and the development of this tech- nique will be pursued. A surprising observation with respect to machine sensitivity was made during direct fluorination experi- ments. It was observed that the addition of fluorine gas to the instrument followed by pressurizing to 1 atm with dry nitrogen returned the machine sensitivity to its initial value or higher. This technique will be useful in avoiding the time-consuming machine cleaning (about three days) after each run. It further emphasizes the need for an internal standard, since the sensitivity may increase as well as decrease during different steps of a given experiment. It was previously reported® that work is proceedmg' toward refining the infrared adsorption spectra of molybdenum oxyfluorides, since they exist as im- purities in MoFs samples and, hence, complicate the derivation of its spectrum. Efforts to synthesize the oxyfluorides of molybdenum for this purpose led to a study of the reaction of MoQ;(s) with MoF4(g) in a Knudsen reaction cell. The cracking patterns for MoF,, MoOF,, and MoO, F, have been reported earlier.!®-21’ More detailed fragmentation patterns, including doubly _ ionized species, may be seen in Figs. 11.9-~11.11, and ionization efficiency curves in Fig. 11.12. The reaction 19R. A. Strehlow and J. D. Redman, MSR Program Semiann, Progr. Rept. Aug. 31, 1967, ORNL-4191, pp. 144—47. 20p. A. Strehlow and J. D. Redman, MSR Program Semiann. Progr. Rept. Feb, 29, 1968, ORNL-4254, pp. 134—43. ‘2R, A, Strehlow and J. D, Redman, Regctor Chem. Div. Ann. Progr. Rept. Dec. 31, 1967, ORNL-4229, pp. 37-39. 10 3 o 45 50 55 60 - 65 107> MOF2+ CELL PRESSURE (torr) ~4 10 ; - 90 100 Ho 120 130. 127 70 140 atomic mass units/charge ORNL-DWG 70-6760 s 80 85 90 95 100 atomic mass units/charge MoFz* 50 160 170 180 190 200 Fig. 11.9. Fragmentation Pattern of MoFg at 75°C. was first carried out in a copper cell, which was unsatisfactory. Although very little copper-containing material was observed in the effusing gas, cell reaction was indicated by mass transfer of copper, the reduction of MoO; to MoO,, and the reduction of MoF, to MoFs and MoF,. The reaction of the MoO; residue with M0F6 at 700 C produced Species Vapor Cdmpositidn (%) MOF6 | ’ 1.0 MoFs . 180 MoFs 230 MoOF 4 - 570 MOOze, o . 1.0 The reaction of MoO;(s) with MoF4(g) was again studied using a nickel cell. In this case no nickel- containing molecules were observed in the effusing gas, no mass transfer of nickel was detected, the residue was pure MoO3, and no lower-valence fluorides of molybde- num were seen. Since there was no evidence of reduction of the molybdenum compounds and since .oxidation of hexavalent molybdenum cannot occur, it - uncertainty is concluded that all of the oxyfluorides observed were of hexavalent molybdenum, This resolves the earlier 1921 in assigning the MoOF, cracking pattern and allows analysis of the products of both of the above experiments. The composition of the effusing gases may be seen in Fig. 11.13. Note that conditions exist in which either pure MoO, F, or M00F4 may be synthesized. In earlier reports®? we have described the d1spropor- tionation of pure RuF; in the temperature range 600 to 700°C as: 4RuF;(s) > Ru(s) + 3RuF4(g) . Tonization efficiency curves and appearance potentials for Ru®, RuF,*, RuF;*, and RuF," are given in the above references as well as the cracking pattern for RUF4 . V 10 107> Cm 45 50 55 €0 65 70 5 | w € MoOF + A AND & MoF,* a. - T —d w © 20 100 110 120 130 140 atomic mass units/charge ORNL-DWG 70-6761 75 80 85 9 95 100 atomic mass units/charge MoOF2+ AND MoFy* 150 160 170 180 190 200 - Fig. 1110, Fragmentation Pattern of MoOF,4 at 125°C. Recently the reaction of fluorine with ruthenium metal contained in a nickel Knudsen effusion cell was investigated. Fluorine gas was passed through the cell over the temperature range 25 to 800°C. Below 450°C no reaction was observed. Throughout the range 450 to 629°C the effusing vapor was a mixture of F, and RuF,. Hence the reaction was Ru@) + 2Fa(6) > RuFa(e) . The simplicity of the vapor in the case of ruthenium fluorides contrasts markedly with the analogous “molybdenum fluorides, where MoF,, MoFs, MoFq, Mo, Fyo, and Mo;F,s are observed simultaneously in the vapor phase. Above 629°C, NiF, appeared in the spectrum, while-the RuF, vapor_ pressure was observed to be at a maximum near 675 to 700°C, as shown in * Fig. 11.14, This behavior of nickel in the presence of fluorine has been observed by others.2?:23 At lower temperatures a layer of NiF, forms and protects the . nickel substrate from further attack. In the range 620 to 700°C the NiF, desorbs, and cell corrosion occurs. In our case the nickel appeared to react with both the F, and RuF,, reducing the effusion of both species. No polymers, oxyfluorides, or ruthenium fluoride frag- ments with mass greater than RuF," were seen. 22§ p_McKinley, J. Chem, Phys. 45, 1690 (1966). 23R, L. Jarry et al., J. Electrochem, Soc. 110, 346 (1963). j i : ; CELL PRESSURE (torr) 129 - ORNL-DWG 70-6762 1073 -4 107 _ 45 50 55 60 65 70 75 80 85 90 95 100 atomic mass units /charge ‘MoO,F * AND _MOO F2+ 90 100 110 120 - 430 140 150 160 170 180 190 200 otomic mass units/charge Fig. 11.11, Fragmentation Pattern of MoO;F, at 300°C. 130 . ORNL-DWG 70-6763 . 11 10 + . MoOF3 (x57) MoDF g \..- o —t & -—n—_" — ION INTENSITY {orbitrory units) o o ‘——H [ ——t ] / ] | v 1 / :/ | * - r'. R~ A WA o '." =,. . : . -," =’ . . 14 #6 8 20 22 24 26 28 14 t6 . 18 20 22 24 : ELECTRON VOLTS (ev, CORRECTED) T 2 '!l / Fig. 11,12, Ionization Efficiency Curves for the Principle Fragments of Molybdenum Oficyfluoriéles. 131 ORNL-DWG 70-6764 o P i e el 2y, < g m X o o o — & — 8 = O o ~ — Y P e el e e, m & = e + _ = _— o’ e e 7 e i % L w w = , 5 8 VSIS LTI LI LS LS F IS LS 7HT I - .m . d o g oW s — — o & L - . . 8] (P - & ez ‘ o B Q L o © 5 O = 8 . o > g o . - " 2 = v 2 =] = - O 0 ) ye o = ) o O o O Q o o Q9 o Q o o @ o < N Q o Mw < N © (%) NOILISOdWOD HOdVA I*xr 132 _ ‘ o : ORNL-DWG 70-6765 10° ' 10° . 2 10° 5 (N » 7 . ~ 2 x 52 kcat/mole _ : ‘ . — 1% . 5 63 kcal/mole ' 2 , : 10! 0.932 1.028 1424 1.220 1.316 1000/r ok Fig. 11.14. Reaction Products from the Reaction Ru(s) + F2(¢) in a Nickel Cell, . 12 Properties of the Alkali Fluoroboratcs 12.1 PHASE EQUILIBRIA IN THE ALKALI METAL FLUORIDE-METAL FLUOROBORATE BINARY SYSTEMS: THE SYSTEM RbF-RbBF, L. O. Gilpatrick C. J. Barton Investigation of the system RbF-RbBF,, which is very nearly complete, represents an extension of studies we have made of systems containing NaBF, and KBF,. Differential thermal analysis was the chief technique used in this investigation. The RbBF, used was pre- . pared by S. Cantor using an aqueous technique, mixing RbCl and NaBF, to precipitate the slightly soluble RbBF,. This compound melted quite sharply at 582°C and showed a polymorphic transition at 247°C. Rubid- ium fluoride was used as received from a commercial supplier except for dehydration. This material melted at 783°C, 12° below the best literature value for this compound (795°C). . | Equilibrium phase transition data were obtained for ten compositions of RbF-RbBF; mixtures, and an equilibrium phase diagram of the system (Fig. 12.1) was constructed from the results. This system was found to exhibit a single eutectic; it is found at 31.5 mole % RbF and melts at 442 + 2°C, It is quite similar in this respect to the NaF-NaBF, and KF-KBF, systems. The poly- morphic transition of RbBF, was detectable in all mixtures studied, and the temperature at which it . occurred was independent of composition, showing a ‘probable absence of solid solutions. There was no evidence of compound formation. A lowering of the eutectic temperature in RbF-rich compositions may be due to impurities such as hydrolysis products in the impure RbF. This possibility is currently being investi- gated by treating the commercial RbF with ammonium bifluoride. ' It is apparent from Fig. 12.1 that this system does not have a low enough melting point to be attractive for use as a coolant salt in molten-salt reactors that are - designed for maximum economy in conjunction with supercritical steam generator plants. ORNL~-DOWG 70-3032 800 ' ) n..__‘\ -] I I l "\ ' LIQUID A LIQUIDUS 700 | . o SOLIDUS <[ * POLYMORPHIC TRANSITION — | \ IN RbBF, 600 LIQUID + RbF N | = \“ : 7 —t :,3 ' \ _ /‘n/_ W 500 2 \“AIQUIDM:RbBE, g ol T § 400 o a RbBF, + RbF 300 . * . 200 | | B ROBF, +RbF 100 L RbF {0 20 30 40 50 60 70 80 90 RbBF, " RbBF, (mole%) Fig. 12,1, The System RbBF4-RbF, 133 122 SOLUBILITY OF Na,CrF, IN SODIUM TETRAFLUOROBORATE MELTS C. J. Barton Green crystals of Na;CrF¢ have been observed in corrosion test loops in which NaBF,-NaF (92-8 mole %) salt mixtures were circulated. No data on the solubility of this species in fluoroborate solvents exist, but extensive studies of the solubility and stability of chromous and chromic fluorides (Cr** and Cr*) in several other fluoride mixtures were made in connec- tion with the ANP program some time ago. The compound NayCrFg was mixed with sufficient NaF and NaBF, to give a 10 mole % mixture and a solvent composition of 92 mole % NaBF,. The Na;CrFs (prepared earlier by B. J. Sturm) was added in _ ORNL-DWG 69-13697 TEMPERATURE ‘("C) 700 600 500 400 1.0 08 . 086 04 0.2 01 0.08 NasCrFg (mole %) ‘006 004 ¢ FIRST SET OF SAMPLES ® SECOND SET OF SAMPLES 002 - 001 09 10 U 1.2 13 14 15 100077 (ek) Fig. 12.2, Na3CiF¢ (mole %). 134 preference to C1F; because the combination of CrF, with solvent NaF would change the solvent composition - significantly if CrF; were very soluble, The procedure followed in CeF3 solubility studies was employed in the present investigation except that no tracer was used, and sole reliance for solubility - measurements was placed on wet chemical analysis of filtered samples removed with copper filter sticks. Two sets of samples were taken because of scatter of data in the low concentration range in the first set of samples {open circles, Fig. 12.2). Sodium, boron, and nickel were also determined in most of the samples to “establish the solvent composition and possible corrosion of the nickel container. The average of six Na and B analyses, corrected for the Na in Na;CrF,, indicated a solvent composition of 95 mole % NaBF, instead of the intended 92 mole % eutectic mixture. Additional studies will be required to establish the effect of solvent composition on chromium solubility, but the writer does not believe that it will be markedly different in a 92 mole % NaBF, solvent from the observed values shown. in Fig. 12.2. Nickel values were low (20 to 30 ppm) in most samples, but two samples taken at temperatures close to 700°C gave values of 300 and 390 ppm. The solubility varied from 0.46 mole % Na;CrF at 700°C to 0.046 at 500°C, and extrapolation to the eutectic melting point (384°C) shows that the solubility would be less than 0.01 mole % at the freezing point of the mixture., The calculated apparent heat of solution of Na; CrF in this solvent was 17.0 kcal/mole. 12.3 PREPARATION OF PURE SODIUM TETRAFLUOROBORATE L. O. Gilpatrick Ralph Apple A 1.5kg charge of NaBF, was prepared by a dry technique to give a product of greater purity than any previously available. In particular, special efforts were made to reduce the oxygen content to very low levels. - The best material currently available shows an oxygen content of 200 ppm or greater as measured by the KBrF, fluorine displacement method. A reaction vessel was constructed of welded nickel and lined with graphite components in such a way that - the molten salt would come in contact solely with graphite during processing. Into this unit was placed a charge of NaBF, of the best available purity (recrystal- lized from 1.4 M HF solution and then vacuum dried). The vessel was connected to a mutual supply of pure He, BF3;, and HF gases. Precautions were taken to remove all traces of water from the HF by bubbling fluorine through several liters of distilled and condensed reagent-grade HF. The saturated HF liquid was evapo- rated to form the HF reagent gas supply used in the preparation.. A stream of purge gas was passed through the salt. Flow rates were adjusted at 1 atm to attain a vapor-phase composition of 25% HF, 25% BF;, and 50% He. The unit was heated to a temperature of 425°C, approximately 20°C above the melting point of NaBF, . Purification was continued for 30 hr, and then 135 within a silica vessel whose vapor space can be sampled (see Fig. 12.3). An initial experiment has been con- ducted in which coupons of chromium metal were in contact with coolant salt with reported analyses of 120 + 60 ppm H,0 and 310 + 50 ppm oxygen; the latter analysis includes the oxygen in H, O. The closed capsule and the silica vessel were outgassed at 200°C for several days. The assembly was then heated fairly rapidly (about 20°C/min). When the temperature reached . 370°C, we noted a steady rise in -pressure in the the system was allowed to cool under the static atmosphere of gas. The unit was evacuated cold, remelted with the system closed, and reevacuated cold to degas the product. Recovered material was packaged under He in a dry box before examination. Analysis indicates that the oxygen content of the NaBF, was reduced to a value near 20 ppm, a tenfold increase in purity over the starting material. There is evidence from differential thermal analysis and chem- ical analysis that some loss of BF3 occurred, so that the product is not a stoichiometric mixture of NaF and - BF3 as we had hoped. It should be possible, however, to reduce this BF; loss to any desired extent by adjusting experimental conditions as employed in a study of NaBF, dissociation." The results obtained in the expenments described - above demonstrate the objective of this procedure, to show the feasibility of a process which has potential for application to large-scale systems. Such procedure could be employed continuously, if necessary, to maintain the concentration of oxygen-containing species in fluoro- - borate coolant salts at a very low level for the purpose of controlling corrosion or, alternatively, to scavenge hydrogenous species from the salt. 12.4 INVESTIGATION OF OXIDATION OF METALS BY FLUOROBORATE COOLANT 'S. Cantor A series of experiments is under way to determine: (1) the oxidizing properties of the fluoroborate coolant salt (92-8 mole % NaBF,-NaF), (2) the retention of moisture or other hydrogen-containing species by the coolant. If the coolant retains some moisture, then tritium passing into the coolant from the reactor core could be isotopically exchanged and retained in the coolant circuit. In an experimental run, evacuated mckel capsules —contammg coolant salt and metal coupons are heated 13, Cantor, Reactor Chem. Div. Ann. Progr. Rept. Dec. 3I ' 1967, ORNL-4229, p. 56. assembly. The pressure eventually reached about 8 mm after the capsule had been heated to and maintained at 625°C for 2.5 hr, The furnace was then shut down and the assembly slowly cooled to room temperature. It was unfortunately not possible to analyze the contents of the gas sampler until five days later. Gas ORNL-DWG 70-6766 PRESSURE RECORDER GAS SAMPLER oo ——{<}— VACUUM - SEALED 1 THERMOCOUPLE~—__| SILICA NICKEL VESSEL VESSEL— | 3in. HIGHx 14 in. DIAM A, ’////I/////‘/ ool l"////////////// 7 CHROMIUM COUPONS ook ol ke L L Ll L L /MARARAARARAARADRA sdddHUbdEdEEY S Fig. 12.3. Apparatus for Determining Corrosion of Chromium in Fluoroborate Coolant. chromatographic analysis indicated the composition in the sampler to be 20% by volume hydrogen, the remaining 80% being air. The capsule was removed and cut open. The walls of the nickel container had not been attacked by the salt or its vapor. The chromium coupons were mostly covered with adherent dark-green crystals. There were additional green crystals within the coolant salt that had frozen close to the coupons. The green crystals were identified by x-ray diffraction as Na3CrF¢. Micro- 136 scopic examination of crystals removed . from the coupons also showed brown and black solids, as yet " unidentified. - The chromium coupons sustained a weight loss in the experiment. From the analyses for H,(g) and for -moisture, approximately one-fourth of the Cr weight loss can be assigned to the reaction - Cr+%H,0 + 2NaF + NaBF, ~ Na;CrFg + %,B,0; + %H,(g) . At 900°K, this reaction has a AG® = —30 % 10 kcal. The uncertainty of 10 kcal is primarily the uncertainty in ‘the free energy of formation of Na; CrF,. . If one represents the total analyzed oxygen as O,(g), then four-fifths of -the weight loss of Cr can be associated with the reaction Cr+ 3/402(3) + 2NaF + NaBF, > Nas CrFg + % B,0; - The AGjqo for this reaction is —100 + 10 kcal. It is 1. Some of the moisture in the salt reacts with ‘chromium metal to yield hydrogen. 2. The hydrogen thus produced passes through the nickel container in a matter of minutes. 3. One certain corrosion product is NagCrFs. 12.5 BORON TRIFLUORIDE PRESSURE OVER FLUOROBORATE COOLANT SALT ADMIXED WITH MSBR FUEL SALT D. M. Richardson J. H. Shaffer Several experiments have been reported in which the pressure of BF; over mixtures of NaBF,-NaF became higher than normal due to the presence of fuel fluoride salt.!»* Greater pressure elevations occurred when the proportion of fuel fluoride was greater. In the MSBR design the fluoroborate salt is maintained at a higher pressure than the fuel salt in the primary heat ex- changers, so that any leakage would be inward. BF; in the fuel salt as the result of a small leak would be removed with the normal reactor off-gas. In the event of a large leak a provision for additional venting capability to relieve BF; pressure might be required during the course of shutting down for repair. For an ~early design of MSBR primary heat exchanger it was estimated that in the extreme case that a tube suffered total offset at both ends, the fluoroborate salt would be injected into the fuel salt in the volume ratio of 1 to 3. This postulated condition formed the basis for a simple exploratory experiment to observe the release of BF, - gas when fuel and fluoroborate fluids are mixed. likely that only a small fraction of the analyzed oxygen in the salt was in the form of molecular oxygen. It is ‘possible that the oxygen analysis actually provides a more reliable H,O analysis than the routine Karl Fischer method. The chief uncertainty in the H,O analysis is whether or not all the moisture in the sample has been removed by the pryldme which is subse- quently distilled into the Karl Fischer reagent. (It is known that not all of the salt sample dissolves in the - pyridine.) No firm conclusions concerning the nature of the oxidation reactions of chromium by the coolant salt can be reached on the basis of this first experiment. It is anticipated that experiments currently in preparation will provide more quantitative information, especially of the hydrogen passing through the capsule. However A stainless steel autoclave was constructed from 2Y%-in. sched 80 pipe, end cap, and 2%- by l-in. reducer and was installed in a vertical Marshall furnace. The vessel was connected above the furnace by a 1-in. sched 160 pipe to a 300-1b flange with metal O-ring. Design and specifications were supplied by the Design Department, Reactor Division (Dwg 10564-D). Design conditions were 100 psig, 1300° F Hydrostatlc pressure ' test was at 1300 psig. Temperature measurements were made with three * stainless-sheathed Chromel-Alumel thermocouples ob- - stainless shim plates. They were positioned 2 in., 7Y, tained from the MSRE instrument group. They were held against the autoclave vessel wall with spot-welded from this initial rexpenment the following may be concluded: 'A. N. Smith and P..G. Smith, MSR Program Semiann. Progr. Rept. Aug. 31, 1968, ORNL-4344, p. 75. 2D, M. Moulton and J. H. Shaffer, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 127~28. ih., and 13% in. from the bottom of the vessel and indicated, respectively, the fuel salt, fluoroborate salt, 137 and gas-phase temperatures. Although the temperatures . were printed by recorder, actual readings of the recorder pointer position were more accurate and were used. Pointer indications were found to check within 5°F with measurements made with an ESI potenti- ometer, model 300. The autoclave was connected by flange to a Y-in. stainless tubing manifold for evacuation and for pres- sure measurement after the system was isolated. Metal- seated 1500-psi valves with welded bellows were used. Vacuum integrity was established, using roughing pump and thermocouple gage, with the autoclave vessel at 1300°F and with the upper manifold under heat lamp. Pressure measurements were made with a 200-psig Heise gage, with uniform graduations, that was adjusted to read 0 psia when under vacuum. The salt additions were made under dry argon. The fuel salt was transferred as a liquid from a batching vessel: 1158 g of LiF-BeFThF, (72-16-12 mole %) and UF,; to produce 0.5 mole % were charged. After cooling, the fluoroborate was added as crushed salt: 215 g of NaBF,-NaF (92-8 mole %) was charged. The ‘estimated volume ratio of fluoroborate to fluoride at 700°C for these additions is 1 to 3. The system was sealed again at the top and evacuated. The autoclave was heated to 150°C and pumped for 3% hr. Little off-gassing was observed. The system was then valved off, remaining open only to the pressure gage. One of the valves was connected to a % -in. tube for venting to the stack filters, if necessary. When the fluoroborate was first melted (385°C) a. pressure rise of 0.3 psia was observed. When the fluoride was first melted (513°C) the pressure was 8 psia. Subsequently, at the same temperatures, these pressures were not reproduced and higher pressures - were found. Two distinctly different sets of data were obtained in this experiment. Initially, during a period of six days, pressure readings were made repeatedly over a range of temperatures. Some of the points were obtained after - equilibration for 16 to as much as 46 hr. Due to the slow rate of thermal equilibration of the furnace- autoclave combination, however, most of the points were obtained at a slow heating rate of 20 to 40°F/hr. These data are shown as circles in Fig. 12.4 in a plot of log P against 1/T°K. With the exception of one point the data lie on a straight line with apparent heat of reaction of 15 kcal/mole. Temperatures of the bottom thermocouple are plotted. The point at 148.5 psia was discovered 25 min after the last of the series of points at the upper end of the line was taken. The heating rate had been 5 to 10°F/hr for several hours. The temperature record just prior to this showed that the bottom thermocouple (always the hottest of the three, during slow changes) had dropped 12°F and that the middle thermocouple had risen 6°F in an interval of about 4 min. These temperature ORNL- DWG 70~ 86767 TEMPERATURE (°C) 400 . 500 600 - 700 200 - I | l | L1 T b SLOPE = 20 kcal moley’ e . 100 f 90 » i 1 80 § / ' if ‘ w 60 s g & / o0 e |F ) @ W 50 e a ® . 8 o g - ’ ‘0' 3 40 . v / L_L . - . * o = e g 3 30 = & e 4 ‘ . * - 20 Y / 9 o SLOPE = 15 kcal/mole 9 {0 - . 1.5 1.4 1.3 t.2 1.¢ 1.0 ’ GOOO/r(oK) 'Fig. 12.4. Pressure of Coolant and Fuel Mixture. 138 changes suggest that sudden mixing or stirring of the two fluids had occurred. The heaters were turned off, dampers at the top and bottom of the furnace were opened, and the system was vented down to 100 psia. A cloud of white smoke came from the roof stack. Although the autoclave was cooling, the pressure rose again to 108.7 psia in 10 min but dropped steadily afterward. The total volume of the isolated system was 1600 cc. Of this, 300 cm® was outside the furnace. When the pressure was 95 psia the bottom temperature was 647°C, and the combined salt volumes are estimated at 482 cm®, leaving 818 cm® in the furnace. Using the ideal gas law, the amount of BF; in the gas phase was 0.15 mole at 647°C, or 8% of the total inventory (1.895 “moles BF3). Subsequently, the autoclave was heated up and then cooled to ambient temperature three times. In each case an identical line was obtained with slope of 20 kcal/mole on the log P versus 1/T°K plot. These data are shown as solid dots in the figure. When extended the Hp — H,95 (cal/mole) are given in Table 12.1. Table 12.2 summarizes the data for the heats and entropies of melting and transition. KBF,4, RbBF,, and CsBF, all undergo a solid state transition from the BaSO,-type orthorhombic structure ~ to the high-temperature cubic structure.)”® NaBF,, on the other hand, exists as the orthorhombic (pseudo- tetragonal) CaSQ, type of structure® at room tempera- ture and in a noncubic form above the transition temperature.*> This latter structure has recently been reported® to be monoclinic with four molecules per unit cell. However, a lowering of the symmetry at higher temperature without a lowering of the number _ of molecules per unit cell is quite unlikely. The structure was derived from a powder pattern, and the authors state that their assignment is not necessarily correct or unique, We believe that this inconsistency . may be avoided by showing that the x-ray spacings line is slightly below the originally observed maximum pressure of 148.5 psia. At the lower temperatures each reheat yielded a different set of pressures that varied - with the residual pressure at room temperature before the heatup was started. In summary, it is apparent that the originally ob- served “discontinuity” was probably the result of sudden spontaneous mixing of the materials in the autoclave that had not previously been in actual equilibrium, since no stirring or shaking was provided. It was not possible in this experiment to determine either that two immiscible liquid phases were always present or that a solid phase had not appeared. The only definite result is that the pressure of BF; over the mixture was more than an order of magnitude greater than that over NaBF,-NaF (92-8 mole %). The effect of the relatively large gas volume in the experimental system would be to reduce the mole fraction of NaBF, and to shift the mixture. closer to the liquid-liquid immiscibility boundary. A very thorough experimental investigation is required to define this complex system adequately. 12.6 HEAT CONTENT OF ALKALI METAL FLUGROBORATES A.S.Dworkin M. A, Bredig We have cbmpleted our high-temperature heat con- tent measurements of NaBF,, KBF,, RbBF,, and CsBF,. The equations which represent our results for (Table 12.2 and ref.5) are compatible with the assump- tion of a mechanical mixture of the orthorhombic low-temperature phase with a high-temperature phase of hexagonal, rather than monoclinic, structure. These data give a ¢/a = 1.55, which may be compared with the }ugh-temperature form of CaSO,, also shown® to be hexagonal rather than cubic, with a c¢/a = 1.54, that is, slightly distorted from the ideal “close packed” sym- metry, ¢/a = 1.63, as in wurtzite, ZnS, Although the entropies of fusion of NaBF, and KBF, are similar (Table 12.2), the entropy of transition of KBF, is much larger than that of NaBF, (5.93 to 3.13 e.u.). This may be explamed qualitatively on the basis of the differing structures of both the low- and high-temperature solids which in turn are due to the large difference in the size of the cations. The NaBF, structure is more disordered below the transition, which is reflected by the fact that at SO0°K Sy — S;9s is more than 1 e.u. larger for NaBF,; than for KBF,. In addition, the high-temperature cubic structure of KBF, most probably is more compatible with anionic rota- tional or librational disorder than is the structure of NaBF, of lower symmetry. ' - 1M, J. R. Clark and H. Lynton, Can. J. Chem. 417, 2579 (1969). 2D, 1. Huettner, I, L. Ragle L. Skerk; T. R. Stengle, and H., c Yeh,J. Chem. Phys. 48,1739 (1968). 3C Finbak and O, Hassel, Z, Physik. Chem, 323, 433 (1936). ‘G. Brunton, Acta Cryst. B24, 1703 (1968). 5C.W.F.T. Pistorius, J. C. A. Boeyens, and J. B. Cla:k Htgh Temperatures-High Pressures 1, 41 (1969). SM. A. Bredig, Chem, Div, Ann, Progr Rept. May 20 1968, 0RNL-4306 p. 129, KBF,, RbBF,, and CsBF, are isodimorphous and therefore can be considered as a series separate from NaBF,. Table 12.2 shows that although the enthalpies and temperatures of melting are similar, there is a decrease in temperature, entropy, and enthalpy of transition of 21, 25, and 41%, respectively, with increasing cation size in this series. The particularly ' latge relative change in the enthalpy may be attributed to the decrease in lattice energy with increasing size of the cation, which also facilitates the rotation or libration of the fluoroborate ion. The large decrease in enthalpy as compared with the entropy of transition also explains the relatively low temperature of transi- tion found for CsBF,. Table 12,1. Equation Coefficients for Enthalpy Data for Equation: | Hp — Hyog (cal/mole) =g+ bT +cT> +dT™ -3 . S : 2 -4 Average Temperature Com?ound e X 10 b ¢ X 10° ‘d X10 Per Cent Exror Range °K) NaBF, —-3.820 3,148 3.703 -12.17 - 03 298-516 -9.785 36.48 : 0.1 516—-679 -8.605 39.52 , 0.1 _ 679-750 KBF4 —6.325 15.62 : 1.943 -1.737 0.2 298-556 ' -7.800 34.95 0.2 556843 - =1.710 39.94 0.1 843-900 RbBF,4- -7.430 18.77 1.697 +9.677 _ 0.2 298-518 -1.897 34.34 0.1 518-855 - —7.985 - 39.92 L 0.1 855-1000 CsBF,4 - -7.614 17.63 2,009 +17.01 0.2 298-443 —8.673 34.18 ' 0.1 443-828 - —8.390 39.36 0.1 828-1000 Table 12.2, Heats and Entrofiies of Melting and Transition of Alkali Metal Fluoroborates Compound '{m _ - AH,, : Asm T, .. AH, AS,, Cp(li-qluld) . (K (kcal/mole) (eu/mole) K) - (kcal/mole) (eu/mole) (cal deg "mole ) NaBF, 679 3325 4.78 516 1.61 3.1 39.5 KBF,4 843 4.30 5.1 556 3.30 5.9 39.9 RbBF4 855 - 4.68 5.5 518 - 2.86 5.5 39.9 CsBF; - 828 458 5.5 443 1.94 4.4 39.4 13. Physical Chemistry of Molten Salts 13.1 EQUILIBRATION OF RARE-EARTH- CONTAINING MOLTEN FLUORIDES | WITH VARIOUS SOLIDS C.E.Bamberger C.F. Baes, Jr. We have continued our attempts® to prepare insoluble compounds of rare earths for possible use as rare-earth ion exchangers by reactions in molten LiF-BeF,-ThF, (72-16-12 mole %). The new tests have included phosphates and various insoluble oxides as possible reactants. In addition to CeF;, used previously with 144Ce as a tracer, YF; was used in current tests because it chemically resembles the rare earths in the middle of the series and can be determined spectro- graphically with good sensitivity. As previously, tested reactants were stirred with the molten salt containing the dissolved rare earth. From time to time samples were taken to determine if rare earth had been removed from the salt and whether new solid phases might have formed. , o ' A number of the reactions which are most likely to occur if the compounds currently being tested are effective extractants for rare-earth cations are listed in Table 13.1. In previous tests, reaction 1 was found not 1, E. Bamberger and C. F. Bacs, Jr., MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL 4449, p. 131. to occur with CeF; in the -presence of excess NiO, added to suppress the direct oxidation of the nickel . container by V,O;. Because of concern that NiO might instead have reacted with V, 05 to form Ni(VO;),,an additional test was run with V,0Og in the absence of NiO. As Table 13.2 shows, there was no evidence of cerium or yttrium removal. As previously, considerable ~ corrosion of the nickel container was noted. Reactions 2 to 6, which similarly suggest the forma- tion of oxide phases containing rare earths by reaction of sparingly soluble oxides, similarly were found not to occur (Table 13.2). Since one of the most abundant rare-earth minerals is monazite (thorium-—rare-earth phosphates), we at- tempted the synthesis of such. compounds from the melt by the addition of sodium phosphate (reaction 7). Neither a significant reduction in the !44Ce concentra- tion nor any other evidence of appreciable reaction between the phosphate and the molten fluorides was detected. Calcium fluoride was added to the system in an attempt to. precipitate a rare-earth-containing cal- cium phosphate (reaction 8). This seemed a possibility since a systematic study of rare-earth minerals by Mineyev? includes examples of the replacement of Ca®* by Ln* plus OH™ or F~ jons. The results obtained 2D. A. Mineyev, Geokhimiya No. 2, p. 237 (1968). Table 13.1. Possible Reactions for Extraction of Rare-Earth Cations from Molten LiF-BeF,-ThF, N A WN o - . WO3(e) + W(c) + . ¥,v,05(c or ) + % ThO, + Ln* =% Th* + LnVO,(c) . 2Ti0y(c) + %, ThO, () + 2Ln* = ¥, Th* + Ln,Ti, 04(c) . ¥,Nb,05(c) + % ThO,(c) + Ln* > ¥ Th*" + LnNbO, (c) . %,C1,03(0) + % ThO,(¢) + Ln® > % Th* + LnCi0,4(c) . 3WO03(c) + % ThO,(c) + 2Ln> > Th* + Lny(WO4)3(c) 3x+1) 3+_>3(x +1) ThO,(c) + 2xLn Th* + 2Ln WO, (c), where x <1 . NasPO4(c or ) + Ln® - LnPO, (c) + 3Na* . Caz(PO4),(c) + Ln® + F™ - Ca,Ln(P0,),F(c) + Ca® . nCa® + mLn* + (3m + 2n)F ~ ~> nCaF,'mLnF; (ss) " 140 Table 13,2, Tests of Rare-Earth Extraction from Molten LiF-BeF,-ThF, (72-16-12 Mole %) by Various Compounds Reagent Added to _ Ph.a.ses P.resenf After » Temperature RS:nr;?nni‘ng Rt;t;l:un:lg 7 Solute Concentration in Melt the Melt® Equilibration with Melt o in Melt (%) in Melt (%) (Mole %) 80 V,O5 + 14 ThO, V,05(]) + ThO, + [Ni(VO3),] 660 96 100 Ce*=0.1,Y*= o. 9, V¥ = 0.2, Ni* = 0.4 60 TiO, TiO, 600 100 100 Ce* = 0.7, Ti"* = 0.005 +TYFy TiO4 600 100 100 - Ce*=0.7,Ti" =0.009, Y*=07 +27 K, TiFg +40ThO; TiO 720 100 100 - Ce* =0.7, Ti"* = 0.005, Y* = 0.7 20 K, NbF, + 40 ThOp Nb,0s 600 100 100 Ce* =0.1,Y* = 1.0, Nb™ = 0.06 +21 CtF3 + 18 ThO, [Nb,O5 + Cry03] 600 100 100 T Ce*=0.1,Y* = 1.0,N6™ = 0.01,Cr* = 0.13 +11 NiF; +6 ThO, [Nb,O5 + Cr,03 + NiO] 600 100 100 Ce* =0.1,Y* = 1.0, Nb™ = 0.01, Ni** = 0.04 18 Na, WO, + 20 ThO, [WO3] + ThO, 600 100 100 Ce*=0.1,Y¥=1.0,wt =121 W [WO3 + W] + ThO, 600 100 100 Ce*=0.1,Y*=10,w =095 60 Na3PO,4 | 600 97 - ND¢ Ce¥*=0.1,P0,* =3.7 + 100 CaF, [CaFa-YF3} (ss) 600 7 ND - Ce™=0.1,Ca%=8.9,P0,> =56 130 CaF, CaF, 600 | Ca¥*=100 | | +0.001 CeF3 CaF» 600 98 . Ce* =0.1,Ca%" = 10.0 - +100 CaF, CaFp 600 99 Ce* =0.1,Ca™ = 10.0 +120YF3 [CaF2"YFs3](ss); xCapz(ss) = 0.69 600 78 53 Ce*=0.1,Y*=5.1,Ca® = 10.0 “Numbers indicate millimoles of reagent added per mole (63.2 g) of LlF-Bng-ThF4 (72-16-12 mole %). bPhases in brackets are presumed to be present but were not identified. °ND: not determined. o 1 revealed an appreciable extraction of rare earths. Analysis of the salt phase also suggested the presence of Ca** and PO, % in solids, but these were not further identified (Table 13.2). Reaction 9 was suggested by the phase diagram of the CaF,-YF; system, wherein a solid solution containing as much as 55 mole % YF; as well as the compound CaF,-4YF; is reported.® The solubility of calcium at 600°C in the presence and absence of PO, (8.8 and 7.8 mole %, respectively) and the percentage of cerium remaining in the melt (78 and 76) may be interpreted to mean that the solid phases present in both experi- ments were solid solutions of CaF,-YF; with some CeF3. Under experimental conditions studied, the | distribution of cerium seems insufficiently favorable and the solubility of the calcium-containing solids too high to render reactions 8§ and 9 a promising basis for a separation method for rare earths. | While these tests have all given essentially negative results, they have produced some useful information. As expected, TiO, and Nb,Os, currently being studied for the first time, along with Cr, O; and Fe; 0, , studied previously, react only to a small extent with the molten fluoride. It appears that the orthophosphate ion is stable in such systems and does not act as an oxidant toward nickel containers. Finally, the solubility of CaF, in the salt used (72-16-12 mole % LiF-BeF,- ThF,) is evidently 10 mole % at 600°C. ' 132 THE OXIDE CHEMISTRY OF PLUTONIUM IN MOLTEN FLUORIDES C.E.Bamberger R.G.Ross C.F.Baes,Jr. Since plutonium ‘is being considered as a fuel for moltensalt reactors,! it is important to examine its chemistry in molten fluoride systems containing oxide. Such information is needed to establish in a reactor fuel salt the conditions of redox potential and oxide concentration under which plutonium would remain in solution as PuF; . It might also be of value in developing reprocessing methods for such fuel salts. From the available data®™® it is possible to construct the Pourbaix diagram shown in Fig. 13.1, which summarizes the predicted behavior of plutonium in molten Li, BeF,; as a function of the oxidation poten- tial (the ordinate) and the oxide ion concentration (the abscissa). In addition to the three oxides of plutonium (PuO, 5, Pu0, 4, and Pu0O;) whose stabilities have been determined,® it is possible that PuOF may be 33, Short and R, Roy, J. Phys. Chem. 67(9), 1861 (1963). 142 formed in the present system,* but there is insufficient evidence to judge its stability. The boundaries of the region of dissolved PuF; in this diagram are calculated for two concentrations; one is the hypothetical unit ‘mole fraction and the other is 0.002 mole fraction, which is approximately the concentration of PuF, which would be used in a molten=salt reactor fuel. These boundaries are uncertain by as much as plus or minus one log unit along each axis from the uncertamty given for the free energy of formation of PuF;.* Within ‘such a wide uncertainty one cannot rule out the possibility that plutonium might precipitate as PuQ; s from a molten-salt reactor fuel which becomes contami- nated with sufficient oxide. At lower oxide concentra- tions, but under relatively oxidizing conditions, PuO, clearly could be precipitated. With the ThF,4 present in an MSBR fuel, the formation of a PuO,-ThO, solid . ‘solution becomes possible, and this should lower the oxide level or the oxidation potential at which a plutonium oxide-containing phase might precipitate. Because of the uncertainties in the behavior repre- - sented by Fig. 13.1 we were prompted to attempt a direct measurement of one of the equilibria which establish the boundaries of the dissolved PuF; region in a diagram such as this. The result would not only improve the accuracy of such a diagram for plutonium; it should also yield an improved value of the formation free energy of PuF;. To various molten fluoride mixtures containing PuF;, stepwise additions were made of an oxide (BeO, ZrO,, or ThO,) known to be the normally stable saturating oxide phase. The equilibrations were performed in a specially designed leak-tight stirred vessel, described elsewhere,® which had been assembled inside a glove box suited for work with alpha emitters. Filtered samples of the fluoride phase and unfiltered samples of both oxide and fluoride phases were taken for analysis. _ 1R, E. Thoma, Chemical Feasibiliiy of Fueling Molten Salt Reactors with PuF3, ORNL-TM-2256 (October 1968). C. 1. Barton, J. Phys Chem 64, 306 (1960); J. A. Fredricksen, L. O. Gilpatrick, and C. J. Barton, Solubility of Cerium Trifluoride in Molten Mixtures of LiF, Ber and ThF4, ORNL-TM-2335 (January 1969). 3W. T. Ward, R. A. Strehlow, W. R..Grimes, and G. M. Watson, J. Chem. Eng. Data 5, 137 (1960). 4F. L. Octting, Chem. Rev. 67,261 (1967). 5c. F. Baes, J1., p. 617 in Symposium on the Reprocessing of " Nuclear Fuels (Nuclear Metallurgy vol. 15), ed. by P. Chnottl, USAEC, 1969. C. E. Bamberger, C. F. Baes, Jr., T. J. Golson, and J. Nicholson, Reactor Chem. Diy. Ann. Progr. Rept Dec. 31, 1968, ORNL-4400. 143 - / / . OoRNL-DWs 70-6769 g PUF3 (d) I i b i fog XU‘H’/XUS"' —6‘— B — Xpyg,=1-0 -0 — 0.002 — -6 Pu(c) -4 o [ ] - a -6 |- _ | | 1 I i 1 ] | -8 -7 -6 -5 -4 -3 =2 -1 o log on- ' ' Fig. 13.1. Pourbaix Diagram for Plutonium in Li;BeF, st 600°C Based on Literature Data (Refs. 13 of Sect. 13.1and Refs. 1 - ‘and 9 of Sect. 13.2). Boundaries of the PuF3 and PuF, regions indicate conditions of oxidation potential and oxide concentration under which these components, dissolved at the concentrations indicated, are at equilibrium with another phase. The heavy vertical In the following molten salt mixtures, which con- tained up to 0.54 mole % PuF,, no detectable amount of dissolved plutonium was found to be precipitated by additions of the indicated stable oxide phases: LiF-BeF, -ZtF, (65.7-29.3-5 mole %) + ZrO, LiF-BeF, (66.7-33.3 mole %) + BeO line at the bottom indicates the upper limit of oxide concentration set by the solubility of BeO. LiF-BeF, (66.7-33.3 mole %) + BeO + NiO 'LiF-BeF, -ThF, (72-16-12 mole %) + ThO, However, when NiO was added to the last salt in the presence of ThO,, there was a decrease in the dissolved PuF, concentration, presumably as a result of the reaction ' ' . 144 Table 13.3 . Equilibrium Data for the Reaction PuF3(@) + % ThO, (ss) + | NiO(c) ®Pu0, (ss) + % ThF4 (@) + %,Ni’ () : X X. 3/4 Temp:rature .. XpyuF ¥ PuO, ( ThF, o bt Pu0, XPqu XTho, 590 0.00210 * 0.000119 0.0894 ¥ 0.00282 9.44 ¥0.84 : 0.00216 £ 0.000119 0.0878 ¥0.0028? 9.00 ¥0.80 0.00253 % 0.00015% 0.0781 ¥0.0039¢ - 6.76 ¥0.72 £ 0.00175 £0.00019% 0.0983 ¥0.0049¢ 12.58 ¥1.87 | | Weighted mean 8.51+0.44 690 0.00469 * 0.00024% 0.0219 ¥0.0061% 0.97 ¥0.34 0.00475 * 0.00004% 0.0204 ¥0.00104 0.89 ¥0.05 Weighted mean 0.89 £0.05 “From analysis of filtered fluoride phase. By material balance. “From lattice parameter measurements of the oxide phase by x-ray diffraction, dFrom alpha pulse-height analysis of the oxide phase. Pqu (d) + 3/4 ThOg (SS) + 1/2 NIO(C) = PuO, (ss) + ¥ ThF, (@) + % Ni(c) (1) (d, ss, and c denote, respectively, the molten fluoride, solid solution, and pure crystalline phases). Samples of the equilibrated oxides, washed free of fluoride, gave x-ray powder patterns characteristic of the (Pu-Th)O, solid solution, confirming that the above reaction had occurred. - Equilibrium was assumed to have been attained when the PuF, content of filtered samples became constant with time within the analytical scatter of ~+5%. Final values of Xp g, along with values of Xpyo, calculated by material balance are shown for two temperatures in Table 13.3. The table includes, as well, oxide samples whose composition was determined by alpha pulse-height analysis or from the lattice parameter determined by x-ray diffraction. The equilibrium quotient of reaction (1), _(XThF4) 0=\ % Th02 was calculated from these results and from values for Xrnr, and Xppo, which were known from material 4y , Pu0O, (2) XPuF3 . balance. It can be 2seen in Table 13.3 that the largest analytical uncertainties occurred in the composition of the equilibrated (Pu-Th)O, phase at 590°C, and these are reflected in the corresponding uncertainties of XpuF, and Q. The temperature has a marked effect on the value of the equilibrium quotient; between 590°C and 690°C, Q changes by a factor of 10. The resulting values of Q for reaction (1) were used to estimate the free energy of formation of PuF; in the Li, BeF4 reference salt. The free energy of reaction (1) in this solvent is given by ' AG, =—RTIn (Qg%{;‘Fn;/gP“F?') > (3) -wherein activity coefficients in the solid solution are assumed to be unity and gry,p, and gp, y, are activity - coefficients in the molten salt used for the measure- ment, LiF-BeF,-ThF, (72-16-12 mole %). These are presently estimated as 0.40 and 0.35 respectively (Sect. 13.6). Combining AG, with the following formation free energies: AGpyo, = —252.67 +45.90 (7/1000) £ 2 kcal/mole (ref 4) AG°ThO =-292.40 +44. SS(T/ 1000) %1 kcal/mole .'(ref. 7) AGY;0 = —56.26 +20.35(T/1000) + 0.2 kcal/mole (ref. 8) AGip, (L2 B) = —491.19 + 62.41(7/1000) | + 22 _kcallmole (ref. 5) 0. Kubaschewski, E. L. Evans, and C. B. Alcock, Metallur- gical Thermochemistry, Pergamon Press, 1967. ®). F. Elliott and M. Gleiser, Thermochemistry for Steel- making, American Iron and Steel Institute, Addlson-Wesley Publ. Co., Reading, Mass., 1960. we calculate the following formation free energies of PuF; in Li, BeF, in the hypothetical unit mole fraction standard state: AGyp, g, (L, B) = ~326.96 + 2.6 keal/mole (590°C) = —325.67 + 2.6 kcal/mole (690°C) The assumption of ideality of the PuO,-ThO, solid solution probably introduces a negligible uncertainty, as does any error in the estlmates of actlv:ty coefficients in the salt phase. Combining the above values of AGp, g 4 With the free energy of solution of PuF;, , AEPuFa(]'Q B) - AG;an =14.41 — 6.05(T/1000) 0.09 (based on Barton’s solubility measurements?), we ob- tain the following formation free energies of PuF;(c): " AGpyp,(€) = —336.11 + 2.6 keal/mole (590°C) = _334.46 £ 2.6 kcal/mole (690°C) The values are 16 to 20 kcal more negative, in the temperature range studied, than values reported by Oetting in his review on thermodynamic properties of plutonium compounds.® The values of Oetting are those calculated by Buyers and Murbach?® in 1957 from measurements of the equilibrium 1 Buyers and Murbach used the free energies of formation per fluoride of both UF, and UF; available at that date (88 and 89 keal at 1573°K, respectively) and concluded that an accurate determination of n was not necessary since it did not affect significantly the calculation of AGPuF However, using more recent estimates of the correspondmg free energies of formation of UF, and UF, (86.5 and 93.0 at 1573°K),® it is apparent that UF; was the most stable fluoride present under the -experimental conditions of Buyers and Murbach. This information together with their experimental data yields values of AGp, g, which are consistent with the present estimates. %A. G. Buyers and E. W Murbach Nucl. Sci. Eng 2 679 (1957) J K. Dawson, R. M. Elliot, R. Hurst, and A. E, Truswell, J. Chem. Soc. 1954, 588. TURO+sh+inmelun. @ By combining our value of AGPuF with the equi- libria 4PuF, + 0, 2226 =——=3PuF, +Pu0, ;) vacuum 7 studied by Dawson et al.!® we obtain the following free energy of formation of PuF,(c): AGp,p, =~—387.92 +9.95(T/1000) + 2.8 kcalfmole . A revised Pourbaix diagram which incorporates these new formation free energies of PuF; and PuF, is shown in Fig. 13.2. The considerably more negative values of these free energies cause the boundaries between the dissolved fluorides and the solid oxides of plutonium to shift about three decades to the right, toward higher dissolved oxide concentrations. This result makes it abundantly clear why PuQ, s could not be precipitated from Li, BeF, saturated with BeO. Figure 13.3 shows the predicted behavior in the LiF-BeF,-ThF, (72-16-12 mole %) melt, with various compositions shown for the Pu0,-ThO, solid solution. It is apparent that under the mildly reducing conditions (Xy g, /Xy, ~ 100) of an MSBR fuel, even if an oxide phase were inadvertently precipitated it would contain no significant amount of PuO,. 13.3 THE CeF;-ThF, SYSTEM L.O.Gilpatrick H.Insley C.J, Barton Exploration of the CeF;-ThF; system was under- taken as part of the longrange investigation of the LiF-BeF,-ThF, -CeF3 system. It should, however, have considerable scientific interest because, except for thLe system UF;-UF,,! little is known about the binary -systems of the heavy-metal tri- and tetrafluorides. Our studies have been limited to gradient quenches because the high liquidus and solidus temperatures for the system exceed the capability of our present differential thermal analysis equipment. The end mem- bers were purified by fusion with ammonium bifluoride to remove oxides, and weighed quantities of the purified materials were blended to make mixtures containing 10 to 95 mole % ThF,. The mixtures were sealed in the standard thin-wall nickel quench tubes and heated in a furnace having a platinum—10% rhodium heating element. The maximum temperature of these 1G. Long _émd R. E. Thoma, Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, p. 5. 146 12 T 1T 1 | I | 1 [ 10 |- ' o § .“v o 8 - E| | x| | 6 |- | PUE () — 1 4 4 | - 2—.— 0..... ORNL-DWG 70-6770 _4 | — ] — PuF3 (d) i’: , ~ — 5 X -0 |~ —12 N o Pu(c) L4 2] | | | I I | | | | I -7 -6 -5 -4 -3 2 - 0 1 2 log Xg2— Fig. 13.2. Revised Pourbaix Diagram for Plutonium in Li;BeF, at 600°C Based on Present Results. The boundaries of the PuF; and PuF, regions are shifted downward and to the right, reflecting changes of the free energies of formation for these compounds to more negative values. quenches was about 1200°C, and the -equilibration period varied from 1% to 3 days. | Rather limited data have been obtained because of ‘difficulty experienced with self-welding between the nickel quench tubes and the nickel block that surrounds the tubes. The data obtained to date show that ThF, is the primary phase in mixtures containing more than 81 mole % ThF,. There appears to be a eutectic composi- tion formed between ThF, and a new phase, pre- sumably a compound that is as yet unidentified, meltmg at 975°C. Another unidentified phase appears as the first crystals to separate from melts containing 67 and 70 mole % ThF,. There is also evidence of solid - solution of ThF, in CeF;, but the extent of solid solution formation remains to be determined. We have modified our high-temperature quench fur- nace by introducing a thin graphite liner into the nickel quench block in an effort to overcome the self-welding problem. If this is successful, study of this system - should be completed in the near future. ORNL-DWG 70-67T74 — — qo— Xqu4= 10 |— . PUF4(d) o | Xoury Xpur, =10 1 Xpuy=0.002 XPUF4/XPUF =2.2 310_ , ¥ log Py /P"'zz I N | PuFy(d) Xing,=0.12 -12 |- - Pulc) L 1 ‘ 26 10 iy o B0 —— e — i ~ Tho, (c) o log X4+ /X 3+ | NN Thie) -7 ' -6 - -5 . log Xo2=- ' -4 o =3 Fig. 13.3. Pourbaix Diagram for Plutonium in LlF-Bng-ThF4 (72-16-12 mole %) at 600°C Based on the Present Results, The upper limit of oxide concentration is set by the solublllty of ThO,. Pqu-Th()g solid- solutions are formed at the hlgher oxidation potentlals ' 13.4 EQUILIBRIUM PHASE RELATIONSHIPS IN THE SYSTEM LiF-BeF, -CeF L.O.Gilpatrick H.Insley ~C.J.Barton Investigations of phase equilibria in the system 'LiF-BeF, CeF3; have continued. Advances were made principally through use of the thermal gradient quench technique, This system is of interest in part because of its presumed analogy with the system LiF-BeF,- PuF,,!»? which has some interest as 2 possible fuel composition for molten-salt reactors.’ Composmons W. T. Ward, R, A. Strehlow, W. R. Grimes, and G. M. Watson, Solubility Relations Among Rare Earth Fluorides in Selected Molten Fluoride Solvents, ORNL-2749 (1959). 2C. J. Barton, M. A. Bredig, L. O. Gilpatrick, and J. A. Fredncksen, Inorg. Chem. 9, 307 (1970). 3R.E. Thoma, Chemical Feasibility of Fueling Molten SaIt Reactors with PuF3,ORNL-TM-2256 (Jan. 20, 1998) _ ~were chosen to help define the lower-melting pfiase boundariés. This resulted in studying compositions containing less than 10 mole % CeF;. The cerium fluoride primary phase field dominates the system, as might be expected from its high melting point of 1460°C and from the fact that LiF and CeF3; do not form intermediate compounds. This field approaches within 0.5 mole % of the low-melting LiF-BeF, binary composition ranging from 33 to approximately 60 mole 148 - studying the binary system -UF,-UF;. % BeF,, where the freezing temperatures of LiF-BeF, compositions are less than 480°C. An earlier estimate® that the maximum solubility of CeF3 and PuF; in an LiF-BeF, melt containing 52 mole % BeF, is approxi- mately 0.5 mole % at 550°C is therefore confirmed. "The present findings are also consistent with the results of solublhty measurements with CeF; in LiF-BeF,- - ThF, melts.’ Because of the low solubility of CeF; in the ternaryr ‘system in this composition region, there is little difference between the minimum melting point of the " LiF-BeF, binary system (360°C)® and that of the crystal transition at a temperature of 841°C. Their observations were made by thermal analysis while Since poly- morphic crystal transitions have not been previously reported in this widely studied compound this report aroused some interest. Attempts were made to confirm this observation using differential thermal analysis apparatus which we developed recently for molten-salt studies.3 A sample of depleted UF, prepared at the Oak Ridge Gaseous Diffusion Plant and designated SR-1909, batch 4, was examined first, in the temperature range 700 to 1010°C with a differential temperature sensitivity of 50 pvfem (~1.25°C/cm) and at a rate of 1°/min. A laboratory-purified lot of UF, used as a spectrographic standard was also examined under identical conditions. No evidence of a thermal phenomenon could be found at or near 841°C on repeated thermal cycling. However, a single endothermic peak was found at '958°C on heating the Sr-1909 batch 4 material during LiF-BeF, CeF; ternary, currently estimated to be . 358°C. The composition of this ternary invariant point is yet to be determined accurately, but it lies near the composition LiF-BeF, -CeF; (46.5-53-0.5 mole %). Some evidence has been found for two immiscible liquids in equilibrium in a small region of this system which is rich in BeF; and near the CeF; -BeF, eutectic composition, currently estimated to contain 6 + 1 mole % CeF 5. This eutectic has a melting point of 540 + 5°C. An equilibration time longer than the one to three weeks currently used may be needed to attain equi- librium in this system, where equilibrium is attained . slowly at those compositions high in BeF, due to the high viscosity of the melts. Further work will be needed to substantiate these findings and to define the boundary of the region of immiscibility. 13.5 AN INVESTIGATION OF POSSIBLE POLYMORPHIC TRANSITIONS IN URANIUM TETRAFLUORIDE L.O. Cilpatrick Crystalline uranium tetrafluoride - was reported by Khripin and co-workers'*? to undergo a dimorphic 4C. 1. Barton, J. Phys. Chem. 64,306 (1960). the first cycle only. The spectroscopic standard material showed no. thermal effect below the melting point. Melting temperatures for the two materials were 1010°C and 1020°C respectively (literature value 1036 C4 %), and meltmg occurred over a wide tempera- ture range, indicating lack of purity. Khripin and co-workers noted that the oxygen con- tent in their system was between 1 and 3 wt %. The two specimens of UF, we have examined contained 0.18 and 0.078 wt % oxygen, respectively representing 1.7 and 0.7 mole % UQO, . ‘The present evidence seems to mdlcate that no polymorphlc crystal transition exists in pure UF,4 at or near 841°C, at least within the sensitivity limits of the . present measurements, and that the observed effects are most probably due to oxygen species undergoing | chemical reaction such as : 2UOF, - UO, +UF, , 11, A, Khripin, Yu. V. Gogarinski, and L. A. Lukyanova, Izv. Sibirsk. Otd, Akad. Nauk SSSR, Ser. Khim. Nauk (3) 1, 14 (1965). 2L. A. Khripin, S. A. Poduzova, and G M. Zadneprovsku Russ. J. Inorg. Chem. 13(10), 1439 (1968). 5C. 1. Barton, L. O. Gilpatrick, and J. A. Fredricksen, MSR ' Program Semiann. Progr. Rept. Feb. 28, 1_969, ORNL-4396, p. 163. | ®R. E. Thoma, H. Insley, H. A. Friedman, and G. M. Hebert, J. Nucl. Mater, 21(2), 176 (1968). v _3L. 0. Gilpatrick et al., p. 85 in Thermal Analysis, vol. I, R. E. Schwenker, Jr., and P. D. Gorn, eds., New York, 1969. *A. D. Ryon and L. P. Twichell, Rept. No. H-5 385.2 (TL-7703), Tennessee Eastman Corp., Oak Ridge, Tenn. (1947). 5S. Langer and F. F. Blankenship, J. Inorg Nucl. Chem, 14(1/2), 2631 (1960). for example, followed by solution of the UO, to form a binary system melting at lower temperatures. 13.6 ESTIMATION OF ACTIVITY COEFFICIENTS IN LiF-BeF, -ThF, MELTS C.F. Baes, Jr. Shaffer et al.! have collected a considerable amount of data on the following reductive extraction equilibria over a wide range of composmons in L1F-BeF, -ThF, mixtures: : LiF(F) + % Th°(Bi) = % ThF, (F) + Li® (Bi) CeF3 (F) + % ThO(Bi) = % ThF, (F) + Ce® (Bi) (wherein F and Bi denote, respectively, the molten fluoride and the bismuth phase). Making the reasonable assumption that activity coefficients in the dilute bismuth phase are unity, the equilibrium constants for these reactions may be written DLi)( hF4) | | - 1) (D'Hl fuiw . DCe) f%‘hF | , is approximated as the fraction of the total fluoride ' in'the system which is present in species i. The heat of mlxmg is of the form AHm =RTZB,-jt,n,(tj¢j/bj) s ‘ (8) where the summation is over all the pairwise interac- tions of the four different kinds of species (1) F~, (2) Be F,(b—-2a)- (3) MF,(?=¥)-, (4) NF,(®~2)~, The coefficients t; and t; are the numbers of nonbridging fluorides on each kind of species: : ty =1, t,=2b—4a, t3=bs, t4=b, . The product #;n;(#; ¢;/b;) is intended to be proportional to the number of contacts between nonbridging fluorides of species i and j. The six coefficients §;,, Bi13,B14,B23, P24, Baa, reflect the heat effect of such contacts. - A Proceeding as previously, the activity of each species is obtained by the partial differentiation (AH,, [RT - anl AS,/R) In(@)= ) Substitution of the resulting expressions for each activity into Eqgs. (4) and (6) yields expressions for the - ~ volume fractions of all species in terms of the adjustable x, B, and 7y parameters and the volume fractions of free fluoride, the monomer BeF,?", complexes of MY* and NZ*, These four volume fractions — ¢, PBeFy q)MFc, and ¢MFd — are determined by an iterative computation involving ma- ‘terial balance conditions which are specified by the composition of the mixtures. With the four volume fractions determmed the activities of the four components in a given mixture are entiation of AH,, and AS,, and the lowest fluoride - obtained from the proportionalities UFR=E) (BeF,) & (BeF42)/(F ) (MF,)) & (MF (e~ ))/(Fy~Y (NF,)« (NF (-2 )(F Y- wherein the activities of F~ , BeF4 %", MF (¢~ J’)" and NF4(4-2)- are obtained by appropriate. partlal differ- [Eq. (9)}. The proportion- ality constants will be chosen-so that the activities of LiF and BeF, are unity in 2LiF-BeF, and so that the activity coefficients of MF,, and NF, approach unity as Xy p. and Xy F, approach zero in this solvent. In applying this model to the reductive extraction and solubility results!»2 in LiF-BeF,-ThF,-CeF; melts, it was found that the data could not be accounted for by the assumption of only one thorium and one cerium species in solution. An acceptable fit was obtained, however, by assuming the followmg scheme of species (Table 13.4): ThF, , ThFs~, ThF¢ %~ CCF2+ ’ CCF3 s CCF4_ Here it was necessary to adjust only the o3, 04 ,and B34 parameters of the model; a, and §,, were known already from the fit of the model to the binary LiF-BeF, mixture. The other § parameters were set somewhat arbitrarily at the values given in Table 13.4. Thus the first fit was obtained by using three adjustable parameters of the model, along with the three ad- ~ justable equilibrium constant values. The standard error of fit was 1.32 times the estimated random error of the data. No significant improvement was obtained by assuming 7y; = ¥4 and adjusting this additional param- eter. A slightly improved fit (0 = 1.24) was obtained with the scheme ThF* , ThF,*, ..., ThF,* CeF*',CeF,*,. .., CeFgs3" and an adjustable y (= y3 = v4) parameter. This scheme is essentially the same as the previous one, since the principal species over the composition range involved are the same, but it is preferred as being less arbitrary. The small number of parameters of the model which require adjustment in order to fit the data suggests that the model is a reasonable representation of the nature of molten LiF-BeF,-ThF; mixtures and should provide Table 13.4. Polymer Model for LiF-BeF,-ThF,-CeF; Melts at 600°C _ Scheme I ,'.Scheme II Species assumed? Be,F,(®—20)-plus: ~ ThF,, ThFs, ThF¢?, ThF*, ThF,?,.., ThFs> " CeF,*, CeF3, CeF4” CeF%*, CeF,t, ..., CeFg¥ Parameters qdjustedb - 05 : 2.24 3.75 0y 1.67 442 V3,74 0) . 