. ) B Lo . - . , L o oy - - T e ¥ J, e e . vk et e e e L e TG . T TR A el ‘ v e -, c Sl : Wb L e P D i e ["S:“L‘ - ! ) A i e e < - _f‘.,‘r‘._ ., mfig} CRHD MG R i‘ Y e et /w h. :"_;.fi:"(—." - ., . Sk TN ‘ & ’) ' v & 4 3 T ,.— % o ¥ T e ,s n » 5, b Tl - L~ A - . = Joa T o - g g b R o - - S k- s . ’?‘f It PERET 'g o e AR e R ' CoAT e F ,"v‘:i if'v' . L - gi;:‘"'; w5 s L2 . 3‘ FeL .o t s L . te 3T i . Y s f Dt e b J e o v 55T el - Fr e o - - - r o - . r .k UNGLASSIFIED i MARIETT A ENERGY 5 STEMS LIBRARIES SRR w260 o 47 3 445 03klzap 2 UC81- Reactors—~ MOLTEN-SALT REACTOR PROGRAM - QUARTERLY PROGRESS REPORT FOR PERIOD ENDING OCTOBER 31, 1957 und; .. fl\é hbrary OAK RIDGE NATIONAL I.ABORATORY OPERATED BY UNION CARBIDE NUCLEAR COMPANY Division of Union Carbide Corporation POST OFFICE BOX X * OAK RIDGE, TENNESSEE UNCLASSIFIED $1.50 Printed in USA. Price' ___ — __ cents, Available from the Office of Technical Services U, S5, Department of Commerce Washington 25, D. C. LEGAL NOTICE This report was prepcred as an account of Government sponsored work, Neither the United Stotes, nor the Commission, nor any person acting on behalf of the Commission: A, Maokes any warrenty or representation, express or implied, with respect to the accurccy, completeness, or usefulness of the information contained in this report, or that the use of ony information, apparatus, method, or process disclesed in this report may not infringe privately owned rights; or B. Assumes any liabilities with respect to the use of, or for dumages resulting from the use of any information, apparatus, methoed, or process disclosed in this report. As used in the above, ‘‘person otting on behalf of the Commission*’ includes any employee or contractor of the Commission to the extent that such employee or contracter prepares, handles or distributes, or provides occess to, any information pursuant to his employment or contract with the Commission, UNGLASSIFIED ORNL-2431 UC-81 =~ Reactors~Power Contract No. W-7405-eng-26 MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT For Period Ending October 31, 1957 H. G. MacPherson, Program Director DATE ISSUED FEB7 1958 OAK RIDGE NATIONAL LABORATORY Operated by UNION CARBIDE NUCLEAR COMPANY Division of Union Carbide Corporation Post Office Box X Ock Ridge, Tennessee INCLASSIFIE Hllllllflllflllfllflflfllwtl]lflfillllzflliljfllllflfil@lfllle vl . VNS RE LD 37. 38. 39. 40, 41, fipi>>.’£flmi&—)>§-ITI>CJS—EI"%'HE—EOOSOU?UmOflme'nCImF" M. F. B. M. . G. Alexander . S. Bettis . S, Billington . F. Blankenship . P. Blizard . L. Boch . J. Borkowski . E. Boyd . J. Breeding B. Briggs . 0. Campbell . W, Cardwell . H. Carr . |. Cathers . E. Center {K-25) . A, Charpie H. Coobs . L. Culler H. DeVan . B. Emlet (K-25) . K. Ergen Y. Estabrook . E. Ferguson . P. Fraas . A, Franco-Ferreira H. Frye, Jr. . T. Gresky L. Gregg . R, Grimes . Guth . S. Harrill . W. Hoffman . Hollaender . S. Householder « H. Jordan W. Keilholtz . P. Keim T. Kelley Kertesz W. Kinyon E. Lackey 42, J. A. Lane UNCLASSIFIED INTERNAL DISTRIBUTION 43, 44, 45, 46, 47, 48, 49. 50. 31. 52, 53. 4, 5. 56. 57. 58. 59. 60. 61. 62. 63. 64, 65, 66. 67. 68. 69. 70, 71, 72, 73. 74. 75, 76. /7. 78. 79. 80-82. 83-90, 91-95, I——205-98. EXTERNAL DISTRIBUTION 99. R. F. Bacher, California Institute of Technology 100. Division of Research and Development, AEC, ORO 101-630. Given distribution as shown in TID-4500 (13th ed. Rev.) under Reactors-Power category (75 copies — OTS) UNCLASSIFIED ORNL.-2431 UC-81 — Reactors—Power ivingston cPherson nly ann cDonald Mchlly ilford . Li . Ma . Ma . Man .M .M xfn_;uhir-mix;u ‘-—va’am>moom * . Murphy P Murray (Y-12) . L. Nelson . Patriarca . M. Perry . Phillips . M. Reyling T. Roberts . T. Robinson . E. Romie . W. Savage . W. Savolainen L. Scott . D. Shipley . J. Skinner . H. Snell A. Swartout . Taboada . H. Taylor . VonderLage . Watson . Weinberg . Whatley . Whitman . Williams . Winters J. Zosler ORNL Y-12 Technical Library Document Reference Section Laboratory Records Department Laboratory Records, ORNL-RC Central Research Library * .3>§>f“m>£—>§m.‘—>1-n§:—'uo>'u§t—m moomz O NoO® UNGLASSIFIED CONTENTS FOREWORD ... ettt et ate she s e st sttt sa e e e ssen et e e b e e ea b esbenbeereberaeatessenterentes SUMMARY Lttt b et ettt e a et e e ke e e se et e s s ne e b e et et ea b e s e ete e e e e e ebe st etesrenes PART 1. REACTOR DESIGN STUDIES NUCL EAR CALCULATIONS. ..ottt sttt sttt sttt s e s et s e e st ese s CAMMA HEATING OF CORE VESSEL .. et se e et ves e ene et b e HEAT TRANSFER SYSTEMS o ettt s st e as et e v n s eeeba e ees METALLURGY oottt et sttt st st e eass et e e s e te st etabesatasabesa e b besartan e naen et easesasen saenesrerens Dl Gt O oo ee ettt s ettt e e ettt eeeeeeeeeeetaee e e e eaeeeeaeae ae s e taee——eae e t—te .t taertaae ar————ttttr ., atanaanrans Mt eria] PO CUIEMONt ..o i ittt ettt e e e e e e e e s easeaee s e e e es e aeaaensrarssnsaaaeseseen b Cation OF TN R -8 o et e eete e ee e e v e e e e e e eessanseesesresse s aseeemsas eesaeasessessessseeesassnnnseeses TR AN S FOTMOTION KM TICS toviieereeretseeeeeereereeseerrereeesesseasesessasessessssseseeessansesasssseseassatassesenssasaseesas e, Welding and Brazing. ..ottt et et e en et e eaeer s eeseeneas DY NAMIC COFOSION Lottt vttt ettt et e e eas et e e e s e b st eae et b e st eb e baat s eeseaaterasaabbeseeneeosansaebstaarensan T @St Program et st et et as e et a e st e e e s s etk e e et rae e ereaaeareeree Thermal-Convection Loop Tests ittt cte et s te s e sbe s neste b e s sbeseresenna s Forced-Circulation Loop Tests ..ottt e eaera et e e et st e s es b e ssreran sasensasenas RADIATION DAMAGE .ottt ettt e et e et e st ase e e eve ke e ee s saeeeases e beassseantesesaseasaeseannese s besnn sneesnessans [N=P il LO0D T @SS ceiereriiiieiiiieeie ettt cte e e e vt e et et b e sttt e s a et s s ses e ra bt est e bt eaaebesanabeenasre s st nrananeobensenens S1atic IN-Pile Capsules ...ttt et bt sre e st aan e anesne e CHEMISTRY oottt ettt e et e sttt et ebe s vt a s eatees e b et et oSk etec b e s en b ek e e b e bt e et enmecssabans Phase Equilibrium StUdies ..o..oo i e SOIVENT SYSTEMS woovitititeie et bbb bbb s b s e st Systems Containing UF .o s Systems Containing ThF ;..o s Systems Containing Plutonium Fluorides ... Chemistry of the Corrosion ProCess ...ttt FisSion-Product BehaVior ...ttt st ere et drecen b e e sen s e e bt s b Solubility of the Noble Gases ...t Solubility of Rare-Earth Fluorides ..o Production of PuUrified Molten Salts . iiie it iieserae e ste et e e e e e rar e e s ase et te s e rtessssneeeasaaes saransssnnes Methods for Purification of Molten Sl et vctrcreessesresessresessesteeaeerarsssesss srsseesatntaneansessnenas UNCLASSIFIED UNCLASSIFIED FOREWORD This quarterly progress report of the Molten-Salt Reactor Program records the techni- cal progress of the research at the Laboratory, under its Contract W-7405-eng-26, on power-producing reactors fueled with circulating fused salts. The report is divided into two major parts: 1. Reactor Design Study and 2. Materials Studies. Until July 1, 1957, the Molten-Salt Reactor Program was largely a design study, with only token expenditures for experimental work. As of July 1, the program was expanded to include experimental work on materials. A further augmentation of the program occurred on October 1, 1957, when personnel and facilities for additional research and experimen- tation became available. As a result of this transition, the scope of this quarterly report is considerably broader than that of the previous report, particularly with respect to metallurgy and chemistry. Similarly, it is expected that future quarterly reports will present the activities of the Experimental Engineering group. UNCLASSIFIED MOLTEN-SALT REACTOR PROGRAM QUARTERLY PROGRESS REPORT SUMMARY PART 1. REACTOR DESIGN STUDIES Nuclear Calculations Additional calculations were made of the nuclear characteristics of two-region, homogeneous, molten-salt, converter reactors. Critical-inventory calculations revealed that for a 9-ft-dia core, the minimum inventory would be about 100 kg of U235, The volume of fuel in the external system was taken to be 340 ft3 for a power level of 600 Mw (thermal). Regeneration ratios were obtained as a function of inventory for a 600-Mw system, with thorium concentration as a parameter. From an analysis of the ratios, an envelope was found which is the focus of points of maximum regeneration ratio for a given fuel inventory., From this envelope it may be concluded that (1) with o fuel inventory of a trifle over 100 kg, a regeneration ratio of 0.4 can be obtained (in an 8-ft core), (2) by doubling the inventory a regeneration ratio of 0.6 can be ob- tained (also in an 8-ft core), (3) by doubling it again a regeneration ratio of 0.65 can be obtained (in a 10- to 11-ft core}, and (4) further investment of fuel would have a negligible effect on the regeneration ratio, A design-point selection is to be attempted by balancing fuel savings by regeneration against inventory and processing charges; however, it appears at present that 400 kg of U233 may be an economical maximum for these systems. A few calculations on systems fueled with U232 have indicated much lower critical inventories and better regeneration ratios than those obtained for U233 fuel, Gamma Heating of Core Vessel It was estimated that for operation of the Reference Design Reactor at 600 Mw in a core vessel 6 ft in diameter with T mole % ThF , in the