3 4u5sk p515889 7 ORNL-4119 UC-80 — Reactor Technology MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT FOR PERIOD ENDING FEBRUARY 28, 1967 LIBRARY LOAN COPV DO NOT TRANSFER TO ANOTHER PERSON 1f you wish someone else to see this document, send in name with document and the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION A ORNL-4119 UC-80 — Reactor Technology Controct No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT For Period Ending February 28, 1967 M. W. Rosenthal, Program Director R. B.-Briggs, Associate Director P. R. Kasten, Associate Director JULY 1967 OAK RIDGE NATIONAL LABORATORY Qak Ridge, Tennessee operoted by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION ! This report is one of a series of periodic reports in which we describe briefly the progress of the program. Other reports issued in this series are listed below. ORNL-3708 is an especi'o”y useful-report, because it gives a thorough review of the design and construction and supporting development work for the MSRE. It also describes much of the geheral technology for molten-salt reactor systems. Period Ending January 31, 1958 ORNL-2474 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 Januory 31 and Aprii 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 SUMMARY oottt et e e e b e e b b e et e e et e et n e e et e e s e e e e et ae e et b et e 1 INTRODUCTION ....ooovecerrnrneesmnesensseseressesssesmsoess oo OSSOSO 9 PART 1. MOLTEN-SALT REACTOR EXPERIMENT 1. MSRE OPERATIONS ........ccoiiii e et PRV 11 1.1 Chronological Account of Operations and Maintenance ...................... e 11 1.2 Reactivity BAIANCE .........cccooiieieiiieieiieeie ettt ettt st e e ea e e et 14 OB 4 0 1=3 8 1= 1 Lo =T OO P OO PO SRS TSP P PP TP PP PP DRSO ‘14 Circulating Bubbles ... ...ooo.vi v e e e e 17 1.3 Thermal Effects of Operation........c..ccccvveiiiiiiiiiiiiiiicieee e 18 " Radiation Heating .......cocooieoiieee et e e 18 Thermal Cycle HiStOIY ..ottt bbbt e e - 19 1.4 Reactor DYNAMICS ....cocovieiiiieieeeiie e eeteieae sttt eoiis e ebesses sear e e era s e e s e s et s e e s e s e e st bt n e s ens e 20 1.5 Equipment PerformancCe............cccoouvriviiienviciccien e et e et e aa e 21 Heat Transfer .......cccocciiviiiiiiiii e e e FOTURUUP 21 Main BIOWELS ..ooooiiiiiiiiiiieeeieeeeeeeee e e e e e reraeen e 25 | S Ts BT Tae) gl D8 Vel L 1-3 1 b L= OO UUS PSP RRPPP PR 26 Off-GaS SYSLEIMS ....vrv vt everoseee e ee e eeeeeessse e es b s s 12 et ems s es et ettt een e s aa e 27 Cooling-Water SyStems ........c.ccocooervinccrincnnn et e eee et e aaeeba e ere et b e a et ee e 31 Component-Cooling SYSOM . ... oot et e bbb 32 Salt-Pump Oil SYSEEMS ... e et e 34 N =Yoo s (ot D RS AT (=Y 1 FOUU U U O ORI UTU PSP PPSPP IO PR PPPTP P 34 | 3 (R 1 c< ST U OO OO PP PP PPTT PP PISTOPPPPPPS 35 Control Rods and DIIveS ..........ooooiiiiiiiiii i e 36 Samplers ..........ccocoeeennns TSROSO PR P PP PO VPPPPPPPP R 36 (OFe T N R et 11 = « LAUUURUUTUUOT U PPN 37 2. COMPONENT DEVELOPMENT .........oooeoorooooeseeesooecereoessssssesooees o sosmoes oo sciers o S S 39 2.1 Sampler-Enricher et et e v 39 COMEAIMATIA IO oo e e e e e e e e et e et e e e et et e e ne e e nre e e 39 CaAPSULE SUPPOTt WATE. ..o oot iiitietitsieie et etetes et ceteet et es et et etses s na s s ekt - 40 Seal Leakage ..o SO U OO PTOOP PO 40 | R Te =y 111 Lo - B UUUTUTUUUTUTU U OO OO VOO U OO OO OO PRSI - 40 2.2 COOLANE SAMPIET ..o eeee et eeeeerae e seses s s b s bbb 41 2.3 Fuel Processing Sampler ..............ccccunne. s e et 41 2.4 Off-Gas Sampler............. OO e e re e e et eeeeaeaeaertreaeareee e e et e eeennnte s 41 2.5 Off-GAS FIIter — MK Il oot eeie oot e et e e ettt e e e e e s eb s e se e s emta e e e e e e e e et e eat srn e e e 42 Confe nts iii 2.6 2.7 2.8 2.9 Feltmetal Capacity TOS Lottt eee e ASSUMPLIONS ..ot et e TR U Experimental Procedure ... e oo _ RESUIES ..o et e ettt a e e e e e e e e naeeas evetsieenies Examination of the Mk I Off-Gas Filter ............... s e Pressure DIOP TeStS ...ttt et e e e s e e e e ars e e s aen e nrasre e e nteeraare Disassembly ... e R ObserVations .......................... ettt eeetheeete e e o e e et ee—eae e tet e s eh et e e e s e tesho et e et t e et esene et e e e e ereenne Migration of Short-Lived Gaseous Products into the Graph1te ..................................................... Remote MaintenanCe ............oooiiiiiiiiiie e e e e Summary of Remote Maintenance Tasks Performed ...........ovoeiiiiomeee e e, Radiation Levels ..........ccooooiiiiiiiiice e et e e, Contamination..............cccccevvvvvennn.n. OO TTOTO L TR Conclusion ............... ettt e Eeeeeieteeieeeeeeeettsineeeneontsntehhteeaeeeethateee s beteteeetaaeibeeeteeee ne e aenn e eeteesaes 3. PUMP DEVELOPMENT ... e ettt ettt e et e 3.1 3.2 MSRE PUMPS . ....o .o oo et e e e s et e e s e e e ee s e oo oo oo s e Molten-Salt Pump Operation in the Prototype Pump Test Fac1l1ty .............................................. MEK-2 FUEL PUMP....oioiii ettt e et et et e et e et e et ete et b st e e s Stress Tests of Pump Tank Discharge-Nozzle Attachment ... Spare Rotary Elements for MSRE Fuel and Coolant Salt Pumps............coovvviiveeeeennn. SUSUTU Lubrication R 253 (= 1 TP USSR PSSR URUROR SRR Other Molten—Salt Pumps .......................................................... Fuel-Pump High-Temperature Endurance Test Fac1l1ty .......... JEUTRUORR e e 4. INSTRUMENT DESIGN AND DEVELOPMENT ..ot e e e 4,1 4.2 4.3 4.4 Instrumentation and Controls DeSIEN .......ccovviiiviiiiiiie e ettt Off-Gas Sampler Design ........ et e T TSRS Control System DeSign ... e e et o MSRE Operating EXRpPerienCe ... i e e Control System Relays ..., vt e Temperature Scanner SYSEEM ........c.ccooviiiiii it VaBLVES Lo e e e e TR IMOCOUPLES oo e e e e e Coolant-Pump Radiation Momtor ...................................................................................................... Safety SYSLEM ..ot e e Data System ...... et eeeeteeeeteeeetaeetntaeeaieeern——entaeaeas e et e Instrument Development ..................... P PPRUB e et eae et s Performance — GENeral............ccccooiiiiiii ettt sttt TEMPEIAUIE SCAMMET ........ooviiiiviiiieieeeeee oo ettt ee e et eee et ese e st re et e setereees High-Temperature NaK-Filled Differential Pressure 'I‘ransmltter ................................... e Ultrasonic Level Probe . ... e e e e, MSRE Fuel Distillation System LeVel PrOBE ..oooooooeoeeeeee e Bell-Float-Type Level Indicator for the Mark II Pump........ OO OSSO OIO ' 5. MSRE REACTOR ANALYSIS ............ ettt USRS RURUTRURTUSPUN 79 5.1 Neutron Reaction Rates in the MSRE Spectrum...............c.cooiin e e 79 5.2 Isotopic Changes and Associated Long-Term Reactivity Effects During Reactor Operation ..............cceoeviiiiiiiiiiiereiereecni e s e e e 83 5.3 Analysis of Transient 135Xe POiSOMIME.....c.ccovioieiiieiiet et ettt n et e 86 " PART 2. MATERIALS STUDIES 6. MOLTEN-SALT REACTOR PROGRAM MATERIALS............ccccoiiirincann e 95 6.1 MSRE Surveillance Program — Hastelloy N ... e 95 6.2 Mechanical Properties of Hastelloy N ... SRRV RPUOPPUURRRIN 103 6.3 Precursors of MSBR Graphite.......ccccccooviiviviiinniiii e e e 108 6.4 Graphite Irradiations ............ccccooviviininiciinn et e e, e 110 6.5 Brazing of Graphite ... FO TP ORPR SR OUUUUURURT e 111 Large Graphite-to-Hastelloy-N Assemblies .............ccccoviiiiiiiiiiiii 111 6.6 Corrosion Resistance of Graphite-to-Metal Brazed Joints........cccocciviiiiiiiiiiiiiiini e 111 6.7 . Thermal Convection Loops ........ccooovvivieiiiieiiiic e e e 115 6.8 Evaluation of MSRE Radiator Tubing Contaminated with Aluminum ............ e, 116 A O 7§ 0 0 K3 1 5 SO PO U 118 7.1 Chemistry of the MSRE ......oocooivioueieeereeeeeesoeresse e, et 118 Fuel Salt Composition and Purity................... e J SO ST PR ORRUPPTTPRRY 118 MSRE Fuel Circuit COorroSion ....c.ccocoiiiiiiiiiimiiinii e N 120 Extent of UF, Reduction During MSRE Fuel Preparation ..., e 121 Adjustment of the UF, Concentration in the MSRE Fuel Salt..............cooioiii 123 7.2 Fission Product Behavior in the MSRE ... e 124 ' Long-Term Surveillance SPeCIMEens ........ccoiiviiiiiiiiiiiiii e e e 125 Uranium Analyses of Graphite Specimens ................ccccooeeiiniiiienn, e et e et es 128 FUEL SALt SAMPIES ...ooimioii ittt e e et ettt b et et et e 130 Effect of Operating Conditions ... OO PSPV TO VPP oS 131 Effect of Beryllium Additions .........ccooeoiiimiiiii oot e 131 Pump Bowl Volatilization and Plating TeStS ..o 134 Uranium on Pump Bowl Metal Specimens ........cccocciiiiiiiniviiiicii e 138 Freeze Valve Capsule EXPEriments .........ccccovvoricriiiiiiiiiiiiiiciin st 138 Special Pump BoW]l TeSES ...ttt sttt 141 General Discussion of Fission Product Behavior.............cooooviiiiiiiiin et 142 7.3 Physical Chemistry. of Fluoride MeItS ............ccocoviiririiiiioriiieicess e et e s s 144 . The Oxide Chemistry of LiF-BeF ,-Z1F , Mixtures ... e 144 Solubilities-of SmF ; and NdF, in Molten L1F BeP, (66-34 mole %) ..o 144 Possible MSBR Blanket-Salt M1xtures .............................................................................................. 146 7.4 Separations in Molten FIUOrides ........ococcoooiiiiiiniiiiiiec s 149 Removal of Rare Earths from Molten Fluorides by Precipitation on Sol1d UF, e, 149 Extraction of Protactinium from Molten Fluorides into Molten Metals............... SSUUUTRTRUUR 150 Extraction of Rare Earths from Molten Fluorides into Molten Metals .............cccccoiiiinninnnns 152 7 Protactinium Studies in the High-Alpha Molten-Salt Laboratory .................. e e 153 Preliminary Study of the System LiF-ThF ,-PaF,............. TR UPRSUPTOUROORO 155 vi 7.5 Development and Evaluation of Analytical Methods for Molten-Salt Reactors ......................... 156 Determinations of Oxide in MSRE Salts ...........c.cccccoevrivennnen. e e, 156 Determination of U3*/U%* Ratios in Radioactive Fuel by a Hydrogen Reduction MEEHOM . ... e e e et et b et e e bttt te e eneaas tevrernrrrenes 158 EMF Measurements on the Nickel-Nickel(Il) Couple in Molten Fluorides...........ccoooveinnnn. 162 Studies of the Anodic Uranium Wave in Molten LiF-BeF -ZrF ... 163 Spectrophotometric Studies of Molten-Salt Reactor Fuels .. ..o 163 7.6 Analytical Chemistry Analyses of Radioactive MSRE Fuels ... 164 Sample Analyses ..........c........ et s et 165 Quality Control Program .........c.ccooooviiiiiiiiiiniinen et e e n e e e e aaes TR 165 8. MOLTEN-SALT CONVECTION LOOPS IN THE ORR ........ccccocoiiiiiiiii 167 8.1 Objectives and DESCIIPLION ......ccoooiiioiiiieeieceee ettt ettt st st saesr s srerenesneresaseeee 167 8.2 First Loop EXPEIIMENt . ..ottt e ee e e e e et st e e ste e et eeneb e e enbe et e anennes 168 In-Pile Irradiation ASSEMDBLY ........oooiomiiioteeeete sttt st sttt et es e st ess et etes s sresst et es st eennees 168 Operations....... e te e areeeee iiiteesiiieeeeeseerterteieionererieeatht i bt sbeeeane aeeee e aaaneeaeseenannrees e 168 Chemical Analysis of Salt ... ..., e, 168 [000] ¢ o131 + WU USSR " b e bttt ieraaaanaaaaaraaaaaaaaas 169 Fission Products ...........c.ccooveveinninnn... N s eeerarereeteresrere e b nbeto tteeasaen i bebes e ratba b et aeanan e e srnretene 169 Nuclear Heat and Neutron FLUX ...t e st e ee s sra e s e sree e 169 Hot-Cell Examination of COMPONENTS ..........ooeiiiiiioreiiie i s eeie e et et ee e s et e enerne e 169 8.3 Evaluation of System PerformancCe .............cccocvvriiiiiiiiiniiiei e 171 HeaterS ....cciiiiiiiiieie et e e e b e breeaeeesaasnsreeaaneaeeesraraaenae e 171 0o Yo =3 <=3 U OO O ST UUUUPRIOPPPR PR 171 Temperature Control ... e e n e .1 Sampling and Addition ............cccooveviiiiiiiniii OO U UD PSPPI 171 SAIt CIrCULALION ..oooiiieiiiii e et ettt et et e e et e e s e etrtn e e e e s r e e e e 172 8.4 Second In-Pile Irradiation ASSEmDbIy.......c..cciiiiiiiiiiiir e e 172 OPEIAtioN ...oeeeeeeeeieeeciiiee et e s s e 173 . PART 3. BREE}ER REACTOR DESIGN STUDIES 9. MOLTEN-SALT BREEDER REACTOR DESIGN STUDIES ........coooiomvvioreeeeeeeeeeeeee e e 174 0.1 GENEIAL ..o e e et ee e et e e s ettt et e s snaseennenas 174 9.2 Flowsheet ........cccoeeeviinnes e R s ettt 177 9.3 Reactor Cell Component Arrangement............occeerenees R s eeteererteeereetreeaeeesetreareeeearasanen 179 9.4 Component Design ................. SRRSO T veterennreernreenransaareseeansannraeeaarnreeeeernns 182 REACEOr VESSEL.. .. ittt ettt e ettt et e e e et e e e n bt e et anen et nmr e 182 "Fuel Heat EXChanger...........cccccooiviiiiiiei et sree e e erreuresratesaesmee i aaeseanreans £ sneeneeent£ens 186 Blanket Heat EXChanger ... 190 0.5 ReACtOr PhYSICS .ottt ettt et 193 9.6 MSBR Gas Handling SYStEmM .......oooueereeereoreeeeeeeeeseeer e e et ettt 199 Xenon Removal...........cccooiiiiiiii S SRS ORUORPO 199 Mechanical DeSiEn. ..ot ittt et et et e e e 200 10. wvii Gas Injector Systém ........................................................................... e et e 200 BUbble Separator SYSLEML.. .. c..ooiuieiieitiieeieneeteet et esee e ete sttt e eae e iee e e ree et e e e .. 202 Volume Holdup SYSEEM . ..oooiiiiiiiii et e e s te e e snie e e e s e 202 NONCritiCal COMPONENES ..iivviiieiiieiieieei e et e e ettt e e e et ee e e e e eaeeeaee e aeesamseeeeeseeaneeeen e sneinbeeeeean 203 MOLTEN-SALT REACTOR PROCESSING STUDIES .. ... ettt e 204 10.1 Continuous Fluorination of a Molten Salt.............. et e, 204 " NONPIOLECEA SYSEEIM ..ottt ettt et et e ee et e e e e 204 - Protected SYStOM .. ....occoiiiiiiiii et e e e e 205 10.2 Molten-Salt Distillation Studies.............c.cccceeiinne, PO OEO DR PTOORPUPUROPPSORRY. s 206 Relative Volatility Measurement.................... bt eht b e er e eae e et en e aee s .. 206 - Vaporization Rate StUAIiES .........cccciiiiiiiiiioiiii et e et e ch et etnen e ae e 206 Buildup of Nonvolatiles at a Vaporizing Surface ... 208 10.3 Vacuum Distillation Experiment with MSRE Fuel Salt................ccoovi 211 Summary PART 1. MOLTEN-SALT REACTOR EXPERIMENT | 1. MSRE Operations The maintenance and other shutdown operations started during the last period were completed, and power operation of the MSRE was resumed in October with only one of the main radiator blowers in serviée. After 17 days’ operation (run 8) at the maximum power attainable with one blbwer, the second blower was installed, and a 12-day run (run 9) ‘was made with both blowers in service. _Thé nextv run (run 10) began in December and was continued at full power for 30 days without interruption. A fourth power run (run 11) was in progress at the end of the report period with 31 days accumulated. ' Further refinements were made in the reactivity balance, and on-line calculations were used as a guide during operation. Application of these refinements showed that the unaccounted-for reactivity change was only about +0.05% 0k/k through the end of run 10 (16,450 Mwhr). Addi- tional dynamics tests were performed at the start of run 11, which showed that the dynamic char- acteristics of the system were unchanged. The performance of equipment in the priméry and auxiliary systems was ge»nerélly satisfactory. Detailed study has revealed no change with time in the heat-transfer coefficients of the main heat exchanger or the radiator. Salt plugs in the reactor off-gas line, from an accidental overfill of the fuel purfip, caused some difficulty, but the line was finally cleared before run 10, and this problem has not recurred. Partial plugging of the particle trap and charcoal beds in the off-gas system was also encountered, but this did not restrict power operation. A new particle trap, of different design, was installed, and no further plugging was noted at that location. Only routine mainte- nance problems were encountered with other equipment. 2. Component Development The sampler-enricher was used to isolate thirty-six 10-g samples and eleven 50-g samples on a routine basis. In addition, eight special samples for use in cover-gas analysis and five cap- sules for beryllium addition to the fuel were handled. In-an effort to rc?'duce the contamination ‘ llevel in the sampler and adjacent areas, the inside of the sampler was»cleaned with s'p'onges, and an additional ventilation duct was installed near the transport cask position. The design of the sample capsules was changed to provide 5;1 nickel-plated magnetic steel top instead of the orig- inal copper top, so that a magn.elt can be used for retrieval in fh“e event a capsule is dropped. A mechanical method of assuring that the maintenance valve is. closed is being substituted for the existing pneumatic system, which has given difficulty because I}of the gradual increase in leakage of buffer gas through the upper seat. The valve itself will not be replaced at this time. Several minor maintenance tasks were performed, including replacementli of the manipulator boots, inspec- tion of the vacuum pumps, énd replacement of a small hinge pin and cotter key which had worked loose inside the sampler. : | , Eight 10g coolant salt samples were isolated. The valve seats of the removal valve in the coolant sampler were replaced. | Installation of the fuel process sampler is complete except for the shielding, and the opera- tional checks have been cornpléted. _ The off-gas sampler was altered to include a hydrocar‘bon analyzer section in pléce of the chromatographic cell, which was not ready. The intemal piping was also rearranged to permit sampling upstream of the 522 line filter. | A redesigned MK II off-gas filter was in-stalled in place of the old particle trap and charcoal filter assembly. The filter was enlarged from 4 to 6 in. ID and was arranged so that the Yorkmesh entrance section can be heated above the temperature of the rest of the filter. There were other changes in the arrangement of the intemals, which were made to correct deficiencies found in the first filter. Measurements showed that the pressure drop was less than 1 in. H,O at three times _rated flow and that both filters had efficiencies greater than 99.9%. Tests of the Feltmetal sec- tions of the filter indicated that the expected life might be as little as three weeks or as great as 70 years, depending on the character of the particles in the rleactor off-gas system. The old patticle trap was taken to the hot cells for testing and examination. Tests indicated an increase in the pressure drop by a factor of 20 over the cleanw filter and that the Fiberfrax sec- tion was essentially clean. We concluded that most of the pressure drop was in the Yorkmesh entrance section, where material had filled in the space betweér:{ the wires and plugged the open- ing of the inlet pipe. Since the inlet pipe expanded longitudinally due to fission product decay heat, it is believed that operation at power caused the inlet pipe; to push into the plugging ma- terial, thereby increasing the resistance to flow in the manner of a thermal valve. Metallographic examination of the deposit area of the Yorkmesh showed heavy carburization of the wire, and there were indications that the wire had been heated to at least }200°F. The deposit itself con- tained much carbonaceous material, as well as high Ba and Sr fractions. There was very little. Be or Zr, indicating that there was essentially no salt carried to this point, and most of the fis- ‘sion products found were daughter products of Xe and Kr. The Fiberfrax section was very clean except for a small deposit of oil in the first layer. :} A model was developed which evaluates the concentration of “‘very short lived’’ noble gases in the graphite while the reactor is at power. Reasonable agreement with measured concentration distributions was obtained for 14°Ba, 1“Ce, and °'y. [\ Among the maintenance tasks completed were: (1) replacement of all of the air line quick disconnects in the reactor cell with metal compression fittings after the elastomer in the orig- inal disconnects became embrittled from radiation, (2) replacement of the particle trap in the off- gas line, (3) removal of frozen salt obstructions in several of the lines coming from the pump bowl, (4) replacement of the core sample array, and (5) replacement of a control rod and drive. 3. Pump Development The restriction in the annulus between the pump shaft and shield plug in the MK-1 prototype pump, which was discussed in the previous semiannual report, was found to have been caused by the salt level being raised accidentally into the annulus during a fill of the system. The MK-1 fuel pump tank was removed from the prototype pump facility and installed on a _test stand. This stand was constructed for room-temperature measurement of the stresses pro- duced in the weld attachment of the discharge nozzle by forces and moments imposed by the pump-tank discharge piping. During initial testing, a crack was found in the heat-affected zone of the weld attachment. The crack was repaired by welding, and further exploratory tests are being made to measure the stresses. The spare rotary element for the MSRE fuel pump was prepared for reactor service and is being held in standby. The shaft-seal oil leakage problem on the spare rotary element for the MSRE coolant pump was resolved, and the assembly is being completed for standby service. The ‘ lubrication pump endurance test was continued. Shaft defllection' and crit\ical speed tests were completed on the MK-2 fuel pump rotary element, and the MK-2 pump tank is about 70% fabricated. - 4. Instrument Design and Development _The design of instrumentation and controls for the off-gas sampler is essentially complete. Changes in the design of the sampler system requitred changes in the instrumentation and controls design. Performance of developmental instrumentation has continued to be generally satisfactory. No problems were encountered that required redesign or initiation of new development. Results of investigation indicate that none of the commercially available solid-state multi- plexers are suitable for use as a direct replacernerit for the mercury switches used in the MSRE temperature scanner. No further progress has been made in determining the cause of failure of an NaK-filled differential-pressure transmitter in MSRE service. The effectiveness of modifications to ultrasonic level probe circuitry has not been determined because no salt has been transferred to the fuel storage tank. A new conductivity level probe was developed for use in a fuel distillation system. Design of this probe is similar to that of probes used in the MSRE drain tank but differs in that it is smaller in size and will provide a usable indication of level changes. 5. MSRE Reactor Analysis I Neutron energy spectra in the MSRE were calculated and used to estimate important isotopic changes and associated long-term reactrvrty effects during power operation. The changes con- sidered include depletion of ?3*U, 235U, and ?*3y, production of ?3°U and ?3°Pu, burnout of initial °Li in the salt and %8 in the graphite, and production of tritium and '®0 as products of n,a reactions. With the exception of 23%U, the principal contributors to long-term reactivity changes were found to be the ®Li burnout and 23Pu productron Further studies were made in the correlation of the observed time behavior of the '3%Xe poi- soning in the MSRE with calculations from a theoretical model.il Graphic comparisons are given of calculated buildup and removal of '?*Xe reactivity, following changes in power level, with some of the experimental reactivity transients observed from op“eration to date. The results, in N good accord with previously reported evidence, point to the cofi‘clusion that a small amount of undissolved helium gas is in circulation with the salt, which enhances the mass transfer and . removal of xenon from the reactor. In addition, the transient arralysis supports the assumption of a fairly high efficiency of removal of xenon directly from the}gas bubbles by the external . stripping apparatus. Approximate rarrges of the circulating bubble volume fraction and bubble- stripping efficiency obtained from the analysis were 0.10 to 0.1;5 vol % and 50 to 100% respec- tively. ’ PART 2. MATERIALS STUDIES 6. Molten-Salt Reactor Program Materials ‘There was no microscopically visible corrosion or coatings ion the Hastelloy N reactor-vessel- wall surveillance specimens exposed to molten fldoride fuel in the core during a 7800-Mwhr opera- tion in which the specimens accumulated a thermal neutron dose of approximately 1.3 x 10°° nvt, ~ However, a carbide deposit about 0.001 in. thick was found on specrmens in' contact with the graphite. ' A loss in ductility of the irradiated specimens wao found at elevated temperatures, as ex- " pected. In addition, however, there was an unexpected 20% redrrction_in ductility at low tempera- ture, which is thought to be related to extensive grain-boundary lcarbide precipitation. The doses received by the metal specimens are higher than the reactor vessel is anticipated to receive over its lifetime, and the test results give reassurance that the mechanical properties of Hastelloy N are more than adequate for the service planned. | | A family of curves was obtained from tests of Hastelloy N at various strain rates and tempera- - tures. - These curves will allow one to predict the strain rate sensitivity of the ductility of Hastel- loy N at any given temperature. The strain rate sensitivity changes markedly wit}r temperature. We examined seven experimental grades of isotropic graphiteli. None satisfied all the require- ments for molten-salt breeder reactors, but one had a good pore spectrum, and four appeared to have potential for MSBR use. 1} 1] 5 Experiments have been designed and are being fabricated which will pemit irradiation of _graphite to the high exposures that will be incurred in an MSBR. Irradiations in HFIR, DFR, EBR-II, and ORR are planned. The search is continuing for corrosion-resistant alloys that are suitable for brazing graphite to Hastelloy N. A furnace for brazing large graphite-to-Hastelloy-N assemblies is being constructed. The lofig-term thermal convection loops of Hastelloy N and type 304L stainless steel have continued to circulate fused salts, acquiring 43,024 and 31,749 hr respectively. Weight losses from specimens inserted in the stainless steel loop are less than what was measured on earlier samples. 7. Chemistry ‘ o ! _ Chemical analyses of the uranium concentration in the fuel salt show a measurable decrease. This results from dilution of fuel salt by the remnants of flush salt that remain in the reactor after flushing and from the transfer of about 7 kg of uranium from the fuel to the flush salt in each drain- flush-fill sequence. _ | The chromium conéentration has remained s‘teady at about 60 ppm, indicating the absence of corrosion in the reactor fuel circuit. | | | At termination of MSRE run 7, 1.66 gram-atorris of uranium had been Burned,v and, as a conse- quence, about 1.66 equivalents of oxidizing species had been produced in the fuel. To neutralize this oxidizing effect, and to make the fuel more reducing in character, the fuel was treated with beryllium metal to reduce a small amount of UF, to UF,. To date, 27.94 g of beryllium has been introduced; this has converted 0.65% of the UF, to UF ,. Most fission products behaved as expected with the exception of rather noble metals, which continued to show an apparent tendency to volatilize and to plate on metal surfaces. Attempts to decrease the volatilization by chemical reduction of the fuel with elemental beryllium were unsuc- cessful. Detectable volatilization apparently continued for long periods after shutdown; a three- day shutdown reduced the volatiiization of molybdenum by a factor of only 5. Further studies of the soiubility of oxide in fuel—flush-salt mixtures have been carried out. A minimum solubility occurs when the mole fraction of Z1F , reaches 0.01. In connection with a study of methods of reprocessing MSBR fuel, the solubilities of SmF and NdF , in fuel solvent have been measuréd as a function of temperature. \ Salt compositions for possible use as a blanket for the MSBR have been reviewed. The feasibility of removing rare earths from fuel by precipitation on solid UF , has been studied, and the results are moderately favorable. Activify coefficients associated with the reductive extraction of rare earths from fuel into bismuth amalgams have been measured. The . ptocess appears quite attractive. The use of reducing agents for protactinium removal from blanket melts was investigated further. Electrolytic reduction gave disappointing results, but the use of thorium as a reducing agent gave good results, especially when there was a large surface area of iron metal available to receive the protactinium. | In addition to the regular salt samples, several special samples were analyzed. These in- cluded capsules used to make beryllium additions to the fuel éalt, MSRE off-gas samples, and highly purified LiF » BeF2 samples. The absolute standard délviation for oxide determined in ten radioactive fuel samples taken from the MSRE over an eight-month period was 8 ppm. A transpiration method has been developed for the determination of U3 /U ** ratios in radio- active fuel samples. The method is based on the measurement of the HF produced by the reduc- tion of oxidized species when the molten fuel is sparged with hydrogen. Increases in the udt/utt “ratio from about 0.0005 to 0.005 were observed when metallic geryllium was added to the fuel in the reactor. | An experimental reference electrode, consisting of an Ni/Ni 2* half-cell electrically connected to the fluoride melt through a wetted boron nitride ‘““‘membrane,’ exhibited satisfactory Nernstian reversibility. On the basis of limited stability tests, this elecfrode appears to be suitable as a reference for electrochemical measurements in molten fluoridesl. An anodic oxidation wave result- ing from the voltametric oxidation of U** at +1.4 v was studilzd and found to have properties most consistent with oxidation of U** to U5, followed by catalytic disproportionation of the US* The _spectrophotometr'ic determination of U3" in molten fluoride salts was investigated by a new technique in which U3* is generated voltametrically in th!e optical path of a captive-liquid cell. The feasibility of determining 50 ppm of U3" in the presénce of 2% U*" was demonstrated. Measurements of the absorption spectra of Er3+, Sm 3+, and Ho 3+ in LiF-BeF2 indicated that these ions would not interfere with the determination of U3, 8. Molten-Sali Convection Loops in The ORR Irradiation of the first molten-salt thermal convection loop experiment in the Oak Ridge Re- search Reactor was terminated August 8, 1966, because of a leak through a broken transfer line. A power density of 105 w/cm® was achieved in the fuel channeis 6f the graphite core before fail- ure of the loop. A second loop, modified to eliminate causes of failure encountered in the first, began long-term irradiation in ]-anuary 1967. An average core pnower density of 160 w per cubic centimeter of fuel salt was attained and maintained in the first ORR irradiation cycie. PART 3. BREEDER REACTOR DESIGN STUDIES 9. Molten-Salt Breeder Reactor Design Studies | Breeder reactor design studies have been concerned primarily with making a choice of the basic reactor on which design effort will be concentrated. The'modular concept has been chosen, and the power for which the module is to be used is set for the moment at 556 Mw (thermal). The . »; avérage core powet density, and therefore the flux, has been arbitrarily cut from the 80 kw/liter used in previous studies to 39 kw/liter to give greater core life expectancy. X Further optimization studies have been made on reactor parameters. A durable core config- uration has been established. The core is 10 ft high and contains 336 fuel cells. The volume is 503 ft 3, of which 16.5% is fuel salt, 6% fertile salt, and 77.5% graphite. A blanket 11/4' ft thick axially and 11/2 ft thick radially surrounds the core. A 6-in. graphite reflector is placed between the blanket region and the container vessel. The fuel cells are joined to the dished head plenum by pipe thread connections. The fuel heat exchanger and the blanket heat exchanger are flanged into place, reducing the number of pipes to be remotely cut and welded if replacement of these items is necessary. A con- centric coolant line connects the primary and blanket coolant circuits. Flowsheets and design criteria are being developed for the gas sparging system and the off-gas system. The layout of the reactor cell has been revised to eliminate some stress préblems that were found in the original layout. A first attempt at a better mounting for reactor cell components has been made and is being analyzed. Some of the more basic MSBR nuclear calculations have been started, and from the first re- sults some changes have been made in the unit cell dimensions of the core. The reactor as now contemplated has a yield of e;pproximately 6% per year, a breeding ratio of 1.07, and a fuel cycle cost of 0.43 mill/kwhr on an 80% plant factor. Work on reactot physics included (1) a series of cell calculations performed to examine the the sensitivity of the MSBR cross sections and reactivity to various changes in cell structure ° and composition and (2) several two-dimensional calculations of the entire reactor. The refer- ence cell contained ~0.2 mole % 233U in the fuel salt and 27 mole % 23271 in the fertile salt. The fuel volume fraction was 16.48% and the fertile volume fraction 5.85%. The results of the cell calculations indicated a reactivity advantage dssociated with increasing cell diameter, and a nominal diameter (flat to flat) of 5 in. was selected. Detailed radial and axial flux distributions were obtained from the two-dimensional calculations. The radial and axial peak-to-average flux ratios calculated from these distributions were 1.58 and 1.51, respectively, giving a total peak-to- average ratio of 2.39. | | The central cell of the reactor was examined for reactivity control purposes. If a completely empty graphite tube of 5 in. OD and4 in. ID is filled with fertile salt, the change in reactivity is 8k/k = —0.018%. If the empty tube is filled with graphite the reactivity change is 8k/k = +0.0012%. Thus there appears to be a substantial amount of reactivity control available by vary- ing the height of the fertile column in the tube. 10. Molten-Salt Reactor Processing Studies The concept of an integral processing plant based on a fluorination and distillation flowsheet has matured in the last year. Studies on continuous fluorination techniques have ascertained that high recoveries and good fluorine utilization are feasible, and the measurements of relative vola- tilities for the distillation stép have been highly encouraging.- Further analysis of the operations has revealed no new problems which could thwart this approach. Continuvous Fluorination of a Molten Salt. — The recovery of uranium from the fuel salt of an MSBR by continuous fluorination embraces two §ignificant problems: (1) the establishment of an adeciuate concentration gradient in the tower to effect both higfi recovery and reasonable fluorine utilization and (2) the operation of the system with a frozen layer of salt on all surfaces to pro- tect them from oxidation by fluorine. Studies with nonprotected? systems using l-in.-diam towers Etanium with fluorine utilization of 15%. Studies on column protection involve the construction of a 5-in.-diam nickel tower with have demonstrated steady-state recoveries up to 99.9% of the u provision to generate heat fluxes to create a frozen wall of salf. Molten-Salt Distillation Studies. — Relative volatilities measured at 1000°C and 0.5 mm Hg pressure for CeF,, LaF,, NdF ;, and SmF , with respect to LiF were 3 x 1073, 3x10"% 6 x 10=% and 2 x10~* respec-tively. The consistency of the results assures that these relative vol- atilities are accurate. Data have been acquired on rate of vaporization as a function of system pressure which show that the processing rates necessary in an MSBR system can be achieved in stills of reasonable size. However, analysis of the buildup of nonvolatile salts at the vapor- izing surface indicates that some method of salt circulation is mandatory. Vacuum Distillation Experiment with MSRE Fuel Salt. — An experiment is planned in which about 48 liters of MSRE fuel salt will be processed by vacuum distillation after the 235U has been removed by fluorination. The equipmen't, which has been t!:lesigned and is being fabricated, wi‘ll be used in an experimental program with nonradioactive sait to study still perfformance before it is installed at the reactor site for use with irradiated salt. / [ Introduction The objective of the Molten-Salt Reactof Program is the development of nuclear reactors which use fluid fuels that are solutions of fissile and fertile materials in suitable carrier salts. The program is an outgrowth of the effort begun 17 years ago in the ANP program to make a molten-salt reactor power plant for aircraft. A molten-salt reactor — the Aircraft Reéctor Ex- periment — 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 23“Q‘UF4 or 235UF4 dissolved in a salt of composition near 2LiF-BeF ,. The blanket would be ThF , dissolved in a carrier of similar composition. The technology being developed for the breeder is also applicable to advanced converter reactors. - Our major effort at present is being applied to the operation and testing 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 reactots and to provide several years of experience with the operation and maintenance of a small molten-salt reactor. ‘ The experiment is demon- strating on a small scale the attractive features and the technical feasibility of these systems for large civilian power reactors. The MSRE operates at 1200°F and at atmospheric presdsure and produces about 7.5 Mw of heat. Initially, the fuel contains 0.9 mole % UF ,, 5 mole % ZrF ,, 29.1 mole % BeF , and 65 mole % LiF, and the uranium is about 30% 235y. The melting point is 840°F. In later operation we expect to use highly enriched uranium in the lower concentration typical of the fuel for a breeder. The composition of the solvent can be adjusted in each case to retain about the same liquidus temperature. The fuel circulates through a reactor vessel and an external pump and heat-exchange system. All this equipment is constructed of Hastelloy N, ! a nickel-molybdenum-chromium alloy with ex- ceptional resistance to corrosion by molten fluorides and with high strength at high temperature. The reactor core contains an assembly of graphite moderator bars that are in direct contact with - the fuel. The graphite is a new materialv2 of high density and small pore size. The fuel salt does not wet the graphite and therefore does not enter the pores, even at pressures well above the operating pressure. . Heat produced in the reactor is transferred to a coolant salt in the heat exchanger, and the coolant salt is pumped through a radiator to dissipate the heat to the atmosphere. - A small facil- ity installed in the MSRE building will be used for processing the fuel by treatment with gaseous HF and F,,. Design of the MSRE was begun early in the summer of 1960. Orders for special materials were placed in the spring of 1961. Major modifications to Building 7503 at ORNL, in which the reactor is installed, were started in the fall of 1961 and were completed by January 1963. . 1Also sold commercially as Inco No. 806. 2Grade CGB, produced by Carbon Products Division of Union Carbide Corp. 9 10 Fabrication of the reactor equipment was begun early in .1962. Some difficulties were experi- enced in obtaining materials and in making and installing the equ.ip.ment, but the essential instal- lations were completed so that prenuclear testing could begin in August of 1964. The prenuclear testing was completed with only minor difficulties in March of 1965. Some modifications were made before beginning the critical experiments in May, and the reactor was first critical on ]une_ 1, 1965. The zero-power experiments were completed early in July. Additional modifications, main- tenance, and sealing of the containment were required before the reactor began to o'pegaté at ap- preciable.power. This work was completed in December. Operation at a power of 1 Mw was begun in January 1966. At that power level, trouble was experienced with plugging of small ports in the control valves in the off-gas system by heavy liquid and varnish-like organic materials. These materials are believed to be produced from a very small amount of oil that leaks through a gasketed seal and into the salt in the tank of the fuel circulating pump. The oil vaporizes and accompanies the gaseous fission products and helium cover gas purge into the off-gas system. There the intense beta radiation from the krypton and xenon polymerizes some of the hydrocarbons, and the products plug small openings. This dif- ficulty was largely overcome by installing a specially designed filter in the off-gas line. Full power — about.7.5 Mw under design conditions — was reached in May. The power is lim- ited by the heat-removal capability of the salt-to-air radiator heat-dump system. The plant was operated until the middle of July to the equivalent of about one month at full power when one of the radiator-cooling blowers — which were left over from the ANP program — broke up from me- chanical stress. While new blowers were being procured, an array of graphite and metal surveil- ‘lance specimens was taken from the core and examined. o Power operation was resumed in October with one blower; then in November the second blower was installed, and full pbwer was again attained. After a shutddwn to remove salt that had acci- dentally gotten into an off-gas.line, the MSRE was operated in D%—:-cember and January at full power for 30 days without interruption. A fourth power run was begun l‘iate in January and was still in ' i i In most respects the reactor has performed very well: the fujel has been completely stable, progress after 31 days at the end of this report period. the fuel and coolant salts have not corroded the Hastelloy N cont:ainer material, and there has been no detectable reaction between the fuel salt and the graphit!e in the core of the reactor. Me- chanical difficulties with equipment have been largely confined tfo peripheral systems and auxil- iaries. Except for the small leakage of oil into the pump bowl, tl:le salt pumps have run flawlessly for over 10,000 hr. | \ Because the MSRE is of a new and advanced type, substantial research and development ef- fort is provided in support of the operation. . Included are engineering development and testing of réactor components and systems, metallurgical development of materials, and studies of the chem- istry of the salts and their compatibility with graphite and metals both in-pile and out-of-pile. Conceptual design studies and evaluations are béing made of large power breeder reactors that use the molten-salt technology. ‘Some research and development is being directed specif- ically to the requirements of two-region breeders, irfcluding work on materials, on the chemistry of fuel and blanket salts, and on processing methods. i» 1Y Part 1. Molten-Salt Reactor Experiment 1. MSRE Operations P. N. Haubenreich 1.1 CHRONOLOGICAL ACCOUNT OF OPERATIONS AND MAINTENANCE . Guymon H. C. Roller R. H J. L. Crowley R. C. Steffy T. L. Hudson V. D. Holt P. H. Harley A. 1. Krakoviak H. R. Payne B. H. Webster -R. Blumberg C. K. McGlothlan The reactor shutdown that started in July! continued through September. The first of the two specially construc_:téd replacement blowers was delivered on September 28, ten weeks after the reactor was shut down. Meanwhile, the time was fully occupied with a host of other jobs that were completed about the time the replacement blower was received. These included removing and re- placing core samples, work on control-rod drives, replacement of the special fuel off-gas filter, modification of the radiator door seals, and repairs and modifications in the cooling-water system. The first step in the reactor startup in run 8 was seven days of flush-salt circulation (see Fig. 1.1). During this time the salt that had frozen in the sampler line at the pump bowl was thawed. After the temporary heaters for this job were removed, the reactor cell was sealed, and the leakage ~ was shown to be acceptable by a test at 10 psig. By this time the first blower was ready to run, but delivery of the second replacement unit was not expected for several weeks. Therefore, nuclear operation was resumed early in October with only one blower. The reactor was operated for 17 days at the maximum power attainable with one blower: 5.8 Mw. During this time the pressure drop across the new off-gas particle trap increased to several psi.. The inlets to the main charcoal beds also became restricted and had to be relieved by back- blowing with helium. Two days after the start of power operation, the fuel off-gas line became plugged near the pump bowl, causing the off-gas to be diverted through the overflow tank. This 1MSR Progr&frl Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 9. 11 12 ! . ORNL-DWG 66-12792R INVESTIGATE GO TO OPERATE REMOVE CORE SPECIMENS REMOVE OFF-GAS PLUGGING FULL POWER AT'POWER REPLACE MAIN BLOWER SALT PLUG. CHANGE REPAIR THAW FROZEN LINES ’ CHECK FILTERS AND VALVES SAMPLER TEST CONTAINMENT CONTAINMENT LOW-P — DYNAMICS CHECK \ OPERATE TESTS CONTAINMENT : AT POWER 8 36 2 NUCLEAR POWER Tl 1 ; 4 o) o 2 ! 0 III | 3 [ ] T F SALT CIRCULATING | FYEL @ (o] ! !nlm:ml[ ¢ e —n J F M A M J J A Fig. 1.1. MSRE Activities in 1966. ’ complicated the routine recovery of salt that gradually 'accumulétes in the tank, and during some recovery operations, activity was forced into the line that drains oil leakage away from the pump. The reactor was shut down to install the second blower unita, which had just been delivered, and to attempt to relieve the plugged line at the purfip bowl. The system was flushed, the reactor cell was opened, and heat was applied to the line. While the line was hot, pressure was applied and the line opened up. Tests showed that, within the accuracy of the available instrumentation, the pressure drop was then normal. The heaters were removed, ‘%the cell was sealed, and nuclear operation was resumed seven days after the fuel was drained for the shutdown. Also during this time, a gas flow element in the vent from the oil catch tank was'li removed and an alternative flow measurement provided. The flow element had become plugged \x;hen, as the last step in the fuel drain, the overflow tank was emptied and gas from the pump bowl again vented out through the oil drain line. | When the power was raised on November 7 for run 9, the temi)erature of the off-gas line showed that it was already plugged and that the gas was again bypassing through the overflow tank. To prevent the transfer of activity into the oil drain line when salt was recovered from the overflow tank, the nuclear power was reduced several hours before eaéh t:fransfer. Limitations on the amount of salt that can be tolerated in the overflow tank and the heel th“at remains after a recovery re- quired that the power be reduced for a transfer every two days after the initial heel had accumu- lated. Nuclear operation was continued in this fashion for 12 days while heaters, tools, and procedures for more positively clearing the off-gas line were devised. Then the reactor was shut down to work on the line and also to check what appeared to be a high inleakage of air into the reactor cell. 13 Search for the restriction in the off-gas line revealed thin plugs in the flanges at both ends of a removable section near the pump bowl. The plugs were easily poked out, and the line was then shown to be clear by viewing, probing, and pressure-drop measurements. The plugs were attributed to flush salt almost completely blocking the line as a result of the overfill in July. An incon- sequential amount of this salt was also seen in the 4-in. holdup pipe. The high inleakage into the reactor cell proved to be from valve-operator pneumatic lines. Since these lines are protected by automatic block valves, the leaks did not violate containment. Therefore,‘flowmeters were installed so their input could be taken into account in the routine monitoring of cell leak rate during operation. During this three-week shutdown, we also made some repairs and modifications to the cofilpo— nent cooling blowers and removed and repaired an air valve in the reactor cell. "~ Run 10 began on December 14 and continued for 30 days at full power. During the first two weeks of the run, the pressure drop across the particle trap in the fuel off-gas line repeatedly built up and had to be relieved by forward- or back-blowing with helium. Examination of th_e first partiéle trap, removed in August, had shown that heating the central inlet tube would tend to jam it into the first-stage filtering medium (see p. 47). To test the effect of reducing the heating, for the last two weeks of run 10, the off-gas was delayed on its way to the particle trap by routing it through the two empty drain tanks. When this route was followed, the particle trap pressure drop came down and stayed down. ' The inlets to the main charcoal beds had to be back-blown during the first week of run 10 but not afterward. . _ During this run the UF ; concentration in the fuel salt was increased by the addition of 16 g of beryllium metal through the sampler-enricher. One purpose of inbreasing the reducing power of the salt was to investigate the effects on volatile fission product compounds (see p. 123). Another was to alleviate concern over possible corrosion. Practically no corrosion had been seen (<0.1 mil of generalized corrosion in 20 months), but the absence of cotrosion depends on maintaining a reducing environment, and a larger margin was desired. Operation at full power was to be interrupted after 30 days to permit inspection of the new blower hubs and blades, which had by then been run over 1000 hr. But toward the end of the run two conditions developed which caused us to drain the reactor and extend the shutdown. The heat exchanger between the treated-water and tower-water systems began to leak at an increasing rate, and the leakage from the air lines in the reactor cell became so large that the measurement errors clouded the determination of the cell leak rate. When the reactor was shut down, inspection showed that the blowers were in excellent condi- tion. The leaks in the air lines were traced to deterioration of neoprene seals in some quick- disconnects. All the disconnects in the reactor cell were replaced with metal-compression fit- tings, and the leakage was stopped. The heat exchanger leaking water was replaced. The filter assembly and the pressure control valve in the fuel off-gas line were removed, and a new filter assembly was installed. This consisted of two filters in parallel, each with much larger frontal ‘ area than the old filter (see p. 42). 