1.81 B3a -0.421 -0.438 Ky _ ~0.00316 0.00328 . K, - 1.60 1.67 K3 0.00583 0.00589 Parameters not adjusted € -2.32 By2¢ 0.43 B13,B14 0.43 823,624 0 Agreement factord 1.32 ' 1.24 %The species Be Fb(b —24)~ include all such polymeric anions which can be formed by sharing of corners and edges of BeF4 tetrahedra. DThe subscripts 1—4 of a, 8, and v refer, respectively, to the four kinds of species (1) F7, (2) BegFp (029, (3) ThF,®—, and (4) CeF,—3)~. The three K values are defined in Eqs. 1-3. ‘@, and B, were determined by a previous fit of the mode! to binary LiF-BeF, mixtures (ref. 3). 9Defined as (B o) X, is a value of Dy /DY or Do /D3I? determined by Shaffer ef al. (ref. 1), or it ;s a CeF; solubility measured by Barton et al (ref. 2); X is the corresponding calculated quantity; o, is the estimated experimental uncertainty in X o' Np is the " number of observations; Np is the number of parameters adjusted. a good representation of the activity coefficients of the components. Accordingly, activity coefficient curves have been generated (Fig. 13.4) using the parameters of scheme 11, Table 13.4. All curves are plotted vs X ;r, which — except in the case of fo.p, — seems the most - important composition variable deterrmmng activity coefficients in these mixtures. All curves are adjusted so that aL‘F, @peF 4 fThF4, and fo.p, are unity in Li; Be Since the CeF; curves depend on the relatively more accurate solubility measurements,? they should be the most accurate, with an estimated uncertainty not exceeding 10% at compositions farthest from the reference composition. The LiF curves depend partly, and the ThF4 curves depend almost entirely, on the less accurate reductive extraction measurement! and have uncertainties which may reach 30% at the extremes of the composition range. The BeF; curves and LiF curves depend in part on accurate emf data for the binary system3:4 and should generally be of intermediate accuracy. As more data become available from current emf and other studies, it should be possible with the aid -of this model to generate activity coefficient curves for other solutes in these LiF-BeF,-ThF,; melts. This should greatly enhance our ability to predict the effect of melt composition on many important equ1hbr1a in MSBR fluoride melts. 0667+ fiF 152 ORNL-DWG 70-6772 0.333-fge, 0.5 The, - XLiF XLiF Fig. 13.4. Activity Coefficients of Components in LiF-BeF,-ThF, Mixtures Containing a Low Concentration of CeFj at 600°C. Curves are based on parameters in scheme I1, Table 13.4. Standard states are defined such that X1 ;F/LiF» X 18 A § BeF,n JThF,» and fCeF3 are all unity in 2LiF-BeF, (66. 7—33 3 mole %). ‘ * 13.7 ESTIMATION OF ACTIVITY COEFFICIENTS OF ALKALINE EARTHS IN MOLTEN BISMUTH _. AND AS CHLORIDES OR FLUORIDES - J. Braunstein A promising modification of the reductive extraction process for removal of rare-earth fission products from. molten-salt breeder reactor fuels employs an acceptor salt such as molten lithium chloride contacting the bismuth phase.! Thermodynamic data suggest that the fission product alkali metals, Rb and Cs, probably will concentrate in the chloride phase.? Ba and Sr should have a smal]er tendency to transfer to the chloride, but in trying to calculate distribution ratios of Cs, Rb, Ba, and Sr in the process, activity coefficients are needed for the fluoride salts, chloride salts, and the metals in the bismuth phase, Direct experimental measurements of the needed thermodynamic quantities are not yet 153 available. However, recent advances in molten-salt solution theory have been made® useful guide to the estimation of activity coefficients from data in related systems. Estimates are given in Table 13.5 of the activity coefficient; vy, of SrCl, and BaCl, as dilute solutions in molten LiCl, of StF, and BaF, as dilute solutions in LiF, and of Sr and Ba as dilute solutions in molten bismuth. Also listed are the excess chemical potentials, uE = RT In v, at infinite dilution of the above solutes in LiCl, LiF, or Bi. " Free energy data have been summarized for mixtures which provide a of KCl or of NaCl with CaCl,, S1Cl;, or BaCl, and of - LiCl with MgCl,.* Conformal ionic solution theory® predicts a linear relation between the excess chemical Table 135, Activity Coefficients and Excess Chemical Potentials for Dilute Solutions of Sr or Ba as Chlorides, as Fluorides, or in Molten Bismuth uE (kcal per Solven_t Solute mole of solute)® 7b LiCl SiCly +1 1.7 BaCl, +2 3,0 ,LiF S1F, +1 L7 BaF, +2 3.0 Bi Ba —52t0—44 2X107'? to1x10710 Sr —44 1071 “Soluhon at infinite dilution, 650°C, for supercooled liquid when below melting point of solute or solvent. For the fluoride mixtures, it was assumed that activity coefficients of divalent solutes in LiF-BeF, mixtures are similar to those in LiF.” Further, asym- metry in the free energy of mixing in alkali fluoride— alkaline earth fluoride systems was neglected, and =0 E . (Msolute)x colute was estimated as (“solvent/ X:olu te)X amlute'o > again using the correlation provided by conformal ionic solution theory.5-6 _ Free energy data for Sr and Ba in molten bismuth are - sparse and are difficult to obtain because of the activity potential at infinite dilution of divalent salt and an interionic distance parameter. This correlation has been ~ applied to the chlonde mixtures® and permits estima- tion of U, and ”Bam in molten L1Cl W1th a short linear extrapo ation. 1E. 3. Smith, J. J. Lawrance, and C. T. Thompson, MSR Program Semiann. Progr. Rept. Feb 28, 1969, ORNL-4396, p. 285 | . 2W.R. Gnmes private commumcatxon. 3. Braunstein, Statistical Thermodynamics of Mo!ten Salts and Concentrated Aqueous EIectrolytes 0RNL-4433 (1970) (in press). %G. D. Robbins, T. Fetland, and 'r 9’stvold, Acta Chem. ‘ Scand. 22, 3002 (1968). SH T. Davis, J, Chem, Phys. 41, 2761 (1964). 63, Braunstein, K. A. Romberger, and R, Ezell, 21st South- eastern Regional ACS Meeting, Richmond, Va,, Nov, 5-8, 1969, paper 86. of these alkaline earth metals. Previous estimates of vz, range between 107'® and 107%. Estimates of the activity coefficients can be obtained® from a plot of enthalpies of formation (per gram atom of metal) of the compounds BaBi, Ba;Bi,, StBi, Sr3Bi,, and Sr,Bi* ! vs composition by extrapolation of the slope to infinite dilution. The activity coefficient of Mg in Bi, estimated in the same manner, is within a factor of 2 of the measured activity coefficient, - - C.F. Baes, pnvate communicatlon 8y, J Egan, Brookhaven Natlonal Laboratory, private com- munication. - %0. Kubaschewski and H. Villa, Z Elektrochem 53, 32 (1949). 10g A, Shchukarev, M. P. Morozova, Kan Kho-Yu, and G. V. Kokosh, J. Gen. Chem, URSS 26, 1705 (1956). 115, A. Shchukarev, M, P, Morozova, Kan Kho-Yu, and V, T, Sharov, J. Gen, Chem, URSS 21, 321 (1957). 154 13.8 POTENTIOMETRIC STUDIES iN ~ MOLTEN FLUORIDES B. F..Hitch C.F. Baes, Ir. We have continued our studies of the cells " Be(c)ILiF,BeF, IRE. - 1 Pt(c) 3,UF4,'R.E. | | I iFBeF, | - Ni(¢)|NiF,, _}R.E. - I ,BeF, in which the reference electrode is nickel immersed in —molten Lig BeF.; saturated with NiO and BeO: _ RE. = [NiF,(~107*),[BeO(c);|Ni(c) LiF(0.67), . [NiO(c) BeF, (0.33) In previous measurements of these cells, this reference electrode, contained in a fritted silica tube, has proven relatively reliable and especially convenient for po- tentiometric studies of molten fluorides; however, the gradual attack of the silica by the fluoride melts, with the production of gaseous SiF, and dissolved oxide, has caused difficulties, In the current measurements, we have investigated the use of sintered BeO as a container material for the reference electrode. The compartments used, obtained from Brush Beryl- lium Corporation, were 99.8% BeO. They had been closed at the lower end by a disk of the same sintered BeO. In initial tests it was found, fortunately, that the seal between the disk and the end of each tube was not “complete, but in fact contained holes or cracksas large as 2 p. The electrical resistance measured with molten salt both inside and outside these compartments was quite low, typically 75 ohms, Yet salt leakage through these compartments appeared to be very small. These compartments have been exposed to Li, BeF; melts for periods up to three weeks with no detectable weight loss or visible signs of attack. Cell 1. — Since the BeO compartments seemed practically impervious to salt we were concerned with ~ the possibility of BeO acting as a solid electrolyte. If we assume that the Be?* ion or the 0% is carrying the current, the cell reaction for cell I above would be Be’(()+ NiO()>BeO@+No(@) (1) and for either case the cell potential would be in- dependent of melt composition. However, if there is liquid contact and we assume that all the current is _carried across the liquid junction by the Li* ion' we ‘may then write the following cell reaction: Be®(c) + NiO(c) + BeF,(d, R) + 2LiF(d) > Ni°(c) + BeO(c) + BeF,(d) + 2LiF(4, R); (2) R designates solutes in the reference compartment, Since the reference solution closely approximated the composition of the compound Li, BeF,, the standard state we have adopted for LiF(d) and BeF,(d), then 8Lir(r) 2nd 2p.p, () are unity, and the cell potential is given by : S RT dBeF, . | ’ To check the effect, or lack of effect, of melt composition on the cell potential we varied the com- position of the LiF-BeF, imixture on the left in cell 1. Results are shown in Fig, 13,5, where they are plotted against activity data from previous measurements.? As can be seen, the cell potential does vary with melt composition, showing a dependence on @y, /a F which agrees quite well with Eq. (3), represented by the line in Fig. 13.5. This result suggests strongly that the - current is being carried from one haif-cell to the other by Li* ion through the molten salt, rather than by mobile Be?* or O*” jons in the solid BeO., A series of measurements of cell I containing Li, BeF, was made over the temperature range 500 to 700°C. The results differ by about 5 mv from those obtained previously with the use of silica compartments The present results give 'E‘;=l.892-—0.044(Tl1000),_ @ where T is the temperature in degrees Kelvin. This “expression for the potential of cell I, when combined with our previous measurements of the Be?*/Be® couple vs the HF/H,, F~ electrode,? gives the expression in Table 13.6 for the potential of the NiO, Be?*/BeQ, Ni® reference electrode vs the HF/H, , F~ electrode. 1g. A ‘Romberge: and J. Braunstein reported a transference number of #;; > 0.9 for LiF-BeF, mixtures over the range of 0.33 to 0. Sl molc fraction BeF, in Reactor Chem. Div. Ann. " Progr. Rept. Dec. 31, 1968, ORNL-4400, p. 12. 2B F. Hitch and C. F. Baes, Jr., J, Inorg. Chem. 8(2), 2031 - (1969). - _ 155 Table 13.6. Electrode Potentials vs HF/H,, F~ Be?* + 2¢~+Be(c) - U“+e -*Ua" HF+e “>F"+ %hH, . Ni** 2¢™ = NI(¢) oW Be + NiO(c) + 2¢ —>BeO(c) + Ni(c) E° = _2.460 + 0.694(T/1000) E°=-1.106 (610°C) E°=0 E° = 0.568 +0.650(7/1000) E°=0.343 (605°C) The above measurements were made over a period of three weeks, and a single Be® electrode was used. There was no tendency of the cell potential to run down, nor was there any other indication that the Be® electrode was being “poisoned” by the reduction of reducible metals on the electrode surface. This is strong evidence that there was no significant leakage of Nl ion from the reference electrode compartment. ORNL-DWG 70-6773 60 40 k;\ | 20 10 2 t\éE \ . 8 \ <& - @ o . \ 170 174 178 182 186 1.90 : £ {v) : ' Fig. 13.5. Potential Dependence of Cell I on the Ratio of aBeF:/aLlFO Cell 11, the U**/U** Couple. — Studies of cell II are continuing with the reference electrode compartmented in BeQ instead of silica. Present measurements give a value of 1.110 v at 610°C for E° in the Nernst expression . \ E= E°——ln( “F“ 5) XUFg for Li, BeF, as the solvent. This may be compared with our previous measurements of E°, which gave 1.067 v, ‘and the predicted value of 1.139 v. The uncertainties in the present measurements should be no more than £15 mv. The potential of the U* /U couple vs the HF/H,, F~ couple (Table 13.6) was calculated by combining the measured cell potential of cell II and the half-cell potential of the reference electrode. We plan to make further measurements using the BeO compartments in further attempts to resolve the inconsistencies noted above as well as to establish a more accurate standard reference potential for the cell. Cell 111, the Ni**/Ni® Couple. — The cell reaction for cell 111, above, with Li,BeF, in both half-cells and at low concentrations of NiF, may be written ' BeO(c) + NiF,(d) - BeF,(d) + NiO(c) . (6) Then the cell potential should be given by RT S | E=E°+-2InXyip, - 0 Cell potentials with the reference electrode compart- mented in silica were determined by titrating weighed amounts of anhydrous NiF, into an Li, BeF, melt at 605°C. A plot of the results is shown in Fig. 13.6. The ‘measured cell potentials have an estimated uncertainty of £10 mv and differ by some 30 mv from those based on the equilibrium measurements of Blood.? Measure- ments have been attempted using BeO compartments for the reference electrode but have not yielded - consistent results; further investigations are planned. 3C. M. Blood, The Solubility and Stability of Structural Metal Dzfluortdes in Molten Fluoride Mixtures, 0RNL-CF-61-5-4 (September 1961) 156 ORNL-DWG 70- 6774 0.01 0.005 0.002 NiFp X 0.001 0.0005 0.0002 10.0001 ‘ 0. . 0.05 0.10 0.15 S Ely) Fig. 13.6. Measured Cell Potentials of Cell II. The measured cell poiential in volts at 605°C is given by : S E=0343+0087log Xyjp, ) - ® The potential for the Ni**/Ni® couple vs the HF/ H,, F~ electrode (Table 13.6) was calculated by combining the potential of cell III and the reference electrode half-cell potential. 13.9 ELECTRICAL CONDUCTIVITIES AND IONIC MOBILITIES IN THE 'MOLTEN LiF-BeF, SYSTEM G. D. Robbins J. Braunstein In addition to its relevance to the molten-salt reactor program, investigation of electrical transport in molten LiF-BeF, mixtures is of considerable theoretical inter- ‘est, The binary LiF-BeF, system is comprised of pure components exhibiting transport properties which dif- fer by over seven orders of magnitude. Both in its transport properties and its thermodynamic properties it exhibits characteristics of network liquids at the BeF, end and of a typical ionic liquid at the LiF end.'™ 1C. F. Bacs, Ir., J. Solid State Chem. 1, 159 (1970). 2g, Cantor, W. T. Ward, and C. T. Moynihan J. Chem. Phys. 50 2874 (1969). 3J. Braunstein, K. A. Romberger, and R. Ezell, Southeastern Regional ACS Meeting, Richmond, Va., Nov. 5-8, 1969, papcr 86. The results of electrical conductivity measurements in LiF-BeF, mixtures ranging in composition from 34 to 70 mole % beryllium fluoride presented here contain data from two previous experiments?-5 and afford an analysis of the variation of electrical conductivity with ~ both composition and temperature. Ionic mobilities have been reported - previously® at two temperatures and two compositions using transference number® and density” data. The new conductivity data reported here permit an extended evaluation of the composxtlon dependence of internal ionic mobilities. Cell constants of the four cells employed in these studies ranged from 98.55 to 150.4 cm ™! as determined by previously described techniques.4-8 Conductances in the molten fluorides were measured with a specially constructed ac bridge which contains a variable resist- ance and capacitance, in series, in the balancing arm and has the capability of variation of voltage and frequency of the applied sinusoidal potential.® Applied potentials were 25 to 50 mv peak-to-peak across the cell. For - resistances of the order of 100 ohms the bridge inaccuracy is <*0.1% over the frequency range 0.1 to 30 kHz and increases toward %0.2% as resistances approach 10 or 500 ohms. | - As has been previously discussed,®:? the frequency range over which resistance is independent of frequency (or varies in a regular mauner) must be determined for each experimental arrangement and electrolyte solu- tion, This was again demonstrated in the current investigation, in which a frequency range could be found over which the molten KNO; used in cell constant determination exhibited no variation of resist- ance with frequency, within the $0.2% uncertainty limits of the bridge, for each of the four cells (the ranges were 1 to 20, 1 to 10, 0.5 to 3, and 3 to 30 kHz), With molten fluorides in the cells, two of the mixtures showed a frequency-independent region (3 to -30 kHz and 5 to 50 kHz), while in the other two, “G. D. Robbins and J. Braunstein, “Electrical Conductivity Measurements in Molten Fluoride Mixtures, and Some General Considerations on Frequency Dispersion,” in Molten Salts: Characterization and Analysis, G. Mamantov, ed., Marcel Dekker, New York, 1969, ' ; ' ' 5G. D. Robbins and J. Braunstein, MSR Program Semiann. Progr. Rept, Aug. 31, 1969, ORNL-4449, pp. 14142, 6K, A, Romberger and J. Braunstein, MSR Program Semiann, Progr. Rept. Aug. 31, 1969, ORNL-4449, PP 138-41 Inorg Chem. 9, 1273 (1970). 7D. G. Hill, S. Cantor, and W, T. wm,.r Inorg. Nucl. Chem. 29, 241 (1967)." 8G. D. Robbins and J. Braunstein, J, Electrochem, Soc. 116, 1218 (1969). 9G. D. Robbins, J, Electrochem, Soc. 116, 813 (1969). resistances were approximately linear in 1A/F (2% variation, 0.5 to 5 kHz, in one; 0.5% variation, 1 to 10 kHz, in the other). In these melts measured resistances ‘were extrapolated to infinite frequency vs 1 l\/fi_ Specific conductances are listed in Table 13.7, These were obtained from four sets of experiments in dif- ferent cells (see Fig. 13.7). The upper temperature limit of approximately 550°C was imposed by use of molten potassium nitrate as a molten salt bath. In one experiment (cell 2) no bath was present, permitting an ~ upper temperature limit of 647°C to be attained. However, as the data obtained with cell 2 reflect (Fig. 13.7), the absence of a bath around the cell resulted in greater scatter, probably due to temperature variation in the conducting region of the cell. ' 157 As can be seen from Fig. 13.7, the specific con- ductance data vary smoothly as a function of tempera- ture. The curves shown represent computer-fitted least- squares equations that were used to generate the isotherms of specific conductance plotted in Fig. 13.8 as a function of composition. The points shown are calculated values (at 25°C intervals) for each experi- mental composition and extend to the next 25°C be- yond the actual data range. It is believed that the di- vergence of the data at X = 0.540 from a smooth curve _ at higher temperatures is due to experimental scatter. That the data obtained at X = 0.500 with cell 4 form a smooth curve with data from cell 1 at compositions on either side, lends added credence to the results, Table 13.7. Specific Conductance as a Function of Temperature and Composition ' o for Molten LiF-BeF, Mixtures 10 xohms ™ cm™) £C°0) khms? em™) 0 k(ohms™ cm™) XBeF, = 0.340 XpeF, = 0.500 X=0.575 468.0 1.342 369.6 0.236 445.3 0.266 476.3 1.390 379.7 0.265 449.7 0.277 480.5 1.424 389.3 - 0.295 459.6 0.300 499.3 1.540 3992 0.327 470.4 0.328 530.1 1,753 425.7 0.417 491.8 0.382 . 562.0 1,957 - 450.2 - 0.505 515.7 0.445 588.0 2,132 475.5 0.603 542.3 0.522 619.4 2.330 503.8 0.715 . X=0.600 647.2 2.513 ggg: g:ggg | 460.1 0239 Xper, = 0.380 : o 479.9 0.282 465.5 1.100 . X=0520 | 500.1 0.325 486.1 1,226 398.7 0.271 525.4 0.388 502.7 1,322 '418.4 0.329 550.2 0.450 533.2 1504 434,3 0.379 - X=0.650 " Xpep, = 0420 :;g‘; -3;;‘_%2, 480.6 0.169 455.3 - 0.844 498.6 0.602 5211 0,230 . 476.5 0.962 513.6 0.661 551.5 0.276 5014 1.097 528.5 - 0.715 ‘ . X=0.700 5286 1241 ( - X=0.540 501.2 | 0.113 - Xpep, =0.470 3802 0.183 521.9 0.133 425.2 : 0.500 390.0 0.206 549.1 0.160 446.2 0.587 3999 . 0,230 . 4754 0.719 409.9 0.255 499.7 - 0.830 - 419.5 0.280 5259 0.954 4395 0.334 | 460.3 - 0.395° 4789 1 0.451 - 499.2 0.514 524.2 0.595 550.2 0.675 * parameter relation is ORNL-DWG 69—15723 * CELL1 #CELL 2 S CELL 3 e CELL 4 360 380 400 420 440 460 480 500 520 540 560 ti*C) Fig. 13.7. Variation of Spedfic Conductance of LiF-BeF, Mixtures with Temperature, - - Table 13.8. Pamineter Values for the Function ba X —ble 2 K = (a1 + a, X) exp m =1 2 T in degrees Kelvin Parameter Value ay 7.1845 as 7.7458 by - 252.57 2 3.2965 c1 467.24. ¢s ' - ~333.96 Rather than reporting the 49 parameter values re- quired for interpolation employing the individual equa- tions shown in Figs. 13.7 and 13.8, it is more convenient to represent the specific conductance as a function of temperature and composition by a single expression with coefficients derived by least-squares fitting of all the data simultaneously, One such six- T—(er +e,X) T is in degrees Kelvin. Coefficients for Eq. (1) are listed in Table 13.8. The standard error of fit is 0.007. o Bax kK =(a; +a,X) exp [—-—-_—'P-'-’-e——] ' '(1) Based on the assumption of randomness and inde- pendence of errors with an estimated uncertainty in of 0.4% from cell constant determination, 0.2% from resistance measurement, 0.5% from resistance vs 1/\/f extrapolation, and 0.3% from temperature measure-- relation!® 158 ~ ORNL-DWG 69-13724R . ' -CELU '8 —\ 8 . . CELL 2 | °\\ o CELL 3 8/ o 0.4 7, = 0 NN \\ ~ .\, 02 s '\:-\'..,\'§ %% 06.3 | 6.4 0.5 06 07 xBer Fig. 13.8. Specific Conductance of LiF-BeF, Mixtures vs Composition, - Lo ment, the calculated probable error is 0.8%. We believe the uncertainty level of these results to be of the order of £2%. _ Mobilities of lithium ion relative to fluoride (or relative to beryllium ion, since beryllium is immobile relative to fluoride®) may be calculated from the where C;, is the concentration of lithium ion in equivalents per cubic centimeter, F is the faraday, and ty; is the transference number of lithium ion relative to fluoride, which is unity.® The mobilities of lithium relative to fluoride have been calculated at 25° intervals and are shown in Fig. 13.9. These, extrapolated to 650°C, are compared in Fig. 13.10 with the mobilities of Li. relative to halide in the binary LiCl-KCl,!? . LiBr-KBr,'? and LiCI-PbCl,"'? systems, and K relative 104, Klemm, “Transport Properties of Molten Salt” in Molten Salt Chemistry, M. Blander, ed., John Wiley and Sons, New York, 1964. ' ' 11¢, T, Moynihan and R. W, Laity, J, Phys. Chem. 68, 3312 (1964). ' : . 1200, P, Mehta, F. Lantelme, and M. Chemla, Electrochim, Acta 14, 505 (1969). ' 13w. K. Behl and 1. J. Egan, J. Phys. Chem, 71, 1764 (1967). 159 to halide in the KC1-CaCl, ! 3 and KCI-MgClg 13 systems In the LiF-BeF, system, values of by ; are smaller than those in the other binary system of halides, as might be expected'with increased network character of the liquid. ORNL-DWG 69— 13721R 5.0 \ " 45 . S ‘\\\ N 40 o — .\‘ \ 860-0 PN SOk 30 \ \\ 300 \ . L S . N &‘g\: " = 05 - 0 - 03 0.4 05 - 086 o7 xBer Fig. 13.9. Internal Mobility (Lithium Relative to Fluoride) as -a Function of Composition in the System LiF-BeF,. . ~ ORNL-DWG 70-32MR {(x10°%) 20 -KBr 6350°C LICI-PbCl, 650° -MgCl, 825° b (em? v=! sec ) o KCI~CaCl, §25° pas LiCI-KCI 640° ,2 LiF-BeF, 650° 0 - 0O ot 02 03 04 05 0607 08 09 1.0 X Fig. 13.10. Internal Mobilities (Lithium or Potassium Rela- tive to the Anion) vs Mole Fraction of the Second Component, 13.10 ELECTRICAL CONDUCTIVITIES OF PROPOSED MSER FUEL COMPOSITIONS "IN THE LiF-BeF,-ThF, SYSTEM G.D. Robbins - A. S. Gallanter" The absence of reliable theories for predicting electri- cal transport in molten salts renders a priori estimation of electi'icaI conductance in ternary mixtures extremely uncertain® Hence, specific conductances and their temperature dependences from near the liquidus to 650°C have been measured for four proposed MSBR fuel compositions. Two additional LiF-BeF,-ThF, compositions having equal molar ratios to individual proposed fuel compositions were also investigated to determine the effect of additions of each component. Starting materials®> were of compositions (in mole percent LiF-BeF,-ThF,) 68-20-12, 63-25-12, and 72-21-7. These had been previously treated with an H,/HF mixture followed by reduction with pure hydrogen. A liquid aliquot of each starting material was delivered into a nickel vessel under argon, frozen, and opened in a dry box; the entire sample was ground to 30 mesh particle size and mixed on a rotating tumbler for at least 3 hr to ensure homogeneity. Two additional compositions (60.5-28-11.5 and 71-18.13-10.87) were obtained by addition of either clear pieces of lithium fluoride, which had been fused under a hydrogen-argon mixture and slowly crystallized, or sublimed beryllium fluoride glass.* A sixth composition (73-16-11) was prepared from the component starting materials, the ThF, having been treated with H, /HF followed by H,. The silica conductance cells and measuring bridge are described elsewhere.® Cell constants ranged in value from 93.09 to 156.12 cm™' and were independent of frequency (£0.1%) from 10 to 50 kHz as determined in molten potassium nitrate.’ With molten fluorides in the - cells, some showed no. frequency dependence of the resistance, while the resistances of others varied as 1 I\/_ and were extrapolated to infinite frequency. Assigned uncertainties in the determination of specific con- ductance are 0.4% from cell constant determination, 10RAU Summer Student Trainee, 1969, Brooklyn Col]cgé, Brooklyn, N.Y. 23ee, for example, A, Klemm, ‘Transport Properties of Molten Salts,” in Molten Salt Chemistry, M. Blander, ed., John Wiley and Sons, New York, 1964. 3These were obtained from J, H. Shaffer, ORNL. *Prepared by B. F. Hitch, ORNL. 5G. D. Robbins and J. Braunsteln, J. Electrochem. Soc. 116, 1218 (1969). 160 Table 13.9, Specific Conductance Data for LiF-BeF,-ThF, Mixtures to ~ x(ohms™ cm™) t(°C) x (ohms ™! em™) 73-16-11 Mole % LiF-BeF,-ThF, 63-25-12 Mole % LiF-BeF,-ThF, 528.8 1.627 5551 . 1.390 538.9 1.693 565.7 1.458 542.3 - 1.721 584.3 1.576 552.5 " 1.796 599.3 1665 599.8 2.134 622.2 1.799 629.7 2,337 641.0 1.919 651.6 2495 645.1 1.943 , 72-21-7 Mole % LiF-BeF,-ThF, 71.000-18.125-10.875 Mole % LiF-BeF,-ThF, 5406 ., 1.865 - 553.6 1.682 557.8 '1.990 576.6 1.851 570.4 . 2.079 598.0 2.003 592.8 . 2,243 620.4 2.184 611,0 2377 . o T 60.5-28.0-11.5 Mole % LiF-BeF,-ThF 625.9 2.492 e % LiF-BeF,-ThEa 641.1 2.609 5715 1.375 599.0 - 1.558 68"20‘12_M013 % LiF'BeFQ'ThF4 614.5 1.647 544.0 , 1.454 631.8 1.757 551.5 1.503 | 566.9 1.615 583.6 1.728 '603.7 1.861 632.5 2.055 651.7 2,178 Table 13,10. Analytical Representation of Specific Conductance of LiF-BeF,-ThF4 Mixtures Mole % LiF-BeF,-ThF, Equation (z in °C) 0 73-16-11 k=1422+7,077X 1073 (+-500) 0.003 72-21-7 k=1.562+ 7.386 X 10 (r ~ 500) 0.004 68-20-12 k =1.160+ 6.740 X 10> (¢ — 500) 0,004 63-25-12 x =1,056+ 6,111 X 102 (¢ — 500) 0,003 71.000-18.125-10.875 k= 1,278 + 7.481 X 102 (£ — 500) 0.005 £ 60.5-28.0-11.5 x=0.926+6.317 X 1073 (¢ — 500) 0.007 Table 13.11. Effect of Variation of One Component in LiF-BeF,-ThF, ‘Mixtures at Constant Molar Ratio of the Other Two Curves of Constant Constant katio Variation of the Effect of Variation at — Component Ratio? ‘ , . Third Component 550°C 650°C A->D ‘ LiF/BeF; = 3.42 (10.02) ThF,4 up S mole % x down 22.5% x down 18.7% D=>C. BeF,/ThF4 = 1.67 LiF up 3 mole % x up 10.4% x up 10.5% E->F - LiF/ThF4 = 5.25 BeF3 up 3 mole % k down 8.8% x down 5.0% - - 4See Fig. 13.11. 161 ORNL-DWG 70-3212 2.8 : . mole % LiF-Ber-Thfi; 2.6 72-21-7 73-16-11 2.4 ] ‘ ‘/ /0’ /./ 2.2 : /‘/ _~n-18425-10.875 ¥ L~ -~ /- 68-20-127 | 5 / ~ |- T 20 . ~ . . ‘ - g / / / | -» 63-25-12 ¥ » » : ‘ ; A ] L~ A _/ - 1.~ 1~ - - 60.5-28-11.5 1 ¢ / - L~ 16 |-B ° /. / | AT A P /' o s 1.4 v — | 1.2 520 540 560 580 600 620 640 . 660 680 ¢ {°C} Fig. 13.11. Specific Conducta_nce of LiF-BeF,-ThF4 Mixtures vs Temperature (° C). 0.2% from resistance measurement, 0.3% from tempera- ture measurement, 0.5% from 1/+/f extrapolation, and 1.5% from temperature gradients resulting from the absence of a molten nitrate bath due to the high temperatures. Assuming randomness and independence of errors, the calculated probable error is 1.7%. Con- sidering additivity of the assigned uncertainties, the estimated overall uncertainty is of the order of +3%. Specific conductance data for six LiF-BeF,-ThF, mixtures are given in Table 13.9. The temperature variation of specific conductance is well represented by linear equations, as can be seen from Fig. 13.11. The computer-fitted least-squares equations represented in Fig. 13.11 and the standard error of fit, o, for each composition are listed in Table 13.10, Three pairs of curves have a constant ratio of two of the components. The effect of variation of the molar concentration of the third component is shown in Table 13.11. At the compositions noted, addition of LiF at constant molar ratio of BeF, to ThF, results in an increase in the specific conductance. Addition of either BeF, or ThF, at constant molar ratio of the remaining two components lowers the specific conductance. These - measurements supersede estimates of conductance for four proposed MSBR fuel compositions.® ®G. D. Robbins, pp. 1417 in Physical Properties of Molten-Salt Reactor Fuel, Coolant, and Flush Salts, S. Cantor, ed., ORNL-TM-2316 (August 1968). 13.11 DETERMINATION OF LIQUIDUS TEMPERATURES IN THE LiF-BeF, SYSTEM FROM EMF MEASUREMENTS OF TRANSFERENCE CELLS K.A.Romberger J. Braunstein- Emf measurements of cells with transference provide a method by which, in favorable cases, precise deter- minations can be made of liquidus or solid saturation temperatures in multicomponent systems. The tech- nique, to be useful, requires that the emf be stable, - reproducible, and responsive to the change of activity in the melt accompanying a phase change occurring in one of the half-cells. The phase change may be engendered ~ either by changmg the temperature at constant total composition or by changing the composition at con- stant temperature. In either case, plots of emf vs composition or emf vs temperature show discontinuities of slope at the liquidus composition or temperature. Our studies of the emf’s of the cell with transference? |uF ||ug | | _ Be Bng BCFQ_ Be (l) 1 || o | 1%, A. Romberger and J. Braunstem Inorg Chem, 9, 1273 (1970); MSR Program Semiann, Progr. Rept. Aug, 31, 1 969, 0RNL-4449 p. 138; MSR Progmm Semiann, Progr. Rept. Feb, 28, 1969, ORNL-4396, p. 180. : 162 ORNE-DWG 70-4107 350 T T — ' T ] : SATURATED, LiF 607.5°C HOMOGENEOUS _ A - — \|\ ISOTHEBM 2o L7 . 340 — \ N _ N\ E | T N\G s : \ 5 . 16— NN, Vn 330 \ ‘O@ g ' \ ) '97 x = 7 5 o 7 ® A w BeF, ; o : ‘ N < 'o 0.2652 £0.0003 % W t= . ol g \A\D 1o 12 [ 530.2 +0.2°¢— o] R 320 G s w o o \ Wl | 1NN |9 N 310 \\ ‘\ 300 L19 - A T(s) 0.25 0.26 0.27 0.28 0.29 X BeF, 0.30 540 530 520 510 TEMPERATURE (°C) Fig. 13.12, Data Plots lllusttatmg the Two Graphical Methods for the Determination of Liquidus Temperature, (a) Plot of EMF vs salt composition. Temperature c¢onstant at 607.5 + 0.1 °C. (b) Plot of EMF vs temperature. Total bulk salt composition 0.2997 mole fraction BeF,, indicated that the system I.iF;Ber ‘was particularly amenable to precise phase boundary determinations with the cell (1). Measurements were made in which the ‘reference solution (I) remained homogeneous while the melt in the bulk compartment (II) was transformed from a homogeneous solution to one containing the equilibrium saturating solid, either by additions of one component isothermally, or by lowering the temper- ature of the entire cell. The measurements have yielded liquidus temperatures in the LiF-BeF, system for compositions between 0.12 and 0.58 mole fraction BeF,. Experimental detalls of the emf measurements are given in ref. 1. Liquidus temperatures and composmons were derived from plots of emf vs the composition of compartment II at constant temperature, as in Fig. 13.124, and from plots of emf vs temperature at a fixed overall compo- sition of compartment II, as in Fig. 13.12b. These plots are representative of the sensitivity of the method, namely, about 0.0002 to 0.0003 mole fractlon BeF, in composition and +0.2°C in temperature. = . ‘The' liquidus-temperature—composition results are shown in Table 13.12. These data were obtained from two independent experiments, which overlap along the LiF liquidus for approximately a 50°C temperature interval. The internal consistency of each set of points is within $0.0002 mole fraction BeF;, but there appears to be an offset of approximately 0.0015 mole fraction BeF, between the two sets. This probably “results in part from an uncertainty of +0.001 mole fraction BeF, in the correction of the composition, in one of the experiments, for material withdrawn from the bulk compartment into the small reference com- . partment. - S An interesting aspect of the LiF-BeF, system has been whether the solid compound Li,BeF,; melts congruently, as do the other alkali flucroberyllates, or incongruently. Three different solid phases have been shown to exist in equilibrium with the liquid phase of LiF-BeF, : LiF, Li, BeF,, and BeF,. If, when the total bulk composition is 0.3333 mole fraction BeF,, the solid in equilibrium with the liquid is Li, BeF,, then the compound . melts congruently, and, in addition, an ' LiF-Li, BeF, eutectic exists (although it might coincide - with the melting point). If the equilibrium solid at 0.3333 mole fraction BeF, is LiF, then melting of Li, BeF, is incongruent, and there is a peritectic. The latest résumé of this system by Thoma et al? inter- preted the available data to suggest that Li,BeF, melted incongruently. We believe the precision and accuracy of the emf method, however, to be capable of resolving the question virtually unequivocally, and the new results indicate that Li, BeF, melts congruently. “The evidence for this conclusion is presented below. When the emf is measured of a cell in which a. homogeneous solution is cooled. until precipitation occurs, the observed emf indicates the composition of the solution phase regardless of whether solid is present or not. The direction of the emf change after the onset of precipitation is determined by whether the com- position of BeF, in the melt increases or decreases. 2R, E. Thoma, H. Insley, H. A, Friedman, and G, M. Hebert, J. Nucl, Mater, 21(2), 176 (1968). 163 Table 13.12, Liquidus Temperatures in the LiF-BeF, System Between 0.12 and 0.59 Mole Fraction BeF, Mole Fraction Liquidus Mole Fraction Liquidus BeF, ‘Temperature BeF, - Temperature (Xger,) C) (Xper,) 0 Equilibrium Saturating Phase: BeF, 0.2140 688.2 0.5869 o 438.1 0.2220 678.0 \ - | 0.2301 663.7 Equilibrium Saturating Phase: Li; BeF, 0.2331 662.5 . 0.3415 458.8 0.2358 658.3 0.3495 458.0 0.2407 647.6 0.3580 . 456.4 0.2515 633.0 0.3625 : 455.5 . 0.2525 628.8 0.3663 | 454.2 0.2525 631.6 0.3746 451.0 . 0.2619 611.9 0.3811 4484 0.2652 | 607.5 0.4267 424.9 0.2747 384.8 0.4476 410.7 0.2826 | -568.5 0.4755 3936 0.2875 3576 S | 0.3012 530.5 Equilibrium Saturating Phase: LiF 0.3056 516.7 0.1188 - 784.7 0.3061 518.2 0.1215 7829 0.3119 502.8 0.1371 7706 0.3141 | 495.9 0.1404 7679 0.3179 486.3 0.1574 7532 0.3189 , 4837 0.1625 7419 0.3208 480.5 0.1853 7239 0.5216 477.2 0.1894 718.0 0.3242 4101 0.3285 458.2 0.2066 698.0 Comparing a series of mixtures whose initial com- positions cross the composition of the compound, as indicated schematically in Fig. 13.13, the cell emf - increases during precipitation for initial compositions on one side of the compound composition, and decreases on the other side. For an incongruently melting compound the emf’s will always change in the same direction. Thus a compound, peritectic, or eutec- tic point can be identified. It should be noted that this - argument depends only on the response of the Be electrodes to BeF, concentrations and therefore yields - the topology of the liquidus line independent of possible errors in values of temperature or composition. For compositions within +0.005 mole fraction of the Li; BeF4 composition, the liquidus curve was found to be very flat; that is, the melting temperature was quite insensitive to composition change. Hence a large frac- tion of the total salt present had to precipitate or dissolve to lower or raise the temperature an experi- mentally meaningful amount, say 0.1°C. Because of this heat load, it was not feasible to increment the equilib- rium temperature when a large fraction of solid was present, : ' Hence, rather than trying to maintain temperature equilibrium, we cooled the solution slowly at a nearly constant rate, The emf was recorded as a function of time both before and after precipitation occurred. ‘Figure 13.14 shows a plot of emf vs time where all the initial emf’s have been. shifted to a common zero value - in order to emphasize the changes. The fingered pattern shows that the solution composition moved first toward higher and then toward lower BeF, compositions as the - initial overall BeF, concentration was lowered. LiF additions 96 and 97 showed very little emf change ~ while precipitation was occurting. The calculated com- positions of these solutions were 0.3337 and 0.3329 mole fraction BeF,, compared with the theoretical - value of 0.33333 for Li; BeF,4. X-ray data show that 164 ORNL-DWG 70- 4406 : | i i | ¢ ] i ] ! | | i ! ¥ 1 ! | i i : . - PERITECTIC ~ EUTECTIC-STABLE COMPOUND START AT T AND COOL : START AT Tp AND COOL Ky Xp X3 Rgq Xg ® ®B. B x X3 X 1w kg VX O 2O oy O 152 %3 % 5 % YT F FF T am N B WD S =X X X x = = To To & ol N o' & To n ‘-4 Ty T2 T3 Ta TEMPERATURE —e= ' Y2 Xz Xq Xy (o) Fig. 13.13. Idealized Temperature-Composxtxon and Temperatuxe-EMF Dlagrams as Observed with (a) Congruently and (b) Incongruently Melting Compounds. TEMPERATURE —= -y o anf s [soLiD 1 SOL_',D ' Ts + : . . souorz | Y'9VP ASSUME x5 AS Fsolid 11 . . Xy K2X3 Kg X5 -— {E| EMF (5) ORNL-DWG 69-13936R COOE NO. xg,p,{BOOK) 2 92 0.3356 93 0.3354 4 94 0.3346 95 0.3344 6 96 0.3337 97 0.3329 98 0.3324 RN 99 0.3320 100 0.3314 TIME ELAPSED COOLING (min) /// o NN o AENNRS o ost00 TN N\ [ o 96\ 98 {100 \\ 104 | W / 93 %95%_97 99\ \ f03_ | / VT VNN . | 101 \Mo2 -2 - 0 v 2 ~—TOHARS HIGHER AE(my) TOWARDS HIGHER—- Fig. 13.14. EMF-Time Plots for Composition near to 0.3333 Mole Fraction BeF,. crystallme Li,BeF, is a st01ch10metnc compound to at least 0.33333 mole fraction BeF, .2 The phase diagram for the region near the melting ' point of Li, BeF, is shown in Fig. 13.15. The series of dots near the Li,BeF, composition are the results of 33, H, Burns and E.K. Gordin, Acta Cryst. 20, 135 (1966). thermal halts determined at the same time as the emf-time curves were being recorded. These data verify the small temperature-composition dependence for this composition region. The LiF-Li;BeF, eutectic is be- lieved to be at 0.328, £ 0.0004 mole fraction BeF, ata temperature of 458.9 + 0.2°C. The temperature at the Li, BeF, maximum is 459.1 +0.2°C. The temperature ORNL-DWG TO-4105 '4.,0 \ \ " '« THERMAL HALTS . l\ ‘© EXPERIMENT NO. 6, 1st CYCLE ® EXPERIMENT NO. 6, 2 nd CYCLE — — FROM THOMA ¢7o/., /. MUCL. MATER. 27, 166 (1968) g \ W 465 \‘ | uiQuio __} 2 . g S ‘ 1 -EUTECTIC £ LiF || LiF: LigBeFs [ l 458.910.2°C i LIQUID \| 0.32800.0004 460 } |- mole fraction BeF, ‘ X \ - Soporerrren—X . , T b T _ar | ~~ T~ ~0,2°C ‘ LI23EF4 L|283F4 T a \ ‘ -+ "'-.\ S~y ass LiouD ~ 0.32 . 0.33 0.34 0.35 0.36 XBeF, {mole fraction) Fig. 13.15. The LiF-BeF, Temperature-Composition Diagram . for Compositions Between 0.32 and 0.365 Mole Fraction BeF,. difference between the maximum and the eutectic is 0.15 + 0.05°C. These new data indicate that the LiF phase field is about 0.01 mole fraction smaller in extent at 460°C than previously believed, the discrepancy decreasing , with increasing temperature up to 700°C (~0. 21 mole fraction BeF,). 13.12 COORDINATION EFFECTS ON U(IV) SPECTRA IN FLUORIDE MELTS L. M Toth The spectrum of U(IV) in molten fluorides has been previously reported! and is currently being used for spectroscopic determinations of U(IV) concentrations. This present investigation is aimed at correlating changes in the stability of U(III)-U(IV) mixtures to changes in the number of fluoride ions associated with the uranium cations. A variation in the number and positioning of fluoride ions around a cation is expected to alter its visible and ultraviolet spectrum; but until - now, little attention has been paid to these effects in molten fluorides. It is of interest to view this aspect of coordination chemistry in detail because the nature of the fluoride ion environment should affect not only the spectrum but also the reactivity of the coordinated metal ion. These coordination effects are expected to play a more significant role in polyvalent systems than in monovalent ones. Dilute fluoride solutions of U(II)-U(IV) provide the most pertinent systems for initial study to properly assess both the measurability of coordination effects and their influence on stabiliza- - 165 tion of valence states. Information gained from this system will serve as a guide when less familiar poly- valent systems are undertaken in detail (e.g., Nb and Mo). The spectra- of U(dV) in Figs. 13.16 and 13.17 demonstrate the influence of solvent composition and temperature changes respectively. The concentration of UQV) in all solutions is approximately 1 mole %. Curves 4, B, and C of Fig. 13.16 show U(IV) at 550°C in Flinak (LiF-NaFKF, 46.5-11.5-42.0 mole %), LiF- BeF, (66-34 mole %), and LiF-BeF, (48-52 mole %) respectively. Curves 4 through D of Fig. 13.17 show U(V) in LiF-BeF, (66-34 mole %) at 460, 500, 550, and 690°C respectively, One species of interest is found at high free-fluoride concentrations and low tempera- tures (curve A, both figures). Preliminary results from crystal spectra of U(IV) in known environments lead to the assignment of this species as 8-coordinated U(IV). It is characterized by intense, well-resolved peaks at 9200, _ 15,400, and 20,500 cm™', Lesser peaks belonging to this species are also found at 6200 and 7200 cm™ and are perhaps more useful in identifying 8-coordinated U(IV) because other species of U(IV) do not absorb here appreciably. - As either the free fluoride ion concentration is decreased or the temperature is increased, there is a shift to a second species (trend of curves A-C, Fig. 13.16, and curves A—D, Fig. 13.17) which is character- ized by less-intense, broader bands at 9200 and 15,800 em™! and no bands at 6200 and 7200. It is believed that this species is of lower coordination number, presumably in equilibrium with the 8-coordinated species: UFg* < UF 4—x+(s —%)F". Although the UF, 4~ species has not yet been identi- - fied, it is presumed to be 7-coordinated U(IV). By comparison with 9-coordinated U(IV) crystal spectra, there is no evidence to suggest that 9-coordinated U(IV) exists in any molten fluoride solutions thus far studied. The most striking effect of decreasing coordination number in U(IV) fluoride spectra is a decrease in the intensity of the 9200 atm™ and that for HF (ORNL-3872, p. 123) about 1.3 X 1074, | _ [HF] salt . 400, 5 [H2 ] salt ' ( ) At low pressures of H, and low concentrations of UF; a large fraction of the hydrogen should therefore be present as HF in the salt in a molten-salt reactor and in the gas phase in contact with it, such as the gas in the MSRE pump bowl. The HF may react with unplated metals to produce hydrogen atoms, which then escape via permeation, but the rates are not known. The use of chemical gettering by means of oxide in salt quite probably would lead to a corrosion problem.? The presence of moderator graphite and almost certainly the presence of oil should yield, additionally, methane. However, at 1000°K ' P Kp = =0.1. (6) Hy Consequently, unless the hydrogen pressures approach an atmosphere, only the presence of the radiation field should permit appreciable reaction of the graphite moderator to yield methane. | B. J. Wood and Henry Wise* studied the reaction of graphite with hydrogen atoms over the temperature range 450 to 1250°K and found that although hydro- gen recombination dominated reaction with the graph- ite, a carbon reaction rate is given by: R=10° (_.__‘.’Tin_) [H,] [H]O-S moles!/2min /. even at the relatively high temperature of 1000°K. The rate of carbon reaction had a maximum of about 5 38, Cantor, sect, 12.4, this report, 4B. J. Wood and Henry Wise, J, Phys, Chem. 73, 1348 (1969). times this value at 835°K and a hydrogen pressure of 1 torr, but due probably to the thermodynamic in- stability of methane, at elevated temperatures the rate was lower. The reaction to produce hydrocarbons (methane primarily) was attributed to a reaction of hydrogen atoms with the edges of basal planes in the graphite sample. Although this work indicates that the hydrogen-graphite interaction could occur in a molten- salt reactor, predicting what will occur at reactor conditions is not yet possible. Similar considerations gases if present in the reactor should be expected. The solubility of H, determines the stripping capa- bility of a sparge process. No measurements of H, solubility were found for salt. Applymg the model for rare-gas solubility,’ 1 ~18.087%y . o K, -R—T“CXP(T>, ' B (7) for this admittedly different case leads to estimates ranging from 8 to 20 times the helium solubility, or 8 X 107 to several times 107¢. The uncertainties in r (0.9 A to 1.1 A) and in @ (0.3 to 0.5) are significantly large. Expressions for permeation found in the literature 168 where 1/2 < n < 1, with the value of » increasing at low pressures (less than 107 atm for Inconel). This type of dependence with n = 1/2 has been shown to apply to Mo and W for hydrogen at pressures as low as 107° torr,® although at lower pressures the value of # should increase. A relationship such as is shown in Eq. (9) is expected for the metals of interest to molten-salt reactor designers, but the exact pressure dependence ~ has not been determined at low pressures. “indicate that equilibration with oils or hydrocarbon had profoundly dlfferent pressure dependence Dushman’s expression,® Cp —x1/2 . 0=Kp 1+Cp’ where Q = permeation rate, p = the pressure, and C and K are constants, yields a 3/2 power dependence at relatively low pressures. This would have made it difficult to expect permeation to be significant at the expected - pressure in molten-salt reactors, since Q approaches zero faster than p in this expression. Unfortunately, this relation was obtained from the literature without the qualifications which the original authors’ specified. A detailed analysis of the perme- ation pressure dependence produces the results that Q=K. ) O M. Blander et l., J. Phys. Chem. 63, 1164 (1959). 6, Dushman, Scientific Foundanons of Vacuum Techmque Wlley, New York, 1963, 7C. J. Smithels and C. E. Ransley,Proc Roy. Soc. A150,172 (1935), 8R Frauenfelder, Permeation and Diffusion of Hydrogen in Tungsten and Molybdenum, Westinghouse Research Laboratory report WERL 2823-27 (June 23, 1967). @® Inadequate information also exists on diffusivities as a function of composition in various alloy systems as well as, of course, in the salt systems of interest. 13 14 CRYSTAL STRUCT URE OF THE COMPLEX - FLUORIDE (Na,Ll)-] Th5 F3 1 George Brunton D. Richard Sears Among the complex compounds formed from the alkali fluorides and monoclinic tetrafluorides of the ‘heavy metals, an unusual stoichiometry, 7AF-6MF, {(where A is an alkali metal and M is a heavy metal, such as Zr, Hf, Th, or U), occurs. Heretofore, all members of this class of compounds were found to produce crystals from the melt which were rhombohedral, space group R3, Investigations of phase equilibria in the system LiF-NaF-ThF, by Thoma et al.! disclosed the occur- rence of a ternary solid solution, with a host structure unlike those commonly .observed in such fluoride systems. Single crystals of this phase were isolated from frozen fluoride mixtures and subjected to x-ray and neutron diffraction analysis. The results of this examination? showed the existence of the compound (Na,Li), ThgF3, , which crystallizes in the space group P3cl with g, =9.9056 + 0.0003 A and co = 13.2820 + 0.0005 A, 25°C (Fig. 13.18). The calculated density is 6.045 g/cm® and Z = 2, There are 18 atoms in the asymmetric unit. Two independent Th** ions and F(1) through F(10) are on the general position 6(d); Th(1) is coordinated by nine F~ and Th(2) is coordinated by ten F~. Ion F(11) is on position 2(g). The three sodium ions Na(1), Na(2), and Na(3) occupy positions 2(c), 2(&), and 2(a) respec- tively. The ion Na(3) is surrounded by 12 F~, and Na(2) and Na(1) are surrounded by six F~ at distances less than 3.0 A. Ions Na(1l) and Na(2) have five and 1R, E. Thoma, Reactor Chem, Div. Ann. Progr Rept. Jan, 31, | 1965, ORNL-3789, p. 23. " 2G, D. Brunton and D. R. Sears, Acta Cryst B25, 2519 .(1969). 169 ’ surrounded by seven F ™ nearest neighbors less than 3.0- A distant. Positional parameters and temperature fac- tors for (Na,Li); ThgFa, are listed in Table 13.13. three more F~ nearest neighbors, respectively, between 3.0 and 3.5 A. Ion Li(1) is octahedrally coordinated at position 2(a); Li(2) is at general position 6(d) and is - Table 13.13. Positional Parameters and Temperature Factors for (Na,Li); TheF3 POSITIONAL PARAMETERS AND TEMPERATURE PACTORS FOR (Na,Li),TH,F,, ATOM X Y Z . ) | B;2 B, By s By, B:s no? 0.0782(4)? 0.39%4(4) 0.0% 6.0012(3} 0.0012(3) 0.0006(2) 0.0006{3) 0,0002(2) -0.0001(2) may M 0.0786(2) 0,3954(2) [ ] 0.0035(2) 0.0029(2) ©0.0005(1) 0.0016(2) 0.0000(1) 9.0000(1) X0 0,.0847(2) 0.4094(2) 0.0 0.0044(2) 0.0039(3) 0.0007(1) 0.0024(2) 0.0000(2) -0,0003(1) NX 0.0850(4) 0.4088(4) 0.0 0.0016(4) 0.0013(4) 0.0003(2) 0,0010{3) -0.0004(2) -0.0002(2) no 0.9148{4) 0.5911(4) 0.2059(2) 0.0008(3) 0.0017(3) 0.0003(2) £8.0007(3) 0.0001{2) 0.0003(2) ey XN 0.9153(2) 0.5906(2) 0.2060(2) 0.0033(3) 0,0035(2) 0.0006(1) 0.0020(1) 0.0002(1) 0.0000(2) %0 0,.9214(2) 0.6045(2) 0.2061(2) 0.0036(2) 0,0035(2) ©¢.00035(9) 0.0017(2) 0.0000(1) 0.0002(1) X 0,9220(4) 0.6043(4) 0.2059(2) 0.0011(4) 0,0013(4) 0.0006(2) 0.0005(3) 0.0002(2) «~0,0001(2) NO 0.1372(;)' 0'429}(5) -o.lssi(s) o.ooz?(}) 0.0318(1) 0,0007(3) 0.0003(6) -0.0005(4) -0.0001¢4) 1) XN 0.144(4 0.443(4 ~0.192(3) 0.005(2 X0 9.139(4) 0.441(4) =0,181(3) 0,002(1) d X 0.1335(9) ° 0,4435(9) =0.1803(6) 0.0036(8) 0.0041(8) 0.0008(3) 0.0010(7) 0.0002(4) 0.0006(4) NO 0.8658(8) 0,5552(7) 0.3853(6) 0,0036(7) 0.0030(7) 0.0009(3) . 0.0007(6) -0,0008{4) -0.0002(4) r(z) XN 0.869(3) 0.564(3) 0.376(3) 0,003(2) d X0 0.864(4) 0.563(4) 0.388(3) 0.002{1) d X 0.8615(8) 0.5692(8) 0.3916(6) 0.0025(7) 0.0037(8) 0.0057(3) 0.0001(6) -0.0002(4) 0.0005(4) : nO 0.1191(8) 0.3061(8) . =0.3759(6) 0.0031(7) 0.0039(8) 0.0015(4) 0.0020(6) 0.0011(5) 0.0002(4) 3 xn 0.114(4) 0,306(3) -0,374(3) 0.005(2) a X0 0.129(5) 0.315(4) -0.374(3) 0.003(2) 4 X ©.1197(9) 0.3088(9) =0.3770(6) 0.0023(7) 0.0041(9) 0.0016(4) 0.0018{7) - 0.0004(5) =0.0009(5) ¥o 0.8781 0.6893(8) 0.5827(8) 0.0037(8) 0.0028(7) 0.0012(4) 0.0015(6) 0.0006{4) =0.0004(4) r4) XN 0.874(4) 0.682(4) 0.586(3) 0.005(2) d X0 0.886(5) 0.690(4) 0.587(3) 0.003(2) d X 0.8784(9) 0.6912(9) 0.5811(6) 0,0028(8) 0.0027(8) €.0015(4) 0.0013(6) 0.0012(5) 0.0003(5) no 0.3431(6) 0.5103(7) =0.0444(5) 0.0015(6) 0.0047(7) 0.0022(3) 0.0001(5) =0.0005(3) 0.0005(4) 7(5) ™ 0.340(4) 0.517(4) ~0,053(3) 0.006(2) d X0 0.357(5) 0.459(5) -0.062(3) 0.004(2) d X ©.3512(8) 0.4488(8) -0,0555(6) 0.0031(7) 0.007(1) 0.0025(4) 0,0004(7) -0,0003(4) -0.0006(5) No 0.6479(8) 0.5494(7) 0.2615(5) 0.9069(9) 0.0038(7) 0.0021(3) 0.0006(7?} 0.0004(4) 0.0003(4) r(6) X 0.646(4) 0.541(4) 0.258(3) 0.008(2) d X0 0.668(4) 0.493(5) 0.249(3) 0.004(2) d NX 0.6564(7) 0.4906(8) 0.2501(6) 0.0016(6) 0.0054(8) 0.0021(3) =-0.0001(6) =0.0007(4) 0.0007(4) XO 0.1033(8) 0.1966(7) -0.0655(6) 0.0059(8) 0.0025(7) 0.0015(3) 0.0031(7) -0.0001(4) -0.0006(4) D N 0.112(4) 0.202(4) «0.071{3) 0.005(2) d . X0 0.079{(4) 0.196(4) =0,061(3) 0,0009(9) d X 0.0808(8) 0.1893(8) =-0.0588(7) 0.0033(7) 0.0010(7) 0.0024(4) 0.0000(6) =-0.0001(4) -0,0010(4) NO 0.9191(7) 0.8111(7) 0,2643(T) 0,0014(6) 0.0031(7) 0.0018(3) 0.0005(6) 0.0009(4) 0.0001(4) r(s) ™ 0.918(3) 0.812(3) 0.257(2) 0.002(2) a X0 0.893(4) ©.803(3) 0.267(3) 0.0003(9) d *X 0.8985(9) 0,8058(8) 0.2712(6) 0.0053(8) 6.0038(8) 0.0017(3) 0.0039(7) =0.0006(4) =0.0010(5) »0 0.1295(8) 0.5851(7) ~0.3531(6) 9.0026(7) 0.0020(7) 0.0017(4) 0.0003(6) -0.0014(4) 0.0004(4) r9) x 0.125({4) 0.584(4) -0.364(3) 0.007.(2) 4 X0 0.117{4) 0.576(4) ~-0.347(3) 0.001(1) d X 0.117(1) 0.5781(8) -0.3459(4) 0.0049(9) 0.0014(7) 0.0014(3) 0.0007(6) «0.0012(5) 0.0004(4) ¥0 0.8836(8) 0.4214(8) 0.5508(6) 0.0022(7) 0.6033(7) 0.0013(3) 0.0003(6) ~0.0007(4) 0.0008(4) regy X¥ 0.879(3) 0.421(3) 0.543(3) - 0,004(2) d X0 0.868(5) 0.417(4) 0,555(4) 0.002(1) da »X 0.8716(9) 0.4153(8) 0.5578(6) 0.0029(8) 0.0035(8) 0.0016(4) 0.0010(7) -0,0013(5) 0.0005(5) no 0.66667 '0.33333 =0.306(1) 0.0038(7) 1 0.008(1) f t f ray 0.66667 0.33333 -0.321(5) 0,008(3) d X0 0.66667 0.33333 =0,321(8) 0.012(4) d .2 0.66667 0.33333 -0.309(2) 0.017(3) 1 0.011(3) t f 1 ®O 0.66667 0.33133 «0.047(2) 0.002(1) - 1 0.003(2) f t 1 wa(l) N 0.66667 0.33333 -0,045(4) 0.003{2) o X0 0.66667 0.33333 =0,055(4) 0.001(2) d X 0.66667 0.33333 «-0,054(2) 0.005(2) 1 0.0005(8) t t ¢ "0 0,333 0.66667 0.258(2) 0.005(2) 1 0.0006(8) 1 f 1 N2y o™ 0.33333 0.66667 0,254(4) 0.005(2) d x0 0.33333 0.68667 0.248(4) €.001(1) 4 - ‘WX 0.33333 0.66667 0,251(2) 0.0005(9) f 0.003(1) 1 t f ¥O 0.0 0.0 0.121(2) 0.006(2) ? 0.002(2) 1 f T mp N 0.0 0.0 0.118(3) 0.008(2) d f X0 0.0 0.0 0.084(5) T 0.009(3) 4 x 0.0 0.0 0.083(2) 0.006(2) t 0.005(2) 1 1 1 o 0.0 0.0 =0.139(3) 0.003(2) 1 0.002(2) 1 t 1 Ly X 0.0 0.0 «0.19¢2) 0.02(1) d X0 0.0 0.0 =0,15(2) 0.008(6) d NX 0.0 0.0 ~0,156(3) 0.004(2) t 0. 000(2) t -t 1 O - 0.233(4) 0.283(3) «0.188(3) 0.027(6)% 0.004(3) 0.014(4) «0,011{4) 0.012(4) -0,001(3) : N 0.28(2) 0.29(2) «0,20(2) 0.01{1) d Li(2) xo o.272(9) 0.310¢9) -0.221(8) 0.003(3) d nx 0.286(8) 0.250(7) -=0,13(2) 0.04(2) 0,019(7) 0.07(2) «~0.030(9) 0.04(2) -0.04(1) SCoefticients in the Temperature Factor: EXP-(B)h®+f; k?+B,, 07 +28; sht+28, k1428, ,hk). 'Bundafll Error (in parentheses) Corresponds to Last Bignificant Digit in Parameter. Defines Origin Along Z. B ‘x.ny Temperature FPactors for F, Na, and Li are Constrained to be Isotropic.. f Only those Bi; Adjusted as Independent Variables are Shown, are: "|"g;12’|‘ and ’. 3*;.'0- (L.', (1956)). ' Cremperature factor is mot positive-definite. Sysmetry Conatraints Upon the ’11 of all Atoms in Special ~m_|nm hyo - orixisal X-ray Paraweiers and Data. WO - Original Neutron Parameters and Data. XN « X-ray Data, NO Parameters fX = Neutron Data, XO Parameters. 170 : . ORNL-DWG 69-2079 Fig. 13.18, The Cation Contents of One Asymmetrical Unit of (Na,Li),ThsF3,; Showing the Coordination Polyhedron Around Each Cation. One-fourth of a unit cell is outlined. : 14. Chemistry of Molten-Salt Reactor Fuel Reprocessing Technology 14.1 DISTRIBUTION OF CERIUM, EUROPIUM, AND STRONTIUM BETWEEN BISMUTH AND LITHIUM CHLORIDE " D.M.Moulton J. H. Shaffer When solutions of lithium in bismuth are used to remove rare-earth fission products from an MSBR fuel salt, thorium will be extracted in about the same proportion. However, the rare earths can then be oxidized from the bismuth into a chloride without disturbing the thorium.! The separation is possible - because of the relative instability of thorium chloride. We have examined the behavior of cerium, europium, ‘and strontium in the bismuth—lithium chloride system. In an earlier experiment an LiF-BeF, salt containing 58.1 g Th and 0.235 g Ce (as fluorides) had been reduced with a lithium-bismuth solution at 600°C until all of the Ce and Th were removed from the salt. Analytical results corresponded to roughly half the cerium and 10% of the thorium (the latter corre- sponding to saturation) in the 3 kg of metal. The pot was cooled and cut open. The top part of the bismuth slug was cut off to avoid fluoride salt contamination, and the rest, weighing 2.279 kg, was put into a graphite-lined vessel with a molybdenum dip tube. Pure lithium chloride was prepared from reagent- grade salt by first heating it in argon to 150°C and holding for 24 hr to remove water. Then HC1 was added to the Ar, and the temperature was raised slowly to 575°C. The sweep gas was changed to 10% HC1in H,, and the salt was melted and taken to 650°C. The H,-HC1 treatment lasted 30 hr, during the last 10 of which the HCl concentration was doubled. A salt sample taken after 20 hr showed less than 10 ppm nickel. Another sample taken after the full treatment when dissolved in water showed no basicity to phenol- phthalein, Finally the salt was reduced with pure hydrogen at 700° till the HC1 content of the effluent gas was below 0.01 megq/liter. LL.E. McNeese, this report, sect. 1.1. 