2 o ol ORNL-1154 This document consists of 225 pages. Copy §i of 203 . Series A Contract No. W-7405, Eng-26 AIRCRAFT NUCLEAR PROPULSION PROJECT QUARTERLY PROGRESS REPORT for Period Ending September 10, 1951 R. C. Briant Director, ANP Project C. B. Ellas Coordinator, ANP Project Edited by: W. B, Cottrell DATE 1SSUED: OAK RIDGE NATIONAL LABORATORY Operated by CARBIDE AND CARBON CHEMICALS COMPANY A Division of Union Carbide and Carbon Corporation Post Dffice Box P 0ak Ridge, Tennessee ORNL-1154 Progress Report INTERNAL DISTRIBUTION 1 G. T. Felbeck (C&CCC) 37 R. W. Stoughton 2-3 Chemistry Library 38 F. R. Bruce 4 Physics Library 39 H. W. Savage 5 Biology Library 40 W. K. Eister 6 Health Physics Library 41 A. S. Householder 7 Metallurgy Library 42 C. B. Graham 8-9 Training School Library 43 R. N. Lyon 10-13 Central Files 44 C. P. Keim 14 C. E. Center 45 W. R. Gall 15 C. E. lLarson 46 A. J. Miller 16 W. B. Humes (K-25) 47 R. W. Schroeder 17 W. D. Lavers (Y-12) 48 D. S. Billington 18 A. M. Weinberg 49 E. P. Blizard 19 E. H. Taylor 50 C. E. Clifford 20 E. D. Shipley 51 G. H. Clewett 21 £. J. Murphy 52 A. D. Callihanr 22 F. C. VonderLlage 53 R. S. Livingston 23 R. C. Briant 54 W. D. Manly 24 J. A, Swartout 55 J. L. Meem 25 C. B. Ellis 56 C. D. Susano 26 A. H. Snell 57 W. B. Cottrell 27 A. Hollaender 58 W. M. Breazeale 28 F. L. Steahly 59 W. R. Grimes 29 K. Z. Morgan 60 A. Brasunas 30 D, W. Cardwell 61 N. M. Smaith 31 M. T. Kelley 62 H. F. Poppendiek 32 E. M. King 63 F. C. Uffelman 33 C. E. Winters 64 D. D. Cowen 34 J. A. Lane 65 P. M. Reyling 35 J. H. Buck 66-75 ANP Library 36 J. P. G111l 76-81 Central Files (0.P.) 82-103 104-113 114 115-122 123 124-126 127 128 129-134 135 136 137-141 142-145 146 147 148-151 152 153-156 157-159 160 161-162 163-166 167 168-169 170-171 172 173 174-175 176-179 180-188 189-203 ORNL-1154 - Progress Report EXTERNAL DISTRIBUTION Aircraft Nuclear Propulsion Project, 0Oak Ridge Argonne National Laboratory Armed Forces Special Weapons Project (Sandia) Atomic Energy Commission, Washington Battelle Memorial Institute Brookhaven National Laboratory Bureau of Aeronautics Bureau of Ships Carbide and Carbon Chemicals Company (Y-12) Chicago Patent Group Chief of Naval Research duPont Company General Electric Company, Richland H. K. Ferguson Company Hanford Operations Office Idaho Operations Office Towa State College Knolls Atomic Power Laboratory Los Alamos Massachusetts Institute of Technology (Kaufmann) Mound Laboratory National Advisory Committee for Aeronautics, Cleveland National Advisory Committee for Aeronautics, Washington New York Operations Office North American Aviation, Inc. Patent Branch, Washington Savannah River Operations Office University of California Radiation Laboratory Westinghouse Electric Corporation Wright Air Development Center Technical Information Service, Oak Ridge data as of 1946. 111 iv Reports previously issued in this series are as follows: ORNL-528 Period Ending November 30, 1949 ORNL-629 Period Ending February 28, 1950 ORNL-768 Period Ending May 31, 1950 ORNL-858 Period Ending August 31, 1950 ORNL-919 Period Ending December 10, 1950 ANP-60 Period Ending March 10, 1951 ANP-65 Period Ending June 10, 1951 Ll TABLE OF CONTENTS SUMMARY _ Part 1. REACTOR THEORY AND DESIGN 1. THE AIRCRAFT RFACTOR EXPERIMENT . Core Design | Fluid Circuit Heat-disposal circuit : Contamination of helium system Control of the ARE Liquid-fuel control system Solid absorber rod design Electronic computor design Electrical Circuit Remo te-Handling Equipment Building Facility for the ARE 2. EXPERIMENTAL REACTOR ENGINEERING Liquid-Fuel Systems Pump Development Centrifugal pumps for figure-eight loops ~ARE pump design Canned-rotor pump Level tank pump Electromagnetic pumps Test Loops Calibration loop Sodium manometer loop Seal Tests Frozen-sodium seal Bellows seal Graphitar and tool steel seal Instrumentation Stress-Rupture Tests Self-Welding Tests Valve-Packing Experiments 17 PAGE ~NO. 10 10 10 11 11 11 12 12 12 12 i4 14 15 15 16 17 17 20 20 20 20 21 21 21 23 23 24 24 Contamination of Liquid-Metal Systems Cleaning of liquid-metal systems Purification of liquid metals Purification of inert gases Analytical results with sodium Investigation of Sodium Condensation ARE Component Tests Heat-Exchanger Tests Building Modifications and Experimental Facilities Alkali Metals Manual 3. REACTOR PHYSICS IBM Calculations Production report The multiregion reactor problem Cylindrical multigroup calculation The effect of the borom blanket in the ANP reactor Statics of the Aircrafc Reactor Experiment Summary of calculations on the ARE Estimated critical mass of the ARE Control rod effectiveness Kinetics of the Aircraft Reactor Experiment Mass-reactivity coefficient Fuel conductivity Neutron lifetime Start-up accident Coolant temperature Preparatory Physics Calculations Correction to "Wigner formula for resonance escape" probability Effect of "Wigner formula for resonance escape’™ correction The transmission coefficient of the B,C curtain in the ANP and ABE reactors Heating in the boron carbide curtazin in the AMP reactor Temperature savings in the ANP reactor The Sodium Hydroxide Reactor 4. CRITICAL EXPERIMENTS Critical Assembly of Air-Water Reactor PAGE ND. 25 25 26 26 26 27 27 28 28 30 31 32 32 33 33 33 34 34 36 37 48 49 49 56 36 60 60 64 66 66 67 70 19 79 Critical Assemblvy of Graphite Reacter Part II. SHIELDING RESEARCH BULK SHIELDING REACTOR Reactor Operation Mock-Up of the Unit Shield Mock-Up of the Divided Shield LID TANK DUCT TEST SHIELDING CALCULATIONS Analysis of Bulk Shielding Reactor Neutron Data Interpretation of Lid Tank Gafima-flay Data Shielding Calculations for the ARE Activation of nitrogen and helium in the ARE reactor pit Activation of impurities in BeO : Detection of leaks in the fuel elements by means of radiocactive tracers ‘ NDA Divided-Shield Studies Use of NH, as a Shielding Material Mechanical requirements of an ammonia shield Calculational procedure & Comparison of shield weights using NH; and H O NUCLEAR MEASUREMENTS The 5-Mev Van de Graaff Accelerator Measurement of the (n,2n) Reaction in Beryilium Bremsstrahlung from Li8 Radiation detection equipment Bremsstrahlung activity Time~of-Flight Néutron Spectrdmeter . PAGE NO. 80 83 B3 83 83 87 89 91 91 92 92 92 93 93 93 94 95 95 96 97 97 97 98 98 98 100 S vil Part III. MATERIALS RESEARCH 10. CORROSICN RESEARCH Static Corrosion by Fluoride Fuels The pretreatment process Corrosion of structural metals Corrosion of platinum Static Corrosion by Moderator Coolants Capsule technique with hydroxides Corrosion by sodium hydroxide Corrosion by potassium hydroxide Corrosion by other hydroxides Corrosion by binary hydrogenous systems Static Corrosion by Fluoride Coolant Mixtures Static Corrosion by Sodium Static Corrosion Test of a Reactor System Mass Transfer Phenomenon in Static Corrosion Chemical-reaction mechanism for mass traansfer Experimental evidence Dynamic Corrosion Tests in Thermal-Convection Loops Corrosion by lithium and lead Corrosion by seodium Dynamic Corrosion Tests in Forced-Convection Loops 11. PHYSICAL PROPERTIES AND HEAT-TRANSFER RESEARCH Investigation of Free Convection Within Liquid-Fuel Elements Theoretical analysis of natural convection Measurement of the fuel-element temperature distribution Measurement of the fuel-element velocity distribution Physical Properties Heat capacity Thermal conductivity of liquids Thermal conductivity of solids Falling-ball viscometer Brookfield viscometer Vapor pressure Density Heat-Transfer Coefficients vili PAGE NO. 103 104 104 106 110 112 113 113 113 118 118 118 118 122 122 122 123 124 124 129 131 132 132 132 133 134 134 134 135 135 136 136 136 137 137 12. 13. Heat transfer in fused hydroxides and salts Heat transfer in boiling-liquid-metal systems Heat transfer in molten lithium METALLURGY AND CERAMICS Welding of Inconel Tensile tests Fatigue tests All-weld-metal tensile tests Corrosion of welds Special weld tests Welding of Molybdenum Creep of Metals in Controlled Atmospheres Creep of 316 stainless steel Creep of inconel : Creep of niobium Creep test of loaded inconel tube Stress-Relaxation Tests Fuel-Element Fabrication Hot rolling Mechanically formed matrix Loose-powder sintering Rubberstatic pressing Compatibility of potential fuel-element materials Control-Rod Fabrication Metal Cladding of Beryllium Oxide Refractory Metals Ceramics Laboratory CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS Fuel Development Low-melting fluoride systems Preparation of UF, Homogeneous fuels Moderator~Coolant Development Preparation of pure sodium hydroxide Preparation of other hydroxides PAGE 137 138 139 141 141 141 143 144 145 145 145 145 146 146 146 146 147 147 147 149 149 149 151 151 152 152 152 154 154 155 159 161 162 162 165 1X Decomposition pressures of hydroxides Binary hydroxide systems Hydroxide-fluoride systenms Hydroxide-borate systems Coolant Development 14. RADIATION DAMAGE Creep Under Irradiation Radiation Effects on Thermal Conductivity Irradiation of Fluoride Fuels Pile irradiation of fuel capsules Cyclotron irradiation of fuel capsules Corrosion of Iron by Lithium Under Cyclotron Irradiation Liquid Metals In-Pile Loop Part IV. ALTERNATE SYSTEMS 15. SUPERCRITICAL WATER REACTOR Outline of a Specific Design Fuel Elements and Assemblies Reactivity Stabilaty Pressure Shell Shield 16. CIRCULATING-MODERATOR-COOLANT REACTOR: HKF Operational Characteristics Reactor Characteristics Core Design Reactor Physics 17. CIRCULATING-MODERATOR-COOLANT REACTOR: ORNL Fluid-Circuit Specifications PAGE NO. 165 166 167 168 168 170 170 171 173 173 173 174 174 179 179 180 180 182 183 183 185 185 186 186 187 188 188 Fuel-Element Design Reactor Design 18. HIGH-TEMPERATURE POWER PLANT STUDIES Sodium-Liquid-Vepor Cmmprassorljet Helium-Cooled Reacters Part V. APPENDIXES 19, ANALYTICAL CHEMISTRY Analysis of Reactor Fuels Spectrographic results Development of colorimetric metheds Determination of platinum Oxygen in Sodium Oxygen in Lead Oxygen in Helium . Oxygen and Nitrogen in Lithium Carbon in Lithium Uranium Trifluoride in Uranium Tetrafluoride Identification of Residue 1in Lithium-Metal Coolant System Analytical Services 20. LIST OF BEPORTS ISSUED ‘PAGE NO. 188 188 192 1923 194 199 200 200 200 201 201 201 201 1202 202 202 203 - 204 205 X1 3.4 3'5 8.1 10.1 10.2 10.3 10.4 12.1 12.2 12.3 13.1 13.2 13.3 13.4 13.5 13.6 13.7 x11 LIST OF TABLES TITLE Composition and Design Data of Core 93 Volume Fractions of the Materials in the Side and Bottom Reflectors Summary of Calculations on the ARE Core Having 25 1b of Uranium and Reflector 520 Calculation of Uranium Requirement for the ARE Design of 10 June 1951 B4C Transmission Coefficients Comparison of Shield Weights Using NH; and H,0 Summary of Fluoride-Corrosion Data Obtained in 100 hr at 1000°C (1830°F) Summary of Corrosion Data Obtained in 100-hr Tests at 815°C with PAGE NO. 36 37 38 41 67 96 111 Mixtures of Sodium Hydroxide with Sodium Hydride, Sodium, or Water 119 Corrosion and Operational Data on Lithium-Containing Thermal- Convection Loops Corrosion and Operational Data on lead-Containing Thermal- Convection Loops Room-Temperature Tensile Properties of Inconel Tube-to-Header "Pairs" Inconel All-Weld-Metal Tensile Values Results of Investigations of ZrO, as a Control Material Summary of Promising Fluoride Fuel Systems Solubility of Uranium in Hydroxides Solubility of Uranium in Mixtures Consisting of Sodium Hydroxide and Sodium Tetraborate Purification of NaOH by Recrystallization from Ethyl Alcohol Purification of NaQOH by Recrystallization from H,O Decomposition Pressures of Ba(OH)z'(1.9)H20 Low-Melting Non-Uranium Fluoride Eutectics 125 126 142 144 152 156 161 162 164 164 166 169 TABLE PAGE NO. TITLE NO. 14,1 Results of 31-Mev Alpba Irradiation of Lithium in Iron Capsules 175 15.1 Summary of Reactor Design-Point Values 181 17.1 Design Coolant Condition for Maximum and Cruise Power 190 17.2 Temperature Throughout Sodium Hydroxide Core 190 19,1 Determination of Uranium Trifluoride by Two Methods 203 19,2 Backlog Summary | 204 X111 FIGURE NO. 2.3 2.4 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 X1V LIST OF FIGURES TITLE Fluid-Bearing (Canned-Rotor) Pump Comparison of General Electric and ANP Electromagnetic Pump Data (750°F, Sodium) Centrifugal Pump with Frozen-Sodium Seal Experimental Facilities Schematic Drawing of ARE Core Arrangement kefif vs. Uranium Weight for the ARE Reactivity of the ARE Core Backed by Various Reflectors or Additional Core Material Reflector Savings of Various Reflectors Backing the ARE Core Spatial Distribution of the Lethargic Average of the Fast Flux Spatial Distribution of the Thermal Flux in the Core Flux Spectrum in the ARE Leakage Spectrum from the ARE Core to the Reflector l.eakage Spectrum fromthe ARE Reflector Spatial Power Distribution in the ARE Radial Thermal-Flux Distribution in the ARE Reactor with Seven Control Rods; Placement No. 1 Radial Thermal-Flux Distribution in the ARE Reactor with Seven Control Rods; Placement No. 2 Radial Thermal-Flux Distribution in the ARE Reactor with Seven Control Rods; Placement No. 3 Response of Flux to a Step Increase in Reactivity for Various Reactivity Changes Response of Fuel Temperature to Step Increase in Reactivity for Various Reactivity Changes Phase Plot of Flux vs. Fuel Temperature for Various Step Increases in Reaectivity Power Response of ARE to Step Reactivity Change of 0.009125 with Various Mass-Reactivity Coefficients PAGE NO. 18 19 22 29 35 37 39 40 42 43 44 45 46 47 418 48 48 50 51 52 53 FIGURE NG. 3.18 3.19 3.21 3.22 3.23 3.24 3.25 3.26 3.27 3.28 3.29 5.1 502 6.1 - 7.1 1.2 8.1 9.1 9.2 10.1 10.2 TITLE Response of Fuel Temperature to a Step Increase of Reactivity of 0.009125 for Various Effective Mass-Beactivity Coefficients Phase Diagram of Flux vs. Fuel Temperature for Response of ARE to a Step Increase in Reactivity of 0.009125 Response of Flux to a Step Increase in Reactivity of 0,002 for Various Average Neutron lLifetimes Phase Diagram of Relative Fxcess Power (Flux) vs. Fuel Temperature Phase Plot of Flux (or Puwér) vs. ¥uel Temperature Neutron Abscorption vs. Penetration with the B,C Layer of the ARE Abserption Spectrum for thfi 200-megawatt Sodium Hydroxide Reactor Leakage Spectrum for the 2fi0~megawatt Sodium Hydroxide Reactor Fission Spectrum for the Bare Water-Moderated Reactor Leakage Spectrum for the Bhre Water~Mederated BReactor Fission Spectrum for the Rgflected Water-Moderated Reactor Leakage Spectrum for the Reflected Water-Moderated Reactor Preliminary Gamma-Bay Specfirum at 130 em from the Water- Reflected Reactor | Bulk Shielding Facility Camparison of Shields for:3~ft Reactor NeUtTOH Attennation in Water Areund Duct §Water Patch Around Duct ;Typical Composite Shie]d!§H20~PthH3 Bremsstrahlung {rom Efl“pimfl Lithium Loop Absorption of Bremsstrahl@ng from [.i? Beta Rays Inconel Specimen from Pretreating Pot in Which Fluoride Bath Mixture was "Deactivaced”™ at 950°C {for 100 br Corrosion of Inconel by 3INa¥F-UF, PAGE NO. 54 55 57 58 59 68 73 T4 15 76 7 78 85 86 88 89 90 95 99 99 105 107 xy FIGURE NO. 10.3 10. 4 10.5 10.6 10.7 10.8 10.9 10.10 10.11 10.12 10.13 10.14 10.15 10.16 10.17 10.18 10.19 10,20 10.21 10.22 10.23 11.1 11.2 XVi TITLE Corrosion of 347 Stainless Steel by 3NaF-UF, Heat-Treated Inconel Specimen Corrosion of Inconel by Pretreated Fluoride Fuels Heat-Treated 316 Stainless Steel Specimen Corrosion of 316 Stainless Steel by Pretreated Fluoride Fuels Corrosion of Inconel by Dehydrated Commercial Sodium Hydroxide Corrosion of 316 Stainless Steel by Dehydrated Commercial Sodium Hydroxide Corrosion of Nickel A by Sodium Hydroxide Corrosion of Inconel by Dehydrated Commercial Potassium Hydroxide Corrosion of 316 Stainless Steel by Dehydrated Commercial Potassium Hydroxide Corrosion of Inconel by Strontium Hydroxide Corrosion of Iron by Strontium Hydroxide Corrosion of 318 Stainless Steel by Strontium Hydroxide Corrosion of Inconel by Pretreated Fluoride Coolant Corrosion of 316 Stainless Steel by Pretreated Fluoride Coolant Corrosion of 316 Stainless Steel by Sodium Metal Crystal Formation in Loop Containing Lead Failure in 316 Stainless Steel Loop Containing Lead Thermal -Convection Loop Operated with Lead Cold Zone of 446 Stainless Steel Loop Showing Attacked Surface and Metal Crystals Which Form in Lead (Dark Areas) Cold-Zone Weld in 1010 Steel Loop Showing Metal Crystal Formation Adjacent to Pipe Wall Covered with Iron Oxide, Presumably Formed During Loop Fabrication Schematic Diagram of Free Convection Apparatus Thermal Conductivity of Copper PAGE NO. 107 108 108 109 109 114 114 115 115 116 116 117 117 120 120 121 127 128 128 130 130 133 136 FIGURE _ : fi : PAGE NO. | - TITLE | NO. 11.3 Schematic Diagram for Determining Heat-Transfer Coefficients 138 11.4 Comparison of ORNL Lithium Heat- Transfar Data with Those of Other Investigators | 140 12.1 Effect of Penetration of Fatigue Life on Inconel Tube-to- Header Sp301mens 143 12.2 Effect of Cold-Working Stainless Steel-~U0, Cores 148 12.3 Effect of Cold-Working Irofi-U02 148 12.4 Effect of Particle Size of U0, 150 12.5 Die for Rubberstatic Pressing 151 13.1 The System NaF-LiF-UF, | | 157 13.2 The System KF-LiF-UF, | | 158 13.3 The System RbF-LiF-UF, 159 13.4 The System NaF-KF-RbF-UF, 160 13.5 Sodium Hydroxide Purification Apparatus | 163 13.6 The System Sr(OH),-Ba(OH), 167 13.7 The System NaOH-LiOH | 168 14.1 Composite Plot of Three Bench (o) and Three In-Pile (o) _ Cantilever Creep Curves _ 172 14.2 Liquid-Fuel Sample Container with Thermocouple Well and : Pressure Fittings 173 15.1 Supercritical Water Reactor Flow Arrangement 180 15.2 Cross-Section of Fuel-Elemeat Assembly | 182 16.1 Schematic Drafiing of Reactor Core | 185 17.1 Sodium Hydroxide Cooled and Moderated Reactor 189 17.2 Temperature Throughout Sodium Hydroxide Core 190 17.3 Sodium Hydroxide Velocity and Film Coefficient Throughout Core 191 XV11 SUMMARY This quarterly progress report of the Aircraft Nuclear Propulsion Program at the Qak Ridge National Laboratory i1s divided into five parts: - RBeactor Design, Shielding Hesearch, " Materials RBesearch, Alternative Systems, and Appendixes, each of which 15 discussed separately in this summary. Part I. REACTOR DESIGN The Aircraft Reactor Experiment {$ec. 1). The design of the high- temperature Aircraft Reactor Experi- ment has been modified to use the 30- 1n. core instead of the 36-in. core, as proposed in the last report. All reactor components are on order with deiivery expected this coming quarter. Design of the external fluid circuit is essentially complete, and many of the major components are on order. Construction of the building facility is on schedule. Experimental Beactor Engineering {Sec. 2)., Recent developments have included a frozen-sodium seal for a centrifugal pump, satisfactory cper- ation of an electromagnetic pump and of a canned-rotor pump with sodium, operation of a liquid-metal valve at 1500°F without self-welding, and oper- ation of a mock-up of the reactor liquid-fuel system. Technigues have been developed further for the handling and purification of sodium and sodium systems to minimize oxygen contami- nation; this bas contributed appreci- ably to the successful eoperation of these systems at high temperatures. Reactor Physics (Sec. 3). The resctor physics calenlations have been dz=voted largely tostatics and kinetics of the ARE, although some analysis of a sodium hydroxide reactor and a full aircraft-size BeO reactor were com- pleted, The uranium investment 1n the ARE 1is 29.2 lb, the slight increase over previous estimates being due largely te increased structural and material poisons in the completed design. As the change in reactivity for normal control is relatively insensitive to the location of the six cuter control rods, they have been located to maintain the most unifeorm flux distribution. Calculated kinetic responses of the ARE to arbitrary changes in reactivity, as a result of possible accidents or failures, show the system to be well damped. Pre- liminary calculations of a sodium hydroxide reactor designed at ORNL have yielded data on leakage and absorption spectra of the core and indicate a critical mass of approxi- mately 32 1b. Critical Experiments (Seec. 4). The preliminary reactor assemblies which are being, or have been, investigated include simulations of the air-water aircraft reactor, proposed by General Electric, and a graphite-uwranium assembly to study power and flux distributions. Measurements of this latter reactor were still being made at the end of the quarter. The two modi fications of the air-water reactor demonstrated the savings in critical mass to be gained by decreasing the thickness of the water lavers, ‘ Part Ii. SHIELDING RESEARCH Bulk Shielding Reactor (Sec. 5). The measurements on the mock-up of the ANP PROJECT QUARTERLY PROGRESS REPORT ideal unit shield have been completed and indicate a weight of 1Z4,000 1b for a 3.8-ft spherical reactor. Con- struction of the mock-up of the divided shield i1s essentially complete, and preliminary shadow-shielding tests have been completed. Lid Tank (Sec. 6). Recent measure- ments on the lead-—borated water unit shield, i1n which the boron concen- tration has been 1ncreased to 1.3 wt %, result in a unit shield weight for a 3-ft spherical core of 101,200 1b, 6000 1b lower than that calculated by the Shielding Board in ANP-53, ago. a year Duct Test (Sec. 7). A practical treatment for liquid-metal ducts 1in a reactor shield has been demonstrated by the use of an extra comical layer of water around the duct as it emerges from the reactor shield. The weight of such 2 patch for 2 single 6-1n. sodium duct surrounded by a l-in. air- filled annulus is approximately 1000 1b. Shielding Talculations (Sec. 8). Analysis of Lid Tank data has yielded "removal cross-sections" for lead, boron, and water of 3.4, 2.0, and 0.8 barns, respectively. The theoretical analvsis of the divided shield 1s again being scrutinized very care- fully. It seems likely that, when some of the approximations are better defined, greater the weight may be somewhat than previously supposed, although not prohibitively s0. A weight advantage of several tons by using ammonla i1n the divided shield in preference to water has been demon- strated. An analysis of some possible radiation hazards associated with the ARE has also been made. Nuclear Measurements (Sec. 8). The 5-Mev Van de Graaff generator is inm operation, and preliminary experiments indicate its useful range of energies to be from below 0.2 Mev up to 6 Mev, An in-pile lithium loop has yielded data on Bremsstrahlung intensities, and measurements of the (n,2n) reaction in beryllium were made, The instal- lation of the neutron time-of-flight spectrometer at the LITR has progressed although fabrication of the assembly has been seriously delayed during the quarter, Part II1I. MATERITALS RESEARCH Investigation of the materials problems of a high-temperature reactor continues to comprise a major part of the effort of the Aircraft Nuclear Propulsion Project. 1In addition to the empirical research on corrosion, radiation damage, cation, materials fabra- and reactor chemistry, this program includes the determination of the basic thermal and physical con- stants associated with these materials at reactor temperatures. Corrosion Research (Sec. 10). Cor- rosion studies have been expanded to include corrosion by hydroxides and possible fluoride coolants as well as corrosion by the liguid metals and fluoride fuels already under 1in- vestligation for some tilme. The extensive corrosion tests with sodium, both statiec and in thermal convection loops, conclusively demonstrate that sodium causes negligible corrosion to elther inconel or a number of stain- less steels at 1500°F. Static cor- rosion tests of the pretreated fluoride fuel mixtures indicate that containing of these liquids 1s likewise possible under the same circumstances. Cor- rosion attack by such pretreated fuels FOR PERIOD ENDING SEPTEMBER 10, averages about 1 mil on stainless steel and 2 mils on inconel after 100 hr at 1500°F. Corrosion attack encountered with hydroxides and with hydroxide- bearing materials, however, 1s a great deal more severe. Both commercial and specially pure sodium hydroxide are extremely coerrosive to inv@nel and to the stainless steels. Corrosion by potassium hyvdroxide 1is 5emewhat less severe. Tests with barium and strontinm hydr@x1fiea indicate these canstics to be somewhat less corrosive than @h,“ of the alkali hydroxides, but sxili 50 severe as to preclude their use at this time. Rigid controel of the purity of the waterial, which was necessary with the fluorides, has not yet been realized but should reduce the present hydroxide cerrosion rates. Physical Properties and Heat-Trams- Fer Research {Sec. 1i1). Of particula interest in the design of the reactor are i%n measuremenis of the physical properties of probable reactor ma- Epridis as well as the studies of heat-transfer phenomena and assoriated coefficients. Heat-capacity measure- ments have been completed for 316 stainless steel, lithium, molybdenum, and zircoeaium in the range 300 to 1006°C, The thermal conductivity and density of the flvorade fuel have also region of interest, measurement of the been defined in the Apparatus for the viscosity of high~temperature liguids and the heat-transfer coefficieats of hydroxides and fused salts is now easentially complete. Investigation of the free-convection mechanism within quiescent liquid-fuel elements 1is underway. ‘ Metallurgy and Ceramics {(Sec. 12). The metallurgical processes involved construction and assembly of a in- in the high-temperature reactor core, 1951 cluding welding of tubular fuel elements, fabrication of s0lid fuel elements, creep of metals, fabrication of control rods, and cladding of beryllium oxide moderator, are cur- rently being investigated. It is now apparent that tube-to-header welds having tensile strength comparable to that of the parent metal may be resdily gffected with znconsel by an electro- magnetic "come-arc” technigue. Added to the list of technigees by which solid-fuel pfiwments may be fabricated s that of "rubberstatic pressing."” in all technigues the uss of 2 screensd fraction of sintered UQ, secems to he desivable. The creep laber atory and stress-corrosion laboratory are now in operatlon and are emphasizing tests on incenel and stainless steel specimens. A ceramic labersatery has besen set up and 13 bheing equipped and staffed. Chemistry of Gigh-Tenperature Liguids (8sc., 133Y. The chemical research on reactor fluids has been extended to clude study of nen- metallic liquids for use as moderators and/er heat-travnsfer fluids, in ad- dition to the development of liguids for use as high-temperatnre fuels. Nine fluoride fuel systems, both ternary and quaternaty, are singled out as covering a useful rangs of uranium concentration ing satis reactor and possess- sfactery melting points. In the development aof hmmfigenfl@ua reactor fuels, solmtinnsfififi) in NaOH-Ne B, O, show premise, in dd&l tien $o the NaOH-1.10H system previously developed. The recent investigations for moder ~ ator coolants have indicated that several bimary hydrogenous systems appear to be satisfactory as far as liquid ranges, moderating ability, and heat-transfer properties are concerned. Preparation of sodium hydroxide of greater than 99, 8% NaOH by weight was ARP PROJECT QUARTERLY PROGRESE REPORT required in the above and in the assocliated corrosion studies. The list of nonmetallic coolants now imn- cludes 11 fluoride systems of usable liguid range and low corrosiveness, although their heat-transfer properties are not sufficiently well known to permit evaluation of their usefulness. Radiation Damage (Sec, 14). Although a number of fuel capsules have been irradiated in both the pile and cyclotron, complete data are not yet available ewing to the residual activaty of the capsule. From preliwinary results 1t appears unlikely that radiation will have a2 significant effect on the fuel. On the other hand, the thermal conductivity of incoenel appears to decrease by a factor of 2 iu less than a week of exposure in the X-10 pile. The flux devendence of this decrease has not been de- termined, but a temperature anneal of the effect has been demonstrated. As regards creep under irradiation, there 1s a definite reduction in primary creep due to airradiation with 347 stainless steel. However, after long periods of strain, 200 hr and irradiated specimens exhibit significantly greater elongations than abeove, control specimens. Twe preliminary experiments on irradiating lithium ia iron capsules showno appreciable added corrosion ascribable to radiation effects. Part IV, ALTEBNATIVE SYSTEMS The major effort of the Aircraft Nuclear Propulsion Project at Qak Ridge Natienal lLaboratory is directed toward a 1500°F liquid-fuel reactor. However, research is 1in progress both here and at associated laboratories on several other reactor systems which show promise. Among these alternative systems recelving consideration are the supercritical water reactor and two configurations for a cirvculating- moderator reactor. Supercritical ¥aisr Reactor (Sec. 15). The design of a supercritical water reactor 1s being amalyzed by Nuclear Develcpment Associates, Inc. Their present concept of the reactor and shield 1s a 2.5-ft square-cylinder active core surrounded by an 11-ft- diameter sphere of water. The water makes two passes through the core which contains fuel elements of the sandwich design previously proposed. Studies of reactivity show that uniform thermal flux i1s simultaneously con- sistent with minimum critical mass, approximately 25 1b. The designed to deliver 400 megawatts wath a maxlmum wall temperature of 1290°F, reactor 1S Circulating-#oderator-Coolant Reactor: HKF (Sec. 18). A circulating- moderator-coolant reactor ewploying circulating sodium hydroxide as both moderator and coolant and a fixed ligquid fuel of a mixture of fluorides has been proposed by The H. K. Ferguson Co. The reactor is designed to deliver 140 megawatts at the desigu point of 0.8 Mach and 35,000 ft in a modified B-52 airplane, with a maximum power at of 230 megawatts. The uranium investment in this reactor is high, 187 l1b. This 1s directly at- tributable to the poor heast-transfer characteristics of sodium hydroxide, which necessitates a large amount of inconel for heat-transfer surface. sea level Circulating—¥Moderater-Coelant Reastor: ORNL (Sec. 17). A design of a circulating-moderator-coolant reactor has also been advanced by the Reactor Design Group at Oak Ridge Nationsal Laboratory. The working fluid FOR of this moderator-coolant reactor is likewise specified as sodium hydroxide. A 2.5-ft spherical core using liquid fluoride fuel 1s expected to deliver 200 megawatts with a maximum wall temperature of 1500°F. An essential feature of this design is the use of eannular fuel elements to attain the high ratio of heat-transfer surface to fuel volume necessary with the hy- droxide coolant. ' High-Temperature Power-Plant Studies {Sec. 18). North American Aviation, Inc., has conducted an investigation of high-temperature (above 1800°F) helium and sodium liguid-vapor power cycles with regard to their application to a Mach 1.5, 45,000-ft-altitude aircraft. They conclude that a helium- cooled reactor cannot achieve the supersonic propulsion of aircraft even 1f reactor temperatures as high as 3300°F and helium pressures in the range 1000 to 2000 psi should be used. On the other hand, a liquid-metal- cooled reactor operating in conjunction with a sodium liquid-vapor compressor Jet system does appear feasible for the supersonic aircraft. PERIOD ENDING SEPTEMBER 10, 1951 Part V. APPENDIXES Analytical Chemistry {(Seec. 18). Chemical analysis is required in almost every phase of the reactor program. Although some of these analyses are routine, the development of many new analytical techniques is required. In all, over 400 samples were submitted for analysis during the last quarter and over 1700 determi- nations were made. The n-butyl bromide me thod, which has been developed for the determination of oxygen in sodium, 1s extremely accurate for oxygen contaminations down to 0.015%. The oxygen content of argon and helivum 1s now determined by a colorimetric method which gives excellent precision below 25 ppm. Methods are also being developed for the determination of metallic corrosion products in fluoride fuels and metallic coolants and for the determination of oxygen in lithium and lead. ' List of Reports Issued (Sec. 23). The reports issued during the past quarter include scme fifty reports on all phases of the ANP program. 1. THE AIRCRAFT REACTOR EXPERIMENT W. M. Breazeale G. A. Cristy L. F. Hemph1ll 5. V. Manson R. W. Schroeder ANP Division The Aircraft Beactor Experiment (ARE) is a 3-megawatt reactor designed to provide first-hand experience with a high-temperature (1500°F) reactor. Recent modifications of the ARE design have been permitted for convenience where this has not affected the operat- ing temperature of the core. In particular, the smaller core, proposed in the last report, (') has been adopted and consideration 1s being given to the use of NeK since this coolant would eliminate the need of preheaters. The external-fluid, electrical, and control circuits have been generally established and detailing is now 1in process. The design of the fluid circuit is sufficiently well defined to permit ordering of heat exchangers, blowers, and associated tubing. The reactor control signal will now be obtained from both inlet and coutlet coolant temperature. The reactor will contain si1x control rods symmetrically dispersed around a central safety rod. Construction of the building facility for the AREis proceeding as scheduled. The excavation 1s complete and most of the concrete foundation has been laid. {I)Nd M. Smith, Jr., “*‘Becommendation on Alternative Loading,’ Aircraft Nuclear Propulsion Project (uarterly Progress Report for Period %ggigg June 10, 1951, ANP-65, p. 55 (Sept. 13, CORE DESIGN The core arrangement has been re- designed to provide for an active lattice 30 in. in diameter and 34 1in. in length, as suggested by the Reactor Physics Group.{!? (The smaller core and thick reflector favor critical mass and power distribution.} The original core-tube and fuel-tube sizes have been retained, providing a maximum fuel capacity of 31 1b of U?3%%, using fused fluorides (NaF—KF-[W;)containing 150 1b of U?3® per cubic foot. The current core—pressure shell assembly provides for seven control thimbles, located in the core, and two instrument chambers, located in the reflector. Material and component orders have been revised to include the changes and features referred to above. Detail and assembly drawings for the entare core-—pressure shell assembly are being prepared. The orders for core components which were placed during the preceding quarter®?) are still outstanding. Delivery of most items is expected within the next twomonths. After extensive negotiations with the Norton Coempany, the Brush Company, AFC, and Norris, it has been deczded that all moederator and reflector beryllium oxide blocks will be hot- pressed by Norton and that the original bilock si1ze {(3.75 flats? in. BCross (2)8, W, Schroeder, "Design of the Aircraft Reactor Experiment’ ANP-65, op. cit., p. 10.° ANP PROJECT QUARTERLY PROGRESS REPORT will be retained. Initial shipments of BeO have been made. FLUID CIRCUIT Detailed layouts of all aspects of the external fluid circuit are currently being made. The availability of the various components from either at-hand stock or commercial suppliers 1s being determined, and in some cases the procurement of these i1tems has been initiated. The use of NaK as the reactor coolant will eliminate the need of preheaters in the coolant system. The helium monitoring circuit, the reactor room and pump-room space cooling circuits, and the control-rod cooling circuits are being reanalyzed in an attempt to obtain dual usage of some of the components. Heat-Disposal Circuit. The heat- disposal systems provide for heat transfer from NaK to helium to ethylene glycol to water. The NaK-to-helium heat exchangers and the helium-to- glycol heat exchangers have been studied in detail, and the configu- rations which appeared to be most attractive have been reviewed with two prospective suppliers, the Griscom- Russell Company and the Vulcan Copper and Supply Company. FEach of these companies confirmed our basic approach and have submitted formal gquotations for the heat exchangers. As each of the prospective suppliers has agreed to quote on the basis of l-in. inconel tubing in the NaK heat exchanger, the tubing has been requisitioned to favor the optimum delivery date. The afore- mentioned companies also confirmed the calculated helium flow rates and pressure drops, enabling the release of requisitions for the helium blowers. Invitations to bid have been extended to six blower manufacturers; the bids 10 ‘released to the stack. have been received and currently are being evaluated. The glycol-to-water heat exchangers have been found inY-12 surplus, and their custody has been transferred to the ARE project. It has been decided to revise the NaK system to provide for upward flow from reactor to an expansion tank, then down from the expansion tank to the heat exchangers and down from the heat exchangersto the low point immedi- ately above the dump tank, and from there to the pump and the reactor. This revision places the pump on the low-temperature side of the heat ex- changer and obviates the need for employing the pump as a surge chamber. It 1s planned to maintain pump level control by means of a forced helium flow through an orifice, the area of which 1s adjusted by the NaK level. It has also been decided to employ a mechanical valve in the dump tank line rather than to maintain the liquid level by gas pressure as previously contemplated. Contamination of Helium System. Failure of a fuel element 1s expected to release xenon and krypton to the NaK circuit and in turn to the helium present at the various NaK free sur- faces. Studies have indicated that the helium so contaminated must be held for several days before being After several possible scrubbing arrangements had been investigated, 1t was decided to make this helium circuit a closed loop with a valved line to the stack, permitting release after any desired waiting time. This system will 1in- clude a low-pressure reservolr, a scrubber, a small compressor, amonitor, aod a high-pressure reservoir. FOR PERIOD FENDING SEPTEMBER 10, 1951 CONTROL OF THE ARE F. 8. Bettis, Research Director’s Division The control of the ABRE will in- corporate solid absorber rods 1in addition to the liquid-fuel control apparatus which decreases the fuel volume in the core. The control signal 1is now taken from both the inlet and outlet coolant temperatures rather than from the outlet tempera- ture only. Detailed design of the absorber control rods has been com- pleted, anda rod and actuator assembly is being fabricated. A high-tempera- ture fission chamber has been designed, and a satisfactory multiplier has been the reactor developed for dynanic computor, Liguid-~Fuel Control System. The elementary diagram of the control system has recelved major emphasis quarter. A completed elementary control diagram is now ready . The control-room layout, in- cluding operating console, relay and during this instrument racks, conduits, has pleted. Detailing of these items 1is now being completed. and interconnecting been partially com- A fundamental change in the control the source of the servo regulating signal was changed. method was made when This signal no longer originates from the cooclant outlet temperature but 1is obtained from both coolant ocutlet and inlet temperatures., The equation for this signal 1s given by E = (T, T - T, + K(T, - T) where error signal actuating the regulating rod T, = coolant inlet temperature T, = temperature of reactor core circult at equilibrium condition at start-up with reactor power output essentially zero K = constant determined by ratio of outlet temperature rise to inlet temperature drop from start-up valuae T T, ™ coolant outlet temperature the the Using this equation a solution of reactor kinetics is 1in process by Nuclear Physics Group. : The test rig, for testing the operatien of bellows for moving high- temperature liquids, has been completed and 1s in the Experimental Engineering building. It has not yet been put into operation, and the tests will not be started until some of the more important loop tests have been com- pleted. The been 80lid Absorber Rod Design. number of absorber rods has established, pending verificatvion by the critical experiments. Present designs show six rods symmetrically disposed about the center with an additional red at the center of the reactor. Detailed designs of the control rods, thimbles, and actuators have been completed. The fairst work order for one cemplete rod and actuator assembly will go into the shop early in the next quarter. i1 ANP PROJECT QUARTERLY PROGRESS REPORT Calculations on the heat generated in the absorber rods show that these cooled satisfactorily with The rod design, which has been completed, provides for the coolant helium to escape i1nto the reactor pit and provides a monitoring atmosphere for the pit. Gas adequate to cool B,C rods are being provided even though the possibility still can be helium. volumes of using hafnium absorbers ex1sts. A design of a fission chamber for use at elevated temperatures has been completed and such a chamber is being fabricated. This chamber will be ready for testing 1n about six weeks. Electronic Computor Gesign. The first step toward the solution of the computor problem has been completed. A satisfactory multiplier for taking the product of two analogue voltages has been constructed. Work will continue on the computor design at the present level of endeavor. ELECTRITAL CIRCUITY Finalization of some of the major fluid-circuit features has permitted an estimate of electric-power reguire- Detailed investigation of following the failure of ments. conditions external power has indicated that the heat capacity of the system 1s very large compared to the post-scram heat generation. Accordingly, 1t 1s un- necessary to provide for continued operation of the main heat-disposal circuit. It 1s planned to provide for operation of the NaK pumps, the monitoring and control rod coolant hlowers, and the space cooling systems, following failure of outside power. 12 On this basis it appears that the maximum battery power requlirement will be less than 10 kw. The a-c—d-c motor generator set appsars to be governed by the design point condition, which requires approximately 50 kwd-c power. Warm-up of the entire system from room temperature to 1500°F, within 24 hr, appears to require 100 kw a-c¢ and 10 kw d-¢. Circuit diagrams are being prepared on this basis. REMOTE-HANDLING EQUIPHENT Conferences have been held with the Timken Bearing Company regarding bear- ing selection and other design features of the remote cutting machinery. On the basis of these discussions final layouts are being made, and requisi- tions have been released for the larger procured components. FExperiments are in progress involving welding, cutting, rescarfing, and rewelding a 2-in.- thick stainless steel plate. These tests will be repeated with a 2-1n.- thick inconel plate when the material currently on order is received. BUILDING FACILITY FOR THE ARE The facility design has been com- pleted by the Austin Company and the work has been awarded to At the present construction the Nicholson Company. time the building excavation 1s com- plete, the concrete pouring 1s near- ing completion, and the building steel and the concrete reinforcing steel are on order. The Atomic Energy Commission has suggested that certain portioms of the facility work that Oak Ridge "FOR PERIOD ENDING SEPTEMBER 10, 1951 National [Laboratory had intended to perform could advantageously be con- tracted to the Nicholson Company in the i1nterest of hastening the final completion date. Accordingly, AFEC has been advised of several services that Nicholson can perform and that other such items probably will become indi- cated when ORNL facility engineering has progressed further. 