fuel, core gamma rays will liberate 13.4 w/em? in the core vessel wall, Heating by gamma rays emitted in the blanket was found to be 0.97 w/cm3, and capture gamma rays originating in the wall were found to contribute 1,63 w/cm? to the heating. The calculations were made for a pure nickel core vessel, and the results are somewhat lower than those to be expected from the calculations now being made for an INOR-8 alloy vessel. Heat Transfer Systems Two thermodynamic systems for producing power from the heat from a molten-salt reactor are being considered, and components and conditions repre- senting preliminary optimization, with respect to cycle efficiency and component sizes, have been selected. Particular attention has been given to limiting thermal stresses. One system being studied ftransfers heat from the fuel salt to a coolant salt to sodium to water, and the other substitutes mercury for the sodium, The electrical output from a 600-Mw (thermal)} reactor would be 258.6 Mw with the sodium system, and 295.8 Mw with the mercury system. PART 2. MATERIALS STUDIES Metal lurgy A coordinated program is under way for the in- vestigation of container materials for molten salts that will permit operation of a molten-salt reactor for long periods of time at temperatures up to 1300°F, Of the materials investigated, the nickel- base alloys are the most suitable to the specifica- tions. Inconel is the best suited of the commer- cially available materials; but, since its corrosion resistance and high-temperature strength are marginal, the alloy INOR-8 has been developed. Inconel is being studied for comparison and as a secondary choice, The metallurgical program for investigating these materials will include material property studies and fabrication development, Techniques for remote welding and inspection are also being studied. The material property studies are, at present, concerned primarily with the procurement of the INOR-8 shapes (that welding wire, etc.) needed for corrosion tests, and is, pipe, sheet, tubing, for physical and mechanical property tests, Pre- liminary studies of cold-rolled sheet material have MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT indicated that the decrease in ductility of INOR-8 is greatest during the first 40% reduction in thick- ness. The recovery and recrystallization charac- teristics of the cold-worked INOR-8 alloy are being investigated., Mechanical property studies of welded joints of a nickel-molybdenum alloy similar to INOR-8 have indicated that the alloy is weldable and that the mechanical properties of the joint are satisfactory both in the as-welded and the aged conditions. Thermal-convection loop tests of Inconel and INOR-8 are being made in order to provide data on the corrosive properties of beryllium-bearing fuels as compared with the properties of zirconium-base fuels, the corrosive properties of fuel mixtures containing large quantities of thorium for breeding, and the corrosive properties of non-fuel-bearing fluoride mixtures for use as secondary coolants. These studies will be supplemented with forced- circulation loop tests at flow rates and tempera- ture conditions simulating those of an operating reactor, Radiation Damage An in-pile thermal-convection loop is being pre- pared for operation in the LITR in order to obtain information on the effect of radiation on corrosion of the container materials by the various fuels being considered. Analyses of the fission products in the fuels will also be made. An improved de- sign hos been worked out for the installation of thermocouple leads in the air annulus, and a mockup test is being prepared. Calculations were made for the equilibrium distribution of fission gases in the system, and the charcoal trap to be included in the system was sized accordingly, Preparations are being made for irradiations of lithium-beryllium-uranium flucride fuels in INOR-8 capsvules in the MTR. Chemistry A review has been made of the extensive body of information available on fused salts in order to orient the development of fuel solvents, fuels, breeder blankets, and secondary coolants for use in molten-salt power reactors. The systems of most promise appear to be those which contain BeF, with LiF and/or NaF. Efforts were made to resolve discrepancies in reported melting points of BeF,, and additional confirmation of the reported value of 545°C re- sulted. vealed mixtures that may be of interest as fuel solvents and carriers and as coclants in the sys- tems LiF-BeF,, NaF-BeF,, and NaF-LiF-Bef,. The NaF-LiF-UF, system does not appear to be useful directly as a fuel, but low-melting mixtures of interest are available in the LiF-BeF -UF, system. A number of ThF ,-containing systems are being studied, and tentative phase diagrams have been prepared for the BeF ,-ThF,, LiF-ThF, and NaF-ThF, systems. The LiF-BeF,-ThF, and NaF-BeF,-ThF, systems have been shown to be remarkably similar to their UF ,-containing analogs. It thus appears that a relatively wide choice of useful breeder materials is available. Studies of containing plutonium fluorides Phase-diagram investigations have re- systems initiated, An analysis of the corrosion mechanism in systems in which fluoride fuels are contained in nickel-base alloys which contain molybdenum and chromium was applied to the MSR (Molten-Salt Reactor) system. |t appears virtually certain that with small chromium activities such as those in inconel and the INOR alloys, and with small temperature drops such as those contemplated for most reactors, chromium deposition will not result if fuel mixtures based on the BeF,-containing system are used, The available information on fission-product behavior in fluoride fuels has been analyzed in terms of the MSR program. Preliminary experi- ments with NaF-BeF, {57-43 mole %) have shown the solubility of CeF, in this mixture to be con- siderably less than in NaF-ZrF, mixtures and to be more temperature sensitive, were The anticipated, nearly ideal, solid solution behavior of mixtures of rare earths in the BeF,-containing system, along with reduced solubility, should make the fission- product partition process quite attractive, Part 1 REACTOR DESIGN STUDIES NUCLEAR CALCULATIONS L. G. Alexander In further nuclear calculations critical inven- tories and regeneration ratios were obtained as functions of core diameter and thorium concentra- tion in the core. For these calculations the new intermediate-concentration thorium were used. cross sections for Also, minor corrections to some of the previously reported cases have been obtained. Relevant specifications for the systems studied are given in Table 1.1, The results of the calculations are summarized in Table 1.2; analyses of the results are being made. A graph of the critical concentrations estimated for these reactors is presented in Fig. 1.1, It may be seen that the curve for 1 mole % ThF , does not conform to the general pattern. The calculations have been checked and the results are believed to be correct. In an effort to clarify this behavior, the flux-averaged cross sections for fission of y23s Table 1.1. Specifications for Two-Region Molten-Fluoride-Salt Reactors Core Diameter 5t0 10 ft Carrier salt 69 mole % LiF, 31 mole % BeF2 Mean density 2.0 g/cm3 Li composition 0.01% Li6 U238 concentration 7.5% of U235 concentration s See Table 1.2 concentration Core Vessel Thickness ]/3 in. Composition INOR-8 alloy Blanket Thiekness 2 ft Composition 25 mole % ThF4, 75 mole % LiF Mean density 4.25 g/cm’® Geometry Spherically symmetric J. T. Roberts and capture in thorium were computed from the flux spectra: g = .1-020 fvc No ¢(u) du dV , where V _ = volume of core, N = atoms per cubic centimeter. The results for reactors having zero and 1 mole % ThF, in the core are shown in Table 1.3. It is thought that increasing the thorium concen- tration in the 10-ft core decreases the fission cross section more or less uniformly so that the critical concentration rises regularly. In the 6-ft cores the spectrum is already hard. Adding thorium hardens it further, but the effect on the fission cross section is less., There is a regenerative effect at work, The spectrum is hardened in the following three ways: (1) by decreasing core size, (2) by increasing thorium concentration, and (3) by increasing uranium concentration, If the spectrum is hardened relatively less in the 6-ft core by the addition of thorium, the de- crease in uranium cross section is relatively less; hence, less uranium is needed, and the spectrum UNCLASSIFIED ORNL~LR —DWG 24920R n n n N ~N Q @ & s n 1 mole % ThF, CRITICAL CONCENTRATION [atoms/(cm®x 15 '*)] B o co 6 N CORE DIAMETER (f1) Fig. 1.1. Critical Concentrations of U235 for Two- Region, Homogeneous, Moclten-5alt Reactors. MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT Table 1.2, Nuclear Characteristics of Clean Two-Region Molten-Flucride-Salt Reactors at a Fuel Temperature of 1150°F Case number Fuel Thorium concentration in fuel (mole %) Diameter of core (ft) Critical concentration (mole % UF,) Critical mass (kg) Critical inventory* (kg) Regeneration ratio Core Blanket Total Neutron balance Captures in Fuel U238 } Th in core Th in blanket Li+ Be in core F in core Core vessel Li + F in blanket Leakage from blanket Neutron yield,* 77 87 53 54 55 106 86 107 61 88 89 90 9 92 93 63 84 62 87A U235 U235 U235 U235 U235 U235 u23.