171 | Approkimately 1.842 kg of this salt was added to the bismuth at 650°C. The system was sparged with H,-HCl at-2 meq/min, and samples were taken after every 10 meq, using copper filters for the salt and stainless filters with graphite inserts for the metal. Analyses were made radiochemically for Ce and spec- trographically for Li and Th. The cerium-lithium ratios are shown in Fig. 14.1, Excluding the values indicated by open circles, for which the lithium numbers are out of line, we find log (D¢./D3;) = 8.312 £ 0.384. The thorium distributions could not be found accurately because the salt concentration was too low. At the end of the experiment the hydrochlorination was run for several much longer periods. The cerium content of the salt rose to 100% of that expected from the mass balance, but only about 10% of the thorium appeared, suggesting that the remainder was not transferred with the bismuth slug. - The distributions of eufopium and strontium were studied in the usual fashion, starting with salt solutions and extracting them into bismuth. About 2.155 kg of LiCl containing 76.7 g SrCl; and 0.153 g EuCl; was prepared by treating LiCl, StCl; -6H, O, and Eu,0; ina manner like that for the Ce experiment. The salt was ‘added to 3 kg of Bi at 650°C and reduced first with 6.09 g Th and then 17.16 g Li in increments. Samples were taken with copper filters and graphite dippers and analyzed as usual (spectrographic Sr). Strontium and europium distributions are also shown in Fig. 14.1. Again the thorium values were unusable. The distribu- - tion coefficients (Dy,/D};) are given by log Kg, = ' 2.284 1 0.025 and log Kg, = 1.024 £ 0.141. From these experiments it is obvious that all three of these fission products will distribute "strongly into lithium chloride, leaving the thorium behind. Since it will be very difficult to reextract either europium or strontium into a second bismuth pool, other methods will probably need to be used to clean up the chloride salt. The cerium experiment showed that a controlled oxidation can be used to strip the bismuth. By doing this ‘it should be possible to reach a rather high 172 ORNL—DWG 7O~ 6776 0 ,r "‘Eu ' -9 o — v A// / & : _g A/ . A /‘/ . . . ce 1| ‘J/ i /'./ : : — . . / _ -3 -2 -1 0 { 2 log &y Fig. 14.1'. The Distribution of Ce, Eu, and Sr Between LiCl and Bi at 650°C as a Function of D,. _ concentration of rare earths with a small amount of lithium chloride, so that distillation or some other purification scheme could be practical. 14.2 EXTRACTION OF CESIUM FROM LITHIUM CHLORIDE INTO BISMUTH BY REDUCTION ' WITH LITHIUM AT 650°C ~ D.M.Richardson J. H. Shaffer Experimental determination of the distribution of cesium and lithium between bismuth and lithium chloride when lithium metal is added for reductive extraction has been made. Knowledge of the behavior of fission product cesium in this stage of the metal transfer process is important not only to confirm that cesium is easily extracted from bismuth into lithium-chloride, " but also to determine the concentrations of lithium -which would be required for the back extraction to bismuth in the final contactor. The chloride salt was purified by sparging with 4 meq HC1 per liter of hydrogen in a nickel vessel at 650°C. Approximately 1 kg of this salt containing 1 mole % CsCl (labeled with !37Cs) was transferred to a standard ‘4-in. vessel containing 3 kg of hydrogen-fired bismuth in 2 mild steel liner. Salt samples were taken with copper filter sticks, and metal samples were taken through a mild steel dip tube into the bismuth using ~ graphite ladles. Weighed additions of lithium metal were made to the bismuth through a separate mild steel dip tube, -and equilibration was promoted by . argon sparging. The dlstnbutlon coefficients obtained for cesium were: log Dpg = log X, ; — 1.503, in which Xj ; is the mole fraction of lithium in the bismuth phase. The data are plotted in Fig. 14.2 and are shown with the derived theoretical line with slope of 1. The point at 0.0054 mole fraction lithium was not used in positioning the line since the counting rate of that bismuth sample was . only 21 counts min~! g~!. For comparison, contamina: tion with 100 ug of salt would elevate the counting rate by 50 counts/min. In several cases one of the pellets of a sample pair exhibited a higher counting rate by about '50 counts/min and was not included in the results shown. The equilibrium quotient obtained was: Dcyf/Dy ; = 0.031. This value compares favorably with " an estimate of 0.019 by D. M. Moulton.! Spectrochemical analyses of bismuth in the salt- phase were 15 to 35 ppm and indicate that little or no Li; Bi was dissolved in the lithium chloride. On the other hand the high contaminant level in the bismuth (1500 ppm ~copper, 1200 ppm nickel) might be the cause, in the salt phase, for the oxidation of any bismuthide ions that were formed. - : 11, M. Moulton, sect._ 14.4, this report, ORNL-DWG TO—6777 1073 DISTRIBUTION COEFFICIENT OF CESIUM —q 1073 2 5 02 2 5 10! LITHIUM CONCENTRATION. IN BISMUTH (mole fraction) . Fig. 14.2. Dlstnbutlon Coeffic:ents of Cesivm in LiCl and Bismuth Solutions at 650°C. ' 14.3 REMOVAL OF CHLORIDE FROM SIMULATED ., MSBR FUEL SOLVENT BY REACTION WITH ANHYDROUS HYDROGEN FLUORIDE F.A.Doss W.R.Grimes J.H.Shaffer According to current MSBR fuel reprdcessing tech- nology, lithium chloride will be employed for the back extraction of rare earths from bismuth to effect their separation from thorium.! This close coupling of . molten LiCl to the reactor system could possibly result - in the contamination of the fluoride fuel mixture with chloride ion. Although this event would probably not affect the physical and chemical properties of the fluoride mixture, chloride concentrations as low as 20 ppm (1% of Nyr, ) would have a significant effect on the neutron economy of the breeder reactor. Therefore an experimental program has been conducted to ex- amine the feasibility of an HF-H, sparge treatment for removing chloride from a simulated MSBR fuel solvent mixture and to provide a basis for preliminary engi- neering design of a process system, Specifically, equilib- . rium quotients and velocity constants were determined for the reaction Cl"+HF=F-+HCl . o GV - as functions of temperature, HF concentratxons in H2 , and gas sparge rates. - 11 . E. McNeese, this report, Chap. 21. 173 The reaction vessel was constructed from a 16-in. ~ length of 4-in, IPS nickel pipe with welded end closures of Y-in. nickel plate. Penetrations through the top plate provided for insertion of a ¥ -in. gas sparge tube and a thermowell, for gas exhaust, and for withdrawal of salt samples. Approximately 2.7 kg of LiF-BeF,- ThF4 (72-16-12 mole %), which had been previously purified, was contained in a 3%-in.-diam nickel liner within the reaction vessel. The gas sparge tube extended to within % in. of the bottom, so that a gas bubble path of about 5 in, in the salt was created; no other provisions for salt agitation were made. Chloride con- centrations of about 1 wt %, introduced as NaCl, were maintained for analytical expediency. Samples of the salt phase were withdrawn for chlonde analyses at the beginning and end of each reaction period. Acid gas concentrations in the influent and effluent streams were determined periodically by titra- tion with standard caustic solutions referenced to the flow of hydrogen and reaction time. Aliquots of the gas effluent titration solutions were submitted to the “Analytical Chemistry Division for chloride analyses. ~ Intermittent concentrations of chloride in the salt phase were determined by graphical integration of data from these gas-phase analyses. Seventeen reaction periods of 4 hr each were conducted at 650 and 750°C, at gas sparge rates of 0.5, 1, and 1.5 liters/min of H, (STP), and at nominal HF influent concentrations of 3, 5, and 7 meq of HF .per liter of H;. Operation of the experimental assembly was terminated at this point because of corrosion of the top plate of the reaction vessel. Reaction Velocity Constants. — The reaction noted by Eq. (1) would be expected to follow the second- order rate expression dN . . “—gf-l‘klNcWHF o ¢ 2 where N is the respective mole fraction of chloride ion and HF. However, for a gas sparge experiment where . the HF concentration in the gas influent is maintained constant, the reaction rate can more simply be defined by the first-order expression | Each experiment where reaction rate data could be obtained showed very good agreement with the inte- grated form of this equation. Typical results are shown in Fig. 14.3. The dependence of k; on HF concen- tration for these pseudo-first-order reaction conditions “becomes C kz ='k1NHF. ._ ‘ -' _ . ) o (4) The dependence of the reaction rate on gas flow rate is related to stripping efficiency and mass flow rate of HF, ‘and in these experiments to the relative agitation of the melt, Values for k, [Eq. (4)] for data obtained at 650 " and 750°C are plotted vs the mass flow rate of HF and related to nominal HF concentrations in Fig. 14.4. - Equilibrium Quotients. — Although these experiments were not conducted under ideal equilibrium conditions, 174 calculations of equilibrium quotients for each experi- mental point showed that steady-state conditions be- " tween the gas and liquid phases did prevail.. These values, based on the equation PHCl NF Q PHFNCI ‘ were 'derived from analyses of the gas effluent and from the chloride content of the melt, hypothetically calcu- lated as the cation mole fraction of NaCl. An average value of K, was obtained for each experiment. These _ : ORNL-DWG 70-6778 - 0.35 T T I i fSE ‘n(Nclo/NCLr) | NOMINAL HF CONCENTRATION: 5 mitliequivalents/liter Hp 0.30 F— H, SPARGE RATE: 1.5 liter/min — o DATA AT 650°C ® DATA AT 750°C I / 0.25 : — /) A 045 — / - - ‘ | / / ./ 0.10 . , : 7 : o ./ « 0.05 . . / '“(”uo/”a,) o 3 0 40 80 120 160 200 240 ' REACTION TIME (min) Fig. 14.3. Relative Rate of Removal of -Chloride from LlF-BGFg-'I'IlF4 (72-16-12 Mole %) by HF-H, Sparge at 650 and 750°C. ORNL-OWG 70-6779 NOMINAL HF CONCENTRATION (mnlluequuvolenis/mer Hz) TEMPERATURE {°C) o 3 650 - o 5 650 . A 7 . 650 ® 5 750 A T - 750 0.8 : — 7 _ z . . E / / s os / > 1. : ( 5= . 81 + U2 ogq / //./ o g& // e .: ® ° . / . O g / // A" a & o2 W A Ny 0 0 2 q 6 8 10 MASS FLOW RATE OF HF (milliequivalents/min) Fig. 14.4. Effect of Gas Sparge Conditions on the Removal of Chloride from LlF-Ber-ThF;, (72-16-12 Mole %) by Reaction with HF at 650 and 750°C. apparent equilibrium constants should approach true equilibrium conditions as the gas residence time in the salt phase becomes infinite (zero flow rate). However, values for K obtained at 650°C varied from about 6 to 10 and did not show a satisfactory correlation with gas sparge rates. Values for K, at 750° C were constant at about 4.4 £ 0.5. 14.4 CALCULATED BEHAVIOR OF SODIUM, RUBIDIUM, AND CESTUM IN THE MOLTEN-SALT BREEDER REACTOR REPROCESSING PLANT D.M. Moulton There will be moderate amounts of rubidium and cesium fission products in the molten-salt reactor, _occurring as stable and soluble fluorides everywhere except in the reprocessing stages. A calculation of their distribution between bismuth and both fluoride and chloride salts has been made as a guide to their behavior until experimental data become available. A similar calculation has been made for sodium because of the possibility of mixing of the fuel and coolant salts. First, standard free energy for the reaction - MX+Li=M+LiX ' () at 650°C where all substances are pure liquids was evaluated by adding integrals over the heat capacity and latent heats to the room-temperature enthalpy and entropy in the usual fashion. For those substances melting above 650°C, the integration was carried out to the melting point and then readjusted to 650°C using the heat capacity of the liquid. The values were taken from Kubaschewski et al.! A value of the heat capacity - for liquid RbF was arbitrarily assigned as C =17.7 cal deg™! mole™, rather than that generated from the formula in ref. 1, because of reservations concerning the original data.? It was also necessary to use a few other estimated values. | Activity coefficients were obtained from a varlety of sources. That for Li in Bi came from Foster et al.? and was evaluated at infinite dilution of Li. No data were found for Rb and Cs in Bi, so the excess free energy of K in Bi was taken from Hultgren et al* and assumed independent of temperature. The excess chemical po- tential of Na in Bi was taken from the data of Fischer et al® and again assumed independent of temperature. Kleppa and coworkers® have measured heats of mixing of the lithium halides with those of the heavier alkalies, 10. Kubaschewski ef al., Metallurgical Theimochenustry, Pergamon Press, Oxford, 1967. 2C. E.Kayloret al., J. Am. Chem. Soc. 81, 4172 (1959). 3M. S. Foster et al., Inorg. Chem. 3, 1428 (1964). 4R. R. Hultgren et al., Selected Values of Thermodynamic Properties of Metals and Alloys, John Wiley and Sons, New York, 1963, SA.K. Flscheretal J. Phys. Chem, 7! 1465 (196‘7) 175 all as liquids. Their results are expressed in the form ™ = N;N,(a + bN, + cNN,), where subscript 1 refers to 'Li, and it can be shown that the partial heat of mixing at infinite dilution of component 2 is given by a + b. It was assumed that all of the excess chemical potential was due to the heat of mixing, which was itself temperature-independent; Kleppa presents some evi- dence that would justify these assumptions. The free energy change for the reaction MX(diss. in LiX) + Li(Bi) =M(Bi) + LiX(lig) (2) is the sum of these values. If the final assumption is made, as suggested by J. Braunstein,” that the ratio of activity coefficients of LiF and MF does not change much on the addition of beryllium and thorium fluorides, then these numbers should allow prediction of the behavior of sodium, rubidium, and cesium on both sides of the metal transport step. Table 14.1 shows the standard free energies, activity coefficients, equilibrium constants, and distributions at X, {(Bi) = 0.002 and X; ;o; = 1, Xj ;5 = 0.72. The free energy is given for reaction (1) and does not include the activity coefficients. Experimental investigations of these extractions were initiated recently.® The calculations should be viewed with some caution since they involve small differences of rather large L. S. Hersh and O. J. Kleppa, J. Chem. Phys. 42, 1309 (1965); J. L. Holm and O, J. Kleppa,J. Chem. Phys. 49,2425 : (1968) J Braunstein, pnvate communication. sD M., Richardson and J. H. Shaffer, sect, 14.2, this report. Table 14.1. Standard Free Energies, Activity Coefficients, Equilibrium Constants, and Distributions at X (Bi) = 0.002 for the Reaction at 650°C of MX(LiX) + Li(Bi) = M(Bi) + LiX, Where All Are Liquids AG® is the standard free energy for all pure liquids A6® ko YLi(Bi) ) - o AG™ (kcal) m TYgalta o Kq Dy = X/ Xmx , L Fluorides o Na -11.67 . 0.138 1 0.34 27 0.075 Rb -16.48 1.71 0.047 650 1.8 Cs -21.61 1.71 0.090 20X 10° 56 , Chlorides - ' ' Na . -L78 - 0.54 £ 0,20, 4x107° Rb 138 0.052 0.042 g8x10°% Cs 1.85 0,030 0.019 4%x107° “yLix = L 176 numbers as well as arbitrary assignment of several values for unmeasured quantities. Nevertheless, on the basis of these one can make two predictions. First, the heavy alkalies will be extracted rather well from the fluoride in the rare-earth removal process. In fact, much of the - cesium will be detained in the protactinium isolation system but will probably find its way out eventually. Sodium will not be extracted very well. Second, all of these elements will tend to transfer into the chloride and will not be extracted appreciably even at high Li concentrations in Bi until their concentration in the salt gets to be much greater than that of lithium, where the activity coefficient ratios change. The very considerable difference between fluorides and chlorides is due mainly to the extreme stability of LiF, making lithium metal by far the most powerful reductant in the fluoride system. LiCl is actually less stable than RbCl or CsCl; and this is augmented by the rather substantial heats of mixing of the chlorides, so that the heavier alkalies tend to stay in the salt phase. 14.5 REDUCTION OF URANIUM AND PROTACTINIUM WITH TITANIUM D.M.Moulton -J.H.Shaffer W.R. Grimes An attempt was made to use metallic titanium as the reductant for U and Pa with the expectation that TiF; would be formed and could subsequently be removed as volatile TiF, after oxidation with HF, leaving the salt . solvent unchanged. Such a process could eliminate the need for an electrolytic cell. To a solution of 9.74 g UF,, 0.16 g Z1F,, and 1 mc of 233Pa in 2.91 kg of - LiF-BeF,-ThF, (72-16-12 mole %), contained in graph- ite with 3.00 kg of Bi, was added incrementally 5.05 g of Ti, or about 2.5 times the equivalence of the U + Zr. There was no appreciable extraction of Pa either at 600 or 700°C, its maximum distribution being estimated at 0.003. The Pa balance remained close to 100%. Then 0.4 equivalent of Li was added, and the Pa was extracted as usual. The P2 balance was excellent until the last addition, where 70% was in the metal but only 15% was in the salt, It may have adsorbed on the Ti, which was substantially above saturation. No other analyses were made in this system since it was quite clear that Ti is not a strong enough reductant to be - useful. 14.6 BISMUTH-GOLD ALLOYS AS EXTRACTANTS FOR RARE EARTHS 'AND THORIUM - D.M.Moulton J. H. Shaffer Some earlier experiments had suggested that the addition of gold to bismuth would increase the solu- bility of thorium, thereby raising both its distribution and that of the :rare earths. About 2.79 kg of LiF-BeF,-ThF, (72-16-12 mole %) containing 2.79 g of CeF; labeled with *#*Ce was placed in contact with 790 g of Bi in a well-bottom pot. All material in contact with the liquids was graphite except for a molybdenum dip tube. Thorium metal was added until the Ce distribution reached the saturation value for thorium in bismuth. Then enough gold (39 g) was added to bring its content in the metal up to 5 at. %, and the system was hydrofluorinated until all the cerium was oxidized, whereupon the extraction was repeated. Two more gold additions brought the gold percentage up to 10 and 20 at. %, doing an extraction and hydrofluorination at each. Dipper samples were used for both phases. In the first extraction without gold, a cerium distribu- tion of 0.055 was reached at 600°, agreeing with values obtained from earlier experiments. At 5, 10, and 20 at. % gold the maximum values were 0.044, 0.103, and - 0.052. No analyses were made for lithium or thorium. When gold was present the measured cerium concentra- tions were somewhat erratic, suggesting that a solid phase may have formed; this did not happen with pure bismuth. The addition of gold does seem to increase cerium extraction, but the improvement is rather small considering the amount of gold needed. This program is now being directed toward the use of other alloying metals to increase the thorium solubility. 14.7 BISMUTH-PLATINUM SOLUTIONS FOR THE EXTRACTION OF FISSION PRODUCTS . FROM MSBR SALT - D.M.Moulton J.H.Shaffer =~ W.R. Grimes We have continued to study the effect of adding platinum to the bismuth phase upon the reduction of various fission products. As discussed earlier,! the 1D, M. Moulton, J. H. Shaffer, and W. R. Grimes, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 151. formation of an insoluble platinide with zirconium may be a good way to separate that element from uranium. An experiment was run in which *$Zr was extracted into bismuth; a melt containing. 3.48 kg of LiF-BeF,- ThF, (72-16-12 mole %) and 0.1 wt % Z1F,, labeled with ®5Zr, was contacted with 800 g of a metallic Bi-U solution, all at 600°C. When about one-third of the zirconium had been removed, 11.32 g of Pt was added. In Fig. 14.5 are shown the percentages in the salt and the total balance (metal and salt, but excluding any precipitate) of U and Zr. Several samples were taken between the Pt addition and the last U addition, and the results are plotted on the abscissa over the break in the scale. Filter and dipper samples of the metal gave different results, indicating a solid phase, The solubility product X, X3, for the unfiltered samples was about 3 X 10712 except for the last, which was 2 X 10™**, This - is in reasonable agreement with the value found earlier. Filter samples would have given still lower values. It is clear that platinum will separate zirconium from uranium by forming an insoluble platinide. The ura- nium balance seems not to be affected. The lithium and thorium contents of the metal were not greatly altered by the platinum addition, and the equivalents of Zr lost from the salt were about equal to the equivalents of U gained if the valence of U is taken as 4+. In this - experiment the amount of platinum was slightly less than stoichiometric to form ZiPt;. Further work is in progress to determine the effect of excess platinum on the uranium balance. Strong compound formation with platinum is not limited to zirconjum; any of the metals in the early part of the periodic table should do so, especially those in the third and fourth groups with empty d orbitals. Asa result of discussions with Professor Leo Brewer we decided to see whether this could be used to separate thorium and rare earths. LiF-BeF, (66-34 mole %) with 888 ppm CeF; ‘was used as the salt so that the thorium balance could be followed. Six grams of thorium metal in 1-g increments was used to reduce 2.74 kg of this salt into 800 g Bi. Now platinum metal was added, first 8.23 g and then 7.29 g more, The changes in lithium, thorium, and cerium distributions and balances are shown in Table 14.2. Cerium was simply reoxidized into the salt. There was no measurable change in the thorium content of the salt, indicating that it probably precipitated. The change in the lithium concentration was quite unexpected, and if the reductant balance was maintained the lithium must have precipitated as well. About half of the added platinum disappeared, but this cannot be stoichiometrically related to the lithium and thorium disappearance since their systems with plati- num are complicated and not well characterized. - More lithium was added till all of the cerium was transferred from the salt to the bismuth. It did not seem to form a precipitate, Thorium analyses of the salt ORNL-DWG 70-6780 120 ' : T U BALANCE l Zr BALANCE ' @ — ¢ 100 =i "_'_-‘%. — == — \% i ’-O\ * 3 | N i = U IN SALT - i : Zr IN SALT ’ o 60 2 B E . g 40 - o ' A‘\‘i ) . ™ 20 — A 0 . | - | o > 4 6 8 10 12 116 _ URANIUM ADDED (g} Pt ADDED NO U ADDITION IN THIS INTERVAL Separation of U and Zr in Bi by Addition of Pt. Fig. 14.5. Separation of Uranium and Zirconium in Bismuth by addition of Platinum. Table 14.2. The Effect of Platinum on the Distribution and Balance of Li, Ce, and Th in LiF-BeF; and Bi ' 1st 2d NoPt sadition Addition "2 Dy, ' 0.011 0003 - 0.001 0.049 Reductant balance, % 29 7.1 <3 35 De, 3.4 0.1 0.3 169 Ce balance, % 101 97 97 90 Dpp 2.7 0.5 <0.2 Thbalance,% - 103 89 92 Pt balance, % 53 61 13 were not completed at this writing, but a reasonably constant and low concentration in the metal even under strongly reducing conditions suggests precipitation. The platinum balance fell to about 10% by the end of the experiment, and the product X2 . X, during this part of the experiment was fairly. constant. The addition of platinum to bismuth does not seem to help in the extraction of rare earths, but its interaction with thorium and lithium will be important in the uranium- zirconium separation, 148 EXTRACTION OF CERIUM AND THORIUM FROM LiF-BeF, (66-34 MOLE %) INTO BISMUTH AT 600°C J.H. Shaffer D.M.Moulton Previous experiments have examined the extraction of rare earths into bismuth from LiF-BeF,-ThF, mixtures “having constant thorium concentrations of 3, 6, 9, and 12 mole %.! As a continuation of this program, the extraction of cerium at low and varying thorium concentrations in the salt phase has been studied. The purpose of this experiment was to simulate the salt- bismuth conditions in the proposed electrolytic cell and to provide data relating to the thermodynamic activity of ThF, in the salt phase. - The starting salt mixture was 2.205 kg of LiF (65.75 “mole %), BeF, (33.87 mole %), and ThF, (0.38 mole %) with a cerium content of about 2.35 X 10~ mole fraction. This mixture was placed in a graphite-lined extraction vessel to which 3.00 kg of bismuth had been added. A total of 10.08 g of lithium metal was added in 5. H. Shaffer, D. M. Moulton, and W. R. Grimes, MSR Program Semiann. Progr. Rept. Aug. 31, 1968, ORNL-4344, p. 176. | | 178 tared increments to reduce cerium and thorium from the salt phase. Samples of each liquid phase were withdrawn at equilibrium after each addition of lith- fum. Salt samplés were obtained in copper filter sticks, and metal samples were filtered through stainless steel and collected in graphite cups. The results of this experiment were derived from chemical analyses of the salt phase for Li, Be, and Th, spectrochemical analyses of the metal phase for Li and Th, and radiochemical analyses of both phases for cerium. The distributions at equilibrium of cerium and tho- rium between the two liquid phases are related to the similar distribution of lithium by the equations CeF; + 3LiJ; = Ce}, + 3LiF I ¢} and ThF, + 4Li}; = ThY, + 4LiF . (2) The value for the equilibrium quotient for the re- duction of cerium by lithium [Eq. (1)] was 3.26 X 10°. The corresponding value for the reduction of thorium by lithium [Eq. (2)] was 2.59 X 102, The separation of cerium from thorium (& = D¢, /Dr,,) Was evaluated by the equation Ina=% InK—% InDy, , as shown in Fig. 14.6. The equilibrium quotient for the reduction of cerium by thorium was 6.41. ORNL—-DWG 70— 6781 ind =¥a In K —Ya In Op, =064 Do /Dyn=a, SEPARATION FACTOR 0.2 0.5 i 2 5 o Dp= Nmelul/”sulf , THORIUM‘ DISTRIBUTION Fig. 14.6. Effect of Thorium Distribution on_the Separation . of Cerium from Thorium During Reductive Extraction from LiF-BeF; (66-34 Mole %) into Bismuth at 600°C., Calculations of the material balance for thorium and cerium during the experiment showed a substantial loss of cerium from solution in the presence of solid thorium. Within the accuracy of these data the mole ratio of cerium to thorium in the solid phase was about 0.0022 while measurable concentrations of thorium and cerium were in solution in the salt phase, but it increased to 0.015 after the last addition of lithium to the extraction system. No further examination of this ~ behavior was made; however, evidence of a solid phase containing both thorium and cerium was noted by changes in their concentrations in solution on heating the bismuth slug from 600 to 700°C in the absence of a . salt phase during a subsequent experiment. Although this possible finding of solid solution formation of cerium and thorium has no practical significance to the proposed MSBR reprocessing scheme, it probably precludes further interest in sepa- rating cerium from thorium in bismuth by cold-zone deposition of thorium.? The relatively low cerium- thorium separation factors obtained in this experiment further substantiate the “free fluoride” effect on the ‘activity coefficients for thorium in the salt phase. Values for the equilibrium quotients obtained from this - experiment will be considered with others obtained in this program for possible description of the salt composition effect on the rare-earth extraction process. 2D, M. Richardson, W. R. Grimes, and J. H. Shaffer, MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL-4396, p. 193. 5. Development and Evaluation of Analytlcal Methods for Molten Salt Reactors " 15.1 DETERMINATION OF OXIDE IN MSRE SALT R.F.Apple J.M.Dale A.S. Meyer During the operation of the oxide determination apparatus at the High Radiation Level Analytical Laboratory, increasingly high pressures were required to maintain the flow of gas through the system during hydrofluorination. This increase occurred across the . NaF trap used to separate HF from water in the -effluent gas. It was assumed that during the repeated absorption and regeneration steps the NaF particles degraded to form a more tightly packed bed, partic- ularly at the trap entrance. As the more closely packed particles swell during HF absorption, the flow is restricted. A further examination of the system also revealed that two control valves had become inoperable. The hot cell was entered, and the NaF trap and the valve compartment were replaced. Three MSRE salt samples were received for analyses and were stored in the hot cell until the necessary repairs to the oxide apparatus could be made. The results of the analyses are listed below. A 96% recovery of oxide from an SnO, standard sample indicated that the repaired apparatus was functioning properly. Sample Oxide (ppm) FP 20-24 flush 71 FP 20-4 fuel 58 " FP 19-69 fuel a 4Unable to remove the salt from the transport container. 15.2 VOLTAMMETRIC DETERMINATION OF [U**]/[U] RATIOS IN MSRE FUEL J.M.Dale R.F.Apple A. S. Meyer Work on the zirconium reductions of the simulated - MSRE fuel at the Y-12 site was temporarily dis- continued in order to free the voltammeter for use in the determination of [U“*]I[U*] ratios in MSRE fuel samples. This work will be continued as time permits. Fuel sample FP-19-45 was taken on October 6 and held in a heated carrier until October 16, when the [U**]1/[U%*] ratio determination was made. Prior to taking this sample an amount of beryllium equivalent to 0.4% [U*]/[ZU] had been dissolved in the fuel. The voltammetric determination showed the presence of 0.13% [U*]/[ZU]. Another beryllium addition was " then. made, which increased the total beryllium U?* equivalent to 1.16%. After this addition, sample FP-19-60 was taken on October 20, and the determina- tion was made on October 22. This determination showed the presence of 0.07% [U*]/[ZU]. Because . the determination showed no increase in U*" concentra- tion, a third sample, FP-19-75, was taken on October 30 and run-on November 11. The determination on this ~ sample showed 0.02% [U*"]/[ZU]. During run 20 of the MSRE, November 25 to December 12, two beryllium additions were made to the fuel salt. The U equivalent of the amount of beryllium dissolved in the first addition was about 1.2%, and that of the second addition was about 1.7%. It was assumed that at the start of run 20 the U** was negligible, so 2.9% [U*]/[ZU] was the maximum amount expected to be present. After the first beryl- lium addition, sample FP-20-14 was taken and showed the presence of 0.11% {U"]/[ZU]. Sample FP-20-26, - taken - after the second beryllium addition, indicated that 0.20% U>" was present. The reasons for the discrepancies between the results of U analyses and - book values are still not completely understood. This problem will receive further investigation. ' 15.3 SPECTRAL STUDIES OF MSRE SALTS J.P.Young G.Goldberg The acquisition and installation of the hot-cell spec- trophotometer and related equipment for spectral studies of molten-salt reactor melts has been completed. 180 A description of the apparatus and technique originally devised for carrying out these studies has been given previously.l»2 The unsatisfactory operation of the s sample loading furnace, reported in the last progress report, coupled with the planned shutdown of the MSRE, caused a marked change in development efforts . - and required a major design change in sampling and sample-transfer technique. The original plan required the removal of a 5-g sample of MSRE fuel from the reactor. This sample was to be cooled to 200°C for transfer to the sample loading furnace for remelting and separation into five separate spectral cells, windowed or windowless. The cells were then to be transferred to an optical furnace for spectral study. The entire operation was to be performed without exposing the MSR fuel to the atmosphere. The multisample approach had to be abandoned. Instead, the removal of a sample of MSRE fuel directly in a spectral cell was considered. Two such methods of single sample removal were devised. In one, a windowless or screen cell' would be immersed in the MSRE pump bowl fuel. Although liquid would run out 3 P Young, MSR Program Semiann, Progr. Rept Aug 31, 5 ' 1969 ORNL-4449, p. 160. ' 23, P. Young, MSR Program Semiann. Progr Rept. Feb 28, - 1969, ORNL-4396, p. 202; Aug. 31, 1968, ORNL-4344, p. 192. of the cell as it is withdrawn from the fuel, a sufficient quantity would remain if flow rates and reservoir - volumes were properly selected. In the laboratory, this approach worked well when such a cell was rapidly withdrawn from a melt, but its performance seemed marginal at the withdrawal rate of sample capsules from the MSRE pump bowl (0.5 in.fsec). The sampling device used for obtaining MSRE samples is shown in- Figs. 15,1 and 15.2. In the exploded view, the small spectral (windowless) cell is shown at the bottom. This cell fits in the sample capsule body, the large component to the left. A sleeve, shown next to the cell, fits over the body and holds the cell in the capsule body by means of the tab on the sleeve which fits over the cell. The whole assembly is held together by the T-shaped key. The assembled sampler is shown in Fig. 15.2. This capsule assembly is designed so that the nonwetting fuel salt will flow freely away from the capsule as it is withdrawn from the MSRE fuel; likewise, clearances are such that no salt should be trapped in the various crevices. The spectral cell is designed, however, to hold melt in two ways. As the capsule is lowered into the fuel and withdrawn, melt is caught in the reservoir in the top part of the cell. On remelting, the cell is turned so that this reservoir would be up; thus the melt will run into and fill the . - - - _— " INCHES Fig. 15.1. MSRE Spectral Cell Sampler Assembly: Exploded View. 182 INCHES Fig. 15.2. MSRE Spectral Cell Sampler Assembly. lower part of the cell. The capsule is designed so that it can be disassembled with the one manipulator located in the sample-enricher station of the reactor. The spectral cell can then be placed in the sealed transport container for shipment to the hot-cell spectro- photometric facility. Although the capsule assembly did not work as well in the reactor as it had in the laboratory tests, it was possible to collect three of four planned spectral samples and have them transferred properly to the transport container. One of the four cells could not be removed from the capsule. The reasons for this are not clear. The original capsules and cells were made from copper and were hydrogen fired prior to use. Dipping the capsule in the salt is probably After installation of the in-cell apparatus and spectro- photometer components, a nonradioactive MSRE-type sample was melted to check out the optics, furnace, and spectrophotometer. The system worked quite well, and spectra of U(IV) in the molten salt were recorded. In attempting to transfer the first MSRE sample, a sample that should have contained no U(III), as it was taken just after the MSRE went to power on its last run, the sample loading furnace could not be used to open the transport container. It was subsequently found that this transport container had, in all probability, not main- tained an inert atmosphere around the sample. It was finally necessary to open this transport container out of a cleaning operation. It is conceivable that the close fit of the cell to the capsule body, the elevated tempera- the effect of high radiation levels caused the com- ponents to self-weld. After this experience, the cells “were made of nickel. The spectral cells were placed in the sealed transport containers, transferred, and stored at temperatures above 200°C in the hot cell until the spectro- photometer and sample transfer and handling apparatus were éompleted. The revised system was designed so - that the spectral cell was positioned in a bottom cap of each transport container. This sealed bottom cap was to be removed inside the sample loading furnace. The cell would then be-shifted on a turntable to another - position within this furnace to be picked up by a sealable lid assembly for final transfer, under inert atmosphere, to the spectral furnace for meltmg and spectral study. ture, the cleanliness of the similar metals, and perhaps | the sample loading furnace. The spectral cell was transferred in air to the spectral furnace. Although the furnace atmosphere’ around the sample was carefully evacuated and purged prior to melting the sample, on melting no real spectral evidence of U(IV), or U(III), was observed. It .is possible that the U(IV) was precipitated as UO, while it was stored in the faulty transport container. : After repair and mechanical strengtherung of parts of the sample loading furnace, another sample, MSR - 20-30, was brought to this furnace. The transport container which held this sample was opened without problem, and the spectral cell, a nickel windowless cell, - was transferred as planned to the optics furnace. This | particular sample was taken at a time when the U(IIT)/U(IV) ratio was believed to be about 2.8%, based on additions of Be®. At this level, the visual color of the melt should have been quite yellow. On melting, the solution appeared to be light green in color, definitely not “yellow. Although the solution appeared -to be somewhat turbid, spectra were obtained that demon- strated the existence of U(IV) in the melt. The turbidity of the solution, with attendant light loss, resulted in some loss of resolution of individual peaks, but over all the spectrum of U(IV) in the wavelength range 300 to 1600 nm was quite similar to that reported for U(IV) in nonradioactive LiF-BeF,s> this sample of MSRE fuel, approximately 200 mg in size, had a radiation level of about 75 r. The heights of the absorbance peaks were consistent with a U(IV) concen- tration in the sample of about 0.12 mole %. Further studies of this sample are planned to see if U(IH) can be grown into the MSR fuel by radiolytic evolution of F, gas at room temperature or by chemical reduction of U(IV) by uranium metal. 15.4 TRITIUM IN THE EFFLUENT GASES OF THE MSRE J.M.Dale R.F.Apple A.S.Meyer "During the operation of the MSRE in the latter part of run 19, analytical equipment was set up in the MSRE - vent-house sample station in order to measure the concentration of tritium in the various reactor effluent gas streams. A fractional part of the gas stream being analyzed was passed over a bed of hot copper oxide in order to convert the tritium to water. In general, two copper oxide temperatures were used, one at ~340°C and the other at ~800°C. The higher temperature allows the oxidation of methane, while the lower temperature does not. After oxidation the gas stream was passed through a bubbler tube which contained a few milliliters of water and then through an empty trap which was cooled to about —80°C. Two sets of traps ‘were used to ensure that all of the tritium was being collected. Sample streams were controlled at about 100 ‘em®/min to which 3 cm®/min of natural hydrogen was added as a carrier for the tritium. The amount of tritium collected for a specified period was determined by a liquid scmtlllatlon technique. - ~ The system was checked with a standard gas- of known hydrogen and methane concentrations. It was determined by both gas chromatographic and mass spectrographic analyses of the standard effluent from the copper oxide at 800°C that the hydrogen and methane were completely oxidized. Table 15.1 shows the results obtained durmg the last - week of run 19. 31, P. Young, Inorg. Chem. 6, 1486 (1967). Table 15.1, Tritium Analyses for Run 19 of the MSRE . CuO Tritium Da_lte Gas System . Temperature Co (curies/day) 10/24 Fuelsalt 340 9,3 10/27 = Fuelsalt - 640 1m1 10/30 Coolant salt - 800 0.62 10/30 - Cell _ 800 0.003 11/1 Fuel salt 800 22.7 An attempt was also made during run 19 to determine the concentration of tritium in the coolant stack air with the existing analytncal setup. This was precluded by the residual amount of tritium in the amalytical system and sampling lines from previous analyses. Although the amount of tritium found in these analyses was comparatively small (microcuries), the factor for converting to curies per day (1.38 X 10%), due to the large air flow in the stack, greatly magnified any sampling errors. For this reason a new sampling system was installed during the shutdown period between run 19 and run 20 so that.a more concentrated sample could be taken. A sheath was installed on a 2-ft section of a radiator tube and was fitted with tubing so that air could be drawn through the sheath for sampling. On November 21, about three weeks after the end of mun 19, two samples were taken from the fuel off-gas system. At this time the pump bowl was empty of salt, ~and the helium flow through the system had been dropped from 4.2 liters/min to 2.4 liters/min. The results obtained from the low- and high-temperature analyses were 3.15 and 7.44 curies/day, respectively, an appreciable amount compared with that found when the reactor was at full power. - Table 15.2 lists the fuel off-gas analyses obtamed during run 20, © The analyses made on December 2, after about one week at power, indicated that the total tritium concen- tration had not built up to the values found in run 19 Table 15.2. Fuel Off-Gas Analyses for - Run 20 of the MSRE Date - CuO o - Tritium Temperature ( C) o (curies/day) 12/2 P 800 , 11.6 12/2 ' ' 340 ‘ _ 9.7 12/11 ‘ ‘ 800 , 15.0 12/12 340 - 30.8 ? | ! 184 Table 15,3, Tritium Analyses of Air Drawn Through Radiator-Tube Sheath Date Sheath Air Flow Sample Volume Tritium (curies/day)” o (liters/min) (liters) Maximum Minimum? 12/6 : : 4.0 5.59 0.41 : 0.32 12/9 95 6.04 | 048 0.28 12/9 5.5 5.92 o038 0.26 12/10 . _ - 134 6.02 ‘ ' 0.42 0.14 12/10 = 7.0 6.2 0.69 : - 0.55 12/11 : 5.6 11.95 - 0.60 0.54 "V alues listed are adjusted for the entire 3600 ft of radiator tubing. Mlmmum values are adjusted for a blank which was determined from runs made on pure helium. (Table 15.1). This was also true for the sample taken on December 11. The reason for the jump in concentration of the sample taken on December 12, the day of the reactor shutdown, is not known. Table 15.3 lists the results of analyses made on air drawn through the 2-ft sheath on the radiator tube during run 20. Another approach was made to determme the amount of tritium being expelled through the radiator tubes. ~ This- involved taking 1-liter cooling air stack samples into evacuated bulbs which contained about 1 ml of - water. The gas samples, taken on December 11, were dynamic studies in molten fluoride salts. A nickel— nickel fluoride reference electrode? contained in thin- walled boron nitride was developed for emf measure- ments in molten fluorides at a working temperature of about 500 to 550°C. At hlgher temperatures, the BN deteriorates rapidly. _ In view of the limitations of the BN container material at the higher temperatures, a somewhat new 'concept for a reference electrode for fluoride melts is " under investigation. The authors independently fol- allowed to equilibrate with the water, and the tritium . was counted. These results are shown in Table 15.4. Samples 1 and 2 were taken through a tube which extended 3 ft into the base of the stack. The tube extended 18 ft into the stack for samples 3 and 4. The only apparent explanation for the difference between the results of Table 15.3 and Table 15.4 is that the sampling was not equivalent for the two methods. 1 Table 15.4. Tritium Analyses of Cooling Air Stack Samples Sample Tritium (curies/day)? 1 4.6 2 45 3 3.3 4 4,4 @V alues assume a total Stack_ flow of 203,000 scfm. _ 15.5 REFERENCE ELECTRODE STUDIES IN MOLTEN FLUORIDES D.L.Manning H. R. Bronstein - A simple and stable reference electrode would be very useful for electroanalytical measurements and thermo- lowed rather similar paths in this study, and only the most promising electrode system which evolved from combining our individual experiences is reported here. The new approach to the problem involves the use of a - single crystal of LaF; as the ionic conductor between the reference compartment and the melt being in- ‘vestigated. At the temperatures of interest, 500°C and above, single-crystal LaF; has sufficient ionic con- ductance, almost completely supplied by the mobile fluoride ion, so that any liquid-solid junction potential would tend to be minimal if the melts on either side of the crystal were essentially the same. Two models of the electrode are presently being tested. One model is similar to Egan’s approach® where an Ni/NiF, pellet is press-fitted with a nickel electrode against one face of the LaF; crystal (~% in. in diameter and % in. thick). The other face of the crystal fits against a fine-porosity (<5 u) nickel frit welded into the bottom of a nickel sheath. The electrode assembly is contained inside the nickel sheath (some are copper) but separated from the sheath with boron nitride machined to provide insula- tion between the nickel container and the crystal. The purpose of the fine-poromty nickel frit (% in. in 1 ORNL Chemistry Division, 2H, w. Jenkins, G. Mamantov, and D. L. Mannmg, J Electroanal Chem, 19, 385 (1968). 33, 1. Egan and R. J. Heus, Z, Physik, Chem 49, 38 (1966). diameter and %, in. thick) is to protect the crystal, as much as possible, from undue etching by the molten fluoride. Once the melt inside the frit at the frit-crystal interface becomes saturated with LaF;, then further attack on the crystal should be minimized due to the slow rate of diffusion of the melt in the frit. The other form of the electrode is similar except that the LaF, crystal is in the form of a small cup which contains the same molten fluoride salt as the test melt but saturated with NiF,. Contact to the melt in the cup is achieved with a small nickel electrode. The saturated NiF, solution (10”2 mole fraction) in the cup would hardly alter the slight solubility of LaFa (1 to 2 mole % by analogy to CeF3). Preliminary observations were made by immersing the electrodes in molten LiF-BeF; and LiF-BeF,-ZrF, saturated with NiF,;. The emf of the cell should be zero, Essentially zero (+10 mv) was obtained in molten LiF-BeF,; however, an excess potential of ~190 + 10 mv was observed in molten LiF-BeF,-ZrF,. This potential was stable for several days. The stability of the electrode is encouraging; however, the source of the extra potential remains to be resolved. ' Another test was made to observe the effectiveness of - the nickel frit in protecting the LaF, crystal. The crystal lost 5% of its weight after 40 days exposure to molten LiF-BeF,-ZrF, at 500°C. These studies will continue, and it is also planned to test the stability of the electrode in a reducing melt, namely, a melt that contains a high [U*]/[ZU] ratio. 15.6 REMOVAL OF OXIDE FROM NaBF, R.F.Apple A.S. Meyer The study of methods for removing oxide and water from molten NaBF, was continued. A pot containing 2600 g of NaBF, was treated with He-BF;-HF (50-50-180 ¢cm®/min) for 16 hr at 900°F, Titrating the effluent gas from the melt with Karl Fischer reagent indicated that about 100 mg H,O per hour was being removed by this method. Samples were removed prior to and after hydrofluorination and analyzed for oxide and water: . Water (ppm) | | _ Oxide (ppm) Prior to 1700 “ 1600 A_fter 1400 300 A batch, 1550 g, of NaBF, specially prepared by L. 0. Gilpatrick and contained in a graphite-lined nickel pot was treated with the He-BF;-HF (100-45-55 cm®/min) gas mixture for approximately 30 hr at 800°F. The salt was then cooled and evacuated to remove excess HF and BF;. The melting point data curve of this material indicated that it contained about 1 mole % of NaF. A portion of the salt was taken for oxygen and water analysis to check the efficiency of the purification: | Oxide (ppm) Water (ppm) Prior to ' © 300 ' 300 After : <90 40 15.7 A PRELIMINARY STUDY OF VOLATILE AICl, COMPLEXES J.D.Lodmell' A.D.Horton | A.S.Meyer A series of experiments were performed to determine whether volatile AICl; complexes of the chlorides of MSRE fuel constituents might be applicable to MSRP technology. The basis for this investigation was a report by Gruen and @ye? that complexes of the type NdCl,:(AICl3),, exhibit significant vapor pressures (~17 mm- at 700°C) at AL Clg pressures of about 3 atm. These preliminary studies were made to determine whether the complexes could be used for analytical separations of rare-earth fission products and to in- vestigate the possibility of separation of fission products from chloride reprocessing streams generated by the Metal Transfer Process.? The volatilization tests were made in a quartz tube that contained weighed quantities of chlorides (NdCl;, EuCl;, and LiCl) in a platinum or quartz boat. These chlorides were exposed to a stream of helium (~100 cm®/min) that had been contacted with AICl; near its sublimation temperature. Because of plugging problems . the concentration of AICl; in the helium carrier was limited to about 10%. The chlorides were exposed to the gas stream for periods of about an hour at ~ temperatures up to 750°C. With one exception, in which the exit end of the tube plugged and .an unknown pressure of AICl; was generated, no significant distillation of the rare-earth chlorides was observed. Traces of rare earth were ~detected in the AICl; condensate, but weight losses of 10RNL Summer Technical Student. ip. M. Gruen and H. A. @ye, Inorg. Nucl Chem. Letters 3, 453 (1967). _ 3F J. Smith, ). J. Lawrance, and C. T. Thompson, MSR Program Semiann, Progr. Rept. Feb. 28, 1969, ORNL-4396,.p. 285. 186 ~ the chloride samples were negligible. (Minor losses may have been obscured by deposits on the boats.) This -behavior at lower AICl; pressures appears consistent with the complex being formed with the unstable dimer Al Clg and with @ye’s suggestion® that several Al,Cl, molecules are involved in the complex. Conversely, LiCl distilled quite readily. A llqllld condensate was observed in the cooler portions of the tube when the temperature of the LiCl reached about 500 to 550°C. Distillation of 0.5 g of LiCl was . complete in less than 1 hr at 600°C. The condensate was analyzed and found to correspond to LiAICl, with about 20% excess AlCl;. A 1:1 volatile complex of NaCl with AICl; has been reported.® The LiCl was quantitatively removed (~99%) from 0.4 g of a 30:1 mole ratio Li/Nd mixture of the chlorides in about 40 min at 600°C. _ On the basis of the above experiments it was suggested that the system be investigated for potential ‘application to MSBE fuel reprocessing. In a possible separation scheme the chloride salt from the metal transfer process (principally LiCl containing rare earths in concentration enhanced over thorium) would be subjected to AICl; distillation, with the rare-earth chlorides remaining in the residue. The volatilized LiAICl, would then be decomposed, with the LiCl returned to the metal transfer equilibration system. Evaluation would require equilibrium constant measure- ments of the various chloride complexes including - thorium, a study of the effects of AlCl; contaminant on the efficiency of the metal transfer process, and an evaluation of the possibility of contamination of the fuel by aluminum. An evaluation by the Chemical Technology Division® indicated that the application of this process to the main reprocessing stream of the reactor would require the distillation of prohibitive quantities of salt. Accord- ingly, further study of these complexes has been assigned a low priority. . “H A fi'ye, pnvate communication. *E. W. Dewing, J. Am. Chem. Soc. 71, 2639(1955) SL. E. McNeese, private communication. ‘There was no work done on this activity during the report period. 187 ~ Part 5. Materials Development Our materials program has concentrated on the development of graphite and Hastelloy N with im- proved resistance to irradiation damage. We have approached the graphite problem by studying the - dimensional changes during irradiation of several com- mercial graphites. These studies have revealed which - graphites are most stable and, additionally, have shown what preirradiation properties are important in making some experimental graphites that likely will have improved dimensional stability during irradiation. The graphite used in an MSBR must also be sealed to reduce “the permeability to gaseous fission products. We are developing techniques for sealing graphite with carbon. The fast flux seen by the Hastelloy N is quite low, and the irradiation damage to this material is associated with the production of helium from the transmutation of 1¢B by thermal neutrons. The threshold boron level required to cause embrittlement is too low to be obtained commercially, so we sought to reduce the problem by changes in alloy chemistry. Modified compositions of Hastelloy N containing additions of Ti, Hf, and Nb look promising. The improved properties are associated with the formation of a fine dispersion of type MC carbides, and we are studying how the carbide type changes with alloy composition and aging. The compatibility of Hastelloy N with fluoride salts continues to receive attention; the main emphasis is currently placed on the proposed coolant salt, sodium fluoroborate. ‘This salt is more corrosive than other fluoride salts; we attribute this aggressiveness to the presence of adsorbed moisture. The variation of cor- rosion rate with moisture content and methods of removing moisture from a flowing salt stream are being studied. ' | . Some work has begun on developing structural materials suitable for use .in the chemical processing - plant. In this application the materials will be exposed to both salt and bismuth. Nickel-base alloys are highly soluble in bismith, and iron-base alloys mass transfer very badly. Molybdenum seems compatible, but it is difficult to fabricate. One avenue of research involves methods_of coating steels with molybdenum or tung- - sten to protect them from the bismuth. A second area of endeavor is that of finding brazing alloys that are compatible with bismuth for joining molybdenym. 16. MSRE Surveillance Program H.E.McCoy W.H. Cook - We maintain a surveillance program in the MSRE to follow the property changes that occur in the graphite and Hastelloy N used in fabricating the experiment. There are surveillance positions in the core and outside the reactor vessel. Graphite and Hastelloy N are included in the core surveillance facility. The deposition of fission products on both materials is studied (see Part ~ 3 for the results). The fluence received by the graphite is quite low, and the dimensional changes have re- 188 mained undetectable by the measuring techniques that we have available in the hot cells. The fluence is high enough to change the mechanical properties of the ~ Hastelloy N, and we have run both tensile and creep tests to follow these changes. Metallographic studies are performed to follow corrosion. Only Hastelloy N samples are in the facility outside the vessel. The samples more nearly reflect the condition of the reactor vessel. Hastelloy N of the same heats used in fabricating 189 the MSRE has been used in the surveillance program, and some additional alloys of modified compositions have been included. The Hastelloy N samples removed to date from the | MSRE are summarized in Table 16.1. The details of tests on the first three groups of samples have been presented previously,'™® and the present discussion will concentrate on the Hastelloy N samples demgnated as group 4. 16.1 PROPERTIES OF I-IASTELLOY N SAMPLES REMOVED FROM THE MSRE . H. E. McCoy As discussed previously' the metal samples removed after run 18 (designated group 4) were discolored. The ‘straps that held the surveillance assembly together were examined and found to contain cracks up to 3 mils - deep. We examined unstressed samples from the sur- veillance rods and found that they also had a modified structure to a depth of 1 or 2 mils (Fig. 16.1). Samples exposed to static barren salt for the same period of time did not have the modified structure (Fig. 16.2). To ~ determine qualitatively what effect this structural change would have on the mechanical propertiés, we bent a test sample at room temperature. The photo- micrographs in Fig. 16.3 show what occurred. On the tensile side of the sample, cracks formed that pene- trated about 4 mils. On the compression side there was no cracking. Several of the mechanical property samples were - examined metallographically after testing to determine the behavior of the modified microstructure under uniaxial testing at 25 and 650°C. At a test temperature of 25°C, the fracture was primarily intergranular in the irradiated sample (Fig. 16.4) and primarily trans- granular in the unirradiated sample (Fig. 16.5). There was a large amount of edge cracking to a depth of about -4 mils in the irradiated sample (Fig. 16.4), but there - was none in the unirradiated sample (Fig. 16.5). At a test temperature of 650°C the fractures were inter- granular in both unirradiated and irradiated samples. 1H. E. McCoy, Jt., An Evaluation of the Molten-Salt Reactor Experiment Hastelloy N Surveillance Specimens —~ First Group, ORNL-TM-1997 (November 1967). . 2H, E. McCoy, Jt., An Evaluation of the Molten- SaltReactor " Experiment Hastelloy N Surveillance Specimens — Second Group, ORNL-TM-2359 (February 1969). 