ORNL-4658 UC-80 — Reactor Technology CHEMICAL ASPECTS OF MSRE OPERATIONS Roy E. Thoma OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION DISTRIBUTION OF THIS DOCUMENT IS UNLIMITED Printed in the United States of America. Available from National Technical Information Service U.S. Department of Commerce 5285 Port Royal Road, Springfield, Virginia 22151 Price: Printed Copy $3.00; Microfiche $0.95 This report was prepared as an account of work sponsored by the United States Government. Neither the United States nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights. DISCLAIMER This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency Thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. DISCLAIMER Portions of this document may be illegible in electronic image products. Images are produced from the best available original document. Contract No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM ORNL-4658 UC-80 — Reactor Technology CHEMICAL ASPECTS OF MSRE OPERATIONS Roy E. Thoma REACTOR CHEMISTRY DIVISION NOTICE This report was prepared as an account of work sponsored by the United States Government, Neither the United States nor the United States Atomic Energy Commission, nor any of their employees, nor any of their contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, com- pleteness or usefulness of any information, apparatus, product or process disclosed, or represents that its use would not infringe privately owned rights, DECEMBER 1971 OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee 37830 operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION DISTRIBUTION OF THIS BUC ¥ . CONTENTS ABSTRACT EXECUTIVE SUMMARY ACKNOWLEDGMENTS KEY WORD INDEX I INTRODUCTION 2 CHEMICAL BEHAVIOR IN THE FUEL AND COOLANT SALT SYSTEMS DURING PRENUCLEAR OPERATIONS 2 1 Preoperational Procedures 2 2 Flush Salt 2 3 Coolant Salt 24 Fuel Salt 2 4 1 On Site Preparation 2 4 2 Uramum Assay 2 4 3 Structural Metal Impurities 2 4 4 Oxide Contaminants 24 5 Analysis of Helium Cover Gas 24 6 Lithium Analysis 2 4 7 Examination of Salts after Zero-Power Experiment 2 4 8 Appraisal of Chemical Surveillance in Prepower Tests 3 CHEMICAL COMPOSITION OF THE FUEL SALT DURING NUCLEAR OPERATIONS 31 Introduction 3 2 Component Analysis 33 Oxide Analysis 3 4 Uranium Concentration 3 5 Structural Metal Impurities 3 6 Chemical Effects of Reprocessing 3 7 Materal Balances for 2°*U and 2?2 U Operations 371 Recovery of 23°Uand ?3%U 3 7 2 Inventores for Stored Salts 37 3 Salt Loss from Leakage 111 10 14 15 15 16 19 26 30 31 31 33 36 36 50 50 52 53 55 57 57 58 59 10 11 12 1V CHEMICAL COMPOSITION OF THE FLUSH SALT DURING NUCLEAR OPERATIONS 4 1 Role of Flush Salt Analysis in the Determination of Salt Residue Masses 4 2 Transfer of Uranum and Plutonium to Flush Salt in >*3 U Operations 4 3 Flush Salt Loss to Off-Gas Holdup Tank CHEMICAL BEHAVIOR OF THE COOLANT SALT 51 Composition Analysis 5 2 Corrosion Behavior CORROSION IN THE FUEL CIRCUIT 6 1 Modes of Corrosion 6 2 Corrosion in Prenuclear Operations 6 3 Corroson in Power Operations 6 4 Additions of Reductants and Oxidants to the Fuel Salt 6 5 Effect of Uranium Trifluoride on the °° Nb Concentration of the Fuel Salt DETERMINATION OF REACTOR POWER 7 1 Power Estimates with 23°U Fuel from Heat Balance and Other Methods 7 2 Power Output of the MSBR Based on the Isotopic Composition of Plutonium 7 3 Isotopic Composition of Uranum during 233U Qperations 7 4 Isotopic Composition of Uranium during 22 °U Operations PHYSICAL PROPERTIES 8 1 General Properties 8 2 Denstty of Fuel and Coolant Salts 8 3 Crystallization of the MSRE Fuel INTERACTIONS OF FUEL SALT WITH MODERATOR GRAPHITE AND SURVEILLANCE SAMPLE MATERIALS CHEMICAL SURVEILLANCE OF AUXILIARY FLUID SYSTEMS 101 Water Systems 1011 Cooling Tower Water 10 1 2 Treated Water Supply 10 1 3 Vapor Condensing System 10 2 Hehum Cover Gas 10 3 Reactor Cell Air 10 4 Oil Lubncation Systems TRANSPORT OF MATERIALS FROM SALT TO COVER GAS SYSTEMS 11 1 Fission Products 11 2 Restrictions in the Off Gas System 11 3 Tritwum Transport in the MSRE IMPLICATIONS OF THE MSRE CHEMISTRY FOR FUTURE MOLTEN SALT REACTORS 61 61 64 64 65 66 69 71 71 74 74 79 94 99 99 101 108 110 112 112 112 116 118 121 121 122 122 123 127 128 128 131 131 132 138 139 CHEMICAL ASPECTS OF MSRE OPERATIONS R E Thoma ABSTRACT In this report are tabulated all resuits of laboratory analyses performed in surveillance of MSRE salt, water, cover gas, and o1l systems Excepted are analytical data pertaining to fission product and trittum distribution and transport The report recapitulates conclusions derived from chemical analyses performed from the 1964 preoperational test period until termination of power operations 1n 1969, modified by the results of postoperational examination of the reactor compo- nents Surveillance results were evaluated with respect to their significance as indicators of the performance of the MSRE and as indicators of the need and potential for development of specific in-line methods of analysis for molten-salt power reactors As judged from chemical data, the MSRE was highly successful as a materials demonstration The flowing salts did not wet their containment systems, fuel salt netther wetted nor penetrated the graphite moderator surface The cumulative generalized corrosion within the fuel circurt resulted in the removal of chromium from the alloy to an average depth of 04 muil, while that in the coolant system was undetectably low The results of postoperational examinations, although cor- roborative of predicted corrosion, also indicated finite but shight intergranular attack In operations which successively employed 235U, 223U, and 23°Pu as sources of power 1n the reactor, the circulated fluids remained chemically stable, free of radiation damage, and free of contamination The average full-power output of the reactor, as computed from experimental results of 1sotopic dilution mass spectrometric analysis of fissile species and subsequently confirmed by cap- ture-to-absorption ratio measurements, was shown to be 7 4 MW(t) EXECUTIVE SUMMARY The Molten Salt Reactor Experiment (MSRE) was conducted durnng the period from 1965 to 1969 as the first extensive demonstration of the operability of molten-salt reactors Durnng this period, continuous survetllance of the chemical behavior in the circulating fluid salt, water, cover gas, and o1l systems was maintained through a program of laboratory analysis The principal function of the program was to ensure that the Experiment would proceed with the freedom from chemucal problems that was anticipated from the results of prior supporting research and development programs The results of chemical analyses were also used for assistance in developing operational plans for nuclear engineering experiments with the reactor In these experiments, the reactor was employed to a hmited extent as an experimental chemical facility to obtain chemical data that were not otherwise available Al of the results of laboratory analyses (except the mass of data on fission product and tritium distribution and transport) performed 1n surveillance of the MSRE salt, water, cover gas, and o1l systems are summarized 1n the current report In this archive record are recapitu lated various conclusions derived from the laboratory analyses performed from 1965 to 1969 as modified by the results of postoperational examination of the reactor components Surveillance results were evaluated with respect to their significance as indicators of the need and potential for development of specific in-line methods of analyses for molten-salt power reactors Examination of metal- graphite assemblages removed periodically from a flow channel 1n the graphite moderator confirmed inferences from chemucal data that materials compatibility was excellent This report 1s divided into chapters that pertain to the chemical behavior of flush, coolant, carrier, and fuel salts 1n the prenuclear operational period, the chenical composttion of the fuel and flush salts during nuclear operations, the results of corrosion surveillance, power estimates from chemical and 1sotopic dilution data, and the results obtained from analysis of samples from auxiliary fluid systems Experience with the MSRE throughout the shake- down periods preceding power operations confirmed that the molten fluoride salt mixtures were intrinsically noncorrosive to Hastelloy N and that effective proce- dures were employed to prevent serious contamination of the salt circuits during this period Upon mitiatton of power operations, orifices in the fuel off-gas system became restricted Investigation showed that the cause was organic material o1l that seeped nto the fuel pump Improved filters successfully alleviated the plugging problem, but the continued passage of hydrocarbons through the pump constituted a chemucal factor that introduced a degree of un- certainty to some interpretations of chemical behavior in the MSRE In operations which successively employed 23°U, 233y, and 2??Pu as sources of power in the reactor, the circulated fuel salt remained chemaically stable, free of radiation damage, and essentially free of contami- nation. While chemical data alone were useful as statistical indicators of trends in the concentration of fissile species 1n the fuel salt over extended penods of power operation, they were of secondary importance 1n day-to-day operations because on-site reactivity balance measurements proved to be some ten times more sensitive to changes 1n the concentration of fissile material than the mdividual chemical results The combined results of chemical and mass spectrometric analysis, however, furnished information that was uniquely suited to use in establishing numerous abolute values and were applicable for determination of trends m performance, in computation of inventory, and in establishing the distnbution of uranmwum between the fuel- and flush-salt systems After three years of operations with 2352380 fyel salt, uranium was removed from the carnier salt in preparation for tests of **>*U as a fuel for molten-sait reactors Uranmium was removed from the carrier salt by fluorination, the uranwm hexafluoride product was absorbed on NaF beds Recovery of the uramum from the NaF absorber beds yielded less uramum than expected A painstaking investigation was made which led to a refinement in the material balances of fissionable species in the reactor from the outset of operations From these 1t was deduced that the dis- parity was caused by retention of 0.8 kg of 233U (2 48 kg ZU) 1n the chemical reprocessing facility at the MSRE. Fuel salt was prevented from becoming increasingly oxidizing as burnup of fissionable material proceeded by the addition of small amounts of beryllium metal to the salt flowing through the pump bowl. The results of these experiments showed that the disposition of > Nb between the containment materials and the salt could be used as an indicator of the freedom from our development of a potentially oxidizing condition 1n the fuel salt. Disposal of gaseous tritium emanating from the MSRE posed no radiological hazard, consequently, no program for completely defining 1ts distribution was instituted at the outset of operations After recognition of the importance of tritium control 1n large molten-salt reactors, studies of tritium in the MSRE were actively pursued. Results of these studies are described in other MSRP reports and are not treated extensively here. As judged from chemical data, the MSRE was highly successful as a materials demonstration. The flowing salts did not wet their containment systems, fuel salt neither wetted nor penetrated the graphite moderator surface. Chemical analyses showed corrosion with the 2L1F-BeF, coolant system to be neghgible. This has been borne out by subsequent examination of the salt side of the tubes from the air-cooled radiator and of the coolant side of the primary heat exchanger. Sumilar but more numerous analyses suggested that corrosion with- i the fuel system was slight (but observable). The cumulative generalized corrosion within the fuel circuit resulted 1n the removal of chromium (the most chemi- cally active constituent of the alloy) from an average depth of 0.4 mul, some ten tumes less than was anticipated from the preoperational laboratory mea- surements of self-diffusion coefficients of chromium n Hastelloy N. It 1s inferred that the major fraction of this corrosion resulted from interactions of atmospheric oxygen retained in the graphite moderator after periods of reactor maintenance. Postoperational examination by metallographic tech- niques confirmed the low generalized corrosion but disclosed a grain-boundary effect near surfaces exposed to the fuel which resulted 1n cracks to a depth of one grain 1n strained specimens. This hitherto unobserved phenomenon 1s currently being mvestigated. A salient conclusion from the chemical studies de- scribed 1n the current report 1s that development of automated 1n-line methods for determination of redox potential ([U**]/[ZU]) of fuel salts, for dynamic assessment of corrosion rates, and for measurement of the presence of oxides at low (<50 ppm) concen- trations 1n flowing salt will be required for operation of larger reactors Operation of the MSRE served to demonstrate the practicality of the molten-salt reactor concept, its safety, reliabihty, and tractability to simple mainte- nance methods, These operations confirmed that the molten fluonide salts are immune to radiation damage, equally serviceable with varnous fissie species as energy sources, and tolerant of the buildup of fission and corrosion products The MSRE thus fulfilled its role and demonstrated chemical compatibility of matenals, stmple refueling and reprocessing of salts, and the potential need for automated in-line analysis as part of the operational controls system in molten-salt reactors. ACKNOWLEDGMENTS The data recorded in this document were obtamed through the efforts of a large number of people in several Divisions of the Oak Ridge National Laboratory: Analytical Chemistry, Chemistry, Chemical Tech- nology, Metals and Ceramics, Operations, Reactor, and the Solid State Divsions. We are especially indebted to the staff of the Analytical Chemistry Division, for only through the enthusiastic commitments of these chemists to the needs of the Molten-Salt Reactor Program was it possible to obtain the excellent quality of chemical information pertaining to MSRE operations that we now have. The efforts of this group were led by R. F. Apple, R. E. Eby, J. M. Dale, C. E. Lamb, A. S. Meyer, Jr., and W. F. Vaughan, with the administrative support of L. T. Corbin, I. C. White, and M. T. Kelley. We were assisted as well by E. M. King and his associates in hot-cell examinations of specimens from the MSRE. Counsel and advice was provided throughout the period of reactor operations by members of the MSR program staff, =~ M. W. Rosenthal, R. B. Briggs, P. N. Haubenreich, J.R. Engel, R.H. Guymon, and B. E. Prince, and by the chemists who have been my associates in the program. Of this group W. R. Grimes, E. G. Bohlmann, F. F. Blankenship, and H. F. McDuftfie supplied special assistance and provided significant contributions to the conclusions offered here, KEY-WORD INDEX *analytical chemistry + *burnup + *chemical properties + *chemistry + *corrosion products + *fluorides + *fuels + *fused salts + *MSRE + *power measurement + *primary salt + *reactors + *sampling + *secondary salts + *single-fluid reactors + *surveillance + *uranium fluorides + *uranium-233 + *uranium-235 + beryllium fluoride + coolants + cover gas + fuel preparation + graphite + Hastelloy N + inventories + lithium fluoride + MSRP + nickel alloys + oxides + oxidation + oxygen + phase equilibria + plutonium fluorides + primary system + reactor vessel + zirconium fluoride 1. INTRODUCTION During the last two decades a great number of reactor concepts have been proposed to fill the foreseeable need for electric power toward the end of the century and to conserve supplies of fissionable materials. Of these concepts, only a few remain of potential signifi- cance to the nuclear economy. Foremost among this group are the Liquid-Metal Fast Breeder (LMFBR), the Gas-Cooled Fast Breeder (GCFBR), the Light Water Breeder (LWBR), and the Molten-Salt Breeder (MSBR). The initial efforts to develop the molten-salt system began more than 20 years ago at the Oak Ridge National Laboratory. A detailed examination of the program which followed is described in a series of papers published early in 1970.! By 1964 development of MSR’s had culminated in the construction and operation of the Molten-Salt Reactor Experiment (MSRE) as a demonstration of the practicality of these reactors. It was designed to employ, as nearly as was feasible, the same materials that were proposed for use in molten-salt breeder reactors. Thus the MSRE was constructed to circulate uranium fuel as UF, dissolved in a molten fluoride mixture within a Hastelloy N circuit. The fuel mixture was pumped at a rate of 1200 gpm through a graphite core matrix contained in a cylindrical core vessel (Fig. 1.1). Dry, deoxygenated helium was supplied at 5 psig to the pump bowl. A flow of this gas carried xenon and krypton out of the pump bowl to charcoal beds. When the reactor was operated at full power, fuel entered the graphite core at 632°C (1170°F) and was heated to 654°C (1210°F). The salt was then dis- charged through the shell side of a tube and shell heat exchanger, returning through a fuel inlet to the reactor vessel. A coolant salt circulated through the heat exchanger, through the air-cooled radiator to the coolant pump, and back to the heat exchanger to complete the circuit. At full power, the temperature of the coolant salt varied from 546°C (1015°F) to 579°C (1075°F) in this circuit. Design parameters of the MSRE are summarized in comparison with those for larger molten-salt reactors in Table 1.1. Here it is noted that two fuel salt compositions were employed in the MSRE. The reactor was operated initially with a 235U fuel charge; the uranium from this charge was recovered and replaced with 233U for the latter period of reactor operations. Nuclear characteristics of the MSRE with its 2351 fuel charge are listed in Table 1.2. From the inception of operations with the MSRE in 1965, the performance of the MSRE was positive indication of the technical feasibility of molten-salt reactors. The MSRE has shown that a molten-fluoride reactor can be operated at temperatures above 1200°F without attack on either the metal or graphite parts of the system; that reactor equipment in the radioactive parts of the plant can be repaired or replaced; and that xenon can be stripped continuously from the fuel. Operations with the MSRE were terminated as plan- ned late in 1969. The reactor was operated for a cumulative period of 13,172 equivalent full power hours during the period of nuclear operations from June 1, 1965, to December 12, 1969. A chronological ORNL-LR-DWG 6i09TRIA FLEXIBLE CONDUIT TO CONTROL ROD DRIVES GRAPHITE SAMPLE ACCESS PORT COOLING AIR LINES ACCESS PORT COOLING JACKETS ha FUEL OUTLET REACTOR ACCESS PORT SMALL GRAPHITE SAMPLES - HOLD-DOWN ROD OUTLET STRAINER CORE ROD THIMBLES LARGE GRAPHITE SAMPLES CORE CENTERING GRID FLOW DISTRIBUTOR - VOLUTE GRAPHITE - MODERATOR \ STRINGER FUEL INLET / _ T~- CORE WALL COOLING ANNULUS REACTOR CORE CAN — REACTOR VESSEL —1 ANTI-SWIRL VANES - VESSEL DRAIN LINE - MODERATOR SUPPORT GRID Fig. 1.1. MSRE reactox vessel. Table 1.1. Comparnson of MSRE, MSBE, and MSBR design data’ % MSRE MSBE MSBR Reactor power, MW( 1) 73 150 2250 Peak graphrie damage flux 3x 1013 5% 1014 3x 1014 (E, > 50 keV), neutrons cm 2 sec ™! Peak power density, W/cc Primary salt 30 760 500 Core including graphite 66 114 65 Peak neutron heating in 02 26 17 graphite, W/cc Peak gamma heating 1n 07 63 47 graphite, W/ce Primary salt Volume fraction 1n core 0225 015 013 Composttion, mole % TLiF 65 (64 5)° BeF, 29 2(302) 715 717 ThE,4 0(0) 16 16 235 238yp, 0 83 (0) 12 12 233UF, 0 (0 14) 05 03 Zr¥, 502 None None ~ Liquidus, °C 434 500 500 Liquidus, °F 813 932 932 Density, Ib/ft3 at 1100°F 141 211° 210 Viscosity, Ib 171 hr ™! at 1100°F 19 29¢ 29 - Heat capacity, Btu/1b°F 047 032 032 Thermal conductivity, 83 075 075 Btu hr™? £t oE ! Volumetric heat capacity, 66 66 66 Btu ft ™3 °F ! Temperature, °F Inlet reactor vessel 1170 1050 1050 Outlet reactor vessel 1210 1300 1300 Cuculating primary salt volume, ft> 70 266 1720 Inventory fisstle, kg 76 (32)b 396° 1470 Power density primary salt 4 20 46 cairculating average, W/cc Y rom J R McWherter, Molten Salt Breeder Experiment Design Bases, ORNL-TM 3177, p 3 (November 1970) Figures n parentheses refer to the second fuel loading, contamning 223 UF, €206 at 1300°F, 212 at 1050°F 916 4 at 1300°F, 34 2 at 1050°F €2334 1taal Table 1.2. Nuclear charactenstics of MSRE with 235U fuel Thermal neutron fluxes,a neutrons cm "2 sec”! Maximum 379 x 1013 Average in graphite moderated regions 148 x 1013 Average n circulating fuel 474X 1012 Reactivity coeffiaentsb Temperature, (°F) ! 77% 107 235y concentration 0253 Fuel salt density 023 Graphite densty 053 Prompt neutron lifetime, sec 28x%x 107 At operating fuel concentration, 7 4 MW bAt mafial cmfical concentration Where umits are shown, coefficients for variable x are of the form (1/k)/(8k/ax), other coeffictents are of the form (x/k)/(8k/ax) history of reactor operations 1s summarized in Fig 12 Detailed accounts of these operations are described 1n the Molten-Salt Reactor Program semmiannual progress reports In operation, the MSRE employed three salt mix- tures fuel, coolant, and flush salt (that was used to scavenge mpurities from the fuel circuit and from surfaces of the graphite moderator before and after the fuel containment system was opened) Fuel and coolant salts for use in the MSRE were selected on the basis of considerations which are discussed 1n detail 1n an earlier report > The fuel salt consisted essentially of a carrier muxture into which suitable amounts of fissile material could be dissolved to produce fuel salt The carrier was selected to be a maxture of 7LiF BeF,-ZrF, such as to provide the optimum physical and chemical properties of the fuel salt Some of the criteria included 1n optimuzing the carrier composition were liqudus temperature, vis- costty, and zircomum (included to ensure that UO, could not be precipitated from the molten fluoride solution) concentration The phase diagram of the LiF-BeF,-ZrF, system®** shown in Fig 1 3 illustrates the options available for choice of carrier salt The salt compostiron selected on the basis of the considerations mentioned was ' LiF BeF,-Z1F, (65-30-5 mole %) An additional measure was adopted 1n the selection of salt composition of the fuel to minimize the possibility that troublesome deposits containing fissionable ma- terial might segregate from the molten-fluoride fuel solution This was the choice to constitute the uranium fuel charge of about two-thirds 2°®U and one-third 235U, 1t was based on chemical considerations, and arose from uncertainty as to the probable value of the oxidation-reduction potential that would prevail n the fuel salt 1n normal operations From Long and Blanken- ship’s* results, it was concluded that the dispropor- tionation of UF, would not proceed to the extent that the amount of metallic uranmmim produced would precipitate or cause problems by alloying with the fuel salt containment system If, for some reason, however, the [U*}/[ZU] concentration ratio were to rise above 50%, formation of uranmum alloys and carbides was foreseen as possible This difficulty was recognized 1n the 1nexactness of our information concerning the value of the average total cation-anion balance that would result from one fission event in the reactor environ- ment If the tendency was toward a slight excess of cations, the potential for reduction of U* — U3" and disproportionation was increased Increasing the total inventory of uramum would reduce proportionately the rate of development of unfavorable {U**]/[ZU] con- centration ratios For these reasons, the choice was made to specify that the concentration of uranium in the fuel salt would be 09 mole %, even though <0 3 mole % of highly enniched uranmium would have been sufficient to make the MSRE cntical Such measures, 1t was shown, were overly conservative in affording protection which 1t 1s recognized now was not essential Notwithstanding, they, like the meticulous operational methods which were employed, comprse the margins of safety which were appropriate to the experiment Under simlar considerations, the coolant salt was chosen as a muxture corresponding to the fuel composi- tion but contamning neither fissile material nor zur- conum Salt of the composiion ’LiF-BeF, (66-34 mole %) was used both as the coolant and as the flush salt Samples of the MSRE fuel mixture and (less fre- quently) the coolant mixture were analyzed routinely durning all periods when salts were circulated m the reactor On each occasion of 1ts use, the flush salt also was analyzed The concentrations of the salt constit- uents, oxide contaminants, and fission product species were momtored on a continung basis Chemical analyses were performed regularly with samples re- moved from the circulating salts in order to evaluate the utility of a contmuous surveillance program as well as to fix goals for in-line analytical controls for future molten-salt reactors The MSRE provided the imtial experience 1n these respects, although a molten-salt reactor, the Aircraft Reactor Experniment,® previously demonstrated the operability of molten-salt reactors, the scheduled period of its operation was brief and did not include a program of chemical analysis For the MSRE, however, we sought to demonstrate through a long period of operation the stability of such reactors, SALT IN FUEL LOOP POWER DYNAMICS TESTS INVESTIGATE OFFGAS PLUGGING REPLACE VALVES AND FILTERS RAISE POWER REPAIR SAMPLER ATTAIN FULL POWER CHECK CONTAINMENT g gy T ST FULL - POWER RUN -— MAIN BLOWER FAILURE REPLACE MAIN BLOWER MELT SALT FROM GAS LINES REPLACE CORE SAMPLES TEST CONTAINMENT RUN WITH ONE BLOWER INSTALL SECOND BLOWER - ROD OUT OFFGAS LINE CHECK CONTAINMENT 30—dey RUN AT FULL POWER } REPLACE AIR LINE DISCONNECTS SUSTAINED OPERATION AT HIGH POWER REPLACE CORE SAMPLES TEST CONTAINMENT } REPAIR SAMPLER SALT IN FUEL LOOP POWER J . erchan et P g S e g L. e gt ORNL-DWG 69— 7293R2 XENON STRIPPING EXPERIMENTS MAINTENANCE } INSPECTION AND REPLACE CORE SAMPLES TEST AND MODIFY FLUORINE DiSPOSAL SYSTEM PROCESS FLUSH SALT PROCESS FUEL SALT LOAD URANIUM-233 REMOVE LOADING DEVICE 233, 7ERO - POWER ' PHYSICS EXPERIMENTS INVESTIGATE FUEL SALT BEHAVIOR CLEAR OFFGAS LINES REPAIR SAMPLER AND CONTROL ROD DRIVE 233 DYNAMICS TESTS INVESTIGATE GAS N FUEL LOOP HIGH-POWER OPERATION T0 MEASURE 233y 4 /o, REPLACE CORE SAMPLES REPAIR ROD DRIVES CLEAR OFFGAS LLINES INVESTIGATE COVER GAS, XENON, AND FISSION PRODUCT BEHAVIOR ADD PLUTONIUM IRRADIATE ENCAPSULATED U MAP F.P DEPOSITION WITH GAMMA SPECTROMETER MEASURE TRITIUM, SAMPLE FUEL REMOVE CORE ARRAY PUT REACTOR IN STANDBY 0 2 4 6 8 0 Fuel S5 POWER (Mw) ruee NS POWER (Mw) FLusH 1 Flusd 1 Fig. 1.2. Chronological outline of MSRE operations. PRIMARY PHASE AREAS: ® uF LigBefgZrfg © Li,BeE, @ LigZrig ® LizZrF, ®) LizZr,F Bef, © ® zrf, 80Q LigZryfg . ~- 520\ £-507-, \‘ré ORNL-DWG €6—7321R3 TEMPERATURES IN °C COMPOSITION IN moie % 2-LIQUIDS of fused fluorides having the composttion ' LiF BeF, ZrF, UF,; (65029250083 mole %), fissionable 22°U comprsed about one third of the urantum nventory To provide for an orderly approach to cntical operation of the reactor and to facilitate fuel preparation, the fuel was produced from the three salt mixtures described above. The enriched fuel concentrate mixture, m which all 233U was combined with "LiF as UF, 93% enriched in 235U, to form the bmary eutectic mixture (27 mole % UF, ), was prepared in six small batches (15 kg of #3°U each) for nuclear safety and for planned incremental additions to the reactor fuel system. The balance of the uranium required for the fuel was provided as a chemically wdentical muxture with UF; depleted of 2°°U. The third component mxture, the barren fuel solvent, conststed of the remamning constituents of the reactor fuel and had the chemucal composition ” LiF-BeF,-ZrF, (64.7-30 1-5.2 mole %). The reactor fuel for the zero-power experiments was produced subsequently by adding small increments of "LiF-235UF, into this carrier salt m the drain tanks and finally into the pump bowl The composition of the salt was fixed at this point by the amount of 2**UF, required for criticahty to be sustained with one control rod completely mserted. The addition of the enriched fuel concentrate mixture to the MSRE to within 1 kg of 233U of cnticality was accomplished during the latter part of May 1965. The first major addition of enriched fuel concentrate con- sisted of the transfer of about 44.17 kg of 23%U from three containers directly into the fuel drain tank Three subsequent additions of 2?3 U to the reactor drain tank increased 1ts 235U inventory to 59.35, 64.42, and finally 68.76 kg The transfer of less than batch-size quantities of 225U was made by inserting the salt transfer hne to a predetermuned depth in the batch contamer. It was anticipated that the composition of the fuel at this point would be 7 LiF-BeF,-ZrF,-UF, (65 0-29.17- 5.0-0.83 mole %), the composition calculated from the weights of the carrier and enniching salts added to the reactor was changed steadily throughout the zero-power experiments as capsules of the enriching salt were added to the pump bowl. Complete results of all analyses performed with MSRE salts during the precritical and zero-power experiments are summarized in the following sections of this chapter. 2.2 Flush Salt Part of the LiF-BeF, mixture which was furnished for use m the MSRE was employed to flush the fuel system imtially and later, on occasions before and after maintenance periods 1if 1t were likely that atmospheric contaminants could have entered the system. Samples of each of the batches of "LiF-BeF, mixtures for use as 10 flush and coolant salts were analyzed 1n the facilities of the ORNL Analytical Chemistry Division prior to use in the MSRE. Analytical results for the 61 batches of salt used as flush salt and primary coolant are listed in Table 2 1. Although the nominal composition was ’ LiF-BeF, (63-34 mole %), the average composition as determined by chemical analysis was " LiF-BeF, (63.56 = 0.005— 36.44 * 0.005 mole %). Examinations of the material balance of the production operations indicate that the disparity 1n composition was due to an analytical bias. Efforts to deterrmine whether such a bias actually existed were made by the analytical chemists, but were unfruitful. Before salt was charged into the reactor, the dramn tanks and salt loop systems were carefully dried by evacuating and purging with dry helum. A detailed description of the procedures employed 1s given else- where.? Thereafter, late in 1964, approxmmately 4187 kg of flush salt was charged into the fuel drain tank; it was then transferred among the tanks in the drain tank cell. These operations served to calibrate the weighing devices, to check elevations and volumes, and to establish the operating requirements of the freeze valves. Throughout the prenuclear test period, speci mens of the circulating flush salt were obtamned for chemical and spectrochemical analysis. The results of spectrochemical analysis are listed in Table 2.2, chemi- cal analyses are listed in Table 2.3. Spectrochemical data were of principal value to ensure that during the early stages of operation the salt charge was free from any unexpected cationic contaminants in the contain- ment system. These analyses were very useful 1n that they provided the assurance required, but with ex- tended experience with the reactor they became of less significance and were discontinued. The data shown in Table 2.3 suggest that the operations connected with transfers of the flush salt among the storage tanks in the drain tank cells 1n November and December of 1964 did not introduce appreciable amounts of structural metal contaminants mmto the salt, the analytical results for 1ron suggest, rather, that since its apparent concentration in the salt samples removed from the drain tank was lower by about one standard deviation than the average concen- tration in the salt before use and in use, 1t was present in the flush salt mitially as metallic particulates which were precipitated from the salt during this period. Little or no change 1s evident 1n the chromium concentration before circulation of the flush salt, the data in Table 2.3 can be interpreted, however, to indicate that after completion of the flushing operations in January— March 1965, the chromium analyses represent an 11 Table 2.1. Chemical analyses of LiF-BeF, (66-34 mole %) produced for use as MSRE flush and coolant salts 7 Batch Net wt. Li Analyses No. of salt Assay Li Be F Cr Ni (kg) (wt %) (ppm) 116 117.3 99.992 13.76 9.61 77.3 18 14 204 117 119.3 99.992 13.77 9.75 77.5 32 14 171 121 119.8 99.992 13.29 9.51 77.2 62 6 125 125 117.8 99.991 13.20 9.59 77.3 16 5 35 126 119.8 99.992 13.27 9.69 76.9 15 32 103 127 119.6 99.992 12.98 9.86 76.8 7 12 123 128 119.6 99.992 13.06 9.81 76.9 9 45 112 129 117.1 99.992 13.20 9.53 77.2 <5 <5 97 131%* 117.2 99.991 13.30 9.67 77.0 14 31 104 132 117.7 99.992 13.00 9.78 77.1 <5 18 76 133 124.8 99.992 12.41 9.30 76.3 8 79 i6l 134 126.1 39.991 12.80 9.84 77.3 6 72 151 135 120.4 99.991 12.90 9.82 77.5 7 21 123 136 120.0 99.991 12,20 9.97 77.2 8 11 180 137 119.4 99.991 12.76 9.49 77.8 ) 10 72 138 119.2 99.991 i2.70 9.90 77.0 15 23 85 139 105.5 99.991 12,91 9.78 76.2 6 66 119 140 125.3 99.991 12.70 g.81 77.1 21 91 130 141 117.7 99.991 12.70 9.86 77.4 37 36 138 142 111.0 99.991 12.90 10.00 77.2 23 5 130 143 124.4 99,993 12.70 9.97 77.3 19 34 79 l44% 118.1 99.994 12.90 9.98 77.2 14 56 110 145 119.8 99.994 12,80 9.96 77.0 15 <5 142 146 119.8 99.994 13.00 10.00 77.4 23 134 135 147 118.8 99.994 12.90 10.00 77.1 18 107 117 148 119.1 99.993 12.90 10.10 77.3 24 12 119 149 117.5 99.993 12.80 9.89 77.4 i3 105 166 150 118.7 99.993 13.20 9.70 76.7 18 34 86 151 124.6 99.993 12.90 G.82 77.4 13 12 143 152 112.3 99.994 12.50 9.62 77.4 9 27 172 153 125.6 99.994 13,00 9.84 76.7 17 31 132 154 95.9 99.994 11.10 9.55 77.2 26 47 106 155 128.1 99.994 13.90 8.92 76.2 27 91 117 156 119.1 99.993 12.70 9.55 77.3 17 28 92 157 118.8 99.993 13.20 9,80 77.0 11 29 23 158 119.4 99.993 13.20 9.80 77.1 17 29 124 159 119.2 99.993 13.10 9.74 76.9 13 108 163 160% 119.6 99.994 13.20 9.77 77.0 13 11 99 161 118.9 99.994 13.20 9.80 77.2 8 i8 75 * Production Excess. Table 2.2. Spectrochemical analysis of MSRE flush salt during mtial use Sample Impunity concentration? No Al B Bi Ca Cu Mg Mn Na Pb S1 Sn Zr FD-2-20 FD-2-3 FD-2-4 FD-2-5 FP-1¢ FP-2 FP-3 FP4 FP-§ FP-6 FP-7 FP-8 FP-9 FP-10 FP-11 FP-12 FST-14 FST-2 FST-3 FD2-7 FD2 8 FD2-9 o W WmW W o0 mmmmwo OO0 N AN A mwWmOOOOOOCOO0OO0000O0O0n oloRoNeNoNoRoNaloReoNoNoNaNONONO NGRS Ne! >A >A >A Wm W SRR R W o s e o e e~ e e - e e 2 4Ly and Be omitted A = 100 to 1500 ppm, B = 10 to 100 ppm,C=1i0 10 ppm bED designates samples removed from the fuel drain tank CFP designates samples removed from the fuel pump bowl dEsT designates samples removed from the fuel storage tank 12 average final value of 60 ppm If 1t 1s assumed that the increase resulted from scavenging of moisture or oxides from the system, one would anticipate a corresponding increase of 12 ppm of oxygen in the salt (see Sect 6 1) Results of oxygen analyses of the flush salt exhibited perplexing vanance The larger values are believed to reflect adsorption of moisture inadvertently introduced through the handling procedures after samples were removed from the reactor If the oxygen concentrations were representative of either contamination of the salt by water or oxidation of the container alloy by atmospheric oxygen, the occurrence of chemical attack on the Hastelloy N walls of the circuit should have been reflected 1n increased concentrations of chromium in the salt For example, if an increase of oxide concentra- tion m the order of 200 ppm were to have been caused by such reactions, an equvalent increase 1in chromium concentration of 60 ppm should have been observed To test whether a large bias might have been mnvolved i the fluorination assay of oxygen concentration (the KBrF, method), large (50 g) samples of the salt were obtained and analyzed by a newly developed method which, like the punfication procedures, depends on the removal of water by sparging the molten salt with an H,-HF stream For the three samples that were ana lyzed 1n this manner, the average concentration found was 75 ppm After the flush salt was drained from the reactor on completion of the tests in which 1t was used, the flush Table 2.3. Composition of MSRE flush salt in prenuclear operations Weight percent Parts per milhion Date Sample U 11 Be Zr = Ie Cr Ni Pu o4 Book Analytical Charge salt®? 1268 962 11/30/64 DC-1 1312 968 12/8/64 DT-1 12/11/64 DT?2 12/15/64 FD-2-3 1/12/65 FD-2-5 1/12/65 FP-1 1/13/65 FP-2 1365 983 1/14/65 FP3 1/16/65 P4 1355 935 1/18/65 | ] 1/20/65 FP-6 1350 996 1/23/65 FP-7 1365 946 2/3/65 FP-8 2/3/65 I'pP-9 1335 998 2/11/65 FP-10 949 2/23/65 FP-11 3/4/65 FP-12 Av 1347 968 76 51 99 01 137 17 34 7708 99 94 45 45 7 432 59 16 24 419 48 18 <3 390 44 22 49 555 140 24 10 35 80 52 104 62 i <10 33 56 46¢ 79 34 102 27 212 62 30 74 72¢ 80 07 103 56 125 <10 <20 180 77 85 101 00 180 54 <20 150 106¢ 75 80 9917 210 57 <20 142 75 05 128 60 <20 15604 No analyses performed 144 7796 10166 118 38 22 4K BrF4 method unless otherwise noted bAverage for 61 batches CHF purge method dSample stored 48 hr m capped plastic container before analysis salt was reprocessed > The salt was transferred by gas pressure to the fuel storage tank in the fuel reprocessing system and sparged with a mixture of H, and HF It was concluded from measurements of the HF concen- tration 1 the off-gas stream that the reprocessing operations were effective in removing 115 ppm of oxide from the salt > The oxide concentration should have been reduced by this process to the same concentration which resulted from preparation and purification of the salt, ~55 ppm®* The results of chemucal analysis indicated that the chromium concentration of the flush salt increased by 43 ppm 1in use, corresponding to an increase of 180 g of chromium, and equivalent to 55 5 g of oxide, or to an increase of 13 ppm in oxide concentration The H,-HF analytical data indicated that the average concentration of oxide 1n the {lush salt during 1ts use was 75 ppm, or that the oxide concentra- tion of the salt increased by 20 ppm mn use This increase, while shightly greater than would have been anticipated from the above discussion, may be ration- alized by the following considerations An indetermui- nate fraction of the oxide removed in reprocessing was introduced from the oxide scales present on the surfaces of the chemical reprocessing facility containers, the fuel storage tank, and the walls of the fuel circunt Prior purge of the system with dry 1nert gas was performed before the facility was used for reprocessing the flush salt, but the system was not previously flushed with a scavenger salt It might therefore have been anticipated that the oxide concentration of the salt i the reprocessing system, after its initial use, would yield a hagher concentration of oxide than that removed from the salt in the fuel circulation system Thus, if 1t 1s assumed that in the reprocessing experiment ~0 4 kg of oxide scale was removed f{rom the system, good agreement exists among the analytical data, except for measurement of oxide concentration by the KBrF, method One potential use of the flush salt that had been anticipated was 1o remove any oxides that might deposit 1n the heat exchanger 1f the fuel salt were to become seriously contaminated Concurrent with the beginning of MSRE operations, chemical data ap- peared® which showed that the solubihity of the least soluble oxide i the LiF-BeF,-ZrF, carner salt was some three times that in LiF-Bel, mixtures It was thus necessary to regard the flush salt as an agent which might have limited capacity to scavenge oxide impu- rities from the fuel circuit unless, if saturated, 1t were reprocessed Its principal function after imitial use was to remove residues contaiung uranium fluoride from the fuel circuit after 1t was drained Concurrently, the 13 capacity of the flush salt to contamn dissolved oxides increased with each use in the reactor, because the fuel restdues carried into the flush salt added about 2 5 kg of zircontum to the salt each time 1t was used It would probably have questionable value 1f 1t were expedient to dissolve precipitated oxide deposits from the heat exchanger This did not preclude the use of LiF-BeF, flush salt for such application, because at 650°C the solubility of the saturating oxide phase, BeO, 1n the salt was known to be ~200 ppm, corresponding to a capacity of the flush salt to dissolve and retain 1 135 kg of ZrO, 1n solution at this temperature Successive repurtfication and removal of oxides that were picked up 1n the salt would permit it to be used to clear the heat exchanger of an oxide plug if necessary No such requirement developed A preferable alternative was also available if needed The increased solubility which small amounts of zircomum fluonde provide, increasing the oxide solubility to 800 ppm for a few mole percent of ZrF,, showed that if 1t were expedient to do so, the composition of the flush salt could have been changed simply by the addition of "LiF ZrF,; so as to enhance its solvent capacity for precipitated oxides Operation with the flush salt also shed lLight on another aspect of the use of molten fluorides mn a arculating system, namely, the appearance of salt (or salt constituents) in the cover gas Analysis of the hehhum cover gas by Million and Pappas (see Sect 2 4 5) indicated the presence of fluorides m the fuel off gas, but this probably represented salt droplets rather than any decomposition product The principal difference in the design of the fuel and coolant pumps 13 that the fuel pump contains a spray nng which functions to remove fission product gases from the fuel The major salt flow into the pump tank 1s through a bypass that s taken from the volute discharge line into a toroidal spray ring in the upper part of the pump bowl From there the salt sprays out through two rows of holes and mmpinges on the sait surface in the tank to prowvide gas-liquid contact for gas stripping ¢ This difference 1n design of the two pumps seems to account for the tact that it was occasionally necessary to remove solids from filters, valves, and hines n the fuel system, but rarely necessary with the coolant system Material recovered from the coolant off.gas system showed only traces of salt constituents, but the fuel off-gas solids generally contained minute beads of frozen sait One experiment conducted dunng the precritical operation pertod consisted of krypton stripping tests For these tests, a special mnsert was installed into the side of the off-gas flange nearest to the fuel pump bowl 14 PHOTO 4863 -71 oy Fig. 2.1. Fluonde glass beads removed from MSRE off-gas line after precritical experiment PC-1. Later, the flange was opened to remove the insert, inspection showed that small amounts of small glassy beads were deposited between the flange faces Toward the end of the first set of experiments with flush salt, fuel pressure control became erratic (see Chap 11) After the salt was drained from the circuit in March 1965, the gas pressure control valve far down- stream at the vent house was found to be partially plugged The valve body was rinsed with acetone, which was found to contain beads (1 to 5 y n diameter) of a glassy material After a week of carner salt circulation 1n May 1965, the small control valve again began to plug This time, it was removed and cut open for examination A deposit on the stem was found to be about 20% amorphous carbon, and the remainder was 1- to 5-u beads having the composition of the flush salt A photograph of these beads, obtamned with the petrographic mucro- scope, 1s shown in Fig 2 1 The glass beads were found to have a refractive index of 1315 As computed by interpolation of the refractive indices of the crystalline components LiF and BeF,, its composition would be LiF-BeF; (54-45 mole %) Chemical analysis showed, however, that 1ts actual composition was LiF-BeF, (66-34 mole %) The anomalous refractive index of the glass results from tts glassy character, that s, the hiquid phase at LiF-BeF, (66-34 mole %) 1s of significantly lower density than the crystalline solids of this COMmpo- sition The beads were glass rather than crystalline, mplying rapid cooling of the molten-salt mist The carbon was presumably soot from oil that had been thermally decomposed in the pump bowl 2 3 Coolant Salt The coolant salt mixtures for use in the MSRE were prepared as described elsewhere ' Twenty-four batches of the salt muxture were allocated for use as the coolant The composition and purity of these mixtures are listed 1n Table 24 A charge of 2610 kg of "LiF-BeF, (66-34 mole %) was delivered i molten form to the coolant drain tank in October 1964 The coolant system was preheated and purged of moisture in a similar manner to the fuel system, but was not Table 2.4, Fluoride production for MSRE coolant salt mixture LiF-BeF, (66-34 mole %) Chenncal analyses 7 Batch of salt L1 Weight percent Parts per No assay million (kg) | ] Be F Cr Nt Fe 101 118.1 99991 1368 943 769 8 32 133 102 116 2 99.991 14.23 9,14 772 5 8 175 103 121.2 99991 1381 932 767 10 58 182 104 1195 99990 13.68 982 772 17 11 208 105 113.2 99.990 1283 957 771 27 45 216 106 1190 99.990 1343 966 772 11 68 309 i07 1150 99990 1312 958 769 10 18 94 108 1145 99990 12.81 965 773 T 13 246 109 1199 99990 1373 927 770 8 5 62 110 1279 99.990 1315 957 766 19 12 94 111 121.2 99990 1399 921 765 13 8 142 112 1200 99991 13.81 947 770 11 48 218 i13 117.6 99.991 13.80 9.53 77.0 <5 2§ 212 114 1209 99991 1314 979 770 9 22 196 115 120.8 99.991 1374 954 774 14 5 131 118 125.0 99991 1306 991 773 9 20 134 119 114.5 99,991 12.41 10.16 769 10 18 182 120 125.5 99.991 13.65 9.52 771 8§ 35 127 122 117.8 99991 1326 9.48 773 39 70 126 123 117.9 99991 1337 948 769 48 10 113 124 118.5 99.991 1335 945 771 86 14 154 130 119.9 99,990 13,10 955 774 40 13 188 preflushed with molten salt. The system was filled and drained 18 times durning the time the MSRE was operated. The chemical behavior of the coolant salt was not noticeably different during prenuclear operations from those that followed, and will be described 1n detail later in this report. 2.4 Fuel Salt 2.41 Onssite preparation. After circulation in the reactor in March 1965, the flush salt was dramed to fuel drain tank FD-1. The fuel salt was then constituted within the reactor. The "LiF-BeF,-ZrF, fuel carrier mixture was charged into fuel drain tank FD-2, starting April 21, 1965. On-site records show that 4558.1 kg (the contents of 35 shipping containers) of this carrier salt was melted and transferred to the dramn tank; to this was added 236.2 kg of "LiF-*3®UF,, contamning 147 6 kg of *3®U (depleted in 23°U) In our current review of the significance of the results of chemical and mass spectrochemical analyses, 1t becomes evident that these values are in munor error. The "LiF-BeF,-ZrF,- 238UF, as described above was circulated through the fuel curcuit for some 250 hr during precritical test PC-2. 15 Each of the two fuel dran tanks on the MSRE incorporates two pneumatic weigh cells for estimating the inventory of the salt in these tanks. Calibrations of the weigh cells in the drain tanks were made carefully during the 1mtial stages of operations. The precision of these measurements was found to be about £0.8% (£28 kg).” Continued experience with this equipment mdi- cated long-term drifts i the readings. It was necessary, therefore, to determine inventories on the basis of analytical chemical data, and on computations involving salt densifies and volumes at certamn reference points (the circulating loops filled to the pump bowls or the tanks filled to the level probe points) to eliminate the effects of extraneous forces on the indicated weight.? During the imtial fill operations, "LiF-BeF, flush salt was admutted to all parts of the fuel circutt system. On draining the reactor, flush salt remained on the pump rotor and in the freeze valves, as well as residue 1n the heat exchanger. At this point, some 35 kg of salt was unaccounted for by the weigh cell data and was presumed to have remamned in these locations. The "LiF-BeF,-ZrF,-2*3UF, salt mixture was introduced mto the same circuitry at the beginning of the PC-2 test and was, therefore, diluted slightly by flush salt. Later in this report, 1t is shown that the increased concentra- tion of uramum in the flush salt after use indicates that the salt residue left in the fuel circust system after 1t 1s drained 1s about 20 kg. Estumation of the nominal concentration of uranium 1n the fuel during precritical test PC-2 thus requires the assumption that ~20 kg of flush salt was added 1mtially to the fuel. As noted 1n Sect. 2.4.2, however, the best agreement between expected and observed concentration 1s found when it 18 assumed that a negligible amount of flush salt residue remained in the system at completion of PC-2. This assumption 1s untenable. An alternate assumption, that less carrier salt was delivered to the reactor than was credited 1n the loading operation, 1s more credible as an explanation of the apparent anomaly. Chemucal analyses were conducted durning the precriti- cal experiments for the purpose of establishing analyt- ical base lines for use in the full-power operating period. Prior to the zero-power experiment, analyses of all salt samples were conducted with conventional equipment i the facilities of the ORNL Analytical Chenustry Division. Thereafter, the fission product activity made it necessary to perform all analyses in the ORNL High-Radiation-Level Analytical Laboratory (HRLAL). The facilities of this laboratory have been described elsewhere.” In order to provide continuity which would relate the results of prenuclear operations with those that followed, specimens of the fuel salt obtained during the precritical test period were analyzed concur- rently in both facilities Mass spectrometric analyses were performed routinely throughout #33:238( operations Samples of the fuel salt were analyzed before, duning, and at the comple- tion of the ?