| CENTRAL RESEARCH LIBRAR | DOCUMENT COLLECTION - BT ' o~ Tt - - 4456 0353310 2 [0 IR fi_}"é/fi A == 4 Sy AIRCRAFT NUCLEAR PMOJECT QUARTERLY PROGRESS REPORT CENTRAL RESEARCH LIBRARY DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON If you wish someone else to see this document, send in name with document and the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY OPERATED BY CARBIDE AND CARBON CHEMICALS COMPANY A DIVISION OF UNION CARBIOE AND CARBON CORPORATION ucc| POST OFFICE BOX P OAK RIDGE. TENNESSEE SEceN SECURITY ENFORMATION ORNL-1513 This document consists of 200 pages Copy é;;—of 247 ecopires Series A Contract No W-7405-eng-26 ATRCRAFT NUCLEAR PROPULSION PROJECT QUARTERLY PROGRESS REPORT for Period Ending March 10, 1953 R. C. Briant, Director J. H. Buck, Associate Director A, J. Miller, Assistant Director Edited by W. B. Cottrell DATE ISSUED R 8 1853 0AK RIDGE NATIONAL LABORATORY . Operated by CARBIDE AND CARBON CHEMICALS COMPANY A Division of Union Carbide and Carbon Corporation Post O0ffice Box P 0ak Ridge, Tennessee MARTIN MARIETTA ENEAGY SYSTEMS LIBRARIES A - .. oo - g“x‘ V.W%M"" « 1k "‘g Fl + * + 117 4Tk« L ] ] %‘ R e i i e i T P Ag‘ s LY ~ - d - £ W xi"- - : " Wy i a4t T ey 3 ‘ PRE A TErt 2 ST 54 Yy 2 RRA AN T L T OO =1 O\ U P LI R PR UFPFOIAEEMSE=QDIITNPUION-oDIOR o= llaender PEOPAEOOEFFEEOIPELOEUDEOOCOSRUDFEAIIZIARIOEOOTO ZVUEHEIRETVEDD < \‘k- Adamson Affel Baldock Barton Bettis Billington Blankenship Blizard Bredig Briant Briggs Bruce Buck Callihan Cardwell Cathcart Center Cisar Clewett Clafford Cottrell Cowen Eister Emlet (Y-12) Ergen Fraas Gall Graham Grigorieff (Cj Grimes Household-,; Humes (K—;5) Kellholr;' Keim 4 . Kelleg ertes f/ . Kigf L:f%on ;7v1ngston pLyon ” o INTERNAL DISTRIBUTION 42, 43. 44, 45, 46, 47. 48. 49. 50 > Asultant) 0 0 M > O MmUY ORNL-1515 Progress W. I Manly L. g Mann W. . McDonald JL. Meem AN J. Miller Z. Morgan J. Murphy F. Poppendiek M. Reyling W. Savage D. Shipley S1sman P. Smith (consultant) H. Snell L. Steahly W. Stoughton D. Susano . A, Swartout E.N\{l. Taylor F.'ég Uffelman E. RVanArtsdalen F. C. §onderLage J. M. Wgde A. M. Wepberg J. C. Whi¥g E. P. Wign& (consultant) H. B. Willa\g G. C. Wallaa X\ J. C. Wilson W C. E. Winters § ANP Library Biology Library Central Files Health Physics Libr#y Metallurgy Library Reactor Experimental Engineering Library Technical Information Department (Y-12) Central Research Library 111 1lv 99-101, 102412, 114-12 122, 123-127. 128, 129, 130-131, 132, 133, 134-138, 139-160, 161-164. 165, 166. 167-174, 175. 176-183. 184-185, 186-188, 189, 190, 191-193, 194-197, 198. 199-200,. 201-202, 203, 204, 205-206. 207, 208, 209.210. 211, 212-215. Y SR Y EXTERNAL DISTRIBUTION Air Force Engineering Office, Oak Ridge Argonne National Laboratory (l copy to Kermit A Armed Forces Special Weapons Project (Sandia) Atomic Energy Commission, Washington Battelle Memorial Institute ookhaven National Laboratory Buyeau of Aeronautics (Grant) Bugkgu of Ships Cala ornia Research and Developme Ch1cag Patent Group Chief of Naval Research duPont d%Tpany 4 General Electric Company, ANP’ Hanford Operations Office USAF-Headquar% rs, Office Idaho Operations, Office f Iowa State College 4 dderson) "Company gf Assistant for Atomic Energy copy to Phillips Petroleum Co.) \ 4 Knolls Atomic Power Lgboratory Massachusetts I gtltute of Technology (Kaufmann) to A, Sllv/rst.eln) National Adv1sory Committee for Aeronautics, Washington New York Operations Office North Amefican Aviation, Inc, NuclearBDevelopment Associates (QDA) Patent Branch, Washington Rand Cdrporation (1 copy to V. G. nning) Savangah River Operations Office, Adjgusta Vifro Corporation of America W“étlnghouse Electric Corporation 216-232. 233-247, Wright Air Development Center 2 — = N = 1 copies to B, Beaman copy of Col. P. L. Hill copy to Lt. Col. M. J. Nielsen copies to Consolidated Vultee Aircraft Corporation copy to Pratt and Whitney Aarcraft Divasaon copy to Boeing Airplane Company copy to K. Campbell, Wright Aeronautical Corporation Technical Information Service, Oak Ridge, Tennessee vl Reports previously 1ssued 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 Period Period Period Period Period Period Period Period Period Period Period Period Period 3 ., 4 "{.: Endaing Endaing Endang Ending Ending Ending Ending Endaing Ending Endaing Endaing Endang Endang -~ e 1 3":'}3“‘ f“l’.«"" YR . PR ™ November 30, 1949 February 28, 1950 May 31, 1950 August 31, 1950 December 10, 1950 March 10, 1951 June 10, 1951 September 10, 1951 December 10, 1951 March 10, 1952 June 10, 1952 September 10, 1952 December 10, 1952 TABLE OF CONTENTS FOREWOBD L] * L ] L ] L] L - L] L ) L] . . . . . L - . L ] . L L ] . L ] L] PART I. REACTOR THEORY AND DESIGN IN'TRODUCT I ON AND SlJ.MMA-RY L] L] L] L] L] L ] L] . L] L] L ° * L] L] ] L ] ].c 2- CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT , . ., . Fluid Circult « & ¢ o « ¢ o ¢ o o « ¢ o o s « o o o Stress Analysis . ¢ o ¢ ¢ s ¢ ¢ 3 e e s e e 4 0 e Reactor « « v v ¢ ¢ ¢ ¢ o o o o o ¢ o« s o o o « o o Instrumentation « « o « o ¢ o s o o o o« o o « « « o Off-Gas System . ¢ ¢« ¢ ¢ ¢« « o« o ¢ o « o o o o o« o Reactor Control ¢« & & ¢« ¢« ¢ ¢ ¢ ¢ o ¢ ¢ « o o o o Electrical System « « o ¢« ¢ ¢ ¢« s o o ¢« o « o o o o EXPERIMENTAL REACTOR ENGINEERING . . . . . . « . . & Pumps for High-Temperature Liquids . . « « « « « & Centrifugal pump with combination packed and frozen for fluoraides . ¢« « « ¢« ¢« o ¢« ¢ o & . . Allis-Chalmers centrifugal pump for l1qu1d metals Laboratory-size pump with gas seal . . . « « . & ARE-s1ze sump pump « « « « o« « ¢ o « o o o o o o ARE centrifugal pump .« « ¢« ¢« « ¢ « ¢ ¢ o« o « o« o Electromagnetic pump .« « « « ¢ « o« o o o o s o = Rotating-Shaft and Valve-Stem-Seal Development . . Mechanical face seals . ¢« v & ¢« ¢« ¢ ¢ ¢« & ¢ « & & Combination packed and frozen seal with MoS, . Combination packed and frozen seal with graphite Graphite packed seal ., . . .« ¢« & ¢ ¢ ¢ « ¢« ¢ o« & Rotating-shaft seal test . ¢« « « ¢« ¢ ¢ ¢ ¢ & o« & Packing penetration tests « « o « o « o o o o o Frozen-sodium and frozen-lead seal for NaK . . . . ARE valve test .« ¢ o ¢ o« o « ¢ o « o o o o o o Heat Exchanger Systems .+ . ¢« ¢« 4 o ¢ o o ¢« ¢ « o & Sodium-to-air radiator tests « o+« o« « o o ¢ ¢ o Bifluid heat transfer loop .« « « ¢ ¢ ¢ ¢ ¢ o o & ARE Fuel Circuit Mockup « « ¢« o ¢ s ¢ ¢ ¢ o ¢ o« « & Vortexes and bubbles .« ¢« ¢« ¢« ¢« ¢ o ¢ o ¢ o o o & Operational instabilaities « + ¢« &+ ¢ ¢ ¢ o ¢ o o & Instrumentation « « « « ¢ « o o ¢ s ¢ s o ¢ s & o Pressure measurement . . . . « o v s v s 6 e » ARE leak detection i1ndicator tests . « « « o o« o seal e o o « o s . « s o n « o e o« o . o e e e .« . Vil [ L ] Slip rings for temperature measurement . o o o o o« o o o o Rotameter type of flowmeter « 4« o « o « « « o o o « o o o & Handling of Fluorides and Liquid Metals . . . & ¢« « &« o o o . Distazllatzon of NaK & o & & v v ¢ v v 4 v v v v 4 v o o o Flame tests for NaK vapor 1n helium « + « & v & ¢ & « o o« & 3 L ] BEACTOR PHYS ICS . Ll L] L] [ ] Ll [ ] . L] L] L] L] L] L] L] L] L] L] . . L] L] 4. REFLECTOR-MODERATED CIRCULATING-FUEL REACTOR . . . « & +« « « Stat1c Physics & v 4 v 4 v 4 4 b e e e e e e e e e e e e e Reactors with various reflector-moderators . . . + « o« o & Reactors with beryllium reflector-moderators . . « « « o & Summary tables and graphs . « v ¢ ¢« ¢ ¢« ¢ ¢ 4 e 0 4 e . Critical Experiments .« « ¢« o 4 ¢ ¢ o o « o o o o « o o o s First critical assembly . & ¢« ¢ ¢« 4 ¢« ¢ ¢ o o « o o o o s o Second critical assembly . . ¢« ¢« ¢« ¢ ¢« ¢ ¢ 4 0 0 e 0 o . Design Characteristics « « o+ o « o « o« o o s « o o o s o o o Factors affecting core diameter « . « « « o« « « o« « « « o« 4 Temperature, pressure, and StTeSS « o+ o« « o« o o « o o « o &« Reflector heating « + ¢« 4 & ¢ & 4 ¢ ¢ ¢ ¢ ¢ o o o o o o o o Shi1elding o o 4 4 4 ¢ ¢ ¢ ¢ ¢ o o o o ¢ o o o o ¢ o o o s o Power Plant Des1gn « ¢« ¢« &« ¢« 4 4 ¢ ¢ ¢ o ¢ « o o o o o o o « PART II. SHIELDING RESEARCH INTRODUCTION AND SUMMARY . . . & & & ¢« v ¢ ¢ o o o o ¢ « ¢ o « « o 4 5. LID TANK FACILITY [ ] [ ] [ ] . . . Ll L] - L] - ] L] * L ] L} * L} L] L] * . L ] Effective Fast-Neutron Removal Cross Sections « « « « « o o & Mockup of the Unit Shield of the Reflector-Moderated Reactor 6. BULK SHIELDING FACILITY . v 4 ¢ ¢ ¢ ¢ o o« o « « o o o ¢ o o o« s Air-Scattering Experiments . . ¢ ¢ ¢« ¢ ¢ 0 0 4 0 e e e e Irradiation of Animals &« ¢ ¢ o 4 & o ¢ o o o o o o « o s o o Neutron Spectroscopy for the Divided Shield . . . . . . « . . Gamma Spectroscopy for the Divided Shield . . . . . « « « . . Fission Energy and Power in the Bulk Shielding Reactor . . . 7 L] TOWER SHIE]-JDING FACILITY L] L] L ] - L] * - . e L] . - L] e L] L] L] - L] Tower Des 1gn L ] [ ] L ] L L ] [ ] L ] L ) L ] L ] L ] L ] L ] L ] L ] L ] [ ] L L ] * L ] L ] - a Construction Schedule « « ¢ ¢ ¢ % o « « o« o « « o s s o s o 8 L ] NUCLEAB thASI.IREMENTS L] -* L ] -* L ] L ] L ] L] - . * L] L ] L ] [ ] L ] L] L] 9 L] k] Fast-Neutron Scintillation Spectrometer . .+ « « « « « « « o o Measurements with 6-Mev Van de Graaff . . ¢« ¢« ¢« ¢« ¢+ ¢« ¢« « o & vill 36 36 36 36 37 39 41 41 44 46 51 53 53 36 61 64 66 69 74 79 87 89 89 89 91 91 91 92 92 92 97 97 97 101 101 101 PART II1., MATERIALS RESEARCH INTRODUCTION AND SUMMARY . . ¢ v & ¢ 4o ¢ v o ¢ « « « o o & 9. 10, CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS ., . . . « . . . Fuel Mixtures Containing UF, v ¢ v v v v v « v o LiF-ZrF, -UF, . . ¢ v v v v 6 v v e e e e 0 e KF-LiF- BeF -UF ¢« s v e e 4 e e e e e e e e Fuel M1xtures Contalnlng ThF « o e s 0 s s e e 4 e Fuel Mixtures Containing UCl s o s s = % s & s a NaCl-UCl, . & ¢ & 4 v v v v o v o o o o o s o o KC1-UCl, . . .. . .. s e e e & s e s e s e o Fuel Mlxtures Containing UF e s s 4 s e e e e e s Addition of Reducing Agents to ARE Fuel . . . . . . Coolant Development « ¢« « « « « « & o « « o « o o LiF-ZrF, ¢ ¢ v v v v v i e e e e e e e e e CsF-ZrF, . ¢« 4 ¢« v v v 0 v 0 v 0 e e vt e e e KF-LlF-ZrF4 s 5 @« * e & 8 8 & 4 e ® s = & 9+ o+ o4 . RbF-LiF-ZrF, . . ¢ ¢ ¢ v ¢ v ¢ v v 4 ¢ ¢ o 0 o o KF-LiF-BeF, . . . . . s 4 4 e 4 s e e e e s Production and Purlflcatlon of Fluoride Mixtures . Laboratory-scale fuel preparation « « « « « « « & Pilot-scale fuel purification « « « ¢ « « ¢ o + & Fluoride production facility ., . « ¢ ¢« « « « « . Hydrofluorination of Zr0Q,-NaF mixtures . . . . . CORROSION RESEARCH . . & ¢ ¢ ¢ ¢ ¢ ¢ « ¢ o o « o o « Fluoride Corrosion of Metals i1n Static, Seesaw, and Rotating Tests « ¢ ¢ ¢« ¢ ¢ o o ¢ o o o o o o« Crevice COrroS10n o« o « o « ¢ o o o o« o o o o « o High-temperature pretreatment of Inconel . . , . Effect of time of exposure of Inconel to fluoraide mixture with and without ZrH, added . . . . . Structural metal fluoride additives « « « 4+ « o Hydrogen fluoride additive . « « ¢ « o« o« « o & Fluoride corrosion i1n a rotating rig . . « « o+ o Static tests on Incoloy and Inconel in fluorides Fluoride Corrosion of Metals i1n Thermal Convection Loops Effect of fluorade batch s1ze « « ¢« « « & « « o & Fluoride pretreatment . « o« o o« o o o o o o o o o Hydride additives o+ + 4o ¢ ¢« ¢« ¢ ¢ o ¢ o & « o o « Metal addatives o« o o & ¢ ¢ ¢ ¢ ¢ ¢ o « o o o o o Temperature dependence . ¢ ¢« ¢« ¢ ¢« ¢ « ¢ o o o Crevice COTTOS10N « ¢« « « ¢ « « o o « o ¢ o « o o Inserted corrosion samples .« o « o « o+ & o o o & 23 105 107 107 107 107 107 108 109 109 109 110 112 112 112 112 112 113 113 113 114 115 115 117 118 118 118 118 119 120 121 121 121 121 122 123 123 123 124 124 1X 11. . —w- — Effect of exposure time . . . . . . . Nonuranium bearing mixtures . . . . . Liquid Metal Corrosion of Structural Metals . Seesaw tests of sodium-lead alloys . Static tests with lathium . . , . . . Incoloy 1n sodium, lithium, and lead Corrosion by lead i1n thermal convection loops « « « « « . & Corrosion of Ceramics by Fluorides and Liquid Metals ., ., . . Cermets in fluorides and liquid metals Static tests of the ARE-BeO blocks in Na and NaK . . . . . Seesaw tests of BeO in NaK . . . . . Convection loop tests of BeO in NaK . . . . . . . . . . . . SOIUblllty Of Beo in NaK « * e & e » Effect of Atmosphere on the Mass Transfer of N].Ckel in HydrOXIdeS . . . . " e . Fundamental Corrosion Research . . . . Identification of corrosion products from Preparation of special complex fluoraides Air oxidation of fuel mixtures . . . METALLURGY AND CERAMICS . . . . . « « . . Welding and Brazing Research . . . . . Cone-arc welding . +« « « ¢« « « o« .« . Fabrication of heat exchanger units , Brazing of copper to Inconel . . . . Evaluation Tests of Brazing Alloys . . Corrosion of brazing alloys by fluoraides Effect of brazing time on joint strength Effect of spacing on brazed joint strength L] [ ] [ ] L] * L] L] L ) . convection loops Strength of brazed joints with various base metals ., . . . Creep-Rupture Tests of Structural Metals Preoxidized Inconel in argon ., . . . Peened Inconel 1in argon « « « « . . . Effect of Inconel grain size on time to rupture . . . . .+ . Effect of environment on strain vs. time curves for Inconel Type 316 stainless steel tube burst tests Fabrication of Control and Safety Rods Control rods for the G-E R-1 Reactor Tower shielding facility safety rods Columbium Research .« + « v & v o o « & Gaseous reactions ¢+ s 4 & e e 4 s Ox1dation & « « ¢ o o o « o o « o o 4 Fabrication of Metals . . . . « ¢« « « . High-conductivity metals for radiator fins in argon s e e SOlld‘phase bondlng * ¥ ¢ 8 e ® & ¢ 8 e &6 4 4 e+ & s e e e = 124 124 125 125 126 126 128 129 129 131 132 132 133 136 136 136 137 137 139 140 140 140 142 143 143 143 144 145 145 145 146 147 147 148 148 148 150 150 150 151 152 152 152 Extrusion of high-puraity tubing . . « « ¢« ¢« ¢ v ¢« ¢« ¢« ¢« ¢ « o & 152 Hot-Pressed Pump Seals . ¢ ¢« & ¢ ¢« ¢ ¢ ¢ ¢ ¢ ¢ ¢ o ¢ o ¢ « o o s 153 Ceramlcs € 8 @ e & © 4 e @& 4 8 & %8 & 4 8 & 8 8 6 8 e e s e e = s 154 Ceramlc Coatlngs fOI‘ Shleldlng s 8 & 8 & 8 & & & 8 8 4 & & =8 154 Development of cermet fuel elements « « ¢« ¢« « « o o ¢ o« o o o« o 154 Reduction of porosity of ARE beryllium oxide . « ¢ ¢ o ¢ & & & 155 12, HEAT TRANSFER AND PHYSICAL PROPERTIES . . . ¢ ¢« ¢ ¢ ¢ ¢ ¢ o« o« s o 157 Thermal Conductivity of Liquids + ¢ ¢ ¢ ¢« ¢ ¢« o « ¢ ¢ o o o « o « 157 Heat Capacity of Liquids . . . . ¢ s s s e e s e e e 157 Viscosity and Density of Alkala Hydrox1des c t s e a4 e et e 158 Vapor Pressures of Fluorides .« o « ¢ ¢ « o « o o ¢ o o o « « « & 159 Prandtl Moduli of Various Materials . « ¢« ¢« ¢ ¢« ¢« ¢ ¢ ¢ « ¢ ¢ o« « 160 Specific Reactor Heat Transfer Problems . . . . « ¢« ¢« ¢« ¢« « « ¢« & 160 Turbulent Convection 1n Annuli & & ¢ 4+ ¢ ¢« ¢« v 4 ¢ ¢ o ¢ o o o o 160 Clrculatlng“Fuel Heat TranSfer *+ 8 # 4 8 8 8 e 8 s 8 5 8 8 s s » 162 High Temperature Reactor Coolant Studies + « + « o ¢ ¢ o ¢ o o & 163 13 * BADIATION DAMGE L] * * L] L ] * - . * L ] L ] * L ] . L ] L ] a L ] L ] . L ] L L] . L ] 1 65 Irradiation of Fused Materials . & « « ¢ o o o o o s o o o s o & 165 In-Reactor Circulating Loops « « « & « ¢ o o o o o o s s o o o & 167 Creep Under Irradiation . « « ¢ o o v ¢ o o o o o o s o o o o s 167 14. ANALYTICAL STUDIES OF REACTOR MATERIALS . e e e e e e e e e e 171 Chemical Analysis of Reactor Fuels and Contamlnants e s 8 s e e e 171 ZITCONIUM o o s o o o « o o o o o o s s s s s s s o o s o s o s 172 Reduction products « « o « o « o o o ¢ o s o o o o s 2 o o o o 172 Ox1des & v ¢ & o 4 ¢« o 4 e s e s e s e e s e e s e e e e e e 173 Hydrogen fluoride . . ¢« & ¢ ¢ ¢ ¢ 4 ¢« ¢ ¢ o o o o o s o s s o = 173 Corrosion products « « « o ¢ s o s o o o s o o o o o o 4 o o 175 Preparation of Uranium Tetrachloride . « &« « ¢ ¢« o« ¢ ¢ ¢ o « o & 176 Petrographic Examination of Fluorides . « « ¢« ¢« ¢« ¢« ¢ ¢« ¢« ¢« « « « 176 UCl, -NaCl system .+ & ¢« ¢« o o o o o o o 2 o o o o s o s o s o s 176 UC1,-KCl System o« « o « o o o o o o o o o s o « o o o o o o o & 176 X-Ray Daffraction Studies . « ¢« « ¢ ¢ o ¢ o ¢ ¢ s o s o s o o s 177 Service Chemical Analyses 4 4 ¢ ¢ o« ¢ o o o s « o s s o o s s o« & 177 PART IV, APPENDIXES 15. LIST OF REPORTS ISSUED DURING THE QUARTER . . . . . ¢« ¢« « ¢« « &« &« & 183 TECHNICAL ORGANIZATION CHART OF THE ANP PROJECT . . ¢ ¢ ¢ ¢ « « ¢ « & & 187 ANP PROJECT QUARTERLY PROGRESS REPORT FOREWORD This quarterly progress report of the Aircraft Nuclear Propulsion Project at OBRNL records the technical progress of the research on the circulating-fuel reactor and all other ANP research at the Laboratory under 1ts Contract W-7405-eng-26. The report 1s divided into three major parts I. Reactor Theory and Design, II. Shielding Research, and III. Materials Research. Each part has a separate introduction and summary. The ANP Project 1s comprised of about three hundred technical and scientific personnel engaged i1n many phases of research directed toward the nuclear propulsion of aircraft. (The Project organization chart 1s included as an appendix.) A considerable portion of this research 1s performed in support of the work of other organizations participating in the national ANP effort. How- ever, the bulk of the ANP research at OBNL 1s directed toward the development of a circulating-fuel type of reactor. The nucleus of the effort on circulating-fuel reactors 1s now centered upon the Aircraft Reactor Experiment - a 3-megawatt, high-temperature prototype of a circulating-fuel reactor for the propulsion of aircraft. This reactor experiment 1s now being assembled, 1ts current status 1s summarized in sec. 1. However, much supporting research on materials and problems peculiar to the ARE will be found in other sections of Part I and Part III of this report, along with the general design and materials research contained therein. A survey report on a reflector-moderated type of circulating-fuel reactor 1s contained 1n sec. 2. Since the feasibility of such a reactor was indicated by critical experiments completed during the previous quarter, considerable research, as well as design effort, has been devoted to problems associated therewith. The shielding research report, Part II, 1s devoted almost entirely to problems of aircraft shielding. e CABERE Rt Bidind g ay _— ;:4 ek L ‘,fi gi X y BT g e SRR AL W 4 geaent W% A '%II“ ¥ k3 - y - » Lfli T oy N + "3" A Pty ™ - s r Part | REACTOR THEORY AND DESIGN INTRODUCTION AND SUMMARY The Aircraft Reactor Experiment (sec. 1) 1s rapidly taking form as equipment 1s received and installed in the ARE Building. There were no significant changes 1n design or concept during the quarter, and the installation 1s proceeding according to schedule. The fuel system heat transfer loop and the fill and flush tanks are completely installed, and the sodium heat transfer loop 1s nearing completion. Installation of fuel and sodium piping 15 now 1n progress. The control and graphac panels have been completely instru- mented and checked. The formal presentation on the hazards involved in the ARE was made to the Reactor Safeguards Committee earlier in the quarter, and in view of the subsequent discussion of the Committee members, 1t 1s not expected that any modifi- cations of the experiment will be required. Valves, pumps, i1nstrumentation, and other components of the fluoraide- fuel and sodium-coolant systems are being developed and tested for the ARE (sec 2). Both centrifugal and electromagnetic pumps of ARE capacity have been operated, the latter will be employed i1n the by-pass loop for preliminary tests, with NaK, of the external fuel system. Impeller designs for the centrifugal pumps are con- ventional, but the pump seal, particu- larly -for fluoride fuels, remains a problem. The packed-frozen seals on fuel pumps have widely varying leakage rates, and adequate temperature control of the frozen zone has not been obtained. Both packing and face-seal materials are being examined 1n an effort to find materials with the desired service and life charac- teristics. Gas face seals for vertical- shaft pumps are being investigated. Considerable success has been experi- enced with frozen-sodium seals 1n sodium pumps, and frozen-lead seals are being tested for use with fluoraides. Frozen-sodium and frozen-lead seals have proved to be unsatisfactory for use on NaK pumps. The ARE bellows type of valve was tested at 1100 and 1300°F and found to be satisfactory, although excessive leakage occurred after operation at 1500°F. The 1initial mockup of the ARE fluid circuait revealed instabilities and gas entrain- ment that were elaiminated by mainor system alterations. The rotameters, level indicators, and pressure trans- mitters have operated satisfactorily for long periods of time under the conditions expected in the Aircraft Reactor Experiment. The reactor physics studies (sec. 3) were concerned with the statics of the reflector-moderated reactor and damping of power oscillations. The applicability of the multigroup method to the calculation of the reflector- moderated reactor was demonstrated by the good correlation obtained waith critical experiments. The damping of power oscillations 1n a reactor 1s demonstrated for a nonlinear case, with delayed neutrons included, by assuming (1) constant power extraction or (2) a special case of cooling with circulating fuel. The reflector-moderated reactor and an aircraft power plant assembly employing this reactor are described (sec. 4). In general, the reactor consists of lumped regions of fuel and moderator or reflector. An essent1al element i1n this reactor 1is the use of a thick beryllium reflector that not only contributes to neutron economy but also constitutes the farst layer of the aircraft shield. This reactor and the compact, spherical arrangement permit realization of what 1s practically a unit shield for a 200-megawatt reactor and heat exchanger combination that weighs about 80,000 pounds. Craitical experai- ments on a mockup of thais reactor ANP PROJECT QUARTFRLY PROGRESS REPORT assembly indicate a craitical mass of A preliminary weight estimate of the about 20 1lb and a total uranium in- complete power plant assembly, in- vestment of 100 lb when the compact cluding the turbojets, 1s about 115,000 external fuel system 1s 1included. pounds. PERIOD ENDING MARCH 10, 1953 1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT J. H Buck, Research Director’'s Division, E. S Bettis, ANP Division e The major effort in the ARE project during the quarter has been 1n 1instal- lation of equipment in Building 7503. The arrival of the heat exchangers permitted the i1initiation of a large amount of construction work that had been postponed for several weeks Valves began to arrive toward the end of the quarter, and work could proceed on the i1nstallation of the maain plumbing system. The installation of equipment will no longer be dependent upon the arrival of components. The installation work 1s not ahead of schedule, but thus far 1t appears that the completion goal of June 1953 will be realized No new technical crises have arisen 1n the project. Tests of valves, seals, pumps, etc. continue (sec. 2) and, to date, reveal no insurmountable difficulties. Seals have operated satisfactorily, but their performance has not been so good as might be wished. The same condition exists with respect to the valve tests, but the valves have been shown to be usable. Although detailing of installation drawings continues to be a problem, 1t has not yet actually retarded the job. There have been the usual modifications dictated by problems arising i1n the field, but they have been no more numerous or serious than would be expected 1n a system of such complexity as the ARE. FLUID CIRCUIT G. A. Craisty Engineering and Maintenance Division No significant design changes were made in the fluid circuit. However, since 1t appeared 1nadvisable for the long drain lines 1n the fuel circuat to stand full of stagnant fuel, two valves were added to permit draining An these long lines so that they w¥Il be empty during operation of the experiment, ot The heat disposal loops for the fuel circuit, that 1s, the fuel-to- helium heat exchangers, the helium-to- water heat exchangers, the thermal barriers, the helium ducts, and the helium blowers, are now completely installed. Figure 1.1 shows these loops 1nstalled 1n the pat. Three sections of the fuel pipe, together with annuli, thermocouples, and heaters (where possible), have been prefabricated and can be welded 1nto the system when the valves for these lines are ready. Figure 1.2 shows one of these sections of fuel pipe. All the fill and flush tanks have been put 1n the pit, and work has been started on installing heaters on these units. These tanks are shown in Fig. 1.3. The by-pass, which enables the system to be run without the reactor in the circuit, has been designed. The electromagnetic pump, which wall be i1nstalled 1n this by-pass for circulating liquid metal, has been tested by the Experimental Engineering Group (sec. 2). The surge tanks have been completed, and tests of the mockup of the hydraulic system are continuing (sec 2). The helium-to-water heat exchangers for the sodium coolant caircuit have been received and are being installed. The li1quid metal-to-helium heat exchangers are to be delivered the first week of March. STRESS ANALYSIS R. L Maxwell J. W. Walker Consultants, ANP Division The reactor by-pass circuit was checked to see that previous pre- stressing calculations would not be ROD - COOLING] DUCTING IHEUUM DU BARRIER MOTOR| ¢ (RAISE -LOWER)! HELIUM- TO-WATER g4 HEAT EXCHANGERS HELIUM BLOWER THERMAL BAF?RVERS £ FUEL TO-HELIUM L/ HEAT EXCHANGERS 140434 SSIYO0Hd ATHALYVNO LOAf0¥d dNV Fig. 1.1. Fuel System Heat Disposal Loops. Fig. 1.2, Fuel-Pipe Section. UNCLASSIFIED PHOTO 10852 d INSULATION 0N CLAM SHELL ELECTRIC HEATERS ‘0T HOYVW ONIGNI doI¥dd €561 P ‘UNCLASSIFIED H g PHOTO 10907 TAL Fig. 1.3. Fill and Flush Tanks. LYOdA SSAYI0Ud ATHALYYND LDAf0¥d ANV invalidated by 1ts 1ncorporation in the circuit A report on the complete analysis of the stresses 1s beaing prepared A design has been developed for 1ncorporating strain gages 1n the system at the anchor points to check the computed stresses These gages w1ll be 1nstalled prior to the pre- stressing, and the magnitude of the prestressing forces will be measured The stress relief as the temperature 1s raised will be measured until the temperature reaches about 200°F, above 200°F the gages will no longer be usable REACTOR The reactor 1s ready for operation, except for welding of the serpentine coils and final assembly. Figures 1 4, 1.5, and 1.6 show the "status of parts of the reactor All preliminary work on the pressure shell has been com- pleted The coi1ls have been received and sample coi1ls are now being test welded 1n ORNL shops It 1s planned to assemble the reactor i1nto a test circuit at the site and to circulate li1guid metal through the fuel coils at desi1gn temperature so that the reactor can be i1ndependently checked prior to 1ts i1nstallation Thais pretesting of the reactor core will also serve to clean the fuel passages with hot liguid metal prior to 1ts 1ncorporation 1nto the system, INSTRUMENTATION R. G. Affel, ANP Division The only real design modification in the i1nstrumentation 1nvolved a change 1n the leak detection circuit. The change was a simplification to minimize trouble without changing the principle of operation of the system. All the control room instruments have been mounted. The graphic panel has been 1nstalled and checked with a dummy setup of pneumatic switches that simulated valve operation. PERIOD ENDING MARCH 10, 1953 Pneumatic lines have been run to valves and sensing points 1n th'% pits, but a small amount (2%) of work remains to be done. Thermocouple designs have been completed and orders have been placed Several of the thermocouples needed to allow instal- lation work to proceed have been made by the Instrument Department All instruments and actuators have been labeled The enunciator panels have been 1nstalled and checked, except for connection to the sensing points. OFF-GAS SYSTEM There has been no change 1n the off-gas system, and installation in the building 1s practically completed. The final design for the shielding of the radiation monitors has not been completed, but this shielding will be quite simple and will be assembled 1n the field by using a stacking of lead bricks around the monitors REACTOR CONTROL The model fission chamber for use at high temperature has operated satisfactorily for several hundred hours at a temperature of 1400°F This test has proved the acceptability of the i1nsulator for this chamber. The live chamber, with a U?3?% sleeve, has not yet been tested. The control room relay panel has been completed. The wirang of the console and the control board instru- ments console has progressed as far as the i1nterconnection terminal blocks No interconnecting cables have been run between the amplifier cabinet, the relay cabinet control board, and the control actuator assembly. This work had just started at the end of the quarter. The mechanism for control rod actuation 1s being assembled 1n the crane bay area for testing. All parts for this assembly are complete and on hand i1n Building 7503, but no signi1ficant progress had been made 1n the assembly. 11 14 b | 235 in. 0D PHOTO |o!95| LY0dAd SSAYI0Yd ATHALYVNO IDAL0¥d dNV PERIOD ENDING MARCH 10, 1953 REFLECTOR COOLANT INLET SAFETY ROD HOLES) BOTTOM OF| PRESSURE SHELL: ITEMS NUMBERED 1 THROUGH 6 ARE 3 FUEL OUTLET LINES Fig. 1.5. Reactor Pressure Shell. 13 R ANP PROJECT QUARTERLY PROGRESS REPORT P&o 10899) WELD JOINT TO PRESSURE SHELL o I r— . ' FUEL INLET! ‘ HEADER / L. 1 REFLECTOR COOLANT| 14 ELECTRICAL SYSTEM ” The electrical work 1s proceeding satisfactorily. The large job of completing the heater circuit panels has been accomplished. Heaters are being installed on the fill and flush PERIOD ENDING MARCH 10, 1953 tanks, and leads have been run from the control panels to two of the faill and flush tanks. Both motor generator sets have been received and checked. It was found that these sets raised the noise level 1n the basement tp an untolerable level, consequently they w1ll be moved out of the building. 15 2. EXPERIMENTAL REACTOR ENGINEERING H. W. Savage, ANP Division The work of the past three months on the liquid circuits of the ARE has been devoted to minimizing corrosion to values acceptable for 1000 hr of continuous or 1ntermittent operation at elevated temperature, determining and applying a process of blending fluorides for the ARE that will yield a product with a quality equal to that of the mixture used i1n the most satis- factory corrosion tests, providing pumps for the circulation of fluorides and ei1ther sodium or sodium-potassium alloy, developing seals for the pumps that are reliable for 1000 hr of operation, providing other circuit components, and studying circuit modifications that will assure accepta- ble and operable fluid dynamics. This work was carried out cooperatively with the Metallurgy Division, the Materials Chemistry Division, and the ARE Section of the ANP Division, Both centrifugal and electro- magnetic pumps of ARE capacity ~ 50 to 100 gpm - have been tested. The conventional pump i1mpellers are satis- factory, but seals for mechanical pumps are still a problem, A 100-gpm, a-c, double-cell, electromagnetic pump has proved to be entirely satisfactory for pumping sodium-potassium alloy at 1500°F, and 1t 1s being tested for pumping sodium, Packings for horizontal-shaft centrifugal pumps have not yet been proved reliable. The most promising packing was graphite and/or MoS, retained with braid or graphite rings, or possibly by magnetic means. Leakage rates have varied widely, and friction demands, 1n some cases, have been excessive. Thermal control of the frozen zone, needed for low-friction operation and for remotely stopping and starting the pump, been achieved. has not yet O1l-lubricated, low-temperature, face-type, gas seals have been reliably operated for thousands of hours in high-temperature, centrifugal, sump- type pumps. An ARE-size sump pump, similarly sealed, 1s undergoing a water test, and there have been no major difficulties, to date. Frozen-sodium or frozen-lead seals for NaK have proved to be impractical, Frozen-sodium seals for ARE-s1ze sodium pumps (2 1/2-1n.-d1a shaft) are now being tested, and the difficulties experienced to date appear to have been due to thermal distortions, Frozen-lead seals have been used satisfactorily with fluorides, and in several tests with molten fluoride 1in contact with molten lead, there has been no mixing of the fluoride and the lead. High-temperature face seals for use with molten fluorides are being lnvestigated, available. Other components that have been operated successfully for long periods are high-temperature rotameters and high-temperature liquid-level 1nda- cators (both the rotameters and the liquid-level 1ndicators use variable- inductance pickups for signal gener- thin-walled (0.005 1in.) bellows 1n pressure stransmitters 1in contact with fluorides, a pneumatically- actuated, bellows-sealed, Stellite- seated valve, and the slip-ring and brush arrangements for removal of the thermocouple signals used for measuring temperatures 1n moving parts. A study of the parallel circuits and the pumping and expansion tank arrangements of the ARE has revealed operational steps that will have to be taken 1n the event of failure of one of the pumps so that loss of circu- lation from the remaining pump and expansion tank can be prevented. The but data are not yet atlon), 17 ANP PROJECT QUARTERLY PROGRESS REPORT study also revealed the need for piping modifications to eliminate gas entrainment at the free surface in the surge tanks. Sodium-potassium alloy has been distilled from closed systems to simulate circuit cleaning of the ARE, negligible quantities of alloy remained in 1solated pockets. PUMPS FOR HIGH-TEMPERATURE LIQUIDS W. B. McDonald A. G. Graindell W. G. Cobb G. D. Whaitman W. R. Huntley A, L. Southern J. M. Trummel P. W. Taylor ANP Divisaion Centrifugal Pump with Combination Packed and Frozen Seal for Fluorides A 50-gpm pump with a Stellite No. 6 coated, 2 1/2-1n.-di1a shaft was operated for 85 hr with the ARE fuel, NaF-ZrF,-UF, (50-46-4 mole %). The seal consisted of a packing of 1/4-by 1/4-1n. monel braid wrapped in nickel foi1l., Operation was unsatisfactory and the test was terminated because of shaft seizure i1n the packing area. Examination of the shaft after the run showed that very severe scoring had occurred. Some grooves were 0,049 1in. deep. It may be concluded from this test that small packings alone are adequate. Compressive loading to minimize leakage will be high and waill result 1n excessive wear because the poor penetration of liquid fluoraides will prevent their acting as a sort of lubricant. The packing i1n subsequent tests 1ncluded braid as a component and either graphite or MoS,, each of which 1s a lubricant, The pump has been repacked waith Dixon’s Microfyne flake-graphite powder, which 1s retained by close- fitting, solid; APC-graphite rings at in- each end of the packing gland., Initaial dry runs, without fluorides present, showed this packing to be quite sensitive to compressive loadaing. Approximately 35 psi was found to be a 18 maximum packing loading for smooth, dry operation without the seal heating excessively. The repacked pump has now operated at a shaft speed of 700 rpm for 300 hr at 1200°F, The pressure against the seal has been varied from 5 to 18 psi. Leakage of solid fuel from the seal has varied from 60 to 75 g/day during operation with pressures in the 5~ to 10-ps1 range. Power 1input to the driving motor has been quite smooth, with a variation from 2.6 to 3.0 kw, and under these con- ditions only 1.5 to 1.8 kw 1s being lost as friction in the pump seal, Alli1s-Chalmers Centrifugal Pump for Liquid Metals. The Allis-Chalmers hydrostatic-bearing pump previously reported(!) was shut down during the quarter for alterations. A new, one- piece shaft has been fabricated, the bearing surface on the i1mpeller has been hard coated with Stellite, and a frozen-sodium gas seal has been 1n- corporated i1n the pump. All fabri- cation 1s complete and the pump 1is ready for reassembly and further testing. Laboratory-Size Pump with Gas Seal The laboratory-size gas-sealed cen- trifugal pump operated 996 hr waith fluoride fuel, NaF-ZrF,-UF, (46-50-4 mole %), at a temperature of 1500°F. The pump has thus far operated a total of 2300 hours. The pump was run at 3600 rpm and produced a 50-ps1 head at 8 to 10 gpm. Figure 2.1 shows the upper assembly of the pump after 996 hr of operation, The ZrF, condensation does not appear to be excessive, however, 1in the original assembly, probe shorting was serious and the two probes in the pump were rendered 1noperable after 50 hr of operation, The shorting may have been a result of fuel freezing around and bridging the probe and radiation shield assembly, shown 1in Fig. 2.1, when an i1nadvertent liquid-level surge (l)ANP Quar 1439, p 19 Prog Rep Dec 10, 1952, ORNL- PERIOD ENDING MARCH 10, 1953 UNCLASSIFIED PHOTO 6-6074 Fig. 2.1. Gas-Seal Pump Assembly After Operation. 19 ANP PROJECT QUARTERLY PROGRESS REPORT occurred during operation or fillaing. This condition has been alleviated by 1increasing the clearance between the probes and radiation baffles and in- serting a half cylinder in the cutaway section 1n the baffle assembly, The pump operated gquite satis- factorily without probes, and the liquid level 1n the pump bowl was established by noting the pump per- formance and temperature of the face plate. The reason for termi- nation was again a broken quide vane 1n the discharge bowl. The vanes have been rewelded tothe bowl several times, but the latest failure occurred in the parent metal of the vane rather than in the weld. The reliability of this assembly can be greatly increased by employing thicker vanes and adding a reinforcing ring around the top edge of the six vanes, An 1dentical pump has been in- stalled and has operated approximately 700 hr 1n a fluoride system that in- cludes a fluoride-to-NaK heat exchanger. Several design changes were made to reduce or nearly eliminate the o1l leakage past the rotary gas seal, It was found that when the gas volume above the pump was vented 1nto a si1ght glass, the discharged helium was heavily laden with o1l vapor. It has not been ascertained whether this vapor 1s harmful to fluorides, but, at least, the contamination was un- desirable. In the original pump, o1l was used to cool the drilled shaft and to flood the rotary gas seal assembly both for cooling and for lubrication. In the present pump, the shaft and the stationary seal ring are water cooled by using a closed system. The rotary face seal can be lubricated, 1f necessary, by adding small amounts of o1l, as liquid or mist, through an opening 1n the bearing housing. The rotary, Graphitar No. 30, gas- sealing ring was replaced with a si1lver-impregnated graphite (Morganite MYIF) ring, and an attempt was made to 20 run the dry seal against hardened tool steel. However, at a bearing pressure of approximately 60 psi, this combi- nation would not run unlubricated. It has been necessary to add small amounts of 011 to reduce gas leakage and to prevent small power excursions 1in the recorded demand of the pump motor. The seal has required an average of 6 to 8 drops of o1l per day, and the gas leakage has been negligible. It 1s planned to use a Keller Aar Line lubricator to atomize the lubrai- cation i1nto the seal region through the bearing housing. This unit may be operated remotely as frequently as experience dictates. ARE-S1ze Sump Pump A modification has been made to the existing model DA pump(z) to convert 1t to a sump pump with a shaft labyrinth seal from which leakage wi1ll return by gravity to the pump’s sump tank. The modified pump 1s now being tested with water and other room-temperature ligquids, and 1t wi1ll be tested with liquids that simulate the density and viscosity of the ARE fuel. Studies are being made of the labyrinth seal during pump operation, the effects of the pump inlet suction bell, and the volume of fluid and the bafflaing required in the sump tank for satisfactory pump operation, The operation of the labyrinth seal (0 009-1n. radial clearance) was 1in- vestigated by measuring the leakage of the seal at various pump speeds and heads. As expected, the fountain 1ncreases with increase in pump speed and in pump head. However, the amount of the bypass fountain flow varied from 2 to 10% of pump flow, which 1s a reasonable fraction. A minor problem was encountered in that the labyrinth leakage splashed the surface of the lIiquid 1in the sump tank and some of the gas bubbles were carried deep enough to be entrained in the fluid entering the impeller. (2)y 6. Cobb, A G Whitman, ANP Quar ORNL-1375, p 17. Grindell, and G D Prog Rep Sept 10, 1952 The pump has been tested 1n two types of sump tank. The first tank was a rectangular, plexiglas box, 14 by 16 by 30 in., in which the fluaid 1inlet to the tank was considerably offset from the eye of the impeller, The second tank was a plexiglas cylin- der, about 12 in. 1n diameter and 14 in. high, in which the fluid inlet 1s below and directly 1nto eye of the 1mpeller. When continuously submerged, even 1f only to a small fraction of an inch, the pump i1nlet suction bell appeared to function satisfactorily in elther sump tank under all conditions of pump speed and flow. Pump priming 1s not difficult to accomplish 1n either tank. In the 1nitial test setup, the volume of gas 1n the system piping at startup was approximately equal to the volume of Ii1quid, and 1n the cylindrical tank, the prime was broken by a sudden surge of gas i1nto the impeller eye. Placing an obstruction over the fluid inlet to divert most of the gas bubbles away from the eye to the free surface was sufficient to maintain the prime. 