3H. E. McCoy, Jr., An Evaluation of the Molten-Salt Reactor '. Experiment Hastelloy N Surveillance Specimens — Third Group, "ORNL-TM-2647 (January 1970). The irradiated sample (Fig. 16.6) had edge cracks to a depth of about 12 mils; the unirradiated sample (Fig. ' 16.7) had fewer cracks. The fact that the irradiated sample strained only 5% and the unirradiated sample 24% before failure should be considered in comparing ‘these photomicrographs. That the sample that strained the most had fewer edge cracks indicates. that the structure modification of the irradiated samples caused added embrittlement near the surface. - We know that the oxidation potential of the salt was relatively high during some of the time that the samples were in the reactor. This should remove chromium from the specimen surfaces, and indeed we observed that the - chromium concentration of the salt did rise. However, this should not embrittle the metal, nor should it pro- duce the modified microstructure. We are using several techniques to identify chemical changes near the surface. This work is incomplete, and we will report the - - findings later. ~ The tensile properties of the standard Hastelloy N samples removed from the core with group 4 were reported previously.? The creep tests have been com- pleted on these materials, and the properties will be compared with those obtained at other fluences. The stress-rupture properties at 650° for heat 5085 are shown in Fig. 16.8 for various fluences. The data for the control samples show that the rupture life decreased with increasing aging time at 650° to 15,289 hr and then increased slightly with further aging for 22,533 hr. The irradiated samples show a decrease in rupture life with increasing thermal fluence up to 9.4 X 10%° neutrons/cm?. No further deterioration occurred with ~an increase in fluence to 1.5 X 10%! neutrons/cm?. This behavior is as expected, since these two fluences represent boron burnups of 87 and 90% respectively. This small difference should not be detectable in the properties that we measure. The minimum creep rate ‘measurements shown in Fig. 16.9 show that neither - irradiation nor aging has an appreciable influence on the - minimum creep rate. The possible exception is at very high stress levels; however, these samples failed at such low strains that they likely are still in the primary creep - stage and have not reached the secondary or minimum - creep rate stage. The fracture strains for these samples . are shown in Fig. 16.10 as a function of creep rate. The - fracture strain decreases progressively ‘with increasing - 1w, H, Cook, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 165-68. ‘2H, E. McCoy, R. E. Gehlbach, and W. H. Cook, MSR . Program Semtann. Progr Rept. Aug. 31, 1969, ORNL-4449, pp. 168 70. " Table 16.1. Summary of Exposure Conditions df Surveillance Sampl‘es" : Group 1 Group 2 Group 3 "Group 4 Core, Core, Vessel, Core, " Core, Vessel, Core, Core, Vessel, Standard Modified Standard Standard Modified Standard Standard Modified Modified Hastelloy N Hastelloy N Hastelloy N Hastelloy N Hastelloy N Hastelloy N Hastelloy N Hastelloy N Hastelloy N Date inserted 9/8/65 9/13/66 8/24/65 9/13/66 6/5/67 8/24/65 9/13/66 4/10/68 5/7/68 Date removed 7/28/66 5/9/67 6/5/67 4/3/68 4/3/68 5/7/68 6/69 6/69 6/69 Mwhr on MSRE at time 0.0066 8682 0 8682 36,247 0 8682 72,441 36,247 of insertion _ ‘ ' Mwhr on MSRE at time 8682 36,247 36,247 72,441 72,441 72,441 92,805 92,805 92,805 of removal ' ' Temperature, °c 650+10 650110 65010 65010 650 £10 65010 65010 65010 65010 Time at temperature, hr 4800 5500 11,000 15,289 9789 20,789 22,533 7244 17,033 Peak fluence, neutrons/cm . : o ' Thermal (<0.876 ev) 1.3x10%® 4.1 x10%° 1.3x10'? 94x10%° 53x10%° 26x10? 15x10*! 51x10%° 2.5 X 10*°? Epithermal C>0.876 ev) 3.8 X 10 1.2 x 10?! 25x10'? 28x10%! 16 x10%! 5.0x10'% 37x10®! 91x10%° 3.9 x 10'? C50kev) 12X 102 37x10%° 2.1x10"° 85x10%° 48x10%° - 42x10"° 1L1x10*' 1.1x10%° 3.3x 10'° C1.22Mev) 3.1x10"° 1.0x10%° 55x10'% 23x10%° 1.3 x10%° 1x10'? 31x102® o08x10°° 86x10'® C2.02Mev) 1.6X% 10" 0.5 x10%° 3.0x10'% 11x10%° 0.7x10%° 60x10'% 15x102° 04x10*® 35x10'® Heat designations ~ 5081, 5085 21545,21554 5065, 5085 5065, 5085 67-502, 67-504 5065, 5085 5065,5085 7320,67-551 67-504 9 nformation compiled by R, C, Steffy. Revised for full-power operation at 8 Mw. 061 S 191 = | R-48864 1 " 2 : wid : = x.- LS é * . g § - 8 4 * 5 6 1z * A Fig. 16.1, Typical Fhotomicrographs 6t’ Héstefldy N (Heat 5085) Exposed to the MSRE Core for 22,533 hr at.650°C. (a) As . polished; (b) etched, glyceregia. S00X. T - , ' 192 fluence. However, the remarkable fact is that the unirradiated samples have fracture strains of 15 to 30% and that the lowest fluence studied reduced the fracture strain to 2.5%. This fluence produced a helium content of about 1 ppm in this heat of matérial. All of the results that have been presented are for heat 5085 (used to make the cylindrical shell of the MSRE). Heat 5065 (used to make the top and bottom heads of the MSRE) has been included in the program, and its performance parallels very closely that of heat 5085. Several heats of modified Hastelloy N have been included in our program, and detailed mechanical property tests similar to those just described for heat 5085 have been run. The results have been reported,* S but Fig. 16.11 summarizes the performance of some of these new alloys. All of the modified alloys have rupture lives approaching those of standard Hastelloy N in the unirradiated condition. The minimum fracture Fig. 16.2. Typical Photomicrographs of Heat 5085 After Exposure to Static Barren Salt for 22,533 hr at 650°C. (a) As polished; (b) etched, glyceregia. strains vary from 0.5% for standard Hastelloy N to 6% for alloys modified with 1.1% titanium or 0.49% hafnium. None of these alloys shows any evidence of higher corrosion due to the addition of these elements that are easily oxidized. However, as we have discussed previously, most of these alloys do not retain their good properties when irradiated above 700°C. Thus further chemical modifications are needed to obtain an alloy suitable for service at 700°C (see Chap. 18 for further details on this work). 3H. E. McCoy, Jr., An Evaluation of the Molten-Salt Reactor Experiment Hastelloy N Surveillance Speczmens — First Group, ORNL-TM-1997 (November 1967). *H. E. McCoy, Jr., An Evaluation of the Molten-Salt Reactor Experiment Hastelloy N Surveillance Specimens — Second Group, ORNL-TM-2359 (February 1969). SH.E. McCoy, Jr., An Evaluation of the Molten-Selt Reactor Experiment Hastelloy N Surveillance Specimens — T}urd Group, ORNL-TM-2647 (January 1970). 0.007 INCHES 500X 193 PHOTO 92445 10.005 in. . Fig. 16.3. Typical Phdtomicrographs of a Hastelloy N (Heat 5085)ASample Exposed to the MSRE Core for 22,533 hr at 650°C. The sample was bent in a vise. (¢) Compression side, 100X; (b) compression side, S00X; (c) tension side, 100X; (d) Tension side, 500X. Etchant, glyceregia. ‘ _ ' 194 PHOTO 99446 10.010 in. 100X 10.030 In. Fig. 16.4. Photomlcrographs of a Hastelloy N (Heat 5085) Sample Tested at 25°C Aftet Being Exposed to the MSRE Core for 22,533 hr at 650°C and Irradiated to a Thermal Fluence of 1.5 X 10°! Neutrons}cm (g) Fracture, etched, lOOX (b) edge, as polished, 100X, (¢) edge, etched, 100X Etchant, glyceregia. 195 PHOTO 99447 0.035 INCHES I 100X Iy Fig. 16.5. Typical Photomicrographs of a Hastelloy N (Heat 5085) Samplé Tested at 25°C After Being Exposed to Static Barren Fuel Salt for 22,533 hr at 650°C. (a) Fracture; (b) edge near fracture, 100X. Etchant, glyceregia. 196 4 | . PHOTO 99448 {0.010 in. I 0.035 INCHES 100X 10,030 in. Fig. 16.6. Photomicrographs of a Hastelloy N (Heat 5085) Sample Tested at 650°C After Being Exposed to the MSRE Core for 22,533 hr at 650°C and Imadiated to a Thermal Fluence of 1.5 X 10%! Neutrons/cm?, (2) Fracture, etched, 100X; (b) edge, as polished, 100X, (c) edge, etched, 100X. Etchant, glyceregia. ’ : _ . 197 - 0.035 INCHES N 100X | ] | : Fig. 16.7. Typical .Photomicrographs of a Hastelloy N (Heat 3085) Sample Tested at 650°C After Being Exposed to Static Barren Fuel Salt for 22,533 hr at 650°C, (@) Fracture; (b) edge near fracture, 100X, etchant, glyceregia. 198 ORNL-DWG 69—4470R fl\\\ 60 UNIRRADIATED — ™ . \IQ \‘1» | A \\‘:\ ‘\\\\ 50 e - IR o] TR g. \‘\.."‘ [~ o ™ 1~ T \§\\ O 40 = = o - i ¥ Wy N » . L' g L T | [T o THERMAL R Saoka o @ 30 - FLUENCE TIME AT / STk 2 (neutrons/cm?) 650°C (hr) \évc-..___ 1ol (& » e UNIRRADIATED: O / T~ . 5o |- @ UNIRRADIATED 4800 0 To , a UNIRRADIATED . 15,289 o T\{ L 4 UNIRRADIATED 22,533 | - a b ol @13 10'9 11,000 1.3 x 10%° nedtrons/cm? (MSRE) AND a26 x 10° 20,789 3-5 x 1029 neutrons/cm? (ORR) o 1.3 x 10?° 4800 - e ol ¢924 x 102 15,289 o 1.5 x 102! 22,5633 Lo b 1 | 1o~ 10° 40! 102 103 104 Fig. 16.8. Postirradiation Stress-Rupture Properties of MSRE Surveillance Specimens (Heat 5085) at 650°C. RUPTURE TIME (hr) ~ ORNL—DWG €9-44TiR2 70 / LA " 60 v off |4 A i Fa 50 ,/ — o J' A & D — Q|0 ; o o 40 Ank 41 € L1 S L L“ [a THERMAL = JN !/ / FLUENCE TIME AT 0 O gl o Llee ¢ (neutrons/cm?) 650°C (hr) g 30 AT ® UNIRRADIATED O 1l = /r o 1.3 x 0% 14,000 Lo 0 a 26 x 10° 20,789 20 20 " Le[i °o 1.3 x 10 4800 Il ) " el o 9.4 x 102 15,289 10 o 1.5 x 102! 22,533 || = UNIRRADIATED 4800 & UNIRRADIATED 15,289 | o ¢ UNIRRADIATED 22,533 1074 1073 1072 0™t 10° 10! Fig. 16.9. Minimum Creep Rate of Hastelloy N (Heat 5085) Surveillance Specimens from MSRE at 650°C. MINIMUM CREEP RATE (% / hr) 199 ORNL-DWG 69-4472R b 5 RANGE 2-5x102° neutrons/cm2 (ORR) . RN é 3 ~1.3%40'% neutrons/cm? w 0 _ § 21~ 2.6x10'° 3x102° E -fleuh‘OflS ma : neutrons/cmz. [T Ty 4 9.4 x4020 : 1.5x102! o neutrons /cm? - neutrons /cm?2 10~4 10-3 40~ 10—+ 109 4ot MINIMUM CREEP RATE (%/hr) 7 " Fig. 16,10, Variation of Fracture Strain with Strain Rate for Hastelioy N (Heat 5085) Surveillance Specimens at 650°C. ORNL-DWG 70-3980 .70 7T T TTTTITT T TT7 ™M\ STANDARD HASTELLOY N UNIRRADIATED \Q:\, \ 60 at 67-502, 5.3:10295'%\_ Ny . \ ~ NN N 50 s NN \ . h : SN I 2;, SN 67-554, 5.1x102° = p 7320, 5.110 N 2 M \'\\\ o - \ | I o Pl o 40 - % g \\ 1 ™~ ! . \\an. - by by $ - """'-u,.....l Py \ \ . ' ™~ A w . \"'Ny o, q\ b ) ~20 1 h"""-n 1.3x10°° STANDARD HASTELLOY N | | 67-504 . : ' \"'.""'-. 5 leoal : 20 9.4x40°OSTANDARD HASTELLOY N . : I 10 0 . 11 10° 0! 10? 100 10* .RUPTURE TIME (hr) Fig. 16.11. Stress-Rupture Properties at 650°C of Several Modified Alloys That Have Been Included in the MSRE Surveillance Program. The first number by each curve is the alloy designation, the second number is the fracture strain range, and the third number is the thermal fluence in neutrons/cm?. The alloy modifications are: heat 67-502 contains 2% W and 0.5% Ti, heat 67-504 contains 0.49% Hf, heat 67-551 contains 1.1% Ti, and alloy 7320 contains 0.5% Ti. 17. Graphite Studies W. P. Eatherly The purpose of our graphite program is to develop graphites that are suitable for use in molten-salt reactors. The graphite in these reactors will be exposed to high neutron fluences and must maintain reasonable dimensional stability and mechanical integrity. Further- more, the graphite must have a fine pore spectrum that will exclude salt and a very low permeability to gaseous fission products, notably !3%Xe. Our initial approach to the problem of dimensional stability during irradia- tion has been to evaluate all of the commercial products that look potentially acceptable. This study has been done carefully and systematically, so that the respective importances of several variables have emerged. Physical property measurements, x-ray studies, and electron microscopy have greatly supplemented the interpreta- tion of the results. We have found commercial graphites that wfll per- form acceptably in a moltensalt breeder reactor, but our studies offer encouragement that graphites with improved resistance to irradiation damage can be _ developed. Our ideas are being used in a small fabri- cation program to produce test samples of potentlally better graphites. The problem of obtaining low gas permeability is being approached by gas impregnation with a hydro- carbon. This problem is not independent of the dimensional stability of graphite, since differential . changes in dimensions of the impregnated material relative to the base graphite can increase the per- _meability. We have developed several techniques for obtaining low permeabilities, but the stability of the structures under irradiation will largely govern the choice of process. 17.1 FUNDAMENTAL STUDIES OF RADIATION - DAMAGE MECHANISMS IN GRAPHITE D K. Holmes S.M. Ohr W.E. Atkinson T.S. Noggle The earlier studies of “in situ” damage in graphite due to the ijlluminating beam of the 200-kv electron microscope had demonstrated that extensive damage H. E. McCoy, J1. occurred; however, the detailed nature of the observed structures could not be determined, since the contrast behavior did not systematically vary in a manner that permitted analysns Recent work employing a dark-field technique to monitor and study the damage has considerably improved the sensitivity to the appearance ~ of damage clusters and provides unambiguous evidence that the damage clusters are d1slocat10n loops of ~ interstitial type. These pbservatlons are being extended to evaluate the effect of irradiation temperature, dose, and electron energy on the size, density, and nature of the damage clusters and the incubation time for their appearance. The increased sensitivity of the dark-field method to presence of the damage clusters compared with bright- field observations leads to apparent incubation periods approximately a factor of 2 shorter than previously reported. In addition, it has been found that irradiation - with 150-kv electrons leads to damage clusters; how- ever, the incubation period is 2 to 3 times longer than .the incubation time for 200-kv irradiations. This obser- vation of damage from 150-kv electrons suggests that the threshold energy for. displacement of carbon atoms - is lower than the value of 33 ev reported in the literature. At this time the experimental techniques being employed are able to generate information which can be compared with the theoretical model developed which treats the kinetics of the nucleation and growth of interstitial clusters. This model is based on the work of Mayer, who advanced the hypothesis that boron plays a critical role in the nucleation of interstitial clusters. For the experimental observations to test or suggest changes in the theoretical treatment, it is ~ necessary that graphite containing known amounts of . boron be studied. The availability of good graphite crystals has been limited, and boron-doped material has not been available. To obtain an adequate supply of material, rock containing graphite single crystals was obtained from the Lead Hill graphite mine at Ticonderoga, New York, and chemical dissolution of the calcite matrix is being carried out to free the natural single crystals. Preliminary x-ray studies have indicated 200 that the material obtained contains excellent single crystals. Purification, followed by doping with boron, ~ will be carried out on this material to give specimens suitable for testing the theoretical model as to the role of boron in the formatlon of the interstitial damage clusters. S 17.2 PROCUREMENT OF NEW GRADES OF GRAPHITE W. H. Cook .One of the initial objectives of our graphite irradia- tion studies was to obtain types of graphite that were typical of all variations in nuclear graphite technology. To a large extent we have accomplished this. For this reason our rate of procurement has slowed. Our current procurement 'is directed toward (1) obtaining newly developed grades and (2) acquiring additional stock of the few grades that have shown the most resistance to radiation damage. The latter procurement is to supply materials for more detailed investigation of these grades. The different grades of graphite are used to extend our knowledge of the behavior of graphite in neutron irradiation. This involves new startmg materials and new fabrication processes. ' We have acquired three grades of pitch-bonded lampblack material, grades $20,! L31,' and SA45.? Grade S20 has been fired only to a carbonization temperature, while grades L31 and SA45 have been fired to graphitizing temperatures. Grades S20 and L31 are being evaluated in our irradiation studies. The purpose is to determine the potential of these poorly crystallized materials for use in graphite bodies under MSBR conditions. ‘We have a sample of grade HS-82-3 fine-grained 201 isotropic graphite! that has a low density (nominally 1.62 g/cm®) but high strength. This has also been. L _mcluded in our irradiation studies. . 17.3 GENERAL PHYSICAL PROPERTY ‘'MEASUREMENTS W.H. Cook We rafi a cursory set of physical property tests on the new graphites to determine if they are potentially cursory tests help us choose the grades that warrant more extensive testing. Table 17.1 is a summary of some of the cursory test data. _ Group 1 in Table 17.1 is a series of materials from Poco Graphite, Inc. Grades AXF-5Q! and AXF-5QBG have been the most resistant to radiation damage at 715°C of all grades tested to date.? The latter is an impregnated and graphitized product of grade AXF-5Q. Their good resistance to irradiation damage warrants more extensive studies. The grade AXM-5Q is of particular interest in the series, since it was crushed to- pro'vide the filler materials for the study of special binders.> The grade AXF-5Q-UFG is the same as the standard grade AXF-5Q except that the starting ma- terials have appreciably smaller grain sizes. Group 2, from Great Lakes Carbon Corporation, is a- comparison of “raw coke” graphite made in the laboratory, grade H364, and in a pilot plant, grade H337, by impregnating and graphitizing the base stock grade JOZ. Grades H364 and H337 have been the second most resistant to radiation damage.? The im- pressive point is that the base stock had an apparent density of only 1.57 gfem®, which was increased to 1.94 to 2.00 g/cm® by multiple impregnations, and that the final product showed relatively good resistance to radiation damage. The grades of group 3, from Speer Carbon Company, as a class are not as isotropic as the materials in the first two groups; however, they were included in the irradiation studies because of special types of startmg materials. Group 4, from the Graphite Products Division of Carborundum Company, consists of all fine-grained - grades of graphite that are even more anisotropic than the previous three groups. This degree of anisotropy makes these have low potential for MSBR core material. The limited gas permeability data continue to em- phasize the problems of attaining low permeability. The common grades have helium permeabilities greater than 1072 ¢m? /sec, and the special grades tend to be in the useful for MSBR’s or if they can contribute to our understanding of irradiation damage in graphite. These !Supplied by the Stackpole Carbon Company, St. Marys, Pa. 2Supphed by the Carbon Products Division of the Umon Carbide Corporation, 270 Park Avenue, New York. 107* cm? /sec range for these four groups. Future work will involve measurements of the permeabilities and other physical propertles of some of these grades after uradlatlon - 1Grades AXF-5Q ‘and AXF arte essentially the same; grade - AXF is used as the reference material for our gas sealing studies. . - 2C.R. Kennedy, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 175-177. ‘w. H. Cook, MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL4369, pp. 217-19. 202 Table 17.1. Some Physical Propertics of Various Grades of Graphite? : Apparent . Specific . Permeability Grade Type? Density Orientation® Resistance to Helium : (g/cm®) ' (microhm-cm) (cm?/sec) Group 19 AXZ-5Q NI 1.55 (45) ~ 2335(6) : , . - 2330(6) 2330(6) - _ - S [2330](18) AXM-5Q NI 17445) . 1670(6) : : ' ' : 1555(6) 1680(6) _ e o ‘ [1635}(18) AXF5Q NI 1.82 (39) ' 1410(6) \ - 1310(6) ' - _ ’ [1370](12) 3.5X 1072(4) AXF-5Q-UFG NI 1.87 (44) ' 1695(6) 54 x 1073 ' o - 1475(6) 1.7X 1072 1485(6) : [1550](18) AXF-5QBG NI 1.90 (3) ' ©[1160](12) 6.2X107? ' ' 1.8 X 107% . . Gmupzd , 1 JjozZ NI 1.57 (28) 2295(6) _ ~ . - 2220(6) . J H364 NI 1.94 (3) ¢ 830(2) . 92X%X1073 . ' . z 880(15) o H337 NI 2.00 (30) 740(3) 3.8 X 107%(4) ' : 800(3) 54X 107%4) [770]¢6) Group 39 : : 9948 A 1.90 (6) wg - 920(7) T 48X 107%(3) : : ag 1074(9) 6.3 X 1072(3) 9950 A 1.72 32) wg - 850(10) | ag ©1020(4) 9972 A 1.82 (32) T wg 830(4) 2.0X1072 | , ag 1005(10) HL-18 A 1.87 (45) wg . 1705(12) : ag 1750(6) . Group 49 Graph-i-tite “A” A 1.90 (28) z, Wg ~ 1000(6) ' _ c, ag 1385(6) Graph-i-tite “G” A 1.88(28) z, Wg ~ 1020(6) ’ . o c, Wg 1275(6)“ , : “The numbers in parentheses following data are the number of values averaged BNI = near isotropic and A = anisotropic. CThe directions in which the specific resistivity and helium permeability were determined: wg = with gram, ag = across grain, z = parallel with long dimension, and ¢ = chord direction. dGroup 1 is material obtained from Poco Graphite, Inc.; group 2 is material obtained from Great Lakes Carbon Corporation; group 3 is material obtained from Speer Carbon Company; and group 4 is material obtained from the Graphite Products Division of Carborundum Company. . ized by metallography, chemical analyses, particle size 203 17.4 GRAPHITE FABRICATION R. L. Hamner Our goal is to fabricate high-density isotropic bodies for molten-salt reactor applications. We fabricated bodies by warm uniaxial pressing and by isostatic pressing. These bodies were then impregnated to in- .crease density. The characteristics of - the starting materials and the results of our fabrication studies follow. : 17.4.1 Characterization of Matetials Four grades of filler materials were used: 1. JOZgrade graphite powder (Great Lakes Carbon Corporation), derived from a petroleum coke, un- impregnated, base stock produced commercially. Our interest in this material was evoked by the irradiation behavior of grade H337, which is made from the impregnated base stock. 2. Graphitized Robinson coke, air-blown graphite powder (Carbon Products Division, UCC) which does not form a needle-ike structure and, therefore, may be amenable to the fabrication of isotropic bodies. : 3. Santa Maria coke (Collier Carbon and Chemical Corporation), which is reported to result in bodies similar in unirradiated properties to the Poco grades 4. Thermax (R. T. Vanderbilt Company, Inc.), a submicron spherical carbon black, which is used frequently to enhance the packing characteristics of graphite particles in a fabricated body. These flours (excépt Thermax) have been charactet- distribution, helium density, and surface area. Inter- layer spacings and crystallite sizes are now being determined by x-ray diffraction. : By metallographic examination the JOZ powder appears -to ‘be the most graphitic of the group but contains more needlelike particles in the finer frac- ~ tions; the larger particles of all the materials: are relatively blocky; the Santa Maria flour has peculiar “pinwheel”shaped areas which tend to break down into short needles in the finer fractions. Other charac- terizations are shown in Table 17.2. A coal tar pitch - 11 ASL Progress Report-9 LA 4171-MS, February—Apnl '1969. Table 17.2. Characterization of Graphite Flours (Filler Materials)®® Particle Size Particle Size Range (u) Distribution . ; %) joz Robinson Santa Maria 100 <100 <120 <180 90 <97 . <91 <93 50 <43 <55 <34 10 <7 <. <3 “Helium density (glcm3): JOZ, 2.21; Robinson, 2.12; Santa Maria, 2.04. bSurface area (mzlg) JOZ 4.17; Robinson, 4.18; Santa Maria, 3.14. (Allied Chemical Company, grade 15-V) was selected as a binder for initial work, but only the fraction soluble in benzene at 80°C was used. As determined by our experiments, about 20% insolubles, which are presumed to be impurities and higher-meiting pitches, were filtered out of the pitch. The coking value as de- termined in an open boat in flowing helium was 33% at - 1000°C. Thermogravimetric analysis showed that ap- proximately 89% of the volatiles escaping during carbonization were given off between 200 and 500°C. 17.4.2 Fabrication by Hot Pressing The technique of fabricating by warm uniaxial pressing was selected, since it yields a product similar to extruded bodies because of the orientation effect but requires less time and materials. Billets nominally 3 in. in diameter and 2.75 in. high were pressed in a graphite - die at 1600 psi and approximately 1100°C; these were then graphitized at 2800°C (unpressurized) and im- pregnated ‘with pitch to increase density. The charac- . teristics of the pressed billets are shown in Table 17.3. The densification due to Thermax additions is apparent. Crystalline anisotropy was determined by a qualitative ~check which compares the maximum diffracted in- tensities obtained from the basal planes in flat thin specimens having the same thickness and whose faces are perpendicular to and parallel with the fabrication axis. This technique is probably sufficient for the initial development stages of our fabrication studies. The isotropy of the Robinson billets was far superior to that of either the Santa Maria or the JOZ billets: the latter showed the greatest anisotropy, probably because of the needle-like particles in the flour which promote preferred orientation, ' ‘We performed multiple impregnations on specimens from these six types, as shown in Table 17.4. Our - 204 - Table 17.3. Fabrication Conditions and Results for Wafin—PreSsed Graphite Billets? SM-1 SM4 ROB-1 ROB-5- JOZ2 JOz4 Graphite powder, % 100 85 = 100 85 100 85 Thermax, % N 15 0 15 0 15 Total amount of filler, g - 500 . 450 450 450 400 500 ‘Binder level, pph 20 20 20 20 31 20 Bulk density,? g/cm® 1.56 166 1.57 1.77 1.63 1.78 Crystalline anisotropy factor 1.09 1.085 1.00 ~1.00 134 1.22 “Pressed at 1080°C, 1600 psi; graphitized at 2800°C, unpressurized. bAs determined by dimensional measurements and weights. Table 17.4. Bulk Densii:y Increase with Successive Impregnations . Bulk DenSityd (g/cm?) for Given Number of Impregnations o 1 2 3 - 4 5 6 JOZ2 169 175 182 188 1.89 1.89 1.94 JOZ4 1.77 1.83 1.89 195 1.96 SM-1 156 165 172 179 185 188 1.88 SM4 165 1.74 181 185 1.87 ROB-1 156 - 165 174 181 184 185 ROB-5 177 1.82 188 193 1.96 aps detemuned by dlmensmnal measurements and weights of 0.5-in. -dlam rods. impreghation Ilarocedurewas to evacuate the specimens for 1'%, hr, then impregnate the specimens with coal tar ~ pitch in an autoclave at 160°C and 500 -psi for 1 hr. _ After the final impregnation the specimens were heated - at 2800°C to graphitize the impregnant. In general, for the same number of impregnations, specimens having - initially higher bulk densities had higher final densities and reached their maximum density with fewer im- pregnations. It is well known that carbon ylelds from binder (or impregnant) materials can be extremely variable and seem to depend on carbonizing conditions as well as on filler materials. The major variables affecting carbon yield are not well defined. OQur carbon-yield experi- ments on the 15-V coal tar. pitch alone have produced values ranging from 25 to 33%, even though the samples were from the same batch of pitch and run under the same conditions. _ We determined carbon ylelds of the pltch after each impregnation. Although there are some anomalies in the data, we observed generally that as the density in- creased, the carbon yield increased. We have no explanation for this, but we feel that such phenomena must be studied in the interest of improving fabncatlon techniques. The permeability of the JOZ-2 family was measured on a machined HFIRtype specimen at 1 atm He. The value obtained was surprisingly high in view of the bulk density of the specimen. Pertinerit data for the spec- imen are compared in Table 17.5 with those obtained on extruded Varcum-bound specimens which are now being irradiated in HFIR. Specimens for measurements by mercury porosimetry were cut from the same rods from which the permeability specimens were machined. These data show that bulk density alone is not a good indicator of permeability and that the materials in- volved must be considered. It appears that the mode of filling open pores is different for Varcum and pitch, as might be expected, Varcum being the better sealer material. Here, again, the need is pointed out for study of the behavior of binder and/or impregnant materials. Specimens from the ROB-5 body were submitted as candidates for irradiation because of their favorable anisotropy ratio of 1.00. 17.4.3 Fabrication by Isostatic Pressing ~ We selected the JOZ and Santa Maria materials for isostatic molding since these materials had an unsatis- factory anisotropy factor when fabricated by warm pressing. Compositions were formulated with and with- out Thermax additions and with 25 pph coal tar pitch as binder. The flowsheet for processing the isostatically pressed bodies is given in Fig. 17.1. - - After isostatic pressing, the pellets were nominally 1 in. in diameter and 2.25 in. long. The pellets expanded slightly during carbonization (approximately 3 vol %), probably because of gas release from the pitch. Im- pregnation procedures were the same as for the warm- pressed specimens except the pressure was increased to 1000 psi. 205 Table 17.5. Physical Property Measurements on Experimental Graphites Famil Bfilk Density? Apparent Densityb Open Pore Permeability Y (glem®) (g/cm3) Volume (%)?° (cm?/sec) Joz-2 1.94 2.18 11.23 8.2 X 1072 Y586 1.84 2.002 , 6.51 27X 1074 Y588 ©1.82 1999 6.87 47X 107% “By mercury porosimetry; JOZ-2 at 1 atm, Y586 and Y588 at 1.8 psia. bBy mercury porosimetry at 15,000 psi. ORNL-DWG 70-8238 I FILLER MATERIALS AND PITCH-BENZENE SOLUTION SLURRY BLEND TO STIFF MUD STAGE WHILE EVAPORATING BENZENE ‘ REMOVE REMAINDER OF BENZENE FROM MIXTURE AT {50°C UNDER VACUUM ' , ; Table 17.6 shows the increases in bulk densities of the specimens with successive impregnations. Note that the densities of the unimpregnated specimens are much lower than the same compositions that were warm pressed; this is to be expected. Also the densification due to Thermax additions is apparent, although with successive impregnations this effect became less pro-. nounced. Carbon yields from the impregnant were determined after each impregnation. As in the case of the warm-pressed specimens, the carbon yields in- creased with successive impregnations. : 17.44 Cbnclusfons The following conclusions are evident from even these : LCOOL TO ROOM TEMPERATURE : ' early results of our fabrication development work: . . : l 1. Isotropic bodies can be fabricated by techniques | GRIND IN HAMMER MILL TO —60 mesh | that tend to orient the filler particles, but this l FORM INTO PELLETS 1.25in. DIAM., 2.5 in. LONG BY UNIAXIAL PRESSING AT 500 psi l ISOSTATlCALLY PRESS PELLETS AT 20,000 psi AND $10°C. 7 : | CARBONIZE ON 72-hr CYCLE TO 4000°C IN HELIUM | | GRAPHITIZE AT 2800°C : | i —-—I IMPREGNATE WITH PITCH AT 1000 psi AND 460°C l Fig. 17.1. Flow Sheet for Prooessmg Isostatically Pressed u . Specimens. depends upon the characteristics of the filler par- ticles. - 2. The addition of Thermax is definitely beneficial in increasing bulk densities of graphite bodies. 3. High bulk densities can be attained by impregnating ‘low-density bodies. > Carbon yields from binder or impregnant materials are variable and need extensive exploration. In general, as the bulk density for a given family _increases with successive impregnations, the carbon yield from the impregnant increases. 17.5 MEASUREMENT OF THE THERMAL ] ol ! CONDUCTIVITY OF GRAPHITE - | - ) . | L—-{ CARBONIZE AT 1000°C ] J.P.Moore D.L.McElroy ‘ ' ' - l An apparatus has been constructed for determining ? o IGRAPHITIZE AT 2800°C o B I the effects of irradiation damage on the thermal _conductivities of graphites of nuclear interest. This apparatus, which is essentially a guarded linear device, is shown in Fig. 17.2, where the numbered positions 206 Table 17.6. Average Bulk Density Increases of Isostatically Pressed Specimens with Successive Impregnations Bulk Density (g/cm® at Given Number of Impregnations o 1 2 3 4 s Santa Maria-1 (SM-1): 100% Santa Maria flour, 25 pph 15-V pitch 1.335° 1476 1.601 1710 1.773- 1.791 Santa Maria-2 (SM-2): 85% Santa Maria flour, 15% Thermax, 25 pph 15-V pxtch 1486 1.61 1.71 1.768 1.824 JOZ-1: 100% JOZ flour, 25 pph 15-V pitch JOZ-2: 85% JOZ flour, 15% Thermax, 25 pph 15-V pitch 1438 1.583 1.700 1.799 1.862 1.572 1.696 1777 1.823 1.829 ORNL OWG 70-1944 SPECIMEN HEATER ASSEMBLY 0 OUTER FURNAGE o HEATER CO000000C O OO . No \\\\\\\_\\\\\\\\\\\\\\\\\\\\}\\\\\\\\\\\\\\\\\\\\\\\\\:\\\\\\\\\\\\\\\\\\\\\\W Noooooocooo GUARD CYLINDER HEATER ASSEMBLY . ~i2cm ; ™~ GAS-POWDER ANNULI SPECIMEN I~ GUARD CYLINDER | RS BASE HéATER/ NICKEL TUBES TO WATER COOLED HEAT SNk 4 Fig. 17.2. Schematic Drawing of Apparatus for Measuring Therma! Conductivity of Graphite. Numbers indicate the locations of thermocouples. denote the locations of Pt vs Pty oRh,, thermocouples. The specimen is a 1.0- We determined the changes in anisotropy of ‘ extruded stock near the braze from spherical samples whose poles were at the positions indicated in Figs. 17.3 and 17.4. The discontinuity in the curve could be . correlated with a discontinuous temperature gradient -during brazing. - Computer programs have been written to correct the x-ray data obtained from the Bragg maxima for calculation of lattice parameters and crystallite sizes. The data are corrected for background scatter, Lorentz polarization, structure, temperature, and sample trans- parency factors. Sample transparency is a very major contribution to the position and shape of x-ray line ~ profiles.* Instrumental contributions to the profiie ~width and shape are subtracted out by.the method of . ‘Stokes.> Average crystallite sizes are calculated accord- ing to the method of Warren and Averbach.® The standard used was a thin layer of Madagascar natural flake material sprinkled on the surface of a piece of 3The joints were prepared tay F. E. Clark, Chemical Engineer- . ing Department, Development Division, Y-12 Plant. “MSR Program Semiann.- Progr. Rept. Aug. 31, 1969, ORNL4344, pp. 228-30. . SA.R. Stokes,Proc ‘Phys. Soc. (London)Gl 382 (1948) B, E. Warren and B. L. Averbach, J. Appl. Phys. 21, 595 (1950). ) ROG?BVV | /\' 208 aluminum, ‘which makes a negligible contribution to background scatter. Calculated crystallite sizes are greatly influenced by the choice of sample standard, and it is believed that this is the best standard sample material we could obtain, Some of the unit cell ¢ lattice parameters and crystallite sizes calculated by this technique are shown in Table 17.9. It is concluded that the corrections used are adequate, since the lattice o ORNL-DWG 70-8239 0.78 , » ot Ry / , 0.7 / N\ o N\ 1 0.75 - O —-8 0.2 0.4 0.6 0.8 1.0 DISTANCE FROM JOINT (in.) Fig. 17.3. Anisotropy vs Distance from Braze for G.L. 1008 Graphlte Flour, " ORNL-DWG 70-8240 0.679 0.677 DISTANCE FROM JOINT (in.) ~ Fig. 17.4. Anisotropy vs Distance from Braze for Poco AXM Gtaphxte Flour, ¥ o 0.2 0.4 0.6 0.8 1.0 parameters calculated from both the (002) and (004) peaks are in substantial agreement. The center of gravity will not necessarily coincide with the peak position because of the asymmetry of the profiles. Figure 17.5 is a plot of the crystallite size vs ¢ parameter in which there is a linear relat:onshlp over this range of sizes. . With these careful x-ray analyses on graphites before and after irradiation, a better understanding of the material behavior can be obtained. | ORNL-OWG 70-8241 400 \ — < -wlo e 200 - .\\ 100 0 — - 6710 6.72 6.74 6.76 6.78 ¢ (X) Fig. 17.5. Plot of Crystallite Size (L ) vs Lattice Parameter c. Table 17.9. Lattice Parameters (¢) Calculated " from Centers of Gravity and Peak Positions and Crystallite Sizes (L) c(A) 002 ' 004 : ' Grade (002) ’ (004) L(A) Center of Center of ' ‘Gravity Peak Gravity - Peak AXF 0 ' 6.779 - 6.764 6.776 6.763 230 AXF (3000 C) v 6.758 6.752 6.751 6.738 330 AXF-5 ABG o 6.761 6.752 6.755 6.750 280 AXF-5 ABG (3000 C) 6.754 6.740 6.757 6.737 . 380 H364 _ . 6.737 6.725 6.742 6.723 430 H337 6.737 6.725 6.744 6.721 440, 17.7 ELECTRON MICROSCOPY OF GRAPHITE C.S.Morgan C.S. Yust ‘Specimens of hot-pressed - pyrolytic graphite and Ticonderoga natural flake graphite have been examined in an effort to relate the features observed in these more ‘perfect lattice arrangements to those seen in the complex polycrystalline structures. Dislocation images and extensive moir€ patterns were readxly observed, as were the modification of these features by variation of the diffraction conditions. The thicker the foil ex- ‘amined, the greater the number of dislocations and moiré patterns observed, due to the increase in the number of planes on thch dlslocatlons and misorienta- tions can exist. ~ The exammatlon of polycrystalhne samples by trans- mission electron microscopy involves the passage of the electron beam through many layers of graphite between which substantial misorientations may exist. The elec- tron waves transmitted through the polycrystalline foil would be expected to experience some modifications which might obscure detection of some of the defects which may exist in the lattice. To date, however, there has not been any observation of dislocations in the polycrystalline foils, a result which is not fully under- stood but which may result from the propensity of dislocations in small grains to move to the grain surface. Polycrystalline specimens of graphite examined by electron transmission microscopy are found to contain regions of aligned lamellae, or domains. Within the 209 difference between specimens thinned by mechanical polishing and specimens thinned by electrochemical methods. The sharply reduced evidence of lamellae in the mechanically polished irradiated foils suggests that the graphite is harder after irradiation. Irradiated specimens exhibit many dots of the orderof 1000 Ain diameter, the origin of which has not been determined (but see Sect. 17.1). They are found to be reduced in number by a 1-hr anneal at 2400°C. 17.8 GAS IMPREGNATION OF GRAPHITE C. B. Pollock We have been studying techniques for manufacturing graphite with very low helium permeability by sealing the surface with pyrolytic carbon. Our objective is to obtain a graphite with a surface permeability of 107° cm? [sec or less to prevent the absorption or retention of agapr'ecmble amounts of fission product poisons such l Sxe Commerclally avallable graphltes suitable for use in the core of an MSBR presently have helium per- meabilities of 10™* cm?/sec or greater. Beatty de- “veloped a technique by which graphite can be sealed domain regions the ¢ planes are approximately parallel, but wide variations of orientation may exist around the ¢ axis. Lamellae illustrating domains are made evident by microcracks occurring during mechanical polishing but reflect substructure properties of the graphite. Studies have indicated that. the expansion due to neutron irradiation may be related to the domain size. For example, the domain size of AXF graphite is smaller than that of H337 graphite, and the macro- scopic effect of irradiation is less in AXF than in H337. One of the prominent features observed in the * polycrystalline foils is 2 banding effect on the lamellae. Study of the moiré patterns in the more perfect lattices suggests that at least in some cases the bands are moiré - patterns. The fact that the band patterns are limited to the width of the lamellae may be related to the overlap of adjacent layers. ‘ The examination of irradiated graphlte has been extended to several graphites irradiated to a neutron fluence around 3 X 1022 neutrons/cm?. The transmis- sion electron micrographs of these specimens show little with pyrolytic carbon to the desired levels." We have now completed the third irradiation of graphite speci- mens sealed in this manner. Results from two earlier irradiation tests were misleading in that the cleaning operation (ultrasonic cleaning) is thought to have been harmful to the specimens. In the third experiment, 27 specimens were irradiated to fast-neutron fluences ranging from 5.2 to 139 X 10?! neutrons/cm? at a temperature of 715°C. Base stock graphites investigated in this experiment included Poco AXF graphite and Great Lakes Carbon Company grade H337 graphite. The specimens were gas impregnated by the vacuum- pulse technique to helium permeabilities ranging from 35X 1077 to 2 X 107 c¢m? [sec before irradiation. The first measurements taken after irradiation were the helium permeabilities. Results were found to range ‘from 4.27 X 107® to as great as 10™* c¢m?fsec. Helium permeability of the specimens as a function of fluence is plotted in Fig. 17.6. The gasimpregnated H337 graphite had an appreciable -change in helium per- meability at about 8 X 10%! neutrons/cm?®, while drastic change in the gas-impregnated Poco graphite appeared to occur at about 1.0 X 10?? neutrons/cm?. Helium permeabflntles of the spemmen appeared to be IR.L. Beatty, MSR Program Semiann. Progr. Rept Feb, 29 1968, ORNL-4254, p. 191. HELIUM PERMEABILITY (cmZ/sec) ~ the specimen to calculate ‘the permeability. _ ORNL-DWG 70-1598 10° o S o N 4] n g8 .2 4 8 8 10 12 14 (x10%) FLUENCE (neufrons/crn2 £>50keV) Fig. 17.6. Changes in Hehum Permeabihty w:th Neutron Irradiation at 715°C. Samples were sealed by gas unpregnatlon with butadiene. leveling off at about 105 cm?/sec after irradiation to the higher fluences. No variable other than fluence - appeared to exert any appreciable influence on the postirradiation helium permeability values. _ Superficially the increases in helium permeability observed in this experiment are. somewhat disappoint- ing. However, let us consider how the helium per- meability was obtained. The equations used to calculate the helium permeability and the sample geometry are shown in Fig. 17.7. We used the entire wall thickness of If we assumed that the permeability was controlled by a surface layer of perhaps 0.004 to 0.04 cm, then the calculated values for helium permeabfllty of that layer would decrease by one or two orders of magnitude below the values plotted in Fig. 17.6. Our preirradiation . values for helium permeability would decrease to less than 10 cm?/sec, and the postirradiation -values would be less than 10”7 cm?/sec, or very close to the desired levels. Previous radiographic studies of sealed samples indicated that the highest density of deposited 210 carbon is very near the surface.® Thus there is strong justification for using the smaller values for the thick- ness £, Dimensional changes were also measured on these samples as a function of fluence. Figure 17.8 is a plot of the fractional length and diameter changes as a function of fluence for Poco graphite, and Fig. 17.9 is a similar plot for H337 graphite. The graph includes data on 2R. L. Beatty, -MSR Program Semiann. Progr. Rept Aug 3I 1969, ORNL-4449, p. 174. ORNL-DWG T7O0-2274R e— 1,02 cm ——] 0.32 cm PHe = ' . N = o cm3/sec 127¢m M emem : SURFACE AREA =-2.67 cm?2 . * THICKNESS = 0.348 cm K=T.67cm LEAK RATE X SURFACE AREA AT THE MIDSECTION OF SAMPLE THICKNESS OF SAMPLE = cm2/sec Fig. 17.7. Sample Geometry and Equanons Used to Compute the Gas Permeablhty ORNL-DWG 70-1599 2.0 . : . POCO GRAPHITE N AXF-5QBG AND AXF-5QBG-3 1.5 —— © DIAMETER CHANGE A LENGTH CHANGE ® UNIMPREGNATED BASE STOCK ' . _ . 1.0 b — 2 . 2 . ' P e Je°° : NCE % 0 3 s 1 A b . o § ‘P2t E L —2 T ] s o ¥ . .” -0.5 L e -1.0 : 0 5 10 15 20 25 (0% FLUENCE (neutrons/cm? £ > 50keV) Fig. 178. A Comparisbn of the Dimensional Changés that .Occur in Gas-Impregnated Samples and Base Stock of Poco Graphite During Irradiation at 715°C.’ 211 ORNL-DWG 70~-1604 3.0 . GREAT LAKES GRAPHITE ' GRADE H337 . . o ' 2.5 [——t HL DIAMETER CHANGE : © HR DIAMETER CHANGE A HR LENGTH CHANGE 20 & - HL LENGTH CHANGE * . : e UNIMPREGNATED BASE S$TOCK . L L5 ¢ _ g 2 < 1.0 . ~J < a 0.5 o E A A @ . ’ e 4 . a * 9 o & . » ‘E ® 0‘ o a a0 o O A Y A, *°, -0.5 o -1.0 i 0o.. 5 10 15 20 25 (x102)) FLUENCE (neutrons/cm?2, £ >50keV) Fig. 17.9. A Comparison of the Dimensional Changes that Occur in Gas-Impregnated Samples and Base Stock of Great Lake Carbon Company H337 Graphite During Irradiation at 715°C. These samples were made from a pipe wall, and the designation HL and HR indicate that the sample axes were in “the longitudinal and radial directions respectively. ' gas-impregriéted_ graphite and unimpregnated base stock material. The data points on gas-impregnated graphite - were taken from this experiment, and the data points on ummpregnated base stock graph1te were collected from a number of other experiments.> We conclude that the dimensional changes of the gasimpregnated graphite are equivalent to those for - unimpregnated graphite out to the maximum fluence studied. Apparently the amount.of carbon being added to the graphite (maximum of 10% by weight for Poco graphite) does not cause appreciable change in the behavior of the graphite bodies at these fluences. The irradiation testing of a number of specimens from this experiment is to be continued in the next HFIR irradiation experiment. We wish to observe the behavior of the specimens under higher neutron doses for several reasons. First, we would like to know if the helium 3c. R. Kennedy, MSR Program Semiann. Progr. Rept. A.u'g.' 31, 1969, ORNL-4449, p. 175. permeabilities of the specimens continue to increase beyond the present level of 1075 cm?/sec or if they level off as Fig.-17.6 suggests. Second, we would like to observe the progressive changes in helium' permeability in several specimens presently exposed to lower neutron doses. From these results one could draw conclusions regarding the manner in which fast-neutron fluences affect helium permeabilities. Are we observing changes _in the impregnated carbon, or is the graphite matrix opening up? Hopefully, future results will shed light on this question. Pyrolytic carbon coatings that are potentially useful at moltensalt reactor fluences have been developed recently by D. M. Hewette of the Carbon Development group. Samples coated with isotropic pyrolytic carbon derived from propylene have been exposed to fast- neutron fluences of up to 2 X 10?2 neutrons/cm? at 715°C with remarkable dimensional stability (see Sect. 17.10). A number of graphite samples have been coated with similar pyrolytic carbon coatings and then therm- ally cycled from 25 to 1800°C with no apparent damage. Since the properties of the coatings and the graphite base stock are similar in the temperature and fluence range anticipated for MSBR’s, coating may be a better way to achieve low permeability than impregna- tion. For the next HFIR experiment we have prepared a - number of coated specimens for irradiation testing. All were coated with isotropic pyrolytic carbon derived from propylene at 1300°C in a fluidized bed and then were heat treated at 1800°C after coating. All speci- - mens have helium. ‘permeabilities less than 107 cm? Jsec. Figure 17.10 is a photomicrograph of the out81de coating of such a sample, illustrating the lack of porosity and the excellent bond between base stock material and coating. Figure 17.11 is a polarized light view of the coating and shows that it is isotropic. We plan to investigate the influences on permeability of the - base stock material and coating thickness in this experiment. We have continued to study scalmg up the gas-pulse carbon impregnation system. One significant problem has been the extremely poor efficiency of gas utiliza- ‘tion. With the present system only that portion of the - hydrocarbon that is in close proximity to the work piece is utilized for carbon deposition. The remainder * of the hydrocarbon and the hydrogen generated in the . process are wasted. To improve the efficiency we developed a simple gas-recycle system. The exhaust gas from the system is collected in a condenser maintained at a temperature below the boiling point of butadiene. A vent line allows the hydrogen to escape. The 212 Y-97999 0.014 INCHES ~ *0.003 in. 0.007 INCHES 500X . 10.007 in. 1 l‘ 10.005 in. 1 10.005 in. 250X 10.010'in. T r ~ Fig. 17.11. Polarized Light Photomicrograph of Isotrbpic Pyrolytic Carbon Coating on Poco Graphite. 250X, unetched. - . butadiene thus collected is then recycled, and gas . utilization efficiency has increased at least tenfold. -Also, some preliminary experiments have been con- ducted using a steady-flow gas-impregnation system rather than the gas-pulse system. We have found that in order to avoid excessive tar formation, it is necessary to use a very dilute hydrocarbon--carrier-gas mixture (0.5 to 2.5% by volume of hydrocarbon). Our studies show that Poco graphite can be sealed to helium per- meabilities of less than 107 cm? /sec in approximately 3 to 5 times as long as by the gas-pulse system. However, a system of this type could be easily designed to handle a number of moderator blocks, whereas the gas-pulse technique could handle more than one only with difficulty. It would appear to be much simpler to scale up a steady-flow system than a gas-pulse system. 17.9 HFIR IRRADIATION PROGRAM C.R.Kennedy HFIR graphite éxperiments 9 and 10 have beefl ‘ removed from the reactor and have been replaced by experiments. 11 and 12. The newer materials added to 213 the irradiation program and included in these experi- ments are given in Table 17.10. ‘Graphite grades 586, 588, ABS-5, ABS-6, YMI-3, ™ 350-10, YMI-13, and YM 350-11 are all made with isotropic cokes and fabricated by molding or extrusion. These grades were made to evaluate the binder systems in an attempt to obtain materials with properties similar to those of the Poco grades but fabricated by methods that could be used to make large moderator elements. The grinding of the coke or the fabrication process did impart some degree of anisotropy into these graphites. As noted by the results in Fig. 17.12, there is very little difference between the respective grades 586 and 588, YMI-3 and YM 350-10, and YMI-13 and YM 350-11. We also note that these grades have lost the initial delay in the densification process that is characteristic of the Poco grades. Therefore the lifetime for these materials would be considerably less than the Poco graph1tes and similar to more conventional graphites. We also note that the densification is less for these materials than for the conventional needle or acicular coke graphites. This is shown in Fig. 17.13, where the densification is shown to be a function of the original Table 17.10. Graphite Grades for HFIR Irradiation Progtam Graphit Bulk Forming Graphitis raphite orming o . raphitizing Grade Source? Den31 ; MethodP Filler® Bmderd Temperature o) 586 A 184 E AXF Varcum 2800 588 A 1.82 E 2033 , Varcum. 2900 YMI-3 A "1.62 WM AXM : ITX 2800 YM350-10 A 1.65 WM AXM - ' 350 Pitch 2800 YMI-13 A 1.72 HM AXM ITX 2800 YM350-11 A 1.75 HM AXM ' 350 Pitch 2800 AES-5 B 1.83 E Santa Maria, 15% Thermax Varcum 2800 ABS-6 B 1.89 E ~ Santa Maria, 15% Thermax ‘Varcum 2800 9950 C 1.71 M P : ' P Joz D i 1 : Raw coke — base stock for H364 and H337 grades HL-18 C 1.86 P - AXMATX A 1.75 M AXM, 15% Thermax = ITX 2800 XM-ITX A 1.75 M Santa Maria, 15% Thermax ITX 2800 TITX A 1.55 M Thermax ‘ - ITX 2200 T-V A 1.55 M Thermax Varcum 2200 S-20 . E 1.50 M Lampblack - Pitch 980 L-31 . E 1.66 M Lampblack Pitch 2800 HS-82 E 173 - M P : ‘ P 2800 Glassy carbon F 1.45 Cellulose ' 3000 4Source: A, Development Division, Y-12 Plant; B, Los Alamos Scientific Laboratory; C, Aitco Speer Carbon Company; D, Great Lakes Carbon Company; E, Stackpole Carbon Company; F, Tokal Electrode Manufacturing Company, Ltd. Formmg method: E = extruded; WM = warm molded 1400°C; HM = hot molded 2800 C; M = molded. Filler;: AXF, AXM, and 2033 were obtained from ground Poco grades AXF and AXM and Stackpole grade 2033 respectively; P = propnetary dBinder: Varcum = furfuryl alcohol; ITX = isotruxene; P = propnetary density. There are two significant exceptions to this behavior: grades YM 350-11 and YMI-13 do not densify. However it is seen that these graphltes made from isotropic particles do not, as a class, have the same potential for densification as the needle or acicular coke ‘graphites. This behavior suggests that the porosity responsible for shrinkage is external to the particle or the binder shrinkage cracks. This is because the growth within an isotropic particle would havé a -greater potential for removal of the porosity within the particle ORNL-DWG 70-8242 /Jsss YMI-13 ‘ / // YM 350-14] 2 100 In (14 AW/ 1) ; » 0 - P, YMI-3 POCO GRADESM 2 \\F\: -4 21 o 5 10 15 20 25 (x10°7) FLUENCE (neufrons/crnz. £ > 50 kev} Fig. 17.12. Volume Changes of Grades 586, 588, YMI-3, YM 350-10, YMI-13, YM 350-11, and Poco Graphites After Irradiation at 715°C. ORNL~-DWG 70-8243 ® > 3 z g o YMI-13 YM350-# 7~ AXF-5QB6 5 1 -_% AXF '?Bs 7 = 8 . E T YM350-10 o @ -Sf ‘ - a //NEEDLE AND ACICULAR § COKE GRAPHITES = o = -10 ’ ‘ : Y LT 1.8 . 1.9 20 ORIGINAL BULK DENSITY tq/cmS) Fig. 17.13. Max:mum Densxficatton of Graphite Grades When Irradiated at 715°C. 214 than a highly aligned particle. Also, much of the “differential growth within the isotropic particle is absorbed, and the result is that only small particle shape changes are available for reducing the binder shrinkage cracks. On- the other hand the needle coke growth, because of the high alignment of the particle, is ~ available for closure of the binder shrinkage cracks. This behavior, coupled with the fact that the two hot- molded graphites, grades YM 350-11 and YMI-13, do not have binder shrinkage cracks and do not exhibit a densification, suggests strongly that the porosity- - controlling densification is a result of binder shrinkage during processing. These observations suggest that the increase in life by a delay in the densification process is most likely to be obtained by use of a binderless isotropic graphite such as the Poco grades or the raw coke H364. It is also interesting to compare the growth rates of these graphites with the needle and acicular particle graphites. We can again use the growth rates when the volume - change with respect to fluence is zero' for comparison with R, the preferred orientation param- eter.! This is done in Fig. 17.14, showing that the growth rate for the Poco graphltes is about 3 times 10. B. Cavin, MSR Program Semiann. Progr. Rept. Aug. 31 1969, ORNL-4449, p. 172. ORNL-DWG 70- 8244 {(x1072% T T T ‘ i \POCO COKE GRAPHITES 8 - . | 5 \ i . v \ o~ : \ \ 8 . 3 N " . — - ?Jlg‘ 2 \“ 0 \ NEEDLE , . AND ACICULAR S COKE GRAPHITES 7 0.8 09 10 GROWTH RATE (lla 0.3 04 0.5 06 . O Fig. 17.14. Crystalhte Growth Rates of Graphite Grades During Irradiation at 715 C ‘ Table 17.11. Quantitative Microscopic Resu!ts on Irradiated Graphites ; . Exposure Bulk Volume Void Void , Ggg&i:e [neutrons/cm Change Area Number (>50 kev)] % » (No./mm?) | - x 10! X 10° AXF 12.9 -0.32 8.67 . 22.5 AXF-5QBG 38.2 6.22 13.62 30 AXF-5QBG-3 11.5 . 1.95 10.75 23.8 AXF-5QBG-3 383 4.40 17,79 32.1 AXFJUFG , 4.9 1.09 10.20 39.3 AXF-UFG 25.2 3.43 15.40 - 398 BY-12 o 204 314 9.10 289 ‘RY-12-29 : 24.9 8.6 21.35 - 23.8 ATI-S ' ~23.1 0.51 13.00 17.5 1425 . 24.4 1.83 10.48 254 H364 ' 192 0.47 7.94 - 204 higher than for the needle coke graphites. This behavior is consistent with previous observations of Bokros? showing that the growth rate increases with decreasing crystallite size. Preliminary measurements® indicate the Poco graphites and flour have a much smaller crystalhte size than the needle coke materials. The microstructures of several grades of graphite were examined by both the optical and the electron micro- scope. The samples were also evaluated using the quantitative microscope (QM), which yields the fraction of void area and the number of voids. The results are given in Table 17.11. The void area fraction generally is in very good agreement with the percentage of acces- sible void volume that we determined by helium density measurements. The number of voids is not an absolute number because of the irregular shapes of the voids; the same void can be counted a number of times. The num- - ber is only relative and is not significant except for large _changes. The fact that the void volume results obtained by the QM are in excellent agreement with helium den- sity and bulk density changes is fairly direct evidence that virtually all of the ac_cessible void volume is visible in the microscope. It is the inaccessible void volume which is ill defined and.is not observable by avallable microscopic techniques. A major result of these xmcroscoplc exammatlons is the observation of the irradiation effect on low-density 2], C. Bokros and R, J. Price, “ljhnensiofial Changes Induced . in Pyrolytic Carbon by High-Temperature Fast Neutron Irradia- tion,” Carbon §, 301 (1967). 30. B. Cavin, personal communication. components or phases in the graphite structure. Upon irradiation these very low-density phasés (approxi- mately equal to 1.5 gfcm®) densify to about 2.1 gfcm® and leave a large void volume in the graphite. If there is a large amount of the low-density phase, the increase in void volume is not observable from bulk volume changes but can be deduced from helium density measurements. The magnitude of this type of behavior is illustrated in the micrograph of grade 1425 in Fig. 17.15. The bulk density change of the irradiated sample is only 1.8%; however, the void volume has increased by over 9%, An example of a material where the increase in void volume is equal to the bulk volume change is grade | BY-12 (Fig. 17 16). 17.10 TRRADIATION BEHAVIOR OF PYROLYTIC CARBONS D. M, Hewette II C.R. Kennedy Bulk graphites consist of at least two graphitic constituents, filler and binder, that have widely dif- - ferent properties: large crystallite size and ‘high density for the filler and smaller crystallite sizes and low to intermediate densities for the binder. To predict how these bulk graphites behave under irradiation, one must know how the different constituents will behave. - Deposition of pyrolytic carbons in fluidized beds offers 2 means of obtaining homogeneous structures having properties that scan the extremes found in bulk graphites. The pyrolytic carbons, however, are not completely graphitic and have a turbostratic disorder of the layer planes. Irradiation of these types of materials furnishes a means of obtaining densification and 216 Y—-97047 l—0.025 INCHES—>| UNIRRADIATED IRRADIATED 2.4 x10%2 neutrons/cm?® Fig. 17.15. Photomicrograph of Grade 1425, Y-97043 UNIRRADIATED ~ IRRADIATED 2.0x 102? neutrons/cm? Fig. 17.16. Photomicrograph of Grade BY-12, dimensional-change data for monolithic structures that have properties varying from a density of 1.4 g/em? with a crystallite size of 50 A to a density of 2.07 g/em?® with a crystallite size of 160 A. A description of 217 the methane-derived coatings exhibit significant den- sification at low fluences followed by swelling and large dimensional changes at the higher exposures. Specimens ~‘of most of the structures tested in this experiment have the structures that were prepared for this study is given - in Table 17.12. Thin sheets of material of these various structures were irradiated at 715°C to fluences ranging from about 5 X 10! to 2 X 10?2 neutrons/cm?® (>50 kev) in HFIR. We have evaluated the density and dimensional changes, and they are presented in Figs. 17.17—17.20. These results indicate that at exposures up to 2 X 10?? neutrons/cm?, the high-density isotropic propylene-derived coatings are quite stable (i.e., undergo only .slight changes in density and dimensions). On the other hand, at the same conditions ORNL~DWG 70-4H9 ' FLUENCE (neutrons/cm2, £>50 keV} ' 0O 02 04 06 08 10 12 14 16 18 (x10%2) I T T JT T TT 1T 1 2.4 ‘/ \\ - O.1B Mev) Fig. 17.17. Effect of Fast Neutron Exposure at 715 C on the Density of Methane-Derived Pyrolytic Carbons. Al been reloaded for further irradiation. The stability of the propylene-derived structures, in which the dimensional changes in the parallel direction reach a maximum of about 2%, has great significance to the work on reducing the permeability of graphite (see ORNL - DWG 70-4917 FLUENCE (neutrons/cm2,> 50 keV) 0 5 10 15 xi0? 