13 ANP PROJECT QUARTERLY PROGRESS REPORT 2. EXPERIMENTAL REACTOR ENGINEERING H. W. The ANP Experimental Enginesering Group has the responsibility for developing and/or testing materials, methods, and components applicable to the ARE and ANP reactors. Presently under development are pumps, valves, level and pressure indi- preheaters, heat etc., and, in addition, studies are being made on static and dynami¢ corrosion, methods for handling liquid fuels, purification and handling methods for ligquid metals and blanket- ing gases, full-scale test facilities, test Information derived from this program provides fundamental engineering data for component design and provides a basis for establishing adequate operational procedures to demonstrate the reliability of both laboratory and full-scale systems. flowmeters, cators, seals, ex- changers, and operational techniques for equipment. Work has continued on liquid-{fuel systems with the result that a non- enriched sample of fluoride fuel mixture has been filtered and trans- ferred from one fuel container to another for the first time. A frozen- sodium seal had been in operation for 700 hr at the end of the period, seal- ing agalnst 26 psi 1in a system which contained sodium at 1500°F. A pump incorporating this frozen-sodium seal had also been put 1nto operation svccessfully. In other tests, a graphite gas seal showed promise. Satisfactory operation was obtained with a General Electric type G-3 electromagnetic pump during tests to reproduce G. E. performance curves; good agreement was obtained up to 750°F. 14 Savage, ANP Division Stress-rupture and self-welding tests were well underway, with early resnlts indicating that moderately stressed inconel and 316 stainless steel suffered very little, i1f any, lncrease 1n corrosion from sodium at 1500°F. Inconel tends to weld to inconel only slightly in sodium at 1500°F; steel. the same 1s true for stainless However, zirconium shows no tendency to weld to stainless stsel. Methods were developed by which aging and double filtration reduced the oxygen contenl of sodium to less than 200 ppm, and blanketing-gas purifi- cation equipment reduced the oxygen content to below 15 ppm. Specafi- cations were drawn up and engineering drawings are being prepared for full- scale testing of ARE components, both individually and as a unmit, LIQUID-FUEL SYSTEMS E. Wischhusen and D. R. ANP Diavision Ward, Present plans call for the first ARE reactor to employ fuel sealed inside i1ndividual pins and inserted in the reactor. A second rszactor core may contain the fused fluoride fuel in a fuel system consisting of perhaps ninety fuel tube clusters in which tube clusters would be filled and emptied by means of differential gas pressure. Full-scale mock-ups of portions of the system are being made to reveal in- formation on such i1tems as bubble formation, filling and dumping charac- teristics, valving, and operating techniques. - FOR PERIOD ENDING SEPTEMBER 10, 1951 tlock-up No. 1 was a crode-but-guick system constructed froem glass tubes, rubber stoppers, etc. and using colored water to simulate the fuel. It served primarily to indicate valving arrange- ments, gage requirements, and operat- ing techniques which would be needed for working with mock-up No. 2. Mock-up No. 2 was a carefully con- structed all-glass full-size fuel system employing tetrachlorvethylene (sp. gr. 1.6) to simulate the liquid fuel. Work with this system revealed data on bubble formation, filling and empiying rates, flushing methods, etc., all of which were applicable to mock- up No. 3. Mock-up No. 3, which bas been assembled, 1is an all‘meta] system con- structed from 316 stainless steel 3/16-in.-0.d. tubing and is capable of operating at temperatures up to 1500°F. In a preliminary test using tubing of this size the fuel mixture of fluorides has been successfully filtered and transferred through the tubing in the molten state. In conjunction with uranium fuel transfer experiments, it was decided to investigate the containing of helium in equipment fabricated from 316 stainless steel and 1nconel of 0.030 in, wall thickness at tempera- tures of 1500°F or greater. These experiments were conducted on weld- free type 316 stainless steel and 1nconel tubing at 1600°F under 54 psi helium pressure differential and using a Westinghouse mass spectrograph helium leak detector for detecting helium. It was foudd that no helium diffused interstitially through the tubing walls during 150 hr, the length of the test. ‘ PUMP LEVELOPMENT Pomps currently under development for the ARE and experimental loops include a variety of electromagnetic and centrifugal pumps as well as the unique canned-rotor pump and Mlevel- tank” pump. Performance curves of two electromagnetic punps have been obtained up to 1000°F. At this temperatnre a high rsontact resistance created an open The centrifogal pumps currently development are distinguished by the type of shafe seal. The centrifugal pump discussed below ‘embodied a gas seal; other promising pump seals are discussed under "Seal Tests."” The canned rotor pump has operated for 90 hr with NaK at temperatures up to 400°F with no discernable wear. Awong the pump developments are a Duriron pump which operates immersed 1n the to-be-pumped fluid and the two-stage electro- magnetic pump. Other pump information 1s mentioned under "Loop Tests™ and "Seal Tests.™ circuilt. under A loop for testing either electro- magnetic or mechanical pumps of ARE size has been designed which contains approximately 40 ft of 2J-in. pipe and utilizes 7.5-kw tube furnaces for heating the circulating medium up to 1500°F. The loop volume, including a 3-ft3 sump tank, 1is approximately 5 ft®. Space is provided for electro- magnetic or venturi type flowmeter in- strumentation. Centrifugal Pumps for Figure-Eight Loops (W. G. Cobb, ANP Division). The centrifugal pump for liguid metals described in the previous reportf?) (1)w. G. Cobb, “Pumps,” Aircraft Nuclear Propulsion Project Quarterly Progress Beport for Period Ending June 10, 1951, ANP-635, p. 167, esp. p. 168 (Sept. 13, 1951). ANP PROJECT QUARTERLY PROGRESS REPORT was 1nstalled in a figure-eight loop. Teflon packing rings were used for shaft sealing, and level indication and control were taken from the liguid surface in a tank mounted at the side of the pump housing. Flow observed during operation was 7.5 gpm at ap- proximately 45 ft of sodium, but the difficulties of simultanecusly con- trolling the three liquid levels 1in the system under dynamic conditions inhibited smooth operation. The initial run lasted for 3% hr at a sodium temperature of 795°F before failure of the shaft seal; operation the final % hr. A second run lasted 16 hr, reaching a temperature of 900°F, but however, was smooth for terminated when the pump became starved and hence was unable to pump smoothly. Examination of both these runs after shutdown i1ndicated a failure of the ligquid-level control system. In the first run the pressure equalizer line between the pumpand level control tank became plugged, allowing sodium to rise through the seal. This satu- ation was corrected by additional heaters and insulation on both the ligquid and gas lines between the pump and level control tank. The explanation for the intermittent pumping 1n the second test was failuve of the level indication and control equlpment to maintailn a satisfactory liquid level in the pump. Subsequent checks did reveal considerable variation between the two liquid levels during operation. Difficulties encountered 1n maintaining constant liquid level in the surge tank and in the pump indicated that this method of control was impractical. ARE Pump Design (W. G. Cobb and J. F. Haines, ANP Division). A con- tract has been let with Allis-Chalmers 16 Company for design and fabrication of a sump type internal-bearing direct- driven centrifugal pump for testing and possible use with the ARE. This and other pumps of ARE capacity being designed by ANP Experimental Engincer- ing personnel, are being guided by the following considerations: 1. The pump is preferred to be located on the cold side of the reactor to minimize high-temper- ature effects on pump structure. 2. Separate pump sump and surge tanks are to be provided. 3. An 1individual level control is to be provided for the pump. Tests are being performed to determine the feasibility of controlling pump level from a dynamic surface. 4. A gas seal for the pump shaft 1s to be used. 5. Flow direction through the pump 1s to be comnventional with suction through the lower side and discharge from a concentric casing. 6. The to be carried by an overhung vertical shaft with no internal bearings. impeller 1is 7. Pump cooling 1is to be ac- complished by flowing helium, which 1s to be monmitored for coolant leakage. 8. Bearing lubricant 1s to be circulated continuously; this feature provides cooling for a rotating face seal. 9. The controlled liquid surface FOR PERIOD ENDING SEPTEMBER 10, 1951 in. the pump is to be as small as compatible with stability. 10, The initial pump for testing is to use a cast impeller with casing and other parts to be fabricated and machined. Canned-Rofor Pump (A. R. Frithsen and M. Richardson, Reacter Experimental Fngineering Division). The 1%-hp fluid-bearing pump (Fig. 2.1}, de- scribed in the last report, (?) has been operated with water and NaK without detectable wear. The first test was made in a water~circulating system and continued intermittently for a total of 1200 hr running time. Examination of the pump parts after dissassembly showed no visible or measurable wear or corrosion. All clearances were the same as before the test was begun. : In a subseguent test the diametrical clearance between the motor rotors and motor cans was reduced from 0.017 to 0.012 in., and the pump was then used to circulate NaK for a total running time of 90 hr. During most of this period the temperature was held at 200 to 300°F, although short runs were made at room temperature and also as high as 400°F. No indication of wear or corrosion was found on any of the parts upon dissassembly, Lut a considerable accumulation of alkali oxides was found at the tabs of the motor cans and close to the walls of the pump volute where the liguid 1is believed to have been relatively stagnant. A Fischer-Porter Flowrator was used for flow measurements and operated with complete satisfaction even at the higher (400°C) tempera -~ tures. : (2)/-\. R. Frithsen and M. Richardson, “Canned Rotor Pump,”” ANP-65, ep. cit., p. 170. The limitation on operating temper- ature for this pump 1s believed to be determined by thermal damage to the insulation in the motor windings. To increase the permissible operating temperature, an order has been placed with the General Electric Company for winding two motor stators with class H wire which 1s capable of withstanding a temperature of 500°F. 1In addition, a proposal has been requested from Allis-Chalmers Company for special wire which can be used at TO0°F. With the latter winding and by cooling the liquid fed into the motor cans, it is expected that liquid metals or salts can be pumped at 1000°F or higher. A 3-hp fluid-bearing pump has been designed and has been partially fabra- cated to test a number of revisions in the design of the smaller pump and to determine the practicability of seal- ing up the smaller model. This pump is expected to be ready for check ‘runs with water about October 1. | Level Tank Pump (W. B. McDonald, ANP Division). A Duriron conventional pump has been modified for operating while immersed in liquid metals. This pump has a capacity of approximately 12 gpm at a 40-ft developed head. A conventional packing gland is used to seal the shaft, and a labyrinth fitting closely about the shaft minimizes the amount of liquid metal by-passed to the level tank. This pump will operate while tmmersed in liquid metal in a level tank and will be driven by a motor on an overhung shaft mounted on top of the tank. ‘Actual testing awalts the availability of a test loop. Electromagnetic Pumps (J. H. Wyld and A. L. Southern; ANP Division; A. G. Grindell, Engineering and Maintenance Division). A General 17 UNCINS¥FIED ANP PROJECT QUARTERLY PROGRESS REPORT s FLOWRATOR Fig. 2.1 - Fluid-Bearing (Canned-Rotor) Pump. " UNEEFIED FOR PERIOD ENDING SEPTEMBER 10, Electric a-c electromagnetic pump was modified by Experimental Engineering personnel to utilize a flexible second- ary conductor. Preliminary tests gave flows of approximately 16 gpm of sodium at 5 psi and 890°F. In a combination performance and endurance test, a G. E. type G-3 electromagnetic pump was operated in the calibration loop (see below) for 173 hr. Performance curves were obtained at temperatures of 300, 500, and 750°F for comparison with data supplied by G. E. The pump HEAD (psi) : L o o n 50 voiis @ | I 0 0 2 4 6 8 —_— Its @~_-:itf4LT1r~o-.__“. oo 4 t— ““‘--— -0 T ——n [———=H 1951 failed at 1000°F owing to a current lug melting loose from the cell wall. This failure occurred at 250 volts input but after performance curves had been obtained for that temperature. General Electric supplied data for their type G-1 pump whose performance closely resembles that of the type G-3 pump procured by ANP. A graph compar- ing operation of the G-1 and G-3 pumps 1s shown in Fig. 2.2. The efficiency UNCLASSIFIED DWG, 12876 ® ANP DATA, PUMP 3216680, G-3 B G-E DATA, PUMP 9159849, G- 1 0 12 14 16 i8 20 FLOW (gpm) Fig. 2.2 - (750°F, Sodium). Comparison of General Electric and ANP Electromagnetic Pump Data N I L0 ANP PROJECT QUARTERLY PROGRESS REPORT of these electromagnetic pumps in- creases essentially linearly from 0.5% at flows of 2 gpm to about 3% at flows of 14 gpm. Although performance curves were taken at lower temperatures, the larger part of the operation was con- ducted at 750°F, which was the maximum temperature recommended for the pump cell in use. Operating temperature was increased to 1000°F to obtain performance curves and to determine the upper limit of temperature allow- able with the pump cell supplied. Engineering drawings for the two- stage electromagnetic pump were com- pleted and submitted to the shop. Fabrication is incomplete. TEST LOOPS A. G. Grindell, Engineering and Maintenance Division The "calibration" loop and the "sodium manometer' loop, both of which were designed to calibrate and other- wise check flowmeters, have been operated with liquid metals during the past quarter. To date the more sig- nificant application of these loops has been in conjunction with pump tests. Calibration Loop. The calibration loop, which was designed primarily to check electromagnetic flowmeters against venturi types and to check the performance of electromagnetic pumps, has been equipped with heaters, thermo- couples, pressure measuring devices, a lock valve for operating on a con- stant weight of sodium, and a static cold leg for continually removing oxygen from the circulating sodium. 20 The system was degreased with per- chloroethylene, evacuated, outgassed at 800°F, and flushed twice with sodium at 800°F. 1In this condition it has operated satisfactorily as a test loop for the G. E. electromagnetic pump described above. Sodium Manometer Loop. The sodium manometer loop is an experimental loop designed to afford means of developing testing, and gaining experience on flow-nozzle manometer type flowmeters and simultaneously developing and testing electromagnetic flowmeters. The first attempt to pump sodium dur- ing the period resulted in electro- magnetic pump cell failure and sodium fire during the start-up. The loop was dismantled, repaired, and restarted. In this test the electromagnetic pump was provided with a flexible secondary conductor, and during the first 2 hr of operation at constant power input, flow increased from 2.5 to 9 gpm. This increase in pumping rate was attributed to 1in- creased wetting of the pump cell wall, and increased pump 1nput power over- heated the flexible secondary con- ductor. Pump power was decreased, but loop temperature was taken up to 890°F. The pump failed after27 hr of operation owing to the inadequate current capacity of the flexible secondary conductor; another conductor 2% times the cross- sectional area of the first 1is now being fabricated. SEAL TESTS Development of seals for pumps handling liquid metals or other coolant at 1500°F has received attention throughout the period. Tests have FOR PERIOD ENDING SEPTEMBER 10, 1951 been conducted simultaneously with pump experiments in some instances, while 1n others tests have been con- ducted in simulated equipment. The seal types i1nclude a frozen-sodium seal, metal-to-metal seals, and metal- to-metal seals at the end of a bellows. Frozen~Sodium Seal (W. B. McDonald, ANP Division). A seal employing the principle of solidifying a column of sodium around a rotating shaft is under development. The testing device consisted of a pot containing sodium at 1500°F. A sleeve extended from below the sodium level i1n the pot to the outside, and a shaft rotated inside this sleeve. Sodium was forced by gas pressure into the annular space between the rotating shaft and the sleeve. A refrigeration coil, to cool the sodium to below its freezing point, was wound around the outside of the sleeve and soldered to 1t to increase thermal conductivity. The rotating shaft was 1% in. in diameter with a diametrical clearance of 0.032 shaft and sleeve. By the end of the period this seal had operated for approximately 400 hr against a pressure of 18 psi1. The coolant employed was kerosene refrigerated by passing through a coil immersed in a dry ice and ethylene glycol bath. in. between the At the end of the period a second test using sodium at 1500°F but with the pressure differential increased to 26 psi was underway. Also, a con- ventional centrifugal pump was modified by adding a frozen-sodium seal and was installed in a pump test loop (Fig. 2.3). The relatively short space between the pump and the bearings 1is expected to limit the temperature to which the circulating sodium can be raised; however, this pump was de- livering 15 gpm of sodium at 650°F at 2000 rpm at the end of the period. Bellows Seal (A. P. Fraas and M. E. LaVerne, ANP Division). A Dureg stain- less steel centrifugal pump(3): was reworked to provide radial holes in the rear face of the impeller to keep the seal cavity dry and was operated for 80 hr with sodium at temperatures up to 1000°F. Tests with water showed that the design is effective in scavenging liguid that leaks into the seal cavity through the labyrinth seal on the rear face of the impeller. The test with sodium was discontinued when finely divided sodium was found in the silicone 01l circulated over the outer side of the bellows seal. Since only gas should have been i1n contact with the inner side of the seals, and since the liquid level contrel had given difficulty with sticking of a solenoid valve, the rig was disassembled for inspection., All parts were found in good condition except the bellows which found to have two axial cracks on the periphery of two con- volutions. The test will be repeated with another bellows. was Graphitar and Tool Steel Seal (W. B. McDonald, ANP Division). A seal consisting of a Graphitar (U. S. Graphite) ring sandwiched between a stationary member and a rotating member made of Ketos tool steel sealed suc- cessfully against a gas pressure of 10 psi at a seal temperature of 350°F for 450 br. Shaft speed was approximately 2000 rpm. This test was termimated at 450 hr in order to make changes in the test equipment to enable NaK vapors to be i1ntroduced to the sealing members. (3)A. P. Fraas, Progress Report on Stainless Steel Acid Pump Reworked To Test Special Features for Operation with Liguid Metals, ORNL, Y-12 site, report Y-F15-6 (Feb. 1, 1951). 21 UNCE®FIED ANP PROJECT QUARTERLY PROGRESS REPORT Fig. 2.3 - Centrifugal Pump with Frozen-Sodium Seal. ” UNCPASSBIIED FOR PERIOD ENDING SEPTEMBER 10, 1951 These tests were underway at the end of the period. INSTRUMENTATION J. F. Bailey, Consultant to ANP Division The development of suitable instru- ments for indicating flows, pressures, levels, ete. 1s an 1mportant con- sideration for the adequate testing of liquid-metal systems and for the successful operation of the ARE. Hecently considerable thought has been given to these instrumentation problems, with the result that a study has been made of all types of level control devices to determine the advantages and the disadvantages of each. The types studied include electroresistance, electromagnetic, electroreluctance, gas flow, and float, and recommended applications for each have been com- piled. This work has been done with the view to designing a vigorous re- search program to develop adequate instruments for application by both the Experimental Engineering and the ABE groups. spark plug type probes are being used to control levels. This method is simple but unreliable because of periodic short- ing out of the probes. FElectro- resistance probes for: level indication are to be investigated 1n an experi- mental system which is presently under- going fabrication. FExperimental work 1s to determine the merits of this type of probe as a possible replace- ment for the spark plug type. At the present, A model of the gas-flow control has been set up in the Liquid Models Test Laboratory, and the operating charac- teristics are being studied with water as the liquid. Experimental work on other types is in the planning stage. During the gquarter two magnets for use with électromagnetic flowmeters were made by using two U-shaped Alnico V pieces and cold-rolled steel spacers. These magnets givea flux of 1500 gauss across a 1l4%-in. gap. Using L-in. stainiess steel tubiang with these magnets, a flow of 20 gpm pgives an induced voltage of 10.6 mv. The low-flux magnets have advantages in that their use permits existing equipment to be used for vecording high flows, and braking effect in the tlow circuit 1s reduced. When {flows are below 10 gpm, the magnet strength should be above 2500 gauss to allow the same recording egquipment to be used with higher accuracy. STRESS-RUPTURE TESTS J. L. Gregg, Consultant to: ANP Division Stress-rupture tests a¥e being con- ducted by the Experimental Engineering Group to determine the strength of materials at elevated temperatures. For testing, sections of a tube of type 316 stainless steel and one of inconel were machined to 0.015 in. wall thickness, pressurized to 110 psi with argon, and immersed in sodium at 1500°F. The gas pressure produced a hoop stress of approximately 1500 psi in the thinned section of the tube. Ry the end of the quarter, these tubes had been under test for approximately 500 hr without failure, indicating that attack of sodium on moderately 23 ANP PROJECT QUARTERLY PRCOGRESS REPORT stressed inconel and stainless steel 1is rather small. Equipment is being assembled for testing sheet metal 1in sodium at elevated temperatures. SELF-WELDING TESTS J. L. Gregg, Consultant to ANP Division A. P. Fraas and M. E. LaVerne, ANP Division Self-welding tests are being con- ducted in sodium at 1500°F to provide information pertinent to seals, valves, hearings, etc. for the ARE. Preliminary results show that inconel has only a slight tendency to weld to inconel under these conditions. true for stainless steel. shown no tendency to weld to zirconium in any experiment conducted thus far. The same 1s Inconel has Consideration of their atomic structure had indicated that zirconium should be as unlikely to weld to 1iron- chrome-nickel alloys as any metal availlable. In order to test this supposition, a 304 stainless steel standard globe valve was reworked to provide a zirconium washer floating on the valve stem hetween the seat and the valve disk of 347 stainless steel. The reworked valve was operated 1in sodium for 142 hr, 48 at 1200°F and and 94 at 1500°F. Operation con- sisted 1in a number of closings and openings, the required opening torque being taken as a measure tendency toward welding. of any No indication of welding was obtained during the test, and visual examination after completion of the test disclosed no signs of welding between the zir- conium washer and either the seat or disk. A second test 1s being made 24 with three valves in parallel on the same rig, one valve having a zir- conium washer, one a molybdenum washer, and one just the standard stainless steel valve disk. VALVE-PACKING EXPERIMENTS W. C. Tunnell, ANP Division Since the ABRE system as designed requires valves to operate at 1500°F, a program for determining valve design characteristics 1s underway. If con- ventional valves should prove un- satisfactory, 1t 1s likely that a bellows type valve would be a suitable alternative. Such a valve would re- quire packing to provide safety features and also to permit temporary use of the valve 1n case of bellows failure. Experiments to determine satisfactory packing materials for use with sodium are in progress. Following isa list of the materials tested and the pressures and tempera- tures reached during preliminary in- vestigatlions: 1. Amosite asbhestos: held sodium for 21 hr at 1500°F under 15 psi before leaking. 2. Graphite powder: leaked sodium at 600°F under 10 psi pressure. 3. Met-1.-X: held sodium for 2 hr at 1500°F under 15 psi leaking. before 4. Lead-mill slag: held sodium for 10 days at 1500°F under 15 psi pressure; no leaks had occurred before experiment was terminated. 5. Nickel metal powder: held sodium for 10 days at 1500°F under 15 FOR PERIOD ENDING SEPTEMBER 10, 1951 ps1 pressure; no leaks were observed before experiment was terminated, 6. Soda ash: leaked sodium at 1300°F under 10 psi pressure. Since this experiment was considered a preliminary screening operation for selecting possible packing materials, no effort was made to operate the valve stem when the sodium reached 1500°F. Other egquipment is being designed with smaller clearances be- tween the sealing parts of the pack- ing glands to afford more conclusive tests of powdered or granular materials. CONTAMINATION OF LIQUID-METAL SYSTEMS The corrosion of metal containers by liquid metals, in this case primarily is attributed to the oxygen Nominal sodium, contamination in the system. precautions to exclude oxygen from ligquid-metal systems, including the use of 1nert-gas blankets and cleaned systems, had been taken, but they did not succeed 1in limiting corrosion to the desired level. It became apparent that even the small remaining oxygen contamination was a fault, and conse- quently provisions for its reduction have been developed. The three main sources of oxygen, barring leaks, are from scale in the mechanical system, from the sodium charged to the system, and from the inert-gas atmosphere used with the system. Oxygen contamination from cach of these sources is minimized by technigques now in use. | Cieaning of Liguid-ymetal Systems and H. R. Bronstein, ANP Ligoid-metal systems re- (K. Devenish Division). quire careful cleaning prior to being put into operation to remove welding scale, oxides, absorbed gases, etc., to reduce contamination of circulating coolant. These systems are now being degreased with hot perchloroethylene, after which they are evacuated and outgassed at approximately the temper- ature of operation. An experiment 1is underway 1in the calibration loop to test the efficiency of preconditioning a systemwith sodium prior to operation. In this experiment, in addition to degreasing and evacuating, sodium was circulated at 800°F and removed, and clean sodium was introduced. This second batch of sodium was circulated for 3 to 4 hr at 800°F also and then dumped. Clean sodium was provided for operational use. By the end of the period, the loop had operated for approximately 175 hr with no apparent difficulty due to faulty cleaning procedures. Equipment has been installed for descaling stainless steel alloys by the "Virgo Salts" process promoted by Hooker Electrochemical Company. 1In this process the metal 1s immersed in a molten caustic, which converts scale to salts soluble in hot dilute hydro- chloric acid. After these salts are removed by hydrochloric acid, parts are brightened and passivated by a dip in nitric acid. No results show- ing the efficiency of this process were available by the end of the quarter. Removal of sodium from systems after operation is a problem also receiving attention during the period. Sodium 1s now removed from small tube passages by immersing the entire assembly in a water bath which is heated up to 100°C by live steam. When the residual sodium hés melted, 31t i1s forced from the small passages by gas pressure. 25 l'lllli'ii'lh[ ANP PROJECT QUARTERLY PROGRESS REPORT Of course, some burning and small explosions occur which require that the operation be conducted in the open, but no danage to equipment has been experlienced. Sump tanks are now emptied by heating to above the melt- ing point of sodium and flowing it either by gravity or gas pressure into an ash can immersed im a heated o1l bath which keeps the sodium from pro- jecting above the surface, thereby pre- venting a combination oi1l-and-sodium fire, Purification of Liguid Metals (.. A. Mann, ANP Division). Methods of purifying liguid metals to reduce the oxide content to acceptable levels for use in convection loops, figure- el1ght loops, pump test loops, etc. have been 1mproved. In the method now 1n use scdium or NaK 1s held at approximately 250°F for at least 24 hr to allow suspended oxide particles time to agglomerate. The metals are then filtered through a S5-p sintered stainless steel filter into another fill tank. From this tank the metal is again filtered through a 5-u filter into the operating system., Analytical results indicate that double filtration consistently reduces the oxide content to below 200 ppm. The results also indicate that no appreciable reduction 1s achieved with further filtration. Purification of Inert Gases (L. A. Mann, ANP Division). 1Inert gases used for blanketing ligquid metals in operat- ing systems require purification to minimize oxXygen content. Argon gener- ally contains on the order of 20 to 50 ppm oxygen, and helium cylinders have shown as high as 65 ppm (chemical analyses are reported in Sec., 19). A purification apparatus has been put into operation which removes oxygen from blanketing gases by bubbling 26 through a column of NaK. Tests thus far indicate that use of this system consistently reduces the oxygen con- tent of gases used in operating systems to below 15 ppm, which is considered a safe level, As an added precaution, helium cylinders have been set aside for the exclusive use of the ANP Experimental Engineering Group, and these cylinders are never permitted to go below 100 psi to prevent air from being introduced. These methods, which are currently 1in use, satisfactory. are considered Analytical Results with Sodium (H. R. Pronstein, ANP Division). Sam- pling techniques instituted with existing equipment during the period removed lnconsistencies 1n analytical results, making possible the standardi- zation of limits of detection of impurities in sodium. analytical results Evaluation of (see Sec. 19) on sodium samples taken at the time of fill and showed that filtered sodium generally contained approximately 0.02% oxygen, but the oxide content upon termination appeared to be a function of the size of the operating equipment, fill to termination, termination From the time of the oxygen content of sodium 1n thermal-convection loops rose from 0.02 te 0.03%, responding results from figure-eight loops were 0.02 and 0.06%, respectively. These findings indicated the need for more thorough cleaning and outgassing while cor- procedures; consequently the technique of repeated flushing of a system with hot sodium prior instituted, to operation was Design of a sampling device to take repeated samples of a coolant directly from the operating system at any temperature up to 1500°F 1is virtually completed. Two small loops for testing FOR PERIOD ENDING SEPTEMBER 10, 1951 the reliability of this sampling device prior to its installation with an operating system are under construction. INVESTIGATION OF SODIUM CONDENSATION H. R. Bronstein, ANP Division Condensation of sodium in gas- pressurizing lines leading to operat- ing systems was investigated further during the quarter. Electrostatic precipitation showed that plugging was due to a solid aerosol phenomenon, and that quantities of sodium removed by the precipitator were of such a magni- tude that a means of returning this condensate to the main tank should be sought. Consequently, the precipitator was redesigned to allow for periodic or continuous melting of the sodium collected. The experiment was rerun on a dynamic gas flow basis with traps placed after the precipitator to col- lect any sodium getting through it. No sodium was detected beyond the electrostatic precipitator. NaK traps have been suggested as a possible solution for the problem of gas-line plugging. Models of several types have been built, and their operating characteristics are being studied with air and water. From information obtained from these tests, a trap for use with operating systems may be designed. ' A group fromthe MIT Practice School investigated three methods of reducing or preventing sodium plugging of blanketing-gas lines during the period. These methods were (1) electrostatac precipitation, {(2) a sodium condenser coil immersed in a het oil bath, and {3) a NaK bubbler. Although all methods tested proved to be partially effective, results of work by this group indicated that the oil-bath trap was the most practical. Design improvements would increase the efficiencyofall methods, however. The oil-bath condenser causes particles to condense on the walls of the tubing because of its long length and curvature, and, upon condensation, the sodium flows back into the main tank by gravity since the oil tempera- ture is maintained just above the melting point of sodium. Additional experiments to determine the effect of variations in tubing length and size and bath temperature were recommended. ARE COMPONENT TESTS H. P. Kackenmester and D. L. Salmon, ANP Division Specifications have been written by the ARE Components Testing Committee and engineering drawings are being prepared for construction of facilities for full-scale ARE components testing. Five all-steel completely enclosed test cells are to be erected, four of which are 1dentical 1n design and are to be used for measuring physical and operating characteristics of individual components such as pumps, pipes, valves, instruments, heat ex- changers, etc. prior to their use with the ARE. The other test area, larger in size, is to be used for testing components 1in combination and will include facilities for testing one complete heat-exchanger system under simulated operating conditions. A gas-fired furnace delivering 3,000,000 27 ANP PROJECT QUARTERLY PROGRESS REPORT Btu/hr is to be used to simulate the reactor., Design of pilping systems and test loops for individual equipment items 1s progressing satisfactorily, and procedures for components testing are being established. An apparatus has been set up for determining the pressure drop of different orifice configurations 1n connection with coolant-tube orifice investigations. A detail drawing has been completed for lucite headers to adapt an ARE fuel-pin cluster assembly for a water flow test, and detail drawings are heing prepared of a full- si1ze reactor flow passage mock-upto be used for measuring the pressure drop and flow distribution with water. HEAT-EXCHANGER TESTS A. P. ANP Division Fraas, The heat-exchanger model designed for testing in the figure-eight loop has been tested with NaK at 800, 1000, and 1200°F, and a curve of heat- transfer coefficient as a function of the flow rate has been obtained. The experimental data give values as much as 40% below those calculated from Lyon's formula ¢(*) at low flow rates, but check closely at high flow rates. As soon as endurance testing 1s com- pleted, a report will be 1issued on the the flow and heat-transfer tests which have been carried out on heat-exchanger and fuel-element models. (4)R. N. Lyon, Forced Convection fleat Transfer Theory and Experiments with Liquid Metals, ORNL- 361, p. 21, Eq. 34 (Aug.19, 1949). 28 BUILDING #ODIFICATIONS AND EXPERIMENTAL FACILITIES P. L. Hill, USAF Building modifications and the pro- curement of experimental eauipment and facilities 18 progressing on a schedule compatible with that of the ARE. The 120-ft hood has been completed and 1is in use for testing intermediate-silze liquid-metal systems. To allow maximum flexibility, no fixed partitions are provided inside the hood, bwt pro- visions have been made for placing barriers around equipment undergoing test. The exhaust system 1s designed to give a linear facial velocity of 100 to 150 ft/min. Four power-control panels similar to those 1n for thermal-convection-loop operation have been installed i1n front of the hood and are ready for use. Each panel provides up to 24 kw of controlled heater power for use on systems under- going tests. An overall view of this installation 1s shown in Fig. 2.4, use Areas for ligquid and air test equipment have been completed, and these facailities will permit the use and/or air in testing the flowmeters, and of water performance of pumps, heat exchangers 1in so far as these tests may be conducted using air and water as working fluids. Design for a one-man standard chemical laboratory and a special laboratory for testing high-temperature liquid metals has been completed and equipment has been ordered. The cleaning arca for de- greasing and/or removing trace residues of ligquid metals 1s designed and partially completed. Bulk residues of liquid metals or other hazardous ma- terials are being removed in a facility outside the building. This facility I HNBEWION 1 Fig. 2.4 - Experimental Facilities. 120 ft. is across the top. The pump test is PHOTO 6 339 SoDIUM o ;,‘,)‘M FIRES " ALT A bl Y U S0ODIM 1y i “ ' THje w xa & e “li i in the left foreground; the hood, ‘0T HAAWALdAS ONIANA dOoIydd ¥O0d4 IS61 AIGRRARION ANP PROJECT QUARTERLY PROGRESS REPORT is incomplete but was in partial operation at the end of the period. ALKALI METALS MANUAL R. Devenish, ANP Division P. L. Hill, USAF During the quarter the Y-12 Alkali Metals Guide, which sets forth safety 30 considerations to be observed when working with alkali metals or liquid- metal systems, was revised. The revised guide 1s being prepared for general distribution. In this con- nection, colored movies have been made of sodium fire extinguishing operations and liquid-metals disposal. The film has been edited, and a dupli- cate is being made for showing to interested groups. FOR PERYOD ENDING SEPTEMBER 10, 1951 3. REACTOR PHYSICS N. M. Smith, Jr., The principal subjects of activity have been the calculation of the static characteristics of the design referred to here as the "ARE of 10 June 1951" and extensive calculation of the kinetic behavior of the ARE by machine tcchniqfies. The design of 10 June 1951 refers to an ARE of two- thirds the core volume of that in the design of 10 March 1951, Besides reduction of core size, manufacturing specifications and tolerances required a reduction of the average demsity of moderator in both core and reflector and an increase in structure and coolant volume, The effect of the densi1ty changes overrides the effect of reduction of core size, resulting in a net 1ncrease 1n uranium inventory requirement. Minor changes 1in the static coefficients which resulted are summarized form, in tabular and graphical The uranium investment required is estimated to be 27.8 lb +10%, -20%, for the depleted, hot, poirsoned, instrumented reactor with control rbd thimbles in place. Allowance for expansion of liguid fuel through the boron curtain adds a requirement of 1.4 1b to the above figure, yielding a total inventory requirement of 29.2 b of uranium. : The effectiveness of mutual shadinhg effects of a hexagonal array of seven 2-in, BR,C control rods has been calcu- lated using three-group procedure made equivalent to the 3Z2-group pro- cedure for a homogeneous reflected reactor. With one rod in the center, and six on a ring surrounding the Physics Division center,. the total reactivity ef- fectiveness is insensitive to changes in the radius of the ring between 6.5 and 10.5 in. Absorption in the perman- ently placed thimbles will reduce the differential effect of the seven rods to 0.22 1n Ak/k. Muatual shading to the extent of a reduction of ef- fectiveness of seven isolated control rods by 14% 1s indicated by the calcu- lations, The kinetic response of the ARE to arbitrary changes of multiplication constant and to changes in inlet coolant temperature for various values of neutron lifetime and of fuel temper- ature coefficient has been studied and graphs are presented. The results show that the fuel temperature responds in an overdamped fashion to reactivity changes well over prompt critical. A start-up accident throwing the reactor prompt critical at a low (300-watt) level was studied, re- sulting in a tolerable transient, These calculations have been made by starting with steady-state initial conditions and by numerical integration of the set of nonlinear differential equations by IBM machine procedure. Work bas continued on the development of other IBM calculations: multiregion problems, reactor solutions in cy- lindrical coordinates, and the effect of a boron layer between core and reflector. | : Variocus problems relating to the physics of reactor calculations or to the physics of the ARE arediscussed. These include a proper formulation for 31 ANP PROJECT QUARTERLY PROGRESS REPORT the age theory of the relation between neutron flux and slowing-down density, the transmission coefficient of a boron curtain, and heating in the boron curtain, Calculations have been started on the physics of hydrogeneous reactors, and in particular for a NaOH moderated and cooled reactor. Results are quoted. IBM CALCULATIONS Production Report (F, C. Uffelman and P. Johnson, Uranium Control Department). Reactors. During the period ending September 1, 1951 the IBM section completed calculations on 84 reactors. Included in these 84 reactors were four bare reactors calculated by the end-point linear approximation method and three adjoint reactors. This brings the total number of reactors calculated by the EPLA method since calculations were begun in February to 182 reactors, Programing for the calculation of reactors with an absorbing layer be- tween core and reflector 1s very nearly complete. PBecause the method{(!) used consists in making two parallel calcu- lations for flux distributions in the reactor and combining the results of these calculations for each group be- fore going on to the next group, the calculation of a reactor blanketed with B,C will take between five and six times as long as the calculation of a regular reactor. (1. B. Mills, IBW Procudure for B,C Layer Between Core and Reflector by the Coveyou Method, ORNL, Y-12 site, report Y-F10-61 (June 28, 1951). 39 Cores and Reflectors. During the period ending September 1, 1951 average cross-sections and constants were calculated for 26 cores and reflectors, and "constant'" variations were calcu- lated for 34 cores and reflectors. Of the cores calculated, 16 were hy- drogenous and required calculation of the following factors in addition to those calculated for nonhydrogenous cores: TsH_ (1) fUT ’ . Tr BN = (2) ’ Ry + 0.71 (T/S,p)¥ N 82 ———| (3) 3507 rp B 1 N z .fN = . - (4) TR tan'l B/ZTR l/zTR Z;;T? is calculated twice, once for a constant B and once for the variable B derived in formula (2) above,. A new set of cross-section boards and procedures 1s being set up for calculation of special cross-sections and constants, The new set-up will be more flexible but slower in operation than the specialized set-up used for regular nonhydrogenous cores, so that the old set-up will continue to be used for regular cores and reflectors. FOR Kinetic for 16 different sets of The length number of Kinetic Calculations, calculations conditions were completed. of the time interval, the functions, and the initial starting conditions were varied from set to set. Boards and procedures are now being set up for a kinetic calcu- and both spatial variations are made, lation in which temporal The Multiregion Hezctor Pruhiem?Z} (C. B, Mills, Physics Divistion). A two-group calculation has shown that the effect of material cutside the reactor reflector is i1mportant for the fast neutrons leaking out of the reactor. A formula was developed(?’ for use with the IBM multigroup pro- cedures to include the effects of of these "outside” materials. A new identxfy a factor eliminated in the development of the presently used formula must be retained as a factor in both the solution of the homogeneous equation, A , and the corrective term, Pn, This factor is the ratio of the values of o* {(defined in ANP-58)¢3) un each side of the interface through which the neutron being solved, to a of reactor index number 1s required to the different regions. Alse, flux equation 1is power given by the number space points between the first core- reflector interface and the i1nteviace Otherwise the eqguations form to the two- of 1nterest, are i1identical 1n region solution. Cylindrical Multigroup Calculation (N. Edwonson, Physics Division)., It is desired to calcolate the neutron (Q)Abstracted from the report by C. B, Malls, The Multiregion Reactor Problem as Applied to the Multigroup Method, ORNL, Y-12 site, report Y-F10-68 (Aug. 16, 1951;. : 3. K. Holmes, The Hultigroup Method as Used by the ANP Physics Group, ANP-58 (Feb. 15, 19561). PERIOD ENDING SEPTEMBER 10, 1951 and the effective multi- plication constant k for a reflected cylindrical reactor with bare ends by transforming the partial differential Fermi age equation. discribution The nettron flux ¢(r,u) and che slowing-down density g(r,u) are assumed symmetrical relative to the axis of the cylinder. This assumption reduces the variables to r and z. 1t 1is further assumed that 1n the reactor equation r and z may be This the expression of the z dependence by a factor space separated. leads to Ly 9(H + d) where H 1s the half-cylinder length and 4 is the linear extrapolation length, 1t 1s assumed that d has the same value for all lethargies and for both the core and the reflector. It 15 assumed that the neutron flux @lr,u) is = 0 and isequal to zetro at the extrapolated boundary finite for r of the reflector. The core-reflector boundary conditions avre: (1) ¢{(r,u) and (2) -D(op/Or) is Work on this problem is is continuous cOontinuous. being carcied on as rapidly as possible, and the formulation is in process of publication in report form. The Effect of the Boron Blanket in the ANP Reactor (C. B. Mills, Physics Division}. The a tveactor with an method of solution for energy~dependent core and reflector(*? been transformed) invto the system absorber between has (A)RA R. Coveyou, Spherical Reactor with Absorbing Interface. [I[, ORNL, Y-12 site, veport Y-F10-52 {April 30, 1951). ANP PROJECT QUARTERLY PROGRESS REPORT required by IBM multigroup procedures. This consists of computation 1in this order: 1. Compute a flux distribution in core and reflector with a source term in the core only. 2. Repeat step (1) for a source term in the reflector only. 3. Ceompute separately currents due to sources and reflectors at the core-reflector interface, Compute four currents. 