1In a system i1n which there 1s a small gas volume compared with the liquid volume at startup, priming 1s accomplished by bleeding the trapped gas into the sump tank or by filling under vacuum. In the ARE system, the trapped gas 1s forced out. Tests were made 1in which the fluid was gassed by admitting air, helium, or argon 1nto the eye of the suction bell until the fluid became milky. The time required for the fluid to become relatively clear was then noted, In all tests, 1including tests with flow rates of up to 100 gpm and tests 1n which the cylindrical tank without baffles was used, the fluaid degassed 1nless than 5 min of operation. When the 1inlet of the cylaindrical tank was offset about 1/2 1n. with respect to eye of the impeller, the entire holdup volume of liquid in the tank was given a rotary motion that 1in- creased 1n velocity at high impeller PERIOD ENDING MARCH 16, 1953 speeds and low flow rates. When the baffles were 1n place, this circular motion occasionally caused the for- mation of vortexes and gas entrainment in the liquid. Without baffles i1n the cylindrical tank, the rotary motion was present but no vortexes were observed. The gas seal for this pump will be either a frozen-sodium or frozen-lead seal or a mechanical face seal. ARE Centrifugal Pump An ARE model FP pump(3®’ has been constructed and 1incorporated 1n a test loop. The most critical feature of this pump was the frozen-sodium seal for which sodium from an external source was fed to a sealing annulus and frozen in rings on each si1de of the annulus., Three tests have been run in which NaK was pumped for about 15 min, 4 hr, and 1 hour., FEach test was terminated because of excessive seal leakage. Chemical analyses of samples of leakage material 1ndicated the presence of from 1 to 7.5% potassium. Since thas pump does not operate successfully with NaK, 1t was modified for testing with sodium. For operation withsodium, the external supply of sodium to the seal wi1ll not be needed and a larger surface wi1ll be presented for cooling of the seal. Prior to installation of the new sealing gland, a series of tests was run on the externally supplied seal with helium i1n the pump loop. During these tests, pump speeds and seal temperatures were systematically varied, and considerable data were taken to provide more adequate under- standing of the frozen-seal mechanism. Successful operation, as a gas seal, was obtained over the available range of speeds, 400 to 1920 rpm, and the temperature of the sodium in the gland groove varied from 300 to 400°F. Speed and pressure conditions were determined under which the seal would fail and under which 1t would operate reliably. (B)ANP Quar 1439, p 21 Prog Rep Dec 10, 1952, ORNL- 21 ANP PROJECT QUARTERLY PROGRESS REPORT The sealing gland that has been installed i1n the pump for operation with sodium has a length-to-diameter ratio of 2.3, 1n contrast to the ratios of 0.6 and 0.2 for the 1insade and outside glands of the previously used externally supplied seal. Also, the radial shaft-to-gland clearance has been increased from 0.015 an. for the externally supplied gland to 0.030 1n. for the longer gland. Initaial tests of this gland and seal indicate a satisfactory seal, but they also 1indicate the presence of excessive thermal distortions that will result in displacement of parts and eventual bindaing. Electromagnetic Pump The two- stage electromagnetic pump that waill be used to circulate NaK in testing and cleaning operations of the fuel circuit of the ARE has been tested with NaK at 1500°F and 1s undergoing similar tests with sodium. With Nak, the pump delivery 1s approximately 90 gpm at approximately 10 psi, per- formance curves are shown in Fig. 2.2, The pump has givenentarely satisfactory performance, except for a hard-soldered Joint failure in the current loop, which was easily repaired, cooling, which was removed. and excess ROTATING-SHAFT AND VALVE-STEM-SEAL DEVELOPMENT W. B. McDonald R. N. Mason W. C. Tunnell P. G. Smith W. R. Huntley ANP Division R. E. Engberg Long Range Reactor Planning Group Mechanical Face Seals High- temperature mechanical face seals for li1guids such as fluorides are being 1investigated, Successful operation of such seals may be dependent upon dry- film lubrication, boundary lubrication, or adsorbed-film lubrication, but research has revealed only empairical 22 knowledge about the principles of similar seals successfully used 1n applications much less severe than those i1mposed by ANP requirements. The problem appears to be one of selecting appropriate materials that wi1ll retain optically flat surfaces at high temperatures and will have the proper degree of hardness and softness relative to each other, Graphite 1s being considered for the soft material and materials such as tungsten carbade, titanium carbide, and cermets are being considered for the hard materaial, The methods of fabrication and the maintenance of high ambient tempera- tures and low thermal gradients to avoid distortions are also pertinent problems. The work of others has shown that a principal factor in determining wear and surface damage 1n sliding 1s the tendency of the sliding or rubbing materials to alloy. Although other factors, such as the hardnesses of the materi1als, the melting points, the rubbing velocity, and the temperature of operation, are considered as being of secondary i1mportance, 1n some instances they may become of controllang 1mportance. The use of MoS, as a high- temperature lubricant 1s intriguing, and considerable work has been done to obtain surfaces coated or impregnated with this material. Table 2 1 lasts the attempts to establash a MoS, layer by adding MoS, to the vehicle to make a paste. It should be noted that MoS, will begin to oxidize at about 700°F 1n air, but 1n the absence of oxygen 1t 1s stable to 2000°F. Tests of face seal materials, made by rotating them 1n contact with the other specimen are described in Table 2.2, The chatteraing noted may have been due to forces resulting from the geometry used 1n the test apparatus. These experiments indicate that of the materials tested, graphaite 1s best for one of the materials, under these test conditions, when air 1s present, PERIOD ENDING MARCH 10, 1953 \ UNGLASSIFIED DWG 18713 * I T lC RRENT l INPUT VOLTAGE TO CU NT LOOP '\-Cl\ \ A 100 \r\ B 125 i6 Lt ) \ 0 150 \ ® 175 EUTECTIC NaK LOOP AT 750°F 12 ~ g \A\\ \ \D\ O] > \ \ \ 4 0 0 10 20 30 40 50 60 70 80 a0 FLOW (gpm) UNGLASSIFIED DWG 18712 20 INPUT VOLTAGE TO GURRENT LOOP ® 25 O 50 16 A 75 A 100 W 125 \‘.X 8 ‘50 175 g2 -~._E}__-- I fi‘"”“--.‘~h\~\ EUTECTIC NaK LOOP AT 1500°F E "T}~§§\‘ [15] h‘h“fi- *75\‘\“‘\55\\\\\\;::\\&\\\\\ ox \ 0 10 20 30 40 50 60 70 80 90 FLOW (gpm) Fig. 2.2. Performance Curves for Double-Stage Electromagnetic Pump with Nak. Input voltage to magnet loop, 125 volts. 23 ANP PROJECT QUARTERLY PROGRESS REPORT TABLE 2.1, DESCRIPTION OF ATTEMPTS TO OBTAIN A MoS, SURFACE BASE METAL VEHICLE TREATMENT AND APPLICATION REMARKS 1 Type 316 G-E enamel Brushed on, air drled,o Good coating obtained, atainless steel furnace heated to 700 F apparently 2 Type 316 Colloidal Brushed on, air draed, , Did not stick to base metal stainless steel s1lver furnace heated to 700 F 3 Type 316 High- Brushed on, air dried, . Varnish burnt off, dad not stick stainless steel temperature furnace heated to 700 F varnish 4 Type 316 No 1202 varnish | Brushed on, air dr1ed,0 Varnish burnt off, did net stick stainless steel furnace heated to 700°F 5 Type 316 Keystone grease Base metal heated to 700°F, Did not stick to base metal stainless steel paste brushed on 6 Type 316 Permatex Ne 2 Base metal heated to 700°F, Showed promiae, fair coating atainless steel paste brushed on obtained 7 Type 316 Ruby flux Base metal heated te TOOOF, Showed promise, fair coating stainless steel psste brushed on ebtained 8 Type 316 Aquadag No 200 Base metal heated to 700°F, Did not stick to base metal stainless steel paste brushed on 9 Type 316 Permatex No 2 Base metal heated te TOOOF, Remained soft, possibly net stainless stesl thinned waith paste brushed on enough heat, indicates promise 10 Lapped tungsten carbide 11 Etched tungsten carbide 12 Type 316 stainless steel 13 Tungsten carbide 14 Tungsten carbaide 15 Type 316 Sebacate None None None Handi-flux Ruby flux Permatex No 2 Heated to 700°F, dipped in dry M082 Heated to TOOOF, dry Hn82 sprinkled on Heated to 700°F, dipped 1n dry H082 Paste applied, heated with torch Paste applied, heated waith torch Base metal heated, brushed Thin layer obtained, apparently Thin layer obtained, apparently obtained Good layer Adhered to uneven base metal, but Soft and flexible, apparently Good layer, apparently Brass burnt, but apparently stainless steel with Sebacate on 1n furnace 16 Brass Handi-flux Brushed on base metal, heated 1n furnace 17 Type 316 Corn syrup Brushed on base metal, stainless steel heated 1n furnace 18 Type 316 Celluflux Brushed on base metal, stainless steel heated 1n furnace 19 Type 316 Pliobond Brushed on base metal, stainless steel heated 1n furnace some layer remained Fair coating, shows premiase Good coating, but uneven, shows promise Fair coating, hard and brattle An attempt has been made to establash an upper temperature limit for silver- impregnated graphite rubbing against type 316 stainless steel. The assembly ceased to seal helium at about 450°F, and although the temperature was 1in- creased slowly to 910°F, sealing daid not occur again, When the temperature was decreased to 450°F, sealing was 24 again accomplished. w1ll be continued, and varicus sealing pressures will be tried. Another experiment 1s 1n progress 1n which an attempt 1s being made to seal against water with silver- impregnated graphaite vs, type 316 stainless steel atvarious differential pressures and bearing pressures. To This experiment TABLE 2.2, PERIOD ENDING MARCH 10, SUMMARY OF TESTS OF FACE SEAL MATERIALS BASE METAL NO. {cf , Table 2.1) MATERIALS RESULTS Operated for 2 hr before chattering occurred Operated for 2 hr before chattering occurred Operated for only a short period Operated for 5 hr before chattering occurred Operated for 3 hr before chattering 1953 11 Etched tungsten carbide vs type 316 stainless steel 10 Lapped tungsten carbide vs type 316 stainless steel 13 Tungsten carbide vs tungsten carbide 16 Brass vs tungsten carbide 12 Type 316 stainless steel vs tungsten carbide 14 Tungsten carbide vs type 316 stainless steel 15 Type 316 stainless steel vs tungsten carbide C-18 graphite vs carbide Graphite (No carbide tungsten 40) 1mpregnated with MoS, vs tungsten carbide APC graphite vs tungsten C-18 graphite vs tungsten carbide C-18 graphite vs tungsten occurred, addition of dry MoS, stopped chattering for about 15 min Operated for 7 min before chattering occurred Operated for 50 min before chattering occurred Chattered immediately, surface not flat Operated for 12 hr at self-generated temperature of 225°F Attempted to coat graphite with fuel 27, operated for 4 hours Heated with torch to 475°F, chatter developed Dry MoS, added, heated to 913°F wath torch, operation smooth, except between 400 and 450°F. when chattering occurred date, the seal has not been consistent, 1t has leaked at times and in varying amounts. Other materials have been ordered, such as boron nitride compacts, MoS,- 1impregnated copper, silver and stain- less steel compacts, and a sintered molybdenum compact, that will be treated to form a MoS, surface. Combination Packed and Frozen Seal with MoS,. Because MoS, has high- temperature lubricating properties, an attempt was made to use 1t as a high-temperature packing material for sealing fluorides, fluoride leakage was slight, operation was smooth, and there were no power surges. However, as 1n previous experiments, the MoS, was too fluid and could not be con- tained. The retainers used were Graphitar No. 14 rings. The material slowly leaked out of the stuffing box and around the gland until the gland was completely 1inserted, For the next attempt to seal waith this material, a packing of Inconel 25 ANP PROJECT QUARTERLY PROGRESS REPORT braid impregnated with MoS, was used. This experiment was terminated after approximately 925 hr of operation with fuel No. 30 at 10 psi and 1750 rpm. Excessive leakage occurred because the seal reached too high temperatures when an attempt was made to restart operation after binding had occurred. The operation of the seal corresponded to operation of seals 1n other tests 1in that the leakage of the seal was apparently temperature sensitive. When the seal was too hot or too cold, leakage and power surges resulted, but under certain conditions operation was smooth and there was no detectable leakage, There have been nointentional start-stop tests, although the unit has been accidentally stopped on several occasions, and power surges have overloaded the motor relays several times, The unit has been restarted with little difficulty on each occasion by heating the packing area and using a wrench to ‘““break’” the shaft loose. Combination Packed and Frozen Seal with Graphite. Thecontinuous operation of the graphite-packed seal, mentioned 1n the previous report, was terminated after 725 hours. The seal had shown no signs of failure, and after another 150 hr of periodic and stop-start operation, the test was again termz- nated. During the latter period of operation, attempts were made to control the seal temperatures and, thereby, the zone in which freezing was occurring, since 1t was believed that most satisfactory operation would occur 1f freezing took place within the graphite region. Two fairly stable temperature ranges of operation were found, one set of temperatures corresponded to that maintained durang the 725-hr period of continuous operation and the other set was about 150°F lower. at intermediate temperatures. Operation was unstable At the lower temperatures, operation was smooth, there was no detectable leakage, and the unit could be stopped 26 for as long as 20 min and started again by motor power only. The seal used 1n this test was similar to the seals used 1n other tests i1n that leakage occurred when the seal was too hot or too cold. Graphite Packed Seal. In an attempt to determine the benefits of holding the transition from fluid to solad within the graphite, a test was set up that i1ncluded a long (4 1in.), graphite- packed, stuffing box. In order to obtain more complete temperature data, four thermocouples were installed in the rotating shaft, and signals were taken off through slip rings. After a dry run, fluorides were i1introduced and the seal i1mmediately started leaking graphite. The Graphitar rings used as retainers did not contain the graphite at the top of the seal. Approximately 3/4 in. of Inconel braid was then added, and the graphite was successfully retained. The motor for this test was a 5-hp unit. Operation started off fairly normally, but the temperatures adjacent to the seal began to increase. Cooling of the bearing housing and the gland was started, but the temperatures continued to rise. The temperatures never became stable, they fluctuated over a range of 940 to 2190°F throughout the seal regiron. The hottest point was apparently near the middle of the packing, and the lowest temperatures were 1n the regions of the bearings and the fluoride. The molten fluoride was acting as a coolant in this test. The temperatures continued to increase until the test had to be terminated after approximately 50 hr of operation because the shaft .froze. The shaft could be restarted by motor power afterit cooled, but would not continue to operate when higher temperatures were reached. At no time was there any detectable leakage, however, severe shaft scoring occurred. A picture of the seal and shaft cut through the center line 1s shown in Fig. 2.3 The extreme scoring of the PERIOD ENDING MARCH 10, 1953 g 2 we Lo ] 8o 3o 3L £2 5 Shaft and Graphite Seal. Section View, 2.3. Fig. A 27 ANP PROJECT QUARTERLY PROGRESS REPORT 1 3/16-1n -dia shaft 1s quite evident, other portions of the shaft (also at very high temperatures) showed no markings whatsoever. The condition of the graphite 1tself 1s of interest, since 1t apparently packed i1n more or less “concentric” radial layers that varied i1n thickness along the length of the seal. The layers adhered at times to the shaft and rotated, and at other times they adhered to each other and remained stationary. The powder was exceptionally fine and very “greasy”, the “concentric” layers were very slimy, and there was no visible evidence of fluoride penetration. Rotating-Shaft Seal Test. The apparatus for the tests of the rotating- shaft seal consists of a container for the packing material being tested, an upper compression gland that exerts pressure on the packing, a lantern gland located midway in the container, and a shaft extending through the assembly. The shaft rotates in a bearing beneath the assembly and 1s driven by a motor mounted above The packing container 1s mounted on four rollers and 1s restrained from turning with the shaft by two tension springs The torque necessary to turn the shaft 1s 1ndicated by a pointer and a calibrated scale on the assembly Fluorides are forced under pressure from a fill tube into the container through the lantern gland During the tests, the temperature of the midpoint of the packing container was 1500°F, the temperature of the lower shaft opening was 1400°F, the pressure was 20 psi. The packing materials tested were MoS, and powdered graphite The MoS, packing began to leak after 1 hr of operation. The torque required to turn the shaft at 350 rpm was 2 ft-1b. The graphite test ran for 8 hr before leakage occurred, and the same torque was required for rotation. There was considerable wear on the upper packing gland. The shaft had to be started and stopped a number of times because of severe vibration. 28 Packing Penetration Tests. The apparatus used 1n the packing pene- tration tests and the results of several tests were described prevai- ously ¢*) Further tests have been made and the results are presented in the following Stainless steel braid impregnated with MoS, was compressed and heated twice to the annealing temperature, with further compression after each heating This packing held the fluorides for 1/2 hr at 2 psi, but when the pressure was i1ncreased to 5 ps1, leakage occurred. Examination showed that the leakage path was along the walls of the containers, the screw stem, and the strands of the braid It could not be detected that any fluoride mixed with the MoS, or penetrated through 1t. A second boron nitride test was run 1n an assembly with much smaller clearances than those used for the previous boron nitride test. In this test, 1t was possible to raise the pressure to 30 psi, and operation was continued for 234 hr before leakage occurred, A test with J. T. Baker Chemical Co. powdered graphite previously reported(?) has continued for 240 hr with no leakage. Analysis showed that the packing was approximately 30% amorphous carbon. Three more graphites have been tested, and ain all cases 1t was possible to raise the pressure to 30 psi1i without leakage occurring. The analyses and results are given 1n Table 2.3. There should be a minimum of amorphous carbon 1in the graphite because 1t would be expected to increase the friction between a rotating shaft and the packing, however, the graphite with the highest amorphous carbon content sealed for the longest periods. That the packing i1n the first test did not leak whereas the packing in the later (4)4ANP Quar Prog Rep Dec 10, 1439, p 23 1952, ORNL- PERIOD ENDING MARCH 10, 1953 TABLE 2.3. PACKING PENETRATION TESTS ANALYSIS OF AMORPHOUS HOURS OF OPERATION PACKING MATERIAL CARBON CONTENT BEFORE LEAKAGE (%) OCCURRED Dixon No. 2 flake graphite 10 22 Dixon Microfyne graphite 50 75 Amer Graphite Co. No. 620 powdered graphite 50 63 tests did, although the physical characteristics of the graphites were the same, might be the result of fuel No. 14 being used 1n the fairst test and fuel No. 30 in the later test, Frozen-Sodium and Frozen-Lead Seal for NaK. It was reported previously(%? that tests were under way to determine the feasibility of sealing a NaK pump with an externally supplied frozen- sodium or frozen-lead seal. Although the preliminary tests were encouraging, subsequent tests revealed that seal life (the time before failure because of alloying) was, 1n general, 1inversely proportional to NaK temperature and that the seals were unreliable for periods of operation greater than 100 hours. In the frozen sodium seal, the sodium alloys with the NaK and forms a low-melting-point alloy in the seal that cannot be frozen with room- temperature cooling water, as a result the seal 1s completely lost. Re- frigeration, with temperatures below the melting point of eutectic Nak (12°F), would be required to assure seal retention. In the frozen-lead seal, the lead alloys with the NaK and forms a high- melting-point alloy, as a result there 1s shaft seizure and 1t 1s impossible to operate with reasonable power requirements. (S)AHP Quar Prog Rep Dec 10, 1439, p. 26 1952, ORNL- ARE Valve Test. Preliminary gas and liquid leakage tests have been run on the ARE pneumatically driven valve supplied by Fulton-Sylphon Co. The valve 1s constructed of Inconel and has a Stellite-to-Stellite valve and seat. The shaft-sealing member 1s a four-ply Inconel bellows The valve 1s adaptable to being eather normally closed or normally opened, and the position 1s reversed by application of 15 psi1 of air pressure to a large bellows actuator. For the 1nitial test, the valve was mounted vertically and adapted for normally closed operation. Waith the valve at room temperature, gas leakage was small with test pressures from 30 to 60 psi1 that were first applied up against the valve seat and then down on 1t. Raising the valve temperature to 1050°F caused a slight i1ncrease 1n gas leakage, however, the leakage remained small. Testing for liquid leaks with the ARE fuel, NaF-ZrF,-UF, (50-46-4 mole %), at 1100°F revealed that there was no leakage with test pressures of 30 to 60 psi applied both above and below the seat. This test was repeated several times. In a test at 1300°F, slight leakage (0.3 an.3/hr or less) occurred at various 1ntervals However, 1t was concluded that the leakage at thais temperature was not serious. The initi1al checks at 1500°F were very successful. Subsequent checks, 29 ANP PROJECT QUARTERLY PROGRESS REPORT however, showed a very serious leakage rate, 120 in.3/hr, that could not be corrected by reseating the valve, It was concluded therefore that the valve might be usable for several cycles at 1500°F but that 1t could not be expected to hold tight 1f cycled indefinitely. Difficulty at the elevated tempera- ture was experlenced 1n that the valve stuck when left closed for short periods However, 1n all cases to date, the valve has eventually become operable by increasing the pressure up to 30 psi1 on the valve actuator In one instance at 1100°F and another at 1500°F, 1t was necessary to vibrate the valve body to get the valve open. When the valve was cut for 1inspection, 1t was found that some binding of the Stellite faces had occurred HEAT EXCHANGER SYSTEMS G. D. Whitman D. F. Salmon ANP Division Sodium-to-A1r Radiator Tests. The sodium-to-air radiator tested had a core element with 30 fins per inch, (®) This test ran for 200 hr and was then terminated because of poor heat transfer performance. The maximum, over-all, heat transfer coefficients obtained were of the order of 6 Btu/hr*ft?+°F, which 1s far below the predicted performance It was discovered, after i1nlet sodium temperatures of 1500°F were reached, that very few of the Nicrobrazed fin-to-tube joints were good. The radiator could be seen through the air exhaust, and 1t was observed that with minimum air flow less than 10% of the nickel fins showed any heat color. The small amounts of braze material used to avoid closing the 0.025-1n. fin-to-fin gaps were ap- parently i1nsufficient to completely join the fins and tubes. No oxide plugging occurred 1n the radiator (6 )ANP Quar 1439, p 21 Prog Rep Dec 10, 1952, ORNL- 30 tubes during this relatively short run, and the bypass filter caircuzit was operated without diffaiculty. A Roots Connersville gas pump driven by a 2-hp Varidrive has been installed 1n place of the Buffalo centrifugal blower so that a wider range of air flows and better flow regulation can be obtained. Bi1fluid Heat Transfer Loop. The b1fluid heat transfer loop,‘’’ which has been 1n operation for approximately two months, transfers heat from fuel No 30 to NaK. The heat transfer takes place 1n a concentric tube section, the center tube contains the hot fluoride and the NaK 1s 1in the annulus. The center tube has an inside diameter of 0.269 in, and an L/D ratio of 40, whereas the annulus has an L/D ratio of 22. Heat transfer data on the fluoraide fuel have been taken over a Reynold’s number range of from 5,000 to 20,000, which corresponds to velocities of from 8 to 30 fps. The maximum fluoride temperature was 1400°F, and the temperature drop across the heat exchanger was varied from 15 to 40°F. The NaK-side data were taken at Reynold’'s numbers from 20,000 to 100, 000. The over-all heat transfer coef- ficient for heat fluxes from 300,000 to 500,000 Btu/hr*ft? has varied from 1000 to 2500 Btu/hr /) k AW pd j —_— ———— FLUID FLOW Fig. 2.4. Sketch of ARE Fuel-Circult Mockup. 31 ANP PROJECT QUARTERLY PROGRESS REPORT surge tank circuits - one for moderator coolant and the other for fuel. Each circuit includes two parallel circuits, each of which includes, 1n series, a pairofheat exchangers (inparallel), a pump, and a surge tank. As designed, each surge tank was to receive and deliver the entire flow of 1ts par- ticular pump. However, the surge tanks of the moderator coolant circuits have now been provided with bypasses to reduce the flow through them to low levels and thereby reduce gas- entraining turbulence to acceptable values. The moderator coolant flow rate 1s, 1n general, several times higher than the fuel flow rate. However, in the fuel circuit there are strong 1ncentives to retain complete flow control of all fuel, 1inasmuch as the system will be filled initially with nonuranium-bearaing salts and wi1ll have U?35_bearing salts added to 1t until criticality 1s reached. Once addition starts, no &alt can be removed or drained, and complete and uniform mixing must occur. Consequently, the expansion or surge tanks must provide sufficient capacity to receive the enriched fuel (about 15% of the system volume) and also to provide for fuel thermal expansion from 1000 to 1500°F (about 10% of system volume). The thermal expansion w1ll be about 200°F more than that anticipated. There 1s also a strong incentive for minimum volume holdup of enriched fuel in each surge tank. Vortexes and Bubbles. The system was operated 1initially with water, and the surge tanks were 1nvestigated for vortexing above the discharge lines. The surge tanks had no baffles, and entrained gas was observed in the discharges at all flow rates above 10 gpm with the tanks half full of liquid. This gas entrainment was due to vortex- 1ing of the liquid leaving the tanks. At lower liquid levels, vortexing occurred at flow rates of less than 10 gpm. Many geometries were tried for breaking up this vortexing, and a 32 baffle design was finally developed that allowed water flow rates of up to 40 gpm to be put through the tanks without visible gas entrainment 1f a minimum level of approximately 4 ain. was malntained in the 10-1in.-di1a tanks. This antivortexing baffle consists of a flat plate, 6 by 8 1n., attached to the bottom of the probe well that 1s centered over the discharge laine. The probe well was cut off so that the plate could be located 2 in. from the bottom of the tank. There are three vertical fins attached to the plate, one 1n the center and one near each edge, so that they are approximately 1/8 1n. from the bottom of the cy- lindrical tank. This device serves to guide the flow from the inlet to the outlet with less velocity loss than in the open tank, and the flow 1s re- strained from going 1n on the sides and edge near the end of the tank where the more severe vortexing always occurred. When the glycerine solution was tried, there appeared to be less tendency to entrain gaé 1n the surge tanks. At rated flow, 40 gpm, lower fluid levels could be maintained 1in the surge tanks without there being visible gas entrainment 1in the dis- charge. It had been felt that Froude’s modulus should be applicable, since the fluid mockup and the ARE fuel circuits were similar, and that gravity forces should predominate on the free surfaces in the tanks. Since the mockup surge tanks are full ARE si1ze, the ARE fuel flow rates were maintalned with the water and the glycerine-water solution for this in- vestigation i1n order to maintain hydraulic similarity, The moderator circuit in the ARE 1s quite similar 1n design to the fuel circuit. The Na or NaK flow (about 100 gpm) cannot be put through a surge tank of ARE dimensions without there being very severe gas entrainment 1in the discharge. This condition can be remedied by bypassing most of the flow around the tanks or by using the tanks as stand pipes only. The first remedy would be the more desirable, since gas removal would be more effective, It was found that the surge tanks removed visible gas bubbles in the water and glycerine-water solution. Air bubbles of the order of 1/8 in. 1n diameter and larger were removed 1in a very few cycles, and those so small as to be barely visible were removed after an hour or so of circulation at low or rated flows., When the glycerine-water solution was first transferred to the loop, 1t was nearly opaque because of the finely divided gas bubbles, but the solution was clear after circu- lation through the surge tanks for approximately 1 hour. The system tested for gas removal had a capacity of 15 gallons. Operational Instabilities The system has also been i1nvestigated with respect topump failures and operational instabilities. During operation with each pump delivering 30 to 40 gpm at approximately 45 psi head, 1f one pump 1s shut down the remaining pump 1im- mediately becomes gas bound, starts pumping 1intermittently, and severely gasses the loop. The gassing of the pump that remains operable results from the transfer of liquid from the surge tank 1n the operating loop to the surge tank in the loop that was shut down, that 1is, when one pump 1is shut off, 1ts associated surge tank, which 1s essentially at pump suction pressure, 1s subjected to the dais- charge pressure of the operating pump and therefore gas volume 1s compressed and causes transfer of liquid from the operating loop to the loop that was shut down. When a pump operating at or near rated flows stops, eirther the loop must be immediately 1solated from the system or the remaining pump stopped. The loop containing the 1noperable pump must be valved out of the system before flow 1s restarted. Also, at relatively high flow or pressure rise PERIOD ENDING MARCH 10, 1953 across the pumps, impeller speeds must be the same for both pumps, or excessive amounts of liquid will be transferred from one surge tank to the other, In an attempt tomaintain equilabrium between the two tanks 1n case of a pump failure, they were connected by a liquid line. It was never possible to operate satisfactorily with this line open because one pump would develop a slightly greater head and take over the entire load and thus 1induce back flow through the other pump. At such times, there was heavy flow from one tank to the other through the liquad line connecting the two surge tanks. Flow through the heat exchanger cair- cuits would not drop off apprecaiably. The pump carrying the load could easily supply both heat exchangers and a small back flow through the other pump. The head flow characteristics of the pumps are shown in Fig. 2.5, It will be noted that between shut-off and 50 gpm there 1s little or no drop in head, and at some speeds a very slight pulsation point can be detected. These flat curves explain the abilaty of one pump to handle the total flow of the reactor and heat exchanger, as previously mentioned., One pump run- ning at a slightly higher speed can easily buck the other pumps and handle the total loop load without losang sufficient head to fall below the shut-off head of the other pump. During operation below 25 gpm per pump, oscillations have been observed 1n the flow. These have never been serious and could always be reduced to negligible amplitude by increasing the flow rate of the system. The osc1l- lations are evidenced by an 1ncrease 1in flow rate and liquid loss 1n one pump circuit and a simultaneous de- crease 1n flow rate and liquid gain 1n the other circuit. The condition then reverses and the period of the oscillation 1s approximately 2 seconds. The maximum change in flow rate has been t2 5 gpm at flow rates of 20 gpm per pump. Thas disturbance can be 33 ANP PROJECT QUARTERLY PROGRESS REPORT HEAD (psi) UNCLASSIFIED 60 | DWG 18715 ® NO 1 PUMP O NO 2 PUMP 35 A ARE PUMP 3880rpm 50 J 363Orpmd J i S— — i — —_————)—————_ e SE—— —— — a5 b) Q ) P O o 40 ZF_-——- —— —igierpm 3300rpm —_H e —_— 'P——* —l—-._._._‘ !5- . e —— . ! 35 30 2780 rpm —_— Y o W W o W 25 2475rpm 20 — 1400 rpm 15 I20(3rpm — \ - - —- --.“. - ZS-‘--_ . | 1900 rpm ) S ——— ——— — — — —— —— — C— S —— — c _““D_—c)— ———-—O—-— 10 1335 rpm P 0 o 5 10 15 20 25 30 35 40 45 50 FLOW (gpm) Fig. 2.5. Head Flow Characteristics of Pumps. 34 explained by the pump characteristics and the absence of check valves in the discharge lines of the parallel pump arrangement. In this particular system, when one pump takes the entire load, there 1s a subsequent pressure rise i1n the surge tank of the stalled pump. This 1s then seen as an increase 1in discharge pressure of the stalled pump, whach then overpowers the pump handling the load and momentarily reverses the Ssi1tuation, The actual reversal of flow occurs only 1n the line connecting the surge tanks through the pumps, that 1s, the suction lines and the common dis- charge line., The flow through the heat exchangers 1s unsteady but does not reverse or stop. The flow through the reactor appears tobe quite steady, even though 1t 1s being supplied by alternate pumps. An Esterline Angus recording watt- meter was 1nstalled on one pump motor, and 1t was established that duraing these oscillations the pump alternately carried full load and shut-off head demand. This condition can be ef- fectively corrected by 1nserting a resistance 1n the common pump dis- charge line sothat stable pump charac- teristics are obtained. A 2-psi 1n- crease 1n pressure 1s sufficient to result i1n a discharge curve which that the head decreases with flow from shut-off to maximum flow. This was demonstrated by partly closing one of the throttle valves 1n the pump discharge line during an oscail- lation, the disturbance immediately stopped. shows The system 1s now being tested waith tetrabromoethane, which has physical properties nearly the same as those of the fluorade fuel (viscosity, 9.27 cps, specific gravity, 2,95 at T7°F). The mockup wi1ll again be used to check gas removal and vortexing in the surge tanks, filling and draining the reactor, pump characteristics and flow i1nstabilities. PERIOD ENDING MARCH 10, 1953 INSTRUMENTATION W. B. McDonald D. R. Ward P. W. Taylor A. L. Southern P. G. Smath ANP Divaision Pressure Measurement Two tests of Moore Nullmatic pressure transmitters, operating with the 0.005-2n. wall bellows completely submerged in the ARE fuel, NaF-ZrF, -UF, (50-46-4 mole %), at 1100°F, have logged enough operating time to make 1t seem likely that thas method of pressure measurement 1s the most advantageous ofany tried to date. The transmitters are ‘“upside down’' in the sense that the cavity containang the bellows and the liquid under pressure 1s open at the top. A nickel transmitter has been cycled every 1/2 hr between 10 and 30 psi1 for 1900 hours. A type 316 stainless steel transmitter has operated similarly for 1500 hours. A third transmitter of type 316 stain- less steel filled with lead and topped with a layer of the molten fuel has 1500 hr of operating time, to date, and 1s running slightly cooler than the others. The initial zero shift ain each case was approximately 1% of full scale at the operating temperature, and dri1ft over the 1500-hr period was approximately 2%. However, the zero position can be adjusted easaily, af necessary. Several other Moore transmitters have been installed in dynamic loops for circulating fluorides or liquad metals. These transmitters operate “right side up,’’ and there 1s trapped helium around the bellows to protect them from the liquid. This method of protecting the bellows, although operable 1n most cases, does not seem as satisfactory as complete submersion. The ARE fuel 1in the nickel transmitter mentioned above has been rapidly frozen and then rapidly remelted three times, with no apparent damage to the 1nstrument. This 1s not possible with the trapped-gas instrument because of fouling of the bellows. It was also 35 ANP PROJECT QUARTERLY PROGRESS REPORT found that when the pressure transmitter 1s filled with static fuel, the bellows temperature can be maintained at 1100°F or below. Operation of the bellows at the reduced temperature reduces the corrosion rate. Failure of the bellows i1n one of the type 316 stainless steel trans- mitters after 500 hr of operation at 1200°F was apparently caused by oxi- dation, Helaium instead of air 1s now used as the continuous-flow-balancing gas 1n the pressure transmitter tests, ARE Leak Detection Indicator Tests There 1s i1nterest in developing a secondary method for detecting leaks in NaK or fuel systems in the ARE to supplement the halogen leak detector. Experimentation revealed that a so- lution of phenol red can be made very sensitive to changes in pH by adjust- 1ng the solution to a reddish brown. When helium 1s passed over a surface of hot fuel and then over the surface of the i1ndicator solution, the solution changes color. Likewise, 1t was found that a cold lump of the fuel dropped into the indicator solution changed the i1ndicator toward the acid direction, Slip Rings for Temperature Measure- ment Three tests of slip raings for measuring the temperature i1nside a rotating shaft were started. Because si1lver-graphite brushes were notavail- able, carbon brushes were used. One test was completed i1n which copper rings and carbon brushes were used. When the brushes were well run-ain, an accuracy 1n the temperature measure- ments of +1°F at 1200°F was obtained. The combination of carbon brushes and carbon rings 1s satisfactory 1f a wiping brush 1s mounted on the raing. If the ring 1s not kept clean, the resistance across the contact increases with time, and the error may increase to t40°F or more. A second combination of materaials, coin silver and silver-impregnated graphite, was placed 1n operation before being tested, and performance has been satisfactory for several 36 hundred hours. In a third test that 1s being conducted with type 316 stainless steel rings and carbon brushes, an error of +20°F has de- veloped, but 1t 1s a constant error. The carbon brushes will be replaced with si1lver-i1mpregnated graphite brushes, which should reduce the error slightly. Commerical slip ring as- semblies have been ordered that waill work on either a 7/8- or a 2 1/4-1n. Shafto Rotameter Type of Flowmeter. Testing has continued on the rotameter type of flowmeter reported previously, (%) and approximately 1000 hr of operation has been logged 1n the temperature range of 1100 to 1300°F. This 1instrument will measure flows of fuel No, 30 with 10% accuracy over the range of 9 to 60 gpm. Two i1nstruments of this type have been tested, and the results 1indicate that this instrument will be reliable for ARE operation. HANDLING OF FLUORIDES AND LIQUID METALS L. A. Mann J. M. Cisar F. M. Grizzell ANP Division Distillation of NaK. NaK has been chosen as the precleaning fluid for the ARE before introduction of fluoride salts. Since only a small fraction of NaK (less than 0.5 wt %, cf., sec. 9) can be tolerated in the fuel without reduction of UF, to UF3, studies have been made to determine how much NakK remains after the system 1s drained and how effective distillation 1s 1n removing 1t. Since the vapor pressure of potassium at any temperature 1s higher than that of sodium, upon distillation the alloy undergoes rapad depletion of potassium, and there 1s a corresponding enrichment of sodium, Therefore each distillation 1s pro- gressively more difficult, The boiling point of sodium 1s 1621°F at 1 atm, hence, vacuum distillation 1s used to avolid excessive temperatures. An () 4NP Quar 1439, p 29 Prog Rep Dec 10, 1952, ORNL- absolute pressure of 100 mm Hg, which corresponds to asodium boiling temper- ature of 1285°F, was selected for the distillation. After some 1nitial difficulty with gas leaks, several 5-1b batches of NaK were distilled, and the following observations were made 1. There was no difficulty an completely distilling the Nak. 2. Most of the distillation occurred in one surge 1n each of the four runs, and a part of the condenser was heated to almost the distilling pot tempera- ture. 3. The adjustable probe (a probe wired through a stuffing box of s1licone-rubber “washers’'’) worked excellently, and gave reproducible level readings that checked to within 1/32 an. or closer. NaK has been distilled from ARE pump loop, but because of excessive oxidation of the type 316 stainless steel 1n the loop, a sufficiently low pressure could not be maintained on PERIOD ENDING MARCH 10, 1953 the system for a good test of the efficiency of the distillation process to be obtained. Pockets of NaK re- mained 1n some of the traps, par- ticularly the pump. In the Inconel ARE system, such corrosion should not be encountered, hence, essentially complete removal of NaK by distillation should be effected. Flame Tests for NaK Vapor 1n Helium (D. R. Ward, ANP Division). Tests are being conducted to determine the reliabality of flame tests for detect- 1ng the presence of NaK 1n helium. One method used was to introduce helium, from the system from which NaK was being removed by the distallation process, 1nto an otherwise colorless hydrogen flame, the brightness of the resulting flame was then measured by a phototube circuit. The results of six tests indicate that this method may be used to determine the degree of cleanliness of a system from which Nak has been removed by the distillation method. 