RUN MARTETTA £ gvs T 3 445k 03L4525 O ORNL-3282 UC-80 - Reactor Technology TID-4500 (17th ed.) Contract No. W-TLO5-eng-26 MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT FOR PERIOD ENDING FEBRUARY 28, 1962 R. B. Briggs, Program Director Date Issued OAK RTDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the U.5. ATOMIC ENERGY COMMISSION VR 3 yysk 03k4525 0 SUMMARY PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE Design Introduction.~-There was no significant change in the MSRE design concept during the past six months; nearly all the design ideas described in the preceding semiannual report remain unchanged. This report period wag concerned with detailing the many parts of the system, the kind of activity that precludes lengthy description here. Ma jor Components.--Design of all major components (reactor, pump, heat exchanger, radiator, and drain tanks) was completed, and shop draw~- ings were issued. A model of the control-rod drive was built, and tests are just beginning on the operation of the mechanism. Since some ques- tions arose regarding the stability of unclad graphite in the core, a design study was made of a core that incorporated clad graphite. Secondary Components.--The designs of the thermal shield, radiator- door drive mechanism, heat exchanger support, and pump support were com- pleted in detail. Revisions were made to the layout of electrical, instrument, and auxiliary gas and water connections to the reactor cell. Pump-~lubrication package drawings were completed. Pregssure-Suppression System for the Reactor Cell.--Further study of the hazard associated with a salt spill in the reactor cell indicated the desirability of providing a pressure-suppression system in order to ensure that no accident could raise the cell pressure to an unsafe level. A pre- liminary design was made for such a system. Electrical Distribution System.--The system was thoroughly investi- gated, and the complete electrical supply including emergency and stand- by power was dlagrammed. Detailing of the electrical heater circuits for bringing the system up to operating temperature is 50% complete. Instrumentation and Controls Design.--Wide-range servo-positioned figsion chambers will be used in the system in order to provide continuous coverage of reactor operations from startings to full power. Compensated ion charmbers will be used to provide an accurate, linear monitor of reactor power and to provide a flux signal for the servo controller. The concep- tual design of a servo controller, which will be used to control reactor temperatures at low power level, was tested on the analog computer and performed satisfactorily. The preparation of instrument-application drawings is nearing com- pletion. Nine drawings have been approved. The tabulation is almost 75% complete. The design of electrical control circuitry is continuing. The layout of the main control area was revised to provide more space in the data room. Auxiliary instrument panels will be located in a transmitter room adjacent to the reactor, in the service room at the north end of the service tunnel, and in the cooling-water room. The main control-board layout was approved after several revisions, and a l/h«scale model was constructed. Specifications are being prepared for the procurement of a digital data system that has the capability of performing on-line computations, detecting and alarming abnormal conditions, and logging process data auto- matically and on demand. The MSRE instrumentation and control system is being designed to ensure that failure of the data system will not compro- mise the safety or operability of the reactor. Requirements for process and personnel radiation monitors are being investigated. Specifications have been completed for five types of instrument com- ponents, four of which have been approved. Preparation of seven specifi- cations is nearing completion. Instrumentation components are being procured. OSome equipment will be obtained from Homogeneous Reactor Experiment No. 2. 2., Component. Development Freeze-Flange Development.--Two sets of MSRE prototype flanges were fabricated of INOR-8, and testing was started. After completion of thermal distortion and seal leakage measurements, the flanges will be installed in the thermal-cycle test facility and in the prototype pump test loop. Experiments were conducted to determine the temperature distribution and distortion characteristics of a 6~in. Inconel freeze flange, using electrical-resistance heating in the bore to simulate operation with salt. The freeze position of the molten salt was 4.8 in. from the bore when the bore was at 1400°F, whereas the predicted position was 2.7 in. from the bore. The maximum observed distortion varied from 0.024 in. at 1000°F to 0.063 in. at 1400°F. No permanent distortion was noted after 40O cycles to 1300°F. Control-Rod Development.--A simple chain-driven control-rod mechanism, using gas as a coolant, is being developed. Complete unit replacement is planned for maintenance. A 28-in.-long, U200-w section of removable heaters for 5 in. pipe was built and tested. A heat loss of 450 w per linear foot of heater was measured at operating conditions. Heater Tests.~-~The reactor vessel heaters have operated satisfacto- rily for 3750 hr, with TOO hr at 1425°F. However, the heater supports and Inconel reflector buckled due to an inadequate support design. Cor- rections will be made and the test continued. The prototype cooling bayonets for removing fuel after-heat in the drain tanks have operated for 3860 hr and 40 thermal shock cycles, with- out significant damage. Each bayonet can remove 6.3 kw from 1300°F salt. Sampler-Enricher System.--A few simplifying changes were incorpo- rated in the design of the sampler-enricher system in order to reduce maintenance problems and increase system reliability. A double-sealing gate valve was opened and closed 100 times by hand, with leak rates of less than 0.3 cc of helium per minute at operating conditions. A double- sealing, buffered, sliding seal for oxygen exclusion during insertion and removal of the sample transport container was mechanically cycled 1400 times, with a leak rate <1 cc/min with 15-psig helium buffer pressure. MSRE Core Development.--The pressure vessel for the full-scale MSRE core model was installed in its test loop, and measurements were made of the flow distribution in the entrance volute. The results agreed with previous measurements made on the l/5-scale model to within the precision of the scaling factor. Helium Purification.-~-Construction of a full-scale oxygen-removal unit is about 50% complete, and an electrolytic type trace oxygen analyzer was purchased for testing the oxygen-removal unit. MSRE Engineering Test loop.--Interpretation of the effectiveness of the oxide flushing runs in the Engineering Test Loop was complicated by the presence of zirconium fluoride, which had been inadvertently included in the salt mix. The slow-drain difficulity was the result of an addition of BeO pellets which migrated to the drain line and formed a viscous plug, later removed by raising the line temperature 60°F. Oxide sludge was manually removed from the pump bowl after efforts to dissolve it had failed. The treatment of the salt in the drain tank with a mixture of hydrogen and HF removed the equivalent of 1025 ppm of oxyzen without excessive corrosion of the container. The 8-in,.-diam graphite container was fabricated from INOR-8 and installed in the loop. A dry-box for loading and unloading graphite samples was fabricated and is undergoing test. MSRE Maintenance Development.--A program to produce an inventory of tools and procedures to cover MSRE maintenance problems was initiated. A hydraulic actuator system was successfully demonstrated for tightening the clamps on the freeze flange. The problem of removal and reinstalla- tion of flange ring gaskets was resolved, and the design of tools was started. Tools were tested for aligning flanges, jacking pipe, and re- moving the pump. A l~5/8~in.wdiam wide-angle periscope was superior to similar smaller-dilameter devices previously tested. Vi Brazed-Joint Development.--The tapered braze Joint for l-l/Q-in, sched-40 pipe was modified to use a 0.005~in.-thick sheet braze preform which reduces to approximately 0.001 in. after the braze is formed. Ultrasonic and metallographic inspection of completed prototype joints indicated 81 to 86% bonding. A representative braze joint was held at 1250°F and intermittently exposed to reactor salt for T4 hr. Mechanical-Joint Development.--Tools for remotely cutting, tapering, and brazing the 1-1/2-in. pipe joint were received and testing started. A mechanical joint, using trapped gas to keep salt out of the gasket area during intermittent use, is being designed for use in crowded areas. A moisture separator for use in a re-entrant-tube steam generator- superheater was tested and found to carry over approximately W% water. Further work is postponed. Pump Development.--The design drawings for the MSRE fuel pump were approved, and the thermal analysis of the fuel and coolant pump tanks was completed. Water testing of the model of the cooling pump was completed. Fabrication of the rotary element for the prototype fuel pump was com- pleted, and fabrication of the pump tank was nearly completed. Design drawings for the lubrication stands and drive motors were submitted for review, Additional INOR-8 castings of impellers and volutes for the fuel and coolant pumps are being made, and dished heads for the pump tanks are being inspected. MBRE Instrument Development.--A series of tests 1s being made to evaluate methods of attaching thermocouples. A test rig was assembled for developmental testing of mechanical thermocouple attachments for use on the radiator tubes in the MSRE. Development of a thermocouple scanning system, using a mercury-jet commitator, is continuing. A noise problem was eliminated by the use of make~-before~break switch action. Investigation of methods of economically obtaining signals which re- liably indicate the operating status of freeze flanges and freeze valves is continuing. A monitoring system, manufactured by the Electra Systems Corporation, and a control relay, manufactured by Daystrom Incorporated, are being evaluated. Development of a continuous-level element for use in measurement of molten salt levels is continuing. Several high-temperature differential transformer designs have been investigated, two alternate level-element designs were developed, and a level test facility incorporating the two level element designs was fabricated. Testing of the prototype level elements is underway. vii Testing of a prototype single-point level indicator was continued. A prototype of the single-point level indicator to be used in the MSRE is belng fabricated for test. 3. Reactor Engineering Analysis Reactor Physics.--Further analysis of the MSRE temperature coeffi- clents was done to obtain estimates of the effect of retaining fission- product xenon and samarium in the core graphite, and of inserting a non- 1/v absorber such as rhodium. Since the principal contribution to the temperature coefficient is the increase in leakage with increasing temper- ature, and since the major effect of poison insertion is a change in the temperature variation of thermal utilization, the calculated temperature coefficients are not appreciably changed by poisoning in the core. Results of simplified reactor-kinetics calculations indicate that a peak pressure rige of 6l psi will result from a 0.3 step reactivity in- sertion; a 1% step produces a 210-psi peak in the pressure rise, and a 0.6% step produces a 13-psi peak in the pressure rise. In the large step additions the principal removal of reactivity results from heating the fuel salt. An IBM TO90 program, 2DGH, for the calculation of gamma-ray heat- deposition rates was checked out and put into service. Results of a heating survey in the top head of the MSRE vessel show a maximum heat generation of 0.12 W'/cm3 at the lower end of the outlet pipe. PART TI. MATERTALS STUDIES 4. Metallurgy Dynamic Corrosion Studies.--Tests are in progress to determine the effect of oxidizing impurities on the corrosion behavior of fused salts. A loop containing NaF-ZrF, and contaminated with HF had uniform attack to a depth of 1/2 mil after approximately 200 hr of operation. Examination " an INOR-8 thermal convection loop containing molybdenum and graphite of inserts was completed and corrosion data are reported. The previously reported embrittlement of molybdenum specimens appears to be associated with surface contamination. A series of tests were initiated to evaluate the effects of CFq4 on MSRE core materials. No significant reaction was noted between CF, and INOR-8 at the temperatures studied. An analysis was made of data obbtained from corrosion tests containing inserts; the analysis indicates that chromium, although a small fraction of the total metal loss, plays a major role in the corrosion process. Welding and Brazing Studies.--Improvements were made in the design of the tube-to-tube-sheet joints for the MSRE heat exchanger. Weld "roll- over" was minimized and the braze trepan was deepened in order to permit viii the preplacement of an adequate amount of brazing alloy. Small mock-up test sections were brazed, and it appears that satisfactory brazes can be obtained over a range of temperature rise from 75°C/hr to 225°C/hr. A sleeve-type braze Jjoint was developed for MSRE use, and suitable brazing conditions were determined which will permit good bonding along the Jjoint length. Inspection methods are being developed for this Joint. Initially, difficulties were experienced in qualifying welders for INOR-8 work, but refinements in welding procedures have resulted in the production of satisfactory welds. The room- and elevated-temperature mechanical properties of dissimilar-metal welds containing INOR-8 were determined. In general, these welds exhibited satisfactory integrity, and failures occurred in the stainless steel, nickel, or Inconel, or at the interface between those metals and the weld metal. Mechanical Properties of INOR-8.--Mechanical properties of cast INOR-8 are being determined. Cast INOR-3 had shorter rupture life and higher minimum creep rates than wrought INOR-8. Evaluation of MSRE Graphite.--A sample of CGB-X graphite similar to graphite to be used for the MSRE was evaluated, using MSRE evaluation tests. Salt permeation tends to be restricted to shallow (less than 0.1 in.) penetration below the surfaces at a pressure of 150 psi, about three times that expected in the MSRE. There appears to be a slight heterogeneity in the accessible pore spaces. Tests indicated that a simple mercury impregnation test at room tem- perature can ve a suitable quality-control test for relating standard molten fluoride permeation into graphite. Twenty-hour-long purges with the thermal decomposition products of NH.F.HF in the temperature range from 1300 to 930°F removed oxygen con- tamination from high- and moderately low-permeability grades of graphite (AGOT and R-0025) to such an extent that there was no detectable uranium oxide precipitation from a molten LiF-BeF--UF, mixture when it was exposed to the purged graphites for 4000 hr at 1300°F. Lower purging temperatures and smaller quantities of NH.F-HF also appeared promising for the removal of oxygen from these grades of graphite. A precursory test showed that refractory monoclinic ZrOs can be con- verted to a less refractory fluoride compound by exposing it to NH4F.HF at 1300°F. 5. In~Pile Tests Interaction of Fissioning Fuel with Graphite: Test No. ORNL-MIR- 47-3.~-An experiment designed to determine whether fissioning fuel in con- tact with graphite exhibits interfacial characteristics different from the nonwetting behavior shown out of pile was operated for 1580 hr at a fuel power density of 200 w/cc. The fuel, LiF-BeFo-7rF4-ThF4-UFy (69.5-23-5- 1-1.5 mole %), contained in encapsulated graphite boats, reached maximum temperatures of about 900°C in undergoing 8.5% burnup of U?2% and still remained nonwetting toward graphite. The graphite was virtually undamaged. Supplemental observations, some of which are still in progress, re- vealed that CF4 was produced, that the frozen fuel appeared black because of discoloration by beta radiation, and that several other unusual or puzzling phenomena had occurred. The persistence of CF4 was contrary to thermodynamic equilibrium and may have been favored by the experimental arrangement. The escape of CF4 in the offgas is potentially a serious problem, mainly because the effect is the same as if a strong reducing agent were acting on the fuel. In-Pile Testing.--A description of the apparatus and the test con- ditions for in-pile test No. ORNL-MTR-4T7-4 is given. The test is de- signed to tell whether CF4 can exist in the cover gas over the fuel, when the graphite 1s submerged in such a way that the CFs formed at the fuel- graphite interface must pass through the fuel before escaping into the cover gas. Also, the experiment was planned to further demonstrate the compatibility of the fuel-graphite-INOR-8 system under thermal conditions at least as severe as those expected during MSRE operation. 6. Chemistry Phase-Equilibrium Studies.--Progress was made in the elucidation of phase equillibrium relations for the system LiF-BeFp-ZrF4, which provides an analogue of initial freezing behavior of the MSRE fuel. A partial phase diagram for the system and a number of invariant points were estab- lished. The ternary compound, 6LiF.BeFn+ZrF,, the first phase to separate (at 441°C) on cooling the MSRE fuel, was subjected to crystallographic study. Equilibrium behavior in an important composition section from the five-component system which contains the MSRE fuel was studied; the section shows the effects of diluting the fuel with the coolant and of removing 2LiF.BeF, by distillation. References are given to other phase equilibrium and crystallographic studies of fluoride systems and com- pounds of interest. Oxide Behavior in Fuels.~-~The purification of fluoride melts was demonstrated by the removal of an estimated 1200 ppm of oxide contami- nation from a charge of LiF-BeFo-ZrF4 that had been studied in the Engl- neering Test Loop. The oxide was removed as water by treating the melt at 565°C for 70 hr with gaseous HF containing 20% hydrogen. ILittle cor- rosion was observed as a result of this treatment. ILaboratory studies of the behavior of sulfate, a common contaminant of fuel raw materials, showed that sulfate ion is not stable in fluoride melts at temperatures between 500 and 800°C. The behavior of various inorganic sulfates in LiF-BeFs melts was explored at 600°C. Physical and Chemical Properties of Molten Salts.--The estimation of the densities of molten fluorides to within 2% of reported values was achieved by refinements of a method based on the additivity of molar volumes. Densitles of solid, complex, metal fluorides were estimated to within about 5% by the use of analogous assumptions; such estimates have facilitated the cholce of the number of molecules per unit cell for com- plex fluorides under study. Cryoscopic and calorimetric studies of fluoride systems were extend- ed to include freezing-point depressions in sodium fluoride containing uni- and trivalent solutes and in mixtures of NaF-LiF, and enthalpy changes from 8TL.0°C to 0°C for KF, LiF, and various mixtures in between. Refined thermodynamic values for the equilibriwm, NiFo + Ho == Z2HF + Ni, at elevated temperatures, were obtained and interpreted. The theory of molten~-salt behavior was extended by studies using molten nitrate systems. Graphite Compatibility.--The behavior of CF4 in contact with normal and partially reduced MSRE fuel at 600°C was studied in static and in gas- recirculation systems. FEvidence was obtained that CF4 can react with oxide contaminants to produce COp. Based on indirect evidence, the solubility of CF4 in MSRE fuel at 600°C cannot be greater than 1 x 1078 moles of CFgq per cubic centimeter of melt per atmosphere. Chemical Aspects of MSRE Safety.~--Chemical aspects of MSRE safety were studied. The sudden injection of molten fuel into water gave no substan- tial hydrolysis of the fuel, according to petrographic or x-ray diffraction examination of the products of the reaction, but titration of the offgas indicated a 2% yield of HF. The solubility of MORE fuel salt components in water was studied from 25 to 90°C; the rate of uranium solubility and the amount of uranium dissolved implied that neutron poisons should be provided in any water which might come in contact with the fuel. To permit safety calculations on criticality hnazards arising from segregation by partial freezing of the fuel, a hypothetical crystallization path was defined, and density of the concentrated fuel was estimated. Fluoride Salt Production.--The time required for the purification of fluoride salts in the Fluoride Production Facility was shortened by the adoption of a combined HF-Hs treatment. Current modifications in the size of salt-transfer containers will give a 50% increase in the production rate. The addition of premelting furnaces is being studied as a possible means to a further increase of 6L4% in the production rate. Analytical Chemistry.--Analytical studies included (1) evaluation of methods for the determination of oxygen in the nonradioactive MSRE fuel and (2) a survey of methods applicable to the complete analysis of fuel samples from the reactor during critical operation. T. Fuel Processing MSRE Flowsheet,~~Spent MSRE fuel will be fluorinated to recover uranium. Methods for disposal of the excess fluorine are being investi- gated. Fluorine Disposal Tests, Using Charcoal.--Fluorine disposal by ve- acting fluorine with charcoal was shown to be nonexplosive and greater than 99.99% effective in seven runs. Solid and 1iguid products were trapped from the offgas, which contained 52,6 mole % CFy4. xi CONTENTS MMARY S 0 0 0 0 ¢ 000 02 OO O OO OO OO P O OO OO S E OO 0P ONE S eSS e SO eSO OO GG iii PART I. MOSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MORE DESIGN 4t eececesoescoasovoseocssssossssscosossscssoosoasoss INtroduction ccececesesescacecanscscosscsscsoscssonssnasese Reactor Core and VesSsSel ..eeeecresscooscosccocnscococsocs Primary Heal EXChanger ...ecececescccocococscsoscscsssone Radiator coeseseecseescoersersesesceosasacsssaccescscacacss Fuel-Salt Drain Tanks ..eeeeecscsccscccsscscoosscososons Equipment Layoul cceeecececsecccesessscscccsscsssssasoas Cover-Gas OYStemM seecsececssosesosesossssrsscscscsasossssne Oystem Healers teeeeceesecsessscscsscssesosccosccosss cevene Maintenance Design, Assenmbly Jigs, and Fixtures ........ Reactor Procurement and Installation ....ccceeccocccceces 1.10.1 Major Modifications to Building 7503 .veeecocens 1.10.2 Construction Outside Building T503 ceieeveosones 1.10.3 Planning and Scheduling of the MSRE * o H\0 0= O\ =W O O - OTIIT0O O DWW Installation ..... ® 0 O 0 @& O ¢ O OO B PO O 0P8BS P9 O C 0O 0 S s 9 1.10.k Procurement of Materials ....ececececoscssceas .o 2 1.10.5 Procurement of Components eeeeceesecess cecssceocss 10 1.11 Reactor Instrumentation and Controls System .eececececess 10 1.11.1 Nuclear Control System .eeeeeeececsssesescsosccnas 10 Instrument-Application Diagrams ...eeeeecccceses 14 Electrical Control CircultYy .eeceececccecossnccos 15 Layout of Instrumentation and Controls SYystem .eeeveeeoeecocosoesssoseacsscass 15 5 Control Panels seeeeesescecscsscsssoscscassoanes 17 6 Data Handling ceeeeeececeoeeccooscesccsscssoonss 18 .T Process and Personnel Radiation Monitoring ..... 18 8 Procurement StALUS .eeeeeeeeerreccosceccosccnsos 19 i =l Sy =w 2, COMPONENT DEVELOPMENT +ecececocscessaccesoesosascanssosscscsscoaces 21 2.1 Freeze-Flange Development .cevieeeceerecccccsocnccsscnccs 21 2.1.1 MSRE 5~in. FlangesS .eeeeccececsaccessccass cecocse 21 2,1.2 Freeze-Flange-Seal Test Facility ..... sessccncse 2L Control-Rod Development c.eeeecssescoscscssscsssssssssns Ry N w O Heater TeSES ceeseeeoccceeasosesscssscesssssocsssscssesscs 26 2.3.1 Pipe Heaters .eceeeeeserscescsscccosscssscssscscans 26 2.3.2 Reactor Vessel Heabers ,...eeeeescsscccccccessos 29 2.1 Drain~Tank COOLETS eeceescecccerssensccsooccsssoscsescosss 29 2.5 Sampler-Enricher SySTeM ceceeesscessssseccsscossssonocoas 29 2,.5.1 General Concept of Sampler-Enricher .eeeeceeceses 29 285.2 Operationalvalve Q2 6 0 0 O 0O DD S O O8O O P 9SO e SN e O ’3—]‘ 3. s <11l .12 .13 oo Xil 2.5.3 Removal Seal for Sample Container ...... 2.5.4 Detail Design of Sampler-Enricher .......... ce MSRE Core Development ...eeceeceeoccosecns ceceeesesoaans 2.6.1 Full-Scale Core Model ....eeeeeeveeecocoennnns .. 2.6.2 Core-Inlet Flow Distribubion ..eeveeeceeeeens . Helium Purificalion ....eceereecerseossocccnseoscssansns MSRE Engineering Test Loop (ETL) ........ 2.8.1 Toop Operalions ..eeeveeeereeenoeeees R .. 2.8.2 Opecration of Freeze Valve .eeeeeeeeeeoocccconns 2.8.3 Oxide Removal v..eeeeeeeen. Ce et ceeee 2.8 0 HEF TreatmMent veeeeeeeeeeeseoeeeesananceeeaoneses 2.8.5 FTL Graphile Facility voeeeeeeeeeeeereeceecenns MSRE Maintenance Development ....ceeevee ceeeccesrcsaan s 2.9.1 Placement and Removal of Freeze-~Flange Clamp ... 2.9.2 Flange Alignment and Pipe-Jacking Tools ........ 2.9.3 Gasket-Replacement Procedures .....eeeeeeseen . 2.9.h Miscellaneous Disconnects ..eeeeseces.. ceeen 2.9.5 Remote Viewing ..eeeeceeeceees . 2.9.6 Component Removal .......... Ceeecesseeseeeeees . Brazed-Joint Development ....eveveeveeosoereoooncsns oo 2.10.1 Joint Desigll seeeeeseescsosoesosasesaanocsnasas 2.10.2 Brazed-Joint Testing ..eeeeeveeeceneconosas ceosoe 2.10.3 Remote Fabrication of Braze Joint ...ceeeeeeeen. Mechanical-Joint Development ....ceeeeeecosss e esen ceee Steam Generator ...ceeeeseess cesens S e essenussssasencases Pump Development ..eeeeeeresersrscncessescsssssosscoanosns 2.13.] MORE Fuel Pump +..veveverececrereroscsoooonaosss 2.13.2 MSRE Coolant PulMD .eveveeeccecoosoceccooononccses 2.13.3 Advanced Molten-Salt PUmMpPS ceeeeeeoccenss . MSRE Instrument Development ....eeeeeeeseesocosccccocnas 2.1.1 Thermocouple Attachments ..v.eeeeeeeeesoonenss .. 2.14.2 Temperature SCANNET «evveevereveeeeroreeones 2.14.3 Single-Point Temperature-Alarm System .......... 