= 1] JJ\JI“M\I‘M‘H‘l‘h“‘l“fl‘h\‘ l; ORNL-3529 UC-80 — Reactor Technology TID-4500 (24th ed.) MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT FOR PERIOD ENDING JULY 31, 1963 OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY .DO NOT wTRANSFER TO ANOTHER P SON. 2 Printed in USA. Price: $2.75 Availble from the Office of Technical Services U. S. Department of Commerce Washington 25, D. C. LEGAL NOTICE This report was prepored as an account of Government sponsored work. Neither the United States, nor the Commission, nor any person acting on behalf of the Commissio A. Makes any worranty or presentation, expressed o impliad, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of eny information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or B. Assumes any liab es with respect to the use of, or for damages any information, apparatus, method, or process disclosed in this report. As used in the cbove, “‘person acting on behalf of the Commission’" includes any employes or contractor of the Commission, or employs sulting from the use of of such contractor, to the extent that such employee or contractor of the Commission, or employee of such contractor prepares, ates, or provides access to, any information pursuant to his employment or contract with the Commission, or hi employment with such contractor. ORNL-3529 Contract No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM SEMIANNUAL PROGRESS REPORT For Period Ending July 31, 1963 R. B. Briggs, Program Director Date Issued DEC 30 1963 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee cperated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION I | HRTR 3 4456 013413y H#J e iii SUMMARY Part 1. MSRE Design, Engineering Analyéis, and Component Development 1. MSRE Design, Procurement, and Construction Major effort on the mechanical and process design of the MSRE con- tinued to be on the revision of drawings to incorporate recent results of development work and to facilitate the fabrication and construction work. An overflow tank for installation beneath the fuel pump bowl in the pump furnace is being designed to replace the overflow line that was to have been installed between the fuel pump and the fuel drain line. A fuel pump with a larger volume for expansion of fuel in the pump bowl is being designed. This pump will replace the pump that is being in- stalled in the reactor. A strainer is being designed for installation in the reactor vessel to prevent pieces of graphite that are larger than 1/8 in, in diameter from being circulated with the fuel. The layout of the instrumentation and control system remains essen- tially the same as previously reported. One new panel was added, four panels were relocated, and the thermocouple routing was revised. Loca- tions for personnel radiation monitors were established. Instrument application diagrams and tabulations were revised to incorporate recent design changes. Criteria for control and safety circuits were extensively reviewed, and some control-system block diagrams were approved. Schematic dia- grams of freeze valve circuits, instrument air compressors, and lube- 0ll pumps were approved. Design of 40 instrument panel sections and fabrication of 28 panels are complete. Design of interconnecting wiring for the annunciator system and for Foxboro ECI instruments was completed. Wiring diagrams for 16 control- circuit terminal boxes were completed. Design of thermocouple installation and interconnections is nearing completion. Process and personnel-radiation-monitoring systems designs are nearing completion. Additional analog studies of reactor fill and drain transients were made. Results of these studies were used to determine the size and loca- tion of capillary restrictors in the helium supply and bypass lines. iv Fabrication of the reactor vessel is nearing completion, but final assembly must await the delivery of the moderator graphite in October. The radiator and fuel and coolant pumps are nearing completion. The heat exchanger and the fuel and coolant drain tanks were completed and delivered to the reactor site. A contract was awarded for the fabrication of removable heaters for the fuel and drain tank systems. Other electrical procurement is 0% complete. Except for those items required by recent revisions, preparation of specifications and initiation of procurement of process instrument com- ponents was completed. Most of these components are now on hand, and delivery of the remainder is expected within the next three months. Proposals received from four vendors for a computer data-logging system are presently being evaluated. Installation is progressing in all areas. Installation of reactor equipment and auxiliaries in the reactor cell, drain tank cell, and coolant cell is approximately MO% complete. Auxiliary equipment outside these cells is being installed by a contractor, and his work is approxi- mately 10% complete. Work was begun on the installation of instrumentation and control equipment in all areas. 2. Component Development The test on the prototype cooling bayonets for removing after-heat from the MSRE drain tank was terminated after 2600 temperature cycles (1300 - 212OF) because of a leak in the steam system. The test was equiva- lent to several years of operation in the MSRE. Severe thermal fatigue cracking was found in both Inconel tubes of the bayonet, particularly near the weldments for the centering spacer bars. A prototype bayonet of an improved design was constructed of INOR-8 for further testing. A prototype drain-tank heater was operated for 2876 hr at l2OOOF ~average temperature without mechanical difficulty. The electrical circuitry that connects the individual heater elements was changed to permit separate control of the upper and lower zones of the heater and thus obtain a more even temperature distribution. The all-metal prototype hgaters for 5-in. pipe were operated satisfac- torily for six months at 1400°F. A small loss in insulating properties was attributed to damage of one of the heaters early in the test. A method of calibrating the control rod remote position indicator with the actual position of the bottom of the rod at one point was demonstrated. The prototype control rod drive mechanism was received and is being in- stalled for testing. ' i Measurements in the test loop of the helium purification system indi- cated complete removal of oxygen, with inlet concentrations of up to 225 ppm. However, a breakthrough occurred at 100 ppm after the titanium bed reached 15% of saturation while operating at lOOO . PFurther tests will be run with the bed at 12000F An electrolytic oxygen analyzer was tested and found to agree within 10% with the results of mass spectrometer analysis. Fabrication of the sampler-enricher system mockup and its installation into the engineering test loop was completed, and testing was started. A sample was successfully isolated and removed from the pump bowl and trans- ferred to the Analytical Chemistry ILaboratory for analysis. The entire pro- cedure required 3 hr, which is acceptable. The flanged disconnects for use on part of the reactor sample transfer tube were successfully assembled. The motor-operated valves that isolate the sampler from the fuel pump bowl were received, and tests indicated that the leak rates through the valves were acceptable. Final design has started on the fuel loop sampler~enricher system. The Engineering Test Loop (ETL) was placed in operation with fuel salt after two oxide additions and removals by HF treatment had been performed on the flush salt and after the fuel salt had been treated with HF and H, in the drain tank. Seventy-two percent of the oxygen of the first addition and eighty percent of the oxygen of the second addition was collected as water in the cold traps of the off-gas line during the treatment of the flush salt. The equivalent of 625 ppm oxygen was removed from the fuel salt during 168 hr of treatment with HF and H,. FExamination of the kinetics of the water stripping indicated that continuous treatment with HF was not necessary since the time-consuming step in the process was the resolution of the precipitated oxide to make it available to the HF. Agitation by bubbling with H, and helium was sufficient to redissolve the precipitate after the concentratlon was reduced below saturation by short periods of HF treatment. After 2000 hr of operation with fuel salt at lEOOOF and with salt in the graphite access joint molten, the loop was drained, cooled, and opened at the access joint to remove samples of solids that were deposited at the liquid-gas interface. The solids, while apparently free of U0,, contained corrosion products and crystals of the salt that was formed by selective freezing of the lithium, beryllium, and zirconium fluoride phases. The apparent corrosion rate was initially higher than that during previous opera- tion with the flush salt; however, by the end of the period the rate was back to normal. Study of the xenon distribution throughout the MSRE system was continued, and preliminary results of an experiment to determine the xenon-removal rate at the pump bowl indicated that at reduced salt flow-rates the stripping efficiency was as low as 17%. Fabrication and testing of the tools for use in maintenance of the reactor system continued, with special emphasis on a freeze flange which had overhead interferences. Testing of an improved version of a general- purpose light for remote illumination was started. vi Data from tests with the MSRE prototype pump showed that the diffusion of radioactive gas up the shaft annulus to the region of the shaft seal could be made acceptably low by flowing helium down the annulus and through the pump bowl at rates of 100 to 1000 liters/day, respectively. TFlow rates of 3300 liters/day in the annulus and 4600 liters/day through the pump bowl are available in the MSRE. Preliminary measurements indicated that the circulating salt in the pump loop contained 1 to 2 vol % of undissolved gas, presumably entrained by the stream of salt that is circulated through the xenon-removal appa- ratus in the pump bowl, The pump test was halted after 4700 hr of operation to modify the test facility. It was then resumed to observe the behavior of the flex- ible mount for supporting the fuel pump, the bouyancy liquid-level indi- cator, and the MSRE disconnect flange in the pump tank off-gas line, and to continue Investigations with the device for measuring the concentration of undissolved gas in the flowing salt. Endurance tests were continued with the PKP test pump, the test pump having one molten-salt-lubricated journal bearing, and the test pump for the MSRE lubrication stand. The design of a fuel pump to accommodate a larger volume of thermally expanded fuel was initiated. ' Six additional months of satisfactory operation were accumulated on a prototype model of a two-level single-point probe for indicating the liquid level in molten-salt systems. Performance is unchanged from that reported previously. The design of the MSRE probe assembly was modified to incorporate a secondary containment barrier. Two float-type level transmitters have now operated satisfactorily for 18 months in molten salt at 1200°F. A third transmitter was installed on the MSRE prototype pump test, and initial performance was satisfactory. A temperature scanner system being developed for use on the MSRE was demonstrated on the level test and on the ETL. This system has been accepted for use on the MSRE, and the design of a five-channel system for this application has been completed. Three alternative methods for using closed-circuit television for viewing during remote maintenance operations are being investigated. Eight MSRE prototype thermocouples continue to perform satisfactorily after 5000 hr of operation on the ETL. Ten similar thermocouples accumu- lated 5300 hr of satisfactory operation on the MSRE prototype pump test loop. Drift of six MSRE prototype thermocouples operating at 1200 to 1250°F_air remained at less than #2°F after 18 months of operation. Thermal shock testing of ten thermocouples was completed. Two ther- mocouples were still functioning after 2630 severe thermal cycles. vii Techniques and'prOCQdures were developed for use in installation and testing of in-cell thermocouple extension cable seals. A satisfactory material was found for sealing the ends of mineral- insulated sheathed thermocouples that will be located in the reactor and drain cells, and technigues and procedures were developed for mass pro- ducing these seals. Development of a high-temperature NaK-filled differential pressure transmitter was continued. Further investigations were made of the plugged dip tube in the developmental MSRE bubbler system. 3. MSRE Reactor Analysis The nuclear characteristics of the reactor were calculated for each of three fuel-salt compositions. In each case critical concentration of uranium, reactivity coefficients, and flux distributions were calculated. Predicted control rod worth ranges from 5.6 to 7.69% Sk/k, depending on the fuel composition. The rods are required to shim only about 1 to 3% 6k/k, leaving ample shutdown margin. There is a large uncertainty in shim requirements due to lack of information on the properties determining the xenon removal from the core. Alpha particles from the uranium will interact with beryllium and fluorine of the fuel salt to produce about 4 X 10° neutrons/sec in the core. The internal photoneutron source will be 107 to 10° neutrons/sec after power operation. A review of the biological shield design showed that after the addi- tion of concrete-block shielding at some points, the biological shielding will be adequate. Part 2. Materials Studies 4. Metallurgy The tube-to-tubesheet joints for the MSRE heat exchanger were success- fully brazed using a method previously developed for this purpose. Visual inspection and hydrostatic and ultrasonic testing were performed on this heat exchanger, with no evidence of leakage. The thermal-fatigue characteristics of INOR-8 were determined from the conventional Coffin-type test, and a low-cycle-fatigue equation was found to describe the plastic-strain - fatigue relation. Thermal-fatigue data showed good agreement with those previously reported for isothermal fatigue. ' viii A program was initiated to examine the postirradiation tensile prop- erties of INOR-8. Subsize tensile specimens have been irradiated to 7 x 10%9 nvt (5L Mev) in the temperature range 1100 to 1400°F and will be tested for strength and ductility. MSRE moderator bars are being tested to establish the physical and mechanical properties of the material useful to the MSRE. The chemical composition and oxygen content of these bars met specification requirements. Test specimens from these bars were rapidly heated to 1800°F and cooled, with no deleterious effects on the graphite and no propagation of salt- filled cracks. An irradiation program for this material has been started. The use of molten salt to remove oxygen from graphite was demonstrated under conditions similar to those proposed for the MSRE. Mixtures of Gd,03 and Al,03 were successfully cold pressed and sintered into cylinders to 959 of calculated densities by working with prereacted powders. Distortion of the cylinder was minimized by the use of a holding fixture of the same composition. 5. Radiation Effects Continued examinations of parts from earlier irradiations of MSRE fuel and graphite showed that xenon could be recovered from graphite cores even after several months' storage and that, although there was virtually no penetration of the graphite by fuel, uranium and lithium in small amounts were found in the graphite, probably as a result of oxi- dation~reduction processes associated with the release and recombination of F» during shutdown. Although the yield of CF, from fissioning fuel containing submerged graphite was too small to be measured with certainty, a possible interpre- tation was that CF, could be expected in the MSRE at 109 of the rate of production of long-lived xXenon. Such a rate would be favorably low and would constitute no problem in MSRE operation. The rate of recombination of Fp; with fuel from which F> had evolved as a result of radiation damage from decay energy was found to depend on the pressure of Fz, but to be even more strongly dependent on both the condition of the fuel and the temperature. The behavior of fuel under gamma irradiation resembled that of in-pile capsules in response to decay energy, whereas electron irradiation gave somewhat lower yields of Fp, and x rays even lower. The x-ray experiments showed a noticeable difference in the response of individual components of the fuel. 6. Chemistry ‘The fuels for operation in the MSRE correspond closely in LiF and BeFy proportions to compositions found along the 450°C isotherm that is associated with the LiF-rich region of the LiF-BeF,-ZrF, ternary system. This prov%des the maximum LiF content consistent with a melting point below 450°C, thereby leading to the most favorable physical properties for a given content of quadrivalent fluorides. As a safeguard against 1] ix possible alterations in chemical oxidation-reduction potential of the fuel, normally buffered by the small fraction of UFs; formed as UF, comes to virtual equilibrium with the containing metal, the UF, content of the initial fuel has been set at ~l mole ¢ rather than the ~0.1l5 mole ¢ re- quired for clean criticality. Since this is accomplished by decreased enrichment of U235F4, there is a diminished heating effect in the un- likely event that uranium is deposited in the core. The inclusion of 5 mole ¢ of ZrF, as a scavenger for oxide ion led to the composition con- taining LiF-BeF,-ZrF,-U?35F,-U?28F, in the proportions of 65-29.17-5- 0.29-0.54 as representative of the fuel for initial criticality. This fuel will be blended from prepurified concentrated fuel, LiF-UF, (73-27 mole %), and solvent, LiF-BeF,-ZrF,; (64.43-30.44-5.14 mole 4), the latter of which will have been used in the prenuclear operation. A region of liquid-liquid immiscibility was found in the LiF-BeF;- ZrF, system, in composition ranges that melt characteristically to glassy liquids; containing BeF,, the miscibility gap includes ZrF, concentrations from 25 to 70 mole ¢ and LiF concentrations from 5 to 20 mole ¢. The exact boundaries, though far removed from regions of interest in reactors, are still under study in gquenching experiments. The structure of a recently discovered fluoride of xenon, XeF,-XeFy, has been determined by x-ray diffraction. Conditions favorable for the existence of this addition compound may have occurred at some stage during the decay period following in-pile tests with sealed capsules that evolved xenon and subsequently F». Oxide equilibria in flush-salt - fuel-salt mixtures have provided additional evidence that there is no appreciable solid solution formation of U0, with ZrO, when excess oxide ions are scavenged from melts as Zr0O,. Indications in some ranges of a dependence on both concentration and temperature were found for the limiting ZrF4/UF4 concentration ratio required for scavenging oxide as ZrOs, but the better than 5/1 ratio design value previously specified for the MSRE contains an ample margin as far as operating fuel is concerned. In connection with sulfur removal during fuel purification proce- dures, several bench-scale studies were carried out to compare the po- tential usefulness of alternate methods and to learn more of the detailed mechanisms of sulfur behavior with respect to both removal and corrosion. Relatively little change in current practice involving H-HF mixtures was indicated except that an associated use of elemental beryllium as a reducing agent, which should also be effective for other impurities as well, showed some promise of improved overall efficiency. Among the factors that influence the interfacial behavior of MSRE fluoride melts with respect to graphite, trace amounts of H;0 of 10 ppm or less in the atmosphere over sessile drops caused marked spreading or superficial wetting of graphite but no penetration. In comparison with H>0, 0O, caused relatively little alteration of the normal nonwetting behavior; neither was saturation with dissolved oxides in the fluoride melt of any consequence in changing the nonwetting behavior. Recent measurements of viscosity as a function of temperature gave viscosities of 8 centipoises for the MSRE fuel at 650°C and 10 centi- poises for the coolant at 570°C; these temperatures are averages for normal reactor operation. Rates of removal of reducible impurities from MSRE melts were stud- ied under various conditions; for treatments with H,, both higher flow rates (10 liters of H, per minute) and higher temperatures had a signif- icant effect in shortening the time required for purification. Recent measurements on strongly reduced fuels gave unexpectedly low values for the solubility of UFs;. The disagreement between these and previous measurements and other puzzling data that were obtained suggest that defined equilibrium conditions were not achieved. A pyrolytic method was developed for determining the fluoride con- tent of the fuel salts and was tested in the hot-cell mock-up. The salt is reacted with moist oxygen at 1000°C in the presence of U30g in nickel equipment, and the fluoride that is evolved is trapped in sodium hydrox- ide. A relative standard deviation of 19, was obtained when nonradio- active salts were analyzed by use of remote techniques. An amperometric method was developed for titrating Cr(VI) with ferrous sulfate to analyze for chromium in fuel salts. The precision of bench-type analysis is l%. Testing of the method in the hot-cell mock-up has begun with nonradiocactive samples. A radio-frequency concentrator was incorporated in the ignition chamber of the modified Ieco analyzer that is being used to study the determination of oxygen in fuel salts by inert-gas fusion procedures. Samples of uranium and zirconium oxides in graphite capsules were re- duced to carbides within a 5-min ignition period in the initial tests. The statistical evaluation of the spectrophotometric method which uses dimethylglyoxime and potassium persulfate oxidation for determi- nation of nickel was completed in the hot-cell mock-up. The relative standard deviation of the method was 2.89. Prototypes of some of the equipment for handling and processing fuel samples from the reactor were also tested in the hot-cell mock-up, and some improvements were made. 7. Fuel Processing The fuel-processing system design was reviewed, and minor revisions are being made to the drawings. Safety of the system was reevaluated, and an activated-charcoal trap was added to improve iodine and tellurium removal from the off-gas. xi SUM]VLARY ----- S 2 6 8 6 0 B S 0 E 8 BT S0P T RS BT A AN SIS AR SRS iii Part 1. MSRE DESIGN, ENGINEERING ANALYSIS, AND COMPONENT DEVELOPMENT 1. MSRE DESIGN, PROCUREMENT, AND CONSTRUCTION «eveeseneanss ceasss Status of Design ..... Tt e eesessatsests et e acat et et ean ot anns Mechanical and Process DeSifgll seveeeocnossrssssssessrsocsnsas Instrumentation and Control Design ........ Cerrestereanesns Status of Fabrication of Major Reactor Components ...vieeeeess Reactor Vessel and Control Rod Thimble Assembly .c.ceeeecns Heal EXChan@er veveesesscoscesesasessnasssssosscsnsssnssssoss Fuel and Coolant Pumps seeveesenss e s e s e Radiator and Radiator EnclosSure .eesececeescesossscassees ceu Salt Storage TAnKS seeeeesssscsossrssssacccsossnccssssoasesscs Freeze FlangesS seeeesescssssscsasrsstsensstsatassssssnsenss 15 Major ProCUrement tuseesesssosonsecesscsssessossrssasacsosncrens 15 Moderator Graphite ..eieeiessseetesonssstosassnscsssanssons 15 Salt Piping and Component Heating Equipment ...ceceeececses 15 Renmote Maintenance Equipment s.iceeiieiacssvsssensrtonsosanen 15 Reactor Auxiliary SyStemS .eveeecececsccsscsssesecnsncsases 15 Instrumentation Fabrication ....cevveeuen.. et sie e 15 Fuel-Salt Sampler and Enricher ....veeeees besersasenanensns 16 Status of Construction ...eeeeieviieeiecitenncsceressnssosnes v 16 Reactor Cell ........ cesisenseane ceennannan e reetasaesscens 16 Fuel-Drain-Tank Cell .veeveveescosssscasassssncosonsrasneas 16 Coolant Cell teevereecsssoanrsosnsseasossssosasssssssssssssnnns 16 CPFF Construction «eeesessceessasaarsosesssssscsssssessscnns 21 Procurement and Installation of Instrumentation ...ceeeeeeeesns 21 VOOV OOV O RnWW W 2. COMPONENT DEVELOPMENT .« « v v v eeeseonnnnnesonnesesnsconsnasnnnnnns 23 Drain-Tank Co0ler TEeST tveieevrsncessrssssssscscsasnassssnssnsocns 23 Heatel TeStS weeeersssanvissssosscasssnsssossasssosssassssocsnsss 23 Drain-Tank Heater teveesesoesscasrsssesnssscssassssosssasans 23 Pipe Heater with Reflective Insulation ...eieeveerscsnccans 30 CONETYOL ROA weveeeocsocnosssssaassassssssssssssnsansesnsssscescses 30 Control Rod Prototype seevevececensssacsssacness Ciesesannen 34 Helium Purification System seeeeeececscresscrscsnscascssssssses 34 Sampler-Enricher System MOCKUD +ceeeteoraessocosssssnanassonss 35 Flange DiSCONNECES sevseencecescasscsssasessssessssssssssnssss 37 Operational and Maintenance ValvesS suveeeesssessnssassnnnsonns 38 Design ceeeeannns Gt e e eeeeeneecectestet ettt atacans s 38 Engineering TeSt L.OOD sevevescocososcssosassssscsassssssansosss 38 Treatment of Flush Salt with HF in ETL Drain Tank Following Oxide AAAitiONS seeevvesoonesccsessesocesnsaanas 38 Treatment of Fuel Salt with HF in ETL Drain Tank ......ee.. 38 Xii Loop Operation with Fuel Salt and Graphite .. Operation of the Graphite Container Access JOoInt ceeeecescs Analysis for Chromium as an Indication of Corrosion ....... Xenon Transport in MSRE System .....ceveveeescesas et taseorran Prototype Pump Testing Facility ..... Water Test Pump LOOP etvercicasocsasosassacessnsnsssssnsossses Maintenance Development .ieiiesessettssessacsssscscnsonesns *o e Pump Development ..civiiecessescccascsasreoscncssassonsssessnees Prototype Pump Operation and Testing PKP Fuel Pump High-Temperature Endurance Test Test Pump with One Molten-Salt-Lubricated Bearing ....¢.... Lubrication-Pump Endurance Test ..... F'Llel lelp MI(-2 ooooooo LI R I B A B R B R A A .. Instrument Development .eiieceecesecasssasssasesosseacsscasss Single-Point Liquid-Level Indicator ..cieieeecectoccncosane Pump-Bowl Liquid-Level Indicator .... Temperature Scanner ..... tecnseranaan Thermocouple Development and Testing Closed-Circult Television for Remote Maintenance Viewing «seeeeee T R L High-Temperature NaK-Filled leferentlal Pressure Transmitter .vveceeocscrsesensnnsse .o Bubbler-Type Liquid- Level Indlcator .......... cessesrenraas MSRE REACTOR ANALYSTIS cevcecococnnoscnas Nuclear Characteristics of Core iveeeees Controcl ROA WoOTrth sevvivieersstnsescstosssenscssssocccsncsnosnses "Method of Calculation eeeesvssesccsccscescanvans teeesasanan RESULLS seeeeveeeeeionvacsassecnsssssasaasnsnasosnssssocssss Reactivity Shimming .e.eeveeeenssesssecssecssssascssssancnssses Inherent Neulron SOUI'CES cereceertssecssrsosrsesssrsesssscssosanss Biological Shielding .c...eeieeeeesecaacnsosassssssasasassansonse Part 2. MATERIALS STUDIES METALLURGY evevevesvnosaonsns ceesenns oo Heat Exchanger Fabrication ...cccevees et s e ssesserssensassan s MSRE Primary Heat Exchanger Brazing .... Ultrasonic Inspection of Tube Joint Brazing . Mechanical Properties of INOR-8 .eeeevwn. Thermal Fatigue ......... secstesscnnn Postirradiation Tensile Testing ..... Evaluation of MSRE Graphite ........ vese Chemical Composition and Oxygen Content ..... . Thermal Cycling of Salt-Impregnated Graphite seieeceseceese Irradiation Effects on Grade CGB Graphite ... Removal of Oxygen from Graphite with Molten Salts ............ Sintering Characteristics of Gd;03-A1,04 40 40 41 A A v 45 47 47 52 54 54 54 54 54 55 55 56 57 57 58 60 60 62 62 62 64 66 66 71 71 71 71 73 73 75 75 75 76 76 77 78 xiii R.ADIATIONEFFECTS. ------- l.l.ll.;.l'..l..l...l'l....I.'.I..I... Postirradiation Examination of Assembly ORNL-MIR-47-4 ..veeens Xenon Recovery from Graphite Cores ........... cesene ceeenaa Nonvolatile Constituents in Graphite Cores ..vieveccescesss Effect of Irradiation in Assembly ORNL-MIR-47-5 .eiveevon.. ces Production of CF, Under Operating Conditions .............. Postirradiation Examination of Assembly ORNL-MTR-47-5 «eeveenen Fluorine Generation from Decay Energy ........ cenescasasaee Qut-of-Pile Irradiations of Solid MSRE S8lts ceveessersescossne Gamma Irradiations .ieeeecescessens e X-Ray Trradiations ceeeerescessssscstcescssessssasasscssnsssse Irradiation with Van de Graaff Electrons ...ieeieesecececss Irradiation of CFy eevecrevesrrccsssnssassnssssstsssscssnnns CHEMSTRY a & ¢ 0 & s & 0 8 * 4 & 9 8 9 % 0 8 S % O % S 0 R 9 s 8 & S S S P S RS eSS LSS Phase Equilibrium Studies .ieeeeierssersscassssssssssccosacsss Phase Behavior in MSRE Fluoride Systems ...cvececescacscceas Liquid Immiscibility in the System LiF-BeF,-ZrF, ....c0v..s Petrographic Examination of Irradiated MSRE Salt «..veeeven Xenon Fluoride StUQIeS tveeevosasorssscscossosssssssossssnas Core and Blanket Fluids for Future Reactors .....ciieeceens Oxygen and Sulfur in Molten Fluorides ..iieieececeevecscosaass Oxide Behavior in LisBeF,-ZrF,-UF, ccivereeracisssesassenns Removal of Sulfates from LioBeF, «ceviieensccciscccnnnnneess Physical Properties of Molten Fluorides ..ieiiesecancsrcascans Interfacial Behavior of Molten Fluorides with CGraphite ceveeeenrseenssossenssssssasssocsssssssssssssssossses Viscosities of MSRE Fuel and Coolant ...ceecerccceccsccaacs Oxidation-Reduction Reactions in MSRE Melts (iveicevecerenccaes Chemical Reduction of Dissolved Fluorides of Structural Metals +ececesececcessnscasrssssossoaanssssss cees Apparent Solubility of UFs Produced by Reducing Fuel with Excess Zirconium Metal .ieeeeerecncrcaccosecsasasancs Development and Evaluation of Methods for the Analysis of the Radioactive MSRE Fuel ..i.iivieeesccrseserssnsscnsssssasnss Fluoride ....... cee st gses s esanse e sessas e s assan e s o Chromilum sevesessessecsnes ceeseaanas ceteir et s et s ettt neny sa OXygen ...... e checiescasessanescarnesttaans oo .o NiCKEl ceveverasscssacsnsossonnssccnsnnansnsss venen Development and Testing of Equipment in the Hot-Cell MOCKUD +eecovescorosnosssssssossscssassoonsssassssassssassssssssss Er[-]ELPROCESSING ? ® 8 B P @ P S # 9B B D SO E NSNS R s Bessoas 80 80 80 81 83 83 86 86 94 94 96 98 103 105 105 105 106 108 110 110 111 111 117 125 125 129 130 130 134 135 136 136 136 137 138 141 Part 1. MSRE DESIGN, ENGINEERING ANALYSIS, AND COMPONENT DEVELOPMENT 1. MSRE DESIGN, PROCUREMENT, AND CONSTRUCTION The MSRE design, procurement, and construction was 78% complete on July 31, 1963. otatus of Design Mechanical and Process Design Design of the mechanical systems for the reactor consisted mainly in modifying existing drawings to incorporate recent results of develop- ment work and to ease fabrication and construction problems. New design work was started on an overflow tank for the fuel pump, a new fuel pump with a larger bowl, and a strainer for the reactor vessel. The present design of the fuel pump was based on experience with components that existed early in 1960 and included enough margin of safety to ensure that a reliable pump could be developed in a short time. As a result the diameter and overhang of the shaft and the diameter of the bowl were limited to values which, in turn, restricted the free volume in the pump bowl to less than 2 ft°. The pump bowl is also a surge tank for the fuel system; and this small volume, although adequate for normal operation, provided very little margin for accommodating abnormal condi- tions. Initially, an overflow line was included in the design to return excess liquid from the pump bowl to the drain tank. However, the design of the overflow line was considerably more difficult than had been ex- pected; and since its operation presented several uncertainties, an over- flow tank was substituted for the line. The arrangement of the fuel pump and the overflow tank is shown in Fig. 1.1. The tank is a toroid, with a volume of about 5 ft°. Tt is located near the suction line directly below the pump bowl in the pump furnace and is coupled to the pump by a short overflow line. The tank is equipped with bubblers to indicate the liquid level. Liquid can be returned to the fuel system by pressurizing the tank with helium through the bubbler lines or the vent line. Spare parts to be supplied for the MSRE will include replacements for most of the major components. In some instances changes will be made in the replacements to improve the design and performance of the reactor. A replacement for the fuel pump is being designed. The design will include a larger pump bowl to eliminate the overflow tank and a float- type liquid-level indicator to substitute for one of the bubbler-type indicators. Experience with the prototype pump indicates that the shaft overhang can be increased sufficiently to provide a fuel expansion volume of about 6 ft° without changing the shaft diameter, the seals, the bearings, or the bowl diameter, and without significantly reducing the reliability. The float-type level indicator was developed too late to be included in 3 UNCLASSIFIED ORNL—DWG 63-6474 © °) PUMP MOTOR—— __| L ] [ | BEARING HOUSING | LIQUID-LEVEL- ; INDICATQR LINES — a2 o’ | VENT LINE C ' o | | B — i T Ko T = 15 r.—.::_JI:|§ rfi gjz P < OVERFLOW TANK —— PUMP SUCTION LINE OISO KOS AR A X X < » X \/ P X < . "I"‘tN = Fig. 1.1. TFuel Pump and Overflow Tank Assembly. the original pump; but it appears to be an important improvement over the bubbler devices and will be used on the replacement. The MSRE was designed on the basis that there would be no attrition of graphite in the core; therefore no chips of graphite would circulate with the fuel. Inspection of the graphite for the reactor indicates that some spalling can be expected; so a screen (strainer) was designed for installation in the reactor vessel to remove pieces larger than 1/8 in. in diameter from the circulating fuel. The installation i1s shown in Fig. 1.2. Part of the screen is welded to the head of the reactor vessel, but most of it is incorporated in an extension of the plug and control rod thimble assembly that enters through the nozzle on the top of the wvessel head. This assembly is replaceable, with some difficulty, after the reactor has been operated at power. Instrumentation and Control Design System Layout. Except for minor revisions the layout of the instru- mentation and controls remains the same, as described in previous reports, One process instrument panel was added in the auxiliary control room. Multipoint temperature recorders with alarm switches which will monitor the reactor drain-line temperatures are located in this panel. Two auxil- iary control panels serving the lube-oil packages were relocated in the service room to provide more space for other purposes. The two control panels for the helium cover-gas purification system were moved to the west end of the diesel house near the helium supply station. Locations of additional conduits between field-mounted equipment and the nearest wireways were determined. ILocations of thermocouple and instrument dis- connects within the reactor and drain tank cells remained firm, but the assignment of individual thermocouples to disconnects was considerably revised. These revisions are the result of a continuing review of remote- maintenance reguirements and the addition of an overflow tank for the fuel pump bowl. The locations of personnel radiation monitors and their inter- connecting wireways were determined. Flow Diagrams. All instrument application diagrams, instrument tabu- _ lations, and design drawings were revised in accordance with recent design changes. The diagram of the chemical processing system is not yet approved for construction.. Recent changes to the design of the fuel-pump-bowl cover- gas system, the addition of a pump-bowl overflow tank, and additional instrumentation dictated by control-circuit design will require further revisions to these diagrams. Simulation of Reactor Fill and Drain Transients. Further studies were made of an analog computer model of the fill and drain system.2 The fill-system control was designed to fill the reactor in a reasonable time and yet ensure a salt flow rate of no more than 0.5 cfm when the core is half full, according to the requirements of the postulated fill accident.3 A capillary restrictor was designed and calibrated for insertion in the helium fill line to so limit the flow under worst-case conditions. A simu- lation incorporating the characteristics of this capillary indicated that the time required for a normal fill would be approximately 3-1/2_hr. UNCLASSIFIED ORNL— DWG 636475 r's —FS\ — : N N Q -wyt N PLUG \\\ Q N ._fi" PN s 2 o — FUEL OUTLET wara ACCESS N \ NOZZLE — N A o VA4 L TOP HEAD OF REACTOR VESSEL o o © o o} o o o o ollo o of _ gl === - - r TS STRAINER ___lf:_ji_ /___:Iz\ . (16-gage PLATE WITH ¥,,-in. N i T ~diam. HOLES ON Yg4-in. L _ CENTERS) 2=9- SRl — NEBEE Lo g -_l____‘l_ | fl‘l'": dl | =l i L = =———2-in. GRAPHITE MODERATOR - il /: : T | GRAPHITE SAMPLES I / = j | | | | | | NN AN\ NN NN Fig. 1.2. Installation of Strainer in Reactor Vessel, Recent studies of various means of preventing pressure surges in the pump bowl resulted in the incorporation of a single capillary restrictor in the common bypass line between the pump bowl and the drain tanks. Further investigation of the undesirable results of pressure surges2 (possible dam- age to pump-bowl o0ll seal, reversal of purge-gas flow, and erroneous level instrument indication) led to a compromise between these effects and the retarding of a dump and to a relaxing of the requirements for limiting the surges. It was decided to design the restrictor such that the surge would not affect the pump-bowl level indication so long as the valve in the line to the spare drain tank vwas open. Table 1.1 summarizes the drain times for various conditions, where in each case the initial bowl pressure was 5 psig, the initial drain tank (No. 1) pressure was 27 rsig, and the bowl purge flow rate was 0.15 scfm. The freeze valve was assumed to open 5 min after the start of the run, and the drain times shown started when the freeze valve opened. The maximum rate of pressure rise in the bowl that will not affect the level indication was calculated to be 8.5 psi/min. Control Circuitry Design. The control and safety circuit require- ments were studied in great detail. Firm criteria for their design were .established, and the preliminary block diagrams were extensively revised. Design of the circuits proceeded in those areas where criteria were avail- able. OSchematic drawings of the 12 freeze valve circuits were completed and approved. Schematic drawings of the instrument air compressors and lube-0il pump circuits were also approved. Block diagrams showing control logic for process fill and drain, coolant salt, containment, freeze valve safety, and auxiliary process control systems have now been approved. To- gether these diagrams describe the criteria for control and protection during startup and operation of the process system. Additional block dia- grams of the control rod and radiator load control systems have been com- pleted and issued for comment. Design of the control circuitry is continuing. Table 1.1. Predicted Primary System Drain Times Maximum Rate Time to Time for Drain Tank 2 Drain Tank 1 Drain Tank 1 of Pressure Drain 25% Complete Bypass Bypass Vent Rise in Pump of Core Drain (HCV-545) (HCV-5L4Y) (HCV-5T3) Bowl (psi/min) (min) (min) a a Open Open Open 8.5 13.7 39.4 Open Open” Closed 8.5 16.6 45.8 Open Closed Opena None 15.2 59.4 Closed open® Open” 15 13.8 41.3 Closed Open? Closed 15 16.3 43 Closed Closed Opena None 19.8 127 aPosition at start of run. Control Panels and Cabinet Design. Designs of 40 of the 53 panel board sections presently required are complete. Seven of the remaining 13 are either partially complete or held for revision. Preliminary design studies have been made on the remaining six panels. The design of three thermocouple alarm panels, two thermocouple scanner panels, and one main board panel was completed and approved. Minor revisions to two main board panels were also completed. Work on two pro- cess radiation monitor panels is nearing completion, and the control-circuit relay cabinet design is under way. The fuel-pump level transmitter panel is being redesigned to accommodate the new equipment required by the addi- tion of the pump-bowl overflow tank. Field Installation Design. Drawings of the routing of interconnecting wiring and cables for the annunciator system and for the Foxboro Electronic Consotrol Instrumentation system were completed and approved for construction. Wiring diagrams of 16 junction boxes serving all valve position switches, instrument air system alarm switches, cooling water system, special equip- ment area, and the helium cover-gas system were completed and approved. The design of the thermocouple interconnections, including wiring dia- grams for the large junction boxes just outside the cells, were exten- sively revised to incorporate changes resulting from reassignment of thermocouples to disconnects within the reactor and drain cells and addi- tion of in-cell thermocouples. Thermocouples. Detail design of the thermocouple system, including interconnections, is ~80% complete. Design of the thermocouple installation has proved tc be more diffi- cult than originally expected. Owing to the requirements of remote main- tenance and the' compactness of the in-cell reactor system, the location ' of disconnects and the routing of leads as well as the disconnect assign- ment and the point of attachment of each in-cell thermocouple have had to be carefully selected. DNumerous changes in thermocouple location and other drawings were required before the many conflicting requirements were satisfied. The design of this system is now considered to be rea- sonably firm. Thermocouple locations in the chemical processing system and the fuel loading and transfer system were determined. The thermocouple tabulation was revised and approved for construction. A proposed design for the two thermocouple wells, one each to be installed immediately upstream and downstream from the coolant salt radiator, was submitted for comment. Cal- culations indicate that the proposed well is strong enough to withstand the conditions of temperature, pressure, velocity, and vibration existing in the system. The design of the thermocouple scanner system is complete except for the addition of a signal-identification device. Process and Personnel Radiation Monitors. The number and type of process monitors remain as previously reported. The chemical plant moni- tors have not been specified. Panel, shield, and detector installation drawings were completed, and all equipment was ordered. Personnel monitoring instruments remain as previously reported. A drawing showing individual instrument locations was issued. The instruments were ordered and most have been delivered. The criteria for incorporating the personnel monitors into the building evacuation system are being de- veloped. The criteria for the off-gas stack monitoring system were established. Status of Fabrication of Major Reactor Components Reactor Vessel and Control Rod Thimble Assembly The reactor vessel (Fig. 1.3) is ~90% complete; final assembly will be done when the graphite core blocks are received in October. After some temporary suspension of work on the control rod thimbles, pending the re- sults of development work, drawings are being revised and fabrication work is being restarted. Heat Exchanger The primary heat exchanger was completed in the Y-12 Shops and de- livered to the reactor site (Fig. 1.4). Fuel and Coclant Pumps Fabrication of coolant pump bowl was completed (see Fig. 1.5), and final machining of the fuel pump bowl is in progress. Work is continuing on the fuel pump cooling shroud. Fabrication of the overflow tank is in progress, and a float type of level indicator unit is being made for the .coolant pump. The finish machining of the bearing housings for both pumps was completed, and the pump lubricating-oil stands are complete except for the installation of the instrumentation. The first of the contalnment vessels for the pump drive motors is being finish-machined prior to installation of the motor components. Vessels for three more motors are in various stages of weld repair of laminar defects. Radiator and Radiator Enclosure The salt-to-air radiator was completed, and its installation in the radiator enclosure is nearing completion (Fig. 1.6). The electrical heating eguipment is also being installed in the radiator enclosure. Salt Storage Tanks The fuel-salt flush tank, the coolant-salt storage tanks, the two fuel-salt storage tanks (Fig. 1.7), and the two steam domes and bayonet assemblies for cooling the fuel-salt storage tanks were completed and delivered to the reactor site, Building T7503. Salt level indicator probes are being fabricated for each of the four salt storage tanks. UNCLASSIFIED PHOTO 39332 Final Machining of Reactor Vessel. e &) Py Fig. 1l.4. Completed Shell-Tube Heat Exchanger for MSRE. UNCLASSIFIED PHOTO 39597 TT UNCLASSIFIED PHOTOQ 39954 Fig. 1.5. MSRE Coolant Pump Bowl After Final Machining. UNCLASSIFIED PHOTO 39935 g = EEEET — T - Ll 120 Fig. 1.6. Salt-to-Air Radiator Installed in Radiator Enclosure. 15 Freeze Flanges Four of the 6-1/2 pairs of freeze flanges were completed and delivered to the reactor site. Fabrication of freeze flange clamps is in progress. Major Procurement Moderator Graphite The National Carbon Company has produced graphite bars which are satisfactory for the moderator core blocks. Machining is in progress, and delivery 1is scheduled for October. Salt Piping and Component Heating Equipment The coolant-salt storage-tank furnace, fuel-pump furnace, and four salt storage-tank furnaces (less removable electric heaters) were completed. Fabrication of 12 heater control panels, reactor heaters, and drain-tank removable heaters is continuing. Procurement was completed for numerous electrical items, including heaters, cable, transformers, thermal insulation, wire, seals, terminal blocks, and other materials. Procurement is approximately 90% complete for all electrical items except the special pipe heaters. A contract was awarded Mirror Insulation Company for the detail design and fabrication of special units for heating the salt piping and heat exchanger in the reactor and drain-tank cells. : The heater supports in the drain-tank cell were completed;_ Remote Maintenance Equipment Procurement is complete for special optical tooling equipment. Vendor fabrication of the large, portable, sliding shield for the maintenance facility is nearing completion. The fabrication of pipe alignment brackets for freeze flanges is nearing completion. Fabrication was started on the graphite sampler equipment. ' Reactor Auxiliary Systems Young Radiator Company completed the three space ccolers for removing heat from the reactor and drain tank. The two helium preheaters and a leak- detector valve cabinet were finished. One stainless steel expansion tank, one stainless steel condensate tank, and other miscellaneous equipment were obtained from vendors. Fabrication of a treated helium surge tank is in progress. Instrumentation Fabrication Four stainless venturi flow elements for measuring flow in the lubri- cating-oil packages for the salt pumps were delivered. The neutron instru- ment tube extension for the reactor is complete. The neutron tube harp assembly and six process line detector lead shields are being fabricated. 16 Fuel-Salt Sampler and Enricher The fuel transfer tube and positioning jig assembly were completed. The special motorized valves were received from the vendor. TFabrication of the main unit of the sampler-enricher awaits completion of a redesign -of some pileces on the basis of results of development tests. Status of Construction Reactor Cell The thermal shield was installed, and the interconnecting piping was completed (Fig. 1.8). The support steel was installed for the fuel pump, heat exchanger, fuel piping, and auxiliary piping. The disconnects were attached to the heater cable, the cable was installed, and the disconnects were mounted. Water piping was installed for the component cooling system and the space coolers. About one-half the auxiliary piping has been fabricated and installed for the fuel pump. All penetrations for leak-detector tubing, valve air lines, and electrical cable were installed, and the welding was completed. The Jjig was assembled for the reactor heat exchanger and fuel pump; optical tooling is being used to locate flanges, equipment center lines, etc. The heat exchanger was located on the jig along with two pairs of its flanges (Fig. 1.9). Fuel-Drain-Tank Cell All support steel for the drain tanks, drain piping, electrical dis- connects, etc., was installed, and the welding was completed. Steam and water lines were installed for the steam domes. The drain- tank furnaces were installed (Fig. 1.10). The fuel flush tank was fitted to the jig and located in the cell (Fig. 1.11). The steam dome for the drain tank No. 2 1s being fitted into the jig, and this operation is 90% complete. Ninety percent of the helium priping and the drain-line piping has been fabricated. Coolant Cell Installation of the o0il catch tanks, waste oil receivers, intercon- necting piping and valving is ~50% complete. UNCLASSIFIED PHOTO 62899 LT Fig. 1.8. View of Reactor Cell During Installation, Showing Thermal Shield and Auxiliary Piping. UNCLASSIFIED PHOTO 62079 8T Fig. 1.9. Optical Tooling and Jigging for Accurate Assembly of the MSRE Fuel System. Fig. 1.10. N Drain-Tank Cell, Showing Assembly of Steam Dome Cooling System for Fuel Drain Tanks. (‘yI 20 UNCLASSIFIED PHOTO 61732 Fig. 1.11. Coolant Drain Tank Assembled on Locating Jig. 21 The water piping was installed between the water room and the reactor cell penetrations. The supports were installed for the coolant drain tank and the coolant bump . CPFT Construction Cable trays and other wiring'and conduit supports are being installed in the basement of Building 75C03. This work is ~lO% complete. The dismantling and removal of switch-gear panels, circuit breakers, etc., in the diesel building west of Building 7503 was started. Existing wiring and electrical gear to be used in the MSRE are being checked out; this work is ~10% complete. The installation of the component cooling tanks and compressors and miscellaneous auxiliary piping outside the reactor, fuel drain tank, and coolant cells has begun. Procurement and Installation of Instrumentation With the exception of some additional instrumentation required by recent revisions and instrumentation for the chemical processing system, preparation of specifications and initiation of procurement of process instrument components were completed. Most of these components are now on hand, and delivery of the others is expected within the next three months. Several orders for components that required special development or procurement effort were completed. The special components include the weld-sealed tranmmitters and valves for radioactive helium gas service. Freeze-flange and freeze-valve temperature alarm switches were received from the Electro Systems Corporation, and acceptance tests were completed. Vendors quotations for weld-sealed solenoid valves are being evaluated. Purchase orders have been placed for all components of the thermocouple scanning system. ' AEC approval was obtained for purchase of a computer data-logging system. Proposals received from four vendors are presently being evalu- ated. Most of the equipment required for the process radiation monitor system is on hand, and the remainder is on order. Personnel radiation monitor instrumentation is on order, and procurement of nuclear instru- ment components is under way. Fabrication of 28 instrument panels was completed, and 11 additional panels are being made. Installation of completed panels and other equip- ment at the reactor site was started in June and is rapidly continuing. 22 The installation of 149 thermocouples in the radiator assembly was completed. These installations were particularly difficult owing to the special nature of the mechanical attachment to the thin-walled radiator tubes and the compactness of the radiator assembly. Thermocouple dis- connects and multiconductor extension cables in the reactor cell are being installed. Installation of control valves and signal transmitters in the coolant drain cell is also under way. The ORNL Shops are fabricating the special alarm discriminator units for the temperature scanner system. References 1. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNI-~3419, p 49. 2. Tbid., pp 52-5kh. 3. J. R. Engel, P. N. Haubenreich, and S. J. Ball, Analysis of Filling Accidents in MSRE, ORNI, TM-L49T7. 2. COMPONENT DEVELOPMENT Drain-Tank Cooler Test The drain-tank cooler test was terminated after 2600 cycles between 1300 and 212°F owing to a leak in the steam system. This test is be- lieved to be equivalent to several years of service in the MSRE. TFigure 2.1 shows one of the three Inconel cooling tube assemblies after it was opehed for examination. The l-l/Z-in.-diam, 0.065-in.-wall thimbles which were in contact with the carbonate salt on the outside and air on the inside were in good condition. Two of the three l-in.-diam sched-40 pipe cooling tubes that were inserted in the thimbles were cracked, as were the l/2—in.—diam, 0.049-in.-wall water tubes contained in the cooling tubes. The third tube was not opened. Figure 2.2 shows the most severe cracking, which occurred 38 in. from the bottom of the cool- ing tube approximately 36 in. below the surface of the salt. There was additional severe cracking on this tube at the same elevation, but on the opposite side. Figure 2.3 is a detailed photograph of the water tube shown in Fig. 2.1, indicating the severity of the cracking. This cracking was at the same elevation as that shown in Fig. 2.2, and the resulting jet of water possibly contributed to the damage of the cooling tube. Figure 2.4 shows the lower part of the cooling tube from Fig. 2.1. This figure shows how the l/l6—in. spacer bars that centered the cooling tube inside the thimble had cracked. It also shows another crack at a weld point. The thermocouple shown had failed previously, as indicated in the thermo- couple test described later in this chapter. Repeated, severe thermal stressing of the tightly confined compo- nents appears to have caused the cracking in all cases. However, samples of the damaged sections are being metallurgically examined to confirm this. The design of the spacer fins shown in Figs. 2.3 and 2.4 was changed, and a prototype cooling tube of INOR-8 was assembled for further testing. Heater Tests Drain-Tank Heater A prototype drain-tank heater was built and is being tested to demon- strate the structural strength and thermal stability of the stainless steel container. Figure 2.5 is a photograph of the unit during assembly, showing the eight individual ceramic-encased heating elements which have a total rated capacity of 4.55 kw. The container is 101-3/4 in. high, 2 in. thick, and is curved on a radius of 28 in. Each heater unit is equipped with a remote electrical disconnect and pickup bail. There are sixteen heaters of this type in the heating system of the three drain tanks. 23 1¥5-in. DIAM, 0.065-in. WALL THIMBLE PoT0 H8 FIG 2.4 FIG 2.2 1-in. COOLING TUBE (SCHED-40 PIPE) i 3 SS CLAD THERMOCOUPLES 2.1. Y/o-in. WATER INLET TUBE, 0.049-in. WALL Drain-Tank Cooler Assembly After Test. e UNCLASSIFIED PHOTO 39628 Fig. 2.2. Failure of SPACING RIBS \ THERMOCOUPLES UNCLASSIFIED PHOTO 39804 9¢ moHES bl CRACKS Failure of 1-in. Cooli AN SPACING RIBS 6-in. Elevation. UNCLASSIFIED PHOTO 39625 L2 28 UNCLASSIFIED ELECTRICAL DISCONNECT PHOTO 39334 PICKUP BAIL CERAMIC ENCASED HEATING ELEMENT COVER PLATE Fig. 2.5. Drain-Tank Heater. 29 The unit has operated for 2876 hr at 1200°F average temperature without difficulty. The temperature distributions are shown in Fig. 2.6. Temperatures on the heater face were measured at the approximate center of each ceramic element with thermocouples that were cemented on the outer sur- face of the ceramic, under the outer metal cover plates. Temperatures on the cover plate were measured on the surface of the inner metal cover plates at the indicated positions. The large temperature difference between the top and bottom of the heater was considered undesirable; and the electrical circuitry is being changed to put the upper and lower zones on separate controls. An addi- tional 4 in. of insulation is being added to the top of the enclosure to UNCLASSIFIED CRNL—DWG 63—6476 28-in. R DISCONNECT OUTSIDE FACE INSIDE FACE o | 695 X1240 h X ¢ 1290 X X )\ 1365 1360 1305 1285 X X X X 1345 1350 1285 X X1230 1230 X X X 1295 1205 1210 1160 X X X 1150 1190 X 160 X X 980 840 HEATER FACE COVER PLATE TEMPERATURE (°F) TEMPERATURE (°F) POWER INPUT: TEST HEATER — 5.36 kw Fig. 2.6. Drain-Tank-Heater Temperature Distribution. 30 make the total thickness 8 in. to reduce the large heat losses at this point. Testing will continue when the changes are completed. Pipe Heater with Reflective Insulation The reflectively insulated heater boxes for the 5-in. pipe were operated as a life test. Data from the beginning and the end of six months of continuous operation at 1400°F are compared in Fig. 2.7. The increased heat loss is due to a change in the emissivity of the lamina- tions and damage to the silver plating of one of the units, as previ- ously reported.t Silver plating of the undamaged heater is in excellent condition. Both units under test appear to be in good condition, with little warping. A contract has been let to the Mirror Insulation Company of Lambertville, New Jersey, for the detailed design and fabri- cation of similar heaters for the reactor. UNCLASSIFIED ORNL-DWG 63-6477 1400 0/.- ./ o) 1200 o 44r’°"' 4% SEPT. fff;/"’ ”’d,jb”’ ,/”’/’ "”,/’1WAY 1963 1000 / o y ® 7,= CERAMIC HEATER TEMPERATURE Ty=Ta (OF) T,= AMBIENT AIR TEMPERATURE 800 | l 300 400 500 600 700 HEATER TEMPERATURE (w/lineur ft) Fig. 2.7. Comparison of Insulation Performance of Removable Reflective-Insulated Heater Boxes (Mirror Mfg. Co.) for MSRE 5-in. Pipe. Heaters had been op- erated for six months at 1400°F. Control Rod The variation in length of the flexible control rod after a scram operation continues to be a problem. The accuracy of the rod position meafurement needed for use in analysis of reactivity data was established as =0.z2 in. ©Since the position-indicating devices operate at the upper end of the rod while the active section is at the lower end, any varia- tion in length due to distortion, thermal expansion, or slip in mechani- cal linkage will produce an error in measurement of true position. 31 UNCLASSIFIED PHOTO 62054 POWER- AND POSITION- 47 INDICATOR PACKAGE | 4— CHAIN DRIVE CONTROL ROD DRIVE UNIT HOUSING TOW BLOCK AND 4 ROD DISCONNECT 90Y-in. THIMBLE ADAPTOR FLANGE N Fig. 2.8. Control Rod Drive. 32 UNCLASSIFIED PHOTO 62055 e 37-PIN DISCONNECT POSITION-INDICATOR SYNCHROS DRIVE SPROCKET GEAR REDUCTION b AND CLUTCH UNIT s— .} CHAIN DRIVE 25-w AC LOW INERTIA SERVO MOTOR SHOCK ABSORBER AND TOW BLOCK Fig. 2.9. Control Rod Power- and Position-Indicator Package. UNCLASSIFIED DRIVE UNIT ADAPTOR FLANGE PHOTO 39863 ROLLER HOUSING 2 x0.065-in-WALL INOR-8 TUBE — COOLING AIR EXIT REPRESENTS 10-in. ACCESS NOZZLE FLANGE HANGERS POSITIONING SLEEVES Fig. 2.10. Control Rod Thimble. 34 A positive position indicator is being tested which will detect the passage of the lower end of the rod past a known point. It is a simple ailr gage that makes use of the cooling air which normally flows down the hollow control rod. An air-exhaust nozzle with radial ports is attached to the bottom end of the rod, and a throat (or restrictor) is welded into the guide bar cage at a known point in the thimble. When the nozzle passes through the throat, a clear signal is indicated on the recording of the pressure drop across the rod. By calibrating the synchro indicators of rod position with the pressure signals when a rod is installed, any change in the rod length during operation is indicated as a change in the calibration. Preliminary results show that the error in indication of the single position is < 0.1 in. Control Rod Prototype The prototype rod drive mechanism was received from the Vard Corpo- ration of Pasadena, California,in mid-July. Figure 2.8 shows the complete drive unit together with its housing, and Fig. 2.9 shows the power- and the position-indicator package. One of the position-indicator synchros was found to be electrically grounded internally and was returned to the contractor for replacement. Installation of the prototype control rod test facility is essen- tially complete. Figure 2.10 shows the complete control rod thlmble, in- cluding the drive unit adaptor flange. The section of thimble from the hangers down will be immersed in salt in the reactor and will be heated in a furnace to reactor tempera- tures for the test. The overall length of the thimble is 11 ft, 3- 1/2 in. The complete control rod system will be tested when the control cir- cuitry is checked. Helium Purification System Construction of the test loop for the helium purification system was completed, and proof testing of the dryer and oxygen-removal unit was started. The loop consists of a molecular sieve dryer for water removal and a high-temperature titanium spongebed for oxygen removal, plus asso- ciated flow, pressure, and temperature controls and analytical equipment for determining the moisture and oxygen content of the gas stream. The dryer and oxygen-removal unit are full-scale replicas of units designed for the MSRE cover-gas system. The efficiency of the system is checked by introducing helium at known impurity levels and monitoring the effluent gas. A number of short ( the liquid level in the joint was controlled by trapped gas pressure and no frozen seal was formed. The increase in chromium concentration in salt samples removed during fuel-salt operation indicated a corrosion rate three times that experienced during previous flush-salt operation. The loop was shut down in June, after 2000 hr continuous operation, for examina- tion of the joint and in preparation for testing the operation of the sampler-enricher mockup.