37 - 3. REACTOR PHYSICS W. K. Ergen, From the standpoint of the physics of nuclear reactors, the main event during the past quarter was the correlation between the IBM multaigroup calculations and the critical experi- ment on a fairly realistic, reflector- moderated, reactor mockup. Because of the approximations i1nvolved in the calculations, 1t previously had not been clear how close this correlation would be. That the experimental critical mass turned out to be only slightly higher than the computed value and was well within the limits acceptable with respect to uranium investment and fuel chemistry, consti- tutes a milestone 1n the development of this type of reactor. Furthermore, there was, except at the boundaries, good correlation between the computed and the measured flux Thais applies to the flux distribution in space and in energy, but the only experimental indications of the energy distribution are the thermal value and the cadmium ratio, which are effectively only two points on the energy distribution ANP Division curve. Some significant discrepancies exi1st at the boundaries, but the simple diffusion theory employed in the calculations obviously cannot be expected to hold at the i1nterfaces between widely different materials. The nuclear studies pertinent to the reflector-moderated reactor are included 1n sec 4, ‘“Reflector- Moderated Circulating-Fuel Reactor.” As to reactor kinetics, 1t was shown that the delayed neutrons introduce damping of reactor power oscillations, even 1n the case of large initial amplitudes, and, further- more, one typical example indicates that this damping does not interfere destructively with the damping caused by fuel circulation. Also, the damping can be demonstrated i1n a typical case in which the power and flux vary along the path of a circulating fuel. The techniques employed are very similar to those reported i1n the previous ANP quarterly reports and are therefore not repeated here. 39 - ¢+ 3W# 4. REFLECTOR-MODERATED CIRCULATING-FUEL REACTORS A. P, Fraas, ANP Division C. B. Mills, ANP Division A, D, Callihan, Physics Division Homogeneous reactors are the most simple nuclear reactors with respect to 1nternal structure because there are no structural elements 1n the reactor core. Although the con- s1derable experience achieved with liquid fluoride fuels i1ndicates that they do not lend themselves to the design of small, homogeneous reactors because of their poor moderataing properties, 1t has been shown that, by using thick and efficient reflectors, 1t 1s possible to design a small, fluoride- fuel reactor that will be relatively free of structural com- plexities. Further, the combination of such a reactor with a spherical- shell heat exchanger would provide a reactor, heat exchanger, and shield package that would be much smaller and lighter than any reactor now being considered for the nuclear propulsion of aircraft, This section summarizes the work recently carried out on this type of power plant. The first part covers the work done on the static physics of the reactor, including the effect on the reactor of core size and the use of various materials 1n the core and reflector, The second part presents the results of the critical experi- ments and correlates them waith the multigroup calculations. The thaird part covers the mechanical design envisioned and the developmental work 1nitiated to provide a basis for the detalled design of a full-scale reactor. The fourth part covers the shielding work todate, and the last part presents some possible full-scale aircraft power plant arrangements, including engines and radiators, (I)D K Holmes, The Multigroup Method as Used by the ANP Physics Group, ANP-58 (Feb 15, 1951) STATIC PHYSICS C. B. Mills, ANP Division The most simple reflected reactor 1s a small sphere of fissionable ma- terial surrounded by a concentrac sphere of moderating material. The fast neutrons leave the fuel-bearing region easily and diffuse ainto the reflector. If the reflector 1s relatively thick and does not capture the essentially thermalized neutrons, the chain reaction can be supported because of the sufficiently great probability that the low-energy neutrons will diffuse back i1nto the central fuel-bearing region. The fission cross section,as well as the other cross sections 1n the fuel region, 1s so haigh that the slow neutrons re-entering the reactor cannot easily re-escape. This reactor has been briefly evaluated with respect to moderator, fuel, and geometry effects to obtain a qualitative understanding of neutron diffusion and loss. The analysais involved the use of the ORNL-ANP multigroup method,(!? supported by critical experiments. Neutron diffusion 1n a reflector- moderator can be most easily described by referring to the results of the first multigroup calculation for a 32-1n.-dia fuel region surrounded by a 12-1n,-thick beryllium oxide reflector. Figure 4.1 shows the leakage spectrum as a function of lethargy for neutron leakage from the central fuel-bearing region i1nto the reflector. The net current 1s very high across the boundary for high-energy (low-lethargy) neutrons. Most of the fission neutrons move immediately into the reflector, where they are moderated to thermal energy. 41 ANP PROJECT QUARTERLY PROGRESS REPORT DWG 18T7i6 025 020 0 REACTOR 97 (SEE TABLE 4 1) 3 N 4 e”r i |._ 2 y 235 u 251ib OF U g 015 ey (L) vl < I = -0 2509 - THERMAL '% 0 05 b———| ESGCAPE AT r z 17 =18 6\ ) o u'l B (4] < = Y 0 LH —— -005 20 18 16 14 2 10 8 6 4 2 0 LETHARGY, v Firg. 4.1. They then either diffuse back i1nto the fuel region, since the fuel region at the center 1s a sink for thermal neutrons, or out of the reactor across the outer reflector boundary. Figure 4.2 shows the spatial distra- bution of nevtrons at three lethargies. The thermal flux 1s very high in the nonabsorbing berylliumoxide reflector. The lethargy distribution of fissoning absorptions and neutron escape for an assumed neutron source distribution 42 Leakage Spectrum to Reflector vs. Lethargy. ! (Fig. 4.3) 1s given on Figs. 4.4 and 4,5. Intermediate lethargy fissioning processes are 1mportant, but 50% of the fissions results from thermal neutrons streaming into the central region from the reflector. The escape spectrum, which 1s i1mportant for the reactor shield, shows that avery small fraction of fast neutrons escapes from the reflector. The relative values of fast-neutron escape through a l-in.- thick layer of boron carbide for a ¢ L% ) PERIOD ENDING MARCH 10, 1953 DWG 18717 REACTOR 97 (SEE TABLE 41) ¢"vs n 25 |b OF U235 / \ v=18 6, THERMAL 10 v=7 TO 10 v=17015 = \ NEUTRON FLUX FOR AN AVERAGE OF ONE FISSION /cm® sec 5 // ‘\\\\‘--_ 0 2 4 6 8 io 12 14 {6 18 20 22 24 26 28 30 32 34 36 SPACE POINTS,n(r=214 n,cm) Fig. 4.2. beryllaum reflector-moderated reactor as compared with a reactor with a 6- in.-thick beryllium oxide reflector and a moderating core are shown ain Fig. 4.6. The fast escape 1s even smaller for some of the subsequent reflector-moderated reactor designs described below. In summary, the neutrons are born 1n the central region, spend the major part of their life in the reflector, and, after moderation, diffuse into and out of the fuel-bearing region. The prompt-neutron lifetime 1s about 2 x 10°% which 15 a relatavely high value. There 1s a positive component 1n the temperature coefficient of re- activity of this reactor because there 1s self-shielding of the fuel and a sec, Neutron Flux Distribution for Three Lethargy Values. part of the fissions 1s caused by fast neutrons 1n the “Doppler-region,’ The positive component arises because the increased thermal motion of the U??° atoms effectively broadens and flattens out the resonance peaks 1in the fission cross sections, particularly those peaks 1n the high-neutron-energy region. This decreases the self- shielding with respect to neutrons of energles corresponding to the peaks and i1ncreases the reactaivity. Theoretical estimates and critical experiments have thus far failed to guarantee that this positive component of the temperature coefficient of reactivity 1s small compared with the negative component expected from the fuel expansion. However, this reactor 1s noworse with respect to the Doppler 43 ANP PROJECT QUARTERLY PROGRESS REPORT DWG 18718 th3 7z 3 22 20 ® * NORMALIZED TO ONE SOURCE NEUTRON PER ¢m® OF CORE REACTOR 97 (SEE TABLE 4 1) 16 25 |b OF U238 / o8 I ASSUMED e T 77 NUMBER OF NEUTRONS PRODUCED PER cm3® OF CORE % n 06 l CALCULATED [ 0a | fotl— CORE ' REFLECTOR ~——t——1m | 0 2 4 6 8 10 12 14 ie 18 20 22 29 26 28 SPACE POINT, n Fig. 4.3. Spatial Power Distribution. temperature coefficient of reactivity than the other reactor being studied. A small amount of U238 3dded to the fuel or placed 1n the structure near the fuel, would compensate for a positive temperature coefficient, be- cause with U238 3t 1s the absorption that 1s self-shielded, and the decrease in the self-shielding with increasing temperature wouldreduce the reactivity, Very low absorption and very good moderation must be obtained in the reflector. Structural material at the core-reflector interface should also have a low absorption cross section, although the importance of the cross section at the i1nterface 1s reduced by the small probability of several thermal-neutron transits because of 44 the high absorption and low albedo of the fuel region. The structural ma- terial in the fuel-bearing region will compete with the uranium for thermal neutrons, Reactors with Various Reflector- Moderators Computations have been made for a number of possible reactor core si1zes and reflector-moderator compositions, and the results are summarized i1n the following 1. The leakage of fast neutrons from the reflector as a function of the slowing-down power, 523, of the reflector 1s shown in Fi1g. 4.7. A factor of 1/5 1n gzs results 1n a factor of 400 in the total escape of fast neutrons, that 1s, neutrons with lethargies smaller than 4. Figure 4.8 PERIOD ENDING MARCH 10, 1953 &N DWG 18719 012 2 o1 Q REACTOR 97 (SEE TABLE 4 1) 5 010 w N - n VGF = 17 Z 235 = 008| 92% THERMAL 251b OF U al x 007 a % 006 G - a—j 005 |-'1 = | i | ~ 004 2 x 003 Ll a w 002 = =} @ 00t * 0 19 18 17 16 15 14 13 12 1 1O 9 8 7 6 5 4 3 2 1 0 LETHARGY, v Fig. 4.4. Fi1ssion Spectrum vs. Lethargy. (.£ DWG 18720 (=] el = -} L P-4 3 REACTOR 97 (SEE TABLE 4 1) = v = 0 2985 £ & 0007 — THERMAL v 035 x ESCAPE AT 251b OF U o 0006— u=186 [ > -‘"qu [©] € 0005 = I T 1| ly 0004 - . - et Z 0003 — & W 0002 Ll 2 ooo 4 u 0 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 {4 O LETHARGY, v Fig. 4.5. Leakage Spectrum to Shield vs. Lethargy. shows the reflector thickness for a given neutron current escaping from the reflector as a function of the moderating properties of the reflector material, These properties are charac- terized by the age-to-thermal in the reflector material, 2. The absorption in the reactor fuel 1s sensitive to the slowing-down power, £, of the reflector only for 45 ANP PROJECT QUARTERLY PROGRESS REPORT s \ a vtz DWG 18721 4 8X10 — 4 4 CIRCULATING -FUEL ANP REACTOR S 40 (REACTOR 69) — L] “;:; 36 e = =me == Be0 REFLEGTED=-MODERATED REAGCTOR & (REACTOR 129, SEE TABLE 4 1) - g 32 5 £ = 28 S o 1 x 24 L&) Z 20 [ 5 16 2 B -] 12 os 04x10'2 o —— eamam Spmass m— _1|_-_“--‘_fl 19 18 {7 16 15 14 13 M 10 9 8 7 6 5 4 3 2 { o} Frg. 4 6. LETHARGY, v Comparison of the Neutron Current from Two Reactors Into the Bprate%-Water Shield at the 400-Megawatt Power Level. 3 R poor moderators, The absorption de- creases rapidly for &2 < 0,08 (Fag. 4.9), 3. The escape spectrum vs. lethargy 1s shown in Fig, 4.10 for the two best reflectors, beryllium and beryllium oxide. The differences in the spectra are characteristic of the differences in the cross sections and the moderat- ing power of the two materials, Beryl- Iium 1s a much better reflector ma- terial than beryllium oxide. 4. The effect of reflector material and thickness on the escape of fast neutrons 1s shown in Fig. 4.11. 1In order of desirability, the materials are beryllium, beryllium oxide, 6 1in,. of beryllium plus 6 1n, beryllium oxide aggregate, 4 in, of beryllium plus 8 in. of carbon, and carbon, NaOD 1s a good reflector only with regard to neutron escape, The critical mass with a NaOD reflector 1s high (60 1b) because of the sodium ab- sorption, of carbon, 46 5. The effect of the moderator indicates that there 1s little de- pendence of the power distribution on the reflector material in a saimple system. The presence of structure and the effect of operating temperature are expected to favor beryllium as the reflector-moderator, 1f a minimum ratao of peak-to-average power density 1s to be obtained., 6. The effect of reflector thick- ness on reactivity (Fig. 4.12) shows that the beryllium or beryllium oxide reflectors must be over 10 in. thick but that little 1s to be gained by exceeding 16 in, in thickness. Reactors with Beryllium Reflector- Moderators. The results given above, as well as other considerations, indi- cate that beryllium 1s the best re- flector-moderator. Several beryllium reflector-moderated reactors differing 1n si1ze and design were studied, and the reactivity coefficients were estimated, Although the numbers given PERIOD ENDING MARCH 10, 1953 v R DWG 18722 010 W st CARBON oY b " = > o oS g 0 ol 3 & 3 8 l‘ z £ \ [ s £ > ‘ oD o 2 ¥ \ LR e{-kBERYLLIUM OXIDE v e u g \ [« TS 2 - \ < ¥ ooo \ o ©° AY w o AY E F AN 2 AN L N\ ™ § ‘\\3::,—BERYLLIUM S ~~ \\ 000 — - O 02 03 2 Fig. 4.17. Total Neutron Leakage of F1ssion~Energy Neutrons vs, Slowing- Down Power. DWG 18723 30 ‘ CARBON"——mgy 25 ,/' E /,‘-’ DWG 18724 20 ’,/’ g 10 [92] P - : g 35 e—BERYLLIUM OXIDE g F15 /’ w - : & F~—BERYLLIUM z o T e ° 510 50° 3 o - B w h ol o Qo 5 2 ) = S 0 = 0 0 100 200 300 400 0 01 0z 03 04 FERMI AGE {cm?) % Fi1g. 4.8. Reflector Thickness for Constant-Fast Neutron Leakage vs. Age- Frg. 4.9, Neutron Absorption 1n to-Thermal of the Reflector. the Fuel vs. Slowing-Down Power. 47 ANP PROJECT QUARTERLY PROGRESS REPORT OWG 187!5 e (] ut N 3 Z & S g oot " S 0009 THERMAL S e ESCAPE AT BERYLLIUM OXIDE . ° 0008 v=186 (REACTOR 101, TABLE 4 o EJ Be0=033t3 g » 0007 Be = 02982 | / 5 8 0006 Y a l // =~ w ) _I A % 3 - o 0004 I Ll zZ 0003 Th, 8 “—u...._‘_ 5 o002 1 w S z N\BERYLLIUM T 0001 (REACTOR 104, TABLE 4 1) Ifi o | | = 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 LETHARGY, v Fig. 4.10 Neutron Escape Spectrum to the Shield for Beryllium Oxide and Beryllrum. below refer to the specific reactors considered, the general conclusions are believed to be valid for a varaiety of reflector-moderated reactors. 1. The ratio of the peak-to-average power density can be decreased by re- pl'acing the central volume of the fuel with a moderator, The ratio of the peak-to-average power can be reduced to about 1.3, and the critical mass can be reduced to about 1/2 the corres- ponding value for a comparable reactor without a center 1island. 2, The effect of changes 1n various components on the ratio of peak-to- average power density 1s of interest, For a basic design, in which the d1- ameters of central moderator, fuel region, and reflector were 14, 22, and 46 i1n., respectively, and there was 2.8 vol % Inconel in the NaF-UF, fuel- coolant (to simulate structure between 48 fuel and reflector), the relative changes are indicated by the fractions gf%en bel ow. fractional change i1n the parameter noted. The numerator refers to the The denominator 1s the ratio of the peak power density to the average power density and represents the fractional change of this ratio caused by the change indicated in the denominator. PARAMETERS NP/P Amount of U233 in the fuel ———z—-= ~ 0,16 N/M NP /P Fuel layer thickness -——1—-= ~ 0.34 AT/T Struct ght fract /P 0.067 ructure wei raction —— =~ (), No/p AP/P Radius of central moderator —— = ~ 0,117 OR/R OWG 18726 5 \\ %ARBON 2 R\ \ 2 \ =—4m Be+8in C 0 N—A—— < ALY 1 I v AN AN i | 3 g ST \-BeC AGGREGATE v VY 3§ o BeO Py \ W \\\ \ h g \Y‘& \F—Slln Be+6mn C : ° A\ W AN, a \V—A‘NUOD Zz 102 HIGH-DENSITY N o CARBON —— s =) X w A £ . % \ < u. 2 \-—BERYLLIUM o) 4 \‘ \ 5 \ . \\ 1073 (o} 4 B 12 16 20 REFLECTOR THICKNESS (in) Fig. 4.11. Fast-Neutron Leakage vs. Reflector Thickness. 10 REFLECTOR 120 = X 0 < R\REFLECTORS 147 AND 101 REFLECTOR 110 -10 o} 10 20 REFLECTOR THICKNESS (in) Fig. 4.12. Effect of Reflector Thickness on Multiplication Constant. PERIOD ENDING MARCH 10, 1953 The computed values of peak-to-average power density at the two sides and at the center of the fuel-coolant reg:on are tabulated in Table 4.1. 3. The reactivity coefficients of desi1gn interest are, where p1s density, M 1s weight, and T 1s thickness Ak /R For core structure, ——— =~ 0,052 , Do/ Ak/k For fuel layer thickness, —— =~ -0.17 |, AT/T Ak /R For beryllium reflector, ——— =~ (.52 |, bo/p Ak /k For U?35 content, ——JL- =~ 0,22 , /M Ak [k For NaF coolant, =~ 0,106 . Do/p 4, The substitution of 6 or 8 in. of carbon for beryllium 1n the re- flector increases the critical mass by 15 and 25%, respectively, 5. Most of the structural materials that might be used between the fuel and moderator have a high absorption cross section, and therefore some of the thermal neutrons coming from the reflector are absorbed. The loss of all the thermal neutrons would increase the critical mass by a factor of at least 10, A 1/4-in.-thick layer of Inconel might result in a reactivity loss of as much as 0.24, which implaies an 1ncrease 1n Crltlcal mass Of the order of 100%. 6. A negative component of the temperature coefficient of reactivity 1s obtained from the thermal expansion of the fuel, for two reasons (1) the loss of scattering centers for fast neutrons and (2) the loss of fuel from the active volume. If an expansion coefficient for the fuel of 2x10°*/°F 1s assumed, the temperature coef- ficients resulting from these two effects are, respectively, 0.2 X 10°*/°F and 0.4 x 10°*/°F, a total of 0.6 X 10*/°F, The first of these reactivity coefficients 1s smaller for some other reactors (for example, the 49 08 TABLE 4.1. STATIC PHYSICS OF SEVERAL REFLECTOR-MODERATED REACTOR DESIGNS REACTOR FUEL REGION REACTIVITY COEFFICIENTS POWER DENSITY RATIOS ESCAPE CALCU- REACTOR REFLECTOR Outarde | Tnarde | Cmaach Freatal | Thermal Base U3 Mass Over all(®) | Inside | Outside | Mim TOTAL OF FAST LATION TYPE (1fe thick) Diameter | Diameter (15 ) %) Ok/k Ak/k Ok/k Peak Peak- man NEUTRON NEUTRONS(® ) NUMBER o . te to ° ESCAPE o2 10 (2o} Gn ) &8 F) DM/M Ag(°F) Average | Average | Average ¢ Lo Mev) 103 Circulating fuel NaOD 22 14 a 60 40 96 x 109 0 20 126 116 0 90 0 397 0 00228 3 regions 121 Clrcul:(fl.}ing fuel Be(Na cooled) 22 25 45 06 x10° 0 20 -7 x 108 2 16 0 52 0 299 0 0002996 (str €¢)) 2 regions 124 Fuel plates (Ne Be{Na cooled) 19 9 17 k1] 75 %x 108 0 25 119 1 59 0 75 0 264 0 0002596 cooled) 3 regions 117 Circulating fuel Be(Na cooled) 22 14 14 8 59 65 x 10 ® 0 42 -8 x 10§ 118 135 083 0 299 0 000291 (etr )} 3 reglons 118 Carculating fuel De (Na cooled) 19 11 12 58 176 x10° 0 27 114 1 29 0 85 0 335 0 000328 (str } 3 regions 119 Circulating fuel Be(Na cooled) 22 11 17 52 69 x 108 Q26 1 14 129 0 85 0 303 0 000280 (atr )} 3 reglons 120 Carculating fuel Be{Na cooled) 16 ain 22 14 13 60 115 = 10 ¢ 0 25 115 1 29 0 86 0 222 0 0000149 (str ) 3 thick regrons 105 Circulating fuel Be(Na cooled) 22 14 9 65 87 x10°¢ 0 30 -8 x 10 % 114 128 0 86 0 308 0 00029 3 regions 95 Circulating fuel BeQ 16 22 28 44 x10° Q 27 -7x10 % 1 95 0 50 0 447 0 0043 2 regions 97 Carculating fuel BeO 32 18 5 53 -21 %10 ¢ 0 47 -12 x 190 % 17 0 60 0 345 0 0020 2 regions 101 Carculating fuel BeO 22 14 9 4 60 -0 5 x 10 ¢ 0 30 -8 x 10-% 110 1 30 0 8BS ¢ 190 0 0025 3 regions 109 Carculating fyel BeO 48 24 160 40 14 x10° 2 34 211 0 55 (air cooled) 3 regions ~ 108 Fuel plates {Na BeO 22 14 a5 40 19 x10 ¢ 0 20 -9x 105 121 1 69 071 0 352 0 0029 cooled) 3 regions e 115 Circulating fuel BeO aggregate 22 14 15 30 7 x 10 § 0 20 107 121 0 90 ¢ 473 0 01037 3 regions 102 Circulating fuel c 32 20 33 40 -17 x 10 ¢ 0 20 -5x 10 % 13 12 0 90 QO 450 0 040 3 regions £22 Carculating fuel C Be (high p} 22 14 13 57 65 »10°8 0 56 115 129 0 86 0 297 0 00167 (etr ) 3 regions 127 Circulating fuel Be{(4 1n ) C(8 1n ) 16 31 35 408x10% 0 20 1 86 0 52 0 488 0 010744 2 regions 128 Circulating fuel Be(6 1n ) C(6 1n ) 16 25 5 kL] 70l x10°% 0 20 19 0 49 0 440 0 00436 2 regions 129 Circulating fuel Be (No cooled) 19 8 23 47 8 02 x10°* 016 -4x10°* 0 92 1175 0 65 0 343 0 000501 (str )} 3 regions 137 Critical experi- Be 16 10 30 47 100 x 10 ¢ 127 1 00 o 87 0 367 0 001191 ment 20 mal foil end sodium cans 3 regions 140 R 118 with p{NaF) Be(Na cooled) 1% 11 15 36 105 x10 % 0 96 1 34 0 83 0 343 0 0003866 X 1/ 3 regions (a)poes not include the Doppler effect (b)Through 1 1n of boron carbide c)In these calculations, structure was added to the fuel to simulate structure between fuel and moderator LUYOdIY SSTYI0Ud ATHALYVNO I1DAf0Yd dNV ABE¥$%§¢§ factor of 10, but the second 1s approximately the same, 7. The mean lifetime of the prompt neutrons 1s about 4 X 10°* sec for the beryllium-reflected reactor with a center 1sland. About 1/2% of the prompt neutrons has a mean lifetime of the order of 10°? sec because of the low absorption and leakage probability 1n the reflector. 8. A cooling system must be pro- vided for the moderator because of gamma and neutron heating, about 4% of the total power appears in the moderator volume. Insertion of such a system means an i1ncreased amount of structure 1n the reflector, which, an turn, causes a loss 1n reactivity and a gain 1n gamma-ray intensity., This problem 1s being studied. 9, One aspect of importance from the shielding standpoint 1s the energy spectrum of the neutrons escaping from both the surface of the thick reflector and through the fuel-circulation passages at the reactor ends. Thas spectrum 1s strongly thermal for the reflectors of interest. The addition of 1 1n. of boron carbide at the reflector boundary reduces the total leakage through the sides of the reactor to the values listed in Table 4.1 as “escape of fast neutrons.” The use of a nonpoisoning, heavy material (lead or bismuth) as the reflector coolant may be beneficial from the shielding standpoint because 1t should moderate fast neutrons by 1nelastic scattering and also serve as gamma shielding. The fast leakage through the fuel passages and from fissioning 1n these passages must be minimized by suppression of fission and by the addition of extra shielding materaial, This problem 1s not easily adaptable to calculation, and hence craitical experiments will be performed to evaluate leakage control methods. 10. A feasible reflector-moderated reactor can be built by replacing the circulating-fuel coolant (for example, PERIOD ENDING MARCH 190, 1953 NaF-UF,) withuranium-bearing stainless steel fuel plates and sodium coolant. A design 1ncorporating this feature has been proposed by A. S. Thompson. Shim control for such a reactor can be accomplished by varying the concen- tration of potassium 1n the sodium coolant, By this means, the relative burnup could be made very large. These results imply that the goal of a simple, structureless core 1s 1im- practical because the fuel solvent 1s 1nadequate as a moderator. An 1island of solid moderator in the center of the core seems to be desirable, and the usefulness of such an 1sland 1s not entirely nuclear. As will be shown later, the hydrodynamics of a re- flector-moderated reactor are also aided by a central 1sland. Summary Tables and Graphs. The multiplication constants of two- and three-region reflector-moderated reactors as a function of uranium weight in the fuel region for several core and 1sland sizes and moderator compositions are given 1in Fig. 4.13. The various reactors referred taq 1n this figure are described in Table 4,1. The figure summarizes a number of point values for reactivity. To faci1litate the extrapolation, the curves drawn through point values indicate the manner in which the multiplication constant varies with urantum weight for the partaicular reactor to which the point value refers. If only two-region reactors are considered, the shapes of the curves vary systematically with the position of the point value 1n the uranium weight vs. keff plane, Hence, only a few of the curves were actually computed, and the others were drawn by analogy. Likewise, a few keff Vs, uranium weight curves can be computed for three-region reactors and then simi1lar curves can be drawn through the point values. The shapes of the curves for three-region reactors are, of course, different from those for 51 ANP PROJECT QUARTERLY PROGRESS REPORT DWG (18728 12 l//pw4 123 1 97. //}05 101 122 { ‘o / ) V&' ® / «° /108 g 1104 Z // 14 % g 09 //nemfl Q = o — I o -J a 08 |_ - e ) = SEE TABLE 41 FOR DESCRIPTION OF REACTORS BY NUMBER 15 07 103 o6 o 5 10 15 20 25 30 35 URANIUM WEIGHT (Ib) Fig. 4.13. Uranium Weight vs. keff for Several Reflector-Moderated Reactors. the two-region reactors. (Three- region reactors have center 1slands.) The curves described above are not expected to be exact. It should be noted that the IBM multigroup method starts with an assumed power distrai- bution, and, 1fthe calculation results 1n a power distribution different from the assumed one, an 1teration must be made by using the computed distrai- bution as a start, Such 1terations were carried out only where essential. Whenever possible, fuel-coolant con- stituents were held constant to emphasize the effects of main interest. It 1s to be noted that any fuel self- shielding or poisoning effect, or any change 1n reactor size, 1s reflected by a rapid change 1n reactivity and a 52 rapid change i1in the slope of the curves of keff vs. U235 pass, Typical results of the multigroup solution of the neutron diffusion processes are given in Figs. 4,14 to 4,17, Figures 4.14 and 4.15 give spatial power distribution and flux spectra, respectively, for one extreme case — a reactor with a small (8-in.- dia.), central, moderator region, a thick (5 1/2-1n.), fuel-bearingregion, and sodium-cooled fuel plates. The relatively high peak-to-average ratio 1in the power distribution curve, Fig. 4.14, emphasizes the value of a thin fuel layer. Figures 4.16 and 4.17 give the computed power and flux distribution for the first craitical PERIOD ENDING MARCH 10, 1953 DWG 18729 20 | i ¥ let— ASSUMED w 18 | | x S let—CALCULATED w16 " REACTOR 129 (SEE TABLE 41) ©14 o (V8] N \ a2 (& o] \ // ¥NORMALIZED TO ONE SOURCE % 10 \ / NEUTRON PER c¢cm3 OF CORE 2 \ \ // % g Ar=2032 06 @] o @04 BeO } | | 3 e CENTRAL | FUEL -l -l 2 MODERATOR | REGION i BeO REFLECTOR | o) 2 4 6 8 10 12 14 16 18 20 22 24 26 28 SPACE POINT, n{r=nlAr.cm) Fig. 4.14. Spatial Power Distribution for a Three-Region Reactor with a Thick Fuel Annulus. experiment calculation, for which 1- by 3- by 3-in. sodium-filled cans and 10-m1l-thick, 3-1in.-dia, U235 fuel disks were used. A comparison with the experimental results 1s given 1in the next section, CRITICAL EXPERIMENTS D. V. P. Wailliams R. C. Keen J. J. Lynn Physics Division Dunlap Scott, ANP Division Experiments C. B. Mills, ANP Division Computations First Critical Assembly. A pre- liminary craitical assembly of the reflector-moderated circulating-fuel reactor was described previously. (%) Some i1nformation already reported 1is repeated here to givea unified picture of the first reflector-moderated reactor critical experiment., The fuel consisted of 0.020-1n.-thaick pieces of U?3% petal lumped between l-in. layers of sodium canned 1n stainless steel. This arrangement was shown to be 1nefficient 1in neutron utilization because of the self- shielding in the thick uranium layers. The self-shielding of the fuel was measured experimentally by replacing one of the 20-m1l fuel disks with 10 disks that were each 2 mils thaick with aluminum catcher foils between them. The activaity of the catcher foi1ls showed that the 20-mi1l layers 2)p ¥y P Willaems, R C Keen, J J Lyan, D Scott, and C B Mills, ANP Quar Prog Rep Dec 10, 1952, ORNL-1439, p 48 53 ANP PROJECT QUARTERLY PROGRESS REPORT 30 £ DWG 18730 22 u 20 & N\ O REACTOR 129 (SEE TABLE 41) / \ LETHARGY GROUP ./ = 92 . [ 1\ 2] € 16 14 /' \ g, Ar=2032cm \ 2 : / \ D2 o / \ W \ Z {0 O z / \ o 8 N\ % N N=3 ! 3 6 \\ T T— w N=12 Iy 55 S E 4 a N=20 ! 0 l — ‘ ::_= \ 0 2 4 6 8 10 12 {4 16 18 20 22 24 26 28 SPACE POINT, An{r=nar,cm) Fig. 4.15. Spatial Flux Distribution for a Three-Region Reactor with a Thick Fuel Annulus. were 66% effective, the corresponding theoretical estimate was 63%. The critical mass was found experi- mentally to be 15 kg of U?35, With this loading, amultiplication constant of 1.03 was calculated by the multai- group method. The usual self-shielding correction for the lumped fuel was used. The calculation was carraied out for a spherical shape, whereas the actual geometry was rectangular. This difference was accounted for by reducing the fuel volume for calcu- lation purposes by 10%. The computation did not include any poisoning effect of i1mpurities i1n the materials. 54 The neutron-flux traverses measured radially through the mid-plane of the reactor were given previously.(3) The activations of bare- and cadmium- covered-i1ndium foils and their differ- ences were shown. These results are typical of those obtained along other traverses, and they confirm the prediction of high neutron flux in the moderator 1sland and reflector. This effect 1s, as expected, particu- larly pronounced for thermal neutrons. The cadmium fraction, derived from the 1ndium-fo1] activation data and defined as the ratio of the activation (3) Ibid , Fag 5 3, p 51 PERIOD ENDING MARCH 10, 1953 DWG 18731 16 t 4 »* W REACTOR 137 (SEE TABLE 4 1) © / S 12 Q u Aslsumsol / m_ {0 | | g CALCULATED \ 47 "] & osg o S * NORMALIZED TO ONE SOURCE NEUTRON §ose PER c¢cm® OF CORE [+ a 204 o = ar=1757 D Wwo?2 [T o & 0 m = 3 | =2 CENTRAL s | | o MOBERATOR —J-_ COR? et REFLECTOR _7 1 } 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 SPACE POINT, n{r=nar,cm) Fig. 4.16. Calculated Spatial Power Distribution for the First Critical Assembly. by neutrons of energy below the cadmium cut-off to the activation by all neutrons, was given previously.(*) The values of this cadmium fraction, as calculated from the IBM multigroup computations, were also plotted. Theoretical and experimental values practically coincide i1n the center of the fuel and in the bulk of the reflector. The discrepancy at the fuel surface 1s not surprising, since all practical calculation methods, including the "age” multigroup method, are not exact near boundaries, The fissioningdistribution through- out the fuel region 1s shown 1in Fig. 4.18. The experimental values are catcher-foi1l measurements. Since the measurements were not normalized, (4) Ibud., Fig 55, p 53 there 1s no significance in the close agreement of the absolute values of the theoretical and experimental curves. However, the agreement 1in the peak-to-average ratios of power density does seem significant, The discrepancy in the shape of the curves inside the fuel layer can be attributed to the fact that there 1s more moderator in the rectangular center island than was assumed i1in the computation based on spherical geometry. To determine the reactivity loss expected from structural material at the core-reflector 1nterface, an experiment was performed (Fig. 4.19) with one-half of one of the four outside surfaces of the core covered with 1/4-1n. stainless steel plates. The loss 1n reactivity was 220 cents, which 1s consistent with the theoretical 55 ANP PROJECT QUARTERLY PROGRESS REPORT AxT. s OWG 18732 20 m > 18 t REACTOR 137 (SEE TABLE 41) O o '8 oV G \ LETHARGY GROUP, ¥ =92 x4 \\ a 2 L \ TN 2/ W10 \ / \ L) g / \\ (NS o 5 6 N o \\ \\\ fi 5 g yd SN x N=12 N a N=20 — R EXPERIMENTAL 5 (NOT NORMALIZED) = l ] (U] 2 &05 n W w 0 9 10 " 12 13 14 SPACE POINT, 2 (r=1 757 n,cm) Fig. 4.18. Measured Fissioning Dis- tribution 1n the Fuel Region of the First Critical Assembly. : Neutron flux distributions measured with bare- and cadmium-covered-indium foi1ls along a horizontal traverse lyang in the mid-plane of the assembly are shown 1n Fig. 4.22. Note the spatial distribution of the cadmium fraction, that 1s, the fraction of neutrons, detected by indium, that have energies below the cadmium cut-off. Both the experimental values for the cadmium fraction and those computed by the multigroup method are presented. The difference 1s withan experimental error, except at the core-reflector interface where 1t 1s about 40%. The experiments indicate a neutron energy lower than that computed for this point. (Recent data aindicate that this diafference may be caused by PERIOD ENDING MARCH 10, 1953 DWG 18734 225 /) 200 / 175 / 150 1g LOSS OF REACTIVITY (cents) ro (] T o L / o] n 6 '/B 3 16 ‘/4 THICKNESS OF STAINLESS STEEL (i) Fig. 4.19. Reactivity Loss vs, Thickness of Stainless Steel at the Fuel-Reflector Interface of the First Critical Assembly. the void at the interface, which 1s peculiar to the structure of the critical experiment.) The experimental cadmium fractaion curve shown in Fig. 4.21 1s repeated as curve A 1n Fig. 4.23, which gaves the distribution of the fraction of neutrons, detected by indium, waith energies below the cadmium cut-off. Curve B of Fig. 4.23 was obtained from the first critical experiment. Comparison of the two curves shows the similarity of the spectra an the centerof the fuel i1n the two assemblies. It 1s to be remembered that the fuel layer 1n the first mockup was only 3 in. thaick. A preliminary measurement of the neutron leakage spectrum gave values of the cadmium fraction for indium- detected neutrons at the outside edge of the reflector, both on the axais and at the side at the mad-plane. The respective values were 0.4 and 57 ANP PROJECT QUARTERLY PROGRESS REPORT L B & s DWG 18735 - ) BERYLLIUM ---------------------------------------------------------------------------------- i Fig. 4.20. 0 9 A measure of the fission rate or power distribution i1n a direction parallel to the reactor axis has been obtained from the i1mprovisation 1llustrated in Fig. 4.24. Uranium metal disks, 1.4 i1n. 1n diameter and 0.002 in. thick, 1in contact with aluminum disks, 0.005 in. thick, were placed adjacent to one of the fuel containers, and the activity on the o8 Vertical Cross-Section of the Second Critical Assembly. aluminum, caused by recoiling fission fragments, was counted. The low value at the 17-i1n abscissa point probably resulted from the shielding of ex- ternally reflected neutrons by the layer of fuel that was thicker there than at the 20-i1n abscissa poaint, In a similar manner, an attempt was made to determine the fission rate pattern i1n a direction perpendicular Fig. 4.21. PERIOD ENDING MARCH 10, 1953 Photograph of Second Critical Assembly at Section AA of Fig. 4.20. 59 ANP PROJECT QUARTERLY PROGRESS REPORT 4 DWG 18736 - = T . < BERYLLIUM I FUEL - BERYLLIUM - GRAPHITE ™ iISLAND VoD REFLECTOR To 27 RADIAL INDIUM TRAVERSE 30 © BARE INDIUM, EXPERIMENTAL A CADMIUM-COVERED INDIUM, EXPERIMENTAL * EXPERIMENTAL CADMIUM FRACTION 25 A CALCULATED CADMIUM FRACTION " o | oS W T, ‘o s = E o} 2 — 5 < i v c [P g 15 \ L = © = | s (o] < O 10 o 05 s A / [ ] 0 — - o 0 5 10 15 20 25 DISTANCE FROM REACTOR AXIS (tn ) Fi1g. 4.22. Radial Indium Activation and Cadmium Fraction i1n the Second Critical Assembly. | 0 DWG 18737 BERYLLIUM FUEL BERYLLIUM GRAPHITE ISLAND REGION REFLECTOR /o— REFLECTOR 08 = o o 06 & B - w v (o] S 04 ® o 2 T rheo T S 51 O "One 0& ~o 4 \‘“0 /7 02 ‘\% 6 4 2 0 2 8 10 12 14 16 18 DISTANCE FROM CENTER OF FUEL REGION (in) Fi1g. 4.23. Comparison of the Cadmium Fractions i1n the Two Critical Assemblies, 60 PERIOD ENDING MARCH 10, 1953 DWG 18738 BERYLLIUM REFLECTOR x4 L i ~ BERYLLIUM ISLAND -~ 7727777777727/ o 722222 2227277777702 T~ 0 002-in URANIUM DISK AND ALUMINUM CATCHER FOIL ; / AN T H DN FISSION FRAGMENT ACTIVITY {counts per sec x10°3) (9] o n 0 5 10 15 20 25 30 35 DISTANCE FROM REACTCR INTERFACE (in) Fig. 4. 