‘.5 U235 U235 u235 U235 U235 U235 U235 U235 U235 U235 U233 0 0 0 0 025 025 025 0.25 0.5 0.5 0.5 075 075 075 1 1 1 0.25 5 6 8 10 6 7 8 10 6 8 10 6 8 10 6 8 10 5 0.453 0,110 0.047 0.033 0.243 0.160 0.167 0.073 0.572 0.185 0.113 0.772 0.318 0.180 0.875 0.492 0.281 0.147 82 47 49 67 102 80 80 106 180 138 164 243 237 260 275 367 409 27 509 189 110 11 408 230 181 175 720 313 270 972 537 428 1100 831 674 165 0.039 0.025 0.014 0.012 0.170 0.201 0.224 0.256 0.229 0.332 0.390 0.261 0.378 0.459 0.288 0.394 0.485 0.160 0.541 0.531 0.408 0.303 0.432 0.410 0.350 0.262 0.368 0.305 0.233 0.338 0.267 0.206 0.322 0.236 0.183 0.718 0.580 0.556 0.422 0.315 0.602 0.611 0.574 0.518 0.597 0.637 0.623 0.599 0.645 0.665 0.610 0.630 0.668 0.878 0.5578 0.5210 0.4983 0.4918 0.5513 0.5227 0.5105 0.5007 0.5716 0.5279 0.5120 0.5781 0.5497 0.5262 0.5781 0.5663 0.5449 0.4574 0.0072 _ 0.0220 0.0128 0,.0060 0.0935 0.1053 0.1145 0.1283 0.1308 0.1752 0.2000 0.1510 0.2071 0.2416 0.1657 0.2236 0.2643 0.0000 0.0729 0.3013 0.2768 0.2029 0.1490 0.2380 0.2124 0.1788 0.1310 0.2104 0.1612 0.1195 0.1955 0.1466 0.1086 0.1863 0.1136 0.0995 0.3281 0.0239 0.0686 0.1587 0.2254 0.0301 0.0636 0.0949 0.1414 0.0164 0.0552 0.0889 0.0116 0.0306 0.0573 0.0096 0.0191 0.0347 0.0329 0.0470 | 0.0741 0.0950 0.1208 0.0072 0.1278 0.0871 0.0942 0.1013 0.0986 0.0708 0.0805 0.0796 0.0638 0.0660 0.0663 0.0603 0.0574 0.0566 0.1087 0.0046 .79 1.93 2.01 2.03 1.81 .91 19 2,00 1.75 189 1.95 1.73 1.82 1.90 173 176 1.84 2.19 94 U233 0.25 0.092 29 116 0.175 0.592 F 0.767 0.4538 0.0896 0.2686 0.0696 0.1184 2.20 95 233 0.25 0.056 41 94 0.252 0.421 B 0.673 0.4 502 0.1133 0.1894 0.1298 0.1173 2,22 96 233 0.25 10 0.043 63 103 0.285 0.306 0.591 0.4486 {ome) 0.1371 0.1763 0.1100 2,23 *For 600-Mw system with 340 #3 of fuel in system outside core. Table 1.3. Flux-Averaged Cross Sections in Molten-Salt Converter Reactors Core Thorium Average Cross Diameter Concentration Section (1) in Core (barns) (mole %) U235 Th 10 0 94 3.0 1 14 1.9 é 0 36 2.7 1 8 1.8 is hardened less, etc, |f the effect of the thorium decreases sharply around 0.75 mole % ThF, be- cause of resonance shielding, the deviation in the results for the 6-ft core with 1 mole % ThF , in the fuel would be satisfactorily accounted for, It is planned to calculate the cross sections for cores with 0,25 and 0.75 mole % ThF, in order to verify this speculation. The critical masses were computed from the core diameters and critical concentrations, The results are listed in Table 1.2 and plotted in Fig. 1.2, together with four additional points resulting from preliminary calculations for 5-ft cores. The apparent deviation of the 6-ft core with 1 mole % ThF, is more pronounced than in Fig. 1.1. It should be remembered, however, that regardless of the thorium concentration the critical masses of all these reactors must decrease as the diameter decreases to below some critical value. Two points, one for a small, homogeneous, molten-salt, beryllium-reflected critical experiment and the other approximating a metallic uranium critical assembly, have been plotted in the lower left-hand corner of the figure. All the curves in Fig. 1.2 probably pass near these two points and through the origin; a family of curves consistent with this assumption is shown in Fig. 1.3. The calculated points appear to correlate reasonably well on this basis. Further calculations will be performed to verify some parts of the curves, especially parts of the curve for 1 mole % ThF . As will be shown later, however, the reactors corresponding to the portions of the curves to the left of the 6-ft ordinate are not economically attractive, and the details of the peaks will not be investigated. The solid portions of the curves in Fig. 1.3, which are believed to be correct, and the curve for PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED ORNL—LR—DWG 26533 450 ] —! [ i i 1 O PRELIMINARY CALCULI{ATIO NS FOR 5-ft CORES 400 350 300 250 200 CRITICAL MASS (kg of UZ>%) 150 100 50 CORE DIAMETER (ft) Fig. 1.2. Critical Masses for Molten-Salt Reactors, 1 mole % ThF,, which was accepted provisionally, were used to compute the critical inventories listed in Table 1.2 and shown in Fig. 1.4, (The values at 7 and 9 ft were obtained by interpolation on a graph of the regeneration ratio vs diameter and ThF, concentration,) The volume of fuel in the external system (pump, pipes, heat exchangers) was taken to be 340 ft3, which corresponds to a power level of 600 Mw (thermal). It may be seen that a minimum inventory of about 100 kg of U23° is obtained at a core diameter of 9 ft, The corre- sponding power density in the core is 55 w/cm’ (average) and the specific power is 6 Mw per kilo- gram of fuel in the system, The regeneration ratio is shown in Fig. 1.5 as a function of inventory (for a 600-Mw system), with thorium concentration as a parameter, It is thought MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT UNCLASSIFIED ORNL—LR—DWG 26534 450 - 400 350 300 250 200 CRITICAL MASS (kg of U23%) 100 50 CORE DIAMETER (ft) Fig. 1.3. Tentative Correlation of Critical-Mass Data for Molten-Salt Reactors Based on Assumption that All Curves Pass Through Origin. that for a given thorium concentration the critical inventory falls as the core diameter increases because the concentration falls more rapidly than the volume increases., Simultaneously the ieakage of neutrons to the blanket decreases, but this decrease is, at first, more than offset by the in- crease in captures by the thorium in the core be- cause of the increase in the ratio of thorium con- centration to uranium concentration, Eventually, however, the uranium concentration tends to level off ot the finite concentration corresponding to the infinite reactor, and the volume increase becomes the important factor. As a result, the inventory passes through a minimum. At about UNCLASSIFIED ORNL-LR-DWG 24967R 1100 1000 900 8C0 700 800 500 400 CRITICAL INVENTORY (kg of UZ3%) 300 200 100 *1 4 5 6 7 8 9 10 CORE DIAMETER (ft) Fig. 1.4, Salt Reactors. Critical Inventories for 600-Mw, Molten- the minimum point, the thorium captures in the core level off, But, since the leakage to the blanket continues to fall with volume increase, the regeneration ratio falls off. Also, parasitic captures in fluorine, etc., increase as the spectrum softens. Thus the curves in Fig. 1.5 turn down- ward and eventually must turn back to the right., The downturn of the curves is surprisingly sharp, It appears that the ‘‘elbows’’ define an envelope which is the locus of points of maximum regenera- tion ratio for a given fuel inventory, This envelope is shown in Fig. 1.6, The principal conclusions to be drawn are that {1} with a fuel inventory of a trifle over 100 kg, a regeneration ratio of 0.4 can be obtained (in an 8-ft core), (2) by doubling the UNCLASSIFIED ORNL-LR-DWG 24921R 0.8 0.7 o g xr 06 g 5 1o & 0.5 | = Il [ Ly / 0.4 0.3 * J O 500 400 600 800 1000 CRITICAL INVENTORY (kg of U%3°) 1200 Fig. 1.5. Regeneration Ratios in Two-Region, 600-Mw, Molten-Salt Reactors. UNCLASSIFIED ORNL—-LR-DWG 24922R 0.8 0.6 ’ REGENERATION RATIO 04 02 A 0 200 400 600 800 CRITICAL INVENTORY (kg of U239 Fig. 1.6. Optimum Regeneration Rotios for Two- Region, 600-Mw, Molten-Salt Reactors. 1000 PERIOD ENDING OCTOBER 31, 1957 inventory a regeneration ratio of 0.6 can be ob- tained (also in an 8-ft core), (3) by doubling it again a regeneration ratio of 0.65 can be obtained (in a 10- to 11-ft core), and (4) further investment of fuel would have a negligible effect on the re- generation ratio, The optimum diameter appears fo increase steadily after the inventory increases beyond 200 kg. A plot of 3 — 1 is also included in Fig. 1.6 to give the upper limit on the regeneration ratio. The interval between the envelope and the 7 — 1 curve represents parasitic absorptions in fluorine, etc, These absorptions decrease with increasing in- ventory, because of increased competition from uranium and thorium, and also because of hardening of the spectrum. Judging from the trends of the curves in Fig., 1.6, however, it would seem that, although n — 1 approaches a minimum at an inven- tory of 600 kg, further significant reduction in parasitic absorptions can be obtained only by enormous increases in the inventory. Selection of a design point for reactors of this type will consist in balancing fuel savings by re- generation against inventory and processing charges. This will be attempted; however, it does not seem likely that it would pay to invest more than 400 kg of U?3% in such systems. It must be remembered that the results discussed above apply only to the initial, clean state; that is, they apply only in the absence of fission products and uranium isotopes (other than U238y, The accumulation of these substances will of course affect both the critical inventory and the regeneration ratio. However, the performance of the system may not be impaired significantly for some time after startup. In the first place, the U233 generated in the blanket will have a fission cross section in these epithermal reactors about twice as high as that of U235, Adding this MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT U232 15 the core will tend to offset the ac- cumulation of poisons, and it may not be neces- sary to add UZ33 for some time, Further, 7 for U233 is much better than that for U23°, especially in the intermediate-energy range, and replacing U235 with U233 will reduce the parasitic absorp- tions in the fuel. These effects will be studied by means of the Sorghum program being prepared for the Oracle, A few calculations on systems fueled with y233 have been completed. Four cases, in which core diameters range from 5 to 10 ft and the ThF , con- centration is 0.25 mole %, are described in Table 1.2. The results are compared with those for cor- responding U233 systems in Fig. 1.7. The critical inventories are much lower with U233 fuel, and the regeneration ratio is much better. It is hoped that regeneration ratios well above 0.9 can be achieved at inventories below 500 kg, The calcu- lations will be extended. UNCLASSIFIED ORNL —LR—-DWG 24923R 1.0 U233 | \: > wn w = 0 x 20 %, REDUCED T I-—“-'_—l__._.