® Treatment of Flush Salt with HF in ETL Drain Tank Following Oxide Additions Following the treatments of the flush salt with HF and H, to remove oxide that was introduced by the graphite, two additional treatments were performed to evaluate the oxide-removal and cold trap collection effi- ciency. Beryllium oxide was added in an amount equal to ~100 ppm oxygen in the salt. The HF treatment began ten days later (the fifth treatment in this tank), and 72% of the added oxygen was collected as water in the cold traps in the off-gas line. Following the installation of an im- proved moisture trap in the off-gas line, a second oxide-addition test indicated removal of 80% of the oxygen that was charged. Treatment of PFuel Salt with HF in ETL Drain Tank * In preparation for loop operation with fuel salt, the salt was treat-~ ed with HF in the alternate ETL drain tank. An equivalent of 625 ppm of 39 oxygen was removed from the 162 kg of salt during 168 hr of treatment with a mixture of H, and HF. The large total amount of oxide that was removed and the low solubility of oxide in this salt” indicate that a large por- tion was a precipitate. Figure 2.13 shows the water collection rates observed during the HF treatment of the fuel salt. Each of two periods of HF treatment was followed by a period of stripping with H, and He. The kinetics of water stripping is characteristic of a process where the metal oxide (ZrOE) is present as two phases in the salt mix. The active phase is undoubtedly dissolved in the molten salt and reacts rapidly with HF. The inactive phase is probably a solid oxide which dissclves slowly as the salt becomes unsaturated. The newly dissolved material can now react further with the HEF. UNCLASSIFIED ORNL-DWG ©63-6478 50 T I l PHASE T ( | | | | | ] | | ! | | 40 - | H, +He ] DA " STRIPPING I HF+ M T 30 i ] Ay — 20 | | H2+HeA“__j____ | I STRIPPING _ | : = ' ! £ 10 T L ! | = | - OBSERVED | @ 6 COLLECTION | o RATE = |\ | © 4 1 I © \ A | . NG | | ‘*6 2 I i { ° ‘ z | | | L PHASE 11 | | | z ! i N N i - 08 { N | . N — S — = | LL E 08 \ \‘\ R\ RATE +— “ \ N N\ | 61 04 N I | 8 \ N \ | i T | 2 \ S | & o2 \j\\\XCF i (s ' | —N— | 0.08 1 i T\ | HAS 004 l PPHASEL N | \ | 0.03 4 0 40 80 120 160 200 2 .TREATMENT TIME (hr) 0 Fig. 2.13. Removal of H,0 During HF Treatment in ETL Fuel-Drain Tank. 40 Analysis of the data indicates that ~68% of the total oxide was orig- inally in the inactive phase. After the treatment the salt mixture was stripped by sparging with H, for 38 hr and with He for 60 hr before it was loaded into the loop. Loop Operation with Fuel Salt and Graphite After the graphite was sampled through the dry box, the graphite access joint was opened to the atmosphere so that the deposit which re- mained from the flush-salt operation could be removed. This was done to provide a clean surface for any subsequent deposition during fuel-salt operation. After the loop was reassembled, it was evacuated at 200°F and then cleaned with flush salt. To prevent contact with the cleaned access joint surface, the flush salt was kept out of the annulus by a continuous bleed of helium into the joint. The loop was placed in operation with fuel salt on April 4 and oper- ated at 1200°F until June 27 for a total of 2018 hr. Operation of the Graphite Container Access Joint The access joint was again operated without a frozen salt seal so that samples might be obtained if any deposit collected due to cold trap- ping. After the loop was filled with salt, the liquid level in the joint was adjusted to the position and temperature of the previous operation with flush salt,® but without the stabilizing benefit of the frozen salt cake at the top of the joint. For the first 500 to 600 hr of operation, the salt level in the Jjoint rose slowly upward, necessitating periodic additions of gas to maintain the desired level detected by thermocouples. Tesgsts with a helium leak detector showed that this effect was not caused by a loss of the trapped gas due to leakage. The possibility that the loss of gas was due to transport from the joint to the lower pressure of the pump bowl as dissolved gas in the molten salt was tested by introducing argon into the joint annulus. The much lower solubility of argon in molten salt caused the loss of gas to be reduced to an insignificant amount; and the liquid level remained reasonably stable (¥1/8 in. estimated), with no additions of gas being necessary during the final 1400 to 1500 hr of operation. By cooling the upper portion of the joint with air, the container wall at the liquid-gas interface was maintained at ~lOOOOF, 200° below the bulk salt temperature. The loop was drained with cooling air on the upper joint and with- out additional heat. The joint was examined via the removable dry box facility,9 and samples of the deposit were removed for analysis. Unlike the previous joint that was exposed to flush salt, the removable plug was completely clear of any deposit or discoloration below the liquid level. At the liquid level, there was a deposit on a portion of the cir- cumference of the plug, with only a discoloration on the remainder. The deposit occurred on the side corresponding to the salt outlet of the graphite container, indicating it to be flow oriented, as was noted dur- ing previous operation. The deposit appeared in definite layers, begin- ning next to the metal wall with a dark, metallic layer that was covered 41 with salt crystal bands of green, clear (white), and orange. The pre- dominant layer contained the white crystals and occurred at the liquid level. This crust formed, evidently to the exclusion of uranium bearing phases, by selective freezing of the lithium, beryllium, and zirconium phases identified in Table 2.1l. The presence of the ZLi-BeF,, which melts at 8500F, would indicate the temperature to be this low at this location of the joint. This is possible, even though the outside wall temperature was recorded at approximately lOOOOF, because the tempera- ture gradient at this point is very steep. Zirconium oxide was identi- fied petrographically, but no uranium oxide was detected. Chemical analyses of two of these samples are given in Table 2.2 along with the analysis of a typical salt sample that was removed from the circulated fluid during operation. Table 2.1. Phases Identified in X-Ray Diffraction Examination® of Samples Removed from ETL Access Joint After Fuel-Salt Operation Sample Phase and Approximate Freezing Temperature Dark, metallic Metallic iromn, ZrQO,, ZLiF-BeF, (8500F), 2LiF-ZrF, (1100°F), Lig-Be-Zr-Fi, (990CF) White crystal 2LiF-ZrF, (1100°F), 2LiF-BeF, (850°F), Lig-Be-Zr-Fy, (990°F), ZrO, Orange crystal Identical to white crystal Table 2.2. Chemical Analyses of Samples Removed from ETL Access Joint After Fuel-Salt Operation Sample Ni Cr Fe Zr U Th Li Be Ne Dark metallic 213 ppm - 983 ppm 13.6% 35.6% 0.97% White crystal 22 ppm 45 ppm L.6% 0.15% 11.9% 7% 0.09% Typical sample from 5 ppm 870 ppm 65 ppm 9% 0.61% 0.2% 11.0% 6.1 1.2% circulated fluid Analysis for Chromium as an Indication of Corrosion Since corrosive attack by trace impurities is directed toward chro- mium in the container walls,l most of the salt samples removed from the loop during operation are routinely analyzed for chromium and iron (analysis for nickel usually shows <10 ppm). 42 It was noted that there was a much more rapid increase in chromium during the fuel-salt operation than during flush-salt operation. Figure 2.14 shows chromium and iron during both operations for comparison. Al- though the loop and graphite container are fabricated of INOR-8, an esti- mated 35% of the wetted loop surface area is Inconel in the Dana pump. If it is assumed that chromium is removed to a uniform depth from both Inconel and INOR-8 surfaces, the increase of 500 ppm during fuel-salt operation shown in Figure 2.14 represents the removal of chromium to a depth of 0.7 mil during 2000 hr of operation. The initial sharp increase of chromium and iron during fuel-salt op- eration is difficult to account for by the reaction with moisture in the loop or by residual HF in the salt after treatment. However, it is pos- sible that the available iron and chromium in solution, which were oxi- dized to the 3% state by the HF treatment, were not reduced owing to in- sufficient hydrogen stripping. Thus the initial 200 ppm Fe 3t and 40O UNCLASSIFIED 1200 ,\/\/ ORNL—- DWG 63~ 6479 T T T T O CHROMIUM ® GEROMIUM o IRON ~ ® IRON | ' 1400 — e | FLUSH-SALT : : N FUEL-SALT | OPERATION WITH | i OPERATION 1000 ——1 ' ' | | GRAPHITE AT | | T[T WITH GRAPHITE AT 7 | 1200°F L 1200°F : 900 l— 4540 hr —s=rq {HF |} fe——— 2018 hy ——=] ' | R | (P | | ]I ’4lfi1 500 ! | Bl ."" | | IR —_ , | 3 700 | | T e | | :l | e | | L o | I I I | ' | = | I 2 500 | | | : | = I 8 : | ! | ? // T 400 l : i |/ F S | [ | ‘ 300 o 0° | n | ee P | [ a 3 [ O 0| TANK Feddm 200 © oy | \ Sy 100 TANK Gr =t < = ~ g B \\.,___. e — 1 T 0 1 AUG SEPT OCT MARCH APRIL MAY JUNE 1962 1963 Fig. 2.14. Results of Chromium and Iron Analysis from ETL During Operation with Flush Salt and Fuel Salt. ] 43 ppm Cr 3t would produce an additional 100 ppm FeF, and 200 ppm CrF, by the reaction 2FeF, + 2CrFs + Fe© + Cr° — 3FeF, + 3CrFy. The remaining increase in chromium and the concomitant decrease in iron would then be an indication of corrosion due to FeF, + cr® - Fey4 CrF,. This would indicate the need for additional hydrogen stripping to reduce corrosive impurities after treatment with HF. It will be noted from Fig. 2.15 that the chromium leveled off near the end of operation as the iron was depleted. UNCLASSIFIED ORNL -DWG 63-6480 100 o0 | 80 \ POWER LEVEL: 10 Mw CORE FLOW: 1475 gpm BYPASS FLOW: 65 gpm xe'3% DIFFUSION COEFFICIENT: 70 Dyg=1.527 x40~ #t¥ne xe '35 MIGRATION TO GRAPHITE (%) H [+)] (@) (o] / o, o ~ C 20 40 60 80 100 120 STRIPPING EFFICIENCY (%) Fig. 2.15. Migration of Xel?’ to Graphite in MSRE. by Xenon Transport in MSRE System Preliminary studies have shown that the xenon distribution through- out the MSRE system (and the Xel3® poison level) is strongly dependent on the removal efficiency in the pump bowl (Fig. 2.15). Experiments have been planned and are being conducted for determining this removal effi- ciency. Prototype Pump Testing Facility The initial experiments for measuring removal efficiency were at- tempted in the MSRE prototype pump testing facility. Since there was no effective means for measuring the concentration of dissolved tracer gas in molten salt, it was necessary to monitor the tracer activity in the pump bowl gas space. The experiments consisted in saturating the system with a tracer gas (Kr®%), purging the pump bowl with helium, and monitor- ing the decay of the tracer concentration in the pump bowl. The insolu- bility of Kr®5 in molten salt prevented determination of the removal efficiency. At saturated equilibrium the volumetric concentration of Kr8% in molten salt is 1/1600 the Kr®> concentration in the gas space. Thus, during the purging of the gas space, the Kr® (dissolved in the liquid) was removed; but its effect on the Kr8% gas concentration in the pump bowl could not be detected. The gas concentration in the pump bowl must decay by a factor of 1000 before the gas concentration in the liquid will affect it and a difference in stripping ability can be realized. These results are shown in Fig. 2.16. Such low tracer concentrations could not be monitored, and subsequent experiments were conducted in the wvater test pump loop. Water Test Pump LooplZ2 The problems of low solubility of tracer and sensitivity of analysis are reduced by using carbon dioxide in solution in water for the test. The information from the COs-water experiments will be applied toward the determination of the efficiency of stripping xenon from molten salt. This experimental facility is similar to the prototype pump testing facility in all necessary requirements for conducting a stripping experiment, namely, size and configuration of pump bowl and stripping device. Carbon dioxide is dissolved in the water and is removed by an air purge through the pump bowl. The main difficulty has been caused by the production of bubbles of air, water vapor, and carbon dioxide in localized low-pressure zones. The presence of bubbles complicates the analysis of the experimental data. However, a low-pressure zone caused by a valve, orifice, or pipe bend is an excellent stripping device and may provide an effective means of xenon control. BSystem modifications and operation with low flow rates were necessary measures taken toward the elimination of bubbles. The first bubble-free results, taken with a flow rate of 850 gpm and a 2.2% bypass fl%w through the pump bowl, indicated a stripping efficiency as low as 17%. 45 Future experiments are planned for determining stripping efficiency under various conditions of stripper flow rate and air purge rate to ob- tain a full understanding of the xenon-removal mechanism within the MSRE pump bowl. UNCLASSIFIED 14 ORNL-DWG 63—-6481 10 T I T T ] 1 { 1 | I ] | 4 95 % REMOVAL EFFICIENCY | e 2 % REMOVAL EFFICIENCY — \ 3\ AY \ PURGE RATE = 4000 cm3/min \ \ 2 L\ \ [ 103 < A Y — Y "E \ <7 \ @ \ 2 \ s 2 z \ E 1012 L% xr . “ S \ w 5 AY Q \ 5 \ 0 \ CD‘_ 2 \ X 14 . \/' 10 o N S LN AN , \\\ 1010 \ \e' 0 60 , 120 : 480 TIME ( min) : Fig. 2.16. Kr8> Tracer Decay Rate in MSRE Pump Testing PFacility. Maintenance Development After a successful demonstration of the freeze flange maintenance procedure, the design of the tools was completed, incorporating the im- provements indicated by the demonstration. Special consideration was given to the flange in line 100, which connects the reactor and the fuel pump. Overhead interferences with another salt line on one side of the work area and a pump support beam on the other side requires a more com- plex flange-clamp operator tool. 46 This tool differs from the others tested in that the applied load is offset from the center line of the hydraulic cylinder, thus putting a moment on the ram. This effect was reduced by using the outside wall of the cylinder as a bearing surface; no difficulty was encountered after repeated tests. Additional alterations were made to the support mast ex- tension to reduce the problem of engaging the tool onto the clamp. A device was fabricated to permit tilting such masts to a more convenient angle for working under interferences. The pipe alignment tool was also modified to permit working with these overhead interferences. Test of an improved version of the HRT rod light for remote illumin- ation was started. With this design (shown in Fig. 2.17) the electrical hookup is entirely mechanical so there will be no solder to melt off as the lamp heats up. The bulb guard is stronger and provides more mechani- cal protection. The added expense of fabricating these lights will be offset by the savings in maintenance time. A tool to handle the thermocouple disconnect was designed with align- ment guides to protect the pins of the disconnect during assembly. A prototype heater, including the power lead-in and disconnects, was assem- bled; and tests'? indicate that it is a satisfactory arrangement. unct Assriep atsiny N BULB GUARD MAST FITTING CERAMIC INSULATOR BLOCK 750-w PROJECTOR BULB Fig. 2.17. Improved Light for Remote Illumination. 47 Pump Development Prototype Pump Operation and Testing High-Temperature Circulation of Molten Salt. The prototype pumpl4 test continued with circulation of the salt LiF-BeF,-ZrF,-ThF,-UF, (70-23-5-1-1 mole %) at 1200°F and 1200 gpm. Test operation of the pump was temporarily halted after 4700 hr to modify the test facility. During the run, tests were made to determine the back-~diffusion rate of radio- active gas up the shaft annulus against the helium purge; and a program was started to measure the concentration of undissolved gas in the circu- lating salt. ' Diffusion of Radioactive Gas Up the Pump Shaft Annulus. One of the functions of the helium purge down the pump shaft in the MSRE fuel pump is to prevent or minimize the diffusion of radioactive gas from the pump tank into the region of the lower seal and catch basin, where it could cause the oil that leaks past the shaft lower seal to polymerize. The damage to this oil is not important, per se, but gradual blockage of the drain could result in forcing the leakage into the pump shaft annulus and could eventually cause seizure of the shaft. The radiation tolerance of the oil used with the MSRE pumps is 107 rads/g. With this amount of irradiation the oil will flow and drain from the catch basin. The permissible source in the catch basin of the MSRE fuel pump for a dose of 107 rads/g of oil with 1-Mev beta radiation is 0.13 curie. This is based on the conservative assumption that the leakage through the catch basin is 1 g/day and that all the energy of the betas is absorbed by the oil. Since the volume of the catch basin is 423 cm3, the maximum permissible concentration in it is 3.1 x 1074 curies/cm>. Tests were performed on the prototype pump to investigate diffusion of radioactive gas up the pump shaft annulus, which is identical to the annulus in the MSRE fuel pump. The path for flow of gas through the annulus and the catch basin is shown in Fig. 2.18. The concentrations of Kr85 in the pump tank and the catch basin were determined with count rate meters. The results of the diffusion tests on the prototype pump are shown in Table 2.3. There was no detectable diffusion of Kr®® up the shaft annulus with purge rates as low as 100 liters/day when the concentration in the pump tank was as high as 3.54 x 107® curie/cm3. The limit of de- tection corresponded to a concentration of 0.95 X 10710 curie/cm? in the catch basin. Thus, the maximum permissible concentration of 3.1 x 1074 curie/cm3 should not be exceeded in the catch basin if the concentration in the pump tank is maintained at less than 11 curies/cm3. The concen- tration of radiocactivity in the pump tank is related to the purge flow, as shown in Fig. 2.19. The data indicate that the purge flow should be >1000 liters/day to ensure that the concentration in the pump tank will be <11 curies/cm3. A purge flow of 3300 liters/day down the shaft annulus and 4600 liters/day total through the pump bowl are available in the MSRE. 48 by I /// L o (1 o LJ ¥ O N ] TNT z 2+ OE - ASn|u_._._ QE= o C352 I NS = > 'S L83 ° 2533 oxg a O 1///// mb Q st X = a 2 ) o ' \fiK Shaft Annulus of MSRE rge Gas in Fig. 2.18. Diagram for Flow of Pu Prototype Pump: 49 Table 2.3. Back-Diffusion Tests Shaft Purge Catch Basin Purge Kr85 Concentration® in Pump Tank (liters/day) (1liters/day) (curies/cm3) X 10-© 406 L87 0.89 Lo6 487 0.89 430 Lo7 2.06 382 481 2.77 366 354 3.28 215 360 1.85 108 360 1.21 173 673 1.65 109 360 0.75 109 681 0.90 104 681 3.54 aConcentration values for catch basin were below the limit of detection. UNCLASSIFIED ORNL-DWG 63—6483 20 i | | =~ . 10 Mw POWER £ ° 100 % STRIPPING EFFICIENCY e \ 50 gpm BYPASS FLOW ] [e4] T \ 5 psig PUMP TANK PRESSURE (8] ) \ P 3 |_ = 5] \ = 8 \ 8 \ 2 \ g 4 D 4] [V 0 0 1000 2000 3000 4000 5000 PUMP TANK PURGE (liters /day) Fig. 2.19. Fission-Gas Concentration in Pump Bowl vs Purge Flow Rate. 50 Measurement of Undissolved Gas Circulating with the Molten Salt. The MSRE fuel-pump tank is equipped with a spray ring and uses a helium purge in conjunction with it to remove xenon poison from the circulating fuel. The spray flow of about 50 gpm churns the pump-tank liquid sufficiently that some bubbles of blanket gas enter the pump intet along with bypass flow. Under these circumstances the circulating salt carries undissolved gas in it at all times. A program was started to measure the concentration of undissolved gas in the circulating salt in the prototype pump loop. A 10-curie cesium source was placed on one side of the inlet pipe to the pump where the flow- ing salt is at low pressure, and a detector was placed on the opposite side. The intensity of the radiation at the detector was expected to in- crease exponentially with a decrease in the density of the flowing mixture of salt and gas. Under steady operating conditions, undissolved gas should be distributed nearly homogeneously in the flowing salt and should be detectable as a decrease in apparent density of the fluid. Preliminary test results showed that the output signal from the detector decreased, as expected, with a decrease in salt temperature and a corresponding increase in salt density but that there is need for additional calibration data and study of the data. The device was sensitive to changes in density of ~0.1%, as determined by changing the temperature of the salt with the pump stopped. When the pump was started, an apparent decrease in density of the salt was noted. As shown in Table 2.4 the apparent change in density was equivalent to the entraimment of 1 to 2 vol % of gas and reached a steady value about 2 min after the pump was started; and MSRE pump con- ditions of 1200 rpm, 1200 gpm, and 1200°F salt temperature were attained. : Investiéations are continuing to obtain a calibration of the appara- tus with and without molten salt in the loop and to determine how the con- centration of undissolved gas is affected by salt temperature, pump speed, and gas-blanket pressure. Table 2.4. Undissolved Gas Concentration Tests Test Data Detector Output Undissoclved Gas Test N Change for 1% Detector Output Change in Circulating Salt €8s O- Density Change Flow Stopped Flow Started Flow Stopped Flow Started (divisions) (1200 to O gpm) (O to 1200 gpm) (1200 to 0 gpm) (O to 1200 gpm) (divisions) (divisions) vol % vol % 1 25 Ly 39 1.76 1.56 2 25 by 40 1.76 1.60 3 24 4y : 42 1.83 1.75 b 24 41 h3 1.71 1.79 5 26 43 Ly 1.65 1.69 6 26 45 42 1.73 “1.62 7 26 L5 L6 1.73 1.77 Continued High-Temperature Operation. The three modifications made to the test facility at e conclusion of the 4700-hr run included in- stallation of the MSRE fuel-pump support (Fig. 3.20), a buoyancy-type con- tinuous liquid-level indicator in the pump tank, and an MSRE disconnect flange in the pump-tank off-gas line. During the modification, fresh salt was added to the system. The present composition was calculated to be -ZrF,-ThF,-UF, (67.9-24.9-5.3-1.0-0.9 mole %). pump support provides for limited movement of the pump along each of three mutually ndicular axes and was designed to impose relatively small forces and moments on pump suction and discharge nozzles. The pump tank was mounted on the support, and the loop was anchored at one point and supported at two other points with spring hangers. The anchor point was chosen so that the displacements of the pump along one horizontal axis and the vertical axis would be similar to, but smaller than, those which the reactor fuel pump will experience. Upon heating the test loop to leODF, the pump tank moved in the hori- zontal direction chosen but did not move in the vertical direction. Two tests were made to investigate why the support would not provide for the UNCLASSIFIED PHOTO 63472 Fig. 2.20. MSRE Fuel-Pump Support. 52 vertical displacement of the pump. The loop pipe anchor was removed, and the pump was lifted with an overhead crane (1) with the counterbalance springs, which provide for the vertical displacement set at zero force and (2) with these same springs set to provide for 8600 1b of upward force. The difference in the total forces required to 1lift the pump in these two tésts and the signs of rubbing noted on lifting-link components indicated the existence of unwanted rubbing forces in the components. The support was removed for further inspection and testing and for alterations to re- move the cause of sluggish operation in the vertical direction. It was replaced with the rigid supports used previously. Pump Inspection and Assembly Prior to Current Operation. The pump rotary element was closely inspected following the 4700 hr of operation. There was evidence of o0il leakage past the copper ring gasket at the joint between the shield plug and the bearing housing. This leakage 'had drained down the outside surface of the shield plug and formed a black carbon deposit over the lower two-thirds of the shield plug. The impeller was also discolored with a black deposit. Droplets of fuel were found on the lower end of the shield plug. A small deposit of fuel was found in the lower reaches of the annulus be- tween the shaft and the shield plug. This deposit was probably formed during periods when the purge rates in the shaft ammulus were low. The ring-joint gasket groove in the pump tank was discolored along approxi- mately 1 in. of the periphery, possibly indicating a leak. Otherwise, the pump rotary element was found to be in good condition. The pump was assembled with new preloaded bearings, and the shaft seals were replaced. Shaft deflection and natural frequency measurements were made, indicating that the preloaded bearings increased the shaft stiffness by approximately 15%. The new shaft seals contain bellows assem- blies with the Graphitar nose pieces attached by a clear baking varnish. Cold shakedown of the shaft seals was performed for 100 hr prior to installation of the rotary element in the prototype pump loop. There was no measurable leakage from the lower seal; the upper seal leaked at a rate of 44 cm3/day. The pump rotary element prior to installation in the loop is shown in Fig. 2.21. PKP Fuel Pump High-Temperature Endurance Test This pump testl® was interrupted by a failure in the electrical in- sulation of ‘the pump drive motor. The pump had operated continuously at 1950 rpm for 9816 hr, circulating molten salt LiF-BeFp-ThF,-UF, (65-30=-4=-1 mole %) at 1225°F and 510 gpm. Insulation on the stator and rotor windings was found to be burned and was replaced. The pump rotary element was disassembled, inspected, and reassembled for further endurance tests. The shaft lower seal was re- placed with one having the Graphitar nose piece attached to the bellows assembly with a clear baking varnish. Cold shakedown will be performed on the rotary element before it is installed in the hot-test facility. 53 UNCLASSIFIED PHOTO 62042 Fig. 2.21. Prototype Pump Rotary Element Without Drive Motor. 