MATRRTAET 3 445k 038388Y 5 ORNL DOCUMENT REFERENCE LIBRARY, Y-12 BRANCH LOAN COPY ONLY Do NOT transfer this document to any other person, If you want others to see it, attach their names, re- turn the document, and the Library will arrange the T as requested. #) T IL»] £ ORNL-1816 This document consists of 177 pages. Copy ?3 of 218 copies. Secwimsrimm Contract No, W-7405-eng-26 AIRCRAFT NUCLEAR PROPULSION PROJECT QUARTERLY PROGRESS REPORT For Period Ending December 10, 1954 W. H. Jordan, Director S. J. Cromer, Co-Director R. I. Strough, Associate Director A, J. Miller, Assistant Director A. W, Savolainen, Editor DATE RECEIVED BY INFORMATION AND REPORTS DIVISION (JANUARY 6, 1955) DATE ISSUED JAN 21 1955 OAK RIDGE NAT!ONAL LABORATORY e “Operated by CARBIDE AND CARBON CHEMICALS COMPANY © A Division of Umon Carbide and Carbon Corporation ' Post Office Box P QOak Ridge, Tennessee (EARATARTIIR 3 4yuhsk D383884 5 e 1. G. M, Adamson 2. R. G, Affel 3. C. R. Baldock 4. C. J. Barton 5. E. S. Bettis 6. D. S. Billington 7. F. F. Blankenship 8. E. P. Blizard 9. G. E. Boyd 10. M. A. Bredig 11. F. R. Bruce 12. A. D. Callihan 13. D. W. Cardwell 14. J. V. Cathcart 15. C. E. Center 16. G. T. Chapman 17. R. A, Charpie 18. G. H. Clewett 19. C. E. Clifford 20. W, B. Cottrell 21. D. D. Cowen 22. S. Cromer 23. R. S. Crouse 24. F., L. Culler 25. L. B. Emliet (K-25) 26. D. E. Ferguson 27. A. P, Fraas 28. J. H. Frye 29. W. T. Furgerson 30. W.R. Grimes 31. E. E. Hoftman 32. A. Hollaender 33. A. S. Householder 34. J. T. Howe "~ 35. R. W. Johnson 36. W. H. Jordan 37. G. W. Keilholtz 38. C. P. Keim 39. M. T. Kelley 40. F. Kertesz 41. E. M. King 42. J, A, Lane 43, C. E. Larson 44, M. E. LaVerne ORNL-1816 Progress INTERNAL DISTRIBUTION 45, 46. 47. 48. 49. 50. 51. 52. 53. 54, 55, 56. 57. 58, 59, 60. 61. 62. 63. 64. 65. 66. 67. 68. 9. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84-93. 94-133. 134. 135-137. . Livingston . Lyon . Maienschein . Manly . Mann . McDonald . McQuilken Meem ' . Miller . Morgan . Murphy . Murray (Y-12) . Nessle . Oliver . Patriarca . F. Poppendiek M. Reyling W. Savage . W. Savolainen . D. Shipley . Sisman . P. Smith . H. Snell . |. Strough . D. Susano . Swartout . Taylor Trice . Van Artsdalen . YonderLage Warde . Weinberg White . Whitman . Wigner (consultant) . Williams . Wilson . Winters . Zerby X-10 Document Reference Library (Y-12) Laboratory Records Department Laboratory Records, ORNL. R.C. Central Research Library DOE-EMEE-PETIMEMEND>OOMPITIVIO-MAPL-TECENDD WUV N-T=E=wm>o0Z0 omONwoOzzTNnoP TP TEewE “ . =} =) 138. 139. 140, 141, 142-152, 153.157. 158. 159. 160. 161-164, 165 166-169. 170-171. 172-173. 174. 175. 176-177. 178. 176. 180. 181, 182. 183. 184-189. 190-201. 202-216. 217. 218, EXTERNAL DISTRIBUTION Air Force Plant Representative, Burbank Air Force Plant Representative, Seattle Air Force Plant Representative, Wood-Ridge ANP Project Office, Fort Worth Argonne National Laboratory Atomic Energy Commission, Washington (Lt. Col. T. A. Redfield) Bureau of Aeronautics (Grant) Chief of Naval Research Convair, San Diego (C. H. Helms) General Electric Company, ANPD Glen L. Martin Company (T. F. Nagey) Knolls Atomic Power Lockland Area Office Los Alamos Scientific Laboratory Materials Laboratory (WADC) (Col. P. L. HI”) National Advisory Committee for Aeronouhcs Cleveland North American Aviation, Inc. Nuclear Development Associates, Inc. Patent Branch, Washington Powerp[anfLaborofor'y (WADC) (A. M. Nelson) Pratt & Whltney Aircraft Division (Fox F’r0|ect) USAF Project Rand USAF Headquarters Westinghouse Electric Corporation (Bettis Laboratories) Wright Air Development Center (Lt. John F. Wett, Jr., WCOSI-3) Technical Information Service, Oak Ridge Division of Research and Medicine, AEC, ORO Atomlc Energy Commisswn - Eas’r Hartford Area i e e e ORNL-528 ORNL-629 ORNL-768 ORNL-858 ORNL-919 - ANP-60 ANP-65 'ORNL-1154 ORNL-1170 ORNL-1227 ORNL-1294 ORNL-1375 ORNL-1439 ORNL-1515 ORNL-1556 ORNL-1609 ORNL-1649 ORNL-1692 ORNL-1729 ORNL-1771 Reports previously issued in this series are as follows: Period Ending November 30, 1949 Period Ending February 28, 1950 Period Ending May 31, 1950 Period Ending August 31, 1950 Period Ending December 10, 1950 Period Ending Mcrch-IO, 1951 Period Ending June 10, 1951 Period Ending Sep’remrber 10, 1951 Period Ending December 10, 1951 Period Ending March 10, 1952 Period Ending June 10, 1952 Period Ending September 10, 1952 Period Ending December 10, 1952 Pericd Ending March 10, 1953 Period Ending June 10, 1953 Period Ending September 10, 1953 Period Ending December 10, 1953 Period Ending March 10, 1954 Period Ending June 10, 1954 Period Ending September 10, 1954 1L v [ .k o) FOREWORD This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL re- cords the technical progress of the research on circulating-fuel reactors and all other ANP research at the Laboratory under its Contract W-7405-eng-26. The report is divided into three major parts: . Reactor Theory, Component Development, and Construction, Il. Materials Research, and ill, Shielding Research, The ANP Project is comprised of about 400 technical and scientific personnel engaged in many phases of research directed toward the achievement of nuclear propulsion of air- craft, A considerable portion of this research is performed in support of the work of other organizations participating in the national ANP effort. However, the bulk of the ANP research at ORNL is directed toward the development of a circulating-fuel type of reactor. The effort on circulating-fuel reactors was, until recently, centered upon the Aircraft Reactor Experiment. This experiment has now been completed, and the operating ex- perience is described in Section 1 of Part |, The design, construction, and operation of the Aircraft Reactor Test (ART), formerly the Circulating-Fuel Aircraft Reactor Experiment (CFRE), with the cooperation of the Pratt & Whitney Aircraft Division, are now the specific long-range objectives. The ART is to be a power plant system that will include a 60-Mw circulating-fuel reflector-moderated reactor and adequate means for heat disposal. Operation of the system will be for the purpose of determining the feasibility, and the problems associated with the design, con- struction, and operation, of a high-power, circulating-fuel, reflector-moderated aircraft reactor system. The design work, as well as the supporting research on materials and problems peculiar to the ART (previously included in the subject sections), is now re- ported as a subsection of Part |, Section 2, ‘‘Reflector-Moderated Reactor." ‘ ‘V’ifl T "~y L CFOREWORD v o et eeeeen e e e e e e e o SUMMARY & i i ettt i i it e i . e e e PART |. REACTOR THEORY, COMPONENT DEVELOPMENT, AND CONSTRUCTION 1. CIRCULATING FUEL AIRCRAFT REACTOR EXPERIMENT ........... ... ..., Operation of the Aircraft Reactor Experiment ... ......... e e . Low-power experiments . ..ttt it ot eetoens sttt onenens e e e e High-power experiments ... v v v vt v v ittt ian ot enossotoanosssnononens Mathematical analysis of approach to critical ... ..o i, P e e b e e s Operation of ARE Pumps « v v i v it i ittt it ittt sttt et ie st s oaronnenen Loading of the ARE v 4 vttt it ittt et eeteosoneononseenenoeenenenes Anclysis of Ledk in Sodium System . . ih i v e e o st e s e ettt Preliminary Estimates of Corrosion in ARE . ..... St e e e e e 2. REFLECTOR-MODERATED REACTOR .t vvuvvuennenunnnnn, P .. Expunsion-Tank;n’d'Xenon?Removal System ........ e e e e e Control Rod Desrigrn ConSIerations « v v e v v ensneenenenennn, B Fill and Drain System « v v v e v e e nenenns et e e e e e e Design Physics v v v v v vttt i i i ittt tennnonnnnsennns e Ac'rlvcmon of the Inconel core shells Gt a e e e et e ettt e e 3. EXPERIMENTAL REACTOR ENGINEERING + vttt vvtennerorteeeneneenennes In-Pile Loop Component Development ....... e e e e e e e Horizontal-shaft sump pump . o v v vttt it ittt ittt ittt ittt sttt nnannsse Heaot exchonger .. .o v ittt ittt it i ittt st i it v et et an et nasansans Pump Development Program .« v o v v v v v vttt et et e ettt et e e, ARE-type sUMP PUMPS « v v v v v et e o v v o v e o e e e e e Ve e e Ve Large-scole pump developmenf co e -------------------------- Demgn and Operqtlon of Forced- Clrculahon Corrosmn and Mass Transfer Tests Beryihum-sodtum-lncone! mqss frcnsfer - Notura!-gas hect sourc:es for forced-curculahon Ioops oooooooooooooooooooooooooo 'Pump for heat exchcnger tests Gas- Furnace Heat Source 'Deve'l'opment ceeeas " Natural-gas burner .. ....... ettt ee e e e e Trap for Fluoride Vapors « v v v v ittt vttt ittt et o nstneeoooeesnosnsosenesss 'Hec:f, Exchanger Deve[opment oo RS e e e e e e “Sodium-cooled’ 100-kw furnace tests . e 11 1. 12 13 15 18 18 19 21 21 24 25 26 26 28 29 41 41 41 43 43 43 43 44 Fused salts in Inconel . .......... e e aeaes 44 Sedium in mulhmetcfl loops e e e ee e e T o . 45 e ~~Heat exchcnger tests . ettt e e T U 46 Header ook tost ... LoL oLl 46 46 47 47 50 50 vii 4. CRITICAL EXPERIMENTS &0t oteretereneeesenenenaneneanananennns 52 ~ Reflector-Moderated Reactor . v v v v ev e ti e oo rnsenesosoesonssasonssnsss 52 . PART Ii. MATERIALS RESEARCH 5. CHEMISTRY OF MOLTEN MATERIALS ... ovvvrinteitennenennennen, weeeee. 57 Yy Phase Equilibrium Studies « v v v v v v v it ettt i ettt eneeeenas BRI 57 Solid phase studies in the NaF-Zr ,-UF system . .....coovvivn it e 57 Phase relationships in UF _-bearing systems . .o v v v vt i iv e vvntenonnoonnoonnns - 58 Solubility of UF, in NaF-KF-LiF eutectic . ....ovvviveriiiii iy 59 UF , stabilify v v e e it ittt i ettt eneenenssossnneroonssnessnsnooessns 60 Dif\?erenfialr’rhermcl ANAlYSIS & i v ittt e it e e e e e 61 Chemical Reactions inMolten Salts .« . . v v ittt i ittt i sttt e it e 61 Reduction of FeF, by H, in NaF-ZrF, systems .. ....oooviniiiiii ., 61 Reduction of Ul':4 to UI:3 influoridemelts v v v v i s ittt et te et i e e 61 Electrochemistry of fused salts .. .. i ittt i it ie it vessonnnnnneesss 62 Stability of chromous and ferrous fluorides in molten fluorides ......... ... .cvvuy 63 Reduction of UF, by structuralmetals ... oovvviv e, 64 Production of Purified Molten FIUotides « v e v e v e vononensnnnsonsnsns e e _ 66 Electrolytic purification of zirconium-base fluorides + v v e v v v v et v n i vnronssons 66 Fluoride production facility . . v v i ittt ittt e it ittt e s e cennennsnnnens 68 Alkali fluoride processing facility ..... e it e s e e 69 Purificationof KF and RbF .o o i i i i i i it ittt iiei ittt ennenns cees e /0 Preparation of various fluorides + v v v v v v et vt v eveeenens Cesve o c et 71 - Chemistry of Alkali Hydroxides v uuveeueeuenneenrenreneensoennenesennnns 71 j Purification of hydroxides v v v it inenienietvannosononsnsonrsosnsas 71 - Effect of additives on hydrogen pressure over NaOH-Ni system . .. v vt e v et v vveanns 71 § Fundamental Chemistry of Fused Salts + v v v v ettt it i in it oo neenonoanannsnos 72 Solubility of xenon in fused salts . ........ Gt ettt et ettt 72 X-ray diffraction studies in salt systems ... v it vi v et nnennas cevees e e 72 Physical chemistry o vu e i v et innstennetnneseeessosennsssenssssnnos 73 Viscosity measurements o4 v v v oo v v s aeeoeessnossosersossssosossanas s 75 6. CORROSION RESEARCH ................. ® & & & 5 5 & # 8 2 F & & & 2 B " b . ..... V'.- 76 Thermal-Convection Loop Corrosion Studies v v v e v v v evsenovvosensonoenenss e 76 Inconel foop containing UF; in alkali-metal-base mixtures ... ... .o 0ol., e 76 Type 316 stainless steel loops containing UF, in alkali-metal-base mixtures . ... .. ce 76 Effect of induced potentials ... v i i ii ittt e essensar e 77 Hastelloy B thermal-convectionloops ...t veive i v e e ettt 77 Special alloy thermal-convectionloops .+ vttt v vt vt e v eonnenen e e e PP 4 Sodium-Beryllium-Inconel Compatibility ... ciivitie vttt sneesasnsassssnnnsss 78 Static Corrosion Studies .o e ittt it ittty e e 80 Brazing alloys on stainless steel .. ... .. o0y et he e 80 Brazing alloys on nickel .. ... .. .. t e et e e e em et e e 80 Screening tests of carbides v v v i v vttt i i e e e raee s e R 82 : g‘ Magnesium-lithium alloy inwater . ..o i ittt ittt ie it iiieeeeneenonns 86 . Rubidium Investigations « .. vuvevon. C et et e e . 87 - « [ viii < -y n ] . 'RADIATI'ON_DAMAGE . e e T R T MTR Stahc Corrosmn Tesfs e ' ~ Effect of lrrodlahon on UF -C F e | Minldture 1n- P|le Loop N Ty v eewnn i e e Fundamental Corrosion Research v v v e v ch e e b b e et i ettt Mass transfer in liquidlead « o v v oo v v v v v et et ettt st e ce e Fused hydroxides as acid-base analog systems . ......... e d e e e Flammability of alkali metal ‘solufions at high ’remperotUres e ceeees Chemical Studies of Corrosion v veveeeeeeensos et s e e s et e sttt Effect of temperature on the cortosion of Inconel melts with N|F2 additions ... .o Effect on chromium additions on fhe corrosion of Inconel and fluoride melts with NIF AddifioNS v vt v st eevesttnncancsssans bt e s et et st e e Effecf of chromlum valence state on corrosionoflnconel .+ v v v v e v v s st oo e v cvosoeos . METALLURGY AND CERAMICS +.vuvenuvennnennn, e e - Development of Nickel-Molybdenum Base Alloys ... v i i eiinanevenns coene Fabrication experiments « .o e vt e ot evovsossoonvenes et e ettt e New alloys ...... c e eesa et et s s e st et an e C e sttt Oxidation and oxidation protection + .. vv s os. Gt ettt et et e s Radiator fabrication i v v it v et oot aeneanssarosessosssstosesassssonss Welding « v ot vt vttt sttt snsosonsssnssssessosstosstossatsaroerocesss Mechanical properties studies v oevee et eonesnnoessns e e cons Special Materials Fabrication +iveeseeoeeeseesoosoesoatoaanssesoosnnsss Duplextubing v iiinennoaesen Gt i e s e e et e oo e Boron carbide shielding ...... e e e e e v e et et e s et es et Tubular fuel elements «vvvveennensas Ceees s esa e e esena Cee et Control rods v v vvvoes v e et ee s Gt et et e e i e st e Al-UO, elements for shielding experiment . .... ettt et r e e e e e e Brazing Alloy Development ... ... C et e e cesecs e a e ce s e Heat Exchanger Fabrication . . ... ... et ee et cr et e e Bery”iumoxideFUelElemenfs -oo.oocc ooooooo lol.otlnooocooooo'o;no-oooo . HEAT TRANSFER AND PHYSICAL PROPERT[ES .................... coenus Fused Salt Heat Transfer +uvvuveeensenn et ees s e e vee b e s e e s In-Pile Loop Heat Exchanger Analysis ...... s ee s s et s sasesasane s e s e nen e Free Convection in Fluids with Volume Heat Sources . ..veeeeveeenne. e v Heaf Transfer Effecfiyenes.s of Reactor Coolants v v v v vt o v v en v ensnse e e ART Core Hydrodynamics ... ..vuu ee s et . ART Heot Trunsfer » @ .» *» e . ......... g . * ’. . * L] . . L L . - ® & & 0 @ - . . ..-‘.. . * .. . .. ‘. _':-..7&.‘. ' Physmc[ Properhes Measurements . e e e e e e e e e Heat capacity « oo e e v v e v eonnss et i e e i e e e e e e e " VISCOSIi‘y ceee e s e e e e et e Bench £S5t + o v o o o v e v v o s s oo e oo oo oottt ss e oot Heat transfer calculations . v v v o v vt vttt ettt et es s vt ot aoanonosnoassssen 88 88 92 96 98 98 - 98 99 100 100 100 102 103 103 104 106 107 107 107 108 108 108 108 109 111 112 112 112 113 114 114 115 N6 116 117 L7 S 121 122 122 122 Removal of Xe!35 from Molten Fluoride Fuels « v v v v v i vnvevnenene IEERE .. .' R 124 b LITR Horizontal-Beam-Hole Fluoride-Fuel Loop .o v v v v v iv v iins ceeee e vee. 125 . Crégp and 'Sfress-Corrosion T eSS v i e v i et et ittt et i e e e e 126 - Remote Me:tdllography ..................... e e e Ry e ]27 - ‘ Mass Speétrogmphic Analyses ...ttt eri i it e e e . 127 10, ANALYT]CAL STUDIES OF REACTOR MATERIALS P DR e . 129 o :Anafy'rlccl Chemlstry of Reactor Materials « v v v e v v s vevnnenneeeneeenns e 129 Determination of uranium metal in fluoride salt mixtures .. ... . i, 129 , Determmatlon of trivalent uranium in fluoride fuels ...... e et 129 , Determination of oxygen influoride fuels .. v v e it i it it i et e 131 : -~ Determinationof sulfur .. .. i i i i i s i i e 131 : Deterin_incl_tion of fluoride in NaF-KF-LiF-base fuels ............. e veee e 132 Petrographic Investigations of Fluoride Fuels .. ....ovvinen . e e 133 - ANP Serwce Laborafory ................................... e .-_. . 133 11, RECOVERY AND REPROCESSING OF REACTOR FUEL ... vvvvenns e e 134 _Flssmn Pr_oduc? Removal v vvnin it ittt iiiiieiaanes T 7 ‘ Appliéafions of Fused Salt~Volatility Processes v v v v v v s vt v vt s oncnoeenens .. 137 Aircraftreactor fuels . oo v v ittt i i it e it i i e e i e 138 Heterogeneous reactor fuels .. ... . ... oo e e s et 138 PART Ill. SHIELDING RESEARCH . 12, SHIELDING ANALYSIS &+ v et v veeeeeeeeeneneeieneaeenenns eeee.. U3 | :; Slant Penetration of Composite Slab Shields by Gamma Rays .« o v vv v v v v v nnseannen 143 Emergy Absorption Resulting from Incident Gamma Radiation as a Function of Thickness of Materials with Slab Geometry . . . . . i ittt it i i et ieenen 144 :';: A Formu\a_fion for Radiation Injury to Include Time Effects .. .. v v v i v i iiiiiien 145 Analysis of Some Preliminary Differential Experiments .. .. ... ... 0o 146 Interpretation of Air and Ground Scattering at the Tower Shielding Facility .. .......... 147 i 13. LID TANK SHIELDING FACILITY .. \vuvintiniineentanennanennns S | Effective Neutron Removal Cross Section of Lithium , ... ... ... e 151 VGVE-ANP Helical Air Duct Experimentation . ..o vv v v esnenenonnnnnnns e 153 Reflector-Moderated Reactor and Shield Mockup Tests .. ..... it e st . 155 14. BULK SHIELDING FACILITY vttt et it eiite e eencaoannnneenes e - 156 " A Search for Short-Half-Life Nuclear Isomers in K39, Rb83, Rb87 and Zr90 e 156 , 5 15. TOWER SHIELDING FACILITY & tttttttteeeeeeeeeeeeaneeneenneenenenns 158 o} | TSF Experiment with the Mockup of the GE-ANP R-1 Shield Design . ................ - 158 ] ' R | PART 1V. APPENDIX . o : ORGANIZAT]ON CHART OF THE AIRCRAFT NUCLEAR PROPULSION PROJECT ......... 67 - E . SECRET » ¥ __«) : .-;._L,_._durmg fhe rennre‘_ ' ANP PROJECT QUARTERLY PROGRESS REPORT SUMMARY PART I. REACTOR THEORY, COMPONENT DEVELOPMENT, AND CONSTRUCTION 1. Circulating-Fuel Aircraft Reactor Experiment The Aircraft Reactor Experiment was operated successfully for over 100 Mw/hr. During this time the maximum equilibrium outlet fuel tempera- ture was 1580°F, and the maximum power level was over 2.5 Mw. The temperature coefficient of reactivity 2 min after the introduction of a pertur- bation was —5.5 x 10~3, although its instantaneous value was considerably higher. As a result, the reactor was extremely stable while operating at power, and there were no control problems. The critical mass was 32.5 Ib of U235 and thus was in reasonable agreement with the calculated value of 30 ib. 7 Operation of the process systems and their instrumentation was exceptionally trouble-free during the six days of operation rom the time the ‘equipment was sealed in the pits until the sodium and the fuel were dumped. It is believed that most of the fission-product - gases evolved from the liquid fuel. In a 25-hr run, it was determined that no more than 1 part in 30 of the xenon poison remained in the fuel. An unexpectedly steep initial slope that was observed in the usual plot of [1 = (1/multiplication constant)} was found to be due to the formation of a_thermal-flux maximum near fhe flssaon chombers\_“ "“dumps in heat exchanger fes'rs.' Imenswe work has been done on the development of a hydraullc “:'cn'cun fhat prowdes a fuel pump, a “fuel expansion __tank, an& a xenon-removcl system clnd that performs _the varlous funct:ons requ:red ‘of the sys’rem. ) lnmal fesfs of a “Lucite f]ow model have given “oromisingresults. Des;gn criteria were esfabl:shed_» - f'for the contro| rod fho will .as the mulhpllcahon increased. Once the shape - ‘of fhe spct;al distribution of the thermal-neu’rron:‘%_’\' flux was set up, the flssuan chambers reglsferedi“f '_,,ron!y ‘the generci increase in flux |eve| and the “Tf:,‘;‘:kcoun’r ra’re mcrecsed slow!y ' 'rransmlflmg subjected, it was interesting to note that all t ;_-'_.‘for remot power runs. Smce this was the hlghest “ambient temperature to which ARE pumps have been 'systenis. rotary elements functioned satisfactorily. Fuel samples were withdrawn from the ARE during operation below criticality and at very low power. As was anticipated, the chromium content of the fuel increased after each addition of fuel concentrate. Just before the first additions of fuel concentrate were made, the chromium concentration of the carrier material was 100 ppm and the carrier had been circulated for 155 hr. Since the system had been operating isothermally, the chromium content of 100 ppm was considered to represent 0.5 mil of attack on the Inconel. By the time the final sample was taken, after a total of 307 hr of operation, the chromium content was 445 ppm, which represents about 5 mils of attack. A plot of the data obtained showed that the chromium content of the fuel had started to level off. Since it was not possible to sample the fuel during full power operation, no information on mass transfer can be obtained until the equipment can be sec- tioned. 2. Reflector-Moderated Reactor The reactor assembly and the installation for the Aircraft Reactor Test (ART), formerly the Circulating-Fuel Reactor Experiment (CFRE), have been the subject of analytical layout and detailed design studies, Component development tests are proceeding as scheduled, and specifications have been complefed for radiators to be used as heat r be prmcapally a means for confrolhng ‘the meda ing formu]ated Desngn work 'is Under way on @ f[l@;and drain sys’rem “for the reactor assembly that will provlde a qunck disconnect coupling suitable’ operahon w:th hlgl’l-iemperafure llqwd An analysis has been made of the activation of the Inconel core shells that would come from two péi’qture level of the h:mrecc’ror fuel. Plans for coohng the control rod are oo i s ANP QUARTERLY PROGRESS REPORT main sources: the activity of the cobalt in the Inconel and the activity of the fission fragments that strike the Inconel and remain there. The analysis showed that little would be gained by trying to obtain cobalt-free Inconel, since the activity could not be reduced below that caused by the fission fragments. Adjustments are being ‘made in the constants used in multigroup calcu- lations in order to decrease the disparity between the calculations and the results of critical experi- ments. ' The various types of facilities suitable for the “operation of the ART have been considered, and it is now planned to install the reactor assembly in a double-walled container with a water-filled annulus. The type of container was selected after careful con- sideration of the hazards involved in the operation ~ ofthe reactor. This typeof installation provides for containing the products resulting from a nuclear accident and/or chemical reaction of all combusti- bles in the installation, minimizing the likelihood of serious damage from an explosion caused by sabotage or bombing, and removing and disposing of volatile fission products evolved during the course of operation, 3. Experimental Reactor Engineering Progress has been made in the design and testing of components for a horizontal, entirely enclosed loop for insertion in an MTR beam hole. The loop will be used to circulate proposed reactor fuels in the MTR flux so that radiation effects on the fuel can be studied. A layout arrangement of the loop has been made and is being improved consistent with specification and component changes. Tests of two horizontal-shaft sump-type pumps were made with the fluoride mixture NaF-ZrF, circulating at a temperature of 1350°F for 500 and 1000 hr, respectively. The Graphitar-lapped steel face plates wore considerably in both tests, but they did not fail. A single-tube salt-to-air heat ex- changer that will fit inside the beam hole has been fabricated. - Sump pumps of the type used in the ARE are being evaluated for application in large-scale heat exchanger tests, and pumps for ART use are being developed. One ARE-type sump pump is being tested at K-25 for performance and life determina- tions. The pump is operating at pump speeds of " up to 1500 rpm at a maximum capacity of 40 gpm, ~and it is circulating the fluoride mixture NaF-ZrF,- UF, at a maximum pump inlet temperature of 1350°F. Operating time now exceeds 3000 hr, The pumps for the ART inherently possess many imposing design problems because of the large capacity requirements: fuel pumps, 650 gpm, 50-ft head; sodium pumps, 430 gpm, 125-ft head; NaK pumps, 2800 gpm, 280-ft head. Investigations of impeller cavitation characteristics and inlet and entry conditions are being made, along with studies of sealing, cooling, lubricating, and driving problems, _ Forced-circulation corrosion and mass transfer tests with fused salts in Inconel systems are under way. Considerable difficulty has been experienced with failures of the thin-walled tubing (l/z-in. oD, 0.020~in. wall thickness) at bends and welded joints. The loop design has been altered by elimi- nation of the economizer and by using 0.045-in.- wall tubing. One loop of the modified design is now being operated and has accumulated about 150 hr at a Reynolds number of 10,000 and a tempera-~ ture gradient of 200°F. A third beryllium-sodium-Inconel mass transfer test has been completed following 1000 hr of operation. The maximum operating temperature (at the beryllium section) was 1300°F, and the Reynolds number was about 190,000. Another similar loop has been fabricated and is being assembled. A loop in which sodium will be circu- lated in type 316 stainless steel is being fabri- cated. The heat exchanger test program has included a header leak test and a series of pump tests with an ARE moderator-coolant type of pump. The header leak test (fuel to NaK) showed that self- plugging of an NaF-ZrF ,.UF , fuel through 0.002- in.-dia leaks can occur. The pump tests showed that speeds of up to 3700 rpm could be safely achieved with this type of pump. Developmental work on a gas-furnace heat source that can be used in large heat exchanger tests was continued. A sodium-cooled 100-kw furnace was tested for about 120 hr before a gross leak termi- nated further operation. An output of 85 kw was obtained at a sodium outlet design temperature of 1500°F. Two natural-gas burners, based on @ design development of the Esso Marketers, were built and tested. Heat releases of 400 kw at 3100°F were achieved with the first burner, and 1 Mw ot 3300°F was obtained with a second burner that was improved on the basis of the daia obtained from the first burner. ol e i « N A trap for fluoride vapors was developed that can be used in any high-temperature molten fluoride system to prevent plugging of the gas lines leading to or from the vessel containing the fluoride, particularly if the salt contains ZrF ,. Successful operation of this device was achieved during the ARE pump tests at K-25. ' 4. Critical Experiments The second critical assembly of the reflector- moderated reactor was constructed that incorporated a beryllium island. Criticality was achieved with 4.66 kg of U235 at a density of 0.092 g of U235 pe cubic centimeter of fuel annulus. This Iocding was shown, by calibrated controf rods, to contain 0.92% excess reactivity, equivalent to about 0.31 kg of U235, and therefore the critical mass of the unpoisoned assembly was 4.35 kg of U235, The experimental results thus did not agree well with the calculated critical fuel loading of 7.16 kg of U_235 or 0.142 g of U235 per cubic centimeter of fuel annulus. A series of measurements of neutron and fission-rate dls’rrlbuflons and of reactivity _coeffuments will be made prior to the incorporation, in the structure, of additional materials to simulate reactor components. PART Il MATERlALS RESEARCH 5. Chem:stry of Molten Mufermls Additional data obtained in solid phase studles of the NaF-ZrF ,-UF, system have essentially confirmed the previous findings. There was some indication from thermal analysis of mixtures along the Na_U,F ., -Na Zr ¢F3y ioin that these com- ponents do not form a complete set of solid solu- ‘tions. A new quenchlng appcratus is bemg used for the studies that is expecfed to’ ‘accelerate 'rhe ~‘rate ‘at which data can be obfamed Sfudles of ’rhe;_\ _ fi-,_;phase relahonshlps m UF -bearlng systems haved EE ?_con’rlml S in the NaF KF LIF eufectic to provide a usable R SYSfem for the reflector-moderated circulating- el recctor. EXberl ,Of UF as 0 fun‘ ' P S a.wn ‘that UF is more stable at elevated’ temperatures than eshmofes of its thermodynamic properties had indicated. It was also fournd in studies of \ o d, nd work was s'rarted on NaF-LaF, and - . ‘KF- LGF “systems, since Lul':3 is known to be a “stable ¢ stand-m ' for the more drfflcultly handled UF,. Experlmem‘al evudence was obtained which lndicc'red that UF is probcbly sufflmenfly soluble PERIOD ENDING DECEMBER 10, 1954 UF ; mixed with KF and heated in Inconel ccp'su.les that disproportionation of UF, occurred at a lower temperature than would have been expected in the absence of KF. For differential thermal analysis of the uranium-bearing mixtures, improved graphite containers with provision for the thermocouples to be immersed directly in the melt are now being used, Investigations of the chemical reactions in molten salts are continuing in an effort to under- stand the mechanisms involved in the purification of fluoride mixtures and in the reduction of UF to UF, in fluoride melts. Similar studies of the corrosion products of Inconel in fluoride mixtures and of the reduction of UF, by s’rructurai metals are under way. Production of purified zirconium-base fluorides mixtures has been continued on a large scale in the 250-lb-capacity facility. |t is anticipated that by full-time operation of this facility by a crew of one engineer and eight technicians, a sufficient stock pile of material will be available by December 31, 1954 to allow termination of this operation for some time, This stock pile {to be ~7000 Ib) should be sufficient to meet the requirements of Pratt & Whitney Aircraft, other outside requesters, and those of the Laboratory for research in this field. The pilot-scale processing facility continues to be used for small-scale development of purification processes for new fuel compositions of interest. Considerable effort has been expended on attempts to prepare UF ,-bearing mixtures of NaF-LiF-KF. Careful control of operating conditions has allowed production of sufficiently consistent UF j-content ""m'crz"fér‘i'dl to permit its release for corrosion testing. E[ectrolyhc purlflcchon of NaF ZrF mixtures hcs made poss:ble a &gmficam‘ cut in processmg fime of these ma'rerict!s._ Mechnc impurities have been found to be removed by electrolysis in a fraction of the time requured by hydrogen strlppmg It has not been possdaie to cut the processing “time of UF -bearing mixtures by using electrolysis, bfb'bd‘b[y'bkecause of the simultaneous reduction of 4 1o UF at the cathode and oxidation of UF 'cn‘ the anode. ‘The tentative values previously obtained for the solubility of xenon in fused salts were essentially “confirmed by data obtained with new, improved apparatus, The new values were 8 x 10~8 and 9.3 x 10~ 8 mole of xenon per milliliter of melt at T e ANP QUARTERLY PROGRESS REPORT 260°C and 8.9 x 10~8 and 9.6 x 10~8 at 450°C. X-ray diffraction studies of the binary systems of alkali fluorides with uranium trifluorides are under way, 6. Corrosmn Research Stud:es of the corrosion of Inconel, type 316 stainless steel, Hastelloy B, and several special alloys when exposed to the fluoride mixtures of interest were continued through the use of thermal- convection apparatus. In the studies of Inconel loops containing UF, in alkali-metal-base mixtures, considerable difficulty was encountered in making the mixtures, in controlling the total uranium con- tent, and in determining the ratio of UF, to UF,. Conflicting data were obtained, and efforts are now being made to achieve better control of the variagbles involved. An alkali-metal-base mixture ~ containing UF, was also circulated in type 316 stainless steel loops, and, as in the Inconel loops, ‘considerable mass transfer occurred. Thermal-convection loops constructed of Hastel- oy B were operated with NaF-ZrF -UF (50-46-4 moie %), with NaF-KF-LiF (11 5-42 0- 46 5 mole %) containing 12 wt % uranium as UF,, and with sodium as the circulated fluids. The loops which circulated the fluoride mixtures showed only small amounts of corrosion — 1.5 to 3 mils, The period of operation of the loop has not had a noticeable effect on the corrosion; therefore it is felt that the small amount of corrosion that occurs takes place in a short time and may be a function of the condi- tion of the original surface. The Hastelloy B loops operated with sodium exhibited considerable mass transfer. Since the loops were known to be covered with an oxide deposit and were not cleaned before they were filled, additional loops are to be operated to determine whether the mass transfer was caused by contamination. Loops constructed of several modified Inconel-type alloys were operated with NaF-ZrF ,-UF,. The data obtained from these loops indicate that reduction of the chromium content of the alloy to 5% or less greatly reduces corrosion. ~Tests of the compatibility of sodium, beryllium, and Inconel in a system have indicated that at 1300°F the flow rate of the sodiumand the spacing -~ between the beryllium and the Inconel are the con- trolling factors, Beryllium surfaces exposed to flowing sodium are unattacked, but surfaces ex- posed to relatively stagnant sodium at 1300°F show dissimilar metal mass transfer between the beryllium and the Inconel. Tests are under way to determine the minimum spacing that can be tolerated between beryllium and Inconel when exposed to slow-moving sodium, Several type 304 stainless steel T-joints brazed with experimental alloys prepared by the Wall Colmonoy Corporation were fested in sodium and in NaF-ZrF ,-UF ,. Most of the alloys had ‘good resistance to the fluoride mixture, but only the 10.2% P-13% Cr-76.8% Ni alloy had good resis- tance in both sodium and the fluoride mixture. Several T-joints of A-nickel were also brazed with various alloys and tested in the fluoride mixture NaF-ZrF -UF , and in sodium hydroxide. All the alloys tested hod good resistance in the fluoride mixture, except the 69% Ni-20% Cr-11% Si alloy, whereas all the alloys had poor resistance to sodium hydroxide, except the 82% Au-— ]8% Ni alloy. The carbides of titanium, zirconium, chromium, and boron were corrosion tested in fused fluorides, in sodium, and in lithium. For the most part, these carbides showed fair resistance in the various media, except B,C, which was rather severely attacked in sodium and in lithium. The agueous corrosion resistance of an 80% Mg-20% Li alloy which has been proposed as a crew-compartment shielding material is being studied. Dynamic tests of the resistance of Inconel to attack by rubidium are under way. Preliminary results showed maximum attack in the hot leg to a depth of 1 mil, Previous tests on the mass transfer character- istics of container materials in liquid lead indi- cated that alloys in which intermetallic-compound formation was possible showed a marked increase in resistance to mass transfer as compared with the pure components of the materials. Additional tests of this hypothesis have now been conducted. The work on fused hydroxides has centered on the development of a systematic chemistry of these substances. By an application of an acid-based theory, at least 12 types of acid-base analog reactions are predicted for the fused hydroxide systems. Each of these types is briefly discussed in the text. Work on the flammability of alkali- metal solutions at high temperatures has continued, and additional data have been obtained for bismuth- rich alloys with sodium which provide a more com- plete picture of the reactivity of this system. The effect of water vapor on the reactivity of these alloys has also been studied, L ‘ ‘f;‘k ¥ R In further chemical studies of corrosion, informa- tion was obtained on the effect of temperatures and of chromium additions on the corrosion of Inconel by fluoride melts with NiF, additions. 7. Meta"urgy and Ceramics The nickel-molybdenum-base alloys are being studied extensively as possible reactor structural materials with qualities superior to those of Inconel. Attempts are being made to improve the ductility and the fabricability of commercial Hastelloy B through purification and to find another suitable and improved nickel-molybdenum-base alloy that has the strength and corrosion resistance of Hastel- loy B. The results of attempis to improve the fabricability of the materials are being evaluated by extrusion experiments because extruded seamless tubing will be required if any of these alloys are to replace Inconel as a reactor structural material, Temperature-cycling tests in air were performed with Hastelloy B and other nickel-molybdenum alloys to determine the behavior to be expected in an NaK-to-air heat exchanger such as will be required for aircraft application. These tests have shown that oxidation protection will be required. Three radiator test specimens have been con- structed of Hastelloy B for studying the fabrication methods and for determining the thermal shock and oxidation resistance of a complicated assembly, The specimens contained 20 tube-to-header joints, which were inert-arc welded by using semiautomatic welding equipment, and 3 in. of Inconel-ciad copper fins spaced 15 fins per inch and brazed to the tube with Coast Metals alloy No. 52. The individual fube—to-header ]omfs were buck-brazed The fcbrl-___ "f)these test ussembhes was rouhne, and '..f'no unusuol dlfflcuihes were encountered Prlor: . to evoluatlng the properhes of Hostelloy B weldl‘ixs"l "f.-f»_'-_-:."metal “the sfrengfh ‘and ducflllty properhes offl'” ca’rlon commercually avmlabie Hcs’re”oy B plate were .anneohng and (2) solution cnnealmg and aging cn‘h.’ 1950°F, These » very prellmmary experumen'rs hqve ~shown that the room-fempera’rure duc'rlln‘y can be greatly lmproved by an aging freatment prior to 7 service at 1500°F, S ; P "ma (::Hl‘ escnd o g Six new 'Iv on the creep properties of Hastelloy B, Inconel, and other alloys for use at high temperatures. “additional” 'rube-burst units _Have been mstol[ed for studying the effect of Hluoride mixtures and sodium PERIOD ENDING DECEMBER 10, 1954 Several of the new nickel-molybdenum alloys are being tested to determine their load-carrying abili- ties, : A study of the flow of metals during an impact extrusion is being made as a part of the efforts to produce duplex seamless tubing that will have good oxidation resistance on the outer surface and good corrosion resistance on the inner surface. Several suitable B ,C-containing compositions have been developed for the shield for the ART heat exchanger; various nonmetallic bonding ma- terials can be used. Thirty-five control rods are being prepared for the GE-ANP project that contain a mixture of 50% aluminum powder and 50% B,C. Fuel plates are being prepared for a study of the effect of delayed neutrons on the over-all shield weight of circu- lating-fuel reflector-moderator reactors, A finned surface for a fused salt-to-air heat exchanger being considered for use in conjunction with an in-pile forced-circulation loop was fabri- cated and tested for corrosion resistance to the fused salt., The corrugated fins were formed from 0.010-in,-thick nickel sheet and were brazed to a 3/a-in.-OD Inconel tube with an 82% Au-~18% Ni brazing alloy. A 0.020-in.-dia nickel wire was used as a spacer between each fin segment, and it served to provide a capillary path for the braze alloy, which was preplaced at the fin interlock joint only. The braze alloy used was found to have excellent resistance both to air and to the fused salt. Beryllium oxide ceramics were produced, proba- bly for the first time, by casting from a basic slip. ’Hifhen‘o an aud had been considered to be neces-_ “sary to assure a well- defloccuhfed ‘slip. This “":work is belng done in an efforf to prepore berylllum ';oxlde fuel elemenfs. 8. Heat Transfer cmd Physlcul Properties | Preliminary forced-convection heat transfer " experiments with the ARE fuel, NdF;ZrF4¥UF4 "~ (55.5-40-6.5 mole %), in an Inconel tube, yielded heat transfer coefficients which differed from expected values by only 24% after 115 hr of oper- ation at abou'r T300°F no corrosion deposits were “observed on the inner tube wall at the end of the eXperlment., The performance charqcterlshcs of an “in-pile loop heat exchanger were determined. The mathematical velocity and temperature distributions for several free-convection volume-heat-source ANP QUARTERLY PROGRESS REPORT systems were compiled, and apparatus was de- veloped for measuring the over-all radial tempera- ture difference (or heat transfer coefficient) for a modified version of the flat-plate system in which it will not be necessary to measure fluid temper- atures. _ " Several approximate mathematical temperature "soiuflons for forced-flow volume-heot-source en- fronce systems were developed which can be used to predict the temperature structure of struts or screens located in circulating-fuel reactor flow passages. The question as to whether electric currents, which generate heat in the circulating fluids of experimental volume-heat-source systems, affect the fluid flow characteristics was investi- gated. It was found that the hydrodynamic structure was not mfluenced by the presence of the electric currents. The enthalpies and heat capacities of NaF- -ZrF - (53-43-4 mole %) were determined; the heat _ éqpacity in the solid state over the temperature ~range 70 to 525°C was found to be 0.18 cal/g.°C, and the heat capacity in the liquid state over the temperature range 570 to 885°C was found to be 0.26 cal/g-°C. The enthalpies and heat capacities of LiF-KF-UF, (48-48-4 mole %) were also ob- tained; the heat capacity in the solid state over the temperature range 125 to 465°C was found to be 0.234 + (0.95 x 10=4) cal/g-°C, and in the liquid state over the temperature range 565 to 880°C it was found to be 0.657 - (3.93 x 10~4); cal/g-°C. The viscometry equipment used earlier was modified so that the accuracy of liquid vis- cosity measurements was significantly increased. The more accurate measurements obtained for i\iaF‘-ZrFfl-UF4 (53.5-40-6.5 mole %) were 30% lower than the preliminary values reported previ- ously. The thermal conductivity of molten NaF- ZrFA-UF4 (53.5-40-6.5 mole %) was found to be 1.2 Btu/hr-ft.