4., Compute the corrective term for reflector flux due to a core source and the term for core flux due to a reflector source. 5. Add flux and correction 1in core and reflector, and use the total flux to determine the source term for the next higher lethargy group as in the regular multi- group procedure, STATICS OF THE AIRCRAFT EXPERIMENT REACTOR Summary of Calculations on the ARE (B. T. Macauley, USAF). Multigroup reflected-reactor calculations on the latest ARE design, hereafter referred to as the design of 10 June 1951 and 1llustrated in Fig. 3.1, were made during the past quarter. The core of this design, having auranium inventory of 25 1b, is referred to as Core 93. The composition of the core and the design data are given in Table 3.1. Two series of reflected reactor calculations were made; the first consisted of a spherical core having the properties listed in Table 3.1 and surrounded by the side reflector (Reflector 519), and the second series 34 of calculations consisted of a spheri- cal core having the properties listed in Table 3.1 and surrounded by the bottom reflector (Reflector 520), The constituents in both the side and bottom reflectors are given in Table 3.2, Calculations were made using the compositions of the ARE core (Core 93) except for varying the uranium con- centration and using Reflector 520. The results were plotted showing how keff varies with uranium investment (Fig. 3.2). Also, calculations were made using Core 93 (25 1b uranium investment) and each of Reflectors 519 and 520, varying the core thickness to show how k varies with the reflector thickness %gig. 3.3). Graphs of &, § VS, reflector thick- ness for the 3tft right cylindrical corec design (Core 84; 1incomplete curves previously reported in the last quarterly), as well as this core re- duced in volume by one-third (Core 91) and one-sixth (Core 92) have been drawn and show the same general relations as indicated in Fig. 3.3 for the ARE core (Core 93). The reflector saving for the ARE core (Core 93) is given in Fig. 3.4. Numerical calculations were made to determine various design and nuclear coefficients, The results are summa- rized in Table 3.3. Figure 3.5 1s a plot of the spatial distribution of the lethargic average of the fast flux in the core, and Fig, 3.6 is a plot of the spatial distribution of the thermal flux 1in the core. Plots of the flux spectrum at various radii are shown in Fig. 3.7. Figures 3.8 and 3.9 show the leakage FOR PERIOD ENDING SEPTEMBER 10, 1951 DWG. 12879 SECRET jl ) ANP~-PHY-210 ?/IE 69 % REFLEGTOR ! PHYSICAL DATA TOTAL . : ZONE | ZONE 2 REFLECTOR ZONE 2 % REFLECTOR 935 775 91.55 ‘ % COOLANT | 6 8.20 2 a4 _ % STEEL 0.37 1.65 0.52 ALL DIMENSIONS ARE IN INGHES o, VOID 4.50 12.40 5.49 Fig., 3.1 - Schematic Drawing of ABE Core Arrangement. 35 ANP PROJECT QUARTERLY PRCGRESS REPORT TABLE 3.1 Comgosition and Design Bata of Corc 23 Core geometry Cylinder Core diameter (in.) 29,45 Core height (in.) 34.00 Core volume (ftd) 13.411 Core composition: VOLUME FRACTION VOLUME (£t%) (%) Moderator (ReOQ) 10.872 81.10 Liquid fuel 0.185 1.38 Coolant (sodium) 0.621 4.63 Fuel-tube metal 0.130 0.97 Coolant-tube metal 0.078 0.58 Control-tube metal 0.221 1.64 Void 1.304 9,72 Fuel-element dimensions 225 1in. o.d. and 0.025 1in. wall thickness No. of fuel tubes 65 No. of control elements 1 Thickness of side reflector (in,) 8.75 Thickness of bottom reflector (in.) 8.00 spectrum from the core and the re- There results a critical mass of flector, respectively, Figure 3.10 1is the spatial power distribution with a uranium investment of 25 1b, Estimated Critical Mass of the ARE (N. M. Smith, Jr., Physics Division, and B. T. Macauley, USAF). The results of the previous section have been employed to estimate the critical mass of the ARE design oif 10 June, The calculations are summarized 1n Table 3.4. 36 27,8 1b of uranium. This 1s somewhat higher than the value of 21.51b gquoted in the last quarterly report (ANP-65) for a reactor of two-thirds volume of the design of 10 March 1951. This 1in- crease has been brought about by the engineering of the core of a more of BeO block clearances, and structural densities., The design of 10 June contains less Be( and more inconel and realistic estimate densities, sodium, FOR PERIOD ENDING SEPTEMBER 10, 1951 TABLE 3.2 Volume Fractions of the Materials in the Side and Boettom Reflectors CONSTITUENT (fe3) (%) VOLUME | VOLUME FRACTION Reflector 519 (Side Reflector) Reflector {(Be() 19.676 94.89 Coolant (sodium) 0.41 1.98 Structure {(inconel) 0.074 0.36 Yoid G.590 2.85 Reflector 520 {(Bottom Reflector) Reflector (BeD) 2.56 81.1 Coolant { sodiam) 0.220 6.98 Coolant-tube metal 0.0185 0. 58 Control metal ©0.0520 1.64 Void 0.3072 9.72 Contrel Rod Effectiveness (J, W. Webster, Physics Division, and R, J. Beeley, Oak Ridge School of BReactor Technology). The last guarterly report (ANP-65) gave the results and method of calculation on the effectiveness (Ak/k) of an axial 2-in. B,C control rod. The heating due to the (n,a) reaction and gamma absorption was also reported, This quarter it was decided that seven 2-in. control rods would be necded in the ARE tooffset the effects of temperature and fission product poisons with a margin of safety. Caleulations were made by the Nordheim- Scalettar Method(®"7) to determine the best placement for these rods to obtain the maximum effectiveness in k. The physics work of the last six months on the statics of the ABRE control system has been written and distraibuted as report Y-F10-71.¢%) "Included in this report is a discussion of three-group theory when, of the three solutions of DWG. 12880 SEGRET T 1t {07 [ ANP-PHY-95 ] REAGTOR G77, STANDARD 1,05 f—- > 8 1,03 F—- 3 REAGTOR 1024 x - x 0 o Moo - x = o h .. A Q1 ut w T 0499 t— e t [} e .S ] o ur o [ W 0.97 - @ { REAGTOR 1023 3 ‘:- e —— L. Q A 0gs e A CORE RADIUS 43.807 om REFLECTOR THIGKNESS 18.8C3 cm Q.93 P REFLEGTOR 520 -1 81.10 % Be0, 2.22% INGOREL, 6.98% Na ! o9l ) ' | J ."J J [ J [ B 16 8 0 22 24 26 URANIUM WEIGHT {10} : ¥Fig. 3.2 - keff vs., Uraaniom Weight for the ARE. (5)R. Scalettar and L. W. Nordheim, Theary of Pile Control Rods, MDDC-42 (decl. June 17, 1946). (6)C. BR. McCullough, Summary Report on Design end Development of High-Tenperature Gas-Cooled Power Pile, MonN-383, Appendix {Sept. 15, 1547); Criticality and Control, NEPA-6, Appendix (Ove. 1, 1948). (7)J. W. Webster, {ontrol-Rod Effect Reactivity and Power Distribution, NEPA 1C-50-2-52 (Feb. 7, 1950). , (B)J. W. Webster and R. J. Beeley, Ph Y Calculations on the ARE Control Heods, ORNL, ¥-12 site, report Y-F10-71 (Aug. 29, 1951). 37 ANP PROJECT QUARTERLY PROGRESS REPORT TABLE 3.3 Summary of Calculations on the ARE Core Yaving 25 1lb of Uraniomand Reflector 520 Nuclear coefficient of reactivity: From start-up to normal operation Ak/kE = -0,007995 Temperature coefficient Ak/k/°F = 5.04 x 1076 . .. . Ak/k Uranium mass coefficient of reactivity = +0.300 Am/m Density coefficient for core materials: AL /T Moderator f—lll = +0.4894 Do/ L) peo Ak [k Coolant [—“*—} = -0,00636 Do/ P ke Ak /& Structure [mmm*] = -0.1140 AVL/p inconel Density coefficient for reflector materials: Ak/w Moderator {—-fll = +0.2246 Do/p BeO (Ak /& Coolant "fi““] = -0.00093 \Bo/o Na Ak/k Inconel [ } = -0.03836 A¥Z/F) inconel Lifetime (Core 93 with Reflector 519) 1.535 x 10°? sec % thermal fissions 59.8 Spatial and lethargic average of fast (0 < u < 18.6} flux, &&a“ 16.12 neutrons/cm?/sec per unit of lethargy per fission per second per cubic centimeter of core Spatial average of thermal (u = 18.6) flux, ¢ 47.99 neutrons/cm?/sec per fission per second per cubic centimeter of core thermnl Integrated flux at 3 megawatts at central position 1.682 x 10'* neutrons/cm?/sec FOR PERIOD ENDING SEPTEMBER 10, 1951 OwG. 12881 . SECRET AUGMENTED CORE RADIUS (em) a4 48 52 56 80 &84 €8 T2 S 76 80 84 88 92 130 _ . T rrTr Tt syttt : - ANP-PHY- 196 | 126 CORE 93, kg = 1.664 : / : ] 1,22 fe-es : ' - m BARE CORE ' : = (1) 1BM GALCULATION e 118 = - - “REFLECTOR 519 ™ e l— ST e 3. e A0 - -t 3 . o . e — - w g . w108 P - i a2 Lo “~ REFLEGTOR 520 —_— w . : o Mooz — ] b oo wd £ [~ ] < )u‘. o 098 t— ] w b~ o i - = = o ' o S 094 — ] & R Lo ] 0.90 b COMPOSITION ~ VOLUME FRAGTION ] L1QUID FUEL BeD NGONEL Na — GORE 93 - 0.0138 0.8110 CO3IS £.0463 — (W, = 0.0000716 % I07* atoms /ce) 086 . REFLEGTOR 519 0.9489 00036 00198 : = / : REFLECTOR 520 0.810 0.0222 0.0698 — ] 0.8 b - 0.78 ~ ‘ — Cora _ 1 l l o 5 1G 15 20 25 30 35 40 45 50 AUGMENTED REFLECTOR THICKNESS (cm) Fig. 3.3 - Reactivity of the ARE Core Backed by Various Reflectors or Ad- ditional Core Material.. ' 39 DWG, 12882 " SECRET [ 1+ 1 17 17 1 1T 71 | T I—" L e ANP-PHY~-203 —] 20 — ] REFLEGTOR 519 — - 6 — ] g — — ] z REFLECTOR 520 > & g b ] & 5 | w COMPOSITION (VOL FRAGTION) T E ol W, FUEL B8e0 INCONEL Na CORE 93 0.716 x 10%° 0,038 0810 00319 0.0463 o REFLECTOR 519 0.9489 0.0036 0.0198 —— REFLEGCTOR 520 0.8110 00222 00698 — 4t —] b 0 4 8 12 5 20 24 28 32 36 AUGMENTED REFLECTOR THIGKNESS (cm) 40 Fig. 3.4 - Reflector Savings of Varions Reflectors Backing the ARE Core,. the buckling obtained, two are complex of six. conjugates, instead of the uswual situation where all three are real,one positive and two negative. This case was encountered in the control rod studies, 31.3% for 9.5 in., and 29.3% seven rods in each case was: for a radial distance of 6.5 1in., 31.1% for 7.5 in., 32.0% for 8.5 1in., for 10.5 The total Av/v (effect of thimble not deducted) obtained for the 29.1% in. In regard to the shadowing effect of the rods on each other, numbers are to be compared with 7 X 5.3 = 37.1%. Thus the minimum shadowing effect of (37.1 - 32.0)/37.1 The discussion of the effectiveness of the seven control rods given in Y-F10-71 follows: The total effect on reactivity of the insertion of seven 2-in. rods was investigated. The pattern was one axial rod and a ring of six at the same radius equally spaced around the center rod, Five different distances from the axis were tried for the ring 40 at a radial distance of 8.5 1in. the axis and the net effect of the seven rods, deducting the effect of the thimble, is 22% in reactivity. It follows from the results the total reactivity effectiveness these = 13.7% occurs with the ring of six of f that FOR PERIOD ENDING SEPTEMBER 10, 1951 TABLE 3.4 Calculation of Uranium Requiremént for the ARE Design of 10 June 1851 ; RELATIVE WVEICHTED URANIUM MASS ) i : k R ib CORE REFLECTOR POSITION YALU& eff eff (1b) 93 None Top 1/6 0.8045 0.1341 93 520 Bottom 1/6 1.0730 0.1788 93 519 Side 2/3 1.1585 0.7722 Weighted value of keff for clean homogeneous reactor 1.085 Seven control rod thimblés at 0.014,each:!ess 14% for mutual shading ; ' -0.078 Net ARE, clean, with control-rod thimble 1.007 Total keff in clean reactor desired for fission poison override, experiment, and control instrumentation 1.041 Additional & needed for above 1.034 T Teff Ak Additionaj uranium needed for above, usihg —— = 0.300 _ 2.8 . . Ar/a Uranium in Core 93 25.0 Total uranium in reacting volume (critical mass) ' 27.8 5% allowance nbove and inside B,C curtain 1.4 Total uranium inventory requirved 29.2 _ Estimated probable error +10%, -20% 15 not sensitive to rod placement in the region investigated, 1.e., 6 to 10 in. from the axis. This 1s caused by the fact that two opposing factors atre acting. When the rods are put close to the axis they are inserted into a tegion of higher undisturbed flux but the shadowing effect of rods on each other is large; on the other hand, when the rods are put farther from the axis they are inserted into a region of lower undisturbed flux, but any one rod does not tend to feel such a large flux depression caused by the insertion of neighboring rods. The two effects counteract to givea rather broad region of good rod placement. It is known from previous calculations (3 however, that if the rods are very close to the axis or very far oput toward the reflector, a sharp loss in combined effectiveness is incurred. Since the hexagon moderator blocks are about 3% in. across flats, it develops that the possible radii awvailable for rod placement are 3%, 7.5, 11.235 in., ete. The closest to optimum place- ment is 7.5 in., and this has been adopted in the current ARE design. 41 = 8o FLUX (neutrons/cmz/sec per unit of lethargy per fission) 3C.0 25,0 2C0 18.0 10.0 5.0 DWG. 12883 SECRET e e e e LATTICGE POINT, n Fig. 3.5 - Spatial Distribution of the Lethargic Average of the Fast Flux,. — ANP-PHY-204 — L REACTOR 977 o - COMPOSITION (VOL FRACTION) L LIQUID Na N FUEL BeD INCONEL COOLANT T CORE 93 0716 x 10%° 00138 0.8110 00319 00463 ] REFLECTOR 520 0.810 00222 00698 T f b, ridr T . . T _ space __ Spatiai average of integrated fast flux, d)'fcas? = frzdr . L = 16.12 neutrons/cme/sec per unit of lethargy _ = S, (v)du n | where = $n =70 _ . For 3 megawatts, spatial avercge of integroted fast flux =0.904 X IOB neuTrons/cma/sec ] thermal flux =1.086 X (02 neutrons/cmZ/sec — Ratio of peak to agverage =1.884 —] Ar =2.067 cm e sy b ] 2 4 6 8 10 12 i4 5 i8 20 22 LY04AY SSTAU90¥d A THIALUVAD ID3If0oHd JdNV - 75.0 70.0 65.0 60.0 55.0 45.0 FLUX (neutrons/cm - "% CONTROL -ROD PATTERN o1 4 - g S o 3 ! & 3 = ¥ | NORDREIV -SCALETTAR o5 1 5 z APPROXIMATION - © o -] o 11 1 I 0 10 20 30 10 50 g0 70 RADIAL DISTANCE FROM AXIS (cm) Fig. 3.12 - Badial Thermal-Flax Distribution iz the ARE Reactor with Seven Control Rods; Placement No., 2. Figures 3.11, 3.12, and 3.13 show the distribution of the thermal flux along a radius vector through a rod and in between two rods for three different placements corresponding to I o 1n., of 6.5, 8.6, and 10.5 radiil A8 PROGRESS REPORT DWG, 12891 SECRET 25 L— ANP-PHY -146 | 20— 15 — =< 3 T GONTROL -ROD ol PAT TERN Q Q o x o o @ o =} a < [ = & 0.5 5 2 | norRDHEIM-SCALETTAR 7 © o APPROXIMATION L 5} P b L | 0 10 20 30 a0 50 50 70 RADIAL DISTANCE FROM AXiS (cm) Fig. 3.13 - Radial Thermal-Flux Distribution in the ARE Esacitor #ith Seven Control Rads; Plaocement No. 3. respectively, for the circle defined by the centers of the six outer rods. KINETICS OF THE ATRCERAFT REACTOR EXPERIMENT M. J. Nielsen, USAF J. W. Webster, Physics Division In the last qguarterly report 1t was mentioned that solution of the non- linear kinetic equations for the ARE was in progress on the IBM machines, Three graphs were included showing the calculated response of flux (or power) and fuel temperature toa step increase in reactivity of 0,009125 (25% over prompt critical). This quarter the adaptation of the kinetic equations to solution by IBM machines was continued. Three types of problems were comsidered: (1) the response to astep change 1n reactivity with the coolant inlet temperature held constant, (2) the response to a change in the coolant inlet temperature with no control rod motion, and (3) the response to a change in the coolant inlet temperature with the control rod being activated by the change in coolant inlet and outlet temperature, The equations and method of solution have been written up in ANP-68,(9) Solutions for the ARE (3-megawatt) and ANP (200-megawatt) reactors and for variations of parameters in the neigh- borhood of those for the ABE and ANP will be distributed in report form in the near future., The results of kinetic calculations completed to date by this method are presented i1n the form of curves of fuel temperature vs. time, and phase plots of flux vs. fuel temperature. The latter type of plot has the advantage that the new asymp- totic steady-state flux and fuel temperature can be shown. flux vs. time, Most of the curves have to do with the response of the flux and fuel temperature to a sudden change of con- trol rod position or to an accident such as a deformation that might cause a step increase in k. The changes in reactivity inserted are 8k = 0.002, 0.004; 0.006, and 0.009125 where the last represents a 25% increase over prompt critical. The corresponding flux and fuel-temperature response are illustrated in Figs. 3.14, and 3.15. A phase plot combining these two figures is given 'in Fig. 3.16. Various parameters are changed to study their effect, for example, the neutron lifetime, the uranium mass-reactivity coefficient which has to do partly with the effectiveness of the B,C (g}M. J. Nielzen and J. W, Webster, Solution of Kinetic Equations of Cylindrical Liquid-Fuel Reactors, ANP-68 (Sept. 18, 1951). PERIOD ENDING SEPTEMBER 10, 1951 curtain at the top of the core, and the conductivity of the fuel. The reactor is assumed to be at its design power of 3 megawatts at time zero except in one calculation — investigation of start-up accident -~ 1n which the power 1s taken as 300 watts at t = (. In none of the cases considered does the energy release seem sufficient to damage the reactor, and 1in all cases the fuel temperature responded in an overdamped fashion, The fuel temperature for the case of &k = +0.009125 stabilizes out at a new temperature 186°C hotter than the design value of 800°C when no counter control rod action is taken. There is, of course, some magnitude of 5k, which would cause temperatures result- ing in failure of the fvuel tubes or fuel solution before the stabilization occurred, Mass-Reactivity Coefficient. The investigation of the effect, on the response, of assuming different values for the uranium mass-reactivity coef- ficient (Figs. 3.17 and 3.18) demon- strated the strong sensitivity of the response to this quantity. It 1is clearly very important that the B C curtain be as opaque to neutrons as it is possible to make 1t. ' Fuel Conductivity. The increase 1in the fuel conductivity by a factor of 10 had remarkably little effect on the response to a ok_ (Fig. 3.19). The explanation seems to:be that the heat input to the fuel due to fission 1s rather large compared toc the loss by heat transfer to the coolant during the time of the power pulse. Hence a change in the coocling term {(which is about a factor of 2 rather than 10 since the conductivity affects only 49 o < (¢—¢O)/gb0, RELATIVE EXCESS FLUX, OR POWER 200 8.0 6.0 14.0 2.0 100 8.0 6.0 40 2.0 DWG. 12892 SECRET ol T I e ool — ANP-PHY- 175 — b REACTOR PROPERTIES (IDENTIFICATION, A-1) — i GOMPOSITION (VvOL FRACTION) FUEL SOLUTION BeO Na INGONEL CORE 84 0.01705 0.89000 0.07:80 0.02i80 T e, —————— —n 84=0.009125 8k=0006 Ek=0.004 REFLECTOR &I13 0.88910 0.006358 0.09i8C CORE DIAMETER = LENGTH= 3 1 REFLECTOR THICKNESS = 5.57 in. THERMAL FISSIONING =68% AVG. FLUX, ¢,, AT STEADY-STATE DESIGN POWER =6.5 x 10" DESIGN POWER =3 megawatts STEADY-STATE FUEL TEMPERATURE AT DESIGN POWER =300°C CONSTANT INLET COOLANT TEMPERATURE NEGTRON LiIFETIME = 1.4 x 10™ % sec 84 =0.002 | ] T T : : I T T ; 7 1 T T | | : i ] ! ] { ! | i I { | 2.0 2.4 2.8 3.2 36 4.0 4.4 4.8 TIME (sec) 5.6 6.0 Fig., 3.14 - Response of Flux toc a Step Increase in Reactivity for VYarious Reactivity Changes. IHOdAH¥ SSAH90¥d ATHALYVAO 103alodd 4NV DWG. 12893 0.0 SECRET 20, | 7 T T T T 1 Y 1 T T T T T T T T T T T T T P oo T L REACTOR PROPERTIES {iDENTIFICATION, A-1) o ANP-PHY-178 _| COMPOSITION {VOL FRACTION) 8.0 — FUEL B SOLUTION BeO Na INGONEL GORE 84 0.01705 0.B9000 0.07160 0.02:80 w " REFLECTOR 513 0.88910 000836 0,09160 >3 3 180 i— — > = L ] © CORE DIAMETER=LENGTH= 3 f1 ‘ W a0 REFLECTOR THICKNESS =557 in. | w THERMAL FiSSIONING = 68% g L : AVG. FLUX, ¢, AT STEADY-STATE DESIGN POWER =66 x i0" _ % DESIGN POWER = 3 megawatts I . s 120 = STEADY-STATE FUEL TEMPERATURE AT DESIGN POWER = 800°G — < CONSTANT INLET COOLANT TEMPERATURE H — NEUTRON LIFETIME = 1.4 x 07 % sec B - B Z 100 ‘ 4 - _ — | : — - £ o = - 34 = C.009125 ] -~ & | u o 80 F— o — e = — e == o $#=0.006 st - = S - - & - = u:_J 60 t— o . : — T ' 3 s . T 8420004 - > L = © oo 2 4.0 - — A = @ a e ] 8 o 2.0 — ’ 84 =0.002 — o y - 2 — : i : P | ! ' z s o1 | + | | | o b g 0 0.4 0.8 L2 1.6 2.0 24 2.8 32 38 40 4.4 4.8 5.2 56 6.0 o TIME (sec) - = _ Fig. 3.15 - Response of Fuel Témfietfitnré to Step Increase in Reactivity for Various Reactivity LR P Changes. IS6t1 A o DWG. 12894 SECRET | T ¥ T T T T T T T T T T T 16.0 —— ANP-PHY-I73 __] 015 L REACTOR PROPERTIES (IDENTIFICATION, A-1) ] COMPOSITION (VOL FRAGTION) 14.0 — FUEL | SOLUTION Se0 Na INCONEL 0.20 CORE 84 0.01705 0.89000 0.0760 0Q.02180 h REFLECTOR 513 0.86910 0.00636 0.0960 |20 —— —_— CORE DIAMETER = LENGTH= 3 ft REFLECTOR THICKNESS= 5.57 in. THERMAL FISSIONING = 68% 06 AVS. FLUX, #,, AT STEADY-STATE DESIGN POWER=6.6 X 10" ' DESIGN POWER = 3 megawatts STEADY-STATE FUEL TEMPERATURE AT DESIGN POWER=800° T CONSTANT INLET COOLANT TEMPERATURE = 623°C — NEUTRON LIFETIME =14 x 10 % sec 8.0 LY0ddY¥ SSAYIOHd ATHIALEVND 12Af0¥d 4NV (¢>—¢0)/¢0, RELATIVE EXCESS FLUX, OR POWER 84 =0.009125 —] 0.40 20 0 20 40 60 80 100 120 140 i60 180 200 DEPARTURE OF FUEL TEMP. (°C) FROM STEADY~STATE DESIGN VALUE Fig. 3.16 - Phase Plot of Flux vs. Fuel Temperature for VYarious Step Imcreases in Reactivity, DWG. 12895 SECRET 100 | 17T 1 1T 1 ' f ] 1 T T ] ANP-PHY- 129 90 }— ] E ap REACTOR PROPERTIES (DENTIFICATION, A=) ] 2 COMPOSITION {VOL FRAGTION) e | FUEL &, 70— %‘=oé_fl? SOLUTION BeO Na INCONEL ~ — < / 7 CORE 84 0.01705 0.89000 007160 002180 b REFLECTOR 513 0.88910 0.00636 0.09/60 w,, 60 t— _— - < wvs L . GORE DIAMETER=LENGTH=3ft_ o o S x 50 — Bk _ o opp A REFLECTOR THICKNESS =5.57 in, ] 2 w < kYR, THERMAL FISSIONING = 68% - ul v a0l AVG. FLUX, ¢, , AT STEADY-STATE DESIGN POWER=6.6 x 10" _ rm < CONSTANT INLET COOLANT TEMPERATURE = &y Ak A Ak/k = FRAGTIONAL CHANGE IN EFFECTIVE MULTIPLICATION CONSTANT = B 20 L— 21 —pi64 20 | S & 3 k "m Am/m=FRACTIONAL CHANGE IN URANIUM INVENTORY — < NO MOOERATOR EFFECTS & & LK - pa6 A - ‘ 20 — k ‘ m — 2 & =z A ..i:ojgaé_’? % 10— k — :3 S I 1 i 4 4 . 4 £ z O I [ 1 1 i 1 i T 3 oy r T g O 04 08 {2 16 20 24 28 32 36 40 44 48 52 56 80 - ~ TIME (sec) o = Fig. 3.17 - Power Response of ARE to Step Reactivity Change of 0.009125% with VYarious Mass-Reactivity Coefficients. ' g5 ICel wn » P DWG. 12895 % SECRET T T T T T T T T T T T T T 1T T 1T 1T T 17 L g 520 ANP-PHY-154 —] S REAGTOR PROPERTIES {IDENTIFIGATION, A-1] — = wi -— > CORE DIAMETER = LENGTH= 3 #t d £ 400 REFLECTOR THICKNESS = 5.57 in. ] g - THERMAL FISSIONING = 689 = 2 AVG. FLUX, ¢, AT STEADY-STATE DESIGN POWER = 6.6 x 107 — ! b DESIGN POWER = 3 megawatts w 360 STEADY-STATE FUEL TEMPERATURE AT DESIGN POWER = BOOD°C — ~ 2 CONSTANT INLET COOLANT TEMPERATURE = 623°C g o NEUTRON LIFETIME = 1.4 x 1079 sec T & § 300 g Ak/k =FRAGTIONAL GHANGE IN EFFECTIVE MULTIPLICATION GONSTANT ] = = = . o g Ak/k =0.082 Am/m Am/m= FRACTIONAL CHANGE IN URANIUM MASS IN CORE ™ th q "__,——" — ‘m Q - CE, I //’ wn [ied 280 ~ — — = = < = & -~ —_ o |‘ o Aktk=0.164 A m/m g & 240 7 — 2 ! s = 2 i =~ g ! / — T e S 7 4 g f T Akzk=0.246 Am/m 2 |/ - _ o ’/ T — - B 160 I ;/ e e T 7 s / v T —_ € S e Aksk=0D.328 Am/m E |20 !, //’ ey & Ilf /'/ o ll_/' — 1 80 Ifif — ] 40 — i ; : ; ; i ! | i { ' i i , 0 L bbbt e b [ | 04 0.8 1.2 16 2.0 2.4 2.8 32 36 40 44 ag 5.2 56 6.0 TIME (sec) Fig. 3.18 - Response of Fuel Temperature to a Step Increase of Reactivity of 0.009125 for Various Effective Mass-Reactivity Coefficients. DWG. 12897 SECRET 1T T 1 1 1 1 1 ' 1 1 7 T T T o~ T _I 3 6 L ANP-PHY- 51 REACTOR PROPERTIES ({IDENTIFICATION, A1) B COMPOSITION (VOL FRACTION) - L ‘ ‘ ' " FUEL o o — 0.5 SOLUTION BeQ Na INGONEL 6 — GORE 84 0.01705 0,89000 O0.07i160 0.02180 _| 020 REFLECTOR 513 ' ' 0.88910 000836 0.09160 CORE DIAMETER = LENGTH = 3 1 ] W4 — REFLECTOR THICKNESS =5.57 in. i z 4 THERMAL FISSIONING = 68% ‘ I i - AVG. FLUX, ¢, AT STEADY -STATE DESIGN POWER =6.6 x10'® ] o STEADY-STATE DESIGN POWER = 3 megawatts < 12— ASYMPTOTIC FLUX =20.4 x 10'? — = 0.10 ’ ASYMPTOTIC POWER = 9,3 megawatts . CONSTANT INLET COOLANT TEMPERATURE = 623°C — @ 0 25 Y AVG. FUEL TEMPERATURE AT DESIGN POWER = 800°C Q10— ) : ASYMPTOTIC AVG. FUEL TEMPERATURE = 1017°C ] > g S > & = g [ |2 — - o4 gl el - T = -] W o o & o ~ & 0.30 ] < 5 | — S = S 4 1005 ] 2 & — 0.40 — o A o LI 0.60 : - ) ) .00 500 C o eed | | - Z L Ll by 3 O : rpy 0 20 40 60 80 oo 120 140 160 180 200 220 DEPARTURE OF AVG. FUEL TEMP (°C) FROM STEADY-STATE VALUE ' = ~Fig. 3.19 - Phase Diagram of Flux vs. Fuel Temperature for Response of ARE to a Step Increase in Reactivity of 0.009125. Fuel conductivity arbitrarily increased by a facter of 10, ' | 1S61 i & ANP PROJECT QUARTERLY PROGRESS REPORT one term of several in the heat- transfer expression) does not markedly affect the shape or magnitude of the power pulse, Neutrom Lifetime. Variation of neutron lifetime | was calculated only for the case of 8k less than the delayed neutron fraction (and with the stabilizing effect of the fuel expansion neglected in order to see more clearly how the lifetime affects the growth of flux). For 8k, larger than [, the delayed neutron fraction (and no self- stabilization), the flux grows, of course, approximately like 5oc o PHeTA) N When 8k _ is less than the fraction of delayed neutrons, the flux also increases very rapidly initially (as if all fission neutrons were prompt) with an initial slepe inversely pro- portional to the lifetime., However, after a time which 1is directly pro- portional to the lifetime, the flux curve flattens out with a rather sharp "knee"™ and rises from then on with the "stable'" reactor period {(when there are no self-stabilizing features). The stable reactor period is a function of the 8k _ but not of the neutron lifetime. The rapid initial rise 1in the flux is due to the fact that the rate at which delayed neutrons are being returned to the population 1is almost equal to the rate at which delayed-neutronemitters are being created, while the flux 1s still near the steady-state value. The reactor thus behaves momentarily as 1if all fission neutrons were prompt. These facts are borne out qualitatively 1in comparing the flux response to a dk_ = 0.002 for three different assumed 56 lifetimes (Fig. 3.20). Equation 11.35.3 of TID-386¢1%) provides a mathematical interpretation of the flux behavior as obtained (Fig. 3.20). Although the effect on reactivity of the fuel expansion was not included in these calculations, it is apparent that for step changes in reactivity less than the fraction of delayed neutrons the peak power 1n the power pulse (when fuel expansion is allowed) will be roughly independent of the neutron lifetime for lifetimes smaller than 10°* sec. Start-Up Accident. In regard to the investigation of the response to a start-up accident, the results (Fig. 3.21) indicate that a step change in reactivity at low power will be less serious than the same step change at high power from the point of view of fuel temperatures reached, Coolant Temperature. The response of the flux and fuel temperature to a step increase i1n the inlet coolant temperature of 100°C is shown in Fig. 3.22. Since the rise 1in the coolant temperature as the cooclant passes through the ARE core (at design power of 3 megawatts) 1s 194°C, the drop through the heat exchanger is roughly the same. A rise in reactor inlet coolant temperature of 100°C implies a rise of heat-exchanger outlet temper- ature of approximately 100°C. The temperature drop through the heat exchanger must have been 94 instead of 194°C., This implies that the external power requirement was suddenly reduced from 3 to (94/194) x3 = 1.45 megawatts, Calculations on the reactor were made holding the reactor inlet coolant (IO)S. Glasstone and M. C. Edlund, The Elements of Nuclear Reactor Theory,. (November, Part III. TID-386 1950). DWG. 12898 SECRET 20 1 ; I f ! o ‘ i ’ 0T . o R b L T T T T - _ - _ _ S S - ANP-PHY-185 — 18— REACTOR PROPERTIES (IDENTIFIGATICN, A-1) — COMPOSITION (VOL FRAGTION) FUEL T e b SOLUTION BeC Na INCONEL CORE 84 0.0ITO5 083000 00760 002180 — REFLECTOR 513 0.88810 0.0083e QQ960 CORE DIAMETER={LENGTH=3 ¢ N REFLEGTOR THICKNESS = 5.57 in. THERMAL FISSIONING=88% 2 — AVG. FLUX, ¢, AT STEADY-STATE DESIGN POWER =66 X o' —_ DESIGN POWER = 3 megawotts T STEADY-STATE FUEL TEMPERATURE AT DESIGN POWER = 800°¢G T CONSTANT INLET GCOOLANT TEMPERATURE . - NEUTRON LIFETIME = 1.4 X 1077 sec (¢ —dy)/,, RELATIVE EXCESS FLUX, OR POWER o ‘0T HAGWILLAS ONIONH dOI¥Ad HOo4 NEUTRON LIFETIME = 0.4 %1077 sec — NEUTRON LIFETIME =14 % 10" % se [¢] § } i ' { | ! t ) i » : 1 ) ! i : ! P b b Lo e 2.8 e 44 4B 5.2 55 8.0 TIME {sec} [ Ca [ ™ < m o ot i o M 4 i m ) 0 Fig. 3.29 -~ Response of Flux to g Step Yncrease in Reactivity of 0.002 for Various Average Neuiron Lifetimes. HReactivity change due to thermal expansion 3 of the fuel not included. 1861 o o] (p—d,)/¢,, RELATIVE EXCESS FLUX, OR POWER DWG. 1289¢ SECRET o s : s 6.0 — ANP-PHY-139 0.30 REAGTOR PROPERTIES (IDENTIFICATION, A-1) — COMPOSITION (VOL FRAGTION) - 0.80 FUEL 50 — SOLUTION BeO Na INCONEL — CORE 84 0.01705 0.89000 0.07160 ©.02i80 — REFLECTOR 513 0.88910 0.00636 0.09160 — 40 — CORE DIAMETER=LENGTH= 3 t1 ] REFLECTOR THICKNESS =557 in. . 0.75 THERMAL FISSIGNING = 68% __ AVG. FLUX, ¢, AT STEADY-STATE DESIGN POWER=6.6 x 10" 30 b— DESIGN POWER =3 megawatts — AVG. FUEL TEMPERATURE AT DESIGN POWER =800°C - CONSTANT INLET COOLANT TEMPERATURE = 623°C — NEUTRON LIFETIME = 1.4 x 10™ % sec 2.0 |— — 10— —] - 0.64 | o O . — 048 -1.0 — 0 — o -200 -180 - 180 -140 -120 -100 -80 ~80 -40 -20 0 DIFFERENCE OF AVG. FUEL TEMPERATURE (°C) FROM DESIGN VALUE Fig. 3.21 - Phase Diagram of Relative Excess Power (Flux) vs. Fuel Tempera- ture, For response of ARE to a step increase in reactivity of 0.009125 when reactor power at 300 watts and initial fuel temperature are the same as the de- sign coolant inlet temperature. 20 LY0dA¥ SSAY90Ud ATHALYVNO XLJAf0Ud dNV FOR PERTOD ENDING SEPTEMBER 10, 1951 DWGE. 12900 : SECRET + .05 ’ I l I 0 ‘ , ANF-EHY-155 000 - -005 ] ~0.10 - t o L = 2 & -0.5 x . < REACTOR PROPERTIES (IDENTIFICATION, A~1) _:)_, COMPOSITION (VOL FRACTION) L. e o —020 — FUEL ) ] & : SOLUTION Be Na INCONEL 2 CORE 84 O.OI705 282000 007160 0.02180 L REFLEGTOR 513 0.839I10 000635 0.03160 Lul ; : Zo-025 = CORE DIAMETER =LENGTH =3 ft! = oL G REFLECTOR THICKNESS = 8.57 in. & THERMAL FISSIONING = 68%, SL030 RIG.FLUX, ¢, AT STEADY-STATE DESIGN POWER=6:6 % 107 = DESIGN POWER = 3 megowatts & NEUTRON LIFETIME = 1.4 2 107 * sec o STEADY-STATE FUEL TEMPERATURE AT DESIGN POWER =800°C 7 o 5 ASYMPTOTIC FUEL TEMPERATURE AFTER CHANGE = 815.6°C ' / 035 e _ _ — ” ASYMPTOTIC POWER AFTER CHANGE = i 58 megawatts //// e =040 - // e DOTTED PORTION OF GURVE ESTIMATED '/,"/ : -"‘\>// _ - ~045 — — - CALGULATED T POINT e . K. C. 1020 0 | | B | | | | L O 10 20 30 40 50 DEPARTURE OF FUEL TEMPERATURE FROM STEADY-STATE DESIGN VALUE (°G) Fig. 3.22 - Phase Plet of Flux (or Power) vs. Fuel Temperature, After 1in- crease of 100°C in inlet coolant temperature, ' : 59 ANP PROJECT QUARTERLY PROGRESS REPORT temperature at 100°C hotter than the steady-state design value, and it was found that the reactor would, long time, after a steady out at a new power of 1.575 megawatts, assuming no change in control-rod setting. This is very close to the new power demand, but, even 1f the coolant circuit from heat- exchanger outlet to reactor inlet was long enough to allow the reactor to settle to this new power, several more circuits of the coclant accompanied by small incremental temperature adjust- ments would be required before the power output settled to the requirement. new Actually, regulating control rod in motion by an error signal which 1s such a function it 1s planned to set the of the 1inlet and outlet coolant temper- atures that the point in the reactor coolant pass which 1s at 1250°F under design power remains at this tempera- ture (this point is about one-third of the way through the reactor). The inlet coolant temperature, outlet coolant temperature, and fuel tempera- ture will adjust themselves to new conditions when a new power reguire- ment 1s made, but always with the fulcrum point of the coolant at 1250°F, The scheme ensures that the coolant pumps which are located at the reactor coolant inletendwill not be subjected to high-temperature coolant {(approxi- mately 1500°F) when the reactor is at low power. This would occur 1f the control rod was not moved. An interesting feature of the curves showing the response of the flux to a step increase 1n coolant temperature of 100°C is the "thermometer" effect of the fuel-tube expansion. This 60 expansion lets more fuel into the core momentarily and the power increases. The fuel 1tself soon feels the thermal effect of the coolant temperature, however, at which point fuel 1s lost to the core owing to expansion above the B,C curtain, and the power starts down. PREPARATORY PHYSICS CALCULATIONS Correction to "Wigner Formuia for Resonance Escape" Probability(!!) (M. C. Edlund, Physics Division)., In the present multigroup method the age- velocity equations without a source are made homogeneous in the slowing- down density by assuming that the flux per unit lethargy 1s related to the slowing-down density by q(u) d(u) - . £z, + 3, + DB?) (1) This expression, correct for hydrogen moderator, does not hold for heavier moderators. A somewhat better approxi- mation than that given by Eq. (1) for nonhydrogenous moderators can be obtained from the integral represen- tation of gq. (ll)This procedure for obtaining a correction to the Wigner formula, and in particular the method of eliminating the d[Z (u)]/du from the Taylor’'s series expansion of zqu(u') was polnted out to the author by E. Greuling and G. Goertzel. An NDA report by G. Goertzel is forthcoming. FOR PERIOD ENDING SEPTEMBER 10, 1951 The slowing-down density in a mix- ture of N elements 1s given by Ii - 1~ . J {E%i Plu) T - . i N n ~ \ Zsi{ug) dlu’) glu}) = e ) I - a. diza- PHlu)] , ~ t A s - {ua'- u) (e "8 a.)du'.(3) pnl ey du * ' [:”””—-ai du’ (2) LA N Integrating, here . e Iy =8 25 dlu) + -1 Fa, M. - 13?2 B L 4:1,1- - e e .‘., I Mi 1 fl.i:il l d[gsi (b(U)] + €, + (4) . : 2 1 - a. an M. = mass number of the 1th scatter- l ing nucleus, €. = In (1/@i)~ Each Z_. ¢pu') in Eq. (2) 15 ex- panded as a Taylor's series about lethargy u. 1If there was no capture or leakage, 2_ ¢{u’) would be constant, Hence if the capture cross-sections do not vary much in a lethargy interval ¢, the variation of ES P{u’) mav be given quite well by just the first two terms of the expansion. The integrals in the summation Fq. (2) after expansion of ES H{u’) are of the form 61 ANP PROJECT QUARYERLY PROGRESS REPORT Again expanding %_ ¢{(u’') aboutu and taking only the first two terms, Eq. (5) becomes dg d 1 —=N 5. ) - du 1-a. =1 d[ZS Plu)] + -_— (' - a)l e® v gyt (6) du (6), N d ZE:? &, 2y elu) i=1 Integrating Eq. dq (7) du du and substituting Eq. (4) into Eg. (2), we obtain N g ‘—5& &, 2y, Plu) i=1 N B B; g; Plu) (8) du 62 where a.€. 1 2 l-—ai Defining average £ and 8 as N 7 E‘::i Zs 1 i=1 é’ e 2 s (9) N } Si Zs t T Bz 2 where 2_is the total macroscopic scat- tering cross-section, FEgs. (7) and (8) become dg g d[f 2. Plu)l du du d& _dlE; ¢(u}] = 2, p(u) + & »(10) du du _ dIB %, ¢(u)] g ~¢&2 o) - ——y s du FOR or df3 P g =& Z, plu) ~—2 @lu), du _ dZ, plu)] - B - . (11) du The derivative termS'hmis ¢(u) can now be eliminated from Eqs.(10) and (11), yielding dq B du S p). (12) In a finite system, the rate of change of slowing-down density with lethargy is equal to the total rate of loss of neutrons 1in the system; 1.e., d . 29 - (s + DB?) #(u) du a (13) where DB? #(u) is the rate at which neutrons leak from the system, Equations (12) and (13) now de- termine the relation between ¢(u) and g{u). Eliminating dq/du, we obtain - glu) d)(u) T”ES + ')/(Za 4 DBZ)’ (14) PERIOD ENDING SEPTEMBER 16, 1951 where If all the Esi(g)’s vary the same way with energy, df3/du and d&/du become equal to zero, and Eq. (14) becomes _ q(fi) P T ETG, + DRz) If we have a single moderator, 1 - a - ae ~aé/2 y o BN R (15) 1 - a - ae recalling that &€ = 1 - ae/{(l-a) and € = -In a. Rewriting Egq. (15}, - a toa . 2 r 1l -atalna - {aln Q)/zn(lfi) e - a In a In the limiting case of hydrogen moderator, a = 0 and lim vy = 1, a0 which checks with the rigorous result. 63 ANP PROJECT QUARTERLY PROGRESS REPORT In the limit of infinite mass scat- tering nuclei, a = 1 and lim vy = 0. a1 In this case there 1s no slowing down, and from Eq. (14) ¢ = 0. The resonance escape probabilaty in an infinite medium of a single type moderator may be obtained by eliminat- ing ¢(u) from Egs. (12) and (13) and integrating. In this case B? = 0, and S(u), the neutron source, is taken to be zero in the resonance region. The resonance escape probability 1is u a > p(u) = exp ~| —m-— &x, t yz a du.(17) 0 A. M. Weinberg has calculated the correct resonance escape probability for the case of constant cross- sections {1?) the result is p(u) = e-,u.u For carbon moderator and EG/ZS = 0.2, (= 1.118817. According te Eq. (17), 1 y_: = E — = 1.116024 sgs/za Ty (12)A. M. Weinberg and L. C. Noderer, Theory of Neutron Chain Reactions, ORNL CF-51-5-98, p. 37 (Aug. 10, 1951). for the corresponding case. If the resonance escape probability 1s given by the Wigner formula, = zfl. p(u) = exp - | ————— du, . E(Z, + 2)) _ 1 Lo — = 1.056401, E(Z /5, + 1) and, if the weak capture formula 1s used, 1) zd du, plu) = exp - £ 0 p = E fES = 1.267681. It is apparent that for the regions of heavier capture the best result 1is given by Eq. (17) for the case 1in which the absorption cross-section variles only slowly with energy. Effect of the "Wigner Formula for Resonance Escape" Correction (C. B. Mills, Physics Division). It has been shown by Edlund in the preceding section, using a method of Greuling and Goertzel, that a more correct relation FOR PERIOD ENDING SEPTEMBER 10, 1951 between the neutron slowing-down DBr+'Za : og(u) density g and the flux ¢ is — ; glu) + { * gzs oy, DB?) ' du | g(u) Plu) = — ‘ = fission source. EE_ + y(E_ + DB?) | where ~ : The corrective term y(Z_ + DR?) mus t | : be inserted into the solution of the usual recursion formulas. Thus the average values required for multigroup B 1 - a - ae - a€2/2 constants are : y E-m 1l -a - a€ N for a single moderator., Here 1 € = 1n l/a, ‘ . : 3§zsztr ¥ 7(32aztr + BY M-1y? | | N a = [-~] where M is the atomic ' zb M+1 number, : ' 2 fis + y{DB% + Za) £ = average loss in log of energy | per collision, ‘ : N ¥ = scattering cross-section (cm™1), z, s : : : - 2, = absorption cross-section (cm™ 1), ' gzt vy + DB?) D = diffusion constant, B? = geometrical buckling. The values of & = & and v for the Introduction of this definition into beryllium-moderated critical experi- the bare reactor eguation{!3) gives . ment are Uy I 0.2078, (13)M. J. Nielsen, Bare Pile Adjoint Solution, ORNL, Y-12 site, report Y-F10-18 (Oct. 27, 1950). Y il 0.1427. &5 ANP PROJECT QUARTERLY PROGRESS REPORT A hand calculation was made 1n the usual way for keff' The values are Corrected kcfif = 0,91617, Uncorrected kqfi = 0.90347. The effect of the correction is seen to be small but significant, The Transmission Coefficient of the B,C Curtainin the ANP and ARE Reactors (C. B. Mills, Physics Division). A boron carbide curtain 1s part of the design of the ANP and ARE reactors. The transmission coefficient as a function of lethargy of this curtain is required for determination of criticality coefficients and the neutron lethargy and space distribution. The B,C curtain is 2 in. thick in the ANP and 1 in, thick in the ARE reactor. In both cases i1t 1s a dense layer thickly studded with coolant- tube holes. For purposes of calculation the transmission of a solid slab is first computed and then a constant correction due to the total aperture of the holes is added. The transmission coefficient is almost unity for the high-energy neutrons. For thermal and near-thermal neutrons, the holes provide the only leakage through the curtain. The transmission 7y of a layer of thickness T and macroscopic absorption cross-section 3_ is given by the integral e~ Y y = 2(2,T)? 66 This function has been integrated numerically,(!*) The leakage through the holes i1s evaluated by assuming a cosine distraibution of neutron ve- locities (as for ) with a probability p{x) of being x cm off the axial line at each hole being 27x. A projection of the solid angle subtended by the reflector side of the hole permits evaluation of the probability of a neutron traversing the opening. A table of the fraction of the neutrons penetrating the B,C curtain at each lethargy group is given in Table 3.5. This table will be used for a trial solution for the neutron flux in the vicinity of the absorber, An iteration, requiring new values of v to fit the more accurate form of the flux, may be required, Hieatinmg in the Boron Carbide Curtain in the ANP Reactor (C. B. Mills, Physics Division). A solution of the reflected ANP reactor, including the effect of the boron carbide curtain intended to shield the fuelreservoirs, is not yet complete. The possibility of some secondary effect preventing the use of this curtain 1s of some interest prior to a more exact calcu- lation., One such possibility is the heating in the curtain due to absorption by the B!'%(n,a)Li’ reaction, with a release of 2.88 Mev. The model used for an approximate value of total heating per square centimeter and power distribution is a bare reactor with the ANP power distri- bution, The boron curtain side of the ANP reactor will be essentially bare because of the high absorption and low reflection. (14)B. R. Coveyou and J. E. Bradley, Tabulation of F-Functions, &1-1629 (May 1944). ?Gfi PERTOD ENDING SEPTEMBER 160, 1951 TABLE 3.5 B,C Transmission Coefficients ANP (2 in. B,C) ARE (1 in. B,C) v (holes} : v (holes) _ GROUP | v (slab) | (%-in. dia.) |y (total) v{slab) (%-in. dia.) y {total) 1 0.95671 0.01024 0.967 0.97800 0.00211 0.980 2 0.94504 0.01301 0.958 0.97200 0.00268 . 0.975 3 0.93051 0.01643 0.946 0.96428 0.00342 0.968 4 0.91235 0.02073 0.933 0.95464 0.00434 . 0.959 5 0.88990 0.02603 0.916 0.94244 0.00551 0.948 6 0.86228 0.03257 0.895 0.92720 0.00697 0.934 7 0.82874 0.0300 0.839 0.90828 0.00878 0.917 8 0.78802 0.0250 0.813 D.86492 0.01101 0.896 9 0. 57498 0.0150 0. 590 0.64038 0.01610 0.656 10 0. 14577 0.0065 0.151 0.17710 0.02150 0.1986 11 0.00580 0.00539 0.0112 0.05238 0.02610 - 0.07858" 12 0.00012 0.00539 0.00551 0.00524 0.03276 0.0380 13 0 0.00539 0.00539 0.00028 0.03276 - 0.0330 14 0 0.00539 0.00539 0. 60002 0.03276 0.03276 15 0 0.00539 0.00539 0 0.03276 - 0.03276 16 0 0.00539 0.00539 0 0.03276 0.03276 31 0 0.00539 0.00539 0 0.03276 0.03276 91 0 0.00539 0.00539 0 0.03276 ° 0.03276. 92 0 0.00539 0.00539 0 0.03276 . 0.03276 03 0 0.00539 0.00539 0 0.03276 | 0.03276 The absorption distribution was thickness of the B,C layer. The powér determined for each of the 31 lethargy groups by the usual current attenuvation method expressed by the F-function where b = Z d; Z absorption cross sectlon and d is the 15 the macroscopic dlstrlbutlon was determlned'byafldltlon of the contrlbutlon_of each lethargy group at several points through the B C curtain. The total power was 46 watts/cm? for the ANP reactor. A similar calculation gave 0.69 watt/cm? for the ARE reactor, The power distri- bution is given in Fig. 3.23. Most of this power is concentrated near the core surface of the curta1n. ‘ Temperature Savings in the ANP Reactor (R. J. Beeley, Oak Ridge School . of Reactor Technology, and C. B. Mills, 67 ANP PROJECT QUARTERLY PROGRESS REPORT DWG. 12901 - SECRET 100 §— — - ANP-PHY-137 ] 5 50 ||— . (0 / ] — tharr,-1) — - a. ABSORPTION = 2—70---—— B o H— j.—:o X ] L m e @ 8 1 ? ] @ dil,~1) 210()() g2 < ot e | e &n WHERE e 7 \7 3 a7 é x/L - !_—) 0 i~ ] % i X = DISTANCE FROM CORE SIDE OF THE B,C LAYER ] ~_ 1 N . = NEUTRON MEAN FREE PATH FOR ABSORPTION 7 N 5 I\ Io= [o(M)=INCIDENT CURRENT INTENSITY AT ] IR SURFACE OF B,C LAYER - < B TOTAL POWER REPRESENTED BY THIS ABSORPTION | o ASSUMING SAME CURRENT LETHARGY DISTRIBUTION | ] =N FOR BOTH ARE AND ANP REACTORS: = - ARE, 0.69 wafis/cmz; ANP, 46.0 wafi's/cmz J = THIS CURRENT WAS COMPUTED ON SAME BASIS 8 OF | FISSION PER cubic centimeter PER second @ l_ O 10— o [-—f z " - 9 . 5 ol © 05— . a | { W - : ] = - > 0 i = = o4 |— — : : 3 : - B W VoL % B ~ z 012 — NaCH, COOLANT AND MODERATOR 75.0 =— = & B 316 STAINLESS STEEL, STRUGTURE AND FUEL TUBES 20.23 L S Z FUEL, UZ*° 150 Ib/£+3 4.77 r- | e = 00,10 — f — 3 ] & % r | ;.__ s v MODERATION BY ALL CONSTITUENTS | ! = > L — @ 008 — MODERATION BY HYDROGEN ONLY | 2 o — | X — - I— L - 0,06 — ] = Z - - _ =] o W 004 |— . . tad o L. e 2 s e s o DWG. 12905 : SECRET - 2 7 T : : = X P N - ~ < 070 — o - — - 3 ANP=PHY -9} ; ¢ [ 1 g © 060 (— | = = ! 