54 Test Pump with One Molten-Salt-Lubricated Bearing This pump'® was placed in operation at 1200 rpm, circulating the salt LiF-BeF,-UF, (62-37-1 mole %) at 1225°F and 100 gpm. The test was shut down after 58 hr of operation, when high current to the drive motor inter- rupted the electric power. Disassembly of the pump revealed that the im- peller and shaft, respectively, had jammed against the volute and molten- salt bearing. It is believed that the incident was caused by the loss of two of the four fulcrum pins, which provide for the angular displacement in the gimbals mount that supports the bearing. The molten-salt bearing was damaged and is being replaced prior to further test operation with the pump . Lubrication-Pump Endurance Test The lubrication pump'” was shut down after operating 7344 hr at 3550 rpm while circulating turbine-type oil at 160°F and 70 gpm. A short had developed in the insulation in the stator windings of the motor. The motor was rewound twice (it shorted after approximately 1 min of operation after the first rewind), and the pump was placed back in operation and has operated 250 hr at the above conditions. Fuel Pump MK-2 Degsign of a new fuel pump for the MSRE wag initiated. It differs from the present design in that the pump-tank volume available for thermal expansion of the fuel will be 6.3 ft>, compared to the present 1.9 ft3. The hydraulic design of the impeller and volute will not be changed, but the length of the shaft will be increased approximately 8 in.; the diam- eter remains at 3 in. The pump tank will remain 36 in. in diameter, but its height will be extended to provide the increased expansion volume. The internal baffles that control the liquid behavior in the pump tank will be developed in a water test mockup, which is being fabricated. Instrument Development Single-Point Liquid-Level Indicator Testing of the prototype of a two-level conductivity-type level probe for use in molten salt continued. This test has been in operation for approximately one year. There has been no change in performance from that reported previously.l® Design drawings for the probes to be installed in the MSRE were com- pleted and approved; however, the design is being revised to incorporate secondary contaimment of the electrical leads. This secondary barrier is being obtained by the use of ceramic-to-metal hermetic seals on individual leads that penetrate the mounting head. " 55 The probability that the probe would be damaged or ruptured if the salt in the drain tanks were allowed to freeze and were subsequently melt- ed was studied. Results of this study indicate that damage is not likely if the normal precautions, necessary to prevent damage to other parts of the tanks, are observed. Pump-Bowl Liquid-Level Indicator Testing of two float-type level indicators, started in February 1962, is continuing.2®,20 During the 18 months since the start of testing, one differential transformer has been at lZSOOF, except for short periods when the temperature was raised to 1300°F or dropped to 1000°F. The other (mounted outside the furnace, but insulated) has operated continuously, with one end at 1000°F and the other at 300YF. Both indicators continue to operate satisfactorily. A third indicator of similar design, except that a hollow INOR-8 float is used instead of a solid graphite float, was constructed and in- stalled on the MSRE prototype pump test. Testing of this device began July 26, 1963, and initial performance was satisfactory. Testing will continue to determine the long-term reliability of this design and to ob- serve performance under dynamic conditions approximating those of the reactor system. Contingent on satisfactory performance of the pump-test-loop installa- tion, the ball-float level transmitter has been accepted for use in the MSRE system, and all necessary design drawings have been completed for its installation on the MSRE coolant pump bowl. Since extensive piping re- visions would be required, this type of transmitter will not be installed on the MSRE fuel pump bowl unless serious difficulties are experienced with the bubbler level system during precritical operation or unless the new pump, which is presently being designed to provide greater expansion volume in the pump bowl, is installed. Studies are being made to deter- mine the feasibility of adapting the present level indicator design to the new pump. The major problem to be resolved is that of increasing the range from 5 to 16 in. The feasibility of installing the float directly in the pump bowl instead of in an ocutboard chamber is also being investi- gated. Temperature Scanner Development testing of the scanner system on the level test facility was continued.2* The use of the scanner during the heating and cooling -of the test facility tanks and piping was demonstrated for operations per- sonnel. : In April, the scanner was moved and installed on the ETL where it was used in startup, operation, and shutdown. Standard recorders were retain- - ed as backup instruments during the demonstration. The demonstration lasted approximately one month. During this period, the scanner was inoperative several times because of switch and amplifier 56 failures. The switch failed owing to oxidation of the mercury caused by loss of the nitrogen purge gas. The causes of the amplifier failures were not determined. After approximately 900 hr of operation on the ETL, the test was terminated. After the test was discontinued, a report was issued by the loop oper- ator.22 The report stated that the scanner was a valuable tool for use in operating the loop. It was recommended that some recorders be used in con- junction with the scanner and that a signal marking device be provided to positively identify each thermocouple displayed on the oscilloscope. A signal marking system is being developed. At present it appears feasible to assemble a satisfactory, low=-cost marker system from commercially avail- able modules. The temperature scanning system has now been accepted for use on the MSRE. The MSRE system will have five 100-point scanner channels and will be flexible enough to permit blocking and transferral of thermocouples to other scanning channels, either individually or in groups of 25. All major components will have plug-in connectors. This feature will greatly . reduce downtime in the event of component failure. The design of this system has been completed. Components have been ordered, and panels are being fabricated. Thermocouple Development and Testing Ingineering Test Loop Thermocouples. Eight MSRE prototype surface- mounted thermocouples on the ETL facility continued to be checked periodi- cally for performance. All the thermocouples were still functioning proper- ly after 5000 hr of operation at temperatures up to 1200°F. Prototype Pump Test Loop Thermocouples. Ten MSRE prototype surface- mounted thermocouples on the prototype pump test loop continued operation at temperatures up to 1200°F. A1l the thermocouples had accumulated 5300 hr of satisfactory service when the loop was shut down for pump modifica- tions in May 1963. Bayonet Thermocouples. The testing of thermocouples in the drain-tank test facility for endurance under thermal shock in the temperature range 1350 to 200°F was completed. The failure of two more thermocouples at 2630 cycles plus previously reported failuresZ>® left only three of ten thermocouples still functioning. The bayonet tubes to which the thermo- couples were attached were removed from the rig and submitted to Metal- lurgy for examination. Several of the thermocouples were broken at one or more points along the portion of their length which had been subjected to the thermal shock. It is believed that some of this damage was in- flicted during the removal operation. Drift Test. The observed drift in the calibration of six Inconel- sheathed MgO-insulated Chromel-P — Alumel thermocouples, operating in air at 1200 to 1250°F, remained the same (f2°F) after 18 months, as was previ- ously reported. No further drift tests will be conducted with the agbove thermocouples since they were not made with material to be used in the 57 MSRE. Future tests will be conducted with thermocouples made with MSRE materials. Thermocouple End Seals. The testing of materials and techniques for use in sealing the ends of mineral-insulated thermoccouples and copper- tube-sheathed thermocouple extension cables was continued.Z% Experiments were conducted with Physical Science Corporation 0900 glaze compound to develop procedures for mass productionof this type of seal on the ends of mineral-insulated thermocouples that terminate at disconnects inside the reactor and drain-tank cells. Several satisfactory seals were made with this material under laboratory conditions. A shrinkable tube made by Rayclad Tubes was tested for use in sealing the ends of mineral-insulated thermoccuples located outside the reactor and drain-tank cells. Test seals made with this material were leaktight to helium at 100 psig. A typical thermocouple disconnect and extension cable assembly was prepared for testing the susceptibility to radiation damage. Closed-Circuit Television for Remote Maintenance Viewing The feasibility of using closed-circuit television viewing during remote maintenance operations at the MSRE is being investigated. Methods being considered include single- and two-channel sterec (three-dimensional) viewing and multichannel (one-dimensional) viewing, using two or more ‘cameras located to view the subject from different angles. Both types of stereo systems were installed and compared in a remote maintenance test facility. Operation of a multichannel one-dimensional system was observed at the Atomics International Santa Susano Site; results to date are inconclusive. Rach type of system has certain in- herent advantages and disadvantages. The single-channel system, which is manufactured by the Stereo-tronic Company, has the inherent disadvantage that the horizontal field of view is restricted and a high lighting level is required to obtain a satisfactory picture. The two-channel stereo system requires more equipment and is therefore more subject to failure. At present it appears that the two-channel system is more difficult to adjust and may produce more operator fatigue than the Stereo-tronic system. The one-dimensional viewing system has the inherent advantages of simplicity and lower cost but requires an unrestricted field of view from at least two directions. Evaluations of all three types cf systems is continuing. Final choice of the system to be used at the MSRE will be based on results of these tests and on remote maintenance cperations re- gquirements. High-Temperature NaK-Filled Differential Pressure Transmitter Development of a high-temperature NaK-filled differential pressure transmitter continued.®> This work is being done by the Taylor Instrument Company under the terms of a purchase agreement with ORNL. The major 58 problem during this report period has been the design of the high-tempera- ture INOR-8 diaphragm seal heads to satisfy ORNL requirements for weld quality and minimum material thickness without sacrificing performance characteristics. Major difficulties have been resolved, and design draw- ings are being prepared for ORNL approval. Bubbler-Type Liquid~Level Indicator As previously reported,&® one of the two bubblers undergoing develop- mental testing on the MSRE prototype pump test facility became plugged with salt due to leakage in an improperly installed fitting. Previous efforts to remove this plug had been unsuccessful. During a recent shutdown of the MSRE prototype pump test loop, the plugged bubbler line was explored with a flexible rod. It was found that the line was open down to the semicircular pipe in the pump bowl and as far along this line as the rod could reach. Since there was no gas flow through this bubbler when pressure was applied, it must be assumed that the dip leg is still plugged. There are two types of dip legs being tested in this pump bowl. One is an open pipe with a V-notch cut in the side of the open end. The other has a closed end with a l/8—in.-diam hole in the side of the pipe just above the end. The latter is the one that is plugged. Since the dip leg involved is inaccessible, the cause of plug- ging cannot be determined until the leg can be inspected, when the pump is dismantled. Design of the bubblers for the MSRE fuel-pump installation has been revised to use the V-notch construction instead of the l/8—in. hole. The design of the MSRE coolant pump bubbler could not be changed since fabri- cation of this pump has been completed. Very little operational data was obtained during this period because the dip leg that was still clear was needed for back-diffusion studies with krypton. This work is complete, and testing of the operational bub- bler has resumed. References 1. MSRP Semiann. Progr. Rept. Jan. 13, 1963, ORNL-3419, pp 17-23. 2. Ipid., p 2k. 3. Ibid., p 27. L. TIbid., pp 33-3L. 6. Ibid., pp 23-25; 10. 11. 1z. 13. 1h. 15. 16. 17. 18. 19. 20. 21. 2z. 23. 2k, 25. 26. 59 Ibid., pp 110-116. Tbid., p 31. MSRP Semiann. Progr. Rept. Aug. 31, 1961, ORNL-3215, p 57. Letter from D. R. Sears, H. Insley, and F. F. Blankenship to J. L. Crowley, Cold Trapping Behavior Encountered at the Access Port of FETL, MSR-63-29 (July 11, 1963). J. H. DeVan and R. B. Evans 1II, Corrosion Behavior of Reactor Materi- als in Fluoride Salt Mixtures, ORNL TM-328 (Sept. 19, 1962). P. G. Smith, Water Test Development of the Fuel Pump for the MSRE, ORWL, T™~79 (Mar. 27, 1962). Letter from R. Blumberg and J. R. Shugart to D. Scott, Recommendations and Results on the Mechanical Assembly of the MSRE Heater Electrical Disconnect, MSR-63-27 (July 1, 1963). MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 37. Tbid., p LO. MSRP Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, p 66. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 4O. Ibid., pp L4O-41. Ibid., pp 41-42. MSRP Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, p 68. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORWNL-3419, pp Wk-L45. Memo from J. L. Crowley to R. B. Briggs, 'Use of Temperature Scanner on ETL," May 7, 1963. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 47. Ibid., p 47. Tbid., pp 45-46, Fig. 2.23. Ibid., pp 4h-=b45. 3. MSRE REACTOR ANALYSIS Nuclear Characteristics of Core The nuclear characteristics of the reactor were calculated with each of the three fuel salts described in Table 3.1, 8Salt A is a thorium- containing fuel originally proposed for the MSRE. Salt B contains a min- imum of highly enriched uranium and is representative of the core fuel for a two-region breeder. 8Salt C contains sufficient U232 to increase the uranium concentration in the fuel to 0.8 mole %; at this concentration there should be 1little concern about the chemical stability of the fuel. Salt C will be used in the first operation of the MSRE. In the calculation of critical uranium concentrations and flux dis- tributions, the reactor was represented by a multiregion, cylindrical model, similar to one described previously.l Critical concentrations were computed with MODRIC, a one-dimensional, multigroup code. Resonance cross sections for the MODRIC calculations were generated for each core composition with the GAM-1 code.? Thermal-neutron cross sections for MODRIC were also computed for each separate case, taking into account the thermal spectra and spatial disadvantage factors. In addition to the critical concentrations, the MODRIC calculations gave relative neutron fluxes for each of 34 groups and two-group constants Table 3.1. MSRE Fuel Salts for Which Detailed Nuclear Calculations Were Made Fuel Type A B C Salt composition (mole %) LiF™ 70 66.8 65 BeFs 23.7 29 29.2 ZrF, 5 L 5 ThF, 1 0 0 UF, (approx) 0.3 0.2 0.8 U composition (at. %) | UE34 1 1 0.3 U==3 93 93 35 U=38 1 1 0.3 U=38 5 5 6l L4 Density at 1200°F (1o/ft3) 1kkL 134 143 %99.9926% 1i7, 0.0074% 1i®. 60 Table 3.2. Nuclear Characteristics of MSRE with Various Fuels 6l (1200°F, graphite density 1.86 g/cm®) Fuel A Fuel B Fuel C Uranium concentration (mole %) Clean, non-circulating 235 0.29 0.18 0.29 Total U 0.31 0.19 0.83 Operatinga 235 0.34 0.20 0.35 Total U 0.36 0.21 0.89 Uranium inventorjb (kg) Initial criticality 235 79 L3 11 Total U 85 52 220 Operatinga U=33 91 25 92 Total U 98 59 230 Thermal-neutron flux® (neutrons cm™2 sec™t) Max imum 3.3 x 1013 5.6 x 1012 3.3 x 1013 Average in graphite- . 1.4 x 1013 2. x 1013 1. x 1012 moderated regions Average in circulating fuel 4.0 x 1012 6.8 x 1012 4.0 x 1012 Reactivity coefficientsd Fuel temperature (°F-1) 3.0 x 1075 =5.0 x 10~3 —3.3 x 107> Graphite temperature (°F~1) 3.4 x 107 4.9 x 107> ~3.7 x 10°° Uranium concentration 0.25 0.30 0.18¢€ o.21% Xel35 concentration in core -1.3 x 10° —2.0 x 10°8 —-1.3 x 10°® (atoms/barn-cm)~* Xel35 poison fraction —0.75 ~0.69 —0.75 Fuel-salt density 0.19 0.35 0.18 Graphite density . 0.76 0.73 0.77 Prompt-neutron lifetime (sec) 2.3 x 107% 3.5 x 107* 2.k x 107¢ ®Fuel loaded to compensate for 4% Bk/k in poisons. Ppased on 73 £t2 of fuel salt at 1200°F. At operating fuel concentration, 10 Mw. QAt initial critical concentration. ficients for variable x are of the form: cients are of form: (x/k)(dk/dx). e . . . ey - Based on uranium isotopic composition of clean critical reactor. fBased on fully enriched uranium. (1/x)(3k/ox ). Where units are shown, coef- Other coeffi- 62 for each region of the reactor model. The two-group constants were em- ployed in calculations with Egquipoise-3, a two-dimensional, two-group code, which gave various coefficients of reactivity, reactivity importance func- tions, neutron lifetime, and absolute neutron fluxes. Coefficients of reactivity for uniform changes in fuel and graphite temperatures were computed for a one-region model of the core, taking into account lattice effects and the changes in neutron spectrum with fuel or graphite temperature. A summary of the predicted nuclear characteristics of the MSRE is given in Table 3.2. ' Control Rod Worth The reactivity worth of the control rods was calculated with each of three fuels in the core. Method of Calculation The Equipoise-3 program was used in the control rod calculations. The cross section of the core, including the control rods, was represented in x-y geometry as shown in Fig. 3.1. Dimensions were chosen to give the same transverse geometric buckling in the square core model as in the actual cylindrical core. In the calculation it was assumed that the 1.08- in.-0D Gdz05-Al-04 poison cylinders are black to thermal neutrons and transparent to neutrons of higher energy. Transmission through the thim- bles and across the air gap was calculated by a P-1 approximation. Results The total worth of the three control rods inserted all the way through the core wasg calculated for the clean, critical core at 1200°F with fuels A, B and C. The worth is different for each fuel because of differences in the neutron diffusion properties of the core. The worths of various combinations of fully inserted or fully withdrawn rods were calculated for fuel A. Results are shown in Table 3.3. Fractional worth as a function of distance inserted was calculated and reported earlier.® 63 UNCLASSIFIED ORNL-DWG 63 -6484 A X 5.03 | 1.41 | 1.00 | 154 | 1.00 | 1.41 | 5.03 |—-ax 12 4 6 8 40 (2 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 43 { 2 4 T 6 . 21.78in, 8 12 14 2.21in. i6 + 8 2 3 1.57in. 2 4 22 243in, 24 4 26 { S 1.57in. 2 4 30 2.21in 32 34 36 42 43 B o no }4———‘4 @ 5 Fig. 3.1. Cross-Sectional Model of MSRE Core Used in Equipoise Calculations of Control Worth. AX is mesh size in centimeters, and S is sample holder, Table 3.3. Worth of Fully Inserted Control Rodsa Worth Fuel Rods In Rods Out (% Sk/k) C All three None 5.7 B All three None 7.6 A All three None 5.6 1 and 3 2 b4 2 and 3 1 h.1 1 2 and 3 2.4 2 1l and 3 2.3 %Rod 2 is opposite sample assembly. 64 Reactivity Shimming The control rods are required to make the reactor subcritical at times and to compensate for the changes in reactivity which occur during a cycle of startup, power operation, and shutdown. Slow changes in reac- tivity due to growth of corrosion products and low-cross-section fission products will be compensated by increases in uranium concentration. The rod worth required for shimming the reactivity depends on the type of fuel. Table 3.4 lists predicted rod shim requirements with fuels A, B, and C in the reactor. ' Table 3.4. Rod Shim Requirements Effects (% ®k/k) Cause Fuel A TFuel B Fuel C Ioss of delayed neutrons 0.3 0.3 0.3 Entrained gas 0.2 0.b 0.2 Power (0-10 Mw) 0.06 0.08 0.06 Xel35 (equilibrium at 10 Mw) 0.7 0.9 0.7 Samarium transient 0.1 0.1 0.1 Burnup (120 g of U=23) 0.03 0.07 0.03 Total 1.k 1. 1.k A loss of delayed neutrons occurs when the fuel circulating pump is started. At the same time, gas is entrained in the circulating fuel by the action of the stripper in the pump bowl. The gas effects in Table 3.4 assume 2 vol % gas at the pump suction or 1.2 vol % in the core. - The power shim is the amount of reactivity required if the core out- let temperature is held constant. (The shim would be 0.2 to 0.3% 6k/k if the mean of the inlet and outlet were held constant.) The xenon poisoning depends on the flux, the stripper efficiency, the xenon diffusivity in the graphite, and the fuel-graphite xenon transfer (see Fig. 2.15, Sec 2). The tabulated values of the xenon effect were calculated for a stripper efficiency of 50%, xenon diffusivity in the graphite of 1.5 x 107° ftZ/hr, and a mass transfer coefficient of 0.08 ft/hr. Considering the uncertainties in these values, the xenon effect is expected to be between one-third and twice the values tabulated. The reactivity effect of sm'*® and similar high-cross-section, long- lived fission products levels off at about 1.1% Bk/k at high power. If the flux is reduced to zero after high-power operation, the poisoning effect increases, because of short-lived precursor decay, by 0.1% ak/k. The poisoning decreases by the same amount when the power is again raised. 0 65 The last item in Table 3.4 (burnup) is the reactivity change due to this cause between normal fuel additions. Any of the control rods can be used as a servo-driven regulating rod while the other two are controlled manually. All three rods are used for shimming and all can be used for shutdown. 1In order to provide the max- imum shutdown margin, the uranium concentration of the fuel is limited to barely that required for full-power full-poison operation with all rods at the upper limits of their operating ranges. At the upper limits, which are set to avoid the low sensitivities at extreme withdrawal, the rod poison amounts to about 0.5% 8k/k. The useful worth of the rods, from full insertion to the upper end of their operating ranges, 1s therefore less (by 0.5% ®k/k) than the total worth (Table 3.3). The minimum shut- down margin provided by the rods is the difference between the useful worth of the rods and the shim requirements {(Table 3.4). Minimum shut- down margins for fuels A, B, and C are 3.7, 5.2, and 3.8% 8k/k respectively. These margins are equivalent to reductions in critical temperatures of 580, 530, and 5%0°F respectively. A typical program of rod positions during an operating cycle 1is shown in Table 3.5. While the core is being filled, the rods are partially withdrawn to provide rapid shutdown if abnormal conditions should appear. The rods are fully inserted before the pump 1s started so that no 'cold- slug" accident could make the reactor critical. As the reactor is operated at power, the two manually controlled shim rods are kept well above the regulating rod to avoid a reduction in the sensitivity of the regulating rod by '"shadowing' by the other two rods. Table 3.5. Rod Positions During Typical Operation, Fuel C Rod Position Rod Poisoning (in. Withdrawn) (% dk/k) Condition Regulating Shims Regulating Shims Total Filling core (1% subcritical) 28.5 28.5 0.9 1.9 2.8 Starting fuel pump 0 0 1.9 3.8 5.7 Going critical, no Xe, peak 28 .4 S 1.2 0.1 1.3 om, no burnup At 10 Mw, no Xe, peak Sm, no 29.4 5t 1.1 0.1 1.2 burnup At 10 Mw, equilibrium Xe and 39.3 5 0.6 0.1 0.7 Sm, no burnup At 10 Mw, equilibrium Xe and 39.9 P 0.5 0.1 0.6 Sm, 120 g of U233 burnup 66 Inherent Neutron Sources The MSRE fuel salt will always contain an inherent source of neutrons. Alpha particles from the uranium interact with the beryllium and fluorine to produce neutrons, and there is also a contribution from spontaneous fission. After high-power operation the inherent source will be much stronger because of photoneutrons from the beryllium, produced by the fission product gamma rays. Predicted source strengths are given in Table 3.6. The spontaneous fission source is mostly from UZ3® while the U®®% alphas are responsible for nearly all the g-n source.® Thus these sources differ for each fuel. The photoneutron neutron source depends on the fission product activity - and the amount of beryllium in the core, which is about the same for each fuel. The presence of the strong internal source contributes materially to the nuclear safety.® - Table 3.6. Inherent Neutron Source in Core™ Strength (neutrons/sec) Source Fuel A Fuel B Fuel C a-n ‘ 5x 105 3 x 105 L4 x 10° Spontaneous fission iNe) 20 700 Photoneutron 8 nP L x 10° 14 8 x 108 7T 4 2 x 108 30 4 3 x 107 an Effective"” core, containing 25 ft3 of fuel. For o-n and fission sources, clean critical con- centrations of uranium were assumed. bTime after operation at 10 Mw for one week. Biological Shielding A review of the MSRE biological shielding design was completed. Special attention was given to irregularities in the annular shield around the reactor cell, and it was found that several areas will re- quire additional local shielding. Most of this is required to lower the radiation level in the blower house, where the dose rate during 10-Mw i» 67 operation would be ~140 mrems/hr with no additional shielding. Stacked concrete block can be placed conveniently to eliminate the contribution from the reactor cell, but levels in the blower house during 10-Mw oper- ation will range up to about 30 mr/hr from the coolant salt in the radi- ator and piping. A block wall can be added around the blower house if necessary. The solid portion of the overhead shield will limit the dose rate to less than 2.5 mrems/hr except directly over the core, where the estimated dose rate will be 16 mrems/hr. Much higher dose rates could exist over points where gaps between blocks overlap. Polyethylene and steel strips in the gaps and stacked concrete block can be used where needed to pre- vent excessive radiation through the top shield. References 1. MSRP Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, p 22. 2. G. D. Joanau and J. S. Dudek, GAM-1: A Consistent Py Multigroup Code for the Calculation of Fast Neutron Spectra and Multigroup . Constants, GA-1850, June 28, 1961. ' 3. MSRP Semiann. Progr. Rept. Feb. 28, 1961, ORNL-3122, p 73. 4. P. N. Haubenreich, Inherent Neutron Sources in Clean MSRE Fuel Salt, ORNL TM-611 (in preparation). 5. S H. Hanauer, Nucl. Safety 4(3): 52 (March 1963). Part 2. MATERTALS STUDIES L. METALLURGY Heat Exchanger Fabrication ) MSRE Primary Heat Exchanger Brazing After the tubes in the MSRE heat exchanger were welded to the tube- sheet at the front face, they were brazed into the tubesheet to increase the strength of the Jjoint and to reduce the probability of a leak. The brazing was done at the Wall-Colmonoy Corporation by a methodt developed for this purpose and tested on the sample heat exchanger.® Preplaced 82 wt % Au — 18 wt % Ni brazing alloy was used in ring form. The tube bundle was sealed in an all-welded retort to minimize the possibility of air leaks, and the system was evacuated and then purged with dry hydrogen (-80 to —85°F dewpoint) at a rate of 145 cfh. The use of a large, gas-fired pit furnace enabled the entire unit to be heated to braz- ing temperature, thus minimizing distortion due to differential thermal expansion. The brazing cycle was similar to that used on the sample heat ex- changer. The temperature was raised at a rate of about 300°F/hr to a thermal equalization temperature of 1650°F. After about 20 min the temper- ature was again increased, and that at the tubesheet end of the heat ex- changer was held between 1850 and 1885°F for 1 hr. A slow furnace cool at a rate of about 300°F/hr was employed. Visual inspection of the completed tube bundle showed that all com- ponents were bright and all visible joints had good fillets (see Fig. 4.1). After the tube bundle was welded into its shell, both a helium~leak test and an 800-psi hydrostatic test were conducted. No leakage was found. Ultrasonic Inspection of Tube Joint Brazing The %15 tube-to-tubesheet joints in the MSRE heat exchanger were tested ultrasonically for unbonded areas using the Lamb-wave technigue and mechani- cal devices described previously,3 and the heat exchanger was found accepta- ble for the MSRE. A number of these Jjoints contained areas which gave indi- cations comparable to, or larger than, the signal from a reference 5/52-in.— diam flat-bottomed hole. Although some of these indications covered ex- tensive areas of the joints, analysis of the amplitudes indicated that the signals were not large, continuous voids but possibly were due to groups of small discontinuities. Since indications of this type and size had not been observed in the earlier work, there was no metallographic reference to which they could be compared. An attempt is being made to establish such a reference by braz- ing joints with discontinuities of approximately the right size. This reference should allow a better interpretation of the data and of the braze condition in the heat exchanger. 71 UNCLASSIFIED PHOTO 39431 Fig. 4.1. Completed Tube Bundle Mounted in Inspection Fixture. In- set shows appearance of braze fillets at tube-tubesheet juncture. 73 Mechanical Properties of INOR-8 Thermal Fatigue The thermal-fatigue characteristics of INOR-8 have been determined from the conventional Coffin-type test. The low-cycle-fatigue equation can be used to describe the plastic-strain — fatigue relation: AepN =K . (1) For INOR-8 the constants a and K were evaluated as 0.88 and 0.52 respective- ly. Thermal-fatigue data for this alloy reported earlier® were in error. The correct data are shown in Fig. 4.2. It has been common practice to assume that the value of a is constant, being 0.5 for all materials. Recent results of tensile tests at elevated temperature and the thermal-fatigue results show a to be related to the strain-hardening exponent n or, a = 1 3 ’ (2) where n is determined from the stress-strain relation. The thermal-fatigue characteristics show good agreement with the isothermal-fatigue data avail- able on the alloy. Tests results at maximum temperatures of 1300 and 1600°F show no appreciable difference in the Ae_ or N_ relation. Hold times of up to 40 min at the maximum temperature shoW some beneficial effect on the plastic-strain — fatigue behavior. Relaxation of elastic stress did occur during the hold time, and the plastic-strain value per cycle was therefore increased. As predicted by Eq. 1, the number of cycles was reduced by the increased plastic-strain range. Equation 1 does not consider the manner in which the plastic strain is produced. Thus, the time-dependent effects cannot be shown by this equation. A significant part of the plastic strain may result from relax- ation; therefore, the thermal-fatigue merit of a material cannot be based solely on the evaluation of the constants of Eg. 1. In order to coampare the thermal-fatigue behavior of INOR-8 to that of other common reactor ma- terials, the rapid-thermal-cycling conditions were selected. The range of temperature cycles to failure for INOR, Inconel, and type 304 stainless steel is shown in Fig. 4.3. INOR-8 is shown to be superior to both the other materials by a substantial amount. These tests were per- formed under similar conditions; the specimen geometry, testing machine, cycle time, environment, and maximum temperature for all tests were nearly identical. When hold time is added at the maximum temperature, the curves are displaced to the left. ILimited data on INOR-8 indicate a displacement of approximately one log cycle for a 40-min hold time at 1300°F. Data showing hold-time effects were not available for all the materials. 74 UNCLASSIFIED - ORNL-DWG 63-6485 0 e HASTELLOY-N [0 S < N £ N o . N 2 N <1 - r 10 2 N z X g [ n o O ) = ] 0 o || ]e q . ‘\ D z 9N o N § o3 ® 1600°F 7ny N o NU % o 1300°F TML\X \" > a ™~ - '\\ 0 5 N \\ \.. - AN 4 10° 10’ 10° 10> 10 N, CYCLES TO FAILURE Fig. 4.2. Strain Fatigue of INOR-8. UNCLASSIFIED ORNL-DWG 63-6486 1200 \\\ \ N 1000 PN ™ \ \\ \ [ \\ N w N \ 2 800 ™ \\\“*~~|N0R-8 B I V o \ Ll \ o ™~ 2 600 <{ \ ™~ o \\ o B ~L S INCONEL < THERMAL FATIGUE N E 400 1300°F T, ) 304 STAINLESS e MAX STEEL 3 { min PER CYCLE 200 0 q 5 2 3 o) 2 5 10 2 5 10 2 5 10 AN, CYCLES TO FAILURE Fig. 4.3. Comparison of Range of Temperature vs Cycles to Failure of INOR-8, Inconel, and 304 Stainless Steel. 75 Postirradiation Tensile Testing A program was initiated to examine the postirradiation tensile proper- ties of INOR-8. Subsize tensile specimens which were 0.125 in. in diameter over a gage length of 1.125 in. were prepared from INOR-8 heat 5081. These specimens were annealed at 2150°F for 15 min and were quenched rapidly in argon to simulate the condition of the INOR-8 in the MSRE. Twenty specimens will be irradiated in ORR Experiment 109 at the temperature range 1100 to 1400°F and a dose of about 7 x 10F° nvt (>1 Mev). The postirradiation tensile testing will be conducted at room temperature and within the irradiation temperature range. Both strength and ductility of the irradiated specimen will be compared to similar properties of unir- radiated control specimens. Evaluation of MSRE Graphite Graphite bars (grade CGB) selected from material produced for the MSRE moderator are being evaluated for compliance with specifications and to establish properties data useful to the MSRE. This testing is for the purpose of establishing the general level of the property values and is not part of the quality control called for in the specification. Most of the tests were made on specimens from three graphite bars. Chemical Composition and Oxygen Content A composite sample from the three bars was analyzed for chemical puri- ty. The results of the analyses and the values specified for the MSRE are given in Table L4.1. Table 4.1. Chemical Analysis of Grade CGB Graphite Specified Maximum Analyzed (wt %) (wt %) Ash 0.07 0.0005 Boron 0.0001 0.00008 Vanadium 0.01 0.0009 Sulfur 0.005 0.0005 A measure of the tenaciously held oxygen of grade CGB graphite was determined by placing specimens in a closed system, evacuating the system to <10™° torr at room temperature, and then measuring the STP volume of carbon monoxide evolved from the graphite at 1800°C. The oxygen content of three lots (eight specimens) was determined to be 6 cm® of CO (STP)/ 100 cm® of graphite by this procedure. This was well under the maximum of 30 cm® of CO (STP) /100 cm® of graphite permitted by the specification. 