2.14.4 Pump-Bowl-Level Indicator ....... et eeacnann 2.14.5 Single-Point Level Indicator .e.eeo... N REACTOR ENGINEERING ANALYSTS ..eveeieveescrescossvosonne 3.1 Reactor PhySicCs teeeeeeescccscesssesssessaosssnscsancssnns 3.1.1 Analysis of MSRE Temperature Coefficient ....... 3.1.2 Reactor-Kinetics Studies ....eeeeecorercconcenes 2.1.3 Gamma-Heating SUTVEY ceeeeeersesveesosescoconcas PART TIT. MATERTALS STUDIES METALLURGY 4t eeoeoeeossaceecaosenaccnansnns ceerecans cecensnennn 401l Dynamic~Corrosion SLUALES weeeeeeeeeeeenorsenessanananns 4.1.1 Fluoride-Salt Contamination Studies ..... . . 4ol.2 Molybdenum~-Graphite Compatibility Tests ...... . 4.1.3 Corrosion Effects of Carbon Tetrafluoride ...... ol 4 Bxamination of Corrosion Inserts from TNOR-8& Forced~Convection LOOPS veeeeseon. ceesesacaanne 4.2 Welding and Brazing Studies ve.eeeeeees. cheenen ceveseas 4.2.1 Heat Exchanger Fabricabion ...eeececee. ceraene . 4.2.2 RemMObLe BraZiNg tveeeeeeeeeeeeeeanneennaanns oo o & o 0 0 0 0 0 31 32 32 52 32 33 34 35 35 37 37 41 41 4l 41 43 43 4 45 45 46 51 51 51 52 55 56 58 58 58 60 6l 66 68 68 68 68 70 2 72 72 4 7l 79 81 g1 8/, xiii 4.2.3 Welding of INOR=8 sceveevevscoscoscssoscesacscas 86 L.2.4 Mechanical Properties of INOR-8- Dissimilar-Metal WeldS seeeeescecescccsscoccns 88 4.3 Mechanical Properties of INOR-8 ..ceeeeeecoesconscnsaess 89 4.h Evaluation of MSRE Type Graphite .eieecececesscecsssccas 89 h.4.1 Comparison of the Permeation of Graphite by Mercury and Molten Fluorides ....cceeceecee 92 h.h.2 Removal of Oxygen Contamination from Graphite with Thermal Decomposition Products of NH4AF 'HF .eieeecececcecscosooscnsnscs 94 b,k.3 Reaction of ZrOs with Thermal Decompo- sition Products of NH4AFHF ...eeveeeeonces ceaes 94 5. IN"PILE TESTS S 8 6 0 0 6 0 6 6 6 0 8 6 O 8 00 P O 0 OO OO E OO SO L S OO e PO 0N 00N e 9’7 5.1 Interaction of Fissioning Fuel with Graphite: Test No. ORNL-MIR-UT-3 .iveiveeeoerecesaeas 97 5.1.1 Description of Experiment c..eeeeecescsccesvence o7 5.1.2 Dismantling of In-Pile Assembly ....... cesesens . 99 5.1.3 Temperatures ....ce.. cevecersocs cesessesesesasss 100 5.1.4 Gas Analyses .e.ceeeeecees Ceesesecscececaseas ... 100 5.1.5 Test Effects on Graphite ...ieeveseeecescnss ceees 105 5.1.6 Analyses of Graphite .ceeeseeeeccsecosesscecoses 106 5.1.7 Test Effects on COUPONS +evvevsescasssssssscsess 106 5.1.8 Test Effects on FUELl vvveeeeereeceaannnns cesess. 106 5.1.9 ConclusSions .eeeeecceessessccsssssoscsscsssasssas 110 5.2 MSRE In-Pile Testing eveececceocssssscscessosssessssecess 110 6. CHEMISTRY +vvvcevocnsoncescacsossosaasssososcassssansos cerasees 114 6.1 Phase-Equilibrium Studies ....ccoeeveveeeee ceececenneeee. 114 6.1.1 The System LiF-BeFo-Z1F4 sevececcesceococenseess 114 6.1.2 The System LiF-BeFo-ZrFs-ThF4-UFs .vvvvveeanee.s 116 6.1.3 Phase Equilibrium Studies in Fluoride OYSTEMS sveeeescossescscososesesoasscsssnsasans 110 6.2 Oxide Behavior in FUELS .eeveeececcssossccsosssssssssseas 117 6.2.1 Removal of Oxide from a Flush Salt ...eesveesess 117 6.2.2 The Behavior of Sulfates in Molten FIUOTIAES +oveeoscecscosnsooscesovasosansenens 118 6.3 Physical and Chemical Properties of Moplten Salts ....... 121 6.k Graphite Compatibility seeeceescecoerscacsssscsnsasseses 122 6.%.1 The Behavior of Carbon Tetrafluoride in Molten FluoridesS ceevececscesosccoessnesess 122 6.5 Chemical Aspects of MSRE Safeby veeeeececcercccceneeasss 124 6.5.1 Physical Effects of Mixing Molten Fuel and Water ..eecesececceecscoassossscaesoas 124 6.5.2 Solubility of Fuel-Salt Components in Water s.vseececeeecrssscccosoosscsscceancaes L27 6.5.3 Solubility of MSRE Coolant in Water .eeceeeeeees 129 6.5.4 Partial Freezing of MSRE Fuel .veevevereen ceeees 130 6.6 Fluoride-S5alt Production ..eeceeecereeceecoeracesaeeaess 130 6.6.1 Production RateS .eveeeeeeeoceosscscocossoncosees 130 6.7 Analytical Chemistry coeeeececcessoessscoscoscssseacesss 131 6.7.1 Oxygen in Nonradioactive MSRE Fuel ...eeveeeeee.. 131 7. Xiv 6.7.2 Adaptation of Analytical Methods to Radioactive Fuel ...i.iceveeeencens e ¢ ¢ &6 & 0 0 0 0 mL}?ROCESSD\]G @ & 5 & 6 0 0 6 0 0 00 & b SO SO S O 0L S e 0SB s 0000 » 8 6 06 ¢ 0 6 4 0 0 0 7.1 T.2 MSI{EFlO.‘JSheet ® & & & & & 0 8 0 O 5 00 " 8 P8OOSOt e e e0e e Fluorine Disposal Tests, Using Charcoal ......... ® & o & 0 0 0 132 134 134 134 PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE DESIGN 1.1 INTRODUCTION Accomplishments in the past six months were mainly in the field of working drawings, and there were very few design studies since no change was made 1in the system concept. A little more work than was anticipated resulted from moving the pump off the top of the reactor. Design of a satisfactory pump mount, which provided sufficient degrees of freedom for the pump, took more effort than expected. This mount has now been detailed and will be tested on a hot-pump~test stand. Since no lump-sum bidder could be found for component fabrication, these items were scheduled for local fabrication. This necesgitated the making of shop drawings for components showing weld details, etc., which otherwise would have been left to a vendor. While access to the top of the reactor is relatively unencumbered, the control-rod thimbles must bend slightly in order to clear the graphite- sampling-port flange. This bend made 1t necessary to have some flexibil- ity in the control-rod drive mechanism. Since cell atmosphere is used to cool the control rod, some means of containment of fuel salt in case of a control ~-rod~thimble rupture imposed certain restraints in the control rod and its drive mechanism. It was decided to build and test a model before issuing formal design drawings for the drive mechanism and associated cool- ing and containment piping. The model has Jjust been completed, and testing has started. The auxiliary reactor-cell work is nearing completion. The thermal- shield design is finished. Most of the support steel within the cell has been detailed. Analyses of stresses in hold-down bolts and containment- cell skirt cylinder were made. Pipe-support detalls were designed, but the drawings are not yet finished. Penetrations, disconnect locations, and assembly tool accessories have been established. Circuit layouts for the electrical system are approximately 75% com- plete, and no problem remains other than preparing the necessary drawings. Layouts have been made for the drain-tank heaters. Power and control points for the heater circuits have been designated on electrical one- line diagrams. Drain-tank-cell piping has been layed out and slightly simplified. Drain-tank support structure with weigh cells has been completed. Work has not yet started on the detailed layout of piping and support in the coolant cell. However, the radiator with its enclosure and heater cir- cults is completed. Instrumentation and control design effort was concentrated on the final phases of conceptual design and on specification writing, tabula- tions, and similar procedures required to producc detailed design drawings. Conceptual design was virtually completed, and, in some areas, component procurement was initiated. 1.2 REACTOR CORE AND VESSEL The reactor core and vessel underwent no design changes. For infor- mation, the cutaway drawing shown in the last report is repeated in Fig. 1.1. UNCLASSIFIED ORNL-LR-DWG 510978 FLEXIBLE CONDUIT TO CONTROL ROD DRIV SAMPLE ACCESS PORT" i ES —— 2 ?///v"COOLING AIR LINES CORE CENTERING GRID FLOW DISTRIBUTOR % GRAPHITE-MODERATOR STRINGER —— | FUEL INLET REACTGR CORE CAN REACTOR VESSEL— e i i B il 'v‘\ I I ANTI-SWIRL VANES— . The volume of the coolant holdup in the heat exchanger is 5.7 ft°. 'The heat exchanger when installed and full of fuel and coolant weighs 3500 1b. Heat exchanger mounting was designed to allow for movement of the exchanger with load and with thermal expansion of the system. 1.4 RADIATOR Work on the radiator, =2xcept for thermocouple attachments, has been finished for several months. The radiastor enclosure and door drive mech- anism required additional detailing. Attachment of heaters inside the radiator enclosure required many drawings. This was true also of the door drive mechanism. The entire radiator package is now complete, and some fabrication on the support structure has already taken place. Thermocouples were added to each of the 120 radiator tubes near the cold end of the radiator in order to monitor as accurately as possible the low salt temperature in order to prevent freezing in the radiator tubes. 1.5 FPUEL-SALT DRAIN 'TANKS Some changes were made in the drain tanks. The volume of the tanks was too small to allow ample margin in the freeboard in case unforseen excessive temperatures should obtain. Therefore the diameter of the tanks was increased from 48 to 50 in. to provide ample margin of safety in free- board volume for every reasonable eventuality. One other change in the drain-tank design involved the addition of a penetration in the tank dome for the insertion of ligquid-~level probes. With these changes, the shop drawings were completed and issued for construction. The support structure for the drain tanks, including the weigh- cell -mount detalls, was completed. 1.6 EQUIPMENT LAYOUT The concept of employing a batching operation in the Fluoride Vola- tility Pilot Plant for fuel processing was abandoned in favor of eventual installation of facilities for on-site fluorination of the fuel in the spent-fuel storage tank. This will result in the elimination of the "fuel transfer cell" from the layout plans. A waste tank will be required for storing carrier salt after removal of uranium. Layout within the reactor cell is complete but will be undergoing constant modification to a slight degree as more complete information on maintenance tools and practices becomes available. An overflow pipe was added from the pump bowl to the drain line downstream of the freeze valve to eliminate any possible danger of overfilling the pump gas space because of some inadvertent or unforseen rise in fuel temperature. Layout in the drain-tank cell is in the same status as that within the reactor cell. The salt piping here has been completely redone, because the first layout was thought to be intolerably crowded and conplex. All auxiliary systems have now received some attention. The punp- lubrication layout 1s complete. The component cooling system, water cooling system, and electrical service to the cell are better than 90% complete., Work has barely started on support details for the coolant piping. A vapor-suppression system was added as part of the containment (Fig 1.2). The large reactor-cell ventilation line has a tee which takes off from the cell side of the two large valves. This line goes to a water-filled tank so that discharge gases from the line pass through a large volume of water. The water-filled tank vents into another closed volume in the form of a shielded tank for containing noncondensable gas which escapes from the first tank. Temperatures, pressures, required UNCLASSIFIED ORNL-L.R-DWG 67462R 40 ft FROM REACTOR BLDG. . ELEV. 858 ft 2-in-DIA VENT LINE - TO FILTERS AND STACK / SPECIAL EQUIPMENT ROOM RN IN REACTOR BLDG.- FLOOR ELEV. GROUND ELEV, FILL 852 fi 851 f1-6in. A/ A ELEV. 848 1 1] 5 S \10—in;~DIA RELIEF LINE 10-1 DIA x 54—t LONG 3900 13 BURSTING DISK R f foem ~ AOO ] 4-in. LINE ELEV. 836 fi 30-in-DIA REACTOR CELL VENTILATION DUCT | T I ELEV, 830 ft 10~ft DIA x 23-ft HIGH N |~ 1800 ft3-1200 13 OF WATER%; ) 1l i ELEV. 824 fi Fig. 1.2. Diagram of MSRE Pressure-Suppression System. holdup capacity, and shielding requirements were calculated, and a pre- limipnary layout was made showing arrangement, pipe and tank sizes and locations, and shielding. 1.7 COVER-GAS SYSTEM Work was continued on detailing of system piping and components. Design memorands were issued on the oxygen-removal units, the helium surge tank, and the secondary containment and shielding for the off-gas line and charcoal adsorber pit. The cover-gas system is about 00% com- plete. Routing of pipe between major components constitutes the major remaining design. 1.8 SYSTEM HEATERS Heaters for the coolant circuit were layed out for the entire pliping system. These will be conventional clamshell-type ceramic heaters. They are not required to be remotely removable; so conventional-type insulation will be installed over the heaters on the pipe. For pipes requiring remote maintenance, that is, all the salt piping in the reactor and drain-tank cells, some kind of removable heating and insulating "package'" must be made. Design has taken the empirical approach. A full report on pipe-heater design is given in the Chap. 2. Design of the actual pipe-heater module to be used in the MSRE has Jjust started. Drawings will be based entirely on the mockup unit. Heaters for the reactor vessel also followed the empirical approach, and design drawings are being made of a thoroughly tested prototype heater. Heaters have been decsigned for the dralin tanks. These consist of standard heater units arranged in sectors which, when combined around a tank, effectively place the tank within an electrically heated furnace. Since the pump, with its rather long inlet pipe, constitutes an oddly shaped unit, a different heating method has been employed. Here an insulated furnace is constructed around the pump, enclosing the inlet pipe. Into the top of this enclosure and around the periphery of the pump bowl, arrangements have been made to insert commercial-type bayonet heaters. This makes the pump-heating plan similar to the one used for the reactor vessel, except that the actual heater units will be of com- nercial design and will be purchased items. All heating designs for the salt systems have now been determined; drawings are in various stages of completion, but none are yet ready for issue. 1.9 MAINTENANCE DESTGN, ASSEMBLY JIGS, AND FIXTURES The design of the Jjigs and fixtures Tor the assembly of the initial and replacement components was halted, awaiting firmer system design. It has been restarted and 1s proceeding on the basis of using optical tooling methods for the precision location of mating parts. Optical tooling will be used to make in-cell measurements of mating points at the time that component replacement becomes necesgary. This procedure will reveal any deformation that may have occurred from operation. 1.10 REACTOR PROCUREMENT AND INSTALLATTON 1.10.1 Major Modifications to Building 7503% The prime contract to make major modifications to Building 7503 wa.s awarded to the Kaminer Construction Corporation of Atlanta, Georgia. The major portions of this work include modifications to the 24-ft-diam containment vessel, modification of the drain-tank cell, modification to the coolant-system cell, installation of the building ventilation system, and erection of the secondary contalinment walls inside the high- bay area. Construction work started in November 1961 and is scheduled for com- pletion about October 15, 1962. To date, the following work has been accomplished by the contractor: 1. Excavation of the drain-tank cell was completed. The structural steel hold-down beams, both vertical and horizontal, were welded in place, and a concrete leveling pad was poured on the floor of the drain tank. The reinforcing steel is being set preparatory to pouring the concrete floor of the tank. 2. Demolition of the penthouse wall, above the coolant-system cell, wa.s completed; the structural steel was set in place and the forms were erected Tor pouring the new concrete wall for the penthouse. 5. Off-gite Tabrication of the stainless steel grid and liner for the drain-tank cell is almost complete. This will be installed in March. 4. Off-site fabrication of the T-ft extension of the 24-ft-diam containment vessel is in progress, as are the stainless steel diaphragm and the shielding support beams. On-site work on this containment vessel is scheduled to start in March. 5. 0Off~site fabrication of the carbon-steel structure and walls for the secondary containment is in progress. Figure 1.5 shows construction progress in the drain-tank cell. UNCLASSIFIED PHOTO 56727 i in the Drain-Tank Cell Down Beams lold- ion of | Installat 3. 1. g F 1.10.2 Construction Outside Building T750% Design drawings and specifications for the construction work exterior to Building 7505 are completed, and the work will be advertised for bids in March. The major portions of this work are the off-gas filter pit, the erection of a 100-ft-high carbon-steel stack, and the installation of a cooling tower. The work is expected to be done concurrently with the existing work under contract with Kaminer Construction Corporation and is expected to be completed at about the same time. 1.10.5 Planning and Scheduling of the MSRE Installation A detalled review has been made of the activities required for pro- curing and fabricating components for the MSRE and for installing the reactor system in Building 7505. The information was used to develop a critical~path diagram and schedule for the MSRE work. The information was coded for the IBM 7090 computer, and preliminary results indicate a completion date of July 1963%. A more detailed review is being made of the critical and near-critical activities to verify this estimated com- pletion date and to determine if an earlier date is possible. The schedule is based on the following assumptions: 1. That the lump-sum contract work will be completed and the cells will be available for equipment installation by November 1, 1962. 2. That about 80% of the reactor and drain-tank process and service piping will be prefabricated between July 1, 1962, and December 30, 1962. 5. That about 50% of the process piping, service pliping, power and thermocouple cable will be installed in the reactor and drain-tank cells prior to the installation of the process equipment. The reactor components are to be preassembled on a jig outside the reactor cell. In order to minimize the effects on the schedule of this preassembly, as well as the effect of any unanticipated delay in the completion of the building-modification work, the process and service lines will be prefabricated as indicated. 1.10.4 Procurement of Materials Delivery of INOR-8 plate, rod, and weld rod started in November 1961; by February 15, 1962, about 80% of the material had been received. Approx- imately 111,000 1b of this material was manufactured by Haynes Stellite Company at an average price of $%.38 per pound. Contracts were awarded in September 1961 to International Nickel Company, Michigan Seamless Pipe and Tube Company, and Wall Tube Company for manufacture of 18,000 linear feet of INOR-8 seamless pipe and tubing. This material weighs approximately 18,000 1b and is being manufactured for an average price ol about $18 per pound. The pipe and tubing are scheduled for delivery in May 1962. 10 Taylor Forge Company was awarded a contract in September 1961 for the manufacture of 420 INOR-8 pipe and tubing fittings for an average price of $260 each. Delivery of these fittings is scheduled to begin in March 1962, with completion by May 1962. The National Carbon Company was awarded a contract in September 1961 for the manufacture of approximately 9000 1b of moderator graphite, completely machined to specifications and tolerances, for an average price of $21 per pound. It is scheduled for delivery in July 1962. Contracts have also been awarded for the manufacture of 150 flexible Inconel metal hoses for the fuel drain tanks. 1.10.5 Procurement of Components Since efforts to obtain lump-sum bids for fabricating MSRE compo- nents were unsuccessful, because of industry's lack of familiarity with TNOR-8, the fabrication of the major MSRE components (reactor, radiator and enclosure, heat exchanger, fill and drain tanks, and pump parts) is being performed in the Y-12 Machine Shops, where experience has been years. Fabrication 1s started in the shop as material becomes available and as fabrication drawings, prepared for outside vendors, are modified for use in the Y-12 Machine Shop. Work was started in January on the radiator enclosure and in February on the reactor. By February 15, about 18 men were working on these two jobs. This manpower load will increase to a peak of about 125. Completion of the fabrication of all the major reacltor components is scheduled for October 1902. Subcontracts for fabricating INOR-8 have been awarded to the Taylor Forge Company for forging the heat-exchanger tube sheet; to the UCNC Paducah Machine Shop for forming the reactor inlet volute, radiator headers, and dished heads for the heat exchanger, radiator, and reactor vessel ; and to the Lukens Steel Company and Phoenix Steel Company to form dished heads for the various fill and drain tanks and pumps. The dished heads for the radiator, heat exchanger, and pumps have been delilvered. Scheduled delivery of the other TNOR-8 fabricated items is for March and April 1902. 1.11 REACTOR INSTRUMENTATION AND CONTROLS SYSTEM 1.11.1 Nuclear Control System The nuclear instrumentation and controls system is block-diagrammed in Fig. 1.k, The wide-range servo-positioned fission chamber channels? are expected to provide continuous coverage of the full range of reactor 11 8 UNCLASSIFIED { —_— Afow Afreactor ORNL-LR-DWG 67585R | RANGE i FLOW "5 ¢ 7 8 i CHANGE : % ?1 2 o i ¥ COMPENSATED 1 \ {ON CHAMBER ! POWER SUPPLY \LOG» LOG: TWO PEN Q-1045 | a99s / RECORDER MICRO-MICRO : LEVEL | RANGE ¢ SERVO POWER ( ™ AMMETER B CONTROL AMP MEAN COMPUTER TEMPERATURE L05; SET POINT E 17 LINEAR Z POWER RANGE S F Y S | CHANGE > ‘ COMPENSATER ALTERNATE CONNECTIONS SO THAT [ FLUX SET POINT 1ON CHAMBER SUFF EITHER CHANNEL MAY BE USED TO ‘ &-1045 POW;}';%U LY ; PROVIDE SERVO INPUT { MICRO-MICRO i . LIiIPNTIE'R | AMMETER ’ L ~2 LOWER 1 'd LIMIT | | | LIMIT ! | SW. UPPER LOWER o —_— LOG LIMIT LIMIT COUNT RATE | POWER | SET POINT | RECORDER [ R POWER SERVO | LOG POWER _— O — | T [~ AMP, AMP, . | - oo} i : . ONEAR PERIOD COMPUTER| \'J | i manuaL | COUNT RATE ! PERIOD { POSITIONING | METER CORDER -t - ! LINEAAR LOG 1 PREAMP | PULiED MP. COUNT RATE 0-2263 | METER FISSION CHAMBER ] DISCRIMINATOR ‘ Q-2217 PREAMP - UPPER POWER SUPPLY SCALER CimIT Q-2263 WIDE RANGE FISSION CHAMBER AND PREAMP. Q SERVO < FLEXIBLE ASSEMBLY RC 2-12-3 FISSION LOWER CHAMBERS r 7 LIMIT ? l POSITION ' : I ’ UPPER i 1 COUNT RATE : LIMIT { i f SET POINT ™ i POWER SERVO LOG POWER | AMP, amp. | 1 LOGE i LINEAR PERIOD COMPUTER i COUNT RATE METER - * LINEAR i} | PULSE amp l METER § A FISSION CHAMBER PREAME { DISCRIMINATOR Qe POWER SUPPLY SCALER N Q-2263 . ANNUNCIATOR l + T \ | .~ ALTERNATE CONNECTIONS {— ‘ ! LOG] DATA LOGGING POINT — i / NEUTRON —— SENSITIVE AMPLIFIER LOG . — CONTINUITY MONITOR \Efi%fi\ — AMP. FS ION CHAMBERS ’\ ANN. %;/ REACTOR QUTLET 1 2aN TEMPERATURE ———t=di FUEL SALT FLOW RATE OTHER INPUTS J {FLOW, RADIATION, LEVEL, ETC.) TRIPS LOGIC Fig. 1.4, Block Diagram e t SHIM ROD ? U NO.3 MANUAL SHIM CONTROL RELEASE ALL RODS of MSRE Nuclear instrumentation and Control System. SHIM ROD NO. 2 &) TORQUE-SYNCHRO POSITION RECE!VER AMPLIFIER WITH POWER SUPPLY INDICATING LAMP | TORQUE-SYNCHRO POSITION TRANSMITTER CLUTCH (ELECTRICALLY OPERATED} ELECTRICAL CONNECTION ————— ALTERNATE ELECTRICAL CONNECTIONS ———— MECHANICAL POWER CONNECTIONS 12 operation, from startup to full power. The linear channels, with com- pensated chambers as the signal source, are the most accurate nuclear information channels and are used to monitor reactor power and o provide the flux signal for the servo controller. These channels have a span of four decades; and, by combining range switching and chanmber relocation, they are useful over a wide range. 'The limit channels will be used to provide a "reverse' mode flux signal which will call for all the rods +to be Inserted at an average rate of 0.040% Ak/k per second. Figure 1.5 diagrams the reactor power ranges of the above nuclear instrument. During initial critical periods, two temporary counting UNCLASSIFIED ORNL-LR-DWG 68575 FISSION-CHAMBER CHANNELS SCALER LOG CR LINEAR NEUTRON LOG POWER FLOW CHANNELS INVERSE POWER LIMIT CHANNELS A} (SEE NOTE 1) (SEE NOTE 2) (SEE NOTE 5) 10° S - S SEE NOTE 4 ol %-—m—-~1w~\ T S — z © 1t — L e e e ] 101 | — e o 4 mofi4w\\{r ——— 10‘1\\;i; e ] L - b S U 40”2._.‘fi [ (O S A 10_3L~—1mwf‘]\4-~—-' e 1. Ceonsists of compensated chamber to micromicroammeter. Used for servo flux input signal as well as recorder input. Lower threshold, possibly one decade below that shown, if all conditions optimized. Range switching not shown, 2. These are servo-operated fission channels and can be used to provide all readouts shown. 3. Under ideal conditions {low elecirical noise and relatively clean fuel) the servo will control in the 1C-Mw region. 4. Range of servo-operated fission channel dependent on movement in attenuating medium (H 20), pro- vided that no gross flux distortion exists in medium traversed by chamber. 5. The two-decade span may be located over a wide range, Fig. 1.5. Ranges of MSRE Power and Nuclear Instruments. 13 channels with BFs chambers located in the composite shield will be used. Except for the BFy counters, all nuclear instruments will be located in the large 36-in.-diam water-filled penetration, inclined approximately L5° (ref 2). The servo control system will be derived from a conceptual design (Fig. 1.6) proposed by E. R. Mann. Servo control of the M3RE is required, because its extremely conservative operation yields a very high value for the ratio of heat capacity to power density. Inherent self-control by temperature coefficient is therefore sluggish and productive of power oscillations which, while not dangerous, are not acceptable from an oper- ational standpoint. The need for servo control will diminish rapidly if reactor power is allowed to exceed the present 10-Mw limitation. UNCLASSIFIED ORNL - LR-DWG 68576 FLOW POWER REMOVED- e e o e o e e e e e e e e — — —— —— - RE ACTOR POWER- ¢ T T T T T T T T T T MEAN SERVO 6; — "] | TEMPERAT URE MOTOR | \ lERROR SIGNAL cowen | € ' "+ B, AMPLIFIER S e e 1y | ‘ | l | " l | ~ Ll | REGULATING | | | SET POINT, 8gp | ROD } 'I | ! | | €=4 [qb~ F(6,—6)~¢ (/)] -8 [BSP - 2] Fig. 1.6. Block Diagram of MSRE Reactor Controller, Functionally, the servo controller simply augments the temperature coefficient of the reactor; in the power range it is not used to control reactor power on operator demand. The conceptual design was tested on the analog computer® and per- formed satisfactorily. The translation of the controller from conceptual design to hard- ware is in process and consists of substituting and packaging suitable industrial ~grade components for computer-operational amplifiers and the 14 addition of a power amplifier to match the shim-rod motor. The control package will be thoroughly tested and will be used in conjunction with the analog simulator for operator training. 