24. Assembly. to the axis along a horizontal line lying 1n the mid-plane. Uranium disks and small (5/16 in. in diameter) aluminum catcher foils were placed between the fuel containers The results show a center-to-edge activity ratioof 0.45, that 1s, a 55% depression in the power across a fuel layer 3.75 in. wide. By direct measurement with 20-mi1l cadmium 1n the first critical assembly, 1t was found that 70% of the fissions was caused by neutrons with energy below the cadmium cut-off. From the similaraity of the spectra, 1t 1s suspected that about the same distribution occurs in the second case. No direct measurement was made. DESIGN CHARACTERISTICS A, P. Fraas, ANP Divaision Although the many complex and interrelated considerations underlyaing Measured Power Distribution Parallel to Ax1s of Second Critical the design of the full-scale reactor shown 1n Fig. 4.25 should logaically precede a detailed description, an exposition of these considerations will be greatly samplified by early reference to a fairly specific system. Therefore, this section begins with a descraiption of the reflector-moderated reactor design and subsequently takes up the more significant of the many factors that have entered 1nto that design. Much work remains to be done on reflector-moderated reactors to give better bases for detailed designs, but the construction i1ndicated by Fig. 4.25seems to be the most promising one considered, to date, for an aircraft power plant. This cross section through the reactor core, moderator, and heat exchanger shows a series of four concentric shells, each of which 1sa surface of revolutzion, 61 ¢9 INCONEL PRESSURE SHELL HEAT EXCHANGER SODIUM 7/ EXPANSICN TANK 7 T SODIUM PUMP FUEL EXPANSION TANK — R CONTROL ROD BaC / CONTROL ROD BaC INCONEL CORE SHELLS . I3 Yg-1n THICK \\ ¢ Bfin FUEL PUMP L4 SODCIUM TG Nak HEAT EXCHANGER NaK OQUTLET NoK INLET 10 6 12 INCHES F1g. 4.25. Three Region Reflector-Moderated Reactor. DWG 4830 LHOddY SSAYI0Ud ATHALYVNAO 12dAf0dd dNV Thé Twd® inner shells surround the fuel region at the center, that 1s, the core of the reactor, and separate 1t from the beryllium 1sland at the center and from the outer beryllium reflector. The fuel cairculates downward through this region in which the fissioning takes place and then downward and outward to the entrance of the spherical shell heat exchanger that lies between the moderator outer shell and the main pressure shell. The fuel flows upward between the tubes 1n the heat exchanger into two mixed flow pumps at the top. From the pumps, 1t 1s discharged inward to the top of the annular passage leading back to the reactor core. The total fuel volume 1n the system for thais design would be approximately 7 ft?3, of which approximately 1 1/3 ft3 would be 1n the reactor core. Most of the remaining fuel would be i1n the 1n- terstices between the tubes 1n the heat exchanger. The moderator was designed to be cooled by sodium flowing downward through the annular space between the beryllium and the enclosing shells and back upward through passages i1n the beryllium. Two centrifugal pumps at the top circulate the sodium first through the moderator and then through the small toroidal sodium-to-NaK heat exchangers around the outer peraiphery of the pump and expansion tank region. A horizontal section through the pump and expansion tank region 1s shown 1n Fig. 4.26. Note that sump pumps with gas seals are used. A pump of this type recently completed 1000 hr of very successful operation in a fluoride system, with pump inlet temperatures of about 1400°F, The praimary construction material 1s Inconel because 1t seems to be the material that 1s most resistant to fluoride corrosion, Beryllium was chosen as the moderator material, partly because 1t seemed to be the best material obtainable, from both shielding and craitical mass stand- points, and partly because 1ts physical PERIOD ENDING MARCH 10, 1953 properties seem to be superior to those of any other material that maght be used in that location, other than graphite. It was decided that the vase-shaped 1sland in the center should be used, partly because 1t reduced the critical mass and improved the power distribution 1n the fuel regiron, and partly because hydro- dynamically 1t promised to give the simplest and most desirable fuel passage The 12-in.-thick beryllium reflector followed by 1 in. of boron carbide was chosen originally to keep the neutrons escaping to the heat exchanger region to a level approximately equal to that of the delayed-neutron flux from the circu- lating fuel in that region. It has since proved to be nearly an optimum configuration from both cratical mass and shielding standpoints. The spherical shell heat exchanger, which makes possible the compact layout of the reactor-heat exchanger assembly, 1s based on the use of tube bundles curved i1n such a way that the tube spacing 1s uniform, irrespective of latitude.®3) The i1ndividual tube bundles terminate in headers that resemble shower heads before the tubes are welded i1n place. Thas arrangement facilitates assembly because a large number of small tube-to-header assemblies 1s made leaktight much more easily than one large unit. Furthermore, these tube bundles give a rugged flexible con- struction that resembles steel cable and 1s admirably adapted to service in which large amounts of differential thermal expansion must be expected. This basic tube bundle and spacer construction was used 1n a small NaK-to-NaKheat exchanger that operated satisfactorily for 3000 hr with a Nak inlet temperature of 1500°F. (%) Two {(5)A P Fraas and M. E LaVYerne, Heat Ex- changer Design Charts, ORNL-1330, Dec 7, 1952 (6)G H Cohen, A P Fraas, and M E LaVerne, Heat Transfer and Pressure Loss trn Tube Bundles for High Performance Heat Exchangers and Fuel Elements, ORNL-1215, Aug 12, 1952 63 ANP PROJECT - - QUARTERLY PROGRESS REPORT D!G 18739 Na TO ISLAND FUEL TO CORE —— N S & > <$ NN Na K QUTLET Fig. 4.26. sodium pumps and two sodium-to-NakK heat exchangers are provided so that failure of one pump or one heat exchanger will not completely disable the reactor. Factors Affecting Core Diameter, The first factor considered i1n es- tablishing the reactor core diameter (the fuel region) was the effect of diameter on critical mass, much work remains Although to be done to establish the core diameter, 1t appears, at this time, that the critical mass 1s essentially independent of the core diameter. A more important factor affecting critical mass 1s the 64 FUEL PUMP IMPELLER —_/ <% q%mmmflfl/ B, FUEL EXPANSION TANK » _~ <> L s““‘ < =0 8 ;52 w owv | n < ud & - > J u- w " | ™ g " Na PUMP OUTLET LAk gt Nak INLET MP OU TEMP :=1100°F TEMP = 1150 °F | / PRESS =100ps! PRESS = 60psi E [ i v - IN " N 7 f ’ Al il“) PRESS SHELL TUBE WALL gggs;uggfl-' i § sk Tsap - 1130°F *IVpst No FROM REFLECTOR FUEL TO PUMPS STRESS = 450 pst TEMP = 1150°F TEMP = {190°F PRESS = 20ps PRESS =10 psi FUEL TO CORE TEMP = t190°F [=] PRESS = 45ps! [=] [+ ) (o] [ .— = [e] [ ] BERYLLIUM FUEL PEAK TEMP =1200°F INCONEL CORE SHELLS Yg-n THICK INNER CORE SHELL TEMP - 1200°F - STRESS = 240 pst [ ECY 1‘1 OUTER CORE SHELL TEMP = {200°F STRESS = 360pst PRESS SHELL TEMP = {500°F TUBE WALL TEMP ={530°F STRESS = 370 pst \ STRESS =100 pst NaK OUTLET TEMP =1500°F PRESS =60psi FUEL TO HEAT EXCHANGER t 0 6 12 TEMP =1590°F INCHES PRESS = 40 ps Fig. 4.27. Temperatures, Pressures, and Stresses Throughout the Reflector- Moderated Reactor Operating at 200 Megawatts. 67 ANP PROJECT QUARTERLY PROGRESS REPORT from the fluoride to NaK. It is expected that, ultimately, six, small, heat exchangers will be built for investigating the effects on endurance life of operation at various tempera- ture and pressure levels. Figure 4.28 shows one tube bundle for this heat exchanger. Six of these bundles will be used in a 5-in.-dia cylindrical annulus instead of in the larger and more complex spherical shell array envisaged for the full-scale reactor. The design temperature level in the sodium circuit for cooling the moderator region is quite tentative. - Tube Bundle for Spherical (a) Tube-to- (b) One Fig. 4.28. Reactor Heat Exchanger. header welded connections. end of a completed tube bundle and “shower head’’ assembly. 68 The temperatures specified in Fig. 4.27 were based on the highest temperatures that appear reasonable from the stand- pointof compatibility of the beryllium- sodium-Inconel system. Higher tempera- tures would be likely to give trouble because of mass transfer and because of attack of the Inconel canning material by the beryllium. The temperature levels required should not be difficult to attain, since the moderator heat would be removed at a high temperature level. Two engines could be fitted with small auxiliary radiators placed between the compressor and the main radiators. These could be employed to subcool a portion of the secondary circuit NaK to, perhaps, 900°F. This cooler NaK could then be fed to the Na-to-NaK heat exchanger at the top of the reactor. There, by cooling the sodium from the moderator cooling circuit, it would be heated to about 1100°F so that it would then flow to the main heat exchanger at the same temperature as that of the NaK from the other radiator circuits. Thus the weight penalty attached to cooling the moderator region would be only about 400 lb of radiator core plus possibly 200 1b of extra reactor shield and 100 1b of pump and line weight. Lead, bismuth, a nonuranium bearing fluoride, sodium, and NaK were all given serious considerationas coolants for the beryllium moderator. The metallurgists felt that lead or bismuth would be likely to pose a serious mass transfer problem. The relatively high neutron absorption cross section of the potassium in the NaK made it quite undesirable from the critical mass standpoint. Rubidium might be used in place of potassium but, because of little demand, it is currently very expensive. Thus sodium seemed to be the best choice for a moderator coolant. Since corrosion and mass transfer are likely to occur in a Be-NaK-Na system, it seemed essential that the beryllium be clad 1n some fashion. Work at Battelle(®’ indicates that beryllium can be chrome-plated to give satisfactory resistance to sodium attack at 932°F, However, this work was carried out on small specimens, and there 1s a question as to whether adequate protection to the very large surface areas required would be practical. An alternate possibility would be to can the beryllium in thain-walled Inconel cans and to fill the small 1nterstices between the beryllium and the can with stagnant sodaum. This arrangement appears to be the more promising of the two, but both possi- bilities are being investigated. The reflector could be constructed of two large hemispheres of beryllium 1f the canning technique were used Cooling passages couldbe rifle-drilled through the beryllium and lined with thain- walled tubes, which could be welded into headers at the ends The Brush Beryllium Co. has indicated that the fabricationof these large hemispheraical shells would probably be no more difficult than the fabrication of large flat slabs The personnel of the Y-12 beryllaum shop state that 1t would not be difficult to rafle- dri1ll holes 3/16to 1/4 i1n. 1in diameter and as much as 40 i1n deep, with the hole diameter held to within 0.001 in. and the hole center location held to within 0.010 inch. Since 1t would be very difficult to plate the 1insides of holes, an alternate construction that appears attractive, 1f chrome-plated blocks prove practicable, 1s the use of a large number of wedge-shaped segments, which would be shaped much like the sections of an orange. These sections could be made with shallow grooves 1n their surfaces that would form passages for cooling streams of sodium 1n the assembly. The choice of a secondary coolant was di1fficult., Careful examination of the components i1n the external system (B)J G Beach and C L Faust, Electroplating on Beryllium, BMI-732 (Apr 1, 1952) PERIOD ENDING MARCH 10, 1953 indicates that about 40 ft® of fluad would be required as a minimum. This means that 1f lead or bismuth were used, the weight required would be an the range of 30,000 pounds. A number of molten salt mixtures was considered, but the melting points 1n all cases were above 300°F, which 1s objectionably high. Other media were rejected be- cause of corrosion considerations. Sodium or NaK seemed to be definitely superior to any other coolant con- si1dered except for the objectionable neutron activation and resultant gamma activity of sodium, and the fare hazard associated with a possible leak. Shielding characteristics indicate that for the most promising distribution of shielding material between the reactor and the crew, gamma activity from the sodium would be less troublesome than the prompt-gamma activity from the reactor core. Con- si1derable experience with sodium has indicated that 1f the plumbing 1s designed to keep the stresses low so that burst types of failure will not result, leaks that do occur as the result of fatigue cracks or corrosion pits develop slowly and do not result 1in serious fires. In fact, saince sodium does not explode (so long as no water 1s present), 1n many ways 1t seems no more hazardous than gasoline. After all the factors were weighed, the most promising secondary fluad appeared to be NaK., A 56% sodium and 44% potassium alloy, which has a melt- 1ng point of 56°F, was chosen for the analysis. Its low melting point, excellent heat transfer characteristaics, good corrosion characteristics 1n 1iron-chrome~-nickel alloys, and low densi1ty made 1t appear to be the best choace. Reflector Heating (A. H. Fox, R. W. Bussard, ANP Division). The heating of the reflector of a reactor by the absorption of gamma radiation and the slowing down of neutrons originating 1n the core presents a heat removal task that 1s only an order of magnitude 69 ANP PROJECT QUARTERLY PROGRESS REPORT smaller than that of cooling the core 1tself., The distribution of the heat generated 1n the moderator depends on the strength and location of the sources and on the attenuation of the radiation between the sources and the point 1n question. For the gamma-ray energies of interest, the mechanism of degradation i1nvolves principally the Compton collisions. The mechanism of degradation of energy of the core gammas 1s complicated, even with the most simple geometry of source and reflector, for amore complex geometry, such as that of the reflector-moderated reactor, many simplifying assumptions are necessary before even an approxi- mate estimate of the rate of heat generation 1in the beryllium may be obtained, The first approximation used 1in- volved the straight-ahead theory of gamma absorption for whichit 1s assumed that Compton collisions merely degrade 1in energy and do not scatter the photons. Only exponential attenuation with a coefficient that 1s charac- teristic of the material through which the photons pass need be considered. When a photon passes from one medium to another, no refraction 1s con- sidered. Use of a buildup factor to allow for ordinary scattering 1s being investigated. Another saimplifying assumption concerned the source distraibution. Only the total amount of gamma flux was considered, and variations from point to point 1n the core were neglected, Thus an average power density for the core was determined simply by dividing the total reactor power output by the volume of the fuel region of the core. (An 1investigation of sources distributed as indicated by the critical experiments showed only a small variation in results.) The particular case considered was that of an 18-1n.-dia, spherical, fluoride-fuel region surrounding a 9-1n.-d1a, central, beryllium 1sland 70 and enclosed by al2-in.-thick beryllium reflector, A 3/16-in.-thick layer of Inconel was considered as being the separating material between the fuel region and each of the two moderator regions. The reactor power output was taken as 200 megawatts, and the energy evolved 1n the gammas was taken as 12 Mev/fission, which gave a total power from gammas of 12 megawatts and a gamma source density i1n the fuel region of 275 watts/cm>. The average gamma energy was assumed to be 1 Mev, and the reciprocal attenuation lengths for the fuel, Inconel, and beryllium- moderator regions were taken as 0.09, 0.30, and 0.16 cm™!, respectively. The power density was computed at points spaced about 1 in, apart along a radius from the center of the island to the outside of the reflector. The resulting values are given 1in Table 4,3 (case A) and Fig. 4.29., Table 4.4 (case A) shows the integrated values of the power, in the various regions of i1nterest, obtained by graphical integration of the power density curve. The fact that the total gamma heating indicated by Table 4.3 1s only 10,6 megawatts 1nstead of the 12 megawatts assumed 1s due, 1n part, to leakage from the reflector and, the approximations made for the computations. However, 1t 1s believed that the distribution of the gamma heating 1s approximated quite well by Fig. 4.29., One important result of this work 1sthat for this configuration more than 60% of the gamma energy goes into heating the fuel. A second computation was made with different values for the reciprocal attenuation lengths, that 1s, 0.06 and 0.13 cm™! for the fuel and beryllium, re- spectively, and the results are shown 1n Table 4.3 as case B, As can be seen, this relatively large change had no great effect on the distribution of the gamma heating. The total power in part, to for both case A and case B 1s given 1n Table 4.4. PERIOD ENDING MARCH 10, 1953 DWG 18740 450 ISLAND FUEL REFLECTOR 400 ri "2 350 [t—INCONEL SHELL— fm——Na ANNULUS —=4 INCONEL CAN ~ 300 L] E (5] ~ @ 3 250 % TOTAL HEATING > }/ TOTAL HEATING tT') \< = / wi & 200 N x w & TOTAL HEATING 150 y HEATING” "\ % L / I / \ NEUTRON /| HEATING y, 100 ’//// \ NEUTRON HEATING #:;fijéiiEAflNG &\ \ y HEATING 50 N --____ I, \ NEUTRON S HEATING o | 0 2 4 6 8 10 t2 14 16 18 20 RADIAL DISTANCE FROM REACTOR CENTER LINE, » (in) Fig. 4.29. The power density resulting from neutron moderation within the beryllium of the 1sland and reflector can be obtained directly from multigroup results by using the flux distribution ¢(r,pn) or (W:/n) C. F. (correction factor).¢!) The energy loss for each lethargy group 1s the average energy loss per collision times the number of collisions i1n that group at a given radius or space point, The spataial distribution, t h W; cE . . £2., 107 Radial Power Density from Neutron and Gamma Heating. 1s normalized to the total power lost by moderation (2 1/2% of reactor power) by the use of the integration operator Q. The power densities resulting from neutron moderation and from gamma heating, as obtained by the methods described above, as well as the resultant total power density, are shown i1n Fig. 4.29. Severe peaks 1in the Inconel shells and a rapid drop with distance from the fuel region are shown, A number of factors must be con- si1dered 1n the design of a system of 71 ANP PROJECT QUARTERLY PROGRESS REPORT s . TABLE 4.3. POWER DENSITY IN VARIOUS REGIONS RADIAL DISTANCE FROM RATE OF HEAT GENERATION CENTER OF FROM GAMMAS REGION REACTOR CORE (watts/cm3) r (im } r (cm) Case A Case B Island beryllium 0 0 47 5 61 9 2 5 08 60 65 3 7 62 80 74 4 10 16 101 93 4 31 10 95 120 142 Island Inconel 4 31 10 95 223 330 4 5 11 4 354 497 Fuel 45 11 4 106 99 5 12 7 167 137 55 14 186 148 6 0 15 2 182 152 6 75 17 1 184 151 8 20 3 157 132 8 5 21 6 141 116 9 22 9 73 65 Reflector Inconel 9 22 9 215 325 9 19 23 3 150 227 Reflector beryllium 9 19 23 3 80 98 9 5 24 1 53 7 57 10 25 4 37 7 44 6 10 5 26 7 25 4 31 9 11 27 9 19 0 24 4 12 30 5 10 3 14 0 13 33 0 5 4 8 3 14 35 6 29 49 15 38 1 1 67 30 21 53 3 0 066 0 19 TABLE 4.4. TOTAL INTEGRATED POWER moderator cooling passages. The volume IN VARIOUS REGIONS TOTAL POWER REGION {megawatts) Case A Case B Island beryllium 0 46 0.51 Island Inconel 0 24 0 29 Fuel 6 71 5 58 Reflector Inconel 0 57 0 85 Reflector beryllium 3 14 3 44 12 of both the sodium and, especially, the Inconel must be minimized to keep parasitic neutron absorptions withain reasonable limits. The neutron flux curves show that from the neutron economy standpolnt coolant passages 1n the beryllium are much more i1mportant several inches from the fuel region than at the fuel-moderator interface. Thermal stresses will be set up by temperature variations in the beryl- lium. Since an elongation of about 30%.ca&n be obtained in beryllium at the operating temperatures envisioned, thermal stresses should not lead to cracking but might cause distortion that could become a problem after a number of thermal cycles of the system. For this reason, the temperature variation i1n the beryllium between adjacent coolant passages was held to 50°F. The pressure drop through the various coolant passages was limited to 40 ps1 to keep pressure-induced stresses low, The maximum beryllium- sodium 1interface temperature was held o o) o Q © © O 0 o O ) o) ?O0o0o C 0 o ) ()OOOO OOOO ) 0 = Q. 28 i B . . O O "-\.\ 2 o O N ::‘-\»-\.9 © ‘\. q00%800 o ISLAND ooooooo 0500 %R OOOOO%%O O.Oe.(2 i INCONEL SHELLS AND Na ANNULUS Fig. 4.30. PERIOD ENDING MARCH 10, 1953 below 1200°F to reduce mass transfer 1in the Be-stagnant-Na-Inconel system. Several detail designs were 1n- vestigated that favored first one and then another of the various require- ments, that 1s, minimum poison, minimum variation 1n beryllium temperature, minimum beryllium-sodium interface temperature, minimum sodium system pressure drop, etc. Figure 4.30 shows the hole pattern in the beryllium for a promising arrangement, and the re- sulting temperature distribution 1is shown i1n Fig. 4.31. DWG 18744 PASSAGE FOR INLET Na TO REFLECTOR (2 REQ'D) 026~ 1n -DIA HOLES IN Be, CONTAINING 025-i1n ~OD TUBES FOR Na COOLANT FLOW Cooling Hole Distribution in Reflector-Moderator. 73 ANP PROJECT QUARTERLY PROGRESS REPORT S DWG 18742 1600 ' ISLAND FUEL 1400 REFLECTOR 7 TEMPERATURE 1200 “nuj'/ /U U\" DISTRIBUTION Um (\f\(\/‘\;/ __ 1000 n y & \\\ < Na COOLANT & HOLES Nao COOLANT HOLES | '.-. % 800 <~ INCONEL SHELL—— &fi e, g i Na ANNULUS ——=it M INCONEL CAN 600 400 200 0 0 2 4 6 8 12 14 16 18 20 RADIAL DISTANCE FROM REACTOR CENTER LINE (n) Fig. 4.31. Temperature Distribution Across Midplane of 200-Megawatt Reflector-Moderated Reactor. SHIELDING A. P. Fraas, ANP Division J. B. Trice, Solid State Division The shield design philosophy that has been the basis of the work on the reflector-moderated reactor shield has been that an operational airplane, 1in the military sense of the word, can be achieved only with great difficulty with the use of a divided shield because a divided shield would give an unprotected man 50 ft from the reactor a lethal dose i1n 15 seconds. Those experienced in aircraft operation and maintenance know that perhaps 50% of the maintenance work 1nvolves non- 74 Sodium 1inlet, 1000°F. scheduled operations, the character and detailed nature of which are 1mpossible to predict without an extensive background of experience. Provision of an adequate set of equip- ment and facilities to take care of all contingencies promises to 1nvolve expenditures of the same order as those required for the nuclear power plant 1tself, It was felt that 1f the radiation dose at full power could be cut to about 10 r/hr at 50 ft from the reactor, the greater majority of the nonscheduled maintenance jobs that might delay a flight at the last minute or require emergency attention immediately following a landing could be handled without special equipment. For this reason, effort has been directed toward the development of something closely approaching a unit shield. It 1s fully realized that a welght penalty 1s inevitable. It was found that the 1nitaial startup and warmup of the reactor can probably be carried out at power levels that, i1n general, do not exceed 10% of the full power of the reactor, thus the radiation dose during startup and warmup would be 10% of that at full power. If in the tuneup process the operations must be carried out close to the reactor, the reactor might be 1dled down from 10% to perhaps 1% of full power output so that the radiation dose at any partai- cular point would be only 1% of the full-power dose., It 1s true, however, that after a long flight the accumu- lation of fission products will constitute a substantial gamma dose even after the reactor 1s shut down completely. This dose w1ll still be only of the order of 1 or 2% of the total radiation dose at full power output because the decay gammas are fewer 1n number and softer than the prompt gammas. A drain valve will be provided at the bottom of the reactor so that the fluoride fuel can be drained 1into underground tanks. It 1s hoped that about 98% of the fuel can be removed by a simple draining operation and that most of the remaining fuel can be flushed out by using a non- uranium-bearing fluoride. Precisely what can be done 1n this direction 1s difficult to predict, but one of the major 1tems of information ex- pected from ARE test work 1s the extent to which the radiation from a circulating-fluoride-fuel reactor can be reduced by draining and flushing. The degree to which the Inconel structure wi1ll be activated and waill represent an i1mportant source of radiation after shutdown 1s difficult to estimate, If the shield 1s not disassembled, rough estimates indicate PERIOD ENDING MARCH 10, 1953 that the Inconel activation should not be a problem. Here again, the ARE w1ll yield much valuable information. The shield for the 200-megawatt reflector-moderated reactor has some characteristics that are peculiar to this particular reactor configuration. As 1ndicated i1n an earlier section, the thick reflector was selected on the basis of shielding considera- tions.{%? The two major reasons for using a thick reflector are that a reflector about 12 in., thick followed by a layer of boron-bearing material w1ll attenuate the neutron flux to the point where the secondary gamma flux can be reduced to an unimportant level for a quasi-unit shell., Thas thickness also reduces the neutron leakage flux from the reflector to the heat exchanger to the level of that from the delayed neutrons ap- pearing 1n the i1ntermediate heat exchanger as delayed neutrons from the circulating fuel. An additional advantage of the thick reflector 1is that 99.98% of the energy developed in the core will appear as heat 1in the high-temperature zone inside the pressure shell. This means that very little of the energy produced by the reactor must be disposed of with a parasitic cooling system at a low temperature level. The materaal in the spherical-shell i1ntermediate heat exchanger 1s about 70%as effective as water for the removal of fast neutrons, so 1t too 1s of value from the shielding standpoint. The delayed neutrons and the decay gammas from the circulating fuel 1n the heat exchanger region might appear to pose a serious handicap. However, each has an attenuation length that 1s much shorter than the corresponding attenu- ation length for radiation from the core. Thus, from the outer surface of the shield, the intermediate heat exchanger appears as a much less (g)A P Fraas, Three Reactor-Heat Exchanger- Shield Arrangements for Use with Fused Fluoride Circulating Fuel, ORNL Y-F15-10 (June 30, 1952) 75 ANP PROJECT QUARTERLY PROGRESS REPORT 1ntense source of radiation than the more deeply buried reactor core. It 1s true that this arrangement violates one of the precepts of the matched shield, namely, that each layer of heavy material should be placed as close to the reactor core as possible, However, the loss in weight associated with the disparaty between this i1deal case and the shield layout of Fig., 4.32 1s not too great, whereas the engineering advantages derived are of crucial importance. A series of Lid Tank experiments 1S 1nh progress to determine the optimum arrangement and thickness of the various beryllium, boron carbide, 1ron, lead, and borated water layers for a shield of the type shown in Fig. 4.32, which represents the best configuration tested to date, When this series of tests 1s completed, the resulting shield will be used as a basis for comparison. Further tests will be run with more unusual materaials, Rough calculations i1ndicate that as much as 10,000 1b of shield weaght mi1ght be saved through the use of some of the special materials. When the test work 1s completed, 1t 1s hoped that a fairly sound basis for a decision on the use of these special materials will be provided by a com- parison of their cost and attendant welght savings relative to the simpler shield configuration of Figure 4,32, One of the major pieces of 1in- formation obtained from the Lid Tank experiments has been the confirmation of the original estimates of the degree of activation of sodium 1n the heat exchanger region. It had been suspected that in a relatively than slab of material having the poor moderating properties of the heat exchanger matrix, neutrons of 0.1 Mev or higher would tend to escape from the slab long before they could be slowed down. Although 1t 1s not possible to simulate the effects of the cir- culating fuel i1n the Lid Tank experi- 76 ments, the spectrum of the delayed neutrons 1s not too different from that of the bulk of the neutrons that escape from the fuel region through the thick beryllium reflector and the heavy boron carbide curtain between the reflector and the heat exchanger. For this reason, 1t 1s felt that the sodium activation data from the Lid Tank experiments give a good basis for estimating the activation of the sodium 1n a full-scale reactor of the type shown 1n Fig. 4.25, On this basais, 1t appears that the sodium activity after a long period of full-power output will give a dose of about 1 r/hr at 50 ft from the reactor. The activaity of the potassium in the Nak would be much less than that of the sodium.(!%) The NaK could be drained after operation and the activated sodium allowed to decay. Several weeks should suffice to bring the activaity of the sodium down to a negligible level, The design described 1s believed to give enough shielding all around the reactor to give a dose of 5 r/hr at 50 ft from the center of the reactor at full power. An additional lead layer would be employed at the front of the reactor to act as a light shadow shield for the crew. Further shadow shielding would be employed at the back of the crew compartment to attenuate the radiation from both the reactor and the NaK radiators to give a dose of approximately 1 r/hr 1nside the crew compartment., Tables 4.5 and 4.6 give detailed data for such a configuration for several reactor core diameters, power out- puts, and power densities. Further work covering the effects on shield weight of various dose levels for both ground and flight crews 1s under way. (1°)w. K Ergen, Potasstium as Coolant tn Con- nectton witth Ground-Safe Shields, ORNL Y-F20-.4 (Dec 6, 1950) LEAD SHIELDING REACTOR PRESSURE SHELL THERMAL INSULATION MAKEUP FUEL TANK VALVE THERMAL INSULATION NaK DUCT 1 ra l”’fll/lll///”l/fll//l”/I///I///// ra THERMAL N _—— 0'6? o7 > N [V o~ L ™ t"‘\a ~ O \X 3 °o =k H—:-}ILL_— - NaK DUCT REACTOR PRESSURE SHELL \\ THERMAL INSULATION LEAD SHIELDING PERIOD ENDING MARCH 10, PUMP QUILL SHAFT CONTROL ROD ACTUATOR ,/PUMP QUILL SHAFT llllllll A A A A A4 Z 7 w7 27 z TUBE BUNDLE SELF—SEALING RUBBER TANK HORIZONTAL SECTION BORATED WATER \\WEB OF CANTILEVER BEAM SUPPORTING REACTOR AND SHIELD 0 6 112 24 | l | - 1 SCALE IN INCHES COOLING PASSAGE BORATED WATER DRAIN VALVE FUEL DRAIN LINE THERMAL INSULATION TUBE BUNDLE SELF-SEALING RUBBER TANK VERTICAL SECTION F1g. 4.32. Shield Configuration for 200-Megawatt Reflector-Moderated Reactor 1953 77 TABLE 4.5. DOSAGES AT VARIOUS LOCATIONS FOR THE SHIELD OF TABLE 4.6 DOSE (r/hr) ® & LOCATION At At Full At Shutdown Startup* Power NaK not NaK Drained Drained 50 ft from reactor (except 1n front cone) 0.5 6 1 0.1 Crew compartment 0.1 0.2 0.02 16 ft from reactor (except 1n front cone) 6 60 10 1 Shield surface (except 1in front cone) 60 600 30 10 *Ten per cent of full power **Fifteen minutes after shutdown from long period at full power TABLE 4.6. PRELIMINARY SHIELD WEIGHT ESTIMATES SHOWING EFFECTS OF VARIATIONS IN REACTOR POWER, CORE DIAMETER, AND POWER DENSITY IN THE FUEL REGION®* Power, megawatts 200 200 400 200 400 800 Reactor core diameter, in, 18 22.7 22,7 28.5 28,5 28,5 Power density, kw/cm? 5.5 2.78 5.5 1,37 2.75 5.5 Pressure shell outside diameter, 1in. 59 62.3 68.8 66,6 72.4 8l.6 Lead layer outside diameter, 1n. 72 75.3 82 79.6 85.4 94.6 Reactor shield outsaide diameter, 1in. 120 119 128 121 129.5 140 Total weaght, reactor, heat exchanger, and shield, 1b 74,000 (84,000 | 108,000 {93,000 (115,000 | 147,000 Shadow disk at crew (7 ft daia), 1b 3,600 3,600 3,600 3,600 3,600 3,600 *Computed from Lid Tank data for the beat Be-B4C-NaF-B4C-Fe-Pb-H20 layer configuration tested up to March 13, 1953 POWER PLANT DESIGN A. P. Fraas, ANP Division Some early work on nuclear power plants 1nvolved detailed designs of power plants in which the engines were widely separated from the reactor. The acute problems of differential thermal expansion, thermal lag and system control, and the plumbing weight associ1ated with these arrangements 1ndicated that i1t would be advantageous to use compact units, Thus, the current trend toward compact power packages, followed by both the General Electric Co. and the Pratt and Whitney Aircraft Division, seems 1in order. 79 ANP PROJECT QUARTERLY PROGRESS REPORT In considering the types of appli- cation of greatest 1nterest, 1t appeared that there are three repre- sentative cases namely, aB-36 flight- test-bed installation, a Mach 0.9 sea- level bomber, and a Mach 1.5 high- altitude bomber. In sketching pre- liminary layouts, 1t appeared that there was little difference 1n the requirements for the first two, and the third differed 1n not too serious a fashion., As a result, the layout of Fig. 4.25 was prepared for the Mach 0.9 sea-level bomber on the basis that 1t would require a reactor heat output of 200 megawatts. presumed that the same power plant might be used for preliminary test work at a reduced power, say 130 megawatts, in a B-36 flight-test-bed installation. Table 4.7 gives the weight of various components of the installation shown in Fig. 4.33. If an 1increase 1n power were made, the weight of the reactor-heat -exchanger-shield assembly It was would i1ncrease as indicated in Table 4.6, and the weight of engines, radiators, pumps, and lines would be roughly proportional to the power A to be required at an altitude at whach the air-swallowing capacity of the engines was cut in half, the turbojet engine weight would be doubled. The weight of the radiators would be increased by, perhaps, 50%, but the weight of the pumps and lines would remain substantially the same. The power plant of Fig. 4.33 represents only a preliminary layout for showing the proportions that might be expected. The thrust that can be obtained from this power plant 1s heavily dependent on the allowable reactor operating temperature. By careful design of the moderator-cooling system, 1t will be possible to keep the temperature of all structural parts down to the temperature of the secondary fluid leaving the intermediate heat ex- changer. This should make 1t possaible to operate such a power plant, con- structed of Inconel, at NaK outlet temperature of about 1500°F for 500 hours. A take-off or war-emergency rating of 1600°F might be permitted for perhaps 20 to 50 hr of the 500 hours. A brief description of the essential features of the design shown 1in Fag, output. If the same power output were 4.33 follows. The reactor 1s the same TABLE 4.7. MAJOR WEIGHT COMPONENTS OF THE 200-MEGAWATT AIRCRAFT POWER PLANT OF FIG. 4.33 NO WEIGHT OF UNITS COMPONENTS (1b) 4 Turbojet engines 12,400 16 Radiators (filled with NaK) 6,000 16 NaK pumps, header tanks, and turbine drive assemblaes 1,600 16 Piping for NaK caircuits 2,000 Thermal insulation for NaK caircuats 1,300 NaK 1n NaK circuits (except in radiators) 2,400 4 Fuel and sodium pump drives 300 Turbojet i1nlet and outlet ducts and support structure 8,000 Total power plant weight excluding reactor and shield 34,000 80 18 e OWo. (8744 FUEL PUMP DRIVE TURBINE e e A Na PUMP DRIVE TURBINE AIR TURBINE NoK EXPANSION TANK WEB OF CANTILEVER BEAM FROM REAR WING SPAR ~\( NoK TO INTERMEDIATE HEAT EXCHANGER (100°F) f— EXTERNAL SHIELD (RUBBER CONTAINER FILLED WITH BORATED WATER) LEAD SHIELD INSULATION REACTOR NaK PUMP NaK TO ENGINES (1500°F) TAIL PIPE NaK-TO-AIR RADIATOR HELICAL BAFFLE ENGINE DATA MODIFIED WRIGHT TURBOJET COMPRESSION RATIO 4:4 (CORRECTED FOR SEA LEVEL) ", AIR FLOW 220 Ib/sec (CORRECTED FOR SEA LEVEL) DIAMETER = 44 Y in. LENGTH =140 in. ENGINE WEIGHT = 3100 Ib (WITHOUT RADIATOR) RADIATOR WEIGHT = 1500 Ib (WITH NaK) REEE 6 COMPRESSOR Fig. 4.33. Aircraft Power Plant (200 Megawatt). ‘0T HOMVW ONIGNYI aorydd €S6T ANP PROJECT QUARTERLY PROGRESS REPORT as that shown i1n Fig. 4.25. The shield was designed so that 1t might be i1nserted through an opening 6 ft 8 1n., wide in the bottom of the fuse- lage. Thus 1t would lend 1tself to installation in the bomb bay of the B-36 just aft of the main spar. Two cantilever beams placed on 6-ft centers were employed as the primary structure, The front ends of these beams could be bolted or pinned to fittings on the rear spar, and the reactor and the shield structure could be attached to the rear ends. The shield structure would be divaided sections, each of which would consist of an i1nner layer of lead and an outer tank of rubber that would be filled with borated water. The top and bottom sections would be separated from the four sections in the central region by surfaces of revolution for enclosing the NaK ducts through the shield. A large, bowl-shaped, rubber tank would be placed at each side. These tanks could be deflated while the power plant was being inserted 1in the airplane and could be filled with water after the power plant was 1in place. The top section of the shield would be penetrated by five tubes. Four of these tubes would carry quill shafts to transmit power from bleed- off air turbines to the pumps at the top of the reactor, and the fifth tube would carry the control rod actuating mechanism. into s1x major Each of the engine radiators would be coupled to a separate tube bundle 1in the i1ntermediate heat exchanger in such a way that there would be 16 separate secondary fluid circuits. The expansion tank, filter, and pump for each of these circuits would be located at the point of greatest elevation and lowest temperature ain the system, that 1s, at the top of the shield and just outside 1t. Thus the NaK would leave the pumps and pass i1nto the shield, through the 1ntermediate heat exchanger, back out 82 through the shield down to the ra%fator, and back up to the pump. Thejpipes passing to and from the radiator would be approximately 3 1/2 in, in diameter, The pressure drop through the radiators would be about 35 psa, and that through pipes fromthe reactor would be about 7 psi. In all cases, the pressure has been kept low to keep the stresses low and, hence, to minimize the likelihood of a burst type of failure., Thus, even though small leaks might occur as a result of fatigue cracks or small corrosion pits 1n welded or brazed joints, 1t seems likely that a major leak (except for gunfire) could be avoided. The turbojet engines indicated in Fig. 4.33 are modifications of a new engine being developed by the Wraight Aeronautical Corp. The original engine 1s a high-compression-ratio two-spool machine. Since fuel economy 1s less important than engine weight for a power plant of the type shown, (11) an estimate of the weight of a modified engine was prepared by the Wraght Aeronautical Corp., and estimated performance curves for these engines are given i1n Fig, 4,34, The radiators used 1n these turbojets are of the same type as those that have been tested at ORNL in the form of core elements during the past year. One of these radiator cores operated for over 1000 hr at 1500°F, or above, including over 40 hr at 1700°F.¢(11) T¢ ;5 expected that developmental work now in process will produce lighter units with better heat transfer performance than the radiators of this preliminary design, The procedure that might be followed to assemble the complete peower plant would be to mount the reactor on the cantilever beams that support 1t and the shield assembly, The turbojet engines would then be mounted to (ll)w S Farmer, A P Fraas, H J Stumpf, and G D Whitman, Preliminary Design and Per- formance Studies of Sodiua-to-Air Radiators, ORNL- 1509 (to be published) AIR FLOW {ib/sec) HEAT CONSUMPTION {kw) TOTAL THRUST (Ib) 320 280 240 200 160 120 80 40 64,000 56,000 48,000 40,000 32,000 24,000 16,000 8000 Fig. . PERIOD ENDING MARCH 10, SRR OWG 18745 I f | { 1 | ! ! { CORRECTED COMPRESSION RATIO= 40 1 ENGINE WEIGHT=3100 {b ENGINE LENGTH =140 n 15,000 ft —— 35,000 ft e |t | 45,000 ft —T 55,000 1 T i RADIATOR WEIGHT =4500 Ib RADIATOR DEPTH =1{2.n No TEMPERATURE THROUGH RADIATOR = 400 °F | - SEA LEVEL " _—-—-"'/ ——’ 15,000 11 | odue—t 35,000 ft |t " | 45000 ft 55,000 ft ENGINE DIAMETER = 44 5 n SEA LEVEL AIR FLOW =220 Ib/sec RADIATOR INLET FACE AREA =16 7 ft2 T [ SEA LEVEL Pl 15,000 f1 I l 35,000 ft 45000 11 | 55,000 f1 o] 03 06 o9 12 15 MACH NUMBER 4.34. Estimated Turboj)et Engine Performance Curves. 