MARTINMARIET TA ENEREY SYSTEMS LIBRARMES LRGN ORNL-1692 This document consists of 154 pages. Copy fi&‘ of 264 copies. Series A. Contract No, W-7405-ang-26 AIRCRAFT NUCLEAR PROPULSION PROJECT QUARTERLY PROGRESS REPORT For Period Ending March 10, 1954 R. C. Briant, Director W. H. Jordan, Associate Director A. J. Miller, Assistant Director A. W. Savolainen, Editor DATE ISSUED OAK RIDGE NATIONAL LABORATORY Operated by CARSBIDE AND CARBON CHEMICALS COMPANY A Division of Union Carbide and Carbon Corporation Post Office Box P Quak Ridge, Tennessee NMABTIN MARISTTA ENERGY SYSTEMS LIBRARIES 1R 3 445k 0349kL23 ¢ o R i VPN DA W= "'2?’?0.“‘-—."”.'“.3:9?’?”?'3”’.”‘??’??’.'”??U?PP?UPP?T”?U?F”T“PF”PQ?UP INTERNAL DISTRIBUTION . A. Mann ORNL-1692 Progress M. Adamson 43. W. B ficDonald Wffel . J. L Meem R. B&yock 45. A 4. Miller J. Barto ¢ 7. Morgan S. Bettis JE. J. Murphy S. Billington ‘ #J. P. Murray (Y-12) F. Blankenship # G. J. Nessle P. Blizard £ ). P. Patriarea A. Bredig _5]. H. F. Poppendiek C. Brignt }'" 52. P. M. Reyling R. Bruce . & 53. H.W. Savage D. Callihan %\ 54, A, W, Savolainen W. Cardwell 55. E. D. Shipley E. Center i 56. Q. Sisman A. Charpie N 57. G. P. Smith H. Clewett " 58, L. P. Smith (consultant) E. Clifford & 59. A. H. Snell B. Cottrell ¥ %0. C. L. Stors G. Cochran N C. D. Susano D. Cowen & 62.%). A. Swartout L. Culler 63. M H. Taylor B. Emlet (K-25) b4. J. W Trice K. Ergen & 65. E. R%Yan Artsdalen P. Fraas 66. F. C. Wgnderl.age R. Grimes 7 67. J. M. Wale E. Hoffman & 68. A. M. Weierg . Hollaender 69. J. C. White % . S. Householder & 70. G. D. Whitmari T. Howe 71. E. P. Wigner (cdgsultant) .wmqug‘ 72. G. C. Williams 73. J. C. Wilson W. Keilhol 74. C. E. Winters i 75-84. ANP Library 85. Biology Library 86-90. Laboratory Records Depurf et . 91. Laboratory Records, ORNL R Al 92. Health Physics Library 93. Metaliurgy Library WL ivingston 94. Reactor Experimental Lyon Engineering Library . D. Manly 95-97. Central Research Library 98, " 99. 100, 101, 102. 103-114. 115, 116, 117-125. 126. 127-132. 133. 134, 135-136. 137-144. 145, 146. 147. 148. 149. 150. 151, 152-155. 156. 157. 158-160. 161-164. 165. 166. 167. 168-175. 176-177. 178-179. 180. 181. 182. 183-185. 186-189. 190, 191. 192-193. 194.195, 196-201, 202, 203-210. 211-222. 223-232. 233-234. 235. EXTERNAL DISTRIBUTION Air Force Engineering Office, Oak Ridge Air Force Plant Representative, Burbank Air\Force Plant Representative, Seattle Air Porce Plant Representative, Wood-Ridge ANP Pxoject Office, Fort Worth 4 Argonne Yational Laboratory (1 copy to Kef# Armed Forxes Special Weapons Project | Armed Forcds Special Weapons Project, Atomic Energ ,Commission, Washingtg Battelle Memorl Institute : Brookhaven Natioka!l Laboratory Bureau of Aeronautixs (Grant) Bureau of Ships California Research and®Develog@ent Company Carbide and Carbon Chem) ols Chicago Patent Group Chief of Naval Research Commonwealth Edison Co Convair, San Diego (C, Curtiss-Wright Corporatigf Department of the Navyg Op-362 Detroit Edison Compal" duPont Company, Augbsta Anderson) tishington (Gertrude Camp) t. Col. M, J. Nielsen) elm Foster Wheeler C General Electric £ompany, ANPP General ElectrigtCompany, Richland Glen L. Murhnf{:nmpany (T. F. Nagey) Hanford Operghions Office lowa State u'llege Knolls Atorfic Power Laboratory Lockland Area Office fs Scientific Laboratory jsetts Institute of Technology (Kaufmann) s Laboratory (WADC) (Col. P. L. Hill) Monsarfto Chemical Company Natighal Advisory Commiitee for Aeronautics, Cleveland (A.%ilverstein) Natifnal Advisory Committee for Aeronautics, Washingien '5';5?_1‘ Nayal Research Laboratory ': atv York Operations Office e brth American Aviation, Inc. Auclear Development Associates, Inc. Patent Branch, Washingion Phillips Petroleum Company Powerplant Laboratory (WADC) (A. M. Nelson) Pratt and Whitney Aircraft Division (Fox Project) Rand Corporation (1 copy to V. G. Henning) San Francisco Operations Office 2ad loglcal Defense Laboratory 239-240. University gy 'forma Radlohon l.aboratory, Berkeley Qhouse Electric Corp s 250-264. Technical Information Service, OOI( Ridge 244249, Wesf vi Reports previously issued in this series are as follows: ORNL-528 ORNL-629 ORNL.-768 ORNL.-858 ORNL-919 ANP-60 ANP-65 ORNL-1154 ORNL-1170 ORNL.-1227 ORNL-1294 ORNL-1375 ORNL-1439 ORNL-1515 ORNL.-1556 ORNL-1609 ORNL-1649 Period Ending November 30, 1949 Period Ending Februcry 28, 1950 Period Ending May 31, 1950 Peried Ending August 31, 1950 Period Ending December 10, 1950 Period Ending March 10, 1951 Period Ending June 10, 1951 Period Ending September 10, 1951 Period Ending December 10, 1951 Period Ending March 10, 1952 Period Ending June 10, 1952 Pericd Ending September 10, 1952 Period Ending December 10, 1952 Period Ending March 10, 1953 Period Ending June 10, 1953 Period Ending September 10, 1953 Period Ending December 10, 1953 FOREWORD This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL records the technical progress of the research on the circulating-fuel reactor and all other ANP research at the Laboratory under its Controct W-7405-eng-26. The report is divided into three major parts: |. Reactor Theory and Design, ll. Materials Research, and lll. Shielding Research. The ANP Project is comprised of about 300 technical and scientific personnel engaged in many phases of reseorch directed toword the achievement of nuclear propulsion of aircraft. A considerable portion of this research is performed in support of the work of other organizations participating in the national ANP effort. However, the bulk of the ANP research at ORNL is directed toward the development of a circulating-fuel type of reactor, The nucleus of the effort on circulating-fuel reactors is now centered upon the Aircraft Reactor Experiment — a high-temperature prototype of a circulating-fuel reactor for the propulsion of aircraft. The equipment for this reactor experiment is now being assembled; the current status of the experiment is summarized in Section 1 of Part I. The supporting research on materials and problems peculiar to the ARE ~ previocusly included in the subject sections ~ is now included in this ARE section, where convenient. The few exceptions are referenced to the specific section of the report where more detailed information may be found. The ANP research, in addition to that for the Aircraft Reactor Experiment, falls into three general categories: (1) studies of aircraft-size circulating-fuel reactors, (2) moterials problems gssociated with advanced reactor designs, and (3) studies of shields for nuclear aircraft. These phases of research are covered in Parts |, 1, and i, respectively, of this report. CONTENTS FOREWORD-..-V -------- ® 2 5 & 5 ¢ 02 5 0 B B S e A Ao SUMMARY ......... tes s ee e s s e s e st e s e e e PART I, REACTOR THEORY AND DESIGN 1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT . The Experimental Reactor System. . v v v v vs i v tonoensseans Reactor Physics o v v v v e nnnrsnnirnnnosenceaaas Neutron femperature .+ oo o v v s v s v ao v s s o cesaers » Criticality in the ARE fill-and-flush tank. ..... e ers e The ARE critical experiment. . .. v v v v v it v v o Production of Fuel Concentrate . . . v vttt v e v nnenneen PUMPS + i ittt i it et t e s e s s e Acceptance test program « o e v s v v s s e s s o s e s Pumpassembly. . ..o v v i i it iieinen bee e impeller fabrication ond testing . .. .. e i ae e Fuel Recovery and Reprocessing ... v v v i vt Solvent extraction processing + v . s a4 et es e 2. EXPERIMENTAL REACTOR ENGINEERING , o0 v v vvs . In-Pile Loop Component Development . ... .. e e e s Vertical-shaft sump pump + v« v e vt e e s v v v oo nusossas Pump with a centrifugal seal .. ... . 000 ceoae s Horizontal-shaft sump pump « v v e v e v v v s v aen ceoan e Heat Exchanger and Radiator Tests ..o v v vvv s v e e e o Heat exchanger test . v v v v v vt i s v e s s ennssosssas Sodium-to-air radiafor test . . v v v v v s v s et a e e Forced-Circulation Loops + vt v v v vt e vt niovs oo Alr-cooled loop. « v v o vt v i it st ts i e st i et Beryllium-NaK compatibility test v v v oo s v v v v vv s v s o Bifluid heat transfer 1oop + « + c v v e v vt vt st v v s nese High- Temperature Bearing Development . . . ... et e e Materials compatibility tests. o v v v vt i et v i e enen Bearing tester design . . s o v vss s v o enncesnonneas Rotary-Shaft and Valve-Stem Seals for Fluorides « oo oo 0 v v Spiral-grooved graphite-packed shaft seals. . .. .00 o0 Graphite-BeF , packed seal .. . .o o vnvii oo Varingseal o0 i e viii v e P ecrseretean s Chevronseal . ... .. cee s Gt e et aansarsaens . Graphite-Baf, packed seal . ..... .0, Shaft seal for in-pilepump . . c o0 v v b v ee s I Packing penetration tests .. o0 oo v e asne s oo Valve-stem packingtests ... ieie oot vscoenocons Self-Bonding Tests of Materials in Fluoride Mixtures . ... .. Removal of Fluoride Mixtures from Equipment « v v v v v s v v 0 0 3. REFLECTOR-MODERATED REACTOR. . o v ev v e nv o Effects of Reactor Design Conditions on Aircraft Gross Weight Core Flow Experiment ooooooooooooooooooooooo ....................... FUeIdISSOiUfiono llllllllllllll ¢ o 3 s 0 8 0 8 A A b e o e ....................... lllllllllllllllllllllll ....................... ----------------------- ooooooooooooooooooooooo ooooooooooooooooooooooo ooooooooooooooooooooooo ooooooooooooooooooooooo 6. Reactor Calculations ... ... .... Gt s n e s e e s e s e et e e e e L e e e s e s e e s an e Reactor Dynamics . o .. . . S e e e s e s e e e E e s e e s et e adeseatsa e asasetbancnoa Computational Techniques v v v vt v v v an et e e e s e e e Beryllium Cross Sections « v v v v ittt ittt ittt ittt it e nnessosnnsnosnoenesnss Chemical Processing of Fluoride Fuel by Fluorination .. ..... et i e s et e CRITICAL EXPERIMENTS . ... ...... C e et e et Supercritical-Water Reactor . v v v v v v v v v e v nnnne Afr-Cooled Reactor v v v v ittt ittt i ot ittt nnenuetoseostoetsnnessenenens . Reflector-Moderated Reactor .. . ... .. ... e Gttt e e e e e e PART . MATERIALS RESEARCH CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS . . . i it i i ittt et evnnononnnonss Thermal Analysis of Fluoride Systems. ... .. et e e e NaF-LiF-ZrF -UF oo oo oo L et e e s e e e e e s ceeas NaF-BeF -ZrF -UF, ... oo, et e e e e e e e NaF-LiF-BeF,-UF, ..... e e e e et e e et e e Thermal Analysis of Chloride Systems ., . ... ettt e e RbCIUCH, o.vvvvint e e NaCl-ZrCl, oo e v ot e r e Quenching Experiments with Fluoride Systems . ....... et e et et e e e e e NaF-UF, ........ C et e e et e e e e e e C e e e s e ce s NaF- ZrF ....................................................... Filtration Ana!ysus of Fluoride Systems ................................ heee A NaF-ZrF4 UF, «ooviea C e et i et i e ae st e e NaF-KF-ZrF ~UF o i i i e e e e Production of Purified Fluoride Mixdures . . . it it ittt it it en v vnnnnnean e L.aboratory-scale production and purification of fluoride mixtures + o v oo v v v v e e e v vnoe . Purification of fluoride mixtures for phase studies. v v v v o vttt et o ittt seenonnasens Experimental production facilities ... ..... e e e e e et Production-scale facility . . ... .. ... . oL, e e e e e Purification and Properties of Hydroxides . ... ..o o i i i i iii it et Purification of hydroxides ... .. C e e Ch et te e e e Reaction of sodium hydroxide with carbon . .. .. et e e e et Chemical Reactions inMolten Salts . . .. .. ittt ittt ittt it ittt eeteaannsan Chemical equilibrig in fused salts . ... ..... ® 55 s 04 e na e nsaeeenss Chetaeas Formation of UF, in NaF-ZrF melts o o i i iiiieiiaenaany Treatment of molten NaZrF . with strong reducing agents . .......... C e e et Reduction of Nif, and FeF, by hydrogen .. ... it iiii i Spectrophotometry of supercooled fused salts . . oo v v it it e et i i e ceen EMF measurements in fused salts ... ... P e e e c e a e e e CORRQOSION RESEARCH . e ee et s e e Fluoride Corrosion in S’rcmc. and T:I'rmgni"urnace Tests. v v v e h e e et Cee s Effect of methods of manufacturing the fuel on the corrosionof lnconel . . .. o oo oot Corrosion of ceramic materials . ....... e et e s e e e e s Effect of graphite additions + v v v v v v e v i e e s et e et e Etfect on nickel and Inconel rods exposed to fluoride mixtures in@ir « ..o v e s enn e Thermal Convection Loop Designand Operation v v v v i ie v ittt et noetnnnasoness . Fluoride Corrosion of Inconel in Thermal Convection Loops « v v v v v vttt ettt ot e v v vnan Effect of exposure time on depthof attack . ... oo oo v v v vt e e it e . 39 40 40 41 42 45 45 45 45 49 49 50 50 50 50 51 52 52 52 53 54 4 >4 55 55 55 55 35 56 56 36 56 56 60 61 62 62 63 65 66 66 66 66 67 67 68 68 Effect of exposure time on chemical stability of the fluorides . ... Effect of exposure time on corrosion by nonuranium-bearing fluorides Effect of variation in uranium concentration « v v v o v oo o s e s v oo Effect of temperature v v vt vt i v vt nranereonnsens Effect of variations in loop size and composition .. ..o vvvensn Effect of additions to the fluorides, . . .. v v v i v vt v et i v e e Corrosion of various metal combinations . . v v vv v v i vt ev e Fluoride Corrosion of Stainless Steel, lzett lron, and Hastelloy Loops Cotrosion by High-Velocity High-Temperature Fluorides . .. o0 v v u Forced-circulation corrosion Joop « v v v v v s v e v vt cn o nn e vens Oscillating-furnace studies . oo s v v et s s it i e s e e nen Static Tests of Brazing Alloys in Fluorides and Sodium ,........ Mass Transfer in Liquid Lead .. ... Gt e et sttt e e METALLURGY AND CERAMICS .. i iiiiicinnnnnaan Mechanical Properties of Metals « v v i v v ittt i vt v v oo v oo Stress-rupture tests of Inconel + . oo v ettt i i ce e Creep Tests of Columbium ... oo vttt ii e Brozing Research . oo vttt tn ittt e i nnenneeanonsosnns . High-conductivity radiator fins « v v v e vt cn vt v i annnnesen Fluoride-to-air radiator . v v v o e v v v v n et e R Helium-afmosphere brOZing ® ® & p © B w8 B 4 % A F & ¢ & 5 F P & P s e Nickel-phosphorus brazingalloy v cvevvi i iva i ¢ 2 & 2 8 ¥ * 4 » & & & & 4 2 & @ * 2 9 . & . o 0 . & 4 % & 3 & & ooooooooooo & & & @ * o 2 @ - "™ e @ . *» 2 @ L » L ) . 9 b 0 . e . *» & & 8 2 * 4 * + @ . e & & . » Nickel-germanium brazing alloy + v v vt o v it it ittt e it ittt i i e e e High-Conductivity Metals for Radiator Fins . ... .00 v ceeen Diffusion barriers . .o v oo v v e v v uns C it e e s a e e ae s Diffusioncouplesll’I.......-.I..lfli..l..l...ll...l Commercially clad copper v v v v v v i v iva s e esaaen oo Fabrication of Special Materials .. v v it vt i in ittt tranonns Clad columbium disks « v oo v it i vttt i it ea e v nosns s Clad molybdenum ond columbium tubing ... v v e v i e v e v un Drawn tubular fuel elements . . v o v vt it i e st ennvass CeramicReseGrch¢..|I|..l...l...l'......I...'l.l...l e 5 & e 2 . .9 * 3 & & F & 3 4 @ - . ¢ & P * a Ceromic container for fluoride fuel « o . v v v v vt il ittt i it s i e Ceramicpump bearings .. v ittt ittt snt et ossnsnsosnssasonasesssasass Boron carbide-iron cermets v v o v v ot v v st s s v oot os s s o st e st aassees e HEAT TRANSFER AND PHYSICAL PROPERTIES ...t eiinenernoans . Physical Properties Measurements « . vt v v v et et i st s annunsesonsnnoason C e e Heat capacitys o v v o v v ov e s s vooneens e e et ee s cosess v Thermal conductivity @it vttt vt it et ten st onensoaassasocnsssansesaasenns Density, viscosity, and surface tension « . v v v oo v e e s v s oo Electrical conductivity of moltensalts .. ..o cv v i v v a Fused-Salt Heat Transter . . v oo v i it ittt iiet et anon s Hydrodynamics Research v e v e v v et to vt nssnnsonsennsas Circulating-Fuel Heat Transfer . oo v v i vt ie v v it nnsonns RADIATIONDAMAGE . ........ ... ven.. et e Radiation Stability of Fused-Salt Fuels , . o v v e v v i v v i onns Petrographic examinationof fuels . ... v eivviiveii v Temperature control in static capsule tests ¢ v v v vv v oo Miniature circulating loops « o . . . .. e ra e e e e e oooooo * 8 4 8 8 & B @ . 8 . & 8 . 8 s 5 2 B . S & 8 3 & *» P . * & & . ¢ o @& ? & & & 2 @ L f s @ . ® 0 9 ® a4 & & 3 0 & 2 @ 4 % & 8 e & @ 8 & . - - - . & 69 69 70 71 71 72 73 73 74 74 75 75 77 84 84 84 87 87 87 89 89 90 91 9 91 21 92 92 93 93 93 93 94 94 95 @5 95 95 97 98 98 98 99 102 102 102 102 103 xi 10. 11, 12. 13. 14. 15. xii Stress-Corrosion Apparatus 103 Creep Under Irradiation « v vt vt ittt ettt nosnosennesenasasoneossssssonsses 103 ILITR Fluoride-Fuel Loop v v v v v v v e it it it sttt s an s s sesnsansasonsaseesnas 105 ANALYTICAL STUDIES OF REACTOR MATERIALS . o v vt i i it it i e s s e s ansnn 107 Analytical Chemistry of Reactor Materials . o0 v st i it i s it ittt ittt s it vonanns 107 Oxidation states of chromium and iron in ARE fuel sclvent, NaZrF, . oovvvviv v oot 107 Oxidation states of chromiumand uranium in ARE fuel solvent, NaZrF . ..o o vivvvinn s, 108 Separation of UF fromUF 5 . oo v ii i it i s c i i 109 Fluoride Fuel Investigations « o v v v v i vttt it ne st s s aornseesastonoananensans 112 Summary of Service Chemical Analyses . o v v it i ittt et il it e i e e s 112 PART I, SHIEL.DING RESEARCH SHIEL DING ANALY SIS i ittt it ittt sttt nnetsassnstsossssonesanonss 115 Estimates of Removal Cross Sections Based on the Continuum Theory of the Scattering of Neutrons fromNuclei « v vov v i ittt il i ittt et i s 115 Critique of Lithium Hydride as aNeutron Shield . ... .o v i i i it ittt i it e it e 116 LID TANK FACILITY i it it i ittt o e tistetnsnsnssssnasnsssssnsenssanssaees 117 Air-Duct Tests o v it it ittt ittt tstnnosotneesssssensessnsnaossossnsasas 117 Shielding Tests for the Reflector-Moderated Reactor . v oo v v e it i it enne oo 117 Lid tank dose measurements corrected to designed reflector-moderated reactor shield . . .. ... 118 Shield weight calculations . v v v v i ittt it il ittt i it e it s s s s s oeasss 120 Removal Cross Sections « v o v v oo ot s ot s s et oo tnossssanansnessasaonsssssssons 122 INSIruUmMENtatioN v v v v it o o ot v o oot s e b ettt s e e e et et s 122 BULK SHIELDING FACILITY 4 o vt et v it ettt it e st seoesnsesosesasonssensnnans 123 G-E Test Fixture Duct Experiment « . o i vt it i ittt ittt nnsnnnennoneeeonnas 123 GE-ANP Shicld Mockup Experiments « « v v v o v oottt aes toetoorsassssssennssososeos 124 TOWER SHIELDING FACILITY 4t vt ittt ettt st s nnssoennasssassasasnnsoens 135 Facility Construction. o « v v v v vttt ittt ot oot tessoetneesnsessssssnssossssas 135 Radiation Detection EQUIPMEnt v v o v v vt ot c s vt v o s sasetnesstrsnsveossosasssno 135 Thermal-neutron fluX . o o v vt it it i it o st s ettt stosansssnssoeanenossssssnaas 135 Fast-neUIron dOSE . v o v v v et e o vt st e st s st s na s oo sasensoesnnsossansansosass 135 GOMMGA-TAY dOSE ¢ 4 v s v s v st s s o s oo s s sossnoonssoosssssensssososassssenss 135 Other gamma-ray deteCiors & v v v v v vt s it ot s ettt ot oo st onesanssssnssasnsssssase 136 Isodose Plotter v . vt o vttt s ot ot et s o s caesesoassessseasnensnsenonesoes 136 Estimate of Neutron and Gamma Radiation Expected ot the Tower Shielding Facility .. ... ... 137 PART IV. APPENDIX LIST OF REPORTS ISSUED DURING THE QUARTER . . i i it v ittt v e e s e s s oo sann 141 W ANP PROJECT QUARTERLY PROGRESS REPORT SUMMARY PART I, REACTOR THEORY AND DESIGN Assembly of the Aircraft Reactor Experiment is proceeding, the fuel and the sodium circuit pumps are being assembled and tested, and the methods for recovery and reprocessing of the fuel are being developed (Sec. 1). The reactor was installed in the pit, and the fuel and the sodium circuits were completed except for the pump assemblies. An examination of a spare heat exchanger, however, confirmed doubts as to the reliability of these units, and they are now being rebuilt. The oper- ations manwal has been completed, and training of the operating personne! is continuing. The production of the ARE fuel concentrate was com- pleted, and the analytical results were verified. An analysis indicated that the fill-and-flush tank would be subcritical if it were necessary under emergency conditions to admit the fuel to it. The ARE pumps are being inspected and tested, and the pump tanks and covers have been made avail- able for installation. The faobricated impellers designed to replace the faulty cast impellers were found to be satisfactory. The experimental reactor engineering work in- cludes the development of pumps for circulating fluoride mixtures in in-pile loops, the design and construction of forced-circulotion loops for cor- rosion testing, developmental work on high-temper- ature bearings, and tests of rotary-shaft and valve- stem seals for fluoride mixtures (Sec. 2). The pumps being developed and tested for circulating fluoride mixtures in in-pile loops for studying radiation damage of the fluoride mixture and the loop material are of two types: pumps for use inside reactor holes and pumps for use external to reactor holes. Work is under way on a vertical- shaoft sump pump, an in-pile pump with a cen- trifugal seal, and o horizontal-shaft sump pump. Since the in-pile pump with a centrifugal seal is not sensitive to its orientation with respect to gravity, it also shows promise as an aircraft type of pump. Limited test operations have been com- pleted on the l-megawatt regenerative heat ex- changer which employs some of the design features of the proposed heat exchanger for the reflector- moderated aircroft reactor. A sodium-to-air radi- ator is used as the heat dump for this heat ex- chonger test, and radiator performance and en- durance data are being obtained. Two all-inconel corrosion testing loops and one beryllium and Inconel loop are being developed that will provide high liquid velocities and high temperature differ- ences between the hottest and coldest parts of the circuit, Design work and developmental in- vesfigations are continuing on o hydrodynamic type of bearing which will operate at high temper- atures in o fluoride mixture. The work thus far has been primarily concerned with determining the compatibility of materials with respect to wear and corrosion. Additional tests were made of rotary- shaft and valve-stem seals for fluoride mixtures, and a program for testing the self-bonding of ma- terials in fluoride mixtures was planned. Methods for removing fluoride mixtures from equipment are being developed. Preliminary tests show promising results in rapid dissolution of NaF-ZrF ,-UF , with a nitric acid-boric acid solution. There was only Jight attack on the Inconel container walls, Studies of the reflector-moderated reactor were primarily concerned with the effects of reactor design conditions on aircraft gross weight, but these studies also included a core flow experi- ment, reactor physics considerations, and the de- velopment of g method for chemically processing the fuel (Sec. 3). A parametric study was made to determine the effects on aircraoft gross weight of reactor temperature, power density, and radi- ation doses inside and outside the crew com- partment. The results of analyses of three design conditions show that the gross weight of the airplane is relatively insensitive to reactor design conditions for Mach 0.9, sea-level operation but that it is quite sensitive for the much higher performance, supersonic tlight conditions. Also, an increase in reactor temperature level of 100°F is more effective in reducing gross weight than is a factor-of-2 increase in power density. Pre- liminary tests of a full-scale Lucite core model indicated the need for cutoffs in the pump volutes and turning wvanes oround one quadrant of the impeller periphery to obtain uniform flow distri- bution at the core inlet and also the need for a set of turbulator vanes in the core inlet passage to assure axially symmetric flow and no flow ANP QUARTERLY PROGRESS REPORY separation ot the core shell wall. Specifications were prepared for studies of the effects of the geometry of the reoctor on the physical quantities of interest and for precalculations of several pro- posed critical experiments. [n addition, develop- mental work was done on a new, nonagueous method for processing fuels of the NaF-ZrF -UF, system which consists of (1) recovery of the vranium by converting the UF, in the molten NaF- ZrF4-UF4 mixture to the volatile hexafluoride, using elemental fluoride; (2) gas-phase reduction of the partly decontaminated UF, to UF,; and (3) refabrication of the molten salt fue! from this UF,. Critical experiments were assembled to test the designed core dimensions and core compositions of the supercritical-woter reactor ond the G-E, air-cooled, water-moderated reactor. Also, the critical experiment program for the reflector-moder- ated reactor was re-evaluated (Sec. 4). Reflector- moderated assemblies of simple geometry are to be constructed, and material variations will be made to check consistency with theory and the fundamental constants. PART I, MATERIALS RESEARCH The chemical research on liquids for use in high- temperature reactor systems has beezn primarily a continuing effort to obtain fuels with physical properties superior to those of the mixtures in the I\hJF-Zri:“-UF:4 system (Sec. 5). Thermal analysis techniques were used for siudies of the NaF-KF-ZrF ,-UF , NoF-LiF-ZrF -UF ,, NaF- BeF,-ZrF -UF ,, NaF-LiF-BeF,-UF ,, RbCI-UCI,, and NaCl-ZrCl4 systems. The thCl-ZrCl4 system is of interest because a low-melting-point water- soluble mixture in this system would be useful in ad- dition, the quenching technique applied previcusly has been used along with x-ray and petrographic for removing fluorides from equipment. examination of slowly cooled specimens to obtain a better understanding of the complex relation- ships in the NaF-UF, and NaF-ZrF, systems. High-temperature phase separation has also been used for studying systems, such as NaF-ZrF4, NqF-ZrF4-UF4, and NaF-KF-ZrFA-UFA, in which solid solutions are expected. Experimental studies of the reaction of chromium and iron in the metallic state with UF , in molten NaZrF_ have shown that the deviation of the activity coefficient from unity is probably due to the formation of complex ions such as UF ™, UF™7, and FeF, ™. Two methods were developed for preparing NaZrFS melts con- taining UF3; one sample was found to contain 3.54 wt % UF, and another contained 5.95 wt % UF,. Evidence was accumulated which shows that the presence of ZrF, in a nonuranium-bearing fluoride melt does not afford a mechanism for storing latent reducing power and that ZrH, does not impart a hydrogenous character to the melt, Rates of reduction of Nif, and FeF, by hydrogen in nickel reactors were determined at 600 ond 700°C. Additional spectrophotometric determi- nations of absorption spectra for UF A and UF3 in guenched fluoride melts were carried out. The facilities for the production of experimental fluo- ride mixtures have been modified and expanded, and the large-scale production facility has been reactivated. The corrosion research effort has been deveted almost entirely to studies of the effects of various parameters on the corrosion of Inconel by fluoride mixtures, although some work has been done with ceramic nickel, various 400 series stainless steels, lzett iron, Hastelloy, and varicus brazing alloys (Sec. 6). Corrosion testing in thermal convection loops is being accelerated by the addition of loop facilities and modifications of loop design, Additional tests have confirmed that the corrosion of Inconel increases with oper- ating time when exposed to either NaF-ZcF ,-UF or to NaF—ZrF4 and that the attack is due to the mass transfer of chromium metal. It was alse established that the depth of attack in I[nconel increases slowly, but linearly, with increasing uranium content in the fluoride mixture. However, increasing the surface-to-volume ratio in the thermal convection loops decreases the depth of attack in Inconel., The fluoride mixture NaF-ZrF - UF , was circulated for 500 hr at 1500°F in severc! 400 series stainless steel loops, and metal erys- tals were found in the cold leg of each loop. At- tempts were made to circulate NaF-Zer-UF at 1500°F in Hasteloy B loops, but all attempts failed either because of catastrophic oxidation or mishandling. Static corrosion tests hove been made of many brazing alloys, and it has been materials, found that the nickel-phosphorus and the nickel- phosphorus-chromium alloys have the most satis- factory corrosion resistance in NqF-ZrF4-UF4. The study of corrosion and mass transfer charac- teristics of container materials in liquid lead has been extended to include tests with cobalt, be- ryllium, titanium, and Hastelloy B. The work in metallurgy and ceromics included, in addition to the important study of high-conduc- tivity metals for radiator fins, stress-rupture tests of Inconel, preparations for creep tests of co- fumbium, the brazing of various materials to In- conel, the fabricotion of clad columbium disks and of columbium and molybdenum tubing, the drawing of tubular fuel elements, the development of ceramic materials for pump bearings and for containers for fluoride fuels, and the preparation of boron carbide—iron cermets for use as shielding in connection with in-pile loop studies (Sec. 7). The studies of the effect of environment on the creep and stress-rupture properties of Inconel have continued, and it has been demonstrated that a small change in the composition of the lnconel has a major effect on its creep properties. Methods were developed for preparing o nickel-phosphorus brazing alloy powder that can be applied in the conventional manner. In the study of high-conduc- tivity metals for radiator fins it was shown that claddings of types 310 and 446 stainless steel on copper are satisfactory with respect to diffusion. However, it may be necessary to use Inconel as the cladding material, and therefore various ma- terials were tested as diffusion barriers between [nconel and copper. None of the refractory ma- terials tested were satisfactory, but diffusion barriers of iron and types 310 ond 446 stainless steel were successful in preventing copper dif- fusion for up to 500 hr at 1500°F. Radiator fins of various materials were brazed to Inconel tubing for heat tronsfer tests, and an Inconel spiral-fin heat exchange unit was fabricated for use in radi- ation damage experiments, of beryllium oxide with minor additions of other materials were tested and were found to be un- satisfactory as containers for fluoride fuels, Various composites The physical properties of several fluoride mix- tures and other materials of interest to aircraft reactor technology were determined, and the heat transfer characteristics of reactor fluids were studied in various systems (Sec. 8). The feasi- bility of a small-scale system to be used in studying the rutes of deposition of the films or deposits found at interfaces between Inconel and NaF-KF-LiF has been demonstrated, and the system has been constructed. The system was designed to operate at Reynolds moduli of up to PERIOD ENDING MARCH 10, 1954 10,000 and therefore may alse be useful as a dynamic corrosion testing unit. A technique has been developed for measuring fluid velocity pro- files in duct systems by photographing tiny par- ticles suspended in the flowing medium. This method is to be used for determining the hydro- dynamic structure in the reflector-moderated re- actor flow annulus. Additional forced-convection, laminar flow heat transfer experiments have been conducted in pipe systems which contain circu- lating liquids with volume heat sources; the heat sources are generated electrically, The laminar flow data fall dabout 30% below the values pre- dicted by previously developed theory. included ad- ditional studies of fluoride mixtures in Inconel capsules and determinations of the creep of metals under irradiation, as well as the design and con- struction of in-pile circulating loops (Sec. 9). Data obtained from recently completed petrographic examinations support the previous indications that fluoride mixtures in the NaF-ZrF -UF, system are chemically stable under reactor radiation. Concurrent tests have shown that the corrosion of Inconel by these mixtures increases from 1 to 2 mils of subsurface-void attack for out-of-pile tests to 5 to 6 mils of intergranular attack for in-pile high-temperature tests. Miniature loops are being designed for circulating fluoride mixtures The radiation domage program in such space-restricted locations as the vertical holes of the LITR. The designs are based on detailed studies of fuel flow rates, power den- sities, and rates of heat removal. A new in-pile stress corrosion apparatus has been developed with which it will be possible to determine, simul- taneously, the corrosion effects on stressed aond unstressed portions of a tube. Tests in the LITR and in the ORNL Graphite Reactor have shown that the creep behavior of Inconel is not seriously affected by neutron bombardment. Similar tests are to be made in the higher flux of the MTR., An in-pile loop for circulating fluoride mixtures in the LITR is 80% complete. The analytical studies of reactor materials in- cluded investigations of methods for determining the oxidotion states of the corrosion products, iron, chromium, and nickel, in fluoride mixtures and methods for the separation and determination of UF3 and Ul:4 in fluoride mixtures (Sec. 10). It was established that the solubility of UF, in NaZrFs involves, first, oxidation to tetravalent vranium and, then, dissolution. The conversion of UF,; and UF, to the corresponding chlorides by fusion with NaAlCl, was accomplished. The uranium chlorides are readily extracted by ccetyl- acetone-acetone mixtures. Samples of the fluoride mixture NaF~ZrF4-UF4 that were exposed to mois- ture before being canned were examined with the polarizing microscope after having been exposed to gamma radiation in the MTR conal for 265 hours. In comparison with fuel stored and canned in o helium atmosphere, no effects of irradiation could be observed. PART IH. SHIELDING RESEARCH The work of the Shielding Analysis Group has been concentrated mainly on the investigation of neutron shields (Sec. 11). Calculations were made according to the continuum theory nuclear model to obtain a better understanding of the relationship of differential fast-neutron cross sections to ef- fective removal cross sections, as measured in The calculated values agree well with Lid Tank data for the elements calcu- lated, namely, aluminum, iron, copper, lead, and bismuth. Other heavy-element cross sections can be calculated with relative ease and fair assurance of accuracy. This approach is not, however, useful for the calculation of cross sections for light nuclei. Recent calculations by NDA on the UNIVAC show that a slab of lithium hydride used as a neutron shield would weigh only 63% as much as a thickness of water which would give the same attenuation. Independent estimates made at ORNL confirmed the NDA value. It was shown that the cooling of a lithium hydride shield for the reflector- moderated reactor would not be difficult, bulk experiments. The effects of one and two air ducts on the fast-neutron dose received outside a reacior shield were further investigated at the Lid Tank Facility (Sec. 12). Each of the two ducts consists of one to three 22-in.-long straight sections. The sec- tions were joined ot 45-deg angles. The experi- mental results show thot when two ducts are parallel and in the some plane the dose is in- creased. If they are porallel but in different planes, the dose is the same as for a single duct. In the case where the axis of a neorby short straight duct intersects the middle of a long duct, the dose is again increased. The Lid Tank radi- ation dose measurements made behind 82 reflector- moderated reactor and shield mockups were ana- lyzed. Although there ore still uncertainties about the air and structure scattering, the heat exchanger region, ducts ond voids, and optimization of the shield size and weight, o preliminary calculation indicates that the weight of the basic, designed reactor and shield assembly (excluding the crew- compartmeni shield) will be 44,500 pounds. For this estimate, the core diameter was 18 in., the power was 50 megawaits, and the dose rate cutside the crew shield at 50 ft from the reactor center was 10 rem/hr, shows the removal cross section of B,0, to be 4.4 +0.14, Two bulk shields for the GE-ANP program were tested at the Bulk Shielding Fuacility (Sec. 13). The first test was of a mockup of a duct system ond shield to be used at the G-E Idaho reactor test facility., The second mockup consisted of two sections of the shield for the R-1 reactor. While the data from these tests have not been completely analyzed, it appears thot the results agree gquite well with calculated estimates for the designs. The Tower Shielding Facility is neoring com- pletion and operaiion should begin during the month of March (Sec. 14). A complete set of instruments has been collected for the experi- mental program; especially developed for use ot this focility. Con- sideration of the Tower Shielding Facility for use in a biological program for establishing dose rates for pilots in a prompted a study of radiation doses which can be achieved. A recent Lid Tank measurement some of the instruments were nuclear-powered aircraft has Part | REACTOR THEORY AND DESIGN 1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT E. S. Bettis J. L. Meem ANP Division The reactor for the Aircraft Reactor Experiment was installed in the pit and the fuel and the sodium circuits were completed except for the pump assemblies. An examination of a spare heat exchanger, however, confirmed previous doubts as to the reliability of these units, and they are now being rebuilt. The work necessary to make these heat exchangers as reliable as the rest of the system is expected to require approximately four months. The operations manual has been com- pleted and training of the operating personnel is continuing. The production of the ARE fuel con- centrate, NOZUF , was completed, and the ana- lytical results were verified. The possibility of the fill-and-flush tank going critical if the fuel were admitted to it under certain emergency conditions was explored. A pertinent comparison with an earlier critical experiment and with a machine calculation indicated that the fill- and-flush tank would be subcritical even if it were necessary to admit all the fuel to it. In addition, the calculated and experimental values for the cadmium fraction as obtained from the critical experiments for the reactor are presented. Eoch ARE pump is being subjected to component inspection, cold shakedown testing with air ot room temperature, testing at operating temperatures with sodium, and visual inspections both after assembly and after each cold and hot shakedown test. All pump tanks and covers have been fabri- cated and made available for installation. The fabricated Inconel pump impellers designed to replace the faulty cast impellers have been tested and were found to be satisfactory. Recovery and reprocessing of the ARE fuel has been studied further, and it is planned to dissolve the solid fuel in an aqueous solution, extract the uranium with 5% tributyl phosphate, and isolate it by ion exchange. A dissolution rate of 5 kg of uranium per day is anticipated. Batch ion- exchange tests have been made, and colculations indicate that a Higgins continuous ion-exchange contactor 3 in. in diameter and 6 ft in length will permit a processing rate of 9 kg of uranium per day. However, the capacity will be limited by the fuel dissolution rate. THE EXPERIMENTAL REACTOR SYSTEM The fuel and sodium heat exchangers for the ARE were fabricated by an outside contractor over a year ago. There has been considerable concern throughout the project as to whether the welding was up to the standards maintained on the re- mainder of the system., As acheck on these units, a spare heat exchanger was cut up and examined. The examination showed that the welds at the lugs for stress relief of the tube-to-header plate joints had been made with coated rod and were entirely unsatisfactory. In addition, it was found that an accidental strike had been made on the end of one tube bend with o metallic arc. These findings further emphasized the doubts about the heat ex- changers, and therefore a decision has been made to disassemble and rebuild the heat exchangers to eliminate all the original welds. The moterials required are available, and it is estimated that this work will require approximately four months. The reactor is installed in the pit and both the fuel and the sodium circuits are complete except for the pump assemblies. There is a bypass around the reactor in the sodium circuit which will not be removed until after the sodium circuit has been tested with water, since it would be undesirable to contact the beryllium oxide blocks with water. The pumps for the sodium circuit have been tested at operating temperature and are now ready for installation in the system. A modification was effected in the fuel circuit by removing one of the fill-and-flush tanks. This tank became superfluous when the liquid metal cleaning of the fuel circuit was eliminated from the procedure. The tank was removed from the system in order to eliminate a potential source of trouble associated with the valve in the line to this tank. A simplification was effected in the pump aux- iliaries when the dibuty| carbitol system for pump seal cooling was replaced with a direct water cooling system that requires no circulating pumps or heat exchangers. This cooling system was checked and was found to be satisfactory in the tests run on the pumps. The lubricant cooeling systems for the main sodium and fuel pumps have ANP QUARTERLY PROGRESS REPORT been completed and two of the systems have been installed in the pits. Each of these units is com- prised of an oil-to-water heat exchanger, an oil pump and a spare, ond the changeover solenoid valves for selecting the operating oil pump. Installation of electrical heaters and insulation on the pipe lines are 90% complete. All valves are now heated on individual circuits with variac controls for each valve. The heating circuits have been checked out on all fuel tanks in the tank pit, and strip heaters on the hot fuel dump tank have been replaced with Cromolox tubular heaters. The dump tank ‘‘chimney’’ was equipped with « damper in order to make possible adequate temper- ature control of this tank. The sodium circuit helium loops have been checked out and the hydraulic drives for these blowers have been put in operation. The 3-hp electric motors for driving the hydraulic pumps were exchanged for 5-hp units to provide ample power should the ducts not be completely filled with helium. Also, the helium ducts were welded at the joints to reduce the helium legkage and to improve the ‘‘open pit’" operation prior to the power run of the system. The operations manual has been completed and issued. Definite training sessions for operating personnel are being held on schedule and oper- ational procedures are becoming firm, REACTOR PHYSICS W. K. Ergen J. Bengston C. B. Mills ANP Division Neutron Temperature As mentioned previously,] there is some un- certainty as to the effectiveness of xenen puoisoning in a high-temperature reactor, particularly in a reactor with strong thermal absorption such as that in the ARE., This uncertainty results from the rapid drop of the xenon absorption cross section with increasing neutron energy. The high temper- ature of the reactor means a high average energy of neutrons in thermal equilibrium with the moder- ator, and the average neutron energy is even higher than this eguilibrium energy because some neu- trons are absorbed during the slowing-down process "W, K. Ergen, J. Bengston, and C. B. Mills, ANP Quar. Prog. Rep. Dec, 10, 1953, ORNL.