/G [NEFANTD 3 445k 0361553 1 .u. . :“l led F % QOGUMER] ORNL=-3014 UC-81 - Reactors = Power MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT FOR PERIOD ENDING JULY 31, 1960 CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON If you wish someone else to see this document, send in name with document and ‘the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION Printed in USA. Price .$_2£5__. Available from the Office of Technical Services Department of Commerce Washington 25, D.C, ¥ LEGAL NOTICE This report was prepared as an account of Government sponsored work. Neither the United States, nor the Commission, nor any person acting on behalf of the Commission: A. Makes ony warranty or representation, expressed or implied, with respect to the accuracy, completeness, or usefulness of the information contained in this report, or that the use of any information, apparatus, method, or process disclosed in this report may not infringe privately owned rights; or B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of any information, apparatus, method, or process disclosed in this report. As used in the obove, “person acting on behalf of the Commission'' includes any employee or contractor of the Commission, or employee of such controctor, to the extent that such employee or contractor of the Commission, or employee of such controctor prepares, disseminates, or provides occess to, any information pursuant to his employment or contract with the Commission, or his employment with such contractor. ORNL-301kL Uc-81 — Reactors — Power TID-4500 (15th ed.) Contract No. W-T405-eng-26 MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT FOR PERIOD ENDING JULY 31, 1960 H. G. MacPherson, Project Coordinator DATE ISSUED 15108 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION " Hfl‘{!flin "I’W ifl]flr‘i "fl‘ i"‘m""i‘vs‘m il 3 4456 0361553 1 SUMMARY PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE Design The Molten-Salt Reactor Experiment is a logical extension of the program con- ducted at ORNL during the past ten years in the investigation of molten fluoride mixtures and container materials for circulating-fuel reactors. The major objec- tives of the MSRE are to demonstrate the safety, dependability, and serviceability of a molten-salt reactor and to obtain additional information about graphite in an operating power reactor. Conceptual designs of the MSRE have been made in which the core is construc- ted of vertical graphite stringers in which fuel passages are machined. Flow of fuel is two-pass through the core, with downflow through an annulus along the core-vessel wall and upflow through the moderator passages. The heat-removal cycle is from fuel salt to a secondary coclant salt to air. No control rods are needed because the fuel has a temperature coefficient of re- activity of approximately -4 x 1077 (&k/k)/°F. A dual-purpose sampling and enriching system is provided for the removal of semples for chemical analysis or the addition of enriched fuel. This sampler is located at the pump expansion volume. In the design of the core, consideration was given to the Poppendiek effect, and the flow passages were designed to minimize the resulting temperature rise. The primary heat exchanger is designed for 10-Mw duty. It is a conventional U-tube heat exchanger in a 25% cut baffled shell. The fuel is in the shell and the coolant salt is in the tubes. The radiator is designed to fit the existing space in the air duct in Build- ing 7503. It is a bare-tube radiator with large tubes to minimize the possibility of salt freeze-up. Four drain tanks are required for the MSRE, two for the fuel salt, one for the flush salt, and one for the secondary salt. A steam boiler is provided for after-heat removal in the fuel drain tanks. The system layout was designed to facilitate remote maintenance in the primary circuit and in the fuel drain tank pit. A cover-gas system is being designed to remove fission products from the gas stream and minimize the radiation damage effects on the pump-lubricating oil sys- tem for bearings and seals. A study of the requirements for a remote maintenance system for the MSRE is in progress. Additional studies are being made of the modifications to Building 7503 required to accommodate the construction of the reactor. 2. Component Development A program of component development and testing in support of the MSRE design is designed to improve the reliability of the freeze flanges, freeze valves, sam- pler and enricher, gas-handling system, and fuel and coolant pumps. A facility for thermally cycling freeze flanges between room temperature and 1400°F was designed, and fabrication is nearing completion. This facility will also be used to test improved gas seals for the flanges. Two freeze valves, which have no moving parts, are being fabricated for test; one of these is heated by electrical resistance heaters, and the second is heated by a high-frequency induction coil. The problem of the removal of fuel samples and the addition of enriched fuel is being studied. Test equipment is being fabricated to test the feasibility of using a metal freeze seal that can be broken and remade whenever the sampler and enricher system is used. Fabrication of a one-fifth-scale plastic model of the MSRE core is nearing completion. The model is to operate with water at 40°C with a flow of 50 gpm. This simulates the MSRE fluid velocity and Reynolds number. Work is continuing with the remote-maintenance development facility. Follow- ing operation with molten salt, disassembly, and reassembly, an evaluation of the tools and procedures was made. These were found to be generally satisfactory, al- though some suggestions for improvement have resulted. Operation of forced-circulation corrosion loops continued. Nine INOR-8 loops and two Inconel loops are in operation. Operating times range from 3000 to 18,000 hr. An engineering test loop is being designed and will be constructed to evalu- ate such components as the sampler and enricher, gas-handling system, dual drain tank, freeze valves, level-indicating devices, and heaters. The system will operate isothermally at temperatures up to expected maximum MSRE temperatures. Design and testing work on the MSRE primary and secondary pumps continued. Design of a water test was completed and fabrication started. Procurement was started for some pump components. A layout of the primary pump was completed and is being reviewed. Analyses of the thermal stresses in the pump and the nuclear heating in the structural material are being made. An experiment was designed to determine the extent of back diffusion of fission gases up the pump shaft to the region occupied by the oll seal. Testing of in-salt bearings continues. The pump which incorporates this bearing development has operated for 35000 hr and has undergone 63 stop-start tests. 3., Reactor Engineering Analysis Graphite undergoes shrinkage at MSRE temperatures in a neutron flux of ener- gies greater than 0.3 Mev. Calculations were made to determine the effects of this shrinkage in the MSRE core. Both axial and transverse shrinkages were taken into consideration. v An analysis was made of the temperature effects in a graphite-moderated core with round and flat fuel channels. The hot spots resulting from the Poppendiek effect were found to be considerably reduced in the case of the flat channels. The effects of completely blocking a fuel channel were analyzed; results indicated that if one fuel channel in the region of greatest power density were completely blocked, the fuel temperature in that channel would probably rise no more than 4YOO°F above the mixed-mean temperature in adjacent open fuel passages. Reactor physics calculations were performed for the MSRE. For a cylindrical core 54 in. in diameter by 66 in. high, graphite-moderated with 8 vol % fuel salt containing 4 mole % ThF), the calculated critical loading was 0.76 mole % uranium (93.3% US32); the associated critical mass in the core was 16 kg of U232, At a reactor power of 5 Mw, the peak power density in the fuel salt wai 60 w/cc and the average was 24 w/cc. The computed peak thermal flux was 3.6 x 102 neutrons/cme- sec, and the average was 1.2 x 1010, Camma heating produced a power density of 0.1 w/cc in the core wall at midplane and 0.2 w/cc in the support grid at the bottom of the core at the reactor center line. The fast flux (above 1 Mev) in the center of the 2-in.-square graphite blocks was calculated to be about 2% less than the value at the edge in contact with the fuel channel. An analog-computer analysis was made of the loss of flow in the MSRE primary system. For this study the primary flow was decreased exponentially on periods of 1.5, 3, 6, and 10 sec. No after-heat, convection cooling, or moderator tempera- ture coefficient of reactivity was simulated in this analysis. Using a temperature coefficient of reactivity of -9 x 10'5(£k/k)/°53 the flow was decreased from 100% to zero. On the 1l0-sec period the maximum fuel tempera- ture leveled off at 1375°F after 80 sec. Using a temperature coefficient of -4.5 x 1072 (8k/k)/°F, the temperature leveled off at 1445°F after 80 sec. PART IT. MATERIALS STUDIES L, Metallurgy The last of three INOR-8 corrosion inserts was removed after 15,000 hr from an INOR-8 forced-convection loop. Weight-loss evaluations indicated the insert to have lost 1.7 mg/cm2, which corresponds to a wall thickness decrease of 0.08 mil if uniform removal of the wall i1s assumed. The compositions of thin corrosion films found on several of the long-term INOR-8 corrosion loops were investigated by means of an electron-beam microprobe analyzer. Results of analyses of the film indicated an increase in molybdenum content and virtually complete depletion of chromium and iron, compared to the composition of the base metal. Two types of solidified metal seals have been developed for use with molten fluorides at elevated temperatures one contalning an alloy sump with a tongue-and- groove Jjoint design, the other having an alloy-impregnated metal-fiber compact. Seals on components made according to both methods have been made and broken a number of times in an argon atmosphere, and subsequent helium leak tests indicate both seals to be leaktight. Several potential braze metals are being investigated for use in these seals. A method was devised and equipment was built for the leak testing of graphite- to-metal braze joints. Initial testing to 60-psi pressures was done with various braze alloys, utilizing isopropanol as the testing fluid. A technique was vi established to improve bonding, which calls for oxidizing the ends of the graphite prior to brazing. Welding and back-brazing procedures are being developed for tube-to-tube- sheet joints for the MSRE heat exchanger. A seven-tube sample has been fabricated successfully which contains trepanned areas on both the welded and brazed sides and includes a braze-metal sump with feeder holes in the tube sheet. This mini - mzes the effect of different heating rates for thick and thin metal sections. A1l the mechanical-properties data for INOR-8 were reviewed in order to es- tablish design values for the alloy. Design stresses for temperatures below 1050°F were selected on the basis of two-thirds of the 0.2% offset yield strength. Above 1050°F, the design stresses were based on the stress to produce 1% creep strain in lO5 hr. A table with the selected design strengths for temperatures up to 1L400°F is included. Tensile tests were performed to determine the effect of low creep strains on the strength and ductility of INOR-8. No effects were observed that could in- fluence the structural integrity of reactor components made from INOR-8. Tests on specimens from selected locations and orientations in large pileces of R-0025 and MHA4Im-82 graphite indicated relative uniformity within each grade of (1) apparent densities and (2) permeation by molten fluorides at 150 psig in 100-hr exposures at 1300°F. Permeation of S-h and AGOT graphite with LiF-BeFp-ThF)-UFy (67-18.5-14-0.5 mole %) at 1300°F at pressures of 25, 65, and 150 psig in 100-hr exposures indi- cated that (1) there were small differences in salt permeation of these grades for the different pressures used and (2) actual and theoretical salt permeations were practically the same except for the 25-psig permeation of grade S-k. Grades S-i and AGOT, respectively, are moderately low- and high-permeability grades of graphite. A single series of five precipitation tests was made with AGOT graphite and molten LiF-BeFp-UF) (62-37-1 mole %); only the volume of the graphite was varied in order to determine the relationship of graphite volume to uranium precipitation. For volume ratios of graphite to fuel of 27:1 to 5:1, the uranium precipitated per cubic centimeter of the bulk volume of the graphite remained approximately con- stant and averaged (1.5 + 0.4) mg to (1.3 - 0.3) mg. Additional tests were made in order to confirm data indicating that the thermal decomposition of NH}F+«HF removes oxygen contamination from graphite to such an extent thet it could contain molten LiF-BeFo-UF) (62-37-1 mole %) at 1300°F without causing the usual UO, precipitation from the fuel. No carburization was detected on unstressed INOR-8 specimens after exposure to LiF-BeF,-UF), (62-37-1 mole %) - graphite system for 12,000 hr at 1300°F. 5. Chemistry A fuel composed of LiF-BeFo-ThF)-UF) (65-30-4-1 mole %; m.p. 450°C) has been selected as representative for the MSRE design work, but there is a strong possi- bility that 5 mole % of ZrFy will be included as an oxide scavenger in a revised composition. The change is pending confirmation of current experiments which in- dicate that ZrOp is more insoluble than UOp, and thus provides protection against the precipitation of U0y as a result of accidental contaminations wilth oxide. vii The coolant composition, LiF-BeFp (66-34 mole %; m.p. 465°C), affords a low viscosity in combination with a suitable melting point. Treatment with hydrogen as a means of removing oxide from graphite proved relatively ineffective. The rate of permeation of graphite immersed in a wetting salt is slow, at least in the later stages, presumably because gas trapped in the graphite voids can be replaced only as fast as it leaves by diffusion through the salt. 6. Fngineering Research The surface tensions of two NaF-BeF, (57-43 mole %) mixtures have been de- termined to fall between 200 and 150 dynes/cm over the temperature range 500 to 300°C. A 6% discrepancy between the two measurements may relate to differences in contaminant content of the two samples due to differences in exposure time in the circulating loop from which the samples were drawn. The results are 1n rea- sonable agreement (although somewhat lower) with data obtained with an NaF-BeFy (63-37 mole %) mixture. An analysis of the precision of the measurements indi- cates that residual effects due to errors in pressure and geometrical measurements have been reduced to about *3%; however, a large uncertainty (as much as an addi- tional *3%) still remains in the salt density as used in evaluating the data. Corrections to the original data and the inclusion of more recent results have yielded a revised correlation of the mean heat capacity of BeFE-containing salt mixtures. Heat-transfer studies with LiF-BeF,-UF)-ThF) (67-18.5-0.5-14 mole %) in Inconel and INOR-8 tubes have been interrupted after 5560 hr of operation to re- place the circulating pump. Damage appears to be restricted to the upper shaft- bearing. Analysis of the salt (pre- and post-operational) shows a composition differing from the nominal composition; this will necessitate a re-evaluation of the data using corrected values of the thermal properties. 7. Fuel Processing Preliminary studies indicated that ThFy in molten-salt reactor fuel may be decontaminated from rare-earth fission products by dissolution of the rare-earth fluorides in SbF5-HF. The LiF of the fuel must be removed first, by dissolution in OF, to prevent precipitation of the antimony, probably as LiSbFg. P v R D T T e A e b i o CONTENTS SL]WARY. ---------- P R T NI I I I I I R I T I R R B N I I I A [ A BB B Y R RS N N B R R R I N N N L iii PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE DESIGN ......... Cetsaseen fhreesensaas Chereeaaenas Creerassssenennn 1 1.1 Introduction seseevecsvassesesssansanns cesessscscsesenstesesenans 1 1.2 MSRE Objectives s.eeierecsascsesssncssanns cesesisesesesansesrenns 2 1.3 Conceptual Designs seeeseesasssascssascaancs cesateassearnsseanns 2 1.4 The Reactor Core and Vessel «e.eeeveeeses Ceeedracesenstersrnenns L 1.5 The Primary Heat Exchanger ......ceee... Ceteisasescasarasesnsanas 8 1.6 Radiator ......... Cietisensesanaesssanssasararecaaans B 1 1.7 Drain Tanks «eec... csesesessernstssas st nesetonerane cetesesesess 11 1.8 Fquipment Arrangement +.ceeeececesencacns S 24 1.9 Design of Cover-Gas SyStem .eeeeeseiecesccacan cerieccerscesensas 18 1.10 Design of the Remote-Maintenance System for MSRE «vieevsrsscsess 19 1.11 Modifications to Building 7503 ...ceceecennn =24 2. COMPONENT DEVELOPMENT ..vvvvevanernannas cieeens Cereesieieiaeees cevess 24 2.1 Freeze Flange Development su.eeseveeeacnns = 2.2 Ireeze Valves .cviveescerercscssnssasansns =) 2.3 Sampler-Enricher Development ...... 25 2.4 MSRE Core Development +..eeces.. - 2.5 Remote Maintenance Development Facility ..cceceeeen.. crscaesasnss 27 2.6 TForced-Circulation Corrosion LOODPS secsececcceccccasnens cevessess 2B 2.7 Pump Development ..ieeiviescersescrsnsscressnscescscnsons ciessese 29 2.7.1 MSRE Primary Pump ..ccesrecnccanetscansnnes cevssenssases 29 2.7.2 MSRE Secondary PUmp eecvsrerennarassanencns Ceteiesesanaas 30 2.7.3 Advanced Molten-5Salt Pumps «sevevecenn. cesseas teseneesrsss 30 2.7.4 MF-F Pump Performance LOOD eeeevevececnaanss cesesassenes 31 2.7.5 Frozen-Lead Pump Seal sieseveectciccccenannn ceveesesrees 31 2.7.6 MSRE Engineering Test LOOD svecsccacseecosenns cessessens 31 3. REACTOR ENGINEERING ANALYSIS ....... Gt eseerrseanaata s G - 3.1l Effects of Graphite Shrinkage Under Radlatlon in the MSRE Core . 