1953 83 structure 'attached to these same cantilever beams. The pipes connecting the engine radiator to the reactor and heat exchanger assembly could then be welded i1nto position, along with the NaK pump and header tank assemblies, The entire system could then be pressure checked carefully and examined closely for leaks. It would probably prove desirable to install NaK in the secondary circuit, sodium 1n the moderator circuit, operate the pumps, and check all components carefully for leaks before the thermal insulation was 1nstalled. An auxiliary supply of compressed air would be required to run the sodium and NaK pumps at perhaps half speed for this check. The system temperatures 1in the Nak circuits would not have to exceed room temperature, but an auxiliary burner arrangement should be provided to heat the two radiators coupled to the two sodium-to-NaK heat exchangers for the moderator cooling system. These heaters could be used to bring the sodium system temperature up to 300 or 400°F., Upon satisfactory completion of this test work, thermal insulation could be installed around the reactor pressure shell and the connecting lines and pumps of the sodium circuit, By using the same heat source, the sodium system could be brought up to perhaps 1000°F and a nonuranium- bearing fluoride salt could be pumped into the reactor. The fluoride-fuel system could then be checked and drained, and the lead shielding and the rubber tanks for the borated water could be i1nstalled to complete the assembly. After the power plant unit had been installed in the airplane, the rubber tanks could be filled with borated water, and the necessary connections could be made for the instruments and controls. 84 In starting up the reactor, 1t 1is expected that the NaK circuits would be used to heat the sodium to bring the reactor temperature up to about 1000°F. A fluoride fuel melt con- taining less than the desired amount of uranium would then be poured ainto the fuel system and circulated. Gradual additions of a fluoride melt containing a rather high percentage of U%3% could be made until the reactor became critical. The temperature of the reactor could then be allowed to 1ncrease to the desired value - probably an average temperature of 1300°F - and the load could be gradually increased by starting up first one and then another of the turbojet engines. It will probably be possible to depend upon the self-stabilizing characteristics i1nherent 1in the fluoride fuel reactor to keep the mean fluoride temperature essentially constant so that, i1in effect, the reactor would become a slave to the turbojet engines and would act as a constant-temperature heat source, Control of the individual turbojet engines could be accomplished by varying the amount of air allowed to bypass the radiators through adjustable louvers i1n the helical baffles between In effect, this would give a variable, mean, turbine-air- inlet temperature. Coupled with this control, 1t would probably be necessary to have an adjustable jet exit nozzle to give good performance and to avoid compressor stall over a wide range of operating conditions. Fine control of reactor temperature could be accomplished through the use of one or two control rods in the central 1sland or i1n the reflector region. Coarse shim control could be obtained by varying the uranium concentration in the fluoride fuel, radiator banks. Part 1l SHIELDING RESEARCH INTRODUCTION AND SUMMARY E. P. Blazard J. L. Meem, Associate Physics Divasion With the development of a reactor design 1n which the fissions are confined to a small volume, and an which 1t 1s hoped the radiators can be kept relatively nonradioactive, the hopes for a laght unat shield have been revived. Accordingly, a unit-shield mockup was recently installed i1n the Lid Tank Facilaty and 1s now being optimized with regard to the location of the shield com- ponents. In addition, an intensive search 1s beaing carried out for promising materials that could be used 1n the unit shield. Measurements of relative shielding effectiveness 1indicate lathaium to be a promisaing component for neutron shielding, and inquiries have been 1nitiated to determine whether uranium could be made available and fabricated as a gamma shield (sec. 5). The air-scattering experiments at the Bulk Shielding Facility have been completed. Earlier difficulties have been ascribable primarily to unsuspected background effects, and 1ndications now are that the weights described 1n the report of the Shieldaing Board (1950) were approximately correct. The neutrons were somewhat low and the gamma rays somewhat high, with the weight differences approxai- mately cancelling. The i1rradiataon of animals in the Bulk Shielding Facility 1s now complete, and the animals involved have been returned for long-range observation of the biological effects of these exposures. Neutron spectroscopy on the divaded shield i1n the Bulk Shielding Facilaity has now commenced, and the first few spectra have been measured with the proton-recoi1l spectrometer. The results are only prelaminary, hence, no data are yet available. However, the gamma spectral measurements from the divided-shield mockup are beang tabulated for machine calculations (sec. 6). The Tower Shielding Facilaity design 1s being developed rapidly. The tower structure design 1s final and a building 1s fairly completely laad out. The Reactor design features and some of the instrumentation are yet to be developed. Indications are that the facility will begin operataion by the end of the calendar year, and tests with the first shields will commence shortly thereafter (sec. 7). A neutron spectrometer 1s being developed that utilizes scintillation in a lithium fluoride crystal. This apparatus gives promise of appreciably greater sensitivity than recoil-proton spectrometers 1f developmental diff1- culties can be overcome. No cross- secti1on measurements were completed on the 6-Mev Van de Graaff during the quarter because the machine was being moved to 1ts permanent location an the X-10 Area (sec. 8). 87 5. LID TANK FACILITY J. D. Flynn G. T. Chapman J. N. Miller F. N. Watson Physics Division The Lid Tank Facility has been used during the past quarter praimarily for measuring the fast-neutron removal cross sections of a number of con- ventional and potential shieldaing materials. In addition, a mockup of the unit shield of the reflector- moderated reactor has been assembled. EFFECTIVE FAST-NEUTRON REMOVAL CROSS SECTIONS The following effective fast- neutron removal cross sections have been measured Al 1.19 barns/atom Be 1.12 barns/atom Cu 2.08 barns/atom Fe 1.93 barns/atom LiF 2.80 barns/molecule W 3.08 barns/atom The data on lithium fluoride are particularly interesting. If fluorine 1s assigned a cross section of 1 barn, as would be indicated from a comparison with previously measured values for oxygen and carbon, 1t would appear that lithium exhaibats a cross section of 1.8 barns. Thas is exceptionally high for such a light nucleus, and, 1f verified, reveals a real i1ncentive for using this element in aircraft shields. Accordingly, 1t 1s planned to measure the cross section of fluorine as soon as possible by usingslabs ofa saturated fluorocarbon. Measurements have also been made on the effect of replacing water near the source with a slab of transformer o1l (CH,). By comparison of these measurements with those for graphite, 1t 1s possible to determine the effect of the oxygen that the o1l replaced 1n the water close to the source. This represents the first measurement of the true effective removal cross section for this element. Previous estimates have been dependent upon a somewhat different definition, because the oxygen that was measured was spread throughout the shield as in water It 1s 1nteresting to note that, as might be expected, the oxygen exhibits a somewhat greater effectave removal cross section 1in the locataion near the source. It had been proposed by Sleeper of Brookhaven National Laboratory that deuterium might prove a more effective shi1eld component than hydrogen because of the more nearly isotropic scattering (laboratory system) 1t exhibits because of 1ts i1ncreased mass. To explore this point, a tank of D,0 was inserted next to the source 1n the Li1d Tank Facility, and an effectave removal cross section was measured Results 1ndicate a difference 1n effective removal cross section between normal and heavy hydrogen of 0.1 barn. This corresponds very closely to the difference 1n thear total cross sections 1n the range of about 2 to 5 Mev and indicates that no added premium accrues from the increased angle of scattering or, at least, that this 1s counterbalanced by the greater energy degradation of the light hydrogen. MOCKUP OF THE UNIT SHIELD OF THE REFLECTOR-MODERATED REACTOR Recently a mockup has been installed in the Lid Tank Facility to simulate the reflector heat exchanger and the unit shield of the Fireball reactor. The beryllium reflector 1s simulated by a large slab of material supplied by KAPL plus some additional material 89 ANP PROJECT QUARTERLY PROGRESS REPORT that was available at the Y-12 sate, The heat exchanger 1s simulated by sodium fluoride and aron, the sodium fluoride 1s loaded into large, than, 1ron boxes. The prototype for the shield will be of lead and water, and tests currently under way are designed to determine the optimum location of the lead within the water. 90 An investigation will also be made of the desirability of borating part of the water. These experiments have just begun, and therefore no firm data are yet available, however, there are indications that the over-all reactor- shield weight will be very low. This 1s a direct result of the small volume to which the fissions are confined. PERIOD ENDING MARCH 10, 1953 6. BULK SHIELDING FACILITY J. L. Meem R. G. Cochran M. P, Haydon K. M. Henry H. E. Hungerford E. B. Johnson J. K. Leslaie T. A. Love F. C. Maienschean G. M. McCammon Physics Division The air-scattering experaiments and the program of i1rradiating monkeys have been concluded. In addition, some preliminary neutron spectral measurements have been made, but further work will be postponed untal measurements have been completed on the mockup of the top plug of the SIR shield. The gamma spectral measurements on the reactor part of the divided shield are being applied to calculations of a divided shield. AIR-SCATTERING EXPERIMENTS The air-scattering experiment at the Bulk ShieldingFacility, originally carried out last summer,(!*2?) has been extended and i1mproved 1in an effort to understand the serious discrepancy that appeared to be extant between this experiment and the calcu- lations of the Shielding Board.(?) With the 1increased power now available, 100 kw having recently been approved for the BSR, 1t 1s now possible to eliminate much of the extrapolation previously required. In addition, 1t was discovered that the radioactaivaity 1in the pool water was responsible for a large part of the observed dose 1n the mocked-up crew position. Measure- ments were taken to determine the effect of spurious radiations scatter- 1ng from the pool walls, as well as from the reactor support structure. (L) ANP Quar Prog. Rep June 10, 1952, ORNL- 1294, p 46. (Z)J L, Meem and H, E, Hungerford, Air- Scattering Experiments at the Bulk Shielding Facility, ORNL CF-52+7-37 (July 8, 1952) (3)Report of the Shielding Board for the Aivrcraft Nuclear Propulsion Program, ANP-353, p 64 ff {Oct 16, 1950) These were found to contribute only negligibly to the observed neutron and gamma doses. Although the experiment has been completed so recently that 1t as 1mpossible at this time to give a complete report, the following con- clusions can be drawn 1. The neutron dose 1n the crew compartment appears to be lower by a factor of 5 than the ANP-53 calcu- lations i1ndicate. The advantage accruing from this amounts to about 5000 1b for the standard crew-shield design. 2. The gamma-ray dose in the crew compartment appears to be higher than indicated by the ANP-53 design by a factor of 3.5. The weight penalty associated with this 1s about 6000 pounds. The conclusions differ from the earlier ones both i1n the magnitude of the discrepancies and 1n the amount of material required to make up the added attenuation. Since this experi- ment 1s basically so very crude, further exploitation of the Bulk Shielding Facility in this type of work 1s not considered worthwhile, the present comparison with the ANP-53 calculations 1s considered adequate. IRRADIATION OF ANIMALS In addition to the two groups of monkeys mentioned 1n a previous quarterly report,¢*) a third group has been i1rradiated by the Bulk Shielding Reactor. The complete series of experiments 1s summarized 1n Table 6.1. (4)ANP Quar ORNL- 1375, p 66. Prog Rep Sept 10, 1952, 91 ANP PROJECT QUARTERLY PROGRESS REPORT \ TABLE 6.1. IRRADIATION OF MONKEYS IN THE BULK SHIELDING FACILITY IRRADIATION CONDITIONS SERIES 1 SERIES II SERIES III Exposure rate,* rem/hr 1 0.25 4 Number of exposures 8 16 8 Time per exposure, hr 16 8 16 Time between exposures, days 7 7 7 Total exposure, rem 128 32 512 Number of animals used 12 12 12 *One-half dose in neutrons and ene-half dose 1n gamma rays All exposures have been completed and the animals have been returned to the USAF School of Aviation Medicine at Austin, Texas, where they will be held for observation., It 1s expected that the third group, which received 512 rem, will develop eye cataracts, but that the first and second groups will not If the experiment turns out as anticipated, the threshold for cataracts should be definitely bracketed A report 1s being prepared in cooperation with the Health Physiecs Division that gives details of the dosimetry during the experiment. Figure 6.1 shows one of the animals ready to be placed 1n a watertight cage for submersion 1n the pool. NEUTRON SPECTROSCOPY FOR THE DIVIDED SHIELD Some preliminary neutron spectra have been run with the proton-reco1l spectrometer developed by Cochran and Henry.{(5) Fast-neutron data have also been taken with nuclear plates and threshold detectors. All results are 1n preliminary form and are not vet suitable for reporting. The experiments will be continued as reactor time permits. (S)R G Cochran and K M Henry, A Proton Recoil Type Fast-Neutron Spectrometer, ORNL-1479 {in press) 92 GAMMA SPECTROSCOPY FOR THE DIVIDED SHIELD The data on the energy and angular distribution of gamma rays from the divided-shield mockup are being tabulated for machine calculations. The energy, angular distribution, and total intensity of the direct, as well as scattered, photons arriving at the crew compartment will be computed. FISSION ENERGY AND POWER IN THE BULK SHIELDING REACTOR A report on the power distribution 1n the reactor with a beryllium oxide reflector has been completed, (%) and a report on the determination of the energy released per fission 1s beaing prepared. Figures 6.2 and 6.3 1llus- trate the time decay of neutrons and gamma rays from the reactor after shutdown. The data are normalized to unity for operations at 1 and 100 kw and plotted against time after scramming the reactor. It 1s interesting to note how the (¥,n) reactionon beryllium keeps the neutron level fairly high,. Without the beryllium oxide reflector, the neutrons decay with periods that are characteristic of the delayed- neutron emitters. (G)J L Meem and E B Johnson, Determination of the Power of the Bulk Shrelding Reactor, Part I, OBNL-1438 (an press) PERIOD ENDING MARCH 10, 1953 PHOTO 10829 Fig. 6.1. Monkey and Irradiation Cage for the Bulk Shielding Facility Experiment. 93 14 FRACTION OF FULL POWER DWG 17500 \ | _~GAMMA DECAY AFTER 1-kw OPERATION FOR 2 hr | L 5 [T —— _—GAMMA DECAY AFTER 100-kw OPERATION FOR 16 hr D s 5 —-NEUTRON DECAY AFTER 1w OPERATION FOR 2 hr 10-3 NEUTRON DECAY AFTER 100-kw OPERATION FOR {6 hr 5 o —— 106 0 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300 320 340 360 {hr 2 hr 3 hr 4 hr 5 hr & hr TIME AFTER SHUTDOWN {min) Fi1g 6.2. Decay of Neutrons and Gamma Rays from 0 to 6 hr After Shutdown LU0ddY SSIY90dd ATHILYVNO LIIf0¥d dNV S6 FRACTION OF FULL POWER 107! 107 -« DWG 17504 TIME AFTER SHUTDOWN (sec) Fi1g 6 3 Decay of Neutron and Gamma Rays from 0 to 400 sec After Shutdown - GAMMA DECAY AFTER 1kw OPERATION FOR 2 hr \ D — e P —— e [ — v P~ GAMMA DECAY AFTER 100 kw OPERATION FOR 16 hr \\ \ /NEUTRON DECAY AFTER 1-kw OPERATION FOR 2 hr \ T —— = AY ‘\ N\ \\ AN ~. N J ™~ \ \ NEUTRON DECAY AFTER 100 kw OPERATION FOR 46 hr \\ \ Ti/// o] RAMP =1\ \ HOIST HOUSE JX \ ©, \ ° g bl \ \ Q 8 . ) e A \1 & o / ¢} / X %3 / POOL — } 7 ‘Tl \Q POOL, TOWER, o PAD |0 AND PAD\ " | HOIST HOUSE & 0 / / o o // / - /5’ ; & é@/ // a /// / TOWER STRUGTURE / ‘ E_4 _1/ / // o / ]/ ’/ / [/ / | &= -1-J / |-n; Tj u) 6////// T -—gff///,//>pz_,/i:i 9t // g \96 /t 7101 ft © /S ° /S s S/ /S / ;r_/M J/ //// 30 0 30 60 90 == — __——____— SCALE IN FEET Fig. 7 2. 100 Tower Shielding Facility (Plan View) PERIOD ENDING MARCH 10, 1953 8. NUCLEAR MEASUREMENTS A fast-neutron scaintillation spectrometer employing UF crystals, which have low gamma-ray response, 1s being developed, The 6-Mev Van de Graaff has been set up in the High- Voltage Laboratory, but no research has yet been undertaken in the new location. FAST-NEUTRON SCINTILLATION SPECTROMETER J. Schenck F. J. Muckenthaler Physics Division The work on the development of a scintillation spectrometer for fast neutrons has appeared so promising for shielding work that two research personnel have been assigned to the project to accelerate the work. Although europium-activated lithium 10odide crystals have been used to date, lathium fluoride would be much more desirable because of 1ts lower gamma-ray response. Jhe work 1s now concentrated i1n an effort to grow a lithium fluoride crystal with suitable scintillation properties. MEASUREMENTS WITH THE 6-MEV VAN de GRAAFF H.B. Willard, Physics Divisaion The 6-Mev Van de Graaff was moved during the quarter from 1ts temporary location in Building 9201-2, Y-12 site, to 1ts permanent quarters in the High-Voltage Laboratory, Building 4500, X-10 site. A beam was fairst obtained late in the quarter, but no cross—section measurements were undertaken. 101 Part 111 MATERIALS RESEARCH INTRODUCTION The research on fused fluoraide systems (sec 9) has been largely concerned with the development of techniques for purifying and preparing enriched ARE fuel and fuel carrier The fluoride treatment procedure 1involves successive purging of the liquid with HF, H,, HF, and finally He. The treated fuel then contains only traces of HF, NiF,, and FeF,, which are primary factors i1n fluorade corrosion The beneficial effect on the corrosion of the reducing agents has been substantiated, and the quantity of the various reductants that may be tolerated 1n the fuel has been defined. The consequences of the attendant reduction of some UF, to UF; are being investigated. For the longer-range fuel development program, several ternary and quaternary fluoride systems containing LiF are being 1nvestigated, as well as some binary chloride mixtures containing ucCl, The corrosion rksearch (sec. 10) has been primarilyon the determination of the corrosion characteristics of fluoride mixtures. The 1ncreased attack experlenced 1n recent tests with the ARE fuel mixture 1s believed to be due to faulty purification rather than to handling techniques Zirconium hydride as an additive to the fuel mixture reduces the fluoraide attack on Inconel by a factor of 2 Apparently the mechanism 1s the reduction of the metallic fluorides Ni1F, and FeF;, which are present 1in the fuel, because addition of these fluorides to the fuel 1ncreases the corrosion. Measurements of the concen- tration of NiF,, FeF,, and CrF,; as a function of exposure time of Inconel in the fuel have been made as an 1index of corrosaion The nickel content remains approximately constant and the 1ron content drops sharply to a constant value, however, the chromium AND SUMMARY content rises during the first 100 hr and then levels off After the addition of ZrH,, the only significant change 1n these concentrations was 1n the chromium content, which then changed 1n the same manner as the 1ron content. The addition of chromium metal was not effective i1n inhibiting fluoride corrosion, apparently because of the low solubilaty of chromium 1in the fluoride Although Inconel and stainless steel are severely attacked by lead 1n tests 1in convection loops, molybdenum and columbium are relatively 1mmune However, static tests of lead-sodium alloys show that sodium additions decrease the otherwise severely corrosive action of the lead on both Inconel and stainless steel The stability of BeO 1n NaK 1s stall being investigated. It appears that there 1s some solubility of BeO 1in NaK The metallurgy and ceramics research (sec 11) includes the fabrication of various reactor components for use at high temperatures, creep rupture tests of structural metals, the development of special alloys, cermets, and ceramic coatings, and tests of welding and brazing alloys Control or safety rod inserts are being prepared for the G-E reactor and for the Tower Shielding Facilaity Chrome-plated, high- conductivity fins have not proved adequate for high-temperature radiator use, but Inconel or stainless-steel-clad copper fins may now be satisfactorily brazed. Recent creep tests of stain- less steel and Inconel have shown that surface oxidationis not the controllang factor i1n the longer rupture times observed 1n air as compared with the rupture times 1n hydrogen and argon The cone-arc welding technique 1s being adapted to the welding of dished headers for heat exchangers because of the unequal arc distances encountered. Because of dilution and embrittlement of the Nicrobraz used 1in 105 ANP PROJECT QUARTERLY PROGRESS REPORT brazing a heat exchanger element other alloys have been considered for this application. The evaluation of high-temperature alloys with respect to temperature, strength, and corrosion resistance 1s continuing The physical property measurements were made primarily for the determi- nation of various properties of several fluoride mixtures, and the heat transfer studies were concerned with experimental systems, as well as aircraft components (sec. 12). The heat capacity of the ARE fuel was found to be 0 31 cal/g*°C from 550 to 850°C. Vapor pressure measurements of the enriched fuel mixture were lower than expected and i1ndicated the formation of complex 10ons at high temperatures. Preliminary measurements were made of both the thermal conduc- tivity and the viscosity of sodium hydroxide at elevated temperatures. A minimum-weilght analysiswas completed for an aircraft radiator with the optimum fin spacing and thickness and the optimum tube spacing and diameter Experimental apparatus to determine the fluid velocity profiles and the fluid temperature structure 1s being assembled One study 1s under way to determine temperature distrabution 1in entrance regions, whereas another study will compare the various high- temperature heat transfer fluids. Creepunder 1rradiation and corrosion by sodium 1n an in-reactor loop are 106 being investigated, however, 1irradiation of fluoride mixtures comprises the greater part of the radiation damage studies (sec 13) In recent 1r- radiations of the ARE fuel mixture 1in the MTR, i1n which the power generation 1s 20 times that expected in the ARE, corrosion and changes 1n the fuel composition occurred that were 1n excess of those observed 1n control tests. However, the Inconel and the fuel were at temperatures 1n €XcCess of 1500°F during the tests The recent i1n-reactor creep measurements confirmed previous conclusions that irradiation had little effect on the creep strength studies of reactor materials (sec 14) included chemical, petrographic, and x-ray diffraction 1dentification of impurities, corrosion products, reduction products, and constituents of reactor fuels Volu- metric methods for the determination of zirconium 1n the presence of uranium and for the simultaneous determination of uranium trifluoraide and zirconiummetal have been developed. An apparatus has been built to aid in the determination of traces of oxides 1n fluorades With petrographic and x-ray-diffraction studies, 1t has been possible to determine and define compositions of the compounds and eutectics present ain UCl,-NaCl, UC1,-KCl, and NaF-ZrF,-UF, systems. Analytical PERIOD ENDING MARCH 10, 1953 9. CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS W. R. Grimes, Materials Chemistry Division The major effort of the ANP Chemistry group continues to be devoted to the study of fluoride mixtures for use as aircraft reactor fuels and coolants., A large fraction of this effort has been devoted to the specific materials that comprise the fuel solvent, the concentrated fuel, and the final fuel mixture for the ARE. The final processing plans for preparation of the fluoride mixtures 1n pure form are nearly finished Progggsires for removing all except traces of HF, N1F, and FeF, from these materials are in the final develop- mental stages. Contrary to previous beliefs, hydrogen does cause partial reduction of UF, to UF; 1n these melts at 800°C. The beneficial effect of reducing agents on the corrosion of fluoride mixtures has been well sub- stantiated, and the quantities of various reductants thatmay be tolerated 1n the fuels has been defined. Phase equilibrium and other studies of systems 1n which all or part of the UF, has been replaced by UF; are still being carried out as a part of the study of the effect of added reducing agents and as an aid 1n the 1dentifi- cation of corrosion products Adda- tional materials have been synthesized that contain trivalent uranium and seem to be 1dentical with species produced during corrosion. In the hope of developing other and superior fuel mixtures for future applacations, an extensive study of phase equilibria 1n systems containing lithium fluoride 1s being carried out. Some study of UCl, systems 1s also 1n progress FUEL MIXTURES CONTAINING UF, L. M Bratcher C. J. Barton Materials Chemistry Division L1F-ZrF,-UF,. The L1F-ZrF, -UF, system was investigated with UF, con- centrations of up to 45 mole % and ZrF, concentrations of up to 50 mole %. The lowest melting point observed was 440°C for the mixture containing 68.5 mole % L1F, 26.5 mole % UF,, and 2.0 mole % ZrF,. Thermal halts were noted with this composition, as well as with other compositions nearly the same, at about 400 to 415°C, however, 1t cannot be ascertained definitely from the data that have been obtained whether this effect 1s due to a eutectic of unknown composition or to a solid transition KF-LiF-BeF,-UF,. The effect on the melting point of varying the UF, con- centration of one composition 1n the KF-Li1F-BeF, -UF, system has been determined. The data show that when UF, 1s added to the 40% BeF,, 20% KF, and 20% LiF mixture the melting point goes through a minimum at a low con- centrationof UF, and then rises rather rapidly wath increasing UF, concen- tration. The thermal effect noted at 330 to 340°C may indicate the existence of a low-melting-point eutectic 1n the four-component system, a great deal of work will be required to determine the composition of the eutectic. The lowest melting tempera- ture observed to date 1s 380°C at 2.5 mole % of UF,. FUEL MIXTURES CONTAINING ThF4 L. M Bratcher C. J. Barton Materials Chemistry Divisaion Other 1nvestigators{(!) have shown that the LiF-ThF, equilibrium diagram 1s quite similar to that of the LiF-UF, system, and unpublished data obtained 1in this laboratory indicated that the BeF, -ThF, system 1s similar to the BeF,-UF, system Since the LiF-BeF, binary 1s common to both systems, 1t (I)J O Blomeke, An Investigation of ThF4- Fused Salt Solutions for Homogeneous Breeder Reactors, ORNL-1030 (June 19, 1951) 107 ANP PROJECT QUARTERLY PROGRESS REPORT seemed reasonable to assume that the Li1F-BeF, -ThF, equilaibrium diagram would be comparable to the diagram for the LiF-BeF -UF, system described an the previous report.{(2?) The data shown 1n Fig. 9.1 confirm this ex- pectation. The higher melting point of the L1F-ThF, eutectic (550°C com- pared with 490°C for LiF-UF,) 1s reflected 1n the contours. A com- position containing 5 mole % ThF,, 67 mole % LiF, and 28 mole % BeF, melts (2)L M Bratcher and C. J Barton, ANP Quar Prog Rep Dec 10, 1952, ORNL-1439, p 105 -~ at approximately 430°C., From the limited data obtained with the lower ThF, concentrations, 1t appears that compositions i1n this system that melt below 400°C will contain a low con- centration of ThF, and a high concen- tration of BeF,. FUEL MIXTURES CONTAINING UCI4 R. J. Sheal C. J. Barton Materials Chemistry Division The problem of obtaining pure UCl, has not yet been completely solved. Efforts to prepare the pure material . W \ s Lo \ 5 ‘0 ., \G DWG 18746 }&\ Y \& 1114 °C \"x g \ 1} 30 X (¥} U\T- ‘" . A M X 3\‘-1;’- .~ u‘/ - 1000°¢ r-". »,\.lb '% i ?y- s \D ’ ‘\‘{ o ' \ \Va J/a/ 950 900 850°C g00™° 550°C 50T LiF 543°C Fig. 9.1, 108 LipBeF, 845 475°C The Systenm LlF-Ber-ThF4. by sublimation from the available crude UCl, have not been successful. Synthesis of UCl, by liquid-phase chlorination with hexachloropropyl- ene{3) has been used to provide material for recent studies. Although analytical data from which to evaluate this material are not yet available, 1t appears that the material 1s not of good quality The data reported briefly below may therefore be revised when purer UCl, becomes available. NaCl-UCl,. Thermal data on 18 com- positions i1n the NaCl-UCl, system have been reported by Kraus.?4) He con- cluded, on the basis of the data obtained from cooling curves, that the compound 2NaCl-UCl, was the only compound formed 1n this system and that 1t melted 1ncongruently at 430 £+ 5°C. It also appeared from hi1s data that there was only one eu- tectic, which melted at 370 * 5°C. The present,lncomplete investigation seems to show the existence of two eutectics and two compounds in this system. The first eutecticis believed to beat about 30 mole % UCl,, and 1t appears to melt at 420 %+ 10°C. One of the compounds, Na,UCl,, 1s believed to melt congruently at 440 * 10°C. The second eutectic 1s believed to be at approximately 45 mole % UCl,, and 1t appears to melt at 370°C, 1n good agreement with Kraus’' results. In the 65 to 90 mole % UCl, range, a series of breaks at 420 + 10°C probably represents the i1ncongruent melting point of a compound such as NaU,Cl, or NaU,Cl,, The pure compound has not yet been prepared, and 1t cannot be definitely 1dentified on the basais of information available at this time. Kcl-Ucl,. Thermal data on 14 com- positions i1n the KCl-UCl, system, 1n the 20 to 75 mole % UCl, range, have also been reported by Kraus.(*) These (3)B M Pitt et al , The Preparation of TCl4 vith Hexachloropropylene, C-2 350 3 (July 27, 1945) (4)C A Kraus, Phase Diagraa of Some Complex Salts of Uranium with Halides of the Alkal:i and Alkaline Earth Metals, M-251 (July 1, 1943) PERIOD ENDING MARCH 10, 1953 data showed that there was one com- pound, K,UCl ., which melted congruently at 650 + 5°C, and two eutectics 1in the system. The KC1-K,UCl, eutectic was stated to be at 25 mole % UCl,, and 1tsmelting point was given as 5601 5°C, the K,UC1 ,-UCl, eutectic was 1ndicated to be at 53 mole % UCl,, and 1ts melting point was 350 + 5°C. Thermal data and petrographic examination of the soli1d phases show strong evidence for the existence of three compounds and three eutectics 1n this system. The present results are in fair agree- ment with those of Kraus on the location and melting point of the KC1-K,UCl, eutectic, the melting point of K,UCl,, and the melting point of the higher eutectic. However, 1t appears that a K,UCl,-KUCl, eutectic with 42.5 £ 2 5 mole % UCl, occurs 1n this system and that 1t melts at 320 £ 10°C The compound KUCI, probably melts congruently, but a reliable value for 1ts melting point has not yet been obtained. The thermal evidence for the existence of a com- pound with more than 1 mole % UCIL, per mole of KCl 1s a series of breaks 1in the 70 to 85 mole % UCl, region at 400 + 5°C. The thermal effect at thas temperaturemay indicate the 1incongruent mel ting point of this compound of unknown composition. As mentioned 1n the section on NaCl-UCl,, and 1in line with experience with fluoride systems, 1t has been found rather difficult to prepare incongruently mel ting compounds in sufficiently pure state to permit their positive 1dentification. The composition of the KUCl.-KU Cl, ., eutectic has not been accurately determined. FUEL MIXTURES CONTAINING UF, V. S. Coleman C. J. Barton Materials Chemistry Division Work on the NaF-UF3 and KF-UF3 systems has not yet revealed com- positions that melt below 700 and 680°C, respectively. Complications 109 ANP PROJECT QUARTERLY PROGRESS REPORT from UO, and UF, 1n the melts have been minimized byusingalkali fluorades that have been treated with HF at high temperatures and by minimizing the top temperature to which the melt 1s heated to decrease disproportiona- tion of the UF,. These systems are present, studied, at in nickel capsules welded under an argon atmosphere. The examination 1s hampered because the KF-UF, compounds decompose readily in the laboratory air. Petrographic examination 1s accomplished by grinding the material i1na dry, inert atmosphere and transferring 1t to microscope slides under o1l. Although the thermal data are as yet 1nconclusive, 1t appears that the compound formed 1n NaF-bearing melts 1s 3NaF-2UF,. ADDITION OF REDUCING AGENTS TO ARE FUEL J D. Redman V. S. Coleman L. G. Overholser D. C. Hoffman K. J. Kelly F. F. Blankenship Materials Chemistry Division Decreased corrosion of Inconel 1in thermal convection loops has been demonstrated to result from small additions of NaK, ZrH,, or Zr° to fluoride mixtures of i1nterest as ARE fuels., When these reducing agents are added 1n large quantities to materials containing UF,, they produce UF,, which 1s only sparingly soluble in the melt at 600°C. Therefore the reactions of possible ARE fuels with various reducing agents have been studied i1in an effort to discover a safe upper limit to the amount of such additives. Previous experiments had demon- strated that addition of sufficient reductant resulted i1n separation of UF, as a discrete phase, which could be removed by filtration at 600°C. The resulting filtrates revealed no trace of the easily detectable UF, after solidification, but they showed considerable quantities of various 110 complex compounds of UF,. Identical mixtures (with the same reducing agents) that were not filtered showed considerable quantities of UF,. If a sufficiently small quantity of reducing agent was added so that no UF, could be filtered at 600°C, there was no UF, detected in the cool, solidified material Various ZrkF,-bearing mixtures, with and without added UF,, have been equilibrated with predetermined small quantities of NaK (78 wt % K), ZrH,, and Zr° 1n sealed capsules of Inconel by rocking for 16 hr at 4 cpm with the ends of the capsules at 800 and 650°C, respectively. The capsules were then heated at 800°C 1n a vertical position for 2 to 6 hr to permit settlaing of insoluble material and, especially, unreacted reducing agent., Cooled samples were sectioned longitudinally for petrographic examination. The upper tolerance laimits an number of equivalents (one equivalent 1s the quantity of reductant needed to reduce all UF, to UF,) are listed in Table 9.1. Reducing agents added to NaZrF; 1in amounts 1n excess of 0.2 wt % showed opaque nonmagnetic masses, and chemical and x-ray tests indicated the materaial to be, at least in part, Zr°. Appar- ently, trivalent zirconium does not exi1st 1n fluoride melts under these circumstances. Verificationof the tolerance values for ZrH, 1n the ARE fuel mixture has been obtained by equilibration of ZrH, with the fuel mixture at 800°C 1n nickel equipment, followed by equila- bration at 600°C and filtration of the resulting melt at 600°C through sintered nickel. The data obtained by chemical analysis of the filtrate and by petrographic examination of the filtered residue are shown in Table 9.2, The data indicate that no solad materi1al exists i1n the liquid at 600°C 1f less than 0.9 wt % ZrH, as added to the mixture. Agreement between the data shown 1s probably within the error of such experiments. The failure to observe UF,, as such, in any of the filtrates affords con- firmation of the assumptions used to interpret the previous data. When small quantities (up to about 0.4 eq) of reducing agent are added, an olive-drab phase (average refractave index of 1.558) occursin the NaF-ZrF, - UF, mixtures. Reducing agents added to compo- sitions 1n the NaF-ZrF,-UF, system produced an olive-drab phase (average PERIOD ENDING MARCH 10, 1953 refractive 1ndex of 1.558) at the 0 2-eq concentration. The quantaty of olive-drab phase generally increased upon addition of 0.4 eq of reducang agent, With larger additions of reducing agent, a decrease 1n the olive-drab phase was observed, along with an i1ncrease inthe red-orange phase (UF,-2ZrF,). Synthesis and petro- graphic examination have i1ndicated that the olive-drab phase may be UF3-2ZrF4, in which tetravalent uranium has been partially substituted for TABLE 9.1. AMOUNTS OF REDUCING AGENTS TOLERABLE IN FLUORIDE MIXTURES AMOUNT TOLERABLE FLUORIDE USED NaK ZrH, 2c® Equivalent Weight % Equivalent®| Weight % Equivalent Weight % NaZrF5 02 02 02 NaF-ZrF4-UF4 (46-50-4 mole %) 06 07 12 10 12 09 NaF-ZrF4-UF4 (50-46-4 mole %) 0.4 035 08 07 12 10 *Hydrogen considered neutral in calculation of equivalent weight, TABLE 9. 2. DATA OBTAINED BY CHEMICAL ANALYSIS AND PETROGRAPHIC EXAMINATION OF ARE FUEL®* AFTER TREATMENT WITH ZrH2 LRANIUM CONTENT er2 ADDED OF FILTRATE PETROGRAPHIC OBSERVATIONS (wt %) (wt %) 0 8.62 0.2 8.72 Uranium reduction occurred, 0.5 8.68 no discrete UF; observed 0.7 8.65 0.9 6.58 UF; found on filter and in 1.0 7.12 reaction chamber 1.5 5.95 1.5 8.64 Presence of Zr0O, indicated an oxygen leak ‘NaF-ZrF4-UF4 (50-46-4 mole %). 