50 { ) | I l (1 b ! STRUCTURE PARALLEL PERPENDICULAR NOQ. DIRECTION DIRECTION 40 - 12 o . — 13 o a . 82 14 < < /'-"_ - 18 A s : w 30 — . ¢ 3 < . 3 | / 20 q : / 2 Z / ® 90 —t— ' & / /. = ¥ L o.. 0 (/ ’1" a—— — ” -qo y '?-.\ u - T - . s PP -.._. + o L\ ~~ S — 9 i——.{ £ \\\ . /1:/ \ ‘_fl' ek ~20 : TN ~8 ~30 0 2 4 6 8 10 12 4 (x10% FLUENCE (neutrons/cm?, > 0.18 MeV) Fig. 17.18. Effect of Fast Neutron Exposure at 715°C on - Dimensional Changes of Methane-Derived Pyrolytic Carbons. Table 17.12. Description of Pyrolytic Carbons Irradiated in HFIR Experiment MSR-8 18 Methane 1600 : - Deposition . Annealing Bacon Apparent Dsetr's:g:i::;:n So(;l;r:e Temperature Temperature Density Anisotropy Crystallite _ ' _ . ( C) { C) o 7 Factor Size (A) 6 © Propylene ' 1250 1900 2,07 S~ - 80 7 - Propylene ' 1250 2200 207 C~11 120 8 Propylene 1250 2500 2.07 ~1.1 160 12 ~ Methane 2000 ~ None 1.88 1.0 110 13 Methane . 2000 None 2.00 ~1.3 120 14 - Methane - 2000 None 2.08 - ~4 150 15 . Propylene 1250 None 199 o~ 30 6 Propylene 1250 1900 - 204 ~1.1 120 17 Propylene 1250 2200 2.04 ~1.1 . 160 - None .15 1.0 50 ks T e R b 1T L T s e In (1+2£), DIMENSIONAL CHANGE (%) Sect. 17.8). After exposures of 2.5 X 10?2 neu- trons/cm?, the dimensions of impregnated and base 218 stock Poco graphite specimens changed less than 1%, while grade H337 graphite specimens showed maximum changes of 3% (see the figures in the above section). Thus propylene-derived coatings deposited to seal the surface of these graphites may retain their integrity and sealing characteristics at fast fluences greater than 2 X 1022 neutronsfcm? (>>50 kev). Specimens of these two ORNL—DWG 70-1600R ' FLUENCE (neutrons /cm?, £> 50 keV) 0O 02 04 06 08 10 {2 14 16 18 (x10%) 2.2 N I I I I T T 17 0 5 " | © . ’/ 84 . 20 12 3l v et _| ~ 2 4 RYs " g - . 510 > IRRADIATION TEMPERATURE: 745°C 5 I I | l l 3 18 g P 1T 7 L T NG RE PROPYLENE DERIVED COATINGS 15 AS-DEPOSITED AT 4250°C 16 AS-DEPOSITED AT 1250°C, ANNEALED AT 1900°C 16 [ {7 AS—DEPOSITED AT 1250°C, ANNEALED AT 2200°C s | | 0 2 4 6 8 10 2 14 (x102%) FLUENCE (neutrons/cmZ, £ 048 MeV) Fig. 17.19. Effect of Fast Neutron Exposure at 715°C on the Density of Propylene-Derived Pyrolytic Carbons. ORNL-DWG 7049518 FLUENCE (neutrons/cm2> 50 keV) 0 5 10 15 (x10%Y) STRUCTURE PARALLEL PERPENDICULAR NO. DIRECTION DIRECTION i5 o ® 16 o . 17 o . £ 0 2 4 6 8 10 12 14 {x10%) FLUENCE (neutrons/cm3 > 0.18 MeV) Fig. 17.20. Effect of Fast Neutron Exposure at 715°C on Dimensional Changes of Propylene-Derived Pyrolytic Carbons, graphites sealed with isotropic propylene-derived py- rolytic carbon have been prepared by Pollock for uradlatlon testing. 17.11 CALCULATION OF LIFETIME AND INDUCED STRESSES IN MSBR GRAPHITE CORES S.J. Chang To analyze the effects of fast-neutron damage on the core graphite of a moltensalt reactor, three methods have been developed and discussed earlier.! The im- pulse-type analysis, the most general of the three, has been further developed to lift restrictions on the crosssectional shape. The method can thus be used to analyze effects of geometrical shape, boundary traction, thermal properties, and neutron-induced dimensional changes provided only that the creep coefficient and the neutron flux not be sensitive to position over a given cross section. The analysis is based on the constitutive equatlons of the form? Ex =(l —uzy*(d(’x - 1—_;(109 +(1 +p)aT + ) — peq M Ey =(l —#2)J* (doy —-i_--—p.dox) + (1 +pXaT + ) — pe, 7xy.= 2(1 + ”}I * doxy s where J(D) is the uniaxial neutron creep function J(D)=%+K(T)D+2—IE(1 _e 40Py and the symbol . denotes the convolution operation defined by J*da-f J(- D)-—dD In the above equations €, and o, denote, respectively, the strain and stress in the x direction, ¢ is Poisson’s ratio, E is Young’s modulus, €, and A, are constants, 1 MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL- 4396, pp. 229-31, 3. 1. Chang, C. E. Pugh, and S E. Moore, Viscoelastic Analysis of Graphite Under Neutron Irradiation and Tempera- ture Distribution, Fifth Southeastern Conference on Theoretical and Applied Mechanics, Raleigh, N.C., April 1970. is the coefficient -of thermal expansion, T is the temperature distribution, ¢ is the neutron-induced dimensional change, D is the neutron dose, and K(7)) is the creep coefficient from the uniaxial creep test. If the creep coefficient K(7T) is assumed to be a constant, independent of temperature, and the tempera- ture-dependent neutron-induced dimensional change is given by : VO, T)= 4, D + 4,(DD, then the stress, and therefore the strain, of the two-dimensional problem of arbitrary shape can be expressed as a linear combination of several fictitious ~elastic solutions with the same cross section. From this result the problem of an arbitrary shape can be solved, provided that solutions of the elastic problems are available. To apply the method to the- prehrmnary design calculation, numerical results were obtained for the core graphite, which was assumed to be in the form of long cylindrical tubes. Then the design lifetimes of the cylinders were evaluated based on two criteria, one by volumetric distortion and the other by axial strain, The specific problems solved were concentric circular cyl- inders with inner radius g, outer radius b, length L, and b/a = 6.667. Several values of b ranging from 4 to 6 cm were used to illustrate the geometric influence. The material parameters used were: ' Young’s modulus £=1.9 X 10° psi , Poisson’s ratio 4 =0.27 , coefficient of thermal expansion ¢ = 6.20 : X 1076 °c™ constant 4o = 2.0 X 107** neutron 1 em?, creep coefficient K(T) = (53 — 145 X 1072T ) +14X107°7%) 1077 (psim)” where T is Chosen at the inner radius 4, and neutron- induced dimensjonal change is given by . ¥(D,7)=A(D[(107**D)* + B(TX107**D)] , where A(T)-—(O 11-70X 10"51')/(5 7-60 X 10'31‘)2 per (1022 nvt) -3 and B(T)=2X (60X 1073T—5.7). In all of the above equations the unit of T was the degree centigrade. : Parameters related to the reactor were salt temperature Ty), = 625 — 75 cos ( %) °C thermal conductivity of the graphite K, TCK) —1 61 =Q. 358[,773 ] wattscm™ C7 , where T at n =g was used, heat transfer coefficient H =a —0-2{1.444 X 1073T(°C) — 0.228] wattscm ™2 °C™! , flux © (E > 50 kev) =4.5 X 10'4 sin (%) nvt, and gamma heating 0 =1.2 + 9.0 sin (%) watts/cm® . The numerical results are shown in several figures. Figure 17.21 shows a typical temperature distribution at Z/L = 0.6, a crucial section as can be seen from the lifetime curves in the subsequent figures. At this same cross section, the circumferential strain at r = b as a ORNL -DWG 69-7799 760 750 7 N 740 . / b=5¢m : \ 730 b/a=686T ZfL =086 \ 720 / \ 710 — \\ 700 : - \ 690 — ' \ 680 . ' 670 TEMPERATURE (°C) 0 0.2 04 08 08 . 1.0 (r-a)/({b-a) Fig. 17.21. Radial Temperature Profile at the Axial Position Z[L = 0.60. function of neutron dose increases rapidly after D=1 X 1022 nvt, as shown in Fig. 17.22. The increase of the axial stress 0, as D increases is seen to approach a straight line for D > 10*? nwt in Fig. 17.23. In fact, this - linear dependence of 0, on D at high fluence has been shown analytically in ref. 2 as a property valid for all ' ORNL-DWG 69—7796 20 S */ \ R // JN S \\ 1/ / €a= Ez (70) 0 1 2 3 NEUTRON DOSE (1022 mwt) Fig. 17.22. Circumferential Strain at QOutside Surface as a Function of Fluence Level at Z/L = 0.60. ORNL—DWG 69-383 125 ' A &=5cm . / 100 ———“1=08 : 75 : / T | | ¥ o 1 T 2 ) 3 4 NEUTRON DOSE (1022w} -25 Fig. 17.23. Axial Stress at the Outer Surface as a Functxon of Fluence Level at Z/L = 0.60. ' 220 stress components. Therefore, above D > 1022 nyt, the stress components can be calculated easily if their values at D = 1022 nyt are known. As the definition of graphite life, it has norma]ly been assumed the material will generate macroscopic cracks . '. after it begins to significantly expand beyond its original dimensions. We may define two ad hoc life- " times, first the fluence at which the material returns to its original volume and second, the fluence at which its axial strain returns to zero. The resulting lifetimes as a function of axial position in the core are shown in Figs. 17.24 and 17.25. It is seen the critical section occurs at Z[L = 0.57, approximately, and that the axial strain criterion to define lifetime is slightly more severe. As mentioned previously, the creep coefficient K(T) ‘was assumed to be constant in this analysis. More realistically, the above method is being generalized to account for a temperature-sensitive K(7). The analytical analysis has been completed, and numerical results are being computed. ORNL-DWG 69- 3835 40 | [ Tor . = 62575 cds (L) °C SALT T = 4.2 + 9.0 sin (-Z,_—') watts /¢cc ’/, 6.667 35 z \ | | / £ o g% / w "b=4cm / w . E \ ~—— W : ’ 6 / m e x/ 5 - : 0.4 0.5 0.6 0.7 0.8 Z/L . Fig. 17.24. Lifetime of MSBR Graphite Core Cylinders as a Function of Axial Position According to the Volumetric Distortion Criterion. Fig. 17.25. Lifetime of MSBR Graphite' Core (firlinders as a Function of Axial Position According Criterion. to _the Axial Strain 221 LIFETIME ( months) 8 A ORNL-DWG €9-3834 AT r=b c9= € T — . o4 05 06 Z/L 0.8 18. Hastelloy N H. E. McCoy The main shortcoming of Hastelloy N for use in molten-salt reactors is its embrittlement due to neu- tron irradiation. We have found that this embrittlement can be reduced by the addition of Ti, Hf, and Nb. Our work during the past months has concentrated on optimizing the concentrations of these elements. Small ‘lab melts of about 2 Ib have been used for this purpose, and some 50--and 100-b commercial melts are being obtained of the compositions that look promising. Our evaluation includes postirradiation mechanical property tests, long-term aging, and welding studies. Two compatibility questions are receiving attention. - First, the compatibility of Hastelloy N with steam must be evaluated. Samples are being exposed in two test facilities. Second, the compatibility of Hastelloy N with our new proposed coolant salt, sodium fluoroborate, must be evaluated. This work involves capsule tests, thermal-convection loops, and a single forced-convec- tion loop. Since the corrosion rate seems to be influenced by the water content of the sodium fluoro- borate, work is in progress to develop a process for removing water from the salt. 18.1 AGING OF MODIFIED ALLOYS C. E. Sessions Our previous studies have involved a comparison of the aging effects at 650 and 760°C of commercial heats of titanium-modified Hastelloy N. We showed' from a statistical fit of the tensile data that the results predicted the following: (1) a post-age ductility increase with increasing titanium concentration in the alloy, (2) a small decrease in strength and ductility with aging times greater than 1500 hr, (3) significantly lower ductility after aging at 760°C for each titanium-modi- fied alloy, and (4) a minimum in strength and maxi- mum in ductility between 0.5 and 1.2% titanium. 1€, 8. Lever and C. E. Sessions, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 18384, We have further investigated the effects of titanium and molybdenum content on the aging behavior of small laboratory- heats in order to check these trends. Figure 18.1 shows the 650°C yield strength plotted as a function of titanium and molybdenum content before and after aging at 760°C. In an Ni—12% Mo-7% Cr—0.06% C base composition, the yield strength increases for the first 0.1% titanium but does not show a continuous strengthening trend between 0.1 and 1.0% titanium. For most titanium concentrations, however, the 650°C yield strength is increased by aging 1500 hr at 760°C. In contrast to this behavior, the yield strength increases rapidly with increasing molybdenum concentration from 10 to 20% in the Ni—7% Cr—0.06% C base alloy (Fig. 18.1). The yield strength is also increased by aging 1500 hr at 760°C in these a]loys with varying molybdenum content. The tensile ductility values of these two alloy series are compared in Fig. 18.2. For tests in the assolution- annealed condition (1 hr at 1177°C), there is no effect of titanium on the 650°C tensile ductility. However, after aging 1500 hr at 760°C, the beneficial influence of titanium is apparent. As in the case of commercial ORNL— DWG 70— 6782 (x10%) ] 1 s SRR/ 7 40 AGED 1500 hr 7~ T & AT 760 °c\74 L/ I V A ' 2\ // /! © 35 Hf# £ 4 Z / \ s I 4 e | \ L/ | A //7/ 0 e . a 30 1 / s SOLUTION / > ANNEALED 7 [L—soLuTioN 25 AT 477 °C £ ANNEALED - . | Jh aTwrTeC | 21b LAB MELTS /7 l 20 L ) c4 08 12 168 12 16 20 TITANIUM (%) " MOLYBDENUM (%) - Fig. 18.1, Effect of Titanium and Molybdenum Concentratlon on the Yield Strength of Hastelloy N at 650° C. 222 223 , : } _ORNL—DWG 70— 6783 a5 > 45 s’ 7 30 : /soumon —1 40 - | o / o/ ANNEALED s / ATHT77%C 25 = 35 |get g | 7/ e I ' AN = 20 ,_"'AGED {500 he AT 76°°c ~ AT wr7 oc ' g .’l' . 30 f ) . AGED 1500 h § \r r Q 15 25 ¥ d . '? ’ \{ ' ‘ /’ " ¢ / 10 20 N . N, 21b LAB MELTS 5 15 | o? 0 10 0 04 0.8 1.2 i6 8 {2 16 20 ‘ TITANIUM (%) . MOLYBDENUM {%) Fig. 18.2. Effect of Titanium and Molybdenum Concentratxon on the Tensile Ductlhty of Hastelloy N at 650°C. Solution anneziled 1 hr at 1177°C and tested at a strain rate of 0.002 min_ alloys,!** the post-age ductility increases with titanium. The results indicate that for titanium contents <0.5 the ductility decreases on aging at 760°C and for titanium contents >0.5 the ductility increases. This behavior as a function of titanium was previously shown to result from the stability of the MC-type carbides at 760°C as determined by the titanium concentration in the alloy.? The influence of molybdenum on the ductility after aging is also shown in Fig. 18.2. In the solution- - annealed condition there is a ductility decrease at 20% molybdenum for alloys containing between 10 and 20% molybdenum. When aged 1500 hr at 760°C, each alloy shows a significant loss in ductility, with the exception . of the alloys with the lowest (10%) and highest (20%) molybdenum contents. The carbide phases precipitated in these alloys at 760°C are M,C> for concentrations ~ up to 20% molybdenum, where we get M C carbides at 650°C* and likely at 760°C also. Thus apparently we - get the smallest ductility loss on aging when MgC is -precipitated, although when M,C is precipitated, the detrimental effect is smallest at the lower molybdenum concentrations. This might possibly indicate that the influence of titanium content on the 760°C aging 2C.E. Sessions, MSR Program Semiann, Progr. Rept. Feb. 28, 1969, ORNL-4396, pp. 233-35. 3R. E. Gehlbach, C. E. Sessions, and S W, Cook, MSR Program Semiann, Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 193-95. 4R. E. Gehlbach et al., this report. behavior would be greater at higher molybdenum concentrations than we find from our studies of the 12% molybdenum alloys (Figs. 18.1 and 18.2). This conclusion is based on the observed larger ductility loss attributable to M,C precipitation at 14.5% molybde- num in Fig. 18.2. Additional aging studies are in progress on new alloy compositions which will be discussed at a future date. The microstructures developed on aging the commercial titanium-modified alloys have been discussed previ- ously; however, some recent results in Figs. 18.3—18.6 give the optical microstructures after 10,000 hr aging at 650 and 760°C. Figure 18.3 shows samples aged after a solution anneal of 1 hr at 1177°C, Fig. 18.4 shows samples annealed 1 hr at 1260°C and aged, Fig. 18.5 shows a sample with stacking fault precipitates, and Fig. 18.6 shows two alloys aged after a pretreatment of 1177°C plus 10% strain at room temperature. In general the precipitation that occurs at 650°C in each alloy is finer than that at 760°C. The grain boundary precipi- ‘tate appears to be coarser and more widely spaced at 760°C. The structures produced by solution anneals at 1177 and 1260°C are not appreciably different, as found by comparing Figs. 18.3 and 18.4, with one exception. In Fig. 184 the acicular structure after the 760°C age for the 1.2% titanium alloy corresponds to precipitates on stacking faults,® as shown by electron - microscopy,? and this microconstituent did not develop as profusely if the material was given a preage anneal at a lower temperature of 1177°C (Fig. 18.3). A lower-magnification optical photograph of precipi- tates on stacking faults in a heat of Hastelloy N containing 1.2% titanium after 3000 hr at 760°C is- shown in Fig. 18.5. The concentration of stacking fault . precipitates is fairly heavy near grain boundaries, and the presence of twms does not appear to affect their formation. ~ The influence of prestraining 10% at room tempeta- ture on the microstructure developed during aging is shown in Fig. 18.6 for the heats containing 0.15 and 1.2% Ti. Precipitation appears to be heavier at 650°C in ‘the low-titanium heat as compared with Fig. 18.3, but the effect of prestrain is insignificant at 760°C. In the - heat containing 1.2% titanium, prestraining has en- hanced precipitation at both 650 and 760°C, but particularly at 760°C. The effect of prior straining, as judged from Fig. 18.3, has been to promote the acicular type of precipitate which corresponds to precipitates on stacking faults. Thus stacking fault precipitation is enhanced by prestraining after an 1177°C anneal as well as by raising the solution annealing temperature from 1177 to 1260°C., | | 224 Y-98226 AGED 10,000 hr AT €50°C AGED 10,000 hr AT 760°C 0.45% Ti o 0.27 % Ti 0.45 % Ti Fig. 18.3. Microstructures Developed in Ni—12% Mo-7% Cr—0.07% C Alloys Modified with Titanium and Solution Anncaled 1 hr at 1177°C Before Aging. 650X, ‘ ¥Y-98225 \ AGED 10,000 hr AT 650°C AGED 10,000 hr AT 760°C 045% Ti 0.27 % Ti 0.45% Ti | 1.2%Ti Fig. 18.4. Microstructures Developed in N —12% Mo-7% Ct—0.07% C Alloys Modified with Titanium and Solution Annealed 1 hr at 1260°C Before Aging. 650X. 225 Y-98382° | 10.001 in. *0.003 in. Q.007 INCHES 500X I 15.608 . 10.007 in. F:g 18.5, Preclpxtates on Stackmg Faults in 1.2% Ti Heat of Hastelloy N Solution Annealed 1 hr at 1260°C and Aged 3000 hr at 760°C. 500X, 18.2 EFFECT OF CARBON AND TITANIUM ON POSTIRRADIATION PROPERTIES C.E.Sessions H.E. McCoy Defining the influence of alloying on the high-temper- ature irradiation damage is one objective of our alloy development studies on Hastelloy N. To study optimum titanium and carbon concentrations in Hastelloy N, an irradiation was conducted in the ETR at a design temperature of 600°C to a thermal fluence of 2 X 10?° neutronsfcm?®. The design temperature, however, was not achieved, and the temperature of the irradia- tion was determined from observations on melt wires to be approximately 550°C. Postirradiation creep-rupture tests were conducted at 650 and 760°C after this irradiation, and the results are. presented in Figs. 18.7—18.9. Six commercial heats of Hastelloy N with the titanium and carbon contents given in Table 18.1 were used. They included three levels of titanium (0.6, 09, and 1.2%) and two levels of carbon for each titanium level, with the high carbon content being nominally 0.08% and the low carbon content 0.03%. Figure 18.7 gives the postirradiation stress-rupture properties of the low-carbon alloys for each titanium level. The numbers associated with each data point are - the total fracture strains. At the low carbon level, the lowest-titanium heat (67-553) has a shorter rupture life at both 650 and 760°C, as expected. However, the two higher titanium levels show approximately the same rupture ‘lives. Thus at low carbon levels no particular “advantage is gained by increasing the titanium from 0.9 to 1.2% when the irradiation temperature is as low as 550°C. Generally, the ductility is greater at lower stress levels for a given alloy. In Fig. 18.8 the results for the high-carbon heats for these same three levels of tita- nium are presented. The results are very surprising, - since they indicate a maximum postirradiation creep- rupture life at 650 and at 760°C for the 0.6% titanium 0.0035 INCHES ! 10.003G n. ———— ety 000X 10.004C in. L Fig. '18.6. Microstructures Developed in Agmg 10, 000 hr After a 1-hr Solution Anneal at 1177 C and 10% Prestr:umng at Room Temoperature 1000X. (a) 0.15% Ti Aged 10 000 hr at 650° C; (b) 1.2% Ti Aged 10,000 hr at 650° C; (¢) 0.15% Ti Aged 10,000 hr at 760°C (d) 1.2% Ti Aged 10, 000 hr at 760°C. heat (67-550). Equally surprising is the fact that the lowest creep-rupture lives were found for the 0.9% titanium heat (67-549) and that properties intermediate between the 0.6 and 0.9% titanium heat were measured for the 1.2% titanium heat. Since the scatter in these test results does not appear to be excessive for a particular alloy (i.e., good straight lines can be drawn through the data), we must conclude that the role of titanjum at high and low carbon contents is quite different. These results are consistent with the hypothe- sis that above 0.6% titanium in Hasteloy N the irradiation damage (postirradiation properties) at 550°C is independent of the titanium level; however, we have previously shown that this hypothesis is not true for irradiations at 650 and 760°C. Nevertheless, the nature of the irradiation damage produced at 550°C has not been previously evaluated in detail, and the possible complication of carbide precipitation during postirradi- ation testing (for a 550°C irradiation) could mask our previously observed beneficial mfluence of higher tita- nium concentrations. Flgure 18.9 shows the effect of carbon content at the 1.2% titanium level on the postirradiated properties of Hastelloy N. These data, replotted from Figs. 18.7 and ) 227 ORNL- DWG 70- €784 (x10%) [ l 1.4 1.6 -'--—-. ™ " . .-"'-::::h.._ A4'2 - o 4.6 . \--'--.- ;-J- "*---\_____ 2.9 P\‘--.._:Emg -.-.""'n-. "--\ 40 -_-"'-.._ ' [S— 3.5 11 ,. --\-__--l 2 7.5 1 v 30 @ 72 '-U; 7.34 \“.'8.5 s NN || 70 N 20 ) 6.1 4P TIPAE.7 TESTED AT Ti CONTENT NN 5.6 650°C 760°C HEAT (%) 5.6 '\‘\% ol © o 67-551 1.4 N T a A 67-570 0.9 o & 67-553 - 0.6 SRR i o 1o° ! 10° 10° Fig. 18.7. Effect of Titanium Content and Test Temperature on the Postirradiation Creep-Rupture Propertres of Low-Carbon Heats of Haste!loy N. Samples solution annealed 1 hr at 1177°C and irradiated at 550°C to a thermal fluence of 2 X 102° RUPTURE LIFE (hr) neutrons/cm The numbers by each point mdrcate the fracture strain, ORNL-DWG 70- 6785 (x10%) l | I ” +~Llli6.0 - i ""C\ 0g8.9 -_-.----' 4.8 |'."--:-., : 50 — . 4.6 SR =T . ) ~ -—__i‘T -5 \'7:'60-::53..2... 40 g | N 16.8 § 30 ey 4.8 £ \ *"'\ "U-) m.4 ‘\12-1 e ,....Iq.. . 12.9 T~ 20 S .:2\.‘_;\ 5.4 || ’ : N st 1™ TESTED AT “Ti CONTENT ~\l‘ | 650°C 760°C = HEAT (%) 70" 10}— o ° 67-548 1.2 - A A 67-549 09 ' c - 67-550 . S IIIIJH | 0! ® 10 102 03 ‘Fig. 18.8. Effect of Trtamum Content and Test Temperature on the Postlrradratxon Creep Rupture Propertics of High-Carbon Heats of Hastelloy N. Samples solution annealed 1 hr at 1177°C and irradiated at 550°C to a thermal fluence of 2 X 10°° RUPTURE LIFE (hr) neutrons/ em?, The numbers by each point indicate the fracture strain. 228 - ORNL-DWG 70- 6786 e | o | LTI 1 1.6 e 6-0 ’ o | "T HIGH CARBON U T s 50 : _ | _76 W¢" _-bl.-..--.-'- a0 | LOW CARBON ~_ |35, : et 0 30 2168 . 4 T% 2 ||| o 8.5 g:i 'HIGH CARBON o v 20 11 NI -5 - - TONL g ~ TESTED AT CONTENT (%) o At o | e50°C 760°C HEAT Ti C 5.6 o}— 8 A 67-548 1.2 0.8 LOW CARBON N .o e 67-551 1.1 0.03 N 10°! 0% - 10' Fig. 18.9. Effect of Carbon Content on the Postirradiation Creep-Rupture Proper annealed 1 hr at 1177°C and irradiated at 550°C to a thermal fluence of 2 X 10 RUPTURE LIFE (hr) 10° 103 ties of Titanium-Modified Hastelloy N. Samples neutrons/ cm?, Numbers by each point indicate the fracture strain. Table 18.1. Chemical Analyses of Commercial Heats of Titanium-Modified Hastelloy N Irradiated in the ETR Chemical Analysis (wt %) Element 67-550° 67-553 67-549 67-570 67-548 67-551 Ni Bal Bal Bal Bal Bal Bal Mo 11.9 12.0 11.7 11.7 12.4 12.2 Cr 1.05 7.03 7.02 7.04 . 7.09 1.02 Mn 0.11 0.12 0.13 10.02 0.13 0.02 Fe 0.02 0.02 0.03 0.02 0.13 0.02 C 0.094 0.028 0.08 0.028 0.09 0.028 Ti 0.64 - 0.64 0.93 0.92 1.2 1.1 Zr 0.002 0.005 0.002 0.003 0.002 0.0005 Al 0.07 - 0.07 <0.05 0.07 0.08 <0.05 Si 0.01 0.01 0.02 0.02 0.03 0.02 B 0.002 0.0007 0.0002 0.0002 0.00002 0.0002 ZAlloy heat number. W‘_...mnw‘.._._., 229 18.8, show clearly a large influence of the higher carbon concentration on both the rupture life and the creep ductility at each stress level and at both test tempera- tures. Thus for irradiations at $50°C the creep rupture lives and ductilities of titanium-modified alloys are enhanced by higher carbon contents. These results again ¥ 18.3 EFFECTS OF IRRADIATION AT 760°C ON THE CREEP-RUPTURE PROPERTIES H. E. McCoy 'OF MODIFIED HASTELLOY N ~ C.E. Sessions | Several small laboratory melts of modified Hastelloy emphasize the complexity of ‘the role of alloying N have been irradiated in an effort to optimize the additions on the high-temperature irradiation damage. - chemical composition. Table 18.2 lists some results for VTablVe 18.2. Postirradiation Creep-Rupture Properties of New Modifications of the Ni—12% Mo—7% Cr—4% Fe—0.2% Mn—0.06%C Base Alloys® Postirradiation Creep Properties Alloy Alloy Additions (wt %) Stress Rupture Fracture Mg;ier:um Number Ti . Nb Zr . Hf sib o8 Life Strain P (psh (b @ nate (%/hr) ' x10° 284 0 0 0 0 15 1547 0.2 0.0008 : - 10 355.4 0.6 0.0009 286 0.2 0.5 005 0 - 40 9.4 0.9 0.070 ' 35 7.2 0.5 0.037 289 1.0 0 0. 0 30 350 4.56 0.011 . 40 11.2 1.4 0.095 290 105 0 0 0 30 648.6 5.3 0.003 | 40 20.8 2.0 0.067 1291 20 0 0 0 40 388.3 4.7 0.008 : 47 168.6 6.8 0.027 292 24 0 o - 0 40 - >1100 1.4¢ 0.0006 293 30 o0 0 0 | 47 1513.8 1.0 300 08 13 004 07 016 47 141.0 6.1 0.038 301 0 0 006 08 40 339.6 10.8 0.021 1302 o 0 0.05 0.5 40 226 9.4 0.032 303 0s . 08 0 0o 40 271 .63 0.019 , o ) | 55 8.8 5.0 0.442 315 0.5 0 005 0 40 48,1 4.4 0.081 - | L 35 158.8 4.7 0.026 181.. 0.5 1.8 0 - 0 40 175.3 4.25 0.02 - , 35 666.2 428 0.008 o S . 27 15804 359 - - 0.002 184 - 12 0 0 1.2 0.2 . 47 235 13.7 0.15 | o | o 35 2292 22.7 0.005 22 0. 0 0 1202 40 ~ 405.5 13.3 0.022 o - 27 2035.0¢ - 6.9 10.002 "Alloys annealed 1 hr at 1177°C, irradiated at 760°C to 3 X 102° neutrons/cm and then creep tested at 650°C bsilicon content of <0.01% unless specified otherwise. CStress raised to 47,000 psi and sample failed in 10 hr. dTest discontinued before failure. 230 ORNL-DWG 70-344R (x103) 1 60 ¢ 303 (5.0) _ 50 | 18437 '/,291(6.8) | ' ~d | | 291(e8)e ¢ | 300 (6.0)@ o 3(')3'(!5'3!)' 293 (1.0) - » +Hl 289{1.4) 302 (9.4) / '291(4.7) - 315(4.4) |-301{10.8) = 40 286(0.9) o 5.2 | I! _otesl oo & 117 , i81a.2) | || [|[202t1-4 g 286 (05)p : 1] _-® | 181(4.3) e 184(22.7 W 290207 | 315 (4.7 289045 £ 30 : - l l i \Mzeots.s STANDARD HASTELLOY N__L>>{] IRRADIATED ' S . S~ 20 : 284(0.2) e 10 284 (0.6) 0 : 102 103 - 104 100 10! RUPTURE LIFE (hr) Fig. 18.10. Postirradiation Stress-Rupture Properties of Several Alloys of Modified Hastelloy N at 650°C. All materials were annealed for 1 hr at 1177°C and irradiated at 760°C to a thermal fluence of 3 X 10%° neutronslcmz. All alloys have a base composition of Ni—12% Mo~7% Cr—0.06% C, and the carbide-forming additions to each alloy are shown in Table 18.2. The numbers in parentheses indicate the fracture strain. new alloys that were recently tested in postirradiation ‘creep rupture. Each alloy has the base composition Ni—12% Mo—7% Cr—4% Fe—0.2% Mn—0.06% C, with various minor additions of Ti, Nb, Zr, Hf, or Si as listed in Table 18.2. We can evaluate the relative benefit of a given combination of alloying elements by comparing the creep properties after irradiation. Figure 18.10 compares the results of stress-rupture testing for these new alloys with those for standard irradiated Hastelloy N. These data were obtained at 650°C after irradiation at 760°C to a thermal neutron “fluence of 3 X 10?% neutrons/cm?. These particular test conditions have been shown to reflect the sensi- tivity of postirradiation properties to alloy content.! As expected from previous results, alloy 284, which contains no Ti, Nb, Zr, or Hf, had inferior properties compared with standard Hastelloy N. Alloy 286, containing additions of 0.5% niobium + 0.2% titanium, was no better than standard Hastelloy N. However, 1Y, E. McCoy et al., MSR Program Semiann, Progr, Rept, Aug. 31, 1969, ORNL-4449, pp. 184-92, - altoy 301 (0.8% hafnium) and alloy 315 (0.5% titanium + 0.05% zirconium) both appear to be considerably better than the standard alloy in terms of rupture life and fracture strain. . _ Alloys 289 and 290 both contain about 1.0% tita- nium, but the properties were no better than standard _ Hastelloy N. As discussed previously! the specimens with approximately 1.0% titanium showed a large variation in creep properties after irradiation at 760°C that we do not yet fully understand. However, these two alloys at 30,000 psi stress had high postirradiation ductility, For increasing alloy content, alloy 232 (1.2% hafnium) is particularly good, with a 13% fracture strain in a creep test at 40,000 psi. Alloy 303 is an outstanding alloy composition for additions below ' 1.5%. This alloy contained 0.5% titanium and 0.8% niobium and was tested at 47,000 psi. The postirradia- tion rupture life was a factor of 100 higher than that for standard Hastelloy N. Even at 55,000 psi the ductility and rupture life were excellent for this alloy. Alloys 181 (1.8% niebjum + 0.5% titanium), 184 (1.2% titanium + 1.2% hafnium), 291 (2.0% titanium), 231 and 300 (0.8% Ti—1.3% Nb—0.7% Hf) had excellent - postirradiation creep properties. Of these four alloys tested to date, alloy 184 is far superior, with 23% fracture strain for creep at 35,000 psi. These recent results on alloys with 2% titanium are also quite encouraging and indicate that titanium additions alone can probably produce the desired mechanical behavior during service at 700°C. ‘Alloys with higher concentrations of titanium of 2.4 and 3.0% (alloys 292 and 293) exhibited very high creep resistance at this test temperature. The very low creep rates and low rupture ductilities of these two alloys probably indicate that these compositions pre- cipitate gamma prime (Ni; Ti) type intermetallic com- pounds during irradiation at 760°C. If this proves to be . true, we would not be interested in alloys with such high titanium concentrations, 18.4 ELECTRON MICROSCOPY OF MODIFIED HASTELLOY N ALLOYS | R.E.Gehlbach S.W. Cook Electron microscopy studies of Hastelloy N have concentrated primarily on characterizing the effects of alloying additions on the microstructure of several modifications of the alloy after exposure at elevated temperatures, namely, 650 and 760°C. The resistance to radiation damage is altered markedly by the addition of small amounts of Ti, Nb, and Hf. Silicon is also important in determining the type of precipitate formed. The concentration of this element is controlled largely by the melting practice, being high (0.5%) for -alloys prepared by electroslag remelting! or air melting and low (<0.1%) for vacuum-melted alloys. The molyb- denum concentratlon in the base alloy is also nnpor- _ tant. Laboratory Heats. — A number of 2-Ib laboratory melts were prepared to evaluate the influence of Ti, Nb, Hf, and Si on precipitation and the resistance to irradiation damage. Several of these alloys have been aged at 650 and 760°C. Precipitates were extracted electrolytically and identified by x-ray diffraction. The alloys with their compositions and the pl‘eClpltateS detected are listed in Table 18.3. ‘The effects of titanium additions on precipitation have been discussed previously.?*® Precipitation of MC carbides® often occurs in a stacking fault morphology at -titanium concentrations up to the 1.2% level. Examination of an alloy containing 2.4% titanium (292) that had been aged 200 hr at 760°C revealed the absence of this morphology, although much MC (g = 4.29 A) was present as films and very fine particles in the .jagged grain boundaries and as particles in the - matrix. Gamma prime (Ni;Ti) was not detected in this material. H. E. McCoy et 4l., MSR Program Semiann, Progr. Rept. ~ Aug 31, 1969, ORNL-4449, pp. 186-92, 2R. E. Gehlbach and S. W. Cook, MSR Program Semiann, Progr Rept. Feb, 28, 1969, ORNL-4396, pp. 240—-42. ' 3R. E. Gehlbach, C, E, Sessions, and S. W, Cook, MSR Pro- gram Semiann, Progr. Rept. Aug 31 1969, ORNL-4449, PP 193-95, 4The “M” designates metallic atogs in the carbides, Table 18.3. Alioy Compositions and Precipitate Types Observed in Several Heats of Modified Hastelloy N . _ o . : . b Alloy _ Composition? (%) o Aging Temperature _ B : o L o - MC M,C MgC ‘MC M,C M¢C 292 24 <0.005 <0.005 . 0.02 . X 285 008 0.5 <0.005 . 0.08 x¢ : X 306 0.01 0.55 - 0.002 0.27 x4 x x 286 023 05 <0.005 0.02 x® X x€ X 287 0.12 0.6 <0.005 0.14 X X X x¢ X X 1310 015 0.57 0.54 <0.01 X 313 <0.02 <001 07 . 0.02 X X 314 0.65 1.3 065 0.35 X X 9AM alloys have a nominal composition of Ni-12% Mo—7% 1—4% Fe-0.2% Mn-0.06% C. : ‘bAged 1000 hr at the indicated temperature except alloys 292, 310, 313, and 314, which were aged 200 hr at 760 C. . "Plus unidentified phase. 9Very fine unidentified precipitate observed by electron microscopy. €Stacking fault morphology observed by electron microscopy. 232 { YE-A0001 Fig. 18.11a. Microstructure of Hastelloy N Modified with 0.5% Nb. 10,000X. The effects of small amounts of niobium, of titanium and niobium, and of titanium, niobium, and silicon on the microstructure after aging at 760°C are shown in Fig. 18.11. The addition of 0.5% Nb alone (alloy 285) resulted in relatively coarse M, C (Fig. 18.11a), which is typical of the base alloy. No stacking fault precipitate was observed in these alloys. A combination of 0.5% niobium and 0.23% titanium (alloy 286) resulted in a finer grain boundary precipitate (Fig. 18.115b) than that observed in the alloy without titanium (285). The ‘combination of 0.14% Si, 0.12% Ti, and 0.6% Nb (alloy 287) resulted in the formation of large blocky M¢C-type carbides in both the grain boundaries and the matrix, as shown in Fig. 18.11¢. These silicon-rich N 233 Fig. 18.11b, Microstructure of l;lasiefloy N Modified with 0.5% Nb and 0.23% Ti. 10,000X. carbides are very stable and are not dissolved during annealing as described previously.® After the silicon is consumed in the MgC, which contains about 3.3% silicon, further grain boundary carbide precipitation is similar to that in the silicon-free alloy (286), although smaller amounts are present. Relatively small carbides *R. E. Gehlbach, MSR Program Semiann. Progr. Rept. Feb. 29, 1968, ORNL-4254, pp. 20613, - ' also precipitated in the matrix in alloy 287. Some precipitate occurred in the stacking fault morphology in alloys 286 and 287, containing titanium, but not in 285 or 306, which do not contain any titanium, There is some question regarding the structure of the stacking fault precipitate observed in the alloys containing niobium. Aging at 650°C resulted in much finer M, C than that generated at the higher temperature (760°C). In addition to large quantities of M4C, a small amount i | ! | 4 (¢ 234 B YE-10004 | Fig. 18.11c. Microstructure of Hastelloy N Modified with 0.5% Nb, 0.12% Ti, and 0.14% Si After Aging 1000 hr at 760°C. 10,000X. of very fine MC was dispersed throughout the matrix in 306 after aging at 650°C; however, only M;C is present after aging at 760°C. ' The MC formed in hafnium-modified alloys has a morphology different from that in the titanium- modified alloys. Figure 18.12z is typical of an alloy containing 0.7% hafnium (alloy 313) aged 200 hr at 760°C. As shown in Fig. 18.12b, 0.15% titanium and 0.57% niobium in combination with hafnium (alloy 310) did not significantly affect the morphology of the carbides formed at 760°C. A comparison of Fig. 18.12b with Fig. 18.11b shows the effect of hafnium at comparable levels of titanium and niobium. A similar microstructure was previously shown for an alloy containing 1% hafnium and 1% titanium (alloy 184).2 The presence of 0.35% silicon in an alloy containing v 235 Fig. 18.12a. Microstructure of Hafnium-Modified Hastelloy N After Aging 200 hr at 760°C. 0.7% HE. 10,000X. 0.65% Ti, 1.3% Nb, and 0.65% Hf (alloy 314) resulted in large stable Mg C in the matrix and grain boundaries with subsequent precipitation of fine MC (Fig 18.12¢) similar to the silicon-free alloys. Due to the high level of silicon in 314, much of the carbon was tied up in the M,C, with little available for MC precipitation during aging. No stacking fault morphology was found in the hafnium-modified alloys. ' - Commercial Alloys. — We have examined the phases present in four new commercial alloys (Table 18.4). Two are electroslag remelted heats containing 0.77% titanium and 1,7% niobium (heat 69-344) and 1.1% titanium and 0.92% hafnium (heat 69-345). Both contained high concentrations of silicon and had large amounts of MgC after annealing. Small quantities of MC were also present in both alloys. ()8 236 YE-1000 Fig. 18.12b. Microstructure of Hafnium-Modified Hastelloy N After Aging 200 hr at 760°C. 0.54% Hf, 0.57% Nb, and 0.15% Ti. 10,000X. Precipitates in the two vacuum-melted heats were examined after aging at 650°C. One contained 0.92% titanium and 2.0% niobium (heat 69-648) and the other 1.3% titanium and 0.60% hafnium (heat 69-641). Both have only trace levels of silicon, and no MyC was found. Two MC carbides are present in 69-648 and three in 69-641. Molybdenum Series. — We have identified the pre- cipitates formed in Hastelloy N as a function of the molybdenum concentration. These alloys have silicon levels of about 0.01%. Only M,C precipitated at 650°C for molybdenum levels of 10 through 16%. At 19% molybdenum, the only precipitate formed was a high- parameter MgC (@ = 11.23 A), probably Ni;Mo,C. " 237 Fig 18. 12c. Microstructure of Hafmum-Modlfied Hastel!oy N After Agmg 200 hr at 760°C. 0.65% Hf, 1.3% Nb, 0.8%Ti, and 0.35% Si. 10, OOOX Discussion. — Our observations on the effects of alloy modifications on the microstructure of the heats described above are consistent with our previous obser- vations. The formation of particular types of precipi- tates is dependent on the alloying additions. The M,C-type carbide is stable in the basic alloy, with additions of Ti, Nb, and Hf in sufficient quantities giving MC-type carbides. The M,C is coarser at 760°C than at 650°C; however, it is the higher temperature which is of primary interest to us, since the outlet fuel salt temperature of an MSBR will be about 700°C. The postirradiation mechanical properties deteriorate mark- edly in ‘materials which form coarse M,C at the irradiation temperature. 238 Table 18.4. Precipitates in Commercial Modified Hastelloy N spe . d Alloy Composition? (%) Phases? Number Zr Ti Nb Hf Si Fe 69-344¢ 0.001 0.77 1.7 <0.01 030 40 MgC@O)+MCW) 69-345¢ 0.3 1.1 <0.01 092 025 40 MgC(S)+MC(W) 69-6489 <0.05 0.92 2.0 <0.05 005 03 MC 69-6419 <0.05 1.3 <0.05 0.60 <001 03 MC 2 All alloys have a nominal composmon of Nn—— 12% Mo—7% Cr—0.2% Mn-0.05% C bg = strong, W = weak, CAnnealed 1 hr at 1177°C, dAnnealed 1 hr at 1177°C plus 1000 hr at 650°C. The beneficial effect of MC carbides on the postirradi- ation mechanical properties appears to depend on the morphology of the precipitate, or at least on the alloying addition used. Titanium alone in sufficient quantities improves the postirradiation creep rupture life, but the fracture ductility may be too low at high ~creep stresses. The most obvious microstructural modi- fication in the MC-forming titanium-modified alloys is the formation of massive stacking fault precipitates. The niobium-modified heats discussed do not contain sufficient quantities of niobium to stabilize MC. Based on microstructural observations, we would not expect these alloys to have good properties after irradiation. Higher concentrations of niobium would be expected to exhibit good properties if MC is formed in a desirable morphology. We will be studying several alloys with larger niobium additions, particularly alloy 303, which looks very promising (see Table 18 2). “Additions of hafnium appear to control the type of microstructure even with large quantities of titanium and niobium in the alloy. These alloys are characterized by small discrete particles in the matrix and grain boundaries, and the postirradiation mechanical proper- ties are outstanding. ~ The effect of small quantities of silicon is to cause the _ formation of a silicon-rich stable McC that is very coarse. This is undesirable in that it results in a grain boundary structure approaching that of standard Has- telloy N rather than permitting the generation of a controlled microstructure by judicious alloying and ~ heat treatment. Thus the electroslag remelt process is virtually eliminated from further consideration as a production method for an improved Hastelloy N. However, small amounts of silicon can likely be tolerated, particularly when hafnium is present. 18.5 WELDABILITY OF COMMERCIAL ALLOYS B.McNabb H.E.McCoy Two of the prime requirements for an alloy to be used in a nuclear reactor are fabricability and weld- ability. In scaling up small lab melts to commercial-size melts, these properties must be evaluated and experi- ence gained through commercial vendors for promising alloys. Because of financial restraints our commercial experience must be limited to small 50- and 100-1b melts from commercial vendors. Evaluation of weldability was carried out by making welds under highly restrained conditions and then using side bend tests. The weld side bend tests were made according to ASME Boiler and Pressure Vessel Code Sect. 9; they used four side bend specimens % in. thick bent 180° around a % -in, radius. Tensile specimens will be made from the welds to further evaluate the weld metal properties. The four 501b heats from Allvac Metals Company (69-641, 69-648, 69-714, and 70-727) were received as 1, -in.-thick plates 4 in. wide by 10 in. long in the rolling direction. Stnps 1, in. square were sawed from the plates in the rolling direction and swaged to % and %, in. weld wire. The plates were beveled 50° each, giving a 100° included angle weld, with a % 4-in. land at . the root pass. The weld direction was parallel to the rolling direction in order to get the maximum weld length, The two 100b heats from the Materials Systems Division of Union Carbide Corporation (69-344, 69-345) were % X 9 X 10 in, in the rolling direction. These plates were prepared as above for the Allvac heats, except the weld direction was made perpendicular to the rolling direction so that any stringers in the base metal would be aligned axially in 239 Table 18.5. Vendor’s Analysis of Modified Hastelloy N Heats Method? S ' _ Relative Alloy of Ni Mo Cr Fe Mn C Si Ti Nb Hf Zr v Welding Melting Rating? 69-641 A Bal 139 69 0.3 035 0073 500 ppm), we have found that elements other than chro- - mium may be oxidized by the salt, and as a result the lw. D. | Manly et al, “Metallurgical Problems in Molten Fluoride Systems,” Progr. Nucl. Energy, Ser. IV 2, 164-79 (1960). corrosion rate is higher and the attack is more uniform. These cases will be pointed out in the discussion of the individual systems. The status of the thermal-convec- tion loops in operation with fluoride salt is summanzed _in Table 18.7. 18.6.1 Fuel Salts ' Loop 1255, constructed of Hastelloy N and con- taining a simulated MSRE fuel salt plus 1 mole % ThF,, continues to operate without difficulty after 7.9 years. - .Loop 1258, constructed of type 304L stainless steel and containing removable insert specimens in the hot leg, has operated about 6.6 years with the same salt as loop 1255. A plot of the weight change of the specimens in the hot leg as a function of operating time at various temperatures is given in Fig. 18.13. The corrosion rate at the highest temperature; 688°C, assuming uniform wall removal, has averaged 1.1 mils/year and is continually decreasing. Even though the weight losses are relatively large, there has been no indication of plugging. Because of interest in iron-base alloys W1th lower chromium contents for possible containment of molten salts, a maraging steel (12 Ni—5 Cr—3 Mo~bal Fe) specimen has been exposed to salt in loop 1258. Table - '18.8 gives a comparison of the weight losses and corrosion rates of several alloys, including the maraging steel, under similar exposure conditions. As expected, because of the lower chromium content, the maraging steel shows better corrosion resistance than the stainless steels, but the corrosion behavior of the Hastelloy N is still superior to both. The test is contmumg, and other ‘iron-base alloys will be tested. Loop NCL-16, constructed of standard Hastelloy N with removable specimens in each leg, has operated with the two-fluid MSBR fuel salt for over 17,800 hr. The corrosion rate at the highest temperature, 704°C, in NCL-16 assuming uniform attack is 0.04 mil/year. For this system, as with all others studied to date, titanium-modified Hastelloy N specimens continue to have smaller weight changes than standard Hastelloy N specimens under equivalent conditions. This has been Table 18.6. Composition of Hastelloy N Chemical Content (wt %) Allo ‘ y Ni Mo - Cr " Fe Si - an Ti Standard Hastelloy N Bal 17.2 Titanium-modified Bal 13.6 Hastelloy N 7.4 45 06 0.54 0.02 73 <01 <001 014 05 & 241 Table 18.7. MSR Program Natural Circulation Loop Operation Through February 28, 1970 - Loop _ ~ Salt Maximum Operating Number Loop Material Specimens Salt Type Composition Tem%eratm'e Co Time . ‘ (mole %) (hr) . 1255 Hastelloy N Hastelloy N + 2% Nb%D Fuel LiF-BeF,-Z1F4-UF4-ThF4 704 90 69,300 (70-23-5-1-1) : - _ 1258 Type 304L SS Type 304L stainless steel®¢ Fuel LiF-BeF,-ZrF4-UF4-ThF, 688 100 58,000 , (70-23-5-1-1) NCL-13A Hastelloy N Hastelloy N; Tn-modlficd Coolant NaBF4-NaF (92-8) 607 125 11,800 Hastelloy N controls®:9 _ ' NCL-14 Hastelloy N Ti-modified Hastelloy NS4 Coolant - NaBF4-NaF (92-8) 607 150 20,300 NCL-15A Hastelloy N Ti-modified Hastelloy N; Blanket LiF-BeF,-ThF4 (73-2-25) 677 55 13.400 ' Hastelloy N controls®4 : NCL-16 Hastelloy N Ti-modified Hastelloy N; Fuel ~ LiF-BeF,-UF,4 (65.5-34.0- 704 170 17,800 : Hastelloy N controls &9 0.5) NCL-17 Hastelloy N Hastelloy N; Ti-modified Coolant NaBF4-NaF (92-8) plus 607 150 5,900 ' Hastelloy N controls&9 - ~ steam additions NCL-18 Hastelloy N Ti-modified Hastelloy N; Fertile- LiF-BeF,-ThF4-UF, 704 170 11,600 ~ Hastelloy N controls&9 Fissile (68-20-11.7-0.3) ‘ NCL-19A Hastelloy N Hastelloy N; Ti-modified Fertile- LiF-BeF,-ThF4-UF4 (68- 704 170 100 . Hastelloy N controls®9 Fissile 20-11.7-0.3) plus | bismuth in molybdenum hot finger NCL-20 Hastelloy N Hastelloy N; Ti-modified Coolant NaBF4-NaF (92-8) 687 250 1,700 . Hastelloy N controls®9 : %Permanent specimens. . bHot leg only.. , “Removable specimens. ¥ dHot and cold legs. 0 ORNL-DWG 68-60878R 8 { CIMEN REPLACED : WEIGHT CHANGES CONTINUED WEIGHT LOSS (mg/em?) o o H o 60 12 LOOP 1258 ® EQUIVALENT TO 1mil/year O EQUIVALENT TO 1.5 mil/year 14 16 SPECIMEN TIME IN SYSTEM (1000 hr ) 22 24 - Fig, 18.13. Weight Loss of Type 304L Stainless Stee! Specmens as a Function of Operatlon Time and Temperature in LiF-BeF,-Z1F4-ThF;-UF, (70-23-5-1-1 Mole %) Salt. Table 18.8. Gompanson of Wexght Losses of Alloys at~663°C After ~2490 hr in Similar Flowing Fuel Salts in 2 Temperature Gradient System Salt mixture: LiF -Ber-ZrF4-UF4-ThF4 (70-23-5-1-1 mole %) | Allo | Weight Loss Average Corrosion , y (mg/em®) Rate (mils/year) Maraging steel 3.0 0.53 Type 304 stainless steel 6.5 1.1 Hastelloy N 0.4 0.06 'generally attributet_l to the absence of iron in the- modified alloys. The chromium content of the salt, 242 “glazed” with a coating that is impossible to remove without damaging the metal. This has made weight change measurements difficult; however, there has been little change in the chromium concentration of the salt. 18.6.4 Coolant Salt - Loop N_CL-13A,.constructed of standard Hastelloy'N - with removable specimens in each leg, has operated for 11,800 hr with the fluoroborate salt proposed as the MSBR secondary coolant. Figure 18.14 gives the weight changes of specimens at various temperatures as a currently 400 ppm, continues to show a small increase - with time, while the iron content has apparently - stabijlized. Earlier the FeF, content in the salt had decreased, suggesting that the reaction FeF, (d) + Cr(s) = CrF, (d) + Fe (deposited) , where d = dissolved in salt, s =solid solution, was responsible for part of the mass transfer. Since no changes in FeF, are now seen, the UF, corrosion reaction must now be the primary cause for the chromium removal. The overall process is still con- trolled by solid-state diffusion of chromium in the alloy. ' : 18.6.2 FertileFissile Salt Loop ‘NCL-18, constructed of standard Hastelloy N with removable specimens in each leg, has operated . with the single-fluid MSBR salt for over 11,600 hr. The weight change at the highest temperature, 704°C, is —1.1 mg/cm?. Assuming uniform attack, the corrosion - rate is 0.05 mil/year. The titanium-modified Hastelloy N specimens show smaller weight losses than standard specimens in equivalent positions, and, as'in the case of function of operating time. This figure is typical of the plots of total weight change of the specimens vs time that we obtain from all the natural-circulation loops. It is clear that material is removed in the hot sections and deposited in the cold sections. It is also obvious that these changes are temperature dependent and that the corrosion and deposition rates decrease with time. The corrosion rate at the maximum temperature, 605°C, is 0.6 mil/year. The attack in this case may involve other alloy constituents besides chromium, since analyses of ~ the salt show approximately 1000 ppm water and 1000 ppm oxygen. Large amounts of these impurities portend high corrosion rates due to the formation of strong oxidants such as HF. The chromium concentra-’ tion of the salt initially was 253 ppm and has remained near that value throughout operation. Loop NCL-14, constructed of standard Hastelloy N with removable specimens in each leg, has operated for 20,300 hr with the fluoroborate coolant salt. The weight changes of the specimens at the various tempera- tures as a function of operating time are given in Fig. 18.15. Two changes of corrosion rate are noted in the plot. These changes were traced to a defective gas line and a leaking standpipe ball valve, and their effects have ‘been discussed.2:3 The overall corrosion rate at the ‘maximum temperature, 605°C, is 0.55 mil/year. Again, NCL-16, the chromium content of the salt is increasing - (150 ppm at the present), while the iron content is stable. ‘ 18.6.3 Blanket Salt Lbop NCL-15A, constructed of standard Hastelloy N with removable specimens in each leg, has operated 113,400 hr with the LiF-BeF, salt containing 25 mole % Specimens exposed to this salt are often . ThF,. because of impurity-controlled attack, the corrosion is generally uniform. It appears that under extreme oxidizing conditions when all constituents of the alloy are removed, the nickel and molybdenum that are removed as fluorides deposit as metals in the cold portion of the loop. Larger amounts of chromium and iron fluorides remain in the salt. After the extreme oxidizing conditions pass, the iron fluorides oxidize ‘more chromium from the alloy. It is encouraging that, 25, W. Koger and A. P. Litman, MSR Program Semiann. Progr. Rept. Feb, 28, 1969, ORNL4396, p. 246. , 3. W. Koger and A. P. Litman, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 200. 243 ORNL-DWG €9-4763AR WEIGHT CHANGE (mg/cm?) 0 1000 2000 ‘ 3000 4000 5000 6000 7000 ~ 8000 9000 - TIME OF OPERATION (hr) Flg 18.14. Welght Change vs Time for Standard Hastelloy N Specimens in NCL-13A Exposed to Fluoroborate Salt at Various * Temperatures. 10 ORNL-DWG 69—12625R 465°C . # 470 8 =T 0480 5 . . _/ 0485 - e () e ‘ =1, . —_ 050 ) N s S A & © ‘BE. EO.:E:'« e . . - o O, § =R~ =] ‘E \\ . \0-,____- 'P\.—__ 40 w =5 * i R b @ ~ i g ~— \° . & o .‘\ \o . = -10 . : ) \ .-'\- \___‘?65 35 . o ~ - Y _ \o N\ _ . , \\ o\ ._ -‘5 D‘-l - * ‘ . \-o ) \ - | \\ N 580 . - -20 : _ N - B \ | ’ ' | ' ' ‘ \ o - - -25 : _ 4 508 * _ : 0 2000 4000 . 6000 8000 10,000 = 12,000 4,000 16,000 18,000 TlME OF OPERATION {hr} Fig. 18. 15. Weight Change vs Time for Titanium-Modified Hastelloy N Specimens in NCL-14 Exposed to Fluoroborate Salt at \ / Various Temperatures. I | i i | { | i i 30 , : 20 — STEAM INJECTION , _ { .,___.-—-——-I527-493°c «-5 10 1 S o 544°C £ | ~ 0 § * 560°C = ———— 571°C S -0 o & * 582°C = -20 , '--..:‘-\-v 593°C _ * 607°C -30 -40 in spite of two known problems which increased the corrosion rate, loop NCL-14 has operated for over two years with the fluoroborate salt and that the incre- mental corrosion rate is currently less than 0.5 mil/year. Loop NCL-17, constructed of standard Hastelloy N with removable specimens in each leg, is being operated to determine the effect of steam inleakage in a flowing fluoroborate salt—Hastelloy N system. This experiment has run for over 5900 hr and is continuing in order to provide information on the immediate and long-range corrosion of the system after steam injection. The loop was operated for 1000 hr, the specimens removed and weighed, and steam forced into the flowing salt system through a 16-mil hole in a closed Y -n. Hastelloy N tube, simulating a leaking heat exchanger. Steam was forced into the system until the pressure began to increase, thus indicating that no more steam was soluble in the salt. - | _ Table 189 summarizes the results to date. The corrosion rate continues to decrease with time, and the incremental .rates have now fallen to 1 mil/year. Figure 18.16 shows the weight changes and temperatures for “the specimens in NCL-17. As usual, the changes are temperature dependent, and the rates are decreasing with time. Figure 18.17 shows micrographs of the specimens in the hottest and coldest positions. The ORNL~-DWG 70~-4936 - 0 1000 2000 3000 4000 5000 6000 TIME OF OPERATION (hr) Fig. 18.16, Weight Change vs Time for Hastelloy N Specimens in NCL-17 Exposed to NaBF4-NaF (92-8 Mole %) at Various Temperatures. attack seen in Fig, 18.17a is general, as expected for impurity-controlled mass transfer processes. Micrometer measurements show a loss of approximately 1 mil from specimen surfaces at the maximum temperature posi- tion. In Fig. 18.17b we see a large amount (~2 mils) of deposited material. This material has been analyzed using an electron beam microprobe, and Table 18.10 gives the results for the matrix and the deposit. These results show that very little chromium has deposited. The nickel and molybdenun concentrations had reached ~a maximum of 74 and 56 ppm, respectively, 288 hr - after the steam addition, decreased, and quickly leveled off to about 10 ppm each. From the microprobe ~evidence much of the nickel and molybdenum ap- parently has deposited in the cold leg, although not in the same ratio as they exist in Hastelloy N. A similar ~ deposition was also found and discussed for NCL-14. One may conclude that under highly oxidizing condi- tions where large weight changes occur (NCL-14 during the air leak and NCL-17 after steam inleakage), most of Table 18.9. Weight Changes of Hottest Specimen (607°C) from Steam-Injected Fluoroborate Salt Loop NCL-17 Specimen Exposure Airerag’ Overall . - e —flhfl—— Weight Ch%““ " Corrosion Rate Before After (mg/cm”) (mils/year) Steam Steam 1054 ‘ -0.5 0.2 239 -12.0 , 19.5 424 - -15.2 13.9 663 - =171 10.0 1474 - =220 58 2888 -26.3 3.5 Table 18.10. hficroprobeoAnalysis of Specimen in Coldest Position (493 C) in Loop NCL-17 Exposed to fluoroborate salt for 4000 hr; steam injected into salt 1000 hr after . beginning of run Composition? (wt %) - Element — Matrix Deposit Ni - 73.7 348 Mo o 14.3 30.6 Cr 6.8 <0.5 Fe ‘ 38 2.9 P = 3.0 Undetermined ~30 Corrected for absorption, sccondary fluorescence, and atomic number effects. 7 bThin layer close to sample surface. 245 l v- 98296 10.00% in. '0.003 in. 0.007 INCHES 500X 10.005 in. 10.007 in. e 1 10.001 in. 10.003 in. 0,007 INCHES 500X 10.005 i [ 10.007 in. . - Fig. 18.17. Microstructurc of Standard Hastelloy N Exposed to NaBF,-NaF (92-8 Mole %) in NCL-17 for 3942 hr, Steam injected into salt after 1054 hr. (a) 607°C, weight loss ~26.8 mg/cm?, as polished; (b) 495°C, weight gain +14.8 mg/cm?, as polished, ~ j There is a relatively spongy surface layer on this sample about 2 mils thick. the changes are attributable to the movement of the normally stable nickel and molybdenum. Any increased chromium removal from the hot leg is noted just as an increase in the chromium concentration in the salt. The ‘phosphorus found in the deposit was apparently from a phosphate impurity in the steam. For the last 2500 hr the concentrations of impurity constituents in the salt,. Fe, Cr, Ni, Mo, H, 0, and O, , have remained fairly con- stant. The present chromium level is 320 ppm. The con- tinuing results of this test are quite significant in that ~ we have shown that the fluoroborate salt and the Hastel- loy N can withstand an accidental steam inleakage con- dition and that a system consisting of these components could continue to operate without extensive damage even if the salt were not repurified. Loop NCL-20, constructed of standard Hastelloy N ‘with removable specimens in each leg, has circulated a fluoroborate coolant salt for 1700 hr and is being operated at temperature conditions very near those - proposed for the maximum (687°C) and minimum (387°C) salt-metal temperature (primary heat ex- changer and steam generator respectively) of the MSBR secondary circuit. Forced air cooling (as opposed to - ambient air cooling used on other thermal-convection loops) is used on the lower half of the cold leg, and the- practical operating temperatures obtained were 687°C maximum and ~438°C minimum: a AT of 250°C. This AT is thought to be the largest obtained at ORNL in a molten-salt thermal-convection loop, and the maximum temperature is the highest for fluoroborate salt in a - loop. The weight change of the specimen at the highest temperature was —0.3 mg/cm? (0.15 mil /year assuming uniform attack) after 700 hr. Salt analyses showed an -increase in chromium content of 30 ppm and an H,0 content of 550 ppm. 