2 — ———— MODERATION BY ALL CONSTITUENTS — = & 050 — ————— MODERATION BY HYDROGEN ONLY _ & 2 = i — - o o W N 040 . - & CYLINDRICAL CORE, TIAM. 25.4 ¢cm, HEIGHT 34.0 cm % O o — ATOMIC RATIO N, /N 235 = 52,9 g > © 030 |— =3 < T ’_‘ —— Led - L 0.20 — = D — x ul & 010 — 0 v 0032 — —_ 3 o L ] W (é}h o028 — ] S CYLINDRICAL CORE, DIAM. 25.4 cm, HEIGHT 22.4 om 5 | ATOMIC RATIO Ny /N ,235 =328.7 - o L Zz 0024 — — = o v - — o “ 0020 — ] Wy i - ~ S tad & = 0,016 — — > =i T - }_ __________ 1 ] g 2 ; FRACTION THERMAL ! — - i FISSIONS =(.813 { & wooolz — e 4 — = z — \ _ ™ - \ = 008 ] > o \\ a w3 : Y, L —] w & \ 4 b - g = 1 { . ! O O S T O T = I8 16 14 2 o o= LETHARGY, v . Fig. 3.28 - Fission Spectrum for the Reflected Water-Moderated Reazctor. Moderation by all constituents, LL IS61 -3 oc LEAKAGE PER UNIT LETHARGY PER 2.5 FISSION NEUTRONS, Vo o ~ < C60 0.50 0.40 0.30 G.20 0.10 O DWG. 12207 SECRET ool CYLINDRICAL CORE, D!AM. 25.4 ¢m, HEIGHT 22.4 cm L ATOMIC RATIO N /N je3s = 328.7 ANP-PHY-193 — Moderation by all constituents, — 18.6 THERMAL ] l —— ] = i 1 ] i | 8 i6 |14 12 10 8 6 4 2 0 L ETHARGY, u Fig. 3.29 - Leakage Spectrum for the Reflected Water-Moderated Reactor. JHOJHY SSAU90Hd XA'MILYVNO LDAT0Ud ANV FOR PERIOD ENDING SEPTEMBER 10, 1951 4. CRITICAL EXPERIMENTS A. D. Callihan, Physics Division The Critical Experiment Group, which is responsible for the experi- mental investigation of preliminary reactor assemblies, has completed the investigation of one assembly and is now taking data on another. The first mock-up was that of the aircraft reactor, proposed by General Electric, which employs water as a neutron moderator and air as a heat-transfer medium. - Two modifications of this reactor with the same hydrogen-to- carbon-to-uranium atomic ratlo were considered, and the distribution of thermal neutrons throughout the com- ponents was compared. The second 1in- vestigation, still in progress, 1s a study of the power and flux dis- tributions through a graphite-uranium reactor with carbon-to-uranium atomic ratio of 990. ' CRITICAL ASSEMBLY OF AIR-WATER REACTOR The study of the proposed air-water- cvele aircraft reactor, initiated at the request of the General Electric Company and executed with the co- operation of some of their personnel, has been completed for the present. In this reactor metallic uranium of high U??® enrichment was inter- spersed among blocks of graphite which simulated the proposed uranium—silicon carbide core material, and a meth- acrylate plastic served as an adequate substitution for water. The plastic was in 1%-in.-thick horizoental layers between 4%-in.-thick layers of uranium- bearing graphite. The critical mass of the assembly was about 29 kg. Measurements of the thermal-neutron distribution through a unit cell showed a high concentration in the hydrogenous material. The thermal flux in the plastic was 1.8 times the average in the graphite and 3.6 times that at the surface of the uranium. The thickness of the individual uranium pieces was 0.01 1in.; this introduced some self-shielding {rom incident neutrons. In an experiment to evaluate the extent of this effect, a standard piece of uranium was re- placed by five pieces, each 0.002 ia. thick, with aluminum foil separating adjacent ones. In this technique recoiling fission fragments collect on the aluminum, and their activity is a measure of the fission rate in, es- sentially, the surface layer of the uranium adjacent to the aluminum. A comparison of the activities on the successive aluminum surfaces gives, relatively, the fission rates at 0.002-1in. intervals throughout a fuel piece. In the reactor under study the average fission rate throughout a 0.01-1in, fuel piece was 84% of that occurring at its surface. The observed large thermal-neutron concentration in the plastic implied that the moderator-fuel inhomogeneity was causing a significant increase 1in the uranium requirement. Conseguently, the plastic thickness was reduced to % in. and was placed between the graphite-uranium layers, maintaining the same H:C:U atomic ratio as before. because of the Tt was necessary, 79 dimensions of available materials, to alternate the ¥%- plastic layers between 3- and 1%-in.-thick graphite layers, thereby probably not achieving optimum fuel-moderator homogeneity. The critical mass of this modification of the original assembly was 18.5 kg, and a more uniform distribution of thermal neutrons throughout the com- ponents resulted. The data are reported more fully elsewhere. (1) CRITICAL ASSEMBLY OF GRAPHITE REACTOR The inaugural program for the Critical Mass lLaboratory was the study (1), Hunter, Report on Critical Experiments for a Water Moderator {CA- 2), G.E. ANP Report DC-51-9-11 (September, 1951) 80 of simple reactors of good geometry for cowmparison with theoretical pre- dictions. The first of these, de- scribed in the-twoe preceding ANP guarterly reports (ANP-60, ANP-65), was a bare cube having a beryllium moderator. Further simple reactor studies were postponed in order to carry out the air-water-cycle reactor investigation described in the above paragraphs., On completion of these studies a graphite-uranium reactor was assembled, consisting of the 0,01-1n, thick uranium metal disks separated by 4-1n.-thick blocks of graphite, the C:U atomic ratio being 990, It 1is a rectangular parallelepiped 45 by 45 by 44 1in. Insufficient materials were available to make the system critical without a reflector so 3 1in. of graphite surrounds four sides of the core. The loading 1s about 45 kg of uranium. Measurements of power and flux distributions are being made. . 5. BULK SHIELDING REACTOR J. L. Meem, Physics Division The mock-up of the unit shield has been completed with the water-reflected Bulk Shielding Reactor. Preliminary analyses of the data indicate that an ideal unit shield for a 200-megawatt reactor with a 3-ft right square cylindrical core will weigh 124,000 1b. The construction of the divided shield is nearly completed, and preparatory shadow shielding measurements are reported. The spectroscopic instru- ments required for the divided-shield measurement appear to be satisfactory. This reactor; formerly known as the "swimming pool" or the "Shield Testing Reactor," is now identified as the Bulk Shielding Reactor (BSR). REACTOR OPERATION All measurements during the past quarter, including those on the unit shield, have been made with the water- reflected reactor.('? Because of the urgency of the unit shield measure- ments, the check on the power of the reactor by measuring the heat rise of the water through the fuel elements has not been completed. Preliminary measurements using the gamma-ray scintillation spectrometer (see below) indicated that to measure the spectrum at the face of the reactor a complete set of cold fuel elements would be required. These cold fuel elements have been obtained. ' (I)J. L. Meem and E. B. Johnson, Determination of the Power of the Shielid-Testing Reactor. 'I. Neutron Flux Measurements in the Water-Reflected Reactor, ORNL-1027 (Aug. 13, 1951). MOCK-UP OF THE UNIT SHIELD During the past quarter the measure- ments on the mock-up of the unit shield have been completed and a report is being written. Preliminary analysis of the data indicates that an 1deal unit shield around a 3.8-ft-diameter spherical reactor (approximating a 3-ft right square cylinder) operating at a power of 200 megawatts will weigh about 124,000 1b. Final evaluation of this weight will be given in the above- mentioned report. The limits of error of all the measurements will be analyzed so that the weight for this mock-up can be definitely bracketed. Measurements were made of gamma-ray ionization, thermal-neutron flux, and fast-neutron dosage along the center- line beginning at the face of the shield and proceeding outward into 'the water. HRuns were made with different concentrations of borated water up to 0.4% boron by weight. With the borated water in the shield, it was observed that the neutron-flux distribution in the reactor was altered. Since this would change the fast-neutron leakage, neutron-flux measurements with gold foils were made throughout the reactor following the procedure in ORNL-1027. MOCK-UP OF THE DIVIDED SHIELD The construction of the divaided shield mock-up by General Electric 1is - 83 ANP PROJECT QUARTERLY PROGRESS REPORT nearing completion. A suggested pro- gram for the measurements and calcu- lations has been outlined. ?’ The development of the spectroscopic in- struments has progressed to the point where the measurements appear quite feasible. Using an underwater model of a multiple-crystal scintillation spectrometer, (®’ a preliminary gamma- ray spectrum at 130 cm from the water- reflected reactor has been obtained.® The spectrum 1s shown in Fig. 5.1. For details of the measurement see RBefer- ence 4, An additional experiment, pre- liminary to the divided-shield measure- ments, has been completed. (%) This was a shadow shield test in which two lead (Q)J. L. Meem and R. H. Ritchie, Suggested Program for Divided Shield Measurements and Calculation o{ Air Scattering, OBANL CF-51-8-7 (Aug. 1. 1951 (3)5. K. Bair, F. C. Maienschein, and W. B. Baker, Multiple Crystal Gamma-Ray SEectroscopy Using Nal-ThI Crystals, NEPA-1701 (Fekh. 3, 1951). (4)F. C. Maienschein and R. H. Ritchie, Pre- liminary Gamma-Ray Spectral Measurements at the to be Bulk Shielding Facility, ORNL CF report No. assigned. 84 slabs, each approximately 6 ft long, 5 ft wide, and 1% in. thick, were placed symmetrically along the center- line north of the reactor. The rear of the first lead slab was 17% 1in. from the reactor with 2 in. between the slabs. Vertical traverses were made behind the slabs with a gamma ionization chamber. The same traverses were repeated after removing the lead slabs. The gamma traverses behind the lead are shown in Fig. 5.2. A reporton the nuclear plate camera for fast-neutron spectroscopy has been issued.(®? For further details of spectroscopic instrument development work, see the Physics Division Quarterly Progress Report for the quarter ending in September. (S)H. E. Hungerford, Jr., Experiment 5 at the Bulk Shielding Facility — the Shadow Shield, ORNL CF-51-8-252 (Aug. 20, 1951). (6)J. .. Meem and E. B. Johnson, A Nuclear Plate Camera for Fast Neutron Spectroscopy at the Bulk Shielding Facility, ORNL-1046 (in declas- sification). FOR PERIOD ENDING SEPTEMBER 10, 1951 | : DWG 12468 5 \ . SEGRET 10° — | — - | ¥ COMPTON DATA 0’ ] o © PAIR DATA 0 o 2T8S . — ¥ \ X COMPTON DATA 30 1 g & ‘ o f : fi\@ + PAIR DATA 30 oL\ 238 0% — : - . - x\! \ 5 — s -\ 233 ~ ~. I P ] s b X 645 v [ % | o 5 \ 1 ':,-, [ f 1 /A T ~ x \ . ~ o BT 2 3 oA 07 b \ % ] O o i T JiH] - L — 2 - I B ] o L— / i e [ o — s 1: - = . \r ~ e A - - IT L 5 qu o \t r 2107 |— \ = : L N <> - 2 N ¢ L + \T -] | “I\]’J — 1 ] - 4N 10 — | f POINTS WITH O COUNT RATE ] + | P T 7 v | | | | l l 0 2 4 6 8 10 12 14 GAMMA -RAY ENERGY (Ey, Mev) Fig. 5.1 - Preliminary Gamma -Ray Spectrum at 130 cm from the Water-Reflected BReactor. ‘ 85 ANP PROJECT QUARTERLY PROGRESS REPORT . SECRET DWG. 12253 R~ As,TrmRe 55 t— A 1IZONTAL DISTANGE FROM REAR OF SHADOW SHIELD (cm) HOR CURVE 1—-TRAVERSE S5cm FROM SHIELD CURVE 2-TRAVERSE |5em FROM SHIELD CURVE 3-TRAVERSE 35cm FROM SHIELD CURVE 4-TRAVERSE €0cin FROM SHIELD CURVE 4'-CURVE 4 PLOTTED TO SAME SCALE BASE AS OTHER CURVES q SGALE TO RIGHT OF EAGH GURVE GIVES 2 ¢ INTENSITY IN r/hr/wat! I o g 50cm LINE A-LINE CONNEGTING POINTS OF MINIMUM 17%in. S [ S s INTENSITY % L L5 em LINE B- LINE GONNECTING POINTS OF MAXIMUM S \ 5 cm . INTENSITY o SCALE DIACRAM OF SUADOW= | | INE G- SHADOW LINE OF SHIELD FROM ?rj A VERTICAL T;"IAVERSES WERE CENTER OF REAGTOR ) MADE AND THE RELATIGN OF LINES A,8, AND G TO REAGTOR G A /B A I R 0 20 40 60 30 100 120 14C 160 180 200 220 240 260 VERTICAL DISTANCE FROM GCENTERLINE (cm) Fig. 5.2 - Bulk Shielding Facility. A pictographic view of gamma radiation intensities existing along vertical traverses measured behind the shadow shield. 86 FOR PERIOD ENDING SEPTEMBER 10, 1951 6. LID TANK E. P. Blizard J. D. Flynn C. E. Clifford M. Marney T. V. Blosser R. Burnett T. Hubbard Physics Division The unit shield of iead and borated been reached in the unit shield,:and H,O kas been optimized for a 3i{t spherical reactor using double the boron concentration, now 1.3% boron: hy weight, of the shield presented ANP-53, (1) Preliminary Lalculatlgns now indicate that this results in: an ideal shield of 50.6 tons, represent- ing a weight saving of about 3 tons over the corresponding shield weight of 53.6 tons, using the data reported in ANP-53, Appendix B, corrected to a J-ft reactor. This new low weight value for a unit shield differs from that in the preceding section because of the difference in the core size as well as in the boron concentrat1on of the water, The saving in weight is primarily the result of the addition of boron which, by further reducing the number of high-energy capture gamma rays pro— duced in the shield, allows the lead to be placed at smaller radii for the same attenuation., It is doubtful that the optimum concentration of boron has ‘Vpepors of the ANP Shleldtng Board, NEPA- ORNL, ANP-53 (Oct. 16, 1950). therefore work with higher concen- trations will be continued using potassium metaborate. With this salt concentrations of up to 10% boron by weight can be investigated, The compositions of the present shield, the shield of ANP-~53, and the 1dealized shield for halt dose are presented in Fig. 6.1. The shield had a nearly constant B2l and therefore should be very close to. the minimum weright for this concentration of boron. (%’ Approximately ten con- figurations were measured before achieving the constant R%1 It is of 1nterest that an additional 6 fions in shield weight (i.e., 56.5 tons) will further attenuate both the gamma rays and neutrons by a factor of 2 (gémma rays from 1.10 to 0.55 x 10°° r/hr and neutrons from 3.6 to 1.8 X 10°* mrep). (2)g, P Blizard, ‘‘A Method for Experimental Shield Optxmlzatlon,’ Aircraft Nuclear Propulsion Project (QQuarterly Progress Report for Period fi;gé?g August 31, 1950, ORNL-858, p. 17 (Dec. 4, (B)Appendlx B of ANP-53,:op. cit. 0 oo {a) LEAD-0.6 % BORATED Hy0 SHIELD {FOOTNQTE 3} 3—-ft REACTOR WITH 6.6-¢tm STRUCTURAL Fe LID-TANK GAMMA DOSE=1.ix10" 5r/hr NEUTRON DOSE =3.6x10"* mrep SHIELD WEIGHT = 53.6 TONS {b) LEAD-4.3% BORATED Ho0—-PURE H,0 OPTIMIZED SHIELD 3—f: REACTOR WITH 6.6-cm STRUCTURAL Fe LiD-TANK GAMMA DOSE = 1.1x 102 r/hr NEUTRON DOSE = 3.6 x 10" mrep SHIELD WEIGHT = 50.6 TONS {c) LEAD-1.3% BORATED H»0-PURE H,0 OPTIMIZED SHIELD {HALF STANDARD DOSE} 3~f1 REACTOR WITH &.6-cm TRUCTURAL Fa L1D- TANX GAMMA DOSE = 0.55x 1072 t/hr MEUTRON DOSE = 1.8 x10°* mrep SHIELD WEIGHT = 56.5 TONS 2.54-cm Pb SLABS {8} SECRET DWG. 12909 ¢ 1.00-cm Pb SLAB B, O3 SAT. SOLN. - REACTOR —2.22-cm Fe SLABS{3) f / (457—cm RADIUS) N\ % 3 a4 & e {0.6 % BORON BY WEIGHT) 0 ® R NN N N T R AN S e OOO O OOOOO Oo Q \\k \“ : \Q s - e Oy 8, O’ ce A R N BN ARy N t. . ' OO OO O /"\.‘.' K%\ - AL - : . Wt 'OO' O O N R RN RN NN . ot 088 05880 NN AN N N - T O.OO ol NN oy b - : - ! ' . : | | o b 1 o M@ g o Mmoo 0 QN ~ ~ RB B T B £ a g & § & @ SCALE IN GENTIMETERS 0.3-cm Fe Fe TANK WALLS {(2)~— Fe TANK WALLS {2) TANK WALL 3.8-cm Pb om Fe TANK Wa REAGTOR 2.2-cm Fe SLABS / 7 SLABS i5) Kz 8904 SOLN. O-sem :URE .0 - (457 cm RADIUS) /g 3 3 / 5 {1.3 % BORON) J ‘2 AT Q' D e o e OOO . O Ll T L -l L S . Y ] °O O oy Se0 e e § SR LN L OO Oo" o 00 e IIRANT Y. N = ' 0?0 @00 0s AV RN SN o) o-o 895300 ¢ O AN < I AR N L | . P | | © e~ N W M o o ~ g 88 3 - e & 8 @ SCALE IN CENMTIMETERS _ 3.8-cm Pb SLABS(4) 0.3-cm Fe 2.2-cm Fe SLABS (3) j./ Fe TANK WALLS {2) 0.3-cm Fa TANK TANK WALL / @ Ko B. 0. SOLN WALL REAGTOR i 2 54-cm Pb Fe TANK , 259294 . (45.7-cm RADIUS) Wi /G)/\ SLABS WALLS ¢ {1.3% BORON BY WEIGHT) .~ PURE H,0 .0. o0 ¥ 7 N-"_".:'- E-. . . I ,_./ 38 020 e Pe NN Y N NN 20 ///‘ ] \\‘-..\‘.- E . - > . ST O Op oo° O OMANININ: R RN CO TN ; L 00 o O oo j %\ ‘ : s S ‘.' : s e 2 Ry i RN Ll O O Q OOOOO.O fi\%\ ‘.::-: i N ol J o i o o NN NG~ e o N 2 S ¢ 858 8 2w & 2 & 2 @ SCALE iN CENTIMETERS Fig. 6.1 - Comparison of Shields for 3-ft Reactor. LUO4AY SSTUO0Ud ATHALMVAO LDALOUd dNV FOR PERTIOD ENDING SEPTEMBER 10, 1951 7. DUCT TEST E. P. Blizard C. E. Clifford M. K. Hullaings M. Marney Physics Division A simplified design for the patch required to prevent excess neutron leakage from ducts perforating a divided (water) shield can be made on the basis of the duct test measurements completed to date. The weight of a patch needed for a single 6-in. aluminum- filled duct surrounded by a 1-in. air-filled annulus was found to be approximately 1000 lb exclusive of duct or container. ' The calculation for this patch weight was based on the data and con- figuration presented in Figs., 7.1 and 7.2. The duct 1s straight and extends beyond the shielda distance sufficient to attenuate the neutrons to the same level as 105 cm of water, which is given in ANP-53¢!) as the thickness of the divided shield on the rear of the reactor. The extension is surrounded by a cone of water which attenuates nentrons scattered radially to the appropriate level. This level is the intensity at the edge of the shield multiplied by the ratio of the area of the shield covered by the cone to the area of the cone. Some assumption had to be made with regard to the attenuation radially from the duct since experimental limitations prevented measurement of the thickness required for such a patch, The attenuation of water for the Lid Tank source was used as an (l}Be ort of the ANP Shzeldmg Board, NEPA- OINL ANP-53 (Oct 16, 1950). upper limit for extending these measure- ments. The water curve is normalized at Z = 0 in Fig. 7.1 to the curve of the attenuation along the duct axis. The ratioc of the heights of these two SECRET vo? DWE. 12910 [ [ ] ! | | i | | | | 5 o ATTENUATION 8Y BENT OucT o] A ATTENUATION BY STRAIGHT DUGT # NORMALIZED ATTENUATION IN Hy0 21 FOR 28-in. FISSION BGURCE ] 8 & NORMALIZED TRAVERSE, AT RIGHT 15% f-- ANGLE TO DUCT AXIS ] 51— ] _~EDBE OF DLCT 2 - ] 10 20 3D cm FROM DUGCT AXIS ol b ST -] RELAXATION LEKGTH = 12.1 cm - RELATIVE THERMAL NEUTRON FLUX ON DUCT tOcm FROM END OF DUCT 103 |— - 5 |~ - er (3505100 - 10% [— ] 5 —q 2 _ - 10 L4t 1 ol 0 20 40 &0 BO GO 20 140 160 118G 200 220 240 TOTAL LENGTH OF DUCT AXiS {cmi Fig. 7.1 - Neutron Attenuation in Water Around Duct. 6-in., aluminum- filled duct with 1-in. air-filled annulus. ' 89 ANP PROJECT QUARTERLY PROGRESS REPORT curves at the edge of the shield (105 cm), corrected by the ratio of the areas, 1s taken to be the radial attenuation required of the patch. The traverse measurements close to the endsof the ducts are approximately a7 5 CM =~ the same shape for all configurations measured, and this shape was fitted to the Lid Tank water curve in a region where the slopes were similar over the 30 cm measured. The measured traverses were then extrapolated to the desired attenuatiaon, SECRET OWG. 12911 WEIGHT OF PATCH = 960 Ib CENTERLINE OF DUCT PASSES THROUGH CENTER OF SPHERICAL REACTOR; PATCH CONSTRUCTED SYMMETRICALLY ABOUT CEN_IERLINE Y I SCHEDULE 20 STEEL 1-in. ANNULUS {AIR-FILLED) SCHEDULE 40 STEEL {Ocm A¢(SURFACE AREA OF TRUNCATED CONICAL PATCH)= 29,450 cm? A5 (AREA GOVERED BY PATGH ON SPHERE) = 16,570 cm? SPHERICAL SHELL OF SHIELDING SURROUNDING REAGTOR Fig. 7.2 - 90 Water Patch Around Duct. 'FOR PERIOD ENDING SEPTEMBER 19, Analysis of the data from shield 1951 8. SHIELDING CALCULATIONS being made. The data(l) for HzO, testing facilities has permitted the semiempirical establishment of im- portant radiation attenuation relations. An expression was derived for attenu- ation of neutrons from a point source in water by analysis of data taken at the Bulk Shielding Reactor. Similarly, analysis of Lid Tank data has yielded an approximate formula for the total centeriine gamma dose which has been applied to subsequent experiments, In the continued investigation of radiation hazards in the ARE, further calculations were made on the acti- vation of the Be(O moderator, the detection of leaks in the fuel elements by means: of radioactive tracers, and the activation of nitrogen and helium in the reactor pit. Calculations show the divided- shield weight to be higher than origi- nally estimated, owing chiefly to symmetry and angular distribution considerations, Some new effects have been computed relative to the shadow- shield theory which may prove that this device 1s much less useful than previously believed. A shield using ammonia in place of water has been considered which would provide a decrease in shield weight of about 7% at takeoff and 13%at landing. ANALYSIS OF BULK SHIELDING REACTDR NEUTRON DATA S. Podgor, Physics Division An analysis of fast-neutron dosimeter data of the Bulk Shielding Reactor 1is V(May 7 1951); Pb-H,0, and Fe-H,0 were treated by a method ‘similar to the one used by Welton®?) in the analysis of Lid Tank dosimeter data. A point source of neutrons was there calculated to be attenuated in water as follows: A » . P(S) = == &% o-BS e'zJa?S‘ 4nS*? where P(S) 1s the uncollided dose at a distance S from the source, A is a normalizing constant to be adjusted, and B8 is the oxygen “removal" cross- section in reciprocal centimeters which is assumed constant and to be adjusted. a comes from the flSSlon spectrum representation N(E) g e 0-75E where N{E) is the number of neutrons of energy E Mev per neutron emitted per Mev, and @ = 0.75. ¥ comes from the hydrogen cross - section repre- sentation 0.73 o(E) = 2738 - E + 1.66 where v ¥ 0.735. These two repre- sentations are taken from a paper by Blizard and Welton.(3) (1)3.:6. Cochran et al. to J. L. Mecm, Fast- Neuvtron Dosimeter Measurements, ORNL CF-51-5-61 OBNL CF-51-5-73 {(May 11, 195]1). Report Of the ANP Shteldtng Board, NEPA-ORNL, ANP 53, p. 120 (Oct. 16, 1950). ' (3)E P. Blizard and T. A. Welton, “ The Shle]d— ing of Mobile Remctors. I,)” Reactor Science and Technology (to be publlshed) 91 ANP PROJECT QUARTERLY PROGRESS REPORT The source was taken to be the leakage from the side of the reactor facing the detector. To simplify the analytical work the source is here assumed to be circular instead of rectangular. With these assumptions 1t was possible to fit the shape of the experimental attenuation curves by using removal cross-sections for lead and 1ron of 3.4 and 2.0 barns, re- spectively, which values are the same as those obtained in the analysis(*%) of Lid Tank thermal data. The oxygen cross-section neecded was 0.8 barn. The neutron build-up was assumed to be that for pure water, as given 1n pre- liminary form in an earlier quarterly report. (5) INTERPRETATIGN OF LID TARK GAMMA -RAY DATA A. Simon, Physics Division An approximate formula for the total centerline gamma dose 1in the Lid Tank has been applied to several experiments. The variation in gamma dose with the number of slabs in a uniform lead-water shield was re- produced satisfactorily for several positions of the detector. In addition, the variation of gamma dose with degree of boration of a uniform lead- water shield was also reproduced. Differences between experimental and theoretical normalization factors can be explained 1n terms of neglect of gamma build-up in water. (4)g, Podgor, Analysis of Lid Tank Neutron Data for Lead and Iron, OBNL-895 (Jan 23, 1951). (S)g, Podgor and T. A. Welton, ‘““Neutron Build- up in Water,” Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending March 10, 1951, ANP-60, p. 169 (June 19, 1951). 92 It is planned to apply the method to the computation of an iron-water shield and to observations 1in the BSF. A complete report will be pre- pared in the near future. SHIELDING CALCULATIONS FOR THE ARE W. K. Ergen, Physics Division Consideration of the radiation hazards associated with the ARE has led to the analytical evaluation of several such possible hazards. The instantaneous release of the nitrogen and helium in the reactor pit could possibly, but not probably, expose personnel at nearby sites to a dose equal to 20 min exposure to the accepted "maximum permissible concen- tration" of C'? in air.(® A means of detecting leaks in the ARE fuel elements by the use of radiocactive tracers has been proposed. A re-evaluation of the activities of the impurities in BeO, based upon a later analysis of the impurities and a revised Co(n,7y) cross- section, still indicates no handling problem after shutdown. Activation of Nitrogen and Helium in the ARE Reactor Pit.¢’) It is pro- posed to fill the ARE resctor pit with nitrogen during operation as a pre- cautionary measure in the event of a sodium leak. Assuming certailn average operating conditions, about 1.3 curies of C" would be formed by the N*&,p)Cct? reaction. This activity, released 1in- stantanecusly through a 100-ft stack (G)K, Z. Morgan, Maximum Permissible Concen- centrations of Radioisotopes in Air, Water, and the Human Body, Subcommittee on Internal Dose of the National Committee on Radiation Protection, report to be published. (T)Abstracted from the report by W. K. Ergen, Activation of Nitrogen and Heliunm Reactor Pit, (July 13, in the ARE ORNL, Y-12 Site, report Y-F20-16 1951). FOR under extremely unfavorable and very rare meterclogical conditions, would give personnel at X-10 and the HRE site a dose equal to a 20-min exposure to the sccepted maximum permissible concentration of C'* parts in air, Twenty per million (by volume) of argon in the nitrogen would give the same biological dose as the C'*. The other probable impurities of nitrogen as well as the N'%(n,¥)N'® reaction, eld negligible If the pit were filled with pure well helium, we would get, by the Hed (n, p)H® reaction, 5.6 % 1072 times the biological dose computed above for nitrogen filling. activities. Activation of Impurities in Beo.(®) Since Memo Y-F20-11 was issued on this subject, a more accurate chemical analysis of the BeO for the ARE showed the presence of greater amounts of impurities; also, the cobalt activation cross-section has been found to be 34 barns, and not 22 barns, as assumed 1in ¥Y-#20-11 on the basis of then avail- able data. Nevertheless, the acti- vation of the iwmpurities in BeO wall not cause handling problems after the shutdown of the ARE, under the con- ditions specified. Betection of Leaks in the Fuel Elemenls by Means of Radigactive Tracers. (%) A gamma-emitting sub- stance added to the fuel elements 1in the ARE could be used as a means of detecting leaks of fuel into the coolant before the ARE is operated. In to just get a definite detection, 2ven order (B}Absti«cLed from the report by W. K. Frgen, Activation of Impurities 1n Be), ORNL, Y-12 S:ite, report Y-F20-14 (May 21, 1981). \g}fiastractfid from the report by W. K. Frp Detection of Leaks in the Fuel Elements by of Rudicactive Tracers, OBRNL, Y-12 Siie, Y-i20-15 (July 10, ]951}. DIl Heans report PERIOD ENDING SEPTEMBER 10, 1951 if any one fuel tube should break 1in a system of 500 fuel tubes, a totval gamma activity of about 10 mc quired. is re- How much additional radiation precaution would have to be assumed in using this method 1s not cervain, but the possibility of leak detection without bleedingoff coclant makes this method seem attractive. Investigation 1s also being made of a2 photofluori- metric method of leak detection. NDA DIVIDED-SHIELD STUDIES Nuclear Development Associates, Inc. An assumptions 1involved in calculations of the split shield. A sazmplified examination is being made of the geometry has been assumed thus far, with the reactor a point source sur- rounded by a spherical shield. The represented by a spherical cap of 20° half angle. The following are tentative conclusions gamma shadow shield 1is emerging out of the calculations: ANP-53 scatter- scatter- ing with penetration normal the crew compartment conservative. Jt appears likely that the scattered neutron dose 1s one-half or less of the figure there obtained,. 1. The assumptions made 1in in regard to neutron air ing (isotropic, elastic to sides) are 2. Gamma rays can be scattered 1in the plastic at the crew com- partment 1n such & manner as to increase greatly their subse- guent penetration through the fead. Gamma shielding by the plastic 1s hence sharply reduced. It may be advisable to move the tic back to the re plas actor 93 94 ANP PROJECT QUARTERLY PROGRESS REPORT shield, a change which will not affect the weight of the ANP-53 neutron shield. The first-scattered dose at the crew, with any appreciable amount of lead crew shielding, is nearly independent of initial source energy down to around 1 Mev. Below that it drops off very rapidly owing to the in- creased lead absorption. The SFAC calculations by the National Bureau of Standards indicate that the radiation leaving the reactor shield does not contain any extremely large proportion of its energy 1n low-energy photons. The current practace of assuming the source energy flux to be concentrated at 3 Mev thus does not lead to error. Scattering in the lead at the crew shield has been calculated using build-up factors computed by The Rand Corporation. The resulting increase in dose 1s a factor of 2.5 for 1.55 cm of lead and about 4 for 2.5 cm of lead. It has usually been assumed that in the zone not shielded by the shadow cap the gamma rays emerge radially from the reactor shield. However, i1f they deviate from the radial direction, the effect of the shadow is partly nullified. Thus, i1if the emerging rays are uniformly distributed in angle from the radial out to 30° zero thereafter, the advantage gained by the shadow shield drops (in a typical case) from around 16 to 2. This appears to be one of the most significant and effects uncovered so far. 6. Streaming of photons around the shadow cap so as to emerge 1in regions assumed to be covered has been examined. There 1s some indication, both theoretical and experimental, that this i1s not so Important as the deviations from radial discussed above, 7. In the absence of the effects mentioned in 1tems 5 and 6, the shadow shield is quite effective in decreasingthe first-scattered dose. However, it has little effect on the second and higher scatterings. Some preliminary calculations 1indicate that the second scattering can then be quite significant, USE OF NH, AS A SHIELDING MATERIAL (1) E. P. Blizard and H. L. F. Physics Division * % Wyld, ANP Division Enlund, J. H. An analysis has been completed of the weight savings realized by re- placing water in a reactor shield by anhydrous ammonia., Although the number of hydrogen atoms per cubic centimeter is about the same for ammonia and water, the ammonia density 1s less than 70% that of water. The shield weights for a 4-in.-diameter reactor delivering 200 megawatts were calcu- lated for a divided shield employing water as the hydrogenous material and for a similar shield with ammonia. The composite (ammonia) shield is about 7% (IO)Abstracted from the report b H. L. F. Enlund, and J. H. Anmonia as a Shielding Materzal (in preparation) cm loan from Gibbs and Cox, Inc. On loan frem Reaction Motors, P. Blizard, Yd The Use of ORNL CF report Inc. FOR lighter at takeof{ and, owing te the evaporation cof the boil-off ammonia, is some 13% ligheer on landing. The welght savings from use 6f ammonia thus appear to be appreciable, but whether this advantage would be offset by the obvious disadvantage of carrying a large volume of a hard-to-handle ma- terial is still a question. Mechanical Requirements of an Ammonia Shield. Anhydrous ammonia has a vapor pressure of 1 atm at -28°F and must therefore be refrigerated to take care of inleakage of heat from the atmosphere as well as that due to neutron and gamma radiation from the reactor. The mechanical requirements for maintaining an ammonia shield are refrigeration equipment, a pressure shell to maintain the liquid at 120°F in case the refrigeration fails, and insulation to reduce heat influx. Since the heat flux into the shield due to nuclear radiations from the reactor is quite large, a higher temperature inner shield layer of water, con- ventionally cooled, and lead 1is pro- vided. The lead in this inner shield is inserted as a "heat trap" to lower the heat flux into the ammonia. Approximately 1 ft of water 1s pro- vided to remove neutron heat as well as to reduce secondary gamma production in the lead heat trap. The heat pro- duced in the ammonia is calculated from the gamma flux 1t will receive. A section of this shield assembly is shown 1in Fig. 8.1. Calculational Procedure. The total heat energy (H, inBritish thermal units) entering the ammonia from the gamma heating of the iron and lead heat trap during a 25-hr flight may be expressed in terms of the thickness (T, in centimeters) of the lead {(assum- ing a 3-in. ircon pressure shell): PERIOD ENDING SEPTEMBER 10, 1951 SECRET DWG. 12812 A -3-in. STEEL SHELL 12 in, 20+n. LEAD SHELL l/8“'!1'\. STEEL '.-. < 1-in. INSULATION § i, _: CORE /a-in. STEEL —1.375 in 43 375-n. R 79.0-n.R.__ “H,0 ' Vgrin, STEELY J7 tin. INSULATION | ", NHs Fig. 8.1 - HZ{}-Pb-NHE . TypicaiCump@site Shield, H = 2,36 x 107 ¢-T7/0.95 Since 589 Btu may be removed by evap- orating 1 1b of ammonia, the required ANP PROJECT QUARTERLY PROGRESS REPORT "boil-off" of ammonia (B, in pounds) to dissipate the above heat is: B = 4,01 x 10% ¢~T7/0.95 The lead heat trap is assumed to replace some of the lead (for gamma shielding) on the rear crew shield disk. But since increasing the thick- ness of the heat trap increases the total weight and also decreases the amount of NH, required for boil-off refrigeration, the shield must be optimized with respect to weight for these two design variables., The thickness of ammonia neutron shield* was first calculated to be 35.6 1in., and the expression for the equivalent in the two lead When this in- formation is combined, the total weight of the ammonia shield, which for the purposes of this calculation includes the lead heat trap, the rear crew shield, and the boil-off ammonia, may be expressed in terms of the thickness of the lead heat trap. The weight of the shield is then minimized by taking the first derivative of the variable weights with respect to the thickness of the lead trap. The thickness of this trap becomes 1.875 in., and the thickness of the rear crew shield 3.36 1in. gamma attenuation shields is derived, Comparison of Shield Weighis Using NE, and H,0. A comparison of shield *In a divided shield the neutron shielding is most effectively accomplished by hydrogenous material at the reactor. 96 TABLE 8.1 Comparison of Shield Weights Using NH, and H20 3 H,0 (1b) NH; (1b) Water 79,800 6, 520 Ammonia 42,300 Lead heat trap 15,450 Crew disk 9, 400 8,730 Stee]l shells 6,200 8, 300 Insulation 600 Total (fixed) weight 95, 400 82,900 Ammonia boil-off radiation 5, 570 Thermal conduction 380 Savings, takeoff condition 6,550 1b Landing condition 12,500 1b weights as calculated above, with water in one case and anhydrous ammonia and associated equipment in the other, is given in Table 8.1. It will be noted that the composite shield 1is 6550 1b lighter at takeoff and, owing to the evaporation of the boil-off ammonia, it will be 12,500 1b lighter on landing. The appeal of the boil-off scheme for refrigeration lies in its sim- plicity and in 1its low weight for landing, the most critical period of the flight. 1In addition, 1t 1s possi- ble that the vapors boiled off could be burned in the engines for added thrust, On the other hand, 1t 1s entirely possible that the refrigeration could be accomplished by mechanical means with considerably less weight. The Air Force at Wright Field has agreed to investigate this possibilaty. FOR PERIOD ENDING SEPTEMBER 10, 1951 9. NUCLEAR Continued efforts on the 5-Mev Van de Graaff accelerator research program have satisfactorily succeeded in lowering the machine background, providing an energy calibration accu- rate to a few kilovolts, and improving the ease of operation below 1.5 Mev. Data taken from a neutron source with and without beryllium surrounding the source are being analyzed to de- termine the (n,2n) reaction in beryl- lium. ~ An in-pile lithium loop operating at 1000°F for one week yielded data on Bremsstrahlung caused by beta particles from the Li’(n,B8)Li® reaction. The average intensicvy of the Bremsstrah- lung was about 2 X 10° Mev/cc/sec with an average energy of only 0.06 Mev. THE 5'MEV VAN DE GRAAFF ACCELERATOR H. B. Willard, Physics Division Both the neutron and gamma-ray backgrounds of the 5-Mev Van de Graaff accelerator have been reduced by the extensive use of only tantalum in the vicinity of the beam. Additional tantalum slits have been provided during the past guarter in order to further collimate the proton beam and reduce these backgrounds to a satis- factory level. In addition, the X-ray intensity coming from the reverse electron current up the tubes has been found sufficiently low to permit unshielded operation of a Nal-Thl crystal counter in the target area. Gamma-ray yield curves resulting from the proton bombardment of fluorine MEASUREMENTS and beryllium have been obtained. These data provide a large number of energy calibration points in addition to new 1nformation regarding levels in the compound nuclei involved. By proper manipulation of the machine, 1t was found possible to cbtain data for proton bombarding energies as low as 0.158 Mev. In preparation for neutron gross- section work, the Li7(p,n)Be’ yield in the forward direction has been extended to above 5 Mev. The response of the Bonner type neutron counter has been measured. Neutron yvield curves for protons on boron and beryllium have been obtained from threshold to above 5 Mev. MEASUREMENT OF THE (n,2n) REACTION IN BERYLLIUM ' E. D. Klema, Physics Division David Martin and David Ott, Oak Ridge School of Reactor Technology An experiment has been carried out in which the neutrons from an {(a,n) source are detected by means of aglong counter. The neutron source is sur- rounded by a sphere of beryllium, and the counting rates with bare source and source surrounded by the sphere are measured. Similar measurements are made with a carbon sphere to estimate the neutron absorption in the bervyl- lium sphere. Measurements have been carried out with polonium-boron and pelonium-~ beryllium reutren sources. Beryllium spheres of 4 and 9 in. diameter and carbon spheres of 4 and 5% in. diameter have been measured. 97 ANP PROJECT QUARTERLY PROGRESS REPORT The data obtained are being reduced to the (n,2n) cross-sections for the neutron spectra used, and the as- sumptions made 1n the calculations are being considered in detail in order to set limits on the values of the average (n,2n) cross-sections for the two neuntron spectra. BREMSSTRAHLUNG FROM Lis C. D. Baumann W. W. Parkinson R. M. Carroll 0. Sisman Chester Ellas Physics of Solids Institute Lithium at 1000°F has been circu- lated in the X-10 reactor for one week in a 316 stainless steel loop and sub- sequently examined. The primary purpose of this experiment 1s to study the Bremsstrahlung activity produced by beta decay of lithium. Li?, created by the reaction Li7(n,B)Li8%, decays with a 0.88-sec half-life, giving a 12.7-Mev beta ray. This beta ray 1is absorbed in the lithium stream and the container walls (1/16-1in. stainless steel ), producing the Bremsstrahlung activity. Preliminary analysis of the data indicates the average energy of the Bremsstrahlung to be about (.06 Mev with an average intensity of 2 x 10° Mev/cc/sec. Actual operation of the loop 1s discussed in this report in the section on radiation damage, where it may be noted that there was no significant amount of corrosion,. Radiation Detection Equipment. The 1on chambers used 1n this experiment were cylindrical in shape (4 1in. in diameter, 4 in. high) with one of the bases a thin-walled aluminum "window." The chamber was filled with air at atmospheric pressure, and its walls were maintained approximately 150 volts above ground. The chambers and their amplifying and recording equip- ment were bench-tested with Co®? of various strengths so as to simulate the line sources of the circulating lithium in the external pile loop. A lead brick with a 2-1n.,-diameter hole was placed between the source and the counter to act as a collimator for radiation coming into the counter. The in situ calibration curves were approximately linear and of the form S = AR - b where S 1s the source strength per unit length of pipe, R 1is the recorder reading (the amplified chamber response), A i1s a constant inversely proportional to the distance of the source from the counter, and b is the background correction which varies directly as pile power level, the distance of the and includes the natural background. The data have nct yet been corrected for the ef- ficiency of the counter for radiations of this energy range, and the neutron flux 1s still in the process of being determined. inversely as counter from the pile, Bremsstrahlumng Activity. The Bremsstrahlung activity, uncorrected for counter efficiency and neutron flux, at the three ionization chambers (or three different distances from the active lattice) for various velocities of the lithium stream is shown in Fig, 9.1. By extrapolating back to infinite velocity we find the Bremsstrahlung activity for zero decay to be 2 x 10° Mev/sec per cubic centimeter of natural lithium (at a flux of about 2 x 10'! neutrons/cm?/sec), Figure 9.2 shows absorption curves for the Bremsstrah- lung using iron and copper absorbers. The variation 1n activity with these copper and 1ron absorbers, between the tube containing the circu- lating interspersed lithium and No. 2 counter, FOR UNCLASSIFIED PSl—A~208 s DWG. 12178RI 3x10 9 10 o Chamber No. | X Chamber No. 2 v Chamber No.3 BREMSSTRAHLUNG ACTIVITY [ Mev/ {ceXsecl | 5 wm 0 0.l 02 03 0.4 1/ VELOCITY (sec/ f1) Fig. 9.1 - Bremsstrahlung from In- Pile Lithium Loop. gives the half-thickness of copper to be 0.025 in. and that of the iron 0.030 1n. Using these figures, in conjunction with the mass absorption PERIOD ENDING SEPTEMBER 10, 1951 UNCLASSIFIED PSl-A—209 Lo DWG ~— 12IT9RI 1, 0 - o Copper Absorber, Velocity 5.533 fi/sec x: Copper Absorber, Velocity 4.98 ft/sec Velocity 3.38 fi/sec v iron Absorber, 0 ©o002 004 Q.06 008 0.10 THICKNESS OF ABSORBER (in) Fig._9.2 - Absorption of Bremsstrah- lung from Li® Beta Rays. coefficient for X-rays!’? and ignoring the scattering cross-section, the average energy of the Bremsstrahlung was found to be 0.063 and 0.058 Mev, respectively, Thus the average energy of the Bremsstrahlung arising from the circulation of lithium in type 316 stainless steel tubes 1is of the order of 0.06 Mev. The radiation is identi- fied as that from Li® since it shows the 0.88-sec half-life by its decay while moving from one chamber to the next. During the week of operation of the loop, the velocity of the lithium stream was maintained most of the time at approximately 3 ft/sec in the external loop (approximately 1% ft/sec in the 1n-pile loop because of the larger diameter tubing). The velocity was changed only during the periods when runs of activity vs. velocity {1)W. S. Snyder and J. L. Powell, Absorption of ¥ Rays, 0H¥L~4§1, Supplement 3 (March 14, 19501, 99 were made and when the above absorption curves were determined. The velocity increased for the absorption measurements 1n order to increase the starting activity. The curves of activity vs. velocity were taken periodically during the entire week of operation, but, because no significant change occurred, 1s presented. wa s only one set of curves Since the intercepts and slopes of these curves showed no significant variation with time, 1t may be concluded that there was no appreciable corrosion of the tube by the circulating lithium. That there was no significant amount of corrosion 1s further shown by the counting the results of Ilithium from a 9-1n. section of the tubes adjacent to the active pile lattice. The total count as obtained 1n a "100% geometry" counter was only twice background. TIME-OF-FLIGHT NEUTRON SPECTROMETER G. S. Pawlacki, Oak Ridge Institute Of Nuclear Studies F. C. Smith, Physics Division A time-of-~-flight neutron spec- trometer for operation 1n the energy 100 range up to several thousand electron volts is being constructed. ?? Final assembly has been delayed by diffi- culties associated with the fabri- i.e., brazing, of the rotor. Preparations for the installation of the spectrometer at the LITR will be completed early in October. cation, Serious delays occurred 1n the fabrication of the rotor because of a brazing operation 1in assembly. The use of copper brazing was abandoned in favor of silver-copper eutectic alloy. The brazing was successfully completed i1in an 1mprovised hydrogen- atmosphere furnace. The brazed sub- assembly 1s being finished-ground by an outside contractor who previously ground the shroud ring for the rotor. When the subassembly is returned, the rotor will be assembled and balanced at the laboratory. The present esti- mate of completion date 1is the end of October 1951. (Q)G. S. Pawlicki and E, C, Smith, "Time-of- Flight Spectrometer,”” Aircraft Nuclear Propulsion Project Quarterly Progress Report for Pertiod Ending June 10, 1951). 