*°U loading operation The results reflected the 1sotopic difution of 2*®U precisely and were used in tests for corroboration of the agreement between estimated and analytical values of the concen tratton of uranmum in the fuel salt and to refine estimates of the amount of depleted uranium that was charged 1nto the fuel system The composition of the fuel salt was changed steadily throughout the zero power experiments (run No 3) as capsules of the ennching salt were added to the pump bowl Compositional analysis during this perniod served, therefore, to permit evaluation of the fuel composition dynamucally rather than as a statistical base for ref- erence during the power run Most of the uramum required for enrichment was loaded i four charging operations to one of the fuel dramn tanks, FD 2 After each addition the salt was transferred to the second drain tank and back again to ensure thorough mixing The mixed salt was loaded 1nto the reactor system after each charging operation, and count rate data were taken at several salt levels in the core and with the reactor vessel full These data were compared with the barren-salt data to momtor the neutron multiphcation and to estabhsh the size of the next addition '® Extrapolation of inverse count-rate plots with the reactor vessel full showed that the loading after the fourth addition was within 0 8 kg of 235U of the critical loading when the salt was not circulating and the control rods were withdrawn to their upper hmts The remainder of the *3°U was added directly to the circulating loop with enriching capsules These were inserted into the fuel-pump bowl via the sampler-enncher to increase the loading by 85 g of uranium at a time Count rates were measured after each capsule with the fuel pump off and the control rods withdrawn The reactor became critical after the eighth capsule with the pump off, two rods fully withdrawn, and one rod poisoning 0 03 of 1ts available worth These zero-power experiments (run 3) were com- pleted on July 4, 1965, after 764 hr of circulation of the salt The fuel was then drained and mixed with the salt remaining in the drain tanks After the fuel loop was drained, 1t was filled with flush salt, which was circulated 1 3 hr, sampled, and drained Analysis of the flush salt led to the conclusion that 0 77 kg of uranium remained n the fuel circurt on completion of run No 3 16 Final preparations were made for operation of the reactor at power In December 1965, low power exper1 ments were begun As soon as containment testing was fimished, the instruments, controls, and equipment were given the check outs required prior to startup The fuel system was then heated, and flush salt was circulated for three days Samples of the flush salt were taken Then fuel was charged to the loop for tests in the reactor at various low power levels, 100 and 500 kW and finally 1 MW The experiments that were con- ducted during this time extended over a one month period, providing the single period of operation when the composition of the fuel salt remained nominally constant Samples of the fuel salt were obtained regularly during this period in a continuing effort to appraise the utility of the chemical surveillance program and in order to verify that the reactor system was in suitable condition for the mnitiation of power operations Good agreement between anticipated and observed results was found The results of chemical analyses indicated that the system had remained free of corrosion during all preliminary operations, and it seemed that the analyses would serve as a reliable measure of fuel stability and corrosion 2 42 Uranium assay Chemical analyses and on site nuclear measurements were accumulated regularly when fuel salt was circulated in the reactor in order to evaluate as many aspects of the behavior of the molten radioactive fuel salt as feasible in contact with the Hastelioy N containment system and with the graphite moderator A number of important characteristics of the reactor soon became evident For example, by the time that statistically significant populations of analyti cal chemical data were produced from the HRLAL it had become dapparent that on site neutronic measure ments of the reactivity balance were sufficiently sensi tive to detect changes of as little as 0 1% i the concentration of fissile material circulating in the fuel system ' This represented at least a factor of 10 greater sensitivity than was indicated by the statistically significant results of chemical analysis In addition, reactivity balance measurements were made dynami cally and, essentially, on a continuous basis While chemical data alone were useful as statistical indicators of trends in the concentration of uranwum over ex tended periods, the combined results of chemical and mass spectrometric analyses furnished information that was umquely suited for use in estabhishing numerous fundamental values for use in subsequent evaluations of MSRE performance, and 1n accounting for the changes in the distribution of urammum between the fuel and flush salt systems In order to compare the concentration of urantum in the fuel salt as measured analytically with expected values, we have evaluated the results of chemical analysis, mass spectrometric analysis, the weights of the fluonides as measured in the production facility, and on-site measurements of the weights of the salt con- tainers delivered to the MSRE From the weights of the enriching salts credited to the reactor and from the isotopic composttion a uranium material balance was computed Progress reports for this period of MSRE operations indicate that satisfactory agreement between nomunal and analytical values was found As the results of mass spectrometric analyses were accumulated, however, comparisons of the nominal 1sotopic composi- tion of uranium during initial loading operations at the MSRE with the results of mass spectrometric analyses disclosed that closest agreement 1s reached when the assumption 1s made that the initial "LiF-**®UF, charge contained 145 6 kg of uramum rather than 147 6 kg as credited (see below) This lesser amount of uramum, carried by 234 5 kg of "LiF-?33UF, enriching salt and 17 added to the amount of "LiF-BeF,-ZtF, carrier salt credited to the MSRE (4558 1 kg), would produce a salt mixture in which the concentration of uramum 1s 3038 wt % The average concentration observed was 3 044 wt % In the early stages of analysis, a bias of —0 8% was discovered 1n the amperometric data'? which, when used to correct the analytical results, causes the average observed concentration in precritical run No 2 to be 3068 wt % Thus the nomnal concentration of uranium 1n the fuel salt ts inexplicably lower than the averages of the observed values if the amount of carrier salt credited to the MSRE (4558 1 kg) 1s assumed to be accurate A computation of the material balance for the MSRE fuel salt in prepower tests, using this value, 1s summarnzed in Table 2 5, which shows that the analyti- cal values tend to indicate slightly greater average concentrations than nominal If, however, we assume that ~40 kg of the carrier salt was not charged into the fuel circuit, and the fuel charge during the precritical experiment was constituted from 145 60 kg of uranium Table 2.5, Material batance for MSRE fuel salt in prepower tests Inventory (kg) Uranmam concentration {wt %) U 7L1F-UF4 Totat Nomuinal Analytical Run PC-2 (av) 145 60 2345 4792 6 3038 3 068 Run 3, FP-3-1,2 Added +47 49 765 Total 193 09 3110 4869 1 3965 3998 Run 3, FP 3-3 Added +2143 34 53 Total 214 52 345 53 4903 6 4 375 4 390 Run 3, FP-3-4 Added +4 67 +7 52 Total 219 19 35305 4911 1 4 463 4 446 Loop charge 195 23 4374 3 4463 Left in drain tank 2396 536 8 4 463 Run 3-F Added +7 83 +12 74 % charge 227 02 45237 4611 Loop charge 203 06 4386 9 4 629 Loop residue -0 77 165 To drain tank 202 29 4370 4 4 629 Already in drain tank +23 96 +536 8 4 462 FD-10-3 226 25 4907 2 4611 4 648 Run 4 I loop charge 208 15 4514 6 4611 Flush residue +18 0 Loop charge 208 15 4532 6 4592 46420028 Run 5-1, drain tank 226 25 49325 2 4593 Runs 5 7 (av of 14 samples) 4629+ 0026 Run 4 (av of 22 samples) 4642+ 0028 Runs 4 7 (av of 36 samples) 4638 + 0025 and 4754.5 kg of salt, the nominal concentration of uranium should be 3.062 wt %, as compared with an analytical value of 3.068 wt %. Comparably improved agreement 1s found between the nominal and observed concentrations of uranmum for all other prepower measurements when the assumption 1s made that the above quantittes of uranium and salt were charged to the MSRE fuel system. A uranium material balance for the prepower period was computed on the basis of these values, and 1s given i Table 2.6. We deduce, therefore, that the salt charge which was used for the precritical test, PC-2, consisted of 145.6 kg ***Um a total salt charge of 4754.5 kg. The results of the chemical analyses performed with samples of the fuel salt during this test are hsted in Table 2.7 Most of the uranium required for enrichment was loaded mn four charging operations to one of the fuel drain tanks, FD-2. Samples of the fuel salt were obtamned frequently during this period 1n a continuing attempt to appraise the utiity of the chemucal surveiilance program, and to 18 verify that the reactor system was in suitable condition for the imitiation of power operation. As the concentration of uranium 1 the fuel salt was mncreased in preparation for operating the reactor at tull power, fuel-salt samples were obtamned for 1sotopic and compositional analysis after groups of three or four capsules of enriching salt were added. Changes in the 1sotopic composition of uranium in the fuel salt during mitial loading operations are summarized in Table 2.8. Compositional analyses were obtained over a period when the incremental change in the concentration of urantum 1n the fuel salt was large 1n comparison to that in the normal operating mode of the reactor. They serve, therefore, to permut evaluation of the fuel composition dynamically rather than as a statistical base for reference duning the power run. The results of these analyses are listed 1n Tables 2.9 and 2.10. As noted previously, the salt mixture used in the precritical test, PC-2, was formed from purnfied salt mixtures of the nominal compositions ” LiF-BeF,-ZrF, Table 2.6. Material balance for MSRE fuel salt in prepower tests Inventory (kg) Uranium concentration (wi %) U 7L1F-UF4 Total Nominal Analy tical Run PC-2 (av) 145 60 234.5 4752.6 3.064 3.068 Run 3, FP-3-1,2 Added +47.49 76.5 Totat 193.09 311.0 4829.1 3.998 3.998 Run 3, FP-3-3 Added +21 43 34.53 Total 214.52 345 5 4863.6 4410 4.390 Run 3, FP-34 Added +4.67 +7,52 Total 219 19 353.05 4871 1 4 500 4.446 Loop charge 195 23 4338.6 4500 Left in dramn tank 2396 532.5 4 500 Run 3-F Added +7 83 +12.74 Z charge 227.02 4883 8 4.648 Loop charge 203.06 43513 4.667 4.665 Loop residue -0.77 -165 To dram tank 202 29 43348 4.667 Already m dram tank +23 96 +532 5 4 500 FD-10-3 226.25 4867.3 4 648 4.648 Run 4-1, loop charge 208 15 44779 4 648 Flush residue +18 0 Loop charge 208.15 44959 4.630 4,642 + 0.028 Run 5-1, drain tank 226 25 4885 3 4,631 Runs 57 (av of 14 samples) 4,629 = 0.026 Run 4 (av of 22 samples) 4,642 + 0.028 Runs 4—7 (av of 36 samples) 4,638 + 0.025 19 Table 2.7. Summary of MSRE salt analyses, expeniment No. 2, fuel sait Nominal composition of salt LiF-BeF ,-Z1F 4-238 UL 4 (64 86-29 54-5 07 0 53 mole %) Weight percent Parts per million Date Sample L U Nominal z Fe Cr Ni O Analytical 5/11/65 FP-2-13¢ 1030 699 1132 3062 5/12/65 FP-2-14¢ 1020 655 1149 3062 5/12/65 kP 2-15 1014 709 1142 3062 5/13/65 TP-2-16 1050 660 1209 3062 5/13/65 ¥P-2-17 1030 654 1108 3062 5/13/65 FTP-2-18 1050 654 1181 3062 5/14/65 FP2-19 1020 616 1141 3062 5/15/65 FP-2-20 1053 664 1136 3062 5/16/65 FP 2-21 1030 638 1108 3062 5/17/65 ¥P-2-22 1020 621 1117 3062 5/18/65 kP 2-23 030 616 1123 3062 5/18/65 FP-2-24 1078 684 1118 3062 5/19/65 PP 2-25 1052 650 1136 3062 5/19/65 FP-2-26 1060 667 1127 3062 5/20/65 FP-2-27 1046 644 1128 3062 5/20/65 FP 2-28 1055 657 1105 3062 5/21/65 FP 2-29 1051 730 1111 3062 5/21/65 ¥P-2-30 1061 659 1166 3062 3 043 67 50 98 66 74 26 34 20455 3041 7190 10318 149 <50 36 3023 95 19 18 4715 3 044 71 60 10384 271 <50 33 3053 67 40 98 38 75 21 21 155 3053 7340 105 31 152 <50 43 3 061 7140 102 22 130 26 47 3023 68 20 99 76 120 23 21 660 3039 71 00 101 81 133 25 49 3052 7110 10174 221 28 55 3033 70 50 101 23 159 28 42 3014 68 60 100 42 76 23 <5 215 3014 67 40 98 40 110 24 17 3038 7140 102 98 113 24 <20 3063 67 10 98 35 27 23 14 70 2 998 72 30 103 47 123 34 <20 3053 67 70 9968 203 28 20 405 3025 719 10370 175 27 <20 95 23 165 +21 42 44 2 Analyses performed under supervision of W F Vaughan bAnalyzed by KBrF4 method “Analysis performed under supervision of C E Lamb (64 7-30 1-5 2 mole %) and "LiF-UF, (73-27 mole %) The results of chemical analyses of the individual batches of these salts are listed in Tables 2 11 and 2 12 Results of the analyses of the carrier salt indicate that its composition was ’ LiF-BeF,-ZrF, (62 25-32 47-5 28 mole %), or that the Li/Be ratio 1s skewed as noted previously 1n connection with the coolant and flush salt Whale the results from chemical analyses show similar bias 1n determination of lithum and berylhum when compared with nomunal values, they are 1n good agreement when the analyses of the salts are compared with those for the starting materials Reexamination of the material balance in the preparation of these salts indicated that dewiations from the nominal composi- tions are much less than is indicated by the analytical data '3 In order to estimate how closely the composi- tion of the fuel-salt muxture matched the design composition, 1t is necessary to compare the analytical results with a nominal composition computed from the analyses of production plant samples The composition of the salt mixture used 1n the precritical run compares as shown in Table 2 13 with the nonmunal composition which would result from (1) combining the carrer salt and L1F-2*#UF, of the design compositions and (2) combining compositions as indicated from chemical analysis In the foregomng discussion, we have estimated the quantities of carrier salt and LiF-UF,; mixtures used to constitute the MSRE fuel for power tests The chemical compostition of the salt circulated in the fuel circuit at various stages during the zero power and in all prenuclear operations 1s compared with the results of chemical analyses in Tables 2 14 and 2 15 The values listed 1n these tables comncide within the precision limits of the analytical chemical results, the mass spec trometric methods, and the estimates of the physical properties of these fluids We have therefore adopted the nominal composttions hsted in Table 215 as reference compositions for appraisal of results through- out the remainder of MSRE operations 2.4.3 Structural metal impurities. One of the features which originally suggested the potential feasibiity of molten-salt reactors was the recognition that if the physical and mechanical properties of any one of several mickel-based alloys, such as those composed principally of Ni-Mo-Cr-Fe, were suitable for reactor construction, the alloys would almost certainly be Table 2.8. Isotopic composition of uranium in the MSRE fuel salt during initial loading operations — Loop Inv. 234, 235, 236, 238, 234, 235, 236, 238, (kg) (kg) (kg) (wt %) LiF-2 50, Loading 145.60 - 0.323 - 145.277 LiF-“77UF, to 5/25/65 47.49 0.452 44,143 0.187 2.708 Sample No. FP3-2 193.090 0.452 44,466 0.187 147.985 0.234 23.029 0.097 76.640% 23 0.234 22,501 0.099 77.086= LiF-?3%F, to 5/30/65 +26.100 0.249 +24.261 +0.102 +1.488 219.190 0.701 68,727 0.289 149.473 0.320 31.355 0.132 68,193 195.240 0.625 61,218 0.257 133.140 Sample No. FP3-7 +0.820 +0.008 +0.762 +0.003 +0,047 196.060 0.633 61.980 0.260 133.187 0.323 31.613 0.133 67.932 0.247 23.133 0.103 76.,437% Sample No. FP3-12 41,094 +0.010 +1,017 +0.004 +0.063 197.154 0.643 62,997 0.264 133,250 0.326 31.953 0.134 67.587 0.335 3L.784 0.138 67.743 Sample No. FP3-17 +1,789 +0.017 +1.663 +0.007 +0.102 198.943 0.660 64,660 0.271 133.352 0.332 32.502 0.136 67.030 0.340 32.606 0.136 66,918 Sample No. FP3-22 +1.787 +0.017 +1.661 +0.007 +0.102 200.730 0.677 66,321 0.278 133.454 0.337 33,041 0.138 66.484 0.352 33.122 0.141 66.386 Sample No. FP3-27 +1.030 +0.010 40,957 +0.004 +0.059 201.760 0.687 67.278 0.282 133.513 0.340 33.346 0.140 66.174 0.355 33.340 0.141 66.165 Sample No. FP3-32 +1.274 +0.012 +1.184 +0.005 +0.073 203.034 0.699 68,462 0.287 133.586 0.344 33,720 0.141 65.795 0.355 33,719 0.139 85.787 Run 3-Residue -0.77 -0.003 -0.241 -0.001 -0.525 202.264 0.696 68.221 0.286 133.061 23,950 0.077 7.510 0.032 16.332 Run 4-I 226,214 0.773 75.730 0.318 149.393 0.342 33.477 0.141 66.041 0.354 33.553 0.145 65,9488 0.350 33.250 0.140 66.260% 2Bold face type indicates nominal values, b . —Results of mass spectrometric analyses. c —fnomalous results; possible misidentification of sample. . i d —Average of 3 analyses, e —Average of 4 analyses, 0T 21 Table 2.9. Summary of MSRE salt analyses, experiment No. 3, fuel salt Weight percent Parts per million 2354 in fuel Date Sample L L e & £ Fe C Ni o crouit(ke) Nominal Analytical? 5/27/65 FP-3-1 10.45 6.4 10.90 4.01 3.984 66.8 98.62 115 28 38 335 39.3 5/27/65 FP-3-2 10.7 6.14 11.19 4.01 3.822 68.63 10049 161 30 44 3350 39.3 5/30/65 FP-3-3 10.7 6.51 11.08 4.42 4.355 69.23 101.88 323 24 53 3915 56.9 5/31/65 FP-3-4 10.8 6.57 11.10 451 4.411 68.4 101.33 371 36 48 60.93 5/31/65 FP-3-5% 10.35 6.71 11.04 4.51 4.450 67.6 100.15 162 28 30 490 60.93 6/2/65 FP-3-6 10.75 6.16 11.60 4.52 4,457 69.69 102.66 112 31 39 330 61.46 6/3/65 FP-3-7 10.6 6.62 1118 4.52 4,416 70.9 103.72 121 31 44 61.46 6/4/65 FP-3-8 10.65 6.39 11.57 4.53 4,448 70.10 103,16 130 28 49 520 61.62 6/6/65 FP-3-9 16.6 6.29 11.43 4.53 4,448 69.35 102,13 113 34 6l 320 61.87 6/7/65 FP-3-10 10.75 6.75 11.39 4.54 4.464 70.12 10345 122 34 25 875 62.05 6/8/65 FP-3-11 10.65 6.46 11.11 4.54 4.444 7090 103.56 123 35 62 3715 62.22 6/9/65 FP-3-12 1055 6.70 11.69 4.55 4,461 68.21 101.61 238 42 68 1745 62.47 6/10/65 FP-3-13 1070 6.56 11.06 4.56 4.494 70.08 102.89 150 33 53 2465 62.89 6/11/65 FP-3-14 1050 6.38 11.32 4.57 4.439 69.72 10236 141 38 60 1760 63.23 6/12/65 FP-3-15 10.70 6.55 11.27 4.57 4.454 69.25 10222 116 38 39 1115 63.31 6/13/65 FP-3-16 10.30 6.36 11.23 4.58 4.462 68.15 100,90 185 48 34 555 63.81 6/15/65 FP-3-17 1050 6.28 11.11 4.59 4.523 68.94 101.56 151 44 44 855 64.14 6/16/65 FP-3-18 10.80 6.46 11.62 4.59 4.463 6948 102.82 143 42 68 64.30 6/17/65 FP-3-19 10.00 6.53 11.13 4.60 4.566 67.12 99.54 139 36 71 605 64.71 6/18/65 FP-3-20 10.50 6.68 10.75 4,60 4.499 68.57 10L03 105 49 61 423 65.12 6/19/65 FP-3-21 1045 6.31 10.69 4.61 4.504 68.30 100.32 97 31 74 705 65.46 6/20/65 FP-3-22 10.50 6.54 11.24 4.62 4.535 68.90 101.82 134 35 73 820 65.80 6/21/65 FP-3-23 990 6.67 10.99 4.63 4.568 68.09 10047 122 35 84 1570 66.13 6/22/65 ¥pP-3.24 10.30 6.50 11.17 4.63 4.582 68.99 101.54 77 37 47 840 66.13 6/23/65 FP-3-25 10.60 6.80 11.12 4.63 4.587 67.99 101.10 135 38 52 580 66.21 6/24/65 FpP-3-26 10.45 6.51 1148 4.63 4.557 68.20 101.21 135 34 90 440 66.37 6/25/65 FP-3-27 10.40 6.44 10.27 4.64 4,633 68.68 10042 167 53 28 310 66.80 6/26/65 FP-3-28 10.65 6.26 10.85 4.65 4.610 67.43 99.80 105 38 34 475 67.05 6/27/65 FP-3-29 1040 6.52 10.18 4.66 4.640 68.54 100.28 110 36 44 520 67.56 6/28/65 FP-3-30 10.35 6.55 10.54 4.66 4.651 68.25 10034 121 37 40 935 67.90 6/29/65 FP-3-31 10.10 6.56 10.84 4.67 4.642 67.83 99.97 128 39 49 326 67.98 6/30/65 FP-3-32 10.10 697 11.20 4.67 4.605 71.66 104,55 149 42 93 1235 67.98 7/2/65 FP-3-33 1040 6.47 1124 4.67 4.563 68.20 100.87 170 40 88 965 67.98 7/3/65 FP-3-34¢ 4.67 67.98 7/4165 FP-3-35 1040 6.85 11.15 4.67 4.597 67.79 100,79 158 48 74 1070 67.98 Av(34) 148 37 55 2Corrected for 2*3U enrichment. bCriticality. €50-g sample obtained for A. S. Mever. compatibie with molten-fluoride mixtures. This recogni- tion developed from the fact that numerous metal fluorides were more stable than the fluorides of these alloy constituents and could be used for fuel and coolant mixtures. By the time conceptual plans for the MSRE were formulated, an alloy of this type was developed at ORNL specifically for application in molten-salt reactor systems, and is now designated as Hastelloy N. Its average composition is 70Ni-16Mo- 7Cr-7Fe-0.05C wt %. The selection of salt constituents with respect to their compatibility with such alloys has been reviewed by Grimes'* and does not require further elaboration here. Values for the standard free energies of formation of the structural metal fluorides, CrF,, FeF,, and Nik,, signify that, as impurities in the reactor salts, iron and nickel should be present in metallic form. In the molten-salt reactor fuel salt, the fluid contains small amounts of UF; generated in the equilibrium reaction Cr? + 2UF;= 2UF; + CrF,. In power operations with the MSRE the concentration of UF, was adjusted occasionally by in-situ generation of U?* 22 Table 2.10. Comparison of nominal and analytical values of uranium concentration in the MSRE fuel circuit — zero-power experiment Sample No. Reactor Inventory Loop Inventory Additions Concentration U U Salt U Salt U Salt Nominal Observed® (kg) (wt %) Run 3-I 145.60 4754.5 145.60 234.5 3.044 3.068b FP 3-1 193.09 4831.0 47.49 76.5 3.997 4.016 FP 3-2 3.822 3.853 FP 3-3 214.52 4865.5 21.43 34.53 4,409 4.390 FP 3-4 219.19 4873.0 4,67 7.52 4,498 195,24 4340.4 5.498 4,446 FP 3-5 195.24 4340.4 4,498 4.486 FP 3-6 195.70 4341.1 0.4581 0.7394 4.508 4.493 FP 3-7 196.06 4341.7 0.3622 0.5847 4.516 4,452 FP 3-8 196.24 4342.0 0.1813 0.2926 4,520 4,484 FP 3-9 196.51 4342.5 0.2687 0.4337 4,525 4,484 FP 3-10 196.70 4342.7 0.1853 0.298% 4,529 4,500 FP 3-11 196.88 4343.0 0.1847 0.2981 4.533 4.480 FP 3-12 197.15 4343.5 0.2742 0.4426 4,539 4.497 FP 3~13 197.60 4344 .2 0.4527 0.7306 4.549 4.530 FP 3-14 197.97 4344.,8 0.3632 0.5862 4,557 4,475 FP 3-15 198.06 4345.0 0.0893 0.1441 4,558 4,490 FP 3-16 198.59 4345.8 0.5323 0.8589 4,570 4,498 FP 3-17 198.94 4346 .4 0.3516 0.5674 4,577 4,559 FP 3-18 199.12 4346.7 0.1723 0.2782 4,581 4.499 FP 3-19 199.56 4347 .4 0.442% 0.7136 4,590 4.602 FP 3-20 200.01 4348.1 0.4494 0.7253 4.600 4,535 FP 3-21 200.37 4348.7 0.3648 0.5887 4.608 4,540 FP 3-22 200.73 4349.3 0.3581 0.5779 4,615 4,571 FP 3-23 201.09 4349.8 0.3558 0.5742 4.623 4,604 FP 3-24 201.09 4349.8 4,623 4.618 FP 3-25 201.17 4350.0 0.0883 0.1426 4,625 4.623 FP 3-26 201.35 4350.3 0.1756 0.2833 4,628 4,593 FP 3-27 201.80 4351.0 0.4494 0.7252 4.638 4.669 FP 3-28 202.16 4351.6 0.3599 0.5760 4.646 4,646 FP 3-29 202.61 4352.3 0.4538 0.7324 4,655 4.676 FP 3-30 202.98 4352.9 0.3675 0.5932 4,663 4.688 FP 3-31 203.07 4353.0 0.0918 0.1482 4,665 4,679 FP 3-32 263.07 4353.0 4,665 4,642 FP 3-33 203.07 4353.0 4,665 4,599 FP 3-34 203.07 4353,0 4.665 FP 3-35 203.07 4353.0 4,665 4.634 Run 3-F 203.07 4353.0 4,665 Run 3-F circuit residue -0.77 -0.77 16,50 4,665 202.30 4353,0° Drain tank charge 23.95 532.6 Run 4-1 226.25 4885.6 4,631 3Corrected for -0.8% bias. Average for Precritical Experiment, PC-2. cNeglects probability that mass of fuel and flush residues are not identical. Table 2.11. Chemical analyses for MSRE fuel carrier salt 23 TLiF-BeF,-ZrF4 (64.7-30.1-5.2 mole %) Batch Net wt. 7Li Analyses of salt Assay Li Be Zr F Cr Ni Fe (kg) (wt %) (ppm) F-162 130.8 99.994 10.6 7.43 11.5 70.0 23 17 38 F-163 131.6 99.994 10.6 7.36 12,0 69.9 21 21 54 F-164 133.8 99.996 11.0 7.43 11.6 69.9 14 24 145 F-165 132.0 99.996 10.8 7.30 12.1 70.1 29 24 39 F-166 137.0 99.996 11.3 6.98 1l1.5 70.0 23 8 124 F-167 125.2 99.996 10.9 7.23 12.1 70.1 19 <3 26 F-168 133.1 99.996 10.8 7.30 12.0 70.0 13 3 33 F-169 132.6 99.995 11.1 7.38 12,0 69.6 15 16 36 F-170 134.5 99,995 10.5 7.37 11.8 70.3 20 7 46 F-171 132.9 99.995 10.69 6.98 11.90 70.3 17 17 91 F-172 134.2 99.995 10.61 7.33 11.72 70.3 26 16 89 F-173 133.1 99.995 10.50 7.29 12.12 70.0 18 24 79 F-174 133.8 99.995 10.66 7.57 11.99 70.0 31 17 141 F-175 133.6 99.995 11.17 7.05 11.97 70.37 19 <5 120 F-176 133.8 99.995 11.65 7.27 11.84 69.65 13 <5 80 F-177 133.8 99.995 10.58 6.95 11.92 69.77 29 26 90 F-178 131.1 99.995 11.03 7.37 12.09 69.91 28 7 84 F-179 131.1 99.995 16.79 7.47 12.24 69.75 33 11 106 F-180 133.4 99.995 11.19 7.49 11.72 70.45 29 40 198 F-181 134.8 99.995 10.84 6.90 12.48 69.1 25 29 141 F-182 134.2 99.995 10.73 6.89 12.44 69.3 27 29 240 F-183 133.0 99.994 10.99 7.21 12.35 69.80 22 8 53 F-184 132.8 99.995 10.90 7.16 12.28 69.27 23 <5 61 F-185 133.2 99.995 10.92 7.30 11.99 70.24 22 <5 58 F-186 133.5 99.995 10.99 6.86 12.10 69.59 19 <5 53 F-187 133.4 99.995 10.87 7.13 12.25 69.71 27 9 136 F-188 133.8 99.995 10.86 7.14 11.84 69.86 24 19 113 F-189 134.2 99.995 10.87 7.32 11.93 70.74 22 6 63 F-190 134.1 99.995 10.89 7.26 11.78 70.62 22 8 88 F-191 132.0 99.995 10.35 7.23 12.0 71.2 17 13 69 F-192 134.3 99.995 10.31 7.19 11.5 72.0 28 28 150 F-193 133.4 99.995 9.75 7.40 11.4 71.4 13 10 220 F-194 137.3 99.995 10.82 7.33 12.1 70.3 21 12 137 F-195 134.1 99.994 10.56 7.34 12.2 70.5 19 16 173 F-196 131.1 99.994 10.48 7.34 11.4 71.5 24 11 151 F-197 132.6 99.99%4 10.88 7.39 11.18 70.6 21 10 130 F-198A 146.4 99.994 4.68. U-62.26 33.60 8 29 66 F-198B 136.6 99.994 4,67 U-62,.82 33.82 10 <5 33 mn the salt or by addition of FeF, as an oxidant Generally, an effort was made to hmut the [U>*]/[ZU] ratio to ~0 0l as a means of retarding the corrosion reaction The reaction Ct® + 2UF, = 2UF; + C1F, has an equilibrium constant with a small degree of temperature dependence Since the equilibrium constant increases with temperature, operation of a loop with a significant temperature gradient causes the chromium concenira- tion of the alloy surface that 1s at high temperature to decrease, transferring metal to the cooler regions Since chromium comprises a small fraction of Hastelloy N and 1ts availability to the salt 1s diffusion limited, attainment of equilibrium by this reaction 1s very slow In prenuclear operations with nonuraniferous salts, the salt loops were operated in essentially 1sothermal conditions Under these conditions, corrosion of the circuit walls could arise only from contaminants In all the activities associated with the startup of the MSRE, stringent measures were taken to remove all traces of moisture from the newly constructed system and to prevent subsequent introduction of impurities Table 2.12. Chemical analyses for TLiR-? 35UF4 (73-27 mole %) fuel concentrate for the MSRE Chemical analysis Batch No Weight percent Parts per mllion L: U F Cr N1 Fe E 201 479 6153 330 17 <5 26 E 202 4 80 6202 332 27 89 74 E 203 492 62 26 334 29 47 11 E 204 4381 6121 333 24 9 69 E 205 468 6159 330 21 14 22 E 206 470 6151 327 <5 21 <5 24 During October and November 1964, the fuel system was readied for use, 1t was purged of moisture, heated to 650°C, and tested for leak tightness The purge gas was high-punty hehum To ensure that it was free of oxygen and moisture 1n use, the gas flowed through a series of dryers and preheaters and through hot tita nium sponge before 1t entered the reactor From the results of analyses of the individual salt batches produced for use in the MSRE (see Tables 2 11 and 2 12) the average concentrations of Cr, Fe, and N1 in the blended salt charge (" LiF-BeF,-ZrF4-2*® UF, ) at the beginning of the precritical experiment PC-2 should have been 21, 99, and 14 ppm respectively The salt samples removed from the fuel pump during this test were supplied to the ORNL General Analytical Labora tory (GAL) as well as to the HRLAL so the overlapping results would show whether shight biases might result as new equipment and necessary modifications in methods were employed As determined by the GAL, the average concentrations of structural metal impurities in the PC-2 fuel circuit salt were Cr 19, Fe 99, N1 23 ppm, while from the HRLAL, the average concenirations were found to be Cr 37, Fe 163, N1 34 ppm (Table 27) Following run PC-2, "LiF ??3UF, enriching salt was added to the carrier salt to constitute the final 235U fuel mixture The average concentrations of the impurities on completion of this experiment as deter- mined by the HRLAL were Cr 37, Fe 148, N1 55 ppm (Table 2 7) With extended experience, we learned that one standard deviation for Cr analysis was ~7 ppm, there fore, the difference in chromium concentration as measured 1n the two laboratories can be regarded as a bias correction — that 15, HRLAL results could be expected to average about 18 ppm higher than those from the GAL Table 2.13. Composition of the MSRE fuel salt 1n the precritical examnation L1 Be Zr 237,934y Total Carrier (norminal composition), kg 51691 309 62 541 36 45200 7L1F—238UF4 (nomunal composition), kg 1155 145 60 234 5 Total fuel salt (nominal), kg 528 46 309 62 541 36 145 60 47545 Weight percent 1112 651 1139 3062 Weight percent analytical (18 samples) 1042 6 60 11 35 3 068 Carrier (analytical composition), kg 487 75 327 61 539 06 "L1F-238UF, (analytical composition), kg 10 96 146 34 2345 Total fuel salt (analytical), kg 498 71 32761 539 06 146 34 Weight percent 1049 6 89 11 34 3077 Weight percent analytical (18 samples) 1042 6 60 1135 3 068 25 Table 2.14. Summary of MSRE salt analyses, experiment No. 4, fuel salt Weight percent Parts per mitlion Date Sample ya Ia Be VA b2 Te Cr N1 QO Nominal Analytical 12/21/65 FP-4-15 800 12/21/65 FP-4-16€ 12/22/65 FP-4-17 655 12/24/65 FP-4-18 10 25 668 1124 4651 6744 10026 144 43 44 859 12/22/65 FP-4-19¢ 12/23/65 FP 4-20¢ 12/24/65 FP-4-21 10 47 640 1082 4671 66 79 99 15 121 60 52 12/25/65 FP-4-22 10 54 654 1095 4 664 64 68 97 37 116 44 84 12/26/65 FP-4-23 1027 674 1086 4 642 65 06 97 57 99 46 35 12/27/65 FP-4.24 10 65 665 1096 4 655 67 66 10058 116 48 45 12/31/65 FP 4-25 10 60 637 1141 4 646 65 44 98 47 26 35 42 1/1/66 FP-4 26 10 60 663 1120 4 642 66 90 99 97 89 48 47 1/2/66 FP-4-27 10 55 653 1107 4618 67 68 10045 222 50 41 1/3/66 FP-4-28 10 60 642 1110 4 663 66 19 9897 211 49 34 1/5/66 'P-4-29 1070 671 1154 4 654 69 75 103 35 11T 39 31 1/7/66 FP-4-30 10 63 681 1119 4 661 67 32 100 58 83 49 27 1/10/66 FP-4-31 10 30 663 1126 4632 68 51 10133 190 37 41 1/12/66 FP-4-32 10 55 671 11 80 4625 6766 101 34 173 43 33 1/14/66 P-4 33 11200 675 1107 4596 66 25 99 97 55 50 39 t/17/66 FP-4-34 11355 649 1113 4 601 6820 10182 164 58 16 1/19/66 FP-4-35 1136/ 668 1086 4721 69 35 103 01 74 54 <5 1/21/66 FP-4-36 11255 632 1120 4632 6676 10016 125 47 <5 1/22/66 I P-4-37 11300 633 1108 4622 69 35 10268 189 53 80 1/24/66 FP-4-38 10 70 654 1146 4 608 67 25 10056 311 51 25 1/26/66 P4-39 @Values corrected to 33 696 at % >>°U bur purge method “No sample obtamed dKBrF4 method €For amperometric analysis T Erroneously high because of weak batteries in automatic pipet On completion of the fueling operation in July 1965, the reactor was dramned and flushed Radiation levels were Jow enough to permit mamntenance and mstalla- tion work to begin mmmediately in all areas, n preparation for high-power operation In August 1965, the assembly of graphite and Hastelloy N surveillance specimens, which had been 1n the core from the beginning of salt operation, was removed While the reactor vessel was open, mnspection revealed that pieces were broken from the horizontal graphite bar that supported the sample array The pieces were recovered for examination, and a new sample assembly, designed for exposure at high power and suspended from above, was installed The fuel-pump rotary element was re- moved 1n a final rehearsal of remote mamtenance and to permut inspection of the pump nternals It was remnstalled after inspection showed the pump to be in very good condition Tests had shown that the heats of Hastelloy N used 1n the reactor vessel had poor high-temperature rupture life and ductiity in the as-welded condition Since the vessel closure weld had not been heat treated, the entire vessel was heated to 760°C (1400°F) for 100 hr, using the installed heaters, to improve these properties The experience which was developed with the MSRE throughout the shakedown peniod preceding power operations confirmed that the molten-fluoride salt mixtures were intrinsically noncorrosive to Hastelloy N and that effective procedures were employed to prevent serious contarmnation of the salt circuits during this period By December, maintenance and modifications were completed, and flush salt was circulated through the fuel system for a period of three days On December 20, 1965, the fuel salt was loaded into the system, and the zero-power experiment began On completion of 26 Table 2.15. Chemical composition of the MSRE fuel salt in prenuclear operations 1 Be Zr U b3 Carrier salt, kg 516 91 309 62 541 36 45200 TLF-238UTF,, ke 1155 145 60 234 5 TLiF-235UF, (bulk charge), kg 584 7359 118 55 Total fuel salt, kg 534 30 309 62 541 36 21919 4873 05 Loop charge (89 07%), kg Carrner salt 460 41 27578 482 19 LiF-238yF, 10 29 129 69 LiF 235UF, 520 65 55 Additions to loop 6 32 7 83 12 736 Total 43530 Loop charge, wt % 1108 6 34 1108 4 665 FP 3-32-35 (3), wt % 10 30 676 1120 4 624 Fuel salt residue in loop, kg 183 105 183 017 165 Salt to drain tank, kg 480 39 274 73 480 36 202 30 4336 5 Dramn tank charge, kg 58 40 3384 5908 2395 5326 53879 308 57 539 44 226 25 4869 1 Drain tank charge, wt % 1107 6 34 11 08 4 647 Charge to loop for run 4 (92%), kg 495 69 283 88 496 29 208 15 4479 6 Flush salt residue m loop, kg 251 1 66 180 Total fuel salt, kg 498 20 285 54 496 29 208 15 4497 6 Total fuel salt, wi % 11 08 6 35 1104 4628 Analytical (69 samples), wt % 10 785 + 0 064 6571 +0076 11197 20260 4641+ 0026 this period the concentrations of the structural metal mmpurities had average values of Cr 48 £ 7, Fe 131 % 65, N1 40 = 20 ppm Here, for the first time, there was evidence that an increase in the concentration of chrommum in the fuel-salt mixture had taken place As will be noted later, such mcreases followed a regular pattern, occurring almost exclusively after periods of maintenance when the reactor was opened to the ambient atmosphere for some period of time It now seems that circulation of the flush salt for periods of only a few days was insufficient to negate the effect of the exposure completely 2 4.4 Oxide contaminants. Among the summary reports on various aspects of MSRE operations, no separate review of analytical chemical developments is included Except for a few analyses, for example, for oxide and for lithium, the methods in the ORNL Master Analytical Manual were employed routinely and found to be satisfactory In response to the need for improved methods for deternunation of the concentration of oxides in the MSRE salts, a rehiable and accurate method was devised and applied concurrent with early MSRE operations This development was an important aspect of MSRE chemical development and 1s therefore described 1n detail here Long before operations with the MSRE were started 1t was demonstrated 1n laboratory tests that precipita- tion of oxide as the saturating phase in the molten fuel, flush, or coolant salt was very unhkely However, assurance that the concentration of oxide as a con- tarminant of these salts was not increasing could not be obtained from imtial measurements using conventional methods '® Many of the results of early analyses of MSRE flush and carrier salts were anomalous, as shown in Tables 2 2,2 7, and 29 At face value, the analytical data shown 1n these tables might suggest that the concentration of oxide in the samples examined ranged sporadically from 100 to 4000 ppm If real, the higher values would represent the presence of more than 0 5% Z10, 1n the specumens Each of the 52 specimens obtained for analysis durning precrnitical run No 2 and duning the zero-power expermment was subjected to petrographic exanmunation With the carner and LiF- BeF, salt, the sensitivity for detection of well-formed crystals of ZrQ, by the petrographic methods 1s well below 100 ppm, neither crystalline ZrO, nor oxy- fluorides were found to be present in any of the specimens examined The sampling procedures employed at the MSRE effectively protected salt specimens from contact with moisture up to the pomnt where they were transferred into a transport container, subsequent handling proce- dures, however, were not so stringent and are presumed to have allowed moisture to contact the surfaces of the fluoride specimens on occasion Beryllium fluoride and LiF-BeF, glassy phases are hygroscopic Crystallization of each of the MSRE salt mixtures can produce minute quantities of these phases, typical sampling conditions effect rapid cooling of the samples and are therefore even more conducive toward the production of the hygroscopic phases than equilibrium crystallization of the salt muxtures These hygroscopic phases, once moistened, cannot be redried completely by ordinary desiccants at room temperature The anomalously high values for oxide analyses during the early stages of MSRE operations are thus indicative of exposure of the salt samples to moisture-laden atmospheres An effort was nitiated to develop improved methods for oxide analysis that would be adaptable for use n the HRLAL After a study of several methods which might have application under conditions such as those which prevailed in the HRLAL and where no provision for atmospheric control m the hot cells was made, the analytical chemists'® concluded that a hydrofluorina- tion-transpiration method based on the reaction O? + 2HFE(g) = H,0(g) + 2F was potentially the most useful Their development of the new method 1s described 1n the following paragraphs In princaple, the amount of water evolved from purging a molten-sait sample with an H,-HF gas mixture would serve as a measure of the quantity of oxide 1n the molten-salt sample Since the water evolved and the HF consumed are both proportional to the amount of oxide present, either compound would serve as an indicator species Water momitoring methods were selected because of their greater convenience and reliability than those for HF The apphication of this method to the analysis of radioactive samples required the development of (1) a sampling technique which minimized atmospheric con- tamination, (2) the incorporation of 4 water-measure- ment techmque which was convenient for hot-cell operations, and (3) the fabrication of a compact apparatus for conserving the himited space available in the hot cells It was necessary to adapt the sampling techniques from methods already developed to obtain samples for wet analysis By using a copper ladle of nearly the same dunensions as the enriching capsules, approximately 50-g samples were obtamed which could be transported in the existing transport container Exposure to the atmosphere was minumized by remelt- ing and hydrofluonnating the entire sample n the 27 sample ladle The ladle was sealed 1n a nickel-Monel hydrofluorinator, with a delivery tube spring-loaded against the surface of the frozen salt Before the salt was melted, the apparatus was purged with a hydro- fluorinating gas mixture to remove water from the inner surface of the hydrofluorinator and from the exposed surface of the salt As the sample was melted, the delivery tube was driven to the bottom of the ladle by spring action for efficient purging of the sait In all preliminary tests the water in the effluent purge gas was measured by Karl Fischer titration While the Karl Fischer reagent was shown to be remarkably stable to radiation, the titration would have been excessively difficult to perform routinely 1n the hot cell As an alternative, an electrolytic mosture monitor was adapted for these measurements Since such equipment 1s subgect to interference and damage from HF,' 7 1t was pracfical to include a sodwm fluoride column 1n the effluent gas traimn In operation at about 90°C 1t removed HF from the etfluent gas without significant holdup of water A schematic flow diagram of the appatatus mstalled in the HRLAL 1s shown in Fig 22 The apparatus 15 pictured m Fig 23 A modular design was selected to facilitate any changes found necessary during com ponent testing and to permit necessary repairs in the hot cell Except for the hydrofluorination furnace, all hot cell components were contained within a cubical compartment, 16 1in on a side Samples of the flush and fuel salt taken during the December startup of the reactor were analyzed for oxide Table 2 16 summanzes the results Figure 24 152 recording of the data output from analysis of the first fuel-salt sample taken afier the fuel was loaded into the reactor The results of the analyses by the hydro- fluorination method were in good agreement with those by the KBrF, procedure The KBrF, values paralleled the trends shown by the hydrofluorination method but Table 2.16. Oxide concentrations of flush and fuel salt from the MSRE Tume of salt crculation Oxide concentraion Sample (hr) (ppm) Flush salt 247 46 29 1 72 47 6 106 Fuel salt 34 120 238 105 322 30 525 65 28 ORNL-DWG 65-8855A MOISTURE / MONITOR Pb SHIELD / He SPLITTER > (¢ Dyl - U T PURIFIER / N\ LHe BACK FLUSH / 'f? | NoF - HF TRAP N N\ N\ " N CELL WALL § \ D HYDROFLUORINATOR Fig. 2.2, Schematic flow diagram of the apparatus for the determination of oxide in MSRE fuel by hydrofluorination. averaged shightly higher. This bias was not unexpected, since the pulverized samples required for the KBrF, method were easily contaminated by atmospheric mois- ture. Subsequently, 50-g fuel-salt samples were taken at various power levels from zero to full power. With an in-cell radiation monitor, the initial sample read 30 R at 1 ft. This activity increased to 1000 R at 1 ft at the full-power level. Results of the oxide analyses are given in Table 2.17. The oxide in the coolant salt sample, 25 ppm, is, for practical considerations, identical with a value of 38 ppm obtained for a coolant salt sample taken on January 25, 1966, and analyzed in the laboratory after three weeks’ storage. The fuelsalt analyses are in reasonable agreement with the samples analyzed in the development laboratory before the reactor was oper- ated at power. The oxide concentration in these nonradioactive samples seemed to decrease gradually from 106 to 65 ppm. Between the FP-6 and FP-7 series the sampler-enricher station was opened for mainte- nance, and thus the apparent increase in oxide concen- tration (ca. 15 ppm) may represent contamination of Table 2.17. Oxide concentrations of coolant and fuel salt from the MSRE QOxide concentration Sample Code (ppm) Coolant salt CP-4-4 25 Fuel salt FP-6-1 49 FP-64 53 FP-6-12 50 FP-6-18 47 FP-7-5 66 FP-7-9 59 FP-7-13 66 FP-7-16 56 the samples by residual moisture in the sampling system; however, the number of determinations was not at the time sufficient to establish the precision of the method at these low concentration levels. In an attempt to determine whether radiolytic fluo- rine removes oxide from the fuel samples, sample FP-7-9 was removed from the transport container and 29 nd & ® © T o T a cell apparatus. . Hot 3 .2, Fig 30 ORNL- DWG 66-402BA 40 H, FLOW, 200 cc/min e HF, 0.2 atm E FUEL SALT, 48.7¢q > OXIDE FOUND, 120 ppm 3 30 - n,,300°C o |.:I->| —»Ha-HF,300°C ~SALT MELTED i 20 \ r‘\ e 2 o T 10 \ v | ¥ 0 __’__/ — 0 20 40 60 80 100 TIME (min) Fig. 2.4. Determination of oxide in MSR fuel sample FP4-11 by hydrofluorination. stored in a desiccator for 24 hr prior to analysis. Since the oxide content, 59 ppm, is comparable with that of the remaining samples for which analyses were started 6 to 10 hr after sampling, no significant loss of oxygen is indicated. A more direct method of establishing the validity of these results by measuring the recovery of a standard addition of oxide was used in subsequent operations. A tin capsule containing a known amount of SnQ, was heated to 550°C in the hydrofluorinator as hydrogen passed through the system. The SnO, was reduced to the metal, and the water that was formed passed on to the electrolysis cell. Two standard samples of Sn0O, were analyzed after a four-month interim, and oxide recoveries of 96.1 and 95.6% were obtained. The slight negative bias was attributed to momentary interruptions in the flow of the hydrofluorinator effluent gas through the water electrolysis cell. Diffi- culty with cell plugging was encountered throughout the period of development of the oxide method. As an attempt to eliminate the negative bias and also to provide a replacement cell for the remote oxide apparatus, it was deemed necessary to find a method of regenerating the electrolysis cell that would permit a steady gas flow at relatively low flow rates. The water electrolysis cell contains partially hydrated P, O; in the form of a thin viscous film in contact with two spirally wound 5-mil rhodium electrode wires. The wires are retained on the inside of an inert plastic tube forming a 20-mil capillary through which the sample passes. The 2-ft-long tubing element is coiled in a helix inside of a %-in.-diam pipe and potted in plastic for permanence. During the course of the investigation of the cell, it was found that a wet gas stream in itself did not cause the electrolysis cell to plug. It was also necessary for curtent to be flowing through the cell for flow interruptions to occur. This indicated that the hydrogen and oxygen evolving from the electrodes create bubbles in the partially hydrated P, Qs film, which then grow in size sufficiently to bridge the capillary and form an obstructing film. After many unsuccessful approaches, an acceptable solution to the problem was obtained by means of a special regenerating technique which employed dilute acetone solutions of H3PO, as the regenerating solu- tions. This provided a desiccant coating sufficient to absorb the water in the gas stream and gave a minimal amount of flow interruptions during electrolysis. The cells were thereafter successfully regenerated in this manner and yielded oxide recoveries of 99.6 + 1.3% from standard SnQ, samples. 2.4.5 Analysis of helium cover gas. As noted in Sect. 2.4.4, a reliable means for the determination of oxide in the fuel salt did not exist when the initial experiments with the MSRE began. The results of the oxide analyses were perplexing, and they were almost certainly invalid as indicators of contamination by moisture, since the concentrations of structural metal contaminants in the salt remained essentially constant. In order to obtain some direct assurance that negligible inleakage of moisture was occurring in the fuel system, temporary measures were devised and applied. The method con- sisted in bleeding off a small flow of the helium gas in the pump bowl and passing it through a monitoring system in the high-bay area. Mean values for the concentration of HF in the helium cover gas were obtained during the PC-2 and zero-power experiments by adaptation of a continuous internal electrolysis analyzer for gaseous fluorides Through the cooperation of ORGDP personnel, data were obtained regularly throughout prenuclear opera- tions The results showed that HF was evoived from the salt at a maximum mean value of 150 to 200 ug/hiter ' ® The validity of this value 1s somewhatl questionable because of the probable positive bias that particulate fluorides would contribute The results are, therefore, conservative to an indeterminate degree, but correspond to the introduction of no more than 1 ppm of oxide into the salt per day through the reaction 2H, O + Zrk, = 210, + 4HF | an amount which would escape detection by other methods 2 4 6 Lithium analysis In Sect 2 4 2 we noted that chemical analyses of the fuel salt routinely indicated that Be/Li ratios were significantly greater than for the nomunal composition of the salt A careful examination of the resulis for all the salt mixtures prepared for use m the MSRE shows that if 1t 18 assumed that the analytical results are correct for berylhum in the coolant and flush salts, for beryllium and zirconium in the carrier salt, and for urammum in the LiF-UF, salt muxtures, and if a correction factor of 5 to 6% 1s added to each of the results of hithium analyses, the average results of composition analyses nearly coincide with nomunal values When this disparity was first noted, a reexamination of the production records was made but did not disclose any dispartty in the material balances that would account for the anomaly in Be/Li results Evidence from the purification-plant mventory data indicated that the composition of the salt delivered into the reactor system was of the design composition, 66 0 *+ 025 mole % LiF, 34 0 £ 0 25 mole % BeF, A sample of the MSRE coolant, taken from one of the batches loaded into the reactor, showed a hquid-sohd phase transttion temperature of 457 6°C, thus 1s within 0 1°C of the temperature of the equilibrium hquid-solid reactions which can occur in LiF-BeF, muxtures richer than ~33 mole % LiF Chemical analyses of this material indicated 1ts composition to be LiF-BeF, (63 63-36 37 mole %) The thermal data indicated, however, that the matenal contained at least 65 5 mole % LiF, and thus confirmed the composition indicated by the weights of the maternals used in 1ts preparation Application of the same correction factor to the results of hthium analyses for either LiF-UF, or LiF-BeF, mixtures effects equal improvement of the 31 match of analytical results with nominal values, we must conclude therefore that a negative bias of 5 to 6% existed m the lithium results Lithium was determined in the MSRE salt samples by flame spectrophotometric methods The results are compared with calibrated standards, and regarded by the analytical chemisis to match satisfactorily with these standards However, view of the comncidence of values which results when a correction of ~5% 1s applied to the results, we must infer that a negative bias of that magnitude affected the MSRE results The lithtum used to produce the salt charges for the MSRE was selected from stockpiled LiOH in which the ®L: had been depleted to 001% or less Assays were available on each batch of LiOH, these analyses served as criterta for selection of the material used 12-2° The assays of the batches of LiOH which were to be used for the MSRE ranged from 00072 to 0 0085% °La, with the average of the batches which were to be used for the flush salt and the fuel carrier salt as 0 0074% After each converston of hydroxide to the fluoride,’ the hthium 1n each product batch was again assayed before the LiF was used to make up the coolant, flush, or carrier salt The assays of the batches of LiF used to make up the fuel salt ranged from 0 004 to 0 006% and averaged 0 0049% 21 In view of the relation of tritium production in the MSRE to the 1sotopic composttion of lithrum, Hauben- reich?? recently completed a full review of the possible sources of triium and of the analyses on which production estimates were based In connection with that review, S Cantor obtained new analyses for ®Lt in unused LiF-BeF,-ZrF, carrier salt The results comn cided with those obtatned imtially for LiOH, but were, for reasons still unknown, substantially hugher than the ~50 ppm which was previously used as the basis for estimation of production rates 23 247 Examination of salts after 7ero-power experi- ment. At termination of the zero-power experiment, the fuel salt was drained for temporary storage n fuel drain tank FD-2 Flush salt was circulated through the fuel system for a 24-hr period and drained for storage Four specimens of fuel salt were obtained from the fuel dramn tank at intervals during the first 300-hr penod of storage This practice was not continued after power generation began because of the hazards which would be mcurred by attempts to provide temporary access for samphing radioactive salt One purpose of the effort to sample salt from the dratn tank was to confirm that the composition of the salt in the tank conformed to that expected from mixing the newly constituted salt in the fuel circuit with that of lower uranium concen- tration m the dramn tank Any mismatch noted at this time would serve as a base-hne correction to data obtamed later during power operations A second purpose was (o investigate the effectiveness of quiescent storage of the salt in reducing the amounts of sus- pended particles of the structural metals won and nickel In retrospect, this period afforded the oppor- tunity for securing a wide vanety of base-line data, an opportunity that was exploited only to a modest extent, 1 that only four salt samples were removed for analysis Their compositions, as determined from chem- ical analysis, are hsted in Table 2 18 The data shown in Table 2 18 indicate that approxi- mately the same bias noted previously n the analysis of carrier constituents continued to be evident and that the disparity between the analytical data and the nominal concentration of uranium was ~0 3%, which 1s within the precision normally found for uramium analyses A comparison of the values for 1ron and nickel in the samples drawn from the drain tank and those previously 32 obtained from the pump bowl shows no significant difference Ewidence confirmung this expectation was provided later by analysis of two samples removed from the pump bowl before the reactor was operated at power Samples were withdrawn from the reactor n enricher ladles and were transferred under an inert atmosphere to the graphite crucible of an electro- chemical cell assembly for electrochemical studies by D L Manning The cell assembly and electrodes developed for these electrochemical studies are described else where 2% Average total concentrations of iron and nickel in the melt, as determmed by conventional methods, were about 125 and 45 ppm respectively Iron and nickel, as cationic species in the molten fuel, are electroreducible in the melt and can thus give voltam metric reduction waves By voltammetry and by stand- ard addition techniques, Mannmg found?