1.11.2 Instrument-Application Diagrams During the past report period, the 15 instrument-application flow diagrams listed in Table 1.l were revised and issued for comment. In general, the result of the revisions was a reduction of the number of instruments and thermocouples required. Nine drawings were approved for construction. It is estimated that this basic part of the instru- mentation effort is 90% complete. Table 1.1. Instrument-Application Flow Diagrams No. of Flow Diagram Application P-AA-B-40500" Fuel-salt clreuit 40501 Coolant-salt system Lo502™ Fuel, flush, and drain-tank fault system MO50§a Cover-gas system Lo50L" fuel-salt-pump lubricating-oil system MO505a Fuel sampler-enricher system 40506" Liquid-waste system 40508 Coolant-salt-pump lubricating-oil system MOSOQa Water system MO5ZLOa Off-gas system 40513 Fuel, fill and transfer system Los1 kL Instrument~air and service-air systems h0515 Containment air aProjectmapproved drawings. A tabulation of all instruments shown on the flow diagrams, giving identifying numbers, location, function in process, and a brief descrip- tion, was also revised in accordance with the latest revision of the application diagrams. This tabulation is nearly 75% complete. 15 Although 11 thermocouple-location drawings were approved for con- struction, they were revised because of the reduction in the number of thermocouples required., A list of all thermocouples, showing the type and location of the readout device for each one, was issued for comment. 1.11.5 Electrical Control Circuitry Control-circuit design is continuing. Elementary schematic drawings showing basic electrical control and the safety circuits were nade for the sampler enricher, fuel- and coolant-pump lubricating-oil system, water sys- tem, contalnment-air system, and the coolant-salt radiator, 1.11.4% ILayout of Instrumentation and Controls System The layout of the instrumentation and controls system i1s continuing according to the general scheme outlined in the last report.? The lLayout will permit all routine operations to be performed in the main control- roonm area. An effort is being made to centralize instrumentation, and Tfield panels will be used only where the nature of the operation dictates that controls and instruments be located in the field. The arrangement of the main control area was revised as shown in Fig. 1.7. The data- room area is larger by 225 square feet, allowing more space for the data- handling system. The main control and auxiliary control area were shifted to the south. The instrument shop was moved from the location shown pre- viously® to the southeast corner of the building at the same (852 ft) level. A second major control area is the transmitter room, which 1is located on the 84O-ft level adjacent to the reactor. Instruments for the leak-detector system, weligh system, and the gas-purge systems for the fuel and coolant circuits will be mounted on auxiliary panels in this area. Solenoid valves through which air is supplied to salt-system gas~control valves are to be located in this area, along with ampliflers and power supplies for the process-~variable transmitters. Two instrument panels are located in the service room at the north end. of the service tunnel. These panels house instrumentation and con- trols equipment associated with the lubricating-oil systems for the pumps that circulate the fuel and coolant. Both pumps are located in the serv- ice tunnel. Radiation-monitoring instruments serving the service tunnel will also be mounted on these panels. One more instrument panel will be located in the water room near the radiator blowers at the southwest corner of Building 750%. Instru- ments connected to the cooling-water system are to be located on these panels. UNCL ASSIFIED ORNL-LR-UWG 68577 OPERATIONAL AREA PLAN VIEW ELEVATION: 852 ft Fig. 1.7. Operctional Area of Main Contro! Room and Auxiliary Control Room. f 2 3 4 5 6 7 (8 9 - I { R Y — 21 ft — o 24t e e 24§t AT f1 3N, e {7 ft Gin - AT £ B ! ' ! (T e b e I *‘f'f”ii‘?—lr’)y“ c = — - — -~ ‘ ; "~ 1 ; EemE e Em T s *‘r'frfffm‘ | | | T | iINSTRUMENT SHOP = i | | AND STORAGE = o }L ‘1 ~ /‘/’ ] 1 : . » 1 s = — — | ‘ | AREA: 441 fi 1 = | | | \ i “GLASS WALL L ‘ | | ; | | r Y - X { I 2 RMOCOUPLE | :r ’ I ! I I - CA. B0 12 THERMOCOUPLE | : ; AREA: 1160 ft Chemer | — ‘ ‘ DATA AND SUPERVISOR ROOM |- ~ ‘ : ; : = - K b e e e | : \‘ i S AT IXILIARY CONTROL ROOM i \ " ; 2 AREA: 475 #12 MAIN CONTROL ROOM *\/\ NALCX'E : PCOE OL ROO 1\ . | : AN UCLEAR PANELS | |, - j | N \/ ,,,,,, e 0 e e - | Cppry i IS LT ot L RELAY CABINET--~ PRl = arel - 1 bk i CONTROL POWER DISTRIBUTION CENTERS W - ) = FUTURE PANELS 91 17 1.11.5 Control Panels The main-control-board layout was approved after several revisions. A 1/h-scale model of this board as revised is shown in Fig. 1.8. The design of the control console was issued for comment and is shown in Fig. 1.9. Detail drawings of the control-circult relay and the ther- mocouple cabinets located in the auxiliary control area were also approved., Detailed fabrication drawings showing instrument layouts, panel-cutout detalls, and wiring and piping diagrams are almost 60% complete, These drawings will not be completed until exact dimensions of purchased equip- ment are known. SUNMCLASSIFIED O PHOTO 37959 Fig. 1.8. One-Fourth-Scale Mode! of MSRE Control Panel, UNGLASSIFIED ORNL~ LR~ DWG 68578 Fig. 1.9. Conceptual Design of MSRE Console. 18 1.11.6 Data Handling A study of M3RE data-handling requirementss was completed and issued Tor comment. The study indicated that the MSRE could advanta- geously use a digital data-collecting and -handling system to process the data from the reactor. The data system would be used in conjunction with the conventional instrument and control system. The study report was reviewed by MSR project personnel, and a decision was made toO procurc a digital data system with the capability of performing on-line compu- tations, detecting and alarming off-normal conditions, and logging the process data automatically and on demand. The MSRE instrumentation and control system is being designed Lo ensure that failure of the data system will not compromise the safety or operability of the reactor. All data necessary for the operation and safety of the reactor will be available from the conventional instrumen- tation system. The function of the data-handling facility will be o implement the conventional system by recording data needed for experiment analysis in a more convenient form. Provisions have been made in the design of the instrumentation system to permit expansion of the conven- tional data display and recording system should the need arise. The study of the MSRE data-handling requirements was extended and examined in more detail in order to prepare specifications for the data system. A rough draft of the specifications was completed and distri- buted for review and comment. The reactor operation and experimental analysis groups are reviewing their respective data collection and display requirements. These requirements and the comments on the preliminary specifications will be incorporated in the final specifications. At this time it appears that approximately 2L2 process variables will be transmitted to the data system. 1.11.7 Process and Personnel Radiation Monitoring An initial study of the MSRE radiation instrumentation requirements was made, and a preliminary tabulation of the detectors” was issued for comment. The memorandum was reviewed by project personnel, and the comments were incorporated into the instrument application diagrams. The process monitors are those detectors used to monitor the reactor- system process lines, some reactor components, the two stacks, the reactor cell, the drain-tank cell, and the radiator pit. The personnel monitors are those normally required for personnel protection throughout the inhab- ited areas of the building. These monitors would normally be such ltems 19 as hand and foot counters, door monitors, continuous alr monitors, etc., as required by health physics practices. The requirements for both types of monitoring is again being reviewed in order to determine the specific detector and associated electronic equipment required for each location. The possible detector locations are being studied in order to determine the ambient operating conditions so that adequate shielding, cooling, and cholce of materials and compo- nents can be made. 1.11.8 Procurement Status Specifications were written for the following eguipment during the report period.: . Weigh transmitters. Thermocouple patch panels. Valves for radiocactive-gas service. Pressure and differential -pressure transmitters, weld sealed. Venturi flow element for the coolant-salt system. Mineral-insulated Inconel-sheathed thermocouples. v s O OH The specifications for items 1, 2, 3, M, were approved, and a purchase order was placed for item 2. Purchase requisitions were written for items 1 and 3. Ttems 5 and 6 were issued for comments. Instrument cabinet requirements were determined and orders for cab- inet frames, panels, and associated hardware were issued to the ORNL stores department. (These are ORNL stores catalog items.) Writing of specifications for the following items is continuing, and they will be issued for approval early in the next report period. l. Standard valves. . Standard pressure and differential-pressure transmitter. . Variable-area flowmeters. Pressure regulators. Recorders, indicators, and controllers. Pressure switches. Differential-pressure transmitter, NaK filled for high- temperature (1500°F) service on the venturi flow element. —J O\ WO Equipment that can be economically salvaged from HRT facilities and that meets MSRE specifications will be used. ‘N 20 REFERENCES J. L. Anderson and R. E. Wintenburg, Instrumentation and Controls Division, Ann. Progr. Rept. July 1, 1960, ORNL-2787, pp 145-150. MSRP Progr. Rept. Aug. 51, l96l, ORNL-%215, p 9, Fig. 1.5. O. W. Burke, MSRE Analog Computer Simulation of the System with a Servo Controller, ORNL CF-062-12-50. MSRP Progr. Rept. Aug. 3L, 1961, ORNL-5215, p 21. Ibid., Fig. 1.15, p 22. G. H. Burger, MSRE Data Collecting and Handling Requirements, A Study Report, MSR-61-112. G. H. Burger, MSRE Process and Personnel Radiation Monitoring, MSR-61-108. 21 2. COMPONENT DEVELOPMENT 2.1 FREEZE-FLANGE DEVELOPMENT 2.1.1 MSRE 5-in. Flanges® Two sets of INOR-8 flange forgings and four ring-joint gaskets were purchased for production and testing of prototype MSRE flanges. Some difficulty was encountered in machining the surfaces of the ring grooves to precise dimensions. Successive cuts of less than 0.0l5-in. resulted in rapid tool wear, and frequent sharpening was required. A pair of the flanges is Dbeing installed in the seal~test facility2 for thermal-distortion and gas-leakage measurements. After completion of these tests, the flanges will be installed in the thermal-cycle test loop 3 and the prototype-fuel-pump test facility.? A set of gages was designed to assist in obtaining a consistent over- all thickness of the flange-gasket assembly. This thickness control is necessary because oOf the load-deflection characteristics of the spring clamps used to load the freeze flange. These gages were Tabricated and used in the fabrication and inspection of the prototype flange parts. 2.1.2 Freeze-Flange-Seal Test Facility5 Testing of the 3—1/2~in. freeze flange, using various ring gasketls and a resilient clamp, was discontinued after it was found that insuffi- cient strength in the flange body permitted the ring-groove dimensions to be changed by repeated thermal cycling. Experiments were conducted to determine the temperature distribution and the distortion characteristics of the 6-in. Inconel freeze flange. The flange and clamp were similar to the 5-in. MSRE freeze-flange assembly previously described, except for the following: pipe size (6-in. sched-LO for 5-in. sched-40), material (Inconel for INOR-8), and the absence of spring action in the clamp. Gap width between flange Taces was maintained by a 20—3/8—in. diam ring of 1/16-in. Inconel wire. Two separate heat sources were necessary to obtain even heating of flanges and pipe extensions. Twelve silicon carbide heating elements were connected in series and mounted in the bore, centered at the f{lange faces. Clam-shell resistance heaters were aligned along the exterior of the pipe stubs beginning 2—1/& in. from the flange face, with the insula- tion beveled at a 45° angle away from the flange faces. Various combina- tions of Calrods, clam-shell heaters, and induction heaters were tried in 22 arriving at this combination but they failed either to reach temperature goals or to maintain them for extended periods. As shown in Fig. 2.1, the static temperature distribution was meas~- ured for bore temperatures of 850, 1000, 1100, 1200, 1300 and 1L400°F. The freeze position of the molten salt was 4.8 in. radially outward from the flange bore when the bore was at 1400°F. The average increase in radial position of the freezing temperature was 7/8 in. for each 100KC increase in bore temperature. UNCLASSIFIED ORNL-LR-DWG 68579 | 1400 S - Lo | NO COOLING AIR CIRCULATED 1200 | 1000 850°F | FREEZING POINT 800 600 TEMPERATURE (°F) 400 e 200 e 0 2 4 6 @ o DISTANCE FROM THE CENTER LINE OF PIPE (in.) Fig. 2.1. Temperature Distributions in the Central Plane of the 6-in. Freeze Flange for Yarious Bore Temperatures. At 1L00°F, the results were checked against the Sturm-Krouse® analyt- ical predictions. Experimentally the freeze point was 7.8 in. from the pipe center line, whereas the analytical prediction was 5.7 in. The as- sumptions of continuity between flange and clamp and of radial temperature variation are sources Ior the discrepancy. 23 With a bore temperature of 1300°F and of 1400°F, all power to the assembly was cut off to simulate a salt dump and to note the effect on temperature distribution. The bore cooled to the salt freezing point of 850°F in 27 min and in 34.5 min after shutdown, respectively. When the flanges were at eclevated temperatures, there was a dimen- sional distortion of the plane faces of the flanges, so that they tended to assume a more conical shape. This is illustrated in Fig. 2.2, which shows the distortion when the bore temperature was 1300°F. The maximum distortion varied from 0.024 in. at a 1000°F bore temperature to 0.063 in. at 1400°F; this is relative to an initial spacing of 0.062 in. between the flange faces. No permanent distortion was noticed, although the flanges were cycled 40 times from 200°F to 1300°F. UNCLASSIFIED ORNL -~-LR-DWG 68580 MALE SECTION FEMALE SECTION . ‘ | ‘ RING POSITION i O NO AR THROUGH THE CLAMPS READING ® 15 cfm THROUGH EACH CLAMP HALF i £ RADIAL DISTANCE FROM PIPE CENTER LINE {in.) - i Q g1 o ] 5 \ B \ Q ® ./\'-~ »»»»» _____ —AL-850°F: FREEZE POINT 1 \ WITHOUT AIR ® e "."’850°F: FREEZE POINT \ WITH AIR ® 5 120 80 40 0 40 80 120 GAP (thousandths of an inch) Fig. 2.2. Distortion of 6-in, Freeze Flange Operating at a Bore Temperature of 1300°F, The transient distortion characteristics were checked during a shut- dovn from 1300°F. Fifteen minutes after power shutdown, the maximum flange distortion had decreased from 0.061 in. to 0.034 in., while the position of the freezing-point temperature moved in 1 in. The signifi- cance of the rapid decrease in distortion, compared with the slow change in salt freeze position, is that the amount of salt lodged between the flange faces is reduced during cooling, and the tendency for permanent distortion of the flange gap is lessened. 24 Lastly, the effects of cooling air were noted by passing 15 cfm of instrument air through each clamp half (bore temperature of 1300°F). (See Figs. 2.2 and 2.3.) Whereas the distance from the pipe center line to the position of the freezing point temperature decreased from 6.5 in. to 5.8 in., the distortion increased from a maximum of 60 mils (without air) to 107 mils (with air). It was concluded that the temperature gradient was the dominant factor leading to the distortion of the flange. A development program similar to the one just described will be con- ducted on the INOR-& 5-in. freeze flange. Also, the sealing properties ° of two nickel ring gaskets (one oval and the other octagonal) will be determined when used with the spring clamp. UNCL.ASSIFIED ORNL-LR-DWG 88581 1400 ‘ \ BOGRE TEMPERATURE 1200 +— Y§ - e — AIR 1000 | — — FREEZE POINT, 850°F\\\\k - b LN O BOO | e e e e e e e e fen L o 2 | g r {5 cfm BLOWING THROUGH / o S 600 | EACH CLAMP HALF ”fif' < L = ® \ @ 400 } : : — 200 {- --—- | | o | | 0 2 a 6 8 10 DISTANCE FROM CENTER LINE OF PIPE (in.) Fig. 2.3. Cooling Effects on 6-in. Freeze Flange. 2.2 CONTROL-ROD DEVELOPMENT A simple chain-driven control-rod mechanism is being developed for use in the MSRE. Figure 2.4 shows the system which is being built for testing. The weight of the flexible metal hose and the poison elements 25 UNCLASSIFIED ORNL~-{ R—-DWG 68582 REVERSIBLE-DRIVE MOTOR MECHANICAL CLUTCH GEAR ARM ... — POSITION INDICATOR SYNCRO TRANSMITTER FIXED DRIVE SUPPORT AND 3-in. CONTAINMENT TUBE o= £ v pR w“N \ 6! COOLANT EXHAUST -~ 2-in. CONTAINMENT THIMBLE Fig. 2.4, MSRE Control-Rod-and-Drive Assembly. combined is 6 to 8 1b. The drive unit is overdesgigned, with the capabil- ity of exerting a 20-1b downward thrust and a 25-1b upward pull. The total travel of the poison elements is 66 in., at a rate of 1 in./sec. 26 The rod position is indicated by a calibrated digital voltmeter con- nected to a synchro-operated linear resistor coupled to the chain-drive sprocket shaft. A 12-v dc supply is furnished to the resistor. The rod position is read out as a voltage, Tor convenience in data logging. For maintenance, it is planned that the entire unit will be replaced and direct maintenance performed in a "hot" shop. The entire control-rod assembly can be removed by unbolting the cooling-gas container and drive- unit base from the fixed support structure. Manually operated electrical and air disconnects will permit withdrawal of the unit, leaving the thimble and chain container in position. Reactor cell air will be pumped to the gas container in order to cool the drive motor and the control rod. Part of this stream is diverted into the flexible-metal drive hose by means of a loose-fitting tube mounted concentrically in the hose. All the gas leaves through a common exhaust at the top of the containment thimble. 2.3 HEATER TESTS 2.3.1 Pipe Heaters A full-scale section of removable heaters for 5-in. pipe was built and tested. The test section is shown in Fig. 2.5. Kach heater unit is 1 BT " % : o JNCLASSIFIED PHOTO 37519 Fig. 2.5. Pipe-Heater Test. 27 28 in. long and contains six flat-plate 700-w heating elements, which can be seen in Fig. 2.6. A 6-in. yoke piece which is lapped to fit the heater-box ends 1s inserted between heater sections to make a snug closure. EFach heater unit can be replaced by first withdrawing the yoke at each end of the heater, and then simply lifting the box. The heater base is shown in Fig. 2.7. 1t contains no heaters and is a permanent part of the support structure. _UNCLASSIFIED PHOTO 38320 Fig. 2.6. Pipe-Heater Section. 1600 1400 {200 1000 i L x 2 £ 800 he w o > |i¥] ’_ 600 400 200 0 Fig. 2.8. Temperature Distribution in 5-in. Pipe Heater and Insulation. 28 Fig. 2.7. Pipe Shown Mounted on Heater Base. UNCLASSIFIED ORNL-LR-DWG 68583 \ T f _PIPE-WALL AVERAGE TEMPERATURE P /;/REVERSE SIDE OF HEATER ¢ ; ! | : POWER INPUT (watts per linear foot) i o | o 52 ? ® 437 ........... A 345 | A 270 i N a4 UNION ASBESTOS & RUBBER CO. (1200°F RATING) ® N g UNARCOBOARD | \ | INSULATION ‘ OUTSIDE SKIN . TEMPERATURE i | 0.005-in. ALUMINUM FOIL\\JV } SUPERTEMP INSULATION ‘ l - (1900°F RATING) fl_‘ L N\e i (EAGLE - PICHER CO.) | o L L ‘ ‘ l 0 1 2 3 4 DISTANCE FROM PIPE WALL (in.) UNCLASSIFIED PHOTO 37975 29 Tests indicate a maximum 10°F variation in temperature around the pipe surface, due to the unheated base section. Figure 2.8 is a plot of temperature distribution through the insulation at various power inputs. There is a heat loss of about 450 w per linear foot of pipe at operating temperature. 2.3+2 Reactor Vessel Heaters”’ The reactor vessel heaters have operated a total of 3750 hr. The heater-surface was held at 14250F for 700 hr to maintain 1325°F on the heat~-sink face. The system was then opened for visual inspection, and the general condition was the same as before. However, the upper guide- tube rack was inadequately supported to carry the load of the guide tubes at the higher temperature, and it had buckled. The buckling had not affected the heater operation but had caused misalignment between the guide tubes, the heater pins, and the thermal-shield penetrations. This condition could create serious difficulties for remote malntenance and possibly cause burnout of the heaters. After the rack is realigned, addi- tional supports will be installed and the test continued. The 30-gage Inconel reflector had buckled to some extent in spite of precautions taken to avoid it. Hanging the reflector material and banding it in place will be tried in order to reduce this distortion. 2.4 DRAIN-TANK COOLERS The prototype cooling bayonets® for removing fuel afterheat in the MSRE drain tanks have operated for a total of 3860 hr. Fach l~l/2~in. cooling thimble has the capability of removing 6.3 kw when inserted into 1300°F molten salt to a 60-in. depth. The bayonets have been thermally shock tested through 40 cycles, without apparent damage to the cooling tubes except for minor warping of the l-in. Inconel boiler pipe. Thermal shocking was accomplished by drying the steam system out completely, allowing the bayonet temperature to approach the salt temperature, and then adding water to the steam drum. The thermocouples for measuring pipe temperatures were shielded in stainless steel tubes and strapped to the pipe surfaces. Attempts to use bare-wire thermocouples arc-welded to the pipe were not successful. The test system 1s being modified to permit automatic thermal shock- ing as described above for life testing of the bayonets. 2.5 SAMPLER-ENRICHER SYSTEM 2.5.1 General Concept The present concept of the MSRE sampler-enricher system is shown in Fig. 2.9. The prinecipal changes made to the system previously reported® are: 1. A buffered, double-sealing gate valve replaced the solder freeze valve as the primary containmment valve during sampler maintenance. 30 TRANSPORT CASK UNCLASSIFIED ORNL~LR— DWG 63318R2 AREA 4 CELL ” SAMPLE TRANSPORT CONTAINER ILLUMINATOR / REMOVAL VALVE VIEWING —EXHAUST SYSTEM s o) Lol by R SR ZRIERS : m\g PR CABLE DRIVE RS X TR " MECHANISM PCXOKS SO, N ":’s‘z’s’ ‘KHELIUM ;sts\’s‘ PURGE T0 b VACUUM LI PUMP MANIPULATOR . 70 URANIUM VACUUM SHIELDING P PUMP OPERATIONAL j r“\ VALVE R S LEAD - SHIELDING | AREA 2B N TIPSR At AREA 2A -OUTER STEEL SHELL DISCONNECT P //// FLANGES ®- EXPANSION JOINT CONTAINMENT'J//// LATCH STOP PUMP BOWL VESSEL SHELL Fig. 2.9. MSRE Sampler-Enricher System. 31 2. A flanged seal isolates the upper manipulator area (3) from the area of the valves (2B) in place of a sliding seal. 3. The seal between the contaimment vessel (area 2A) and area 2B was flanged. k. The periscope was changed to a mirror system external to the outer compartmwent, with a quartz window for viewing. The maintenance valve was changed because of the unreliability of the solder freeze valve tested earlier and the complications produced by the gas controls necessary to make a double solder seal.® A double-sealing gate valve similar to the operational valve (sec 2.5.2) will be used. The other changes were made to simplify maintenance procedures and to increase general rellability. 2.5.2 Operational Valve A bellows-sealed, 2-in., Crane double-sealing gate valve considered for use as the operational and maintenance valves was received and tested. As received, the Stellite-faced gate did not seat properly against the Stellite-faced seats and had to be lapped to fit. After the valve was reworked, 1t was opened and closed by hand 100 times. A torque wrench wag used to apply a reproducible torque of 7) ft-l1b on closing, considered to be an acceptable closing force. Periodic checks of buffer-gas leakage to atmosphere through the seals were made. With a buffer pressure of L0 psig of helium, the total leakage was 0.3 cc/min. After 100 cycles the leak rate did not appear to be increasing. A motor operator was obtained for the valve, and leak rates will be determined for motorized operation. 2.5.3 Removal Seal for Sample Container A buffered, double-seal is being tested for use as the sample- container removal seal. The seal, located belween the sample-transport- container removal valve and the transport cask, prevents oxygen contami- nation of the outer compartment while the capsule is being inserted or removed from the outer compartment. The seal consistsg of two rubber O-rings, lubricated with a light film of silicone grease and buffered with 15-psig helium. For the seal test, the sample transport container or rod is moved through the seal by a hydraulic cylinder; it pauses for 1 min and is then withdrawn from the seal. Total leakage of buffer gas through the O-rings is measured periodically while the rod is at rest. After 1400 cycles, total leakage was o | | } wT / £ | ! - 3 i | | Ll | ‘ > 12 b— S— ] -~/ - o ~ | | / / - : 3 L/“/ | 50% DESIGNED FLOW RATE (612.5 gpm) / // L { ( [ 8 : ‘ } i Ly O TTe— v '25% DESIGNED FLOW RATE (305 gpm) | }N@\O / - ! ! G : ‘ s | . i O'N"B"'“ T .| P ‘ ‘ ! | (O *"\fll// | 1 . Lo ‘ 0 45 90 135 180 225 270 315 360 8, ANGLE FROM INLET TANGENT (deg) Fig. 2.10. Center-l.ine Velocity Distribution in Volute of MSRE Full-Scale Core Model. 33 bince water was used for this test, the MSRE Reynolds number was not reproduced exactly. However, the Reynolds numbers in the volute are well into the turbulent range, and since the various coefficients involved are weak functions of the Reynolds number at these high values, the error in velocity distribution is expected to be small. The velocities reported here are about 10% higher than those predicted from runs made in the l/5~scale model.t This difference is attributed to inexact scaling of the model. The uniform decrease of velocily in the volute with tangential position indicates the desired uniform supply of fluid to the vessel. 