76 Thermal Cycling of Salt-Impregnated Graphite Tests were continued® to determine if salt-filled cracks in CGB graphite would be propagated by thermal cycling through a temperature range that includes the melting point of the salt. In previous work> impregnated specimens were not significantly altered by 100 cycles be- tween 390 and 1300°F. Similar results were obtained on additional speci- mens that were cycled to 1800°F. The volumetric expansion of salt on heating from the liquidus temperature (930°F) to 1830°F is approximately 14%. Test specimens at the higher temperature were impregnated with LiF- BeFo-ThF,-UF, (67-18.5-14-0.5 mole %) at 1300°F and 150 psi to an equiva- lent graphite bulk volume of 0.1%. A radiographic inspection indicated that the salt was primarily confined to the cracks in the graphite. BREach impregnated specimen was subjected to a high temperature cycle having the rate of rise summarized in Table 4.2. Table 4.2. Rate of Temperature Rise for the Thermal Cycles of Salt-Impregnated Graphite (Grade CGB) Time to Reach Temperature Maximum Temperature (min) ) Specimen 1984 Specimen 198-5 590 0 0 1500 3.7 3.4 16830 36.0 £3.0 There were no detectable changes in the graphite specimens or their cracks as a result of these high-temperature thermal cycles. They exceeded the maximum temperature (1350°F) expected in the MSRE. The maximum pre- heating rate® of the core in the MSRE is expected to be 54°F/hr. On this basis the slowest of the high-temperature thermal cycles was more than 200 times faster in reaching 1300°F than the maximum preheating rate. The results of the preceding tests indicate that thermal cycles of the MSRE should not damage the graphite. Irradiation Effects on Grade CGB Graphite The effect of neutron irradiation on the mechanical and physical properties of grade CGB graphite is being investigated. Fifty specimens have been irradiated in the ORR to dose levels from 2 to 5 x 107° nvt (E > 0.18 Mev). The irradiation temperatures of the specimens varied with the flux and were from 700 to 1050°F. The effect of this irradiation i on the dimensional stability, electrical conductivity, thermal expansion, fracture stress, fracture strain, and modulus of elasticity will be de- termined. The electrical resistivity of grade CGB graphite was determined at 27°C in the directions perpendicular to and parallel with the extrusion direction (length) of the bar, on control specimens for the irradiation test. The thermal-conductivity values were calculated from the electrical resistivity’ and are compared with those originally estimated for the MSRE graphite. The results are summarized below: Perpendicular Parallel Electrical resistivity (pohm-cm) 1256 644 Thermal conductivity (Btu £t~% sec™t £t™2 °F~1) Calculated 60 117 Original estimation L5 80 This comparison indicates that the thermal conductivity of the graphite may be better than was originally expected and that the graphite is slightly more anisotropic than was expected. Actual measurements on the irradiated graphite will be necessary to determine its thermal conductivity as functions of temperature and dosage. Removal of Oxygen from Graphite With Molten Salts The use of molten NaF-ZrF,-UF. (50-46-4 mole %) to remove oxygen from a moderately permeable graphite, grade R-0025, was demonstrated under con- ditions similar to those proposed for purging the MSRE. Work was concen- trated on this salt because it appeared to be more effective than the LiF-Bels-UF, or LiF-BeFs-ThF,-UF, salts for removing oxygen from graphite.® In early work the ratio of the volume of the molten salt to the bulk volume of a graphite crucible was 1:27, and the purging salt was allowed to freeze in the crucible before it was removed. There was some scatter in the results from duplicate tests, presumably due to incomplete removal of the frozen salt and/or to the relatively small quantity used. The proposed flush-salt operation of the MSRE differs from this method in that the salt will be introduced and removed from the reactor in the molten state; thus the ratio of the total volume of the molten salt to the bulk volume of the graphite in the reactor will be 1:1. The larger guanti- ty of salt and its removal in the molten state as proposed for the MSRE offered more effective conditions for cleaning the graphite than did the conditions of the early laboratory tests. 78 Approximate MSRE conditions were produced on a laboratory scale with four crucibles of grade R-0025 graphite. 1In each test the graphite cruci- ble was completely submerged in the salt for 500 hr at 1300°F. The 500-hr period was chosen because this was the estimated time that the MSRE flush salt would remain in the reactor. The purged crucibles were then charged with the oxygen-sensitive LiF-BeFs-UF, (62-37-1 mole %) salt in an inert atmosphere and held at 1300°F for 4000 hr. (This salt readily precipitates uranium dioxide in the presence of oxygen.) Radiographic tests did not detect any uranium oxide precipitate forming. Similar tests will be made with R-0025 and CGB graphite and the proposed MSRE coolant (and purging) salt LiF-BeFs (66-34 mole %) as soon as this salt is available. : Sintering Characteristics of Gds03-Al1-505 A study was completed to develop procedures for fabricating thin- walled cylinders of Gdz0Os that contain up to 30 wt % Al.05. These cylin- ders are of interest as control rod elements for the MSRE, and their composition was designed to withstand moisture in the atmosphere at temper- atures up to 1400°F. The Al,05 was added to minimize hydrolysis and subse- quent deterioration of the Gd,0s. In preliminary work,® severe distortion resulted from heating cold- pressed mechanical mixtures of T0% Gds0s5 — 30% AloOs to 1650°C, indicating the formation of a low-temperature-melting compound, identified by x-ray diffraction analysis as the primary perovskite-type phase GdAlO5. The distortion was minimized by the use of prereacted powder;lo however, some flairing of the ends of the cylinders still occurred owing to frictional restraint during sintering. A technique was investigated in which a pressed specimen of prereacted material was supported on a similar, but shorter, "dummy" specimen so that the frictional restraint was sustalned by the supporting member. By this technique it was possible to sinter at temperatures up to 1645°C and to ob- tain undistorted pieces with bulk densities of 5.1 to 5.2 g/em®. Densities to 5.26 g/cm® were attained, with minor dimensional changes, by thermal cycling the parts between 1635 and 1660°C; however, this technique is not considered practical for making large quantities of the material. A prereacted mixture of 80 wt % Gds0z — 20 wt % Al,05 was also investi- gated using similar techniques, except that the mixture was reacted initial- ly at 1750°C rather than at 1650°C. No deformation or tendency to weld was observed in specimens sintered at temperatures as high as 1750°C for 1 hr. An x-ray diffraction analysis of these materials revealed GdA105; perovskite and an, as yet, unidentified phase. The results of this work indicate that shapes of controlled size and density can be fabricated by the use of standard cold-pressing and sinter- ing techniques if prereacted powders are used. Distortion due to frictional restraint of the shrinking parts during the sintering operation can be mini- mized by the use of supporting members also made of prereacted powders. 10. 79 Reflerences R. G. Donnelly and G. M. Slaughter, Fabrication of Heat Exchanger Tube Bundle for the Molten-Salt Reactor Experiment, ORNL-3500, in press. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, pp 65-68. MSRP Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, pp 90-91. ITbid., pp 95-96. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, pp T1-Tk. Tbid., p 123. The Industrial Graphite Engineering Handbook, p 5.B.01.06, National Carbon Company. MSRP Progr. Rept. Feb. 28, 1961, ORNL-3122, p 97. MSRP Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, pp 98-99. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, pp T76-T79. 5. RADIATION EFFECTS Postirradiation Examination of Assembly ORNL-MTR-L7-U The effects of the Fo evolved in frozen fuel in response to decay energy have been further explored. No evidence has been found of Fs or of significant amounts of CF, under operating conditions. Xenon Recovery from Graphite Cores Assembly ORNL-MTR-47-4, which has been described previously,!™3 contained capsules of INOR-8 enclosing graphite and MSRE fuel. In four of the capsules, l/2-in.—diam cylindrical cores of graphite (5 g) were immersed in 25 g of MSRE-type fuel. After an exposure of about 107C nvt in the MTR, the capsules were punctured, and the cover gas was collected for analyses; much gaseous Fs was found. In conformance with results from earlier irradiations, in which gaseous Fs must also have been pre- sent, there was little if any of the expected xenon from fissioning. This was attributed to the formation of solid compounds such as XeFp, XelFsXeF,, and XeF,, which have relatively low vapor pressures. After the graphite cores were removed from capsules 45 and 12 and examined, they were stored in screw-cap plastic bottles for five months. The necessary manipulations involved many hours of exposure to hot-cell air. On removal from storage, the graphite was heated under vacuum, and the evolved gases were collected for analysis by mass spectrometry. The gas samples were taken following periods of at least 1 hr each at 600, 800, and 1000°C. In the case of capsule 45, the xenon released from the graphite at the successive temperatures was 0.031, 0.025, and 0.055 cmS (STP), totaling 0.111 em® or 7.5% of the calculated 1.5-cm® yield from fissioning. A considerable amount of SiF, (~3 cm®) was also in the gas; the source of the silicon, other than what might have come from the stain- less steel vacuum tank and lines and from the glass manifold, is not known. In the case of capsule 12, there was a small air leak during the heating under vacuum; 0.065 cm® of xenon was recovered at 600°C along with insignificant amounts at higher temperatures, which again amounted to 7.5% of the calculated 0.87-cm® yield from fissioning in capsule 12. This time the fluorine appeared as CFy (~0.4 em®), and little SiF, was noted, in contrast with the behavior in the absence of an air leak. The xenon fluorides should not have remained as such in the graphite during the long exposure to air because of both their sublimation pres- sure and their reactivity to form the nonvolatile XeOz. The XeOg molecule has a dipole moment and should adsorb on graphite much more readily than the symmetrical XeF,; however, the dipole moment of XeOsz is probably not large because of the compensating effect of the unshared electrons on the xenon atom at the apex of a triangular pyramid with three oxygens at the base at Xe-O distances of 1.76 A and 0-Xe-0 angles of 103°. Oxygen that might have resulted from decomposition of XeOz during the heating was apparently converted to CO. 80 81 Graphite from an unirradiated control capsule was subjected to the same procedure as the irradiated graphite, and the gas (~2 cm® total) was predominantly hydrogen, along with other common constituents from the de- gassing of graphite. As recovered from the heating in stainless steel apparatus, the graphite from capsule 45 was found to be coated with a silvery film re- sembling in appearance some previously encountered surface deposits of chromium.and/or iron carbides on graphite. In an attempt to identify the film, the top of the graphite core was placed in a mounting as for metallographic preparations and subjected to x-ray analysis. The x-ray pattern resembled that for U0z, which led to the conclusion that the film had been inadvertently removed and that the x-ray pattern resulted from the presence of uranium in the surface graphite beneath the film. The occurrence of uranium in the graphite was confirmed by chemical analyses described below. Nonvolatile Constituents in Graphite Cores Because of the repeated cycling through strongly oxidizing and re- ducing cycles as a consequence of Fo generation during reactor shutdowns, the contents of capsules from assembly ORNL-MTR-47-4 had an uncertain environmental history. As a check on possible alterations, chemical analyses were made of the graphite cores from three irradiated capsules and a control capsule. Large pieces, 0.8 g or larger, in which the surface~to-mass ratio was similar to that of the original, were dis- solved in a boiling mixture of sulfuric and nitric acids. The uranium was analyzed by a fluorometric method; the other constituents were ana- lyzed by emission spectra except for lithium, which was analyzed by flame photometry. The analyses presented in Table 5.1 show increasing uranium contents with capsule power, which in turn correlates with the intensity of the off-equilibrium cycles associated with shutdowns. The unirradiated sample, which had been exposed to a thermal history the same as that of the ir- radiated samples, did not absorb any appreciable amounts of uranium or . lithium. Graphite from capsules 12, 45, and 3A was subjected to bakeout of gases at 1000°C, which may have caused changes in the constituents analyzed, and graphite from capsule 36 was not analyzed because it was involved in an attempt to recover fuel by melting. On a per atom basis, there is more lithium than uranium in the graphite cores, while the other ma jor constituents of the fuel are virtually absent. Speculations regarding the mechanism by which uranium occurred in the graphite have dealt mainly with two possibilities: (1) Transport of Ufg seemed possible as a consequence of the Fo atmosphere, but graphite alone was not regarded an adequate reducing agent for fixation of UFg in the cores. (2) Strongly reduced salt was probably in brief contact with the graphite during startup periods before the released Fy could be re-equilibrated into the melt. Autoradiography of metallographically mounted cross sections of graphite revealed that the radiocactivity in the cores was highly concentrated in a narrow band at the surface. Table 5.1. ORNL-MIR-47-4 Graphite-Fuel Compatibility Test Analyses Percentage U Ii Zr Be Th Ni Cr Fe Mn Mo Original fuel 3.89 11.6 10.1 4.8 5.8 0.0025 0.0024 0.00%96 Capsule 12, fuel 2‘82 <0.018 <0.004 <0.004 <0.001 <0.013 Capsule 45, fuel g‘gi 0.016 <0.007 . <0.007 <0.007 <0.025 Capsule 12, graphite 0.136 0.0% <0.017 0.0008 N.D.? 0.013 0.014 N.D. 0.019 0.062 Capsule 24, graphite 0.214 0.052 0.0054 0.003 0.015 <0.012 <0.015 <0.005 0.089 Capsule 45, graphite = 0.523 0.062 <0.016 0.002 N.D.? 0.01 0.04 N.D 0.01 0.062 Capsule 3A, graphite 0.00088 <0.001 0.00003 0.012 0.004 0.067 <0.014 <0.035 a N.D., not detected. ¢s 83 Since much of the fission product activity from any uranium in the gra- phite would have remained embedded in the vicinity of the uranium, the autoradiographs could tentatively be construed to mean that uranium, if present in graphite during exposure, was largely confined to the outer L mm of the cores rather than distributed throughout. Further trials with autoradiography before and after neutron activation are in progress. Fission Product Activities in Graphite Exposed to Fissioning Fuel. Analyses by gamma spectrometry of the fission product activities of Ce, Cs, Ru, and Zr found in the graphite were of interest for comparisons of the calculated burnup of the uranium in the graphite with that in the fuel. The conversion of activity to apparent percent burnup, shown in Table 5.2, gave considerable scatter, much of which was due to transport of the fission species after formation. Cesium activity may have entered the pores of the graphite as xenon gas, and ruthenium activity may have been reduced from the salt phase and deposited on the graphite. The zirconium activity was unreliable because of the low activity level at the time of analysis. Cerium activity showed a rather low but constant value and was probably the most reliable index; the implication was that the averaged fraction of the exposure undergone by uranium in the gra- phite was one-third to one-sixth that for the fuel or that most of the uranium found in the graphite had been there only during the later ex- posure cycles in the MIR. Table 5.2. Apparent Percent Burnup Based on Specific Fission Activity Capsule Calculated Calculated Burnup of Uranium in Graphite Based on No. Burnup of Fuel Ce Cs Ru Zr 12 5.5 1.80 7.76 38.6 27.7 24 7.0 1.89 10.8 63.6 2.6 45 9.7 - 1.37 2.48 8.96 Q.77 Effect of Irradiation in Assembly ORNL-MTR-47-5 Production of CF, Under Operating Conditions As described earlier,%’5 two of the capsules (0.67 and 0.34 mole % UF, fuel respectively) in irradiated assembly ORNL-MTR-47-5 were like those in the previous irradiation experiment, except that they were pro- vided with gas sampling lines through which capsule gas was swept into evacuated, 250-ml collection cylinders after having accumulated in the capsule under a nominally static helium pressure (12 psia). Conditions during the accumulation periods ranged from ambient temperature, during shutdowns, to temperatures of 815°C at fission-power densities of 85 w/em® in LiF-BeFo-ZrF,-UF, (~67-28-4-0.67 mole %). 84 Of the 61 samples collected, 33 were obtained from fissioning molten fuel, and of these about 14 from the 0.67 mole % UF, fuel and 15 from the 0.34 mole % fuel appeared to be useful. In the digest of results at operating temperatures presented in this section, four samples have been neglected; although three of these contained measurable amounts of CF,, their history and analyses involved irregularities that were not clearly resolvable. The varying radioactivity of the samples, which in many cases neces- sitated several transfers before introduction of marginally small amounts into the mass spectrometer, led to considerable variation in the sensi- tivity of analyses. An estimated approximate lower limit for measurement of about 5 ppm of CF, had been considered feasible on the basis of calibration trials with the collection apparatus. Most of the samples were analyzed for Fp at K-25 and twice again for CF, and other consti- tuents by different mass spectrometry laboratories (ORNL and Y-12). Noticeably greater sensitivity of measurement was attained on the later reanalyses, where radiocactivity was longer a problem; in general, the Jlatest figures were adopted as most representative, although again there were large fluctuations from sample to sample in the amount that could have been present (below the limit of detection) in the 18 samples that gave negative results. Partly because the amounts of CF, produced were generally too low to measure accurately, there was much scatter in the results on any of the bases on which correlations were attempted. The most enlightening information was obtained from plots such as Figs. 5.1 and 5.2, wherein UNCLASSIFIED ORNL-DWG 63-662 T T | T T i 2 [-BALANCE WITH FLUORIDE EXCESS FROM FISSIONING ] N = 05 o Cf/ RATIO BY MASS ; FOR NEGATIVE SAMPLE (AS RATIQ) o 0.0! 0.02 003 004 0.05 006 007 Xe IN SAMPLE (%) Fig. 5.1. Production of CF, from Fuel and Graph- ite at MSRE Power Densities but at Slightly Lower Temperatures. 85 the amount of xenon (long lived) in a sample served as an internal dosi- meter and measured the exposure in terms of the relative amounts of fission energy that were released while the samples accumulated. The amounts of CF, in a sample were presented as ratios of CF, to xenon in the same sample for the following reasons: (1) the primary generation of CF4 1s probably strongly influenced by the energy, as is of course certainly true of xenon; (2) comparison of mass spectrometric peaks on the same sample eliminates many possible discrepancies that might stem from unrecognized differences in sampling; and (5) also a convenient reference factor is available for assessing the chemical effects on fuels of any given loss of fluoride as CF,. The excess of fluoride ions aris- ing from fissioning UF,, because the fission product cation valences for equilibrium with the container do not accommodate the available anions from the fissioned UF,, gives a small increase in oxidizing power of the fuel with increasing burnup. Ordinarily, this oxidation should in prin- ciple be manifested as a relatively insignificant increase in corrosion product concentration. However, if the moles of CF, evolved from a fuel were approximately equal to the gram atoms of stable xenon produced during the same interval, the oxidation-reduction level would tend to remain favorably balanced. The exact ratio for balance varies with conditions of operation but for purposes of estimation can be considered as unity or a little higher. Thus, the cross-hatched bands at the top of Figs. 5.1 and 5.2 denote the upper limit of the CF4/Xe ratios below which the removal of CF, is a corrosion problem of negligible proportions for the MSRE. For higher rates of fluoride removal, that could conceivably re- quire occasional re-treatment of the fuel with HF, a chemical margin of tolerance (due to the approach to sufficient reduction to deposit uranium) -would be encountered only after an appreciable reduction of UF, to Ufs. [y UNCLASSIFIED ORNL-DWG 63-6614 2 |- BALANCE WITH FLUORIDE EXCESS FROM FISSIONING _| ] o, 0.5 o CFy/y, RATIO BY MASS SPECTROMETRY ® ESTIMATED LIMIT OF DETECTION| | FOR NEGATIVE SAMPLES 04— {(AS RATIO} { 11HIN [l Cfs/ve IN GAS SAMPLE 0.005 , s ) | ANTICIPATED SENSITW 0.002 OF ANALYSES 0.0014 ' ' | 0.02 004 0.06 008 0.10 .42 0.44 Xe IN SAMPLE (%) Fig. 5.2. Production of CF, from Fuel and Graph- ite at MSRE Temperatures but at Slightly Higher Power Densities. - ' 86 Estimated limits of sensitivity for each of the samples that gave negative results have been plotted as filled circles in Figs. 5.1 and 5.2 for use as upper limits. The shaded area shows the locus of points cor- responding to about 5 ppm CF,, and the open circles correspond to samples for which definite CF, contents were found. Taking into account both positive and negative samples, if CF, were present in amounts proportional to xenon production, the quantity was too small to be measured with cer- tainty. On the other hand, there was a possibility that CF, underwent decomposition and/or reaction with the wall because of the radiation field in the capsule; thus the amount present in samples collected over long periods could represent a steady state between formation and dis- appearance. The straight dashed lines in Figs. 5.1 and 5.2 represent trial extrapolations to zero time, effectively at least, and thus show a possible relative primary rate of generation of CF, compared to xenon in the vapor space in the capsule. In principle, a curve for CF, concen- tration should pass through all open circles and beneath all filled ones; but, because of the scatter of the data, the curves remained indeterminate. However, the trial values obtained from the dashed extrapolations were considered representative of the rate of production of CF, that could have prevailed. Thus the conclusion would be reached that CF, was produced in the vapor space at about 6% of the rate of xenon and that a steady-state concentration at, or just under, 5 ppm CF, was approached. A simple basis, though not a rigorous or necessarily correct one, for scaling such results from capsules to a reactor could be the assumption that for a given power density in the fuel the CF, production was proportional to the area of the interface between graphite and fuel. Then, since xenon production varies with the fuel volume in the core and the MSRE core contains 7 x 10° em® of fuel in contact with about 10® cm® of graphite, compared to the capsules with 10 cm® of fuel in contact with 12 cm® of graphite, the scaling factor would be 10° cm® . 10 cm> , 12 em? 7 x 10° em® or roughly unity. Thus to the extent that the assfimptions involved in the foregoing consideration are justified, the loss of CF, from the MSRE is not expected to cause difficulty. Postirradiation Examination of Assembly ORNL-MTR-47-5 Fluorine Generation from Decay Energy After 2000 hr of irradiation in the MIR at an average thermal-neutron flux of 2 x 10'° neutrons cm™@ sec”!, the 47-5 experiment was withdrawn, and valves were attached to the gas sweep lines in preparation for ship- ment. Within 11-1/2 days after termination of the exposure, the assembly was installed in an ORNL hot cell with pressure-measuring devices attached to both gas-swept capsules. A gas sample withdrawn from the low-power capsule (0.34 mole % UF,) contained 38% He and 55% Fo by mass spectrome- tric analysis. It also contained large amounts of tellurium and iodine activities (4O r/hr at contact through a nickel wall). The high-power 87 capsule (0.67 mole % UF,) was not sampled because of the activity problem. Both capsules were then evacuated, and pressure-rise measurements were begun (Figs. 5.3 to 5.6). (as samples were taken frequently to check the UNCLASSIFIED ORNL-DWG 63-598R 14 l , [ | 2 \\ - 7\ cURVE For 6; =004 7 . \ 3 8 | - @ a LI_N 5 6 — | M E (&) 2 n 4 2 l - \\ 1 \ | S 0 1 | 3e°c| s0° | se°|34e|-70°| 30| 470 | 290 | e9 |87 2ge " N S N TR RN S NS B 0 10 20 30 40 50 60 70 80 90 100 COOLING TIME (days) Fig. 5.3. Postirradiation Fluorine Generation in MIR-47-5 Capsules. UNCLASSIFIED ORNL -DWG 63-599R COOLING TIME (days) 10 20 30 40 50 60 70 80 90 > | | | ) | | ! 1 | 4 65 w/cm3 CAPSULE — \ /___/ - sse |N-PILE < 3 -~ -——IN TRANSIT 8 / — OUT-OF-PILE g / O // J » . |L|_ y I e I/ ( { f— / ’,/ 35 w/cm3 CAPSULE . -~ -~ . -~ . 7 O :“‘o.—/ 36 50 | 86 |34l-70| 3 1.7 29 69 187128 TEMPERATURE (°C) Fig. 5.4. Radiolytic Fluorine Loss from MIR-47-5 Capsules. 88 UNCLASSIFIED ORNL-DWG 63-3423 83°C 54 °C 88 °C a5 33°C | | 88°C | 34°C || 74°C | | 38°c | B2 ....oooo.oo""ooo.. L] 2 -401"CAPSULE NO. 4 =, 3 E (GMCITI ) ...........o.O......... - Seeiene,, oo 16 o _3-5 [ XX XN _LAJ > '\.——" ""'--.___ -—] . / FLUORINE PRESSURE, +—SAMPLED | = CAPSULE NC.4 g 3.0 F» ADDED L N —I8 2 ™ b o w —-2.5 1 atm OF F» EVACUATED —= o NORMAL F» A\ 2 L e o ACCUMULATION ] 8.4 o -2.0 0 3 : oooooloooooloo0000000oooooooooooooooooooooooooooooooooooouoooooooo L CAPSULE NOC. 3 (35 w/ecm3) sl T 95 105 15 425 135 145 455 COOLING TIME (days) F> PRESSURE (psia) Fig. 5.5. Recombination of F, with Radiolytically Reduced (Frozen) MSRE Fuel. 33 °C 54 | 88 °C 34 °C l | UNCLASSIFIED ORNL—DWG 63—-3424 L] 8 74 °C ] l 38 °C T ; T 7 > NORMAL F, EVOLUTION | 2888338 = [} o] . k] : \ ; Ty \ I-\ ’;"‘e. ""; Y .“" * \ } g ° S -2 |F, ADDED TO 1 psia<<-pi3 'J‘. } = ‘ { atm CF, ADDED— J = Z -3 H CF, EVACUATED = ) 3 . © -4 [—F, ADDED TO 05 atm @ 4—F, PRESSURE CONSUMED TO 0.25 atm ' o -5 F, ADDED TO 4 atm - 2 e 0 CAPSULE NUMBER 3 (35 w/cm3) u .’ e CAPSULE NUMBER 4 {65 w/cm3) [ ] = y > . |* q - -8 95 105 15 125 135 145 155 Fig. 5.6. Effect of Temperature and Fluorine Pressure on Rate COOLING TIME (doys) Recombination of F, with Reduced (Frozen) MSRE Fuel Salt. of g9 purity of the evolved fluorine and to substantiate the absence of air leaks into the capsules. Farly samples contained as much as HO% Os, from reaction of Fo with oxides on the system walls. Later samples reached Fp purities of 96.7%. After measurement of the room-temperature fluorine evolution rates, the variation of rate with capsule temperature was studied. Rates were measured at 50, 86, minus 70, and 2°C, with intermittent room-temperature determinations to provide a normal for comparison. The data for the first 95 days are shown in Fig. 5.3, together with a calculated fluorine- generation curve based on a G value of 0.04% molecule of Fo per 100 ev of fission product decay energy (calorimetrically determined data® for decay power were used, since energy absorbed in the fuel was needed). 1In Fig. 5.4, the same data on fluorine liberation are represented in integral form, and as indicated by the points near the origin, there were losses of Fp from the capsules during shutdowns between reactor cycles; these losses were equivalent to increments of chemical reduction of the fuel. The ac- cumulated effects, shown in terms of the extent of expected reduction of UF, to UFg in Table 5.3, caused the fuel (rather early in its history) to be reduced to the point that some disproportionation of UFgs could have been expected; appreciable deposition of uranium, particularly in the cap- sule with higher power density, must have occurred before the exposure was terminated. Further evolution of Fs after termination led to cumulative losses, in terms of the total fluoride content of the fuel, of 2 and 4% for the low- and high-power capsules respectively (nominal experimental design powers were 35 and 65 w/cm®). For the first 50 days the capsfile pressures ranged to 1 atm but were decreased to zero at each sampling. Pressures less than 1 psia prevailed between day 50 and day 114, Some of the more significant conclusions to be drawn from Figs. 5.3 and 5.4 are as follows: 1. Fluorine evolution rates showed a broad maximum between O and 50°C, with a peak near 35°C or higher, which shifted slightly with time. . ' 2. At -70°C, the rates decreased to low values (0.67 mole % UF,) or zero (0.34% mole % UF.). 5. At 85°C, the net rates in both capsules dropped to zero, rapidly for the highly reduced salt but slowly for the less reduced salt. i, The F» production rates initially matched a G value of 0.Ok. Subsequent G values decreased to 0.02. 5. In most cases, when the capsule temperature was abruptly raised, a burst of fluorine was noted. Conversely, after a sudden drop in capsule temperature to room temperature, several days were required for recovery to a steady rate. Table 5.3. Calculated Conversion of UF, to UF; Due to F,; Loss Induced by Decay Energy During Pile Shutdowns While Irradiating MSRE Fuel (ORNIL-MIR-47-5) 0.34 mole 9 UF, in Fuel 0.67 mole % UF, in Fuel Induction Cumulative F» Loss Induction Cumulative F, Loss Period (hr) Std em® % UF, to UFs Period (hr) Std cm® % UF, to UFs Shutdown Interval - Date Duration (hr) 10/29 98 36 7.8 32 8 21.4 46 11/13 53 30 - 10.5 43 107 32.7 71 11/19 64 >4, N.Y.? N.y.® 6 55.5 120 12/1 47 207 12.0 49.5 0 75.5 163 12/21 fi 47.5 >47.5 N.Y.2 N.y.® >47.5 N.y.? N.Y.