®? that the salt contained ~10 ppm of wron as Fe?" and that if nickel were present as a cationic species, 1ts concentration was below the limit of detection by voltammetry (<1 ppm) Table 2.18. Composition of MSRE fuel-salt samples obtained from fuel dratn tank FD-2 Sample Composition {wt %) Composition (mole %) No Li Be Zr U Lit BeF, ZiF, UF4 FD-2-10 1013 6 54 1110 4 580 FD-2-11 10 18 6 54 11 27 4 638 FD-2-12 10 55 6 68 11 95 4614 FD-2-13 10 55 622 1095 4613 Average 10 35 641 11 32 4611 63 38 30 46 532 0836 Nomunal 65 19 29 00 501 0 808 Sample Parts per mllion No Cr Fe N1 FD-2-10 A? 37 141 83 B? <30 53 <30 FD-2-11 A 37 97 72 B <30 63 <30 FD-2-12 A 26 91 31 B <70 50 7 FD-2-13 A 38 147 132 B <70 50 6 Average A4 35 119 80 A(3) 33 110 62 Run 3 A 3748 154 £55 48 £ 19 ZAnalysis by wet chemical methods bAnalysm by spectrochemmcal methods 33 PHOTO 80844 INCHES Fig. 2.5. Flush sait residue from volute of the pump bowl. Since all previous transits of the salt were conducive to retaining suspended metal in the salt, we felt that possibly the finely divided metal (some threefold more dense than the fluoride mixtures) would gradually setile in drain tanks on conclusion of the zero-power experi- ment. In an attempt to test this possibility, samples of the static fuel salt were obtained from the fuel storage tanks throughout a two-week period following the zero-power experiment. The concentrations of the structural metal impurities were not found to change during this storage period. We concluded, therefore, that thermal convection in the storage tank was sufficient to prevent settling of fine metallic particles in this tank and that the concentrations of these metals 1n the salt might be expected to reman essentially constant throughout the remainder of MSRE opera- tions. On removal of the fuel pump rotor for examina- tion,?® approximately I kg of flush salt was found to have been retained in the rotor flange area (Fig. 2.6) and 1n the volute of the pump bowl (Fig. 2.5). Additional small fragments of salt were found in various adjacent locations: in the grooves of the shield-block O-ring and in line 903, through which pump-bowl gas was vented to an HF analyzer that was installed temporarily in the prenuclear operational period to analyze the composition of the off-gas. All salt speci- mens were submitted for microscopic and/or chemical analysis and were found to be entirely free of oxides or oxyfluorides. Results of the chemical analyses per- formed with these samples are listed in Table 2.19. A minor oil leak allowed oil to enter the shield-block Table 2.19. Chemical analysis of fuel pump salt samples Chemical composition (wt %) Sample locatton L1 Be Zy U I Labyrinth flange 1336 9.381 0.15 0.0177 7805 Pump volute 1370 9.71 0021 0.0195 77.68 Upper O-ring groove 9.81 6.9 10.87 4.294 62.57 Lower O-ring groove 12.1 0.910 0.295 72.76 Pump suction No chemical analysts run. Contained dispersion of fine nongraphitic carbon. 1025 671 11.11 4.602 68.87 13.12 968 O 0 77.08 Fuel salt Flush salt section of the pump. Thermal decomposition products of the pump oil were in contact with the salt specimens in these areas and were found as partial films on some of the salt (Fig. 2.5) or dispersed through fragments of other specimens. In all locations where crystallized salt deposits were found, it was evident that the molten salt had not adhered to the metal surfaces. This was particularly evident in the volute deposit, where the molten specimen cooled rapidly and maintained the surface angle tangency characteristic of sessile drops which are free of contaminant oxides. 2.4.8 Appraisal of chemical surveillance in prepower tests. From a chemical standpoint, operations of the MSRE during the precritical and zero-power experi- ments were performed in a manner that maintained the purity of the salt charges during all transfer, fill, and circulation operations. The results of the chemical analyses obtained with samples from the fuel and 34 o o g nr o LA PHOTO 81120 Fig. 2.6. Flush salt in rotor flange area. coolant systems reinforced the longstanding conclusion, drawn in laboratory and engineering-scale experiments, that pure molten-fluoride mixtures are completely compatible with nickel-based alloys. The utility of routine analysis of salt samples removed from the circulating systems came to be realized as of questionable value with respect to several of the constituents after brief experience with the MSRE. It became quite clear, for example, that only easily recognized operational perturbations were likely to cause the average concentrations of carrier or coolant salt constituents to vary beyond standard deviations in the analyses and that without major refinement in the analytical methods for determining the concentration of uranium in fluoride mixtures these analyses would not provide operational control. Nonetheless, the potential value for applications of statistically signifi- cant results justified, in our judgment, a continued and active program of chemical analysis. Furthermore, by the time the zero-power experiment was completed, the value of chromium analysis as the single quantitative indicator of corrosion or of corrosion-free performance was solidly established. Analysis of salt samples for this impurity involves wet chemical dissolution techniques, which account for some 90% of the cost of a complete analysis. Because of this factor, there was little reason to omit analysis of more than one component, for such omission would preclude determination of overall com- position. Several aspects of the relation of analysis to MSRE operations proved to be discomfiting by the completion of the zero-power experiment. Annoyingly, chemical analysis of the flush-salt samples removed from the system indicated that the amounts of salt residues remaining in the system after drain operations were somewhat more than could be accounted for by the on-site estimates of the volumes of probable void space where it was deduced that salt could reside. Ultimately, the application of neutron activation and mass spec- trometric analysis to this problem confirmed that estimates made from chemical data were the most nearly correct. Until this problem was resolved, con- tinued exchange of fuel and flush salt by drain-fill operations suggested a negative bias in the analysis of uranium, Methods were developed for slight refine- ments in the uranium analysis, but not surprisingly, these did little to resolve the fundamental problem. Weigh cell data, in which some reliance was placed originally, proved to be of little value with continued experience, and accordingly, increasing reliance on chemical analyses was developed. For a lengthy period, there was no absolute method for establishing the level of contamination of the salts by small amounts of oxide ion. This assay was performed in a partly satisfactory manner during the precritical and zero-power experiments, but the method was not adapted for application to highly radioactive materials until well into the power period. The lack of dynamic methods for obtaining various analyses of the gas and vapor above the surface of the fuel salt proved to be a severe limitation in the interpretation of the chemical behavior which occurred in the fuel system. In the absence of on-site gas analysis, either by mass spectrometry or gamma-ray spec- trometry (which was developed later for postopera- tional examinations), interpretation of several aspects of molten-salt reactor behavior remained less than completely resolved — in particular, of greater signifi- cance in power operations, the experimental investi- gation of factors controlling the transport of noble metal fission products. In general, the surveillance program was, to this stage, a success principally because of the careful procedures which were employed to ensure its trouble-free opera- tion and because of the intrinsic stability and compati- bility of the materials used in its construction. References 1. J. H. Shaffer, Preparation and Handling of Salt Mixtures for the Molten-Salt Reactor Experiment, ORNL4614 (January 1971). 2. MSR Program Semiannu. Progr. Rep. Feb. 28, 1965, ORNL-3812, p. 5. 3. R. B. Lindauer, Processing of the MSRE Flush and Fuel Salts, ORNL-TM-2578 (August 1969); G. Long and F. F. Blankenship, The Stability of Uranium Trifluoride, ORNL-TM-2065 (November 1969), 4. R. B. Lindauer, private communication. 5. C. F. Baes, “The Chemistry and Thermodynamics of Molten-Salt Reactor Fluoride Solutions,” ITAFA Symposium on Thermodynamics with Emphasis on Nuclear Materials and Atomic Transport in Solids, Vienna, Austria, 1965. 35 6. A more detailed description of the hydraulics in the system is given in the report by J. R. Engel, P. N. Haubenreich, and A. Houtzeel, Spray, Mist, Bubbles and Foam in the Molten-Salt Reactor Experiment, ORNL-TM-3027 (June 1970). 7. MSR Program Semiannu. Progr. Rep. Feb. 28, 1965, ORNL-3812, p. 10. 8. MSR Program Semiannu. Progr. Rep. Feb. 28, 1965, ORNL-3812, p. 10. 9. L. T. Corbin, W. R, Winsbro, C. E. Lamb, and M. T. Kelley, “Design and Construction of ORNL High- Radiation-Level Analytical Laboratory,” /1th Hot Lab Conf. Proceedings, ANS, November 1963; “High- Radiation-Level Analytical Laboratories,” Anal. Chem. 37(13), 83A (1965). 10. B. E. Prince et al., Zero-Power Physics Experi- ments on the MSRE, ORNL-4233 (February 1968). 11. J. R. Engel and B. E. Prince, The Reuactivity Balance in the MSRE, ORNL-TM-1796 (March 1967). 12. MSR Program Semiannu. Progr. Rep. Feb. 28, 1966, ORNL-3936, p. 122. 13. J. H. Shaffer, internal correspondence, Oct. 9, 1964, 14. W. R. Grimes, Nucl. Appl. Technol. 8, 137 (1970). 15. G. Goldberg, A. S. Meyer, Jr., and J. C. White, Anal. Chem. 32,314 (1960). 16. A. S. Meyer, R. F. Apple, and J. M. Dale, ORNL Analytical Chemistry Division. 17. W. S. Pappas, Anal. Chem. 38, 615 (1966). I8. Results summarized in a memorandum from J. G. Million (ORGDP) to R. E. Thoma, July 14, 1965. 19. P. N. Haubenreich, “Selection of Lithium for MSRE Fuel, Flush and Coolant Salt,” internal corre- spondence, Apr. 10, 1962. 20. H. F. McDuffie, “Selection of "Li Batches for MSRE Fuel, Flush and Coolant Salt,” internal corre- spondence, May 22, 1962. 21. J. H. Shaffer, personal communication. 22. P. N. Haubenreich, A Review of Production and Observed Distribution of Tritium in MSRE in Light of Recent Findings, ORNL-CF-71-8-34 (in press). 23. P. N. Haubenreich, Tritium in the MSRE: Calcu- lated Production Rates and Observed Amounts, ORNL- CF-70-2-7 (Feb. 4, 1970). 24. D. L. Manning, internal correspondence, Jan. 19, 1966. 25. C. H. Gabbard, Inspection of Molten-Salt Reactor Experiment (MSRE) Fuel Pump after Zero Power Experiments (Run 3), ORNL-CF-66-8-5 (Aug. 4, 1966). 3 CHEMICAL COMPOSITION OF THE FUEL SALT DURING NUCLEAR OPERATIONS 3 1 Introduction The MSRE was not the first reactor to use a molten fluoride fuel mixture The ARE (Aircraft Re- actor Experiment) was operated as the 1nitial demon stration of a molten-salt reactor in November 1954, using a fuel solution of NaF-ZiF,;-2?°UF, (53-41-6 mole %) ! The time during which the molten fluoride fuel was circulated was brief, ~1000 hr, and did not include a program of chemical surveillance In contrast, the MSRE was intended to provide the experience and operational data on which design and plans for molien- salt reactors of 30-year operating lifetimes could be based Plans for its operation, therefore, included a chemical surveillance program that was comprehensive n scope Moreover, chemical information and quantita- tive data pertaiming to a number of phenomena that could only be obtained from operating reactors were lacking, even though a thorough program of laboratory, engineering-scale, and in-pile tests had established the principal parameters required to design and operate the MSRE Among the most important of these are the chemical factors which control the distribution of fission products within the fuel and off-gas systems, and the redox chemistry typical of dynamuc fuel systems in which fuel burnup and replenishment proceed concur- rently It was also important to establish the corrosion resistance of Hastelloy N under realistic and typical operating conditions, to examine the capability of a molten-salt reactor to operate under circumstances where contaminants might enter the salt occasionally or chronically, and to study the possible effects that operation of numerous chemical components of the reactor might have on fuel and coolant chemustry It was the mtent of these efforts to evaluate, on a continiing basis, the adequacy of the surveilance program to assess the performance and safety of the MSRE as the prototype of a molten-salt breeder reactor Imtial activities of the program of chemucal surveil- lance served to establish that the behavior of the fuel and coolant salts would conform with that expected from laboratory and engineering tests We sought information that would indicate whether the fuel and coolant salts remained chemically stable and noncorro- sive, to what extent the consumption of fissile matenal deviated from that expected from nuclear considera- trons, and whether or not urammum losses could be detected by chemical examination of salt samples In 36 the beginning stages of MSRE operation, samples were taken on one-day intervals to assess the chemical stabiity of the fuel, to establish whether or not the concentration of oxide contamunants remained below the saturation lumts (~700 ppm for the average operating temperature of 650°C), and to correlate concomitant corrosion, if 1t occurred, with possible introduction of contaminants into the system Finally, we anticipated that through measurement of 1sotopic changes 1n the uranium composition of the fuel it might be possible to obtain accurate comparisons of the amounts of uranium burned with those expected from nuclear considerations After a significant amount of uranium had undergone fission, procedures were imtiated to monitor the rela- trve concentration of trivalent uranium 1n the fuel salt Studies of the chemical effects of uranwum fission 1n molten-fluonde muxtures had led to an early estimate that the fission reaction would be mildly oxidative and would produce ~0 8 equivalent of oxidation per gram atomic weight of urantum consumed This estimate was based on an appraisal of the yield and chemucal stability of the state of the fission products Since 1t was possible to evaluate only the equilibrium states of all the various fission product species — for example, the rare earths, iodine, tellurium, and the “noble” metals, including Mo, Nb, Ru, Ta, Re, and T¢ — nonequilibrium behavior in highly radioactive environments of the MSRE would not have been surprising and would conceivably intro- duce uncertainties into interpretation of fission product behavior 1n the MSRE Previous experience with fission product behavior had been explored almost exclusively through in-pile capsule tests, although some appraisal of the ARE fission product distribution was made 2 A vigorous 1nvestigation of the fate of fission products in the MSRE was therefore imnitiated and pursued through- out the entire period of reactor operations The results of those studies are summarized elsewhere 3 Concur- rently, laboratory studies were developed 1n support of the on-site efforts to establish the relationships of fission product distributions in the MSRE with chemi- cal parameters During the course of MSRE operations it became 1ncreasingly evident that munor variations in the oxidation-reduction potential of the fuel salt almost certainly had a real effect on some of the fission product distributions within the system and that accurate means for the determination of the concentra- tion of trivalent uranum 1n MSR’s would be needed Active work was devoted thereafter to the development of means for analysis of U3+/ZU Based on the considerations discussed above, a pro- gram of chemical surveillance was formulated and Table 3.1. Composition of the MSRE fuel salt (Weight %) (ppm)_ 3+ Date Sample Equiv. Li Be Zr F Z Fe Cr Ni U in Au (kg) L Eq. L Eq. [Uu” /Iu] Full N Obs Circula- Oxida~ Reduc~ % Pfi¥?r om. : tion tion tion Nom. 2/13/6¢ Run 5-1 4.631 209,950 0 0 3.633 0.41 2/14/66 FP 5-1 4 10.65 6.53 11.45 4,631 4,625 68.18 10Ll.44 68 51 15 0 0 3.633 0.41 2/16/66 Run 5-F 475766 Run 6-1 a 476766 FP 6-1 Concentration of oxide: 49 ppm 4/6/66 - Fuel circuit dralned because of restriction in valve No. 561. 4/8/66 - Fuel circuit filled 4/9/66 FP 6-2 11.43 6.22 10,80 4,631 4.622 68.97 102,07 77 57 63 209.950 0 0 3.633 0.41 4/13/66 FP 6-3 Sample for oxide analysis: analysis unsuccessful 4/14/66 FP 6-4 Concentration of oxide: 55 ppm® 4/15/66 FP 6-5 15 10.45 6.37 11.12 4.630 4.605 69.85 102.40 234 65 35 209.944 -0.006 -0.005 3.628 0.41 4/16/66 FP 6-6 15 10.55 6.46 11.32 4.630 4.625 68.56 101,52 101 46 49 4/17/66 FP 6-7 10.32 6.41 11.42 4,630 4,647 69,67 102.47 71 46 58 4/19/66 FP 6-8 30 10.43 6.49 11.29 4,630 4.655 68,90 101,77 89 45 58 209.938 -0.012 0.039 3.594 0.41 4/20/66 FP 6-9 39 10.48 6,66 11.52 4,630 4,684 68.92 102.26 168 58 55 209.934 -0.016 0.050 3.583 0.40 4f25/66 FP 6-10 55 10.30 6.76 11.54 4.630 4,595 68.81 102.00 129 42 36 209.928 -0.022 0.070 3.563 0.40 5/9/66 FP 6-11 115 10.30 6.42 11.28 4.630 4.612 70.09 102.70 95 59 125 209.904 =0.046 0.147 3.486 0.40 5/10/66 FP 6-12 138 Concentration of oxide: 50 ppm® 5/11/66 FP ©-13 147 10.50 6,41 11.06 4,629 4.628 66.79 99.38 101 52 54 209,891 -0.059 0.188 3,445 0.39 5/13/66 FP 6-14 166 10.50 6.45 11.33 4.629 4.617 68.18 101.08 107 48 74 209.884 ~0.066 0.212 3,421 .39 5/14/66 FP 6-15 190 10.55 b6.86 11.35 4,629 4.601 68.18 101.54 94 45 52 209.874 -0.076 0.243 3.390 0.38 5/18/66 FP 6-16 254 10,50 6.58 11.31 4.628 4.629 67.88 100,90 84 45 36 209. 848 -0,102 0.326 3.307 0,37 5/23/66 FP 6-17 322 11.44 6,64 11.05 4.628 4.652 68.26 102,04 122 49 54 209.822 -0.128 0.411 3.222 0.36 5/25/66 FP 6-18 Concentration of oxide: 47 ppm? 5/26/66 FP 6-19 400 10,40 6.88 11.72 4.627 4,667 69.10 102.77 99 39 39 209.790 -0.160 0.512 3.121 0.35 5/28/66 Run 6-F 400 4,627 209, 790 6/12/66 Run 7-1 400 4.627 209.806 6/12/66 FP 7-1 400 10.60 6.78 11.16 4.627 4.647 69.26 102,45 108 51 70 209, 806 6/15/66 FP 7-2 Sample for oxide analysis: sample unsatisfactory for analysis 6/17/66 FP 7-3 487 10,55 6.56 11.44 4,626 4.656 68.24 101.45 148 50 51 209,756 -0.194 0.623 3.010 0.34 6/20/66 FP 7-4 526 10, 50 6.40 11.61 4.626 4.640 67.82 100,97 110 52 38 209.740 -0.210 G.673 2.960 0.34 6/22/66 FP 7-5 Concentration of oxide: ppm? 6/24/66 FP 7-6 658 10.60 6,63 11.35 4.625 4,614 69,22 102.41 115 o6l 66 209.687 -0.263 0.843 2.790 0.32 6/26/66 FP 7-7 706 10.55 6.65 11.13 4.624 4.641 69.60 102.57 78 46 54 209.668 -0.282 0.904 2.729 0.31 7/1/66 FP 7-8 725 10.60 6,59 11.67 4.624 4.663 67.87 101.39 94 49 52 209.661 -0.289 0.927 2.796 0.31 7/4/66 FP 7-9 Concentration of oxide: ppm*2 7/6/66 FP 7-1C 846 10,63 6.91 11,21 4.623 4.609 68.20 101.5¢6 36 49 <20 209.612 -0.338 1.084 2.549 0.29 7/10/66 FP 7-11 944 10,45 7.00 11.22 4,622 4.630 69,48 102.78 58 39 36 209,573 -0.377 1.210 2.423 .27 7/13/66 Fp 7-12 1012 10,50 6.65 11.04 4,622 4,640 68.44 101,27 84 46 45 209.546 -0.404 1.300 2.333 0.26 7/15/66 FP 7-13 Concentration of oxide: ppm? 7/18/66 FP 7-14 1032 10.55 6.50 11.26 4,622 4.660 68.79 101.76 53 44 63 209.538 -0.412 1.321 2.312 0.26 7/20/66 FP 7-15 1047 16.55 6.71 11.60 4.621 4.638 67.59 101,08 83 44 28 209.532 -0.418 1.341 2.292 0.26 7/22/66 FP 7-16 Concentration of oxide: ppm@ 7/25/66 Run 7-F 1047 4,621 209.532 -0,418 1.341 2.292 0.26 9/25/66 Run 8-1 1047 4.603 208,711 10/8/66 FP 8-5 1047 11.10 6.23 11,21 4,603 4,643 71,21 104,39 139 108 208.711 -0.434 1.341 2.292 0.26 10/11/66 FP 8-6 1100 10,95 6.37 10.97 4,603 4,624 70.24 103,15 133 61 72 208.706 ~0.439 1,408 2,225 0.25 LE Table 3.1 (continued) (ppm) {(Weight %) " Date Sample Equiv Li Be Zr U F L Fe Cr Ni LU in AU (kg) L Eq. £ Eq. ([u” /ru] Full Nom Oos Circula- Oxida- Reduc- z nger ! ’ tion tion tion Nom. 10/13/66 FP 8-7 Concentration of oxide: 44 ppm? 10/17/66 FP 8-8 1166 10.95 6.88 11.23 4,603 4.638 69.94 103.64 219 65 35 208,680 ~-0.465 1.491 2.142 0.24 10/19/66 FP 8-9 1200 11.00 6.60 10,79 4.602 4.626 72.27 105.29 124 66 42 208,666 -0.479 1.536 2.097 0.24 10/21/66 Fp 8-10 1228 11,00 6.43 11.41 4.602 4.630 67.55 101.02 89 73 119 208.655 -G, 490 1.572 2.061 6.23 10/24/66 FP 8-11 Sample for oxide analysis: analysis unsuccessful 10/26/66 FP 8-12 1357 11.15 6.62 11.20 4,601 4.650 69.01 102.63 106 51 26 208.603 -0.542 1,738 1.895 0.22 10/28/66 FP 8-13 1380 14,25 6.49 11.29 4.601 4.623 68,82 105.47 84 63 60 208,594 -0.551 1.767 1.866 0.21 10/31/66 FP 8-14 1401 13,85 6.65 11.17 4.601 4.620 69.08 105,37 82 63 25 208.586 ~-0.559 1.793 1.840 0.21 10/31/66 FRun 8-F 1401 4.601 208.586 11/7/66 Run 9-1 4.582 207,754 11/7/66 FP 9-1 1401 10.93 6.55 11,70 4,582 4.618 68.39 102.19 140 66 50 207.754 -0.559 1.793 1.840 0.21 11/9/66 FP 9-2 1449 Concentration of oxide: 44 ppm? 11/11/66 FP 9-3 1480 10,95 6.60 10,97 4,582 4,621 69.06 102.20 140 65 58 207.723 ~0.591 1.895 1.738 0.20 11/14/66 FP 9-4 1528 [u3t/zu] = 0.10% 207,703 -0.610 1.956 1.677 0.19 11/16/66 FP 9-3 1544 10.95 6.64 10.96 4,581 4.618 67.96 101,13 145 56 26 207,697 -0.616 1.976 1.657 0.19 11/17/66 FP 9-6 Concentration of oxide: ppm 11/19/66 TFP 9-7 1662 - 6.71 10.95 4.580 4,557 68.79 176 57 74 207.650 -0.663 2.126 1,507 0.17 11/20/66 Run 9-F 1662 4.580 207.650 -0.663 2.126 1.507 0.17 12/12/66 Run 10-1I 1662 4.561 206.815 12/12/66 FP 10-3 1662 Sample for determination of organics in offgas; sample faulty 12/12/66 FP 10-4 Sample for determination of organics in offgas by Cu0; 600 ppm HF 12/14/66 FP 10-5 1662 11.10 6.47 11.00 4,561 4.608 66.65 99.83 181 58 93 206,815 -0.663 2.126 1.507 0.17 12/16/66 FP 10-6 1662 11,05 6.25 11.13 4,561 4.612 67.57 100.61 168 62 55 206,815 -0.663 2.126 1.507 0.17 12/19/66 FP 10-7 1677 11.20 6.82 11,27 4.561 4.605 66.98 100.88 174 53 65 206.809 -0.669 2,146 1.487 0.17 12/22/66 FP 10-8 1746 11.20 6.69 11.13 4,561 4.604 67.21 100.83 149 56 62 206.781 -0.697 2.235 1.398 0.16 12/24/66 FP 10-9 1794 11.05 6.51 10.8% 4.560 4,575 70.20 103.22 167 62 49 206.762 -0.716 2.296 1.337 0.15 12/27/66 FP 10-10 50 g sample for oxide analysis; no results available 12/27/66 FV 10-11G Freeze valve capsule-gas sample 12/28/66 FP 10-12 1813 11.10 6.66 11.24 4,560 4.577 67.72 101.29 91 63 48 206.754 -0.724 2.322 1.311 0.15 12/30/66 FP 10-13 1846 11.30 6.84 11.09 4.560 4,628 65.82 99.68 162 61 82 206.741 -0.737 2.364 1.269 0.15 1/1/67 FP 10-14 1870 Be addition: 3 g as powder 206.732 -0.746 2.393 1.906 0.22 1/2/67 FP 10-15 Special sample (50 g}; not subjected to chemical analysis 1/3/67 FP 10-16 1920 Be addition; 1 g as powder 206,712 -0.766 2.457 2.064 0.24 1/3/67 FP 10-17 1920 11.15 6.47 10.88 4.559 4.633 69.03 102.16 102 64 166 206.712 -0.766 2.457 2,064 0.24 1/4/67 FP 10-18 1948 Be addition; 1.63 g as rod 206,701 -0.777 2,492 2.391 0.27 1/6/67 Fp 10-19 1968 11.15 6.64 11.03 4.559 4,621 67.75 101.19 145 62 62 206.693 ~0,785 2.518 2.365 ¢.27 1/9/67 FP 10-20 2161 11.08 6.40 10.82 4,557 4,622 69.30 102,22 164 57 57 206.615 -0.863 2.768 2.115 0.24 1/11/67 FP 10-21 50 g sample for oxide determination 1/11/67 FV 10-22G Freeze valve capsule~gas sample 1/13/67 FP 10-23 2247 Be addition: 10.65 g as rod 206,581 -0.897 2,877 4.369 0.50 1/13/67 FP 10-24 2247 11,23 6.55 11.21 4.556 4.606 68.73 102.36 151 56 87 206.581 -0.897 2.8717 4.369 0.50 1/15/67 FP 10-25 2288 50 g sample for UM/IU analysis; UX/IU = 0.66%C 206.565 ~0.913 2.928 4.318 0.50 1/16/67 Run 10-F 2288 4.556 206,565 -0.913 2.928 4,318 0.50 1/28/67 Run 11-~I 2288 4.556 206,585 -0.913 2.928 4,318 0.50 1/28/67 Fp 11-1 2288 11,18 6.28 10.90 4,556 4.603 67.46 100.42 131 66 54 206,585 -0.913 2.928 4,318 0.50 1/30/67 FP 11-2 2308 11.10 6.27 10.65 4.556 4,599 69,16 101,78 112 75 63 206.577 -0.921 2.954 4.292 0.49 2/1/67 FP 11-3 2331 16.42 6.41 11.08 4,556 4.606 66.72 99.03 150 61 64 206,568 -0.930 2.983 4,263 0.49 Q 8¢ Table 3.1 (continued) (Weight %) (epm) 3+ Date Sample Equiv. Li Be Zr u ¥ L Fe Cr Ni U in AU (kg) L Eq. L Eq. [U” [EU] Full T o6 Circula- Oxida- Reduc- A Paver ome s tion tion tion No. 2/3/67 FP 11-4 2372 11.10 6.33 11.10 4.556 4,592 67.87 100.99 145 62 22 206.551 ~-0.947 3.037 4.209 0.48 2/6/67 FP 11-5 2468 50 g sample for U'/IU analysis; UI*/IU = 0.60%¢ 206.513 -0.985 3.159 4,037 0.47 2/8/67 FP 11-6 2516 11.25 6.31 10.97 4,554 4.555 67.44 100.52 131 67 33 206.494 ~1.004 3.220 4.026 0.46 2/10/67 FL 11-7 2541 11.38 6.70 11.27 4.554 4,558 69.92 103.83 172 62 50 206.484 -1.014 3.252 3.994 0.46 2/13/67 FP 11-8 2614 11.50 6.62 10.81 4.553 4.569 68.89 102.39 312 54 107 206,455 -1.043 3.345 3.901 0.45 FV 11-9G Freeze valve capsule-gas sample 2/15/67 FP 11-10 2663 Be metal addition; 11.66 g as rod 206,435 ~-1.063 3.409 6.424 0.74 2/171/67 FP 11-11 2779 10.67 6.57 10.87 4,552 4,551 69,94 102,60 165 73 43 206.389 -1.109 3.557 6.276 0.72 2/21/67 FP 1i-12 2811 10.93 6.27 10.88 4,552 4,567 69.96 102.60 76 75 34 206.376 -1.122 3.598 6.235 0.72 2/22/67 FP 1i-13 2835 50 g sample for U¥t/IU analysis; UX/IU = 0.69%C 206.366 -1.132 3.631 6.202 0.71 2/24/67 FP 11-14 2884 10.98 6.35 11.26 4,551 4,525 67.83 100.95 98 78 75 206.347 -1,151 3.691 6.142 0.71 FP 11-14 2884 4,551 4.539 206.347 -1.151 3.691 6.142 0.71 2/28/67 FP 11-15 2950 11.11 6.49 10.74 4,551 4,552 68.89 101.78 71 62 37 206,321 -1.177 3.775 6.055 0.70 2/28/67 FV 11-16G Freeze valve capsule-gas sample 3/1/67 FP 11-17 2994 11.33 6.47 11.21 4,550 4,553 70.25 103.81 67 58 47 206,303 -1,195 3.833 6.000 0.69 3/2/67 FP 11-19 3014 10.73 6.67 11.17 4.550 4,576 68.68 101.83 120 56 43 206.295 =1,203 3.858 5.975 0.6% 3/3/67 FP 11-19 3038 10.47 6.57 11.11 4.550 4.589 67.12 99. 86 117 68 42 206.285 -1.213 3.890 5.943 0.68 3/6/67 FP 11-20 3109 10.51 6.72 10.98 4.549 4.561 67.60 100.37 122 59 49 206,257 -1.241 3.980 5.853 0.67 3/8/67 FP 11-21 3126 10.50 6.36 10.89 4,549 4.576 66,83 99.16 168 63 46 206.250 -1,248 4,003 5.830 0.67 3/9/67 FP 11-22 3126 10.55 6.57 10.92 4.549 4,572 66.05 98.66 104 62 63 206.250 -1.248 4,003 5.830 0.67 3/10/67 FP 11-23 3126 10,53 6.49 10.95 4,549 4,583 69.70 102.25 136 63 55 206.250 ~-1.248 4,003 5.830 0.67 3/13/67 FP 11-24 3169 10.55 6.51 10.90 4,549 4,547 68,50 101.01 173 67 63 206,233 -1,265 4,057 5.776 G.66 3/16/67 FP 11-25 50 g sample for oxide analysis; analysis unsuccessful 3/20/67 FP 11-26 335% 10.43 6.49 10,91 4,547 4,570 67.17 99.57 118 52 53 206,157 -1.341 4.301 5.532 0.64 3/21/67 FP 11-27 3380 10.52 6.85 10.85 4,547 4,577 66,04 98,84 72 63 80 206,149 -1.349 4,327 5.506 0.63 3/21/67 FP 11-28 Concentration of oxide: 58 ppm 3/22/67 FP 11-29 3407 10.48 6.46 10.92 4,546 4,584 64,62 97.06 126 64 56 206.138 -1.360 4,362 5.471 0.63 3/23/67 FP 11-30 3460 10.48 6.39 11.02 4.546 4,597 64,51 96,99 115 66 71 206,117 -1.381 4,429 5.404 0.62 3/27/67 FP 11-31 3524 10.53 6.58 11.06 4,546 4.559 67,06 99.79 80 64 50 206.092 -1,406 4,509 5.324 0.61 3/28/67 FP 11-32 13548 50 g sample for UM/IU analysis; U3*t/IU = 0.045%C 206.082 -1.416 4.541 5.292 0.61 3/29/67 FP 11-33 3571 10,53 6.43 11,14 4,545 4,567 66.38 99,04 142 72 72 206.073 -1.425 4,570 5.263 0.61 3/31/67 FP 11-34 3619 10.55 6.33 11.37 4,545 4,582 68,99 101.82 146 64 64 206.054 =1.444 4,631 5.202 0.60 4/3/67 FP 11-35 3690 10.53 6.35 11,12 4.544 4,566 67.24 99.81 194 73 64 206,025 =1.473 4,724 5.109 0.59 4/4/67 FV 11-36G Freeze valve sample; capsule penetrated by leaching solution 4/5/67 FP 11-37 3739 10.55 6.33 10.75 4,544 4,541 65.93 96,10 79 80 49 206.006 =1.492 4.785 5.048 0.58 476/67 FP 11-38 3763 50 g sample for U/IU analysis; U3t/LU = b 205.996 -1.502 4.817 5.016 0.58 4/7/67 FP 11-39 3856 11.57 6,44 10.92 4.543 4.536 66,55 100,02 182 69 52 205.959 -1.539 4.936 4.897 0.56 4/10/67 FP 11-40 3856 Be addition: 8.40 g as rod 205.959 -1.539 4.936 6.761 0.78 4/10/67 FP 11-41 3856 10.42 6.37 10.77 4.543 4,579 68.58 100.72 135 56 58 205.959 -1.539 4,936 6.761 0.78 4/11/67 FV 11-42G Freeze valve sample - wt gain: O g Pump off 40 min. Sampler port 2" above salt surface 4112767 FP 11-43 3904 50 ¢ sample for /Ly analysis; sampler machined improperly; no results obtained 414767 FP 11-44 3937 10.50 6.60 11.01 4.542 4.561 69.88 102.55 140 59 44 205.927 -1.571 5.038 6.659 0.77 4/17/67 FP 11-459 4007 10.58 6.50 10.65 4,541 4,548 67.13 99.41 88 54 41 205.899 -1.599 5.128 6.569 0.77 4/18/67 FV 11-46G Freeze valve sample. Pump on. Sampler port above salt surface &/21/67 FP 11-47 4167 10.95 6.48 10.96 4,558 4.604 66.65 99.64 169 71 75 206.678 -1.639 5.257 6.440 0.74 L/24/67 FP 11-48 4172 10.45 6.52 10,85 4,558 4,578 67.23 99.63 210 49 58 206,652 -1.665 5.340 6.357 0.73 4/25/67 FP 11-49 4196 50 g sample for UX/ZU analysis; »>2000u moles HF 206.642 ~1.675 5.372 6.325 0.73 6¢ Table 3.1 (continued) (Weight %) (ppm) 3+ Date Sample Equiv Li Be Zr U F ) Fe Cr Ni LU in aU (kg) I Eq. Z Eq. {U” /3y] Full o Ohe Circula- Oxida~ Reduc- A nger ’ ’ tion tion tion Nom. 4/26/67 FP 11-50 Tandem graphite specimens - wire showed 30 X activity at salt-gas interface 4/28/67 FP 11-51 4274 11.20 6.45 10.97 4,557 4,571 69.61 102,80 114 6l 61 206.611 -1.706 5.471 6,226 0.71 5/1/67 FP 11-52 4341 11.33 6.45 10.79 4.556 4,566 69.27 102.41 80 60 25 206.584 -1.733 5.558 6.139 0.70 5/3/67 FV 11-53G 4388 Freeze valve sample. Pump on, level low (51%) 206,566 ~-1.751 5,616 6.081 0.70 5/5/67 FP 11-54 4436 16.93 6.63 10.95 4.556 4.551 69.91 158 61 63 206,547 -1.770 5.677 6.020 0.69 5/8/67 FP 11-55 4507 50 g sample for use in hot cell experiments 206.518 -1.799 3.770 5.927 0.68 5/9/67 FP 11-56 50 g sample for oxide analysis - poor results obtained; estimate: 50-150 ppm 5/10/67 FP 11-57 Sampler was found to be empty 5/10/67 FP 11-58 4513 10.48 6.66 11.27 4,556 4,607 70.85 103.97 131 81 42 206.516 -1.801 5.777 5.920 0.68 5/10/67 Run 11-F 4513 4.556 206.5106 -~1,801 5.777 5.920 0.68 6/19/67 Run 12-1 4513 4,536 205.679 6/19/67 FP 12-5 45,13 11.2 6.74 10.94 4,536 4.550 66,32 99,75 123 52 60 205.679 -1.801 5.777 5.920 0.68 6/21/67 FP 126 4513 50 g sample for U3*/ZU analysis U3t/Lu = 0.71%C 205.679 -1.801 5.777 5.920 0.68 6/21/67 ¥V 12-7G Freeze valve sample for Run 12 baseline data 6/21/67 ¥P 12-8 Be addition: 7.933 g as rod 205.679 -1.801 5.777 7.676 0.89 6/23/67 FP 12-9 4558 Be addition: 9.840 g as rod 205.661 -1.819 5.834 9.807 1.13 6/26/67 FP 12-10 4602 11.6 6.91 10.57 4,536 4,525 67.27 110.87 134 71 72 205,643 -1.837 5.892 9.749 1.12 6/29/67 FP 12-11 4674 50 g sample for U3T/SU analysis U3t/iu = 1.3%¢ 205.615 -1.865 5.981 9.660 1.10 6/30/67 FP 12-12 4706 11.6 6.54 10.91 4,535 &4.545 66.66 100,76 113 64 62 205.602 -1.878 6.023 9.618 1.11 7/3/67 FP 12-13 4739 Be addition: 8330 g as rod 205.589 -1.891 6.065 11.424 1.32 /57167 FP 12-14 4787 11.5 6.50 11.22 4.534 4,557 67.95 101.73 145 82 47 205.569 -1.911 6.129 11. 360 1.31 7/6/67 FP 12-15 4811 Be addition: 11.677 as rod 205.560 -1.920 6.158 13.922 1.61 7/7/67 FP 1216 4836 11.4 6.40 10.62 4.534 4,567 68.27 101.26 269 110 68 205.550 -1,930 6.190 13.890 1.60 7/10/67 FP 12-17 4885 11.3 6.40 10.66 4.533 4,532 66.92 99.81 216 144 53 205.530 -1.950 6.254 13,826 1.60 7/11/67 FP 12-18 4909 Concentration of oxide 57 ppm 7/11/67 FP 12-19 4917 11.5 6.19 11.00 4.533 4,522 65.05 98.26 100 102 62 205,518 -1,962 6.292 13.788 1.59 7/12/67 FP 12-20 4933 10.6 6.36 10.76 4.533 4,557 65.76 98.04 81 64 L4 205.511 -1.969 6.315 13,765 1.58 7/13/67 FP 12-21 4981 50 ¢ sample for UX/LU analysis: U3H/IU = 1.0%C 205.492 -1.988 6.376 13.704 1,58 7/13/67 FP 12-22 4981 10.5 6.52 10,43 4,532 4,566 66.18 98,19 247 90 50 205,492 -1.988 6.376 13.704 1.58 7/14/67 FP 12-23 4998 10.6 6.68 10.58 4,532 4,526 65.26 97.64 154 78 76 205,485 ~1.995 6.398 13.682 1.58 7/15/67 FP 12-24 5022 11.38 6.36 10.67 4.532 4,567 66.46 99.44 176 67 56 205.476 2,004 6,427 13.653 1.58 7/16/67 FP 12-25 5046 16.7 6.46 10.53 4.532 4,496 66,10 98.29 208 68 62 205.466 -2.014 6.459 13.621 1.57 7/17/67 FV 12-26G Freeze valve capsule-gas sample 7/17/67 FP 12-27 5070 10.7 6.61 10.78 4,532 4,550 67.50 100.14 195 75 72 205.457 -2.023 6.488 13.592 1.57 7/18/67 FP 12-28 5097 16.7 6.44 10,66 4,531 4.569 69.00 101.37 150 68 52 205. 446 -2.034 6.523 13.557 1.56 7/19/67 FP 12-29 5121 16.7 6.41 10.87 4,531 4,529 67.11 9G.62 177 84 78 205.436 -2.044 6.555 13.525 1.56 7/19/67 FP 12-30 U Addn. caps. no. 99 7/19/67 FP 12-31 U addn. caps. No. 100 7/20/67 FP 12-32 U addn. caps. no. 101 7/20/67 FP 12-33 U addn. caps. no. 102 7/10/67 FP 12-34 U addn. caps. no. 103 7/21/67 FP 12-35 U addn. caps. no. 104 4,543 205.983 =2.044 6.555 13.525 1.56 7/21/67 FP 12-~36 5169 10.5 6.68 10.96 4,543 4.554 66.40 99.09 156 84 129 205.964 -2.063 6.616 13.464 1.55 7/21/67 FP 12-37 U addn. caps. no. 105 7122767 FP 12-38 U addn, caps. no. 106 7/22/67 FP 12-39 U addn, caps. no. 107 7/22/67 FP 12-40 U addn. caps. no. 108 4,551 206,328 -2.063 6.616 13.464 1.55 oy Table 3.1 {continued) (Weight %) (ppm) 3+ Date Sample Equiv. Li Be Zr U F I Fe Cr Ni. iU in AU (kg) I Eq. L Eq. [U” /EU] Full o ohs Circula- Oxida- Reduc- A Power * ’ tion tion tion Nom. 7/23/67 FP 12-41 5229 10.6 6.52 10.68 4.550 4,562 68.00 100.36 110 64 192 206.304 -2.087 6.693 13.387 1.54 7/23/67 FP 12-42 U 6.52 capsule No. 109 7/23/61 FP 12-43 U addn. caps. no. 110 FP 12-44 U addn. caps. no. 111 FP 12-45 U addn. caps. no, 112 FP 12-456 U addn, caps. no. 113 4.560 206, 760 ~2.087 6.693 13,387 1.54 7/25/67 FP 12-47 5266 10.6 6.50 10.50 4.560 4,586 66.34 98.33 120 72 70 206,745 ~2,102 6.741 13.339 1.53 FP 12-438 U addn. caps. no. 114 FP 12-49 U addn, caps. no. 115 FP 12-50 U addn. caps. no. 116 4.566 207.020 -2.102 6.741 13.339 1,53 7/26/67 FP 12-51 5296 11.22 6.60 10.32 4.566 4,588 66,32 99.05 94 72 39 207.008 ~-2.114 6.780 13.300 1,52 7/28/67 FP 12-52 5366 10.7 6.47 10.7F 4.565 4,594 65.98 98.45 119 72 92 206.980 -2.142 6.870 13.210 1.51 1731767 FP 12-53 5433 10.5 6.39 10.66 4,565 4.503 62,48 97.53 182 72 60 206.954 -2.168 6.953 13,127 1.50 8/1/67 FP 12-54 5433 50 g sample for isotopic analysis 8/2/67 FP 12-55 5463 4,564 206,942 -2.180 6.991 13.089 1.50 10.75 6.44 11,12 4.564 4,577 65.72 98.61 156 72 300 206.942 10.80 6.42 10.95 4,564 4.575 64.57 97.32 136 58 720 206.942 8/3/67 FP 12-56 5492 Be addition: 9.71 g as rod 206.930 -2.192 7.030 15.205 1.74 8/3/67 FP 12-57 5496 4,564 206.928 ~2.194 7.036 15.199 1.74 8/4/67 FP 12-58 5500 11.33 6,59 11.08 4.564 4.587 65.14 98.73 156 54 170 206,927 -2.195 7.040 15,195 1.74 11.3¢ 6.74 11.28 4.564 4.549 66.53 170.40 160 64 424 206.927 B/4/67 FP 12-59 5500 11.00 6.58 10.77 4.564 4.600 66.62 99,57 138 74 66 206.927 ~-2,195 7.040 15.195 1.74 8/5/67 Run 12-F 5500 Sampler cable severed 9/15/67 Run 13-I 5500 4,543 205.984 -2.195 7.040 15,195 1,74 9/15/67 FP 13-4 5568 10.98 6.90 11.21 4,543 4,552 65,88 99.55 141 82 82 205.957 -2,222 7.126 15.109 1.74 9/15/67 FP 13-5 5568 50 g sample for U3+/IU; u3*/zy = 1.60%¢ 205.957 -2,222 7.126 15,109 1.74 9/18/67 FP 13-6 5666 10.95 6.42 10,78 4.542 4,587 67.01 99.77 118 66 76 205.918 -2.261 7.25% 14,984 1.72 9/18/67 Run 13-F 4,542 205,918 9/20/67 Run 14-I 4.542 205.924 ~-2,261 7.251 14,984 1.72 3/20/67 FP 1l4~1 Sample capsule suspended in empty pump bowl for 10 minutes 9/21/67 FP 14-2 No sample obtained; sampler found to have no sample port. 9/22/67 FP 14-3 50 g sample for isotopic analyses 9/25/67 FP l4-4 5757 10,70 6.32 11.52 4.541 4.577 66,16 99.30 104 76 46 205.887 -2.298 7.370 14,865 1.71 9/26/67 FP 14=5 50 g sample for isotopic analyses 9727767 FP 146 50 g sample for isotopic amalyses 9/28/67 FP 14-7 5828 10.70 6.24 10.98 4.540 4.572 66.19 98.71 102 79 63 205.859 -2.326 7.460 14.775 1.70 10/2/67 FP 14-8 5923 10.80 6.13 10.98 4.540 4,584 66,54 99.12 98 74 68 205.821 ~2.364 7.582 14,653 1.69 10/3/67 FP 14-9 5933 Sample for oxide analyses; no results obtained 10/5/67 FP 14-10 5995 10.90 6.74 11.40 4.539 4,587 65,48 99.13 150 80 76 205.792 -2.393 7.675 14.560 1.68 10/9/67 FP 14-11 6086 10.80 6.20 11.06 4.538 4,579 65.18 97.78 128 88 66 205,756 -2.429 7.790 14,445 1.66 10/12/67 ¥P 14-12 6155 10.90 6.88 11.01 4,538 4,579 66.82 100,22 176 60 60 205.728 ~-2.,457 7.880 14,355 1.65 10/16/67 FP 14-13 6249 10.50 6.36 11.08 4.537 4,575 65.73 98.27 146 65 57 205.691 -2.494 8.000 14.235 1.64 10/20/67 ¥P lé-14 6344 10.65 6.52 11.25 4,536 4,480 66,88 99.81 110 80 112 205.653 -2.532 8.120 14.115 1.63 10/23/67 FP 14-15 50 g sample for hot cell experiments 10/24/67 FP 14-16 6440 10.80 6.30 11.28 4.535 4,556 65.75 98.72 158 76 105 205.615 -2.570 8.242 13.993 1.61 10/26/67 FP 14-17 6440 10.80 6.26 10.81 4.535 4,557 66.14 98.59 124 80 55 205.615 -2.570 B.242 13,993 1.61 10/30/67 FP 14-18 6506 10.53 .00 11.00 4.535 4.553 65.24 97.35 112 70 62 205.594 ~2.591 8.310 13.925 1.61 184 Table 3.1 (continued) (Weight %) (ppm) 3+ Date Sample Equiv. Li Be Zr U F L Fe Cr Ni ZU in AU (kg) I Eq. L Eq. [U”/1U] Full Circula- Oxida~- Reduc- A Pgief Nom, Obs. tion tion tion Hom. 11/3/67 FP 14-19 6604 10,73 5.92 11.25 4,534 4.553 66.68 99.13 108 70 51 205.549 -2.636 8.454 13.781 1.59 11/3/67 FV 14-208 Freeze valve capsule - salt sample 11/6/67 FP 14-21 6678 10.60 5.94 10.69 4.533 4,557 65.88 97.69 128 76 61 205.520 ~2.665 8.547 13.688 1.58 11/7/67 FP 14-22 6702 10.60 5.86 11.32 4.533 4,529 66.74 99.07 106 72 74 205.510 -2.675 8.580 13.655 1.58 11/8/67 FP 14-23 6726 10.43 6.04 11.29 4.532 4.534 66.89 99.21 94 68 56 205.501 -2.684 8.608 13.627 1.57 11/9/67 FP 14-24 6751 10.63 6,18 11.34 4,532 4.567 66.42 99,17 169 72 72 205,491 -2.694 8.640 13.595 1.57 11/13/67 FP 14-25 6848 10.40 6.12 10.92 4.531 4,548 66.19 98,20 124 72 62 205.452 ~2.733 8.765 13,470 1.55 11/14/67 FP 14-26 6872 10.63 6.10 11.00 4.531 4,572 66.34 98.66 103 66 60 205,442 -2.743 8.797 13.438 1.55 11/27/67 FP 14-27 6913 10.20 6.19 11.12 4,531 4,553 66.53 98.62 126 72 69 205,426 -2.759 8.848 13,387 1.54 11/30/67 FP 14-28 7020 10,53 6.19 11,18 4.530 4,589 66.79 99.31 135 72 60 205,383 -2.802 8.986 13.249 1.53 12/4/67 FP 14-29 7142 10.40 6.34 10.88 4.529 4,564 66.42 98,63 121 10 12 205.335 ~2.850 9.140 13.095 1.51 12/5/67 FV 14-30S Freeze valve capsule - salt sample 12/6/67 FP 14-31 50 g sample for use in hot cell experiments 12/11/67 FP 14~32 7313 10.40 6.45 10,94 4.527 4,550 66.52 98.88 120 70 58 205,266 -2.919 9.362 12,873 1.49 12/12/67 FP 14-33 10 g sample exposed to atmosphere in Line 928, 1.5'ft. above pump bowl 12/13/67 FP 14-34 10 g sample exposed to atmosphere in Area 1C for 10 min. 12/14/67 FP 14-35 50 g sample for isotopic analysis 12/15/67 FP 14-36 50 g sample for isotopic analysis 12/18/67 FP 14-37 50 g sample for isotopic analysis 12/19/67 FP 14-38 7464 10.80 6.38 10.56 4.526 4,561 66.06 08.39 174 82 96 205.206 -2.979 9.554 12.681 1.47 12/20/67 FP 14-39 Concentration of oxide: 46 ppm 12/26/67 FP 14-20 7597 10.30 6.28 10.87 4,524 4.531 66.31 98.32 113 79 69 205.103 -3.032 9.724 12,511 1.45 1/2/68 FP 14-41 7732 10.30 6.34 11.12 4.524 4,536 66,23 98,55 134 84 59 205.099 ~3.086 9,897 12.338 1.43 1/8/68 FP 14-42 7799 10.50 6.10 11.16 4.523 4,531 66.41 98.73 148 87 73 205.072 -3.113 9.984 12,251 1.42 1/11/68 FP 14-43 7849 10.10 6.46 10.96 4.523 4,524 66.36 98,33 118 T4 70 205.052 -3.133 9.984 12.251 1.42 1/15/68 FP 14-44 7917 9.80 6.42 10,45 4.522 4,466 66.51 97.67 120 9c 68 205.025 -3.160 10,13 12,105 1,40 1/16/68 FP 14-45 7939 50 g sample for USt/LU analysis; 205.016 -3,169 10.16 12.075 1.40 1/17/68 FP 14-46 7953 10.20 6.50 11.16 4,522 4.516 67.06 99.47 152 84 72 205.011 -3.174 10.18 12.055 1.39 1/18/68 FP 14-47 7970 10.50 6.40 11.01 4,522 4,536 66.78 99.25 136 81 53 205.005 ~3.180 10.20 12.035 1.39 1/22/68 FP 14-48 8038 10.25 6,42 ig.92 4.521 4,557 66.96 99,13 111 0 51 204.977 -3.208 10.29 11,945 1.38 1/25/68 FP 14~49 8040 10.28 6.56 10.85 4.521 4.538 66.92 99.17 108 74 56 204,976 -3.209 10.29 11.945 1.38 1/29/68 FP 14-50 8040 10.20 6,38 11,38 4,521 4.541 66.24 98,76 121 79 70 204.970 -3.209 10.29 11.945 1.38 2/1/68 FP 14-51 8057 10.3 6,70 10.96 4.521 4.54]1 66.43 98.96 132 81 73 204,969 -3.216 16.31 11.925 1.38 2/5/68 FP 14-52 8129 10.15 6.56 10.90 4.520 4,541 66.57 98.74 110 68 48 204,941 -3.244 10.40 11,835 1.37 2/6/68 FP 14-53 Concentration of oxide: 58 ppm 2/8/68 FP 14-54 8186 9.88 6.03 10.38 4.520 4. 446 66.58 97.35 105 80 50 204.918 -3.267 10.48 11.755 1.36 2/9/68 FP 14-55 Ni rod suspended in Ni basket - 2 hours 2/12/68 FP 14-56 8236 16.28 6.36 11.08 4.519 4.533 66.53 98.84 117 90 68 204,898 ~-3.287 10.54 11.695 1.35 2/13/68 FP 14-57 50 g sample for UX*/IU analysis; UH/SU = ~0.35% 2/15/68 FP 14-58 8305 10,30 6.40 106.93 4.515 4,529 66.70 98.89 94 82 58 204.870 -3.315 10.63 11,605 1.34 2/19/68 FP 14-59 8401 10.23 6.58 10.7¢ 4.518 4,527 65.93 98.09 101 82 42 204.832 -3.353 10.75 11.485 1,33 2/20/68 FP 14-60 50 g sample for F. P. experiments - S. S. Kirslis; 10 g obtained 2/22/68 FP 14-61 8456 - 6.21 10.96 4.517 - 66,63 - 74 204,810 -3.375 10.82 11.415 1.32 2/26/68 FP 14-62 50 g sample for use in hot cell experiments 2727768 FV 14-638 Freeze valve capsule - salt sample 80 2/28/68 FP 14-64 8602 10.30 6.24 10.80 4.516 4,552 66.04 97.94 144 80 56 204.752 =-3.433 11.01 11.225 1.30 3/4/68 FP 14-65 8697 10.2G 6.40 11.08 4.515 4.510 66.58 98.80 63 83 36 204,714 ~3.471 11.13 11.105 1.29 cr Table 3.1 (continued) | . (Weight %) {ppm) Date Sample Fquiv. Li Be Zr U F ¥ Fe Cr Ni Pu_ LU in AU (kg) £ Eq. L Eq. ¢[U3F/zu) Full Nom Obs Circula- Oxida- Reduc- y Power ’ ’ tion tion tion Nori. 3/5/68 FV 14~-665 Freeze valve capsule - salt sample 3/6/68 FV 14-67G Freeze valve capsule - gas sample 3/7/68 FP 14-68 8742 10,09 6.24 1G.85 4,515 4.474 66.41 98,10 118 78 72 204,696 -3.489 11.19 11.045 1.28 3/8/68 FP 14-69 50 g sample for use in hot cell experiment 3/11/68 FP 14-70 Malfunction of sampler; no sample obtained 204.588 -3,594 11.59 10.645 1.23 3/28/69 FP 14-71 50 g sample capsule; no sample obtained 3/28/68 Run l4-F 9005 -3.615 8/18/68 FST-19 50 g sample for UF3 analysis (after transfer to FST) 8/19/68 FST-20 Isotopic analysis for ETA measurement 8/19/68 FST-21 Isotopic analysis for ETA measurement 8/19/68 FST-22 Isotopic analysis for ETA measurement 8/19/68 FST-23 Isotopic analysis for ETA measurement 8/19/68 FST-24 Isotopic analysis for ETA measurement 8/21/68 FST~25 6.92 1054 4,306 131 170 36 8/21/68 FST-26 6.80 10.46 4,293 132 164 35 8/29/68 FST=27 26 ppm 400 420 840 at finish of Fy 9/2/68 FST-28 8 ppm 430 420 1540 filtered; after hr I, 9/3/68 FST-29 3 ppm 400 420 530 filtered; after 33 hr Hy 9/4/68 FST-30 11.42 4 ppm 360 440 530 not filtered 380 460 180 9/4/68 FST-31 50 g sample for oxide analysis 9/6/68 FST-32 11.52 2 ppm 110 100 10 9/7/68 FST-33 no results; filter end empty 9/14/68 FP 15-5 10 g sample for analysis of total concentration of reductants 9/14/68 FP 15-6 10.40 6.86 11.11 0.513 0.516 64,38 93.30 140 42 46 113 21.887 - - - 0 9/15/68 TP 15-7 Be addition: 10.08 g as powder - - 2.24 0 Scrapings of metal deposit from Be cage 12,0% 0.11% 8.03% - 9/15/68 FP 15-8 10 g sample for total reducing power; lost by explesion in 1ab %/17/68 FP 15-9 10.70 6.74 11,40 0.663 0.648 69,32 98.84 131 50 75 116 28.373 +6.486 9/19/68 FP 15-10 10.45 6.44 11.28 0,768 0.764 67.18 96.13 125 29 60 102 32,862 +4. 489 9/20/68 FP 15-11 Capsule No. 30 Enrichment No, 1 10/1/68 FP 15-12 ~ 6.18 11.07 0.770 0.764 66.68 - 109 31 70 120 32,952 +0.090 10/2/68 FP 15-13 Capsule No, 28 Enrichment No, 2 10/2/68 FP 15-14 Capsule No. 25 enrichment No., 3 10/2/68 FP 15-15 Capsule No. 26 Enrichment No. 4 10/3/68 FP 15-16 Capsule No. 24 Enrichment No. 5 10/3/68 FP 15-17 Capsule No. 23 Enrichment No, 6 10/5/68 FP 15-18 10.63 6.62 10.94 0.780 0.822 67.61 99.66 159 36 82 106 33,430 +0.478 10/6/68 FP 15-19 Capsule No. 20 Enrichment No, 7 10/6/68 FP 15-20 Capsule No. 19 Enrichment No. 8 10/6/68 FP 15-]1 Capsule No. 27 Enrichment No, 9 1G/7/68 FP 15-22 Capsule No. 29 Enrichment No. 10 10/8/68 FP 12-34 Capsule No. 16 Earichment No. 11 10/9/68 FP 15-24 Capsule No. 21 Enrichment No, 12 134 Table 3.1 (continued) (Weieght %) (ppm) 34 Date Sample Equiv. Li Be Zr U F Z Fe Cr Ni Pu ZU in AU (kg) I Eq. I Eq. [U7/LZu] Full Circula- Oxida- Reduc- i Poggr Nom. Obs. tion tion tion Nok 16/ /68 FP 15~25 10 g sample for analysis of total concentration of reductants: % = 0.6%, [UST]/[LU] equivalent: 0.097% 16/10/68 FP 15-26 11.60 6.72 10.95 0.794 C.804 71.36 102,24 140 50 51 34.009 +0.579 16/12/68 FP 15-27 Capsule No. 22 Enrichment No. 13 34.009 10/12/68 FV 15-28% Freeze valve capsule - salt sample; no sample obtained 10/12/68 FV 15-29G Freeze valve capsule - gas sample 10/13/68 FP 15-30 Be addition; 8.34 g as rod - - - 0 Scale from Be rod 36.5% 50.3% 13.2% Scale from Ni cage 78.6% 4.0% 17.4% 16/15/68 FP 15-318 Capsule No. 35 Enrichment Neo. 14 34.189 +0,090 10/15/68 FV 15-328 11.75 6.44 11.17 0.795 0.785 - - 134 59 - 34,189 - 10/17/68 FP 15-33 11.55 6.36 11.16 0.795 0.797 68,36 98,23 158 56 61 133 34,189 - 106/18/68 FP 15-34 Capsule No. 42 Enrichment No. 15 10/19/68 FP 15-35 Capsule No. 18 Enrichment No. 16 10/19/68 FP 15-36 Capsule No. 17 Enrichment No. 17 10/20/68 FP 15-37 Capsule No. 36 Enrichment No. 18 10/20/68 Fuel circuit drained 10/23/68 Fuel circuit filled 10/23/68 FP 15-38 11.45 6.76 11.45 0.804 0.797 66.72 97.21 154 68 43 146 34,443 +0,571 - - 0 10/26/68 FP 15-3%@ 0.804 0.795 10/28/68 FP 15-40 12.37 - - 0.804 0.820 - - 676 139 74 10/28/68 FP 15-41 Capsule No. 41 Enrichment No,19 34.533 +0.090 - - ¢ 16/29/68 FV 15-428 - 6.13 11,09 0.804 0.779 - - 132 60 143 135 34.533 - = - ¢ 10/29/68 FV 15-43G Freeze valve capsule - gas sample 10/30/68 FV 15-44 Capsule No. 38 Enrichment No. 20 10/30/68 FP 15-45 Capsule No. 34 Enrichment No. 21 10/31/68 FP 15-46 Capsule No. 45 Enrichment No, 22 10/31/68 FP 15-47 Capsule No. 39 Enrichment Ne. 23 11/4/68 FP 15-48 Capsule No. 32 Enrichment No. 24 11/4/68 FP 15-49 Capsule No. 31 Enrichment No. 25 11/5/68 FP 15-50 Capsule No, 37 Enrichment No. 26 11/6/68 Fv 15~518 - - 11.18 0.821 0.798 35.144 - 11/6/68 FV 15-52G Freeze valve capsule -~ gas sample 11/6/68 FP 15-53 Cu capsule containing a magnet; exposed to salt 5 min. 11/9/68 FP 15-54 Capsule No, 44 Enrichment No. 27 35.234 +0.0%0 0 11/11/68 FP 15-55 11.30 6.53 10.99 0.823 0.813 68.00 97.68 181 80 140 35.234 +0.070 - - 0 11/11/68 FP 15-56 Capsule Wo. 32, from PF 15-48, for laboratory tests 11/12/68 FV 15-57S Freeze valve capsule ~ salt sample filtrate 0.823 0.736 2510 1290 340 residue 0.823 0.794 183 76 76 11/12/68 FV 15-58G Freeze valve capsule - gas sample 11/13/68 FP 15-59 Segmented magnets - suspended in salt 1 hr; pump off 11/15/68 FP 15-60 11.25 6.64 11,15 0.823 0.828 68.50 98.41 176 62 31 148 11/15/68 FP 15-61 Segmented magnets - suspended in salt 1 hr; pump on 11/15/68 FP 15-62 Be addition: 9.38 g as rod 0 Scale from Ni cage cap 4.72% 3.26% 4.71% Scale from Ni cage body 2.74% 4.53% 0.10% 11/16/68 FP 15-63 11.50 6.50 11.04 0.823 0.818 69.52 99.42 143 62 46 153 I 4 4% Table 3.1 (continued) (Weight 7) (ppm) + Date Sample Equiv. Li Be 7T U F n Fe Cr Ni LU in AU (kg) I Eq. L Eq. [U3 JiU] Full Nom obs. Circula- Oxida~ Reduc- % Power hx ‘ tion tion tion Nom. 11/19/68 FP 15-64 Magnet No. 4 (segmented) immersed 4 times 11/20/68 FP 15-65 11.80 6.99 11.23 0.823 0,843 66.91 97.90 98 63 52 171 11/20/68 FP 15-66 Segmented magnets, separated by Be metal spacers. Be addition: 1.0 g. 11/21/68 FP 15-67 Magnet No. 6; immersed 5 min. 11/22/68 FP 15-68 11.68 6.73 11.02 0.823 0.816 68.70 98.99 148 62 26 168 11/25/68 FV 15-69G Freeze valve capsule - gas sample 11/26/68 FP 15-70 50 g sample for oxide analysis; specimen not usable 11/27/68 FV 15-71S Freeze valve capsule - galt sample 11/28/68 Run 15-Ff 12/11/68 Run 16-1 12/12/68 FP 16-1 Magnet capsule 12/13/68 TFP 16-2 Magnet capsulel2/13/68 FV 16-3S 12/13/68 FV 16-38 12/16/68 FV 16-4S ~ 6.85 10.90 0.823 0.807 67.88 - 152 84 30 12/16/68 Run 16-F 1/12/69 Run 17-1 t] 1/12/69 P 17-1 G 11.35 6.39 10,94 0.823 0.780 70.63 100,09 116 52 57 140 35.234 4] G 0 0 1/14/69 Fv 17-28 2 - 7.02 11.08 0.823 0.812 66,18 - 122 68 53 35,233 -0.0005 1/16/69 FP 17-3 84 50 g sample for oxide determination - overheated - not usable 35.207 1/21/69 FP 17-4 11.58 6.82 11,32 0.822 0.812 67.89 98.46 112 62 56 144 35.207 -0.0266 1/21/69 Fp 17-5 50 g sample for oxide determination: 61 ppm 1/22/69 FV 17-68 Freeze valve capsule ~ galt sample 1/22/69 FV 17-75 124 Freeze valve capsule - salt sample 35.195 -0.039 0.13 -0.13 -0.09 1/22/69 FV 17-8 124 Be addition: 8.57 g as rod 1.77 1.17 1/27/69 FP 17-9 148 11,35 6.88 11.08 0.822 0.823 66.64 96.97 149 58 64 158 35.187 ~0.047 0.15 1,75 1.16 1/28/69 ¥V 17-105 172 6.68 10.3% 0.822 0.598¢ 162 128 270 35.180 -0.053 0.17 1.73 1.15 1/30/69 FP 17-11 220 cr® rod exposed to fuel for 6.5 hr, 4.73 g dissolved (0.18 eq.) 35.165 -0.070 0.23 1.85 1.23 2/6/69 FP 17-12 374 11,50 6.58 10.71 0.820 0.815 70.09 99.74 148 78 60 145 35.116 -0.118 0.38 1.70 1.13 2/8/69 FP 17-13-16 50 g samples for mass spectrometric analysis 2/10/69 FV 17-17G Freeze valve capsule - gas sample 2/12/69 FP 17-18 467 11.60 6,90 11.14 0.819 0.820 67.63 98.13 163 70 54 156 35.087 -0.148 0.48 1.60 1.07 2/19/69 FP 17-19 543 11,40 7.00 11.12 0,819 0.817 68.50 98.88 125 70 56 156 35.063 -0.172 0.56 1.52 1.01 2/26/69 FP 17-20 698 11.30 6.92 10.86 0.818 0.817 69.03 98.97 143 63 54 145 35,014 -0.221 0.72 1.36 0.91 2/26/69 FP 17-21 Dumbell shaped N1 rod exposed to fuel galt for 30 seconds 2/31/69 Fv 17-225 818 Freeze valve capsule - salt sample 34.976 -0.258 G.84 1.24 0.83 3/5/69 FP 17-23 839 11.60 6.75 10.60 0.817 0.814 132 70 58 132 34,969 =0.265 0.86 1,22 0.81 3/13/69 P 17-24 921 11.50 6.64 10.62 0.816 0.816 136 76 40 144 34,943 -0.291 0.95 1.13 0.75 3/17/69 FV 17-25G Freeze valve capsule - gas sample 3/17/69 FP 17-26 1006 50 g sample for determination of U3*/ZU concentration 34.916 -0.318 1.04 1.04 0.69 3/19/69 FP 17-27 1048 11.50 6.72 11.03 0.815 0.806 98 70 77 141 34,903 -0.331 1.08 1.00 0.67 3/26/69 FP 17-28 1147 11.93 6.82 11.56 0.814 0.805 147 71 33 159 34.872 -.0362 1.18 0.90 0.60 3/26/69 FV 17-295 1171 Freeze valve capsule-salt sample 34,864 -0.370 1.21 0.87 0.58 4/1/69 FP 17-30 1292 il.40 7.01 11.09 0.813 0,817 68,60 98.96 148 69 34 148 34.826 -0.,408 1.33 0,75 0.50 4/2/69 FV 17-318 10,460 7.24 7,69 0.813 0.972 4/3/69 FV 17-325 1340 Freeze valve capsule - salt sample 34.811 -0.423 1.38 0.70 0.47 474769 FV 17-33G Freeze valve capsule - gas sample 4/7-8/69 FP 17-34~40 50 g samples for mass spectrometric analysis 4/9/69 Run 17-F 1538 34,748 ~0.486 1.58 0.50 0. 