2.7 HELIUM PURIFICATION Work on the construction of a full-scale oxygen~removal unit is about 50% complete. The design, illustrated on Fig. 2.11, provides for a titanium getter tube encased in a cylindrical resistance heater which is in turn surrounded by an outer pressure vessel. ‘The outer wall is pro- tected from excessive temperatures by an annular layer of high-temperature insulation. The pressure vessel is L4-in. sched-40 pipe and is about 30 in. long. The unit will operate at 250 psig, lZOOOF, and 10 liters/min of helium flow, that is, at design conditions for the reactor gas system. UNCLASSIFIED ORNL~ LR—-DWG 68585 THERMOCOUPLE (1) GETTER TUBE (2) HEATER, 1000 w (3) HIGH-TEMPERATURE INSULATION (3) PIPE, 4-in. SCHED-40 SS (5) FLANGES, 4-in. 1500-1b SS, g in. WELDING NECK, RING JOINT @ r ®) INSULATION 2tin. @— (7) REFLECTOR THERMOCOUPLE -1 1JI1 [} |— THERMOCOUPLE THERMOCOUPLE — | |~ THERMOCOUPLE ? N0 N @0\ @}W“ . I @»-»-— LTHERMOCOUPLE THERMOCOUPLE Fig. 2.11. Oxygen-Removal Unit, MSRE Cover-Gas System. 34 An electrolytic~-type trace-oxygen analyzer was purchased for use in monitoring helium samplcs from the oxygen-removal unit. Tests with helium of known oxygen concentration showed the analyzer to work satisfactorily in the range of O to 10 ppm of oxygen, provided that fluctuations in sample gas temperature are small. 2.6 MSRE ENGINEERING TEST LOOP (ETL) The first run of the ETL (Fig. 2.12) was intended to evaluate the effectiveness of the [lushing operation by following the oxide content of the salt during extended operation.12 This objective was compromised by the discovery, after 3150 hr of operation, that some zirconium fluoride UNCLASSIFIED ORNL -LR-DWG 54492A ~——SAMPLER-ENRICHER MOCK UP _—SAMPLER-ENRICHER TRANSFER TUBE —DANA PUMP 7 Il _.. - - GRAPHITE HANDLING DRY BOX MANUAL SAMPLER GRAPHITE ACCESS ( FREEZE SEAL) —_ — \ / I BOS S =y U = ,/// ; FUEL TANK ;LVfé?““\FLUSH TANK MANUAL SAMPLER FREEZE VALVES Fig. 2.12. MSRE Engineering Test l.oop, The original loop did not contain the sampler-enricher mockup, the graphite container, and the dry box. had inadvertently been included in the salt mix, and ZrOs had precipitated. The balance of the run was devoted to an investigation of various means of removing ZrOs from the system and the accomplishment of the removal. 35 2.8.1 Loop Operations Operation with the ETL included (1) tests to verify the cause of the initial slow drain,*® (2) attempts to dissolve precipitated ZrOs by tem- perature cycling of the salt and by increasing the ZrF, content, and (3) removal of ZrOs manually from the pump bowl and chemically (with HF) from the drain tank. After the treatment of the salt in the drain tank with HF, the loop was shut down for major alterations, including the installation of the INOR-8 graphite container. A summary of the entire ETL operation is given in Table 2.1. The elapsed time begins with the startup of the loop on April 20, 1961. 2.8.2 Operation of Freeze Valve In the previous period it was reported that the initial draining operation required an excessive amount of time.t3 Tests were performed to separate two possible causes. The first test ran for a long period, during which drain-line tem- peratures were kKept at or below previous settings to see whether the line would again plug as the result of normal operation. The loop operated continuously for 560 hr, after which the loop was drained in a normal time with no difficulty. The second test duplicated the BeO addition made previously. An addition of 30 g of BeO in pellet form was made to the pump bowl as before, requiring over 200 hr. At 3150 hr, or 320 hr after the BeO addition, dif- ficulty was again experienced in draining the loop. Extra heat applied to the valve did not help drain the loop. However, when extra heat was applied to the l/Z—in. drain line to raise its temperature from the range 920 to 1130°F to the range 980 to 1190°F, the loop drained. (The freezing point of the salt was 850°F.) From this and other evidence, it was concluded that the difficulty was due directly to the BeO addition. Portions of the pellets (original size, O.4 in. diam x 0.5 in. long) left the capsule, entered the pump suction and settled in the static salt of the l/2~in. drain line, there forming a sludge due to segregation and reaction with the Zrf, in the system. Examination®** of the sludge formed directly beneath the addition port revealed the presence of Zr0s, 2Li¥-*Befs, and 2ZLiF-Zrt,. The 2L1iF-Bels phase melts at 850, and the 2LiF*Zrf, phase melts at 1105°F. The presence of the higher-melting 2Z2LiF-ZrF, phase, along with the ZrO0s, formed enough of a plug so that the loop would not drain. By raising the temperature of the drain line 6OOF, the sludge became sufficiently less viscous to allow the loop to drain. In total, the loop was drained and refilled Th times, without difficulty except for the two above-mentioned occasions. Table 2.1. 36 Operating History of the ETL Time (hr) Remarks 0 Beginning of the operation, April 20, 1961 940-1140 First addition of BeO pellets 1280 Loop drained to begin freeze valve tests. Excessive time required to open valve® 1280-1690 Operational tests performed on freeze valves 1690-2250 2250-2830 2830-3050 3159 3150 3200~3500 3820 4180 4180-4380 4380-4680 l'700-5200 5200-5860 Operation for 560 hr to attempt duplication of the previous slow drain Additional operational tests performed on freeze valve becond addition of BeO pellets Salt sample ETL-69 revealed prescnce of Zr0s crystals Drain time again excessive after BeO addition Temperature cycling of salt Trom 1050°F in drain tank to 1200°F in the loop, for oxide- transfer attempt Additions of “Zri, were begun in order to in- crease oxide solubility Third addition of Zrl', was made, bringing salt composition to LiF-BeFs-ZrF,, 062-3L-4 mole ¢ Circulation for 200 hr at temperaturcs up Lo 1200°F Loop drained, cooled, and pump removed for examination salt treated in drain tank with mixture of Ho + HF Drain tank kept hot for additional salt samples; final sample (E1L-133) extracted g5ee ref 13. and Loop operation terminated for alterations 37 2.8.3 Oxide Removal Sludge had been seen and sampled through the lwl/Zuin. sampler connec- tion to the pump-bowl lid while the system was at temperature (llOOOF). Two methods of redissolving the sludge were tried, and visual inspection was used as the method of estimating their effectiveness. First a "cold trap" type of removal was attempted by circulating salt over the sludge at 1200°F, draining and cooling in the drain tank to 1050°F. This cycle was repeated three times during the interval 3200 to 3500 hr (see Table 2.1). Second, ZrF, was added in the form of ZrF,-LiF (49-51 mole %) on three occasions to increase the expected oxide solubility. According to the visual observations, the sludge deposit was not affected appreciably by either of these attempts. The loop was drained into the flush tank for cooldown and manual removal of the sludge from the pump. The results of the oxide analysis of samples taken between the time that ZrF, was discovered (3150 hr) and the pump was removed (4400 hr) averaged 563 ppm oxygen, as shown in Fig. 2.13. 2.8.4 HF Treatment Prior to the complete shutdown of the loop for alteration, apparatus was set up and operated by the Reactor Chemistry Division for the treat- ment of the salt in order to remove the precipitated oxide believed to be in the drain tank. The treatment, consisting of passing a mixture of hydrogen and HF through the salt, is described in Sec 6 of this report. The treatment took place over a period of 500 hr and removed the equiva- lent of 1025 ppm of oxygen from the salt. Figure 2.13 shows the results of the oxide analysis of salt taken during the treatment period (4700 to 5200 hr) and afterwards (until 5900 hr), while the drain tank was kept hot. The change in oxide concentration and the amount of oxide removed by treatment indicates that a considerable amount was precipitated in the drain tank. The amount of concomitant corrosion in the Inconel drain tank is in- dicated in Fig. 2.14 by the increase in the chromium content of the samples taken between 4700 and 5200 hr. This increase is equivalent to the removal of all the chromium to a depth of 0.0005 in. over the entire of drain-tank surface; although the partial depletion of chromium extends deeper, the attack is not excessive. Examination of five Inconel dip tubes used for bubbling the HI and Hp through the salt revealed only very light to moderate surface roughening and pitting. 2.8.5 ETL Graphite Facility The &-in.-diam graphite container,® with a longitudinal frozen-salt seal access, wags made of INOR-8. The container was installed into the loop as indicated schematically in Fig. 2.12. The upper portion of the container is shown in Fig. 2.15. At the same time, most of the loop piping was changed from Inconel to INOR-O. 1300 1200 1100 1000 900 800 700 OXIDE {ppm) 600 500 400 300 200 :- 100 Fig. 2.13. Resuits of Oxide Analysis — Engineering Test Loop, Oxide content of salt vs number of hours CIRCULATED SAMPLES == -+ DRAIN-TANK SAMPLES i HF + H, TREATMENT ZrF, ADDITIONS UNCLASSIFIED ORNL-LR-DWG 68586 TERMINATED 12-20-64 — T T T T — T T Y Y 1 T e wivy dyn 99 ‘ 1 D i i 1 - - — ~ , - . - — ] © O OPEN (UNFILTERED) SAMPLE I O B ® FILTERED SAMPLE [ 77 | — S i &5 . . O i 1 O o | | ¥ O ! L e i _ | 5 AVERAGE OF 37 SAMPLES o ® 7 ——— FROM 2150 TO 4400 hr o ] e e 2 : - - T O O © o o ; o o8 © o - D - 6\ o ® — O o . ! O 3 ® ; ; o o 9 { AVERAGE @ | T © | e e T e oc% , e e | S, - : [ i O ® i 1 i - SEPTEMBER : OCTOBER NOVEMBER } o ! DECEMBER ! : i ! ; i , i | i | | | i | | | 1 30 32 34 36 38 50 42 44 46 48 50 52 54 56 58 (x10%) TiME {(hr) ZrF4 additions and HF treatment (four months), since beginning of test, including 8¢ 39 UNCLASSIFIED ORNL~LR~DWG 68587 600 S A® eee; - e wp | 500 :”. - Tis® . E ole® s o .5"‘.‘&‘ 2 * ® . > ® ~ 400 ® e ® 2 : I E ° 0.,&1. z ¢ . HF TREATMENT S 300 o el , IN DRAIN TANK - o] © a op oS z s = g Q200 5 S ® e 100 0 0 1000 2000 3000 4000 5000 6000 CHROMIUM CONTENT (ppm in flush salt) vs HOURS SINCE BEGINNING OF ETL OPERATION Fig. 2.14. Results of Chromium Analysis ~ Engineering Test Loop. Chromium content (ppmin flush salt) vs hours since beginning of test. £ UNCL ASSIFLED i G‘:\’AP;HIYTE ' s . ,‘ . ::’ S S | : : . 7 PHOTO 37826 (65 InLONG) -~ ‘g SR e ; N 1 « : S GASKET MAKEUP . i ; o 5 TO DRYBOX ~OVAL RING GASKET | AND FLANGES < SALT OUTLE o EXTERNAL "~ COOLING COILS GRAPHITE HOLDDOWN GRAPHITE SUPPORT PLATE INTERNAL COOLING BAFFLE Fig. 2.15. Engineering-Test-Loop Graphite Container, Showing Freeze Joint, 40 The dry box has been fabricated and has recelved preliminary testing. Figure 2.16 shows the loop and graphite container during construction, and the dry box in the unmounted position. After operution of the loop, the dry box would be mounted over the container for access to the graphite through the flange. UNCL ASSIFIED PHOTO 37919 i DANA PUMP AND SUPPORT GRAPHITE CONTAINER Fig. 2.16. Engineering-Test-l.oop Graphite Facility in Construction. 41 2.9 MBRE MAINTENANCE DEVELOPMENT The objective of the maintenance development program is to ensure the capability of replacing any item of equipment which fails after the reactor has become radiocactive. Reacltor-cell eqguipment problems are being studied in the maintenance mockup.'® The product of this program will be (1) an inventory of tools or tool designs which have been appropriately tested and demonstrated and (2) procedures and techniques for performing the MSRE maintenance within the capabilities of the available handling equip- ment and the known operating requirements. The following sections (2.9.1 to 2.9.6) summarize progress made on specific operations. 2.9.1 Placement and Removal of Freeze-Flange Clamp The first concept for putting clamps on freeze flanges called for using 1—1/2~in. bolts and trunnion nuts. This method was abandoned because of repeated severe thread galling. A second method, utilizing hydraulic cylinders (see Fig. 2.17), vas successful. The clamps were driven on and off the flanges many times, with no trouble. A further advantage is that this method does not require the closely held parallelism between the upper and lower clamps that the screw-thread system does. The equipment required to operate the clamps remotely is being designed and fabricated. 2.9.2 Flange Alignment and Pipe-Jacking Tools These are used to provide force to move the flanges, as shown in Fig. 2.18. The jack moves axially to open or make up the joint and to hold the joint closed while the flange clamps are driven on, and the align- ment tool provides force vertically and laterally, to overcome distortion and preloading of the spring supports. Both tools were tested and were able to move the pipe flanges of the simulated MSRE line in the prescribed directions and amounts. 2.9.3 Gasket-Replacement Procedures The time interval in the maintenance procedure while the flanges are open is critical because of the possibility of damage to the gasket and groove, because of the danger of spreading particulate contamination from the reactor internals to the cell atmosphere, and because of the danger of contaminating the reactor system with materials in the cell atmosphere. It is believed that four tools will be required for gasket replacement: (1) some form of dust catcher must be utilized until the openings can be covered, (2) a tool capable of freeing a stuck ring gasket must be used to remove the existing gasket, (3) two separate covers must be placed to seal the two pipe openings, and (h) a handling tool is needed to position and lock the new gasket in place. The flange covers were received and are being improved with respect to handling, space requirements, and sealing ability. The stuck-ring tool is being fabricated. The remainder of the tools will be designed in the near future. 42 UNCLASSIFIED PHOTO 37988 R C R HYDRAUL | OPERATO | ic Clamp Operators. Hydraul 7. 2 ig. F 43 UNCL ASSIFIED PHOTO 37993 FLANGE ~ALIGNING. | TOOL PIPE JACK o Fig. 2.18. Flange Alignment and Pipe-Jacking Tool. 2.9.4 Miscellaneous Disconnects Several components are directly accessible in the plan view, and their removal constitutes unit remote operations that do not require co- ordination with other operations. These are: power, thermocouple and electronic disconnects, and pipe line and reactor heaters. Remote han- dling of one of these components generally consists in engaging a hook of some sort, imparting either a lifting or twisting motion to free the unit and then transporting it to the storage position. Several of these hook tools have been built and tested. A test of an electric-power disconnect indicated the need for refining the viewing and handling techniques, with special emphasis on protecting the insulated leads. 2.9.5 Remote Viewing In some maintenance work, remote viewing will be required. A 1.625-in.-dianm wide-angle periscope was borrowed for testing disconnect operations. This device was superior to an 0.875-in.-diam periscope used earlier, especially in its ability to "see" in dimly 1lit areas. Prelimi- nary quotations have been received on a similar modular periscope capable of working in the MSRE radiation environment. 44 2.9.6 Component Removal The only work done on this phase has been to practice the direct- vision handling of the circulating pump, using the special 1ift sling and the overhead crane (see Fig. 2.19). All auxiliary piping will be added to the pump mockup as final details become known, and a full replacement procedure will be tested. #® UNCL ASSIFIED PHOTO 37995 Fig. 2.19. Simulated MSRE Pump with Lifting Sling. 45 2.10 BRAZED-JOINT DEVELOPMENT7 2.10.1 Joint Design The tapered braze Jjoint for l-l/z-in. sched-40 pipe was further modi- fled (Fig. 2.20) to use 0.005-in.-thick sheet-braze preform that places braze metal throughout the joint. Axial pressure is maintained on the Joint members during brazing so that when the braze metal melts, the joint thickness is reduced to 0.001l to 0.002 in. UNCLASSIFIED PHOTO 37965 COMPLETED LAPERED JOINT 005 INCH THICK SHEET BRAZE METAL WAS PREPLACED IN THE JOINT. AXIAL PRESSURE WAS MAINTAINED ON THE ASSEMELY. WHILE IT WAS HEATED TO 1900°F, HELD for ONE MINUTE, AND PERMITTED TO COOL. Fig. 2.20. Brazed Pipe Joint, Showing the Sectioned, Completed Joint (Above) and Starting Materials (Below). 2.10.2 Braze-Joint Testing Ultrasonic and metallographic inspection of completed joints made with prototype equipment on the bench indicated 81 to 86% bonding. One completed l-l/2~in. Jjoint was subjected to tensile loading, with the following results: at 1300°F it held a 28,000-1b load, held a 10,000-1b load at 1500°F for 5 min, an 8000-1b load at 1600°F for 5 min, and failed at 1700°F under an 8000-1b load. Visual inspection of the separated joint indicated complete or nearly complete wetting of the mating INOR-8 sur- faces by the braze metal. 46 A representative braze joint was held at 1250°F and exposed to reactor salt for a total of 74 hr of intermittent exposure, simulating drain-line conditions. At the completion of the test, the joint will be examined metallographically. 2.10.3 Remote Fabrication of Braze Joint Tools for the various mechanical operations required to fabricate a braze joint remotely were received and satisfactorily operated. The func- tions of the tools and the sequence of operations are as follows: 1. The traveling vise is lowered over a continuous run of pipe and bolted to the base plate (Fig. 2.21). The vise jaws are then closed on the pipe. 2. The pipe cutter is lowered over the pipe onto its sup- port on the traveling vise (Fig. 2.22). 3. The pipe-cutter drive is operated, carrying the cutter around the pipe. The knife is automatically fed by the star wheel and pin device (Fig. 2.23) about 0.009 in. each revolution. After the cut is completed, the cutter is opened and removed. i, After the component is removed, the tapering tool is placed on the base plate. The traveling vise feeds the pipe into the rotating shaped cutter (Fig. 2.24), which machines a 6° included-angle taper to a depth of 1 in. When the cut is finished, the pipe is withdrawn and the tapering tool removed. 5. One portion of the furnace can now be placed over the ta- pered male-pipe stub. 6. The new component with the female-joint half, contalning the braze preform, and the remainder of the furnace are installed. 7. The fixed vise is installed over the female-joint half to hold and align it (Fig. 2.25). 8. 'The sliding vise then inserts the male into the female pipe joints, the furnace is assembled over the joint, and the thermocouple leads are connected to the record- ing instruments. After an adequate inert-gas purge, the joint is inductively heated to 1850°F while axial force is maintained; it is held at that temperature for a minute and then permitted to cool. R R S UNCLASSIFIED PHOTO 37963 % P 47 i 3 ipe. inuous P ing Vise Being Lowered over Cont 21. Slid @ Fig. 2 48 o LJ L %21 %) < J O = o PHOTO 37964 R s St ipe, ise over Continuous P ing V ipe Cutter Being Mounted on Slid P Fig. 2.22. 49 UMCL ASSIFIED FHOTO 37971 e REMOTE DRIVE SHAFT Fig. 2.23. Pipe Cutter. Drive shaft rotates cutter around pipe. Star wheel and pin feed the knife with each revolution, 50 UNCL ASSIFIED PHOTO 37968 REMOTE DRIVE SHAFT Fig. 2.24, Pipe-Tapering Machine, UNCLASSIFIED PURGE LINE =~ »- , A t g THERMOCOUPLE PHOTO 37967 FEED SCREW f ’. LEADS ORIVE VISE-JAW | ) } a '''' ! : e V| SE - JAW DRIVE - {NDUCTION - HEATING VISE § o Fig. 2.25. Joint and Furnace Ready for Assembly. 51 2.11 MECHANICAL-JOINT DEVELOPMENT The overflow line to the fuel-pump bowl described in the design section of this report is routed through an area too crowded to permit making a braze joint after component replacement. Therefore, a mechanical jJoint is being developed for this service. Since the Jjoint will be ex- posed to salt for only short periods during reactor-fill operations and during infrequent overflow from the pump, a trapped-gas pocket will be used to keep salt out of the region of the gasket. Initially, the over- flow line will be continuous with joint halves installed, and preparations will be made for cutting at the appropriate points for installation of the mechanical Joint. A prototype will be tested. Z2.1l2 STEAM GENERATOR A steam separator for a bayonet-tube steam generator-superheater was constructed and tested in the separator test chamber previously reported_.l8 The flow areas within the test separator were as follows: "Boiling" annulus — 0.312 in.< Swirling-separator annulus — 0.0638 in. Steam-outlet nozzle — 0.0283 in.=2 Water-return tube — 0.1452 in.?® 2 Swirl vanes l-7/8—in. long and having a 300 angle with the separator axis were used. Alr and water were circulated through the separator at various flow rates and ratios of flow rates. 1In all cases, the water carryover was very high. As an example, with an air flow of 3 ¢fm and a water flow of Os5 gpm, the water carryover was approximately 0.2 gpm. Further work was postponed due to a shortage of manpower. .13 PUMP DEVELOPMENT The design drawings for the MBRE fuel pump were approved, and the thermal analysis of the fuel and coolant pump tanks was completed. Water testing of the coolant pump model was completed. Fabrication of the rotary element for the prototype fuel pump was completed, and fabrication of the pump tank was nearly completed. Design drawings for the lubrica- tion stands and the drive motors were submitted for review. Additional ITNOR-8 castings of impellers and volutes for the fuel and coolant pumps are being made, and dished heads for the pump tanks are being inspected. 52 Z.13.1 MOSRE Fuel Pump PK-P Pump Hot Test This test pum;pl9 was placed back in operation to obtain more testing of the resistance of the lower shaft seal and of the impeller to cavi- tation damage and to obtain additional observations on lubricant inven- tory. The pump circulated LiF-BeFo-ThF,-UF, (65-30-4-1 mole %) at 1225°F, 510 gpm, and 1950 rpm, and has operated for (88 hr. The differential pressure across the lower shaft seal was maintained at 3/M psi, and no measurable oll leakage was noted. Prototype Fuel Pump and Hot-Test Facility Fabrication of the rotary element®® was completed, and dimensional checking, assembly, and bench testing were started. Fabrication of the pump tank is about 80% complete. Modifications were made to the support structure of the test facil-~ itygl to accommodate the flexible mount designed for the fuel pump. Sev- eral other modifications were completed to accommodate testis of various items including (1) the bubble liquid-level device, (2) the freeze flange, (3) a section of pipe heater, (4) operation of the sampler-enricher device in the pump tank, and (5) the comparison between temperature readings for thermocouples installed in a thermowell and attached to the external sur- face of the piping. Thermal Analysis of MSRE Fuel and Coolant Pumps The thermal stress and strain Tatigue ana1y51522 of these two pumps has been completed. The thermal stresses in the fuel pump during reactor operation at 10 Mw which were previously reported have been revised to include the effects of the meridional heat flow along the surface of the pump tank. The maximum principal stresses are shown in Figs. 2.20 and 2.27, and in Figs. 2.28 and 2.29 for the fuel and coolant pumps, respec- tively. Calculations were completed, showing that a constant cooling— air flow of 200 cfm on the upper surface of the fuel pump tank will provide adequate life; therefore, the previously proposed automatic air-flow- control system will not be used. A forced-convection air-cooling systen is not required for the coolant-salt pump. The predicted total usage factor of the fuel and coolant pumps are 0.30 and 0.57, respectively, com- pared with a safe design value of 0.08. . 2 . . . . . A detuiled report®® covering the analysis was written and is being reviewed. MSHR ruel Pump The design of the rotary element®? was completed and approved. Four each of the rotary elements were placed Ior fabrication by off-area fabri- cators. The design of the pump tank was completed and approved, and 53 UNCLASSIFIED ORNL-LR~DWG 64496 30,000 VOLUTE SPHERE Ti 20,000 /- JUNCTION -—/7\ 10 Mw OPERATION 10,000 f-or-rrm ffod 200 acfm COOLING AIR INTERNAL CYLINDER EXTERNAL CYLINDER Iy -10000 —A—— 1 ) \“/Q\\ -------------- O- POWER OPERATION 200 acfm COOLING AIR STRESS (psi) o N — -20,000 OP FLANGE -30,000 -6 -4 -2 0 2 4 6 8 AXIAL POSITION (in.) Fig. 2.26. Principal Thermal Stresses in Cylinder of Fuel Pump. UNCL.ASSIFIED ORNL. ~LR—-DWG 64497 4000 2000 \ O POWER, 200 acfm COOLING AIR CYLINDER T ,_——_—-——‘—"M JUNCTION \4 -2000 / fi \M ~4000 \ / 10 Mw, 200 acfm GOOLING AIR -6000 ~8000 I///, ~10,000 ~12,000 STRESS { psi} 0 | 2 3 4 5 6 MERIDIONAL POSITION (in.) Fig. 2.27. Principal Thermal Stresses in Sphere of Fue! Pump. partial fabrication will precede the delivery of the INOR-3 volute castings and the requisite INOR-8 pipe and tubing. The dished heads for the pump tank were received and are being in=- spected. Three pairs of impeller and volute castings of INOR-8 are being poured by the founder of the coolant salt pump castings. Design drawings of the drive motors were reviewed and returned to the manufacturer for revision. 