'Mention was made in the last semiannual mport4 of capsule experiments designed to determine the uptake of chromium from standard Hastelloy N into relatively pure NaBF;-NaF (92-8 mole %) containing approxi- mately 400. ppm each oxygen and water at 427°C (800°F), 538°C (1000°F), 649°C (1200°F), and 760°C (1400°F). Tests were conducted for 1200 hr, after which the salt was analyzed for chromium and other impurities. Figure 18.18 shows the logarithmic variation of chromium content as a function of the reciprocal of the absolute test temperature. The iron concentration was originally about 200 ppm and decreased by an amount proportional to the increase in chromium. The final iron content of the salt at the highest temperature (760°C) was 60 ppm. An Arrhenius-type relationship appears to hold be- tween 538 and 760°C. This would be expected if the - 246 . ORNL-DWG 69— 12236A TEMPERATURE (°C) '103 800 700 650 600 550 500 450 400 10 Cr CONCENTRATION (ppm} 10' 090 40 1 12 13 14 15 10%%r w1 Fig. 18.18. Temperature Dependence of Chromium Concen- tration in Sodium Fluoroborate Salt Exposed for 1200 hr in Hastelloy N Capsules. The dashed line is computed from the measured diffusion rate of chromium in Hastelloy N and assumes that the surface concentration of chromium is zero. rate of ch_romium buildup in the salt were controlled by - solidstate diffusion of chromium to the capsule wall. ~ Using diffusion data for chromium in Hastelloy N obtained by Evans, DeVan, and Watson® and assuming the chromium surface concentration to have been reduced to zero by the salt, we obtain the predicted chromium buildup shown by the dashed line in Fig. 18.18. The close agreement between the slopes (activa- tion energy) of the predicted and experimental curves gives credence to the diffusion-controlled assumption. - The observed behavior of chromium and the in- significant changes in iron and nickel concentrations in the salt indicate that when the H, 0 and oxygen con- centrations of the fluoroborate are at a level as low as 400 ppm, the corrosion mechamsm is the selective re- moval of chromlum 4. W. Koger and A. P. Litman, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 201. SR. B. Evans, J. H. DeVan, and G. M. Watson, Self-Diffusion of Chromium in Nzckel Base AIons 0RNL-2982 (Jan. 20, 1961). o et kAt 18.6.5 Summary of Corrosion Studies This section will act as a summary of the current developments involving the compatibility of fused salts and Hastelloy N. Table 18.11 compares the corrosion rates of standard Hastelloy N in tests of 5000 hr or longer involving several fluoride salts. Except for certain of the fluoroborate tests the overall corrosion rates have been relatively small. Thermal-convection loop NCL-14 has operated success- fully for over two years and the pump loop MSR-FCL-1 247 nearly one year with the fluoroborate mixture. Compar- ' ison of the rates experienced by NCL-13A, NCL-14, and MSR-FCL-1, all circulating the sodium fluoroborate mixture, indicates a velocity effect on the mass transfer. We believe that the effect of velocity on corrosion in this system is a function of declining importance as the purity level of the salt improves. The impurity effect per se has been discussed in detail previously. Because of the importance of the impurities on the corrosion of Hastelloy N by fluoroborate, efforts have begun on methods of purifying the salt. Also to be considered are better methods for analyzing and identifying the impurities. At some purity level, perhaps 200 to 400 ppm water andfor oxygen, as analyzed, solid-state diffusion of chromium in the Hastelloy N will likely - control corrosmn as it does in the fuel salts This is most important, since it suggests that an entire MSBR of either the one- or two-fluid variety can be operated with none of the main circulating channels suffering more than a few tenths of a mil per year corrosion attack. Experimental proof of this is one of our major near-term goals, 18.7 FLUOROBORATE PURITY TEST J.W.Koger R.F. Apple Attempts to purify the NaBF,4-NaF (92-8%) coolant salt have continued. A schematic of the process is given in Fig. 18.19. A mixture of BF;, He, and heated HF was passed into a Hastelloy N vessel containing the impure (~2000 ppm H, O) fluoroborate salt (2.7 kg) at 480°C for 15 hr. The exit gas entered a 90 vol % methyl alcohol—10 vol % pyridine solution, where the mixture was titrated with Karl Fischer reagent to the dead stop end point. The titration indicated that about 600 ppm H, O was removed. Chemical analysis disclosed that the O, content decreased from 1700 to 1400 ppm and that the H,O content changed from 2000 to 300 ppm. Changes in the BF; content of the salt appeared to be minimal. Further work is planned using a nickel vessel to eliminate corrosion products resulting from the reaction of HF with Hastelloy N. ' 18.5 FORCED-CONVECTION LOOP MSRFCL-1 'H.C. Savage J W. Koger W.R. Huntley The MSR-FCL-I forced-c1rculat10n loop is being operated to evaluate the compatibility of standard - Hastelloy N with NaBF,-NaF (92-8 mole %) coolant salt at temperatures and flow rates similar to those which existed in the MSRE coolant circuit. Salt velocity in the %-in.-OD by 0.042-in.-wall Hastelloy N loop tubing is normna]ly 10 fps, and Hastelloy N corrosion " Table 18.11. Comparison of Conosnoh Rates for Standard Hastelloy N in MSR Systems After More than 5000 hr Operahon Loop Salt - T max e T Velocity Cfrcrl:g:::il{ta te Designation Type (O (0) -~ (fps) (mils/year) NCL-13A" Coolant? 605 145 0.1 - 06 NCL-14 Coolant? 605 145 01 0.55 MSR-FCL-1 Coolant? 588 78 10 1.2 'NCL-16 . ~ Fuel® 705 170 01 - 004 NCL-15A Blanket® - 675 5 . 01 0.03 NCL-18. Fertile- = 70§ 170 0.1. 0.05 Fissiled o ' 9NaBF4-NaF (92-8 mole %), 1000 ppm each water and oxygen. bLiF-BeF,-UF4 (65.5-34.0-0.5 mole %), <200 ppm each water and oxygen. €LiF-BeF,-ThF4 (73-2-25 mole %), <200 ppm each water and oxygen. dLiF-BeF,-ThF4-UF,4 (68-20-11.7-0.3 mole %), <200 ppm each water and oxygen. 248 35 cm3 /min BF3 3, . 60 cm /m?n o 3 HF 125 cm™/min 00O *NaBF, — NaF (92-8 mole %} 450 - 510° C ORNL-OWG €9-142547R KARL FISCHER REAGENT . END POINT INDICATOR 90% METHYL: ALCOHOL 10% PYRIDINE | ORIGINAL — 2000 ppm HO AFTER 15 hr SPARGING =300 ppm H,0 Fig. 18.19. System for Removing Water from Coolant Salt, test specimens are exposed to the circulating salt at * three temperatures — 510, 555, and 588°C. The third 2000-hr run (total accumulated time of 6098 hr at design conditions) was completed on October - 22, 1969, and corrosion specimens were ~ removed for metallographic examination and weight change measurement. Test specimens were reinstalled in the loop, and on November 23, 1969, operation was started on the fourth run. As of February 28, 1970, the loop had accumulated 8100 hr of operation at design’ conditions. Salt samples have been taken for chemical analysis about every 500 hr. 18.8.1 Metallurgical Analysis Wéight Changes. — Table 18.12 details weight -changes of the standard Hastelloy N specimens in the loop and the changes in concentration of the metallic impurities ‘in the salt during the operating life of the loop. It is apparent from the weight changes that the corrosion rate is decreasing and now averages about 1.2 mils/year at the hottest position, assuming uniform attack. This rate is higher than that obtamed in thermal—convection loops under similar conditions; the apparent effect of velocity on corrosion strongly suggests that corrosion processes are occurring other than the selective removal of chromium and iron. Salt Chemistry. — The data in Table 18. 12 show that “there was a large increase in chromium fluoride com- pound (corrosion product) in the salt during the first operating period with little subsequent increase. During .each period the iron fluoride has decreased. This leads one to beheve that the reactlon Cr(s) + FeF,(d) = CrF,(d) + Fe(deposited) , where s = sblid solution, d = dissolved in salt, is responsible for a portion of the corrosion. The salt contained a large amount of iron fluoride at the beginning (407 ppm), and each time the salt is dumped Table 18.12, Average Weight Change of Specnnens in MSR-FCL-1 and Changes in Concentratlon of Chromium and Iron in NaBF4-N aF (92-8 Mole %) Average Specimen Weight Corrosion Rate at 588°C Incremental Time of Operation Change at Ind:cated Assuming Uniform Loss Impurity Change (hr) Temperature (mg/cm?) (mils/year) in Salt (ppm) 510°C 555°C 588°C - Overall Incremental cr . Fe 27412 | +1.0 _27 ~1L1 6 . 16 +234 ~155 4755 _ +2.5 —4.2 - -1713 1.4, 1.2 . +40 -263 67642 To+3. -5.2 -—21.0 1.2 ' 0.7 0. ~190 % ncludes 659 hr isothermal operation, - b6098 hr at design conditions. 249 Y-9761F i 10.004 in, 10.003 in. 0.007 INCHES 500X 10.005 in. b 10.007 in. Fig. 18.20. As-Po!xshed Photpmicrograph of a Standard Hastelloy N Specimen from MSR-FCL-I Exposed to NaBF4 -NaF (92-8 Mole %) at 588°C for 6764 hr, Weight Loss 21.4 mglcm . S00X, back into the fill tank with the residual salt it picks up more iron fluoride. The results of the oxygen and water analyses have been quite scattered, with the values around 1000 ppm for each. These fairly high numbers and the velocity effect indicate that impurities such as HF also play a large part in the corrosion. With no increase of corrosion products in the salt at the present time, all the material removed in the hot section is deposmng in the cold section. Evidence at the present ‘shows that the deposit is metallic and adherent on the metal surface and is not in the form to cause plugging very rapidly. Meta!lography - anures 18.20—18.22 show micro- graphs of specimens from several locations in MSR-FCL-1 after more than 6000 hr salt exposure. Figures 18.20 and 18 21 are the specimens exposed to salt at 588 and 555°C respectwely The uniform attack is seen on both specimens, with the rougher surface seen on the iugher-temperature sample shown in Fig. 18.20. In Fig. 18.22 the mounting material removed some of the deposit, but portions can still be seen. Pump Deposits. — Visual examination of the pump and pump bowl after over 6000 hr of operation showed no obvious corrosion. None of the green Na3CrFg corrosion product that had been observed previously was present. o ~Cold Finger.. — During the present run a cold finger corrosion product trap, similar in design to one pre- viously used in sodium fluoroborate test loop PKP-1, 1 was inserted into the salt in the pump bowl in an attempt to induce preferentlal deposmon of corrosion products from the salt circulating in the loop. Such pref- erential deposition was observed on the cold finger in- serted into the pump bowl of loop PKP-1. ! The cold finger used in MSR-FCL-! is a closed-end nickel cylinder, 1% in. long, % in. OD, with a 0.070-in.-thick wall. Cooling is by means of an argon- water mixture injected into the cylinder ID and then discharged to the atmosphere. The metal wall tempera- ture is.measured and recorded by two 0.020-in.-OD sheathed, ungrounded Chromel-Alumel thermocouples inserted in two 1-in.-deep axial holes (0.023 in. diam) in the 0.070-in.-thick wall of the cold finger. 1A. N. Smith, P. G. Smith, and R. B. Gallaher, MSR Program Semiann. Progr. Rept. Feb. 29, 1969, ORNL-4396, p. 102. 250 Y-97616 loooim, ‘|0003h. 0.007 INCHES 500X [ 15605 ™ fre— 10.007 in. Flg. 18.21. As-Pohshcd Photonucrograph of a Standaxd Hastelloy N Specimen from MSR-FCL-1 Exposed to NaBF4-NaF (92-8 Mole %) at 555°C for 6764 hr. Weight loss 4.7 mg/cm?. 500X. Eight tests were made in which the cold finger was inserted into the salt in the pump bowl of test loop MSR-FCL-1 and cooled to temperatures (as indicated by the thermocouples in the wall of the cold finger) ranging from 493 to 140°C. The duration of the tests ranged from 1.5 to 5.3 hr. The temperature of the salt in the pump bowl was 510°C. In contrast to the deposits of material containing Na, CrF which were found on a cold finger in PKP-1 loop at metal wall temperatures of 400, 460, and 477°C (ref. 2), no significant deposit of any kind was seen on the cold finger tests in loop MSR-FCL-1. Even in the three tests where indicated wall temperatures were below the salt liquidus temperature (385°C), the surface. of the cold finger was essentially clean as visually observed when withdrawn into a sight glass. Occasionally, small patches (%4 to % in. across) of The cold finger was then mstalled in the PKP-l fluoroborate loop in an attempt to duplicate the previous cold finger test results of this loop. The chromium concentration of the fluoroborate salt in loop PKP-1 is about 500 ppm, while that in MSR-FCL-1 is about 250 ppm. Two tests were run in which the indicated cold finger wall temperature was about 150°C (salt temperature in the pump bowl was 548°C). Test times weré 1 and 4.5 hr. No deposition on the cold finger was seen after withdrawal into a sight glass. Tests at ‘higher temperatures were not made because of temperature control difficulties with the cold finger. Since the cold finger previously used in the PKP-1 loop contamed grooves on the outside surface, the surface (Iower half only) of the MSR-FCL-1 cold finger was scored with file marks ~0.010 in. deep, and a third test run at about 150°C lasting 6% hr was made in the whlte material, estimated to be a few mils tthk were - seen on the surface. - 2R, B,_ Gallaher and A. N. Smith, MSR Program Semiann. Progr. Rept, Aug. 31, 1969, ORNL-4449, pp. 74-75. PKP-1 loop. In this run a deposit was obtained.. Generally, the entire surface of the cold finger was covered with a white deposit, with an overlay of bright ‘green material on the lower half of the cold finger. The - cold finger was allowed to stand overnight under a [~= ETY 0.007 INCHES 500X I Fig. 18.22, As-Pohshed Photomlcrograph of a Standard Hastelloy N Specimen from MSR-FCL-1 Exposed to NaBF,-NaF (92-8 Mote %) at 510°C for 6764 hr. ‘Weight gain 5.0 mg/cm 500X. The light material away from the surface of the sample is part of the surface layer that was separated from the sample during the mountmg operation. - : helium atmosphere, and by this time the deposit had begun to separate from the cold finger surface. In moving the cold finger for photographing, the entire deposit spalled off, leaving a clean metal surface. The total deposit weight was 1.63 g, of which 1.26 g was mostly green material (complete separation was not possible). Chemical analysis of the green material disclosed 2.93 wt % Cr, 1.03 wt % Fe, 370 ppm Ni, and <500 ppm Mo, with the remainder Na, B, and F. Stoichiometric calculations show 11 mole % Na;CrFg ‘and 4 mole % NazFeFg, with the remainder a mixture of NaBF, and NaF. | - Subsequent operation at the same condxtlons in loop 'MSR-FCL-1 did not produce a deposit. On the basis of these tests we have concluded that a wetting problem “exists with the salt and that a grooved or roughened surface is required in order to obtain a deposit. We are trying various designs to improve adherence of the deposit to the cold finger and to lessen the probability of accidental removal during withdrawal from the pump bowl. ~ 18.8.2 Forced-Convection Loop MSR-FCL-2 Work is in progress on a second molten-salt forced- circulation loop, designated MSR-FCL-2, which is to be ~used to study the corrosion resistance of Hastelloy N and the mass transfer properties of Hastelloy N and fluoroborate-type coolant salt systems at conditions proposed for the MSBR. Other salts proposed for the ‘MSBR could also be circulated in the loop if desired. The loop design is generally similar to the presently operating - forced-circulation corrosion-test loop,3 MSR-FCL-1. A new pump?* (designated ALPHA) de- signed for variable salt flow rates up to 30 gpm and heads to 300 ft will be used to provide increased salt 3p. A. Gnadt and W. R. Huntley, MSR Program Semiann. Progr. Rept. Feb. 28, 1968, ORNL-4254, pp. 226-27. A G. Gnndell et al, MSR Program Semiann, Progr Rept Aug. 31, 1969, ORNL4449, p. 78. 252 velocities up to 20 fps. Other important new features of - the loop design include: (1) three corrosion specimen assemblies designed for easy installation and removal without draining salt from the loop, (2) two in- dependently controlled heated sections for flexibility in ~controlling the salt temperature at each of the three corrosion specimen assemblies, (3) provisions for de- termining heat transfer coefficients in one of the heated sections, (4) bulk fluid AT to 165°C, which requires ~125 kw at the expected flow rate of ~4 gpm, and (5) an auxiliary expansion tank to allow for expansion of the salt and to provide space for salt sampling, liquid level indicating probes, installation of a “cold finger” ~device to study - preferential deposition of corrosion products, and gas purging for ondine salt purification. ~ 18.9 CORROSION OF HASTELLOY N | IN STEAM B.McNabb H.E.McCoy Specimens were removed from the Bull Run Cor- rosion Facility? on January 12, 1970, after 2000 hr in a steam environment at 3500 psi and 538°C. Weight gains were about 0.1 to 0.4 mg/cm? (0.04 to 0.14 mil/year assuming uniform corrosion) for all compositions of standard and modified Hastelloy N. The oxide films were adherent, with no sign of spalling. The specimens were removed for weighing at 372, 1000, and 2000 hr. The oxidation rates appear to be linear after the initially faster rate up to 372 hr. The same trends are continuing that were observed at the first removal. We observed that surface treatment has an effect on the weight gains of the specimens. After 2000 hr, surface- ground specimens (400-grit paper) had the largest weight gains of 04 mg/cm?, as-received gained 0.2 mgfcm?, and electropolished specimens gained 0.1 mgfcm?. Most of the specimens were returned for further exposure, but some were removed for detalled examination and testing. ~ Some chromlum steels were also mcluded in the facility for comparison. The oxidation rates of the steels were much higher than Hastelloy N, having weight gains from 3 to 6 mg/cm? for 1000 hr exposure to the same environment. These steels contained from 1.1 to 8.7% chromium and ~1% molybdenum. As ‘shown in Fig. 18.23, the oxidation rates of these steels varied by a factor of only 2 in the steam environment at 1B, McNabb and H. E, McCoy, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 205-9, ORNL-DWG 70-6787 8 - .. _AMARS.0 o £ / LAl > / 3 Cr 1.90 £ 8 Cr 4,20 w4 7 “‘"1 & & s0x I T = o i~ 10.010 in. 0.045 INCHES G.0%0m. 75X 10.020 i, ’ Fig. 195, Molfbdenum T-Joints Brazed with Two Different Fe-Mo-C-B Brazing Alloys, Excellent wetting was obtained. Note the cracks in the powder-metallurgy base metal. As polished. 258 Y-97842 Fig. 19.6. Molybdenum Lap Joint Brazed with Fe-C-B and Tested in Bismuth 700 hr at 600°C. As polished. 62X. The bismuth used in these tests was grade 69 obtained from Cominco American,: Oxygen contamination in the ‘bismuth resulted in reactions tliat caused the formation of a plug in the first.two loops; but this problem was eliminated in subsequent tests by bubbling hydrogen through ‘the molten metal for 2 hr at 350°C and filtering - through a 10 type 304 stainless steel filter prior to loading into the loop. One of the most promising materials appears to be molybdenum, Loop 3 contained two tensile samples of ~molybdenum and one molybdenum and one TZM (M0-0.5% Ti—0.1% Zr) tubular sample in each of the L reduction in thickness was detected, " and the weight . changes were within experimental error. Specttographic analysis of the bismuth before and after test showed an increase in molybdenum concentratlon of only 0.3 ppm. Loop 4 contained niobium and Nb-—l% Zr samples and was operated at 705°C with a AT of 75 + 5°C for a period of only 115 hr before plugging occurred. The - effect of bismuth on these samples at this temperature high- and low-temperature regions. After 3000 hr of - scheduled operation at a maximum temperature of 700°C and a AT of 95 * 5°C, we found very little reaction between molybdenum or TZM and the molten bismuth. In fact it did not appear that the bismuth wet ‘the metal surfaces very well. Figure 19.8 shows the appearance of the samples before and after test. No was catastrophic, as indicated by Fig. 19.9. The thickness of the tensile samples was reduced from 0.0195 in. to 0.013 in., indicating dissolution by the bismuth, X-ray diffraction analysis of a sample taken from the plug showed the presence of several phases: - one had the same structure and lattice parameter as high-purity bismuth, and one was identified as alpha- zirconium, There were also four weak unexplained -diffraction maxima. None of the diffraction lmes could be attributed .to a niobium type of crystal structure. "FROZEN _BISMUTH 259 . MOLYBDENUM 0 Fig. 19.7. Molybdenum T-Joint Brazed with an Fe-Mo-C-B Brazing Alloy and Tested in Bismuth. Cracking must have occurred after testing, since the cracks extend through the frozen bismuth. As polished, 62X. However, since the atomic radius of niobium is the same as that of bismuth, it could occupy substitutional lattice sites in the bismuth unit cell without affecting the lattice parameter. A semiquantitafive spectrographic analysis of the plug indicated that in-addition to bismuth it contained 0.5% Nb, 100 ppm Zr, 300 ppm Cr and Ni,-and 500 ppm Fe. The chromium, nickel, and jiron undoubtedly were picked up during filtering of the bismuth through a stainless steel filter prior to test. In the future we plan to determine the compatibility of other potential container materials such as tantalum and graphite with molten bismuth. In addition we will also evaluate the effect on compatibility of other metallic ions such as lithium or thorium in the bismuth. 19.5 CHEMICAL VAPOR DEPOSITED COATINGS L.E.Poteat J.I. Federer . Application and evaluation of tungsten coatings on materials for fuel processing requirements have con- tinued. Tungsten, which has good corrosion resistance to liquid bismuth, is applied by hydrogen reduction of WF¢ at 500 and 600°C. Coatings measuring 0,005 to 0.009 in. thick were deposited on the materials shown in Table 19.1, Initially we believed that the degree of adherence of the coatings was associated with the difference in coefficient of expansion between tungsten and the substrate, However, our recent work has shown that the composition of the substrate determines the degree of adherence. Differences in thermal expansion, on the other hand, may influence the service life of the coating, . L A qualitative assessment of the adherence of the tungsten coating to the materials in Table 19.1 was made immediately after coating. We observed that the coatings cracked and spalled from the steel and stainless steel specimens upon cooling from the deposition temperature, but remained intact on iron-nickel alloys, “A” nickel, Hastelloy C, and Inconel 600. The non- adherence of tungsten coatings to steel or stainless 260 Table 19.1. Composition and Thermal Expansion of Subsmfe - Coefficient of Substrate ppermg) Expansion,? Pnnc:p‘al ?;r:;s Heuents Materal — 2510600C @ —— ——— : | [pm.m. ( C)-l] Fe Cr Ni Steel 145 99+ Stainless steel ' Type17-7PH- -~ ~17.1%2 c 17 7 Type 304 185 ¢ 18 8 Type 405 ~11.2 ¢ 11.5-135 - Type430 11.2 ¢ 14.0-18.0 Type 442 117 ¢ 18.0-23.0 “A” nickel 133 ' 99 Hastelloy C4 13.3 5 15 58 Inconel 600 : - 153 9 . 12-15 175 Fe-35%Ni . 10,0 e 35 Fe-40% Ni - 100 c 40 . Fe-45% Ni 100 c 45 Fe-50% Ni 100 c 50 _ “Tungsten4 6 pin, in, ™} ( ‘0. bgolution-treated condmon, lower in the aged conditions. “Major elemént. 217% Mo, 5% W. o) (8) Fig. 19.8. Sample Amangement Used in -Loop 3. () Before test, (b) after 300 hr in bismuth at 700°C All tensile samples are molybdenum, _steels is attributed to the formation of stable iron andfor chromium fluorides on the surface by reaction .with WF4. These fluorides are neither volatile nor reducible by hydrogen at 600°C. We know that the tendency for nickel to react with WFs to form a fluoride is much less than for iron and chromium, and a continuous fluoride film probably does not form on the surfaces of iron-nickel and nickel-base alloys. Utilizing this information we obtained adherent tungsten coat- “ings on stainless steels by first applying a thin nickel plate by electrodeposition. It was necessary to bond the nickel plate to the stainless steel by annealing at 800°C prior to deposition of tungsten; otherwise the tungsten coating would not adhere. Typical specimens are shown in Fig. 19.10. Note that the coating cracked and separated from a plain type 430 stainless steel speci- men, but was intact on either a nickel-plated type 430 specimen or an Inconel 600 specimen. The adherence of the coatings to sheet-type speci- mens was further evaluated by thermal cycling, spiral | ~ bend testing, and tensile testing. We found the coatings - uncracked and intact on the iron-nickel alloys, “A” nickel, Hastelloy C, and Inconel 600 after 15 cycles between 25 and 600°C. Types 405, 430, and 446 stainless steel specimens which had not been nickel plated prior-to tungsten coating were also tested, but the coating cracked and spalled from types 430 and 446 after four cycles and began to separate from type 405 after 15 cycles. In the spiral bend test at 25°C, specimens are manually bent around the jig shown in Fig. 19.11.! Bent specimens of types 304 and 430 stainless steel ‘(nickel plated before coating), Hastelloy C, and Inconel 600 are shown in Fig. 19.12. The stainless steel specimens were about 0,025 in. thick and the other two specimens were about 0.063 in. thick prior to coating. Each specimen was bent to a radius of curvature of 1 in. or less, and different amounts of springback determine - the final shape of the specimens. Although the tungsten coatings cracked, they did not spall from any of the specimens. _ Tensile testing of the coatmg utilizes the specunen - arrangement shown in Fig. 19.13. The coated specimens are brazed to the ends of steel bars, which are pulled in a tensile machine. Tensile adherence strengths greater "than 16,000 psi have been measured for coatings on Inconel 600. Coatings on stainless steel are expected to have similar strengths, but they have not yet been evaluated, 15 Edwards, “Spiral Bend Testing for Electrodeposited Coatings,” Trans. Inst. Metal Finishing 35, 101-6 (1958). i i ! i i { { | i | i i | (c) 0 INCHES 1 () Fig. 19.9, Tensile Samples of Niobium (Top) and Nb-llr (Bottom) from Bismuth Loop 4. (¢) Hot leg; (b) cold leg; (¢) before test. Coating of substrates with molybdenum by hydrogen reduction of MoF¢ has begun. This process requires temperatures of approximately 800°C or higher. At lower temperatures an impure, noncoherent deposit occurs. One experimental disadvantage of higher deposi- tion temperatures is that the vessel to be coated must be strong enough to withstand a differential pressure of about 1 atm at temperature, since pressures less than 10 torrs are required to produce a deposit of uniform thickness. A second disadvantage could result from = reaction of molybdenum with the substrate material. Although both tungsten and molybdenum form binary alloys with iron, chromium, and nickel, molybdenum would be expected to alloy more readily with these ‘elements, especially at the higher deposition tempera- tures that are needed. Molybdenum can form inter- metallic compounds with iron and nickel, and this could decrease . the adhesion of coatings. However, preliminary investigation has ‘indicated that molyb- ‘denum coatings are more ductile than tungsten coatings and, therefore, should be less likely to crack during thermal cycling. 19.6 SCALING RESISTANCE OF CARBON STEELS B.McNabb H.E.McCoy Vessels used to contain bismuth for molten-salt fuel processing studies have been made from carbon steel, but they have been found to form an oxide scale in service at 650°C. To evaluate the effectiveness of various commercial coatings in preventing scaling, 262 Y-98673 i : O 1 2 . - “ INCHES Fig. 19.10. Typical TungSten-Coated Specimens. (z) Type 430 stainless steel, coating cracked and separated. (&) Type 430 stainless steel, nickel plated prior to coating. (c) Inconel 600. o ‘ Fig. 19.11, Spiral Bend Test Jig. i 263 Y-98674 INCONEL 600 0 , o HASTELLOY C INCHES Fig. 19.12, Tungsten-Coated Bend Specimens. Y-98672 COATED SPECIMEN FARE R RN INCHE Fig. 19.13, Tensile Specimen for Testing the Adherence of Coatings. several coatings recommended for use in this tempera- ture range for protection of steels from oxidation were compared. These were Markal CR-9 (Markal Company, Chicago, Illinois), Speco H-170, a silicone-base alu- minum paint, Speco H-170, a silicone-base zinc paint (Speco Inc., Cleveland, Ohio), a silicone-base aluminum paint (unknown vendor), and as-received tubing with no coating (designated M Cr-9, SPE AL, SPE ZN, SIL Al, and AS REC respectively). The specimens coated were 1-in.-long, 0.375-in.-diam tubes with 0.058 in. wall thickness. They were first cleaned in acetone and then the coatings applied by dipping and air drying 16 hr. They were then heated in an air furnace for 1000 hr at " 650°C, cycling to 25°C every 25 hr. | Results are summarized in Fig. 19.14, The as-received specimen gained weight until the oxide layer spalled after about 400 hr. A net weight loss was measured after 600 hr, but then after that time the specimen started to gain weight again. The specimen coated with * Markal CR-9 initially lost weight, presumably due to a loss of volatile elements in the coating, then gradually’ gained weight until after 1000 hr it had the smallest weight change of all. The specimens coated with the siticone-base aluminum and zinc paints gained weight at approximately the same rate as the as-received speci- men, but they did not spall upon cooling and, there- fore, had the largest weight gains after 1000 hr. The remaining unaffected metal (wall thickness) in the various specimens was M CR-9, 0.049 in.; SPE AL, 0.041 in., SPE ZN, 0.038 in.; SIL AL, 0.037 in.; and AS REC, 0.035 in. This value was inversely related to the observed weight increase except for the as-received specimen, where spalling occurred. An alternative to using carbon-steel tubing to contain bismuth in a fuel processing vessel might be to use 264 ORNL-DWG T70-6789 60 SiCA /j. 50 — ,7. - | ./ SPE Zn 40 ./ : & N ¢ — 30 pa AS RECEIVED / WEIGHT CHANGE (mg/cm?) ‘\\ N N - N\ 20 V;/ /— \ x - 10 / / ’/ ( e /’!r ___‘_./' L/ i _ ) 0 %" | (1/ o™ X\~ = N7 Cr-9 o " | 0 200 400 600 800 1000 TIME (hr) Fig. 19.14. Weight Changes in Coated Stee! Samples Exposed to Air at 650°C and Cycled to 25°C Every 25 hr. Croloys or chromium steels that have improved oxida- tion resistance. Croloys are widely used in steam plants, but not much is known about their spalling resistance in air at 650°C under cyclic temperature conditions. To evaluate their spalling resistance, samples of 14, -in. to 3-in. pipe of Croloys having 1.1, 1.9, 2.0, 4.2, and 8.7 wt % chromium along with a maraging steel (Ni, 12%; Cr, 5%; Mo, 3%) for comparison were tested. We found that small additions of chromium to.iron increased scaling but that additions greater than 5% greatly improved oxidation resistance. 20. Support for Components Development Program H.E. McCoy The culmination of a development program is to use the basic information that has been developed to power plant will require pumps, heat exchangers, steam generators, and numerous other components. At this stage of development, work on components is not receiving a high priority. However, certain basic capa- bilities are needed to proceed with building components when this part of the program is reached. One of these areas involves the ability to make remote cuts and welds ‘construct an operating engineering system. A nuclear . Figure 20.1 (composite) shows a cross section and photomicrograph of a root pass in Hastelloy N piping. The weld is extremely sound and free from any type of microfissuring. Figure 20.2 shows the general ap- pearance of welds in 6-in.-diam pipe after the root pass (left) and completion of filler passes (right). A major problem in remote maintenance of an MSBR ~ or any reactor is that of obtaining correct alignment for maintenance purposes. Our involvement in this program has been primarily with the choice of welding configurations and parameters. A second area of work concerns the development of a bearing to operate in molten salts. Our present concepts of fuel and coolant pumps do not require such a bearing, but .small auxiliary pumps may be desired that utilize such a bearing. A bearing for this operation must be a cermet (ceramic-metal composition). This cermet can be solid or simply a coating on a metal. Our current work involves the evaluation of four cermet coatings that - were applied by plasma spraying. A third area is involved with assuring adequate instrumentation for operating a nuclear plant. The examination of a pressure-measuring device in a sodium fluoroborate loop has led to a better appreciation of the quahty control needed for such de\nces 20.1 REMOTE WELDING DEVELOPMENT T R.Housley G.M. Slaughter | and fitup of mating components. If very accurate alignment is not a necessity, the design criteria for pipe aligning mechanisms can be simplified. We are determin- ing the amount of mismatch which will be permissible, and Fig. 20.3 shows a cross section of a successful pipe weld which was misaligned by about % ¢ in. The adaptability of the equipment to this condition is excellent. We are also improving the welding torch for this equipment, and an automatic self-adapting arclength control has been developed. This is in preliminary checkout evaluation and looks very promising. 20.2 SUPPORT FOR SYSTEMS AND ~ COMPONENT DEVELOPMENT J. W. Koger , Failure Analysis of the PMD in the PKP-1 Pump - Loop. — An evaluation of the salt leak that occurred ~ near the Inconel PMD (pressure measuring ‘device) of Through a cooperative program with the Reactor Division (sece Chap. 7), we are developing optimized - joint designs and welding procedures for remotely welding the large Hastelloy N pipes in an MSBR. For economy and expediency we have done much of our “preliminary work with - austenitic stainless steel. A description of the welds made and the types of joint . designs and consumable inserts studied is presented in Sect. 7.4. This discussion describes the metallurgical evaluation of the test welds. the PKP-1 pump loop (see Chap. 7) was made using radiographic, metallographic, and chemical techniques. The system had been in operation approximately eight years and contained NaBF.; -NaF (928 mole %) dunng " the last two years. The salt leak occurred near a weld joining two sections of sched 40 Inconel pipe. The weld was about " % in. from the PMD diaphragm flange and about 18 in. from the PKP-1 main loop. Records showed that the 265 PMD was purchased in 1956. According to the manu- facturer, the only test for quality of the weld where failure occurred was a helium mass spectrometer leak check. Our analysis disclosed that the weld almost g N o N ~ N l S — o -y | v 5000 |_X002 W OO'0] SIHINI BIO™ 0 W G10°0) mns e e e is t soundness ]oih » icrograph of Root Pass. Excellent ection of Root Pass in Hastelloy N Pipe; (5) Phot Fig. 20.1. (3) Cross S evident. Etch, HCl, H,0,. 267 Y-98196 INCHES Fig. 20.2. Welds in 6-in.-Diam Pipe After the Root Pass (Left) and Completion of Filler Passes (Right). INCHES Y-98194 Fig. 20.3. Successful Weld in Piping Misaligned by About %4 in. -completely miséed the intersecti.on of the two poftic‘ms of pipe due to misalignment of the matching ends. It is postulated that this was discovered by ORNL personnel after purchase, and that a repair was attempted by back brazing from inside the tube. Pieces of suspected brazing alloy were removed from inside the tube and analyzed. The compositions of the remaining braze alloy and the Inconel are given in Table 20.1. This brazing alloy was much less compatible with the fused salt than Inconel due to the presence of B and higher Cr and Fe concentrations. Thus we believe that corrosion was the cause of failure. _Figure 20.4 shows the failed pipe section in both the as-polished and etched conditions. The porosity of the braze material is evident in Fig. 20.4a. The weld metal and the pipe intersection are evident in Fig. 20.4b, 268 Table 20.1. Compositions of Inconel and Braze Alloy "Ni Cr Fe Ti Mn Si Al Mo B Na Inconel 77 14 74 0.3 Brazealloy 57 10 30 0.2 02 0.1 0.2 <0.03 03 005 0.2 0.3 1.0 0.5 Table 20.2. Composition of Induction Probe Sheaths Alloy Constituents (wt %) Ni Fe Cr Mo Mn Si C S P Hastelloy N Bal 4.22 6.96 16.17 0.47 0.62 006 0.010 0.2 1.3 0.7 0053 0.027 0.021 Type 304 Stainless 9.8 Bal 18.0 Table 20.3. Effect of Applied Electric Current on Corrosion in Fluoroborate Salt at 550 C Sample . Exposure 1000-Hz - Wall Thickness (mils) 0. Material Time (hr) Cuflent‘ Before (Nominal) After Change 336 Hastelloy N 1430 Yes 35 33.2 -1.8 337 Hastelloy N 1430 No 35 34.0 -1.0 338 Type 304 SS 70 No 20 12.5 -17.5 339 Type 304 SS 70 Yes 20 13.0 -7.0 %Probe electrical conditions were 0.04 v rms and 0.1 amp. although the etchant removed the braze metal. Figure .20.5 shows a portion of the tubing 180° away from the failure. Again porosity is evident in the braze metal. - The weld here was properly aligned. Figure 20.6 shows the condition of the Inconel pipe wall away from the failure. Subsurface voids are evident to about 7 mils in depth. - Another PMD, bought at the same time, was cut apart and analyzed. In this device the weld was sound and no braze alloy was present. Thus we believe that this brazing operation was an isolated repair. We have ~ recommended, however, that all PMD’s purchased on the same order and serving in critical facilities be tested by radiographic methods to determine their condition. Effect of Applied Electric Current on Corrosion in Fluoroborate Salt. — Four liquid level induction probes exposed to salt in the Inconel PKP-1 loop were examined in order to evaluate the effect of excitation current on fluoroborate corrosion. Two of the probes were sheathed with Hastelloy N and two with stainless steel. The compositions of the respective alloys are given in Table 20.2. The stainless steel probes were exposed for 72 hr and the Hastelloy N probes for 60 days. For each material, one probe served as a control, no current imposed, while the other saw 1000 Hz 0.04 v rms and 0.1 amp. The salt temperature in all cases was 550°C. The results of the tests are summarized in Table 20.3. : ' The visual appearance of the probes after test is compared in Fig. 20.7. The two stainless steel probes . showed heavy attack and were noticeably discolored. The corrosion rate of the stainless steel averaged 0.1 mil/hr. Hastelloy N showed no visible attack, but measurements of the wall thickness of the sheath indicated a corrosion rate of 7 X 10™* mil/hr. Wall thickness changes (Table 20.3) and metal- lographic examination both indicated that the excita- tion current had a small effect on Hastelloy N corro- sion, although precise measurements in a more highly controlled environment are necessary to confirm this behavior, Such an effect was not noted for the stainless steel probes, which suffered 143 times the corrosion loss of the Hastelloy. The heavy attack of the stainless steel probes in this test paralleled that which led to 269 < » [+ 0 o i > . ? (a) As polished, 20X 1 Loop. Device on PKP- ng to Pressure Measuring Leadi c Inconel P; in 12X Failure Near Butt Weld Fig. 20.4. (b) etched, HC1, HNO,, lactic acid, 270 Y~96896 Fig. 20,5, Appearance of Weld 180° from Failure Area Shown in Fig, 20,4, (a) As polished, 20X; (b) etched, HC1, HNOj3, lactic acid, 12X, ' ' i o Y-96891 15.001 m. 10.003 in. 0.007 INCHES 500X I 15.005 m, i = 10.007 in. Fig. 20.6. Inconel Pipe Joining PMD and PKP-I Loop, Exposed to Fluoride Salts for Many Years. Etched, HCI, HN03, lactic acid, 500X.: failure of a stainless steel probe during an earlier PKP-1 loop run.! Although stainless steel is nominally poorer in resistance to salt attack than Hastelloy N, attack of the former alloy in the PKP-1 system is magnified by dissimilar metal interaction with the large Inconel o surface compnsmg the system. 20. 3 COATED BEARING SPECIMENS W. H. Cook Tests on high-temperature bearihgs using alkali metals or fluoride salt as hydrodynamic lubricants have gen- erally indicated that hard surfaces working opposite ~ hard surfaces give the better performance in boundary (rubbing) lubrication tests.!»2 Boundary lubrication conditions will exist in certain eccentric loadings or 1y, W Koger and A. P. Litman, Catastrophic Corrosion of Type 304 Stainless Steel in a System Circulating Fused Sodium Fluoroborate, ORNL-2741 (January 1970). during the starting or stopping of hydrodynamically lubricated bearings. Damage usually results from solid- phase bonding. Our present concepts of large molten-salt pumps do not require bearings that operate at high temperatures. However, there are small auxiliary pumps and valves in which it probably will be desirable to use materials that - will resist solid-phase bonding (galling, self-welding) and ~ wear in salt up to temperatures of 700°C. We feel that metal-bonded carbide cermets offer the best potential for these kinds of operatmg conditions, and we cur- rently have a small evaluatxon program. . The least 'W. H. Cook, “Materials for Valves and Bearings in Molten Metals and Fused Salts,” pp. 762—74 in Reactor Handbook, 24 ed,, vol. 1, ed. by C. R, ‘Tipton, Jr., lntersmence, New York, 1960. 2. L. Moor and R. G. Frank, Materials for Pdtasfiumf Lubricated Journal Bearings, Vol. V. Friction and Wear Studies (FmaI Report), for the Period from April 22, I963-September 22, 1966, GESP-100, pp. 7, 8; 11. Fig. 20.7. Liquid-Level Induction Probes Exposed to NaBF,-NaF (92-8 Mole %) in PKP-1 Loop To Determine Effect of Applied 272 "1 Y-96516 INCHES Current. Nos. 336 and 337: Hastelloy N, 1430 hr, 336 had 1000-Hz current Nos. 338 and 339: type 304 stainless stéel, 70 hr, 339 had 1000-Hz current. expensive and least complicated design approach is to apply such hard-surface coatings as thin layers on bearings or valve faces and surface and grind them to the required surface finish. We are currently evaluating four such coatings that have been applied to Hastelloy N cylinders 1 in. diameter and 1 in. long by plasma spraying. These were surface ground to make them 0.003 in. thick. Their designations and nominal compositions are listed in Table 20.4. These coatings, represented by six specxmens per coating composition, survived 100 thermal cycles be- tween 100 and 700°C in pure (99.996%) argon.® The heating and cooling parts of each cycle were 20 and 40 min respectively. These thermal cycles caused no detectable changes in the sample diameters. None of the 3W. H. Cook and L. R. Trotter, MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL-4396, pp. 267—68. coatings spalled, but all showed some degree of failure in the fluorescent dye penetrant examinations made on them by the nondestructive testing (NDT) group before and after the test. These coated bearing specimens are listed in Table 204 in the decreasing order of their resistance to the thermal cycling tests as mdxcated by these NDT tests. - Metallographic examinations of as-recelved (control) and selected thermal-cycled (tested) specimens have been made.? Their microstructures are compared in Fig. 20.8, and comments on these are given in Table 204, There was no evidence of any separation between the coatings and Hastelloy N, even though all showed scattered oxide at this interface as part of the fabrica- tion problem. The presence of oxides in the matrices of the coatings tended to be a fabrication problem for all *The ine_tallographic preparations were made by M. D. Allen of General Metallography. - Table 20.4. Metal-Bonded Carbide Coatings on Hastelloy N Cylinders 1 in. in Diameter by 1 in. Long Ground Surface Finish? Phases Identified by besignatidi\ Nominal Composition of Coating (emin) X-Ray Analyses® Remarks on Microstructure® Linde LW-5 25% tungsten carbide and mixed - 14-23 WC,CrsCy W-Cr carbides plus 7% nickel - : , LW-1 - Tungstencarbideplus7-10% - 0.6-9 - WC, 6:W,,C, Co3C : : cobalt . : : . Metco 81NS 75% chromium carbide plus 25% 30-63 Cr7C3 Cr,C.Cr3Ca, nickel-chrome alloy : _ 7 Ni-42 at. % Cr alloy MTI o 84% tungsten carbide—10% 42-125 oW, C, Co3C, Mo, MoC, Mo,C = molybdenum plus 6% cobalt K _ There is some scattered oxide at interface of Hastelloy N and coating with essentially . none in the coating. Most carbide particles are massive. One cannot distinguish between the WC and Cr3C, in the as-polished state, Some small cracks observable in carbide " particles appear to be in binder, The binder is a minor phase. The ground sur- face is smooth, There is no difference in the control and tested specimen. . There is some scattered oxide at the inter- face of the coating and Hastelloy N. There is a small amount of oxide, more than in - Linde LW-5, and the carbide particles appear to be thin, wavy layers parallel with the Hastelloy N surfaces. There are stringers - of small carbide-like particles in the binder. These contribute to this having a -rougher surface than Linde LW-5. There is, essentially, no difference in the control and tested specimen. There is scattered oxide at the mtcrface of the coating and the Hastelloy N, The carbides appear to have been plastic. There is oxide throughout the coating with some forming lenticular layers between the carbides. The carbides have cracks in them. The original surface is rough due to cracks and porosity. There is no difference in the contro! and tested specimen. There is scattered oxide at the interface between the Hastelloy N and coating. It is also throughout the coating. The coating is very porous. The tested coating appeared to fall apart more than the control, which suggests that the former may have been 9The range of values as determined on 12 specimens by Mechanical Technology Inc. weakened by the thermal cycling. bReported for reference; reported previously by 1. E. Epperson, MSR Program Semiann. Progr. Rept. Aug. 31 1968, ORNL-4344, pp. 285-86. ©The “control™ refers to as-received specimens, and “tested specimens”™ refers to those that were subjected to 100 thermal cycles between 100 and 700 C in which the heating and cooling required 20 and 40 min respectively, n ~] w 274 PHOTO 99452 (port 1) (g-1) ‘ (CONTROL) NO. %2 (o0-2) (TEST) | NO. 5 LW-5 (NOMINALLY: 25% TUNGSTEN CARBIDE AND MIXED W-Cr CARBIDES PLUS 7% NICKEL) (b-1) (CONTROL) NO.7 (£6-2) TEST NO. 6T LINDE LW-{1 (NOMINALLY: TUNGSTEN CARBIDE PLUS 7-10% COBALT) Fig. 20.8. The Microstructures of 0.003-in.-Thick Metal-Bonded Carbides on l-in.-diam by l-in-Long Hastelloy N Cylinders a-1, b-1, c-1, and d-1 (Controls, As Received) and a-2, b-2,c-2, and d-2, After 100 Thermal Cycles Between 100 and 700°C. As polished. 500X. ' (c-1) (CONTROL) METCO 81 NS (NO 275 PHOTO 99452 (part 2) NO. 5 | NO.7 (c-2) (TEST) MINALLY: 75% CHROMIUM CARBIDE PLUS 25% NICKEL-CHROMIUM ALLOY) (g=1) (CONTROL) , (TEST) MTI (NOMINALLY: 84% TUNGSTEN CARBIDE-10% MOLYBDENUM PLUS 6% COBALT) NO. 7 (o-2) NO. 6 N.&_.._.._._...__-__,.-._____... four coatings, ranging from just detectable to easily detectable with the microscope in the order listed in 276 Table 20.4. The presence of the oxides in the coatings makes it even more important that these be tested for their compatibility with molten fluoride salts at 700°C, because past experience has shown that oxides are normally attacked. The removal of the oxides might seriously. weaken such coatings as Metco 81 NS and MTI, which have the larger quantities of oxide con- tamination. The presence of oxides in these coatings helped create this fault, but it cannot be discounted. The detail on this anomalous fault is warranted in that it illustrates the problems-that one must cope with in ensuring the absolute integrity of coatmgs for valves and bearing surfaces. _ The characteristics of these coatmgs as currently ~ determined, desplte some shortcomings, are encourag- could be prevented by better control of the atmosphere dunng coating. A “blister” fault as shown in Fig. 20.9 developed on one-out of the six specimens for Linde LS-1 after the 100 thermal cycles test. The pushed-up region of “blister” in Fig. 20.9b appears to have separated in the coating in our first examination, and there are three gray oxidedike particles in the break. It seems remote that ratcheting of the oxides during the thermal cycles - Fig. 20.9, A Raised or “Blister” surface, and (b) a longxtudmal section (through the short diameter of the crescent traced by the cracks). Both 100X. Fig. 209 isa metallographxc section. 0.C30n. ing. However, a more comprehensive program is re- quired to determine their overall potential. These and improved coatings (such as being oxide free) must be thoroughly examined on the basis of microstructure, development of improved better NDT evaluations, thermal cycling resistance, heat treating (which is normally higher than the 700°C maximum operating temperatures), compatibility in molten fluoride salts, design, and bearing tests in molten salt, involving boundary (rubbing) and hydrodynamic lubrication tests. PHOTO 99453 Defect in Linde LW-1 Specimen No. 5. (¢) Unshadowed appearance on the ground bearing & Part 6. Molten-Salt Processing and Preparation Part 6 deals with the development of processes for the isolation of protactinium and the removal of fission products from molten-=salt reactors. This reporting period has seen a significant revision in the conceptual flowsheets that form the basis for our development effort. The reductive extraction scheme for the removal of rare earths from the reactor fuel salt was the weak required exploitation of separation factors between thorium and the rare earths in the range of 1.2 to 3.5 in the fluoride-salt—bismuth system. The large electrolyzer for generating the extractant was required to operate in an undesirable mode, with very low flows of electrolyte and virtually complete removal of the thorium from the salt. The discovery that the thorium distribution coef- ficient in an LiCl-bismuth system is high and that the rare-earth distribution in the same system is low (a separation factor of the order 10° or better) has allowed us to turn to a metal transfer process that is much more-tractable and does not require an electro- lyzer. The major part of our effort in chemistry during this period was devoted to the definition of the distribution of the system components between the LiCl acceptor salt phase and a reducing bismuth phase. Analysis of flowsheets using this new concept has led reductive extraction step. Calculations made for the combination of the fluorination—reductive-extraction and the metal transfer flowsheets promise good system - performance. - link in processes proposed for the MSBR, because it Our development program has not, however, requlred ' drastic revision, since many of the operations required for the new processes are the same as those required for the old ones. Our work on contactor development has continued both with the flow-through system using bismuth and salt and with the mercury-water simulation “of the salt-bismuth system. Our axial mixing studies with the simulated system have led to a design of back- mixing preventers which seems to be quite effective at flow ratios (dispersed phase/continuous phase) of 50 or higher. We have increased our efforts to develop a continuous fluorinator that will be protected from corrosion by layers of frozen salt deposited on its metal surfaces. The work on electrolytic cells has been - deemphasized, but not terminated, since an onstream us to consider alternative protactlmum removal methods which do not require an electrolyzer. The obvious and most desirable alternative is to return to fluorination for removing the uranium before the method of producing lithium-bismuth solutions would be very useful and since the old reference protactinium flowsheet would still be a very attractive alternative if an operable cell were available. - We have initiated work toward the development of a continuous process for purifying reactor-grade salt for both an initial reactor charge and for our experimental - program, We have also assisted the MSRE operation by preparing capsules loaded with 23°PuF; for use in " adding 23°Pu to the fuel. . 21. Flowsheet Analys1s 211 RARE-EARTH REMOVAL USING THE METAL TRANSFER PROCESS ' L.E. McNeese We have devised a new process called the metal transfer process for removing rare-carth and alkaline- earth fission products from the fuel salt of a single-fluid - MSBR. In tlus process, blsmuth contammg thorium and 277 lithium is used to transport the rare-earth fission products from the reactor fuel salt to an acceptor salt. Although LiCl is the preferred acceptor salt, LiBr or LiCl-LiBr mixtures could also be used. Both thorium and rare earths transfer to the bismuth; however, because of favorable distribution coefficients, 278 only a small fraction of the thorium transfers with the rare earths from the bismuth to the LiCl. The effective thorium—rare-earth separation factors for the various rare earths range from about 10% to about 108, The final step of the process is removal of the rare earths from the LiCl by extraction with bismuth containing 0.05 to 0.50 mole fraction lithium. The new process does not require an electrolytic cell. This is an important advantage over the earlier reductive extrac- tion process, which also had the disadvantage of rare-earth—thorium separation factors near unity. The conceptual process flowsheet (Fig. 21.1) includes - four contactors that operate at about 640°C. Fuel salt from the protactinium isolation system, which is free of uranium and protactinium but contains the rare earths at the reactor concentration, is countercurrently con- tacted with bismuth containing approximately 0.002 ORNL DWG TO-4517 PROCESSED ' SALT TO ——— REACTOR : o t EXTRACTOR 1§ s 3-6 STAGES : ! FUEL SALT ) b FREE OF | ' Bi U AND Pa 0 : | ' ! ) _ ) EXTRACTOR 2 ! 3-6 STAGES : 1 S Lo __4 -——— Bi-Li r (0.5 MOLE FRACTION Li) * EXTRACTOR 3 i 2-3 STAGES Lic i L — — — & Bi-~Li 1 + DIVALENT F——=—= RARE EARTHS l . bt - e = Bi-Li F | {0.05 MOLE EXTRACTOR 4 |} FRACTION L) 1 STAGE : g ke o Bi-Li t ¢ +TRIVALENT -—- RARE EARTHS ' Fig. 21.1. Metal Transfer Process for Removal of Rare Earths from a Single-Fluid MSBR. mole ffaction lithium and 0.0025 mole fraction tho- fum (90% of thorium solubility) in contactor 1. Fractions of the rare earths transfer to the downflowing ~ metal stream and are carried into contactor 2. Here, the bismuth stream is contacted countercurrently with LiCl, and fractions of the rare earths and a trace of the thorium transfer to the LiCl. The resulting LiCl stream is routed to contactor 4, where it is contacted with a ~ bismuth solution having a lithium concentration of 0.05 ‘mole fraction for removal of trivalent rare earths. About 2% of the LiCl is routed to contactor 3, where it is contacted with a bismuth solution having a lithium concentration of 0.5 mole fraction for removal of divalent rare earths (samarium and europium) and the alkaline earths (barium and strontium), The LiCl from contactors 3 and 4 (still containing some rare earths) is then returned to contactor 2. Calculations were made to identify the unportant system parameters. It was found that there is consider- able latitude in choosing operating conditions which -will yield a stated removal time. The number of stages required in the contactors is low: less than six in contactors 1 and 2, three or less in contactor 3, and one in contactor 4. The process appears to be essentially insensitive to minor variations in operating conditions such as flow ratios, reductant concentrations, and temperature. The required salt and bismuth flow rates depend on the desired rare-earth removal times. Conditions will first be considered which give rare- earth removal times equal to or somewhat shorter than removal times presently assumed in reactor evaluation calculations which assume a 50-day removal time for all rare earths except europium, which has a 225-day removal time, Table 21.1 shows the effect of number of stages (V) in contactors 1 and 2 for a typical trivalent rare earth (Janthanum) and a typical divalent rare earth (europium). Operating conditions include a 3-day proc- essing cycle (3 gpm salt flow), a bismuth flow rate of 8.4 gpm through contactors 1 and 2, and an LiC] flow Table 21.1. Variation of Rare-Earth Removal . Time (in Days) with Number of Stages in -Contactors 1 and 2 for Typical Trivalent and Divalent Rare Earths ‘ Removal Time (days) for N% of — Rare Earth -1 2 3 4 5 6 La 52.1 380 340 321 31.0 30.2 Nd 58.0 427 384 36.5 354 34.8 Eu 2224 2196 2194 2194 2194 2194 @N = number of equilibrium stages in each of contactors 1 and rate of 11.2 gpm. Contactor 4 was assumed to consist of one stage, and it was assumed that a bismuth-lithium solution (0.05 atom fraction lithium) was fed to the contactor at a rate of 22.7 liters/day. Contactor 3 was not used. Thus with two or three stages in contactors 1 and 2 and with volumetric flow ratios near unity, one can obtain satisfactory removal times with relatively small processing streams for a three-day processing cycle, If packed columns were used as the contactors, the diameters would be approximately 6 in. Since the number of stages needed is low, the use of mixer- settlers should not be dismissed, particularly for experi- ments aimed at demonstrating the process concept. The rate at which ’Li is removed from the system in the withdrawal stream is 52.4 g-moles/day. The rate at which thorium is removed is less than 0.069 g-mole/ day. If the bismuth withdrawal stream were hydro- fluorinated or hydrochlorinated in the presence of a 279 suitable waste salt in order to recover only the bismuth, the cost of the 7Li lost would be 0.0018 mill/kwhr. It is apparent that one can obtain removal times much shorter than those discussed previously by increasing the throughput and hence the size of the processing plant. A three-day processing cycle will be retained in the following cases; however, other flow rates will be increased to 10 times the values discussed above. Thus the bismuth flow rate in contactors 1 and 2 will be 84 gpm, the LiCl flow rate 112 gpm, and the bismuth- lithium solution flow rate 16 ft3/day. The effects of number of stages in contactors 1 and 2 and of the concentration of lithium in the bismuth- Rare Earth Table 21.2. Variation of Rare-Earth Removal Time (in Days) with Lithium Concentration in Contactor 3 and Number of Stages in Contactors 1 and 2 Removal Time (days) for N% of —- 1 2 3 ° 4 5 6 Li Concentration in Contactor 3 = 0.05 Mole Fraction La 775 526 443 403- 379 3.64 Nd 836 585 505 467 446 434 ' Eu 166 147 141 138 137 136 Li Concentration in Contactor 3 = 0.20 Mole Fraction La 759 5.10 428 387 363 348 Nd 822 571 491 454 433 420 Fu 871 681 623 598 58 5.79 Li Concentration in Contactor 3 = 0.50 Mole Fraction La 759 510 427 3.87 363 348 Nd 822 571 491 453 433 420 Eu 827 637 579 554 542 5.35 4N = number of equilibrium stages in each of contactors 1 and of 0.0115. It appears that significant increases in breeding ratio are within reach if reasonable means are available for removing uranium and protactinium from the fuel salt prior to removal of the rare earths and if ‘means for removing the rare earths from LiCl are lithium solution in contactor 4 are shown in Table 21.2. It is seen that rare-earth removal times of 3.5 to 15 days can be obtained depending on the conditions chosen. It should be noted that one cannot afford to discard lithjum at the rate associated with this mode of operation. Hence, a means for recovering lithium from the - discard stream from the contactor would be required. An alternative would be removal of the rare earths from the LiCl by other means such as contact with zeolitic exchangers. ‘Variation of breeding ratio vfith rare-earth removal time was calculated and is given in Fig. 21.2. The changes are given with respect to the base case, which has a removal time of 50 days for all rare earths except available which do not result in discard of large quantities of lithium. 