® 1/4 85 30¢ 17.5 72 207 08.8 214 1/21 8 07 18.2 75 0 107.3 232 a N.Y. indicates no yield. 06 91 6. Fluorine evolution rates were much less strongly influenced by the degree of reduction of the fuel salt than by the tem- perature; the chemical reduction indicated in Fig. 5.4 would have, in molten fuel, caused violent chemical changes such as the precipitation of all of the uranium and some of the other cations as metals. 7. At 87°C an actual consumption of fluorine (negative genera- tion rate) was noted in the 0.67 mole % UF, fuel. The more highly reduced state of this salt might account for its greater tendency to recombine with fluorine. In order to explore further the nature of the recombination of fluo- rine with reduced salt at elevated temperatures, several additional variations were made in fluorine pressure as well as capsule temperature. Fluorine pressures from a fraction of 1 psia to 1 atm and capsule tem- peratures from 33% to 88°C were studied. A final test was run to determine whether the radiolysis of CF, at 1 atm of pressure by 150-day-cooled fuel salt might be detectable. The results from these variations in capsule conditions are presented in rate form in Fig. 5.5 and as cumulative fluoride loss or gain by the fuel salt in Fig. 5.6. Also given in Fig. 5.6 is a plot of the Fz pres- sure in the 0.67 mole % UF, capsule on the same time scale. Similar but not identical pressures obtained in the 0.34 mole % UF,4 capsule at any given time, The following detailed description explains in part some of the pecu- liarities and complications in Fig. 5.5. Both capsules were evacuated after the 87°C interval that ended on day 95, and fluorine evolution with no added pressure was followed for 15 days at 33°C to furnish a "base line". The observations under these conditions were extended over 15 days because the evolution rate continued to rise slowly, in line with pre- vious observations that room-temperature rates were not immediately re- covered ‘after the salt had been heated. On heating to 83°C, the 0.67 mole % UF, salt rapidly consumed the generated Fp, the pressure falling from 0,39 psia to 0.07 psia. In the same interval the generation rate in the 0.34 mole % UF, capsule slowly dropped to zero, confirming previous behavior. On day 113, the temperature was raised to 88°C, the capsules were evacuated, and 1 psia of fresh Fy was added to each. The initial consumption rate in the 0.67 mole % UF, capsule with higher reduction more than doubled its value for 0.3 psia and then decreased. Fluorine was again added to 1 psia, and the consumption rate nearly reproduced the previous gyration. At 1 psia and 88°C, the 0.34k mole % UF, salt slowly consumed Fo. The capsules were sampled and evacuated, and 0.5 atm of Fy was added to each. The more concentrated fuel consumed Fo at approximately twice the l-psia rate for three days, then less rapidly as the Fp pressure de- creaged to 0.25 atm. A similar doubling of its 1l-psia rate occurred in the other capsule. - 92 - The capsules were again sampled and evacuated, and a full atmosphere of Fo was added to each. Again the consumption rate in the concentrated fuel did not quite reach the previous maximum rate at 0.5 atm Fo, possibly because salt accessible to Fo had become partially saturated. A similar rate decrease was noted in the other capsule. On day 122, the assembly temperature was reduced to 3H°C, leaving slightly less than 1 atm of Fp in each capsule. In both capsules, fluo- rine generation was rather quickly resumed at rates similar to the initial base-~line rates. Curiously, there was now little if any change in rate in 12 days of continued evolution; otherwise, the addition of F +to the fluoride-deficient salt apparently had not enhanced subsequent Fo liber- ation. In fact, the slightly lower rates could perhaps have been consid- ered indicative of slight back-reaction at 1 atm of Fo at 34°C. With the Fy pressure maintained near 1 atm, the capsule temperature was raised to 54°C. The 0.67 mole % UF, capsule began to consume Fo slowly. The other capsule exhibited the small burst of Fo often observed after temperature rises, followed by a decrease of generation rate. When the temperature was raised to T74°C, the 0.34 mole % UF, salt showed ano- ther small Fo burst and then leveled at a low positive generation rate. The 0.67 mole % UF, salt consumed Fo at a moderate rate at 74°C, showing an erratic rate behavior; such irregular behavior is common in gas-solid reactions involving successive formation and rupture of partially protec- tive films. Finally, when the temperature was again raised to 88°C, the consumption rate for the more-reduced salt increased to a value only one- fourth its previous value at 88°C and 1 atm Fo and then fell off strongly. This is again suggestive of saturation of the outer surfaces of the fuel salt by Fn. At 88°C, the 0.34 mole % UF, salt gave off another slow Fo burst that lasted more than two days before consumption dominated. It appeared that the 0.34 mole % UF, capsule was more prone to such bursts ~at 1 atm of Fo overpressure than at 1 psia overpressure. The capsules were again sampled, evacuated, and cooled to 38°C. An atmosphere of CF, was added to both capsules and allowed to remain for four days. The apparent negative rates for both capsules for the first day were probably due to a slight drop in capsule temperature, causing a pressure drop. (The measured pressure drops were only 0.0% and 0.05 psia out of a total pressure of 1Lk.7 psia.) For the remaining three days, approximately normal generation rates were observed. Analyses of the CF, before and after radiolysis were identical except for the expected small Fo concentration from salt radiolysis. After evacuation of the CF, from both capsules, the observed Fo generation rates were about the same as with CF, present. The main conclusions of the behavior demonstrated in Figs. 5.5 and 5.6 may be summarized as follows: 1. MSRE salt, chemically reduced to the extent correspbnding to 3.5 to 4% fluorine loss, reacted with fluorine at a rate which was very low at room temperature, easily measurable at 54°C, and 93 rapidly increasing as the temperature exceeded 88°C. With salt reduced only to 2% fluorine loss, the recombination was measur- able only above 88°C. 2. Continued reaction of fluorine with the strongly reduced salt resulted in a very marked decrease of consumption rate before even 0.5% of its 4% fluorine deficit had been restored. With the other salt, a similar deceleration was noted. 3. In extrapolating the above conclusions, temperatures well above 100°C would probably be required in order to restore radiolyzed MSRE salt to its original oxidation state by fluorination. In the case of the sealed in-pile capsules in recent MSRE irradia- tion tests, the probability is thus enhanced that some of the Fo generated during reactor shutdowns could have survived to elevated temperatures at startups; meanwhile reduced components of the fuel could have reacted with the graphite. b, For reasons implicit in the foregoing discussion, exact rela- tions between Fo pressure and recombination rates are not derivable from the available data. Qualitatively, however, the initial reaction rate of fluorine with a highly reduced fuel was approximately proportional to fluorine pressure. As reaction continued, the reaction order fell to 0.5 and less. 5. Radiolysis of CF, was not detectable in the presence of irra- diated fuel salt. The salt had cooled 150 days at the time of testing, and the high-power capsule should have been generating beta activity at the rate of 0.1 w. Conclusions Pertaining to Reactor Operation. Fluorine generation 1s expected to be negligibly slow from solid MSRE fuel salt which had been fissioning at 65 w/cm®, provided its temperature is maintained above, e.g., 200°C. The possibility of Fo release from molten fuel is even more re- mote, in view of the higher temperature and improved kinetics for fluorine back-reaction in the liquid state. Mechanism of Salt Radiolysis. The fluorine generation and consump- tion data discussed above are consistent with the following kinetic pic- ture of salt radiolysis. The net fluorine liberation rate depends on the balance of at least three effects: (1) the primary radiolytic generation rate, which is probably temperature independent; (2) chemical recombination of fluorine iniide the crystals with the reduced radiolytic species, e.g., Li-, Beo, U®", and U°; and (3) a temperature-dependent diffusion step for escape of internal F° or Fo from the crystal. At low temperatures the slowing of diffusion would decrease net Fo liberation, perhaps simultaneously build- ing up higher internal pressures of Fo which give rise to bursts of re- leased gas when the salt is warmed. At high temperatures, accelerated O recombination of Fso with reduced products would result in the observed zero or negative formation rates. Schematically, this results in an overall pattern like that in Fig. 5.7; but, because the onset of the re- combination reaction depends on the condition of the fuel and the pre- sence of Fo, the temperature scale in Fig. 5.7 should be viewed as merely illustrative. UNCLASSIFIED ORNL-DWG 63-663 LI_N W (@] z o '_ o J o > o 0.6 mole % g (.5‘ UF4 FUEL s k g ; & | DIFFUSION § CONTROLLED? 4 a P ’ i 1 8 5 _i 4 i a W 5 a Al E % S o 2 ; o ‘ /! - = ; . = : > ';' Vg \ BACK—- REACTION = g .z CONTROLLED? < < I I e # T T H o 2Bf—=— 0.3 mole % UF; FUEL y }#J : ? -60 —-40 -20 0 20 40 60 TEMPERATURE (°C) Fig. 5.7. Schematic Summary of the Effect of Temperature on Postirradiation Release of F, from MSRE Fuels (5% Burnup). OQut-of-Pile Irradiations of Solid MSRE Salts ’ Gamma, Irradiations The production of fluorine-containing gases in reactor-irradiated MSRE fuel-salt compatibility tests raised the question of whether such behavior arose from the effect of radiation (possibly'internal) on the solid salt. One phase of the experiments exploring this problem con- sisted in exposing such a fuel salt, otherwise unirradiated but in con- tact with graphite, to gamma irradiation in the 10,000-curie Co®° facility of the ORNL Chemical Technology Division. 95 A 25-ml INOR-8 container holding 35.3 g of salt (14.1 ml) and 2L-1/L- in. graphite spheres (3.1 ml) was exposed in the facility for 1870 hr at temperatures of 35 to 67°C. Composition of the salt was as follows: LiF (nat) 69.2 mole % H50 <250 ppm BeFs 2%.0 mole % Oxygen 305 ppm ZrFy 5.2 mole % Nitrogen 61 ppm UF, (nat) 1.1 mole % Sulfur 32 ppm ThF, 1.7 mole % Chromium 8 ppm Iron 130 ppm The 35.3 g of salt contained 0.76 g-atom of cations and 1.12 g-atoms of fluoride ions. The container was equipped with gas tubing and valves, pressure- measuring instruments, and a sampling tank assembly to obtain gas samples; the complete apparatus involved in pressure measurement and gas collection was prefluorinated by K-25 Plant personnel. After the salt was loaded around the graphite, melted and refrozen in a glove box under a dry inert atmosphere, the 1id was welded in place, and the assembly was flushed and filled with purified helium at a pressure slightly above atmospheric. Dose rates to the salt were calculated from previous calibratioms, with due consideration of the attenuation resulting from the materials and geometry involved. The center-line dose rate was estimated to be 0.76 x 1018 ev g™t min"t. Thus, for fluorine production corresponding to a pressure increase of a given number of pounds per square inch per hour, Gp, = 7.5 X (psi/hr) = molecules of Fs per 100 ev. - The salt was exposed in the Co8® facility at ambient temperature (50 to 60°C). For about 600 hr no significant variation (<0.3 psi) in gas pressure was observed (GF2 = 0). After this time pressure began to rise at a mildly accelerating rate. At hour 1343 the pressure had risen from 16.7 to 18.7 psi (GF, = 0.02). The temperature was lowered to 35°C and held there until hour 1724. The pressure changed during this pericd from 18.4 to 20.4 psi (G, = 0.0k). The salt temperature was then held at 67°C {ambient) until hour 1870. The pressure changed during this period from 21.2 to 23%.0 psi. A small, sharp increase in pressure was noted for a few hours after the temperature increase to 67°C at hour 1724; similar pressure bursts were encountered with reactor-irradiated capsules followifig temperature increases during the decay period. The steady-state G value for the sub- sequent period was gbout 0.06. At hour 1870, the pressure-monitoring manifold was connected to the sampling system, which included four 60-ml collection tanks. The gas volume of the irradiated container and manifold was 30.6 ml. 96 Successive samples were removed into the tanks, with helium being admitted to the manifold after withdrawal of the first two samples. The samples, as analyzed on a K-25 Plant mass spectrometer, gave the following results (helium and air components not included): Gas std cm® (total) F» 1L k2 Excess 0o 1.41% SiF, 0.02 CF4 0.26 COF2 0.03 COz 0.39 "Not including estimated 0.6 to 1.0 std em® involved in second sample removal, which had some air leakage. There were no other gases found except air components in proper proportions. It is possible that the CF, and COFo came from COso or other organic contamination during sampling or analysis. The equilibria of the fol- lowing reactions are strongly to the right. Fo + COz » COFs + 1/205 Fs + COFs = CF4 + 1/202 The other possibility is that the CF, was produced by the reaction of fluorine with the irradiated graphite, although in the absence of ir- radiation no measurable reaction would have been expected. The increases in pressure due to the evolution of Fo are consistent with the observation of radiation damage to LiF and other alkali halides, as summarized by Billington and Crawford.’ Displacement of fluorine atoms by any of various forms of radiation ultimately leads to escape of fluorine atoms or molecules from the crystal. Since return to the va- cancies in the solid lattice is probably an extremely slow process for molecular fluorine, the accumulation of a substantial fluorine pressure after an induction period would not be unlikely. Similar effects have evidently been encountered in the reactor-irradiated salts due to fission product activity during the decay period. The autoclave has been repressurized with helium and is being ir- radiated to further explore the effect of temperature. X-Ray Irradiations Fuel-salt mixtures and individual component fluorides were irradiated with x rays using two types of x-ray units (GE 250 kvp, 30 ma, and Muller 97 250 kv, 3 kva). All experiments were made with the maximum input to the machines at 250 kvp (GE) or 250 kv (Muller). In spite of the use of direct current in the Muller unit, dosimetry indicated an absorbed dose rate of only 0.3 relative to that for the GE unit. Irradiation vessels were fabricated from nickel with diaphragms of 0.010 or 0.020 in. of nickel to allow passage of the maximum amount of energy into the salt. The inside surfaces were well fluorinated at 150°C before introduction of the salt. The salts were handled only in a helium dry box and were recrystallized by melting in nickel crucibles in order to keep the carbon content as low as possible. Generally, a 75- to 100-g sample of salt was sufficient to form a layer about 0.75 in. deep over the area of the x-ray beam. Dosimetry measurements were made with ferrous sulfate. The dosimeter solution (5.5 liters) was placed just below the irradiation vessel in a container with a solution depth of 8 in. (large enough to intercept the total beam). The amounts of energy absorbed by the test salts were de- termined by the difference between that energy passing through an empty vessel and that passing through the vessel containing the sample. This dosimeter is satisfactory for use with photon energies as low as 50 kv. However, the depth of solution was insufficient to stop all the high- energy X rays. Calculations showed that the error in energy absorbed by the salt due to solution depth is insignificant. The fuel salt used in these experiments contained LiF-BeFs~ZrF,- ThF,-UF, (70-23-5-1-mole %). Two irradiation experiments were carried out with this salt in helium at 1 atm using the GE machine. A 100-g sample absorbed energy at the rate of 5 x 10*® ev/min; but, owing to difficulties with the machine, it was not possible to irradiate for long periods of time. The first experiment (135 min, 6.6 x 107 ev absorbed) resulted in liberation of 267 ppm COFs and 174 ppm CF4 in 2L ml of helium, as determined by mass spectrometry, and indicated a G value of 0.006 mole- cule of Fo liberated per 100 ev of energy absorbed. A second irradiation of the same salt (425 min, 1.7 x 10°% ev absorbed) resulted in liberation of the same compounds, but with a G value of 0.009. In the first experi- ment the CF4/COF2 ratio was 1.3, while in the latter it was 2.0. The yield of Fs used for calculating G values was the total found in the CI4 and other compounds of fluorine. There were no significant differences between the irradiated and as-received salt as determined by carbon and oxygen analyses, x-ray diffraction, and microscopic examination. 1In both cases the oxygen content was ~200 ppm and the carbon content was <500 ppm. Further irradiations were carried out on another sample of the same fuel-salt mixture with the Muller machine. With an input to the tube of 250 kv at 12 ma, the dose absorbed by the fuel salt was about 1.3 x 102 ev/min for a 110-g sample. The salt was loaded into the fluorinated vessel in a helium dry box and was treated with Fo at 150°C until the consumption rate was low. During this fluorination, appreciable amounts of CF4, COFs, and SiF, were detected. After the fluorination the vessel was thoroughly purged with helium and left with 1 psig of helium over- night before analysis. Mass spectrometer analysis of the internal gas showed no trace of Fp, CF4, or COFs. 98 The salt was irradiated for 22 hr (1.6 x 10?2 ev absorbed) in helium at 1 atm. Gas analysis showed 280 ppm F» and 230 ppm CF,, with only traces of COFp and S5iF4, in 20 ml of helium. The G value from these data was 0.002. Irradiation of the same salt for an additional 44 hr (3.3 x 10°% ev absorbed) gave similar results, with a G value of 0.00k. ‘ The .vessel containing the above salt was heated for 16 hr at 250°C at a pressure of 3 to 4 p to remove any adsorbed gases. Two subsequent irradiations (51 hr, 4 x 10%% ev; 93 hr, 7 x 10°% ev) yielded no Fs but did yleld appreciable quantities of CF, and COF-. The calculated G values were 0.001l3 and 0.002 respectively. Analysis of the salt after removal from the vessel indicated an oxygen content of about 1800 ppm. If this resulted from in-leakage of air during the heating, it could ex- plain the failure to detect Fo in the gas after irradiation. A 75-g sample of ZrF, was irradiated on the Muller unit (250 kv, 12 ma) for a total of 269 hr (1.4 x 10%® ev absorbed) without evidence of fluorine evolution. This sample of salt, in its irradiation vessel, was thoroughly fluorinated to remove any carbon, sulfur, or silicon impurities and to convert metal oxides to fluorides after the above exposure. It was then irradiated for 65 hr (3.4 x 10°% ev absorbed) without detectable radiolysis. A 50-g sample of LiF was irradiated for 333 hr under similar condi- tions. Owing to the low mass absorption coefficient of LiF, this exposure resulted in absorption of 3 x 10°% ev. Gas analysis failed to detect liberation of fluorine or gaseous fluorine compounds in the helium cover gas. This LiF was fluorinated in its irradiation vessel and consumed appreciable fluorine before the consumption rate became negligible. . Evidence to date implies that x-ray photons of 250 kv and less can cause radiolysis of metal fluoride mixtures under certain conditions. It appears that the individual component salts may be less susceptible to decomposition than the mixtures, and it 1s entirely possible that the pre- sence of impurity salts may drastically affect the G value for fluorine generation. Irradiation with Van de Graaff Electrons As mentioned previously, experiments to investigate the effect of beta radiation on fluorine evolution from solid, MSRE-type fused salt have been in progress.® These were initiated to establish whether irradiation of the solid salt results in evolution of fluorine in amounts required to explain the amounts found in in-pile capsule tests; but an important ob- Jective was the evaluation of Gp, as a function of several variables, including irradiation dose, dose rate, and crystal size in the salt. Ex- periments have been completed which yield values of Gy, as a function of dose, at a constant dose rate, up to a total dose of about 6 x 1022 ev per gram of salt. Observations have also been made on the effect of dif- ferent dose rates on Gp_, and on changes in crystallite size in the salt as a result of the irradia%ions. The experiments and results to date are summarized in the following paragraphs. 99 Experimental. The procedure involved exposing particles of solid fused salt to fast electrons and determining gain or loss of fluorine in the irradiation cell (Fig. 5.8) due to the irradiation. The cell was filled to atmospheric pressure with a mixture of known amounts of helium and fluorine (4 to 6% fluorine) prior to irradiation. Control experiments on fluorine loss or gain and the development of methods and technigues to UNCLASSIFIED ORNL—DWG 63—6487 tCOLD AIR IN GAS our\ | NICKEL CELL BODY Ya-in. Cu < L§ 77777 N\ 7 zZZZ N \ \/—-NmKEL FLANGE H,0 OUT T § § , ‘ \ \ g ANNNNN NN § § \ | Wi w coofr?gpg[{ocx_____§ N &”Q/ NICREL WINDO \ N N \ EENPEENIN \ 2cm \/ \ \ \ YRR N 1\ NN § o L NICKEL WiNDOW 0.37-mm DEEP B /\\ N RECESS FOR SALT § N g R THERMOWELL % v, Wiz e § / CELL VOLUME BETWEEN VALVES-5.75 cm? He IN Fig. 5.8. Irradiation Cell, 100 control loss or gain comprised a major part of the experimental work. The addition of a fluorine-helium mixture was adopted as a method of minimiz- ing fluorine loss to the cell and of favoring the formation of a steady state with respect to the amount of fluorine sorbed by the salt. A summary of some other experimental conditions is given in Table 5.k, Fluorine analyses were made by spectrophotometric measurement of the amount of iodine formed during passage of a gas sample through KI solution. A few check analyses of the composition of cell gases by the mass spectro- gréphic method were obtained also. Provisions were made with auxiliary equipment for (1) evacuation of cell and connecting lines, (2) fluorination of system, (3) addition of purified helium, (4) addition of fluorine-helium control samples, and (5) sweeping of gas from cell through KI solution. The cell was removed from the auxiliary equipment for filling with salt and for irradiations. One group of experiments was made to determine the Gp, as a function of dose. A total dose of 6.1 x 10%% ev/g was accumulated in a series of 2-hr exposures at 0.97 pa. Another group of experiments was made in which the salt was irradi- ated for 8.67 hr at 0.97 pa before suitable control measurements were made. Following this, l-hr exposures at 0.97, 0.49, and 1.9L pa were made along with control measurements. Other experiments were made to determine the fluorine loss or gain in the cell under irradiation but without salt. Two samples of salt which had been exposed to a total dose of about 7.5 x 1022 ev/g in two different experiments were examined by C. F. Weaver for crystallite size. Results. The results of the exposures of the cell without salt showed that fluorine was released at a low rate throughout the 12 hr of irradiation. The rate increased with dose to a maximum of about 10 ug/hr. (It is likely that this fluorine was released from the Teflon gasket.) The control experiments made at intervals between irradiation experiments showed a continuing loss of fluorine to the cell and/or salt throughout each group of experiments. 1In the group of experiments to determine fluorine evolution as a function of dose, the ratio of the amount of fluo- rine lost, as indicated by control experiments, to the increase observed in a radiation experiment was always <0.2. In the other group of experi- ments, this fraction was higher (about 0.7 at maximum). However, in this latter case, relative release rates were of primary interest, and these were considered accurate to about + 10%. A mass spectrographic analysis of the gas contained in a cell after a l-hr irradiation at 0.97 pa showed the following percentage composition: Fo, 5.16; No, 3.2; 0s, 1.0; COs, 0.04; CF,, 0.06; U, 0. The ratio of Np to O indicates air contamination. This contamination could have occurred during the transfer to the mass spectrograph. Table 5.4. Summary of Some Experimental Conditions Employed in Van de Graaff Experiments Salt samples Approximate com.positiona Size Bulk density Weight in path of electrons Total weight exposed Total weight in cell Crystallite size in unirradiated material Intensity of electrons impinging on salt in usual experiments Estimated average rate of energy deposition in salt® at 0.97 pa Estimated temperature in salt at 0.97 pa - 69 IiF, 23 BeF,, 5.2 ZrF,, 1.1 UF,, 1.7 ThF; (mole % 30- to 50-mil particles from crushed solid 1.0 g/cm’ 0.37 g/cm? 0.29 g 0.7 g 10up, compared with was brownish purple, and the more intensely colored material had a range of refractive indices. The color change was observa- ble at dose levels too low to cause measurable changes in refractive 109 indices.®”1 The compound 2LiF-ZrF, had reduced refractive indices but remained colorless.® The occurrence of these color effects in electron- irradiated salt® that remained crystalline at near room temperature indi- cated that they are radistion effects rather than the results of doping with fission products or other nuclear products, or contamination by the container. No UFs has been detected in these samples®™® in spite of the reducing conditions ensuing from the loss of Fo induced by radiation damage. The expected UF -containing compounds have been detected.”?8,11 In an earlier observation of irradiated salt that had been exposed to air in a melt-out operation, much UQ, was present.l® (xides were rarely observed in samples from subsequent irradiation experiments.®” 82121 The change in composition of the liquid by evaporation, as evidenced in a series of samples by a decrease in the relative amount of low re- fractive index material (<1.%93%) and the presence of well-chrystallized LiF,6 was attributed to a wvapor composition richer in BeFs than the fuel. The appearance of the distillate indicated that the composition change occurred by selective removal of 2LiF-BeFs, BeFo, and 2LiF-ZrF, (ref. 6) and was most extensive in the most highly irradiated rr1a1:erials,6’8 pro- bably because of the higher temperatures involved. In a given series, capsules which were most highly irradiated con- tained the best crystallized material.® This may have been caused by a reduced rate of cooling as a result of greater fission product afterheat. However, a similar crystal growth was observed in electron-irradiated samples for which the temperatures remained well below the solidus.? This observation was surprising; accordingly, more than 40 each of irra- diated and unirradiated samples were contrasted in establishing the va- lidity of this correlation between exposure and crystal size, and no exceptions were noted for the in-pile or electron-irradiated crystals. The effect was not detected in crystals that recelved x rays at relatively small doses such as 102° ev/g. The growth, apparently a consequence of radiation rather than a heat effect, remains unexplained and will require further study. The range of particle sizes found in the in-pile capsules suggests that the surface area of the solids could have differed by a order of mag- nitude from capsule to capsule. This may favor a recombination of released fluorine, as for example in capsule 6 of ORNL MTR-L47, which contained very very finely divided material® and little Fo (ref. 13), according to Weaver.'% The large proportion of opaque and fine materials, as well as the extensive radiation damage, provide a need for x-ray diffraction data to supplement the petrographic results, since, in addition to identification of phases, a knowledge of structural changes in the radiation-damaged materials might have both theoretical and practical interest. For instance, if it were established that fluorine is released selectively from one phase rather than from the entire salt, perhaps a composition for MSR fuels could be selected which would not yield the Fo-generating phase on freezing. 110 Xenon Fluoride Studies Further studies on the xenon fluorides by x-ray diffraction techniques established the existence of a compound with the formula XeFs-XeF,, and its crystal structure was determined.'® This substance was readily crystal- lized from vapor containing a mixture of XeFp and XeF,, and crystals of about 1 mm were grown. No other intermediate compound in this binary sys- tem could be detected; the compound should have been present at some stage in irradiated MSRE capsules, but too little is known of its chemical be- havior to confirm the surmise that XeF, or XeFg was the species present at the time gas samples were taken from capsules. The crystal structure consists of an array of discrete XeFs and XeF, molecules whose symmetries and dimensions are the same as in crystals of the component phases. Since contacts between molecules in XeFs -XeF, are not significantly closer than intermolecular distances in either of the components, the new phase is described as a molecular addition compound. Core and Blanket Fluids for Future Reactors Investigations of temperature-composition relations in several molten- salt systems of potential use in the development of a molten-salt fast- breeder reactor were initiated recently. The guiding requirements were (1) that the core and blanket fluids remain as homogeneous fluids at tem- peratures above 550 to 600°C; (2) that the core fluid, a chloride mixture, contain 35 to 55 at. % of fissionable species; and (3) that the blanket fluid, a fluoride mixture, contain ~25 at. % of thorium and be capable of dissolving an additional amount (5 to 7 at. %) of uranium and plutonium at operating temperatures. Although several of the binary systems of UCls, UCl,, and PuClg with the alkali chlorides or alkaline-earth chlorides are well known, information for somewhat more complicated mixtures of UCls and PuCls was necessary to obtain the requisite combinations. Accordingly, the behavior within three-component systems is under study. Planned in- vestigations of chloride ternary systems (at ORNL) were preceded by the construction of predicted phase diagrams of the chloride systems of UClg and PuCls to facilitate early discovery of optimal compositions for reactor use. The diagrams suggested that the required 35 to 55 at. % of uranium and plutonium could be dissolved at 550 to 600°C by a proper choice of ternary chloride compositions. The simplest system which appeared to be capable of providing a range of fluoride melts suitable as blanket fluids and as homogeneous liguids at temperatures above 550 to 600°C is NaF-KF-ThF,. As the blanket re- mains in use, the composition could change sufficiently that it effectively becomes a mixture of NaF-KF-ThF,-UFs-PuFs. Preliminary data obtained in an investigation of the system NaF-KF-ThF, showed that the minimum melting temperaturel® occurs within the subsystem NaF-2NaF*ThF,-NaF KF-ThF, at the eutectic composition NaF-KF-ThF, (67-11-22 mole %), melting at 535°C. The contiguous primary phase fields at this invariant point are NaF .2ThF, solid solution, NaF, and NaF+KF:ThF, solid solution. Substitutional solid solutions are readily formed between NaF-ThF, and KF-ThF, solid phases. 