111 ANP PROJECT QUARTERLY PROGRESS REPORT trivalent uranium. Petrographlc examinations revealed that UF,-UF,- 4ZrF, was an olive-drab, pleochroic phase, with an average refractive 1index of 1.564, a birefringence of about 0.008, and an optic angle of 80 degrees. Another olive-drab material (refractive index of 1.574) was obtained when these materials were mixed to correspond to UF,-3UF,- 8ZrF,. A mixture corresponding to 3UF,-UF,-8ZrF, produced striated crystals of olive-drab and red-orange materials. The 50 mole % NaF composition developed free UF, upon addition of 1.2 eq of Zr®° or ZrH,. The 46 mole % NaF composition formed free UF; upon the addition of 1.6 eq Zr° or ZrH,. The usual olive-drab and orange-red phases were observed 1n these com- positions with less than 1.2 eq of reductants. It was noted throughout this study that the 46 mole % NaF, 50 mole % ZrF,, 4 mole % UF, composition toler- ated slightly more reduction without precipitating UF, than did the 50 mole % NaF, 46 mole % ZrF,, 4 mole % UF, composition. Further, 1t was noted that both compositions are more tolerant to Zr° and ZrH, than to NaK. These observations support the mecha- nism suggested previously, (%) which postulates that NaF and UF, compete for the ZrF, present. The Na(K)F formed by the NaK reduction reaction 1s capable of breaking the UF,:2ZrF, complex to form Na(K)ZrF, and free UF,. On the other hand, ZrF4 formed by the Zr° or ZrH, reduction reaction 1s capable of complexing more UF,. This suggests that a fluoride com- position with a ZrF,-to-NaF ratuio larger than unity would be desirable 1f only because of 1ts capacity to complex UF,. (S)F F Blankenship, D C Hoffman, K J Kelly, and T N McVay, ANP Quar Prog Rep Dec 10, 1952, ORNL-1439, p 119 112 COOLANT DEVELOPMENT .. M. Bratcher C. J. Barton Materials Chemistry Division LiF-ZrF,. Study of the LiF-ZrF, system, mentioned 1n the previous progress report,(s) has been continued by thermal analysis andby petrographic and x-ray-diffraction examination of solidified melts. Although Li,ZrF 1s the only compound that has been positively 1dentified, there 1s some evidence for one additional compound with less than 33 mole % ZrF, and one with more than 50 mole % ZrF,. CsF-7ZrF,. Both Cs,ZrF, and CsZrF, have been 1dentified by petrographic examination of melts in the CsF-ZrF, system, and three eutectics have been shown to exist. The lowest-melting- point (412 * 10°C) eutectic contains about 40 mole % ZrF,. Further work 1s 1n progress to complete the equi- librium diagram for the system. KF-L1F-ZrF,. The KF-LiF-ZrF, system has been examined over a ZrF, concen- tration range of 5 to 60 mole %. A ternary eutectic melting at about 385°C appears close to the KF-ZrF, eutectic. The exact composition of this eutectic has not yet been estab- lished. RbF-L1F-ZrF, Melting-point deter- minations have been made for com- positions in the RbF-LiF-ZrF, system that have 5 to 60 mole % ZrF,. The results were similar to those obtained with KF-LiF-ZrF, mixtures. The lowest melting point shown to date was 370°C for amixture containing 45 mole % RbF, 45 mole % ZrF4, and 10 mole % La1F, this composition 1s probably close to a ternary eutectic. Thermal effects at about 550°C were observed with a number of compositions i1n this region. These effects are believed to be due to the presence in the fused mixture of complex oxyfluorides that may result from the hygroscopic nature of the RbF. (6) M Bratcher and C J Barton, ANP Quar Prog Rep Dec 10, 1952, ORNL-1439, p 113 KF-LiF-BeF,. The data obtained thus far with compositions 1n the 4 to 57 mole % BeF, range of the KF-LiF-BeF, system seem to indicate that only slight lowering of the melting point results from the addition of KF to LiF-BeF, mixtures. The lowest melting point that has been confirmed by both heating and cooling curves 1s 350°C for the mixture with 54 mole % BeF,, 10 mole % KF, and 36 mole % LiF, and 1t 1s only slightly lower than the melting point (365°C) of the 50 mole % L1F, 50 mole % BeF, composition. PRODUCTION AND PURIFICATION OF FLUORIDE MIXTURES F. F. Blankenship G. J. Nessle Materials Chemistry Division H. W. Savage, ANP Division Study of the hydrogenation-hydro- fluorination procedure previously described¢’) for fuel preparation has been continued on laboratory and pilot scales. Operating conditions to provide optimum purity of the molten mixtures are now reasonably well defined. Construction of the production- scale equipment for producing 3000 ib of ARE fuel solvent (NaZrF.) has been delayed to some extent by lack of manpower. This equipment should be avallable for testing early in Aprail, and the material should be prepared during May and June. All raw materials for this operation have been received, and chemical and spectrographic analyses of each batch indicate that the materials are of satisfactory purity. Laboratory-Scale Fuel Preparation. (C. M, Blood, F. P. Boody, R. E. Thoma, Jr., Materials Chemistry Division). Corrosion by selective removal of chromium from Inconel 1n fluoride melts may occur by any combination of several mechanisms. Three of the (TYF F Blankenmship and G J Nessle, ANP Quar Prog Rep Dec 10, 1952, ORNL-1439, p 122 PERIOD ENDING MARCH 10, 1953 most likely of these processes are Cr® + 3UF, < CrF, + 3UF; , (1) 2C:° + 3HF < 2CrF, + 3H, (2) and 2Cr° + 3N1F, < 3N1° + 2CrF, 2Cr® + 3FeF, < 3Fe® + 2CrF, , (3) Cr® + FeF, < Fe° + CrF, Addition of the strong reducing agent ZrH, would be expected to minimize the corrosion, regardless of whaich mechanism was effectave. Corrosion by the first mechanism can be minimized by reduction of UF to UF, by some reductant such as ZrH, to the extent tolerable without excessive deterioration of the physical properties of the fuel. Corrosion by the other mechanisms can, 1nprancaiple, be eliminated by the proper techniques for removing the reagent materials. The hydrogenation-hydrofluorination process for fuel purification was originally chosen as the one most likely to produce very pure fluoraide liquids. The schedule for this process was prescribed to provide what was considered to be large excesses of time for reaction and stripping at each stage. In recent experiments, attempts were made to quantitatively evaluate the chemical changes during each process stage. The need for such data 1s obvious. Incomplete treatment at any stage 1s reflected by poor corrosion characteristics, however, long treatment times are uneconomical 1n terms of equipment laife, production rate, material consumed, and manpower required and, in addition, result in slight alteration in fuel composition because of volatilization of the ZrF,. Stripping with 1inert gas should serve as a rapid and reasonably com- plete method for removing HF, because the acid fluorides are unstable at temperatures above 500°C. Since any residual HF should lead to enhanced corrosion, the progress of HF removal 113 ANP PROJECT QUARTERLY PROGRESS REPORT has been studied i1ntensively by analyzing the exit gas stream for HF. Initial experiments revealed that plots of the logarithm of HF concen- tration in the helium vs. volume of helium passed were essentially linear, with half-life volumes between 40 and 70 liters when 0.5 to 1.2 liters/main of helium was passed through about 1 liter (3 kg) of the melt. From tests of the stripping of the HF by He, 1t may be concluded that typical batches of fuel prepared by the standard procedure described previously should have been stripped to about 10°% mole of HF per liter of He and should have contained about 10 ppm of HF in the melt, There 1s no detectable difference in stripping rate with He between the NaZrF, and the NaF-ZrF,-UF, mixture containing 4 mole % UF,. The effect of flow rate on attainment of equi- librium vapor concentration has been only partly explored. When hydrogen 1s used to sweep the melt after passage of HF, the acid content of the effluent gas 1s con- siderably higher than when helium 1s used. The following reactions are almost certainly responsible NaF, + H, = N1° + 2HF, AF® 000k = —65 keal , 2FeF,<+ H, — 2FeF, * 2HF, AF° ;0% = =31 5 keal , FeF, + H, — Fe® + 2HF, AF° 000% = 1 5 kecal , 2UF, + H, = 2UF; + 2HF, AF®;,00°g = 22 O keal . In treatment of NaZrF, with hydrogen at 800°C half-li1fe volumes of about 70 liters are found. Since most of these experiments are performed 1in nickel containers, appreciable amounts the HF result from reduction of NiF,. In stripping of the ARE fuel mixture with H,, the contrast between He and H, 1s even more noticeable. When H, 1s used, the HF content of the gas drops sharply, as expected, but 1t levels off at about 10°* mole HF per liter of H,. The reduction of UF, to UF; 1n the solid state 1s considered negligible below 900°C. However, UF, 114 has been shown by petrographic tech- niques to be present in ARE fuel batches after prolonged H2 treatment at 700 to 800°C. There seems to be no doubt that the high values for HF i1n the exit gas are a consequence of this reaction. From the rate of HF evolution, 1t appears that the rate of reduction 1s about 3% of the UF, present per hour. Filtration of the melt after hydrogen treatment and HF stripping serves to reduce the nickel content of the fuel batch to less than 150 ppm. It has not been possible, to date, to bring the i1ron content of the melt to such a low level. It appears, however, that low values for N1F2, FeF,, and HF can be achieved without excessive exposure times and without undue reduction of the UF,. Pilot-Scale Fuel Purification (G. J. Nessle, J. E. Eorgan, Materials Chemistry Division). During the past quarter, a total of 326.4 kg of mixed fluorides, comprising 24 small batches of approximately2 kg each and 11 large batches of approximately 25 kg each, has been processed by the group in Building 9928 and dispensed to re- questing personnel in the ANP Division. Heating the control panels to 125°F has resulted 1n marked improvement 1in the smoothness and efficiency of operation. The addition of two pressure gages, one directly on the reactor can and one directly on the receiver can, has eliminated the difficulty i1n discovering possible danger points, and minor failures during a run are now practically nonexistent, The process for treatment of mixed fluorides has been modified to permit 1.5 hr of stripping with H, and 1.5 hr of stripping with He after the HF treatment and before transfer of the molten fluorides. A gas throughput of about 15 liters/min 1s accomplished by pressurizing to 10 lb (gage) and releasing to 0 1lb (gage) 30 times per hour. The HF content of the straip gas 1s usually about 10”% mole/later at the time of fuel transfer. Thais probably corresponds to about 10 ppm of HF 1n the melt. The combaination of H, treatment and filtration through sintered nickel has reduced the soluble nickel 1n the melt from 900 to less than 150 ppm. Fluoride Production Facility (G. J. Nessle, Materials Chemistry Division, H. W. Savage, ANP Division). The production facility for preparing in excess of 3000 1lb of fuel solvent 1is scheduled for completion during the month of March. Shakedown runs and other preliminary tests of the appara-~ tus should require less than four weeks, the production operation could, 1f necessary, be finished by mid-June. The low-hafnium ZrCl, has been con- verted to ZrF,, and the equipment for this process has been cleaned and put into standby condition for future use, 1f needed. At present, the stockpile contains 3000 1b of ZrF,. All batches of the ZrF, have been sampled and analyzed by chemical met hods for zirconium and fluoraine. Complete spectrographic analyses, with special shots of each batch, for hafnium and boron are available. The boron content of the batches ranges from 1.8 to 0.3 ppm, with an average of 0.8 ppm, and seven of the nine batches show less than 85 ppm hafnium. The other two batches, comprising about 25% of the total, show 150 and 550 ppm hafnium. PERIOD ENDING MARCH 10, 1953 About 1000 1b of NaF from one production lot has been stockpiled for use 1n fuel production. Chemical and spectrographic examinations of samples from three containers indicate that the material 1s satisfactory for use Hydrofluorination of Zroz-NaF Mixtures (R. E. Thoma, Jr., C. M. Blood, Materials Chemistry Division). The feasibility of producing NaZrF, by a short treatment of NaF-ZrQ, mixture with HF at a high temperature has been demonstrated. Best results are obtained when some previously prepared NaZrF; 1s added to the charge so that a liquid phase 1s present during the 1nitial stages. An amount of NaZrF¢ that 1s equivalent to 10% of the total charge appears to be suffai- cient to bring about complete con- version of the mixture in 4 hr at 800°C. Utilization of about 30% of the HF delivered has been attained. There are still some problems in connection with the equipment for thas operation. Frequent rupture of the HF input line occurred when nickel tubing was used. Metallographic examination revealed that the attack was due to sulfur, which was probably present in the Zr0O, feed. Graphaite liners 1n the nickel reactors have been used satisfactorily, however, there has been difficulty in the use of graphite dip lines for HF because of breakage during runs. Further study of this procedure, which promises to reduce materially the cost of ZrF, as NaZrF,, 1s planned. 115 10. CORROSION RESEARCH W. D. Manly, Metallurgy Division W. R. Grimes, Materials Chemaistry Division H. W. Savage, ANP Division Corrosion phenomena are usually examined first 1n static or seesaw tests, and then selected systems or phenomena are examined 1n the convection loops. Crevice corrosion, effect of exposure time, and the effects of additives have been studied in both static and dynamic tests during the past quarter No significant increase 1in corrosion was noted ain static tests However, the crevice corrosion found 1n thermal convection loops was about three times that found 1n adjacent tubing, the mechanism by which crevices 1ncrease corrosion 1s not yet understood. Corrosion, as evidenced by the chromium metal content of the fluoride, has been shown to be fairly constant after the first 100 hr of exposure Convection loops that were run for 1000hr at 1500°F 1ndicated that the amount of corrosion decreased with time, although there was a small increase 1n the depth and intensaty of attack. The work on the dynamic testing of the fluoride fuels 1n Inconel thermal convection loops has continued, with a major effort being made to determine why the corrosion 1n the fluoride systems has been 1increasing. It has been shown that neither the increase i1n fluoride batch size nor poor handling techniques during filling were responsible, retreating the fuel batches has produced contra- dictory results. Evidence indicates that the i1ncrease 1n corrosion 1s being caused by faulty purification, since the corrosion was deeper 1n loops 1n which hydrofluorinated fluoride mixtures were circulated than 1t was 1n the loops that used as-melted fluoride as the fuel. One Inconel thermal convection loop was Inconel - operated with a hot-leg temperature of 1650°F, and no large increase 1n the depth of attack was noted 1n comparison with that found 1n loops run at 1500°F A whirlagig rig, patterned after one developed by NACA, with which fluoride flow velocities of 10 fps can be obtained, has been placed 1n operation. Preliminary testswith this apparatus with fluoride fuels 1n Inconel show only a slight 1ncrease 1n corrosive attack compared with that found 1n static capsule tests. It was confirmed that ZrH, 1s an effective corrosion inhibitor when used with NaF-ZrF, -UF, (46-50-4 mole %), TiH, will also inhibit corrosion, but not quite so effectively as ZrH,. It now appears that the solubilaty of chromium in NaF-KF-LiF-UF, (10.9-43.5- 44.5-1 1 mole %) 1s about 3000 ppm, which 1s not so high as originally expected Considerable attack was found 1n an Inconel loop 1n which the fluoride had been saturated with chromium metal prior to circulation. The compatibility of BeO in Nak continues to be examined, and a correlation between the density of the BeO and the corrosion behavior 1s being studied Various surface treatments for the BeO are being evaluated. There 1s some evidence that BeO 1s soluble 1n NaK to an extent that 1s dependent on the temperature of the system The corrosion of two cermets was examlned 1n both fluorides and liquid metals. Al though the tests 1n liquid metals showed little or no attack, at least one fluoride test resulted 1n signifi- cant corrosion The mass-transfer characteristics of various metals in high-purity lead are being studied by usaing 117 ANP PROJECT QUARTERLY PROGRESS REPORT quartz thermal convection loops The results show that both columbium and molybdenum have very good re- si1stance to mass transfer and corrosion in liquid lead On the other hand, severe corrosion and mass transfer were found i1n lead-Inconel systems, even thoughthe lead had been carefully deoxi1d1zed Materials to be tested in liquid lead include type 304, high and low carbon, stainless steels, type 410 stainless steel, and Armco 1iron. Tests of stainless steel and Inconel 1n sodium-lead alloys continue to show that additions of sodium to lead decrease the severe corrosive action. The effect of a hydrogen atmosphere in minimizing mass transfer of nickel 1n sodium hydroxide was demonstrated by using a 50-50 mixture of helium and hydrogen However, the use of atmos- pheres of CO,, forming gas, wet hydrogen, or dry air under a hydrogen pressure of one-half atmospheric pressure resulted 1n considerable mass transfer. A number of simple and complex fluorides of the structural metal has been prepared for studies of thear effect on corrosion by fluorade mixtures The complex fluorides formed 1n corrosion tests are 1dentified by either petrographic or x-ray- diffraction techniques,or both The oxidation of molten NaUF, yielded no uo,r,, UO,, or UF,. However, 1t 1s evident from the resulting products that the container plays an important part in the mechanism of air oxidation of uranium 1n the molten fluorides. FLUORIDE CORROSION OF METALS IN STATIC, SEESAW, AND ROTATING TESTS D C Vreeland J E. Pope E. E Hoffman L R. Trotter Metal lurgy Division F Kertesz C R. Croft H. J. Buttram R. E. Meadows Materials Chemistry Division Crevice Corrosion. In connection with the crevice corrosion tests 118 being run, 1t was suggested that a more realistic method of testing would employ a tapered or V-shaped crevice 1n the bottom of the testing tube Accordingly, a ji1g was designed for putting tapered crimps 1n test tubing With this jig, crevices can be made that are approximately 1 1/4 in long and vary 1in width up to approximately 0 3 inch. Several static tests with various materials for containing the NaF-ZrF,-UF, (46-50-4 mole %) have been rumn to test the crevices Corrosion 1n the crevices seemedto be somewhat errataic, with some surfaces being apparently unattacked and others havaing the usual subsurface voids However, accelerated corrosionain these crevices was not noted, the attack that was found 1n the crevices was not beyond what might be expected for the materials tested. High-Temperature Pretreatment of Inconel. On the basis of results from some previous tests with NaF-KF- LiF-UF, (10 9-43 5-44 5-1 1 mole %) at high temperatures (1200 to 1300°C), 1t was reported that there was no surface layer attack on Inconel. It was thought that perhaps a pre- treatment at high temperature would render Inconel i1mmune from corrosion at the usual temperature of 816°C. Accordingly, two Inconel tubes were loaded with the fluoride, heated for 4 hr at 1250°C, and then heated for 250 hr at 816°C. Light attack was noted 1n each tube 1n the form of subsurface voids that extended to a depth of 1 mil in one tube and 0.5 m1l in the other tube. Effect of Time of Exposure of Inconel to Fluoride Mixture with and Without ZrH, Added. When the ARE fuel mixture NaF-ZrF,-UF, (50-46-4 mole %), without added material, 1s tested 1n Inconel under an atmosphere of helium, the severity of attack and the depth of penetration seem to 1ncrease very slightly with exposure times of between 100 and 250 hours, Differences between corrosion observed after 25 and 100 hr seem to be real. A plot of the structural metal content of the fluoride mixture vs. exposure time 1s shown 1in Fig. 10.1. It 1s obvious that the chromium content shows a rapid i1nitial rise and then tends to level off at a high value, whereas the 1ron content falls rapidly at first and then very slowly. It seems certain that reactions such as FeF; + Cr —— CrF; + Fe, with consequent deposition of the metallic 1ron on the walls, are responsible fora part of the corrosion observed. When this fuel mixture, with 0 5 wt % ZrH, added, was tested 1n Inconel, the corrosion obtained was considerably less severe, but the corrosion changes with exposure time PERIOD ENDING MARCH 10, 1953 were similar., Figure 10 2 shows the results of analysis of the fuel mixture for the structural metals. In this case, FeF; and N1F, contents dropped to very low values because of the strong reducing agent, and, apparently, most of the CrF, was also reduced. However, with increased exposure time, the CrF; concentration increased slightly. These results are encouraging 1n that they indicate that the relatively large corrosion rates 1nitially observed are drastically reduced at longer exposure times Structural Metal Fluoride Additives. Chemical analyses of fluoride materials before and after corrosion testing 1in Inconel have repeatedly demonstrated that the 1ron and nickel contents of the mixture are reduced and the D 1600 WG 18656 CHROMIUM — /# 1400 // T 1200 Q b .— & W 1000 =z (o] o | 2 800 LJ = - < £ 600 ...,. Q 0 x 7 400 200 | IRON NICKEL 0 0 100 200 300 400 500 EXPOSURE TIME (hr) Fig 10.1 Structural Metal Content of NaF-ZrF4-UF4 (50-46-4 mole %) As a Function of Exposure Time. 119 ANP PROJECT QUARTERLY PROGRESS REPORT 1000 - DWG 18657 STRUCTURAL METAL CONTENT {ppm) M~ - ——--IRON _ | CHROMIUM 0 NICKEL 0 100 200 300 400 500 EXPOSURE TIME ({hr) Fig. 10 2. The - e chromium content 1s 1ncreased. reactlons 3FeF, + 2r —— 2CrF, + 3Fe and 3N1F2 + 2Cr ——> 2CrF,; + 3Na are obviously involvedin the corrosion mechanism, Since both 1ron and nickel fluorides are likely to be present to some extent i1n the final ARE fuel and since considerable amounts of these compounds have been present 1n some of the batches tested, a serles of experiments has been conducted to demonstrate the effects of the fluorides of 1ron, chromium, and nickel i1n considerable amounts on the corrosion of Inconel by the ARE fuel and the ARE fuel solvent. Addition of either FeF, or NaF, 1increases the severity of attack and depth of penetration. The addition of CrF, apparently has no effect on the corrosion rate. When 1ron or chromium fluorides are added, metallo- graphic examination reveals metallic deposits on the capsule walls, Chemical analyses of the solidified melts, after exposure, reveal that the 120 Structural Metal Content NaF-ZrF,-UF, (50-46-4 mole %), with 0 35 wt % Zrflz,As a Function of Exposure Time. addition of CrF; does not appreciably alter the amounts of 1ron and nickel normally found i1n the mixture. The addition of NiF,, however, greatly increases the amount of CrF; found an the melt, the very large quantities of soluble nickel added are nearly completely reduced and are apparently deposited on the walls of the capsule so that they are not removed with the fluoride mass. The addition of FeF, also serves to raise the chromium content of the fluoride mixture, but only a part of the FeF, seems to be reduced to metal. The handlaing, samplaing, and ana lytical techniques are not sufficiently preclise to permit obtaining accurate material balances from the equations shown above, i1n general, however, the balance 1s within 120%, from which it appears that a considerable fraction of the corrosion 1s, as was expected, due to other actions. Hydrogen Fluoride Additive. Hydro- fluoric acid 1s one of the reagents used 1n the process for purifying the fluoride melts. This reagent may be present in the melts at low concentrations, even though con- si1derable careis used i1n the stripping operation. Study of corrosion of Inconel by a batch of ARE fuel, NaF-ZrF4-UF4 (50-46-4 mole %), to which HF had been added as NaHF,, showed that heavy subsurface void formation resulted when as much as 0.3 wt % HF was present. This attack was shown to be as much as 3 mils deep, whach 1s at least twice the depth expected without addition. However, the use of 0.1 wt % HF resulted i1n less severe attack, with penetration to only 15 mls, and the addition of 0 04 wt % HF yielded a light metallic deposit and relatively light corrosive attack. The fuel mixture used 1n these experiments had been treated waith H, and, presumably, contained small amounts of UF;. It 1s possible that the reaction 2UF, + 2HF =— 2UF, t H, may have masked the deleterious effect of small HF additions. Fluoride Corrosion i1n a Rotating Rig. The whairligig ri1g¢!’ 1s now running and operational ‘‘bugs” are gradually being worked out In thais device 1t will be possible to run relatively high-velocity corrosion tests with the fluoride mixtures contained 1n Inconel and other suitable tubing. At the present time, a fluaid velocity of approximately 10 fps 1s being employed. Upon metallographac examination after a 100-hr test waith NaF-ZrF,-UF, (46-50-4 mole %) 1n Inconel at a temperature of 816°C (no appreciable thermal gradient employed), 1t was noted that the attack was no more severe than that often encountered 1n ordinary static corrosion tests, The attack was quite uniform and was in the form of subsurface voids to a depth of 2 mils. (I)L G. Desmon and D R Mosher, Preliminary Study of Circulation i1n an Apparatus Suitable for Determining Corrosive Effects of Hot Flowing Liquids, NACA-BM-E-51D12 (June 29, 1951). PERIOD ENDING MARCH 10, 1953 Chemical analysis of the fluoride after the test revealed the presence of 0.025% Fe, 0.003% N1, and 0.141% Cr. Other tests are being run with various additions to the fluoride mixtures, and attempts are being made to vary the heating arrangements on the tubing so that sizeable temperature gradients can be attained. Static Tests on Incoloy and Inconel 1n Fluorides. Incoloy (32 % N1-21% Cr-47% Fe) was tested in NaF-ZrF,-UF, (46-50-4 mole %) for 100 hr at 816°C. In a few places, subsurface voids to a depth of 2 mils could be noted. A photomicrograph of the Incoloy, taken after testing, 1s shown 1in Fig. 10.3. The descaling properties of the fluoride on oxidized Inconel were tested at 816°C, and 1t was found that the oxidized Inconel was cleaned very well at the 1-, 4-, and 8-hr test times that were tried. FLUORIDE CORROSION OF METALS IN THERMAL CONVECTION LOOPS G. M. Adamson, Metallurgy Division Effect of Fluoride Batch Size. As discussed 1n the previous report, ( ?) the corrosion has been i1ncreasing 1in the Inconel loops circulating NaF- ZrF,-UF, (46-50-4 mole %). The increases 1n themselves were not enough to cause trouble, but they are dangerous when regarded as a trend that 1s not understood. Consequently, a series of loops was run to determine whether the 1ncrease in batch size 1n the production of the fluoride mixture was responsible for the corrosion increase. Loops were run with fluoride mixtures prepared in 5- and 50-1b batches for 100- and 500-hr periods. In both the 100- and 500-hr tests, more corrosion was caused by the small batches than by the large batches. Obviously, the i1ncrease 1in batch s1ze was not the reason for the (Z)G M Adamson, ANP Quar Prog Rep Dec 10, 1952, OBNL-1439, p 133 ff 121 ANP PROJECT QUARTERLY PROGRESS REPORT Fig. 10.3. hr at 816°C. increase in attack. Furthermore, even though extreme care was taken during filling of the loop, the attack was still extensive; therefore poor handling during filling was also eliminated as the cause. Fluoride Pretreatment. Two small batches of NaF-ZrF,-UF, (46-50-4 mole %) were taken from the 50-1b batch and given a complete retreat- ment.¢?) When these batches were circulated in Inconel loops, the attack in one instance was considerably greater than average, whereas in the other it was less; therefore it may be concluded that some major variable still has not been controlled. G)c, M, Blood, A. J. Weinberger, F. P. Boody, and G. J. Nessle, ANP Quar. g Rep. June 10, 1952, ORNL-1294, p. 97. 122 v Static Test of Incoloy in NaF-ZrF,-UF, (46-50-4 mole %) for 100 The attack found in Inconel loops in which NaF-ZrF,-UF, (46-50-4 mole %) circulated was much less than had been found when NaF-KF-LiF-UF, (10.9- 43.5-44.5-1.1mole %) was the circulated material.(*) Since the NaF-ZrF,-UF, had received more extensive pre- treatment with H, and HF, its higher purity is a logical explanation for its lower corrosiveness. Accordingly, two special batches of the NaF-KF-LiF- UF, mixture were given the same special pretreatment. These batches produced attack on Inconel that was almost twice that heretofore observed. Two additional batches of the same fluoride were then specially prepared and circulated in Inconel loops, and (4)G, W, Adumson, ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL-1375, p. 103 ff. the attack was even worse than 1in the first test, In all the tests with pretreated fuel, the operator reported a strong odor of hydrogen fluoride during the filling of the loops. Since hydrogen fluoride, as an additive 1in static tests, has been shown to 1ncrease corrosion, 1t 1s most probable that the HF used at one step 1n the fuel pretreatment, as well as the NaiF produced during this treatment, 1is not being completely removed and 1s causing the increased corrosion. These tests show that poor purifi- cation may do more harm than good. Hydride Additives It has been confirmed that zirconium hydride wall reduce the attack of NaF-ZrF, -UF, (46-50-4 mole %) on Inconel. An addition of 1/2% zirconium hydrade 1n one thermal convection loop reduced the depth of attack to between 1 and 4 mils. This attack 1s lower than normal but 1s deeper than that found when zirconium hydride 1s added to NaF-KF-LiF-UF, (10.9-43.5-44.5-1.1 mole %). The coolant underwent considerable transformation, as evidenced by the presence of a brown material. Titanium hydride (1/2%) was added to the NaF-KF-LiF-UF, mixture and circulated 1n Inconel. The maximum penetration of 2 5 mils, although much less than that found when no additive was used, was not quite so low as that obtained when zirconium hydride was added. Surface layers were found 1n both the hot and cold legs These layers must receive additional study 1f any hydride 1s to be used as an inhibitor. Metal Additives. The three most common 1mpurities found in the fuel are nickel, 1ron, and chromium. One loop 1n which chromium metal powder had been added to NaF-KF-LiF-UF, (10.9-43.5-44.5-1.1 mole %) was discussed 1n the previous report.(?) This test was repeated, and again no change 1n the average depth of attack PERIOD ENDING MARCH 10, 1953 was found, no unusual layers were found, and the chromium content of the fluorides from the loop was normal, so again the chromium metal did not go into solution. Chromium metal powder, in sufficient quantity to make 0.5% in the solution, was then added to a transfer pot before the fluorides were placed anto 1t. After the fluorides had been added, the pot was kept hot and helium was blown through 1t for 30 minutes. Again, average attack was found after this materaial had been circulated in an Inconel loop. The chemical analysis of thas fluoride showed a chromium concen- tration of 2500 ppm, which 1s about twice the normal concentration, Further work 1s scheduled to determine whether the solubility of chromium 1is much lower than-was expected. This 1s likely, since the chromium content did not 1ncrease greatly during circulation and yet attack had taken place As another approachto this problem, a section of Inconel pipe was chromium plated and welded 1nto the hot leg of a loop. After the NaF-KF-LaF-UF, mixture had been circulated, the depth of attack was the same 1n adjacent plated and unplated sections. No traces of the chromium plate remained on the walls. A metallic appearing layer was found in the cold leg. The chromium content of the fluorides varied from 2500 to 3000 ppm, whaich, although higher than normal, 1s about the same as that found in the loop discussed above. Temperature Dependence. A loop filled with the fluoride fuel, NaF- ZrF,-UF, (50-46-4 mole %), was operated for 500 hr with a minimum cold-leg temperature of 1500°F and a hot-leg temperature of 1650°F., The hot-leg attack was moderate to light, with a maximum depth of 11 mils and an average depth of 8 mils. A concen- trationof voids in the grain boundaries had taken place. The results may be compared with those of a similar 123 ANP PROJECT QUARTERLY PROGRESS REPORT loop test in which the hot-leg temper- ature was 1500°F The hot-leg attack 1n this loop was from 3 to 10 mails deep. Comparison of these two loops indicates that the corrosion mechanism 1snot extremely temperature sensitive. One disturbing fact 1s that with the hotter loop there was some evidence of uranium segregation in the cold leg. This 1s now being checked in another loop. Crevice Corrosion. Another loop operated 1n the study of crevice corrosion¢?) was an Inconel loop, with two crevices built i1nto the hot leg, 1n which NaF-ZrF,-UF, (46-50-4 mole %) was circulated. The upper hot-leg section showed a maximum attack of 16 mils and an average attack of 12 mi1ls, whereas the attack in the lower hot-leg section had a maximum depth of 9 mils. The maximum attack i1n the crevices was 19 mails and the average was 12 mils. The intensity of attack 1n the crevice was several times that in the straight section. This confirms previously obtained results i1n that the attack 1n the crevice 1s only slaightly deeper but the i1ntensity of attack 1s greatly increased. Inserted Corrosion Samples. Me- tallographic data received for the previously discussed(?) loops 1n which thermocouple tubes were i1nserted into the center of the hot and the cold legs confirm the results obtained with flat samples. With both NaF- ZrF,-UF, and NaF-KF-LiF-.UF,, the attack on the loop walls i1n the hot leg was normal, but the hot-leg inserts were attacked to a depth of onlyl mil, no satisfactory explanation has been offered for this phenomenon. Effect of Exposure Time. An Inconel loop was used to circulate NaF-ZrF,-UF, (50-46-4 mole %) 1000 hr, and the resulting hot-leg attack was moderate to heavy in 1ntensity and extended to depths of 5 to 11 mils. The attack was present both as general attack and as a concentration of the 124 voids 1n the grain boundaries For comparative purposes, these data are summarized i1n Table 10 1, along with simi1lar data for loops previously operated for 100 and 500 hours Although the depth and intensity of the attack i1ncreases with the exposure time, the rate of attack definitely decreases TABLE 10.1. EFFECT OF EXPOSURE TIME ON DYNAMIC CORROSION OF INCONEL BY NaF-ZrF4-UF4 TIME OF OPERATION | MAXIMUM ATTACK | cpposyoN INTENSITY (mils) (hr) 100 4 Light to moderate 500 10 Light to moderate 1000 11 Moderate to heavy Nonuranium Bearing Mixtures. Since NaF-ZrF, (50-50 mole %) will be used during the 1nitial testing of the ARE, several loops have been operated to study 1ts effects These are standard loops, that 1s, Inconel loops operated for 500 hr with a hot-leg temperature of 1500°F. The upper section of the hot leg of one loop showed light attack that extended to depths of 4 to 8 mils, the lower section was practically unattacked The second loop 1s still operating. A third loop, 1n which NaF-ZrF, (52-48 mole %) was circulated, ylielded scattered i1ntergranular attack of from 3 to 8 mils 1n the top section of the hot leg. The lower section showed general pitting 1 mil deep, with an occasional patch of pits up to 6 mils deep No deposit was found in the cold leg Zirconium hydride (1/2%) was added to another batch of this coolant, which was then subject to the standard loop test The hot-leg surface of this loop was pitted, but the attack, as shown by the subsurface voids, extended to a maximum depth of only whereas 1.5 mils. A two-phase layer 1 mil thick was present on the hot-leg surface, and a tightly adhering metallic-appearing layer 0.1 mil thick was found in the cold leg. Again the zirconium hydride acted as an inhibitor, but hot-leg deposits were formed. One Inconel loop circulated 500 hr with NaF-BeF, (57-43 mole %). The maximum hot-leg attack was 9 mils and the average was 4 mils. More voids were concentrated in the grain bounda- ries than are normally found, and the voids were larger. A nonmetallic deposit was found in the cold leg. With both NaF-ZrF, and NaF-BeF,, the attack mechanism seems to be the same as for the other fluorides. 10.4. 100 hr at 850°C. Fig. 100x. PERIOD ENDING MARCH 10, 1953 LIQUID METAL CORROSION OF STRUCTURAL METALS E. E. Hoffman J. V. Cathcart L. R. Trotter D. C. Vreeland J. E. Pope Metallurgy Division G. P. Smith Seesaw Tests of Sodium-Lead Alloys. A series of seesaw tests of sodium- lead mixtures has been run in Inconel and types 430 and 304 stainless steel tubing for 100 hours. The lead for these tests was purified by bubbling hydrogen through molten lead at 750°C for 1 1/2 hours. All Inconel tubes were attackedin the hot zone. Severity of the attack decreased with increasing sodium additions, but the attack was still appreciable with a 30 wt % sodium addition, as shown in Fig. 10.4, UNCLASSIFIED Y-7804 [ 5% =y N, < S Inconel After a Seesaw Test ina30 wt % Na-70 wt % Pb Mixture for 125 ANP PROJECT QUARTERLY PROGRESS REPORT which illustrates the type and extent ofattack of these sodium-lead mixtures. In no case was attack or deposition of crystals noted in the cold zone upon metallographic examination; however, in several of the tubes, crystals could be detected in the bath material In the test with the type 430 stain- less steel tubing, there was very little corrosive attack by the sodium- lead mixtures. Type 304 stainless steel showed exceptionally good resistance to these corrodants. A section of the hot zone of a tube after exposure to 5% Na-95% Pb is shown in Fig. 10.5. There was es- sentially no attack in any of the tests with type 304 stainless steel. A summary of these tests is given in Table 10.2. Fig. Pb Mixture for 100 hr at 830°C 100X. 126 Static Tests with Lithium. Two- component static tests of several materials have been run with lithium at 1000°C to furnish a basis of comparison with three-component tests run previously. Severe attack had been noted in the three-component tests, 10 to 15 mils penetration being quite common. It should be emphasized, however, that the more recent tests were static tests and the attack might be less severe than in the dynamic systems. Results of these tests are presentedin Table 10.3. Incoloy in Sodium, Lithium, and Lead. Incoloy (32% Ni-21% Cr-47% Fe) was tested for 100 hr at 816°C in sodium, lithium, and lead. No attack was noted in either sodium or lithium, although some crystal deposition was UNCLASSIFIED Y-8144 10.5. Type 304 Stainless Steel After a Seesaw Test in a 5 wt % Na-95 wt % TABLE 10, 2. PERIOD ENDING MARCH 10, 1953 SEESAW TESTS OF VARIOUS MATERIALS RUN IN SODIUM-LEAD MIXTURES FOR 100 HOURS BATH COMPOSITION AVERAGE ¥FHPERAJUBE MATERIAL (%) (¢ METALLOGRAPHIC NOTES Na Pb Cold Zone Het Zone Inconel S 95 530 §20 Terminated after 18 hr, no cairculation, 12 mils of intergranular attack in het zone 10 90 330 870 Intergranular attack entirely through 35-m1l wall 2n hot zone 15 85 315 860 15 m1ls of heavy intergranular attack in hot zone 20 80 395 855 12 mi1ls of intergranulsr attack in hot tone 25 75 470 850 S mils of intergranular attack in hot zone 30 70 485 850 6 mils of intergranular attack 1n het’ zone Type 304 stainless steel 5 95 560 830 Essentially no attack 1n either hot or cold zonme i1n any of the type 304 stain- lesas steel tests 10 90 575 840 15 85 545 830 20 80 585 825 25 75 600 835 30 70 600 835 Type 430 stainless ateel 5 95 565 850 Essentially ne attack in either hot or cold zone 1n any of the type 430 gtaine less steel tests 10 90 580 845 15 85 575 820 20 80 575 810 25 75 550 805 30 70 610 835 127 ANP PROJECT QUARTERLY PROGRESS REPORT TABLE 160.3. RESULTS OF TWO-COMPONENT STATIC TESTS WITH LITHIUM AT 1000°C MATERIAL TIME(EE TEST METALLOGRAPHIC NOTES Type 316 stainless steel 100 No si1gn of attack on sections exposed to liquid phase, 2 5 mils of intergranular attack on sections exposed to vapor phase Inconel 100 Some mass-transfer crystals could be noted on the surface, no attack in vapor zone, 3 to 4 mi1ls of subsurface voids on tube i1n bath zone Type 304 stainless steel 400 l to 2 mils of intergranular penetration and transformation from austenite to ferrite apparent 1n the test with lithium. In the test with lead, 1 to 2 mils of intergranular penetration could be noted. Corrosion by Lead 1n Thermal Convection Loops. A series of tests has been made to determine the extent of mass transfer and corrosion 1in thermal convection loops containing liquid lead. The systems lead-Inconel, lead-niobium, and lead-molybdenum have been studied. Results are also avalilable for the lead-type 304 stainless steel system, but are incomplete. Since the results of previous investigations of mass transfer and corrosion 1n lead thermal convection loops have been subject to criticism because of oxide i1n the lead or on the metal tubing, 1n the present series of experiments, the liquid lead was deoxidized with hydrogen prior to contact with the test metal. Care was also taken to avoid heavy oxidation of the metal sections of the loops. The apparatus and experi- mental procedure were described in the previous report. (5) The quartz tubing used in the loop 1s virtually i1nsoluble 1n lead at the test temperatures, the concen- tration of 3510, 1n the lead after one (5)ANP Quar Prog Rep Dec 10, 1952, ORNL- 1439, p 148, 128 test was only 17 ppm, and the S10, concentration 1n a lead sample that had not been 1n contact with glass or quartz was 15 ppm Loop tests were run with Inconel, columbium, molybdenum and type 304 stainless steel samples suspended 1n the circulating lead The results for Inconel have been reported previ- ously, (%) the results for the other metals are summarized i1n Table 10 4. The columbium specimens exhibited no mass transfer and virtually no corrosion even after almost 600 hr of loop operation. The small decrease in wall thickness was probably due to the solution of the small amount of columbium required to saturate the lead. No mass transfer occurred with molybdenum, but the corrosion, although slight, was a little greater than that found for columbium. Little or no intergranular penetration was observed. The reason for the dis- crepancy 1n the wall-thickness measure- ments for the second lead-molybdenum test 1s not known. The molybdenum specimens used were not actually sections of tubing but were made of molybdenum foil bent into a cylinder. The foil had a specified thicknessof 10 mils, which was checked by micrometer readlngs at two points on the foi1l. It 1s possible that there was a varlation 1n thickness 1n the PERIOD ENDING MARCH 10, 1953 TABLE 10.4. CORROSION OF STRUCTURAL METALS BY LEAD IN QUARTZ THERMAL CONVECTION LOOPS TEST TEMPERATURE (°C) SPECIMEN WALL THICKNESS (in.) TIME OF METAL NO OPERATION * | Hot Zone | Cold Zomne.