32 3.2 Temperature-Rise Effects in MSRE Cores with Round and Flat Fuel Channels cevecsoeseccescononsseanss G L 3.3 Temperature of Fuel in a Blocked Passage in the MSRE ....cc00nee 39 3.4 MSRE Reactor Physics ..e.veeevonn. O 'S | 3.4.1 Core CalcUlations seussecsssuesesosessasvsonesnssssennsns 41 3.4.,2 Gamma-Heating Calculations s.cevecenrconarsssnaacescnsans i 3.4,3 Drain Tank Criticality ....... Ceeeretecenaneasanannns ee. Lh 3. 4,4 Pump-Bowl Fission Product Activities ...eeveneen. . 3,4.5 Cell Calculations sevesesseesssosssnsansaasnnass S [ 3.5 Analog Computer Study of MSRE Primary-System Flow LoSS sveaeesse 45 3.5.1 Description of the System Simulated .....cviveeecsasesss k6 3.,5.2 Analog Computer Program .eeesescsesassaoes B ¥ 3.,5.3 Simulator Operation «ersessesesescescancocsanasoannsns R [T 3.5.4 Conditions Used to Obtain CUIrveS cvevevvreanecenans ceanas L9 x PART II. MATERIALS STUDIES b, METALLURGY +eetevovceaatoooonostsonssssasnsonaancsssosnsssosaseonssnsnsas . 55 4,1 Dynemic-Corrosion Studies s.eeeieveesacasnn ettt teteeee e 55 4.1.1 Forced-Convection Loops «.ee... C e ieeaseresaieeaean 55 4.1.2 Microprobe Analyses of Surface Fllm .................... 56 4.2 Welding and Brazing StUdi€s +veeirrieeieerenensnennreeasconanonns 58 4,2.1 Solidified-Metal-Seal Development Ceteiater ey 58 4.,2.2 Brazing of Graphite ...... Ceeerseeseterarteceatatecennans 59 4,2.3 Heat Exchanger Fabrication ........ Ceenarsasecenaaeaneae 63 4.p.4 Mechanical Properties of INOR-8 .......... et teresaenean &l 4.3 Permeation and Apparent-Density Uniformity in Large Pieces of Graphite iveeeisrentrennnas e e e s e n s eenneene s 67 4,35.1 Permeation of AGOT and S-4 Graphites by Molten Salts at Different Pressures ....... Ceie i chaeasen 67 4.4 Precipitation from Molten Fluoride Fuel in Contact with Various Volumes of Graphite .......... creerneesraasas e 69 k.4.1 Removal of Contamination from Graphite ......ccievveeean. 69 L.k,2 INOR-8 - Fuel - Graphite Carburization Tests ........... 70 4.5 Tn-Pile TeStS seeeereerenenanennans f et et ee ettt T1 5. CHEMISTRY vieevieerviestsatsosensssronsanassas teeensaranans Pesecescsnna T2 5.1 Phase Equilibrium Studies +evvoeereeonsesss B - 5.1.1 MSRE Fuel and Coolant seievsnenrntnnencscnna s et et es e T2 5.1.2 Systems Containing ThFy .......c.0v.n. Cereesrerseans veens T2 5.1.3 The System ZrFU-ThF) «.evtevreteretvestessnsssssrsavesnsnse Th 5.2 Effect of Tetravalent Fluorides on the Freezing Point of Scdium Fluoride ............ Sececesiesitaeeretteennaens N ) 5.2.1 Phase Diagram of Fluoride Systems .........0... Ceeeeases 7T 5.2.2 The System LiF-YFz .....v0vnnnns Ches et iesaressensanen s 78 5.2.3 Melting Point of NiFo et iiinitieereveenensnns 79 5.3 Oxide Behavior e.svevesee cheeeaes Pesesesencresontaonsacnn cheasas 79 5.3.1 Oxide Behavior in Fuels .cieieivntenneerenens Cerreecnens 79 5.3.2 Zirconium Oxyfluoride and Attempted Preparation of Urancus Oxyfluoride .......cceinieeeicnersannrsncaas 80 5.4 Graphite Compatibility cveieieieieriereennaeaeasasannsaansnssanns 80 5.4.1 Removal of Oxide from Graphite by Treatment wWith Hydrogen cvieveerinestiecssrienssecssssansassssnase 80 5.4.2 Behavior of Graphite when Wetted by a Molten Fluoride .. 81 5.5 Preparation of Purified Materials ............. e o1 1 6. ENGINEERING RESEARCH ...ivuivrieienenenenereenenenensnancnansenonnennnnns 83 6.1 Physical-Property Measurements ....c.eeieeeieneeeeenncnarnncnans 83 6.1.1 Surface Tension and Density «eveeeeeeeenercrenenennansas 83 6.1.2 Heat Capacity ce.eeeve.. e Ceeeenaaas Cetetrecireaae 85 6.2 Heat-Transfer STUGIES .eieeeiecererrasoroaoasnsesncsaesonaseaans 86 7! F[JEIIPROCESSING LR R R B N B I B D O R N N B R R Y R AN A L R R DR R N N B R R R B R DL R I I I I I B B A 88 PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE DESIGN 1.1 INTRODUCTION The concept of using molten fluorides, one of which is UFy with highly enriched uranium, as a liquid fuel for reactors has been actively studied at ORNL for about ten years. Many design studies have been made, and one reactor was constructed and operated successfully as a high-temperature low-power ( X2.5 Mw) reactor experiment, the ARE. Much experimental work has been done on phase diagrams involving molten fluo- ride fuel systems. The work included some in-pile as well as many out-of-pile experiments in which test loops of circulating fluoride melts were operated for thousands of hours with substantial imposed temperature differences in the loops. This work, most of it under the ANP program, has developed fluoride-fuel technology to a point where it is ready for serious consideration for high-temperature power reactors. Container material has also been the concern of an extensive metallurgical program for about ten years. Stainless steels, Inconel, Hastelloy, and other alloys have been studied for corrosion resistance to the molten fluorides; the experimental reactor, which was run successfully, employed Inconel. While the corrosion resistance of Inconel was adequate for a short-term experiment, it was considered to be unsatisfactory for long-term service. INOR-8, an alloy, was developed specifically for compatibility with the fluo- rides. This alloy was made into components for pumped test loops, and thousands of hours of corrosion tests were performed. Here also some in-pile as well as many out-of-pile tests were run, and it was concluded that INOR-8 was a satisfactory material for the construction of the components of a high-temperature, molten-salt- fueled reactor. Having established a satisfactory fuel and satisfactory container material, the next question concerned the kind of moderator which would be acceptable for such a system. Theoretical calculations, for which no experiments were performed, indicated that the fluorine in the salt could be employed, thus making possible a homogeneous molten-fluoride reactor system. Many design studies were made of this concept. Tt was obvious that, while a homogeneous system was possible, the neutron economy of this type reactor was such as to preclude any serious attempt at thermal breeding. Therefore some studies, both theoretical and experimental, were begun in order to determine the possibility of employing unclad graphite as a moderator for the fluoride-fuel reactor. An estimate of the breeding potential of the graphite- moderated molten-salt reactor can be found in an ORNL report.l 2 Compatibility tests between graphite and the fluoride fuels were run, and no chemical incompatibility between graphite and the fluorides was found. While some fuel penetrated the graphite sample, it was also determined that(neglecting the effects of radiation)not more than 1 vol % of the graphite would be permeated if especially low permeability graphite were used. In view of the favorable results of the long development program on materials, the design and construction of a molten-salt reactor experiment has been initiated. 1.2 MSRE OBJECTIVES The primary objectives of the MSRE are to determine whether a molten-fluoride, circulating-fuel reactor system can be made to operate safely, dependably, and ser- viceably. By this it is meant that no credible accident can endanger personnel, that the system is capable of achieving a very high percentage of operational time, and that any component can be removed from the system and a replacement made, per- mitting return to normal operation. The reactor will, however, provide answers to many questions of importance to future molten-salt reactors. It will provide long-term irradiation tests of fuel, INOR-8, and graphite under actual service conditions. From the behavior of graphite with respect to the absorption of fuel and fission products, important questions with regard to the feasibility of breeding in molten-salt reactors can be answered. 1.5 CONCEPTUAL DESIGNS The reactor consists of a cylindrical vessel in which a graphite matrix con- stitutes about 88% of the volume. Fuel enters the vessel at an annular volute around the top of the cylinder and passes down between the graphite and the vessel wall. A dished head at the bottom reverses the flow and directs it up through rectangular passages in the graphite matrix into a dished head at the top, from which it goes to the suction line of a sump-type pump mounted directly above and concentric with the reactor vessel. Flow through the reactor is laminar, and the passage width is narrow enough to prevent an excessive Poppendiek effect. The reatangular passages were chosen because they gave, with the design flow, a much lower hot-spot temperature than did cylindrical fuel passages. From the pump dis- charge, the fuel flows through the shell side of a cross-baffled tube-and-shell heat exchanger and thence to the reactor inlet. The tube side of the primary heat exchanger contains the secondary coolant, which is also a binary fluoride melt (LiF-BeFp, 66-34 mole %). This fluid is cir- culated by a sump-type pump through the tubes of the heat exchanger and through the tubes of an air-cooled radiator. Air is blown across the unfinned tubes of this radiator and up a T70-ft stack by two axial blowers. A basic flow diagram of the primary and secondary salt systems is shown in Fig. 1.1. Control of this reactor is quite simple and does not call for internal control rods of any kind. Fast control is effected by the volumetric temperature coeffi- cient of the fuel, which produces a temperature reactivity coefficient of approxi- mately -4 x 10-5 (ak/k)/OF. Slow control is accomplished by fuel enrichment, with the provision of poison addition (e.g., ThF)) if desired. Fuel addition in gross amounts for the original loading will take place in the drain tank. Subsequent addition for burnup and fission-product poisoning will UNCLASSIFIED ORNL-LR-DWG 50409 COOLANT PUMP FUEL 850 GPM PRIMARY SALT PUMP HEAT EXCHANGER 110Q °F o - > LiF-61% — I L BeF-34% % 10251°F ) ' ThF- 4% ! i ‘ UR- 1% ” SECONDARY SALT X LiF-66% | i BeF-34% 1225 °F ¥ 1450 GPM D ] ——— L : m : :: 5 T wzs=F ¥ i » 1 i 3 i REACTOR ": LJ ; CoRe 1 VESSEL ;5 l E:‘ AIR e g b 160,000 RADIATOR ! FJ/ ! E SCFM —» 300 °F P {00 °F L REACTOR CELL noot Iy [ e - FREEZE PLUG SPARE FILL & DRAIN FLUSH TANK COOLANT FILL & DRAIN TANK (60 cuft) DRAIN TANK TANK (60 cuft) (30 cu ft) (60 cuft) Fig. 1.1. Basic Flow Diagram of Primary and Secaondary Salt Systems. be made through an enricher assembly communicating with the gas space in the pump bowl. This enricher assembly will also be used in essentially the reverse manner for taking fuel samples at any time during operation of the systen. The fuel system is to be an all-welded system except that each component will employ a freeze-flange connection to the piping system. The rotary element of the pump will also employ a conventional metal-gasketed flange with provision for inert- gas buffering to ensure inleakage of gas in case of imperfect closure. The secondary coolant system will be all-welded and will not employ the freeze flanges. This is made possible because direct maintenance can be effected on this circuit at all times. The molten salt in both the primary and secondary circuits will be sealed off from the respective drain tanks by means of freeze plugs in the drain lines. Pro- bosed designs of the freeze flanges and the freeze plugs are now under construction, and tests of these units will be made to determine the ultimate design. Heating of the primary and secondary salt systems will be accomplished by means of electrical heaters around all lines and components of these systems; several types of commercial heaters have been studied. It now appears that the heaters will be locally constructed, and Tnconel or stainless steel pipe will be used. Designs of these units are being made, and tests will be run to verify the advisability of using this kind of heater. The off-gas system has been planned as a recirculating system, to minimize the consumption of inert cover gas. Holdup capacity for radioactive decay and filters for cleanup of the gas will be incorporated. Provision is made for continuous ex- haust through charcoal beds to the stack, in case the recirculation system fails at any time. The fuel dump tank must be provided with heat for keeping the clean fuel molten and must have cooling available to remove afterheat from radioactive fuel. The tank is provided with electric heaters similar to those on other parts of the salt cir- cuits. Cooling is provided by means of thimbles penetrating the tank through the top; the thimbles have coolant tubes inserted in them which are spaced away from the thimble wall. The coolant tubes will be filled with water which will be boiled by the radiant transfer of heat from the thimble walls. 1.4 THE REACTOR CORE AND VESSEL The physical structure of the reactor consists of a containment vessel; an inner, open-ended TINOR-8 cylinder serving both as a separating baffle for the cool- ing annulus and support for the graphite core; the composite graphite-moderator matrix with positioning and support members; and various flow regulating devices. Figure 1.2 shows the concept of the reactor. Significant geometrical and flow characteristics are listed in Table 1l.l. A heat-generation rate of 0.2 w/ce in the INOR-8 wall of the reactor vessel requires a heat removal of 23 kw. The wall cooling will be accomplished by the fuel flowing along the wall. Turbulent flow is desirable in the annulus in order to mini- mize the Poppendiek effect. With the design flow rate of 1450 gpm in the 1l-in.-wide cooling annulus, the Reynolds modulus is 13,500, and the temperature of the outside wall surface is less than 5°F above the bulk stream temperature. The moderator graphite of the core is built up from 2 x 2 x 66 in. stringers. The size (area) of these matrix stringers is limited by the size of impervious graphite available at present. The stringers are pinned in beams at the bottom of the core. A graphite band will hold the matrix together, and an INOR-8 yoke is provided for centering. A coarse screen prevents possible graphite fragments from leaving the core. If the graphite is packed tightly in the inner can at room temperature, the differential expansion between the TNOR-8 and graphite will open a radial clear- ance of 3/16 in. at operating temperature. It is expected that the fast-flux dis- tribution in the core may tend to cause radial bowing in the graphite. In order to reduce the nuclear effects of bowing and shrinkage, the graphite stringers will be banded over the middle two quarters with molybdenum bands. Flow passages in the matrix are provided through rectangular channels machined into the faces of the graphite stringers (see Fig. 1.%). The tabulated channel dimensions provide a fuel volume fraction of 12%. The specified channel configura- tion is the product of intensive studies into the temperature effects asgociated with the relatively slow flow through the core (about 2 fps). These effects are UNCLASSIFIED ORNL-LR-DWG 52034 FUEL OUTLET GRAPHITE SAMPLE BLOCK S T <] 5 o QO >z e MS o < o U] FUEL INLET () (X B0 % () ) ) 5 g o 0 e ¥ o o % ave R BERGEX ) J0e ) OO VAN AV A O S \ FUEL INLET VOLUTE ~~— REACTOR VESSEL ; GRAPHITE-MODERATOR — STRINGER B % 0 Y AV () " (] % X 4 o A CCRE YOKE AND SCREEN REACTOR CORE GAN FUEL PASSAGE CORE-POSITIONING GRAPHITE BEAMS VESSEL DRAIN LINE SWIRL VANES ANTI CORE GRID SUPPORT MSRE Reactor. Fig. 1.2. 6 Table 1.1. MSRE Reactor-Vessel Design Data Inlet pipe 6 in., sched 40 Outlet pipe 8 in., sched 40 Core vessel oD 58-3/8 in. ID 57-1/4 in. Wwall thickness 9/16 in. Design pressure 50 psi Design temperature 1300°F Fuel inlet temperature 1175°F Fuel outlet temperature 1225°F Inlet Volute Annulus ID 54-1/2 in. Annulus OD 56-1/2 in. Over-all height of core tank 8 ft Head thickness 1l in. Graphite core Diameter 54 in. Core blocks (rough cut) > x 2 x 67 in. Number of fuel channels 1064 Fuel-channel size 1.2 x 0.200 x 63 in. Effective reactor length ~ 65 in. Fractional fuel volume 0.120 Core container ID 54-1/8 in. 0D 54-5/8 in. Wall 1/k in. Length 73-1/b4 in. a composite of the Poppendiek gradient and a less significant temperature gradient in the graphite produced by the internal heat generation. With reasonable flow rates in a once-through multichanneled core, the flows are laminar or at best un- stably turbulent. Ior laminar-flow conditions, the Poppendiek effects become rather severe with wide channels. The radial temperature gradient of a circular channel providing 10% fuel volume in a 2- X o>-in. graphite section is 357 times as large as that of a 0.1-in.-wide channel with equal fuel fraction. With the selection of the present 0.200-in.