14 Nuclear operation was resumed in run 11 on January 28 anqi continued without interruption (except for 2 hr to investigate a false temperature alarm) through the end of the report period, February 28. No difficulties of any consequence were encount‘_lered, and the program of adjusting the fuel UF , concentration and observing the effect on volatile fission products continued. Details of operations and maintenance during this report per1od are given in the sections which follow. Although the emphas1s tends to be on the troubles the reactor was in operation most of the time, and the operatmn was in most respects quite/satisfactory. Table 1.1 summarizes some of the history. Salt was circulated in the fuel and coolar%.t loops for 60 and 82%, respectively, of the time in this report period. The reactor was critical 53% of the time. 3 Table 1.1, Summary of Some MSRE Operating Data Aug. 31, 1966 Feb. 28, 1967 . Increase Time critical, hr : 1775 4092 ' 2317 Integrated power, Mwhr 7823 _' 21,514 13,691 Salt circulation, hr ) ‘ Fuel system 4691 7337 2646 Coolant system 5360 1 8946 3586 1.2 REACTIVITY BALANCE J. R.'Engel The purpose of the reactivity-balance calculation dufing po%ner operation of the reactor is to provide current information about the nuclear condition of the sj}stem. During this report period, improvements were made in the calculations, making it possible to detect vety small anomalies in reactivity behavior; none was observed. Calculations made éluring previous ’periodsz of opera- tion did not include the !33Xe poisoning term because the mathematical model, with the coef- ficients then available, did not adequately reflect the xenon behavior in the reactor. When power operation was resumed in October 1966 (run 8), we included a calculation‘of the xenon effect to provide complete reactivity balances. Subsequently, the overall calculation was improved by modifying some xenon stripping parameters to improve the descr”1pt1on of the xenon transients, and by including long-term isotopic change effects that had beep previously neglected. Experience Figure 1.2 shows the results of some .of the on-line reactivify balances calculated in runs 8, 9, and 10. During runs 8 and 9 the parameters used to calculate the xenon poisoning were 2Ibid., pp. 10—13. 15 ORNL—DWG 66—41944R +o4 REACTIVIT o ] Y j 4008 e A "&L* sy { | 005 BT — 4 i z 0 3 ‘M\h ,4 v, \"“'J"E‘ o % & sl ..-“ii . Lg. 4 3 > ] : D i PRIV R @ _ 4 . o ¢ = o0 : s » o0 REACTIVITY 04 s -0 | | | -045 : *L. -0.45 = : 8 POWER : 8 ; i | [ ] ] z 6 gl 6 —H—t I a— x 4 c < 3 =% 4 — 2 i 2 PO‘|NE_R o . - 0 8 10 12 14 16 18 20 22 24 26 28 30 7 9 1 13 15 17 19 ) OCTOBER, 1966 . . , NOVEMBER, 1966 +040 | l ‘ ‘ T T 7 (—— | +0.05 |—= 5 REACTIVITY L. w __| 41* l_ s, ey 0 WBhe $upa, P ! PRI L m;.r A i 3 L’ O AR T NPT & Z_ 005 ‘“‘l Ford !‘7* —'|F' il : : ~ g ) i . o F 2 -0.10 + -045 —t - COMPUTER OUT * i -020—} OF SERVICE : ot I I B : . I E | { g ] x4 POWER B = 2 0 : ‘ 14 18 18 20 22 24 26 28 30 1 3 5 7 9 T 13 15 DECEMBER, 1966 JANUARY, 1967 Fig. 1.2. Residual Reactivity During Power Operation in Runs 8, 9, and 10. found to describe the steady;state condition reasonably well, but the transients were not well described. This difference between the calculated and actual xenon transients produced the cyclic variation in the residual reactivity in November when the nuclear power was cycled between 0 and 7.2 Mw. Detailed analysis of the observed xenon transients led to changes in some param- eters which produced much better agreement between the calculated and observed transients in run 10. However, there is still a small difference (~0.03% 8k/k) between the calculated and observed steady-state poisoning at 7.2 Mw. This causes the apparent change in residual reactivity between zero power and the condition with steady-state xenon poisoning. The larger negative reactivity transients in Fig. 1.2 are all due to off-normal operating condi- tions not accounted for in the reactivity balance. In all except one of these transients, the nega- tive reactivity resulted from an increase in the circulating void fraction while the salt level-in the fuel-pump tank was abnormally low. The low level, in turn, was caused by low system temperatures which followed instrument-initiated power reductions. In all these cases the excess voids were stripped out and the reactivity recovered when the normal pump level was restored by increasing the system temperature. The negative reactivity excursion on October 23, 1966, was caused by an abnormally high salt level in the pump tank which reduced the efficiency of xenon removal by the pump spray ring. The negative reactivity was produced by the transport of additional 135Xe into the graphite moderator. Since the removal of xenon from the graphite is a slower process than bubble stripping, the recovery in this transient was slower than in the other cases. 16 Operation at power was resumed on January 28, 1966, and has continued without major inter- ruption through the end of this report period. After the initial blilildl.lp of 135Xe poisoning, the residual reactivity has been between 0 and + 0.04% &k/k except.fo'r one incident in which excess circulating voids were introduced by abnormal operation (low salt level in the pump tank). The long-term drift in the reactivity balance is summarized in Fig. 1.3, which shows the residual reactivity as a function of integrated power from the start of power operation through run 10 (ending January 15, 1967). The reference condition for this figure is the system condition at the start of run 4 (December 20, 1965). Because of the remaining small uncertainty in the !35Xe term, the represeritative results shown here are taken at zero polwer with no xenon present. There appears to have been a positive shift of about 0.05% 8k/k during the first 1000 Mwhr of operation that has remained relatively constant since that time. No speci‘fic cause has yet been established for this shift. However, the change is nearly as small as the efitimated confidence limits of the calculation (£0.04% &k/k) and is much smaller than the operating limit on the reactivity anomaly, which is %0.5% 8k/k. . Previous reports of the reactivity behavior? suggested the possibility of a significantly larger positive reactivity anomaly. However, those tentative conclu»sio‘ns were based on data from early results which have since been corrected. The earlier balances did not include the reactivity effects of isotopic changes (other than ?35U) or of flush-salt dilutions. Correction for these effects made a net reduction in the magnitude of the apparent an‘omaly. (The calculation of the - isotopic-change effects is described on p. 83 of this report.) In .addition, preliminary analysis of some pressure-release experiments indicated a circulating void f!r_action of 1 to 2% by vélume in the fuel salt.? If such a void fraction had been present at steady state, the negative effect of the 31bid., pp. 22-24. ORNL-DWG 67-1076A 0.08 ® [ ] = . \ . ® 004 s 3 PR * (3 ° - [ ] E 3 S : 5 0 " < wi ax .} | S—— . w . 400 : : 0.04 2 5 erezre 22 =z ’ 7727077777, Z7772 - s o ] L re % W 300 [ S 003 2 (@] ) . HTI = HEAT BALANCE POWER . g 100 REACTOR CUTLET TEMPERATURE — RADIATOR OUTLET TEMPERATURE I.:.;:J 0.04 APR MAY JUN JUL AUG SEP ocT NOV DEC JAN FEB MAR 1966 1967 Fig. 1.5. Observed Performance of MSRE Main Heat Exchanger. open by changing the bypass-damper position. A direct measure of the radiator-air outlet tem- perature can be made only when the bypass is fully closed, and even then the air from the an- nulus blowers would cause an error in the measured temperature rise across the radiator. The outlet air temperatures for the other conditions were calculated from air and salt heat balances around the radiator. Corrections were applied for the effects of the annulus blowers. However, the measured temperature rise of the air across the radiator tubes was about 17% higher than the calculated rise in the two cases where the bypass damper was‘ closed and a measurement was possible. This discrepancy indicates that either the air-flow Jneasurernent, the temperature meas- urement, or the reactor heat balance is incorrect. The heat balance and the temperature measure- ments were assumed to be correct, and the calculated outlet air temperatures for the other radiator conditions were ;:orrected to be consistent with the values measured when the bypass damper was closed. The coefficients for the two conditions when the bypass damper was closed were 28.5 and 38.5 Btu hr—! ft—2 (°F)~!, These are the same values that w“ere calculated from June 1966 data for similar operating conditions. ‘ The apparent error mentioned.above is about the same as the 15 to 16% discrepancy between ‘the air heat balance and the salt heat balance reported previously!® for similar conditions. This 19pid., pp. 37—38. 23 seems to indicate that the discrepancies are the result of some specific error in measurement rather than random variations in the data. The radiator air instrumentation was not intended to be accuratej enough to permit a precise heat balance on the air. The stack is not;sufficiently long to ensure either a uniform velocity distribution or a uniform temperature distribution across the stack cross sectién. Since the velocity and tempetrature are measured only at a single point in the stack, any flow ot temperature asymmetry could cause a significant error. There is also the pos- sibility that the initial calibration of the air flow instrument was incorrect. Methods of Improving the Heat Transfer. — A stu?i\y was completed to determine whether the maximum power capability of the reactor could be raised by some convenient method. For normal operating conditions, an upper limit of 1210°F has been placed on the reactor outlet temperature. This temperature was selected on the basis of thermal stress cycling and © ORNL-DWG 67-4763 100 /3 . a s —_ i ' / s & . | - ! 7] A o | R 3 / / (] E:E 60 i — ‘ : &o . Eao A 25 | |/ / &340 / : & / A a 2 | / [ ] (] 20 / / 2 ' TWO MAIN = ‘ l BLOWERS H RUNNING 77 0] & - A » g / © 10 { 2 5 iC ~ 20 AR PRESSURE DROP (in. HZO) . . i I Fig. 1.7. Effect of Air Pressure Drop on Radiator . Heat Transfer Coefficient, | stress rupture life of the reéctor system. The minimum coolant éalt temperature in normal opera- tion has been se‘t at 1000°F, in order to reduce the probability of freezing the radiator in case salt flow is interrupted. These two temperature limitations and the heat transfer capability of the main heat exchanger limit the reactor power to about 7.4 MW.!i In the summer months this power limitation about coincides with the capacity of the coolant radiator. However, during colder weather the radiator heat transfer capacity increases, and a portion of the cooling air must be bypassed to avoid overcooling the coolant salt. \_ The reactor power can be increased only by improving the hqiat transfer performance of both the radiator and the main heat exchanger. Other than replacement with a larger unit, the heat- exchanger capability can be increased only by increasing the flow rates of the fuel and coolant systems. Increasing the temperature difference between the fuel and coolant system is undesirable because of adverse effects on the thermal-cycle and stress-rupture life of the reactor system. that could be obtained and to A study was completed to determine the maximum flow rates determine the increase in heat transfer performance that would result from these flow increases. The pumps are capable of accepting larger-diameter impellers and of operating at a higher horse- power rating so that increased flow is possible. The flow can also be increased by using a ,higher-freqfiency power supply to increase the rotational Speed.olf the pumps. Slightly higher flow rates can be obtained with the higher rotating speed, because the horsepower rating of the drive motors can be increased at the higher speed. Calculations indicate that the maximum possible flow rates for the fuel and coolant systemé would be 1530 and 1644 gpm respectively, assuming that the pumps themselves are the limiting factor. Using these maximum flow rates and operating between the temperature limits that now permit 7.2 Mw, the heat fransfer capability of the heat exchanger would be increased to 8.1 Mw. Even this modest increase might not be practical, how- 25 ever, because of undesirable effect_s of increasing the flow. For example, development tests suggest that increasing the flow causes more gas bubbles to be introduced into the circulating salt by the stripper jets in the pump bowl. This would be undesirable, because it would introduce more uncertainty into the reactivity balance, In the radiator, increasing the salt flow has very little effect, since over 95% of the heat transfer resistance is on the air side of the tubes. There is no way to improve the radiator per- formance without major expense. Additional surface area could be provided by adding more tubes or by adding some type of fins to the tubes, but either would be difficult and time consuming. Additional air capacity could be provided, but this too would be a major undertaking. The radiator air flow would have to be incrgased by a factor of about 1.8 to remove 10 Mw. This would in- crease the air pressure drop to ~35 in. H,O and the power requirements' of the blowers to about 2900 hp, as compared with 500 hp for the present blowers. Neither the present blowers nor-the building electrical system is capable of meeting these requirements. Oné additional blower could possibly increase the radiator heat removal to the 8.1-Mw level, which would be consistent with the maximum possible power of the heat exchanger, However, the blowers would be operating very close to their surge limit, new drive motors might be required to avoid an overload condition, and the existing building electrical system would be unable to supply the third blower. In conclusion, the difficulties.in raising the power capability of the reactor far outweigh any advantages that could be gained from the relatively small power increase that can be reasonably achieved. Since the objectives of the MSRE can be met with the present heat removal system, no attempts to increase the power capability are planned. ‘ " Main Blowers C. H. Gabbard. The main blowers, MB-1 and MB-3, which had failed at the end of run 7,!! were rebuilt by the manufacturer and returned to normal service. By the end of the report period, MB-1 and MB-3 had operated without difficulty for 2100 and 1675 hr, respectively, since they were repaired. Main blower 1 has been inspected twice since it was repaired, and main blower 3 was inspected once, the first inspection of MB-1 coming after 350 hr of operation -while MB-3 was being installed. Bot-h blowers were inspected at the end of run 10 — after 1350 and 930 hr of operation. The inspections included dye-penetrant inspections of the blading and hubs, a visual inspection of the coupling, an alignment check, and a retorquing of all the bolts. With the exception of a few bolts that were retightened, both blowers were completely satisfactory. We attempted to find the cause of the failures at the end of run 7, but the specific cause is uncertain. The best explanation appears to be that one of the blades failed first along one of several existing cracks and that the resulting unbalance, or impact with the blade fragments, rbid., p. 40. 26 caused the remaining damage. The origin of the existing cracks is also unknown, but they probably occurred either during fabrication or during the overspéed_ test. The rebuilt units have reinforced hubs and have magnesium alloy blades that are 35% lighter than the original aluminum ones. The new units were also giveP a 30% overspeed test, with a dye-penetrant inspection of the hubs and blading before and after the test. The manufacturer had difficulty in casting the magnesium alloy blades free of heavy 'surface porosity and cracks. Rotor assemblies were rejected on three occasions because of ¢racking in the blades after the overspeed test. In each of these cases, cracking had been preshent prior to the test, but it had been removed by surface grinding. The units that were accepted, including a spare unit, were free of objectional defects. “ The rebuilt rotor assemblies were installed in MB-1 and MB:3 under the supervision of the manufacturer’s service engineer. The rotor and drive-motor sha%ts were carefully aligned, and the rotors were dynamically balanced in place. Instrumentationiiwa's provided to monitor the vibrations and the bearing temperatures of the two blowers whil(“-:‘ they are in operation. Vibra- tion normally runs below 1 mil, although greater vibrations deve'iope(‘i on two occasions, once when ice built up on the blades of MB-1 and once when its bearings became very cold (below 15°F). Radiator Enclosure M. Richardson | ' ‘i The coolant-salt radiator operated continuously with salt cifculating at temperatures between 1000 and 1200°F for the last five months of the report period. The modifications of the door seals'? proved effective in reducing air leakage and consequent heat losses to a satisfactory minimum. Although close examina%ion was not possible after the radiator went into operation, external examination on January 171; showed little deterioration in more than 3400 hr at high temperature. The inlet door was in ex!cellent condition, with good con- tact at top and bottom between the linked hard seals and the sofjjt seal on the face of the enclosure. On the outlet door the hard seals appeared to be in good condition. There was a tighi seal across the top but a gap of 1/8 to 3/8 in. across the bottom of the door. Ah 1-ft section of the soft seal that had become detached was found in the outlet duct. | Heat leakage through penetrations in the top and sides was no problem. The thermocouples and power wiring showed no evidence of overheating, v As a result of the blower failure in July 1966, some antimissile proteétion was provided ‘steel aircraft cables, shock between the blowers and the radiator. A grid of 1/4-in. stainless mounted, on 2-in. centers was installed just downstream of the blowers. Heavy wire screens with 0.4-in. mesh were installed just ahead of the radiator, 121bid., pp. 67—70. 27 The performance of the magnetic brakes on the door-lifting mechanism became marginal toward the end of run 10, and some slipping occurred when door positions were adjusted, Preparations were made for replacement of the worn and broken brake shoes at the next shutdown or if the brakes become inoperable. Of#f-Gas Systems P. N. Haubenreich The fuel off-gas system continued to présent some problems throughout the period. One prob- lem area was associated with the overfill of the fuel pump that got flush salt in some gas lines near the pump. The other was a continuation of the difficulty that appeared when the reactor first operated ‘at power: the accumulation of polymerized cfil residues in small passages. The former interfered with operations and required considerable work in the réaétor cell to remedy. The latter was a nuisance through run 10, but after the installation of a new particle trap caused no more‘ | trouble. | Plugging near Fuel Pump. _ After the accidental overfill of the fuel pump, ! 3 the bubbler reference line was cleared by remotely applied, external heaters. Salt in the sampler line was melted the same way, but ran down and froze at the junction with the pump bowl. The obstruction that formed here cleared out when the pump was heated up with salt in the bowl. No further trouble was encountered with these lines. The off-gas line, shown in Fig. 1.8, was a different matter. 131bid., pp. 24~25. ORNL-DWG 67 — 4764 ——{— . E— e m— } 41_5 JUMPER OFFGAS HOLDUP P VOLUME , 4in. DIAM ;| TO PARTICLE TRAP ‘ J AND CHARCOAL BEDS —~—}§ j PUMP BOWL—»—| |- OVERFLOW PIPE —=] OVERFLOW TANK Fig. 1.8. Off-Gas Piping Near Fuel Pump and Overflow Tank. 28 Although some salt entered the off-gas line, as indicated by a I‘temporary rise in temperature at TE-522-2, pressure drop measurements showed no significant difference from the clean condition. This was attributed to the blast of compressed helium from the drain tank that was released back- ward through the off-gas line, before the salt had time to freez? completely, when the overfill triggered an automatic drain. Therefore the only action taken before the startup in September was to replace the short, flexible ¢ ‘jumper’’ section of the line, where the convolutions would certainly hold somé salt. The off-gas flowed freely, with no unusual pre‘s.sure drop through 26 days of circulating helium, flush salt, and fuel salt at low power. Then, two days after poflwer operation was resumed at 5.8 Mw, a plug developed in line 522 somewhere between the pump bowl and the junction of the over- flow tank vefit with the 4-in. holdup line. The first indication was a decrease over a few hours from 225°F to 160°F at TE-522-2, as the plug caused the off-ga:ls to bypass through the holdup tank. The presence of the plug was confirmed when HCV-523 was closed to build up pressure to return salt from the overflow tank to the pump bowl: pressure in the pump bowl also built up. Efforts to remove the restriction by applying a 10-psi differential either forward or backward were unsuccessful. 4 | The bypassing of the off-gas through the overflow tank did not hi_nder operation, except for one specific job: recovery of salt from the overflow tank. Salt slowly but continuously accumulates in the tank during operation, and it is therefore essential to return salt to the pump bowl two or three times a week in order to maintain proper levels. With the plug in the line from the pump bowl, it was necessary to greatly reduce helium flows into the pump so that the overflow tank pressure could be increased faster than that in the pump bowl to make the salt transfer. Through the remainder of run 8, salt was returned from the overflow tank six times, and on at least four of these occasions some fission product activity was blown or diffused up the pump shaft annulus into the oil collection space. This was a consequence of the reduced helium purge down the shaft annulus and the unavoidable, sudden pressfirization of the pum;;- bowl that occurred at times in the procedure. The charcoal trap in the vent from the oil catch tank prevented any serious release of activity to the stack. The activity level in the oil tank and the‘; line increased, however, and the last two releases from the pump bowl caused the flow element in the vent to plug partially and then completely. | After run 8 was terminated, stéps were taken to clear the plug from the off-gas line so that the normal salt recovery procedure could be used. Frozen salt was suspected of causing the plug, so the fuel loop was flushed to reduce radiation levels, the reactor cell was opened, and specially built electrical heaters were applied to the line between the purwnp' bowl and the first flange. Heating alone did not clear the plug,‘ but when, with the line hot, 10 psi was applied backward _ across the plug, it blew through. The pressure drop came downjas more helium was blown through until it became indistinguishable from the normal drop in a clea;l pipe. The temporary heating apparatus was then removed. , 29 - In run 9 the power was raised only 8 hr after fuel circulation had commenced, but TE-522-2 came up to only 150°F, indicating that the line was already plugged again. Plans were immedi- ately set in motion to do a more thorough job of clearing. While tools and procedures were being devised, the reactor was kept in operation, but great care was taken to avoid getting fission products or salt spray up the pump shaft annulus again. This entailed lowering the power to 10 kw, 24 hr before the overflow tank was to be emptied, then stopping the pump 4 hr beforehand to let the salt mist settle. Then the fuel p:lmp was vented through the sampler and the auxiliary charcoal bed during the transfer. After three cycles of this, the reactor was drained and flushed again in preparation for working on the off-gas line. This time heat was applied to the short section of line between the second flange and the top of the 4-in. decay pipe. When heating to about 1100°F did not open the line, the flexible jumper was disconnected to permit clearing the obstruction mechanically. In the flange above the 4-in. line, the 1/z-irl. bore was completely blocked, but the weight of a chisel tool broke through what appeared to be only a thin crust of salt. Borescope inspection showed that the rest of the vertical line was practically clean, and there was only a thin layer of salt in the bottom of the horizontal run of the 4-in. pipe. Helium was blown through the line at five times the normal flow, and the pressure drop indicated no restriction, Turming then toward the pump bowl, we saw a similar plug in that flange. This too was thin and wlas easily broken out. A 1/4-ir1. flexible tool was then inserted all the way into the pump bowl to prove that a good-sized passage existed. A new jumper line was installed and operation was resumed. No further difficulty was encountered with this section of the off-gas system. ’_ ‘ Because obtaining samples remotely without spreading contamination would have been most difficult, no analyses were made of the material in the flanges. But it appeared that salt had frozen in the line, almost completely blocking it, during the overfill. Material in the off-gas stream during operation then plugged the small passages. The heaters melted the salt out of the pipe, but the flangés were not as hot, and a thin bridge remained. Particle Trap. — The first particle trap was designed on the basis of a rather brief period of investigation and development after plugging in the off-gas system halted the planned approach to full power. It served its purpose in that it protected the pressure control valvg and to some extent "the main charcoal beds, permitting the experimental program to proceed. When the pressure drop across the trap began to build up, an identical replacement was prepared, so that the first could be removed and examined as an aid in designing a better trap while operation continued. This re- placement was deferred until after rufi 7, because the pressure drop across the trap never became prohibitive, | | . The examination of the first particle trap and the design of the new model are described on p. 47. The second unit served through runs 8, 9, and 10. This unit behaved in runs 8 and 9 much as had the first particle trap. The pressure drop occasionally built up to 5 to 10 psi, beginning two days after power operation startéd in run 8, but backblowing with helium was effective in reducing the pressure to 2 to 4 psi. ‘In run 10, however, after the first week of power operation, backblowing ~ 30 effected only temporary relief at best. Various tactics were used to get gas to the particle trap with as little delay as possible, to see if increasing the fissi?n product heating would drive off organic material from the place where it was causing plugging% After these efforts proved inef- fectual, the opposite approach was used: the off-gas was delayed as long as possible. This was done in recognition that heating caused the central inlet tube to expand farther into the Yorkmesh filtering material. The delay was obtained by routing the gas through an equalizing line to the empty drain tank, through the tank and the salt fill lines to the other tank where it bubbled up through several inches of salt heel, then out through the drain-tank vent line to the particle trap. ‘This gave the gas about 8 hr of delay and also bubbled it throtilgh salt in the freeze valves and drain tank before it reached the particle trap. The pressure dr-fop across the trap was 16 psi when the gas was rerouted, but within a few hours it was below 2 ps:i. For the last two weeks of run 10, the pressure drop across the particle trap remained below 2 psi when the delayed route was followed, but began to build up almost immediately when the original route was used.. The charcoal filter, just after the particle trap, held up fission products, as shown by the heating during power operation, but the temperature profile showed no significant change to indi- cate poisoning by organics. The pressure control valve was operated full open, with the pump- bowl pressure determined by the drop in other parts of the sys{jern. Therefore it appeared that nothing would be lost by removing these items to make room fc;r new, dual-particle traps. The new (third) particle trap had been in service for 5100 Mwhr of reactor operation by the end of the report period. The measured pressure drop showed no increase at all, remaining below 0.1 psi. At 7.4 Mw the temperatufe measured near the Yorkmesh was around 275°F, the thermocouple nearest the Feltmetal (but shielded by a pipe and a bellows) average 108°F, and the lowest couple indicated 61°F, only 10°F above the temperature of the water §urrounding the trap (see p: 42 for a description of the new traps). | . Main Charcoal Beds. — The main charcoal beds continued to perform their function of holding up the fission products: there was no breakthrough of activity other than the normal 10-year 85Kr. Some difficulty was encountered at times, however, when the pressure drop at the inlet built up to . . an inconvenient level. . ' Bed sections 1A and 1B had been used almost exclusively during earlier runs, and they were on line when run 8 started. As had happened before, the pressfure drop at the inlet of these sec- tions began to increase a few hours after the power was raised!. When the pressure drop reached 7 psi, sections 2A and 2B were also brought on line. The preésure drop through the four sections in parallel remained below 1 psi through the end of the l;un, but the pressure drop through 1A and 1B remained high, and backblowing did not bring their pressure drop below 4 psi. A situation similar to that in 1A and 1B also existed in the auxiliary charcoal bed, where the pressure drop was abnormally high and did not respond to backblowing. Tests showed that the restriction was near the inlet end, and it was strongly susp‘ect(;;ed that it was due to organic mate- rial clogging the steel wool at the opening of the inlet pipe into the bed. Therefore, a remotely placeable assembly of electric heaters was designed for this section, and in the brief shutdown 31 betweenA‘runs 8 and 9 it was tried. Some improvement was observed when the heater temperatures reached 720°F, and after these temperatures were raised to 1235°F and then cooled, the -pressure drop was down by a factor of 5, to a satisfactory level. In light of this success an attempt was made to clear up the restriction in section 1B by heat- ing the inlet with a torch. (Electric heaters to fit this bed were not then available.) Although the temperature in the bed reached 870°F, there was ho improvement in pressure drop. In run 9 all four sections of the main bed were operated in parallel, with no further effort to clear 1A and 1B and no indication of any change in pressure drop. After this run, electric heaters were .installed on the inlets of 1A>ar1d 1B. Heating to 750°F for 8 hr brought the pressure drop back to the normal range for clean beds. Although the heaters could not be used in normal operation because the cooling-water level is above the inlet sections, they were left in place with leads brought out through the shield. | Run 10 began with 1A ;and 1B on line, but after one day of power operation the pressure drop began to build up, and the flow was switched to 2A and 2B. Three days later, for the first time, the pressure d‘rop across these beds began to increase, reaching 5 psi before they were backblown to reduce the drop below 2 psig. The flow was returned to 1A and 1B, after their pressure drop had been reduced below 3 psi by backblowing.i For the last three weeks of run 10, the pressure drop did not again build up. In run 11, through February, only 1A and 1B were used. Thel pressure drop remained about 2.5 psi for two weeks and then began slowly to increase. After four weeks of operation it had reached 5 psi. | 4 and the re- The early finding of organic material in the original charcoal-bed inlet valves! sponse of the pressure drop to heating are gobd evidence that most of the trouble with pressure drop was caused by organics. Why the restriction built up so slowly in run 11 is not known. Designs were completed during run 11 for more positive remedies for the high pressure drop. These include a particle trap just ahead of the bed inlet manifbld and piping to bypass the inlet section of each bed where the plug is believed to be. Cooling-Water Systems R. B. Lindauer Treated-Water Cooler. — Befc;re run 8, 17 leaking tubes in the treated-water cooler were plugged. ! In runs 8 and 9 and through part of run 10 the leakage from the treated-water system averaged 2 gal/day or less. Then early\in January the leakage increased to about 6 gal/day and a few days before the end of run 10 began to increase again, reaching about 50 gal/day by the shutdown. Leakage in the cooler to the tower;water system was confirmed by the appearance of lithium in the tower water. Before run 11 the cooler was replaced by an available surplus heat 14M’SR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL~-3936, pp. 124-—-29. - 15SR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 47—48. 32 exchanger. The new exchanger has only two-thirds the tube area of the old cooler, but the heat transfer coefficient is higher, resulting in about the same performance. Through the end of February there was no evidence of leakage in the new cooler. Space Coolers. — Leaks had occurred in both reactor-cell space coolers at brazed joints on the brass tubing headers, and repairs had proved to be difficult because repair of one joint tended to open adjacent joints. Prior to run 8, both coolers (and all otfier in-cell components) were proved to be leak-tight, but when operation was resumed, condeqsed moisture began to appear in the system recirculating the cell atmosphere. Although leakage; from the space coolers was suspected, they could not be isolated for leak testing because of the high temperatures in the cell. In view of the discouraging experience with repairs, two replacément coolers were procured in Which the headers and nipples wete made of copper and the troublesome joints were welded. (The Heliarc welding of copper to copper was done at ORNL.) | Radiolytic Gas. — As reported previously, '* when the reactor is at full power about 3 scfh of radiolytic hydrogen and oxygen is evolved in the treated-water slystem. Steady-state accumulation in the system was reduced from 8 ft3 to 6 ft3 by supplying the tl‘lermal shield slides with \;rater " pumped from a small vented tank. Before run 8, a 350-gal degassing tank was installed in the return line from the main thermal shield and slides. With this tank in the system, the gas accumu- lation at full power was reduced to 3 ft3. The holdup was the same whether or not the slide supply tank was used in conjunction with the main degassing tank. The radiolytic gas being stripped from the water is diluted with air to below the explosive limit ifi the degassing tank and is then vented outside the reactor building. , _ Chemical Treatment. — In the course of testing and repair of leaks in the treated-water cooler, cooling-tower water leaked into the treated-water system. Before the reactor was started up in September, the resultant sodium contamination was reduced by dilution with condensate, after which the required corrosion inhibitor concentration was restored by adding more lithium nitrite. Power 6peration was resumed with 1.6 ppm sodium in the treated water. Activation of the sodium raised the radiation levels around the water system, but did not' interfere with operation. After September, no more additions of corrosion inhibitor were requirgd until January, when some makeup was needed because of leaks. | ‘ ' In the cooling-tower system, corrosion inhibition was by daiily additions of balls of Nalco 360 (sodium chromate and phosphate). To eliminate troublesome delposition of calcium phosphate on flowmeter tubes, the inhibitor was changed to a mixture of potassium dichrofnate and zinc sulfate. Continuous-feed addition equipment was also installed at the cgoling tower. J Component-Cooling System P. H. Harley During thié report period the component-cooling pumps, CCI‘D-I and CCP-2, operated 1490 and 2159 hr respectively. Although some difficulties were encountered with the system, the output was adequate and there was no interference with reactor operation, | 33 ~ Previous operatién of the pumps had indicated insufficient capacity, and the speed of both units was ihcreased during the shutdown after run 7, Operation was satisfactory in run 8, but in run 9 it became evident that there was a malfunction either in the output-pressure transmitter or in the pressure-control valve, PdCV-960. During the shutdown after run 9, inspection of PACV-960 showed that the valve stem was stuck with the valve partly open. Deterioration in the operation of this valve probably also accounted for at least part of the apparent loss of capacity of the pumps. The valve oberated satisfactorily after the valve was repacked and the stem polished and lightly lubricated., A drain was installed at the discharge of the gas cooler which should reduce moisture collection in the valve bonnet, and PACV-960 is cycled periodically to ensure free stem operation. The pressure transmitter was also checked and found to be operating properly. How- ever, the piping was rerouted to eliminate sections where condensate could collect. During this same maintenance period, the butterfly check valve from CCP-1 was found to be inoperative and was repaired. The failure occurred at a silicone rubber hinge which supports the two wings of the check valve. This check valve would prevent CCP-2 from developing pressure, because the gas could short-circuit back through CCP-1 to the blower suction. This was the second reported failure of a check valve; it had been in service between 1548 and 3074 hr. A similar check valve had failed on CCP-2 in January 1966, after 1640 hr of operation, and had been replaced by one made at ORNL as there was no sparé on hand. The locally fabricated check valve was replaced in September 1966, although it had not failed after ~.1800 hr. Pump CCP-1 had to be stoppéd during run 10 because of low oil pressure, and CCP.-Z was used for the remainder of the run. After run 10, a cracked copper fitting in the oil system was repaired, ~and 2 gal of oil was added to bring the oil level back to normal. The drive belts on the component-cooling pumps caused intermittent difficulties through run 7.- Prior to run 8 the motor mounts were strengthened to reduce flexing, and deflectors were installed to try cooling of the belts with the incoming gas. Even with the deflectors, a thermocouple near the belt of CCP-1 indicated an ambient temperature of 150 to 160°F. Although this is above the 130°F for which the belts are rated, inspection in December, after 1400 hr of operation, showed thét the belts were in good condition and only moderate tightening was required. There has been no evidence of belt slippage' since the motor mounts were reinforced. The rupture disk installed to eliminate the pressur!e relief valve leakage would not withstand the shock of starting the blowers and has been removed. No inspection has been made on the strainer installed in the gas piping in August. However, there has been no increase in system pressure drop to indicate collection of material in the strainer. Pump CCP-3, which cools out-of -cell freeze valves, operated the entire period without trouble. 34 - Salt-Pump Oil Systems J. L. Crowley The systems supplying oil for lubrication and cooling in the salt pumps operated.continuously for the last 51/2 months of this report .period. Operation was practically free of trouble. Oil leaking past the lower shaft seal in a salt pu.mp collects“in an external tank. Oil from the fuel pump, which had accumulated in earlier months at rates up to 20 cc/day, !® accumulated in October and November at about 5 cc/day. There was no accumuiation in December, but in January | the oil began to collect again at about 5 cc/day. Seal leakage fi-om the coolant pump, which had accumulated previously at aBout 2 cc/day, increased to about 17lcc/day. These rates are far below the 1000 cc/day that was originally set as acceptable, I; Leakage of oil into the salt pump bowls is, in principle, detectable by changes in oil in- ventories. The inventories in the two systems both indicated a ;mall, unaccounted loss over the 51/2 months of operation, as shown in Table 1.3, In four mon%hs of earlier operation,’® the - calculated loss from the fuel system was —0.3 liter (an apparent gain), and from the coolant sys- tem there was an apparent loss of 0.3 liter. The significance ofjthese changes is somewhat ques- tionable because uncertainties in the inventories are relatively farge (around 1 liter or more). But during the last months of operation, the decrease in reservoir levels was rather steady, and there was no evidence of more oil being trapped in the motor housings. Table 1.3. Oil Systems Inventory Changes i September 1966—February 1967 Changé (liters) Item Fuel Coolant - System System Removed in samples 6.49 ” 5.64 “Accumulation in catch tank 0.57 “ 1.35 _ - Total accounted for - 7.06 . 6.99 Decrease in reservoir 8.34 | 7.56 Apparent loss 1.28 0.57 . - Electrical System T. Mullinix T. L. Hudson R. H. Guymon During the shutdown prior to run 8, the entire electrical system at the reactor site was over- - hauled and put in good operating condition. Most of the major electrical breakers were tested and 16/bid., pp. 50—51. '35 calibrated to assure proper operation. One of the emergency-power diesel generators (DG-3), which had a cracked engine block., was replaced with a unit of the same size obtained from another instal- lation. In additioh, the 250-hp.motors for the main radiator blowers and the two 48-v dc generators were inspected. and reconditioned, | ‘ In general, the electrical system performed satisfactorily during this period. The emergency- power generators are operated under load routinely for test purposes, and they have always oper- ated when required. On September 30 a major power failure in the X-10 area caused an automatic transfer to the alternate feeder for the MSRE. During this outage, two of the MSRE diesel gen- erators were operated in parallel with TVA to help supply the X-10 electrical load. This opera- tion was entirely satisfac‘tory.' The only difficulty encoun'tered with the electrical system was a fuse failure in the high-reliability ac power supply (static dc-ac inverter) which.resulted in an un- scheduled control-rod scram. Heaters T. L. Hudson During the shutdown (in September) a resistance check was made on all in-cell heaters at junction boxes located outside the reactor and drain-tank cell. Other than heater HX-1, which had partly failed before the shutdown, only one heater, H102-1S, was found that had failed. This was one of three installed spare heaters on'the vertical section of pipe under the heat exchanger. No repairs were required, since there is adequate capacity in the other heaters on this section of pipe. ~ The resistances of the coolant-system heaters located outside of the reactor cell were also checked. Three heaters were found that had failed. One was on the coolant system 5-in. piping, one on the coolant-pump furnace, and the other one, a spare heater, was ofi the fill-line piping. "It was necessary to replace only the heater on the 5-in. piping, since there was adequate heater capacify at the other locations to preheat the coolant system. Aside from heater-element failures, a remote disconnect for one of the fuel-pump heaters (heater FP-1) was damaged during remote operations associated with the tha\;ving of the salt plug in the fuel-pump off-gas line. The heater was plu'gged into a spare disconnect to restore it to ) service. _ ’ Some additional heater-element failuréé héve occurred since the operation of the reactor sys- tem was resumed. As mentioned above; part of the elements in heater HX-1 had failed earlier, On October 12, additional elements failed in this heater, and on October 28 the last of the ele- ments failed, Using the other heaters, it has been possible to keep this section of the heat ex- changer adequately heated. This was aided by the coolant system, sihce coolant-salt circulation was maintained during the period. ‘When the fuel and coolant systems are drained, tests will be made to determine whether this heater must be replaced. ' One of the elements on heater H-106-4 failed on January 19, 1967. However it has been pos- sible to maintain adequate temperatures using the other elements and adjacent heaters. 36 Control Rods and Drives R. H. Guymon M. Richardson The performance of .the control rods and drives has been within the operating limits throughout the operation. None of the rods has ever failed to scram on request, but some difficulty was en- countered in withdrawing one of the rods. After run 7 we found that the drop time for control rod 3 had increased slightly.!? While we were investigating this problem during the shutdown before run 8, the drop time increased still further and approached the limit of 1.3 sec. In addition, occasional hanging was observed when the rod was raised 2 in. from its fully inserted position. We removed the rod-drive assembly and found that the long drop times were caused by a bent cooling-air tube. The air tube presumably was bent when the drive unit was last reinstalled (February 1966). The hanging on rod with- drawal was attributed to interference between sharp edges at the bottom of the rod and a protru- sion of ‘a warped guide in the thimble. We replaced both the rod drive and the control rod with spare units after first rounding the sharp corners at the lower end of the spare rod. This relieved the hanging problem and gave an initial average rod-drop time (for 35 drops) of 0.92 sec with a maximum drop time of (.95 sec. Subsequent measurements of the drop for rod 3 gave 0.85 sec. (The drop time normally decreases because of increasing flexibility of the rod with continued use.) Shortly after the new assembly for control rod 3 was put in service, we observed a shift of about 0.4 in. in the reference zero position as indicated by a special single-point indicator near the bottom of the rod thimble. There has been no subsequent change in the reference position, and the cause of the shift is not definitely known. However, when a similar shift occurred pre- viously,!® it was attributed to slippage of a drive chain on a sprocket gear. The lower-limit switch on rod 3 sometimes sticks after a scram, and the rod must be fully withdrawn to dislodge it. This may be a recurrence of a similar, previously reported difficulty,!® but it presents no difficulty in normal operation. | On February 13, 1967, the fine synchro on rod 2 failed. The coarse synchro is f‘uvnctioning properly and provides sufficient information for continued operation. Samplers R. H. Guymon R. B. Gallaher The fuel and coolant salt samplers met the requirements of the experimental program without delays or serious difficulties. As reported previously, flush salt that froze in the fuel sampler line during an accidental overfill was melted with external heaters. However, some obstruction femained at the top of the pump tank until the system was heated when there was flush salt in 171bid., p. 53. . '8MSR Program Semiann. Progr. Rept. Feb, 28, 1966, ORNL-3936, p. 54. 