-1649, p. 12, before they reach equilibrium. Only data from experiments at room temperature are available for studying this problem, ond they are not applicable to the ARE problem, The neutron energy spectrum in an infinite ab- sorbing moderator can be computed according to the Wigner ond Wilkins method? in which one integration must be performed numerically. This numerical integral is being coded for machine calculation, The calculations are to be carried out for the following moderators: H, D, Li7, Be, C, O, H,0, and F. Constant scattering and 1/v absorption will be considered. The ratio of ab- sorption cross section (taken, for instance, «at the energy in equilibrium with the moderator temper- ature) to scattering cross section will assume several representative values. Various moderator temperatures, T, can be taken care of by intro- ducing a dimensionless parometer, neutron energy per kT, Criticality in the ARE Fill-and-Flush Tank It was not intended originally that the final ARE fuel mixture would ever be admitted to the tank provided for flushing the ARE fuel system and for tilling the systemn with the nonuranium-bearing fuel carrier. However, during the construction of the ARL the desirability of being dable to admit oll the fuel ond carrier to the fill-and-flush tank under certain emergency conditions became ap- parent, At that stage of construction it was not possible to replace the tank with a tank designed for definite subcriticality nor was it possible to perform a critical experiment on the tank. Some of the cross sections and other nuclear dota for determining the criticality of the tank are somewhat in doubt, but if all these data are ad- justed (within the limits of present knowledge) to be most favorable for criticality, the possibility of the tank going critical could not initially be excluded. However, it was found that a pertinent comparison could be made with data extrapolated Criticality was not achieved in this experiment, but extrapo- lation of the ieasured curve of multiplication from an earlier critical experiment.’ 26. P. Wigner and J. E, Wilkins, Jr., Effect of the Temperature of the Moderator on the Velocity Distri- bution of Meutrons with Numerical Calculations for H as Moderator, AECD-2275 (Sept. 14, 1944; declassified Sept. 13, 1948). 3C. K. Beck, A. D. Callihan, and R. L. Murray, Critical Mass Studies, Part [, K-126 (Jan. 23, 1948). constant vs. uranium investment showed that to be critical the assembly would have had to have been at least as large as the fill-and-tlush tank. Since the shapes of the critical assembly and of the tank are different, the reciprocal of the buckling wos used as the “'size.”” The scattering and moderating properties of the critical assembly and of the tank are essentially the same, but the extrapolated assembly contained very much more fissionable material than the tank. On this basis, the tank would be suberitical. A machine calcu- lation was performed which confirmed this con- clusion, The ARE Critical Experiment Computed values were compared with experi- mental data? for cadmium fraction in the cylin- drical ARE critical assembly with o bare top, an Inconel-covered bottom, and BeO-reflected sides. Spherical coordinates were used for the calcu- lations. Figure 1.1 shows median-plane rodial values for critical assembly CA-8, which had o reflector thickness of 7],5 inches. Similar data for critical assembly CA-9, which had a reflector thickness of 12 in., are shown in Fig. 1.2. The depression at the center of Fig. 1.2 is due to the control rod assembly, Figure 1.3 shows the axial values in CA-8, and Fig. 1.4 shows similar values for CA-9., The discrepancy between the experi- mental and the calculated values in Fig. 1.3 indi- cates that an incorrect value was assumed for the 4D, Callihan and D. Scott, Preliminary Critical As- sembly for the Aircraft Reactor Experiment, ORNL-1634 (Oct. 28, 1953). o;wfis% | | REACTOR 500 i}A & EXPERIMENTAL (REE 4, TABLE IV, p.28) : % o CALCULATED Qo Qo400 i e 5 TR 3oz b e e e ASSUMED CORE RADIUS 16.5 in. -- ,,,,,,,, IR N Lo b ] ol _IJ_ 0 4 8 12 15 o0 REACTOR RADIUS (in.) Fig. 1.1. Assembly CA-8, Radial Cadmium Froction in Critical PERIOD ENDING MARCH 10, 1954 radius in the spherical system used to represent the diffusion equation in the axial direction, PRODUCTION OF FUEL CONCENTRATE G, J. Nessle J, E. Eergan Materials Chemistry Division F. A, Doss ANP Division The production of the ARE fuel concentrate, No,UF ., was completed and the analytical results were verified as acceptable on December 14, 1953, A ORNL - LR—-DWG. 345 0.0 ] 'REACTOR 509 | j _® CALCULATED | i 0.40 CADMIUM FRACTION |--:—-——CONTROL ROD POSITICN 0AD [ 1o ! NS S — - ASSUMED CORE RADIUS, 12 38 in. o | el 0 2 4 6 3 o 12 14 16 REACTOR RADIUS (in) Fig. 1.2, Radial Cadmium Fraction in Critical Assembly CA.9, oa‘mul_nu NG 346 DISTANCE FROM REACTOR END {in) . 15 10 5 0 OUO A j’ """""—""—"}""T' ] REACTOR 535 | 050 beane T e o EXPERIMENTAL (REF 4, TABLE V, p.29) | z » CALCULATED ! | B2 040 f b el e I 0 : ‘ ] < o e e e Ot ] | & 530 et T~ A s - ""\"& 1 5 | | '\.\’f 020 e oo \ 104 is due to greater reliobility of analytical results gained by a tenfold increase in initial fission-product activity spikad into the dissolved fuel, Further study of tributyl phosphate solvent ex- traction processing of ARE fuel solution is under way for determining the optimum extraction con- ditions. Preliminary results (Table 1.1) on the effects of variations in feed nitric acid concen- trations over a range of TBP concentrations show that the test decontamination with 0.5 M HNO, is obtained in the range 5 to 7.5% TBP in Amsco 125-90W, With 3 M HNO, feed, the maximum de- contamination occurs at a lower TBP concen- tration, In both cases, at low TBP concentrations, ruthenium beta activity is much less important than zirconium, These decontamination factors are tenfold lower than those obtained with ARE-type fuel, primarily because of the high ARE-type fuel feed fluoride content, over 1 M, while the feeds used in these experiments contained no fluoride. A high fluoride content is particularly important in increasing zirconium decontamination, and the complexing action of fluoride lowers to a somewhat less extent the extractability of other fission products and vranium, The optimum ratio of fluoride to alu- minum is one of the most important factors yet to be studied in ARE solvent extraction processing. TABLE 1.1. EFFECT OF FEED ACIDITY ON DECONTAMINATION IN BATCH COUNTERCURRENT TESTS Ratio of feed to scrub to extractant: 3:1:4 Uranium saturation of solvent: 20% Feed: 0.5 or 3 M HNO, plus aluminum nitrate for extraction factor of ~ 4 at feed plate; gross beta activity of '[07 counts/min/m! Scrub: aluminum nitrate for extraction factor of 2 at fourth scrub stage TBP CONCENTRATION GROSS BETA ACTIVITY OF PRODUCT IN AMSCO 125-90W DECONTAMINATION FACTOR (% of gross) (%) OF PRODUCT Ruthenium B Zirconium fi' 0.5 M Nitric Acid 3 7.0 x 103 0.01 94 5 2.6 % 104 32 40 7.5 3.1 x 10* 41 37 15 5.3 x 103 85 7.3 30 920 86 5.1 3 M Mitric Acid 3 3.1 x 104 7.7 67 5 2.4 % 103 0.6 89 15 2.7 x 10° 99 2.3 14 PERIOD ENDING MARCH 10, 1954 2. EXPERIMENTAL REACTOR ENGINEERING Ho. W, Savage, ANP Division Two types of pumps are being developed for in- pile test work — pumps for use inside reactor holes and pumps for use just external to reactor holes. The downflow, gas-sealed pump for external use has demonstrated the required performance char- acteristics with no gassing. A preliminary water test indicated promising performance of a centri- fugaily sealed pump model. Development work is continuing eon packed seals for horizontal in-pile pumps, ond designs are being developed for hori- zontal-shaft gos-sealed pumps. Development of corrosiontesting units empleying, simultaneously, high liquid velocity and high temperature difference between the hottest and coldest points in the liquid circuit was continued. The units being developed include two all-lnconel units that will operate with NaF-ZrF .UF, (50-46-4 mole %) as the liquid and will be cooled with air and NoK, respectively, ond one beryllium and Inconel unit that will operate with NaK as the fluid and will be cooled with a heat economizer, The l-megawatt regenerative heat exchanger has operated for 1680 hr, including 1370 hr of fluoride pump operation. A high-performance sodium-to-air radiator was employed as a heat sink, Additional tests were made of high-temperature bearing materials and of rotary-shaft and valve- stem seals for fluorides. A program for testing the self-bonding of materials in fluorides has been planned, and a preliminary test of Inconel vs. Inconel is under way., Methods for removing fluoride mixtures from equipment are being de- veloped, All pump housings and covers have been fabri- cated, inspected, and delivered to the ARE for installation. Intensive proof-testing and inspection of all pump parts is in progress and will continue until all ARE pumps have been delivered. Details of the work on these pumps are presented in Sec. 1, *'Circulating-Fuel Aircraft Reactor Experiment.” IM-PILE LOOP COMPONENT DEVYELOPMERNT W, B. McDonald J. A, Conlin C. D, Baumann D. F. Salmon W. G. Cobb D, R, Ward ANP Division Components are being developed for in-pile loops which are to be operated in the LITR and the MTR for studying the effect of radiation on fuel stability and corrosion of contoiner materials {cf., sec. 9, ““Rodiation Damage’’). The components to be developed are compact fused-salt pumps whick can be operated inside reactor beam holes, high- performance heat exchangers for removal of fission heat, flow= and pressure-measuring devices suitable for in-pile service, and other equipment essential for the operation of in-pile loops. The in-pile loops are to operate with fuel power densities of from 1 to 5 kw/em?, flow rates in the turbulent range, & maximum fuel temperoture of 1500°F, and temperature differentials of from 100 to 300°F, Development of the pumps is now under way. Vertical-Shaft Sump Pump Three vertical-shaft gas»sealed sump pumps have been constructed ond are undergoing tests. Initial testing with water revealed the entrainment of large quantities of gas in the pump discharge ot most pump shaft speeds and flow rates, The pump is of the downflow type, with the fluid entering the impeller around the rotating shaft. Vortexing of the fluid around the rotating shaft was found to be o major cause of gassing. In addition, poor distribution of the fluid upon entrance into the suction chamber caused turbulence of the free fluid surfoce. Design changes were made (Fig. 2.1) which corrected both these difficulties and eliminated gas” entrainment in the fluid ot shaft speeds and flow rates well above the design condi- tions. The pump is sealed with a Morganite (MYIF) face seal, and the shaft and seal are cooled by circulating oil. One pump was tested for 218 hr with fused salts at 1200°F and 305 hr at 1400°F, The normal operating conditions during this test included a shaft speed of 3200 rpm and o flow rate of 1.3 gpm with o developed head of 26 feet, The second and third pumps are undergoing woater fests pre- paratory to testing at 1500°F with sedium, Pump with o Centrifugal Secl A Piexiglas test model of an in-pile pump with no mechanical liguid seal has been constructed (Fig. 2.2). The sealing in this pump is accome plished by centrifugal action of the pumped liquid. This is analogous to a sump pump with the gravite- tional field replaced by a centrifugal field, 15 ANP QUARTERLY PROGRESS REPORT DISTRIBUTION BAFFILE ACROSS FLOW ENTRANCE -~ ANNULAR FLOW DISTRIBUTION CHANNEL FLOW PATH ANTI-FLOWBACK SEAL IMPELLER CASTELLATED NUT NOTE: PARTS SHOWN THUS: ARE THE RECENT M- PORTANT REVISIONS UNCLASSIFIED ORNL-LR-DWG 334 SPRING HEAT BAFFLES SHAFT LIQUID LEVEL ANTI-VORETEXING RING { ALSO SERVES TO REDUCE THE VOLUME OF LIQUID IN THE PUMP} FEEDBACK PASSAGE IN SHAFT ( KEEPS ANNULAR OPLENING ARQUND SHAFT FILLLED WITH LIQUID AND PREVENTS VORTEX- ING NEXT TO THE SHAFT, WHICH WOULD PERMIT GAS TO BE DRAWN DOWN TO THE IMPELLER ) Fig, 2.1, Vertical-Shoft Sump Pump (Medel LFA) Showing Modificotions Made to Prevent Gassing. The inlet side of the impeller is conventional; however, the back side is extendsd to form an annular chamber which is partially filled with the pumped fluid which rotates at high speed during operation and forms an onnulus of ligquid with @ free surfoce that never reaches the pump shaft, The annular chamber is pressurized with gas. Since this gas pressure contributes to the absolute prassure throughout the system, it must be such that o no place in the system does the total 16 pressure drop below the vapor pressure of the fluide The face seal shown in Fig. 2.2 is a gas seal for retention of the pressurizing gos. The small radial holes in the annular section of the back of the impeller are designed to permit any gas which might be entropped due to turbulence between the rotating fluid ond the stotionary wall of the pump housing to be centrifuged out., En- tropped gas will thus be prevented from entering the main fluid system, The pump was tested with water as the working fluid and air as the pressurant. The sealing characteristics were very good, and there was no observable pump leakage regardiess of pump orienfation, even when the pump was inverted. The test also indicated that the loop can be easily filled. A bypass from the pump discharge to the onnulaor chamber at the centrifugal seal removes entroined gas in o short period of operation. Since PERIOD ENDING MARCH 10, 1954 this pump is not sensitive to its orientation in respect to gravity, it also shows promise as an aircraft type of pump., This possible application is being further explored. Horizontal-Shaft Sump Pump A horizontal-shaft sump pump (Fig. 2.3) is come templated for use with in-pile loop experiments in UNCLASSIFIED ORNL-LR-DWG 335 — RADIAL HOLES FOR DEGASSING BY PASS TO ANNULUS CHAMBER -~ - PUMP HOUSING —INLET LLINE FOR GAS PRESSURE TANGENTIAL DISCHARGE - T ANNULAR CHAMBER ~-FACE SEAL [GAS) / - EREE FLUID SURFACE WHEN OPERATING Fig. 2.2, Centrifugal Seal Pump. UNCLASSIFIED ORNL-LR-DWG 236 /-—--~ BEARINGS —- N\ £ N _/ N /-GAS SEAL DISCHARGE ) dpaann J AN / PRIVE MOTOR N ™, L \ LIQUID LEVEL- o e -~ e SHAFT ................................... —_ —_— \\. o \\ \‘ . INLET - E wanp - - . \ T \; Oy h SUCTION \— BAFFLES Fig. 2.3. Horizontal-Shaft Sump Pump. 17 ANDP QUARTERLY PROGRESS REPORT which it is desirable to have all equipment con- tained within the reacctor access holes and existing shielding. The pump takes its suction from a con- trolled volume of liquid below the pump shaft, and a rotary face seal is used to maintain an inert=-gas atmosphere above the liquid in the pump. A mockup has demonstrated that a pump of this design can be primed by increasing the pressure on the liquid surface in the pump to force fluid into the impeller, HEAT EXCHANGER AND RADIATOR TESTS A, P. Fraas R. £. MacPherson R. W. Bussard He Jo Stumpf ANP Division Heot Exchonger Test Limited test operations have been completed on the l-megawatt regenerative heat exchanger which employs some of the design features of the hedat exchanger proposed for the reflector-moderated air- croft reactor. The test apparatus was described previously,! The scope of the test program was limited by several factors, 1. Original designs of the heat exchanger were based on the use of NaF-KF-LilFF (11,5-42.0-46,5 mole %) as the circulating fluoride medium. Since this mixture was not available at the time required in sufficient quality or quantity, it was necessary to use NaF-ZrF , (53-47 mole %), a more dense and more viscous fluid, which put serious limitations on the attainable Reynolds numbers on the fluoride side of the tube bundles, 2, A design error in the area of the sodium flow path through the heat exchanger made it necessary to hold sodium flow rotes to one-half to one-third the desired levels in order to prevent serious over- heating of the electromagnetic pump cell, 3. Circuit limitations on the power supply to the resistance heater which supplies heat to the sodium circuit limited heat inpul to a maximum of 50 kw, that is, approximately one-half the desired value, These limitations prevent the drawing of re- licble conclusions from the performance data ob- tained on the heat exchanger, However, con- siderable confidence has been gained in the practicability of such a heat exchanger design.? 'R, E. MacPherson and H. J. Stumpf, ANP Quar, Prog. Rep. Dec. 10,1953, ORNL-1649, p. 28, 2ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556, Fig- 8»], pu 72¢ 18 Since initial startup, the heat exchanger has been on heat of temperature levels varying from 1200 to 1600°F for 1680 hours. The fluoride pump was in operation for 1370 hr of this time, and the entire bifluidsystem was in complete operation for a total of 712 howrs, This period of trouble-free heat exchanger operation was interrupted by binding of the fluoride circulating pump., Upon removal and inspection of the pump uassembly, it was foundthat the binding was caused by the condensa- tion of zirconium fluoride crystals in an onnular area around the pump shaft above the liquid level in the pump sump. Considerable crystal formation was also present on the pump sump wall above the liquid level where it was cooled by the lower flange of the top head assembly (Fig. 2.4). Inspec- tion of the internal surfaces of the heat exchanger showed them 1o be in excellent condition., Modifi- cations are being mode in the test loop, and future operation will be concerned mostly with endurance testing of the component parts. Sodium-tc-Air Radiator Tests R. E. MacPherson H. Jo Stumpf ANP Division J. G, Gallagher Americon l.ocomotive Company The design of the intermediate heat exchanger test loop '+ indicated that o sodium-to-air radiafor weould be the most convenient form of heat dump and of the same time would provide an opportunity to obtoin additional radiator endurance and per- formance data with little extra cost and setup time. A strip-fin and tube radiator designed to dissipate 100 kw of heat was built for the rig? and was operated for 1013 hr at sodium inlet temperatures ranging from 600 to 1550°F, A leak in the sodium heater coil made it necessary to interrupt the test. The heater coil is being replaced ond the test is to be resumed, The air flow through the radiator was varied from 0.58 Ib/sec-ft? to 17.32 lb/sec-ft2 at the inlet face areq, with corresponding Reynolds numbers of 327 to 11,100. The radiator geometry was the same as that of core element No. 3 described in ref, 5, 3B. M. Wilner and e Jo Stumpf, Intermediate Heat iExch)ange:r Test Resufts, ORNL CF-54-1-155 (Jan. 29, 954), 4p, Patriarca, G. M. Slaughter, and J. M. Cisar, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNI.-1649, p. 84-89. Sw, s, Former, A, P. Fraas, H. J. Stumpf, and G. D. Whitman, Preliminary Design and Performance Studies of Sodium-to-Air Rediators, ORNL-1509 (Aug. 3, 1953). PERIOD ENDING MARCH 10, 1954 Fig. 2.4. Zirconium Fluoridé Crystal meufia;n in Sump Pump Used to Circulate RaF-ZxFfi’ for 1370 hr at 1200 to 1600°F, except that the fins were interrupted every 26 inch, The performance data plotted as Colburn modulus (j) ond friction factor (f} vs. Reynolds number are given in Fig. 2.5. A plot of aireside film coefticient (h) and over-all heot transter coeffi- cient (U) vs. Reynolds number is given in Fig. 2.6 for radiator No. 3, which is a plate-fin type of core element with the fins interrupted every 2 inches.” Reduction of the test data for radiator No, 3 to plots of j and f vs. Reynolds number indicated that more frequent interruption of the fins increased both the heat transfer modulus j and the friction factor f and resulted in no net gain in performance, A more complete analysis of the strip-fin radiator and comparisons with other compact surfaces will be given in a forthcoming report. 5D, F. Saimon, ANP Quar, Prog, Rep. Dec. 10, 1953, ORNL.-1649, p. 29. £ FORCED-CIRCULATION LOOPS D. F. Salmon ANP Division Air-Cooled Loop A second air-cooled forced-circulotion loop was fobricated from (0,15in.-0D by 0.025in.wall Inconel tubing.® Difficulty waos encountered in filling and storting the loop with the fluoride mixture NUF-ZI’F,rUF" (50-46-4 mole %). The tubing was ruptured in the process, and, ofter the necessary repairs were mode, the same difficulty was again experienced. Freezing ot an electrical- resistance heater connection, followed by thawing with a torch, caused the failures. Another loop, fabricaoted from lé-in.-OD by 0.035 ine~wall tubing, started with ease affer it was filled. This loop operated for 100 hr before « brush failure in the pump moter coused the cold 19 ANP QUARTERLY PROGRESS REPORT UNCLASSIFIED ORNL_-LR-DWG 337 e I"#}} l— h T 7T | AR P P b i | — . FQI"T ON F/—‘«CTOR —_ k' b e rSOD\UNI INLET TEMPERATURE 1200 TO 1350 °F | (SOD\UM INLET TEMPERATURE 500 TO 600°F 10! 0 I =L " N o = 2 5 a2 ' . +0 - COLBURN MODULLS ] somum INLET TEMPERATURE 500 TO 600°F;’j;j " SODIUM INLET TEMPERATURE 1200 T0 1350° F7;¥ e e gy AL | , ‘ / | S | b B S ‘ g | ‘ | I ‘ : | i ! by ‘ 103 [ | [ [ i 10° 2 5 107 2 5 1) ey /407 G Mg ) Fig. 2.5, Sodium-to-Air Radiator Performance Data Plotted as Colburn Modulus and Friction Factor vs. Reynolds Number, leg to freeze. The resultant excessive heating in the hot leg caused a tube rupture. The first 50 hr of operation was at 1400°F (maximum) with a temperature differential of 120°F, and the remaining operation was at 1500°F (moximum) with a tempera- ture differential of 160°F. The fluid velocity {calculated by heat balance} during this test was 2 fps (Reynolds number, 1155), performance with time was found. No deteriorationof Use of the larger diameter tubing corrected the filling and starting troubles, but, to obtain the desired turbulence aond temperature difference, a pump with greater head must be obtained and the magnitudes of the heat source and heat sink must be increased. 20 Beryllium-NaK Compatibility Test The compatibility of beryllium metal with NaK is being tested to determine whether protection of the beryllium would be required if these materuals were used together. A sample of beryllium / in. thick and 1 in. in diameter with ten ]/8 in.dia flow passages was inserted in an Inconel figure 8 loop and exposed to circulating NaK for 1100 hours. The sample and the NaK at the sample were main- tained at 1450°F, and the velocity of the MNaK through the sample was 20 fps (Reynolds number, 8600). The minimum temperature in the cold zone of the loop was 900°F. The beryllium sample is now being analyzed metallurgically; it had a heavy, black scale when it was removed from the loop. Biflvid Heat Transfer Loop The bifluid loop for transferring heat between the fluoride mixture NaF-ZrF -UF , (50-46-4 mole %) and NaK in two double-tube heat exchangers has been completely rebuilt for a third test. The primary purpose of this loop is to determine cor- rosion and mass-transfer effects in an all-Inconel system with the fluoride mixture circulating at definitely turbulent flow (Reynolds number, >5000) and with a total temperature difference inthe fuel in excess of 100°F from the hottest to the coldest point. (The results of metallurgical examination of the loop after the second test are presented in sec, 7, “‘Corrosion Research.’’}) The systemconsists of the two double-tube heat exchangers operating between the hot and cold legs of a figure 8 loop. The NaK circulating in the annulus of one heat exchanger heats the fluoride mixture, and the NaK circulating in the other heat exchanger cools it by the same amount. The center tubes of the heat exchangers are made of 0,225-in.-OD by 0.025-in.- wall Inconel tubing 45 in. long. A model LFA, centrifugal, sump-type pump is used to circulate the fluoride mixture {cf., Fig. 2.1), Preliminary cleaning by circulating a fluoride mixture for 2 hr at 1200°F has been accomplished. Startup with the regular charge of the fluoride mixture has been delayed because a leak developed in the NoK side of one heat exchanger and that heat exchanger is being replaced. PERIOD ENDING MARCH 10, 1954 UNGLASSIFIED 100 ORNL-LLR-DWG 338 ______ } J{_ SRR SRS N S SO N o b b e e e o FILM COEFFICIENT, A, DETERMINED BY ASSUMING TUBE waLL anp | 1717777777 Lo SODIUM FILM RESISTANCE NEGLIGIBLE = ot e L 50 - e L | | e ---------------------------- " FILM COEFFIGIENT, # SODIUM IN AT 1200 TO 1350°F-.[ SODIUM IN AT 500 TO 600 °F -] _ 20 e - C | =] o £ S 2 o 10 > jas] = <1 = | OVER-ALL HEAT TRANSFER COEFFICIENT, ¢| | 5 Sl SODIUM N AT 1200 TO 1350 °F ot TTSODIUM IN AT 500 TO 600°F, . 2 e { b ¥4 | l } 102 2 5 10° 2 5 10* A 4r, & R, ( A ) Hr Fig. 2.6. Air-Side Film Coefficient and Over-All Heat-Transfer Coefficient vs, Reynolds Number for Sodium-to-Air Radiator No. 3. 21 ANP QUARTERLY PROGRESS REPORT HIGH-TEMPERATURE BEARING DEVELOPMENT W, C, Tunnell J. W, Kingsley P. G. Smith ANP Division W, K. Stair, Consultant University of Tennessee R. N. Mason Design work and developmental investigations are continuing on a hydrodynamic type of bearing which will operate ot elevated temperature in a fluoride mixture. The work to date has been con- cerned with determining the compatibility of me- terials with respect to wear and corrosion, Materials Compatibility Tests Tests were continued with the use of the equip- ment described previously,” in which a rotating plate specimen is maintained in contact with a stationary pin specimen under a known load. Tests of a chrome carbide pin and a titanium carbide plate, a titanium carbide pin and a titanium carbide plate, and an Inconel pin and a graphite plate were operated at 1200°F for 2 hr each in NaF-ZrF ,-UF (50-46-4 mole %), In each test, metal buildup of material between the specimens prevented proper contact. This metal buildup has been identified as iron, presumably from the type 316 stainless steel container material, Figure 2.7 shows the metallic buildup on a chrome carbide pin specimen. lnvesti- gation is under way to determine whether this con- dition is due to having a multimetallic system and whether it would be possible to sufficiently reduce the metal transfer by using different container materials, It is probable that any bearing design or application will have at least one material other than the container material in the fluid system. One test for which a nickel-plated pot and an Inconel sump pot with chrome carbide and titanium carbide specimens were used also showed an un- identified magnetic deposit that is being analyzed. Bearing Tester Design Design calculations for bearings operating in fluoride fuels have been made for a number of ex- pected operating conditions. The deflections of Inconel shafts ot temperatures of up to 1500°F have been evaluated for various diameters of up to 7Ra N. Mason et al., ANF Quor. Prog. Rep. Dec, 10, 1953, ORNL-1649, p. 25, 22 LASSIFIED o PHOTQ-21 147 Fig. 2.7. Metallic lron Buildup of Chrome Car- bide Bearing Material Tested at 1200°F for 2 hr in Fluoride Mixture NqF-ZrF4nUF4 {50-46-4 mole %) . 2 in. and for shaft overhangs of up to 22 inches. These length-to-diameter ratio of as low as 0.5 may have studies indicate that bearings having a sufficient load capacity to he used in circulating pumps at fuel temperatures of up to 1500°F and that for these low length-to-diameter ratios, the bearings may be usad without self-aligning mounts. In all cases computed, the maximum load capacity was determined by the bearing limitations rather than by the shoft delflection. ROTARY-SHAFT AND YALVE-STEM SEALS FOR FLUORIDES W, C. Tunnell J. W, Kingsle; P. G, Smith ANP Division R. N. Mason Spiral-Grooved Graphite-Packed Shaft Seals Seal test No. 20 for which a graphite packing was used oround a ]3/16-in.-difi shaft with o machinesd spiral groove was terminoted after 3487 hr of opera- tion because of a heater failure. There had been no detectable leakage of the fluoride mixture (NaF-ZrF ,-UF ,, 50-46-4 mole %) during this period, and only a small amount of graphite leaked during the early part of the run. There was essentially no required or attention given to the apparatus during the last 2500 hr of operation. The powert requirement was more regular and smooth than had been experienced with other packed or frozen seals, When the apparatus was disassembled, it was found that the shaft was worn and that the packing had been impregnated with fluorides. Figure 2.8 shows the shaft after disassembly. Seal test No. 29 was a further attempt to verify the nonwetting charocteristic of graphite in a rotating shaft seal for sealing NaF.ZrF, UF (50-46-4 mole %). The shaft was 2/ in. in dmm- eter and it operated in a horlzonml position, in contrast to a previous test® in which a lflé-in.-diq shaft operated satisfactorily in a vertical position for a period in excess of 3000 hours. The apparatus consisted of a rotating shaft, Fig. 2.9, containing a spiral groove that operated in a conventional *stuffing box arrangement, to which was attached a fluid container with a means for pressurizing the fluid. The packing material for this test consisted of a mixture of 90 wt % Asbury graphite and 10 wt % MoS The retainers for this material were bronze wooi ' This test was made with the fluoride mixture NaF-ZrF ,-UF , (50-46-4 mole %) at 1100 to 1180°F and with the shaft rotating at 560 rpm for a peried of 47 hours. The packing temperature was from 100Q0°F down to 600°F, and the pressure on the fluoride mixture was 5 psi. Ledkage occurred during the last 17 hr, and the test was therefore terminated for examination. maintenance Post-run exgmination revealed that the bronze wool retainer at the hot end of the packing had failed and allowed the packing material to be con- veyed into the seal pot. This packing failure is believed to have been the reason for fluid leakage from the seal. Further tests on this seal are planned. Graphite-BeF 5 Packed Seal The previously reported® test No. 22 in which a BeF ,-graphite mixture was wsed as the packing BANP Quar Prog. Rep. Sept. 10, 1953, ORNL-1609, pe 236 PERIOD ENDING MARCH 10, 1954 Fig. 2.8, Spiral-Grooved Graphite-Packed Shaft After Testing in Fluoride Mixture NaF.ZeF UF, {50-46-4 mole %) for 3487 Hours, material was terminated because of binding of the 1 3{6-in.-diu shaft ofter 4530 hr (over six months) of practically continuous operation at speeds of up to 2500 rpm. The total leakage of the flucride mixture NaF-ZrF -UF (50-46-4 mole %) during this period was less fhan 10 in.2, and the maintenance time was essentially zero. The power requirement during the test was variable, buf it is not known how much of this variation was due to intermittent metal-to-metal contact. When the apparatus was disassembied, there was no visible evidence that 23 ANP GUARTERLY PROGRESS REPORT Fig. 2.9, Spiral-Grooved Shaft, could explain the sudden binding and stopping of the shaft. The shaft was worn, as can be seen in Fig. 2.10, and the places where the Stellite coating was worn through can be eosily seen. Y-Ring Seal The Vering seal test No. 27 reported previously® was operated again on a 21{,-in.-dia shaft to seal helium at a high temperature. Gas leakage below 1 em3/sec was readily obtainable over an extended period of time, and therefore this seal will be tested with Hluorides, The leakage appeared to be sensi- 24 Fig. 2.10, Continuous Operation for 4530 hr ot Speeds of Up to 2500 rpm in Flvoride Mixture NaF-ZsF -UF (50-46-4 mole %). Graphite-BeF, Poacked Shaft After Power surges were not experienced in this test os in previous tests. The pressure differential across the seal has been low in all tests on this apparatus. tive to the shaft temperature, Chevron Seal In an attempt to restrain powdered packing ma- terials, a 1%-in.—diq Chevron {Skinner) seal {(Fulton- Sylphon Company) was tested (test No. 30). This seal is all metal and was designed for use around reciprocating shafts or pistons. The sealing ele- ment consists of a series of 0.003-in.-thick sheets These elements are made so that the sealing edges ore deflected when the seal assembly is engaged ex- ternally in a cylinder or internally on a rod or shaft (that is, interference fits). It is claimed that because these seals are flexible they provide seal assemblies that have relatively low friction. Figure 2.11 is a sketch of a possible arrangement of these seals for retaining powdered packing. Such a seal has been tested as a means of re- taining a mixture of 50 wt % MoS, and 50 wt % graphite. The test operated for over 750 hr, with no heat added. There was some wear on the Colmonoy coated shaft and some leakage of the packing. In this test, the packing material was introduced into the stuffing box region by means of Q screw conveyer. formed into flexible V or reed-type rings. A pressure differential was established across the seal by helium gas pressure, and leakage of this gas was apparently a function of the packing pressure, since the gas leakage increased with graphite leakage but was reduced -3 ,,,,,,, ';./ SEAL BOOY L i CHEVRON SEALS (SKINNER) - I/L“i PERIOD ENDING MARCH 10, 1954 when the screw conveyer was operated. Operation of the screw conveyer caused small increases in power requirements for the seal. The seal was heated to about 1000°F, but no detectable dif- ference in operation was noted. |t is planned to test this design with fluorides. Graphite-BaF , Packed Seal Seal test No. 28 was made in an attempt to verify the lubricating properties of barium fluoride when it is used as an additive with graphite for a packing material, The packing consisted of a mixture of 10wt % BaF , and 90 wt % National Carbon graphite No. 2301. The retainers for this packing were bronze wool, 4 The apparatus consisted of a horizental rotating shaft 2]{‘ in, in diameter inserted into a seal pot through a stuffing box, The seal was pressurized by helium pressure applied to a surge pot attached to the seal pot. The shaft was driven by a 5-hp Varidrive motor that was belt-connected to the driven shaft, UNCI_ASSIFIED DWG. BEK-159T6A S RN L : LOWER SEAL RETAINER - _ > | o [y} """ o ,,,,, o g __________ < a ...... i - e e [N T = - =3 | 5 A s COLMONGY NO. & SPRAYED " | HARD FACING oo™ \ S A A in PACKING e J e UPPER SEAL ASSEMBLY Fig. 2.11. Chevron (Skinner) Seal Arrangement for Retoining Powdered Packing. 25 ANP QUARTERLY PROGRESS REPORT This test operated for 290 hr with a shaft speed of 760 rpm. The fluoride mixture used was Naf- ZrF -UF, (50-46-4 mole %) at temperatures of 1100 to 1150°F, and the pressure on this fluid was 5 psi. The temperature along the packing was from 1000°F down to 600°F. The power trace was typical of that of a frozen seal with excessive leakage. Post-run examination showed that the shaft surface was in very good condition where the packing was located ond thot the only oppreciable wear was in the region of the pocking gland. The packing maferial was coked and quite slick to the touch. Other tests with barium fluoride as an additive to the packing are planned for the near future. Shaft Seal for In-Pile Pump Seal test No, 33 is being made to determine the feasibility of sealing fluorides in an in-pile pump rotating shaft with a packed type of seal. The packing for this test consists of a mixture of 95 wt % Asbury graphite and 5 wt % MoeS, ond is retained by bronze wool at each end of the packing. The apparatus being used consists of a seal pot to which a stuffing box is attached and through which « ]/z-in.wdic Inconel shaft is inserted into the pot. The shaft is driven by a ‘/‘g-hp constant- speed motor with a rated speed of 3450 rpm. The test unit is constructed with the stuffing box off- center on the seal pot so that the container will perform both ds a sump tank and a seal pot. The seal pot functions as o sump tank before and after the test. The system is dumped by pivoting the test to o vertical position. The pot is filled to between half full and full to provide a gas space for pressurizing the fuel against the seal. Heat is applied externally through the wall of the pot; cooling is by natural convection, To date, this test has operated for a period of 360 hr with NaF-ZrF -UF, (50-46-4 nwole %) at a temperature of 1180 to 1250°F, The temperature drop along the seal is from 1100 to 550°F, and the pressure on the seal is 2 psi. Operation has been quite successful in that no difficulties have been encounfered. The power required by the seal has been of the order of 100 to 200 watts for most of the test; however, for a period of one to two days, the power requirement was of the order of 200 to 300 watts. This increased power appeared to have been the result of a slight lowering of the tempera- ture of the seal. The average leakage rate is 26 about 35 cm3/day, with lower rates of times for periods of two to three days., The power trace is quite variable and resembles those previousiy encountered with frozen seals, When this test has been completed, another test will be made with this equipment with o hard-faced shaft. Packing Penetration Test A packing penetration test was made with MaF- ZrF -UF, (50-46-4 mole %) ogainst an Asbury graphite that was similar to Asbury 805 with re- spect to amorphous carbon content and spectro- graphic analysis, but the particle size was smaller and more uniform. The operating conditions were the same as in previous tests.’ The test was terminated after 1125 hr, and there had been no fluoride seepage through the packing material, Yalve-3term Packing Tests Two additiona! valve-stem packing tests were made this quarter, One test was made under the same operating conditions as those used for the previous tests.'? The material tested was Asbury 805 graphite with Fel Pro C-5 high-temperature lubricant against Naf-ZrF ,-UF , (50-46-4 mole %). This test was terminated at 350 hr because of leakage. The other test was made under actual operating conditions, A ]/4-in. stainless steel valve in a fluoride transfer line was backed with the Asbury graphite used in the packing penetration test. The fluoride used was a mixture of NaF-Zri -UF, (50-46-4 mole %) and Naf-KF-LiF-UF ; (10.9-43.5- 44.,5-1.1 mole %). There were no modifications in the valve. The original pocking was removed and replaced with the graphite, and a thin layer of bronze woo! was used at the bottom of the stuffing box. Operation of the valve was quite satisfactory. The valve was under fluoride pressure about 1 hr, during which time it wos opened and closed 32 times; there was no leakage. When the transfer was completed, the line and valve were blown clear of all fluorides. After coeling, the valve wos not frozen or stuck and could be cycled easily. . B. McDorald et al., ANP Quor. Prog. Rep. Dec. 10, 1952, ORNL-1439, p. 23. 10R, N, Mason, P. G. Smith, and W, C, Tunnel, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p, 25. SELF-BONDING TESTS OF MATERIALS IN FLUCRIDE MIXTURES G. F, Petersen ANP Division The successful operation of some equipment may depend on the resistance of some of its ports to self-bonding, or self-welding, at high temperatures. An example is valve operation, in which the seal may become stuck, The purpose of this experiment is to check the validity of the design of the test apparatus, a modified stress-rupture tester. The long-range purpese is to test materials and establish criteria for selection of materials couples {metals, ceramics, cermets). A preliminary test is being made of a couple of Inconel against Inconel in NaF-ZrF .UF , (50-46-4 mole %) with a total load of about 50 pounds. The 3/B-irl.-dicx by %-in.-long Inconel cylinders have flat contact surfaces. In the present test, the fluoride temperature is 1375°F, and the system pressure (helium) is 1 to 1 lé psig. The test will be operated for 100 hours. Alterations in design, if any, will depend on inspection of the sample and equipment upon completion of the test, It is intended that many combinations of materials will be tested. 11y, W, Dobratz et al., Tuballoy Process Research Memo, N-34 (Apr. 9, 1943). PERIOD ENDING MARCH 10, 1954 REMOYAL OF FLUORIDE MIXTURES FROM EQUIPMENT L. A. Mann G. F. Petersen ANP Division A laboratory-size project is in progress for de- termining methods for removing fluoride mixtures from equipment. The rate of attack on Inconel of a nitric acid~boric acid solution vs. acid concentra- tion is being studied.!! Preliminary tests show promising results in rapid dissolution of NaF-ZrF ,- UF ( (50-46-4 mole %) with only very light attack on the Inconel container walls. There ore some indications that the solution attacks Inconel faster if the Inconel has previcusly been exposed to molten fluorides. Quantitative tests are being made to determine rates of solution of the fluorides and rates of attack on Inconel under various conditions of concentration of the acid solution and temperature on both untreated Inconel samples and samples pretreated with the fluorides. In a bench-scale test with 50 mi of 18% HNQO; and 10 g of H,BO, per 150 ml of water at 180°F and 1 gpm flow through o Ié-in. Inconel pipe laden with a ]/s—in. thickness of fluorides, about three.fourths of the pipe area wos cleaned down fo the metal in 1 hr; a variable-thickness film of fluorides was left on the remaining pipe area. The uncleaned area may have been portially protected by bubbles, Further bench-scale tests will be made. 27 ANP QUARTERLY PROGRESS REPORT 3. REFLECTOR-MODERATED REACTOR A. P. Fraas, ANP Division A parametric study was made to determine the effects on aircraft gross weight of the reactor temperature, power density, and radiation doses inside and outside the crew compartment. A chart was prepared for use in the calculations which gives the weights for the reactor, the reactor shield, the crew shield, and the propulsion machinery as functions of aircraft gross weight and useful load. The results of the calculations for three design conditions are presented, and it is concluded from this study that the gross weight of the airplane is relatively insensitive to reactor design conditions for Mach 0.9, sea-level operation but that it is quite sensitive for the much higher performance ond supersonic-flight conditions. Also, an increase in reactor temperature level of 100°F is more effective in reducing gross weight than is a factor-of-2 increase in power density, An evaluation of the results of preliminary tests of a full-scale Lucite core model indicated the need for cutoffs in the pump volutes and turning vanes around one quadront of the impeller periphery to obtain uniforin flow distribution at the core inlet. Studies of flow in the core have shown that a set of turbulator vanes is required in the core inlet passage to assure axially symmetric flow and no flow separation at the core shell wall. Specifications were prepared for a parametric study of the effects of the geometry of the reflector- moderated reactor on the physical quantities of interest, such as critical mass, required mole per cent of uranium in the fuel, and power distribution. The reactivity coefficients of Inconel, sodium, and beryllium as functions of radius were completed for a 50-megawatt reactor design. Specifications were prepared for precalculations of several proposed criticol experiments., The calcu- lations are being made on the UNIVAC according to the multigroup, nine-region procedure coded by the ORNL. Mathematics Panel. Techniques for performing reactor statics calcu- lations on the ORACLE and a method for computing the ‘‘age-to-indium’’ and the kg for beryllivm- moderated systems are described. Developmental work was done on a new, nonaqueous method for processing fuels of the NaF-ZrF ,-UF , system. reactor 28 EFFECTS OF REACTOR DESIGN CONDITIONS ON AIRCRAFT GROSS WEIGHT A. P, Fraas ANFP Division B. Wilner Aerajet-General Corp. The costs of construction, operation, and mainte- nance of aircraft are directly proportional to the gross weight., Thus it is important to know the effects on airplane gross weight of reactor tempera- ture, power density, and radiation doses inside and outside the crew compartment. survey' was carried out by using the quite com- plete set of shield weight data prepared in the course of the 1953 Summer Shielding Session and the engine performance data given in a recent A parametric Wright Aeronautical Corporation report, 2 The basic method used by the Technical Advisory Board, North Americon Aviation, lnc., and the Boeing Airplane Company was used to prepare a set of tables and charts to facilitate aircraft performance calculations. The engine compression ratio was taken as 6:1 and the pressure drop from the compressor to the turbine was taken as 10% of the compressor outlet absolute pressure. The specific impulse and specific heat consumption were taken from Figs. IX-1 through 1X-12 of the Wright report; engine compressor and turbine weight were taken from Fig. 1-19Y and engine air flow from Fig. [-18. Engine nacelle drag was taken from Fig. 67 of ANP-57,% except that 50% submergence of the nacelles in the fuselage was assumed. The weight of the engine tailpipe, cowling, and support structure was taken as 25% of the compressor and turbine weight., The weights of the NaK pumps, lines, and pump drive equipment were calculated on the same bases as were the estimates given in YA. P. Fraas ond B. M. Wilner, Effects of Aircraft Reactor Design Conditions on Aireraft Gross Weight, ORNL CF-54-2-185 (Feb. 26, 1954). ZR. A. Loos, H. Reese, Jr., and W. C. Sturtevant, Nuclear Propulsion System Design Analysis Incorperoting a Circulating Fuel Reactor, WAD-1800 (January 1954). 3A. P. Fraas, Effects of Major Porameters on the Performance of Turbejet Engines, ORNL ANP-57 (Jan. 24, 1951}, ORNL-15154 The radiator cores were designed to give 1140°F as the turbine air inlet temperature, with a peak NaK temperature of 1500°F, and an air pressure drop across the radiator core equal to 5% of the compressor outlet pressure. The resulting weight of the NaK system was somewhat higher than would be obtained from the Wright report. Table 3.1 presents the results of this survey. Table 3.2 gives the installed weight of the pro- pulsion machinery and the reactor power as functions of thrust. The weight of the reactor plus the reactor shield was given as a function of reactor power in Tables 3.1 through 3.4 of ORNL-1609%.° These data were plotted to give charts similar to Fig. 3.3 of ORNL-1609. The basic equation relating aircraft gross weight to the weight of the aircraft structure, the useful load, the shield weight, and the weight of the propulsion machinery is the same as that used by the Technical Advisory Board, North American Aviation, and Boeing: Wg e Wst + UL + st + W pm ! where Wg = gross weight, Ib, We, = structural weight {including landing gear), lb, UL = useful load, Ib, st == shield weight, ib, W = propulsion machinery weight, 1b. i Thsij weight of the structure was taken as 30% of the gross weight, a valve in keeping with pro- portions used by the TAB, North American Aviation, Boeing, and the bLockheed Aircraft Corporation. While this value would probably be closer to 25% for subsonic aircraft (except for aircraft using low specific-impulse power plants, such as the super- critical-water cycle), the value used seemed representative and adequate for the purpose of this analysis. In solving the equation for gross weight, it was found most convenient to prepare a chart such as that in Fig. 3.1, which gives the total weight for the reactor, the reactor shield, the crew shield, and the propulsion machinery as a function of aircraft gross weight and useful load. The useful load was considered to include the crew, radar equipment, ‘A, P, Fraas, ANP Quoar. Prog. Rep. Mar. 10, 1953, ORNL-1515, p. 79. A, P. Fraas, ANP Quor. Prog. Rep. Sept. 10, 1953, ORNL-1609, p. 33 . PERIOD ENDING MARCH 10, 1954 armament, bomb load, and other such items. Since the shield weight used was for a dose of 1 ¢/hr in the crew compartment, the useful load can also be construed to include any extra crew shielding required to reduce the crew dose to less than 1 r/he. The solution for gross weight was obtained graphically by plotting the weight of the propulsion machinery plus reactor and shield against gross weight, as in Fig. 3.1. The lift-to-drag ratio for the aircraft was taken as a function of the flight design condition, with allowance for the fact that the flight design condition would, in general, not give the optimum lift-to~drag ratio obtainable with the airplane, because take-off, landing, and climb requirements would necessitate wing loadings lower than those for minimum drag. The values used for the various flight conditions considered are given in Table 3.3. The lift-to-drag ratios given are for the airplane configuration without nacelles, an allowance for nacelle drag having been deducted from the specific thrust in Table 3.1. Thus the lift-to-drag ratio with nacelles would be lower than that indicated in Table 3.1, particularly at high Mach numbers, The results of calculations for three design conditions ‘are shown in Figs. 3.2, 3.3, and 3.4. A number of important conclusions can be deduced from these curves, and perhaps the most important is thatthe gross weight of the airplane is relatively insensitive to reactor design parameters for the Mach 0.9, sea-level conditions, but it becomes quite sensitive for the much higher performance supersonic-flight conditions. [t is also evident that an increase in reactor temperature level of 100°F is more beneficial than a factor-of-2 increase in power density. It should be noted that the turbine air inlet temperature will be lower than the peak fuel temperature by roughly 400°F, depending on the heat exchanger proportions, Thus a turbine air inlet temperature of 1140°F corresponds to a peak fuel temperature of about 1540°F, In genera], it appears that the aircraft gross weight is not very sensitive to the degree of division of the shield, except in the range of reactor shield design dose rates below 10 r/hr at 50 feet. The effect is greater at dose rates below 10 r/hr, partly because the incremental weight of a given radial thickness of shielding material in- creases as the square of its radius and, hence, the shield. weight increases at a progressively more rapid rate as a unit shield is approached. A second 29 0t TABLE 3.1, Compressor Pressure Ratie = 6:1 CALCULATIONS FOR POWER PLANT SPECIFIC QUTPUT Ratio of Radiator Outlet Pressure to inlet Pressure = .90 i/3600)0)' ’ f a b c d e f g= fi ——— |- h f=— i k= i+ \3413/¢ e , Specitic g . . . Propulsion Turbine Specific Turbojet Engine MaK System Mach Alitude Inlet Specific Thrust Less Heat Constants installed Weight Weoight Machinery No. ) Temperature Thrust Nacelle Bre7 - 7 v " {1b/1b of Weight o CF) (Ib-sec/1b) Drag fu/secs “w/ib ; N {1b/1b of (Ibesec/1b) of thrust of thrust of air of thrust thrust) 0.6 Sea level 1140 25.7 25.2 6.22 6.69 15.25 0.605 0.494 1.099 1240 30.7 30.2 6.07 6.51 15,0 0.496 0.481 4.977 1340 35.5 35.0 6.04 6.46 14.81 0.424 0.478 9.902 0.6 35,000 1140 40,8 40,3 5.35 5.71 56.6 1.405 0.532 1.937 1240 45 44,5 5.54 5.91 55.9 1.255 0.550 1.805 1340 4B8.3 47.8 5.6 5.97 55,1 1.154 0.555 1.709 0.9 Sea level 1140 19.4 8.6 6.86 7.52 12,65 0.648 0.549 1.197 1240 24,8 23.8 6.73 7.40 11.89 0.500 0.533 1,033 1340 29.3 28.3 6.55 7.15 11.73 0.414 0.515 0.929 1540 37.5 36.5 $5.35 5.88 11.40 0.313 0.496 0.809 0.9 35,000 1140 35.8 34,8 5.63 6.11 43,5 1.250 0.555 1.805 1240 490 39 5.75 6,22 43.0 1.103 0.565 1.668 13490 43.5 42,5 5.8 6.26 42.4 0.998 0.579 1.568 1540 50.8 49.8 5.85 6.29 41,4 0.831 0.572 1.403 1.5 35,000 1140 24,5 20.0 6.30 .14 24,6 1.231 0.635 1.8465 1240 28,5 24.0 6,26 7.84 24,3 1.01¢ 0.612 1.622 1340 33 28.5 6,20 7.57 24.0 0.843 0.590 1.433 1540 40.5 36.0 6.16 7.32 23.6 0.656 0.571 1.227 1.5 45,000 1140 24.5 20.0 6.30 8.14 39.6 1.979 0.748 2727 1240 28.5 24.0 6.26 7.84 39.0 1.625 0.720 2.345 1340 33 28.5 6.20 7.57 38.6 1.353 0.696 2,049 1540 40,5 36.0 6,16 7.32 38.0 1.055 0.673 1.728 JEOdIE SSTADOUL ATHILRVND 4NV 1€ TABLE 3.2. PROPULSION MACHINERY WEIGHT AND REACTOR OUTPUT FOR YARIOUS THRUST REQUIREMENTS Ratio of Radiator Qutiet Pressure to Inlet Pressure = 0.90 Compressor Pressure Ratic = 6:1 TURBINE THRUST (b} MACH | ALTITUDE INLETY 10,000 15,000 26,000 25,000 30,000 40,000 50,000 60,000 NO, () TEMPERATURE T : - (°F) W | P Mo | P | W | P | W | P U W | P W b P f W, L P R, P 0.6 | Sea level 1140 10.99 | 66.9 | 16.48 | 100.4 | 21.95{ 133.8 | 27.45 | 167.2] 32.95| 200.7 | 43.95 | 267.6 | 54.95 {334.5| 65.90 | 401.4 1240 9.77 | 65.1.| 14.65 ] 97.6| 19.53| 130.2| 24.40 | 162.8| 29.30| 195.3 | 39.10| 260.4} 48.85 {325.5 58.60| 390.6 1340 9.02 | 64.6 | 13.52 | 96.9| 18.03| 129.2} 22.52 | 161.5} 27.05] 193.8 | 36.05 | 258.4] 45.05 {323.0| 54.05| 387.6 0.6 35,000 1140 19.37 | 57.1| 29.05| 85.6 38.70| 114.2| 48.40 | 142,8| 58,05| 171.3 | 77.45| 228.4 | 96.80 {285,5] 116,2 | 342,6 1240 18.05 | 59.1 | 27.05 | 88.6| 36.10| 118.2| 45,10 147.8 54.15{ 177.3 | 72.20| 236.4| 90.2 |295.5| 108.3 | 354.6 1340 17.09 | 59.7 | 25.65| 89.6{ 34.20| 119.4| 42.75| 149.2| 51.30| 179.1} 68.40| 238.8| B5.50 |298.5| 102.5| 358.2 0.9 | Sec level 1140 11.97 | 76.2 | 17,95} 114.3} 23.95] 152.4] 29.95| 196.5| 35.90| 228.6'| 47.90| 304,8| 59.90 |381.0{ 71.80 457.2 1240 10.33 | 74.0 | 15.50 | 111.0 20.65| 148.0| 25.80 | 185.6] 31.00| 222.0] 41.30{ 296.0| 51.70 |370.0] 62.00| 444.9 1340 9.29 | 71.5| 13.95| 107.2 18.60| 143.0| 23.20| 178.8{ 27.90| 214.5 | 37.15| 286.0| 46.45 |357.5| 55.75| 429.0 1540 8.09| 68,8 | 12.13| 103.2] 16.20| 137.6| 20.20 | 172.0} 24.30| 206.4| 32,35} 275.2| 40.40 {344.0] 48.50| 412.8 0.9 35,000 1140 18.05| 61.1| 27.05| 91.6| 36.10| 122.2| 45.10| 152.8| 54.15| 183.3| 72.20| 244.4| 90.2 |305.5(108.3 | 366.6 1240 16.68 | 62.2 | 25.00 | 93.3| 33,35| 124.4| 41.70| 155.5| 50.00| 186.6| 66.70| 248,8{ 83.40 |311.0/1060.0 | 373.2 1340 15.68 | 62.6 | 23.50| 93.91 31.35{ 125.21 39.20] 156.5| 47.00| 187.8 62,70] 250.4{ 78.30 {313.0| 94.00| 375.6 1540 143 | 62.9 | 21.05| 94.3] 28.10| 125.8 35,10} 157.2] 42.20] 188.6| 56.20] 251.6{ 70.25 | 314.4| 84.30| 377.2 1.5 35,000 1140 18.66 | 81.4 | 28.00 | 122.1] 37.30| 162.8| 46.70| 203.5| 56,00} 244.2] 74.70| 325.6| 93.30 |407.0|104,5 | 488.4 1240 16.22| 78.4 | 24.35| 117.6] 32,50} 156.8| 40.60 | 196.0] 48.70| 235.2| 64.90| 313.6| 81.20 [392.0] 97.30| 470.4 1340 14.33] 75.7 | 21.50| 113.6] 28.70| 151.4| 35.85| 189.2| 43.00| 227.1| 57.30] 302.8| 71.70 {378.5] 86.00| 454.2 1540 12.27 1 73.2{ 18.40{ 109.8 24.55! 146.4| 30.701 183.0{ 36.80| 219.6| 49.10} 292.8| 61.30 | 366.0] 73.6C| 439.2 1.5 45,000 1140 27.27} 81.4| 40.85] 122.1] 54.501 162.8{ 68.20| 203.5| B1.B0| 244.2{109.0 | 325.6{136.3 |407.0} 163.5] 488.4 1240 23,451 78.4| 3520} 117.6} 46.95] 156.8| 58.70| 196.0| 70.45| 235.2| 93.90| 313.6{117.30 | 392,0] 141.0| 470.4 1340 20.49% 75.7 | 30,75} 113.6{ 41.00] 151.4| 51,30| 189.2| 61.50] 227.1| 82.10} 302.8{102.5 |378.5! 123.0| 454.2 1540 17.28| 73.2| 25.90| 109.8{ 34.55| 146.4] 43.20| 183.0| 51.80| 219.5| 69.15| 292.8| 86.40 | 366.0| 103.6| 439.2 *W__ = Propulsion machinery weight, 1073 Ib, **P = Reactor power, megawatts,
    x >0, (x -~ 3w Yo = 675 + 175 cos .————;5—“” . 1 18 > x > 165 . For straight-line joins: x = 13.5, y_ = 5.585t0x = 16.5, c Yo = 5,135, **Equations for island: Yy = 3, 18 > x> 9, xIT ¥ = 3.75 + 0.75 cos ~9- . 8 2>2x2>0 37 ANP QUARTERLY PROGRESS REPORT ducted with air as the working fluid to facilitate changes in the vital pump volute—core inlet region, The disiribution and the direction of flow were determined by the use of tufts. A careful evaluation of the tuft patterns indicated the need for cutoffs in the pump volutes and turning vanes around one quadrant of the impeller periphery to obtain uniform flow distribution at the core inlet. After experi- menting with a number of arrangements, it was found that the configuration of Fig. 3.7 was the most satisfactory because it operated with no appreciable flow separation. This arrangement is the same asthat envisioned in the designs presented in the previous report, except in the details of the shape of the turning vanes. After the pump discharge—core inlet region had been investigated, a qualitative study was made of the flow in the core by using a tuft on the end of a wire probe. The results indicated that the rovnded flow nozzle shown at the core inlet’ is necessary to avoid regions of unstable and asymmetric flow in the core. A set of turbulator vanes placed in the core inlet passage ond designed to produce the vortex patternshownin Fig. 3.6 gave better velocity distribution through the core. From all indications, it appears that this arrangement satisfies the flow requirements, namely, axially symmetric flow and no separation at the core shell wall. Fabrication of the equipment required to convert the system to permit flow tests with water has been started. With the water system, it is expected that 7R. W. Bussard and A. P. Fraas, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL.-1649, p. 31 f£. ORNL--LR--DWG 199 SECTION A-A IMPELLER CUTOFRF TURNING VANES L IMPELLER DIFFUSER RING Fig. 3.7. Core Header Region of Reflector-Moderated Reactor. 38 a thorough and quantitative study will be made. Pitot-tube traverses of the core and header regions should establish the velocity distributions of interest and hence enable a meaningful heat transfer analysis to be made of the system. In addition, it should be possible to determine the hydraulic characteristic of the pump impellers for various operating conditions. REACTOR CALCULATIONS M. E. LaVerne ANP Division C. S. Burtnette USAF Specifications were prepared for a parametric study of a set of 48 related reactors in which the parameters of core diameter, reflector thickness, and fuel thickness will be varied over u wide range. fn this study the effects of the geometry of the reflector-moderated reactor on the physical quanti- ties of interest, such as critical mass, required mole per cent of uranium in the fuel, and power distribution, will be ascertained. Calculations on this sef of reactors are about two-thirds complete; the remaining calculations are being made as rapidly as available machine time permits, As part of a study of a 50-megawatt reactor with an 18in, core, a 4-in.-thick fuei annulus, and a 12-in,-thick reflector, reactivity coefficients have been obtained for Inconel, sodium, and beryllium as functions of regions within the reactor. In com- puting these coefficients, the amount of material M (of Inconel, sodium, or beryllium) was increased in each region by an amount AM without displacing any of the other components., The resulting change in reactivity produced by this idealized experiment can be of practical significance in critical experi- ments where voids already exist. In an actual reactor, the over-all change in reactivity could be obtgined by summing algebraically the effects of adding one material and removing another, The reactivity coefficients of Inconel, sodium, and beryllium are presented in Figs. 3.8, 3.9, and 3.10. Other data of interest are: 40.7 Ib 4.1175 % 10% cm? Critial mass Volume of reactor core Power density (normalized to 1 fission per cm3) 1.29672 At surface of island PERIOD ENDING MARCH 10, 1954 Minimum 0.58781 Atouter surface of fuel region 1.90198 Reactivity coefficient for p233 Ar/{AM/ M) 0.19526 Per cent thermul fissions 33.14% Critical experiment precalculations were speci- fied for (1) two-region reactors with 16- and 21-in.- dia uranium~-foil and Teflon-laminate cores and beryllium reflectors and (2) a three-region reactor with a beryllium reflector and island and o 21-in.- OD core (with and without Inconel core shells) similar to those in the two-region reactors, Caleu- lational difficulties make a firm estimate of critical mass impossible at this time; however, preliminary results indicate a critical mass of about 16 1b for the two-region reactor with a 16-in.-dia core. Addi- tional calculations have been requested for a more precise determination, ORNI_~LR-DWs 200 0 e ceemnamn T o l] ‘.. I o | o i i | : 1 . i 0.0t b e _.....:... cobeee b e | [ I I ! I I I I 1 | | I 1 1 | { 0,02 pr— {4 L brolrenenon 4o ] ' P | ) i 1 EL | b oo REFLECTOR o] REGION 3 Lo | ! g _00'5 v mmmmmmee e e - ____1.._ ._1._.. ...g._._._._ ,.'t‘.‘ o : | ISLAND . ; ] : | ' REGION 1 | H i o | REGION | REGION | il | 1 | -0.04 o Jl{8 12 -~ : . ; Si= ” | \ \ r I ~2 | "REGION 7 ‘\(\ [ | 3 CORE SHELL \ -0.05 b REGION 2 A "REGION & ... ] ' REGION 5 | O OB — A ] \ ,,,,,,,,,,,,,, e fiovc)? ----------------------------- sy e ————————— 7,-] _0.0a ............................. B A — e — i e ] e CORE SHELL REGION 4 -Qf)g s mmaa e ._._,AAA-: ......... 1 ........................... J 0 10 20 30 a0 50 REACTOR RABIUS {cm) Fig. 3.8. Reactivity Coefficient of Inconel as a Function of Reactor Radius., 39 ANP QUARTERLY PROGRESS REPORT ORNL—LR—-DWG 201 N ; | J L 000315 | | | - B ] Pl Pl ey I I ! ~ 0.0001 | Bl e 1 t 53 e | I ! T o Pl i ' r <] : 1wl o ~ | — 0000 I e e Lliers s ! 00002 CHBIE LS 202 - t ] z § =z IR 2 2 ~0.0003 fmem byt - s { o : & { -0.00048 i 1 FUEL .- —----—i»—-—-———------l—--— : - REGION 3 ‘ | | | —0.0005 1 : ,-«‘rrr ] 4 i | I l —0.00CE : @ L fod | IS_AND | i ; REGION 1 i i —oo007 50 | * B . - 1 i | i —00008 S \ - REFLECTOR - 1. el ;U | ~ | i —00009 b : T . N CORE | [~CORE SHELL | SHELL. | i, REGOON 4 | TOOC | gEgoN 2 1T H | ! ’ | ] ‘ ‘ —0.0011 t—m et L 0 10 20 30 40 50 €0 REACTOR RADIUS (cm) Fig. 3.9. Reactivity Coefficient of Sodium as a Function of Reactor Radius. REACTOR DYNAMICS W. K. Ergen ANP Division The results of previous investigations on the kinetics of circulating-fuel reactors were summa- rized previously,® but, in the mathematical sense, the results could not be proved rigorously at that Brownell of the Institute of Advanced Study was consulted in the matter and he prepared a paper? which proves some of the points with mathe- time, 8w. K. Ergen, The Kinetics of the Circulating-Fuel Nuclear Reactor, ORNL. CF-53-3-231 (Mar. 30, 1953). A somewhat improved version of this memorandum has been accepted for publication by the Journal of Applied Physics. 9At Brownell's suggestion, this paper will be sub- mitted to the Journal of Rational Mechanics under the title A Theorem on Rearrangements and Its Application to Certain Delay Differential Equations,” by Srownell and W. K. Ergen. 40 ORNL-LR-DWG 202 012 z I | ; ! i QA1 feeeee i i . i . .. . | P —— REFLECTOR —— i CORE SHELI Lo | ' 010 [-REGION 2\ oy | |l il I { | | 009 :; ,l : b bt i t - N i i [ - I1] FUEL i cos *‘hREGIDNa Fpor % IL* f* I | I i 0.07 {H__ — _‘T{ — i ] —-]I t N vy > 0.06 | — ____H - __.‘ } } \ i { :m 5 . i - i | i = ” ‘ I l i | |l iz J_u@ t gf 005 v e S i oy =18 oy 3 o I tetsl i — Q04 ca b et e ____.iL_._ ISLAND I | 2| & rg : ; REGION | | & | | L | | | oce : i S { P Jl I L 001 fommrmem ——t} | | | 1 Ny | o c o q N _ | . i CORE SHELL _| _ 00! REGION 4 | ! \ —002 b oo 0 10 20 30 40 60 REACTOR RADIUS (cmn) Fig. 3.10. Reactivity Coefficient of Beryllium as a Function of Reactor Radius, matical rigor. |t is expected that the new rigorous methods will be extended in the future to cover all the tentative results in the previous work, The considerations regarding the inhour formula of a circulating-fuel reactor, as mentioned in the previous report,]o 11 were summarized in a memo- rcmdum, COMPUTATIONAL TECHNIQUES R. R. Coveyou ANP Division R. B. Bate Corps of Engineers, U, S, Army The methods of reactor statics computations de- scribed by Holmes'? are being modified for use on the ORACLE. Since the old methods were designed for the then available [BM machine and since the ORACLE exceeds these machines in speed of 10y, K. Ergen, J. Bengston, and C. B. Mills, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL.-1649, p. 12, ”W. K. Ergen, The lnhour Formula for a Circulating- Fuel Nucleor Reacior with Slug Flow, ORNL CF- 53-12-108 (Dec. 21, 1953). 12h K. Holmes, The Multigroup Method as Used by the ANP Physics Group, ORNL ANP-58 (Feb. 15, 1951). computation, the modification of the technique will include the introduction of new features which tend to improve the accuracy of the results, The follow- ing such features will probably be incorporated: 1. A *'P, approximation’ of the angular distri- bution of the neutron flux, which is an improvement over the previously used “P, approximation,”’ will be used to increase the reliability of the technique, particularly near the interfaces between media of different characteristics. In thin layers, such as those that occur in the various reflector-moderated reactor designs, each peint is necessarily close to such an interface. Consideration has been given to the replacement of the thin layers by special bound- ary conditions, but these special boundary conditions will probably become unnecessary with the improved treatment of the thin laoyers. 2. For each element, the cross sections will be averaged over each group and then the group aver- ages will be averaged over the various elements present, Previously, the opposite procedure was used; that is, for a number of lethargy points within a group the cross sections were averaged over the elements and then the group average was obtained. The new procedure greatly simplifies the compu- tation, 3. Self-shielding'? will be coded into the machine computations to eliminate the necessity of com- puting the self-shielding corrections to the cross sections by hand. 4. Inelastic scattering will be treated in a manner analogous to fission. The neutron is “‘absorbed” in one lethargy group and each such absorption gives rise to sources of neutrons in other lethargy groups. 5. The ‘‘Fox technique will be employed. This is o mathematical trick which eliminates the necessity of the introduction of ‘‘trial functions™ for the flux distribution in each lethargy group and the necessity of repeated iteration. The ‘‘Fox technique’’ gives the solution by traversing the space points in each lethargy group twice. The actual coding of the technique described above will be attempted in the near future, 14 ]3W. J. C. Bartels, Self-Absorption of Monoenergetic Neutrons, KAPL-336 (May 1, 1950). e R Coveyou and R. R. Bate, Three-Group Five- Region Spherically Symmetrical Reactors with Thin Sheél)s Between Regions, ORNL CF-53-11-136 (Nov. 23, 1953). PERIOD ENDING MARCH 10, 1954 BERYLLIUM CROSS SECTIONS C. B. Mills ANP Division it has been difficult to compute both '‘age-to- indium”’ and k_ for beryllium-moderated systems. Direct use of tabulated values of o and o has given 62 cm? for the age and 1.00 for the k¢ of a small reactor. Use of a p-scattering correction to adjust the oge to 80.2 cm? results in a kg value of 0.90. This inconsistency is not serious for most reactors of design interest, for which the error in kog¢ with p-scattering is about 2%, Therefore it has been sufficient to assume an inelastic and an (n,2n) cross section of the proper magnitude to absorb the k_ error. The cross sections in the 10 > E > 0.2 Mev neutron energy range were then adjusted to give a neutron escape value the same as that with o_ or 0, 2, reactions, The effects of angle-scattering corrections on o, and £ for several assumptions on symmetry of scattering in the center of mass system are: Case 1. For s-scattering, isotropic in the center- of-mass system: fi(cos@) =1, Utr = “ hb)as ! 2 b o= — 34 alna = = 1 4 -, & =&, tTT i 2 A+ 1, Case 2. For p-scattering (groups 1te 8§ 10 > E > 0.2 Mev): plcos ) = 1 + Ccos 8, O = (] Eb)as ' 2 (1 1 ) b= opef - — 1, 34 3 542 1 ] = 1 - - . €= % C(l——a 2§0> 41 ANP QUARTERLY PROGRESS REPORT Case 3. For d-scattering (groups 1 to 5, 10 > ¥ > 0.8 Mev): plcos 6) = 1 + 2 cos? @, 2 2a 34 154 b o= -, d 1 + — +§;C—;)(a+5)(a-—l)} . The cross-section curves presented in AECU- 204013 and the p- and d-scattering effects described above can be used to obtain o correct ‘‘age-to- indium'’ (80.2 cm?) value if a p-scattering correction value is assumed that varies uniformly from 0 at 0.2 Mev to 1 at 10 Mev. At 10 Mev, 45/{,:0 = 0.640 and (1 — 8)/(1 ~ b,) = 0.654, where £, = 0,208 and by = 0.0745. To compute the correct ket (1.00) for a small (B? = 0.0085) beryllium-moderated critical experiment, an inelastic scattering cross section of near 0,076 barns must be assumed [an (n,2n) reaction cross section can be one-half this value]l, With this assumption, the various numbers can be com- puted reasonably well. In particular, 7= 80 cm?, kg = 1.00, and @, and &, in the 7 > I > 1 Mev region are 2.12 ond 145 in comparison with ex- perimental values of 2,18 + 0,05 and 1.37 £ 0.11 barns.'® The cos? & term adds 1.3 em? for a = 0.55; the first flight correction is 3.9 cm?, and the last flight adds 0.6 emZ2, ]6Em T. Jurney, Inelastic Collision and Transport 1(‘.'.'9?15)5 Sections for Some Light Elements, 1L A-1339 (Dec. 17l:Jechns.trcm:ad on a pilot-plant scale at K-25; S. H. Smiley, D. C. Brater, and R. H. Nimmo, Metal Recoveiry Processes, K-901, Part1 (Mar. 10, 1952). 1sDeva|oped and demonstroted by the Materials Chemistry Division; G. J. Messle et al., ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-160%, p. 15. 42 CHEMICAL PROCESSING OF FLUORIDE FUEL 8Y FLUORINATION F. N. Browder D. E. Ferguson G. |. Cathers £. O, Nurmi Chemical Technology Division A new, nonaqueous method for processing fuels of the NaF-ZrF,-UF, system from an actual aircraft reactor has been shown to be feasibla. of three steps: (1) recovery of the vranium by converting the uranium tetrafluoride in the molten Naf-ZrF -UF ; mixture to the yolatile hexafluoride, using elemental fluorine; (2) gas-phase reduction of the partially decontaminated UF, to UF4;]7 and (3) refabrication of the molten salt fuel from this UF4.18 Scouting runs made by the Chemical Technology Division on the first siep of the process have shown that more than 99% of the fission products and less than 0.05% of the y23s would remain in the original NaF-ZrF, mixture and be discarded to waste., The process appears to be attractive from the standpoint of cost, inventory of fissionable material, and radiocactive waste volume, The major chemical cost is for the hafnium-free zirconium fluoride used in the fuel, which is not recovered, and would be about 25¢ per gram of U235 processed, at present prices. The fluorine would cost only 1¢ per gram of U23% processed, based on a 4% fluorine efficiency and a unit cost of $1.00 per pound for fluorine. The process appears to be less hazardous than the BrF, or It consists CIF3 processes for uranium recovery, since it can be operated at atmospheric pressure or under a slight vacuum. In the scouting runs, fluorine gas was passed through 100 g of NaF-ZrF -UF, (50-46-4 mole %, 8.5 g of uranium per charge) at temperatures above the melting point (530°C), and the volatilized UF was recovered in a dry ice trap. In some cases, the product was resublimed under vacuum. The fluorination was carried out in an lnconel vessel in a table furnace with o dip tube for bubbling fluorine through the molten atmospheric pressure was used to minimize any hazard from leaks. Unused fluorine passed into o chemical trap of soda lime and alumina. salt; The flucrine flow, which was approximately 100 ml/min over a period of several hours, was controlled by ¢ needle valve feeding into o glass Rotameter type of flow meter, but pressure changes in the supply tank of known volume were considered to bz more reliable than the flowmeter for estimating the total amount of fluorine used in each run. In the first two runs (Toble 3.5) with 4 and 13 times the theoretical amount of fluorine, 43 and 99.7% of the uranium was volatilized, as determined by analysis of the NaF-ZrF, residue. In the next three runs, in which more than 20+fold the stoichi- ometric amount of fluorine was used, 99.93 to $9.97% of the uronium was volatilized and good material balances were obtained. The low fluorine etficiency was probably partly due to the poor contact between the gas and the molten salt. There no evidence in any of the experiments of volatilization of ZrF . The line leading from the fluorinator to the trap was kept at 70°C to prevent deposition of UF,. The charge in the last three runs was spiked with fission products to the extent of 2 x 10° beta counts/min. Gross beta decontamination factors of 100 to 270 were obtained in the fluerination step (Table 3.6). The major contaminants of the Wdas PERIOD ENDING MARCH 10, 1954 product were ruthenium and niobium, ond radio- chemical analysis of the residuve showed that over 90% of the ruthenium and 60 to 80% of the niobium had followed the uranium. Sparging of the molten salt prior to fluorination would therefore lead to better decontamination from ruthenium and niobium, as well as from the more volatile short- lived fission products. The possibility of loss during nuclear operation of some ruthenium and niobium, as well as of halogen and rare gas fission products, is also indicated. ' In two runs, part of the UF, product was re- sublimed into a second dry ice trap under vacuum, An over-all gross beta decontamination factor of 4000 to 5000 was obtained in both cases {Table 3.6). The poor yield of 33% in one case was due to partial hydrolysis of UF, in the first trap be- cause of faulty drying ond conditioning of the apparatus. Better conditioning would undoubtedly improve the yield in both the fluorination and the sublimation steps. TABLE 3.5, URANIUM RECOVERY INFLUORINATION OF FLUORIDE FUEL Initial chorge: 8.5 g of vranium in 100 g of NuF-ZrF“-UFA (50-46-4 mole %) FLUORINATION RATIO OF FLUORINE URANIUM IN AMOUNT OF URANIUM TEMPERATURE USED TO THEORETICAL RESIBUE RECOVERED BY RESUBL IMATION (°C) REQUIREMENT (% of initial charge) (% of initial charge) 665 4:1 57 585 to 605 13:1 0.25 600 to 435 24:1 .02 29 620 to 655 26:1 .07 80 585 to 600 40:1 0.03 97 ANP QUARTERLY PROGRESS REPORT TABLE 3.6, DECONTAMINATION OF URANIUM BY FLUORINATION OF FLUORIDE FUEL Initial charge: 8.5 g of uranium in 100 g of NaF-ZrF4-UF4 (50-46-4 mole %} URANIUM * BETA DECONTAMINATION FACTORS RESUBLIMED PROCESS — Gross Ru Zr Nb TRE (% of initia charge) Fluorination at 600 to 635°C 270 14 4.2 x 10° 13 2.1 x 104 Fluorination at 620 to 655°C 100 6 1.4 x 104 6 3.0 % 104 Resublimation 5 % 10° 250 8.4 x 10 |5.8 x 10° | 4.6 x 10° 33 Fluorination at 585 to 600°C 230 17 1.9 x 103 5 6.0 x 10° Resublimation 4.4 x 10° 270 4.5 x 104 360 5.6 x 10° 77 *Decontamination factars for resublimed material include decontamination obtained in the fluorination step. PERIOD ENDING MARCH 10, 1954 4. CRITICAL EXPERIMENTS A, D. Callihan Physics Division SUPERCRITICAL-WATER REACTOR E. L. Zimmerman Physics Division J. 5. Crudele J. W. Noaks Pratt & Whitney Aircraft Division A critical experiment was assembled to test the designed core dimensions and core composition of the supercritical-water reactor, A 38-in., equi- fateral, cylindrical, aluminum tank is used for this experiment. An organic liquid (C5H402), which has a hydrogen density similar to that of water in the supercritical state, serves as the neutron reflector and as part of the moderator. The fissionable material, enriched uranium in an aqueous solution of UO,F,, is contained in 1-in.-dia stainless steel tubes. The effective loading for initial eriticality was about 5 kg of U?3% without stainless steel inserts in the core. Stainless steel is inserted in the core by loading %6-in.