°F. The thermal conductivities of the alkali fluoride mixture NaF-KF-LiF with and without UF, were compared in the liquid and solid states. - 9. Radiation Damage The program of MIR capsules containing fused fluoride fuels has been continued, and irradiated capsules containing both irradiations of Inconel - UF,- and UF ,-bearing mixtures have been ex- -3 ) . " amined. There was practically no corrosion of the irradiated UF j-bearing capsules nor of one of the irradiated UF ,-bearing capsules. There were no significant differences in the iron, chromium, or nickel contents of the irradiated fuels, as com- pared with the starting fuel batches, and there was no evidence of segregation of either uranium or impurities, An improved version of the capsule irradiation facility has been put into service, and, to speed up the program, arrangements have been made for irradiating two capsules simultaneously in separate, but adjacent, facilities in the MTR. Examinations of a welded nickel capsule con- taining a UF ,«C,F | solution that was irradiated in the ORNL Graphite Reactor indicated the un- suitability of the solution for use as fuel in a mockup of a circulating-fuel reactor. A miniature in-pile loop designed for insertion in a vertical hole in the LITR was bench tested and was found to be satisfactory. The bench test included four freezing and melting cycles in 260 hr of operation at a maximum temperature of 466°F, Since it was possible to freeze and melt the fused salt without causing failure of the loop, it may be advisable to fill the in-pile loop before it is in- serted in the reactor and then to melt the fuel mixture after the loop is in position, A Delco motor has been rebuilt to withstand radiation and the high temperatures to be used for the in-pile joop. Heat transfer calculations that predict the thermal behavior of the loop were completed. A second model of the loop for insertion in the LITR is being constructed. Plans were completed for an in-pile study of the removal of xenon from molten fluoride fuels under ART conditions, The horizontal type of fuel-circulating loop designed for irradiation in the LITR has been operated out-ofpile with a non-uranium-bearing salt and is now being inserted in the HB-2 facility of the LITR. The creep-test apparatus for testing Inconel at high temperatures in the MTR is being bench tested. The stress-corrosion apparatus for LITR operation has been successfully bench tested, and an in-pile apparatus is being constructed, Remote metallographic studies of solid fuel ele- ments were continued, and additional information on the relationship between UO, particle size and radiation damage was obtained, 10. Analytical Studies of Reactor Materials Methods were developed for the determination of uranium metal and UF3 in NaF-KF-LiF-base 1) ) reactor fuels. Uranium metal is determined by converting it to UH, with hydrogen at 250°C, then increasing the temperature in an atmosphere of carbon dioxide to 400°C to decompose the hydride, and finally measuring the volume of gas liberated as a consequence of thermal decomposition. Under these conditions, CO, was reduced to CO by urani- um metal and also by UF,. A trap of 1,0, was incorporated to oxidize CO and thus remove this source of error, A solution of methylene blue was found to oxidize trivalent uranium quantitatively to the tetravalent state without liberation of hydrogen. A procedure in which methylene blue is used as the oxidant was developed for the determination of trivalent uranium in a variety of materials, The method appears to be applicable to routine analysis. Two other reagents, cupric chloride and titanium tetrachloride, will, under selected conditions, oxidize trivalent uranium to the quadrivalent state only. The latter reagent can probably be adapted to an automatic coulometric titration procedure for this purpose, ' Calibration of the apparatus for the determination of oxygen as oxide in fluoride reactor fuels was completed for quantities of oxygen up to 235 mg/liter, In this range the relationship log k/c = A\Jc + B isvalid; k is the specific conduc- tivity of water in HF, ¢ is the concentration, and A and B are constants, A colorimetric method was adapted for the de- termination of sulfur as sulfate or sulfide in fluoride salts. The sulfur is used to form mefhylene blue, an intensely colored dye, for which the absorbancy is readily measured. The method was also applled_ to fhe defermmcmon of sulfur m sodlum. A Semi-i_"j_"_‘ in off-gas streams wds set up ‘that is based on the ""rurbldn‘y of bismuth sulflde in a buffered solution, -~ The specfrophofometrlc htra’non “of fluorlne |nf:" ~ fluoride fuels w1'rh a zwcomum ‘complex was in- “vestigated ‘and was found fo ‘be unschsfacfory for routine application, F’etrogrcphtc examinations were made of severalq' ~ hundred samples ‘of fluonde melts, Most of the " samples were from “alkali fluoride systems “eon- The Anolyflcai ‘Service Laboratory” taining UF .. reported 1587 samples which involved 11,541 de- terminations, PERIOD ENDING DECEMBER 10, 1954 11, Recovery and Reprocessirfig of Reactor Fuel A plant for recovering (in seven batches) the uranivm from the ARE fuel and rmse by the fluoride-volatility process is being demgned and construction is scheduled for completion by December 31, 1955, It is estimated that the amount of material to be processed will be 12,4 #3 of NaF-ZrF ,-UF, containing 65 kg of uranium, This plant will demonstrate, on a pilot-plant scale, the feasibility of the fluoridesvolatility process as applied to the processing of a circulating-fuel aircraft reactor, The feasibility of the process has been established on a laboratory scale. The basic equipment, as now envisioned, will consist of a fluorinator, an absorption column packed with NaF, a cold-trap system, and a fluorine disposal unit, This method of recovery and decontamination can also be used for processing heterogeneous reactor fuel elements of the type that can be dissolved in fused fluoride salt by means of hydrogen fluoride. Compactness of plant, operation at atmospheric pressure, and economical waste disposal are some of the advantages of this type of process, PART I, SHIELDING RESEARCH 12. Shielding Analysis Calculations made by the Monte Carlo method with the use of the ORACLE were complefed for the attenuation of gamma rays in the sides of a a two-component crew shield and the heating by gamma rays in the beryllium slabs adjacent to the gamma-ray slab source. The resulis of the gamma- ray hecmng study are of interest for calculations of thermal stresses and consequenf coohng require- ments for the reflecfor _region of the CIrculatmg- fuel reflector-moderoted reactor. Attempts were made to_incorporate the Biology Planning Chart _No. 1 of the ANP Med:cal Adwsory Group and other :_l_recommenda"hbrnws into shield designs. It was demons’rroted that the chcrt could be supplcmfed by a fhree-parameter mathemahcol ‘formulation which lends considerable f]exrblll’ry in the oppllccmon of _the tolerance limits. Anaiyses of the Tower Shield- _ing Fccnln‘y (TSF) data mclude a phenomenologlcal - ancly5|s of the behavnor of neutrons which Ieove “the reactor shleld cmd dre cnr scaflered mto the' crew shield. Relatlvely ‘simple ‘models are shown to explain some of the values of scattered flux s | incident on crew shield sides and rear. In ad- dition, more detailed calculations of the TSF data were made to estimate the fraction of scattered radiation at the maximum altitude which is attribute able to ground scattering. There is evidence that for many measurements it will be about 1%, and for highly anisotropic shields such as the GE-ANP R-1 it will be no more than about 5%. The con- clusion is reached that the TSF towers are high enough for experiments applicable to high-flying alrplcnes. - 13, L:d Tank Shielding Facility At the Lid Tank Shielding Facility (LTSF) the effective neutron removal cross section for lithium was determined as 1,01 + 0.04 barns, The meas- urements were made in a medium of oil behind a solid slab of lithium. The experimentation with _GE -ANP hehcal air ducts was continued, and measuremenfs have now been made beyond an array ‘of fhlrfy-flve 3-in.-dia ‘ducts both in a medium of water and in a gamma shield medium of steel Raschig rings and borated water. The presence of the ducts in plain water increased the thermal- neutron flux by a factor of 300; in the gamma shield medium the ducts increased the flux by a factor of 3000, Preparations for the reflector-moderated reactor and shield mockup tests are being com- pleted. It is now planned to use solid UO, as the simulated fuel in the mockups. The UO, will be mounted on a movable belt. 14. Bulk Shielding Facility The Bulk Shielding Facility (BSF) was shut down during much of the quarter for modifications to the reactor pool. These modifications will per- mit the use of demineralized water in the pool, and thus they will eliminate corrosion of the reactor fuel elements. Several BSF staff members participated in a program to discover any short- period isomeric gamma-ray transitions which might be present in K39, Rb85, Rb87, and Zr?0. Measure- ments were made at the ORNL Graphite Reactor, and no previously unknown short-period isomers were found in K39, Rb®3, or Rb®7. Zirconium-90 yielded a 2.30 £ 0.03-Mev gamma ray with a half- life of about 0.83 * 0.03 sec. This may be of importance in design considerations for a circu- lating-fuel reactor. | 15. Tower Shielding Facility Measorements made around the GE-ANP R-1 reactor shield in the TSF pool were compared with similar measurements made earlier at the BSF. The variation of the thermal-neutron intensity as the altitudes of the mockup and TSF detector tank were varied simultanecusly was also measured. The data indicated that ground-scattered neutrons are still observable at the maximum altitude for this particular reactor-shield combination. For tests at the 195-ft altitude, lead was placed in the rear of the detector tank to simulate the shielding in the crew compartment, and the thermal-neutron flux and gamma-ray dose rate distributions within the tank were determined. This experiment was interrupted for o period of 2"/2 weeks so that the TSF could participate in an Air Force Project in which a group of monkeys were exposed to massive neutron radiation doses. I‘l : W ) Part | REACTOR THEORY, COMPONENT DEVELOPMENT, AND CONSTRUCTION " wl it 0 8 ) -y o 4 © 1, CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT E. S. Bettis J. L. Meem Aircraft Reactor Engineering Division OPERATION OF THE AIRCRAFT REACTOR EXPERIMENT - The Aircraft Reactor Experiment was success- fully operated during the quarter. Uranium in the form of molten Na,UF, was added to the barren carrier, NaZrF., with which the fuel system was initially filled, to make the reactor critical. The fuel composition at initial criticality was 52.8- 41.5-5.7 mole % (NaF-ZrF,-UF,), which has a melting point of 990°F, whereas the final fuel mixture {which included excess uranium) had a ~ composition of 53.2-40.5-6.3 mole % (NaF-ZrF,- UF,) and a melting point of 1000°F, It was initially intended to remotely add the concentrate to the fuel system from a large tank which contgined all the concentrate, after first ‘passing it through an intermediate transfer tank. This system was discarded when temperature- control and continuous-weight-measuring instru- mentation on the transfer tank proved to be un- satisfactory. Instead, o less elaborate, but more direct, method of concentrate addition was em- ployed. This enrichment operation invoived the successive connection of numerous small concen- trate containers to an intermediate transfer pot, which was, in turn, connected to the fuel system by a line which injected the concentrate in the pump above the liquid level. Each of the concen- trate containers was weighed before and after a transfer in order to determine the amount of uranium injected into the system. The concentrate was supplied in batches in cans containing from about 0.25 ib of N°2UF5 (for rod calibration) up to about 33 |b (as was used during the first subcritical loading). In the enrichment operation the pump bowl served as a mixing chamber and uniformly distributed the concentrate inté the circulating stream. (For details of the loading operation see following subsection on ‘‘Loading of the ARE.”) .The first concentrate addition was made on October 30, but the reactor did not become critical until three days later (3:45 PM, November 3). Most of the intervening time was spent in clearing the end of the transfer line at the pump, which, because of limitations inherent to only this par- ticular design, was difficult to heat and even more difficult to service. The approach to criticality was carefully charted after each fuel addition. The resultant curve of reactivity [1 — (1/multiplication constant)] as a function of the addition of fuel (in terms of pounds of U235 per cubic foot) is presented in Fig. 1.1. The data from three different ionization chambers are presented; meters Nos. 1 and 2 were fission chambers located in the reflector, and the BF3 counter was located external to the reactor at the mid-plane of the cylindrical side. The unique shape of these curves (which, when first extrapo- lated, suggested a much lower critical mass than was actually required) is believed to be due to the particular radial flux distribution of the reactor at the location of the chambers and the change in this distribution as criticality was approached. The calculated volume of the carrier in the fuel system (before concentrate addition) was 4.82 3. (The only significant check on this value was -EURET ORNL-LR-DWG 3852A 1.0 = METER NO. 1 2 . 0.9 \}'4 e 7 ” Vi / 0.8 / METER / . NO.2/ > 0.7 / 2 /f 2 /- 8F, 2 os A £ / -~ =2 = > 05 T / 0.4 £ 0.3 0.2 : o 4 8 12 16 20 24 28 U CONCENTRATION (Ib/ft3) Fig. 1.1. Approach to Criticality. 11 i B g et s oo b ANP QUARTERLY PROGRESS REPORT obtained from subsequent analyses of fuel samples, together with the known amounts of concentrate added.) While the total amount of vranium (U23%) added to the system in order to make the reactor critical was approximately 135 |b, because of the amounts withdrawn from the system for samples and in trimming the pump level, the uranium concentration at criticality was 23.7 Ib/#3; or, since the calculated volume of the 1300°F core was 1.37 3, the ‘‘cold,”” clean critical mass of the reactor was 32.5 1b of U235, ~ Low-Power Experiments Several ‘‘experiments’’ were performed on the critical reactor at low power, including reactor power and rod calibrations. In addition, the effects of the process system parameters on re- activity were noted, and a preliminary measure- ment of the temperature coefficient was under- taken. The tests were started on the morning of November 4 and were completed by noon on November 8. The regulating rod was calibrated both by the addition of fuel and by a determination of the resultant pile period upon withdrawing the rod (as derived from the inhour equation). The value of the rod was first obtained by noting the amount of rod insertion required to maintain a constant power level as a finite amount of fuel was added to the system. This information, together with a calculated value of the mass reactivity coef- ficient (Ak/k)/(Am/m) of 0.232, permitted a de- termination of the value of the rod. The technique of rod calibration by pile period was also employed both at design fuel flow (48 gpm) and with no fuel flow. The data from each of these tests are presented in Fig. 1.2. Although there is con- siderable scatter, the data from the different rod calibration techniques appear to be mutually con- firmatory, and a rod value of 0.032 Ak/in, obtained from the data was used throughout the remainder of the experiment. It should be noted, however, that the period calibration with no flow is believed to give the best data, since the inhour equation is applicable without correction and the resultant value of the rod is not dependent-on a reactivity coefficient, , The reactor power was first estimated from the fission chamber counting rate, but attempts were made to confirm the estimate by operating the . SECRET ORNL-LR-DWG 3900A 0.04 A A ° | . ® A a A A 0.03 % A X X X A £ A < 0.02 x o ® 0.0 A FUEL ADDITION @ PILE PERIOD ZERO FLOW X PILE PERIOD 48gpm FLOW . ’ 0 2 4 6 8 10 12 14 ROD POSITION (in.) | - Fig. 1.2. Calibration of Regulating Rod. 12 i3 " 0 T hR) o b @ ..' R . e 4 " A ":'vented\ by“ operating the pit at s clean reactor af a low power for a 1-hr period and then withdrawing a fuel sample and taking a count of the sample. This calibration was attempted first at an eshma’red power of 1T w and then at 10 w. The fuel activity from the 1- whr run was too low for an cccurate ‘count to be made, but that from the ]0 whr run indicated a power of 13.5 w. The nuclear instrumentation was cali- brated on the basis of this power determination. t developed later that almost all the volatile, as well as the gaseous, fission products were ap- parently continuously removed from the fuel at the pump, and consequenfly the actual power was probably much greater than ‘that mdlcafed by the fuel sample. ‘Attempts were made to measure the temperature 'coefflcrent when the reactor was subcrmcql and ‘again during the low-power operation. In both instances it was estabhshed that the coefflctenr wds hegative, and, in the latter case, it was ~ determined that the magnitude was approximately 5 x 103 Ak/F, A more accurate determination of the magnitude of the temperature coefficient was deferred until the high-power runs were made. As a part of the low-power operation, the shim rods were calibrated in terms of the regulating rod. Each of the three shim rods had approxi- mately 0.15% Ak/in. for most of their 36 in. of travel. ngh Power Experiments The reac’ror was fmally taken to hlgh power (estlmated at 1 Mw from the nuclear instrumen- tation) at 6:20 PM, November 9, some six days after it first beccme cri’ncal durmg which there were perlods of opercmon ‘with power levels of ]0 ]00 end 500 kw, and fmcl!y”' -1 Mw. Power incre ined ' hcnpai‘ed merely by mcreqsmg ’rhe speed o * heat exchanger.f"j' whlch ‘cooled fhe “fuel pressures and remotely exhausting the pit gases to the atmosphere. This power level was attained after a 30—hr perlod of operohonj“ ' _._“_‘_wh[ch recor o bt ai ined. “":i.'of cpprOleOTeIY 2.5 Mw) After about 18 min at msamas'pfie'rsc* PERIOD ENDING DECEMBER 10, 1954 Once high power was attained, the reactor was operated at various power levels during the next several days, as required, to complete the desired tests, These tests included measurement of the temperature coefficient of reactivity, a power cali- bration from the process instrumentation, and a determination of the effect of large increases in reactivity, and they were concluded by a 25-hr run at full power to determine whether there was a detectable buildup of xenon. The temperature coefficient of reactivity was determined simply by placing the regulating rod on the flux servo and then increasing the speed of the blower cooling the fuel. The change of rod position (converted into reactivity) divided by . the change in the reactor mean temperature de- termined the reactor temperature coefficient. The absolute value of this coefficient was initially quite large, and it decreased after 2 min to a relatively constant value of —5.5 x 107°% Ak/°F Further analysis of the data is under way to ascertain the precise value of the instantaneous fuel temperature coefficient, which is, of course, the most important characteristic affecting the control of a power reactor. It is certain that this coefficient was considerably larger than was ex- pected and that the reactor was exceptionally stable. , In this, as in any potential power reactor, the reactor behavior as a result of large increases in either reactivity or power demand is of par- ticular interest. With a circulating-fuel reactor operating to produce power, insertion of the safety rods reduces the reactor mean temperature. The power level, on the other hcmd is controlled by fthe rate at which heat is withdrawn from the J_ACIrcula'rmg fuel ‘The effects of severol of these 'operaflons are :llustrcn‘ed graphlcol]y in Fig. 1.3, ) l,the reqcfor mle'r and outlet fuel emperaiures o er a 100-min ‘test. The 12 thermo- couple recdmgs on rhe lnle’r ‘and outlet of each ' '\‘h'efstx paralle[ fuel c1rcunrs through 'rhe reactor _ e}iown in fhls ftgure.. ‘At zero time and with "'d ‘mean femperurure of ‘about 1320°F,\ the fuel "_:her“ - shown la’rer,_corresponds to a reoctor power high power the helium blower motor was turned off, and the temperature gradient was eliminated 13 _.blower speed was graduclly increased over"' min perrod untll a reactor AT of 250‘°F_ was__” (Thls ’rempercfure difference as w1[|4 ANP QUARTERLY PROGRESS REPORT 1900 Y REDUGED SPEED 1800 SHIM ROD. SUBCRITICAL SHIM RODS 1700 TEMP. INTERLOCK BLOWER (FULL SPEED) !690 1500 1400 1300 1200 REAGCTOR INLET AND OUTLET TUBE TEMPERATURES (°F) 1100 1000 ON FUEL HELIUM {FULL SPEED) . BEOREI- ORNL-LR-DWG 3916 SPEED TO MAXIMUM|| CRITICAL: SHIM RODS FUEL HELIUM SPEED TO ZERO AT 72 Mw SLOWLY INCREASED FUEL 104 100 96 92 88 84 80 76 72 68 64 60 56 52 48 44 40 36 32 28 24 20 6 122 8 4 O TIME {min) Fig. 1.3. Power Excursions. in about 8 min. However, by this time the reactor mean temperature was 1380°F, and it was there- fore decreased (to 1350°F) by inserting the shim rods. The reactor was again brought from low power to high power in 2 min, and after 8 min at high power the shim rods were withdrawn to in- crease the mean temperature. It is most sig- nificant that during the 2-min interval required to bring the reactor from a nominal power of 100 kw to 2.5 Mw the shortest recorded pile period was only 14 sec. Furthermore, when at high power, group withdrawal of the shim rods (0.02% Ak/sec) . resulted in a pile period of only 10 sec., These ““two limiting periods were consistently repro- ducible. | Additional insight into the behavior of the negative temperature coefficient is also afforded by Fig. 1.3. By time 84 min, the blower was off and the reactor power was reduced to about 100 kw. The shim rods were then inserted to make the reactor subcritical, At 86 min, the fuel helium blower was turned on. The higher density 14 cooled fuel made the reactor critical, and in 2 min a AT of about 200°F was obtained. At this time, however, the blower speed was automatically reduced by a low-temperature signal. The blowers were subsequently shut off and the shims inserted. Although the reactor power level had been cali- brated against the activity count of a fuel sample, the actual reactor power remained in doubt through- out the experiment, not only because of uncertainty regarding the retention of fission products by the fuel but also because of discrepancies in heat balances in the process systems, that is, heat removed from fuel and sodium vs heat picked up in their water heat dumps. While the causes of these discrepancies are now being analyzed in detail, the most reliable estimate of the reactor power is believed to be that obtained from the temperature differences and flows in the fuel and sodium systems. During one typical period of power operation, the fuel AT of 370°F at 45 gpm accounted for more than 1.9 Mw in the fuel, while the sodium AT of 115°F at 150 gpm accounted 148 *y »k " tr 4 At i for more than 0.6 Mw in the sodium, which result in a total reactor power in excess of 2.5 Mw. The temperature differences quoted were obtained from the fuel pipe temperatures to and from the reactor at the time the reactor inlet and outlet core tube temperatures (as shown in Fig. 1.3) recorded a AT of only 250°F. While there are several possible phenomena which could contribute to ‘these low core tube temperatures (external cooling, surface layers, radiation, etc.}, there were 10 to 15 thermocouples on the inlet and outlet pipes which gave consistent readings. The outlet pipe temperature was at an equilibrium of 1580°F and was in excess of 1600°F during transients. The last scheduled experiment to be conducted on the reactor was a measurement of the xenon buildup during a 25-hr run at high power. The amount of xenon buildup was observed by the amount of regulating rod which had to be with- drawn in order to maintain a constant power level. However, during the 25-hr power run the regulating rod had been withdrawn only 0.3 in.,, or one- thirtieth the amount calculated on the assumption that the xenon would remain in the fuel. The scheduled tests were completed by 8:00 AM, November 12. The reactor operation was then demonstrated for all those who attended the ANP Information Meeting on Friday, November 12. During the 12-hr period of 8:00 AM to 8:00 PM, the reactor power was cycled 21 times. The resulting temperature cycling was probably as severe as that to which an aircraft reactor would be subjected. At 8:00 PM, November 12, with the scheduled experimental program completed and over 100- Mwhr of operation logged, the reactor was made subcritical, but circulation of the fuel and sodium was contmued The followmg morning, the’ fuel qnd ’rhen 'rhe sodmm were. dumped lnfo ' , thelr respecflve dump tunks. ' ' Maihemutlcal Anu|y5|s of Approach to Crmcul L W. E Klnney Alrcraff Reoctor Engmeermg DIVISlOn o When 'rhe ARE ‘was brought to critical by suc-f ” cessive fuel addlflons, it was observed ‘that the‘_ ' ~ usual plo'f of [] - (]/multlpllcohon constant)] vs “granium concen'rrahon lncreased at flrs'r very ' rapldly, but when ’rhe curve go’r close to, l the_. is observed in many reactors, but the ARE ex- hibited the effect to an unusual degree. In order PERIOD ENDING DECEMBER 10, 1954 to explain this, the ORACLE three-group, three- region code was modified so that flux shapes at successive fuel additions could be calculated. Group constants were obtained by flux weighting with fluxes from an Eyewash calculation on the ARE. Figure 1.4 shows the space distribution of the thermal flux for no fuel and for runs 2 through 6. The effect of the reflector as fission neutrons become more numerous can be seen. Figure 1.5 compares experimental and calculated startup curves for the fission chambers which were located in the reflector as indicated in Fig. 1.4. In the caleulation, where CR is the counting rate of the fission chamber, ;. is the fission cross section for group i i, and ¢ is the group 7 flux. For the ARE startup CR e where m is the multiplication constant, CR. is the counting rate on run j, and CR is the counting rate with no fuel. |t seems, then, that the unex- pected, rapid initial rise in the counting rate of the fission chamber and the {1 ~ (1/m)] curve is due not only to the general rise in flux level but also to the formation of the thermal-flux maximum near the fission chambers. Once the shape of spatial distribution of the thermal neutron flux is set up, the fission chambers register only the general increase in flux level and the count rate increases slowly. OPERATION OF ARE PUMPS W ‘G. Cobb A. G. Grindell ... W.R. Huntley Alrcraft Reactcr anmeermg Division The operation of fhe ARE pumps durmg the . prenuclear and nuci’ear phases of the experiment ‘can be regorded as successful A ’rotc] of 635 hr of operatlon at h:gh temperature was accumulated ‘ with the sodlum sys'rem pump and 462 hr with the fuel system pump The latter total includes 169 hr ~with fuel carrier, 220 hr with eritical fuel bu’r no “heat removal, and 73 hr with heat removal, The preoperation check of the ARE pumps consisted in a functional check of the seal gas- 15 T TEEE T s e e o e Sy ™t £ b el S ANP QUARTERLY PROGRESS REPORT GPeRET ORNL-LR-DWG 4450 CHAMBERS > N — B - | o . - ez B | { > é I PRESSURE 3 ] | SHELL S REGION z CORE ' REFLECTOR i | | x 2 | L | L T o | | O \0\0“"0——0——7—’1’/-0—()\ | 2 \S\?\& I 0 . 0 20 30 40 50 60 70 RADIUS (em) Fig. 1.4. ARE Thermal Flux vs Radius. balancing systems, installation of a vacuum sink on the reactor-fuel inlet-pressure-transmitter pilot valve to protect the pressure transmitter during the vacuum fill of the fuel system, installation of a gas nozzle antiplugging device on the fuel pump tank to reduce the zirconium fluoride vapor- condensate plugging problem, and a functional check and setting of low-flow alarms on water and lubricating oil systems. After about 24 hr of hot prenuclear operation, some noises from the bearing housing of the main ,'--'ff_?_____fu'eri pump were noted. This disturbance was first ““heard on the crystal noise pickups. After many increases and decrease. in intensity, the noise finally subsided, and the bearing housing per- formed quietly during the sealed pit operation. No detectable increase in drive-motor power oc- curred during the time the noises were heord. 16 Durin.g the filling of the fuel system 'cmd‘s‘ub':-'fi. sequent nuclear operation, there was some leakage from the gas space of the main fuel pump tank. This pump tank served as a fluid expansion and degassing chamber, as well as the pump sump tank. The gas leakage may have been associated with any or all compression joints (spark plug probes, Swagelok fittings), with any welded gas flange seal, or with the two compression joints in the rotary element, The water and the lubricating oil auxiliary systems on all four pumps (including two standby pumps) functioned without incident during the entire experiment. The seal gas-balancing systems on the pumps also functioned without incident. The liquid level indicating probes on the two sodium pumps and on the main fuel pump (the standby fuel pump was isolated from the fuel bk LY is, 1h . ‘)_.J. Ve wh ier ¥ “‘1{'_ PERIOD ENDING DECEMBER 10, 1954 ORNL—~LR—DWG 4451 _‘____,__.-—'O 1.0 — Rzt el /__—_—-—f:" — T e T T — P T fi\ - = ® ’/// 0.8 0.6 ; E ® CHAMBER 1 = A CHAMBER 2 | © COMPUTED 0.4 0.2 0 : 0 2 4 6 8 10 12 14 16 18 U2 CONCENTRATION (ib/ft°) Fig. 1.5. {1 - (l‘/m)] vs U23% Concentration for ARE. system) functioned satisfactorily., The maximum level probe in the main sodium pump tank shorted occasionally, but it subsequently cleared. Appli- cation of the *‘short burner’’ (high current source to separate accumulated condensate from the probe) had no effect., The maximum operating level probe in the main fuel pump tank shorted several times during early stages of hot operation. Each time, the short was cleared instantaneously by use of fhe short burner. After approxlmafe!y' five such occurrences, the circuit was arranged for continuous burning, cnd no furfher mdlcahon' of shortmg appeared 'The pump power-’rransmmmg V-belts functioned ‘sahsfocfonly durmg the entire experiment. The effect of radiation damage upon the belts may be determined upon inspection. The power traces indicated good operation of the pumps undér load. The voltage regulcmon of the pump motor supply (an’a-c~d-c motor gener-: ator set) became somewhat unsteady with passage’ of time. This unsteadiness of voltage regulchon could be seen in pump speed and motor current indications. The pump tachometer generators performed well after some initial troubles were corrected. The tachometer generator on the main fuel and main sodium pumps failed during the prenuclear hot operation. These failures, it is believed, can be traced to improper fit-up of the generator shaft into inner races of ball bearings and, possibly, to excessive tension on the V-belt. Indications were that the ambient temperature was approximately 175°F in the pump region during the nuclear and power runs. This is the highest ambient temperature fo which ARE pumps have been subjected, but the rotary elements - functioned satisfactorily at this temperature. The lubricating oil and the water temperatures were maintained af about the values experienced during hot shakedown development and acceptance testing. - - The vapor trap installed in the main fuel pump of the ARE system was used successfully during the enrichment procedure. Several hundred cubic feet of gas was vented through it. The heated line and trap were still open ot the termination of the experiment. A report on the vapor trap design 23 ’\_ !»“ Ly 4«?. ]7 N o * ANP QUARTERLY PROGRESS REPORT and operation is being prepared. Modifications of this vapor trap design could be used on other fused salt systems in whuch reliable gas con- nectlon is imporfcn'f ' LOADlNG OF THE ARE e _ " G. J. Nessle | . C.M. Blood J. E. Eorgan UL P Boody N. V. Smith C.R. Croft J. Truitt Mcn‘erlais Chemlstry DIVISIon. ' R Wiley | A:rérafi Reactor Englneermg Division Fmal prepcranons ‘and installations for fher 'Ioadmg of the ARE were completed during the month of October.. Transfer of the barren carrier (NaF ZrF 50-50 mole %) into the ARE fill tank . No. 2 was completed in approximately 43 hr w:fhouf mmdeni’ Approximately 2750 Ib of carrier | looded into the fill tank, and sufficient : maferta[ was on hand to fill the fuel system to its operating liquid level. The enriched fuel storage tank had previously been loaded with 60 Ib of carrier material for practice injections into the reactor. Tests indi- cated that the semiautomatic enrichment system was inadequately heated, and therefore equipment for addition of portions of the fuel concentrate (Na,UF,) directly to the circulating fuel pump was designed and assembled. In order to provide adequate means for ap- proaching criticality safely, some of the original 30-1b batches of concentrate were reduced to 10-Ib batches, and, similarly, to provide accurate measurements for calibration of the control rods after criticality, 12 small batches of approximately 0.5-1b size were prepared from a 30-1b batch. The injection system consisted of a main con- centrate’ can (containing 30, 10, or 0.5 lb as required) connected to an intermediate transfer can by a resistance-heated 1/4-in. Inconel tube. The intermediate transfer can was designed so that it could be filled to a predetermined level (approximately 5.5 b) when the 10- or 30-Ib batches were in position to be transferred. The 0.5-1b batches were transferred through a smaller mfermed:ate can, which contained the entire small batch each time of transfer. The intermediate can was, in f't._ayrn','connec’red to the fuel pump. The main transfer line to the pump was heated by “calrod units, WHi_le the nozzle within the pump s was heated by resistance heating. The close tolerances of the injection nozzle resulted in an inherent tendency to short out and cause the nozzle tip to be cold. " This, coupled with the drainage of high-melting-point material from the fong transfer line, always presented the possi- bility of a frozen-fluoride plug. The first 30 b of concentrate was mlected into the fuel pump without apparent difficulty. After this injection, however, difficulties in the form ~of line plugs and venting troubles frequently occurred. After 48 hr, most of these problems were under control, and loading of the enriched concentrate proceeded smoothly. The reactor was brought to criticality, and small injections of 0.5-1b batches were made for the control rod “calibrations. At the completion of the final small addition, the transfer line broke and simultane- ously plugged ot the injection nozzle. Efforts to repair the damage were fruitless, and the final injection of 30 Ib of the concentrate to allow ARE operation at power was made through the sample line into the pump. ANALYSIS OF LEAK IN SODIUM SYSTEM E. E. Hoffman W. H, Cock C. F. Leitten Metallurgy Division The sodium system of the ARE was filled initially on September 26, 1954, but the sodium had to be dumped on September 27 because of a leak at a tube bend in the sodium purification system. The section of austenitic stainless steel pipe in which the leak occurred was located just ahead of the filter traps. The system had been filled with sodium for 37 hr and had reached a maximum temperature of approximately 1150°F before the leak was detected. A portion of the section removed and the location of the leak are shown in Fig. 1.6. The leak occurred where a metal thermocouple protection tube had been at- tached to the stainless steel pipe by Heliare welding. The attack around the weld was due to the formation of sodium oxide as the sodium leaked through the weld. The inside of the pipe at the weld may be seen in Fig. 1.7. An ex- cessive amount of penetration was obtained during the welding operation. The large weld nugget cracked during cooling after welding. A cross section of the weld crack is shown in Fig. 1.8. Examination of the pipe in this section revealed -AY i ge 2 de i docpd Rie . R ohl: Kol o iy i L Cojo = i 0 2 i W o ¥ o -p % E T T Fig.‘ o . Fig. 1.7. fweld 12X, Redueed 27%. Where Thermocouple Well Had Been Attached. B UNCLASSIFIED Y-13310 Inside View of Crack in Thermocoup!e‘ PERIOD ENDING DECEMBER 10, 1954 UNCLASSIFIED Y-13298 (INSERT Y-13309) 1.6. Sect'ion' of Tube Bend Showing Location of Sodium Leak in Type 347 Stainless Steel Pipe Inset shows location of leak at higher magnification. “a variation in wall thickness from 76 to 135 mils, and the spectrographic examination revealed that two pieces of pipe in this section were type 347 stainless steel, while the other two were type 316 stainless _steel. Chemical analyses of three sodium samples taken from this section indicated an 'average sulfur conte of 31 ppm. The actual cause of the exce;suve penetrahon can probably be aflnbufed to a lack of knowledge on the we!der s part of ‘the vqnahons in pipe wall 'rhlck- _ nesses m thts sectlon.ww PRELIMINARY ESTIMATES OF CORROSION IN ARE W, R Gfi.mes DA R Cfiheo Materials Chemlstry DlVlSlon It was possible to wnhdrow' samples of the ARE fuel for analysis during operation of the reactor 19 5 ., b st Ty ; & ‘*} 4 ! 1 Lo ANP QUARTERLY PROGRESS REPORT B UNCLASSIFIED) 3 y-3418 | Fig. 1.8. Cross Section Through Thermocouple Weld Crack. below criticality or at very low power levels. Samples were withdrawn by pressurizing the pump bowl sufficiently to cause flow through a heated Inconel tube into a sampling device, which was constructed to allow line flushing before col- lection of the sample. Analyses of the samples were carried out by the ANP Analytical Chemistry Group end are shown in Table 1.1. As may be seen, the chromium content of the fuel was essentially constant at 100 ppm after 155 hr of circulation of the fuel carrier through the reactor. If corrosion of the Inconel is assumed to have been uniform over the surface of the system — and since it was operating isothermally during this period, such an assumption is reasonable — the chremium con- centration of 100 ppm would represent about 0.5 20 TABLE 1.1. ARE FUEL SAMPLE ANALYSES Hours After Uranium Chromium Filling with Barren Carrier (wt %) (epm) 19 81 60 90 110 102 155 100 157 1.84 150 178 3.45 190 182 5.43 200 205 9.54 205 242 12.21 300 246 12,27 320 268 12.54 378 286 12.59 420 307 13.59 445 mil of attack on the Inconel. As was anticipated, the chromium content of the melt was found to increase a few hours after each addition of fuel concentrate (Na,UF ;). By the time the fing! sample was taken, after 307 hr of molten fluoride circulation, the chromium con- tent was found to be 445 ppm. Again, if a uniform rate of attack of the Inconel is assumed and if the removal of circulated material from the system during this time (to maintain desired pump liquid level) is taken into account, this chromium content would represent about 5 mils of attack on the Inconel system. I|f the data in Table 1.1 are plotted as chromium confent vs time, it is seen that the chromium content had begun to level off by the time the final sample was taken. A pro- jection of the slope of the chromium _cur-ve- indi- cates a maximum chromium concentration of about 600 ppm after several hundred hours of operation. There is no doubt that mass transfer of chromium metal began soon after power operation started, since there were then large temperature differ- entials across portions of the system. Sampling wa_s'not possible at this time and could not, of course, have given information on mass transfer. Heat transfer characteristics of the ARE did not appear to have changed during its 150 Mwhr of operation, and so whatever mass transfer occurred was unimportant. Details of mass transfer will be studied when sectioning of the equipment becomes possible. - R b il i - r¥ C» 1954 PERIOD ENDING DECEMBER 10, 1954 2. REFLECTOR-MODERATED REACTOR wr A. P. Fraas Aircraft Reactor Engineering Division -~ Analytical studies, layout work, and detail de- sign have proceeded on both the reactor assembly - and the installation for the Aircraft Reactor Test (ART), formerly the Circulating-Fuel Reactor Experiment (CFRE). Work has also continued on the component development tests outlined in the previous report. | Reports have been completed on the high-power-density beryllium thermal stress test and the first fuel-to-NaK heat exchanger test.2:3 Specifications have been completed for radiators to be used as heat dumps in heat ex- changer tests, It is expected that essentially the same specifications can be used for procurement of the radiators for the ART heat dumps and that endurance test experience gained in the course of these small heat exchanger component develop- ‘ment tests will establish the reliability of the product of at least one vendor. | EXPANSION-TANK AND XE NON- REMOVAL SYSTEM # , G. Samuels | Aircraft Reactor Engineering Division W. Lowen, Consultant ¥ Recent work on the expansion-tank, fuel-pump, and xenon-removal system has been directed primarily toward the development of a hydraulic circuit that performs all the various functions re- quired of the system. The basic components and their principal functions were described in an earlier report4 and may be summarized as fo“ows (1) an orifice for bypassing a mefered qmoun'r of”v fuel from the mcnn system for processmg, @ a - 'swul chamber for producmg a large gas- ]quId ffec'r the prompt releuse of xenon £ lA P. Fraas, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 25. m'Beryllzum Tesi No. 1, ORNL CF 54 10-106 (Oct. 25, “Stress Study of the Intermediate Heat Exchanger Test, ORNL CF-54-11-69 {to be issued). ‘4R, W. Bussard and A. P. Fraas, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, p 39. 