111 Therefore, segregation among the crystallizing prhases in the system NaF- K¥-ThF, is diminished, as should also be the case for mixtures of NaF-KIF- ThF4-UF,. A static molten blanket constituted of NaF-KF-ThF, appears to have favorable properties and, from the standpeint of crystallization be- havior, should be relatively free of hazardous segregation of the uranium and thorium phases during freezing-thawing cycles. An expanded investiga- tion of fast-breeder blanket materials is planned. Oxygen and Sulfur in Molten Fluorides Oxide Behavior in LioBeF,-7ZrF,-UF, In a previous progress report,’ results on oxide behavior in flush- salt - fuel-salt mixtures (LigBeF4 containing principally UF, and ZrF4) were presented which were consistent with the simple dissolution equilibria U0a( &) » UL 4+ 202" (1) 7r0o(s) » zrit 4 2027 . (2) It fias foyund that Zr0, could be precipitated by adding oxide in the presence oi U%" without detectable U0z precipitation until the ratio [zr*"]/[U*"] in solution fell to about 2.7 (at 600°C). With further oxide additions both UQz and ZrQO. were precipitated, with the ratio of [zr*"]/ .[U4+] remaining nearly constant. This suggested strongly that UO- and ZrO> were precipitated separately as pure oxides; i.e., that no (Zr-U)0s solid solutions were formed in these systems. The relation of the separate concentrations of 74" and U%t in the melt to the estimated oxide ion concentration in the melt was consistent with the simple mass action behavior expected for the above two reactions. This implied that no significant complexing of U4t or zZrtt with 02~ occur- red in the concentration ranges tested earlier; but continued studies suggest that, over larger concentration ranges, the behavior is more com- plicated. The recent results were obtained during further exploration of the behavior of U0z and ZrOs in equilibrium with LisBeF, melts containing UF, and Z2rF, which were initially directed toward the possibility of oxide solid solution formation. Published phase diagramsl8 are in agreement that a virtually continuous solid solution is formed between UO- and tetra- gonal Zr0s, at elevated temperatures (1800°C) and that, although an immis- cibility region in the tetragonal solid solution prevails at lower tem- peratures, tetragonal ZrQO, stabilized by UO- persists far below the transition of Zr0Oz to monoclinic, near 1000°C. Also, in regard to the substitution of U for Zr%t in monoclinic Z2r0s, these diagrams show, though not with agreement in detail, appreciable regions of solid solution. Such solid solutions have not been manifest in the monoclinic ZrO- pre- cipitated in the presence of UF, from MSRE-type melts at operating 112 temperatures; but, because of the potential importance to fluoride reactor fuel performance, further investigations were made of the behavior of UQo with Zr0Oz in equilibrium with LioBeF, melts containing UF, and ZrF,. In one of the series of experiments that demonstrated the absence of solid solutions, a weighed quantity of LisBeF, (~2.5 kg) melt containing about 1.3 moles of ZrF, per kilogram of melt was purified in a nickel container by HF-H, sparging. Over a period of about 70 days, successive weighed additions of UOz, Zr0s, and UF, were made. Weighed samples of the melt were withdrawn from time to time with copper filter sticks. Contin- uous helium sparging was used to provide agitation and to blanket the melt. The results of melt analysis and the estimated solid levels are summarized in Fig. 6.3, wherein the various additions to the melt are labeled A through J. (A-B) Upon successive UOp additions (A and B), the U*" concentration rose and the Zr%" concentration fell by the calculated amounts (solid lines) for the complete metathetic reaction UOs(s) + Zr*t - Ut & 2roo(s) . (3) (C-E) Over a period of 36 days, the resulting melt was equilibrated with the precipitated ZrOs and with two additions (C and D) of finely divided ( po———d | U4* @ i g£06 anl + < 204 D—D—-fl c.2 ua+ 2 =0 § i I LI L | J A B C DE F G H 1 J ¢ : ADDITIONS w UOa UOZ ZFOZ Zro UF4 UF4 UOZ UF4 UF4 s 17 1 ] { g | ~ SOLIDS LEVEL 2140 [ Zro; 1 - g ] — 2 T —— % 2 o 10 20 30 40 50 60 70 80 TIME (days) Fig. 6.3. Dissolution of U0, in Li,BeF,—5 mole % ZrF, at 600°C Followed by the Precipitation of UO, by UF, Additions. 113 approximately to that of the MSRE fuel salt, was unchanged over this ex- tended period. In particular, the constancy of the U4 concentration in the presence of up to 1.5 moles of Zr0- per kilogram of melt indicated no appreciable takeup of yst by Z2r0-. At point E, a sample of melt plus solid was taken by means of a dip stick. Microscopic and x-ray examina- tion confirmed the absence of UO- in this sample. Along with the finely divided ZrO, which had been added, well-formed ZrOs crystals grown from the melt also were found. (F) At point F, sufficient UF, was added to decrease the [Zr**]/ [U%*t] ratio to 2.2 in the melt. Although on the basis of earlier results it was expected that UOs precipitation could have intervened under these conditions, the agreement of the analyses with the solid lines in Fig. 6.3 indicated that no detectable precipitation occurred. ( The 7zr**T concentra- tion decreased only in accord with the diluting effect of the added UFy4.) (G-J) At point G, sufficient UF, was added to produce a [Zr*+]/[U%t] ratio of about 1. The subsequent rise in 7r%t concentration and the dis- appearance of U4t from the melt, shown by the divergence of the dashed lines from the solid lines in Fig. 6.3, indicated that precipitation of UO> had now occurred by the reverse of reaction (3), 72r0s(s) + U - z2r*t + UOs(s) . (4) Conformance with this indication was obtained at point H by the addition of a small amount of UO», which produced no change in the melt composition. At points I and J, still further U0z precipitation was produced by UF, sdditions. After the UF, additions at G, I, and J, the [Zr*"]/[U%"] ratios were 1.57, 1.47, and 1.34, respectively, at 600°C, indicating that the solubility ratio of Zr0O- to UOs fell with increasing concentration of Uttt and Zr%* in the melt. Previous results at much lower concentrations of U4t and zr*t gave a ratio of ~2.7. During the period I to J, equilibra- tions were carried out at 500 and 700°C as well as at 600°C; as indicated in Fig. 6.4, the effect of temperature on the solubility ratios was small in this concentration range (~1 mole of ZrF, per kilogram of melt). At the conclusion of this sequence, the melt was drawn from the nickel vessel through a filter of sintered nickel. The residue on the filter was ground and then washed with water to remove the slightly soluble fluoride salts and the finely divided insoluble material. The residue, examined microscopically, was found to contain principally well-formed, colorless Zr0- and transparent, red UOz crystals. There was no indication here of (U-Zr)0s solid solution formation nor of nucleation of overgrowth of one crystal form on the other. In a subsequent experiment the evident variation of the ratio [Zr%T] /[U%*] as a function of [Zr*"] when both Zr0s, and U0, simultaneously saturated the melt was examined further. An LisBeF, melt, initially free of Zr**t and U4+, was purified in the usual way. Additions of U0, and Zr0Os followed by UF, and ZrF, were made in amounts that were adjusted so that 114 UNCLASSIFIED ORNL-DWG 63-2619R 50 I [[1]1 o 500°C a+ s 600°C L+—zr4*] ESTIMATED « 700°C 20 57 FOR BeQ SATURATION a PREVIOUS RESULTS + : AT 600°C - = T o = —~, -+ T 5 N, v A 3 .‘ 'y o = AP o 3 r __“‘ad-—-_.____ op 2 . R T e—a A g T L § ~ 0.005 001 002 0.05 0. 0.2 0.5 1 [2+4+] (moles/kg of melt) Fig. 6.4. Ratio [Zr**]/[U%T] at U0, and Zro, Saturation of Li,BeF, as a Function of [Zr%T]. both U0z and Zr0p were expected to be present as solid phases throughout the experiment. The [Zr®*"] concentration was varied in the range 0.007 to 0.6 mole/kg. The temperature was varied from 500 to 700°C. Agitation was provided by a mechanical stirrer which was sealed at the top of the nickel vessel by a Teflon cone seal. In general, equilibration times of several hours to one day were allowed after each addition to the melt or after each temperature change. As a test of the rate of approach to equilibrium, a few samples taken less than 1/2 hr after a composition change were found to be close to the new equilibrium. The saturation ratios, [Zr4+]/[U4+], found in these measurements are plotted vs [Zr*"] in Fig. 6.4. Also, the results from the previous series, at [Zr4t] approximately 1 mole/kg, as well as some earlier data are in- cluded in Fig. 6.4. Although there was an increasing scatter in the data at the lower concentrations of U4t and Zr4+, there was a consistent trend. The previously reported saturation ratio, [Zr**]/[U4"] = 2.7, had been found in the range 0.01 to 0.03% mole of 74t per kilogram of melt. When both U0z and Zr0p saturate a melt, an equilibrium involving both reactions (1) and (2) applies, with the result that the activity ratio 4+ S [ZI' ]7ZI'4"+ KZI'OE 4+ = .8 is a constant. As the composition of the melt varies from nearly pure LioBeF, to LipzBeF, - 5 mole % ZrF, - 3 mole % UF,, some variation in the ratio 7Zr4+/7U4+ is to be expected, and this presumably accounts for the relatively gradual variation in [Zr4"]/[U4"] in the range 0.02 to 1 mole of zrt per kilogram. 115 Below 0.02 mole of zr*™ per kg, the concentration ratic rises sharply with decreasing 7, r4t concentration, the increase being strongly influenced by the temperature. The reason for this variation in the saturation ratio is not known at present. Two alternative explanations suggest themselves: 1. The saturation limit of U*T in the presence of increasing oxide ion concentration may fall more rapidly because UOo is not the saturating phase; e.g., a mixed oxide or some new U%T-02- phase (M,UO,) is formed in which M is Be2t and/or Iit and y is greater than 2. The possibility of a solid uranium-containing phase in which the activity of UO- was less than unity gained support from previously reported experiments'® in which it was found that U*t could be reduced to very low concentrations in fluoride melts ( was converted from black aggregates of UQs,,- to well-developed, transparent garnet-red crystals of similar size. The absence of color in the Zr0- suggests that little or no Ut is present in these crystals; the 116 UNCLASSIFIED PHOTO 62267 o ;(C_)-,\fi N o e Fig. 6.5. Photomicrographs of (a) U0, i, Starting Material, (b) Zr0, Starting Material, and (c) Mixture of Well-Crystallized U0, and ZrO, Resulting from Equilibration with Li,BeF,-UF,-ZrF, at 500 to 700°C. 117 garnet-red color of the UO- is usually associated with oxygen to uranium ratios no greater than 2.05. An x-ray diffraction pattern was made of this mixture using Cu KX radiation and a 0.25° 26/min scanning rate with subsequent hand settings for increased precision. Diffraction peaks were produced to approximately 135° 26. An internal standard of crystalline LiF was used. Within the limits of the method the lattice constants of the monoclinic Zr0O, and cubic UOs,gpo were not detectably different from the available standards. It is inferred from these and previous results, which show no evi- dence of (U—Zr)02 solid solution formation, that the five published phase diagrams of the Zr0s-UO> system are considerably in error at temperatures below 1000 to 1100°C. A reinvestigation of the system at lower temperatures, using a suitable fluoride fused salt as an equilibrating medium, would be of considerable interest. For the present it appears quite probable that the MSRE fuel will not be subject to uranium deposition as (U—Zr)02 as a result of oxide contami- nation. To further confirm this, (U-Zr)02 solid solutions--presently being prepared by UOs-Zr0s reaction at 1750°C--will be equilibrated in LipBeF, melts. Removal of Sulfates from Li-BeF4 Sulfates are usually present as impurities in molten-salt raw mater- ials. During normal removal by the Hp-HF purification process at the production facility, wherein equipment of copper and nickel is used, there have been instances of severe sulfide corrosion of nickel containers and piping. The study of sulfate behavior in molten fluorides 1is continuing, therefore, in order to investigate means of reducing the potential cor- rosion hazards during molten-salt purification and also to assess the possibility of significant sulfide attack on INOR-8 by sulfur impurities which could conceivably remain in the molten salt produced for the MSEE. Preliminary studies of sulfate removal by the usual HF-Ho purlflca- tion procedure, 20 g5 well as sulfate decomposition and solublllty, have already been reported. In the present measurements, removal of sulfate from molten LisBeF4, by He, Hz, and Ho-HF sparging are compared. Reactions of Sulfate. A large number of reactions are of possible interest in the removal of sulfate ion from molten fluorides, and they iead variously to the formation of S0z, S0-, metal sulfides (by‘corrosion), S-0-F, and S-F compounds. The results reported previously indicate that one mode of decomposition in the absence of reacting gases at 500 to 800°C is the formation of 50g, 50,2 = s05(g) + O°7 . (5) Although the position of this equilibrium is not known, this reaction may be used as the starting point in a consideration of possible subsequent 118 reactions. Thermodynamic data available in the literature®®:23 lead to the following reactions (AF values are for a temperature of 900°K, 627°C): At low partial pressures of SOz, partial decomposition to SOo occurs, S0 (g) = SO0z(g) + 402(g) AF = 3.45 xeal (6) 502, 503, and 8042' all may be expected to oxidize nickel to form oxides and sulfides.* ' - S0-(g) + 3Ni(s) —» 2Ni0(s) + Nis(s) AF = -28.1 kcal* (7) S0a(g) + LNi(s) - 3Ni0(s) + Nis(s) AF = -62.6 kcal* (8) 80,2" + Ni(s) - 3Ni0(s) + NiS(s) + 02" (9) In the presence of hydrogen, it has been observed that 8042- is re- duced to HsS, ' S0,%" + UHo(g) = 07 + HzS(g) + 3H=20(g) (10) S0z and SOs also are expected to be reduced, S0=(g) + 23Ho(g) » HoS(g) + 2Ho0(g) AF = -U45.0 kecal (11) S0s(g) + UHo(g) -~ HoS(g) + 3H20(g) AF = -67.9 kcal HoS, in turn, can react with Ni to form NiS,* HoS(g) + Ni(s) = Wis(s) + Ho(g) AF = -11.9 kecal (13) *Since a number of nickel sulfides exist and, often, they are non- stoichiometric compounds, the formula "NiS" is used here and elsewhere in the test to represent whatever nickel sulfide might form. In reactions 7-9 and 13, "NiS" represents the nickel’ sulfide with the lowest sulfur content formed at 900°K. Its composition is approximated by the formula NigSz. The free energies listed in reactions 7, 8, and 13 are based on the PHES/PHg ratios (Fig. 6.6) observed by Rosenquist.Z3 119 The corresponding reactions expected for copper are 4.4 keal I SO=(g) + Cu(s) = CusS(s) + 20u0(s) AF S0s(g) + 8Cu(s) - Cuss(s) + 3Cus0(s) AF = 26.1 keal S0,27(g) + 8culs) - CusS(s) + 2Cus0(s) + O~ HoS(g) + 2Cu(s) = CusS(s) + Ha(g) AF = -13.7 kecal (1%) (15) (16) (17) Thus, both in the presence and in the absence of hydrogen, sulfate may corrode nickel and copper. Nickel and Copper Corrosion. Penetrating sulfide attack on nickel can occur as the result of a eutectic which is formed between nickel and sulfur (Ni - 33 at. % S) which melts at 645°C. Above this temperature the liquid formed apparently can penetrate the metal rapidly along grain UNCLASSIFIED ORNL-DWG 63-2T749 1072 LNiyS, + Hy —= £Ni + HpS [0 © 2 I {107 /— i T 75 4 Q 8 / / / 6 g /CupS + Hy —= 2Cu + Hp5$ J4 / 400 600 800 1000 TEMPERATURE (°C) Fig. 6.6. Ratio of H,S to H, Pressures Required to Produce Sulfides of Nickel and Copper. Upper curve from Rosenquist;23 lower curve calculated from free- energy data.22 120 boundaries.®4 The equilibria (reaction 13 above) in which nickel sul- fides can be formed from Ni and HoS have been studied by Rosenquist.®3 He finds for solids containing less than 33 at. % 5, the equilibrium Py,s/Pp, ratios plotted vs temperature in Fig. 6.6. Provided the Pg_g/Py, ratio falls below this curve, no nickel sulfide will be formed, and any that is present will be reduced to the metal. Included in Fig. 6.6 is a curve (calculated®® from AF for reaction 17) which shows the Pu,s/Pp, ratio required for Cu»sS formation. From this it appears that the usual HF-Hs treatment would be most suitable for sulfate removal if the resulting PH2S/PH2 ratios do not ex- ceed the levels required for nickel and copper sulfide formation. Severe nickel attack might be avoided by holding the temperature below 645°C, thereby avoiding a liquid eutectic in grain boundaries. The alternative approach is to use a strongly oxidizing gas such as HF or Fs in the hope that S05 and sulfur oxyfluorides can be driven off rapidly before exces- sive oxidation of the metal surfaces can occur. Helium Sparging. Several runs have been made in which small batches (~250 g) of LisBeF, were melted and purified by Hso-HF sparging, and then known weights of LisS0, were added and helium sparging was continued over extended periods. Copper-lined nickel containers were used in these ex- periments. The rate of sulfate removal was followed by periodic sampling of the melt with copper filter sticks, followed by analysis for total sulfur, The results of two such runs are shown in Fig. 6.7, indicating the dependence of the removal rate on temperature. In another run at 750°C, UNCLASSIFIED ORNL-DWG 63-2720 3000 € a = 200052 ar [o—to~Jl_ 600 °C, 413 cm® OF He PER MINUTE w -;\'B‘-—. o ™ 0 H 1000 w = 800 AN S M S 600 — 4 o 750 °C, 500 ¢m>® OF He PER MINUTE (1 J\C =2 a - 75] \ | s o el 200 ' o} 2 4 .6 8 10 TIME (days) Fig. 6.7. Removal of Sulfate from Li,BeF, by Helium Sparging. 121 the rate was not noticeably influenced by changing the gas flow rate from 500 to 250 cm®/min. In a later test the rate of sulfate removal from the melt was much faster at 750°C than is indicated in Fig. 6L7, falling from 1100 to 800 ppm in 3 hr and to <100 ppm in 70 hr. However, in the same run, acidic gases (collected in a KOH scrubber) were given off slowly over a period of many days, showing a decay half-time of about 50 hr. The total amount ultimately collected was equivalent to only about one-third of the sul- fate initially present if this gas were S0s. These results indicate that the removal of sulfate can occur by several processes, which presumably include copper and nickel sulfide formation as well as S0g evolution. Even if corrosion could be prevented, the evolution of SO5 appears too slow a process to render inert-gas sparg- ing an attractive removal method. Hydrogen Reduction. In a subsequent experiment a purified Li-BelF, melt to which LisSO, had been added (to 2000 ppm S) was sparged with hydrogen at various flow rates at 600, 700, and 800°C. A nickel con- tainer without a copper liner was used. The HoS5 in the effluent gas was absorbed by ammoniacal CdCls, scrubbers, which were subsequently analyzed for sulfide. The course of this run is shown in Fig. 6.8. At the beginning of the experiment the influence of temperature on the removal rate was clearly evident, the rates being in the approximate ratios 4:7:10 at 600, 700, and 800°C respectively. The increase in removal rate with increasing Ho flow was also evident, particularly during the period from 12 to 20 hr, though the relation is less than a direct proportionality. During this same period (800°C, average Ho flow of 345 cm®/min), the sulfur was removed with an approximate half-time of 6 hr. At the end of the run the con- tainer was sectioned lengthwise, and the frozen contents were removed, ground, and sampled. The samples contained 0.0095 mile per kilogram of sulfur, all as sulfide, and 0.027 mole per kilogram of nickel. ©Since, with Ho sparging, NiZ* should not be present to a significant extent in the melt, this nickel was probably present as NiS and NiO solids sus- pended in the melt. From the material balances for the run (830 mg of sulfur added, 329 mg.of sulfur removed as HxS, 149 mg of sulfur found in the frozen melt, 360 mg of sulfur unaccounted for), only 40% of the sulfur added was removed as HoS; the remainder was removed as SOs5 or by reaction with the nickel container. The sectioned container was found to be in good condition. The upper wall of the vessel which had been exposed to the gas phase was clean and shiny. Below the surface of the melt the wall was covered with a light, blackish coating; however, there was no sign of severe penetrating attack. From this it seems clear that a portion of the L43% of the total sulfur un- accounted for was lost to the walls. Thus it appears that hydrogen alone can remove sulfate more effectively than an inert gas. The principal advantages are that removal is faster mi|limoles/min x 10% TOTAL S EVOLVED TEMPERATURE (°C) 122 UNCLASSIFIED ORNL-DWG 63-2721 400 HYDROGEN FLOW RATE 100 420 H25 EVOLUTION RATE ] 1 ] .)|. B L. e m C ® . » — B Q a o—0-0—0o -—e-® e —9-g H,S EVOLVED { millimoles } ™ 800 700 TEMPERATURE 600 0 8 16 24 32 40 48 56 64 72 80 88 96 TIME { hr) Fig. 6.8. Removal of Sulfate from Li,BeF, by Hydrogen Sparging. 123 and, although corrosive attack is not prevented, it is much less severe than with an inert gas. Removal by Ho and HF., Since, even in the presence of hydrogen, some corrosive attack on nickel was evident, yielding nickel sulfides which were not readily reduced by hydrogen, it was decided in the subsequent Ho-HF sparging experiment to add beryllium metal at the beginning as a means of rapidly reducing sulfate to sulfide., This would prevent the un- wanted reduction of sulfate by nickel with resulting nickel corrosion. Once in the form of sulfide, sulfur would present no corrosion hazard under reducing conditions., It would then be removed by HF, SHF(g) + S°7 = HoS(g) + 2F . (18) Of course, HF will oxidize nickel, Ni(s) + 2HF(g) = WNi®* + 2F + Ho(g) (19) and this, at first sight, might appear an alternate route to the formation of NiS. By combining reactions 18 and 19, H-S(g) + Ni(s) = Io(g) + Ni®T + 827 (20) it is clear, however, that these two processes are eguivalent to reaction 1% and that, provided the PHES/PHE ratio is low enough (Fig. 6.6), the activity of NiS in the melt Will Tot exceed that required to produce nickel attack. Similar arguments would apply in the case of copper corrosion. To a mixture of LisBeF, (500 g) and beryllium powder (0.621 g, 0.138 mole/kg) at 600°C in a nickel container without a copper liner, LisSO4 was added (to 500 ppm, 0.015 mole/kg). After sparging with helium for 1 hr, a filtered sample was withdrawn and analyzed, giving 390 ppm IS (0.012 mole/kg) and 259 ppm S5~ (0.0081 moleékg). After 22 hr, a sample cave 97 ppm =S (0.0030 mole/kg) and 93 ppm S5~ (0.0029 mole/kg). The reduction thus was virtually complete, producing a slightly soluble sul- fide, presumably BeS or LipS. At this point, Hy and HF sparging was begun (the course of the run is summarized in Fig. 6.9). Although it was not possible to resolve the effects of all the variables during the subse- quent sparging period, the following observations could be made. 1. During the initial sparging, the high HF flow rates used for the purpose of oxidizing the unreacted beryllium metal evidently produced ex- cesses of HoS which caused some nickel attack. This is indicated by the fact that, subsequently, HoS evolution occurred when only Heo sparging was used. Indeed, the Hp flow seemed to be the most important variable. 124 UNCLASSIFIED ORNL-DWG 63—-2722 ~Q 272 200 y A A yore O HF cm3/min L A H2 crn3/rnin 160 c A—ta E S 120 £ [&] 80 K A A 40 A A ‘ILO o HYDROGEN AND HYDROGEN FLUORIDE FLOW RATE A __E i E | | O Loaeb— g_ | l 1 t So 80 ® L4 ) o v 60 / * H,S EVOLUTION RATE o ; ‘5..) “0.\. E \ T~ 40 [ @ g ] “—.\ :;—: @ [ ] ..'. E | \ [ A N\ ¢ |3 o ¢ ° J ¢ OL_» ® ® 6 ** oo " oy """ | o e® i ) J g ° o 7 wo .. o £ % J°E ® zf2 > H,S EVOLVED el = a)”/ P Py 0@ 800 2 w 700 x D }_ < & % 600 TEMPERATURE w '_ 500 0 8 16 24 32 40 48 56 c4 72 80 88 =15} TIME {hr) Fig. 6.9. Removal of Sulfate from Li,BeF, by Hydrogen and Hydrogen Fluoride Sparging (Cubic Centimeters of HF per Minute Calculated Assum- ing Ideal Monomeric (as). 125 2., The HoS evolution rate increased with temperature. This is clearly evident at 24 hr, when the temperature was increased from 600 to 700°C. At 68 hr the increase of temperature to 800°C produced relatively high HoS evolution rates even though the amount of sulfur remaining was relagtively small. | 3. Of the 7 millimoles of sulfate added, about 6.5 was recovered as HoS. A filtered sample of the melt at the end of the run showed 12 ppm 2SS and <5 ppm S Nearly all this removal was accomplished in approxi- mately 24 hr, accumulated with Hs flow rates above 60 cm®/min. Although the container has not been examined for evidence of sulfur attack, it is apparent from the good material balance that little cor- rosion occurred. This may be attributed mainly to the use of beryllium metal reduction of the sulfate. In future tests, lower excesses of beryllium metal will be used and/or it will be removed from the melt be- fore Hs-HF sparging is begun. This should allow a more careful study of the separate effects of Ho and HF flow rate on the HoS evolution rate. It was of particular interest that the use of beryllium metal led to nearly complete removal of sulfate without appreciable loss of sulfur to metal corrosion. This could be of value in the fuel production opera- tions, wherein many batches of raw material are processed in a single vessel and cumulative effects of metal corrosion could possibly be serious. If further tests confirm the effectiveness of beryllium metal in pre- venting sulfide corrosion, its utilization in purification of fluorides would seem quite feasible. For mixtures in which UF4 was absent, beryl- lium metal could be added during the initial meltdown operation, and any excess would remain in the premelting vessel when the molten salt was subsequently transferred to the processing vessel for purification by HF-H- treatment. Physical Properties of Molten Fluorides Interfacial Behavior of Molten Fluorides with Graphite The wetting behavior of molten salts on graphite is well understood only for specific systems and usually only if the systems are not com- plicated by the presence of impurities. For example, MSRE fuels, even with additional fluorides that can occur because of corrosion or fission- ing, do not wet graphite in the absence of extraneous impurities. Among the impurities that are known to have a pronounced effect, however, oxygen is outstanding--presumably as combined oxide at the graphite-fuel interface. One of the most sensitive indicators of wetting behavior is the contact angle. In the sense used here, wetting is defined in terms of the contact angle, and 90° marks the transition from wetting to non- wetting. To study wetting behavior, a vacuumtight apparatus for observing sessile drops of fuel on graphite plaques was assembled and equipped to operate with a static or flowing atmosphere of various gases. Initial 126 experiments with LiszBeF, at 650°C in nominally pure helium showed an ex- treme sensitivity of the salt droplet to trace amounts of oxide or mois- ture at impurity levels so low that accurate measurement became a major problem. The changes in drop behavior were much less drastic and rapid when the amounts of impurities were decreased either by improved purifi- cation or by the use of static rather than flowing atmospheres. For thermodynamic and kinetic reasons, Ho0 is much more reactive to- ward MSRE melts than Os; in fact, 0o reacts with the coolant mainly through intermediary reactions with the container. Less than 10 ppm H-0 in a flowing helium atmosphere caused the formation by hydrolysis of an oxide scum on drops of LioBeF,, and at only slightly higher H-0 concentrations the scum became a shell with sufficient rigidity to prevent further change in droplet shape. Even with 400 ppm 0o at 650°C in thoroughly dried helium, changes in contact angle were not detected. No important dif- ferences in behavior were noted when moist Hs was substituted for helium or when MSRE fuel was substituted for LioBeF,. Also three types of gra- phite (AGOT, CGB and pyrolytic) were tried; neither the type nor the orientation had any detectable effect. Experience in earlier MSRE studies and extrapolation of current results both led to the conclusion that the drops remained unchanged for indefinitely long periods in sufficiently pure helium. In static helium, purified with hot titanium metal sponge, the con- tact angle, 8, between graphite and either MSRE fuel or coolant was 150 to 160° at 640°C. Thus, neglect of cos 8 (cos 155° = 0.9) involves a 10% error in estimating capillary effects of fuel in graphite, and for many purposes this is insignificant. ' The wetting, such as that shown in Fig. 6.10, when it occurs, takes place on the surface with virtually no permeation of the graphite. The field for the metallographic cross section illustrated in Fig. 6.11 was chosen to show an exceptional case where the large void near the center of the picture was partially penetrated. Although the failure of the wetting mechanism to continue into graphite pores has not been explained, Plugging by oxide and depletion of HoO available from the gas in the pores may play significant roles. In any case, chemisorbed oxygen on the gra- phite appeared to be of relatively small significance in all the experi- ments, provided that degassing of the graphite was carried out at 650°C or above., Another point of considerable significance for MSRE operation was that previous saturation of a salt droplet with oxide had little or no effect on the wetting; only when additional oxide was formed on the droplet interface, by introduction of moisture, did wetting ensue. This is in agreement with a proposed mechanism requiring oxide formation at the three-phase contact for wetting to proceed. Quantitative measures of the rate of change induced by 10 ppm H-0O are shown graphically in Fig. 6.12. 