54 UNCLASSIFIED ORNL~LR-DWG 64498 30,000 VOLUTE T 20,000 - O POWER NO EXTERNAL COOLING 10,000 |- Q o 10 Mw NO EXTERNAL COOLING J w ul O | | o % 10 Mw NO EXTERNAL COOLING 240,000 fboreoreeeer e e e O POWER NO EXTERNAL SPHERE JUNCTION COOLING -20,000 \\ TOP FLANGE -30,000 I -6 -4 -2 0 2 4 6 8 AXIAL POSITION (in.) Fig. 2.28. Principal Thermal Stressesin Cylinder of Coolant Pump. ORNL-LR-DWG 64499 UNCLASSIFIED 60 10 Mw NO EXTERNAL COOLING O-POWER NO EXTERNAL COOLING @ Q O Q Q ; 20 e e i wn & - CYLINDER n JUNCTION 0 | | ! -20 { 0 1 2 3 MERIDIONAL POSITION (in.) 4 Fig. 2.29. Principal Thermal Stresses in Sphere of Coolant Pump. MSRE Iubrication Stands The design of the Jlubrication stands for the MSRE fucl and coolant and valves have been pro- punips has been completed. cured for the two stands. The pumps, filters, 55 A stand was constructed for proof testing the canned-rotor-type pumps to be used in the lubrication stands, and one pump was operated for 1428 nr, circulating a turbine-type oil at 160°F, 70 gpm, and 3500 rpm. Discoloration of the oil was noted, and examination of the pump revealed a varnish~like coating on most of the surface of the rotor, indicating an insufficient flow of cooling oil through the motor cavity. The pump support bearings were modified by adding three axial grooves spaced at 120° to to reduce resistance to the flow of cooling oil through the motor cavity. The flow was increased several fold and the temperature of the external surface of the motor stator was reduced from 270 to 195OF. An endurance test will be made. 2.13.2 MSRE Coolant Pump Coolant-Pump Water Tests The MSRE coolant-pump water tests®> have been completed, and the impeller diameter has been tentatively set at 10.3 in. to provide a Tlow rate of 850 to 920 gpm, assuming the actual head loss in the coolant-salt system is within t10% of the design value of 78 ft. The hydraulic char- acteristics of the pump are shown in Fig. 2.30, and the efficiency of the pump at operating conditions is approximately 78%. UNCLASSIFIED ORNL~LR-DWG 68588 1O HEAD —mh\ 100 /— \ ] T PREDICTED FLOW. RANGE -\\\\\\\ WITH SYSTEM RESISTANCE 90 ~ — AT B850 gpm =78 ft 10% 80 ™ 2 70 \\\\ ) 2 N ¥ 60 N = \\\\ £ - 50 [ ] é / x SHAFT Hp AT SG =1,92‘,,,/f”“’ n 40 320 // (,//f””/ IMPELLER 0.D: 10.27 in. 20 frr CALCULATED FROM HYDRAULIC TEST DATA ~ ] WITH IMPELLER 0.D: 9.90 in. {0 0 0 200 400 600 800 1000 1200 1400 CAPACITY. (gpm) Fig. 2.30. Hydraulic Characteristics of MSRE Coolant Pump. 56 Reactor Coolant Pump The design drawings are being checked prior to review. One satisfac- tory volute and two satisfactory impeller castings of INOR-0O were received. Weld repairs are being made to one impeller and volute to raise the casting quality to Class I level. The other impeller is Class I quality, without repair. An additional volute casting of INOR-8 was poured by the founder, who reports that radiographic and dye-penetrant inspections indicate an acceptable casting. 2.13.3 Advanced Molten-Salt Pumps Pump with One Molten Salt Lubricated Bearing The test of this pump®® was terminated at 12,500 hr of operation during which LiF-Bels-UF, (62-37-1 mole %) was circulated at 1225°F, (5 gpm, and 1200 rpm; 92 start-stop operations were sustained. For the last two start-stop operutions, rubbing could be detected while the pump shoft was rotated by hand; however, no evidence of the rubbing could be found in the recorded trace of the power input to pump-drive motor. Examination of the Journal and bearing surfaces indicated slight rubbing. The bearing assembly is shown in Fig. 2.31. Each of the two sets of gimbals supporting the hearing were satisfactorily operable at the conclusion of the test. During disassembly, difficulty was experi- enced with the removal of the INOR-8 journal sleeve from the Inconel pump shaft. The difficulty resulted from self-welding of the INOR-8 sleeve and the Inconel shaft, and is shown in Fig. 2.32. The bearing and journal will be replaced for further endurance testing. UNCL ASSIFIED PHOTO 37393 TkT’/’l'l",'lk‘l'ri""f‘l"‘"l"}f"f T . 1 2 3 [ INCHES \ Fig. 2.31. Gimbals-Mounted, Molten-Salt-Lubricated Journal Bearing. 57 UNCL ASSIFIED PHOTO 37896 AR R Welding Between Inconel Shaft and INOR-8. 0320 Self- Figo 2 58 2.1% MSRE INSTRUMENT DEVELOPMENT 2.14.1 Thermocouple Attachments Welded thermocouple attachments made with Heliarc welded, TNOR-8 adaptor lugs of different weights and shapes27 are being ecvaluated in a series of tests. Initial tests were made in order to determine the bonding strength and effects, if any, on the calibration. Attachments made with side lugs welded along the edge parallel to the axlis of the sheath®® withstood the greatest pull, bend, and prying force before sepa-~ ration occurred. Seven thermocouples were attached to a section of INOR-G pipe by the methods tested above and then checked for calibration accuracy under static conditions between 900 and 1400°F. The greatest error noted for any couple tested was +6°F, which is within the 3/4% tolerance specificd for the Chromel-Alumel material used in this test. These thermocouples, which are now being soaked at 1200°F, will be rechecked in several weeks. Additional thermocouples are being prepared for testing under simu- lated operating conditions. They will be tested on the Engineering Test Loop, freeze-valve test, and a pipe heater section in the pump-test loop. A test rig was assembled for developmental testing of mechanical attachments for use on the radiator tubes in the MSRE. 'The procedure con- sists of heating a L in. section of 7/8-in. 0D x 0.065-in.-wall INOR-8 tube with a cartridge~type heater which was inserted in a silver-plated copper slug located inside the tube. Reference thermocouples are located between heater and slug, slug and tube wall, and on the outside wall of the tube. The latter is a 30-gage bare-wire thermocouple, spol welded to the tube wall. The results of a preliminary test made with a l/8uin.wOD Inconel-sheathed, MgO-insulated thermocouple are as follows: Outer-Wall Temp.(°F) Air Flow Inner-Wall Temp. (°F) Reference Couple Test Couple 0 1230 1150 1060 80-90 fps 1230 1040 850 As indicated by the results, the inner-wall thermocouple was not measuring true wall temperature. This will be corrected before further tests are conducted. Also, the heated section will be purged with inert gas to prevent cracking of the plating on the copper slug. Test thermo- couples have been made with the junction ground to the sheath's wall, which will be placed next to the heated surface. 2.14.2 Temperature Scanner Development of a thermocouple scanning :system,,g9 commutator, is continuing. using a mercury-jet 59 During the early development of the system, the method of switching created considerable noise. The switch was then being operated in a break-~before-make switching mode. Operation in this manner resulted in the generation of 0.25 v and greater noise pulses, with pulse durations about 1 to 2 pusec. Further examination of the switch output signal re- vealed that the noise pulses always occurred during the break portion of the switch action. It was postulated that the pulses were caused by static charges generated by the mercury jet in the insulated region between the signal pins. The noise pulses did not appear when the jet was in contact with the signal pin because the source impedance (thermo- couple and lead-wire resistance) was about 100 ohms or less. However, when the Jjet moved away from the signal pin, the input impedance became very large, and the nolse pulses appeared in that edge of the output signal. The switch manufacturer verified the cause and existence of the noise pulses. The problem was eliminated by changing the switch operation to make- before-break, resulting in a low input impedance at all times. The output signal from this type of operation is different from the break-before-make action because, during the make period, the mercury jet is in contact with two signal pins simultaneously for about 10% of the outpubt-signal width. During this time, the output signal is the average of two adjacent input signals. However, in the break portion, the jet is again on a single signal pin, and the output signal is equal to the emf of a single thermo- couple. This type of output signal is acceptable as an input to the pro- posed alarm discriminator and does not materially alter the display and identification of signals on the oscilloscope. The development has progressed to the point that a preliminary design has been completed. The system design appears to be workable, and detailed circuit design is in progress. An alarm-detector circuit is being designed and is expected to be completed by April. One complete 100-point scanning system is being built for testing. It appears that at this point in instrument-system design, approxi- mately 400 thermocouples will be required to be scanned by the scanning system. The radiator-temperature monitoring system will require 120 points. These will utilize two scanners of 60 points each, with provisions for detecting and alarming on low radiator-tube temperature. The remaining 280 points will utilize three scanners. One scanner each, with provisions for producing high or low alarms will be used on the fuel system, the coolant system, and the drain tank system. The final arrangement, operation, and integration of the five scan- ners into a system will await testing of the first prototype unit, now being built. It should be completed in April. 60 2.14.3 Single-Point Temperature-ilarm System Investigation of methods of economically obtaining signals which reliably indicate the operating status of freeze flanges and freeze valves is continuing. Results indicate that the Electra Systems Corporation monitoring system, described previously, 2° will be suitable for monitoring freeze- flange temperatures but will not be sultable for the freeze-valve opera- tions because of the lock-in (seal) feature inherent in this system. 1In the freeze-flange monitoring system, only high-temperature-alarm monitor- ing is required, and since all temperatures would normally be below the alarm point, an alarm would occur if any frecze-flange temperature rises above the alarm point. However, in the frecze-valve-monitoring opecra- tions, both high- and low-temperature signals arce required to indicate whether the valve 1s open or closed, and the monitor is required to indi- cate immediately when the operating state of the valve is changed. The lock-in feature would prevent the change of state of the alarm monitor until the monitor is manually or automatically reset. This reset action is considered undesirable as it would produce spurious signals which could interfere with the operation of safely interlocks and which would be annoying to the operator. Other methods of monitoring the freeze valves arc being investigated. One device which appears promising is a magnetic relay, Daystrom Magsense Control Relay, Model A-82, manufactured by Daystrom, Lncorporated, Ia Jolla, California. The unit is a complete solid~state device consisting of a magnetic amplifier with an isolated winding for the input signal and a silicon- controlled rectifier to furnish the output switching action. The relay operates from a 25 to 30 v dc supply. The signal input range is O to 100 pa dec, standard. The input resistance is a nominal 360 ohms. The relay has a response time of 200 msec, standard. The alarm setpoint is adjustable with tl% of' the full rated input range by a poten- tiometer or by external means. The output is analogous to a single-pole double-~throw relay rated at 1 amp at 25 to 30 v de. This can be considered as two outputs, one energized above the set point and one below set point. The repeatibility of the alarm set point is *1 pa or +1% of the input range. 'The hysteresis is less than 2 pa and can be increased or decreased by use of an external resistance. The operating temperature range is 35 to 100°F at full output rating and at stated operating specifications. An Klectra Systems Corporation monitoring system has been placed on order. It will be tested to determine whether it mects {the operational and reliability requirements of the MSRE. One A-82 control relay is on hand and being tested. The results to date indicate that it performs to specifications and would be sujtable for use 1in monitoring the freeze valves. 61 The freeze-valve-monitoring operation will require three relays for each valve. Integration of three A-82 relays into a monitoring system would require additional design to supply indicator lamps, power supply, and a calibration method. However, at this time i1t appears that this type of system offers the greatest reliability and best operating characteris- tics at the least cost. 2.14.4 Pump-Bowl-Level Indicator Development of a continuous-level 31 eclement for use in measurement of molten salt levels in the MSRKRE fuel and coolant salt pump bowls has continued. During the past report period several high-temperature differ- ential transformer designs were investigated, two level-element designs were developed, and fabrication of a level test facility incorporating the two level-element designs was completed. Testing of the prototype level elements 1s underway. The major effort in the program during this report period was devoted to the development of a differential transformer which would operate reliably (without excessive shifts in characteristics) in the temperature range from 850 to 1300°F. Several variations in the design of the transformers were investigated, and three designs were selected for fur- ther testing. These included the nickel-wound lava-insulated transformer previously reported,sl'a transformer similar to the nickel -wound transformer but with Inconel windings, and a guartz-insulated transformer with Hastelloy C windings. Although previous tests of the nickel-~wound transformers 3% were prom- ising, the known and tested durability and low temperature coefficient of resistivity of Inconel, the lower resistivity (compared with Inconel) of Hastelloy. C, and the negligible temperature coefficient of resistivity of Hastelloy C indicated that these materials should be tried. Neither Inconel nor Hastelloy C were as satisTactory as nickel. The Inconel transformer was durable, but the temperature effects on the output signal of this transformer were greater than on the nickel-wound transformer (see Fig. 2.33). The Hastelloy-C transformer also exhibited excessive temperature effects on the output signal. "The wire in this transformer became very brittle and broke during a Tirst inspection made after one heat cycle to 1LOOCF. Two unforseen difficulties were encountered while fabricating and testing the transformers. One was caused by the use of unfired Fiberfrax insulating material, the other by the magnetic characteristics of the tube furnace used in the test. All transformers tested were encased in an Inconel coil form which gave mechanical protection to the coils and insu- lation. Fiberfrax paper was placed between the transformer and the sur- rounding case to absorb mechanical shock. When the nickel transformer assembly was heated, the organic binder in the Fiberfrax insulation evap- orated, condensed on the lava forms and on the ceramic insulating beads on the lead wires, permeated the insulation, and carbonized. The result- ing carbon deposits effectively shorted the transformer; measured resist- ance to ground from either primary or secondary was less than 10 ohms. 62 UNCLASSIFIED ORNL-LR-DWG 68589 50 40 | uw INCONEL | 2 } Q w i T — 5 ! - ’LM._ p © T ? w»sszzr.':i,:::m»m@ f S.[) Z NICKEL | . or ] r<:I: ‘ i z CONSTANT VOLTAGE EXCITATION BOTH TRANSFORMERS ADJUSTED AT 1050°F 1 ,,,,,,,, - - ] 0 CORE MOVED WHILE TRANSFORMER WAS AT 1050°F SO THAT RECORDER INDICATED 25% OF SCALE ’ ‘ \ & o | | | 600 700 800 900 1000 1100 1200 1300 TEMPERATURE (°F) Fig. 2.33. Temperature Characteristics of High-Temperature Differential Transformers Wound with Nickel and Incone!l Wire. Constant voltage excitation; both transformers adjusted to 1050°F; core moved while transformer was at 1050°F so that recorder indicated 25% of scale. The nickel winding was att.cked and, for all practical purposes, destroyed. The trouble was corrected by completely dismantling the transformer, baking the lava forms at lSOOOE and then reassembling, using new insulating beads, nickel wirce, and Fiberfrax paper which had been prefired in air to drive out the organic binder. During initial testing of the trunsformers considerable difficulty was encountered in controlling the furnace temperature and in obtaining useable data. 1t was noted that, when the furnace was on, the Honeywell temperature recorder-controller indicated temperature variations of about 200°F but did not control the furnace. It was also noted that the Dynalog recorder, used to indicate the differential transformer output, becamc insensitive and locked in one position when the heaters were on and that the temperature of the transformer core was 30 to 50°F higher than the outer transformer temperature. Investigation showed that the heaters in the Marshall tube furnace were spirally wound. The 60-cycle magnetic field produced by this spiral configuration induced voltages in the secondary of the transformer and in the thermocouples sufficient to saturate the amplifiers of the Honeywell and Dynalog recorders. The difficulty was avoided in subsequent tests by heating the trans- former above the temperature required, turning off the heaters, and taking 63 data as the transformer cooled. (Care will be taken in the design of reactor-system heaters to ensure that this difficulty does not occur in the reactor level-element installation.) During testing of the transformers, an interesting phenomenon was observed. When the transformers were heated above the curie point of the i1ron core and then allowed to cool so that the temperature passed back through the curie point, the signal output from the differential trans- former increased very rapidly to a high peak value at the curie point of the iron and then decreased rapidly to a much lower value, thus producing a gpike on the recorded trace at the curie point of the core material. The spikes occurred repeatedly at the same temperature (see Fig. 2.3L4). There 1s a possibility that this phenomencn could be used to obtain an accurate temperature calibration check point or as the basis for a simple high-temperature-alarm device. The lemperature at which the spike occurs could be varied by the use of core materials of different curie points. UNCLASSIFIED ORNL-LR-DWG 68590 60 50 40 30 DYNALOG CHART READING 20 1300 1400 1300 1400 TEMPERATURE (°F) Fig. 2.34. Temperature Effect on OQutput Signal of Differential Transformer as lron Core Passes Through lts Curie Point, While the transformers were being tested, the design of the level test facility was modified to permit testing of an additional level- indicating system. A new tank and head assembly was added, and provisions 64 were made for the detection of molten-salt level with a spark plug probe. Provisions were also made for future addition of a bubbler-type level device. As shown in Iig. 2.35, the new head assembly is designed to mount a differential transformer above the tank. The core for this transformer projects from the top of the float through the gas space above the float UNCLASSIFIED ORNL -LR-DWG 68591 0 e ) SPARK PLUG // PROBE PRESSURE _# INDICATOR LEVEL RECORDER f———-- NO. 2 DIFFERENTIAL TRANSFORMER NO. 2~ . ~CORE NO. 2 Jfi Prirenn TEMPERATURE ~ RECORDERJ ] ]_HELIUM X S - FLOAT NO. 2 o/ LEVEL RECORDER NO. 4 - DIFFERENTIA: TRANSFORMER NO. 4 | | 1 | | | | l | | | | | | | | ( l o / e CORE NO. 4 } FURNACE = ; | | | S | I l | | l | I | | [J CHEATERS- - - . - ’, - A FREEZE VALVE d FREEZE VALVE - NO. 3 - NO. 2 = —— | f-- o8 Fig. 2.35. Simplified Diagram of MSRE Level Test Facility. - 65 and into the core tube inside the differential transformer. This differ- ential transformer is insulated but at present 1s not heated. The end of the transformer next to the furnace is operating between 980 and 1000°F, the other end between U480 and 500°F. While unnecessarily severe, this operating condition does provide every opportunity for the level system to fail. If there are to be any deposits of solids from the vapors above the molten salt in the temperature range 1200 to 500°F, the proper condi- tion exists. If there are deposits, they presumably will accumulate on the core and core tube to the extent that the core will no longer move. Construction of the level test facility and fabrication of the two prototype level-element assemblies was completed, and testing of the level elements is proceeding. A photograph of the completed test facility, made prior to installation of insulation, is shown in Fig. 2.36. In this photo- graph, the differential transformers can be seen above the tank on the left, and below the tank on the right. | UNCLASSIFIED { PHOTO 37897 Fig. 2.36. Molten-Salt Level Test Facility. 66 The system has operated three weeks atl temperature. 5Salt was added after the first weck of operation at temperature. Preliminary results are very cncouraging. They indicate that the temperature effects on span and zero, over the range 800 to 1200°F, are acceptably low and that there is no pressure effect and negligible hysteresis. There has been no evidcnce of salt absorption by the graphite float. The tests are being continued to determine the effects of continuous high-temperature operation on the characteristics of the transmitters. 2.14.5 Single-Point Level Indicator The test of the single-point level indicator previously reported32 was terminated after two months operation. At the end of this period there was an 11% reduction in the amplitude of the signal. Due to the on-off characteristics of the device this was not objectionable. When removed from the salt pot for inspection, the portion of the probe which was inside the pot but above salt level had an even deposit of green material adhering to it. The portion which was below salt level was clean and shiny. The probe was damaged during the inspection, and teslts were dis- continued. A prototype of the probe assembly that will be used in the MSRE 33 is being fabricated and will be tested. REFERENCES 1. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, pp 28-32. 2. Ibid., p 28. 3. MSRP Quart. Progr. Rept. July 31, 1960, ORNL-301k, p 20, L. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 5l. 5. TIbid., p 27. N Sturm-Krouse, Inc., Analyses and Design Suggestions for Frecze Ilange Assemblies for MSRE (Nov. 30, 1960). 7. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 33. 8. 1Ibid., p 37. 9. 1Ibid., p 37- 10. Ibid., p Lo. 11. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 3(-39. 12. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, pp 54-55. 26. 27. 28. 67 lbid., p 5>. R. E. Thoma, "2240 Service Samples (10/19/61)," private communication. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 57. Ibid., p 58-61. Ibid., p 61l. Ibid., p 65. Tbid., p 46. Ibid., p 48. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 49. lbid., p 50. C. H. Gabbard, Thermal Stress and Strain Fatigue Analysis of the MSRE Fuel and Coolant Pumps, ORNL-TM-78 (to be issued). MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 50. Ibid., p 5k4. Tbid., p 5h. Ibid., p 78. Tbid., p 79, Fig. 2.41. Ibid., D 77 Ibid., pp 25-27. Ibid., pp O7-72. Tbid., pp 73-T7h. Ibid., p 76, Fig. 2.39. 68 3. REACTOR ENGINEERING ANMNALYSIS 3.1 REACTOR PHYSICOH 3.1.1 Analysis of MSRE Temperature Coefficient The previously reported analysisl of the temperature coefficients - of reactivity in the MSRE was extended to include the effects of retaining fission-product xenon and samarium in the core graphite and the effect ol including a small amount of a nonml/v absorber for the purpose of increas- : ing the size of the temperature coefficient. Since there are low-lying resonances in ooth Xel®% and Smlég, the av~ erage thermal -absorption cross section Tor these nuclides decreases with increasing temperature above about 200°C; in certain situations this can lead to a positive contribution to the temperature coefficient, due to the reduced poison absorptions relative to fuel absorptions. There are cer- tain nuclides (e.g., Rh*©3) for which the average thermal cross section decreases slowly with increasing tewperature; insertion of such an absorp- er would lead to an increase in the size of the temperature coefficient due to the change in thermal utilization with temperature. The principal contributions to the MSRE temperature coeffilcients are due to the increase in thermal leakage with increasing temperature; alter- ation of the thermal utilization with temperature is not a large effect. Therefore the cnanges in temperature coefficients resulting from either - fission-product buildup or poison insertion are small. Results of calcu- lations for rhodium insertion are shown in Fig. 3.1; salt, graphite and total temperature coefficients are plotted vs the ratio of the thermal- macroscopic~apsorption cross section of tne unpoisoned core. I the as- sumption is made that all fission-product Xe*3® and Sm»*° are retained in the core grapnite, then the grapaite tewmperature cocificient 1s changed from ~6.0 x 1075/°F to -5.2 x 107°/°F; the salt temperature coelficient is changed from -2.8 x 1075/°F to -3.1 x 1075/°F, so that the total is changed from ~8.8 x 107°/°F to -8.3 x 1077/°F. 3.1.2 Reactor~Kinetics Studies Prelininary results were obtained from simplified reactor-kinetics calculations®’ 3 of the response of tne MSRE to step-reactivity insertions. A modified version of the PET-I 7O pr@grama was put into operation on the IBI 7090; additions to the program include revised input and outpult formats and the treatment of separate fuel-salt and graphite temperatures and cem- perature coefficients of reactivity. Power, fuel-~salt temperature rise, and grapnite-temperature rise are plotted as a function of time aflter o 69 % step insertion (starting from 10 Mw) in Fig. 3.2; peak pressures are plotted as a function of step size in Fig. 3.3. UNCLASSIFIED ORNL-IL.LR-DWG 68592 {x1075) -10 p SUM o —9 W —— =8 . o o~ s = GRAPHITE L o L L L Q o bt T -4 ’_ To evaluate the susceptibility to salt penetration for CGB-X graphite, the pore~-entrance-diameter distributions of the accessible pores of two specimens were determined by using a mercury porosimeter. The data indi- cated that at approximately 65 psig, the maximum operating pressure ex- pected in the MSRE, 0.12% of the bulk volume of the graphite should be permeated by salt. (Under the conditions of a standard fluoride screen- ing permeation test, 0.16% of the bulk volume of the graphite should be permeated. ) The pore-entrance-diameter distribution indicated that most of the accessible pores of CGB-X graphite had entrance diameters less than 0.3h u. Since theoretically, the molten-salt pressure would have to be 265 92 psig to force salt through a 0.3h4-pu capillary, the pressure required to permeate the major portion of the accessible pores would have to he greater than 265 psig. These are qualitative data because of the limited sampling (only two specimens) of the graphite, the inherent limitations of the mercury poro- simeter, the lack of accurate data on the surface tension of the salt, However, the results should represent the worst conditions, because the specimens were small and oriented in such a way as to expose the maxi- mun quantity of accessible pores. The overall volume permeated by molten fluorides in a full-scale MSRE graphite bar should be less than the quali- tative, theoretical values above because of (1) better pore-space orien- tation in the bar and (2) because the shallow permeation by salts as described above in the radiographic examinations would be "diluted" by the large salt-free core of the graphite bar. L.4,1 Comparison of the Permeation of Graphite by Mercury and Molten Fluorides The permeation of a particular grade of graphite by a molten salt is the preferred way to determine the suitability of the graphite to be used in the special conditions found in a molten-salt reactor. However, this would be a somewhat cumbersome quality-control test for a graphite vendor. A mercury-impregnation test was proposed as a method that could be related to the molten-salt permeation of the graphite. Calculations showed that a U52-psig pressure on mercury at room temperature should cause it to permeate graphite to the same extent as does LiF-BeFs-ThF4-UF, (67-18.5- 1%4-0.5 mole %) salt at 150 psig and at 1300°F in the standard salt-perme- ation screening test. Tests have been made with various grades of graph- ite to evaluate the permeability relationship between the mercury and salt. In these tests, the specimens were evacuated to 50 ¢ Hg prior to being submerged in mercury at room temperature; pressure of U452 psig was applied and held for 20 hr. Similarly, pieces of the specimens were sub- jected to the standard fluoride-salt-permeation screening test. The re- sults are summarized in Table L.13. Comparison of the bulk volumes of the different grades of graphite permeated indicated that the mercury-impregnation test was approximately equivalent to that of the standard screening test. 