9 interface area by v:olem‘ cgl'rahon “and gas en'rrcnn- for the main fuel system with a liquid surface that is stable for all attitudes (in flight); (3) a centri- fuge cup integrally mounted on the back of each fuel pump impeller for degassing the processed fuel, pumping the processed fuel back into the main system, and providing a seal for the fuel pump. Several Lucite models that differ in component design and circuit arrangement have been built and tested, and they have led to the design illustrated in Fig. 2.1. In this model the swirl chamber is mounted between the two fuel pumps and raised above the centrifuge cups to assure their priming under starting conditions. The swirl in the chamber is produced in two ways: by nozzleslo- cated in the swirl chamber floor and by swirl pumps which are needed to maintain a high swirl velocity when the fuel level rises above the 25% full condition owing to thermal expansion in the main system. The nozzles and swirl pumps are arranged to deliver high-velocity fuel jets tangenti- ally at the periphery of the swirl chamber to give good agitation and to assure a strong centrifugal field and a stable free surface. The nozzles serve both to meter the bleed flow through the expansion tank and to control the fuel pump suction pressure relative to the swirl chamber gas pressure. This important function is perhaps more evident in the schematic circuit analog pre- sented in Fig. 2.2. By tracing the bleed flow, it can be seen that the pressure drop across the _ nozzles approximates the heat exchanger resis- . tance. Consequently, the fuel pump suction pres- sure is maintained near the helium pressure exisfing in the swirl chamber. ..The configuration of nozzles and parts is so proportioned as to oversupply the centrifuge fuel Aflow demand in_order to minimize the possibility of the cen’rrlfuge becommg emp’ry, “especially during speed transients, o centrifuge is dlverfed by the flow-split stator to R, W, BUSSGrd und R E MccPherson Tbermal Stresses The excess fuel to the the swirl pump, which, as previously described, ejects this fuel back into the sw:rl chamber. Thls_ "arrangemenf of mcorporcn‘mg a sec0ndcry circuit within the bleed circuit contributes toward stabi- lizing the three free surfaces inherent in this layout. m ! y paven i ,' ) P i‘ ,’A STy T £ ; Y » ¥ 3 T o6 & ST 21 4 @ SWIRL CHAMBER SERVES AS EXPANSION.TANK PRODUCES GAS ENTRAINMENT. PROVIDES NEGATIVE 6 STABILITY SINGLE SHAFT —OVERHUNG ‘/{ 1 " FUEL UNDER HELIUM PRESSURE UNIT ASSEMBL:/ REMOVABLE L 7 //////// - FUEL PUMP - R -y LARGE RADIAL AND D'SCHAFSEOF;,RS.ESSU E mend AXIAL CLEARANCES A A Fig. 2.1. Expansion-Tank and Xenon-Removal System, X “SECRET ORNL-LR-DWG 4497 SLINGER [MPELLER SEALS AGAINST PRESSURE ENABLES INVERTED CPERATION ) AI———— SWIRL PUMP PROVIDES ADDITIONAL ENERGY FOR SWIRL AGITATION SERVES AS GAS TRANSFER PUMP ASSURES VARIABLE SPEED STABILITY | T——— FLOW-SPLIT STATOR PRIMES CENTRIFUGE ASSISTS IN KEEPING CENTRIFUGE SUPPLIED WITH FUEL CENTRIFUGE SEAL IMPELLER MAINTAINS PRESSURE IN SYSTEM PREVENTS CENTRIFUGE BY-PASS q_,jjk CENTRIFUGE REMOVES BUBBLES ACTS AS A TRANSFER PUMP OF THE BLEED FLOW 1dOdIY SSIYOOYd ATIFLIVNO dNV PERIOD ENDING DECEMBER 10, 1954 HELIUM PRESSURE sEewET ORNL—LR~DWG 4498 £ | ] . O e | : Ap~60psi VSLJNGER & IMPELLER -BUCKING 777 T I IMPELLER LEAKAGE SWIRL PUMP = | - Ap~30psi > 2 | ' 2|9 XENON- REMOVAL SYSTEM < o e - BLEED CIRCUIT -/ > < fi— \SWIRL STATOR * s )] x% s CHAMBER 1 SEAL FLOW T & BUCKING PUMP o313 & - NOZZLES 5* o ™ CENTRIFUGE 229 SEAL .J IMPELI_ER . W@ CENTRIFUGE FUEL PUMP * REACTOR CORE < MAIN FUEL CIRCUIT™ 23 A R i i Cold, ;e e i ANP QUARTERLY PROGRESS REPORT The addlhon of 'rhe swirl pump made it neces- sary, however, to replace the sllnger ring, previ- ously employed, with a slinger impeller in order to buck any leakage from the swirl pump. Although the complete assembly has not yet been tested in ~an inverted position, the slinger impeller seems to be very effective. A similar arrangement is now ‘being tried as a means of controlling the leakage from the main pump discharge through the clearance _ between 'rhe impeller and the casing past the cem‘rlfuge. More conventional sealing arrange- ments were found to be ineffective because of the abhb'rmoily 'ldrge‘ radial clearances (0.060 in.) specified in the design objectives. Initial tests currently in progress with the centrifuge seal impeller have been quite promlsmg Although no detailed studies of the off-gas system and fuel pill addition sysfem have been made, the changes in fhe xenon—removal system will make it neces- sary that the off-gas system connection and the fuel pill addition be made through the swirl chamber rather than through the pump housing as illustrated previously.> CONTROL ROD DESIGN CONSIDERATIONS The design of the ART has been predicated on the belief that no fast-moving control rods will be required because of the inherent stability of the reactor that will arise from the strong negative temperature coefficient of the circulating fuel. It is further expected that only about 1% Ak/k will be required to compensate for xenon poisoning, since it is anticipated that with the bypass ex- pansion tank and gas-scrubbing system planned, the xenon will be removed almost as rapidly as it is formed. |t is expected that compensation for burnup and the accumulation of fission-product poisons can be effected by adding pills of solid fuel having a high percentage of U235, Thus the principal function of the control rod will be to control the mean temperature level of the reactor fuel. Since the temperature coefficient will probably be approximately 5 x 10=3 Ak/k per °C and since a temperature level variation of the order of 150°C will probably be required, provision of a Ak/k of 1% in the control rod for temperature level control should be more "rh‘an adequate. ln reviewing the various requirements for control, thus follows that a control rod having a total k e ff of 2% should prove to'be adequate. This reactivity Sibid., Fig. 3.4, p 40. 24 canbe readily incorporated in a single slow-moving control rod. [t would seem entirely adequate to move this control rod at a rate of 1% Ak/k per minute, a slow and conservative value. The size and stroke of the rod are important considerations, Two items of experimental data are availoble fo indicate the size of the rod re- quired. In the course of the second reflector- moderated reactor critical experiment (that making use of powdered fuel placed inside 1%-in.-square aluminum tubes), it was found that with an 18-in.- dia core and a 9%-in.-dia island a control rod effectiveness of about 0.6% Ak/k was obtained from a 346-|n.-0D stainless steel tube filled with powdered boron to a density of about 50% and inserted in one-half of the critical experiment assembly. A second indication of the effective- ness of control rod material is available in that a piece of gadolinium oxide approximately the diameter and thickness of a 25-cent piece was in- serted at the center of the island in the three- region octahedron critical assembly (without core shells). The value of this gadolinium oxide wafer was found to be approximately 0.2% Ak/k. Some additional data are also available from multigroup calculations. All these data seem to indicate that a rod approximately 3’/8 to 1/2 in. in diameter having a stroke of approximately 20 in., about 71/2 in, of which would be below the equator, should prove adequate to give a Ak/k of 2%. The heat generation to be expected in the con- trol rod will establish the amount of cooling that will be required and, to a large degree, much of the detail design of the rod. The heat generation, in turn, is dependent, in part, on the number of neutrons absorbed in the rod at full power. Molti- group calculations indicate that 1.3% of the neutrons produced will be absorbed in a control rod having a Ak/k of 1%, if it is assumed that the average effectiveness of the control rod is half that of the portion at the center of the island. Thus a control rod having an effectiveness of 2% Ak/k would, when in the ““full in’’ position, absorb 2.6% of the neutrons. [f a value of 3 Mev per neutron absorbed is assumed, the energy associated with neutron absorption in the control rod would amount t0 0.039% of the total power generated in the reactor. This would total 24 kw for a 60-Mw power level. The control rod cooling requirement would be a minimum if gadolinium rather than boron were used as the absorbing material in the rod. This 3 —(’ - ;.’ s Cal pust ) 5 - =, T ey Hcnn ool e 1, < el | v BmBe “§ W a e _amount to about 20 w/emS, ' quire'd. g frangi would mean that most of the energy associated with the hard capture gammas emitted from the gadolinium would appear as heat in the surrounding moderafi-ng material rather than in the gadolinium. Thus a gadolinium rod would have the advantage 'rhcn‘ relatively little of the heat associated with neutron captures in the rod would appear in the red, and therefore little provision for cooling would be required. On the other hand, if boron were used, the very short range ofpho emitted from neutron captures in boron would cause virtually all the energy associated with the neutron capture to appear in the boron. ‘The principal heating in a gadolinium rod would be that induced by gammas from the fuel region. For a 60-Mw reactor it appears that the power ~density in the gadolmlum from this source would Estimates indicate that about 20 cm® of gadolinium would be re- The total power generation in the rod from gammas from the fuel region would amount to opprox1mate|y 400 w, or about one-sixtieth the power ‘generation in a boron rod. It should be noted that gadolinium oxide is available at a cost of about $4000/1b and that it can be fabricated readily by using conventional ceramic techniques. Cooling of the rod presents quite a number of problems. If the rod were immersed in flowing sodium, cooling would present no problem. How- ever, the control rod octuotmg mechanlsm would have to operate, to some degree at least, in sodium, and self-welding might present some difficult problems. If the rod were placed in a thimble in the center of the core and an atmosphere of helium were mamtamed in the 1h:mb|e,7|f mlghfr be posmble to effect coohng through rodlahop‘_or:;i S conduchon from the rod materlal to the sodiv cooled wo!ls of 'rhe thlmble._” crrongemehf;mlght prove to be scflSfccfory WIth o 'umfomcle or ‘f' ‘Haf u might be circulated over the rod to cool it, in which case the outboard end of the control rod Thus_‘ sor'r of an 1 " Valves and thetr consequent enfralnment in la'rer _”\fllhng and drommg operohons “and’ por'rly to mini- n"'m:zeu the _amount ‘of raduoochvfiy ‘that might be »Jilnclude’ prov151on for accurafe m quantity of fluid in the “drain tank at all times PERIOD ENDING DECEMBER 10, 1954 mechanism could include a heat exchanger so that the heat would be removed from the circulating helium and transmitted to the shield water, Un- fortunately, such a system would be dependent on the satisfactory operation of quite a number of moving parts, always a likely source of trouble. FILL AND DRAIN SYSTEM A good, reliable, relatively simple fill-and-drain system incorporating a quick disconnect coupling suitable for remote operation with high-temperature liquid systems is clearly needed. In the design of the ARE, fixed tanks withremotely operated valves were used in order to avoid a remotely aperated coupling. This approach led to a quite complicated collection of tanks, plumbing, and valves, and to somewhat clumsy arrangements for system drainage. The requirements of such a system are fairly straightforward. In the first place, a good coupling that can be operated remotely and that will be dependably pressure-tight is required. This coupling must be relatively insensitive to align- ment both as to concentricity and to parallelism of the axes of the mating flanges or surfaces. A reliable pressure-tight valve on either side of the coupling must be provided so that, after the coupling has been made, the space between the valves can be evacuated and then purged with helium. The valves on either side of the coupling can then be opened and the filling or draining operation carried out. Upon completion of the operation, the upper valve can be closed and the space between the valves cleared of liquid by sevérol short blasts of helium. The lower valve can then be closed and the coupling broken. Scavengmg the spoce befween the volves is very mportant, partly to mlmmlze the possnbihty of ox:dotlon,of droplefs of mcn‘erml betw<=en the mitted fo ’rh'e All this implies that "'rhe volume ‘and surface dreu be’rween the two volves should‘bo_mlmmlzed ‘and fhat the geomefry k fa "I'tote drcunage and to ,surve}nen'r of fhe‘ during either the filling or the training operation. This is particularly important in connection with STy e 25 T £ "ANP QUARTERLY PROGRESS REPORT reactor fuel systems because it is essential that the exact amount of fuel in the reactor be known - at all times. One good way of measuring the quantity bf liquid in the tank would be to support the tank on two Hagan Thrustorq units. This should give an instantaneous measure of the weight of the tank to an accuracy of 1% of full-scale reading. Design work is under way on a system to fill the above-specified requirements. Complete tests of mockups of proposed designs will be made before the final design is chosen. DESIGN PHYSICS W. K. Ergen Aircraft Reactor Engineering Division ' The comparison between multigroup calculations and critical ‘experiments on three-region reflector- moderated reactors has shown a certain amount of discrepancy, which must be overcome by adjusting the constants used in the caleulations. Investi- gations were made with the ultimate aim of ob- taining simpler expressions for the critical mass and power distribution for reflector-moderated reactors. This was done in the hope that these simpler expressions could be fitted to the experi- ments by adjusting of constants and that this fitting might be as satisfactory as the one employed for the multigroup calculations. The three-group ORACLE calculations of the approach to criti- cality of the ARE (see Sec. 1) gave a satisfactory explanation of the strange results observed during the experiment. An example of the calculations performed for the ART is the following calculation of the activation of the Inconel core shells. Activation of the Inconel Core Shells "H. W. Bertini Aircraft Reactor Engineering Division An attempt was made to estimate the activity of the Inconel core shells on the inside of a 60-Mw reflector-moderated reactor after the reactor had been operating for 1000 hr. The activity will come from two main sources: the activity of the cobalt in the Inconel and the activity of the fission 26 fragments that strike the Inconel and remain there. The method described below for obtaining the estimate gives an order of magnitude of the total activity. Reasonable approximations are used to simplify the calculations, and, in this way, many of the details that a more complete analysis would require are circumvented, and yet the required accuracy of the calculations is maintained. To obtain the activity due to the cobalt, the total flux at the surface of the Inconel is obtained, This flux is multiplied by the product of the microscopic cross section of cobalt times the total number of cobalt atoms in the core shells. This gives the total number of cobalt atoms being activated per second. Multiplication of this number by 1000 hr gives the total number of cobalt atoms activated during the reactor lifetime. Since the half life of cobalt is so long (5 years) com- pared with the reactor lifetime, the decay of the excited cobalt atoms can be assumed to be con- stant, and therefore the disintegration rate can be taken as AN, where A is the decay constant of cobalt, and N is the total number of cobalt atoms activated during the lifetime of the reactor. The following values were used for the calcula- tions: Reactor power = 6 X 107 w Total mass of U235 in the core = 2.5 x 104 g Density of Inconel =8 g/cm3 Weight per cent of cobalt in the Inconel = 0.1% Half life of cobalt =5 years Average flux at the surface of the reactor, (’és =2X (]Sreacfor 1 curie =3 x 1010 decays/sec 0’5‘,J = 550 barns GSO = 35 barns The total mass of Inconel in the core shells of the reactor is approximated by the mass of two spherical Inconel shells, one with a 15-cm radius and the other with a 25-cm radius, each shell being 0.3 em thick. Then, + In Wy PERIOD ENDING DECEMBER 10, 1954 (total fissions/sec) h“& ! ft N X crfu % (total U233 gtoms) 9 6 x 107w x 3 x 1019 (fissions/w-sec) % . 6 x 10 x 2.5 x 104 <550 x 10—=24 ~ - * (fissions/neutron.cm?) | 235 -1 x 10”(neut’rons/cm2»sec) The foi;al- lr'wmber‘ of cj:o.boh‘ atoms inl'rhe‘: liner is | th'?at = [8 (g of Inconel/em®) x 4m (252 + 152) (em?) x 0.3 cm | X 10'3 (g of Co/g of Inconel) '« 6 x 1023 (afom; of Co/g-at. of Co)l + 59 (g of Cp/g-cflr‘° of Co) - - 2.6 x 1023 - The activations per second are Co = Co P ¢A x Ntotal A I 1074 x 2.6 x 1023 x 35 x 10~24 i 1] 9.1 x 104 (activations/sec) . The total number of cobalt atoms activated is P x 1000 (hr) = 0 o, i = 9.1 x 109 (activations/sec) x 1000 (hr) x 3.6 x 103 (sec/hr) - =33 k102, | The activity of the shells due to _cobdh‘ activation is given by =A% Nf") .nr(rai;s'iihlzf.e;;j‘l“a'rions/sec) R ;-'-fvr;z:‘,-_-3.,~3'-;--1021-, L 2 (h/day) x 3.6 x 102 (sec/hr) 3 x 101 5 (years) x 365 (days/yea 27 ANP QUARTERLY PROGRESS REPORT In calculating the activity of the fission frag- ments, use is made of Fig. 2 in ORNL-53,% which is a curve of the dose from a uranium slug exposed ~ for three years plotted against time after exposure. The dose is given in mr/hr at 1 meter per watt of exposure, ‘ Calculations now show that 1 curie of 2-Mev gammas gives a dose of approximately 1 r/hr at 1 meter. |f it is assumed that the average energy of the decay gammas from the fission fragments is 2 Mev (an overestimate), the curve can be used to obtain an estimate of the curies of activity for anv time after exposure. However, it is first necessary to calculate the watts of exposure in the regions of the reactor close enough to the Inconel core shells so that the fragments due to fissions will strike the Inconel. This is done by finding the range, R, of the fission fragments in _ the fuel and again approximating the fuel annulus by a spherical shell of 25 ecm OD and 15 em ID “to find the total power generated in the regions adjacent to the Inconel. This power is divided by 2 to account for the fact that approximately one-haif the fission fragments generated in this region will impinge on the shells. Since the curve was plotted for a slug exposed for three years, any valves obtained by using this curve will be overestimates of the fission fragment activity of a 60-Mw reflector-moderated reactor. The range of fission fragments is estimated to be 0.001 ¢m, and the 1 watts of exposure = 6 x 107 w x — 2 47 (252 + 152) x 10-3 X The last factor of 2 in the above equation is due to the flux at the surface of the Inconel being twice that of the average flux. Some values of the total activation of the Inconel core shells obtained by using the above methods are given in Table 2.1. One conclusion to be drawn from Table 2.1 is that little would be gained by trying to obtain special Inconel with especially low cobalt content; the activity could not be reduced below that caused by the fission fragments, TABLE 2.1. ACTIVATION OF INCONEL CORE SHELLS Activity (curies) Days After Exposure Cobalt Fission Total ' Fragments 10 480 390 870 20 480 260 740 100 480 80 560 Mathematical Models for Reflector-Moderated Reactors L.. T. Anderson, Consultant The thermal absorption rate has been calculated for an idealized reactor consisting of a nonmoder- ating spherical-shell fuel region surrounding a moderator and surrounded by an infinite moderating reflector with a fission source of 1 neutron/sec in the fuel region, The thermal capture cross section of the fuel region was taken to be infinite, 4 x 2 The nonthermal capture cross section of the fuel 3 7 (253 - 159) region was assumed to be zero, and the thermal and nonthermal absorption cross sections of the 4 moderafor were likewise assumed to be infinitely = 1.3 x 10%w . . small, The neutrons slowed down according to age theory, and the thermal neutrons diffused accordin $G. Ascoli and O. Sisman, Absorption o{l Radiation df);’ . h d from an "X Slug by Lead, Flg. 2, p 7, ORNL-53 (Mdy to diffusion t eory. 1948). The absorption rate is, for the source located at the inner radius of the fuel region, 1 - 2 e i (1) Ay =— dx , 7 2 2 0 x R, 1 +{— = 1] xcotx| + cot?x Ry 28 Doy T 034 W #ea g o where R, is the island radius, R, s the outer radius of the fuel region, and 7is the thermal age. Evidently the absorption rate depends only on two parameters: R,/R, and T/Rf. For the source located at R,, the result is PERIOD ENDING DECEMBER 10, 1954 and this curve is also shown in Fig. 2.3, The ““thermal absorption rate’’ plotted is a direct measure of the neutron leakage — the higher this rate the lower the leakage. R, 2 L R - -(T/R%)xz 1+ —1— 1] x cot x e (2) A2 = — ——— dx . 7 R2 2 2 0 x R 1 | — ~ 1) xcotx| + cot?x R? For a uniform source throughout the fuel region, the result is 1 Rl (R2 + ZR-‘) (3) A = A, +3 (A; - A,y . Equations 1, 2, and 3 were evaluated by numerical integration for R,/R, = 1.67 and 'r/R2 = 0.33, 0.44, and 0.55. The plots are shown in Flg. 2.3. For the case R, =0, Eq. 1 reduces to 1/2 -7/ R% 71/ e erfc , sEeRET ORNL-LR-DWG 4499 ‘ t/R? 0.222 0333 0.444 0,555 0.666 09 g ey w 5 5 e ~ 08 : [ ® i . ) . \\4 | —SOURCE AT A, z A L1 UNIFORM SOURCE - Q - - g = o “\‘~\ e \ a "\ * % 0.7 9., 5\ — T o~ = SOURCE AT 7, | P *- 3 ' 2 ~ S [ NO ISLAND T Pl " os : L — , 0.20 024 05 om _ ox o /R, Fig. 2.3. Thermadl Absorpfi.onr Rate in Mathe- matical Model of ReflectorsModerated Reurciors. 2 2 R] +R1R2+R2 PROPOSED ART INSTALLATIONS A, P, Fraas F. R. McQuilkin Aircraft Reactor Engineering Division A study has been made of the various types of facilities suitable for the operation of the Aircraft Reactor Test. In addition to the obvious require- ments for shielding, heat dumps, and auxiliary equipment, it is essential from the hazards stand- point that provision be made to contain the products resulting from a nuclear accident and/or chemical reaction of all combustibles in the installation, minimize the likelihood of serious damage from an explosion caused by sabotage or bombing, and remove and dispose of volatile fission products evolved during the course of operation. With these criteria in mind, a series of four basic reactor installations was considered; each differed in some fundamental characteristics from the others, The four installations are as follows: (1) an open test unit mounted over a water-cooled pan at the National Reactor Testing Station (NRTS), (2) a circular Quonset type of hemispherical building, (3) a water-walled tank, and (4) a reactor submerged in a pool of water. rl‘Each of ’rhe msfalichons mcluded f;ve malor w"{:m'[oonen'rs “"a shielded reactor assembly, radi- ators, blowers, a control system, and a fill-and- drain system. These components would operate 29 T T T i G ANP QUARTERLY PROGRESS REPORT as an integroted system, irrespective of the type of test installation chosen. In all but the fourth installation, the assembled reactor would be sur- rounded by an aircraft type of shield of lead and borated water, and, in all installations, it would be coupled to heat dumps consisting of banks of aircraft-type NaK-to-air radiators through which cooling air would be circulated. Appropriate con- _frol systems, along with auxiliary shielding, equip- -ment, and services, would comprise the balance of - the test installation. A generalized flow sheet for th:s setup was shown in Fig. 2.1 of the previous " quarterly progress report.” All but the first of the above-listed installations ‘would be located in Oak Ridge. The first installa- tion is illustrated in Fig. 2.4. It was devised to permit operation of the reactor surrounded by an © aircraft-type shield with a heat dump on either side " fo simulate the turbojet engines. Of the installa- tions considered, it offers the most compact “arrangement of the equipment with the least amount of shielding and minimum provision for contain- ment, This scheme was developed with the thought that it could be built in Oak Ridge so that all the welding of high-temperature-liquid piping could be made, inspected, and pressure tested and some preliminary testing carried out, probably including a hot critical experiment, before the unit was shipped to NRTS, The dimensions of the unit are such that it would fit on a flat car and comply with standard railroad side and overhead clearance regulations. To do this, it would be necessary to dismount certain elements, such as the pump drive motors, the blowers, and the blower drive motors. This could be done easily, since only bolted con- nections would be involved, The reactor would be ~ set up with a heavy, water-cooled pan beneath it, - This pan would catch, hoid, and cool the fuel in ‘the event of an accident, A control room would be built as a unit and shipped to NRTS on a second “flat car. The contrél room and the reactor would “probobly be placed a quarter of a mile to a mile apart, and the two would be coupled by telemetering . _equipment, The pumps in the layout are shown as - -being driven by d-c electric motors, but air turbine ‘motors would serve equally well if a source of , " compressed air were available. If the tests were . _run at NRTS and Air Force, portable, gas-turbine- S fipé"dir 'com'pressors were used, a compressed air .- 7A, P. Fraas, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 21. 30 source might be more easily arranged than a d-c generator set. ‘ ‘ In examining the NRTS installation design, a number of points became evident, First, the prob- lems associated with operating a reactor at NRTS seem to be rather serious, particularly from the standpoint of the amount of time that would be lost in maintaining an operation 2000 miles from Oak Ridge. The distance would make it particu- larly difficult to cope with unforseen problems. Any relatively small difficulty that might arise would be likely to introduce a major delay if that difficulty were not forseen. It appears therefore that an NRTS installation would entail a loss of at least six months in getting the reactor into opera- tion, The second major point that developed in connection with the examination of the design was that the major hazards appeared to be much less serious than had originally been presumed. The very compact installation achieved through careful design directed toward simulation of a full-scale aircraft type of power plant led to a very low in- vestment of sodium and NaK, about one-twentieth of that required for the KAPL-SIR reactor designed for the same power level, Further, the use of circulating fuel with its high negative temperature coefficient gives a reactor in which a nuclear explosion seems almost out of the question, In view of the relatively small amounts of energy released under conditions of a total reactor tragedy, designs were prepared for the installation of the ART in a closed building. The layout shown in Fig. 2.5 envisions a sort of circular Quonset building about 200 ft in diameter. The test unit would be built and operated on the test floor behind supplementary shielding, and a heavy, water-cooled pan would be placed beneath the reactor, This pan would catch, hold, and cool the fuel in the event of an accident. Such an arrange- ment would give plenty of floor area in a relatively inexpensive building that could be sealed to con- tain any fission products that might be released, in line with the philosophy underlying the use of the Hortonsphere at KAPL., A similar type of building, but shaped as a hemisphere rather than as the circular Quonset building, might be used in order to reduce the amount of steel required in the framing of the building. The arrangement shown in Fig. 2.6 follows a quite different philosophy. It provides for con- taining the reactor assembly within a pressure A g o ) e ~ow w b g wha? 0 1€ ) . . . © \ - oo o~ " wia " - s “SEURer™ ORNL-LR-DWG 1391 FUEL PUMP MOTOR Na PUMP MOTOR THERMAL INSULATION NaK PUMP, HEADER TANK AND 60 H.P. DRIVE MQTOR THERMALLY INSULATED DOORS s = AIR INLET SCREEN AXIAL FLOW BLOWER NaK TO RADIATORS | 0 1 2 3 ‘ e e —] FEET NaK RADIATORS SUPPORT FRAME 26 x 7 ----- bl 7 ] - : | ] Lt i j AUXILIARY BLOWER AND OIL BURNER FOR PRE-HEATING THERMALLY N - }-in. THERMAL INSULATION INSULATED BOORS Fig. 2.4. Open Reactor Test Unit. ye6l ‘0l ¥39W3D30 ONIANI AOId3d R e e A - PR — VI T T A ST AT i - € UNCLASSIFIED " . ORNL-LR-DWG 4500 e 8Ot Oin ———] . —— i—-— 4t Cin. 1¥OdI¥ SSIIO0Ud ARILIVNO dNV - 112 ft Oin, LONGITUDINAL SECTION HORIZONTAL SECTION ELEVATION T Fig. 2.5. Circular Quonset Building. w) PERIOD ENDING DECEMBER 10, 1954 Fig. 2.6. Perspective of Proposed ART Installation in Addition to Present ARE Building Showing Double-Walled Tank. vessel submerged in water, while the heat dump temperature of the inner tank should not exceed equipment is located outside, but nearby, A the water temperature by more than 40°F, even if double-walled tank from 20 to 30 ft in diameter and the bottom of the inner tank were covered with 2T e N T L from 26 to 36 ft fbl[ is 'enw's.loned "The s space | molten material, In the event “of. _sorrie'rhing so between the wa[ls wou‘ld be offlthe order of 18 in, serious as a meltdown, it seems likely that a part wcmd would be f:lled w:th wcter. The mner wall ~ of the water shield surrounding the reactor pres- sure shell would be ruptured and that water would tend to collect in the bottom of the tank with the fuel - This water would remove heat from the fuel, ‘vaporize, and then condense on the tank walls _ so that the atmosphere within the tank would the heat gwen off by 'rhe decay gamma achvnty ' probably not get hotter than around 300°F, "':would\ Ibe carried off by vaporization ‘of the water ~ The bolt flanges shown in Fig. 2.6 near the tops be-fween (fhe tank walls, Since the rate of surface- of the tanks could be placed about 3 ft above floor to- yvatel_' hect_fransfer’under boiling condmon's. “Jevel. The reactor installation and preliminary would be exceedingly high (on the order of 400,000 shakedown testing could be carried out with the Btu/hrft2) and since the thermal conductivity of tops of the tanks removed. Since the reactor shield the fluoride fuel is relatively low, the water-side would be quite effective, the space inside the 33 ANP QUARTERLY PROGRESS REPORT tank wou.ld be shielded fairly well, and it would be possible for aman to enter the inner tank through ' a manhole at the top for inspection or repair work, even after the tank had been closed and the ve- “actor had been run at high power. The water capacity of the spcice between the tanks, together ~with the water in the reservoir above the inner tank (apprommotely 10 ft deep), would be of the order of 110,000 gal, Boiling of this water would suffice to carry off all the heat generated by the fission products for about two days without any fresh water being supphed to the outer tank. All the vcrlous wires, plpes, tubes, etc, connected to the reactor and its auxiliaries would pass through carefully laid-out junction panels such as the one shown in Fig, 2.7. Such a panel might be installed in the wall of the tank shown in Fig, 2.6 with a pressure-hght gaske’red flanged ;uncnon. The various thermocouples, power wiring, etc. could be installed on the reactor assembly in the shop and fitted with Cannon plugs so that they could be plugged into the panel in a short period of time after the reactor assembly had been lowered into position in the test facility. This should minimize the amount of assembly work required in the field. The layout investigated as the fourth proposed installation would involve placing the reactor in- side o sort of swimming pool with water-tight thermal insulation surrounding the pressure shell, lines, and pumps. The lines could then be brought out the top of the water tank to the instruments and the heat dumps, In the event of a severe accident that resulted in a meltdown, there would be sufficient heat capacity in the water to absorb the heat from the fission products for some days before a seriously large amount of water would have been boiled out of the pool. Such an arrange- ment could be enclosed, of course, in an air-tight building. This might not be necessary, but it seems likely that, in the event of an abrupt melt- down type of failure, the high heat capacity of the region inside the thermal insulation might put so ~much heat into the water in a very short period of “time that bubbles would boil violently to the ".surface and disperse entrained fission products, g 'Aftéf,v‘- perhaps, 10 or 15 min, it seems unlikely that such entrainment would prove to be a problem, The unshielded reactor assembly will weigh “approximately 10,000 Ib, the lead gamma shield approximately 30,000 Ib, and the water in the ~shield approximately 34,000 1b, The first two of 34 these. ttems could be handled: convemenfly with a 20-ton crane, while the borated water could be ~pumped in after the rubber fonks hud been mstaHed for the water shield.’ : An armngement such as the double-walled tank shown in Fig. 2.6 should prove to be cdequate fo take care of any accident not involving sqbofcge or bombing. The same should be true for either the hemispherical or circular Quonset 'rype of building. Only the NRTS installation would, be- cause of the remote location, present a not- too- serious hazards problem if effectively sabotaged. To evaluate the merits of each of the installa- tions considered, an attempt was made to envision as many accidents as possible that might prove to be serious during the course of operation of the ART. The worst natural accident that could be envisioned would result in a meltdown, The only source of an explosion that has been envisioned would be either sabotage or bombing. If the volatile fission products are removed during the course of operation, the major hazard to the sur- rounding area would be from the fission products that might be dispersed in the course of violent boiling or from an explosion. Although in the preliminary hazards analysis consideration was given to operational hazards, operational sabotage, fire, earthquake, flood, windstorm, and bombing, the controlling considera- tions appeared,to be those associated with a total reactor tragedy. The total reactor tragedy is con- sidered here as being an accident in which all the heat that could possibly be released from the chemical combination of various materials in the reactor and associated system would be released, together with the heat from the fission products accumulated after extended operation at full power and the energy released in an extreme nuclear runaway. 1he principal hazard associated with an ultimate reactor catastrophe is the dispersion of the fission products that would have accumulated from extended operation at high power. The key data on these parameters are presented in Table 2.2, In examining the data on the heat that could be released from the combustion of various materials in the reactor installation, it is immediately evi- dent that if kerosene or another hydrocarbon were used in place of water in the shield a very large amount of heat could be released, an amount almost one hundred times greater than that from any other ST " o A2 ¥ Ay PERIOD ENDING DECEMBER 10, 1954 UNCLASSIFIED ORNL-LR-DWG 4502 © & OIL IN -1 © o O THERMOCOUPLES O o © @ e & o © o o O o o + O o O o @ CONTROL LIGHTS FUEL NUCLEAR PUMP ROD AND ADDITION INSTRUMENTATION SPEEDS ACTUATOR RECEPT. MACHINE ! O i o @‘ $ @ \‘h o O 0o o M ~ FUEL NaK PUMP AND Na Na PUMP DRIVE=-1 DRIVE-2 DUMP VALVE SPARE DRIVE-{ DRIVE -2 0O O 0 0 O 0 o =="0 o O 0 0 SHIELD COOLING . IN IN SPARE SPARE SPARE SPARE SHIELD COOLING ouT ouT XENON VENTS He SUPPLY * © © \&J. EXPANSION TANK 3 PRESSURE LIQUID LEVEL SPARES © © OIL OUT-2 @ @ e & & O 35 TP T AT A et e - e 'ANP QUARTERLY PROGRESS REPORT TABLE 2.2, SUMMARY OF KEY DATA ON ART HAZARDS Sources of Energy Heat from combination of 1000 |b of Na and NaK with water Heat from combination of 1000 1b of Na and NaK with air "Heat from combination of 1200 Ib of zirconium-base fuel with NaK Heat from combustion of 34,000 |b of shield kerosene with air Heat from éxtreme nuclear accident Fission-product decay gamma heat emitted during first 2 hr after shutdown {assuming no fission-product removal) Sources of Radicactivity Total fission-product activity for saturation 2.1 x 10% Bt 2.4 x 10% Bt 0.13 x 10% Bty 635 x 10% By 0.3 x 108 Beu 8 x 10% Bty 6 x 108 curies Gaseous fission-product activity for saturation 107 curies Sodium activity in moderator circuit for saturation 10° curies Sodium activity in NaK circvit for saturation 35 curies “,Temperafures Associated with Accidents #I_qme temberfifure for stoichiometric NcK*-H20 reaction 3070°F Temperature of reaction products for Na*—~zirconium-base fuel reaction 3300°F _Tefnpératur_e of a‘t'mo’sphere in ]2,000-“3 tank (assuming uniform dispersal of 310°F .--.com_i.fopstiidn prbduéffi from NG-H2O reaction) ’Tempérafurje rise of NQK-H20 reaction products and shield water if uniformly mixed 62°F Major Radiation Sources and Doses for Typical Conditions** Equivalent source for all fission products Equivalent source for inert gases only Rate of generation of activity in the form of inert gases (assuming 20-min holdup in atmosphere of expansion tank) Same as above but for 420-hr holdup time Dose at shield surface during operation Dose at shield surface 15 min after shutdown Dose at shield surface 10 days after shutdown Dose outside 24-in.-thick concrete wall 15 ft from center of reactor during operation Same 15 min after shutdown 4 x 108 curies 108 curies 1000 curies/sec 1 curie/sec 100 r/hr 4 r/hr 1r/hr 0.025 r/hr 0.001 r/hr *|nitial temperature assumed to be 1500°F, **These key data for typical conditions were developed for the major radiation sources and doses by assuming 1000 hr of continuous operation of the ART at 60-Mw power level with only the gaseous fission products coming off and by assuming the reactor to be encased in an aircraft-type shield which gives a dose of 1 r/hr at 50 ft. source of energy. The heat that could be released from the combustion of the sodium and the NaK in the moderator and heat dump circuit is relatively ‘small; thus it seems that water should be used as Vt'he' shielding material rather than kerosene, be- - cause, even if the water were to combine in stoi- chiometric proportions with all the sodium and the NaK in the system, the resulting energy would still not present a difficult problem, The heat that would be released from a nuclear accident depends in large measure on the character of the accident. However, the value presented in Table 2.2 is that for the worst accident that can be 36 envisioned, that is, one in which uranium abruptly begins to precipitate out of the fuel in the core so that the fuel is carried into the core at the normal rate but no uranium leaves with the exit stream of fluoride. While hardly a credible accident, this does seem to represent the maximum rate of in- crease in reactivity and, hence, the extreme nuclear accident conceivable for this reactor. The second portion of Table 2.2 presents the equivalent radiation sources for the various fluid circuits. The values given for the sodium and NaK activities have been obtained from multigroup calculations. Py ey FEe YT o 4 £ i o - -5 4} A " ma At It should be noted that the presence of the shield water would provide a strong damping effect on any temperature rise associated with an accident, The third part of Table 2.2 shows the temperature rise associated with the release of energy from the various sources for various conditions. As shown in the table, if all the heat from the combustion reactions were to be confined in the combustion products, high temperatures would result from stoichiometric reactions, Even in such instances, much of the heat would be dissipated to the atmos- phere within the containing vessel so that the temperature would be substantially lowered, If the reaction were quenched by the shield water, still lower temperatures would result. It is clearly essential that provision be made to contain the products of a total reactor tragedy except, possibly, if the reactor is installed at NRTS. A comparison of key data for several types of container is given in Table 2,3, For this com- parison the worst set of conditions applicable to each case was presumed. As may be seen, the double-walled tank compares favorably with the hemispherical and ellipsoidal buildings. While difficult to evaluate numerically, the double-walled tank also appears superior in that it would be less subject to sabotage. Even if both walls were ruptured and the reactor melted down, the residue would tend to sink to the bottom of the tank pit, PERIOD ENDING DECEMBER 10, 1954 where it would be flooded by the water that had filled the region between the walls. This water would both absorb the heat of any reaction and act as a shield to reduce the radiation level at the top of the pit. After careful review of these and a host of lesser considerations, the 24-ft-dia double-walled tank was chosen as the most promising test facility. Although the double-walled tank type of test installation is adaptable to several types of reactor test facility, the most promising possibility is to install it in an addition to the ARE Building. Such an arrangement would permit the use of services and facilities that already exist in the ARE Build- ing, for which no plans have been formulated now that it has served its original purpose. For ex- ample, items such as the control room, offices, change rooms, toilets, storage area, water supply, power supply, portions of experimental test pits, access roads, security fencing, and security lighting are available. Of the several schemes considered for modifying the ARE Building to accommodate the ART, the ‘most attractive plan is to construct an addition on the south end to effect a 64-ft extension of the present 105-ft-long building by extending the ex- isting roof and side walls. Figure 2.8 shows a preliminary design. The floor level of the addition would be at the ARE basement floor grade, with the double-walled tank to house the ART sunk in TABLE 2.3. COMPARISON OF KEY DATA FOR SEVERAL TYPES OF REACTOR INSTALLATION CONTAINERS 24-ft-dia 17-ft-dia _ _ Doub[e-Walled - Double—quled 209..”