127 UNCLASSIFIED PHOTO 61167 3.27 mm= ‘r_ 644°C 2 hr 17 min AFTER MELTING 642°C 10 hr 37 min AFTER MELTING SALT: LipBefy GRAPHITE: CGB GAS: HELIUM WITH < 10 ppm Hy0 FLOW RATE: ~ 50 cm3/min Fig. 6.10. Effect of Moisture in Helium on Wetting of Graphite by MSRE Coolant for Different Elapsed Times After Melting. For experiments involving LipBeFy, which was preferred because of its transparency, the precipitated oxide was presumed to be BeO. To con- form that BeO is wetted by LioBeF4, several trails were carried out; with the exception of one as yet unexplained case, wetting ensued shortly after melting of fluorides on BeO plaques in four experiments. Fig. 6.11. Wetting Layer of Li,BeF, on AGOT Graphite After Freezing in Place. Reduced 26%. UNCLASSIFIED ORNL-DWG 63-64 130 ——— T ANL=D) L L1 LL] CGBL-L,B-4 TEMPERATURE : 480°C FLOWING HELIUM: 50 cm®min | f—++++ CONTAINING 10 ppm OF H,0 ‘ CONTACT ANGLE (deg) 3 5] oL HEIGHT OF DROP| I B e [ 3 Tt | | E 2 1 T | 72 BASE OF DROP | | | e et | | b il INERR | | | | ol IR L AT | o1 o2 o5 2 5 0 20 50 100 TIME AFTER MELTING (hr) Fig. 6.12. Effect of Moisture on the Characterization of a Drop of LipBeF, on Graphite. Sessile 129 The possibility that protons rather than oxide from H-0O served as the surface-active agent was not favored by demonstrations that NH,HF> and KHFs did not wet graphite, at least initially; however, KHFs, evolving HF as it melted, developed contact angles of less than 90° in a few minutes. Viscosities of MSRE Fuel and Coolant The viscosities of an MSRE-type fuel mixture, LiF-BeFo-ZrF,-UF, (66.5- 28.7-4.0-0.8 mole %) and coolant, LiF-BeFso (66~ 31 mole %) were measured over the range 475 to 625°C with a commercial rotational viscometer (Brookfield). The viscometer was calibrated with glycerine-water solutions. The data for the salt mixtures are given in Fig. 6.13. Linear log 1 - (l/T) relations appear to represent the data adequately. The equations of the correlating lines in Fig. 6.13 are 1 = 0.0916 exp (4098/T) for the fuel mixture and 1 = 0.116 exp (3755/T) for the coolant, where 1 is the viscosity in centipoises and T is temperature 'in °K, The fuel mixture on which these measurements were made was an interim composition which was derived from the specified’ fuel for- highly enriched uranium (0.2 mole % of UF.) by increasing the UF, to: '9:85 mole % and cor- respondingly decreasing LiF and BeFs by equail decrements.. The fuel sche- duled for operation in the MSRE, LiF-BeF,-ZrF,-UF, (65.0-29.1-5.0-0.9 UF4), was estimated to have a viscosity of 5% higher than the foregoing mixture, #, TEMPERATURE (°C) 30 625 600 550 500 475 | T T T ® FUEL. MIXTURE (LiF-BeF,-ZrF,-UF,, 66.5-28.7-4.0-0.8 mole % ) o o~ L] 20— O COOLANT (LiF-BeF,, 66-34 mole % ) i ’./i;:/ / ey ,/’/’ 7, VISCOSITY (centipoises ) . L1 | | 1.0 1.1 1.2 1.3 1000/ (o Fig. 6.13. Viscosity-Temperature Relation for MSRE Fuel Mixture and Coolant. 130 and hence to follow the equation n = 0.0962 exp (L4098/T). At tempera- tures that are pertinent to MSRE operation, the viscosities are 8 centi- poises at 650°C for the fuel and 10 centipoises at 570°C for the coolant. Oxidation-Reduction Reactions in MSRE Melts Chemical Reduction of Dissolved Fluorides of Structural Metals Structural metal fluorides in molten fluoride fuels or coolants in- fluence the corrosion behavior if present in other than equilibrium amounts, and in large excess they could conceivably change the pattern of oxide -precipitation. Consequently, in the preparation of molten fluoride mix- tures for the MSRE and in the subsequent fluoride processing operations, good control of the concentrations of such impurities is important. Ac- cordingly, the rates at which the cations of structural metals can be removed from MSRE-type melts by hydrogen sparging and by using stronger reducing agents, such as beryllium metal and zirconium metal, have been studied. Other objectives of these experiments were to improve fluoride purification techniques and to serve as a guide for fuel processing in the MSRE. Additional studies of the corrosion rates associated with the use of various HF-H, mixtures for oxide removal were planned. Reduction of FeF- by Hydrogen. The first experiments were with ap- proximately 2 kg of LiF-BeFs (66-34 mole %) to which FeFo along with a small amount of Fe’’ as a tracer had been added. Following dissolution of these materials by hydrofluorination, the rates of removal were deter- mined by analysis of salt samples withdrawn during various conditions of hydrogen sparging. The results shown in Fig. 6.14 indicate the importance of hydrogen flow rate in providing better liquid-gas contact and of the temperature of the melt in which diffusion in the liquid film is probably the controlling process, Since the reduction rates were controlled by gas-liquid contact con- ditions that were not well defined, the small-scale study could not be directly scaled to larger equipment. Accordingly, studies were also made in the large production facility with ~110 kg of the mixture LiF-NaF-ZrF, (27.5-27.5-45.0 mole %) initially containing about 2000 to 3000 ppm by weight of iron. Since small-scale experiments demonstrated much faster reduction rates at higher temperature, the production facility experi- ments were conducted at a "maximum advisable temperature for sustained operations" of 700°C. The rates, presented in the form of first-order reaction constants as a function of H, flow rate in Fig. 6.15, indicated that hydrogen sparging rates of about 10 liters/min would accelerate the reduction rate by about 300% over those obtained with the hydrogen flow rate of 3 liters/min customarily used in the past. Also, there was little to be gained with flow rates greater than 10 liters/min. On MSRE melts, an initial iron concentration of 200 to 300 ppm in ~70 kg of LiF-BeFs (66-34 mole %) was used; the melt temperature was again maintained at T00°C. The saving in time at faster flow rates as- sociated with the rate constants shown in Fig. 6.16 demonstrated the UNCLASSIFIED ORNL-DWG 63-6489 150 l T T TEMPERATURE AV (°c) [Fe®* ], (ppm) e 800 600 o 700 : 550 125 A 600 700 ol o ] w REMOVAL RATE OF Fe2* (pmn/nr) |\ : \ \, 25 y V4 / oy A 0 0.5 1.0 15 HYDROGEN FLOW RATE (liters /min) Fig. 6.14. Reduction of Fe?t in LiF-BeF, (66-34 mole %) at 600 to 800°C by Hydrogen Sparging. advantage of using higher hydrogen sparging rates for removing structural metal impurities from molten fluorides and more than compensated for any increased volumes of Ho that might be required. Reduction of Structural Metals by Beryllium and Zirconium. The use of strong reducing agents such as beryllium or zirconium for removing structural metal impurities from molten salts should under many circum- stances be more effective than the current technique of hydrogen sparg- ing. However, the use of these elements has frequently led to obscure side effects that must be evaluated before the adoption of strong re- ducing agents in routine fluoride purification processes, and experiments toward this end have been initiated. In one experiment, weighed increments of beryllium metal turnings were added to a molten mixture of LiF-BeFs (66-34 mole %) at 600°C which initially contained about 800 ppm of iron and about 200 ppm of chromium. The analyses of salt samples withdrawn prior to each addition of beryl- lium are shown in Fig. 6.17. A reaction period of about 4 hr between additions was assumed to be adequate for complete reaction; the melt was agitated by helium sparging during the entire reaction period. 1In a similar experiment (Fig. 6.18), a relatively large concentration (5000 ppm) of chromium was removed from solution in LiF-BeFs (66-34 mole %) at 600°C by additions of zirconium metal turnings. 132 UNCLASSIFIED ORKL-DWG 63-6490 0.10 £ ¢ z / ® % 0.08 < / - w < Q (8] L 008 Q Q ] ul > P4 2 0.04 5 / QUANTITY OF SALT #Okg (ASSUMED W 1st ORDER REACTION DEPENDENCE x ON Fe2* CONCENTRATION ) > £ 002 . - w x *U 0 0 5 10 15 20 25 HYDROGEN FLOW RATE { Iifers/min ) Fig. 6.15. Effect of Hydrogen Flow Rates on the Relative Reaction Rate for Reduction of Fe?t from Solution in LiF-NaF-ZrF, (27.5-27.5- 45,0 mole %) at 700°C in the Fluoride Production Facility. UNCLASSIFIED ORNL-OWG 63-6494 N QUANTITY OF SALT: 70kg (ASSUMED {st ORDER T REACTION DEPENDENCE ON Fe?* CONCENTRATION) Z INITIAL Fe2* CONCENTRATION; 200-300 ppm = 0.08 = < - w = o] © [ = Z o006 " G / o] ® - w0 / > /. = o £ 004 // Q L4 wl x Y /' £ 002 / - Lt o O 0 0 2.5 5 7.5 10 12.5 HYDROGEN FLOW RATE (liters /min ) Fig. 6.16. Effect of Hydrogen Sparge Rates on the Relative Rate of Reduction of Fe?t from Solution in LiF-BeF, (66-34 mole %) at 700°C in the Fluoride Production Facility. 133 UNCLASSIFIED ORNL-DWG 63-6492 25 1 I O TOTAL Cr, Ni AND Fe ® Fe A Cr 20 | A Ni WEIGHT OF MELT: ~2 kg o\‘ \0\ .\é%\i\\\\ \\\‘ o\\e bt ® \i\\: LY — \ 0] - (=] [ 5 METAL FOUND IN SOLUTION ( moles /kg x10%) N A A 2 04 0.2 0.3 0.4 0.5 BERYLLIUM METAL ADDED(g) o) Fig. 6.17. Reduction of Structural Metal Fluorides from Sclution in LiF-BeF, (66-34 mole %) at 600°C by Beryllium Metal. UNCLASSIFIED ORNL -DWG 63-6493 . o _—END POINT FOR Cro* — ¢e?* = - E . = g0 E 3 ® =z o E 60 | O W = S 35 40 (@] [T = [ ] S = 2 20 \ g @ z END POINT FOR : . cr?t— ¢ 0 [ ] o) 20 40 60 80 100 ZIRCONIUM METAL ADDED (millimoles/kg) Fig. 6.18. Removal of Chromium from Solution in LiF-BeF, (66-34 mole %) at 600°C by Additions of Zirconium Metal. 134 Both these experiments gave favorable results for effective removal of structural metal contaminants from molten fluorides, but the apparent stoichiometry of the reacting species was probably complicated by the fact that iron and chromium may be present in both their divalent and trivalent chemical forms. Because of these and other complications re- lated to corrosion and fuel stability, further studies of the effects of elemental beryllium and zirconium are continuing. Apparent Solubility of UF5 Produced by Reducing Fuel with Excess Zirconium Metal Continuing investigations of the behavior of molten fluorides con- taining UFg (ref 25), which have a bearing on the limits of chemical re- duction consistent with stability in MSRE fuels, included determinations of the solubility of UFs. Approximately 1.75 kg of the mixture LiF-BeFs-ZrF4-ThF,-UF, (70-23- 5-1-1 mole %), containing about 6 wt % of uranium, was heated to 600°C in a nickel reaction vessel and treated with anhydrous HF and hydrogen to further ensure melt purity. Approximately 25 g of specially prepared zirconium metal turnings was added in weighed increments to the cooled reaction vessel. The zirconium was expected to form ZrF, by reducing UF, to UF3. Following each addition, the mixture was heated to 800°C and sparged with dry helium. Filtered samples of the salt were obtained after each reaction period, using sintered copper filters, and analyzed by a hydrogen~-evolution method that does not necessarily distinguish trivalent uranium from other reducing agents, but no others should have been pres- ent. At 800°C approximately 73% of the uranium was found as U3" by the analytical method, and no significant loss of total uranium was noted; the final zirconium metal additions corresponded to a cumulative 500 ex- cess for the complete conversion of all UF, to UFs, and hence some metal should have remained. After the final solubility measurements, the liquid at 600°C was decanted from the reaction vessel; the solid material which remained contained UFg apparently as a primary phase according to petro- graphic and x-ray diffraction examinations. The uranium content of the solid was found to be about 40 wt %, compared to 80% for the uranium con- tent of UFz. Solubilities shown in Fig. 6.19 for 600 and 700°C were obtained after 8 and 20 hr at temperature. Equilibrium temperatures, read with a Pt vs Pt -~ 10% Rh thermocouple, were approached from both heating and cooling directions. Solubilities at 650 and 550°C were found after 8 hr at equilibrium températures reached by cooling conditions. Con- verted to mole percent, the solubilities range from about 0.1 at 550°C to about 0.7 at 700°C. The heat of solution, calculated from the slope of the line in Fig. 6.19, was 21.6 kcal/mole. One puzzling aspect of the analyses was that the U*T concentration apparently remained constant at about 2 wt % over the temperature range studied and hence that the expected equilibrium LUF3 = 3UF, + U°, involving increasing proportions of UF, to UFg at higher temperatures, was not recognizable. Also unexplained were the lack of a more complete reduction in response to the excess of zirconium metal as well as the difference from previously reported results by another technique®® for a similar composition in which solubilities of UF5 of 0.45 mole/kg at 471°C were implied. 135 UNCLASSIFIED ORNL-DWG 63-6494 TEMPERATURE (°C) 700 6 0.30 50 GCI)O 550 0.20 \\ 0.15 AN SN N 0.08 q\\ 0.06 \\\ 0.04 \ o | | \ 10.0 10.5 1.0 1.5 12.0 12.5 1000/7 (ox U** CONCENTRATION FOUND (moles per kg of salt ) Fig. 6.19. Effect of Temperature on the Apparent UF; Content of LiF-BeF,-ZrF,-ThF, -UF, (70-23-5-1-1 mole %) After Reduction with zr0. - Development and Evaluation of Methods for the Analysis of the Radioactive MSRE Fuel Efforts have continued since the last report period on the develop- ment and evaluation of methods of analysis of the radiocactive MSRE fuel. The "mockup" hot-cell facility has been used in most of this work in order to simulate the operating conditions necessary to analyze highly radio- active materials. The evaluation of a pyrolysis method for the determina- tion of fluoride has been successfully completed in the mockup facility, and an amperometric method for the titration of chromium is being tested. A modified technique for the determination of oxygen by inert-gas fusion is being investigated. 136 Fluoride The pyrolytic determination of fluoride®? is carried out in an appa- ratus designed for remocte operation. A nickel reaction tube is first heated to 1000°C, and moist oxygen is passed through it at a rate of about 2 liters/min. A test portion, usually 100 mg of the salt, which has been pulverized for ease in sampling and to ensure homogeneity, is placed in a nickel boat and mixed with about 3 g of powdered UzOg, which acts as a catalyst for the evolution of fluoride. The boat is inserted in the heated reaction tube, which is immediately sealed to prevent loss of fluoride. The evolved fluoride is trapped in a known gquantity of sodium hydroxide. After 40 min the absorber solution is removed from the apparatus, and the excess sodium hydroxide is titrated with standard hydrochloric acid to a potentiometric end point. The fluoride is equivalent to the sodium hydro- xide neutralized by the absorbed hydrogen fluoride. In analyses performed remotely on nonradioactive materials, a relative standard deviation of 1% was obtained. Chromium An amperometric method for the titration of Cr(VI) with ferrous sul- fate has been developed for the determination of chromium in MSRE fuels. A titration cell assembly, which includes a pyrolytic graphite electrode and a standard cglomel electrode, was specially designed for this method. Chromium, in 5- to 50-ug quantities, is oxidized to Cr(VI) with argentic oxide in 0.5 M HoS04. The Cr(VI) is subsequently reduced to Cr(III) by titration with ferrcus sulfate; this reduction causes a decrease in dif- fusion current, which is measured amperometrically. The progress of the titration is followed with an ORNL model Q-1160 polarograph, with the pyrolytic-graphite electrode at +1.0 respective to the S.C.E. The titra- tion is continued until a sharp deflection in the titration curve is observed. When the MSRE fuel is dissolved in sulfuric acid, the titration of chromium can be performed without any prior separations. Application of the method is now being tested in the mockup facility with nonradio- active samples. The precision of '"bench-top" analysis is 1%, the ultimate precision that can be expected. Oxygen A radio-frequency concentrator was incorporated in the ignition chamber of the modified Leco analyzer that is being used to study the determination of oxygen in MSRE fuels by inert-gas fusion procedures. The purpose of the concentrater is to provide temperatures sufficient for the evolution of oxygen as carbon monoxide from uranium and zirconium oxides present as contaminants in the fluoride fuels. The new ignition chamber was fabricated from a type RF-20 Sylvania radic-frequency lamp in which the rf energy from an external coupling coil is concentrated to heat a titanium carbide emitter (target) to tempera- tures as high as 4100°C. The concentrator consists of a water-cooled metal cylinder (2 in. 0.D. by 3 in. long) with its lower end open and its upper end perforated with a 5/8—in. beveled hole in which the rf energy is 137 concentrated on the target. The cylinder is slotted on a radius. The lamp was adapted by sealing a helium entrance tube to the bottom of the envelope, removing the target, and sealing, over the hole in the concen- trator, a vertical tube that serves as a loading port and gas exit line, In operation the graphite capsule containing the sample is suspended in the center of the hole in the concentrator from a threaded graphite rod, which also serves to seal the capsule. The position of this capsule is determined by an enlarged spindle that is machined on the top of the support rod to fit the loading tube closely. At the end of the analysis the ignited sample is removed from the ignition chamber by increasing the gas flow to raise the capsule through the loading tube. The maximum temperatures obtained with the concentrator have not been messured. However, from observations of color temperature made with the Leco generator operating at fractional power, the new ignition chamber provides better heating than a more elaborate assembly in which the coup- ling coil is placed within the envelope.®® Also, because the concentrator partially shields the envelope from the heated capsule, the chamber can be constructed of borosilicate glass rather than quartz, and a large ground glass Jjoint is eliminated. In initial tests with samples of uranium and zirconium oxides, re- duction to the carbides was obtained within a 5-min ignition period. Thus temperatures in the order of 2400°C are indicated.®® Determination of oxygen in MSRE samples which have been analyzed by the KBrF,3° method will be performed by comparison. ' Nickel The statistical evaluation of nickel in MSRE fuel has been completed in the mockup. The dimethylglyoxime spectrophotometric method with potas- sium persulfate oxidation was employed. A solution of 1 g of MSRE fuel in 100 ml of 0.5 M HoS50, was used as the test soclution. Into a 50-ml flask was pipetted MT5 ml of the test solution containing 39 pg of nickel. The reagent solutions were added, with slurrying between each addition, as follows: 30 w/v % ammonium citrate, 5 ml; saturated potassium persulfate, 10 ml; after 10 min, concentrated ammonium hydroxide, 10 ml; 1 w/v % dimethylglyoxime in 95% alcohol, 2 ml; and distilled water (amount re- quired to make final volume to 50 ml). The solution was allowed to stand for 30 min, and the absorbancy was measured against a blank prepared in the same manner, including the test solution but without the dimethyl- glyoxime reagent. This sample blank was necessary to cancel the inter- ference due to the uranium color and the slight change in pH caused by the acid content of the test sample. The absorbancy was obtained with the remotely operated filter photometer, fitted with a 450-mp filter., The nickel content of the test sample was obtained by comparing its absorbancy with a standard curve prepared in the same manner, containing the same acid content but without the uranium interference. The standard curve was prepared in this way to eliminate the need for a pH meter since the method is pH dependent. Test solutions containing nickel concentrations as low 138 as 0.5 ug/ml can be determined by this method. The relative standard de- viation of the method was 2.8% Development and Testing of Equipment in the Hot-Cell Mockup The prototype equipment for handling the transport container to re- move the sample-filled copper ladle, known as the Transport Container Decoupling Device, has been completed and tested. Upon continued testing, the gear box of the bear-box-motor section proved weak. Its correction is being studied. In separating the powdered sample from the crushing container (pulverizer-mixer) and copper ladle, several difficulties were encountered with additional testing in the method previously outlined. To overcome these, two devices were designed: one to hold and align the pulverizer- mixer during assembly prior to pulverizing the sample and another to effect the transfer of the powder to the polyethylene bottle from which powdered aliquots will be taken for analysis. These new devices are called the Pulverizer-Mixer Aligner and the Powder Transfer Device re- spectively. Their prototypes have been tested and modified, and the improved versions are being fabricated. The pyrolytic apparatus for fluoride determinations has been satis- factorily tested on MSRE fuel in the mockup. It consists of a specially adapted 8-in. tube furnace and nickel combustion tube. The adaption enables remote insertion and withdrawal of the nickel boat containing the sample and catalyst plus the plugging and unplugging of the combus- tion tube. The tube furnace is fitted with a Chromel-Alumel thermoccuple. The thermocouple output is led outside the mockup cell and connected to a "Sim-ply-trol" pyrometer control and a variac, thus permitting a tempera- ture control of + 10° at 950°C. In order to combine into a single apparatus the electrochemical cells needed to perform amperometric, coulometric, and polarographic de- terminations in the hot cells, a variable electrode cell system was devel- oped. It is approximately 80% complete and is presently being used in the ampercmetric method for the determination of chromium. Other phases of the work necessary for handling and analyzing the MSRE fuel samples (such as weighing and its associated problems and deter- minations of iron, chromium, etc.) in the hot cell are in progress. 10. 11. 12. 13. 1k, 15. 16. '17. 18. 139 References MSRP Quart. Progr. Rept. Jan. 31, 1958, ORNL-2474, p 1. MSRP Semiann. Progr. Rept. Feb. 28, 1961, ORNL-3122, p 97. Ibid., p 96. MSRP Semiann. Progr. Rept. Feb. 28, 1962, ORNL-3282, p 116. U. M. Korenev and A. V. Novoselova, Dokl. Akad. Nauk SSSR 1k, 5 (1963). C. F. Weaver, "Petrographic Examination of Irradiated MSRE Fuel (ORNL- MTR-47-3)," Nov. 8, 1961, unpublished internal communication. C. F. Weaver, "Petrographic Examination of the MSRE Salt, LiF-BeFso- ZYF4-UF,-ThF, (70-23.3-5-1-0.7 mole %), in Capsule 36 of ORNL-MTR- Lr-b," Oct. 31, 1962, unpublished internal communication. C. F. Weaver, "Petrographic Examination of Irradiated MSRE Salt (ORNL-MTR-47-4)," Apr. 19, 1963, unpublished internal communication. C. F. Weaver and R. E. Thoma, "Petrographic and X-Ray Diffraction Observations of Electron-Irradiated MSRE Salt," Apr. 30, 1963, un- rublished internal communication. C. F. Weaver, "Petrographic Examination of an Irradiated MSRE Salt (ORNL-MTR-L44-2)," Dec. 5, 1961, unpublished internal communication. C. F. Weaver, "Petrographic Examination of the MSRE Salt in Capsule No. 24 of ORNL-MTR-47-L," Sept. 21, 1962, unpublished internal communication. C. F. Weaver, "Petrographic Examination of Salt Sample from Test of Graphite Compatibility under Irradiation (ORNL-MTR-47-2)," Oct. 6, 1960, unpublished internal communication. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 82. C. F. Weaver, personal communication. J. H. Burns, R. D. Ellison, and H. A. Levy, J. Phys. Chem. 67, 1569 (1963). MSRP Semiarnn. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 109. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 110. I. Cohen and B. E. Schaner, A Metallographic and X-Ray Study of the UO5-Zr0s System, WAPD-253 (June 1962). 19. 20' 21. 22. 23. 2k, 25. 26. 27. 28. 29. 0. 140 Reactor Chem. Div. Amn. Progr. Rept. Jan. 31, 1961, ORNL-3127, p 8. MSRP Semiann. Progr. Rept. Feb. 28, 1961, ORNL-3122, p 120. MSRP Semiann. Progr. Rept. Feb. 28, 1962, ORNL-3%282, p 118 J. F. Elliott and M. Gleiser, Thermochemistry for Steel Making, vol I, Addison-Wesley, Reading, Mass., 1960. T. Rosenquist, J. Iron and Steel Inst. 1954, p 37. J. H. DeVan, ORNL Metals and Ceramics Division, personal communica- tion, March 1963. MSRP Semiann. Progr. Rept. Aug. 31, 1962, ORNL-3369, p 100. MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 109. R. H. Powell and O. Menis, Anal. Chem. 30, 1946 (1958). MSRP Semiann. Progr. Rept. Jan. 31, 1963, ORNL-3419, p 1k4O. Edgar J. Beck and Forrest E. Clark, Anal. Chem. 33, 1760 (1961). G. Goldberg, A. S. Meyer, Jr., and J. C. White, Anal. Chem. 32, 31k (1960). 7. FUEL PROCESSING The flowsheet for the MSRE fuel-processing system is shown in Fig. 7.1. Design of the system was reviewed, and minor changes are being made to the drawings. Changes include provisions for remote replacement of the NaF bed, since plugging with chromium is a possibility during fluorination, and addition of an activated-charcoal trap to improve iodine and tellurium removal from the off-gas system. A water analyzer for the H,-HF treatment is being developed by the ORGDP Analytical Instrumentation Section. It is hoped that a commercial hygrometer can be used following an NaF trap for removal of HF. UNCLASSIFIED ORNL—DWG 63-3123A SALT 200°F NaF ABSORBERS IN GUBICLE SAMPLER W ————— Iy ‘/|--——H20 ANALYZER . [ IN GUBIGLE SALT CHARGING ,J H 5 E 5 H 4 s HIGH BAY AREA H —- D OONNELLY MANTHOS* TAYLOR" YENARD* . G. BRYSON* . L. GRIFFITH METALLURGY IXIITTIFIIT X FLOW MODELS FTF¥OLPRNET MECHANICAL DESIGN BRADFORD, JR.* BROWN HURTT JONES KERR NICHOLSON ROBERTS ROBINSON L. ROUSER PORINET . G. STERLING WMORNAODADAODWAAR R. J. KEDL* - B. J. YOUNG e REMOTE MAINTENANCE Az SITE PREPARATICN E. DUNWOODY™" E. NORTHUP* E. SALLEE P&E P&E PRE R. L. MOORE 15C J. R. TALLACKSON 1&C 5. J. BALL® 1&C H. R. BRASHIER" 18C J. R. BROWN B&C G. H. BURGER 1&C T. M. CATE 1B8C D. G. DAVIS 1&C s. 1. DITTO : 1&C P. G. HERNDON 1&C B. J. JONES* 1&C J. W. KREWSON 1&C J. L. REDFORD** 18C C. E. STEVENSON 1&C R. WEIS 18c ANALYSIS P. N. HAUBENREICH R J. R. ENGEL R B. E. PRINCE R FUEL PROCESSING R. B. LINDAUER cT E. C. HISE* R. BLUMBERG P. P, HOLZ* C. JONES J. R SHUGART DMARADODR COMPONENTS SCOTT AC ANALYTICAL CHEMISTRY DIVISION CT CHEMICAL TECHNOLOGY DIVISION D DIRECTOR'S DIVISION I1&C INSTRUMENTATION AND CONTROLS DIVISION M METALS AND CERAMICS DIVISION P&E PLANT AND EQUIPMENT DIVISION RC REACTOR CHEMESTRY DIVISION R REACTOR DIVISION * PART TIME ON MSRP “* DUAL CAPACITY *** EURATOM EXCHANGE SCIENTIST W. B. McCDONALD ® PROCUREMENT AND FABRICATION C. K, MCGLOTHL AN R F. L. ROUSER R J. M. TEAGUE R CONSTRUCTION B. H. WEBSTER R N. E. DUNWOODY™ P&E R. 5. JACKSON* P&E J. P, JARVIS P&E J. W. FREELS P&E M. L. LOVE P&E L. P. PUGH R w. D. TODD P&E R. SMITH, JR. SCHEDULING W. W. GOOLSBY* P&E INSTRUMENTATION AND CONTROLS INSTALLATION J. L. RECFORD** 18C R. W. TUCKER 18C RICHARDSON N. SMETH** 5. WEBER C. A. GIFFORD TFFED AWV LOOPS L. CROWLEY B. GALLAHER R. E. CARNES W. H. DUCKWORTH A - DWW ™WT PUMP DEVELOPMENT A. G. GRINDELL" C. H, GABBARD* P. G. SMITH a0 REACTOR CHEMISTRY EITA » proTTm BEALL* GUYMON HARLEY** PAYNE** ULRICH*~ MSRE OPERATIONS Tomom™ D W..R. GRIMES* RC H. F. MCOUFFIE- RC F. F. BLANKENSHIP RC C.,F. BAES" RC E. L. COMPERE* RC H. C. SAVAGE* RC J. H. SHAFFER RC R. E. THOMA" RC . F. A DOSS RC . E. EORGAN RC H. A. FRIEDMAN* RC G. M. HEBERT RC S. S. KIRSLIS RC P. J. KREYGER"** RC . W. K. R. FINNELL RC | B. F. HITCH RC ¥ W. JENNINGS RC w. P. TEICHERT RC - ANALYTICAL CHEMISTRY 1 L. T. CORBIN" AC ). C. WHITE* AC R. F. APPLE AC M. J. GAITANIS AC Ww. L. MADDOX" AC ] RADIATION TESTING ‘ 1. A CONLIN R W, I;i!. MIXON R S VT S (] O O W oo H HE R Y ON P ENEOL P INGE LT C RSN N N OuEEEE YN OE O Adamson Alexander Baes Beall Bettis Billington Blankenship Blizard Boch Bohlmann Bolt Borkowski Boyd Breeding . Briggs . Bruce antor Cardwell Conlin Cook Corbin Cristy Crowley Culler DeVan Donnelly Douglas Engel Epler Ergen . Fraas Frye, Jr. Gabbard Gall Gallaher Grimes Grindell Guymon Harrill Haubenreich . Hibbs (Y-12) . Hill . Hise QUWuuuEHGEQEHEYEH O LEHAGQRZ OBU*ij-ZED'ffl_CD'EU_UJ?U_m'm*U?i*'U':U:D@mIT'b':J>l—:Im:l>2 145 ORNL-3529 UC-80 — Reactor Technology PTD-4500 (24th ed.) Internal Distribution P E RO R ORISR0 AU AN ORI N EEEEINS 0N B TR e W. P. Hof fman Holz Hollaender ':-L:l':',UI.:-Lj:'Z'.I._—"'UONL_'OW'TQUJ'IJUQHmbflflbmblfl%wmzm Householder Howell Jordan Kasten Kedl Kelley Lane Larson {K-25) Lincoln Lind Lindauer Livingston Lundin MacPherson Manly . Mamm McDonald . McDuffie . McGlothlan . Miller . Moore . Morgan Moyers Murray (K-25) Nelson Nestor Northup Osborn Parsly Patriarca R. Payne Phillips B. k. Pike Prince Richardson = = = E P! Robertson Roche Rosenthal Savage Savolainen 146 J 88. D. Scott 105. W. C. Ulrich ’ 89. H. E. Seagren 106. D. C. Watkin 90. J. H. Shaffer 107. G. M. Watson 91. E. D. Shipley 108. B. H. Webster 92. M. J. Skinner 109. A. M. Weinberg 93. G. M. Slaughter - 110. J. H. Westsik o4k. A. N. Smith 111. J. C. White 95. P. G. Smith 112. L. V. Wilson 96. A. H. Snell 113. C. H. Wodtke g7. I. Spiewak 114, Biology Library 98. C. D. Susano 115-116. Reactor Division Library 99. J. A. Swartout 117-120. ORNL—Y-12 Technical Library, 100. A. Taboada Document Reference Section 101. J. R. Tallackson 121-123. Central Research Library 102. E. H. Taylor 124-153. Laboratory Records Department 103. R. E. Thoma 154. Laboratory Records, ORNL R.C. 10k. D. B. Trauger External Distribution 155-156. D. F. Cope, AEC, ORO 157. R. W. McNamee, Manager, Research Administration, UCC, New York 158. R. L. Philippone, AEC, ORO 159. W. L. Smalley, AEC, ORO 160. M. J. Whitman, AEC, Washington 161. Division of Research and DeveloPment AFC, ORO 162-776. Given distribution as shown in TID- H5OO (24th ed.) under Reactor Technology category (75 copies - OTS) [¢ ) ORNL-~24Th ORNL-2626 ORNL-2684 ORNL-2723 ORNL-2799 ORNL-~-2890 ORNL-2973 ORNL-301k4 ORNL-3122 ORNL-3215 ORNL-3282 ORNL~3369 ORNL-3419 147 Previous reports in this series are: Period Ending Jamuary 31, 1958 Period Ending October 31, 1958 Period Ending January 31, 1959 Period Ending April 30, 1959 Period Ending July 31, 1959 Period Ending October 31, 1959 Periods Ending January 31 and April 30, 1960 Period Ending July 31, 1960 Period Ending February 28, 1961 Périod Ending August 31, 1961 Period Ending February 28, 1962 Period Ending August 31, 1962 Period Ending January 31, 1963