37 the pump.. Thereafter there was no interference with sampling, nor any mechanical‘ trouble of any consequence throughou:lt the report period. Contamination, although adequately controlled, did become more of a problem because of the higher activity of the salt and the nature of some of the special devices that were lowered into the fuel pump. | During this report period 66 samples were removed from the fuel system and 8 were removed from the coolant system. Of the fuel-system samples, 43 were routine 10-g salt samples, 6 were large salt samples (50 g) for oxide analyses, and 17 were special-purpose samples related to studies of the oxidation state of the uranium and fission product behavior. Containment P. H. Harley H. B. Piper Reactor Cell Leakage. — The secondary containment, consisting of the reactor and drain-tank cells and the closure devices in their penetrations, was shown to have acceptably low leakage at “the beginning of the report period. It remained so throughout the period, but the routine meas- urement of cell inleakage during operation gave erroneously high results in run 9 and was some- what uncertain in run 10 because of leaking pneumatic lines, o In a comprehensive test of penetration block valves and check valves before run 8, only 8 of 160 valves showed leakage above the specified. allowable rate. Metal filings and bits of Teflon pipe tape were found in some of the valves, and cleaning and replacément of some soft seats cor- rected all of the leaks, o The reactor and drain-tank cells were tested at 10 psig three times during the period.' Before run 8 the measured leak rate was 65 scfd; before run 10, 43 scfd. These rates are factors of 3.0 and 4.5 below the acceptable rate at 10 psig. The 10-psig test before run 11 was only long enough to show that the leak rate was acceptably low; no exact figure was obtained. For normal operation the cell is held at —2 psig, and the permissible inleakage has been set conservatively at 85 scfd. (This was based on an extrapolation from accident conditions which assumed orifice flow through all leaks.) In run 8 the measured rate was 65 scfd, within the prescribed limit, but by a much smaller factor than was obtained in the 10-psig test just before the run. In run'9 the calculated inleakage increased to more than 300 scfd, and one reason for termi- nating the run was to investigate this problem. Several valve-operator pneumatic lines were found to be leaking in the cell, but when the run 9 measurements were corrected for their contributions, the net rate was only 62 scfd. In run 10 the pneumatic-line leaks were measured continuously and deducted frorfi the total inleakage, giving a fiet inleakage rate around 50 scfd. In run 11, after the air-line leaks were stopped, the measured inleakage. was only about 10 scfd. | Leaks in the pneumatic lines do not represent possible routes for activity release in the event of a spill in the cel‘l, because block valves just outside the cell automatically shut off the lines if the cell goes above atmospheric pressure. Of course theseleaks must be taken into account in the calculations of the inleakage through other routes, and, during run 10, flowmeters were used in ‘the air lines to obtain the necesséry information. The leakage from eight lines became quite large, 38 however (up to 3500 scfd), causing considerable uncertainty in the cell inleakage measurement. Nitrogen was substituted for air in the leaking lines to keep the oxygen concentration in the cell below 4% (to rule out fire in the event of an oil leak). After run 10 the pneumatic-line leaks were traced to quick-disconnects in which the neoprene seals had become embrittled. There were 18 disconnects with elastomer seals; eight disconnects, all near the center of the reactor cell, were leaking. Seventeen of these disconnects were replaced with special adaptors sealed at one end by an aluminum gasket and at the other by a standard metal-tubing compression fitting. (One disconnect was left because it was on a line that is always at cell pressure.) After the replacements, all 17 lines were proved to be leak-tight. Activity Releases. — Operation of the reactor was the direct cause of only two minor releases of activity to the ventilation stack. These occurred while capsules containing gas from the pump bowl were being removed through the sampler-enricher and involved 1 and 2 mc of iodine. There was also minor contamination of the work area at the sampler-enricher, which was easily cleaned up. Continuous monitoring of the ventilation stack showed that a total of 206 mc of iodine and 9 mc of particulate activity were released during the six-month report period. Most of this (192 mc) was released in December while the fuel off-gas line at the pump bowl was being probed and inspected. Replacement of the fuel off-gas particle traps in September and again in January reéulted in considerable contamination of the interior of the vent house. The area was cleaned to below con- tamination-zone levels before operation was resumed, but after the January incident, this cleaning necessitated replacing the floor grating and the top layer of stacked concrete block under the grating. Stack Fans and Filters. — Stack flow was maintained continuously at a safe level throughout the period. The east stack fan was improved by the installation of new bearings, sheaves, and belts as had been done earlier on the west fan, Before fun 8, after other maintenance was finished, the roughing filters were replaced to reduce pressure drop and increase stack flow, 2. Component Development Dunlap Scott The effort of the development group consisted in work on the fission product off-gas removal system, on the sampler-enricher, and evaluation of the radioactive maintenance of the components, Several problems with the evaluation of their probable causes and solutions are given below. 2.1 SAMPLER-ENRICHER R. B. Gallaher Forty-seven routine samples were taken during reactor runs 8, 9, 10, and part of 11. Of these, 11 were 50-g samples for oxide and U3 */U** analysis. In addition to routine sampling, the sampler-enricher was used to perform a number of special sampling and addition operations as follows: 1. three samples containing absorptive wire for investigation of cover gas constituents, 2. two CuO capsules for determination of hydrocatbons in the cover gas, 3. three gas bombs for general cover gas analysis, 4, five capsules for addition of beryllium to the fuel. Refer to Chap. 7 of this report for a detailed discussion of the special samples. Cumulative totals to date for the sampler-enricher are 189 samples (including special samples) and 87 enrich- ments. Operation of the sampler was for the most part satisfactory. A brief summary of nonroutine operating and maintenance experience is given below. Contamination There has been a gradual increase in the radiation level inside the samplef. This is atfributed to small particles 6f salt which cling to the outside of the capsule and are dislodged during han- dling in the 1C and 3A areas. Some difficulty has been encountered with the spread of contami- - nation in the area a;ljacent_ to the sampler, caused apparently by the release of small quantities of gas or solid material during transfer of the sample into the transport cask. Two steps were taken to reduce this problem: (1) the interior areas of the sampler were decontaminated by wiping with adhesive tape and damp sponges, and (2) a ventilation duct, connected to the main building venti- . lator system, was installed near the tra.nsport cask position. 39 40 Capsule Support Wire As an empty capsule assembly was being attached to the drive unit latch, the wire which con- nects the capsule to the key pulled free from the key. Fortunately, the operator retained his grip on the capsule so that it was not drop})ed into the sample withdrawal pipe. It was found that the _knot on the end of the wire was too small and had pulled through the key. Since then all existing capsule assemblies have been reinspected. Also, new capsules have been fabricated using nickel- plated mild-steel tops in place of the original copper top-s, so that a magnet can be used for re- trieval in the event a capsule is dropped into the withdrawal pipe. Seal Leakage There has been a gradual increase in leakage of buffer gas through the upper seat of the main- tenance valve. This condition is appérently caused by an accumulation of salt particles on the seating surface. While the leak rate is relatively small (approximately 25 cm?3/min), it is suffi- cient to cause difficulty with proper opération of the interlock which indicates that the valve is closed. Since cleanup of the seating surfaces would be difficult, and since replacement of the valve is not warranted by the leak rate alone, a mechanical method of assuring that the valve is closed is being substituted for the existing pneumatic system. The désign of this section of the sampler-enricher does permit ready replacement of the maintenance valve should this become necessary. Maintenance The sample withdrawal pipe (line 999) was plugged with flush salt on July 24, 1966 (see Chap. 1, this report). The plug, which was located about 2 ft above the pump bowl, was melted by means of electrical heaters clamped to the outside of the pipe. The melted salt drained down the pipe and formed a new plug immediately above the pump. This second.pl‘ug was finally melted out by a combination of external heaters and heat from the pump bowl after the reactor system was brought up to normal operating temperature.’ ‘ - Manipulator Buffer System. — At some time during the maintenance period, a small leak de- veloped in the manipulator buffer system, and both boots were replaced. A small ‘‘pinhole’’ was found in the atmosphere—side boot. The cause of the hole was not determined. This set of boots had been used for 35 sample cycles ~ Access Door Mechanism. — Durmg one sampling cycle, a cotter key was noticed on the floor of area 3A. Close examination of the access port door mechanism showed that the key had come out of the top of the lower hinge pin.. The hinge pin had then worked down out of the top of the hinge. Using only the one hand manipulatof, the pin was repositioned properly in the hinge, a new IR. B. Gallaher, Removal of Flush Salt from Line 999, MSR-66-33 {Oct, 31, 1966). 41 cotter key was inserted into the 0.100-in.-diam hole in the pin, and the key was then spread to lock it in place. In doing this the lower key was knocked out and had to be replaced also. This repair work did not interfere with reactor operation or the sampling schedule. Yacuum P‘umps. — The radiation level in the vacuum-pump enclosure was checked after a two- month cooling-off period (following shutdown on July 17, 1966). It was found that the radiation level was less than 100 mr/hr, which is low enough to permit direct maintenance work in this area. The oil level was checked on both pumps, and oil was added to one of thé pumps. Spare Parts. — In an attempt to minimize the time required for haintenance, arrangements have been made to stock certain critical subassemblies. These include a manipulator arm, an area 1C subassembly, and an operational valve assembly. 2.2 COOLANT SAMPLER R. B. Gallaher Eight 10-g samples were isolated by the coolant sampler. A total of 53 samples, including two 50-g samples, have been taken with this equipment. ‘ The leak rate of the removal valve-seat buffer increased after one sample. This was corrected by replacing the valve seats. Also, an extra washer was added to the valve to permit more pres- sure to be applied to the seals. 2.3 FUEL PROCESSING SAMPLER R. B. Gallaher Installation of the fuel processing sampler is complete except for the shielding. Leak testing and operational checks have been completed. The system was found to be satisfactory except for a defective pai't in the containment buffer system. The éhielding has been designed and fabricated and is ready for installation. 2.4 OFF-GAS SAMPLER A. N. Smith R. B. Gallahe_r A system is being fabricated for installation in the reactor off-gas stream to pemit analysis for fission produét and other gases. Originally? it was planned that the system would include a thermal conductivity cell, a chromatographic cell, and a refrigerated molecular sieve trap. Three changes have been made to the sampler system since it was last reported: 1. Several developfiént problems associated with the chromatographic cell could not be re- solved by the scheduled startup date for the sampler. The chromatograph has therefore been omitted from the system. 2MSR Program Semiann. Progr. Rept. Feb, 28, 1966, ORNL-3936, p. 67. 42 2. A hydrocarbon oxidizer, a C02-H20 absorber, and a second thermal conductivity cell have been added to the system. These components, acting in concert, will provide a measure of the level of hydrocarbon contamination in the off-gas stream. In addition, the oxidizer unit, by virtue of its position upstream of the molecular sieve trap, will serve to protect the latter from fouling by organic material. 3. The original design provided for sampling the off-gas stream at a point immediately down- stream of the particle filter in the reactor off-gas line, specifically line 522. The piping has now been rearranged to provide an additional sample point immediately upstream of the 522 filter. Except for the approximate 40-min delay time, samples from the upstream point should be identical with pump bowl exit gas. Also, by using the hydrocarbon detector (see item 2 above), results from the two sample points may be used to judge the relative effectiveness of the off-gas filter. Gas will be routed to the sampler by w;'zly of existing lines (533, 561, and 538), so that the changes asso- ciated with the new sample point will be restricted to piping internal to the sampler. A diagram of the revised sampler system is shown in Fig. 2.1. Structural and piping work on the sampler proper is complete. Instrument and electrical work is approximately 90% complete. 2.5 OFF-GAS FILTER - MK I A. N. Smith The off-gas filter installed in line 522 in March 1966 consisted of a particle trap, or prefilter, and a charcoal filter (see ORNL-4037). Experience during operations plus hot-cell examination of the particle trap after removal from the system (see elsewhere, this section) indicated that the particle trap was too small and that the charcoal filter was unnecessary. Accordingly, the par- ticle trap and the charcoal filter assembly were replaced with two particle traps in parallel. The new particle traps, shown in Fig, 2.2, are similar to the old unit except for the following revisions: 1. The trap housing was increased from 4 to 6 in. ID, resulting in an increase in cross-sectional area of 225% in both the Yorkmesh and Fiberfrax sections. 2. The unit was in effect turned upside down so that the Yorkmesh section is at the top of the , unit and the Fiberfrax section is at the bottom. This change was made to permit heating of the Yorkmesh section (using beta decay heat and lowering the water level) while still maintaining cooling on the other two sections, Pipe connections were adjusted to compensate for this change, so that the flow path is now downward through the filter. 3. The disposition of the Yorkmesh was modified to provide increased frontal area in the direction normal to the flow. 4, Only the coarse Feltmetal was used. The original particle trap used a coarse section (removes 98% of particles <1.4 ) and a fine section (98% <0.1 f) in series. Since the first trap had shown very little loading in the final section, it was not believed that the fine Feltmetal was ) needed. 43 SAMPLER CONTAINMENT BOX / REFERENCE GAS (HELIUM)\ VENT « f————1 THERMAL- 4—’ @ CONDUCTIVITY S Dt CELL o & <1 THERMAL - CONDUCTIVITY HOLDUP '—E::: FUEL prd PUMP MAIN CHARCOAL BED HCV533 VALVES | @_T: v ORNL-DWG 67-4765 100 cc/min 3 psig \ DRAIN TANK VENT HEADER Fig. 2.1, Schematic Diagram — MSRE Off-Gas Sampler. CHARCOAL BED AUXILIARY £ 5. The total area of the filter elements was increased to 288 in.2. The two new traps, how- ever, differ in that the No. 1 unit has the support screen on the upstream side of the Feltmetal, resulting in an effective filter area of 30% of the total area. The original particle trap had an ef- fective filter area of 22 in.2, so that the relative increase is 4% for unit No. 1 and 13 x for unit No. 2. ORNL-DWG 67-4766 YORKMESH T.E. FELTMETAL [T‘ E . FIBERFRAX TE. . N NWIF 7 7 F T 7 7 77 7 7 F 77 7 7 7 F 7 fF f 7 SN ) f I, = = Y=t & [ v v/ - - —— — ' - — e e NN N : \ \ | y /4 1IN — ; N H V1 —_— ‘:'l —— .t\\\}\?-jr—'___ = ‘{7 \\'5\\1“ — \_‘\\\u\.l\\\\\\\\‘| I = e ——————— = X M ; 8 7. 7 /’ ™~ ™~ ™~ A \ i ¥ ] V i N ) | | | I ' = — S\ VA A AW 4 £ L S S L 7 yd NICKEL BAFFLES { o 1t 2 3 INGHES Fig. 2.2. Line 522 Particle Trap Mark Il, Particle Trap Subassembly. 144 oy 45 6. The depth of the Fiberfrax section was reduced by about 50% (from 6%, to 3% in.), because examination of the first trap did not show much oil, even in the first Fiberfrax section. Specific design data are as follows: Yorkmesh Section Material . ) ,304 S5 Overall size 6 in. in diam X 4 in. deep Total weight 0.62 1b Bulk density 0.16 g/cm3 Wire diam - 0.011 in, Calculated surface area _ L 894 in.? Feltmetal Section Material ' Huyck No, FM 225 Removal ratihg — gas service 98% <1.4 Size of inner element 3 in. in diam x 101/2 in, long Size of ocuter element 51/2 in. in diam x 101/2 in. long Effective surface area Unit No. 1 (effective area is reduced 85 in. 2 by perforated support plate) ‘ Unit No. 2 281 in.? Fiberfrax Section Material ‘ Carborundum Co. Long Staple Fiberfrax Mean fiber diam Spu Packing density _ . . 81/2 to 9 1b/ft> Total weight ' 213.5 ¢ Geometry ‘ Six annular compartments; two each at .1/4 in., 1/2 in., and 1 in. deep Efficiency and pressure drop tests were made on the new particle traps prior to installation. The results were as follows: . Unit No. DOP* Efficiency Pressure Drop at 15 liters/min 1 _ 99,588 <1 in. H,O S 2 . 99,960 <1 in. H20 *Dioctylphthalate-polydisperse mean particle size, 0.8 [Lt. 2.6 FELTMETAL CAPACITY TEST A. N. Smith The particle trap used in the MSRE off-gas system® (line 522) contains three filter elements in series: a loose-knit bundle of stainless wire (Yorkmesh), a porous stainless steel cylinder IMSR Program Semiann. Progr. Rept. A‘ug. 31, 1966, ORNL-4037, pp. 7477, 46 i (Huyck Feltmetal No. FM 225), and a bed of inorganic fibers (Fiberfrax long staple). Sizing cal- culations for the unit have been hampered in part by a lack of information as to the loading capacity of the Feltmetal. Accordingly, a bench test was performed to develop data which might be used to estimate the life expectancy of the Feltmetal under reactor conditions. rd Assumptions _. In designing the experiment the following assumptions were made regardirig the reactor off-gas stream: _ 1. The character of the suspended particles is variable, ranging from crystalline tnoble—g:sls daughter product) to tarlike (oil decomposition products), with various intermediate mixtures. 2. The particle size is on the order of 1 p. . 3. The concentration of particleé in the off-gas stream is such that the rate of accumulation on the filter is on the order of 1 g/day. This estimate is jbased on the assumption that salt en- trainment is negligible, the daughter product deposition rate is ~0.1 g/day (approximately 1% of 7.5 g of 235U per day), and the pump oil leakage averages 4 g/day, of which 25% is converted to solids which are deposited on the particle trap. Experimental Procedure Particles suspended in a helium carrier stream were pas.sed through samples of the Feltmetal (see Fig. 2.3). Ducting upstream and downstream of the filter was glass, permitting direct obser- vation of the flowing stream. The geometry was such that particles on the order of 5 y and larger could be expected to settle out before reaching the filter element. The particles were generated by evaporation of hexane from very fine droplets of hexane solu- tion containing the solid under study, which was aspirated into the flowing gas stream. The con- centration of these solutions and the flow conditions at the aspirator were held constant, such that the particle size could be expected to be the same for the three solids studied. Dave Moulton of Reactor Chemistry was responsible for conception and development of the particle generator. The test was divided into two phases: ' 1. The change in load capacity as a function of particle character was checked by using three different solids: naphthalene, stearic acid, and paraffin. 2. Using stearic acid, the effect of support screen position on load capacity was measured. The support screen, a thin perforated metal sheet, is normally placed on the downstream side of the Feltmetal to provide mechanical support in the event of excessive pressure drop across the filter. In the MSRE, however, routine backblowing operations had made excessive pressure drop in the reverse-flow direction more likely. It was thus desirable to know the effect of providing mechanical support on the upstream side of the filter. ‘ Helium flow during all tests was maintained at 5.3 liters/min. - Collection of solids was con-, tinfied until the pressure drop across the Feltmetal rose to abofit 3.5 psi. The quantity of solids collected was determined by weighing. t 47 ORNL—DWG 67 — 4767 : VENT HELIUM HEADER J 20in. FILTER . SAMPLE |1 Il II ] SECONDARY FILTER | | il . b T_ 1 ] * . 2-in. DIAM GLASS PIPE FEED . \ SOLUTION —=| |~ e =+ PARTICLE N GENERATOR — . — |=— MERCURY MANOMETER 7 2 = N N GAS FEED /0.0|3in, _\ HOLE DETAIL OF PARTICLE 4 § 2in " GENERATOR p 2 RN Z N N AT TSI —u—— SOLUTION FEED LINE N N\ N ) N N Fig. 2.3. Capacity Test — Feltmetal (Huyck No. FM 225). . Results The test results are summarized in Table 2.1, The relative capacity of the Feltmetal for naphthalene/stearic acid/paraffin was found to be about 1000/5/1. If these figures are applied to the reactor system at a collection rate of 1 ‘g/day,‘the life expectancy of the new MSRE particle trap ( see Sect. 2.5) might be as little as 3 weeks or as much as 70 years, depending on the char- acter of the material collected. This estimate is based on a Feltmetal area of 280 in.2 and a permissible pressure drop of 5 psi. The capacity of the Feltmetal with the support plate on the upstream side was found to be about 25% of the capacity obtained with the support plate on the downstream side. This result is in reasonable agreement with the calculated value of 30% free area for the support screen. 2.7 EXAMINATION OF THE MK | OFF-GAS FILTER ; Dunlap Scott - A. N. Smith The Mark I lparticle trap,* which had been installed in the fuel system off-gas line in April 1966, was replaced in September 1966 with a new trap of identical design. In October 1966 the old. ‘MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 75, 48 Table 2.1. Loading Capacity of Huyck Feltmetal No. FM 2257 Weight Collected Final AP Unit AP 2 i . : e Type of Solid (g/m.z) (psi) : (psi g in.”) Capacity for Different Types of Solids Naphthalene 0.96 0.058 0.06 Stearic acid 0.31 3.6 . ‘ 12 Paraffin - 0.06 3.9 65 / Support Plate Final AP (psi) Weight Collected (g) Capacity vs Support-Plate Position Upstream 3.6 0.186 Downstream 3.5 0.73 dcarrier gas was helium at a flow of 5.3 liters/min, trap was taken to the hot cells for testing and examination, Pressure drop tests were performed, after which disassembly, examination, and sampling were done; the samples were analyzed by the Analytical Chemistry Division. A description of this work is given below. Pressure Drop Tests The setup for the pressure drop tests is shown in Fig. 2.4. Helium was metered to the particle trap through a calibrated flowmeter at a metering pressure of 10 psig. Flow was controlled at V-1, and the pressur;a drop was .assumed to be the PI readin‘g, since the gas was vented to the hot cell (cell pressure controlled at — 1.5 in. HZO) . After the total drop through the unit was determined at various flows, three 14-'in.-diarn holes were drilled through\the wall of the trap to permit the flow tq short-circuit to the cell without passing through the Fiberfrax section, and the pressure drop meas- urements were repeated. The test results are plotted in Fig. 2.5. Also shown in Fig. 2.5 are the results of the pressure drop measurements made on the clean filter before installation in the MSRE off-gas system, | The pressure drop data before and after reactor service indicate the following: 1. The pressure drop after service was about 20 times the ‘‘clean’’ pressure drop. Assuming orifice-type flow, this would represent a reduction in flow area of about 80%. 2. The bulk of the pressure drop occurred in the Yorkmesh and Feltmetal sections of the filter. Thus the Fiberfrax section was essentially clean. The particle trap was in service in the reactor off-gas line dufing the operating period from April 1966 through July 1966. During this time the pressure drop across the particle trap exhibited wide fluctuations., Increases were usually, but not necessarily, associated with periods of power \ 49 ORNL-DWG 67-4768 (@ YORKMESH 304 STAINLESS STEEL | , ® COARSE FELTMETAL ~ ® FINE FELTMETAL . F @ FIBERFRAX 2 QIPI \' E7 p_9 ¢ ® % ’ 'kilvfél——h %_%}l ) 1-—HOT CELL km~d_77 - = 7] BLDG 3026 T ———— HELIUM CYLINDER T HOLES DRILLED THROUGH (/| PIPE WALL AT THIS POINT TO OBTAIN PRESSURE DROP = (&L EXCLUSIVE OF FIBERFRAX Fig. 2.4. Pressure Drop Test — MSRE Particle Trap No. 1. CRNL-DWG 67-4769 10 77 1/ 5 - (® OATA TAKEN 10-24-66 74 L SZ ‘ ' 4 ol - ,// >(3) DATA TAKEN /' 10-25-66 / 1 S —— (1) CLEAN FILTER BEFCRE INSTALLING B —— (@) TOTAL DROP ACROSS FILTER AFTER REMOVAL = — FROM MSRE OFF-GAS SYSTEM a 05— / Q (® SAME As CURVE (2) EXCEPT PRESSURE )4 o - DROP THROUGH FIBERFRAX IS EXCLUDED / o _ i 4 ' =2 7 / o o2 : 1 a . / NORMAL MSRE e ot OFF-GAS FLOW—— ) /! 4 V4 0.05 VAN )4 (1) DATA TAKEN 3-24-66 7 . 0.02 METER NO. A-2784 USED FOR ALL TESTS METERING : PRESSURE CURVE NO.{ AMBIENT; METERING PRESSURE CURVE NO. 2 AND 3 10 psig | R 0.01 IR o L 0.2 05 A 2 5 10 20 50 100 HELIUM FLOW (liters [STP) /min) Fig. 2.5. Pressure Drop Results — MSRE Trap No. 1. 50 operation. Decreases were sometimes unexplained and at other times were associated with delib- - erate attempts to clear the trap, such as reverse-blowing with helium. Table 2.2 shows steady- state values for the particle-trap pressure drop at ;larious times during the April to July 1966 period. The fluctuations in pressure drop seem to indicate that material was alternately deposited and then removed by increased pressure drop or breakdown (such as by thermal or radiation ef- fects). The maximum pressure drop of 9.7 psi would indicate a flow area reduction of about 95%. Disassembly The particle trap was disassembled to permit examination of the three main sections of the trap. The cuts were made with a band saw which is part of the hot-cell equipment in the HRLEL. All operations o‘fhandling, cutting, shifting, disassembling, sampling, weighing, photographing, and preparing samples for shipment were done by the experienced operators of the HRLEL using manipulators with direct and periscopic viewing, and the work suffered very little from the remote handling requirement, . The first cut was made at section A-A, 5 in. above the lower weld (Fig. 2.6). This point was chosen to pemmit removal of the bellows section of the inlet pipe as well as the upper section of the coarse and fine filter sections. ' A second saw cut was made through the upper filter at section B-B to get a closer look at a deposit on the lower surface of the upper end flange. Observations Yorkmesh Section. — The area of the Yorkmesh which had been immediately below the inlet pipe was covered with a blue-gray to black mat which had completely filled the space between the wires of the mesh (Fig. 2.7). The shape of the mat corresponded to the bottom of the inlet ORNL-DWG 66-11444R FINE METALLIC FILTER FIBERFRAX COARSE METALLIC FILTER . (LONG FIBER) THERMOCOQUPLE (TYP) LOWER WELD SAMPLE 97 "' w—— () I!”W///////////' ///[/f/////' 7 /////////////////////////////,/,l Er7: 2 II I "z?’(, i ' 7 fi S | R | BT dfz}.yfi# B } ; = (1 il — [TIETS ATIAAVTETRRNIVIARL N 4 2z Iz rrr == SU———— CYTTTTTTTrTOTYY T m fwcfil M[fl 7. }/ = - 7 _ _ L T %’ | 4 | %/,g ’ ;‘( it ' ) ] G l ,m,.mmm”/////// /// SAMPLES ‘A LsampLe 3A L —-"g" 58 AND 6 STAINLESS STEEL MESH NICKEL BAFFLE (8)- Fig. 2.6. Location of Sofiaple Points in Particle Trap. 51 R32824 Fig. 2.7. Inlet to Yorkmesh Section MSRE Particle Trap Mark I. 52 Table 2.2. Pressure Drop Ex'perience — Particle Trap Mark | Particle Trap Pressure Drlop : Reactor Power at Date and Time at 4.2 liters of Helium per min Time of Reading (psi) . ' (Mw) Before installation ‘ 0.06 4-4-66 - k <0.1 0 4-8-66 | 1.1 0 4-9-66 ) | 0.9 0 4-10-66 . 2.8 0 5-20-66 9.7 6 5-31-66 2.2 0 6-8-66 0.8 0 7-16-66 ‘ 8.7 _ 7.1 7-+16-66 2.8 ‘ 0 9-16-66 1.3 (2 months after. shutdown) ! Hot-cell test : 1.2 pipe, and it is likely that this material was the major restriction to gas flow during the operation in the reactor. Since the inlet pipe temperature probably increased several hundred degrees during power operation of the reactor, it is believed that this restriction behaved much like a thermal valve. This could account for the unexpected increase in pressure drop while at power and the decrease - in pressure drop when the power was reduced. In laboratory tests® using a sample of the porc;us metal to filter an oil mist, a rise in temperature caused a decrease in the pressure drop of a loaded filter. This predictable-result was attributed to boil-off of the oil. . A radiation survey made around the outside surface of this lower section of the trap gave readings ranging from 6900 to 8000 t/hr. The probe read 11,400 r/hr when inserted into the position formerly occupied by the inlet pipe. The radiation level dropped off to 4100 r/hr at the bottom of the trap. The bottom of the inlet pipe had a deposit of blackish material which corre- sponded to that in the Yorkmesh. The inside of the inlet pipe at the upper end of the bellows ap- peared clean and free of deposit. The exterior of the bellows had some of the light-yellow powdery material on a background of dark brown. SB. F. Hitch et al., Tests of Various Particle Filters for Removal of Oil Mists and Hydrocarbon Vapor, ORNL-TM-1623 ( Sept. 7, 1966). ' ' ’ - 53 The outer shell of the lower section was slit at the welti, and the Yorkmesh was removed. It was found that the surface of the outer wires of the mesh bundle was covered with a thin layer of amber-colored organic material.. Much of this material evaporated from the heat of the floodlamp used to make the photographs. As the bundle was unrolled, the color of the film on the wire changed from amber to brown to black near the center. The black material was thicker than the wire by a factor of 2 or 3. This material was brittle, as was the wire, and much of it came loose as the wire was flexed. Samples of this material, designated Nos. 5B and 6, were taken for examination and chemical analysis. A piece of the wire covered with the black material was removed and mounted for metal- lografihic examination, and it was found to be heavily carburized with a continuous network of carbide in the grain boundaries. There was no evidence of melting of the wire; however, the grain growth and other changes indicated operating temperatures of at least 1200°F. The nonmetallic deposit observed on the wire mesh was apparently of a carbonaceous nature and appeared to have been deposited in layers. These ‘‘growth rings’’ were probably the result of off-gas température and flowrate changes. Porous Metal Section. — The perforatedrplate of the coarse filter section and the lower flange of the filter assembly were covered with a stratified scale (view A-A, Fig. 2.6). The color varied from a very light yellow to orange. One stratum in the lower flange area appeared gray, almost black. A sample (No. 9) was taken of the light-colored material, including some of the black. The perforated plate of the fine filter section was covered with a thin, dark-brown coating, which seemed to be evenly distributed over the surface of the plate. The inner surface of the outer wall of the trap was covered with a light-amber coétin'g, which was also evenly distributed. It is be- lieved that these coatings were deposited by condefisation and sputtering of the oil from the adja- cent filter and that the dark-brown color indicates that the porous metal screen had operated at a much higher temperature than had the outer wall. The lower surface of the upper flange (view B-B, Fig. 2.6) contained a deposit which had the appearance of organic residue. The deposit was amber colored, and the fractured edge (Fig. 2.8) gave the impression that the material was brittle. There is as yet no explanation for formation of this deposit or how it came to be formed in this particular location, The radiation level on the outside of this section of the filter read from 200 to 360 r/hr. A piece of the inner (coarse) Feltmetal and perforated plate (sample 3A) was removed and examined in the hot cell. ' ' Fiberfrax Section. — The upper section of the filter containing the Fiberfrax was slit open, and the layers were examined. The material at the entrance showed an oillike discoloration, but there was no evidence of any significant accumuzlation of material. Corfiparison of the weights of the. | different layers with the weights of the material originally loaded indicated changes’'of less than 0.2 & | An interesting observation relates to the very low radioactivity level in this material, even in the entrance section which is separated only by the Feltmetal filter from an area containing ma- terial with activity levels of thousands of r/hr. The only detectable activity above the examination CYN 1015 FRACTURED EDGE Fig. 2.8. Upper End of Porous Metal Section — MSRE Particle Trap Mark I. 55 cell background (4.2 r/hr) was at the discharge end of the Fiberfrax section. It is probable that ~ this activity resulted from pressure transients which could have carried gaseous.decay products from the charcoal trap back upstream into the particle trap: Even so, the activity level was only 1.8 t/hr above background. Analytical Results A total of four samples from three different locat1ons (F1g 2. 6) were subJected to a variety of analytical tests. The samples were identified as follows: Sample No. Taken from 3A Coarse section of porous metal filter .9 Scale on lower flange of porous metal section 5B Mat at inlet to Yorkmesh section 6 Mat at inlet to Yorkmesh section For sections of sample 3A it was found that about the same weight loss (0.2% of sample weight) resulted from heating to 600°C in helium as from dipping in a trichlorethylene bath. The material removed by heating was cold trapped and found to be effectively decontaminated; however, the trichlorethylene wash was contaminated with fission products. ‘ Samples 5B and 9 were compared for low-temperature volatiles; at 150°C No. 9 lost 5% and No. 5B lost none. When raised to 600°C, the weight losses were about the same, 35% for sample 5B and 32% for sample 9. Analysis of the carbon content gave none for sample 9 and 9% for ‘ sample 5B. This indicates that sample 9 had not reached as high a temperature as had sample 5B. The mass spectrographic analysis of sample 6 indicated that there was a very high fraction of fission products. These are estimated to be 20 wt % Ba, 15 wt % Sr, and 0.2 wt % Y. In the same analysis the salt constituents Be and Zr were est1mated to be 0. 01 and 0.05 wt % respectively., In addition the material in samples 5B and 6 contained small quant1t1es of-Cr, Fe, and Ni, while sample 9 did not. The reliability of these values is compromised by difficulties caused by the presence of organics and small sections of wire in the sample, The gamma-ray spectrographic work indicated the ptesence of the following isotopes: 137Cs, 89g; 103Ry or 196Ry, 119mAg 95N, and ! *°La. ‘ All three samples were chemically analyzed for Be and the level was below the detectable limit of 0.1%. Attempts to analyze for Zr were compl1cated by the presence of large quant1t1es of Sr. ' The fraction of soluble hydrocarbons was determmed using CS " and the values were 5B, 60%; 6, 73%; and 9, 80%. The extract solutions from samples 5B and 6 were allowed to evaporate and a few milligrams of the residue was mounted between salt crystals for infrared analyS1S The sam- ples were identical and were characteristic of long-chain hydrocarbons.__There was no evidence of 56 any functional groups other than those involving carbon and hydrogen. Nor was there any evidence suggesting double or triple bonds. There was an indication of a possible mild cross-linkage. It:is likely that there is more cross-linkage ‘of the organic in the gas stream than appeared in these sam- ples, and the low indication could be due to the insolubility of the cross-linked organic and the high operating temperatures of the wire mesh, which would cause breakdown of the organic into elemental carbon and volatiles. | The following conclusions can be drawn from the results of the examination: 1. Since the spectrographic analysis indicated that the concentrations of Be and Zr were very small compared with those of the fission products Ba and Sr, the amount of entrained salt mist / carried to the filter was hegligible_. 2. The high activity and the large amount of barium and strontium in the inlet. section of the particle trap indicate that a large fraction of the solid daughters of Kr and Xe which decay in the line is carried down the line with the off-gas stream. 3. The distribution of activity indicates that entrance areas "'of the Yorkmesh were unexpectedly effective in trapping the solid fission products. Decay heat in this area resulted in temperatures above 1200°F. | 4. The collection of hydrocarbon mist on the Yorkmesh had probably enhanced the collecting efficiency of the mesh for solid particles. 2.8 MIGRATION OF SHORT-LIVED GASEOUS PRODUCTS INTO THE GRAPHITE R. J. Kedl | The measured concentrations of various fission products in the MSRE graphite samples were listed as a function of depth in the samples in the previous progress report.® Three of these ma- terials are daughters of very short-lived noble gases, '*°Ba from (16-sec) 4%Xe, 141Ce from (1.7-sec) '41Xe, and °'Y from (10-sec) °!Kr. A model has been developed which evaluates the concentration of “‘very short-lived’’ noble gases in the graphite while the reactor is at power. It can be shown that with the reactor at power, —xy\/ EA _ . vD €/D. e " /e - e = V5 ¢/ where - C s = noble-gas concentration in graphite (atoms per volume of graphite), Q = noble-gas generation rate in fuel salt (atoms/time, volume of salt), A = noble-gas decay constant, D = noble-gas diffusion coefficient in molten salt, s Dg = noble-gas diffusion coefficient in graphite, "€ = graphite void fraction available to gas, SMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 181—83. R T 57 x = distance in graphite. The very short-lived noble-gas concentration distribution in gréphite is approached very shortly after the reactor is brought to power. We can therefore compute the longer-lived daughter product concentration distribution as a function of time with the reactor at power. Also, the daughter product concentration can be taken thr0ugh the period when the reactor is shut down by consider- ing decay only. This model assumes that once a noble gas decays, its daulghter is immediately adsorbed by graphite and does not migrate. ‘ Concentration distributions were computed for 14°Ba, 141Ce, and °'Y in the MSRE core graph- ite samples using values of Dg obtained from the slope of the distribution curve. The computed values of the concentrations at the surface of the sample are shown in Figs. 2.9, 2.10, and 2.11, together with the measured distributions. For sake of clarity, the computed concentration is shown only at the surface, but the computed internal distribution would be roughly parallel to the measured internal distribution. 2 ORNL-OWG 67-4770 + SAMPLE FROM TOP OF CORE—WIDE FACE ® SAMPLE FROM MIDDLE OF CORE—WIDE FACE X SAMPLE FROM BOTTOM OF CORE—-W|DE FACE CIRCLED POINTS AT DEPTH=0C INDICATE COMPUTED © < - o ® VALUES 52 & Q Q © \ e 10! LR TN it \\. N <5 N MIDDLE 3 o~ N\ TOP\ E \X \ ’g BOTTOM T 2 | + : 1 AN S ‘+ X \ @ z \ = X {0 X 8 0 \\ \\‘\ ’ = N N\ @ (] — + E N\ AN \‘X S s AN = Q \\ N (&) n \ \ m ¥ \ N\ 2 . \+\ 10? 0 0.01 0.02 0.03 0.04 0.05 0.06 DEPTH IN GRAPHITE (in.) Fig. 2.9. 140p . Distribution in MSRE Graphite Samples. {4 co CONCENTRATION IN GRAPHITE (dpm /g-graphite at 4100 on 7-17-66) 58 CRNL-DWG 67-47T1 10 @ 5 \ - \. + SAMPLES FROM TOP OF CORE—WIDE FACE 2 ® SAMPLES FROM MIDDLE OF CORE—WIDE FACE — \ X SAMPLES FROM BOTTOM OF CORE—WIDE FACE N CIRCLED POINTS AT DEPTH=0 INDICATE COMPUTED 5)6 VALUES ' \[#] 10 MIDDLE + \ v \ \ N\®, TN > \\TOP \ '\ BOT TOM 9 , 10 \ AY AN \\ \\ AN _ e \ N ° \\ \\ \.\ \ N \ \ \\ 2 + ‘\ X 8 M 10 \ AY \ 5 + 2 10 0 0.0 0.02 0.03 0.04 0.05 0.06 DEPTH IN GRAPHITE (in.} Fig. 2.10. 141¢e Distribution in MSRE Graphite Samples. 59 ORNL-DWG 67-4772 s [\ DATA NOT AVAILABLE FOR TOP AND BOTTOM SAMPLES __| \ CIRCLED POINT AT DEPTH=0 INDICATES COMPUTED | \ . VALUE vd oL (=] / \ —X \\ 5 AV 2 B 10° . N\ \, \ 5 \ \MIDDLE L 91y CONCENTRATION IN GRAPHITE (dpm/g-grophne at 4400 on 7-17-66) o7 \ 10a \ N X ) \ 5 AN ® » , - \ 107 : 0 0.01 0.02 0,03 0.04 7 005 0,06 DEPTH IN GRAPHITE {in.) Fig. 2.11. 1Y Distribution in MSRE Graphite Samples. 2.9 REMOTE MAINTENANCE R. Blumberg The reactor shutdown which began in July 1966 extended. through October 30. " After that the reactor cell membrane was cut three separate times to carry out maintenance operations in the cell. During this period, power production increased from 7822 Mwhr to 16,277 Mwhr at the beginning of the last maintenance operation. 60 Summary of Remote Maintenance Tasks Performed August 26, 7822 Mwhr. — A removable section of off-gas line 522 was replaced. This spool piece was replaced because it was suspected of having a plug of frozen salt. This was the hottest piece removed from the reactor, approximately 2000 t/hr. It was removed with blind flanges ofi both ends for containment. September 7, 7822 Mwhr. — Both a control rod and a rod drive from position 3 were replaced. This was routine from a handling standpoint. The lower end of the rod measured 250 t/hr on contact. | September 9. — The west space cooler that had been removed the previous period for repair of a water leak was replaced. The assembly had some wipable contamination which was. re- moved at the decontamination facility. There were some local spots of induced radiation of up to 400 mr/hr at 6 in. The leak, which existed in a brazed joint, was rebrazed directly, under Health Physics supervision. r September 10, October 30. — The sampler withdrawal pipe (line 999, see Sect. 2.1) was thawed. By using a series of remotely manipulated external heaters, we managed to move the frozen salt down the pipe to within 3 in. of the end. The final thawing of this line required using the heat from the circulating flush salt plus that from several small ceramic heaters which had to be remotely positioned in the complex area around the pump tank afid line 999. This was a totally unanticipated job, and much of the difficulty was due to the crowding of various equipment into | one area. September 15. — The line 522 particle trap was removed and replaced. The trap which was re- moved was stored in a dry well at the burial ground using blind flanges for containment. This unit was later retrieved and examined at the HRLEL. : » September 16, 7822 Mwhr. — We completed installation of the new graphite and Hastelloy N surveillance samples, This went very smoothly and quickly. (The above operation completed the shutdown which started in July.) November 5, 10,067 Mwhr. — We installed three external heaters on line 522 at the pump in an attempt to thaw a salt plug (see Chap. 1). This proved to be ineffective. The radiation level at open tool penetration in the portable maintenance shield was 70 r/hr. - November 23, 11,236 Mwhr. — The plug in line 522 was removed by direct means. External heaters, chisel-type rods, a flexible shaft snake to clear a pipe with two 90° bends, and a borescope were used on this job. The flexible pipe spool piece was replaced in order to make sure that this line was open; 70 r/hr at tool penetrations was a definite factor in slowing down- the work. | December 5, 11,236 Mwhr. — We removed, repaired, and reinstalled valve 960, which controls the component cooling air. Gamma radiation was significantly lower here than for the job on line 522. Induced radiation in the valve (less than 20 mr/hr) was low enough to allow contact repair of the valve stem. wites] | ~J 61 Jonuary 16, 16,277 Mwhr. — We replaced 17 “sfiap-tite” quick disconnects in lines which supply air to operate valves. These disconnects had developed leaks due to radiation embrittle- | ment of an elastomer seal. The replacement fitting does not have ény elastomers. The work was started with only three days of decay, and the radiation at open tool holes in the maintenance shield was 60 r/hr. The difficult part of this job, from a mechanical handling standpoint, con- sisted in making up a 3/8-ir1. tubing joint on horizontal tubing at a distance of 20 ft. While the air lines were being refitted in the reactor cell, a particle trap of improved design was installed in the off-gas system. In this operation all of the existing parts were removed from the area of the old trap, leaving only the inlet and outlet flanges. The new assembly was fabri- cated in jigs to assure that it would fit the existing flangeé. -During the course of this job the working area became severely contaminated and had to be cleaned several times. Radiation Levels Radiation levels increased si gnificantly in almost every situation. These are described in Table 2.3, and a plot of in-cell radiation levels vs time is given in Fig. 2.12, There are several things of interest that one may note from these curves: 1. Both the reactor cell and drain cell radiatioh levels are increasing as more power is. produced. ‘ . } 2. The radiation level in the reactor cell, with the fuel salt drained, decays at a faster rate than the radiation level in the drain cell with the fuel in the drain tank. . 3. The dramatic increase in radiation level in the drain tanks during the period that the off-gas was vented through the drain tanks (December 23—January 14) gives some indication of how radio- actively hot the off-gas is. | 4. On December 11 after some 23 days of decay, we find that the radiation level at the north wall of the reactor cell is down to 250 r/hr from a level of 2000 r/hr on November 18. This is an indication of the residual radiation effects that-may be of consequence in a shutdown involving handling of major components. Table 2.3. Description of Radiation Levels That Affect Maintenance General background just above portable maintenance shield (PMS) when in place. Usually less than 5 mr/hr; however, over the off-gas line it was 10 to 20 mr/hr. At’an open tool penetration. Up to 100 r/hr recorded. Generally 70 r/hr over the off-gas line. 60 r/hr on PCV 919 air disconnect work. High bay levels when two roof blocks are off. The handling of these blocks and movement of the PMS slide must now be done routinely from the remote maintenance control room. During this operation, there were readings of 20 mr/hr in the reactor control room and 40 to 60 mr/hr in the hot change house. Hot spots on top of reactor cell before removal of lower shield blocks but after removal of the upper blocks. We have run into local hot spots as high as 400 mr/hr before, but during the most recent shutdown, there was an area reading 2 r/hr located alor_:g the south edge of block R above the heat exchangér. ORNL DWG 67-4773 -—— iympW 20G'6H 62 [N - JyMW 002°9) =K 141 o - q3and Hltm 1714 GNY T7132°38 4IA0 NHOM LHYVLS JIONVYNILNIVAN NI938 17130 N3dO 3SNOH AN3A NI A -4—— 13Nn4 NIvd0 B — Jymy 0029+ ¥2ILINDANS N NN Z AN R MR ¢ RN // NN AN ///WWW 7/ N TIME-SCALE ——— Ne—e S | i \/\"“'\/ N 1I%Ss HSN1d 3LvIN2HID aNY 714 . 225 3NN -H— ONI99NTdNN LYVLS N A \«<-PT NO. 6 BETWEEN THE TWO DRAIN TANKS (DRAIN CELL) bW e lv -4—— SX3078 dO0L 3IAOW3Y CENTER OF FUEL CHANNEL 0.04 ‘ ' /‘e/ 0.876ev ’ ’ l | || : A [ E)dE=0.9760 A L —6/—"" . el 0! : 0001 0.002 0005 001 0.02 005 04 0.2 0.5 1 ENERGY (ev) Fig. 5.2. Calculated Thermal Neutron Spectrum in the MSRE. to that in a single-region reactor, consisting of a'homogeneofis mixture of graphite and fuel salt, in which corrections have been introduced for the fuel channel geometric self-shielding of the absorption resonances in 238U. The physical neutron spectrum corresponding most closely to this theoretically calculated flux should be the spectrum within the graphite stringers. From Fig. 5.1 it may be seen that the flux spectrum above about 0.1 Mev exhibits the general characteristics of the fission spectrum. At lower energies (increased lethargies), there is a transition to the spectrum of scattered and moderated neutrons. The irregular de- pendence on energy of some of the group fluxes in the transition region can be identified with | the effects of neutron resonance scattering in the light elements in the fuel salt (’Li, °Be, and 1°F). | | Figure 5.2 éhows the flux spectrum in the energy region below about 1 ev (0.876 ev has been chosen as the effective upper cutoff energy for the thermal group in the MSRE calcula- tions). In this energy region the cross sections of the important nuclides in the MSRE do not exhibit strong resonance effects, and therefore a smooth curve has been drawn through the calculated group fluxes. Comparison of the curves in Fig. 5.2 shows that, within the fuel salt channel, there is a slight depression or flux disadvantage, together with an energy ‘‘hard- ening,’’ or preferential removal of low-energy neutrons, relative to the graphite. This spatial flux depression is relatively small from the standpoint of estimating reaction rates (maximum depression in the energy-integrated flux is less than 5%, volume averages over the graphite and fuel regions less than 3%). The local flux spectra shown in Fig, 5.2 are normalized such that the energy-integrated thermal flux, averaged over the total volume of fuel channel and associated graphite, is unity. Cross sections for the important nuclide constituents in the MSRE, averaged ovér the spectra shown in Figs. 5.1 and 5.2, are listed in Table 5.1. To obtain the effective cross 82 Tabie 5.1. Average Cross Sections for Thermal and Epithermal Neutron Reactions in the MSRE Spectrum Cross Section Averaged Cross Section Averaged Effective Cross Secticn Over Thermal Spectrum, Over Epithermal Spectrum, Nuclide Energy <0.876 ev Energy > 0.876 ev in Thermal Flux (barns) (barns) (barns) 6pi® 417.5 17.4 457.6 Boron 330.4 13.8 362.4 149gm 3.57 x 10* 90.8 3.6 x 10° 151gm 2.63 x 103 126.7 2.9 x 103 135xe 1.18 x 10° 84.6 1.18 x 10° Nonsaturatingé 20.0 (barns/fission) 10.0 (barns /fission) 43.1 (bams/fission) fission products 234U 39.7 35.4 121.4 233y (abs) 271.9 24.8 329.1 235y (v x fission) 555.4 37.4 641.8 236y 2.6 17.7 43.5 238y 1.2 9.4 22.9 239py (abs) 1404.0 20.5 1451.3 239py (v x fission) 2411.9 36.7 2496.7 58ped 0.53 0.036 0.61 59¢cod 16.3 2.6 22.25 4Cross section for the reaction 6Li(n, O'.)