~OD tubes into the UO,F, solution. In the current experiments the designed dimensions and composition of the core are being approached by varying the height of the organic liquid reflector and moderator (assumed to be the effective core height) as a function of the number of fuel tubes loaded at increasing steel-to- uranium ratios. AIR-COCLED REACTOCR D. V. P. Williams J.J. Lynn D. F. Cronin C. Cross Physics Division J. D. Simpson R. C. Evans W. Baker H. E. Brown General Electric Co., ANP Division A preliminary assembiy of the AC-100-A, air- cooled, water-moderated reactor of the General Elec- tric Aircraft Nuclear Propulsion Project was made, The fuel is enriched-uranium metal disks placed be- tween steel disks. The disks are mounted inside aluminum tubes, 4 in. in diameter, The fuel section is 30 in. high. Thirty-seven aluminum tubes, in a pattern designed te give uniform radial power, con- stitute the core. The core is immersed in water, which serves as the neutron moderator and as an effectively infinite reflector. In order to make the system initially critical, it was necessary to deviate significantly from the prescribed loading by increasing the uranium from 26.6 to 45.6 kg and by decreasing the steel content by about one-half. A series of measurements is being made to ascertain the cause of the discrepancy between the reactivity of the experiment as designed and that which could be made critical. The details of the experiments will be reported by the General Electric Company. REFLECTOR-MODERATED REACTOR D. Scott B. L. Greenstreet ANP Division The critical experiment program for the reflector- moderated reactor has been altered to provide more fundamental information than that obtained by previous experiments. The earlier work was designed for direct measurements on rough mockups of possible reactors, with the purpose of es- tablishing design parameters. These mockups were, in general, of complicated geometry and usually contained materials unique to the unit being studied. The effort in the immediate future will be centered on reflector-moderated assemblies of simple geometry, and material variations will be made to check consistency with theory and the funda- mental constants. These results should alse aid in the evaluation of previous reflector-moderated critical assemblies. | It is now planned, first, to build a basic reflector- moderated reactor with two regions ~ fuel and reflector. The fuel region is to contain uranium and a fluorocarbon plastic, Teflon, to simulate the fluoride fuels, and the reflector region will contain beryllium. The fuel region is to be rhombicubocta- hedral (essentially, a cube with the edges and corners cut away) to approximate a sphere within the limitations imposed by the shape of the availa- ble beryllium. The purpose of this first experiment is to check machine calculations. The program is set up to then follow either of two alternatives, depending .on the results from the first assembly. 45 If the experiments confirm the theoretical calcu- lations to a sufficient degree, three-region octa- hedrons with a berylliom "‘island’’ separated from the reflector by the fuel will be byilt, The first of these assemblies will have no Incone! core shells, the second will include Inconel, and the final one will mock up the reactor, including the end ducts. In the event that poor agreement between theory and the first experiment is found, a second two-region assembly of different core size will be constructed. 46 In all cases, the fuel region will be built of alter- nating sheets of uranium metal and Teflon to permit some variation in the uranium density. The uranium sheets are to be 0.004 in. thick, and they will be coated with a protective film to reduce surface oxidation. The Teflon sheets will be 1/16 and %2 in. thick to make possible fairly homogeneous distribution of the uranium. The reflector and the reflector-moderator will be beryl- lium metal. Part Il MATERIALS RESEARCH 5. CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS W. R. Grimes Materials Chemistry Division A number of four-component systems with low UF, concentrations have been re-examined in the continuing effort to obtain fuels with physical properties better than those of the NaF-ZrF ;-UF, system. The low viscosity reported for a NaF-KF- ZrF -UF; mixture has served to stimulate interest in this system, and additional thermal data are being obtained. Other systems being studied by thermal analysis are NaF-LiF-ZrF,-UF,, NaF- BeF,-ZrF ,-UF,, NaF-LiF-Bein-UF‘i. Thermal analyses were also made of the RbCI-UCI; and NaCl-ZrCl, systems. The study of the NaCl-ZrCl, system was initiated to verify the low melting points reported in the literature for this system, A low-melting-point water-soluble mixture of this type would be of interest for removal of fluorides from equipment. Also, the gquenching technique applied previously to NaF-ZrF, mixtures has been used along with x-ray and petrographic examination of slowly cooled specimens to obtain a better understanding of the complex phase relationships in the NaF-UF, and NaF-ZtF, systems. High- temperature phase separation has proved to be a useful tool for studying systems in which solid solutions are expected and for which petrographic and x-ray examination can, accordingly, give only rough approximations of the composition of the separated phases. The systems NaF-ZrF,, NaF- ZrF ,-UF,, and NaF-KF-ZrF -UF ; were studied by using this technique. The experimental production facilities for the preparation of fluoride mixtures have been modified and expanded for supplying the research materials required for the long-range ANP program. The 250-1b capacity equipment has also been reactivated to supply the fluoride mixtures orderedby the Pratt and Whitney Aircraft Division of United Aircraft Corporation and the sizeable demands of the ORNL-ANP program. Analyses for silica in two purified batches of Sr(OH)}, confirmed the finding that less than 500 ppm of silica is introduced by passage of the material through o fine sintered-glass filter. Additional studies of the reaction of sodium hy- droxide with carbon have confirmed that graphite is oxidized by NaOH at elevated temperatures. Conse- quently, the carbonate content of molten NaOH will increase if any carbonaceous matter is present. The experimental study of the reaction of chromium metal wifhf UF, in molten NaZrFg was repeated with the use of improved techniques, and the reaction of iron with UF, in this solvent was also studied. These experiments have shown that the deviation of the activity coefficients from unity is probably due to the formation of complex ions such as UF;7, UF™7, and FeF;~. Two methods for producing NaZrF. melts contsining UF,; were developed. Cne sample, filtered at 600°C, was found to contain 3.54 wt % UF,, and another, filtered ot 700°C, contained 5.95 wt % UF,. Evidence was accumulated which showed that no mechanism for storing latent reducing power in a nonuranium-bearing fluoride melt is afforded by the presence of ZrF and, as a corrolary, that there are no prospects for imparting hydrogenous character to a melt by means of ZrH, solubility. Rates of reduction of NiF, and FeF, by hydrogen in nickel reactors were determined at 600 and 700°C. Further determinations of absorption spectra for UF, and UF,; in quenched fluoride melts were carried out with o Beckman DU spectrophotometer, and ad- ditional decomposition potential measurements in KCl melts in a hydrogen atmosphere were made. THERMAL ANALYSIS OF FLUORIDE SYSTEMS C. J. Barton H. Insley Materials Chemistry Division In the continuing effort to obtain fuels with physical properties better than those of the NaF- ZrF ,-UF, system, a number systems with low UF, concentrations have been re-examined. The low viscosity reported' for a NaF-KF-ZrF -UF, mixture has served to stimulate interest in this system, and additional thermal data are being obtained. of four-component ]H. F. Poppendiek, Physical Property Charts for Some Reactor Fuels, Coolonts and Miscelluneous Material: Third Edition, ORNL CF-53-3-261 (Mar. 20, 1953). 49 ANP QUARTERLY PROGRESS REPORT NoF.Lif-ZrF -UF, It was previously reported? that the ternary mixture NaF-l.iF-ZrF, (40-20-40 mole %) melted at 426°C; however, data presented in the same report showed that the melting points increased con- siderably with increases in UF, concentration, Additional data have been taken to confirm these tindings and to explore the effect of U, on other ternary compositions. The addition of 5 mole % UF, to the 40-20-40 mole % composition mentioned above gave a mixture witha melting point of 485°C, vhich is in reasonable agreement with the previous data. The lowest melting point so far established at the 4 mole % UF, level is 470°C for NaF-ZrF,- LiF-UF, (24-38.4-33,6-4.0 mole %). This mixture has a melting point that is nearly 50°C less thon that of NoF-ZrF4-lJF4 (53-43.4-3.6 mole %). NoF-BeF ,-ZrF -UF, Some thermal data were obtained with mixtures in the NaF-BeF,-ZrF ,-UF , system in connection with a study of BeF, pump seals.®? These data were collected by mixing various amounts of BeF, with a ternary mixture NaF-ZrF4-UF4 (50-46-4 mole %) and running cooling curves with the resulting mixtures. A limited thermal analysis of this four- component system has been conducted by adding UF, in amounts of up to 10 mole % to six low- melting-point ternary mixtures and observing the thermal effects when the fused mixtures are allowed to cool. In all cases the melting points of mixtures 2. M. Bratcher and C. J. Barton, ANP Quor. Prog. Rep. Dec, 10, 1952, ORNL-1439, p. 114, . M. Bratcher, R. E. Traber, Jr., and C. J. Barton, ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL.-1375, p. 78, 4. M. Bratcher, J. Truitt, and C. J. Barton, ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556, p. 40. containing 2 or 2.5 mole % UF, were higher than that of the ternary mixture. The minimum melting point at this uvranium level was about 450°C. At the 4 mole % leve!, the minimum melting point was near 460°C., Although this investigation can be regarded as only preliminary, the melting points observed do not sesm low enough to justify a detailed examination of the system unless it seems to be attractive because of other physical properties. NaF-LiF-BeF .UF, Twe compositions in the NaF-LiF-BeF,-UF, system were swubjected to thermal oanalysis in connection with the preparation of ternary mixtures for viscosity determination. The thermal effects observed on cooling curves, which may be too low because of supercooling, are shown in Table 5.1, The mixture NaF-LiF-BeF, (35-20-45 mole %) was chosen for viscosity tests to be made by the Physical Properties Group in the near Viscosity datc obtained from ancther installation, which indicate that LiBeF, has o much higher viscosity thon that of NaBeF,, sirongly influenced the choice. future. 5 THERMAL ANALYSIS OF CHLORIDE SYSTEMS A, B. Wilkerson R. J. Sheil C. 1. Barton Materials Chemistry Division The investigation of chloride fue! systems during the past quarter was limited to the binary alkali chloride-UCl; and -UCI, systems for which satis- factory equilibrium diagrams had not been obtained. 5Le‘H‘er, J. K. Davidsen to C, J. Barton, Density- Viscosity Data, ORNL CF-53-5-100 (May 13, 1953). TABLE 5.1. THERMAL EFFECTS WITH THREE- AND FOUR-COMPONENT BeF, MIXTURES COMPOSITION (mole %) THERMAL EFFECTS (°C) NaF LiF BeF, UF, 20 35 45 0 323, 295, 245 19 33.3 42.7 5 395, 288*, 237 35 20 45 0 330* 33.3 19 42.7 5 420, 307* w Indicates supercooling occurred. 50 The UCI, systems, particularly KC!--UCI4 and RbCl-UCld, have shown poorer reproducibility of thermal effects than the UCI3 systems, and the liquidus for these two systems cannot at the present time be located with certainty in the low-melting- point regions extending from about 40 to 60 mole % UCl,. It appears that other techniques, such as quenching, filtration, or differential thermal analy- sis, will be required to determine accurately the melting points and phase relationships in these regions. Petrographic examination has been less satisfactory for determining compound compositions in fused chloride mixtures than in the fluoride sys- tems becouse of difficulty in handling the very hygroscopic samples, reaction of the solid material with some refractive index oils, and lack of pure, crystalline compounds for standards. Oxide phases are apparently not so easily observed in chloride melts as in fused flucrides. An investigation of the NaCl-ZrCl, system was 900 ................... [ GO R + TEMPERATURE [°C!) RbCl 10 20 30 40 PERIOD ENDING MARCH 10, 1954 initiated to verify the low melting points reported in the literature for this system.® A low-melting- water-scluble mixture of the type found in the NaCI-ZrCl, system would be of interest for re- moval of fluorides from engineering equipment or, possibly, for final flushing of the ARE fuel circuit after operation. RbCI-UCI, Preliminary data for the RbCl UCI, system were reported previously.” A tentative diaqrom for this system, based only on thermal analysis dato, is shown in Fig. 5.1, Three compounds are indicated: RbSUCl , which melts congruently at 745 £ 10°C; Rb UCIS, which melts incongruently at 560 * 10°C; and RbUCH,, which melts incongruently at 550 = 6H. A. Belozerskn and O. Chem. {U.5.5.R.} 13, 1552 (1940). 7¢C. J. Barton and S. A. Boyer, ANP Quar, Prog. Rep. Dec. 10, 1953, ORNL.-1649, p. 52. J. Appl. A. Kucherenko, ORNL -LR-DWG 340 50 50 70 &80 920 UCly UCly {mole %) Fig. 5.1. . The System RbCI-UC|3 (Tentative). 51 ANP QUARTERLY PROGRESS REPORY 10°C. |t is possible that the latter compound melts congruently, but incongruent melting behavior seems to be more likely from the available data. The lowest melting point observed was 513 £ 5°C at approximately 45,5 mole % UCI . NaCi-ZrCl Eutectic composifions that melt at 390, 220, and 162°C ond the compounds Na4ZrC|8 and NaZrCls which melt at 535 and 330°C, respectively, were reported for the NaCl-Z¢Cl, system.® Neither the compounds nor the eutectic temperatures reported have been verified in this loboratory. Some compo- sitions in the range 11 to 45 mole % ZrCl placed in glass capsules equipped with thermocouple wells so that, in addition to thermal analysis, visu- al observation could be made. effect thermal were The lowest thermal observed was 355°C. The data obtained in this to indicate a congruently melting compound at 33.3 mole % ZrCi, with a melting point of about 615°C. The identity of the compound was con- firmed by petrographic examination of the fused melts; a single phase was indicoted at this compo- sition, Study of this system is continving. incomplete laboratory seem QUENCHING EXPERIMENTS WITH FLUORIDE SYSTEMS R. E. Thoma R. E. Moore M. S. Grim C. J. Barton Materials Chemistry Division G. D. White H. Insley, Consultant Metallurgy Division The quenching technique previously applied to NaF-ZrF, mixtures has been used along with x-ray and petrographic examination of slowly cooled specimens to obtain a better understanding of the complex phase relationships in the NaF-UF, and NaF-ZrF, systems. The materials used in these studies were, in each case, prepared in small batches by the hydrofluorination-hydrogenation technique described in subsequent poragrophs of this section. The quenching procedure was described in a previous report.® The slowly cooled specimens consisted of 10 to 15 g of the material contained in sealed capsules of nickel. These samples were SC. J. Barton et al., ANP Quar, Prog. Rep., Dec. 10, 1953, ORNL-1649, p. 54. 52 heated in a furnace to considerably above the liguidus temperature and allowed to cool to room temperature over a period of 2 to 4 hours. Data from o large number of experiments of each type are presented paragraphs. and correlated in the following Ncfl:-ljF"4 Previous therma! analysis of this system ? showed two compounds: Na,UF ., which melts incongruently, and NalF ., which melts congruently, Zachariasen also reported!® the compound N03UF7, and recent thermal analyses have confirmed the existence of this material, The compound Na,UF, appears to melt incongruently a few degrees above the eutectic temperature, While it appears that the major phase fields are well established in this system, the range of existence of the several stable modifica- tions of N02UF6 is still in doubt; there are also some questions regording possible modifications of Na,UF .. Zachariasen reported three forms of Na, UF a (cubic), B, (hexagonal), and ¥ (orthorhombics. However, slowly cooled compositions containing 22 to 50 mole % UF, prepared in this laboratory show a fourth form designated as S, (hexagonal). Data obtained by petrographic and x-ray diffraction of nine such specimens are shown in Table 5.2, When material of the: NazUFé composition was quenched from temperatures in the range 390 to to 747°C, it was not possible to obtain a glass without some crystalline material. The a (cubic) form of N02UF6 was the predominant phose at 610°C and above. At 604°C and below, the B4 form was the principal phase, One sample quenched at 590°C, however, showed nearly pure 3, crystals. Further studies will be necessary to determine the stability range of the S8, form and to explain the absence of the v form, The observance of crystalline phases with the NG3UF7 structure with refractive indices ranging from about 1,414 to 1.428 indicates possible solid solution formation between Na;UF, and Na,UF,. Crystalline phases with approximately the same refractive index but with different degrees of bire- fringence, along with some crystals which show anomalous interference colors, are also observed. crystalline ). P. Blakely et al., ANP Quor. Prog, Rep. Mor. 10, 1951, ANP-60, p. 128, mW. H. Zachariasen, J. Am. Chem. Soc, 70, 2147 (1948). TABLE 5.2. PHASES PRESENT IN SLOWLY COOLED NaF-UF , MIXTURES COMPOSITION PHASES IDENTIFIED BY (mole % UF ) PETROGRAPHMIC AND mate 7 PV 4’ | X.RAY DIFFRACTION ANALYSIS 22 Ba-Na UF, Naf 25 flS-NuQUFé, Na3UF7, NaF (trace) 28 B4-Na,UF ¢ (major), NagUF, 30 f3-Na,UF , (major), NajUF, 33.3 B3-NayUF 37 B4-Na UF o (major), NaUF ¢ 40 83-N02UF6, NaUF . 43 B3-Na,UF, NallFg 50 NaUF ¢ (majar), BS-NGQUFé, UF,, U02 (trace) it appears likely that there are also different crys- tolline forms of Na UF.. Thermal analysis of compositions in this region showed thermal effects well below the solidus temperature. Quenches with 37, 40, and 43 mole % UF , compo- sitions ot 690 to 7'!9°C preduced only |sotrop|c material believed to be glass. Quenches with the 50 mole % NaF-50 mole % UF , mixture in the same temperature range produced fibrous birefringent crystals, The refractive index of these crystals was approximately the same as that reported for NOUFS. Glass may have been present in some of the quenched samples, A melting point of 710°C was reported earlier for the 50-50 composition,’ but a more recent determination gave 714°C; this is the value reported by Kraus. ! NaF-ZrF The methods and apparatus described previously have been used for further studies of the NuF Lk, system. Because the published thermal data'?2 seem to furnish an adequote picture of the equilibrium diagram from pure NaF to 30 mole % ZrF,, quench- ing experiments have been largely confined to work ”C. A. Kraus, Phase Diagrams of Some Complex Salts of Uranium with Holides of the Alkali and Alkaline Earth Metals, M-251 (July 1, 1943). 12¢. . Borton et al., ANP Quar. Prog. Rep. Dec, 10, 1953, ORNL 1649, p. 54, PERIOD ENDING MARCH 10, 1954 with materials of higher ZrF , content. Petrographic and x-ray diffraction examinations of slowly cooled preparations have led to postula- tion of a compound at 40 mole % ZrF This ma- terial, for which published dcfi'a13 show some evidence, is nearly isotropic and often exhibits a distinctly Fibrous structure; its refractive index is 1.470 (birefringence, 0.004). Na,ZrF appears in preparations containing as much as 42 mole % ZrF ,, the 40% compound (Na,Zr,F, I) must melt incongruently, Examination of quenched specimens of 43 mole % ZrF ,, which were previously equilibrated after cooling from higher temperatures, has shown this material to be close to a eutectic composition; the melting point appears to be 495°C, Since anisotropic Slowly cooled specimens of the 50 mole % ZrF material have produced mixtures of two phases. One of these phases, with refractive indices of O = 1.508 and E = 1.500, was previously believed to be NaZrF the other material, with refractive indices « = 1 420 and y = 1,432, is believed to be NGBZrde.” In addition, a third phase known as R-3 (refractive indices of O = 1,445 and F = 1.417), which is described below, occurs frequently in compositions containing 47 to 57 mole % ZrF,. Studies of this system are complicated by the foct that liquidus and solidus temperatures are very close together in this region and, in oddition, liquidus temperatures appear to depend on previous thermal history of the specimen. Somples quenched after heating to a high temperature (750°C) followed by equilibration at temperatures near the liquidus show lower liquidus temperatures than those which have never been above the equilibration tempera- ture. However, the following observations seem to be justified. Data from quenching experiments, in agreement with previous thermal data, show the liguidus temperature at 57 mole % ZrF, (near the Na Zr F ¢ composition) to be 530°C. The primary thSe from 52 to 57 mole % ZrF appears to be Na ‘?_'r‘,,’l:1 Below the solidus temperature (510°C when ap- proached from above, 519°C when approached from below), NaZf“F] 9 and the material formerly believed to be Nc:ZrF5 are found over the composition iterval 52 to 57 mole % ZrF4. By M. Brmcher and C. J. Barton, ANFP Quar. Frog. Rep. Dec. 10, 1952, ORNL -1439, p. 112, ldR. E. Moore, C. J. Barton, and T. N. McVYay, ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609, p. 1., 53 ANP QGUARTERLY PROGRESS REPORY However, if NaZrF5 and N032r4F]9 coexist be- low the solidus at 50 mole % ZrF4, then Na,Zr F g must be the primary phase. Numerous quenches at this composition have shown that this is not the case. When equilibrium is approached from higher temperatures with 50 mole % ZrF,, only glass or glass with R-3 is found above 511°C. Below 511°C, “NoZrF "’ and NayZr,F, o are found, When equi- librium is approached from below, ‘‘NaZrF." and Na,Zr,F,, coexist at as high as 519°C; above this temperature, glass or glass and R-3 are found. When equilibrium is aopproached from higher temperatures with 47 mole % ZrF,, the liquidus temperature is 510°C; “*NaZvF.'' and liquid co- exist at as high as 523°C when approached from below, The solid phase is "“NaZrF,," alone, in either cuse, While other explanations are possible, it appears that the crystalline material of refractive indices O = 1308 aend £ = 1.500 is not NaZrF, but NagZrF , i this material probably melts congruently at 523°C and has a eutectic with NajZr F o at about 52 mole % Zrf,. The NayZr F | compound forms solid solutions with Na,Zr, P, in which the optical properties vary in a uniform manner with composition, The unidentified phase R-3 which has appeared over the 45 to 57 mole % ZrF, range has almost without exception been associated with o gloss phase. In one case, o sample containing 49.9 mole % ZrF, that was quenched from o temperature far above the liquidus was nearly pure R-3, Samples of this material equilibrated at 485 and 498°C be- fore quenching vyielded crystals too small for petrographic examination; x-ray diffraction indi- cated that the matericl had decomposed completely into NagZr F . and Na,Zr, F, . It is possible that R-3 is the compound NchFs. Samples of Zrf, containing 1 to 5 mole % NaF, cooled slowly from temperatures above the liquidus, show coexistence of ZrF4 with small amounts of N%Z'fw It seems obvious that no complex compounds of ZrF, content higher than 57 mole % exist in this system. ]SR. J. Sheil and C. J. Barton, ANP Quor. Prog. Rep. Sept. 1, 1953, ORNL-1609, p. 61. 16¢. 5. Barton and R. ). Shzil, ANP Quor. Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 55. 54 FILTRATION ANALYSIS OF FLUORIDE SYSTEMS C. J. Barton R. J. Sheil Materials Chemistry Division High-temperature phase separation by filtration, as described in previous reposrts,!3+16 has been applied to & number of materials during the past quarter, This technigue has, in gereral, been reserved for studies in which solid solutions are expected and for which petrographic and x-ray examingtion can, accordingly, give only rough ap- proximations of the compositions of the seporated phases, Since enough material is used in the fil- tration experiments to provide samples adequate for chemical analysis, the filtration technique is a useful tool for studying solid solutions, NQF-Z:’F4 In filtration of NaF-ZrF samples containing 37 mole % ZrE, at 560°C, 96% of the charge was re- covered as filtrate; similar filtration at 537°C yielded 75% of the charge. |In each case, the residue was found by petrographic examiration to be predominantly Ma,ZrF,. Chemical analysis of the residue at 537°C confirmed this finding., The filtrate cooled in each case to a mixture of a nearly cubic phose and NGQZrFé; the cubic phase is, presumably, NGSZrzF”, as described above, There seems to be no evidence for solid solutions in the range 33 to 40 mole % ZrF . NaF-ZrF -UF Twe filtrations with NaF-ZrF ,-UF samples con- taining 85 mole % NaF and 7.5 mole % ZrF4 indi- cate thai NaF is the primory phase in this region, When NoF precipitates sufficiently to reduce the concertration of this moaterial to 81 moke %, « second phuse appears which is rich in ZrF4,' composition of the liquid then moves toward that of the NaF-UF, binary. Additional studies in this region will be made os time permits, since it op- pears thot previous therma! analyses do not reliably indicate the ligquidus temperature, Some data on the pseudo binary system Na,UF ;- Nassz4 were presented in a previous report.'é Additiona! filtration dota obtained during the past quarter have shown essentially complete miscibility of these compounds in the seolid state, ot least to 92 mole % Ma UF,. Behavior of materials of higher uvranium cortent will be determined as time permits. NaF-KF-ZrF +UF Quite low melting points can be obtained with the ternary system NuF—KF-ZrF4.]7 However, ther- mal analysis indicated that the melting point of such low melting compositions was raised considerably by small amounts of UF,. This point has been checked during the past quarter by filtration of material containing NaF-KF-ZrF, (5-52-43 mole %) to which excess UF, had been added. Only 1.2 mole % UF, and 2.4 mole % UF, were dissolved at 426 and 503 C, respectively. Whtle these values do not necessarily represent the maximum solu- bility obtainable ot these temperatures, they do verify the low solubility of UF,, as shown by thermal analysis, PRODUCTION OF PURIFIED FLUORIDE MIXTURES F. F. Blankenship G. J. Nessle Materials Chemistry Division L.aboratory-5cale Production and Purification of Fluoride Mixtures G. M. Watson C. M. Blood F. P. Boody F. F. Blankenship Materials Chemistry Division During this quarter, a total of 21 batches of various fluoride mixtures was prepared or purified in laboratory-scale apparatus. Seventeen of these batches were special preparations to be used in reduction and kinetic studies for which a higher than normal degree of purity is required. The de- sired purity was obtained by hydrogen reduction of the structural metal fluorides at 800°C to approxi- mately 3.0 x 1073 mole of HF per liter of effluent gas, Purification of Fluoride Mixtures for Phase Studies F. P. Boody Materials Chemistry Division Nine batches of small samples of fluoride mix- tures with a total of &3 different compositions in amounts of 25 g each were purified during the past quarter.'® The occurrence of large differences in ‘71.. M. Bratcher, R. E. Traber, Jr., and C. J. Barton, ANP Quar. Prog. Rep June 10, 1952, ORNL-1294, p. 84. 18 F. Blankenship, C. M. Blood, and F. P. Boody, ANP Quar. Prog. Rep. Dec. 10, 1953 ORNL-1649, p. 5. PERIOD ENDING MARCH 10, 1954 melting point and in vapor pressure with small changes in composition made it necessary to re- strict the mixtures in a single batch to a narrow composition range (10 mole % in the case of Zer) Usually, the samples, contained in platinum cruci- bles, were treated with HF for about 90 min at temperatures up to 700°C and then treated with H, for 90 min. at 500°C and allowed to cool under helium, This procedure was effective in removing oxides and hydrolysis products; however, a small amount of unidentified black scum appeared on the surface of each melt. Experimental Production Facilities G. J. Nessle J. E. Eorgan J. P. Blakely F. A, Doss C. R. Croft R. G. Wiley J, Truitt F. H. DeFord Materials Chemistry Division A total of 3485.5 kg of fluoride mixtures was pre- pared during the quarter in the equipment in Building 9928, Of this quantity, the greatest portion con- sisted of NaF-ZrF -UF, (50-46-4 mole %) and NaF- Zrl“:‘4 (50-50 mole %). In order to eliminate the use of any fluoride mixture of unknown purity which could possibly invalidate important research, a policy has been instituted which does not allow release of any material until all analyses have been completed and evaluated, Planning for conversion of the facilities in Build- ing 9928 for producing berryllium-bearing fluoride mixtures is essentially complete, with actual work scheduled to begin in a few weeks. Recommenda- tions of the Y-12 Health Physics Division have been followed in the planning of this equipment revision, ' The additional experimental fluoride facility being installed in the basement of Building 9201-3 is progressing rapidly, insofar as the processing equipmemnt is concerned. Installation of utilities to serve this equipment is expected soon. These changes and equipment increases are due to the need for greater experimental versatility in fluoride preparations. Until recently, the entire efforts of this group have been directed toward pro- viding sufficient quantities of processed fluorides to allow all testing necessary to the ARE program. Since the ARE program is nearing completion, the fluoride production group will be able to direct its main efforts to exploration of new fluoride compo- sitions which will be important in the long-range 55 ANP QUARTERLY PROGRESS REPORY ANP program. Present plans call for the experi- mental production to become less routine and to deal mainly with new fuel and coolant mixtures, Any large, routine processing required in the future will be done with the 250-Ib units in Building 92013, Production-Scale Facility J. E. Eorgan F. A, Doss R. Reid R. G, Wiley M. S. Freed Materials Chemistry Division The 250-1b capacity equipment for the production of fluoride mixtures has been reactivated to supple- ment the production of the pilot-scale equipment, About 750 b of NaF-ZrF, (53-47 mole %) was processed and dispensed to requestors of this material, This operation included repackaging the material from lorge (250-1b) receivers to receivers ranging in size from 10- to 600-1b capacity, As indicated above, this equipment is to be utilized for the production of large guantities of fuel and coolant mixtures. For some months to come, a large order placed by the Pratt and Whitney Aircraft Division of United Aircraft Corporation plus sizeable demands by the ORNL-ANP program will make it economically desirable to operate the 250-1b unit on a nearly continuous schedule, Facilities are to be built to allow reducing the 250-Ib batches to any convenient size that may be requested, PURIFICATION AND PROPERTIES OF HYDROXIDES Purification of Hydroxides i, E. Ketchen L.. G. Overholser Materials Chemistry Division The effort devoted to the purification of hydroxides during this quarter was considerably reduced from previous levels. Only two batches of Sr(OH):2 were purified; the purified material presently on hand totals approximately 5 Ib of S¢{OH), containing less then 0.2 wt % H,0O and fess than 0.1 wt % SrCO,, Five batches of NaOH were purified by filtering a 50 wt % aqueous solution of NaQH through a fine sintered-glass filter to remove Na,CQO, prior to dehydration. The purified material contained less than 0.1 wt % H,O and Na,CO,; these values are in agreement with those obtained for earlier batches, 56 Analyses for silica in these lots confirmed the finding that less than 50 ppm of this material is introduced by passage through a fine sintered-glass filter., Reaction of Sodium Hydroxide with Carbon E. E. Ketchen L. G. Overholser Materials Chemistry Division In a previous report,w preliminary results were presented which indicated that carbon or carbe- naceous matter would react with NaOH at 700°C to yield Na,CO,. Further studies have shown that graphite is easily oxidized by NaOH at temperatures of 500°C or higher. Graphite (0.2 wt %) was added to NaOH contained in nickel capsules, which were then sealed under inert atmospheres and heated for 24 hr at 500°C. This treatment increased the NOZCO the caustic from 0.05 to 0.3 wt %. !n similar tests at 700°C, the Nc:2CO3 concentration increased from 0.06 to 0.6 wt %, That this increase is not due to NiO on the capsule walls or to diffusion of air through the capsules has been shown by using hydrogen-fired capsules and enclosing the sealed capsules in a quartz envelope filled with purified helivm, The results leave no room for doubt that graphite is oxidized by NaOH at elevated tempera- tures, Consequently, the carbonate content of molten NaOH will increase if any carbonaceous matter is present, content of CHEMICAL REACTIONS IN MOLTEN SALTS . F. Blankenship L. G. Overholser W. R. Grimes Materials Chemistry Division Chemical Equilibria in Fused Saolts L. G. Overholser J. D. Redman C. F. Weaver Materials Chemistry Division Some preliminary results from experimental study of the reaction of Cr° with UF, in molten NaZrF . were reported previously.?? These studies have been repected with the use of improved techniques, and similar, but probably more accurate, values 19g. E. Ketchen and L. G. QOverholser, ANP Quor, Prog. Rep. Dec, 10, 1953, ORNL.-1649, p, 63. ZOL, G. QOverholser, J. D, Redman, and C. F. Weaver, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 82. hove been obtained. In addition, the reaction of Fe® with UF, in this solvent has been studied, and some equilibrium dota have been obtained at 400°C and at 800PC. Preliminary study of the reaction of Cr® with FeF, in this solvent has shown that the reaction proceeds nearly completely to CrF, and Fe®; consequently, values reported for equilibrium constants for this reaction are approxi- mations only, Study of the reaction 2UF ((solution} + Cr(crystal) e=== 2UF , (solution) CrF2(50|ufion) was confinved by using the apparatus previously described, In these experiments, approximately 2 g of chromium metal was placed in the nickel charge bottle and treated at 1200°C with dry hydrogen for - PERIOD ENDING MARCH 10, 1954 2 hours, The required quantities of UF, and pure NaZrF. were loaded into the bottle in a vacuum dry-box to avoid exposure of the container or con- tents to air or water, The system was assembled, tested for lecks, and heated to the predetermined temperature, After equilibration, the molten mix- ture waos filtered, and the solidified filtrate was prepared for analysis, The results of o number of experiments are shown in Table 5.3. The K_shown in Table 5.3 is de- fined as 2 X UF:3 CrF2 G XZ u F4 where the X values are the concentrations of the species af equilibrium expressed as mole fraction. TABLE 5.3. EQUILIBRIUM DATA FOR THE REACTION OF C:° WITH UF, IN MOLTEN NoZrFg EXPERIMENTAL CONDITIONS CONCENTRATION OF Time Temperature UF, Added Ce*t IN FILTRATE K (h) (°c) (moles/kg of melt)(@ (ppm){ 5 800 0 195 5 800 0 240 3 600 0.363 2310 4.0 x 10~4 5 600 0.363 2400 4.6 x 1074 5 500 0.363 2290 3.9 % 10~4 3 800 0.363 2880 9.5 x 10=4(e 3 800 0.363 2330 4.0 x 10~4 3 800 0.363 2410 4.7 x 10~4 5 800 0.363 3250 1.5 x 10=3(d 5 800 0.363 2450 4.8 x 10~ 5 300 0.363 2300 3.9 x 1074 5 800 0.181 1470 4.1 x 1074 5 800 0.181 1420 3.5 x 10~4 to) Total weight of UF, + NaZrFg = 40 grams. (b)gjank of 200 ppm to be subtracted from determined values in calculations. (C)Kx calculated from mole fractions of UF ;, UF,, and CrF,, assuming ane formula weight of NaZrFg =2 moles. These values omitted in subsequent discussion by Chauvenet’s criterion. ‘57 ANP QUARTERLY PROGRESS REPORT Since there is, as vet, no reliable method for ana- lytically determining UF; in the presence of Cr't, the UF; and final UF4 concentrations were calcu- lated from the measured Cr'' concentration. All indicates that the dissolved chromium is divalent; petrographic examination reveals that UF is formed. The K_values have been calculated on the basis that one formula weight of NaZrF, is 2 moles, that is, NaF plus ZrF ; this choice permits the use of rational activities based on mole fractions of the simplest pure components for all constituents of a melt. This is convenient, since estimates of the free energy of formation of pure components in their most stable states are readily available?! and are widely used for predicting feasibility of chemical reactions, Since for the previously reported?? values for K the formula weight of NchF5 was assumed to represent ] mole, the values given here are smaller by about a factor of 2, When compared on the same basis, the K_ values reported here are only slightly smaller than the previous values; the agreement in AF® for the reaction, as written, is within 500 calories. This difference is well within the ac- curacy of the estimates of free energy of formation available evidence 2]L,, L. Quill (ed.), The Chemisiry and Mefcr”un_fir of Miscellanecus Materials, NNES 1V.19B, McGraw-Hill New York, 1950. of the pure compounds at these temperatures. The values listed here are probably more accurate, since the soluble chromium “blank’ has been reduced from 900 to 200 ppm. The reaction 2UF ,{solution) + Fel(erystal) =—— 2UF3(so|ution) + Fer(solufion) has been evaluated by using similar techniques. Iron wire, hydrogen fired at 900°C, was used for these studies. The system was virtually free of FeQ; only 100 ppm of soluble iron was found in blank runs, In nearly all cases, analysis indicated that all the dissolved iron was in the ferrous state, and no UF , was detected petrographically. It is possible that the equilibrium concentrations are so low that no discrete crystals of the material are formed. The experimental dota, along with the K_ values calculated in the manner described above, are shown in Table 5.4, From the data shown in Tables 5.3 and 5.4, it is possible to estimate the relative magnitudes of the activity coefficients of materials dissolved in NaZrF’5 for use with free energies of formation of the pure substances, For the reaction UF, + Mg —22UF, + MF, , it would be possible to tabulate free energies of TABLE 5.4, EQUILIBRIUM DATA FOR THE REACTION OF Fe® WITH UF, IN MOLTEN NaZrF, CONCENTRATION OF TIME TEMPERATURE UF, ADDED ++ 4 Fe ' IN FILTRATE K {(hr) (°C) (moles/kg of melt)}* x {ppm) 5 800 0 100 3 600 0.363 770 0.8 x 10™° 5 600 0.363 990 1.7 % 1077 3 800 0.363 540 2.1 x 1078 5 800 0.363 790 3.5 x 108" 5 800 0.181 310 1.1 x 1078 5 800 0.72 710 1.3 x 1078 * Total weight of UF , + NeZrFg = 40 grams, * * This value can be omitted by Chauvenet's criterion. 