'The greatest differ- ence between the percentage of the bulk volume permeated by mercury or by molten salt was observed in CGB-X graphite. Here mercury took up more volume than the molten salt. It is believed that this may be due to the heterogeneity of the pore distribution of this grade and/or that the sur- face tension assumed for the molten salt is low. In the Specification for Graphite Bar for Nuclear Reactors, MET-RM-1 (5~10m61), the mercury pressure required is W70 psig, and the maximmun permissible weight galn in the graphite specimen due to mercury impreg- nation is 3.5%. On this basis only grades B-~1 and CGB-X would meet the mercury requirement of the specification. Table 4.13, 93 Comparison of the Permeations by Mercury and Molten Fluorides into Various Grades of Graphite Test Conditions:® Molten SaltP Mercury Temperature, °F: 1300 70 Test period, hr: 100 20 Pressure, psig: 150 L52 Bulk Volume of Graphite Welght Gain of Graphite Grade of Dimensions of Specimens Permeated (%Qd Permeated with Graphite (in. )¢ Salt Mercury Mercury (%) CGB-X 1.50 x 0.50 diam 0.02% 0.09° 0.6 CoB-X 0.125 x 1.50 x 1.50 0.01 0.05 0.k B-1 0.125 x 0.9 diam 0.05 0.06 0.4 Sl -IB~D 0.75 x 0.4 diam 0.4 Sl -LB-D 0.125 x 0.9 diam 0.5 0.6 b.5 CS-112-8 0.125 x 0.9 diam 0.k 0.8 5.9 RHA-1 0.125 x 0.9 diam 0.6 0.7 5.6 S-l-LA-C 0.75 x 0.4 diam 0.8° Selt-TA-C 0.125 x 0.9 diam 0.8 0.9 6.8 R-0009 1.50 x 0.50 diam 2.0° 2.0° 1.1 R-0009 RG 1.50 x 0.50 diam 2,7° 3,2 22,5 R-0025 1.50 x 0.50 dian 5.7 5.2° 37.2 AcoT 1.50 x 0.50 diam 13.9%7 1h,b3 115.8 aSpecimens were evacuated, submerged in the molten salt (or mer- cury), and then the pressure was applied to the molten salt (or mercury). Pryis was the TiF-BeFao-ThF,-UF, (67-18.5-14-0.5 mole %) mixture. C . - - Nominal dimensions. dThe superscript number showvn on the figures in the column indi- cates the nunber of values averaged; the absence of the superscript denotes that only a single value is involved. 94 h.h.2 Removal of Oxygen Contamination from Graphite with Thermal Decomposition Products of NH4FHE Tt was previously r‘@ported,l4¢ that six graphite crucibles were purged with the thermal decomposition products of NH4F-IF and then were exposed to "oxygen-sensitive” LiF-BeF--UFs (62-37-L mole %) salt at 1300°F in order to establish the effectiveness of the purge. (This salt readily precipitates UOs in the presence of small quantities of oxygen.) These crucibles have now accumilated L4000 hr of such exposure with- out any UO, precipitate being detected by radiographic monitoring. This indicates that the purge with the thermal decomposition products of am- monium bifluoride was equally effective in removing the oxygen contami- nation from the graphite crucibles even though the R-0025 graphite is a moderately low permeability grade and AGOT is a relatively high perme- ability grade. The tests were terminated for examination of the LiF-BeF o~ UFs mixture and the graphite crucibles. The success of the preceding tests indicated that lower purging temperatures might be successful. The lower temperatures would have the advantage of being easier to attain in large systems and perhaps would limit the reaction of the decomposition products with TNOR-8, the struc- tural material for the reactor. Therefore, an R-0025 graphite crucible was purged simllarly to the preceding tests except that the temperature was T50°F. The crucible has since held the LiF-BeF--UF, melt at 1300°F for more than 2000 hr, without any precipitate being detected. A similar test 18 in progress in which an R-0025 graphite crucible was purged at 390°F. OCther tests have been made by purging graphite at 1300°F with varying quantities of NiH4F.HF. Two AGOT graphnite crucibles of the same bulk volumes as the preceding were purged with 0.2 and 0.4 g of NH,F-HF, respectively, for 20 hr at 1300°F. The purges appear to have been suc- cessful in removing the oxygen from the graphite because the crucibles have held the LiF-BeF»-Ul'y melt for 1000 hr at 1300°F without any detect- able precipitate forming. Additional exposure time through LO00 hr is planned. h.h.3 Reaction of Zr0, with Thermal Decomposition Products of NH4F.HF An undesirable refractory sludge found in the MSRE Engineering Test Toop™™ was tested to establish its composition and to determine a method for removing it from the loop. X-ray diffraction analyses indicated that the sludge consisted mostly of Zy0»o (monoclinie) and Zr¥F4, plus some Belo and traces of other materials that could have been fluorides. Chemical analysis showed the following: W11 7r, 2k.6h F, 15.9% 0, 4.85% 14, 3.29% Be, and 0.01% Na.>® One gram of the sludge was exposed for 20 hr to the thermal decompo- sition products from 1 g of NH4F.HF crystals at 1300°F to determine 2 95 whether the oxygen of the ZrOp could be replaced by fluorine. This would reduce the refractoriness of the sludge and make it easier to remove from the pump bowl, The particulate sample of the sludge reacted with the NH4F.HF decom- position products during exposure, became molten, and wetted its Inconel container. X-ray analysis of this material indicated that it was pri- marily 2LiF«ZrF4, plus a moderate amount of 3LiF-4ZrF; and a small quantity of unidentified material. There was no evidence that any monoclinic ZrOs remained, indicating a maximum of 5 wt % of ZrOo. The conversion of the refractory oxide sludge to the fluoride form as described above should permit its removal from the loop by the use of a molten-fluoride flush salt. REFERENCES lo Ro Bo Briggs et alo, I"/ISRP ngro Reptc Augo 31, 1961.’ ORNIJ"'3215, p 99-100. 2. C. M, Blood, Solubility and Stability of Structural Metal Difluorides in Molten Salt Mixtures, ORNL CF-61-5-4, p 23 (Sept. 21, 1961). 3. R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 96-99. 4. R. B. Briggs et al., MSRP Progr. Rept. Feb, 28, 1961, ORNL-3122, W6 p -6, 5. R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 93-96. 6. H. G. MacPherson et al., MSRP Progr. Rept. July 31, 1960, ORNL-301lk, p 63 ""65 ° T. X. V. Cook and R. W. McClung, Development of Ultrasonic Techniques for the Inspection of Brazed Joints (to be published), 80 R. Bo Briggs et 8.1., MSRP PI‘OgI'o Reptn Al)go 31, 1961’ ORNIJ“BE]—-S, pp 109, 111-12. 9. The International Nickel Co., Inconel Welding Electrode "182" and Inconel Filler Metal "82" (April 1961). 10. R. W. Swindeman, The Mechanical Properties of INOR-8, ORNL-2780, P 29-30. 11. 12, 13. 96 H. G. MacPherson et al., MSRP Progr. Rept. Apr. 30, 1960, ORNL~-2973, P 53. R. B. Briggs et al., MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 93. H. L. Ritter and L. C. Drake, "Pore-~Size Distribution in Porous Materials," Industrial and Eng. Chem., Analytical Ed. 17(12), 782 (Dec., 1945). — R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 11k, R. B. Briggs et al., MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, p 5k, Private communication from J. L. Crowley. 97 5. IN-PILE TESTS 5.1 INTERACTION OF FISSIONING FUEL WITH GRAPHITE: TEST NO. ORNL-MTR-47-3 Since MSRE fuel and graphite are chemically inert and thermodynami- cally compatible, with respect to each other, in the absence of radiation, relatively extreme exposure conditions were used to accentuate the effects of fissioning in the irradiation test ORNL-MTR-47-3.1 A power density of 200 w/cc was developed by adjusting the concentration of fully enriched U235F, to 1.5 mole % instead of the <0.3 mole % to be used in the MSRE. Consequently the four 3-week cycles in the MTR, accumulating 1580 hr at power, gave a burnup of 8.5% in comparison with the 6% per year antici- rated for the MSRE at 10 Mw. Because so little is known about why some molten salts wet graphite and some do not, and about changes in interfacial behavior that might occur in a fissioning fuel, the primary purpose of the experiment was to tell whether or not permeation of the graphite by fuel was to be expected. Out-of-pile tests had indicated that it was not and that in clean systems penetration occurred only in response to pressure in the manner expected from the pore spectrum of the graphite and the surface tension of the nonwetting liquid. The contact angle of the fuel meniscus was chosen as a convenient and reliable index of possible changes in wetting. A vertical blade of graphite dipping into a pool of fuel in a graphite boat allowed room in the same capsule for coupons of INOR-8, pyrolytic graphite, and molybdenunm. The choice of a boat with a dished inside contour was influenced by the need for a container which could accomodate freeze-thaw cycles without stress. Capsules in previous experiments (47-1, 47-2)2 had ruptured from freeze-thaw cycles. A step along a portion of the length of the bottom or submerged edge of the blade extended to within 1/16 in. of the floor of the boat; normal fuel, with a surface tension of almost 200 dynes/cm, does not penetrate such a small crevice and thus another device for detecting wetting behavior was provided. 5.1.1 Description of Experiment As described elsewhere in greater detail,3 each of four sealed INOR-8 capsules, depicted in Fig. 5.1, contained a 3/16-in.-thick R-0025 graphite blade (7.3 g) dipping into a shallow pool of fuel (11.4 g or 5 cc) held in a graphite boat 3 in. long and 1-1/4 in. wide. Two of the boats were of R-0025 graphite (75 g), a relatively impervious grade, and the other 98 UNCLASSIFIED ORNL-LR-DWG 56754 METAL LOW -PERMEABIILITY METAL S SPECIMENS SPECIMENS GRAPHITE O J’\\\\.' 5 - 77 0N < A 7 —MOLTEN- ! SALT FUEL GRAPHITE BODY \THERMOCOUPLE WELL 0 Ya 4/2 3a 1 - Lo b 0] INCH Fig. 5.1. MSRE Graphite-Fuel Capsule Test ORNL.-MTR-47-3, two were of AGOT graphite (66 g) which had been preimpregnated with 9 g of fuel. The capsules, horizontally aligned in a vertical diamond array, were contained in a sodium bath which served as a heat transfer medium. Thermocouple wells, which were affixed through one end of each capsule, measured the graphite temperatures at a point roughly midway between the bottom of the fuel and the capsule wall. Coupons of INOR-8, pyrolytic graphite, and molybdenum were attached to the blade and were partially submerged in the fuel. The fuel was purified by treatment at 800°C with HF and with H,; oxide was removed from the graphite by degassing in a vacuum at 2000°C. All sub- sequent manipulations were carried out in a helium-atmosphere glove hox to avoid contamination by the atmosphere, and several extra capsules were . assembled at the same time to provide control specimens. The design neutron flux was 2.3 x 10+3. This figure and 8.5% burnup have been adopted for purposes of calculation since somewhat discrepant data from dosimeter wires average to a lower value, while a comparison of the ratio of uranium isotopes, available from mass spectrometry, indicates a higher value. The disassembly and initial observations were carried out in hot cells at the Battelle Memorial Institute under the supervision of ORNL Reactor Chemistry personnel, and later the separated pieces were sent to ORNL for further study. The wvapor pressure of the fuel increases by approximately a factor of 10 for each 100° interval between 700 and 10000C, smounting to about 0.2 torr at 1000°C. At MSRE temperatures the vapor pressure is negligible, and although the control capsules showed no evidence of distillation, a considerable amount of vapor-phase transport (enhanced by the temperature gradient toward the cool walls as well as by the high fuel temperature) 99 did occur in the in-pile capsules. The composition of the vapor seems to be slightly to the BeFo-rich side of the stoichiometry for LisBeF4, and although the vapor is much poorer in quadrivalent cations than the liquid fuel, ZrF4 is present perhaps to an extent of a few tenths of a percent. To a very good approximation, distillation of the fuel corresponds to the removal of LisBeF4. If an estimated 2.5 g were lost from the lig- uid fuel in capsules such as No. 15 or 16, leaving 8.9 g of a fuel, the LiF~-BeF ,-ZrF,-ThF,-UF, proportions would have been altered from 69.5-23- 5-1-1.5 mole % to about 70-17-7.5-1.5-2 mole %. Several effects of the distillation were noted. The condensation occurred predominantly at cool metal walls, firmly cementing the boats to the capsule walls in a manner that required that the walls be sawed off in strips to open the capsules. This involved some slight mechanical damage to the contents. FEnough distilled salt was lost and fuel salt scattered to prevent obtaining a satisfactory mate- rial balance on either the amount distilled or the amount remaining in the fuel pool. Some condensation occurred in cooler ports of the graphite and precluded interpretation of the weight changes of the graphite in terms of liquid permeation. The concomitant change in fuel composition engen- dered a crystallization path with LiF as the primary phase. A considerable amount of the distilled salt was found in the form of condensed droplets, resembling pearls, that evidently dislodged from the capsule walls and fell onto or among the broken pieces of frozen fuel. 5.1.2 Dismantling of In-Pile Assembly After irradiation was campleted, on July 27, 1961, the in-pile assem- bly was shipped to the Battelle Memorial Institute Hot Cell Facility for dismantling. The outer water Jjacket was sawed off, and the sodium tank was cut by a machining device which avoided tumbling the capsule contents. The sodium was melted under mineral oil, and the capsule assenmbly was lifted out. The capsules appeared to be in good condition. The individual capsules were cut loose from the sodium-tank bulkhead and the dosimeter wires were removed. Micrometer measurements of the outside diameters of the capsules showed no change from original dimensions. No bulging, bow- ing, nor distortion was noticeable. After drilling for gas sampling, each capsule was opened by cutting through the welds on the recessed end caps with a specially designed cutter in which the capsule was clamped in a stationary vise and the cutting tool revolved around the work. The end caps were pried off with the help of a prying bar and a split collar with a tongue seating in the groove milled by the cutter. As mentioned previously, the graphite boats were sealed tightly into the INOR-8 capsules by salt which had volatilized into the narrow space between them; therefore, the boats could not be pushed out of the capsule. A remotely operable cutting machine was designed and built which permitted the capsule to be clamped horizontally on a milling table and to be moved 100 past a side-milling cutter turning at low speed. Fach cylindrical capsule wall was cut longitudinally into three sections which then were pried loose, exposing the graphite boat. Three identical unirradiated capsules, which had been thermally cycled in a manner similar to the irradiated capsules, were likewise subjected to gas sampling and then opened to provide blank samples of salt, graphite, and metal with which to compare the irradiated materials. 5.1.3 Temperatures Because of thermal convection in the sodium bath in which the capsules were immersed, the upper capsule, No. 3, operated at higher temperatures than its duplicate, No. 8, which was at the bottom of the diamond array. Capsules 15 and 16 were cooler because of the absence of preimpregnated fuel in the boat. The operating temperatures are given in Table 5.1. The temperatures ran 40 to 45°C higher initially than at the end; almost half the decrease occurred in the first 70 hr, possibly as a consequence of improved. heat conduction as salt-vapor condensate accumulated in the gas gap between the graphite and the capsule wall. 5.1.4 Gas Analyses The nominal volume of the gas space in the sealed capsules, not count- ing about 4.5 cc of voids in the graphite, was about 16 cc; the initial amount of capsule gas, though not known, was estimated as about 17 cc (STP) by assuming an average gas temperature of 50°C at the instant of sealing. Gas samples were obtalned by drilling the end of the capsules in an evacuated chamber which enclosed both the drilling apparatus and the cap~ sule. Before drilling was started, leak rates were reduced to acceptably low values. The released gas was transferred by a Toepler pump and an associated manifold into sample bulbs equipped with "break-seals." A condensed summary of the results from mass spectrometry is given in Table 5.2. As reflected in amounts shown in the second column of Table 5.2, some of the capsules gave samples with much-smaller-than- expected pressures. The difficulty stemmed from a varying combination of lack of sufficient Toepler-pump cycles, lack of adequate instrumenta- tion, and lack of information on volumes of various parts of the system. But, even after these points were at least partially remedied, there was evidence for a slow rate of gas release from the capsules. A partial blockage due to the pressure of condensed salt vapor may have contributed to the difficulty in transferring the expected volumes of gas. The Generation of Carbon Tetrafluoride The relatively large amounts of CF4 shown in column 4 of Table 5.2 represent a noneduilibriuwn condition wnich had not been fully anticipated; usually temperatures as high as the 800 to 900°C prevailing in the capsules supply sufficient activation energy to prevenl the accumulation of species Table 5.1 Time-Averaged Operating Temperatures and Identification of Irradiated Capsules Capsule Average Operating Temperature (OC) Number Graphite? Pretreatment Fuel, Max. Blade-FuelP Boat-Fuell Thermocouple (estimated) (estimated) (estimated) in Graphite 15 & 16 R-0025 2000°C in vacuum 835 790 730 710 PN 8 AGOT Preimpregnated with 850 £ 810 750 730 9 g of fuel o 3 AGOT Preimpregnated with g5 00 850 825 9 g of fuel ; 8R-0025 graphite is relatively impervious compared to AGOT. PInterface temperatures; no allowance for films at the fuel interface was made. 101 102 Table 5.2 Off-Gas Compositions in Volume Percent Averaged values from mass spectrometric results obtained at ORNL and BMI Standard ccd Components Capsule No. Sampled AirP He CF, Xe Kr o, Ar® 15 10 5.84 80.6 09.85 0.012 1.40 0.0k 2.29 16 19 22.75 61.45 8.73 0.003 1.1k4 4,73 1.62 8 7 5.36 76.9 2.32 11.82 1.96 0.11 k.20 3 17 L.57 79.6 0.67 11.43 1.78 0.10 0.96 Contro1d 10 7.38 87 <0.0003 0.17 5.5k4 aApparent volume of gas transferred from gas-release chamber. bMainly inleakage to the evacuated release chamber while drilling the capsules. CArgon was supplied as a blanket gas for the welding arc used in a helium-atmosphere glove box to seal the capsules. d'I‘he control was heated through cycles roughly corresponding to the thermocouple reading vs time for capsules 15 and 16. that are thermodynamically unstable toward reaction by several tens of kilocalories. The main source of the CF4 was the graphite-fuel interface. The main sink for CF4 is the fuel. Most of the CF4 from the source was probably consumed immediately by the fuel, but some fraction of the production dif- fused into the graphite and thus was bypassed into a reservoir where it was preserved until it could again come in contact with fuel. The voids in graphite are interconnected and, of course, commmnicate to the gas space above the boat. In the voids or the gas space, conditions were not favorable for access of CFy to a reactive surface. Ehe frequently encountered kinetic inertness of CF, in many of its reactions™ was also a contributing factor. Since a higher concentration of CF was found in the capsules (15 and 16) in which the generation rate was smaller, the consump- tion rate evidently was controlling as far as the steady-state concentration in the capsule gas was concerned. Conditions which may have diminished the consumption rate include the films of condensed salt vapor, essentially Li.BeF,, which coated the cap~- sule walls. Hotter regions appeared to have been covered by a film resembling a carbon or graphite deposit from the pyrolytic or radiolytic decomposition of CE,. Even in the control capsules a film of black dust, 103 presumably graphite machining dust, accumulated on the surface of the fuel ingot. In any case the consumption reaction 4id not proceed rapidly enough to restore equilibrium conditions. A possible reaction mechanism which accounts for the faster consump- tion rate in the capsules that contained the prepermeated boats proposes that CFy reacts most rapidly by means of three-phase contact (gas, graphite, and fuel) which allows a heterogeneous reaction; this mechanism leans on the fact that the reduction of CEy even by "unreduced" fuel is thermodynam- ically favored. Regions of three-phase contact were much more abundant in the prepermeated boats. The higher temperature of the fuel-graphite interfaces in the prepermeated cases was also of importance in accelerating the consumption reaction. The consumption of CFs by dissolution in the fuel and subsequent homo- geneous reaction is probably slow, but even for this mechanism, the area of fuel-gas interface was greater in the prepermested boats. The point of concern about the CE, generation is the removal of fluo- ride ions from the fuel-—the reduction of the fuel as CF4 is carried away in the off-gas. If this removal of fluoride occurs in a system with sub- merged graphite, like the MSRE, the most readily recognized manifestation of the reduction would be the conversion of Ufy to UFz. As the concentra- tion of UF, increases, the disproportionation reaction 4UFs — 3UF, + UPY leads to the formation of metallic uranium alloys with the container and also to the formation of uranium carbides by reaction with graphite. Although the presence of a measurable amount of reducing power in the fuel from the capsules has not been satisfactorily confirmed, it is instruc- tive to compare the amounts of CF4 accumulating in the capsules on the basis of the calculated percentage conversion of UFs to UFs in the fuel. This was done in Table 5.3; no allowance has been made for the anticipated reduction due to the fact that the fissioning process produces a total cation valence requirement greater than can be matched by the four equiv- alents of fluorides from a gram atom of fissioned uranium. Xenon In Table 5.4 the amount of xenon found in each capsule is compared with the amount expected for 8.5% burnup of the U®®. For the two pre- impregnated cases, 3 and 8, the proper amount was found, but in the other two cases only about 0.1% of the expected yield was found. The missing xenon has not been located, but there is a possibility that it was somehow selectively absorbed in some L feet of gum rubber tubing that was used in the gas-collecting system. Also an attempt is underway to analyze the irradiated graphite for xenon, but since the parts have been exposed to air several months there is small probability that the graphite analyses will resolve the question. There were no signifi- differences in the isotopic distribution of the xenon recovered from the four capsules. 104 Table 5.3 Calculated Conversion of UF, to UF,, Based on CF4 Evolution Average Capsule Vol % CF,& Total VolumeP Yield Calculated® Found (std. cc) (std. cc CFy) % U*" Reduced 3 0.67 23.5 0.157 0.45 8 2.32 21.1 0.49 .41 15 9.8 16.5 1.62 8.3 16 8.73 20.5 1.79 9.1 aAverage of BMI and ORNL mass spectrometer analyses. Ppased on krypton analyses and 8.5% burnup. CAccording to stoichiometry of MUF4 + C — CFy + UFs, including impregnated UF4 in calculation. Table 5.4 Comparison of Results of Xenon Analyses With Theoretical Yield Capsule Vol % Xe® Total Volume® Total Xe Theor. Total Xe Found (std. cec) (Std. cc) (std. cc) 3 11.43 23.5 2.69 2.86 8 11.82 21.1 2.52 2.83 15 0.012 16.5 0.0020 1.59 16 0.007 20.5 0.0014 1.61 aAverage of BMI and ORNL analyses. bBased on krypton analyses. c Based on 8.5% burnup. 105 Ezzgton The behavior of the xenon is even more puzzling in view of the fact that the krypton yields were normal, both in amount and in the proportions of the isotopes. 