-?1'0 11-5-f1-d:la Tank with 11t Tank with 9-f = 1ipseidal - Hemispherical Strmght Secflon Srrmghf Sectton Building Building Heat released, Btu | o 1'0"5' 2.4 % 106 ' 107 107 conm.ner vo[ume, &3 Sy 200 4600 1.2x108 0.4 x 108 ”\Confumer surface clr_ea ft2 N o 2640 . ]400 4.3 x 104 2.1 x 10% . Peak Qus femperaturé,* OF A 44'4' o 970 T 130 ' 220 Peak gas pressure, ps;lg 1 ) | 37 T 1.4 3.9 7_ _Requnred shell 'rhtckness (for allow- T e e k & T ' ~able s"?:’S.f,]S,O 00 p51), in. o 0.217 _ 0.251 0.15 - 0.09 : We;ght ofsfeelmsheii, o R e L 5 ]34** SRR *Eission-product afterheat was not includedin caleulating peak gas temperature and pressure where adequate, provi- sion was made for cooling the container wall. **lncludes steel frame. 37 I i s i e 8¢ s, UNCLASSIFIED ORNL-LR-DWG 4471 e 14 Y5 in. L HATCH POURED GONCRETE SOLID CONGRETE BLOCKS [777] CORED GONGRETE BLOGKS Fig. 2.8. 20 4 8 12 16 20 FEET oo A 4—| . i 1 i 106 ft 1in. 64 ft O in, ADDITION——‘ —r——~ /\l /\‘ /\ I/\ - - W WY B e O RN AN 835 KT I LOCKER ROOM & oea w INSTRUMENT COUNTING OFFIGE CONFERENGE > 3 ROOM ROOM ROOM - 7 :’oj Zw — E 20 INSTRUMENT A N & E = SHOP =0 I A CORRIDOR 3 % 2 DOWN& g —] 2 n/\}—"b——r\/\y g - £ i 0 up £ fi [+2] GENERAL CONTROL, ROOM w SHOP 3 & FIELD MAINTENANGE SHOP = < & £ +1 - = ] T _‘L * + i K| ] [i2] 3 M H = R ZZeaTTTT. Ty T A e T T A L -EXISTING 12-X 12-ft 5- x 8-ff CONTROL INSULATION ROLL-UP DOOR, 3-x e T T T TRENCH 7-it SWING DOGR AND COLUMN TO BE RE- MOVED. J k ! <7 2 |><] N = d b ol b b ) e e o) b b e g%uzl - T R | L o[ 4 [ o jjg c G L NEW 26-ft-HIGH X o ) t J 30-ft-WIDE ROLL- =& W UP DOCR. = = -t £t 4in. | P~ o 1 " L 11 1t i3l 1 )I( ") 1 1 e 1 1 1 11 1|1 1 1 m :O ] F 7 T 7 il PART FIRST FLOOR PLAN EXISTING BUILDING £ MONORAI LEGEND AROV PART BASEMENT PLAN 101t - ) 14OdIY SSFIO0Yd ATIILIVNO NV Hin ”"L"‘:‘ ‘. N the floor up to 3 ft below the bolting flange level. The reactor tank location would be approximately centered in the 42-ff-wrde, 64-ft-long high-bay ex- tension directly in line with the ARE experimental bay. The reactor would be positioned so that the top of the shield would be at the building floor elevation, . To permit use of the experimental pits for instal- lation of auxiliary equipment, to permit possible underwater reactor disassembly work after reactor operation, and to provide a large entry door to the ART areaq, the south wall of the ARE experimental bay would be removed, the overhead crane facility would be revised from 10- to 20-ton capacity, and the truck door in the nor'rh wall of the ARE Building would be enlorged. The double-walled tank described above is essenhclly the container conSIdered for this instal- lation. The inner tank, or pressure vessel, would be approximately 24 ft in diameter with a straight section about 11 ft long and a hemispherical bottom and top. The joint between the removable top and the lower portion of the tank would probably be « bolting flange, with provision for sealing the joint with a low-melting-point alloy, Thus the top would be used only as required by the operating program. The outer tank would be a right circular cylinder approximately 27 ft in diameter and about 47]/2 ft high. About 26 ft of this water-containing tank would be above floor grade, and this portion of the tank would be ofl'ached by a flange or weld joint only when the operating program requrred that the upper hemisphere of the inner tank be in place. The inner tank would be set coaxially with the outer tank, and the bottom of the inner tank would < be 1. St above fhe bof'rom of the . outer fcnk T ' ".would prowde o ]5& cnnulus befween 'rhe fwo\ “dp “ere Pr fype shleld ‘with @ gross W ghr “of about 35 tons, would be mounted on verr:cal col‘umns ‘with "the reactor center 3 ft off center fromflt e vessel axis and about 12 ft above an open- groted floor. This positioning would provide needed space for PERIOD ENDING DECEMBER 10, 1954 movement of the portable fluoride fuel and sodium moderator coolant containers to their operating stations under the reactor. The off-center [ocation would also serve to minimize the length of NaK piping that must run from the reactor through a bulkhead in the pressure vessel to the heat dump radiators outside the tank. Quite a variety of shields has been considered for the ART. The most convenient seems to be one functionally the same as that for an aircraft requiring a unit shield, namely, a shield designed to give 1 r/hr at 50 ft from the center of the reactor. Such a shield is both the lightest and most com- pact that has been devised. It makes use of non- critical materials that are in good supply, and it will provide useful performance data on the effects of the release of delayed neutrons and decay gammas in the heat exchanger, the generation of secondary gammas throughout the shield, etc. While the complication of detailed instrumentation within the shield does not appear to be warranted, it will be extremely worthwhile to obtain radiation dose level data at various points around the pe- riphery of the shield, particularly in the vicinity of the ducts and the pump and expansion tank region. Several arrangements have been considered as a means for disposing of the heat generated in the reactor. The most promising is one that resembles a turbojet power plant in many respects. It em- ploys radiators essentially similar to those suit- able forturbojet operation. Conventional axial-flow blowers would be employed to force cooling air through the radiators. This arrangement is flexible and as inexpensive as any arrangement devised. ’WI“ give thermal capacrtles cnd flurd transit times essen’rml[y the same as those in a full-scale ireraft | power plom‘. It will also gwe some very a!uabie expenence ‘with ‘the operation of high - temperature hqurd#ro-cur hear exchangers that em- “"body features of construchon ‘and fabricating ~ ‘techniques suitable for aircraft use, It should be mem‘loned fhcf ‘the type of radiator currently en- ed i |s_yone that has been tested’ at ORNL and proposed ro “date. While’ r‘rsb performance leaves “something to be desired, it is felt that the basic ' cpnflgurotion is sound “and that modifications can ”be made Tater - fo give improved performance.. ' ‘As shown in Fig. 2.8, 16 NaK-to-air heat dump radiators would be mounted in an air tunnel which 39 , AN h o= = s Vg wend o e T i e pedrs "tombe fhe ‘most rehable of any radiator kiR e B S S0 i cne s ANP QUARTERLY PROGRESS REPORT would be placed on a diagonal across the south- “west corner of the addition. Each radiator core would have an inlet face 24 by 24 in. and a depth - of 6 in, These radiators would be located at floor grade over the NaK pipe lines and the NaK fill and drain pit. Four 75,000-cfm axial-flow blowers de- signed to give a head of 10 in, H,O would force air through the radiators and out through a discharge stack. The fuel fill-and-drain system envisioned for the ART incorporates two shielded dump tanks. One would be coupled to the reactor with a remotely - operated coupling, while a second emergency dump “tank would be welded directly to a discharge pipe ~from the reactor. The remotely operated coupling to the first fuel dump tank would give flexibility*in ' the operation of the reactor, since it would make it possible to bring fuel to or remove it from the site . expeditiously and would keep the footage of pipe ~and the number of valves to a minimum. For- handling the heavy, shielded fluoride and sodium containers inside the inner tank, a track would be installed on the floor and inside the wall. Wheels would be mounted on both the bottom and one end of the tank dolly so that the assembly could be lowered by the overhead crane to the floor track, with the end wheels on the dolly riding against the vertical track. Once on the floor track, each dolly would be moved to its operating station under the reactor. Each track pair in this area would probably be mounted on a {ift for raising the tank connection nozzle to the contact position within the reactor shield. In addition to the NaK piping bulkhead mentioned above, a control junction panel such as that shown in Fig. 2.7 would be installed as a part of another bulkhead through the tank below the building floor grade to pass the tubing and the electrical con- ductors required for operation, control, and moni- toring. Two more bulkheads, in the form of man- holes, would probably be installed in the upper portions of the inner and outer tanks. One manhole would be about 3 by 5 ft and located just above the inner tank flange to allow passage through both container walls and thus provide entrance to the - 40 inner tank subsequent to placement of these sections of equipment. The second opening would be a manhole about 5 ft in diameter in the hemispherical top of the inner tank to provide overhead crane service after placement of the top. Sufficient cat- walks, ladders, and hoisting equipment would be installed within the inner tank to provide easy access for servicing all equipment, as shown in Fig. 2.6. A - The control bulkhead in the doubl_é-wdl[ed'tank would be located so that the associated control junction panel and the conirol tunnel would be on the control room side of the tank. The tunnel would extend to the auxiliary equipment pit (formerly the ARE slab storage pit), where it would t_erminofé. The tubes and conductors from the junction panel ‘'would be channeled from the tunnel either to equip- ment in the pit and basement or to thé control room (formerly the ARE control room). The pit and basement equipment would include such items as the lubricating oil pumps and cooler, borated shield water pumps, cooler and makeup equipment, vacuum pumps, relays, switch gear, and emergency power supply. The reactor off-gas flow would probably be piped through the NaK piping bulkhead to a disposal facility outside the building., Such a system would probably consist of an activated-charcoal absorption bed contained within a pipe which would be long enough to provide the required delay period to bring the activity of the krypton (which would not be adsorbed by the charcoal) to a tolerable level. Field maintenance and laboratory facilities would be installed in the area east of the ART test bay and south of the low bay of the ARE. This area and the ARE experimental bay would be partitioned from the ART test bay with about a 16-in.-thick shield wall of stacked solid concrete block, This wall would not be erected until after placement of the upper sections of the double- walled tank, The only other major rework of the ARE facility to accommodate the ART would probably be that of modifying and equipping one of the ARE experimental pits for underwater dis- assembly work on the reactor after operation. B e ey £ o A 3 N oy - ¢ ST a5 fe R T g ey “loop test results, PERIOD ENDING DECEMBER 10, 1954 3. 'E'XPERl_MENTA_L' REACTOR ENGINEERING H. W. Savage Alrcrcft Reactor Englneerlng D|v15|on -Design work is under wcxy on an in- pt|e test [oop for operation in a horizontal beam hole of the MTR. This work is being done in cooperation with the Solid State Division and Pratt & Whitney Aircraft. The specifications for the loop have been established. Tests are under way on a horlzonfal shaft sump pump for the loop, and a salt-to-air heat exchanger has been fabricated. ‘The use of ARE-type sump pumps for high-flow heat exchanger tests is being investigated. The K-25 loop test of ARE components is being con- tinued as a life test of the ARE type pump, and over 3000 hr of operation hcs been [ogged De- velopmental work is under way on pumps for the ART, with particular emphasis being given to impeller fabrication and performance. Three Inconel forced-circulation loops have been operated with NoF-ZtF;‘-UFI‘.' The first two loops tailed after 48 and 3 hr, respectively, but the third loop has operated for 150 hr with a Reynolds number of 10,000 and a 200°F temperature gradient, A third beryllium-sodium-Inconel mass transfer test has been completed and has been submitted for metallographic examination, and components have been fabricated for tests of sodium in multi- metal loops. o Developmental work is under way on components, such as a pump and a gas-furnace heat source, for a loop for testing fused salt-to-NaK heat ex- changers. A leak test has indicated that smq[l ; |eaks of fluoride fuel into NaK are self-plugging. ~ IN-PILE LOOP COMP ature conditions within the range of out-of-pile forced c1rcu|otlon exper:ments, data should be By Teshng ot flow cmd 'remper- obtoined on Which to base extrapolations of the more readily obtainable out-of-pile test data. Of course, space limitations of the reactor beam holes somewhat limit the conditions ob’ramctb!e in a safe and reliable experiment. The proposed test unit will consist of an 8-ft loop of Y%-in. schedule-40 seamless pipe with a little over 1 ft in the active zone of the reactor. The ratio of the total volume to the volume in the high flux zone, known as the dilution factor, is approximately 10. Flows in the turbulent region, 'Reynolds number approximately 4000, with a maximum temperature of 1500°F have been pro- posed. A temperature differential of 300°F and heat transfer of 30 to 50 kw are expected with a power density of approximately 2 kw/cm®. These " conditions probably can be achieved with a fuel composed of f\lcfl“:--Zrl:A-L'H:4 (53.5-40.0-6.5 mole %). Design work is proceeding on this loop and its components to meet the above specifications. Handling equipment is planned which will fa- cilitate isothermal operation of the loop prior to insertion into the reactor and which will also receive the radioactive materials after in-pile operation, Horizontal-Shaft Sump Pump J. A. Conlin ~ D.F. Salmon Aircraft Reactor Englneenng Division --Two, ldenhccl horizontal-shaft, centrifugal-type sump pumps fabrlcc'red cccordmg to the design "descrlbed prev:ously ‘were tested with the fluo- ride mixture. ‘NaF ZrF (50-50 mole %Y at 1350°F dnd a shaft speed of 6000 rrpm for 500 and 1000 hr, ; .fespectlvely. Both pumps showed consuderable T WEGr “on the Grcphltar lapped ‘steel face pla'res," b_vbut nelther ‘seal had failed complefely. p, F. Salmon, ANP Quar. Prog. Rep. June 10, 1954, ORNL-1729, p 19. 41 .>ome m! ' the shaft arrangemen’r':' o been altered to prowde ' i ! L L ANP QUARTERLY PROGRESS REPORT a lower seal temperature; also, the shaft will be operated at a lower speed. The pump is a “‘regenerative-turbine’’ type, but it differs from ‘conventional design in a number of points. There is no positive shaft seal, but, rather, there is close clearance between the shaft and the impeller housing. Some fluid leaks along the shaft laby- rinth into the small combination sump and ex- ‘pansion chamber and is then drawn back into the impeller through a hole connecting the sump to the pump suction. To accommodate thermal ex- pansion, total impeller side clearances have been ~increased to 0.060 in., as compared with the usual UNCLASSIFIED . ORNL—LR-DWG 4522 64 o IMPELLER SIDE CLEARANCE (in.) 60 SYMBOL | FRONT [BACK|TOTAL . 0030 |0031 1006/ a 0022 0041|0063 . b6 ] . \\ o 0043 0020|0063 52 M- & 1S \\OOO IR 48 LN \ N 44 NN N 40 NN 36 2 N T~ HEAD (f1) 28 \0§\\¢ X %0 NN 24 <90, e ~ ",o,b . 20 \%i}\&\ = 5 ;;\\ 12 \'\\ -0 O 02 04 06 08 10 12 14 16 18 ' "FLOW {gpm) Fig. 3.1. Performance Characteristics of a RegenerativesTurbine Type of Horizontal-Shaft Sump Pump Tested with Water. 42 0.005 to 0.010 in. The size of the pump envelope has been reduced by locating the inlet and dis- charge ports axially rather than radially at the periphery and by decreasing the annular passage around the impeller. A Lucite model incorporating the changes made as a result of the tests described above has been tested with water. The curves of Figs. 3.1 and 3.2 give pump performance characteristics at dif- ferent speeds and impeller side clearances. As may be noted, the total impeller end clearance is a primary factor in determining pump per- formance, while the distribution of clearance, front to back, is relatively unimportant. This is fortu- nate, since the end clearance can be fixed in fabrication, but the impeller position is affected by thermal expansion and bearing alignment. ’ UNCLASSIFIED ORNL-LR-DWG 4523 72 { o8 IMPELLER CLEARANCE {in) POINT [FRONT |BACK [TOTAL | 1 0024 [0020[0044 64 ; 2 |0030 (0030|0060 . 2 10030 (0031|0061 60 N a |ooz2|0041]0063 - \ 5 |0.043 0020|0063 56 f"oo 6 |oo019 |oosi|oos0 7 |0058 0021|0079 5 \A% 52 48 \ 2 44 N\ 3 3 L N 40 < 4 FLOW: 0.92 gpm HEAD (ft} ” \\ 32 28 D 1 . 4 24 o0, T~ \ Ly : & . 20 \ 7 \2 4 o . \Z. t2 6' 8 . 0.040 0050 0.060 0070 C.0BO TOTAL IMPELLER END CLEARANCE {in} Fig. 3.2. Effect of Total Impeller End Clearance on Performance of the Regenerative-Tuthine Type of Horizontal-Shaft Sump Pump Tested with Water. TPy e ,‘;‘\ “*’ ¢ LT S 3R () ») LY The pump has unusual ability to remove gas. In water tests, air was injected into the impeller at a rate of several cubic inches per minute and was ~ continuously removed with only a slight “increase in the amount of gassing. Under normal operation there is usually no more than 1 vol % ‘of gas in the fluid. The pump will satisfactorily prime and fill the test loop in which it is installed if fluid is added to the sump while the pump is running and the total impeller side clearances are mamfalned at 0.040 in. or less. However, priming becomes errafic with the greater clecrcnces required for operation of the pump at high temperatures. Several mefhods have been demonstrated for ob- taining a posmve 'dlsp]acemenf of fluid into the foop with Iarge pump clearances. Each method, however, requires either complicated freeze valves or pressure controls. Vacuum filling appears to be the most atiractive solution to the problem but has not been ‘rested Heot Exchanger D. F. Salmon' L. P. Carpenter Aircraft Reqcfor Engmeermg D|V|5|on A 5|ng|e-tube, longitudinal ly fmned, salt-to-gir heat exchanger has been fabricated for use in an in-pile loop. Since the staggered-fin arrangement required to ensure adequate heat transfer with air resulted in a very high pressure drop, it seems imperative to use a lead shielded recirculating system with a secondary gas-to-water heat ex- changer for safety in the event of a salt break into the gas stream. The use of helium would simplify both heat exchangers, ‘dnd therefore a " ““ticipated in the orlgmal desagn, ’rhat IS, above | 2200 rpm. The crn‘lcal speed rcmge was found to 'sult-’ro-hehum exchanger is no bemg desng chcnges and to evaluate the probable trouble-free life based on actual use. "’estabhsh'\performunce ranges without ma|or deSIgn PERIOD ENDING DECEMBER 10, 1954 The test loop, set up at K-25 for eS:omining ARE fuel-tube hair-pin sections and an ARE heat exchanger, was stopped after 2047 hr of nearly trouble-free operation for removal and examination of the ARE test components. At the time of removal of the ARE test components, only a cursory inspection of the pump was made because operation had been trouble-free except for the plugging of gas nozzles with zirconium fluoride vapor condensate. An inspection of the inside surfaces of the pump tank cover revealed sta- lactites of salt that covered nearly all riser and nozzle skirts and bolts and protruded beneath the lower heat radiation baffle intoe the pump tank gas space. These stalactites are not deleterious to pump operation. The loop was modified to consist only of the pump, loop piping, throttling valve, and the previously used venturi element, and the pump has now been operated for an ad- ditional 1000 hr at 1350°F, 40 gpm, and 1500 rpm. Slight noises were noticed at the pump bearing housing after a total of 2847 hr of operation. However, these noises had not greatly increased when the total operating time had reached 3000 hr. The gas-line-plugging problem which was en- countered after 2047 hr of operation was eliminated completely by a newly designed vapor trap. Oper- ation of the loop is being continued as a life test of the pump. | Pump performance data (Fig. 3.3) were obtained over the speed range 600 to 1400 rpm with the fluo- ride mixture NaF-ZrF,-UF, (53.5-40-6.5 mole %) at 1300°F., Other tests of ARE-scale pumps were made to determine critical speeds and per- formance dcta at speeds higher than fhose an- be bef ween 2850 NO SO ':olcm’r pumps th copacmesé of“l430'gpm each at a ‘1’25 -t heod “ahd two or more NaK primary-coolant pumps with capacities of 2800 gpm or less at a 280-ft head. The problems with all these pumps will not be 43 ;nd 3250 rpm “and is fhus well cbove The speeds requtred for “heat exchanger ' b i T ANP QUARTERLY PROGRESS REPORT STONFOENTIAL . ORNL-LR-DWG 4524 €0 60 50 1400 rpm >0 —'——_—-—-.__—-_-—_~ 3 \ 20 \EFFJCIENCY AT 4200 rpm {200 rpm 63 — —'_—'_-'—-—_______‘~~ — ht " > < 30 0 : / T~ 08 = \ 5 L / w / ——___'000 rpm \ / 800 rpm ‘\ "’ / *——'_‘-—-.\ / 600 rpm \ "o 0 . o 10 20 30 40 5 0o 60 70 80 20 FLOW (gpm) Fig. 3.3. Performance of an ARE-Type Sump Pump with NaF-ZrF ,-UF , (53.5-40-6.5 mole %) at-'l-300°F. unlike those encountered in developing ARE pumps, but they will be magnified by the higher capacity requirements, by the resulting higher -power requirements, and by a number of special restrictions imposed because of the proposed locations of the pumps. The location of the fuel pumps imposes problems in impeller and in inlet and discharge housing design, since it will be necessary to match the fluid dynamic requirements of passages leading from heat exchangers into the pumps and from the pump discharge housings into the core. Various impeller des‘igns"rure being investigated and will be used to evaluate cavitation characteristics. Previous experience with ARE impellers indicates that casting is not a promising means for fabri- cating Inconel impellers in the present state of the art, and therefore a brass impeller has been 'made for studying fabrication problems and ob- taining initial performance data. Preliminary planning has been started for the fabrication of .44 models for study of pump discharge and entry conditions, - Seal design, bearing structure, cooling, lubri- cating, and driving means are other problems being studied, and the reliability of these components and operations is being ascertained through a succession of mechanical shakedown, water per- formance, and high-temperature tests. DESIGN AND OPERATION OF FORCED- CIRCULATION CORROSION AND MASS-TRANSFER TESTS Fused Saits in Inconel .. A. Mann W. B. McDonald 7 W. C. Tunnell Aircraft Reactor Engineering Division The Inconel forced-circulation corrosion test loop previously described? as loop No. 2 was put 2y, C. Tunnell, W, K, Sfcir, and J. F, Bailey, ANP Qua{. Prog. Rep. Sept. 10, 1954, ORNL-1771, Table 3.1, pdi. RN . e Myt S Pl . s ) ) ‘reducing the power input into the heater section, fuel being used is NaF-ZrF4-UF4 (53.5-40-6.5 el Reocer Enginaaring Division A third test® has been completed in which molten . “sodium was pump : : " taining a beryllium insert. Approximate test con-_ . lemperstire, 1300°F; minimon flowing. sediom <14 _temperature, less than 1000°F; Reynolds number =~ _ " at minor diamefer of beryllium insert, 190,000, D. R. Ward L. A. Mann PERIOD ENDING DECEMBER 10, 1954 into operation with a Reynolds number of 10,000 and a temperature gradient of 200°F in the fuel mixture NoF-ZrFA-UF4 (50-46-4 mole %). The design of this loop is illustrated in Fig. 3.4. A failure occurred in the heater section of this loop when the pump was stopped after 48 hr, without UNCLASSIFIED ORNL-LR-DWG 4525 to take a faulty instrument out of operation. This loop was constructed of 0.5-in.-0D, 0.020-in.-wall Inconel tubing. Three unsuccessful attempts have been made to get the loop back in operation, but, in each case, failures of the tubing occurred during the cleaning operation with barren fuel mixture. Another loop (loop No. 3) constructed of 0.5- in.-OD, 0.040-in.-wall Inconel tubing was suc- cessfully started and a Reynolds number of 10,000 and a temperature gradient of 300°F were attained. After operation for 3 hr, the test was terminated because of failure of tubing near a welded joint. The data from the startup revealed conclusively that the loop could be operated without a heat economizer; therefore construction of the loop can be simplified. A third loop, without the economizer, was built with 0.5-in.-OD, 0.045-in.-wall Inconel tubing. This loop included a sump pump (model LFB), a cooler, and a heater. The loop has been in operation over 150 hr with a Reynolds number of 10,000 and a 200°F temperature gradient. The e , HEATER mole %). The maximum fuel temperature of 1500°F is produced by a power input of about 72 kw. This loop is scheduled to operate for 1000 hr. Beryll i'qusbd ium-Inconel Ma s_s' Trdnsfg‘r W BM‘cDonald ) :Fi‘g. 3.4. Fdrced-Circuhfio_n 'Cr_ollfrbsii:ilh Test " l.oop. ‘ah an Ineanal leas cons and other sections of the loop have been sub- ough an Inconel loop con- " LS B SUR EER TENEORETE ST et oo 2o mitted for metallographic examination, A fourth ow rate, 3.5 gpm: berglliom test loop of similar construction has been fabri- Do o P oo cated and is being instqlled for testing. s Sodium in Multimetal Loops i .. W.B.McDonald - ... . Aircraft Reactor Engineering Division . The design of a |oE>p4 fqr feéfihQ éofn_bincfions of structural metals in contact with high-velocity 3L. A. Mann and F. A. Anderson, ANP Quar. Prog. 4. A Mann, ANP Quar. Prog. Rep. Sept. 10, 1954, Rep. Sept. 10, 1954, ORNL-1771, p 32. ORNL-1771, p 41. 45 Y, T e Vo i o] 'ANP QUARTERLY PROGRESS REPORT turbulent liquid metals under high temperature differentials has been completed, and components have been fabricated. The first series of tests will include combinations of Inconel and type 316 stainless steel as loop materials, with sodium as the circuvidted liquid metal. Nutural Gas Heui‘ Sources for Forced-Cnrcuiahon Loops L. A, Mann Alrcrafi Reactor Engmeermg Division R. Curry 7 ER Dytko Profi & Whl'rney Alrcraf’r Several ’rests of burners were made in v order to obtom prellminary data for design of a natural-gos heat source for forced-circulation corrosion test Ioops. It was found that exhaust gas temperatures of between 2800 and 3300°F could be obtained without initial air preheating, and, by using air at a rate of 65 to 650 cfm, the equivalent of 100 to 1000 kw of heat was released. Maximum system thermal efficiencies of 20 to 60% appear to be feasible. Tests with eleciric-resistance heated units indi- cate a maximum wall temperature of 1750°F with the tubing lengths required by pressure-drop and flow considerations. Somewhat higher tube-wall external element temperatures or proportionately longer heater sections would be required with gas heating. An economic study indicates that there is little choice between providing electrical power supplies and providing large-capacity air supplies for gas heating. Therefore, since development of a gas-heated unit specifically for this purpose would be expensive and since there is no present proof of reliability of components and operation of gas-fired loops, further study of gas firing for these test loops appears to be unwarranted. HEAT EXCHANGER DEVELOPMENT ‘ Heat Exchanger Tests R. E. MacPherson - Aircraft Reactor Engineering Division R. D. Peak Pratt & Whitney Aircraft A number of tests of fused salt-to-NaK heat exchangers are to be made for determining the fabricability, operability, operating parameters, 7.and reliability of heat exchanger tube bundles such as those proposed for the ART. The test 46 loops are to include reliable pumps, reliable heat sources, and adequate heat sinks. A preliminary heat exchanger test reported previ- ously® indicated reasonable structural integrity of the exchanger (1682 hr of operation), but little data on operating parameters were obtained. For the next experiment, a regenerative system is to be used in which the heat exchanger will be an economizer that will transfer about 4 Mw of heat, while a heat input and release of only 1 Mw will _be reqmred Header Leak Test R. E MacPherson A:rcraft Reuctor Engmeerlng DIV!SlOl"l R. D, Peak Pratt & Whitney Aircraft It has been postulated that small leaks which develop in metal walls separating NaK and fluo- ride fuels may be self-plugging as the result of the precipitation of high-melting-point reaction products. A program of tests is currently under way to check this possibility. To date, one test has been run, which indicates that self-plugging can occur with NaF-ZrF,-UF, fuels. The leaks are installed between three cells as shown in Fig. 3.5. The center cell is pressurized to leak into the end cells under the pressure differentials shown, 5 and 50 psi. The test rig was filled with NaF-ZrF ,-UF , fuel in the center cell and NaK in the end cells. The leaks were about 0.002 in. in diameter. After 12 hr at 1300°F, there was no apparent leakage, and x-ray pictures taken at intervals during the test showed no changes in liquid levels. Pump for Heat Exchanger Tests W. G. Cobb W. R. Huntley A. G, Grindell R. E. MacPherson Aircraft Reactor Engineering Division A series of tests is being made in order to determine whether an existing model DANA pump (ARE moderator-coolant-type sump pump) is suit- able for the high-speed operation which will be required in the intermediate heat exchanger test loop. Critical shaft-speed determinations made by K-25 personnel indicated that the pump could be operated saofely at the proposed speed of ap- proximately 3700 rpm. SA. P. Fraas, R. W. Bussard, and R, E. MacPherson, ANP Quar, Prog. Rep. June 10, 1954, ORNL-1729, p 22. i, L ¥ ’? » - iy "flnveshgatlbjns of gas fired furnfl‘aces E:cpable Qf transferring h_ea’r to [quId mefal‘uf a hlgh ‘rate } reoson. PERIOD ENDING DECEMBER 10, 1954 UNCLASSIFIED PHOTO 22735 : . = CENTER END CELL f CELL END CELL Fig. 3.5. .H..eai E.x(:hanger-l'!e;dder Leak Test, ' The rotary element of the pump wdsr'rfhen oper- combined in forthcoming ‘ex_perimenfs. The problem ated at 3700 rpm in the cold shakedown test stand, is also being explored with vendors of gas-fired and the bearings and mechanical oil seals were equipment. found to operate satisfactorily. The pump is now . being observed for performance qnd mgassmg at Sodium-Cooled 100-kw Furnace Tests hrgh speeds in fhe wo’rer test stand. R, E MocPherson o R W, Bussmd nay 'be requrlred H’&Vé’hgweh'—n : gas-f:red sodwm heat source wds operaf@d wwh : an ou?put of up to 85 kw at sodlum oufle'r temper- coson gtures of ]500°F The mstaliatton 1!8 “shown in SN _.,.‘ A LT AT nlf" volfm'\e‘wnthout exfenswe heat stor : F|g 3.6. A compcrlson of &emg,n cgnfltf_q_qns w:fh al ¢ ) hr of e was ac pmu- lated uyhd‘ér ‘the condlhons “given in Table 3.1. _ “"Since the efficiency was lower than desired, the Two approaches have been investigated experi- system became limited by gas flow because of mentally, and the best features of both are to be high pressure drops on the gas-supply system af 47 b et s 4 o e e NATURAL GAS MANIFOLD S £ Fig. 3.6. Gas-Fired Heat Source Test Installation. 48 UNCLASSIFIED . PHOTO 22793 ¥ v . »~ 2 Ky [ h 5""'furo| " weakening rather than by sample melting. The weakening of the walls was probably caused by TABLE 3.1. GAS-FIRED FURNACE OPERATING CONDlTIONS De;ign Aéqul Sodium outlet femperofure, °F | o 1500 1500 ) VSOdIUm inlet femperofure, °F | 1100 1215 | Sodium AT, °F 400 285 Sodium flow, gprfi: S ‘ 7.3 8.7 Heat output, kw 100 85 o heat removed Efficiency : heat relecsed X 100), % 66 55 the 85-kw level and test operated at lower femperatures (1100 to 1400°F), higher efficiencies were obtained, not only because of the increased temperature differ- ential between-the flue gas and the sodium but also because of the clean condition of the inside surfaces of the furnace. " Over 100-kw output at essentially design effucnency was attained through- out operation in the lower temperature range, but, because of an unsohsfoctory flame-failure .safety system, sustained operation was impossible. After the unsuhsfoctory condmon had been corrected, continuous operahon ‘at 1500°F was resumed, and it was found again that 55% was the best ef- ficiency obtainable. The furnace performed exceptionally well as a compact, high-capacity heat source. At the maximum operating condition of 166 kw total heat released, the release rate based on the combustion chamber volume was ~over 1.5 Mw/#t3, or over 5, 000 OOO Btu/hr/ff3 When 'rhe unn‘ was f!rst fired It is fe[f that ’rhls heat- 'Vrelease rote is hlgh smce ull combushon was probably not accompllshed |n “the 0.1-#3 volume“:“ ‘ of ’rhe combushon chamber, qs ie"\l:denced by 'rhe“ PERIOD ENDING DECEMBER 10, 1954 local boiling of the sodium coolant, by accumu- lation of a gas pocket in the line, by external corrosion in the hlgh—temperature exhaust gases, or by higher-than- predlcfed local gas- -side heat transfer coefficients. The coolant tubes were arranged in ten pora“ei passes that were mani- folded into two ‘banks of five passes each. It was observed in operation that the sodium outlet temperature from one bank of five tubes was as high as 1550°F, while that from the other was only 1480°F. Temperature variation from tube to tube within each bank of five tubes could con- ceivably have been equally bad and have resulted in hypothesized sodium temperatures of 1600 to - 1650°F. The bmlmg point of sodium is 1620°F at utmospherlc pressure and about 1800°F at 2 atm, and thus local boiling of the sodium, together with consequent reduction in the local heat transfer coefficient, is certainly a con- ceivable possibility. Although it seems possible that small amounts of gas were trapped in the spirally wound tube coils d‘uting initial filling with sodium, the location of the tube failures makes it difficult to visualize the accumulation of gas in stable gas pockets as contributing to the failures. Some external corrosion and erosion of the tubes were observed, but there was no strong evidence that they were advanced enough to have been the cause of failure. The gas-side local maximum heat " transfer coefficient could have been con- siderably higher than the predicted value of 60 Btu/hr.ft2.°F because of nonuniformity of gas flow and combustion, and burning could have taken place on or very near to the tubes in question. __This seems p055|b!e in view of the location of _the fcn[ures relative fo the oufle‘r bends of the tube c01| ‘mcxmfolds. These outlet tubes bend _‘ichrectly across the chamber perpendlcular to and __jns:de the lnner ‘row ‘of wound tubmg. In this "_7‘_)4|ocaf|on, ’rhey could have’ acted ‘as flame stabi- ~ lizers for the uncombusted gas’ mixture “and ‘thus __,_;_',_:caused hngh local coeffncuent_s : V'T‘the first row behmd them. |'r_seems‘"probcfl:le that “the tubes in combma’rlon of non f!ld‘w ‘through _fhe' ten parallei flow tubes and a hlghpr thcn- pred:cfed local heat trqnsfer coeff:menf was the -~ “cause of failure. ~The “pertinent ‘parts “of the as- sembly are being examined by the Metaliurgy Division in order to obtain more information re- garding the failures, 49 c 4 s e ot bt b Bl i il ANP QUARTERLY PROGRESS REPORT Natural-Gas Burner L. A. Mann Aircraft Reactor Engineering Division R. Curry Pratt & Whitney Aircraft The exploratory tests on gas burners, described previously,® were used as a basis for new burner developments. Designs for obtaining better mixing of air and gas and higher pressures in the burning chamber have been investigated, and some tests have been made in which preheated air and gas were used. The test units were based, in general, ~on designs developed by the Esso Marketers in the past few years. |In these units, the burning chamber serves also as a flame holder and a " burning-rate controller. A schematic diagram of the basic design is shown in Fig. 3.7. The first burner that was built was operated at approximately 3100°F and 400-kw heat release without. a pressure-forming barrier at the exit of the burner. It was then operated at about 3200°F and 250-kw heat release at about I/3 atm pressure drop from burner to air to simulate a heat ex- changer after the burner. A second burner was designed on the basis of data from the first burner, plus specific provision for preheating the air and gas. Tests were con- ducted in which a temperature of about 3300°F and 1-Mw heat release were obtained by using unpreheated air and gas. Auxiliary tests indicated that premixed air and gas fuel could be preheated to 1600°F or more without serious preignition. TRAP FOR FLUORIDE VAPORS W. G. Cobb A. G. Grindell W. R. Huntley Aircraft Reactor Engineering Division In operating a high-temperature system with a molten fluoride salt containing appreciable quan- tities of zirconium fluoride and blanketed with inert gas, it is difficult to prevent plugging of the gas lines leading to or from the vessel. A .. __.vapor-trap device has therefore been developed b".-_rlv\'rhkich provides a sufficiently large vapor con- " :denser to prevent plugging in an operating period ~“i.of several thousand hours with reasonable gas " movement (no excessive continuous bleed) and to 'f_e:_rix'a‘B'le adequate heating of the gas line between 6. A, Mann and L. F. Roy, ANP Quar. Prog. Rep. " Sept. 10, 1954, ORNL-1771, p 41. 50 UNCLASSIFIED ORNL-LR-DWG 4526 BURNED GAS S Ss SY ] e — —_—— —_—————— e Ul GAS-AIR MIXER Fig. 3.7. Natural-Gas High-Intensity Burner, the trap and the fluoride-containing vessel. Gas line plugging with zirconium fluoride vapor condensate occurred frequently at the pump tank of the ARE component test at K-25, Plugging was eliminated by heating the gas line to 1300°F and by installing the vapor trap so that no cold surfaces for vapor condensation were available until the gases reached the large cross-sectional area of the trap. This type of unit was first tested Dyiv ey m Py . 74 w NP F i & 5 while connected to a small tank of relatively . stagnant NaF-ZrF -UF, (53.5-40-6.5 mole %), at 1400°F. The unit operated for 500 hr, during which time approximately 125 t3 of helium was . bled into the tank and out through the vapor trap. No plugging of the trap or of the Z—in.‘ gas lines had occurred at termination of the test. Exami- nation of the vapor trap and heated tube revealed that all heated surfaces remained cledr and that w? C by o | PERIOD ENDING DECEMBER 10, 1954 condensation had occurred on only the cooled surfaces. A second unit was then installed in the test loop at K-25, and approximately 70 3 of vapor-laden helium was periodically vented from the pump tank without plugging. The trap has since operated over 900 hr with more than 100 periodic vent operations without plugging. Similar traps installed in the ARE fuel pump were also successful. 51 T T T T T ey [T "ANP QUARTERLY PROGRESS REPORT 4. CRITICAL EXP ERIMENTS A D Calhhcm - " Physics Dwns:on REFLECTOR MODERATED REACTOR ot B. L. Greenstreet Alrcmff Reacfor Engmeermg Division ~{'-:?:'-"7 J J Lynn D. V. P. Williams o F’hysms Division , R. M. Spencer _ Unn‘ed STafes Air Force J S. Crude]e E. V. Sandin .70 40 W. Noaks o Pratt & Whitney Aircraft The second of two critical assemblies of the reflector-moderated reactor was constructed as a part of a program presented previously.! The reactor is composed of a beryllium reflector and a " fuel region of enriched uranium metal and Teflon (CF2)n. The first assembly, an essentially spheri- cal fuel core surrounded by the reflector, was described in the previous report.2 Since it con- tained no extraneous structural materials, the experimental results could be compared with the results of multigroup reactor calculations. The results of the second experiment were also to be correlated with the calculations, since the second assembly, too, was unpoisoned by structure, The assembly has a central polyhedral core, or isfand, of beryllium about 10 in. in diameter enclosed by a 4.5-in.-thick uranium-Teflon fuel annulus. The fuel is surrounded by an effectively infinite be- ryllium reflector. Aluminum shells, 0.065 in. thick, separate the fuel from the island and from the reflector and provide structural stability, The assembly is completely described elsewhere,® and a summary of its composition is given in Table 4.1, Figure 4,1 is a view of the mid-section. The calculated? critical fuel loading was 7.16 kg of U235, or 0,142 g of U235 per cubic centimeter of fuel annulus. The first critical experimental array of 7.39 kg at a density of 0.146 g of U233 ID. Scott and B. L. Greenstreet, ANP Quar. Prog. Rep. Mar, 10, 1954, ORNL-1692, p 45. 2D, Scott et ale, ANP Quar, Prog. Rep. Sept. 10, 1954, ORNL-1771, p 44. '3B. L. Greenstreet, Reflector-Moderated Critical ' Assembly Experimental Program — Part 1I, ORNL CF-54. i ,10-119 (Oct. 19, 1954). W.V_E. Kinney, private communication, Oct. 25, 1954, 52 per ‘cubie centimeter of fuel cnnulus confomed about 7.5% excess reactivity, In ‘order to com- pletely fill the fuel region without excess control rod poison, it was neceéssary to decrease the loading to 4.66 kg of U235 at a density of 0.092 g of U233 per cubic centimeter of fuel annulus. This latter loading has been shown, by calibrated control rods, to contain 0.92% excess reactivity, equivalent to about 0.31 kg of U235, which gives 4.35 kg of U235 as the critical mass of the unpoisoned assembly, ' A series of measurements of neutron and fission- rate distributions and of reactivity coefficients will be made prior to the incorporation, in the structure, of additional materials to simulate reactor components, TABLE 4.1. COMPOSITION OF SECOND REFLECTOR-MODERATED REACTOR CRITICAL ASSEMBLY Beryllium Island VYolume, ff3 0.37 Average radius, in. 5,18 Fuel Annulus {exclusive of aluminum core shells and interface sheets) Volume, 2 1.78 (50.4 liters) Average inside radius, in. 5.18 Average outside radius, in, .57 Mass of (kg) Teflon 99.38 Uranium 5,00 U235** 4.66 Uranium coating material 0.05 Scotch tape 0.12 Core Shells and Interface Sheets Mass of aluminum, kg 5.85 Reflector Volume, ft° 22,22 Minimum thickness, in. 11.5 Mass of (kg) Beryllium 1155.0 Aluminum 29.2 *Density, 1,97 g of Teflon per cubic centimeter of fuel annulus. **Density, 0,092 g of U235 per cubic centimeter of fuel annulus, @ PERIOD ENDING DECEMBER 10, 1954 PHOTC 22782 - ical Assembly. i ™ o deiutéd Reactor Cr $Second Reflector-Mo Flgo 4.10 - ™ * < 5 & o e G- Part 1l L ® “t e . 4 T TR v TN T X e =» o fe, A Oua 1954 ORNL- 1771, Flg.5'| p55. 