‘?'H. bNatural-enrichment boron (19.8% 10B). ®Estimated from L. L. Bennett, Recommended Fission Product Chains for Use in Reactor Evaluation Studies, ORNL-TM-1658 (Sept. 26, 1966). ' d . . - Used in flux wire monitoring. sections given in the final column of this table, we have used reactor theory to calculate the average ratio of integrated epithermal to themmal flux over the moderated region of the core. These effective cross sections, when multiplied by the magnitude of the thermal . flux, give the total reaction rates per atom for neutrons of all energies in the MSRE spec- trum. Unless otherwise specified, the cross sections listed in Table 5.1 are for (n,y) re- actions. Those nuclide reactions which are of importance in the interpretation of reactivity changes during operation, plus two nuclide reactions which have been used for thermal-flux wire monitoring, are listed in Table 5.1. Certain high-energy reactions, also of interest in interpreting MSRE operations, are listed in Table 5.2, These include two reactions which have been used for high-energy flux wire monitoring, and those reactions which produce isotopic constituents of interest from the stand- point of fuel-salt chemistry. In general, the reactivity effects of isotopic changes’due to re- 83 Table 5.2. Average Cross Sections and Fluxes for High-Energy Reactions in the MSRE Cross Section, Averaged o T Over Energy Spectrum Total Neutron Flux Above EO, Nuclide Reaction Energy Cutoff Above Eo _ Averaged Over Core Volume 11 -2 1 —1 - (Mev) (bams) (10" " neutrons c¢cm™ “ sec” ~ Mw ) %Be n,2n 1.83 0.234 ' 0.750 °Be n, @ 1,00 _0.0408 1.563 g n, p 4.50 ~0.0250 0.0974 PBg n, a 3.01 0.0988 0.271 54pe® n, p 2.02 0.1644 0.629 58N 2 n, p 1.22 0.1262 \ / 1.288 238y - n, 2n 6.06 0.2490 ©0.0201 a l . . . Used for flux wire monitoring. actions in this group are quite small com;;ared with those associated with the reactions listed in" Table 5.1. As indicated in Table 5.2, the cross sections are averaged ovér that portion of the high- energy spectrum of Fig. 5.1 for which the magnitude of the cross section is significant. The lower cutoff energy listed is that value chosen from the GAM-II group structure which corre- sponds most closely with the physical cutoff energy for the particular reaction considered. The upper cutoff energy is 15l Mev for all réactions. The calculated maghitudes of the neutron fluxes in these énergy intervals, averaged over the volume of the core, are also given in Table 5.2, - The effects of isotopic changes due to the reactions considered in Tables 5.1 and 5.2 are described in the following section. 5.2 1SOTOPIC CHANGES AND ASSOCIATED LONG-TERM REACTIVITY EFFECTS DURING REACTOR OPERATION ‘ In the MSRE the rate of depletion of the 23°U fuel charge is sufficiently low that some con- venient first-order approximations can be used in calculating the changes in isotopic composi- tion of the fuel salt. We have assumed that the magnitude and energy spectrum of the neutron flux remain constant during operation ét a given power level, and we have used the calculated effective cross sections listed in Tables 5.1 and 5.2. For all isotopic changes occurring in the fuel salt, account has to be taken of the ‘‘flux dilution’’ effect of the time the fuel spends in that part of the circulating system extemal to the core, out of the neutron flux. Thus the 2 calculated volume-average thermal flux for the entire fuel loop is 6.7 x 10'! neutrons cm™ sec™! Mw—!, whereas the average thermal flux over the core region is 2.0 x 10!2 neutrons 84 cm~? sec™! Mw~!. Of the isotopic changes being considered here, only the burnout of the residual !°B in the graphite is determined by the latter flux level. Rates of production or depletion of important isotopic constituents of the fuel salt, cal- culated on the basis outlined above, are listed in Table 5.3. The light-element reactions explicitly considered for this tabulation are i \ _ 19F(n,p)1°0 ——— ('H production) , 29 sec - ®Li(n, a)3H "(®Li depletion and 3H production) , °Be(n, 0)’He ——— > SLj (°Li production) , 4 sec 19F(n,a)1N——> 160 (180 production) . 7.4 sec To a good approximation, all the rates given in Table 5.3 may be assumed constant during a large fraction of the exposure life of the current fuel charge. For example, after one year’s operation of the MSRE at 7.5 Mw, the maximum corrections due to saturation effects would reduce the concentrations of 23°Pu and tritium, calculated according to the linear approxi- mation, by about 11 and 4% respectively. Table 5.3. !sotopic Changes in the Fuel Salt During Reactor Operation - Rate of Production (+) or Depletion (~) Nuclide . in Fuel-Salt Circulating System® (g/Mwd) lq +1.3x 10° 3H . +4,1%x 1074 51 N Depletion ,—/8.3 x 1074 Producéion - +0.4x 10”4 Net depletion —-7.9x10™* 160 +2.2x 1074, 234y -52x%x10"3 235y —1.30 236y +0.26 238y _ —0.19 23%py +0.19 “Calculations based on fuel salt volume of 70.5 ft°. 85 In addition to the bumup of 235U during power operation, the changes in the concentrations of 5Li, 234U, 236y, 238y, 239Py, and the nonsaturating fission products, all in the salt, and the residual !°B in the graphite induce reactivity changes which are important in the analys1s of the nuclearbehavior of the core. We shall exclude 235U from consideration here, since separate detailed account is taken of its inventory-reactivity changes in the analysis of op- erations. Most of the remaining effects given above manifest themselves as very slowly de- veloping positive reactivity changes, dependent on the time integral of the power or energy generated, The exceptions are 235U (for which there is a slight increase in concentration resulting from radiative capture in 235U) and the buildup of nonsaturating fission product poisons.v . ‘ As a typical example, reactivity effects at 10,000 Mwhr (the approxima.te exposure at the end of MSRE run 8), corresponding to these isotopic changes, were estimated from reactor theory. The results are given in Table 5.4. In this gfoup the terms of largest magnitude were found to arise from the production of 23°Pu and the bumout of the ®Lj initially present in the fuel salt. In addition to the results listed in Table 5.4, an algebralc formula was de- veloped for the irradiation dependence of all reactivity effects included in this group. This is being used in the on-line calculations of reactivity changes during MSRE operation. > SJ. R. Engel and B. E. Prince, The Reactivity Balance in the MSRE, ORNL-TM (in preparation). Tu.ble-5.4.. Reactivity Effects Due to Isotopic Changes in the MSRE" Approximate Reactivity Nuclide . Effect at 10,000 Mwhr ‘ (% Ak/k) 6L 0.017 10ge 0.007 234yb _0.001 236y —0.003 - 238y . 0.004 239p, . . 0.051 Nonsaturating -0.005 fission products Total 0.072 “The boron concentration present initially in the MSRE graphite was estimated from MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, p. 376. The calculation of the reactivity effect neglects a correction factor accounting for the spatial dependence of the boron burmnout in the graphite. Includes react1v1ty effects of both depletmn of 234y and associated production of 235U, 86 5.3 ANALYSIS OF TRANSIENT '35Xe POISONING Studies were continued with the purpose of correlating the time behavior of the 135Xe poisoning observed in the MSRE with calculations from a theoretical model. The mathematical model us;ed for this purpose has been previously described in refs. 6 and 7. In the present section we will give only a qualitative description of the main aspects and assumptions in the theoretical model. We will then compare graphically the calculated buildup and removal of 135Xe reacti\}ity, following changes in power level, with some of the experimental re- activity transients which have been observed from operation to date. Finally, we discuss some tentative conclusions which can be drawn from the currently available evidence. The theoretical model is an extension of the steady-state model described in ref. 8, to include the transient behavior of the 135Xe reactivity following a step change in reactor power level. In the model chosen, we have assumed that all the 13‘51 produced from fission remains in circulation with the salt. After decay to '3Xe, the xenon migrates to the acces- sible pores of the graphite at the boundaries of the fuel channels and also to minute helium bubbles distributed through the circulating salt stream. An effective mass transfer coefficient was used to describe the transfer of xenon from solution in the circulating salt to the interface between the liquid and the graphite pores at the channel boundaries. Equilibrium Henry’s law coefficients were used for the mass transfer of xenon between the liquid phase at the interface and the gas phase in the graphite pores. The numerical value used for the mass transfer co- , efficient between the circulating salt and the graphite was based on krypton injection experi- ments with flush salt circulating in the fuel loop, performed prior to nuclear operation of the MSRE. ® | Similar considerations were assumed to apply to the mass transfer of xenon from liquid solution to the gas bubbles. The coefficient of mass transfer from the liquid to a small gas bubble, of the order of 0.010 in. in diameter, moving through the main part of a fuel channel, was estimated from theoretical mass transfer correlations. The equilibrium !3%Xe poisoning was found to be relatively insensitive to the bubble diameter and mass transfer coefficient, over a reasonable range of uncértainty for these parameters.’ The computational model provides for different efficiencies of removal by the extemal stripping apparatus of xenon dissolved in the salt and that contained in the gas bubbles. The efficiency of removal (fraction of xenon removed per unit circulatedvthrough the spray - ring) of xenon dissolved in the salt was estimated to be between 10 and 15%, based on some early mockup experiments to evaluate the performance of the xenon removal apparatus. ®MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, pp. 82—87. "MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 13-21. 5Rr. J. Kedl and H. Houtzeel, Development of a Model for Computing 1335xe Migration in the MSRE, ORNL-4069 (in preparation). 87 Calculations of the steady-state '3°Xe poisoning, reported in ref. 7, indicated that the low apparent poisoning as a function of power level could be described by a variety of com- binations of circulating bubble volume fractions and bubble stripping efficiency. One of the purposes of the transient calculations, therefore, was to attempt to separate those parameter effects which could not be separated in the steady-state calculations. A number of transient reactivity curves were calculated, with the aid of an IBM 7090 pro- gram, based on the theoretical model described above. These calculationis were compared with experimental data logged by Ithe reactor’s BR 340 on-line computer. The apparent transient 135Xe poisoning was determined by subtracting all other known power-dependent reactivity effects from the reactivity change represented by movement of the regulating rod following a step change in the power level. In Figs. 5.3-5.13 we have compared some of the transient reactivity curves obtained from this analysis with some experimental transients, in the chronological order in which they were obtained. In each of these figures the solid curves represent the calculated behavior, and the plotted points show the observed experimental reactivity effect. At this date, only a few rela- ~ tively clean experimental transients corresponding to step changes in power level (for which the 7090 program was devised) have been obtained. However, several characteristics of the 135Xe behavior are indicated from these curves. These will be discussed by considering the figures in order. Figure 5.3 shows the calculated and observed xenon transients for a step increase in re- actor power from 0 to 7.2 Mw. The calculations (solid curves) were made for a variety of - circulating void fractions (ab) to show the effect of this parameter on the xenon poisoning, A single bubble stripping efficiency (¢,) of 10% was used for this figure. This relatively low efficiency is approximately equal to the efficiency estimated for the stripping of xenon dissolved in the salt; it was considered to be a lower limit and a reasonable first approximé— tion for this parameter, in the absence of strohg evidence for assuming a higher value for the bubbles. The effectiveness of the circulating gas in reducing the poison level is due to the combined effects of the large overall gas-liquid surface area for mass transfer to the bubbles and of the large xenon storage capability of the bubbles (because of the extreme insolubility of xenon'in molten salt). Thus the bubbles compete effectively with the graphite for removal of xenon from the liquid, and xenon in the circulating fluid is a less effective poison than that in the graphite because about two-thirds of the fluid is outside the core at any instant. The plotted points represent the observed 13%Xe reactivity transient at the beginning of run 7 (July 1, 1966). The data indicate that the low apparent xenon poisoning can be explained by a large void fraction (between 0.5 and 1.0 \;01 %) and a low bubble stripping efficiency. Note, however, that the transient buildup is not closely fitted by these parameter values. In Fig. 5.4 the curves indicate the calculated effect of increasing the bubble stripping efficiency for a fixed, relatively small (0.1 vol %) circulating void fraction. The plotted points are for the same reactor xenon transient shown in Fig. 5.3. A comparison of Figs. 88 5.3 and 5.4 'shows that the steady-state xenon poisoning is desc.ribed as well by a low void fraction with a high bubble stripping efficiency as it is by a high void fraction with a low. stripping efficiency. However, the shape of the experimental transient poison buildup is de- scribed more closely by the parameter values in Fig. 5.4. . Figures 5.5 and 5.6 show the calculated and observed transient buildup of '*5Xe poisoning after a step increase in power from 0 to 5.7 Mw in run 8 (October 1966). The ranges of values of a, and €, used in these calculations are the same as those used in Figs. 5.3 and 5.4. Again, one finds that the shape of the observed transient is matched more closely by the calculations which assume a low void fraction and a high bubble stripping efficiency. Figure 5.7 shows the calculated and observed !3°Xe realctivity transients for a power re- duction from 5.7 Mw to 0, with the !35Xe initially at equilibrium. Only the experimental data for the early part of the transient are shown in this figure, since the reactokr was made sub- critical before the complete xenon transient could be recorded. However, the calculated curves reveal an important characteristic of the transient xenon behavior, which is due to variations in the overall xenon distribution that result from the choice of values for a, and €,. If the circulating void fraction is low, most of the poisoning effect is due to xenon in the graphite, and only a small amount of xenon is in the circulating fluid. Xenon that is produced in the fluid from iodine decay continues to migrate to the graphite for a period of time after the power (and, hence, the burnout rate in the neutron flux) has been reduced. This produces a shut- down peak in the xenon poisoning. Eventually, the stripping process reduces the xenon con- centration in the fluid so that some of the xenon in the graphite can escape and be stripped out. - This results in a more rapid decrease in xenon poisoning than simple radioactive decay.' As the circulating void fraction is increased, a larger fraction of the xenon inventory (or poisoning) is associated with the bubbles, and there is less xenon migration to the graphite. In this case -the shutdown peak tends to disappear. This characteristic makes the shutdown transients some- what more sensitive to the values of the bubble parameters and, potentially, more useful in the analysis of the' xenon behavior. ‘ | For the same experimental decay transient as that plotted in Fig. 5.7, Figs. 5.8 and 5.9 show the effect of increasing the bubble stripping efficiency, with the circulating void fraction held fixed at two representative low values (0.10 and 0.15 vol % tespectively). Although the section of experimental data for this transient is too limited to allow a valid comparison to be made, it is again seen that the high bubble stripping efficiencies and low circulating void fractions also provide reasonable representation of the observed data. The lack of any apparent shutdown peak in the experimental data, however, seems to suggest that a larger amount of gas may have been in circulation at the beginning of the shutdown (termination of run 8) than was apparent at the beginning of the run (Fig. 5.6). A second '33Xe stripping out decay transient, somewhat longer than the preceding, was ob- served during run 9 (November 1966), following reduction in the powér level from 7.4 Mw to 0. This transient is shown in Figs. 5.10 and 5.11, where the calculated curves are again based on the assumption of relatively high bubble stripping efficiencies and low circulating void fractions. 89 In this case a slight rise in the apparent poisoning following shutdown was indeed observed. From comparisons of the results given in these two figures, it appears that, under nomal op- -erating conditions, o, and €, might be bracketed between 0;1 and 0.15 vol %, and 50 to 100% ~ respectively. ( ‘ Finally, in Figs. 5.12 and 5.13, we show the most recent shutdown transient obtained at the termination of run 10 (Jan. 14, 1967). In this case the apparent !3°Xe reactivity transient was recorded for more than 40 hr after the reduction in power level. The results are also in good accord with the conclusions indicated above. Although substantial progress has been made in interpreting the xenon behavior in the MSRE, _the experimental data which have thus far been accumulated for the transient behavior of the 135¥e poisoning are as yet insufficient for any final conclusions to be drawn concerning the ‘“‘best’’ values of the circulating void fraction and bubble stripping efficiency. As one ex- ample, it should be noted that, if gas bubbles are continuously being ingested into the main circulating stream as the evidence indicates, the volume of gas in circulation is probably not constant, but rather is a élowly varying quantity depending on the level of the liquid in the fuel-pump tank and th’e transfer rate of salt to the overflow tank. This dependence is as yet not well understood, and future operation is expected to shed further light in this area. Other refinements of the model for the xenon behavior may b_e required as reactor operating data are accumulated. Theserefinements are not expected to strongly affect the conclusions indicated in the preceding description. CRNL-DWG 67-1077 4 ’ ap, VOLUME PERCENT o e EXPERIMENTAL DATA, OBTAINED CIRCULAEING ' DURING MSRE R™™N NO.7 / © 1.0 0.8 C.05 iy g 7 /’ / C.4 REACTIVITY MAGNITUDE (% 5k/4) 0.2 4 28 32 36 40 TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr) Fig. 5.3. Effect of Yolume of Circulating Gas on Transient Buildup of 135%¢ Reactivity. Step increase in power level from 0 to 7.2 Mw; bubble stripping ef- ficiency, 10%. REACTIVITY MAGNITUDE (% 34/k) REACTIVITY MAGNITUDE (%Sk/k) 90 ORNL-DWG 67-1078 0.7 - T T I l l I I &, , BUBBLE STRIPPING ® EXPERIMENTAL DATA, OBTAINED EFFICIENCY (%) 0.6 R .7 DURING MSRE RUN NO P 0.5 — I / /_____é-é-- 0.4 // /,__/ /] / 50 0.3 S / —— | "] X ' L] ®s0 (0 ® 0.2 ) e . & ———T"00 o _— / A / /.‘./ 0.1 74 /-/ / 28 o o L&&° . 0 4 8 12 16 20 24 28 32 36 40 TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr) Fig. 5.4, Effect of Bubble Stripping Efficiency on Transient Buildup of 135y, Reactivity, Step increase in power level from 0 to 7.2 Mw; vol % circu- lating bubbles, 0.10. ORNL-DWG 67—1079 0.7 0.6 [ & EXPERIMENTAL DATA, OBTAINED : DURING MSRE RUN NO.8 ap, . VOLUME PERCENT . CIRCULATING BUBBLES | oot -—-‘——_- 0.5 = 040" 0.4 / o3 / L oosT ’ / /, / — | ~0.50 (0] 2 / _‘.—-—'—'___- ‘s o [ ] d Py . . / P /’ o ] '{OO ///__-o-.-'l—’-.-.-‘ —— /i . (oK - / //T-“ P N °* e * ¢ 0 0 4q 8 12 16 20 24 28 32 36 40 TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr) Fig. 5.5. Effect of Yolume of Circulating Gas on Transient Buildup of 135x¢ Reactivity. Step increase in power level from 0 to 5.7 Mw; bubble stripping effi- ciency, 10%. REACTIVITY MAGN&TUDE (T 84/k) ( o REACTIVITY MAGNITUDE (% S4/4) 91 ORNL —-DWG 67—-1080 07 1 0.6 ¢,» BUBBLE STRIPPING ——— EFFICIENCY (%) e EXPERIMENTAL DATA, OBTAINED 1o =" 0.5 DURING MSRE RUN NO.8 P ot 7 20 / /——_ / / C.3 - : Iy . . . . o 0.2 / 2 ¥ 23 —100 / ] A—r.-.-fa"'fi"“ 0.1 / / e L 2 . / / . -.‘ib ° ¢ fi’-.‘. O . 0 4 8 12 16 20 24 28 32 36 40 TIME AFTER INCREASE IN REACTOR POWER LEVEL (hr} Fig. 5.6. Effect of Bubble Stripping Efficiency on Transient Buildup of 135 Xe Reactivity. Step increase in power level from 0 to 5.7 Mw; vol % circu- lating bubbles, 0.10. ORNL-DWG 67-1081 ' EXPERIMENTAL. DATA, OBSERVED DURING MSRE RUN NO. 8 0.5 | . AN a,, VOLUME PERCENT — \ \cmcuu\flNG BUBBLES '-——.'-.-0-.* \ | P00 - 0.0 0.2 . \-& \ . . ‘ | \ \N o5 \ . \ R 7 EO.SO 0 0 4 -8 12 16 20 24 28 - 32 36 40 TIME AFTER DECREASE IN REACTOR POWER LEVEL ({hr) Fig. 5.7. Effect of Volume of Circulating Gas on Transient Decay of 135xe Reactivity. Step decrease in power level from 5.7 Mw to 0; bubble stripping effi- ciency, 10%. ORNL—DWG 67-—14082 0.7 e EXPERIMENTAL DATA, OBTAINED DURING MSRE RUN NO. 8 \ ) o 3 | \\ \\ \ ) / BUBBLE STRIPPING REACTIVITY MAGNITUDE (%o84/k) [®] D Eb' ¢ eolee \ EFFICIENCY (%) ——| \ 02 T 0. >\\ \\\1o~.. l\\\ \20 \ g 0.1 ; \--.._\\ 50 |~ S 0 : : 0 4 8 12 16 20 24 28 32 36 40 TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr) Fig. 5.8. Effect of Bubble Stripping Efficiency on Transient Decay of 135ye Reactivity. Step decrease in power level from 5.7 Mw to 0; vol % circulating bub- bles, 0.10. A ORNL-DWG 67 -1083 06 . 0.5 / AT g e EXPERIMENTAL DATA, OBTAINED < DURING MSRE RUN NO.8 2 04 ™ & N Z 03 ~ ‘ 2 \ . = ? ®ececoe \ . > '“_‘:JLLA. \ \ &, BUBBLE STRIPPING = 0.2 L . EFFICIENCY (%) 5 —Z. L = P ———— ‘g L \\"‘\ \ : 10\ I&J o \-‘_\\ 20\\ ] 150(2)____----____ o | 0 a 8 §2 16 20 24 28 32 36 . 40 TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr) Fig. 5.9. Effect of Bubble Stripping Efficiency on Transient Decay of 135y, Reactivity. Step decrease in power level from 5.7 Mw to 0; vol % circulating bub- bles, 0.15. ORNL-DWG 67-1084 1 1 1 Y iy ® EXPERIMENTAL DATA, OBTAINED DURING / \\ MSRE RUN NO. 9 0.6 \\ I 7 B AN AN \ Y -\\ \ ~. ~ P T T~ ' €,, BUBBLE STRIPPING REACTIVITY MAGNITUDE (% 84/4) ™ N AN N ™~ N S~ ~J \ .3 — EFFICENCY | ) e [ ] o o 4 A 0.2 . = ® =~ \10\ P ' 20 ® \ \\ \\ 0 0 q 8 12 {6 20 24 28 32 36 40 TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr) Fig. 5.10. Effect of Bubble Stripping Efficiency on Transient Decay of 135)(e Reactivity. Step decrease in power level from 7.4 Mw to 0; vol % circu- lating bubbles, 0.10. ORNL DWG 67-1085 07 - 06 /’—-“\ _ - ~g N \ . - 05 N e EXPERIMENTAL DATA, OBTAINED 3 \ DURING MSRE RUN NO. 9 (71 ] 2 ~ ~ T , § \ \ _ = 03 : \\ ¢, sBUBBLE STRIPPING > LB SR S ol N : \ EFFICIENCY (%) — .";-._ \ N o S e o \\ 0.1 - \\___ 100 T 0 0 4 8 12 16 20 24 28 32 36 40 TIME AFTER D.ECREASE IN REACTOR POWER LEVEL (hr) Fig. 5.11. Effect of Bubble Stripping Efficiency on Transient Decay of 135)(e Reactivity. Step Decrease in power level from 7.4 Mw to 0; vol % circu- lating bubbles, 0.15. ORNL-DWG 67-1086 0.8 / 0.7 ™S .. / \ o EXPERIMENTAL DATA,OBTAINED \ DURING MSRE RUN NO. {0 x 0.6 3 N\ > ) \ w 0.5 | =~ Q \ jum = 4 N é ’ \ N fi - T~ N | ¢,» BUBBLE STRIPPING = . \ . EFFICIENCY (070) - [ L Y \ 10 Q * ] . » [ J P \ \ I?J ™ |\ i \ x 0.2 L B ™~ 20 ’ AR \\ N * “.\ 50 \\ ® ¢ T~100 o L M . .-..,__'___ — \ ' M e ® 0 o 4 8 12 16 20 24 28 32 36 40 135 TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr)” " Fig. 5.12. Effect of Bubble Stripping Efficiency on Transient Decay of | Xe Reactivity. Step decrease in power level from 7.4 Mw to 0; vo!l % circu- lating bubbles, 0.10. ORNL-DWG- 67-1087 0.7 < 06 A ® EXPERIMENTAL DATA ,OBTAINED — | = / \ DURING MSRE RUN NO. 10 @ , £ os AN W \ = 5 04 //__‘.\'"“‘ \ 2 ' \ N . > \ V .BUBBLE STRIPPING S 03 N EFFICIENCY (%) 5 IS i Bl T \ < e L 2 - \ N0 T 02 S e . / '*....‘ \ ~ \ [ ) Iy \ (o} ) !:n,..__‘:}(g L 100 e oL | 0 4 8 12 16 20 - 24 28 32 36 40 TIME AFTER DECREASE IN REACTOR POWER LEVEL (hr) Fig. 5.13. Effect of Bubble Stripping Efficiency on Transient Decay of ]35Xe Reactivity. Step decrease in power leve! from 7.4 Mw to 0; vol % circu- lating bubbles, 0.15. Part 2. Materials Studies 6. Molten-Salt Reactor Program Materials 6.1 MSRE SURVEILLANCE PROGRAM — HASTELLOY N W. H. Cook H. E. McCoy Several stringers of Hastelloy N test specimens were located in an axial position near the center line of the MSRE for sur;/eillance purposes. These specimens were from heats 5085 (cy‘lindrical v_esrsel) and 5081 (miscellaneous parts)., They were removed after 4800 hr at 645 + 10°C, during which the reactor had operated 7612 Mwhr. The peak thermal dose was 1.3 x 10%° nvt, and the peak fast dose (>1.2 Mev) was 3 x 10'° nvt,! The peak-to-minimum thermal flux over the length of the test specimen array was a factor of 5, but this was not found to be a signif- icant variable in this evaluation. The details of the removal of the specimens have been de- scribed previously,? but we want to mention again that there was some bowing at various points along the surveillance assembly. We used an optical comparator to eliminate those specimens bowed by more than 0.001 in. over the gage length. The surveillance control specimens were exposed to MSRE-type fuel salt and duplicated the thermal history of the in-reactor specimens. They were also bowed, and the ones bowed By more than 0,001 in, were not used. ‘ Samples of both reactor and control specimens were metallographically examined. While the réactor samples had structures that were dirtier than the controls, no major .changes in structure were found. No evidence of attack or deposition was found within the gage lefigths' of'any of the samples. However, all test specimens from heat 5085 (both irradiated and control) were found to have a surface layer on the 14 -in.-diam portion of the samples on the sides which were in con- tact or near contact with the graphite. Such layers were not found on the smaller l/s-in.-diar'n. gage length of the specimens. Figure 6.1 shows the nature of the surface layer on these samples. 1 private communication with H. B. Piper. . 2MSR Program Semiann. Progr. Rept, Aug. 31, 1966, ORNL-4037, p. 97. — 95 = = V.00 INLHES T5500% Fig. 6. loy N in near contact with graphite. Hastelloy N Surface from Exposed MSRE Surveillance Samples. Surface deposit from Hastel- Note also the extensive carbide precipitation that has occurred along the grain and twin bound- aries during the exposure. A similar surface reaction layer was noted on samples from heat 5081 but did not occur as frequently. The 20-mil-thick Hastelloy N straps (heat 5055) that bound the graphite specimens also had such a reaction product on the sides that contacted the graphite. These straps showed limited ductility when bent at room temperature. They tended to break intergranularly in the reaction region when bent sharply. Although we have not made a positive identification of the surface layer, we assume that it is a carbide produced by a reaction be- tween Hastelloy N and graphite. Both tensile and creep-rupture tests have been conducted on the surveillance specimens. The creep-rupture testing is not yet complete, so only the results of the tensile tests will be reported. ! is shown as a The total elongation at fracture when deformed at a strain rate of 0.05 min™ function of temperature in Fig. 6.2. Both heats show some reduction in ductility in the irradiated condition. Heat 5085 exhibited a slight reduction in ductility when tested unirradiated at room temperature, while the irradiated specimen of heat 5081 exhibited the same effect. With the ex- L\ 97 ception of these values, the ductilities remained essentially constant up to temperatures of 500°C. At temperatures above 500°C, the ductility of the irradiated and the control material decreases with increasing temperature, with the irfadiatéd material showing a greater loss in ductility. At temperatures above 650 to 700°C, the control material exhibits improved ductility, whereas the ductility of the irradiated material continues to decrease. Figure 6.3 compares the properties of the irradiated and control specimens at a lower strain rate, 0.002 min~—'. Qualitatively, the behavior is very similar to that noted at a strain rate of 0.05 min~!. However, the loss in ductility is magnified at the lower strain rate. The variations in the properties of the two heats of material have been reduced to where the differences are minimal. ) ] \ Figures 6.4 and 6.5 show how the ratio of the irradiated to the unirradiated tensile property varies for heats 5081 and 5085, respectively, as a function of temperature. For both heats, the yield strength is unchanged by irradiation. The ultimate tensile stress is reduced about 8% up to a temperature of about 500°C, where the reduction becomes considerably greater. The reduc- tion in ductility at room temperature and at elevated temperatures is also clearly demonstrated. We have compared the irradiated and control surveillance specimens. Let us now look at how the properties of the control specimens have changed during their 4800 hr of thermal exposure to molten salt. Table 6.1 shows representative properties of heat 5085 after several different heat treatments. The first group of tests was run with the material in the as-received condition (mill annealed 1 hr at 1177°C). Annealing for 600 hr at 650°C had no appreciable effects on the properties. The MSRE surveillance specimens were given a 2-hr anneal at 900°C before insertion ORNL-DWG 67-2452A 70 60 w o \|. > 9 A PSS = ‘ \\‘\ / [l ‘_,:/ \ \ ] / z - . 4 E 40 =I/ \ y Vi £ 40 8 —— AT % —_— . . \ e o [ ] \ Al\ ‘// d \ z|\\ - 30 N \ 7 . VAN Z CONTROL . IRRADIATED N o N o 8 o 5081 N\ s o 5085 Y 1 ¢=0.05 min™! \\“ ~ 10 \\\,,\ 0 0 100 200 300 400 500 600 70O 800 900 000 TEST TEMPERATURE (°C) Fig. 6.2. Comparative Tensile Ductilities of MSRE Surveillance Specimens and Their Controls ot a Strain Rate of 0.05 min-l. 98 ORNL-DWG 67-2453A 50 P, - > Y o / \ o \ "1" \\ , / . CONTROL IRRADIATED \\ // l 1Y A s o 508 N = ' A e 5085 \ * _.-/ &= 0002 min' , ‘\’\M - N (@] TOTAL ELONGATION (%) \ w o) N . [ 10 AN \;\g\____;o_ 0 0 {00 200 300 400 500 e00 700 800 ‘ 900 100C TEST TEMPERATURE (°C) Fig. 6.3. tomparative Tensile Ductilities of MSRE Surveillance Specimens and Their Controls at a Strain Rate of 0.002 min-I. ORNL-DWG 67-2458A 1.2 1 O s s m—_fy mm & [ ] —T T AT e . S S~gT —4, > I>—- e - NA Sa ElE 0.8 g \‘\ b, &S ' \ ||\" --..__: ola Ve la ] \ A o ’ s \ o olu 0.6 [] == : alz \ al2 e YIELD STRESS \ EE oa A ULTIMATE STRESS M > ; w TOTAL ELONGATION - \ &=0.05 min | N 0.2 :\ N\ ~ ~N - O S 0 100 200 300 400 500 600 700 BOO 900 1000 TEST TEMPERATURE (°C) Fig. 6.4. Comparative Tensile Properties of lrradiated and Unirradiated MSRE Surveillance Specimens, Heat 5081. into the reactor, and the properties of the material in this state are also indicated in Table 6.1. The properties at 25°C were unchanged, and slight reductions in tensile strength and ductility were noted at a test temperature of 650°C. After exposure to salt for 4800 hr at 650°C, there is a further reduction of 10% in the tensile strength and the ductility’. Table 6.2 shows that heat 5081 underwent similar changes at elevated temperatures but did not suffer the reduction in ductility at 25°C noted for heat 5085. ] ‘. 99 ORNL—DWG €7-2459A 1.2 -4— 0.8 = —y=———] - iy NP B 06 IRRADIATED PROPERTY UNIRRADIATED PROPERTY 04 e YIELD STRESS 4 ULTIMATE STRESS B TOTAL ELONGATION é=0.05 min ' 0.2 0 100 200 300 400 500 600 700 TEST TEMPERATURE (°C) 800 900 1000 Fig. 6.5. Comparative Tensile Properties of Irradiated and Unirradiated MSRE Surveillance Specimens, Heat 5085, Table 6.1. Tensile Properties of Hastelloy N — Heat 5085 Test Strain Yield Ultimate Uniform Total Reduction Specimen Heat Treatment Temp Rate Stress Stress Elongation Elongation in Area No. °C) (min~!) (psi) (psi) (%) @) %) .76 As received 25 0.05 52,200 116,400 51.3 52.5 ' 56.3 77 As received 427 0.05 30,700 102,900 57.6 59.4 49.9 284 As received 600 0.02 32,200 85,100 46.6 47.6 40.4 78 As received 650 0.05 28,700 80,700 35.5 36.7 33.2 285 As received 650 0.02 31,900 65,000 25.8 31.8 27.2 283 As received 650 0.002 30,500 64,300 22.6 24.1 28.8 79 As received 760 0.05 32,100 61,500 24.7 27.0 31.9 80 As received 871 0.05 30,700, 42,300 9.0 31.8 33.4 81 As received 882 0.05 23,100 23,100 1.8 40,2 43.9 1339 A.R. + 600 hr at 650°C 25 0.05 41,100 115,700 52.5 52.8 41.7 . 1340 A.R. + 600 hr at 650°C 650 0.05 33,800 76,000 36.8 37.5 36.4 4295 A.R. + 2 hr at 900°C 25 0.05 41,500 120,800 52.3 53.1 42.2 4298 A.R. + 2 hr at 900°C 500 0.05 32,600 94,800 51.2 54.1 40,9 4299 A.R. + 2 hr at 900°C 500 0.002 33,500 100,200 52.0 53.3 41.7 4296 A.R. + 2 hr at 900°C 650 0.05 29,600 75,800 31.7 33.7 34.6 FC-3 , 25 0.05 45,500 111,200 46.8 . 46.8 31.5 DC-19 A.R. + 2 hr at 900°C + 500 0.05 33,600 94,300 48.8 . 49.3 41.5 DC-26 4800 hr in 500 0.002 33,400 91,700 43.3 44.3 37.8 DC-14 MSRE salt at 650°C 650 0.05 31,800 70,100 25.8 26.8 30.0 DC-25 650 0.002 31,500 62,500 22.8 24.3 27.2 100 Table 6.2, Tensile Properties of Hastelloy N — Heat 5081 - . Test Strain Yield Ultimate Uniform Total "Reduction Spe:;men Heat Treatment Temp Rate Stress Stress Elongation Elongation in Area o (Cc) (min~!) (psi) (psi) (%) (%) (%) 4300 A.R. + 2 hr at 900°C 25 0.05 52,600 125,300 56.7 59.5 50.5 4303 A.R. +2 hrat 900°C - 500 0.05 32,000 100,300 57.8 60.7 44.4 4301 A.R. + 2 hr at 900°C 650 0.05 32,200 81,800 31.7 33.9 29.9 4302 A.R. + 2 hr at 900°C 650 0.002 32,900 74,900 29.0 29.5 31.8 AC-8 25 0.05 47,700 118,700 55.9 57.6 48.8 AC-19 A.R. + 2 hr at 900°C 500 0.05 35,800 97,800 53.6 56.6 46.2 BC-9 + 4800 hr in 500 0.002 36,200 95,300 46.2 47.0 38.1 AC-27 |MSRE saltat 650°C | 650 0.05 32,400 68,400 23.8 24.6 23.1 AC-17 - 650 0.002 33,600 66,700 22.8 23.2 21.6 Some of the test specimens have been e};amined metallographically. Figure 6.6 shows the microstructure of control specimen No. AC-8 from heat 5081, which was tested at 25°C. This specimen exhibited good ductility, and this is reflected in the intragranular, shear-type fracture and the lack of intergranular cracking. Figure 6.7 shows the irradiated specimen from the same heat that was tested at 25°C. The fracture is largely intergranular, and there are numerous intergrarular cracks. The control and irradiatéd specimens from heat 5085 tested at 25°C ex- hibited the same geheral characteristics as those illustrated in Fig. 6.7. The other specimens examined were tested at 650°C, The failures were all intergranular, with the irradiated speci- mens exhibiting the typical characteristics of a high-temperature, low-ductility, intergranular fracture. ' | We compared the ductilities of the surveillance specimens with those for specimens irradiated in other -experiments without salt present. Heat 5081 had been irradiated previously in the ORR. The ORR experiment was run at 700°C to a thermal dose of 9 x 10*° nvt, and the matetial was in the as-received condition. The MSRE surveillar-lcer specimens weré run at 650°C to a thermal dose of 1.3 x 1029 nvt, and the preirradiation anneal was different. However, none of these dif- ferences are thought to be particularly significant, and the results compare rather well. Figure 6.8 shows that the postirradiation ductilities of heat 5081 after both experiments are very similar. The most important question to be answered concerning these data is how they apply to the operation of the MSRE. The surveillance specimens were exposed to a thermal dose of 1.3 x 10?°% nvt. (The MSRE vessel will reach this dose after about 150,000 Mwhr of operation.) This burned out about 30% of the '°B and produced a helium content of about 10~ 3 atom fraction in both heats. The high- temperature tensile properties are exactly what we would expect for this dose. Our in 100X 0.0 Fig. 6.6. Microstructure of Hastelloy N (Heat 5081) Exposed to Salt for 4800 hr at 650°C. Tested at 25°C at a strain rate of 0.05 min™ | (specimen No. AC-8). Etchant: glyceria regia. (a) Fracture. (b) Edge about Y in. from fracture. 102 0.035 INCHES > 100X o T 100X T Fig. 6.7. Microstructure of Hastelloy N (Heat 5081) Exposed for 4800 hr in the MSRE at 650°C. Tested at 25°C ot a strain rate of 0.05 min™! (specimen No. D-16). Etchant: glyceria regia. (a) Fracture. (b) Edge about Y in. from fracture. L) 103 ORNL—DOWG 67—4779 1.0 [ ] <>\ 0.8 l o ¢ = 0.002 min~" g &2 5 = N\ : O THIS STUDY, b, =1.3X10%° nvt, T=650°C 2 06 ® ORNL —TM ~ (005, ¢ _=9x102% avf, —] r 7= 700°C w - (= 5 = ] ol® W= ok Lo B8 alz o|d sle x|z ® s } — 0.2 \ \ | ‘N\” -0 400 - 500 600 700 800 900 ' {000 T TEST. TEMPERATURE (°C) Fig. 6.8. Comparative Effects of Irradiation in MSRE and ORR on the Duc. tility of Hastelloy N, Heat 5081, lives).* We are reasonably confident that our tests on the .surveillance specimens will indicate a similar behavior. Our work also seems to indicate a saturation in the degree of radiation damage at a helium atom fraction of about 107, and we feel that the properties of the material will not deteriorate further. The low-temperature ductility reduction was not expected. It is probably a result of grain-boundary precipitates forming due to the long thermal exposure. Ir- radiation play;s some role in this process that is yet undefined. The low-temperature properties are not ‘‘brittle’’ by any standards but will be monitored closely when future sets of surveil- lance specimens are removed. 6.2 'MECHANlCA_L PROPERTIES OF HASTELLOY N H. E. McCoy | - Since the surveillance assembly holder was deformed, it was not possible to remove only part of the specimens and return the remainder to the‘MSRE; therefore, all 162 specimens were removed. The number of specimens required for actual surveillance purposes was rather small, and we used the others for learning‘ more about the general radiation damage éharacteristics of Hastelloy N. These data are included along with other tests in the following discussion. ‘*MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL.-4037, pp. 118 and 121. 104 Table 6.3 shows the postirradiation tensile properties of several heats of Hastelloy N. Heat 5085 shows pr0pérties that are very similar to those observed for the surveillance specimens. Using a test condition of 650°C and a strain rate of 0.002 min~ ', heat 5065 has a total elonga- tion of 10 to 14% if irradiated cold and about 6% if irradiated hot. This is slightly less than observed for heats 5085 and 5067. Note that heat 5085 is the only heat irradiated at an elevated temperature that shows a significant reduction in ductility at 25°C. Several of the si)ecimens from heat 5081 were tested at various strain rates and temperatures to determine how the properties varied with these parameters. The variation of the total elonga- tion with strain rate for the control specimens is shown in Fig. 6-9; The ductilities at 400, 500, 760, and 850°C are reasonably independent of strain rate. At 650°C, the ductility is very dependent on strain rate down to rates of about 0.05 min~!, Figure 6.10 shows that the ductility of the irradiated material is only slightly strain-rate-dependent at 400°C. At 500°C (Fig. 6.11), the ductility decreases markedly with decreasing strain rates below values of about 0.05 min~1, This probably corresponds to the transition from transgranular to intergranular fracture. At 650°C (Fig. 6.12), the ductility depends very heavily upon strain rate, showing a rapid decrease at rates of 2 min~! and not having reached a constant value at a strain rate of 0.002 min~—!. At 760°C (Fig. 6.13), the ductility is a strong function of strain rate down to rates of 0.005-min~1!, where the dependence on strain rate becomes much less. At 850°C (Fig. 6.14), the ductility is strongly dependent on strain rate down to rates below 0.05 min™!, where it beco’mes fairly con- stant. This family of curves will allow one to predict the strain-rate sensitivity of the ductility of Hastelloy N at a given temperature. ORNL DWG 67-2455A 70 60 - __a300°Ce 3 A s A < ~500°C ! ! : 2 50 yam po =T — N S pra §850°C, : V P ,,” / 9 I _L I z"f :: 40 (— 0760°Co : [e) ,i ) e % ' : // o >0 P . LA . 1 w650°Ca | 8 HEAT 5081 = . INERl o 20 L TEST TEMPERATURE e 400°C ®650°C 10 A 500°C © 760°C v 850°C i U 11 103 2 5 402 2 5 0! 2 5 10° 2 5 10" STRAIN RATE (min~ ") Fig. 6.9. Influence of Strain Rate on the Ductility of MSRE Surveillance Control Specimens, Heat 5081. Table 6.3. Postirrodiation Tensile Properties of Several Heats of MSRE Hastelloy N After Irradiation in a Helium Environment Boron Irradiation nvt, Test Yield Ultimate Uniform Total Reduction Heat Condition Level Experiment Temp Thermal Temp E Stress Stress Elongation Elongation in Area - Specimen No. (ppm) No. ©c) Dose (°C) (psi) (psi) %) %) ) No.. x 1020 5065 A.R. 20 ORR-149 43 8.5 - 25 0.05 102,900 135,100 31.5 35.5 52.0 570 5065 AR. 20, ORR-149 43 8.5 200 0.05 82,300 119,600 36.1 39.2 54.8 571 5065 A.R 20 ORR-149 43 8.5 650 0.05 44,600 76,900 26.3 27.3 24.5 572 5065 A.R. 20 ORR-149 43 8.5 650 0.002 40,800 57,400 12.2 13.1 21.9 574 5065 A.R. 20 ORR-149 43 8.5 871 0.05 33,500 36,400 1.7 1.8 3.25 573 5065 A.R. 20 ORR-149 43 8.5 871 0.002 23,200 23,400 0.9 1.8 6.07 575 5065 20 ORR-149 43 8.5 25 0.05 101,500 132,200 30.0 33.6 55.2 581 5065 8 hr 20 ORR-149 43 8.5 200 0.05 78,600 118,300 37.3 39.8 45.9 582 5065 > at 20 ORR-149 43 8.5 650 0.05 35,100 77,700 25.3 25.8 22.6 583 5065 8710Cfi 20 ORR-149 43 8.5 650 0.002 38,100 66,200 13.7 14.3 17.3 585 5065 20 ORR-149 43 8.5 871 0.05 37,100 38,500 1.7 1.9 2.94 584 5065 20 ORR-149 43 8.5 871 0.002 25,900 25,900 0.8 1.6 5.26 586 5067 A.R. 20 ORR-155 500-700 1.4 25 0.05 59,000 123,700 .49.4 51.2 45.4 2289 5067 A.R. 20 . ORR-155 500-700 1.4 650 0.05 38,400 66,800 14.0 14,0 20.5 2290 5067 A.R. 20 ORR-155 500-700 1.4 650 0.002 36,400 59,900 9.6 9.6 16.0 2201 -5085 A.RI. 38 ORR-}SS - 500-700 - 1.4 25 0.05 46,300 110,600 41.1 41.2 40.3 2285 5085 A.R. 38 ORR-155 500-700 1.4 650 0.05 30,800 64,400 18.0 21.2 19.1 2286 5085 AR. 38 ORR-155 500-700 1.4 650 0,002 30,200 51,800 9.3 9.9 15.4 2287 5065 A.R. 20 ORR-155 500-700 1.4 25 0.05 50,100 117,000 54.4 56.1 53.2 1857 5065 A.R. 20 ORR-155" 500-700 1.4 . 650 0.05 37,400 62,100 12.3 12-.4 17.2 1858 5065 A.R. 20 ORR-155 500-~-700 1.4 650 0.002 34,800 49;1 00 6.4 6.5 15.3 1859 5065 A.R. 20 ETR-41-31 600 £ 100 3.5 550 0.002 . 49,100 68,600 9.1 9.4 1273 5065 A.R. 20 ETR-41-31 600 100 3.5 600 0.002 42,200 56,300 8.2 8.5 1276 5065 AR. 20 ETR-41-31 600 £ 100 3.5 650 0.05 41,200 59,000 10.8 11.3 1270 5065 A.R. 20 ETR-41-31 .600+ 100 3.5 650 0.002 42,600 51,800 5.9 6.1 1271 5065 A.R 20 ETR-41-31 600 £ 100 3.5 760 0.002 41,900 46,000 2.8 2.8 1274 5065 A.R 20 ETR-41-30 <150 5 650 0.05 46,700 76,200 21.6 22.2 28.8 383 5065 A.R. 20 ETR-41-30 <150 5 650 0.002 37,700 53,300 9.3 10.0 10.1 380 5065 A.R. 20 ETR-41-30 <150 5 650 0.002 40,000 57,500 il.4 11.6 17.5 384 SOT1 106 ORNL—DWG 67— 2449A 60 50 L * o | 1| b — 7] . 40 [ ] | ] - | | & ! (] | s =z [ ] Z 30 = » HEAT 508 400° C 20 4 UNIFORM ELONGATION e TOTAL ELONGATION o = REDUCTION IN AREA O ‘ 10° 2 5 0?2 2 5 10! 2 5 1° 2 5 1o STRAIN (%) STRAIN RATE (min ') Fig. 6.10. Influence of Strain Rate on the Ducfi'lity of MSRE Surveillance Specimens at 400°C, Line based on total elongation values. ORNL-DWG 67— 2504A 60 o o \ o] lo] A [T AY N (@) HEAT 5081 500°C e TOTAL ELONGATION 10 . 4 UNIFORM ELONGATION ’ REDUCTION IN AREA a 0 . L 2 2 5 o' 2 5 1o° STRAIN RATE (min ") 2 5 10’ Fig. 6.11. Influence of Strain Rate on the Ductility of MSRE Surveillance Specimens at 500°C. Line based on total elongation values. 107 ORNL-DWG 67-2450A 24 | ] / | 4 A . y , 20 A A '] / 9 e L A . -~ | 1"’fl" = // = 2 E 12 [ ] [ n 1;,/"' A ’ pE “/ . g =" HEAT 5081 650°C A UNIFORM ELONGATION . e TOTAL ELONGATION 4 n REDUCTION IN AREA 0 . ' — ‘ 03 2. 5 402 2 5 410! 2 5 10° 2 5 10! STRAIN. RATE (min™") Fig. 6.12. Influence of Strain Rate on the Ductility of MSRE Surveillance Specimens at 650°C, Line based on total elongation values. ORNL-DWG 67-2451A 24 ’ [ ] 20 HEAT 5081 760°C / A 4 UNIFORM ELONGATION ‘ // ® TOTAL ELONGATION ' /) £ 187 » REDUCT!ON IN AREA p = ] z 1 a ] o o : / 8 affil e T s 1 N ST 1 ' [ ] [y 0 -2 —4 0 | 103 2 5 1072 2 5 10 2 5 10 2 5 10 STRAIN RATE {min~!) Fig. 6.13. Influence of Strain Rate on the Ductility of MSRE Surveillance Specimens at 760°C. Line based on total elongation values. 108 ORNL DWG 67-2500A 12 1 {0 HEAT 5081 850°C 9 4 UNIFORM ELONGATION ;: . 8 ¢ TOTAL ELONGATION — / § 7 ® REDUCTION IN AREA 7 z )i g © / E 5 |// «n s i 4 [ ] i 3 3 gl | eTT] 2 a A a A i ? :l ] ,_ o . 103 2 5 102 2 5 40 2 5 40 STRAIN RATE (min™%) Fig. 6.14. . Influence of Strain Rate on the Ductility of MSRE Surveillance Specimens at 850°C. Line based on total elongation values. 6.3 PRECURSORS OF MSBR GRAPHITE W. H. Cook Basically, the isotropic graphite sought for molten-salt breeder reactors (MSBR) should have these properties: Permeability to helium 1077 t6 1073 cm? /sec Pore entrance diameter ' None larger than 1 {4 Electrical resistivity <1000 microhms cm? cm™ ! Coefficient of thermal expansion Approx 4.5 X 1078 (DC)—1 Ash content ' <150 ppm Boron ‘ ' <1 ppm Bulk density >1.86 g/cm3 In addition, it should be a well-crystallized graphite without fillers such as lampblack or carbon black. Other parameters, such as mechanical properties, probably would follow satisfactorily if the preceding requirements were met. Our immediate needs are modest quantities (50 1b) for general evaluations and for the specific and important radiation damage studies. We have made initial examinations on seven experimental grades of graphite. None satisfy all the basic requirements listed above; however, one shows a good pore spectrum and four ap- pear to have potential. Two do not show promise, 'y 109 The Chemical Engineering Development Department of the Y-12 Plant® has made some ex- perimental MSBR-type graphite that has a microstructure as shown in Fig. 6.15. This has been fired to 3000°C and has not had any impregnations to decrease its pore sizes. Except for some minor flaws, the structure is unusually tight. This is also shown by the pore spectrum shown in Fig. 6.16, in which the major portion of the pore entrance diameters are less than 0.3 . The larger pore entrances shown may be the result of the small cracks and voids visible in Fig. 6.15. It was fabricated as a small piece, 1.6 in. diam x 0.8 in. long, and there are some problems in fabrication, purity, and high electrical resistivity, but it does look encouraging in these early stages of development. It is the only one of the seven different grades that was fabricated specifically for MSBR requirements. The other grades are materials that were originally developed for other purposes. As one might expect, the major problems with these are high gas permeabilities and pores that are too large. All these had pore entrance diameters as large as 3 y, whether they had uniform or only surface graphitic impregnations. Four of the grades, which had uniform impregnations through- Sy-12 Plant, operated by the Union Carbide Corporation for the Atomic Energy Commission, Post Office Box Y, Oak Ridge, Tenn. Y—-76831 = 0.023 INCHES T~ 100X Fig. 6.15. An Experimental Isotropic Graphi: As polished. 100x. Grade B-5-1, Showing Low Porosity and a Few Flaws. 110 ORNL-DWG 67-4780 BULK VOLUME PENETRATED (%) o 1.0 20.9 12.4 1A 9.60 8.10 6.80 5.60 4.30 3.15 2.10 0.92 0.8 0.7 0.6 0.50 0.40 0.30 0.20 0.100 0.090 0.080 0.070 0.060 0.050 0.040 0.030 " 0.020 0.018 0.014 PORE DIAMETER {u) Fig. 6.16. A Pore Spectrum Plot of an Experimental lsotropic Graphite, Grade B-5-1. out the structure, had sufficient properties to warrant further investigation. Two are being purchased in 50-1b quantities for irradiat\ioh studies and general evaluations: An interesting feature of six of the grades of graphite above that had pore entrance diameters as large as 3 pu was the_xt the impregnations appeared to decrease the accessible void volumes but did not lower the range of the pore entrance diameters. Decreasing the accessible void volume is helpful, but having the pore entrance diameters less than 1 pu is of greater importance. Both MSBR and impregnatibn requirements appear to dictate that the base stock must have the major amount of its accessible voids with pore en- trance diameters approximately 1 p. 6.4 GRAPHITE IRRADIATIONS C. R. Kennedy by Irradiation experiments to demonstrate the ability of graphite to sustain massive neutron ex- posures have been designed and are being fabricated or are in progress. Graphite of MSBR quality will be irradiated in the DFR, HFIR, EBR-II, and ORR. The major irradiations will be obtained from capsules placed in the HFIR, The specimens will be located in two rod assemb-lies, which will replace two of the californium production rod assemblies. All the 32 graphite ring specimens in the rod assembly are designed to be irradiated at 700 + 25°C. After a one-reactor- cycle experimént to verify the design temperature, the specimens will be removed, examined, and recycled alternately every six months until a total irradiation time of 21/2 years is obtained. At this time, the total irradiation exposure will range from 5 to 10 x 10?2 vt (E > 0.18 Mev). The specimens will be examined for dimensional stability, gas permeability, and mechanical in- tegrity. 