| Original | Hot Zone | Cold Zone (hr) Columbium 1 800 575 0.035 | 0,032 0.034 270( ) 2 800 565 0.035 0.033 0.033 572(%) Molybdenum 1 800 450 0.010 | 0.006¢¢) | o0.006(¢) 473(®) 2 800 595 0.010 | 0.011 0 010 305(8) Type 304, high carbon, stainless steel(?) 1 800 500 111 (a)Stoppage caused by failure of hot leg heater, no plug formed (b)Scheduled termination (C)D1screpancy probabl y caused by variation 1n tube thickness. (d)Examxnat1on incomplete fo1l and that the original foi1l used CORROSION OF CERAMICS BY FLUORIDES 1n test No. 2 was slightly greater than 10 mils thaick. The results of the test with type 304 stainless steel are i1ncomplete. On the basis of the thermal record of the experiment, however, 1t appeared that circulation 1n this loop was also stopped by plug formation. The type 304 stainless steel specimen used had a carbon concentration of 0.087%. In summary, both columbium and molybdenum showed very good resistance to mass transfer and corrosion in liquaid lead. On the other hand, severe corrosion and mass transfer occurred 1n the lead-Inconel systems despite the careful deoxidation of the lead with hydrogen. The tests have i1ncluded a nickel-base alloy (Inconel), columbium, molybdenum, and an iron-base alloy (type 304 stainless steel). It 1s planned to extend the study to other iron-base alloys, such as type 410 stainless steel, to a low-carbon type 304 stainless steel, and to Armco 1iron. AND LIQUID METALS D. C. Vreeland L. R. Trotter E E Hoffman J E. Pope Metallurgy Division E. E. Ketchen L. G. Overholser Materials Chemistry Division L. A. Mann J. M. Cisar D. R. Ward ANP Division Cermets 1n Fluorides and Liquid Metals. Two cermets were prepared by the Ceramics lLaboratory, one consisted of TiC plus 20% N1 and 15% CbC plus TaC, and the other consisted of ZrC plus 20% 1ron. These cermets were tested 1n various corrosive mediums by sealing them,together with the corroding agent, under vacuum 1n Inconel tubes. After the tubes had been heated for 100 hr at 816°C, specimens were removed, checked for weight and dimensional changes, and examined metallographically. 1In the tests with sodium and lead, there was little or no attack (Fig. 10 6). The 129 ANP PROJECT QUARTERLY PROGRESS REPORT UNCLASSIFIED PE Y8209 ol Vit § Fig. 10.7. Titanium Carbide Cermet After Exposure to NaF-ZrF,-UF, (46-30-4 mole %) for 100 hr at 816°C. TABLE 10.5. RESULTS OF STATIC TESTS OF TiC AND ZrC CERMETS IN SODIUM, LEAD, AND SOME FLUORIDE MIXTURES CERMET CORRODANT b i METALLOGRAPHIC NOTES TiC | Na -0.0065 No attack apparent Pb Pb adhering to specimen No attack apparent NaF-KF-Li F-UFy Fluoride No attack apparent (10.9743.5-44.5-1.1 mole %) | adhering to specimen NaF-ZcFy-UF, -0.0119 Bonding material appeared to be (46-5024 mdle %) leached from specimen to a depth of 5 mils zC | Na -0.0037 Specimen showed that some uniform solution had taken place, but no thickness change was involved NaF-KF-Li F-UFy 0.0158 Surface appeared to be slightly (10.9°43.5-44.5-1.1 mole %) roughened 130 Fig. mixture NaF-KF-LiF-UF, (10.9-43.5- 44.5-1.1 mole %) appeared to have little effect on these materials, but in the one test with NaF-ZrF,-UF, (46-50-4 mole %), the specimen was attacked to adepth of 5 mils (Fig. 10.7). No significant dimensional changes could be noted in any of the tests. The weight loss data and the results of metallographic examination are given in Table 10.S5. Static Tests of the ARE-Be0 Blocks in Na and NaK. Two tests have been performed with ARE beryllium oxide moderator blocks heated for several hundred hours at 810°C in a type 347 stainless steel pot containing Na or NaK. In the first test, the beryllium oxide was in 255 immersed in.? of PERIOD ENDING MARCH 10, 1953 L3 S UNCLASSIFIED .‘ . Y-8210 8 x 2 005 % ol 3| 10.6. Zirconium Carbide Cermet After Exposure to Sodium for 100 hr at 816°C. sodium for 213 hr at 816°C. Although no crackingor spalling of the beryllium oxide was noticeable immediately after the test, the surface im- perfections apparent in part a of Fig. 10.8 occurred after the test specimen had been soaked in alcohol and water. In the second test, the NaK was transferred to the pot with the beryllium oxide when both the NaK and the beryllium oxide were at room temperature. The test was run for 500 hr at 816°C with the beryllium oxide immersed in 255 in.® of NaK. After the test, the beryllium oxide was badly cracked, as can be seen in part b of Fig. 10.8. The weight and dimensional changes noted for both tests are given in Table 10.6. 131 ANP PROJECT QUARTERLY PROGRESS REPORT Fig. 816°cC. 10.8. (a) In Na for 213 hr; Seesaw Tests of Be0 in NaK. Be- ryllium oxide specimens 1/4by 1/4 by 1 in, were tested in NaK in the rocker furnace for 100 hr, with a hot-zone temperature of approximately 830°C and a cold-zone temperature of ap- proximately 450°C. Specimens were crimped into the hot end of the tube to prevent their sliding from the hot end to the cold end. Specimens indicated in Table 10.7 as‘‘low density” had a density of approximately 2.27; those indicated as “high density” had a density of approximately 2.83. 132 UNCLASSIFIED Y-8297 ARE Beryllium Oxide Blocks After Static Tests in Na and NakK at (b) in NaK for 500 hours. Specimens indicated as ‘“dehydrated” were dried for 24 hr at 125°C before test. The information derived from these experiments is tabulated in Table 10.7. It is evident from these data that after the specimens were stripped in alcohol there was some slight evidence of cracking in the high-density specimens and some slight evidenceof spalling in the low-density specimens. Convection Loop testsof Be0 in NakK. Eight specimens of BeO 1/4 by 1/4 by 1/2 in. were placed in a wire basket 1n the surge tank of a thermal con- vection loop, and NaK was permitted to flow over the specimens at a velocity of about 8 fpm for 212 hours. The temperature of the NaK was 1150°F in the cold leg and 1000°F in the hot leg, and the samples were 1n contact with NaK at 1500°F. The NaK specimens 1n the second test were pretreated before exposure to the hot NaK i1n the following manner- specimens 1 and 2 received no special pretreatment, specimens 3 and 4 were treated with MgNO,, specimens 5 and 6 were 1mpregnated with AlF,; and CaF,, specimens { and 8 were treated with BeNO,. The test results are given 1in Table 10. 8. TABLE 10.6. PERIOD ENDING MARCH 10, 1953 It 1s noteworthy that specimens 1, 3, 5, and 7 were located above speci- mens 2, 4, 6, and 8, respectaively, 1in the basket during exposure to the Nak, and 1n each case the upper specimen suf fered the greater loss 1n weaght. Special treatment of BeO to increase 1ts resistance to hot NaK appears to offer possibilaities, and further tests are to be conducted. The effect of the density of the specimen could not be correlated to the weight losses 1n these tests, Solubility of Be0 1n NaK. Various workers have 1nitiated experiments to determine, 1f possible, the effect of NaK on BeO at temperatures of approximately 800°C 1n both static and WEIGHT AND DIMENSIONAL CHANGES OF ARE-Be0 BLOCKS AFTER STATIC TESTS IN Na AND NaK AT 816°C BEFORE AFTER CHANGE AFTER CLEANING IN CHANGE TEST TEST (%) WATER AND DRYING (%) First Test, 213 hr ain Na Length, in 5914 5917 +0 05 Width, ain 3 706 3 708 +0 05 Weight, g 3 154 3 221 +2 1 3 205 +1 6 Second Test, 500 hr in NaK Length, 1in 5 909 5 899 -0 17 Width, ain 3714 3712 -0 05 Weight, g 3 160 3 205 +1 4 3 187 +0 85 TABLE 10.7. RESULTS OF SEESAW TESTS OF Be0 IN NaK AFTER 100 hr AT 816 °C WEIGHT WEIGHT | TOTAL WEIGHT TOTAL BeO FOUND BY ANALYSIS SPECIMEN LOSS LOSS LOSS OF NaK FILTER AND CONTAINERS (g/an.?) (%) (ng) (mg) Low densaty, as received* 0 0454 2,38 30.7 14.26 High density, as received 0.0286 1 25 18 9 4,97 Low densaty, dehydrated 0.0893 5.15 47.0 9.91 High densaty, dehydrated 0 0104 0.52 6.1 6.02 *This test run for only 16 hours. 133 ANP PROJECT QUARTERLY PROGRESS REPORT dynamic test systems. The weight losses of the BeO blocks observed in a number of 1nstances prompted a search for the beryllium that had been removed from the block. 1In those tests for which the NaK was filtered at room temperature, the beryllium found 1n the filtrate corresponded to approximately 3 ppm of Be(O, which was a very small part of the total beryllium removed from the blocks. Further examination revealed that the be- ryllium or the BeO, or both, was located pramarily on the walls of the containers to which 1t adhered tightly, The question then arose as to‘'whether or not the beryllium was removed from the block and deposited on the wall by some process involving solubilaty of Be or BeO i1n the NaK at 800°C. Two tests were performed to measure the solubility i1n NaK at 800°C. 1In the first test, BeO was contacted with NaK at 800°C for 6 hr in the reaction- filtrationrig previously described. () The NaK was filtered at approximately 800°C, and the filtrate was collected in a nickel receiving tube. Based on this single experiment, the solubility of berylliumin NaKat 800°C corresponds to roughly 100 ppm, but 1t 1s not known whether the material dissolved was Be or BeO. (6)ANP Quar Prog BRep Dec 10, 1439, Fag 10 7, p 121 1952, ORNL- TABLE 10.8. In the second test, a stainless steel thermal convection loop was used to circulate NaK over BeO blocks suspended in the surge tank After 212 hr of circulation at a flow rate of 6 to 10 fpm, a weight loss of about 25% was measured for the BeO blocks (original weight of 4 g). After the NaK had been drained, the loop was sectioned as shown in Fig. 10.9. The operating temperatures of the loop are given on the figure, along waith the analytical results for the BeO determinations on the respective 3 1/2-1n sections. The analytical results show that the three coldest (and lowest) sections of the loop contained the greatest concentrations of beryllium. Since the analytical data showno accumulation of BeO fragments i1n the coldest part of the tank, 1t 1s suggested that a dissolution process, governed by the temperature gradient, 1s responsible for the transfer of the material from the BeO block to the walls of the cold region. Such a mechanism could be represented by the following reaction 2NaK + 2Be0 == Na,0 + K,0 + 2Be , which, 1f the reaction proceeds slightly to the right at 800°C and the beryllium formed 1s soluble 1n NaK, would account for removal of RESULTS OF TESTS OF Be0 IN Nak SPECIMEN NUMBER 1 2 3¢e) | 4le) 5 6 7 8 Average loss 1in dimensions per surface, 1in 0 006 0 004 [ 0002 (0002 | O 000 | 0 001 | 0 005 |0 004 Loss of weight, g 0 152 0 120 | 0 091 |0 076 | O 069 | O 062 | 0 162 |0 119 Loss of weight (%) 11 § 9 6 8 6 0 49 47 11.1 9.6 Density of specimen 2 60 2 50 54 2 45 2 82 2 65 2.66 2 45 (a) alcohol bath 134 Much or all of the weight loss may have been due to the shedding of a thin surface film 1n the PERIOD ENDING MARCH 10, 1953 UNCLASSIFIED DWG 48658 /\/\/\,PM 0] 20 22 WEIGHT OF BeQ FOUND IN VARIOUS SECTIONS > AVERAGE 19 SECTION NO BeO (mq) TEMPERATURE (°F) _1 - 1 113 1500 3 137 1360 3 5 552 18 - 6 1250 7 217 9 2208 4 1 10 248 4 1140 17 1" 70 8 13 5 76 ] — 14 % 1860 5 15 364 1370 \ 17 414 — 6 18% 1440 19 474 1800 20 1480 24 672 22 27 48 ¥ 14 and 18 LOCATED UNDER HEATER Fi1g. 10 9. Schematic biragram of Convection Loop for NaK-Be0 Test Showing Temperature and Be0 Recovered 1n various Sections. 135 ANP PROJECT QUARTERLY PROGRESS REPORT beryllium from the BeO block. If the solubi1lity of beryllium 1s less at 600°C than at 800°C, deposition or alloyingin the cold area would occur. EFFECT OF ATMOSPHERE ON THE MASS TRANSFER OF NICKEL IN HYDROXIDES H. J. Buttram C. R Croft F Kertesz Materials Chemistry Division The 1inclined-tube technique previ- ously described¢’’ was used for further study of the effect of various expera- mental factors on the mass transfer of nickel by hydroxides. With a tempera- ture gradient of 150°C (800°C at the bottom and 650°C at the top) under helium atmosphere, sodium hydroxide transported enough nickel to the liquid level of the inclined nickel tube to plugit completely i1n 48 hours. As was previously mentioned, hydrogen gas sweeping over the surface of the molten hydroxide reduced the mass transfer considerably, only polishing of the hot end and roughening of the cold endof the tube could be observed. The temperature gradient was then increased to 300°C (500°C at the top and 800°C at the bottom), and a hydrogen atmosphere was used. Similar satisfactory®results were obtained, the hydrogen largely suppressed any crystal deposit at the liquid level. Attempts were made to determaine the minimum pressure at which mass transfer 1s effectaively i1nhaibited by hydrogen. A 50 vol % helium-50 vol % hydrogen mixture over the hydroxide gave satisfactory results. On the other hand, with an absolute hydrogen pressure of 38 cm Hg - established as a manually adjusted vacuum - the hydrogen system affordedno protection, possibly because of the technique employed. A number of other atmospheres was also applied over sodium hydroxide 1in the 1nclined tube test. The use of (T)ANP Quer Prog Rep Dec 10, 1439, p 142 {f 1952, ORNL- 136 carbon monoxide and forming gas (10% H, + 90% N,) did not have a beneficial effect. Water in the hydrogen, even 1in small quantities, resulted in considerable mass transfer, Bubbling the hydrogen through water at various temperatures resulted in nickel crystal deposits 1n amounts similar to those obtained under a helium atmosphere. A continuous vacuum over the system (which would remove any hydrogen formed during the test) resulted 1n extremely heavy corrosion and mass transfer, and left the hydroxide with a nickel concentration of up to 3% However, when the tube was evacuated and sealed, the mass transfer was much less severe. Dry air was definitely less harmful than wet hydrogen. The use of commercial sodium hydroxide, which contains up to 2% water, resulted in greatly reduced mass transfer under dry hydrogen. One effect of the dry hydrogen may be to strip the water from the hydroxide. In addition to the tests with sodium hydroxide, experiments were made with sodium hydroxide-potassium hydroxide eutectic, with pure potassium hydroxide, and with lithium hydroxaide. With none of these materials could the i1inhibiting effect of hydrogen on the mass transfer of nickel be ob- tained, the reason for this 1s not yet clear. More closely controlled experiments are planned. FUNDAMENTAL CORROSION RESEARCH Identificationof Corrosion Products from Convection Loops (D. C. Hoffman, Materials Chemistry Division). Cooling jets that impinged on the cold leg of an Inconel thermal convection loop in which the LiF-NaF-KF eutectic containing 1.8 mole % of UF, was circulating, produced partial plugs which persisted after the cooling jets were removed and could not be liquefied at temperatures of up to 650°C. X-ray analysis of the plugs showed the presence of KF and some material that gave a slightly shifted spectrum of K,NaCrF, Petrographic investigation substantiated the presence of this compound and also revealed some bright metal that was shown to be nonmagnetic, Inspecti1on under binoculars (45X) showed a network of crystals, presumed to be the K,NaCrF,, growing inward from the tube walls. The compound K,NaCrF, has a solu- bility 1n the NaF-KF-LiF eutectic of 450 ppm Cr(III) and 2000 ppm Cr(III) at 500 and 650°C, respectively. It therefore appears probable that K,NaCrF, has been responsible for plugging in this loop, as well as 1n others, 1n which the amount of metal transferred was insufficient to account for the phenomenon. ©Of the ten stainless steel harps that were plugged while circulating this mixture, s1x showed higher chromium content 1in the cold leg, one showed higher chromium content in the hot leg, and the other three showed no systematic variation, The K,NaCrF, crystals apparently form an open network that gradually stops the flow without developing, under usual operating conditions, easlly detectable local concentrations. The observation that increased operating temperatures postpone the onset of plugging 1n stainless steel loops 1s consistent with this explanation of the plugging phenomenon. Preparation of Special Complex Fluorides (B. J. Sturm, L. G, Overholser, Materials Chemistry Division). Prepa- rations of K,LiCrF,, KNaLiCrFy Li,NaCrF,, Na,LiCrF,, and NaCrF; have been made to furnish specimens for x-ray and optical crystallographac data. Continued study of the complex fluorides of nickel has not yvet established whether definite compounds or solid solutions exist in the composition range between K,NiF, and Na,N1F,. Also, several simple fluorides of the structural metals, including chromofluoride, ferrofluoride, nickel fluoride, and ferrous oxide, have been PERIOD ENDING MARCH 10, 1953 prepared for studies of their effect on corrosion by fuel mixtures. Chromic fluoride can be prepared by thermal decomposition of (NH,),CrF, at temperatures not exceeding 900°C some CrF, 1s formed at higher tempera- tures. Chromous fluoride 1s prepared by heating (NH,),CrF, or a mixture of hydrated CrF; and NH,KF, to 1200°C 1in graphite crucibles. The product from either of these procedures 1s i1dentical with that obtained by hydrofluorination of metallic chromium at 800°C, Ferric fluoride 1s prepared by hydrofluora- nation of “anhydrous” FeCl; at 200 to 300°C. Ferrous fluoride may be prepared by heating (NH,) FeF, at 750°C or, preferably, by hydrofluorai- nation of FeCl, at a maximum tempera- ture of 500°C. Nickel fluoride 1s best prepared by thermal decomposition of (NH,),N1F, Ferrous oxide 1s prepared by decomposition, under an atmosphere of CO,, of ferrous oxalate precipitated from an aqueous solution of ferrous ammonium sul fate by ammonium oxalate. The ferrous oxide produced 1s very finely divided and 1s easily oxidized by room temperature air, all handling of this compound requires a dry, 1inert atmosphere. Alr Oxi1dation of Fuel Mixtures (R. P. Metcalf, Materials Chemistry Division). Solad UF, 1s known to yield UO,F, and UF; as predominant reaction products when treated wath oxygen at high temperatures.,(%:9) However, when molten NaUFS in a nickel boat was treated with 40 to 60 cc/min of dry air for 2 to 6 hr at 750°C, no gaseous uranium compounds were observed X-ray-diffraction techniques reveal the presence of N10, U;0,, and Na,NiF, among the solid products of the reaction. In addition, several un- 1denti1fied products that may be (8) 5. Fried and N R Davidson, The Reaction of UF, with Dry 0 A Nev Synthesis of UFG' AECD- 2981 ?Nay 1945). S S Kirslis, T. S8 McMillan and H. A. Bernhardt, The Reaction of Urantum Tetrafluoride with Dry Ozxygen, K-567 {March 15, 1950) 137 ANP PROJECT QUARTERLY PROGRESS REPORT double fluoraidesof nickel are observed. The absence of UO,F, and UO, and the lack of evidence for volatilization of UF, are noteworthy. These results, as well as the rapid corrosion rate of nickel an contact with molten fluorides ain air, are consistent with the followaing reaction mechanism 138 2N, + O2 ——> 2N10 2N10 + UF4 —> U0, + 2N1F, 3UO2 + 0, ——> U,0, 2NaF + N1F} —_— NaleF4 It appears that the container plays an 1mportant part in the mechanism of air oxidation of uranium 1n the molten fluorades. 11. Wo Dn Manly . PERIOD ENDING MARCH 10, 19353 METALLURGY AND CERAMICS J. M. Warde Metallurgy Division Cone-arc welding 1s now being applied to the production of the many tube-to-header joints needed 1n the construction of the complicated heat exchangers for aircraft reactors, Since the heat exchangers are to be operated at a moderately high pressure for the temperature i1nvolved, dished header sheets are desired. The dished header sheets i1mpose a new problem of unequal arc distances 1n cone-arc welding., Studies are under way to select the proper conditions tominimize this effect so that consistently sound welds can be produced. In the fabri- cation of large heat exchanger test units by Nicrobrazing, difficulties due to dilution and embrittlement by the brazing alloy were encountered. It 1s possible to substitute other alloys for this brazing cycle that will not cause severe dilution and embrittlement, and the evaluation of high-temperature brazing alloys 1s continuing. The melting point of the palladium-nickel brazing alloy was lowered by additions of germanium. The effects of time at the brazing temperature and of joint spacing on the strength properties of Nicrobrazed Joints have been determined, A summary of the butt-braze tensile strength data obtained on the various brazing alloys 1s presented. A search for a satisfactory brazing alloy for the joining of stainless- steel- or Inconel-clad copper fins to Inconel and stainless steel tubes has proved fruitful, and the oxidation characteristics of the alloys that show good flowability on Inconel are now being studied. However, 1t has been demonstrated that a chrome-plated, high-conductivity, copper fin does not have sufficient oxidation resistance for this application. The effects of environment and of surface treatment on the creep and stress-rupture properties of Inconel are still being studied. Preoxidized specimens were tested i1n argon, and 1t was found that the creep properties were quite similar to those determined in argon, therefore surface oxidation 1s not the controlling factor in the longer rupture times observed i1n the air tests. Control rod inserts of a mixture of boron carbide and i1ron are being prepared by powder metallurgy tech- niques for the GE-ANP program. The control rod inserts must be metal- lurgically bonded to the outside stainless steel tube, and various brazing alloys are being studied for this application. Boron carbide safety rods for the Tower Shielding Facility were also produced. Special heats of a high-puraty Inconel are being prepared for cor- rosion tests, In the subsequent extrusion of Inconel tubes, one of the biggest difficulties to overcome 1s the selection of a proper extrusion lubricant. A satisfactory tube has been produced by the use of glass as a lubricant. In other studies, the oxidation characteristics of columbium are being i1nvestigated, and some exploratory runs have been made on the production of materials for use as pump seals of the face-seal type being studied by the Experimental Engineering Group. An enamel with high boron content was applied on a shield test plate, but a part of the surface was not well covered and a second attempt was necessary. Two cermet fuel elements are described. Impregnation of hexagonal beryllium oxide blocks with beryllium and magnesium nitrates and 139 ANP PROJECT QUARTERLY PROGRESS REPORT calcium and aluminum fluorides was attempted. The fluorides filled the pores, and the magnesium impregnation shows promise. WELDING AND BRAZING RESEARCH P. Patriarca V. G. Lane G. M. Slaughter C. E. Schubert Metallurgy Division Cone-Arc Welding. The design of recent fuel-to-NaK heat exchangers for reactor research has necessitated the fabrication of small, Inconel, tube- to-header subassemblies that can be built individually and then be welded into one, larger, test assembly. It is obvious from the large number of tube-to-header joints to be heliarc welded that a semiautomatic method would be extremely desirable. As a result, the cone-arc welding technique was applied to preliminary experiments in this subassembly fabrication. Since the heat exchanger is to be operated at a moderately high pressure for the temperature involved, dished header sheets are desirable. If the plane of the bottom of the header is kept level during welding, it is obvious that unequal arc distances will prevail around the periphery of many of the tubes. Thus the ability to make con- sistently sound welds on the dished header depends, to a large degree, upon the selection of conditions that will minimize these variations in arc distance. Offsetting the tungsten electrode from the center of the tube prior to welding and using a ball-and- socket joint arrangement from which the header can be pivoted are obvious means of equalizing the arc distance around the tube periphery. Preliminary experiments on this problem have consisted of investigating the variables of arc current, arc time, and electrode distance required to produce consistently leak-tight welds. An experimental, cone-arc welded, test specimen is shown in Fig. 11.1, which illustrates the 140 An Experimental Heat Fig. 11.1. Exchanger Subassembly After Cone-Arc Welding. tendency for uneven melting around some of the tube peripheries when the tungsten electrode was centered over the tube with the header plate axis perpendicular to the electrode axis. The diameter of the header plate is 2 1/4 in. and the plate thickness is 0.125 inch. The tubing has an outside diameter of 0.148 in. with a 0.025-in. wall, and the hole center-to-center distance is approximately 0.378 inch. Some difficulties have been en- countered with joints that are not pressure-tight, but it is expected that further experiments will enable the production of subassemblies that will be completely sound. In each test assembly that has been fabricated thus far, only one or two joints were unsatisfactory; the other joints were pressure-tight to air at approximately 60 psig. This indicates that an improvement in technique or a slight variation in welding conditions should help to improve the quality of these cone-arc welded joints. Fabrication of Heat Exchanger Units A second large sodium-to-air heat exchanger assembly was fabricated by Nicrobrazing and the over-all appear- ance, as shown in Fig. 11.2, was much better than that of the first as- sembly.(') Changes incorporated in the brazing procedure for this second unit were the use of lesser quantities of Nicrobraz, the use of several aspirators to promote more even hy- drogen flow between the fins, the buildup of the whole assembly off the the drastic initial heating rate, and the use of a slightly lower brazing temperature. Pressure testing with helium, however, revealed the presence of a leak in the tube-to-fin matrix. A technique was developed to seal off this tube on each side of the leak by rebrazing, can bottom to overcome (L)ANP Quar. Prog. Rep. Dec. 10, 1952, ORNL. 1439, p. 164, UNCLASSIFIED Y-8580 Fig. 11.2. Sodium-to-Air Radiator After Nicrobrazing PERIOD ENDING MARCH 10, 1953 but upon performing this operation, several other leaks appeared in other tubes. It is hoped that the damaging effects caused by subsequent Nicro- brazing operations can be eliminated by adopting a modified design that would permit the heliarc welding of the tube-to-header joints and the joints in the heavier manifold sections. By the use of the single-braze method, any leaks in the tube-to-fin matrix after brazing can be eliminated from the coolant circuit by plug welding the proper tubes. The tube-to-header welds could then be made by manual heliarc welding, as could the other joints in the manifold circuit. With some minor changes in radiator design, it seems probable that completed units could be fabricated with less chance of obtaining leaky joints. By elimi- nating several high-temperature brazing operations, grain growth in the stain- less steel tubes could be minimized, as could embrittling of the base metal by brazing alloy diffusion. It is well known that certain Nicrobrazed stainless steel joints exhibit brittleness to a high degree. This factor, coupled with the alloying away of the tube wall by brazing alloy dilution and diffusion, is thought to be responsible for a major portion of the leaks encountered in heat exchanger fabrication. Since it seems probable that joint brittleness may also be associated with dilution and diffusion phenomena, a systematic study of this problem has been initiated. Nicrobrazed tube-to-fin joints are being prepared by using 0.010-in.- thick, type 302 stainless steel fin material and 0,150-in.-OD, 0.016-in.- wall, type 304 stainless steel tubing for studying, by metallographic ex- amination, the effects of (1) the quantity of brazing alloy used, (2) the brazing temperature for a given time, and (3) the brazing time for a given temperature. Small, medium, and large amounts of brazing alloy and 141 ANP PROJECT QUARTERLY PROGRESS REPORT the time at temperature, which may vary from 10 min to 18 hr (for an overnight brazing cycle), are being investigated., The choice of brazing temperature 1s being varied from 20°C above the melting point to a maximum of 120°C above the melting point. Since 1t may be advisable to find a substitute for Nicrobraz in the con- struction of sodium-to-air heat ex- changer, two other brazing alloys are being subjected to similar diffusion and dilution i1nvestigations. An 82% Au-18% N1 alloy 1s typical of the lower-melting, ductile, oxidation- resistant brazing alloys that, un- fortunately,are i1ncompatible with sodium but may be used for tube-to-fin construction 1f dilution and diffusion can be controlled. A 60% Mn-40% N1 alloy has been shown to be compatible with sodium, but i1t 1sslightly attacked in high-temperature oxidation tests. It 1s also somewhat brittle in the as- brazed condition. Brazing of Copper to Inconel The need for a satisfactory brazing alloy for joining copper fins to Inconel tubing and for edge-sealing sheared, Inconel-clad, copper fins has been be resistant to oxidation at 1500°F, should preferably serve as a diffusion barrier against copper penetration into the Inconel during service, and should have a relatively high strength at 1500°F. It 1s likely that some alloy other than a copper-base alloy would best fill these requirements. An experimental, modified, Nicrobraz alloy, which melts at approximately 1850°F, appears to be very promising, as do alloys of 82% Au~18% N1 and 90% Au-10% Co. Oxidation tests at 1500°F on brazed Inconel joints indicate no appreciable attack., A list of several alloys, currently being 1investigated, that melt i1n this medium temperature range 1s given 1in Table 11.1. Flow- abi1lity tests on Inconel have been made, and the results of these tests are also listed in Table 11.1. Edge-sealing experiments on chromium- plated copper disks are being con- ducted, and techniques are being studied for obtaining a satisfactory edge seal on these disks. At the present, the fin edges are being suspended 1n a slurry of the modified, low-melting-point, Nicrobraz alloy and slowly rotated until an even coating emphasized. The resultant braze should of brazing alloy 1s deposited. The TABLE 11 1 PROPERTIES OF THE MEDIUM-TEMPERATURE BRAZING ALLOYS CURRENTLY BEING INVESTIGATED BRAZING MELTING POINT FLOWABILITY METHOD OF TEMPERATURE BRAZING ALLOY (°F) 2°F;U ON INCONEL APPLICATION 95 5% Cu-4 5% Be 1590 1750 Poor Powder 92% Cu-8% S1 1530 1700 Poor Powder 75% Cu-25% Sn 1470 1600 Good Powder Low-melting-point Nicrobraz 1800 1850 Good Powder 82% Au-~18% Na 1800 1850 Excellent Sheet 90% Ag-10% Cu 1600 1800 Poor Sheet 90% Au-10% Co 1830 1870 Good Sheet 95% Ag-5% Ge 1650 1700 Poor Sheet 90% Cu-10% Ge 1800 1870 Good Sheet 142 fin is then edgebrazed in a hydrogen atmosphere. Another method to be investigated for covering these exposed copper edges is the preplacing of a fine wire of brazing alloy in the crevice formed by chemically etching away a small amount of copper in nitric acid. Tt is expected that this latter technique, if successful, will prevent fin distortion, which is believed to be due to unavoidable nonuniform wetting of the fin periphery by brazing alloy applied as a slurry EVALUATION TESTS OF BRAZING ALLOYS P. Patriarca V. G. Lane G. M. Slaughter C. E. Schubert Metallurgy Division Corrosion of Brazing Alloys by Fluorides. In a further attempt to 11.3. Fig. hr at 1500°F in NaF-KF-LiF-UF, (10.9-43.5-44.5-1.1 mole %). regia. 100x. PERIOD ENDING MARCH 10, 1953 lower the melting point of the 60% Pd- 40% Ni brazing alloy, 5% germanium was added. This alloy can be consistently brazed at 2150°F, which is in the brazing range of the more common Nicrobraz alloy. Static corrosion tests on brazed joints in the molten fluoride salts indicate that there is only slight attack by molten NaF-KF- LiF-UF, (10.9-43.5-44.5-1.1 mole %), as shown in Fig. 11.3, whereas severe attack is present on the sample immersed in NaF-ZrF,-UF, (46-50-4 mole %), Fig. 11.4. Effect of Brazing Time on Joint Strength. A review of results of recent and previous experiments con- ducted to evaluate the effect of time at the brazing temperature on the joint strength of Nicrobrazed Inconel joints shows that the effect of time UNCLASSIFIED Y-8629 .0t INCHES (100X) o & Inconel Joint Brazed with a 60% Pd-35% Ni-5% Ge Alloy After 100 Etched with aqua 143 P eSS ANP PROJECT QUARTERLY PROGRESS REPORT Inconel Joint Brazed with at 1500°F in NaF-ZrF,-UF, (46-50-4 mole Fig. 11.4. is not so important as the previously reported limited data indicated. The room temperature strengths of these joints are generally much less than the strength of the base Inconel and do not vary appreciably with time at temperature. Only occasionally does a joint exhibit high tensile strength at room temperature, and the previous results were apparently distorted by these infrequent high values. A summary of the data is given in Table 11.2; four tensile bars were tested for each brazing time. It is expected that the brazing time will have a greater effect with stainless steel, because the width of the diffusion zone is generally larger. The composition of Inconel is similar 144 UNCLASSIFIED 1 ¥-8701 a 60% Pd-35% Ni-3% Ge Alloy After 100 hr %). Etched with aqua regia. 100X, to that of Nicrobraz, with respect to nickel, chromium, and iron contents, and therefore diffusion may be some- what hindered by the lack of large concentration gradients. The tensile strengths of Nicrobrazed Inconel joints at 1500°F have all been consistently high, as previously reported, and the fractures often occur in the base metal. Effect of Spacing on Brazed Joint Strength. An investigation was con- ducted to determine the effect of joint spacing on the short-time, room-temperature, tensile strength of Nicrobrazed Inconel joints. Four butt-brazed tensile bars were prepared for each joint spacing of 0.005, 0,010, 0.015, and 0.020 inch. The results of PERIOD ENDING MARCH 10, 1953 TABLE 11 2 EFFECT OF TIME AT BRAZING TEMPERATURE ON STRENGTH OF NICROBRAZED INCONEL JOINTS TIME AT 2150°F AVEBRAGE ROOM- TEMPERATURE TENSILE STRENGTH OF JOINT EFFICIENCY (min) BRAZED JOINTS (psi) (0 252-in specimen) (%) 5 34,900 40 10 40, 100* 46 20 32,900 38 30 31, 400 36 Includes one value of 69,800 psi this i1nvestigation indicate that joint spacing 1s not a critical factor 1in the room-temperature tensile strength of Nicrobrazed Inconel joints, at least within the ranges 1investigated. This range should cover nearly all applications, since the effect of a shrink fit would tend tobe lost during the furnace brazing operation, and a spacing of over 0.020 in. would result 1n extremely poor fitup. Strength of Brazed Joints with Various Base Metals Butt-braze tensile data on the 73.5% Ni1-16.5% Cr- 10 0% S1 alloy indicate that the base- metal composition may be a very 1mportant factor in determining the subsequent tensile strengths of brazed joints., The average of the room-temperature tensile strengths of the Inconel joints was 33,700 psa. Similar tests on brazed type 316 stainless steel showed an average tensile strength of 64,400 psi, which 1s nearly double the value recorded for Inconel. The elevated-temperature tests for both base metals gave excellent results. Joint efficiencies of 99% were obtained for Inconel, and efficiencies of 98% were obtained for type 316 stainless steel, Nicrobrazed joints on stainless steel also had higher tensile strengths than Nicrobrazed joints on Inconel, The average, room-temperature, tensile strength for stainless steel joints was 68,800 psi, whereas the corres- ponding value for Inconel was 34,900 psi. The 1500°F tensile tests again gave evidence of joint efficiencies approaching 100% for both base metals. A summary of a major portion of the butt-braze tensile data obtained thus far 1s presented 1n Table 11.3. The value listed for the room-temperature strength of Nicrobrazed Inconel joints 1s that resulting from a series of tests inwhich some specimens fractured at values nearly equal to the tensile strength of Inconel., Hence, this average value 1s somewhat higher than the values shown in Table 11.2, 1n which no data were recorded that indi- cated such high trends. Also the values of joint efficiency for stain- less steel joints are only approximate because the 0.252-1n.