34 4/14/69 Run 18~I 1538 Sy Table 3.1 (continued) (Weight Z) (ppm) 3+ Date Sample Equiv. Li Be Zr U F I Fe Cr Ni Pu IU in AU (kg) I Eq. L Eq. 4[U”/LU] Full m o Circula- Oxida- Reduc- % ?Oflgr om. S- tion tion tion Nom 4/14/69 FP 18-1 1538 11.30 6,50 10,65 0.812 0.800 66.50 95.78 151 86 33 157 34,787 =0.486 1.58 0.50 0.34 4/14/69 FV 18-25 1538 11.36 6.70 14.16 0.812 0.897 Nb: 42% 34,787 -0.486 1.58 0.50 0.34 4/18/69 FP 18-3 1564 Zr rod exposed to fuel for hr; 20.24 g dissolved (0.89 eq.) 34.779 -0.494 2.61 1.36 0.91 4/19/69 FV 18-4S5 1565 10.50 6.62 13,82 0,812 0.861 34.779 -0.494 2,61 1.36 0.91 4/18/69 FDE-A 4/22/69 FDE~B 9.50 5.14 17.68 0.179% 2.73% 0.037% 4/23/69 FDE-C 5.49 5.67 13.39 0.510% 1.61% 0.227% 4/23/69 FP 18-5 1718 11.35 6.53 11,41 0.811 0.805 69.10 99.23 100 77 50 156 34,730 -0.543 2.77 1.20 0.81 4/23/69 FV 18-6%5 6.28 6.19 9.3 0,8if 0,755 4725769 FP 18-7 1766 Zr Rods exposed tc fuel: 24,04 g dissolved (1.05 eq.) 34,715 -0.558 2.25 1.77 1.18 4/26/69 FP 18-8 50 g sample for isotopic analyses 4/29/69 FP 18-9 50 g sample for isotopic analyses 4/29/69 FP 18-10 1855 11.50 6.38 11,18 0.810 0.826 69,40 99.32 119 79 44 139 34.687 ~-0.586 1,91 2,11 1.41 4/30/69 FP 18-11 Empty Ni cage, exposed to salt for 10 hours 5/2/69 FP 18-12S 9.61 5/5/69 FP 18-13 1979 10.50 6.21 10.95 0.809 6.781 69.80 98.28 119 65 50 147 34,648 ~0,625 2.04 1.98 1.33 5/6/69 FV 18-14G Freeze valve capsule ~ gas sample 5/6/69 FV 18-15G Freeze valve capsule - gas sample 5/8/69 FP 18-16 2044 50 g sample for U3*/LU analysis. USt/IU = 0.4% 34.627 ~0.646 2.10 1.92 1.28 5/8/69 FP 18-17 30 g FeFy added to pump bowl. (0.64 oxidation equiv}. 5/9/69 FP 18-18 Interfacial tension measurement 2,74 1,28 0.86 5/9/69 FV 18-195 7.16 10.61 0.672 5/12/69 FP 18-20 Cu cage exposed to fuel salt for 10 hr. 5/13/60 FV 18-21G Freeze valve capsule - gas sample 5/14/6% FP 18-22 2224 11,23 6.89 11.32 0.807 0.804 70.10 100.38 107 65 33 154 34.570 -0.703 2.93 1.09 0.73 5/15/69 FP 18-23 2248 Be addition: 5.68 g as rod (1.26 eq) 34.563 -0.710 2.95 2,33 1.57 5/17/69 FP 18-24 Interfacial tension measurement 5/17/69 BV 18-25G Freeze valve capsule - gas sample 5/17/69 FV 18-26G Freeze valve capsule - gas sample 5/19/69 FP 18-26 Fission product deposition sampler (ligquid phase test), 1 hr. exposure 5/20/69 FP 18-27 2344 10,50 7.10 16.82 0.807 0.799 68.20 97.46 149 79 33 133 34.436 =0.740 3.05 2.23 1.51 5/20/69 FP 18-28 2344 Interfacial tension measurement - Be® present in sampler (3.17 g:0.704 eq) 34.532 =0.740 3.05 5/21/69 FV 18-29G Freeze valve capsule - gas sample 5/21/69 FP 18-30 50 g sample for isotoplc analyses (48.1 g obtained) 5/22/69 FP 18-31 50 g sample for isotopic analyses (41.7 g obtained) 5/22/69 FP 18-32 50 g sample for isotopic analyses (36.5 g obtained) 5/23/69 FP 18-33 50 g sample for isotopic analyses (44.1 g obtained) 5/23/69 FP 18-34 30 g sample for isotopic analyses (40.4 g obtained) 5/23/6% FP 18-35 50 g sample for isotopic analyses (37.4 g obtained) 5/26/69 FP 18-36 50 g sample for isotopic analyses (42.6 g obtained) 5/26/6% FP 18-37 30 g sample for isotopic analyses (42.0 g obtained) 5/271j69 FP 18-38 50 g sample for isotopic analyses (41.0 g obtained) 5/27/69 ¥FP 18-39 50 g sample for isotopic analyses (39.7 g obtained) 5/27/69 FP 15-40 50 g sample for isotopic analyses (44.5 g obtained) 5f27/69 FP 18-41 50 g sample for isotopic analyses (41.3 g obtained) 5/28/69 FV 18-42G Freeze valve capsule - gas sample 5/28/69 FP 18-43 2464 10.40 5.92 10,40 0.806 0.791 70.60 98.16 199 72 38 153 34,495 -0,778 3.17 9t Table 3.1 (continued) | . (WEiEht z) (me) I+ Date Sample Equiv. Li Be Zr U F L Fe Cr Ni pu U in AU (kg) I Eq. L Eq. {U” /fIU] Full Yom Obs Circula- Oxida~ Reduc~- A Pog;r : : tion tion tion Hom. 5/29/69 FV 18-44C Freeze valve capsule - gas sample 6/1/69 FV 18-45G Sample taken 30 min. after shutdown 6/1/69 FY 18-45G Sample taken 6 hr after shutdown 6/6/69 Run 18-F 2547 0.805 34.468 -0.805 3.26 8/11/69 Run 19-1 2547 0.802 3.26 0 8/16/69 FP 19-8 2547 10.50 6,12 10.90 0.802 0. 850 78.5 106.82 186 78 92 igég 34. 360 -0.805 8/18/69 FV 19-95 Freeze valve capsule - salt sample 66 Nb: 34% 8/19/69 FP 19-10 Fuel enrichment - 162,047 g salt = 100.145 g U 8/19/69 FP 19-11 Fuel enrichment - 155.387 g salt = 96.G29 g U 8/21/69 FV 19-12 2547 Fuel enrichment - 155.383 g salt = 96.027 g U IU = 292,201 34,652 ~0.805 o 8/21/69 FV 19-13G Freeze valve capsule - gas sample e 8/21/69 FV 19-14G Freeze valve capsule - gas sample 8, 8/21/69 FV 19-15G Freeze valve capsule - gas sample 2 8/21/69 FV 19-16G Freeze valve capsule ~ gas sample 2 8/27/69 FP 19-17 2629 11.0 5.69 10.94 0.809 (0.811 72,2 100.68 221 83 48 l;;: 34.626 -0.830 g 8/28/69 FP 19-18 2646 11.4 5.85 11.01 0.809 0.797 68.0 97.09 197 75 42 124 34,621 -0.836 8 9/4/69 FV 19-19G Freeze valve capsule - gas sample n 9/4/69 ¥V 12-20¢ Freeze valve capsule - gas sample 8 9/5/69 FP 19-21 2727 10,90 7,20 11,20 0.808 0.795 6%9.80 99.92 177 89 15 98 34,595 -0,861 e 9/9/69 FP 19-22 2793 10.95 6.92 11.30 0.807 0,799 68.50 98,50 147 87 20 104 34,575 -0.882 2 9/10/69 FV 19-23G Freeze valve capsule - gas sample + 9/10/69 FV 19=245 Freeze valve capsule -~ salt sample 17 0,81 0.55 9/12/69 FP 19-25 PuF4 addition: 31.6 g PuF3 (Zr: 0.62 g; = 0.03 eq.) 9/19/69 FP 19-26 PuF3 addition: 35.6 PuF3 0.84 0.57 9/19/69 FP 19-27 2793 10.70 7.12 10.80 0.807 0.792 70.10 99.55 183 96 50 144 34,575 -0.882 0.84 0.57 9/23/69 FV 19-28G Freeze valve capsule - gas sample 9/23/69 FV 19-29G Freeze valve capsule - gas sample 9/24/69 FP 19-30 10.70 7.38 10,90 0,807 0.781 66.30 96.10 227 97 48 142 9/24/69 FP 19-31 PuF3 addition: 39.2 g PuFj 9/25/69 FP 19-32 PuF3 addition: 40.8 g PuF3 9/25/69 FP 19-33 PuFq addition: 42,2 g PuFj 9/26/69 FP 19-34 PuFq addition: 39.2 g PuFj 165 9/29/69 FP 19-35 2968 10,75 7.07 11,10 0.806 0,775 69,00 98,73 214 97 52 ¢ 34,519 -0.938 0.18)% 0.71 0.48 9/29/69 FV 19-368 Freeze valve capsule - salt sample 97 &g%n 9/30/69 FV 19-37G 2992 Freeze valve capsule - gas sample 34.512 -0.945 (0.21) 0.68 0.46 9/30/69 FV 19-38G 2992 Freeze valve capsule - gas sample 10/1/69 FP 19-39 Surface tension exp. no Be present 10/2/69 FP 19-40 3040 Surface tension exp. - 2.87 g Be dissolved (0.64 eq.) 34,497 -0.960 (0.25) 1.28 0.87 10/2/69 FP 19-40 Img: 1171 0.54 22.4 <0.01 10/3/69 FV 19-41G Freeze valve capsule - gas sample 10/3/69 FV 19-428 Freeze valve capsule - salt sample Nb: 0% 10/5/69 FP 19-43 3106 10.60 6.61 11.40 0.805 0.767 71.70 183 96 45 165 34,476 =0,981 (0. 32) 1.21 0.82 10/6/69 FV 19-44S Freeze valve capsule - salt sample 10/6/69 FP 19-45 3130 50 g for YI/Iv 34,468 =0,989 (0.35) 1.18 0.79 10/7/69 FV 19-46G Freeze valve capsule - gas sample 10/7/69 FV 19-475 3154 Freeze valve capsule - salt sample Ly Table 3.1 {continued) (Weieht %) (ppm) 3+ Date Sample Equiv. Li Be Zr U F E Fe Cr Ni Pu LU in AU {kg) L Eq. % Eq. [U” /Zu] Full Circula- Oxida~ Reduc=- A Poger Nom. Cbs, tion tion tion Nom. X 10/8/69 FP 19-48 3178 Be addition: 4.91 g as rod (1.09 eq.) 34,461 -0.996 0.37) 2.25 1.52 8§.58 6.95 14,7 0.804 0.849 (65.11) 0.217% 3.53% 0.07%Z 161 34.453 -1.004 (0. 40) 2.22 1.50 10/9/69 FP 19-49 Graphite assembly exposed to pump bowl wvapor 10/9/69 FP 19-50 Graphite assembly exposed to pump bowl salt 16/9/69 FP 19-51 Graphite assembly exposed to pump bowl salt 10/13/69 P 19-52 Copper rod exposed to pump bowl salt 10/13/69 FP 19-53 3321 11.48 6,79 11,60 0,804 (.786 68.50 235 102 40 164 34,411 ~1.049 (0.54) 2.08 1.41 10/14/69 FV 19-54G Freeze valve capsule - gas sample 10/14/69 FV 19-558 Freeze valve capsule - salt sample 10/15/69 FV 19-56G Freeze valve capsule - gas sample 10/17/69 FV 19-57sS Freeze valve capsule - salt sample 10/17/69 FV 19-58S Gas sample takenm 1 hr after reduction of power to 10 kw 10/18/69 FvV 19-59S Gas sample taken 4 hr after reduction of power to 10 kw 10/20/69 FP 19-60 3458 50 g sample for uH/tu analysis; uH/y = 0.07%h 34.365 -1.092 (0.68) 1.94 1.31 10/21/69 ¥FP 19-61 Nb foil exposed to fuel salt 45 min 10/22/69 FV 19-62G Gas sample with intake port at bottom 10/22/69 FP 19-63 3506 11.25 6,78 10,90 0.802 0.802 70.0 99.78 136 95 33 173 34,350 ~1.107 0.73) 1.89 1.29 10/22/69 FV 19-64G 10/23/69 FV 19-656 Freeze valve capsule - gas sample 10/24/69 TFP 19-66 Capsule for FP plating experiment - 10 hr exposure 10/24/69 ¥FP 19-67 Capsule for FP plating experiment - 3 hr exposure 10/27/69 F¥P 19-68 Capsule for FP plating experiment - 10 min exposure 10/28/69 FP 19-69 3627 50 g sample for oxide analysis; sample not removable from transport container. 34,311 -1.146 (0. 86) 1.7¢6 1.20 10/28/69 FV 19-70G Gas sample with intake port at top 10/28/69 FP 19-71 50 g sample for hot cell tests 10/29/69 FP 19-72 50 g sample for hot cell tests 10/29/69 FV 19-73G 10/30/69 FpP 19-74 3698 11,20 7.31 10.70 0.801 0.796 69. 30 99.35 124 101 20 168 34,289 -1.168 (0.93) 1.69 1.15 10/30/69 FP 19-75 50 g sample for U3*/IU analysis: UST/IU = 0.02%h 10/31/69 Fv 19-76S Freeze valve capsule - salt sample 10/31/69 Fv 19-77S Freeze valve capsule - salt sample 11/2/69 FV 19-78G Gas sample obtained 1 hr after drain Run 19-F 3777 34,267 -1.193 (1.01) 1.61 1.10 Run 20-1 11/26/69 FV 20-15 0.800 0.784 34.274 -1.193 - - ¢ 11/26/69 FP 20-2 3777 Surface tension experiment - control sample, 2 hr exposure 11/27/69 FP 20-3 7LiF-233yF, addition U = 97.2 g 11/27/69 FP 20-4 50 g sample for oxide determination: 58 ppm 34.371 -1.201 11/28/69 FP 20-5 Sample for UX/IU analysis by spectrophotometric method 11/28/69 FP 20-6 3825 10.00 7.19 10,90 0.802 0.783 72,80 101,71 199 98 72 117 34, 356 ~1.208 0.15 0.1 11/29/69 FP.20-7 3849 Be addition: 6.974 as rod (1.55 eq.) 34.384 -1.216 ( - ) 1.53 1.04 White Salt 9.92 7.42 11.95 0.803 0.799 -— 0.17% 1,88% 0.04% Dark Crust 0.10% 6.447 0,047% 11/30/69 FP 20-8 Cu0-Ni assembly exposed to fuel pump vapor 8 hr 12/1/69 FP 20-9G Freeze valve capsule - gas sample 12/1/69 FP 20-10 Ni powder in Ni capsule, exposed to fuel pump vapor 8 hr 12/2/69 Fp 20-11 Electron microscope screen assembly, exposed to fuel pump vapor 2 hr 12/2/69 FP 20-12G Freeze valve capsule -~ gas sample 12/3/69 FP 20-13 Ni rod, exposed to fuel pump vapor & hr 514 Tablie 3.1 (continued) | | . (Weight %) (ppm} I+ Date Sample Equiv. Li Be Zr U F z Fe Cr Ni Pu U in AU (kg) L Eq. L Eq. [U” /ZU] Full Circula=- Oxida- Reduc- % power Nom. Obs. tion tion tion Nom., Hi 12/3/69 ¥P 20-14 50 g sample for U/IU analysis: 0.11% 12/3/69 FP 20-15 Sample for U3*/IU analysis by spectrophotometric method 12/4/69 FP 20-16 CuD-Ni assembly exposed to fuel pump vapor 8 hr 12/4/69 FP 20-17 Cu0-Ni assembly exposed to fuel pump vapor 8 hr 12/4/69 FP 20-18 'LiF-233UF, addition, U:92 g 12/5/69 FP 20-19S8 Freeze valve capsule - salt sample 12/5/69 FP 20-20 Ni bar exposed to fuel pump 8 hr 12/8/69 FP 20-21 Cu0-Ni assembly exposed to fuel pump salt 8 hr 12/9/69 FP 20-22 4093 Be addition; 9.874 g as rod (2.19 eq.) 34,271 ~1.293 (0.25) 3.54 2,41 12/9/69 FP 20-23 Surface tension experiment, with Be®, 4 hr exposure 12/9/69 FP 20-24 Sample for U3'/IU analysis by spectrophotometric method. ! 12/10/69 FP 20-25 Ni bar exposed to fuel pump vapor 8 hr 12/10/69 FP 20-26 50 g sample for U'/IU analysis; U/SU = 0.20% 12/10/69 FP 20-27G Freeze valve capsule - gas sample 12/10/69 FP 20-28 Cu0-Ni assembly exposed to fuel pump salt 8 hr 12/11/69 FP 20-29 Cu0-Ni assembly with Pd windows exposed to pump bowl vapor 8 hr 12/11/69 TFP 20-30 4140 Sample for U/IU analysis by spectrophotometric method 34,256 -1.308 (0.30) 3.49 2.37 12/12/69 FP 20-31 4165 10.80 .80 10,90 0,800 0.801 71.60 101.03 265 92 160 34,249 -1.316 (0.33) 3.46 2.35 12/12/69 FP 20-32¢G Obtained 1 hr after fuel drain 12/12/69 Run 20-F 4167 0.800 34,249 -1.316 (0.33) 3.46 2.35 *Analyzed after storage for 30 hr at room temperature, a MSRP Semiann. Prog. Rept. for P/E Feb. 28, 1967, ORNL-4119, p 156. b MSRP Semiann. Prog. Rept. for P/E Aug. 31, 1967, ORNL-4191, p. 168. 0 R. E. Thoma and J. M. Dale, ORNL-4396, Feb., 28, 1969, p. 134, d 0.819 kg U added; sample addition capsules not designated by FP-numbers Samples removed from the fuel storage tank e £f 50 g sample for determination of uranium concentration by fluorination. & Increase in oxidation equivalents after FP 19-26. ORNL-4548, p 180 h Sample was not usable for U3+/EU analysis. Oxide concentration determined to be 71 ppm. 6 maintained throughout the entire period of reactor operations with the MSRE. All salt, metal, and gas specimens removed from or introduced into the fuel system were assigned sample designation numbers. Analytical data obtained with these specimens are summarized in Table 3.1, which will serve as the basis for the discussion in the remainder of this section. 3.2 Component Analysis As soon as a statistically significant number of fuel-salt samples were obtained it became evident that the average variation in the analytical values for the components Li, Be, Zr, and F within one standard unit of deviation (1 o) would preclude the use of the results of chemical analyses for these components for opera- tional control of the reactor. Of special interest in this connection is the component zirconium. It might seem that if reduction in the concentration of zirconium in the salt were detectable analytically, such reduction would signal that the oxide saturation limit was exceeded. However, assuming that the salt was origi- nally completely free of oxide and that the solubility limit is ~700 ppm at 650°C, the fuel salt could accommodate 0.34 kg of oxide before it became saturated. If the source of the oxide was from moisture, saturation by oxide would correspond to complete reaction of 0.23 kg of HF with the containment vessel, from which 0.29 kg of Cr would be leached into the salt, increasing its concentration there by 59 ppm. It is thus evident that unless highly sensitive techniques were available for measuring the oxide concentration of the salt introduced from contamination of the salt by moisture, the initial indication of such contamination would be the increase of chromium. The [ ¢ sensitivity limit for zirconium corresponds to *0.24%, which corresponds to 1.75 kg of ZrO,. Although the solu- bility of ZrQ, in the fuel salt is sufficiently temperature dependent to expect that on saturating the molten-salt solution with oxide and that further formation of oxide would result in the deposition of oxide crystals at the coldest point in the circuit, it has not been established experimentally that, with rapidly circulated streams, such crystal growth would actually occur. If not, significant amounts of zirconium might be carried in the salt in slurry form and an analysis of samples removed from the pump bowl would not signify a loss of zirconium from the salt as one might expect otherwise. It is thus clear that by the time it would have been possible to establish from zirconium analysis that the 10 limit had been exceeded, at least 1.41 kg of ZrO, would have been precipitated and probably 50 carried by the salt as a slurry. Since, of all the fuel salt constituents, zirconium is the one whose reduction in concentration would independently reflect increasing amounts of oxide impurity, only perfunctory attention was given to the precision of the analyses of the carrier constituents. Analysis of these constituents was con- tinued on a routine basis because of their use in providing general confirmation of the inference that the fuel was chemically stable and because the principal analytical costs were incurred from operations related to handling and processing the radioactive samples; laboratory tests which followed dissolution of the salts for analysis comprised a minor fraction of the overall charges. Mean values of the MSRE fuel-salt composition as determined from analysis of samples removed from the pump bowl are given in Table 3.2. 3.3 Oxide Analysis Until the chemical equilibria involving H,, HF, and oxides in LiF-BeF, melts were defined completely,? no satisfactory analytical method existed for determina- tion of oxide in MSRE salts. The KBrF; method,> while occasionally satisfactory, produced results that reflected the frequent contamination of samples after they were removed from the reactor by quantities of water adsorbed in transit and during storage, and were for the most part generally incredibly large and erratic. Subsequently, an improved method of analysis was developed® based on the equilibrium 0% + 2HF(g) = H,O(g) + 2F ", which occurs when a molten-salt sample is purged with an HF-H, gas mixture. Although this hydrofluorination method was regarded as less sensitive than the inert-gas fusion method” or the KBrF, method, satisfactory sensitivity was achieved by ~50-g samples. After development, only this method was used for analysis of MSRE salt samples. During the period when the MSRE was in operation it was important to know the approximate concentration of oxides in the fuel and coolant salts, but the information was not considered to be particularly relevant to a need to develop highly sensitive in-line analytical methods for reactors in which on-line fuel reprocessing is done. This arises from the fact that on-line reprocessing methods are likely to employ chemical processes in which, for rare earth removal, there is a complete turnaround in the uranium inventory. If this is done, the turnaround will be accomplished by fluorination, and the reconsti- tuted stream will be adjusted in uranium (IV, III) concentration. The fuel-salt mixtures which are most suited for use in breeder or converter systems employ a fertile carrier Table 3.2. Mean values of the MSRE fuel composition from chemical analysis Sample No, of Li Be Zr U F LiF BeF7 ZrFy UFy, Cr Fe Ni Nom, Aver Group Samples wt 7) (mole %) {(ppm) Wt 4 U Carrier® 36 10.79+0.33 7.25+0.18 11.93+0.31 - 70.1740.62 62.24+0.90 32.48+0.90 5.28+0.16 22+5 102+54 1449 Run 22 8 10.4440.19 6.79+0.31 11.2630.13 3.0440.02 67.50%0.62 62.84+1.08 31.46+1.12 5.16+0.11 0.54+0.01 1948 98+51 23+3 3.038 Run 2 10 10.3940.17 6.44+0.02 11.43+0.34 3.05+0.02 71.65+0.80 63.72+0.43 30.40+0.44 5.33+0.14 0.55+0.01 37413 163+49 34+11 3.038 Run 3 51 10.45+0.21 6.49+0.21 11.22+0.35 3.98+0.71 69.80+1.72 63.46+0.83 30.60+0,84 5.23+0.17 0,71+0,12 3748 154455 48+19 Run 4 22 10.51+0.14 6.55+0.16 11.14+0.30 4.642+0,028 67.17+1.44 63.36+0.57 30.65+0.58 5.15+0.12 0.83+0.01 48+7 131+65 40+20 4.648 Run 5-7 14 10.52%0.28 6.54+0.19 11.32+0.23 4.629+0.026 68.72+0.86 63.36+1.03 30.58+0.91 5.23+0.14 0.82440.013 50+7 10844 54+25 4.627 Run 4~7 47 10.52+0.18 6.57+0.18 11.2440.27 4.638+0.025 67.96+1.36 63.29+0.72 30.70+0.70 5.19+40.13 0.824+0.011 49+7 114+55 46421 Runt 8 8 11.78+1.41 6.53+0.20 11.16+0.19 4.6324+0,011 69,77+¢1.49 65.84+2.49 28.57+2,12 4,8240.35 0.771+0.058 64+7 122445 61+36 4.605 Runt 9 4 10.99+0.10 6.63+0.067 11,15+0.37 4.603+0.031 68.55+0.48 64.17+40.10 30.04+0.15 4.99+0.19 0.794+40.011 61+5 150+17 52420 4,587 Run 10 10 11.14+0.08 6.58+0.19 11.05+0.15 4.609+0.020 67.82+1.38 64,65+0.45 29.6430.48 4.92+0.06 0.791+0,010 60+4 150+30 74¥35 4.567 Run 11aA 31 10.80+0.35 6,46+0,15 10.97+0.18 4.570+0.018 67.81+1.46 64.30+0.90 29.88+0.82 5.02+0.13 0.804+0.018 64+6 131+48 54416 4,557 Run 11B 6 10.89+0.36 6.53t0.09 10.97+0.17 4.579+0.023 68.94+1.66 64.27+1.00 29.96+0.85 4.97+0,14 0.799+0.017 64+11 144+46 54+18 4.565 Run 1l 38 10.82+0.35 6.47+0.15 10.9740.17 4.571+0.019 67.99+1.53 64.30+0.90 29.89+0.81 5.01+0.13 0.803+0.018 64+7 133+47 54416 Run 10-11A 41 10.8940.34 6.49+0.17 11.0040.17 4.580+0.024 67.83+1.43 64.34+0.82 29.81+0.76 4.99+0.13 0.801+0.017 63+6 136+45 59+24 Run 10-11 48 10.88+0.34 6.49+0,.16 10.99+0.17 4.579+0.024 67.95+1.47 64.3740.83 29.84+0.76 4.99+0.13 0.801+0.017 63+7 136+44 58+23 Run 12-4 15 11.004+0.45 6.50+0.18 10.75+0.21 4.544+0.023 66.79+1.10 64.65+0.96 29.70+0.90 4.85+0.13 0.790+0.020 7149 166+54 61+11 4.535 Run 12-B 11 10.85+0.32 6.54+0.11 10.82+0.29 4.570+0.028 66.10+0.90 64.21+0.57 30.08+0.49 4.92+0.14 0.800+0.017 69+8 139+26 226+209 4.571 Run 12 26 10.93+0.40 6.5230.15 10.78+0.24 4.555+0.027 66.50+1.06 64.46+0.83 29.86+0.77 4.88:0.14 0.794+0.021 7040 154446 110+ ? Run 13-14A 24 10.65+0.20 6.28+0.28 11.104+0.21 4.561+0.024 66.27+0.52 64.51+0.74 29.51+0.80 5.16+0.14 0.81740.017 63+6 123422 68+15 Run 14B 20 10.35+0.28 6.39+0.16 10.914+0.24 4.523+0.030 66.49+0.32 63.55+0.92 30.48+0.86 5,1540.14 0.821+0.012 82+7 121+147 62414 Run 14 44 10.51+0.28 6.33+0.24 11.014+0.24 4,5430.032 66.3740.45 64,07+0.95 29,95+0.95 5.15+0.13 0.819+0.015 77+8 123+43 65+15 4,548 Run 4-14 176 10.7540.50 6.48+0.21 11.04+0.27 4.585+0.046 67.44+1.49 64.09+1.11 30,05+1.00 5.06+0.19 0.809+0.024 64+13 130+45 67+67 Run 16-18 21 11.33+0.39 6.71+0.30 10.994+0.29 0.809+0.011 68.71+1.26 65.08+0.91 29.93+0.93 4.85+0,14 0.137+0.004 7249 135+24 49+12 0.8155 Run 19 11 10.95+0.34 6.83+0.48 11.08+0.27 0.791+0.013 69.40+1.66 63.94+1.67 30.96+1.74 4.97+0.18 0.137+0.005 93+8 186+38 38+13 0.8045 Run 19-20 13 10.86+0.40 6.85+0.45 11.05+0.26 0.791+0,012 69.83+1.86 63.68+1.75 31.21+1.79 4.97+0.17 0.137+0.005 93+8 193+41 41+16 Run 16-20 33 11.14+0.45 6.76+0.37 11.02+0.28 0.802+0.015 69.18+1.60 64.53+1.46 30.43+1.46 4,9040.16 0.13740.004 8C+14 157+43 46+14 Run 2-20 258 10.79+0.50 6.530.25 11.0540.28 3.881+1.462 67.87+1.65 64,10+1.24 30.16+1.08 5.,05+0.19 0.685+0.261 62+17 136+58 62+58 a Analyses from the General Analysis Lab., all others from the HRLAL. 1S 52 salt of the approximate composition ’LiF-BeF,-ThF, (72-16-12 mole %). It has been deduced recently that in uraniferous fuel salts constituted from this base, the oxide tolerance at operating temperatures is in the range 30 to 70 ppm.® Although there is no need for incorporation of an oxygen getter, such as ZrF,, in the salt directly, this low oxide tolerance indicates that there will be a need for swift, satisfactory, and preferably in-line methods for determination of Q% (and UF;) concentrations in the fuel after storage and during startup and operation. Further, it will be necessary to incorporate a demonstrated continuous process for removal of 0% to a satisfactorily low level from a substantial side stream from the reactor. The results listed in Table 3.1 show that the concentration of oxide never exceeded ~10% of saturation and were thus regarded, along with the corrosion data, as satisfactory evidence that the fuel system did not experience contamination during periods when fuel was circulated in it. Other material pertinent to this discussion is described in Chap. 6, concerning corrosion and anomalies in the behavior of UF ;. 3.4 Uranium Concentration One of the purposes of the Molten-Salt Reactor Experiment was to examine the applicability of various techniques for rapid, accurate, reliable means of estab- lishing the inventory of uranium in the fuel-salt system. We projected that the experience gained with the MSRE ORNL-DWG 71-9987R I 468 o *BASED ON INITIAL LOADING OF ) 226.25 kg U, 4885.3 kg 2 SALT, . U CONSUMPTION RATE=0.3991042 g/EFPH [7.4 Mw (th)] 466 ° > © © . ® o ) 52 B o _— . z 464 r-X 7 - R — L | \.. [ ] ® w b \d ") a2 - - - L @ & Z ® e o o B 'R [ ® e Z 460 2 ~~ hd o = b ® ® ilj ® o) s Qoo o ¢ (] 9 .' £ 458 s > o s 3o ® 3 ° % e ® ® =z ® ® = ® e ® ® = o & &g .fi ® D 4.56 o ® @ & £ = o 8 oq [ | e e b =z ~ 9 8 o e ¢ ® o ® D % oF @ @ @ > iy~ L . 5 4.54 @ ~ :" > NOMINAL VALUE®” @ o S~ ' S o % E . . : \\o 4,52 —] - § Z Lt @ s ® O 4,50 O ® 4.48 & RUN NO. ® 5-7 8.9 10 1 12 13-14 o , T T T T o) 2 4 6 8 (x10%) equivalent full power hours Fig. 3.1. Comparison of analytical and computed values of uranium in the MSRE fuel salt (*2%U and 23U tuel charge). would fix the alternatives to be explored for the development of larger molten-salt reactors. When it was noted early in MSRE power operations that the sensitivity of on-site determination of uranium concen- tration coefficient in the reactivity balance was some tenfold greater than the statistical variation in chemical analysis, further consideration of the applicability of laboratory analysis of individual samples for operational control was abandoned. Other sections of this report discuss uranium burnup rates (Chap. 7), fissile inventory (Sect. 2.4.2), and material balances (Sects. 3.6 and 3.7), the results of which led to the final estimates of the concentrations of uranium in the fuel salt as listed in Table 3.1. After final adjustments in the estimated amounts of salts contained in the reactor system and of the power 53 production rates, comparisons of the analytical data for uranium concentration were made with the nominal values for 233U and 232 U fuel, respectively, in Figs. 3.1 and 3.2. It is evident in Fig. 3.1 that the analytical data contain a small positive bias in reference to nominal values. The bias is of little consequence, for with the average uranium inventory a difference in the estimated carrier salt mass of 1 kg changes the nominal concentra- tion of uranium by 0.01 wt %. 3.5 Structural Metal Impurities In the development of molten-salt reactor technology, continuvally recurring needs for accurate values of the properties of moving fluids become apparent. Conse- quences of ever increasing amounts of structural metal ORNL-DWG 71— 9988R 0.830 0.825 0.820 0.815 € Q.810 0.805 & ® 0.8C0 2 URANIUM CONCENTRATION {w! %) 0.795 0.790 0.785 0. 780 0.775 2 3 4 (x10%) equivalent full power hours Fig. 3.2. Concentration of uranium (*>>U) in the MSRE fuel circuit. impurities that might be generated in the MSRE and that might alter the properties of the flowing sait streams into which they were released were envisioned at the outset of operations. With the possibility that impurities would exist both as ionic and metallic species, it was desirable to examine the interrelation- ships of their development with respect to corrosion of the container circuits, fission product transport, cavita- tion effects, changes in the flow characteristics, and their impact on MSBR reprocessing capability. Since all but the initial aspect of this behavior comprises an ongoing effort, the scope of the present report includes consideration of structural metal impurities in the MSRE only with respect to the implication of changes in their concentration as related to corrosion. Specific discussion of the chemical significance of the amounts of impurities and changes in the concentrations is deferred for consideration under Sects. 5 and 6, in which corrosion in the fuel and coolant circuits is discussed, and for subsequent reports where the rela- tionships of physical and chemical properties of the MSRE to metallurgical behavior and to fission product behavior are assessed. Midway through the experiment, phenomena were observed that led to the speculation that changes in the surface-related properties of the impurities were of possible significance to the changes observed. It was not experimentally feasible at that time to utilize the reactor as the experimental tool for studies of the interfacial behavior in salt-metal systems. By the time that laboratory studies had advanced to the stage where a program of studies with the reactor could be formulated, the schedule for operation of the reactor called for its termination. The combined experience from laboratory investiga- tions and from an extensive engineering test program in which various fluoride mixtures were circulated in thermal convection and forced circulation loops led to the expectation that the concentration of structural metals in the MSRE salts would increase only slightly throughout the planned period of reactor operation. As anticipated, only modest increases occurred. The MSRE fuel was produced, as described in Sect. 2.4, from carrier salt and fuel concentrate, for which the chemical analyses indicated average concentrations of impurities were quite low. Chemical analyses were performed with samples of fuel salt removed regularly from the fuel pump bowl, and occasionally from the fuel drain tank. The average concentrations of the structural metal impurities found in these samples are listed in Table 3.2. Individual analyses are listed in Table 3.1. 54 Table 3.3. Average concentration of structural metal impurities in the MSRE fuel salt Number of Impurity concentration (ppm) Sample group samples Cr e Ni Fuel concentrate Run No. 2 3 34 377 14859 55%18 4 69 53+8 122451 52%27 57 14 507 108*44 54125 8 8 647 122*45 61%36 9-10 i1 38 647 133%47 54%16 12 18 72t9 160+52 85%73 13-14 Early stages 24 73+t6 125%£22 68*15 Latter stages 20 827 12124 62%14 Total 44 778 123+23 6515 15 1618 21 7229 135+t24 49*12 19-20 13 93+8 193*41 41*16 The methods employed by the analytical chemists for determination of structural metal impurities were de- scribed previously.? Each of the methods employed is capable of approximately the same precision. Yet, as shown by the results in Table 3.3, the experimental precision obtained was much better for chromium than for either iron or nickel. Corrosion reactions in the MSRE were expected to produce only CrF, as a dissolved species in the salts; iron, nickel, and molybde- num, if present, were expected to persist in the metallic species. The differences in observed standard deviations for the three elements suggest that behavior very similar to that anticipated did result. Inspection of the indi- vidual results in Table 3.1 shows that the relative concentrations of iron and nickel found among indi- vidual samples rarely were constant. This is a somewhat surprising result since the excellent mixing conditions which existed in the circulating salt should have been conducive for allowing these metals to produce equilib- rium solid solution alloys. The possibility that the metal particles might have deposited from the salt also was expected but seemingly did not occur either during circulation or during storage. In retrospect, these observations are compatible with the early conclusion that the pure salt fluids do not wet metal surfaces and with the fact that the low viscosities of the fluoride mixtures enable thermal convection currents to main- tain the fluids in well-mixed condition during storage. That thermal mixing occurs in these tanks was substan- tiated by the first composition analysis in PC-2. As evidenced by the composition of a sample which was obtamned from No 2 dram tank shortly after 7LiF- 238JF, was added to the carrer salt, virtually com- plete mixing of " LiF-UF, and carrier salt was achieved before intertank transfers were made to ensure homoge- nization Comparably efficient mixing 1s also probably adequate to prevent settling of the very fine (below the limit of microscopic detection) metallic particles of iron and nickel in the drain tanks A quality control program was imtiated by the Analytical Chemistry Division as chemical surveillance of the reactor salt began The results of this survey showed that comparable accuracy and precision were achieved i the analysis of structural metal impurities in the salt samples 19 From the beginning, molybde- num was not detected in the samples Its absence then and later was interpreted to signify that such corrosion as occurred 1n the reactor was essentially diffusion finmted 3.6 Chemical Effects of Reprocessing The MSRE ncluded a fuel processing plant for removal of oxides, if necessary, from the fuel, flush, and coolant salts The plant was also intended for use 1n recovering uramum from the fuel and flush salts after termination of expertments with the MSRE A detailed description of the methods, operating procedures, and components 1s given i the MSRE Design and Opera- tions Report 't It proved to be unnecessary to repurify any of the salt charges in MSRE maintenance opera- tions, although flush salt was repurified after its mitial use and prior to nuclear operations 1n order to remove from the salt the oxides that 1t had accumulated during its 1nitial use 1n the fuel system As planned, the 235,2380) fye] salt was fluorinated mn the chemucal reprocessing plant after completion of experiments with 235y fuel salt Decision to utilize 233U as fuel entailed not only fluorination of the 22> U salt but subsequent reduction of the structural metal impunities that enter the salt as 1t 1s fluonnated Decision to reuse the carrier to constitute 2**U fuel made 1t exigent to remove metal particulates from the salt carrier before returning it to the fuel system Procedures for reduction of the fluonides were evaluated and tested, and an optimum filter medium was selected 12 A sintered metal filter was then designed?3 and incorporated in the return hine from the processing plant Flush salt was reprocessed first to remove the uramum 1t had accumulated 1n use (see Sect 4 2) Following fluorination, the fuel was treated, as was the flush salt, to reduce the structural metal fluorides present in their metallic forms 55 The amounts of uranium which were recovered from the fuel and flush salts were estimated from the increases in weights of the NaF absorbers in which UFy was collected The actual weights of UF, were not measurable because fluormation of the salts in Hastel- loy N processing vessels converted small amounts of molybdenum to 1ts volatile fluondes, which were absorbed along with UF4 on the sodmum fluoride beds The approximate amounts of molybdenum absorbed on these beds were estimated from the proportionate amounts of 1ron, chromium, and nickel that were dissolved mto the flush and carrier salts after fluorina- tion Estimates of the weights of uranium recovered were then obtammed by subtraction of the probable amounts of molybdenum absorbed from the total weights gained by the absorbers In a summary of experience with the processimng plant, Lindauer has estimated that the amounts of uranium recovered from the flush and fuel salts at this time were ~6 5 and 217 85 kg respectively 14 The sodum fluoride absorb- ers were subsequently dehvered to the Goodyear Atomic Plant in Portsmouth, Chio, where wet chemical procedures were used to recover the uranium charge from the NaF absorbers The results of these efforts and their significance are discussed 1n Sect 3 7 The reprocessing procedures were notably successful i decontaminating the uranum charge from fission products The only activity collected in any measurable amount on the absorbers with UF, was ?°Nb None of the filled absorber vessels exceeded maxmmum permis- sible radiation levels, each vessel could be removed from the processing plant without protective shielding At termination of operations of the MSRE with 2°°U fuel, the concentration of structural metal contami- nants in the flush and fuel salts was scarcely greater than when they were fust used Fluormation of the salts 1 the processing plant mcreased the concentra- tions of chromium, iron, and nickel contaminants temporarilly On completion of reprocessing procedures, therr concentrations were reduced once again (o satis- factorily low values Changes in these concentrations as a result of processing the salts are histed 1n Table 3 4 For operation of the reactor with 2**U fuel, a charge of only about 40 kg of urantum was required as compared with the some 220 kg of uranmuum that was used during 23°U operations In order to provide a sufficient volume of the fluid fuel mixture, the reactor was loaded with a supplementary charge of fuel carrier salt amounting to 129 9 kg Methods for removal of the lanthanide element fission products from the fuel streams of molten-salt reactors using reductive extraction techniques were not fully Table 3.4. Structural metal fluoride concentrations? Concentration (ppm) Crt10 Fetd40 Nit1s Flush salt? In reactor system Before fluorination After fluorination After 10.8 hr of H, sparging 76 150 52 104 133 210 516 No sample taken After 604 g of Zr and 9 hr of H, sparging 100¢ 174¢ 50¢ After 1074 g of Zr and 25 hr of H, sparging No sample taken After filtration 76¢ 141c 26 Fuel salt? In reactor system 85 130 60 Before fluorination i70 131 36 After fluorination 420 400 840 After 17.1 hr of H, spaiging 420¢ 430¢ d After 33.5 hr of Hysparging 420¢ 400¢ 520¢ After 51.1 hr of H, sparging 460¢ 380¢ 180¢ After 5000 g of Zr and 24 hr of H, sparging 1006¢ 110¢ <1¢¢ After 5100 g of Zr and 32 hr of H, sparging After filtration No sample taken 34 110 60 From R. B. Lindauer, Processing of the MSRE Flush and Fuel Salts, ORNL-TM-2578 (August 1969). ng sparging times are cumulative. ‘Filtered sample. dContaminated sample. developed until after the MSRE had operated for some time. Until then the most promising method for their separation was the high-temperature low-pressure distil- lation of fuel carriers after removal of uranium. Tests of this method with irradiated salt were scheduled to take place at the MSRE on completion of 23U operations after removal of uranium from the fuel salt. Approxi- mately 48 liters of irradiated carrier salt was allocated for this experiment; when the charge was delivered, however, it was found that only a total of 12 liters was transferred. Partly because of this anomalous behavior and because of its importance in the evaluation of the reactor fuel inventory, a considerable effort was made to obtain an accurate measurement of the volume of salt remaining in the fuel storage tank. At the beginning of 33U operations, samples of fuel salt were removed from the MSRE pump bowl and analyzed regularly by colllometric methods which had been applied previously and which were checked against standards on a routine basis. Their precision and accuracy is regarded to be ~+0.5%. The results of these analyses were, on the average, greater than the nominal values by 0.008 wt %, ~1% of the nominal value, and indicated (disregarding precision limits for purposes of this calculation) that the net weight of the carrier salt during this period was 4639 kg rather than 4708 kg, or that an additional mass of 69.0 kg (1.13 ft*) of salt, that is, a total of 2.99 ft* of the original carrier salt, was not returned to the reactor to constitute the 233U fuel charge. In an attempt to support the conclusion that some 3.0 ft* of carrier salt was left in the fuel storage tank, an isotopic dilution experiment was performed in April 1969, in which two samples of ® LiF (35 g) were added to this tank. Retrieval of samples from the tank using a windlass and 10-g sample ladles was subsequently successful in removing small amounts of salt (which incidentally were found to be encrusted with metal residues) only through repeated attempts. The total weight of ®LiF added was 66.28 g (15.8 g of Li). Prior to ®LiF addition the carrier salt was found by analysis to have an average " Li/ZLi concentration of 999905 £ 0.0015 wt %. The samples recovered after the addition had a concentration of 99.914 + 0.024 wt %15 which would indicate that the volume of the salt is 3.0 £ 0.84 ft>. However, samples of the carrier salt delivered to the still-pot section of the distillation apparatus were found to have concentrations of °Li/ ZLi = 1.59 and 2.85 wt %, indicating that the ® LiF was not dispersed homogeneously in the storage tank but instead had dissolved preferentially in the salt fraction delivered to the still pot. Homogeneous dispersal would have resulted in a higher average ®Li concentration in Table 3,5. Chemical composition of the MSRE carrier salt as indicated by chemical analysis 57 Number Chemical composition (mole %) Specunen of Samples LiF BCFQ ZI'F4 Carrier salt, as produced 36 62241090 32482090 5282016 Carrier salt, assunung transfer of 13 93 kg Z:4 62 33 32 58 515 Runs4 14, all samples 176 6461112 3029%10 510019 Run 14 (only) 44 6461 095 3020095 5192013 Runs 17 20, all samples 33 64 53+146 3043 %146 4901016 2Assumption of the removal of 13 93 kg of zirconium from this salt permits comparison after its dilution by six flush salt residues the sample recovered from the fuel storage tank, and a lesser volume of salt would have been computed from the results of the 1sotopic dilution analysis experiment Thus, although its lower limit cannot be deduced unequivocally from the 1sotopic data, the volume of salt retained 1n the storage tank cannot have been as high as 3 84 ft? The efforts to obtain samples from the storage tank provided additional informatton that allows a separate estimate of the salt volume Sometimes small quantities of salt were obtained, sometimes none If this 1s interpreted to mean that the salt level was at the point in the dished head directly below the samplers, the pool contained ~3 0 ft> of salt Assuming that the storage tank contained 3 0 ft* of salt rather than 1 13 ft* used before, we correct the earlier result 4707 — 1826 (30 ft3) kg + 1130 (1 13 ft?) kg and conclude that the drain tank contained 4638 kg of carrer salt at the beginming of 233U power operations Thus the net weight of the fuel charge at the begmning of ***U power operations was 4638 kg + 38 3 kg = 4676 kg, and the uranium concentration was O 819 wt % Ample thermodynamic evidence existed to indicate that reprocessing procedures would not change the average composition of the fuel carrer salt This expectation was confirmed by comparnson of the average composition of the fuel salt before and after reprocessing and by determunatton that the uranwum concentration m the 233U fuel mixture conformed to our expectations A discussion of the uranmum assay duning this period may be found m Sect 73 The results of chemucal analyses shown 1n Table 3 5 support the conclusion that the average composition of the fuel carrier salt remained essentially constant throughout the entire period of MSRE operations 3.7 Material Balances for 23°U and 2*® U Operations 37.1 Recovery of 235U and 2°%U. A matenal balance of the fuel and flush salts was maintained on a continuous basis from the beginning of experimental operations with the MSRE With refinements 1n physi- cal property data for the salts, in neutron absorption cross-section data, and 1n analytical methods, the quality of the balance mmproved steadily The accuracy afforded by these improvements was established finally by recovery operations at ORNL and at the Goodyear Atomic Corporation, Piketon, Ohio There, the uranium hexafluoride removed from the 22°U fuel and flush salts by chemucal processing at ORNL was recovered by wet chemical methods The amounts of uranium recov- ered using these methods were 6 420 = 0 006 kg from absorbers 6, 7, and 8 {from the flush salt) and 214 572 + 0204 kg from the remaimmng 25 absorbers The procedures employed 1n the recovery of uranium were described 1 detail in a report from the Goodyear Corporation as follows !© The NaF matertal (25 absorbers) was processed using the continuous dissolver and resulted in 120 batches of solution, tor a total of 46,049 liters contamnng 214,456 grams of uranium These batches were measured and sampled, i duphcate, using the solution recovery accountability measurmng columns Batch samples were proportionally composited, in duphcate, repre senting from 10 to 20 batches of solution All composiie samples were analyzed for total uramwum and weight percent U 235 by thermal mass spectrometer In addition, the contin uous dissolver operation generated 1,588 pounds of filter cake, contaiming 105 grams of uramum This matenial was also sampled in duplicate and composited on a weight basis Vent losses resulting from steam effluent were calculated to be 11 grams of uranmum The lumts of error used for the measurement of solution and filter cake resulting from processing the remaiming sodmum fluonde were 0 18 and *10 percent of the reported value per composite analyzed The hmit of error for the total uranium recovered was based on propagation of error applicable to each composite group measured Also, ncluded in the hmat of error 1s a bias estimate for the volume measurements of £( 25 hiter per batch, equivalent to 126 grams of uranum These lumts of error are based on previously determined precision and accuracy estimates Statistical evaluation of the analyses obtained was not performed at this time since the variation between samples indicates the precision will be well within the estimated values. Analyses of standard samples for comparison {0 known uranium value were made during the processing period and showed excellent agreement. The following precautions and special consideration were taken while processing the NaF material to minimize measurement uncertainty: 1. The continuous dissolver and solution recovery systems were dedicated to processing NaF material exclusively from mid-March through June, 1970, to minimize rinsing of equipment as well as possible crossover of other material in process. 2. The accountability measuring columns were recalibrated on a weight basis and the recycle time required for representa- tive sampling was determined by test data. 3. The accountability measuring columns wese rinsed twice, 30 liters per rinse, between each batch measured and sampled. 4. AH containers, plastic, and miscellancous scrap generated during opening the 25 absorbers were deconiaminated and the resulting solution processed through the dissolver. 5. Vent losses, even though minimal, were calculated based on total steam flow to atmosphere and the uranium concentra- tion determined by analyses of condensate samples through- out the processing period. 6. System clean-out was initiated after processing the solution generated from the last three absorbers which were esti- mated at zero uranium content. This solution, 4,210 liters, contained an average of 0.033 grams of uranium per liter. Final rinse of the system averaged less than 0.002 grams U/liter, A disparity between the ORNL material balance and the Goodyear results of 0.83 kg of 235U was evident. In an attempt to resolve this discrepancy we requested the Goodyear Corporation to retain the absorber vessels until investigation of the cause of the disparity was identified. In a telephone conversation between R. B. * Lindauer and J. G. Crawford,!7 two pertinent details were disclosed: the few fines remaining in the absorber tanks were removed (by wiping the interior surfaces with moistened tissues) and added to the process solutions. Further, the interior surfaces of the absorbers were noted to be bright and shiny, a condition which seems to preclude the possibility that significant amounts of uranium were retained on these surfaces. The overall results strongly suggest that rationaliza- tion of the disparity in the amount of uranium which we anticipated would be recovered by the Goodyear plant and that which was actually recovered would be possible only through a careful review of MSRE operations. In this connection, two possibilities were suggested, by J. R. Engel and R. B. Lindauer, respec- tively.!8 Engel noted that the particle filter (a 9-ft filter designed to remove corrosion product solids from the 58 fluorinated salt before its reuse in the reactor) in the line between the fuel storage tank and the processing tanks could, after treatment of the flush salt was completed, have contained an unknown amount of zirconium metal, delivered to this location as the processed flush salt was returned to the reactor system. It is difficult to assign high probability to the events which could have reduced the uranium from the fuel charge as, subsequently, it passed through this filter, in such a way that some 2 kg of uranium remained in the filter; however, the possibility cannot be excluded and merits further examination. Another possible site where uranium may have been retained, as has been suggested by R. B. Lindauer, is the high-temperature sodium fluoride absorber bed which is positioned between the fuel storage tank and the NaF absorbers. The design temperature for operation of this absorber is 750°F, based on previous laboratory studies.!? The laboratory studies indicate that this absorber would not retain UF, at the operating temperature; the possibility that temperature gradients prevailed within the absorber at periods near the end of fluorination operations which allowed the retention of some uranium within the absorber should now be examined, The possibilities that such errors as might be ascribed to misestimates of power output of the reactor, reactivity anomalies, and implications of short-term trends in the results of chemical analyses, have been reexamined; we conclude that their possible contribu- tion to the disparity described is negligible. The possibility that the retention of uranium in either site might be established with certainty by application of newly developed neutron interrogation techniques using a californium source seemed favorable. Mockup experiments were devised and tested in the.remote maintenance practice cell at the MSRE with a neutron source and an ORR fuel element. These tests showed, however, that the method could not be used without considerable modification, including procurement of a more intense source, and the effort was abandoned. Although confirmation of the results of the existing isotopic dilution analyses would be desirable by direct experiments such as were planned, the data which were used to monitor the transfers of uranium and pluto- nium within the reactor system appear to be suf- ficiently reliable to estimate the amounts of 233U and 238 retained in the reprocessing system and for computation of final inventory distribution. 