21.2 PROTACTINIUM ISOLATION USING FLUORINATION-REDUCTIVE EXTRACTION L. E. McNeese The fact that the new process for removing rare earths from a single-fluid MSBR does not require an electro- Tlytic cell has led us to consider protactinium isolation europium, which has a 225-day removal time. It should be noted that if all rare earths were removed on a five-day cycle, an increase in the breeding ratlo/ of 0.0126 would result; decreasing the removal time to - three days would result in an increase of 0.014, and a removal time of seven days would result in an increase systems that do not require electrolyzers. One protac- tinium isolation method by which an electrolyzer can be avoided is fluorination—reductive extraction. In this method most of the uranium would be removed from the fuel salt by fluorination prior to isolation of the protactinium by reductive extraction. . The fluorination-reductive-extraction system for iso- lating protactinium is shown in its simplest form in Fig. 21.3. The salt stream from the reactor first passes through a fluorinator, where most of the uranium is removed by fluorination. Approximately 90% of the salt leaving the fluorinator is fed to an extraction column, where it is countercurrently contacted with a 0l6 280 ORNL DWG. €9-12121 Oi4 - 012 T oo .068 - 006 - , l.004 ~ 002 - INCREASE IN BREEDING RATIO -004 1 i N I | .OO| 3 REFERENCE CASE/ 225 DAY Eu REMOVAL { i 1 .t {1 1 1 4 S 11 1 1 6 7 & 9 i 10 20 30 40 50 €0 70 80 90 100 RARE EARTH REMOVAL TIME, DAYS Fig. 21.2, Effect of Rare-Earth Removal Time on Breeding Ratio. ‘bismuth stream containing lithium and thorium. The uranium is preferentially removed from the salt in the lower extractor, and the protactinium is removed by the upper contactor. A tank through which the bismuth flows is provided for retaining most of the protactinium - in the system. ' The bismuth stream leaving the lower contactor contains some protactinium as well as the uranium not removed in the fluorinator and the uranium produced from - the ‘decay of protactinium. This stream is con- tacted with an H,-HF mixture in the presence of approximately 10% of the salt leaving the fluorinator in order to transfer the uranium and the protactinium to the salt. The salt stream, containing UF, and PaF,, is then returned to a point upstream of the fluorinator, where most of the uranium is removed. The protac- tinium passes through the fluorinator and is sub- sequently extracted into the bismuth. Reductant (lith- ium and thorium) is added to the bismuth stream leaving the oxidizer, and the resulting stream is returned to the upper contactor. The salt stream leaving the upper contactor is essentially free of uranium and protactinium and would be processed for removal of - rare earths before being returned to the reactor. Calculations have shown that the system is quite stable with respect to variations as large as 20% for concentrations, and number of extraction stages. The required uranium removal efficiency in the fluorinator is less than 90%. The number of stages required in the extractors is relafively low, and the metal-to-salt flow ratio (about 0.26) is in a range where the effects of axial mixing in packed column extractors will be negligible. Since the protactinium removal efficiency is very high and the system is quite stable, materials such as 231Pa, Zr, and Pu should accumulate with the 233py. These materials can be removed by hydro- fluorination of a small fraction of the bismuth stream leaving the lower extractor in the presence of salt which is then fluorinated for uranium recovery. Because sufficient decay time would be allowed for most of the 233p; to decay to 233U, the 233Pa and %23V losses would be acceptably low. Operating conditions that will yield a ten-day protac- " tinium removal time include a fuel salt flow rate of 0.88 most of the important parameters: flow rates, reductant gpm (ten-day processing cycle), a bismuth flow rate of 0.23 gpm, two stages in the lower contactor and six to eight stages in the upper contactor, and a decay tank volume of 200 to 300 ft3. The required quantity of reductant is 340 to 420 equivalents/day, which will cost 0.012 to 0.015 mill/kwhr if "Li is purchased. The remainder of this section is devoted to a discussion of the effects of important system param- eters. Unless otherwise stated, the operating conditions 281 ORNL -DWG-69-13169 EXTRACTOR DECAY (8i) REACTOR EX TRACTOR OXIDIZER REDUCTANT ADDITION SALT CONTAINING UF; AND Pafy LI, Th Fig. 21.3. Protactinium Isolation by Fluorination — Reductive Extraction. include an operating temperature of 640°C, a process- ing cycle of ten days, a reductant addition rate of 429 equivalents/day, a protactinium decay tank volume of 1300 ft?, a maximum thorium concentration equivalent to 50% of the thorium solubility at 640°C, and a uranium removal efficiency during fluorination of 98%. * The effect of number of equilibrium stages in the upper extractor is shown in Fig. 21.4. Less than two stages are required in the lower contactor; however, more stages can be used to advantage in the upper contactor. Approxlmately five stages are adequate, and - little benefit is obtained from use of additional stages. Values of the uranium inventory in the decay tank, which is approxnmately 1% of the reactor mventory, are also shown. : The effect of the fraction of uranium not removed during fluorination is shown in Fig. 21.5. High uranium removal efficiencies are not required, since there is no _protactinium removal time is shown in Fig. 21.8. No effect if as much as 12% of the uranium remains in the salt. : - The effect of reductant addition rate on protactinium removal rate is shown in Fig. 21.6. No effect of - decreasing the addition rate is seen until one reaches a value of 390 equivalents/day, after which a gradual but - significant effect is observed. The changes shown are also approximately the changes which result from variations in the bismuth flow rate for a constant inlet reductant concentration. Hence, there is little effect of variations in the bismuth flow rate. , The effect of processing cycle time on protactinium removal time is shown in Fig. 21.7. The removal time is equal to the processing cycle time (100% removal efficiency) for processing cycle times greater than about ten days.- As the processing cycle time is decreased below this point, the system loses efficiency slowly. This figure shows that the system is relatively insensi- tive to minor variations in the salt feed rate. - The effect -of protactinium decay tank -volume on change is seen as the tank volume is decreased from the 282 ORNL DWG 69-13174 16 ' l | | - ’ o g 15 —10 = S o - e e v 14 —0.8 3 * -8 i w X y TEMPERATURE €40 °C 2 = PROCESSING CYCLE 10 DAYS = 4 13 Li REQUIREMENT 429 MOLES Li/pay —{06 g DECAY TANK VOLUME 300 ft.3 O o METAL TO SALT FLOW RATIO 0.26 & oo 2 o l_ > e ¥ z 11— —10.2 5 = <1 @ = 10 | l | ! | 0 -3 4 5 6 7 8 9 10 NUMBER OF STAGES IN UPPER COLUMN Fig. 21.4, Effect of Number of Stages in Upper Extraction Column on Protactinium Removal Time and Uranium Inventory in Protactinium Decay Tank. nominal tank volume of 300 ft* until a value of 280 ft3 is reached, after which the system loses efficiency slowly. Smaller tank volumes could be used without loss of removal efficiency by increasing the reductant concentration in the bismuth fed to the extractors. The effects of operating temperature and reductant addition rate on protactinium removal time are shown in Fig. 21.9. The optimum operating temperature is about 640°C, and the system is relatively insensitive to minor temperature variations. The reductant costs associated with the addition rates shown decrease in increments of 0.001 mill/kwhr from a value of 0.015 mill/kwhr for an addition rate of 429 equivalents/day. 21.3 MSBR PROCESSING BY - FLUORINATION--REDUCTIVE EXTRACTION AND THE METAL TRANSFER PROCESS L.kE. McNeese ~ We have now adopted for study a;floWsheet that uses - fluorination—reductive extraction for protactinium iso- lation and the metal transfer process for rare-earth removal. These processes were described in detail in earlier sections of this report; this section will be limited to a description and an analysis of the flowsheet shown in Fig. 21.10. ' Salt withdrawn from the reactor is fed to a fluori- nator, where most of the uranium is removed as UFg. Most of the salt leaving the fluorinator is fed to a reductive extraction column, where the remaining uranium is removed and the protactinium is extracted into a bismuth stream. The bismuth stream containing the extracted protactinivm flows through a tank of sufficient volume to contain most of the protactinium in the reactor system. Most. of the bismuth stream leaving the extraction column is contacted with an - H,-HF mixture in the presence of about 10% of the salt leaving the fluorinator in order to transfer materials such as uranium and protactinium to the salt stream. This salt stream is then recycled to the fluorinator. The bismuth stream leaving the lower column also ‘contains several materials that must be removed for satisfactory operation of an MSBR. The most important of these are fission product zirconium, which canbean important neutron- absorber, and corrosion ‘ product nickel, which forms an intermetallic nickel-thorium compound having a low solubility in bismuth. These materials and others that do not form volatile fluorides ) 283 ORNL DWG 69-13179 T 1T T 17 T 1T T T T 7 7 > S 7 —Ji1.e F = fas _ > z 16 x —1.2 o 4 < - § & © i5|— —1.0 g & = 3° = N °. 3 — X g M- o —8 Z o TEMPERATURE 640°C = * DECAY TANK VOLUME 300 {13 4 X @ STAGES IN UPPER COLUMN € o g 13- STAGES IN LOWER COLUMN 2 -6 © METAL TO SALT FLOW RATIO 0.26 z REDUCTANT FEED RATE 428.6 eq/day - > . o [»] ' ra— - 2F 4 £ ) W > - z " 2 3 : Zz i < —— @ . ) . =2 ob—t 1t 1 1 | 4 01 g, 0 .02 .04 .06 .08 10 A2 A4 FRACTION OF U NOT‘REMOVED 8Y FLUORINATION Fig. 21.5. Effect of Fraction of Uranium Not Removed by Fluormauon on Protactuuum Removal Time and Uranium Inventory in Protactinium Decay Tank. during fluorination are removed by hydrofluorination, in the presence of a salt stream, of a small fraction of ~the bismuth stream exiting from the lower column. The salt is then fluorinated for removal of uranium. Sufficient time is allowed. for the decay of 233Pa so that the rate at which this material is lost is acceptably low. The remaining materials, including Zr, Ni, 23! Pa, and Pu, are withdrawn in the salt stream from the - fluorinator. Oxidation of part of the metal stream leaving thé lower contactor is chosen as a means for removal of these materials, since this results in discard of no beryllium and very little lithium or thorium; - discard of salt from other points in the system would result in much higher removal rates for the major components LiF, BeF, ,and ThF,. : The bismuth streams leaving the hydrofluonnators are then -combined, and sufficient reductant (lithium and thorium) is added for operation of the protactinium isolation system. Effectively, this stream is fed to the upper column of the protactinium isolation system; actually, it first passes through a captive bismuth phase in the rare-earth removal system in order to purge uranium and protactinium from this captive volume. The salt stream leaving the upper column of the protactinium isolation system contains negligible amounts of uranium and protactinium but contains the. ~ rare earths at essentially the reactor concentration. This stream is fed to the rare-earth removal system, where fractions of the rare earths are removed from the fuel carrier salt - by countercurrent contact with bismuth ‘containing lithium and thorium. The bismuth stream is then contacted with LiCl, to which the rare earths, along with a negligible amount of thorium, are trans- ferred. The rare earths are then removed from the LiCl 284 -~ ORNLDWG 69-13(78 ‘1 T T T T T T T T T % __lfi — 1.0 g 0. < < [~ e a ‘im B TEMPERATURE €40°C ° - o8- = DECAY TANK VOLUME 300 f13 oZ = | STAGES IN UPPER COLUMN 10 d 2% 5 - STAGES IN LOWER COLUMN 2 _ =2 § PROCESSING CYCLE 10 DAYS E“" ail- o : —~0.655 = = w ot o - - w . >} a Za :O S8 = = < a = nl o2~ T — " ol 1 1 14 1S L, b, . 300 320 340 360 380 400 420 440 REDUCTANT FEED RATE {eq/DAY) Fig. 21.6. Effect of Reductant Feed Rate on Protactinium Removal Time and Uranium Inventory in Protactinium Decay Tank. by contact with bismuth containing a high concentra- tion of 7Li, Separate contactors are used for removal of - the divalent and trivalent rare earths in order to minimize the quantity of ?Li required. Only about 2% of the LiCl is fed to the contactor in which the divalent materials are removed. _ , Calculations have been made for a range of operating conditions in order to evaluate the flowsheet just described. In making these calculations the MATADOR code was used to determine the reactor breeding ratio for each set of processing plant operating conditions examined. Data are not. available on the cost of processing for this flowsheet or for the reference flowsheet for the processing system that uses electro- lyzers in both the protactinium and rare-earth removal systems. In the absence of these data, we examined processing conditions which would result in the same reactor performance (ie., the same breeding ratio) as “that obtained with the previous reference flowsheet. - Although the optimum operating conditions which will result in a breeding ratio equal to that of the . reference reactor and processing system (1.063) have not been determined, the following conditions are believed to be representative. The reactor was processed on a ten-day cycle, with the complete fuel salt stream (0.88 gpm) passing through both the protactinium isolation system and the rare-earth removal system. The resulting protactinium removal time was ten days, and the reductant requirement was 350 to 430 equivalents/ day, which costs 0.012 to 0.015 mill/kwhr. The protactinium isolation columns were less than 8 in, in diameter, and the total number of required stages was less than 10. The protactinium isolation system also ~ resulted in a ten-day removal time for materials that are more noble than thorium but do not have volatile fluorides. These include zirconium, 23!Pa, plutonium, the seminoble metals, and corrosion products. The rare-earth removal system consisted of three - primary contactors: (1) a 7.1-in.-diam, six-stage column in which the rare earths are transferred from the fuel salt to a 12.5-gpm bismuth stream, (2) a 13-in.-diam, six-stage column in which the rare earths are transferred from the bismuth to a 33.4-gpm LiCl stream, and (3) 2 12.34in.-diam column in which the trivalent rare earths O ORNL DWG €9-13176 T T T T 28— 26— 24— N N I Pa REMOVAL TIME (days) S | TEMPERATURE 640° C DECAY TANK VOLUME 300 #t3 STAGES IN UPPER COLUMN 6 STAGES IN LOWER COLUMN - REDUCTANT FEED RATE 428.6 eq/day | T T 2 16 ] L& ] URANIUM INVENTORY IN DECAY TANK, % OF REACTOR INVENTOR® e —3 14-— - iz-— —:.I |o°- — ' ; VV . ls é.. V'. |lo - |Iz l -;40 4. _ PROCESSING CYCLE TIME (days) - Fig. 21.7. Effect of Processing Cycle Time on Pmt:_:cfiniufn Removal Time and Uranium Inventory in Decay Tank. are transferred from the LiCl to an 8.1-gpm bismuth stream having a lithium concentration of 0.05 mole fraction. Two percent of the LiCl (0.69 gpm) was con- . ‘tacted with a bismuth stream (1.5 ¢cm®/min) having a lithium concentration of 0.5 mole fraction for removal of the divalent fission products such as Sm, Eu, Ba, and Sr. The total lithium consumption rate-for the rare- earth system was 119 moles/day, which costs 0.0042 mill/kwhr. o : The rare-earth removal times ranged from 15 5 days for cerium to 50.4 days for europium. The distribution data for neodymium, which are believed to be conserva- ~ tive, were used for rare earths for which distribution data were not available (i.e., Y, Pr,Pm). The costs for reductant in both the protactinium isolation system and in the rare-earth removal system - constitute only a small fraction of the total processing costs; however, they indicate that one can purchase - reductant rather than use an electrolytic cell for producing this material. As data become available on - processing costs, we will determine operating conditions that result in the most economic operation of the processing plant. 286 ORNL DWG 65-13177 24 9>~ @ o - O - , Z 22_ — w > z - 71 x 1 8 _ 20— ~0Q » W : P - m s w w _ o ¥ 18— —J8 3 - X 3 - 4 2 x 4 S TEMPERATURE 640 °C - 3 e REDUCTANT FEED RATE 428, € aq/dayle > = TAGES IN UPPER COLUMN 10 ‘ g w STAGES IN LOWER COLUMN 2 o1 o L PROCESSING CYCLE 10 days _ o & ‘ z 14— ez Qo - - -4 Z W Z 12— —_i2 = & L_ 2 4 Z - & o a0 b b g 0 180 200 220 240 . 260 280 300 320 Pa DECAY TANK VOLUME, tt3 Fig. 21.8. Effect of Protactinium Decay Tank Volume on Protacumum Removal Time and Uranium lnventory in Protactuuum ‘ Decay Tank. 1 287 ORNL DWG &9-13180 T T T T PROCESSING CYCLE 10 DAYS - 9 STAGES N UPPER COLUMN 40 STAGES IN LOWER COLUMN 2 {8 DECAY TANK VOLUME 300FT3 17 16 15 343 eq REDUCTANT /DAY Po REMOVAL TIME, days 3 13 12 " 10 [ _ 600 60 60 630 640 650 660 60 TEMPERATURE - Fig. 21.9. Effects of Operating Temperature and Reductant Addition Rate on Protactinium Removal Time. 288 ORNL DWG 70-2833 saLt || URe PROCESSED SALT PURIFICATION || REDUCTION , + DIVALENT RARE EARTHS ‘ FLUORINATOR Fo 7] te-—=siu | ot EXTRACTOR | (005 MOLE I |-— FRACTION Li) IT,-I — —»Bi-Li L _ J +TRIVALENT o “""'I"" =1 I T ' I EXTRACTOR ! | H2 i : SALT CONTAINING RARE EARTHS ' : 1 5 e mm———— — e e o e —— ——— -{ |-.— ——————— = UFg | | : COLLECTION : : ! f "—"l | exTracToR : : : EXTRACTOR I71 li : | |1_J _J | 1 : ) L -—- Bi-LI : REACTOR ] 16 ——| DECAY Liel ._I'L"l FRACTION L) : , ! i | I L'] ‘ EXTRACTOR ‘ I - ‘ | Il EXTRACTOR UFg ' __er‘ : i b — —&Bi-Li { I I i ) i I I | REDUCTANT RARE EARTHS | SALT CONTAINING ADDITION " PaF, AND UF, 1 Ha-HF Li, Th Fig. 21.10. Flowsheet for Processing a Single-Fluid MSBR by Fluorination—Reductive Extraction and the Metal Transfer Process. 22. Measurement of Distribution Coefficients in Molten-Salt-Metal Systems L. M. Ferris The main effort during this reporting period was in * obtaining data in support of the development of a metal transfer process for the removal of rare-earth and other fission products from MSBR fuels. This process (which is described in detail in Sect. 21.1) involves the reductive extraction of rare earths and thorium from the fluoride fuel salt into liquid bismuth and the subsequent selective oxidative extraction of the rare ‘earths into an acceptor salt such as LiCl or LiBr. The ‘rare earths, and the trace of thorium present with them, are stripped from the acceptor salt by a lithium-bismuth solution in which the lithium concentration is at least 5 at. %. A rare-earth—thorium separation factor of at least 10* is possible by this techmque Distribution coeffi- cients, mole fraction of M in bismuth phase Dy = mole fraction of M in salt phase ¥ were measured at 600 to 700°C for several lanthanide and actinide elements and for barium using LiCl, LiBr, LiCl-LiF, and LiBr-LiF salt phases. The systems con- taining LiF were studied to determine the effect of fluoride on the process in the event that the acceptor salt becomes contammated with some fuel salt. - Expenments were also conducted to deterrmne the . extractability of barium from LiF-BeF, -ThF, solutions and to determine the solubility of bismuth in MSBR fuel salt. Some preliminary work with fluoride salts contalmng BiF; was done in support of the develop- ment of electrolytic cells for the reductwe extractxon process IM. E. Whatley et al, Nucl. Appl. Technol. 8, 170 (1970). 22.1 METAL TRANSFER PROCESS STUDIES ' F.J.Smith C.T.Thompson L.M,Ferris J.F.Land The equilibrium distribution of several lanthanide and actinide elements and of barium between liquid bis- muth solutions and the potential acceptor salts LiCl and LiBr has been determined. At a given temperature the distribution coefficients for each element could be expressed as log Dy, =nlog X ; +log Ky , in which Xj; is the mole fraction of lithium in the bismuth phase, » is the valence of the element in the salt phase, and log K§; is a constant. The apparatus and general technique have been described elsewhere.l»2 The two-phase salt-bismuth systems were contained in either molybdenum or mild steel crucibles. In experi- ments with lanthanum, neodymium, and europium, dried acceptor salt and bismuth were heated together to the desired temperature under an atmosphere -of pure argon; then a small amount of thorium was added to- remove residual oxide from the salt. Finally, a piece of the desired rare-earth metal was added to the system. The amount added was such 'that about 50% of the " rare-earth metal was converted to the halide and dissolved in the salt, while the remainder was retained in the bismuth. The rare earth in the salt was extracted 'L. M. Ferris et al., “Isolation of Protactinium from Single- " Fluid' Molten-Salt Breeder Reactor Fuels by Selective Extrac- tion into Li-Th-Bi Solutlons,” to be published in the Proc. Thtrd lntem Protactimum Conf 2 M. Ferris et al., “Equilibrium Distribution of Actinide and Lanthanide Elements Between Molten Fluoride Salts and Liquid Bismuth Solutions,” J. Inorg. Nucl. Chem. (in press). - 289 into the bismuth phase in increments by the periodic addition of lithium-bismuth alloy to the system. In some experiments, barium metal was added to the “system when extraction of the rare earth was practically complete; then the addition of lithium-bismuth alloy was continued. A period of at least 4 hr was allowed for equilibration after each addition of alloy. A filtered sample of each phase was taken after each equilibration. Analyses of these samples provided the basis for the 290 calculation of the distribution coefficients. In other experiments, bismuth and dry salt, along with the appropriate amounts of thorium and uranium metals (or samarium oxide), were loaded into a molybdenum crucible. After the two-phase system had been heated to the desired temperature, the system was treated with either an HCI-H, or an HBr-H, mixture to convert the metals or oxide to their respective halides (which dissolved in the salt) and to remove residual oxygen from the system. After treatment with pure hydrogen for several hours, the system was sparged with pure argon. Extraction of the thorium, uranium, and sa- marium from the salt was effected by the incremental addition of lithjum-bismuth alloy to the system using the procedure outlined above. In some experiments the thorium that was added to the crucible had been irradiated to provide about 1 mc of 233Pa. Gamma counting of samples from these experiments produced the protactinium data presented below. Some of the distribution coefficient data obtained in these studies are shown in Figs. 22.1 and 22.2. Data obtained for uranium, neodymium, and europium when LiCl was used as the salt phase are presented (Fig. 22.1) as plots of log D vs log X; ;. The slopes of the lines indicate that n is 3, 3, and 2 for uranium, neodymium, and europium respectively. Plots of log Dp vs log Dy . and of log Dy; vs log Dy, using data obtained with LiCl at 640°C are shown in Fig. 22.2. Since uranium was trivalent, the slopes of the respective lines show that both thorium and protactinium were tetravalent in the - salt under the experimental conditions employed. Values of log K* obtained thus far in our studies using a -variety of salts are compiled in Table 22.1. These results ‘show that, from a chemical viewpoint, LiCl and LiBr will be equally good as acceptor salts and that the distribution behavior of most of the elements will be rather insensitive to temperature changes. Inspection of the data in Table 22.1 reveals that the presence of fluoride in either LiCl or LiBr causes a change in distribution behavior, particularly that of thorium. This is illustrated by the data obtained for thorium and lanthanum'in LiCl-LiF and LiBr-LiF solutions at 640 and 600°C respectively (Fig. 22.3). ORNL DWG 70-4501 Nd3*- (640°C) DISTRIBUTION COEFFICIENT 5 | T N 102 3 Eul* ) ] (630°C) i 10'3 L A llllll! L 1 llJ__LI_I_l l 1 11-11-' 1 2 ot L ill 1% 104 1073 102 10! MOLE FRACTION LITHIUM IN BISMUTH PHASE Fig. 22.1. Equilibrium Distribution of Uranium, Neodymium, and Europium Between Molten LiCl and Liquid Bismuth Solutions. These data indicate that the thorium-lanthanumrsep‘ar_a- tion factor will be decreased if the acceptor salt becomes contaminated with fluoride fuel salt. The - effect of fluoride contamination of the acceptor salt on the efficiency of the metal transfer process is presently being analyzed. : S 22.2 EXTRACTION OF BARIUM, RARE EARTHS, AND THORIUM FROM SINGLE-FLUID MSBR FUELS J.C.Mailen F.J.Smith C.T.Thompson Measurement of the distribution df fission products and thorium between LiF-BeF,-ThF,; MSBR fuel salts and liquid bismuth solutions has been continued in ORNL. DWG T0-4502 - . ’0. Y v oo T rrrroeg T v iy T v rerrg rr rrry - - PROTACTINIUM {SLOPE =1) 3 ™ < ™Y URANIUM URANIUM AND PROTACTINIUM DISTRIBUTION COEFFICIENTS . {SLOPE=3/4) o't 3 e 3 n.! i L lllllll L A lljllll. i L Jllllll i L lIIIIlI L L L1 il w3 w0t 10! 1.0 0 102 THORIUM DISTRIBUTION COEFFICIENT Fig. 22.2. Distribution Coefficients for Thorium, Uranium, and Protactinium Obtmned with Molten LiC1 and Liquid Bismuth Solutions at 640°C. - support of both the reductive extraction process' and the metal transfer process (Sect. 21.1). Recently, we determined distribution coefficients for banum at 600 to 690°C using LiF-BeF,-ThF, (72-16-12 mole %) as ~ the salt phase. The data at each temperature could be - expressed as. ‘ - log DBa =2 log)(Li +log K*. The values obtamed for Iog K* are as follows Temperature logK* - Co e 600 - 4.049 645 3678 - 690 - 3530 The distribution coefficients measured for barium 'are‘ very nearly the same as those obtained previously? for europium; thus the behavior of these two elements in the reductive extraction process will be nearly identical. 291 Table 22,1, Values of log K* Derived from Distribution Coefficient Data log D =nlog Xy ; +log K* Temperature ©0) Salt Element logK* 630 LiCl Eu? 2.301 640 LiCl Ba? 1.702 La* 7.973 Na* 8.633 Sm?* 2.886 Th% 15.358 Pa* 17.838 U 11.278 640 LiCI-LiF (98.1-1.9 mole %) Th% 13.974 640 LiCl-LiF (96-4 mole %) Th* 1290 : Pa¥ 147 - U 10.80 640 LiCl-LiF (90-10 mole %) La¥ 7.288 Th¥ 11.309 600 LiCI-LiF (80-20 mole %) La¥ 7.235 Nd¥ 7.644 : o Th* 10.964 640 LiCI-LiF (80-20 mole %) La* 7.124 ' Th% 10.629 700 LiCL-LiF (80-20 mole %) Nd>* 6.732 Th¥ 9.602 575 LiBr Ba? 1.497 600 LiBr Ba?* 1.443 La¥* 9.079 Nd¥* 8.919 Th* 16.16 640 LiBr La* - 8.266 Na¥* 8.834 650 LiBr Ba?* 1.358 - 700 LiBr Ba?* 1.316 Nd* 8.430 600 LiBr-LiF (90-10 mole %) La¥* 8.158 Th* 12.380 600 LiBr-LiF (80-20 mole %) La% 7.840 Th* 11.373 It was suggested® that the rare-earth—thorium separa- tion in the reductive extraction process might be -enhanced by substituting NaF for some of the LiF in the MSBR fuel salt. This was tested by measuring the distribution of lanthanum and thorium between LiF- NaF-BeF,-ThF, (62-10-16-12 mole %) and bismuth M. E. Whatley et al., NucL Appl. Technol. 8,170 (1970). - 2L_M. Ferris et al., “Equilibrium Distribution of Actinide and Lanthanide Elements Between Molten Fluoride Salts and Liquid Bismuth Solutions,” J, Inorg. Nucl. Chem. (in press). 3G. 1. Cathers, Chemical Technology Division, personal communication, September 1969. 292 ORNL DWG 70-4503 17 —TTT | L L l 1 l 1T 11 I_I_l T'EMP ELEMENT SALT (°C) A Le® LiBr-LiF " 600 16 O Lo® LiCI-LiF 640 - A Th* LiBr-LiF 600 ® Th** LiCI-LiF 640 - Log K* 0} , - 0 0.05 o4 045 0.2 MOLE FRACTION LiF IN SALT Fig. 22.3. Values of log X* Obtained for Lanthanum and Thorium Using LiBr-LiF and LiCI-LiF as the Salt Phase. solutions at 600°C. The equilibrium data obtained be expressed as : log Dy, =3log X;;+7.380 and | log Dy, = 4log Xy ; +9.568 . When the bismuth phase was saturated with thorium, the lanthanum and thorium distribution coefficients were 0.07 and 0.014, respectively, corresponding to a separation factor of 5. This separation factor is about ‘71ll|||l|1|||lj;||1||ti twice as high as that obtained using LiF-BeF,-ThF, (72-16-12 mole %) as the salt phase at 600°C. However, this increase in separation factor would probably be offset by a loss of neutrons in the reactor due to the presence of NaF in the salt. ' 22.3 SOLUBILITY OF BISMUTH IN SINGLE-FLUID MSBR FUEL SALTS L.M.Ferris J.F. I.and_ During the chemical processing of MSBR fuel by either the reductive extraction or the metal transfer process, the fuel salt will be contacted with liquid bismuth before it is returned to the reactor. Since nickel-base alloys are corroded by bismuth, the pres- ence of bismuth in the salt could cause serious damage to a reactor vessel constructed of Hastelloy N. There- fore a means for ensuring that the fuel salt being - returned to the reactor is free of bismuth must be devised. We have attempted to measure the solubility of ‘bismuth in LiF-BeF,-ThF, (72-16-12 mole %) as a preliminary step in the development of such a method. About 125 g of salt and 5 g of bismuth were heated to 700°C in a mild steel container under an atmosphere of purified argon. Filtered samples of the salt were taken periodically and were analyzed for bismuth by both a polarographic and a spectroscopic method. After 75 days at 700°C, the system was cooled to 600°C, where it was held for an additional 35 days. Although the analytical data obtained in this experiment are er- ratic and inconsistent (Table 22.2), it is obvious that the solubility of bismuth in the salt was low (<5 ppm) under the conditions used. The cause of the disparity between the two analytical methods is not known and must be found before meaningful data can be obtained. 22.4 STUDIES INVOLVING BiF, C.E.Schilling J.F.Land Bismuth trifluoride is expected to be produced at the anode in the electrolytic cells required for the reductive extraction process.! Therefore, information on the solubility of BiF; in molten fluoride salts and on the corrosiveness of salts containing BiF; is desirable. In preliminary experiments we have found that BiF; is soluble to the extent of at least 4 mole % both in LiF-BeF, (66-34 mole %) and in LiF-BeF,-ThF, (72-16-12 mole %) at 600°C. Graphite was the only IM.E. Whatley ef al., Nucl. Appl. Technol. 8, 170 (1970).- 293 Table 22.2. . Analytical Results Obtained in Experiment to Determine the Solubility of Bismuth in LiF-BeF,-ThF, (72-16-12 Mole %) Bismuth Analysis of Salt (ppm) T . Time at emperature °0) Temperature Sample Polarographic Spectrographic (days) Method Method 700 8 1 10 9.2 2 43 15 3 4.5 0.6 : 4 0.55 40 5 5.0 _ 6 0.5 56 - 7 2.0 2.8 8 2.0 2.0 75 9 <1 1.9 10 1.5 1.5 600 14 11 1.5 29 12 7.0 44 35 13 0.2 14 0.56 material tested that was not rapidly attacked by the BiF; dissolved in the salt solution. The BiF; used in these experiments was prepared from commercial BiF;, which had the approximate - composition BiF;-BiOF (62-38 mole %). Treatment of this material with gaseous HF, initially for about 18 hr at 150°C and then for 5 to 6 hr at 500°C, resulted in a product that contained less than 3 mole % BiOF. The weight increase during the HF treatment was in good agreement with that calculated from the initial and final oxygen analyses of the material. Other chemical analyses showed that the F/Bi atom ratio increased from about 2.3 to 3.0 during hydrofluorination. The product had a melting point higher than 700°C and did not attack graphite. When the HF treatment was conducted in a platinum vessel, a small amount of metallic bismuth was formed and the platinum became- very brittle, Embrittlement of platmum by BiF; has been observed previously.? Two experiments were conducted to obtain informa- tion on the solubility of BiF; in molten fluoride salts and to assess the corrosiveness of melts containing - dissolved BiF;. In the first experiment, BiF; and LiF-BeF, -ThF, (72-16-12 mole %) were heated under argon in a graphite container to 600°C. A solution that - was 20 mole % BiF; would have been produced if all the trifluoride had dissolved. Attempts to obtain - filtered samples of the liquid at 600°C were futile; the stainless steel samplers were so badly corroded in only _ 2D. Cubicciotti, J. Electrochem. Soc. 115, 1138 (1968). 1- to 5-min exposures that they would not function. Metallic bismuth was one of the corrosion products. When a nickel thermowell was immersed into the system, the bottom was corroded away by reaction with BiF; and/or by dissolution in liquid bismuth, and salt was ejected from the tube. The frozen salt had the apparent composition LiF-BeF,-ThF,-BiF, (68.1-14.7-11.5-5.7 mole %) and did not appear to contain metallic bismuth. The LiF/ThF, and LiF/BeF, mole ratios were 5.9 and 4.6, which are about the same as those in LiF-BeF,-ThF, (72-16-12 mole %). The graphite container showed no visible signs of attack and underwent no significant weight change. In the second experiment, LiF-BeF, (66-34 mole %) and enough BiF; to produce a solution 5.2 mole % (24 - wt %) in BiF; were loaded into a graphite vessel and were heated to 605°C under argon. After about 28 hr a filtered sample of the liquid was taken; then the temperature of the system was varied between 497 and 600°C, with filtered samples being taken at selected temperatures in this range. Molybdenum samplers were “used throughout this experiment. The system was exposed to molybdenum only during sampling (less than 10 min for each sample) and during temperature “measurement (when a molybdenum thermowell was introduced). As shown in Table 22.3, after a slight initial decrease the bismuth content of the solution remained constant at about 20 wt % (equivalent to about 4.1 mole % BiF;) throughout the first part of the experiment, even when the temperature was lowered to 497°C. In this part of the experiment, it was noted that the amount of metallic bismuth in the samples, al- 294 Table 22.3. Results of Experiment Made to Test the Solubility of BiF; in LiF-BeF; (66-34 Mole %) e Cumulative Time at ‘ Analyses of Cumulative Time Sample 7 Temperature Time of Indicated Samples (wt %) of Exposure Nod Co) Experiment Temperature U — to Molybdenum (hr) (hn) . Bi Mo (min) 1 605 ' 28.5 28.5 29.6 ' : 10 2 550 41.25 12,75 120.2 ' : 170 3 497 124.75 83.5 20.3 195 4 524 | 145.75 21.0 20.8 ' : 491 5 578 163.75 18.0 19.5 . 541 6 600 187.75 24.0 ' 13.3 2.22 613 600 - 196.0° 8.25% ' 1108? ~ %In chronological sequence. BThermowell was penetrated at this time, and experiment was terminated. though small, increased with the number of samples taken, that is, as the time of exposure to molybdenum increased. After the fifth sample had been obtained, the system was heated to 600°C and held there for 24 hr (Table 22.3). During this time the bismuth concentra- tion in the salt decreased to about 13%;the amount of molybdenum found in the salt (2.2%) was only about two-thirds of that expected for the reaction BiF; + Mo - MoF; + Bi. After the sixth sample had been taken, the molybdenum thermowell was immersed in the liquid, and the system was held at 600°C until the thermowell was penetrated (i.e., for about 8 hr), apparently by reaction with BiF;. The results of these preliminary studies lead to the following conclusions: (1) Oxygen that is present in BiF; as BiOF or similar compounds can be removed by hydrofluorination. (2) The solubility of BiF; in fluo- ride salts, particularly LiF-BeF,-ThF, (72-16-12 mole %), appears to be high enough for efficient operation of the electrolytic cell required in the protactinium isola- tion portion of the reductive extraction process.! (3) Graphite is the only material encountered that was not attacked rapidly by fluoride meits containing BiF;. 23. Engineering Development of Process Operations L. E. McNeese 23.1 REDUCTIVE EXTRACTION ENGINEERING STUDIES B. A.Hannaford C.W.Kee L. E.McNeese Reductive extraction studies with the low-carbon- steel flow-through system have been continued. The modifications to the system reported in the last semiannual report! were effective in improving the flow control of both the bismuth and the salt streams. These consisted, in essence, of an expanded chamber in the line which conducted salt from the column and a bypass with a freeze valve which could pass the bismuth trapped by the chamber to the line which conducted the bismuth stream from the column. Measurement of A1pMsR Program Semiann, Progr. Rept, Aug. 31, 1969, ORNL-4449, o _ pressure between two points in the chamber provides an indication of large quantities of bismuth draining through the chamber. By closing a freeze valve at the bottom of the chamber, it should also be possible to detect low bismuth entrainment rates. Four runs were attempted during this period, no one of which was free from difficulty, Mass transfer of iron by the bismuth continued to be a major problem, “despite the reduction in feed tank temperature to lower the concentration of dissolved iron in the bismuth feed. The transfer line in which most of the iron deposits have occurred is the bismuth drain line between the column and the bismuth sample — a line that is normally filled with bismuth. However, on one occasion iron deposits also completely restricted the salt line between the jackleg and the column; this line contained " bismuth only during periods of cyclic salt and bismuth - { 1 flow. Figure 23.1 shows a segment removed from this line following run 7. The restricted lines (3% in. in PHOTO 97584 2 1 J INCHES Fig. 23.1. Section of Salt Transfer Line Between Salt Jackleg and Column Inlet, Plugged with Dendritic Iron. Darkening of lower specimen occurred during examination, when it was accidentally exposed to air while hot. 295 diameter) were replaced with % -in.-djam tubing. Fol- | lowing run 7 we attempted to inhibit mass transfer of iron by adding sufficient Zircaloy-2 to the bismuth feed tank to give about 100 ppm of zirconium in the bismuth, o The results from run 5, which is the most successful ~ run to date, are shown in Figs. 23.2 and 23.3. The run 296 was started with a bismuth flow of 90 ml/min; the salt flow was started about 30 min later. The pressure at the- bottom of the column, as measured in the salt jackleg, became steady about 5 min after the salt jackleg was filled and remained steady for about 12 min, Then - there followed in succession: a gradual pressure rise, another 12-min steady period, another pressure rise, ORNL-DWG 70-2828 6 r I 1 ] T l ¥ - [ T ' ¥ jfi ¥ ,7 v ‘_l_ T ‘ : i * BISMUTH + | 3 + SALT , . . k .r 5) . L i 4 L -l 5 2 . 8 . . £ + + w sl N A | ) + s + 3 1 N P\J wat + \ ~ o = - o . \ ‘ . 1 e . P S [ 0 20 40 60 80 100 120 140 160 180 200 TIME (min) Fig. 23.2, Flow Rate of Bismuth and Salt from Feed Tanks During Run 5. ORNL-DWG 702829 100 e e —— [ 80 - ® a' . % - Z a 60 n a . - o . T i = a0} - g J o i : - I——————/ ) . T . g e C : - 0 i —t 4 L i 1 i L 1 e 1 i J 2 1 . 20 40 60 80 100 120 140 160 180 200 TIME {min) . Fig. 23.3. Bismuth Holdup During Run 5 Based on Pressure in Salt Jackleg. 4o 297 Table 23,1, Apparent Bismuth Holdup at Several Bismuth and, finally, a 41-min steady period. Since the pressure and Salt Flow Rates During Run § at the bottom of the column is predominantly due to the static pressure of both phases, the apparent bismuth holdup can be calculated for each of these steady Steady Bismuth Flow Salt Flow Bismuth Holdup Perlod (tnl/min) (mlfmin) (rol %) “periods. The holdup values are shown in Table 23.1. 1 (12 min) 75 75 16 During the last steady period, the salt flow rate was 2 (12 min) 81 I 2 doubled with no apparent accompanying change in 3 (41 min) 87 75-150 28 column pressure, When the salt flow reached 150 ml/min (i.e,, a bismuth flow rate of 87 ml/min), PHOTO 98387 Fig. 23.4. Section of 0.82-in.-ID Column Showing Dendritic Iron Deposit (Upper Photograph). The lower half of the column was potted in epoxy resin, and the steel pipe was stripped away. flooding of the column caused oscillahon in the salt flow rate. A short time later 2 metal chip caught in the valve that controls the level of the salt jackleg and caused even greater oscillation. Run 6 was successful in that countercurrent flow was obtained in the column. The bismuth flow in the period of most nearly steady flow gradually increased from 155 to 175 ml/min and then fell steadily to about 120 ml/min over a 30-min period. During the same period the salt flow slowly increased from 65 to 80 ml/min, Also during this time the apparent bismuth holdup, as 298 Orifice Calibration. — Bismuth and salt calibration runs were made with an orifice meter of the same design as that calibrated earlier? with mercury and water. The discharge coefficients for salt were 0.28, - 0.34, 0.36, and 0,37, respectively, for flow rates of 130, 195, 245, and 295 m!/min, Further experiments with bismuth provided an average discharge coefficient of 0.73 (standard deviation = 0.07). _ The off-gas connection from the orifice chamber now includes a mercury seal. In future experiments the . additional off-gas pressure (above and below the orifice) indicated by column pressure, increased from 70 to 82% and then decreased to 70%. An iron deposit in the bismuth exit line during run 7 : caused bismuth to overflow into the salt exit line. After the affected line had been replaced, the flow of bismuth and salt was still intermittent (run 8). Although the region in which operation was attempted was near flooding, much of the problem may have stemmed from iron deposits in the column. We decided to replace the column as a result of the observed decrease in flooding rate, Radiographs of the column made before and after run 8 showed a tight packing arrangement of the - Y%-in.-diam molybdenum cylinders but did not clearly show the presence of the dendritic iron deposits, which - . we found in the lower half of the column when it was sectioned for direct examination (Fig. 23.4). Air oxida- tion of the carbon steel column was severe, amounting to a loss of about 0.050 in. of wall thickness. The high-temperature aluminum paint applied initially appeared to have given little protection against oxida- tion; components and transfer lines that were replaced after the column was removed were painted with Markal CR, which the Metals and Ceramics Division had shown to be generally superior in cyclic tests. A new column, packed with ¥, -in.-diam molybdenum ~ Raschig rings, was installed in the system, The 84% void fraction of the new column is much higher than that of the original column (~40%) and should permit higher flow rates and minimize the effect of a moderate amount of iron precipitation. The bismuth-salt dis- engaging section at each end of the column was modified to improve scparation of the phases; minor improvements were made in the design of the entrain- ment detector located in the salt overflow loop. X rays were made of the packed section, and pressure drop will aid in forcing the fluids through the drain line and will prevent the orifice drain chamber from filling with liquid, This 1s expected to reduce the scatter in the data. g 23,2 DESIGN OF A PROCESSING MATERIALS - TEST STAND AND THE FIRST MOLYBDENUM REDUCTIVE EXTRACTION EQUIPMENT E.L.Nicholson W. F. Schaffer, Jr. L. E. McNeese E. L. Youngblood Difficulties have occurred frequently in the engineer- ~ ing development program due to the use of carbon steel as the material of construction for the experimental equipment.! The difficulties arise from three sources: (1) the low strength of carbon steel at elevated - temperatures (600°C) and the additional loss of strength that may occur as a result of graphitization upon prolonged exposure at elevated temperatures, (2) the poor resistance to air oxidation on external surfaces at elevated temperatures, and (3) plugging of the experimental equipment by mass transfer deposits of iron. These problems were recognized when the equip- ment was being designed,? and they are being partially circumvented so that experimental work can continue. It has always been evident that carbon steel is not a " satisfactory material for extended use. The refractory metals molybdenum, tungsten, and tantalum show satisfactory resistance to salt and bismuth, but the cost and fabrication problems are too formidable to permit their use for the first experimental systems. Molyb- ‘denum appears to be the best refractory metal for measurements across the column were made with - flowing argon, These measurements will serve as a base line for future diagnostic tests of the column. 2MSR. Program Semiann. Progr. Rept Aug. 31, 1969, ORNL-4449, p. 223. reprocessing applications, and tubing, plate, and billet stock are commercially available in a limited range of ApMsR Program Semiann, Progr. Rept. Aug. 31, 1969, " ORNL-4449, pp. 229-30. 2y, Suskind et al., Corrosion Studies for a Fused Salt—Liquid Metal Extraction Process for the Ligquid Metal Fuel Reactor, BNL-585 (T-146) (June 30, 1960). - sizes. Recent® and continuing development work (see Part 5, Materials Development) by the Metals and Ceramics Division indicates that it is feasible to fab- ricate small-scale equipment from molybdenum, We have prepared a preliminary design and cost estimate and have received approval for a reprocessing materials test stand and the first molybdenum molten- salt reprocessing equipment system to be operated in the test stand. The first test will demonstrate the ~ fabrication of a complex refractory metal system, test the suitability of molybdenum as a material of con- struction for reprocessing equipment, and provide engineering data on packed-column performance with the bismuth-salt system, The test stand will be placed in 2 beryllium contain- ment area in Building 4505, Its associated equipment will include instrumentation, high-purity gas supply ‘systems, a fill-and-dump vessel for the salt and bismuth used in the test equipment, and the test stand con- tainment vessel. The containment vessel will be about 20 in, in diameter by 14 ft high and will be heated and filled with an inert gas to protect the molybdenum equipment during high-temperature operation. , The first equipment to be installed in the test stand ‘will be a simple reductive extraction system consisting of a l-in-diam by S-ft-high packed column with 3% -in.-diam upper and lower disengaging sections. Salt and bismuth will be circulated countercurrently through the column by gas-lift pumps, which will pump the streams to elevated 3% -in,-diam head pots for gas separation, flow measurement, and gravity flow back through the column, Special instrumentation will include an interface detector for the lower disengage- ment section of the column and a device for measuring ‘pressure drop across the packed column, Samplers are’ provided for each stream, and reducing agents (e.g., thorium ‘metal) or oxidants (e.g., hydrogen fluoride) can be added to the system to study their effects on column performance, The molybdenum equipment will be suspended by hangers from the top flange of the test-stand containment vessel so that the loop can be lifted out for repair or for modifications that would allow other materials, instrumentation, or flow systems ' to be demonstrated. ‘We have begun to prepare a full-sxze layout of the test stand and the molybdenum equipment to pinpoint the problems of clearances and orientation of vessel nozzles and lines. This will enable Metals and Ceramics Division personnel to recommend forming and joining tech- 3SMSR Program Semumn. Progr. Rept. Aug 31 1969, ORNL-4449, pp. 210-13. 299 - niques for each vessel and line. We will then prepare the final design of the system, which will be fabricated and assembled in the test stand under the supervision of the Metals and Ceramics Division, We have nearly completed the detailed design draw- ings for the head pots. After they have been reviewed, we will verify performance by constructing and testing a plastic model. The design incorporates a mechanical assembly method for the internal baffles, process and _ instrumentation lines, and an orifice assembly for flow measurement that minimizes the need for welding. Only one circumferential weld will be required to join the two halves of the vessel that are produced by a back-extrusion forging operation. The spacing and the size of the nozzle bosses, which constitute an integral part of the head forgings, are such that the bosses can be machined to permit attachment of the process and instrument lines by welding, brazing, or by mechanical fittings. Final choice of the attachment method will be specified by Metals and Ceramics Division personnel. ' Mechanical fittings may simplify field assembly procedures and facilitate the repair of damaged compo- nents or the addition of new components to the loop in the future. The Gamah Division of Stanley Aviation Corporation is investigating the problem of couplings for molybdenum equipment at no cost to us. They produce a patented high-pressure metal seal ring coupling for aerospace use in which the sealing 'is accomplished by deformation of a flat annular-disk seal ring. The seal ring can be reused, and only a very low torque is required to assemble the connector. They have developed a swaging process for attaching the connector to pipe or tubing. This process eliminates the need for welding molybdenum and would produce a strong leak-tight joint as opposed to a welded joint that is ‘highty fragile and brittle or a brazed joint with reduced corrosion resistance. We are examining a sample swaged joint (prepared by the Gamah Division) consisting of %-in. molybdenum tubing swaged into a simulated },-in,-diam connector hub, No flaws have been detect- able by radiography or dye-checking procedures, and the joint is leak-tight to the limit of sensitivity of our helium leak-detector equipment (5.9 X 107 cc/sec per scale division). Metallurgical examination of the joint is under way. Gamah Division personnel have also pro- vided us with a % -in. tubing test assembly consisting of an all-molybdenum connector with 4-in. lengths of molybdenum tubing swaged into the connector. They report no leakage of the molybdenum joint with 400-psi helium at room temperature or after exposure at 1300°F for 7 hr. The joint is now being evaluated by Metals and Ceramics Division personnel. - Various small items have been fabricated from molyb- denum for use in MSBR reprocessing development work. The experience has been valuable, although sometimes frustrating. We wanted to build an electro- magnetic helical induction pump of molybdenum for circulating bismuth, Such a pump is not available commercially, and the refractory metal flow element would have to be fabricated by ORNL in any case.Ina - preliminary test of the . fabrication method for the helical flow element, /s-m -OD molybdenum tubing was wound at 450 to 500°F on a heated mandrel to give a 2%-in.-OD, eight-turn coil of satisfactory quality. Because an actual pump cell would require a coil having many more turns than this, two commercially available lengths of tubing were welded together. All attempts to hot-bend the coil failed due to a break in the tubing in the brittle heat-affected zone ‘of the weld. The con- tinuous electrolytic cell testing system is arranged so that the exit streams from the electrolytic cell flow into - 2-in.-diam sampling cups before going to the mixer- settler *ank. Both the sampling cups and the exit tubing are made of molybdenum. The first two sampling cups cracked during the welding assembly operation. Some weld joint designs were modified, and the sequence of 300 assembly was altered, Three satisfactory units and one additional unit with a minor defect resulted from a total of six attempts. We also learned that repairs to a defective weld have a high probability of causing complete failure of the structure by cracking.. ' 23.3 BISMUTH-SALT INTERFACE DETECTOR ~ J.Roth L.E.McNeese A bismuth-salt interface detector is required for control of the interface location in salt-metal extraction columns. The possibility of using an eddy-current detector is being explored, and equipment has been fabricated for testing this type of detector. Detectors having two configurations are to be developed; these consist of (1) a2 configuration in which the working fluid (salt or bismuth) is located inside the detector coil and (2) a configuration in which the working fluid is located outside the detector coil. In each case the detector consists of a primary winding (through which an alternating current is passed) and a secondary winding (in which a current is induced). The induced current is dependent on the conductivities of the materials adjacent to the primary and secondary coils; since the conductivities of salt and bismuth are quite different, the induced current reflects the presen(:e or absence of bismuth, : ‘The principal problem associated with this type of detector results from the low permeabilities of materials that can be used with bismuth (e.g., molybdenum or - low-carbon steel), since these materials will separate the coils from the bismuth and salt. We are attempting to use a duplex detector tube consisting of a thin-walled molybdenum tube (wall thickness of 20 mils or less) and a type 316 stainless steel tube. Since type 316 stainless steel has a relatively high permeability, the effective permeability of the resulting tube should be satisfactory. The electronics system for testing the two detector configurations has been assembled, and the detector coil for the first configuration has been fabricated. The detector coil for the first confi guratlon shown in Fig. 23.5, consists of a bifilar winding of 30-gage platinum wire wound into grooves that have been machined into the surface of a !% -in.-OD, % ¢-in.-ID tubular Lavite form, The machined grooves are 0.015 in. wide and 0.015 in. deep, with a round bottom, and are separated by 0.035 in. of Lavite, The winding, which is 12 in. long, contains ten turns of each coil per inch, An end collar 1% ¢ in. in diameter and 1 in. long is located at the end of each coil. The connecting leads consist of a twisted pair of wires in order to ensure minimal connecting lead inductance. The entire assem- bly is coated with a ceramic glaze to reduce the probability of external shorting of the coils. _ The power input to the primary coil is supplied bya Wavetek function generator, model 110, and is approxi- ‘mately 35 kHz at an output of about 15 v. The output . of the secondary coil will be amplified to about 5 v, rectified, and filtered, The differential voltage (about 1.5 mv dc) will be recorded. Preliminary tests of the detector system ‘at room temperature and at 600°C (nominal operating tempera- ture) have been made, and required system modifica- tions have been completed. The equipment has been installed, with provision for positioning an argon- bismuth interface at known points within the duplex tube (molybdenum inside 316 stainless steel) around which the detector is located, Equipment for containing and purifying the molten bismuth, controlling system temperatures and pressure, and supplying argon and hydrogen to the system 1s also provided. 301 ' ' T ¥ PHOTO 96825 i = ] i - - i i 302 23.4 CONTACTOR DEVELOPMENT: PRESSURE DROP, HOLDUP, AND FLOODING IN PACKED COLUMNS J.S.Watson L.E.McNeese A study of pressure drop, holdup, and flooding rates in a 2-in.-ID packed column was made by a group of -MIT Practice School students.! Mercury and water were used to simulate bismuth and molten salt. We have previously reported? similar measurements with several packing materials in a 1-in.