—l--lI‘—"'_'. i,! - _—r—\ /A-/ A 300 \; O G 10, REDUCED ) 200 100 : o 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 ANNEALING TEMPERATURE {°F) Fig. 2.4, Hardness (VHN) of INOR-8, Heat SP-16, Deformed by Rollifig and Annealed 1 hr at Indicated Tem- perature. 20 PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED ORNL-LR~-DWG 24041 70 [ 60 e / '_ 2 W 50 A Ll o & / o - g £ 40 A [aN] Z ! 10% REDUCED | 5 1 T > | o | uwJ F P z L 50 % REDUCED "] & 20 - a ) a 10 / 40 %, REDUCED L 609, REDUCED T | 80 % REDUCED 0 | 0 200 400 600 800 1000 1200 1400 1600 1800 2000 2200 ANNEALING TEMPERATURE (°F) Fig. 2.5. Room-Temperature Elongation (2-in. Gage Length) of INOR-8, Heat SP-16, Deformed by Rolling and Annealed 1 hr at Indicated Temperature. of test, termed the circular-groove test, utilizes a 3/32-in.-wide, 2-in,-diq, %-in.-deep, circular groove machined in a 4 x 4 in. specimen of ]/2-in. An inert-arc fusion pass (no filler metal addition) is made around the groove and the absence or presence of weld-metal cracks is observed, Test samples of heats SP-16 and 30-38, together with samples of four other alloys for comparison, are shown in Fig. 2,7. It may be seen that the restraint of the specimen caused complete circumferential cracks in the Haynes SP-16 specimen, while no cracking can be ob- served in the ORNL 30-38 dlloy. As was ex- pected, the Inconel specimen cracked extensively, because no columbium-modified filler metal (known as INCO 62) was added. plate. Another type of crack test, designated as the open-end slot test, utilizes a 2% x 5-in, specimen of ]/2-in. plate. A 3-in,-long, ]/]6-in.-wide slot is cut from one end, and a fusion pass is made from the open end of the slot toward the closed end, As may be seen in Fig. 2.8, cracks were found in the Haynes SP-16 alloy but not in the ORNL 30-38 alloy. As a means of further evaluating the weldability of the Haynes SP-16 alloy, the preparation of typical weld-test plates (Fig. 2.9) for radiographic, mechanical-property, and metallographic studies was initiated. The first test plate, NiMo plate No. 36, was fabricated by using SP-16 alloy filler metal which had been sheared from ]/16-in. sheet. Preparation of this plate was terminated after 2] MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT {x 403) UNCLASSIFIED ORNL~LR-DWG 241450 220 200 ~ A 4 N - N — N, A 2 N w480 o 1_. » A wJ 1 5 \ = o - 160 N 5 N = \ = N — < M 60 5 a sy 140 120 ELONGATION (%) 50 100 200 500 1000 2000 ANNEALING TIME (min) Fig. 26. lsothermal Annealing Curves for INOR-8, Heat SP-16, for 60 and 80% Cold-Reduced Specimens. deposition of the first few passes because gross weld-metal cracking had occurred. The extent of the cracking is evident in Fig. 2,10, Metallo- graphic examination of the weldment revealed weld-metal cracks of the type shown in Fig. 2.11 and base-metal cracks of the type illustrated in Fig. 2.12. In view of the unsatisfactory results with the SP-16 filler metal, another weld-test plate, NiMo plate No. 37, was fabricated by using Hastelloy W filler wire. This plate is shown in Fig. 2.13. Although no weld-metal cracks were found by visual and dye-penetrant inspection, metallo- 22 graphic examination of the weld joint revealed severe base-metal cracks of the type shown in Fig. 2.14. Extensive fusion-line porosity of the type illustrated in Fig. 2.15 was also found, Metallographic examination of the SP-16 heat revealed an abnormal number of stringers, which may be responsible for the porosity and extensive base-metal and weld-metal cracking. On the other hand, ORNL alloy 30-38 appeared to be immune to weld-metal cracking, and thus melting practice may be responsible for the poor performance of the Haynes heat. Additional tests are being conducted to evaluate weld-metal and base-metal cracking tendencies of the ORNL heat, INOR-8 HAYNES SP—16 (HT 8284) PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED Y-24145 HASTELLOY B Fig. 2.7. Circular-Groove Test Specimens for Studying Weld-Cracking Susceptibility. DYNAMIC CORROSION J. H. DeVan J. R. DiStefano R. S. Crouse Test Program A basic question underlying the development of a molten-salt reactor for power production concerns the long-time compatibility of structural materials with the intended fluid heat-transfer mediaq, fluoride salts and sodium. As stated above, two structural alloys, Inconel and [INOR-8, are presently being studied as possible container materials for these appli- cations. Corrosion experiments utilizing thermal- convection loops and forced-circulation foops fab- ricated of both alloys are presently under way. The test program will, in general, involve three phases. In the first, the relative corrosive properties of the several fluoride mixtures listed in Table 2.2, are to be determined in thermal- convection loops operated for a period of 1000 hr., These tests will provide data on the corrosive properties of beryllium-bearing fuels as compared with the properties of zirconium-base fuels, the corrosive properties of fuel mixtures containing lorge quantities of thorium for breeding, and the corrosive properties of non-fuel-bearing fluoride mixtures for use as secondary coolants, The second phase of testing, for which thermal- convection loops will again be used, will subject those systems which appear to be compatible in phase-1 experiments to more extensive investi- gations at longer time periods and ot two temper- atures, 1250 and 1350°F. The third and final phase of this out-of-pile testing will be conducted in forced-circulation loops at flow rates and temperature conditions simulating those of an operating reactor system, The temperature con- ditions to be employed in all three phases of testing are shown in Table 2.3, The status of 23 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT UNCLASSIFIED Y-24190 | INOR-8 | HAYNES SP—16 (HT 8284) | o Fig. 2.8. Circular-Groove and Open-End Slot Weld-Cracking Test Specimens. 24 tests which have been conducted thus far under these programs is described below, Thermal-Convection Loop Tests A second Inconel loop has completed 1000 hr of operation under the phase-1 program and has been examined metallographically, This second UNCLASSIFIED ORNL-LR-DWG 22383R Fig. 2.9. Weld Test Plate Design. PERIOD ENDING OCTOBER 31, 1957 loop circulated fuel mixture 123, NaF-BeF,-UF, (53-46-1 mole %), and operated in accordance with the temperature schedule given in Table 2.3 Hot-leg sections of the loop showed light-to- moderate surface roughening and surface pits to a depth of 0.00025 in., as shown in Fig. 2,16, The appearance of the cold-leg specimens was similar to that of the hot-leg specimens, and they were entirely free from layers of metal crystals. The results of chemical analyses of samples of the fused salt taken before and after the test for the metal constituents are given in Table 2.4, Other Inconel thermal-convection loops are now being operated to complete an initial evaluation of all the salt mixtures listed in Table 2.2, An INOR-8 loop is also being operated with the salt mi xture NoF-Ber-UF4 (53-46-1 mole %), and additional INOR-8 loops are being fabricated for operation with the remaining fluoride mixtures listed in Table 2.2. UNCLASSIFIED Y-24048 Fig. 2.10. NiMo Test Plote No. 36. 25 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT ' _ " UNCLASSIFIED L Y-24179 AL A L A o . L . - # " 4 : [ 5] & - .'_ p : " - .02 ¥ ® - ¢ - e - /7 """ - - » _‘ . - 4 = ) to T Foa o ~ ; - oamm— | . = - e ¢ . L N s . / / - e -~ . - - . - i ’ o 1 » . " —ad g ) T ) ; - " g _“ o - . 4 . . » - - " » - " po—u oA 2 Fig. 2.11, Weld-Metal Cracks in Haynes SP-16 Weld Metal Welded to SP-16 Base Plate. Etchant: chromic ocid and HCI. 100X, UNCLASSIFIED Y-24180 , e INCHES Fig. 2.12. Base-Metal Cracks in Haynes SP-16 Welded with SP-16 Weld Metal. Etchant: chromic acid and HCI. 100X. 26 PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED Y.-24049 1 Fig. 213, NiMo Test Plate No. 37. . - . ' UNCLASSIFIED z , Y.24182 , ' , . . { i L i ' : > / ' . : r . . v . * \ . 8 \ . - ‘ oo . . * \' - ‘fl-\‘. . -, . . v\‘! : ‘ . . ~ 1 ' vy ‘ L . . . Cs \ ) . o’ o "\ ' ¥ « . + T ‘ ‘ 1\ . . v \ 1 . ' -8 . - Y 1 Y r INCHES jo E I 1 1 o - 3 01 Ole 013 e Fig. 2,14, Base-Metal Cracks in Haynes SP-16 Welded with Hastelloy W Filler Metal. Etchant: chromic acid and HCL. 250X. 27 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT o : o 'UNCL ASSIFIED : ’ L o : : : | Y-24181 / : ; . , .7 J : / ‘ Ir b ‘ ;' ¥ ’ ” . -‘ . / ! /, A ‘e I s P ! 4 ’ , ] ' ‘ ? { p " . y i ff ,L;. .} A /] : Ee * ey T R I N ! ' B3 ' .