3.7.2 Inventories for stored salts. Later in this report, methods for establishment of the power output of the MSRE are described (see Sect. 7). By application of the 59 Table 3.6. Inventory of residual uramum and plutonium n the MSRE? A. Uranium 233y 234y 235(; 236 238y; sy Fuel circuit inventory, run 20-1, kg 28 782 2 545 0876 0036 2035 34 274 Total inventory, run 20-1, kg 31052 2746 0945 0039 2196 36 978 Drain-tank inventory, run 20-I, kg 2270 0201 0069 0003 0161 2704 Fuel-circuit mventory, run 20-F, kg 28 739 2 563 0876 0 036 2035 34 249 Transfer to flush salt, run 20-F, kg 0411 0 037 0013 0 001 0029 0491 Charged mto drain tank, run 20-F, kg 28 328 2 526 0863 0035 2 006 33758 Drain tank residue, kg 2270 0201 0 069 0003 0161 2 704 Final dramn tank inventory, kg 30 598 2727 0932 0038 2167 36 462 U/sU, wt % 83918 7479 2556 0104 5943 B. Plutonmm 239p, 240p, 238,241,242p, sPu Luel circuit mventory, run 20-1, g 6258 6181 2 39 690 0 Total mventory, run 201, g 6802 67 19 2 60 749 99 Dramn tank mventory, run 20-1, g 544 538 021 59 99 Fuel-circunt inventory, run 20F, g 6156 6543 2 37 683 4 Transfer to flush salt, un 20-F 2 g 618 6 57 023 68 6 Charged into dramn tank, run 20-F, g 5538 58 86 2 14 614 8 Drain tank residue, g 544 538 021 59 99 Final drain tank inventory, g 608 2 64 24 235 674 79 Pu/ZPu, wt % 90 13 952 035 IWeights are based on comparisons of analytical results and computed values These comparisons mdicate that maximum power output was 7 4 MW(t) Final estimates assume 4167 EFPH (equivalent full-power hours) at 7 4 MW(t) Brhis item makes the simphifying assumption that the total amount of plutonium estimated to be transferred to the flush satt was transferred during the final flush of the fuel circust 1sotopic analyses used for this determination, together with other analytical methods, it became possible to estabhish the power output of the reactor with good preciston, and accordingly to compute final materal balances for the fuel and flush sait systems. The results of these computations are listed in Table 3.6 In storage, the fuel salt 1s divided equally between the two drain tanks. Both fuel and flush salt were frozen in storage and are maintamed between 232 and 343°C to minimize the evolution of fluormme from the frozen salts An 1nventory of the uranium and plutonmum con- tamned 1n the drain tanks, based on data Iisted in Table 3.6, together with results of mass spectrometric analy- ses, 1s shown in Table 3.7. Using the values in this inventory the composition of the fuel salt was calcu- lated. Nommnal values are compared with analytical data in Table 3.8. 3.7.3 Salt loss from leakage. Experiments with the MSRE were termunated on December 12, 1969. The fuel and coolant salts were dramned for storage in the drain tanks. After the fuel was dramned, and while the freeze valves were being frozen, am increase n the radiocactivity 1n the reactor containment cell was ob- served, which indicated that a very small leak had occurred 1n the prumary containment. After several hours the activity began to decrease after having driven the momitor on the recirculating cell atmosphere to a maximum of 35 mR/hr. The fuel loop and drain tanks were pressurized without causing an apparent effect on the activity. A sample of the cell atmosphere indicated that the activity was caused prmmanly from '?3Xe Since leakage had apparently stopped, flush salt was charged into the fuel system and circulated for 17 hr Before flush salt was dramned from the circuit the system was pressurized to 20 psig for 2 hr without showing any signs of leakage It was concluded after further investigation that the leak was in freeze valve FV-105 or in the immediate vicinity of this valve To conclude operations with the fuel circuat, flush salt was transferred through the fill line to ensure that FV-103 and the adjacent line were filled with salt Table 3.7. Material balance for the MSRE fuel and flush salts Uranium concentration Inventory (kg) (Wt %) Uranium Fuel salt Carrier salt Calculated Observed Total uranium charged to the MSRE, 1964 229.020 Correction to 238U inventory4 -2.0 Fuel inventory at termination of run No. 3 227.02 4883.8 4656.8 4.648 4.648 Total uranium burned in 23°U operations, 9057 EFPH at 7.4 MW()? —3.615 —-3.615 0 Uranium added as fuel replenishment +2.461 +3.074 +0.613 Total uranium removed in samples —0.256¢ ~5.583 —5.327 Uranium returned to circuit with 233U charge -1.935 0 0 Transfer balance, fuel to flush salt —6.420 -12.0 -12.0 Maximum amount of uranium recoverable 217.255 4865.68 4648.42 Maximum amount of uranium recovered (ORNL)4 217.99 Maximum amount of uranium recovered (Goodyear)€ 214.776 ORNL inventory minus Goodyear recovery value ~-2.479 Carrier salt retained in storage tank (3.0 ft3) ~182.6 Supplementary addition of carrier salt +129.9 LiF-233UF,4 (including 1.935 kg 235:238(J residue) 38.298 57.367 +21.00 Fuel salt charge at beginning of run No. 16 38.298 4654.80 4616.50 0.823 0.820 Total uranium burned in 233U operations ~1.316 -1.316 0 Uranium added as fuel replacement +0.389 +0.630 +0.241 Transfer balance, fuel to flush salt —-(0.289 ~0.289 0 ‘Fotal uranium removed in samples —-0.025 -3.167 —3.142 Fuel inventory at termination of MSRE operations 37.057 4650.5 4613.6 0.797 0.792f ISee sect. 2.4.2. PBurnup rate: —0.3991042 g/EFPH. See B. E. Prince, MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4449, p. 25. €J. R. Engel, MSRE Book Uranium Inventories at Recovery of 235U Fuel Charge, MSR-68-79 (May 15, 1968). dR. B. Lindauer, Chem. Technol. Div. Annu. Progr. Rep. May 31, 1969, ORNL4422, p. 40. €See sect. 3.7.1. f, Average of four samples obtained during run 20. Table 3.8. Composition of fuel salt stored in the MSRE drain tanks TLiF BeF, ZiF, 233%2ygp, 23%1lipyp, Mole Percent Nominal 64.50 30.18 5.194 0.134 Analytical®? 64.53 30.43 4.90 0.137 2.38 % 1073 Weight Percent Nominal 41.87 3544 21.67 1.033 0.0177 Analyticat? 41.37 35.27 20.19 1.06 “Current calculations do not include corrections for transfer of carrier solvent residues to flush salt or flush salt residues to fuel. Disparity between nominal and analytical values for zirconium will be reduced by introduction of this correction factor. bRun 17-20, average of 33 samples. References 1. E.S. Bettis et al., Nucl. Sci. Eng. 2,841 (1957). 2. ). A. Lane, H. G. MacPherson, and Frank Maslan (eds.), Fluid Fuel Reactors, p. 590, Addison-Wesley, Reading, Mass., 1958. 3. F. F. Blankenship, E. G. Bohlmann, S. S. Kirslis, and E. L. Compere, Fission Product Behavior in the MSRE, ORNL-4684 (in preparation). 4. A. L. Mathews and C. F. Baes, Jr., Oxide Chem- istry and Thermodynamics of Molten Lithium Fluo- ride—Beryllium Fluoride by Equilibration with Gaseous Water—Hydrogen Fluoride Mixtures, thesis, ORNL-TM- 1129 (May 7, 1965). 5. G. Goldberg, A. S. Meyer, Jr., and J. C. White, Anal. Chem. 32,314 (1960). 6. MSR Program Semiannu. Progr. Rep. Aug 31, 1965, ORNL-3872, p. 140. 7. MSR Program Semiannu. Progr. Rep. Feb. 28, 1965, ORNL-3812, p. 160. 8. C. E. Bamberger and C. F. Baes, Jr., MSR Program Semiannu. Progr. Rep. Aug. 31, 1970, ORNL-4622, p. 91. 9. MSR Program Semiannu. Progr. Rep. Feb. 28, 1966, ORNL-3936, p. 168; ibid., Aug. 31, 1966, ORNL-4037, p. 200;ibid., Feb. 28, 1967, ORNL4119, p. 165. 10. M. T. Kelley, Statistical Quality Control Report, January—March 1969, ORNL-CF-69-4-30; April-June 1969, ORNL-CF-69-7-65;July—September 1969, ORNL- CF-69-10-36; October—December 1969, ORNL-CF-70- 1-37. 11. R. B. Lindauer, MSRE Design and Operations Report, Part VII, ORNL-TM-907 (May 1965). 12. R. B. Lindauer and C. K. McGlothlan, Design, Construction, and Testing of a Large Molten-Salt Filter, ORNL-TM-2478 (March 1969). 13. J. H. Shaffer and L. E. McNeese, Removal of Ni, Fe, and Cr Fluorides from Simulated MSRE Fuel Carrier Salt, ORNL-CF-68-4-41 (April 1968). 14. R. B. Lindauer, Processing of the MSRE Flush and Fuel Salts, ORNL-TM-2578 (August 1969). 15. Analyses performed by J. R. Sites, Analytical Chemistry Division. 16. Correspondence from W. B. Thompson, Supervi- sor of Process Engineering, Goodyear Atomic Corpora- tion, Piketon, Ohio, to the Oak Ridge National Labora- tory, September 2, 1970. 17. Goodyear Atomic Corporation, Piketon, Ohio. 18. Personal communication. 19. G. L. Cathers, M. R. Bennett, and C. J. Shipman, MSR Program Semiannu. Progr. Rep. Aug. 31, 1968, ORNL-4344, p. 321. 4. CHEMICAL COMPOSITION OF THE FLUSH SALT DURING NUCLEAR OPERATIONS 4.1 Role of Flush Sait Analysis in the Determination of Salt Residue Masses The potential utility of a flush salt to scour contam- inants from the MSRE fuel circuit was recognized early in the MSRE development program; the sagacity of the choice to use this salt was confirmed immediately after preliminary experiments with the reactor — experi- ments in which the flush salt was repurified after its first period of circulation through the fuel system.! The repurification procedures employed at that time showed that in its first use the salt had served to remove 61 a considerable amount of contaminants from the fuel and moderator graphite system. Not foreseen, however, was the fact that it would function later to have a significant influence on evaluations of reactor perform- ance. This arose from the fact that the quantities of residues of fuel and flush salt, intermittently cross- transferred before and after reactor maintenance, could not be adduced from on-site inferences. Soon after experiments began with the circulating fuel salt, it became evident that only the results of highly accurate chemical analysis of the flush salt would make it possible to establish reliable values for the amounts of fuel and f{lush salt residues left in the reactor fuel system after it was drained. Accuracy in these values was also necessary in order to compute changes in nominal values of the concentration of fissionable material in the fuel circuit as the reactor experiment proceeded. Results of chemical analyses indicated that after the fuel salt was drained, circulation of the flush salt removed an average of 20 kg of fuel salt from the drained circuit. Refinements in the values for the physical properties of the fuel and flush salts allowed an estimation from the relative density of the flush and fuel salts that the average mass of flush salt residues would be ~17.5 kg. With these values, nominal values for the composition of the fuel salt were calculated for the period when experiments were conducted with 235¢ fuel. The total mass of flush salt originally charged into the MSRE was about 4190 kg. From averages of the increments in the changes in concentration of uranium which developed in the flush salt in use, it was expected that after its final use with ?°3U fuel the flush salt would contain uranium at a concentration of 1500 ppm, representing a total pickup of 6.29 kg of uranium. Analysis of the flush salt after its final use to remove 23517 fuel residue from the fuel circuit showed that the concentration of uranium was 1488 ppm and, there- fore, that the amount of uranium which was recover- able from the salt was 6.23 kg; the actual amount of uranium recovered was 6.420 kg, With the final amount of uranium in the flush salt established as 6.420 kg, the average increments were established as 0.917 kg, and the average mass of fuel residue as 20.0 kg. Complete results of the chemical analyses performed with the flush-salt samples removed from the MSRE are listed in Table 4.1. The average concentrations of uranium in the samples removed from the systemn during 2?3 U operations are shown in Fig. 4.1. Flush salt was circulated in the MSRE fuel-salt circuitry 13 times as a part of the 23°U operations and 62 Table 4.1. Summary of MSRE salt analyses, flush salt Weight % Parts per million Date Sample U4 Li Be Zr F b3 Fe Cr Ni ob Book Analytical PC-1 13.12 9.68 77.08 5993 45 45 7 4320 FP-3-36 13.70 9.71 0.021 0.0195 77.68 101.17 144 64 75 12050 12/15/65 FP-4-1 35b 12/16/65 Fp-4-2 . 13.65 9.83 <0.0025 0.0210 80.52 104.04 116 <19 33 562 12/16/65 FP4-3. 46° 12/16/65 FP-4-4 13.55 9.35 <0.0025 0.0207 79.34 102.30 212 62 30 740 12/17/65 FP-4-5 72¢ 12/17/65 FP-4-6 13.50 9.96 <0.0025 0.0200 80.07 103.58 125 <19 <20 180% 12/17/65 FP-4-7 13.65 9.46 <0.0025 0.0241 77.85 i01.03 i80 54 <20 1502 12/17/65 FP-4-8 106¢ 12/17/65 FP-4-9 13.35 9.98 <0.0025 0.0221 75.80 29.20 210 37 <20 1425 12/18/65 FP-4-10 9.49 <0.0025 0.0230 75.05 98.26 128 60 <20 13002 9/28/66 FP-8-1 12.40 9.79 0.02 0.0409 80.70 10299 262 75 31 9/29/66 FP-8-2 12.40 9.66 0.25 0.505 78.14 101.01 197 82 225 9/30/66 FP-8-3 225 9/30/66 FP-8-4 12.43 9.02 0.22 0.458 75.75 98.93 382 53 74 11/2/66 FP-8-15 9.34 <0.01 0.0578 78.74 128 65 <15 11/3/66 FP-8-16 13.50 9.24 <0.10 0.0715 78.01 100.94 125 61 <15 11/4/66 FP-8-17 13.40 8.62 <0.10 0.0556 73.63 95.81 76 52 <15 11/24/66 FP-9-8 13.35 9.59 0.24 0.0840 72.15 95.43 175 53 26 12/11i/66 FP-10-1 13.15 8.90 0.145 0.0828 72.67 9498 169 71 53 12/11/66 FP-10-2 13.10 8.41 0.190 0.6747 73.03 9484 194 75 42 5/10/67 FP-11-59 13.43 7.57 (.345 0.0292 76.10 97.50 222 68 40 5/10/67 P-11-59 0.9304 5/16/67 FP-11-59 0.1080¢ 511/67 FP-11-60 13.40 9.36 0.260 0.0268 79.23 102.30 119 74 34 6/16/67 ¥P-12-1 13.40 8.64 <0.2 0.0778 80.22 102.56 108 66 26 6/16/67 Fp-12-2 13.70 9.59 <0.2 (.0793 75.40 98.89 104 70 23 6/17/67 FP-12-3 13.60 9.38 <0.2 0.0826 77.40 100.68 93 68 24 6/17/67 FP-12-4 50-g sample for oxide analysis 9/8/67 FP-13-1 12.70 8.42 <0.2 0.1392 75.14 96.64 106 82 260 9/10/67 FP-13-2 12.80 8.84 0.72 0.1084 76.12 98.72 140 60 250 9/11/67 FP-13-3 12.80 10.04 0.70 0.1069 7772 10142 202 76 232 4/28/68 FP-14-72 12.53 9.66 0.42 0.1507 73.85 9664 150 76 52 4/28/68 FP-14-72 0.1488¢ 8/2/68 FST-11 0.1172 104 (before F5) 8/4/68 FS8T-12 24 ppm 176 119 (after Fy) 8/4/68 FST-13 7 ppm 232 112 516 (after F5) 8/5/68 FST-14 6 ppm 174 133 543 (after F5) 8/8/68 FST-15 No sample obtamned 8/8/68 FST-16 0.30 206 136 36 8/8/68 FST-17 50-g FVS capsule after Zr addition 174 100 50 8/10/68 FST-18 No sample obtained 8/14/68 FP-15-1 13.43 9.16 0.546 0.0036 152 76 26 8/15/68 Fp-15-2 50-6 sample for eta experiment — no sample retrieved 8/15/68 FP-15-3 25-g sample for eta experiment 8/16/68 FP-15-4 25-g sample for eta experiment 8/11/69 FV-19-15 8/12/69 FP-19-2 12.43 9.64 0.51 0.0072 79.7 102.30 128 81 23 8/13/69 FP-19-3 50-g sample for oxide determination 8/13/69 FP-19-4 12.50 9.53 0.41 4.0065 79.2 101.68 197 74 25 8/13/69 FP-19-5 11.55 7.55 0.49 0.0075 80.5 100.14 241 80 40 12/13/69 FP-20-33 11.55 8.19 0.430 0.0118 7740 103.59 146 93 49 12/13/69 FP-20-34 50-g sample of flush salt for oxide analysis 7t 12/14/69 FP-20-35 11.50 8.45 0.540 0.0108 7790 1034 155 88 46 aValues corrected to 33.696 at. % 23°U. dRecheck by fluorometric method. bKBrF4 method emploved unless otherwise noted. €Delayed neutron activation analysis. ¢HF purge method used. 63 . ORNL-DWG 71-9989 CTTTTTITTITTITITTTT] ® FLUOROMETRIC ANALYSIS 8 | _liiss ‘ A FLUOROMETRIC ANALYSIS, CONFIRM ® DELAYED NEUTRON ACTIVATION ANALYSIS 1400 = () EXPERIMENTAL AVERAGE ] e e b e {341 (1272) 1200 (7ss) T T'°7 B @ $ 1000 oo T 1 — & £ O 5 ) ? @ 800 iobzsol (808) & - ® — 603 : 600 & (616) l ® @ “nll‘a?“ unm-426 400 ® (424) & by o oo i o feEs o 1T G ---.249 (215) 9 & ¢ 200 (g F-26 » O o P TT T T YT T I TRt an s Mt ITTTTODOODoDoO Lo NN 0n T 0 o O o a o 000 o o0 oo ad ol 0O oo 4o A | VRS 1 T W R I SN W SV N A U S TN I TN NN N 5 SN ¥ N S ¥ A I ENO ¥ W ¥ N W UV I ER § SO N SAMPLE NUMBER Fig. 4.1. Uranium concentration in MSRE flush salt. Calculated average mcrement = 177 ppm (includes transters of five fuel residues totaling 11.34 kg from fill line). three times as part of 233U operations, once on each occasion before the fuel was constituted in the reactor system, and thereafter whenever the fuel system was to be opened for maintenance. If during the maintenance period, the possibility developed that the fuel circuit was exposed to ambient cell atmosphere, the circuit was again washed with the molten flush salt before resump- tion of experiments with the fuel salt. In use, the flush salt was circulated for brief periods, usually 24 to 48 hr, because it was assumed that molten LiF-BeF, flowing in the metal circuit at 650°C would be extremely effective as a solvent for contaminants that may have entered the system during the previous maintenance period. Gross exposures of the circuit to potentially contaminating atmospheres were avoided during the maintenance periods first by flushing the reactor system with argon, and subsequently by maintaining a nitrogen atmosphere in the standpipe column through which access to the reactor core was gained. Neither blanket gas was specially purified for this application. It was judged that the flush salt was circulated for sufficient periods of time to be effective, because, in use, major increases were not observed in the concen- tration of contaminants in the salt nor were there major increases in the concentration of chromium in the fuel salt on initiation of the subsequent experiments with fuel salt. After each of these occasions, however, the chromium concentration of the fuel salt gradually rose to a higher value than had preceded the maintenance period. The significance of these changes is discussed in Sect. 6. 4.2 Transfer of Uranium and Plutonium to Flush Sait in ?*2 U Operations Reclamation of 6.420 kg of uranium from the flush salt indicates that the average increase in uranium concentration developed from the removal of a single fuel-salt heel during 2*°U operations was 219 ppm rather than the average 215 ppm we had observed from analyses performed during the course of reactor opera- tions. With the 233U fuel salt, a reduction of the conceniration increments was anticipated proportionate to the lower concentration of uranium in the fuel salt, and should have resulted in an average increase of ~40 ppm per flush operation. The average of the two results obtained with the flush salt after completion of run No. 20 shows that the final concentration of uranium in the salt was 113 ppm, or that for each of the three times flush salt was used to remove fuel salt residues, the average increase in the concentration was 38 ppm, in good agreement with the anticipated value. 64 Table 4.2. Results of mass spectrometric analysis of MSRE flush salt In weight percent Uranium 233 234y 235y, 236y 238y Sample No, FP-20-33 40.55 3.73 17.67 0.25 37.80 FP-20-35 38.20 3.47 17.06 0.24 41.03 Plutonium Sample No. 233py 240py 241py 242py FP-20-33 94,74 4.78 0.44 0.04 FP-20-35 94.63 4.87 0.45 0.05 After termination of experiments with 233U fuel the reactor was drained and flushed. Two samples of flush salt were obtained (FP-20-33 and -35) in which the average concentration of uranium was 113 ppm, an increase of 42 ppm over that found during the single previous use in which it removed a fuel-salt heel in 2331 operations (the beginning of run No. 19). The single flush-salt sample obtained during the startup operation preceding the introduction of 233U fuel into the system was reported to have contained 36 ppm of uranium. The three groups of data are thus consistent — an indication that the concentration of uranium in- creased as predicted. They also suggest that the flush salt contained 35 to 40 ppm of uranium when it was returned from the chemical reprocessing plant. Isotopic analyses obtained from the last two samples of flush salt are shown in Table 4.2. The isotopic composition of the uranium contained in these samples indicates that (correcting for the contribution from the fuel of 0.023 kg and neglecting a correction for 238U burnout) the flush salt contained ~50 ppm (0.21 kg} of uranium. This is somewhat more than one would expect considering the excellent agreement between expected and observed amounts of uranium previously recovered from the flush salt, and from analyses obtained during the uranium recovery experiments with flush salt in which results indicated the residual concentration was no higher than 24 ppm, and possibly as low as 4 ppm. 4.3 Flush Salt Loss to Off-Gas Holdup Tank The results of the chemical and mass spectromeiric analyses described above provide a basis for the conclusion that only very small amounts of either fuel or flush salt could have been lost from the primary containment or came in contact with the cell atmos- phere through the leak that developed adjacent to the freeze valves (see Sect. 3.6.3). If cell air were aspirated into the fuel salt as it drained from the reactor at termination of run No. 20, a probable consequence would have been that ZrO, deposits may have formed in the region of the break, later to be dissolved in the flush salt. Possibly, some uranium would have been deposited as well, along with the ZrO,. Small amounts of ZrF, are transferred to the flush salt as a part of fuel residues, yet, as shown in Table 4.1, the concentration of zirconium in the flush salt after its final use does not seem to have increased. Second, sizable amounts of uranium deposited in the area of the break would both increase the anticipated concentration in the flush salt and alter the isotopic ratio in favor of an anomaious increase in 223U fraction. The isotopic analyses shown in Table 4.2 denote a slight anomaly in isotopic ratios of uranium, but of opposite character than suggested above; that is, they indicate that slightly more 235U and 23%U were present than was expected from the prior reprocessing operations. In one previous incident, the possibility developed that flush salt was inadvert- ently? lost from the salt system. In Piper’s account® of operations in July 1966, he notes that the fuel loop was filled with flush salt in preparation for maintenance work in the reactor cell. Flush salt was transferred into the overflow tank to flush out the residual fuel and to check the indicated level in the pump bowl at the overflow point, and transfer began when the indicated level was 9.6 in. (In January the indicated level at the overflow point had been 9.2 in.) After approximately 0.6 ft*® of salt was transferred, the fuel pump level suddenly rose off scale. Rising level in the overflow tank triggered a drain, but not before salt had entered some of the lines connected to the top of the pump bowl. This behavior resulted from an operational error which allowed the salt level in the flush tank to be lowered too far, and also allowed the pressurizing gas to enter the fill line and pass up into the reactor vessel. The gas expanded rapidly as it rose through the salt, causing salt to flood the pump bowl. The pump bowl reference line was plugged with frozen salt, and enough salt was frozen in the sampler tube to obstruct passage of the latch. A thermocouple indicated that some salt also entered the off-gas line, but this line was not plugged. Salt also froze on the annulus around the fuel-pump shaft, preventing its rotation. In subsequent maintenance operations electric heaters were applied to the outside of the lines to melt out the salt in the bubbler reference line and the sampler tube. The short flexible portion of the off-gas line was replaced because of uncertainty over the possible effects of salt in the convolutions. 65 We noted above that for the increments observed in the change of concentration in the flush salt the projected amount of uranium recoverable for the salt was ~0.3 kg, and that a total of 6.42 kg was recovered. The projected value was calculated from the final analytical values, 1488 ppm X [4187 + 7 X 2.6 kg (the average mass increase from fuel-flush cross transfer)] = 6.257 kg. The fact that 6.42 kg of uranium was recovered suggests that if the analytical data are free of bias, only a small volume of flush salt could have transferred to the off-gas in the overflow accident. For example, if 50 kg of flush salt were lost by transfer the maximum amount of uranium which could be re- covered from the flush salt at a concentration of 1488 ppm would not exceed 6.18 kg. Although it cannot be contended that the above results unequivocally demon- strate that the amount of flush salt transferred was negligible, they provide convincing evidence that the amount was indeed small. In retrospect, the flush salt has served as the source of several kinds of valuable information pertaining to the performance of the MSRE. It now seems evident that it could have been used more effectively as a more valuable source if the periods of its use had been extended and designated for additional chemical investi- gations. References 1. R. B. Lindauer, Processing of the MSRE Flush and Fuel Salts, ORNL-TM-2578 (August 1969). 2. H. B. Piper, MSR Program Semiannu. Progr. Rep. Aug. 31, 1966, ORNL-4037, p. 24. 5. CHEMICAL BEHAVIOR OF THE COOLANT SALT Of some 7000 kg of LiF-BeF, (66-34 mole %) prepared for the MSRE, 2610 kg was designated for use as the coolant salt and the remainder as flush salt. This coolant was circulated by a 75-hp motor and pump through a circuit which included a tube and shell heat exchanger and an air-cooled radiator. Detailed descrip- tions of this coolant system are given in the MSRFE Design and Operations Report.' At termination of the Molten-Salt Reactor Experiment the salt had circulated in the system for a period of ~26,000 hr. With the reactor at full power, coolant salt entered the tube bundle of the heat exchanger (Fig. 5.1) at a tempera- ture of 546°C (1015°F) and at 77 psig, removing heat from the fuel salt as it circulated through the shell side of the exchanger, and flowed on to the air-cooled === p——= sz cEr ) e THERMAL-BARRIER PLATE e == TUBE SHEET = \\\\\\\\\\\ COOLANT INLET %‘ = T-STREAM COOLANT QUTLET LAN SEPARATING BAFFLE FUEL OUTLET i) \\\\\@\\ 66 ORNL.-LR-DWG 52036R2 FUEL INLET i/2-13.-CD TUBES = CROSS BAFFLES 16.4-1n 0D x 0 2-in. WALL x 8-ft LONG Fig. 5.1. Primary heat exchanger. radiator (Fig. 5.2). On reentry to the heat exchanger, the salt was at a temperature of 579°C (1075°F). The coolant was circulated under these conditions for approximately half the period of its use. At zero or low power the salt was circulated isothermally, at about 650°C (1200°F). The coolant pump operated at con- stant speed, developing a salt flow of 6286 kg/min (793 gpm). 5.1 Composition Analysis The salt was circulated for some 1200 hr during the prenuclear test period. Since flush and coolant salts were supplied from the same inventory, the coolant salt was analyzed during this period only for impurities. At the end of prenuclear testing, concentrations of the structural metal impurities, chromium, iron, and nickel, were established to be approximately 30, 90, and 7 ppm respectively — remarkably low considering that the circuit had not been flushed previously with molten salt. The results of spectrochemical analyses showed that no significant amounts of additional impurities were introduced during this period. As the reactor was brought to full power early in 1966, coolant salt was again circulated, sampled regularly, and subjected to compositional and impurity analysis. The results of these analyses, together with all other analyses of the coolant salt, are listed in Table 5.1. It was stipulated that the chemical analyses include a determination of zirconium concentration as part of the procedures for coolant-salt analysis. Although such analyses were not conceivably applicable as primary indicators of fuel leakage, their incorporation supplied base-line data for the contingency that possible analysis 67 ORNL-LR-DWG 55841R2 PENTHOUSE DRIVE CHAIN . [/ DOOR DRIVE MOTOR AND GEAR REDUCER | ’k SUPPORTING STEEL ~fl < WIRE ROPES — w~ M B =y DOOR CAM GUIDE 4 3 BLDG 7503, FIRST FLOOR (ELEV 852 ft-0n) SHOCK SPRING B R . N 0 . AIR BAFFLES — ] ) i . | / | a \/\;«\ . AIR INLET DUCT ~~_ " §T~AIR QUTLET DUCT - / . 3 /" AIR DUCT FLANGE — Al Goane e L ‘ b ‘ 3 | / { - ~ . MAIN AIR BYPASS DUCT fi 5 RADIATOR ENCLOSURE RADIATOR TUBES AIR FLOW - DAMPER Fig. 5.2. MSRE radiator coil and enclosure. Table 5.1. Summary of analytical chemical data for the MSRE coolant circuit 68 Date Total Number Sample Li Be Zr F Cr Fe Ni 0 Hr, in circuit Design (wt %) 1/9/65 0.3 CP1-1 - - - - 20 117 8 420 1/10/65 24 CP1-2 - - - - 25 181 <5 433 1/11765 b4 CP1-3 - - - - 26 184 <5 472 1/13/65 94 CP1l-4 - - - - 27 96 12 580 1/15/65 143 cPl-5 - - - - 27 83 <12 325 1/17/65 186 CPl-6 - - - - 27 93 <12 525 1/20/65 267 Run 1-F 1/22/65 267 Run 2-F 1/22/65 277 CP2-1 - - - - 30 63 5 388 1/27/65 393 Cp2-2 - - - - 29 101 7 270 2/1/65 514 CP2-3 - - - - 28 65 25 263 2/11/65 757 CP2-4 2/18/65 913 cp2-5 2/24/65 1062 CP2-6 3/5/65 1249 cp2-7 3/5/65 1249 Run 2-F 6/26/65 1249 Run 3-1 6/26/65 1250 CP3-1 6/30/65 1342 CP3-2 6/30/65 1364 CP3-3 12.87 9.53 - 33 78 38 695 7/1/65 1370 Run 3-F 1/8/66 137¢ Run 4-I 1/8/66 1371 CP4~1 36 65 <5 - 1/18/66 1375 CP4-2 26 83 <5 130 1/20/66 1423 CP4-3 13.78 8.91 16.50 40 40 40 190 1/20/66 1431 CP4-4 50 g sample for oxide analysis 25b 1/25/66 1544 CP4-5 14.20 8.87 <0.002 76.80 50 50 20 180 1/25/6% 1558 CP4~6 50 g sample for oxide analysis 38P 1/27/66 1593 Run 4-F 2/7/66 1593 Run 5-I 2/7/66 1600 CP5-1 13.86 8.85 0.0015 76.70 35 5 <10 150 2/10/66 1664 CP5-2 13.82 8.55 0.0016 76.60 35 <2 <10 110 2/15/66 1784 CP5~3 12,76 10.51 0.0062 76.04 46 57 64 65 2/18/66 1865 CP5-4 12.60 9.17 0.0175 76.80 35 119 14 <20 2/23/66 1977 CP5-5 12.60 9.15 0.0153 76.70 42 87 28 120 3/2/66 2176 CP5-6 13.01 10.06 0.0050 77.20 20 100 8 260 3/4/66 2219 Cp5-7 13.17 9.49 0.00%9 76.40 24 112 <2 171 3/8/66 2331 CP5-8 12.70 10.22 0.0069 77.20 29 36 <2 71 3/11/66 2406 CP5-9 12.77 9.43 0.0105 76.70 32 44 <2 35 3/15/66 2482 CP5-10 12.97 5.59 0.0016 76.90 33 31 <2 <20 3/20/66 2560 CP5-11 12,92 9,61 0.0032 76.40 36 49 53 <20 3/22/66 7648 CP5-12 12.75 9,74 0.0030 76.40 44 57 8 <20 3/25/66 2722 CP5~-13 13.10 9.98 0.0997 76.20 32 73 10 162 3/25/66 2730 Run 5-F 3/29/66 2730 Run 6-1 3/29/66 2825 CP6~1 13.09 9.73 0.0031 76.10 30 43 41 <20 4/1/66 2907 CP6-2 12.68 9.56 0.0220 76,40 38 28 9 <20 4/20/66 3321 CP6-3 12.82 9.13 - 76.80 45 76 47 210 4/24/66 3431 CP6~4 12,76 9.63 0,0055 76.00 57 68 15 210 5/16/66 3946 CP6~5 12.70 9.79 0.0095 74.80 20 57 6 135 6/8/66 4499 CP6-6 - 9.65 0.0050 74.80 28 37 6 <40 6/8/66 4499 Run 6-F 6/28/66 4499 Run 71 6/28/66 4807 cP7-1 13.64 8.72 0.0032 14.65 32 60 <2 <20 6/29/66 4826 cp7-2 12.16 9.88 0.0070 74.50 31 51 <2 300 7/5/66 4967 CP7-3 12,39 10.24 0.0096 74.80 51 80 <2 145 7/12/66 5134 CP7-4 12.70 9.26 0.0008 75.90 46 24 28 85 7/22/166 5374 Run 7-F 10/2/66 5374 Run 8-1I 10/2/66 5376 CcPg-1 12.90 9.94 0.0008 74.55 17 23 <5 53 10/25/66 5929 cPg-2 12.40 9.35 0.0013 75.70 24 19 <2 <25 10/25/66 Run 8-F 11/8/66 Run 9-1 11/8/66 6264 CP%-1 12,30 9.71 0.0012 17.40 4 30 29 73 11/22/66 6608 CP9-2 12.65 9.55 <0.0005 74.23 33 25 <29 75 11/22/66 Run 9-F 12/7/66 Run 10-1 12/7/66 6962 CPi0-1 i1.90 9.64 ¢.0033 74.80 30 36 <5 75 1/5/67 7655 CP10-2 12.80 9. 40 0.0032 74.50 20 653 34 475 1/5/67 Rur 10-F 2/2/67 Run 11-1 2/2/617 8570 CP11-~1 12.50 9.73 <0.,0020 77.00 18 78 <10 330 2/16/67 8798 CP11-2 11.90 9.57 0.0017 76.50 23 35 <10 530 3/7/67 9258 CP11-3 13.20 9.46 0.0090 76.00 13 105 15 380 3/21/67 9751 CP11-4 4/13/67 10145 CP11-5 12.90 9.70 <0.002 76.00 32 24 <10 - 5/11/67 10836 CP11-6 13.50 9.46 0.0060 76.10 35 35 14 314 5/11/67 Run 11-~F 6/16/67 Run 12-1 6/16/67 10848 CPl2-1 12.70 9.30 0.0074 77.40 39 168 10 - 7/6/67 11346 CP12-2 14.90 9.10 <0.0026 75.10 26 33 <20 88 7/11/67 11460 CP12-3 12.73 9.40 <0.0040 74.30 30 27 26 230 7/11/67 Run F-12 69 Table 5.1 (continued) Date Total Number Sample Li Be Zr F Cr Fe Ni 0® Hr. in circuit Design wt %) (ppm) Run 13,14-I g;;?égy 12150 CP13-1’ 12.40 9,27 0.0045 73.80 <15 62 zgg igg 9/29/67 12550 CP1l4-1 12,60 9.87 0.0098 72.80 24 <10 20 e 10/25/67 13174 CP14-2 14.30 8.86 0.0011 76.32 24 25 o o 11/27/67 13966 CP14-3 15.20 9.35 0.0100 75.50 24 61 o g 1/18/68 15214 CPl4-4 12.50 10.10 0.0019 76.10 25 2; Pt o 3/20/68 16702 CP14-5 14.60 9.61 0.0078 17.20 <15 3 3/28/68 16894 Run 1l&~F 9/ /[68 Run 15-1 9/17/68 17714 CP15-1 11/27/68 19414 Run 15-F 12/ /68 Run 16-1 12/10/68 19524 CP16-1 12/ /68 Run 16-F 12/6 Run 17-1 ;§4§é99 20076 CpP17-1 13.10 9.26 <0.0020 76.60 26 21 <1; zggg 3/6/69 20796 CP17-2 13,40 9.49 0.0021 77.00 25 15 Run 17-F Run 18-I 4/27/69 22044 Cp18-1 13.00 9.52 ¢.0013 77.60 25 12 ;5 ;gg 5/29/69 22812 CP18-2 13.70 9,22 0.0014 77.50 25 14 <10 22860 Run 18-F 2860 Run 19-1 8/10/69 52370 {rie-1 13.40 9,44 <0.0019 77.30 35 <5 10 1820 9/14/69 22966 Cr19-2¢ 13.40 9.25 0,112 77.60 35 143 2 9/14/69 22966 CcP19-3¢ 13.85 9.13 0.103 78,30 35 128 11/2/69 23084 Run 19-F 11/26/6 23084 Run 20-I llfiZGifig 23094 CP20-1 14,00 9.55 0.0019 77.20 46 32 11 3200 12/10/69 23430 Cr20-2 13.70 9.31 0.0206 77.20 60 114 10 1700b 12/11/69 23454 CP20-3 Ni bar, exposed to coolant for tritium experiments 12/11/69 23454 CP20-4 N1 bar, exposed to coolant for tritium experiments 12/12/69 23479 CP20-5 Cu0 specimens, exposed to coolant for tritium experiments. 12/16/69 23479 CP20~6 Cul specimens, exposed to coolant for tritium experiments. 12/12/69 Run 20-F Lnless otherwise noted, oxygen analyses were obtained by the KBrF a b Results obtained by HF~Hp transpiration method, < Samples analyzed bv HRLAL personncl. of fuel-coolant mixing accidents might arise. Results are omitted from Table 5.1, because in none of the analyses of the coolant salt was zirconium detected. As noted in Sect. 3.3, procedures used for the determi- nation of the concentrations of oxides in MSRE samples were unsatisfactory until development of the HF-H, transpiration method was completed. Not until late in the program of reactor operations was this method applied to coolant-salt samples, and then, for the single sample available, it produced questionable results. The results of analyses for oxides listed in Table 5.1 were obtained with the KBrF, method, a method which is generally satisfactory for nonhygroscopic materials, but which was applied to samples which are hygroscopic and had not remained isolated from ambient atmo- spheres after their removal from the reactor. In view of the established absence of corrosion in the coolant circuit, their credibility is dubious. Analyses of coolant salts were conducted by the General Analysis Laboratory of the ORNL Analytical Chemistry Division, while those for the fuel were obtained from the High-Radiation-Level Analytical Laboratory of that Division. 4 method. 5.2 Corrosion Behavior Corrosion of Hastelloy N as a container for molten LiF-BeF, mixtures may originate from a very limited number of sources: from impurities in the melt, from oxide films on the metal, and from mass transfer of metal constituents in the fluoride. Of these sources, only the latter, which is caused by the differential temperature coefficient for solubility of metals in salts, affords a mode of continuous attack in reactor systems that are protected from inleakage of contaminants. Since the procedures adopted for MSRE operations assured that the circuit would be free from chronic sources of contaminants, it was anticipated that surveil- lance of the coolant salt would indicate imperceptibly low corrosion throughout the experiment. In the early stages of MSRE operations, samples of the flush salt were removed from the circuit at a rate of one per week. However, as our experience developed, and it was confirmed that chroni¢ sources of oxidizing impurities were absent, the frequency was decreased, finally to intervals of a month or longer. In contrast to the fuel system, the coolant salt system was not ORNL-DWG 71~9990 &0 50 b 40 e S 20 2 @ 0 5 10 15 20 { x10%) CUMULATIVE CIRCULATION PERIOD (hr) Fig. 5.3. Concentration of chromium in the MSRE coolant circuit salt. exposed to the cell atmosphere during the course of MSRE operations. Under these circumstances, it was reasonable to expect that corrosion would not occur in the circuit, and these expectations were borne out in MSRE operations as indicated by the results of chemi- cal analyses, and later in the postoperational examina- tions of the radiator (see below). The results of the chemical analyses performed with coolant-salt samples showed that no measurable in- crease in the concentration of Cr, Fe, and Ni (with average values of 32, 64, and 20 ppm respectively) developed after the coolant salt was first charged into the MSRE. The concentrations of chromium in the coolant-salt samples, as determined in chemical analy- ses, are shown in Fig. 5.3. The fact that the chromium concentration remained unchanged is remarkable, since it indicates that within the limitation of the analytical precision (7 ppm) no corrosion, excepting that pos- sibly resulting from mass transfer, occurred in the coolant circuit during the entire period of MSRE operations. The demonstrated compatibility of the coolant salt with its containment alloy is believed to be unmatched in any prior experience with either molten salts or liquid metals as recirculating heat-exchange media. The average value of the chromium concentration in the coolant salt remained at 32 ppm for the entire period of reactor operation. The apparent trend toward higher values toward the end of the use period, as is reflected in Fig. 5.3, tends to discount the credibility of the view that the system remained free of contami- nants. However, the results of postoperational examina- tions support this view completely and indicate thereby that continued use and analysis of the salt would have shown that the data were not indicative of a trend, but rather represented normal scatter. As part of the postoperational examination of MSRE components, sections of the radiator tubing were removed from the inlet and outlet ends and tested by ORNL metallurgists. Their analysis of the surfaces of the alloy that were exposed to the coolant salt by electron microprobe techniques disclosed that no compositional variations in Cr, Mo, Ni, or Fe existed within 2 u (the allowable working range on these samples) of the surfaces.? Comparisons of the chemical composition of sections of this tubing with machined turnings showed a higher concentration of carbon at the salt interface after service, as was suggested from examinations of the microstructure.> However, changes in the tensile strength of the alloy were found to be slight and did not indicate serious embrittlement of the alloy. The chemical behavior of the coolant salt in the MSRE is a unique demonstration of the compatibility of pure fluoride mixtures with nickel-based alloys, and shows as well that, with normal precautions to ensure that oxidizing contaminants are prevented from enter- ing these systems, corrosion will not occur. The behavior of the coolant salt in the MSRE is a benign indicator of the potential application of LiF-BeF, based systems, but has only marginal implication with respect to the coolant salt of larger-scale molten-salt reactors in which lower liquidus temperatures are mandatory for the coolant, and where the capital cost of the coolant salt will have a pronounced effect on power economics. These factors seem to preclude the choice of 7LiF-BeF, mixtures as coolants in such reactors. The current choice for the MSBR coolant salt is the NaF-NaBF, eutectic mixture. The highly success- ful performance of the molten fluoride salt "LiF-BeF, in the MSRE is, however, of special significance in that it leads to the recognition that the possibilities of materials combinations are more flexible than was estimable prior to operation of the MSRE. References 1. R. C. Robertson, MSRE Design and Operations Report, Parr I, ORNL-TM-728 (January 1965). 2. H. E. McCoy and B. McNabb, MSR Program Semiannu. Progr. Rep. Aug. 31, 1970, ORNL-4622, p. 120. 6. CORROSION IN THE FUEL CIRCUIT 6.1 Modes of Corrosion Early recognition that numerous inorganic fluorides are among the most chemically stable materials in nature supplied the initial impetus for investigation of the chemical feasibility of molten-salt reactors. It was noted that the fluorides of commonplace alloy con- stituents such as iron, nickel, chromium, and molyb- denum were less stable than the various compounds which might serve as components of molten-salt reactor fuels and coolants, and thus good alloys were poten- tially available as containers for molten fluoride mix- tures (see Table 6.1). Subsequently, a materials develop- ment program was initiated culminating in the develop- ment of the alloy now designated as Hastelloy N specifically for use in constructing the MSRE; its approximate composition was Ni-Mo-Cr-Fe (71-17-7-5 wt %). The basis for the selection of this exact composition is described by Taboada' in a detailed review of the alloy development program. Since the major components of the MSRE salts, LiF, BeF,, ZrF,, and UF,, are much more stable than the fluorides which can result from the corrosion of Hastelloy N, that is, MoF;, NiF,, and FeF,, compati- bility of the molten-salt mixtures and the alloy was essentially assured. The most probable mechanisms by which corrosion might occur in the MSRE were identified and examined extensively prior to operation of the reactor. The chemistry of such processes was reviewed recently by 71 Table 6.1. Relative stability? of fluorides for use in molten-salt reactors Free energy Absorption cross Compound of formatoion N}I?eoiit;r;g section? for at 1000°K ©C) thermal neutrons {kcal per F atom) {barns) Structural metal fluorides CrF, 74 1100 3.1 FeF, ~66.5 930 2.5 NiF, —58 1330 4.6 MoF g =50 17 2.4 Diluent fluorides CaF, —125 1330 0.43 LiF —125 848 (.033¢ Bak, —124 1280 1.17 StF, —-123 1400 1.16 CeF4 —118 1430 0.7 YF; ~113 1144 1.27 MgF, —-113 1270 0.063 RbF -112 792 0.70 NaF —112 995 0.53 KF - 109 856 1.97 BeF, —104 548 0.010 ZrF, —94 903 0.180 AlF 4 -90 1404 0.23 SnF, ~62 213 0.6 Pbl, —-62 850 0.17 BiF, -50 727 0.032 Active fluorides ThE, —101 it UF,4 -95.3 1035 UF, -100.4 1495 9Reference state is the pure crystalline solid; these values are, accordingly, only very approximately those for solutions in molten mixtures. bOf metallic ion. €Cross section for 7Li. Grimes.? In the absence of impurities, corrosion pro- ceeds through two mechanisms, by mass transfer and by selective oxidation of chromium, the most chemically active constituent of the container alloy. Both types of corrosion are limited by self-diffusion of chromium in the alloy and by temperature, which in the MSRE fuel reached extremes of 654°C (1210°F) and 632°C (1170°F). Preponderantly, oxidation-reduction reactions are re- sponsible for all the corrosion in molten-salt reactor fuel systems. The principal reaction which controls the process is UF,4(d) + %, Cr(e) = UF5(d) + % CrF, (d) (1) for which the standard free energy at 1000°K (727°C) is +15.1 kcal; at this temperature the equilibrium constant for the reaction is Ny, Netk, ) K= =5 X 1074, (2) Nyr, ac; where NV is mole fraction of a reactant or product and « is the activity coefficient. For purposes of examining the way this equilibrium affects corrosion behavior, assume conditions similar to those existent with 233U fuel, where Ny, = 1.38 X 1073, Cr** = 5.15 X 1073 (150 ppm), an, = 3 X 10722 The equilibrium concentration of UF; thus becomes 2.832 X 107° and Nyg,/Nyp, =2.832 X 107°/1.38 X 107 = 0.205% . These results illustrate that the reaction tends to come to equilibrium with a small fraction of the uranium in the trivalent state. Chemical factors which remove UF, from the salt system promote the forward reaction (1) and tend to remove chromium from the container atloy by selective oxidation. Fission in molten-salt reactor fuel systems causes a gradual increase in the oxidation potential in molten- salt fuels (based on UF,) as uranium is consumed. From estimates of fission yields* and probable oxida- tion states of the species produced, the effect of fuel burnup on corrosion can be evaluated. A recent appraisal of this balance indicates that when reducing conditions are maintained and xenon and krypton are removed rapidly from the fuel, the sum of the electrical charges on the fission product cations is less than an average value of +4 per mole of uranium burned, and ~0.76 equivalent of oxidation results from the fission of one gram atomic weight of uranium.’ It was useful for this reason to increase the Ny /Nyp, concen- tration in the fuel salt occasionally as a means of IINITNIZIng corrosion. Self-diffusion coefficients of chromium in nickel-base alloys in the temperature range 600 to 900°C were determined by Watson and co-workers® by monitoring the total intake of %! Cr by the alloys exposed to salt solutions containing this radiotracer and by measuring the tracer concentration profiles through successive electropolishing of the specimens. They found from the loop experiments that the diffusion coefficient of 5'Cr in Hastelloy N at 650°C (1200°F), the mean temperature of the MSRE fuel, was 1 X 107'* cm?/sec. In recent examination of metal surveillance specimens from the MSRE core,” evidence 72 ORNL-DWG 70-4833 7 == -1 — & ,// // \e 8 y Z / 14 2 S s / MEASURED (0 = 4 x 107 '® cm2/sec}——— l_ < / S = 4 7~ 7L CALCULATED (£ = 2 x 10~ ** cm2/sec) 2.0/ O / / 2 !/ = 4 ~ o / [nag T / [GI )/ 0 0 1 2 3 4 DISTANCE FROM SURFACE (mils) Fig. 6.1, Chromium gradient in Hastelloy N sample exposed in MSRE core for 22,533 br. was found that indicated the diffusion coefficient of chromium in the core specimens may have been as high as 4 X 107'* ¢m?/sec (Fig. 6.1). The fact that the data from the MSRE samples were of necessity obtained from hot-cell operations, whereas no such constraints were imposed in the original laboratory measurement of self-diffusion coefficients, lends somewhat greater credi- bility to Watson’s values. Both values may be accurate, however, because Watson noted that for Inconel the self-diffusion coefficients of chromium were strongly dependent on annealing conditions at low temperatures. Conditions leading to large grains led to low diffusion coefficients and vice versa. Extrapolation of Watson’s data for Hastelloy N (see Fig. 6.2) indicates that the self-diffusion coefficient for Cr in Hastelloy N at the lowest temperature of the MSRE fuel circuit is 8 X 107! ¢m?/sec. The rate of mass transfer is limited by the rate of diffusion of chromium into the alloy in the coolest area of the system. Corrosion from mass transfer was therefore expected to be of little consequence in the MSRE.* Fick’s equation, M, = 2Cy(Dt/m)'/? (where C, = concentration of the diffusing element in the bulk species, D = diffusion coefficient, # = time) can be used to predict the quantity of material removed by dif- *Postoperational examination of sections of Hastelloy N from the heat exchanger disclosed no evidence of an enrichment of chromium at the salt-metal interface. This observation has led to a suggestion by F. F. Blankenship that since the coolest location in the fuel circuit is in the area of freeze valve 103, it would be useful in a future examination to determine the profile of chromium concentration in the alloy at this location. 73 ORNL-LR-DWG 46403 TEMPERATURE, (°C) 850 800 750 700 N L A COEFFICIENTS BASED ON TOTAL SPECIMEN COUNTS ® COEFFICIENTS BASED ON Cr3' CONCENTRATION — GRADIENT MEASURED BY COUNTING ELECTROLYTE © COEFFICIENTS S8ASED ON Cr®' CONCENTRATION @ GRADIENT MEASURED BY COUNTING SPECIMEN -2 |8 ’g @ 3 \\\O\ @ oo e LN © 0 L A \ Q 4a A\ka\\i‘ 2 A g 8 . T~ T N A o] \A fa OBTAINED UNDER NON-CORROSIVE CONDITIONS ¢ ° T8 BY ADDITION OF Cr*F, TO NoF-Zrf, l -14 8.75 9.00 9.25 9.50 9.75 10.00 10.25 10.50 ‘O’OOO/TPK_') Fig. 6.2. Self-diffusion coefficients for >1¢r in INOR-8; loop 1248, (from ref. 3). Table 6.2. Comparison of predicted and observed amounts of chromium in 2350 fuel salt Diffusion coefficient = D Predicted Time (Grams Parts per million Observed o) po1x10® p=2x107* p=4x10"* p=1x10* p=2x10""* p=4axio* rm) cmy/sec cmz/sec cmz/sec cmz/sec em? [sec cm2/sec 1,000 263 372 525 54 76 108 48 2,000 371 525 743 76 108 152 48 4,000 548 775 1097 112 159 225 48 6,000 643 910 1286 136 186 263 48 8,000 743 1050 1486 152 215 304 62 10,000 337 1184 1674 171 242 343 64 12,000 909 1286 1819 186 263 372 64 14,000 982 1387 1965 201 284 402 70 16,000 1050 1486 2101 215 304 430 72 18,000 1114 1575 2227 228 322 455 82 20,000 1174 1661 2349 240 3490 480 82 fusion under conditions where the surface concen- tration of the diffusing element is zero. Table 6.2 lists the amounts of chromium which might have been expected to appear in the fuel salt during 23°U operations if all the chromium accessible to the salt were oxidized to CrF,. It is apparent from these data that generalized corrosion, as inferred from increases in chromium concentration of the fuel salt, was only one-fourth to one-third of that expected from the diffusion coefficients. By contrast, the apparent corro- sion rates that were observed during the initial stages of MSRE operations with 233U fuel and again in the beginning of run 19 were approximately 5.4 and 3.84 times as rapid as predicted using a diffusion coefficient of 2 X 107** cm?/sec. Such rates were inconsistent with previous observations and more rapid than the maximum allowed by the maximum value observed for the diffusion coefficient, 4 X 1071% c¢m? /sec. Since the rates of corrosion on these occasions were unlikely to be permitted by chromium diffusion, it seems likely that bulk diffusion involving other constituents of the alloy was operative. From the following argument it is concluded that the anomalously rapid corrosion during these periods and occasionally during 22>U operations was caused by contaminants introduced into the fuel system during periods of maintenance. 