1951, ANP-65, p. 129 (Sept. 13, 16. CORROSION RESEARCH ‘W. D, Manly, Metallurgy Division H. W. Savage, ANP Division "W. BR. Grimes and F. Kertesz, Materials Chemistry Division The corrosion program has been ex- panded to include corrosion by hydrox- 1de and fluoride coolants as well as the corrosion by liguid metals and fluoride fuels already under investi- gation for some time. Although the majority of the corrosion data are from static tests, a moderate number of data on liquid metals are available for thermal-convection loops. Forced- convection loops are being developed, but this development has not yielded many significant corrosion data as yet. In general, corrosion of incornel and stainless steel by pretreated and carefully handled fuel mixtures in the absence of radiation is not sufficiently severe to cause failure in operation for several hundred hours at 800°C. Limited experience at higher tempera- tures suggests that corroesion at 1000°C will not be excessive. Platinum, contemplated as thermocouple shield material, 1s not appreciably attacked by the fuel mixture contained in:a platinum or an inconel capsule. Welded or brazed joints of platinum to inconel are attacked by the molten eutectic. Experience with fluoride coolants indicates that corrosion, at least 1in the absence of radiation, will not be a serious problem with either inconel or the stainless steels. Corrosion problems encountered with hydroxides and with hydroxide-bearing materials, however, are a great deal more serious. Sodium hydroxide, eather commercial or specially pure material, 1s extremely corrosive to stainless steel or inconel. Copper, monel, and nickel are much more resistant although mass transfer occurs with these ma- terials. The corrosion by commercial potassium hydroxide which has been freed from excess water 1s noticeably less severe than that by sodium hydrox- ide but is sufficiently bad to preclude use of inconel and stainless steel at the preéent. Some tests with barium and strontium hydroxides conducted with dehydrated commercial material have that these materials are less corrosive than the dalkalies. These results are preliminary, and the lack of reproducibility in multiplicate tests 1s a serious problem, : indicated While the results obtained to date with hydroxide-bearing systems are not encouraging, it should be noted that (1) only in the case of sodium hydroxide has pure hydroxide been available for use; (2) inert, rather than reducing, atmospheres have been used: and (3) the metal walls have been in the as-received condition, While it is not obvious that changes in technigue will result in improved corrosion resistance, 1t 1s suggested that many variables remain to be tested before 1t can be stated that corrosion of stainless steel and inconel by the alkalies is necessarily so severe as to preclude their use. An extensive number of corrosion tests, both static and dynamiec, have conducted with sodium. It 1s now apparent as a result of this work that been 103 ANP PROJECT QUARTERLY PROGRESS REPORT sodium may be contained with little if any corrosion. Recent static corrosion tests with sodium at 1000°C in 316 stainless steel for periods up to 1000 hr revealed no surface reaction and did not indicate any significant weight change. Operation of thermal con- vection loops of inconel and several of the stainless steels with sodium for 1000 hr at 815°C produced very little corrosion,. Indeed, 1t now appears that little more can be learned of sodium corrosion by operating additional loops of inconel or stain- less steel with sodium. It remains to be demonstrated, however, that a circulatory failure will not occur with saodium in a forced-convection loop 1in which flow rates are higher and temper- ature changes greater. Metallographic examination of several lithium- and lead-containing thermal- convection loops, operated during the past two quarters, 1indicates severe corrosion of the hot leg and somewhat less severe corrosion of the cold leg. Only one of nine such loops completed the scheduled 1000-hr test; all but one of the other eight failed because of an obstruction in the cold leg of the loop. This obstruction is a build- up of corrosion products, STATIC CORROSION BY FLUORIDE FUELS evaluation of metals The experimental static corrosion of structural by fluoride fuel mixtures has all been conducted by the sealed-capsule technique described in previous re- ports.{173) This technique has been demonstrated by experience to be rapid (l)fl. E. Moore, G. J. Nessle, J. P. Blakely, and C. J. Barten, ‘“Low-Melting Fluoride Systems)” Airecraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending December 10, 1950, OBNL-919, p. 242 (Feb. 26, 1951). 104 and completely satisfactory for handling the fluorides, since mixtures thus treated have consistently demon- strated low corrosion rates on inconel and a wide variety of stainless steels. The penetration of inconel by a typical fuel mixture in 100 hr at 800 and 1000°C does not excezed 4 mils. The penetration of the stainless steels (304, 310, 316, 330, 347, 446) in 100 hr at 800 and 1000°C did not exceed 2 mils and was generally less than 1 mil., The corrosion of platinum, in- troduced in thermocouples, has been shown to be negligible in the fluorade fuel at these temperatures, The Pretreatment Process (G. J. Nessle, H. S. Powers, and C, J. Barton, Materials Chemistry Division). 1In a previous report(3?’? the improvement in corrosion which resulted {from treat- ment of the fluoride fuel with pieces of stainless steel and incomel at B850 to 900°C prior to filling of the cor- rosion capsules has been described. While no mechanism for this improve- ment has been demonstrated, it appears likely that this treatment removes traces of hydrogen fluoride which result from hydrolysis of some of the UF, in the fuel. A photograph of an inconel sample which was used to “"scavenge' the active 1mpurities 1in a fluoride bath is shown in Fig. 10.1. The fluoride thus freed from theactive impurities, and containing some 1iron, nickel, and chromium 1n solution, was used for making the static fluoride corrosion tests, (Dp, . Hagelston, “Static Corrosion by Fluoride Melts,” Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period %;gi?g March 10, 1951, ANP-60, p. 212 (June 19, (S)F. Kertesz, F. A. Knox, H. J. Buttram, S. D. Fulkerson, and J. A. Griffin, “Static Cor- rosion by Fluoride Melts on Metal Containers,” Aircraft Nuclear Propulsion Project Quaerterly Progress Report for Period Ending June 10, 1951, ANP-65, p. 148 (Sept. 13, 1951). FOR PERIOD ENDING SEPTEMBER 10, Fig. 16.1 - 1951 UNCLASSIFIED { v 4686 Inconel Specimen from Pretreating Pot in Which Fluoride Bath Mixture was "Deactivated" at 950°C for 100 hr. The smallest divisions offthe scale in the photograph are 1/100 in. A comparison test has shown that corrosion by untreated fluorides which had been vigorously outgassed and then sealed with a helium atmosphere enclosed was not appreciably different from that obtained when the capsules were sealed with an air atmosphere. : Pretreatments of the metal surface has not been successful in minimizing corrosion., Pickling, oxidation, and decarburization of the metal surface have seemed to result in increased corrosion., The oxide film on the surface as received seems to have reasonably good corrosion resistance properties, ‘ ‘ It has been found that some of the fuel samples, whether pretreated with structural metals or not, do contain appreciable quantities of uranium oxide. Sedimentation experiments in sealed capsules have shown an in- crease in uranium content in the lower section of the capsule without a cor- responding change in the fluoride - In content of the wvarious sections. Note very severe intergranular attack. addition, filtration of the molten fuel through sintered stainless steel has removed significant amounts of black uranium oxide. However, uranium oxide did not increase the corrosion rate. ' ' Pretreatment of the ligquid fuel with structural metals serves to introduce appreciable quantities of iron, chromium, and nickel 1nto the fuel melt., Chemical analysis of such fuel mixtures indicates the presence of 600 to 1200 ppm of irom, 200 to 1000 ppm of chromium, and 30 to 200 ppm of nickel after pretreatment although the compounds actually involved have not been identified. The variation in concentration observed indicates that not all these impurities are dissolved. This is borne out by the fact that some sedimentation of the iron seems to occur during the I00-hr corrosion tests, It also seems evident that the nickel content of the pretreated fuels is reduced during the subsequent cor- rosion testing while the chromium 105 ANP PROJECT QUARTERLY PROGGRESS REPORT content 1s not, Further studies are required, but 1t appears that con- siderable chromium is dissolved in the fuel and that the nickel present after pretreatment may be electrolytically deposited on the corrosion specimens at the expense of chromium from the test metal, The amounts involved are small enough to make demonstration of this phenomenon difficult. It seems likely, however, that pure chromium may have some advantages as the pre- treatment metal, Corrosion of Structural Metals (H. J. Buttram, N. V. Smith, C. R. Croft, and J. A, Griffin, Materials Chemistry Division; A. D. Brasunas, L. A. Abrams, and E. E. Hoffman, Metallurgy Division). It appears from studies to date that corrosion of inconel and stainless steel by care- fully handled fluoride fuels is not a serious factor. Furthermore, improve- ment of the pretreatment process and of handling techniques for the pre- treated material should result in further improvements. Three different fuel mixtures, the binary system NaF-UF, (75-25 mole %), the ternary system NaF-KF-UF, (46.5-26-27.5 mole %), and the ternary system NaF-Bel, -UF, (76-12-12 mole %), have each been tested at 800 and 1000°C with no significant difference in corrosion. In general, the corrosion of inconel specimens in 100 hr with each of these fuels was of the order of 2 to 4 mils; that of several stainless steels was around 1 to 2 mils. By NaF-UF,. Several static cor- rosion tests have been completed using the NaF-UF, fluoride mixture with a composition 75 mole % NaF and 25 mole % UF,. Tnconel and 347 stainless steel were exposed for 100 hr at 1000°C (1830°F) to the fluoride cutectic 106 which had been pretreated with these metals for 150 hr at 150°C, The attack by treated fluoride on inconel, in the form of subsurface was noted to a depth of 0.004 in. as shown in Fig. 10.2. The 347 stainless steel was less severely attacked as may be seen in Fig. 10.3, where the subsurface attack was limited to 0.002 in., The nature of the cor- rosion product has not been determined An analysis of the fluoride bath after testing gave the following results: volds, ) AMOUflT (%) IRON NICKEL CHROMIUM Inconel 0.25 Not detected 0.03 347 stainless 0.2 Not detected 0.05 steel (%) By NaF-KF-UFé. The results of corrosion testing of stainless steel and inconel with the NaF-KF-UF, fluoride fuel have been very satisfactory. The pretreatment technique has reduced the corrosion of these metals to mnearly negligible proportions with very little intergranular penetration of the specimens, The photomicrograph re- produced as Fig. 10.4 shows a heat- treated incoenel specimen and Fig. 10.5 shows the type of corrosion te be expected from the NaF-KF-UF, eutectic in 100-hr exposures at 800°C. Figure 10.6 shows a heat-treated 316 stain- less steel specimen, and Fig. 10.7 shows the corrosion in stainless steel after 100 hr exposure to NalF-KF-UF, eutectic at 800°C, The inconel speci- men shown, attack toadepth of 2 mils, is by no means the best specimen of this metal tested. The stainless steel specimen, attacked to a depth of 1 mil, is typical of the 304, 310, 316, 330, and 347 stainless steel samples tested. FOR PERIOD ENDING SEPTEMBER 10, 1951 Fig. 10.2 - Corrosion of Inconel by 3NaF-UF,. 250X. of exposure at 1000°C to 3NaF-UF,. fluoride attack to depth of 0.004 in. | Specimen after 100 hr Voids adjacent to surface represent E .40 N . ~ - # . . - « . 4 " - - “ ' L > = i LAY - R -‘.x e « - - . A - - & . “ o . o B cY L - . 5 . £ i ° - . L ! R <. ¥ ‘ RS v 5 L s - c. : - i ., ok, . . . .. o sy . 67 w - i LA e L s ff ’ - ! > . ~ Ty 5 -, . - ; . . . , . Ry o 3 : . s R ’ * < A, \ 3 A . . T, L e i - & 3 Y - y 7 w”"? Ty ~ : S T 4 AT i « L ; . . o - gy : 8 . - . / ¥ L 4 - - ¢ -~ Fig. 10.3 - Corrosion of 347 Stainless Steel by 3NaF-UF‘. of specimen after 100 hr of exposure at 1000°C to 3NaF-UF,. 250X, Surféce Note formation of voids to depth of 0.002 in. 107 ANP PROJECT QUARTERLY PROGRESS REPORT s T W T, .. L . _ 'Y ‘ & P W ® ey e - - -”*, S . a # oy . " . 9 & * Fig. 10.4 - Heat-Treated Inconel Specimen. 250X. T " f"-l' Lo A ' - Fig. 10.5 - Corrosion of Inconel by Pretveated Fluworide Fuecls. 250X, Specimen after 100 hr at 800°C in NaF-KF-UF, eutectic; slight attack, to depth of 2 mils. 108 FOR PERIOD ENDING SEPTEMBER 10, 1.951 P * e ,,‘ o / ’ ~ - "‘,/ - - / - . W - » o~ ¥ ’ - - ” \ - > P "~ r . * " CE - L T e / c _ o M 3 : - % @ Tt ;,‘"‘ | e Y Yo R . < - ’ - . « " P ot A & ] 3 » . oy [ {/ ‘\b % R - P L . - ek o . : x S " i - ¢ i . - - . . . Lot e N ol -~ ¥ e ¥ " 5 - - . - 5 o ~ st ¥ o % . . . i 5 . X 1 ¥ . - kT e a - _"’ [ A - o+ . - . Y 7 4 B e ha s . - x - : * ¥ e o - ! & ¥ o ' 4 e a * ' . st L ¢ A PO . " . w D ly,n - * \1 . £ . S i ¢ ';- { g " i f IS i » 1 . Y * f 4 . . ._,r a 4 Pos > fi # $ * . - “E . x - oy T - ~ L - - - 1 \"4\ : e - * \ * .. 4 LN 4 ¥ e . Fig. 10.6 - Heat-Treated 316 Fig. 10.7 - Corrosion of 316 Stainless Steel by Pretreated Fluoride Fuels. 250X. Specimen after 100 hr at 800°C in NaF-KF-UF, eutectic; no appreciable attack. 109 ANP PROJECT QUARTERLY PROGRESS REPORT In multiplicate experiments 1inm which individual capsules were removed at intervals for examination 1t has been shown that the major fraction of the corrosion observed occurs in the first few hours. Little additional effect 1s apparent after 1000 hr of exposure. By NaF-BeFQ-UFé. The ternary fluoride NaF-BeF,-UF, (76-12-12 mole %) was also used 1n corrosion testing. A 100-hr test with 310 stainless steel at 1500°F (815°C) lographic evidence of attack, although an analysis of the fluoride bath indi- cated 0.2% iron, 0.02% nickel, and no chromium to be present. The presence of nickel and absence of chromium 1is the reverse of the findings listed above for the binary fluoride mixture, NaF-UF4. However, these differences may be 1llusory because tests described below indicate that gravity segregation may exist in the container, and the composition of the fluoride mixture will depend on the location from which the test sample was obtained. showed no metal- A number of fluoride-corrosion tests were made on commercial alloys using a mixture of 10.7 mole % BeF,, 68.0 mole % NaF, and 21.3 mole % UF,. The materials tested are listed 1in Table 10.1 together with notes taken while the exposed specimens were being studied on the metallograph and results from bath analyses. The attack in all cases was not severe. Some decar- burization was observed, but this 1s not a highly objectionable type of surface instability. Weight losses are undoubtedly caused by solution of the metal sample into the fluoride bath as indicated by bath analyses which are also reported in Table 10.1. Weight gains, 1in spite of 110 some loss by solution, must be in- terpreted as absorption of one or more bath constituents (fluoride or metal) into the metal specimen. Comparison of results of analyses of the fluoride bath at various levels of the tube after test with results of the original fluoride analysis indicate a definite beryllium, a slight de- and very little, af decrease 1n crease 1n sodium, any, decrease 1n the uranium concen- tration. These changes 1in bath analysis may be directly associated with the weight gains observed on several test specimens. Corrosion of Piatinum (D. G, Hill, Consultant, Materials Chemistry Division). Thermocouples are to be inserted in a number of ARE fuel tubes to determine the center temperatures of these structures. Since platinum was suggested as the material for the tubular thermocouple wells, 1t has been necessary to examine the resist- ance of this metal and of some welded joints to the fluoride melt. Platinum slugs immersed in the raw or pretreated fuel, NaF-KF-UF4, in platinum capsules showed noappreciable corrosion as measured by weight changes. Microscopic examination of prepared specimens revealed occasional shallow pitting which may have been due to local impurities. 1In a similar test using liquid sodium metal instead of the fluoride fuel mixture, the platinum metal was completely consumed. These test data will be reported in detail later. At this time 1t appears that pure platinum will prove satisfactory although perhaps nickel plating of the material will be desirable. The brazed Joints must certainly be kept well above the liquid level. TABLE 16.1 Summary of Fluoride-Corrosion Data Obtained in 100 hr at 10600°C (1836°F) (10.7 mole % BeF,, 68 mole % NaF, 21.3 mole % UF,) BATH ANALYSIS AFTER TEST (ppm) | WEIGHT CHANGE MATERIAL TESTED METALLOGRAPHIC NOTES _ TOP BOTTOM {g/in.?) Nickel A Only evidence of attack was precipitation of fine | Ni 350 Ni 40 +0.002 partlcles along surface to depth of 0.002 1 304 stainless steel |Shallow surface voids observed to depth of <0.001 -0.005 : ste in. , , . , . 316 stainless steel Very few shallow surface voids observed to depth Fe 5,500 te 1, 500 0.000 of <¢.001 in. Ni 420 Ni 40 Cr 560 Cr 390 347 stainless steel |[Surface voids observed to depth of §.001 in.; no Fe 10,900 Fe 9, 800 -0.003 other visible evidence of attack Cr 2,400 Cr 1,300 S 310 stainless steel |DPecarburized to depth of 0.002 in.; few scattered -.003 voids in this area 330 stainless steel |Surface voids to depth of (.001 in. fe 4,500 Fe 7,200 -0.003 Ni 20 Ni 800 Cr 520 Cr 320 446 stainless steel |{Sigma-like precipitate throughout ferritic matrix | Fe 1,440 Fe 1,410 -0.9008 except at surface zone 0.00% . deep; few voids | Cr 3G Cr 190 observed to depth of 0.001 in. Hastelloy B Decarburized to depth of 0.005 in.; lightly Fe 5,800 Fe 1,200 +0.003 etched phase persistent throughout Ni 310 Ni 70 Mo <10 Mo NiO + 2Na + H, or 2Ni + 2NaOH —> 2Ni10 + 2Na + 1, Experimental evidence for these re- actions is found in the work of Villard who demonstrated the distillation of FOR sodium and of H, by heating a nickel boat containing sodium hydroxide in a Sodium and H,, quartz tube. oT some - times Nall, could be collected at the cooled end of the tube, the products depending on the conditions. Thus the second reaction shown should in- clude the possibility that the actual product is NaH, This work demonstrates that, although the free energy change is unfavorable to the reactionsg, reaction does occur so that, 1f a further process displaced the equi- librium, continued attack on the metal might take place. The further reaction may be the disproportionation reaction 2N10 = Ni + Ni0, Under the more common conditions of aqueous solution and strong acids, the equilibrium in this type of reaction is far to the left. Tt is considered possible that, in the extreme con- ditions of high temperature and very strongly basic solvent, the equi- librium 1s shifted to the right. Such reversals of equilibrium waith a change in conditions are well known in organic reactions, The scanty information now available on reactions in fused sodium hydroxide does not forbid the theory for this unexplored mediunm. : further reactions are certainly possible: NiQ, + 2NaOH = Na,NiO; + H,0 NiQ, + 2Na = NiD + Na,O The water produced in the first of these possibilities would react with metallic sodium, regenerating sodium PERIOD ENDING SEPTEMBER 10, 1951 hydroxide for further corresion, while the Na,0 produced in the second would be immediately available for reaction with more nickel. Thermodynamically the total process is only the transfer of nickel from impure wall metal to very pure nickel found in the deposit; such a process operates with a decrease of free energy even at a constant temperature. If the temperature varies from place to place or from time to time, the differing temperature coefficients of the several reactions might operate in such a way as to increase the rate of the process, One factor the experimental observation that NiO remains inert at the bottom of a platinum crucible under fused NaOH at 350°C, but at 750°C it 1s either suspended or more probably dissolved in the melt. If a melt of sodium hydroxide in platinum is treated with nickel at 750°C and the contents of the crucible are gquenched rapadly Lo room temperature, Expeyimental Evidence. worthy of note 1s the melt gives off a small amount of a gas on treat- ment with water. On analysis this gas proves ta be O,. This is very daffi- cult to explain on any basis other than that suggested here, since 1t 1s known that either Ni0O, or*Na,NiQO, with water to give oxygen. reacts The same type of reaction may apply to all metals that are known to exhibit mass transfer 1n sodium hydroxide. Fven the noble metals silver and copper should react toaminute extent to . give sodium, and only a very small amount may be needed to permit the dis- proportionation to operate., Work is in progress to study the reactions by several weans, particularly by polarc- graphic technigues and by potential measurements, 123 ANP PROJECT QUARTERLY PROGRESS REPORT DYNAMIC CORROSION TESTS IN THERMAL-CONVECTION LOOPS E. M. Lees, ANP Division J. L. Gregg, Consultant, ANP Division R. B. Day, Metallurgy Division During the period June 1 to August 31, 1951, the ANP Experimental Engi- neering Group has continued testing dynamic corrosion of liquid metals on various materials in conjunction with the Metallurgy Division. Tests of this nature conducted in thermal- convection loops afford a practical means of evaluating the usefulness of certain materials with liquid metals or other circulating coolant, supply fundamental information on the suita- bility of fabricating techniques, afford opportunities to test certain instruments under development, and provide experience from which adequate techniques for operating larger scale dynamic systems may be formulated. Metallographic examination of six lead-containing and three lithium- containing loops operated during the preceding quarter¢®’ have been com- pleted. Operation of all but one of these loops was prematurely (before 1000 hr) terminated, and all but one of these failures was due to internal obstruction in the cold zone of the loops. In general, the corrosive attack of both lead and lithium on the variety of steels and alloys tested 1s most severe in the hot zone of the loop with some attack usually apparent in the cold zone. The solid particles released to the circulating lead and Jithium streams are apparently "trapped” in the cold lead, accumulate, and eventually plug the loop. (5)"Dynamic Corrosion by Liquid Metals,” ANP-65, op. cit., p. 150. 124 During the past quarter, primarily sodium was tested in the thermal- convection loops. JIn no case was there a failure with a sodium-con- taining convection loop in which the failure was due to sodium corrosion. The successful operation of loops with sodium at 1500°F and preliminary re- ports of metallographic examination on loops sectioned after 1000 hr of operation indicate quite strongly that inconel and types 310 and 316 stain- less steel suffer very little cor- rosion by slowly flowing sodium at temperatures up to 1500°F. The successful operation of a 310 stainless steel loop fabricated from Y%-in.-o.d. tubing with a wall thickness of only 0.020 in. may be cited to indicate the relative inertness of stainless steel to sodium at 1500°F. Corrosion by Lithium and Lead. A number of thermal-convection loops which were operated with lead and lithium were examined after failure. The hot-leg temperature was approxi- mately 1500°F, although much higher temperatures were obtained momentarily when internal plugging obstructed the natural convective flow. Failure occasionally occurred at these higher temperatures, not from internal cor- rosion by the molten metal but rather from external causes, such as arcing or high-temperature air oxidation. As flow restriction, due to corrosion, was quite frequent with both lead and lithium, so were the i1ntervals of excessive temperature, and this probably contributed appreciably to the failure rate with these loops. The life of the various loops tested with lithium and lead, the compositions of the loops, and metallographic notes are given in Tables 10.3 and 10.4, respectively. The metallographic TABLE 16.3 corrosion and Operational Data on Lithium-Containing Thermal-Convection Loops LOOP TEMP. (°F) ANALYSIS OF LOAD AFTER TEST (%) HOT COLD i LIFE OF HOT ZONE COLD ZONE LOOP ZONE ZONE [LOOP (hr} REMARKS METALLOGRAPHIC EXAMINATION Cup cup No. 20, 1100 1020 34.5 Loop plugged (?) Intergranular attack observed Fe 0.0025 Fe 0.0075 1010 steel during operation; | in hot and cold zones te failed in hot depth of 0.020 in.; hot weld zone region attacked to depth of 0.040 in.; no metal crystals observed in cold zone - S = ~ No. 35, 1500 1375 280 Loop plugged (?) Intergranular attack noted to Fe 0.01 e Haynes internally; depth of 0.0608 in. in hot- Ni 0.05 3 Alloy failed in hot zone pipe, and to depth of Cr 0.75 g No. 25 zone 0.016 in. in the hot-zone Co 0.75 , weld; only 0.005 in. attack Man 0.02 % evident in cold-zone pipe, — none in cold-zone weld = < N o No. 38, 0.0 Loop failed at Some decarburization noted Fe 0.08 Fe 0.04 3 V-36 start of test Ni 0.0025 Ni 0.011 ; because heater Cr (.001 Cr 0.9624 o element arced Co <0.0025| Co 0.0035 = across hot-zone . e pipe &~ 1S61 o o T ot N N TABLE 10. 4 Corrosion and Operational Data on Lead-Containing Thermal-Convection Loobps LOOP TEMP. (°F) ANALYSIS OF LOAD AFTER TEST (%) HOT COLD LIFE OF HOT ZONE HOT ZONE COLD ZONE COLD ZONE LooP ZONE ZONE | LOOP (hr) REMARKS METALLOGRAPHIC EXAMINATIOM cup PIPE cup PIPE No. 19, 1010 1100 885 1000 Temperature gradient | Hot and cold zone tube walls showed very little Fe 0.0049] Fe 0.013) fe 0.0081 Fe 0.0009 steel in tube increased evidence of attack; metal crystuls formed in gradually, indicat- cold rones ing partial plugging No. 18, 1010 1400 62 Loop plugged in- Tube walls showed very little evidence of attack; Fe 0.39 Fe 0.017 fe 0.0079 stee] ternally during meta) crystals formed in cold zones {Fig 10.23) Cr 92.071 {(enamejed aperstion exterior) No. 32, 316 1500 1300 7 Loop plugged in- Intergranular attack to depth of 20 mils noted in Fe 0.057 Fe 8.8 Fe 0.C14 stainless ternsliy in cold hot zone; none noted in cold zone; metal crystals Ni 0.064 Ni 1.5 Ni 0.35 steel zone; faijure in observed in cold rone (Fig. 10.20) Cr 0.022 Cr 3.6 Cr D.026 hot zone due to ex- Mo 0.020 ternal causes Mn 0.01 Mo 0.13 Mn C.11 Ne. 26, 316 1500 1390 138 Loop plugged in- Hot zone much wore severely attacked than cold Fe 0.15 {Fe 0.04 Fe 4.6 Fe 11.2 stainless terpally during tone; metal crystals observed in cold zone Ni 0.05 |Ni 0.03 Ni 9.6 Ni 5.0 ateel operation Cr 0.001|Cr 0.001 Cr 1.39 Cr 0.03 Mo 0.012 { Mo 0.004 Mo 0.005 | Mo 0.78 G, 0.048 } O, 0.005 No. 28, 446 1500 1300 62 Loop plugged in- 0.00% in, intergranuler attack obaerved in hot Fe 0.03 |Fe 0.045 Fe 7.8 Fe 5.1 stainless ternally during sone; 0.002 in. observed in cold zone; wmetal Cr <0.001 jCr <0.001 Cr 2.92 Cr 4.64 stee] operation crystals observed in portions of coid leg (Figs. 10.21 and 10.22) Mo. 39, ¥-36 1500 1200 34 Loop plugged in- Genera! attack in hot zone only; cold rone showed Fe ¢.025 F= 0.0014 Fe 0.046 Alloy ternally during evidence of metal crystal formation {Fig. 10.19) Ni 0.21 iNi ¢.11 Ni 0.39 operation Cr 0.22 {Cr 9.022 Cr 0.81 Mn 0.004B{Mn 0.0018 Mn D.01T Co 0.52 |Co 0.08 Co 0.79 No. 3§, Haynes 1500 1290 699 Loop plugged in- Hot zone revealed more general attmck than cold Fe 0.0036 Fe 0.0083|Fe 0.016 Alloy No. 25 ternally; feiled zone; mets]l crystals observed in cold zone Ni 90.073 Ni 0.055 {Ni 0,067 in hot zone Cr 0.018 Cr 0.031 !Cr 0.030 Mn 0.0039 ¥Ma 0.0018iMm 0,0059 Co 0.14 Co 0.09 |[Co 0.15 1044 SSTYI0™d ATHALAVAO LDATOMd dINV FOR notes were made after examination of random samples of the hot leg, cold leg, and welded sections. The extent of corrosion was dependent upon the materials involved, the temperature, and the time of interaction. It should be kept in mind that the temperature was not actually controlled but merely attained by initially selecting a certain power input to obtain approxi- mately 1500°F at the beginning of the test, when convective flow was un- obstructed, When partial or complete plugging developed as a result of metal crystal formation in the cold leg, as shown 1in Fig. 10.19, heat losses were reduced and the hot-leg temperature was increased. This was corrected manually and the test was Fig. divisions on 10.19 - Metal Crystal Formation in Loop Containing Lead. the scale are 1/100 in. PERIOD ENDING SEPTEMBER 10, 1951 allowed to proceed until the loop was completely plugged or until it actually failed., A point of failure 1s shown in Fig. 10.20. Althongh it is accepted that the corrosive attack in the hot zone is appreciable, there is a general mis- conception that attack in the cold zone is negligible or even nonexistent, Figure 10.21 1llustrates an attack 1in the cold zone of a 446 stainless steel loop by lead which appears to be inter- dendritic 1in nature. Such attack releases solid particles to the circu- lating lead stream and may cause or accelerate plugging in this manner. A mass of metal crystals observed in the cold leg of this 446 stainless steel UNCLASSIFIED Y. 3945 ¢ B LS A The smallest {(z) Transverse and (b} longitudinal sections showing formation of metal crystals adjacent to cold leg wall of V-3¢ thermal-convection loop containing lead after 84 hr of operation. 127 UNCLASSIFIED ANP PROJECT QUARTERLY PROGRESS REPORT Fig. 10.20 - Failure in 316 Stainless Steel Loop Containing Lead. 3X, Hot leg failure after 7 hr of operation at 1500°F of 316 stainless steel thermal- convection loop containing lead. Note that failure resulted from attack on outer surface of tube wall, . - . - « W Fig. 106.21 - Thermal-Convection Loop Operated with Lead. Original photo- graph taken at 1000X, reduced 31% in reproduction. View of internal surface of 446 stainless steel cold leg from thermal-convection loop operated with lead (area at approximately 1300°F). Note that interdendritic attack releases solid particles to lead stream, UNCLASSIFIED 128 FOR PERIOD ENDING SEPTEMBER 10, 1951 loop is shown 1in Fig. 10.22. These crystals need not necessarily be bonded directly to the tube walls but may merely be trapped there. The opxide layer shown in Fig. 10.23 separates the metal crystals from the loop material, making bonding im- possible. ' An attempt was made to determine the composition of the metal crystals by quantitative analysis. Neglecting lead, the following elements were detected: | In the cold zone of a 316 stainless steel (18% Cr, 10% Ni, 3% Mo) loop after 138 hr of operation: 67% Fe, 6% Cr, 24% Ni, 3% Mo. In the cold zone of a 446 stainless steel (26% Cr) loop after 62 hr operation: 63% Fe, 37% Cr. The composition varied somewhat and these values are averages of analyses made in different areas of the cold zones. Corrosion by Sodium. Materials tested with sodium during the quarter included inconel and types 310, 316, and 347 stainless steel. FEighteen loops were taken out of service, 14 of which completed the scheduled 1000- hr tests with sodium at 1500°F. Three stainless steel (two 347, one 310) %-in. thin-walled tubing loops failed after less than 150 hr of operation owing to leaks in the welds, and one inconel loop failed after 350 hr owing to a plugged gas line 1n the liquid level control system; thus none of these failures are directly attributa- ble to the use of sodium. Included in the 14 loops completing 1000-hr tests was one 310 stainless steel thin- walled loop. On the three loops operating with sodium at the end of the period, one 1inconel loop was operated at 1600°F, and one each of inconel and 316 stainless steel at 1500°F. The increase in successful runs over the preceding guarter is due largely to (1) improved methods for purifying and introducing liquid metals into the systems (2) improved methods for degreasing, cleaning, and fabri- cation, and (3) improved methods for heating and temperature control. The sodium-containing loops are now being examined. Preliminary metal- lographic examination of some of these loops indicates very little corrosion, Small test specimens of similar and dissimilar composition were suspended in the hot and cold legs of these loops. Examination of these specimens has indicated that although weaght changes are small, the hot-zone specimens lost weight but the cold-zone specimens gained weilght, Tt is believed that little can be learned by operating additional thermal loops with sodium except for the pur- pose of testing a particular joint or component. Although it hasbeen proved that thermal-convection loops can be operated without plugging and without appreciable attack of the loop material by sodium, it remains to be proved that a forced-convection loop can be operated with sodium in one section at 1500°F and 1000 to 1200°F in another section without plugging. Therefore future tests in figure-eight forced- convection loops will determine whether there will be sufficient attack on inconel or stainless steel to result in plugging in addition to testing simulated components or flow channels. 129 Fig. 10.22 - Cold Zone o9f 445 Sizinless Steel Loop Showing Attacked Surface and Metal Crystals %hich Form in Lez# (Dark Areas). 150X. The crystals restrict convective flow. - ;“"n H; :'.. - . - fixf Fig. 10.23 - Cold Zone Weld in 1910 Sigsl Loop Showing Metal Crystal For- mation Adjacent to Pipe Wall Covered with Yroi Oxide, Presumably Formed During Loop Fabrication. 175X, 130 UWE FOR DYNAMIC CORBOSION TESTS IN FORCED CONVECTION LOOGPS W. B. McDonald, ANP Divisions Figure-eight sized loops are inter- mediate in size between the thermal- convection loops and full-scale ARE components and afford opportunities to test cleaning, handling, operating, and pumping techniques developed from smaller systems, in addition to supply- ing useful data on corrosion. Ex- perience gained from intermediate- sized systems 1s necessary for develop- ing adequate technigques and methods for assuring acceptable performance testing of full-scale ARE components. Operation of these loops to date, however, has afforded very little opportunity to observe corrosion effects. During the past quarter one figure- eight loop comnstructed from 347 stain- Jess steel was tested with a down-1flow centrifugal pump designed by the Experi- mental Engineering Group for circulat- ing sodium. The test was terminated PERIOD ENDING SEPTEMBER 10, 1951 after 7 hr of operation by failure of the pump shaft seal. The seal was repaired, but during the following start-up period external arcing from electrical heaters burned a hole in the main heater section of the loop. The damage caused by the resulting fire was sufficient to require com- plete dismantling of the loop. Total operating time for this experiment was insufficient to test revised cleaning procedures which i1ncluded repeated flushing with hot sodium until the oxide content remained low, In an effort to prevent repetition of loop failure being caused by auxiliary equipment in future tests, a General Electric electromagnetic pump has been tested in a separate loop to determine its reliability prior to its installation in the figure-eight loop. To prevent future failures due to external electrical arcing, specifi- cations have been drawn up for a 70,000-Btu/hr gas furnace for heating figure-eight loops, and these specifi- cations have been submitted to the Purchasing Department for bids or proposals. 131 ANP PROJECT QUARTERLY PROGRESS REPORT 11. PHYSICAL PROPERTIES AND HEAT-TRANSFER RESEARCH H. F. Poppendiek, Reactor Experimental Engineering Division The influence of the free-convection mechanism within liquid-fuel elements upon the fuel temperature structure 1is being investigated both experimentally and theoretically. Mock-up systems which simulate the actual fuel-element systems have been studied. Volume heat sources were produced by means of resistance heating, and temperatures were measured with thermocouples. An apparatus has been built which will be used to study the velocities of the free-convection cells. A theoretical analysis of the free-convection heat transfer existing in the fuel element has been completed for the laminar flow case; the turbulent flow analysis is currently being completed. These solutions are to be used to predict temperature distributions in liguid- fuel-element systems. Five Bunsen 1ice calorimeters for the measurement of heat capacity are now in operation., Experimental work and analyses have been completed for 316 stainless steel, synthetic sapphire, lithium, molybdenum, and zirconium 1in the range 300 to 1000°C, An improved thermal-conductivity apparatus (Deem type apparatus No, 2) for ligunids has been constructed, A value has been obtained with the equipment previously deseribed(!) {Deem type apparatus No. 1) on fuel salt mixture No. 2 (46.5 mole % NaF, 26.0 mole % KF, 27.5 mole % UF,) 1in (1)L. F. Basel, “'Thermal Conductivity of Liquids,' Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending June 10, 1951, ANP-65, p. 161 (Sept. 13, 1951). 132 the range 643 to 700°C as 0.53 Btu/hr-ft-°F., The longitudinal-flow apparatus for the determination of thermal conductivity of solids is in operation and additional conductivity measurements on various densities of boron carbide have been obtained. The radial -flow apparatus has been com- pleted. Two pieces of eguipment have now been constructed for the measurement of liquid densities. The density of the NaF-KF-UF, eutectic has been determined in the range 535 to 1000°C. The apparatus for the measurement of viscosities is scheduled for com- pletion by September. Work has been retarded by delays in fabrication of parts. An apparatus which will be used to determine the heat-transfer coef- ficients of fused hydroxides and salts has been nearly completed, These fluids are potential reactor coolants. Some preliminary heat-transfer data for a boiling-mercury system have been determined and the figure-eight lithium system has yielded some data. INVESTIGATION OF FBEE CONVECTION WITHIN LIQUID-FUEL ELEMENTS Theoretical Analysis of Natural Convection (D. C. Hamilton and H. F, Poppendiek, Reactor Experimental Engineering Division). An analytical solution for the temperature distri- bution i1n the parallel-plate system has been derived. 1In this system a FOR PERIOD ENDING fluid with a uniform heat source 1s in laminar flow between two parallel and infinitely long plane bounding sur- faces. In the system it was postulated that heat was generated uniformiy and that the long convection cells were in the laminar flow regime. A report based on this analysis 1s being pre- pared. A similar analysis wz2ll be made for a system in turbulent flow. These solutions will also be duplicated for cylindrical coordinate systems. The solution indicates that, for a system in which the distance between the parallel plates 1s 0.18 in., the axial temperature gradient is 100°F/ft and the physical properties are identi- cal te the measured values for the NaF-KF-UF, eutectic; the temperature difference which would exist 1f all the heat were transferred by con- duction is reduced by 25.5%. This result is based on a value of 1 centi- poise for the dynamic viscosity. If the viscosity is greater than 1 centipoise, the reduction in tem- perature difference will be smaller than 25.5%. Measurement of the Fuel-Element Temperature Distribution (F. E. Lynch and M. Tobias, Reactor Experimental Engineering Division). A new apparatus has been devised for measuring the effects of free convection on the temperature distribution in tubes filled with fluid in which heat 1is being generated by electrical means. Because of the fragility and com- plexity of previously constructed devices, considerable difficulty has been encountered in making a series of without interruptions due to A diagram of the present apparatus is shewn in Fig., 11.1. Thermocouples are arranged in the central tube as described in previous runs breakage.. SEPTEMBER 16, 1951 UNGLASSIFIED DWG 12938 POTENTIAL LEADS 9% in, CAN TEST SECTION o 4in, BELL Fig. 11.1 - Schematic Diagram of Free Convection Apparatus. reports,(z's) Aluminum wires arve sealed into curved tubes which are (2JF._E. Lynch and P. €. Zmola, “‘Study of Free Convection in Liquid-Fuel Elements,” ANP-65 op. c¢it., p. (3YF. E. Lynch and P. C. Zmola, “Study of Free Convection in Liquid-fuel Elements,’ Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending March 10, 1351, ANP-60, p. 232 (June 19, 1951). : » 133 ANP PROJECT QUARTERLY PROGRESS REPORT Joined to the ends of the straight test section and serve as potential measurement leads. The electric current which generates the heat in the central tube 1s led in by an electrode through the large curved tube at the top and passes along the test section and out through the bell shaped base. The entire apparatus is placed i1nside a stainless steel can which 1s filled with liguid to just below the top of the large curved tube. The central section 1s held vertical by a retaining clamp at the top and by a swall well at the base. The return electrode is clamped to the When used with brine, or of low freezing can 1itself, other substances point, the system is cooled by wrapping a coil of a swall thermostatically controlled refrigerating unit around the can., A stirring motor assists 1in maintaining a uniform temperature 1in the can. - Tests will be run using mercury and brine. Design of a new apparatus which will be used for the fuel salt itself is now being considered., The chief difficulty 1s finding a contailner material which will resist attack by the salt, 1s structurally strong, and is a nonconductor of electricity, Measurement of the Fuel-Element Velocity Distribution (R, Redmond, Reactor Experimental Engineering Division). An apparatus has been designed and constructed for the purpose of obtaining preliminary in- formation about the nature of the free convection patterns 1in the volume heat source system; temperature distri- butions are also to be measured. A flat plate cell which contains dilute acid 1s to be heated by electric current. The acid will have the generated heat removed at the walls of 134 the flat-plate cell by water cooling. Velocity patterns will be determined by tracing the paths of particles suspended 1n the solution, and the distribution will be thermocouple measure- temperature determined by ments. PHYSICAL PROPERTIES A, R, Frithsen, USAF M. Tobias, HReactor Experimental Engineering Division Heat Capacity (William D. Powers, Reactor Experimental Engineering Division). The heat capacities of type 316 stainless steel, syathetic sapphire, lithiuw, molybdenum, and zirconium have been determined. The values which follow give the heat capacity as a function of temperature (Cp is the heat capacity at T°C in calories per gram per degree centi- grade) in the indicated temperature range. TEMPERATURE Cp MATERTAL RANGE (°C) (cal/g/°C) Type 316 150 - 1000 {0.11 + 5.7 x 10”57 stainless steel Synthetic 300 - 900 |0.26 + 4.4 x 10°°T sapphire Lithium 950 - 1100 [1.0 + 2.8 x 10757 Molybdenum 200 - 1100 0.068 Zirconium 150 - 1050 [0.070 + 3.6 x 10757 FOR Heports are forthcoming on each of these materials. The substances now under investigation -are sodium hy- droxide, nickel, sodiuvm flmoride, petassium fluoride, uranium fluoride, potassium hydroxide, and fuel mixtures. Work will be started soon on lead, bismuth, and lead-bismuth eutectic, A modification of the original ice calorimeter has been made and included in the five calorimeters. The heat content of 2 sample is measured by noting the change in volume of a pure ice-water mixture. The mixture is contained in a lucite shell in which = hollow finned copper cyvlinder 1is sealed. The copper cylinder receives the sample at a known temperatnre,. Jece i1s melted in the lucite shell and, by noting the change 1in volume i1n the ice-water mixture, the heat content of the sample way be calculated. To prevent heat from getting into the system from any source other than the sample, the lucite shell was surrounded by ice. However, impurities in the surrounding i1ce lowered 1ts melting point encugh te cause noticeable freezing of the water in the lucite shell., The lucite shell was therefore surrounded by 2 in, of S:1-0-Cel insulation contained in a steel shell. The steel shell was kept at 0°C by surrounding 1t by 1ce. The heat leakage into the lucite shell was small. An appreciable error can occur in the measurement of the actual tempera- ture of the sample. At high tempera- tures, variations of 5 to 10°C have been found on the surface of capsules placed in the most uniform tempera- ture position in the furnace. It 1is felt that tube furnaces longer than those now in use will reduce thzs error. Manufacturers are being con- PERIOD ENDING SEPTEMBER 10, 1%51 sulted to see if furnaces of the desired characteristics can be fur- nished. Thermal Conductivity of Liguids (M. Tobias, A. R. Frithsen, and L. Basel, Reactor Experimental Engineering Division)., The thermal conductivity of the fused salt NaF-KF-U¥F, (com- position 46,5, 26, and 27.5 mole %, respectively)} has been measured using the OBNL Deem type apparatus,¢!? The average conductivity of four measure- ments was 0.53 Btu/br fo-°F at temper- atures ranging from 643 to T700°C. These data were urgently needed and no check runs were made with sodium. As several imperfections an the design of the apparatus have become apparent, measurements with this apparatus have been discontinued. A new apparatus, the ORNL Deem type No. 2, which has a bellows seal instead of the Pb-Bi seal used in apparatus No., 1, is now approximately 95% com- plete and is scheduled for cperation in September, A few check runs will be made on lead, after which data will be obtained on BeF -NaF-¥F salt mixtures. A hood is being installed over the apparatus to protect operators from escaping BeF, vapor. Thermal Conductivity of Solids (M. Tobias, Reactor Experimental EFngineering Division). The thermatl conductivity of copper has been measured nup to 600°C by the long:- tudinal flow method.