0RNb-1976 5. CHEMISTRY OF MOLTEN MATERIALS ~W. R, Grimes 7 Mdteriols Chemistry Division Study of the NaF-ZrF ,-UF, system by the use of various techniques, mcludmg quenching, thermal 'onalysm, and filtration, has continued. Improve- ments that have been made in quenching methods are expected to accelerate the availability of data by this method. Phase studies of the UF,-bearing fluoride systems have continued to be emphasized because of inferest in their favorable corrosion characteristics, It is believed that problems attendant fo handling the easily oxidized UF, will be alleviated to some extent by extendlng phase studies to include LaF;, an isomorphous stable substitute for UF,. Investigations of the chemlcc[ reactions in molten salts are conhnumg in an effort to understand the " mechanisms mvolved in the purlflcahon of fluoride mixtures and in ‘the reduction of UF, to UF3 in fluoride melts. Slmulcr studies of the corrosion products of Inconel in fluoride mixtures and of the reduction of UF, by structural metals are under way. A sufficient stock pile of purified zirconium fluoride-base fluoride mixtures will have been ‘prepared by the end of this calendar year to allow termination of operation of the large-scale (250-1b capacity) production foculr'ry. Careful control of operating conditions in the pilot-scale focrln‘y has allowed production of sufficiently consistent UF, content material to permn‘ its release for corros;on testing. Elecfrolyflc purn‘ncat;on of NaF- -ZrF , mixtures has been adopted, ‘and sugnlflcanf cuts_ _in processing “time have resulted, Howe ot been posmble fo. CUf the processmg time Of»--m allows a temperature gradient to be imposed upon a nickel capsule opproxumafely 4 in, in length that 77 N M SN Y T o O v SN 3 = NN o N = P . =2 Fig. 5.1. Phase Diagram of the NaF-ZrF -UF System, rests on a nickel wire support, which, in turn, rests on a movable nickel platform at the bottom of the hole. Temperatures are determined at 11 points along the sample length by means of thermo- couples and can be varied by adjustment of the various furnace windings. By moving the platform to one side with a handle above the furnace, the capsule and wire support are allowed to fall into an oil quenching bath, Pfidse-Rélutioqships in UF3-Bearing Systems R. E. Moore R. E. Thoma ~ Materials Chemistry Division H. Insley, Consultant Study of UF;-bearing systems has been con- tinved on NaF-ZrF ,-UF,, NaF-UF,, KF-UF,, and Zrl::“-UF‘,-UF3 mixtures, In addition, work was begun on NaF-LaF; and KF-LaF; systems, since LaF, ‘is known to be a stable *'stand-in'’ for the more difficultly handled UF,. Work on the KF-UF, mixtures has revealed that a red phase, formerly designated as K3UF,, may have as much as 75% of the uranium in the tetravalent state; its apparent homogeneity over a wide range of UST/U%Y is possibly due to a solid solution of K;UF, and 58 K;UF,. The existence of the latter has yet to be demonstrated. It is now thought that NaUF ,, rather than Na,U,F,, as previously postulated, is the empirical formula of an incongruently melting compound in the NaF-UF, system. UF, in ZrF -Bearing Systems. Studies of ZrF ,- UF ,-UF, mixtures were reported earlier.3*4 The more recent of the reports? indicated a probable composition of 50 mole % ZrF4-.-25 mole % UF4—-25 mole % UF, for a brownish, slightly birefringent phase noted in this system. Recent studies of mixtures in this system that were prepared by adding uranium metal to ZrF ,-UF, mixtures and stitring the partially reduced melts until they solidified indicated that the mixture with the above composition consists of two phases, while a mixture containing 33.3 mole % ZrF ~33.3 mole % UF,-33.3 mole % UF, is essentially a single- phase brownish crystalline material that is be- lieved to be a compound. Another phase observed at higher ZrF, concentrations, usually described as olive-drab, appeared to be the only phase present in a mixture containing 75 mole % ZrF ,— 12.5 mole % UF,-12.5 mole % UF,. All these compositions are theoretical compositions based upon complete reaction between UF, and uranivm metal in the molten material. The meager analyti- cal data available for such mixtures indicate that the reactions probably did not go to completion in all cases, and so the percentage of UF is likely to be higher and the percentage of UF; lower than are shown by the formula. In the NaF-ZrF -UF; system, four compositions wete prepared along a join connecting the minimum melting compositions in the NaF-ZrF, binary (42 mole % ZrF ,; mp, 500°C) and the NaF-UF; binary {29 mole % UF,; mp, 715°C). Liquidus tempera- tures determined by thermal analysis for the mixtures containing 5, 12, 20, and 25 mole % UF, were 650, 775, 865, and 775°C, respectively., The 25 mole % UF, mixture contained free UF,, in addition to the NaF-UF; complex and Na,ZrF . KF-UF; System. Earlier studies of KF-UF, mixtures5 indicated the existence of two com- plexes. A red crystalline phase observed as the 3V. S. Coleman, C. J. Barton, and T. N. McVay, ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556, p 41. 4C. J. Barton et al., ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL-1609, p 57. 3¢C. J. Barton et al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 59. ' . principal phase in mixtures containing 75 mole % 2 KF was believed to be K;UF, while a blue phase ; " with optical properties snmllar to those of the blue NaF-UF, complex was assigned the formula : K3U,Fg. It was also previously reported® that all | KF-bearing reduced-uranium mixtures that were analyzed chemically contained varying quantities of tetravalent uranium‘, usually in a form that could hot be recognized as such either petrographlcolly or by x-ay diffraction analysis, Some progress was made durmg the past quarter in gaining a better undersfondmg of phase relationships in this KF system, Investlgahon of the KF-UF, system has been considerably aided by begmn:ng the study of the KF- -LaF, system, The compound LaF, is isomorphous with UF,, has almost identi- cal lattice dimensions, and, furthermore, has only one stable valence state (3+) whereas UF maintained with dn‘flculty. Dergunov’ published partial phqse diagrams for the alkali fluoride~LaF, sysfems, and Zocharncsen8 has studied mtxtures of KF and NaF with LaF; by an x-ray diffraction technique. Both w°rkers indicated only a single compound, KLaF,, formed by these two compo- nents, According to Dergunov, it melts incongru- Y “+ & ently ot 770°C, It seems probable, therefore, that the compound that was called K3U,Fy was actually S . KUF,, with an incongruent melhng point of - 820 i 10°C, Tests performed duting the past quarter demonstrated that the red crystalline phase formerly designated K,UF, can have as much as 75% of the uranium in fhe 'retravalenf state. Since this phase is usually cubic, as is K3UF7, and has a refractive index near that of K 4UF,, the most hkely explcnahon of the h mogeneity of the “3+/U4+_ ratios 77560 to 500°C and helps to explain why it is dif- * ficult to detect ||qu1dus temperatures in mixtures " of this type by thermal cnqums., It is probable that the solublhfy of UF, éutectic is htgh enough fo provrde a usable fuel "";"""*sysfem for a reflector-moderated c:rculcflng-fuel reocfor. ' rred phase over a wi - K,UF,, a 4UF in 'rhe presence of cfkah-metal fluorides has shown ~that potassium fluoride has a strong inhibiting effec'f on the complehon of the reaction. It was ':"found that with the KF concentration of the eutectic mixture (NaF-KE-LiF, 11.5-42. 0-465 mole %), a little better than 50% conversion of the UF, to UF3 could be expected. With this in mind, new attempts were made to produce consistent batches, and it soon became evident that the handling technique was quite important. After several changes in processing techniques were tried, the most promising method was chosen as a standard procedure. Briefly, this method utilizes separate purification steps for the main constituents in- volved. The eutectic (NaF-KF-LiF, 11,5-42,0.46.5 mole %) is first heated and stripped with H,. The melt is cooled and the UF4 is added, after which the melt is again heated and siripped. The melt is then cooled, and uranium metal is added. The final heating and stripping are then carried out, and the batch is transferred to a storage can. In this manner, drying and purification of the alkali metal fluorides are accomplished without -danger of hydrolyzing or oxidizing the UF ; also, . thorough mixing of the UF is provided before the ““uranium metal is added. A further advantage is that the batch is well purified and miscellaneous - _side reactions of impurities with the uranium metal .. _are kept to a minimum, A series of batches was processed to demonstrate In previous concentration ranged from 0.5 to TABLE 5.6. RESULTS OF ANALYSES OF NaF-KF-LiF EUTECTIC*.UF, PREPARATIONS Uranium Centent impurities (wt %) (ppm) u3* Total U Fe Cr Ni - 5.39 8.09 155 70 45 4.83 10.9 95 30 85 6.63 12,5 220 30 . 75 4,92 12,3 60 25 35 5,60 10.5 145 5 1510%* 5.02 1.5 140 30 45 5.86 1.2 80 25 25 5.08 1.4 150 30 60 5.70 10.6 430 28 18 5.47 10,9 215 30 135 *(11.5-46,5-42,0 mole %). **Note: to be rechecked. such material will be accepted for corrosion studies. Since it was known that the reaction U + 3UF, —> 4UF, < did not go to completion, the presence of uranium metal in the reactor vessel after completion of a processed batch was expected and was verified by examination of the heels left in the vessel. A rigid program of equipment preparation was also initiated. All reactor vessels and receiver vessels are flange-topped to allow access for cleaning, and the reactor vessels are equipped with nickel liners to allow easy removal and cleansing. Receiver vessels are presently equipped with Inconel liners because attack by HF gas is no longer a problem since HF was eliminated from the process. Each unit is completely disassembled after each run, and all par'rs are cleaned or replaced when neces- sary. * Purification of KF ond RbF ' C. M. Blood Materials Chemistry Division Several batches of KF and RbF have been proc- essed to meet requirements for these materials in purified form. The RbF, when received, ;entdined . .e-\-;-j s ¢ T f 3 . "“"‘55 o S L “-sf i o PERIOD ENDING DECEMBER 10, 1954 5 as its chief impurity about 0.2 wt % sulfur. Purifi- Fluorination of NiF,, CrF,, and FeF, has been s ~ cation of the RbF has proved to be difficult due to discontinued because the analytical = methods > attack of the nlckel purification equipment by the available are not able to establish the fluorine-to- ' v _ sulfur. It has been found possible by rigorous HF metal ratios accurately enough to determine the % and H, treatments to reduce the sulfur content to effectiveness of the BrF, ftreatment in removing " an ccceptcb!e volue, but in the process a large small amounts of oxides that may be present. amount of nickel is picked up by the melt. Con- siderable treatment time is required to reduce the resulting nickel fluoride so that it may be removed L. G. Overholser F. Kertesz from the product, Materials Chemistry Division CHEMISTRY OF ALKAL!I HYDROXIDES Preparahon of Vatious Fluotides Purification of Hydroxides - B, J. Sturm ) E. E. Ketchen‘ | ' | ~ E.E. Ketchen T, D Materlals Chemls‘rry Division Materials Chemisffy Division ' : ~ Approximately 5 kg of NaOH was purified either by filtering a 50 wt % aqueous solution through a fine sintered-glass filter or by decanting a 50% solution to remove insoluble Na,CO,. In either case the material was then dehydrated at 400°C, Both methods produced material of desired purity The preparcflon “and purn‘lcahon of various structural metal fluorides and several complex fluorides derived frbm these and alkali fluorides have been conhnued These materials cmd hydro- fluormuted UF4 are being utilized at increasing rates for various investigations of interest to the ANP program. Chemical analysis, supplemented (~0.1 wt % each of H,0 and Na,CO,). About " & by x-ray and petrographic examination, has been 1.8 kg of high-purity KOH (~0.1 wt % each K,CO,, . - utilized to determine purity and ldenhty of 1he H,0, and Na) was prepared by reacting pure po- " tassium with water, It is not planned to prepare more hydroxides in the near future because the material on hand should supply the present limited requirements. ' N .y . o _muterlais. _ o ,. ‘ Uranium tetrafluoride, as received, may contain small amounts of oxides, moisture, and higher- valence uramum._ Hydrofluormahon at 600°C is being carried out as a satisfactory method of re- * o Effect of Additives on Hydrogen Pressure s moving these impurities, During the quarter, 11 kg Over NaOH-Ni System ' . of the purified material was supplied for studies ' | y requiring high-purity UF ,. Additional batches of F. A. Knox NIF were prepared by the hydrofluorination at Materials Chemistry Division 600°C of partially dehydrated NICl,6H,0. Ane The effects of small additions of Ni, Na,0, hydrous AlF; was prepared by Te thermal de- and mixtures of the two on the hydrogen pressure _composmon Of (NH4) 3AIF at 500°C. The latter e the NaOH-Ni system have been investigated v o was synthes:zeg by heafmg Ang 3/2H O w”h - by using a modlflco’non of fhe prevaously described oo NHy HF2 at 125°C. Two pounds of K N’F was appardfus. 28 It had been found previously that V_w"w"fprepared by reac'rmg fhe apprc;prlcn‘e qucnhttes large additions of NIO resulted in the formunon of | r°f KHF2 and N'F 4H O at 800 C exfracfmg the ‘water vapor with no detectable hydrogen. A smaller ' was synthemzed by -heafm 9 and CrF 3/H O »_mckel capsule ogcnn gcve evtdence of ‘water vopor, “at 800°C the” pressure was only 31 mm, with water :‘f‘prepar'éd by hem‘mg an | excess‘ Of NH HF2 with ™ When 0.5 mole % was added, a hydrogen pressure CrF 3/H ,0 at QOOOC Pcm‘ of fhe (NH ),CrF in excess of soturated water vapor pressure re- .\..-;.;”suh‘ed and the gas collected from the system was to Y'eld C"F3: “and cnofher P°”'6"r “aft found to be ‘about” 95% hydrogen. “Therefore, the - decomposed to CrFy at 600°C, qufl_r‘tre.qfed W”‘h _-_,;;'_Iorger addmons of NlOGre seen as bemg suffimen’r " hydrogen at '800°C 1o give CrF a o - e < Three b“"rChes. of FeFy We“" ””"””f”ed as a 23ANP Quar. Prog. Rep. Sept. 10, 1953, ORNL.-1609, means of removing small amounts of impurities. p 84. o ey . 71 - -\,:-l( v e ? _-,addmon of 3 mole % NIO to purlfled NQOH ina | 'droplets found in fhe cool porhons ‘of the appcrotus.' ANP QUARTERLY PROGRESS REPORT fo cause oxidation of the hydrogen from the NaOH- Ni reaction to water. Addition of 2 mole % Na,O also resulted in oxidation of the hydrogen to water. At 800°C the pressure with this addition was found to be only 33 mm, while the hydrogen pressure of the NaOH-Ni - system was 126 mm at this temperature. A 2 mole % addition of an equimolar mixture of Na,0 and NiO to NaOH gave a pressure of 33 mm at 800°C. Evidence indicates that a substantial : frachon of fl-us pressure is due to water vapor, FUNDAMENTAL CHEM]STRY OF FUSED SALTS Solublhty of Xenon in Fused Salts R. F. Newton Research Director’s Department Smce the tenfohve va[ues for the solubility of - xenon in the NaF KF- LlF eutectic and in the KNO -NaN03 ‘eutectic were reported, 24 experi- "-'menfs have shown that leakage of xenon through the frozen seal was possible. To minimize this. possibility the apparatus was redesigned to have at least 10 cm of liquid above the frozen section on the xenon side and to be about 6 cm long. The new design also permits returning the fused salt by adjusting the pressure difference and melting the frozen seal, While the metallic apparatus for use with the fluoride eutectic was being rebuilt, the KNO,- NaNO, eutectic was reinvestigated in a glcss apparatus that incorporated the new design features described above. The nitrate mixture can be used without significant decomposition up to about 450°C; in 6 hr, ot temperature,about 0,01% of the nitrate was decomposed at 410°C and about 0.3% was decomposed at 500°C, With the new apparatus, values of § x 10~% and 9.3 x 10=% mole of xenon per milliliter of melt at 260°C and 8.9 x 108 and 9.6 x 10-8 450°C were obtained. These values are essentially in agreement with the previous ones, namely '-‘8,5'><. IQ"‘S at 280°C and 10 x 10~8 at 360°C. “The change in solubility with temperature appears 1o be very small, . X-Ray Diffraction Studies in Salt Systems P. A. Agron M. A, Bredig Chemistry Division MF.XF, Binary Systems. The renewed interest in the binary systems of alkali fluorides with uranium trifluorides was emphasized this past quarter. In studying these systems, difficulties arise from partial oxidation of U(HI) to U(IV) and/or through disproportionation of the uranium in solution in the molten alkali fluorides. It thus appeared to be worth while first to obtain un- equivocal x-ray data on the corresponding binary fluoride complexes of some of the ‘‘4f rare earth’’ metals,2? especially of lanthanum, which may be considered as a good “*stand-in’’ for U(H1). Zachariasen reported2é several trifluorides of the “‘Sfrare earth’’ group as having the same ‘“tysonite’’ structure as those of the ‘‘4f'' group, previously determined by Oftedal.?” For the trifluorides, the average La-F and U-F distances are given as 2.50 and 2.56 B\, respectively. The lattice di- mensions of these hexagonal cells differ by only 0.05%. Thus it is not surprising to discover in samples prepared here?® that the structure of the double fluorides?® of NalaF, and KLQF4 had their counterpart in analogous NGUF4 and KUF, structures, According to Dergunov, in MF-LaF, systems all alkali fluorides, with the possible exception of CsF, form 1:1 compounds, and only CsF forms a 3:1, CS3LGF6, congruently melting compound. These data were obtained by visual observation of crystallization from melts, Compounds below the eutectic that are stable and polymorphous transitions would not have been discerned. In the present work, fused mixtures of the binary systems MF-LaF, (except 3NaF-LaF, and LiF mixtures) of the molar compositions 3:1 and 1:1 were made available from thermal-halt measure- ments.2? X-ray diffractometer patterns were ob- tained on these melts, The phases found in the NaF and KF systems are indicated in Table 5.7 and are the phases expected for these binary com- positions., Excellent agreement is shown for the 7 24%. F. Newton and D. G. Hill, ANP Quar. Prog. Rep, Sept. 10, 1954, ORNL-1771, p 70. 23y, H. Zachariasen, Acta Cryst. Am. Chem. Soc. 70, 2147 (1948). 1, 265 (1948); J. 26y, H. Zachariasen, Fluorides of Uranium and Tho- rium with the LaF3 Type of Structure, MDDC-1153 (date of manuscript, June 1946; date of declassification, July 18, 1947). | 27\, Oftedal, Z. psysik. Chem. 5B, 272-291 (1929). 28\/ S. Coleman, W. C. Whitley, and C. J. Barton, Mcferlqls Chemistry Division, L. M. Bratcher and C, J. Barton, Materials Chemlsfry Division. " L] y ™ L3 e o :The relcmve thermcl & 32y, S ‘Coleman and W. €. Whitley, ANP Quar. Prog. PERIOD ENDING DECEMBER 10, 1954 TABLE 5.7. PHASES FOUND IN MF-LoF; BINARY MIXTURES Hexagonal L m‘c\;ice .Scmple Molc‘r. Structure Type Dimensions {A) o Number Composition Observed Repor‘l‘ed3] L MB-108 INaF - 1L aF, BNl aF a = 617 a=6167 % ' o : c = 3.82 c = 3.819 -109 3KF-1LaF, By-KLaF 4 (and KF) a = 6.52 a = 6.524 % ' c = 3,79 c = 3.791 110 3KF~ILaF, B1-KLaF 4 (and KF) 4 = 6.52 a= 6524 % ‘ B ' S e = 3.79 c = 3.791 -107 IKF-1LaFg B1+KLaF 4 (and unknown a = 6,52 a = 6.524 % ' ' phase) c = 3.79 c = 3.7%91 131 & .‘.]kF-—TLOF:; One phase, p.redomina'nfly {same as unknown cbove) *Zcxchc:ficxsen;3‘l the quantity given is the number of molecules per unit cell. lattice rco'hs'tdri"rs";”t(-'l"abie 5.7). The expected complex phase is found in the NaF- LaF, system, The x-ray analysis of the KF-LaF, fuswn (Table 5.7, sample LMB-107) indicates the presence of B1 -KLaF, and an unknown second phase. A second fusmn at this composition (sample LMB- 131), with an lmposed slower rate of cooling, gives predominantly a single phase which corresponds to the unknown structure found in the previous preparation. For the 3KF- -LaF , composition, only the 5,-KLaF, complex has been observed. Petro-_ graphic exummahons?’ indicate that the new phase has optical’ prOperhes snm:lar to fhe B -phase. rmal stabi "tles of the»'e,_’rWO phases “ted fo be E:ubic,3'_ remom Rep. Sept. 10, 1952, ORNL-1375, p 79 V. S. Coleman, W, C. W |tiey, and C. J. Barton, ANP Quar Prog Rep. Dec. 10, 1952, ORNL-1439, p ]09 'rly.ie]ded the data shown in Table 5.8. " “tration of Na; possubly, Na:K = fo"300°K At 298.15°K; c The in- terpretation was aided greatly by comparing these diffraction patterns with the corresponding LaF, salts, The Greek letter assignment is adopted to correspond with the lanthanum compounds as given by Zachariasen. " The ternary salt mixture 1 KF-1 NaF-2 UF, gives the phase 3, “(K,Na)UF,. Here the ordering requires a doubling of the ¢ axis observed in the B,-phase of the binary salt. The extent of the . By~KLaF present in this melt indicates that the former phase is stabilized by the greater concen- = 2:1 in the lomce. Physncal Chem'sfl'y E . R. :Van Arfsdoie” o HO)/T were fobulqted at‘ 10- deg mterva!s up = 22.83 cal/mole- °K, Defcl[s of this work will be published in separate reports and articles by the Chemistry Division. 73 R g L i i‘ b s i Lo ANP QUARTERLY PROGRESS REPORT TABLE 5.8. PHASES FOUND IN MF-UF; BINARY MIXTURES Molar Composition? Ph Structure T Lattice zb olar Composition ases tructure lype Dimensions (A) 72NaF ]UF3 HexqgonalAphase and fiz-NaUF4 = 6.17 3/2 3NaF—2UF3 also NaF = 3.78 ]KF ]UF3 Hexagonal phase and B]-KUF4 a = 6,51 3/2 : : cubic phase® c = 3,76 Fluorite type ag = 5.9 IKF-INaF-2UF, Two hexagonal phases By “(K,Na)UF,, a=621 3 c =771 B-KUF, a = 6.51 ¥ c = 3.76 a.Sampfes prepc':lred by V. S. Coleman and W. C. Whitley and examined petrographically by T. N. McVay. The number of molecules per unit cell. Thls phose may be the cubic phase of KUF4 or the compound UOF with the fluorite structure. and S° = 29,90 e.u., of which 0.53 e.u. was ob- tained by extrapolation below 20°K. The heat capacity of a sample of pure MoS, was measured from about 15°K to room temperature, The heat capacity and the derived thermodynamic functions mentioned above were tabulated at 10-deg intervals up to 300°K, At 298.15°K, C, = 15.62 cal/mole *°K, and §° = 15.34 e.u. The heat ca- pacity of MoS, follows a T2 dependence between 15 and about 70°K of the sort previously reported>4 for MoO,. This behavior of MoS, is consistent with its lcyer structure, which is so pronounced that MoS, actually is a good high-temperature fubricant (Liqui-Moly). Density and electrical conductance measurements were made of molten mixtures of KCl and Kl across the entire composition face. The molar volume, as calculated from the density, is nearly additive for this system, although it shows deviations close to the Kl side. In contrast, the equivalent con- "'_”ductance shows pronounced negative deviations - from addlflwty for all mixtures, with maximum ,—_’:devmhon at high KCl concentration, These data “‘again show that it is inadvisable to interpret ‘maximums and minimums in electrical conductance ~as the consequence of formation of compounds in 34g. R. Van Artsdalen, ANP Quar. Prog. Rep. Sept. . 10, 1954, ORNL-1771, p 72. 74 the melts. The specific conductivity of molten Kl is expressed by the equation k = =1.7100 + 6.408 x 10~% - 2.965 x 10~%:2 mho/cm between 725 and 925°C, and the density is gwen by the equation o= 3.0985 - 0.9557 x 10~3 g/cm® between 680 and 910°C. The temperature, ¢, is in °C. The density and electrical conductance were determined for pure molten NaBr and are given below for the range 750 to 960°C: k = —0.4392 + 5,632 x 10=3; ‘ - 1.572 x 10~%:2 mho/cm and p = 2.9518 — 0.8169 x 10~3; g/cm3. The self-diffusion coefficient of sodium ion in molten NaNO, has been determined by a radio- chemical tracer technique. The heat of activation for self-diffusion is approximately 6 kcal/mole and, therefore, higher than the heat of activation for electrical conductance. Similar work with thallous chloride®5 has shown that the heat of activation 35E. Berne and A. ‘K.lemm, Z. Naturforsch, 8a, 400 (1953). L2 . LR ™0 -4 o wn, 3 L A ) o Xt for self-diffusion of the thallous ion is greater than that for electrical conduction. These results are in accord with the concept that the application of an electrical field lowers the potential barrier restricting migration of charged ions in the ap- propriate direction of the applied field. The self- diffusion coefficient of a given ion can be used in conjunction with the Nernst-Einstein equation®® to calculate the contribution of that particular ion to the total equivalent conductance of the sait. When this is done, it is observed for sodium nitrate and thallous chloride that, in each case, the cation accounts for about 95% of the equivalent con- ductance of the respective molten salt at temper- atures about 25°C above the melting point, This indicates that the fransport numbers of these cations relative to thelr anions in these two salts are opproximately 0 95. Vlscosfly Measuremenfs B F. A, Knox F Ker'resz N Y. Smith Mcn‘erlals Chem:s’rry Dlws:on The automatic coplllary wscometer, which was used previously®? for determining the viscosity of 36/\ = Dz F2/RT where A is the equivalent cone ducfance D the self-diffusion coefficient, z the charge of ions, 'F the Faraday constant, R the gas constant, and T the absolute temperature. - PERIOD ENDING DECEMBER 10, 1954 fluoride melts, has been modified and placed in operation for viscosity determinations c¢f alkali nitrate melts. Effort has been concentrated on determining viscosity changes immediately above the melting points as an adjunct to the work being done by Van Artsdalen to establish the signifi- cance of ion size on physical properties (see section above). In view of the low melting points of the salt mixtures being studied, a glass instru- ment is used, but it is planned to use an all-metal apparatus when the investigation is extended to chloride and fluoride melts. The salt mixtures studied included pure sodium and potassium nitrates as well as mixtures of the two in 12,5 wt % steps. The values obtained for the pure salts were about 0.3 centipoise above the data reported by Dantuma,3® who determined the viscosity by the logarithmic decrement method. The viscosities of the mixtures nearly overlap at higher temperatures, being located between the viscosity curves of the pure components. A plot of recnprocal temperatures vs the fog of viscosity gave an apparent deviation from linearity near the melting point of most of the mixtures, while the pure salts showed a direct linear correlation. 376, A. Knox, N. V. Smith, and F. Kertesz, ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL~1375, p 145, 38F2. S. Dantuma, Z. anorg. u. allgem. Chem. 175, 33-4 (1928). Yo umn ~ N ;‘ wi f Q 75 b, il ce 1 B bt N T gk e s ey . . N s b i o RN - ANP QUARTERLY PROGRESS REPORT 6. CORROSION RESEARCH W. D. Manly G. M. Adamson Metallurgy Division W. R. Grimes F. Kertesz Materials Chemistry Division ‘Studies of the corrosion of Inconel, type 316 -‘sfamless steel, Hastelloy B, and several special qlloys were continued through the use of thermal- ~convection apparatus. Additional information on ‘the compatibility of the sodium-beryllium-Inconel ‘system was obtained from static, whirligig, and 'fhérmql-c;oh‘\)ecfi_on loop tests. In an effort to find a brazing alloy that has good resistance to sodium and to NaF-ZrF - UF4, tests of several brazes on type 304 stmnless steel T-joints were made. Simi- far joints of A-nickel brazed with various alloys were also tested in NaF- -ZrF ,-UF, at 1500°F and in sodium hydroxide at 1100 and 1500°F. Addi- tional screening tests were made of carbides of boron, titanium, zirconium, and chromium in static s'o‘dium, lithium, and NQF-ZrF4-UF4 at 1500°F for 100-hr periods. An 80% Mg—20% Li alloy being considered as a crew-compartment shielding material was corrosion tested in water at various temper- atures; the weight losses were found to be negli- gible. The resistance of Inconel to attack by molten rubidium was studied in a dynamic loop apparatus. Fundamental studies are under way of the mass transfer in liquid lead and of fused hy- droxides as acid-base analog systems, and further mformahon has been obtained of the flammability of sodium-bismuth alloys. In further chemical studies of corrosion, information was obtained on the effect of temperature and of chromium additions on the corrosion of Inconel by fluoride melts with NiF, additions. THERMAL-CONVECTION LOGP CORROSION STUDIES G. M. Adamson A. Taboada Metallurgy Division Inconel Loop Containing UF, in - Alkali-Metal-Base Mixtures - Additional tests were made with alkali-metal-base fluoride mixtures containing UF; and UF , in Inconel o __"'thermal-convecnon loops, but the results obfomed - were not so favorable as those reported previously.! Cons:derable difficulty was encountered in making 76 the mixtures, in controlling the total uranium con- tent, and in determining the ratio of UF; to UF,. Two supposedly duplicate Inconel loops were operated for 500 hr with the UF;-UF, mixture in NaF-KF-LiF (11.5-42-46.5 mole %) at a hot-leg temperature of 1500°F (AT of ~200°F); however, the rough surface of one loop showed maximum penetration to a depth of 2 mils, whereas the other loop was attacked to a depth of 13 mils. In three other supposedly duplicate loops operated for 1000 hr, attack to depths of 8, 13, and 13.5 mils de- veloped. Metallic-appearing layers were found in the cold legs of all these loops. Several loops filled with NaF-KF-LiF with and without UF, were operated to obtain comparative data for use in evaluating the results of these tests, The data obtained are presented in Table 6.1. Type 316 Stainless Steel Loops Containing UF, in Alkali-Metal-Base Mixtures Several batches of the eutectic NaF-KF-LiF (11.5- 42-46.5 mole %) containing mixtures of UF, and UF were circulated in type 316 stainless steel loops. However, the results obtained cannot be used for comparative studies because the UF ,-to-UF ratios in the fluoride mixtures were unknown and the mix- tures were of doubtful purity. One loop plugged after 886 hr of operation at a hot-leg temperature of 1500°F; examination showed heavy, intergranular, subsurface voids to a depth of 11.5 mils in the hot leg and some intergranular attack in the cold leg. Another similar loop operated its scheduled 1000 hr, but a large deposit of dendritic crystals was found in the cold leg. A similar loop in which all the uranium (12 wt %) was present as UF, was operated to obtain com- parative data. This loop plugged after 570 hr of operation, and metallic crystals were found in the cold leg. The hot-leg surface was very rough, and there was penetration to a depth of 5 mils. The 'G. M. Adamson, ANP Quar. Prog, Rep. Sept. 10, 1954, ORNL-1771, p 96. " PERIOD ENDING DECEMBER 10, 1954 TABLE 6.1. RESULTS OF INCONEL THERMAL-CONVECTION LOOP TESTS OF NaF-KF-LiF (1. 5-42-46.5 mole %) WITH AND WITHOUT UF, CIRCULATED AT A HOT-LEG TEMPERATURE OF 1500°F ' Hoste”oy B thermal C >0 ‘& o _;F -UF, (50- 46-4- mole %), Period of Uranium Maximum Metallographic Notes Test Content Attack : (hr) (wt %) {mils) Hot-Leg Appearance Cold-Leg Appearance 500 11.9 1§ Heovy,"infergronulor, subsurface voids Metallic deposit 500 12* 13,5 Hequ, general, intergrufiulur, subsurface voids Metallic deposit 1000 12* 20 Heavy, general, intergranular, subsurface voids Metallic deposit 500 o 7 Mederate, general, intergranular, subsurface No deposit voids Heavy, general, intergranulor, subsurface voids Thin deposit in one section 500 o 75 ‘*Estimated; analyses not yet available. same fluoride mixfure without uranium was circu- lated for 1000 hr without signs of plugging. The hot leg of this laop showed heavy attack to a depth of 10 mils, and some grains were completely removed. Effect of Induced Potential Welding rods for use as electrodes were attached to the upper hot legs and lower cold legs of Inconel thermal-convection loops that were operated to defermlne the effect of a small induced emf on corrosion, Bah‘ery chargers were used to apply direct currents of 5. 4 amp at 0.45 v. The loops circulated NaF- ZrF UF (50-46-4 mole %) at a hot-leg temperature of ]500°F The direction of current flow was expected to have an effect on 'rhe attack, but no such variation could be observed in 'rhe four Ioops that have been vever, the depths of ch‘ack were less e _fhan the depfh of ctfcck found m a Ioop filled from“" the same bcn‘ch of fluorlde mleure cmd oper'd'red/ 'w:fhouf ctn mduced potenhol ’ exam med However , loy B Thermol-Convechon Loops fluoride mixtures, but there is some evidence that -convection loops have been these foops are p‘i'eseh'red in Table 6 2. "“The perlod'fi—,;_, | , " lated NaF- ZrF UF (50 464 mole 7) at a hot-leg - of operation of the |oop has not had a 'no’rlceable ‘effect on corrésion in the loops that circulated the the small attack that occurs takes place in a fairly short time and is a function of the condition of the original surface. The loops operated with sodium were known to be covered with an oxide deposit and were not cleaned before filling., Additional loops that were cleaned with the use of dry hydrogen are now being operated. Special Alloy Thermal-Convection Loops Several loops were fabricated from tubing of modified Inconel specially drawn at Superior Tube Co. from billets that were vacuum cast at ORNL, The loops were standard size, but the tubing was I/2 in. in diameter with a 0.065-in, wall rather than - the standard 3/ in.-IPS schedule-10 pipe; they were - filled with NCIF ZrF ,-UF, (50-46-4 mole %) and operated with a hof-leg tempercfure of 1500°F, . Some of these loops ruptured when attempts were ... made to restart them after the two power failures, .. The data obtained from these loops are given in ' nTc_xb!e 6. 3 I’r mc:y be noted 'rhat a considerable fir'educhon in depth of attack was found in both loops in which the chromium content of the alloy had been reduced to about 5%, ;llntergrcnulor, subsurface-vmd type of attack to a ~ depth of 6 mils in ]000 hr. ThlS loop also circu- “temperature of 1500°F. The attack was less than that usually found with Inconel. 77 “One’ loop of Hustelloy ’C d'eveloped a heavy," - ANP QUARTERLY PROGRESS REPORT TABLE 6.2. RESULTS OF OPERATION OF HASTELLOY B THERMAL CONVECTION LOOPS WITH VARIOUS FLUIDS ' “ . Hot-Leg Temfiérafure g ‘ L . Period O'F Circuloted Fluid Temperature Differential Test Metallographic Notes | | (°F) (°F) (hr) ' NdFr_-.Z'r'F“-'UF4 | 1500 325 1000 rRough surface; mfergrdm..ll'd} penef?c’ridné; subsurface 7 . voids to 3 mils 1500 325 2000 Shghfly rough surface; voids to a depth of 2 mils 1650 400 1000 Rough surface; most voids to depths of 2 mils or less, but occasionally to 3 mils 1650 400 1000 Maximum attack to o depth of 2 mils; slight deposit . in hot leg NaF-KF-LiF-UF4 1500 325 1000 Subsurface voids and intergranular penetration to a ' : . depth of 2 mils; thin deposit in hot leg; intergranu- lar attack in cold leg to a depfh of ].5 mils Sedium 1500 400 1000 Fine metallic crysfuls adhered to cold Ieg when loop drained while at 1500°F 1500 400 1010 Plugged by mass of fine dendritic crystals in hot leg that probably formed in cold leg TABLE 6.3, RESULTS OF OPERATION OF SPECIAL convection loops with beryllium inserts in the top ALLOY THERMAL-CONVECTION LOOPS WITH NaF-Z:F 4~UF 4 (50-46-4 mole %) WITH A HOT-LEG TEMPERATURE OF 1500°F Alloy Composition {wt %) Operating Maximum Time Attack Ni Fe Cr . Mo (hr) (mils) 76 7 17 1000 13 76 14 10 823* 8 76 14 10 1000* 12 76 19 5 647 3 76 19 5 460* 3 83 7 10 1000 13 83 7 10 1000 15 74 10 6 10 1000 3 74 10 6 500 3.5 10 *Leak developed in loop. SODIUM BERYLLIUM-INCONEL " COMPATIBILITY E. E. Hoffman C. F. Leitten ‘A, Taboada Meto”urgy Division As one phase of the sodium- berylllum-lnconel ‘compatibility studies, a series of Inconel thermal- of the hot legs were operated. The inserts were hollow cylinders about 6 in. long with inside diameters the same as those of the Inconel tubing of the loop. The outside of the beryllium was pro- tected by an Inconel sleeve and was separated from it by a 0.050-in. annulus filled with slow- moving sodium. The loops were operated for 500 he with high-purity sodium at hot-leg temperatures of 900, 1100, and 1300°F. In none of these loops was any attack found on the inside surface of the beryllium, and the outside surfaces of the inserts in the loops operated at 900 and 1100°F were similarly unattacked. However, in the loop oper- ated at 1300°F, the outside surface of the beryllium insert showed widely scattered voids to a depth of 3 mils, All the beryllium specimens were darkened, but no deposits were found by metallographic examination. _ , A summary of compatibility tests conducted by two different methods is presented in Table 6.4. In none of the tests was a layer deposited in the cooler portions of the test system that could be detected metallographically. in all the tests, how- ever, beryllium was found to be present on various sections of the Inconel tubes. Figure 6.1 shows the distribution of beryllium around an Inconel whirligig loop following a 270-hr test at a hot-zone temperature of 1200°F and a cold-zone temperature of 1060°F, Since the velocity of the sodium bath -, & O .4 ¥ 2 d e L in this apparatus was 10 fps, the bath temperature was practically isothermal, The temperatures given are the outside tube wall temperatures. The be- ryllium distribution shown appears to be temperature ~ dependent; however, in other dynamic tests, this was not so obvious. The outside and inside sur- faces of a beryllium insert from the hot leg of an Inconel thermal-convection loop are shown in Fig. 6.2. As in nearly all the other compatibility tests, PERIOD ENDING DECEMBER 10, 1954 the heaviest attack was on the outside surface of the insert. The attack was due to dissimilar metal mass transfer between the beryllium and the Inconel sleeve which surrounded it. The result of direct contact between Inconel and beryllium in a system ‘containing sodium at elevated temperatures may be seen in Fig, 6.3, The Be-Ni intermetaliic layer which formed was approximately 20 mils thick and was extremely hard and brittle, Tests are presently TABLE 6.4, RESULTS OF B.E'RYLLIUM-SODIUM-INCONEL COMPATIBILITY TESTS Coneentration of Test o Test T?s Temperature (7F) Be on Surface Metall hic N me Type (hr) Hot Zone Cold Zone of Inconel Tube etallographic Notes . (pg/cmz) - Whirligig 3.00 1100 !.sothermal- 5.3 to 7.9 300 1200 Isothermal 0.05 300 1300 lsothermal 0,07 300 1400 Isothermal 39.6 to 127 270 1200 1060 - Convec on 1300 1130 16,5 to 39.6* Two dark deposits on Inconel which analyzed high in beryllium; layer on Inconel sleeve; beryllium insert ‘showed no attack on inside and 1 mil of attack on outside surface No deposits on Inconel tube; 0.4-mil layer on Inconel sleeve, beryfl:um insert had no attack on inside surfcce and 1 m:l of attack (volc{s) on outside ‘surface Inconel sleeve which enclosed beryllium insert had 1-mil layer on surface; beryllium insert had no attack on inside surface and subsurface voids to a depth of 5 mils on outside surface Deposit of 2 to 3 mils on Inconel sleeve; beryllium insert attacked to a depth of 1 mil on inner surface and to a depth of 3 mils on outer surface Deposfi' of 0. 2 mll on Inconel s!eeve, bery|||um ine serf o'rfacked fo a depth of 2 ml!s on |nS|de surfcce *See Fige 61, TS T e el R 79 e T OGO T Ty g e - - ANP QUARTERLY PROGRESS REPORT UNCLASSIFIED ORNL-LR~-DWG 4650 BERYLLIUM INSERT: LENGTH — 3.75in. ID—0.430in. Flg. 6 '|. Dlstnbuhon of Berylllum Around an ~ Inconel Whlrhg|g Loop Containing a Beryllium “Insert and’ Sodlum. Tesf period, 270 hr. UNCL ASSIFIED 3497 to Sodlum for 1000 hr at 1300°F. (a) Surface of i 'berylllum exposed to sodium in annular space. - (b) Surface of beryllium exposed to flowing sodium. ~ Unetched. 500X. Reduced 37.5%. under way to determine the minimum spacing that can be tolerated between beryllium and Inconel. STATIC CORROSION STUDIES E. E. Hoffman . W. H. Cook C. F. Len‘fen Metallurgy Division Brazing Alloys on Stainless Steel Additional tests have been made in an effort to find a brazing alloy that has good resistance to both sodium and the fuel mixture NaF-ZrF -UF, (53.5-40-6.5 mole %). The results reported in Table 6.5 were obtained from tests of type 304 stainless steel T-joints brazed with the alloys listed by the Wall Colmonoy Corporation. Tests are ‘also being conducted on Inconel T-joints brazed with each of the alloys listed in Table 6.5. All the brazing alloys listed had good resistance to the fluoride mixture, but only alloy C-29 had good resistance to both the fluoride mixture and to sodium. Brazing Alloys on Nickel A series of T-joints of A-nickel were brazed with the various alloys listed in Tables 6.6 and 6.7 in a dry hydrogen atmosphere and were then exposed for 100 hr to static NaF-ZrF -UF, (53.5-40-6.5 mole %) at 1500°F or to static sodium hydroxide at 1100 and 1500°F,. The specimens were examined me- tallographically in the as-received condition in order to evaluate the inclusions and the porosity. The brazed specimens were thoroughly cleaned before and after testing so that valid weight change data could be obtained. In order to prevent the rounding of the fillet edge upon polishing, each specimen was nickel plated after testing. The data obtained from metallographic examination of the specimens exposed to the fluoride mixture are presented in Table 6.6, and the data on the speci- mens exposed to sodium hydroxide are in Table 6.7. Most of the alloys tested appeared to have good corrosion resistance to the fluoride mixture, except the braze alloy with the composition 69% Ni-20% Cr-11% Si. Several of the brazed T-joints, es- pecially those which included copper, gold, or silicon as an alloying element in the braze material, showed numerous voids in the interface between the base material and the braze fillet., These voids are not considered to be caused by the attack of the fluoride mixture but rather by diffusion of a e TR “A"\.:‘ e i LR f PERIOD ENDING DECEMBER 10, 1954 A x x ’ o o © o = o o O 1e lo - : e B kB B & g kB kB B B .