58 formation of the pure compounds as supercooled liquids at various temperatures. By using these supercooled quuids as reference states, and equi- librium constant, K , could be defined by "‘AAFO == RT ln K(Z @ Then for the reaction in sofution in molten NaZrFs, 2 ' YUF, * YMF, Kg = Ky mimm - K, - K, Yur, where the y_ values, defined by Activity (Az.) =y, X mole fraction , represent deviation of the solution from ideality. However, if the standord states, that is, the crystal- line solids, are taken as the reference states, then the activity coefficient will include, ot temperatures below the melting point of the constituent, a cor- rection due to the free energy of melting. Since this correction is in many cases negligible in com- parison with the uncertainties in the tabulated free energy of formation, the conventional standard stafes have been adopted as the reference states. Table 5.5 shows the activity coefficient quotients calculated from data given in Tables 5.3 and 5.4. There is no evidence that K_ is dependent on con- centration over the narrow range studied; hence, PERIOD ENDING MARCH 10, 1954 an average of values at the same temperature would seem to be justified. The quotients of the appropri- ate K, values indicate that the ratio ye- p /yFeF is 140 at 600°C and 2.45 at 800°C, Some measure, although perhaps an optimistic one, of the reliability of these values may be gained by computing AFO for the reaction FeF (solution) + Fe® at 600°C from the experimentally determined value of K_and the value VCer/yFer = 140. Experi- + Cer(soluiion) mental study of this reaction has been attempted by appropriate modification of the technique de- scribed above. Since the reaction, as written, proceeds nearly to completion, accurate values for the constant are difficult to obtain, However, the best value obtained to date is K, =335 Then K, = K, x K, = 35 x 140 = 4.9 x 10°, and from this value, AF° (600°C) = ~14.8 keal, value which is in startling agreement with ‘rhe literature value of ~15 kcal, The following general conclusions can be drawn from these data. Deviation of the activity coeffi- cients from unity is probably due to formation of complex ions such as UFS“, UF™", FeFB“, etc. TABLE 5.5, APPARENT ACTIVITY COEFFICIENT QUOTIENTS FOR ZUFA(solufion) + Mo(crysfai)_%“*-—“"-—"-:—" 2UF3(soiufion) -+ MFz(solufion) 2 (9 CONCENTRATION OF () YUr. * YME. : TEMPERATURE | AF°® (b) 3 2 METAL SALT (mole fraction) K K (°Q) (keal) a x Cr 0.32 | 0.0091 | 0.0050 500 o | 1.0 4.2 x 104 2400 Cr 0.32 | 0.0091 | 0.0050 800 +4 1 0.1 4.3 « 107* 230 Cr 0.15 | 0.0052 | 0.0031 800 +4 | 01 3.8 % 10~4 250 Fe 0.37 | 0.0031 | 0.0017 600 +15 2 x 1074 [ 1.2 x 1073 17 Fe 0.39 | 0.0017 | 0.0011 800 +19 | 1.4 x 1074 | 2.1 1078 67 Fe 0.19 | 0.0008 | 0.0006 800 +19 | 1.4 x 1074 | 1.1 x 1078 127 Fe 0.78 | 0.0024 | 0.0014 800 +19 | 1.4 % 1074 | 1.3 x 10™8 108 (U)Based on free energies of formation of pure crystalline solids, (b)Ka is defined by: ~AF® = RT In K- (c)}/i is defined by: A, = y.X;, where A is activity (unity for pure components) and X is mole fraction. 59 ANP QUARTERLY PROGRESS REPORYT The value 140 for y- o /yg e at 600°C sug- 2 gests that CrF2 is largely uncomplexed at this temperature, while FeF, is strongly complexed, Since this value drops to 2.45 at 800°C, it appears that the complex ferrous jon is not very stable at high temperatures, It seems likely that of the dissociation reactions CGF, w—=CF, + F~ and FeF, we==>FeF, + F~ , the former is nearly complete at 600°C, while the latter is not yet complete at 800°C, H yc,g_ is taken to be nearly unity at 600°C, then y = /¥ £, is about 50 at 600°C and about 15 3 ot 800°C. Obviously, among reactions of the type UFSM?:T}_UF‘! . F° and UF4" \_—A_UFa + Ffi ’ the former are much less complete than the latter at both temperatures, This suggests that UF_ is hardly complexed, even at 600°C, while UF, is more than 90% complexed, even at 800°C, The increase in VlzJF VFer/leJF4 between 600 and 800°C shows that the complex ferrous ions are less stable toward dissociation at higher temper- atures than are the complex ions of U(IV). It is recognized that many additional experiments on these reactions are needed and that there may be alternative explanations of the dota, However, these conclusions are in good agreement with qualitative data obtained from phase equilibria, vapor pressure, and other experiments, It is be- lieved that these experiments are valuable for studying the nature of fused salts, and it is expected that these studies will be continued and, perhaps, augmented in the future, F ormation of UF3 in NaF-ZrF4 Melts C. M. Blood F. P, Boody G. M. Watson Materials Chemistry Division Two methods for producing NaZrF, melts con- taining UF, have been developed and studied, In the first method, dissolved UF, is reduced to UF, by using hydrogen or, preferably, metallic zir- conium or uranium, or zircenium or sodium hydride. 60 In the second method, added uranium metal is oxidized by the melt according to the reaction 4U%(crystal) + 3ZrF j(solution) ———s> 3Zr%(crystal) + 4UF3(so|ution) or, in the presence of hydrogen, according to the reaction 3H,(gas) + 4U°(crystal) + 3ZrF ((solution) —> 3ZrH,(crystal) + 4UF 4(solution) . The reduction of UF, in such melts by hydrogen gas is too slow to be of practical interest. Treat- ment at 800°C of 3 kg of a mixture containing 6.5 mole % UF, with 600 liters of hydrogen gas for 30 hr yielded 4.4 moles of HF per liter of exit hydrogen and, on the basis of HF preduced, should have converted about 5% of the contained UF, to UF,. Traces of free UF,; were visible on petro- graphic examination of the melt., Heretofore, olive-drab and orange-red reduced phases were found, but no free UF, was observed when ftri- valent uranium species were present in such low concentrations. However, a similor treatment of molten Na,UF, yielded a five- to tenfold lower rate of HF production; the amount of reduction produced by 850 liters of hydrogen in this system was not detected on petrographic exemination, These differences in rate, as well as the nature of the trivalent uranium species, are probably related to the influence of free NaF on the activity coef- ficient of UF,. They suggest that measurement of rate of reaction of UF, with hydrogen in various melts can provide a relative rating of corrosivity of the melt. Two batches of material containing UF, with virtually no UF, were prepared by the addition of uranium metal to NaZrF_ in an amount sufficient to produce 2 mole % of UF;. In one case, hydrogen was admitted to the reactor at a low temperature so that UH,; could be formed and subsequently dis- sociated as the temperature was raised. By using this technique, the metal was finely divided and o large surface area was exposed; the reaction was complete in 2 hours, attempt, hydrogen was admitted only at a tempera- ture too high for UH, formation, and after 2 hr of reaction time, only 90% of the uranium had reacted. The latter batch was filtered at 600°C, and « sample of the filtrate was found to contain 3.54 wt % of UF;. This material was circulated for 520 hr in uranium In the other an lInconel thermal convection loop, and no cor- rosion was detected. However, during transfer of the material to the loop, insufficient time was given for the segregated UF;-bearing phases to re- dissolve; accordingly, the material in the loop contained 3.31 wt % (1.23 mole %) of UF . Another sample to which 5.5 wt % of uranium (equivalent to 6.8 wt % of UFB) had been added yielded, on filtration at 700°C, a filtrate containing more than 85% of the added uranium (5.95 wt % of UF ). On petrographic examination, the UF;-saturated NaZirF, melts usually show NagZr F ., os the major phase, with some of the orange-red UF,;- 2ZrF, and small amounts of UF, and an un- identified yellow phase present. Treatment of Molten NeZrF ; with Strong Reducing Agents G. M. Waotson C. M. Blood F. P. Boedy Materials Chemistry Division Evidence accumulated during the past quarter has confirmed that no mechanism for storing latent reducing power in a nonuranium-bearing fluoride melt is afforded by the presence of ZrF, and, as a corollary, that there are no prospects for imparting hydrogenous character to a melt by means of ZrH, solubility, 1t was previously rs\aporha-dz2 that the loss of weight by zirconium metal in contact with liquid NoZrF. contained in graphite at 800°C far exceeds theloss that could be attributed to reaction with impurities. The possibility of zirconium metal going into solution, either as the trifluoride or as dissolved elemental metal, was strongly suggested by the relatively large reducing power (40 to 150 meq/kg) of the filtered melt, as found by chemical analysis. : ' In order to test further for the presence or absence of reducing power in these fuels, a freated and filtered melt having a reported reducing power of 80 meq/kg was treated with 100 meq/kg of nickel fluoride and brought to 800°C in a nickel container After a 24-br equi- libration period with continuous helium stirring at 800°C, hydrogen was bubbled through the melt, and the yield of hydrogen fluoride was measured. The number under a helium atmosphere, of milliequivalents of hydrogen fluoride 22?"'. F. Blankenship, C. M. Blood, and G. M. Waison, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 60. - PERIOD ENDING MARCH 10, 1954 thus obtained matched closely the vield expected for the reduction of the NiF, by H, and indicated that the zirconium-treated melt had little, if any, latent reducing power. This result is in agreement with the belief that intermediate valence states of zirconium are not stable under the conditions of these experiments and that elemental zirconium exhibits no physical solubility in fluoride mixtures at 800°C. A qualitative test for reducing power was made by adding known amounts (1 and 2 mole %) of vranium tetrafluoride to separate portions of a it . ¥y - . . reducing melt. The mixtures were contained in nickel crucibles and kept at 800°C for several hours in a helium atmosphere. The resulting melts were examined petrographically for the presence of reduced phases. Negligible reduction of the UF, was noted. A control experiment with zirconium hydride added at a concentration of 200 meqg/kg to give reducing power showed obvious reduction, even upon visua!l inspection. The conclusion was reached that no latent reducing power can be developed in NaZrFg by treatment with Zr® and that the apparent reducing power obtained by wet analytical methods was in error. Whether or not the formation of ZrC with the graphite liner was responsible for the previ- ously reported disappearance of zirconium is not known. It is apparently impossible to dissolve appreciable amounts of zirconium hydride in NaZrF.. In two trials, zirconium hydride was formed by the addition of 1 mole of sodium hydride per kilogram of NaZrF and subsequent heating to 800°C under sufficient hydrogen pressure to prevent the decomposition of the zirconium hydride. In each case, a residue, identified as zirconium hydride, remained in the reactor after filtration at 700 to 800°C. The presence of zirconium hydride in the filtrate was tested by the addition of UF, followed by equilibration and petrographic examination of the resultant mixture, as described above. This test indicated only o very slight reduction of the tetrafluoride. Since this method is, apparently, very sensitive, no further uh‘empfs at determination of the reducing power, if any, of the melt were made. uranium 61 ANP QUARTERLY PROGRESS REPORT Reduction of NiF , and FeF, by Hydrogen G. M, Watson C. F. Blood F. F. Blankenship Materials Chemistry Division Rates of reduction of nicke! fluoride by hydrogen in nickel reactors were determined at 600 and 700°C; in addition, a brief survey of the behavior of this reaction at 800°C was made. Comparison of the results of these experiments with similor trials in graphite-lined reactors show several differences. in nickel containers, the per cent recovery of hydrogen fluoride was independent of the length of pretreatment with inert gas and could be reproduced to within £4%. The rate of reaction is enhanced 12- to 14-fold at 800°C in nickel, as compared with graphite, although the gas inlet tube is nickel, regardless of the container construction. The increase in the rate of reaction is roughly pro- portional to the approximately ninefold increase in nickel surface in contact with the reacting mixture and suggests a catalytic effect by the nickel surface. An experimental reaction order approaching unity with respect to nickel fluoride concentration appears o hold for a portion of the reaction at 600 and 700°C. The kinetics of this graphite containers appear to be more complex. An energy of activation of about 2000 cal is obtained by comparison of rates at 600 and 700°C in nickel apparatus, The rapid decrease of hydrogen fluoride concen- tration with volume of hydrogen passed (half-life volume about 1 liter at 800°C for 3 kg of NaZrF ) suggests that stripping of the HF produced is the rate-controlling step and that the solubility of HF in the melt is low. A solubility coefficient, (HF)g/(HF)d, of 2 kg/liter is obtained if it is assumed that the experimentally measured HF concentration in the effluent gas represents equi- librium conditions. When rates of reduction of FeF, in molten NaZrFg contained in nicke! were measured, the yield of hydrogen fluoride was reproducible to within 2% and amounted to about 93% of the theoretical value if the ferrous fluoride used was assumed to be 100% pure. A comparison of the rates of reduction of nicke!l and ferrous fluorides at equol concentrations (10 meq/kg) showed that nickel was reduced about 60 times faster than was ferrous fluoride. reaction in 62 Spectrophotometry of Supercooled Fused Saits H. A. Friedman Materials Chemistry Division Determinations of absorption spectra for UF, and UF, in quenched fluoride melts with @ Beckman DU spectrophctometer have been continued.?® As previously reported,u two characteristic spectrum patterns have been found in the NaF-ZrF ,-UF, ternary. Special attention has been given to regions which might show a new pattern or ¢ transition between the two prevalent types. The results withrespect to pattern type are shown in Fig. 5.2, Transitions were found to occur as the ratio of F™ to UF, plus ZrF, changed from 1:1 to 2:1 (that is, between 50 and 67 mole % NaF). Inability to supercool in many composition ranges without forming crystals has hampered the work considerably. In order to obtain rapid cooling, samples which contained only 15 to 20 mg of fluorides were used, and of some 90 quenches performed during the quarter, in 25% of the cases crystals were successfully avoided, It is parti- cularly disappointing that binary mixtures of NaF-UF, containing 50 mole % and more of UF, have not given glasses, since the glasses would make possible an interpretation of some aspects of the NaF-UF, binary system and, perhaps, a determination of whether UF ™ exists in the liquid state. The spectra of a mixture containing 40 mole % UF, and 60 mole % ZrF, were obtained for both a crystalline solid solution and a glass. These spectra are compared with each other and with the spectrum of erystalline UF, in Fig. 5.3, An attempt was made to study the spectra of glasses in the UF -NaF binary system. The only compositions which could be quenched to glass were the 10, 20, 50, and 60 mole % UF; samples, and, even in these cuses, some quench growth occurred. These drab, brown glasses gave very poorly resolved patterns with no prominent maximums and minimums. In a preliminary attempt to produce a glass in the UF;-NaF system, potassium bromide was used as a dispersing medium. Crystalline UF, was ground with dried potassium bromide, and a flat, tfrans- parent window was made by compressing the mixed 23H. A, Friedman and D, G, Hill, ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609, p. 62. 24y, A. Friedman, ANP Quar. Prog., Rep. Dec, 10, 1953, ORNL-1649, p. 56. - PERIOD ENDING MARCH 10, 1954 QORNL. LR --DWG 341 UF, ® PATTERN FOR COORDINATION NUMBER OF FOUR. ¢ A PATTERN FOR CCORDINATION NUMBER OF SIX. N PATTERN FOR TRANSITION FRCM FOUR TO SIX. A OTHER PATTERN. o GLASS OF THIS COMPOSITION COULD NOT BE FORMED . NaF Na, ZrF. ZrFy Fig. 5.2. Distribution of Types of Spectra Exhibited by UF , in LiquidNaF-ZrF ;-UF , Solutions (Glasses). salt in @ vacuum press.25 When the spectrum of a window containing 17 mg of UF was compared with a potassium bromide window prepared similarly but without UF,, the results were disappointing because of the very large subtractions for the blank and an implausible apparent increase in the general absorption with decreasing wave length. Another effort to improve the reproducibility and resolution of the spectra involved the use of a supersonic generator to disperse the ground 25"The preparation of the KBr mounts was carried out by P. A. Staats and H. W. Morgan of the Stable Isotope Division. fluoride glass in a liquid of matching refractive index., Clumping, rather than dispersion, occurred. EMF Measurements in Fused Salts L. E. Topol Materials Chemistry Division Additional decomposition potential measurements in KCl melts in a hydrogen atmosphere were made. All the experiments were carried out at 850°C with nickel or platinum cathodes, nickel or graphite anodes, and Morganite alumina containers. Elec- trolyses of KCl result in decomposition potentials (E) of 3.15 volts with graphite anodes and 2.00 63 ANP QUARTERLY PROGRESS REPORY ORNL-L!- !WG 342 I\ CRYSTAL, OF, | . P / \ ra w/ ‘\ ‘\ QACQ ——e-s ~— CRYSTAL, 40 mole % U, 0.050 GLASS, 40 riit % U, | AND 6C moie ¥ 7rfy Lo | 300 400 500 600 700 WAVELENGTH (mg) OPTICAL DENSITY PER mg OF UF, Fig. 5.3. Comparison of the Absorption Spectra of ZsF ,-UF, (60-40 mole %) in Glass and Crystdl Form and Crystalline UF , volts with nickel anodes; both these values agree with those found by using a helium c:tmosphere,26 With chromium anodes (prepared by electroplating chromium on copper or gold wire) decomposition potentials at 1.45 to 1.52 volts are measured. Typical curves obtained for the electrolysis of KCI by wusing nickel and chromium anodes are shown in Fig.5.4. With the decomposition potential of 1.5 volts obtained with chromium anodes and the value of 3.15 volts obtained with inert anodes, the potential of the reaction CrCl, = G+ Cly in H, is computed to be 1.6 volts, At this tempera- ture the standard emf of CrCl, is 1.36 volts and that of CiCly is 1.08 volts. Thus it appears that divalent chromium is formed when the metal is oxidized electrochemically and that the effective Cr** concentration present is extremely small because of complexing by the KCli. Solutions of CrCl; in KCI (1 to 2 wt %) were electrolyzed at potentials of 0.98 to 1.10 volts with graphite anodes. With nickel anodes, CiCl, solutions show two distinct changes in slope on current vs, voltage curves with repeated FE-I measurements. The first emf ot 0.45 to 0.56 volt seems to be due to the reaction Ni + CrCl2 == NiCl2 + Cr , E° = 0.54 volt , while the second emf at 0.25 to 0.30 volt may be ascribed to the reaction 3Ni + 2CrCly = 3NiCl, + 2Cr, £° = 0.26 volt. 64 UNCIASSIFIED ORNL"LR-CWG 343 050 - ——" ----- | | i | 77777 - i NICKEL CATHODES ‘ | | 0.40 -———- i “ - |'fl*'~~ - | | : | | MICKEL ANODE . ‘ I CHROMIUM ANCDE - . __ ! | 1 0.30 b | ’a \ e - L ~ 0.20 040 oL 0 Fig. 5.4. Electrolysis of KTl at850°CinHydrogen, It should be mentioned that the reaction Ni + 2CrCl, = NiCl, + 2CeCl, will occur ‘‘spontaneous!y.”’ Similar experiments with equal weight mixtures of ZtCl, and KCl yield decomposition potentials at 1.00 to 1.05 volts and 0.55 to 0.70 volt. These values agree well with the decomposition potentials of 1.1,0.4, and Q.7 volts for the following reactions: ZrCl, + 2Ni = 2NiCl, + Zr, 2Z¢Cl, + Ni = NiCl, + fZZrCI3 ; ZrCId + Ni = NiC|2 + ZrCI2 . To increase the vaporization of ZrCl, on heating and to minimize the corrosion of the apparatus in further runs, the compound K,ZrCl, has been pre- pored by heating the appropriate mixture of chlorides to 700°C in an evacuaied, sealed quartz tube. 26L, E. Topol and L.. G. Overholser, ANP Quar. Prog, Rep. Sept, 10, 1953, ORNL-1609, p. 67. PERIOD ENDING MARCH 10, 1954 6. CORROSION RESEARCH W. D, Manly Metallurgy Division W. R. Grimes F. Kertesz Materials Chemistry Division H. W. Savage ANP Division The static and tilting-furnace corrosion testing facilities were used for further studies of the corrosion of ceramic materials, Inconel, and nickel in fluoride mixtures. Tests made in the tilting- furnaoce apparatus indicate that both hydrofluori- nation and filtration are important in the control of the corrosive properties of high-purity fluoride mixtures. Silicon nitride exposed to NaF-ZrF -UF, and NaF-ZrF, for 100 hr at 800°C under static conditions showed weight gains of more than 100%, and therefore tests of this material have been discontinued. Titanium oxide dissolved almost completely under similar conditions. Tests of Inconel specimens exposed to NaF-ZrF, in beryl- lium oxide crucibles showed that less chromium was removed from the Inconel at 1000°C than at 800°C, probably because of the more rapid formation of a protective layer, Tilting-furnace tests in which graphite rod was added to NaF-ZrF4~UF4 in contact with Inconel showed that the graphite caused only a slight increase in the chromium content of the fluoride mixture; however, powdered graphite caused a more- than-threefold increase of the chromium concen- tration under similar conditions. Nickel and Inconel rods were exposed to fluoride mixtures in air to test their svitability as moterials for sensory elements in fuel level indicators for the ARE fuel recovery system. Of these two materials, inconel appears to be the more satisfactory. Corrosion festing in thermal convection loops is being accelerated by the addition of loop facilities and modifications of loop design. Additional tests have confirmed that the corrosion of !nconel in- creases with operating time when exposed to both Nabk-ZrF ,-UF, and NaF-ZrF, and that the attack is due to the mass transfer of chromium metal. The mechanism of this mass transfer is still not completely understood, but it takes place with very fow concentrations of chromivm in the fluoride mixtures and the concentrations do not change with time. Chemical analyses of NaF-ZrF -UF, after in lInconel therma! convection loops have shown that some segregation of the uranium takes place on cooling. In loop tests for which the uranium content of the fluoride mixture was varied, it was established that the depth of attack in slowly, but linearly, with content in the fluoride mixture. circulation Incone!l increases incredasing uranium The hoped for absence of corrosion at low temperatures was not found. There was still considerable attack in the upper hot-leg sections of loops operated ot 1200°F, l.oops were operated to test special, high-purity, low-carbon-content Inconel, and aimost the same depth of attack was found as in loops constructed from commercial tubing. However, these loops, which were constructed of Y%-in. tubing, showed less attack than that found in similar loops con- structed of Y%.in. pipe. The data obtained confirm the previous conclusion that increasing the surface- to-volume ratio decreases the depth of attack. Several 400 series stainless steels were fobri- cated into loops in which NaF-ZrF ,-UF , was circulated for 500 hr ot 1500°F. In each loop, a considerable quantity of metal crystals was found in the cold leg. The mechanism of the attack changes as the chromium content of these steels increases. With the low-chromivm-content alloy, type 410 stainless steel, removal of the entire surface was found, while with the high-chromium- content, type 446 stainless steel, subsurface voids and surface roughening were found. Considerable effort has been exerted to get a Hastelloy B loop to circulate NaF-ZrF -UF, for 500 hr at 1560°F. All loops failed either through catastrophic oxi- dation or mishandling. These alloys are hot short in the temperature range of the test. ' An Inconel forced-convection [loop in which NaF-ZcF,-UF, was circulated at high velocity (5 fps) and with a high temperature drop (200°F) was examined after termination because of plugging. The exomination showed that the plugging was probably caused by gradual freezing of the fluoride 65 ANP QUARTERLY PROGRESS REPORT mixture at a cold spot and not by corrosive attack. [n a further attempt to study corrosion by turbulently flowing fluorides, an oscillating furnace has been constructed in which turbulent flow is simulated by suddenly stopping and cooling a heated, rotating VY-shaped capsule. Preliminary tests indicate that the effectof flow is rather small in comparison with the increase in attack which can be attributed to the temperature effect. Static corrosion tests in fluoride mixtures have been mode of many of the brazing alloys developed by the Welding and Brazing Group, and it has been found that the nickel-phosphorus and the nickel- phosphorus-chromium brazing alloys have good corrosion resistance in NaF-ZrF ,-UF . The study of corrosion and mass transfer charac- teristics of container materials in liquid lead bythe use of quartz thermal convection loops has been extended to include tests with cobalt, beryllium, titanium, and Hastelloy B. Results alreudy obtained for o molybdenum-nicke! alloy (25% Mo-~75% Ni) were rechecked, The influence of oxides in re- tarding mass fronsfer in a system with type 347 stainless steel inserts was further investigated. FLUORIDE CORROSION IN STATIC AND TILTING-FURNACE TESTS f. Kertesz H. J. Buttram N. V. Smith R. E. Meadows J. M. Didlake Materials Chemistry Division Effect of Methods of Manufacturing the Fuel on the Corrosion of lnconel A study was made to ascertain whether cor- rosiveness would be affected by omission of either the hydrofluorination or the filtration step in the preparation of purified fluoride mixtures. Tilting- furnace tests indicated that when hydrofluorination was omitted the depth of subsurface-veid formation and the amount of chromium in the fluoride mixture after testing were higher than those chserved with the control mixture prepared by the usual process, including hydrofluorination and filtration. The results, which agree with previous findings, show that both hydrofluorination and filtration are important in the control of the corrosive properties of high-purity fluoride mixtures. Corrosion of Ceramic Maoterials The study of the resistance of nonmetallic materials to attack by fluoride mixtures has been 66 continued. Silicon nitride exposed to the mixtures Nof-ZrF ,-UF, (53.5-40.0-6.5 mole %) and NaF- ZrF4 (53.0-47,0 mole %) for 100 hr at 800°C under static conditions showed weight gains of more than 100%, These weight gains were probably due to the penetration of the fluorides into the pores of the hot-pressed material., Tests with this material have been discontinued. Titanium oxide (Americon l.ava Corporation body, Al Si Mag No. 192) was found to dissolve nearly completely when exposed to the fluoride mixtures wder the same conditions, It was known from previous tests that beryllium oxide reacts with vranium-containing fluoride mixtures and that a layer of uranium oxide is formed on the beryllium oxide. Tests were run with the mixture NoF-ZrF, (53-47 mole %) in beryllium oxide crucibles fabricated from the ARE moderator blocks. After exposure ot temperatures of 800 and 1000°C for 100 hr in the presence of Inconel specimens, the formation of a ZrO, layer could be ohserved on the crucible walls, Less chromium was found to be removed from the Inconel at 1000°C than at 800°C, probably because the reaction 2Be0 + ZrF, —— 2BeF, + Zr0, is accelerated at the higher temperature and there is more rapid formation of a protective layer, Effect of Grophite Additions The effect of graphite on the behavior of fluoride mixtures in contact with Inconel was studied by adding various amounts of graphite in either powder form or as a rod to NoF-ZrF ,-UF, {53.5- 40,0-6.5 mole %) and exposing the mixtures to the usual tilting-furnace test, The graphite rod caused only a slight increase of the chromium concentration of the fluoride mixture, while addition of large amounts of powdered graphite resulted in a more- than-threefold increase of the chromium concen- tration. Comparison tests were then made in an attempt to determine the effect of absorbed oxygen in the powdered graphite. As-received graphite and graphite which had been outgassed were used. The outgassing wos aftempted by heating the graphite capsules under vacuum to 800°C for 4 hr hefore loading and to 400°C for 4 hr after loading, The vacuum chamber of the furnace was then filled with helium and the temperature raised 1o 800°C 1o carry out the test. Chemical analyses of the fluoride mixtures after the tests did not show changes atfributable to outgassing, Microscopic studies by T. N, McVay revealed the presence of ZrQ,, crystals in the fluoride mixtures used in both tests; so it must be concluded that the attempted outgassing was not effective in removing all the oxygen. Effect on Nickel and Incone! Rods Exposed to Fluoride Mixtures in Air In order to determine a suvitable material for sensory elements to indicate fuel level in the containers fo be used in the ARE fuel-recovery operation, Inconel and nickel rods have been tested for resistance to attack by molten fluorides in air. The first of these tests was carried out with NaF-ZrF -UF, (53.5-40,0-6.5 mole %) heated to 700°C in an open nickel crucible. The Inconel and nickel rods were held 1 in. from the bottom of the crucible for 48 hours. Determinations of the electrical resistance between the crucible and each rod were made periodically. After 48 hr of contact, the rods were examined metallographically, and was found that the nickel suffered a 50% loss in cross section, while the Inconel was attacked to a lesser degree. ln the second fest, the rods were removed periodically for a 5-min cooling period. During 50 immersions, the resistance between the nickel! rod and the crucible ranged from 1.2 to 7.5 ohms, while the resistance between the Inconel rod and the crucible varied from 0.5 to 1.1 ohms. A heavy crust was observed to form on the nickel rod during the test, while only a thin crust formed on the Inconel rod. The nickel crucible suffered considerable attack during the tests. Of the two materials tested, Inconel appears to be the more desirable for use as sensory elemenfs in fluoride mixtures. THERMAL CONVECTION LCOP DESIGN AND OPERATION G. M. Adamson Metallurgy Division The design of the stondard therma! convection loops was changed to eliminate the expansion pot so that only one size of pipe or tubing would be required for a loop. Figure 6.1 shows the new configuration . without Schedule-40, Zain. IPS pipe is still being used as the sfcndard loop material, but it will be replaced by /-un. schedule-10 pipe as soon as a supply is avmiubie. insulation. PERIOD ENDING MARCH 10, 1954 Another change in the design consisted of replacing the sharp bends at both junctions of the legs with curved sections. The curved sections should reduce the flow resistance and give a slight increase in velocity, The emphasis in the loop work is shlfhng from 500-hr operation to operation for from 1000 to 2000 hours. The longer circulation times are required for studies of mass transfer. With the shift to long-time operation, it has been necessary to increase the number of loops. During this quarter, E UNGLASS)S Tr\ Fig. 6.1. Modified Thermai Convection Loop. &7 ANP QUARTERLY PROGRESS REPORY the number has been increased from 14 to 20, and instruments have heen ordered for a line with 34 loop stations. FLUORIDE CORROSION OF INCONEL IN THERMAL CONVECTION LOOPS G. M. Adamson Metallurgy Division Effect of Exposure Time on Depth of Attack The results obtained in tests for determining the effect of exposure time on the corrosion of Inconel by high purity NaF-ZrF (-UF , circulated in a thermal convection laop at 1500°F confirm the results reported previously.] Results from four loops filled from the same batch of fluorides and circu- lated for varying times are tabulated in Table 6.1. The attack continued with time but ot a rate lower than that found in the early stages. The holes grew in size and became more concentrated in the grain boundaries with increasing time. Figure 6.2 presents typical sections from loops 345 and 329. The results given in Table 6.1 show that chromium metal is mass transferred, that the chromium concentrations in the fluorides are very 'G. M. Adomson, ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609, p. 79. low, and that the chromium concentration remains fairly constant. Both the chemical results obtained from examination of loop 328, which operated for 2000 hr, conflict with the results from other tests in this series and with previous results. Check samples from this loop have been submitted for chemical and metallurgical exomination, and the operating record is being closely studied in an attempt to explain the discrepancies. The 20 mils of penetration found after the 3000-hr test with the high-purity fluorides is deeper than the penetration (18 mils} found in the first series of tests after 2850 hr of operation with impure fluorides. The two results are not exactly comparable, because samples were cut from the loop operated with high- purity fluorides froman areaabout 2 in, higher up the hot leg. However, the error resulting from the sam- pling should be less than 2 mils. In any case, the recently completed tests indicate that purifying the fluorides will not result in o reduction in mass transfer. This conclusion is further confirmed by results obtained with another loop that circulated for 3000 hr the same batch of fluorides that had been further purified by the addition of zirconium hydride. This loop showed a maximum penetration of 13.5 mils. While this loop showed a reduction in depth of attack, it showed about the same and metallurgical TABLE 6,1. EFFECT OF EXPOSURE TIME ON CORROSION OF INCONEL BY NaoF-ZrF ~-UF, MAXIMUM AVERAGE CHROMIUM LOOP | EXPOSURE | AMOUNT AND TYPE PENETRATION TRAP AND COLD-LEG CONCENTRATIONS NO, TIME (hr) OF ATTACK | APPEARANCE IN FIL UORIDES (mils) (ppm) 345 500 General, moderate to Dark layer, but neo 620 heavy, with small metal in trap; no layer voids ot cold-leg wall 327 1000 General, moderate 10 Metallic ring around 550 and intergranular wall; pessibly thin layer on cold leg 328 2000 Moderate to heavy; Metallic ring around 250 intergranular wall 329 3000 Heavy; intergranular, 20 Layer of chromium in 620 with large voids trap and layer on cold-leg wall 68 reduction as would be found by comparing loops with and without ZrH, after 500 hr of operation. PERIOD ENDING MARCH 10, 1954 The size and distribution of the holes in this loop were very similar to those found in loop 329. Metdl crystals were found in the trap, and there wos a thin, but continuous, metal deposit on the cold-leg surface. ' Effect of Exposure Time on Chemical Stability of the Fluorides ' Changes have been noted in the uranium analyses of the fluorides both after circulation in thermal convection loops and after tests in seesaw apparatus. These chemical changes have usually been ex- plained as sublimation of zirconium fluoride and as sampling or analytical errors. The chemical analyses of the fluorides circulated in the loops described above are presented in Table 6.2, The uranium percentage increased 1.2%, while the zirconium content dropped 2.9%. Although the zirconium content change is greater than that for the uranium (if the change is allotted in proportion to the amounts of all the ingredients present), the change is not large enough to explain the change in vuranium content as being due to zirconium sublimation, In addition, the change in the fluorine is not large enough to substantiate the postulated However, the zirconiumanalyses are not particularly accurate, and segregation undoubtedly occurred during cooling, although no low-uranium-content phase has yet been found. sublimation of the zirconium as ZrFA. Effect of Exposure Time on Corresion by Nonuranivm-Bearing Fluorides ' Loops filled from a single batch of NaF-ZrF, (50-50 mole %) which had been produced for the ARE were operated for 500, 2000, and 3000 hours. The data from this series of tests are tabulated in Fig. 6.2. Sections from !nconel Thermal Con- vection Loops Showing Effect of Exposure Time on Corrosion By NoF-ZiF -UF, Circulated ot 1500°F, a. Loop 345; exposure time, 500 howrs. b. Loop 329; exposure time, 3000 hours, Etched with aqua regia. 230X Table 6.3, The same conclusions can be drawn from TABLE 6.2. CHEMICAL ANALYSES OF NoF-ZrF -UF, AFTER CIRCULATION IN THERMAL CONVECTION LOOPS CHEMICAL COMPOSITION AFTER TEST LOOP CIRCULATION ' NO. TIME (hr) wt % ‘ ppm _ u - Zr F Ni Cr Fe. 345 500 9.2 37.2 41.7 <20 620 45 327 - 1000 9.6 . 36.5 41.9 <20 550 95 329 3000 10.0 35.2 41.5 <20 620 110 Original composition 8.8 38.1 41.9 40 60 70 69 ANP QUARTERLY PROGRESS REPORY TABLE 6.3, EFFECT OF EXPOSURE TIME ON CORROSION OF INCONEL BY NaF-ZeF MAXIMUM AVERAGE CHROMIUM LOOP | EXPOSURE | AMOUNT AND TYPE PENETRATION TRAP AND COLD-LEG CONCENTRATION NO. TIME (hr) OF ATTACK (mils) APPEARANCE IN FLUORIDES (ppm) 348 500 Light, intergraonulor 5.5 Few scottered patches 130 of deposit on cold-lag wall; dark layer in trap 346 2000 Moderate to heavy, 9 Thin layer en cold-leg 300 intergranular wall; dark layer and metal in trap 347 3000 Heavy, intergranular, 11 Continuous metallic 115 with large voids deposit on wall; chromium metal in trap these data as from those obtained with Naf-ZrF - UF,. There is rapid initial attack followed by continuing attack at a reduced rate. The depth of attack, however, is less than that obtained with the vronium-beoring fluorides. It is apparent, even with nonuraonium-bearing mixtures, that chromium metal tronsferred and that the chromium concentrations in the fluorides are very low. may be mass Effect of Yariation in Uranivm Concentration Two series of tests were conducted for determining the effect on corrosion of changes in the uranium concentration of the fluoride mixture. In the first series, the fluoride mixture NaF-ZrF ,, with uranium additions that varied from 0.5 to 15 wit % (0.2 to 7.2 mole % UF,), was circulated, In the second series of tests, the high-uranium-content fuel (NaF-ZrF4-UF4, 53.5-40,0-6.5 mole %) was circu- lated in three loops. It has now been determined that the fluorides used in both series of tests received inadequate and varying amounts of gas purging during production. The variations in gas purging lead to varying hydrogen fluoride content and make the results difficult to interpret. While the results of these tests cannot be com- sidered as conclusive, the trends can be partially confirmed by other results, The two leops operated with low-uranium-content fluorides, 0.5 and 2.8 wt%, showed much less attack than was expected, and 70 thus the data are not consistent with the data obtained from other loops. The two batches of fluorides used for these tests were probably of higher purity than the batches used for the other tests, In spite of the low attack, a very thin, apparently metallic deposit was found on the cold-leg wall of each loop. In the other three {oops, in which the uranium contents of the fluoride mixtures were 5.1, 9.8, 14.4 wt %, respectively, the attack increased linearly with increasing uranium content. A very thin layer was found in the loop which circulated the mixture, but no layers were found in the other two loops. While depth of attack increases with in- creasing uranium content, the probability of layer formation in the cold-leg decreases. These results arz confirmed if the results obtained with NaF-ZrF, (50-50 mole %), NaF-ZrF,-UF, (50-46-4 mole %), and the lowest penetration with NaF-ZrF -UF, (53.5-40.0-6.5 mole %) are con- sidered. The atiack in each case showed a small, but linear, increase from 5 to 7 mils with increasing uranium content, Two other Inconel loops were operated for 500 hr with NaF-ZrF ,-UF, (53.5-40.0-6.5 mole %, 14 wt % U). These loops showed maximum attacks of 16.5 and 10 mils. Since the mixtures circulated were impure, the attack should be compared with the 9 mils found with impure NOF-ZrF4-UF4 (50-46-4 mole %). |owest-uranium-content In each of the four loops in which NuF-Zer-UF“ (53.5-40.0-6.5 mole %) was circulated, large vari- ations in uranium content were found, as shown in Table 6.4. In each loop, the uronium content varied about 3%, with the lowest values being found in the hot leg. Similar variations have not been found with the other fluoride mixtures., Ad- ditional samples from the lower portion of the hot leg have now been submitted for analysis. The segregation of the uranium undoubtedly takes place during cooling, but it is not certaln at what temper- ature the segregation starts. Effect of Temperature In a loop operated at 1250°F, only the upper portion of the hot leg was attacked, and the attack was to a depth of 3 mils. It was hoped that at slightly lower temperotures all attack would be eliminated. Two loops were therefore operated with hot-leg temperatures of 1200°F. Loop 390 showed a maximum hot-leg attack of 5 mils, and in loop 350, the maximum attack in the hot leg was 4 mils. The samples were cut from an area 2 in. further up the hot legs of these loops, and therefore the cor- TABLE 6,4. URANIUM CONTENT OF NaF-ZrF -UF, (53,5.40,0.6.5 mole %) AT VARIOUS POINTS IN INCONEL THERMAL CONVECTION LOOPS AFTER CIRCULATION AT 1500°F FOR 500 HOURS URANIUM CONTENT (wt %) SAMPLE TAKEN Loop Number 320 326 381 387 Production batch 16.5 13.8 15.6 l.oop sample (before 17.2 14.4 14.6 15.3 circulation) Hot leg Position 1 12,9 13.6 10.8 Position 2 12.9 | 10.0 10.4 13.5 Position 3 . 