111 Irradiations in the DFR are essentially backup experiments to the HFIR irradiations. These irradiations will be used primarily to obtain relative comparisons of experimental grades to more standard grades; The exposures in these irradiations will not exceed 5 x 102! avt (E > 0.18 Mev). Irradiations in the EBR-II will be restrained growth experiments to confirm the ability of the MSBR graphite to sustain plastic creep deformation under irradiation. Again, the exposure ob- tained will be much less than that obtained in HFIR, with a maximum of about 1022 nyt (E > 0.18 Mev). The maximum tensile strains obtainable from these experiments will be about 3%. The experimental prb,gram in the ORR consists of very closely controlled 'creep experiments to obtain quantitative creep coefficients. These data, although obtained at a very low exposure rate and thus low exposures, are essentially for a‘ stress analysis of the graphite bodies. These experiments will be very limited in scope in view of the base of exrstmg information available on the creep behavior of graphite. The current experrment has been constructed and mstalled in the ORR. 6.5 BRAZING OF GRAPHITE W. J. Werner Studies were continued to develop methods for brazing large graphite pipes to Hastelloy N. Our current activities consist of work in the following areas: (1) development of a corrosion- resistant alldy which will readily wet and flow on graphite but which does not suffer from the transmutation problem associated with gold-containing alloys and (2) devising techniques for the manufacture of graphite-to-Hastelloy-N assemblies of the size and configuration required for loop experiments, Large Graphite-to-Hastelloy-N Assemblies A vacuum or inert-atmosphere induction brazing furnace for brazing graphite-to-metal as- semblies ufi to 4 in. in diameter by 12 in. long has been designed and is currently under con- struetion. Several pieces of graphite, molybdenum, and Hastelloy N have been prepared for brazing, using the tapered-joint de51gr1 reported previously.® Figure 6.17 shows the size and configuration of the components. The graphrte is AT] grade due to the una@arlabrhty of MSRE- grade material in the desired size and configuration. 6.6 CORROSION RESISTANCE OF GRAPHITE-TO-METAL BRAZED JOINTS W. H. Cook - It was reported that the braze joining grade CGB graphite to molybdenum did not have any microscopically visible attack after a 5000-hr ex posure to LiF-BeF ,-ZiF 4-ThF 4-UF, salts at ®MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 140. 112 Y-78516 Fig. 6.17. As-Machined Components for Graphite-to-Molybdenum-to-Hastelloy-N Tapered Joint. 1300°F (705°C) contained in Hastelloy N.” The brazing alloy was 60 Pd—35 Ni-5 Cr (wt %). The chemical analyses of the test salt did not show any significant changes from the analyses of the as-received salt. The microstructure of the brazed joint is shown again in Fig. 6.18 for reference purposes. Electron-probe microanalyses show that molybdenum diffused into the brazing alloy, as was deduced by Jones and Werner from its microstructure.® There was some migration of nickel and palladium into the molybdenum, along with a slight amount of chromium. The speckled precipitates in the brazing alloy are primarily Mo, Ni, Cr, and C in descending quantities. The long, acicular crystals located toward the graphite sides of the brazed joint are primarily Mo, Cr, Ni, and C in descending quantities. These are believed to be essentially mixed carbides of molybdenum and chromium. The microstructures suggest that this is a well-formed joint that has not been appreciably altered by the corrosion exposure. A deposit is present on the surface of the braze metal. The deposit has been identified as pure palladium except for some nickel in the region adjacent to the brazing alloy. While it extends slightly over the graphite, it seems to be attached only to the braze metal. Figure 6.19 shows another view of the deposit with a rough surface and a spotty "MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 115—17. B1bid., p. 113. PHOTO 87447 PALLADIUM COATING BRAZING ALLOY LYBDENUM Fig. 6.18. Microstructures of the 60 Pd—35 Ni—~5 Cr (wt %) Brazing Alloy Used to Join Grade CGB Graphite to Molybdenum. (a) After 5000-hr ex- posure to molten fluoride salts at 1300°F (705°C). Etchant: 10% oxalic acid. 100x. (b) Enlarged microstructure of palladium coating located pri- marily on the brazing alloy. Etchant: oxalic acid. 500x. €11 ettt at MOLYBDENUM Fig. 6.19. External Appearance of Palladium Coating on Brazing Alloy (60 Pd—35 Ni~5 Cr, wt %) of the Grade CGB Graphite Brazed to Molybdenum After 5000-hr Exposure to Molten Fluoride Salts at 1300°F (705°C). (a) Elevation view. (b) High magnification (100x) of typical appearance. 1498 115 deposit on the molybdenum. No evidence of palladium was found in the fluoride salts or on the walls of the Hastelloy N container. While the palladium must have been leached from the braze metal and then redeposited, the mechanism for the transfer is not clear. The thickness of the coating appears to be dependent on the test time. Howe.verr,-since it was only 0.001 in. thick after the 5000-hr exposure to the molten fluoride salts at 1300°F (705°C) and seems to be present only on the braze, it seems valid to say that the braze was essentialiy unattacked for this period. An amorphous-appearing, metallic-like coating on the graphite® surfaces exposed to the molten salts was found to be Cr,C, by x-ray diffraction. Electron-probe microanalyses indicated that some vanadium is also present, There is no explanation for the presence of vanadium un- less it came from the salts, because the graphite and Hastelloy N normally contain vanadium in average quantities of 0.0009 and 0.5% respectively. 6.7 THERMAL CONVECTION LOOPS ° A. P. Litman We are continuing to study the compatibility of structural materials with fuels and coolants of interest to the Molten-Salt Reactor Program. Natural-circulation loops of the type described 10,11 ,re used as the standard test in these studies. previously Three loops are now in operation, and details of their service are shown in Table 6.4. The long-term loops, Nos. 1255 and 1258, fabricated from Haistelloy N and type 304 stainless steel, Ivid., p. 117. 10G. M. Adamson, Jr., et al., Interim Report on Corrosion by Zirconium Base Fluorides, ORNL-2338 (Jan. 3, 1961). ' 11 MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNL-3872, pp. 81-87. Table 6.4. Thermal Convection Loop Operation Through February 28, 1967 Maximum ‘ Loop Loop Material Hot-Leg Specimens Heat-Transfer Medium Temp . gT " Hours No. CF) ("F) _Operated 1255 Hastelloy N Hastelloy N + 2% Nb LiF-Ber-ZrF4-UF4- 1300 160 43,024 {permanent) ThF4 (70-23-5-1-1 ' mole %) 1258 Type 304 L stainless steel Type 304 L stainless LiF-Ber-ZrF4-UF4- 1250 180 31,749 steel (removable) ThF, (70-23-5-1-1 mole %) 10 Hastelloy N None Nal“-KF-BF3 - 1125 265 6,734 (48-3-49 mole %) : 9 Type 446 stainless-steel- None Lil"-Ber-ZrF4-UF4 1400 300 5,255% clad Nb—1% Zr ) (65-29.1-5-0.9 mole %) “Loop plugged on 10-1-66. 116 respectively, and circulating MSRE-type fuel salt plus approximately 1 mole % ThF, continue to operate without incident, Recently, we installed specimens in the hot leg of the stainless steel circuit so as to generaté additional data on that system. It is of interest to compare the compatibility of the salt which has now reached maturity in the loop with earlier results. To date, the specimen in the hottest portion of the loop, 1250°F, has experienced weight losses as shown in Table 6.5, 2 Interpolation of these results indicates a rate loss of approximately 3.6 mg cm™ month™ !, This is lower than the rate revealed in the previous test. Table 6.5. Effect of Time on Weight Change of Type 304 L Stainless Steel Specimen in Contact with LiF-BeF,-ZrF ;-UF ;-ThF (70-23-5-1-1 Mole %) at 1250°F , Time ‘ Weight Change (ht) , (mg/cmz) 25 -1.6. 115 -2.1 450 -2.95 1125 —4.4 Loop No. 10, fabricated from Hastelloy N and circulating a fluoroborate mixture, is now scheduled to operate for one year, after which time it will be dismantled and examined. It continues to circulate without difficulty. 6.8 EVALUATION OF MSRE RADIATOR TUBING CONTAMINATED WITH ALUMINUM D. A. Canonico D. M. Haseltine The failure of the aluminum blower blades at the MSRE site and the resultant damage were discussed in the last semiannual report.!? It was concluded that no damage was visually ob- served; however, the\possibility that some undetected aluminum still might be in intimate con- tact with the Hastelloy N tubes did exist. An experiment was conducted to determine the effect of a prolonged exposure to aluminum at 1200°F. Aluminum pieces from the blower blades and a Hastelloy N tube similar to those used in the heat exchanger were placed in intimate contact and held for times up to 1000 hr. The results of the 1000-hr exposure are shown inm Fig. 6.20a. For comparative purposes, the 5-hr exposure is shown in Fig. 6.205. It is evident that the penetration (approx 0.010 in.) is similar in both - 12MsR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 103-7. 117 Fig. 6.20. Metallographic Sections Through Aluminum Cone and Hastelloy N Tube. Samples were held ot 1200°F for (a) 1000 hr and (b) 5 hr. photomicrographs. The Hastelloy N tubing has a wall thickness of approximately % , in. The penetration after 5 hr was about 16%, and after the 1000-hr exposure it had not increased. The microstructure seen in the specimen held for 1000 hr is considerably different from that in the 5-hr sample. The extended exposure has allowed the various phases to grow and has re- sulted in a diffusion couple of aluminum-nickel complicated somewhat by the presence of zinc and other minor elements. This work supports the conclusions reported previously, that the radiator system is satis- factory for further operation. 7. Chemistry 7.1 CHEMISTRY OF THE MSRE Fuel Salt Composition and Purity R. E Thoma More than half the total power generated by the MSRE (21,464 Mwhr) was produced during the " . current report period. Heat balance.and nuclear calculations indicate that a total of 1.083 kg of 235y, 0.473% of the original uranium inventory, has now been consumed by fission. Table 7.1 . summarizes the results of fuel composition and purity analyses for the MSRE fuel in each of the power runs, including the three conducted in the last six months, Nos. 8 to 10, and the current run, No. 11. These data show that the uranium concentration of the fuel salt has decreased appreciably since power operation began. It should not be inferred, however, that burnup is evident in the results of chemical analysis for runs 4 to 10, for virtually all the decrease noted in Table 7.1 is the result of dilutions of the fuel salt by flush salt. Figure 7.1 shows a cfim_— parison of MSRE inventory values ! with the results of chemical analyses ? for runs 6 to 10. Analytical values shown here have been adjusted to compensate for changes in isotopic com- position of uranium and for those periodic variations in analytical bias as determined by the Analytical Quality Control Group.? - Step decreases in the inventory values reflect the dilution of the fuel salt by residues of flush salt remaifiing in the reactor fuel circuit after flushing operations are completed. Fuel salt is- also removed from the fuel circuit by sampling procedures and by transfer of fuel to flush salt. From the-results of analyses of flush salt specimens we have deducéd that each drain- flush-fill sequence results in a net transfer of 7.1 kg of uranium from the fuel to flush salt. The computation of this value has involved a number of assumptions, such as those concerned with the precision of uranium analyses in the 100-to-800-ppm concentration range, as well as the configuration and dimensions of MSRE fuel circuit components where salt residues may reside. Figure 7.1 also shows values which should have been obtained in the absence of dilution-transfer 1H. B. Piper, personal communication. 2Chemical analyses were performed under the supervision of C. E. Lamb, ORNL Analytical Chemistry Divisien. 3G. R. Wilson, ORNL Analytical Chemistry Division. 118 Table 7.1. Summary of MSRE Fuel Salt Analyses Number . U® Inventory \ ' - Run o Concentration (wt %) Values (wt %) Concentration (mole %) Concentration {(ppm) No. Samples 7Li Be Zr u® R TLiF BeF, ZrF, UF ] Fe Cr Ni 4 22 10.51 £0.137 6.55 T 0.161 11.14 £0.205 4.642 £0.028 4.622 4.622 63.36 £ 0.567 30.65 +0.583 5.15 £0.116 0.83 £t 0.011 131 t65 48 t7 40 L20 5 1 10.65 6.53 11.45 4.625 4,622 4.622 63.63 30.30 5.25 0.816 68 51 15 6 13 ©10.51 10,289 6.54 £0.200 11.31 £0.231 4.630 £0.027 4.622 4.619 63.35 £1.072 30.60 10.946 5.23 £0.145 0.825 10.013 111 T 44 50 +8 56 £24 7 11 10.55 £0.054 6.67 ¥0.174 11.34 £0.215 4.640 £0.017 4.619 4.614 63.04 £0.495 30.9510.580 5.20%0.118 0.819 £0.009 88132 4816 48 t16 4-7 . 47 10.523 £ 0.178 6.572 0.179 11.239 £0.271 4.638 £0.025 4.622 4.614 63.290£0.722 30.698 £0.696 5.188 £0.126 0.824 £0.011 114 55 49 +7 46 +21 8 8 11.78 £1.406 6.53 1 0.199 11.16 £0.193 4.632 1 0.011 4.601 4.599 65.84 +2.486 28.57 £2.123 4.82 £0.347 0.771 £0.058 122 45 64 27 61 £36 9 4 10.99 £ 0.099 6.63 £0.068 11.15210.370 4.603 £0.031 4.587 4.586 64.17 £0.006 30.04 £0.147 4.99 £ 0.185 0.794 £0.011 150 £17 61 £5 52 20 4-9 59 10.528 £ 0.159 6.570 £0.175 11.222 £0.266 4.635 £0.026 4.622 4.586 63.312 £0.679 30.685 £0.665 5.179 +0.125 0.824 £ 0.011 118 +52 51 9 48 T 24 10 10 11.14 £ 0.079 6.58 £0.188 11.0510.152 4.609 £0.020 4.575 4.569 64.65 £0.450 29.64 £0.480 4.92 1 0.064 0.791 £0.010 150 £30 6024 74 35 410 69 10.785 £ 0.641 6.571 £0.179 11.197 £0.260 4.631 £0.026 4.622 4.586 63.835 £1.342 30.258 £1.163 5.095 £ 0.212 0.811 $10.029 122 51 539 52 +27 8-10 22 11.345 £0.883 6.569 £0.173 11.106 £0.212 4.616 £0.022 4.601 4.569 64.998 £1.614 29.320 £1.402 4.898 £ 0.226 0.784 0,036 140136 6215 6533 2Corrected to compensate for isotopic composition. x.denotes beginning of run; y denotes end of run. 611 120 ORNL-DWG 67-4782 4.66 - AN 4.64 RUN 7 - RUN 6 . ' / — 1 RUN 8 L ‘ 4.62 i S TT - + RUN 9 "o 1 T/ T E 1 . : . P —— \-__ % L ! % 460 : RUN 10 3 S 2 . ) L~q 2 458 ! 5 e . T iy 456 CHEMICAL ANALYSES CORRECTED FOR ANALYTICAL BIAS — —— BOOK VALUE ADJUSTED FCR DILUTION BY FLUSH SALT - — — BOOK VALUE WITHOUT ABJUSTMENT FOR DILUTION 454 — . BY FLUSH SALT : - 4.52 - 0] 2 4 6 8 10 12 14 16 _ 18 20 ' Mwhr (X 10°) Fig. 7.1. Weight Percent of Uranium in MSRE Fuel Salt During Runs 6 to 10. y ‘ losses; these values indicate that while chemical analyses were heretofore not sensitive enough to reflect burnup losses, it may be possible that such losses will be reflected in subsequent operations with the present fuel salt. MSRE Fuel Circuit Corrosion R. E. Thoma Since oxidative corrosion of the MSRE fuel circuit results in the formation of chromous fluo- ride, the concentration of chromium in the fuel salt serves as the principal indicator of the extent of generalized corrosion. The chromium content of the fuel salt has remained very low throughout: the operating history of the MSRE. In the current report period the chromium concentration of the salt has remained at 62 £ 5 ppm (Table 7.1), corresponding to a uniform removal of chromium from the walls of the reactor circuit from a maximum depth of about 0.1 mil. Prior to the removal and replacement of the original metal and graphite surveillance speci- mens in August 1966, * the average chromium concentration of the fuel salt was 48 +7 ppm (Tablé 7.1), a value wl-mich was attained during the first power operations with the reactor. A sample taken early in October 1966, which followed the change of the surveillance specimens, showed the chromium content had reached about 62 ppm. All subsequent specimens of fuel salt analyzed ‘after that time have shown the presence of approximately 62 ppm of chromium, indicating the 4MSR Program Semiann. Progr. Rept. Aug. 31', 1966, ORNL-4037, p. 97. y 121 introduction of about 85 g of chromium into the fuel since October 1966. The increase in chro- mium concentration from 48 to 62 ppm may possibly be assignablé to corrosion of the new sur- veillance assembly. If so, the corrosion sustained by the assembly must be greater than that - previously experienced by the entire fuel circuit from December 1965 up to the present time. The metal surface exposed to the salt in the surveillance assembly is about 1000 in. 2. To pro- duce 85 g of chromium would require the'leaching of chromium from this surface, if uniform, to an implausible depth of about 10 mils. The MSRE fuel salt contained a very low concentration of U3* at the outset of power oper- ations, with possibly a maximum of 0.16% of the uranium (366 g, 1.54 gram atoms) in the usdt form (cf. section entitled ‘‘Extent of UF , Reduction During MSRE Fuel Pieparati'on”). At the termination of MSRE run 7, fuel burnup had consumed 1.66 gram atoms of uranium. Just how much oxidizing capacity is produced by fuel burnup is uncertain because it involves unverified infer- ences as to the final chemical identity of many fission products, but between 0.6 and 1.0 equiva- lent of oxidizing capacity should result for each gram atom of 23°U fissioned. If 1.0 equivalent was produced, the fuel salt would have become slightly oxidizing by the end of run 7. No such conclusion is justified, however, from the results of chromium analyses. These results indicate rather that oxidation corrosion ceased before the MSRE had generated a total of 1 Mwhr of power. Absence of corrosion iri runs 4 to 7 should probably be attributed at the beginning to the presence of U3+ and perhaps subsequently to the deposition of noble-metal fission product films. At the termination of run 7 (7800 Mwhr) a uniform film formed by the deposition of all the Mo, Nb, and Ru produced would have a thickness of approximately 150 A. The current experiment with the MSRE is scheduled to be terminated at the end of 30,000 Mwhr of operation. At that time the present surveillance specimens willl be removed for inspection and testing. It will be of considerable value tolle‘arn whether the present assetfibly has sustained the amount of corrosion which has been observed in the fuel samples since last October. Extent of UF ; Reduction During MSRE Fuel Preparation B. F. Hitch C. F. Baes, ]Jr. Uranium was added to the barren fuel salt of the MSRE as a binary mixture of 27 mole % UF, in ’LiF. This fuel concentrate had first been purified by the usual sparging with an HF-H, mixture to remove oxide, followed by sparging with hydrogen alone to complete the reduction of structural metal fluorides such as NiF , and FeF ,.5® During this final reduction step, a small portion of the UF , should also have been reduced, the amount depending upon the duration of the treatment and the equilibrium constant for the reaction 1 _ 7 UF ,(d) +5H2(g) = UF,(d) + HF(g) . 5_]. H. Shaffer et al., Reactor Chem. Div. Ann. Progr. Rept. Jan. 31, 1965, ORNL-3789, pp. 99—109. ' 6_]. H. Shaffer, MSR Program Semiann. Progr. Rept. July 31, 1964, ORNL-3708, pp. 288—303. 122 The exact amount of UF , thus introduced into the MSRE fuel has become a matter of special interest with continued operation of the MSRE owing to evidence that significant amounts of some fission products are far more oxidized than would seem compatible with the presence of significant amounts of UF ; in the MSRE fuel. Consequéntly,.the data collected by Shaffer et al.’ during the purification of the fuel salt concentrate at the production facility recently has been . examined in detail in an attempt to determine the equilibrium quotient for the above reaction: PHF XUF3 T PLTX - Tm, TUF, ’ and to determine the extent of UF4 reduction in the LiF-UF4 mixture. For small amounts of reduction, the UF 3,/UF4 ratio may be related to Q and the volume (V) of H, passed per mole of UF, (nU) by® (1'.1UF3/nU)2 - ZQPI:]/: (V/nyRT) + (n{)JFs/nU)2 , provided equilibrium conditions are maintained during sparging. The last term on the right is the initial "UFs/nU ratio. Replacing ”UFs/”U by QP;]/:/PHF , 1 2 ( V) 1 = + . 2 1/2 0 2 B Pir QPH2 nyRT (PHF) In accord with this equation, plots of l/PI?IF vs V, based on data collected at 700°C during the purification of the various batches of fuel concentrate, were found to be linear. All plots could be fitted reasonably well with lines of slopes corresponding to ¢ ~ 0.9 x 10~ % atm 172, From the final value of Py, knowing Q and Py , the average amount of uranium reduction at the end 2 of the hydrogen treatment was estimated to be 0.16%. _ In an attempt to confirm this estimate of Q and the amount of reduced uranium present initially in the MSRE fuel, an 11.4-kg portion of unused fuel concentrate was studied further in the lab- 7Unpub1ished data, supplied by J. H. Shaffer. - | 8Combination of PHF RT dnUF3= dv and nyry; Pyp 1/2 -1 p Q: n to e_liminaie PHF’ followed by integration, gives |4 1 =——2[r+ln 1-nl+cC, n,RT QPY H, where r = nUFs/nU' For small values of r this simplifies to the equation in the text. 123 oratory. >Hydrogen sparging was initiated at 5S10°C. At this relatively low temperature, no sig- nificant reduction of U** to U3* should occur; however, HF evolution was detected immediately and continued at a significant level until 250 liters of H, had passed and 0.0019 mole of HF per mole of uranium had been evolved. This indicated that inadvertent exposure of the salt to oxi- dizing impurities such as water or oxygen had occurred during prior storage, during transfer of the sample to the reaction'vessel, dr in later handling. Since the presence of HF at this tempera- ture in the amounts seen should have quickly oxidized the UF , present, it was not possible to confirm the amount of UF , initially present in the fuel concentrate. In two subsequent H, sparging runs at 700°C, however, data were obtained which permitted improved estimates of Q from plots of 1/ PfiIF vs V. The resulting values of Q are about twice that estimated from the H2 F|ow Temperature (ml min_l kg_l) Q (ofml/2) . Run1 707 53 1.74 x 107° Run2 , 705 35 1.85%10~° salt production data. It is not reasonable to attribute this discrepancy entirely to the differences. in temperature since, judging from Long’s measurements of the temperature dependence of Q in LiF-BeF melts,® more than a 30°C difference would be requiréd. It seems more likely that the discrepancy is due partly to nonequilibrium sparging conditions in the production treatment. The present value of Q = 1.8 x 10~ ¢ atm!/2 determined for the fuel concentrate is somewhat lower than the value ~4 x 10~ ¢ atm'/?2 which may be'estimated for the MSRE fuel salt at 700°C from Long’s measurements. This indiéates that UF, is not as easily reduced in the fuel concentrate as in the fuel salt. Even though equili_brium conditions might not have prevailed during purification of the fuel concentrate, 0.16% reduction of UF4 remains a_yalid estimate since, in effect, it is based upon _ the integrated amount of HF evolved by reduction which, in turn, is related by material balance to the amount of UF3 formed. Adjustment of the UF, Concenfl;ution in the MSRE Fuel Salt W. R. Grimes R. E. Thoma The fission product isotopes of molybdenum, niobium, fechnetium, rufhenium, and tellurium were expected to appear principally in their elemental forms in the MSRE system. While some might be carried as suspended metal or even in solution as moderately unstable fluorides of low valence state, they were expected to precipitate, in large part, on the metallic portions of the reactor. Although this suggested behavior has indeed taken place in the MSRE, appreciable quan- tities of molybdenum, ruthenium, and tellurium (and probably technetium and niobium) have also -QG. Long, Reactor Chem. Div. Ann. Progr. Rept. July 31, 1965, ORNL-3789, pp. 68—72. 124 been observed in the cover gas in the MSRE pump bowl. Substantial fractions of the fission prod- : uct niobium, molybdenum, tellurium, and ruthenium have been found on or in the MSRE moderator graphite. 10 The presence of these materials as gas-phase species suggested that the fuel salt contained, at the outset and until this year, much less uranium trifluoride than intended and very much less than is tolerable. A The fuel as charged into the MSRE for start of the power opération prpbably had NUF4 at very near 9 x 10~ 2 mole fraction and, at most, Ny at1.4x 10~ ° mole fraction, which corresponds to 0.16% of the uranium being UF .. The UF content of the MSRE fuel was determined!! after ap- proximately 11,000 Mwhr of operation by study of the equilibrium corresponding to 1 ' ) ) UF4(d) +—2— Hz(g) = UFs(d) + HF(g) . The result showed that the concentration of UF_ corresponded to less than 0.05% of the total uranium and probably to less than 0.02%. The MSRE fuel salt was considered to be far more oxi- dizing than was necessary or desirable and certain to become more so as additional power was :‘produced unless adjustment was made in the UF, concentration. A program was therefore ini- tiated to reduce 1% of the 228.5-kg inventory of U**, or 9.64 gram atoms, to U3™ by the addition of small quantities of beryllium metal to the circulating salt. Initially, 4 g of beryllium was intro- duced into the quel sa;lt by melting a mixture of 7LiF-BeF2 carrier salt and powdered beryllium in "the MSRE pump bowl sampler cage. Subsequently, three adflitions have been made by suspending specimens of 3/8-in. beryllium rods in the salt in the pump bowl. The capsules used for adding beryllium were similar in size and construction to those used for sampling for oxide analysis but were penetrated with numerous holes to permit reasonable flow of fuel salt. The beryllium rods have reacted with the ffiel salt at a steady rate, dissolving at approximately 1.5 g/hr. To date, 27.94 g of'beryi1i1‘1m has been introduced into the MSRE fuel salt, a quantity correspdnding to the conversion of 0.65% of the U** to U3™, 7.2 FISSION PRODUCT BEHAVIOR IN THE MSRE S. 8. Kirslis F. F. Blankenship "The initial results of tests on the chemical behavior of fission products in the MSRE were reported previously, ! with descriptions of the experimental facilities used and a discussion of the objectives of this work. Most fission products behaved as expected, with the exception of the noble metals, which showed an unexpected tendency to volatilize and to deposit on metal sur- faces and in graphite. This obsetvation, implying that the fuei salt was more oxidizing than was desirable, contributed to the decision to reduce the fuel b‘y repeated small additions of beryllium 10S. S. Kirslis, MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 165. 11A. 5. Meyer, ‘‘Hydrogen Reduction}of MSRE Fuel,”’ Intra-Laboratory Correspondence, Jan. 3, 1967. 125 metal. The special pump bowl tests in this report period were mainly directed toward following the effect of beryllium additions on the volatilization and deposition behavior of the noble metals. Also completed in the period were the radiochemical analyses on the first set of long-term surveil- lance specimens of graphite and Hastelloy N exposed in the MSRE core. Long-Term Surveillance Specimens The bulk of the radiochemical analyses were reported previouslry12 on the graphite and Hastel- loy N specimens exposed in the MSRE core for 7800 Mwhr of power operation. Further analyses were made on selected graphite samples for *5Zr, 25Nb, '4!Ce, !*%Ce, and '37Cs to provide a more complete pictute of the behavior of these isotopes. A few samples were also analyzed for 147Nd and 91'Y because of the interest in rare-earth-type isotopes themselves, as well as in their rare-gas precursors. The ne\'a-v data are tabulated, along with previous results on these elements, in Tables 7.2—7.4. In a few cases, the previous resuits were slightly corrected on final evalu- ation of the counting data. (Similar corrections for the other previously reported fission products seldom exceeded the 10% analytical error; the revised values will not be reported here.) The new data generally followed the indiéations from previous results. The flhat profile of gasZr in the interior of the middle graphite bar, at a level 100 times that of the blanks, suggests a slight volatility of zirconium, since ?5Zr has no long-lived gaseous precursor. Some of the 95Nb in the interior may have arisen from the decay of 95Z1, but this correction is negligibl\e for the first two surface samples. Even in the interior, the concentration of °3Nb in a given layer was always higher than that of 95Zr, whereas the reverse would have been true if most of the 95Nb had been formed in place from °5Zr. Thus the observed high concentrations of 5Nb in the graphite may not be ascribed to precursor behavior. The additional data on '4!Ce, 14%Ce, 'Y, and *’Nd confirmed the previous conclusion that the distribution of the rare earths and alkaline earths in graphite reflects the diffusion behavior of the precursor rare gas. The species with short-lived gaseous precursors showed steep concen- tration gradients, whereas those whose gaseous precursors had half-lives of several minutes showed relatively flat interior concentration profiles. A diffusion model has been developed'? which satisfactorily accounts for the observed distributions of the fission products with rare-gas precursors. | The very flat interior concentration profile of 137Cs in graphite was confirmed by the new results in Table 7.2. A distribution similar to that of 8°Sr was expected on the basis of precursor behavior. The data suggest a mobility of '37Cs itself, in accord with the known volatility of ele- mental cesium over cesium carbide. !4 Since 99Tc is the worst neutron poison after ?>Mo in the noble-metal fission product group, radiochemical analysis for it was attempted on several graphite samples. It proved extremely 1201SR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, pp. 165—91.. 13R. J. Kedl, unpublished communication. 19MSR Program Semiann. Progr. Rept.. Aug. 31, 1966, ORNL-4037, pp. 165-91. 126 Table 7.2. Radiochemical Analyses of Middle Graphite Bar Weight Depth Disintegfations pef Minute per Gram of Graphite Sample of Cut () 95 95 144 141 137 147 91 (mils) Zr Nb Ce Ce Cs Nd 1y Wide Face Exposed to Circulating Fuel 1 0.8463 6.02 1.12x10'° 6.93x10'! 3.08x10° 3.17x10° 1.53x10° 571x10® 1.90x10"° 4 1.2737 9.7 1.37x10% 0.16x10° 8.43x107 6.53x10° 2.02x107 <1x10? 1.83x 108 7 0.0814 7.50 1.16x10° 7.76x10%® 3.80x107 1.43x10° 1.98x107 <4x10® 3.12x10° 11 0.9145 6.94 8.20x107 2.04x10® 2.22x107 7.33x10%® 2.17x107 8.9x10° 2.07 x 10° 14 08962 6.98 1.51x10° 4.28x10% 4.07x107 555x10° 1.86x107 5.1x107 3.33x10’ 17 1.0372 8.25 3.16 x 107 2.46 x 107 3.90x 10% 1.96 x 107 4.26 x 107 2.66 x 10’ 23 0.8176 6.64 1.09x10° Side Face Exposed to Circulating Fuel 2 0.7583 7.68 8.33x 1011 2.48x10% 3.38x10%% 1.27x107 3.82x10% 2.46x10'° 6 0.9720 10.10 1.52x 102% 2.41 x 107 7.26 x10° 5.12 x 10° 8 0.5139 5.43 4.87 x10% 3.00x 107 2.50%10° 12 0.3395 3.65 15 0.6976 7.61 2.54x 107 8.39 x 103 18 0.3737 4.15 ¢ Other Side Face Exposed to Circulating Fuel 3 06135 6.21 6.36 x 1011 2.45 x'10° 3.35x 10! 8.58x 107 2.59x10° 1.60x 10" 5 0.7543 7.84 2.80 % 101% 1.08x 10%® 7.76 x 10° 1.99 x 107 9 0.6198 6.55 2.36x10° 2.18 x 107 3.01 x 10° 13 1.1652 12.53 ' 16 0.8469 9.24 2.04x107 3.77 x 108 19 0.9406 10.45 Face in Contact with Graphite 20 1.1381 9.23 6.68x10° 1.04x10'! 1.56x10° 2.15x 10'°? 4.8%10° Unexposed Graphite Blanks 10 4.48 X105 2.46 x 10° 1.07x10° 1.41x10° 2.90x10*' <1x10® 1.97x10° 24 131 x 10° 3.28x10° 8.06x10° 6.38x10° <8.3x10* <3x10% 1.41x10° Notes: 1. The samples are arranged in order of successive cuts on each face (see Fig. 7.22 of Ref. 25, p. 179). 2. The sample weights given here have been corrected for the average 4.5% loss during milling. 3. The depths of cut were calculated from the sample weights and areas and the known graphite density. g 4. The activities tabulated are corrected to the time of shutdown, 11:00 AM, July 17, 1966. 127 Table 7.3. Radiochemical Analyses of Top Graphite Bar Weight Depth Disintegrations per Minute per Gram of Graphite Sample .y of f-:m 95, 95p 144, . g, 137 147y 4 91y (mils) . Wide Face Exposed to Circulating Fuel 25 0.3602 6.23 1.28x10° 1.56x10'" 4.06 x10® 8.60x10° 2.24x10" <1x10° 3.59x10" 20 -0.4355 7.93 1.06 x 10° S 1.22x107 T <2x107 5.72x 10° 58 0.5260 9,94 9.5 x10® 4.73x10%° 1.87x10%® 1.57x 107 60 0.2916 5.51 1.49 x 107 62 1.0783 20.38 1.11 x10° 3.54x 107 1.49x10’ Side Face Exposed to Circulating Fuel 26 0.4615 11.41 5.97 x10'°% 2.38 x 10® 4.94 x 10° 31 © 0.4564 11.88 2.50 x 10% 5.22:x10% 2.13x 10° ther-Side Face E;tposed to Circfilofing Fuel 28 0.6703 17.11 5.20% 10" 2.46x 10® 5.38x 10° 33" 0.5404 14.43 8.62 x 107 Wide i’oce in Con‘focf with Groplhife 27 0.6422 11.32 1.24x10° 5.59x 10" 2.57x 10® 4.03x10° 2.50x 107 1.85x 10® 1.55x 10° 32 0.5375 9.95 2.28 x 108 - | 1.0x 107 <4x10° 2.79% 10" 59 0.3154 5.96 6.25x10° 1.05x 107 2.56 x 10® 1,87 x 107 61 0.5835 11.03 | 63 0.7310 13.82 2.04x 10° 2.17x 107 9.93 x 10° Notes: The samples are arranged in order of successi\;e cuts on each face (see Fig. 7.22 of Ref. 25, p. 179). The sample weights given have been corrected for the average 18.9% weight loss during milling. 1 2 3. The depths of cut were calculated from the sample weights and areas and the known graphite density. 4. The activities tabulated are corrected to the time of shutdown, 11:00 AM, July 17, 1966. difficult to purify ??Tc from confaminating activities sufficiently well to provide an accurate count of this low-specific-activity isotope. The results indicated qualitatively amounts of °°Tc to be expected from the decay of the determined amounts of its parent *°Mo. In view of this test and the chemical similarity of Tc and Mo, it is fairly safe to calculate 99T¢ concentrations from - ‘observed ??Mo concentrations on the assumption that °9Tc will remain where the ?°Mo was found. The nature of the distribution of various fission products in the graphite surveillance speci- mens is shown graphically in Fig. 7.2. The lines are rougil averages obtained by visually drawing smooth curves through individual points, often widely scattered, representing all of the available - data. The steepness of the concentration gradients of the noble metals and the very different gra- dients of species with noble-gas precursors ('*%Ba and #9Sr) are strikingly apparent. 128 Table 7.4. Radiochemical Analyses of Bottom Graphite Bar Weight Depth Disintegrations per Minute per Gram of Graphite Sample ’ of Cut () ) 95, 95Np 144~ 141~ 137 g 1,47Nd 91y . (mils) Wide Face Exposed to Circulating Fuel 9 11 8 9 : 7 8 8 34 0.8032 15.04 2.40x10 3.34 X 10 6.05x 10" 8.11 x10 2.80x10 _ <3x10 1.20xX 10 38 0.5979 11.64 2.49 X 10° 1.09 x 107 64 0.2323 4.68 3.12x10% 8.40x10° 6.92 x 108 66 0.3120 6.28 69 0.7183 14.49 7.63x 10° 1.84 x 10° Side Face Exposed to Ciréuloiing Fuel . 11 8 ' 10 7 9 35 0.3904 10.70 2.63x10 7.89x10% 1.04 %10 7.27 x10 3.69x 10 39 0.4480 12.98 9.00x10® 1.09x 107 2.29x10° Other Side Face Exposed to Circulating Fuel 11 8 9 37 0.5480 - 15.39 3.05x10 5.81 x 10" 9.19xX10 41 0.3520 9.62 4.49x10% 3.05x10% 7.25x10° Wide Face in Contact with Graphite 36 0.4810 9.12 3.47x10° 4.20x10'! 9.05x10%® 1.07 x 10'° 1.12x10% s5.10%10° 40 0.5936 11.77- 2.50x 109 <3 X 108 7.69 x 107 65 0.4756 9.58 3.07x10% 2.72x10° 1.10x10® 67 0.4025 8.10 68 0.6260 12.61 4.09 X 105 1.03 x 108 Notes: 1. The samples are arranged in order of successive cuts on each face (see Fig. 7.22 of Ref. 25, p. 179). 2. The sample weights given have been corrected for the average 9.1% weight loss during milling. 3. The depths of cut were calculated from the sample weights and areas and the known graphite density. 4. The activities tabulated are corrected to the time of shutdown, 11:00 AM, July 17, 1966. - Uranium. Analyses of Graphite Specimens Since appreciable quantities of uranium had been found in graphite from previous in-pile tests in which the fuel was allowed to cool and radiolyze, a number, of the milled graphite samples were analyzed for uranium by a chemical fluorometric method and by delayed-neutron counting. The chemical method on dissolved samples showed only that the uranium concentrations were less than 30 ppm by weight. The very sensitive delayed-neutron counting method gave the results shown in Table 7.5. The very low surface and volume concentrations of 235U shown in Table 7.5, corresponding to about 1 g in the complete graphite core, could have no discemible nuclear or chemical effect 129 13 ORNL-DWG 67-774 - < . ] ) Y 1 Wi —t L W T e | I | g . L\ o <§t 14 50 BN foed l\ ‘\ N, —-.\\ - W | W VAN —~ a5 o \ NN o — Sr s VN NN - ] . P — z AN \\ = Y \ \ ™~ a 10 L ¥ 11 Y - 432 Ly L} N e Te ) A1 L ~ & ‘\ \\ \\ e T\ ~ —\40gq x \ A SN S 99 © N 20 N z = %2} \ o 409 .Y Y ™~ \ -~ AN \\ 103 \\ RU \\ e 134 \ I 10 0] 10 20 30 40 50 DISTANCE FROM SURFACE OF GRAPHITE (mils) Fig. 7.2. Distribution Profile of Fission Products in lGruphife from MSRE coré. ) on the operation of the reactor. The interior uranium concentrations in the top and bottom graphite were near the blank graphite value, but those in the middle graphite bar were distinctly higher and exhibited no concentration gradient. . It is interesting to compare the depos1t10n of uranium and molybdenum in the first layer of graphite. The avérage value for uranium from Table 7.5 is 0.072 ug of 235U or 0.22 pg of total uranium per square centimeter. The average amount of °?Mo on the graphite surfaces at the time of reactor shutdown was 0.039 pg of ®°Mo per square centimeter. Assuming that 99Mo indicates the deposition behavior of the stable molybdenum fission products there must have been nearly 2 pg of total molybdenum per square centimeter of graphite after 7800 Mwhr of reactor operation. Thus the weight of total molybdenum depositing in the first layer of graphite was nearly ten times that of total uranium. _ | The amounts of seven fission products deposited on top, middle, and bottom samples of Hastel- loy N from the surveillance assembly were previously fepbrted. 15 In addition, the following val- 151bid., p. 53. : . Table 7.5. Uranium in Graphite Surveillance Specimens Micrograms i Micrograms Sample Milled of 235y of 235y No. Graphite Bar Layer per Gram per Square of Graphite Centimetef 25 Top 1 2.72° 0.080 29 2 " 0.15 58 3 0.16 60 4 0.36 1 Middle 1 3.56° 0.090 4 2 1.26 7 3 1.07 11 , 4 0.87 23 6 1.18 34 Bottom 1 0.66° - 0.047 38 2 0.06 64 3 0.07 66 4 0.11 Graphite . <0.082 blanic B aAverage of dup.licate samples which agreed within about 10%,. ues were obtained for 25Nb deposition: top, 2.68 x 1019 dis min—! cm~2; middle, 2.74 x 101° dis min~! cm~?%; and bottom, 3.79 x 10'° dis min~* cm~ 2. If the niobium were distributed uniformly over the 1.2 x 108 cm? of Hastelloy surface in the MSRE, these values correspond to 42, 43, and 90%, respectively, of the calculatefl total 5Nb present at reactor shutdown in the reactor system. Thus, on the average, about half the *Nb produced was deposited 'on the reactor metal walls. A similar average for the deposition of 9°Nb on graphite was nearly half the total present. Cor- respéndingly, analyses for ?5Nb in recent fuel salt samples, after correction for °5Zr decay since sampling, indicated that _only a low fraction of the total present remained in the salt. The Hastelloy N samples were also analyzed for 6°Co, °*Mn, Co, and Fe. The thermal and fast fluxes calculated from these values agreed satisfactorily with values from the analysis of dosimeter wires included in the surveillance package. For ?5Nb, as for the fission products previously reported, there was no correlation between flux and the amount deposited on Hastel- loy N. Fuel Salt Samples Six additional 10-g samples of circulating molten fuel, taken primarily in connection with recent pump bowl tests to assess the effect of beryllium metal additions on the volatilization 131 1 and plating behavior.of noble-fnefal fission products, were analyzed radiochemically for the 13 isotopes listed in Table 7.6. The data from the last of the previous five samples, FP7-12, are included in Table 7.6 for comparison. The noble-metal activities are plotted in Fig. 7.3. Effect of Operating Conditions In the period of reactor operation covered by the seven tabulated samples, there were four reactor drains and a number.of_ shutdowns, the longer of which lasted 83.5 days, 14.5 days, and - 14.1 days. .Sample FP8-5 was taken to test the effect of reactor drain and long shutdown (83.5 days) on the concentrations of fission products. The activities were calculated back to the time of shutdown (July 17, 1966) and should thus be comparable with the results of sample FP7-12, taken a few days before the shutdown. It is seen that the concentrations of alkaline earths, rare - earths, and ?5Zr were not significantly altered. The small rise in '3!I concentration is attributed to a slight contamination of the sample, which would have a large apparent effect when 'multiplied by the large correction factor for decay for more than ten half-lives. " There were, however, sig- nificant decreases by a factor of 4 'in the concentrations of °3Ru, !°®Ru, and !?°Te, suggesting a slow deposition of these species on the metal walls of the drain tank. Similarly, the noble- metal concentrations were lower in sample FP11-22, taken 3.2 days after shutdown, than in the ' previous sample l‘i‘P11-12, taken during power operation. The other species increased in concen- tration due to the 16 days of power operation between the two samples. Effect of Beryllium Additions It is difficult to conclude from the data in Tablé 7.6 that there was a significant effect of fuel reduction on the noble-metal concentrations in the fuel. The c-oncentrations frequently rose rather than fell, as expected, after adding beryllium. It was interesting that the ?°Mo, !%3Ru, 106Ry, and !3?Te showed parallel rises and falls. The °°Mo results were impossibly high for samples FP11-8 and FP11-12. The high values were checked by reruns on fresh samples. If all the °°Mo produced by fission remginedruniformly distributed in the fuel, the calculated concen- —1. A simple calculation shows that if all the %Mo tration would be 1.4 x 101! dis 1;1in‘1 g produced by neutron activation of the ®®Mo in the first 0.1-mm thickness of the Hastelloy N - reactor containment vessel diffused instantaneously into the fuel melt, the increase in ?°Mo conc‘en_tration would be only about 10° dis/min per gram of fuel. It thus appears that a mechanism is required which either concentrates fission-produced ?9Mo and other noble metals in the pump bowl or resuits in 1arge temporal and spatial variations in their concentrations. Dissolved ?°Mo would undoubtedly be uniformly distributed. If the noble metals circulated as a shspension of insoluble metal particles, it is conceivable that they might concentrate in the pump bowl or vary in concentration with pump bowl level, cover gas pressure, and other operating variables. Since clean metal surfaces are not wet by the fuel salt, there might also be a tendency for metal par- ticles to collect around helium bubbles, which are probably most numerous in the pump bowl. The froth or flotation hypothesis can be checked experimentally in several ways. Table 7.6. Analyses of Fuel Salt Samples Experiment FP7-12 FP8.5 FP10-12 F P10-20 FP11-8 FP11-12 FP11-22 Sampling date 7-13-66 10-8-66 12-28-66 1-9-67 241367 2.21-67 3.9.67 Operating time, days® 4.1 off, 11.9 on 11.9 on, 83.5 off 14.5 off, 13.2 on 14.5 off, 25.4 on 14 off, 16.1 on 14 off, 24.1 on 40 on, 3.2 off Nominal power, Mw 7.2 0 - 7.4 7.4 7.4 7.4 0 Accumulated Mwhr 7200 7800 13,800 15,800 19,000 20,400 22,400 Be addition, g 5.5 10.65 11.65 Isotope Fission Disintegrations per Minute per Gram of Sa 1t? Yield (%) c ' 9.6 hr ?1s¢ 5.81 1.32 x 10!! 1.28 x 1011 1‘.31><1011 1.33x 10! 1.53 x 1011 51 day %%r 4.79 3.96 x 1010 3.70x 1010 3.83 x 1010 4.70 x 1010 4.78 x 1010 6.50 x 1019 7,93 x 10!° 33 day !*lce 6.0 6.88 x 1010 7.20 x 1010 4.63 % 1010 4.11 x 1010 9.23 x 1019 1.10x 101! 285 day !*%Ce 6.0 1.91 x 1010 1.79 % 1010 2.43 x 1010 9,68 x 1010 66 hr Mo 6.06 3.15 x 1010 3.56 x 1010 4.77%x 10 3.20x10''> 2.54x10!'!? 9.09x10!° 39.7 day !°3Ru 3.0 7.14 x 10° 1.42 x 10° 8.02 x 108 6.08 x 108 5.22 x 10% 5.33 x 10° 3.66 x 10° 1.01 year '9%ry 0.9 -2.13x 108> 5.00x 107 ~4.0x 107 2,78 x 1072 1.62 x 108 ~1.9%x 108 1.46 x 108 y - 77 hr 132%Te 4.7 3.81x 1010 1.56 x 1010 2.04 x 1010 5.50x 1010 3.76 x 1010 2.80x 1010 33 day 129m71e 0.35 4,94 x 108 1.38 x 108 1.49 x 108 ~3.1x 108 8.05 day 131 3.1 5.36 x 10'° 7.94 x 1010 5.46 x 1010 7.16 % 1010 4.96 x 1010 7.55x 101 8.30x 101° 35 day 25Nb 6.2 (2.45x 10112 4,77 x 1019 6.73 x 10° 2.36 x 10192~ 1.14 x 10° 1.21 x 1010 65 day °°Zr 6.2 6.55 % 101° 6.01 x 1010 5.42 x 101° 5.21 x 10190 9.34 x 1010 9.25x 101% 1,12 x10!! 12.8 day !*°Ba 6.32 1.42 x 1011 9.00x 10! 9,97x10!° 1.30x 101! 2,74 x 1011> 30 year '37Cs 6.0 3.13 x 108 4.04 x 108 “Duration of previous shutdown and of continuous operating time just before sample was tak shutdown, bCalculated to the time of sampling or to the previous shutdown. en; vice versa for the two samples taken during c€l 133 ORNL-DWG &7-4783 41os jo woab uad ajnuiw sad suolypibajuistp [+;] -] ~ 22-hdd = NMOgLNHS Aop-¢ T T xkll T T Y T T T T T A |DCI T 17T — ] /| /| 3 A_”_.r_ 2h=thdd M-t —1 —] h\%l — gttt | —t— — |~ —H T+ 4+ — — ] ‘28 699t} [Tlpm / 7 | =3 mnzn_.._..ll|llmlllllllIIA.|-||-|IIllloifl I T % — 8 ? - |— — | — — / N I.I.l.. / 4 .y 02-0kdd FF—TF—T— —F — — —— T+ — — =i = 0 — 4 — — re— S — — —— eg 6590} - / / \\ o 2l-okdd H H — — +— 4+ — 14 IJ. L e — | A L IV- — _ / N / NOILIQaY - _ ag bg'g §-8dd H —=——F +— +— —F — = | —+V— TV — — — 1117 T Tl e 1t = —t — |Ih|P||r||- et o - NwIN&n_ S Oy — TNMOQJLNHS hOUlm.meI! T Ill!n..‘ Mwu| — — Tt -n._ Hll]A.Ullfl‘lh"l 1 EI.I“I R R ..H_I = 3R <4 = » 8 o g ’ & o n o st n ‘2 o n o ) [Te] o~ @ Q Q o o e} ,:om JO woib Jad ajnuiw J1ad suoipibajuisip 16 20 24 12 Mwhr (x103) F'ig. 7.3. Nobel-Metal Activities in Fuel Samples. Calculated back to the time of sampling or to the previous shutdown. 134 The 9“_‘Nb concentrati,\ons in Table 7.6 varied erratically and did not parallel the behavior of the other noble metals. This is ascribed to analytical difficulties, which are being further in- vestigated. An unavoidable difficulty is that a large correction for ®*Zr decay must be made for each salt analysis. . The other fission products (°!Sr, 8%Sr, 14!Ce, 14%Ce, 131, ?5Z1, and !*°Ba) generally showed approximately the expected yields in the salt phase within analytical error. It was re- ported previously *® that reactor power levels calculated from the observed concentrations of elements like strontium and cerium, which remain in the fuel i)hase, were consistently about 20% lower than the power levels calculated from coolant heat balances. This discrepancy has per- sisted in the analyses of the new fuel salt samples. To date there have been several dozen radiochemical analyses, all of which indicate lower than nominal power levels. Pump Bowl Volatilization and Plating Tests Six pump bowl tests were run in this report period in which metal specimens were expdsed to the gas phase and the fuel phase of the pump bowl for 10 min. ‘The previous technique 7 was modified slightly in that the coils of silver and Hastelloy N were attached beside the capsule cables rather than being wound on them. This facilitated sample disassembly in the hot cell and furnished an additional specimen of stainless steel cable in the gas phase. Each of the four gas-phase specimens (Hastelloy N, silver, nickel-plated key, and stainless steel cable) and the single fuel-immersed specimen (stainless steel cable) were leached or dissolved, and the solutions were analyzed for at least nine fission products and for #°°U. The gross features of fission product behavior in these runs were similar to those reported previously, with heavy deposition of the noble-metal fissiofi products on all specimens and light contamination by °5Zr, rare earths, and alkaline earths. All the analytical data obtained will not be presented since the mass of detail would be confusing. However, to make clear the effect of the three beryllifim additions, the depositions on the Hastelloy N spécimens in the gas phase and the stainless steel samples in the fuel phase are presented in detail in Tables 7.7 and 7.8. The results on the other gas-phase specimens were generally qualitatively similar to those on the Hastelloy N samples. | | _ , It is seen from Table 7.7 that noble-metal volatilization usually increased after the first and third beryllium additions and decreased as expected only after the second addition. Neverthe- less, tfie average deposition on the gas-phase specimens was slightly lower after the three ad-~ ditions than it had been previously. Much larger decreases in volatilization followed reactor, shutdown. Experiment FP8-5 was run just before power operation was resumed after an 83.5-day shutdown, and FP11-12 after a 3.2-day shutdowh. In each case, activities were calculated back to the time of shutdown. The deposition on the gas-phase specimens of elements with stable 1%1bid., p. 168. 17rbid., p. 69. Table 7.7. Fission Produet Deposition on Hastelloy N Experiment FP7-12 FP8.5 FP10.12 FP10-20 FP11-8 FP11-12 FP11-22 Sampling date 7-13-66 10-8-66 12-28-66 . 1-9-67 2.13-67 2+21-67 3.967 Operating time, days?® 4.1 off, 11.9 on 11.9 on, 83.5 off 14.5 off, 13.2 on 14.5 off, 25.4 on 14 off, 16.1 on 14 off, 24.1 on 40 on, 3.2 off Nominal power, Mw 7.2 -0 7.4 7.4 7.4 7.4 ' 0 Accumulated Mwhr 7200 7800 13,800 15,800 19,000 20,400 22,400 Be addition, g 5.5 10.65 11.66 Isotope Fission Disintegrations per Minute on Total Specimenb Yield (%) 66 hr 2o 6.06 3.35 x 1019 4.00 x 10!° 1.36 x 101! 5.61 x 1010 1.51 x 10!! 2.49x 1010 39.7 day '°%Ru 3.0 2.14 x 10° 1.51 x 108 3.60 x 108 1.03 x 10° 4.71 x 108 2.25 x 10° 1.14 x 107 1.01 year 9°Ry 0.9 ~7.2 x 107 4.86 x 106 ~7.0x 108 3.63 x 107 1.5x 107 7.1 x 107 77 hr '32Te 4.7 3.67 x 101! 1,55 x 101! 3.36 x 101! 1.20 x 1011 1.56 x 10! 2.74 x 1010 33 day 1%977¢ 0.35 2.40 x 108 8.79 x 108 2.5 x 10? ‘ 8.05 day !31I 3.1 2.71 x 1010 Y 1.77 x 1010 1.01 x 1010 5.77 x 107 7.58 x 10° 9.68 x 108 35 day 9°Nb 6.2 3.91 x 10° 1.1 x 107 65 day 257¢ 6.2 7.49 x 106 <3.6 x 10° <4.0x 108 $4.5x10° £3.2x 108 ~6.54x% 10° 12.8 day '*%Ba 6.32 1.84 x 108 - 2.73 x 108 1.72 x 108 1.35x 108 7.40 x 107 235y, ug total 2.63 © 0.64 . 091 0.423 1.88 1.79 “Duration of previous shutdown .and of continuous operating time just before sample was taken; vice versa for the two samples taken during shutdown. PCalculated to the time of sampling or of previous shutdown. SET Table 7.8. Fission Product Deposition from Fuel on Stainless Steel Experiment FP7-12 FP8-5 FP10-12 FP10-20 FP11-8 FP11-12 FP11-22 Sampling date 7-13-66 10-8-66 12-28-66 1-9-67 2-13-67 2.21-67 3.9-67 Operating time, days? 4.1 off, 11.9 on 11.9 on, 83.5 off 14.5 off, 13.2 on 14.5 off, 25.4 on 14 off, 16.1 on 14 off, 24.1 on 40 on, 3.2 off Nominal power, Mw 7.2 7.4 7.4 7.4 7.4 0 Accumulated Mwhr 7200 7800 13,800 15,800 19,000 20,400 22,400 Be addition, g 5.5 10.65 11.66 Isotope Fission Disintegrations per Minute per Specimenb Yield (%) 66 hr ?9Mo 6.06 1.30x 10! 2.23 x 10! 3.13 x 101! 7.28 x 10! 4.20x 10'1 7,13 x10° 39.7 day !%3Ru 3.0 3.10x 10° 5.90x 108 2,24 x 10° 9.50 x 10° 1.88 x 1087 1.38 x'10° 1.01 year 196Ry 0.9 6.60 x 107 1.77 x 107 ~9.2 % 107 2.96 x 108 5.6 x 1057 4.04 x 107 77 hr 321e 4.7 2.21 x 101! 