5.1.5 Test Effects on Graphite Dimensional, Weight, and Electrical Resistivity Changes in the Graphite Dimensionsg of the graphite boats and blades did not change within the precision with which the measurements in the hot cell were made. Changes greater than 0.1% should have been detectable. Weight changes of the graphite parts were also not very meaningful, since they represented the combined effect of distilled salt condensed in the pores, broken or lost graphite, and of some sticking salt that was difficult to remove. The blades of unimpregnated boats gained weight, probably as a result of salt-vapor condensation. The preimpregnated boats lost weight by distillation. The electrical resistance of the graphite parts approximately doubled as a result of the irradiation exposure. Such changes have been attributed to the trapping of conductance electrons in defects induced by radiation. Hardness of Graphite Rockwell hardness measurements on the R-0025 graphite increased by 10% from 9 to 108, as a result of the exposure. The AGOT graphite boats increased from about 56 to about 80 as a result of preimpregnation with salt, and the readings after radiation were not significantly higher. No differences were found between regions under the salt pool and else- where in the same boat. Wetting Behavior When the pieces of the fuel ingot, jostled by the disassembly opera- tions, were fitted into their original positions in the boats, the meniscus was observed to be the same as in the control specimen and to definitely show nonwetting behavior toward graphite; the metal coupons were wetted. Fuel was not found in the 1/16-in. clearance between the blade and boat but did fill the region under the blade where the clearance was 1/8 in. The nonwetting behavior is receiving further confirmation from metallog- raphy and autoradiography of the graphite, along with analyses of semimicro cores drilled from the graphite. Information from weight changes of the graphite, though uncertain, was at least indicative of no pronounced per- meation of the graphite by fuel. 106 5.1.6 Analyses of Graphite Autoradiography and Gamma-Ray Spectrometry Cut sections of the graphite parts were autoradiographed with a beta- ray-sensitive film to determine the overall distribution of radioactivity in the bulk graphite. On the basis of information from autoradiography, 77 selected sites were drilled out, furnishing 1/32-in.-diam core samples, which were checked for specific activities (gross beta and gamma) and gamma spectra. Unimpregnated Graphite.--The unpreimpregnated graphite seemed to have not been penetrated by the fuel, although there was considerable radiocac- tivity present. The principal gamma activity was found to be Zr-Nb°S> . (frequently 106 dis min-1 mg-1), which appeared to have been dispersed by distillation of the fuel; relatively high readings in the lower-temperature portions of the graphite were fairly common. The beta activity, which was - roughly paralleled by the Rul®®7109¢ getivity, generally diminished gradu- ally with distance from the fuel interface; there were sharp lines on the autoradiographs indicative of pronounced activity at interfaces, stronger for liquid than for vapor exposure. On the bhasis of experience in an earlier tests with fuel that did not contain 7rF4, Cs*>% and Cs'®7 activity was expected in the graphite at perhaps 10D dis min~t mg“l. Surprisingly, the cesium,activ}ty appears to have been swamped; none was found, though 10” dis min-1 mg — for cesium should have been detectable. The autoradiographs had a grainy appearance, as though the activity had accumuwlated in finely dispersed but relatively large pores. Preimpregnated Graphite.--With the exception of the absence of Cel%% and of ruthenium activity, there were no recognizable abnormalities in the gamma. spectrum of the preimpregnated graphite. Both the beta and the Zr- Nb 2 gamma activity were uniformly present at more than 100 times the intensity encountered with the unimpregnated graphite. Attempts to under- stand the failure to find Ce'** or ruthenium have been initiated. 5.1.7 Test Effects on Coupons Metallographic examinations of coupons on INOR-8, pyrolytic graphite, and molybdenum, which were attached to the blade, have not been completed, but visual inspection revealed that the molybdenum had been severely corroded, having lost about half its thickness, while the other specimens appeared unaffected. INOR-8 wires binding the coupons to the blade had become brittle. 5.1.8 Test Effects on Fuel Petrographic and X-ray Examination Except for discoloration the fuel appeared, under the optical micro- scope, to have no excepltional features when the composition changes were 107 taken into account. Atbtempted x~-ray examination proved unfruitful because of interference from background gamma radiation. Petrographic observation established that fuel composition had shifted sufficiently that LiF was the primary phase and that the main cause of the dark color was a brown discoloration of LiF. The compound Li-BeF,, as expected, was also distin- guishable in the irradiated fuel; scme of the crystals of this compound, ordinarily uncolored, had a faint brownish-purple tint as a result of the radiation. No oxide, no UFs, and no opaque materials were found. Attempted comparisons of samples from the exterior and interior of the fuel ingot revealed no evidence of segregation. Chemical Analyses Because of the bad effects which a strongly reduced fuel would have on MSRE operation, the most important question posed for chemical analysis was whether or not the fuel was reduced to the extent implied by the amocunt of CF, produced. Although the feasibility of such a determination was moot, an attempt was made and is still in progress. To preserve the reducing power of the samples, single chunks or sege ments representing a complete cross section of the fuel ingot were added to the dissolver without grinding. Facilities for grinding, homogenizing, and transferring powder in an inert atmosphere in the hot cell were not available, but complete dissolution of the sample under an inert atmos- phere was attainable in about 4 hr. Dissolution in HC1-HBO, solutions under an atmosphere of helium or argon gave an evolution of hydrogen, the amount of which was ascertained by mass spectrometry of the cover-gas. Six dissolutions producing hydrogen at less than 2% concentration had been carried out in a hastily contrived dissolver and gas colleclting appa- ratus when results from mass spectrometry on the gas samples indicated that difficulties had been encountered. Efforts to establish the correc- tion for radiolytic hydrogen, which accounted for most, if not all, of the yield led to discrepancies and erratic results on the apparent rate of production of hydrogen; the trouble was attributed to the design of the sweep-gas system and to the lack of a means of concentrating the recov- ered hydrogen. Further dissolutions were postponed until improved techniques could be developed and tested. After a two-months development period an assembly was adapted in which COp was used as the sweep gas, and the hydrogen was collected sbove KOH solution in a gas burette from which any insoluble gases could be directly removed for identification by mass spectrometry. Results from dissolutions in the new apparatus are not yet available; the earlier results can be tentatively interpreted as indicating that only in capsule 16 (two samples) was there any evidence of reduced species, and this is not certain. With the exception of O.l% CH, in the gas from capsule 15, for which duplication has not yet been attempted, none of the gas samples gave detectable amounts of hydrocarbons (limit of detec- tion 0.01%) which were sought as evidence of the presence of carbides. 108 There were several reasons for selecting chunks or radially fractured segments of ingots for analytical samples. This was the best way of pre- serving the reducing power. Sampling for purposes other than chemical analysis had already depleted half the ingot from some of the capsules, and no complete ingots were available. Only untested grinding techniques were available, and there was no evidence for, or reason to expect, exten- sive segregation around the ingot. The results of the analyses for major constituents, shown in Table 5.5, did not provide a good basis for conclusions regarding the overall composition of the ingots and suggested that rather extensive segregation has occurred. Six more samples, giving a total of twelve, are scheduled for dissolution, but the results are not yet available. Since the scatter of the results in Table 5.5 was somewhat larger than expected, a program was initiated to validate the hot~cell procedures for uranium analysis, using reference samples from a large batch of normal (unenriched) fuel. A fuel which on the basis of previous routine analyses on three different dates was believed to contain 5.8 + 0.2% uranium (mean deviation) was used for reference samples that were reported by routine facilities to contain 5.57% uranium and by hot-cell procedures to contain 5.8 + 0.1% uranium. The solution which was used in obtaining the 5.57% value in the routine facilities gave a value of 5.79 when checked at the hot-cell site. The implied bias, if any, between routine and hot-cell analyses for uranium was regarded as negligible for ingot samples of pre- sent interest. Results of spectrographic analyses for corrosion products are shown in Table 5.6. The chromium results, if interpreted as Cr®', were compara- ble with, though slightly higher than, equilibrium values that have been found for the isothermal corrosion of INOR~8 in the absence of irradiation; they were at least an order of magnitude higher than expected for a "reduced" fuel. In the absence of irradiation, INOR-8 corrosion equilibrium concen- trations for Fe2" any greater than 20% of the Cr?' concentration have been questioned; the iron concentrations in Table 5.6 were so high that efforts to establish the possibility of accidental contamination are indicated. As Tar as the direct interpretation of the analytical data on CFa and. on corrosion products so far available were concerned, the fissioning process wag oxlidizing. According to fission ylelds, the fissioning pro- cess, even if slightly oxidizing, is not sufficiently so (for 10% burnup ) to correspond to the amount of CF, found in capsules 15 and 16. Analyses of Fuel by Gamma-Ray Spectrometry Attempts to compare the exterior and interior of the fuel ingot by gamma~ray spectrometry showed that ruthenium was concentrated at the surface, particularly at the gas interface. Cerium-1kl, barely detectable in samples from the surface, was prominent in dissolved samples that in- cluded interior portions. In all cases the predominant activity was the mass-95 zirconium-niobium decay chain. 109 Table 5.5 Preliminary Results of Chemical Analysis of Fuel Salt from MSRE Test ORNL~-MTR-47-3 Component (wt %) 232§cga§£ U h 7r® Be® 1iP Original batch® 7.19 6.08 10.4 L.57 10.5 Unéiii?iiged 6.004¢ 5.1 11.1 L.8% 9.6 Hypothetical® (8.3) (7.77) (13.3) (3.3) (9.5) Capsule 16 5.453 5.17 11.0 4.5 13.3 Capsule 16A 6.340¢ --D 10.8 4.8 9.4 Capsule 15 7.861¢ ..h 1h.5 5.4 1k Capsule 3 11.703 8.6 11.5 4.1 15.2 Capsule 3A 6.519¢ ..h 11.3 - 3.2-3.5 15.2 Capsule 8 9.82 7.59 11.8 4.8 16.7 aSpectrographic determination. Prlame photometric determination. CAnalysis obtained from original preparation. dAnalysis of portion of ingot from unirradiated control sample. ePolarographic analysis of wranium (the other uranium snalyses were by the coulometric method). Tealorimetric determination gave 4.60% Be. &Calculated approximate composition based on distillation of 2.5 g of LipBeFy from original 11.4 g of fuel salt and 10% burnup of uranium. hPrecipitation occurred in original solution before analyses vere completed. 110 Table 5.6 Spectrographic Analysis of Corrosion Products Corrosion Wt % of Corrosion Product Found in Capsule Product No. 3 No. O No. 15 No. 16A Cr 0.06kL 0.077 0.06 0.05 Fe 0.095 0.12 0.21 0.15 Mn <(0.014)& <(0.013) <(0.007) <(0.006) Mo <(0.11) <(0.11) <(0.05) <(0.05) Ni <(0.1k) <(0.13) <(0.07) <(0.06) a <(-~~) indicates below the detectable limit shown. 5.1.9 Conclusions Due in part to the difficulties of hot-cell examinations, some aspects of the postexposure examinations have not been satisfactorily culminated, but the main objectives of the experiment have been accomplished. The nonwetting behavior of fissioning fuel toward graphite has been demonstrated, and the nature of radiation effects which might arise dur- ing MSRE operation has been disclosed. Only the evolution of CF,4 promises to be a potentially serious problem, and this effect may have been greatly accentuated by the boat-and-pool con- figuration chosen for the experiment. In any case the gas occupying the voids in graphite exposed to fissioning fuel will contain several percent of CF4; in reactor operation, however, the graphite is submerged, furnish- ing a more favorable arrangement for the reaction of CF4 with the fuel, and the loss of CF4 to the offgas in these circumstances should be consid- erably less, or perhaps negligible. 5.2 MSRE IN~-PILE TESTING The third MSRE experiment, ORNL-MIR-47-3 (refs 6 and T) was irradiated in the MIR reactor from May 5 to July 24, 1961. The graphite used in the experiment was in contact with fuel but not submerged in it. During the postirradiation examination by Battelle Memorial Institute during October and September, CF, was found in the cover gas over the fuel. As a result, an irradiation experiment, ORNL-MIR-W7-4, has been designed to determine whether CF, will exist in the cover gas over MSRE fuel when the graphite is submerged such that CF, formed at the fuel-graphite interface must pass through the fuel before escaping into the cover gas. Also, the experiment 111 will further demonstrate the compatibility of the fuel-graphite—INOR-G system under thermal conditions at least as severe as those expected dur- ing MSRE operation. The L47-4 experimental assembly contains six capsules. Kach of four larger capsules (Fig. 5.2) is 1 in. in diameter and 2.25 in. long and con- tains a CGB graphite core (1/2 in. in diameter and 1 in. long) submerged UNCLASSIFIED ORNL~LR-DWG 67714R Cr—Al THERMOCQUPLE-—-. NICKEL THERMOCOUPLE WELL NICKEL. POSITIONING LUGS {2)~—. NICKEL FILL LINE ~ NICKEL VENT LINE N \— ~INOR -8 CAP HELIUM COVER GAS (3.5cm3) — N ; ‘ PUNCTURE AREA ; 4 A FOR GAS SAMPLING” K| Jy== - INOR~8 CAN - TN % A XX MOLTEN SALT FUEL (25 ¢g) RN R XXX %X 5 T XD 8 N - Al THERMOCOUPLE 7 Fig. 5.2, Submerged Graphite-Molten-Salt Capsules., CGB GRAPHITE | TS & &‘0’0 / 0.0 % ~NICKEL POSITIONING L.UG approximately 0.3 in. in ~25 g of fuel. The fuel will generate from 62 to T6 w/cc, depending upon the capsule location. The expected surface- averaged temperatures at the graphite-fuel interface and the fuel-INOR-8 vessel interface are respectively 1340 and 1120°F. However, the INOR-8 used to fasten the graphite core is also expected to be in contact with the fuel at 13400F. The two smaller capsules (Fig. 5.3) are designed to study the effect of power density and temperature on the formation of CF4. These capsules contain 1/2-in.-diam cylindrical crucibles which contain fuel at 55 w/cc and 130 w/cc and are expected to operate at ~1320°F and ~1650°F respectively. These capsules, which contain exposed graphite in contact with fuel, should also provide a basis for comparison with the previous experiment to evaluate the effectiveness of graphite submersion in preventing the net generation of CF4. The experiment is to be irradi- ated in the MTR from March 12 to June 4, 1962. Qut-of-pile tests with dummy capsules were utilized in the develop- ment of a hot-filling method and apparatus to ensure that the salt properly 112 UNCLASSIFIED ORNL-LR-DWG 67745 r-NICKEL POSITIONING LUG re-—-PUNCTURE AREA FOR HELIUM COVER GAS _ [ GAS SAMPLING (24 ¢cm3) 7 >><1 l SN |~ MOLTEN-SALT FUEL K] - — (10 g) A HELIUM GAS GAP—T ] - Klig—CGB GRAPHITE 3 ] CRUCIBLE “ige 5.3. Crucible-Molten-Salt Capsule, = INOR~ 8 CAN > 152:’\ SNELITSIIRET I I % 3 # Y, ~ NICKEL POSITIONING LUG covers the graphite and to minimize contamination of the fuel to be irradi- ated. Also, the results from three types of thermal cycling tests on dummy capsules indicate that the in~pile capsules should survive the pres- sure stresses created by the expansion of melting fuel (caused by the thermal cycling anticipated from MTR reactor shutdowns and startups dur- ing irradiation). Preliminary conceptual design is in progress for the fifth irradiation experiment, ORNL-MTR-47-5. This experiment will be similar to 47-4 but will be equipped with a system to purge the capsules to permit measurement of the formation of CF, during operation. REFERENCES 1. MSRP Progr. Rept. Aug. 31, 1961, ORNL~3215, p 117. 2. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 108. 3. F. F. Blankenship et al., Fuel-Graphite Irradiation Test, ORNL~h47-3, ORNL~TM-118 (in preparation). L. T. J. Brice, pp 432-33 in Fluorine Chemistry, vol. 1, ed. by J. H. Simons, Academic Press, New York, 1950. 113 5. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 108. 6. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, pp 101-102. 7. MSRP Progr. Rept. Aug. 31, 1961, ORNL-3215, pp 117-18. 114 6. CHEMISTRY 6.1 PHASE-EQUILIBRTIUM STUDIES 6.1.1 The System LiF-BeFs-ZrF, The ternary phase diagram for LiF-BeFs-ZrFs (ref 1) contains a pri~ mary phase field for LiF which persists to temperatures more than 400°C below the melting point of ILiF (84L4°C). A valley lying just outside the LiF phase field can be considered as a prototype of the MSRE fuel region. The MSRE fuel composition, LiF-BeFo-ZrFu-ThF4-UF,s (70-23.7-5-1-0.3 mole %; liguidus temperature, 441°C), contains only a little more than 1 mole % of additional guadrivalent fluoride not found in the LiF-BeFo-ZrF, ternary system, and on freezing the fuel, no ThFs or UF, appear in equi- librium solids until perhaps half the liquid has solidified. Thus the system LiF-BeFo~ZrF4 provides a useful facsimile of the initial freezing behavior of the MSRE fuel. New informalion on liquidus contours in the low-melting regions is included in the expanded, but still incomplete, diagram for the LiF-BeFo- ZrF4 system shown in Fig. 6.1, and values for invariant points are listed in Table 6.1. Of the three ternary systems involving LiF-BeFo, with either Zr¥,, ThFs (ref 2) or UF4 (ref 2), only LiF-BeFo-ZrFs contains a ternary com- pound. This compound, 6LiF.BeFs.ZrF,s melts semicongruently to LiF and liquid at L75°C, and is the first phase to separate (at 441°C) on cooling the MSRE fuel. Single~-crystal specimens of this compound, when studied by x-ray dif- fraction, were found to be body-centered tetragonal with ag = 6.57 A, cy = 18.62 A, The space group was one that is uniquely determined by the Laue symmetry and systematic absences, namely, Ihi/amd--D}. A density of 3.19 g/cc was calculated for a crystalline fluoride of this stoichiometry, on the assumption that the molar volume of the complex compound is the sum of the molar volumes of the components. This value is in good agreement with 3.06 g/cmS, calculated on the basis of the unit cell dimensions and the presence of four stoichiometric units of LigBeZrFis. In order to satisfy the space-group requirements, the Be and Zr ions must lie in planes that are separated by 4.6k A and are perpendicular to the ¢ axis. The locations of the other ions will be the subject of further studies. 115 UNCLASSIFIED ORNL~LR— DWG 65852 i;g PRIMARY PHASE AREAS @) 2LiF - ZeF, ® 3LiF - Z¢F, © 6LiF - Ber,: ZrF, / N © LiF ® 2LiF - BeF, ® 3LiF - azrfy /- ) @BQFZ BLIF - BeF, - ZF, ® 2R 3LiF-4ZrF, P 520 £ 507 AL N LN b N ~ ~ \ \ e ~ N, e \ NN S ~ % \\\\ 2LIF - ZrF, 596 “ £ 570 d ~ "/ 3LiF- ZrF, 662 ~ c e F 2eF, 662 —_ \,( ~ B ’ \\ £ 598 A S s \\\\ . A NN\ ™~ O ,_-l\ @ g A S ® N 0 0 £358° e \\/ SN s N BeF, 648 K\ 2LiF-BeF -~ P 453 £ 370 548 6LIF - Bef, - ZrF, Fig. 6.1. The System LiF-‘Ber-ZrF4. Table 6.1. Invariant Equilibria in the System LiF-BeFo-ZrFg4 So0lid Phases Present gggposigi§2 (mol;réi gggzzizzt Tem%fg?ture 3LiF«ZrFys, LiF, 2LiF.ZrF, 5 5 20 Peritectic ~4180 LiF, 2LiF«ZrF,, OLiF.BeFoZrF, T 12 1h Peritectic 70 LiF, 2LiF.BeFp, 6LiF.BeFo.ZrFy 67 30 3 Peritectic ~35 2LiF«.BeFyo, 2LiF«ZrF4, 6LAF -BeF-ZrFy 64 32 I Peritectic 428 2LiF.BeFp, BeFp, 2LiF.ZrF4 L7 52 1 Eutectic 358 2LiF.Bel'p, ZrF4, BeFyo ~28 ~U5 2l Eutectic 470 116 6.1.2 The System LiF-BeFs-ZrFs-ThF,-UF4 An important composition section from the five-~component system which contains the MSRE fuel is shown in Fig. 6.2. This section, estab- lished with quenched samples from high-~temperature equilibrations, in- cludes 2LiF.BelFs and the nominal fuel composition LiF-BeFo-ZrFg-Thka-UF4 (70-23-5~1~1 mole %). The dilution of the fuel with coolant (approxi- mately 2LiF-BeF2), as a possible result of a leak in the MSRE heat ex- changer, gives compositions that lie between the fuel and 2LiF.BeFs, as shown in Fig. 6.2. The liquidus temperature varies linearly with mole fraction between the fuel and the coolant. The presence of 1 mole of coolant in 25 moles of fuel changes the primary phase from HLiF.BeFo.7rF4 to 2LiF.Bel». UNCLASSIFIED ORNL-LR-DWG 65684 525 - TR i \ g | i U;! —‘ \Q\‘ ) T ! S 500 : ; & | o . D J = @ ! 2 ‘ % 475 PRIMARY PHASE a -LiF- z L .. w 6 LIF-BeF,-ZrFy - @ 450 \‘ 2 L[F-BeFafi_"J/ ‘ o 425 S 0 | UF, ; i . ’_ ) . "Lw_,,w_, e ‘ ThFg ! 1 ‘ &" ‘1 T RS {fizzé,,/~/*’// : o Fig. 6.2. The System LiF-BeF »ZrF -UF -ThF . £ ///'* | \ The section containing MSRE fuel and 2LiF-BeF2. 20 L. ‘ _ 60 §° e 2 70 . UIF T £ ST ; | £ e -—.«f"«‘w,"’ w“‘"-/fi“m" 80 ,,,,,,,,,,,, e N 0 5 10 15 20 25 30 33l BeF, (mole %) Distillation of the fuel, as exemplified in the high-temperature irradiation described in Chap. 5, closely corresponds to the removal of 2LiF.BeF» from the fuel; the effect of this on phase behavior is also shown in Fig. 6.2. Relatively little distillation occurs before LiF becomes the primary phase and the liquidus temperature rises sharply. 6.1.3 Phase Equilibrium Studies in Fluoride Systems Revisions and additions, discussed in more detail elsewhere,3 were made to the phase diagrams for KF-ThF,, BeF-ThF,s, NaF-YFs3, NakF-ThF,-UbFg4, NaF-BeF,-Tht',, Nab-BelFo-ZrFs, LiF-NaF-ThF, and CrF,~CrFa, and crystal structure studies were carried out on LiSblg, Na¥ly, KsUF7z, KaTh¥y, LiRbF5, and chromium(II,ITT) fluoride. 117 Correlations based on relative cation sizes lead to the expectation that RbF.5rFp, CsF.CalFp, and CsF.BaFs could occur as stable compounds. These expectations were confirmed, and some of the optical character- istics of the crystals were established.* 6.2 OXIDE BEHAVIOR IN FUELS 6.2.1 Removal of Oxide from a Flush Salt The use of gaseous HF containing 20% hydrogen to remove oxide from a flush salt was demonstrated on a charge of LiF-BeFs-ZrFs (62-34-4 mole %) for the Engineering Test Loop.® In the course of operation of the loop, treatments of the charge with known amounts of BeO (~40O ppm) had resulted in saturation with oxide.® Presumably 600 ppm of dissolved oxide (a figure from chemical analysis) was present at saturation with ZrOs. When neither temperature cycling nor an increase in the ZrFs concentration from 1 to 4 mole % was effective in dissolving the oxide, an effort to remove dissolved oxide from the fuel with HF was initiated. The equation is 2HF + 0% — 2F + Hz0?. To minimize corrosion of the Inconel drain tank which held the salt mixture, 20% hydrogen was used to help maintain reducing conditions. This concentration of hydrogen was adequate to keep nickel and iron, but not chromium, reduced to the metallic statei7 ‘the induced corrosion, Jjudged both by the 50~-ppm increase in the Cr& concentration in the fuel, and by metallographic examination of dip legs from the drain tank, was within tolerable limits.® The amount of oxide removed from the melt at 565°C over a period of 70 hr, as plotted in Fig. 6.3, was monitored by estimating the amount of UNCL ASSIFIED o~ ORNL-LR-DWG 66932 g 200 ] ° e = HF - Hy RATIO: ~1/5 (vol) e = 50 | H, FLOW RATE: 2.5 liters/min T b TOTAL WEIGHT OF FLUORIDE SALT: ./ > 137 kg J = e 2 100 - AT 7 - i v 5 50 A~ o o 2 - [ned 8 o//// Fig. 6.3. Removal of Oxides from Engineering é 0 10 20 30 a0 50 60 20 Test Loop by HF-H, Treatment at 1050°F. HF TREATMENT TIME (hr) Hs0 in the off-gas. Since there was 176 kg of salt in the tank, the estimated 215 g of oxide removed corresponded to a decrease in oxygen con- tent of about 1200 ppm. Although the material balance on oxide 1s poor, the procedure does seem to have been effective because less than 200 ppm of oxide was found by chemical analysis of the salt after the HF treatment. 118 6.2.2 The Behavior of Sulfates in Molten Fluorides The total oxygen in a fluoride melt can be comprised not only of oxide ions but also of such ions as S04% and C02®". The case of 8047 is of particular interest since 5047 1is the most common form of sulfur contamination in the starting materials for fuel preparation. A complete removal of sulfur is essential for avoiding the very detrimental sulfide embrittlement of nickel and nickel-based alloys, and of course the oxygen accompanying the sulfur in sulfate must also be re- moved. In the regular purification treatment, sulfates are reduced by hydrogen to sulfides, and HF converts the sulfides to fluorides, with the release of HoS in the effluent gas. The effectiveness of this procedure has been demonstrated previously.8 To learn more about the behavior of 504% in fluoride-melt experi- ments, 920 ppm of sulfur as LisS04 was added to a previously purified mixture of LiF-BeFs (63-37 mole %) at 500°C. After maintaining the melt UNCLASSIFIED ORNL-LR-DWG 68597 450 é ‘ l ® L AT 500°C FOR ~1 WEEK 2 4001 ] 350 INITIAL SULFUR CONCENTRATION: < 3ppm — \ SULFUR ADDED AS Li,SO, = 920 ppm TIME AT EACH TEMPERATURE: ~ 24 hr SAMPLED ~ 8 hr AFTER EACH TEMPERATURE ~ 300 ELEVATION fi £ a & = 2 250 A > J O o Z e 200 - I SR D t d 2 w 150 f(——— — B D 100 50 —d et e e 4 S S — <3 ppm\ 500 550 600 650 700 750 800 850 TEMPERATURE (°C) Fig. 6.4. Apparent Thermal Decomposition of Sulfate lon in Molten LiF- Ber (63-37 mole %) in Copper. 