{*? The data (Fig. 11.2) agree to within 4% wath the values given in the International Critical Tables. (4)y. Tobias, “Thermal Conductivity of Salids,” ANP-65, op. cit,, p. 162, 135 ANP PROJECT QUARTERLY PROGRESS REPORT UNCIASSIFIED 250 DWG, 12939R1 = Z 20 3 . [GT GOPPER DATA & b~ ] ! & . m""""'*-:-—-l_...f_h__“‘_&_ L . >, : . Tt = 400 o 3 0 2 a & U ~ 150 3 oL w = = 0 100 200 300 400 500 00 0 TEMPERATURE (°0) Fig. 11.2 - Thermal Conductivity of Copper, Improvement of the preliminary data on the thermal conductivity of boron carbide by the longitudinal flow method 1is continuing. In addition, experiments have been started using the radial flow apparatus to measure thermal conductivity. Falling-8all Viscometer (S. T. Kaplan, Heactor Experimental Engineer- ing Division). The falling-ball viscometer 1s now being checked out with water. High-temperature checks with bismuth are contemplated. Delivery of the electronic eguipment to measure the rate of fall of the radiocactive ball was overdue and has only recently been placed i1n satisfactory operation. The dropping valve of the apparatus(®? has been modified to accommodate balls up to % in. diameter. Tantalum-plugged stainless steel balls, 0.25 1.d. with a densityof 12.2 g/cm?® at room tempera- ture, as well as 0.20-1n.-diameter cobalt-coated pyrex balls with a room- temperature density of 2.6 g/cm’® are (SL‘Viscosity—Measuring Tube,” Atrcraft Nuclear Propulsion Project Quarterly Progress Report for Pertiod Ending December 10, 1950, ORNL-919, Fig. 7.8, p. 200 (Feb. 26, 1951). 136 Additional balls of various densitlies are prepared, presently available. test being Brookfield Viscometer (F. A. Knox and I, Kertesz, Materials Chemistry Division). Some approximate values for viscosity of typical fuel specimens and for some fluoride mixtures of interest as coolants have been obtained by a wmodified Brookfield Synchro- Lectric viscometer fitted with a special spindle., The 1nstrument was and at Toom calibrated 1n distilled water water-glycerol solutions temperature and in molten sodinum chloride at 830 to 1000°C. The viscosity of the NaF-KF-UF, eutectic was found to be about 32 to 35 centipoises at 600°C and 12.5 to 16 The NaF-BeF,-XI under consideration as a centipoises at 1000°C. eutectic, coolant, Since these samples also showed high values. both showed con- siderable guantities of suspended {uranium oxide in the fuel and beryllium oxide in the coolant}, these values may be greatly in error. matter The viscosity of the NaF-KF-1aF eutectic mixture was 1n the expected low range of 2.2 centipoises at 500°C and 1.5 centipoises at 800°C. Brookfield Engineering laboratories are designing a high-temperature rotational viscometer for this work. Preliminary designs were submitted by them for inspection, and modifications are now under study. Vvapor Pressure (R. E. Moore and C. J. Barton, Materials Chemistry Division}). The temperature of the fluoride fuel at the center of the ARE fuel pins i1s still a subject of some uncertainty but will almost certainly FOR be higher than 2000°F. While the components of the ligquid fuel are thermally stable to temperatures well above this figure, determine 1t 1s necessary to the vapor pressure of suitable fuel mixtures in this temper- ature range, Preliminary results indicate that the vapor pressure of the KF-NaF-UF, eutectic 1s less than 3 mm at 1000°C. The apparatus and procedure employed are those reported by Rodebush and Dixon¢®? and moditied by Fieck and Rodebush, ¢(7) The prepared fuel mixture 1s placed in a cylindrical vessel of stainless steel which may be heated in a resistance furnace, The vessel i1s connected by one tube to a manometer and by a second tube to an intermittent pump and toe a The two tubes outside the furnace, source of inert gas. are connected, by a differential manometer containing a light liguad. With the container and sample maintained at the desired temperature, the inert gas pressure 1s reduced 1n small intervals by the pump. The residual inert gas pressure 1s equiva- lent to the vapor pressure of the sample at the highest pressure suf- ficient tomaintain a permanent pressure difference in the differential ma- nometer. This action depends on the fact that when the i1nert gas pressure is equal to or less than the vapor pressure of the ligquid, diffusion of the inert gas 1s the only mechanism for equalization of the pressure 1in (6)W. H. Rodebush and A. L. Dixon, ‘“The Vapor Pressures of Metals; a New Experimental Method,” Phys. Rev. 26, 851 (1925). (7)E. F. Fiock and W. H. Rodebush, “*The Vapor Pressures and Thermal Properties of Potassium and Some Alkali Halades,” J. Am. Chem., Soc. 48, 2522 (1926). ' PERIOD ENDING SEPTEMBER 10, 1951 the two tubes. TIf the tubes are of sultable diameter this diffusion is very slow. Considerable difficulty has been encountered 1n mailntaining temperature control in the resistance furnaces at the high temperatures required. It is certain, however, that the total vapor pressure of the NaF-KF-UF, eutectic mixture 1s well below that of pure UF, at temperatures in the range 0 to 1050°C. These experiments are being continued at present. Pensity (S. I. Kaplan, Reactor Experimental Engineering Division). Following test runson sodium at 200°C, a determination of the density of NaF-KF-UF, eutectic was made over the range 535 to 1000°C (46.5, 27, 26.5 %, respectively) using the buoyancy apparatus described in a previous report.(®> The data can be represented by the equation mo le o = 4.702 - 0.00115T, where £ is the density in grams per cubic centimeter and T is the tempera- ture in degrees centigrade. The test material was blanketed in an atmosphere of argon purified by bubbling through NaK. Determination of the density of various coolant and fuel salt mixtures is projected for the next quarter, HEAT-TRANSFER COEFFICIENTS Bealt Transfer in Fused Hydroxides and Salts (H. W, Hoffman, Reactor (8)5. I. Kaplan, “Density of Liguids,’ ANP-60, op. cit., p. 246. ' 137 ANP DIVISION QUARTERLY PROGRESS REPORT Experimental Engineering Division). The construction of the apparatus for the determination of the heat-transfer coefficients of fused hydroxides and salts 1s approximately 80% complete. Operation of the system 1s anticipated sometime within the next month. The first material to be tested in the apparatus will be the eutectic mixture of NaF-KF-LiF having a melting point of 454°C, Future fluids to be tested include NaOH and the eutectic mixture of Ba(OH)z-Sr(OH)Q. A schematic diagram of the experi- mental system 1s given in Fig. 11.3. The charge is first melted under vacuum in the melt tank and then moved by applying argon gas pressure to tank A. In an experimental run the molten fluid 1s pushed by gas pressure through the heat exchanger and i1nto tank B. Tank B rests on a scale, and cumulative welght readings are made to determine the fluid flow rate, Temperatures of the outside surface of the heat ex- changer are obtained by means of chromel-alumel thermocouples spark- welded to the tube wall. An electric current 1s passed through the wall of the heat exchanger tube, and the heat input 1s determined from the measure- ment of this current and the voltage drop along the heat exchanger. The inlet and outlet temperatures of the test fluid are measured by thermo- couples on the tube wall in the iso- thermal regions immediately before and after the test section. The 1nside wall temperature is calculated from the outside wall thermocouple readings. At the conclusion of a run, lasting approximately 30 min, the fluid 1is pushed through the return line 1nte tank A, The system 1s then ready for another experimental run. Temperature level and flow rate 1n the system can be varied within reasonable limits, Heat Transfer in Boiling-Ligquid- Metal Systems (W, 5. Farmer, Reactor Experimental Engineering Division). The boiling-heat-transfer investi- gations have been continued during this period on the boiling-mercury experiment using a flat plate and on the boiling-sodium and -potassium experiments using a horizontal tube. only were boiling-mercury Preliminary results obtained for the experiment owing to condenser failure in the high-frequency induction heater employed as a heat source. However, using pure mercury, temperature dif- ferences of approximately 100°F were observed between the metal surface and the boiling fluid for a heat flow rate of about 50,000 Btu/hr ft2?, The large temperature difference observed in this case was probably due to non- wetting of the heat-transfer surface by the boiling mercury. In addition, considerable temperature fluctuations were observed in the boiling liguid, accompanied by bumping. The apparatus has been moved to a new location to utilize another induction heater, and further results will be forthcoming. UNGLASSIFIED CWG. 12940 FILTER . HEAT EXCHANGER | {1 TANK + {5 ™ POWER SUPPLY RETURN LINE Fig. 11.3 - Schematic Diagram for Determining Heat-Transfer Coefficients. FOR PERIOD ENDING SEPTEMBER 10, of the horizontal tube boiler has been accelerated during this past quarter. The erection and fabrication of all tanks and piping are complete, Only the duct work and electrical construction remain to be done. This latter work should be completed and preliminary sodium and potassium runs should be possible this coming quarterly period, The construction Heat Transfer im Molten Lithium (C. P. Coughlen and H. C. Claiborne, Reactor Experimental Engineering Division)., During the past quarter the figure-eight lithium system was placed in operation, and preliminary heat-transfer data were obtained. To remedy pump and instrumentation dif- ficulties whichwere encountered during operations, extensive alterations have been made and a new Einstein-Szilard pump has been obtained on loan from Allis-Chalmers Manufacturing Company. The new pump will permit operation over a much larger range of velocities in the test heat exchanger. The experimental results obtained with a concentric-pipe heat exchanger are preliminary because of the 1im- perfect operation of the auxiliary squipment which prevented the attain- ment of complete equilibrium and because of the low flow rates obtainable (1140 to 1600 lb/hr in the annulus and 980 to 1150 1b/hr in the core pipe). 1951 Separation of the i1ndividual heat- transfer coefficients was not possible under these conditions. The overall heat-transfer co- efficients determined experimentally ranged from 4100 to 5160 Beu/hr-ft?-°F, Heat balances checked to within 24% with an average deviation of 12.5%. The major errors in the heat-balance data undoubtedly resulted from in- complete equilibrium and the un- certainty itn the flow measurements. A comparison of the Nusselt modulus computed from the experimental data with that of Lyon(?) and that of Isakoff and Drew('?) is shown in Fig. 11.4. The experimental values are 16 to 80% higher than those found by Isakoff and Drew when using mercury for an L/D = 38. Saince L/D % 30 for the lithium heat exchanger, the Nusselt modulil are expected to be somewhat higher than those for a system with L/D =58, Tt is ewmphasized that the results shown in Fig. 11.4 are based on a rough separation technique of the fluid thermal resistances. A more detalled discussion is given in ORNL CF-51-8-32. (9)R. N. Lyon, “Liquid Metal Heat Transfer Coefficients,” Chem. Eng, Progress 47, 75 (1951). (10)5. E. Isakoff and T. B, Drew, ‘““Heat and Momentuim Transfer in Turbulent Flow of Mercury,” to be published. 139 ANP PROJECT QUARTERLY PROGRESS REPORT UNCL ASSIFIED 2 DWG. 11913 10 ISAKOFF AND DREW, Ha, L/D = 58 A o ' \~ Q@@ e 45 / 21 4 Ve S EXPERIMENTAL POINTS FOR / s LITHIUM, £/0 =30 — - O wn wn 2 o LYON EQUATION (NaK) OO & o 10 4 10° 10 PECLET MODULUS Fig. 11.4 - Comparison of ORBNL Lithium Heat-Transfer Data with Those of Other Investigators. 'FOR PERIOD ENDING SEPTEMBER 10, 1951 _ 12. METALLURGY AND CERAMICS W. D. Manly, Metallurgy Division The installation of eguipment in the creep laboratory, powder metallurgy faboratory, and arc-welding laboratory iz essen tiaily complete. The work on the stress-rupture laboratory to test meta is in a ii1quid-metal medium i1s progressing, and the laboratory should e 12 nhshal %y ctober i. In addition, 2 csramics laboratorv has been eos- tablished and 1s mow 1o pavtial oper- 4 C1ON, Welds of the tube-to-header type have been made using nicrobraz, cone arc, and manuai arn; thase welds were evaluated using tensile tftests and wodiafied fatigue tests. Strengths comparable to the parent inconel stock have been found 1n these jmint51 in creep testing the condition of Lhe specimen is being corvelated wal? the creep resistance of the material. The effect of h@a% treatment of inconel on its creep resisbance is considerable, Fabrication of fuel elements usiug technigues previously discussed’ '’ has sontinvued, A new method of fabrication, rubberstatic pressing, has shown promise in fuel-element fabrication, Work on control-red fabrication has been started using the dafferent powder metallurgy techniques, g, A Adamson, “Fuel-Element Fabrication,”™ Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending June 10, 1857, ANP-65, p. 181 (Sept. 13, 1951). : WELDING OF INCONEL P. Patriarca F. W. Dresten G. M. Slaughter Metallurgy Division The bulki of work to date has been confined rte basic evalustion ot the guality of inconel tube-to-hesader welded joints described in the laszt report. ) A limited amcunt of fabri- cation 0f tvypical tube-to-header assemblies has been undertaken, Tensile tests have shown that tube-te-header welds can be made which compare faver- ably 1n both strength and ductility to the parent inconel tubing. Fatigue tests of walded Joints exhibit astrong dependennsy ou the degree of penetvation of the weld, 2 condition not evident in tenstle data. Two manuwally ineve- arc-welded inconel tube-to-header pairs were Lested in ligquid sodiuw at 1000°C for 100 hr to determine the effect on the room-temperature tensile strength. The only effect nbsarved was a slight lowering of the tensile strength and an appreciable increase in elongation due to annealing. Teasile Testx. The results of room-temperature tests of inconel tube-to-header tensile pairs made using manual-inert-arc, cene-arc, and braziong techniques are summarized in Table 12.1. It may be noted that the strength and ductility of manually and cone-arc welded pairs were quite similar and compared favorably to the {Q)P. Patriarca, “Welding of Inconel,” ANP-65, ep, cit. p. 191 141 ANP PROJECT QUARTERLY PROGRESS REPORT TARBLE 12.1 Room—-Temperature Tensile Properties of Inconel Tube-to-Header '"Pairs" NO. OF |AVERAGE TENSILE | AVERAGE ELONGATION, * DESCRIPTION TESTS |STRENGTH (psi) (% 1n 3 1in.) Menual-inert-arc weld with complete pene- 3 93,000 23.8 tration; tested as welded Cone-arc weld with complete penetration; 3 96, 500 21.5 tested as welded Cone-arc weld with average penetration of 4 95, 500 26.9 70% of 0.062-in. header sheet; tested as welded As-received inconel tubing 2 105, 000 30.8 Nicrobrazed 2 82,200 28.9 Manual-inert-arc weld with complete pene- 2 89, 500 36.9 tration; tested after treatment in liquid sodium at 1000eC for 100 hr *Ténsi]e strength based on tube dimensions of 0.187 in. o.d.; 0.025-in. wall except for pairs treated in liquid sodium; tubing used was 0.188 in. o.d.; 0.030-in. wall. final pair length -- initial pair Jength ** . Percent elongation = x 100. initial pair length strength of the as-received inconel tubing used in the fabrication. It is interesting to note that welded joints with average penetration of 70% of the header thickness of 0.062 in, exhibited strength and ductility comparable to welds with complete penetration. Failures in both types of joints occurred in the heat- affected zone immediately adjacent to the weld, as would be expected since this area 1s that of minimum cross- section. Partial penetration requires considerably less critical control of the variables of the cone-arc method, Complete penetration 1is more 142 difficult to achieve with consistency, and 1t is expected that even more difficulty will be encountered when experiments with thinner header sheets and smaller tubing are attempted. Strength values for joints brazed with nicrobraz in preliminary experi- ments are included in Table 12.1. The lower tensile strength and somewhat higher elongation of these specimens may be attributed to the annecaling received in the brazing operation, Brazing of basic joints for evaluation and comparison with welded joints will recelve major consideration in future work. FOR PERIOD ENDING SEPTEMBER 10, 1951 fatigue Tests, In order to evaluate further the quality of tube-to-header joints, a rotating tubular beam fatigue test was developed. The header plate is immovably clamped with the tube extending into a rotating eccentric cam. The test is effectively torsion- less since the eccentric cam, within which the tube end rides, i1s equipped with roller bearings, Specimens were 5.8 1n, long and were rotated at 1125 rpm at a pre-set deflection. ' A limited number of data are pre- sented as an S-N scatter band curve in Fig, 12.1. Further tests are necessary for conclusive evaluation. It may be noted, however, that welds incorpo- rating complete penetration, whether manual inert arc, cone arc, or brazed, exhibited superioer fatigue life to those joints with partial penetration. This condition was not evident on ex- amination of tensile data. The lower fatigue life of these welds may be attributed to the lack of penetration, and, as a result, the presence of a "notceh."™ The evaluation of fatigue, tensile, and corrosion properties of partially welded cone-arc joints with ‘UNCLASSIFIED DWG~Y-4638-RI 1 T 1 I I [TTl &6C 50 [op] Q > 40 — —— o 2 - - 0 Lel * 30 - n 20 - To) i 1 _I_lLl!I! ] I 11||11|' ] &® © O T ‘ 1 MANUAL INERT-ARC WELD, 1000% PENETRATION CONE-ARC WELD, 70% AVERAGE PENETRATION NICROBRAZ CONE-ARC WELD, 100% PENETRATION 10 100 000 NUMBER OF CYCLES X {0° Fig. 12.1 - Effect of Penetration of fFatigue Life on Inconel Tube-to- Header Specimens, 143 ANP PROJECT QUARTERLY PROGRESS REPORT the notch eliminated by brazing will be the subject of future work. All-Weld-Metal Tensile Tests. The properties of weld deposits in inconel plate are of interest since the ARE reactor shell and other components will be fabricated from thick plate, A limited number of tests have been conducted using Y-in. inconel plate and various filler metals to provide all-weld metal 0,357-1n. tensile specimens., The remaining weld deposits were reserved for quantitative analysis and static corrosion tests in liquaid sodium and a fluoride fuel mixture. TABLE The results of room-temperature tensile tests are presented in Table 12.2. 1t may be noted that the physi- cal properties of the inert-arc and metal-arc weld metal were comparable with one exception. The weld deposit of the Inco No. 42 welding rod was definitely inferior. This rod was designed for oxyacetylene welding; however, chemical analysis being con- ducted may explain these relatively inferior physical properties. Inspection of welds of the tube-to- header type involving a relatively intricate geometry and numerous small components 1s difficult. Radiography 12.2 Inconel All-Weld-Metal Temnsile Values 0.357-1in.-diameter test specimens; 2-in, gage length TENSILE SPECIMEN STRENGTH ELONGATION YIELD STRENGTH, NO. DESCRIPTION {psi1) (% in 2 in.) 0.2% OFFSET (psi) 1 Inert-arc welded; 1/8-in. inconel wire 100,000 40.5 55,000 filler metal 2 Metal -arc welded; 1/8-in. Inco No. 132 rod 85, 500 31 47,200 filler metal 3 Metal-arc welded; 1/8-in. Inco No. 132 rod 97, 400 37.5 58,000 filler metal 4 Inert-arc welded; 1/8-in. Inco No. 62 rod 98,000 44 55, 500 filler metal 5 Inert-arc welded; 1/8-in. Inco No. 62 rod 101, 400 43.6 51,000 filler metal 6 Inert-arc welded; 1/B-in. Inco No. 62 rod 92,000 36 48,000 filler metal 7 Inert-arc welded; 1/8-in. Inco Ne. 42 rod 78, 600 33 33,000 filler metal 144 FOR PERTOD ENDING SEPTEMBER 10, 1951 has been attempted with single speci- mens, and deliberate flaws which were located on the radiographs were much more evident with the naked eye., Tt 15 expected that the use of Dy-Chek (a compound used to check for flaws in welds), followed by microscopic examination of weld surfaces, hydraulic pressure testing of assemblies, and helium leak testing will prove useful in detecting major flaws., This prfiw cedure will be set up as partof future work as a preliminary approach to the problem, Corrosion of Welds. The effect of the tensile stremgth of inconel tube-to-header speci- sodium on we lded mens has been indicated previously. Samples of the weld deposit from specimens 1 and 2 of Table 12.2 were treated in liguid sodium at 1000°C for 100 hr and in a ligquid sodium-—po- tassium—uranium fluoride mixture at 816°C for 100 hr, There was no sig- nificant loss or gain in weight of auny of the sawples. Metallographic ex- amination indicated a weld immunity to attack by these liquid metals during the interval of test, A recent prob- because of de- Special Weld Tests. Jem which has arisen parture from the tube-to-header type of reactor involves sealing off of the inconel tubing under a positive helium pressure, Preliminary experiments have been performed which indicate that pressures of 25 psig of helium can be held within inceonel tubing by appli- cation of mechanical pressure to crimp the tubing. While still under mechani- cal pressure, the tubing is sheared from the pressure source and is then welded shut, Pressure-tight tubes with gages brazed on one end for verifi- cation have been subjected to swaging experiments in an attempt to accomplish a uniform cylindrical closure more suitable to the newly proposed ARE design, Another recent problem involves the use of a hydroxide as a cooling medium in preference to sodium. Negotiations are 1in progress with the International Nickel Company for cladding of welded inconel tube-to-header pairs with various thicknesses of nickel. These test specimens will be evaluated for corrosion properties as plated and atter diffusion treatment, Plating experiments with nickel and copper will also be conducted at Oak Ridge using the facilities of the Research Shops at X-10 to supplement the work at INCO. WELDING OF i A contract with Bensselaer MOLYBDENUM Poly- technic Institute for an investigation of the welding of molybdenum has been granted, Jnitilal research will invelve electric resistance flash welding and w1ll be underway during the next quarter, CREEP OF METALS IN CONTROLLED ATMOSPHERES R. B. Oliver J. W, C. W, Weaver Metallurgy Division Woods The installation of equipment in the creep laboratory is essentially complete, The installation in the stress-rupture laboratory for testing in liquid metal 1s about 7T0% complete and 1t 1s hoped that testing can be started in September. ANP PROJECT QUARTERLY PROGRESS REPORT The creep-testing program was started early in June. The initial emphasis was to traln nine new tech- nicians as well as to determine proper operating conditions for the equip- ment, Preliminary results from creep tests on 316 stainless steel, inconel, and niobium have shown that the labo- ratory is functioning properly. These early results are affected by some inaccuracies which are being overcome. Creep of 316 Stainless Steel. Duplicate sheet specimens of 316 stainless steel (0.500- by 0.065-in. reduced section) have been loaded to 3000, 4000, 6000, 7000, and 8000 psa at 815°C in an argon atmosphere. There 1s considerable scatter of the results, and this work 1s being repeated with closer attention to details. The scatter probably reflects the struc- tural i1nstability of the stainless steel. Creep of Inconel. Sheet specimens of 1inconel in the as-received temper (0.500- by 0.065-in. reduced section) have been loaded to 3000, 5000, 6000, and 8000 psi at 815°C in an argon atmosphere., Duplicate tests check very well, and the results are close to the published data. Also, speci- mens were heated to 1120°C in vacuum for 2 hr and air-cooled; the treatment increased the grain size by a factor of about 16. Following this anneal, specimens were loaded to 5000 and 8000 psi. At B00O psi there was close agreement between duplicate tests, but the rupture life was much shorter than With loading the elongataion with the fine-grained specimens. the 5000-ps1 vs., time curves for the coarse- and fine-grained materials were nearly identical. Loadings of coarse- and fine-grained material will be made at 3000 psi; it 1s expected that the coarser grained material will have a considerably longer rupture life at this and lower stress levels. Severe intergranular cracking has been observed on the inconel sheet specimens. These cracks appear to originate at the edge. To see if this is truly an edge effect, one inconel 0.505-in. bar specimen has been loaded to 3000 psi. Unless the same type of cracks develops on the round bar, it can be assumed that this type of crack- ing is an edge phenomenon and need be considered only in the rare case 1in which exposed and stressed edges occur. Creep of Niobium. Two niobium specimens were loaded to rupture at 815°C in argon. Rupture occurred at 26,000 psi with about 20% elongation. One specimen loaded to 6000 psi has been in test for 650 hr with a negli- gible elongation. Work on niobium has been temporarily suspended pending delivery of additional sheet stock. Creep Test of Loaded Inconel Tube, A study is being made of the effects of multicomponent stress systems on creep. An inconel tube % in. in o.d. with the wall thickness reduced to 0.010 in., for a length of 3% in. was sealed at one end and loaded to 66.6 psi internal pressure with argon. The tube was heated to 1800°F, and the pressure was maintained until the tube ruptured, the failure being longi- tudinal. This loading gave the follow- ing stresses: Tangential 1500 psi (tension) I.ongitudinal 750 psi (tension) Radial 35 psi (compression) FOR PERIOD ENDING SEPTEMBER 10, 1951 Rupture occurred in approximately 400 hr with the following deformations: Length 0.8% increase Circumference 2.8% increase Internal cross- sectional area - 5.7% 6.5% increase increase Volume A similar sheet specimen loaded in tension to 1500 psi at 1800°F ruptured in 200 hr with about 6% elongation. STRESS-RELAXATION TESTS R. B. Oliver, Metallurgy Division Eight simulated stress-relaxation tests for type 316 stainless steel were conducted at stress levels of 5000, 10,000, and 20,000 psi using the Baldwin creep-rupture machine. All tests were made at a temperature of 1500°F. After 16 hr the stress was about 4600 to 4800 psi regardless of the initial stress level. Stress- relaxation curves are also to be obtained for inconel and nickel A, Design work for stress-relaxation tests on inconel bolts in sodium at 1500°F has been completed, but con- struction of the equipment has not been finished. FUEL-ELEMENT FABRICATION G. M. Adamson £. S, Bomar J. H. Coobs Metallurgy Division The work on fabrication of solid fuel elements has progressed along the lines discussed in the last report. (! An additional fabrication technique, rubberstatic pressing, is now also be- ing i1nvestigated., With all fabri- cation technigques the use of a screened fraction of sintered;UO2 seems to be desirable, Hot Rolling. After a plate has been fabricated by hot-rolling it is desirable that it be drawn to final size, Since a suitable drawbench is not available, cold-rolling was sub- stituted, A set of plates was fabri- cated by the procedure outlined in OBNL-987. They were hot-rolled at 1100°C to a reduction of 75%, Six plates, each containing 30% U0, , were made; three had matrices of iron and three of 302 stainless steel. Plates from each set were then given additional cold reductions of 1CQ, 25, and 50% by hand-rolling., The plates with iron matrices rolled very well, showing only minor evidence of segregation of the UO,. However, the units with matrices of stainless steel suffered considerable segregation and striation of the UO,. 1In neither case was the bond between the core and cladding material altered. Figures 12.2 and and 12.3 are representative of this work. To determine the effect of using coarse U0, in the fabrication of cores a high-temperature induction furnace was assembled, A toctal of 330 g of a special high-purity UQ, was sintered at 2100°C in a hydrogen atmosphere. The average weight change amounted to a loss of 0.5%. After being sintered the coarse fragments of U0, were uniformly gray in color and quite hard. The apparent density had been increased by a factor of 3. The sintered product was then ground and divided into various size ranges by screening. 147 u ANP PROJECT QUARTERLY PROGRESS REPORT Y.4110 Fig. 12.2 - Effect of Cold-Working Stainless Steel—U0, Cores, 175X, ‘ Y.4227 Fig. 12.3 - Effect of Cold-Working Iron—-—-uo2 Cores. 175X . FOR PERIOD ENDING SEPTEMBER 10, 1951 The -100, +200 mesh size fraction of the U0, was mixed in two batches with -100 and -325 mesh 302 stainless steel. Compacts were prepared in the usual manner and rolled at 1120°C to reductions of 50, 75, and 90%. Me- tallographic examination indicated segregation of the coarse U0, (Fig. 12.4a) to be less semsitive to the particle size of the matrix stainless steel powder than is normal UQO, (Fig. 12.4b). However, a tendency to agglomerate into localized groups for reductions greater than 75% was noted. There was alsoc some evidence of the U0, particles penetrating into the cladding layers., To conserve on stainless steel an attempt was made to use mild steel tubing instead of stainless as the protective covering. The samples were rolled at 1100°C; however, they were heavily oxidized before receiving the first pass through the mill. On metallographic examination it was evident that the oxidizing conditions had penetrated to the compact and had prevented adequate bonding. Further tests will be made, and it is hoped that a satisfactory atmosphere may be obtained by changing the adjustment on the gas dissociator. Mechanically Formed Matrix. The work in the X-10 Shops on developing a method for punching molybdenum plates and obtaining a high percentage of open area, has been held up waiting the delivery of carbide punches., Other methods of forming a suitable mechani- cal matrix are being placed on an in- active basis until the accompanying bonding problems have been solved. A low-temperature induction furnace 1is being assembled for these solid-phase bonding studies., Bonding will be tried both with and without load. The furnace 1is nearly completed and 1is expected to be placed 1in operation. as soon as the induction furmaces are ready. Loose-Powder Sintering. Work on the sintering of loose powders 1is: being carried out 1n our own furnaces. The first run was to be one of a series now to determine optimum sintering temper- ature and was to be run at 1250°C. Instead of the intended as-received UD,, a size fraction {-100, +325 mesh) of the high-density sintered UOQ, was used by mistake. With 30% coarse U0, in 320 stainless steel a fair bond was found at 1250°C., JIncreasing the temperature to 1280°C did not give a definite improvement in the bond. Another sample using regular UO, was prepared and sintered as above; how- ever, very little bond was obtained. Rubberstatic Pressing. Rubberstatic pressing has been used during this quarter as an additional fabrication technique. A drawing (Fig. 12.5) shows how this procedure may be used to press a powder layer onto the inner surface of a tube. Pressure applied to the steel punches is transmitted through the rubber and presses the powder tightly onto the inner side of the tube. This 1s followed by a sintering operation. This technique is being used as a fabrication method on both 'solid fuel elements and con- ‘trol rods. In the first specimen tried in the fuel-element work the lower half of the powder layer was 302 stainless stee}l, and the upper half was electro- lytic iron. After being pressed under 40 tsi and sintered at 1250°C, the stainless steel was well bonded but the iron had pulled away during sintering, A series of samples was - 149 ANP PROJECT QUARTERLY PROGRESS REPORT Y.4428 Fig. 12.4 - Effect of Particle Size of U0, (a) Coarse vuo, . 175X, (b) Normal U02. 250X, 150 FOR UNCLASSIFIED DWG-IZTI?_ N § | ——STEEL PUNCH STEEL BACKING TUBE AN AN b SOSOSOSNY NS o STEEL DIE STEEL MANDREL RUBBER CORE AN I in. _— Fig. DPie for Rubberstatic Pressing. 1205 - high-density coarse UOQ, _ particle sizes in -325 mesh 320 stain- less steel matrices, All the compacts were pressed at 30 tsi and sintered at 1250°C. The sample with -325 mesh U0, was not bonded, while another with ~200, +325 mesh U0, was bonded only superficially. However, using -100, +200 mesh and -60, mesh U0, were well bonded. Compatibility of Potential Fuel- flement Materials. To determine the PERIOD ENDING SEPTEMBER 10, 1951 compatibility of UQ,-iron or U0, —302 stainless steel cores in contact with 316 stainless steel, cladding sections of hot-rolled plates were sealed 1in evacuated tubes. The tube and ‘i1ts contents were held at 1000°C for 100 hr. Metallographic examination disclosed a diffusion layer of less than 0.00]1 in. between the 1ron cores and the 316 cladding and none between the 316 cladding and the 302 cores. Both of these results were in agree- ment with earlier tests, CONTROL-ROD FABRICATION G. M. Adamson E. S. - J. H. Coobs Metallurgy Division Bomar This group i1s now starting work on the fabrication of control rods for the ARE. While hot-rolling and rubber- static pressing are both being in- vestigated as possible fabrication procedures, the most likely method, at present, seems to be a simple canning of powders. Materials now being in- vestigated as possible control ma- terials are B4C and HfOz, but the investigation 1s being expanded to include development of refractory boron, hafnium, or cadmium compounds, As a preliminary investigation ZrQ, was used as a substitute for Hf02, and some plates were prepared with the standard hot-rolling technique. While these two compounds are similar chemi- cally, there is a considerable differ- ence in their density; however, it is not likely to affect this preliminary work. Cores with 30% by volume of ZrOzijlnickel, iron, and 302 stainless steel were fabricated. They were clad with nickel or 316 stainless steel by hot-rolling at 1100°C. Table 12.3 is a summary of the results of this in- vestigation. 151 AND PROJECT QUARTERLY PROGRESE REPOGHRT TABLE 12.3 Remnlts of Iovestigaiions of Zro 25 2 Contvel Material 2 BOND OF CLADDING CORE (vol %) CLADDING REDUCTION (%) TO CORFE REMATHS 30 Zz0,, TG Ni | Nickel 56 Very good Distribution of Zz0, good 80 Very good Distribution of Zr{)2 good 30 ZrOz, 70 N1 | 316 stainless steczi 92 Good Tendeney teward clumping of oxide particles 30 ZrOzv 70 Fe | 316 stainless steel 53 Fair Distribution good T4 Fair Slight tendency toward clumping 30 Zz0,, 70 Fe 316 stainless steel g2 Good Extreme clustering of Zi( particles; sows penstration of ZyQ, into cladding 30 Zr0,, 70 302| 316 stainless steel 71 Very good Distribution very gocd stainless stee! METAL CLADDING OF BERYLLIUM OXIDE Gerity Michigan Company of Adrian, Michigan, has been attempting to developa plating procedure for cladding beryllium oxide. The only oxide readily available at the beginning of this investigation was some high-fired material left from the Daniels and several crucibles of low-density porous material. The high-faired material had a dark film on the sur- face which prevented any bonding. The porous material introduced considerable difficulty from gases and/or solutions reactor becoming entrapped in the pores and then forming blisters during subsequent heating., If this problem can be solved, the Gerity investigators are gquite optimistic that a successful cladding may be plated on. However, two samples which they gave us and which appeared to be satisfactory did not have enough mechanical bond to allow mounting for metallographic 152 study. They have been asked to give this phase of the work a2 low prioraity and to push their electroforming investigation. REFRACTORY METALS A survey of the literature on molybdenum and niobium has been under- taken in preparation for future work with these metals. Simple tests are being run on available specimens to confirm or disprove some of the questionable data. CERAMICS LABORATORY T. N. McVay, Consultant The ceramics laboratory located at Y-12 has been established as part of the Metallurgy Division. The group will consist of nine men under the direction of Dr. John M. Warde. The FOR PERIOD ENDING SEPTEMBER 16, 1951 work of the laboratory is teo be of three types, namely, basic or long- term ceramic research, engineering development, and service research for other divisions of GCak Ridge National Laboratory. A considerable amount of equipment has been received, and it is expected that a large part of the laboratory will be in operation by October 1. ‘ ' Work has already been started on and the first teo be studied are alumina-—uranium oxide compositions. Some work has ceramic fuel elements, also been done on the use of hot- pressed beryllia for valve parts to handle 1ligquid sodium, as 1t has been found that beryllia hot-pressed at Argonne National Laboratory was re- sis tent to sodium at 1500°F. The Metallurgy Division has a work- ing agreement with the Electrotechnical Laboratory of the Bureau of Mines, Norris, Tennessee, whereby the latter will fabricate crucibles and other small refractory parts for OBNL and will Carry on some engineering development projects 1n the nuclear energy field. 153 ANP PROJECT QUARTERLY PROGRESS REPORT 13. CHEMISTRY OF HIGH-TEMPERATURE LIQUIDS W. R. Grimes, Materials Chemistry Division Research in the ANP Chemistry Group has, as indicated by previous reperts in this series, been con- cerned almost entirely 1n the past with development of liquids for use as aircraft reactor fuels. While such studies stil]l constitute an appreciable fraction of this effort, the program has been broadened 1nmn recent months to include study of nonmetallic liquids for use as moderators and/or heat-transfer fluids (coolants) for such a re- actor. These researches have 1in- cluded a small number of physical property determinations, as well as the phase-equilibrium and thermal- stabi1lity studies necessary to demonstrate the chemical properties of the liquids. Nine fluoride-fuel systems, both ternary and quaternary, may be singled out as covering a useful range of uranium concentration {up to about 200 1b of uranium per cubic foot) and with satisfactory melting points (below 550°C). UF,;, which might be used as a fluoride-fuel component, has been prepared with only 0.8 wt % of impurities. In the development of homogeneous reactor fuels, solu- tions of UO; in NaOH-Na,B,0, show promise, together with the NaOH-L10H system previocusly developed. The moderator-coolant program requires a stable (at 800°C) hy- drogenous liquid with good heat- transfer properties. Yhile sodium hydroxide, a satisfactory moderator, 1s an acceptable heat-transfer medium, the existence of a corrosion- resistant container for this fluid at this temperature has not been demonstrated. For these studies quantities of sodium hydroxide assaying better than 99.8% NaOH by welght are now routinely prepared. In addition, potassium hydroxide, barium hydroxide, strontium hydroxide, lithium hydroxide, rubidium hydroxide, and several binary hydrogenous systems are being considered for this application. The development of nonmetallic coolants has been largely confined to consideration of non-uranium- bearing fluoride systems 1n con- Junction with the fuel development program. Eleven fluoride systems of usable liquid range and low cor- rosiveness are listed, although thear heat-transfer properties are not sufficiently well known to permit evaluation of their usefulness. FUEL DEVELOPMENT Research on liquid fuels 1s still directed toward development of low- melting solutions of UF, 1in alkala and alkaline earth fluorides and toward development of self-moderating fuels, with the former program having the major emphasis. Research on the fluorides has progressed so that 1t is possible to list nmine liquad systems, covering a wide range of uranium concentrations, which have satisfactory melting points. Ul,, which 1s a possible fuel component, has beenr prepared in batch samples which assay 99.2% UF; by weight, FOR PERTIOD ENDING SEPTEMBER 10, Solutions of UQ, in NaOH-LiOH and in NaGH~N32B407both show some promise as homogeneous reactor fuels. Low-Melting Fluoride Systems (J. P. Blakely, L. M. Bratcher, and €. J. Barton, Materials Chemistry Division; D. G. Hill, Consultant). Phase equilibrium studies of systems containing uranium tetrafluoride have been continued by the technique of thermal analysis previously de- scribed.¢?*®) Studies of the ternary systems have been extended to include four component systems in an effort to define systems in which low- melting regions were available in a wide range of uranium concentrations. The system NaF-KF-RbF-UF, is con- sidered the most promising of the four—component systems investigated to date. At present the nine systems listed in Table 13.1 below are those of definite promise as fuels. In ad- dition, at least three systems con- taining PbF, have been shown to have suitable melting points at useful uranium concentrations. Jt 1s mnot likely, however, that Pb¥, can be shown to be stable 1n contact with structural metals. in small-scale separation of the (l)C. J. Barton, R. E. Moore, J. P. Blakely, and G. J. Nessle, ¢ LOW»Me}tlng Fluoride Systems — Thermal Ana]y31s " Aireraft Nuclear Propulsion Project (uarterly Progress Report for Period indigg fugust 321, 1950, OBNL-858, p. 110 (Dec. (z)B. E. Moore, G. J. Nessle, J. P. Blakely, snd C, J. Barton, “Low-Melting Fluoride Syscems,” Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending December 10, 1950, ORNL-919, p. 242 (Feb. 26, 1951). (3)5. p. Blakely, R. E. Moore, G. J. Nessle, and C. J. Barton, “Phase Studies of Fluoride Systems,”’ Aircraft Nuelear Propulsion Project uarterly Progress Report for Period Ending March 10, 1851, ANP-60, p. 127 (June 19, 1951). Lapl : I'he recent success 1951 isotopes of lithium‘*? prompted the investigation of systems containing lithium fluoride. From an inspection of Table 13.1 1t 1s apparent that lithium fluoride would be a valuable addition to the list of possible fuel components. The most valuable mixtures without LiF would seem to be the first three listed in the table. The various fuel systems studied during the period are discussed briefly under appropriate headings below. NaF-LiF-UF, Previous studies(S$) had indicated the existence of a low-melting region in this ternary diagram around 31 mele % UF, and 16 mole % NaF. Recent data have permitted construction of the equi- librium diagram, as shownin Fig. 13.1. The low-melting region covers a wide area on the diagram; two separate eutectic points apparently exist inside the 550°C contour. KF-LiF-UF, Substitution of KF for NalF resulted in higher melting points in nearly all parts of the diagram. The areas showing melting points below 525°C are very small although the 550°C contour encloses a large area of the diagram. The Pqulllbrlum diagram for this system 1s shown 1in Fig. 13.2, RbF-LiF-UF, The equilibrium diagram for this system i1s shown in Fig. 13.3. The 550°C contour in- cludes uranium concentrations from 20 to 42 mole % while the 500°C contour includes an area nearly as (4JL. P. Twichell, “Lithiuf~lsotope Sépa- 31. ration,” ANP-~65, op. cit., p. (SL‘NaF?LiF*UF4 System,” OPNL~919, op. c¢if., p. 246. ' ~" - 155 ANP PROJECT QUARTERLY PROGRESS REPORT TABLE 13.1 Summary of Promising Fluoride Fuel Systems LOWEST MELTING COMP. OF LOWEST RANGE OF UF, CONC. OF MIXTURE POINT FOUND (°C) MELTING MIXTURE COMPONENTS MELTING AT 550°C OR LOWER (mole %) (each 110) (mole %) NaF-P\bF-UF4 24-41 500 32.0 UF 23.0 BRbF 45.0 NaF UF NaF RbF KF NaF-KF-RbF-UF, 18-41 500 25. 56. 15. ~N o W o NaF-KF-UF, 26- 30 530 27.5 UF, 26.0 KF 46.5 NaF KF-LiF-UF, 0-41 520 35.0 UF, 57.0 LiF NaF-LiF-UF, 15-39 450 31.0 UF, 16.5 NaF 52.5 LiF RbF-LiF-UF, 21-42 456 35.0 UF, 9.0 RbF 56.0 LiF LiF-UF, 24-32 480 26.5 UF 73.5 LiF NaF-BeF,-UF, 0-21 480 10.0 UF, 15.0 Ber 75.0 NaF NaF-KF-Bel,-UF, 0-25 478 10. 54. 22. 13. UF4 NaF Ber KF Lty © 9 156 FOR PERIOD ENDING SEPTEMBER 10, 1951 OFFICIAL USE ONLY DWG. 12947 NalF 9925°C Fig., 13.1 - Tie System NaF°L1F~EF4. great. The minimum-melting mixture (456 + 10°C) convaians 82 wt % UF, and 7 wt % RbF. Naf -KE-PbF ,~UF,. This system was studied only in a preliminary fashion because of the reaction of PbF, with metal contaimers at elevated tewpera- tures., The system appears to be less valnable than the NaF-KF-UF, system in so far as high uvrasium concentra- tions are concerned. The lowest melting point so far demonstrated is at 493 £ 107C av 20 mole % UF, and 48 mole % NaF, although there is evidence for another eutectic at 470°C of as vet undiscovered com- position. Nef -KF-BeF ,-UF,. This system has been studied 1n some detail up to JO mole % UF, where the melting 157 ANP PROJECT QUARTERLY PROGRESS REPORT 730 °C KUFg 780°C 750°C A<—750- UF, OFFICIAL USE ONLY 1035 °C DWG. 12948 1000 K3 UF; 970 °C J/ 900 N 200 735°C 5 130 ~800 \) KF \\\ 850°C 490LC 845°C Fig. 13.2 - The System KF'LI'F*UF‘. points have become too high for NaF -KF-RbF-UF,. This four- effective use. The 550°C contour surface in the pyramidal diagram extends from 0 to 25 mole % UF,. At the 10 mole % UF, level the 550°C contour area enclosesa large fraction of the triangle, permitting wide variation 1n the proportions of NalF, KF, and BeF,. This system may be of considerable interest 1f low concentrations of uranium are re- quired. 158 component system has been studied in somewhat more detail than any of the others since 1t provides low- melting mixtures containing up to 40 mole % UF,. A sketch from the model of the contour area enclosing all compositions melting below 550°C is shown in Fig. 13.4. For this purpose the solid model was con- structed using triangular prism coordinates and was viewed from the FOR PERIOD ENDING SEPTEMBER 10, 1951 UF, OFFICIAL USE ONLY Gam oG DWG. 12949 720°C RbF / LiF 795°C 845°C Fig. 13.3 - The System BRbF-LiF-UF,. NalF edge of the model. A study of the data i1ndicates that mixture melting below 550°C may be prepared with uranium content varying between 18 and 41 mole %. The lowest meltaing mixture found in this system (500°(C) contains 25, 56.3, 15, and 3.7 mole % UF,, NaF, RbF, and KF, respectively. Preparation of UF, (W. C. Whitley, Research Participant; C, J. Barton, Materials Chemistry Division). Uranium trifluoride 1s of i1nterest to the Fuel Research Group primaraly as a possible fuel component but also as a possible product of radiation damage to UF,. Since the compound is5 not readily available, a method for its preparation 1in a high state of purity has been developed. 