-wide channels, the graphite temperature at the midpoint will be AO°F above the mixed mean temperature at the nuclear center of the core. This problem is treated more fully in Chap. 3. UNCLASSIFIED QORNL—LR-DWG 52035 PLAN VIEW TYPICAL MODERATOR STRINGERS SAMPLE PIECE X CROSS-COMMUNI - CATING CHANNELS 83? NOTE: NOT TO SCALE ‘\\\\!'A LI Fig. 1.3. Typical Core-Block Arrangement. It is expected that complete blocking of one fuel channel would raise the temperature of the fuel in the channel about 6800F above the level of adjacent channels. Provision will be made for the removal of five graphite samples from the center of the core. These full-length samples will be lifted out through the suction line and the pump bowl. 1.5 THE PRIMARY HEAT EXCHANGER The primary heat exchanger is being designed (Fig. 1.4) for a duty of 10 Mwto be transferred from the fuel salt to the secondary coolant salt. Pertinent data of the preferred design are given in Table 1.2, The design of the heat exchanger follows the configuration of conventional 25%-cut, baffled shell-and-tube units, with greater emphasis on reliability and UNCLASSIFIED ORNL-LR-DWG 52036 FUEL INLET U-TUBE BUNDLE Y2-in ~0D HEAT EXCHANGER TUBE CROSS BAFFLES {25 % CUT) THERMAL-BARRIER PLATE COOLANT INLET L) - COOLANT-STREAM SEPARATING BAFFLE COOLANT OUTLET FUEL OUTLET Fig. 1.4. Primary Heat Exchanger for MSRE. 9 Table 1.2. Primary-Heat-Exchanger Design Data Heat load 10 Mw Shell-side fluid Fuel salt Tube-side fluid Coolant salt Layout 25% cut, cross-baffled shell and U-tubesgs Baffle pitch 10 in. Tube pitch 0.812 in., triangular Active heat-transfer length of shell 5 ft 10 in. Over-all length ~ 7 ft Nozzles Shell side 6 in. IPS Tube side 5 in. IPS Shell diameter 16-1/4 in. ID Shell thickness 1/5 in. Number of U-tubes 156 Tube-sheet thickness 1-1/2 in. Heat-transfer surface area 250 ft2 Fuel holdup ~ 5.5 ft3 Terminal temperatures at design point Fuel salt Inlet 1225°F; outlet 1075°F Coolant salt Inlet 1025°F; outlet 1100°F Effective LMDT 135%°F simplicity of construction than on particularly high performance. The space limita- tlons of the containment area call for a fairly short unit. The heat-transfer and pressure-drop design are based partly on experimental heat-transfer data of Amos, MacPherson, and Sennc (for the tube side) and partly on methods suggested by Kern.? From the heat-transfer point of view it is prefer- able to pass the larger flow of fuel salt through the shell side, and the smaller flow of the coolant through the tubes. The shell side presents less opportunity for retention of gas pockets during filling operations than does the tube side. The fuel salt operates at lower pressure; thus thinner shell walls may be used. The shell side, however, has slightly more liquid holdup. The U-tube configuration results in a much shorter over-all length. The tem- perature effectiveness of the unit is 97.5%, compared with true counterflow. The 1809 bend in the tubes minimizes the thermal expansion problem. The tube and baf- fle pitches were chosen to give an even number of baffles in the shell side within the range of baffle pitches of 0.2 to 1.0 shell diameter, where the methods of Kern have good accuracy. The stresses in the heat exchanger have been analyzed. Because of the low- pressure operation, mechanical stresses are significant only in the tube sheet. 10 With a 50-psi pressure differential on a 16-in.-dia flat plate, 1.5 in. thick, the estimated stress is 3000 psi. In order to minimize thermal stressing of the plate, which would be additive to the mechanical stresses, a thermal baffle is placed about 2 in. from the tube sheet. This baffle will provide a stagnant layer of salt, reducing the thermal gradient across the tube sheet to ~20°F. The stif- fening effects of the stream-separating baffle in the tube-side header would induce high localized stresses if attached to the tube sheet. For this reason, a laby- rinth seal is provided at the tube sheet, and the baffle is welded to the dished head. While mechanical stressing of the tubes is insignificant (only ~228 psi hoop stress), stresses due to thermal gradients across the wall may reach 6500 psi at the fuel inlet end. However, the combined stresses remain well below the T700-psi level allowed for a metal temperature of 12250F, which is the maximum design tem- perature for the fuel. A tube-to-tube-sheet joint of the welded and back-brazed construction is being tested to determine the tube-sheet thickness required in the present exchanger. The first specimens appear to have produced sound joints. Two other heat exchanger configurations were also investigated. An axial- shell-flow exchanger having closely spaced tubes was rejected on the basis of expected manufacturing difficulties. A second loosely spaced heat exchanger with similar flow pattern would have resulted in a much longer unit, not suited to the limited space in the reactor pit. 1.6 RADIATOR In the salt-to-air radiator the thermal energy of the reactor is rejected to the atmosphere. The radiator will occupy a part of the existing air duct in Build- ing 7503, and its over-all dimensions are SO tailored that it can be installed in the available space. The physical design concept is shown in Fig. 1.5, and the design data are listed in Table 1.3. The heat transfer design is based on Curves by Kays and London)“L for the air gide and on data previously reported.2 The performance of the heat exchanger under reduced load conditions is presently under study to determine the most effec~ tive method of power control. Several features were incorporated in the design as protection against freez- ing of salt in the radiator tubes: 1. Large-diameter tubes were used. 5. The heat rate per unit area was kept low by using bare tubes on the air side to take most of the temperature drop in the air film. %, The lowest design temperature of the bulk salt is 1859F above the freez- ing point. 4, An elaborate headering system is proposed to ensure even flow distribu- tion between tubes. 5, Studies are under way to provide adequate heating of the radiator should one or more tubes freeze accidentally. 11 UNCLASSIFIED ORNL-LR-DWG 52037 5-in. SECONDARY-SALT INLET \k\\“%h4n-ODX()OTZ#anALL TUBING B-in. HEADER 10 ROWS, 12 TUBES PER ROW 5-in. SECONDARY - SALT OUTLET 2Y-in. TUBE MANIFOLD AIR FLOW Fig. 1.5, 8Salt-to-Air Radiator. The layout of the tube matrix is expected to allow movement of the tubes under thermal expansion with a minimum of restraint. The sloping tube configura- tion will promote drainability. 1.7 DRAIN TANKS Four salt-carrying drain tanks are under design for the MSRE, as listed below. 1. a duplicate pair of fuel drain tanks, 2. one container for the flush salt, 5. one container for the secondary salt. Items 2 and 3 are simple pressure vessels made of INOR-8, each supplied with a dip-tube fill-and-drain line and gas connections for Pressure filling. Their dimensions are listed in Table 1.L. 12 Radiator Design Data Table 1.53. Duty Temperature differentials Salt Air Air flow Salt flow Effective mean At Over-all coefficient of heat transfer Heat~transfer surface area Tube diameter Wall thickness Tube matrix Tube spacing Subheaders per row Main headers Air-side AP Salt-side AP 10 Mw 1025 to 1100°F 100 to 300°F 167,000 cfm at 15 in. 32 830 gpm at avg temperature 920°TF 53 Btu/hr-ft2-°F 685 £t° 0.750 in. 0.072 in. O pressure 12 tubes per row; 10 rows deep 1-1/2 in., triangular % in., IPS 8 in., IPS 11.6 in. Hy0 6.5 psi The after-heat in the fuel salt requires cooling in the drain tanks if the fuel is dumped a short time after shutdown. Considering time lags in the drain- ing procedure, it is unlikely that the tanks would receive the fuel less than 15 min after shutdown. Accordingly, the heat removal rate of 100 kw would main- tain bulk-salt temperatures below 1350°F during the decay period. To make the heat removal as nearly uniform as possible throughout the tank, 40 immersed bayonet coolers are used, with boiling water as coolant. Water cooling was selected above gas, molten salt, or NaK because of its simplicity and relative independence from utilities failures. To further enhance the reliability of the fuel drain system, two identical tanks will be provided: one in use, the other in standby. Double-wall separation is used between the salt and the water. (It appears that the induced thermal stresses are not excessive, although further in- vestigations will be made.) The thimble design is shown in Fig. 1.6, and the fuel-drain-tank system is shown in Fig. 1.7. The fuel drain tank, filled with salt but without cooling thimbles, would have a multiplication constant of about O.44. Even if the cell is flooded with water (e.g., as an emergency cooling method), the multiplication constant would reach only O.77. 1.8 EQUIPMENT ARRANGEMENT The equipment arrangement for the MSRE has been developed on the basis of the following criteria: 1. All equipment that contains fuel is to be replaceable by remote mainten- ance. Table 1.k4. 13 Drain~Tank Design Data l. 2. 3, Fuel Drain Tank Height Diameter Wall thickness Vessel Dished head Capacity Fuel Gas blanket Max. operating temp. Cooling method Cooling rate Coolant thimbles Number Diameter Secondary Drain Tank Height Diameter Liquid capacity Wall thickness Vessel Dished head Cooling method Flush Salt Tank Height Diameter Capacity Coolant Gas blanket Wall thickness Vessel Dished head Cooling method 35 in. (without coolant headers) 48 in. 1/2 in. 3/4 in. 55 £t b7 £t 1350°F Boiling water in double-walled thimbles 100 kw 14 UNCLASSIFIED ORNL—-LR-DWG 52038 COOLING-WATER INLET\ /STEAM QUTLET ot = JACKET BREATHER HOLES A TANK HEAD T AT TTT N y ¥4 — QUTER CONTAINMENT TUBE STEAM RETURN TUBE — i 7SPACER FINS WATER FEED TUBE — {1\ 7 4 Fig. 1.6. Cooling Thimble for Primary-Salt Drain Tanks. UNCLASSIFIED ORNL-LR-DWG 52039 STEAM CUTLET CONNECTION STEAM DRUM SAMPLING CONNECTICON THROUGH DRUM 1 TYPICAL FOR FLEXIBLE TUBING; QUTER RING FLEXIBLE TUBING; TYP- [CAL FOR INSIDE RING L [} ZT -0 5 €T 5 @ E | <1 | od ic w O N =z Sm < v - 2 Zs & = O \ B //////////./////,////////I/ ) | S D =2 = o — WW Wi o1 = S0 i > =z < c EU @ @ 2N % i A T — Ya ao = > i TYPICAL BAYONET BRACING, FOR INNER AND OUTER RING Primary Drain and Fill Tank. Fig. 1.7. 16 2. Equipment that contains coolant salt is to be replaceable by direct maintenance. 3, The primary fuel loop is to be located in the existing 24_ft-dia contain- ment vessel in Building 7503. 4, The fuel storage tanks are to be located in separate containment. 5. Only vertical movement of the fuel-circulating pump should be permitted. Several studies of possible layouts have been made, and the following conclu- sions have been reached: From the standpoint of piping stresses, the most favor- able arrangement is to mount the pump on top of the reactor, with a minimum length of piping from the reactor to the pump suction. A minimum of 3 ft is required by the hydrodynamics of the pump. It is also desirable to anchor the coolant piping close to the pump so that expansion of the parts of the cooclant system outside the primary-loop containment vessel does not affect the stresses inside the vessel and so that the differential thermal-expansion stresses imposed on the fuel system by the coolant piping are minimized. The arrangement shown in Fig. 1.8 and 1.9 has been developed on the basis of the above criteria. The reactor is placed in the southwest quadrant of the con- tainment vessel, with the fuel pump mounted on top. The heat exchanger is aligned with the pump discharge. The reactor is supported near the top, and the pump is UNCLASSIFIED ORNL-LR-DWG S0410A TOP OF PENTHOUSE Ibaliaibibinlttii N CONCRETE SHIELDING, ; At e STEEL BARRIER, \ MOTOR WEIGH = | SCALES = PRIMARY PUMP. 4 A . 8-in. SUCTION ! . HEAT EXCHANGER i R . S, R R CRBERGA Sk T — ‘-i b , y - : i &1+ Qin. ; . B SHIELDING N —_— | ¥ - | | I D REACTOR 7 AIR PLENUM WALLS . t¥2-in. DRAIN LINE ************ I 07774 .1 1%-in. DRAIN AND T ; : FILL LINE. ! RADIATOR ‘ / i : RS W — PIT P — e S o - EXISTING STEEL{~, ] WATER \ | SHIELD R > : : e LOW POINT FILL AND DRAIN | i OF TANK - ¢ TEST CELL TANK NO. 2 _—FILL AND DRAIN TANK NO. 3 FLUSH i TANK i R T Y - i ¥ Fig. 1.8. General Arrangement of Primary, Secondary, and Drain Systens (Elevation). L ] [} ] L] UNCLASSIFIED ORNL—LR—DWG 52040 — NUCLEAR WATER SHIELDING s v /// / m i pi it 1 li ’t— B '| // AND PRIMARY PUMP | /7 | | |l I “ I /// /PIPE SLEEVES I | ! i L FILL AND DRAIN TANK NUMBER 3 I / 1 I L__ _/ — { 1\ L 4H|EAT EXCHANGER = I { | HOT STORAGE PITS } | l: _— | | N - ! FILL AND YIRS ] | i DRAIN TANK /\g/ NEUTRON SOURCE TUBE RAg:fl OR H I | NUMBER 2 —1 >\ \\\ \ p il 4 7 i | ll X N et /', / N SR A T “f > | | ? NN RS L s k ! I ' FILL AND DRAIN TANK NUMBER ' STACK AND AIR DUCT Il J | W PLENUM | FLUSH TANK— NG S s T ¥ J—L CABLE AND PIPE TRAY ! TR | T H / i AR DUCT [ W / ! ii PLENUM—/’%’: ! ANl 1 I ! o il il ; a - b i S I 1 I h i I Iy / ; U L I / 7 { / / / EXISTING FANS —— = = - ES BELLM . ’ i G % BLOWER HOUSE @‘ - MAX e 1 Mev. Midplane. flux for the exposures that have been tested, and differs between graphites and between axes within a piece. The more permeable graphites appear to have lower shrinkage rates, by as much as a factor of 8, than the less permeable types needed for use with molten salts. For CEY graphite, typical shrinkages for a neutron dose of 10 0.004 parallel and 0.002 normal to of graphite expected to be used in the MGSRE, nvt of energies of 1 Mev and greater were the extrusion axis. The plots of the neutron flux of energies greater than 1 Mev in the MSRE are shown in Figs. 3.1 and 3.2. The value at the center is 1.52 x 10 year at full power of 10 MwT. tends to produce bowing, is 0.0735 x 10t midplane. 32 nv, or a dose of 4.8 x 10 The maximum flux differentisl, radially, nvt for a full which nv per inch, at a 20-in. radius on the UNCLASSIFIED ORNL-LR-DWG 52046 33 The shrinkage may produce four effects, which are considered separately. These are: 1. tension at the surface, caused by the fast-flux depression in the center of a graphite piece; 2. gross axial shrinkage of pieces, tending to shorten the reactor; 5. ‘transverse shrinkage, which tends to increase the fuel volume fraction within the reactor; 4. bowing of the pieces, caused by flux gradient. Calculations have been made on the assumption that there is no yielding or annealing, although there is evidence that shrinkage-induced stresses may be re- lieved. Results of the calculations are shown primarily as strain and second- arily as stress, which is the product of strain and Young's modulus. The strain to produce rupture has been shown to be about 0.001 in./in. for graphite made from coarse particles, and increasing to about 0.003% for types made from fine particles. It is estimated to be greater than 0.002 for CEY graphite. As Young's modulus and strength at rupture both increase with irradiation, and as no data are available, both the stress prgduced by a strain and the allowable stress are very uncertain. However, 1.5 x 10° psi is used for Young's modulus to indi- cate the value of the stresses that might be produced. The analysis presented here is based on a core 66 in. long by 27 in. in radius, made of 2- by 2- by 66-in. pieces, with axial fuel channels cut into each face. Some effects, particularly in relation to bowing, would be entirely dif- ferent for different-size pieces and a different arrangement. The flux depression within a graphite piece is due to the Progressive ther- malization of the neutrons and may be considered separately from the macroscopic variation of the flux across the piece, and the results added. The variation be- tween average and edge, which is proportionel to the tensile strain, amounts to 1.4% of the average flux (see Fig. 3.8). At the reactor center after a full year at 10 MwT, the strain would be 0.00027. With E = 1.5 x 10¥ psi, the stress would be 400 psi in tension. The axial shrinksge of a 66-in.-long Piece at the reactor axis is calculated to be 0.860 in.; for an edge piece the value is 0.122 in. (see Fig. 3.3). UNCLASSIFIED ORNL-LR-DWG 52048 0 // c = 0.5 / . g — = £ 1.0 — — I ) 1.5 : 0 4 8 12 16 20 24 28 RADIUS (in.) Fig. 3.3. Axial Shrinkage vs Radius After One Full-Power Year, 34 Transverse shrinkage, which is half of the axial shrinkage, is greater in the center than at the surface of the reactor, as shown in Table 3.l. This shrinkage would result in opening of the fuel passages, and to a greater extent in the center than at the surface. Table 3.1. Transverse Shrinkage (inches per inch per full-power year) Radial Location Axial Location Center Edge Center 0.0096 0.0010 End 0.0012 0.0002 The flux gradient across the graphite pieces would produce outward bowing at the midplane, allowing an unrestrained reactor to become barrel-shaped. The bowing effect reaches a maximum about three-fourths of the way out from the cen- ter, with the outer pieces less bowed. With the center of the reactor banded, the effect of transverse shrinkage of the pieces must be considered. The bowing would cause inward packing at the ends and outward packing, against the band, at the midplane. The ends would be keyed against rows slipping, with the result that concentric rings would be locked to- gether, with the rings inside relatively free. The pleces would bow outward at the midplane, increasing the fuel fraction most at the center. Deflections expected of unrestrained graphite pieces after a full-power year are listed in Table 3.2 for a radial row. Also tabulated are the shrinkage per 5. x 2-in. piece at the midplane and the strain if the pieces are kept straight. The hoop load to keep them straight, allowing for pieces bowing to teke up shrinkage but not considering the expansion of the band under the load, is 11,000 1b. 3.2 TEMPERATURE-RISE FFFECTS IN MSRE CORES WITH ROUND AND FLAT FUEL CHANNELS The temperature effects in cores with round and flat fuel channels, with the fuel in laminar flow, have been investigated. It has been found that the differ- ence between the fuel-graphite interface temperature and the mixed-mean fuel tem- perature would be excessive if round channels were used but would not be if the same fuel flow were provided in the form of flat channels. The two core configurations considered are jllustrated in Fig. 3.4. In both cases, it 1s assumed that the core is an asseumbly of graphite stringers of square cross-section extending the full height of the core. The fuel flows vertically through the core, making a single pass in both cases. In the round-channel case, the fuel channels are produced by machining a quarter-round of radius r at each corner of the graphite blocks. The blocks are packed in as tightly as possible, but it is assumed that there will be cracks of some width t between the blocks due to machining tolerances (these are exagger- ated in the sketch). 35 Table 3.2. Deflections, Shrinkages, and Strains at Midplane for a Radial Row After One Full-Power Year Piece Deflection Shrinkage Bending Strain No. (in.) (in.) (%) 0 (axis) 0 0.0191 0 1 0.057 0.0190 0.013 2 0.137 0.0186 0.030 3 0.192 0.0179 0.042 4 0.229 0.0170 0.050 5 0.272 0.0159 0.060 6 0.315 0.0145 0.069 7 0.367 0.0130 0.081 8 0.390 0.0113 0.086 9 0.407 0.0096 0.089 10 0.421 0.0078 0.092 11 0.412 0.0060 0.091 12 0.375 0.0042 0.082 13 0.315 0.0027 0.069 In the flat-channel case, the blocks are spaced spart a distance 2r in some convenient manner {e.g., by providing bosses on the blocks). In the comparison, it is assumed that the comparable cores contain equal fractional fuel cross-section f and the same center-to-center block spacing s. Poppendiek and Palm.er5’l'L have presented solutions for the difference between the temperature at the channel wall and the mixed-mean fuel temperature for a fluid in laminar flow containing a volume heat source. Their equations may be rewritten in terms of the geometric parameters of these systems as follows: 2 Q, F f s 1-F (t -t ) - R E { 0.229 [1 e Round nk [f+ =2 (1- 2./f/n )] rFr s bo—= . 