‘b ] ' } [ ¥ v i b f ' ok P 2 . v { ‘ ! - ! ' J v / . ' . ‘ i v 4 i * { l ; . . . . \ ; l. 3 v’, . . i }‘i .., . ’ 4 % . . . ; e 02 oo . . r ) A 1 - » v_ . ’ jo.o2 K ‘ Fig. 2.15. Fusion Line Porosity of Type Found When SP-16 Base Metal Was Welded with Either SP-16 or Hastelloy W Filler Metal. Etchant: chromic acid and HCI. Table 2.2, Compositions of Molten Salt Mixtures to be Tested for Corrosiveness in Inconel and INOR-8 Thermal-Convection Loops Salt Composition {mole %) Designation NaF LiF KF ZrF4 BeF, UF, ThF4 Fuel and Blanket Salts 122 57 42 1 129 55,3 40.7 4 123 53 46 1 124 58 35 7 125 53 46 0.5 0.5 126 53 46 ] 127 58 35 7 128 71 29 130 62 37 1 131 60 36 4 Coolant Salts 12 1.5 42 84 27 38 28 Table 2.3, Operating Temperatures to be Used in Yarious Phoses of Corrosion Testing Operating Temperatures {°F) Difference Between Maximum Hot and Cold Sections Phase 1 Fuel mixtures 1250 170 Coolant salts 1125 125 Phase 2 Fuel mixtures 1250 to 1350 170 Coolant salts 1050 to 1250 170 Phase 3 Fuel mixtures 1300 200 Coolant salts 1200 200 Forced-Circulation Loop Tests An Inconel forced-circulation loop has been in operation for some time in order to determine the corrosion rate of this alloy in a zirconium-base fluoride and in sodium over a pericd of one year, The loop design employed provides for the oper- ation of separate sodium and fluoride circuits connected through a U-bend heat exchanger, The maximum fluoride interface temperature in this foop is 1250°F, while the maximum sodium temper- ature is 1150°F. The loop, as of October 15, had accumulated 4900 hr of operation. Con- struction of additional Inconel forced-circulation loops, as well as INOR-8 forced-circulation loops, is now under way for testing under the phase-3 program. PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED T-13379. st [ 009 ¢ 4 - ' I'-\, - ) ) 10 e * . L ‘ L. . \ - . - Tl . Lo Lt ~ . . .C" A ;\\ " o 00 . S . . o o T ' s * " » ST . . 2 T ‘ 012 o O e A L o TE T - P - s r 3 . 013 . . . ¥ v Sty : %, . o . T T L 014 Fig. 2.16. Hot Leg of Inconel Thermal-Convection Loop 1163 Which Circulated Fuel Mixture 123, NaF- BeF,-UF, (53-46-1 Mole %), for 1000 hr at a Maximum Temperature of 1250°F. 250X. Table 2.4, Metal Constituents of Fused Salt Before and After Circulation in Loop 1163 Major Minor Constituents {ppm) Constituents (wt %) U Be Ni Cr Fe Before-test sample 3,52 8.03 75 375 185 After-test sample Hot leg 3.43 8.32 30 465 180 Cold leg 3.50 8.32 30 435 180 29 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT RADIATION DAMAGE G. W. Keilholtz IN-P{LE LOOP TESTS W. E. Browning R. P, Shields H. E. Robertson J. F. Krause An in-pile thermal-convection loop fabricated of an INOR-8 alloy is being prepared for operation in the LITR with a lithium-beryllium-uranium fuel mixture. The purpose of the experiment is to demonstrate the reliability of this combination of materials and to determine the effects of radiation on corrosion of the container material by fluoride fuel in the MSR design temperature range. The fuel container will be examined metallographically for evidence of corrosion after irradiation, ond an analysis will be made of the fission products in the fuel., Design features and operating con- ditions for the loop are described below: Fuel composition Ber-LiF-UF4 {37-62-1 mole %) with Li” and U233 Power density 50 w/cm® or higher as necessary to achieve desired temperature differential Maximum temperature 1250°F Axial temperature 150°F differential Up to 10,000 hr Duration of in-pile operation The loop will be air-cooled and will have external heaters. It will also have a charcoal trap for removing the xenon gas evolved from the fuel. The mechanical details are similar to those of the miniature, vertical, dynamic corrosion loop recently operated in the LITR. This loop can be converted to a forced-circulation loop with a minimum of modification after suitable pump bearings have been developed. The parts which will be assembled to form the loop are shown in Fig. 2.17. It may be noted that the loop consists essentially of two parallel tubes with connecting bends at the top and bottom. The loop will be cooled by air passing through an air annulus which will, in general, follow the loop all the way around. The air will travel in the same direction as the fuel. 30 UNCLASSIFIED - PHOTO 41790 CHARCOAL TRAP LIPS v ) | — FUEL LEVEL fo = i k] 8 AHCIYHOET™ TENOILTN COOLING AIR Fig. 2.17. Parts for INOR-8 Thermal-Convection Loop To Be Operated in the LITR. Details of the design and planning of the con- struction and operation of this loop have been reported, ! The fuel tube is 30 in. long, % in, in outside diameter, and will contain 42 ecm? of fuel. The total power will be approximately 2 kw. A full-scale mockup of the in-pile loop is being prepared, with electrical resistance heating of the fuel tube to simulate fission heating. The purpose of this out-of-pile test will be to demon- strate the workability of the design and to provide operation information that will aid in an evaluation of hazards in the reactor. The parts shown in Fig. 2.17 have been fabricated for the bench test and are being assembled. Practice welds have been made for the fuel tube by using INOR-8 tubes and INOR-8 welding rod. Fuel for the bench test has been requested. An improved design has been worked out for the thermocouple leads in the cooling-air annulus that should permit greater movement of the fuel tube without plastic deformation of the thermocouple lead wires. Preparations are being made to conduct a thermocouple mechanical-mockup test of this improved thermocouple installation tech- nique. Calculations have been made for the equilibrium distribution of fission gases in the system. It is estimated that 9 x 1074 moles of stable xenon and 1.5 x 10~% moles of stable krypton will accumulate during 10,000 hr of operation at 2 kw. This amount of xenon and krypton will be absorbed in the 500-g charcoal trap to such a degree that ]W. E. Browning, Proposed In-Pile Convection Loop for Molten-Salt Reactor, ORNL CF-57-11-20 (in prepara- tion). PERIOD ENDING OCTOBER 31, 1957 the remaining partial pressure of xenon will be 600 u Hg and that of krypton will be 150 u Hg. These equilibrium partial pressures are sufficiently low for the adsorption efficiency of the charcoal to be unimpaired. The combined decay-heat generation of all the daughters of fission gases having half-lives greater than 10 sec is only 6 w, according to the calculations. Cooling of the charcoal trap will not be a difficult problem. Parts for the in-pile loop are being fabricated and will be assembled ofter the bench test has been started. Fluoride fuel containing enriched U235 and Li7 has been requested for the in-pile test, The INOR-8 tubing required for both the bench and the in-pile apparatus is available, STATIC IN-PILE CAPSULES H. L. Hemphill Preparations are being made to irradiate lithium- beryllium-uranium fluoride fuels in INOR-8 cap- sules in the MTR. Tests simulating in-pile conditions have been conducted on INOR-8 tubing Thermal cycling in air at various temperatures up to 1250°F did not W. E. Browning suitable for use in capsules. damage the adherent oxide coating which protects INOR-8 from atmospheric corrosion. Further tests are to be conducted in high-velocity air streams similar to those that are used in-pile, and equip- ment for testing the durability of thermocouples attached to INOR-8 capsules is being prepared. Fuel material has been requested for these tests. 31 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT CHEMISTRY W.R. Grimes PHASE EQUILIBRIUM STUDIES C. J. Barton R. E. Moore H. F. Friedman R. A. Strehlow H. Insley R. E. Thoma R. E. Meadows C. F. Weaver An extensive and systematic study of phase behavior in fluoride systems has been conducted in the Chemistry Division of the Oak Ridge National Laboratory. In recent months attention has been directed toward those systems of possible use as fuel solvents, fuels, breeder blankets, and secondary coolants. Although the final choice of a fuel system for even the simplest of such reactor embodiments is far from certain, it appears that the systems of most promise are those which contain Bel:2 with LiF and/or NaF. The present status of the knowledge of phase behavior of these various systems is described briefly here, and references are given to more detailed reports. Solvent Systems Melting Point of BeF,. — Several investigators have examined the melting point of BeF,, and widely differing results have been obtained. Quenching experiments by Roy, Roy, and Osborn? established a value of 545°C, which was confirmed by Eichelberger et al. of Mound Laboratory. Recently, however, Kirkina, Novoselova, and Simanov,3 who studied the system BaF,-BeF, by thermal analysis and x-ray diffraction, stated that BeF, crystallizes in a new modification that is stable above 545°C when BaF, is present. They claim that 545°C is o transition temperature between the two forms of BeF, and cite the existence of a eutectic at 600°C between BaBeF , and BeF, as proof. Recent quenches of mixtures containing BeF, and up to 5 mole % BaF, have not confirmed the Russian work. When the samples were equilibrated at and quenched from temperatures above 550°C, only isotropic material and quench growth were 2p, M. Roy, R. Roy, and E. F. Osborn, J. Am Ceram Soc. 33, 85 (1950). 3D. F. Kirkina, A. V. Novoselova, and Y. P. Simanov, Zhur. Neorg., Khim. 1, 125 (1956). 