6.2 Corrosion in Prenuclear Operations As noted in Chap. 2, chemical analyses were per- formed concurrently in the General Analysis Labora- tory and in the High-Radiation-Level Analytical Labora- tory with salt samples removed from the MSRE during the period before nuclear operations began. This pro- cedure reflected that a slight operational bias existed in the results of the two laboratories® for some of the species and enabled us to use those biases as correction factors for assessment of changes in the salt. With respect to chromium analyses as a corrosion indicator, the data corrected by this factor showed that at the average value of its concentration, 37 + 7 ppm, no change in the chromium concentration of the fuel circuit salt had occurred throughout the period ending with the completion of the zero-power experiments. In the low-power experiments following the first main- tenance period, during which time the pump rotor was removed for examination, the new average value of chromium in the fuel salt was 48 * ppm. Since the standard deviations for these two values overlap, it can be inferred that the flushing operation preceding the low-power operation (run 4) effectively removed corrosion-inducing contaminants that might have entered the system during the maintenance period. 6.3 Corrosion in Power Operations In Fig. 6.3 the results of all chemical analyses of fuel salt samples for chromium are summarized for power runs. The scattered data are described most simply as increasing linearly with time and correspond io an increase in chromium concentration in the salt at a rate of 12 ppm per year. Statistical analyses of the indi- vidual groups of data, however, indicate that the average values (least-squares method) are as listed in Table 3.3. In operation of the reactor with 233U fuel, the concentration of chromium in the samples of fuel salt increased sharply only during two periods, during the initial stages of run Nos. 15 and 19. One interpretation of these analytical results is that corrosion rates exceeded those predicted by diffusion data only after ORNL-DWG 70-2164 150 140 ¥ 130 - b — 120 L - 0.30- 10 - 0.45 100 . 0.25- “( = - 0.40 1ol & 90 . e / = , - 0.20- . ff 2 80 — R P -0.35 . , o ; = . 1] & ‘2 /e s . . - 8 70 5 * .| =i O 1 |—_ . T . P B A A ' -0.30 £ . . I -{/‘— o --:" R | . 0.10 S g . o e RN L0.25 | / ] T f' - :: 1.1 o 0.054 / ; { T - -0.20 ( 1N 0- . i RUN: 4-14 15-20 + - | ] E | 3 CORROSION | | | L] 1 i fiu | O T ol | GEEE | (mils) B ] RUNNO. |4 |5 7 8|9i 10 1 12| 113 14 5 |16 T 18 19 20 FLUSH ® [ ] 1 LI II‘ '‘BEIEREL 1 l 1: [ i = l ‘ ' DJFM‘AMJJASOND{JFMAMJJASONDJFMAMJJASGNDJFMAMJJASOND 1965 1966 1967 1968 1969 Fig. 6.3. Corrosion of the MSRE fuel circuit in 2*>U and 233U power operation. .“ some specific event. A significant increase in the concentration of chromium in fuel salt samples was noted only during the initial stages of runs Nos. 4, 8, 12, 15, and 19, with an apparent increase partway through run No. 14. In the beginning periods of run Nos. 4, 8, 12, 15, and 19, we find that in each instance a common set of circumstances existed: the reactor core vessel had been opened previously, either for maintenance or to exchange test arrays positioned within the graphite moderator. Although reasonable measures were adopted to minimize the amount of airborne contaminants that might be introduced into the system during these periods, the possibility that significant amounts of oxidants were introduced into the open vessel cannot be excluded. During the first period of power operation, when a significant tempera- ture differential was imposed on the circuit for the first time (run 4), some corrosion was anticipated as the Cr° + 2UF, = CrF, + 2UF; equilibrium reaction adjusted to the temperature profile of the circuit. Under these conditions the increase in the concentration of chro- mium in the fuel salt could have resulted from the establishment of the equilibrium reaction and was not necessarily a signal of the presence of oxidizing con- taminants. The increase in the concentration of chromium in the fuel salt well after run 14 began (see Fig. 6.3) seems to be inconsistent with the premise that external con- taminants were the principal cause of corrosion. How- ever, in the period preceding run 14 part of the graphite and metal specimens in the core were removed and replaced. It seems quite possible, therefore, that the residual concentration of reductant which was gen- erated within the fuel salt during run 12 was sufficient to offset the combined oxidizing effects of whatever contamination was incurred during shutdown and that characteristic of the fission reaction® only through the early part of run 14 and that the subsequent rise in chromium concentration represents the normal com- pensating shift in the equilibrium corrosion reaction. Generally, it is assumed that the contaminant most likely to be responsible for corrosion is moisture. The inference that moist air was the corrosion- inducing contaminant calls into question the efficacy of the flush salt. As an agent for removal of adsorbed moisture, molten LiF-BeF, flush salt is extremely effective, as demonstrated in numerous laboratory experiments, and unquestionably was an effective mois- ture scavenger. If an oxidizing contaminant or contam- inants were capable of diffusion within the graphite or reacting with species deposited in the surface layers of the graphite, the probability of its removal by brief 75 circulation of flush salt might be slight; instead it might be released into the salt gradually after the moderator was heated to high temperatures. Thus, oxygen (per- haps as CO), rather than water, seems more likely to be the cause of the observed corrosion. This conclusion is supported by the fact that the scale found on the nickel cages which were used to expose Be® to the salt during run 15 was comprised preponderantly of iron, whereas on other occasions the principal structural metal in such scales was chromium. Results of chemical analyses showed that the prior fuel reprocessing treatment was effective in reducing the concentration of chromium in the salt from 133 to 34 ppm and of iron from 174 to 110 ppm. The effectiveness of the reprocessing opera- tions in reducing Cr®>" preciudes the likelihood that significant amounts of Fe?" were delivered to the fuel circuit from the chemical reprocessing plant. The reduction of Fe? to Fe® by Be? suggests rather that Fe?* was generated after the beginning of fuel circula- tion by an oxidizing contaminant contained in the closed fuel circuit. As increasing amounts of Be® were added to the salt mixture, the ratio of metallic Fe/Cr found on the nickel metal cages was reduced until a normal balance was reestablished and corrosion ceased (see Tables 6.3 and 6.4). Laboratory studies of the stability of FeF, have led to the conclusion that at equilibrium little or no Fe?" should exist in the MSRE fuel (see Sect. 2.4.7). The results of these tests confirmed that, of the structural metal impurities in the melt, iron and nickel persist almost entirely as metallic species. The results of other laboratory studies, particularly those concerned with the reduction of Fe?" by hydrogen,® suggest that even in concentrations as large as was found by Manning, the presence of divalent iron was likely to have resulted from reoxidation after transfer and remeiting of the salt samples. Subsequent laboratory tests with the MSRE fuel were precluded by the generation of fission products in the salt. Appraisal of the premise that maintenance operations might possibly permit the ingress of enough oxygen to account for the analytical results requires the following considerations. The cumulative amount of oxidation introduced into the fuel salt (during both 233U and 2330 operations) based on the increases of chromium at the beginning of run Nos. 8, 12, 15, and 19 and on the apparent losses of UF; at the end of runs 7, 18, and 19 amounts to 51.31 equivalents, or 410.5 g of 0*", and corresponds to a cumulative exposure of ~50 ft° of air. During 233U operations, the amount of oxygen enter- ing the salt might have been expected to increase the concentration of oxide by 83 ppm, in excess of the 76 Table 6.3. MSRE fuel salt analyses, run No, 15 Be® added Weight percent % M/z Fe + Cr+ Ni Sample (g) Be Te Cr N1 Fe Cr Ni Carrier, before Z:® addition 0 0380 0 0460 0 0180 372 452 176 Carner, after Zr® addition 00110 0 0034 <0 0010 714 220 65 FP-15-7 10 08 Scale from Ni cage 120 0106 8 03 597 0353 398 FP-15-30 8 34 Scale from Be® rod Z mg 31D (42 8) (11 2) 365 503 132 Scale from N cage zmg @61 493) (21 2) 786 40 174 FP-15-62 9 38 Loose particles 6 06 383 375 144 425 416 159 Scale from top 629 472 326 471 372 257 371 Scale from cage 7 34 274 453 0 096 372 615 13 Salt average, 11/23/68 659 00140 0 0067 0 005 55 26 19 Table 6.4. Relative fractions of Fe? and Cr® reduced from MSRE fuel salt in run No. 15 Sample Equivalents of Corrosion e’ /Cr0 on nickel No Be? added rate (milsf/year) cage? FP-15-7 224 088 113 FP-15-30 4 09 054 195 FP-15-62 617 035 061 2Average Fe/Cr in carrier salt was 2 24 at the incepiion of run No 15 sensitivity hmuts for the analytical method and well above that observed It must be recalled, however, that early 1n *3°U power operations the concentration of 07", as measured expenimentally, dechned from 120 to 60 ppm, suggesting that under power operation 0® 1s partially removed from the salt as a volatide species [t may be concluded, therefore, that the corrosion ob- served in the MSRE 1s likely to have been caused as described above but that the mechanism has, as yet, not been demonstrated unequivocally The rationale proposed above has several implications concerning the behavior of the MSRE dunng 233U operations During the first 16 hr in which fuel salt was circulated at the beginming of run No 15, the salt did not transfer to the overflow tank and behaved as though it contained a neghgibly small bubble fraction Thereafter, Be® was introduced, and the bubble frac- tion began to increase, with further exposures of the salt to Be® the fraction varied erratically ' Certanly the Be® reduced the surface tension and thereby allowed easier transport of gas from graphite to salt, followed probably by oxidation of the metallic iron impurity, which acts as an oxidant to the circuit walls Corrosion would continue until the oxidants were consumed The model of corrosion proposed here has a relation to the changes in bubble fraction The corro- sion data suggest that with respect to its physical and chemical properties, the fuel did not achieve a reference state until the beginning of run 17 That 1ts bubble fraction then was greater than observed m 235U operafions probably was related principally to 1ts lower density The probability that atmospheric oxygen was the primary causative agent of corrosion mn the MSRE was discounted by Grimes in the following appraisal of the observed behavior 1! We find no evidence whatsoever that air was introduced mto MSRE dunng tts periods of normal high temperature operation There 1s, on the other hand, no doubt that air i appreciable guantities was admutted to this reactor system during shut downs when the reactor circuit was at temperatures of 300°F or below Only a small fraction of the oxygen admitted at such low temperatures should have reacted with the reactor metal and should have been available to cause subsequent corrosion Flushing of the reactor circuit with hehum during the reactor heatup should have removed most of the unreacted air (oxygen) Moreover, use of the flush salt before admission of the fuel should have removed some (and perhaps a large fraction) of the reacted oxygen However, some oxidant was admutted, and some corrosion from this source probably occurred in MSRE The evidence suggests that corroston from admitted ar must have been a small fraction of the minor amount which took place Since a detailed deserniption of the corrosion picture in MSRE at all ttmes 1 not yet available, 1t 1s not possible to define the precise amount of corrosion due to ingress of oxygen It 1s easily possible, however, to set upper himits upon the amounts of oxygen that could have been mvolved These amounts are small If all the observed corrosion (as evidenced by rise in CrF, concentratton) i MSRE Runs 1 thru 9 s aitributed to imngress of oxygen the oxide 1on concentration of the fuel should have nsen by about 11 ppm If corroston (by the same indicator) in Runs 10 thru 14 were due to oxygen ingress, and if this oxygen mgress were also responsible for oxadizing all the UF;3 created by the Be® added dunng those runs, the oxide ton concentra tron of the fuel should have risen by about 25 ppm Durnng the entire sequence of runs with 235UF4, therefore, all the chemically observed corrosion could be due (it seems very certamn that it was not) to the mgress of oxygen resulting in less than 40 ppm of added oxide 1on mn the fuel Our systematic chemical analyses for 0% 1n the MSRE fuel showed no evidence of such an mcrease MSRE was provided with no mechanism intended to remove 02_, and we have been unable to postulate much less to demonsirate — an imadvertent mechanism for 0% removal Our methods for determmnation of QF might possibly fail to find this <40 ppm increase It seems much more plausible, however, that the actual increase in 0% was a small fraction of this figure and that the observed corroston was due to oxadation of chromium by UF, through mechanisms similar to those postulated from many years of testing It 1s true that we have drawn detailed curves of the chromium behavior ot the mdividual MSRE runs with 233 UF,, and we have been tempted to speculate about them However the marked scatter in the chromium data gives such curves httle or no statistical signifi cance Rise i the chromium concentrations in MSRE Runs 15 thru 19 (that s in the runs fueled with 233UF,) was more dramatic and the mitial mcreases were almost certainly more statistrcally signtficant Several pieces of evidence suggest that the 233UF4 fuel, which by virtue of reuse of the carrier salt was necessarily less well characterized and probably less pure than was the 235UF,4 wnrtial fuel charge, was relatively oxidizing This may have been due fo oxygen mgress and oxidation of the MSRE metal surfaces duning the long fuel change shutdown between Run 14 and Run 15 If all the corrosion during Runs 15 thiu termunal Run 19, again as adduced from increases in chromium concentration of the fuel and allowmng for oxidation of all reductant added, s attributed to admitted oxygen, the oxude ion concentration of the fuel should have mcreased by some 45 ppm, this increase, again, seems conceivable but unhkely In summary, the corrosion picture shows that the quanfity of oxide 1on which could have entered the fuel during the whole MSRE operation (involving 20 separate shutdowns) was less than 100 ppm It seems virtually certamn that the actual amount which entered was much less than this The ZrF, present i this fuel would have prevented precipitation of appreciable UQ, even if much larger quantities had entered, oxide 1on at the maximum levels suggesied above would not have precipitated Z10, Subsequent corrosion of the rteactor circuit through mgress of oxygen would, however, not have been prevented (or appreciably affectied) by the presence of ZrF, m the MSRE, this would be equally true for ZiF,4 1n the MSBR fuel It has been postulated that one of the principal reasons for the expectedly low corrosion observed is that the metal surfaces of the fuel circuit have been covered with a film of the noble-metal fission products Nb, Mo, Tc¢, and Ru about 10 A thick Results of electron microprobe analysis of the metal surveillance spectmens'? removed from the MSRE m May 1967 lend support to this view in that they did not reveal any 77 change in chromium concentration below a depth of 10 M, the lhimut of measurement, as do the results of electron mucroscopic nvestigation of the specimens,’ ? which 1ndicated that the surface was covered to a depth of several thousand angstroms by those metals Al- though the postulate of a dynamucaily produced and autoregenerative noble metal film 1s an attractive rationale of the very low corrosion observed in 225U operations, 1t fails when the behavior of the MSRE during the early stages of run 19 1s considered Here, rapid corrosion was proceeding concurrently with the generation and presumably the deposition of the same noble metal fission products which seemed earlier to have restricted the corrosion of the fuel circuit walls Another hypothesis has been suggested recently by Grimes' > as an alternate rationale of data that indicate the almost complete absence of corrosion dunng extended periods of operation, while at other times rates were more rapid than could be accounted for by diffusion-controlled mechanmisms It presupposes that the presence of UF; in the fuel salt does not counter the corrosion of Hastelloy N effectively once the concentration of Cr®" in the salt has risen to a level of ~70 to 80 ppm Rather, at concentrations of this order and n the presence of the intense radioactive flux m the reactor core, chromium carbide, Cr3C,, 1s formed at the surfaces of the graphite moderator The forma- tion of such a phase thus causes the graphite moderator to act as a sink for chromium and promotes corrosion of the container alloy at the most rapid rate allowed by the diffusion coefficient of chrommum If such a mechanism operates, the difference in the chromium concentration as observed in the fuel salt samples and the maximum possible concentration represents the quantity of chromium which was deposited at the surface of the moderator graphite On occasions when the oxidation potential of the salt increased substan- tially, this carbide coating would then decompose and release large amounts of Cr?" to the salt, thus causing rapid changes in the concentration of chromium in the salt samples submutted for chemical analysis to suggest that the corrosion rate was mncreasing rapidly This hypothesis 1s attractive for a variety of reasons (1)1t implies that throughout the operation of the MSRE the corrosion rate was invariant and diffusion limited, (2) the [UF3]/]UF,] remains so low 1n such systems at all times that attempts to repress corrosion in MSBR fuels by external adjustment of the [UF;]}/[UF4] are needless, as well as the previously foreseen need to develop in-line analytical methods for 1ts determination, and (3) the hypothesis provides a rationale for the otherwise mnexplicably low rates of corrosion which seemed to characterize operations of the MSRE with 2350 fuel. It seems unlikely that significant quantities of chro- mium carbides will have formed on the moderator graphite surfaces for the following reasons: 1. In run No. 15, the principal constituent of the slag found on the nickel cages used for the introduction of beryllium was iron (see Sect. 6.7). Presumably, this iron deposit was formed by reduction of Fe?* from the melt and represents a condition in the melt which is so oxidizing that, if present, the relatively unstable compound Cr3C: (AG, 5p0°k = —23 keal) should have decomposed completely during this period. The possible amount of chromium which should have been released to the fuel salt by such decomposition, as calculated using diffusion coeffi- cients from 1 X 107'% to 4 X 107** cm?/sec, would increase the concentration of chromium in the fuel salt to values of 240 to 480 ppm (see Table 6.2), that is, to much higher concentrations than were observed at any time during MSRE operations. 2. The coincidence of apparently stepwise increases in Cr followed maintenance periods (as noted previ- ously). 3. The carbides of niobium are more stable than is Cr3C,; in 223U operations the oxidation potential of the salt increased gradually during runs by fission to the point that niobium entered the salt (see Sect. 6.5) and was subsequenily reduced and removed from the salt by Be addition; under these conditions, Cr3C, should have become reoxidized, entered the salt, and increased the chromium concentration of the samples sharply, where, in fact, no perturbation of the average concentration of chromium was noted. In postoperational examinations with one of the stringer bars from the reactor core, a search was made to ascertain whether or not a coating of chromium carbides adhered to the surface. No evidence of such a coating was found. A surprising result of the postoperational tests per- formed with specimens of the MSRE control rod thimbles was the observation of anomalous intergran- ular penetration of the alloy which had been exposed to the fuel salt. Occasionally, inspection of Hastelloy N specimens removed from the reactor core showed grain boundary cracks at the surfaces of specimens subjected to tensile tests. These cracks appeared in material strained at room temperature, which is not normal for Hastelloy N, but they penetrated only a few mils and 78 had no detectable effect of the mechanical properties of the specimens. The cracks in the control rod thimbles, however, were deeper than previously observed, at some locations as deep as 7 mils,’* some ten times the depth of the generalized attack indicated by changes in the concentration of chromium in the fuel salt. Curiously, several fission products, preponderantly tellurium, had penetrated to depths comparable with those of the cracks. It is not possible to ascribe this attack to radiation effects per se, for in-pile tests with salt-graphite-alloy assemblages prior to tests with the MSRE showed no corresponding effects. These unpredicted results have caused a renewal of the investigation of materials removed from the MSRE, with the initial objectives to determine if fission products are indeed the cause of grain boundary separation. While the initial results seem to implicate tellurium as a causative agent, its role is not positively identified. One mechanism, proposed before tellurium was be- lieved to have deposited preferentially, relates the attack to impurities in the following way. Electron microscopic examinations of Hastelloy N show typi- cally that an unidentified group of complex phases comprised of Mo, Cr, C, N, and B, morphologically similar to face-centered M,C carbides, are deposited preferentially at intergranular surfaces. It is also ob- served that the loss of ductility that occurs with Hastelloy N after irradiation correlates with more pronounced intergranular fracture. Such behavior very possibly means that the chemical activity of the carbide-like phases deposited in these locations is increased. If so, it would not be surprising to find that these phases were highly susceptible to oxidative corrosion. It was noted previously that two unusually oxidative regimes occurred in the MSRE, at the beginning of run 15 and again at the beginning of run 19. In an earlier assessment of corrosion of the MSRE fuel circuit we have deduced that the most probable oxidant at those periods was atmospheric oxygen.'® We reach the tentative conclusion, therefore, that the intergranular attack noted was a result of attack by oxygen on the carbide-like phases deposited at inter- granular surfaces. If oxygen were released from the moderator, as suggested previously, the first metal which it might contact would be the control-rod thimbles. It seems likely, therefore, that the inter- granular penetration observed in the postoperational tests resulted from the development of unusually oxidative conditions in the MSRE and is atypical of normal reactor behavior. The overall results of corro- sion surveillance in the MSRE, while not totally unequivocal, appear to indicate that corrosion behavior in molten-salt reactors can properly be anticipated to be negligible over the planned use period of these reactors. Furthermore, the probability of recurrence of the anomalous penetration in future reactors can be assessed in tests with the modified alloys which are currently under development and which will contain a group of carbides perhaps unlike those found in the MSRE alloy. 6.4 Additions of Reductants and Oxidants to the Fuel Salt As discussed earlier in this report (Sect. 6.3) the fuel salt, free of moisture and HF, should remove chromium from Hastelloy N only by the equilibrium reaction ,C1° + UF4(d) = %, CrF,(d) + UF5(d) . When the above corrosion equilibrium was first estab- lished in MSRE power operations, the UF; produced in this reaction, together with that originally added to the fuel concentrate, should have totaled 1500 g, with the result that as much as 0.65% of the uranium of the system could have been trivalent soon after the begin- ning of power operation. The UF; content of the MSRE fuel was determined after approximately 11,000 MWhr of operation to be no greater than 0.05%. The fuel salt was considered to be far more oxidizing than was necessary and certain to become more so as additional power was produced unless adjustment was made in the UF; concentration. A program was initiated early in 1967 to reduce 1 to 1.5% of the uranium inventory to the trivalent state. On the basis that it met chemical criteria and that its introduction into fuel salt could be accomplished conveniently, beryllium metal was selected as the most suitable substance to use for in-situ reduction of U(IV) in the fuel salt. Results of laboratory tests had shown that the metal would be sufficiently reductive to require exposure for conveniently short periods of time but was not so reductive as to cause concern that it might reduce U(IV) to the metallic state. The pure material was available in forms that were easily accom- modated for use with the sampler-enricher device, and the metal afforded a maximum of reductive equivalents per unit mass. Initially, 4 g of beryllium was introduced into the salt by melting a mixture of "LiF-BeF, carrier salt and powdered beryllium in the MSRE pump bowl sampler cage. Subsequently, three additions were made by suspending specimens of ¥-in.-diam beryllium rods in 79 the salt in the pump bowl. The capsules used for adding beryllium were similar in size to those used for sampling for oxide analysis but were penetrated with numerous holes to permit reasonable flow of fuel salt. The beryllium rods were found to react with fuel salt at a steady rate, dissolving at approximately 1.5 g/hr. A summary of all the additions of reductants and oxidants introduced into the fuel salt system via the sampler- enricher apparatus is given in Table 6.5. With 235U fuel salt, all dissolutions of beryllium metal into the fuel salt proceeded smoothly; the bar stock which was withdrawn after exposure to the fuel was observed to be smooth and of symmetrically reduced shape. No significant effects on reactivity were observed during or following the beryllium additions, nor did the results of chemical analyses indicate that the actual concentration of U(HIT) had increased until after run 12 had begun (see Table 6.6). The additions preceding run 12 had increased the U3* concentration in the total uranium by approximately 0.6%. Four exposures of beryllium were made at close intervals during the early part of that run. Samples taken shortly after the last of these four exposures (FP-12-16 et seq., Table 3.1) began to show an unprecedented increase in the concentration of chromium in the specimens, followed by a similar decrease during the subsequent sampling period. Previous laboratory experience has not disclosed comparable behavior, and no well-defined mechanism was available at the time to account satisfactorily for the observed behavior. Speculation as to the cause, partially supported by experimental data, included the following consideration. On the two occasions when the most rapid rates of dissolution of the beryllium rods were observed, chro- mium values for the next several fuel samples, FP-11-10 et seq. and FP-12-16 et seq., rose temporarily above the lo level and subsequently returned to normal. That the increase in chromium levels in samples FP-12-16 to -19 was temporary indicates that the high chromium concentration of fuel samples removed from the pump bowl was atypical of the salt in the fuel circuit and implies that surface-active solids were in suspension at the salt-gas interfaces in the pump bowl. That atypical distribution of species in this location does indeed take place was demonstrated earlier by the analysis of sample capsule support wires that were (1) submerged below the pump-bowl salt surface, (2) exposed to the salt-gas interface, and (3) exposed to the pump-bowl cover gas. The results showed that the noble-metal fission products, Mo, Nb, and Ru, were 80 Table 6.5. Summary of adjustments of [U°']/[=U] in the MSRE fuel salt g . Reductant or oxidant Equivalents of [U3+/2U] Date ample added reductant (%), Number . Form Weight (g) added nominal 2/13/66 Runs 5-1 041 1/1/67 FP-10-14 Be powder 3.0 0.67 0.22 1/3/67 FP-10-16 Be powder 1.0 0.89 0.25 1/4/67 FP-10-18 Be rod 1.63 1.25 0.29 1/13/67 FP-10-23 Be rod 10.65 3.61 0.51 2/15/67 FP-11-10 Be rod 11.66 6.20 0.74 4/10/67 FP-11-40 Be rod 8.40 8.06 0.79 6/21/67 FP-12-8 Be rod 7.93 9.82 0.87 6/23/67 FP-12-9 Be rod 9.84 12.01 1.12 713/67 FP-12-13 Be rod 8.33 13.86 1.30 7/6/67 FP-12-15 Be rod 11.68 16.45 1.59 8/3/67 FP-12-56 Be rod 9.71 18.60 1.72 9/15/68 FP-15-7 Be rod 10.08 2.24 0 10/13/68 FP-15-30 Be rod 8.34 4.09 0 11/15/68 FP-15-62 Be red 9.38 6.17 0 11/20/68 FP-15-66 Be washer 1.00 6.39 0 1/22/16 FP-17-8 Be rod 8.57 8.29 1.19 1/30/69 FP-17-11 Crrod 4.13 8.48 1.19 4/15/69 FP-18-3 Zr rod 20.24 9.37 0.65 4/26/69 FP-18-7 Zr rod 24.04 1042 1.19 5/8/69 FP-18-17 FeF, powder 30.00 9.78 0.52 5/15/69 FP-18-23 Be rod 5.68 11.04 1.20 5/20/69 FP-18-28 Be rod 3.17 11.74 1.59 9/12/69 FP-19-25,26 Zr foil 0.62 11.77 0 9/24/69 FP-19-31-4 Zr foil 1.20 11.82 0 10/2/69 FP-19-40 Be rod 2.87 12.46 0.99 10/8/69 FP-19-48 Be rod 4.91 13.55 1.61 10/21/69 FP-19-51 Nb foil 0.018 13.55 1.37 11/29/69 FP-20-7 Be rod 6.974 15.10 1.16 12/9/69 FP-20-22 Be rod 9.894 17.30 2.49 12/9/69 FP-20-22 Be rod 3.019 17.97 2.95 deposited in abnormally high concentrations at the salt-gas interface. Such behavior suggests that the high chromium concentrations in the fuel specimens were caused by the occurrence of chromium in the pump bowl in nonwetted, surface-active phases in which its activity was low. A possible mechanism which would cause such a phenomenon is the reduction of Cr** by Be® with the concurrent reaction of Cr® with graphite present on the salt surface to form one or more of the chromium carbides, for example, CrzC, (AH Of =21 kecal at 298°K). Such phases possess relatively low stability and could be expected to decompose, once dispersed in the fuel-circuit salt. The possibility that surface-active solids were formed as a consequence of the Be® additions was tested late in run 12 by obtaining salt specimens at the salt-gas interface as well as below the surface. First, specimens were obtained in a three-compartment sample capsule that was immersed so that the center hole was expected to be at the interface (Fig. 6.4). Next, a beryllium metal rod was exposed to the fuel salt for 8 hr with the result that 9.71 g of beryllium metal was introduced into the fuel salt. Twelve hours later a second three-compart- ment capsule was immersed in the pump bowl. Its appearance after removal from the pump bowl is shown in Fig. 6.5. Chemical analyses of the fuel-salt specimens FP-12-55 and -57 did not show significant differences in chromium; however, the salt-gas interface in FP-12-57 was blackened as compared with FP-12-55 (Fig. 6.6). An additional purpose of sampling with the three- compartment capsule was to determine whether foam- like material was present in the sampler area and would be collected in the upper compartment. Globules were noted on the upper part of FP-12-57 (Fig. 6.5), indicating that conditions in the pump bowl were substantially different after beryllium was added to the fuel salt. .‘ Table 6.6. Concentration of UF; in the MSRE fuel salt? Uranium Uranium - o Net Equivalents . 3+ Date SaI:Inp te Megawatt-Hours Consumed Consumed Nethqxf(xlvaI.cnts ggctlaldBe of Reductant Ne; gqgwatlaentts v @ - o (ke {moles) of Oxidation ed (g Added Ob Keauctan Calculated Analytical 0 0 3.13 0.33 11/14/66 FP9-14 12,345 0.632 2.67 2.14 0 3.13 0.99 0.10 0.10 1/1/67 FP10-14 14,950 0.766 3.23 2.58 3 3.80 1.22 0.13 1/3/67 IP10-16 15,050 0.771 3.25 2.60 6 4.46 1.80 0.19 1/4/67 [P10-18 17,100 0.877 3.70 2.96 7.63 4.82 1.86 0.20 1/13/67 FP10-23 17,852 0.915 3.86 3.08 18.28 7.19 4,11 0.43 1/15/67 FP10-25 18,050 0.924 3.90 3.10 18.28 7.19 4.09 0.43 0.66 2/6/67 FP11-5 19,712 1.010 4.26 3.40 18.28 719 3.79 0.39 0.60 2/15/67 FP11-10 21,272 1.090 4.60 3.68 29.94 9.77 6.09 0.64 2/22/67 P11-13 22,649 1.161 4.90 3.90 29.94 9.71 5.87 0.62 0.69 3/28/67 FP11-32 28,342 1.453 6.13 4.90 29.94 9.77 4.87 0.51 0.45 4/10/67 FP11-40 30,900 1.584 6.68 5.34 38.34 11.64 6.30 0.66 6/21/67 FP12-6 36,055 1.663 7.01 561 38.34 11.64 6.03 0.64 0.71 6/21/67 P12-8 36,055 1.663 7.01 561 46.27 13.40 7.79 0.82 6/23/67 FP129 36,416 1.866 7.87 6.30 56.11 15.58 9.28 0.98 6/29/67 FP12-11 37,400 1.932 8.15 6.52 56.11 15.58 9.06 0.95 1.30 7/3/67 FP12-13 37,856 1.940 8.19 6.55 64.24 17.38 10.83 1.14 7/13/67 FP12-15 38,345 1.966 8.30 6.64 76.12 20.02 13.38 1.40 7/13/67 FP12-21 39,500 2.023 8.54 6.83 76.12 20.02 13.19 1.39 1.0 8/3/67 I'P12-56 43,872 2.248 9.49 7.59 85.83 22.18 14.59 1.54 9/15/67 I'P13-5 44,781 2.314 - 9.76 7.81 85.83 22.18 12.42 1.31 1.60 3/26/68 FP14-(I) 72,454 3.743 15.79 12.63 85.83 22.18 9.55 1.01 9/15/63 FP15-7 0 0 10.08 2.23 0 0 10/13/68 ['P15-30 0 0 18.42 4.09 0 0 11/15/68 [P15-62 0 0 27.80 6.17 0 0 11/20/68 FP15-66 0 0 28.80 6.39 0.22 0.15 1/22/69 FP17-8 850 0.044 0.19 0.1 37.37 8.29 1.97 1.31 ?These numbers assume that the 235U salt originally was 0.16% rcduced; that the increase in Cr before initial 235y power operations was real, occurred before 11/14/66, and resulted in reduction of U to U3+; that each fission results in oxidation of 0.8 atom of U3+; and that there have been no other losses of U3+. 18 82 % of ey PHOTO 1865-71 PHOTO 1864-71 Wi R O N Pt St g o el L U o bre [ W 2 © - SEE Pl Bje @ P RCCTR S SN 4 E i s w; o 18 S & e S T w & s lete k& m‘ya ©rmren e fvfifiwMM g Sk, b . I A N SR T B P g il L A ek L, e m Ty @yw?;‘%‘ S o R o R ) i s NS & A A A et 37 QIRS NG 5, Y 5% o5 - ., Fol7EE s 2 = o o g3 = oy - o ns Q Bt w 3 o @ @ & ER: a, [ Q — -+ 2 . 1c Fig. 6.5. Three-compartmeni n MSRE pump bowl after exposure of beryl salt. August 3, 1967, FP- Each of the three compartments 12-57. the capsules shown in F outside by two holes as shown here. 5 was open to the The capsules were 4 and 6. 6 igs. n imn kel capsule suspended ) {4 -compartiment n Fig. 6.4. Three MSRE pump bowl before exposure of berylt positioned approximately so that the center compartment was expected to be at the salt-gas in to 235U fuel um active terface to collect surface- 12-55. salt. August 2, 1967, FP foam 1f 1t were present. No appreciable residue was found 1n the solids. The upper compartment was to serve as a collector of upper compartment. %@@‘&%@m@%flb P &t ST, R-39267 8767 = = - iog A S et Bl oo O o (& i g%fmfisg}&w o e Fet e 5 8 Sk e fimé;;ar’w - R-39266 (£) Fig. 6.6. Surface appearance of fuel salt before and after beryllium exposure to 2350 fuel sait. (@) FP-12-55, (b) FP-12-57. Salt analyses did not reflect significant differences in chromium concentration nor were they indicative of the identity of the material at the blackened surface. 83 Examination of the metal basket that contained the beryllium rod while it was exposed to the fuel showed the presence of dendritic crystals along with a small amount of salt residue (Fig. 6.7). Spectrochemical analysis of material removed from the basket (Fig. 6.7) indicated that the material contained 7.8 wt % chro- mium and less than 10 ppm of iron and nickel. The evidence obtained did not permit inference as to the identity of the phases which formed within the pump bowl as a consequence of the beryllium addi- tions. It strongly implied that nonwetted flotsam can be formed and accumulated temporarily in the MSRE pump bowl. During this period, fuel salt accumulated in the overflow tank steadily during operation and remained there in relative isolation from the fuel stream. At intervals of about one day, part of the salt (60 Ib) was returned to the fuel stream. Recognizing that chromium might be injected into the pump bowl as the salt returned, we performed an experiment in which salt samples were obtained from the pump bowl within an hour after fuel was returned from the overflow tank to the pump bowl. The purpose of the experiment was to determine whether material from the overflow tank contributed appreciably to the perturbations in the chromium concentration. The results were negative, possibly, in part, because sampling and salt-transfer operations were not performed concurrently for safety reasons. In view of the mechanism developed recently'® to account for the transport behavior of the noble metal fission products in the MSRE, we can infer that the exposures of beryllium which generated high concen- trations of chromium in the salt samples simply reduced the Cr*" in the flowing fuel salt to the metal, which in turn became a part of the particulate pool of suspended metals at the surface of the salt in the pump bowl. It remained there temporarily, gradually oxidizing to the fluoride as it reacted with the fuel salt. In their examinations of the behavior of the noble metal fission products, Kirslis and Blankenship! 7 did not observe a significant effect of fuel reduction on the noble-metal concentrations in the fuel. The concentra- tions frequently rose rather than fell, as expected, after adding beryllium, They noted that the °°Mo, '°3Ru, 106 Ru, and '32Te showed parallel rises and falls. The ®°Mo results were extraordinarily high for samples FP-11-8 and FP-11-12. The high values were checked by reruns on fresh samples. If all the ?°Mo produced by fission remained uniformly distributed in the fuel, the calculated concentration would be 1.4 X 10'! dis 84 Fig. 6.7. Residue from berylium addition capsule FP-12-56, August 3, 1967. Capsule was exposed to fuel salt for 8 hr, 9 71 g of Be® dissolved Residue contaned 7 8 wt % Cr and less than a total of 10 ppm Fe and Na . min~! g7'. A calculation showed that if all the *®Mo produced by neutron activation of the *®Mo in the first 0.1-mm thickness of the Hastelloy N reactor contain- ment vessel diffused instantaneously into the fuel melt, the increase in ?°Mo concentration would be only about 10° dis/min per gram of fuel. It thus appeared that a mechanism was operating which either concen- trated fission-produced Mo and other noble metals in the pump bowl or resulted in large temporal and spatial variations in their concentrations. Dissolved ?°Mo would undoubtedly be uniformly distributed. If the noble metals circulated as a suspension of insoluble metal particles, it was reasoned that they might concentrate in the pump bowl or vary in concentration with pump bowl level, cover gas pressure, and other operating variables. Since clean metal surfaces are not wetted by the fuel salt, there might also be a tendency for metal particles to collect around helium bubbles, which are probably most numerous in the pump bowl. The °°Nb concentrations varied erratically and did not parallel the behavior of the other noble metals. This was ascribed to analytical difficulties. An unavoidable difficulty was that a large correction for ?3Zr decay must be made for each salt analysis. From the beginning of operations of the MSRE with 2337 fuel, the behavior of the fuel salt during and after exposure to the beryllium rods was markedly different than previously observed. A detailed description of the changes in fluid salt and gas behavior during these periods is described in a separate report.’ ® Introduction of beryllium metal into the 232U fuel salt apparently caused a major perturbation of salt-gas interactions and resulted in an increase in void fraction of the fuel salt. The exact reasons for development of the greater void fraction that persisted throughout ?22U operations have not been completely determined. In addition to the purely chemical effects which resulted from the addition of beryllium and other reductants, other factors such as pump speed, variation in the solubility of inert gases as a function of hydrostatic pressure, difference in density of the 233U and #33U fuel salt, and changes in interfacial tension were recognized to be related. Interaction of beryllium with the fuel salt during run 15 caused a major change in only one physical property of the fuel — its surface tension. This is the most puzzling aspect of the effect of beryllium on the fuel salt in run 15 since the effect was unprece- dented in the fuel system. However, slight changes in interfacial energies in fluid systems are known to account for marked changes in hydraulic behavior. Since the addition of beryllium in the early part of the 233U operations effected both chemical and physical 85 changes concurrently, the exact way in which these changes altered the void fraction of the salt cannot, therefore, be evaluated exactly, although some aspects of the observed phenomenon can be attributed tenta- tively to the additions. If, as inferred in Sect. 6.3, atmospheric oxygen was the causative agent for the increased oxidation potential of the salt, release of the gas from the moderator graphite might well have been triggered by a reduction in the salt-gas interfacial tension, resulting in an increase in the bubble fraction of the salt. Photographs of the cage assemblies removed from the fuel pump bowl after exposure to the fuel salt are shown in Figs. 6.8—6.23. While none of these cages appeared to have been wetted by salt during 235U operations, each of those removed after treatment of the 233U salt showed evidence that salt had wetted the nickel cage and had adhered to it. A typical example is shown in Fig. 6.17. The presence of bubbles is suggested by the appearance of the upper part of capsule FP-17-8 (Fig. 6.17), on which structures of collapsed bubbles seem to be visible. Of particular significance to the interpretation of the behavior of the fuel salt from August 1968 to January 1969 are the changes in the relative amounts of iron and chromium contained in the deposits on the nickel cages that were used to suspend beryllium into the pump bowl (Table 6.5). From this and other chemical evidence we deduced that corrosion in the fuel circuit was not checked until early in 1969. By that time, the bubble fraction in excess of that observed in 23%U operations was reduced to a fraction that could in general be accounted for by the other factors men- tioned above. It may be speculated that the coincidence of the disappearance of reactivity blips with reestab- lishment of the expected ratio of structural metal fractions in the residues adhering to the nickel cages was indicative of the removal of the last fraction of anomalous excess of bubbles in the fuel salt. Among the conjectures advanced to account for the excessive rates of salt transfer during this period was one based on the recognition that sharp gradients in the density profile of the fuel salt in the pump bowl existed; from this it was suggested that voluminous amounts of foam were developed in the pump bowl. There was no evidence that such foam had existed in 233U operations, but rather that a salt mist was present in the pump bowl. Indication that such a mist existed was obtained by suspension of a cage-rod assembly of nickel into the pump bowl during run 14. Photographs of the assembly after removal from the pump bowl (Fig. 6.10) show that the surfaces of the rod exposed to 86 . 235 -Fig. 6.8. Dendritic crystals and salt on basket of nickel capsule used to expose beryllium to U fuel salt. August 3, 1967, FP-12-56.. ' gt z‘“&? % ):\?“(m B e = iy 2 M - i b & - SRRy a Y k3 s u’w;@f o ——— v ¢ o g # L A g F ri - ke - ] Fig. 6.9. Appearance of nickei basket and rod after suspen- sion in MSRE fuel salt for 2 hy. FP-14-55. the vapor just above the salt were coated with con- densed droplets. Although the vapor space in the sample shroud was probably more quiescent than that in the pump, it was reasoned that foam should also appear here if it were prevalent in the pump bowl. 87 Attempts to produce foam were performed in the laboratory'® but were unsuccessful. Tests for foam production consisted of measurements of the fuel salt level in a 10-in.-diam nickel vessel at intervals during the reduction of essentially all of the uranium to its trivalent state and after stepwise oxidation of the nickel with nickel fluoride. Melt levels were determined by the abrupt change in electrical resistance noted as the insulated nickel rods made contact with the salt mixture., Without exception, the tests showed no evidence of the developmeni of a foam either by increasing or decreasing the oxidation potential of the salt. In an attempt to identify other possible agents which might cause development of foams in fluoride melts, Kohn and Blankenship'?® tested the effect of each of the additives: carbon dust, graphite powder, beryllium metal, finely divided nickel metal, and pump oil vapor. They were unable to promote the development of stable foams with any of these additives in clean melts. Foams were formed only by introducing enough water into the melt, either in the sweep gas or by adding solid hydrates, to give a definite cloudiness. Even so, the foam would collapse very quickly after the purge gas stream was removed. The phenomena observed in run 15 and later remain as puzzling because the opportunity to perform realistic tesis to resolve unanswered questions was lost with the termination of operations of the reactor. It appears that development of a void fraction of unprecedented magnitude took place during a period of rapid corrosion of the containment system and that contaminants which entered the core of the reactor while it was exposed to the atmosphere during maintenance were the source of the corrosion. These contaminants mark- edly affected the interfacial energies of salt-metal, salt-gas, and salt-graphite surfaces but did not generate foam in the fuel circuit. The conditions which allowed these events to occur are avoidable and are atypical of the operating procedures that are projected for future MSR’s. Additional investigation of the physical properties of molten salts will be necessary to establish quantitatively the relationships of interfacial energies in salt systems to the retention of and stability of bubbles in flowing streams. The initial stages of these investigations can properly be carried out in the laboratory; however, the results may prove to be wmrelevant unless the effect of highly radioactive fluxes on the properties is shown to be inconsequential. 88 PHOTO 1867~ 71 Fig. 6.10. Appearance of nickel rod from FP-14-55 after Fig. 6.11. Nickel cage after second exposure of beryliium to suspenston in MSRE pump bowl for 2 hr. (¢) Upper end, 233U fuel salt. October 13, 1968, FP-15-30, § 34 g Bef (b) lower end dissolved from Be rod = i Fig. 6.12. Copper capsule containing several short magnets before exposure to 233U fuel salt. 89 Fig. 6.13. Metallic particles on copper capsule used to expose a magnet to 233U fuel salt. November 15, 1968, FP-15-61. Dendrites on capsule were composed of fine particles of iron. 90 e L 7% < M t8 g —~ = = O 59 2L ~ oy ~ £ a } g nchF o) S ) = : 2% 00} 9 RN o -8 O . =R k- - TR o -9 S 5 T = 223 o o mfl = lvm 2, 2.3 ~ & & 8 o -] o — ) - 5 = o 2 L o0 &% o = e 2 0O - £ o3 Gy O=E < .um Wy 2 B [ 4] 6SM v my . Q= Q. 8 = o = M= mgs = 8 .= = 3 E= el 4”a ¥ E 8 B -t o0 g =3 S w8 e, b & B - w # TS E 5 o% 8 ) o L4 5 S -t PR i ¢ 5y i Houg o oy, g it M) AP CRT A i mfl. o PHOTO 1872—71 L B | el ot o e wmw . = &2 e St B m«mmfi el B B E TR Ny ety @ fihwrmfm%& e o o 91 9.38¢ Be® dissolved from + E flium exposure in 233y fuel salt. November 15, 1968, FP-15-62 d bery 1F kel cage from th ic Ni . 6.16. g F rod 92 . PHOTO 94278 Fig. 6.17. Nickel cage after fourth exposure of berylhum to 233U fuel salt. November 15, 1968, FP-17-8, 8 57 g of Be? dissolved from rod PHOTO 94273 Fig. 6.18. Top of nickel cage from fourth beryllium exposure in 233U fuel salt January 22 1969 FP-17-8, 8 57 g Be® dis solved from rod After leach treatment with Verbocit and nitric acid solutions 93 1 lPHOTO 1873-711 Fig. 6.19. Chromium rod exposed to 2337 fuel sait. January 30, 1969, FP-17-11. e £y 3 5™ g PHOTO 94272 £ Fig. 6.20. Cross section of chromium rod showing surface deposit. January 30, 1969, FP-17-11. &) ' 3 YE T v AR ] piSE AR e 3, ki G s : o g £ 94 § PHOTO 1874 —71 o F{% v £ o 5 o Of e kS %r Y \ i s T ¢ XA 5 EL L BRI e 5 £ T i ”i'g < Y = =3 -, o - iy s e ¥, =) Tl 5% ¥ B b % e 3 5 AT e Bye B i) pg® % £h 2 = ) & s k) o i T 5 o ) i L e P o V.M e L o o i B0 e 5 §W 5 i, i G =% B g R F el " 5 4 s 5 T e ol s T e e o e £ Fig. 6.21. Nickel cage after exposure of beryllium rod to 2337 fuel salt. October 8, 1969, FP-19-48; 4.91 g BeY dissolved from rod. 6.5 Effect of Uranium Trifluoride on the 25 Nb Concentration of the Fuel Salt Minor adjustments in the concentration of uranium trifluoride in the MSRE fuel salt were made occasion- ally during the period when the MSRE was operated with 233U fuel. Their primary purpose was to offset the oxidizing effects anticipated to result from the fission reaction. Within this period the [U*"]/[ZU] concentration ratio was estimated to have varied within the range 0.1 to ~1.7% (Fig. 6.24). No evidence was found that indicated that such variation effected significant changes in either corrosion rate or fission product behavior in the fuel salt within the reactor. This derives from the fact that the ?35U fuel was a highly buffered system in comparison with the 233U fuel used later; that is, the total amount of uranium in the 225U fuel exceeded that contained in the 233U fuel by sixfold. In contrast, operation of the MSRE with PHQOTO 4875 — 74 L 3§ f TR P ol e By SRS ) - 5, ,%5) ‘273-3 ANCE Fig. 6.22. Nickel cage from exposure of beryilium rod to 233y fuel salt. November 29, 1969, FP-20-7: 6.97 g Bed dissolved from rod. 233U fuel showed pronounced changes 1n the corrosion rates and fission product chemistry as the concentration of UF; was altered. Of considerable interest was the appearance of ?*Nb in the fuel salt, noted for the first time in initial operations with 233U fuel.?® Ths observation signaled the potential application of the disposition of ?