8.8 16.7 L.ower horizontal leg 16.6 14.8 13.0 16.0 Cold leg Position 1 16.8 15.4 15.2 Position 2 17.3 14.0 13.0 14.8 Upper horizontol leg 17.6 13.8 PERIOD ENDING MARCH 10, 1954 rosion values are not comparable to those found in the loop operated at 1250°F. In both loops, the attack was found only in the upper portion of the hot leg. The voids were very small and evenly distributed. Figure 6.3 shows a sample from the hot leg of loop 350. A hot-leg temperature of 1200°F is the lowest temperature at which fluorldes may be circulated in present loops. Effect of Yariotions in Loop Size and Composition Two loops constructed from 1/Q-in. tubing instead of the customary lf’z-in.-iF’S schedule-40 pipe were operated with NoF-Zer-UFd (50-46-4 mole %). These loops were operated both os standard loops for the high-purity fluoride vs. Inconel tests and as a part of a surface-to-volume ratio study. A maxi- mum attack of 9 mils was reported previously? for a loop constructed from l-in., tubing with a surface-to-volume ratio of 4.5 in.%/in.?, and an attack of 5.5 mils was reported for a standard loop with a surface-to-volume rcnio of 6.5 in%/in.2. The ratio 'in a loop with /—m. ‘rubmg is 10.5 in.2/in.3. The attack found m the /-m. fublng of loop 359, which had been filled from the same batch of flyorides os that used for the other two loops, was moderate fto heavy, with a maximum depth of 4 mils. Loop 362, also constructed of I/2«in. tubing, showed similar attack to a depth of 3.5 mils, with a few areas extending to 5 mils. Plots of attack vs. the three surfoce-to-volume 26, M. Adamson, ANP Quar. Prog. Rep. Dec. 10, 7953 ORNL-1649, p. 72. Fig. 6.3. Section from Hot Leg of Inconel Loop 350 in Which MaF-ZsF -UF {50.46-4 mole %) Was Circulated ot 1200°F for 500 Hours. Etched Witk aqua regia, 250X 71 ANP QUARTERLY PROGRESS REPORT ratios do not quite yield straight-line relationships, With a thermal convection loop, both the velocity and the temperature drop change with any change in pipe size, and these variotions probably couse the deviations from straight-line relationships. A series billets of high-purity, low-carbon heats of Ilnconel made by the Metallurgy Division were shipped to Superior Tubing Company and drawn into ]/2-in. tubing. This tubing was then fabricated into thermal convection loops. Standard lnconel con. tains 0.08% carbon, whereas these heats contained about 0.014% carbon. To further tie up the carbon, 0.31% titanium was added to one heat. Two loops fabricated from the low-carbon-content Incone! tubing, one with titanium added and one without titanium, were filled from the same batch of fluorides as that used with the ]/1,-in. commercial tubing discussed above. The loop (366) without titanium showed moderate, general, hot-leg attack, with @ maximum penetration of 3.5 mils. The loop with added titanium, loop 367, alsc showed moderate to heavy, general, hot-leg attack to a depth of 3.5 mils, Typical hot-leg sections from these three loops are shown in Fig. 6.4, Within the limits studied, very little, if any, reduction in attack was found. There may have been very slight reduction in depth of attack, but many more loops would have to be tested to confirm this. The Inconel was cleaner than in previous loops and did not contain as many inclusions. One other loop fabricated from low-carbon-content Incone! tubing was filled from a different batch of fluorides, which happened to be one of the several impure batches recently received. After circu- lation of the fluoride mixture, very large grains were found in the Inconel. The hot-leg attack in this loop was heavy and to a depth of 13 mils. This loop is still being studied. Effect of Additiens to the Fluorides Batches of the fluoride mixture NaF-Zrb (50-50 mole %) were held overnight at 1300°F in contact with various materials and then circulated in Inconel thermal convection loops for various times at 1500°F, The pretreatments used and the test results are given in Table 6.5, The standard loop was filled from the same batch of fluerides as that used for the treated loops. These tests do not show the large reduction in depth of attack found with previous zirconium hydride additions. The fluoride batch was quite 72 Fig., 6.4, Loops 359, 386, and 367 in Which MoF.ZrF UF (50-46-4 male %) Wos Circvlated at 1500°F for 500 Sections from Hot Legs of lacone! Hows. a. Loop 359, commercia! tubing, 5. Loop 366, high-purity, low-carbon-content tubing. ¢.lLoop 347, high-purity, low-carbon-content tubing with added titanium, Etched with aqua regio. 250X PERIOD ENDING MARCH 10, 1954 TABLE 6.5, CORROSION BY PRETREATED FLUORIDE MIXTURES CIRCULATED IN INCONEL THERMAL CONVECTION LOOPS FOR 500 HOURS LOOP PRETREATMENT HF CONTENT IN: EXPOSURE MAXIMUM NO ' MATERIAL FLUSHING GAS TIME IN HOT-LEG ATTACK PENETRATION ’ (relative units) LOOP (hr) (mil s} 370 Zr metal* 7 500 .Moderate to heavy, 4 intergranular 37 Zr metal™ 17 500 Light to moderate, 4 intergranular 374 0.3% ZH, 4 500 L.ight to moderate, 6 intergranular 373 0.3% ZrH2 2.4 1000 Moderate to heavy, 8 intergranular 399 None 2 1000 Heavy, intergranular 10 *Black deposit found in treatment pot, impure, but the purity should not have been a factor. Samples taken before and after circulation of these fluorides have been submitted for chemical analysis, Corrosion of VYarious Metal Combinations A series of Inconel loops has been placed . in operation with various combinations of type 316 stoinless steel or nickel inserts in the hot and cold legs. This series of loops is being operated for studying the effects of the presence of various quantities of these metals on the corrosion of Inconel. The results from the operation of the first such loop, reported previously,3 showed that a short length of type 316 stainless steel in the top of the hot leg would reduce the attack on the Inconel but would cause increased mass transfer. Loop 379, the second in this series, was operated with a length of nickel in the hot leg. The Inconel showed heavy subsurface void attack to a depth of '5 mils, and no change in attack in the Inconel was apparent near the nickel insert. The nickel near the weld had a few scattered voids to a depth of 1 mil, while the rest of the insert appeared to be unattacked. [f any even removal had taken place, it could not be determined. No deposit was found in the cold leg. If any reduction in attack was 36. M Adamson, ANP Quar, Prog. Rep, Sept. 10,7953, ORNL-1609, p. 77. caused by ‘the nickel, it was too slight to be con- clusive on the basis of a single oop. FLUORIDE CORROSION OF STAINLESS STEEL, IZETT IRON, AND HASTELLOY LOOPS The results from the first 400 series stainless steel loop- in which NaF-ZrF -UF (50-46-4 mole %) was circulated were presented in the previous report,* The loop was constructed from type 430 stainless steel (0.12% carbon, maximum; 14 to 18% chromium). After the test, the hot-leg surface of the loop was smooth, and there was no visible attack. Metallic crystals were found in the cold leg, and thus there was indication that some even removal had taken place. Two mere 400 series stainless steel loops have now circulated NaF-ZrF -UF, (50-46-4 mole %) satisfactorily for 500 hours, Loop 138 was constructed from type 410 stainless steel (0.15% carbon, maximum; 11.5 to 13.5% chromium). Unlike the other 400 series stainless steels studied, type 410 stainless steel is marten- sitic. Affer operation for 500 hr at 1500°F, the hot leg of loop 138 was very rough, with sharp depressions to a depth of 1.5 mils. There was no evidence of intergranular or subsurface-void types of attack. permit accurate wall The outer surface was too rough to thickness measurements; AG. M. Adamson, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 73. 73 ANP QUARTERLY PROGRESS REPORT however, it seems likely that some thinning of the wall had taken place. The cold-leg wall was covered with a metallic layer, and the fluoride mixture was full of dendritic metallic crystals. The second loop was constructed from type 446 stainless steel (0.35% carbon, maximum; 23 to 27% chromium). After operation for 500 hr, the hot leg of this loop was rough and uneven, but, in addition, a heavy concentration of subsurface voids to a maximum depth of 10 mils was found, The cold leg had a very heavy deposit of metal, and there were dendritic crystals in the fluoride mixture. With both this loop and loop 138, the cold-leg crystals were shown to contain primarily iron, but chromium was also present. It is obvious that a in the chromium content of the alloys causes a change in the corrosion mechanism. With type 430 stainless steel, even removal occurs, while with type 446 stainless steel, both even removal and a subsurface-void type of attack, similar to that found in type 316 stainless steel, are found. change A single loop of lzett iron was filled with Naf- ZrF4-UF4 (50-46-4 mole %) and circulated at 1500°F, This loop was terminated after 39 hr of operaticn because of plug formation, In this short period of operation, the lower sections of the loop had collected many balls of needle-like dendritic iron crystals. These crystals are similar in appearance tothe nickel crystals found in hydroxide systems. Also, a thick deposit was found on the cold-leg wall. The hot-leg surface of this loop was very rough, with areas in which entire grains had been removed. One loop was constructed from 1/2-in. Hastelloy C tubing (58% Ni—-17% Mo—-15% Cr--5% W-5% Fe). This loop was terminated because of a leak in the weld at the top of the hot leg after having circulated NaF-ZrF4-UF4 (50-46-4 mole %) for 456 hr at 1500°F. The hot leg of this loop showed light, subsurface-void formation to a depth of 1.5 mils. The attack was primarily in the second phase found in these alloys. No attack or deposit was found in the cold leg. Attempts were made to circulate NaF-ZrF -UF, (50-46-4 mole %) in several loops constructed from Hastelloy B (62% Ni-30% Mo--5% Fe). The tubes for these loops varied from 0.035-in.-wall, ]/Q-in. tubing to 0.065-in.-wall, 1-in, tubing, but all loops failed. Two failures were due to poor handling, one to a poor weld, and the others to catastrophic 74 This oxidation occurred under the Saurizen cement used to protect the thermocouples. The Haostelloys are hot short in the temperature range 1100 to 1600°F, and thus they are difficult to handle, For the loops now being fabricated, the tubing will be given a high-temperature anneal to minimize these difficulties. One Hastelloy B loop constructed of 1-in, tubing was examined after it had operated ot 1500°F for 91 hr before failing because of catastrophic oxi- dation. The hot- and cold-leg surfaces of this loop were rough. Fabrication cracks were visible in the cold leg but not in the hot leg. Since the surface of the hot leg had a matte finish, it is likely that some general removal had taken place; no inter- oxidation. granular or subsurface-void types of attack were visible. CORROSION 8Y HIGH.VELOCITY HIGH- TEMPERATURE FLUORIDES Forced-Circulation Corrosion Loop G. M. Adamson Metallurgy Division An Incone! forced-circulation loop operated at a maximum temperature of [500°F for 200 hr at a maximum fluid velocity of about 2.5 fps wos de- scribed previously.®> After termination, this loop was subjected to metallurgical examination. The normal, subsurface-void type of attack was found in the hot leg. The attack gradually increased in depth to a maximum of 11 mils in the hottest portion and was both more intense ond more con- centrated in the grain boundaries than that found in the thermal convection loops operated under similar conditions. The depth of attack was only slightly deeper, but, since turbulent flow was not obtained, no large increase would be expected. Considerable carburization of the Incone! was found in the hot leg. The carbides must have coriginated from oil used to lubricate the gas seal. In the hottest portion, a general layer of carbide precipitate was found to a depth of 14 mils, and there was an intergranular layer through the sample. No deposits were found in any of the cold-leg sections examined. Some attack and carburization was found in the first or hottest cold-leg section but not in the other sections. Diffraction and 5D, F. Salmon, AMNP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 29. petrographic examinction of the fluorides, before and after circulation, failed to reveal any changes that might have caused plugging. It seems likely that the plugging of this loop was caused by gradual freezing of the fluorides at a cold spot and not by corrosion. Oscillating-Furnoce Studies F. Kertesz H. J. Buttram J. M. Didlake R. E. Meadows N. V. Smith Materials Chemistry Division Corrosion by turbulently flowing fluorides is being investigated with an osciilating-furnace apparatus. Turbulent flow is simulated by suddenly stopping a rotating V-shaped capsule which has been heated to 800°C by a gas furnace. One of the legs of the rotating capsule is shorter than the other, and, at the instant the liquid is atf the apex (bottom) of the V with its level being below the top of the short leg, the capsule is abruptly stopped; this causes the liquid to rush into the empty part of the leg. When the capsule is stopped, the gas is turned off and the capsule cools. After cooling to 600°C, the capsule is rotated back to its original position, with the apex pointing downward, while the heating cycle is resumed. The complete cycle takes «about 90 seconds. The absence of steady-flow conditions makes it difficult to estimate the actual Reynolds number of the system at the time of stoppage, but in view of the rapid deceleration, it is expected that turbulent conditions should prevail for a short period in the capsule, ' | In order to evaluate the relative importance of flow conditions and of temperature, static tests were run at the same time at 600 and at 800°C, since the welding required in the fabrication of the V-shaped capsules could have changed the properties of the tube walls and thus changed their resistance to corrosion. In another series of check tests, stationary capsules were thermally cycled by alternately heating and cooling from 650 to 800°C. The effect of motion without a temperature gradient was studied by slowly rotating nearly horizontally placed capsules inside an isothermal furnace with the ends kept at 650 and 800°C. Tilting-furnace tests were also run in which the hot- and cold-end capsule wall temperatures were kept at 650 and 800°C, respectively. In addition to the usual 4-cpm - PERIOD ENDING MARCH 10, 1954 tilting-furnace tests, a series of 2-cpm tests was made; the duration of all these tests was 100 hours. The chromium concentrations of the melts after the tests are summarized in Table 6.6. The results given in Table 6.6 indicate that, within the range of the experimental conditions used, the effect of temperature on the amount of chromium dissolved by the fluoride melt is very noticeable. On the other hand, the effect of flow (ranging from static flow to the viscous flow obtained in the horizontally rotated capsules, to the tilting-furnace test, and finally to the pre- sumably turbulent conditions in the oscillating furnace) is rather small in comparison with the increase in attack which can be attributed to the temperature effect. STATIC TESTS OF BRAZING ALLOYS IN FLUORIDES AND SODIUM E. E. Hoffman J, E. Pope L. R. Trotter Metallurgy Division Brazed T-joints have been tested in static NaF- ZrF4-UF4 (50-46-4 mole %) for 100 hr at 816°C, The brazing alloys used were the eutectics 90% Ni-10% P and 80% Ni~10% P-10% Cr, and the base materials were type 316 stainless steel and The details of preparation of the T-joints and properties of the as-brazed joints are given in Sec. 7 of this report, ‘‘Metallurgy and Ceramics.” The results of the static tests are summarized in Table 6.7. As can be seen from the weight-change data, attacked. The greater part of the weight loss may be atiributed, in each case, to attack on the bose material. The excellent resistance of the 90% Ni~10% P brazed joint to attack may be seen in Fig. 6.5. A layer (0.25 to 1 mil in thickness) of nickel-rich solid solution formed on the surface of all the joints tested (Fig. 6.6), except the Inconel joint brazed with the 80% Ni-10% P-10% Cr alloy. These brazing alloys seem to be very satisfactory with respect to corrosion, but the brittleness of the grey Ni,P phase is well illustrated in Fig. 6.7, which shows the brazed specimen before the corrosion test. Inconel and type 316 stainless steel joints brazed with 75% Ni-25% Ge alloy have been tested in NaF-ZrF ,-UF, (50-46-4 mole %) and in sodium. This brazing alloy was corroded to an oppreciable Inconel. none of the specimens were appreciably 75 ANP QUARTERLY PROGRESS REPORT TABLE 6.6, CORROSION OF {NCONEL BY MOLTEN FLUORIDES DURING STATIC AND DYNAMIC TESTS TYPE OF TEST CHROMIUM CONCENTRATION meq Cr/kg meq Cr/kg NaF-ZrF -UF ;* NoF-ZrF ** Static at 600°C 1.6 9.5 Static at 800°C 56.4 25.3 Thermal cycling 58.5 45.6 Horizental rotation, 600°C 18.8 16.4 Horizontal rotation, 800°C 55.4 18.8 Static at 800°C, V-shapad capsules 63.3 40,1 Oscillating furnace, helium atmosphere, 600 to 800°C 75.0 44.1 Tilting furnace, 2 cpm, 650 to 800°C 66.4 28.8 Tilting furnace, 4 cpm, 650 to 800°C 55.9 21.3 Oscillating furnace, vacuum, 600 to 800°C 51.9 49.6 ¥53.5.40.0-6.5 mole %. * % 53-47 mole %, TABLE 6,7, RESULTS OF STATIC TESTING OF BRAZING ALLOYS IN NaF-ZrF -UF (50-46-4 mole %) AT 1500°F FOR 100 HOURS WEIGHT BRAZING ALLOY BASE MATERIAL CHANGE™* METALLOGRAPHIC NOTES (%) 90% Ni-10% P Inconel 0.06 No attock on the braxed joint; 0.25-mil nickel-rich phase on surface of braoze fillet after test, no cracks in joint of either the os-brazed joint or the corrosion-tested joint 80% Ni-10% P-10% Cr lnconel 0.07 Subsurface void formation in braze fillet to a depth of 2 to 3 mils; voids resulted from attack on ths globular particles of the nickel-rich phase 90% Ni-10% P Type 316 0.06 Braze fillet unattacked; 0.5~ to 1-mil layer of stainless steel nickelerich phase on surface of braze fillet; large cracks evident both in the as-brazed and corrosion-tested joints 80% Ni~10% P-10% Cr Type 316 0.18 stainlass steel Scattered subsurface voids in Ni3P phase beneath the 0.5-mil nickel-rich solid sclution on the sur- face; surface loyer unaitacked; no cracks in either the as-brazed or the corrosion-tested joint * Per cent weight change includes weight of brazing alloy (approximately 5% of total) and weight of base material., 76 PERIOD ENDING MARCH 10, 1954 | UNCLASSIFIED CYo10460 t50 X Fig, 6.5. Inconel T-Joint Brozed with 90% Ni-10% P Alloy ofter Exposure for 100 hr at 1500°F in MaF- ZLeF UF, (50-46.4 mole %). Specimen nickel plated after testing to preserve the edge during polishing. Etched with aqua regia. 150X degree in both media. Results of examination of these joints after testing are summarized in Table 6.8. ' MASS’TRANSFER IN LIQUID LEAD C. R. Boston J. E. Fope W. H. Bridges G. P. Smith J. V. Cathcart M. E. Steidlitz E. E. Hoffman L. R. Trotter Metallurgy Division The study of corrosion and mass fransfer in liquid lead has been extended to include tests with cobalt, beryllivm, titanium, and Hastelloy B. As in previous experiments, these metals were in- serted in small, quartz, thermal convection loops® and tested in circulating liquid lead. The results already obtained for the 25% Mo~75% Ni alloy were bW. H. Bridges et al., ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 74, rechecked, and the influence of oxides in retarding mass tronsfer in loops with inserts of type 347 stainless steel systems was further investigated, The loop with beryllium inserts was operated for 456 hr before circulation stopped becouse of plug formation. The hot- and cold-leg temperatures were 815 and 600°C, respectively., As shown in Fig. 6.8, the corrosion of the beryllium specimen was relatively uniform, with very little intergranular attack. The results for the beryllium loop must still be regarded as tentative, however, since the quartz tubing of the loop appeared to have suffered attack. It is possible that part of the mass transfer observed may be attributed to inferaction between the beryllium and the quartz. This problem is being investigated further. Plugging occurred in the loop with fitanium inserts after only 5 hr of operation. Hot- and cold- leg temperatures were 815 and 575°C, respectively. ln addition to the plug formed in the cold leg, « 77 ANP QUARTERLY PROGRESS REPORT l L 0.003 0.002 0.001 INCH 1500 X Fig. 6.6. lncone! T-Joint Brazed with 90% Ni-10% P Alloy after Exposure for 100 hr at 1500°F in NaF- LiF ~UF, (50-46-4 mole %). Note uniform layer of nickel-rich solution which formed on surface of braze fillet during test, Etched with aqua regia. 1500X TABLE 6.8, RESULTS OF STATIC TESTS OF 75% Ni~25% Ge BRAZING ALLOY IN NoF-ZrF -UF, (50-46-4 mole %) AND IN SODIUM AT 1500°F FOR 100 HOURS BASE MATERIAL BATH METALLOGRAFPHIC NOTES Inconel NnF-ZrF4-UF4 Braze fillet attacked to a depth of 7 mils in several areas; attack (50-46-4 mole %} confined to the nickelerich phase formed on exposed surface of the fillet during test Inconel Sodium Very similar to above specimen, with voids to a maximum depth of 4 mils Type 316 Sodium Specimen attacked to a maximum depth of 9 mils; attack confined to stainless steel nickel-rich phase 78 PERIOD ENDING MARCH 10, 1954 1500 X P 0.004 | 0.002 INCH Fig. 6.7. Type 316 Stainless Steel T-Joint Brazed with 90% Ni~10% P Alloy Prior to Testing, All cracks are in brittle Ni,P phase. Eiched with aqua regia, 1500X | 0.002 500X | 0.006 INCH Fig. 6.8, Transverse Section of Beryllium Specimen Exposed to Liguid Lead at 815°C for 456 hr in the Hot Leg of a Quartz Thermal Convection Loop. Folarized light. 500X 79 ANP QUARTERLY PROGRESS REPORT large guontity of mass-transferred material wos de- posited on the hot-leg specimen. Figure 6.9 shows some of the dendritic crystals found on the walls of the hot leg. A transverse section of the hot leg is shown in Fig. 6.10, There is a heavy layer of deposited material at the inside surface of the specimen. Some deposition was also noted on the cold [eg but not to the extent found on the hot leg. Two loops with cobalt inserts have been tested. The first loop plugged after 61 hr of operation with hot- and cold-leg temperatures of 820 and 550°C. Figure 6.11 shows a transverse section of the hot- leg specimen from this loop. A heater failure caused the termination of the second loop after 97 hr of operation with hot- and cold-leg temperatures of 815 and 500°C. The groduol decregse in the cold-leg temperature of this loop indicated that plugging would probably have occurred in another few hours of operotion. Examination of the test specimen from the second cobalt loop has not been completed. ,«'1\ RN 1 . ‘ ii O * < Fig. 6.9. Dendritic Growth Forms Deposited on the Titenium Specimen Expesed to Liquid Lead at B15°C for 5 b in the Hot Leg of a Quartz Therma! Convection Laoop, 30X | 0.002 500X 0.004 0.00% INCH | 0.008 AN Fig. 6.10. Transverse Section of Titanium Specimen Exposed to Liquid Lead ot 815°C for 5 hr in the Hot Leg of a Quartz Thermal Convection Loop., 500X 80 Fig. 6.11. Transverse Section of Cobalt Specimen Hot Leg of a Quartz Thermal Convection Loop. 500X The loop with Hastelloy B (64% Ni-5% Fe-28% Mo-1% Si-1% Mn) inserts was terminated ?on schedule after 504 hr of operation with hot- and cold-leg temperatures of 800 and 600°C. Although some mass-transferred material was found in the cold leg, there was not enough to cause plugging. As is shown in Fig. 6.12, the hot-leg specimen suffered very severe corrosive attack. Lead penetrated almost all the way through the specimen. The cold-leg specimen, on the other hand, was virtually unottacked, with only a thin layer -of mass~transferred material deposited on the inside of the specimen. : The results obtained from a second loop with 25% Mo-75% Ni alloy inserts confirmed the resulis obtained in the first test.” The loop was terminated on schedule after 331 hr of operation with hot- and cold-leg temperatures of 820 and 590°C. Severe intergranular penetration occurred in the hot-leg specimen but little corrosive attack was observed in the cold leg. “W. H. Bridgés et al., ANP GQuar. Prog. Rep., Dec. 10, 1953, ORNL-1649, p. 75. - PERIOD ENDING MARCH 10, 1954 0.002 500X i 0.004 INCH Exposed to Liquid Lead ot 820°C for 61 hr in the Figure 6.13 shows o transverse section from the hot leg of a loop with type 304 stainless steel inserts in which the test specimens were oxidized prior to Ic:vc:ding.7 The loop was operated for 550 hr without the formgtion of a plug large enough to stop circulation. In contrast, loops containing type 304 stainless steel specimens in the as- received condition had to be terminated within 100 hr because of plug formation. ‘ In order to test further the effect of the presence of oxides on mass transfer, a loop with type 347 stainless steel inserts was operated in which the lead was not deoxidized prior to loading, The loop operated for 240 hr before failing because of plug formation; the hot- and cold-leg temperatures were 815 and 500°C. In previous loops with type 347 stainless steel inserts for which the standard deoxidized procedure was followed, plugging occurred in approximately 150 hours. As indicated in Fig. 6.14, the hot-leg specimen showed inter- granular penetration similar to that obtained in other tests with type 347 stainless steel inserts. 81 ANP QUARTERLY PROGCRESS REPORY | UNGLASSIFIED Y1089 Pb-HASTELLOY B INTERFACE Pb—HASTELLOY B INTERFACE = 0.01 0.02 5.03 0.04 {00 X INCH Fig. 6.12, Transverse Section of Hastelloy B Specimen Exposed to Liquid Lead at 800°C for 504 hr in the Hot Leg of a Quartz Therms! Convection Leoop. 100X Pb-TYPE 304 STAINLESS d STEEL INTERFACE ~ . e N ) i _%f 1.” . \?_}YWL'; o TR R :',;flégr' o -0 . TYPE 304 STAINLESS STEEL - - - GTRUED el e T I V K * § ot fif";_._# it Hn ‘ - : e e A: ) :s«:-: - J”‘ ] y’?- ' =, " t' o . '1' , “‘%-; v’ V}j% et A RS T * b * ‘é 0.004 0.006 500X INCH Fig, 6.13. Trunsverse Section of Type 304 Stainless Steel Specimen Exposed to Liquid Lead ot 815°C for 550 hr in the Hot Leg of a Quartz Thermal Convection Loop. Etched with aqua regia, 500X 82 PERIOD ENDING MARCH 10, 1954 | G.002. 500X | ~3.004 l 0.006 Fig. 6.14. Transverse Section of Type 347 Stainless Steel Specimen Exposed to Liquid Lead at 815°C for 240 hr in the Hot Leg of a Quartz Thermal Convection Loop, 500X 83 ANP QUARTERLY PROGRESS REPORT 7. METALLURGY AND CERAMICS W. D. Manly J. M. Warde Metallurgy Division The studies of the effect of environment on the creep and stress-rupture properties of Incone! have continued, and rupture curves are presented for Inconel annealed at 1650°F and at 2050°F and tested in argon and in the fluoride mixture NaF- ZrFé-UF4 at 1300, 1500, and 1650°F. A bar graph is presented to show the time required to reach several different extensions and the amount of de- formation at rupture for Inconel tested in the various environments after being subjected to annealing procedures. The major effect on creep properties of a small change in composition of Inconel is also shown., Tests of the effect on columbium of the purity of argon and helium atmospheres were made preparatory to conducting creep tests of columbium, Methods were developed for preparing a nickel- phosphorus brazing alloy powder that can be applied in the conventional manner. An alloy with satis- factory flowability was prepared, and the optimum conditions forproducing it were established. Alloys in the nickel-germanium system are being investi- gated to determine their suitability for use with liquid metals and fused salts. The 75% Ni~25% Ge olloy used as a master alloy has a melting point of 1151°C and has exhibited favorable flowability at 1180°C. Investigations made in the study of high-conduc- tivity metals for radiator fins have shown claddings of types 310 ond 446 stainless steel on copper fo be satisfactory with respect to diffusion. However, it may be necessary to use Inconel as the cladding material, ond various materiols have been tested as ditfusion barriers between lnconel and copper. Nene of the refractory-type barrier materials tested, nomely, tungsten, molybdenum, zirconium, titanium, and vanadium, were satisfactory. Other tests showed diffusion barriers of iron and types 310 and 446 stainless steel to be successful in preventing copper diffusion for up to 500 hr at 1500°F. Com- mercially produced types 310 and 430 stainless steel-clad copper did not show diffusion of copper in 500 hr at 1500°F. Radiator fins of various materials were brazed to Inconel tubing for heat transfer tests. The fin materials being studied are type 304 stainless steel, nickel, types 310, 430, and 446 stainless 84 steel-clad copper, Inconel-clad copper, a copper- aluminum alloy, and a silver-magnesium-nickel alloy. An Incone! spiral-fin heat exchange unit was fabricated for a fluoride-to-air radiator to be used in radiation damage experiments, results indicate that conventiondl brazing alloys and alloys such as nickel-phosphorus and the precious-metal alloys will satisfactorily flow and wet both Inconel and type 316 stoinless stee! Preliminary in an atmosphere of welding-grade tank helium., The substitution of helium for dry hydrogen will be quite useful for brazing operations in which ceramic or nonmetallic jig materials must be used. Examinations of tubular fuel elements plug-drawn of ORNL indicated thot the core siructures of the elements were greatly improved by decreasing the reduction per pass. The use of nickel as a core material for these elements has been found to be unsatisfactory because of diffusion of the nickel into the stainless stee! cladding. Various composites of beryllium oxide with miner additions of other materials have been tested and found to be unsatisfactory as containers for flue- ride fuels, Special bearing shapes fabricated of beryllium oxide and of high-density graphite were prepared for testing. Six boron carbide—iron cermets were prepared for use os shielding in connection with in-pile loop studies of ANP f{uels, MECHANICAL PROPERTIES OF METALS Stress-Rupture Tests of Incone!l R. B, Oliver D. A. Douglas J. M. Woeds Metallurgy Division Stress-rupture tests of sheet Inconel specimens have been conducted in the fluoride mixture NaF- ZrF ,-UF , (50-46-4 mole %) and in argon at tempero- tures of 1300, 1500, ond 1650°F, Some of the specimens were annealed at 1650°F before they were tested and the others were annealed ot 2050°F, Figure 7.1 shows the rupture-time vs. stress plot for each condition and environment. The rupture curve for the tube-burst tes? in fuel at 1500°F is also shown. The tube-burst curve falls considerably below the curves for the tensile-siressed specimens PERIOD ENDING MARCH 10, 1954 UNCLASSIFIED ORNL -LR-DWG 203 1 Vi i/ 1000 2000 5000 10,000 RUPTURE TIME (hr} 20,000 f———-emespreees e e e | f \ SOLID LINE~-ANNEALED AT 1650°F BROKEN LINE-ANNEALED AT 2050 10,000 - - e ‘ g -~ N o BOOO p—pmme e e e b L L B L N M T o S O L. AN 2 L Lo ,)\\\ ,,,,,, b e e \ ...... \\,\ e ~ ....................... e 2000 ——— et L | | 1000 Lo oL e 10 20 50 100 200 Fig. 7.1. NaoF-ZrF ,-UF, and in Argon. tested at the some temperature (1500°F) and com- pares roughly with the curves for the specimens tested at 1650°F. |t is thought that the tube-burst type of test in which o multiaxial system can be studied is o more realistic test for obtaining in- formation pertinent to the present reactor designs. The rather large differences in rupture properties shown in the data indicate the need for expansion of the tube-burst testing program so that the multi- axial stress system can be more thoroughly studied. In previous reports, the sensitivity of Inconel to both its test environment and its onnealing tem- perature was shown by comparing the times-to- rupture for the various test conditions. However, it should be emphasized that the selection of the best annealing temperature or environment by using rupture time as the criterion would not necessarily provide the best conditions if some other reference point, such as 1% total strain, were the limiting factor, Figure 7.2 shows the time required to reach several selected extensions and rupture. The bars Rupture-Time vs. Stress Curves for Sheet Inconel Specimens Tested in Fiuoride Fuel are arranged from left to right in order of increasing time required to reach 1% elongation. It can be seen that many of the bars are out of order if extension is neglected ond rupture time is used for a comparative basis. The total elongation at rupture is about the same regardless of the anneal- ing temperature or the environment, except in air and in sodium. In sedium; it has been found that the specimen is markedly decarburized during testing, and decarburization would account for the increosed ductility. No explanation has yet been found for the greater elongations reached in air. Examination of Fig. 7.2 shows that the most obvious cases of misfit, with reference to rupture times, are the coarse-grained specimens which were fested in the molten fluoride ond in argon. The explanation for this discrepancy can be found in a comparison of the creep curves for the coarse- ond fine-groined specimens at 3500 psi. Two such curves for tests conducted in argon are shown in Fig. 7.3. Similar curves for tests in fluoride 85 98 UNCLASSIFIED ORNL—LR—-DWG. 204 TIME (hr} 10? o ; o o o e ] i - _ - R I o S _ 5 - - e o ) aso 55% | RUPTURE, o , | 1% ELONGATION 0% 10 — 0% — 50/0 — 5 ‘O . ® i - ™ ~ I . o o \O !l T 0.5 semm e : e T T e e i . o o \\)\ £ < oI~ S 32 ‘* 7 m o 0.4 ] > = £ o SLAB, STEADY STATE @ 2 ® SPHERE, TRANSIENT STATE | =) o ) & S 0.3 [ ‘ —~ EL— Z = | O = 2 | 4, T N R 5 St SO S e -’F»' %)\’& AT SR Ry O oy S A RS ) P 2%, Fig. 8.3. Experimental Forced-Convection Volume-Heat-Source System, 99 ANP QUARTERLY PROGRESS REPORT UNCLASSIFIED DWG. 230244 T 3 10 2 5 10 2 5 104 2 5 1O5 2 3 10 Re Fig. 8.4, Theoretical and Experimental Dimensionless Differences Between Wall and Mixed-Mean Fluid Temperatures (with the Pipe Insvlated). 100 theory in Fig. 8.4, The curves represent the mathematical temperature solutions, and the points represent the experimental measurements. The experimental laminar-flow data fall about 30% below the values predicted by theory. This deviation may PERIOD ENDING MARCH 10, 1954 exist because of free-convection or phenomena. entrance The experimental turbulent-flow data were reported previously; they fall within +30% of the predicted values. TABLE 8.1, EXPERIMENTAL DATA FOR LAMINAR FLOW IN A ?{;Tin.