2t fl-eJfl&)J -oa&'}. S Tor [ 2 { £(1- Fpp) } . = == |2 (1-./1-T 0.485 |1 + - 0.400). (t5-ty) kT [2 ( ) ] [ P (1-4/1-1)V/1-f } In the round-channel case, the terms containing t represent the contribution of the fuel trapped in the cracks. 36 NOTATION a = shorter dimension of rectangular body used in Jakob's6 solution for rectangular solid with uniform volume heat source, ft b = longer dimension of rectangular body, ft ¢ = value for temperature in rectangle tabulated on page 179, ref 6, dimensionless f = fractional fuel channel volume, dimensionless FPF = fraction of power released in fuel F = wall conduction factor in round-channel equation,5 dimensionless F/ = wall conduction factor in flat-channel equation,u dimensionless H = core height, ft k = thermal conductivity, Btu/hr(ft)(°F) P = total reactor power, Btu/hr p/pav = ratio of local to average power in core g’/ = volume heat source term in Jekob's® solution for rectangular body Qg = reactor power denmsity, Btu/hr(ftJ) Q, = fuel power demsity, Btu/hr{ft”) r = radius of round channel or half-width of square channel, ft = center-to-center spacing of graphite blocks in core, ft t = thickness of cracks between adjacent graphite blocks in round channel case, ft = temperature at channel wall 0 tm = mixed-mean fuel temperature 6 = local temperature in infinite rectangular body containing a volume heat source, the temperature at the boundaries being zero (6) = center-line temperature in infinite slab containing a volume heat b = e« source, the temperature at the boundaries being zero If we meke the assumptions that Fpp = 1 and t = 0 so that there is no heat transfer across the fuel-grephite interface, we then find that: (t -t ) °© ™ pound _ Lr (0.062) (to_t ) n(l-\/l-f)2 (0.085) T f1at For £ = 0.1, the approximate value of interest in the present case, (to-tm) Round 357 - (t,-t ) Flat 37 UNCLASSIFIED ORNL-LR— DWG 52049 CHANNEL. FUEL CHANNEL {#) FLAT-CHANNEL CORE Fig. 3.4. Core Assemblies, Thus it is clear that the volume heat source produces a much less severe temperature-rise problem in the flat-channel case. The simplifying assumptions made are favorable to the round channels. For the flat-channel core assumed for the MSRE, the following dimensions were assumed: s = 2in. = 0.167 ft, r = 0.05 1n., from which f = 0.097, W = flR]éDH (p_ffi ’ P = 10 M = 3.415 x 107 Btu/hr, R = 2.5 ft, H = 5.5 ft. 38 The nuclear calculations5 give the radial peak-to-average power as 2.08, and if the axial power distribution is assumed to follow the cosine law, the axial ratio is 1.57 and 1 _3:M13 %300 . 508 4 1.57 = 1.035 x 10° Bru/(hr)(ft0) . (@ > Max (2.5)° x 5.5 =) il Taking k of the fuel = 2.75 and Fpp 0.96, 6 - T2 o (1033 x 10 [0-167 ) [ 0.097 x 0.0k } b <2.75 7 0.007/ | 2 (10095 {0'”85 1+ 57598 (0.05)(0.95), ~0+*° 5 (3.88 x 106) (L.17 x 10"3) %%-(1 + 0.043) - 0.4%0 } = T.2°F. Jakob6 gives a solution applicable to the temperature rise in the graphite. His solution takes the form 6= c (6) O b ’ = (9 ) _ 9/// a2 Ol - o T2k ) For our case, ¢/ = Qg (1- FPF) = 0.04 x 1.033 x 106 = 4.13 x 1oLF Btu/hr(ft5) _ 2 -0.00 k = 12, a = 5095, < 0.083 f4, (6 ) - 4.15 X loh‘ X (O-O(le2 = 11 80F o’y L 2 x 12 : ? and for b/a =1, y/b =0, x/a =0, ref6 gives ¢ = 0.59. Thus 6 = 6.59°F. Therefore it is clear that neither fuel nor graphite temperatures would be excessive in the flat-channel core using laminar flow. For the round-channel core, excessive fuel temperatures would be encountered in laminar flow. Calcu- lations for turbulent flow show that the fuel temperature 1ls not excessive in round channels in that case. For the 0.2-in.-thick 1.25-in.-wide channels used in the actual core design, the fuel fraction in the core is 0.1188. The difference between the fuel-graphite interface temperature and the mixed-mean fuel temperature at the center of the core increases to 23.6°F, while there should be a negligible change in the tem- perature at the center of the graphite relative to that at the surface of the graphite, and the center of the graphite will be approximately 30°F above the mixed-mean fuel temperature at the center of the core. 39 3.3 TEMPERATURE OF FUEL IN A BLOCKED PASSAGE IN THE MSRE An analysis was made to determine the temperature which the MSRE fuel may attain 1f one of the fuel passages becomes blocked so that no fuel flow can occur. If unidirectional conduction is assumed, the problem is fairly straight- forward, although the answers will be ultraconservative. BEssentially it is a superposition of the following solutions: (a) conduction through an infinite slab of fuel with a uniform volume heat source, (b) conduction of the heat released in the fuel across a graphite slab to the adjacent fuel channel, (¢) temperature rise in the graphite due to its volume heat source, (d) temperature rise in the adjacent open fuel channel due to its volume heat source. Part a is the solution of the equation with the boundary conditions . . Wt f(x=0) = 2k For the MSRE core at the point of maximum power density, g = 1.02 x 107 Btu/hr(rt’), k = 2.75 Btu/hr(fit)(°F), £ = 0.05 in. = 0.00417 ft, and t(x:O) = 52.1 F- Part b is the solution of the same equation with boundary conditions: 0 X = £, , 1/q ) tc X (A at x =0 . 40 From the solution to part a, % = -k _t7 , t ! £20) T = -qf = 1.02 x 106 x 4.17 x 10'5 = L2.5 x 1o3 Btu/hr(ftg) : =8 | 3 ;1 (Q) _ Lo.5 x 107 3 tl= (A) = 5 = 3.54 x 10° °F/ft . Solving the equation for part b, 2 3 qct tc(g) = 3.5k x 107 + F3z £ = 1.9 in. = 0.158 ft, L 3 q, = k.56 x 10" Btu/hr(£t’) , k, = 12 Btu/hr(ft)(°F) , 2 4 -1 ) 3 -1 k.56 x 107 x (1.58 x 10 7) t(4) = 3.5% x 10° x 1.56 x 10 ™ + 2 x 12 560 + k7.5 = 607.5°F. Part & is the solution to the laminar-flow equations for flat channels: 2 Qr - £ L Q) _ I 3z 4.56 x 10 x 0.158 ) . (dA = k_ <5.35 x 107 + = ) - 12 <5.55 x 10° + B , 4.67 x 10" , ] Qv = 1.02 x 107 Btu/hr(ft3) for fuel at the center of the reactor, H ! 0.05 in. = k.17 x 10° Tt, Qr = 425 x 10% Btu/nr(rt?) . . L.67 x 104 h.o5 x th Thus F/ = 1 i b b 4 e B 4] Qr2 £t = Y 17F - 14 o m =k 35 ’ 2 1.02 x 107 x (4,17 x 103) < 35.7 - 14 ) 2.75 35 ? 64l x 0.62 = 39.9°F. Thus the temperature in the blocked channel will be 32.1 + 607.5 + 39.9 = 679.5°F above the mixed-mean temperature in the adjacent open channel. It is evident that the temperature rise across the graphite block is the major factor in the above, and it seemed of interest to investigate the effect of two-dimensional conduction in the graphite. Since no analytical solution was found, this problem has been solved by a relaxation procedure, and the graphite temperature drop was found to be 218°F or about a third of that for unidirectional conduction. This answer is low since it is based on the assumption that the graphite - flowing-fuel interface temperature is constant, while it doubtless is somewhat higher near the blocked channel than elsewhere. Based on this solution, the maximum temperature in the blocked channel would be about 290°F above that in an adjacent flowing-fuel channel. The above results suggest that if a single fuel channel were blocked, the fuel temperature probably would be no more than 400°F above the mixed-mean tem- perature in adjacent open fuel passages. 3.4 MSRE REACTOR PHYSICS One-dimensional multigroup and two-dimensional two-group calculations have been performed to obtain estimates of critical mass, flux and power-density dis- tributions, and temperature coefficient of reactivity. Other calculations during the period were concerned with gamma heating in the various INOR-8 structures assoclated with the core, drain-tank criticality, and estimates of the heat de- position and radiation dose due to fission products in the pump bowl. 3.4.1 Core Caleculations For the purpose of survey calculations the reactor core was considered as a bare right-circular cylinder 54 in. in diameter and 66 in. high. The total geo~ metric buckling of the bare reactor was then inserted in the input to GNUT as a transverse buckling, and the reactor was treated as a slab with zero net current on both faces. (This procedure eliminated the iteration on source shape and con- sequently gave a considerable reduction in computing time.) Calculated critical masses for fuel volume fractions of 0.08, 0.10, 0.12, 0.14, and 0.16 are shown in Fig. 3.5. Fuel-salt compositions are listed in 42 UNCLASSIFIED ORNL—LR-DWG 52050 80 160 INVENTORY, COMPOSITION A —_— 70 / 140 [ ] A / 9 120 \ . 50 < INVENTORY,COMPOSHW;;:;"\n\‘~‘ l\ A CRITICAL MASS, COMPOSITION A 30 i — 60 e J— 3 ~ 20 ,,””’ 40 CRITICAL MASS, COMPOSITION B .08 0.10 0.12 0.4 0.16 VOLUME FRACTION FUEL SALT Fig. 3.5. Calculated Critical Masses for Various Fuel Volume Fractions. Table 3.3. Fuel-5alt Compositions (Atomic densities in atoms per barn centimeter) Constituent Composition A Composition B Be 8.820 x 1077 1.002 x 1072 Li6 5.463 x 1077 6.20h x 1077 11! 1.621 x 10°° 2.068 x 1072 (99.997%) F 4.154 x 1072 4,718 x 1072 Zr 0 1.292 x 1077 Th 1.138 x 1077 0 °3? 2.660 x 107" 5,021 x 107 P30 1.850 x 1077 £.100 x 1077 43 Similar calculations were performed for composition A with 8 vol % fuel salt Tor reactor diameters of 3.5, 4.0, 4.5, and 5.0 ft and reactor heights of 5.5 and 10 ft. Calculated criticsl masses are shown in Fig. 3.6. Multigroup one-dimensional calculations were also performed for the reactor model illustrated in Fig. 3.7. These calculations, one along the midplane and one along the center line, were employed to generate two-group conskants for use in two-dimensional calculations with the IBM-704 program Equipoise. In general, good agreement was obtained between the results of critical calculations by the different methods. Two-group flux and adjoint-function distributions were obtained from the Equipoise® calculations and were used to estimate the neutron lifetime and the reactivity loss associated with changing from isothermal startup conditions to power operation. The calculated neutron lifetime was 290 usec; the perturbation- theory estimate of the reactivity loss was 3.9 x 10~ 8k/k, compared to 5.0 x 10-% as obtained from calculation of both the perturbed and unperturbed cases. Part of the above discrepancy was due to neglecting the change in diffusion coeffi- cient in the perturbation-theory calculations. UNCLASSIFIED ORNL—LR—DWG 50592A 50 ) 10 ft H o« < = \\\\\t * ® | b 55 ftH _ (@] g ) = 20 . 2 & - Y QG b el [=] > o 10 < - | & \ ] _.b 0 0 3.5 4.0 4.5 5.0 CORE DIAMETER (ft) Fig. 3.6. Critical Masses and Concen- trations for Various Diameters. 44 UNCLASSIFIEQ ORNL—LR-DWG 50588 - - L can (Vs INOR 8) A ANNULUS | | VESSEL (¥ INOR 8) CORE 2} ° GRID | { 2% 2] GRID 2 | HEADER DIMENSIONS ARE IN INCHES Fig. 3.7. Reactor Model. An estimate of the quantity (8k/k)/(8M/M), where M is the mass of fuel in the reactor, was cobtained from the GNU calculations. Assuming that the only ef- fect of a temperature increase is the removal of fue% from the core, the estimated temperature coefficient of reactivity was -2.4 x 107 (6k/fi)/°F based on a volu- metric expansion coefficient of the fuel salt of 1.0 x 107'/°F and a (sk/k)/(eM/M) value of Q.24 as obtained from the GNU calculations. 3.4.2 Gamma-Heating Calculations The two-dimensional Equipoise calculations were used to provide source esti- mates for the gamma-heating calculations, which were done with the IBM-T7O4 pro- gram.Nightmare.9 Results of the gamma-heating calculations are given in Table 3.4 for a reactor 54 in. in diameter and 66 in. high, with 8 vol % fuel salt in the core, operating at 5 Mw. 3.4.3 Drain-Tank Criticality Criticality calculations were performed7 for a cylinder 5 ft long and 5 ft high containing fuel salt. This cylinder, somewhat larger than the currently proposed drain tank, had a calculated multiplication constant of O.4k4 when bare and 0.77 when reflected with an essentially infinite layer of water. 45 Table 3.4. Results of Gamma-Heating Calculations for MSRE at 5 Mw Heat Generation Location (w/cc) Core can 0.10 Pressure vessel (inside) ] 0.10 midplane Pressure vessel (outside) 0.065 Pressure vessel (center line) 0.19 Upper support grid i . 1.12 } center line Lower support grid 0.66 3.4.4 Pump-Bowl Fission-Product Activities The equilibrium activity due to gaseous fission products in the pump bowl was estimated by the method described by Stevenson.1® The heat generation in the pump bowl amounted to about 6 kw for a reactor power of 5 Mw. 3.4.5 Cell Calculations Three—firoup two-dimensional calculations were performed with the IBM-704 program PDg 1l t6 obtain an estimate of the flux distribution within a moderator block. Sample traverses of the flux above 0.3 Mev are shown in Fig. 5.0, 3.5 ANALOG-COMPUTER STUDY OF MSRE PRIMARY-SYSTEM FLOW LOSS An analog-computer analysis was made of a loss-of-flow accident in the pri- mary system of the MSRE. The temperature changes and the maximum temperatures attained in the system were of primary interest. UNCLASSIFIED ORNL-LR-DWG 52051 ——_ . . X ™~ ~— FUEL SALT N " S — e 7 @) 0.5 o-bc)/ be (%) Q o Se 2R S m Z.os ,,,ff"!/////// _ @ L -1.0 ‘ /// @ (———” s | Fig. 3.8. Three-Group PDQ Cell Calculation. 46 3.5.1 Description of the System Simulated Thermal System.--A preliminary analysis was made of the thermal system, based on preliminary design information (Table 3.5). No after-heat, convection cooling, or moderator temperature coefficient of reactivity was simulated in this analysis. These phenomena will be simulated in subsequent analyses when Table 3.5. Computer Information on MSRE Reactor inlet temperature 1175°F Reactor outlet temperature 1225°F Mean graphite temperature 1250°F Residence time in reactor 3.88 sec Film drop from graphite to fuel Constant at all flows Heat capacity of graphite Loko E%E (cy = 0.425 ffiqu) Prompt 7 and neutron heating in graphite 9% of 10-Mw power Residence time in piping from reactor outlet to H.E. inlet 0.75 sec Residence time in H.E. l.3 sec Heat capacity of metal in H.E. 200 Btu/°F Average film drop between primary coolant and metal at design point 65°F Average drop in metal at design point 66°F Average film drop between metal and secondary coolant at design point 31°F Film drop between primary coolant and metal as function of flow: See graph, displace curve if necessary so that at 10 fps velocity, At = 65°F Mean secondary-coolant temperature at design point 1038°F Residence time in piping between H.E. outlet and inlet (including coolant annulus) 2.88 sec Total circulation time 8.81 sec Temperature coefficlent of reactivity -9 x lO-5 °F‘l Melting point of primary coolant 842°F Melting point of secondary coolant QLO°F Check points . ft hr °F Th 1 ermal resistances ( Bia ) In primary coolant film 3.28 x :I_O-l'L In metal 3,32 X 10"1‘L In secondary coolant £ilm 1.56 x lO“L‘L Temperature differential t -t = 250°F at thermal convec- core HE tion heat removal of 10% of 10-Mw power 47 suitable design information is available. The secondary coolant system was simu- lated merely as a heat dump since the main interest is in the primary system. A very elementary schematic of the system is shown in Fig. 3.9. It is understood that this was a simulation of primary flow stoppage and not one of primary pump stoppage. Pump "run down" information is unavailable at this time. The primary flow rate was decreased exponentially on periods of 1.5, 3, 6, and 10 sec. The heat transfer coefficient between the primary coolant and the heat ex- changer wall was made to vary with primary flow rate in accordance with the curve shown in Fig. 3.10. This curve was derived from the curve supplied by the de- signers, which is shown in Fig. 3.11. Nuclear System.--The delayed neutrons were lumped into one weighted group. Using the values from document LA-211& (dated 1957) for i and Bi, X for the one group was found to be 0.0769 sec™t and B is 6.4 x 10°3. Using the given fuel transit times and the curves in ORNL-LR-Dwg. 8919, B was found to be 3.4685 x 1072 6 where B’/ = ;{1 aiBi and @; is the ratio of the population of the ith group of i=1 UNCLASSIFIED ORNL—LR—DWG 46961 [ Umits oF smutation ] REACTOR REACTOR PUMP INLET, I OUTLET, % — L REACTOR | l | | | | ! PRIMARY LOOP | ! 1 | | | | —— HE.%ETLEE ORIMARY HEAT H.E.INLET, 7, PUMP i EXCHANGER —/VWWWWWWWIWH——= I 1 SECONDARY LOOP —— -~ SECONDARY HEAT EXCHANGER Fig. 3.9. Schematic Drawing of MSRE, UNCLASSIFIED QRNL — LR— DWG 469634 / / HEAT TRANSFER COEFFICIENT oy 0 2 4 6 8 10 PRIMARY FLOW VELOCITY (fps) Fig, 3.10. MSRE Flow Rate vs Heat Transfer Coefficlent. delayed neutrons inside a circulating-fuel reactor to the population of this group in an equivalent stagnant reactor. In the simulation, the delayed-neutron contribution is a function of the flow rate. The method used in the simulation is not precise; however, it appears to be a good approximation for excursions with periods of the same order of mag- nitude as those encountered in this analysis. 3.5.2 Analog-Computer Program The analog-computer program of this simulation is filed as ORNL drawing D-40325. (Copies may be obtained from the print files in the Engineering and Mechanical Division.) 49 UNCLASSIFIED ORNL — LR—DWG 46962A o \\ 60 40 Sl AT =Teyer = The wae CF) 20 0 5 10 15 20 PRIMARY FLOW VELOCITY (fps) Fig. 3.11, MSRE (Tf - Tw) vs Primary Flow Rate for a Constant Heat Transfer Rate. 3.5.3 Gimulator Operation With the simulator in steady-state operation at design-point conditions, the switch was thrown to decrease the primary flow rate exponentially on a given period. This procedure was repeated for various periods. Pertinent information (temperatures, etc.) was recorded versus time. These curves are included in Figs. 3.12, 3.13, 3%.14, 3%.15 and 3.16. Loss of load was also simulated. 