3.59 x 1011 1.96 x 10! 3.84x 1019 3,67 x10!° 2.17x10!° 33 day 129mTe 0.35 3.17 x 108 1.09 x 108 1.34 x 10° | 8.05 day 31 3.1 3.85x 101° 1.23x 10° 1.71 x 1010 1.19% 101% . 3.38x10° 2.15 x 10° 2.63 x 10° 35 day ?5Nb 6.2 7.87x 1010 1.7 %107 65 day 95Zr 6.2 2.84 x 107 <6.5 x 109 <1.2 x 107 S1x107 ~6.3x108 7.2 x 108 12.8 day !*%Ba 6.32 4.81 x 108 1.71 x 108 4.34 x 107 3.25 x 107 3.61 x 107 235y, pg total 24 17.6 1.65 3.57 “Duration of previous shutdown and of continuous operating time just before sample was taken; vice versa for the two samples taken during shutdown, PCalculated to the time of sampling or of previous shutdown. o1 137 fluorides was not affected either by fuel reduction or reactor shutdown. The fact that noble- metal fission products volatilized to a considerable degree after reactor shutdown indicated that the process of fission was not directly related to the vo-latilizing process. As in previous runs, it was observed that deposition on the four different metal specimens was of similar magnitude for all species. Less reaction of noble-metal fluorides with silver in particular was expected since AgF is relatively unstable. In ekperiment FP11-22 a gold wire coil was substituted for the silver one, and it picked up as much of the noble metals as the other metal specimens. This lack of chemical discrimination is puzzling and k:_asts some doubt on the supposition that the noble-metal fission products leave the fuel phase as high-valent fluorides. To'eliminate the possibility that the observed deposition was due to reaction with oxide films on the metal specimens, experiments are planned in which the specimens will be reduced before in- sertion into the pump bowl. These observations of deposition indicate the value of laboratory experiments to identify the chemical nature of the noble-metal species that volatilize from radio- active fuel salt. Such experiments appearto be feasible using either mass spectrometric or gas chromatographic techniques. The data in Table 7.8 show that the average deposition of noble metals was heavier on the stainless steel cable immersed in fuel than on Hastelloy N in the gas phase, although contami- nation by salt, as indicated by °5Zr, 14°Ba, etc., was no greater. It was interesting that the ’plating behavior generally paralleled the volatilization behavior. Thus on the average, the amount of noble metals plated decreased after the second beryllium addition but increased after the first and third additions, as did deposition on the gas-phase specimens. Also, the amount plated was distinctly lower in the two runs made after reactor shutdown. This parallelisrfi is consistent with the theory that high-valent noble-metal species in the fuel take part in both the plating and vol- atilization processes. More information on the nature of the high-valent species in the fuel may be obtained by immersing specimens of other metals besides stainless steel in future pump bowl tests. Becaqse. of the practical interest in the behavior of fhe molybdenum and %%Tc fission p;od- ucts, the results of these pump bowl tests for 99%o are given in detail in Table 7.9. The ques- tionable effect of fuel reduction and the definite effect of reactor shutdown are shown clearly. It is seen that deposition on the nickel-plated key was. relatively heavy, as qbser_ved previously, while depositions on the other ‘specimens were of similar lower magnitudes. The larger size of the nickel specimen probably accounts for the larger total amount deposited. The heavy depqsit on the key suggests the truly gaseous nature of the ?*Mo Spe.cies, since the key is surrounded by the latch and only its bottom tip projects into the pump bowl spéce. The qualitative nature of these tests is shown by the data in Table 7.9. It might be concluded from Table 7.9 that depo- sition of ??Mo on silver was uniformly higher than on Hastelloy N. This is not borne out by previous data, nor by the behavior of the other noble metals. Individual values are probably reproducible only within a factor of about 4. On this basis, only sizable changes in deposition, such as between the last two runs, are to be considered significant. 138 : Table 7.9. Deposition of %Mo on Pump Bowl Specimens Experiment FP7-12 FP10-12 FP10-20 . FP11-8 FP11-12 FP11-22 Sampling date 7-13-66 12-28-66 1-9-67 2-13-67 2-21-67 3-9-67 Operating time, days® 11.9/4.1 13.2/14.5 25.4/14.5 16.1/14 24.1/14 40/3.2 Nominal power, Mw 7.2 7.4 7.4 7.4 7.4 0 ’ = Accumulated Mwhr 7200 13,800 15,800 19,000 20,400 22,400 Be addition, g 5.5 10.65 11.66 Sample ‘ Disintegrations per Minute per Specimenb x 1010 x 1010 x 1010 x 1010 x 1010 x 100 Hastelloy N 3.35. 4.00 13.6 5.61 151 2.49 Silver 12.7 67.0? 16.2 18.2 29.3 2.82 (Au) SSG° ' : 9.57 12.7 8.90 20.2 1.83 Nickel? 17.6 141 105 26.7 54.6 14.4 SSL° 13.0 22.3 31.3 . 72.8 42.0 0.71 : 1 —-1f - Fuel, dis min~ " g 3.15 3.56 4,77 32.0? 25.4? 9.09 ®The slash separates the durations of the previous shutdown and of continuous operating time ]ust before the sample was taken; vice versa for sample FP11-22, taken during shutdown. Calculated to the time of sampling or of previous shutdown. °Stainless steel cablé in gas phase. dNickel-plated key. ®Stainless steel cable immersed in fuel phase. mgMo activity in fuel sample taken simultaneously. Uranium on Pump Bowl Metal Specimens Also recorded in Tables 7.7 and 7.8 are the quanfities of 235U determined by delayed-neutron counting in the leaches of the metal specimens exposed in the MSRE pump bowl. The reported ! results varied from 0.4 to 24 pg of *>°U per sample. Of the total of 25 samples run, only five exceeded 5 ug of 235U, with the remainder averaging 2 pg of ?33U. No correlation was apparent R .between the deposition of urériium and that of fission products'; thus the uranium found did not merely represent contamination of the samples by fuel salt. It is not surprising that uranium - shows different chemical behavior from the noble metals. What is difficult to explain chemically is that it volatilizes or plates at all. Fortunately the extent of uranium volatilization and plating is small (as shown more conclusively in the examination of the long-term surveillance specimens) and is of negligible practical concern. Freeze Valve Capsule Experiments Although valuable qualitative information was derived from the tests in which metal specimens were exposed in the pump bowl, these tests suffered from the drawback that the results could not be interpreted quantitatively. Therefore; a capsule was designed to take a pump bowl gas sample of known volume from whose analysis the gaseous concentrations of fission product species 139 could be calculated. It was required that the sampling device be small enough to pass freely through the bends of the 1 1/2-in.-diam sampling pipe and that it should operate éutomatically when it reached the pump bowl. The device shown in Fig. 7.4 operated satisfactorily to furnish 20-cc samples of pump bowl gas. The capsule is evacuated and heated to 600°C, then cooled under vacuum to allow the Li BeF, in the seal to freeze. The double seal prevents loss of Li,BeF , from the capsule duting sampling. The weighed, evacuated capsule is lowered into the pump bowl through the salf—sampling pipe and positioned with the bottom of the capsule 1 in. above the fuel salt level for 10 min. The freeze seal melts at the 600°C pump bowl temperature, and pump bowl gas fills the 20-cc volume. The capsule is then withdrawn to 2 ft above the pump bowl, and the capsule is allowed to cool. Thete is a slight ‘‘breathing’’ action of the capsule when the initial sample is partially exhaled as the capsule warms from the Li ,BeF , liquidus temperature (457°C) to 600°C. As the capsule cools from 600 to 457°C in the sampling pipe, additional gas is drawn into the capsule. The double freeze seal prevents loss of the salt seal during the exhalation. ORNL-DWG 67-4784 STAINLESS STEEL CABLE 3/ _: ) VOLUME | 3/—in. OD NICKEL 20 e LizBeF, NICKEL CAPILLARY 2Befy - Fig. 7.4. Freeze Valve Capsule. . 140 It isrthou‘ght that the reactive fission product gases first inhaled react rapidly with the interior metal (nickel or stainless steel) walls of the capsule, so that correction need not be made for loss during the exhalation. The final inhalation is 2 ft up the pipe, where the atmosphere should be relatively pure helium. The cooled resealed capsule is withdrawn from the sampling pipe and transported in a carrier to the analytical hot cells. Here a Teflon plug is placed over the protruding capillary and the exterior of the capsule is thoroughly leached free of fission product activities. The top of the capsule is cut off, and the interior metal surface is leached with basic and acid solutions for 1 hr. The bottom of the capsule is then cut at two levels to expose the bottom chamber of the capsule. The four capsule pieces are placed in a beaker and thoroughly leached with 8 ¥ HNO, until the remaining activity is less than 0 1% of the original activity. The three leach solutions are radio- chemically analyzed for the activities shown in Table 7.10 and for 235U by delayed-neutron counting. Table 7.10 gives the total activities in the leach solutions for the freeze valve capsule runs made before and after the first 5.5-g beryllium add1t1on to the fuel. The runs made after the next two beryllium additions, failed since the capsules were lowered a little two far into the pump bowl -and withdrew salt samples. In the two successful runs, the final leach of the segmented capsule contained about nine-tenths of the total of each of the activities. The activities either were | trapped by the sealing salt or reacted very rapidly with the walls of the bottom chamber or the capillaries. The results of Table 7.10 generally confirmed the qualitative indications from the pump bowl tests. For comparison, Table 7.10 includes the ana1y51s of fuel salt sample FP10-20, which was taken between the two freeze valve capsule runs. It is seen that the 20-cc gas samples con- tained more **Mo, '°*Ru, '°Ru, and !*?Te than were contained per gram of fuel salt. The 'slight decrease in most of these activities after adding 5.5 g of beryllium is probably within analytical error. The °°Nb activities were relatively low and may reflect the analytical diffi- culties with this isotope. On the other hand, the 95Zr and 1*%Ba activities were very low, indicating that less than 0.5 mg of fuel salt had entered the capsules. The small amounts of these activities found may have been due to the slight volatility of Z1F , and to the 16-sec - 140xe precursor of 14°Ba. The quantities of 235U found were low, but higher than expected. Although 0.5 mg of fuel salt would more than account for the amount of ?*°U found, this ex- planation is difficult to accept since the values for ®5Zr and !*°Ba were lower in the sample for which the 235U value was higher. Tentatively, the 23°U values are taken as representing volatilization of uranium. The sharp drop in the uranium value after adding 5.5 g of beryllium metal was not reflected in the amounts of uranium found on the metal specimens of the corre- sponding pump bowl tests (see Table 7.7). This correspondence will be examined in future tests. The known volume of the gas samples permits us to calculate quantitatively the molar gas- eous concentrations of the noble metals and of uranium in the helium cover gas, also given in Table 7.10. It is seen that the gaseous concentrations represent significant partiél pressures of uranium, molybdenum, ruthenium, and tellurium. The concentrations for 99Mo multiplied by the 141 Table 7.10. Freeze VYalve Capsule Results Experiment FP10-11 FP10-22 FP10-20 Sampling date 12-27-66 1-11-67 1-9-67 Operating time., daysa 14.5 off, 12.6 on 14.5 off, 27.7 on 14.5 off, 25.4 on Nominal power, Mw 7.4 7.4 7.4 Accumulated Mwhr 13,600 16,200 15,800 Be addition, g J 5.5 Fission Before Be After Be Isotope Yield & ) Fuel Salt b (%) Dis/min Total Ppm Dis/min Total® Ppm (dis min~} g_l) 66 hr ?9Mo 6.06 2.04 x 101 5.4 1.36 x 10*! 3.6 4.77 x 1010 39.7 day '?3Ru 3.0 3.80x 10° 1.4 2.63 x 10° 1.0 " 6.08 x 10° 1.01 year '%%Ru 0.9° ~6.7x10' ~0.24 7.74 %107 0.27 2.78 X 10" 77 tr 13%1e 4.7 5.73 x 1010 1.7 5.08 x 10'° 1.6 2.04 x 101° 8.05 day ‘3'1 3.1 9.75 x 10° 0.75 2.03 x 107 0.18 7.16x101° 35 day ?5Nb #6.2 $3.26 x 107 2.09 x 108 ‘ 2.36 x 101° 65 day °°Zr . 6.2 <2.9x10° <0.0018 $2.2x 107 £0.012 5.21x 10'° 12,8 day '*°Ba 6.32 2.75 x 10° 0.034 3.48 x 10° 0.042 9.00 x 10'° . . , Sy 3.86 ug 46 0.55 pg 6.5 14,000 ug/g “Duration of previous shutdown and of continuous operating time just before sample was taken. bDisintegrations per minute calculated to the time of sampling or of previous shutdown. total cover gas flow (5000 liters/day) correspond to more than half the total ?°Mo produced per day. Since previous material balances indicated that the bulk of the ®°Mo should deposit on the Hastelloy N surfaces or remain in the salt, it is suspected that the gas concentration in the sampling volume surrounded bg} the mist shield is higher than in the remainder of the pump bowl gas phase. A similar calculation for the gas sample before beryllium addition indicates that more than 2 g of uranium per day would be swept out by the cover gas. The next gas sample indicated much lower uranium loss. Clearly, more analyses of freeze valve capsule samples are needed to determine more accurately the volatilization of uranium. Special Pump Bowl Tests A special capsule has been designed (Fig. 7.5) to determine the concentrations of hg}dro- carbons and volatile fluorides in the pump bowl gas. A known volume of the sample gas is made to pass through a weighable stainless steel cloth container containing CuO. From the . weight change and the fluoride ‘analysis of the CuO, the gaseous concentrations of hydrocarbons and volatile fluorides can be calculated. A single test of this type with flush salt in the reactor has been successfully run and showed detectable amounts of hydrocarbons and HF. Quantitative results will be reported when results are available on several tests. 142 ORNL-DWG 67-4785 TOP OF PUMP BOWL [ CONE SEAT — ] MIST SHIELD —— o] T———Cu0 IN STAINLESS STEEL CLOTH BASKET [T~ FUEL SALT __ |l — Fig. 7.5. Hydrocarbon Analyzer. The above test can be run only with the.reactor fuel pump turned off. A variation is being developed which can be run with the reactor at power and the fuel circulating. Another special capsule has been designed for exposing graphite specimens to both the fuel phase and the gas phase of the MSRE pump bowl. This test will permit the evaluation of noble- metal deposition on graphite under various specific short-term operating conditions. - General Discussion of Fission Product Behavior Additional qualitative pump bowl tests and new quantitative freeze valve capsule tests have confirmed the previously reported tendency of the noble-metal fission products t\o leave the fuel by volatilization and by plating on Hastelloy N. The observed behavior was difficult to interpret thermodynamically, as discussed previously. These difficulties were compounded by the currently reported observations that uranium also volatilized and that the behavior of noble metals was not significantly affected by the addition of 27.8 g of beryllium metal to the fuel, sufficient to reduce 143 0.65% of the U** content to U3*. Analyses of the reduced salt for U3+ by a recently developed method indicated that the U4*/U3" ratio was far too low to produce the observed partial pressures of volatile high-valent fluorides of noble metals afid uranium. In order to retain the volatile fluoride hypothesis, recourse must be had to kinetic explana- . tions. It may be postulated that sizable steady-state stoichiometrically equivalent concentrations of US* and U3* are produced radiolytically in the fuel melt. ‘This postulate implies that the oxi- dized and reduced species do not undergo back reaction fast enough for their concentrations to fall to very low values. In the radiolysis of water at high temperatures, steady-state concentra- tions of H, and O, are produced and maintained analogously. U3* and U* are logical choices for the surviving oxidized and reduced species, since U*" is the bulk fuel constituent that is most easily reduced and most easily oxidized. Now the US* can oxidize the noble-metal fission products to high-valent volatile states. Bubbles of helium known to be circulating with the salt can act as kinetic traps to preserve the oxidized s;;ecies from reaction with U3*. The probability of reduction is much poorer in the dilute gas phase. The oxidized fluorides may in this way be delivered to the pump bowl and swept out.' ' A difficulty with this theory, aside from its several assumptions, is that noble-metal volatili- zation was nearly as great three days and 83.5 days after reactor shutdown (Table 7.7) as during reactor-operation, whereas the radioactivity of the fuel melt and consequently its rate of radioly- sis had changed by several orders of magnitude. A very strong dependence of the rate of back reaction of U* and U5 on their concentrations would be required to reconcile this fact to the - theory. An alternate theory postulates that the noble metals circulate as metal sols in the fuel melt. Since the fuel melt does not wet clean metals, the sol would tend to accumulate at the interfaces of the fuel with the gas phase, such as helium bubbles. When the helium bubbles go through the J spray ring and burst, some of the metal sol may be sprayed into the gas phase as a gaseous sus- pension, which would act like a volatile species. Difficult to explain by this theory are the volatilization of uranium and the appreciable pene- tration of noble metals into the surfaces of the graphite surveillance specimens. It is clear that information is needed on the nature of the volatilizing uranium and noble-metal species before a credible account of their behavior can be given. Experiments in this direction are planned. b ' The reduction of the fuel melt by adding beryllium metal will be continued until at least 1% of the U** has been converted to U3*. This additiqnal reduction may have some effect on noble- metal volatilization. The addition of a little hydrogen to the helium cover gas may be effective if volatile fluorides are involved. The effectiveness of hydrogen in reducing noble-metal volatili- zation may be tested in the hot cell by measuring the volatilization from a‘large sample of molten MSRE salt during sparging with helium and with hydrogen. ~ ' 144 7.3 PHYSICAL CHEMISTRY OF FLUORIDE MELTS The Oxide Chemistry of LiF-BeF,-ZrF, Mixtures B. F. Hitch C. F. Baes, ]r.' Previously described measurements®®+19 of the solubility of BeO in LiF-BeF, mixtures and of the solubility of Z1O, in simulated MSRE fuel-salt—flush-salt mixtures have been completed. The results, which have been reported more fully elsewhere,?® may be summarized by the following expressions: In LiF-BeF, saturated with BeO at 500 to 700°C, log X ,_ = —0.901 + 1.547Xp o — 2625/T . 0] 2 In (2LiF-BeF ). + ZiF , satutated with Z10,, at 500 to 700°C, a — 3/2 X s b(XZrF4) ’ O : 1/2 (XZrF4) whereifl log a = ~1.530 — 2970/T , log b=—1.195 — 2055/T . In these expressions, the mole Afraction is defined n, X’ _ I i - B o it "Ban + "sz4 Z0, replaces BeO as the least soluble oxide when X, - exceeds 0,0008. This corresponds to 4 . 1.6 wt % fuel salt in flush salt. With further increases in the amount of fuel salt, the qxi.de tolerance decreases at a given temperature, passes through a minimum at X, - approximately o : 4 equal to 0.01, corresponding to approximately 20 wt % fuel salt, and then increases (Fig. 7.6). Solubilities of SmF; and NdF in Molten LiF-BeF, (66-34 Mole %) F. A. Doss F. F. Blankenship J. H. Shaffer Measurements of rare-earth fluoride solubilities in molten LiF-BeF2 mixtures have been re- sumed to supplement earlier data on rare-earth trifluoride solubility behavior in molten fluoride _ systems.?! This experimental program will examine the behavior of those rare earths which are ;ac. F. Baes, Jr., and B. F. Hitch, Reactor Chem. Div. Ann. Progr. Rept. Dec. 31, 1965, ORNL-3913, p. 20. 19MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 133. 20B. F. Hitch and C. F. Baes, Jr., Reactor Chem. Div. Ann. Progr. Rept. Dec. 31, 1966, ORNL-4076, p. 19. . : 21'I'\’ezulctor Chem. Div. Ann. Progr. Rept. Jan. 31, 1960, ORNL-2931, p. 77. 145 6R NL-DWG 67-769 7(°C) 700 600 500 1000 T T ] Ry /440( 47‘50 . 14 e N4y Je ‘S 200 P o, 29 OXIDE CONCENTRATION (ppm)} ) ® N £ 565° s 50 700 [+ 850 "3, O 033 5 Sy 2 &\ 50 £ 526 LF AVARN BeF 2 845 2LiF- BeFa/sToofatso 400 400 450 500 545 P 465 £ 370 Fig. 7.8. Phase Diagram of the System LiF-Ber-ThF4. | 100 — 1050 [— 1000 — TEMPERATURE (°F) 850 — 800 —— 750 —/—— 950 —— 900 —— TEMPERATURE {°C) 593 566 538 510 454 426 148 ORNL-DWG 67— 4787 3400 /////,! 3000 ////// 2600 LIQUIDUS TEMPERATURE ///, ,///////// 2200 /2/, ’ 1800 //// THORIUM 7 /// 1400 ////// 1000 600 10 15 20 25 30 Tth (mole %) Fig. 7.9. Thorium Content of LiF—Ber-ThF4 Compositions on the Even- Reaction Boundary Curve L — 3LiF . Th F4(ss) +7LiF+6Th F4. Toble 7.11. MSBR Blanket Compositions THORIUM CONCENTRATION (g / liter at 800 °C) Thorium Concentration ' ‘Liquidus Temperature Composition (mole %) (g/liter) Degrees C Degrees F 'I‘hF4 LiF Bos:l:‘2 863 450 842 7 63 30 1207 469 876 10 66 24 1828 500 932 16 69 15 2419 550 1022 22.5 71 6.5 2868 568 1054 29 71 149 7.4 SEPARATIONS IN MOLTEN FLUORIDES Removal of Rare Earths from Molten Fluorides by Precipitation on Solid UF, F. A. Doss H. F. McDuffie J. H. Shaffer Studies of the precipitation of rare-earth trifluorides from solution in LiF-BeF, (66-34 mole %) on solid UF, have continued.2? This program is concerned with the retention of rare-earth fission products on a bed of solid UF, as a possible method for reprocessing the fuel solvent of the reference design MSBR. Experiments conducted thus far have examined the removal of selected rare earths from the simulated fuel solvent upon addition of solid UF,. Results obtained fbr the removal of NdF, are illustrated in Fig. 7.10. Material balance calculations of the system indicate that the composition of the solid phase varied from 0.54 to 0.22 mole %_NdF3 in UF, as UF, was added to the salt mixture. Subsequent changes in the concentration of NdF3 and UF, in the salt mixture, as the temperature was varied from 550 to 800°C, corresponded to a constant solid composition of 0.22 mole % NdF3 in UF, over the temperature region. \ 22MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 145. ORNL-DWG 67— 4788 N N . NEODYMIUM FOUND IN SOLUTION (mole fraction X10%) '\ T \. d 0.4 : o — ® WEIGHT OF SALT.MIXTURE: 1.{{kg 0.2 o] 0 0.2 0.4 0.6 0.8 1.0 1.2~ UFy ADDED (moles) Fig. 7.10. Removal of NdF3 from Li|'=-BeF2 (66-34 Mole %) at 550°C by Precipitation on-Solid UF3. 150 Extraction of Protactinium from Molten Fluorides into Molten Metals D. M. Moulton W. R. Grimes F. F. Blankenship J. H. Shaffer ~ Studies of the extraction of protactinium from the MSBR blanket salt (73LiF-2BeF -25ThF ) by reduction with thorium-bismuth amalgams have continued. In an experiment at 650°C in which the salt originally contained about 600 ppm UF and a trace of PaF , thoriu:fi was added to the bismuth in several successive small increments, and after equilibration the concentrations of Li, U, Th, and Pa-in the bismuth and U and Pa in the salt were measured (Li and Th in the salt stayed constant). Letting Xm and Xs denote mole fractions in the metal and salt phases and with activity" coefficients of unity, we can define a half-cell voltage for each metal: RT X_ 8:83—-r§lnxm, (D | - 8 where 8(1) is a constant to be evaluated later. Plotting In (X_ /X ) for each of the four metals vs the amount of added thorium gave four parallel lines. This was taken to mean that the metals were indeed in equilibrium; therefore, all the £’s were set equal to each other, requiring that the 8; be- adjusted to fit. ‘ ‘ - The stahdard half-cell potential 80,°f a metal is nomally defined for the reaction M = M?* + ne—, where M is the pfire metal in the physical state appropriate to the temperature and M” " is the ion in solution referred to some standard state. Baes?3 has tabulated 80 for a number of elements, using as the standard state of the ion a hypothetical unit mole fraction solution in the MSBR fuel solvent (extrapolated from dilute solution)l, except for Li and Be, for which the standard state is the Li BeF mixture. These values are for the LizBeF4 fuel salt but are taken here as a reasonable approxi- mation of the Li-Th-Be blanket salt. Any variation in the concentration of the ion in the salt can be accounted for by writing ' RT 1 where Ve 18 the activity coefficient in the salt; thus 80 = & at unit ion activity ()/SXS = 1), and Ye = 1 at low concentrations, so 80 is, as stated, the voltage at an extrapolated hypothetical unit mole fraction solution. Highly electropositive metals dissolved in bismuth do not form ideal solutions at all, even at low concentration; these are intermetallic compounds formed with substantial free energies. Never- theless, preserving the form of Eq. (2), we can write RT y X = _ — |pnmm 3 o " 75 "X ) 23¢. F. Baes, Jr., Thermodynamics, pp. 409—33, IAEA, Vienna, 1966. 151 Here y <<1 even in very dilute solutions. At this point an alternative choice of standard states is useful. In the extraction proceés we do not deal with pure metals but with amalgams exclusively, and we are concerned not with ac- tivities but with concentrations. If we write (3) in the form of (1) having ,Xm/XS , the measfifed quantity, as the independent variable, we find E=F —-—ln—-———ln-X—m. (39 Since we are dealing with solutions of constant composition (Li and Th in the salt) or dilute (all the others), all of the y’s may be taken as constant with composition. Then we can define an alternate standard potential 83: ' RT Ym SR CR— P , 0 0 e Y, f and suBstituting this into (3 ’)V we obtain (1), We see at once that 8; is the voltage of the half- cell when X = X_ . Here the standard states or hypothetical unit mole fraction solutions have th‘e_ properties cortesponding to infinite dilution. To evaluate the 8(1)’5 we need know 80 and y_/y, for one of the four metals and just the ratio /Xm/XS'for the others. This will enable us to construct an electromotive series for the system amalgam-salt which will ificlude protactinium, for which we do not know y or 80 . It should be streséed that . and 8; are alternate ways of accounting for the same phenomenon, the formation of intermetallic c.om“pourids with bismuth. | For lithium we have y = 9.8 x 10-5.24 Assuming that this salt acts like LizBeF4, we say that y = 2.2 [ extrapolating from (1)], which gives 8(1) (Li) = 1.803. Now we may use this to find the other Sé’s, which are shown in the accompanying table. ‘ Orry & 1b c M ~& (923 SR -&l ¥, Li 2.601. 1.80 9.8 x 1073 Th ' 1.772 1.47 4% 10”7 \ Pa _ 1.32 ~ U 1.368 ©1.28 4x10° 8—80 (v) at 923°K is half-cell potential for pure metal > ion at hypothetical unit mole fraction in Li, BeF, (except to 0.67 mole fraction for Li). b—gé (v) at 9239K is half-cell potential for amalgam - ion in LiZBeF4 at Xm = Xs . Cym is activity coefficient used to describe activity when referred to standard state described by 80. . - : Values of y_ can be estimated for Th and U. The salt concentration of uranium is low, so y, may be set equal to 1. A rather crude approximation to y_ for Th is an extrapolation of y_ for U in z 24Slightly changed from ORNL internal memorandum of D. G. Hill to W. R. Grimes, June 29, 1966. 152 LizBeF“',23 giving y_ = 7.5, Using these and the tabulated values of 80 , we find y_ as shown. Interpolation of values obtained at Brookhaven?® gives y_ for Th and U as 10" and 1.7 x 10~* respectively. An error of an order of magnitude in ym/ys for Li gives four orders of error in y for Th and U, so considering the approximations in the calculations, the agreement is reasonable. On the other hand, 8(1) changes by the same amount for all valences, so that the difference in 8(1), which is the quantity most directly related to the extraction efficiency, is not affected by errors in y_ . Therefore, although the absolute values should not be taken too seriously, they indicate fairly well the relative ease of extraction of the four elements into bismuth. Extraction of Rare Earths from Molten Fluorides into Molten Metals D. M. Moulton - W. R. Grimes F. F. Blankenship J. H. Shaffer As part of the program of removing lanthanide fission products from MSBR fuel solvent by re- ductive extraction into liquid metal, the study of the activity of lithium in bismuth amalgams has been continued. The method used has been the equilibration of amalgam-salt sys.tems with metallic beryllium, which is insoluble in bismuth and thus has a known unit activity. An analysis of'déta reported earlier?® reveals that, regardless of whether the amalgam was prepared by adding Li or IBe to the bismuth-salt system, only about two-thirds of the reductant metal appeared as amalgé- mated lithium. Because of the difficulty in accurately measuring very small changes in salt com- position, mass balances were made on the basis of the metal phase only, so that it was not pos- sible to tell whether the missing one-third was due to some systematic experimental error or represented an actual reduction process. | ' The use of barium as a reductant has allowed both salt and metal phases to be analyzed ac- curately. In this experiment Ba was added on top of the salt, through which it sank and dissolved in the bismuth. A similar procedure had been used for Be (which floats, however), while Li had been added through a dip leg directly to the bismuth. Here again it was found that only about 60% of the Ba could be accounted for by Ba and Li in the bismuth. Now, however, the lost barium all appeared in the salt phase, with the mass balance generally 95 to 105%. It is clear, therefore, that some unexpected reduction process was going on. The equilibrium constant for the reaction Ba(m) + 2Li* = Ba?" + 2Li(m) is found to be 0.016, corresponding to a AF® of +7.14 kcal/mole [ assuming y(Li+) - 1.4, y(BaZ+) =1]. 8; for the reaction Ba(m) = Bal?t + 2e~ is —1.63 v, using 8(1) of Li as —1.80. All of these values are based on the amalgamated metal at unit mole fraction as the standard and 25Brookhaven National Laboratory Nuclear Engineering Progress Report, July—September 1958, January— April 1959; Annual Report, July 1, 1956. - 26J. H. Shaffer et al., MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 142. 153 not on the pure element. By comparison, 80 for Be is —1.72,%3 and 80 for Ba is probably about —-2.2 to —2.3. ' The only'redugible material present in sufficient quantities to explain this behavior is beryl- lium ion. Excluding for the present the possibilities of Be* and Be® dissolved in the salt, metal- lic beryllium can be present only at unit activity. Thus, at equilibrium a lithium afialgam mole fraction of about 0.18 can be reached before any beryllium is formed, while Ba amalgam should not reduce Be2* at all. However, one can imagine the formation of Be either by the metallic Ba sink- ing through the salt or by Li amalgam at locally high concentration before it is thoroughly mixe;d. If, then, any Be so formed, or Be introduced as such, should be physically disposed in such a way that it no longer can be made to contact the bismuth'by agitation, it will never come to a true equilibrium. For instance, since Be floats, it may splash onto the container walls and not drain back. It is intended to test this explanation by trying to observe beryllium on the container walls, but this had not been done at the time of this writing. At least, it can be shown that the amount of lost reductant is roughly proportionJal to the amount added, which would be expected if this is the case. Unfortunately, some other old work, in which, when salt was contacted with an already equilibrated amalgam, more Li disappeared,?® is still not explained. Protactinium Studies in the High-Alpha Mol ten-Salt Laboratory C. J. Barton H. H. Stone Data given in the previous progress report?’ showed that a large fraction of the protactinium dissolved in molten LiF-ThF, (73-27 mole %) is converted by exposure to solid thorium to a form that does not bass through a sintered copper filter. However, more than half the reduced protac- tinium remained suspended in the molten salt. It was also réported that efforts to convert the dis- solved protactinium into a more readily recoverable form by electrolytic reduction had been un-' successfu}. Further studies of the reduction process and other potential recovery methods are briefly summarized here. Electrolytic Reduction. — Three experiments were performed in an effort to transfer protactinium from molten LiF-ThF4‘ (73-27 mole %) to liquid bismuth by electrolysis. The fraction of the prot- actinium reduced to a form that would not pass through a sintered copper filter varied from 0.5 to 0.95. In all three cases, -analysis of samples of the bismuth layer that had passed through sintered stainless steel filters indicated that only a few pé_rcent of the original protactinium content of the salt phase was dissolved in the molten metal. Because of lack of encouraging results, we have temporarily discontinued the electrolytic approach to protactinium recovery. Thorium Reduction in @ Bismuth-Containing Melt. — An experiment was performed to determine whether transfer of reduced.protactinium from the salt phase to bismuth would be improved by carrying out the reduction in the presence of dissolved bismuth. This experiment was carried out 27C. J. Barton and H. H. Stone, MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037‘, p. 156, 154 in a niobium-lined nickel pot equipped with a niobium dip leg. The 23!Pa, 233Pa.(in irradiated LiF-ThF4) , and 0.55 g of E’.i:!O3 were m@xed with LiF-ThF4 (73-27 mole %) and given a 30-min treatment with mixed HF and H,. After adding bismuth metal to the molten salt, the salt phase was exposed to a freshly cleaned thorium rod, approximately 1/4 in. in diameter, for three consec- utive 45-min periods. After the third exposure the protactinium concentration had been reduced to about 12% of the initial value. _ The crusts collected from the three thorium rods used in this experiment (13.0 g total) con- tained approximately 6.5 g of bismuth, far more than could be accounted for by the change in bismuth content of the filtered samples. The biSmuth. recovered as a separate phase weighed 51 g, as compared with the calculated value of 89 g. Further experimentation will be required to provide a satisfactory explanation of these observations. Thorium Reduction in the Presence of Iron Surface (Brillo Process). — Shaffer reported?8® that when a tracer quantity of *3*Pa precipitated from molten LiF-BeF -ThF , (73-2-25 mole %) in the presence of a large amount of iron surface in the form of steel wool (Brillo), most of the 233pga activity remained with the steel wool when the salt drained away. We have performed a series of experiments that confirm Shaffer’s finding with 2*'Pa concentrations in the approximate range 30 to 100 ppm of 231Pa. The experiments were performed in three steps. A weighed quantity of previously purified LiF-ThF, (73-27 mole %) was mixed with 231Pa and 233Pa in a nickel vessel, treated with mixed HF and H, to remove oxides, and then treated with H2 to reduce NiF2 to metal and PaF5 to PaF4. A weighed amount of grade 00 steel wool (0.068 m?/g surface area) was placed in a mild steel liner which was enclosed in a welded nickel vessel equipped with a stainless steel dip leg. A stream of purified hydrogen flowed through the vessel while it was heated to 800 to 820°C for about 4 hr and then cooled in flowing helium. The dip legs of the two vessels were then con- nected by a short transfer line and heated to about 650‘5C. Helium pressure applied to the salt - vessel transferred the salt to the steel-lined vessel, and helium continued to flow through both vessels to provide mixing of the molten salt. A thorium rod, attached to a l/g'ifl- nickel rod, was ' . inserted in the salt with its lower end about 1/4 in. above the bed of steel wool. After the thorium was exposed in the salt for the desired length of time, it was withdrawn, and a filtered sample of _ - the salt was obtained with a sintered copper filter stick. This procedure was repeated, usually with a clean thorium rod, until a gross gamma count of 233Pa in the filtered sample showed that most of the protactinium had been removed from solution. The vessel connections were then re- versed to transfer the salt back to the nickel vessel. Both vessels were cooled with helium flowing' through them. After it had reached room temperature, the steel-lined vessel was cut through with a small saw and the contents of the liner, consisting primarily of a ball of salt enclosing the steel wool, were removed and weighed. This material was crushed and, in some experiments, separated into mag- 28_]. H. Shaffer, ORNL intemal memorandum to W. R. Grimes, Nov. 15, 1965. 155 netic and nonmagnetic fractions by use of a small magnet. A sample of the unfiltered salt in the nickel vessel was also removed for analysis, Alpha pulse-height analysis was used to determine the 2°!Pa content of all samples, , We found that thorium exposure of 1.5 to 2.2 hr reduced 95 to 99% of the protactinium to a form that would not pass through the sintered copper filter medium. From 93 to 98% of the protactinium remained with the steel wool, along Iwith about 20% of the salt initially present in the experiment. A surprisingly large amount of iron was found in the filtered salt samples before the salt was ex- poséd to the steel wool. The source of this iron has not been determined, but analysis of the filtered samples indicated that, in general, the iron content of the salt decreased as the protactin- ium was reduced by exposure to thorium. This suggested the possibility that iton coprecipitated with the reduced protactinium was carrying the protactinium to the surface of the steel wool. We performed a pair of experiments to test this theory in which we attempted to hold all- the variables constant except the iron/protactinium ratio. The results were not conclusive because répeated analyses of the salt samples have shown such a large scatter in the iron values that it was dif- ficult to calculate reliable iron/protactinium ratios. We are currently conducting experiments with S9Fe tracer in an effort to establish the role of dissolved or colloidal iron in the retention of re- duced protactinium by steel wool and to test the reliability of the currently used analytical pro- cedure for the detemmination of iron in fluoride,samples. Additional experiments will be required to establish the important variables in the Brillo protactinium recovery process and to explore modifications that would make it more readily adaptéble to a large-scale process. Conclusions. — The results of our efforts to transfer protactinium from a molten breeder blanket mixture to liquid bismuth, either by electrolytic reduction or by exposure of the salt to thorium metal, have not been favorable in the small-scale glove-box experiments conducted to date, Nevertheles-s, the encofiraging results obtained in tracer-level experiments with protactinium and rare earths suggest that continued studies are desirable. Preliminary Study of the System LiF-ThF ,-PaF, C. J. Barton . H. H. Stone G. D. Brunton In the protactinium recovery studies described in the preceding section of this report, it has been generally assumed that protactinium is present in the LiF—ThF; (73 mole % LiF) melts as PaF4. This assumption has seemed plausible since the melts have, in every case, received a treatment with H, at temperatures near 650°C, and H, is known to reduce pure PaF _ to PaF at much lower temperatures, (The Pa*" state seems to be the lowest known in fluoride systems.) We have, however, conducted a few preliminary experiments to test this assumption and to see if the phase behavior of PaF , is similar to that of ThF in mixtures with LiF. ~ About 100 mg of .231PaF4 was prepared by evaporating a measured portion of purified stock solution (9 M in HF) to dryness in platinum and heating the residue to 600°C in flowing HF-H, mixture. Conversion to PaF, was confirmed by .weight and by the brown color of the material. A portion of this material was mixed with LiF.-ThF4 (73 mole % LiF) to yield a mix with 68 mole % - 156 LiF and 32 mole % (Th,Pa) F,. Another portion was mixed with LiF and the LiF-ThF , mixture to yield a mix with 73 mole % LiF and 27 mole % (Th,Pa) F4. Both mixtures were admixed with ammonium bifluoride (whose decomposition products on heating help to minimize possible hydrolysis), heated to 650°C, and cooled slowly. | Examination o-f the slowly cooled melts showed that segregation of PaF -rich phases from the bulk of the LiF-ThF material occurred in both cases. Material from the mixture with 68 mole % LiF is believed to be a solid solution of LiPan in LiThFS. One of the phases from the sample with 73 mole % LiF is believed, because of its similarity to the analogous uranium compound, to be Li4PaF8.‘ The PaF does not appear isomorphous with ThF ; the LiF-PaF, system may, in fact, be more like the LiF-UF , than the LiF-ThF , system. Itis obvious that study of the binary LiF-PaF system is needed before attempting further deductions conceming phase relations in the ternary system LiF-ThF4-PaF4. ‘A portion of the LiF-ThF -PaF mixture with 73 mole % LiF was transferred to a small thorium crucible and heated to 650°C in a helium atmosphere. Examination of the material with the polar- izing microscope revealed some Li , ThF_, but a large part of the mixture was in the form of opaque angular fragments, which are probably protactinium metal. X-ray examination will be required to confirm this conclusion. ' ~ 7.5 DEVELOPMENT AND EVALUATION OF ANALYTICAL METHODS FOR MOLTEN-SALT REACTORS Determinations of Oxide in MSRE Salts R. F. Apple J. M. Dale A. S. Meyer The analyses of oxide in radioactive salt samples from the MSRE are summarized in Table 7.12. Table 7.12. Oxide Cencentrations of MSRE Fuel-Salt Samples Oxide Concentration Sample Date Received ' (ppm) , FP-6-1 47-66 ' 49 FP-6-4 ‘ 4-14-66 | 55 FP-6-12 . 5-10-66 | 50 FP-6-18 5-25-66 47 FP-7-5 6-22-66 66 FP-7-9 7-4-66 59 } FP-7-13 7-15-66 ' 66 FP-7-16 7-22-66 56 FP-8-7 : 10-14-66 44 FP-9-2 11-9-66 44 157 The abrupt decrease in concentration between samples FP-7-16 and FP-8-7 coincided with the change from copper to nickel sampling ladles and méy reflect an oxide blank of about 15 ppm. The overall standard deviation of these results, 8 ppm absolute, includes any variation in the oxide content of the fuel in the reactor and in oxide contamination during sampling and transfer. The results of standardization with 5nO, samples?® performed during this period indicate that the pre- cision of measurement of the water evolved by hydrofluorination is about 3 to 4 ppm absolute. During the last week of December the oxide apparatus became inoperative. Because of hy- drogen reduction experiments being pe‘rformed in the same hot cell, it was not possible to enter the cell and investigate the equipment until early February. It was found that the electrolytic moisture- monitor cell was not functioning, and it was removed from the apparatus. With the current MSRE sampling schedules it should be possible to install a new cell with an improved coating®? ‘and re- sume oxide determinations of radioactive salt samples during the month of March. This replacement will be the first major maintenance required for the hot-cell apparatus since its installation in February 1966. The moisture-monitor cell had been in operation for several additional months during laboratory tests of the apparatus. This one-year operation represents a reasonable service life for a moisture-monitor cell, even under normal operating conditions in a nonradioactive environment. During the period when the hot-cell equipment was inoperable, the oxide development apparatus in Building 4500S was réactivated. Several samples of nonradioactive fuel and solvent salt from the second ORR molten-salt loop were analyzed for oxide content. The results of these analyses are summarized in Table 7.13. It is probable that the high oxide levels found in the third and fourth samples were in part due to the contamination during the brief period of exposure when the crushed samples were transferred to the hydrofluorinator. When possible, future samples from this source will be melted into a nickel ladle in a dry box. 294SR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 192. 30, . ’ . o Ibid., p. 193. - Table 7.13. Oxide Concentrations of Fuel and Solvent Salt Samples from the Secand ORR Molten-Salt Loop Salt Condition Sample Weight Oxide Concentration Sample Designation as Received (g) (ppm) 19-29-66 Pellets ' 10.4 200 10+17-66 Pellets 10.1 l 220 11-21-66 Crushed : 18.8 420 F-194 Crushed ' 18.4 820 Solvent salt batch No. 2 Fused into ladle 48.4 115 Solvent salt batch No. 17 Fused into ladle - 13.1 520 - 158 The increase in oxide between the last two samples represehts the oxide pickup when the flush salt was circulated in the loop. Determination of U3*/U** Ratios in Radiocactive Fuel by a Hydrogen Reduction Method J. M. Dale - R. F. Apple A. S. Meyer A proposed explanation of the unexpected distribution of certain fission products in the MSRE system is that the fuel had become sufficiently oxidizing to produce significant partial pressures of volatile fission product fluorides such as MoF , TeF, and RuF which then migrated into the graphite and the blanket gas. Because the accumulation of fission products in graphite is vitally important to breeder reactors, a method for the determination of U? fruet ratios in the radioactive fuel samples was needed. The possibility that a significant fraction of the iron and nickel is present in the fuel as colloidal metal.particles3! precluded any adaptation of the hydrogen evolution method for U’ * because these metallic components would also yield hydrogen on acidic dissolution. An alternate approach, suggested by C. F. Baes, is based on a transpiration method in which a sample of the molten fuel is sparged with hydrogen to reduce oxidized species according to the reaction n—um ME_ + H,— MFr.n + (n — m)HF , in which MF may be UF,, NiF, FeF ,, CrF,, or UF4 in order of their observed reduction po- tentials, The rate of production of HF is a function of the ratio of oxidized to reduced species in the melt. | Some components of the oxide apparatus3? were adapted for the transpiration measurement on ' is connected to a source of radioactive fuel samples. The inlet of a modified ‘‘hydrofluorinator’ thoroughly dried hydrogen and helium (see Fig. 7.11), and the outlet is connected to a heated manifold fitted with six liquid-nitrogen-cooled NaF traps. This arrangement permits the collection of the HF from four successive reduction s’teps with hydrogen and two blank spargings with helium on the same 50-g fuel sample. When the transpiration run is completed, the traps are disconnected from the manifold and removed from the hot cell for desorption and titration of the collected HF. Interpretation of these titrations in terms of U3*/U* ™ ratios in the samples is made by com- paring the HF yields for the reduction steps with those calculated for hypjothetica.l salt compo- sitions. If U*" is the only reducible species in the melt, a derivation based on a material balance and the hydrogen reduction equilibrium yields the rather simple relationship for U3 ¥ as a function of hydrogen exposure: 1 v 3+ Uln -IT—:E';'; - U =Bt: . (1)1 31MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 162. 321bid., pp. 154—62. in which U, udt 159 ORNL-DWG. 67-226 HYDROGENATOR LIID He NITROGEN TRAPS ACCESS AREA S HOT CELL N s © o N B HEATED MANIFOLD § (ORI O OINONONONO! N \ U U UUUU N NoF TRAPS FOR COLLECTING HF Mgfigugénongnr N \ Fig. 7.11. Diagram of the Hydrogenation Apparatus for Determination of u3*/u4* Rotios in MSRE Fuel. ’ are concentrations of total and trivalent uranium expressed as mole fractions H . . . + , . t is the reduction time from ““0’’ U®" concentration, min, 1/2 ) KP-H2 Vo SRT where 3+ U PH atm 1/2 U4 + P1/2 ’ ! H) P!_l = partial pressure of H2, atm, PH = partial pressure of HF, atm, Vs = purge flow, liters/min, S = quantity of fuel sample, moles, R = gas constant, liters atm (°K) ~! mole—1, T = temperature at which the sparge flow is measured, °K. When corrosion products (M].2 *) are present, the relationship assumes a different form as each corrosion product ion successively undergoes reduction to the metal: ~ in which a. = 1 a4 U-a U j=n K2 1 U anU-—U + Uln W + 4 E K2 —_— =Btn', (2) 3+ a =1 n U . 1 1+ (Ky /K) M2T1/2 J ) 160 sz * = concentration, mole fraction, of the jth corrosion product in order of reduction, at t =0, n t., min, is measured from the instant reduction of the nth corrosion product starts, K PHF M. = i M.2+ 1/2 p1/2 (M) Py The yield of HF from any reduction step is calculated from the initial and final concentrations { of U_3+ and corrosion products as follows: . j=n , ‘micromoles of HF ={ (U2F - U3") +2 ¥, [(M].”)i — M2M)| S x10°. ) ; i=1 The relationship of Egs. (1) and (2) is illustrated in Fig. 7.12', in which the yield of HF is plotted as a function of Bt for an oxidized fuel containing U* * and Fe2" in concentrations ap- proximating the elemental analyses of the MSRE fuel. On a logarithmic plot the accumulated HF yield follows a near-linear relationship [ Eq. (1)] derived for an iron-free melt until a critical U3t/U*? 