119 at 500°C for about a week under a continuous helium sparge, the tempera- ture was elevated in 50° increments to 800°C. The welt was maintained at each temperature level for approximately 24 hr, Eight hours after each temperature increase, a filtrate sample of the melt was obtalned for chemical analysis. The decrease in sulfur content, shown in Fig. 6.k, indicates an apparent thermal instability of sulfate ion in the fluoride solvent. In a second experiment LipS504 was added to a purified mixture of LiF- BeFo (66-34 mole %) containing 2 wt % uranium as UFg in an attempt to verify the following reaction mechanism: 80,27 —» 0% + 803 UFy + 2027 - UOp + UF~ Following a 48-hr eguilibration at 500°C, the daily temperature in- crease of 50°C, as in the first experiment, was repeated. Filtered samples of the melt were analyzed for sulfur and uranium, and the effluent gas stream was also analyzed by mass spectrometry. The results of chemical analyses of the salt, illustrated in Fig. 6.5, show the anticipated re- moval of sulfur and uranium from solution, but somewhat more sulfur was UNCLASSIFIED ORML~-L.R-DWG 68598 6000 5000 -1 2.5 {wt %) 4000 e e e e S R e e e e 2.0 3000 Al e 1.5 SULFUR REMAINING IN SOLUTION (ppm) URANIUM REMAINING iN SOLUTION 1000 - : A e 0.5 400 500 600 700 800 200 TEMPERATURE (°C) Fig. 6.5. Decomposition of L.i2504 and Associated Removal of Uranium from Solution in LiF-BeF2 (66-34 mole %), 120 lost than prescribed by the stoichiometry of the postulated reaction mechanism. Analyses of the gas, given in I'ig. 6.6, show the presence of 502 and Ho5 in addition to the anticipated SO0s. UNCLASSIFIED ORNL-LR-DWG 68599 millimoles) { SULFUR EVOLVED FROM SOLUTION 400 500 600 700 800 900 TEMPERATURE (°C) Fig. 6.6. Gaseous Products from Sulfates in Molten LiF-BeF2 (66-34 mole %). Table 6.2. Behavior of Metal Sulfates in Melts of Purified LiF-BeF, (66-34 mole %) at G00°C Equivalent Sulfur Found Sulfate Sulfur Added in Solution Added (ppm) (ppm) 115504 2881 823 NasS04 2232 1900 K504 1818 1810 MgS04 2631 1780 CaS04 2328 1240 PbS04 1045 1130 CuS04 1984 1400 NiS04 2048 1400 FeS04 2085 1650 Cro(504)3 2426 775 A15(S04 )3 2779 1830 Ces(S04)a 1601 936 121 In a third experiment the behavior of some metal sulfates in purified LiF-BeFo (66-34 mole %) was observed visually at 600°C in glass equipment. Filtered samples were teken from each preparation for chemical analysis. The results are shown in Table 6.2. Of these mixtures, precipitation was visually observed after additions of NiSO4, FeSO4, Cro(804)s, and Zr(S04)s. Because of the relative complexity revealed by these exploratory experiments, further studies will be required before conclusions can be drawn. 6.3 PHYSICAL AND CHEMICAL PROPERTIES OF MOLTEN SALTS A nunber of studies of the physical chemistry of molten salts, recent- ly reported in some detail, provided the following summary paragraphs which describe fundamental research of interest as a background for chem- ical problems rising from the use of molten salts in reactors,” All the published density data on molten fluoride mixtures were re- examined in order to develop a more useful method of predicting densities of related systems. To a good approximation, the molar volumes of these melts were found to be expressed by an additive function of the components. By using empirically adjusted values for the molar volumes of the compo- nents of the melt and assuming additivity, the densities could be calcu- lated to within 2% of the reported experimental values. Densities of solid, complex, metal fluorides were calculated with an average error of 5.5% by assuming that the molar volume of the complex is the sum of the molar volumes of the simple fluorides that formed the complex. This method of estimating densities facilitated the choice of the nunber of molecules per unit cell for complex fluorides whose struc- tures weres under investigation. Further studies of freezing-point depressions of sodium fluoride showed that trivalent fluorides cause negative deviations from ideality [the smaller the solute cation size, the greater the deviation, and al- kali fluorides caused mainly positive deviations (attributable to changes in London dispersion forces)]. Alkali fluorides caused freezing-point depressions in lithium fluoride somewhat different from those in sodium fluoride because of differences in coulombic forces. Ixcess partial molal free energies of mixing, calculated from liguidus temperatures in the system NaF-LiF, were expressible in terms of concentration and two constants. The enthalpy changes from 87L4.0 to 0°C were measured with a Bunsen ice calorimeter for samples of KF, LiF, and mixtures containing 0.193, 0.385, and 0.747 mole fraction LiF. The calculated molar enthalpies of mixing when plotted vs composition gave an approximately symmetrical curve, with a maximm at ~50 mole % LiF, Previously reported equilibrium constants for the reaction, NiFs + Hp ==2HF + Ni , 122 at elevated temperatures provided a basis for refined calculations lead - ing to an evaluation of /H°z0a.16 and estimation of AF° values at other temperatures for the reaction with crystalline solid and (hypothetical) supercooled liquid NiF, as reference states. From the true equilibrium constant, obtainable directly from these free energy values, and the Ky values, activity coefficients of NiFy have been obtained. These activity coefficients and their changes with solvent composition, especially those with supercooled NiF, as reference state, offer suggestions as to the nature of such solutions. The temperature coefficients of the association constants K,, Ko, and Kjo for the formation of the species AgBr, AgBrgi, and+AggBr+ in molten KNOs and, K; and Kz for the formation of CdBr , CaT , CdBrs and CdIls- in alkali nitrate mixtures indicate that the entropies of these associations are consistent with the "configurational"” entropy calculated from the quasi-lattice model. The entropy of association of Ag with the polyatomic ion CN™ is much more positive than for the associations with monatomic cations. The influence of solvent on the association constants for the formation of the associated species AgBr or AgCl in the molten solvents NalNOs and XNOs is consistent with the reciprocal coulomb effect. The perturbation theory of Reiss, Katz, and Kleppa for systems with a common anion was extended to the third- and fourth-order terms to yield the equation for the excess free energy of mixing of uni-univalent salt mixtures. ‘M&“‘E‘%‘ = XaXoPOZ + Kife (X1 - X2)Q0% + (XaXeR + XaXz (X3 - Xp)78)8% + ... where X; is a mole fraction of component 1, O is related to the ionic sizes, and P, Q, R, and S are constants. The third- and fourth-order terms are necessary to rationalize experimental measurements in molten salts. An estimate of the contribution of London dispersion energy to the heats of mixing of mixtures of alkali nitrates with AgNOz or T1NO3 is consistent with the observed differences between these mixtures and mix- tures of alkali nitrates. This suggests a method of making estimates of this effect in mixtures containing other polarizable cations. 6.4 GRAPHITE COMPATIBILITY 6.4.1 The Behavior of Carbon Tetrafluoride in Molten Fluorides Recent postirradiation examinations of in-pile test capsules (ORNT~ MIR-4T-3)?© disclosed the presence of CF4 in the cover gas above graphite boats containing MSRE fuel. Carbon tetrafluoride was absent in unirradi- ated control capsules and represented a wmarked departure from thermodynamic equilibrium in the irradiated capsules. 123 To study the kinetics of the reaction of CF4 with the fuel, the CF4 pressure over the fuel (LiF-BeFp-ZrF4-ThF4-UF,, 70-23-7-1-1 mole %) in closed static systems was measured over a period of time. No evidence of reaction with either normal or reduced fuel was found at 600°C. Since the conditions for pressure measurements were not closely controlled and a slow rate of reaction could have escaped observation, more refined experiments were continued. A gas mixture of CF4 and helium was recirculated continuously, bub- bling through a 6-in. depth of melt at 600°C. Accurately measured volumes of gas were initially admitted to the system, and gas samples were peri- odically withdrawn for analysis by mass spectrometry. Based on the CFy- to-helium ratio (about 40:60) in the analyzed samples, there again was no discernible reaction in 25 hr at 600° with either reduced or unreduced fuel, Two factors make these results tentative rather than conclusive. There seemed to be discrepancies in the gas analyses, and control samples of the reduced fuel exhibited none of the characteristic phases, colors, or reducing power expected on the basis of the amount of zirconium metal added as the reducing agent. More definite results were obtained from circulating a CF4 and helium mixture through a salt mixture containing an indefinite amount of oxide. Figure 6.7 shows that, on a per-mole-of-helium basis, 0.15 mole of CFy was consumed, and 0.007 mole of COp was produced. Although the stoichiometry of the reaction 1is quite puzzling (a black deposit, presumably elemental carbon, was noted, and the amount of CO has not yet been determined), the evidence that CF4 reacts with oxide in the fuel is encouraging. UNGCLASSIFIED ORNL-LLR-DWG 68600 0.60 0.020 0.56 oote 3|2 He (&} CO»o He MOLE RATIO —= 0.52 0.012 MOLE RATIO 0.48 0.008 0.44 0.004 0.40 0 10 20 30 40 50 60 70 TIME (hr) Fig. 6.7. Apparent Reoction of CF4 with Oxide Contominants in MSRE Fuel Salt Contained in Nickel at 600°C. 124 Experiments to determine the solubility of CF4 in the molten fuel employed apparatus and techniques previously described in connection with determinations of the solubility of noble gases in molten fluorides.*?t Two determinations of the CF4 solubility were made by saturating the fuel mixture at 600°C with CF4 at 1.3 atm for 6 hr. A portion of saturated melt was transferred to an isolated section of the apparatus and stripped with a known quantity of dry helium. Spectrographic analyses of the strip gas failed to indicate the presence of CF4; however, a small measurable guantity of COz was found. The estimated solubility of CF4 in the fluo- ride mixture was assumed to be no greater than the value corresponding to the COz found in the strip gas, that is, 1 x 107® moles of CF4 per cubic centimeter of melt per atmosphere. Additional determinations of the CF4 solubllity at higher tempera- tures have not been completed. The stripping section of the solubility apparatus has been thoroughly hydrofluorinated in an attempt to eliminate the postulated reaction of CFa with oxide contaminants. 6.5 CHEMICAL ASPECTS OF MSRE SAFETY 6.5.1 Physical Effects of Mixing Molten Fuel and Water The overriding physical effect of suddenly mixing molten MSRE fuel and water would come from the steam pressure generated in the cell. A study made for the Aircraft Reactor Test system included observations of the gross effect of dumping or injecting a sizeable amount (~400 1b) of salt composition No. 12 (NaF-LiF-KF, 11.5-46.5-42.0 mole %) into several hundred gallons of water contained in an open vessel.’® These tests did not include pressure measurements or analysis of gaseous products. The uncertainties in calculating the rate of loss of heat to the MSRE cell structure led to a study of additional methods of relieving the pressure in case of an accident. A small-scale laboratory study of the short-term effects of mixing fuel (LiF-BeFo~ZrF4-Thf'y-UF,, T70-23-5-1-1 mole %) and water was undertaken to provide information regarding possible reaction products and the physical nature of the salt after its interaction with water. The apparatus shown in Fig. 6.8 was constructed to permit the obser- vation of the effects of injecting a small amount of molten salt into water. The hazard involved was considered small since 2 g of salt (0.5 cal g~* °C™*, assumed) at 600°C would convert only 1 g of water to steam. The salt (5 to 6 g) was loaded into the upper chamber (a 1/2-in.-diam x 3-in.-long, flanged, Hastelloy tube). A small furnace mounted around the chamber was used to heat the salt to the desired temperature. Helium pressure was then applied to force the molten salt into water (2 to 5 cc) contained in an inverted nickel cone in the lower Pyrex-pipe chamber, which contained an argon atmosphere. By means of a high-speed (3 in./min) potentiometric recorder and a switching arrangement, pressure bulldup and temperature variations at selected points could be followed. 125 UNCLASSIFIED PHOTO 56520 Fig. 6.8, Salt Jet Apparatus. Data from a recent experiment are shown in Fig. 6.9. The pressure in the water chamber rose some 10 1b above atmospheric and remained there after cooling. Therefore, the increase must have been due to the slight 126 UNGCLASSIFIED ORNL-LR-DWG 68601 122 110 «F THERMOCOUPLE NO. 3 BOTTOM TEMPERATURE ; 86 ) 74 62 i 1 TEMPERATURE (°C 50 ~THERMOCOUPLE NO.1 WATER - CHAMBER WALL TEMPERATURE 38 : o6 o] THERMOCOUPLE NO. 2 VAPOR TEMPERATURE 0 2 4 6 8 10 12 TIME FOLLOWING INTRODUCTION OF MOLTEN SALT TO WATER (min) Fig. 6.9. Temperature Variations in Salt Jet Apparatus Following Introduc- tion of Salt at 650°C (5.51 g of Salt into 2 cc Water). excess of helium used to effect salt transfer. The salt transfer occurred in some 6 to 8 sec and that transferred into the water had an exploded (popcorn) appearance but was unaltered in color. Salt flowing into the nickel cone after the water had all been vaporized appeared solid and dense, with a grayish film. In all experiments, finely divided droplets of salt were found sprayed on the chamber walls and top. No marked pres- sure increases were encountered in eight trials. Petrographic examination of salt from a previous, similar experiment showed a poorly crystallized material containing eutectic-~type growths and exhibiting refractive indices typical of the fuel. Oxides were not detected and if present must have been <1 p in size. The x-ray dif- fraction pattern also indicated very poorly crystallized material--no 7r0ps, 7r0Fo, ThOo, or UOo. Following one experiment, the gas contained in the water chamber after introduction of the salt was swept gently through standard NaOH. Titration of the caustic indicated that some L millimoles of HF or some substance which reacted with NaOH was formed. This represents about 2% of the calculated yield for complete reaction; in none of the experiments was any etching of the glass reaction chamber noted. Future experiments will include much larger volumes of water in order to conserve the heat lost in the finely divided salt seen to spray through- out the reaction charber, but the prospects for obtaining marked pressure rise in this apparatus are poor, presumably because of the great diffi- culty in obtaining sufficiently rapid heat transfer from salt to water. 127 Qualitative and quantitative analysis of the gaseous products is also planned. 6.5.2 Solubility of Fuel-Salt Components in Water Following a hypothetical accident in which molten fuel and water were mixed in the containment cell it might be necessary to wait for weeks or months before the radiation levels decrease sufficiently to permit the approach of people for examination, decontamination, or repair purposes. During this time even a slow rate of uranium leaching from the fuel could, if the equilibrium solubility were sufficiently high, lead to a criti- cality accident. The solubility of uranium tetrafluoride in water is indicated to be low at 25°C (1.6 x louéAl).la This uranium concentration is below that necessary for criticality, but several factors in a real situation may combine to give a higher concentration. The action of the oxygen from alr or the action of peroxide generated by beta-gamma radiation in water could oxidize the uranium to the very soluble hexavalent state. The temperature may well be considerably higher than 25°C for a long time. The acidity of the agueous phase in contact with the frozen salt may be such as to increase the uranium solubility. Anions other than fluoride may be present. Cations other than Uttt are certain to be present. A sol or slurry of uranium-bearing solid could be produced, and it would create the effect of a solution. For these reasons measurements of the solu- bility of MSRE fuel salt in water at a variety of temperatures have been made available for evaluating the probable course of events in the weeks following the postulated accident. To maximize the solubility and the rate of solution, a quantity of nominal MSRE fuel salt (from the Fluoride Production Facility) was ground to powder in a glove box so that it would present a large surface area to the aqueous phase. Portions of it were added to flasks containing water, and the resulting mixtures were stirred at controlled temperatures; samples of the supernatant solution were taken at appropriate intervals for chemical analysis.l4 Figure 6.10 shows that, at 25°C, the uranium concentration rose, within one day, to approximately 0.002 m and then slowly climbed to 0.00275 m in another week; the other components simi- larly rose quickly to values which thereafter changed only slightly. Figure 6.11 shows the solution concentrations after one-day exposures at different temperatures, together with one value for the uranium concen- tration obtained previously.l5 The sequence of temperatures is indicated by the arrows in the figure; since the solution concentrations were not reduced when the temperature was lowered, the equilibrium expected for simple solubility behavior was not re-established on cooling. The tests reported in Figs. 6.10 and 6.11 were made without particular attempts to protect the mixtures from contact with the air. Although the flasks were stoppered during equilibration at a fixed temperature, they were opened to the alr vwhenever samples were taken. Consequently some oxidation of uranium to the hexavalent state probably occurred. 128 UNCLASSIFIED ORNL-LR-DWG 66935 02 ; — — Be ‘ : 1 0.1 T P . 0 S S, L e | 1 1 i 0.08 oy e i - ‘ | ! 0.06 ¢ ; S 004 L — | 2 0.006 ; -~ 3 ; | s f o, 0004 s S ‘ S 2 0.002 ; g ol . , I N L E 0.45 - T , T e — Li ‘ 5 i‘___...., e ) eI Sa $ | . . . . 035~ * ‘ ’ ' T Fig. 6.10. Solution Concentrations of U, Li, Th, : I | : : oosd 1 | , P P | Zr, and Be Found upon Mixing MSRE-Fuel Solid ‘ : T - 1 | with HZO; Concentration vs Time at 25°C, 0.004 L{ . ‘ « , , o ® ® ® i B 0002 1 Ll ¢ : : | ; { 2 3 4 5 6 7 8 9 10 0 TIME (days) UNCLASSIFIED ORNL-LR-DWG 66936 0.3 1 0.2 _ [OR ] 0.20 0.16 02 0.08 c | i 2 0.04 ‘ L ‘ | 3 [ 5 008 — N ‘ ] e W o MQWW & T i S . B | | o 004**'%?-—”%"“’"_'”#——” | . v ! T‘ g i } i g O0— — : | g 006 a— ¢ / J 0.05 ! . /"‘($ ’ ‘ . ! M"”‘—M . ‘ 0.04 . % - - . i : . [ e e \ i ‘v FFige 6.11. Solution Concentrations of U, Li, 0.03 T . APPLE'S DATA J Th, Zr, and Be Found upon Mixing MSRE-Fuel 0012 e e . T . | - + ® ¢ ! > Solid with H2O; Concentration vs Temperature, ‘ u ‘ ! ‘/‘ j : i ) ) | APPLE'S DATA [ 20 30 40 50 60 70 80 30 100 TEMPERATURE (°C) 129 Because of the appreciable uranium concentrations achieved in these tests, neutron poisons will be provided in any water which might come in contact with fuel. Since this provision has been established, solubility studies in the presence of peroxide or radiation have been postponed. 6.5.3 Solubility of MSRE Coolant in Water In experiments similar to those reported above for the fuel salt, the solubility of the coolant salt was ourveybd A gquantity of 2LiF.BeFjs conmposition (containing approximately 2 wt p exces LiF) was obtained from the Fluoride Production Facility and ground to powder in a glove box for use in the solubility experiments. TFigure 6.12 shows the analytical results at 25°C, and Fig. 6.13 the results of experiments over a range of temperatures up to 90°C. To provide information relevant to purification processes for fluorides as well as to the behavior of coolant in contact with water, a more thorough investigation of phase equilibria in the system LiF-BeFo-H»-0 has been initiated. UNCLASSIFIED ORNL-L.R-DWG 66933 &z [FLUORIDE | , . | 25 e té (—'j o (7} ‘ 8% ol | J § Q06 T I ] 3 3 Fig. 6.12. Solution Concentrations of Li, Be, §§ 0.4 ,«f/ 5 B[ERL\:l;L:LLm I AAAAAAAAA A | and F Found upon Mixing Li2BeF4 Solid with %éoz . ‘ x HzO; Concentration vs Time at 25°C. ég \ l ° 0 5 7 10 15 20 25 TIME (days) UNCLASSIFIED ORNL-LR-DWG 66934 30 [ % FLUORIDE ‘ ‘ ‘ [ - PERFECT, PROC. PENN. ACAD. 5C/. o ‘ ¢6 54 19"3?) ‘ = T 8 2.0 A ,,,T, _______________ e ~- L 5 j T “‘%"* ~~ A o A - o 4 A e 4 8 iol “i"'—”"lf'—m‘ T T AT — ] < 4 $ @ A L 2 1 \ o 1 08 T Be AND Li ’ ‘ = i Q I:j— A 3 4 -4_--— R — 2 § g - s | ] Fig. 6.13. Solution Concentrations of Li, Be, o e 1 S & Be (PERFECT’S DATA) and F Found upon Mixing Li,~)8e=:F4 Solid with - o Li | . Jo2l—aA —g | H,0; Concentration vs Temperature, 2 * 1 O e 20 30 40 50 60 70 80 20 100 TEMPERATURE (°C) 130 6.5.4 Partial Freezing of MSRE Fuel A hazard could conceivably be produced upon partial freezing of the reactor fuel while in the drain tank. If the freezing occurs at a suf- ficiently slow rate, the first solids deposited would not contain uranium, and the remaining solution, enriched in uranium, could cause the reactor to be critical when only partly full. To permit tentative calculations of the potential, though unlikely, condition, the following hypothetical crystallization path of the MSRE fuel}6 and estimations of density of the molten liquid remaining along this crystallization path, have been proposed. 1. The compound 2LiF-.BeF, appears as a secondary phase after 20 wt % of the fuel has frozen as 6LiF.BeFo+ZrFs, producing a liquid of compo- sition LiF-BeFp-ZrF4-ThFse-UFs (68.9-26.2-3,3-1.222-0.366 mole %), density 2.246 g/cc at the first appearance of the secondary phase. 2, The compounds 2LiF-2 mesh, 6.6/ between 2 and U mesh, 0.69% be- tween 4 and 8 mesh, 1.23% between 8 and 25 mesh, 0.92% between 25 and 50 mesh, 1.31% between 50 and 100 mesh, and 0.84% <100 mesh (all weight percentages). REFERENCE 1. NASA Memorandum 1-27-59E, February 1959, p 8. 137 OAK RIDGE NATIONAL LABORATORY MOLTEN SALT REACTOR PROGRAM FEBRUARY 4 1962 R. B. BRIGGS, DIRECTOR oA MSRE DESIGN : MSRE OPERATIONS E.S. BETTIS R ! s E.BEALL® ! R.H.GUYMON* ENGINEERING INSTRUMENTATION AND CONTROLS P.H. HARLEY R R. L. MOORE 18C H. R. PAYNE R J. R. BROWN 18C A. N, SMITH R G. H. BURGER 13C W. C. ULRICH R 0. W. BURKE* 13C J. H., WESTSIK* R T. M. CATE 13C L. V. WILSON R D. G. DAVIS 18C P. G. HERNDON 13C B. J. JONES 18C J. W. KREWSON 18C C. E. STEVENSON 18C B. SQUIRES B&R J. R. TALLACKSON 18C MECHANICAL DESIGN R. WEIS cTS W. M. BROWN R J. 7. DICKIE R J. K. DUNCAN R W. C. GEORGE R C. C. HURTT R R. H. JONES R ANALYSIS A. R.KERR R J. 0. NICHOLSON R J. R, ENGEL R L. E. PENTON R P.N. HAUBENREICH R C. A. ROBERTS R C. W.NESTOR R W. ROBINSON R F.L.ROUSER R W. G. STERLING R 1. ALWATTS R FUEL PROCESSING R. B. LINDAUER cT SITE PREPARATION G. B. CHILD E&M W, L. DARIEUX E&M N. E. DUNWOODY E&M T. E. NORTHUP E&M G. W. RENFRO E&M W. E. SALLEE E&M T. A. WATSON E&M MSRE PROCUREMENT AND CONSTRUCTION COMPONENT DEVELOPMENT I SPIEWAK* D. SCOTT, JR. £ FLOW MODELS R. J. KEDL™ B. J. YOUNG ) W. B. MCDONALD R PROCUREMENT AND FABRICATION C. K. MCGLOTHL AN R J. M. TEAGUE R CONSTRUCTION N. E. DUNWOODY* E&M J. P. JACKSON E&M B. H. WEBSTER R L.P. PUGH R SCHEDULING J. P. JARVIS E&M B. N. ROBARDS E&M J. W. FREELS E&M REMOTE MAINTENANCE . C. HISE™ . BLUMBERG P.HOLZ* R. F. BENSON 4. R. SHUGART oo m WD AD COMPONENTS SCOTT, JR. RICHARDSON . SMITH WEBER . JONES I»x0 onz AWV OAD LOOPS . CROWLEY . GALLAHER R. E. CARNES W. H. DUCKYORTH - w AN DA PUMP DEVELOPMENT A. G. GRINDELL C. H. GABBARD P.G. SMITH T METALLURGY ABOGADA . COOK . DEVAN™ . DONNELLY . GILLILAND* H H G G J. LEONARD™ K. T H G J M ROCHE* . VENARD* . WODTKE . D. BRADY* . L. GRIFFITH . A. REDDEN TTFTITIZFETFTETIXTX X 0 BURNS & ROE CHEMICAL TECHNOLOGY DIVISION CONTINENTAL TECHNICAL SERVICES DI ENGINEERING AND MECHANICAL D'VISION INSTRUMENTATION AND CONTROLS DIVISION METALS AND CERAMICS DIVISION REACTOR CHEMISTRY DIVISION RECTOR'S DIVISION REACTOR DIVISION PART TIME ON MSRP REACTOR CHEMISTRY W. R. GRIMES* RC F. F. BLANKENSHIP RC S. CANTOR RC M. J. KELLY* RC W. L. MARSHALL* RC H. F. MCDUFFIE* RC J. H. SHAFFER RC R. E. THOMA* RC D. R. CUNEO RC F. A. DOSS RC J. E. EORGAN RC G. M. HEBERT RC S. S. KIRSLIS RC J. E. SAVOLAINEN RC R. SLUSHER* RC C. F. WEAVER* RC W. K. R. FINNELL RC B. F. HITCH RC W. JENNINGS RC R. G. ROSS RC W. P. TEICHERT RC RADIATION TESTING C. A, BRANDON R J. A, CONLIN R 139 ORNL-~3282 UC-80 — Reactor Technology TID-U500 (17th ed., Rev.) INTERNAL DISTRIBUTION ~ 1. G. M. Adamson W7, P. R. Kasten 2. L. G. Alexander 48, R. J. Kedl 3. S. E. Beall L9, G. W. Keilholtz L. C. E. Bettis 50. M. T. Kelley 5. E. S, Bettis 51. B. W. Kinyon 6. D. 3. Billington 52. R. W. Knight T. F. F. Blankenship 53. dJd. A, Lane 8. E. P. Blizard 5L, C. E. Larson 9. A. L. Boch 55. T. A. Lincoln 10. S. E. Bolt 56. S. C. Lind 11. C. J. Borkowski 57T. R. B. Lindsuer 12, G. E. Boyd 58. R. S. Livingston 13. E. J. Breeding 59. M. I. Lundin 14. R. B. Briggs 60. H. G. MacPherson 15, F. R. Bruce 61, W. D. Manly 16. 0. W. Burke 62, E. R. Mann 17. D. O. Campbell 63. W. B. McDonald 18. W. G. Cobb 6. C. K. McGlothlan 19. J. A. Conlin 65. E. C. Miller 20. W. H. Cook 66. R. L. Moore 2l. G. A. Cristy 67. K. Z. Morgan 22, J. L, Crowley 68. J. C. Moyers 23. F. L. Culler 69. J. P. Murray (K-25) 2, J. H. DeVan 70. M. L, Nelson 25. R. G. Donnelly TL. C. W. Nestor 26, D. A. Douglas 2. T. E. Northup 27, J. L. English T3. W. R. Osborn 23. E. P. Epler . L. F. Parsly 29, W. K. Ergen ™. P. Patriarca 30. A. P. Fraas 76. H. R. Payne 31. dJ. H. Frye, Jr. Tf. D. Phillips 32. C. H. Gabbard 8. W. B. Pike 33. W. R. Gall 9. M. Richardson 34k, R. B. Gallaher 80. R. C. Robertson 35. W. R. Grimes 81. T. K. Roche 36, A. G. Grindell 82. H. W. Savage 37. C. S. Harrill 83. A. W. Savolainen 38, M. R. Hi1l 8k, D. Scott 39. E. C. Hise 85. H. E. Seagren L0, H. W. Hoffman 86. E. D. Shipley 41, P. P. Holz 87. 0. Sisman 42, A. Hollaender 88. M. J. Skinner 43. A. S. Householder 89. G. M. Slaughter Wy, 1. N, Howell 90. A. N. Smith 5. W. H. Jordan 91. P. G. Smith L6, R. G. Jordan 92, A. H. Snell 93. ol 9%5. L aH « & o - 97. 98. 99. 100. 101. 102, 103. * O 9 e . 163-16k., 165, 166. 167. 168. 169-771. Spiewak D. Susano A. Swartout Taboada R. Tallackson H. Taylor E. Thoma B. Trauvuger C. Ulrich C. Watkin M. Weinberg D. F. Cope, AEC, ORO 140 10k, 105. 106. 107. 108-109. 110-113. 114-158, 159. 160-162, J. F. Kaufmann, AEC, Washington R. W. McNamee, Manager, Research Administration, UCC, New York F. P. Self, AEC, ORO J. H. Westsik L. V. Wilson C. H. Wodtke Biology Library Reactor Division Library ORNL -~ Y-12 Technical Library, Document Reference Section Laboratory Records Department Laboratory Records, ORNL R.C. Central Research Library EXTERNAL DISTRIBUTTION Division of Research and Development, AEC, ORO Given distribution as shown in TID-4500 (17th ed., Rev.) under Reactor Technology category (75 copies - OTS)