-—!D TUBE HEAT SOURCE REYNOLDS PRANDTL. AXIAL TEMPERATURE RADIAL TEMPERATURE o o (lew/em®) MODULUS MODULUS RISE (°F) RISE, ty—t_ (°F) 0.0039 560 10.0 13.8 90 0.0077 750 10.5 20,1 15.5 101 ANP QUARTERLY PROGRESS REPOHRT 9. RADIATION DAMAGE J. B, Trice, Solid State Division A, J. Miller, ANP Division Tests of fused-salt fuels in Incone! capsules are being conducted to determine the chemical stability of the fuel and the corrosion resistance of the Incone! under high temperatures in reactor radiation fields, I!n the past, fused-salt fuels in the Naf- ZrF -UF, system were shown by several methods to be chemically stable when they were irradiated for 600-hr periods at fue! power densities of up to 8000 watts/ecm’®. Concurrent tests also have shown that centainer corrosion increases from 1to 2 mils of subsurface-void attack for out-of-pile tests to 5 to & mils of intergranular attack for in-pile high-temperature tests. Data obtained from recently completed petrographic examinations sup- port the previous indications that the fuels are chemically stable under reactor radiation. The occasional oppearance of unusually large grains in irradiated capsules has stimulated an intensive study of the errors involved in temperature measure- ments of the capsules. Miniature loops are being designed for circulating fuel in such space-restricted locations as the vertical holes in the LITR, The designs are based on the results of intensive studies of such variables as fuel flow rates, power densities, and rates of heat removal. Some components have already been constructed and tested. An in-pile stress corrosion apparatus has been developed for obtaining information on the func- tion of stress in corrosion. With this apparatus, it will be possible to determine, simultansously, the corrosion effects on stressed and unstressed portions of a tube, Tests in the LITR and in the Graphite Reactor have shown that the creep behavior of Inconel is not seriously affected by neutron bombardment, Apparatus for similor tests in the higher flux of the MTR is nearly complete. An in-pile loop for circulating fue! in a hori- zontal heam hole of the LITR is 80% complete, and two other loops are being constructed., These loops will be operated to obtain information on the chemical stability of the fuel and the corrosion 102 characteristics of liiconel under reactor irradig- tioi. RADIATION STABILITY OF FUSED-SALT FUELS G, W, Keilheltz M. T. Robinson J, €. Morgon W. R, Willis H. E. Robertson W. E, Browning C, C, Webster M. F. Osborne Solid State Division Petrogrophic Exomination of Fuels Exominations of irradiated fuels by means of an armored petrographic microscope! were made to determine the possible presence of products from fuel decomposition, Eleven samples irradiated for periods as long as 492 hr in the MTR and nine bench-tested conirol samples were examined, For these tests, three salt compositions were chosen that would yield power densities of 2500, 4500, and 8000 watts/em?3, Four irradiated capsules showed the presence of considerable amounts of ZrO, and/or UO,. Oxides were not found in any conitrol samples, in any original salts, or in any other irrodiated cap- sules. Attempts to produce oxides at room tempera- ture by the action of H202 or of gamma radiation in the presence of moisture were unsuccessful, It must therefore be concluded that these four cap- sules leaked. It was found from petrogrophic studies of the fuel from both the irradicted specimens ond the controls that only a very small degree of chemical reduction from U%* to U3 occurred. No evidence of o separate UF, phase was found. The results of the experiments therafore indicate that the presence of reactor radiation, under the conditions chosen for the experiments, has no detrimental effect on the chemical stability of the fuels studied. Tempercture Control in Static Capsule Tests The presence of large Inconel grain sizes to- gether with atypical intergranular cracking in a few Incone! capsules irradiated in the MTR sug- gested the need for careful investigation of un- certainties in such variables as wall temperatures and thermal stresses in the capsules. The temper- ature uncertainty arises from the difference in the wall temperature measured by a thermocouple from that measured without the perturbing effect of the thermocouple. Also, the thermocouple does not measure the true wall temperature because the thermocouple junction is partially cut in the air stream adjacent to the capsule and is therefore cooled by it. The problem:has received con- siderable attention from both calculational and experimental analyses. Miniature Circulating Loops The design: and construction of components for a small in-pile loop te circulate fused-salt fuels is under way, An attempt is being made to obtain an optimum design with respect to requirements for reflector-moderated aircraft reactors, The results of rigorous calculations are being used for designing the loops so that it will be possible to approach a Reynolds modulus of 2000 (turbulent flow) and a temperature drop of 100°F, | The cogent features upon which design and caicu- lations are based are (1) fuel power densities from approximately 500 to 1000 watts/cm? and (2) physi- cal dimensions such that any single loop can be completely canned and then inserted as a unit 18 in, in length by 2 in, in diameter into hole C-48 of the LITR. The principal variabies of the system are rate of cooling-air flow in the annulus around the loop, fuel flow rate through the loop, entrance and exit positions for cooling air, and physical dimensions, such as the surface-to-volume ratio of the Inconel tube. Variations of physical properties of the ait and the fuel with temperature have, as yet, been neglected, as has thermal conduction along the length of the loop metal. The calcula- tions indicate that it will be difficult, but not impossible, to obtain both turbulent flow and a large temperature difference. | Two types of pumps are being studied for use in the miniature loops. One is a stondard sump pump with a Graphitar gas seal; the other is a hydraulic check valve type of pump, as shown in Fig. 9.1, It consists of a rectifier unit containing four check valves, as shown, surge tanks above this unit, and a solenoid-driven piston to supply power. The piston pressure is transmitted to the pumped fluid by helium gas. It has been found in miniature-loop mockup tests that Inconel ball-type valves have a tendency to stick., Ball check volves made of materials such as cermets are being tested in order to find one which will be suitable. PERIOD ENDING MARCH 10, 1954 STRESS.CORROSION APPARATUS W, W. Davis J. C. Wilson J, C, Zukas Selid State Division An in-pile stress-corrosion apparatus designed to reveal the function of stress in corrosion has been built and is undergoing heat transfer tests. The test specimen is tubular and stressed in building. This arrangement should considerobly reduce the number of tests required to obtain information on corrosion, because postirradiation metallographic examination across o single transverse section will permit simultoneous observation, in the same test specimen, of the corrosive effects of the fused-salt {and, incidentally, of the sodium used as a heat transfer agent) on regions subjected to o continuous range of stresses from tensile to com- pressive, as well as on regions of zero stress along the neutral axis of the tubular beam. ' In the preseni test, the t%ubular specimen con- tains the salt, and an annulus of molten sodium surrounds the specimen to conduct the fission heat to an outer Inconel contoiner that is water cooled at its periphery so that it serves as a heat sink. The column of sodiumis cocled at its upper end to minimize vaporization, Initially, no provision is being mode for strain measurements during the test, but the design is readily adaptable to such measurements should they be desired in the future. CREEP UNDER IRRADIATION W, W, Davis J. C, Zukas N, E. Hinkle “J. C, Wilson Solid State Division At the neutron fluxes obtainable in the LITR and in the ORNL Graphite Reactor, the creep behavior of Inconel is not seriously affected by neuiron bombardment? over a range of stresses and temper- atures applicable to presently conceived aircraft reactor designs. Apparatus for tests in the higher flux of the MTR is nearly complete; the apparatus used in the paost for cantilever tests is being modified to give a constant. moment over the full goge length and to provide for use of an exten- someter other than a wmicroformer. The high incidence of microformer failures has caused their 2y, W. Davis, J. C. Wilson, and J. C. Zukas, Solid State Semiann. Prog. Rep. Aug., 31, 1953, ORNL-1606, p. B 103 vol s in NOTE © FUEL FLOW ARROWS SHOWN ) IN ONE DIRECTION ONLY. ~LOWER CAGE MATERIAL © INCONEL - PLUG, WELD FLUSH WITH 80Dy ! DIRECTION OF FUEL FLOW REMAINS THE SAME IN THIS LOOP. Fig. 9.1. Hydraulic Check Yalve Pump. UPPER CAGE UNCLASSFIED OWG D-11280A RHOdIY SSTUD0Ud ATHIAVND dNY use to be discontinued. A choice, which will depend on the outcome of bench tests now in progress, will be made between the Bourdon tube extensometer and a newly developed thermal ex- pansion (bimetal) unit, LITR FLUORIDE-FUEL LOOP W, E. Brundage C. Ellis C. D. Baumann M. T, Morgan F. M. Blacksher A, S. Olson W. W, Parkinson Solid State Division The circulating-fuel experiment for studying the stability and corrosion characteristics of molten fluoride fuels under reactor irradiation was de- scribed in previous reports.” Figure 9.2 shows the major components of the ioop. The design values for the loop are: Temperature of irradi- ated end 1500°F 1400°F Temperature at pump Fission heat generation, 1000 to 2000 watts/cm® maximur density Tota! fission heat gen- eration 10,000 te 15,000 watts Linear flow in irradiated section {0.225 in. D) 10 fps The fuel system is constructed entirely of Inconel, and the fuel will be the fluoride mixture NCIF-ZI‘FA« UF, (62.5-12.5-25 mole %), 30, Sisman et al,, ANP Quar. Prog. Rep. Sept. 10, 70, 1953, ORNL-1649, p. 106, _ PERIOD ENDING MARCH 10, 1954 The fabrication of the shields to protect operating personnel from gamma rays and deloyed neutrons from sections of the loop externa! to the reactor is almost complete. These shields ore designed in sections to facilitate removal of the pump from the loop after the fuel has become active., The pump shield will provide & in, of lead which will be augmented by stacked fead bricks and 12 in. of paroffin. The tube shield between the pump and the reactor face is comparable in thickness to the pump shield, The completed withdrawal shield, into which the loop and water jacket will be with- drawn after removal of the pump, is o 15§ hori- rontal cylinder with an 8]/2»'in. central hole and 5-in. lead walls. The disassembly of the fuel loop after operation in the reactor will require additional tools and « special door in the Sclid State Division hot cells, The design of the door is complete, and the neces- sary tools have been ordered. Design wark to modify the tools for remote operation is 80% cem- plete. Fabrication of parts for three complete loops is about 50% complete. The major effort is on the finishing and testing of parts for a single loop, which is about 80% complefe. Parts for three sump-type pumps, including one for test purposes only, have been completed (cf. sec. 2, ""Experimental Reactor Engineering’’). The test pump, which was modified after being tested with water, is being instalied in o fluoride system. A spirgl-fin type of heat exchanger was developed and found adequate for removal of the fission heat generated in the fuel ot o rate of 10,000 wotts. Approval for insertion of the loop inte hole HB-2 of the LITR has been granted by the Reactor Safety Committee, 105 ANP QUARTERLY PROGRESS REPORT X — "RAGIATOR" SECTION OF LOOP - PROJECTED AND ENLARGED VIEW DwG, 23122 Fig. 9.2. In-Pile Loop. A — LINER, WATER JACKET B — WATER FLOW THROUGH ANNULAR SPACE C —~ COOLING WATER OUT D — COOLING WATER IMN E — AR INLET F -~ AIR OUTLET G —~ FACE PLATE H - PORTS FOR THERMOCOUPLE, HEATER, AND YENTURI LEADS | — FUEL TUBES TCO PUMP J - FUEL TUBE JACKET (HELIUM FILLED) K- YENTURI METER L ~ AIR-TO-FUEL HEAT EXCHANGER AIR-TO-FUEL HEAT EXCHANGER 106 M - N - 0~ Q - R — S . T - U- V. W — X - Y - Z - GRAPHITE STEEL HELIUM-FILLED SPACE CALROD HEATERS (OYER ENTIRE LENGTH OF LOOP) STAINLESS STEEL SHEET WRAPPING HIGH-TEMPERATURE WRAPPED INSULATION B4C-Fe DISKS (NEUTRON SHIELDING) FUEL TUBING (0.404 in. ID, 0.500 in, OD) FUEL TUBING (0.225 in. ID, 0.275 in. OD) STAIMLESS STEEL COVER IRRADIATED SECTION OF LOOP RADIATOR FINS INTERNAL HEATER, NICHROME V HELIX RADIATION SHIELDING PERIOD ENDING MARCH 10, 1954 10. ANALYTICAL STUDIES OF REACTOR MATERIALS C. D. Susano Analytical Chemistry Division J. M. Warde Metallurgy Division R, Baldock Stable lsotope Ressorch and Production Division Investigations were continued on methods for determining the oxidation states of the corrosion products, iron, chromium, and nickel, in ARE fuels and fuel solvents and methods for the separation and determination of UF; and UF, in ARE fuels. A procedure was devised for the determination of chromous fluoride and ferrous fluoride in NaZrF . This method is not appliceble, however, to samples of the ARE fuel. The possibility of selectively oxidizing divalent chremium and trivalent uranium fluorides was investigated. Hexavalent uranium in the form of urany! acetaie was tested as o possible oxidont for trivalent uranium to the tetravalent state., Various media for this reaction were used, but no satisfactory system was found that would prevent oxidation of trivalent uranium by the hydrogen ion., The oxidation of trivalent uranium to the quadrivalent state with iodine in 80% methanol solution wos shown to be essentially quantitative. The preferential dissolution of UF, from UF3 in ARE fuels by means of solutions of ethylenediaminetetrascetic acid and ammenium oxalate was shown to be impractical when applied to actual samples of the fuel. Trivalent uranium fluoride, when heated with NflZny was found to be readily soluble in solutions of ammonium oxa- lote. It was established that the solubility of UF, in these solvents involves, first, oxidation to tetravalent uranium and, then, dissolution. The conversion of UF, and UF to the corresponding chlorides by fusion with NoAICl, was accom- plished. The urcnium chlorides are readily ex- tracted by acetylacetone-acetone mixtures. The structure of the uranium compounds with acetyl- acetone was established by potentiometric titra- tion. Samples of the fluoride mixture NaF-Z¢F -UF, that were exposed fo moisture before being canned were examined with the polarizing microscope after having been subjected to gommo rodiation in the MTR canal for 265 hours, In comparison with fuel stored and canned in o helium atmosphere, no effects of irradiation could be observed. The Analytical Chemistry Laboratory received 917 samples and reported 994 samples involving 9709 determinations. ANALYTICAL CTHEMISTRY OF REACTOR MATERIALS J. C. White Analytical Chemistry Division Dxidation States of Chromium and iron in ARE Fuel Solvent, NoZrF, D, L. Manning Analytical Chemistry Division In cooperation with the ANP Reactor Chemistry Group, which is investigating the equilibrium kinetics of corrosion mechanisms in NaZrFs, methods have been devised for the determination of the concentrations of the products and reactants of the following reaction: NaZrF Cr0 4 FeF, ~——— CrF, + Fel | 800°C Chromium metal and ferrous fluoride are mixed with NaZrFs in a nickel container, heated to 800°C, and filtered through o nickel frit. Samples are taken of the solidified filtrate and the residue, It is assumed that filtration will remove the metal- lic constituents and that the filtrate will contain oxidized chromium and unreacted ferrous fluoride, The residue was observed to contain magnetic particles, that is, metallic iron, and thus the reduction of ferrous fluoride was demonstrated. Since the residue was heterogeneous, only the con- centrations of total iron and chromium were deter- mined in the residue. The filtrate was free from magnetic particles, For the calculation of equilibrium, divalent iron and divalent chromium were determined in the 107 ANP QUARTERLY PROGRESS REPORT filtrate, The following reactions were assumed in making these calculations: Determination of Fe™™ (1) FeF, + H,50,— Fett + 80,77 + 2H" + 2F~ (2a0) Fe't + o-phenanthroline —— (Fe-o-phenanthroline™™) Determination of Crt? (3) CrF, + Fet——= Cr¥ 4 Fett 4 2F- (2b) Fe'" + o-phenanthroline —— (Fe-o-phenanthroline ™) The determination of ferrous iron was accom- plished by extracting the iron with 0.2 M H,S0,, as outlined in a previous report.! An indirect method was used for the determination of divaelent chromium. The somple was dissolved in o 1% (w/v) solution of ferric sulfate, as reaction 3. The ferrous iron formed was then measured colorimetrically as the ferrous o-phen- anthroline complex. The concentration of FeF originally found (reaction 1) was subtracted from the iron found by reaction 2b. The total concen- tration of iron and chromium in the filirote was determined by conventional colorimetric methods. The results indicate thot the reaction proceeds essentially as written and that no other oxidation shown in stotes of these particular metals are present. This procedure is presently being used in the laboratory, The interferences obviously include any ion that will reduce ferric ions to ferrous ions, The presence of uranium makes impractical the application of this method to samples of the ARE fuel. The correction for the tetravalent uranium present (8 wt %} would seriously limit the precision and accuracy of the methods for determining chromium and iron. Work is under way on the development of suitable procedures for application of the ARE fuel. Oxidation States of Chromium and Uranium in ARE Fuel Solvent, NaZrF D. L. Manning Analytical Chemistry Division Work has continued on the determination of the different oxidotion states of chromium and uranium 1J. C, White et al., ANP Quar, Prog. Rep. Dec. 10, 1953, ORNL-1649, p. 107, 108 which occur as a consequence of the following reaction: NaZeF CrF3 + UFaE-'-‘Z‘-‘:”"A Cer + UF 800°C The melt is filtered ot 800°C, ond samples of the filtrate and residue are collected. Previous tests! showed that 0,2 M ammonium oxalate selectively leached both UF, and CrF, from UF, and CrF,, respectively. However, odditional experiments demonstrated that when CrF; and UF, were allowed to react in NaZrF_. ot 800°C, the entire sample was soluble in ammonium oxalate solution. A sample of NaCrF ., formed by heating CrF, and NaF, also dissolved readily in 0.2 M (NH,),C.0,. Evidently, the complex formation of CrF; and NoF render the chromium compound susceptible to dis- solution by oxalote. The absence of divalent chromium in this sample of NaCrF, was established by the ferric sulfate test (reaction 3). As a result of this discovery, the possibility of selectively oxidizing chromium and vranium was investigated. A tentative procedure was formulated 4 . which invelves the following steps. . Determine the reducing power of CrF, and UF, by the use of an oxidant that will oxidize these compounds to the trivalent and quadrivalent states, respectively. 2. Determine the reducing power of CrF,, UF,, and UF, by the use of an oxidant that will oxidize divalent chromium to the trivalent state and UF, and UF , to the hexavalent state, 3. Determine total uranium, 4, Determine total chromium, From the results of these four steps, the four unknown gquantities can be calculated. Steps 2, 3, and 4 present no difficulty. The problem of solving step 1 is more formidable. Hydregen ion (H") can be used to oxidize Cr*™ to Ce3t and U3 to U*': however, adequate precision of the measure- ment of the hydrogen produced has not been at- tained, as yet, In addition, the volume of gas is usually less than 0.5 cm?, for practical purposes. Hence, attention has been turned to other possible oxidants for this step. Oxidotion of UF, by Hexavalent Uranium. Oxida- tion of trivalent uranium to quadrivalent uranium by hexavalent uranium can be represented as 2U3% + U0, " 4 4HY —3 U4 4 2H,0 . An obvious advantage of this reaction is that oxidation proceeds to the quadrivalent state. The difficulty in this reaction is to ensure oxidation by UO,*" and not H¥, as is usually the case. Experiments were conducted in which UF, was dis- solved in excess uranyl acetate solution and the excess UO;’+ was determined polarographically by utilizing the hexavalent uranium reduction wave at about —0.2 volt vs, SCE, Helium was used to sweep the solution before the addition of UF,, and the reaction was conducted in o helium atmosphere. The reaction was vsually complete within 30 min, and green solutions, indicative of quadrivalent uranium, were obtained. The results obtoined in various acid media are shown in Table 10.1. The hydrogen ion apparently is the stronger oxidizing agent in these tests. The oxidation potential of the UO;’*“/U“' couple should be increased in a solution of high hydrogen-ion con- centration, This hypothesis is confirmed by the data obtained, since oxidation of UF, by the U02++ ion increased with increasing hydrogen con- centration from 0 to 44% in sulfuric acid media. However, complex formation of UC)2Jr+ and sulfate TABLE 10.1. OXIDATION OF URANIUM TRIFLUORIDE BY HEXAVALENT URANIUM Urany! Acetate: 0.0021 M UF, (%) UF, (mg) SOLVENT OXIDIZED BY 4+ 4 uo, H 42.6 0.1 N H,50, 0 100 31.1 0.3 N H_,SO, 27 73 54.8 1,0 N H,S0, 44 56 39.2 6.0 N H,SO, 38 62 58.8 1.0 N H,C,0, 47 53 29.2 0.5 N H,C,H,0, 0 100 42.0 1.0 N H,C H,0, 0 100 37.4 2.0 NH,C H,0, 0 100 33.8 4.0 N H,C,H,0, 0 100 51.4 1.0 N HCIO, 0 55.6 2.0 N HCIO, 0 56,7 4.0 N HCIO, 0 40.3 0.2 M (NH,),C,0, o* 0 *UF3 did not dissolve in 2 hours, PERIOD ENDING MARCH 10, 1954 ions was so marked at higher concentrations ‘that no further increase in oxidation potential was observed. The use of perchloric acid media was unsuccessful because the perchlorate ion was reduced by trivalent uranium. None of ‘the hexa- valent uranium was consumed in the acetic acid tests. A preliminary result with an oxalote medium is shown in Table 10,1, Further work is under way with oxalate media, as well as with hydrochloric acid media. Oxidation of Trivalent Uranium to Quadrivalent Uranium with ledine. In an effort to eliminate oxidation by the hydrogen ion, the use of non- aqueocus systems is being investigated. Prelimi- nary work has been concerned with UCI, rather than with UF ; because of its greater solubility in nonaqueous systems. The conversion of UF, to UCI, by fusion with NaAlCl, hos been investigated and is reported below under the heading “‘Sepora- tion of UF , from UF,." lodine is sufficiently soluble in organic solutions to permit its use as an oxidant., In aqueous media the oxidation will probably proceed beyond the tetravalent state, as shown in the following reac- tion, to produce some hexavalent uranium: Ut ¢ L1, ——s Ut 4 I Known amounts of UCl, were dissolved in excess standard iodine solutions of various concentrations of methanol in water, and the excess iodine was titrated with sodium thiosulfate, The results indicate that the oxidation potenticl of iodine decreases with increasing methanol con- centration and that, in absolute methanol, incom- plete oxidation results, In an 80% alcoholic medium the oxidation of U¥* to U*Y is essentially guantitative under the conditions of the test, Further tests are under way, and a more sensitive means of detecting the end point is to be employed. The iodine-starch end point cannot be used in concentrated methonol solutions. A potentiometric titration and the ‘‘dead stop’’ end-point detection systems are to be tested, Separation of UF, from UF, W. 1. Rodd J. L. Mottern Analytical Chemistry Division Efforts to develop methods to separate UF, and UF, in ARE fuels have been continued along two lines: (1) preferential dissolution of UF, and (2) conversion of the fluorides to the corresponding 109 ANP QUARTERLY PROGRESS REPORT chlorides and subsequent separation of the chlo- rides. Preferential Dissolution of UF,. The solubili- ties of UF, and UF, in solvents such as solutions of ethylenediaminetetraacetic acid (EDTA) and ammonium oxalate were studied further. Uranium trifluoride is soluble to the extent of 0.2 to 0.4 mg per 100 mi of selvent in 0,1 M solutions of EDTA et a pH of 6.8 to 6.9 at 25°C after 1 hr of contact. The solubility of UF; is a function of time of contact. The rate of dissolution appears to be about 0.4 mg of UF; per 100 ml of solvent per hour of contact, The absorption spectrum, between 370 and 670 my, of the supernatant liquor of the mix- ture of UF; and EDTA is similar to that of a solu- tion of UF, in EDTA. This indicates that the uranium exists in the same oxidation state in both solutions as was reported previously ! ond that the solubility process essentially involves, first, the oxidation of trivalent uranium to the tetravalent state and, then, dissolution. The time required for complete dissolution of 50 mg of UF ;. in 100 ml of 0.1 M EDTA solution varies from 1 to 20 hr at 25°C, depending upon particle size and previous treatment, Inasmuch as the rate of dissolution of UF , varies so markedly, a signifi- cont amount of UF ; would be dissolved in ensuring complete dissolution of UF,, Any appreciable solubility of UF, would cause large errors in determinations of UF, and UF, in a sample which contained UF , in a concentration greatly exceeding that of UF,, Such an inherent error in the method limits its applicability to the present problem. Further study of the solvent action of soluticns of 0.2 M ammonium oxdlate upon the solubility of UF, ot 100°C in a helium atmosphere indicated a similar time effect. The rate of dissolution of a batch of UF, which had been analyzed as 90.4% UF,, was found to be of the order of 6 to 13 mg per 100 ml when samples of 50 to 100 mg were refluxed for 1 hr under helium. The soluble fraction was approximately 6.5% of the sample. Analysis of the undissolved portion showed that it contained 91.5% UF,. After the initial hour of contact, the rate of dissolution remained constant ot 0.4 o 0.5 mg per 100 m! per hour over a 4-hr period at 100°C. At 25°C, the rate of dissolution, after the initial hour, was of the order of 3 mg per 100 ml per 24 hours. The dbscrption spectra of the UF ;-oxalate ex- tracts are identical to those of UF ,-oxadlate solu- 110 tions. These results reveal that UF, is slowly oxidized to UF‘1 in oxalate media. However, inas- much as 100 mg of UF , rapidly dissolved in 0.2 M ammonium oxalate, preferential dissolution of UF from UF, is more feasible in this medium than in EDTA solutions, Inorder to determine the feasibility of preferential dissclution based on differences in rates of solu- tion, a mixture of 8% UF; in NaZrF, was heated at 800°C fer 3 hr in a nickel container, and the melt was filtered through a nickel frit, The filtrate was olive drob after solidification, and it con- stituted opproximately 60% of the total sample. The residue, retained on the frit, was an orange solid, which was tentatively identified by petro- graphic examination as UF ;.27¢F ,. The composi- tion of each phase is shown in Table 10,2. TABLE 10.2. COMPOSITIOM OF FILTRATE AND RESIDUE FROM REACTION OF UF3 AND NanFs AT 800°C FOR 3 HOURS COMPOSITION (%) FLEMENT os:ii:: Filtrate Residue ud+ 6.46 4.42 4.07 U, total 6.46 6.37 6.42 Zr 40,1 29,9 40.2 Na 101 10.8 10.7 F 43.4 42.1 42.0 Ni 0.07 0.002 Cr 0.002 0.002 Fe 0.0125 0.011 Both the residuve and filtrate were readily soluble in 0.2 M (NH,),C,0,; 200 mg dissolved in 50 ml of solution in 30 minutes. This solubility shows that a radical change in the nature of UF, results when it is heated with NaZrF . in a nickel vessel. Similar behavior was observed with CrF,, as re- ported previously. The data demonstrate con- clusively that a separation of UF; and UF, based on the difference in rates of solution is not feasible. The similarity of the composition of the filtrate and the residue is unique in view of the difference in color, Only the nickel content is different, and this may be explained on the basis of the nickel It was alsc noted that the residue, upon exposure to the atmosphere, slowly changed over a period of two weeks from orange to olive drab. The trivolent uranium content, as expected, decreased to 2.29%. The trivalent uranium con- centration in the filtrate also decreased upon standing to 2.66%. These data indicate that UF, is not stable when it stands in closed, air-filled containers. Additional determinations of UF, will be made to observe further oxidation. Conversion of UF, and UF to the Corresponding Chiorides. The literature? reveals that UCI; and UCH, are soluble in a number of organic solvents, It is also known® that UF, can be converted to UCI, by fusion with NaAICl,. Hence, the separa- tion of UF, and UF, as chlorides merits investi- gation. The reactions involved in the metatheses container used. are 300°C | NaCl + AICI, NaAICI, o 300 3UF, + 3NaAICI, —- > 3UCI, + 3NaAIF, , 4UF , + 3NaAICI, --3-99'-§~'9 4uct, + 4UF|3 + 2NaA|F4 . Optimum conditions for these reactions are being established. A flux ratic of 10:1 and 1-hr fusion ot 300°C are necessary to give greater than 90% yield. Attempts to produce a quantitative reaction in cne fusion are under way. Nenaqueous media have been used to extract the uranium chlorides from the melt. Agueous media would result in immediate hydrolysis and oxidation of the salts. The solubility of UCI, was checked in a number of solvents in which UCI be readily soluble. is known to These solvents ranged from the lower olcohols to acetone and acetylacetone, dimethy lformamide, dioxane, and ethyl acetate. In none of the solvents tested was UCI, insoluble; the lowest solubility was noted in dioxane — 1 mg per 100 ml after ¥ hr at 25°C. The highest solu- bility was in acetone — 68 mg per 100 ml, under similar conditions. Stakble suspensions of UCI, were noted in the alcoholic solvents. 24 Summary of the Properties, Preparation, ond Purifi- cafion of the Anhydrous Chlorides ond Bromides of Uranium, TC-1974 (A-28B) (Sept. 15, 1944). 3‘\/‘, P. Calkins, Dissolution of Uranium Tetrafluoride, Report H-1.740,8, Poper No. 12 (Apr. 1, 1947), os re~ vised by C. D, Susano, AECD-3064 {Oct. 3, 1949). PERIOD ENDING MARCH 10, 1954 A 1.1 mixture of acetone and acetylacetone quantitatively extracts vranium chlorides from the Fifty milliliters of the solvent will extract 100 mg of uranium chlorides in about 2 hours. Aluminum is also exiracted, as is some chromium, iron, and nickel. The behavior of zirconium has not been established quantitatively, but it is assumed that zirconium will also be ex- tracted., A study was initigted to determine whether g3t remains in on unaltered oxidation state after being extracted by acetylacetone. If so, it con be de- termined by potentiometric methods. The poten- tiometric method of determining the chelation nature of a metal chelate developed by Van Uitert et al.? was used to study the siructures of the UGYY and U(IN) acetylacetone solutions. The hydregen ions liberated during chelation were titrated with NaOH to determine the equivalents of acetylacetonate which combined with 1 mole of U3* or U4* in aqueous and alcoholic solutions of UCI,(UCH )-acetylacetone. Theresults indicate that 3 moles of acetylacetone is required to chelate U3* and 4 moles is required to chelate U%* in both oqueous and nonagueous media. Strong evidence is provided of the nature of the complex. Trivalent uranium remains unoxidized in the complex. Attempts to isolate U({ll) acetylacetonate and U(lV) acetylacetonate by precipitation from a neutral solution were made. The solid obtained in all tests to date exhibits very low stability to air and heat, and it has not been purified sufficiently for analysis. fluoride conversion melts. A procedure developed by Sone® has been used to determine the stability of the acetylacetonates, This method is used o determine the relative stability of the chelate as a function of the shift of the absorption band of the acetylucetone spectra; that is, the greater the stability of the cheigating bands, the larger the shift foward higher wave lengths. The close similarity of the absorption bands of UCl -acetylacetene, UCI -acetylacetone, and acetylacetone, with respect to wave length, indi- cates that very unstable bonding exists between 4. G. Van Uitert, B, E. Douglas, ond W. C. Fernelius, A Potentiometric Study of S-Diketone Chelation Tend- encies 1ll, Metal Chelote Comparison, NY0-7276 (May 2, 1951). 5K, Sone, J. Am. Chem. Soc. 75, 5207 (1953). 1t vranium and the chelating agent. Further study of the large differences in molar extinction coeffi- cients between U(lIl) and U(IV) acetylacetonates is contemplated. The possibility of determining U3 in the presence of U** ions in solution is now being studied. Differential oxidation by weak oxidizing agents and the application of potentiometric methods are to be used. FLUORIDE FUEL INVESTIGATIONS G. D, White T. N. McVay, Consultant Metallurgy Division Two samples of the fluoride mixture NaF-ZrF ,- UF,, supplied by the Solid State Division, were studied under the polarizing microscope. These samples had been exposed to moisture prior to canning and subjected to gamma radiation (5,1 x 10%r at 79°F) in the MTR canal for 265 hours. The object of this experiment by the Solid State Division was to determine the effects, if any, that the presence of moisture may have on radiation stability of the fuel in comparison with fuel stored and canned in a helium atmosphere, One sample was canned in the laboratory atmosphere, and the other was exposed in a desiccator containing water for 12 hr prior to canning in the laboratory atmosphere. Both samples were found to contain hydrated N03Zr4F19, in addition to NaZr(U)F., as was found previously when NaF-ZrF ,-UF , was exposed to an atmosphere containing moisture. The irradia- tion produced no effects which could be observed with the polarizing microscope. The usual routine petrographic examinations were made for the ANP Fuels Section. SUMMARY OF SERVICE CHEMICAL ANALYSES J. C. White L. J. Brody A, F. Roemer, Jr. C. R. Williams Analytical Chemistry Division In addition to the usual number of analyses of samples of corrosion tests of ARE fuels and fuel solvents in metal containers, a considerable portion of the laboratory work load was concerned with the determination of oxidation stotes of iron and chro- mium in NanrF5 and the fuel mixture. A number of samples of sodium and of sodium-potassium alloys were analyzed for oxygen and metallic constituents, A total of 917 samples was received, 994 were reported, and 8709 determinations were made (Table 10.3). TABLE 10.3. SUMMARY OF SERVICE AMALYSES REPORTED NUMBER OF NUMBER OF SAMPLES DETERMINATIONS Reactor Chemistry 641 4866 Experimental Engineering 299 3477 Fue! Production 16 204 ARE Fluid Circuit 36 146 Heat-Transfer and Fluid Properties 2 16 Total 994 8709 112 Part |l SHIELDING RESEARCH 11. SHIELDING ANALYSIS E. P. Blizard J. E. Faulkner M. K. Hullings F. H. Murray Physics Division H. E. Stern, Convair Estimates of removal cross sections based on the continuum theory of the scattering of neutrons from nuclei have given values which are in reasonable ogreement with measured values. The shielding properties of lithium hydride and water have been compared by using the concept of the effective neutron removal cross section. If lithium hydride could be substituted for water os the neutron shield in an aircroft reactor, there would be a considerable saving in weight. ESTIMATES OF REMOVAL CROSS SECTIONS BASED ON THE CONTINUUM THEORY OF THE SCATTERING OF NEUTRONS FROM NUCLE! F. H. Murray Physics Division The asymptotic character of the attenuation of neutrons through large thicknesses of material was expressed in a previous progress report! in the form exp(—ox/A), where A is the largest eigenvalue of an infinite matrix. The scattering was assumed to take place ot energies of several Mev (>4) from nuclei sufficiently heavy so that energy losses during the elastic-scattering process could be neglected. In order to apply the results to the calculation of removal cross sections for neutrons from a fission source and for a material between IF. H. Murray, Phys. Div. Semiann. Prog. Rep. Sept. 10, 1953, ORNL-1630, p. 6. TABLE 11.1, the source and a large thickness of water, it is necessary to compute on average value of the attenuation through the material, The “‘uncollided flux” of Welton and Blizard? was employed to calculate the average of the exponential attenuation from the formula fS(E,z:) e"gt/'\ dE =t ' [ste,2) ae with [s(,2) dE = 37.8( + 5% In this calculation, the distance from the source, z, was 120 cm, and the thickness of the material, t, was 10 or 15 cm. The logarithm of the function S(E,z) exp(—ot/A) was plotted and represented in the form —C(E - Eo):E near its maximum, after which the integral was calculated. 8-1.547(24«8)‘/‘ i The total cross sections of various materials were found from the curves of Nereson and Darden.? lron and copper values were adjusted by multiplying their cross sections by a constant factor (0.96) to make their curves pass through the point at 14 Mey, which represents a value of the cross section determined by other authors. The results are presented in Table 11.1, 2T. A, Welton and E. P. Blizard, Reactor Sci. Technol, 2, No. 2, TiD-2002, p. 73, esp. 85 {1952). 3N. Nereson and S. Darden, Phys. Rev. 89, 775, esp. 782-783 (1953). REMOVAL CROSS SECTIONS OF SEVERAL MATERIALS REMOVAL CROSS SECTION, 2 (barns/atom) MATERIAL From Experiment From Caleulation Al 1.31 1.23 Fe 1.95 1.95 Cu 2.08 2.2 Pb 3.4 (estimated) 3.36 Bi 3.43 3.32 115 ANF QUARTERLY PROGRESS REPORT For lead and bismuth the forms of the total-cross- section cuives indicate that the “continuum theory” probably does not apply, but the energies of importance were sufficiently high (rear 10 Mev) to suggest that the caleulation might give values which were nearly correct. If the water thickness were about 60 cm, the energies of importance would be much less? and the application of the continuum theory would be less justified. CRITIQUE OF LITHIUM HYDRIDE AS A NEUTRON SHIELD J. E. Faulkner Physics Division Water has been the most common neutron shield coimtemplated for aircraft use to date, and any compound substituted for it must give greater attenuation for the same shield weight. Some rather lengthy calculations performed on the UNIVAC under the direction of NDA* have shown that @ slab of lithium hydride used as a neutron shield will weigh only 63% as much as a thickness of water which gives the same aHenuation, Ad- ditional comparisons of the shielding properties of lithium hydride and water have now been made on the basis of the effective neviron removal concept.® Calculations made by using removal-cross-section values of 2.4 barns/mole for lithium hydride and 3.02 barns/mole for water® indicate that o lithium hydride shield would weigh 56% as much as a water shield. This is considered to be in reasonable agreement with the NDA results, Lithium hydride is a better neutron attenuator on a volume basis, as well as on o weight basis, and would therefore be 4. . . . Private communication. 3. D, Flynn et al., Phys Div. Quar, Piog. Rep. Dec. 20 1952, ORNL- 1477,p ®This valus for water is, of course, valid for only ane shisld thickness, but it is reasonably constant for the thicknesses considered here, 116 particularly important for spherical reactors where compactness means greater weight saving. the calculations, the density of lithium hydride is taken to be 0.78 g/cm®, the value measured by helium displacement.” For a spherical reactor with a 100-cm-thick lithium hydride shield beginning 100 cm from the core, the thickness of water neces- sary to give the same attenuation would be about 137 cm — a LiH-to-H,0 weight ratio of 0.44, In a practical shield for a reflector-moderated reactor (cf., Sec. 3, ‘'Reflector-Moderated Reactor'”) operating ata power of 600 megowatts, the maximum temperature that would bereached in a 100-cm-thick lithium hydride slab exposed to the resulting neutron flux was calculated to be 315°F., The maximum therms! gradient would be 1.12°C/em. In these calculations it was assumed that both faces of the slab were cocled to 300°F and that the thermal conductivity of lithium hydride was 0.01 cal/°C.cmsec, Lithium hydride decomposes only slightly balow its melting point (1256°F), although it begins to soften at 1184°F, The question of radiation dissociation, as distinct from radiation heating, is In all not well understood, although the outlook is favorable. ln an investigation at Argonne,® lithium hydride under o 50-lb hydrogen pressure was exposed to a flux of 10'! for a period of three months. After irradiction, the pressure was still 50 pounds. This indicated that lithium hydride dissociagtion wunder radiation is completely re- versible, and it might be necessary in a practical shield to use a pressure shell which could with- stand the 50-1b pressure. This would, of course, add 1o the shield weight; however, a pressure shell would be needed for the water shield too, particularly at temperatures of 300°F, T. P. R. Gibb, private communication. BRepon‘ for July, August, ond September, Experimental Nuclear Physics Division; ond Summary Report for April Through September, Theoretical Physics Division, Argonne Nationol Laboratory, ANL- 4208 (Oct. 4, 1948), PERIOD ENDING MARCH 10, 1954 12. LID TANK FACILITY C. L. Storrs GE-ANP G. T. Chapman D. K. Trubey J. M. Miller Physics Division Investigations of the effects of one and two air ducts on the fast-neutron dose received outside a reactor shield have been continued at the Lid Tank Facility, In addition, the radiation dose measurements made behind 82 mockups of the reflector-moderated reactor and shield have been analyzed, ond the effective neutron removal cross section of B,0, has been measured. AIR-DUCT TESTS J. M. Miller Physics Division The air-duct experimentation was continued both os a study of the interference between adjacent ducts and as a study of neutron-streaming through a single duct. The ducts consisted of from one to three straight cylindrical sections (22 in. long and 3‘57]6 in. in diometer) joined at angles of 45 degrees. The first section of each duct was placed adjacent to the source at an angle of 221’2 deg with the normal. At the time the previous studies of interference were reported,] measurements had been made beyond a three-section duct with a one-section duct placed near it. In the first measurement, the one-section duct was parallel to the first section of the long duct and ]331/2 in. from it. in this position, it was in the same plane as the long duct and lined up approximately with the last section. In the second measurement, the small duct was again parallel to the first section but closer (6 in.) and above it in a different plane. It wasnoted that, with the short duct in line, the fast-neutron dose was a factor of 4.5 higher than for the long duct With the short duct not in line, the peak of the dose was the same as that for the long duct alone, although the curve was broadened. alone. An additional measurement was made recently in which the axis of a one-section duct intersected the axis of a three-section duct (offset by an angle 'c. L. Storrs et al., ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL.-1649, p. 121, of 22]/52 deg from the plane of the long duct). In this position, there was neutron-streaming from the short duct into the middle section of the long duct to the extent that the fast-neutron dose was in- creased by a factor of 1.4, It is assumed that the streaming through the one-section duct was con- siderably higher than wos indicated by the dose measurements. The neutrons undoubtedly passed straight across the long duct and scattered in the water beyond. Other measurements have been made on single ducts consisting of cne and two sections. The fast-neutron isodose plots for both these ducts are given in Figs. 12.1 and 12.2. The curves in Fig. 12.2 show that some of the neutrons ‘‘feed out’’ of the duct ot the end of the first section, as would be expected. These experiments will be described in detail in a separate renp-::»n‘.2 SHIELDING TESTS FOR THE REFLECTOR-MODERATED REACTOR F. N. Watson, Physics Division® R, M. Spencer4 F. R. Westfall® In several previous progress reports,>~7 the status of the experimentation on mockups of the reflector- moderated reactor and shield was reported. The tests have now been completed, and an analysis of all the data will be published.?2 A summary of the results follows, 2C. L. Storrs and J. M. Miller, Some Neutron Measure+ ments around Ajr Ducts, ORNL CF-54-2-923 (o be published]. I Now assigned to Tower Shielding Facility. 