32 found by petrographic examination. X-ray dif- fraction of the specimens showed none of the commonly observed form of BeF, but revealed the existence of lines identical with those presented by the Russian investigators for BaBeF,. The ordinary variety of BeF, was found by petrographic examination of quenches from below 550°C, and x-ray diffraction revealed ordinary BeF, along with the compound found at higher temperatures. Since this material coexists in equilibrium with BeF, below 550°C, it is, presumably, BaBeF and it cannot be a high-temperature modification o? BeF .. These experiments present additional confirmation of the 545°C melting temperature of BeF . The System L.iF-BeF,. — The phase diagram presented in Fig. 2.18 for the simple system LiF-BeF2 represents data from Roy, Roy, and Osborn; 4" Novoselova, Simanov, and Yarembash;?3 and the Oak Ridge National Laboratory. The low cross sections and generally good nuclear properties of BeF, and Li’F, along with the very low liquidus temperatures available in mixtures containing from 35 to 70 mole % BeF,, make this system of obvious interest as a fuel solvent or as a coolant. The System NaF+BeF,. — The phase diagram for the NaF-BeF, system, shown in Fig. 2.19, was published by Roy, Roy, and Osborn.® The lowest temperature in the system is below that in the LiF-BeF2 system, but the liquidus temperature rises more steeply as the BeF, concentration is decreased. Mixtures that contain between 40 and 65 mole % BeF, may be of interest to the program. The System NaF.LiF-BeF,. -~ A preliminary and partially complete diagram of the ternary system NaF-LiF-BeF, is shown as Fig. 2.20. The ternary compound has been identified, with a fair degree of certainty, as NaF.LiF.3BeF, by petrographic and x-ray diffraction examinations of 4D. M. Roy, R. Roy, and E. F. Osborn, J. Am. Ceram. Soc, 37, 300 (1954). 5A. V. Novoselova, Y. P. Simanov, and E. |. Yarem- bash, Zhur. Fiz. Kbim, 26, 1244 (1952). 6D. M. Roy, R. Roy, and E. F. Osborn, J. Am. Ceram. Soc. 36, 185 (1953). PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED ORNL~LR-DWG 16426 900 800 — 700 | ————— — ) £ 600 |—— LiF + L1QUID —‘fif ) o 2 5 | < . a L 500 b — L - 400 ( LipBeFy . S +LIQUID R LiF + Li,BeF, Li,BeF, + BeF, (HIGH QUARTZ) 300 —————»———‘——— W — fi—?——‘Lh—?——jfi—v‘-T o 5 ., o | HeBefa LiBeFy + BeF, (HIGH QUARTZ) : @ | | 200 (LiBeFy 4 [iBeF3 + BeFp (LOW QUARTZ) | LiF 10 20 30 40 50 60 70 80 90 BeF; BeF, {mole %) Fig. 2.18. The System LiF-BeF,. previously equilibrated quenched samples con- taining 60 mole % BeF,-20 mole % LiF. The compound melts incongruently to BeF, and liquid at about 287°C. Only single-phase material was found just below 287°C, but the solid-phase reaction in which NaF.LiF.3BeF, is formed proceeds very slowly at only a few degrees lower. The primary phase field of the ternary compound is small and lies near the composition 50 mole % BeF2—25 mole % LiF. Exominations of quenched samples of a number of compositions containing more than 40 mole % BeF, have led to the conclusion that the following are compatibility triangles: NaF.LiF-3BeF, ~ BeF, — 2LiF-BeF, NaF.LiF-3BeF, — 2LiF.BeF, — 2NaF.LiF.2BeF, NaF-LiF-BBeF2 - 2NoF-LiF-28eF2 - NOF-BeF2 l\lcnlz-Lil:-.'.’:BeF2 - BeF2 - Nc:F-BeI:2 Although exact determinations of ternary eutectics and peritectics hove not yet been completed, it is apparent that a region with melting points below 300°C exists in the neighborhood of the primary phase field of NaF.LiF.3BeF,. Compositions in this region (within 5 to 10 mole % of 50 mole % BeF,-25 mole % LiF) may be of interest as low- melting reactor coolants or fuel carriers. Systems Containing UF4 The binary systems LiF-UF, and NaF-UF, have been well known for some years.” Neither 7C. J. Barton et al., "*Phase Equilibria in the Systems LiF—UF4 and NaF-UF4," J. Am. Ceram. Soc. (in press). 33 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT 800 800 NaF + LIQUID 700 UNCLASSIFIED ORNL-LR-DWG 16425 ORNL DATA HIGH QUARTZ = LOW QUARTZ @-Na,BeF, + LIQUID BeF,(HQ) + LIQUID B'~NaBeF3 + LIQUID BeF,{HQ) + 5~ NaBefy B-NaBeF3 + LIQUID BeF,(HQ) + B-NoBefF3 BeF,(LQ) + 8- NaBefy 50 60 70 80 20 Bef, Bef, (mole %) Fig. 2.19. The System NaF-BeF,. o - 600 Lt o o - < o o o = 500 W — a- Na,BeF, + NaF 400 a -Na,BeF, + 8'~NaBeF; 300 ‘a’—NozBeF4 T NaF B -NaBeF3 + y-No,BeF, Y- F4 + NaF 200 NafF 10 20 30 40 UNCLASSIFIED ORNL-LR-DWG 36151 E= EUTECTIC T= TERNARY EULTECTIC Befo TC= TERNARY COMPOUND 542 P= PERITECTIC SUBSQOLIDUS COMPOUNDS SHOWN IN PARENTHESES LIQUIDUS TEMPERATURES ARE IN °C {SNF-LiF- 3BeFy) 320 NaF- BeFy LiF B44 NoF 990 E649 Fig. 2.20, The System NaF-LiF-BeF,. 34 of these systems has sufficiently low liquidus temperatures at low wuranium concentrations to serve as a practical reactor fuel. Liquids suitably rich in UF, for enriching an operating reactor are available, however, with either system. The NaF-LiF-UF , ternary system, although it shows considerably lower liquidus temperatures than those of either binary system, is likewise not useful directly as a fuel. The binary system BeF,-UF, shows no inter- mediate compounds and has a eutectic at a concen- tration of about 1 mole % UF, and a temperature of 530°C. The fernary systems LiF-BeF,-UF, and NaF-BeF,-UF ,, which have been examined in detail at the Mound Laboratory, are shown in Figs. 2.21 and 2.22. It is evident from inspection of these systems that low-melting mixtures of interest as reactor fuels are available over wide composition ranges. PERIOD ENDING OCTOBER 31, 1957 UNCLASSIFIED MOUND LAB, NO. 56-11-29 (REV.) ALL TEMPERATURES ARE IN °C E = EUTECTIC P = PERITECTIC LiF - 4UF, UF,| = PRIMARY PHASE FIELD LiF Fig. 2.21. The System LiF-BeF,-UF ,. 35 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT NaF-2 UF, 7 NaF-6 UF, SNaF-3UF, £ /\ 2 NuF-UF4/' 7 ' 3NaF-UF, A S S NS \ \ R NoF £ / £ 2NaF-BeF, UNC LASSIFIED MOUND LAB. NO. 56-11-30 (REV.) AlLL TEMPERATURES ARE IN °C £ = EUTECTIC PERITECTIC PRIMARY PHASE FIELD A il NaF-BeF, Fig. 2.22. The System NdF-Ber-UF4. Systems Containing 'I"|1F4 A number of ThF ,-containing systems are being actively investigated, but virtually complete phase diagrams can be+ presented at present for only the BeF ,-ThF ,, LiF-ThF ,, and NaF-ThF , systems. The System BeF,-ThF,, - The BeF,-ThF, binary, shown in Fig. 2.23 is very similar to the BeF ,-UF , system. A single eutectic, that melts at 530°C, was found which contains 1.5 mole % ThF4. No intermediate compounds are formed in the system. 36 The System LiF-ThF,. — Four intermediate compounds occur between LiF and ThF . The compound 3LiF.ThF, melts congruently at 580°C, forms eutectics with LiF and 7LiF-6ThF ,, re- spectively, at 570°C with 22 mole % ThF, and at 560°C with 29 mole % ThF,. The compounds 7LiF-6ThF, LiF-2ThF,, and Lil:nfi']'hfi4 melt in- congruently at 595, 755, and 890°C. The phase diagram, which is presented in Fig. 2.24, shows the LiF-ThF, system to have somewhat higher liquidus temperatures than those of its uranium- containing counterpart. A narrow range of TEMPERATURE (°C) TEMPERATURE (°C) 1200 1100 1000 200 800 700 600 500 BeF, 1100 1000 900 800 700 600 500 400 10 20 30 40 50 60 ThF, (mole %) 70 UNCLASSIFIED ORNL-LR-DWG 24554 80 90 Fig. 2.23. The System BeF,-ThF ,. ThE, PERIOD ENDING OCTOBER 31, 1957 LiF-ThF, compositions melt below 1100°F, however, and such mixtures would seem, in principle, to be possible breeder blanket materials. The System NaF+ThF,. — The phase diagram of the NaF-ThF, system obtained from studies in this Laboratory and shown in Fig. 2.25 differs in several respects from one recently described by Emelyanov and Eustyukhin.8 As may be seen in Fig. 2.25, Nan-ThF4 melts incongruently to NaF2ThF, and liquid, and there is only one crystalline form of 2NaF.ThF,. The Russian diagram, in which the liquidus temperatures are in general similar to those of Fig. 2.25, shows NaF.ThF, to be a congruent compound and gives By, s. Emelyanov and Evstyukhin, J. Nuclear Energy 5, 108-114 (1957). UNCLASSIFIED ORNL—LR—DWG 26535 // - \ / A\ s < L_g’ = 'E L.g — = © & & W e L 0 : : 50 LiF 10 20 30 40 50 60 70 80 90 ThF, ThE, (mole %) Fig. 2.24. The Sys"em LiF-ThF4l 37 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT UNCLASSIFIED ORNL—-LR-DWG 18964 1100 / 1000 // oo // W 7~ % 800 / = x w a / = i = 700 \ ,/ \[ l 600 | ' -1 U_cr = ‘ o+ 500 — L <+ ! U . = [ - = — = r—- S L : > < o Ll6 Llc- (% =z =z 400 & ‘ NaF 10 20 30 40 50 60 70 80 20 Thf, ThF, (male %) Fig. 2.25. The System NaF-ThF . two crystalline forms of 2NaF-ThF ,. The ORNL diagram (Fig. 2.25) shows that 4NaF.ThF, decomposes at 568°C. It appears that the Russian workers have interpreted these thermal effects as a transition between two forms of 2NaF.ThF ,, which were suggested by Zachariasen;? one of these two forms of 2NaF:ThF, which may be prepared by Zachariasen's technique, has been shown in this Laboratory to be metastable. The System ThF,.UF,, — The compounds ThF, and UF, form o complete solid solution series for which liquidus and solidus temperatures are not yet completely defined. It appears that the minimum melting point occurs at a concentration of 60 mole % UF4 and a temperature of 1010°C. Ternary Systems Containing ThF4. — None of the ternary systems containing ThF, can be 9W. H. Zachariasen, J. Am. Chem Soc. 70, 2147 (1948). 