*Nb as an in-line redox indicator for molten-salt reactors. Operation of the MSRE with 33U fuel thus gave evidence that pronounced changes in fission product and corrosion chemistry resulted as variations of the concentration of UF; in the fuel salt were made. The 233U fuel was experimentally more tractable for study than the previous charge of 235:238J fuel because of the much lower uranium inventory carried in the 223U fuel. A serious disadvantage was realized, however, when it was discovered that the total amounts of UF; which were obtainable in samples of the 233U fuel caused the previously satisfactory method?! used for e e s Sl 6l P BT 2 i g E:3 Sty s, T o S SRS R e . hmmwmnfimfifiw AR E = L FP-20-22; 9.87 g Be" dissolved. ? 95 PHOTO 1876— 74 s y‘syfi;@ ol 8™ Lo n A N % k3 3 e R 8E T F e w 1969 b Fig, 6.23. Nickel cage from exposure of beryllium rod to 233y fyel salt. December 9 96 ORNL-DWG 7i- 9991R 1.8 1.6 o N o ANALYTICAL RESULTS N (Hy- HF TRANSPIRATION METHOD) 1.4 ) \ N \\ 3 N ~ 1.0 o (u3h/izm 0.4 f\ i 0.2 \\ ~ --------- = 5% %150 222U moles o | LR PP 4% bt b4 FREEZE VALVE SAMPLES | 0 f 2 3 4 5 6 7 8 (x10%) full power hours Fig. 6.24. [U3+]/[EU] in the MSRE fuel salt, runs 5—14, Assumed maximum power, 7.4 MW(t). determination of U3 in the fuel salt to be of little value and required the development of new methods of analysis. Adjustments of the uranium trifluoride concentration of the MSRE fuel salt were made frequently during 2337 operations in order to study the inhibition of corrosion by control of the [U**]/[ZU] concentration ratio and to evaluate the possible application of °5Nb disposition as a redox indicator. The experimental results obtained during this period of operations pro- vided concrete evidence that values of the equilibrium constant for the reaction Cr° + 2UF, = CrF, + 2UF,, as assigned from standard free energy and activity data, are reasonably accurate. Together with niobium distri- bution data they show that after the fuel system had been opened for maintenance the fuel salt subsequently appeared to become oxidizing with respect to the MSRE containment circuit even though it did not seem to have caused corrosion in the drain tanks. Compari- sons of the relative fraction of the ?>Nb inventory which appeared in the fuel salt as the redox potential of the salt changed indicated that when mildly reducing conditions were imposed, as for steady-state operation of the MSRE, niobium very likely became involved in reaction with the moderator graphite to form niobium carbide. During August and September 1968 the 233U fuel charge was constituted from "LiF-2**UF, and "LiF- BeF, -ZrF, carrier salt which had previously contained 235,2381F,. Concurrent with the inception of corro- sion in the fuel circuit, as evidenced by a rapid increase in the concentration of Cr in the fuel salt (Fig. 6.25), niobium began to appear in the fuel salt for the first time?! and persisted there until 28.80 g (6.54 equiva- lents) of Be® had been added. During this period the concentration of chromium in the salt rose from 35 to 65 ppm, indicating the removal of 140 g (5.89 equivalents) of chromium from the circuit walls. Thus, when ?SNb disappeared from the fuel salt after the final addition of beryllium (sample FP-15-62), a total of 97 ORNL DWG 69 5503 1968 - B - o é o o e o a Be T © o = & S (ADDED i mo ° o — i i H I 1 3 J i f ] s 80 & g C ) XIRE: ; i & e 020 — ° " Lz £ —_ g © = = = = EE @ o ] a 2« o - .f I — 60 & o o ) o (TATE! T d © Y o< 9, | = - 5 = 0 w = 040 = S 2 o0 &C b s =0 35 s - 50 ¢ eq h & i z se | &5 e © = Q 55 u —_ e e 40 QD I z o I 530 _m«"f‘r o {0 g SAMPLES & 0 b ig e FREEZE VALVE SAMPLES | o <1 G N 4 i % é o 1 30 I = o o = g L ! 1 | i ) 20 SEPTEMBER OCTOBER NOVEMBER DECEMBER | JANUARY FEBRUARY 1969 Fig. 6.25. Corrosion of the MSRE fuel circuit in run Nos. 15 and 16, September—December 1968. 12 43 equivalents were mvolved in the reduction of Fe?* and establishment of the Cr® + 2UF, = 2UF, + CiF, equilibrium Samples of the salt obtained during the brief pertod of the subsequent run (No 16) as well as the beginning of 233U power operations in run No 17 (FP-16-4 and FP-17-2) showed the presence of 52 and 29% of the ? 3 Nb inventory respectively In Fig 626, nomnal values of [U3]/[ZU] are shown for runs 17—20 These operations comprse nearly the total power operation of the MSRE with 2337 fuel The concentrations of UF; shown m Fig 6 26 are based on the assumption that 0 76 equivalent of oxidation results from the fission of 1 at wt of uranium® and that the maximum power achieved by the MSRE was 74 MW(t) They also assume that the observed increases 1n the concentration of chromium in the fuel salt at the beginning of runs 19 and 20 are indicative of the total loss of U?* from the salt, the nominal values for [U**]/[ZU] at these instants are thus shown as zero The equibibrium constant for the corrosion equibbrium Cr® + 2UF, = 2UF, + CrF, reaction at 650°C, assuming an activity for Cr® mn the Hastelloy to be 0 03, 151 271 X 107 Thus, 1n a regime such as that which prevailed during the 1nitial stages of run No 19, the rate at which Cr° 1s leached from the Hastelloy N circuit gradually decreases as the Cr®* concentration of the fuel salt increases During the mnitial period of run No 19 the Cr?* concentration of the circulating fuel salt rose from 72 to 100 ppm At that point the equhbrium concentration of [U*1/[ZU] 1n the fuel salt anticipated from free energy and activity data 1s ~0 5% The disposition of *>Nb 1n the 22U fuel durnng the mitial period of run No 19 ndicates that when [UP*]/[ZU] was less than ~0 5%, Nb became oxidized and entered the salt, possibly as Nb>" or Nb** Then, as the corrosion reaction Cr® + 2UF, = 2UF; + CrF, proceeded to equilibrium, the U**/TU concentration ratio increased, and at a [U*]/[ZU] value of 0 5%, ?SNb precipitated from the fuel salt Two levels of nominal U**/ZU concentration are given during run Nos 17 and 18, the hugher values based on the assump- tion that corrosion of the fuel circuat during the early stages of run No 17 may have accounted for a fraction of [U*]/[ZU] The extent to which this reaction might contribute to the total concentration of UF; in the fuel at the beginning of run No 17 1s obscure, because the MSRE was operated at full power at the inception of run No 17 Such operation deposits the noble metal fission products on the surface of the Hastelloy N, causing the activity of Cr® at the alloy surface to be effectively reduced The beginning period of run No 19 1s not analogous, for not until the corrosion equilibrium was established was the reactor operated at full power for sustained periods Freeze-valve samples were obtained at the request of E G Bohlmann and E L Compere during runs Nos 1720 Their analyses of the salt removed from the pump bowl showed ?>Nb disposition as indicated by the data points in Fig 626 It 1s evident that, as [U*]/[ZU] decreased below 0 5% 1n run No 17,%°Nb was oxidized and distnnbuted to the fuel salt, to be removed subsequently as this concentration ratio was exceeded Of the data shown in Fig 626 only the 98 ORNL-DWG 6914557 3.0 N 2.8 — N 2.6 U3+/SU NOMINAL CONCENTRATION - = U3T/ZU NOMINAL CONCENTRATION, ASSUMING CREDIT FOR CORROSION AT BEGINNING OF RUN NO. {7 2.4 s 9%Np EXPERIMENTAL DISPOSITION OF 25Nb 2.2 2.0 1.8 & 16 7 . m 1.4 N I y \ i \h\ \\ h\ \\ My \ 2 g \\ < N \ \ 60 Y X \\\ \ \ \ 1.0 NN A hd " 50 \\\\ 4 A \ 0.8 AY . 40 @ AR 3 \ - 0.6 \ A b A AN . 30 ‘z S \\\\\\\\\\\w\\m\{m\m\ AMMHTITET g - oos 0.4 ] y / 20 o \ 0.2 AN 10 /] 4 / I/ 4 0 it / b 0 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Mwhr (x103) \ y I\ . I\ y N R— RUN 17 RUN 18 RUN 19 RUN 20 Fig. 6.26. Effect of U>'/XU on distribution of >°Nb in the MSRE fuel salt. second data point in run 18 appears to be anomalous. It seems likely that this result represents incomplete equilibrium since the sample was obtained just after [UP*]/[ZU] had been adjusted to a nominal value of 0.7%. In view of the fact that Be?, in contact with the molten fuel, generates reduction potential gradients which result in the reduction not only of U*" to U** but of Cr*" to Cr® as well, one might anticipate a kinetic factor to be significant in Be® = Be?", U** = U, Cr** == Cr%, Nb3 = Nb? equilibria in the MSRE fuel salt. The results described in Fig. 6.26 indicate that when the [U>]/[ZU] of the 233U MSRE fuel was poised at ~0.5%, disposition of ?>Nb toward solution in the salt or deposition within the reactor was at a null point, and, as indicated by behavior during ??°U operations, the U>*/°5Nb ratio is the controlling factor. At 0.5% U?* this ratio is 8.5/1. Little is known concerning the chemistry of niobium in the MSRE fuel salt. Preliminary results of laboratory experiments indicate?? that under mildly oxidative conditions niobium assumes an oxidation number of ~3.7. The fact that during run 15, when the oxidation potential of the fuel salt was sufficiently high to permit Fe?" to exist in the salt in significant concentrations, nearly all of the 3Nb inventory of the fuel salt was in solution, whereas in subsequent operations when the oxidation potential was less, no greater than ~50% of the ?>Nb was found in the salt. This behavior seems to indicate that when niobium was deposited on the moderator graphite it reacted to form niobium carbide and that under the various redox regimes which were established in the MSRE during runs 17 to 20 niobium carbide was not removed from the moderator graphite. The prevalence of niobium as the carbide is compatible with the experimental observations by Blankenship et al.'® and as noted by Cuneco and Robertson,?® who found that the concentration of *>Nb at all profiles in three different types of graphite was greater than would have been anticipated if after deposition the isotope re- mained as the metallic species. References 1. A. Taboada, MSR Program Semiannu. Progr. Rep. July 31, 1964, ORNL-3708, p. 330. 2. W. R. Grimes, Nucl. Sci. Appl. Technol. 8, 138 (1970). 3. M. B. Panish, R. F. Newton, W. R, Grimes, and F. F. Blankenship, J. Phys. Chem. 62, 980 (1962). 4, J. O. Blomeke and M. F. Todd, “Uranium-235 Fission Product Production as a Function of Thermal Neutron Flux, [Irradiation Time, and Decay Time,” USAEC report ORNL-2127, Oak Ridge National Lab- oratory (November 1958). 5. C. F. Baes, “The Chemistry and Thermodynamics of Molten Salt Reactor Fluoride Solutions,” Proc. TAEA Symposium on Thermodynamics with Emphasis on Nuclear Materials and Atomic Transport in Solids, Vienna, Austria, July 1965, 6. W. R. Grimes, G. M. Watson, J. H. DeVan, and R. B. Evans, “Radiotracer Techniques in the Study of Corrosion by Molten Fluorides,” Proceedings of the Conference on the Use of Radioisotopes in the Physical Sciences and Industry, Sept. 6—17, 1960, vol. 1Ii, p. 559, International Atomic Energy Agency, Vienna, Austria, 1962. 7. H. E. McCoy, An Evaluation of the Molten Salt Reactor Experiment Hastelloy-N Surveillance Speci- mens — Fourth Group, ORNL-TM-3063 (1970). 8. MSR Program Semiannu. Progr. Rep. Aug. 31, 1965, ORNL-3872, p. 113. 9. MSR Program Semiannu. Progr. Rep. July 31, 1963, ORNL-3529, p. 130. 10. J. F. Engel, P. N. Haubenreich, and A. Houtzeel, Spray, Mist, Bubbles, and Foam in the Molten Salt Reactor Experiment, ORNL-TM-3027 (June 1970). 11. Letter dated February 9, 1971, from D. B. Trauger to M. Shaw, Assessment of Need for Oxygen Getter in MSBR Fuel. 12. Furmnished through the courtesy of C. Crouth- amel, Chemical Engineering Division, Argonne National Laboratory, Argonne, Il 99 13. W. R. Grimes, unpublished work, 1970. 14. B. McNabb and H. E. McCoy, internal corre- spondence, March 31, 1971, 15. R. E. Thoma, MSR Program Semiannu. Progr. Rep. Feb. 28, 1970, ORNL-4548, p. 93. 16. F. F. Blankenship, E. G. Bohlmann, S. S. Kirslis, and E. L. Compere, Fission Product Behavior in MSRE, ORNL-4684 (in preparation). 17. S. 8. Kirslis and F. F. Blankenship, MSR Program Semiannu. Progr. Rep. Feb. 28, 1967, ORNL-4119, p. 131. 18. J. H. Shaffer and W. R. Grimes, MSR Program Semiannu. Progr. Rep. Feb. 28, 1969, ORNL-4396, p. 135. 19. H. W. Kohn and F. F. Blankenship, ibid., p. 137. 20. J. M. Dale, MSR Program Semiannu. Progr. Rep. Aug. 31, 1967, ORNL-4191, p. 167. 21. E. L. Compere and E. G. Bohlmann, MSR Program Semiannu. Progr. Rep. Feb. 28, 1969, ORNL-4396, p. 139. 22. C. F. Weaver, private communication. 23. D. R. Cuneo and H. E. Robertson, MSR Program Semiannu. Progr. Rep. Aug. 31, 1968, ORNL-4344, p. 141. 7. DETERMINATION OF REACTOR POWER 7.1 Power Estimates with 23° U Fuel from Heat Balance and Qther Methods Estimates of the power developed by the MSRE can be derived from nuclear and heat balance data. Both bases were used from the beginning of power operations with the MSRE. As refinements and corrections wete introduced into the physical property and nuclear data, new estimates were made. Consequently, MSR Program progress reports cite various values for the maximum average power of the reactor ranging between 7 and 8 MW(t). The results of the investigation described in this chapter fixed the maximum power of the MSRE at 7.4 MW(t), a value with which other estimates have more recently agreed. The power generation rate of the MSRE was cal- culated routinely by an on-line computer; power production was determined from computation of the heat balance in the fuel and coolant systems. Details of the methods employed are described elsewhere.!*? Once heat balances were established, nuclear instrumen- tation systems were calibrated to correspond to the nuclear power indicated by the heat balance. The results of chemical and isotopic analysis of fuel salt were assessed carefully throughout the period when the MSRE was operated with ?**U fuel and later when the fuel charge was comprised of 223U and plutonium, with the purpose of employing these results to monitor, if possible, the power generated by the reactor. Attempts to use chemically determined values of the concentration of uranium in the fuel-salt samples were generally unsatisfactory because of the point-to-point scatter in these data and because of the overwhelming effect that the use of flush salt in the fuel circuit had on the uranium concentration base lines. The amounts of fuel- and flush-salt residues remaining in the fuel circuit after drains were estimated by comparisons of the chemical analyses of uranium in the flush salt with those in the fuel salt during each period of use. Not until the uranium concentration of the flush salt had undergone several increments was the precision of the average mass of fuel-salt residues narrowed to within 5 kg. This was, however, of insufficient precision to afford a sufficiently accurate base line for computation of the power generation. It was thus evident that the results of wet chemical analyses for uranium in the fuel salt would be of potential use in establishing burnup rates only after long periods of power generation which were free from drain-flush-fill interruptions. Until March 1968, calculations of the heat balance indicated that the maximum power generated by the MSRE was 7.2 MW(t). By that time, a discrepancy between nominal concentration and analytical values for uranium in the circulating fuel salt began to appear; the analytical data over an extended period in run No. 10 showed a negative divergence from the nominal concentration of the uranium in the fuel of about 10%. During this period, however, computations of the reactivity balance did not indicate corresponding or anomalous decreases in the fissile concentration of the salt. Reactivity balance calculations had previously indicated that this method of evaluating reactor per- formance was sensitive to a factor of 10 greater than chemical analysis with respect to detection of changes in uranium concentration of the fuel. The chemical data suggested that the maximum actual power output of the reactor was ~8.0 MW rather than 7.2 MW. However, little credibility could be accorded to this conclusion for the reasons cited above. By early 1968 a sufficient amount of 23°U was generated in the fuel salt to suggest that comparison of the analytical results of mass spectrometric measure- ments would provide a good measure of the integrated power. Tests of this comparison® yielded a slope that was within 1% of the theoretical slope for a power 100 ORNL~DWG 68- 5509 1000 [ ] pd ./ 800 o/ 82" = o = - ® 4 t5 600 4 >_ ) / = 1@ & . / 2 L " 400 < S g 7 & = % 200 0 0 10 20 30 40 50 (x10°) INTEGRATED HEAT-BALANCE POWER (Mwhr} Fig. 7.1. 236y buildup in MSRE vs power production. generation rate of 7.2 MW. The uncertainty in the theoretical slope was regarded to be probably less than 10%, and the actual integrated power was probably within 10% of the value indicated by the heat balances (see Fig. 7.1). Although the 23®U production rate seemed to con- firm estimates that indicated the maximum power output to be 7.2 MW, suspicion was growing that the heat balance calculations were in error, for data collected at different power levels indicated that the value employed for the specific heat of the LiF-BeF, coolant salt was not temperature dependent,® whereas a temperature-dependent relationship was employed in the heat balance.® A program of laboratory measurements of the en- thalpy of solid and liquid Li, BeF4 from 273 to 900°K was completed at this time, by investigators at the National Bureau of Standards,® from which it was shown that the heat capacity was 0.56 cal g 7' (°C)™1. A fluoride mixture was synthesized for confirmation at ORNL of the NBS value. Results of these measurements indicated that the derived heat capacity of the coolant sali was 0.577 £ 0.008 cal g7'(°C)™",” in good agreement with the NBS investigations but substantially higher in the operating temperature range of the MSRE than the previously used value, and showed essentially no variation with temperature. The new value of the specific heat was incorporated into the computerized heat balance computations prior to the beginning of operatton with **3U fuel Calculated full-power level was changed from 72 to 80 MW as a result of the revision in the value for specific heat and m accord with the analytical chemical results Throughout this entire period, changes in the nominal amounts of uranium 1sotopes in the fuel salt were computed based on average crosssection data for thermal and epithermal neutron reactions in the MSRE spectrum that were current to 1965%? The con- sumption and production rates were 234y ~2 167 X 107™% g/MWhr 235y 5417 x 1072 g/MWhr 236y +1 083 x 1071 g/MWhr 238y 7017 X 1072 g/MWhr Total ~5146 X 1072 g/MWhr¢ 4Ret 9 7.2 Power Output of the MSBR Based on the Isotopic Composition of Plutonium The potential use of plutonium as a fuel for molten- salt reactors has been assessed periodically for more than a decade The results of one early study'® showed that a PuF;-fueled two-region homogeneous fluoride salt reactor was operable, although its performance was poor Further development was not pursued for neither the chemical feasibility nor methods for mmproving its performance were obvious Although the thermo- chemical properties of the plutonium fluorides were not well established at that tume, 1t was clear that the most soluble fluoride, PuF,, was too strong an oxidant for use with the available structural alloys The solubility of PuF;, while sufficient for cnticahty even in the presence of fission fragments and nonfissionable 1s0- topes of Pu, was assumed'! to linut the amount of ThF; which could be added to the fuel salt This limitation, coupled with the condition that the con- tinuous use of 23?Pu as a fuel would result 1in poor neutron economy in comparison with that of ?33U- fueled reactors, vitiated further efforts to exploit the plutomum fluornides for application in two region MSBRs Recent developments in fuel reprocessing chemustry and 1n reactor design have established the feasibility of single-fluid molten-salt breeder reactors One of the alternative modes of operating such reactors is to employ plutonium in place of enriched ??3U for the 101 imtial fuel loading and startup of the reactors Assess- ments of plutomum to start up MSBRs 1n this way have concluded that plutomium should prove to be a very satisfactory alternative to enriched uranium, either as initial fuel for a molten-salt breeder reactor or as mnitial fuel and subsequent feed material ' 2714 It now appears that 1t will be possible to operate an LiF-BeF,-ThF,-PuF; single-flnd molten-salt reactor with lower concentrations of thonum and plutonrum than earlier considerations required, for example, with thorium fluoride concentrations of 8 to 12 mole % and with a plutomum fluoride concentration of approxi- mately 25% less than required for 233U loading, that 1s, <02 mole % These conclusions indicated the de- sirabtlity of demonstration experiments to examine, 1n as nearly similar application as possible, the behavior of plutontum m an MSBR The chemical feasibihty of this application was evaluated!® and found to hold pronuse for successful application Therefore, when the MSRE resumed operations with 222U fuel, plans were made to include plutonium as a constituent of the fuel Using pre-1965 cross-section data and assuming that the maximum operating power of the reactor was 8 0 MW, efforts were 1nitiated to establish a material balance for plutonium. During the final period of operation with fuel, a sufficient amount of plutonium was generated for 1ts detection n salt samples to be tractable by standard analytical methods Approximately 600 g of plutonium was generated in the MSRE fuel as 1t was operated with 235U fuel by neutron absorptions in 2381, enough to afford a comparison of the analytical data with anticipated values The results of that comparison, expressed as a matertal balance for plu- tonum, showed that the analytical chemical methods, which were previously satisfactory for determination of the concentration of plutonium in the fuel salt, weire of questionable utility for use with the 232U fuel charge and that 1sotopic dilution methods, using mass spectro- metric analyses, afforded the most satisfactory means of analysis They mdicated, in addition, that at the maximum concentrations in which plutonium occurred in the MSRE fuel salt, 1t existed as a stable chemical entity and, by inference, that Pu, O3 1s not precipitated in the presence of low concentrations (50 to 60 ppm) of oxide 1on The net production rates for plutonium generated n the 235:2381J fuel salt have been estimated’ ¢ to be 235,238y G239 = G238.( 018584(¢—0 605377’ T =0 3318757y G240 = G238.(( 008233~ 0 29166 7° T 0 0648799¢—0 291667 T +0056647¢—0 333187°7) where Gk = mass of k i crculation (g) (k = 238, 239, 240 refers to 238U, 229Pu, and 24 °Pu respectively), T = time mtegrated power (MWhr) At termunation of the %3°U expeniment, 589 g of plutonium should have been generated in the fuel salt with the reactor operating at a maximum power of 8 0 MW For a fuel charge of 4900 kg, this corresponds to 120 ppm The average concentration of plutonium 1n the fuel salt, as determuned from the results of analyses of 18 fuel salt samples obtained during the latter period of power operations with 235238 fuel, was 118 ppm The analytical data do not show a significant trend and are probably not sufficiently precise to use as a basis to mnfer that a real difference exists between calculated and analytical values On completion of these power operations, urantum was removed from the fuel salt by fluorination ' 7 Tt was anticipated that the plutonium would remam 1n the carrier salt Five samples of the carrter salt were removed from the fuel drain after fluorination and found to have an average concen tration of 120 ppm of plutonium, representing a total of 562 g, as compared with an expected value of 125 ppm, that 15, 27 g of plutontum 1s not accounted for by the results of these analyses The concentration of plutonium 1n a salt specimen 18 calculated from the relation Concentration of Pu (ppm) = dpm/g X 3 97 X 107® (338N X 1470+ 239N X 5402 X 107% + 240y X200X 1072 +24 N X 410X 107 +2%2NX 35X 107%), where N 1s the atom fraction of plutonmum as the 1sotope designated The concentration of plutonium 1n each of the 10 g samples of fuel salt taken since the beginning of 233U operations was determined using conversion factors which assumed that the plutonium in the MSRE consisted entirely of 2*°Pu and 2#°Pu The average concentration of plutonum after full loading of the reactor was achieved was found to be 147 ppm (Table 102 7 1) Ths value indicates the presence of a total of 689 g of plutomum, thus the enriching salt might then have been expected to contain 100 g of plutonium In an attempt to substantiate this conclusion a sample of the "LiF 2?3UF, ennching salt was obtained from a section of transfer line at the TURF and submutted for chemical and mass spectrometric analysis The salt residue which was contained in the line was considered to be typical of that delivered for use in the MSRE The results of mass spectrometric analysis showed that this salt contained plutontum 164 wt % of which was 238Pu Rewvised conversion factors were therefore re quired for calculation of the total plutonium concen tration of the TURF salt and 233U fuel salt because of the high specific activity, 6 46 X 10° dis sec™ pg™", for 238Pu The plutomum concentration of the TURF salt appears now to have been 249 ppm using the revised factor and corresponds to the addition of 16 4 g of plutonium along with the "LiF 223 UF, eutectic Current estumates of the nominal concentration of plutonium are now based on the assumption that this amount of plutonium was added to the fuel salt and, therefore, that the MSRE contained ~605 g of plu tonium at the beginning of 233 U power operations A plutontum 1nventory of the MSRE fuel was computed, both before and after loading with 223U fuel, based on Prince’s estimates of the production and fission rates for plutonium and on the assumption that 16 4 g of plutomium was contained 1n the enriching salt These values showed that the plutonium production rate during *3° U operations was greater than previously antictpated and, correspondingly, that the relative changes 1n ?*%Pu and 23°Pu during 233U operations were also shightly different Estimates of the variation of 23°Pu and 2#°Pu during recent power operations require that the quantity and composition of the final plutonium nventory be known accurately As noted previously,'® and as shown m Fig 7 2, attempts to determine the concentration of plu tomum 1n the fuel salt from gross alpha count measure ments were not very satisfactory because of the high specific activity of 238Pu An improved estimate of the plutonium iventory of the system was made from extrapolations of the observed changes in *>°Pu and 24%Pu 1n the beginning stages of power operation with 233U fuel The mitial 24 °Pu/?3°Pu concentration ratio was computed to be 00453, with the plutonium of the reactor at that point as 568 g, approxmmately 2% more than estimated from previous analyses Current esti mates of inventory values have been computed for this revised starting mventory The values obtaining at the time the samples were taken were based on estimated 103 Table 7.1. Summary of MSRE fuel-salt analyses: plutonium Net Weight of Pu in Weight % Pu/ZPu Concentration of Sa;nopie Mwhr Fuel Sait (g) Calculated Analytical Puppm) ' 23%py 240py z 239p, 240p, 23%p, 240p, Calculated? Analytical FP1441 67,767 487 18.6 506 96.32 3.68 96.46 3.31 94 a5 FP14-42 62,305 492 18.8 511 95 120 FP1443 62,705 495 19.0 514 95 125 FP14-44 63,251 498 19.3 517 96 128 FP14-46 63,537 500 19.5 520 96 127 FP1447 63,671 501 19.6 521 97 122 FP14-48 64,211 505 19.8 525 97 126 FP14-49 64,232 505 19.8 525 97 125 FP14-50 64,234 505 19.8 525 97 123 FP14-51 64,367 507 20.0 527 97 119 FP14-52 64,994 510 20.3 530 98 111 FP14-54 65,397 513 20.5 534 99 98 FPi4-56 65,809 517 20.8 538 100 119 FP14-58 66,351 521 21.2 542 101 118 FP14-59 66,982 525 21.8 547 102 120 FP14-64 68,720 537 22.7 560 95.94 4.06 96.05 3.67 104 145 FP14-65 69,481 543 233 566 114 97 FP14-68 69,838 546 235 570 115 102 FP14-Final 72,454 563 25.6 589 120 120 FP15-6 72,454 571 27.1 599 128 113 FP15-9 72,454 573 2.5 602 128 112 FP15-10 72,454 575 27.9 604 129 96 FP15-12 72,454 575 27.9 604 129 112 FP15-18 72,454 575 21.9 604 129 100 FP13-33 72454 575 27.9 604 129 134 FP15-38 72,454 576 28.0 605 129 143 FP1542 72,454 576 28.0 605 129 135 FP15-60 72,454 576 28.0 605 129 129 FP15-63 72,454 576 28.0 605 129 141 FP15-65 72,454 576 28.0 605 129 159 FP15-68 72,454 576 28.0 605 95.85 3.99 95.43 4.12 129 157 FP17-1 72,454 576 28.0 605 129 130 FP17-4 73,127 575 28.0 605 129 134 FP17-9 73,830 572 29.4 602 128 148 FP17-12 75,434 569 304 598 128 138 FP17-18 76,183 564 3L.9 597 127 149 FP17-19 76,791 563 325 596 127 147 FP17-20 78,026 560 32.7 594 127 141 FP17-23 79,154 557 34.8 593 127 130 FP17-24 79,814 555 35.5 592 126 142 FP17-27 80,828 552 364 589 126 140 FP17-28 81,610 550 37.2 588 125 134 FP17-30 82,771 547 38.3 586 125 149 FP18-1 84,741 542 40.1 583 124 160 FP18-5 86,454 538 41.8 581 124 162 FPi8-10 87,267 536 42.5 580 124 145 FP18-13 88,265 533 43.5 578 123 154 FP18-22 89,886 529 44.8 575 92.63 7.20 91.81 7.33 122 164 FP18-27 90,898 527 45.7 574 122 144 FP1843 92,142 524 46.9 570 92.30 7.53 91.38 7.70 122 171 FP18-Final 92,985 524 47.3 570 122 @Based on total fuel charge. 104 ORNL-DWG 69-7561 . i80 170 . lfi . f/ \ \y | i ffi |V —— > o o o g \\\- NOMINAL CONCEN’TRATIONZ \ BEGINNING OF 233U POWER OPERATIONS ~ | 1 | 1 % %l END OF 233y OPERATIONS Y [ PLUTONIUM (ppm) LY 100 f ! o 90 : . 80 5 55 60 65 70 70 75 80 85 90 (x10°) . Mwhr Fig. 7.2. Comparison of nominal and analytical values for the concentration of plutonium in the MSRE fuel salt. Table 7.2. Isotopic composition of plutonium in MSRE fuel salt Sample No. Description 238p, 23%py 2%0p, 24ipy 242p, FP14-41 Z g (calculated) 487 18.6 a a % calculated 96.32 3.68 % analytical 0.011 96.45 3.32 0.22 0.002 FPi4-64 Z g {calculated) 537 22.7 % calculated 95.94 4.06 % analytical® <0.015 96.05 3.67 0.28 0.005 FP15-68 Z g generated in MSRE (calculated) 563 25.6 ¥ g added with 7LiF-?33UF, 0.27 12.9 2.44 0.37 0.33 g 0.27 575.90 28.04 0.37 0.33 .- % calculated 0.04 95.20 4.64 0.06 0.06 % analytical <0.08 95.43 4.12 0.34 0.13 a241py and 2*2py not included; B. E. Prince estimates T g 241,242p, < 1.8 g - bAnalyses performed by R. E. Eby. . average values for the rates of change of 2?°Pu and 240Py in the period between samples (Table 7.2). In the MSRE, fluid fuel was circulated at rates which were sufficiently rapid with respect to changes in the isotopic composition of the fissile species that the salt samples removed from the pump bowl were repre- sentative of the circulating stream. This characteristic of molten-salt reactors makes it possible to use the results of isotopic analyses for a variety of purposes. One potential application, that of appraising the cumulative power generated by the MSRE at various periods, became apparent with the initiation of 2**U opera- tions, for with 223U fuel the isotopic composition of the plutonium inventory (produced partly by that generated in 233U operations as well as from that added later) would change significantly during power production and would possibly serve as an accurate indicator of the power produced. About 600 g of plutonium was produced during power operations with 233U fuel. Thereafter, addi- tional plutonium was introduced into the fuel salt as a contaminant of 7 LiF-2?? UF, enriching salt and later to replenish the fissile inventory of the MSRE during 2331 power operations. 105 Additions of plutonium to the fuel salt as PuF; were accomplished smoothly by use of capsules sealed by 12 zirconium disks. These containers were designed and loaded by Carr et al,'® a typical capsule is shown in Fig. 7.3. In contact with the fuel salt the zirconium dissolved, permitting the PuF; to disperse and to dissolve in the fuel salt. Photographs of one of the capsules after use are shown in Fig. 7.4. Resolution of the analytical problems with plutonium showed that estimated production rates of plutonium in the 235U fuel salt, based on pre-1965 cross-section data, were low. The programs used for reactor physics calculations were revised using post-1965 data, and as a consequence, both consumption and production rates for plutonium and uranium isotopes were revised significantly.?° Samples of the MSRE fuel salt were submitted routinely for determination of the isotopic composition of the contained fissile species. Comparisons of the results of plutonium assays with nominal values that should result from operations at various power levels from ~7 to 8 MW were made. Within this range, best agreement between calculated and experimental values was obtained for a maximum power output of ~7.40 PHOTO 9747 s 5 2 [ ¥ f‘»gd‘éd‘r i @ Fig. 7.3. 239PuF3 capsule for MSRE refueling. e o L ] ey =5 5 k) iy o it o, 4 FP 19-25 TOP SECTION, OPPOSITE SIDE FROM THAT SHOWN IN OVERALL VIEW, FP 18-25 BOTTOM SECTION, OPPOSITE SIDE FROM THAT SHOWN IN OVERALL VIEW. SALT RESIDUE MORE CLEARLY EVIDENT HERE THAN IN OVERALL VIEW, 106 PHOTO 1877~ 71 FP 19-25 FIRST CAPSULE USED TO ADD PuFz TO THE MSRE FUEL SALT. CAPSULE WAS PARTIALLY SUBMERGED IN FUEL SALT FOR FOUR HOURS FP 19-25 CAPSULE VIEWED FROM BOTTOM, SHOWING SALT RESIDUE IN CAPSULE. Fig. 7.4. Capsule sections. Table 7.3. Isotopic composition of plutonium in the MSRE fuel salt at a power generation rate of 7.40 Mw(th) f . Fuel Circuit Inventory Isotopic Composition Isotopic Composition Sample pi:ir — A239py A240p, (Caloulated) (2)? (Calculated) (Analytical) No. e (e per 1000 Mwhe) (gper 1000 Mwho) 355 Tl TR0 WURRWIR gy g, WLERWEPU g g 239Pu 240Pu 239Pu 240Pu Run 17-1 0 0 5415 24,53 5684 9526 4.32 0.0453 FP 17-9 148 537 -3.025 +1,178 539.8 25.16 5674 95.14 4.43 0.0466 9526 4.35 0.0457 FP 17-18 466 2,227 ~2.996 +1.144 534.8 2709 5643 9476 4.80 0.0507 9428 5.16 0.0547 FP 17-19 542 3,656 ~2.949 +1.180 $30.6 2878 S61.8 9444 512 0.0542 9448 5.00 0.0508 FP1720 697 4,492 2,924 +1.105 528.1 2970 5603 94.26 5.30 0.0562 9420 5.5 0.0557 FP 17-23 920 5,862 ~2.894 +1.086 5242 3118 5578 93.97 5.95 0.0595 FP 1727 1047 7,31 ~2.871 +1.072 5205 3254 5555 9370 5.86 0.0625 9358 5.80 0.0620 FP17-28 1145 7,946 ~2.854 +1.063 518.2 3342 5540 9353 6.03 0.0645 9430 5.97 0.0639 FP17-30 1290 8827 ~2.845 +1.054 5157 3435 5515 9351 623 0.0666 93.16 6.18 0.0663 Run 17-F 1536 10,245 ~2.824 +1.039 S11.7 3582 549.9 93.04 651 0.0700 FP 181 1536 10,245 5141 3491 5514 9322 633 0.0679 FP 18-1 1536 10,245 ~2.875 +1.075 Si4l 3491 5514 9322 6.33 0.0679 9292 6.36 0.0684 FP 18-5 1562 11,231 ~2.860 +1.066 S13.5 3511 5511 9318 637 0.0684 9265 6.61 0.0713 FP18-10 1851 12,373 ~2.843 +1,055 507.6 3732 5473 9273 6.82 0.0735 9238 6.84 0.0740 FP18-13 1976 13,873 ~2.826 +1.038 505.0 3826 5457 9254 7.0l 0.0757 9216 7.04 0.0764 FP1822 1976 13,873 ~2.826 +1.038 505.0 38.26 5457 9254 7.0l 0.0757 91.80 7.36 0.0802 FP 1827 2306 15,523 ~2.799 +1.017 498.3 4069 5414 9203 7.52 0.0817 9163 7.49 0.0817 FP 1843 2461 17,281 ~2.773 +0.999 4952 41.81 5395 91.80 7.75 0.0844 9148 17.63 0.0834 Run 18F 2544 18,143 2758 +0.989 493.5 4241 5384 9167 1.87 0.0859 Run 19 2544 18,143 4952 4181 5394 9179 775 0.0844 FP19-17 2625 18,739 ~2.770 +0.993 493.5 4239 5384 9167 7.87 0.0858 9122 7.84 0.0860 FP19-18 2642 19,093 ~2.763 +0.989 4932 42,51 5382 9165 7.90 0.0862 9119 7.87 0.0863 FP19-21 2724 19,449 ~2.755 +0.985 4916 4310 S537.1 9152 8.02 0.0877 9102 8.03 0.0882 FP19-22 2791 19,693 ~2.752 +0,985 490.2 43.58 5363 9142 8.13 0.0889 9090 8.11 0.0892 FP19-24 2791 19,693 2,752 +0.985 490.2 43,58 5363 9142 8.13 0.0889 89.88 8.99 0.1000 FP-19-256 2791 19,693 ~2.760 +0.983 5411 46,74 5906 91.62 7.91 0.0864 FP19-27 2791 19,693 ~2.760 +0.983 5411 46,74 5906 9162 791 0.0864 91.01 8.03 0.0882 FP19-30 2818 19,693 ~2.760 +0.983 540.5 4693 5899 9163 7.96 0.0868 90.89 8.13 0.0895 FP19-31-4 2818 19,693 ~2.750 +0.983 662.6 5451 719.6 92.08 7.58 0.0823 FP19-35 2964 20,961 ~3.820 +1.369 658.6 55.96 7171 91.85 1.80 0.0850 9135 777 0.0851 FP19-43 3102 21,989 ~3.810 +1.357 654.8 5732 7145 9164 8.02 0.0875 9118 7.90 0.0866 FP19-53 3294 23,185 ~3.795 +1.339 649.5 5918 7111 9133 832 0.0911 90.88 8.25 0.0908 FP19-63 3561 24,631 3,775 +1.339 6422 6177 706.4 9091 8.74 0.0962 90.49 8.50 0.0939 FP19-74 3693 26,078 ~3.750 +1.316 639.3 6277 7045 89.09 891 0.0982 90.15 8.78 0.0974 Run19-F 3774 27,069 ~3.720 +1.302 637.1 6354 7031 90.62 9.04 0.0997 Run20-F 3774 27,069 625.8 6181 6900 90.69 896 0.0988 FP 20-6 3820 27,236 ~3.670 +1.309 624.6 6225 6893 90.61 9.03 0.0996 89.89 8.99 0.1000 FP20-31 4159 28,294 ~3.660 +1.293 615.6 6543 6834 90.07 9.57 0.1062 89.36 9.43 0.1055 Run20-F 4159 28,294 ~3.660 +1.293 615.6 6543 6834 90.07 9.57 0.1062 2 Average for period between samples, b A ssumes 92% fuel charge in circulation, LO1 108 ORNL-DWG 706755 0.4% /O G50 a/ . /v /a Q.09 .?: /“ /V./ Ll/' ¢ o 0.08 2 240 /239Pu yd Q.07 i o, * 0.05 / e -] e ’\ 0.06 NOMINAL VALUES OF 2407239 AT 740 Mw (th) — ] 0.04 [ i | | @ RUN RUN 17—————'——-RUN 48—>‘<——RUN 19———-—'*;(7,' i ! | ! 0 1000 2000 3000 4000 equivalent full -power hours Fig. 7.5. [2*°/239Py] in the MSRE fuel salt circuit during 2> >U operations. MW(t). A comparison of calculated and observed values for the isotopic ¢omposition of plutonium which should result from a maximum power output of 7.40 MW is shown in Table 7.3 and in Fig. 7.5. Agreement tests indicate that the standard deviation between calculated and observed values is £0.63% and that the average positive bias in the experimental data is 0.093%. On this basis the maximum power output was 7.41 % 0.05 MW(t). The precision of this value seems to be adequate for related analyses of reactor operations. It is considered unlikely that further refinements in cross-section data for the plutonium isotopes will require any substantial changes in the calculated in- ventories used for this comparison. However, for the purpose of estimating uncertainty in the combined cross-section data and neutronic model used to cal- culate reaction rates, an equivalent of 2% in the power output is judged conservative. The results of experi- ments designed to measure the 233U capture-to- absorption ratio in the fuel of the MSRE were reported recently by Ragan.?! From these results he concluded that the full-power output of the reactor was 7.34 + 0.09 MW, in excellent agreement with our value. Further corroboration of these estimates was provided by Gabbard?®? in a recent reevaluation of the accuracy of the data produced by the MSRE coolant salt flow transmitters. He found that either the pressure trans- mitter or the square-root converter component, which generated 2- to 10-v signals for the computer, was faulty and caused the device to indicate higher than actual flows. Using a corrected flow rate of 793 gpm rather than a nominal flow rate of 850 gpm, as indicated by the computer, to recompute the heat balance power, he estimated that the full-power output of the reactor was 7.65 MW rather than the previous value of 8.2 MW, 7.3 Isotopic Composition of Uranium during 233U Operations For operation of the MSRE with 233U fuel, esti- mation of the power output of the MSRE from measured changes in isotopic composition of the fissile material is achieved with considerably greater precision from analyses of plutonium than from uranium. This arises from the fact that the rate of change in the relative fraction of the most abundant isotopes for plutonium, ?*?Pu and 2*°Pu, is some four times that for the uranium pair, 233U and 2% U. Analyses of the isotopic composition of uranium in the fuel circuit during 233U operations were employed, therefore, primarily to determine whether they atforded approxi- mate confirmation of the power estimate as inferred from the plutonium data (Sect. 7.2). Calculations of the 109 Table 7.4. Isotopic composition of uranium in the MSRE fuel-salt circunif? Sample U/ZU (Wt %) b 234 233 No, EFPH 233U 234U 235U '236U 238U { U]/[ U] Run 17-] 0 84687 6,948 2,477 0.0808 35.807 0.08204 FP 17-18 466 84.590 7.011 2.489 0.084 5.828 0.08288 84.690 6.990 2.470 0.084 3. 771 0.08253 FP 17-24 920 84.489 7.073 2,501 0.087 5.849 0.08371 84.382 7.058 2487 0.089 5.986 0.08304 FP 17-32 1338 84.440 7.131 2.510 0.090 5.867 0.08445 84,445 7.128 2,487 0.087 5.843 0.08440 Run 17-F 1536 84.363 7.152 2.513 0.091 5.875 0.08477 Run 18-1 1536 84,393 7.136 2.511 0.091 5.870 0.08455 Run 18-2 1536 84.199 7138 2.507 0.089 6.067 0.08477 FP 18-4 1563 84.385 7.141 2.511 0.091 5.871 0.08462 &84.249 7.158 2.527 0.091 5.975 0.08496 FO 18-10 1852 84.326 7.180 2.518 0.092 5.883 0.08514 54.269 7.178 2.507 0.091 5.955 0.08517 FP 18-13 1976 84,298 7.199 2.521 0.092 5.890 0.08539 84.060 7.203 2.517 0.087 6.133 0.08568 FP 18-22 2221 84,241 7.232 2.529 - 0.095 5.902 0.08584 54.167 7.208 2.517 0.098 6.016 0.08563 FP 18-43 2461 84.189 7.265 2.534 0.097 5.912 0.08629 84.041 7.232 2.537 0.093 - 6.097 0.08605 Run 18-F 2544 84.169 7.279 2,536 0.098 5.916 0.08648 Run 19-1 2544 84.185 7.267 2.534 0.097 5.913 0.08632 FP 19 10-12¢ 2544 84,224 7.268 2.526 0.097 5.882 0.08629 FP 19-35 2964 84,377 1.349 2.543 0.100 5.919 0.08709 83,987 7.338 2.537 0.099 6.036 0.08737 FP 1943 3102 84.103 7.348 2,539 0.101 5.909 0.08726 83.994 7.328 2,533 0.099 6.047 0.08724 FP 19-53 3294 84.060 7.375 2.546 0.102 5.917 0.08773 83.912 7.358 2.537 0.101 6.092 0.08768 FP 19-63 3561 84.000 7.413 2,553 0.104 5.931 0.08825 83.927 7.408 2,542 0.102 6,021 0.08826 FP 19-74 3693 83.971 7.430 2.555 0.105 5.937 0.08848 83.801 7418 2.569 0.102 6.102 0.08851 Run 19-F 3774 83,953 7.442 2,557 0.105 5.942 0.08864 Run 20-1 3774 83.973 7.427 2.555 0.105 5.938 0.08844 FP 20-3¢ 3774 83.996 7.426 2.549 0.105 5.923 0.08840 FP 20-6 3820 83.986 7.433 2.550 0.105 5.925 0.08850 83.614 7.435 2577 0.104 6.271 0.08892 FP 20-31 4159 83.911 7.482 2.558 0.108 5.940 0.08916 83.742 7.484 2.577 0.105 6.092 0.08936 Upright type indicates values computed on the basis that the maximum power generated by the MSRE was 7.41 Mw(th). The following rates (furnished by B. E. Prince) were used: 233U: —4.643 X 1072 g/Mwhr; 234U: +3.6325 X 1073 g/Mwhr; 23°U: +9.5596 X 1075 g/Mwhr; 236U: +3.0725 X 10~% g/Mwhr; 238U: —2,90 X 10™* g/Mwhr. Results of mass spectrometric analyses are listed in italicized type. bEquivalent full-power hours. “Fuel addition, 110 ORNL—-DWG 70-6756 0.080 P L. /'.,o’l// 0.088 - o/ Afl/ 0.086 QD ./ g /'/ § ./'.o 0.084 > - P _ NOMINAL VALUES OF 224233 AT 7.41 Mw (th) / 0.082 !' ; RUN 17 +—RUN 18~——|-*—F?UN 19 %—Rgg % 0.080 E | t | z \ | n 0 1000 2000 3000 4000 equivalent full — power hours Fig. 7.6. 2347233y in the MSRE fuel circuit during 233y operations. isotopic composition changes of the uranium in the MSRE fuel circuit that should have accompanied operation of the reactor at a maximum power output of 7.41 MW(t) were made and compared with the results of mass spectrometric analyses. The results of this comparison are shown in Table 7.4 and in Fig. 7.6; they indicate that the changes observed in the isotopic composition of the uranium were in excellent agree- ment with those of plutonium. 7.4 Isotopic Composition of Uranium during 235 Operations Once the anomalies in the power production of the MSRE were resolved by analysis of the mass spectro- metric data, and assisted by refinements in the cross- section data, internal consistency developed quickly among various other analytical results. In particular, use of the final estimate of power generation at a maximum rate of 7.4 MW to appraise both isotopic analysis and chemical analyses for uranium during the 23°U oper- ational period showed that these results were entirely consistent with those for uranium and plutonium during 233U operations. In addition, the revised esti- mates in the changes in consumption and production rates for 234U, 235U, 236U, and 238U, together with improved values for the amounts of uranium transferred to the flush salt, brought nominal and analytical values for the concentration of uranium in the fuel into good agreement as shown in Table 7.5. Isotopic composition of the uranium in the circu- lating fuel salt was recalculated on the basis that the reactor had operated at 7.4 MW and that the con- sumption and production rates of the uranium isotopes at this power level were those given in Prince’s revised estimates.?® A comparison of the resulis obtained is shown in Table 7.5. It will be noted that the greatest disparity in nominal and analytical values listed in Table 7.5 is observed for 233U, In Sect. 3.6 it is noted that further analysis of these data led to the conclusion that, of the 23#U nominally charged to the MSRE drain tanks, some 2 kg was probably not delivered. Re- computation of these results in line with that con- clusion would bring the nominal and observed values for the relative fractions of uranium isotopes, as listed in Table 7.5, into even closer agreement. References I. R. H. Guymon, MSRE Design and Operations Report, Part VIII, Operating Procedures, ORNL- TM-908, vol. Il (January 1966). 2. G. H. Burger, J. R. Engel, and C. D. Martin, Computer Manual for MSRE Operators, internal memo- randum, MSR-67-19 (March 1967). 3. R. C. Steffey, Jr., and J. R. Engel, MSR Program Semiannu. Progr. Rep. Feb, 29, 1968, ORNL-4254, p. 10. 4. C. H. Gabbard, Specific Heats of MSRE Fuel and Coolant Salts, internal memorandum, MSR-67-19 (March 1967). 5. R. B. Lindauer, Revisions to MSRE Design Data Sheets, Issue No. 9, ORNL-CF-64-6-43 (June 1964). Table 7.5. Isotopic composition of uranium in the MSRE fuel system? 111 Sample No? — 234 235, 236, 238, - 234, 235, 236, 238, (kg) {wt %) Initial d Loading - 0.325 - 147.272 147.60 - 0.282 99.7785 0.775 75.645 0.320 4.640 81.38 0.952 95.953 0.398 5,702 0.775 75.973 0.320 151.912 228.980 0.339 33,179 0,140 66.342 Run 3-F 0 228,210 Run 4-1I 0 0.711 69.660 0.294 139.288 209,953 0.339 33,179 0.140 66.362 FP 4-14 0 0. 350 53.850 g.140 86. 260 FP 4-28 0 a.350 33,249 0.14L €6.360 FP 4-33 0 0. 344 33.132 0.187 66,387 FP 4-37 0 0.343 33.188 0.138 66.351 FP 6-14 166 0.712 69.594 0.308 139.276 209,889 0.339 33,157 0.147 66.357 6.349 33,534 0.144 65,973 FP 6-19 4060 0.712 69.498 0.327 139.258 209,795 0.339 33,127 0.156 66.378 6. 344 33.445 0.152 66.060 FP 7-8 725 0.710 69.366 0.353 139.235 209.664 0.339 38.084 0.168 66.408 0. 358 33.592 0.176 85.876 FP 7-12 1012 0.709 69,250 0.376 139.213 209.548 0.338 353,047 0.179 66.435 4. 356 35,552 0.188 §5.904 FP 7-15 1047 0.709 69.236 0.379 139.211 209.535 0.338 33,043 0.181 66.438 0.347 35,161 0.177 86,315 FP 8-5 1047 0.707 68.992 0.371 138.657 208,727 0.339 33.053 0.178 66.430 0.348 33.205 0.175 66.214 ¥P 10-7 1677 0.700 68.203 0.414 137.494 206.811 0.338 32.978 0.200 66.484 0.347 33.259 0.198 66.198 FP 11-14 2884 0.698 67.732 0.509 137.409 206,348 0.338 32.824 0.247 $6.591 0.343 32.892 0.241 66,524 FP 11-39 3856 0.697 67.340 0.588 137.337 205.962 0.338 32.695 0.286 66.681 0. 349 32.973 0.278 £6.408 FP 11-44 1937 0.696 67.305 0.595 137.331 205.927 0.338 32,684 0,289 66.689 R 0.351 32,962 6.283 66,404 FP 11l-44 4107 0.704 67.997 0.611 137.365 206.677 0.341 32.900 0,29 66.463 0,348 32,987 0.287 66.378 FP 11-48 4172 0.704 67.971 0.617 137.361 206.653 0.341 32.891 0.299 66.469 0.351 33.067 G.292 86.290 FP 12-5 4513 0.700 67.568 0.627 136.784 205.679 0.340 32.851 0.305 66.504 0.354 33.133 0.296 66.217 FP 1229 5121 0.700 67.321 0.677 136.740 205.438 0.341 32.769 0.330 66.560 . 0.25L 32,963 0,320 66, 366 ¥P 12-51 5296 0.714 68.777 0.697 136.822 207.010 0,344 33.224 0.337 66.095 0.355 33.475 0.325 65.845 FP 12-58 5500 0.714 68.695 0.715 136.807 206.931 0.345 33.197 0.346 66.112 0,353 33.350 (.333 65,964 FP 14-25 6848 0.708 67.788 0.813 136,147 205.456 0.344 32.994 0.396 66.266 0.357 33.390 6.379 65.874 FP 14-46 7953 0.703 67.337 0.903 136.064 205.007 0.343 32.846 0.441 66.370 0.355 33,171 6.422 66.052 FP 14-68 8742 0.701 67.017 6.968 136.008 204.694 0.342 32.740 0.473 66.445 0.35¢ 32,945 6.450 66.251 FP 14-72 9006 0.701 66.910 0.989 135,988 204.588 0.343 32.705 0.483 66.469 0.350 33.083 0.260 66.367 ?Bold faced type indicates values computed on the basis that the maximum power generated by the MSRE was 7.40 Mwth. rates were used: 234U: - 1.665 x 16—3 o /EFPH, 235U: - 4.05150 ¢ 1071 g/EFPH, 236 (see ORNL-4449 , p. 25). Results of mass spectrometric analyses are listed in italicized type. = Equivalent full power hours. [ d Refers to total inventory in fuel system. €0.819 kg uranium added. f1.642 kg uranium added. All following items refer to fuel circuit inventory only. J. H. Shaffer memorandum to R. E, Thoma, Sept., 28, 1970. Us + 8.14 x 1072 g/EFPH, 238U: The following - 7.3691 g/EFPH 6. T. B. Douglas and Wm. H. Payne, J. Res. NBS, Ser. A, 73A,479 (1969). 7. J. W. Cooke, L. G. Alexander, and H. W, Hoffman, MSR Program Semiannu. Progr. Rep. Aug. 31, 1968, ORNL-4344, p. 100. 8. B. E. Prince, MSR Program Semiannu. Progr. Rep. Feb. 28, 1967, ORNIL-4119, p. 79; B. E. Prince, J. R. Engel, and C. H. Gabbard, Reactivity Balance Calcu- lations and Long-Term Reactivity Behavior with *3°U in the MSRE, ORNL-4674 (in preparation). 9. B. E. Prince, MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4449, p. 25. 10. D. B. Grimes, MSR Program Quart. Progr. Rep. June 30, 1958, ORNL-2551, p. 13. 11. J. A. Lane, H. G. MacPherson, and F. Maslan, Fluid Fuel Reactors, p. 656, Addison-Wesley, Reading, Mass., 1958. 12. P. R, Kasten, ORNL Reactor Division, personal communication, 1968. 13. P. R. Kasten, J. A. Lane, and L. L. Bennett, Fuel Value Studies of Plutonium and U-233, unpublished work, 1962, 14. A.M. Perry, unpublished work, 1971. 15. R. E. Thoma, Chemical Feasibility of Fueling Molten-Salt Reactors with PulF';, ORNL-TM-2256 (June 1968). 16. B. E. Prince, personal communication. 17. R. B. Lindauer, Processing of the MSRE Flush and Fuel Salts, ORNL-TM-2578 (July 1969). 18. R. E. Thoma, MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4449, p. 98. 19. W. H. Carr, W. F. Shaffer, and E. L. Nicholson, MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4449, p. 245. 20. B. E. Prince, MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4449, p. 22. 21. G. L. Ragan, MSR Program Semiannu. Progr. Rep. Aug. 31, 1970, ORNL-4622, p. 31. 22. C. H. Gabbard, internal correspondence, Mar. 22, 1971. 