3.5.4 Conditions Used to Obtain Curves The conditions used in the simulation to obtain the reported curves are de- scribed below. For Fig. 3.12, the primary temperature coefficient of reactivity is -9 x 107 (5k/k)/°F. The primary flow rate was decreased exponentially from 100% flow to zero flow on the following periods: Curve No. Period (sec) 1 1.5 2 5 3 6 H 10 No after-heat or convection cooling was considered. L ok o RT3 50 UNCLASSIFIED ORNL-LR~-DWG 46964A 1500 n o l_ wl - l_ =) O o 1400 : - . © ‘ \ <&> __._.—-———_——4\ o — —— w L*%:‘:::—-______\\ 3 ' —— E /‘ : —— ) 1 L 1300 /9/ - - i ' { T o = Ll '-7 > / o Fig. 3.16. Primary Mean Temperature vs Reciprocal of Temperature Coef- ficient. 53 For Fig. 3%.16, the primary flow rate was decreased from 100% flow to zero flow exponentially on a period of 3 sec for all runs. A number of runs were made with different primary-coolant temperature coefficients of reactivity. The maxi- mum primary mean temperature attained in each run was plotted against the recip- rocal of the primary temperature coefficient of reactivity for that run. Each run comprises a point on this curve. 1. 2. 10. 11. REFERENCES D. R. DeHalas, HLO Graphite News Letter No. 2, HW 65642 (June 14, 1960), p kL. J. H. W. Simmons, "The Effects of Irradiation on the Mechanical Properties of Graphite,” Sections E and F, Proceedings of the Third Conference on Car- bon, Pergamon Press, 1959. H. F. Poppendiek and L. D. Palmer, Forced Convection Heat Transfer in Pipes with Volume Heat Sources Within the Fiuids, ORNL-1395 (Dec. 2, 1952). H. F. Poppendiek and L. D. Palmer, Forced Convection Heat Transfer Between Parallel Plates and in Annuli with Volume Heat Sources Within the Fluids, ORNL-1701 (May 11, 1954). J. W. Miller, personal communication. M. Jakob, Heat Transfer, vol. I, p 176-179, Wiley, New York, 1949. C. L. Davis, J. M. Bookston, and B. E. Smith, GNU-IT - A Multigroup One- Dimensional Diffusion Program for the IBM-704, General Motors Report GMR 101 (1957) M. L. Tobias and T. B. Fowler, Equipoise - An IBM-704 Code for the Solution of Two-Group Neutron Diffusion Equations in Cylindrical Geometry, ORNL-2967 (in press). M. P. Lietzke and M. L. Tobias, personal communication. R. B. Stevenson, Radiation Source Strengths in the Expansion Chamber and Off-Gas System of the ART, ORNL CF-57-7-17 (July 1957). G. G. Bilodeau et al., PDQ - An IBM-704 Code to Solve the Two-Dimensional Few-Group Neutron-Diffusion Equations, WAPD-TM-70 (1957 . ot Y PART Il. MATERIALS STUDIES 4. METALLURGY 4.1 DYNAMIC-CORROSION STUDIES 4.1.1 Forced~Convection Loops The final hot-leg insert was removed from INOR-8 forced-convection loop 93544 after 15,140 hr of operation. As previously discussed,l three inserts were installed at the end of the hot-leg section of this loop to provide data on the weight changes of INOR-8 in contact with a beryllium-base fluoride fuel mixture. The operating schedule specified removal of one insert after 5000 hr of exposure, another after 10,000, and the third after 15,000. Operation of the loop began in July 1958 under the conditions: Salt mixture LiF-BeF,UF) (62-37-1 mole %) Max. salt-metal interface temp. 1300°F Min. salt temp. 1100°F AT 200°F Reynolds number 1600 Flow rate 2 gpm Examination of the third insert showed an average weight loss of 1.7 mg/cmg, +6%, along its L-in. length. If uniform removal of surface metal 1s assumed, this weight loss corresponds to a loss in wall thickness of 0.08 mil, *&%. The first two inserts, which were removed from the loop after 5000 and 10,000 hr, showed respective weight losses of 1.8 mg/eme, *2%, and 2.1 mg/cm®, f3%. The corresponding losses in wall thickness were calculated to be 0.08 and 0.09 mil, respectively. The similarity of these values indicates that no significant weight loss occurred after the first 5000 hr of operation. Om all three inserts no loss in wall thickness could be detected from measurements of the inserts before and after test. As previously reported,2 Metallography of the 5000- and 10,000-hr inserts revealed no evidence of attack other than the formation of a thin corrosion film on the surface exposed to the salts. A zone comprised of unusually small grains was also noted below the corrosion film on both inserts to a depth of 1 to 2 mils. Metallography of a transverse section of the 15,000-hr insert revealed light surface roughening, as shown in Fig. 4.1. As in the case of the 5000- and 10,000- hr inserts, the thin corrosion f£ilm and the band of fine-grained material were again found. The corrosion film was the same thickness as found on the 10,000-hr insert; however, the band of fine-grained material had increased in size, compared to the 10,000-hr insert. 35 56 UNCLASSIFIED T-19159 - mat ! . " . / ’ <" 3 , 011 o1z < ? ' 013 % ¢ 4 AN <~ 014 Fig. 4.1. Transverse Section of 15,000-hr Insert Removed from INOR-8 Forced-Convection Loop 9354-4. Etchant: 3 parts HCL, 2 parts H;0, 1 part 10% chromic acid. 250X. The band of fine-grained material is attributed to recrystallization of the matrix near the inside surfaces of the insert. This assumption is supported by the fact that the inserts had been reamed prior to service, which probably induced some cold work along the inside surfaces. The status of two Inconel and nine INOR-8 forced-convection loops which are in operation with various fluoride mixtures is summarized in Sec. 2.6. 4.1.2 Microprobe Analyses of Surface Film The compositions of thin corrosion films found on several of the long-term INOR-8 loop specimens after salt exposure have been investigated by means of an electron-beam, microprobe analyzer. (These analyses were carried out by the Ernest F. Fullam Corp., Schenectady, N. Y.) Included in these examinations were hot-leg specimens from herma%-convection loop 1226iref 3) and from forced- convection loops 9354-4lref 4) and 935&-5(T9f ). Pilms on these specimens, which ranged from 1/4 to 1/3 mil in thickness, were first examined by aiming the electron beam directly on the inner surface, which excited an area of 1.96 x 10-2 eme on the surface to a depth of approximately lu . A scan analysis was then made along a cross section of the specimen from forced-convection test 9354-L4, This latter specimen was sectioned along a plane tangent to the inside surface of the specimen so as to increase the apparent cross section of the tube wall and, correspondingly, the corrosion film. The results of the spectrographic analyses obtained by using both methods are compared with as-received analyses of these specimens in Tables 4.1 and 4.2. Table 4.1. Spot Analyses Made of Surface Films Source of Approx. Temp. Salt Spot Analysis (wt %)* Specimen Location of Specimen Circulated (°F) Ni Cr Fe Mo As-received (Heat SP-16) 71.66 6.99 4.85 15.82 Loop 1226 Hot leg 1248 No. 131 69.2 0 0.4 23.4 (section 2) Loop 9354-4 End of second 1300 No. 130 67.6 0 1.2 23.6 heater leg (10,000- hr insert) Loop 9354-5 End of second 1296 No. 130 66.8 0 0.6 21.5 heater leg (section 11) Loop 9354-5 Recheck 67.0 21.6 * et +cdf Estimated error = I5%. LS 58 Table 4.2. YScan" Analysis Made of 10,00-hr Insert Specimen from Forced-Convection Loop 9354-4 Distance from>’° Composition (wt %)° Inner Surface (Mils) Ni Cr Fe Mo Surface 67.6 0 1.2 23.6 0.5 8.3 2.0 3.2 19.9 1.0 69,5 2.7 4,2 19.9 1.5 68.1 3.3 .2 17.5 2.0 66.1 4.0 .5 17.6 2.5 70.0 .7 4.5 17.6 3.0 69.4 Lh,7 .5 16.4 4,0 69.9 6.7 4.8 15.8 5.0 71.6 6.9 4.8 15.8 aMeasurements made along cross section tangent to inner surface. b Film ends at position just prior to 1.0-mil measutrement. c Estimated error = 15%. As shown in Table 4.1, the spot analyses of the surfaces of all three samples indicated an increase in the molybdenum content and virtually complete depletion of chromium and iron. The single-scan analysis (Table 4.2) shows the depletion of chromium and iron to have occurred to a depth of approximately 3 mils on the "magnified” cross section. Nickel is seen to have decreased and molybdenum to have increased over about the same depth. If these results are adjusted to account for the magnification created by the specimen preparation, this concen- tration gradient is found to occur in about a 1/2-mil-thick surface. As can be seen in both tables, the percentages of Ni, Cr, Fe, and Mo do not total 100%, although the deviation from 100% is within the estimated accuracy of the instrument. To ensure that the deviation was not associated with residual fluorides remaining on the surfaces of the specimens, the hot-leg specimen from loop 9354-5 was cleaned with an ammonium oxalate solution and then re-analyzed. (This solution dissolves fluorides quite effectively without disturbing alloys composed predominantly of nickel.) The results of this re-analysis, as shown in Table 4.1, showed no difference from the first analysis. L.2 WELDING AND BRAZING STUDIES Y.2.1 Solidified-Metal-Seal Development The use of solidified metal seals for elevated-temperature, leak-tight, quick~disconnect joints has been considered for a number of applicgtions includ- ing high-vacuum valve and flange studies for the Sherwood Project. Such seals appear to be especially suitable for molten-salt reactor applications since they are readily applicable for remote handling and since relatively simple equipment may be used for making and breaking the joint. Conventional types of seals containing such materials as rubber, Teflon, and vacuum grease are obviously not possible in view of the high temperature and highly corrosive environment. 59 In general, two designs have been considered, the first containing an alloy sump and a tongue-and-groove joint design and the second containing an alloy- impregnated metal-fiber compact which would be used similar to an O-ring. Three ductile metals, corrosion-resistant to fused salt, were selected as potential sealing materials, and two corrosion-resistant base metals were chosen for the preliminary wetting and compatibility studies. A list of these materials is shovn in Table 4.3. Table 4.3. Materials Under Study for Solidified Metal Seals Melting Base Seal Materials Point (°F) Materials Remarks Gold 1945 INOR-8 Reactor structural Copper 1981 material 80 Au-20 Cu (wt %) 1625 Molybdenum Very limited solubility in liquid Au or Cu Compatibility specimens were made with each of the above combinations, and aging studies are being conducted at 1292C°F and 90°F above the melting point of the seal material. These conditions simulate reactor service temperature and a typical seal opening or closing temperature. These specimens will be examined for base-metal erosion and solid-state diffusion of seal-metal constituents. In an effort to demonstrate feasibility of these types of solidified-metal seals, several small components have been constructed and evaluated. A sump-type seal, fabricated from INOR-8 and sealed with 80 Au-20 Cu, is shown in Fig. 4.2, It was made and broken by induction heating in argon and was helium leak-tested after every sealing. To date, this seal has been found helium leaktight after each of ten cycles of breaking and sealing. Testing will be continued in an effort to determine the lifetime. In addition to the sump-type seal, an unsintered molybdenum-fiber compact was prepared by the Metallurgy Division Powder Metallurgy Group and impregnated with Au-Cu. This was used to make a seal between two molybdenum plates, as shown in Fig. 4.3. The seal was spring loaded for opening and closing while the component was induction heated under inert gas. A drawing showing this method of making and breaking seals is shown in Fig. 4.4, After each cycle of breaking and sealing, the seal was helium leak-checked. To date, the sample has been found leaktight after each of five completed cycles. No solution of the molybdenum prlates is evident, and no additional alloy has been added. Tests of this type will also be continued to determine the lifetime of the Jjoint. As a result of the above tests, it is believed that unsintered molybdenum- fiber compacts impregnated with alloy have sufficient strength at sealing temperature to resist tearing. To make this specimen more applicable to proposed molten-salt operating conditions, a molybdenum-fiber compact will be used between INOR-8 plates on the next seal instead of between molybdenum plates. 4.2.2 Brazing of Graphite Since advanced molten-salt reactor designs will probably include heat exchange systems of graphite, the development of techniques for fabricating such systems has been a subject of study.7:8 Corrosion-resistant brazing alloys in UNCLASSIFIED Y-36196 Fig. 4.2. Test Rig with Sump-Type Seal, the Au-Ni-Ta and Au-Ni-Mo systems have been developed for joining graphite to itself and to metals, and a typical graphite-tube-to-INOR-B header mockup assembly was constructed. A method of leak testing these graphite-to-metal joints was developed and preliminary results were obtained. The test consists of brazing molybdenum caps on low-permeability graphite tubes and filling the assembly with isopropancl, which has a room-temperature viscosity similar to that of the fused salts at operating temperature. An inlet at the top of the assembly provides for a means of internally pressurizing with argon, as shown in Fig. 4.5, 61 UNCLASSIFIED Y-36360 'f(,'] ’l ) -t \,g\'»\)‘- A N AT \% ‘1\ 4\ e 5 O AR 8 Fig. 4.3. Unsintered Molybdenum-Fiber Compact Impregnated with Braze Metal. UNCLASSIFIED ORNL- LR~ 0OWG 52052 MAKING BREAKING SEAL SEAL AFTER BREAKING SEAL T Fig. 4.4. Spring Loading of Impreg- nated Fiber-Compact Seals. 62 UNCLASSIFIED Y-35467 Fig. 4.5. Rig for Testing Graphite-to-Metal Joints. On initial tests, leakage at the brazed joint was observed below 5 psig. However, Jjoints that were more leaktight were obtained by oxidizing the ends of the graphite tubes prior to brazing. It is believed that oxidizing makes the surface of the graphite more porous and provides an additional keying action with the brazing alloy. A similar keying action has been reported in brazing tests on the porous-type AGOT graphite.® Table 4.4t summarizes the results of three tests made in this manner, Table 4.4. Results of Leak Tests on Graphite-to-Molybdenum Joints Part g . Appearance Test No. Tested Brazing Alloy (wt %) 5P PIret LRak 1 Top 70 Au-10 Ni-20 Ta None to 50 psig Bottom 75 Au-10 Ni-15 Ta At 10 psig 2 Top 70 Au-20 Ni-10 Ta None to 35 psig Bottom 70 Au-20 Ni-10 Ta At 20 psig Bottom 75 Au-20 Ni-5 Ta At 35 psig (rebrazed) 3 Top 60 Au-10 Ni-30 Ta None to 60 psig Bottom 60 Au-10 Ni-30 Ta At 35 psig 63 4 .2,3 Heat Exchanger Fabrication The main heat exchanger of the MSRE will consist of approximately 250 INOR-8 (1/2 in. in OD, and 0.045 in. in wall thickness) tubes joined to a 1-1/2-in.- thick TNOR-8 header. Welding and back-brazing procedures are being developed for fabricating the tube-to-header joints in this unit. The basic technigues are those previously used for fabricating heat exchange components containing thick tube sheets.9 Figure 4.6 is a cross-sectional view of the tube-to-header joint configura- tion selected for the heat exchanger. Trepanning on the weld side serves to ensure good heat distribution by providing a uniform weld geometry and to minimize tube-sheet distortion and restraint in the vicinity of the weld Jjoint. In back brazing these joints, there is a unique heating problem because of the large differences in mass of the tube and tube sheet. As a result, the tube- sheet temperature continuously lags the tube temperature during heating, and if UNCLASSIFIED ORNL—LR—-DWG 52053 S/TUBE BRAZE SIDE N TREPAN \ N THREE FEEDER HOLES . WELD 9DE//// TREPAN (o) % (&) Fig. 4.6, Tube-to-Tube-Sheet Joint. (a) Prior to welding and back brazing; (b) after welding and back brazing. 