-y | @ INCH |® ® < & > 5 & 5|2 _ = Fig. 6.3. Diffusion Layer of Be Nl Which Formed on Inconel in Dlrect Contacf wn‘h Berylllum Insert '. in a Thermal-Convection Loop After Exposure for 1000 hr to Sodium at 1300°F. Hardness impressions :F md:ca?egrelahv‘e‘hordnesses. e, Nig, 1300 DPH; Inconel, 180 DPH. Etched with oxalic acid. 500X. constituent from the braze alloy into the base Table 6.7 shows the results obtained from the material. A T-joint brazed with 60% Pd—37% Ni~ metallographic examination of the brazed A-nickel 3% Si and then exposed to the fluoride mixture for T-joints exposed to sodium hydroxide. Of all the 100 hr is s_h_own in Fig, 6.4, Many voids can be braze alloys tesfed only the 82% Au—lB% Ni-alloy “seen along ‘the interface between the base moferlal had good resusrqnce 'ro at’rack ot both 1100 and - and the braze f[llet but the surface of fhe specnmen ) l500°F The specrmen ’rested in sodium hydrox:de " dppears to be*freeof ofl'ock - _ L for l00 hr ot 1500°F is shown in Flg. 6.6, ' ,.’:ln’order to Ver_lf;/ the poss:bilrfy of cllffuswn"“ TR 'rhese rests, as in the fused fluoride tests, ‘base “voids were found along the interface between the base materml and the broze f:llef when the broze c:lloys ‘that contamed copper, gold or silicon as _an olloymg elemem‘ were ‘used, However, fhe same specrnfens 'res’red at ll00°F reveoled no such vo:ds. 7 Tl’ns si 'uahon s Pmbc'bly caused by the ‘manner as were the tos'led T-|omts"\un’c~lv placed in f_‘_on evacuofeclw copsule and whec:tecl “at 1500°F for A « 7 o and 'rl':ef_l; | _ ' " “Voids appears to be independent of environment bemg “Fested in sodlum hydroxlde For 100 hr at and therefore not a result of attack by the fluoride 1100 and 1500°F, respectively. As may be seen o mixture. | in Fig. 6.7, no attack occurred along the braze 81 IR AR RARPTTRUER CORT TN AP TEIRILT Cre T (U T ANP QUARTERLY PROGRESS REPORT { TABLE 6.5, RESULTS OF STATIC TESTS OF BRAZING ALLOYS ON TYPE 304 STAINLESS STEEL T-JOINTS IN NaF-Z¢F -UF , (53.5-40-6.5 mole %) AND IN SODIUM AT 1500°F FOR 100 hr Brazing Alloy Weight o Weight Composition Bath Change* Change Metallographic Notes (wt %) {9} (%) Alloy A-16 NaF-ZrF -UF, —0.0003 =—0.042 Braze fillet unattacked 23 P, 77 Ni _ Sodium -0.0009 ~0.135 Subsurface voids in braze fillet to a depth Lo i . of 5 mils; attack confined to Ni3P phase :i Alloy C-27 NGF-ZI‘F4-UF4 o 0 No attack A 9.6 P, 2,75 Cr, 88.6 Ni Sodium —~0.0006 —0.086 Subsurface voids to a depth of 5 mils; S _ . o Ni;P phase attacked 4 Alloy C-29 NcF-ZrF4-UF4 +0.0011 +0.180 No attack ' ; 10.2 P, 13 Cr, 76.8 Ni Sedium +0.0004 +0.054 Less than 0.5 mil of small subsurface : ' - ' voids ' Alloy B-11 NaF-ZrF4-UF4 ~0.0006 -0.082 Braze fillet attacked to a depth of 1 mil in several areas 10.8 P, 9.2 Si, 80 Ni : o Sodium 0 0 Subsurface voids in fillet to a depth of 4 mils Small subsurface voids in braze joint to a depth of 0.5 mil Subsurface voids to a depth of 7 mils; Alloy J-10 9P, 15 Fe, 4.5 Cr, 71.5 Ni Nc«F‘-ZrFA-UF4 +0.0007 +0.153 Sodium —0.0004 -0.067 Ni3P removed from fillet zones Maximum attack was 0.5 mil in the form of Alloy H-10 10 P, 4.3 Mo, 86.7 Ni Na F-ZrF4-UF-'4 +0.0036 +0.730 small subsurface voids Sodium —0.0006 -0.108 Subsurface voids in braze fillet to a depth _ of 19 mils Alloy 1-10 Nc:F-Zer-UF:4 +0.0008 +0.112 No attack 11.6 P, 6.25 Mn, 82.15 Ni Sodium —0.0018 —0.346 Subsurface voids in braze fillet to a depth | ' of 11 mils f Alloy D-11 NaF-ZrF,-UF, +0.0015 +0.304 No attack 9.9 P, 11.3 Fe, 78.8 Ni Sodium +0.0006 +0.108 Subsurface voids in braze fillet to a depth | of 25 mils *Weight change data include brazing alloy and base material of joint. fillet of the specimen tested at 1100°F and very few diffusion voids were present. However, as may be seen in Fig. 6.8, the specimen tested at 1500°F showed a é-mil surface ottack and also several small interfacial voids. Additional static corrosion tests are being con- ducted on many of these same brazing alloys in sodium in order to evaluate their suitability for "_‘usv'e as aback braze for the sodium-to-fused fluoride heat exchangers. gual Screening Tests of Carbides Specimens of carbides of boron, titanium, zir- conium, and chromium were tested in static sodium, lithium, and NaF-ZrF ,-UF, (53.5-40-6.5 mole %) at 1500°F for 100.hr periods. The measured densi- ties .and apparent porosities of the specimens and the results of the corrosion tests are given in Table 6.8. The data indicate that TiC and Cr,C, had the best resistance to corrosion in each of the mediums; Figs. 6.9 and 6.10 show the typical . ) Sama? - E } 2 X a0 A ARG L M SRS e ARG, S PERIOD ENDING DECEMBER 10, 1954 , TABLE 6.6. RESULTS OF STATIC TESTS OF BRAZED A-NICKEL T-JOINTS IN . NaF-Z¢F,-UF, (53.5-40-6.5 mole %) AT 1500°F FOR 100 hr s, Brdiing Aiioy 7 _ ‘Weight | Wévighfr - Composition Change Change Metallographic Notes (wr %) (@) " 2 _ . 69 Ni—20 Cr-11 Si ~-0.0017 ~0.055 Surface attack to a depth of 16 mils along entire braze fitlet 100 Cu ~0.0006 -0.019 Surface attack to a depth of 0.5 mil along entire braze fillet ’ 82 Au—18 Ni . Z0.001 —0.036 Braze fillet appeared to be unattacked 80 Au—20 Cu —0.0007 0,026 No attack on braze fillet surface T, 90 Ni-10 P ~0.0004 w --0.013 " No attack on braze fillet surface 60 Pd—40 Ni ;0._0016 ' -.-0.06 . No attack on braze fillet surfu;:e 60 Pd—-37 Nil——-3 Si +0.0008 | +0.0'2‘7. " No attack on braze fillet surface TABLE 6.7. RESULTS OF STATIC TESTS OF BRAZED A-NICKEL T-JOINTS IN SODIUM HYD_ROXIDE FOR 100 hr AT 1100 AND 1500°F v _ ' VBrcrxzihg Alloy - Weight Weight Test = _ Composition ‘Change Change Temperature Metallographic Netes . (wt %) () (%) (°F) | 69 Ni—20'Cr—1l_ Si - o 1500 Braze failed completely o +0.0024 +0.093 1100 Braze attacked completely * ’ ' A 100 Cu ~0.0118 -0.408 1500 Braze attacked completely; large voids appear s throughout -0.0011 -0.038 1100 U_ni'form surface attack on braze to a depth of 3 mils 82 Au-18 Ni _0. 004 7 1500 'Nonuniform surffite .a'fl'dck on braze to depth of 1 mil ' ‘--0 0007 1100 No afiack on surfoc'e of braze 80 Au-20 Cu_-‘ V —0-0]06 ’ Umf?rm surfccelfl aflack on enhre brqze to a 2 depth of 3 . . e e mils : _ 7 - ' “attack . ze to a depfh of T mlf : - 60/Pd-40N| i y\ cler.m wuth cn‘tcck in fhe form of 5 oo smoli{sfrlngers runmng to a depth of 4 mils Surface-aflqck toa depth of 0. 5 m|| A ; 'in brittle . : N|3P phase ' ‘ . B'raze atfacked complefely, attack centered in |\l|3 s phase , e , - 60 F’d-—37N|—3S| +0 0023+0.083 o 1500 L Surface aftack on braze fillet fo a depth of 6 mi ||s e R 028 1100 No cftack ";ésen'r ‘On braze surfoce o =, oo U0 Sl f v ey ' 83 ER Ll . < i i ANP QUARTERLY PROGRESS REPORT appearance of the as-received and tested TiC and Cr,C,. These corrosion data are not selective enough to designate whether TiC or Cr,C, is the superior material, If the lower strength and/or more brittle nature of the Cr,C, are objectionable, 0.03 INCH 0.0 ~ Fig, 6.4. A-Nickel T-Joint Brazed with 60% Pd-37% Nl—3% $i and Exposed to NuF-ZrF4-UF (53.5-40-6.5 mole %) at 1500°F for 100 hr. Nofe voids at the interface between the braze material and the base metal. Unetched. 100X. Reduced 38%. UNCLASSI FIED ¥-135 Fig. 6.5. A-Nickel T-Joint Brazed with 60% - Pd-37% Ni-3% Si and A#néaled for 100 hr at -1500°F in an Evacuated Capsule, Note voids at interface between the braze material and the base metal. Unetched. 100X. Reduced 40%. 84 the TiC is superior. The ZrC showed signs of failing only in the fluorlde mrxture, the fine in- cipient cracks in the surfaces of the ZrC particles may be seen in Fig. 6.11. There were indications that the fluoride mixture had penetrated the pore spaces throughout the ZtC; the pore spaces were not en!arged Fig. 6.6. A-Nickel T-Joint Brazed with 82% Au-18% Ni and Exposed to Sodium Hydroxide for 100 hr at 1500°F. Note slight attack at surface of braze. Etched with potassium cyanide. 200X Reduced 39%. Sy NCLASSIFIED: Y1338t - | NICKEL PLATE Fig. 6.7. A-Nickel T-Joint Brazed with 60% Pd-37% Ni-3% Si and Exposed to Sodium Hy- droxide for 100 hr at 1100°F. Note absence of voids at interface. 100X. Reduced 39%. i 3 UNCLASS!F%ED ¥-13348 .04 Nj‘gx!—::i" BLATE 02 INCH Fig. 6.8. ANickel T-Joint Brazed with 60% Pd- -37% Ni-3% Si and Exposed to Sodium Hy- droxide for 100 br at 1500°F. Note surface attack and small voids at inferface between base material qndubrdz'\_e alloy. ‘Unetched. ]00)(._ Réflq_éea 39%. PERIOD ENDING DECEMBER 10, 1954 The reasons for failure of the B,C to withstand the corrosive actions of the sodium and lithium are not known-at this time; powder x-ray analyses of the tested specimens of B,C indicate that B,C is the primary phase and that a new, secondary, unidentified phase is present as a result of both the sodium and lithium corrosion tests, The attack on the B,C by the fluoride mixture is not definite or uniform. The 5 mils of attack found on one side of the specimen was the maximum depth of attack if the 14-mil-thick unidentified phase found on the other side can be attributed to fluoride mixture in a surface imperfection of the test specimen. Fure ther tests will be necessary to properly evaluate B 4C corrosion resistance to the fluoride mixture. These screening, single-run, static corrosion tests of B,C, TiC, ZrC, and Cr,C, indicate that these corbldes, except B,C in lithium or sodium, warrant further and more severe tests in the in- vestigation of materials to withstand the long-term, i e AU Wi Fri:g. 6.9. As:Received TiC, : 500X. (a) (») TiC Afler Exposure to Sochum at 1500°F for 100 hr. Unetched. 85 - ANP QUARTERLY PROGRESS REPORT "TABLE 6.8. SUMMARY OF STATIC TESTS OF CARBIDES AT 1500° F FOR 100 hr A Average Weidht t Material Density ppar?r.x Dimensional e's . Attack : ' 3 Porosity . Change . Remarks Tested (g/cm”) Change {mils) . (%) (%) (%) Tested in Sodium B4C 2.51 3.5 ~6.5 Specimen cracked and fell apart during test TiC ' 4.81 - 1.3 +0.3 0 0 No attack; particles bonded so well that it : s was difficult to distinguish individual _ particles in unetched specimen ZrC 6.9 o 0.8 0 ‘ +0Q.1 0 No attack; particles well bondeg'jl CryC, 6.59 2.2 . 0 ~0.4 0 No attack; particles not well bonded Tested in NuF-ZrF4-U F4 (53.5-40-6.5 mole %) | B4C ' +0.1 ~0.4 5 Unidentified phase on one side of - o ‘specimen that had a maximum thickness of 14 mils ' TiC 0 +0.1 0 No attack; as-recelved and tesfed 7 specimens not bonded nor formed so “well o _ as those tested in sodlum '_ ' - ZrC ' _ o +0.3 Ma;or pornon of pore spaces fnlled wnfh ‘ o fluoride mixture ' Cf3C2 _ | +0.2 No attack Tested in Lithium B4C | -~12,2 Specimen expanded and cracked; original, . ' C ' dark, semimetallic luster dq”ed TiC 0 0 0 - | No attack; mccroscopi.c color slightly lighter Z:C +0.2 -1 0 No attack Cr3C2 +0.2 -1.2 0 No attack; small erack in middle of _specimen; metallic luster disappeared *The average dimensional change (%) is the average of the width, height, and length chunges; whereas the weight change (%) represents a single value. hi gh-temperature corrosion of sodium, hthlum, and NaF-ZrF UF (53 5 40- 6 5 mole %), Magnesmm-L:fluum Alloy in Wo‘l‘er w-gvc‘b"m\'&’mfimm&_«@ An_ 80% Mg-zo% Li auozjwh]‘fi has been proz, g as a possnble crew~compartment shleldmg g e S, ST A oy ez et hflwfl‘fi)‘" ol 86 material was corrosion tested in water at various temperatures, Specimens of this material were carefully cleaned just prior to testing to remove surface films, chiefly Li N, which form on exposure to air. A cleaned specimen was found to gain 0.088 mg/ecm? after standing in the atmosphere for 6 hr. The weight losses of specimens tested at W, ar 12 o ~0ne ‘spec1men was_tested for various l_gngths of Y-12908 PERIOD ENDING DECEMBER 10, 1954 NICKEL PLATE .02 Fig. 6.10. (a) As-Received Cr3C2. (%) C|'3C2 After Exposure to NuF--ZrFA-UF'4 (53.5-40-6.5 mole %) at 1500°F for 100 hr. Unetched. 100X. various temperatures in water for 15 min are tabu- lated below: 7 Témperature Weight Loés (°cy (mg/cm ) o e S50 027 e .80 B L 100 | 0.326 hme in water ‘at ‘100°C and was Found to lose 0. 292 mg/cm in the flrst 2 min of the fes’rs, whlch was 90% of 'fhe‘\'Nél‘gh'r lost in “the ]5-m|n fest. ~ This decrec;.g in_rate of Welght loss may be at- ' :‘frnbufcb[e to 'rhe deple’r:on of hthlum atoms on the . surfuce, since no protective fllm could be detecfed Pure magnesium tesied in water at 100°C for 15 min showed a weight loss of 0.09 mg/cm?. RUBIDIUM INVESTIGATIONS The results of preliminary tests of the resistance of Inconel to attack by molten rubidium were dis- cussed in the previous report.? Static tests at - 1500 and 1650°F have shown a maximum attack of from 1 to 2 mils in ]00 hr. The attack has been " both m’rergranulor and in the form of subsurface 'vmds. Since these tests indicated that under " static lso’rhermal condltlons corrosion of Inconel "‘by rubidium was no serious problem, a dynamic test was conducted, The resistance of Inconel to corrosion by flowing liquid and vaporized ru- bidium in a closed system incorpordting a tempera- ture gradient was determined by means of a dynamic loop apparatys, somewhat similar in appearance to the conventional thermal-convection loops., A 2E. E. Hoffman et al., ANP Quar, Prog. Rep. Sept. 10, 1954, ORNL-1771, p 86. & TTF ey 87 * ANP QUARTERLY PROGRESS REPORT NICKEL PLATE 0.001 0.002 0.003 0.004 0.00%5 INGH 0.006 0.007 - 0.008 p 4 .9‘5-‘,&‘ - ! RS | a1l ¢ Fig. 6.11. («) As-Received ZrC. (b) ZrC After Exposure to Nc:F-ZrF“-UF4 (53.5-40-6.5 mole %) at 1500°F for 100 hr. Unetched. 500X. loop sectioned after a test may be seen in Fig.6.12. The hot legs of the loop were constructed of 0.5- in,-OD, 49-mil-wall tubing, and the condenser section was of 0.25:in.-OD, 35-mil-wall tubing. The loop was loaded with 8 cm® (approximately 12 g) of vacuum-distilled rubidium, which filled it to the level indicated in Fig. 6.12, A vapor- phase heat transfer system of this type will elimi- nate mass transfer in the vapor regions. However, the temperature gradients present in the liquid region might conceivably cause crystal deposition, The maximum attack occurred in the hot leg of the loop to adepth of 1mil, as may be seen in Fig.6.13. The attack was intergranular, and one grain ap- peared to be ready to fall from the wall, A schematic diagram of the apparatus presently used to distill rubidium is shown in Fig. 6.14, | Thus far, 3 Ib of rubidium has been purified. Since the addition of 8% sodium will reduce the melting point of rubidium to approximately 20°F, tentative _ 88 plans call for an investigation of the corrosion properties of the sodiumerubidium eutectic. FUNDAMENTAL CORROSION RESEARCH G. P. Smith Metallurgy Division Mass Transfer in Liquid Lead J. V. Cathcart - Metallurgy Division The results of previous tests® indicated that alloys in which intermetallic compound formation was possible showed a marked increase in re- T sistance to mass transfer in liquid lead as com- = pared with the mass transfer obtained with their pute components. Comparable behavior ‘was not s T observed in alloys such as Nichrome V in which - Y 3J. Y. Cathcart, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 100. A RO . L% o} a3 T ORNL-LR-DWG 4651 BOILER AND VAPOR SEPARATOR SECTION ol N OPERATING DI, N LIQUID LEVEL O, > ' , S t60°F B AT—!360°F - [.||‘1"|!\if>|;|"l||'||||'E|E| OPERATING TIME = 312ty t 2 3 4 5 MAXIMUM CORROSION PENETRATION 0.004 in. NOTE: EQUIUBR[UM LOOF’ SURFACE TEMPERATURES " ARE GIVEN. Fig. 6.12. Sectioned Inconel Loop After Ex- posure to ‘Boiling or Vaponzed Rubidium for 100 hr, this hypofhes:s has been tested further, and, as in earller work e_xper:m m‘sr were"perform d i Il S”UCfiJl’e. mGy'eGstIy be seen in flie pho'rogrdph' 4), V. Catheart, ANP Quar. Prog. Rep. Dec. 10, 1952, ORNL- 1439, P 148. PERIOD ENDING DECEMBER 10, 1954 The first loops in which the 50% Fe-50% Cr specimen was tested in liquid lead failed after 46 hr of operation with hot- and cold-leg tempera- tures of 805 and 550°C, respectively. The second loop failed after. only 30 hr of operuhon with hote and cold-leg . temperatures of 805 and 500°C, re- spectively. Microscopic - examination = showed clearly that a phase transformation took place in the test specimens during the operation of the loops. As shown iin Fig. 6.16, a narrow layer of a second phase formed at the outer surface of the hot-leg specimens, Hardness measurements yielded values of 162 and 1140 (DPH hardness scale) for the new and original phases, respectively. These values correspond closely to those expected for ferrite and sigma phases. The possibility that the new phase was ferritic was supported by the fact that it was magnetic and by a chemical analy- sis of the plugs formed in the loops., Since the plugs were slightly richer in chromium thon in iron, a decrease in the chromium content of the outer layers of the test specimens was indicated. The test temperature was close to the sigma-to- ferrite transformation pomt ‘and therefore even a slight alteration in the original 50% Fe--50% Cr ‘composition would produce an alloy in which only ferrite is stable at the temperature in question. A second alloy, 50% Mo—50% Fe, was also tested. This alloy was chosen because an intermetallic compound predominates at the indicated compo- sition. Circulation of lead in this loop stopped after 520 hr with hot- and cold-leg temperatures o cohpounds are presenf -”Duri‘ng fhepasfquarfer ~ of 805 and 500°C, reSpeéfiver. Examination of this loop has not yet been completed; therefore, the results obtained will not be discussed other han to point out that the pluggmg fime was ap- ely that to be expected for an c”oy in /v/vh:ch an |htermefa[|:c compound existed, - : Specuol comment is requured f_orwthe resui’rs ob- ained with the 50% Fe—SO% ‘ndlg.cafed obove, ‘these specumens had bee-n 'rron-s-;f_' _7foi'm'ed' almosf complefely to sngma phase pr:or to 1esh"hg, “and” ye’r very shorf operatmg tlmes were fi'observed for the loops contammg fhem. Comparable ' loops ™ confcmmg pure iron and pure chromlum T required 275 ) ‘pluggmg occurred. On the other hand the pluggmg “times for other alloys thot form mter ':"'etcillc come X nd 100 hr, respechvely, before fiounds, for example, 45% Cr—55% Co, Hastelloy B (5% Fe-28% Mo—67% Ni), 25% Mo—75% Ni, etc., were all much greater than the plugging times for loops containing their pure constituents. 89 ANP QUARTERLY PROGRESS REPORT 0, a _j‘ e e agi + \ Fig. 6.13. Surface of Inconel Tube Below Liquid Level in Hot Leg of Loop Shown in Fig. 6.12. Note . intergranular attack and grain separated from wall of tube. Etched with aqua regia. 1000X. aormE PHILIPS ORNL-LR-DWG 4652 * VACUUM GAUGE t2-mm VAPOR LINE PIRANI VACUUM GAUGE ; . 3-WAY 10-mm 2 -WAY. 8-m % . VACUUM STOPCOCK VACUUM s'TOPcrgCK HEATING TAPE 100-ml ROUND-BOTTOM PYREX FLASK TO MERCURY : ANOMETER PURIFIED RUBIDIUM . ; : , VMF 20-amp 300-m| ROUND-BOTTOM 2-WAY, 6-mm VACUUM DIFFUSION PUMP PYREX FLASK . STOPCOCK TO HELIUM - : SUPPLY : S , l GLAS-COL HEATING MANTLE TO MECHANICAL PUMP & Fig. 6.14. Apparatus for Distilling Rubidium. 90 PERIOD ENDING DECEMBER 10, 1954 M AL NG _ UNGLASSIFIED '| 0.00 2 , ol g _ S TN : Y- 13215 0.002 0,003 0,004 0,005 0.006 0.007 0.008 0.009 €.010 .04 0.012 0.013 0.014 0.045 0.016 0.017 0.018 0.019 0.020 .024 0.022 g o INCH 5 J ' ' .~ Fig. 6.15. A Tréfisverse Section _of a 50% Cr=50% Fe Al.!oy After a Sigma-Phase Anneul. Note ex- tensive cracks. 200X. ' One " possible explanation for the anomalous one of the test specimens had been completely or behavior of the 50% Fe—50% Cr alloy is suggested partially dislodged by the erosive effect of the by the extensive system of cracks present in the ‘circulating lead, the metal particle couid have original fest spec mens (Fig. 6 15). The cracks been caught in elther the hot- or cold-leg specimen “ho dQubt lncreaéed the surface area o-vo!un:lgrcmo tubes ‘and caused a sufficient slowung down of the | ¢ Tead flow rate to produce prémature failure of the " loop. "Because of the excessive brlftlen ss of the " est specimens, it was not possnble to remove them " “from the loops without crackmg them slightly; there- foré, it is not pésSIb]e to sfcte unequwocably that this explanation is a valid one. However, the ex- " "“fensive cracking of the test specimens and their “extreme friability plus the unknown effect on the ‘phase trdnsformation which ‘took place would ap- _ “b'éc’r' to |us’rlfythe ‘belief that the r'é'su lts obtained ‘small'pl'eces cxlmosf a "’rhe touch with fifhe. 50% Fe-50% Cr alloy may not be com- Thls increase in fnublhty was probably associated parable.fo\\\fhose obtained with other alloys in which wn‘h the penetration of lead info the crack system intermetallic compound formation is possible. ‘referred to above. 1f a small chunk of me’rul from R PR e ;sv ‘ot llkely fl]at such “small’ pl_ugs could' have RN NooT Ao - ol Py 2o B= + H,0 where B~ is taken as a generalized representation of an anionic species. Third, (8) OH- + B-——> BH + O™~ The second type of solvolytic reaction requires that the hydroxyl ion be a stronger proton acceptor than the B~ ion. There are, of course, many examples of this, such as the obvious reaction (9) OH- + HCI—> CI= + H,0 which, in fused hydroxides, is a solvolytic re- action and not a neutralization. Equation 9 repre- sents an obvious reaction because of the ex- ceedingly weak base character of the chloride ion. There should, however, be a series of such reactions involving bases with proton affinities lying between those of CI~ and OH™, such as HP04"" and HCO3"', so that the reaction (10) HCO,~ + OH~—>CO,~~ + H,0 might occur; if it did occur, it would be an ex- ample of Eq. 2. The third type of solvolytic reaction, Eq. 8§, requires that the B~ ion be a stronger base than the O~~ ion. At best, there will be very few " instances of this. Proton n:-Neutru ||zcmon-Type Reactions. Neu- '_i'frdlizcn‘lon reactions in fused hydroxides which ~can be treated by the protonic theory are also of three types: (1) the reaction of water with the oxide ion according to the formula (11) H,0 + 0=~ —> 20H" where the water may come from the dissociation of another compound; (2) the reaction of water with an ion which is a stronger base than the hydroxyl ion but a weaker base than the oxide ion, that is, essentially the reverse of Eq. 7, (12) H,0 + B~—> BH + OH~ and (3) the reverse of Eq. 8§, (13) 0~ + HB—> B~ + OH- An obvious example of the reaction given by Eq. 13 is 7 (14) O~ + HCl— ClI= + OH“ Kmehco“y, this reaction undoubtedly would go by two steps, HCl + OH- — CI= + H,0 H,0 + O~ ~—> 20H~ if the oxide ion were present in low concen- trations, with the second step controlling the rate. However, the equilibrium would be represented by Eq. 14. Equation 13 represents a wider range of reactions than Eq. 8 because the permissible proton affinity range for B~ goes all the way up to that of the oxide ion; therefore the reaction (15 HCO,~ + 0=~ —> CO,~~ + OH- is undoubtedly an example in which the bicar- bonate ion behaves like an acid. Oxidic-System Concept. There are a great many reactions in fused hydroxides which cannot be usefully treated in terms of the protonic-solvent theory. Consider, for example, the familiar re- action ' (16) €O, + 20H~—> CO,~~ + H,0 The application of the Brdnsted theory here re- quires a rather complex, formal treatment. The actual proton exchange is between the bases OH™ and O~~. The weaker base, OH~, serves as the proton acceptor, and the stronger base, O~ ~, " serves as the proton donor. This is the reverse of the neutralization reaction given in Eq. 11 and would not be predicted by the protonic-solvent theory. This seeming contradiction may be re- 30 i . .J g 3 i, L] ¥ " 1] e k) ) B W solved in a formal way, but it is much more useful to apply the following straightforward approach. In Eq. 16 the greater oxide-ion affinity of carbon dioxide compared with water determines the di- rection of the reaction. Lux? and Flood and Férland!® have advanced a definition of acids and bases which was de- signed for tréc'fi_ng the special problems en- countered in nonprotonic fused oxide systems. According to these authors, a base is an oxide-ion donor and an acid is an oxide-ion acceptor; thus (]7) Bcse—\ Acid + O™~ Under this defmmon the convenhonol classifi- cation of slags as acidic or basic is preserved. This acid-base concept has not been applied to solvent systems, but, by a small extension, this may be done. Two competing equilibria of the above type can be considered: AC ————-“AH + O-- o~~ + B“—‘“_ABO (18) AO + B ¥ BO + ATt Base | Acid |l Base Il Acid | The acid-base éqbilibrium shown by Reaction 18 "may be considered to be controlled by a balance between the oxide-ion affinities of the reactant acid and the product acid in a manner that would be analogous to the balance between the proton affinities of the bases in protonic solvents, The self- dlssocu:mon of hydroxyl ions according to Eq. 2 completely fu[fi“s the requsrements of _the Lux deflmtion, _Eq 17. It may, Gf f"'S" seem_:__ SUIASENT o fwo hydroxyl |ons the H. Lux, Z Elektrochem 45, 303 (1939). 10 592, 781, 790 (1947). H. Flood and T. Forland, Acta Chem. Scand 1, PERIOD ENDING DECEMBER 10, 1954 Thus, water, H-O-H, may be regarded as a pyro- analog compound of hydrogen and the hydroxyl ion as the ortho-oxyanion of hydrogen. Because of acid-base relations involving pyroions, such as shown in Eq. 19, two hydroxy! ions will always be taken together to serve as a Lux base analog. Oxidic-Solvolysis-Type Reactions. The solvo- lytic reactions involving the oxides and oxysalt solutes can be classified according to the ioni- zation energy of the atom other than oxygen in the oxide. This method of classification is possible because the ionization energy of an atom is related to its oxide-ion affinity. It is con- venient to divide the elements into four groups according to ionization energy. First, oxides and oxysalts of atoms having high ionization energy should react with the fused hydroxides to give a change in coordination. For example, (200 SO, + 2CH™ &= SO,== + HOH (see also Eq. 16). Second, oxides and oxysalts of atoms of medium ionization energy should react with fused hy- droxides either to form or to depolymerize poly- anionic acids. For example, the reaction (21) P,0,4= + 20H-==2P02~ + HOH will probably take place. The possible reactions of this type are numerous because of the sub- stantial variety of polyanions, For example, there could be the step-wise depolymerization of silicon- oxygen compounds beginning with the three-di- mensional network in silicon dioxide, passing through the two-dimensional net polyanions, " through ‘the one-dlmensronu] chain (and ring) poly- ““anions, the pyrocm:on, cnd funully endmg wrfh " the orthosilicate anion. This second type of solvolytic reaction should ““tend toward depolymerization and, for some parent “\'otoms, go all the way to the ortho-cxyumon be- ““¢ause of the relatively high basic character of " the hydroxyl ion. However, there is no assurance “+ that the ortho-oxyanion will be the most stable <= form in all coses. Third, oxides of atoms of low lomzcmon energy shq_qld react with fused hydroxldes to form oxy- “salts Th:s represen’rs a conversion from a catlon to an cmlon. For example, (22) NiO + 20H-==Ni0,~~ + HOH 95 sy v v o T3y s \,,ai i S Q A ? %.“ ‘-',.} it i ANP QUARTERLY PROGRESS REPORT If a meta-oxysalt rather than an oxide is used as the starting material, in some instances, the meta- oxysalt may be converted into the ortho-oxysalt. Fourth, oxides of atoms having very low ioni- zation energy should not undergo solvolysis at ail. For example, the inert behavior of magnesium oxide toward sodium hydroxide is no doubt due to the low ionization energy of magnesium. There still remain a large number of potential solutes nof treated above. Of these, the only ones which the writer has considered are the relatively ';tomc metolhc ‘salts, These substances will “probably be separated into ions by the highly dipolar hydroxyl ions, and the solvolysis of the cation and the anion can be considered separately. The cation should tend to form oxides and oxy- salts. This reaction should go virtually to com- - pletion except for metals of very small polarization potential. The anionic-solvolytic reaction can '”vary from very complex to nil, and an adequate " “treatment will require the development of a some- what more generalized acid-base theory. 'As an example of the reaction of a metal salt with a fused hydroxide, a small amount of nickel chloride was added to fused sodium hydroxide at 400°C. The reaction proceeded rapidly with the evolution of gaseous water and the formation of a fine black precipitate. The precipitate was separated from the hydroxide and found by x-ray analysis to consist largely of nickel(ll) oxide, together with a small amount of an unidentified compound. This reaction can probably be represented, in large measure, as (23) NiCl, + 20H™ <= NiO + HOH + 2CI~ Previous . experiments indicate that at higher temperatures the reaction represented by Eq. 22 would have followed that represented by Eq. 23. Oxidic-Neutralization-Type Reactions. The neu- tralization reactions of nonprotonic, oxidic solutes can be divided into two classes. First, the neutralization of a base by water takes place according to the reaction (24)' BO + HOH—> B** + 20H- Thls reaction will only occur for nonprotonic 1ac1ds, B** which are stronger than water. Second, reactions of the following type, which “décur in pure oxide melts, would represent a - special type of pseudoneutralization in which no % solvent would be produced (25) Acid + O~ — Base The reaction shown in Eq. 25 includes the entire host of reactions treated in the original Lux theory. When reactions of the type given in Eq. 25 occur in fused hydroxides, the oxide ion concentration can be varied over a wide range of values. This is not possible with the pure oxide melts to which the Lux theory has previously been applied. Ex- amples of the classes of reactions which are of the type represented by Eq. 25 may be obtained by substituting O~~ for 20H~ and deleting H,O in Eqs. 20 through 23. The essential difference between reactions of the type of Eq. 25 and those given in Egs. 20 through 23 is that water is present in the latter reactions but not in the former. The presence of the acid of medium strength should shift the equilibrium considerably for much weaker acids, that is, those derived from atoms of low ionization energy, but only slightly for stronger acids, those derived from atoms of high ionization energy. Mixed-Type Reactions. Finally, reactions of substances of what may be called ‘‘mixed’’ types should be mentioned. These mixed reactions might involve protonic-oxidic salts of alkali metals, and protonic or oxidic salts of nonalkali metals, or combinations of both. The resulting reactions should be capable of analysis in terms of the same considerations used in arriving at the conclusions already stated. The Bronsted protonic theory was originally de- veloped to treat the more useful low-temperature solvents, such as water and liquid ammonia, while the Lux theory was originally developed to treat the metallurgically important, high-temperature, fused-oxide systems. It is of some theoretical interest that both theories may be rigorously applied to reactions in fused hydroxides, although the types of reactions for which these applications are useful form two sets of almost mutually ex- clusive reactions — one set for each theory. This is symptomatic of the need to develop a more general acid-base theory along the lines proposed by Audrieth.!! For many solvent systems, such a theory would be a luxury; for fused hydroxides, such a theory is a necessity. TIL. F. Audrieth and J. Kleinberg, Non-Aqueous Sol- vents, Wiley, New York, 1953, p 2723, ot e P P . ?fr:wv\"a " ) » gt . F‘.g‘ It should be noted that of the types. of reactions discussed above only a few examples have been studied experimentally in a detailed or quantitative way; most have not been studied at all or else are represented experimentally by some of the more obvious cnd less m'rereshng exompies. Flammublllty of Alkah Metal Sofuhons at ngh Temperatures G. P. Sml'rh M. E. Sfeldlltz Mei_'q‘“‘urgy Division In the pre'c:‘eding-hirfi‘epwon‘12 data were given for the flammability of alloys of the sodium-bismuth system, Additional data have now been obtained for bismuth-rich alloys that provide a much more complete picture of the reactivity of this system. The relative reactivity of sodium-bismuth solutions with dry air at 700°C as a function of the mole fraction of sodium is shown in Fig. 6.17. The relative reochvuty scale is “arbitrary. Pure sodium was assigned a reactivity of four units. Unre- active solutions were assigned a reactivity of zero. No measurements were made for mole fractions of sodium between 0.6 and 1.0 because a solid compéund"6"frwébmpositi'o'r'1'-NG;Bi'-p'recipi- tates from solution at a mole fraction of sodium somewhat greater than 0.6. A temperature-composition diagram for liquid sodium-bismuth soluhons is presented in Fig. 6.18, on which is ploh‘ed a curve fhc’r approx:- 2y, E. Steid|ifz, L. L. Hall, and G. P. Smith, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 102. i, UNCLASSIFIED ' ORNL-LR-DWG 4653 Vv 5 4 . > - ’._. S N £ - g 3 - ul b / > 2 o & / - wl . , x o, . Op— — ‘. *o-¢ JV\, 6.17, Relative Reactivity of Sodium. Bismuth Solutions with Dry Air at 700°C as a Function of the Mole Fraction of Sodium. PERIOD ENDING DECEMBER 10, 1954 mately separates the region of reaction from the region of no reaction. The circles and triangles on this diagram represent the temperature-compo- sition values at which tests were conducted. Most of the circles represent two to four tests each. The open circles represent tests in which reaction occurred, while the black circles repre- sent those which showed no reaction. The black triangles represent conditions under which part of the tests showed no reaction, while the rest of the tests showed a slight reaction. The two open triangles represent tests which showed an exceedingly slight amount of reaction. The broken portion of the curve is poorly defined, inasmuch as it may actually pass beneath the open triangles rather than above them, as shown. It may be noted that jets of pure bismuth showed appreciable re- activity at 800°C but not at 750°C, When air was saturated with water vapor, the line of zero reactivity was shifted toward lower sodium concentrations by a small but appreciable amount for sodium-bismuth solutions. This effect is shown for sodium-bismuth solutions at 700°C in Fig. 6.19. It may be seen that, although data for moist air scatter badly, there is unquestionably a small shift toward lower sodium concentrations. Combustion studies have been run on four of the six alkali metal—alkali halide systems for which UNCLASSIFIED ORNL-LR-DWG 4654 900 REACTION g w x 2 2 ° A O-OF-—O0—< E \ a = 1 " \ 600 . ol NO REACTION .\ 500 . 0 0.2 0.4 . : 0.6 MOLE FRACTION OF SODIUM Temperature-Composition |D‘|c|gmm SHmeg Reglons Within Which Jets of Sodium- Bismuth Solutions Did and Did Not Recct with Dry Air, 97 T R T R T ANP QUARTERLY PROGRESS REPORT phase diagrams are available. The results de- scribed in Table 6.9 represent one test of each - system at the temperature indicated. The temper- ‘ature chosen lies in the one-liquid phase region of the phase diagram — generally about 50°C cb'éve' the two-liquid phase region. As may be seen, all tests were conducted in the metal-poor sec'non of the system. UNCLASSIFIED " ORNL-LR-DWG 4655 <, : . & & 1 r 3 . > 5 <1 5, / b = 5 — L { Y m * "0 T~ / 0 02 T 0.4 0.6 MOLE FRACTION OF SODIUM Fig.‘ 6.19. Relative Reactivity of Sodium- Bismuth Solutions with Moist Air at 700°C. The curve represents data for dry air. TABLE 6.9. FLAMMABILITY OF ALKALI METAL-ALKALI HALIDE SYSTEMS Metal Content Test Temperature System (mole %) ©C) Reaction Na-NaF 2 1050 Mild 5 ‘ 1050 Violent 10 1050 Violent NaNaCl 2 850 Little 6.6 900 Violent 30 1050 Vielent Na-Ngal 2 700 Mild 5 800 Mild 10 900 Violent 30 1000 Violent K-KF 2 900 Mild 5 900 Violent - 10 900 Violent 98 ~ CHEMICAL STUDIES OF CORROSION F. Kertesz o | Materials Chemistry Division Effect of Temperature on the Corrosion of Inconel Melts with NiF , Additions H. J. Buttram R. E. Meadows N. V. Smith Materials Chemistry DlVlSion In order to sfudy specn‘tcclly the removal of chromium from Inconel by a known corroding agent in fluoride melts, additions of NiF2 were made to NoF-ZrF ,-UF (53.5-40.0-6.5 mole %) in Inconel capsules. These capsules were subjected to 100-hr static tests in the temperature range 600 to 1000°C. When the chromium concentration of the melt after the tests was plotied against the amount of NiF. added to the melt, a temperature effect at a relatively high concentration of NiF, (690 meq/kg) was noted. Additions of NiF, up to about 70 meq/kg yielded chromium concentrations which were independent of temperature to within +100 ppm Cr**. The mixture to which the large addition of NiF, (690 meq/kg) was made showed, at 600°C, a chromlum content of 280 ppm Cr +4 that was considerably below the stoichiometric quantity; at 800 and 900°C, there were successive increases to 630 and 840 ppm Cr**; while at 1000°C, 550 ppm Cr** could be found. Metallographic obser- vations paralleled the chemical analyses. Pos- sible explanations for the finding of decreased chromium content at 1000°C were set forth ‘in a previous report,'3 Effect of Chromium Additions on f|1e Corrosmn of 7 Inconel in Fluoride Melts with NIF ‘Additions H. J. Buttram R. E. Meodows Materials Chemistry Division The behavior of Inconel capsules exposed to NaZrF, and to NaF-ZrF -UF, (53.5-40.0-6.5 mole %) was studied by using 100-hr tilting-furnace 131, J. Buttram, R. E, Meadows, and N. V. Smith, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 108. ¥ i s gy y " ) & W AL 4 He tests in order to establish the effect of odded nickel fluoride. In the absence of added chromium metal, the plot of the chromium concentration after test vs the nickel fluoride added followed a linear relationship, although the amount found in the case of very large additions of NiF, was smaller than stoichiometric. Metallographic tests showed a heavy subsurface void formation to a depth of 21 mils at the hot end (800°C) and 5 mils af the cold end (650°C), while the corresponding blank capsules (no NiF,) showed only light attack to a depth of 1 to 2 mils. A series of Inconel capsules containing the same two fluoride melts and NiF, with chromium pellets added were subjected to fi'\e same type of testing. The after-test chromium values were above stoichiometric, and metallographic exami- nation of the capsule walls indicated that the NiF., additions had little effect on the amount of attack. The heaviest attack was to a depth of 3 mils, with the average being 1 to 2 mils. The cold ends had thin metal depesits and a slight void formation. PERIOD ENDING DECEMBER 10, 1954 Effect of Chromium Valence State on Corrosion of Inconel H. J. Buttram R. E. Meadows Materials Chemistry Division Additions of chromous and chromic fluorides were made to NaZrF_ and NaF-ZrF4-UF4 (53.5- 40.0-6.5 mole %) in Inconel capsules to determine the effect of chromium valence state. Metallo- graphic examination showed that chromous fluoride additions had essentially no effect even in the case of large additions (18,000 ppm Cr**). The depth of attack was about 1 mil. As would be expected, when chromic fluoride was added, the Inconel showed attack to a depth of 4 mils and some evidence of intergranular penetration. The chromous ion has been demonstrated as the stable state in the NcF-Zer-UF4 systems; therefore, it is reasonable to assume that, when chromic fluoride is the additive, chromium metal will be removed from the Inconel to enable the chromium to be reduced to the chromous state. Chemical analyses made after these tests were inconclusive and difficult to interpret. 99 b s Bt duoty i i Bk ek iz oo S ' ANP QUARTERLY PROGRESS REPORT 7. METALLURGY AND CERAMICS “W. D. Manly J M. Wordé Metallurgy Division The mckel-molybdenum base alloys are bemg “studied extenswely as possible reactor structural "4mu'rer|als ‘with qualities superior to those of _lnconel Efforts are under way to evaluate and improve the existing alloys and to develop better .ones. Radjator test assemblies have been fabrie “cated, and weld stability and mechanical properties ~are being studied. Work has continued on the preparation of duplex tubing, boron carbide shield- ing, tubular fuel elements, and other special materials., The results of oxidation resistance tests of brazing alloys are presented. In addition, the first produchon of beryllium oxide ceramics by ‘castmg from a sllp which is basic is described. - DEVELOPMENT OF NICKEL-MOLYBDENUM BASE ALLOYS The investigdtions under way for the evaluation of nickel-molybdenum base alloys as structural materials for circulating-fuel reactors include attempts to improve the ductility and the fabrica- bility of commercially available Hastelloy B through purification, and efforts to find another suitable and improved nickel-molybdenum base alloy that has the strength and corrosion resistance of Hastelloy B. Fabrication experiments are being used to determine the effects of various treatments on the materials developed, Fabrication Experiments H. Inouye J. H. Coobs Metallurgy Division One of the difficulties experienced in the fabri- cation of Hastelloy B is due to the narrow range of forgeability of the alloy — between 1950 and '2100°F. The upper limit of forgeability is due to impurities which cause grain-boundary melting. “The low ductility at temperatures between 900 and . 1800°F is less understood, but it is believed to be " due to aging and impurity precipitation in grain boundaries. . to eliminate tfrace (tramp) elements by vacuum .’___:'me[hng and the addition of elements to neutralize " their effects. Therefore efforts are being made One of the most critical components of a circu- -r‘rj"."‘rlqfing-'fuel reactor will be the fluoride-to-NaK 100 heat exchanger, for which, because of its com- plexity and fragility, it will be highly desirable to have seamless tubing. Commercially available Hastelloy B tubing, which is now made by welding strip, is not satisfactory because of cracks in the weld, checks on the surface, and, in the dbse'nce of severe working of the weld, nonuniform proper- ties, Therefore several extrusion experiments have been performed in an effort to produce seamless tubing. The initial extrusion experiments were attempts to produce rod from air melts of Hastelloys B and C. When extrusion temperatures above 2100°F and up to 2350°F were used, the a”oys fractured severely in all instances. For the subsequent experiments, the extrusion temperature was lowered to 2100°F, and a reduction ratio of 6.3:1 was established. The results obtained for the various materials are summarized in the following. Vacuum-Melted, As-Cast Hastelloy B. The ingot that was extruded had been homogeneized at 2100°F for 48 hr. The front of the extrusion fractured severely, but some improvement was noted in comparison with the air melts previously extruded, Air-Melted, As-Cast Hastelloy B Plus 0.2% Ti as Ti-Mn-Al-Ni Master Alloy. The ingot was homogeneized 48 hr at 2100°F before extrusion. The exirusion fractured severely, and no improve- ment in comparison with the vacuum melt was noted. Air-Melted, Commercial, Wrought Hastelloy B. The extrusion fractured severely and thus indicated that the cast structure was not responsible for the poor hot forgeability of the alloy. Air-Melted, Commercial, As-Cast Hastelloy C. The extrusion was made through @ cascade die (cone plus shear) to determine the effect of die design. Thus far, the alloy has been extruded through cone dies varying from 45 to 25 deg and through a shear die. In all exfrusions, the rod fractured severely. An extrusion was also made by containing the billet in an Inconel can 0,063 in. thick on the outside diameter and with a %-in.