4U. s. Air Force. SJ. D. Flynn et al.,, ANP Quar. Prog. Rep. Mar. 10, 1953, ORNL-1515, p. 89, 6]. D. Flynn et al., ANP Quar. Prog. Rep. June 10, 1953, ORNL.-1556, p. 119. 7C. .. Storrs et al., ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609, p, 128, : SF, H. Abernathy et al,, Lid Tank Shielding Tests for the Reflector-Moderated Reacter, ORNL-1616 f{to be published}. 117 ANP QUARTERLY PROGRESS REPORT ORNL-LR-DWG 148 e [ | | | i | l 1 mrep/hr —-— - ——+ i S R T e e N | i > N ‘| ' 10 < mrep/nr —\ U ! . \ : 55 | e e | e S 1‘\3.\_._.. S " : ‘ ‘ 10 mrep/hr =~ N —_ ‘ : | AN g 45 .___)‘. _l ,,7,.‘7, Jli P ! . w : | ! | ' o y | TOP VIEW OF DUCT 5 L | IN LID TANK : Zes, N et — oy} S : B L o g ] a — wy e a 1 x> 10 20 30 40 50 60 70 80 90 100 z, DISTANCE FROM SOURCE (cm) Fig. 12.1. Fost-Neutron Isodose Curves Beyond One-Section Duct, ORNL-LR-DWG 119 1, 162 mregp/hr Q 1 mrep/hr A 10 mreasir S0 Cmrearhr - TOP VIEW CF DUCT IN LID TANK . (NLIDTAN NTER LINE 1cm) N on ¥, DISTANCE FROM C SOURCE ro o rspiss 30 40 50 EQ 7O BO 8C 100 1O 120 2, DISTANCE “RCM SCURCE (cm) 10 20 Fig. 12.2, Fast-Neutron lsedose Curves Around Two-Section Duct. The beryllium reflector region should be about 12 in, thick to obtain a good compromise between over-all reactor shield weight and core reactivity requirements. The thickness of the heat-exchanger region has relatively little effect on neutron and gamma doses as functions of the distance from the source plate. 118 For the gamma-shield thicknesses tested (0 to 7.5 in.), the substitution of lead for water has practically no effect on the neutron attenuation well out in the shield. A 0.13-in.-thick layer of g0 (density = 2.1) is as effective as 1 in. of B,C (density = 1.95) in depressing the thermal-neutron flux and consequent capture gammas. Dividing the lead region into layers separated by borated hydrogenous material gives some reduction in gamima dose for a given thickness of lead; however, the full-scale shielddesignis simplified structurally by plocing all the lead together just outside the pressure shell, While lumping the lead increases the lead thickness required, keeping its radius to a minimum largely compensates for this; therefore very nearly minimum over-all shield weights can be obtained in this way for lead thicknesses of up to 6.0 inches, On a volumetric basis, transformer oil is as effective as water in ottenuating the neutron flux, Since the density of the oil is about 0.87, an appreciable saving inshield weight can be obtained by using oil, if it is borated. Some of this weight saving is offset, since the thickness of the lead layer must be increased because the attenuation of the gamma flux is not so great in the oil, Beryllium is more effective than water on a thickness basis for fast-neutron attenuation. ft is important that a boron curtain be used be- tween the heat exchanger and the pressure shell, as well as between the reflector and the heat exchanger, to inhibit capture gamma production and to reduce secondary coolant activation. it does not appear to be worthwhile to use rubidium instead of sodium or NaK as a secondary coolant. Potassium is preferable to sodivm with regard to activation, but it is inferior as a heot- transfer medium. Air and structure scatiering, fission-product decay gammas from the heat- exchanger region, ducts and voids, and optimization of the shield size and weight pose problems that require further investigation. Lid Tonk Dese Meusurements Corrected to Designed Reflector-Moderated Reactor Shield An effective, preliminary, reactor shield design9 was based on the tests described above. Dose-rate 9A. P. Fraas, ANFP Quar. Prog. Rep. Mar., 10, 1953, ORNL-1515, Fig. 4.25, p. 62, and Fig. 4.32, p. 77. curves corresponding to the design were obtained by correcting measurements taken behind the configurations that most closely approximated the design, and these, in turn, were used for the shield weight calculations reported below. PERIOD ENDING MARCH 10, 1954 Configuration 62 and the designed shield are compared in Table 12.1. Configuration 62 closely approximates the designed shield, but in con- figuration 62 there is a 6-in. lead gamma shield, and in the designed shield the thickness of this TABLE 12.1. COMPARISON OF DESIGNED REFLECTOR-MODERATED REACTOR SHIELD ASSEMBLY WITH CONFIGURATION 62 DESIGNED SHIELD CONFIGURATION 62 Thickness Thickness Componenf( al I(in.) Compone nf(a) (i) H,0 1.06 {nconel 0.125 Fe 0.19 Be (Na cooled) 12.0 Be tank 2¢F) 12.20 Inconel 0.0625 !0 0.13 B,C tank(® 1.19 Heot-exchanger S(C) NaF tank l(b) 1.38 region NaF tank Q(b) 1.38 NaF tank 4(0) 2.12 Inconel 0.125 10 0.13 B,C tank 1.19 Inconel 1.50 Fe 1.75 Insulation 0.5 Ph xd Pb 6.00 Plexibor 0.19 Air 1.42 H,0 (1% B) yld) Fe 0.19 H20 (1% B) 24,22 Fe .19 Rubber 2.00 Total 16.6975 + X + ¥ 54.60 a . . ( )Components listed in sequence from surface of source. (b)Be tank 2: 0.64 cm of Al; 7.3 cm of Be (¢ = 1.84 g/cm3); 0.16 cm of Al, 0.03 cm of stainless steel, 7.3 cm of Be; 0.16 em of Al; 0.02 cm of stainless steel; 14.61 cm of Be; 0.15 ¢m of blotter paper, 0.64 cm of Al. B4C tank: 0.16 cm of Al; 2.7 em of B4C (» =1.9 g/cm3),' 0.16 cm of Al. NaF tanks 1 and 2: 0.48 c¢m of Fe; 2.54 em of NaF (p = 0.95 g/cma); 0.48 cm of Fe. NaF tank 4: 0.48 c¢m of Fe; 4.44 ¢m of NaF (o =0.96 g/cm3); 0.48 cm of Fe. (C)Value assumed for these calculations. Actually, this region varies in thickness according to the design power of the reactor. (d)The gamma- and neutron-shield layers are dependent upon the design conditions. 119 ANP QUARTERLY PROGRESS REPORY region will be chosen according to design require- Gamma doses to be expected at various from the designed shield have been determined in the Lid Tank Faocility for lead thicknesses of 4.5, 6.0, and 7.5 in. on the basis of measurements behind configuration 62 and siniilar configurations (61, 65, 73, and 55R). (The cor- rection factors applied to the Lid Tank data are described in detail in ref. 8.) The gamma doses obtained are given in Fig, 12.3. It will be noted that three methods were used to calculate the dose for a shield containing a 4.5-in. lead layer. The results from all three methods agree closely, and therefore there is some assurance that the In Fig. 12.4, the gamma doses at various distances from the shield are given as a function of lead thickness in the gamma-shield region. ments. distances curve presented is realistic, A mean curve of the fast-neutron data taken out to 110 cm behind configurations 39 and 75 was chosen as the basis for a neutron dose curve corrected to the designed shield (Fig. 12.5). Since the limitations of the dosimeter prevent measure- ments of the fast-neutron dose at distances greater TORe ort of the 1953 Summer Shielding Session, ORNL-1575 (o be published). 5 ORNL-LR-DWG 120 L | ~ ~ ! 1 ~ s s oo msmeerceeesmen e m o s ~ . e e - e N - . 5 - 4500 OF LEAD < 4 VETHOD B S 0 METHOD € | b i o . . S 6.0in. OF LEAD # ~_ S 5 e . GAMMA-RAY DOSE {mr/hr) ~ - ~ 7.5 in. OF LEAD *’\“\ -2 Lo 10 100 1o 120 130 140 150 160 170 2, DISTANCE FROM SOURCE (cm) Fig. 12,3, Lid Tank Gammeo-Ray Dose Curves Corrected to Designed Reflector-Moderated Recctor Shield for 4.5, 6,0, and 7.5 in. of Lead. 120 than 110 c¢m, the mean curve was extrapolated out to 140 em by using the trend indicated by thermal- neutron measurements behind configurations 39, 46, 47, and 51. Shield Weight Calcelation The data shown in Figs. 12.3 and 12.5 can be used to determine a specific shield weight by applying the following equation:'° 2 S R b o x LT S 1 LY T YrRMr | 7 T T R SRMR Re h 2 ORNL-LR-DWG 124 W TR T T ! T ] NN T e e Yy e — 5 VO \‘ \ VoA \\ \\\ i I 2 \T——\—\ \ \\ — L \ \ \ DASHED LINES INDICATE \ EXTRAPOLATIONS GAMMA-RAY DOSE (mr/hr} > | CONFIGURATIONS 16,17, AND 18 AND \O z=150 OPTIMIZATION EXPERIMENTS — w2 b L1 L LEAD THICKNESS (in.} Fia. 12.4. Lid Tonk Gamma-Ray Dose Curves Corrected to Designed Reflector-Moderated Reactor Shield with Various Thicknesses of Lead. where Lt = Lid Tank dose, Do g = dose desired at distance x from shielded reflector-moderated reactor, x = distance from center position, S, 1 = Lid Tank surface source strength =9.04 x 10% watts/em?® {a constant, since the data were always normalized to a power of 6 wotts; the self-absorption factor is 0.6, and the source areg is 3970 cn12), Samp = €quivalent reflector-moderated surface source strength, Re= shield outer radius, R = core radius, b = Hurwitz correction = 0.5 + 200/a)? [(z/A) + 11, reactor to crew reactor A = relaxation length for use with Hurwitz correction =7.5 em for both neutrons and gamma rays, ' ° a = radius of Lid Tank source (cm), z=R¢ — Re. For the calculation given below, the following conditions were assumed: 12 in. Beryllium reflector thickness Reactor power 50 megowatts x 50 DRMR 10 rem/hr Gammas ¥ of total dose Neutrons V4 of total dose Re 2 in. (23 cm) (13.2 % 10—5 waffs/cmz) SRMR (calculated for SaFS 9-in. core radius) X (5 x 107 watts) . A solution to the equation was obtained by using successive approximations and the fast-neutron and gamma dose curves presenfed in Figs. 12.3, 12.4, and 12,5, ¥ z is taken as 105 cm, the Lid Tank total dose, D, ., is D ¢+ = 5.87 x 10~4 mrep/hr Then D, ;(neutrons) = 0.25(5.87 x 10-4) rem/hr = 1,47 x 1072 mrep/hr PERIOD ENDING MARCH 10, 1954 ORNL-LR-DWG 122 8 FAST-NEUTRON DOSIMETER DATA (BASED ON CONFIGURATIONS - i 39 AND 75) @ NORMALIZED THERMAL-NEUTRON DATA (BASED ON CONFIGURATIONS - 46,47 AND 51) ’ FAST-NEUTRON DCSE (mrep/hr) 2 - FAST-NEUTRON CURVE CORRECTED 4 cm FOR LOW DENSITY OF HEAT _ EXCHAN 5 2 . 10‘4 ............. [ ........................................................... ,..J._,* 70 30 20 100 110 {20 130 140 Z, DISTANGE FROM SOURCE (cm) Fig. 12.5. Lid Tank Fast-Neutron Dose Curves Corrected to Designed Reflector-Moderated Reactor Shield. and DLT(gummos) = 0.75(5.87 x 10~ rem/hr = 4,40 x 10" me/hr . From Fig. 12.5, it can be seen thot the corrected Lid Tank neutron dose for z = 105 is 2 x 10-2 mrep/hr. Since 2 > 1,47, the approximate value of z is determined by interpolation to be 107 cm. Thus Ry = 107 cm + 23 cm = 130 em The gamma shield necessary con be read directly from Fig. 12.4 as approximately 5.3 in, of lead. The weights of the neutron and gamma shields can be computed after their distonces from the center of the reactor are determined. The neces- sary dimensions are presented in Table 12.1. For a 50-megawatt reactor with a 9-in. core radius, a 1.6-in.-thick heat exchanger seems reasonable.1? The inner radius of the lead laver is approximately 25.3 in., and its weight is approximately 21,600 121 ANP QUARTERLY PROGRESS REPORT pounds. The inner radius of the reutron shield is 30.6 in., and its outer radius is 51.3 in., which gives a neutron shield weight of 15,900 pounds. Thus the total basic reactor shield weight is 37,500 pounds. Approximate weights of the reactor core, heat exchanger, and pressure vessel components may be determined by a volume-density calculation, (Tables presenting such data appear in refs, 10 and 11.) If the weights of these components are added to the basic shield, the basic designed reactor and shield assembly is found to weigh 44,500 pounds. This weight does not include the weight of the crew shield, which is necessary for reducing the 10 rem/hr dose at 50 ft to acceptable tolerances within the crew compartment. There are problems such as the emission of short-lived fission-product decay gammas in the intermediate heat exchanger and the leckage of radiation through passages in the shield that may require small further additions to the shield weight. Such refinements are not treated in this calculation because the Lid Tank experiments do not coniribute to their resolution. REMOVAL CROSS SECTIONS An interesting check on removal-cross-section measurements has been provided by an experiment in which solid B,0; was used in the usual large- slob geometry. A value of 4.4 1+ 0.14 barns was obtained in this experiment for the removal cross section of B,0;, which is to be compared with the ”A. P. Fraas and C, B. Miils, A Reflectfor-Moderared Circulating Fuel Reactor for an Aircraft Power Plant, ORNL CF-53-3-210 (Mar. 27, 1953). 122 value of 4.56 barns calculated by using the previ- ously published figures of 0.87 born for boron and 0.94 barn for oxygen. INSTRUMENTATION Fxperience in operating neutron dosimeters with flowing ethylene has shown that the temperature must be controlled more carefully than has been possible heretofore, and therefore provisions are being made to maintain the temperature of the water in the Lid Tank at some constant value. It is anticipated that the constant woter temperature will also improve the accuracy of measurements with other instruments and will eliminate uncertainties arising from the expansion or change in solubility of shielding materials such as B,0,. In measuring gamma doses of less than 0.1 mr/hr, the sensitivity of the Llid Tank scintillation counters is inadequate when an electrometer circuit is used. Amplificotion of individual pulses followed by an integration of the resulting pulse-height spectrum increases the useful sensitivity of the instrument by a factor of at least 100 and has been extensively used in the Lid Tank and elsewhere. Difficulties encountered in calibration when using artificial sources in agir have been avoided by normalizing the data taken in the Lid Tank to measurements circuit, made by using the electrometer A more serious uncertainty has recently been uncovered, namely, a systematic difference in the obhserved relaxation lengths in water following certain shield configurations when the two tech- niques are employed. The difference is apparently caused by a change in the energy sensitivity of the instrument when the pulse integrator is being used. This effect is being studied further. PERIOD ENDING MARCH 10, 1954 13. BULK SHIELDING FACILITY R, G. Cochran G. McC. Estabrook! J. D. Flynn M. P. Haydon? K. M. Henry H. E. Hungerford E. B. Johnson Physics Division The work at the Bulk Shielding Facility during this quarter was devoted to the measurement of mockups of two bulk shields for the GE-ANP program, The first shield tested (BSF Exp. 18) was a mockup of a duct system and shield to be used at the G-E ldaho reactor test facility., The second mockup consisted of two sections of the reactor shield of the R-1 ANP divided shield. The first of these sections was the rear or oft section (BSF Exp. 19), and the other was the front or forward section {BSF Exp. 20). The minimum reactor power was used for all measurements on these sections so as not to appreciably activate the iron in the shield, since this mockup is scheduled for further testing at the Tower Shielding Facility. G-E TEST FIXTURE DUCT EXPERIMENT H. E. Hurgerford Physics Division The G-E ldaho facility duct mockup (test fixture duct system) consists of two, long, 10-in.-ID steel ducts, a cubical air void or plenum chamber, and two shields. The experimental setup is shown in Fig. 13.1. The reactor was located along one side of the plenum chamber, and one of the ducts emerged from the plenum chamber and wound out through the two shields. special offset reactor with @ 5 by 7 fuel-element loading was used. This loading permitted as much fuel as possible to be located near the plenum chamber, which was separated from the reactor by 3 in. of water. The duct (upper duct) feading from the plenum chamber contained three right-angle bends, one on the horizontal plane and the other two at an angle of 60 deg with the horizontal plane. An auxiliary duct (lower duct) is shown below the upper duct in Fig. 13.1, One of the shields, which consisted of three 5 x 5% ft lead slabs totaling a 2 thickness of 7.0 in. and was encased in stee! slabs For this experiment, a ‘Nee McCammeon. 2F’art time, 1.75 in. in thickness, was located directly between the reactor and the duct system. The other shield was located directly in front of the plenum chamber at the emergence of the upper duct and consisted of iron slabs totaling a thickness of 11.75 inches. The purpose of the experiment was to test some design features of the shield of the Idaho facility. The mockup represented only o portion of the actual duct and shield system. Dose measurements were made along the various traverse lines indicated in Figs. 13.1 and 13.2. Figure 13.2, which is an elevation of the mockup locking west, clearly shows the position of the upper and lower ducts, as well as a theoretical outline of the outer surface of the shield. Points A through P indicate the location of the corres- ponding traverse lines AA through PP, Fast-neutron measurements made along troverse lines AA through PP are shown in Fig, 13.3. The peaks of all curves, except the DD traverse, occurred at approximately 15 ¢m west of the pool center line and corresponded to the final, long, straight section of the upper duct, The corresponding thermal-neutron measurements, shown in Fig. 13.4, represent o more exfensive survey than that made for fast neutrons. The thermal-neutron peaks of the traverses AA through Il showed systematic variation from a location about 20 cm west of the center line for traverses near the upper duct to locations nearer the center line at larger distances from the duct. The gamma-ray meosurements made along the traverse lines AA through PP are shown in Fig. 13.5. These data are of special interest because of the double peak shown on the AA traverse. The larger peck was located 50 ¢m east of the center line, and the smaller peak appeared at about the center line. In succeeding traverses at larger distances from the mockup, the smaller peak dis- appeared and the larger peak shifted nearer to the center line. The larger peck was apparently due to radiation lecking out from between the two shields, and the smaller peak wos due to capture gomma 123 ANP QUARTERLY PROGRESS REPORT LR-DWG 22A b 30,0 == 180 — e QOO e 3,0 % PLENUM . REACTOR FRAME \ CHAMBER Lo 30.0 s~ 175 {RON SHIELD 7 ;- 7.0 LEAD SHIELD \ @ R 28.0 . SUPPORTING STRUCTURES — 1 | / | ~COORDINATE 7 ’ 2.0 ORIGIN - ) | SOUTH i UPPER DUCT - 75 X IR 11.t 5 / ON SHIELD 07500 X ‘ Z REGION OF T J ‘ + EAST-WEST TRAVERSE LOWER DUGT NORTH LINES 44 THROUGH 2P o I REACTOR AND POOL CENTER LINE ~ ! X R A \ 4 ALL DIMENSIONS ARE IN INCHES. EAST WEST NORTH—SOUTH TRAVERSE LINES — Fig. 13.1. Test Fixture Duct. Plan view of duct and shield systems as installed in the Swimming Pool; important dimensions and location of traverse lines indicated, rays from thermal neutrons captured in iron and water near the thermal-neuvtron peak. Results of measurements at the end of the duct taken along traverse lines QQ and RR show a large thermal-neutron peak, a small fast-neutron peak, and no gamma-ray peak (Fig. 13.6). These data indicate that, for the most part, only thermal neutrons were scattered down the duct. The intensity of thermal neutrons ot the end of the duct was about 1% of their intensity at the entrance of the final, long, straight section of the duct. In addition, the plenum chamber was flooded with water and the AA traverse measurements were repeated. The measurements taken are compared with the previous measurements in Fig. 13.7. The contribution of the plenum chamber to the dose along line AA is clearly indicated. The fast- neutron dose was diminished by only about 12% by flooding the plenum chamber, while the thermal- 124 neutron flux was reduced by a factor of 5. The large gamma-ray peck disappeared with the flooding, and the small peak was reduced in intensity by a factor of 2. These dats indicate that about 80% of the thermal neutrons in the duct system originated in {or passed through) the plenum chamber. The results of these measurements agree quite well with calculated estimates for the design. The data are reported in greater detail in ref. 3. GE-ANP SHIELD MOCKUP EXPERIMENTS H. E. Hungerford Physics Division After removal of the duct mockup described above, the rear {(or aft) section of the GE-ANP reactor shield was installed in the pool for testing. 3H. E. Hungerford, The Test Fixture Duct Experiment at the Bulk Shielding Focility, ORNL CF-54-2.94 (to be issuad). \ LEAD AND 90.0 IRON SHIELD- \ 204 \ F OUTLINE OF SHIELD PERIOD ENDING MARCH 10, 1954 LR-DWG 23A ALL DIMENSIONS ARE (N INCHES POSITIONS CF TRAVERSE LINES ARE SHOWN BY POINTS 4 THROUGH 5 ----- up 6‘ T = NORTH ’ /g/ _UPPER DUCT CENTER LINE J \ » \\ 5 \ -FUPPER 30.3 \ 4+ | DUCT ) Voo + 4 M Vo1 | 1A T i1 _—COQRDINATE ORIGIN \ !l REACTOR CENTER LINE | ¥ e A e TEALY / 1L A { ’ [ s e -—0"'-"—9—"—"— mAAmmS TS mmssmsaooo e ST e PLENUM 71 || i \0 F CHAMBER AT LOWER DUCT GENTER LINE - 3 L. | BSR CORE — [ | - LOWER DUC - DuCT 385 41.5 ‘ / \ TP SUPPORTING STRUCTURES | pool . . - CONCRETE BLOCKS ' - 7 t FLOOR” N ¢ 8. v v T Fig. 13.2. Test Fixture Duct. lines indicated. Figure 13.8 shows a plan view of this section of the shield, together with the location of the various traverse made. lines along which measurements were The .main feature of this section is the annular air duct between the inner and outer water shields. Also shown in Fig. 13.8 is a small air void and 3 in, of water located between the reacter and the shield. Dose measurements were taken along traverses north and east of the shield. The results were similar to those taken on the front section (see below) and are not included in this summary. At the completion of the measurements on the rear section, the mockup was removed and the forward or front section was installed in its place. Elevation looking west; important dimensions and location of traverse This section was of the same diameter as the rear section but was thicker., It also had an air void and 3 in. of water between it and the reactor. |In addition, the annular air void contained two extra bends. A lead and steel shadow shield was located within the inner portion of the shield, which also contained water borated to 1.12 wt % boron. To complete the mockup, a shield made of lead, steel, and borated water was placed on the east side of the reactor, A plan view of the setup and the locations of the various traverse lines along which measurements were made are shown in Fig. 13.9, Fast-neutron measurements taken along east-west traverse lines on the north side of the shield are shown in Fig. 13.10, and the corresponding thermal- 125 ANP QUARTERLY PROGRESS REPORT LR - DWG 24A FAST — NEUTRON DOSE (mrep/hr/watt} 60 40 20 0 20 40 60 60 40 20 O 20 40 60 DISTANCE FROM PCOOL CENTER LINE (cm) Fig. 13.3. Fast-Neutron Dose Measurements Along Yarious East-West Traverse Lines. 126 L] THERMAL - NEUTRON FLUX (nvy, /watt) &0 40 20 o 40 680 40 GISTANCE FROM POOL CENTER UNE (om) Fig. 13.4. Thermol-Neutron Flux Measurements Along Various East-West Traverse Lines. LR - DW!! 254 rS6L ‘Ol HOBNVW ONIGN3 gOniad gcl {r/hr/watt) GAMMA—RAY DOSE 1073 3073 1078 100 80 €0 40 WEST‘ 20 0 20 40 60 80 {100 80 60 40 20 0 20 40 80 80 00 BO 60 40 20 DISTANCE FROM CENTER LINE ALONG TRAVERSE LINE {cm) Fig. 13.5. Gamma-Ray Dose Measurements Along Yarious East-West Traverse Lines, 0 LR—DWG 264 &0 eo 400 JHOdIY SSAUO0Ud ATHINVND dNV PERIOD ENDING MARCH 10, 1954 THERMAL-NEUTRON /- - FLUX ON RR - - — GAMMA - RAY DOSE ON QQ —-.. (mrep/hr/watt) OR GAMMA-RAY DOSE {r/hr/wati THERMAL NEUTRON FLUX (nvy,/wait) — FAST - NEUTRON DOSE ON RR FAST-NEUTRON DCS » FAST~-NEUTRON DOS5SE ON QQ ..... e | NORTH SOUTH 50 50 40 30 20 10 O 10 20 30 40 50 80 DISTANGE FROM UPPER DUCT EAST-WEST CENTER LINE (cm) Fig. 13.6. Dose Measurements at End of Duct Along North-South Traverses QQ and RR, 129 ANP QUARTERLY PROGRESS REPORT LR—DWG 29A -1 0 T T T " oI 10 : THERMAL —NEUTRON FLUX, : NOT FLOODED j THERMAL —NEUTRON FLUX, | “FLOODED | i | } . o | FAST—NEUTRON DOSE, FLOODED _ 10~ -3 L S i I | 2 | GAMMA—RAY DOSE,. ! NOT FLOODED \\/ {mrep/ hr/watt) CR GAMMA-RAY DOSE (r/hr/walt) N AN\ _. . FAST—NEUTRON - k—/DOSE, NOT — L NNFL o e i d B 2 A FLOODED - \‘\rA;,,, GAMMA-RAY DOSE \tQ 1 \ NN ol L /_'______ G, _FLOODED 1 ' | ! t THERMAL MEUTRON FLUX (nvm/wafl) FAST —NEUTRON DOSE 8] | i i e ] | oo \ 5 S f-.‘;\.- 107° | et .._.._.__7____74',,,1 ,,,,, ) g . ; o L 10"3 EAST e—1--romees - WEST 1078 e ’ ‘ e j i 10~ t40 120 100 80 60 40 20 0 20 40 60 80 {00 120 140 DISTANCE FROM POOCI. CENTER LINE {(cm} Fig. 13.7. Comparison of Dose Measurements on Traverse Line AA, With ond Without Plenum Chamber Flooded with Water, 130 PERIOD ENDING MARCH 10, 1954 gk LR-DWG-28A PLENUM CHAMBER BSR WEST 62.1 ——~——-——.-1‘ 3.0 == i 89.4 NORTH 1% SHIELD CENTER LINE EAST-WEST TRAVERSE LINES EAST - S O T I T A, Q 20 Q ™M 4 R-12,R-6 ,:\\ R-7 “N\DETECTOR 10 r-g AGAINST SHIELD aeo NORTH-SOUTH 20 TRAVERSE LINES 30 R-10 40 N NOTE: ALL DIMENSIONS DISTANCE FROM IN INCHES. EAST FACE (cm) DISTANCE FROM NORTH FACE {cm) Fig. 13.8. Plan Yiew of Rear Section of GE-ANP R-1 Reactor Shield. Important dimensions and tra- verse lines indicated. 131 ANP QUARTERLY PROGRESS REPOHRT LR“E!! -30A -~ 55.5 - - P REACTOR EAST-WEST e 455 T CENTER LINE 409 ——— \ T —— / 3 RSP : T oy Y e \ S [ : - - ANNULAR 3 @ AIR QUCT ¢ J =z PLENUM YW 3 CHAMBER P T 20 R o 1k p- ! % — o o i \ % k e o o * . 1 [ ! | & e woouw ow o | - g%fi? S N T B e i r ¥ x ol = S | \SHIELD Y o _ CENTER LINE A5 206 Pb SHIELDS i o % ] 5 o e 8D % i Z W e Ko oo e ‘ | N % -l : L W 1 -.,' q 2 / Lid / / = 1.5 STEEI_/ 1.5 Pb i | ¥ o 19 1Q 19 Fo13,7-6 — s F-7 S DETECTOR F-8 . _ 'Y AGAINST SHIELD fo | NORTH-SCUTH 20 I TRAVERSE LINES 30 F1Q — — 75 NOTE: ALL DIMENSIONS IN INCHES. Fig. 13.9. Plan Yiew of Forward Section of GE-ANP R.1 Recactor Shield. Shadow shield, side shield, and position of traverse lines indicated. neutron measurements are presented in Fig. 13.11, In each figure, peaks occurring at points 45 cm from either side of the shield center line coincided with the end of the air-duct region and “‘washed out’’ with distance from the shield. Later, the air void was flooded and some of the measurements The peaks 45 cm from the center line vanished, as was to be expected. Flooding of the air void also reduced the thermal- and fast- were repeated. 132 neutron intensities emerging at the center line by a factor of 4. The gamma-ray measurements north of the shield (not shown) indicate that the dose just outside the shield at the center line was 3.0 x 107 ¢/hr/watt; with the air void flooded, the dose dropped to 1.5 x 1074 ¢/hr/watt. The effect of the exiro thickness of the forward section was to reduce the fast- ond thermal-neutron intensities outside the forward section by a factor of 1000 under the dose outside the corresponding In the case of gamma rays, thereduction factor was nearly 10,000 because of the added shadow shield and boration of the water of the inner shield, portion of the rear section. The measurements taken along various north-south traverse lines on the east side of this section are shown in Fig. 13.12 (fast neutrons),Fig. 13,13 (ther- ma! neutrons), and Fig., 13.14 (gamma rays), ln all cases, a slight peaking effect was observed in the region 115 to 120 c¢m south of the north face of the shield. extended southward as far as the reactor east- west center line, which was located 186.2 cm south of the north face. The south face of the shield proper was located about 140 cm south of the north When the air duct was flooded, the neutron peaks were eliminated, but curiously the gamma-ray traverse still showed a slight peak. It is to be emphasized that this is a very pre- liminary report on the GE-ANP reactor shield. The values reported here are subject to correction factors in reactor power. A more complete report on this shield will be prepared. In some cases, the measurements were face. ORNL-LER-DWG 123 FAST-NEUTRON COSE imrep/hr/watt) 3 2 ,,,,,,,,,,, $078 frmmo g e b N g e e T U booo - >t i F-4, FLOODED =T |oeee. e e L to™’ 0 20 40 80 80 100 120 DISTANCE EAST OF SHIELD CENTER LINE (cri) Fig. 13.10. Fast-Neutron Measurements North of Front Face. PERIOD ENDING MARCH 10, 1954 1wy THERMAL-NEUTRON FLUX (nvy, /watt) 4 S a 20 40 50 30 100 120 DISTANCE EAST OF SHIELD CENTER LINE (cm) Fig. 13.11. Thermai-Neutron Measurements North of Front Section, 4 ORNL-LR-BWG 125 10° T 2 1o 5 5 = S E S~ o E 2 o 240 — = (=] T s o ! T ooz < ' 107% 5 2 10778 40 80} 120 160 200 240 BISTAMCE SOUTH OF SHIELD NORTH FACE (em) Fig. 13.12, Fast-Neutron Measurements East of Front Section, 133 ANP QGUARTERLY PROGRESS REPORT ORKL-LR-DWG 126 . ORNL-LR-DWG *27 e o 7' 5 i = > £ - 5 = . 3 i - v - [ ul ! = _ I i .. S 3 10 . . R | E ; a ' ‘ ! U i = F-&, FLOODED ool ‘ B ! ¥ ‘ : ‘ 210" 5 z T = + X P + e . bt N 3_11 : o — .._.._7;;“ 7 . - s . 37 ‘ ,,,,,, T ! = — : : R ; | { - | 0o ; . : T T | : e . e o ) 4 e —— 3 T | b . L. e _ Smubi ‘ i | | | ‘ od L] ot l Lol o 40 80 120 16C 200 240 Q 40 80 120 160 200 240 DISTANCE SQUTH OF SHIELD NORTH FACE (cm) DISTANCE SCUTH OF SHIELD NORTH FACE {(cm} Fig. 13.13. Thermal-Neutron Measurements East Fig. 13.14, Gammac-Ray Measurements East of of Front Section. Front Section. 134 PERIOD ENDING MARCH 10, 1954 14, TOWER SHIELDING FACILITY C. E. Clitford T.V. Blosser J. L, Hull L. 8. Holland F. N. Watson Physics Division The Tower Shielding Facility (Fig. 14.1) is nearly completed, and operation should begin during March. A complete set of instruments has been collected for the experimental program; some of the instruments were especially developed for use at this facility. Consideration is being given to using the Tower Shielding Facility in a biological program for establishing dose rates for pilots in a nuclear-powered aircraft, and therefore a study is under way of the radiation doses which can be achieved. FACILITY CONSTRUCTION The construction of the T ower Shielding Facility (at the 7700 Area) has been completed, and most of the work of the outside contractors has been ac- cepted. All the ground facilities {control building, roads, fences, grading, reactor pool, etc.) were ap- proved by the Laboratory on February 4. The steel structure, hoist house, and hoist installation have been accepted on the condition that some additional painting will be done by the contractor, The hoists, the last of which was delivered February 1, will require minor modifications. It became apparent during the course of testing that the hoists did not operate properly under no-load conditions; the lines became excessively slack. This condition will be corrected by the addition of weight to the floating sheave blocks., In ad- dition, a slack-line limit switch will be designed and installed which should prevent any domage ot the hoist drums because of a slack line. The mechanical components of the reactor have been tested. It is expected that final installation will have been completed in time for the Tower Shielding Facility to go critical by March 1. RADIATION DETECTION EQUIPMENT T.V. Blosser Physics Division During the past year, a complete set of instru- ments has been collected and checked out for making experimental dose and energy measurements at the Tower Shielding Facility. Some of the instruments are stondard and have been in general use at both the Lid Tank Facility and the Bulk Shielding Facility, Other instruments have been developed or modified to meet the requirements, The instruments used for specific measurements are described in the following paragraphs. Thermal-Neutron Flux 1. BF, Chamber. The instrument is a standard counter, except that the physical dimensions have been made the same as those of the fast-neutron dosimeter to ensure the same geometry for com- porison of measurements, 2. U?3% Fission Chamber. A standard counter was modified to give a flatter plateau on the pulse-height curve. 3. Flow-Type Foil Counter, A standard counter will be used. Fast-Neutron Dose 1. Flow-Type Neutron Dosimeter. A three-cavity dosimeter was developed from the Hurst absolute dosimeter.| It will be used for medium intensities, that is, approximately 0.1 to 20 rep/hr. 2. Multichamber Neutron Dosimeter, Three flow- type neutron dosimeters mounted in paralle! will be used approximately 0.03 to 5 rep/hr. 3. Absolute Dosimeter. A standardHurstabsolute chamber with a built in plutonium alpha source will be used for intercalibration of neutron sources and neutron chambers. 4, Phantom-Type Dosimeter. A standard Hurst dosimeter will be used in high fluxes, that is, approximately 10 rep/hr to limit of recording equipment, for low-intensity measurements, that s, Gamma-Ray Dose 1. lonization Chamber {50 cm?). instrument will be used. A standard ]G. S, Hurst and R. H. Ritchie, Radiology 60, No. 6, 864-868 (1953). 135 ANP QUARTERLY PROGRESS REPORT PHOTO-12198 Fig. 14.1, Tower Shielding Facility, February 12, 1954, 2. Anthrocene Crystal (1/2 in.). A standard instrument will be used for normal dose rates, that is, approximately 2 r/hr. 3. Anthrocene Crysial ('l/2 in.) with Lead Ab- sorption Wheel. A 1/z-in. anthracene crystal will be mounted in a lead box with a small circular opening in one side. A lead wheel, which varies in thickness from 0 to 0.7 in., can be rotated to cover the opening and thus show the effect of adding small thicknesses of lead to a shield. 4. Anthracene Crystal {1 in.). A standard instru- ment will be used for low dose raotes, that is, approximately 200 mr/hr. Other Gamma-Ray Detectois 1. Sodium lodide Crystal (1 x 1 in.). A standard counter will he used. 136 2. Cobalt-Wire Scanner. The scanner, which consists of a single cobalt wire that can be in- serted along the full length cf center fuel plate, was developed for determining flux distribution within each fuel element of the reactor. Upon withdrawal, the wire will be surveyed with a sodium iodide crystal, Isodose Plotter An isodose plotter was developed for surveying the crew compartment in the x,y, and z coordinafes. When the plotter is used with each of the other detection instruments and allied systems, the intensity curves at given positions can be plotted automatically. electronics ESTIMATE OF NEUTRON AND GAMMA RADIATION EXPECTED AT THE TOWER SHIELDING FACILITY C. E. Clifford Physics Division Estimotes of some of the neutron and goamma doses which can be achieved at the Tower Shielding Facility for use in o biological program have been made on the basis of measurements of radiation at the Lid Tank Facility and the Bulk Shielding Facility. The Bulk Shielding Facility measurements show a gamma dose of 16 r/hr/watt for a 17,5.cm-thick water shield and a neutron dose of 1 rep/hr/watt for an 18.4-cm-thick water shield. The Lid Tank Facility data show thata 17.5-em«thick lead~borated water shield (11.4 c¢m of lead) reduces the gamma dose to approximately one-twentieth of that for a pure-water shield of the same thickness. However, TABLE 14.1, PERIOD ENDING MARCH 10, 1954 the neutron dose is increased by a factor of 4.8. Thus the gomma dose at the outer surface of an approximately 18-cm-thick lead—boroted water shield (11.4 cm of lead) would be '%, or 0.8 r/hr/watt, and the neutron dose would be 4.8 rep/hr/watt. This gives a ratio of 6 rep to 1 r, If isotropic emission is assumed and the buildup factors are neglected, the distance from the lead— borated water shield surface at which a 10-rep dose would be received for @ maximum reactor power of 100 kw (area of reactor face = 3 f%; relaxation length of 1.5-Mev neutrons = 300 ft) is approximately 123 feet. At the same point, the gamma dose would be 2.0 r/hr. The distance from an 18-cm-thick pure-water shield at which a 10+ gamma dose would be received (gamma relaxation length in air = 1200 1) is 250 ft; the neutron dose at the same point is 0.33 rep/hr. These and other results are given in Table 14,1, ESTIMATE OF RADIATION DOSES AT TOWER SHIELDING FACILITY FOR A REACTOR POWER OF 100 kw NEUTRON DOSE GAMMA DOSE POINT OF MEASUREMENT (rep/ he/ 100 kw) (+/hr/ 100 kw) For an 18-cm-thick Pure-Water Shield At surface 100,000 1,600,000 123 ft from surfoce 2.08 41.0 250 ft from surface 0.33 10.0 600 ft from surface 0.018 1.0 For an 18+cm-thick Lead—Borated Water Shield (11.4 cm of Lead) At surface 480,000 80,000 123 #t from surface 10 2.0 228 ft from surface 2.0 0.57 294 £+ from surface 1.0 0.27 137 Part IV APPENDIX REPORT NO. CF-54-2-68 CF-54-2-69 CF-53-12-108 CF-53-12-145 CF-54-1-1 CF-54-1-14 CF-54-1-155 CF-54-2.36 CF-54-2-185 CF-53.12-76 CF-54-2-35 CF-54-2-137 CF-54-2-151 CF-53-12-13 CF-53-12-23 CF-53-12-60 CF-54-2-93 CF-54-2-94 ORNL-1616 CF-53-12.41 CF-53-12-112 CF-54-1-170 15. LIST OF REPORTS ISSUED DURING THE QUARTER TITLE OF REPORT l. Aircraft Reactor Experiment ARE Operating Procedures, Parts |, [, and 11} ARE Operating Proceduwres, Part 1Y H. Reflector-Moderated Reactor The Inhour Formula for a Circulating Fue! Nuclear Reactor with Slug Flow Letter on Critical Mass for Two Region Reactors (to C, N. Klahr, NDA) The Behavior of Certain Functions Related to the Inhour Formula of Circulating Fuel Reactors A Memo to A, P. Fraas Re: RMR-RSA Weights Intermediate Heat Exchanger Test Results Shield Designs for the Reflector-Moderated Reactor Influence of Nuclear Power Plant Design Parameters on Aircraft Gross Weight ill. Experimental Engineering Additional Purification of Fused Fluorides Preliminary Discussion of Model-T Component Test IV. Critical Experiments Combination Circulating and Stationary Fuel Reflector- Moderated Reactor An SCW Moderated and Refiected Circulating Fuel Reactor V. Shielding Critique of LLiH as a Neutron Shield Estimate of Neutron and Gamma Radiation Expected at the Tower Shielding Facility Some Estimates of Removal Cross Sections Based on the Continuum Theory of the Scattering of Neutrons frem Nuclei Some Neutron Measurements Around Air Ducts The Test Fixture Duct Experiment ot the Bulk Shielding Facility Lid Tonk Shielding Tests for the Reflector-Moderated Reactor Vi. Chemistry Abstract and Qutline of Paper on *'Fused Salts as Reactor Fuels”’ Analytical and Accountability Report on ARE Concentrate The Sodium-Hydrogen Systam AUTHOR(s) ¥W. B, Cottrell Jo La M'elfim W. K. Ergen C. 5. Burtnette J. Jonos (Lockheed) B‘ M- Wilmr H. J. Stumpf A. P, Fraas A, P. Fraas H. W. Sovage J. G. Gollagher D, Scott J. W. Noakes J. E. Faulkner C. E. Clifford F. H. Murray C. L.. Storrs J. M. Miller H. E. Hungerford W, R. Grimes G. J. Nessle Metal Hydrides DATE IS5UED 2-11.54 2.11.54 12-22.53 12.23.53 1-5-54 1-29-54 2-4-54 2-26-54 12-14.53 2.2.54 2-19-54 2.23.54 2-2-53 12.10-53 12.11.53 Te be issved To be issued 12-8-53 12-18-53 1-1.54 141 REPORT NO, CF-53-12.17¢9 CF-54-2-1 CF-54.2.37 CF-54-2-114 ORNL-1624 CF-53-712-42 CF.52-12-140 CF-53-12-15 CF-54.2-79 ORNL-1647 s TITLE OF REPORT AUTHOR({s) Vil. Heat Tronsfer and Physical Properties Preliminory Measurements of the Density and Viscosity of NaF~ZrF4-UF4 (62.5-12.5-25.0 mole %) Some Preliminary forced-Convection Heot Transfer Experiments in Pipes with Yelume-Heat-Sources Within the Fluids The Measurement of Fluid Velocity by Photography Techniques Heat Capacity of Fuel Composition No. 31 The Mature of the Flow of Ordinary Fluids in a Therinal Convection Harp Yill. Rodiation Domagse The Radiation-lnduced Corrosion of Beryllium Oxide in Sodium at 1500°F Questionnaire for LITR Fluaride Fuel Loop Experiment 1X. Miscellaneous Proposed ANP Program The Curtiss-Wright Reoctor Set Aircraft Muclear Propulsion Project Guarterly Progress Report for Psriod Ending December 10, 1953 S. 1. Cohen . T. N. Jones . Poppendiek G. Winn J. O. Bradfute W. £. Brundage W, Brundage W. Parkinson 0. Sisman R. C. Briant C. B. Mills W, B. Cothral! DATE iSSUED 12-22.53 To be issued To be issued 2-17-54 2-23.54 12-3-53 12-17.53 12-3.53 2.10.54 1-12.54