64 brazing alloys are applied to the conventional location at the tube-to-tube-sheet junction, the alloy preferentially flows to the hottest member, with consequent poor flow on the tube sheet. Placing the braze metal in a trepanned sump removes the problem by allowing the brazing alloy to remain at the temperature of the tube sheet through the entire braze cycle. The sump acts as a reservoir from which the alloy flows to fill the annulus between the tube and the tube sheet and to form a fillet- where the tube meets the tube-sheet surface. The fillet also aids in inspection since the capillary joint must be filled with alloy to form a fillet. The use of back brazing in the fabrication of this unit is unique, since high-temperature brazes are not normally used on tube sheets of this thickness. Optimum welding conditions for the tube-to-tube-sheet joints to obtain weld penetrations of 1.5t and 1t (t = tube-wall thickness) were determined and are listed below. Conditions for Conditions for 1.5t Penetration 1t Penetration Welding current, amp 55 45 Welding speed, in./min 6.3 8.5 Inert gas Argon Argon Arc length, in. 0.050 0.050 Those conditions which give maximum weld penetration of 1.5t also cause severe "roll over" of the weld and thus a constriction in the tube entrance. Therefore, welds with 1t penetration, and less "roll over,' were considered to be more suitable. A seven-tube test assembly was then constructed, and a photograph showing the braze side of the unit is shown in Fig. 4,7. Brazing was conducted in dry hydrogen, using the 82 Au-18 Ni (wt %) alloy. The brazing temperature was 1830°F, and a rate-of-temperature rise of 3OOOF/hr was selected, since it approximates that to be readily obtained with commercial facilities. Good filleting was observed on all seven tubes, and the general visual appearance of the welds and brazed joints was excellent. A metallographic evaluation of several joints is being conducted. A large assembly containing about 20 tubes is being planned to further demonstrate reliability in making a large number of joints. L.2.4 Mechanical Properties of INOR-8 Most of the mechanical-property tests on INOR-8 were completed in 1959, and the results of this program are presented in a topical report.lO A critical review of these data was made in order to establish design values. It was decided that up to 10500F, the design stresses would be based on two-thirds of the 0.2% offset yield strength, adjusted for the minimum specified yield strength. This value was chosen to be 35,000 psi. Above 1050°F the design stresses are based on the stress to produce 1% creep strain in 10 hr. This value was chosen because several heats of material did not exhibit a minimum creep rate. For heats which did exhibit a minimum creep rate, the stress value for 0.1 CRU was above the values based on creep strain. Stresses in the creep range were obtained from Larsen-Miller plots of data at 1100, 1200, 1250, 1300, 1Lk00, 1500, 1650, 1700, and 1800CF. Over 100 creep tests were performed, 65 UNCLASSIFIED Y-35957 Fig. 4.7. Tube-to-Tube-Sheet Sample, Showing Back Brazing. with times ranging from 0.1 to 25,000 hr; the majority were performed between 1100 and 1300°F. The values obtained from the Larsen-Miller curve are conservative with respect to the values obtained by direct extrapolation of creep data. Extrapolation of the rupture stresses to 102 hr revealed that they were well above the creep limits and did not affect the design stresses. The allowable stresses are shown in Table 4.5. Preliminary steps have been taken to obtain approval of these stresses by the ASME Council for The Boiler and Pressure Code. The stress values in Table 4.5 are therefore tentative and may be modified in accordance with the recommenda- tions of the Boiler Code Committee. A series of tensile tests at 1100 and 1300°F were performed to determine the effect of creep strains of less than 2% on the strength and ductility of INOR-8. A general pattern of the results is presented in Table 4.6. The only deleterious effect is on the elongation which, in our case, was a loss of 50% after approximately 12,000 hr of exposure. In spite of this relative loss in elongation, the minimum value was 20% in 2 in. 66 Table 4.5. Allowable Design Stresses for TNOR-8 Wrought and Annealed Sheet and Rod Temperature Allowable Stress, Static (°F) (psi) 100 23,300 200 20, 700 300 19, 100 400 17,900 500 16,900 600 16,200 700 15,600 800 15,300 850 14,900 900 14,700 950 14, 600 1000 14, 600 1050 14,000 1100 10, 40O 1150 7,200 1200 5,200 1225 4, 400 1250 3,700 1300 2,700 1350 2,050 1400 1,550 Teble L.6. Effect of Prior Creep on Tensile Properties of TNOR-8 Conditions Temp. of Change in Tensile Properties Material of Creep Tensile L Exposure ?8;? vield Tensile Strength Strength Elongation TNOR-8 rod 1250°F in air 1100 No change No change 15% loss up to 7500 hr 1250°F in air 1300 No change No change L0% loss up to 8000 hr INOR-8 sheet 1100°F in salt 1100 20% gain No change 50% loss up to 12,000 hr 1300°F in salt 1300 25% gain No change 25% loss up to 12,000 hr 1250°F in air 1300 No change No change 30% loss up to 8500 hr % All strains less than 2%. 67 4,3 PERMEATION AND APPARENT-DENSITY UNIFORMITY IN LARGE PIECES OF GRAPHITE Since there appears to be some advantage in using large pieces of graphite in the design of the molten-salt-reactor moderator, grades R-0025 and MH4LIM-82, which were the largest pieces of low-permeability graphite available, were tested for uniformity of properties. Specimens from various locations and with different orientations were subjected to the standard molten-salt permeation test. The original sizes of both grades, the location and orientation of the test specimens, and the test results are summarized in Table L4.7. In the fabrication of the graphite cylinder of grade R-0025, 1.5-in.-dla holes were drilled through one quadrant of the cylinder parallel to the axis. The piece was impregnated with pitch and then graphitized. The center specimens from this pilece were taken at the location of a hole. The other specimens were taken from a quadrant opposite the one that was drilled. It appeared that the pores of the R-0025 graphite may have been slightly more continuous (accessible) and/or larger in the direction parallel to the molding force (parallel to the cylinder wall) and that the lower porosity of the center specimens extended only 3/h in. from the surface of the center hole. In general, the results indicated that the large pieces of grades R-0025 and MH4YIM-82 were uniform. 4.3.1 Permeation of AGOT and S-4 Graphites by Molten Salts at Different Pressures To determine the effect of pressure ( <150 psig) on molten-salt permeation in graphite, and to determine approximately the relationship between molten-salt permeation and mercury permeation of graphite at room temperature, S-4 and AGOT Table 4.7. Apparent Densities of Graphite Grades R-0025 and MHLIM-82 and Degree of Permeation by LiF-BeF,-ThF),-UF), (67-18.5-14-0.5 mole %) in Relation to Specimen lLocation and Orientation Exposure conditions: 100 hr at 1300°F and 150 psig All values are averages of six; test specimens were 0.500 in. in diameter and 1.500 in. long Specimen Apparent Bulk Density Bulk Volume Permeated Orientation (g/cc) (%) Location with Respect to * ¥ * ' Forming Force R-0025" MH)LM-82 R-0025 MH), LM-82 Center Parallel 1.89 1.82 4.4 7.7 Perpendicular 1.85 6.4 Midway on Parallel 1.88 1.82 5.5 8.0 radius Perpendicular 1.86 6.7 At outside Parallel 1.87 1.81 5.2 8.4 diameter Perpendicular 1.87 1.80 5.7 8.7 * 39 in. in diameter, ll% in. long; size as fabricated. **L9 in. in diameter, L4 in. long; original length not given by vendor. 68 specimens were permeated with LiF-BeF,-ThF}-UF) (67-18.5-14-0.5 mole %) at 1300°F at pressures of 25, 65, and 150 psig In 100-hr exposure periods. A good correlation between the molten-salt and mercury permeation of graphite would contribute to the general indications that the molten fluorides do not wet graphite. Grade AGOT, because it is porous and has good uniformity, was included in the test as a control. Typical mercury permeation (intrusion) data for AGOT and R-0025 graphite (s-4 is a laboratory designation for grade R-0025) were used to plot curves for theoretical bulk permeation of graphite by molten salts.ll These curves were used to obtain the theoretical permeation data that are compared with the actual data in Table L.8. The agreement of the actual permeation values with the typical theoretical values is generally good except for grade S-4 in the 25-psig permeation. Either there is a structural difference in the graphite tested or this is one of the first indications of a slight wetting of graphite by a molten salt. Further investigation is planned. The molten-salt permeation data also show that pressure reduction from 150 to 25 psig does not appreciably decrease the salt permeation into these grades of graphite even though S-U4 is a moderately low permeability grade of graphite. Table 4.8. Comparison of Theoretical and Actual Molten-Fluoride Permeation of S-4 and AGOT Craphites at Different Pressures Test conditions: Temperature, 1300°F Exposure, 100 hr Salt, LiF-BeF,-ThF)-UF), (67-18.5-14-0.5 mole %) Bulk Volume of Graphite Permeated by Salt® (%) Permeation Pressure AGOT gl (psig) Theoretical Aectual Theoretical®® Actual o5 10.5 11.4 1.5 3.7 65 14,2 12.0 4,2 L4 150 15.0 13.2 5.4 h.9 * Each value is an average of six. *¥ These are based on the typical pore spectrum for grade R-0025; however, S-4 1s the laboratory designation for grade R-0025. 69 4.4 PRECTPITATION FROM MOLTEN-FLUORTDE FUEL IN CONTACT WITH VARTOUS VOLUMES OF CRAPHITE A single series of five precipitation tests have been made with AGOT graphite and LiF-BeFp-UF), (62-37-1 mole %) fuel in which only the volume of the graphite was varied in order to determine the relationship of graphite volume to uranium precipitated from the fuel. The test conditions and results are summarized in Table 4.9. The quantity of uranium that precipitated as U0y was determined chemically. The uranium precipitated per cubic centimeter of bulk volume of the graphite averaged (1.3 + 0.4)mg to (1.3-0.3)mg. This would be equivalent to approximately 9% by weight of the uranium in the fuel for a reactor system with a volume ratio of graphite to fuel of 9:1. This estimate is made with the assumptions that the volume of salt and the area of graphite in contact with the fuel do not affect the quantity of precipitate that occurs. Past data for volume ratios of graphite to fuel in the range of these tests support these assumptions. L.4.1 Removal of Contamination from Graphite It has been reportedle that the thermal decomposition of NHyF.HF in the presence of graphite removed its oxygen contamination to such an extent that the graphite could subsequently contain LiF-BeFp-UF), (62-37-1 mole %) at 1300°F without causing the usual UOo precipitation from the fuel. A duplicating test has produced the same results. Work is in progress to determine if INCR-8 is seriously attacked during the thermal decomposition of the NHMF-HF. Table 4.9. Uranium Precipitation from Molten LiF-BeF,-UF), (62-37-1 mole %) Exposed to ACOT Graphite Test conditions: Temperature, 1300°F Test period, 100 hr Atmosphere, vacuum Ratio of the projected surface of the graphite® to the volume of the fuel: 9.6 in.2/in.3 (3.8 cm2/cm3) Volume of the fuel: 0.4334 in.3 (7.102 ce) Ratio of Graphite Weight of Uranium Precipitation Volume to Fuel Total Per Unit Volume of Graphite Volume (mg) (mg/cc) 5:1 3k 0.97 10:1 83 1.2 15:1 177 1.7 20:1 202 1.4 27:17%% 203 1.1 * This is the area of the graphite that the molten fuel would be in direct contact with 1if the graphite were free of pores. *3% This is the ratio for the crucible size arbitrarily chosen for the standard permeation tests. 70 L.4.2 TNOR-8 -- Fuel -- Graphite Carburization Tests The tendency for TNOR-8 and Inconel to be carburized in LiF-BeFo-UF), (67-32-1 mole %) -- graphite system at 1300°F has been investigated in a series of tests conducted in multiples of 2000 hr up to 12,000 hr. No carburization was detected metallographically on specimens from any of the above tests. Sheet tensile specimens were included in each test system, and these were tested after exposure to the fuel-graphite to determine the effect of this exposure on their tensile properties. These tensile-test results are listed in Table 4.10. After 4000 hr of exposure, the corrosion on the Inconel specimens was heavy enough to mask any carburization that might have occurred; consequently, only the results obtained with INOR-8 tensile specimens are listed. Comparing the tensile strengths and elongation values of the long-term tested specimens with those of the short-term control specimens, it is evident that no significant change has occurred. Table L4.10. Tensile-Test Results on INOR-8 Specimens Exposed to Fuel- Graphite Systems Compared with Control Specimens Exposed to Argon for Various Times Time of Test Control Specimens* Test Specimens** (hr) Tensile Tensile Strength Elongation Strength Elongation (psi x 1073} (% in 2 in.) (psi x 10-3) (% in 2 in.) Room-Temperature Tests 2,000 123.2 ho.5 124.8 41.0 Iy, 000 123.0 h2.0 12L.5 43.0 6,000 12k .7 39.0 12k.2 36.0 8,000 127.7 43.5 128.2 36.5 10,000 130.8 43.0 130.3 41.0 12,000 130.6 36.5 126.7 36.3 1250CF Tests 2,000 4.6 18.0 75.6 18.5 4,000 75.9,76.0 19.0,20.0 76.2,T4.3 18.5,18.5 6,000 72.6 16.5 .5 16.0 8,000 76.7 17.5 76.6 19.0 10,000 78.9 15.0 80.8 17.5 12,000 72.8 15.5 75.8 17.0 * Control specimens were exposed to ARGON at 1300CF for the times indicated. % Test specimens were exposed to LiF-BeF,-UF) (62-37-1 mole %) - Graphite at 1300°F for the times indicated. 71 These data indicate that unstressed INOR-8 is not detectably carburized in the fuel-graphite system (described above) at 1300°F for exposures as long as 12,000 hr. 4.5 IN-PILE TESTS Two MSR graphite-fuel capsules (ORNL-MTR-47-1 and -2) were irradiated and were then removed from the MTE on March 11 and June 20, respectively. The first experiment was irradiated for 720 hr and the second for 1600 hr. No serious difficulties were encountered. The two capsules were shipped to BMI for post- irradiation examination. Disassembly is scheduled to start August 15. REFERENCES 1. MSR Quar. Prog. Rep. Jan. 31, 1958, ORNL-217L4, p 31. 2. MSR Quar, Prog. Rep. Oct. 31, 1959, ORNL-2980, p 35. 3. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 33-38. L. MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 55. 5. MSR Quar. Prog. Rep. Oct. 31, 1959, ORNL-2990, p 35. 6. J. W. Tackett, Progress Report — Bakable High-Vacuum Valve and Flange Studies, ORNL CF-59-2-3 (Feb. 6, 1959). 7. MSR Quar. Prog. Rep. Oct. 31, 1959, ORNL-2890, p 33-36. 8. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2799, p 45. 9. R. L. Heestand, ORNL-24L0, p 159-62 (classified). 10. R. W. Swindeman, The Mechanical Properties of INOR-8, ORNL-2780 (to be published). ll1. Private communication from F. F. Blankenship and P. S. Spangler of the Reactor Chemistry Division. 12. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 59. 5. CHEMISTRY 5.1 PHASE EQUILIBRTUM STUDIES 5.1.1 MSRE Fuel and Coolant For design purposes, the MSRE fuel has been chosen at the composition 117F- BeF,-ThF),-UF), (65-30-%-1 mole %; m. p. 4500C). Chemically, this composition can be roughly approximated as Lip-BeFy to which 5 mole % of quadrivalent fluorides are added; however, the exact proportions are based on carefully established phase diagramsl showing the primary phase fields and lowest liquidus temperatures asso- ciated with a given UF) and ThF), content. Fortunately, for concentration ranges of interest as MSRE fuels, the quadrivalent fluorides are interchangeable, with 1ittle effect on the melting point. Furthermore, the LiF-BeFp solvent can hold as much as 15 mole % quadrivalent fluorides in solution without requiring temperatures in excess of 5000C. The preferred fuel compositions for MSRE-type reactors occur just outside the primary phase field of LiF, and thus take advantage of both the unusually large fiieezing- oint depression of LiF caused by strongly "acidice" cations such as Bett, Th*+, and U + and the diminished stability ranges of 3LiF-MF, crystals (as compared with 3:1 coumbinations containing other alkali fluorides). The uranium concentration in the MSRE will be adjusted by adding concentrate to a carrier. For example, the addition of LiF-UFy (73-27 mole %; m. p. 490°C) to the carrier, IiF-ThFj-BeFs (64.75-4.15-31.1 mole %; m. p. 450°C), gives the final fuel composition: 65-30-4-1 mole ¢%; m. p. 450°C. The concentrate is also used for replenishment. During mixing, a complete range of intermediate compositions has at least a transient existence before uniform blending is accomplished, and it is impor- tant that no composition corresponding to a high-melting-point compound or insoluble solid be encountered. Although the concentrate was chosen to conform with this re-~ quirement, to establish with certainty the temperature above which no solid appears, equilibrated samples have been quenched from the neighborhood of the liguidus tem- perature throughout the range between concentrate and carrier. The identified phases in some of the quenched samples are listed in Table 5.1, and the liquidus temperatures derived from these identifications are shown in Fig. 5.1. The absence of any high-melting-point precipitate along the quaternary join conforms with ex- pectations based on previously established ternary behavior. The coolant for the MSRE, L17F—BeF2 (66-34 mole %; m. p- 415°C) closely resem- bles the fuel in that LisBel) can be considered as the predominant constituent. As a consequence, the fuel and coolant are compatible and no chemical interaction ensues, merely dilution of the quadrivalent cations, in case of inadvertent mixing. The pro- portions of LiF and BeF, were selected primarily as a compromise between increasing melting points at greater IiF content and increasing viscosities at greater BeFo contents. Also, a coolant with a higher freezing point than the fuel provides added insurance against accidental freezing of the fuel by too rapid heat removal. 5.1.2 Systems Containing ThFu A realization that the compounds previously reported as K¥:ThFj (ref 2) and RbF-ThF), (ref 3) are actually TKF-6ThF), (ref 4) and TRbF+6ThFy (ref 4) led to a review of KF-ThF), and RbF-ThF) phase relationships. Recent studies of these systems, 72 Table 5.1. Thermal Gradient Quenching Data for the System LiF-BeF2~UFh~ThFh Composition Phase Change (mole %) Temperature Phases Found Just Phases Found Just LiF BeF, UF), ThF), (°c) Above Phase Change Below Phase Change €5 20 1 4 4u8 T ox Liquid Iiquid + 2IiF-BeFs + the 5LiF-ThF)+ Ss¥*¥* 65 30 1 L yoz ¥ o Liquid + 2LiF-BeF, Liquid + 2LiF*BeF + the 5LiF-ThFlL ss + the 7LiF-6(U,Th§Fh ss, with 9 mole % UFy 66.4 24,9 5.4 3.3 w6 T2 Liquid Liquid + the TIiF-6 (U,Th)F), ss, with 21 mole % UF), 68 18.7 10.8 2.5 w6 T o Tiquid Liquid + the TLiF-6 (U,Th)F), ss, with 34 mole % UF) 69.7 12.4 16.2 1.7 w1 ¥ o Liquid Iiquid + LiF + the TLiF-6(U,Th)F), ss, with 38 mole % UFY 71.h 6.2 21.6 0.8 483 %1 Iiquid Liquid + the 7LiF-6 (U,Th)Fy ss, with 43 mole % UF) 1.4 6.2 21.6 0.8 580 I o Liquid + the TLiF-6 Liquid + IAF + the (U,Th)F)y ss TLiF-6(U,Th)F) ss ¥ The uncertainties in temperature indicate the temperature differences between the gquenched samples from which the values were obtained. ¥** So0l1id solution. gL UNCLASSIFIED ORNL- LR - DWG 50121 600 | % ‘ | S 5501——L)4—~'4¢—-*#T—f'1 ) o i; 500 P—F—T ) ‘ - *L; g | uoum;;14j2,,/¥" § 450 ¢o———Ia *JT),ét._ ,7‘7_,_‘,_‘ z \’ | | | | | i 400 —%——l——-}——-l_——j I | | x | | 0 10 20 27 UE, {mole Fo) Fig. 5.1. Quaternary Join Between LiF- UF, (73-27 mole %) and LiF-BeF,-Th¥, (64.75- 31.1-4.15 mole %). now nearing completion, are leading to significant modifications in the phase diagrams. The diagram portraying the phases in the system KF-ThF), based on new data for the composition region 30 to 90 mole % Th¥F), is shown in Fig. 5.2. Invariant equilibria are listed in Table 5.2. The formulas and crystal structures of the jntermediate compounds in this system were reported originally by Zachariasen. The cubic phases, SNaF-3UFy and 3NaF-2ThFy, are structural analogs of a phase which Zachariasen identified as 0-2KF+ThFy; attempts to isolate and identify this phase are continuing. Demonstration of a homogeneous solid phase at TS5 mole % ThF'), corroborates the view of Asker, Segnit and Wylie's that the compound of highest ThFu/KF ratio in the system is KF-3ThF) rather than KF-6ThF), as was previously claimed.2 In fact, it now appears that KF-+3ThF) melts incongruently to ThF) ss and liquid, that it undergoes solid state transitions which account for the exis- tence of three crystal polymorphs, and that it is unstable with respect to KF-2ThF) and ThF) below T775°C. Although relatively detailed work had been made on the phases in the system KF-ThF) before the current studies, the system RbF-ThF) had been investigated in only a cursory fashion.”