0Of the procedures which have been described for preparation of this material, the method using reduction 159 ANP PROJECT QUARTERLY PROGRESS REPORT OFFICIAL USE ONLY DWG. 12950 e e m‘k‘“"""fi =, ‘-——"‘"‘M %M"""%—-m_.‘__‘ e - 50 — 1~ T80 40 ST b T T | a0 2 S @ 2 3 o E & D . 10 h““‘h—vw_._.__.___ ‘__‘._,__J-—"-“'""gojo 50 ] i 0 30 o b T 30 4O 1O 0 10 RbF (mole %) KF (mole %) Fig. 13.4 - The System NaF-KF-RbF-UF,. of UF4 by uranium metal(®) at tem- as yield of UF, 1s concerned. From peratures above 1050°C seemed to be best. While the product of that reaction had been characterized as UF; by X-ray-diffraction methods, no determinations of the purity of the product were made. Preliminary experiments served to demonstrate that temperatures excess of 1000°C were not superior to the range 850 to 900°C in so far in (6).]. C. Warf, Uranium Trifluoride — A Summary Report, AE(D-2523 (March 10, 1949). 160 these preliminary studies products of varying purity were obtained for study. Typical material from these runs was shown to be a black material of specific gravity 8.83 (X-ray density 9.06) which was inscluble 1n all except oxidizing solvents. Chemical analysis showed the material to be about 90 wt % UF,. Experiments 1n cooperation with personnel from the Mass Spectrometer Group of the Stable Isotopes Division showed that 10 to 12 wt % UF, was left unreacted i1n the product. It L] FOR PERIOD ENDING SEPTEMBER 10, 1951 was shown that at 750°C UF, is stable under 10°° mm Hg but that at 800°C the disproportionation into UF, and metallic uranium 1s significant; the disproportionation rate increases rapidly with temperature. Since 1t appeared that low-tem-~ perature reaction for prolonged periods with proper precautions to ensure homogeneity of the reaction mass and to prevent build-up of re- action products on unreacted UF, offered the optimum conditions for preparationof pure UF,, the following procedure was adopted: Pure UF, was vacuum~dried at 450°C for several hours and ground te pass a 200-mesh seive. This UF, was charged into a 10-in.-long by 3~in.~diameter stain- less steel container with the stoi- chiometric quantity of clean uranium turnings and several stainless steel balls. The reagents were again vacuum-dried at 450°C for several hours befere reduction of the uranium to powder by repeated formation and decomposition of UH;. After final decomposition of the hydride, the container was evacuated and sealed. The contents were homogenized by grinding with the steel balls. The reaction was allowed to proceed 16 hr at 900°C, and, after cooling and grinding of the contents, the sealed container was maintained at 900°C an additional 8 hr. The product of this reaction has been shown by chemical analysis teo be 99.2 wt % UF;. This material is of sufficient purity for use 1in phase equilibrium studies. It seems likely that this material is of considerably higher purity than any which has previously been produced elsewhere. A report on this research will be i1ssued in the near future. Homogeneous Fuels (J. D. Bedman and L. G. Overholser, Materials Chemistry Division). The high solubility of uranium as UO; in mixtures of sodium and lithium hy- droxides has been discussed in a previous report.¢’’ This solution might well be of interest as a homogenecus fuel 1f separated lithium isotopes were available and 1f the corrosive liguid could be contained. A small amount of research has been done to ascertain whether similar solubility of uranium can be demon- strated with other hyvdroxide-bearing materials. The data in Table 13.2 indicate that mixtures of barium hydroxide with lithium or sodium hydroxide are of little value in this connection. It 15 established that, at moderately high temperatures, the solubility of uranium 1ncreases with temperature and with increasing concentration of lithium hydroxide; the solubility at 750°C, however, never reaches a value high enough to be of interest as a fuel. TABLE 13.2 Solubility of Uranium in Hydroxides AMOUNT OF HYDROXIDE USED {(wt %) TEMP. SOLUBILITY OF LiOH | Ba(OH), [ NaOH | (°C) | URANIUM (wt %) 25 75 650 0.1 25 75 750 0.2 50 50 650 0.1 50 50 750 0.3 75 25 650 0.4 50 50 750 0.1 (7)“Sol1ibility of Uranium in Mixtures of SodigT and Lithium Hydroxides,” ANP-65, op. cit., p. . 151 ANP PROJECT QUARTERLY PROGRESS REPORT The solubility of UO; 1n mixtures of sodium hydroxide and sodium tetraborate with and without ad- ditional boric oxide 1s indicated by the data 1in Table 13.3. It 1s obvious that considerable amounts of the tetraborate ion must be present 1f significantly large guantities of uranium are to be dissolved. The solubility again shows a sharp tem- perature dependence. Unfortunately, the additionof borate in amounts more than § wt % increases the melting point of the mixture considerably; mixtures containing 15 wt % melt at about 500°C. These studies are to be extended to include other hydroxide- borate systems. TABLE 13.3 Solubility of Uranium in Mixtures Consisting of Sodium Hydroxide and Sodium Tetraborate AMOUNT OF CONSTITUENT IN MIXTURE (wt %) TEMP. | SOLUBILITY OF Na(QOH NazB407 B,0, (°C) URANIUM (wt %) 95 5 650 0.1 95 5 800 0.1 85 15 800 0.4 80 20 800 0.5 55 45 800 5.0 70 15 15 650 0.9 10 15 15 800 2.7 60 15 25 800 4.7 MODERATOR-COOLANT DEVELOPMENT To be effective as a modera- tor—heat-transfer fluid for the aircraft reactor the ligquid must obviously have good heat-transfer properties, must be thermally stable above 800°C, and must contain hy- drogen. Hydrogen-bearing compounds 162 with the required thermal stability, even without regard to probable radiation stability, are not numerous. Development of moderator coolants 1s concerned at present primarily with preparation of very pure alkali and alkaline earth hydroxides and with phase equilibrium studies of various mixtures of these materials with other compounds. In so far as ligquid ranges, moderating power, and heat-transfer properties are con- cerned, several of the materials and mixtures tested could be considered satisfactory. In addition to the pure hydroxides several binary hy- drogenous systems, such as hydroxide- hydroxide, hydroxide-fluoride, and hydroxide-borate, are being examined. Preparation of sodium hydroxide assaying better than 99.8% NaOH by weight has been placed on a routine basis. Pure potassium hydroxide has not yet been prepared in guantity although small-scale preparations using recrystallization from isopropyl alcohol have been accomplished. Pure barium hydroxide, prepared by re- crystallization from aqueous solution followed by dehydration of the re- sulting octahydrate, should soon be available 1n sufficient quantity for corrosion and physical property testing. Methods for purification of lithium and rubidium hydroxides are sti1ll under study. Preparation of Pure Sodium Hy- droxide (I.. G. Overholser, D. E. Nicholson, F. A. Vingiello, Research Participant, and C. W. Harrill, Materials Chemistry Division). Pure sodium hydroxide has been purified by a modification of the method used by FOR TO APPROPRIATE COLD TRAPS AND VROUUM PUMP GROUND Fivtz FRITTED GLADS DISK _ (k , - EW9 RT ' J ‘ % ‘_Ti::T§9uu£TTl“_ ,.“w Q\’)j)?fi VIGRE AUX _ : %0 GOLUMN PERIOD ENDING SEPTEMBER 10, 1951 OFFICIAL USE ONLY WG, 12954 LIPS @g M H ’ § 25/4, %E) 1 355 BALL dom J LITERS g 24{40 3 LITERS - : W” E 0 o e 8 Fig, 12.5 - Scodium Hydroxide Purification Apparalus. workers at the Carnegie Institute of has dissolved. The hot solutinon 1is Technology(s) for use 1n corrosion passed through the fritted glass studies and physical property deter- minations. The purification steps are effected 10 the pyrex apparatus shown in Fig. 13.5. Ethyl aleonhol {1.5 liters) is charged inte flask B and refluxed over magnesium plus a small guantity of 1odine to dehydrate the aleohol. The dryv alcohol is distilled over 1nto flask € which contains 150 g of reagent grade sodium hydroxide pellets and 1is ‘heated until most of the hvdroxide (8)p. E. Sunyder and J. C. . Kelley, The Low~-Temperature Heat Capacity of Sediua Hy- droxide; The Entropy of Sodium Fydroxide at 298. 169K,, Carnegie Institube of Technology, NP-1627 (Junp 1, 1950). filter to remove { sodium insoluble material carbonate, sodium chloride, etc.) and then heated to remove the aleohol until a thick suspension of the sodium hydroxide—~—alcoholate re- in flask D. This flask is removed to a dry beox, and the slurry 1s sucked fairly dry on a fritted filter. The solid 1s then washed with a small volume of cold dry evhyl alecohal, and the solid material 1s transferred to a suction flask. The suction flask 1s heated to about 130°C mains plass under vacuum for at least 24 bt until all the ale oholdte 1s decomposed aund 2 dry powder re- Mains. 163 ANP PROJECT QUARTERLY PROGRESS REPORT Obviously, the atmosphere in the flasks and in the dry box must be free of carbon dioxide and water. This 1s accomplished in the case of the flasks by introducing purified nitrogen. This gas 1s used to force the solution from flask C into flask D and also 1s bled into flask D to reduce bumping during the removal of alcohol. The nitrogen atmosphere prevents the formation of the brown polymerization products which form rapidly i1n the alcohol 1f air 1is admitted at elevated temperatures. The results of analysis of a number of batches of purified sodium hydroxide are given 1in Table 13.4 along with similar data for the starting material. The first three batches were prepared during the early part of the program. Subsequent modificationin equipment and technique resulted in a purer product, as shown by the results for Nos. 4 through 7. At present, sodium hydroxide is being routinely produced with a carbonate content of 0.15% or less and a water TABLE 13. 4 Purification of NaOH by Recrystal- lization from Ethyl Alcohol TOTAL ALKALINITY (%) Na2C03 SiO2 MATERIAL (Calc. as NaOH) (%) (%) Commercial NaCH 97.5 1.5 0.006 96.8 1.4 0.0086 98.0 2.0 0.005 Purified NaCH 1 99.73 0.31 Puri fied NaOi 2 99.81 0.36 Purified NaOH 3 99.83 0.31 Purified NaCil 4 99.90 Purified NaCH 5 99.98 0.13 Purified NalH 6 100.0 0.16 0.008 Purified NaQH 7 99.88 0.07 0.007 content of 0.1% or less. The silica values show that there 1s virtually no pickup of silica in the process used. The yi1eld per batch 1s ap- proximately 65 g, anda total of about 5 1b of purified sodium hydroxide has been prepared by this method. The results of several runs in- dicate that pure sodium hydroxide may also be obtained by removal of carbonate from a concentrated aqueous solution of the hydroxide, followed by dehydration at 450°C under vacuum. For this method a 50 wt % solution of sodium hydroxide is made up in a wax-lined vessel, the suspended sodium carbonate 1s removed by firltration, and the clear filtrate 1s dehydrated in a nickel vessel, first at 200°C and finally at 450°C under vacuum. The results of the analysis of sodium hydroxide purified by this method are given in Table 13.5. A comparison of these data with those given previously indicates that the hydroxide may be about as effectively purified by this method as by the ethyl alcohol process. This method may be applied to quantity production more easily than the ethyl alcohol process. TABLE 13.5 Purification of NaOH by Re- crystajlization from H,0 TOTAL ALKALINITY (%) [Na,CO, |Sio, (Calc. as NaOH) § (%) (%) 164 Purified NaOQH 1 100.0 0.11 0.005 Purified NaOH 2 99.87 0.20 0.004 Purified NaQH 3 99.99 0.16 0.007 Purified NaQll 4 99.71 Purified NaOH 5 100.0 Purified NaQHl ¢ 00.78 Purified NaOH 7 99.94 0.15 FOR PERIOD ENDING SEPTEMBER 10, Preparation of Other Hydroxides (D. E. Nicholson, C. W. Harvill, and D. R. Cuneo, Materials Chemistry Division; F. A. Vingiello, Hesearch Participant, Materials Chemistry Division; and BR. P. Metcalf, Physics Division). Preliminary studies have shown that potassium hydroxide cannot be purified by the ethyl alcohol process used for sodium hydroxide. Apparently the potassium hydroxide~~alcoholate i1s extremely soluble in ethyl alcohol and also is relatively low melting and probably more stable than the cerresponding sodium compound. Limited success has been obtained by using iscpropyl alcoheol, and a study of the use of this compound and of other alcohols 1s underway. Equipment 1s being set up for the purification of strontium and barium hydroxides. The method used entails separation of the insoluble carbonate in aqueous solution, recrystallization as the octahydrate, and finally de- hydration under carefully controlled conditions. The secup, which is constructed of nickel, should permit purification of pound batches of the hydroxides. - ' Purified lithium hydroxide mono- hydrate, availlable commercialiy, was debhydrated at 200°C under vacuum. The dehydrated material so obtained contains approximately 0.1% lithium carbonate and less than 0.,1% water. Rubidium hydroxide, reported to be 99.5% pure, was found to contain approximately 5% carbonate and 15% water. The dehydration of this material is being investigated at present. Decomposition Pressures droxides (E. O. of Hy-~ Price, Research 1951 Participant, Materials Chemistry Division). Experimental work has been confined to a study of the decompesition pressures of barium hydroxide hydrates and anhydrous barium hydroxide. The existence of the monohydrate and octahydrate has been established by numerocus ex- perimenters, but there appears to be some doubt regarding the existence of the trihydrate. The decomposition pressures for the following systems have been reported:(%+19) Ba(OH) , H ,0(s) fing, Ba(DH) ,(s) + H,0(g) Ba(OH) ,* 81,0(s) ;::“.Ba(OH)£H20(s) + THO () These references also include com- parable data for the strontium hy- droxide systems. It mlght be noted that for anhydrous barium hydroxide the decomposition pressure reaches 1 atm at about 1000°C and the cor- responding temperature for anhydrous strontium hydroxide 1s 700°C. The hydrates exert too high pressures at elevated temperatures to be of very much practical value. Pressure measurements have been made on barium hydroxide systems by means of two different types of ap- paratus. One type uses a nickel chamber equipped with a diaphragm for containing the sample, and the pressure exerted against the diaphragnm is measured by balancing against a known pressure on the opposite side £ (J)P M. G. Johnson, ‘“Der NDanpfdruck ven %{ockr;em Salmiak, Z. phystk. Chemn. 61, 457 907 : : (lo)b Tamaru and K. Siomi, *‘Dissoziation von Sr((}H)z, Bafm)z und Hydraren, bestimmt mit Hochtemperatur-Vakuumwaage," Z. physik. Chem. A1T1, 221 (1935). 165 ANP PROJECT QUARTERLY PROGRESS REPORT of the diaphragm. This instrument was used for measuring the pressure of a sample of barium hydroxide ap- proximating Ba(OH),-(1.9)H,0 at relatively low temperatures. Values obtained are given in Table 13.6. These values are 1n fair agreement with those reported in the literature. Measurements were not made at higher temperatures because of difficulties in temperature control. TABLE 13.6 Decomposition Pressures of Ba(QH) ," (1.9)H,0 TEMPERATURE | PRESSURE || TEMPERATURE | PRESSURE (°C) (mm Hg) (°C) (om Hg) 44 6.6 69 19.7 45 6.8 73 21.2 51 8.1 76 24.5 52 9.3 78 25.4 54 10.6 82 27.8 66 16.3 The other type of apparatus employed a tin manometer,using nitrogen to balance the manometer. Experiments with this apparatus using barium hydroxide containing about 2% water showed no measurable pressure at 370°C. For material corresponding to barium hydroxide monohydrate, a pressure 1n excess of 1 atm at 390°C was observed. These measurements are all of a preliminary nature and additional work must be done before any con- clusive results will be forthcoming. Binary Hydroxide Systems (K. A. Allen and W. C. Davis, Stable Isotopes Research and Production Division; C. J. Barton and J. P. Blakely, Materials Chemistry Division)., The 166 system Ba(OH),-Sr(OH), has been studied in some detail by thermal- analysis methods because these materials have been reported to be somewhat less corrosive than the alkali hydroxides. Figure 13.6 shows the completed phase diagram for this system. There is evidence for a solid transition in the com- position range near the eutectic. While these researches were made with dehydrated commercial materials which contained some carbonate, 1t is safe to say that mixtures of usefully low melting point are available in this system., It 1s likely, however, that the decom- position pressures are relatively high at 1500°F. The system L10OH-NaOH has been examined because certailn mixtures of these materials dissolve ap- preciable quantities of UO;. The equilibrium diagram is shown in Fig. 13.7. The low-melting eutectic (219°C) contains 27 mole % LiOH; there 1s evidence of a compound NaOH*LiOH with incongruent melting point. The thermal effects observed at about 180°C are probably due to a solid transition. Phase diagrams for the systems NaOH-KQOH, NaOH-RbOH, and RbOH-KOH have been reported. (1) Phase studies of the other possible com- binations will be deferred until more pressing problems are finished or until it becomes clear that one of the systems offers appreciable ad- vantages from the corrosion stand- point. (11)G. v. Hevesy, “Uber Alkalihydroxyde. 1I,” Z. physik. Chem. 73, 667 (1910). FOR PERIOD ENDING SEPTEMBER 10, 1951 OFFICIAL USE ONLY - DWG. 12952 350 T | 1 I 500 | o W 450 = 4 v & 400 = i b 350 & Q e . ® Sr(OH), 10 20 %0 40 50 60 70 80 90 Ba(OH), Ba{OH), (mole %) Fig. 13.6 - The System Sr(fl“)z-na(OH)z. Hydroxide-Fluoride Systems (W. C. Davis and K. A. Allen, Stable Isotopes Research and Production Division; C. J. Barton and J. P. Blakely, Materials Chemistry Division). Scarpa®!?) has published studies of the NaOH-NaF and KOH-KF systems. According to his experiments the former binary shows the formation of mixed crystals with a solubility break with the melting points of all the mixtures intermediate between those of the pure components. The KOH-KF system 1s similar, with melting points intermediate between those of the pure components but with the (12)g, Scarpa, ‘‘Thermal Analysis of Mixtures of Alkali Hydroxides with the Correspondin Halides,” Atti acced. Lincei 24, [, 738 an 955 (1915). ' compounds completely miscible in the solid state. These data have not been com- pletely confirmed in this laboratory. The incomplete study of the KOH-KF system seems to show a eutectic at about 40 mole % KF melting at 630°C. Itis apparent, however, that materials with a useful melting point may be obtained only at high concentrations of hydroxide; these mixtures would probably be similar to pure hy- droxides in so far as corrosion is concerned. Addition of NaOH to the termary LiF-NaF-KF eutectic to the extent of 10 mole % can be accemplished without great elevation of the melting point. This system and the analogous system 167 ' ANP PROJECT QUARTERLY PROGRESS REPORT OFFICIAL USE ONLY DWG, 12953 500 I s r I 450 400 RATURE (°GC) TEMPE 1o o & q 250 ] . | | | 200 NaOH 10 20 30 40 50 &0 70 80 20 L.iOH LiOH (mote %) Fig. 13.7 - The System NaOH-LiQH. using KOH will be examined in more deta1l. lydroxide-Barate Systems (K. A. Allen and W. C. Davis, Stable Tso- topes Research and Production Divi- sion; C. J. Barton and J. P. Blakely, Materials Chemistry Division). These materials are of 1nterest because of their solvent action on UQ, and also because of the possibility of re- ducing considerably the corrosiveness of the alkali by incorporation of the borate. The data on the sodium hydroxide~—sodium tetraborate system 168 indicate a eutectic at 95 wt % of sodium hydroxide. The liquidus cunrve climbs steeply from this eutectic and shows a break at the composition corresponding approxi- mately to the orthoborate. COOLANT DEVELOPMENT Recently 1initiated studies leading to development of nonmetallic coolants have been largely confined to studies of the alkali and alkaline earth fluoride systems. While mixtures of " FOR PERIOD ENDING SEPTEMBER 10, 1951 useful liquid range and low cor- TABLE 13.7 rosiveness are in hand, the heat- . . : Low-Melting Non-Uranium Fluoride transfer properties are not yet known . .. .. . Eutectics with precision sufficient to permit ~ evaluation of their usefulness. Table 13.7 lists the several fluoride MELTING systems which are discussed in the COMPOSITION (mole %) POINT (°C) literature or which have been studied : ' in the course of the fuel development 50% KF - 50% LiF . 492 program and which show melting poants low enough to be of interest to the S6% KF - 447% AlF, 5?0 program, 60% KF - 40% PbF, 470 33% NaF - 67% PbF 505 It should be noted that these . 2 materials are largely by-products 30% LiF - 70% PbF 460 of the fuel research, and 1t 1is 58% NaF - 42% BeF, 342 likely that a concentrated effort 42% KF - 58% BeF 315 will serve to add to this list and 2 | | to produce materials of lower melting 40% ROF - 607% BeF, 435 point and viscosity. A program of 48% LiF - 52% BeF, , 360 study of other fused salt systems 46.5% LiF - 42.0% KF - 11.5% NaF 454 which are likely to be better heat- _ transfer agents 1s underway. 30% NaF - 5% KF - 65% BeF, 345 169 ANP PROJECT QUARTERLY PROGRESS REPORT 14. RADIATION DAMAGE D. S. Billington, Physics of Solids Institute A. J. Miller, Research Director’s Division Radiation damage and radiation- induced corrosion in the prime con- stituents of the airplane reactor core have been under examination during the past few months. The X-10 graphite pile, the Y-12 cyclotron, and the Berkeley cyclotron have been utilized as radiation sources., The limitations in the significance of the information being obtained are due mainly to the low flux i1n the graphite pile and the lack of methods for correlating effects of accelerator particles with those of pile radiations. In addition, the Berkeley cyclotron has a low beam current, and the Y-12 cyclotron needs a more uniform beam and efficient target~cooling mechanisms that will allow it to be used for irradiating with the high beam While using the presently available sources, the experimental groups are expending the specimens currents that are available. the maximum effort to design and con- struct equipment suitable for use 1in the LITR and MTR and to use higher beam currents 1in the Y-12 cyclotron. In the following description of the experimental procedures and the re- sults to date, the significant items are the stability of the molten {fuel in i1nconel capsules during graphite pile and cyclotron irradiation, the small changes 1n creep rate of 347 stainless steel during graphite pile and the large decrease thermal conductivity in the irradiation, in inconel detected graphite pile. in one experilment 170 CREEP UNDER IERRADIATION J. C. Wilson J. C, Zukas W, W, Davis Physics of Solids Institute A series of three fairly reproducible cantilever creep tests conducted 1in the ORNL graphite reactor exhibited increased total creep strain, to bench tests, compared after approximately 100 hr of operation. Up to this test time there appears to be slightly less creep under irradiation than on the bench, The 1n-pile specimens show about 10% greater creep strain after 200 hr and nearly 20% more strain than the corresponding bench test after about 250 hr, Type 347 stainless steel was the metal tested, the maximum fiber stress was 1500 psi, and the tempera- ature was 1500°F. The creep of 347 stainless steel at 1500°F and a stress of 1500 psi has been measursd by the cantilever creep apparatus in the X-10 pile. The measurements previously reported(!') have been extended, and to date three fairly reproducible 250-hr cantilever creep tests have been made. maximum fiber Up to 100 hr the creep under irradi- ation is slightly less than that of (). ¢. Wilson, J. C. Zukas, and ¥. W. Davis, “Creep Under Irradiation,” Aircraft Nuclear Propuflion Project Quarterly Progress Report for Period Ending June 10, 1951, ANP-65, p. 200, 201 (Sept. 13, 1951}. esp. p. FOR PERIOD ENDING SEPTEMBER 10, the unirradiated control specimens. However, the in-pile specimens exhibit about 10% greater elongation after 200 hr than the control specimen, and nearly 20% greater elongataon after 250 hr. A compesite plot of all the creep data i1s given 1in Fig. 14.1. The reproducibility 1s good except for the test represented by cnrve O, which has been disregarded in this analysis of the data. The diminished creep rate observed during pile shutdowns (plotted on a larger scale in the lower right- hand corner of Fig., 14.1) substantiates the conclusion drawn from the total extensions, namely, that irradiation increases the total creep strain after the initial stages of the creep history, 1t was also the main reason for the summary disregard of curve O in analyzing the data., 1f the divergence between the bench and in-pile curves then the difference in 18 real, creep strain should be guite large, per- centagewise, after longer times. a set of specimens will for 500-hr periods and the total strain measured after removal from the reactor, In this way any possible errors in strain measurement because of radiation effects on the microformer will be eliminated,. A motre complete description of all the 2) Accordingly, be irradiated above work has been given, The observation of diminished creep at shorter times is in agreement with our earlier work.®) QOther workers{4® have reported only decreased creep under irradiation, but in the tests (2)Phy3tC5 of Solids Institute Quartef!y Progfess Report for Period Fndzng July 31, 1851, ORNL- 1148 (in press). (S)Ph sics of Solids Institute Quarterly Progress geport for Period Ending April 30, 1951, ORNL- 1095 (in press). 1951 reported here the duration of the test is much longer and the total strains much qmaller than 1n any other reported creep tests under irradiation. This could, in view of the ignorance of the mechanisms accounting for creep defor- mation at slower rates, be responsible for the apparent difference in results. These creep measurements under irradiation will be supplemented by a tensile creep apparatus which is being designed for use with the LITR. This tensile apparatus will provide amn independent check of the cantilever creep apparatus measurements of creep under irrvradiation. Anocther creep apparatus to be operated in the LITR will provide stress-corrosion creep data. Thisg apparatus, now in the design stage, is basically a wmodified sodium loop. RADIATION EFFECTS ON THERMAL CONDUCTIVITY A. F., Cohen, Physics of Solids Institute Preliminary in-pile tests have been made on one sample of inconel, (%2 ‘The thermal conductivity at 820°C decreased by a factor of 2 1in less than one week of exposure in the X pile and then maintained a fairly constant value. The effect of an in-pile temperature anneal (by increasing the temperature of the specimen 70°C for 40 min and then returning the specimen to 820°C) (4)H P. Yookey et al,, Effect of Cyclotron Irradiation.on Creep of Alumtnum Nnrth American Aviation report NAA-SR-121 (June 8, 1951). (S)Quarterly Progress Report for the Period January 1.~ March 31, 19851, NEPA-1792. (6)A. F. Cohen, "Radiation Effects on Thermal Conductivity,” ANP-65, op. cit,, p. 205. 171 i~ SECRET = ™o PS|-B-180 o DWG - 12450R! | | % = 0.120— ] ] = St 0 N M g L l T ] |\ ' g O!GO"""— i i —_— ; INDICATES PILE ” Q - SHUTDOWN PERIOD /e _ £ s - = < = e 0.080}— ] _g — = < g Z & - W ©0.060— — w — i _.0.100+ = 0 L ;g ] ;g z z 5 wl O o = mad 0.040— 5 — £0.0901 wl _ ld — = < 0.020}— & _ — _ 0.080 O/;LREFER.I‘ENCE SLOPE 7 B0 160 170 18O 190 _ 0 | | | 50 100 150 200 250 300 TIME (hr) Fig. 14.1 - Composite Plot of Three Bench (0) and Three In-Pile (0) Cantilever Creep Curves. 347 stainless steel at 1500°F; 1500 psi maximum fiber stress; irradiated in hole 15 of ORNL graph- ite reactor, fast flux & 4 X 10'° neutrons/cm?/sec, On the ordinate, 0.001 deflection = 15 x 10-6 in./in. strain. FOR was to cause annealing out of approxi- mately 20% of the damage (i.e., the decrease in thermal conductivity). With further irradiation the thermal conductivity returned to the former equilibrium value. Experiments planned to give information on the flux de- pendence of the thermal conductivity are in progress on types 310 and 316 stainless steel and on pure nickel. IRRADIATION OF FLUDRIDE FUELS G. W. Keilholtz, Materials Chemistry Division The effect of radiation on the stability of the molten fluoride fuel contained in inconel is currently under investigation using the X-10 pile and Y-12 cyclotron. Detailed analyses of irradiated samples are not yet available, Preliminary spectrographic analysis of the first cyclotron- irradiated capsule showed no detectable radiation-induced corrosion., Capsules containing fuel prepared with uranium enriched 1in U%35 showed no evidence of pressure build-up under pile irradi- ation. ‘ Pile Irradiation of Fuel Capsules (J. G. Morgan, C. C. Webster, P. R. Klein, R. L. Cooper, D. F. Weekes, and D. D. Davies, Physics of Solids In- stitute). JInconel capsules containing fuel prepared with normal uranium and with uranium enriched in U?%33 tested for pressure build-up due to fuel decomposition in ‘the X-10 pile. The fuel composition was gtandard NaF-KF-U¥F,, 46.5-27-26.5 mole %, The containers were pressurized to 58 t 1 psi with helium at 1500°F in the pile. During 200 hr of irradiation there was no detectable increase 1in the pressure. Figure 14.2 shows the construction of were PERIOD ENDING SEPTEMBER 10, 1951 S Y 4gad 14.2 - tainer with Thermocouple Well and Pressure Fitfings. Fig. Liguid-Fuel Sample Con- the inconel sample container wath thermocouple well and pressure fittings. Examination of the capsules and their contents for radiation-induced cor- rosion and for fuel decomposition 1s in progress, Cyclotron Irradiation of Fuel Capsules (W. J. Sturm and M. J. Feldman, Physics of Solids Institute; R, J. Jones, J. 8. Luce, and C. L. Viar, Electromagnetic Research Di- vision). The cyclotron experiments 173 ANP PROJECT QUARTERLY PROGRESS REPORT have been directed toward the determi- nation of the effects of proton bom- bardment¢’? upon the inconel capsules containing the standard NaF-KF-UF, ARE fuel, Four capsules were bombarded with 2 to 3 pna of 20-Mev protons which pro- vided an energy dissipation of approxi- mately 20 watts per cubic centimeter of fuel. Two l-hr and twa 4-hr runs were made, Radiation cooling was employed, The first capsule melted in but 1n temperatures were maintained 1n the region 1300 to 1350°F. A partial analysis has been made on one of the l-hr samples. No significant changes in fuel composition, trace element content, or corrosion rate due to i1rradiation have been detected., However, some grain growth has been detected i1n the inconel from both the irradiated and an unirradiated control sample. the region of intense beam, subseguent runs One capsule was mounted on a water- cooled block and irradiated with a 20- pa beam current for 1 hr at 1500°F, This capsule has not yet been examined owing to 1its high radioactivity. CORROSION OF IRON BY LITHIUM UNDER CYCLOTRON IRRABPIATION North American Aviation, Inc, W, W. Parkinson, Physics of Solids Institute Bombardments of irom 1mn contact with molten lathium have been carried out in the Berkeley 60-1in. to provide information cyclotron about the (")G. W. Keilholtz, J. C. Morgan, H. Robertson, and C. C. Webster, “Irradiation of Fluoride Fuels,” ANP-65, op. c¢it., p. 207. 174 behavior of structural materials under irradiation while in contact with liguid metals, Corrosion data on the first three capsules irradiated are given in Table 14.1. There is no evidence of accelerated corrosion due to 1irradi- ation. Micrographs were made of cross- sections 1n the i1rradiated and un- irradiated zones of the capsule from run 1. These micrographs(8) show that in some areas, apparently dis- tributed at random through the irradi- ated zone, an 1increase 1n graln size has taken place. No conclusions have been drawn yet as to the nature of this change. The microhardness was found to have an average value of 89.5 Knoop hardness numbers in the irradi- ated zone and 89.3 in the unirradiated, compared with an average of 106.5 over the whole surface before bombardment. Additional irradiations of molten lithium 1n iron capsules have been carried out, and the analytical re- sults are being obtained. This will complete the study of lithium 1in 1iron for the present time. The next series of irradiations will be made on molten fluoride fuels in inconel capsules. LIQUID METALS IN-PILE LOOP C. D. PRaumann 0. Sisman R. M. Carroll W. W. Parkinson C. Ellis Physics of Solids Institute Lithium at 1000°F has been circu- lated in the X-10 reactor for one week. (8) Fig. 14.5, ANP-65, op. ctit., p. 212. FOR PERIOD ENDING SEPTEMBER 10, 1951 TABLE 14.1 Results of 31-Mev Alpha Irra@iatiun of Lithium in Iron Capsules TEMPERATURE | TIME AT EXPOSURE |Fe TN Li CHARGE®) |RADIOACTIVITY OF CHARGE(®) AUN NO. (°C) TEMP. (hr) | (pa-hr) (%) counts/min) 1 625 9.0 25, 1 0.008 2 2 910 6.2 13.6 0.23(c) 2700¢¢) 3 675 17.3 3.8 9.004 - 0.006 2 Control 906 5.7 None 0.004 - 0.006 None (®)Surface to volume ratio, 5.1 cm’/cm® with 1. 24 g of lithium in each capsule. (b)Activity measured on step 1 of Berkeley Scaler, model Decimal 2000. () eak developed in capsule. Jron oxide may have flaked into charge. A detailed description of the experi- mental equipment was presented in the last guarterly report.(®) The princa - pal components of the apparatus are a Y-in.-1.d. 316 stainless steel loop which extends to the center of the pile connected to a 3/l6-in.-1.d. loop external to the pile, an electro- magnetic flowmeter, and three i1oni- ration chambers tomeasure the activity at various points on the external circuit of the lithium loop. The srimary purpose of this experiment was te study the Bremsstrahlung activity '(g)g, 0. Baumann, R. M. Carroll, and O. Sismén, Liguid-Metals In-Pile Experiment,” ANP-65, op. cit,, p. 211. , ' from the decay of 1i% created by the {n,») reaction on Li’. The analysis of this activity is reported in Sec. 9, "Nuclear Measurements," A cursory visual examination of the tube from which the lithium had been removed showed no signs of damage. Also, there was no evidence found of radicactive corrosion products in the Yithium. Forther investigation will be made when the tubes hecome somewhat less radicactive. Rather than pursue the corrosion due to lithium any further at this time, an inconel loop containing sodium is to be operated at temperatures up to 1500°F, 175 15. Nuclear Development Associates, As a result of an extensive analysis of the supercritical water reactor first proposed in WASH-24,(?) Nuclear Development Associates has advanced detailed specifications for a feasible desi1gn, The reactor and shield com- prise an ll-ft-diameter sphere of water with a 2.5-ft square-cylinder active core at its center., The reactor will deliver 400 megawatts maximum wall temperature of 1290°F and with a a fuel investment of approximately 25 1b. The proposed fuel elements are of the sandwich design considered in the last quarterly report,(?) and the core design utilizes a flat neutron flux. The pressure shell introduced some problem of weight, but the overall shield weight, 180,000 1b, remains well within specifications, OUTLINE OF A SPECIFIC DESIGN Apart from some insertions {(pressure shell, thermal shield for pressure shell, gamma-ray shadow shielding cap, flow headers, etc.) the reactor and its shield may be approximately visualized as an 1ll-ft-diameter sphere of water, in the interior of which are spaced some 200 parallel fuel assemblies to form a 2.5-ft sgquare-cylinder (D Abstracted from NDA Quarterly Report on ANP Activities, June 1 to August 31, 1951, Y-F5-55. (2 )Alrcraft Reactor Branch of USAEC, Appli- cation of o Water Cooled and Moderated Reactor te Aircraft Propulsion, WASH-24 (Aug. 18, 1950). (S)Nuc]ear Developmwent :Assccisates, Inc., “Supercritical-Water Reactor,” Aircraft Nurlear Propulsion Project Quarterly Progress Report . Period Ending June 10 1951, ANP-65, p. 2 8 esp. p. 220. fgept 19515 SUPERCRITICAL WATER REACTOR (D Inc. central core. Fach fuel assembly is =2 tube about an inch square, in 'the itnterior of which 1s a fine-scale structure of fuel-bearing plates and water coolant passages, The fuel assemblies occupy about one-fourth of the core volume, the remaining space between them being filled with moderat- spacing between at the axis of the and increases radially cutward* manner as to keep the thermal- ing water, The assemblies 1s least core, in such flux approximately the same for all the assemblies. Thus the heat load is the same for all the tubes, the stainless steel heat transfer surface 1is neutron being used to maximum effectiveness in all the tubes, and the amounts of stainless steel and fuel required 1n the reactor are minimized., Two water flows may he dis in this (1) the high- speed flow throogh the fuel assemblies, which picks up the fission recoil (2} the slow flow of the moderating water between the assemblies, which picks up a smaller amount of heat, principally from neuvtrons and gamma rays. The relative speed of these two flows can be adjusted by a 15.1. shim tinguirshed arrangement heat, and valve as 1ndicated in Fig. This has desirable features for con- trol since it permits varying the temperature rise {and thus the density) of the moderating water without any temperature or pressure changes in the & The cross-section of the assemblies may have te be decreased, and their number increased, near the edge of the core to avoid neutron lasses from B too coarse moderator structurs. This WOJ]d make the number of tubes larger than 200. 179 ANP PROJECT QUARTERLY PROGRESS REPORT SECRET DwWG. 12956 MODERATOR WATER INLET | PRESSURE SHELL OUTLET HEADER EXIT TO TURBINES T =98C°F T FUEL ASSEMBLY INLET FROM PUMPS 7= 480°F j;aBERATDR WATER OUTLET Fig. 15.1 - Supercritical Water Reactor Flow Arrangement,. Table 15.1 summarizes some of the numerical values for the example selected, external power plant system. FUEL ELEMENTS AND ASSEMBLIES Further consideration of the fuel- element possibilities discussed 1in the last quarterly report¢3’ has led to the selection of an oxide type element, The use of powder metallurgy technigques to form a fuel sandwich composed of a sintered UO, and stain- less steel powder layer rolled between two pieces of stainless steel seems particularly attractive. The original work at ORNL‘3***’ produced successful sandwiches of this type with an over- all thickness of 30 mils; however, other research has indicated that this (4L‘Powder Metallurgy — Fuel Element Fabri- cation,” Metallurgy Division Quarterly Progress 1951, Report for Period Ending Januery 31, ORNL-987, p. 52 (June 7, 1951). 180 figure might be lowered considerably, The powder metallurgy technique dis- tributes the UO, in separate tiny pockets in the interior of, in effect, an otherwise continuous piece of stainless steel. It appears that there will be sufficient free volume to accept the fission gases at moderate pressure, Thus, pending in-pile tests of this fuel-element type, we are inclined toward some optimism regard- ing 1ts chances for ultimate success- ful development. A stainless steel (347 or 316 Nb) sandwich of 20 mils total thickness, with the central layer about 10 mils thick and containing, some 10% UO, by volume, 1s suitable. There are many conceivable ways of arranging such plates into assemblies. Figure 15.2 illustrates one arrangement designed to give stiffness, maintain spacings, permit thermal expansions, etc. Here alternate flat and corrugated fuel plates are stacked to fill a square unfueled tube. The contact of the two fueled walls has been estimated from NACA-TN-2257 to run 30°F hotter than the average metal wall temperature, REACTIVITY It appears desirable to keep the thermal-neutron flux radially flat in the core to minimize the amount of stain- less steel needed for heat-transfer purposes, If the amount of fuel needed was dependent only upon the amount of steel used, then this flat- flux design would obviously minimize the critical mass as well, Furthermore it 1s actually found, with appropriate simplifying assumptions, that uniform thermal flux in the core does accompany minimum critical mass, Thus flattening the neutron flux operates directly to. FOR PERIOD ENDING SEPTEMBER 10, 1951 TABLE 15.1 Summary of Reactor Design-Point Values Heat output Core size Water inlet temperature Water outlet temperature Maximum wall temperature Fuel plate‘thickness ' Maximum heat flux from fuel plate Longitudinal maximum/average power ratio Average heat flux from fuel plate Weight of stainless steel fuel plates Percent of unfueled steel added in fabrication Toral weight of stainless steel 1n core Average density of water in core(®? Critical mass in most favorable idealized case(®) Pressure shell inside diameter (sphere) Pressure shell thickness Pressure shell weight Cooling stream hydraulic diameter Cooling stream inlet velocity Cooling stream exit velocity Cooling stream pressure drop Water flow rate Percentage of core volume occupied by: Stainless steel fuel plates Additional unfueled stainless steel Water in high-speed channels . Low-speed moderating water 400,000 kilowatts(?) 2.5-ft square chifider 480°F 980°F 1290°F (P 0,026 1in. 1,250,000 Btu/hr-ft? 1.4 900,000 Btu/hr- ft? 972 1b 22 700 0.73 g/cm? 10 - 12 kg of U35 4.5 ft 1.5 an. 400G 1b 0,065 in. 12 ft/sec 73 ft/sec 25 ps1 430 1b/sec ] LA L N O LN D W (a)Ihe heat output required for level sgperéonic flight at altitude 1s approkimately 300,000 kilo- watts in WASH-24 and NEPA-1843; WASH-24 indicdates a maximum power (sea-level) requirement of 398,000 kilowatts, which figure has not been checked by United Aireraft. (b)lnc}uding allowance of 30°F for hot spots.. (C)When the moderator flow rate and density are as high as possible in Fig. 15.1. d . . . ( )Unpolsoned water density at the above maximum value, no losses from coarseness of structure, ete. 181 ANP PROJECT QUARTERLY PROGRESS REPORT SECRET DWG. 12957 F 1.031n. 0.407in. |0,15?in.| F;—*wwvv*%—wmuogam. ---------- v——fiw——fl SCALE: ABOUT 245, TIMES ACTUAL SIZE Fig. 15.2 - Cross-Section of Fuel- Element Assembly. minimize the fuel content, as well as indirectly by minimizing the amocunt of steel, The problem solved was the follow- ing: Given a region containing moderator and given some fuel (pure fissionable material, or a definite mixture of fissionable material and structure)}, what is the minimum amount of the fuel and how should 1t be distributed in the moderator in order to obtain a ecritical assembly? By assuming that (1) the reactor is thermal, and (2) the volume occupied by the fuel is negligible, 1t was proved that there 1s a region in the moderator in which fuel should be placed, and in this region the thermal flux should be constant. The fuel 182 distribution can be obtained for suitably shaped systems. The above restrictions are well satisfied for pure U?3% immersed in an infinite sea of water, heavy water, beryllium, or carbon. Thus we are now in a position to calculate the minimum critical mass of a chain- reacting assembly moderated by any of the usual moderators. This calculation has not yet been carried out. A calculation for the same power output, and the results of an illustrative calculation using the same power output and with the same assumptions otherwise, comparing a reactor with uniform fuel-tube loading with the flat-flux reactor has been made. Both reactors had cores 75 c¢m long and 67 cm in diameter and reflectors with outer diameters of 100 cm. The flat- flux reactor had 300 kg of stainless steel. The uniform fuel reactor required, for the same power output, 1.54 as much stainless steel and 1.41 as much urantum, STABILITY Calculations of the equivalent bare reactor have been expanded to include the detailed flux shape in the reflected reactor and a fairly detailed repre- sentation of the space variations of flow and heat transfer. First results have been obtained and will be reported as soon as they have been analyzed. The general procedure is to con- sider first small oscillations from equilibrium. The periods of the system are then obtained as the eigenvalues of the problem. A vari- ational principle is used to convert the system of partial differential equations to a matrix problem, which 1y FOR PERIOD ENDING SEPTEMBER 10, 1951 is then handled by methods discussed in a recent report.(%) 1In this method of approach the accuracy of the results for the perioeods and time behavior 15 limited only by one’s diligence and ingenuity in applying the variational principle. : PRESSURE SHELL the pressure shell aond the iayer of heavy Between reactor core is a material to cut down the gamma heat release in the shell. Both this thermal- shielding layer and the shell 1tself are of substantially constant thickness all around the reactor and represent 4 substantial fraction of the Since tend to total gamma-shield weight, divided-shield arrangements concentrate the heavy material at the side of the reactor facing the crew, the pressure shell and 1ts thermal shield reduce the efficiency of the overall shield design by putting substantial amounts of material in the wyrong place. To keep down the weight associated with the pressure shell we may (1) keep the shell nearly spherical and its radius fairly small, (2) reduce the thermal-shielding layer until heat transfer from the shell begins to be troublesome, {(3) make the shell of high-strength alloys such as U. 8. Steel stainless W or Armco 17-7 PH or 17-4 PH, (4) run the pressure-induced stresses quite high in the shell, or (5) permit thermal strains large (S)NDA Quarterly Report on ANP Activities, March 1 to May 31, 1951, OBNL, Y-12 site, Y-F5-47 (June 14, 1951}, : eniough to cause yvielding in the outer fibers.