8. PHYSICAL PROPERTIES 8.1 General Properties Successive refinements in the design or in operational criteria for molten-salt reactors evoke repeated reap- praisals of the accuracy and precision of the available physical property data for reactor materials. As experience with the MSRE grew, most of the values for physical properties of the fuel, flush, and 112 coolant salts were reconfirmed, while for some proper- ties, the values originally employed appeared to be of questionable accuracy. The properties of greatest signif- icance in these respects were viscosity, thermal conduc- tivity, electrical conductivity, phase transition behavior, heat capacity, heat of fusion, density, expansivity, compressibility, vapor pressure, surface tension, solu- bility of the gases helium, krypton, and xenon, isocho- ric heat capacity, sonic velocity, thermal diffusivity, kinematic viscosity, and Prandtl number. Of these, the precision and accuracy are intrinsically variable and depend on methods of measurement or estimation. While the accuracy for some classes of properties, for example, liquidus-solidus temperatures, developed to be of minor concern, others, such as surface tension and heat capacity, had neither been measured nor were measurable with the precision that was desirable for reactor performance evaluations. In response to MSRE program requirements, reappraisals and refinements of physical property data were repeatedly made whenever it was feasible. The final evaluation led to values which, for the coolant and flush salt, are summarized in Table 8.1 and, for the fuel salt, are summarized in Table 8.2. 8.2 Density of Fuel and Coolant Salts Laboratory measurements of the density of molten fluoride mixtures were performed by various groups of investigators before the MSRE was operated. The precision in the results, however, was considered to be less than desirable for use in on-site calculations of reactivity balances and for appraisal of temperature effects. An effort was made, therefore, to make direct measurements with the reactor that would serve these purposes and that would afford independent measure- ment of the salt inventories. An early account gives a perspective on the success of these efforts: ! The weigh cells on all the salt drain tanks were calibrated with lead weights shortly after the equipment was installed and before the tanks and connected piping were heated. Additional data were obtained with the tanks hot during various salt-charg- ing and transfer operations. These data served both to calibrate the weighing systems and to give a measure of the density of the salts at operating temperature. Throughout the operation, salt inventories were computed from weigh-cell readings, using scale factors and tare corrections obtained from the calibration tests. Cold and hot calibration of the coolant drain-tank weigh cells gave scale factors differing by less than 0.5%. Fuel drain tank 2 (FD-2) was calibrated hot twice, with flush salt and with fuel carrier salt; scale factors were within 0.2% of each other but were about 4% higher than the original, cold calibration. The reason for this discrepancy has not been established. The coolant-salt density was measured in the reactor by three different methods; values ranged from 121.3 to 122.3 Ib/ft? at 1200°F, with an average of 121.9 Ib/ft3. When flush salt, which 113 Table 8.1 Physical properties of ithrum fluoroberyllate, Li;BeF,? Property Value Fstimated precision Viscosity n(centipoises) = 0 116 exp [3755/1(°K) ] +7 Thermal conductivity 0010 watt cm ™! °C™! *10% Electrical conductivity K=154X60x 1073 (ohm cm)! at 500°C +10% Melting pomnt? 459 1°C 0 2°C Crystal structure Hexagonal, space group R3,2=13 294, c=891A 001 A Heat capacity Liquid Cp=057calg™t °C™! 3% Sohd Cp=031+361X107*7(°C) cal g™ °C™! +3% Density Liquid pP=2214 ~ 42X 107%7TCC) g/em? 1% = 122 Ib/ft® at 650°C Sohd 0=21953 g/cm3¢ Expansivity 2 14 X 107%/°C at 600°C *10% Compressibility Br(°K)=23X 10712 exp [1 0 X 10737(°K)] cm?/dyne Factor 3 Vapor pressure log P(torrs) = 8 0 — 10,000/7(°K) Factor 50 from 500 to 700°C Surface tension v =260 — 0 127(°C) dynes/cm +30, 10% Solubihity of He, Kr, Xe T(°C) He Kr Xe 2 Factor 10 500 66 013 003 600 i06 055 017 700 151 17 067 800 201 44 20 X 1078 moles cm™ melt atm ™! Isochoric heat capacity, C,, Cy c o r TCC) calg™ calgmole™ calgatom™ — oK—l oK—I oK—l Cv 500 0489 16 , 69, 115 600 043, 15 68, 11g 700 0475 15 4 67, 124 Sonic velocity 500°C 600°C 700°C Thermal diffusivity 500°C 600°C 700°C Kinematic viscosity 500°C 600°C 700°C Prandtl number 500°C 600°C 700°C M= 3420 m/sec M= 3310 m/sec M= 3200 m/sec D =204 % 1073 cm?/sec D=214% 1073 em?/sec D=214 X 1073 cm?/sec v="T744 % 1072 cm?/sec V=434 X 1072 em?/sec P=28¢ X 1072 cm?/sec PI':356 PI=204 PI'=131 9Phystcal properties for the pure compound LiyBeF 4 approximate those for the MSRE flush and coolant salts within limits of experimental uncertainty Unless otherwise noted, property values correspond to those reported in Physical Properties of Molten Salt Reactor Fuel Coolant and Flush Salts S Cantor, ed , ORNL-TM-2316 (August 1968) bg A Romberger and ] Braunstein MSR Program Semiannu Progr Rep Feb 28 1970 ORNL-4548,p 161 €Y H Burnsand E K Gordon, Acre Cryst 20, 135 (1966) 114 Table 8.2. Physical properties of the MSRE fuel salt Property Value Estimated precision Viscosity n(centipoises) = 0 116 exp [3755/T(°K)] 7 Thermal conductivity 0010 watt cm ™! °C™1 *10% Electrical conductivity =_222+681 X 1073T(°C)a +10% Liguidus temperature 434°C 3°C Heat capacity Liquid Cp =057 cal glec! 3% Solid Cp=031+361X107*T(°C) calg™! °C™ +3% Density Liquid p=2575-513x 1074T(°C) 1% =139 9 1b/ft> at 650°C Expanswvity 214 X 1074/°C at 600°C +10% Compressibility Br(°K)=23X 10712 exp [1 0 X 1073 T(°K)] cm?/dyne Factor 3 Vapor pressure log P(torrs) = 8 0 — 10,000/T(°K) Factor 50 from 500 to 700°C Surface tension ¥ =260 — 0 12T(°C) dynes/cm +30, —10% Solubility of He, K1, Xe TCC) He Kr Xe 2 Factor 10 Isocheric heat capacity, €, Sonic velocity 500 66 013 003 600 106 0 55 017 700 151 17 067 800 201 44 20 X 1078 moles cm ™3 melt atm ™! Cy o Cp TCC) calg™ calgmole™ cal g-atom™ R ole OK—l °K'1 v 500 048 16 , 69, 11, 600 048, 159 68, 11g 700 047; 15 4 67, 12 500°C = 3420 m/sec 600°C u=3310 m/sec 700°C p= 3200 m/sec Thermal diffusivity 500°C D =209 X 1073 cm?/sec 600°C D =214 X 1073 cm?/sec 700°C D =215 X 1073 cm?/sec Kinematic viscosiy 500°C v=744 X 1072 cm?/sec 600°C v=4 34 X 1072 cm?/sec 700°C v=128¢ X 1072 cm?/sec Prandtl number 500°C Pr=35 4 600°C Pr=20 4 700°C Pr=13, 2Applicable over the temperature range 530 to 650°C The value of electrical conductivity given here was estimated by G D Robbins and 1s based on the assumption that ZrF4 and UF4 behave identically with ThF4, see G D Robbins and A S Gallanter, MSR Program Serannu Progr Rep Aug 31 1970, ORNL-4548, p 159, 1b1d , ORNL 4622, p 101 18 1dentical with coolant salt, was charged into FD 2, the amount of salt added between two level probes mdicated a density of 124 5 1b/ft3 A denuty of 1209 Ib/ft3 was computed from the pressure requured to hft salt from the dran tank to the fuel loop Weigh cell indications of flush-salt density, using the “hot” calibration factor, ranged from 123 2 to 131 5 1b/ft3 at 1200°F, the “cold” calibration factor would have given 1184 to 126 3 Ib/ft> The data on coolant and flush salt densities thus tend to support the “cold” calibration factor for FD-2 The density of the fuel carner salt, LiF-BeF, Zrk4 (6530 5 mole %) was measured as the salt was being charged to FD 2 This measured density, computed from externally measured weights and the volume between the level probes in FD-2, was 140 6 Ib/ft® at 1200°F Addition of all the uranum added through run 3 would be expected to increase the density by about 53 Ib/ft> Four measurements were made after the uranum was added, using the weigh cells and the level probes Densities based on the *“hot™ calibration of the weigh cells ranged from 149 9 to 152 2 Ib/ft®, with an average of 151 0 1b/ft3 at 1200°F With the “cold” scale factor the average was 145 1 [o/ft3, very close to the expected density Salt densities were computed on several occasions from the change 1in weigh cell readings as the fuel loop was filled In every case the computed density was less than given by other means, suggesting that a full loop volume may not have been transferred The temperature coefficients of density for the salts were computed from the change 1a salt level with loop temperature Measured values of (Ap/p)}/AT were for the coolant salt, -1 06 X 1074 ¢! (average of three measurements), for the flush, 115 x 10™* (°F)7!, and for the fuelsalt, 109 x 107% and ~115%x 107% ¢F)™ (two measurements) The bulk of the inventory data accumulated to date is on the flush salt, because more transfer and fill and dramn operations have been done with this salt Calculated inventories (using “hot” scale factors) have ranged from 1 7% below to 2 6% above the nominal or “book” mventory for no ascribable reason With continued experience 1t became evident that on-site measurements of salt masses were less and less reliable and would have little consequence in appraising reactor performance because of 1naccuracies in the welgh-cell measurements In the absence of such infor- mation the amounts of salts delivered to the circulating system or remaimng in the dramn tanks were approxi- mated from values of the density of the salt mixtures and the dimensions of the container vessels Efforts were 1mtiated to appraise and improve, if feasible, the accuracy and preciston of data pertaimning to the densities of molten fluoride mixtures Relatively few measurements of the densities of liquid salt mixtures were made 1n the development of molten- salt technology at ORNL prior to MSRE operations Experimental values for the densities of a number of Z1F 4-containing nuxtures were made as a part of the Axrcraft Nuclear Propulsion program 2 Sinular measure- ments, which employed the buoyancy principle, as did 115 the ORNL program, were made later with LiF-BeF,- UF,; muxtures by the Mound Laboratory® under con- tract work with ORNL Most density values for the nuxtures used i the MSR program, however, were estimated by the method of muxtures 2 This method was believed to have an accuracy of approximately 5% In an effort to obtain more accurate values for the salt mixtures used m the MSRE, attempts were made to develop new methods for the laboratory determination of the densities of molten salt mixtures Employing a new techmque, Sturm and Thoma# obtained density measurements for the MSRE fuel and coolant salt muxtures Measurements of the depth of salt in a cylindrical container of accurately known dumensions, which was contained in a controlled atmosphere glove box, were obtained using an electrical probe attached to a vernter caliper Contact of the probe with the melt was indicated by completion of an electrical circuit through the caliper and the melt Appropnate correc- tions for thermal expansion at a particular temperature were applied for the contaner and the probe Although some shortcomings were evident in the procedure, prncipally owing to the effects of small amounts of liquid adhering to the probe, 1t afforded a direct measurement of the densities of hiquids which were visually observable during measurement The results of the measurements were found to be in satisfactory agreement with on-site measurements of the densities of the salt stored mn the dran tanks > Laboratory efforts were therefore discontinued Concurrently, a method for esttmating densities which assumes additivity of molar volumes was devised by Cantor © With continued refinement, the method was developed to the extent that 1ts accuracy was within 5% A comparison of the densities of the MSRE fuel and coolant salts, as found using the several methods described above, 1s given in Table 8 3 On the basis of reactor physics analysis, 1t was estimated that the average fraction of the fuel charge circulated during power runs with the MSRE was 92% This value, together with a value of the volume of the crreuit computed from component dimensions, 71 3 ft*, and the mass of the circulated charge as deduced from chemical and 1sotopic analyses (Table 2 6), indi- cates that the density of the *2°U fuel at the beginning of power operations was 139 01 1b/ft®, whereas the value computed by Cantor’s method 1s 139 9 Ib/ft® In more recent application of Cantor’s method to other salt mixtures, excellent agreement has been found between estimated and observed values Such confirma- tion, along with 1ts apparent precision for MSRE fuel 116 Table 8.3. Density of MSRE salt mixtures at 650°C (1200°F) p=a—bTCC)g/cm3 Density parameters p, density Composition (mole %) Source a b g/cm3 b/ft3 x107% LiF-BeF, (66-34) Method of mixtures? 2.24 6 1.85 115.5 Mound Laboratory? 2.158 3.7 1.921 119.9 On-site estimate (see text) 1.89--2.11 118-132 Electrical probe 2.296 4.82 1.983 123.8 Molar volume estimate® 2.214 4.2 1.941 121.8 LiF-BeF,-ZrF 4 (64.7-30.1-5.2) Molar volume estimate® 2471 4.95 2.15 134.2 On-site estimate 2.25 140.6 LiF-BeF,-Z1F 4-UF, (64.7-29.38-5.1-0.82) Method of mixtures? 2.61 7 2.15 134.2 On-site estimate 2.32--2.44 145-152 Electrical probe 2.848 7.69 2.35 146.6 Molar volume estimate® 2.575 5.13 2.24 1399 4S. I. Cohen and T. N. Jones, 4 Summary of Density Measurements on Molten Fluoride Mixtures and a Correlation for Predicting Densities of Fluoride Mixtures, ORNL-1702 (July 1954, declassified Nov. 2, 1961). bMound Laboratory report MLM-1086 (see ref. 3). €S. Cantor, unpublished work, 1969. salt, permits the inference that the most accurate values of MSRE salt densities have been derived by this method. An important characteristic that easily tends to pass unnoticed from attention is that Li,BeF, is unique among the complex fluorides which have significance in molten-salt reactor technology, in that it undergoes the least volume change associated with the melting-freez- ing transition of any of the compounds encountered. This value has not been measured directly but can be estimated by assuming that the linear coefficients of thermal expansion for Li,BeF, and LiF are similar enough to be used interchangeably. If this assumption is valid, the density of Li,BeF,; is 2.064 g/em® at the melting point, about 6% less than at room temperature. The density of liquid of the stoichiometric composition at the melting point, as indicated by the density expression in Table 8.1 is 2.021 g/cm®. These two values indicate that, on melting Li, BeF,, the density is reduced only about 2.07%. It is important to recognize that the MSRE fuel, coolant, and flush salts were nearly of the composition Li, BelF,;, and to some extent the freedom that the MSRE showed from difficulties with freeze valves and from distortion of the radiator tubes under off-specifi- cation cooling conditions”? is due to the very small volume change this compound undergoes at the melting point. 8.3 Crystallization of the MSRE Fuel Laboratory studies of fluoride mixtures were carried on for some years before the MSRE was operated. In these studies, the crystallization behavior of the LiF- BeF,-ZrF4-UF, (65-29.1-5-0.9 mole %) mixture was established to approximate that of the MSRE fuel salt when it operated with *3*U fuel. Crystallization was found to follow the equilibrium crystallization se- quence described below: On cooling the liquid mixture to 434°C, crystalline Li, BeF, is formed. This phase continues to precipitate on further cooling, and at 431°C, Li, ZrF, begins to crystallize; the onset of crystallization by the tertiary phase, LiUF;, begins at 416°C. The liquid portion of the mixture decreases but is present down to tempera- tures as low as ~350°C. When completely frozen, the fuel at equilibrium should be composed of crystalline Li, BeF,, Li,ZrF4, LiUF;, and BeF, in volume frac- tions of 0.735, 0.204, 0.032, and 0.029 respectively. The usual cooling paths for LiF-BeF,-ZrF,-UF,; mix- tures of compositions similar to the MSRE fuel involve a glassing of the BeF,-rich liquids that are low melting and, thus, the last to freeze. Hence, crystals of pure BeF, are generally not expected; rather, the last liquid solidifies as glass which incorporates variable amounts of the phases listed above with BeF, . 117 PHOTO ©67747AR INCH POSITION wt % U 2.94 35.49 4.5 3.32 4.28 4.62 5.74 ~N D WM e Fig. 8.1. MSRE fuel ingot resulting from slower cooling rate. The homogeneity generally characteristic of multi- component salt mixtures in the liquid state is progress- ively destroyed as the mixture undergoes gradual crystallization. The possibility that the MSRE fuel mixture, LiF-BeF, -ZrF,-UF, (65-29.1-5.0-0.9 mole %), might, on cooling in the MSRE drain tanks, experience sufficient solid-phase fractionation to create potentially hazardous conditions was examined in laboratory-scale experiments. Some 650 g of simulated fuel salt was cooled at rates approximating that expected of the entire drain tank assembly and that expected of the fuel alone, 3.46 and 0.387°C/hr respectively. In both cases the radiative cooling geometry was controlled to simu- late as nearly as possible that expected in the drain tanks, even though it was realized that the horizontal AT profile in the cooling radioactive fuel mixture would probably be substantially different from that prevailing in the laboratory experiment. The concentrations of uranium were found to be identical at the top and bottom fractions of each of the two ingots, irrespective of cooling rate. A photograph of the ingot resulting from the slower cooling rate experiments is shown in Fig. 8.1. Chemical analyses of the salt specimens from each of the locations designated in Fig. 8.1 were obtained. For areas 1 to 7, the uranium concentrations were found to be 2.94, 3.49,4.15,3.32, 4.28, 4.62, and 5.74 wt %. In comparison with the nominal concentration of uranium in the MSRE fuel, 5.13 wt %, these experiments show a maximum increase of 23.4% in uranium concentration on very slow static cooling. A fact which accounts for the small degree of segregation of the uranium phases in these experiments is that, at the onset of crystallization of LiUFs, simultaneous crystallization of the three solid phases Li, BeF,, Liy ZrF4, and LiUF; takes place. In addition, the volume of the liquid phase is being reduced steadily, so sharply, in fact, that in this experiment as well as in similar previous ones, some of the liquid phase was apparently occluded among dendritic-like crystals of the solidified phase, a phenomenon which helps prevent compositional variation in the mixture. In the freezing of multicomponent mixtures, maxi- mum segregation of crystalline phases takes place under equilibrium cooling conditions. The segregation repre- sented by the results obtained in the fractionation experiments described here represents, in a practical way, the nearest approach to equilibrium cooling that the MSRE fuel salt may experience in a single crystalli- zation sequence. The crystallization behavior described above pertains exclusively to the fuel-salt mixtures used in the MSRE and is, of course, quite unlike that for fuel mixtures 118 proposed for use in a molten-salt breeder reactor. Crystallization equilibria similar to those for the breeder fuel are described elsewhere by Thoma and Ricci.® References 1. P. N. Haubenreich et al., MSR Program Semiannu. Progr. Rep. Aug. 31, 1965, ORNL-3872, p. 30; see also Sect. 2.4.2 of this report. 2. 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, ORNL-1702 (July 19, 1954, declassified Nov. 2, 1961). 3. Density and Viscosity of Fused Mixtures of Lithium, Beryllium and Uranium Fluorides, Mound Laboratory report MLM-1086 (December 1956). 4. B. J. Sturm and R. E. Thoma, Reactor Chem. Div. Annu. Progr. Rep. Jan. 31, 1965, ORNL-3789, p. 33. 5. P. N. Haubenreich, private communication, Aug. 16, 1965. 6. S. Cantor, Reactor Chem. Div. Annu. Progr. Rep. Jan. 31, 1962, ORNL-3262, p. 38. 7. T. L. Hudson, C. H. Gabbard, and D. M. Richard- son, MSR Program Semiannu. Progr. Rep. Aug. 31, 1966, ORNL-4119, p. 41. 8. R. E. Thoma and J. E. Ricci, Fractional Crystalli- zation in the System LiF-BeF,-ThF,;, ORNL-TM-2596 (July 1969). 9. INTERACTIONS OF FUEL SALT WITH MODERATOR GRAPHITE AND SURVEILLANCE SAMPLE MATERIALS A new method of in-situ analysis for lithium, beryl- lium, and fluorine was invented by Macklin, Gibbons, and Handley' in response to our appeal for their assistance in examination of graphite specimens re- moved from the MSRE core. The method they devised employed proton bombardment of target nuclides and measurement of the yields of neutrons and gamma rays produced. It was found to be applicable for measure- ment of the concentration of target nuclides in graphite at the few-parts-per-million level as a means to deter- mine the extent to which MSRE fuel salt components penetrated into the graphite under radiation. This method has the advantage of applicability to matrix- dispersed samples and in the presence of considerable radioactivity from fission products, features which make it well suited to the examination of MSRE graphite. The results obtained by these workers were reported in detail previously.?*® There remain several puzzling aspects of their findings, and because of this the results are reviewed here. Data were obtained from three samples, (1) a control sample (Y-S5 from CGB bar 635) exposed to nonradio- active salt, (2) sample Y-7, removed from the reactor core after 33,400 MWhr of exposure in May 1967, and (3) sample X 13, removed March 25, 1968, after 66,637 MWhr of exposure. Specimens were moved across a beam of 2.06-MeV protons collimated through a slit of 0.0075 cm width at the ORNL 3-MV Van de Graaff accelerator. Measure- ment of the resulting prompt gamma rays from YE(p,oy)' °O showed that fluorine varied in sample X-13 from 350 ppm near the surface to 60 ppm at the center (Fig. 9.1). The observed ratio of fluorine to lithium was near that characteristic of the MSRE fuel (Fig. 9.2). However, as shown in Figs. 9.1 and 9.3, the lithium and fluorine did not show a simple dependence on depth. Results of the examination of sample X-13 suggested that much of the Li and F came from bulk intrusion, a finding which differed from that resulting from the previous examination of sample Y-7 (Fig. 9.4), where 119 the Li/F ratio became increasingly higher at greater distances from the surface. Equally puzzling is that comparisons with the results of the analysis for 2?°U by Kirslis and Blankenship®* (see Figs. 9.1 and 9.3) showed that the relative concentrations of F and 2°%U became steadily divergent with penetration depth and thus seemed to rule out bulk salt intrusion as a mechanism. Further, in the absence of radiation, a control specimen showed less penetration of both salt and uranium by a factor of 100. The possibility was considered that introduction of a fuel aerosol from the gas phase, as was suggested to rationalize the intrusion of certain fission products,®* was responsible for fuel having penetrated the graphite voids as an aerosol. The cause of this phenomenon is still not resolved. The uranium profiles shown in Figs. 9.1 and 9.3 were derived in the investigation by Kirslis and Blankenship, in which it was found, from delayed neutron activation analysis, that the uranium concentration profile ranges from values as high as 100 ppm at the surface to 6.5 ppm at 50 mils depth, but “in most cases, the total range of variation of the **5U concentration in 50 mils penetration was less than two orders of magnitude. This moderate slope indicates a higher mobility for uranium in graphite than for most fission products [but] ... represents only 1 g of 23U per 1000 kg of graphite.”* ORNL-DWG 68-14530 Ak T T R N O FIRST SURFACE TO CENTER - - & SECOND SURFACE TO CENTER 500 e — @ 2350 IN SIMILAR SPECIMEN e ] Pl s —raA ] \\ | i - ] 1--~A—~CL—1 N }fi ] e X‘fi\ “\ A A ‘;}3.\\ \ V ] N /B s | o Ty 1 P +od /o\ jfi ! \ k"\ c — e s ] A, O -1 - 4 g. 100 - 17 N ] | By a. ab U F o | A b Y ! = - A1t T il v:oqw | \ ‘I’ ko) _Q‘59|: S 50 ] H ¥ ~ W Y ~d PO N { 5&\ e - L | I M 235U 20 N IR Ly \ @ \\ 0 —— |——1+1 - NS - I e ] i ~L _ i g N 5 i i 1 10 100 DISTANCE FROM SURFACE {mils) Fig. 9.1. Fluorine concentration as a function of distance from the surface. 120 ORNL-DWG 68—14532 50 Al o FIRST SURFACE TO CENTER @ SECOND SURFACE TO CENTER|] 20 g [ C o 10 Q = RATIC FOR o i ; o 4 FUEL SALT ¥ . = e | @ ®goo0 | ® ) E 5 @ O [o) : L ) c08 - o » RATIO FOR LiF 1 1 2 5 i0 20 50 100 200 500 1000 DISTANCE FROM SURFACE ( mils) Fig. 9.2. Mass concentration ratio, F/Li, vs depth. ORNL-DWG 68-14534 100 o —— 1000 o FIRST SURFACE TO CENTER ™ & SECOND SURFACE TO CENTER 50 \\\ e 2354 | SIMILAR SPECIMEN 500 < S B e ;jl \b 20 R AR 200 H )—4\& ! J 9 'CrC\l ~ J‘d 8‘5 / g}&%@g - RO ~ 7, - £ ? N Li £ a {0 100 o e o b N 2 < 3 235U»\ N 5 - 50 Y H-—i A\A‘A-A ‘,4‘ o™ \,g‘\ 2 Hs - 20 Bapi 1 10 1 10 400 1000 DISTANCE FROM SURFACE (mils) Fig. 9.3. Lithium concentration as a function of distance from the surface. Although the results of the proton reaction analysis described by Macklin and co-workers do not seem to permit quantitative generalization, they do suggest several points of significance to further development of molten-salt reactor technology: 1. The penetration of the graphite moderator by lithium and fluorine appears to be real, if not massive. Only samples of CGB graphite were ex- amined in the experiments performed by Macklin et al. Whether similar penetrations of more dense or coated graphites can be expected for improved MSBR graphites is not estimable from the current results. Effects of salt-gas-graphite interfacial properties on transport of lithium and fluorine through films is unknown at present. . The current data were derived from a fuel salt with a composition, LiF-BeF,-ZrF,-UF, (65-29.2-5.0-0.8 mole %), whose effect on transport conceivably may differ from that for the MSBR. Whether the trans- 121 QRNL-DWG 68~ 14533 1000 — | — ] sl e [ H 500 — A b — Al 1 I e I \‘T‘\ J_I, Lj_l W‘ — z ™ 200 — N\ | | SN ! ’ = L | SAMPLE X~13 100 F— J . DN - [ - — \\, _— —_ — L S 3\‘\ _ RN M~ ] —- 50 _ N I E 1 1IN L ] 8 N 2 — \ 1 . - _ ful | 20 |— - e SAMPLE Y -7 s\\ 10 —a N - T N ] [ | _S\ i OO S | 5k - N T . . I 4:;,; N ~ | : 2 | S — ? I 3 S— — - | ; g t ‘ w \ | | | 1 1 2 5 0w 20 50 100 200 500 DISTANCE FROM SURFACE (mils} Fig. 9.4. Comparison of fluorine concentrations in samples Y-7 and X-13, a smooth line having been drawn through the data points. port would be markedly affected by the large heavy-metal concentration of the MSBR fuel, con- taining a total of 12 mole % ThF4 + UF,, is of justifiable curiosity. The results obtained indicated with 2?°U fuel salt that there are a number of parameters whose quanti- tative relationship to the transport of salt species to moderator graphites should be examined by further examination of graphite specimens, especially those removed from the MSRE core after completion of the 2337 experiments. Several notable differences between 2337 and *?°U operations existed, including for 233U operation a uranium concentration less than 20% of that with 22°U, a continuously higher void fraction in the fuel salt, and variable differences in the gas-salt interfacial tension. The latter two of these factors were regarded as possibly conducive to enhanced transfer of materials to the graphite moderator and prompted further examination of graphite specimens from the MSRE. In postoperational examinations of a graphite stringer from the MSRE core vessel, Kirslis and Blankenship® obtained spectrochemical and delayed neutron analyses of the graphite milled from the surface of the stringer. Analyses by the proton bombardment method were not performed. The analytical resulis that were obtained were in partial agreement with the earlier results of Macklin and co-workers and indicated bulk penetration of the fuel salt, probably via cracks in the graphite, to a depth of ~2 mils. In contrast, however, they did not indicate penetration to greater depths. Although the results of the two groups of analyses are in good agreement with respect to penetration of the outermost layers, Kirslis and Blankenship did not find anomalous divergence in the relative concentration of any of the components, nor did they find that uranium pene- tration had occurred to greater depths. The absence of detectable amounts of uranium in the graphite at depths where uranium had been detected with samples from ?3* U tests tends to support a model of bulk salt penetration, since the 2?*U concentration of the fuel was less than 20% of that of the >*3 U fuel. Applications of new and increasingly sensitive methods of analysis, such as the proton bombardment method, in molten-salt research and development programs are significant to the continued development of the technology. Continued refinement and adapta- tion of this new method are highly desirable within the framework of future research programs for possible application to MSBR development. References 1. R. L. Macklin, J. H. Gibbons, and T. H. Handley, Proton Reaction Analysis for Lithium and Fluorine in Graphite Using a Slit Scanning Technigue, ORNL- TM-2238 (July 1968). 2. R. L. Macklin et al.,, MSR Program Semiannu. Progr. Rep. Feb. 9, 1968, ORNL-4254, p. 119. 3. R. L. Macklin et al., MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4344, p. 146. 4. S. 8. Kirslis and F. F. Blankenship, MSR Program Semiannu. Progr. Rep. Aug. 31, 1969, ORNL-4344, p. 115. 5. 8. S. Kirslis and F. F. Blankenship, MSR Program Semiannu. Progr. Rep. Feb. 28, 1971, ORNL-4676, p. 73. 10. CHEMICAL SURVEILLANCE OF AUXILIARY FLUID SYSTEMS 10.1 Water Systems Potable water from the ORNL distribution system was supplied to the MSRE for a variety of uses as described in the MSRE Design and Operations Report:! After passing through a backflow preventer, the water is used in the liquid waste system, in the vapor-condensing system, for general cleanup of equipment, as makeup for the cooling tower water system, and for cooling of the charcoal beds. Two 520-gpm centrifugal pumps are provided for circulating cooling tower water, which is cooled by a two-fan induced draft cooling tower. The cooling tower water is used for air compressors, air conditioners, in the chemical plant, for the lube-oil systems, in the charcoal beds, and for condensing steam from the drain tank steam domes. (Process water can also be used for this.) Cooling tower water is also used in a shell-and-tube heat exchanger to provide cooling for the treated water system. Two 230-gpm centrifugal pumps circulate treated water in a closed loop to cool in cell components. Makeup water is supplied by condensing building steam in a shell and tube heat exchanger using cooling tower water as the coolant, Treated water is also used to fill the Nuclear Instrument peneiration. This water is continuously recirculated through a closed loop by a 5-gpm pump to maintain uniformity of the water condition through- out the penetration. In order to minimize corrosion, potassium tetraborate and potassium nitrite were added to the water supply and maintained at concentrations of 500 and 1500 ppm, respectively. Steam condensate is also used io supply water to the feedwater tanks. This untreated water is used in the drain tank bayonets to remove decay heat from the reactor fuel afier the reactor has been drained. The several water systems at the MSRE were sampled periodically and analyzed to determine corrosion rates, buildup of contaminants, loss of corrosion-inhibiting chemicals, etc. Water samples were submitted regularly for chemical analysis in the General Analysis Labo- ratory. At more frequent intervals, on-site tests were performed to give continuity to the control data and to reduce dependency on analyses from the laboratory. Details of sampling procedures, methods of additions of corrosion inhibitors, and of on-ite testing procedures are given in the MSRE Design and Operations Report, Sect. 6-1.1 10.1.1 Cooling tower water. Samples of the cooling tower water were tested on-site daily to check the concentration of its chromate ion, which served as a corrosion inhibitor, pH, and hardness. Samples were submitted on a weekly basis for laboratory analysis, where more complete analyses were performed. In use, the cooling tower water supply was maintained in the pH range 7 to 8, concentration of chromate ion at 30 to 50 ppm, hardness <250 ppm {as CaCO3), and Fe <0.03 ppm. The results of laboratory analyses with cooling tower water samples are listed in Table 10.1. 10.1.2 Treated water supply. A supply of treated water was circulated within a closed system to remove heat from the thermal shield and other equipment in the reactor and drain cells. A program of water system surveillance was incorporated into the procedures em- ployed at the MSRE to minimize the probability that operational difficulties would arise from this auxiliary system. Since standardized methods for control of the water chemistry of auxiliary systems such as those which existed in the MSRE were fairly well stand- 122 ardized, it was anticipated that surveillance of the water supply would require only cursory attention. However, when power was first generated with the MSRE, it was found that the chemicals used in the treated water system had become activated. The source of the activity was identified as *2K, produced from the potassium nitrite—potassium tetraborate mixture used to inhibit corrosion in the system. Extrapolation of the radiation levels observed at low power indicated that if the reactor were operated for extended periods at full power, equilibrium radiation levels of about 400 mR/hr would prevail in the heat exchanger in the diesel equipment room and in the water control room. As the alternative, the potassium-containing inhibitors were discarded and replaced with an analogous lithium- based mixture, highly enriched in the 7Li isotope to minimize tritium production. Lithium nitrite was pre- pared commercially by ion exchange from potassium nitrite and lithium hydroxide for this use. After the 4000-gal ireated-water system was diluted with de- mineralized water to reduce the potassium from 800 to 3 ppm, the desired inhibitor concentration was attained by adding " Li nitrite, boric acid, and 7 Li hydroxide. When the reactor was next operated at power, in run 5, activation again occurred, this time caused by the presence of sodium, ~1 ppm of which was present in the demineralized water from the ORNL facility. Condensate was produced at the MSRE with less than 0.1 ppm of sodium and used to dilute the sodium in the treated-water system to 0.3 ppm. The concentrations of sodium and ®Li were thereafter reduced to sufficiently low concentrations that shielding and zero-leakage containment of the water system were not required. Criteria were established for the treated water after replacement of the inhibitor mixture to meet the following specifications: pH: 7.0 to 9.0, Na* <3 ppm, NO, = 760 to 860 ppm, B> 57 ppm, K™ <8 ppm, AAl = 0 ppm, Fe = 0 ppm, tritium < 2.2 X 10° dis min™ cm™3, AZ hardness (as CaCO;) = 0 ppm. Chemical analyses of treated-water samples were obtained at frequent intervals in order to maintain these specifications and as guides for adjustments of the concentrations of chemicals in the treated-water system supply. On-site tests were also performed with greater frequency than laboratory analyses for control of operations, with supplementary analyses obtained from the Analytical Chemistry laboratory at approximately biweekly intervals. The results of the chemical analyses obtained for treated-water samples obtained from the circulating system are listed in Table 10.2. Those obtained from the nuclear instrument penetration are listed in Table 10.3. N The results listed in Tables 10.2 and 10.3 do not warrant detailed comment or interpretation because they were used primarily as guides for controlling the concentrations of the corrosion inhibitors. It is evident from the results listed that they served satisfactorily for this purpose, and that the treated-water systems were essentially free of operational difficulties. 10.1.3 Vapor condensing system. Among the water supplies which were subjected to routine chemical 123 surveillance was included in the reservoir of the vapor condensing system. This system was incorporated into the MSRE for service only under the accident con- ditions described in the MSRE Design and Operations Report, Part 1.2 1If an accident occurred in which hot fuel salt and the water used to cool equipment inside the cells became mixed, it would be necessary to contain within the MSRE the steam generated by the accident. A vapor condensing system, shown sche- Table 10.1. Results of chemical analysis of MSRE cooling tower water CTW Date Total CrOh_ pH Te CTW Date Total cro, pH Fe Sample Hardness Sample Hardpess No. (ppm) No. (ppm) 8 6/18/65 214 39 8.53 <2, 456 12/2/66 178 12 8.00 < 0.3 9 6/25765 185 13 8.63 <2.0 458 1274766 251 68 8.00 <1.,0 10 7/2/65 284 39 8.70 <2.0 467 12/12/66 197 4 0.3 11 7/9/65 133 141 8.10 - 473 12/18/66 176 14 - <2.0 12 7/23/65 105 21 9,58 <1.0 482 12/27/66 187 28 817 0.3 13 7/30/65 105 30 7.97 <1.0 494 1/8/67 178 16 8.37 0.5 14 8/6/65 185 7z 7.80 <1.0 502 1/15/67 175 18 - 0.3 15 8/23/65 194 - 8.51 <0.3 512 1/30/67 171 9 8.10 0.3 16 9/4/65 248 - 6.90 4.0 516 2/5/67 165 11 8.78 0.3 18 9/17/65 3.0 37 8,77 1.0 524 2/13/67 171 - - 0.3 26 9/24/65 223 25 8.60 - 532 2721767 161 26 8.02 0.3 32 10/8/65 111 44 8.00 0.5 538 2/27/67 175 17 8.10 <0.3 33 10/15/65 30 32 8.10 0.5 546 3/6/67 231 27 8.77 0.3 48 11/10/65 121 14 8.08 0.3 559 3/19/67 223 79 8,36 0.3 55 11/17/65 141 15 8.22 0.2 575 413167 - 23 8.57 0.3 67 11/29/65 161 16 7.0 <0.2 581 418767 342 36 8,60 0.3 79 12/10/65 129 29 7.37 - 588 4/18/67 241, - 8.30 <1.0 83 12/15/65 137 21 8.0 <0.3 596 4/24/67 239 29 8.50 <1.0 97 12/28/65 143 - 7.80 <0.3 603 5/2/67 235 26 8.47 <0.3 105 1/5/66 - - 8.32 Increasing efforts will undoubtedly be devoted to this activity in the future. 2. Corrosion detection. The extent of generalized corrosion in the MSRE was estimated routinely from changes in the concentration of chromium found in the salt samples removed from the reactor. Results of examinations of surveillance specimens removed occa- sionally from the reactor core and from postoperational examinations of the alloy removed from the heat exchanger confirmed the general validity of these estimates. The use of Cr?* concentration as a corrosion indicator thus continues to be uniquely attractive. To verify corrosion-free operation of a large reactor from analyses of numerous samples removed from the reactor would probably be limited by the slow feedback as mentioned above. An alternate means would be far more desirable. One generalization which emerges from MSRE experience is that during periods when the relative concentration of [U**]/[ZU] in the fuel salt was 20.5%, the system appeared to be protected against corrosion. This observation, along with a consideration of the expense, inconvenience, and irrelevance of individual results of chemical analyses, suggests that a more suitable means for establishing that corrosion-free conditions prevail would be to ensure that at all times the proper redox potential existed in the salt stream. 141 As part of the MSRE experiment, evidence appeared to indicate the possibility that the disposition of niobium in the fuel salt could be exploited as a means of monitoring corrosion in molten-salt reactors on a continuous basis. The lack of success we have experi- enced in attempting to determine the concentration of trivalent uranium in the fuel salt when the total concentration of uranium was <0.5 mole % suggests that there is a clear incentive to develop methods of detecting corrosion such as by niobium monitoring by gamma spectrometry, which can provide on-line con- tinuous monitoring of the chemical potential of the fuel salt. As noted in Chap. 6, several aspects of corrosion chemistry in molten-salt reactor systems were left unresolved at termination of MSRE operations. The outstanding example of these was that corrosion attack during ??°U operations might have been anticipated to be some threefold greater than was observed. Further, examinations of metal surveillance specimens removed from the heat exchanger indicated that mass transfer of metal from hot to cold zones was so little as to be undetectable. Temperature differences in the salt cir- cuits were modest, ~50°F, whereas in future molten- salt reactors projected differences run to about 200°F. The absence of base-line data for mass transfer that results from these examinations obviated the possibility of projecting mass-transfer coefficients for larger reac- tors. Generalized corrosion was deduced from chemical measurements of the chromium concentration in fuel- salt samples. Current developments in the attempts to improve the radiation-induced loss of ductility that Hastelloy N experiences after use in high radiation fluxes include the possible inclusion of small amounts of titanium, hafnium, or zirconium in the alloy.* All of these metals are more chemically active than chromium, the most chemically active constituent of the unmodified alloy; their inclusion as constituents in significant concentra- tions may affect the chemical methods for monitoring corrosion on a dynamic basis. 3. Oils and hydrocarbons in the fuel system. The extent of oil leakage into the fuel system must be known accurately and ascertained on a regular basis because of the operational implications of such leakage and because of the chemical effects induced by the presence of hydrocarbon degradation products in the system. Hydrogen, evolved in the thermal degradation of hydrocarbons, will unquestionably be involved in the distribution of tritium in the reactor and will control fission product chemistry, perhaps even favorably if, for example, it enhances the retention of iodine in the system as iodide or tellurium as the element. Surveil- lance of the off-gas by means of a gas spectrometer seems to be very desirable; it may possibly be an important requisite for continuous operation. If the gas analysis by mass spectrometry proves to be adaptable for determination of a variety of species, it could provide the most precise source of control data for the reactor. 4. Miscellaneous. One of the most useful kinds of analysis in the MSRE was that provided by mass spectrometric measurements of the isotopic composi- tion of uranium and plutonium. Occasional removal of salt samples from reactors built in the future, princi- pally for mass spectrometric determination of inven- tory, burnup, etc., could also be used for confirmation by general analyses that the on-line instrumentation was functioning correctly. Little advantage would accrue from any routine efforts to determine chemically whether there were any chronic leaks from fuel to coolant or in the reverse direction, for, in either event, nuclear data would provide operational control. Similarly, laboratory deter- mination of the condition of auxiliary fluids, such as cooling tower water, oil from pump reservoirs, etc., should probably not be done at all. Experience with the MSRE has shown that there will be little difficulty in maintaining these fluids in condition so they will meet physical and chemical criteria over long periods of time. Information pertaining to their condition can be ac- quired simply and quickly and should be obtained at suitable predetermined intervals by in-line automated monitoring equipment. In summary, it becomes evident that operational adjustments for future molten-salt reactors should 142 originate from chemical controls provided almost ex- clusively from on-line instrumentation. It is imperative that such instrumentation be stringently restricted to the minimum for safety, because there should be litile reason to use the information derived for frequent adjustments of the salt systems. These reduce to (1) a continuous method for determining the identity of species in and composition of the gas above the fuel salt; in all likelihood this will entail the development of an automated mass spectrometer; (2) a method of establishing the redox potential of the fuel salt to ensure corrosion-free operation; for this function, fur- ther development of a niobium monitor by on-line gamma spectrometry will be required; (3) verification that bismuth does not enter the fuel stream via the incoming stream from the chemical reprocessing plant; (4) development of a highly sensitive and accurate method for continuous or frequent on-line determina- tion of oxide concentration in the fuel salt. References 1. R. E. Thoma and G. M. Hebert, Coolant Salt for a Molten Salt Breeder Reactor, US. patent No. 3,448,054, June 3, 1969. 2. W. R. Grimes, correspondence to Milton Shaw, “Assessment of Need for Oxygen Getter in Molten Salt Breeder Reactor Fuel,” February 3, 1971. 3. A. S. Meyer and J. M. Dale, personal communica- tion. 4. H. E. McCoy et al., Nucl. Appl. Technol. 8, 156 (1970). ORNL-4658 UC-80 — Reactor Technology INTERNAL DISTRIBUTION 1. J. L. Anderson 41. D. E. Ferguson 82. Dunlap Scott 2. R.F. Apple 42. L. M. Ferris 83. J. L. Scott 3. C.F. Baes 43. A.P.Fraas 84. H. E. Seagren 4. C. E. Bamberger 44. J. H. Frye, Jr. 85. J. H. Shaffer 5. C.J. Barton 45. W. K. Furlong 86. M. J. Skinner 6. J. B. Bates 46. C. H. Gabbard 87. A.N. Smith 7. H. F. Bauman 47. R. B. Gallaher 88. G.P. Smith 8. S. E. Beall 48. L. O. Gilpatrick 89. A. H. Snell 9. M. J. Bell 49, W. R, Grimes 90. Din Sood 10. M. Bender 50. A. G. Grindell 91. R. A, Strehlow 11. E. S. Bettis 51. R. H. Guymon 92. D. A. Sundberg 12. D.S. Billington 52. P. H. Harley 93. J. R. Tallackson 13. F. F. Blankenship 53. P.N. Haubenreich 94. E.H. Taylor 14. E. G. Bohlmann 54. R. F. Hibbs 95--97. R.E. Thoma 15. G. E. Boyd 55. G. H. Jenks 98. L. M. Toth 16. J. Braunstein 56. E.M. King 99. D. B. Trauger 17. M. A. Bredig 57. A.P.Malinauskas 100. G. C. Warlick 18. R. B. Briggs 58. H. E. McCoy 101. C. F. Weaver 19. H. R. Bronstein 59. H. F. McDuffie 102. A. M. Weinberg 20. G. D. Brunton 60. H. A. McLain 103. J. R. Weir 21. S. Cantor 61. L. E. McNeese 104. M. E. Whatley 22. D. W. Cardwell 62. J. R. McWherter 105. J. C. White 23. R.S. Carlsmith 63. A.S. Meyer 106. R. P. Wichner 24, W_ L. Carter 64. R. L. Moore 107, L. V. Wilson 25. E. L. Compere 65. E. L. Nicholson 108. H. C. Young 26. W. H. Cook 66. L. C. Oakes 109. 1. P. Young 27. J.W.Cooke 67. R. B. Parker 110. J. W. Cobble (consultant) 28. L. T.Corbin 68. A.M. Perry 111. P. H. Emmett (consultant) 29. 1. L. Crowley 69. H. B. Piper 112. H. Insley (consultant) 30. F. L. Culler 70. B. E. Prince 113. E. A. Mason (consultant) 31. D. R.Cuneo 71. A.S. Quist 114. R. F. Newton (consultant) 32. J.M. Dale 72. G. L. Ragan 115. J. E. Ricci (consultant) 33. J.H. DeVan 73. J.D. Redman 116. C. H. Secoy (consultant) 34. J. R. Distefano 74. D. M. Richardson 117. H. Steinfink (consultant) 35. S. 1. Ditto 75. G. D. Robbins 118. L. R. Zumwalt (consultant) 36. F. A.Doss 76. R. (. Robertson 119-121. Central Research Library 37. A.S. Dworkin 77—78. M. W. Rosenthal 122. ORNL — Y-12 Technical Library 38. W.P. Eatherly 79. R. G. Ross Document Reference Section 39. 1. R. Engel 80. J.P. Sanders 123—157. Laboratory Records Department 40. R. B. Evans 111 81. H. C. Savage 158. Laboratory Records, ORNL R.C. 143 144 EXTERNAL DISTRIBUTION 159—160. MSBR Program Manager, AEC, Washington, D.C. 161. Merson Booth, AEC, Washington, D.C. 162. Paul Cohen, Westinghouse Elect. Corp., P.O. Box 158, Madison, Pa. 15663 163. D. F. Cope, RDT Site Office, ORNL 164. J. D. Corbett, Jowa State University, Ames, lowa 50010 165. J. W. Crawford, AEC, Washington, D.C. 166. F. E. Dearing, RDT Site Office (ORNL) 167. D. R. deBoisblanc, Ebasco Services, Inc., 2 Rector St., New York 10006 168. C. B. Deering, Black & Veatch, P.O. Box 8405, Kansas City, Mo. 64114 169. A. R. DeGrazia, AEC, Washington, D.C. 170. S. G. English, AEC,Washington, D.C, 171. J. J. Ferritto, Poco Graphite, P.O. Box 2121, Decatur, TX 76234 172. T. A. Flynn, Jr., Ebasco Services, Inc., 2 Rector St., N.Y., N.Y. 10006 173. E. E. Fowler, AEC, Washington, D.C. 174. J. E. Fox, AEC, Washington, D.C. 175. Norton Habermann, RDT, USAEC, Washington, D.C. 20545 176. A. Houtzeel, TNO, 176 Second Ave., Waltham, MA 02154 177. G. M. Kavanagh, AEC, Washington, D.C. 178. H. H. Kellogg, Henry Krumb School of Mines, Columbia U., N.Y., N.Y. 10027 179. M. Klein, AEC, Washington, D.C. 180. S.J. Lanes, AEC, Washington, D.C. 181. Kermit Laughon, AEC, RDT Site Office (ORNL) . 182. J. M. Longo, Esso Research & Eng., P.O. Box 45, Linden, N.J. 07036 183. T.W. Mcintosh, AEC, Washington, D.C. 20545 184. J. Neff, AEC, Washington, D.C. 185. R. E. Pahler, AEC, Washington, D.C. 186. A.J. Pressesky, AEC, Washington, D.C. 187. M. V. Ramaniah, Bhabha Atomic Research Centre, Radiological Labs., Trombay, Bombay-85, AS, India 188. David M. Richman, AEC, Washington, D.C. 189. H. M. Roth, AEC-ORO, Oak Ridge, Tn. 190. R. M. Scroggins, AEC, Washington, D.C. 191. M. Shaw, USAEC, Washington, D.C. 192. T. G. Schlieter, AEC, Washington, D.C. 193. J. M. Simmons, AEC, Washington, D.C. 194. S. A, Szawlewicz, AEC, Washington, D.C. 195. A. R. Van Dyken, AEC, Washington, D.C. 196. N. Srinivasan, Bhabha Atomic Research Centre, Trombay, Bombay 74, India 197. R.C. Steffy, Jr., TVA, 540 Mkt. St., Chattanooga, Tn. 37401 198. A. E.Swanson, Black & Veatch, P.O. Box 8405, 1500 Meadowlake, K.C., Mo. 199. J. A. Swartout, UCC, New York, N.Y. 10017 200. B. L. Tarmy, Esso Research & Engr. Co., P.O. Box 101, Florham Pk., NJ. 07923 201. J. R. Trinko, Ebasco Services, Inc., 2 Rector St., N.Y., N.Y. 10006 202. C. E. Larson, Commissioner, AEC, Washington, D.C. 203. W. W. Grigorieff, Oak Ridge Associated Universities - 204. Leo Brewer, Lawrence Radiation Laboratory 205. R. C. Vogel, Argonne National Laboratory 206. S. Spaepen, Head of Chemical Technology, SCK-CEN, MOL, Belgium - 207. V. K. Moorthy, Metallurgy Div., BARC, Bombay 83, India . 208 209 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222--224. 225--226. 227-231. 232. 233. 234-452. 145 Penneman, Los Alamos Scientific Laboratory Steunenberg, Argonne National Laboratory Reser, American Ceramic Soc., 4055 N. High St., Columbus, Ohio 43214 Sidney Langer, Gulf General Atomic, San Diego, California C. W. Keenan, Dept. of Chem., University of Tennessee, Knoxville Gleb Mamantov, Dept. of Chem., University of Tennessee, Knoxville Rustum Roy, Materials Res. Lab., Penn State Univ., University Park, Pa. Minoo D. Karkhanavala, Chem. Div., BARC, Bombay, India W. Danner, Max-Planck-Institut Fur Plasmaphysik, 8046 Garching Bei, Munchen, West Germany David Elias, AEC, Washington, D.C. Ronald Feit, AEC, Washington, D.C. P. A. Halpine, AEC, Washington, D.C. W. H. Hannum, AEC, Washington, D.C. R. Jones, AEC, Washington, D.C. Director, Division of Reactor Licensing (DRL), AEC, Washington, D.C. Director, Division of Reactor Standards (DRS), AEC, Washington, D.C. Executive Secretary, Advisory Committee on Reactor Safeguards, AEC, Washington, D.C. Laboratory and University Division, AEC, ORO Patent Oftice, AEC, ORO Given distribution as shown in TID4500 under Reactor Technology category (25 copies — NTIS) . RLA. . R K. M. K.