? Results of thermal gradient quenching experiments in the system RbF-ThF) now reveal five intermediate compounds, 3RbF+ThF), 2RbF-ThFy, 7RbF-6ThF),, RbF-3ThFy, and RbF+6ThFy, in contrast with ZRbF+ThF,, RbF«ThFy, and RbF«3ThF) as reported by Dergunov and Bergman. Elucidation of the solld state relationships among the compounds RbF, 3RbF+ThF)y, and 2RbF:ThF) is continuing. 5.1.3 The System ZrFu'ThFu In conjunction with problems arising from the presence of ZrFy in ThF)-contain- ing fuels, thermal gradient quenching experiments of binary ZrF) -Th¥F), mixtures have shown that the system ZrF)-ThFy forms continuous solid solutions with a minimum at approximately 23 mole % ThF), and at 858°. Although the four tetravalent heavy-metal 75 UNCLASSIFIED ORNL-LR-0WG 513914 1200 5:1= 5KF-ThF, 3:4== 3KF-ThF, 4= 2KF-Th 1mohflfié—§EPG$F ~7 o= p / / 1:2 = KF-2ThE, / 1:3 = KF-3ThF, LIQUID // ThF455+ / 1000 Pd LQuio /| at3+uow0\\\ ////, / . ¢/ ThE, ss BL3+L@UM\§"’C— - _::f;- o 716+ LIQUID - \‘at3+Thass _ ////r N~ Lo pr:3+ BY:3+ThE, < . o adi + ///’ % k\\\\uouu)\ // \\\\ /// = 7.6+ 5 BOO LIQUID oy add + hed 1:2+ThE > KF -+ E%Lg LIQUID i3 ’ 4 - LIQUID \\\ l 700 \V I e M KF + a5:1 3: | 764142 o T - B[5:1+ 3+ 600 34 B2 - 1:2+ThE, M A+7: KF+35:4 _ BeA+T6 500 0 - 2 BS54+ R2:| 400 KF 10 20 30 40 50 60 70 80 90 ThE, Thl-;(mole%) Fig. 5.2. The Binary System KF-ThF,. fluorides, ZrFh, HfF), ThF), and UFy are all monocliniec, the unit cell parameters indicate a close similarity between ZrF) and HfF), and between ThF}y and UFy. This relationship is reflected by the temperature minimum in the continuous solid solu- tions between ZrF)-UF) and ZrF)-ThF), while ThF)-UF) forms continuous solid solutions without a temperature minimum. 5.2 EFFECT OF TETRAVALENT FLUORIDES ON THE FREEZING POINT OF SODIUM FLUORIDE Freezing-point temperatures of NalF were measured for solutions containing ZrF), HfF), ThF), and UF}, at concentrations up to 15 mole % MF). This investigation was undertaken to further understand molten fluoride solution behavior, in particular to see if tetravalent fluorides followed the same correlations as divalent solutes.’ Freezing points with a precision of ¥20c were obtainead by conventional thermal analyses techniques, using calibrated thermocouples. Melts were stirred by means of an argon gas stream. Table 5.2. Invariant Equilibria in the System KF-ThF), ThF, in Invariant Ligquid Temperature Type of Phase Reaction at (mole %) (°c) Equilibrium Invariant Temperature 1k 69k Eutectic Liquid <= KF + p-5KF-ThF) 635 Inversion a-SKF+ThF), 3= B~5KF-ThF) 712 Peritectic Liquid + §KF-ThFh -— a-SKF'ThFh 25 875 Congruent Liquid 3= 3KF-ThF) melting point 570 Decomposition §K.'F-ThFh -— B-SKF-ThFh + QKF-ThF)_L 32 652 Eutectic Liquid == 3KF-ThF, + 2KF - ThF) 32.5 658 Peritectic Liquid + 7KF-6ThFh -— 2KF * ThF) 46,2 899 Congruent Liquid == TKF-6th melting point 52 870 Eutectic Liquid 3= TKF+6ThF) + KF-2ThF) 65 929 Peritectic Liquid + Q-KF.3ThF, _— KF - 2ThF) 67 938 Peritectic Liquid + ThF) ss = O-KF-3ThF) 934 Inversion Cx—KF-jThFh -— oz-KF-fiThFu 808 Inversion B-KF+3ThF) -— y-KF - 3ThF, T22 Decomposition 7-KF-5ThFh -— KF-EThFh + ThFh 9L 77 The results, summarized in Table 5.3, show that the values for Zth and Hth agree within experimental error and that these two solutes give higher freezing points than UFy, which in turn gives a higher value than ThFy. Thus, the magnitude of the freezing point depression of NaF increases with increasing radius of the quadrivalent solute cation, in contrast with divalent fluoride solutes for which a decrease in radius is associated with an increase in freezing-point depression. However, a significant difference between the two types of solutes is the stronger electric field around the tetravalent ion and the accompanying tendency toward higher coordination numbers in the liquid state. Assuming eightfold coordination, and according to the crystal radii, only the thorium cation is sufficiently large for all eight fluorides to be in simultaneous contact with the cation. Possibly the smaller cations cannot effectively complex as many fluorides as thorium and, as a consequence, exert a smaller influence on the activity of NaF. In other words, zri+ and Hfu+, having ionie radii that are smaller and nearly identical, have a smaller capacity for fluoride ions, would thus be less acid, and would show smaller but equal negative deviation with the basic NaF. The notion of eightfold coordination of quadrivalent ions in solution is sup- ported by the absorption spectra of UF), in LiF-NaF-KF eutectic6 where the predomi- nant absorption maximum corresponds closely to the maximum for solid UFu;7 the positions of eight fluorides surrounding each uranium in the UFY crystal has been established by x-ray diffraction. Activity coefficients less than unity can re- sult from negative heats of solution or from positive excess entropies of mixing. The heats of solution probably follow previously established correlations with cation radii, but the entropies apparently have more influence in the case of the quadrivalent cations. It appears that NaF is more disordered by ThFy (big cations) as a solute than by ZrF) (little cations), and that this effect is large enough to reverse the sequence of increasing negative deviations expected from only electro- static considerations. 5.2.1 Phase Diagrams of Fluoride Systems A decennial index to the ORNL reports (1950 to 1960) which contain data on fluoride, chloride, and hydroxide phase-equilibrium studies has been issued.? Table 5.3. NaF Liquidus Temperatures (°C) in NaPF-MF) Systems Solute Solute (mole %) ZrF), HEF), ThF), UFy, 0.0 995 995 995 995 2.0 98 985 985 5.0 968 967 96 967 10.0 924 925 911 916 15.0 854 854 823 834 * For ZrF) the measurement was taken at 2.1 mole % and was 983.3. The value given is an extrapolation to 2.0 mole %. 78 5.2.2 The System LiF-YF3z A recently completed phase diagram of the system 1iF-YFz (Fig. 5.3) differs substantially from that reported by Dergunov.10 Results of 8RNL phase studies have shown that the system contains the single binary compound LiF-YFz, melting incon- gruently to YFz and liquid at 815°C. The peritectic invariant point for this re- action occurs at 49 mole % YFz. A single eutectic occurs in the system LiF-YFz at 19 mole % YFz and at 695°C. The structure of the compound LiF-YFz, determined from single crystal x-ray diffraction measurements, has been shown to be tetra- gonal.ll Optical properties of the compound have been established. In conjunction with the ghase studies, the melting point of YFz has been de- termined to be (1144 * 3)°C.le A large exothermic effect in the YFxz heating and cooling curves indicates that YFz undergoes a solid state transition at 10600C. The temperature of this transition has been confirmed by the results of thermal gradient quenching experiments, but it has not been possible to preserve the high- temperature modification of YFz by quenching from temperatures above 10600C. Char- acterization of this form will probably be possible only with high-temperature x-ray diffraction measurements. UNCLASSIFIED ORNL- LR~ DWG 381164 1200 1100 ,////!//,//” 1000 /////// 900 / 3 & L / < 800 i N o = \ / W = 700 N 600 L.I_m bl u. 3 500 400 LiF 10 20 30 40 50 60 70 80 20 YF3 YF3 (mole 75) Fig. 5.3. ‘The Binary System LiF-YFj. 79 5.2.3 Melting Point of NiFo Apparently NiFp melts somewhere between 1400 ang 1500°C, probably about 1425 or 1450°C, but attempts to measure the melting point have been thwarted by the vola- tilization of NiFy, at the requisite temperatures and by some poorly understood phenomena. Previous reported values for the melting point, usually estimated,15'15 have ranged from 1027 to 1450°C.16 The growth of single crystals of NiFs by zone melting is alleged to have occurred with a hot-zone temperature of lh2000.1¥ The necessary temperatures were obtained in a platinum-wound resistance furnace or in a controlled-atmosphere Globar furnace. The furnace had three compartments: a sample heating area, thermocouple well, and a heating-element container. Separate control of the atmosphere in each of these compartments was possible. The enclosed heating elements were made of silicon carbide. Temperatures are measured with a Pt, Pt-10% Rh thermocouple located in a tantalum thermocouple well. Since the temperature is above the melting point of nickel, containment of molten NiF, presents a problem. Sealed platinum capsules of 5-mil wall thickness ruptured from internal pressure developed by NiFp at temperatures of 1400 to 15500C. (The extrapolated vapor pressure of solid NiFp is 650 mm at 1450°C.) In one experi- ment, it was observed that the rupturing of the capsule was accompanied by a flash of light. This effect might be an indication of dissociation to vield Fp. If, as in the case of copper(II) fluoride, fluorine is produced,l8 then a nickel(I) fluo- ride must exist, which has not, as yet, been isolated nor characterized. Graphite was unsuitable as container material because NiFp was reduced to nickel in the inert atmosphere of the high temperature furnace when a graphite liner was tried. Although the equilibrium products of the reactions involved in the dispropor- tionation of chromium(II) fluoride and iron(II) fluoridel9 have been identified, it has not yet been possible to identify all the reaction products from the dissocia- tion reactions of nickel(II) fluoride. 5.5 OXIDE BEHAVIOR 5.3.1 Oxide Behavior in Fuels High-melting-point oxides are generally very sparingly soluble in fluoride fuels, but when compared with each other at reactor temperatures the apparent solu- bilities may differ by several orders of magnitude. The differences provide the basis for metathetical reactions between oxides and fluorides, many of which are potentially useful for reprocessing fuels.20 However, the very low solubility of UO2 leads to hazards associated with unintended contamination by oxide. Recent experiments provide a tentative ranfiing for solubfilities, in solvents like the MSRE fuels, corresponding to Pa't < zrt < UM << ThH¥+ < Bett << Li*, in order of increasing solubility of the oxide. The possibility that ZrOs may be preferentially precipitated in the presence of UFy is under investigation. Quantitative evaluations are not yet available because of unexplained discrepancies in some of the analytical results, but there are plausible indications that the inclusion of ZrF) in the MSRE fuel, at a concen- tration of about 5 mole of ZrF) for 1 mole UFy, results in the solubility product for ZrOp being exceeded at an oxide concentration too low to precipitate UOp. If these results are borne out, the hazard of oxide precipitation in the MSRE could involve ZrO; ratner than UOp, and this much more favorable situation could be 80 achieved without causing a noticeable increase in vapor pressure OTr important changes in other physical properties. Demonstrations of the metathetical reactions involving oxides in fuel mixtures have been carried out in Pyrex apparatus in a furnace equipped for visual observa- tion. The oxide from Pyrex contaminates the melt sufficiently slowly that the oxide content can be considered constant over short ranges of time. The oxide present in clear solvents such as LiF-BeFp, mixtures can be titrated with UFy. A black precipitate of UOp settles to the bottom until the end point is reached, when the green color of UFy (detectible at 300 ppm) becomes apparent. Either ZrFy or ThFy added to LiF--BeF2 containing precipitated UOp caused a dissolution of at least part of the precipitate, and ZrOo seemed to be more insolu- ble than UOp. Analyses of samples removed with a glass filter stick from molten LioBeF), saturated with ThOp gave a solubility of 2.03 wt % or 8.05 x 10-3 moles of ThO» dissolved in 100 g, and the solubility of BeO appears to be about 0.22 mole/lOO g 5.%.2 Zirconium Oxyfluoride and Attempted Preparation of Uranous Oxyfluoride Oxyfluorides were studied because they may result from the contamination of the molten reactor fuel with oxides. Although some early workers claimed to have pre- pared UOFp, the compound has not been substantiated.2l Wright and Warf22 found that U0, and UFy did not combine on heating to 1000°C. The synthesis of ZrOFp from ZrF) and ZrOo was performed and crystallographic data were obtained. Zirconium oxyfluoride also sometimes forms when ZrOo 1s reacted with molten NHpHF5. Although UOF, is not known, an attempt was made to form a mixed uranous- zirconium oxyfluoride. The ionic radius of ut* (0.97A) is sufficiently close to that of Zr+% (0.80A) that a solid-solid solution was considered likely. Reaction of ZrF), with UOp at 1000°C yielded UF), and ZrOp rather than the hypothetical UZr(OF5)p. If a solid solution does form, it is very limited, as UF) was still a major constituent of a fusion corresponding to the composition 47100 . 4ZrF) - UF), - UOs. 5.% GRAPHITE COMPATIBILITY 5.4.1 Removal of Oxide from Graphite by Treatment with Hydrogen The feasibility of a hydrogen treatment for removing oxide from the MSRE moderator graphite appears to be unsatisfactory. A specimen (1—1/2 in. in diameter and 2 in. long) of low-permeability R-0025 graphite obtained from the National Car- von Co. was treated with hydrogen at 6000C for 6 hr at a flow rate of ~2 liters/min. Following this treatment, the specimen was degassed at 1000°C for 21 hr. The results are given in Table 5.4 along with data obtained from two similar specimens of R-0025 graphite which had been degassed thoroughly at 600°C prior to degassing at 1000°C. Because of the variation of the gas content from sample to sample, it is diffi- cult to establish whether the hydrogen treatment had any beneficial effect. At best the treatment removed less than half the surface oxide and, since this would be the portion most readily removed, a satisfactory cleanup would require a prohibitively long period of treatment. 81 Table 5.4. Gases Removed from R-0025 Graphite at 1000°C Volume Composition of Evolved Gas Evolved (vol %) (em>/100 emd Hydro- of graphite) Ho Hs0 carbons co No CO, Ho-fired 6.7 66 1 6 23 3 0.5 Not Ho-fired T.5 55 1 2 34 1 5 Not Hp-fired 8.7 Sk 0.3 2 3l L L A sample of AGOT-NC6, which 1s more pervious, is being hydrogen-treated. Pre- liminary indications here also point toward a removal by the Ho treatment of only about half the oxide in the graphite. 5.4.2 Behavior of Graphite when Wetted by a Molten Fluoride Among the molten fluorides which have been observed to soak through graphite containers rather rapidly, and are therefore presumed to wet graphite, are CsF, PoFp, and SnFp. Of these, SnFp (m. p. 213°C) has been selected as the most con- venient for study. The presumptive evidence for wetting by SnFo was confirmed by the appearance of a thin film of salt on graphite which has been exposed to SnFo melts. In a rough measurement with a DeNouy tensiometer the surface tension of Snkoy at 290°C was found to be about S0 dynes/cm, compared with about 190 to 200 dynes/cm for the MSRE fuel. As an exploratory test, the rate of permeation by SnFp was followed by examin- ing l-in.-long cylinders (1/2 in. in diameter) of AGOT graphite which had been immersed in molten SnF2 for various times at about 230°C. Weight increases, as obtained from density measurements after removing the adherent salt, varied approxi- mately linearly with the logarithm of time in the interval between 10 and 1000 wmin (10, 20, and 27% at 10, 100, and 1000 min, respectively). The rate during this interval was conjectured to have been controlled chiefly by the diffusion of trapped gas, occupying the graphite voids. 5.5 PREPARATION OF PURIFIED MATERTIALS S1x 100-kg batches of LiF-BeFp-ThF)-UF) (65.0-30.0-4.0-1.0 mole %) have been purified in addition to numerous small-scale preparations of a variety of melts. The effectiveness of strong reducing agents, such as zirconium metal, for pretreat- ing coolant mixtures to exhaust oxidizing impurities, seems to be hampered by deposits, as yet unidentified, on the surface of the reducing metal. Experiments along this line are continuing. REFERENCES 1. R. E. Thoma, Phase Diagrams of Nuclear Reactor Materials, ORNL-2548, pp 81, 109 (Nov. 6, 1959), 2. W. H. Zachariasen, J. Am. Chem. Soc. 70, 2147 (1948). 22. 82 E. P. Dergunov and A. G. Bergman, Doklady Akad. Nauk. 5.8.8.R. 60, 391 (1948). R. E. Thoma, Crystal Structure of Some Compounds of UF4 and ThFh with Alkali Fluorides, ORNL CF-58-12-40 (Dec. 11, 1953). S. Cantor, Reactor Chem. Div. Ann. Prog. Rep. Jan. 31, 1960, ORNL-2931, p 48-L9. J. P. Young and J. C. White, Anal. Chem. 32, 793 (1960); also, personal commu- nication, J. P. Young. J. G. Conway, The Absorption Spectrum of UFL and the Energy levels of Uranium V, UCRL-8613 (Feb. 3, 1959). R. D. Burbank, The Crystal Structure of Uranium Tetrafluoride, K-769 (June 6, 1951). R. E. Thoma, Fused Salt Phase Equilibria--A Decennial Index to ORNL Progress Reports, ORNL CF-60-7-52 (July 20, 1960). E. P. Dergunov, Doklady Akad. Nauk, S.5.5.R. 60, 1185 (1948). Met. Ann. Prog. Rep. Sept. 1, 1959, ORNL-2839, p 288. 