38 considered to be complete, but sufficient study of the systems has been made to support the following statements. The system LiF-ThF4-UF4 has neither ternary compounds nor ternary eutectics. Most of the area on the phase diagram is occupied by the primary phase fields of the ThF -UF, the LiF.4ThF ,-LiF.4UF ,, and the 7LiF.6ThF ,- 7LiF-6UF, solid solutions. Liquidus temper- atures, in general, decrease toward the LiF-UF, edge of the diagram. The LiF-BeF,-ThF, and the NaF-BeF,-ThF, systems are remarkably similar to their UF ,- containing analogs. In the LiF-BeF -UF , system, for example, two ternary eutectics occur that melt at 350 and 435°C and contain 1 mole % UF ,-52 mole % BeF2 and 8 mole % UF4--26 mole % BeFZ, respectively, and the ternary eutectics for the ThF ,-containing system are nearly identical in composition and temperature. The very great similarity of the LiF-BeF,.UF, and LiF-BeF,- ThF , ternary systems indicates that a relatively wide choice of useful breeder-fuel compositions should be availablie, Systems Containing Plutonium Fluorides The available thermodynamic data suggest that PuF ; should be more stable toward oxidation and disproportionation than UF , is at elevated temper- atures. Conversely, F’UF4 is more strongly oxidizing and probably is more corrosive than UF,. Previous studies indicated that the solubility of UF, in BeF ,-containing melts was less than 2 wt % at 600°C, and similar low solubilities were anticipated for PuF,. [t seemed desirable, therefore, to determine initially the solubility of PuF , in mixtures of interest for long-lived power reactors, The first solubility determinations were made with the 57 mole % NaF—-43 mole % Bef:2 mixture, which is reported to melt at 360°C (ref 10). The experiments were performed in a stainless steel glove box supplied with low-humidity air at a pressure slightly below that of the laboratory. Nickel filtration apparatus similar to that de- veloped for obtaining UF3 solubility data was used, but it was slightly modified to permit use of smaller amounts (6 g) of fused salts. Mixtures prepared by holding weighed portions of purified solvent and PuF_ in a mixed atmosphere of argon, hydrogen, and HF for about 1 hr at the desired temperature were filtered by application of vacuum and inert-gas pressure. The nickel filter, which was welded into the end of a long piece of :?B-in. nickel tubing, was kept out of the melt prior to the actual filtration operation. The data obtained in the first experiments are shown in Table 2,5, 105, M. Roy, R. Roy, and E. F. Osborn, J. Am Ceram. Soc. 36, 185 (1953). Table 2.5, Solubility of PuF, in NaF-BeF2 (57-43 mole %) Temperature Filtrate Analysis (°C) Py (wt %) PuF 4 (mole %) 538 1.10 0.20 600 1.32 0.25 652 2.16 0.41 PERIOD ENDING OCTOBER 31, 1957 The result obtained ot 600°C was lower than was expected from the results obtained at 538 and 652°C, and therefore the filtration was repeated at 600°C aofter a longer equilibration time. The analysis of this filtrate and those obtained at the same temperature with two other NaF-BeF , compo- sttions (64-36 and 50-50 mole %) have not been completed. Several of the mixtures were also examined spectrophotometrically for the presence of Pu4* in dissolved portions of the fused salt, and nearly all samples have been examined with a petrographic microscope mounted in a glove box adjacent to the stainless steel processing box. The plutonium in all the samples examined to date appeared to be in the trivalent state, but the plutonium compounds present in the mixtures have not as yet been positively identified. The compound PuF, is probably stable in contact with nickel under the conditions of these experiments. CHEMISTRY OF THE CORROSION PROCESS F. F. Blankenship L. G. Overholser R. B. Evans G. M. Watson A reactor fueled with a molten salt and operated at a high temperature will be free from corrosion only if the system defined by the fuel and its container is thermodynamically stable under oll conditions of reactor operation. True thermo- dynamic stability of such a system would not seem to be realizable in practice. The rate at which the system can attain equilibrium can be made to be very slow for many cases, however, and reliable long-term compatibility can be obtained through the use of certain flyoride fuels and nickel-base alloys, such as Inconel or the INOR alloys (de- scribed above under the heading ‘‘Metallurgy’’), which contain molybdenum and chromium. Since most of the experience with molten-salt corrosion behavior has been gained with Inconel, this material will be used as the example in the following discussion. When Inconel is exposed at a single uniform temperature to a molten UF | solution, selective leaching of chromium from the alloy occurs. Re- actions with traces of oxidants, such as FeF, or NiF,, that were originally present in the melt or were formed by reactions of the melt with oxides on the alloy are partially responsible. A more important contributor to the attack is the attainment 39 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT of the equilibrium 2UF, + &r—=—0CrF, + 2UF, In simple isothermal tests no chromium is removed from the metal after chemical equilibrium is attained; diffusion within the alloy presumably proceeds until the chromium activity is equal throughout the metal wall. Such attack can be avoided in principle and coan be minimized in practice by careful purification of the melt, deoxidation of the alloy, and the addition of UF, and CrF, at equilibrium concentrations to the melt. Even when the required purification, deoxidation, and pre-equilibration are perfectly attained, how- ever, a corrosion mechanism still exists when UF4-contoining solutions are circulated, as in a reactor, through Inconel systems which contain temperature gradients. If the iron and nickel constituents of Inconel are considered to be inert diluents for chromium (as is true to a reasonable approximation), the long-term corrosion process can be described simply. When the fuel mixture is circulated rapidly through the polythermal metal system, uniform concentrations of UF,, UF,;, and CrF, are maintained throughout the fluid. The concentrations are those which will satisfy the equilibrium constant VCer’VfiFs NCer'NaF Ka = Ky‘KN = - 3 VCr'YElF4 NeoNoe at N., of the unaltered alloy and at an inter- mediate temperature which depends on the geometry and the temperature profile of the system. In this expression, K_ is the equilibrium constant based on free-energy data, K_ is the activity coefficient quotient, Ky is the equilibrium concen- tration quotient based on mole fractions, y is the activity coefficient based on pure substances as the standard state, and N is mole fraction. The system must, however, at ultimate equilibrium, satisfy K_ at all temperatures and locations in the system with the fixed concentrations of CrF,, UF3, and UF . Since Ky for the reaction shown increases with increasing temperature, the chromium concentration in the alloy surface is diminished at high temperatures and is augmented at low temperatures. These surface concentrations are maintained at the values required by the 40 equations by diffusion of chromium from or into the bulk metal. The equilibrium concentration of chromium at the surface of the alloy is accordingly fixed by the temperature drop in the circuit and the temperature dependence of the equilibrium constant for the chemical reaction. The rate at which chromium is transferred from hot to cold sections of the apparatus is determined by the rate of chromium diffusion in the alloy; this rate depends strongly on the absolute temperature of the metal. The flow rate of the fluid cannot offect the rate of this diffusion-controlied process. If the equilibrium constant changes sufficiently with temperature, the temperature drop is suf- ficiently large, and the activity of chromium in the alloy is sufficiently high, the activity of chromium necessary to satisfy the equilibrium constant at low temperatures rises to near unity, and crystals of chromium can form in the low- temperature regions of the system at the expense of depletion of chromium in high-temperature regions of the system. Inconel systems with high- and low-temperature regions at 1600 and 1200°F in which NoK-KF-LiF-UF , mixtures are circulated and Hastelloy-X systems operated at similar temperatures with the fuel mixture NaF- ZrF ,-UF , belong to the class of systems in which chromium deposits occur. The deposited dendritic crystals of chromium tend to plug the equipment at a rate controlled by diffusion of chromium from high-temperature regions of the system. Such combinations are obviously unsuited for reactor application. If, however, the temperature effect on the equilibrium constant is small, as is true in many ZrF ,-containing systems, the equilibrium is satisfied at all useful temperatures without the formation of crystolline chromium deposits. The corrosion occurs, without restricting the flow passages, at a rate controlled by diffusion of chromium from the surfaces into the interior of the cool region, Measurements of the rate of loss of radiotracer chromium from on inert fluoride melt exposed to Inconel showed the over-all diffusion rate for chromium into Inconel to be 3.7 x 10~15 ecm?