1atio is reached and reduction of Fe2” starts. The reduction of Fezf’ then predominates, with the rate of change of U? * reduced in accordance with Eq. (2) until substantially all of the iron is reduced and the HF yield approaches that derived from uranium reduction alone. A computer program is being developed by the Mathematics Division to calculate HF yields for any preselected reduction steps on any melt composition. This will permit optimization of reduction steps for best distinction of U*/U*? ratios. In the absence of such a program it was possible to select four reduction steps on the criteria that the first two steps at 496°C would re- " duce at least 99.9% of the Fe2' and Ni2 ¥ without materially affecting the U3* concentration and that the last two steps at 596°C would reduce U™ without reducing Cr? *. The latter reductions thus serve as an indication of the ‘‘oxidation state’’ of the fuel. Hydrogen fluoride yields'frorr‘x sample FP-9-4 and from various hypothetical initial fuel compositions are plotted in Fig. 7.13. Broken lines connect calculated yield points, and a solid line connects the experimentally measured HF yields from the sample. The relatively low yields of HF in the first two steps are in agreement with earlier experimental evidence®! that only a small fraction of the iron and nickel is present in the ionic state. The high yields in the last two steps indicate that the fuel is in oxidizing or near- oxidizing condition. On the basis of the HF yield of step III and in consideration of estimated ex- perimental errors, a ratio of U3t/U*? of 0.000 to 0.001 was aséigned. Sample FP-10-25, taken after the addition of sufficient beryllium xfietal to increase the U3* /U*" ratio by 0.0037, was analyzed similarly, and the HF yield at step III corresponded to an applroximate U3*t/Uu*" ratio of 0.005. This represents. an increase of 0.004 to 0.005 over sample FP-9-4, in reasonable agreement with that calculated from the beryllium reduction. Hydrogen fluoride yields from sample FP-11-5, taken after a possible eiposure of the fuel to air, agreed within estimated experimental error with sample FP-10-25. | 161 ORNL - DWG. 67-122 1000 3 F INITIAL . _ CONDITION . B % 143 o +2 . T - No(f,oL(J)), %Fe2) MSRE FUEL SALT, 1.22 moles i ] : \ X5y =80x1073 1 \\ | Xspe=11x 1074 100 \ E 2 \ ] C \ N - \ : \ - \ T w he SAMPLE ; 2 10} f E s [ B(0,0) \ s X 3. X I c(o.oz,o.i)\:;_—:_,::_‘ / % ] n \‘1:':.._," / ! 7 - G0N0 — 4 / = I'// 3 ‘E // ; b " N - D(O.B, H2—>MFm+(n—m)HF, n where MF may be UF_, NiF , FeF , or UF .- The reaction is driven to the right by sparging the molten sample with hydrogen. The liberated HF is collected on NaF traps cooled by liquid nitro- gen. After the HF from the hydrogenation steps is collected, the traps are removed from the hot cell, and the HF is desorbed at 300°C and titrated. The oxidation-state measurements were made in conjunction with beryllium additions to the MSRE fuel salt. To date, three samples have been analyzed. 38]. P. Young, Gleb Mamantov, and F. L. Whiting, J. Phys. Chem. 71, in press. 39MSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 195. 165 ‘Sample Analyses. In addition to the oxidation-state éamples, several other special samples were analyzed. These included MSRE off-gas samples, beryllium addition samples, and six highly purified LiF - BeF, samples. For the past two years, beryllium determinations made on radioactive MSRE fuel-salt samples have exhibited a 2% positive bias from book values. In an attempt to locate the source of the bias, six highly purified LiF - BeF, samples were submitted to the High-Radiation-Level Analytical Laboratory. Three of the éamples were fused at 700°C to simulate the conditions of the fuel salt. - The results obtained on these samples are shown in Table 7.14. Table 7.14. Results Obtained on Analysis of Highly Purified Lithium-Beryllium Fluoride Samples Weight Percent Beryllium Sample - ‘ Difference (%) True ’ Found?® LB 1 3.05 3.07 0.6 LB2 6.98 7.01 | +0.43 LB 3 10.19 10.11 —-0.79 RT 17 13.76 13.74 ~0.15 RT 2° ¢ 15.13 15.10 —~0.20 RT 3° 18.06 17.95 —0.61 @yalues listed as found are average values calculated from approximately 20 determinations. bsamples fused at 700°C. Although the individual determinations exhibited a +2% spread, no significant bias is evident from the data shown in Table 7.14. In addition to the above results, determinations made on syn- Vth-etic solutions similar to dissolved MSRE fuel-salt samples show no appreciable bias in the method. From July 1, 1966, through December 31, 1966, 48 routine salt samples were analyzed as shown in Table 7.15. Several of the MSRE fuel-salt samples were submitted with silver and INOR-8 wires coiled around the stainless steel cable between the latch and ladle. The latch, wires, and cable were separated and prepared for radiochemical analyses. Quality Control Program A surhmary of the MSRE control results for the third and fourth quarters of 1966 is shown in Tables 7.16 and 7.17. The values shown in Tables 7.16 and 7.17 are a composite of the values obtained by four different groups of shift personnel. 166 Table 7.15. Analysis of Radioactive MSRE Fuel Samples from July 1, 1966, Through December 31, 1966 Number of Analyses Made RCA% Msa®? U Zr Cr Be F Fe. Ni Mo Li Prep Prep Oxide 38 38 38 38 38 38 38 2 38 38 8 6 4Radiochemical analysis, bMasS spectrographic analysis. CTwo 50-g oxide samples are being held for future analysis, and two were lost due to a malfunction in the oxide apparatus moisture monitor. ’ Table 7.16. Summary of Control Results for July, August, and September 1966 28 (%) Determination and Determinations . Bias Method Made Fixed " Found (%) Be photoneutron 21 5.0 2.75 +0.43 Cr amperometric 35 15.0 - 14.69 Fe spectrophotometric 33 15.0 6.65 Ni spectrophotometric 33 15.0 8.89 +5.21 U coulometric 24 1.0 1.90 Zr amperometric 26 5.0 4.47 Table 7.17. Summary of Control Results for October, November, and December 1966 Determination and Determinations 25 (%) l Bias Method Made Fixed Found (%) Be photoneutron 146 5.0 1.77 —0.28 Cr amperometric 86 : 15.0 8.06 - —4.33 Fe spectrophotometric 60 15.0 5.43 +2.28 Ni spectrophotometric 54 15.0 5.71 +1.80° U coulometric 172 \ 1.0 0.66 +0.36 Zr amperometric - 77 5.0 ’ 3.21 +1.35 8. Molten—Salt Convection Loops. in the ORR .H. C. Savage E. L. Compere J. M. Baker M. J. Kelly E. G. Bohlmann Irradiation of the first molten-salt convection loop experiment in the Qak Ridge Research Re- actor .beam hole HN-1 was terminated on August 8, 1966, after development of 1.1 x 10'8 fissions/cc (0.27% 235U bumup) in the 7LiF-].3eF2-ZrF4-UF4 (65.16-28.57-4.90-1.36 mole %) fuel. Average fuel power densities up to 105 w per cc of salt were attained in the fuel channels of the core of MSRE-grade graphite. . | : [ Successful operation of the major heating, cooling, tempetature control, and sampling systems was demonstrated; however, leaks developed in two of the four cooling systems. The experiment was terminated after radioactivity was detected in the secondary containment s3lzstem as a result of fuel leakage from a break in the sample line near the loop. Irradiation of a second loop, modified to eliminate causes of failures encountered in the first, was begun in January 1967. Fueled operation began January 30, 1967. Operation is continuing at an average core fuel power density estimated at 160 w per cc of salt. 8.1 OBJECTIVES AND DESCRIPTION The loop is designed to irradiate a representative molten-salt fuel circulating in contact with graphite and Hastelloy N at typical temperature differences and at a core power density of 200 w/cc. In patticular, it allows us to study the interaction of fission products with graphite, metal, and fuel and gas phases, and the chemistry of the fuel salt at high levels of burnup. Provisions for sampling and replacement of both gas and salt pemit conditions in the loop to be determined and to be altered during operation.! =4 1 MSR Program Semiann. Progr. Rept. Aug. 31, 1965, ORNI1.-3872, pp. 106—10. " 2MSR Program Semiann, Progr. Rept. Feb, 28, 1966, 0RNL-3§36, pp. 152-54, 3Reactor Chem. Div. Ann. Progr. Rept., Dec, 31, 1965, ORNL-3913, pp. 34—35. 4I'\’eactor Chem. Div. Ann. Progr, Rept, Jan. 31, 1965, ORNL-3789, pp. 4548, 167 168 8.2 FIRST LOOP EXPERIMENT In-Pile Irradiation Assembly The core of the first loop consisted of a 2-in.-diam by 6-in.-long cylinder of graphite obtained from MSRE stock. Eight vertical 1/4-in. holes for salt flow were bored through the core in an octag- onal pattern with centers 5/8 in. from the graphite center line. A horizontal gas separation tank con- nected the top of the core through a return line to the core bottom, completing the loop. These, and the core shell, were fabricated of Hastel.loy N. The heaters, and the cooling tubes in the core and return line, were embedded in $prayed-on nickel, as was the 12-ft sample tube leading from the loop to the sample station in the equipment chamber at the ORR shield face. Operations The loop was operated with MSRE solvent salt for 187 hr at Y-12, and several salt sa_mplés were taken. It was inserted in beam hole HN-1 of the ORR on June 9, 1966, and operated 1100 hr with solvent salt; during this period the equipment was calibrated and tested, and its performancé was evaluated. The loop was advanced to the position nearest the reactor lattice on July 21, and water injection into the air streams to the tubular core coolers and the jacket around the gas separation tank was tested. One of the two core coolers leaked and was plugged off. Water in- jection was discontinued until after uranium was added. On July 27, after sampling, eutectic 7LiF-UF4 (93% enriched) fuel salt was added to develop a uranium concentration of 1.36 mole %. At this time a capillary tube in the sample removal system broke, precluding further sampling. An associated in-leakage of air impelled solvent salt to a cold spot in the gas sample line, thereby plugging it. During subsequent operation the fission heating rate was determined. Water was released into the loop container during this period from what proved to be a leak in the cooling jacket around the gas separation tank. After a short reactor shutdown on August 8 to permit removal of water accu- -mulated in the container, the irradiation was resumed. That evening, release of substantial radio- activity into the loop container indicated fuel leakage. The lloop temperature was lowered to freeze the salt, and the loop was retracted to 2% flux. It was removed to hot cells for disassembly and examination on August 11, 1966. Chemical Analysis of Salt Samples of solvent salt taken prior to irradiation and after 1100 hr in-pile, and of irradiated fueled salt obtained after dismantling, were analyzed chemically and radiochemically. A sample of salt found between the metal core shell and the graphite was also analyzed. Results are given in Table 8.1 and discussed below. 169 Corrosion The level of corrosion products, particularly chromium and nickel, in the salt increased in the successive samples. This was possibly due to uptake of moisture by the solvent salt prior to loading, with consequent corrosion of the Hastelloy N. The corrosion appears to have occurred in the addition tank, since a sample which was taken directly from the addition tank, without entering the loop, showed similar levels of corrosion products. Fission Products Fission products were counted in a fuel sample after 110 days‘ of cooling, and concentrations are given below as a percentage of the amount produced, calculated on the basis of observed fis- sion heat (4.8 x 10!7 fissions/g). . , Cerium-144 and -141 (77, 64%) and 8°Zr (65%) were somewhat below the calculated production. Cesium-137 (41%) and ®°Sr (42%), both with noble-gas precursors of ~ 3 min half-life, were still lower and could thereby have been lost to the gas space or graphite voids. Tellurium-127 was present in 10% of the amount produced in the salt. Ruthenium-103 and -106, which were expected - to deposit on Hastelloy N surfaces, were not detected (<0.03%) in the salt. Nuclear Heat and Neutron Flux Nuclear heat was measured at various loop insertion positions by comparing electrical heat requirements under similar conditions with the reactor at zero and at full power. Reactor gamma heat with the loop fully inserted was 2900 w (with unfueled salt). With fuel containing 1.36 mole % uranium (93% enriched), fission heat in the fully inserted position was 5800 w. The corresponding overall average fission heat density was 80 w per cc of salt at 650°C, and in the graph'ite core the average fission heat density was 105 w per cc of fuel salt. The average effective thermal neutron flux in the salt was estimated independently from the nuclear heat, from activation of solvent salt zirconium, from cobalt monitors on the loop e);terior, _ and by neutron transport calculation. The results agreed well, ranging between 0.9 and 1.2 x 1013 -2 neutrons cm~? sec™ 1. Hot-Cell Examination of Components After separation from other parts of the package, the loop proper was kept for approximately three months in a furnace at 300°C to prevent fluorine evolution by fission product radiolysis of the salt. At this time it was removed for detailed examination. ‘ The tubular core cooler, type 304 stainless steel, was found to have broken entirely loose with- out ductility at its outlet end as it left the core near a tack weld to the core shell. Intergranular cracks originated on the outer circumference of the coiled tube. - Table 8.1. Loading and Samples from First In-Pile Molten-Salt Convection Loop (analyses as mg/g or mole %) "Li Be Zr vé F Cr Fe Ni Mo Composition as loaded Solvent salt . Composition as manufactured (mole %) b (64.78) (30.06) (5.16) (mg/g) 114.5 68.4 ° 118.9 698.2 ~™~0.020 ~0.,020 ~™0.100 Fuel, eutectic (mole %) (72.46) (27.54) (mg/z) 48.5 619.6 331.9 Fueled loop mixture (calcd) (mole %) (65.16) (28.57) (4.90) (1.36) (mg/E) 106.5 60.2 104.5 74.8 654.0 Hastelloy N — representative analysisc (mg/) 70.400 46.000 696.200 161.000 Analyses of loop samples (mg/g) Hours Hours of Sample Molten Radiation No, 120 0 1 (solvent salt) 98.0 66.5 119.5 683 . 0.310 - 0.275 0.137 n.d. 166 0 3 (solvent salt) 97.5 66.1 122.0 673 0.355 0.285 0.455 n.d. 1260 208 6 (solvent salt) 116.5 © 69.1 120.0 704 0.670d 0.092 0.540 n.d. After shutdown, 1578 329 9 (fueled mixture) 113 58.5 99.4 71.5 0.780 0.258 0.555 <0.015 Graphite—Hastelloy N annulus specimen S-1 105 56.1 101.5 71.7 3.250 3.800 25,300 4,740 292.96% 235y, bj. H. Shaffer, MSR Program Semiann. Progr. Rept. Feb, 28, 1965, ORNL-3812, pp. 150--52. . “Heat SP-19 for comparison, Sler activation gave a Cr concentration of 0,990 mg per g of salt. 0L1 171 The cooling jacket on the gas separation tank leaked at a weld. The fuel leak resulted from a nonductile break in the Hastelloy. N sample line tubing near the attachment to the core bottom. The sprayed nickel was also cracked in this region. ' Fuel salt in the form of a scale a few mils thick was found on the interior of the core shell, between it and the closely fitting graphite core. The analysis in Table 8.1 indicates that the scale is a mixture of fuel salt and Hastelloy N (probably metal debris from cutup operation). Hot-cell metallurgical examination of the interior surfaces of the Hastelloy N comprising the core bottom and core shell wall revealed no evidence of any interaction with salt or carbon, or other change, 8.3 EVALUATION OF SYSTEM PERFORMANCE Heaters _The molten-salt loop package used 21 continuous or intermittent heaters, all 1/8--in.-OD Inconel sheathed, MgO insulated, with Nichrome V elements designed for continuous operation at 870°C. No failures occurred. - Coolers The heat removal rate of the loop coolers was entirely adequate to remove the 8.8 kw of fis- sion and gamma heat, even after the loss (described earlier) of one of the two cooling coils around the loop core section. The air plus water-injection technique appears adequate and resporfsive. The use of water injection was not necessarily the cause of failure of the two cooling units, but only made the failures evident. Temperature Control The response of the heating and cooling systems to rapid changes in the nuclear heat could be tested only under full fission conditions in-pileé. Since this was regarded as important, reactor set- back tests were conducted. Temperature control system response was adequate to maintain the salt molten during a reactor setback with resultant loss of 8.8 kw of nuclear heat, and to return the loop to nomal operating condition during a rapid (11-min) return to full power, Sampling and Additic‘m The sampling and addition system and operating procedures were adequate to permit several ~additions and removals of molten salt while operating the loop in-pile, and to transport shielded samples under inert gas atmosphere to the analytical laboratory. A broken capillary connecting tube prevented additional sampling. 172 Salt Circulation Convective sal-t citculation, at rates of 5 to 10 cc/min, was achieved by causing the return line to operate at temperatures below the core temperature. Flow stoppagés occurred from time to time. These were attributed to bubble formation resulting from different solubility of the argon.cover gas at the varied temperatures around the lo,op'. Salt flow was reestablished by evacuation ~a'nd readdi- tion of cover gas. Loss of flow had no adverse effect on loop operation. 8.4 SECOND IN-PILE IRRADIATION ASSEMBLY A second in-pile molten-salt convection loop,l essentially identical to the first convection loop experiment,® was constructed, and in-pile irradiation began eatly in January 1967. Problems en- countered in the first convection loop experiment and information from subsequent postirradiation hot-cell examinations led to modifications to the second.loop to eliminate these problems. The coolant tubes embedded in sprayed-on nickel around the core section are now 1/“—-in.-OD X 0.035-in.-wall Inconel tubing instead of the 14—in.-0D x 0.035-in.-wall type 304 stainless 'steel used on the first loop. The stainless steel tubing should have been entirely adequate for the service, and no reason for the failure observed has been uncovered; but Inconel is the preferred material for exposure to the high-ten;perature steam-air mixture (~400°C) generated when air-water mixtures are used as coolant. Since the rupture of one of the core coolant tubes occfirred adjacent to a point where the tube was tack welded to the core wall, the tack weld waé eliminated in favor of a mechan- ical strap attachment. An expansion loop to relieve stresses has also been included in each of the coolant outlet lines. A mockup of the modified cooling coil was operated at temperature with air- water mixtures for more than 400 hr, ir;cluding 120 thermal shock cycles (600 to 310°C), with no sign of difficulty. Thermal cycling occurs during a reactor setback and startup, and it is estimated that no more than about 20 such thermal cycles will occur during a year of op‘eration. The two failures which occurred in the capillary tubing (0.100 in. OD x 0.050 in. ID) used in the salt transfer system appear to have resulted from excessive mechanical stress. Consequently, the wall thickness of this line has Been increased to 0.050 in., and additional mechan}cal support hqs been added such that there is now no part of the salt sample line which is unsupported — con- trary to the case in the first loop assembly. The cooling jacket of 1/16-in.-thick stainless steel surrounding the reservoir tank has been re- placed by an Inconel tube wrapped around the outside of the tank and attached by means of sprayed- on nickel metal, as is done on the core section and cold leg. Also, provisions for uée of an air- water mixture as coolant have been added, since it was found that air alone did not provide suffi- cient cooling in the first experiment. - _ Continuous salt circulation by thermal convection was not maintained in the first experiment. It was concluded that loss of circulation was caused by gas‘acc‘umulation in the top 6f the core section. Accordingly, the salt flow channels at the top and bottom of the eight 1/4-in. holes .for salt flow in the graphite core were redesigned to provide better flow conditions at the inlets and exits of 173 - the vertical holes. Further, the top and bottom of the core section, horizontally oriented on the first loop, were inclined at 5° to minimize trapping of gas. Operation The loop was placed under irradiation in beam hole HN-1 of the ORR on January 16, 1967, after 325 hr of satisfactory preirradiation test operation. - Preirradiationv.operat‘i'on included COmpié"te re- moval and replacement of an original flush charge of solvent salt, providing a good demonstration of the operability of the sampling and addition éystem The loop operated satisfactorily under irradiation at 650°C with MSRE solvent salt (’LiF- BeF - ZiF,, 64.8, 30.1, 5.1 mole %) during checkout and calibration measurements, On _]anuary 27, 1967, samples of irradiated solvent salt were taken and submitted for chemical and radiochemical analyses. ' . _ On January 30, 1967, 7LiF—UF4 (63-27 mole %) eutectic fuel (93% 23°U) was added, along with additional solvent salt, resulting in a fuel composition of 7LiF~BeF2-ZrF4-UF4 of 65.26-28.17- 4.84-1.73 mole %. Irradiation began January 30, 1967, with the experiment in a relatively low flux position. Convective circulation was established a't an estimated rate of 30 to 50 cc/min (™~ 2 min circuit time), and the loop was operated in several flux positions. Controls wete tested and ad- justed, and nuclear heat was determined as a function of position. Samples of fuel salt were with- drawn February 6, 1967. Operation in the fully inserted, highest flux position was achieved February 21, 1967, and has since been maintained. The average flux effective in all the salt is estiméted to be slightly above 1 x 1013, The average power density in the fuel salt in the core is estimated to be about 160 w/cc. The total fission heat genera{ion is about 9 kw; with a gamma heat of about 4 kw, the system heat generation is about 13 kw. The loop is operating satisfactorily. * Sampling and addition systems have operated reliably and without difficulty in all cases. It is anticipated that additional samples will be taken and fuel and solvent salt additions made from time to time during the course of the experiment. Part 3. Breeder Reactor Design Studies 9. Molten-Salt Breeder Reactor Design Studies E. S. Bettis C. E. Bettis D. A. Dyslin G. H. Llewellyn W. Terry R. J. Braatz H. F. Kerr T. W. Pickel L. V. Wilson 9.1 GENERAL The Breeder Reactor designnl"3 has been considerably narrowed in scope during this period. A modular concept has been adopted as the basic design, and our effort has been concentrated on more clearly defining the features of one 250-Mw (electrical) module. No work has been done on overall plant design during this period, and no attention has been given to the steam system. * With minor exceptions, all design effort has been concemed with components and layout within the reactor cell. . A layofit was made of the drain tanks with interconnecting piping for all three salt circuits. Figure 9.1 is a plan view of the general arrangement. This preliminary lay- out was made to gef an idea of the magnitude of the ‘‘tank farm’’ problem, and it appears that a satisfactory solution t§ this question will not be overly. difficult. Again, in order to get a preliminary evaluation of the nature of the communication problem between reactor plant and chemical processing plant, one method of effecting this intercon- nection was developed. Basically this plan involves a batch transfer between piants, the batches being moved between the two plants by gas pressure. Transfer to and from.the chem- ical ‘plant is in control of the reactor plant operator. Although the system of communication between reactor and chemical plant has not been worked out in any detail, it appears that the communication scheme can be done with certainty and safety. We believe that the optimization of the reactor is now accurate enough to warrant serious detailing of the désign. .A basic concept of fuel cell, blanket, plenum chambers, and reflector IMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 207. 2Design Studies of 1000-Mw (e) Molten-Salt Breeder Reactors, ORNL-3996. I 3MSR Program Semiann. Progr. Rept. Feb. 28, 1966, ORNL-3936, p. 172. 174 REHEATERS STEAM GENERATORS—|—— 34 ft INSULATION COOLANT PUMP DUMP TANKS 2nd SALT ——— BLANKET FUEL~«\::::::::: o) 43 L2 201t K o7 i1 21 BLANKET HEAT EXCHANGER AND PUMP — | PRIMARY HEAT EXCHANGER AND PUMP ———— 7 4t 8 fi STEAM PIPING 1133 AXA3 I W) ORNL-DWG 67-3082A HX = HEAT EXCHANGER Fig. 9.1. Reactor and Steam Cell Plan. OFF-GAS 34 ft T e, ST e e e e eeror oy LUSH SALT STORAGE 4 { DECONTAMINATIO ________ AND 16 ft &in. STORAGE : b i = o 211 Y P I - AL T sy AN R FUEL D g P S S PROCESSING ! L@ | DECONTAMINATION i 1 s | . TN ol 2 T ¢ A 16 f1 &in. RIS =t ST STORAGE AR : |: ’ (I A :I 1 1.1 [ LI —_— i ) : . REACTOR ¢ ! | ‘\ ," I : S : 14 ft }:—kwA419ft ‘ 18 ft ‘ ot 8 ff et 22— 4 ft —= L*—4ft —*J 4 ' 157 ft - ' I nlAll ' "B".‘__ SL1 176 thickness for a reactor with an average power density of approximately 40 kw/liter has reached a firm status. It is this module which will receive detailed attention. For comparison/purposes a module of equal power but having twice the average powerdensity was calculated. Table 9.1 shows the comparison between these two reactors regarding performance and economy. The blanket and fuel heat exchangers are also considered to be in a sufficiently firm design to justify detailing of these components. Both heat exchangers have undergone significant changes in physical arrangement and will be discussed in some detail later. Table 9.1. Comparison of the Effects of Core Power Density on the Characteristics of a 250-Mw (Electrical) Molten-Salt Breeder Reactor Average core power density, kw/liter . 78 : 59 Power, Mw 556 _ 556 Vessel diameter, ft | ' 11.4 12 Vessel height, ft \ ~12 17 Core diameter, ft 6.34 8 Core height, ft 8 10 Core volume, ft> 252 503 F-raction of fuel in core - 0.164 0.165 Fralction of blanket in core 0.05 0.06 Fraction of graphite in core 0.786 0.775 ] Blanket thickness, ft 2 . 1.5 Fraction of salt in blanket 0.65 - 0.60 Breeding ratio ' 1.06 1.07 Fuel yield, %/year 6.79 6.02 Fuel cycle cost, miils/kwhr .0.42 0.43 Fissile inventory, kg - 175 218 Fertile inventory, 1000 kg | 41 ' 43 Specific power, Mw (themal)/kg 3.18 2.55 Number of core elements 336 336 Velocity of fuel in core, fps 9.7 6 Average flux, e§ 100 kev ‘ . 2.9 x 1014 1.5 x 1014 Fuel volume, ft3 Reactor core - 41.3 83.0 " Plena 10.0 24.0 Entrance nipples 5.6 8.5 Heat exchangers and piping ' 102.0 105.0 Processing 10.5 8.8 Total : 169.4 231.3, Radial peak average flux ratio 1.58 Axial peak average flux ratio 1.51 177 The arrangement of components in the reactor cell as previously repbrted posed some very difficult stress problems. A new arrangement of these components abpears to have removed or greatly lessened these problems. Analyses are currently in progress which will indicate whether or not further changes are required. Operational expériencé with the MSRE has indicated clearly that the handling of fission gases from a molten-salt reactor requires very careffil design. The gas system for the breeder reactor is a most important part of the plant, and a start has been made in the design of this system. A flowsheet which seems to describe this system has been developed. Also, analyt- ical studies of gas stripping efficiencies and parameters which determine or define the nature of the gas sparging system have been initiated. ~ Associated with the gas handling prob\lem is the matter of draining the reactor in the event of a pump stoppage. Modifications in the primary heat exchanger indicated by the sparge gas handling and core drainage influenced us to put an overflow line from the sumpA tank of the pump into a fuel dump tank. These c;hanges in the primary heat exchanger have been rather extensive, although the basic heat exchanger remains very much as the original concept. 9.2 FLOWSHEET Since the reference design is based on the modular reactor concept, the flowsheet of Fig. 9.2 is for one module. The number of the various components in this module is shown. The steam system is shown very sketchily for the 1000-Mw (electrical) plant, and an indication of a steam header fed by the four modules is shown. It will be noted that the maximum coolant-salt pressure is somewhat higher than has been reported previously. This pressure occurs at the discharge of the coolant pump, and, of the 260 psi at this point, 110 psi comes from a gas overpressure at the suction side of the coolant pump. ~ We believe it to be imperative that, in the event of a failure in any part of the salt systems, the pressures should be such that blanket salt would flow into fuel salt, and coolant salt would flow into blanket and fuel salt, depending on the location of the failure. This made it necessary to put a high bias pressure on the coolant salt. The pressures shown on the flow diagram repre- sent design-point pressures and take %nto account static heads and dynamic pressure drops in all cases. In previous flowsheets a separate coolant pump circulated salt through the reheater. We now bypass the discharge of the main coolant pump through a salt-flow regulating valve to the re- heater. This valve is little more than a variable orifice, and we believe the use of such a valve is justified. PROCE SSING FUEL SALT HEAT EXCHANGER AND PUMP FUEL 4 f1t¥day REACTCR J L ——— BLANKET SALT HEAT EXCHANGER AND PUMP COOLANT BLANKET SALT SALT PUMP PROCESSING | | ——— —( FUEL SALT DRAIN TANKS POINT ft3/sec AR O0ROEOEE FUEL psig TEMP (°F) 250 13 250 9 250 137 250 110 25.0 50 25.0 31 250 18 BLANKET 43 125 43 8.0 43 1110 43 200 4.3 145 1300 1300 1300 1000 1000 1000 1250 1250 1250 1150 1130 ' t (% E \BOH_ER SUPERHEATER POINT f1¥/sec BlEBEHEEBGEE COOLANT psig TEMP [°F) 375 375 37.5 5.0 5.0 8.1 8.1 37.5 375 375 37.5 BLANKET SALT COOLANT SALT DRAIN TANK DRAIN TANK 122 1o 260 208 212 252 194 203 198 161 138 1125 1125 1125 125 850 1125 850 850 850 1111 111 PRARDBIRIPODIRDD REHEATERS STEAM 2.52 10.07 7.15 2.92 2.92 513 5.13 1.28 1.28 513 5.00 4.30 7.16 1007 1007 2.52 Fig. 9.2, Module Flowsheet. 3600 3600 3515 3515 3500 600 570 570 540 540 V72 15inHg 3500 3475 3800 3800 REHEAT STEAM POINT 10% Ib/hr psia TEMP (F) 1000 1000 1000 1000 866 552 650 650 1000 1000 706 92 551 695 700 700 ORNL-DWG 67-3081A FEED WATER SYSTEMS PERFORMANCE DATA 1000 Mwle} GENERATION PLANT NET OUTPUT 1000 Mwle) GROSS GENERATION 1034.9 Mw(e) BOILER FEED PUMPS 29.4 Mw BOOSTER PUMPS 9.2 Mw(e) STATION AUXILIARY 25.7 Mw(e) REACTOR HEAT INPUT 2225 Mwlt) NET HEAT RATE 7601 Btu/kwhr NET EFFICIENCY 449 % LEGEND FUEL _— BLANKET = COOLANT STEAM @~ ————— —n—— WATER — FREEZE VALVE DATA POINT X 8L1 179 9.3 REACTOR CELL COMPONENT ARRANGEMENT Because of the relatively high melting point of the blanket salt, it is desirable to connect the coolant circuits of the fuel heat exchafiger and the blanket heat exchanger in series. Coolant salt leaving the fuel heat exchanger at a temperature of 1111°F enters the blanket heat exchanger, where it picks up 14°F to attain the design-point temperature of 1125°F. The coolant-salt line, carrying 37.5 ft3/sec, is a large pipe (™20 in. in diameter) an.d therefore rigidly connects these two heat exchangers. Since the blanket heat exchanger is located high in the reactor cell while the fuel heat exchanger must be located below the bottom of the reactor vessel, there is a shift in the center of gravity of the coupled heat exchangers when they are filled with salt. In the original layout of the reactor cell which had the reactor, fuel heat exchanger, and blanket heat exchanger in a triangular array, a twisting moment resulted about the axis of the fuel connection between the reactor and the fuel heat exchanéer. Also, the method of suspen- sion mounting of the heat exchangers from constant load cell s could not provide accommodation for the differential thermal expansion of the system under all conditions. , A new layout of the reactor cell shown in Fig. 9.3 was made. In this configuration the three components, reactor, fuel heat exchanger, and blanket heat exchanger, are mounted in line. The coolant lines connecting the fuel and the blanket exchanger are made concentric, and the shortest possible spacing between exchangers is used. The connection from fuel heat e€xchanger to re- actor is as short as possible, using concentric piping for this as we have been doing previously. The connection from the blanket heat exchanger to the reactor vessel has been changed. In- . stead of the short concentric pipe used previously, the inlet and outlet lines from reactor vessel to blanket exchanger are run independently. These lines are purposely made long and flexible by running them through four 90°bends. Thus we intend to decouple the interconnection between the reactor and the blanket heat exchanger by use of flexible piping. The three components, reactor, fuel exchanger, and blanket exchanger, are rigidly inter- connected then by only one link for each; the reactor connects by the concentric fuel line to the fuel exchanger, which connects by one concentric coolant line to the blanket exchanger. At de- sign point these connecting lines are very near the ambient cell temperature of 1000°F. The fuel line is actually at 1000°F, and the coolant line is very near 850°F. With the three cell compo- nents in line, a different plan of sfipport is now proposed. Let us point out that this arrange- ment has not yet been analyzed for stresses, so we cannot be sure it is satisfactory in its present conception. ' The plan now involves mounting a stainless steel bed plate in the reactor cell in such a way as to be tied at one end but free to expand with cell temperature. This bed plate is located horizontally below the bottom of the reactor vessel and is mounted through necessary insulated supports to the load-bearing concrete of the cell. The reactor vessel and both heat exchahgers are mounted rigidly to the plate as shown in Fig, 9.4. We believe that this general arrangement will work and that any overstressing found by analysis of the system can be-corrected without departing drastically from the concept. ORNL-DWG 67-3083A BLANKET HEAT EXCHANGER FUEL AND BLANKET PUMP DRIVE MOTORS COOLANT SALT PUMP DUMP TANKS STEAM GENERATORS 8ft Qin. REHEATERS STEAM PIPING 58ft Oin. HEATER Y PRIMARY HEAT EXCHANGER AND PUMP Fig. 9.3. Reactor and Steam Cell Elevation. 081 a [} ORNL-DWG 67-33644 ' 3714t OQin. | ' 22 f1 Qin. ' 12+ Oin. in. 6t Oin. ~=— 15t 10in. ’ N— A : N——" ! N ' - \ BL. HEAT REACTOR EXCHANGER 4%t Qin. (TYP) PRIMARY HEAT EXCHANGER ditatihl :l\‘h-“ L'} I FIXED SUPPORT Il —]_———INSULATION 101t 6in. 141t &in. Fig. 9.4. Reactor Cell Mounting. 181 182 Vertical growth of all components can take place independently, and the only accommodation needed is a low-temperature sealing bellows at each pump-drive penetration through the cell membrane. Any control-rod drive mechanism would involve a similar bellows seal. The treatment of the reactor cell has not received any attention, but it is certain that some formof radiation shield must cover the structural concrete and this shield must have cooling. Inside the radiation shield, between it and the furnace of the cell, there must be some thermal insulation. - 9.4 COMPONENT DESIGN Reactor Yessel The reactor is shown in the elevation drawing of Fig. 9.5. The reactor vessel has a cy- lindrical body 12 ft in diameter and approximately 17 ft high with dished heads on top and bottom. A heavy supporting ring essentially the diameter of the core carries legs which weld to the bed plate and provide the mechanical mounting for the vessel. From the center of the bottom head a concentric fuel line communicates with the fuel heat exchanger. This fuel line has an outside diameter of 24 in. with a concentric retumn liné 16 in. in outside diameter, Fuel flows in these lines at a rate of approximately 18 fps.- Inside the reactor vessel are dished heads forming plenum chambers for distribution of the fuel cells which rise frfom these heads. These plenum chambers are removable from the reactor vessel, being held in place by a flanged mounting ring equipped with clamps. This flange has not been completely designed, but we are considering the use of flat graphite gaskets to pre- vent bonding of the flange surfaces under the high temperature and rather,high loading pressure obtained at desigfi conditions. The inner head communicates with the inner concentric line through a slip joinf which permits movement of the head relative to the pipe. Such movement results from thermal expansion and must be accommodated. In addition, the slip joint allows removal of the core when the flange seal is broken. The core of the reactor is made up of 336 cylindrical graphite fuel cells mounted as close together as tolerances permit to form essentially a cylindrical array approximately 8.3 ft in diameter. The graphite cells are extruded cylinders with center holes 11/2 in, in diameter, surrounded at 120° angles with three 7/a—irl.-diam holes. At the top of each cell there is a graphite cap méchined to provide a smooth communication between the four holes of the cell. Figure 9.6 shows the arrangement of one of these cells. At the bottom of a cell, the graphite cylinder is joined to a transition nipple. This join- ing is done by a combination of a short threaded engagement for mechanical attachment and a rather longer brazed joint for leak-tightness. " The connection to this nipple is a matter of development, and work is proceeding to establish the best practice for effecting this con- nection. The lower end of the nipple is threaded to screw into tapped holes in the upper curved head. 183 ORNL-DWG 67-3078A — 12 ft Oin. DIAM. ; CORE: 8 ft Oin. DIAM x 10 ft Oin. HIGH 336 FUEL CELLS -CONTROL TUBE PURGE ] PLENUM f T0 WOy BLANKET . ‘ X . HEAT 1 AL EXCHANGER 10/9¢0 :.\fl N \. -1 3 . N 4.) . \ g 4-in.-DIAM U H GRAPHITE ® SPHERES—-__i% , | (= T\ 2 7% z 6-in. .0 £k & GRAPHITE % o2 o REFLECTOR—4{// =3 = & © .REACTOR VESSEL _ r 1 1 l / \ ! =Ty | : V. o Dy M ......... ’ FUEL SALT/ i PLENUMS \ TO FUEL HEAT EXCHANGER -Fig. 9.5. The 250-Mw (Electrical) Reactor. 184 {ft 6in. —-‘_fi"——\_——_—_r-_‘ VAN rd LSS Sy ’ L ,(S’/ //// o’ o // //§ \ / " £ oS .7 # ‘ PR ~ // /\ \\\ /—\‘ ,{-\ '\\ N ~ NN EAN N\ N R AL R 101t Qin. (REACTOR) 5.00Qin, DIAM fi’ N N N '\ M § " \ ~ N, N X \ ~, N \ NIR N N N 1ft &in ’L,L’f///;\’ 1ft Qin. ’I’ d N \ By CRNL-DWG 67-267TA 2 ?/3 in. 1Y in. DIAM 3-HOLES, Yg in. DIAM GRAPHITE TO METAL BRAZE 1%gin, OD X 1¥4in, 1D 1% in. OD X 1¥%gin. 1D FUEL INLET PLENUM—/ N NN NN LD TAPERED THREAD 1 -/ “m———— FUEL QUTLET PLENUM Fig. 9.6. Fuel Cell Element. 185 From smaller holes in the inner head, concentric with the upper one, smaller nipples ex- tend through the upper head. These nipples are also attached to the inner head by a pipe thread, each one being screwed into place before attaching the fuel cells above. As a fuel cell is put into place, the 11/2-in. center hole is slipped over the inner nipple.’ This engagement is a slip joint and allows for screwing the cell into place, and in operation it permits the necessary differential expansion between the cell and the inner nipple. After all cells are in pléce, the entire core can be leak-tested, the graphite-to-metal transition joints having previously been tested individually. When the assembled core has been tested, it can be lowered into the reactor vessel and the flanged. joint can be clamped. Just inside the reactor vessel there is a graphite reflector 6 in. thick. This reflector is only on the vertical walls; there is none below or above the core. It is made of rectangular- shaped pieces with mortised joints to form a self-supporting wall resting on a ledge at the bottom of the vessel. The volume between the reflector and the vessel provides a flow passage for the blanket éalt as it discharges through holes in the bottom of the distribution ring located near the top of the vessel. We are considering the use of 4-in.-diam graphite balls in the region between the core and the reflector. These balls have holes in them so that graphite occupi.es 40% of the volume in this ll/z-ft—thick region. A retainer of perforated metal prevents the balls, which float in the blanket salt, from crowding into the top head of the reactor vessel. Instead, by floating up against this retainer, a measure of support is afforded the core structure, the floating balls tending to exert a considerable force on the fuel cells. With this core configuration a two-fluid reactor is achieved. Fuel salt enters the outer concentric pipe at the bottom of the reactor. It flows out to the outer radius of the core and in between the upper inner dished heads, feeding salt into the nipples emerging from the upper head. Fuel passes up through the three 7/8-in. holes of each fuel cell and down through the center 11/2—in_. hole through the inner pipe nipple into the collecting plenum. From there it flows through the slip joint into the inner concentric fuel line to the pump suction at the center of the fuel heat exchanger. The blanket salt flows into a distributor at the top of the reactor, and, emerging through holes in the bottom of this header, it flows behind the graphite reflector to the bottom of the reactor. There it flows up through the balls in the blanket annulus and out of the suction line near the top of the reactor. Some of the blanket salt fills the interstices betweén the fuel nipples and the fuel cells. Circulation of the blanket salt is governed by thermal convection. Heat generation and tem- perature calculations are in progress to determine whether this flow is adequate. - The center location for a fuel cell is left vacant. This is for a control rod. In Fig. 9.6 this control rod is represen\ted as a hollow graphite cylinder 5 in. in outside diameter, 4 in. in inside diameter, and equipped with a gas inlet at the top. By controlling the gas pressure, 186 blanket salt can be positioned at any height within this cylinder, and this blanket volume con- trols the reactivity of the core. Calculations indicate a change of 1.8% in reactivity between the full and empty condition of this cylinder. It may be that, instead of a hollow cylinder, a graphite rod 5 in. or less in diameter will ‘be used to displace blanket salt from this position in the core. Effective control can be ob- tained by driving this rod in a more conventional way. In the event of failure of the drive, | the rod would be lifted out of the core by the more dense blanket, thereby reducing the re- activity. ‘ | The reactor has been optimized as far as nuclear performance is concemed. “There is no very sensitive parame;er, and rather large changes in all dimensions can be tolerated without changing the characteristics materially. Table 9.1 gives the calculated performance of this reactor. This core volume is 503 ft3, with a fuel volume in the core of 83 ft3. This gives an average power density in the core of 39 kw/liter. _ For comparison, the performance of a reactor having a power density of 78 kw/liter is given in Table 9.2. This reactor -has half the volume in the core and would give somewhat better i)erformance. At present we do not know just what the limit is on graphite radiation tolerance. The smaller reactor, whatever the graphite limit, would be expected to have approximately half the core life of the larger one, which is used as the reference design. Some general comments can be made about the reactor as it is currently conceived. As we stated above, the nuclear petformance is so insensitive to parameters that the final reactor design is likely to be close to the present concept. We would prefer to have the core removable in modules rather than having to remove i(t in one piece. This is particularly desirable if we have to maintain the lower power density and if it becomes desirable to make reactor modules larger than the 556-Mw (thefinal) size with which we are concerned at present. We do not yet have what we consider an acceptable way of making these core submodules, but work is con- tinuing to try to develop one. The thickfiess of the heads has not received rigorous attention. At present the top or load- bearing head is 1 in. thick. This is thicker than the load seems to require but is used to give an adequate thickness for tapping for the nipples. The analysis of all stresses in these heads is being made but has not yet been completed. The same is true for the pressure drops. The top of the reactor as shown has a welded joint which can be ground off for removal and rewelded when a new core has been installed. Probably this access joint will Be moved up outside at least one layer of shielding in the final core design. In this case some of the shield- ing blocks would be supported by the top of the reactor vessel. The closure point for the re- actor would, by this means, be more accessible. Fuel Heat Exchanger The heat transfer work which had been done for the Breeder Reactor study was reevaluated. This was made necessary by the results obtained from the MSRE.. It was found that the salt 187 Table 9.2. Fuel Heat Exchanger Data Sheet Fuel-coolant-salt ?xchanger, counterflow, two-shell pass, two-tube pass with donut-type baffles Number required - Rate of heat transfer, Mw Rate of heat transfer, Btu/hr Shell side Hot fluid or cold fluid Entrance temperature, °F Exit temperature, °F Entrance pressure, psi Exit pressure, psi AP across exchanger, psi Mass flow rate, lb/hr Tube side Hot fluid or cold fluid Entrance temperature, °F Exit temperature, F Entrance pressure, psi Exit pressure, psi AP across exchanger, psi Mass flow rate, lb/hr Tube material - Tube OD, in. Tube thickness, in. Tube length, tube sheet to tube sheet, ft Shell material Shell thickness, in. lShell ID, in. Tube sheet material Tube sheet thickness, in. Number of tubes Pitch of tubes, in. Type of baffle Number of baffles Il 1 528.5 1.8046 x 10° Cold (coolant salt) 850 ' 1111 198 161 37 1,685 x 107 Hot (fuel salt) 1300 1000 137 50 87 ‘ 1.093 x 107 Hastelloy N 0.375 0.035 Inner annulus, 13.53 Outer annulus, 14.50 ‘Hastelloy N 1 66.7 Hastelloy N " Top outer annulus, 11/2 Top inner ah.nulus, 2578 Floating head, 33/4 Inner annulus, 4347 Outer annulus, 3794 Inner annulus, 0.600 radial; 0.673 circumferential Outer annulus, 0.625 triangular Donut Inner annulus, 4 - Outer annulus, 10 188 thermal conductivity was less by about a factor of 3 than the value used in the original heat transfer calculations. This difference resulted in degradation of the overall heat transfer co- efficient, necessitating an increase in heat transfer surface of the heat exchanger of approxi- mately 20%. A new heat exchanger design with a larger number of tubes and a slight change in the baffle arrangement was developed. The parameters of this new heat exchanger are given in Table 9.2. B While the new heat exchanger has a higher fuel holdup by reason of the additional tubes, a more favorable arrangement of the lheader and other reductions in parasitic volumes compen- sated for this larger number of tubes, so that the net fuel inventory in the heat exchanger—pump complex did not change significantly from that required in the first heat exchanger used in the reactor study. A significant change in the heat exchanger design concerns the method of flanging the heat exchanger into the coolant circuit piping. The coolant flows up around the heat exchanger and is discharged symmetrically across the outer bank of fuel tubes. In this arrangement the toroidal coolant header is replaced by a jacket around the tubes. Figure 9.7 shows the arrangement of this new heat exchanger. | . After traversing the outside of the fuel tubes in a countercurrent flow pattem, the coolant salt leaves the heat exchanger through a central discharge pipe which connects to the center of the concentric coolant-salt line through a slip joint. The entire heat exchanger can be re- moved from the circuit by opening the large flange, éutting the large fuel pipe communicating with the reactor, and disengaging the center fuel pipe by slipping it out of the slip joint pro- -vided for this purpose. The drain and overflow lines and auxiliary pump lines must also be cut to remove the exchanger. Removing and replacing a primary heat exchanger is a major repair. However, with the de- ‘sign as shown, the difficulties appear to be less formidable than if both coolant lines had to be cut and welded to perform the task. Adequate remote cutting, aligning, welding, and in- specting equipment and methods must be demonstrated for this application. Because of the expansion of the fuel tubes, the toroidal header at the bottom of the ex- changer will move relative to the vessel enclosing the exchanger. For this reason and also to permit removal of the exchanger, a drain line cgnnot be connected to this header. Instead, a dip line for draining fuel extends into this header from the top of the exchanger. This line connects through a freeze valve to the fuel drain tanks for the reactor. When the drain valvé is thawed and the drain tank is vented, the fuel salt in the heat exchanger flows through this line into the drain tank. The gas pressure in the pump bowl is sufficient to effect a complete drain of the fuel salt from the system. | _ In case of a pump stoppage with the attendant necessity to drain the reactor, the following sequence takes place. Upon pfimp stoppage the reactor automatically drains into the pump sump tank and through the 5-in. overflow line; the salt goes to the drain tank. A gas flow into the re- actor inlet plenum is furnished by the gas separator lines from the pump sump cover gas. This automatic drainage takes place in the order of a few seconds. However, the concentric lines 189 ORNL-DWG 67-3080A UPPER BEARING =Y B N W7 SEAL 1 ! GAS CONNECTION / i // : SHIELDING A ) L4 v / /—* — 3£t Oin. START-UP LEVEL //_ M| / &-ft O-in ODTANK —___| att Qin _ M.S. BEARING \ | | HieH 0PR. LEVEL | l—Low OPR. LEVEL IMPELLER % | — L , Sia | 15 in. FUEL TO 7 N \ REACTOR | ~ ~—T""F———1——FILL AND DUMP . 