-di1a stainless steel tensile bars have not yet been tested after they were subjected to the brazing cycle. CREEP-RUPTURE TESTS OF STRUCTURAL METALS R. B. Oliver J. W. Woods D. A. Douglas C. W. Weaver Metallurgy Division Preoxidized Inconel 1n Argon. It was previously reported(?) that the environment surrounding the specimen (2)H B Oliver, D A Douglaa, K W Reber, J W VWoods, and C W Weaver, ANP Quar Prog Rep Dec 10, 1952, OBNL-1439, p 159 145 ANP PROJECT QUARTERLY PROGRESS REPORT TABLE 11.3 BUTT-BRAZE TENSILE DATA FOR SEVERAL HIGH- TEMPERATURE BRAZING ALLOYS TEMPERATURE TENSILE JOINT BRAZING ALLOY BASE METAL (°F) STRENGTH | EFFICIENCY (ps1) (%) Nicrobraz Inconel Room 47,000 54 1500 24,500 100 60% Pd-40% Na Inconel Room 87,600 100 1500 21,000 93 Type 316 stainless steel Room 75, 400 92 1500 22,400 88 73 5% N1-16 5% Cr-10 0% S1 | Inconel Room 33,700 39 1500 23,100 99 Type 316 stainless steel Room 64,400 79 1500 25,000 98 15% Ag-20% Pd-5% Mn Inconel Room 55,900 64 1500 19,700 85 during test would have a significant effect on the creep rate and rupture life. Results obtained during the past quarter continue to indicate that the properties are best when the Inconel 1s tested in air, poorest when tested i1n a hydrogen atmosphere, and intermediate when tested in argon. The rupture life observed 1n tests 1in a hydrogen atmosphere was of the order of one-tenth the life observed when tested i1n air, this observation holds for stresses of 4000 psi or less. One hypothesis regarding the long rupture life 1n air i1nvolved a strengthening action by the oxide film formed during the early part of the test., To test this possibility, aseries of specimens was heated in air for 200 hr at 815°C to form such a film. The preoxidized specimens were then tested at 815°C 1n argon, stressed to 2500, 3500, and 4500 psi. In the three tests, the rupture times were far short of the rupture times 1n aair, and were very close to the life when tested i1n argon., It 1s to be noted 146 that the elongation during test was much greater than that observed 1n other environments, and there was no evidence of cracking i1n the failm. Figure 11.5 shows the strain vs., time relationships of the bright and the preoxidized specimens tested 1n argon, These results could mean either that nitrogen rather than oxygen was the controlling factor or, as 1s more probable, that the oxide film formed stress rises when 1t cracked and that this action, 1n turn, promoted the integranular cracking that ultimately led to failure. Peened Inconel 1n Argon Specimens of Inconel have been heavily peened with fine steel shot and are now being tested i1n argon. This mechanical working of the specimen surface has increased the rupture life in com- parison with the life of specimens that did not receive this treatment. Also, the creep rate for the peened specimen 1s much lower than that observed for other specimens at the same stress 1n any atmosphere. It 1s PERIOD ENDING MARCH 10, 1953 UNCLASSIFIED DWG 18659 [ 7301 /‘ 1% ] z 5 © 2 s 2 © g o / 4 & S O A J/ = O bl i % 7 B_‘t ! / o) za A — / "3‘3 / " S e S ,,//’ <& A 3 4 — NN \\ \\\\\ 0 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 TIME (hr) Fig. 11 3., Strain vs. Time Curves for Control and Preoxidized Inconel Specl- mens at 1500°F 1n Argon. improbable that this effect 1s the result of residual compressive stress induced by the peening action, since such stresses would recover at a temperature lower than the test temper- ature, It 1s hoped that metallographic examination will furnish a clue for explaining the 1mproved properties resulting from the peening. Effect of Inconel Grain Size on Time to Rupture As a general state- ment, fine-grained Inconel (0.105 mm 1n diameter) 1s more sensitive than coarse-grained Inconel (0.250 mm in diameter) to the several test environ- ments, Table 11.4 tabulates the comparative responses to these environ- ments at a representative stress level. Effect of Environment on Strain vs Time Curves for Inconel A fine- grained specimen 1s being tested that was run in argon for the fairst 160 hr and then alternately 1n hydrogen and argon atmospheres. Each atmosphere was maintained for about 100 hours. When the hydrogen was introduced, the strain rate was accelerated, and when the argon was re-introduced, the creep rate during the argon cycle remained approximately the same as the last creep rate i1n hydrogen. These results appear to indicate that the effect of the hydrogen atmosphere 1s one of a permanent nature. Metallographic studies do not clearly reveal what the effect 1s, however, the microstructure 1indicates the possible removal or 147 ANP PROJECT QUARTERLY PROGRESS REPORT TABLE 11 4 EFFECT OF GRAIN SIZE ON TENSILE STRENGTH OF INCONEL TEST CONDITIONS TIME TO RUPTURE AT 3500 psi AND 815°C (hr) Tested 1n air Peened and tested in argon Tested 1n argon Preoxidized and tested in argon Tested 1n molten NaF-ZrF4-UF4 (46-50-4 mole %) Tested in hydrogen Fine-grained Inconel Coarse-grained Inconel ~2500 ~2500 835 730 500 730 - 415 470 250 600 solution of carbides and oxides and a much more freauent i1ncidence of intergranular cracking when the tests are conducted 1n a hydrogen atmosphere. The strain vs. time relationships of fine-grained Inconel tested 1in the several environments at 815°C and stressed to 3500 psi are shown 1in Fig. 11.6. The results are representa- tive of several tests i1n each environ- ment The absence of a linear, second- stage section in the strain vs time curves might indicate that the inter- granular cracking characteristic of high-temperature failure 1s 1nitiated very early i1n the test life. The various test conditions may either hasten or retard the formation and propogation of these i1ntergranular cracks and, hence, alter the creep- rupture properties. The program of testing in molten fluoride mixtures continues, but the ¢ .curacy of the results 1s clouded by temperature control difficulties and by the effects procedure, of the descaling The first series of tests was preceded by a sodium descaling operation. The rupture times were very short, and the microstructures 1indicate a stress-corrosion type of failure. Subsequent tests were pre- ceded by a fluoride descaling process. The rupture times were closer to those observed 1n argon, and the micro- structures revealed subsurface voids 148 or hole formation and very laittle intergranular cracking, Only one test of coarse-grained Inconel 1n fluorides showed a longer ruptu-e life than the best l1fe observed for the fine-grained specimens at the same stress. Type 316 Stainless Steel Tube Burst Tests 1n Argon The tube-burst tests run to date were for type 316 stain- less steel tubes loaded internally with argon under pressure and held in an atmosphere of pure argon, A speci- men so loaded has a multiaxial stress system, and the results indicate that the test life under these conditions will be much less than that found for the same material stressed i1n tension only at a stress equal to the maximum stress 1mposed on the tube. FABRICATION OF CONTROL AND SAFETY RODS E. S. Bomar H. TInouye J. H. Coobs R. W. Johnson Metallurgy Division A. Levy Pratt and Whitney Aircraft Division Control Rods for the G-E R-1 Reactor. A study concerning the adaptation of the 1ron-cemented boron carbide compo- si1tion used for the ARE safety rod slugs to the control rods for the G-E R-1 reactor was undertaken at the request of the General Electric Co. In brief, the specifications for these previous progress report(s) included some results of tests conducted 1n argon purified by titanium sponge. The work during this period has in- cluded 100-hr tests i1in vacuum at 800 and 1000°C. This work will also indicate the pressures tolerable in subsequent heat treatments. The data obtained thus far are given in Table 11.6. Future work 1s to include the effect of recirculating the purified gas over titanium sponge and dryaing agents by using a pump 1n a closed system. Oxidation. Tests have been com- pleted on the oxidation of columbium in air containing 18.7 mm water vapor. This 1nvestigation was undertaken after a run made at 400°C 1n air that inadvertently contained water vapor resulted i1n a particularly rapid rate of oxidation. A constant humidity was obtained by passing the air through a saturated solution of NH4C1 at 25°C, the moisture content checked to withan 0 5% of the saturation value when the water absorbed i1n anhydrone was weighed. The results at 400°C have been checked, and they show that the presence of moisture 1ncreases the oxidation rate by a factor of about 400 at the end of 10 hours. At 600°C, a reverse effect was found i1n that the oxidation (3)5. S Bomar and H Inouye, ANP Quar Rep Dec 10, 1952, ORNL-1439, p 157 Prog PERIOD ENDING MARCH 10, 1953 rate decreased significantly (by a factor of 2). It was also apparent that the oxi1dation rate was not linear. Several breaks in the curve of a plot of weight gain vs. time have been observed and duplicated. At 800, 1000, and 1200°C, the oxidation rates 1n moist air are no different from those 1n dry aar. X-ray data have been obtained for the specimens oxidized in air, and thus far the oxides formed are all modifications of Cb,0,, of which there are three,(%) Inconsistencies have been observed for the oxidation rates at 800°C. In these runs, flat sheet oxi1dized to a white scale on the edges, but the other areas formed a rather dense black oxide. A 0.040-1n, wire became entirely white with oxide. Above 800°C, the oxidation rate showed a slight deviation from linearity,. This may have been the result of partial protection of the underlying metal by the thick oxide or a decrease 1n the surface area. The results at 800, 1000, and 1200°C are being checked by oxidizing columbium rod that 1s clad on the circumference with Inconel. It 1s hoped that the i1nconsistencies observed at 800°C can be eliminated by this method, since 1t 1s believed that the i1nconsistencies were caused by the surface areas not being constant. (4)6. Brauer, Z anorg u allgem amount of the data collected 1s 1in- sufficient to permit drawing any definite conclusions. Zirconium (C, K. Talbott, J. M. McCown, Analytical Chemistry Division). Of the several procedures explored for the determination of zirconium, (!’ the simplest and most accurate 1s a volu- metric method employing mandelic acid, This reagent 1s especially good for determining zirconium 1n the presence of uranium. It was decided to further investigate reagents of the mandelaic acid type, and a volumetric procedure was developed employing p-bromomandelic and p-chloromandelic acids. These acids precipltate zirconium 1n acidic solution 1n the form of zirconium tetra p-bromo(chloro)mandelate, which 1s more easily filtered than the zair- conium mandelate, (It 1s possible to then dry this salt and determine zir- conium gravimetrically,(2+3) but the volumetric method 1s simpler to execute, ) The salt dissolves 1n alkali with subsequent hydrolysis of zirconium. However, when an alkali 1s used 1n a nonaqueous medium to dissolve the salt, hydrolysis does not occur, On this basis, a volumetric method for (I)ANP Quar Prog Rep Dec 10, 1439, p 191 :z’n B Hahn, Anal Ches 23, 1259 (1951) 3) R E OQOesper and J L. Klingenberg, Anal Chea 21, 1509 (1949) 1952, ORNL- 172 determining zirconium has been de- veloped 1n which the p-halomandelate salt 1s dissolved 1n excess standard sodium methoxide solution containing 3 parts benzene and 1 part methanol, and the excess base 1s titrated with standard acetic acid 1n a similar benzene-methanol solution with thymol blue used as the i1ndicator. The equation for the reaction that takes place 1s 4Na p-CI1C4H,CHOHCOO + Zr(OCH,), . which appears to be eliminates the need for the use of empirical factors. The Iimited availability of the reagent has made necessary the synthesis of p-chloromandelic acid. A method proposed by Jenkins(*) has been used for this preparation, a yield of 50% 1s obtained. Reduction Products (W. J. BRoss, Analytical Chemistry Division). The addition of zirconium hydride to inhibit corrosion i1n fluoride reactor fuels has led to a study of the re- duction products that appear as a result of this addition, It 1s postulated that some uranium tetra- fluoride will be reduced to the tri- fluoride and that zirconium metal will be formed. A procedure in which 1t 1s possible to determine both these specles, uranium trifluoride and zir- The reaction, stoichiometric, conium metal, 1n the same sample of reactor fuel has been developed. The procedure 1s based on the vast dif- ferences 1n the rates of the following reactions with 0.2 ¥ HF Zr + 6HF —> H,ZrFg + 2H,, fast, (1) 2UF; + 2HF —> 2UF, + H,, slow, (2) Reaction 1 proceeds quantitatively at room temperature, whereas reaction 2 (4)g (1931) Jenkins, J Aa Chea Soc 53, 2341 exhibits no tendency to react under this condition for some time. A study of the rate of reaction 2 1s incomplete, however, upon heating for 1hr at 90°C, only 5.2% of the theoretical amount of hydrogen 1s evolved. These results 1llustrate the stability of uranium trifluoride i1n this medium. When a mixture of zirconium and uranium trifluoride was treated with 0,2 M HF, the volume of hydrogen liberated was equivalent to the amount of zirconium present Zr + 4HC1—> ZrCl, + 2H,,. slow, 4UF, + 4iCl —> 3UF, + UCl, + 2H2, fast. Zirconium liberates 4.9% of the theoretical volume of hydrogen upon heating for 1 hr at 90°C in 9.6 M HCI. Uranium trifluoride reacts gquanti- tatively, as shown by Manning¢(%) 1in his method for the determination of uranium trifluoride. When a mixture of the two 1s treated with 9.6 M HC1, hydrogen equivalent to the total zair- conium and uranium trifluoride present 1s liberated 1n approximately 20 min at 90°C. This reaction can be ex- plained on the basis of formation of a complex 10on by zirconium with the fluoride 10ns that are liberated by hydrochloric acid attack on the uranium trifluoraide. A two-step oxidation reaction 1s, 1n effect, used to determine zirconium and uranium trifluoride 1n the presence of each other. A solution of 0.2 M HF 1s added to the mixture and the volume of hydrogen evolved from this reaction, conducted at room tempera- ture, 1s calculated as zirconium. A solution of 9.6 M HCl 1s then added, with boric acid, to complex excess fluoride 10ons, and the volume of hydrogen evolved upon heating 1s calculated as uranium traifluoraide. (S)D L Manning, W. K Miller, and R Rowan, Jr , Methods of Determination of Uranium Tri- fluoride, ORNL-1279 (May 25, 1952) PERIOD ENDING MARCH 10, 1953 Preliminary results have been ex- tremely promising. The effect of large excesses of sodium, zirconium, and uranium tetrafluorides on these reactions 1s being studied. oxides (J. E. Lee, Jr., Analytaical Chemistry Division). Metallic oxides react with bromine trifluoride(®) according to the reaction 3M02 + 4BrF; —> 3MF, + 2Br, + 30, , where M represents a quadrivalent cation. The metallic oxide content can be determined by measuraing the amount of oxygen evolved, hence, this reaction 1s of potential interest 1in determining the oxide content of fluoride mixtures. A schematic diagram of the apparatus constructed for use 1n studying thais type of reaction 1s shown in Fig. 14.1. The reactor portion 1s designed so that reaction pressures of up to the order of 1 tsi1 can be permitted. The remainder of the system consists of a reagent transfer arrangement, in whaich transfer may be made under vacuum, and a high-vacuum system with the necessary instrumentation for measuring the oxygen liberated during the reaction. Several samples of zirconium oxide have been fluorinated with chromium tri1fluoride 1n this egquipment at temperatures ranging from 200 to 300°C. Analyses of the reaction residues 1ndicate that the experimental con- ditions employed thus far have resulted 1n fluorinations that are at least 80% complete, Efforts are being made to obtain quantitative fluorination by means of such promoters as uranium dioxaide,. Hydrogen Fluoride (D. L. Manning, Analytical Chemistry Division). A possible explanation of the corrosive- ness of reactor fuels that have been treated with hydrogen fluoride 1s the presence of the gas entrapped in the solidified melt. In order to remove this gas from the molten fuel, an (6)1'[ J Emeleus and A A Woolf, J Chen p 164-168 (1950) Soc , 173 ANP PROJECT QUARTERLY PROGRESS REPORT RELIEF GAGE AN ~a—— REACTOR Fig. 14.1. outgassing-with-helium step has been included 1n the fuel preparation pro- cedure. It 1s desired to test the helium off-gas for hydrogen fluoraide and to monitor the gas stream. The most attractive means of attacking this problem appears to be the measure- ment of the conductaivity of the 174 UNCLASSIFIED DWG 18665 VACUUM <+ SYSTEM s — s—TRAP REAGENT Apparatus for Studying Reactions with Bromine Trifluoride scrubbing solutions through whaich the gases are passed, This type of analysis 1s rapid and has the added advantage that 1t may be made at the site of the experiment. Several scrubbing solutions were studied, and the most promising were calcium hy- droxide and boric acid. Calcium hydroxide reacts with hy- drogen fluoride 1n the following manner Ca** + 20H" + 2HF ——> CaF, + 2H,0 and reduces the conductivity of the solution by removaing hydroxyl 1ons. Approximately 100 pg of hydrogen fluoride reduced the conductivity of 100 ml of 7 x 10°* N calcium hydroxide solution by 7 X 10°® mho. A more sensitive determination 1s possible 1f boric acid 1s used as the scrubbing solution. The reaction that occurs between hydrogen fluoride and boric acid 1s H,BO, + 4HF —> H* + BF; + 3H,0 , 1n which a strong acid (fluoboric) 1s formed, therefore the highly mobile hydrogen 1on 1s introduced 1nto the system and the conductaivity 1s in- creased. The concentration of the boric acid solution should be very dilute, 0.1 wt % or less, so that the conductivity of the scrubbing solution will be extremely low, of the order of 2 x 10°% mho. The conductivity of 100 ml of 3 X 10°3 M boric acid 1s increased 13.5 X 10°% mho by the addition of 100 ngof hydrogen fluoride — a change nearly twice that produced by the same amount of hydrogen fluoride 1n calcium hydroxide solution., These results show that conductivity measure- ments are extremely sensitive and well PERIOD ENDING MARCH 10, 1953 suited for the purpose of determining hydrogen fluorade. Corrosion Products (D, L. Manning, Analytical Chemistry Division). Ex- periments have been conducted to ascertain the effects of molten fluoride salt mixtures on Inconel and other metallic containers. In these experiments, the metal container that has been exposed to fluorides 1s made the anode 1n a sulfuric acid-citric acid bath, Since the metals exist as elements, they are electrolytically dissolved and thereby separated from the compounds present, A similar metal tube 1s dissolved for comparative purposes. The residues from the anodic dissolution have been analyzed chemi- cally and spectrographically. The analytical methods used to date have been unreliable because of flaking of metal particles during dissolution. Refinements i1n the electrode system are being made to eliminate this difficulty. Spectrographic data have been collected for Inconel specimens that were subjected to the standard 100-hr static corrosion test with NaF-ZrF,-UF,, with and without addition of zirconium hydride. The results are given i1n Table 14.1. The amount of the data collected 1s insufficient topermit drawing definite conclusions. Several points of The interest are evident, however. TABLE 14.1. SPECTROGRAPHIC ANALYSES OF RESIDUES FROM ANODIC DISSOLUTION OF INCONEL USED IN STATIC CORROSION TESTS OF FLUORIDE REACTOR FUELS COMPOSITION OF RESIDUE FROM DI SSOLUTION OF INCONEL (%) REACTOR FUEL Al Co Sh T1 Zr U NaF-ZrF,-UF4 t ZrH, 01 01 >20 0 2 10 ND* 0 08 10 0 2 06 ND NaF-ZrF,-UF, 01 20 <01 ND 0 08 01 10 0 2 <01 ND Blank 0 08 01 03 0.2 <01 .None detected 175 ANP PROJECT QUARTERLY PROGRESS REPORT si1licon content i1n the residue from corroded Inconel 1s high and i1ndicates that practically all the silicon in the fuel 1s deposited on the walls of the container, the addition of zair- conium hydride results 1n the deposition of some zirconium on the container walls, and deposition of uranium does not occur 1n either case, PREPARATION OF URANIUM TETRACHLORIDE W. J. Ross Analytical Chemistry Division The preparation of several kilograms of uranium tetrachloride to be used in phase studies by the ANP Chemistry Group has been undertaken. The tetrachloride 1s being prepared by chlorinating UO; with hexachloro- propene, according to the directions given by Pitt et al ¢7) PETROGRAPHIC EXAMINATION OF FLUORIDES G. D. White, Metallurgy Division T. N. McVay, Consultant Routine petrographic examinations of more than 800 samples of fluoraide mixtures were made 1n connection with fuel i1nvestigations. Petrographic examinations were made of compositions 1n the UC14-NaCl and UC1, -KCl systems. The data obtained from these exami- nations were correlated with the thermal data to locate compound and eutectic compositions, UC1 ,-NaCl System In the UC1,-NaCl binary system, two compounds were found 2NaCl-UCl,, and a compound with more than 50% UC1,, possibly NaCl-2UCl1,. The latter compound seemed to have an incongruent melting point, since compositions on the high UCl, side of the binary always contain considerable amounts of free UC],. Optical Data. Na,UCI, Pale green to colorless (7)3 M Paitt et al , The Preparation of TCI4 with Hexachloropropylene, C-2 350 3 (July 27, 1945) 176 Uniaxial negative (tetragonal or hexagonal) Refractive indices E, 1 652 0, 1 664 NaU,Cl, Yellow-green Biaxial negative (probably orthorhombic) with small optic angle Refractive indices a, 1 790 ¥, 1 850 ucl,-KCl system. The UCI, -KCl system produced three compounds 2KC1-UCl,, KC1-UCl,, and an 1ncongru- ently melting compound in the 67% UCI, region of the binary. There were two forms that crystallized in the compo- sition range of 30% to 35% UCl,. They were distinguished by their differences 1in birefringence and interference figures. Optical Data K,UCl 1 Pale green to colorless Uniaxi1al negative Refractive indices E, 1 644 0, 1 654 2 Pale green to colorless Biraxial positive, small optic angle Refractive indices a, 1 636 v, 1 640 KUCI ¢ Pleochroism X, grey Z, blue-green Biaxial positive, small optic angle Refractive 1indices a, 1 692 B, 1 705 v, 1759 KU,Cl Yell ow-green Biax1al positive, with very small optic angle Refractive i1ndices a, 1 740 B, 1 809 v, 1 820 X-RAY DIFFRACTION STUDIES P. A. Agron Materials Chemistry Division The determination of the phases present 1n solidified melts of various compositions 1n the NaF-ZrF,-UF, system has been helpful i1n elucidating the paths of crystallization. The regions of this ternary system that are rich 1n NaF have been examined during the past quarter. Some progress has been made here 1n definition of the ternary and pseudobinary phase regions. The petrographic and x-ray dif- fraction 1dentification of the phases found i1n the solid, upon melting given compositions, are listed in Table 14.2. The phases have been grouped according to the similarities present in the respective solidified melts, thus A-1 and A-2 represent one ternary area, B-1 and B-2, the adjacent ternary area, etc, This division i1into ternary areas 1s 1n fair agreement with the 1sothermal contour plots (to be published)., However, the anomalous situation of four phases 1n samples B-1 and D-2 obviously requires further investigation, Additional compositions of these ternary areas should also be 1nvestigated before mapping the phase boundaries. The examination of a number of fused solids along the composition line joining NaUF, and Na,ZrF, appears to 1indicate a pseudobinary behavior. Approximately 90% of the x-ray daf- fraction pattern can be assigned to these two phases. The cooling curves of a series of these compositions (to be published) i1ndicate a binary eutectic at about 22.5 mole % UF,. The behavior of salts along the composition line between NaUF_; and NaZrF, 1s rather complex beyond 6 mole % UF, content, ®) Further study along this composition line, with the (B)P A Agron, X-ray Studies of Phase Segre- gation in Pseudo NaZrF_-NalUF_ Binary as a Function of Heat Treatment in Radiation Stud:ies, ORNL CF-53-1-332 (Jan., 20, 1953) PERIOD ENDING MARCH 10, 1953 use of liquid sampling and quenching technigues, wi1ill be required to ex- plain this area of the phase diagram. SERVICE CHEMICAL ANALYSES H. P. House S. A. Reed A. F. Roemer, Jr. Analytical Chemistry Division Work in the ANP Analytical Chemistry Laboratory during the period consisted primarily of the analysis of alkala- fluoride eutectic mixtures and zir- conium fuels. Although there was a slight decrease in the total number of samples received during the quarter, there was a significant i1ncrease 1n the number of miscellaneous materials recelved. Important among these was a group of samples that were tested for beryllium content. Included 1in the group were large steel sections of test rigs used to study the compati- bility of NaK and beryllium oxide. The specimens were first treated with methanol to remove excess NaK and then leached with hot, 50% sodium hydroxide solutions to dissolve the beryllium ox1de adhering to the surfaces of the specimens. The beryllium content of the alcoholic and caustic solutions was determined colorimetrically, wath p-nitrobenzeneazo-orcinol as the chromogenic reagent, A number of uranium tetrachloride samples were analyzed for UO, and UOCl,. Separation of UO, from the salt was effected by refluxing portions of the material with dilute ammonium oxalate solution. The 1nsoluble UO2 was removed by filtration, dissolved 1in acid, and estimated quantitatively by potentiometric titration with ferrac sulfate. The UOCl,, along with any UO,, was separated from the UCl, by refluxing with anhydrous methyl acetate. The 1nsoluble residue was dissolved 1n dilute nitric acid, and the UOCl, was calculated from a volumetric determination of the chloride content, 177 8L1 TABLE 14.2. TERNARY PHASE REGIONS IN THE NaF-ZrF,-UF, SYSTEMS CHEMICAL PETROGRAPHIC EXAMINATION!®) X_RAY SPECTROMETER DIFFRACTION ANALYSIS SAWPLE | cowposITION EXPERIMENTAL CONDITIONS!®) (mole %) Phases Refractive Index Intensaty of Strongest Line Phaaes Al 55 NaF Salts fused under helium ges Nan(U)Fs 1 504 90 Nal(Zr)F, 20 UF, and starred until frozen NayZrF, 1 43 40 Na,ZrF, 25 ZrF, uo, (1 te 2%} 35 NaZr(U)F, A-2 55 NaF Salts fused under helaum gas NaZr{(U)F, 1 504 80 Nal(Zr)F, 25 UF, and stirred untal frozen Nazszfi 1 43 50 N‘zz'Fs 20 ZrF, vo, (1 to 2%) 35 NaZr (U}F, B1 62 5 NaF Salts equalibrated in closed NnBZr(U)FT 1 39 75 NaSZr(U)FT 18 75 UF, nickel container and cooled NazUFs 1 49 75 NuU(Zr)FS slowly without stirring (e} 18 75 ZrF, Na,Zr(U}F, 1 43 20 By Na,UF ¢! (also ¥ NayUF,) Ne,ZrF, 143 25 Na, ZrF, B-2 62 5 NaF Salta fused under helium and NaZr(U)Fg 1 504 60 NalU(Zr}F, 20 0 UF, stirred untal frozen Na3Zr{U)F, 1 386 30 NayZr(U)F, 17 5 ZrF, Na,ZrF 1 43 30 Na,ZrF, cl 66 5 NaF Saltas fused under helaium and B85 NaUF; and Nal{Zr)Fq 22 5 UF, stirred until frozen 40 ¥ Na,UFg (also ‘Bl and a NazUFG) 11 0 ZrF, 30 NayZr(U)F, cz2 60 0 NaF Salts fused under helium and NaZr(U)Fg (also 1 50 90 NaUFg 35 0 UF stirred unt1l frozen lower index 40 Na,UF 4 material) By NagUF (also ¥ =and B; NayUFy) 5 02rF, 20 NeyZr(U)F,; c3 67 5 NaF Salts fused under helium and Green phase 146t 0 0] 90 B3 NayUF, 27 5 UF, stirred untal frozen U0, (~2%) 15 NayZr(U)F, 50 ZrF, 10 NaU(Zr)F, D-1 72 5 NaF Selte equilibrated in closed NayZr(U}F; (also 139 to 1 40 90 NayZr{U)Fy 13 15 U nickel container and then an internal green 3 Na,UF Fu cooled slowly without starring phase) s '8’ nzUFs 13 75 ZrF, (slso ¥ and a Na,UF,) 10 NasUF, D-2 72 5 NeF Salts fused under helium and NaaU(Zr)F., 1 413 90 NagU{Zr)F, stirred until frozen 25 0 UF, NaF 20 Na,UF, 2 5 ZrFy U0g (1 to 2%} 12 NaF 5 Nay2c{U}F, E 1 75 NaF Salts fused under helium and NajyZr(U)F; 106 NayZr(U)F; and NajU(Zr)F, starred until frozen 15 UF, NaF 10 ZcFg U0, (~2%) 10 NoF E 2 82 5 NaF Salts equilibrated in closed NayZr (U)F, 139 to 1 40 90 NeayZr (U)Fy nickel conteiner and then 8 15 UF, cooled slowly without stirring NoF 75 NagU(zr)Fy 9 75 ZrF, 12 NoF (a) (&) Samples prepaered by C J Barton's group Petrographic examination by G D White, Metallurgy Division (c )fi3-N02UF6 18 a hexagonal form that has not been previously reported, LU0d3Y SSHYO0Ud ATHALUVAD IJAL0Ud ANV PERIOD ENDING MARCH 10, 1953 TABLE 14.3. SUMMARY OF SERVICE ANALYSES NUMBER OF NUMBER OF SAMPLES DETERMINATIONS REPORTED REPORTED ANP Reactor Chemistry Group 508 4477 ANP Experimental Engineering Group 216 2424 ADP Process Improvement Group 41 97 Heat Transfer and Physical Properties Group 6 86 Electromagnetic Research Division 2 15 General Electric Company 11 114 Total 184 7213 Comparative tests of phenylarsonic acid and mandelic acid as precipitants for zirconium were completed during this period. Spectrographic exama- nation of zirconium oxide residues resulting from the agnition of pre- cipitates of phenylarsonic acad revealed the presence of significant amounts of arsenic, antimony, and 1iron. Zirconium oxide from mandelic acid precipitations was found to con- tain only traces of silicon. The use of mandelic acid as the precipitant resul ted 1n much better separations from interfering 1ons and, i1naddition, a considerable saving 1n time. By the use of this reagent during the latter part of the gquarter, a marked reduction in the number of duplicate determa- nations was effected, and a consider- able i1ncrease 1n the accuracy of this determination was attalned. Of the 784 samples reported during the period, 65% were submitted by the ANP Reactor Chemistry Group and 27% by the Experimental Engineering Group. Over 7000 determinations were made during the quarter, a fourth of which were made on nonroutine samples. Summaries of the ANP analytaical chemistry work for the gquarter are given 1in Tables 14.3 and 14.4. TABLE 14 4. BACKLOG SUMMARY Samples on hand, 11-10-52 108 Number of samples received 808 Total number of samples 916 Number of samples reported 784 Backlog, 2-10-53 132 179 Part 1V APPENDIX REPORT NO CF-53-1-111 CF-53-1-148 CF-53-2-159 CF-52-12-153 CF-53-1-84 CF-53-1-260 CF-53-1-276 ORNL-1461 CF-53-1-64 CF-53-1-267 CF-53-1-317 Y-881 CF-53-2-50 CF-53-2-99 ORNL- 1493 CF-52-12-109 (no number) MM- 54 CF-53-1-83 CF-53-2-79 ORNL-1463 15. LIST OF REPORTS ISSUED DURING THE QUARTER TITLE OF REPORT I General Design Minimum Weight Analysis for an Air Radiator Component Tests Recommended to Provide Sound Basis for Design of Fluoride Fuel Reactors for Tactical Aircraft ARE Design Data AUTHOR(S) S Farmer A P Fraas W B. Cottrell II Experimental Engineering An Improved AC Electromagnetic Pump Cell ARE Regulating Rod Moore Pressure Transmitter Test Summary Components of Fluoride Systems A Simple Electromagnetic Flowmeter for Liquad Metals III Reactor Physics Delayed Neutron Damping of Non-Linear Reactor Oscillations Delayed Neutron Activity in a Circulating Fuel Reactor Delayed Neutron Activity i1n the ARE Fuel Element A Graphite Moderated Critical Assembly The Attenuation of Capture Gammas 1n a Plane Limited Medium of Finite Thickness Heating by Fast Neutrons in a Barytes Concrete Shield General Method of Reactor Analysis Used by ANP Physics Group IV Metallurgy and Ceramlcs Spot Welding of Stainless Clad Fuel Elements Formability and Weldability of Vapor-Deposited Mol ybdenum Final Report Progress Report The FlashWelding of Molybdenum Part I - Temperature Distribution During the Flashing Cycle A High-Temperature Cermet Fuel Element Ceramic Fuel Element Radiation Tests Methods of Fabrication of Control and Safety Elements for the ARE and HRE Reactors .Nunber assigned by ANP Reports Office M E LaVerne E R Mann S H Hanauer P W Taylor W B Cottrell R M Carroll W K Ergen H L F Enlund H L F Enlund E L Zimmerman F Abernathy H L F Enlund F Abernathy H L F Enlund C B Mills G M Slaughter Massachusetts Institute of Technol ogy Rensselaer Polytechnic Institute R Johnson Johnson R H Coobs S Bomar o BN R DATE ISSUED 1-31-53 1-16-53 2-18-53 12.7-52 1-9-53 1-22-53 1-27-53 to be 1ssued 1-7-53 1-27-53 1-27-53 12-7-52 2-6-53 2-11-53 to be 1ssued 12-9-52 10-31-52 12-5-52 1-9-53 2-9-53 to be 1ssued 183 ANP PROJECT QUARTERLY PROGRESS REPORT REPORT NO ORNL-1491 CF-52-11-205 CF-52-12-124 Y-B4-59 CF-53-1-233 CF-53-1-248 CF-53-2-56 CF-53-2-84 ORNL-915 CF-52-11-129 CF-52-12-209 CF-53-1-279 ORNL-1438 ORNL-1471 ORNL-1479 ORNL-1430 Y-B31-390 ORNL- 1453 Y-B32-104 184 TITLE OF REPORT Corrosion by Molten Fluoraides AUTHOR(S) . S Raichardson D C Vreeland ¥ D Manly V Heat Transfer and Physical Properties Estimates of Heat and Momentum Transfer Characteristics of the Two Fluoride Coolants (L1F-48 M%, BeF-52 M%) and (NaF-10 M%, KF-46 M%, ZrF,-44 MZ) Generalized Velocity Dastribution for Turbulent Flow 1n Annula Selected Physical Properties of Potassium and Potassium Hydroxide in the Temperature Range 100 to 1000°C A Literature Search Measurement of the Thermal Conductivity of Fluoride Mixtures No 14 and No 30 Remarks on the Falling-Ball Viscometer Heat Capacity of Fused Salt Mixture No 30 Heat Capacity of Sr(OH), Forced Convection Heat Transfer i1n Thermal Entrance Regions vVl Shielding Materials Research for Mobile Reactor Shieldaing Thermal Neutron Dose at the Crew Compartment The Tower Shielding Facilaity Determination of the Power of the Bulk Shielding Reactor Part 2 Shield Optimization A Proton Recoil Type Fast Neutron Spectrometer VII. Chemistry Modifications of the Dimethylglyoxime Method for the Colorimetric Determination of Nickel Based on the Use of Potassium Persulfate as the Oxidant Analytical Chemistry-ANP Program Quarterly Progress Report for Period Ending November 25, 1952 Dissolution of NaK Minutes of Committee Meeting for the Coordination of Hydroxide Research, Fourth Meeting, December 10 and 11, 1952 H F Poppendiek W B Harraison J 0O Bradfute R V Bailey R L Curtas S J Claiborne R F Redmond S I Kaplan W D Powers G C Blalock W D Powers G C Blalock W B Harraison E P Blizard J L Meem E P Blaizard E B Johnson J L. Meem E P Blizard R G Cochran K M Henry M L Druschel 0O Menas R BRowan, Jr Analytaical Chemistry Division J C White C K Talbott L J Brady F Kertesz DATE ISSUED to be 1ssued 11-29-52 12-19-52 9-23-52 1-8-53 1-14-53 2-6~53 2-9-53 to be 1ssued 11-18-52 1-27-53 to be 1ssued to be 1ssued to be 1ssued 11-3-52 11-20-52 12-2-52 1-7-53 -y REPORT NO ORNL-1376 ORNL- 1490 ORNL-1495 Y-B32-103 CF-52-12-178 CF-53-1-129 ORNL-1476 ORNL- 1500 Y-B31-403 ORNL-1439 CF-53-2-126 CF-53-2-246 PERIOD ENDING MARCH 10, 1953 TITLE OF REPORT Mass Spectrometer Investigation of UF, General Information Concerning Fluorides, A Literature Search, Suppl to ORNL-1252 General Information Concerning Hydroxides, A Literature Search, Suppl to ORNL-1291 Possible Coolants for Solid Fuel Reactors Hydroxide Systems Fused Salt Compositions An Indirect Colorimetric Method for the Deter- mination of Uranium Recovery of Uranium as a Single Product from Fluorides VIII Miscellaneous A Guide for the Safe Handling of Molten Fluorides and Hydroxides ANP Project Quarterly Progress Report for Period Ending December 10, 1952 Objective and Status of ORNL-ANP Reactor Program ANP Information Meeting of February 18, 1953 AUTHOR(S) L O Gilpatraick Russell Baldock J R Sates M E Lee M E Lee W R Graimes M D Banus C J Barton J C White D L Manning C F Coleman Reactor Components Safety Committee W B Cottrell R C Briant W B Cottrell DATE ISSUED 8-29-52 to be 1ssued to be 1ssued 11-25-52 12-23-52 1-15-53 1-13-53 to be 1ssued 1-12-53 1-15-53 2-13-53 3-2-53 185 -1 CHART OF THE TECHNICAL ORGANIZATION OF THE AIRCRAFT NUCLEAR PROPULSION PROJECT AT THE OAK RIDGE NATIONAL LABORATORY MARCH ,1953 Thi maseri | contai | formerien slfacting the aetlan- 14 hense [ the United Stata withl the maenl g § the pronegelw Til 18 USC., T94 the monami o or vel Hon of whi h i eny % an umeutiori od porson | pechibiled by lew ANP PROJECT DIRECTOR R C. Briont RD D Hilyor * Sec ANP ASSOCIATE DIRECTOR FOR ARE H Buc RD ASSISTANT DIRECTOR FOR COORDINATION A J Miller ANP TECHNICAL ASSISTANT ¥ B. Cottrell ANP J M Ciser* ANP P Horman * Sec. ANP ANP LIBRARY ASSISTANT TO DIRECTOR Burlding 9704 | W Cordwall RD L i Cook ANP M. Brown RD P Horman Sec. ANP E Certer RD E Webster RD i STAFF ASSISTANT FOR PHYSICS A 1 Maller ANP STAFF ASSISTANT FOR DESIGN ARE PROJECT ENGINEER 1 H Buck RD, STAFF ASSISTANT FOR METALLURGY STAFF ASSISTANT FOR CHEMISTRY W K Ergen ANP D Hulyer * Sec ANP A P Froos ANP E S Betis ANP P Hormen Sec ANP R C Briom RD W D Monly " ¥ R Grimos WC SHIELDING RESEARCH REACTOR PHYSICS RADIATION DAMAGE GEMERAL DESIGN ARE OPERATIONS EXPERIAENTAL ENGINEERING HEAT TRANSFER AND PHYSICAL METALLURGY CHEMWISTRY uilding 4500 Building 9704 } Building 3025 Building 5704 § Building 7503 Building 9201 3 PROIEE“RIZIES g%fimcn Bunlding 2000 Building 9733 3 1lding E ff"',j:fil § W K. Ergen ANP D S Billingron 55 A P Fraas ANP E S Botnis ANP H W Sovege ANP H F Poppendisk REE ¥ D Menly M W R Grimes MC LID TANK F H Abemath ANP R. G Affel ANP G M Adomson M J Q. Brodiute REE E S. Bomar M P A Agron [ Bu1d g 3000 R B Boveyos AN £ D Bevemrn % R W Bussard | ANP A Crsty Eu 1P Bloksly MC § L Caherr REE W H Budges W C J Barten ue 10 Flynn p C. 8 Mlls ANP W E Brundoge S H. L F Enfynd [ G W Eckerd EM G D Brody M W S Farmer REE J ¥ Cathcort L F F Biankenship NC F G Prohammor* ANP R M Cerroll 55 B E Hill Yz H. L F Enlund* P 4 Cisar ANP D C. Hamiton REE J H. Caobs " C M Biood MC T V Blosser P R Willioms Sec ANP A F Cohon 5 M. E LaVerna ANP B L Greonstroat ANP ¥ G Cobb ANP H ¥ Hoffman REE D A Dougles M R A. Bolomey WC G. T Chapman P C P Coughlen ANP J 1 Lai REE LD r ] F P Bood MC I W Mller P W W Dovis 5 WL Anp B Holuchan & P F E Lysch REE E E Heft u L M. Bratche MC 2L Srome GE COMPUTERS M. J Feldman sS ¥ L Scont ANP H. Homyg GE R E Engberg I;R, E b REE £ Hoffman " v : atcher F N oo A Forbes ANP N E Hinkle ss H ) Stumpf ANP J' P Jackson EM W C George 2 L D Falmer REE H lnouye B v Colemon s atsen . Teagoris ANP G ¥ Koilhaltx 58 A, G Johnson GE A G Grindell AHP D Powers A Cuneo K J Kelly =" CONSUL TANTS ER * I R. Helten ANP T Sefton Sec. REE R B Oliver M D D Davies 11 TECHNICIANS P R Klem 55 E B Perrn ANP W R Huntley ANP TECHNICIANS P Patriorco M J E Eargon W 3 Moynord P 3 G Morgon ed A H Fox Union Callege D Seart ANP E M Less ANP G. M. Sloughter M E E Ketchen MC J N Mocay P T bor 5 R L Mexwall Univ Tem H L Wy LRP R E MacPherson ANP C.G Blalock REE G P Smth M R P Metcolf NC gan G, F Wislicenus Johns Hopkins Univ LA ANP R M Burnent REE D C Vrestand I RE . MC BULK BUELOWG REACTOR - g":' s§ 4 H Myld Reuction Mators Inc § B it e R M. Mazon EM S J Cluibor REE 3 W Woods M L G. Overholsar uC Bu 19 3 301D CRITICAL EXPERIMENTS A E porgnsen 3 E Wrachhusan EW ¥ B McDonald ANP T N Josss REE J Thomas Sec u H.S. Powers MC [ Building S113 HE R;enm 3 M, Milligan, Sec. ANP G. N, Neasle uC 3 Lomes REE J D Redman NC 4 L beom P W T Rebwson 55 ¥ R Oshomn ANP G M ¥um REE TECHNICIANS R. I Sheil uC R G Cochron P A D Callhan P 0 S:sman 3 CONSUL TANTS G. F Petersen ANP G D Brady W B J Stum MC M. P Haydon [ F W Smh oy ¥ R Chombers Unv T D F Solmon ANP J T Eost M R E Thoma Jr MC K. M. Hanry P — R C Keen P ¥ J Stwm b 1 W Walker Unre Tern P G. Smith ANP G i Gonzaler " L E Topol MC H E Hungerford P 4 J Lo P L C Templeten s$ A L Sow ANP L L Hall M R.E Trober MC E B Johnson p J W Noaks P 1B Trice 55 P ¥ Toylo ANP J D Hudson M D E Coldwell Sec. MC J K Leshe P D VP Wilioms P C C. Wobster 55 J . Trummel ANP Building 9766 R. W Johnson M G M MeC p E L Zimmerman p W C, Tunnell ANP ™ ¥ G Lare M CORRGHION LABORATORY ommon W R Willis sS W Pagle cy B.5 Davia Sec Yn 1 C Wlson 55 D R Ward ANP 1M Warde " B, McNobb " By ldung 9764 E Rolor cy 3 C Zukas I REACTOR CONTROL J F Warner* Y12 J E Pope "] F Kertesz MC R Bullard Sec P Budlding 4300 G D Whitman ANP C E Cums* M C. E Schubort " H I Buttrom NC TECHNICIANS e D Alexonder Sec ANP L & Doney M L R Trorter " C R Coh TECHNICIANS E P Epler* D Storay Record Clerk ANF S D Fulkereon* ] C K. Thomas M E Moodows K Honeycuti P C Ells §S D Homis Stanc. ANP J R Johnaon* M C W Weaver M F A Knox MC J L Hll P 6 MEV VAN DE GRAAFF CROSS SECTIONS F M. Blocksher 5§ A E G Bates* Ic A J Taylar® M MY Smih NC D J Kby P Budlding 5500 T E Cale* Ic TECHHNICIANS G D White* M METALLOGRAPHY R M. Simmons P HB Willard P 861 CYCLOTRON IRRADIATIDNS F P Groen* P — 1S Addison ANE TECHNICIAN R S Crouse M CONSULTANTS J K Bau* p Building 92012 $ H Honauer* P G 5. Chilton AN J A Goff M T K Roche [ J . Carter TOWER SHIELDING FACILITY ’ ER P 4 M Coburn ANP e T R P Gibb Willioms Coll J D Kington* P R § Lwvingston* ER n T E Crobt ANP TECHNICIANS \ clloge C E Chiford P P H Statson p J B Ruble* P obtres D G H:ll Duke Univ — A L Boch® ER 1 J Stone* P J R Croley ANP CONSULTANTS WD Allen M T ¥ Blosser P ENGINEER o ER L C. Ookes* P J Didiake MC T M, McVay Umy Al N M, Archioy M CONTRACTORS J Y Estabrock P R ¥ Lamohere* P R J Jones C § Walker* RP F A Doss ANP T Ot Sren E R Boyd M Baitell i L B Holland P phere R L K‘“fh' E: J R Duckworth ANP T S Shevlin Ohio Srate Univ D H Brawor o tolle Memecrial Institvte M K Hullings P TECHNICIAN F H Nedl T L Grogory ANP W H. Farmer " 'JR W Clegg ¥ T Hawton Ic F M, Grizzell ANP R. L, Fitzgerald* M on RADIATION $PECTROSCOPY RESEARCH CONSULTANTS B L Johnzan ANP J C Gowsr " R J[:m' FC Mo H. G MacPherson Mational Carbon Co Inc J W Kingsley ANP E. P Griggs M H, A, Pray renschein P L P Smith Cotnall Unsv DE M:Eny ANP R M ¥ollace M ond others T A Lave P D F Weekes Texos A& M G E Milts ANP Metal Hydrides | F J Mucksnthaler P J 1 Porsons ANP CONSULTANTS l:.D B;n.‘ ne M, A, Redden ANP N J Gront MIT v msonfl:f:: Aavsts F J Schofer ANP J L Gregg, Cornell Univ and athers F H Mooy NOTE Thiz chart shows only the lines of technrcal coordination of the ANP project R. G, Wiley ANP 5 'F: ;‘;:P“ :g: Universaty of Arkansas A Simon P The various sndividuals ond groups of paople hated are engoged mither wholly o port CONSULTANTS EC '"::: Univ Alo :’ ¥ Gugoneff H E Stern cv tima on resoorch ond design which ts coordinotad for tho benofit of the ANP propct in J Smathers L Gordon Sec P J F Boiley Univ Tenn CONTRACTORS ond oihers the manner indicated on the chart Eoch group however s dlso responmible 1o its ¥ R Ctombers Umyy Tean. CONSUL TANT Diviscon Director for the detoiled progross of 115 research and for admimistrative motters J F Haines Boldwm-Lima Hemilton Corp H A Bethe Comell Uniy ¥ K Stawr Univ Tom. F G Tomall he! CONTRACTOR The key to the abbraviations used 13 givon balow It should be ooted that several out and athers AMALYTICAL CHERISTRY Nucloor Development Associotes Inc 2ide argamizations hove parsonnal participating in the GRNL—ANP progrom. Battelle Memonic! Institute Butlding 9733 2 {AEC Controctor) H Goldstain AC Analytical Chomistry Division R K Porke C.D Suaano AC R Aronaon ANP Aurcraft Nuclear Propulsion Division H Sall: - C Chemistry Divizion Z Schoheald H P House* AC cy Consolidated Vultee Aircraht Carp. and others i-? ?{'d » :g EM Enginwering and bowdenonce Division Gerity Michigan Ca, J C. White AC ER Eloc Ressorch . gy e d others romagnetic 1eision G. Geoff an s o aC GE Ganeral Eleciric Ca and others B. Young, tc Instrumemdation and Cortrols Division CONSUL TANT Massachysatts Inahivite of Tochnol LRP Long Range Reactor Planning Geoup 7 Walff oar H H Willord Univ Much. M Metollurgy Division and ofhers NC Materials Chermatry Division SPECTROGRAPHIC ANALYSS P Physics D vision Renzssloer Folytechnic nshitute IR IlhN“Blrild'nm9‘1‘.!-1 o P Pratt and Wnztney Asrcrott Division E F Nippos AN v . st RO Resvarch Derector s Doportmen: and othars and ;Th‘:rl REE Reoctor Exparimental Enginesring Divizien Superior Tube Co. MASS SPECTROMETRY v el = EiT olid Jtate Divikion Y12 Corbide and Corbon Chemscals Co. {¥ 12 Site} Commonwealth Enginsering Ca, of Ohio C. R. Boldeck 1] *Port tuer M. J Hileo i87