/sec at 700°C; the activation energy!! of the process is 66 kcal/g-at. These numbers are, of course, independent of the fuel used. ”G. M. Watson and R. B, Evans {unpublished data). Relatively little information is available re- garding the effect of temperature on the equilibrium constant in BeF,-containing systems, but pre- liminary values obtained for an LiF-BeF, (48-52 mole %) solvent are compared in Table 2.6 with those for Nc:F-ZrF4 solvents for which many corrosion tests have been made. The data indicate that the temperature dependence of K, is sig- nificantly greater for LiF-BeF, systems than for the NaF-ZrF, systems which do not result in chromium deposition in cool zones of dynamic systems. |t appears virtually certain, however, that with small chromium activities, such as those in Inconel and the INOR alloys, and small temper- ature drops, such as those contemplated for most long-term power reactors, chromium deposition will not result if fuel mixtures based on the BeF,- containing system are used. Investigations of the effect of fuel composition on the response of Ky with temperature in BeF,-containing systems of reactor interest are continuing. FISSION-PRODUCT BEHAVIOR J. H. Shaffer R. A. Strehlow N. V. Smith W. T. Ward G. M. Watson When fission of uranium occurs in a molten fluoride solution, the fission fragments must originate in energy states and ionization levels PERIOD ENDING OCTOBER 31, 1957 very far from those normally encountered. These fragments must, however, quickly lose energy as a consequence of collisions in the melt and come to equilibrium as common chemical entities. The valence states which they will ultimately assume are determined by the requirements of redox equilibria among components of the melt and the alloy container and by the necessity for cation- anion equivalence in the melt. The equilibrium valences of several of the fission products are not yet well established, but it seems certain that xenon and krypton will appear as elements and that the alkalies, the alkaline earths, zirconium, and the rare earths will appear as ions of their customary valence. A systematic study of the solubility of these materials in NaF-ZrF, and some other mixtures has been made, and the status of knowledge concerning behavior of these materials is summa- rized below, Solubility of the Noble Gases The solubilities of the noble gases in molten NaF-ZrF, (53-47 mole %) obey Henry’s law, increase with increasing temperature, decrease with increasing atomic radius, and are not appreciably affected by the presence of 10 wt % UF,. The Henry’s law constants at 600°C (in moles of gas per cubic centimeter of NaF-ZrF, Table 2.6. Equilibrium Quotients for the Reaction ZUF4(a’) + Cr{c) fi ZUFs(d) + Cer(d) in Molten Fluorides Reaction Medium Temp:rdture Initial UF4 Concentration KN (- C) (mole %) NaF-ZrF , (53-47 mole %) 600 4.0 (1.3 + 0.1) x 10~4 800 4.0 (2.9 +0.1) x 104 NaF-ZrF , (50-50 mole %) 600 4.1 (3.2 +0.8) x 10~4 800 4.1 (5.9 + 1.8) x 104 NaF-ZrF , (59-41 mole %) 600 3.7 (1.5 +0.2) x 10-3 800 3.7 (1.7 + 0.7) x 10~° LiF-BeF, (48-52 mole %) 550 1.5 2 x 10~6+ 650 1.5 1.9 x 10=5+ 800 1.5 7 x 10~ *Preliminary values. 4] MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT mixture per atmosphere) are 21.6 x 10-8, 11.3 x 108 5.1 x 10-8 and 1.9 x 10-8, respectively, for helium, neon, argon, and xenon. The heats of solution are, in the same order, 6.2, 8.0, 8.2, and 11.1 kcal/mole. The increase in solubility with increasing temperature is in marked contrast to the behavior of HF in this solvent; the Henry's law constant for HF at 600°C is 1.23 x 10=3, while the heat of solution is —4.5 kcal /mole. The solubility of krypton has not been examined, but there is little reason to doubt that it dissolves in NaF-ZrF, (53-47 mole %) at 600°C to the extent of 3 £1 x 10~8 moles/cm3/atm with a heat of solution of 9.5 £ 1.5 kcal/mole. Although experimental determinations are under way, no data have been obtained as yet regarding the solubility of the noble gases in BeF,-con- taining systems. It is anticipated that the solubility of xenon will depend to some extent on solvent composition, (Helium is roughly half as soluble in the LiF-NaF-KF eutectic as in the NqF-ZrF4 mixture described above.) However, it is most unlikely that the solubility will be so different from those listed above as to prevent the removal of krypton and xenon from an operating reactor fueled with a BeF ,-containing mixture. Solubility of Rare-Earth Fluorides The rare-earth trifluorides are sparingly soluble in NaF-ZrF, (53-47 mole %) mixtures. In this solvent the solubility of CeF, is 2.3 mole % at 600°C and 4.3 mole % at 800°C. The product of the solubility (in mole %) and the third power of the cationic radius is constant to within +3% for CeF, LaF, and SmF,. While the solubility changes with the NaF-to-ZtF , ratio, the saturating phase is apparently the simple rare-earth fluoride - (that is, CeF,) over a relatively wide NaF-ZrF concentration range. The solid phase in equilibrium with a saturated solution of two rare-earth fluorides in the NaF-ZrF (53-47 mole %) mixture is a nearly ideal SOFié solution of the rare-earth fluorides. The equality N x N s0 C°F3(d) L°F3(.s's) CoF3 LaF X Neor SE F 3(d) *T3(ss) | {(where N is mole fraction, the subscripts d and ss refer to liquid and solid solution, and $9 is solubility of the single fluoride) is valid at each 42 temperature. This behavior of rare-earth mixtures indicates that partition of fission-product rare earths between the liquid and solid CeF_ in a cooled external circuit should deplete the ?uel of nuclearly harmful species. The reactor fuel would, of necessity, be saturated with CeF, at the reprocessing temperature. |f the fuel were NaF-ZrF4-UF4, considerable CeF, (perhaps 2 mole %) would be carried in the circulating stream. Preliminary experiments on the solubility of CeF, in an NaF-BeF, (57-43 mole %) mixture produced the data presented in Table 2.7. A comparison of the values for the NaF-BeF, and NaF-ZrF, mixtures, as read from smooth curves through the experimental points, is given in Table 2.8, It is obvious that the solubility of CeF, is considerably less in the BeF ,-containing mixture and that it is more temperature sensitive. The saturating phase has been shown to be the simple trifluoride CeF,. No data have yet been obtained with mixtures of rare earths, but it is anticipated that they will Table 27. Solubility of CeF, in NaF-BeF, (53-47 mole %) Temperature Solubility of CeF °0) Ce™™ (Wt %) CeFy (mole %) 718 2.58 0.83 639 1.29 0.41 517 0.39 0.12 434 0.15 0.047 Table 28. Comparison of Solubility of CeF3 in NuF-ZrF4 (53-47 mole %} and in NuF-Ber (53-47 mole %) TempeerUre SO'UbIIH’Y (mole % CEF3) (°C) In NaF-ZcF In NaF-BeF, 750 3.2 1.1 700 2.7 0.65 600 1.9 0.29 550 1.7 0.17 exhibit the nearly ideal solid solution behavior found in the NaF-ZrF, system. Such behavior, along with the much reduced solubility in the BeF, system, should make the fission-product partition process quite attractive. Further studies with NaF-BeF, (53-47 mole %) and other BeF,- bearing systems are under way. PRODUCTION OF PURIFIED MOLTEN SALTS J. P. Blakely C. R. Croft F.A. Doss During the past quarter, 23 batches (described in Table 2.9), including eleven different compo- sitions, were prepared. It is likely that some increase in process efficiency or in equipment capacity will become necessary if the demand for these materials continues as at present. METHODS FOR PURIFICATION OF MOLTEN SALTS J. P. Blakely C. R. Croft F. A. Doss Production efforts have been devoted, in- creasingly, to the preparation in pure form of the various BeF ,-containing mixtures needed for PERIOD ENDING OCTOBER 31, 1957 engineering tests in corrosion loops and similar apparatus. The melts have been processed by using slight modifications of the procedure used to purify NaF-ZrF, and NaF-ZrF ,-UF , mixtures. In this method the blended raw materials are charged into reaction vessels of copper-lined stainless steel that have capacities of 5, 10, or 50 tb. The materials are melted and brought to a temperature of 800°C under a flowing atmosphere of HF to remove water with @ minimum of hy- drolysis. The 800°C melt is sparged with dry hydrogen to reduce penta- and hexavalent uranium compounds, sulfate, and extraneous oxidants. Subsequent sparging with anhydrous HF serves to volatilize the HCl and H,S and to convert to fluorides any oxygenated compounds in the mixture. This HF treatment, however, contaminates the melt with some copper fluoride through reaction with the container metal. A final 15- to 24-hr sparging with hydrogen serves to remove the CuF,, along with any FeF, and NiF, that was originally present in the melt, by reducing the compound to metal, The purified melt is then forced under helium pressure through a transfer line that contains a filter of sintered nickel and a sampler into a clean Table 2.9, Materials Processed During the Quarter Composition Constituents (mole %) Number No. UF, ThF, BeF, LiF NaF KF ZeF, of Batches C-12 46.5 1.5 42.0 3* C-84 38.0 35.0 27.0 2% C-122 1.0 57.0 42.0 1 C-123 1.0 46.0 53.0 2% C-124 7.0 35.0 58.0 1* C-125 0.5 0.5 46.0 53.0 2% C-126 1.0 46.0 53.0 4* C-127 7.0 35.0 58.0 1 Cc-128 29.0 71.0 3 C-129 4.0 55.3 40.7 3* C-130 1.0 37.0 62.0 1 Total 23 *Includes one large batch (> 13 kg); all other batches were small (<6 kg). 43 nickel receiver vessel. By manipulation of appropriate valves and connections the cold receiver is detached from the production assembly and attached for storage to a manifold containing helium under a slight positive pressure. The BeF ,-containing mixtures are substantially more difficult to purify than ZrF ,-containing melts. The difficulties are, in general, a conse- 44 quence of the fact that BeF, contains up to 2000 ppm S, as sulfate, and up to 600 ppm Fe, probably as FeF;. The high sulfur content causes numerous equipment failures, especially failures of the nickel or nickel-alloy dip lines and transfer lines. Studies of improvements which can be made readily in the purification process are under way.