3 \ ) 32% in. ________ l | 4ul mll' | i !H”H[h A‘\\;\ - \IHIJ - | ! dlll"|l|!l" A r”"hih"fltl il M il | —A I f X 1l | m‘ ~48t1 Oin. | OUTER TUBES L L 3794 AT ¥in, oo M I | N | R INNER TUBES ———_ [l |l | @t ein 4378 AT 3gin.OD 6ft 6%in. ODx ! g Kij FUEL DRAIN ~‘ ~ r P l i mm!!!!fllm\ \' AVl Fig. 9.7. Primary Heat Exchanger and Pump. 190 to the reactor and the head space above the fuel tubes in the heat exchanger remain full of salt | and will overheat if not cooled. ' When the pump stops, the pressure in the lower header.of the heat exchanger drops to the static pressure of the salt, about 14 psi. If now the gas valves in the line are opened, the stored gas will flow into the bottom of the heat exchanger, displacing the fuel into the pump sump from which it overflows into the drain tank. By proper sizing of this tank, enough fuel can be displaced f-rom the heat exchanger to bring the level at equilibriumvbelow the top tube sheet of the exchanger. The lines and volumes above the tube sheets will be empty of salt, and no overheating will result. ' i ' We have calculated the cooling required for fuel in the tubes of the heat exchanger under . the stagnant conditions. Because of the geometry of the coolant salt system, there ig con- vective flow of coolant salt even when no coolant pump is running. This convective flow of coolant is adequate to remove aftetheat from the stagnant fuel in the tubes. To provide cooling but to avoid overcooling requires some modification of the steam circuit of a module. We have not studied the modifications in sufficient detaill to justify a discussion at this time. ' | _ i The large sump tank on top of the heat exchanger is not intended to hold the drained fuel from the reactor. Because of the undesirability of installing the afterheat cooling system in the reactor cell and also in order to be able to drain the lines to the reactor, an overflow pipe was installed from the sump tank to a separate dump tank. This tank is equipped with a cool- ing system a{dequate for afterheat removal. The sump tank on top of the heat exchanger appears to be larger than necessary inasmuch as excess fuel flows through the 5-in. drain line to the drai;'l tank. The size of the tank is dictated by the fact that it must hold the reactor volume (83 ft3) upon starting the reactor. Fuel cannot be forced by gas pressure out of the drain-tank fast enough to keep the fuel pump primed on startup. The fuel must be displaced into the sump tank and held there by gas over- ptessure in the drain tank until the reactor is filled by the action of the pump. Not shown in the drawing of the heat exchanger is the gas separator and associated ap- paratus for handling the sparge gas for the fuel system. The plans for handling this gas are being worked out, and drawings of the equipment are not yet ready. The criteria and flowsheet are discussed in Sect. 4. , Blanket Heat Exchanger The blanket heat exchanger is shown in Fig. 9.8. This heat exchanger is similar to the fuel heat exchanger but considerably smaller and less. complicated. There is no need for the gas sparge system, and there is no afterheat 'pro'blem that has to be provided forhere. In addition, the coolant flow is less complicated since counterflow is unnecessary. The femperature\dif- ference is modest between blanket salt and coolant, and the total rise in temperature is only 11°F as the coolant passes through. Table 9.3 gives the characteristics of the blanket ex- changer. 191 _ UPPER BEARING ORNL-DWG 67-3079A —- i St : i %AN:Z SHIELDING %8 !'/ % SEAL il . y 24in. OD 3t Oin. ' / BAFFLE ; | OPERATING LEVEL e / p //— MOLTEN SALT BEARING | == BLANKET "PROCESSING — [— DRAIN —— — \—-. . IMPELLER-20-in. PITCH DIAM . TS ' 1f1 7 in, |s/ ‘\g\ Al —— 4 - BLANKET TO REACTOR 7 R ,&‘\‘\A\‘ \% e . (ZET AL L L 77 S 23 ft 7in. ‘ - FLANGE Vi : GAS CONNECTION GAS SKIRT | OUTER TUBES-827 AT ¥ in. OD 9ft Oin. L INNER TUBES-814 AT % in. OD -4 ft 5in. 0D 6in + 3ft 6in. R R N /—'—— s ) i \ k% . \ f’ COOLANT SALT A Fig. 9.8. Blanket Heat Exchanger and Pump. 192 Table 9.3. Blanket Heat Exchanger Data Sheet Blanket-coolant-salt exchanger, one-shell pass, two-tube pass with disk- and donut-type baffles Number required Rate of heat transfer, Mw Rate of heat transfer, Btu/hr Shell side Hot fluid or cold fluid Entrance temperature, °F Exit temperature, °F Entrance pressure, psi®- Exit pressure, psi® AP across exchanger, psi® Mass flow rate, 1b/hr Tube side Hot fluid or cold fluid Entrance temperature,. °F Exit temperature, °F Entrance pressure, psi'Ei Exit pressure, psi® AP across exchanger, psib Mass flow rate, 1b/hr Velocity, fps Tube material Tube OD, in. ’i‘ube thickness, in. Tube length, tube sheet to tube sheet, ft Shell material Shell thickness, in. _Shell ID, in. Tube sheet material Tube sheet thickness, in. Number of tubes Pitch_of tubes, in. Total heat transfer area, t'tA2 Basis for area calculation Type of baffle Number of baffles Baffle spacing, in. Disk OD, in. Donut ID, in. Overall heat transfer coefficient, U, Btu hr— ! ft—2 4 27.75 9.471 % 107 Cold (coolant salt) 1111 1125 138 129 15 1.685 x 107 » Hot (blanket salt) 1250 1150 111 20 91 4.3 x 10° 10.5 Hastelloy N 0.375 0.035 8.3 Hastelloy N L 40:78 Hastelloy N 1 Center section, 810 Annular section, 810 0.8125 1318 Tube OD Disk and donut “Includes pressure due to gravity head. bPress‘.ure loss due to friction only. 193 9.5 REACTOR PHYSICS O. L. Smith In addition to optimization studies describe;i above, work on MSBR reactor physics in- cluded (1) a series of cell calculations performed to examine the sensitivity of the MSBR cross sections and reactivity to various changes in cell structure and composition, and (2) several two-dimensional calculations of the entire reactor. All of the calculations were based upon the most accurate description of the system that was available as of January 1, 1967. The graphite-moderated portion of the core was 10 ft in length and 8 ft in diameter, contained ~0.2 mole % 233U, 27 mole % 232Th, and had a fuel volume fraction of 16.48% and a fertile volume fraction of 5.85%. The reactor had a 1.75-ft-thick radial blanket consisting of 60% fertile salt and 40% graphite, surrounded by a graphite reflector 6 in. thick. The top axial blanket was 1.5 ft thick and contained 60% fertile salt and 40% graphite. The bottom axial blanket was 1 ft thick and contained 3.18% Hastelloy N, 16.48% fuel salt, and 80.34% fertile salt. A number of structural details below the lower blanket were included in the two-dimen- sional calculations. | Figure 9.9 shows the geometry of a cell in the graphite-moderated core. Graphite dowels of appropriate size are located at the six corners of the cell to yield the desired fertile salt volume fraction. The cell calculations were performed with the code TONG and involved varying (1) cell diameter, (2) fuel distribution (i.e., fuel separation distance, s), (3) 233y concentration, (4) 232Th concentration, (5) fuel volume fraction, and (6) fertile volume frac- tion. Each of these parameters was varied separately while holding the others constant. Table 9.4 and Fig. 9.10 show the effect on reactivity of varying the parameters. The varia- tions are shown relative to a reference cell which had a diameter (flat to flat) of 3.156 in. and a fuel separation distance, s, of 1/8 in. ORNL-DWG 67-4789 ~3\ COS N \ @ ./ ‘ » Fig. 9.9. Cell Geometry. ¢ 194 Table 9.4. Effect on Reactivity of Yarying Cell Properties Percent Case Variation eff 1 Reference cgll 0 2 Cell diameter —36.7 - 0,0271 3 Cell diameter +58.4 +0.0317 4 Fuel separation, s —~100 ~—0.0011 5 Fuel séparation, s + 200 +0.0003 6 233y concentration —25 —0.1031 7 233y concentration +25 +0.0763 8 232, concentrat'ion —25 +0.0936 9 232Th concentration + 25 —0.0768 10 Fuel volume fraction - 25 —0.0851 11 Fuel velume fraction + 25 +\0.0512 12 Fertile volume fraction —~25 +0.1048 13 Fertile volume fraction + 25 —0.0845 ) ORNL-DWG 67—4790 PERCENT FUEL SEPARATION CHANGE -100 0 100 200 0.15 ) [ | | S 0.0 e\p)/(é\ - N\, 55 e o . (&) = N% L DIANETER | 3 . FUEL SEPARATION s | .,t —y .-—-—"--- I < // N edé’ <© NH - N N, < vy Y & \/P \1\3 \ CFy O/ O -010 "—QQ@\’I‘f 7 -40 -20 0 20 > 40 60 - PERCENT CHANGE Fig. 9.10. Effect on Reactivity of Changing Cell Properties. 195 The results of these calculations (compare cases 1 and 3) indicate that there may be a . reactivity advantage (attributable to increased self-shielding of the ?32Th resonances) to .using a cell somewhat larger than the 3. 156-in. reference cell. If the conversion ratio is’ not a'dversely affected, use of a 5-in. cell may, for example, pemit reduction of the 233U inventory. Pending further study, case 3 (which differs from case 1 only in cell size) is considered to represent the current cell dimensions used in the design. Table 9.5 and Fig. 9.11 show information about the flux distribution for case 3. Table 9.5 shows the ratio of the average flux in the fuel to thé cell average flux, the ratio of the a\llerage flux in the graphite to the cell average flux, and the ratio of the average flux in the fertile salt to the cell average flux for the epithermal and fast flux ranges. Figure 9.11 shows the thermal flux distribution in the cell. The results in Fig. 9.11 are based upon an annular approximation to the cell of Fig. 9.9. _ . The two-dimensional calculations used the 5-in.-diam cell composition and cross sections for the core, and they were performed with the diffusion theory code EXTERMINATOR-2. Nine Table 9.5, Flux Ratios in Epithermal and Fast Energy Ranges Energy Range Fuel Graphite ' Fertile ~ 0.821-10 Mev 1.226 0.929 0.878 0.0318-0.821 Mev 1.090 0.984 0.958 1.234--31.82 kev 1.014 0.998 0.991 0.0479~1.234 kev 1.0 1.0 1.0 1.86—47.9 ev 1.0 1.0 1.0 ORNL—DOWG 67-4731 2.0 w 1.5 CELL AVERAGE FLUX g 7 ~~ - — — -t e S ) 5 10 ‘ § —— FUEL _"I"“/ ‘—l' \ii GRAPHITE —-‘ i-¢—~ >:<) GRAPHITE FUEL ?\_FERT;LE —J z 0.5 o] o) i 2 3 4 5 6 : 7 8 9 RADIAL POSITION {cm) Fig. 9.11. Thermal Flux Distribution in Cell (E <1.86 ev). 196 energy groups were used and are shown in Table 9.6. Figure 9.12 shows the radial flux dis- tribution in 'the core midplane for neutron energy groups 1 and 6 separately, and Fig. 9.13 shows the total flux distribution. The radial peak-to-average flux ratio in the graphite-moderated part ofthe core is 1.58. Figure 9.14 shows the axial flux distribution for groups 1 and 6 at a radial distance 18 in. from the axis of the core, and Fig. 9.15 shows the total flux distribution. The . axial peak-to-average flux ratio in the graphite-moderated core region is 1.51. Thus the total peak-to-average flux ratio is 2.39, the peak occurring at the geometric center of the graphite- moderated core. ‘ The central cell of the reactor is intended for control purposes and consists of a graphite tube 5 in. in outside diameter and 4 in, in inside diameter. It is envisioned that control will be achieved by regulating the height of the fertile salt in the tube. If the completely empty tube is filled with fertile salt, .the change in reactivity is &k/k = ~0.018%. If the empty tube is filled with graphite, the reactivity change is dk/k = +0.0012%. Thus there appéars to be a substantial amount of reactivity control available by varying the height of the fertile column in the tube. Table 9.6, Neutron Energy Groups Used in Two-Dimensional Calculations Group Energy Range Pk 0.821-10 Mev 0.0318—~0.821 Mev 1,234-31.82 kev 0.0479-1.234 kev 1.86—47.9 ev 0.776~1.86 ev 0.18—0.776 ev 0.06--0.18 ev O 0 3 O o s W N 0.01-0.06 ev 197 ORNL-DWG 67-4792 {x10'% 2.0 1.6 .4 CORE ~mfe—BLANKE T— 7| -2 GROUP 1 / FLUX (neutrons cm o ® GROUP 6 0.6 \ \ 0.4 0.2 0 { 2 3 4 5 6 RADIAL POSITION (ft) Fig. 9.12. Radial Distribution of Neutron Flux in Groups 1 and 6 at Core Midplane. x10'% -2 FLUX (neutrons cm 198 ORNL-DWG 67-4793 (x10'3) 2.8 \ N IS L~ ) CORE BLANKET sec M Q [o)] = n - FLUX (neutrens em™2 o ® 0.4 : \\ , — 0 1 2 3 4 5 6 7 RADIAL POSITION (ft) Fig. 9.13. Radial Distribution of Total Neutron Flux at Core Midplane, ORNL-DWG 67— 4794 14 - /BOTTOM OF / _ TOP OF CORE / CORE — 12 N [~ 10 [ BLANKET / - \ BLANKET 08 / T e GROUP 6 ‘ Zj: N Y 0] 1 2 3 4 5 6 7 8 9 10 1" 12 13 iq 15 ' AXIAL DISTANCE FROM BCTTOM OF REACTOR (ft) Fig. 9.14. Axial Distribution of Neutron Flux in Energy Groups 1 and 6 at a Distance 18 in. from Core Axis. {x10'3) FLUX (neutrons cm™2 sec !y . 0.8 199 ORNL-DWG 67- 4795 BLANRET-\\‘ 1 / / | A TOPOi—'CORE i/ NEEEER 0 1 2 3 4 5 6 . 7 8 9 10 " 12 13 14 15 ‘AXIAL DISTANCE FROM BOTTOM OF REACTOR (ft) ' Fig. 9.15. Axial Distribution of Total Neutron Flux at a Distance 18 in. from Core Axis. 9.6 MSBR GAS HANDLING SYSTEM Dunlap Scott A. N. Smith R. J. Kedl. The gas handling system for the MSBR serves several functions in the operation of the plant. These include: supplying helium for purging and pressurization of the gas spaces; rapidly removing !?5Xe from the salt; transporting, removing, and storing the radioactive fission product gases; and processing the helium for recycle into the gas supply‘ or disposal to the atmosphere. A scheme for accomplishing these functions is described below. Xenon Removal Preliminary studies of the migration of '*°Xe to the graphite in molten-salt breeder re- ‘actors indicated that it would be possible to reduce the '35Xe poison fraction to an acceptable value of 1/2% if a gas stripping system could be devised which would process the entire reactor " fuel-salt inventory in about 30 sec. Since the reactot systém salt circuit time is about 9 sec, it is required that the entire inventory be processed every 3 to 4 passes around the circuit. The required xenon processing time could be extended by processing the fuel for removal of the precursor, 1*°l. However, the advantages of '*°I processing are limited by the fact that 20% of the total !35Xe produced in the fissioning of 2*3U appears directly, and very high proc- essing rates would be required to obtain even modest gains. For example, if the entire fuel in- ventory were processed for 13°I removal once every hour, the xenon removal process rate would only be extended to about 80 sec, and a very high !l processing rate would increase the xenon 200 processing time to only about 110 sec. Therefore; processing for iodine is not, at this time, being considered as a method of removing !%°Xe. ' The method proposed for removal of 3%Xe involves stripping of xenon-enriched helium bubbles from the fuel stream by means of a gas separator located at the heat exchanger outlet. The bub- bles will be'generated by injecting helium into the salt at the fuel pump suction-, and transfer of xenon from the salt to the gas will be effected during passage through the heat exchanger. . Mechanical Design The flowsheet for the MSBR dff-gés system is given in Fig. 9.16. Helium is injected into the suction of the fuel salt pump and is removed by centrifugal separation at the heat exchanger outlet. The liquid-gas mixture from the separator is then féd into a cyclone separator where the - entrained liquid is removed and returned to the pump bowl.. The gas from the pump shaft purge and the instrument lines is combined with that.from the cyclone separator and fed into the 48- hr xenon holdup system. A-small portion of the flow from this system is fed into the long-term xenon holdup and then through the noble-gas separator to a recycle system for supplying clean helium to the pump shaft purge and the instrument gas lines. The remainder of the gas from the 48-hr xenon holdup system is fed into the gas injector system at the fuel pump suction. It is the salt-powered gas injector that serves as the prime mover for the high-gas-flow recycle system. Some of the design criteria of the critical components are described below. Gas Injector System 1. The gas addition rate shall be sufficient to produce about 1% void volume in the salt at the pump suction, 2. The location and manner of gas injection shall provide a balanced distribution of bubbles in the liquid entering the pump impeller. This consideration affects the bubble distribution in the heat exchanger and the bubble size, as well as the pump denamic balance. 3. There shall be operator control over the gas addition rate into the gas injector. 4. The supply for the gas injector shall be taken from the outlet line from the 48-hr xenon holdup helium recirculation system. There shall be a backflow preventer arrangement at this gas supply point to prévent salt from getting back into the charcoal beds as a result of a- sudden pressure transient in either the salt .or gas system. | ‘ / 5. The method of injecting bubbles into the pump suction shall be a fuel-salt-powered jet pump taking. ité. salt supply from the heat exchanger discharge. The suction pressure for the salt-powered jet pump shall be less than 10 psia at a flow rate of 6.5 scfm of helium. iy BLANKET PUMP CYCLONE SEPARATOR MISCELLANEOUS PURGE GAS ‘\ . 12 25 To mopu 3 4 sl > MODULE HEADER 1234 CW= COOLING WATER POINT ft3/sec ) - o 25 25 25 Q@@ psia 277 '64.7 457 327 " HX = HEAT EXCHANGER TEMP (°F) 1000 1000 1000 1000 LES VOLUME HOLDUP AND FISSION PRODUCTS HELIUM SUPPLY COOLANT PUMP PURGE ——= DRAIN TANK VENT ——] ACCUMULATOR & FILTERS AND PRE - TREAT 93 HI-FLOW AUXILIARY SURGE CHARCOAL TANK BED X CHEMICAL PROCESS PURGE ———= ORNL-DWG 67-3077A HX CW HX cw of . 90-day clflfi%%i\l_ 2G @ . - BEDS [ { ? | X ] CW e 13 113 { 3 { gl [ 3 EITT13 . 91). £ 3 3 3 s T T 13 3 & 3 Kr -85 AND T TRITIUM REMOVAL I PRI psia TEMP (°F) 247 24.0 23.0 18.0 70 247 247 1000 1000 230 230 1000 DATA _U o Z BORO®O®B® MAINTENANCE CONTAINMENT VENTILATION psia TEMP (°F) scfm 200 180 200 165 200 150 2.00° 447 200 44.7. 025 247 50 MAX 17.7 MAX 50 MAX 17.2 MAX SO MAX 147 Fig. 9.16. MSBR Gas Flow Diagram, 100 70 70 70 CIOD1NE TRAP AND ABSOLUTE FILTERS NOTE : TOTAL HEAT LOAD ON 48-hr . DELAY SYSTEM IS ~ 5 Mw 102 202 Bubble Sepnrofor System l _ 1. The bubble éeparator shall be installed within the outer annulus of the fuel-salt line from the heat exchanger to the reactor. ‘ 2. The separator, including the entrance region, shall be kept as short as is reasonably pos- sible, and the separator shall be as close to the heat exchanget as good design pemits. The purpose of this restriction is to pemit removal of the separator along with the heat exchanger and at the same time permit locating the heat exchanger close to the ;eactor. 3. The design of the system shall include assurance that the separator will not admit gas into the salt line to the reactor except when the fuel pump stops. Check valves or liquid sub- mersion of the discharge will not be used to prevent backflow, since the reactor drain scheme will use the separator outlet as a gas source. | 4. The bubble fraction of the salt as it enters the reactor shall be less than 0.1% during steady-state operation. 5. The fuel line outside the gas separator shall be free of protrusions all the way to the heat exchanger outlet. Equipment for cutting and welding the 24-in.-diam fuel line will occupy this space during some maintenance operations. | 6. The gas outlet from the bubble separator and the gas line to the pump ‘bowl shall stay within the salt lirie if possible. This will reduce the number of penetrations through the pipe wall. 7. A modified gas cyclone separator shall be an integral paft of the bubble separator dis- chérge stream. The liquid would be discharged to the pump bowl, and the gaseous discharge would go to the xenon holdup system. - Yolume Holdup System The first stage of the high-flow 48-hr xenon holdup is a gas volume holdup whose several functions are listed below. 1. The first stage shall be a cooled volume for the decay of the very short-lived gaseous fission products. A reentrant tube system similar to the one in the fuel drain tank of the MSRE * will be investigated for providing the cooling. 2. There shall be a provision for a final demisting of the gas as it comes from the cyclone separator. ‘ 3. Since there will be a total of more than a kilogram of fission products produced each day in the reactor complex, and since a sizable fraction is involved in the short-lived gaseous products, a method of collecting, cooling, and conveying to a disposal area must be included. The collecting device should pass only gas which has resided in the volume holdup for at least the specified minimum transit time. | ‘ 4. The volume holdup system shall be sized to reduce significantly the heat load and particulate loading rate from short-half-life fission products in the first stage of the charcoél " bed, which is immediately downstream. 203 Noncritical Components . Many of the remaining components in the off-gas handling system ate reasonably well understood, and, while they must be designed and ultimately tested, it is believed that they will offer no critical problems. Some of these are listed below: 1. High-gas-flow charcoal bed design. Biological charcoal bed design (low flow). 3. Noble-gas separator and disposal system. This includes 85Kr and tritium removal from the helium which is to be returned to the recycle system. . 4. Helium compressor for recycle of clean helium. Gas sampling system for purity control and for surveillance of the chemical condition of the salt. ‘ 6. Instrumentation and controls for operating the gas system. 10. Molten-Salt Reactor Processing Studies M. E. Whatley A close-coupled facility for processing the fuel and fertile streams of a molten-salt breeder reactor (MSBR) will be an integral part of the reactor system. Studies are in progress for obtaining data relevant to the engineering design of such a processing facility. A The fuel processing plant will operate on a side stream withdrawn from the fuel stream which circulates through the reactor core and primary heat exchanger. For a 1000-Mw (electrical) MSBR, approximately 14.1 ft® of salt will be processed per day, which will result in a fuel-salt cycle time of approximately 40 days. The presently envisioned process has been described previously; ! the significant process steps are recovery of the uranium by continuous fluorination, recovery of the carrier fuel salt by vacuum distillation, and recombination of the i)urified UF _ and barren car- rier salt. 10.1 CONTINUOUS FLUORINATION OF AMOLTEN SALT L. E. McNeese B. A._ Hannaford - Uranium present in the fuel stream of an MSBR must be removed prior to the distillation step, since UF | present in the still would not be completely volatilized and would in part be discharged to waste in material rejected from the distillation system. Equipment is being developed for the continuous removal of UF, from the fuel stream of an MSBR by contacting the salt with F, in a salt-phase-continuous system. The equipment will be protected from corrosion by freezing a layer of salt on the vessel wall; the heat necessary for maintaining molten salt adjacent to frozen salt will be provided by the decay of fission products in the fuel stream. Present development work consists of two parts: (1) studies in a continuous fluorinator not protected by a frozen wall and (2) study of a frozen-wall system which is suitable for continuous fluorination, but with an inert gas substituted for fluorine in the experiments. Work on the two systems is described in the fol- lowing sections. Nonprotected System The experimental system has been described previously! and consists of a 1-in.-diam fluori- nator 72 in. long and auxiliary equipment which allows the countercurrent contact of a molten salt IMSR Program Semiann. Progr. Rept. Aug. 31, 1966, ORNL-4037, p. 232. 204 3 205 with F . It is intended to demonstrate the effectiveness of this type of system for uranium recov- ery and to obtain engineering design data on backmixing and F , utilization. The present work has used salt containing no beryllium, but LiF-BeF2 mixtures will be used when suitable facil- ities for handling beryllium are made available. This system also allows development of techniques and auxiliaries required for continuous fluorinators. These include means for controlling and measuring molten-salt flow rates and methods for sampling and analyzing molten-salt streams and gas streams. A gas chromatograph is used for analysis of the fluorinator off-gas for UF _, F,, and N,. Uranium concentration in the salt dufing fluorination is determined from salt sampli:s taken at 15-min intervals. Data obtained from this system have been reported previously;! three additional experiments have been made at 600°C using an NaF-LiF-Z:F , mixture containing 0.37 to 1.16 wt % UF, and having a melting point of ~.475°C. Salt feed rates of 9.8 to 30 cm?®/min and F, rates of 235 to 335 cm3/min (STP) have been used with molten-salt depths of ~.48 in. Uranium removal during one pass through the flu’orinator varied from 99.36 to 99.89%, as determined from salt samples. During the best run' to date, a molten-salt feed rate of 10.2 cm3/min gnd an F, feed rate of 235 cm?®/min (STP) were maintained for a 2.5-hr period. At steady state, 99.89% of the uranium was removed by the fluorinator, as determined from inlet and exit concentrations of uranium in the salt. Later in the run, a molten-salt feed rate of 22.5 cm3/min and a/n-’F2 rate of 270 cm?/min (STP) were maintained for a 2.5-hr period. With these conditions, the uranium removal at steady - state was 99.62%. The uranium concentration in the feed salt was 1.16 wt %, and fluorine utiliza- tions were 15.6 and 30% respectively. The equipment operated smoothly during the run, salt and gas feed rates were constant, and the system was operated at steady state for 2 hr at each of the above conditions. | The available design parameters include the ratio of salt and fluorine flows and the height of the tower. Any recovery can be attained, and an economic optimum will probably fall above 99.9%. Protected System Components for the protected system have been fabricated and are being installed. This system will be operated with a frozen layer of salt on the fluorinator wall and will allow the. countercurrent contact of an inert gas with molten salt in equipment suitable for continuous fluorination. The fluorinator is constructed from 5-in.-diam sched-40 nickel pipe and will provide a section protected by a frozen wall ~6 ft high. Internal heat generation is provided by Calrod heaters blaced inside a 3/4-in.-diam nickel tube éldng the center line of the vessel. The frozen-wall thick- ness will be dependent on the radial heat flux and can be varied from 3/8 to 11/2 in. Frozen-wall thickness will be determined from temperature gradients measured by two sets of four internal thermocouples located at different radii with respect to the vessel center line. A feed tank and receiver vessel will allow the feeding of a salt volume equivalent to approximately ten fluorinator volumes through the system. 206 The system will be operated with a salt mixture (66 mole % LiF—34 mole % Z:F ,) having a phase diagram similar to that of the LiF-BeF , fuel salt for the MSBR and should point out problems associated with the operation of a frozen-wall fluorinator. 10.2 MOLTEN-SALT DISTILLATION STUDIES J. R. Hightower L. E. McNeese Relative volatilities of several rare-earth fluorides with respect to lithium fluoride have been measured in an equilibrium still operated at 1000°C and (.50 mm Hg pressure. Recent work has refined previously reported values for CeF ,, LaF ,, and NdF ; and has added the relative volatility for SmF ,. The rate of vaporization of LiF has been measured at 1000°C at several pressures; these data are use;ful in estimating the probable error in measured relative volatilities and for predicting vaporization rates in eqfiipment suitable for MSBR processing. The results of a study on the buildup of materials of low volatility at a surface where vaporization is occurring are also given. Relative Volatility Measurement A distillation step will be used in the MSBR processing plant to remove rare-earth fission products from the fuel stream. To design the still it is necessary to know relative volatilities of the rare-earth fluorides with respect to LiF, the major constituent of the still pot. The equi- librium still used for these measuremenfs and the operating procedure have been previously de- scribed.? Relative volatilities of four of the rare-earth trifluorides with respect to LiF at 1000°C and 0.50 mm Hg have been obtained from recent experiments and are listed below. Rare-Earth Fluoride Mole Fraction Relative Volatility with Respect to LiF CeF, " 0.02 | 3x 103 LaF, 0.02 3x107* - NdF, 0.05 _ 6x10* _ SmF 0.05 2% 10~4 | b These relative volatilities allow the required rare-earth fluoride (REF) removal efficiencies in a still of simple design without rectification. Vaporization Rate Studies Data on the variation of vaporization rate with total pressure are necessary to assess the error in relative volatilities measured in the recirculating equilibrium still and to predict vaporiza- tion rates in equipment suitable for MSBR processing. A diagram of the equipment used for the 21bid., p. 229. o [ Y x 207- measurement of vaporization rates. is shown in Fig. 10.1. The distillation unit was made from 1- in. nickel tubing bent into an inverted U. The salt was vaporized from a graphite crucible in the left leg of the still, and the condensate was collected in a similar crucible in the right leg. For operation of the still, a mixture oflLiF and a rare-earth trifluoride having a known com- position was placed in a crucible in the vaporizing section of the still, and a second crucible was placed below the condenser. The system was purged with argon while being heated to the desired temperature. When the temperature in the vaporizing section reached 1000°C, the still pressure was decreased to the desited value, and the condenser cooling air was turned on. The condenser temperature of about 500°C caused the salt vapor to solidify on the condenser walls. After a given length of time the St‘ili was pressurized with argon and fhe cooling air turfled off; this allowed the condensate to melt and drain into the crucible below the condenser. Vaporization rates were determined from the decrease in weight of the salt in the vaporizing section of the still, since only part of the salt vaporized was collected in the other crucible. Results of these tests are given in Table 10.1. At pressures near the v;apor pressure of LiF, the LiF vaporization rate increased slowly as the pressure was lowered. At a pressure significantly lower t.han the LiF vapor press‘ure, the rate increased by an order of magnitude. When the total pressure is higher than the vapor pressure of ORNL—-DWG 67 - 15074 1—in. NIGKEL TUBING —= faN ¢ ANANANANAN T CONDENSER . N . ¥ GOOLING AIR ~=—— FUNNEL f'\’ THERMOWELL—e{T] VAPORIZ!NG/‘ :H SECTION CRUCIBLE — CONDENSING SECTION CRUCIBLE : CRUGIBLE . =~ SUPPORTS ARGON ~ Y ' NLET —=C— —— vacu SWAGELOCK VACUUM EEH/ FITTINGS \Efé PUMP Fig. 10.1. Appuratus for Vaporization Rate Measurement. 208 . Table 10.1. Variation of LiF Vaporization Rate with Total Pressure at 1000°C bondenser Pressure®’ Vaporization Rate (mm Hg) (g cm ™2 sec“l) 1.0 7.8 x 107° 0.50 3.3x 1073 0.35 48X 1077 0.1 ' 2.4 x 107* “The vapor pressure of LiF at 1000°C is about 0,53 mm Hg. the salt, the rate of vaporization should be controlied by the rate of diffusion of LiF dand REF through the argon present in the system; the measured rate .at 1.0 mm Hg was comparable with the rate calculated by assuming the rate to be diffusion controlled. Since these data indicate that the recirculation rate in the equilibrium still is controlled by diffusion of the salt vapor through argon, an error in the measured relative volatilities could arise because of differences in the rates of dif- . fusion of LiF and REF vapor. Calculations indicate that the error in the rel\ative volatility from . this source is only about 1% and that other effects such as a nonuniform concentration gradient in the liquid are much more important. Buildup of Nonvolatiles at a Yaporizing Surface " During vaporization of a multicomponent mixture, materials less volatile than the bulk of the mixture tend to remain in the liquid phase and are removed from the liquid surface by the processes of convection and molecular diffusion. Low-pressure vaporization does not generate deeply sub- merged bubbles and therefore provides little convective mixing in the liquid. An appreciable variation in the concentration of materials of low volatility may occur if these materials are removed by diffusion only. 7 Consider as an exemplary case a continuous still of the type showr} in Fig. 10.2. Fuel carrier salt (LiF-BeF 2) containing fission product fluorides is fed to the bottom of the system contin- uously. Most of the LiF-BeF , fed to the system is vaporized, and a salt stream containing most of the nonvolatile materials is withdrawn continuously. The positive x direction will be taken as vertically upward, and the liquid withdrawal point and the liquid éurface will be located at x = 0 and x = [ respectively. Assume that above the liquid withdrawal point, molten LiF containing REF flows upward at a constant velocity V. At the surface, a fraction v/ V of the LiF vaporizes, and the remaining LiF is returned to the bottom of the still. Above the withdrawal point, the concentration of REF satisfies the relation d?’C _ dC - | D— -V _—_=0, . (1) dX2 dX_ : . 209 ORNL-DWG 67-294A PRODUCT SALT v,aCy i RECIRCULATING SALT V-v, Cg —‘~ x=£ ] | | x | | | DISCARD SALT WITHDRAWAL F-v,C, POINT FEED SALT F, Gy Fig. 10.2. Continuous Still Having External Circulation and a Nonuniform Liquid Phase Rare-Earth Fluoride Concentration Gradient. and the .boundary conditions are, at x =1, dC -D = |x=1‘+ VC =vaC +(V - v)C_, and at x = 0, dC ‘ ' -D E|x=O+VCO=(V—v)CS+FCf—-(F -v)C,, where D = diffusivity of REF in molten salt of still pot concentration, cm?/sec, - C = concentration of REF in molten salt at position x, moles of REF per cm? of salt, x = position in molten salt measured from liquid withdrawal point, cm, i | V = velocity of molten salt with respect to liquid surface, cm/sec, C . = concentration of REF at x = I, moles of REF per cm? of salt, (2) 3) 210 -» ‘ : cm? LiF (liquid) F = LiF feed rate, s cm? vaporizing surface - sec em?® LiF (liquid) v = LiF vaporization rate, , cm? vaporizing surface - sec a= relative volatility of REF referred to LiF, C, = concentration of REF at x = 0, moles of REF per cm? of salt, C = concentration of REF in feed salt, moles of REF per cm? of salt. Equation (1) has the solution FCifl — (w/V) (1 = a) [1 — exp (-V{ - x)/D)]} Cx) = . 4 . va+(F =il - /YA - a) [l —exp (-VI/D)]} The fraction of the REF removed by the still is" ' . F - v)C, fraction REF removed = — FC; ' F—v 5) " Fevyva/[l = (w/V)] (1 = @) [1 - exp (-VI/D)} The fractional removal of REF for a continuous still having a perfectly mixed liquid phase is ' 1 fraction REF removed = . (6) . 1+ [va/(F = w)] The ratio of the fractional removal of REF in a system having a nonuniform concentration to that in a still having a uniform concentration will be denoted as ¢ and can be obtained by dividing Eq. (5) by Eq. (6). Thus ' 1+ [va/(F — v)] ' ¢ @ 1+ ve/F - I/ - /M A = @) [1 — exp (-VI/D))} ' Values of ¢ calculated for a still in which 99.5% of the LiF fed to the still is vaporized [v/(F-v) =199] and in which the relative volatility of REF is 5 x 10~ * are given in Fig, 10.3. The following two effects should be noted: 1. The value of ¢ is essentially unity for VI/D < 0.1 for any value of v/V (fraction of LiF vaporized per circulation through still). Within this region, a near-uniform REF concentration is maintained by diffusion of REF within the liquid, and mixing by liquid circulation is not re- quired. 2. The value of ¢ is strongly dependent on v/V for VI/D > 1. Within this region, a near- uniform REF concentration can be maintained only if liquid circulation is provided. For VI/D = 1] Lo & 211 ORNL-DWG 67-30A LO | | I 1Y II 'i' i i ! | SN N — | ~— =0.5 v To.1 0.5 N - . (. \, \.«;—209 0.2 0.4 \ \ 0.05 \ 0.02 \ 0.01 : \ a=5xi0"" V_ =499 i 0.005 Fv 0 \ \L=,| Y 0.002 0.00{ : 0.1 0.2 0.5 1 2 5 10 20 50 100 Ve Fig. 10.3. Ratio of Fraction of Rare-Earth Fluoride Removed in Still Having Nonuniform Concentration to That in Still Having Uniform Concentration. 100, ¢ has a value of 0.0055 with no liquid circulation and a value of 0.99 if 90% of the LiF is returned to the bottom of the still. . ‘ ‘ ' An actual still would probably operate in the region VI/D > 1; so the importance of liquid circulation cannot be overem[;hasized. Liquid phase mixing by circulation is believed to be an essential feature of an effective distillation system. 10.3 VACUUM DISTILLATION EXPERIMENT WITH MSRE FUEL SALT W. L. Carter. Experimental equipment has been designed for an engineering-scale demonstraflon of vacuum distillation of molten-salt reactor fuel. The distillation will be carried out at about 1000°C and 1 mm Hg pressure to separate LiF-BeF, carrier salt from less volatile fission products, primarily the rare earths. Uranium tetrafluoride will not be present during distillation, having been previ- ously removed by fluorination. 212 This experiment is pért of a program to develop all unit operations (see Fig. 10.4) in the processing of a molten-salt breeder reactor fuel. Vacuum distillation is the key step in.the proc- ess because it recovers the bulk of the valuable LiF-BeF carrier, decontaminated from fission products, for recycle to the reactor. Feasibility of distilling fluoride salts was established in batch laboratory experiments by Kelly;? the present experiment will demonstrate the operation on an engineering scale and furnish data on the relative volatilities of the components of the mix- ture. \ | In the interest of simplicity, fabrication time, and economy, no attempt is being made in this experiment to reprodu'ce actual MSBR operating conditions, such as high internal heat generation rate in the still volume or the design of a still that cafi serve in a processing plant for a breeder. Such advances are the next logical step after an engineering-scale demonstration. However, it is the purpose of this experiment to show that molten salt containing fission products can be fed continuously to the still at the same rate at which it is being distilled, with the simultaneous accumulation of fission products in the bottoms. The still can also be operated batchwise to concentrate fission products in some small fraction of the original charge. 3u. J. Kelly, ““Removal of Rare Earth Fission Products from Molten Salt Reactor Fuels by Distilla- tion,’” a talk presented at the 11th annual meeting of the American Nuclear Society, Gatlinburg, Tenn., June 21-24, 1965, ORNL-OWG 66 — B2RA - F, RECYCLE NaF SORBER or REACTOR 100 - 400°C gF, . coLD i | SORBER Y REP BLANKET UFs+ Py LiF-BeF,-UF, : ’ - FLUORINATOR : ~550°C _jet— F2 LiF—BeF,—UF, LiF + BeF, SPENT ) 2 NaF + MgF, , . OFF - AND VOLATILE FP’s GAS poill)— STILL = MAKE-UP - ~{000°C ' UFg LiF -BeF,-UF, ' H REDUCTION - - 2 WASTE | FUEL ‘ ' * MAKE - UP : . DISCARD WASTE FORZr RARE EARTHS REMOVAL INLiF Fig. 10.4. Principal Steps in Processing Irradiated Fuel from a Molten-Salt Reactor. i . Wl 213 - The experimental program is in two parts: About 90% of the time will be devoted to nonradio- active operation, and the remaining 10% to radioactive operation in distilling a small quantity . of fuel from the MSRE. The first phase is expected to log 500 to 1000 hr of operation. The same equipment is to be used in the radioactive experiment after being thoroughly inspected at the conclusion of nonradioactive operation. Radioactive runs will be carried out in an MSRE cell. Components of the experiment are a feed tank (48 liters), still (12 liters), condenser, conden- sate receiver (48 liters), associated temperature and pressure instrumentation, and vacuum sys- tem. The still, condenser, and receiver are fabricated as a unit, The vessels are mounted in an angle-iron frame, which is 3 x 6 x 7 ft high, allowing transport of the entire facility as a unit once instrumentation, power, and service lines have been disconnected. Each process piece is surrounded on all sides by shell-type electric heaters; these in tfirn are enclosed in 4 to 8 in. of thermal insulation. A diagram of the assembly, which shows the feed and sampling mechanisms, is shown in Fig. 10.5. An experiment is carried out by charging a molten mixture of carrier salt and fission product ~fluorides into the feed tank, which is held at a temperature slightly above the melting point. In most cases this is 500 to 550°C. Concurrently the still, condenser, and condensate receiver are heated and evacuated, and the space above the liquid in the feed tank is evacuated. The final pressure is adjusted to about 0.5 atm in the feed tank and 0.5 mm Hg in the rest of the equipment. ORNL-DWG 66—10952 A VENT —— o CoRTESL STILL VENT N J‘p;’: __________ VOL =12 liters P={mmHg r=1000°C ! : ] Np INLET - i o VACUUM PUMP LIQUID N, COLD TRAP CONDENSER 10 in. DIAM X 51 in. CONTAMINATED FEED LiF =65.6 mole % BeF,=29.4 mole % Zrfa = 5.0 moie % FISSION PRODUCTS SAMPLER H FEED TANK =1 VOL = 48 liters RN CONDENSATE TANK F =6 psia : VOL =48 liters 7 = 500°C ) ; F=0.5mmHg T'= 500°C Fig. 10.5. Vacuum Distillation of LiF-Ber-Zde- DECONTAMINATED LiF -BeF,-ZrF, 214 When the still temperature reaches that of the feed liquid, a 12-liter charge is forced into the still through a heated 1iné, and the temperature is raised to 950 to 1000°C, the temperature range in which diétillation begins. The still pot is the highest point of the system and is an annular volume surrounding the top of the condenser. Salt vapors flow into the top of the condenser and condense along its length by losing heat to tlile surroundings. Freezing is prevented by supplying the necessary external heat to keep the éoqdenser surface above the liquidus temperature. In leaving the condenser the distillate passes through a small cup from which samples can be removed for analyses. Liquid-level instrumentation in the still allows control of feed rate to correspond to the dis- tillation rate, which is estimated to be 400 to 500 cm?® of distillate per hour. Determining the actual distillation rate for molten-salt systems is an important part of this experiment. All vessels and lines that contact molten salt are fabricated of Hastelloy N. In the region of the still and upper section of the condenser, the normal use temperature for this alloy may be exceeded by as much as 200°C. Consequently, the vessels are to be examined thoroughiy by dimensional, radiographic, and ultrasonic methods before and after nonradioactive operation. Provision is made for hanging test specimens of candidate metals of ‘construction in the still. g 215 OAK RIDGE NATIONAL LABORATORY MOLTEN-SALT REACTOR PROGRAM FEBRUARY 28, 1967 M. W. ROSENTHAL, DIRECTOR R. B. BRIGGS, ASSOCIATE DIRECTOR P_R. KASTEN, ASSOCIATE DIRECTOR a9on MSBR DESIGN STUDIES E.S. BETTIS . BETTIS* . BRAATZ™ . CARTER** . KERR . LLEWELLYN" . NELMS* . PICKEL™ . POLY . STODDARD* EEIATIQIEDD O EP I -Aram MSRE OPERATIONS COMPONENT AND SYSTEMS DEVELOPMENT 1&C DESIGN AND DEVELOPMENT R. L. MOORE 1&C D. G. DAVIS* 1&C P. G. HERNDON 1&C J. W. KREWSON* 1&C T. M. CATE* 18C - B. J. JONES 1&C NUCLEAR ANALYSIS p—{ B.E.PRINCE R T.W. KERLIN uv - FUEL PROCESSING DEVELOPMENT || M. E.WHATLEY cT W. L. CARTER** cT L. M. FERRIS cT B. A. HANNAFORD cT J. R. HIGHTOWER cT R. B. LINDAUER cT L. E. McNEESE cT E. L. NICHOLSON C1 C. E. SCHILLING cT F. J.SMITH cT J. BEAMS cT J. F.LAND cT C. T. THOMPSON cT i] REACTOR CHEMISTRY EIOMOED TOMTMC< - —-MJP»P»ITX nogoro METALLURGY ADAMSON* CooK* CANONICO" FRANCO-FERREIRA KENNEDY MeCOY* WERNER” L. GRIFFITH* . LANE™ LAWRENCE" SHOOSTER* THOMAS® WADDELL™ WILLIAMS P. N. HAUBENREICH DUNLAP SCOTT R NUCLEAR AND MECHANICAL ANALYSIS COORDINATION ENGINEERING ANALYSIS C. H. GABBARD™** R 4. R. ENGEL R R. J. KEDL R o : C. H. GABBARD** R L H. B. PIPER R J A WATTS R OPERATION INSTRUMENTATION AND CONTROLS R. H. GUYMON R || REMOTE MAINTENANCE J. H. CROWLEY R | J.R. TALLACKSON 1&C R. BLUMBERG R P H. HARLEY R R. K. ADAMS” 1&C J. R. SHUGART R T. L. HUDSON R C.D. MARTIN 1&C A. |. KRAKQVIAK R C. E. MATHEWS* 1&C H. R. PAYNE R J. L. REDFQRD 1&C M. RICHARDSON R R.W. TUCKER 1&C H. C. ROLLER R S.E.ELLIS 1&C R. C. STEFFY R C. E. KIRKWOOD 1&C T. ARNWINE R J.C. LeTELLIER 1&C MSRE COMPONENTS W. H. DUCKWORTH R T. SAMUELSEN 1&C A. M. SMITH R J. D. EMCH R - R. B. GALLAMER R E. H. GUINN R R. E. CARNES 'R T.G. HILL R C. A GIFFORD R C.C.HURTT R 1. C. JORDAN R H. J. MILLER R L. E. PENTON "R CHEMISTRY W. E. RAMSEY R L - R. SMITH, JR. R R. E. THOMA RC J. L. STEPP R ' 5. R WEST R J. E. WOLFE R 5. 1. DAVIS R PUMP DESIGK AND DEVELOPMENT CHEMICAL PROCESSING L] A.G.GRINDELL** R — - P.G. SMITH R _ R.B. LINDAUER cT L v WILSON® 5 DESIGN LIAISON D. D. QWENS R p— C. K. McGLOTHLAN R MATERIALS MAINTENANCE ¥. H. COOK Mec B. H. WEBSTER \_‘ A.F.GILLEN L. P. PUGH AC ANALYTICAL CHEMISTRY DIVISION CT CHEMICAL TECHNOLOGY DIVISION D DIRECTORS DIVISION GEAC GENERAL ENGINEERING AND CONSTRUCTION DIVISION 1&C INSTRUMENTATION AND CONTROLS DIVISION MAC METALS AND CERAMICS DIVISION R REACTOR DIVISION . RC REACTOR CHEMISTRY DIVISION UT UNIVERSITY OF TENNESSEE "PART TIME ON MSRP DUAL CAPACITY W. R. GRIMES* RC F. F. BLANKENSHIP ' RC E. G. BOHLMANN' RC H. F. McDUFFIE* RC MSRE ON-SITE CHEMISTRY R. E. THOMA** ' RC 5. 5. KIRSLIS RC P. 0. NEUMANN® , RC IRRADIATION PROGRAM J. M. BAKER : RC E. L. COMPERE RC H. C. SAVAGE ] RC J. E.SAVOLAINEN | J. R.HART ] RC B. L. JOHNSON RC J. W. MYERS ’ RC R. M. WALLER | RC r PHYSICAL AND INORGANIC CHEMISTRY €. F. BAES, JR.* ! RC C.J. BARTON* ‘ RC 3. BRAUNSTEIN®, RC G. D. BRUNTON* RC S. CANTOR® : RC J. H. SHAFFER* [ RC R. E. THOMA** ‘ RC C. F. WEAVER* ! RC C.E.L. BAMBERGER" RC F. A. DOSS RC H. F. FRIEDMAN* RC L. 0. GILPATRICK* RC D. M. MOULTON : RC J. D. REDMAN® RC K. A. ROMBERGER* RC * D. R. SEARS* . RC H.H. STONE* RC W.K.R. FINNELL . RC B. F. HITCH RC C. €. ROBERTS RC W. P. TEICHERT RC | I b ANALYTICAL CHEMISTRY L. T. CORBIN* AC 1. C. WHITE* AC A.S. MEYER* . AC R. F. APPLE ‘ AC C. M. BOYD* AC J. M. DALE AC F.K. HEACKER AC C. E. LAMB* AC W. L. MADDOX* AC G. MAMANTOY utT D. L. MANNING® AC J. P, | ! ! YOUNG* I AC } | i ———— . e N !lf - VENOUL AN~ O ErE MO EFUVUOONMEIAMECOOMP I NITUMIVIETON OO OO . Adams . Adamson Affel . Alexander . Apple Baes Baker Ball - . Barthold .. Bauman . Beall MmMUo-TNMOoOTX . Bender - . S. Bettis . 5. Billington E. Blanco . F. Blankenship O. Blomeke . Blumberg . L. Boch . G. Bohlmann . J. Borkowski . E. Boyd Braunstein . Bredig . Breeding . Briggs . . Bronstein . Browning . Bruce . Brunton- . Burger . Canonico Cantor PTOAOMAODS- > . W. Cardwell . L. Carter |. Cathers - M. Chandler . L. Compere A. Conlin . H. Cook T. Corbin . B. Cottrell . A, Cristy 217 INTERNAL DISTRIBUTION PAVLSIME-ATINOVOADVDIANPEINIZEOSIPIDEME-EZAN =0T - ORNL-4119 UC-80 — Reactor Technology NMEDVDOETONAFTMOPZVEOIP>PIIOAINMOPAIIP>OEZTMA DI OMOS-I=Ir T . Crowley . Culler Dale . Davis . Davis DeVan Ditto . Donnelly . Dunwoody . Eatherly Engel Epler Ergen Ferguson . Ferris Fraas . Friedman Frye, Jr. . Gabbard Gall . Gallaher- . Gilliland . Goeller: . Grimes .‘Grindell . Guymon . Hammond . Hannaford Harley . Harman Harrill . Haubenreich . Heddleson . Herndon . Hibbs (Y-12) Hightower . Hill Hise . Hoffman - . Holt . Holz . Hor'ron . Householder AAMOLCOCP>PO0NDIIAONZTEOVET TTMEBPAOCrIONrFrA>OAYME-EXTMMEMBDOOMOZ- IO Pr-mMPmMEI-YTTAEZEV-A-20I . L . Hudson . Inouye . Jordan . Kasten Kedl . Kelley Kelly . Kennedy . Kerlin . Kerr . Kimball Kirslis Knowles Krakoviak . Krewson . Lamb Lane . Larson Lawrence . Lincoln . Lindauer . Litman . Liverman . Livingston Lundin Lyon . MacPherson . MacPherson . Maienschein . Martin . Martin . Mathews . McClung . McCoy . McDuffie . McGlothlan McHargue . McNeese Meyer . Miller . Mills . 'Mixon . Moore . Morgan . Moyers Nichols . Nicholson - . Oadkes . Osborn . Parker . Parsly atriarca 158. 159. 160. 161. 162. 163. 164. 165. 166. - 167. 168-244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280, 281. 282. 283. 284, 285. “CITOO=EP>IFTINIV-WIE>I ZARE-PPO-OUEIIOODEME-0OION — U C ] C . W F E nl E H —E>TVOOMPOETM . R. Payne . M. Perry . B. Pike . B. Piper . E. Prince L. Redford . M. Reyling . Richardson @) . Robertson . Roller . Rosenthal . Savage . Savolainen . Schaffer . Schilling ap Scott . Seagren . Shaffer . D. Shipley . J. Skinner . M. Slaughter . N. Smith . J. Smith P. Smith L. Smith G. Smith . H. Sneli . F. Spencer Spiewak. . C. Steffy . E. Stevenson H. Stone . A. Sundberg R. Tallackson B. Trauger W. Tucker C. Ulrich C. Watkin M. Watson S. Watson F. Weaver . H. Webster . M. Weinberg R. Weir . J. Werner . W. West . E. Whatley C. White ) &_v‘ 290 341, 342, 343. 344. 345. 346. 347. 348. 349. 350. 351. 352. 353-354. 355. 356. 357. 358. 359. 360. 361. 362. 363. - 364, 365. 366. 367-368. 369. 370-647. 286. 287. 288. 289, -294. T-D-CMFEMETIITAETNE-ETPO-0D-- — — R | \ G. C. Williams . 295-297. Central Research Library L. V. Wilson 298-339. Laboratory Records Department G. J. Young . "~ 340. Laboratory Records, ORNL R.C. Biology Library . ORNL —.Y-12 Technical Library Document Reference Section EXTERNAL DISTRIBUTION . C. Bowman, Union Carbide Technical Center, 12900 Snow Road, Parma, Ohio 44130 . H. Brannan, Carbon Products Division, 270 Park Avenue, New York, New York 10017 . A. Charpie, UCC Electronics Division, 270 Park Avenue, New York, New York 10017 . F. Cope, Atomic Energy Commission, RDT Site Office (ORNL) . W. Crawford, Atomic Energy Commission, Washington 20545 . A. Douglas, Materials Systems Division, UCC, Kokomo, Indiana 46901 . Giambusso, Atomic Energy Commission, Washington 20545 . J. Larkin, Atomic Energy Commission, ORO A. Lieberman, Atomic Energy Commission, Washington . D. Manly, Material Systems Division, UCC, 270 Park Avenue, New York, N.Y. ]00]7 . L. Matthews, Atomic Energy Commission, RDT Site Office (ORNL) B. McDonq|d Battelle-Pacific Northwest Laboratory, Hanford, Washington W. McIntosh, Atomic Energy Commission, Washington 20545 . A. Rosen, Atomic Energy Commission, Washington 20545 M. Roth, Atomic Energy Commission, ORO Shaw, Atomic Energy Commission, Washington 20545 . E. Sinclair, Atomic Energy Commission, Washington 20545 . L. Smalley, Atomic Energy Commission, ORO . D. Stoughton, UCC, P. O. Box 500, Lowrenceburg, Tennessee A. Swartout, UCC, 270 Park Avenue, New York, New York 10017 . F. Sweek, Atomic Energy Commission, Washington 20545 . W. Ullmann, UCC, P. O. Box 278, Tarrytown, New York 10591 . J. Whitman, Atomic Energy Commission, Washington 20545 James Wright, Westinghouse Electric, P. O. Box 355, Pittsburgh, Pa. 15230 Research and Development Division, AEC, ORO Reactor Division, ORO W. Given distribution as shown in TID-4500 under Reactor Technology category (25 copies — CFSTI) W. Grigorieff, Assistant to the Executive Director, Oack Ridge Associated Universities