- thick disk at the nose of the biilet. The Inconel nose separated from the Hastelloy, but the Hastelloy 0 P ,i»—m:": SN N 3 R H ) ) . ) t{ HeE ~ the extrusion. was fractured only on the front 1 in.” The remainder of the extrusion appeared to be sound. A thin layer of Inconel was found on the whole length of This experiment has not, as yet, been fully evaluated. ' NlckeI-Molybdenum Blnury Alloys. In contrast to the experiments with the Hastelloys, previous extrusion experiments with vacuum melts of the binary alloys in the cast condition showed marked improvement in the quality and ease of fabrication. The tendency to fracture was still apparent to a slight degree and became progressively more noticeable as the molybdenum content of the alloy was increased from 20 to 24%. Two extrusions of tube blanks of the 24% molybdenum alloy were unsatisfactory for further reduction to small-diameter tubing because of the large fractures and deep- defects. 1t was found that the defects could be eliminated by homogeneizing the cast billets for 48 hr at 2100°F, The minimum homogeneizing time was not determined. The data obtained in the extrusions of the nickel-molybdenum binary alloys are presented in Table 7.1. PERIOD ENDING DECEMBER 10, 1954 Since Hastelloy C could not be extruded with any degree of success until it was canned Inconel, it is concluded that lubrication is an important factor in these experiments, At the temperature of extrusion, although the alloy is probably notmolten, the grainboundary is weakened sufficiently by the friction of the flowing metal against the die to cause fracture. By providing a layer of Inconel between the die and the alloy, as by canning, the friction on the alloy is removed. The friction could also be reduced by increasing the extrusion temperature, but, to increase the extrusion temperature, it will be necessary to eliminate the lowsmelting-point impurities from the alloys. An alternative would be to reduce the extrusion temperature and the extrusion ratios. The following tentative conclusions have been drawn: (1) Vacuum melting has little effect on the extrudability of the alloys. (2) The extrudability of the alloy does not depend on whether it is in the wrought or the as-cast condition, (3) Homoge- neizing has a beneficial effect, as evidenced by alloys of the24% Mo~76% Ni composition. (4) Poor TABLE 7.1. EXTRUSION DATA FOR NICKEL-MOLYBDENUM BINARY ALLOYS Atmosphere: Houghton’s salt bath No. 1550 Heating Time: tube blanks, 30 min; rod, 45 min Billets: as-cast, 3 in. in diameter and 3 in. long, homogeneized 48 hr in helium plus hydrogen Die: 30- to '45-deg cone die, olloy CHW Mandrel: 1-in. stem, straight, olloy LPD Lubrication: glass wool in container, Fiberglas sleeving on mandrel stem Dummy Block: 70-30 brass coated with Necrolene Pressure 'R'equiremems (on 3-in. ram}: maximum, 700 tons; minimum, 487 tons Composlflon R T I R AT SOy - Temperature ~ Heat No. (Wf %) Extrusion Remarks Ratio ' '_':'Two fube ex'rruswns, good surfuces “HIOne rov ext.rus:on, good surfoces o B f'One fube extrusnon, good surfoces | : Two fube extrus:ons, minor f|ows ' One rod exfrus:on, good surfoces VEOne ’rube extrus:on, shqttered o One tube extrusion; minor defect 2 in. R FR A A from front 24 32 68 2150 6.3 :1 One rod extrusion; failed to extrude 101 .{"2 Qo i o v g" ;,E ANP QUARTER[Y PROGRESS REPORT fdbr‘icabilify of the cast alloy is associated with a second phase which persists after long-time heat treatments at 2100°F. New Alloys H. Inouye J. H. Coobs Metallurgy Division Several new alloys arebeing studied in an attempt to find an ulloy that is superior to Hastelloy B, “and efforfs are ‘béing made to improve the existing Hastelloys. Prelammary results of experiments o ”:f.w:’rh aHoys of various compositions are presented “in the following. Hastelloy B Plus 0.03% Cerium. A 3-Ib vacuum melt was rolled at 2100°F, and there was con- siderably less cracking than would have been found with Hastelloy B. Numerous small defects . _were present, however, that may be eliminated by - rincreasing the cerium content of the alloy. " Molybdenum-Columbium-Nickel Alloy. A 20% Mo—-5% Cb-75% Ni alloy could be cold rolled from the cast structure to sheet, The structure is two- phased at 1500°F and single-phased at a solution annealing temperature of 1950°F. The tensile strength of the alloy at 1500°F is 36,800 psi (elongation 3.5%). A creep test at a temperature of 1500°F and a stress of 8000 psi is in progress. Molybdenum-Aluminum-Nickel Alloy. A 20% Mo-2% Al-78% Ni alloy could be rolled at room temperature from the cast condition. At 1500°F the tensile strength was 43,500 psi with an elonga- tion of 5%. Molybdenum-Nickel Alloy, A 100-g arc melt of a 24% Mo-76% Ni composition was hot rolled 60% at 2200°F, Following a I-hr anneal at 2100°F, the alloy was cold rolled 72% without cracking. The work hardening data for the alloy are listed below: Cold Work (%) Hardness YPN (20-kg load) Annealed 200 25 347 29.6 | 401 405 423 s 432 0.4 483 720 470 R A 3-1b vacuum melt of this alloy has been rolled to sheet and will be screened for strength and ductility in creep and tensile tests. In addition, an extrusion of the alloy has been rolled to sheet for obtaining more detailed data, A 100-g arc melt of a 32% Mo-68% Ni composition was successfully hot rolled 60% at 2200°F, The alloy was further cold rolled 72% without cracking. The work hardening data for the alloy are listed below: Cold Work (%) Hardness V PN (20-kg load) Annealed 231 21 386 31 444 41 471 54 513 62 502 72 516 A 3-lb vacuum melt of the alloy has been rolled to 0.063-in. sheet and is being machined into creep and tensile test specimens. A 100-g arc melt of a 38% Mo~62% Ni compo- sition was rolled 60% at 2200°F. The alloycracked when cold rolled. In the annealed condition the hardness of the alloy is 396 VPN, The micro- structure of the cast alloy shows a considerable quantity of eutectic, which appears to be similar to that found in Hastelloy B. The following additional arc-melted alloys were evaluated: Alloy Results 75% Ni—20% Mo—5% Al 70% Ni-20% Mo-10% Cb Slight edge cracking during hot rolling, further cold rolled to sheet without difficulty Cracked during hot rolling 78% Ni~20% Mo—-2% V Successfully hot rolled and cold rolled to sheet with no visible flaws 78% Ni~20% Mo—2% Zr Moderate amount of cracking during hot rolling that be- came progressively more severe during cold roelling A D e ok i B Ay Oxidation and Oxidation Protection H. lnouye J. H. Coobs Metollurgy Division Temporary cychng tests from 1500°F to room temperature in air have been performed on Hastelloy B and other nickel-molybdenum base alloys to determine the behavior to be expected in a NaK-to- air heat exchanger such as will be required for aircraftreactor application. The severest operating condition that could exist would be a thermal cycle from operating temperature (1500°F) to room temperature under stresses caused by high-velocity air. The cycling tests of Hastelloy B were evaluated by daily visual examination of the test piece which had been through a heating cycle of 17 hr at 1500°F and then coolingto room temperature in the furnace. Weight changes were also measured daily. In conjunction with these tests, identically prepared specimens were tested at a constant temperature of 1500°F, The results of these initial tests are shown in Table 7.2 and in Fig. 7.1. The results presented in Table 7.2 show that the loss of metal from the Hastelloy B specimen through oxide spalling was prevented by the use of platings. in general, aluminum and aluminum plus nickel are considered to be unsatisfactory platings, however, because of spalling of the protective coating. Several other alloys have been tested for oxi- dation at 1500°F for comparison with Hastelloy B and to determine whether they need oxidation protection during elevated-temperature mechanical and corrosion tests. The results of the tests are showp_ in Fig. 7.2. PERIOD ENDING DECEMBER 10, 1954 At a constant elevated temperature in air, nickel- molybdenum alloys of intermediate compositions form a protective coating of NiMoO,. Upon cooling, regardless of the rate, the coating will spall vigor- ously at temperatures between 440 and 230°F.° Experimental evidence shows that this spalling may be due to a change in the crystalline structure of the NiMoO, coating. Radiator Fabrication P. Patriarca K. W. Reber R. E. Clausing G. M. Slaughter Metallurgy Division J. M. Cisar Aircraft Reactor Engineering Divisicn R. L. Heestand Pratt & Whitney Aircraft A series of test specimens of Hastelloy B radi- ators has been fabricated for use in evaluating the ability of Hastelloy B to withstand thermal shock and oxidation., A specimen that includes approxi- mately 3 in. of Inconel-clad copper fins spaced 15 fins per inch is illustrated in Fig. 7.3. The 20 tube-to-header joints were inert-arc welded by using the semiautomatic welding equip- ment described previously.? The manifolds were manually inert-arc welded, and the assembly was brazed to effect the tube-to-fin joints and the J W. Spretnak and R. Speiser, Protection of Mo lybdenum Against Corrosion at High Temperatures, QOhio State University Research Foundahon Status Report No. 8, 2P Patriarca et al., ANP Quar, Prog. Rep. June 10, ' '1954 0RNL-]729 Flg. 65, p Y6, TABLE\‘7.2 ) RESULTS OF CYCLING TESTS IN AIR oN HASTELLOY B g ,AND PLATED HASTELLOY B SPECIMENS We'igh'r'Chunge iri L ,inf_quingJon Hustéilqy B T gy CYdies Surface Condition - -0.0775 Spalled ™ of chromium ST seomsT 0 Nospalling _0 00] |n of nlckel p|us 0 0001 in. of chromium +0.0035 No spealling 0. 001 ine of mckérl‘ plus a!ummum spray TUR0.0512 " Spalled Aluminum spray +0.0370 Spalled 103 ANP QUARTERLY PROGRESS REPORT WEIGHT INCREASE (10~ 3g/cm?) 0 20 40 60 80 backing up of the tube-to-header joints with Coast Metals_No. 52 brazing alloy. The fabrication of the three test assemblies made to date has been routine, and no unusual behavior or difficulties have been encountered, Although it is expected that the characteristic aging of Hastelloy B and the subsequent loss in room-temperature ductility will affect the longevity of these test specimens, the first assembly has withstood, without failure, as determined by a hehum Ieok test, 100 accrued hours of life with ~one air quench and one water quench directed at the fin surfaces. Othe_r test assemblies are being 'fobncofed with type 310 stainless-steel-clad ~ the evaluation of Hastelloy B, the dimensional ~ stability and oxidation resistance of this material - under thermal shock. 104 copper in order to determine, in conjunction wr’rh‘ UNCLASSIFIED ORNL~LR—~DWG 4426 Ni+Al PLATED HASTELLOY B Al PLATED HASTELLOY HASTELLOY B Ni+ Cr PLATED HASTELLOY B Cr PLATED HASTELLOY B 100 120 140 160 180 200 ELAPSED TIME (hr) Fig. 7.1. Oxidotion of Plated Hastelloy B in Air ot 1500°F, Welding P. Patriarca K. W. Reber R. E. Clausing. G. M. Slaughter Metallurgy Division J. M. Cisar Aircraft Reactor Engineering Division R. L. Heestand Pratt & Whitney Aircraft Preparations are under way for experiments to evaluate the properties of Hastelloy B weld metal as affected by thermal history. A comparative basis for the tests has been prepared through experiments in which the effect of aging time at 1500°F on the room-temperature ductility of wrought Hastelloy” B was determined For these tests, guided bend specimens, /8 X ]/2 X 13/ in,, were cut from /16-|n.~fh1ck wrought Has'relloy B plate e ‘\.-‘ - i, s s & 0 " s " Hastelloy B whenwag’é: PERIOD ENDING DECEMBER 10, 1954 UNCLASSIFIED ORNL—LR—DWG 4427 14 | 12 //‘ ‘° < ~ 7 20 % Mo —80% Ni Sa / o g s Q A" % ' /// © /ln ‘ HASTELLOY B ' &6 — ' 20 % Mo —5 % Cb — 75 % Ni — £ y 20% Mo — 2% Al — 78 % Ni ‘ A | | - = 4 // ] V’ & T | {fl_—&-—- A o : _..,-KJ——V-A‘ - //.’/ : : e O ’——I——— et ——— = 2 ——l = ___-___.—- = . ./7_ 1T " \ 24% Mo ~ 76 T Ni P 7(;‘ = 32% Mo — 68 % Ni O . 0 20 40 60 80 100 120 140 160 180 200 ELAPSED TIME (hr} Fig. 7.2. Oxidation of Nickel-Molybdenum Base Alloys in Air at 1500°F which had been solution annealed at 2150°F and a period of exposure of less than 50 hr. If the quenched. One set of specimens was then aged material is given o stabilizing heat treatment, at 1500°F in vacuum for various perlods of up to "'however, this reduction in ductility is prolonged 300 hr before feshng A" second group oF the "T“for a period of between 100 and 200 hr. The spec:mens ‘wds’ heat treated in hydrogen for /2 hr results of this preliminary investigation indicate . at 2]50°E and then_ for. 30 hr “at 1950°F and wa’rer ___that the application of a stabilizing heat treatment ’quenched bef ] ficml cnd that longer treatments at 1950°F might effect more complete stabilization, ‘A similar group of specimens is being cut from inert-arc butt-welded %wn.-‘rhuck plate in such a "'M‘:'mcnner that the weld metal will constitute the 'ter T/4 in. of fhe bend specimen. Thfea effect of g9 re‘wfhey were aged cn‘ ]500°F ”‘Aerc’rure ducflllty of toom-temperature ‘and 1500°F bend tests, and the ‘fa'r a simulated service results will be compured wuth those for wrought temperature of 1500°F is markedly reduced after Hastelloy B. be seen that the 105 ng-| g time ‘@nd prior thermal hlsfory on these ”"""geomefry and are presenfed in Table 7.3. If'ymaf”'"i'fi'él:'f-arc ‘welds will be determined by using both TomurErT ANP QUARTERLY PROGRESS REPORT ;UNCLASSIFIED Y3617 Fig. 7.3, NaK-to-Air Radiator with Hastelloy B Tubing ond Manifolding and a 3-in. Section of High-Conductivity Copper Fins Spaced 15 Fins per Inch. Mechanical Properties Studies R. B. Oliver D. A. Douglas J. H. DeVan J. W, Woods ‘ Metallurgy Division - M. D'Amore Pratt & Whitney Aircraft The stress-rupture properties of the nickel- molybdenum base alloys are being studied in fused salts, sodium, and various gases. Six new lever arm machines are being used to testHastelloy B sheet specimens, and eight additional tube burst units have Hastelloy B pipe in test. |t is anticipated that design curves for properties in the fused salts at 1500 and 1650°F will be available by February 1955. Data are also being obtained for a design curve for properties in argon. The results obtained in argon at 1500°F are presented in Table 7.4. As Table 7.4 indicates, Hastelloy B has sub- stantially greater strength at 1500°F than Inconel, which will rupture in 1200 hr when stressed to 3500 psi. A Hastelloy B specimen tested in hydro- gen at 12,000 psi ruptured in 1300 hr; o similar time to rupture was found in argon. Thus there is an indication that Hastelloy B may not be sensitive to a hydrogen atmosphere when tested in the so- lution-annealed condition. TABLE 7.3. EFFECT OF AGING TIME ON ROOM-TEMPERATURE BEND DUCTILITY OF WROUGHT HASTELLOY B Tilme at ' 1500°F Bend Angle Before Fracture {deg) (hr) | As-Received Specimen? Heat«Treated Spec_imenb 0 ~ Greater than 135,€ no fracture Greater than 135,l no friacfure_ 25 - Greater than 135, fissuring evident Greater than 135, no fracture 50 92, complete fracture Greater than 135, no fracture 775 90, complete fracture Greater than 135, no fraci‘ufe TOO | 84,> complete fracture Greater than .]35,_.no fracture - 200 80, complete fracture 90, complete fracture | 223 76, complete fracture 82, complete i';rc‘;lcfure 300 | 80, éompiete fracture 75, complete fracture Solution annealed at 2150°F and quenched. bHeat treated in hydrogen for ]/2 hr at 2150°F and then for 30 hr at 1950°F followed by water quench. _ “Maximum bend angle available with apparatus. 106 o4 - " n FTI o TABLE 7,4, STRESS-RUPTURE PROPERTIES OF HASTELLOY B IN ARGON AT 1500°F Stress Time to Rupture Elongation {psi) (hr) (%) 15,000 157 45.0 13,500 180 15.0 12,000 1150 20.0 8,500 2200* 4.0 *Specimen still in test. The new alloys being developed are also tested for stress rupture at 1500°F in argon in order to check the effect on strength which results from the various modifications. The results of tests made to date are presented in Table 7.5, These modified alloys show a decrease in high-tempera- ture strength and elongation in comparison with Hastelloy B in the solution-annealed condition. SPECIAL MATERIALS FABRICATION H. Inouye J. H. Coobs Mefallurgy Division Duplex Tublng The efforts to produce duplex seamless ’rubmg that ‘will have good oxidation resistance on the outer surface and good corrosion resistance on the inner surface have continued. In the previous report,3 tests were described in which hot-rolled and hot-pressed composnes were deep drawn. An alternaflve and more dlrecf method of producing ~ Sept. 10, 1954, ORNL-1771 3, - J. P Coobs and H. lnouyle,QaANP Quar. Prog. Rep, p PERIOD ENDING DECEMBER 10, 1954 tubing is now being used in which the composite billets are extruded. A study of the flow of metals during an impact extrusion is being made, For the study, the effect of the shape of the nose of the extrusion billet was determined on extrusions of Zircaloy canned in copper and extrusions of vanadium canned in steel. In addition, a two-ply and a three-ply billet of stainless steel—carbon steel were extruded, but the extrusions have not yet been evaluated. The results of the tests show that large-grained billets, such as cast billets, produce a rough interface between the outer cladding and the core. A tapered nose on the can and @ square edge on the core result in a short section of can material restricted to the front of the billet; thereafter, the core has a thin layer of the canning material on the surface. A square nose on the can and a tapered core result in a thin section of core begin- ning very soon after the front of the extrusion and increasing in thicknesses toward the end of the extrusion, Boroh Carbide Shielding Suitable compositions for the boron carbide shield for the ART heat exchanger are being de- veloped by The Carborundum Company and by the Norton Company by using various nonmetallic bonding materials, including BN, SiC, and carbon, as well as those mentioned previously.® All these bonding materials have been used to some extent in fabricating useful forms of B,C. Another pos- sible source for fabrication of these shield pieces is Sylvania Electric Products, Inc., which only Vrecenfly became mteresfed in the prob!em. In" a recent conference ‘with representatives of The Carborundum Company it was stated that they ‘Elongation 7 Time foRupture R (hr) (%) gy TR AT g 8,000 510 " 60T 111 L A N 4 vt b wie ks 1l i ANP QUARTERLY PROGRESS REPORT have developed a suitable composition by using SiC as the bonding material. However, specific details and samples of the material have yet to be received, Tubular Fuel Elements Twelve fuel plates have been prepared for forming into tubes, and eight more are being prepared. These tubes are to be drawn at Superior Tube Co. by the plug-drawing technique with low (10%) reductions. Other techniques such as hot drawing ~or hot quging may be used. Control Rods Th:rty-five control rods are belng prepared for the GE-ANP project. They were received as -specially straightened tubes, 0.504 in. ID, 0.625 ‘in, OD, 35 in. long, with one end plugged. The tubes are b'e'ihgwfi“e'd with a mixture of 50% alumi- ~num powder and 50% B 4C, both supplied by the “General Electric Company. The B,C is a special grade containing 81% boron. The mixture is pre- pared with 2% paraffin as a binder and is being packed tightly into the tubes with a pneumatic hammer. Very small increments are used to obtain a high density. Twenty tubes have been filled with average boron concentrations of about 0.8 g/cm3, that is, somewhat in excess of the 0.7 g/cm® required. After filling is completed, the . tubes will be heated and evacuated to remove the paraffin and then cold swaged to final size. Al-UO, Elements for Shielding Expetiment Fuel plates 28 in. long and 21/4 or 21/2 in, wide are being prepared for an experiment to determine the shielding necessary for delayed neutrons. These plates are to be mounted on a belt and rotated at speeds up to 22 fps. The fuel concen- tration of 0.25 g/cm® required for these plates is quife high compared with that usually specified for fuel plates, and therefore new fabrication problems must be resclved. " Several plates have been fabricated by using .UOZ and aluminum powder cores roll clad with aluminum. Three plates having 0.040-in. core ~ thickness wnh 62 wt % UOQ, had the required fuel 'concenfrahon Both Hhigh- flred and steam-oxidized UO2 were used with good results. The edges of the core were quite clean and straight, and the U0, was distributed fairly uniformly, as shown by radiograph examination. Because of the high 108 percentage of UG, it was necessary to hot roll the plates with the covers on at all stages to prevent oxidation. Plates with lower percentages of UO, may be protected with thin aluminum foil pressed on the core compact during cold pressing. The finished plates, as : . «0 K Bt T, L s s o e N ot . W B o0 " ¥ i a much smaller area was obscured by the grid lines themselves. Inherent difficulties have beset attempts to measure the settling velocity of the particles; however, it can be said that the terminal velocity of free fall of Lycopodium spores in water under gravity is certainly less than 0.2 cm/sec and probably much less. Even this is only about 0.16% of the mean velocity which will be measured, The Hewlett-Packard counter which measures the time interval between flashes was adjusted; its absolute accuracy was checked; and it was found to be satisfactory. ' The design of another entrance section which will |mpcrt a rotational component of velocity to the fluid as it flows through the test section has been started. This will produce a velocity struc- ture which is more realistic in terms of what is envisaged for the ART, but it is much more difficult to measure. It is conceivable that a rotational compbhen‘f will help to réduce the 'possfl)le adverse thermal effect of separation regions, perhaps by eliminating these regions. . The velocity distribution in a diverging channel under turbulent flow conditions was studied with the phosphore$centéparfic|e technique. Flow asymmetry at a half-angle of 4 deg was observed. This angle is somewhat less than the 5-deg half- angle given in the classical report on velocity profiles in channels by Nikuradse.? A large flow- visualization system that utilizes the phospho- rescent particle and dye techniques for studying fluid flow characteristics has nearly been com- pleted. - sxs descnlfi)e_d "rhe“ BAET IR AT ‘strdcture ind samp]e forced flow volume heat- “golirce “system,’ 5 Vconducted “on -‘the electrlc':ro! mefhod, oi?&.generchng 1-49 . Nikuradse, Forschungsarbeiten VDI 289, . sandw:ches fiacnd solutlon wn‘h qnd wn‘hout elecfr PERIOD ENDING DECEMBER 10, 1954 volume heat sources within the fluids of experi- mental forced-flow volume-heat-source systems. An experimental study of the factors affecting gas formation at an electrode-electrolyte interface was carried out pursuant to the problem cf gener- ating volume heat sources electrically. The generation of gas at an electrode was found to be dependent upon the material of the electrode and the current density and independent of the applied potential, as well as the type of electro- lyte employed. Platinum was found to have the most desirable characteristic of conducting the maximum amount of current per unit area of elec- trode to an electrolyte for a given degree of gas generation. This effect is most probably the result of an increased surface area because of the known porosity of this metal. For brightly polished platinum, a current density of approxi- mately 8 amp/in.? was achieved before gas liber- ation was noticeable. This value exceeded that of Carpenter 20 stainless steel by two orders of magnitude. Also, tests were carried out to study the compatibility of the more common stainless steels and dilute solutions of strong electrolytes in those portions of an experimental volume-heat- source system where no electric currents are flowing. Carpenter 20, type 347, ond type 316 stainless steels were found to be accordant. Concomitant with the problem of generotmg uniform volume heat sources within annuli of nonuniform cross section, a method for embedding platinum electrodes to form a hydrodynamically smooth surface has been developed. It has been established that several of the new epoxy bonding ..~ plastics adhere to suitably roughened metals in such a manner that the casting of ‘composite of metol _cmd plastic is feasible. Three such expenmenfcl cushngs were fobrlcated !ong p[ashc tubes which were ductmg a' su!furlc AL 7 ‘The results, expressed in terms of the friction factor and Reynolds modulus in Fig. 8.2, indicated that there was no detectable m TN r oy ),,‘ _, v n; ‘rs, el 4 115 (rents, o ANP QUARTERLY PROGRESS REPORT 0.2 — . UNCLASSIFIED ORNL-LR—DWG 4453 OPEN POINTS INDICATE NC CURRENT FLOW SOLID POINTS INDICATE CURRENT FLOW IN FLUID 0.05 {, FRICTION FACTOR . LAMINAR FLOW g 64 Re " Goe Q.01 LENGTH {in.) 20 83 155 {99 BULENT FLOW _0.184 g - Re0-2 10? 2 5 10° Re, REYNOLDS NUMBER Fié. 8.2, Friction Factors With and Without Electric Cutrent Flow in Fluid. difference in pressure drop through the tubes under the two different conditions and that, hence, the hydrodynamic structure was not influenced by the electric currents for the laminar and turbulent ranges investigated. PHYSICAL PROPERTIES MEASUREMENTS T HeatCupacfl'y W, D. Powers G. C. Blalock Reoctor Exper:mental Engineering Division ,7 - The em‘halpies and heat capacities of NaF- ZtF - V_UF4 (53-43-4 mole %) were determined with ’rhe S ‘_Bunsen ice calorimeter in the liquid and solid . states ond wnh the copper-block calorimeter in - ,the hqund state The results are as follows: Sohd (70 to 525°C) Bunsen calorlm_efer __ H = 0.18(2)t 0.18 = 0.01 fi i 16 Liquid (550 to 900°C): Bunsen calorimeter Ht - HOOC = 32 + 0.25¢ ' Cp = 0.25 + 0.02 Liquid (570 to 885°C): copper-block calorimeter H, ~ Hpo = 19.5(0) + 0.265(6) <, 0.265(6) + 0.002 The enthalpies of this salt have also been de- termined by the National Bureau of Standards, and a comparison is made in Table 8.1 of the results, The enthalpies and heat capacities of L:F KF- UF, (48-48-4 mole %) were determined by the c0pper -block calorimeter in the solid and liquid states. Solid (125 to 465°C) H, - Hyo C_0234t+048>< 10-4 2 = 0.234 + 0.95 x 10~4 lLiquid (565 to 880°C) H, - Hgpo s = ~82.54 + 0.657: - 1. 97 x 10=4 42 = 0.657 - 3.93 x 10™* .J (AR m}g Sorendh LD Iy 4 %y X oy iy i Ay 9 L PERIOD ENDING DECEMBER 10, 1954 TABLE 8.1. COMPARISON OF ENTHALPY DETERMINATIONS AT ORNL AND NBS FOR NuF-ZrF -UF (53-43-4 mole %) Enrfh‘alpy (cal/g) Temperature . ORNL ("C) NBS ? Bunsen Copper Block 600 182 178.9 178.9 700 207 205.4 205.3 800 ‘ 232 232.0 231.1 900 257 258.5 256.0 In these expressions H is the enthalpy in cal/g, c, is the heat capacity in cal/g-°C, and ¢ is the temperature in °C. | Viscosity S.‘-I. Cohen T. N. Jones Reac'ror‘Experimen'ral Engineering Division ‘The program of viscometry technique refinement previously discussed® is under way, and measure- ments have been made on one mixture. Studies will be made on others as samples become avail- able. A number of changes have been made in the viscosity apparatus. The length of the Brookfield spindle has been doubled, which allows the speed of rotation of the instrument to be halved without reducing the shear force. This reduction in speed assures the existence of laminar flow. Salts are contained in new tubes with a smaller diameter and consequenfiy a smaller clearonce between the 55. I. Cohen, ANP Quar. Prog. Rep. Sept. 10, 1954, CORNL-1771, p 127. handling of the instruments and improve atmos- pheric control. New calibrating liquids which have densities comparable to the densities of the fluorides are being used for both viscometers. Among these liquids are low-melting-point, easily handled, fused salt mixtures and heavy organic- halogen compounds. Recent measurements made on NaF-ZrF,-UF, (53.5-40-6.5 mole %) are presented in Fig. 8.4. It may be noted that the average values obtained from the recent measurements are about 30% lower than the preliminary values given previously.’ However, when the preliminary measurements were reported, only the more pessimistic, Brookfield values were given. It can also be seen from Fig. 8.4 that the agreement between the two instru- ments for the recent measurements is much better than for the earlier set of measurements. This result is due to improved fluoride preparation and handling techniques, as well as to the refined viscometry methods. Thermol Conduchwty W D Powers R. M. Burnett S. J. Claiborne Reactor Experimen’ral Engmeenng Duvnsuon The\ ’rhermai conduc'rsvn‘y “of NaF- ZrF -UF (53.5-40- 6 5 mole %) in the hqu:d ‘state was found f° be 1.2 BfU/hr-ff~°F at an average sample temper- o "afure of_666°C The varmble -gap devnce was Used T_he conduchvr’ry of ) V‘T-i‘hls salt ‘mixture had’ prewously been estimated box to be used solely for wscosn‘y work is bemg"“ “fo be 1.3 Btu/hr-#+-°F on the basis of an empirical Cbuilt, This dry box should further facilitafe 65.. {. Cohen and T. N, lJones, Preliminary Measure- ments of the Viscosity of Fluoride Mixture No. 44, ORNL CF-53-8-217 {Aug, 31, 1953). 117 L} _ oy o o Ty} Fenh i ANP QUARTERLY PROGRESS REPORT STAINLESS STEEL MELT TUBE FABRICATED FROM SCH 40 1-in.-IPS PIPE CHROMEL-ALUMEL THERMOCOUPLES (4) UNCLASSIFIED ORNL-LR-DWG 4454 BROWN MULTIPOINT RECORDER 7 . =3 SIMPLITROL COPPER LINER BRONZED / : WITH ALUMINUM ——— | . : , . 71 RELAY o N FURNACE ELEMENT —0 | / 4P VARIAC [ 110y . | % 10 /] 1o R . ':g CHROMEL-ALUMEL ]NSULAT|ON // g O CONTROLLING COUPLES it = e // 3‘ ; O % A Ok / T | SIMPLITROL A e Z o J - o POSITION OF THERMOCOUPLE Z / RELAY WELLS IN MELT TUBE WALL ? A% (CHROMEL -ALUMEL COUPLES v /,O IN 0.08-in.- OD STAINLESS . v S STEEL WELLS) / A5 VARIAGC HO v / g ® o Fig. 8.3. Furnace for Yiscosity Measurements. conductivity relation developed for the fluoride mixtures previously studied. The solid and liquid thermal conductivity meas- urements of NaF-KF-LiF (11.5-42-46.5 mole %) and I\lcl"'-KF«LiF--UF4 (10.9-43.5-44.5-1.1 mole %) were reported in part in previous quarterlies. All con- _ducrtivirty ‘data obtained with several types of ““conductivity devices for these two fluoride mix- ‘tures are shown in Fig. 8.5, The liquid and solid ~ conductivity values of the ternary fluoride mixture are nearly the same, and the same is true of the ‘quaternary. 118 A longitudinal thermal conductivity apparatus has been designed and constructed for the meas- urement of the conductivity of metallic beryllium, It is essentially the same apparatus as that re- ported previously’ and is contained in an inert- gas-filled box to protect personnel from beryllium poisoning. 7W. D. Powers, S. J. Claiborne, and R. M, Bumett, ANP Quar. Prog. Rep. June 10, 1953, ORNL-1556, p 86. Pl T37F 9 ™ .-if’,g::ié x_ff:) s P S, #w i Ny, - Sy “ [y = Ay %, 3 SEGRE ORNL-LR-DWG 4455 TEMPERATURE {°K) 600 700 800 1000 1200 100 ‘ | CUP | BROOKFIELD | AVERAGE 50 PRELIMINARY MEASUREMENT MADE AUG. 1853 MEASUREMENT MADE AUG. 1954 o a —_————— * A & (CENTIPOISES) - ™) o C o 300 400 500 700 1000 TEMPERATURE (°C) Frig. 8.4. Vi'scos'ify of NaF-ZrF“-UflF.; (53.5- 40"6.5 m°|e %)o PERIOD ENDING DECEMBER 10, 1954 O ORNL—-LR—DWG 4456 4.00 so1_1o——4—'———- LIQUID —— ] o o 300 ™ £ & . . A = o < A u e 2 .o m - ™~ = —— O v S 5 N \ > "0 . _ 5 £ 2.00 A ~o <3 o ~ ~ 2 ~ 2 ¥ Z ~ o - a Z L:EJ 1.00— METHOD NoF -KF-LiF NcF—KF—LiF—UF,; = ' {11.5-42-46.5 mole %) (10.9-43.5-44.5~11 mole %) - COOLING SPHERE A A (S0LID) FLAT PLATE n o (SOLID) VARIABLE GAP » o {LIQUID} - I | o r 0 500 1000 1500 TEMPERATURE (°F} _' Flg. 8.5. Th-e‘rlfiél;Cofi;h;c-ti‘v'i.fies of NaF-KF-LiF and NaF-KF-LiF-UF, in the Liquid and Solid States. Fo 2 119 :'?lvncf‘s sy ANP QUARTERLY PROGRESS REPORT 9. RADIATION DAMAGE J. B. Trice Solid State Division Ac Jo Mi”ef ANP Project - The program of MTR irradiations of Inconel capsules containing fluoride fuels has continued, and the capsules examined thus far have revealed no evidence of radiation damage. The horizontal type of fuel-circulating loop designed for irradiation in the LITR has been operated out-of-pile with a non-uranium-bearing salt and is now being inserted in the HB-2 facility of the LITR, A small loop suit- able for vertical operation in the LITR lattice has also been successfully bench tested, and a second mode] for in-pile operation is being constructed. The creep test apparatus for testing Inconel at high temperatures in the MTR is being bench tested. The stress-corrosion apparatus for LITR “operation has been successfully bench tested, and an in-pile apparatus is being constructed, Remote metallographic studies of solid fuel elements were continued, and additional information on the relationship between UO, particle size and radiation damage was obtained. MTR STATIC CORROSION TESTS W. E. Browning G. W. Keilholtz Solid State Division Ha Lr Hemphi” Analytical Chemistry Division The program of MTR irradiations of Inconel capsules containing fused fluoride fuels has been continued. Additional irradiations were carried out on capsules containing NaF-ZrF ,-UF, (48.9-49.3- 1.79 mole %) and on capsules containing NaF- ZrF,-UF, (50.1-48.2-1.74 mole %) These fuels generate 1100 w/cm® in the A-38 position in the MTR. To date, five capsules in this series, three with UF, and two with UF, fuels, have been " successfully irradiated for a two-week period with metal-liquid interface temperatures of 1500 £ 50°F, The capsules have been examined metallographi- cally, and the fuels have been chemically analyzed. One out-of-pile electrically heated control capsule containing UF, and one containing UF, have also been examined, and the fuel batches have been chemically analyzed. 120 The irradiated UF3-bearing CO.pSUIé‘.S and one of the irradiated UF ; capsules showed practically no corrosion — that is, they were similar to the control capsules — while one irradiated UF ,-bearing cap- sule, previously reported,! had subsurface voids to a depth of 2 mils. There were no significant dif- ferences in the iron, chromium, or nickel contents of the irradiated fuels, as compared with the starting fuel batches, and there was no evidence of segrega- tion of either uranium or impurities. The chemical analyses of the two out-of-pile controls that have been examined to date were accidentally spoiled, The uranium in the UF -bearing fuel analyzed not less than 96% ,UF3, and the UF4-beuring fuel showed no trivalent vranium. In an attempt to de- tect any minute radiation effects which might be occurring, the irradiation period is being extended to six weeks for the other capsules in this series, Additional examinations were made of capsules containing NaF-ZrF4-UF4 (50-46-4 mole %) that had been irradiated at a nominal temperature of 1620°F and had generated 2700 w/cm®. In this series of irradiations, the in-pile temperature history was quite complex. No evidence of chemi- cal damage could be found in the fuel, and there were no high concentrations of Inconel components. There was some corrosion evident, and therefore out-of-pile control tests will be made in which the in-pile temperature patterns will be duplicated insofar as possible. New experiments with this fuel are to be made in which more closely controlied temperatures and planned temperature excursions will be used, The chemical analyses of samples of fuel taken from irradiated Inconel capsules have at times shown increases in iron such that the iron-to- chromium ratio was greater than that found in the Inconel, and therefore a series of radioactivation analyses of the iron content in the irradiated fuel was made; these analyses demonstrated that the high 'values were due to contamination of the fuel w. E. Browning and G. W. Keilholtz, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 134, i ot ot 0y with unirradiated iron that had entered the fuel during or after opening of the Inconel capsule. The techniques developed by G. Smith (Analytical Chemistry Division) are particularly adaptable to the separation of minute amounts of iron (10 to 100 pg) from very large amounts of fission products. Four samples of fuel from irradiated capsules showed that between 70 and 99% of the iron came from an unirradiated source, whereas samples from an irradiated Inconel copsule showed that less than 6% of its iron content was not radioactive, Since the amount of iron contamination involved is very small (of the order of 50 ug), some possible sources of contamination would include the steel drill used to remove the fuel from the capsules and also the lmpurlhes in the reagenfs used in preparing sample solutions. In the future, samples will be obtained by using a nonferrous drill. Unfortunately, radio- activation techniques cannot be applied to analyses for chromium and ‘nickel because of unfavorable activation yields, insensitivity of detection, and chemical separation difficulties associated with trace amounts of 'rhese elements. An improved and modified version of the capsule irradiation facility is being put into service (Fig. 9.1). Before being shipped from ORNL, each capsule is mounted in a sleeve which provides a sufficiently large annulus for the passage of cooling air between it and the capsule. The bottom of the AIR PATH [NSTALLATION OF IRRADIATION TUBES ToTAN MTR REFI_ECTOR "A" PIECE PERIOD ENDING DECEMBER 10, 1954 sleeve is tapered to fit an inside taper at the bottom of the air tube in the MTR. The advantage of this arrangement is that the capsule can be fitted to the sleeve prior to insertion into the reactor, Several tubes and capsules prepared with this new arrangement have been used satisfactorily in the MTR. In order to speed up the capsule program, arrangements have been made for irradiating two capsules simultaneously in separate, but adjacent, facilities in the MTR. An improved set of temperature controls fabricated by the Instrumentation and Controls Division has been subjected to dynamic performance fests by using an electrically heated capsule. These tests have shown that the new controls will maintain a more constant temperature in the capsule during irradiation, EFFECT OF IRRADIATION ON UF ,.C,F,, W. E. Browning - G. W. Keilholiz Solid State Division Exommahons were made of material from welded nickel capsules that were filled under vacuum with a UF -C,F, ( solution containing 20 wt % UF , and then irradiated in the ORNL Graphite Reactor and in the Tower Shielding Facility. These irradiations were made to determine the suitability of UF - C,F,¢ for use as fuel in a Lid Tank Shielding Facility mockup of the circulating-fuel reactor, - epeme ORNL-LR-DWG 4603 ' %\\\\\\\ X'iwrmcm_ THERMOCOUPL7 / . iCAPSULE , ELLLLLI SIS IL SIS ///11////1//// IRRADIATION TUBE WITH CAPSULE INSTALLED ANNULUS SLEEVE 7 T T r \\\\\\\\\\\\\\\ YL IIII/)II}II/IIIIII L7 PP LA T >I\>>\ fRRADlATION TUBE IRRADIATION TUBE WITH CAPSULE REMOVED Fig. 9.1. New Apparatus for Insertion of Capsu|és in the MTR. 121 ANP QUARTERLY PROGRESS REPORT Three of the capsules were exposed to 1.1 x 10717 -"nvt'f'h.‘ Gray-green solid residues which filled as much as 90% of the volume were found in all the “capsules. The residues in the capsules exposed . to a flux of 1.1 x 10'7 analyzed 92% uranium, and ~the residues in those exposed to a flux of 2.0 x , ]0]5 analyzed 24% uranium. In the capsules given ' fhe higher exposure, gas pressures of between 55 ‘and 100 psi developed. Analyses of these gases ~showed about 30% CF, and about 25% C,F; the * remaining gas was unidentified, 2" ¢ ' - MIN}ATUR‘E IN-PILE LOOP "~ Bench Test W, R.'Willi_s M. F. Osborne H. E. Robertson G. W, Keitholtz v Sd‘lird___State Division The first successful bench test of the miniature in-pile loop designed for insertion in a vertical __hole in the LITR included four freezing and melting cycles in 260 hr of operation. The loop was oper- ated at 1466°F, The linear velocity of the fused salt circulated in the loop was between 3.3 fps (as calculated from the flowmeter) and 3.8 fps (as calculated from the pump speed); thus a Reynolds number of about 3000 was obtained, Since it was possible to freeze and melt the fused salt without causing failure of the loop, it may be advisable to fill the in-pile loop before it is inserted in the reactor and then melt the fuel mixture after the loop is in position. Since, during the bench test, the flowmeter was found to be temperature sensitive, the dependence of the measurement on the temper- ature must be established.’ The bench test was terminated because of a leak in a collar that was welded over a joint in the fuel tube. The two ends of the loop did not butt together “at this location, and the annular space thus formed between the collar and the tube sections trapped the fused salt in such a manner that expansion of the salt caused the collar (not the weld) to rupture. “The metal collar stretched from 0.4 in, in diameter to 0.5 in. before it ruptured. The salt from the " leak, which existed during operation for about 4 “hr, showed a tendency to oxidize and to stick to - the Inconel surface rather than to flow rapidly down ~‘the outside tube wall. Thus, a leak in the colder _portion of a loop that was operating in-pile would ~“ " probably not flow downward into the high-flux region before a safety alarm from the released “radioactivity could scram the reactor, A Delco motor from the group being used in bench tests has been rebuilt to withstand radiation and a higher temperature, The wire on both the rotor and the stator was replaced with glass-insulated wire. All paper was removed, and mica was used in the commutator and in the rotor segments; glass cloth was used, where possible, for other insula- tion. These changes required that the shaft as- sembly be slightly modified to allow more space for windings. The rebuilt motor was tested during the bench test of the miniature loop, and it was found to be satisfactory. In a comparison of the operation of the rebuilt motor with that of the original motor under no load and under load cond i-fio'ns,' the rebuilt motor was found to be slightly more efficient than the original motor. Furthermore, since the rebuilt motor can withstand a higher ambient temperature and since for a given voltage it produces a higher speed, higher fuel velocities can be obtained than were possible previously, Heat Transfer Calculations M. T. Robinson Solid State Division E. R. Mann F. P. Green R. S. Stone Instrumentation and Controls Division D. F. Weekes, Consultant An extensive series of heat transfer calculations has been carried out on the ORNL. Reactor Controls Computer in order to predict the thermal behavior of a miniature in-pile circulating-fuel loop. The derivations of appropriate differential equations and the details of their solution are given in a forthcoming report,? The model finally adopted for the calculations and for the in-pile loop is shown in Fig. 9.2. It was assumed to be mounted in position C-48 of the LITR, with the lower end at the location of maximum thermal-neutron flux.