-ID column which showed that the metal dispersion, and thus the interfacial area, is greatly enhanced by using larger packing sizes (%} in. or. possibly greater). In order to study larger packing materials, a larger column and a larger metal pump were ~ installed. The MIT students obtained the first data with this new equipment. Two different packing matenals were studied: Y%-in. solid cylindrical packing (whlch had previously been tested in the 1-in. column) and % -in. Raschig rings. The Y4-in. packing was tested in order that data from the 1A, 1. Fredriksen, J, J, Protulipac, and S. C. Trindade, Hydrodynamics of a Mercury-Water Packed Column, MIT- CEPS-X-88 (Oct. 17, 1969). %), S. Watson and L. E. McNme, Unit Operations Sect. Quart. Progr. Rept. July—September 1968, ORNL-4366. " ORNL DWG 70-2830 10.0 — 1T T T ¢ > 0o © o o T o o I o o [ = o | PRESSURE DROP, c¢m of Hg MINUS HoO o . o T Vd {(ft/hr) 0 36 — 68 | 102 - 139 | 172 Oa4n00b+ | = = - - - - 1 L] 1 1 L 1 L bl l ol 20 100 600 - SUPERFICIAL WATER VELOCITY, V¢ (ft/hr)r Fig. 23.6. Pressure Drop vs Superficial Water Velocity in-a 2-in. Column Packed with _3/8-in. Raschig Rings. - - 303 l4in, and 2-in. columns could be compared. The agreement between the two sets of data is excellent. Most of the minor differences can be attributed to the small difference in the void fractions of the two columns. Because of wall effects, Ys-in. material packs more densely (i.e., has a lower void fraction) in 2-in. than in 1-in. columns. The results for %;-in. Raschig rings are shown in Figs. 23.6—23.8. The pressure drop and holdup curves are similar to those observed earlier with %-in. packing. The flooding curve shown in Fig. 23.8 is surprisingly close to that observed previously with %-in. Raschig rings. It appears that the flooding data for ¥ -in. solid packing and for '%- and %-in. Raschig rings can be represented by a single curve if the superficial velocities are divided by the void fraction, as shown in Fig. 23.8 and as indicated in Eq. (1). The resulting curve is linear and has a slope of approximately —1 when the square roots of the phase velocities are plotted against each other. These conditions correspond to a constant slip velocity at flooding. Thus the flooding conditions for - these three packing materials, using mercury and water, - can be approximated by a single equation, (1) Vil2+ V32 =31€112 where V, and V; are the continuous- and dispersed- phase superficial velocities, respectively, expressed in feet per hour, and € is the void fraction. This expression indicates that the flooding rate does not depend upon the packing size; however, it is believed that the ORNL-DWG 70-2831 Vg (ft/he) - 36 A | % A 68 O Q n 102 o._ o i 139 Vv = 72 o] ook o - - Q o o o1 ) A - Or = v - ‘ S a4 L:l 17 L1 1 1 x99 by sy oO o S - 100 , 200 ‘ - - ‘SUPERFICIAL WATER VELOCITY, \_/c (ft/hr) Fig. 23.7. Holdup vs Superficial Water Velocity in a 2-in. Column Packed with %-in, Raschig Rings. 304 ORNL-DWG 70-2832 L N T T T T T T T T T T T T T T T | S : ' g 20} — -l : 7 T | ] S, S J O | ] ¥ A 174" SOLID CYLINDERS, - = } MCNEESE AND WATSON ) - O 174" SOLID CYLINDERS : i 0 3/8" RASCHIG RINGS i Or e oo e e b s b bty e e bt e a1 4 ] o 10 20 30 33 % 4 (veselt, (tt7bn)® Fig. 23.8. Flooding Lines for %-in, Solid Cylinders and %-in. Raschig Rings. flooding rate will increase with packing sizes larger than % in. ' : - The model that had been proposed to relate pressure drop, holdup, and flooding rates from the I-in.-column data predicted flooding rates reasonably well for the We are studying column modifications which will decrease the effect of axial mixing and thus improve column performance. The proposed modifications Y, -in. packing, There appears to be some difference . between the slopes of the estimated and measured flooding curves, but the estimated curve is still reason- ably close to the data. With the 34-in. packing, however, involve placing devices at several points along the column to reduce or prevent axial mixing across the column at these points. If the devices are separated by a distance equivalent to one theoretical extraction stage when no backmixing occurs between the stages, the efficiency of the column (height of the column with the predicted flooding curve was significantly higher than the experimental curve. 23.5 CONTACTOR DEVELOPMENT: AXIAL MIXING IN PACKED COLUMNS J.S.Watson L.E.McNeese We previously reported calculations' of predicted reductive extraction column performance which indi- cated that prohibitively long columns would be re- quired because of the axial backmixing in systems requiring high volumetric flow ratios (metal to salt). A rare-earth removal system that exploits the small distribution coefficient of the rare earths between the fuel carrier salt and a reducing bismuth phase is a system of this type. ' - 13, S. Watson and H. D. Cochran, Jr., MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, pp. 238-40. perfect devices divided by the height of the column with imperfect devices) in the rare-earth system when the flow ratio is 100:1 would be greater than 75% if less ‘than 15% of the salt flowing through a particular segment is recycled or backmixed to the segment below. These conditions do not appear difficult to ‘achieve and will result in satisfactory column heights (probably less than 3 ft/stage). The type of device under investigation is illustrated in Fig. 23.9. The metal flows down the annular section of the upper piece into what is effectively an inverted bubble cap. As the metal flows over the weir in the lower piece, it forms a seal that forces the salt to pass upward through a sieve plate in the center of the column. The sieve plate is a restriction which can be ~ designed in such a manner that the salt velocity through the openings will be reasonably high when a low salt flow rate is employed. Thus one would expect relatively little back-diffusion of salt through the openings in the sieve plate, | ORNL-~-DWG 70-4506A METAL FLOW | METAL FLOW o . o SALT FLOW Pof®. | 4 | o o . °¢ e AN A 0% © b © o °° 0 X °:‘,o"o :o°D° o o °° O o .° °° 00 o 0% 0o 0.9 g ° o0, 0.2 %N % o etao 0 ° oo o° 05 oo o 2 00,76 , Fig'. 23.9. Schematic Diagram of an Axial Mixing Preventer. We have tested two designs of “backflow preventers” to date, using mercury and water to simulate bismuth and molten salt. The column was packed with % -in.- diam Raschig rings, and at least one backflow preventer was inserted near the middle of the column. The countercurrent flow of mercury and water was estab- lished, and a tracer solution (cupric nitrate) was injected near the top of the column at a constant rate. The solution in the column was analyzed for the tracer at several points down the column on both sides of the backflow preventers. If the logarithm of the tracer concentration is plotted as a function of distance down the column, a discontinuity is observed at the preventer location, as illustrated in Fig. 23.10. The ratio of the 305 concentrations at the discontinuity is then a measure of the fraction of solution back-diffusing through the preventer. The fraction or percentage of backflow through a preventer appears to be principally a function of the water velocity through the sieve opening and of the diameter of the sieve openings. No dependence of backmixing on mercury rate was evident; however, the relatively few changes in mercury rate and the usual data scatter make this conclusion tentative. It is expected that a significant dependence on metal flow ~rate would be observed if the metal flow rate were increased to the point that complete coalescence was not obtained in the metal downcomer. ) Data obtained with the backflow préeventers aré ORNL DWG 70-4507 1 1 ! N % 10 0.9 0.8 0.7 06 I'll]ll lll"]lllo 0.5 04 03 C/CRrEFERENCE 0.2 0.4 | i I | ] -5 0 5 10 15 20 25 DISTANCE FROM TRACER INLET (IN.}) Fig. 23.10. Concentration Profile Using a Backflow Preventer with a Single J%-in. Sieve Opening. Mercury flow rate, 400 ml/min; water flow rate, 115 mi/min. diam holes. One can see that the extent of backmixing decreases as the water flow rate increases and that backmixing of less than 15% can be expected (water flow rate greater than 90 ml/min or 22 ml/min per hole). The openings were then drilled out to a diameter of Y, in., and the top curve shown in Fig. 23.11 was obtained. Again one can see the decrease in backmixing “as the water flow rate increases. Backmixing did not reach an acceptably low level in any of these experi- ments; however, an extrapolation of the curve predicts “that backmixing would become less than 15% at water presented in Fig. 23.11. Although there is considerable - scatter in the data through which the lines were drawn, the lines are believed to be representative. The first back flow preventer contained a % -in.-deep bubble cap and a Y%-in.-thick sieve plate containing four % ,-in.- ml/min per hole). A second preventer, rates between 150 and 200 m!/min (approximately 50 \‘guich had a 2-in.-deep bubble cap and only a single ¥ ¢-in.-diam opening in the orifice, was tested; however, no back- mixing could be detected at the lowest measurable water rate (23.5 ml/min). The diameter of the opening was mcreased to % in., but agam no backmixing was 306 detected. When the diameter of the opening was increased to % in., the results shown in Fig. 23.11 were obtained. For this preventer, backmixing was less than 15% when the water flow rate was greater than 50 - ml/min. (This value agrees with the results estimated by extrapolating results from the preventer with four % -in, holes.) The diameter of the sieve opening was then further increased to % in., and the results shown in Fig. 23.11 were obtained. Backmixing was increased con- siderably by this change in diameter; however, at '-e ¥ ] sufficiently high water flow rates, backmixing can be reduced to any desired percentage. The only undesirable feature of these backmixing - preventers (other than their complexity relative to simple packed columns) is their reduction of the column capacity or flooding rate. The devices just discussed were operated at metal flow rates as high as 94 ft/hr (column superficial velocity), which is approxi- mately 15% of the flooding rate for a column packed with %-in. Raschig rings. At the lower metal rates an ORNL-DWG 70—4508 o8- FRACTION BACK MIXING .9 0 T o H T - 0.2 ' 1 v 1 i l E 80 100 120 140 WATER FLOW RATE, ml/min Fig. 23.11. Summary of Backflow Preventer Results. gl where accumulation of uncoalesced and coalesced mercury was noted above the backflow preventer, but as the metal throughput was increased, it appeared that 307 back-transfer of water through the preventer occurred. It is expected that the design can be improved in order to allow higher metal flow rates. ' These experiments have demonstrated that devices - can be designed to reduce backmixing to an acceptable level. Although the acceptable level has been defined to be about 15%, one could achieve even lower values if desired. The principal design parameters are the diam- eter of the sieve opening and the salt flow rate per hole. Although openings % in, in diameter or greatér could be used, the required water or salt velocity increases very sharply as the hole size increases. At present, Y,-in.-diam openings appear to be a good choice and will be effective with a flow rate as low as 50 ml/min . per hole, The devices are relatively simple; however, the column capacity is decreased. A greater column capac- ity may be achieved with larger-diameter columns in cases where larger bubble caps can be employed. Estimation of Axial Backmixing in Both Phases in Countercurrent Contactors. — We previously presented! an approximate equation that predicts the effects of axial backmixing in one phase on the performance of a countercurrent contactor. This equation was useful in estimating the column heights required in the protac- tinium and rare-earth removal system. A similar empir- - ical equation has been developed to predict column performance when backmlxmg occurs in both phases. The proposed equation is: : 1 , F | " Npy - 1+F+1/NTU’ 1 =the length of the column. without back- mixing divided by the length of the column with backrmxmg, | m = distribution coefficient, NTU = number of transfer units required to achieve a given separation if the phases are in plug flow. This equation predicts efficiencies to within approxi-‘ mately 0.07 (i.e., the difference between the estimated and the calculated efficiency will be less than 0.07) when NTU=2, n=02, 1 Npex ?F_W 9 and Npey >15. 23.6 AXIAL MIXING IN OPEN COLUMNS L.E.McNeese J.S.Watson A study of axial mixing in a 2-in.-diam open column during the counter-current flow of air and glycerol or butanol solutions was completed by A. M. Sheikh and J. D. Dearth, of the MIT Practice School.! Axial mixing is important in the design of tubular reactors such as continuous fluorinators. Bautista and McNeese? pre- viously reported results obtained in studies of axial mixing in the same 2-in.-diam equipment using air and water. The experimental technique involved photo- metric measurement of the steady-state concentration profile, in the column, of a tracer (cupric nitrate) that was fed continuously to the bottom of the column. The present study was undertaken to obtain data on the effects of two physical properties, continuous-phase viscosity and surface tension, on the axial dispersion ~ coefficient. Molten salts of interest have a viscosity of about 30 centipoises; water (used in previous studies) has a - viscosity of approximately 1 centipoise; and the vis- * Npy = Peclet number in phase x = Jc(HT )/Ex- _ U,(HTU,)/E,, 'HTU, hexght of a transfer unit with phase flow, U = superfiaal velocity of the reference phase, E = axial dlffuslon coefficient, ' F = extraction factor = mUx U, Npey =Peclet number in phase y = cosity of the glycerol solution used in this study is 15 centipoises. The surface tension of molten salts of interest is 100 to 200 dynes/cm, while that of the 1A, M, . Sheikh and J. D, Dearth, Axial szing in an Open Bubble Column CEPS-X-91 (Dec 17, 1969), M, S. Bautista and L. E. McNeese, MSR Program Semiann. Progr. Rept. Aug. 31, 1969, ORNL-4449, p. 240. ORNL DWG 70-4505 I T T T T 1 I T T 1 T T 1 200}~ - O - BUTANOL-WATER SOLUTION, ¢ =37.8 DYNES/CM 8 A -BUTANOL-WATER SOLUTION, o"=53.2 DYNES/CM < A - BUTANOL-WATER SOLUTION, o= 67.4 DYNES/CM (SINGLE RUN) - 8 100I~ - @ -GLYCEROL-WATER SOLUTION, jt=15 CENTIPOISE - . E 80 \ . / - L 50} — 8 4ol ~ & » 30 = o w % a 20 N 10 I P 1111 ! 11 ¢+ 3 1 | ! 4 56 8 10 20 30 40 60 80100 200 1 2 3 AIR FLOW RATE (CM¥SEC) Fig. 23.12. Effect of Gas Flow Rate, Liquid-Phase Viscosity, and Surface Tension on Axial Dispersion Coefficient. butanol solutions varied between 37.8 dyneslcm and the value for water (72 dynes/cm). The results of the study are summarized in Fig. 23.12, which shows the dispersion coefficient as a function of the air flow rate. The dashed line represents the earlier results of Bautista and McNeese, who found that the data fall into two regions. The first region, which covers low air flow rates, corresponds to conditions when individual bubbles formed at the orifice travel up the column without coalescence. The second region, which occurs at air flow rates above 30 cm?®/sec, corresponds to “slugging” flow. Increasing the liquid viscosity to 15 centipoises (by adding glycerol) decreased the dispersion coefficient in the “bubble” (lower) region; however, the transition point between the regions was not changed. The _coefficient increased more rapidly with air flow rate in the slugging region, at least up to the curve by Bautista ~and McNeese. Decreasing the surface tension increased the disper- sion coefficient in the bubble region, although there is little difference between the data obtained from a solution having a surface tension of 37.8 dynes/cm and those obtained from a solution having a surface tension of 53.2 dynes/cm. The transition point occurred at a higher air flow rate, and the bubble region .curve intersected the slugging region curve of Bautista and McNeese. The dependence of the dispersion coefficient on air flow rate appears to be somewhat greater than that observed with air and water, but insufficient data are available to reach a definite conclusion on this point. -This study was especially useful since it represented the first systematic attempt to vary physical properties in a bubble column. The data obtained indicate that higher viscosity and higher surface tension lead to lower dispersion coefficients, which is in the favorable direc- tion for molten-salt systems. However, the variation of physical properties did not cover a wide range of conditions; also, when liquid mixtures are used, there is always a possibility that one component will concen- . “trate at the interface and result in properties in this region which do not correspond to conditions in the bulk liquid. Additional studies in the future may include evaluations of the effects of changing the column and orifice diameters as well as of changing the physical properties of the continuous phase, " 237 DEMONSTRATION OF THE METAL " TRANSFER PROCESS FOR REMOVING RARE EARTHS E.L. Youngblood W.F. Schaffer, Jr. L. E. McNeese E. L. Nicholson J. R. Hightower, Jr. Equipmeht has been fabricated for the demonstration ~and study of the metal transfer process for removing rare earths from single-fluid MSBR fuel salt. The first series of experiments will be made in a carbon steel vessel 24 in. high by 6 in. in diameter, shown in Fig. 23.13. An internal partition divides the vessel into two equal-volume compartments, but terminates % in. above the bottom to allow communication at this level. 309 A bismuth phase saturated with thorium will fill the first 2 in. of the vessel, and above it, in their respective compartments, are a fluoride salt phase (72-16-12 mole % LiF-BeF;-ThF,) and an LiCl phase, each having a depth of about 3 to 4 in. The fluoride salt will initially contain about 0.3 mole % LaF;. The compartment which contains the LiCl accomimodates a double-walled cup, approximately 1.5 in. in diameter and 9 in. high, made of carbon steel. The inner wall of the cup is electrically insulated from the remaindér of the vessel by quartz spacers. During the experiment the cup will contain about 200 cc of 0.4 mole fraction lithijum metal in bismuth. Molten LiCl will be circulated through the cup containing the lithium-bismuth alloy by a pump ORNL-DWG 70-4504 B 1 1 = — T | C ' | _—QUARTZ CARBON-STEEL— 1 PUMP ~ PARTITION I | | o . | 6-in CARBON- - ‘ e STEEL PIPE 24 in | o —LiCI 72-16-12 MOLE %o~ = 'FUEL CARRIER SALT - ) , o 2 _—Li-Bi Th-8i~K e . X Fig. 23.13. Carbon Steel Vessel for Use in the Metal Transfer Experiment. operated with argon pressure. Provision has been made for sampling all phases and for mixing the metal phases. To begin the experiment, the three phases in contact (fuel salt, bismuth, and LiCl) will be allowed to come to equilibrium, and samples will be taken. The LiCl will then be pumped through the reservoir containing the bismuth-lithium solution at a rate of about 25 ¢m?®/min in order to remove LaCl; from the LiCl. The stripped LiCl will overflow the reservoir cup and return to the initial LiCl volume. After a period of about 3 hr, the circulation of LiCl will be stopped, and the system will be allowed to come to equilibrium, Then samples will again be taken. It is estimated that about 20% of the lanthanum initially present will have been transferred to the bismuth-lithium strip solution at this point. The experiment will be continued until the desired fraction of the lanthanum (50 to 90%) has been transferred to the strip solution. A quartz pump with sapphire check valves is being tested for circulating molten LiCl at 650°C. The pump was operated successfully for a few hours, but we have had difficulties with short circuits in the electrical probes that control the pump and with gradual de- terioration of the quartz in contact with LiCl. The deterioration of the quartz is thought to be due to reaction products that are formed by moisture in the LiCl; this moisture cannot be removed by heating. Another test of the pump will be made using LiCl that has been previously contacted with a thorium-bismuth alloy to remove impurities. The pump and its control system are being modified to eliminate the short circuits. 23.8 FROZEN-WALL FLUORINATOR DEVELOPMENT J.R. Hightower,Jr. L. E.McNeese The fluonnat:on of molten salt to remove uranium is requlred at several points in processes being considered ~ for the isolation of protactinium and for the removal of ~ rare earths. The fluorinators will be protected from corrosion by freezing a layer of salt on the metal surfaces that potentially contact both F, and molten salt. AIthough the separate aspects of such an operation _(namely, continuous or batch fluorination and frozen- ~ film formation) have been shown experimentally to be feasible, the testing of a fluorinator protected against corrosion by frozen salt has been hampered by the lack of a corrosion-resistant source for generating heat in the molten salt. The heat will be provided in a reactor fuel processing plant by the decay of fission products in the salt. CONFIGURATION I ORNL DWG T0-2827 CONFIGURATION I COOLING TUBES 1 % z £ Q Q 3 a3 3 the values during the major portion of the run were 24 to 56 times the value measured in the equilibrium still. The effective relative volatility for 14 7’Pm was based on a computed feed concentration and took account of - the MSRE power history. The computed feed concen- tration of '*7Pm was felt to be a more accurate value than the measured feed concentration, which was only about 30% of the expected value, since measured concentrations of other lanthanide fission products agreed well with computed values. As in the case of 144Ce, the relative volatility of **7Pm was low (<7.8 X 107%) at the time the first sample was taken; it then rose sharply to about 3.4 X 1073 for the remainder of “the run. Also, as in the case of !**Ce, the relative ~volatility was low at the time the fifth sample was taken, The relative volatility of promethium had not been measured previously. The variation of the relative volatility of '35Eu during the run. closely paralleled variations of the relative volatilities of !*4Ce and '*7Pm. The value for the first sample was low (less than 1.5 X 1075); however, the values for all the other samples, except the fifth sample, were higher (about 2.2 X 107*). The value for the fifth sample was lower than 2.2 X 107 by a factor of 2.6. The effective relative volatilities during the MSRE distillation experiment (based on a com- 318 puted feed concentration of !$5Eu) were lower than . the value of 1.1 X 107® which was measured in recirculating equilibrium stills, However, the analysis * for '35Eu in the condensate samples is suspect since there was some difficulty in this analysis; thus all of the 155By data obtained for the condensate samples were reported as approximate values. On the other hand, it is run closely paralleled the relative volatilities of ***Ce d 147Pm Figure 24.3 shows the effective relative volatilities of °1Y and °°Sr. During the run the effective relative volatility of 'Y had an average value of 1.4 X 1072} about 410 times the value measured in recirculating equilibrium stills. The variation of ay, v.Lip V3 similar to variations of relative volatilities of the lanthanides; - the low value for the fifth sample was most noticeable. The effective relative volatility of °°Sr (based on the - measured concentration in the feed) had an average value during the run of 4.1 X 1073, about 84 times the value measured in recirculating equilibrium stills. Al though it is not shown in Fig. 24.3, the average value of the relative volatility of 8?Sr (based on a computed " concentration in the feed) was 0.193, about 3900 times the value measured in equilibrium stills. In the con- ‘densate samples the ratio of the ®°Sr activity to the significant that the variation of a; 5 S b u.LiF during the 99gr activity, which should have been about 0.002 in each case, varied from 0.22 to more than 10 (Table 24.1). ORNL DWG 70-4516 — T T T T T T T T T T T T 139 - '_.a_-‘ - 4 = . w - 16 =3 - 45 wN | Ne ~ 14 °o o c = -'33% o — = O o E'fi':" ' 12 8& OZrF4B’5ZI' I.I.Ii o | 2 < 10~ A A |, & - N 7 « u.; 8" a mu_ N %:’z ® N e ° o 7 > ® S ® - R . . . Ber_ 55 ¢ e | t = w < g 3 - - ¢ ® 24 . EE @ BeF, l-l.lj : .AszQ 0@ o'Ezr o 4 [ 1 | L | ! ] 1 L. L I { $ | 0 1 _ 2 3 4 5 6 T 8 LITERS CONDENSATE COLLECTED Fig. 24 1. Effective Relative Volatilities of ZrF,, 9571, and BeF; for the MSRE Distillation Expenment ™ n w ORNL DWG 70-4515 10! T T T T T T T T T T T T T ] T 7 r— - @ ¢ e “Qc 102 ° o b ° ® o _u ° ' 2‘: - — . A - = A A Ty . A A o @ Q A A “Tpm- g A g ] - w a 103 A _ @ Mlc. - A 147Pm - L O |lb€u - o - o o o o . O B o o 0, - o . w Z 107 e — « i - -d - - lfil:-l L .OIY - 5 A gy . i - - - 10—4 L | 1 1 1 ] 1 1 § 1 ] 1 { 1 . 0 1 2 3 4 5 6 7 8 LITERS CONDENSATE COLLECTED | Fig. 24.3. Effective Relative Volatilities of 'Y and ®°Sr in the MSRE Distillation Experiment. ‘causes for the discrepancies include: (1) entrainment of droplets of still pot liquid in the vapor, (2) concentra- tion gradients in the still pot, and (3) contamination of samples during their preparation for radiochemical analysis. Entrainment was suspected for a number of reasons. Entrainment of only 0.023 mole of liquid per mole of vapor would account for the high relative volatilities calculated for the slightly volatile fission products 144Ce, 147Pm, °'Y, and °°Sr. Entrainment rates of this order would not be reflected in- the effective relative volatilities of more-volatile materials (a > 1). The high correlation of the scatter of the calculated effective relative volatilities of different slightly volatile fission products is consistent with the hypothesis that entrainment occurred, Since entrainment was not apparent in the nonradio- active operation of the still,® reasons for entrainment in the radioactive operation were sought to support the hypothesis. Evidence of a salt mist above the salt in the pump bowl at the MSRE and above salt samples removed from the MSRE has been reported,*** and studies have indicated that these mists are present over 31, R. Hightower, Jr.,, and L. E. McNeese, Low Pressure Distillation of Molten Fluoride Mixtures: Nonradioactive Tests for the MSRE Distillation Experiment, ORNL-4434 (in prepara- tion). 48, S. Kirslis and F., F, Blankenship, MSR Progmm Semzann. Progr. Rept. Feb. 29, 1968, ORNL-4254, p. 100. - 38, 8. Kirslis and F, F. Blankenship, MSR Program Semmnn. Progr Rept. Feb. 28, 1969, ORNL-4396, p. 145. s T fwn (w 321 radioactive salt mixtures but not over nonradioactive mixtures. However, examination of data from these studies showed that entrainment rates large enough to explain the results of the MSRE distillation experiment could be ‘obtained only by assuming that the salt concentration in the gas space above the salt during this experiment was equal to that seen above salt in the pump bowl at the MSRE. If the mist formation rate decreases as the power density in the liquid decreases, the concentration of salt in the mist should also decrease with decreasing power density in the liquid. Since the salt used in the MSRE distillation experiment had a much lower power density (400 days decay for distillation feed as compared with less than 30 days decay for salt samples tested for mist formation) than salt samples from the MSRE, it seems unlikely that the concentration of salt in the gas above the salt would have been high enough to explain the high relative volatilities for the slightly volatile fission products. In addition to the argument against the entrainment hypothesis given above, not all discrepancies would be explained by it. For example, it would not account for the variations in the ®°Sr-activity-to-® °Sr-activity ratio or for the low value for the effective volatility of 137CS. Concentration polarization would cause the effective relative volatilities of the slightly volatile materials to be greater than the true relative volatilities. As more- ORNL OWG 70-4513 'oo T I i I 1 ‘ I + T 1T 1Ty r L} I . T I v l 1 A 10.0 10} RELATIVE VOLATILITY WITH RESPECT TO LIF vor.i'l'*'-lrll 3 L b L t 2 3 A & o 5 ; e T 8 LITERS CONDENSATE COLLECTED - : ’ Fig. 24.4, Effective Relative Volatility of ' *7Cs in MSRE Distillation Experiment, Showing Effect of Variation of Assumed Feed Concentration. 322 Table 24.1. Ratio? of 3%Sr Activity to *°8r ~ Activity in Condensate Samples from MSRE Distillation Experiment Ratio? of 827 Activity to 90g; Activity >10 - >10 1.89 2,120 >10 >10 1.21 0.223 0.386 ' 2.30 10 \ , - 0.713 11 ' 1.38 . 4Ratio of activities on July 8, 1969, bpuplicate samples did not agree. C.ondensate Sample No. -hg : WO NL TR W - volatile materials vaporize from the surface of the liquid, the slightly volatile materials would be left behind on the surface at a higher concentration than in the liquid just below the surface. The concentration of these slightly volatile materials in the vapor would then increase, since further vaporization would occur from a liquid with a higher surface concentration of slightly volatile materials. Since effective relative volatilities were based on average concentrations in the still pot, the vapor concentration would be higher than that corresponding to the average liquid concentration, and the calculated effective relative volatility would be higher than the true relative volatility, Concentration polarization would cause the effective relative volatility to be lower than the actual relative volatility for a component whose relative volatility is greater than 1. The extent to which concentration polarization af- fects the effective relative volatility of a particular component depends on the dimensionless group D/vL, which qualitatively represents the ratio .of the rate of diffusion of a particular component from the vapor- liquid interface into the bulk of the still pot liquid to the rate at which this material is transferred by convection to the interface by liquid moving toward the vaporization surface. In this ratio, D is the effective diffusivity of the component in interest, v is the velocity of liquid moving toward the interface, and L is the distance between the interface and the point where the feed is introduced. | The occurrence of concentration polarization is sug- gested by the sharp increase at the beginning of the run in the effective relative volatilities of 144Ce, *47Pm, 155Ey, and possibly of ®!'Y and °°Sr. This increase would correspond to the formation of the concen- tration gradient in the still pot liquid. The effective diffusivities of NdF; in the still pot, calculated from results of the nonradioactive experiments, ranged from 2.1 X 10™ to 18.3 X 10™* cm?[sec and form the basis for estimating the magnitude of the concentration polarization effect in the radioactive operation. During the semicontinuous operation at the MSRE, the liquid velocity - resulting from vaporization averaged 2.2 X 10™* cm/sec, and the depth of liquid above the inlet was about 94 cm. If one assumes that the effective diffusivities of the fission products in the still pot during the MSRE distillation experiment were in the same range as they were during the nonradioactive tests, the observed relative volatilities of the slightly volatile materials would be only 1.8 to 11.5 times the true relative volatility, and the observed relative volatility of 137Cs would be 0.012 to 0.023 times its true value (assuming in each case that the true relative volatilities were those given in refs. 1 and 2). Although concentra- tion polarization may have been significant in the work with radioactive salt, the effect is not great enough to account for the discrepancies between observed relative volatilities and what we consider to be the true values, Also, concentration polarization would not explain the variation in the ratio of ®9Sr activity to ®°Sr activity between samples of condensate. The possibility that the condensate samples were contaminated while they were being prepared for radiochemical analysis is suggested by the wide varia- tion in the value of the 3°Sr/?%Sr activity ratio. Although routine precautions against contamination were taken in the hot cells, where the capsules were cut open, no special precautions were taken, and the manipulators used to handle MSRE salt samples were also used to open the condensate samples. If it is assumed that the source of the contamination was the last salt sample taken from the MSRE before the distillation samples were submitted, only 107 to 1073 g of salt per gram of sample would be required to yield the observed values of the 32Sr/°%Sr activity ratio. ‘Contamination from such small quantities of material would be extremely difficult to prevent. Other observations explained by assuming that the samples were contaminated are the high relative vola- tilities of the slightly volatile fission products and the “high correlation between the variations of calculated o (» - 323 relative volatilities of different fission products. The low relative volatility for ! 37Cs is not explained by this hypothesis. . We conclude that, although several factors may be involved, the discrepancy between the effective relative volatilities of the slightly volatile materials measured in this experiment and the values measured previously is primarily the result of contamination of the condensate samples by minute quantities of salt from other MSRE salt samples in the hot cells. 25. System Calculations 25.1 MSBR PROCESSING PLANT MATERIAL AND ENERGY BALANCE CALCULATIONS Material and energy balance calculations for the reductive extraction flowsheet and conditions shown. in W.L.C arter Fig. 25.1 have been recomputed using an updated library of physical data for the fission product nuclides. Also, the computer program (CALDRON) by which the calculations were made was modified to make it a more accurate model of the process. ‘ These data are for a 2250Mw (thermal) MSBR operating with an active volume of 1461 ft® of Q.25 gpm ORNL-DWG 69-3518 C 185.6 g FP/1t3 2.4 kw/ttd 2.27T gpm Ha+HF 247.52.9 FR/#3 2 : 6.9 kw/1t3 i | oo MAKE UP | I I LiF-BeFa-ThF, LFg—e=UFy | : 0.0025 gpm REDUCTION I | 00 ' . | Ha 1 Satan g FP/tid v I | .36 g FP/M 5 2iuytr RECOVERY | I { 600° f— M ! ! i | BiREMOVAL AND Fe ' 00187 gpmn B } SALT CLEANUP | i -! i }— i | i o | 1 | : 1 | ~1 L I | ! | RARE EARTH ACCUMULATION | | AND B3pg pECAY : ok ; [ 55-DAY HOLOUP RECOVER EACH TANK 39.6 fi i g;g Y : THORIUM i 1922 QFP/Ht : IMF| , REFLUX I 1 1 : 600% I 33 kw 1 ! 00162 gom 200.77 9 FP/i3 M 0.253 gpm 100%-400°%C 1 | 35 kw/f3 42 kw/it3 I L : | { ! 1922 g FPri® I RARE €ARTHI i I lo48759pm | S3aem EXTRACTION | ! I |wzzgreie | 1.2 Ww/fy OO m— e — 1 E3kw/H3 REACTOR | 198649 FP/(t3 i 2250 Mw (thermat) i EXCESS 1 1461 3 FUEL -Po ' UFg 1 LiF-BeF,~ThF,-UF, EXTRACTION \ i 1319 Usd 15 gpm| BR=106 600°C i LiF-ZrFs I I RESERVOIR ! | 5# boirturg I 429 kg FP/f;" A — ;___..___,._r:\_r 2.53 gpm ~ —.} Hp+HF | [ 136 kw/tt i 56 kw/t13 1 ! 233pq DECAY | ¢ ' rwogmnon 200.#3 . I e —— i s | F2 r‘ 507 g Pa/ft i i i } 26,7 kw/ttd i t | BISMUTH | Lo i | RESERVOIR ! 2.4 gpm 1 1 o 543 - — - 376 kw3 § i ¢ 550° FLUORINATION a7.71 g FR/113 ® 0 1 Bi 850°C . 402 w/ftd - 3 . HeHr S8 f Il {HYDROFLUORINATOR f gfi:}:féd“ % ' I { HF 0.9875 gpm 1922 g FR/#3 2 FLUORINATION oo I} 1922 g /113 : u TION | 0.0 Eslpg_\P__ 63 kw/ft3 =1 rewm 0.0245 t¥day B Teooe | 423 kg FR/it3 WASTE TAN 1 133 kw/f13 s ™ Fa I 500 1> : | FEED TANK smyTH | | 543 REBSlERVOlR | SALT STREAM ———Bi STREAM 5.3q Po/ftt3 L el )—J 243,36 g FP/#13 2033 g FP/AH3 16.7 kw/t13 1.3 kw3 Fig. 25.1. Reductive Extraction Process for MSBR Fuel Salt. 324 i " h » " 325 Table 25, l Dominant Mechanisms and Process Cycle Times for Removal of 233Pa and Fission Products from MSER Fuel Salt Cycle Time for Primary Mechanism of Processing from Salt Removal from Salt 233p, _ : 3 days Reductive extraction into bisniuth, plus trapping _ _ - ' in salt for tadmactlve decay Zn, Ga, Ge, As, .- 200days Reductive extraction into blsmuth, plus hydro- Se, Zr,Cd, In, ‘ - fluorination into waste LiF-ZrF 4 salt Sn, Sb Br, 1 . ‘ - 60days Volatilization as fluorides i in pnmary ’ o fluorinator Kr,Xe 50 sec Sparged from salt with inert gas in reactor circulating salt loop Nb, Mo, Tc,Ru, - 50 sec Assumed to be reduced by metal surfaces Rh, Pd, Ag, Te in reactor and to plate out La, Ce, Pr, Nd, 50 days Reductive extraction into bismuth, plus concen- Pm, Sm, Gd, Tb, Dy, ‘ tration in fuel salt for waste discard Ho, Er, Y - : Rb, Sr, Cs, Ba, Eu 3000 days Removed only by discard of fuel salt LiF-BeF, -ThF,-UF, (71.7-16.0-12.0-0.3 mole %) fuel . salt, In the calculations it was assumed that salt for processing is withdrawn continuously from the reactor at 2.53 gpm and that purified makeup salt is added at the same rate. Values on . the flowsheet pertain to equilibrium conditions when fission products and 233Pa are being removed according to the cycle times given in Table 25.1. This table denotes the dominant removal process and cycle time for 223Pa and for each family of fission product elements; for all practical -purposes, most of the fission products are removed by more than one mechanism, For example, _the noble gases are removed primarily by sparging the fuel salt in the MSBR fuel circuit; however, in fuel processing, removal is also accomplished whenever the salt is fluorinated or treated with an H, purge stream. 25.2 EFFECT OF CHEMICAL PROCESSING 'ON THE NUCLEAR PERFORMANCE OF ANMSBR M. J. Bell _ L. E.McNeese A series of calculations has been ‘per‘fo'rined to investigate the effect of the removal of individual fission product elements on the performance of an MSBR. These calculations were made with a computer code called MODROD! which resulted from combining the ROD? (reactor optimization and design) and the MATADOR? (steady-state material balance) codes. This combined code allows an accurate treatment of the neutron production and loss in the reactor core and a more complete description of radioactive decay and neutron capture by the fission products. The calcula- tions were performed for a 2250-Mw (thermal) single- fluid MSBR fueled with 1680 ft* of fuel salt of nominal " composition = 71.7-16.0-12,0-0.3 mole % LiF-BeF,- ThF,-UF,. Further details of the reactor concept are given in ref. 4. In the calculations which were performed, the effect of chemical processing on the neutron poisoning by - individual fission product elements was investigated by varying the efficiency with which an element was removed from the fuel salt with respect to a reference - value. The reference case assumed that: (1) halogens ~ and rare earths were removed on a 50-day cycle, (2) zirconium and seminoble metals were removed on a 200-day cycle, (3) protactinium was removed on a S-day Iy, E. Whatley et al., Unit Operations Sect. Quart. Progr. Rept. April-June 1969, ORNL-4532 (in preparation). 2MSR Program Semmnn. Progr. Rept. Aug. 31, 1968, ORNL-4344, pp. 68—-70. 3MSR Program Semiann. Progr. Rept. Feb, 28, 1969, ORNL- 4396, pp. 275-18, 4MSR Program Semiann. Progr. Rept. Feb. 28, 1969, ORNL- 4396, pp. 77-79. ‘ 326 " cycle, and (4) the salt was discarded on a 3000-day cycle to remove the active metals and neptunjum. In addition the noble gases and noble metals were treated as described in ref, 3. They were assumed to have a 50-sec residence time in the fuel salt and a 110-sec residence time in the helium bubbles; the surface resistance for diffusion of the noble gases to the graphite was fixed to give a poisoning of 0.5% for 135Xe. The removal efficiency for each individual element from Zn to Ho, except for Kr and Xe, was varied from 2% to 27°. These conditions differ slightly . from those used by Baumann® to optimize the reactor design and resulted in a breeding ratio of 1.061 and a fuel yield of 3.2% per annum for the base case. The cost associated with the inventory and replacement of fuel and carrier salt amounted to 0,27 mill/kwhr, Figures 25.2—25.5 illustrate the effect of chemical processing for the elements Nd, Sm, Pm, and Zr on the performance of the reference MSBR. The changes in fuel yield and fuel cycle cost from those determined for the base case are shown as functions of element removal time. The changes in fuel cycle cost result only from changes in the inventory charges for additional fissile material that would be required if an element were removed less efficiently than in the base case. The elements whose -poisoning can be influenced most by . chemical processing are the rare earths, Zr, Ba, and Sr. For a number of materials, primarily the noble metals, no significant variation in reactor performance was observed over the range of removal times investigated. The lack of influence of the noble metals results from the assumption that these materials are removed on a short (50-sec) cycle in the base case. However, if longer removal times were established, removal of these “elements by chemical processing would become im- portant. Additional results and details of the com- putational procedure are given in ref. 1. ORNL DWG 70-4509 00 |- -10 - CHANGE IN FUEL YIELD (% per annum) Li il L 4 el 22 a3l . L e aaal ]‘l’fi'—r'llil | e @ 8 8 CHANGE IN FUEL CYCLE COST (mill/kw hr) 10.04 1 o o ‘N 4000 1 ‘ o 5 10 20 50 100 - 200 500 1000 2000 5000 REMOVAL TIME (days) Fig. 25.2. Effect of Neodymium Removal Time on MSBR Performance. 0.00 w -0.25 ~0.78 CHANGE IN FUEL YIELD (% per annum) ~1.00 -1.25¢% 5 iy 327 ORNL DWG 70-43510 Trevr] 1 st a2 aal 1 —0.04 —0.03 -10.02 CHANGE IN FUEL CYCLE COST (mill/kw hr) 80 100 200 REMOVAL TIME (days) 500 1000 2000 5000 Fig. 25.3. Effect of Samarium Removal Time on MSBR Performance. ORNL DWG 70-4511 0.0 CHANGE IN FUEL YIELD (% per annum) in FrrT] 2 aaald 1 Ledaaad 1 i 1 Lo ol 1 i —o.03 —0.02- ~10.04 CHANGE IN FUEL CYCLE COST (mill/kw hr) .—1=-0.01 10 20 80 1co 200 800 ~ 'REMOVAL TIME (days) 1000 2000 5000 Fig. 25.4. Effect of Promethium Removal Time on MSBR Performance. CHANGE'IN FUEL YIELD (% per annum) 328 ORNL DWG 70-4512 T T T 1 T T ] rvrr] T T ] v 1T T1] 1 Ll 1 2 1 g st 1 2 Ll a5l 1 50 100 200 500 1000 2000 SOOQ 104 2x10* ‘ REMOVAL TIME {days) ' Fig. 25.5. Effect of Zirconium Removal Time on MSBR Performance. 0.03 0.02 0.01 0.00 -0.04 CHANGE IN FUEL CYCLE COST (mill/kw hr) ¥} “w 2] " w 26. Continuous Salt Purifibation System R. B. Lindauer E. L. Youngblood Equipment is being installed to study the continuous purification of moltensalt mixtures as a possible replacement for the present batch method.! The first step to be studied will be the continuous reduction of dissolved iron fluoride with hydrogen. The equipment (Figs. 26.1 and 26.2) to be used in this study consists of a .1%-in.-diam by 7-ft-long nickel column packed with- 1,-in. nickel Raschig rings, a 0.69-ft*> fiber metal filter (50-u rating) for separating the reduced iron from the salt, a flowing stream sampler, and two 20-liter tanks. L.E.McNeese E. L. Nicholson The interface will be positioned at the bottom of the column by means of a control valve in the exit gas stream, We will follow the progress of the reduction step by analyzing salt samples and by continuously analyzing the HF content of the column exit gas stream. A sodium fluoride trap will be used to remove ARGON 6@ | |coLumn SALT - FEED TANK HYDROGEN (& HF) FILTER - \MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL- 3708, pp. 228-303, | BACK PRESSURE 1 - CONTROL VALVE [ ORNL-DWG-69-14298 Y | Y WATER FLLUORIDE ANALYZER - ANALYZER | ] NaF TRAP ’ VENT : —#= OUTSIDE : : ' : BLDG. ABSOLUTE FLAME F ILTEB ~ ARRESTER ~ |Recever - . Fig. 26.1. Equipment Flowsheet for the Continuous Sa!t Purification Systeni. 329 Fig. 26.2. Continuous Salt Purification System in Cell 4B, Building 4505, PHOTO 47229 € " gy {w HF from the off-gas stream before it passes through an absolute filter and is discharged. | ~ Construction of the system in Building 4505 has been completed, and the equipment has been leak tested. Construction of the HF, argon, anid hydrogen supply systems is still in progress. The argon and hydrogen systems will contain oxygen and moisture removal 331 equipment and instrumentation for continuously mon- . itoring trace amounts of these gases. The first steps in the experimental program will consist of charging 15 liters of molten salt (66-34 mole % LiF-BeF,) to the system and determining the . flooding rates of the column, both with the purified salt and with salt containing about 150 ppm of oxide. The oxide will be added to the salt in the feed tank and will later be removed by sparging with H, -HF. It is expected that salt flow rates as high as 200 cm?/min can be used with hydrogen gas rates up to 100 liters/min. Initial operation will be at 700°C. After the study of the iron reduction step has been completed, the oxide removal - step will be demonstrated in the column, Studies with salt containing thorium fluoride will also be made. 27. Désign and Preparation_“o'f 23-9PuF3 Capsuieé _fbr | Small Refueling Additions to the MSRE W.H.Carr W. F. Schaffer, Jr. Specially designed refueling capsules’ were loaded with 23°PuF; powder and added to the MSRE fuel - salt.> Eight capsules were filled with a total of 311 g of 239pyF,; powder in glove-box facilities in Building 3019. The first of these capsules contained 31.6 g of PuF;. As we improved the loading technique, we were lw. H. Cam, W. F. Schaffer, and E. L. Nicholson, MSR Program Semiann. Progr. Rept. Aug 31, 1969, ORNL-4449, pp. 245-46. ZPart 1, this report. D. R. Taylor E.L. Nicholson able to compact more PuF; into a capsule; each of the last six capsules contained 39.2 to 42.2 g of PuF;. No problems were encountered in the filling operation or in packaging the capsules in the MSRE fuel sampler- enrichér charging containers. The exteriors of the loaded charging containers were smeared and found to be free of transferable activity before the containers were sealed in plastic bags. No detectable contamina- tion escaped from the glove-box operation. Similar techniques .could probably be used with much larger powder containers to add plutonium to a barren salt. 332 1) w) 4“1 OAK RIDGE NATIONAL LABORATORY MOLTEN-SALT REACTOR PROGRAM FEBRUARY 28, 1970 M. W. ROSENTHAL, DIRECTOR R [ R.B. BRIGGS, ASSOCIATE DIRECTOR 0 . R KASTEN, ASSOGIATE DIRECTOR R W.P.EATHERLY,” GRAPHITE PROGRAM R H.R.BEATTY,® BUDGET B = T MEBR DESIGN STUDIES COMPONENTS & SYSTEMS DEVELOPMENT INSTRUMENTATION & CONTROLS PHYSICE : MATERIALS . MEBR PROCESSING DEVELOPMENT CHEMISTRY MERE OPERATIONS " DUNLAP SCOTT** . $.4.DITTO" e A.M.PERRY* R J':“:t;:;r.’:i.' e M. E. WHATLEY or - Ac #.N. HAUSENREICH - - . %i.m-“ ;E_ I I MSRE PHYSICS | HASTELLOY N STUDIES CHEWNCAL DEVELOPMENT PORT-OPERATION C.W.COLLING® R B. E.PRINCE R . . GEHLBACH MaC 2.C. MAILEN L. M, FERRIS g SURVEILLANCE AND INSPRCTION P . o oevtsomsnr rroces weTn e AN o siron L xoue o || dvim, & nexcron cremer . e . ’ A.G. GRINDELL n . R.L.MOORE* 8¢ G. M, SLAUGHTER* M&C F.J.SMITH cr kensthp RC " " imm,mm. " . ™ n F.G. HERNDON® 18c O R RSWELL IR n B. MONABE MBC JE.LAND cr MANN" nc 1. R.SHUGART R 3, R MWHERTER" R L. V. WILSON"* R : wh ‘ A C.T. THOWPSON cr A nc SUMMARY REPORTS ON OPERATIONS H. A NELMS® " M.C.YOUNG® R - H.T.KERR n GRAPMITE STUDIES , B H. GUYMON A R. C. ROBERTSON " 0. D. OWENS " ' . 0.B.CAVIN . MaC EXPERIMENTAL ENGINEERING LLSRE ONSITE CHEMISTAY J. R FRANZRES n 2. A TALLACKSON T MBER CORE DESIGN W, H.COOK mac . L.E.MoNEESE er o A1, KAAKOVIAK n L V.WILSON"® R I NUCLEAR INSTRUMENTATION AND M. F. BAUMAN " R. L HAMNER® mac B. A. HANNAFORD cr A :g 3L, STEPP n H. L. WATTS - CONTROLE ANALYSIS 5 1 CHANG' r C. R.KENNEDY MAC J. R HIGHTOWER, JR. cr PANN ‘ W. TERRY ‘A . - : R. W.McCLUNG® MEC C.W. KEE er ERE RC WUCLEAR AND MECHANICAL ANALYSTS W.C. GEORGE ‘R COMPONENTS AND SYSTEMS 4,4, ANDERSON*® Iac MIBR EXPERIMENTAL PHYEICS B. L. McELROY® M&C M5 LN cr E:: RC H.M.POLY R DUNLAP SCOTT** n 0.W.BURKE* 18C £®. MOORE® wac R. B, LINDAVER - cT RFORD RC . tfii'iuno : . B. GALLANER F.H.CLARK* 18 G. L. RAGAN R ¢, B, POLLOCK* MBC J ROTH? cr - :: &A::::y° : : W. H, SIDES 8¢ . C.P.TUNG' . or AL AND INORGANIC CHEMSTRY A HOUTZEEL : A.E. HELMS n . CONSULTANT CHEMICAL PROCESSING MATERIALS J.5. WATSON . cT Rt RC R ifi:f;sm R R.J KEDL* R 7. WKERLIN uT N.C, coLe* o E. L. YOUNGBLOOD cT ;:EHGER RC . : AN BMITH R 4 R.DISTEFAND® MAC 3 BEAMS o : R HEAT TRANSFER AND PHYSICAL €. A. GIFFORD" R .|:I .FEDEI\ER' N&C V.L.FOWLER rcflgmrg?mpwhupgpepm o I 2 3 335 INTERNAL DISTRIBUTION . B, Cavin . M. Chandler - I Clark S. J Cromer (K-25) ). L. Crowley . L. Culler DeVan oy w2 -+ & 0 = -0 ederer - Ferguson o . Ferris . Fraas . Franzreb . Friedman . Fry , . Frye, Jr. C Fuller - . K. Furlong . H. Gabbard EzplfRozm™ 106. 107. 108. 109. 110. 111. 112, 113. 114, 115. 116. 117. 118. - 119. 120. 121. 122. 123. 124, 125. 126. 127. 128. 129, 130. 131. 132. 133. 134, 135. 136. 137. - 138, 139, F - 140. 141. 142. 143, 144, 145. . 146, 147. '148. 149. 150. Oggp?”?FE*>WWFEPWS“F9”F“”PEPWWfiFF“W” W. R. Gall R. B. Gallaher R. E. Gehlbach J. H. Gibbons R. G. Gilliland L. O. Gilpatrick G. Goldberg W. R. Grimes A. G. Grindell . W. Gunkel . H. Guymon . P. Himmond H111 = o e ™ . W. Horton . Houtzeel . L. Hudson .R. Huntley - . ORNL-4548 UC-80 — Reactor Technology 151. 153. 154, 155. 156. 157 158. 159. 160. 161. 162, 163. 164. 165. 166. 167. 168. 169. 170. 171, 172. 173. 174, 175. 176. 177. 178, 179, 180. 181. 182, 183, 184. 185. 186. 187. " 188, 189. 190, 191 192. 193, 194, 195. b 7~ (1] . g @ v . Lyo . Ma cklln . MacPherson . Maienschein D. L. McElroy C. K. McGlothlan 'C. J. McHargue H. A. McLain B. McNabb L. E. McNeese J. R. McWherter H. J. Metz A. S. Meyer C. A. Mills 196. - 197. 198, 199. 200, 201. 202, 203. 204, 205. 206. 207. 208. 209, 210-211. 212. 213. 214, - 215. 216. 217. 218. - 219. - 220. 221, 222, 223. 224. 225. 226. 227. 228-278. 279. -402. 403. 404, 405. 406. 407. 408. - 409. 410. 411. 412, 413, 414, " 415, ngWQZUHHomw;nma>wwsmrfimfimzflcom>w . L. Moore . J. Moorhead . Z. Morgan . A. Mossman . M. Moulton . R. Mueller . L. Myers - . H. Nichol . P. Nichols . L. Nicholson . S. Noggle. . C. Oakes - . M. Ohr . R. Osborn . B. Parker . Patriarca . M. Perry . W. Pickel . B. Piper . B. Pollock M. Poly . E. Prince . P. Raaen . L. Ragan . L. Redford . D. Redman M Richardson . Richardson . D. Robbins . C. Robertson . J. Rose . W. Rosenthal . G. Ross 280. 281. 282. 283. 284. 285. 286. 287. 288, 289, 290. 291. 292, 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309--310. 311. 312. 313. 336 . Roth . P. Sanders . C. Savage . F. Schaffer E. Schilling unlap Scott . Scott . Seagren Sessions Shaffer Skinner ‘Slaughter L E .E. .H. . D, . H. Sides I .M. .N. Smith L .P. .L. (itng Splewak . H. Stone . A, Strehlow D. Stulting . A. Sundberg . R. Tallackson . H. Taylor . Terry . E. Thoma . Thomason .Toth . Trauger Urmzamucpwmn wE My . E. Stevenson - 314. 315. 316. 317. 318. 319, 320, 321. 322, 323, 324, 325. 326. 327. 328, 329, 330-335. 336. 337 R 338, 339, - 340. 341, 342. 343, 344, 345-346, 347-349. 350—400. 401. EXTERNAL DISTRIBUTION J. A. Acciarri, Continental Oil Co., Panca City, Oklahoma, 74601 J. S. V. Andrews, Atomic Energy Attache, UKAEA, British Embassy, Washington, D.C. 20008 J. G. Asquith, Atomics International, P.O. Box 309, Canoga Park, California 91304 Bruce L. Bailey, Great Lakes Carbon Corp., Pine Ave. & 58th, Niagara Falls, N.Y. 14302 N. W, Bass, Brush Beryllium Co., 17876 St, Clair Ave., Cleveland, Ohio 44110 David Bendaniel, General Electric Co., R&D Center, Schenectady, N.Y. J. C. Bowman, Union Carbide Technical Center, 12900 Snow Rd., Parma, Ohio 44130 G. D. Brady, Materials Systems Division, UCC, Kokomo, Indiana 46901 A. L. Travaglini ' - ) R. W, Tucker . ' U Chia-Pao Tung W. C. Ulrich W. E. Unger - D. C. Watkin - . M. Watson . g - .S. Watson : q .L.Watts - . F. Weaver . H. Webster . M. Weinberg . R. Weir ' . J. Werner . W, West . L. Whaley . E. Whatley . C. White . P. Wichner . V. Wilson . J. Young . C. Young J. P. Young E. L. Youngblood F. C. Zapp : , Biology Library (Q ORNL — Y-12 Technical Library > Document Reference Section Central Research Library - Laboratory Records Department Laboratory Records, ORNL R.C. EOI"‘FU"‘Z:ENS'-‘>UUO:E"‘O R. M. Bushong, UCC, Carbon Products Div., 12900 Snow Rd., Parma, Ohio 44130 - Pedro B. de Camargo, Brazilian-Comissao Nacional, Energia Nuclear, Calxa Postal 11049 Sao Paulo, Brazil Premo Chiotti, Ames Laboratory, Iowa State University, Ames, Iowa 50010 ' C Paul Cohen, Westinghouse Electric Corp., P.O. Box 158, Madison, Pennsylvania 15663 | = D. F. Cope, Atomic Energy Commission, RDT Site Office (ORNL) J. W. Crawford, Atomic Energy Comm:ssmn, Washington, D.C. 20545 " at 416, 417. 418, 419. 420. 421. 422, 423, 424, 425, 426. 427. 428, 429, 430. 431, 432, 433, 434, 435, 436. 437. 438. 439. . R. A. Langley, Bechtel Corp., 50 Beale St., San Francisco, California 94119 441. 442, 443, 444445, 446. 447, 448, 449, 450, 451. 452, 453, 454, 455. 456. 457. 458. 459. 460. 461. 462. 463. 464. 465. 466. 467. 337 F. E. Crever, National Nuclear Corp., 701 Welch Road, Palo Alto, California 94304 M. W. Croft, Babcock and Wilcox Company, P.O. Box 1260, Lynchburg, Virginia 24505 Walter A. Danker, Jr., Westinghouse Electric, P.O. Box 19218, Tampa, Florida 33616 C. B. Deering, Black & Veatch, P.O. Box 8405, Kansas City, Missouri 64114 Deslonde R, deBoisblanc, Ebasco Services, Inc., 2 Rector St., New York, N.Y. 10006 A. R. DeGrazia, USAEC, DRDT, Washington, D.C, 20545 D. A. Douglas, Materials Systems Division, UCC, Kokomo, Indiana 46901 Donald E. Erb, Battelle Memorial Institute, 505 King Ave., Columbus, Ohio 43201 H. L. Falkenberry, Tennessee Valley Authority, 303 Power Building, Chattanooga, Tenn. 37401 C. W. Fay, Wisconsin Michigan Power Co., 231 W. Michigan St. Mxlwaukee, Wisconsin 53201 J. E. Fox, USAEC, DRDT, Washington, D.C. 20545 Gerald Golden, Argonne National Laboratory, 9700 S. Cass Ave., Argonne, Ill., 60439 A. Goldmen, UCC, 270 Park Ave., N.Y., N.Y. 10017 W. J. Gray, Battelle-Northwest, 2325 Enterpnse, Richland, Washngton 99352 W. W. Grigorieff, Assistant to the Executive Director, Oak Ridge Associated Universities Norton Habermann, RDT, USAEC, Washington, D.C. 20545 Irving Hoffman, USAEC, Washington, D.C. 20545 'Harry Honig, Babcock & Wilcox, P.O. Box 1260, Lynchburg, Va. 24505 E. E. Kintner, U.S. Atomic Energy Commission, Washington, D.C, 20545 Brice W. Kinyon, Combustion Engineering, 911 W, Main St., Chattanooga, Tenn. 37402 Gene Kramer, Southern California Edison Co., P.O. Box 351, Los Angeles, Calif. 90053 P. M. Krishner, Pioneer Service and Engineering, 400 W, Madison St., Chicago, Ill. 60606 J. Ladesich, Southern California Edison Co., P.0. Box 351, Los Angeles, Calif. 90053 L. W. Lang, Douglas United Nuclear, 703 Bldg., Richland, Washington 99352 D. Manly, Cabot Corp., Stellite Division, Kokomo, Indiana 46901 P. Mays, Great Lakes Carbon Co., 299 Park Avenue, New York, New York 10017 B. McDonald, Battelle-Pacific Northwest Laboratory, Hanford, Washington 99352 W. McIntosh, Atomic Energy Commission, Washington, D.C. 20545 C. McKinley, ACRS Office, USAEC, Washington, D.C. 20545 J. Mordarski, Nuclear Development, Combustion Engineering, Windsor, Connecticut 06095 A. Muccini, Ashland Oil Inc., R&D Building, Ashland, Kentucky 41101 A. Nystrom, Stackpole Carbon Company, St. Marys, Pa. 15857 E. H. Okrent, Jersey Nuclear Co., Bellevue, Washington 98004 , William E. Parker, Airco Speer Research, 47th & Packard Rd., Niagara Falls, N.Y. 14302 Sidney J. S. Parry, Great Lakes Carbon Corp., P.O, Box 667, Niagara Falls, N.Y. 14302 F. N. Peebles, Dean of Engineering, University of Tennessee, Knoxville, Tenn. 37900 H. G. MacPherson, University of Tennessee, Knoxville, Tenn. 37900 Kermit Laughon, Atomic Energy Commission, RDT Site Office (ORNL) C. L. Matthews, Atomic Energy Commission, RDT Site Office (ORNL) A. J. Pressesky, U.S. Atomic Energy Commission, Washington, D.C. 20545 M. V. Ramaniah, Head, Radiochemistry Division, Ahabha Atomic Research Centre, Radxologlcal Laboratories, Trombay, Bombay-85 AS, India David Richman, Research Division, USAEC, Washington, D.C. 20545 w. J. Ww. T. J. Ww. G. W. “J. A. L. Robertson, Atomic Energy of Canada Ltd., Chalk River, Ontario, Canada J. C. Robinson, Dept. of Nuclear Engineering, Univ. of Tenn., Knoxville, Tenn. 37900 T. K. Roche, Stellite Division, Cabot Corp. 1020 Park Avenue, Kokomo, Ind. 46901 M. A. Rosen, Atomic Energy Commission, Washington, D.C. 20545 H. M. Roth, Atomic Energy Commission, ORO : R. O. Sandberg, Bechtel, 220 Bush Street, San Francisco, Callf 94119 R. W. Schmitt, General Electric Co., Schenectady, New York 12301 W. Schréck-Vietor, Kernforschungsanlage Julich, 517 Julich, Germany - 468. 469. 470. 471. 472. - 473 474, 475, 338 R. N. Scroggins, U. S. Atomic Energy Commission, Washington, D.C. 20545 M. Shaw, Atomic Energy Commission, Washington, D.C. 20545 Winfield M. Sides, Northeast Utilities Service Co., P.O. Box 270, Hartford, Conn. 06101 E. E. Sinclair, Atomic Energy Commission, Washington, D. C 20545 W. L. Smalley, Atomic Energy Commission, ORO Earl O. Smith, Black & Veatch, 1500 Meadowlake Parkway, Kansas Clty, Mo. 64114 T. M. Snyder, General Electric Co,, 175 Curtner Ave., San Jose, California 95103 N. Srinivasan, Head, Full Reprocessing Division, Bhabha Atomic Research Centre, Trombay, Bombay - 74, India 476. 477. 478, 479, 480. - 481. 482, 483. . M. Tsou, General Motors, 12 Mile & Mound Rds., Warren, Michigan 48089 485. 484 486. 487. 488, 489. 490. 491. 492, 493. 494, 495. 496. 497, 498. 499, ~500. 501—708. Philip T. Stroup, Alcoa, P.O. Box 772, New Kensmgton, Pennsylvama A. E. Swanson, Black & Veatch, P.O. Box 8405, 1500 Meadowlake Parkway, Kansas Clty, Mo. 64114 J. A. Swartout, UCC, New York, New York . Richard Tait, Poco Graphite, P.O. Box 2121, Decatur, Texas 76234 B. L. Tarmy, Esso Research & Engr. Co., P.O. Box 101, Florham Park, N.J. 07932 C. L. Storrs, Combustion Engineering Inc., Prospect Hill Road, Windsor, Conn. 06095 D. R. Thomas, Commonwealth Associates, Inc., 209 E. Washington Ave., Jackson, Mlcthan 4920 J. R. Trinko, Ebasco Services, Inc., 2 Rector Street, New York, New York 10006 Denis D. Tytgat, Commission des Communautes Europeennes, 200, Rue De La Loi, 1040 Bruxelles, Belgium o J. W. Ullman, UCC, P.0O. Box 278 Tartytown New York 10591 ‘ G. B. von der Decken, Kernforschungsanlage Julich GmbH, 517 Julich, Germany C. H. Waugaman, Tennessee Valley Authority, 303 Power Bldg., Chattanooga, Tenn. 37401 D. B, Weaver, Tennessee Valley Authority, New Sprankle Bldg., Knoxville, Tenn. 37900 R. F. Wehrmann, Poco Graphite Inc., P.O. Box 2121, Decatur, Texas 76234 J. T. Weills, Argonne National Laboratory, Argonne, Illinois M. J. Whitman, Atomic Energy Commission, Washington, D.C. 20545 Mack P. Whittaker, Great Lakes Research Corp., P.O. Box 1031, Elizabethton, Tenn. 37643 H. A. Wilber, Power Reactor Development Co., 1911 First St., Detroit, Michigan 48200 Karl Wirtz, Battelle Seattle Research Center, 4000 NE 41st St., Seattle, Washington 98105 James H. Wright, Westinghouse Electric, P.O. Box 355, Pittsburg, Pennsylvania 15230 C. B. Zitek, Commonwealth Edison Co., 1st National Plaza, Chicago, Illinois 60690 L. R. Zumwalt, North Carolina State University, P.O. Box 5636, State College Station, Raleigh, N.C. 27607 Laboratory and University Division, AEC, ORO AEC Patent Office, ORO Given distribution as shown in TID-4500 under Reactor Technologv category (25 coples — CFSTI) W