* For application in the above- mentioned gamma heating problem, methods have been devised for valculating the the gamma leakage from spherical reactors with vadial wvarization in component concentratrons., 1he gamma build-up processes cccurring inside the reactor have been included approxi- mately, and calculations have also been made of the angular distribution of the emerging photons. Present un- certainties in the energy distribution of gammas produced as a result of fission prevent accurate calculation of the leakage; conservative assumptions have been made, but they could be off by 50% or more. SHIELD Preliminary estimates of divided- shield weights were made for the super- critical water cycle for a range of reactor core slzes, reactor-crew separation distances, and reactor powers., The basic shield used was essentially that described in ANP- 53,(%) modified to include several inches of 1ron near the core because of the pressure shell and its heat shielding. This iren layer was considered useful shielding material, since the fast-neutron flux for this design was considerably smaller at the pressure shell than the primary gamma flux. (The thermal-neutron flux reaching the i1ron was considered stopped by a boron laver.} Then an *While it a pears unlikely that a ductile material can faig by thermal stress, the allowable thermal strains may be limited by practical considerations, such as embrittlement, temperature cycling, gpenings, strain concentration at hot spots, or supporbs. 183 ANP PROJECT QUARTERLY PROGRESS REPORT appropriate thickness of lead was removed from the shadow disks and all the lead was removed from the sides of the compartment, * From the results of this preliminary study, assuming a 44-ft separationm distance crew * Work hasz just started on estimating the nzutron-indnced activity of the circulating water. Preliminary estimates indicate that some of the gsmma shielding wight have to be restored at the sides of the crew compartmeni for designs with condensers far out in the wings (as in a modified B-47, with the engines in pods below the wings). 184 (center of reactor to vear face of crew compartment), the following shield weights for the power and core diameter of Table 15.1 were estimated: Core shield weight 58,500 1b Crew shield weight 21,500 Total shield weight 80,000 1b The shield weight allowed for in the airplane design (WASH-24 and NEPA-1843) 1s 91,600 lb. FOR PERIOD ENDING SEPTEMBER 10, 1951 16. CIRCULATING-MODERATOR-COOLANT REACTOR: HKF Atomic Energy Division, The H. XK. Ferguson Co., Inc. A study of a subsonic anircraft reactor using circulating sodium hydroxide as both mederatoer and coolant and a fixed fuel of UFJ3 dissolved 1in alkalil and alkaline earth fluorides has been completed by the Atomic Energy Divisiocn of The H. K. Ferguson Ce, under contract to OBNL. As a result of this study it was concliunded thatr the circulating-moderator-coeclant reactor can provide ample power for a subsonic airplane. Actually the performance of the aircraft is inter- mediate between the NaOH homogeneous and the UBi circulating-fuel reactors discussed in previous quarterly re- ports. %37 Principal advantages of the circulating moderator-ccclant system over the other two include the use of a moderator of known properties and behavior and the fact that the secondary coolant is not made radio- active, which simplifies the shielding problem. The main drawback, in addition to the fact that a suitable container for sodium hydroxide isnot yet assured, is the multitude of metallic heat- transfer tubes in the core through which heat is transferred from fuel to coolant, The present study was con- fined to the reactor because it was (I)Abstracted from the report by K. Cohen, Circulating-Moderator-Coolant Reactor for Sub- senic Aircraft, HKF-112 (Aug. 29, 1951). This reactor i described in some detail in HKF-112. (Z)H. K. Ferguson Co., “Homogeneous Circu- lating-Fuel ' and Circulating-Moderator Reactory Aircraft Nuclear Propulsion Project Quarterly Praogress Report for Period Ending March 10, 1951, ANP-60, p. 308 (June 19, 1951). : (S)H. K. Ferguson Co., ““Circulating-Fuel Reactor,” Aircraft Nucleer Propulsion Project Quarterly Progress Report for Period Ending June 10, 1951, ENP—GS, p- 224 (Sept. 13, 1951). {1) SEGRET DWG. 129588 NOOH FLOWING INLET HEADER ‘ INTO TUSES i { _ TUBE SHEET | é%ZF?£2¢&Z P L L i 1 | o {TYPIGAL) o SQUARE - « CYLINDER ™ § - i £ | . & o i @ IJI- o CONTROL | — SECTION ! LMt _i_ - wl p] @ i = X - I i 3 i g = 2. £ o I @ TUBE SHEET W 777 7 # 7 d FPZ 77722 L LY RN NAOH OUT OUTLET HEADER SECTIONAL ELEVATION Fig. 16.1 - Schematic Drawing of Reactor Core. : known at the outset that the inter- mediate heat exchanger and remainder of the power plant would be almost 1dentical with those designed for the homogeneous sodium hydroxide reactor and that the power requirements are similar, A schematic diagram of this reactor core is shown in Fig, 16.1. OPERATIONAL CHARACTERISTICS The circulating-moderator-coolant reactor is designed to provide power 185 - 3 - ANP PROJECT QUARTERLY PROGRESS REPORT to operate a modified B-52 type plane crursing at 35,000 ft altitude and Mach 0.8. The maximum metal tempera- ture was set at 1500°F, The problems of a solid moderator are eliminated and a simpler structure 1s adegquate. The liguid fuel has the advantages that it can be drained upon specified landing, which permits a large saving in fuel inventory for a fleet of nuclear the fuel can expand out of which provides a stabilizing aircraft; the core, effect for reactor control. The reactor delivers 140,000 kw at the design point of 35,000 ft and Mach 0.8, with a coclant outlet temperature of 1380°F, inlet 1180°F. The maximum heat flux i1s 530,000 Bru/frt?2-hr, At sea level 1t provides 230,000 kw. This 1ncrease 1s accomplished by in- creasing the temperature rise through the reactor (inlet 1007°F, outlet 1335°F) while increasing the heat flux {maximum 870,000), and i1t permits climb from sea level at 720 ft/min at a speed of 250 knots. REACTOR CHARACTERISTICS The present reactor design has the characteristics of all fixed-{fuel reactors: (1) the heat must be trans- ferred first through the fuel and then through the canning material; (2) the coolant flow must be distributed in accordance with the flux distribution; and (3) the core must contain a compli- cated structure subject to thermal and mechanical stress and to radiation damage. This reactor utilizes the type of ligquid fuel planned for use in the proposed ARE reactor at ORNL, the essential difference being the use of hydroxide as the combined coolant and moderator as compared with the OBNIL use of a fixed moderator of sodium 186 BeO and a coolant of sodium. Sodium hydroxide was specified as the moderator coolant in the HKF proposal because 1t is considered to have the best combi- nation ol properties for this use: adequate slowing-down power; does not freeze, decompose, or have a high vapor pressure 1n the operating temperature range; and 1ts neutron absorption 1s not excessive for this type of reactor. An advantage of this design over the homogeneous sodium hydroxide reactor and the circulating-fuel reactor 1s that the fuel does not circulate ogutside the reactor during operation. This tends to decrease the reactor shield weight, and, more 1lmportant, avoids irradiation of the heat exchanger and outside coolant fluid with delayved neutrons. It also eliminates loss of delayed neutrons The large number of common to all noncirculating-fuel aircraft reactors, 1s a disadvantage compared to the homogeneous reactor, Probably more important than the problems and advantages of the mechanical design are the materials problems. 1{f ma- terials resistant to corrosion by sodium hydroxide with good high- can be outside the core. heat-transfer tubes, temperature characteristics found, the circulating-moderator- coolant reactor appears to be capable of powering the subsonic aircraft, CORE DESIGN The reactor core is a square cylinder 32 in, 1in diameter. The sodium hy- droxide moderator cooclant flows down- ward through small, closely spaced inconel tubes, Fuel occupies the space between the tubes. A jacket of sodium hydroxide around the core acts FOR as a reflector as does the sodium hydroxide in the top and bottom headers. l.Lead provides a gamma shield to pro- tect the pressure shell against over- heating. The sodium hydroxide flows inside the tubes, thereby ensuring against hot spots which could be caused by interruptions in flow, which would occur if the flow were outside the tubes.’ This positive flow path is believed necessary, as sodium hydroxide has poorer heat-transfer properties than sodium, for example. The fuel was designed to be drain- able, in order to decrease greatly the uranium inventory for a group of planes. This advantage, however, brings with it a serious leakage problem, Since the fuel 1s in a common container, it could all escape through a large leak. The effect of small leaks i1s minimized by balancing pressures so that coolant leaks into the fuel, rather than the reverse. REACTOR PHYSICS The uranium weight is high, 187 1b of U**® in the core. This is mostly PERIOD ENDING SEPTEMBER 10, 1951 because of the large amount of inconel, necessary for heat-transfer surface. The reactor could not be much smaller without exorbitant uraniunm content. The uranium could be somewhat by going to larger sizeg, at the expense of increased shield weight, reduced The sodium hydroxide reflector brings up the thermal-neutron flux, and therefore power production, in the The effect toward the center as would be desired, owing to the large absorption in the reactor, and the effect on leveling out the neutron flux is only moderate. core near the boundary. does not extend as far The amount of fuel 1n the middle of the reactor is varied to give a total shim control of 10% reactivity. For inherent temperature stability the fuel in the end zones, which is B0% of the total, can expand to chambers out- side the reactor. No provision is made for xenon override, as draining and replacing the fuel after shutdown makes 1t unnecessary. 187 ANP PROJECT QUARTERLY PROGRESS REPORT 17. L. F. Hemphill CIRCULATING-MODERATOR-COOLANT REACTOR: ORNL R. W. Schroeder H. R. Wesson ANP Division A preliminary design of a hydroxide cooled and moderated reactor has been defined. A 2.5-ft spherical reactor should be able to deliver 200 megawatts with a maximum wall temperature of 1500°F and a maximum coolant outlet temperature of 1450°F, Only 115 megawatts 1s believed adequate for flight at Mach 0.75 at 35,000 ft. An essential feature of the design 1s the use of annular fuel elements to attain a high ratio of heat-transfer surface to fuel volume., These design studies were based on the use of sodium hy- droxide, as 1ts properties are better known than those of the other possible hydroxides. FLUID-CIRCUIT SPECIFICATIONS After exploratory investigations, it was decided to attempt to design a 200- megawatt reactor with a 2.5-ft spherical core, with a maximum wall temperature in contact with the coolant of 1500°F, and with coolant inlet and outlet temperatures of 1100 and 1450°F, respectively., These temperatures imply a permissible coolant film drop of 400°F at the inlet end, and 50°F at the outlet end. The thermal con- ductivity of sodium hydroxide 1is poor relative to the thermal conductivities of liguid metals, causing heat-transfer coefficients and film temperature drops to suffer. Attainment of the 50°F film drop at the outlet end therefore indicated special treatment to be necessary. With all other variables fixed, film drop varies directly with 188 unit power and inversely with V928 where V 1s fluid velocity. Therefore the design should be such as to permit the coolant to exit from a region of minimum power, and such as to permit high coolant velocities at the dis- charge end of the flow path. Con- versely, lower velocities can be employed at stations farther removed from the discharge end, mizing pressure drop. thereby mini- FUEL-ELEMENT DESIGN Fuel-element studies indicated the ratio of heat-transfer surface required to fuel volume required to be high, approximately 1000 ft? surface to 0.33 ft*® volume. (The fuel volume is based on the use of fused fluorides with a uranium density of 150 1b/ft3, and an estimated critical mass of 50 1b.) The use of cylindrical fuel elements with fuel filling the elements would have involved a prohibitive number of very small tubes. The use of annular fuel elements, with fuel in the annulus and coolant moderator flowing inside and outside the element, permitted acceptable fuel temperature gradients, and a mechanical design of enhanced simplicity and ruggedness. REACTOR DESIEGN The foregoing considerations led to the configuration 1llustrated in Fig. 17.1. The entering fluid 1s passed through the reflector region first FREE FLOW 75 % HEAT-- TRANSFER AREA 960 f12 . NO. FUEL ELEMENTS 2446 FOR PERIOD ENDING SEPTEMBER 10, 1951 SECGRET DWG. 12959 FUEL ANNULUS O(H2 =(L377 0D 1 0.OH WALL TUBING {(OUTER TUBE) 0323 0D 1 0015 WALL “0 PER INGH HELICAL SPAGER THREAD (INTEGRAL WITH INNER TUBE) TYPICAL FUEL ELEMENT DETAIL "A" MATERIAL CONSTITUENGY CODLANT-MODERATOR 5.9 17 STRUGTURE 16 £t FUEL 0.37 112 _—MANIFOLD HEADERS TOTAL 7.87i . ~GAS BLEEDER LINES —~DUMP AND EMPTYING LINES SRR SEE DETAIL "A'3—2_ e eR T ey v Mo s NaOH OUTLET PIPE 5-in.IPS AND THERMAL SLEEVE 6 REQ'D EQUALLY SPACED o - o =N Fe N2OH OUTLEY PR :}/‘; ' , o . . _ % NaOH INLET ROTATED AND THERMAL CLERVE : \. © 3G° FROM TRUE POSITION & REQ'D EQUALLY SPACED SLOQUTLET PLUG FOR GAS - BLEED AND. EMPTYING LINES ELEVATION 17.1 - Sodium Hydroxide Cooled ard Moderated Reactior, Dimensiocns are inches unless otherwise specified. 189 ANP PROJECT QUARTERLY PROGRESS REPORT SECRET DWG. 12960 ] i T T T 2101 MAX {n \\\\\////, FUEL TEMP. 026 MAX < (PEAK; AT CENTER ”1 1600 Ry e . /1543 mMaAX 1500 MAX CF FUEL SLAB) . \\\ //” \\\\ ///” > \\\ ’/’,—G—_ ) -9_ _J 1500 \ " \i;///// \\\\i////f:w%;::“, T FUEL SIDE 1;‘:] - . \/ ‘\,/ m—lifip MAX CUEL B P 1400 T T CO .LANjT WAt ] x ; ,/’”//A COOLANT TEMP. a. = 1300 bt ‘?EQT_MAX } k- 7 e __:“’ O g ~—=-0 1200].. e o 1280 MAX fi\l250 MAX SOLID LINE, 200,000 kw 100!l BROKEN LINE, 115,000 kw -J TWO TOP CURVES DRAWN ON TEMP SCALE (2); AlLl. OTHERS ON SCALE () 1000 ) | { L { | | | | i A 8 CD E FG H IJ K LM N 0 REACTOR STATIONS (REFER TO FiG. I7H Fig. 17.2 - Temperature Throughout Sodium fiydroxide Core. (the fluid comprising the reflector) to enable the pressure shell to operate at the lowest temperature in the cycle. It 1s then passed through the core via five parallel passes, starting at the core centerline and discharging from the peripheral pass. flow veloci- and heat-transfer coefficients are plotted against flow path length in Figs., 17.2 and 17.3. Data are plotted corresponding to two design conditions in Table 17.1. The 115- megawatt conditions correspond to the requirements for Mach 0.75 at 35,000 Reactor temperatures, ties, 190 TABLE 17.1 Design Coalant Condition for Maximum and Cruise Power COOLANT COOLANT INLET OUTLET TEMP. (°F) TEME. (°F) Maximum power 1100 1450 200 megawatts Cruise power 1050 1250 115 megawatts FOR PERIOD ENDING SEPTEMBER 190, 1951 SECRET e DWG. 12961 < I T ( I B s Ca 200 - oE INSIDE_TUBE m \(\) i : e e \\‘-.‘ S%ac | / A - 5 é e ________.___'“‘""_'""'—Tfll‘f: ----------------------------------------- / “oh \OUT SIDE TUBE 3 . % 40k | = CINSIDE TUBE ' £ - , >_ 50”’ \\\ e - > _ - S 200 | e 3 l Oh“ ';LWZL:::::::::::M / ) '5‘;’ . QUTSIDE TUBE | [ N b, J C | N [ O3 B o E G H 1 K M N 0 REACTOR STATIONS (REFER TO FIG.{7.1) Fig. 17.3 - Sodium Hydrexade Veloclty and Film Coefficient’ ’Fhruughout fore. Velocity and A for 200,000 kw; same for 125,000 kw. ft, based on NEPA studies for their Design features currently under type IL-14 power plant. The 200- investigation include a fuel header megawatt conditions were selected and fission product gas bleed-off initially for design purposes and were system, and a control system involving intended to include calculated require- displacement of reflector sodium ments with a generous contingency Thydroxide by inert gas. Both systems margin., ' : include negative acceleration features. 191 ANP PROJECT QUARTERLY PROGRESS REPORT 18. North American Aviation, At the request of Oak Ridge National Laboratory, North American Aviation has conducted an investigation of high-temperature (above 1800°F) helium and sodium liquid-vapor power cycles with regard to their application to the Phase II (Mach 1.5, 45,000 ft) arrcraft, This work has been com- pleted, and detailed reports are being prepared for publication. A brief summary of the conclusions reached as a result of these studies 1s presented here, Aerodynamic studies here indicate that L/D ratios of the order of 4.0 to 5.0 represent a maximum attainable range for large supersonic aircraft at Mach 1.5 and 45,000 ft. Values at the higher end of the range can be achieved only with an exceptionally clean airframe and high-impulse engines., (This 1s in sharp contrast to L/D values of the order of 5.7 to 6.5 quoted generally in the literature for systemswithrelatively low-impulse engines.) A consequence of reduced aerodynamic performance i1s i1ncreased power plant impulse, and this in turn requires engineering a higher tempera- 1s presently As a ture power plant than under consideration elsewhere. reference, the necessity of high- temperature power plants for the supersonic propulsion of nuclear air- craft has previously been i1ndicated by NAA.(?) (I)Reprint of an official letter from North American Aviation to Oak Ridge National Laboratory outlining the results of their power plant studies {gglfhc ANP program, NAA-AFR-51A-1653 (Sept. 10, 192 HIGH- TEMPERATURE POWER PLANT STUDIES Inc. North American has investigated compressor Jets as a means for the attainment of the necessary high temperatures. In the binary com- pressor-jet cycles, the working medium in contact with the high-temperature components need not necessarily be an oxidizing fluid. If the operating temperature of the radiator is main- tained below 1800°F, the problems of obtaining suitable high-temperature materials may be significantly reduced, since all high-temperature components are not subject to oxidation. On the basis of experiments con- ducted at North American and elsewhere, there appear toexist materials capable of withstanding high temperatures (up to 3500°F) in nonoxidizing atmos- pheres. DBefore these materials can be considered seriously for engineering application, a major developmental effort 1s required to establish tech- niques by which they can be fabricated into useful components. Furthermore, much has to be learned about the cor- rosion resistance of these materials to the various coolants and the work- ing medium proposed. Some pre- liminary experimental work along these lines has been undertaken at this laboratory. To date, results are encouraging but not yet conclusive. For example, static corrosion rates of 3250°F sodium vapor on molybdenum are apparently well within the required Iimits for the aircraft power plants. (z)J. A. Malone, A. S. Thompson, and H. Schwartz, Remarks on the Necessity of High- Temperature Power Plants for the Attainment of Supersonic Nuclear-Powered Aircraft, North American, NAA-SR-Memo-33 (March 29, 1951). ey FOR North American has conducted an analytical investigation of high- temperature compressor jets to de- termine the plausibility of such systems and to provide a guide for any future experimental work that might be undertaken, The aim in this work has not been to establish the feasibility of systems that might be constructed in the near future. The aim has been rather to determine whether an extensive matertals and component development program for high-temperature com- pressor-Jjet application would be justified. | An integrated reactor, power plant, and airframe study has been conducted for several proposed systems, carried out 1n sufficient detail to determine if a plausible supersonic airplane is achievable., Some systems guickly appeared severely limited in per- formance and onlvy a small amount of analytical detail was reguired. This was true for the gas-cooled reactors. The sodium liquid-vapor compressor jet was developed in considerable detail and appears to be one plausible system, In all cases the analysis was dependent on extrapolaticens of rela- tively meager experimental data on properties of materials and working fluids. The numerical results that have come out of these analyses are not intended to be final and fixed and should be considered in the light of their intended exploratory purpose. SODIUM LIQUID-VAPOR COMPRESSOR JET A liquid-metal-cooled reactor operating in conjunction with a sodium liquid-vapor compressor-jet system has been investigated. On the basis of this study the following specifications for a supersonic aircraft have been established: | PERIOD ENDING SEPTEMBER 16, 1951 General: Mach no. {cruise) - 1.5 Alrvitade {(cruise) 45,000 f¢ L/D (cruise) 5.0 Takeoff speed Landing speed 70, knots TG knots Weight: Gross weight 400,000 1b Power plaat weight 145,600 1b feactor and shield (divided) 120,000 1b Structure weight and miscel Janeous 120,000 b Pay load weight 20,000 1b Reactor: Diameter 4.0 ft Leagth 4.0 ft Moderator BeC + € Fael ‘ U homoge- neously impreg- nated in moderator Conlant Tin Feactor wall temperature ~A000°F Fuel inventory 73 1b Reactor power 535,000 kw Power plant: , Sodium eveporation temperature 260005'{di¥ saturate vapor) Sodium evaporation pressure 290 psia Sodium condensate temperature 1700°F Maximum air temperature ~1450°F Air flow rate 1518 1b/sec Air specific impulse 528 1b thrast/1hb air/sec Several airframe configurations and power plant arrangements were con- sidered, each being optimized to arrive at the best overall performance. In the final configuration the entire power plant is housed within the fuselage, and side inlet scoops are provided for air intake. A shielded crew compartment 1s in the nose of the ship some 50 ft forward of the reactor and capable of housing five 193 ANP PROJECT QUARTERLY PROGRESS REPORT crew members and the necessary control equipment. The power plant is aft of the reactor; the fuselage diameter is approximately 13 ft at this section. The power plant consists of a sodium- vapor generator heated by molten tin, located immediately aft of the shielded reactor. Sodium vapor passes to five turbines mounted circumferentially around and entirely within the fuselage. Each turbine is directly coupled to a 303-1b/sec air compressor which drives atr aft and across a common radiator, the sodium condenser, which occupies the entire fuselage diameter. Air continues aft through a nozzle pro- ducing the propulsive thrust; the sodium condensate 1s pumped forward to the vapor generator, The task of developing such a power plant as described above would require a very extensive effort. However, 1f the appraisal of supersonic aerodynamic performance is correct, and 1f amanned, supersonic nuclear vehicle 1s required, a program of this sort should be under- taken, The sodium cycle is suggested as one possible approach, not neces- sarily aunique one, which on the basis of preliminary studies appears to have no fundamental limitations that pre- clude 1ts development, HELTUM-~COOLED REACTORS in addition to compressor-jet systems in which helium serves both as a reactor coolant and working medium, several turbo-jet systems were investi- gated in combination with helium- cooled reactors. These investigations have indicated that a major development effort uti- for lizing helium-cooled reactors 194 supersonic propulsion of aircraft cannot be justified. In no case does it appear that a supersonic vehicle can be achieved, even when considering reactor temperatures as high as 3300°F and helium pressures in the range of 1000 to 2000 psi. There are fundamental limitations involved having to do with the comparatively poor heat-transfer characteristics of the gaseous coolants. A basic requirement for a high- performance aircraft is light, compact heat-exchange equipment with minimum auxlliary power requirement. Jt 1s well known that at moderate pressures and temperatures gaseous coolants are several orders of magnitude poorer in these regards than the liquid coolants. Increasing gas pressure tends to reduce this difference, as does 1ncreasing temperature differentials between hot and cold fluids. At high pressures equipment becomes more compact and pumping power 1is reduced, but little or no weight saving 1s accomplished since heavier walled equipment 1is required to accommodate the increased pressure. The requirement that temper- ature differentials in heat-exchange equipment be large forces the maximum power plant temperatures to 2 high level without any realizable thermo- dynamic benefit, While temperatures in this analysis as high as 3300°F were considered, air specific impulses remained relatively low (approximately 30 1b thrust/lb air/sec). In all specific designs considered for the helium system, no plausible supersonic vehicle was found. Specific power plant weights in excess of 2.6 1b of power plant per pound thrust were obtained in all instances. Even assuming an L/D ratio of 5.0, which 1s hardly realizable with these low- impulse engines, an aircraft of a gross weight well in excess of 450,000 1b 1s FOR PERIOD ENDING SEPTEMBER 10, 1951 required. ' On this basis it was recom- devoted to the helium, or other gas, mended that no additional effort be cycle. ' 195 19. ANALYTICAL CHEMISTRY" C. D. Susano, Analytical Chemistry Division Satisfactory chemical methods have been developed or suitably modified for the determination of metallic cor- rosion products -~ principally iron, nickel, chromium, molybdenum, and platinum — in reactor fuels composed of eutectic mixtures of fluoride salts., Two colorimetric methods for low concentrations of cobalt in fluoride salt mixtures and alkali hydroxides were found to be too i1nsensitive (sensitivity, 100 ppm) for the purpose. Work is continuing on this problem. A crystalline material, which was deposited in the lithxum at a hot- cold junction in a coolant test systen during an extended shutdown period, was identi1fied by X-ray-daiffraction and chemical analysis as a mixture of slightly contaminated lithium hydride and hydroxide. A method for the determination of carbon in lithium, both carbide and carbonate carbon, has been developed. analvysis A review of the analytical work for the various phases of the ANP Program shows that approximately 50% of the work is concerned with the analysis of reactor fuels of the mixed fluoride salt types, 30% with the analysis of sodium and lithium, and the remainder with miscellaneous samples quite in nature. A total of 1781 (75% chemical, 25% were: made on 310 varied determinations spectrographic) samples. (I)Abstracted from the report by M. T. Kelley and C. D. Susano, Analytical Chemistry — ANP Program, Quarterly Progress Report for Period Ending August’ 31, 1951, ORBNL, Y-12 site, report Y-B31-283 (Aug. 30, 1951). .fi The n-butyl bromide method, which is now being used regularly for the determination of oxygen in sodium, is found to yield excellent results for oxygen concentrations of 0.015% or greater. Studies are being made of refinements of this method so that it may be applied with reasonable accuracy at even lower concentrations. A search is being made to find a liquid medium suitable for use in melting sodium samples from metal tube containers; certain hydrocarbons which are under test appear to be promising for this application. Preliminary tests directed toward the adaptaticn of the method of Pepkowitz and Judd®?) (for oxygen in sodium) to the determi- nation of oxygen and nitrogen in lithium do not appear promising. However, an apparatus 1s under con- struction for use in determining the oxygen content of metallic lead. The method to be used incorporates modifications and refinements of the technique used by Funston and Reed(?? for the determination of oxygen in bismuth. The oxygen content of inert gases {(principally helium) is now being determined by a colorimetric method which gives excellent precision below 25 ppm; at higher concentrations of oxygen the precision drops off sharply. An alternate method will be tested as SZ}L. P. Pepkowitz and W. C. Judd, “Determi- nation of Sodium Monoxide in Sodium,’ A4nal., Chea. 22, 1283 (1950). (3)g. s, Funston and S. A. Reed, “Determining Traces of Oxygen in Bismuth Metal,’ Anal. Chen. 23, 190 (1951). 199 ANP PROJECT QUARTERLY PROGRESS REPORT soon as apparatus fabrication of the required 1s complete. ANALYSIS OF REACTOR FUELS K. Bowan, Jr. J. C. White W. J. Ross Analytical Chemistry Division In the course of the development of methods for the determination of me- tallic corrosion products in NaF-KF-UF, and NaF—Ber—UF4 euntectic fuel mix- tures, a considerable amount of infor- mation has been accumulated as to the solubilities of these mixtures 1n various reagents. This information is being compiled and will be 1issued as a separate report. Methods and techniques have been developed for the determination of metallic corrosion products -— chiefly iron, nickel, chromium, and molybdenum — 1in two reactor fuels, the ternary eutectics NaF-KF-UF, and NaF-BeF, -UF,. At present only the NaF-KF-UF, mixture is of interest from the standpoint of corrosion studies. Platinum and silver are found in the eutectic since thermocouples of these The technigue of from metal metals are used. removing the eutectic sample tubes after heat treatment, which was described in previous re- ports, *+3) has also been applied successfully to samples contained 1in tubes of unusually small size (1/8 1in. i.d.), which have to be used in cor- rosion studied in the X pile, (4)M. T. Kelley and C. D, Susano, Analytical Chemistry -~ ANP Program Quarterly Progress Report, ORNL, Y-12 site, report Y-B31-260 (May 31, 1951). See also ANP-65, p. 234. (S)M. T. Kelley and C. D. Susano, Analytical Chemistry Division (Y-12 Branch) Quarterly Progress Report, URNL, Y-12 site, report Y-B31-270 {June 9, 1951). 200 Spectrographic Results. The eu- tectic mixture was prepared for spec- trographic analysis of corrosion products as described in previous reports. (**5) The agreement of spec- trographic and chemical results was satisfactory for iron, nickel, and chromium. Molybdenum was not de- termined spectrographically. Since the spectrographic preparation is time- consuming and the results essentially substantiate the chemical results, spectographic tests have been dis- continued. Development of Colorimetric Methods, Two reagents for the determination of traces of cobalt in fluoride eutectics and fused alkali metal hydroxides, tetraphenyl arsonium chloride and sodium diethyl thiocarbamate, were investigated. Neither showed sufficient sensitivity under the conditions imposed in this case, 1.e., the limited amount of sample available and the number of other determinations required on the same sample. The lower limit of sensitivity for these reagents 1is of the order of 100 ppm. Other reagents, nitroso R salt in particular, will be tried, and the procedures examined for possible means of inmcreasing sensi- tivity. The procedure for the colorimetric determination of chromium has been standardized to give 5% accuracy 1in the range 100 to 20,000 ppm by adopting the procedure of G. F, Smith(%) which assures holding all the chromium in the same valence state. Excellent precision has been obtained in determining molybdenum in quantities (G)Mixed Perchloric, Sulfuric, and Phosphoric Acids and Their Applications in Analysis, G. F. Smith Chemical Co., Columbus, Ohio (1942). FOR PERICGD ENDING SEPTEMBER 10, less than 100 ppm by using the thio- cyanate-—stannous chloride complex and extracting with butyl acetate, Determination of Platinum. Platinum is currently being used as a thermeo- couple material in studies of the physical properties of the fluoride eatectic fuel mixtures, and 1t 1s therefore necessary to analyze these mixtures for platinum. Samples analyzed for platinum have generally been found to contain less than 100 ppw. Platinum®’’ may be determined colorimetrically as the yellew chlore- platinpus ion, PtClfiéf, the reduction frem Pt{IV)} to Pt(II) being ac- complished by stannous chloride in acidic solution. The color is stakble at acid concentrations of about 0,25 N, Palladium interferes in the procedure. The interference of yvellow U(VY) ion 1is eliminated bv extraction of the uranium with tribautyl phosphate. OXYGEN IN SODIUM Jr., and J. C. White Analytical Chemistry Division B. Bowan, The n-butyl bromide method for the determination of oxygen in sodium is being used on a routine basis samples of oxvygen content of about 0.015% or wmore. Studies are now pro- gressing to determine the accuracy of this and other methods (Pepkowitz and Judd in particular) in the range 0.001 to 0.005%. for (?)E. B. Bandell, Celorimetric Determination of Traces of Metals, p. 494, Interscience, New York, 1950. 1951 GXYGEN IN LEAD D, L. Mananing and W. K. Miller Analytical Chemistry Division The need for s method for the determination of small amounts of oxygen in lead has led to the in- vestigation of the method of Funston and Reed.{(®} This method invelves the measurement of the reducticn in volume of hydrogen when heated in centact with the sample; however, the total volume of the apparatus is so large in comparison with the swall change in volume during the reaction that the method is highly inaccurate. A new apparatus is being designed to correct this failing in which a small known volume of hydrogen is introduced 1into a combustion chamber and allowed to burn with the oxygen from the lead, When the reaction is complete, the remainder of the hydrogen 1s swept out of the chamber with C0,, the CO, absorbed in reaction with KOH, and the volume of the rewaining hydrogen determined, OXYGEN IN HELIUM J. €. White and W, J. Ross Analytical Chemistry Division The method of Brady(®’ discussed in the last quarterly progress report(9®> has been modified and used with good results. The new procedure 1s to measure the velume of test gas re- quired to produce a 5% change 1in transmission of the sodium anthraquinone sulfonate reagent, which changes (B)L. J. Brady, “Determination of Small Amounts of Oxvgen in Gases,” Anal. Chea, 20, 1033 (1948). (QJJ, C. White and W. J. Ross, “Oxy felium and Argon,” gen an Aircraft Nuclear Propulsion Project Quarierly Progress Report for Period %gg{?g June 10, 1351, ANP-65, p. 238 (Sept. 13, 201 ANP PROJECT QUARTERLY PROGRESS REPORT directly proportional to the oxygen content of the gas. The precision of the results is good at oxygen concen- trations of less than 25 ppm, but drops off sharply at higher concen- trations because of the smaller test gas volumes used. This 1s 1llustrated by the following typical results of 26 12 helium determinations made on cylinders: 0, (ppm) TEST VOLUME OF HELIUM (m}) A 3.96 12 2.26 10 11 (avg.) B 0.45 81 0.17 : 130 0.23 300 170 (avg.) A modification of the Winkler method for the determination of oxygen as developed by Pepkowitz(19) will be tried for comparison with the method now 1n use, The necessary apparatus is being fabricated. OXYGEN AND NITROGEN IN LITHIUM R. Rowan, Jr. W. K. Miller D. L. Manning Analytical Chemistry Division An attempt was made to apply the method of Pepkowitz and Judd(?2) oxygen in sodium to this problem with- out success. for The reaction does not proceed until the temperature of the mixture of mercury and lithium 1s raised almost to the melting point of (10), .. P. Pepkowitz, private communication. 202 lithium, much violence. Adaptation of this reaction to a controllable procedure and then takes place with does not seem imminent. CARBRON IN LITHIUM R. Rowan, Jr. W, K. D. L. Manning Analytical Chewmistry Division Miller Most of the carbon in lithium 1is present as lithium carbide, Li,C,, which decomposes on contact with water C,H,; however, some carbon mavy exist as carbonate and graphite., Preliminary investigation is being made of a method for simul- to form acetylene, taneously determining carbide and carbonate which involves dissolving in dilute acid a sample containing lithium carbonate and carbide. The evolved carbon dioxide is absorbed in ascarite and weighed; the acetylene 1is bubbled through an alkaline solution of potassium iodomercurate, K, Hgl,, and titrated as the acetylide (HC=C) Hg; and the graphitic carbon is determined by combustion of the wet residue. URANIUM TREFLUOBIDE IN URANIUM TETRAFLUORIDE W, K. Miller and D. [, Manning Analytical Chemistry Division Uranium trifluoride has recently been prepared in anticipation of 1ts use in esutectic fuel mixtures for nuclear reactors. Since 1ts preparation generally includes reduction of UF,, analytical procedures used to determine UF, purity must be able to separate the two fluorides. Tt was hoped to find an oxidizing agent which would oxidize U(ITT) to U(IV) with no effect on the tetravalent 1on already present. FOR PERIOD ENDING SEPTEMBER 10, Of those mild oxidizing agents which were tried, including water, ammoniacal silver nitrate, and ferric sulfate— hydrofluoric acid, only water appeared promising, although ammoniacal silver nitrate is still under investigation. The amount of UF, may also be determined by measuring evolved hydrogen by the reaction + 3UF, + 2H, 4UF, + 4HC] ——> uCl1, , The apparatus used for this is similar to that for the Dumas nitrogen determi- nation except that a reaction flask fitted with a dropping funnel replaces the Dumas combustion tube., The evolved hydrogen is swept into the gas buret by a stream of CO, Farly results are not too consistent, but it is believed that by reducing the volume of the apparatus better precision can be attained. | ' The most promising method so far considered 'is based on the total reducing power of the sample. An excess of standard ceric sulfate 1s added to a mixture of the sample and aluminum sul fate, and the excess is titrated with standard ferrous ammonium sulfate, Results obtained from this method are compared with those from the hydrogen evolution method on a single sample in Table 19.1. IDENTIFICATION OF RESIDUE IN LITHIUM-METAL COOLANT SYSTEM W. K. Miller and D, L. Manning Analytical Chemistry Division Puring a two-week period of repair one on a lithiuwm-metal coolant system, 1951 TABLE 19.1 Determination of Uranium Trifluoride by Two Methods UF-3 (%) CERIC SULFATE HYDROGEN EVOLUTION TEST METHOD METHOD 1 91.1 97.5 2 91,2 92.5 3 © 89.9 86.2 4 - 89.2 91.5 5 93.5 89.7 6 90.7 98. 1 Avg. . 90.9 92.9 portion of the system remained at the operating temperature of about 500°F while an adjacent section was cooled to room temperature., The entire system was then allowed to cool. Clusters of pink transparent crystals were found embedded in the lithium matrix in the zone of contact between the cold and hot sections. The crystals dissolved completely in water with a vigorous evolution of gas, indicating that they were probably lithium carbide or lithium hydride. A few crystals were mechanically separated from most of the matrix and submitted to the Isotopes Physics Department for X-ray-diffraction analysis. The sample was identified as a mixture of lithium hydride and lithium hydroxide and p0331b1y_ small amounts of lithium nitride. : It is probable that the LiOH was formed by hydrolysis at the surface of the sample and that freshly separated crystals are almost pure LiH. A hydrogen determination by combustion was unsuccessful, probably because%of the presence of some adhering lithium; however, it is planned to determine 203 ANP PROJECT QUARTERLY PROGRESS BREPGRT hydrogen by a gas evolution procedure when the apparatus described above under "Oxygen in l.ead" becomes availa- ble. ANALYTICAL SERVICES L. J. Brady and J. W. Rbbinsan Arnalytical Chemistry Division J. A. Norris, Isotope Hesearch and Production Division Fifty percent of the service analyses during the past quarter involved reactor fuels, 1.e., the two eutectic mixtures UFy -Nal'-KF and UF,-BeF, -NaF; 30% involved sodiumand lithium metals; and the remainder involved beryllium 204 comhnounds, rtesidues, metals, etc. There were 1781 determinations made on 310 samples, 75% chemically and 25% spectrographically. A suwwmary of service analyses performed 1s shown in Table 19.2. TABLE 19.2 Backlog Summary Samples on hand Masy 26, 1951 58 No. of samples reczived 403 Total number of samples 461 No. of sawples reported 310 Backlog as of August 10, 1951 151 FOR PERTIOD ENDING SEPTEMBER 10, 1951 20. LIST OF REPORTS ISSUED DATE REPCRT NO. TITLE OF REPORT AUTHOR( S} ISSUED Design of the ARE Y-¥27-8 ARE Operation with Especial Regard to the Coolant W. M. Breazeale 7-9-51 Circuit Y-F20-15 - Detection of Leaks in the Fuel Elements by Means W. K. Ergen 7-10-51 of Radicactive Tracers Y-F10-57 Activity of Fission Products and Heavy Elements in Glen Putnam 7-12-51 ARE Fuel ' ' ' Reactor Physics Y-¥10-56 = Simple Correction on Multiplication Constant for N. M. Smith, Jr. 5-22-51 Difference Between Assumed and Resulting Fission Distribution in Multigroup Calculations Y-F10-55 = The Contribution of the (s,2n) Reaction to the C. B. Mills 6-5-51 ~ Beryllium Moderated Reactor N. M. Smith, Jr. Y-F10-58 ~ A Discussion of Normalization in IBM Adjoint M. J. Nielsen 6-26-51 Calculations E ‘ : Y-F10-61 = 1IBM Procedure for B4C Layer Between Core and Re- C. B.;Mi]ls 6-28-51 fiector by the Coveyou Method ' Y-F10-60 Recommendation on Alternative Loading N. M. Smith, Jr. 6-29-51 Y-F10-59 The Spherical Reactor with a B,C Layer Between C. B. Mills 7-6-51 Core and Reflector Y-F10-62 The Transmission Coefficient of the B,C Curtain C. B. Mills 7-19-51 - in the ANP Reactor ' Y-F10-65 NeOH Cooled and Moderated Reactor N. M. Smith, Jr. 7-24-51 Y-F10-68 The Multiregion Reactor Problem as Applied to the C. B. Mills 8-16-51 Multigroup Method Y-Fl0-566 ; Numerical Technigque for Criticality Ca]cu]ations J. W. Webster 3-20-51 on Hydrogen Moderated Reactors : : Y-F10-64 = Heating in the B,C Cartain Due to Neutron : C. B. Mills 8-16-51 Absorption and the Blo(n,a)Li7 Reaction ‘ Y-F10-67 Effect on Radicactivity of Fleoding Coolant J. W. Webster 8-14-51 Channels with Borated Water - : Y-F10-70 © Reduction of ?eak Temperatures in Fuel Tubes R. . J. Beeley §-23-51 1205 ANP PROJECT QUARTERLY PROGRESS REPORT REPORT NO. TITLE OF REPORT AUTHOR(8) Y-F10-72 The Calculation of Eigenvalues of Differential R. R. Systems by Numerical Integration Shielding Research CF-51-5-74 Calibration of the Fast Neutron Dosimeter Used at R. G. the Bulk Shielding Facility CF-51-8-7 Suggested Program for Divided Shield Measurements J. L. and Calculation of Air Scattering R. H. ORNL~1027 Determination of the Power of the Shield Testing J. L. Reactor. I. Neutron Flux Measuremesnts in the E. B. Water-HBeflected Reactor ORNL- 1046 A Nuclear Plate Camera for Fast Neutron Spectroscopy J. L. at the Bulk Shielding Facility E. B. CF-5-8-253 Preliminary Gamma-Ray Spectrel Messurements at the F. C. Bulk Shielding Facility R. H. CF-51-8-252 Experiment 5 at the Bulk Shielding Facility — The H. E. Shadow Shield Y-F5-55 NDA Querterly Report of ANP Activities from June 1 NDA to August 31, 1951 — Divided Shield Studies Y-F5-57 The Divided Shield L. A, CF-51-6-53 The Shielding of Mobile Heactors E. P. T. A Heat-Transfer Research CF-51-8-32 Status Memorsadum on the Analysis of Heat-Transfer H. Charecteristics of the Lithium Figure-Eight System 3 Coveyou Cochran Mezem Ritchie Meem Johnson Meem Johnsen Maienschein Hitchie Hungerford Wills Blizard . ¥elton DATE ISSUED 8-30-51 5-11-51 8-1-51 8-13-51 9-14-51 8-27-51 8-8-51 8-31-51 9-17-51 To be published in Begctor Science - and - Technology Claiborne OBRNL- 1040 The Design and Construction of an Ice Calorimeter R. F¥. Bedmond J. Lones Radiation Damage ORNL-928 Physical Properties of Irradiated Plastics 0. Sisman C. D. Bopp ANP-67 Radiation Damage snd the ANP Reactor L. P. Smith 206 8-6-51 8-27-51 6-29-51 7-25-51 REPORT NO. Y-F31.2 Y-B4-16 Y-F31-3 CF-51-8-256 Y-F23-5 Y-¥30-1 Y-F8~22 Y-F31-4 NDA Docu-~ ment HKF-112 Y-B31~260 FOR PERIOD ENDING SEPTEMBER 10, 1951 TITLE OF REPORT Metallurgy Cleanliness of Sodium Circuits Literature Search on Metal-Ceramic Materials Techniques for Molybdenum Plating from Carbonyl Vapor of Solution Literature Survey on Columbium Experimeutal Engineering Tests Required for Beactor Components Report on Test of Extinguishing Agents for Lithium Metal Fires Preliminary Engineering Study of NaQOH Ccoled and Moderated Reactor Testing and Examination on Thermal-Convection Loops Operated with Lithium and Lead Alternative Systems Estimated Divided Shield Weights for Supercritical Water Cycle ' Circulating Moderator Reactor for Subsonic Aircraft Miscellaneous Analytical Chemistry—ANP Program Quarterly Progress Report for Period Ending May 31, 1951 AUTHOR(S) E. . Miller E. P. Carter E. C. Miller W. C. Hagel G. A. Cristy W. C. Tunnel] D. F. Salmon E. 5. Wilsen P. O Nadler B. W. Schroeder . Day . Brasunas > o m NDA H. K. Ferguson Co. DATE 1SSUED 7-4-51 7-11-51 7-27-51 8-21-51 6-13-51 6-18-51 8-9-51 8~20-51 6-8-51 9-1-51 Analytical Chemistry 5-31-51 Division 207