3. Cantor and T. S. Carlton, personal communication. L. L. Quill (ed.), Chemistry and Metallurgy of Miscellaneous Materials: Thermodynemics, p 202 of Natl. Nuclear Energy Ser. Div. IV, vol. 19B, McGraw- Hill, New York, 1950. M. Blander and F. F. Blankenship, ANP Quar. Prog. Rep. Sept. 30, 1957, ORNL-23%87, p 121 (classified). C. Poulenc, Ann. Chem. Phys. 2, b4l (189L4). L. M. Matarrese, The Magnetic Anisotropy of Iron Group Fluorides, Thesis, Univ. Chicago, Aug. 3, 195k. H. Guggenheim, J. Phys. Chem. gi, 9%8 (1960). H. V. Wartenberg, Z. anorg. Chem. 241, 381 (19%9) . J. D. Redman, MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p U45. MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 81. J. L. Katz and E. Rabinowitch, The Chemistry of Uranium, Part I, The Element Tts Binary and Related Compounds, lst ed., p 576-T7, McGraw-Hil1l, New York, 1951. J. Wright and J. Warf, "Attempted Preparation of Uranous Oxyfluoride,” p 19 in See. V of Technical Report, Ames Projects, Chemical Research, Analytical Chem- istry, Report for Period Feb. 1 to Mar. 10, 1944, by C. A. Thomas, F. A. Sped- ding, and H. A. Wilhelm. 6. ENGINEERING RESEARCH 6.1 PHYSICAL-PROPERTY MEASUREMENTS 6.1.1 Surface Tension and Density The surface tensions of two additional NaF-BeFp (57-43 mole %) mixtures were determdined in the temperature range of 500 to 800°C using the maximum-bubble- The results are shown in Fig, 6.1 in comparison with the previously reported measurement® for a relsted mixture (NaF-BeFz; 63-37 mole %).° A least-squares analysis of the three sets of data yielded the linear correla- tions given in Table 6.1; similar equations for two mixtures containing small additions of UF4 and ThF4 to the NaF-BeF> base salt are also tabulated. pressure technique.1 SURFACE TENSION (dynes /cm) 250 UNCLASSIFIED ORNL-LR-DWG 52054 200 - 8OO —“-—-_._____. ‘ A-'—':-'_ —.-‘:"_m::. :‘._._._ | T\“"—-—-—_ — 150 — - SAMPLE NO. SYMBOL COMPOSITION (mole %) NaF BeF, 1 . 571 429 ’ 2 4 57.5 42.5 00 3 0 63.0 37.0 500 550 600 650 700 750 TEMPERATURE (°C) Fig. 6.1. Comparison of Surface Ten- sions of Several NaF-BeF, Mixtures, Table 6.1. Surface Tensions of Seversl BeFa~Containing Fluoride Mixtures * Surface Tension Composition Sample (mole %) o (dynes/cm) ve- NaF BeFz UFy ThFg a b 1 57.1 42.9 - - au6.h 0,116 2 57.5 42,5 - - 269.4 0.136 3 63.0 37.0 - - 293.2 0.165 4 62.0 36,5 0.5 1.0 272.2 0.143 5 62.0 37.0 1.0 - 235.5 0,090 *s=a-bt for t in °C. 83 84 The results for the two NaF-BeFz (57-43 mole %) mixtures show somewhat lower surface tensions than that obtained with the NaF-BeFg (63-37 mole %) salt. Thus, at 650°C, o for sample 1 (see Fig. 6.1) is 8.5% below and ¢ for sample 2 is 2.5% below the value for sample 3. The discrepancy of 6% between samples 1 and 2 (having essentially identical compositions) may relate to the longer operational exposure of sample 1 in the circulating loop from which the specimens were drawn and hence to the increased smount of corrosion-product contaminents. An analysis of the samples for Fe, Ni, and Cr will be made. The effect of impurities on the surface tension of a fluid is illustrated by recent results on o for two distilled- water samples using the meximum-bubble-pressure method. For the first of these, a sample obtained directly from the laboratory distilled-B20 supply line, 0 was meas- ured as 67 to 68 dynes/cm; while for the second (triply distilled and doubly de- jonized), ¢ was found to be T2 to T3 dynes/cm. The surface tension was calculeted from the Schrodinger equa.tion,4 1 1.2 o =2 [too), - )] v -5 0, (1) written so as to make apparent the correction to the maximum bubble-pressure (as indicated by & precision water manometer) arising from the "hydrostatic" pressure due to the immersion of the capillary tip below the salt surface. In Eq. (1), r is the capillary radius, h, the fluid head, and p, the fluid density; the sub- scripts m and s designate the manometer fluid and the salt, respectively. Terms of the Schrodinger equation higher than the second degree in r have been neglected in that they represent an adjustment of only 0.1% in the value of the surface tension. Examination of Eq. (1) shows that the precision of the g-measurement may be related to errors in the pressure measurement, in the determinetion of the cap- illary radius, and in the value of the salt density. It is believed that the residual errors associated with the pressure and geometrical measurements have been reduced to a total of *3%. Thus the biggest remaining source of error is in the salt density; e.g., an uncertainty of +10% in the salt density leads to an pdditional *3% variation in o. Accordingly, an experimental determination of the densities of these mixtures is planned. A comperison of predicted and experimental densities for the NaF-BeFz (63-37 mole %) mixture is given in Fig. 6.2. It is seen from this figure that the curve UNCLASSIFIED ORNL-LR-DWG 52055 2.2 : \ 74_(‘?!NTERPOLATED,r«u_ - 2 ‘*afij\\\‘~a | " _ —~ (A) EXPERIMENTAL, ORNL Q -'---L-__-__._-_ T — o U ——— o0 b= h‘ —+=== t ‘--...__“\ \ | ‘j-—- T ~~(C) PREDICTED, ML B ~ < 1.9 (D) PREDICTED, ORNL- = ==md - .~ .8 600 650 700 750 800 850 TEMPERATURE {°C) Fig. 6.2. Comparison of Experimental and Predicted Densities for NaF-BeF, (63- 37 mole %). etk i 85 (&) representing the mean of the experimental measurements® at ORNL lies about 7% above the line (D) obtained from the Cohen and Jones® correlation of the measured densities of 15 fluoride mixtures of diversified composition. This correlation is reported as showing a maximum deviation of +6%. In similar fashion a general density correlation based on more than 60 measurements by Blanke et al.® at the Mound Laboratory on compositions in the NaF-BeFz-UF4 system yielded curve C. If the density data for NaF-BeFz mixtures alone are abstracted from the total of the Mound data and cross-plotted as a function of the NaF content (for various temper- atures), curve B results. In the calculation of the surface tensions reported in Table 6.1, the correlation developed from the data of Blanke et al, was used to estimate the salt densities, 6.1.2 Heat Capacity The correlation of the heat capacities of BeFa-containing salt mixtures has been revised to correct errors in the original presentation’ and to include addi- tional experimental data, The results are summarized in Fig., 6.3 in terms of the average heat capacity over the temperature range 550 to 800°C and the "mean molec- ular weight" (the ratio of the molecular weight to the average number of atom species in the mixture, M/fi). In obtaining the correlating line, ¢, = kel (/RO T (2) the data point for LiF-BeFa-UFq (70-10-20 mole %) was omitted, The maximum devia- tion of the remaining data from Eq. (2) is %8.5%; LiF-BeFa2-UFq (70-10-20 mole %) falls 16% below the correlation and its heat capacity will be redetermined. UNCLASSIFIED ORNL-LR-DWG 52056 1.0 o0 COMPOSITION , (mole %) -. 08 - {7 12| — |12 |0.325 | 30.34 o — 1 25| 80| — |15 |0.315 | 28 .14 . 07 - | 53| 46| - (1.0 |0.459!18.88 o 53 —1 46| — [1.0 |0.513 | 15.46 S o6 62| — | 37 — [1.0 |0.522 | 15.25 2 7t -1 16 {13 | — lo.312 | 25.87 a N 62| — | 365/1.0 | 0.5/0.491 |15.71 5> % 700 —| 10| — [20|0.246| 3.72 £ 05 . 700 —| 14 16 | — |0.323] 28.27 g \ 70.3| - 1 15.3/13.8| 0.6/0.335| 27.00 o \ u © '\. e ® = [ ] '\\ 3 0.3 = T ° \ 0.2 10 15 20 25 30 35 40 50 60 70 80 S0 100 A&u AVERAGE MOLECULAR WEIGHT V' AVERAGE NUMBER OF ATOMS Fig. 6.3. Correlation of Heat Capacities of Bel2- Containing Salt Mixtures. 86 6.2 HEAT-TRANSFER STUDIES The experimental studya of the heat-transfer coefficients for the salt mix- ture LiF-BeFp-UF4-ThF4 (67-18.5-0.5-14 mole %) flowing in heated Inconel and INOR-8 tubes was interrupted after 5560 hr of circulation to replace the LFB pump. Damage to the pump seems to be restricted to one of the shaft bearings; visual observation of the pump bowl and impeller showed no obvious effects from exposure to the salt. A new pump and belt drive have been installed. Since the flow characteristic of this pump can be expected to be somewhat different from that of the original pump, a recalibration with respect to the turbine flowmeter will be necessary. A large semple of salt was found in the vicinity of the ring seal at the top of the pump bowl. This sample was submitted for chemical analysis; the re- sults are listed in Table 6.2 in comparison with the nominal and preoperational compositions for the mixture, The extent of circulation of the ring-seal sample is unknown, since the salt could have been deposited and frozen in this location at any time during the 5560 hr of operation. A representative semple will be obtained from the sump for analysis. Although no measure of the magnesium con- tent of the original salt is available for comparison, it is believed that the presence of magnesium in the ring-seal semple results from the introduction of a small smount of magnesium-bearing thermal insulation into the system during the replacement of a liquid-level probe located in the pump. Table 6.2. Composition of LiF-BeFz-UF4-ThF4 Salt Mixture Semples Composition (mole %) Constituent Prior to Ring-Seal Nominal Operation Sample LiF 67.0 70.8 69.8 BeFo 18.5 16.6 16.0 ThF4 14.0 12.1 13.7 UF4 0.5 0.5 0.5 ppm Ni | 30 95 Cr 115 510 Fe 305 165 Mg * 2825 Mo * <10 * Not measured. 87 The effect of this uncertainty in composition on the analysis of the data is significant. For example, the use of the analysis for the preoperational salt rather than the nominal in estimating the properties results in a 3% in- crease in the mean heat capacity of the mixture (see Fig. 6.3). Variations could also be expected in the other thermal properties of interest — the viscos- ity and the thermal conductivity — though the changes will probably be of lesser magnitude than indicated for the heat capacity. REFERENCES MSR Quar. Prog. Rep. June 30, 1958, ORNL-2551, p 38. MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 26. Chemical analysis indicates a 63-37 mole % (NaF-BeFz) composition rather than the 6L4-36 mole % constitution previously designated for this sample, J. R. Partington, An Advanced Treatise on Physical Chemistry, Vol. 2: The Properties of Liquids, p 185, Longmans, Green and Co. (1951). S. I, Cohen and T. N. Jones, A Summary of Density Measurements on Molten Fluoride Mixtures and a Correlation for Predicting Densities of Fluoride Mixtures, CRNL-1702 (July 195k). B. C. Blanke et al., Densities of Fused Mixtures of Sodium Fluoride, Beryllium Fluoride, and Uranium Fluoride, Mound Laboratory Memorandum (MIM) -1076 (July 1958). MSR Quar, Prog. Rep. July 31, 1959, ORNL-2799, p 37. MSR Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 46; MSR Quar. Prog. Rep. July 31, 1959, ORNL-2799, p 39; MSR Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 27-29. 7. FUEL PROCESSING Thorium fluoride will be present in MSR fuels either intentionally in a oOne- region reactor or as a result of leakage of blanket salt into the fuel in a two- region reactor. Preliminary investigations indicate that SbFg-HF solutions may be useful for decontaminating the ThF) from rare-earth fission products. Since the 1iF carrier of the fuel salt interferes by precipitating the SbFg, it must be re- moved from the fuel by HF dissolution first, leaching the HF-insoluble residue with SbFg-HF solution would probably decontaminate the ThF), sufficiently for re-use in the blanket. The behavior of fission products other than rare earths has not yet been investigated. This particular solvent was considered because it has been reportedl that SbFg is the strongest "acid" in HF of a number of compounds investi- gated, and that rare-earth fluorides are soluble in SbF:--HF solutions. Other HF dissolution systems for MSR fuel processing are generally basic in nature. In the initial work, HF containing about 20% SbFg was used for the solvent. Rare-earth fluorides dissolved rapidly to give clear brown solutions; the absence of precipitation on partial evaporation indicated a high rare-earth scolubility, probably as the compound RE(SbF6)5. This formula was deduced from the results of adding liquid SbFc to dry powdered rare-earth fluorides. Considerable heat was evolved, and the liquid appeared to wet and darken the color of the powder. These effects were incomplete until 3 moles of SbFg had been added per mole of rare-earth fluoride; above this ratio some excess liquid SbF5 was present. The compound was not appreciably soluble in liquid SbFg but was extremely soluble in HF, to perhaps 200 to 300 g of rare-earth fluoride per liter. If less than three SbFsg molecules are present per rare-earth fluoride molecule, solution is incomplete, but excess SbF5 does not interfere. Addition of small amounts of water to the HF solutions resulted in a precipitate. Thorium tetrafluoride is probably less basic in HF than the rare earths. It did not react with SbFg or dissolve in SbFS-HF solutions. From mixtures of the rare- earth fluorides and ThF)y, SbFg5-HF solutions dissolved the rare earths but not the ThF). The 20% SbF: - HF reacted with LiF, forming an insoluble compound. Addition of SbFg to an HF solution of LiF resulted in precipitation of a white solid, pro- bably LiSbFg. Addition of LiF to an HF solution of SbFS resulted in formation of a white solid, apparently different from the LiF. A1l observations reported here are qualitative since analytical results have not yet been obtained. REFERENCE 1. A. F. Clifford et al., J. Inorg. Nucl. Chem. 5, 27 (1957). 88 OAK RIDGE NATIONAL LABORATORY MOLTEN SALT REACTOR PROGRAM OCTOBER 1, 1960 R. B. BRIGGS, DIRECTOR W. D. REEL, PROGRAM EDITOR D Tl MOLTEN SALT REACTOR EXPERIMENT A. L. BOCH, PROJECT ENGINEER R SPECIAL PROBLEMS W. B. MCDONALD* ADYANCED DEYELOPMENT ENGINEERING RESEARCH COMPONENT DEVELOPMENT FUEL REPROCESSING METALLURGY REACTOR CHEMISTRY H. W. HOFFMAN R . SPIEWAK™ R D. E. FERGUSON cT W. D. MANLY M W. R. GRIMES RC D. O. CAMPBELL cT R. G. DONNELLY M R. E. THOMA RC R. W. SWINDEMAN M G. M. HEBERT RC T. K. ROCHE M W. F. SCHAFFER RC R. G. SCHOOSTER M J. E. EORGAN RC C. M, BLOOD RC W. M. JOHNSON RC B. F. HITCH RC OESIGN METALLURGY REACTOR CHEMISTRY RADIATION TESTING COMPONENT DEVELOPMENT CONSTRUCTION OPERATIONS E.S. BETTIS R A. TABOADA M F. F. BLANKENSHIP RC 1. A. CONLIN R I. SPIEWAK* R W. B. MCDONALD* R S. E. BEALL C. H. WODTKE M B. 5. WEAVER RC W. L. SCOTT R D. SCOTT, JR, R 1. H. DEVAN M S. S. KIRSLIS RC W. H. COOK M F. A. DOSS RC M. A. REDDEN M . K. R. FINNELL RC J. L. GRIFFITH M R. G. WILEY RC W. JENNINGS RC *DUAL CAPACITY €T CHEMICAL TECHNOLOGY DIVISION E&M ENGINEERING AND MECHANICAL DIVISION I&C INSTRUMENTATION AND CONTROLS DIVISION M METALLURGY DIVISION BUILDING AND FACILITIES COMPONENTS SYSTEMS INSTRUMENTATION AND CONTROLS ANALYSIS RC REACTOR CHEMISTRY DIVISION R REACTOR DIVISION L. N. HOWELL E&M 3. H. WESTSIK R L. F. PARSLEY R J. R. TALLACKSON I&C B. W. KINYON R C. E. BETTIS E&M C. H. GABBARD R A. N, SMETH® R R. L. MOORE 14C _C. W. NESTOR R B&R BURNS AND ROE T. E. NORTHUP E&M A. G. GRINDELL R R. C. ROBERTSON R 0. W. BURKE 1&C D. W. VROOM R L LOCKHEED 0. C. WATKIN E&M P. G. SMITH R S. E.80LT R J. R. BROWN 1&C W. H. FORD L . T, WATSON E&M L. V. WILSON R R. E. RAMSEY BaR B. SQUIRES B&R W.L.BREAZEALE L D DIRECTOR'S DIVISION R. W. KNIGHT E&M H. R. PAYNE R R. P. CAPPELLO B&R TI TECHNICAL INFORMATION DIVISION C. B. HAYNIE E&M W. C. ULRICH R A. J. MADAIO B&R S. 0. LEWIS E&M J, C. MOYERS* R 5. 5. SCHILKE B&R 0. T. GIBSON E&M D. E. KLION B&R L. J. SCIRICA B&R G. H. LLEWELLYN E&M C. N. SRSICH B&R W. E. SALLEE E&M R. E. ANDERSEN B&R G. W. RENFRO E&M H. N. GORECKI B&R M. ROSENBAUM E&M $. E. SPERRY B&R J. T. COX E&M R. E. FRANKLIN R J. W, KIDD E&M F. J. UNDERWCOD E&M W. D. HOYLE E&M FLOW MODELS REMOTE MAINTENANCE FUEL HANDLING LOOPS AND TANKS GAS SYSTEMS R. J. KEDL R E. . HISE % AND HEATERS J. L. CROWLEY R AN, SMITH R T. L. CASTILE R C. K. MCGLOTHLAN R 1. C. MOYERS" R R. B. GALLAHER R P. P. HOLZ R M. RICHARDSON R J. R.DICKIE R J. R. SHUGART R G. A. WINDLE R W. H. DUCKWORTH R C. JONES R F. EISEN B&R 89 - ORNL-3014 UC-81 — Reactors-Power TTD-4500 (15th ed.) INTERNAT, DISTRIBUTION 1. G. M. Adamson 46, A, S. Householder 2. L. G. Alexander 47. L. N. Howell 3. S. E. Beall 48. W. H. Jordan 4, C,. E. Bettis 49, P. R. Kasten 5. E. S. Bettis 50. R. J. Kedl 6. D. 8. Billington 51. G. W. Keilholtz 7. F. F. Blankenship 52, C. P. Keinm g§. E. P. Blizard 53, M. T. Kelley 9. A. L, Boch 54. B. W. Kinyon 10. 8. E. Bolt 55. R. W. Knight 11. C. J. Borkowski 56. J. A. Lane 12. G. E. Boyd 57. M. T. Lundin 13. W. L. Breazeale 58, H. G. MacPherson 14. TE. J. Breeding 59. W. D. Manly 15. R. B. Briggs 60. E. R. Mann 16. F. R. Bruce 6l. W. B, McDonald 17. 0. W. Burke 62. C. K. McGlothlan 18, D. 0. Campbell 63. E. C. Miller 19, C. E. Center (K-25) 64. R. L. Moore 20, R. A. Charpie 65. K. Z. Morgan 21, W. G. Cobb 66, J. C. Moyers 22. R. S. Cockreham 67. J. P. Murray (Y-12) 23, J. A, Conlin 68. M. L. Nelson 24. W, H. Cook 69. C. W. Nestor 25. G. A. Cristy 70, T. E. Northup 26, J. L. Crowley 71l. W. R. Osborn 27. F. L. Culler 72. L. F. Parsly 28. D. A. Douglas 73. P. Patriarca 29. L. B. Emlet (K-25) 74. H. R. Payne 30. E. P. Epler 75. D. Phillips 31l. W. K. Ergen 76. W. B. Pike 32. W. H. Ford 77. R. E. Ramsey 33. A. P, Fraas 78. W. D. Reel 34. J. H. Frye, dJr. 79. P. M. Reyling 35, C. H. Gabbard 80. M. Richardson 36. W. R. Gall 8l. R. C. Robertson 37. R. B. Gallaher 82. T. K. Roche 38, W. R. Grimes 83. H. W. Savage 39, A. G. Grindell 84. D. Scott 40, C. S. Harrill 85. W. L. Scott 41, M. R. Hill 86. H. E. Seagren 42, E. C. Hise 87. E. D. Shipley 43, H. W. Hoffman 88. O. Sisman 44. P. P. Holz 89. M, J. Skinner 45. A, Hollaender 90, G. M. Slaughter 1 91. A, 92. P. 93. A. 9%. T. 95. C. 2. J. 97. R. 98. A. 9e. J. 100. E. 101. D. 102. W. 163. 164. 165-"741. N. Smith G. Smith H. Snell Spiewak D. Susano A, Swartout W. Swindeman Taboadsa R. Tallackson H. Taylor B. Trauger G. Ulrich 92 :]_03. 104, 105, 106. 107. 108, 109, 110-113. 114-158. 159. 160- :].62 L] D. C. Watkin A. M. Weinberg J. H. Westsik G. C. Williams L. V. Wilson C. E. Winters C. H. Wodtke ORNL — Y-12 Technical Library, Document Reference Section Laboratory Records Department Laboratory Records, ORNL R.C. Central Research Library EXTERNAL DISTRTBUTION D. H. Groelsema, AEC, Washington Division of Research and Development, AEC, ORO Given distribution as shown in TID-4500 {15th ed.) under Reactors-Power category (75 copies — OTS)