® The maximum power density in the fuel in position C-48 was assumed to be 540 w/cm?3. The computer results were obtained for a variety of different flow rates of fuel (NaF-ZrF -UF,, 53.5-40-6.5 mole %) and of cooling air expressed as Reynolds number of the cooling air stream, Re , or fuel stream, Ref. The heat transfer, y, that is, 2\. T. Robinson and D. F. Weekes, Design Calcula- tion for Miniature Hifib Temperature In-Pile Circulating Fuel Loop, ORNL-1808 {in press). M. T. Robinson, Solid State Semiann. Prog. Rep. Feb. 28, 1954, ORNL-1677, p 27. (e n L1 W gy r 'i{'4 the amount of heat flowing radially through a unit length of the loop, vs the distance, s, from the inlet is shown in Fig. 9. 3 and the devmtlon, . " UNCLASSIFIED - .S5D-B-988 ORNL-LR-DWG 22524 FUEL INLET FUEL OUTLET ' AIR INLET i [ ol / 1 i FUEL LOOP ’ i 1 ¢ \ ] AIR COOLING ’ * ’l ’* ANNULUS g 1N ”‘ ' AIR OUTLET F lg. -- 9.2. ‘Miniature In-Pile 'Clrculuhng-Fuel“ _Loop Model Used for Desngn Culculahon . PERIOD ENDING DECEMBER 10, 1954 of the mixed-mean fuel temperature from its average value vs s is shown in Fig. 9.4. These results meet only. one of the three required boundary con- ditions, namely, that all fission heat be removed into the cooling air stream. The other conditions, that the maximum fuel temperature be 815°C and ~ that the initial air temperature be 30°C, are met by suitable interpolation of the calculated data, Some of the results of the calculations are summa- rized in Table 9.1. SESFEF ORNL—LR—-DWG 37824 60 55 ‘ : A Re, / k 10,000 _ /] 0\ 20,000 — ' 7 N 30,000 2 € 45 AN ) 4/ S \ 60,000 ————— 7,/ z 3 ' | L 40 N\ Vs 5 4 w) . = g J = ' : . : Y T 30 Af— " \\ / / 25 N / /' | \\\\\J/ /| 20 . A 15 : O 20 40 60 80 1025 125 145 165 185 205 s, DISTANCE FROM INLET (cm) Fig. 9.3. Heut Transfer in Miniature In-Plle -7;1:00p for [ Fuel ReY"°|d5 Number of 3000 as @ _F‘uncflon of Dlstunce from fhe lnlet und the ' ’;leferenhol ~Cay Temperature T R Air Flow © (scfm) ’ Te mperat ure {OC)_ - ) “ 6 000 D 15,5b0 R A S e T e 25 ~5;IOV e T R T 20 ST T 10,000 15,500 25 550 20 123 ANP QUARTERLY PROGRESS REPORT SEereT ORNL-LR—DWG 3783A 58 50 45 \ 40 35 20 / o 25 | / . \ 3 ' : I // \\\ 4 2 6000 j X Re, = ] M . 7 {40,000 J’l \ \ ; _ o 15 II \ w - \ 5 10 'I / \\h \‘ \ with Cs'3% and Ce'4] will give a measure of the loss of the important alkali and rare-earth elements through adsorption on metal surfaces, since Cs'3¢ and rare-earth elements will be present only in trace quantities but chemical zirconium will be present in very large quantities, as mentioned above, The data obtained from the ARE fuel sample will be compared with similar data to be obtained from an irradiated sample of NaF-ZrF4-UF4(50-46-4 mole %) that will be maintained solid to retain all the fission products. The cooling period for the irradiated sample will be the same as that for the ARE fuel sample. LITR HORIZONTAL-BEAM-HOLE FLUORIDE-.FUEL LOOP J. G, Morgan M. T. Morgan O. Sisman W, E. Brundage C. D. Baumann A. S. Olson R M Carroll W. W, Purkmson T Solid Stcte DlVISIon ";‘The vs-econ 'floop fcbrlccted f;)r c1rcu|at|ng flyo- ride fuel in the LITR has, been opercn‘ed with a _non-uranlum-bearmg fused sah‘ msxfure for 6 hr in “the Eaborotory and is bemg mserted in the HB-2 After the |oop ‘has been facility of fhe LITR .“ci:'omplefely installed, cnrcu!aflon “of the barren fused salf\wnll ‘be _resumed. W : ' ‘operahon has been obtdlfié‘d, ‘the bcrr-e(n mixture ‘will be drained and the loop will be filled with B atis facfory vTi’ne”desxgn and calculations for the fue T, SRR} R AT e and Physical Properties’ (this report). The curve obtained for pressure of air vs heat removal gives 125 C ] ¢=| fo-alr.; !\hea'r exchdnger used in this loop were substantiated T T Lt e o b ke iE ook gk “canned for insertion in the LITR. AVN'P’ 'OUAR‘TERLY PROGRESS REPORT _.3__‘20 kw at a pressure drop of 50 Ib, 10 kw at 5 Ib, - and 5 kw at approximately 1 Ib. Since the maximum _ _lrhedt removal requ:red is expected to be no more " than 10 kw, the heat exchanger appears to be more : thcm odequate. CREEP AND STRESS-CORROSION TESTS W W Dav:s J. C. Wilson N. E. Hinkle J. C. Zukas ~ Solid State Division " The ‘stress-corrosion apparatus described pre- i V-IZVIidt'Js;lys has been bench tested. The design ap- ‘pears to be sound and therefore a new rig is being Temperature control to within +2°F at 1500°F has been achieved by using a Speedomax air-controller to drive the Variac that supplies the furnace power. Operation _ _'.ové'r"f)'_-e_ridds_ of 500 hr has shown that, if sodium_ " distills out of the hot zone of the furnace, so little ‘-_sodiu_m is iost fhat ifs effectiveness as a heat SW. W. Davis et al.,, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL ]77], p 142, F;g. 9.5. Sandwmh Type UO ,—Stainless Steel Fuel Element After 29% Burnup in MTR (UO Parhcle Size, 3u). 250X. 126 transfer agent is not diminishe_d".r Welding of some joints has been poor, and therefore Dy-Chek is being used in the inspection of all new welds, The LITR Experiment Review Committee has ap- proved the apparatus for irradiation, and the necessary safety circuit for the LITR has been completed, In the bench tests the oufsrde of the spec;men was in air, rather than in sodium as it would be in the in-pile apparatus. The tests were for periods of 432 and 866 hr at 1500°F and a stress of 1500 psi. The specimen exposed for the longer time showed roughly twice as many voids per unit area on the sides of the tube stressed in tension and compression as there were on the sides at zero stress. The maximum depth of voids was about 0.001 in. it was also noticed that the etching (or perhaps staining) characteristics of the salt-metadl interface were not the same for the stressed and unstressed regions when etched in aqua regia, The reproducibility of the creep data obtained was not good, and therefore the specimen shape has been modified in order to concentrate more of the ,. £ » ERY It iR IS a . o i G, - S > creep deformation in the gage length. A specimen tube with thicker walls will be used, and the wall thickness will be ‘r'\eduééd only at that part of the tube which is in contact with the fused salt, This modification should greatly increase the accuracy of creep measurements, The MTR creep test equipment is being bench tested prior to shipment to NRTS, REMOTE METALLOGRAPHY M. J. Feldman R. N. Ramsey W. B. Parsley A. E. Richt Solid State Division Two additional Pratt & Whitney Aircraft capsules, each containing sandwich-type UO,—stainless steel fuel elements, were opened, and the elements were examined metallographically, The stainless steel— uo, particle size was less than 3y for one capsule, and it was between 15 and 44y for the other. The capsules were irradiated at a temperature of about 500°F, and the total burnup for each capsule was 29%. Comparison of the results from the element PERIOD ENDING DECEMBER 10, 1954 with 3-u particle size and the element with 15- to 44-u particle size showed that irradiation induced greater final hardness in the elements with the smaller-pdrticle-size material (Figs. 9.5 and 9.6). Four samples from the smaller-particle-size elements showed cracking of the core without bending. Since cracking did not occur in elements with lower total burnups, it appears that the increased burnup (29%), together with the small particle size, can cause cracking of the core, as illustrated in Fig. 9.5. MASS SPECTROGRAPHIC ANALYSES R. Baldock Stable Isotope Research and Production Division The isotope dilution method has been used during the last one and one-half years for analyzing ARE- type (NaF-ZrF ,-UF,) fuels for their uranium con- tent.% As a result of this experience, a need was 5.. o. Gilpatrick and J. R. Sites, Stable Isctope Re- search and Production Semiann. Prog. Rep. May 20, 1954, ORNL-1732, p 24, UNCLASSIFIED _RMG- Fig. 9.6. Sandwich-Type UO,-Stainless Steel Fuel Element After 29% Burnup in MTR and o Bend Test (UO2 Particle Size, 15 to 44y). 250X, 127 T Mo ™M Mgl ey L bk PR \ i R ' ANP QUARTERLY PROGRESS REPORT felt for further s’rudy of the accuracy of the method. A testing program was inaugurated in which an " enriched sample of U,0p containing 99.86 at. % U235 gnd 0.07 at, % U238, obtained" from B. Harmatz,” was used as the starting material. This material was chemically purified for use as a standard, and the isotope ratio was carefully de- termined. The material was then dissolved in ) HNO3 and diluted to make a standard stock solution, Small portions of this solution were put through the normal ether extraction process used for extracting ‘ uraniuvm from ARE -type fuel samples. Any change ) ?“of |sotop|c composmon was attributed to contami- nation by natural uranium present in reagents and glassware. The observed uranium contamination in these blanks wos less than 0.5 pg per sample, cnd ‘thus it appears “that the normal uranium con- ‘ fcmlncmon found durmg chemlccl separchons of _ B Horma'rz, H. C McCurdy, and F. N, Case, Catalog of Uranium, Thorium, and Plutonium Isotopes, ORNL- 1724, p 2 (May 19, ]954) fuel uranium is neg||g|ble for the usuol fue[ samples which exceed 500 ug of uranium. The second test of the isotopic dilution method verified the concentration of the standard solutions, the accuracy of the spiking, and the reliability of the mass spectrometer. Known mixturés of normal and enriched uranium solutions were examined in the mass spectrometer, and the results were com- pared with the expected U233 concentrations. The greatest difference between calculated percentages and measured percentages was only 0.65%, and the average difference was 0. 37%. | The results of these studies show that the isotopic dilution method is reliable and that good results can be obtained under the conditions which have been used for ARE-type fuel analyses. The method of determining burnup based on measure- ments of the U236 grown in is less subject to error from sample contamination than is the method based on measurements of U%38 and is therefore considered to give the more reliable data. ¥ . LA AR e s i R v ol . S . R it S B e ok i Bk, slscRin, bk il o Ol gy e e # cuhey fond the volume of evolved hydrogen |s measured @ L : ; o 10. ANALYTICAL STUDIES OF REACTOR MATERIALS - j"’ ;o PERIOD ENDING DECEMBER 10, 1954 "oy S vy e 15 -.Wr—’ Y o & C. D. Susano Analytical Chemistry Division J. M. Warde ,Metollurgy Division " The research efforf in analyhcal chemlstry was concentrated primarily on NaF-KF-LiF-base reactor fuel, and, in particular, on analysis after it had been utilized as- a solvent for the reduction of UF, by U° to UF,. Tentative methods for the determination of U° and UF, were developed. Studies were continued on the determination of oxygen as oxides in fluoride fuels. Development work was completed on the determination of sulfur in various reactor fuels and coolants and in off- gases from the production of fluoride fuels. Be- cause of the urgency of other activities, no further work was done onthe determination of alkali metals in NaF-KF-LiF-base fuels.! ANALYTICAL CHEMISTRY OF REACTOR MATERIALS , J. C. White Analytical Chemistry Division Deferm:nutlon of Uranium Metal in Fluorlde Sal't Mleures A S Meyer, Jr. B. L. McDowell ~ Analytical Chemlsfry Division A method based on the measurement of the hydrogen derived from the decomposition of UH3 was developed for the determination of U° in proposed reactor fuels. is converted fo UH by heo’rmg in an atmo _phere’:' is de- of hydrogen at 250°C for 1 hr. "The UH composed by hea'rmg ina stream ‘of CO at 400°C, bY the SOIUfronéof KOH{gkzs produced by ‘rhe reduc-:‘ B 8 lhon °f CO to CO." The’ CO is reduced at femper*’”.% :tweenz400 and 600 OC by both U® and UF,. ). C. White, G. Goldberg, and B. L. McDowell, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 152. In this defermmahon,r UO‘ ‘osphere niodified procedure that is being considered, UH, “'will be decomposed in an atmosphere of c:mmoma ofH SO " "Further’ developmg a titrimetric mefhod for the defermlno-' _, tion of trivalent uranium in the presence of tetrava- “lent uranium in Hluoride salt” mixfures. Under appro- priate conditions, trivalent uranium is selectively The volumes of gas derived from reactions 1 and 2 are reproducible; 1.5 moles of CO is formed for each mole of UH, decomposed when the ignition is carried out at temperatures between 500 and 600°C, and 0.5 mole of CO is formed per mole of UF, under similar conditions. The CO must be removed from the effluent gases before the meas- urement of hydrogen is made. This is accomplished by passing the gases through a tube packed with ] O and powdered pumice, which is maintained at a temperoture of 150°C. The CO is oxidized to CO, in the following reaction: 5C0 + 1,05 —> 5C0, + |, When the method was testedby analyzing samples of pure uranium metal, the hydrogen evolved corresponded to 97% of the theoretical value with a coefficient of variation of 2%. Satisfactory precision has also been obtained for the deter- mination of U° in NaF-KF-LiF-base fuels and in UF,. The concentration limits of the methcds have not been measured experimentally, but, on the basis of the quantities measured, concentrations of U® as low as 0.05% may be determined when 1-g samples are utilized. Since the [,0. tubes may become inactivated after only limited service, a more dependable method of eliminating the CO is needed. In a and the hydrogen will be measured over a solution Determmahon of Trlvaleni Urumum | - in Fluoride Fuels “A. S. Meyer, Jr. W, J. Ross - D. L. Manning B. L. McDowell : Analyhcol Chemistry Division mvestlgahons were carried ouf for oxidized by cupric chloride, CuClz, titanium tetra- 129 A WA é»m? 1.3 $ A ANP QUARTERLY PROGRESS REPORT chloride, TiCl,, and methylene blue, C, ¢HygN5SCHL Of these reagents, methylene blue appears to be the most promising reagent for a possible routine method for the determination. Cupric Chloride (CuCl,). Previous attempts to oxidize UF; by Cu(ll) in H,SO, solution were un- successful. Since the formal oxidation potential of the [Cu(H),Cu(l})] couple is increased in solutions of high chloride concentration, experiments were performed to ascertain whether quantitative oxida- fion of trivalent uranium could be achieved by ~ adjusting the acidity and chloride concentration of the CuCl, solvent solution. j Sainples of UF, were dissolved under an atmos- here of CO in acidic solutions of NaCl which confcmed a mecsured excess of a standard solu- tion of CUC| Affer dissolution of the samples, the unreduced CuCl, was determined iodometrically and 1rhe trlvalent urcmlum was calculated on the bc15|s of the stmchlometry of the equation Cu2* 1 U3t —> Cut + U4t ‘When the dissolution was carried out in a solu- tion 1 M in HCl and 2 M in NaCl, the results were in agreement with those obtained by the hydrogen evolution method.?2 Lower values were obtained when either the chloride concentration or the acidity was altered. Samples of UF3 which had been fused with NaF- KF-LiF evolved hydrogen and subsequently yielded low results when analyzed by the above- described procedure. The method therefore appears to be of limited applicability. Titanium Tetrachloride (TiCl,). It was found that UF, dissolved without evolution of hydrogen in solutions of TiCl, in concentrated HCl to yield dark-brown solutions of TiCl.. The color of these ‘solutions, when diluted, reverts to the rose tint usually associated with Ti(lll) solution. When these diluted solutions are titrated with a standard ~solution of K,Cr,0,, the equivalents which are _consumed at the first sharp change in potential correspond to about 95% of the trivalent uranium as determined by the hydrogen evolution method. If the solutions are then heated to a temperature " of 90°C, the total uranium can be determined by 'ox1d|zmg U(lY) to U(VI) with furfher addition of 'rhe dlchromcn‘e soluhon. Bl e el R, S 2D. L. Mahmng, W. K. Miller, and R. Rowan, Jr., Methods of Determination of Uranium Trifluoride, ORNL- 1279 (Apr. 25, 1952) . 1éo"r While this is the only method which has been found to give a titration for trivalent uranium without subsequent back-titration of excess oxidant, it is not readily adaptable to routine analysis because of the slow equilibrium of the electrode potentials and the rapid oxidation of titanous solution by air after the dilution of the concentrated acid solutions. Experiments are now being conducted to determine whether the TiCl, can be titrated in the concentrated acid with bromine which is generated coulometrically. End points obtained potentiometrically in the concentrated solutions were found to be poorly defined, but sharp breaks with potential changes of about 400 mv were obtained: with polarized platinum electrodes. Titrations between 5 and 10% in excess of the theoretical value were obtained when UF, samples were titrated cou- lometrically. This excess titration is probably a result of diffusion of TiCl, from the cathode com- partment. ‘ Methylene Blue. A proposed method based on the oxidation of trivalent uranium to the tetravalent state by methylene blue was developed for the determination of UF, in a fluoride fuel. In the proposed procedure, the sample is dissolved in @ measured volume of 0.02 N methylene blue solution in 3 to 6 N HCl| under an atmosphere of CO, by stirring for 2 hr at room temperature., The excess methylene blue is then titrated to its reduced state, methylene white, with 0.05 N chromous sulfate solution. The equations involved in the determination are as follows: 2U3* + methylene blue + 2HY — 204" + methylene white 2Cr?* + methylene blue + 2H"—> 2Cr3% + methylene white Because of the intense color of the dye, the visual end point, blue to green, is sharp and re- producible even when the titrations are carried out with 0.025 N chromous sulfate. For the determination of trivalent uranium in UF,, the coefficient of variation is approximately 1%, and the results are in excellent agreement with those obtained by the hydrogen evolution method.? While somewhat poorer precision was obtained in the analyses of NaF-LiF-KF-UF,.-UF, samples, the variations were probably a result of hetero- geneous sampling, as is also indicated by the 0627 134 N L e vader * lrreprodumb:llty of determinations by the hydrogen evolution method. In the ona[yses of some of the NaZrF .-UF samples, low results were obtained becouse fhe samples were not completefy dissolved in 6 N HCI. ~ The methylene blue procedure requires less operational time than the hydrogen evolution method, and, for the samples which have been ‘tested, it appears to give results of comparable precision and accuracy. Tests are being conducted to determine the extent of the reactions between methylene blue and uranium hydride, uranium metal, and other metalllc contaminants of the fluoride samp]es. ‘In order to carry out subsequent determinations of the total uranium concentration on the same samples, the final soluhons from the methylene _‘blue determlnotlon were titrated with oxidizing agents. Titrations with Ce(lV) and Fe(!H) solu- tions did not yield stoichiometric end points. Al_'t»ht_:ug'h'kfhe reaction with K_Cr 07'v'vds'exfreme|y slow, two well-defined potentiometric end points were obtained when the solutions were titrated slowly. The second of these end points corre- sponded to the reoxidation of methylene white to methylene blue plus the oxidation of the U(IV) to U(V!). The stoichiometry of the first break in the “potential curve has not yet been accurately de- fined, but it appears to be consistent with the oxidation of U(IV) to U(V). Further investigation is under way to elucidate the nature of these changes in potential. Determlnohon of Oxygen in Fluonde Fuels ,.A S, Meyer Jr.»{ 4 M Peele 3A. S, Meyer, Jr., and J. M. Peele, ANP Quar. Prog. Rep, Sept. 10, 1954, ORNL-1771, p 148. ~metallic oxides. wafer alcohol mixfure. PERIOD ENDING DECEMBER 10, 1954 is used, a concentration of 1% water corresponds to 235 mg of oxygen. From the above linear relation the calibration curve can be extended to concentration ranges too low for direct measure- ment by weighed primary standards. At the temperatures at which liquid HF could be safely maintained in the reaction vessel, the rate of the reaction between metallic oxides such as uo, and HF was found to be too slow for applica- ’rlon to analytical measurements, The oxides were found to react rapidly with fused KHF, according to the postulated reaction 4KHF , + U0, —> UF,, + 4KF + 2H,0 The procedure has been modified accordingly to carry out the dissolution of the sample in molten KHF . In the revised procedure, the sample is mixed with about five times its weight of anhydrous potassium fluoride, KF, in a platinum crucible, The crucible is then placed in the reaction vessel and the KF is converted to the acid salt by trans- ferring a portion of the HF acid from the conduc- tivity cell to the reaction vessel. After the excess HF has been returned to the cell, the KHF fused by heating the reaction vessel to about 300°C., The water is then transferred to the con- ductivity cell for measurement by repeated distilla- tions with portions of HF. Because part of the KF was carried over to the cell during the distillation, the original apparatus was modified by placing a silver-lined vessel, which is similar in design to a Kjeldahl trap, dlrec’riy above the reaction vessel. The revised ~ procedure is now being tested on pure samples of | ermlnahon of Sulfur J C Whlte G Goidberg Analyhcal Chemustry va15|on Methods were modified ‘and adapted for the deter- 'm‘lnahqn Of 'rraces of sulfur_m ’rhree types of “Qualitative tests with starch-iodide paper were performed to confirm the 0627 138 131 NS PR~ W raere—— pyp—-r g e i 1929). ANP QUARTERLY PROGRESS REPORT presence of sulfur. A colorimetric method? was adapted to determine the concentration of sulfur quantitatively. In this method, sulfur is released from solution as H ,5 by acidification with HCL. The gas is cdsorbed in a solution of zinc acetate solution, and, upon the further addition of ferric 'dm'moriium'Sulfate, a reaction occurs between the organic reagent and sulfur to form methylene blue, an intensely colored dye. The absorbancy of the soluflon of methylene blue is measured with a phofomefer' at 670 my. The method is extremely '4sensmve _wnh a workable range of from 5 to 50 pug per 100 ml of solution. This range corresponds fo a practical lower limit of determination of about 1 part of sulfur per million parts of sodium. Analyses of the sodium revealed that the sulfur was not unn‘orm]y distributed. In particular spots, the concenfrcmon was of the order of 0.1 to 1% or higher. The sulfur content of the sodium sampled from the bulk of the trap was, however, much lower, 50 to 100 ppm, and that of the sodium in the system was less than 1 ppm. These results indicate that sulfur in sodium precipitates rapidly as sodium sulfide and can be effectively trapped and removed from the system. The solubility of sodium sulfide in sodium is evidently extremely small. Sulfur in Flubride Salt Mixtures, The presence of sulfur in mixtures of fluoride salts which are being considered as proposed reactor fuels is con- sidered deleterious, principally from the standpoint of corrosion. The major problem in determining sulfur in fluorides is that sulfur exists in at least two oxidation states, sulfaie and sulfide, and, in order to adapt the methylene blue colorimetric method, the sulfate must be reduced to the sulfide. The reducing mixture recommended by Johnson and Nishita® for sulfate in soils is used for this purpose. About 1 g of the fluoride mix is heated at the boiling point with 4 ml of a reducing mixture composed of 15 g of red phosphorus, 100 ml of hydriodic acid, and 75 ml of formic acid. Reduction of the sulfate is complete within 30 min, and then the colorimetric method can be utilized. When “sulfur is found in fluoride mixtures, the principal | portion of the sulfur is in the form of sulfate rather | than su|f|de as mlgh'r be expected 4. F. Fogo and M. Popowsky, Anal. Chem. 21, 732-4 5C° M. Johnson and H. Nishita, Anal Chem. 24, 736 (]952) 132 Sulfur in HZ-H‘F Gas Streams. The odor of HéS ' has often been noted in gas streams in the fuel production work. Two sources of this sulfur are known: sulfate contaminant in fluoride salts and fluorosulfonic acid, HSO,F, in hydrogen fluoride. Sulfur from the first source is known to be of the order of a few parts per million. A test of the sulfur content of hydrogen fluoride was made by dissolving the gas in a solution of NaOH, deter- mining sulfur in the solution by boiling the basic solution in the presence of H,0,, and precipitating the sulfate as BaSO, with BaCl,. The sulfur concentration was 2.7 mg per liter of HF, A semiquantitative method for determining sulfur as sulfide in off-gas from fuel production was set up in which the gas was passed through a 6% solution of NaOH. An equal volume of bismuth nitrate in glacial acetic acid was added to the scrub solution, and the turbidity, as a result of formation of Bi,S,, was compared with previously prepared standards. The procedure was made more precise by measuring the absorbancy of the turbid solution at 350 my in a 7.5-cm cell with a total volume equal to the volume of the test solution. - The concentration of sulfur found by this pro- cedure ranged from 2 to 35 pug per liter of off-gas. Determination of Fluoride in NaF-KF-LiF-Base Fuels J. C. White B. L. McDowell Analytical Chemistry Division Investigation was continued on the feasibility of a spectrophotometric titration of fluoride based on the decolorization of a zirconium complex or lake, Zirconium alizarin sulfonate and zirconium Erio Chrome cyanine were tested as possible titrants. A titration cell based on the design of Sweetser and Bricker® was fabricated so that the Beckman Model DU spectrophotometer couldbe usedto meas- ure absorbancy. Although the data have not yet been thoroughly evaluated, the technique does not appear to be feasible for application to NaF-KF- LiF-base fuels because of the slowness with which equilibrium is reached. The procedure will be evaluated before further work is done. 5P, B. Sweetser and C. E. Bricker, Anal, Chem. 25, 253 (1953). 0627 135 oW ¥ o Sl . L e b B iR vy [R13 L PETROGRAPHIC INVESTIGATIONS OF FLUORIDE FUELS G. D. White, Metallurgy Division T. N, Mch:y, Consultant Petrographic examinations were made of several hundred samples of fluoride melts. The majority of the samples were from alkali fluoride systems con- taining UF ., | When small amounts of UF, (<20 mole %) are melted with NaF- KF- LiF, KF-NaF, or KF, a red compound is formed which is thought to be K UF,. In melts which contain only KF and UF,, this com- pound is isotropic. However, in the ’rwo with NaF present, the compound exhibits slight anisotropism. This is thought to be due to a slight solubility of NaF in the K sUF (. “Some work was done on the systems KF-LaF,, NaF-LaF,, and RbF- LaF,. The first two systems 7'<:0n’rcm colorless, 1:1 compounds which are uni- axial posmve with refractive mdlces in the wcmn‘y of 1.50. The system RbF- LaF contains a 1:1 compound which is biaxial posmve with a moderate optic angle and refractive lndlces also near 1.50. ANP SERVICE LABORATORY J. C. -Wh"ife' C.R. W||||c1ms ~ W. F. Yaughan Anclynccl Chemistry Division The nonumforml'ry of the NaF-KF-LiF-base fuel samples that have been received in the past necessitated 'a'chcnge in the handling methods. The entire bqfch of fuel resulting from a particular experlmem‘ is now submn‘ted for analys:s. The_w_ ‘-'"“sample |s gro"' ' "'d to pass a No. 50 s:eve quarfered PERIOD ENDING DECEMBER 10, 1954 sampling, and weighing of these hydroscopic samples are carried out in a dry box in order to ensure that the samples are always dry and that the U3* has not changed valence. Since the date of inception of this procedure, more uniform re- sults have been obtained. : fn addition, a change has been made in the method of determining uranium in samples received for analysis. Heretofore, uranium was determined by the zinc reduction, ceric sulfate titration method. The present method? is that based on the reduction with chromous sulfate and final titration of the U(IV) to U(VD) with standard potassium dichromate., This determination is carried out at elevated temperatures with the Beckman Model K automatic titrator. A total of 1,698 samples was received, and 1,587 samples, involving 11,541 determinations, were analyzed and reported. A breakdown of the work load is given in Table 10.,1. 7). s, Decker, Application of Beckman Model K Automatic Titrator to the Determination of Uranium (to be published). TABLE 10.1. SUMMARY OF SERVICE ANALYSES REPORTED _The balance of ~ Tot Number Number - of of Samples Determinations Reactor Chemistry 1,118 8,038 Experimental Engineering 448 3,421 Miscellaneous 21 82 Total 587 0627 137 133 ANP QUARTERLY PROGRESS REPORT 11. RECOVERY AND REPROCESSING OF REACTOR FUEL | D. E. Ferguson G. I. Cathers J. T. Long M. R. Bennett S. H. Stainker W. K. Eister H. E. Goeller R. P. Milford Chemical Technology Division A plant for recovering (in seven batches) the _uranium from the ARE fuel and rinse by the fluoride- volatility process is being designed, and con- ~ struction is scheduled for completion by December 31, 1955, It is estimated that the amount of ma- terial tobe processed will be 12.4 i3 of NaF-ZrF ,- UF , containing 65 kg of uranium, This plant will ‘demonstrate, on a pilot-plant scale, the feasibility -of the fluoride-volatility process as applied to the _ processing of the fuel from a circulating-fuel - aircraft reactor. The feasibility of this process (Fig. 11.1) has been established on a laboratory scale.’*? The basic equipment as now envisioned will consist of a fluorinator, an absorption column packed with NaF, a cold-trap system, and a fluorine disposal unit, The fluoride-volatility process can be adapted for recycling uranium to a fresh fuel concentrate by adding a UF, reduction step and dissolving —-—t D. E. Ferguson et al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 12. ' 2G. . Cathers, Recovery and Decontamination of Uranium from Fused Fluoride Fuels by Fluorination, ORNL-1709 (May 26, 1954). - “Eoner ORNL-LR-DWG 4386 NaF ABSORBENT BED 650°¢C <0.02% U ] / FLUORINATOR 650°C r—— \\\\Q, % ARE FUEL F, WASTE (~ 6 mole % UF, Z IN NaF~ZrF,) ' + COLD TRAP -80°C WASTE SALT 99.95 %, OF U >99% F.P. <0.02% U Fig. 11.1. Fluoride-Yolatility Process for ARE Fuel. 134 the resulting partially decontaminated UFA in NaF-ZrF,. However, if the uranium is to be returned to a diffusion plant or made into heter- ogeneous fuel elements, additional decontamination must be obtained. This may be accomplished by adding a complete UF , fractional distillation step to the procedure shown in Fig. T1.1, This method of uranium recovery and decon- tamination can also be used for processing heter- ogeneous reactor fuel elements of the type that can be dissolved in fused fluoride salt by means of hydrogen fluoride. Compactness of plant, operation at atmospheric pressure, and economical waste disposal are some of the advantages of this type of process, FISSION-PRODUCT REMOVAL Three methods have been tested for removing fission products from the UF ,-F, mixture obtained by fluorination of a wuraniumecontaining fused fluoride salt: scrubbing the gas with a molten salt (NaF-ZrF ), scrubbing the gas with C,F,,, and passing the gas through a solid NaF absorbent bed at650°C. Of the three methods, NaF absorption is the best; it gives an over-all gross beta decon- tamination factor of more than 104, Use of a molten salt scrub, NaF-ZrF , (56-44 mole %) at 650°C, followed by sublimation, for removing fission-product activity from the UF, product of the fluorination step gave, as reported previously,] over-all decontamination factors of 2 % 10% and 430 to 750 for gross beta and ruthenium beta activities, respectively. However, in the scrub step alone, the decontamination factor for the ruthenium, which is by far the most important volatile activity in the long-cooled material, was only 6, as shown by determination of the ruthenium in the molten salt. In a repetition of this experi- ment, the ruthenium decontamination factor in the scrub step was only 2 {run 1, Table 11.1), and the over-all ruthenium decontamination factor was 350. T 6T 128 a i, G o, n T A% ) 19% 5 ¥ PERIOD ENDING DECEMBER 10, 1954 DECONTAMINATION OBTAINED IN THE FLUORIDE-VOLATILITY PROCESS BY SCRUBBING UF, TABLE 111, Synthetic ARE fuel, prepared by hydro-fluorinution of 6 g of irradiated uranium metal in 67 g of NaF- ZrF4 (56+44 mole %), fluorinated at 650°C wn‘h 40- to 80 fold excess fl uorine. _ -R_un L UF6 producf p]us excess fluorme passed ihrough 67 g ' of molten NuF-ZrF at 650°C into a dry-ice trap and then resublimed mto second trap. Runs 2 and 3: UF6 product separated from excess fluorine in dry-ice trap and then volatilized through « C8F16 still into a second trap. Beta Decgntumi nation Factors Activity Run 1 Run 2 Run 3 Over-all Serub* Over-all Scrub* Over-all Scrub* Gross 4800 2.7 290 .2 250 2,2 Ru 350 2.3 16 14 Zr 3.2 x 104 6.8 x 10° 4.7 x 104 Nb 1800 47 1200 86 2 x 108 2 x 108 2 x 10° TRE *Decontamination factor calculated for scrub step alone on basis of activity extracted from scrub material, Scrubbing the UF product from the fluorlnahon and sublimation sfeps with fluorocarbon (C.F ) did not improve the decantamination {runs 2 and 3, Table 11.1). Approximately 9 g of UF, sublimed from a fluorination run, was passed into the bottom of a CSF distillation column (0.5 in. in diameter by 14 in. high, packed with /32-|n. nickel Fenske hehces) that was operctmg at full reflux at about 102°C. The UF6 was taken out of the still as gas’ through an 80°C head. The over-all ruthenium . beta decontammohon factors of 14 and 16 | T in these runs_ s:gh‘[flcanfly grea er‘fhanm the deconfommot‘lon fcctor usuolly obtamed m ’rhe - “In “an explorotory'tr'lul - *solid NaF (Tablg 11,2, 1 (run 1, Table ”, 2), ‘the UF -F stream from the__ 'fluormcl’ror was sed ’rhrough a 25-9 bed of . Y-in. NaF pellets ‘at 300°C.” About 32% of the 8 had to be refluorinated from the NaF bed at 650°C before being trapped. However, the gross decon- “obtained” “Tactivities is probably a filtration effect rather than “runs. “salt waste. uranium was ot absorbed by the NaF, while 38% tamination factors in the two fractions were of the same order, that is, 2 x 10°. The uranium loss on the NaF was only 0.01%. Two experiments were then performed with 30-g beds of 20- to 40-mesh NaF held at 650°C to eliminate the refluorination step (runs 2 and 3, Table 11,2}, The results in both tests showed high absorption of ruthenium and niobium beta activities, and the over-all gross beta decontamination factor was greater than 104, The shght increase ([ess than 10) in decon- “tamination from zrrcomum and total rare-earth on the an absorption effect. The uranium loss _20- to 40-mesh NaF was 0.03%. A poor ‘material balance for ‘ruthenium and niobium “beta actlvrty ‘has been observed in all fluorination In general 1 to 10% of the rufheruum and “about 50% of the niobium remained in the fused After run 3 (Table n 2) the entire nlckel fluormatlon 'vessel was cut into three parts ‘and d:ssolved in dllufe nitric ocui fo obtain the material balance summary “shown in Table 11.3, Nearly all the niobium activity was accounted for in the reactor and fused salt, while 75% of the 135 & S Ty ff} b s bl WP g f 3 w-s ‘uj' oy ., * 'ANP QUARTERLY PROGRESS REPORT TABLE 11,2, DECONTAMINATION OBTAINED IN THE FLUORIDE-VOLATILITY PROCESS BY PASSING UF, THROUGH SOLID NaF ' | Synthetic ARE fuel, prepared by hydrofluorination éf 6g of irradiated uranium metal in 67 g of NaF-ZrF, (56-44 mole %), fluorinated at 650°C., Run 1: UF& product plus 28-fold excess fluorine passed through 30-9 bed of NuF pellets - (/ in.) held at 300°C; 38% (fraction A) was absorbed and then refluorinated off at 650 C into a dry=ice trap, while 32% (fraction B} was not absorbed and was dlrecfly trapped, Run 2: UFéproduct plus 11-fold excess fluorine passed through 30 g of 20+ to 40-mesh NaF held at 650°C and then trapped, <~ Run 3 UF6 product plus 9-fold excess fluorine passed through 30 g of 20- to 40-mesh NaF held at 650°C, trapped, and then resublimed. Beta Decontamination Factors L Run 1 Run 2 o Run 3 - Activity ; : Sl Fraction Fraction i e A B Scrub®* Over-all Scrub* Oversall Scrub* Gross 2200 1600 3.5 2.3 x 104 50 1.6 x 104 24 " Ru 140 120 3.0 5200 90 3100 50 Zr 4300 870 1.6 4.1 x 104 7.3 8.3 x 104 9.4 Nb 430 47 ' 17 3400 280 2500 100 TRE 108 5 x 104 3.5 3 x 10° 7.3 4 x 108 2 *Caleulated for scrub step alone on basis of activity extracted from scrub material, - TABLE 11,3, MATERIAL BALANCE FOR RUTHENIUM AND NIOBIUM ACTIVITY IN A FLUORIDE-YOLATILITY RUN _ Ru 8 Nb 8 Location (% of original) (% of original) Bottom of reactor {in contact with 4 . 0.4 - fused salt) S ;g ...~ Fused salt waste 0.8 69 ' "Tép §f re’achr (not in contact with 8 _ 3___ _fused salt) " Qutlet tube 10 27 : | NqF qbsol;benf 2 . - N .1 _' Total 25 100 , | 3 ]3 . ol . ) ) . o~ oy - , | © 627 140 X, » s mE " i gF » gy o olumma ruthenium was apparently volatilized out of the reactor entirely. Ruthenium and niobium plated out heavily on the metal walls of the reactor. The distribution ratio between metal wall and salt was about 5 for ruthenium activity and less than 0.01 for niobium activity, The behavior of plufomum in the fused salt fluoride-volatility process is important inprocessing fuel that has a high proportion of U238, The results of many fluorination runs at 650°C show that only about 0.01% of the plutonium is carried over with the UF6 product, Further study was carried out on the absorption of UF, in NaF-ZrF, (56-44 mole %) at 650°C. In one trlal 9 g of UF, was absorbed in 30 g of NaF-ZrF , in 1 hr with no noticeable loss. How- ever, passage of helium through the fused salt, ~ either concurrently with or after the UF , addition, produced considerable fuming, A positive test for fluorine with potaséiurh iodide paper was obtained, which indicated reduction of the adsorbed UF to the tetra- or pentavalent form of uranium. A gravimetric test, based on the reaction of fluorine with sodium chloride, showed that the breakdown of UF, to Fé and UF4 amounts to 1 to 2% per hour at 650°C. Engineering information is needed on two steps in the fluorlde-volthlfy process: fluorination and UF, cold-trapping, Effective contacting of fluorine with molten ARE fuel is desirable to minimize fluorine consumption and gaseous waste and to assure complete recovery of the uranium. The cold trap should be operated in such a manner that all the UF, will be condensed and collected on fhe' ) walls of the trap rafher than lost as ¢ smoke. Cons;deruble experience “has’ been - "‘,K’25 in fhe operatlon"‘bf‘ cq[d traps equlpped wu’rhm' catch any UF6 ‘escaping 'fiiust be mmlmlzedfi“_ Also, the experzence'f'riflrUzss ”;'H:'Therma[ Reactor (STR) and the U2 5_.stainless ‘steel fue| elemenfs proposad for ‘the Submarine B !ntermedmte Reaci‘or w(SIR) and_the Army Package bromlde, whlch has physu:al propert!es very neur!y the same as those of molten NuF-ZrF4. No effect” of equipment dimensions on vertical mixing or turbulence was found, and the effects of physical _Lwhlchfl th-e\‘ fl uoride- fostered and expanded “Fot example, fuel element PERIOD ENDING DECEMBER 10, 1954 properties of the liquid could not be demonstrated by a simple correlation of friction factor with Reynolds number, such as is found in ordinary fluid flow through pipes. The acetylene tetra- ‘bromide studies showed that vertical mixing of the fiquid phase was induced by a gas rate of 9.5 cfm in a 12-in,-dia column, Equipment is now being assembled to verify these conclusions with molten NaF-ZrF . Cold traps to be used for a quantitative study of the effect of temperature and gas flow rate on the completeness of UF, removal are being con- structed. The design of these traps is based on a K-25 cold-trap design. APPLICATIONS OF FUSED SALT~VOLATILITY PROCESSES A long-range study has been made to survey the over-all feasibility of fused salt—volatility tech- niques in the chemical processing of ARE-type reactor fuels and certain types of heterogeneous reactor fuel elements. The volumes of radioactive waste from such processes should be much lower than those from the aqueous processes, and processing costs should be low, even though a means of disposing of excess fluorine will have to be provided. Total chemical costs, which in present aqueous processes represent approximately 10% of operating costs, have been estimated to be 20¢ per gram of U235, The operational procedure is much simpler, and the equipment should be inexpensive, even though nickel will be required as the material of construction. Results ofrecem‘ work on fuel element dl':SOIUflon dér V n of fused fluoride salt tech- ' mques' in a vcrlety“of radloochve chemical pro- ‘Cesses. Probable appllcahons in addition fo those for ARE-type fuels include the processmg of the -ercoloy fuel elemen’rs “used in the Submarine vo!ofll‘n‘y techmques may “be dissolution in the conventional nitric acid or sulfuric acid systems is very complex from the g ' 137 fakh b ANP QUARTERLY PROGRESS REPORT standpoint of both operational and corrosion prob- " lems but is efficient and rapid in the high-tempera- ”fure hqmd medlum of fused fluoride salts. Anrcrofi Reucior Fuels The fecovery and reprocessing schedule for ' fluid-fuel curcroft reactors will probably be dictated by aircraft reactor and turbine maintenance sched- ules rather than by the rate of formation of neutron poisons. [t is anticipated that operation will follo’w a schedule such as: (1) one day of oper- ‘ation’ aand one day of downtime for an accumulation “of s seven operahng days, (2) seven days of downtime for minor maintenance, (3) repetition of this sched- ule until 1000 operating hours have been accumu- lated. After 1000 hr of operation, the entire reactor will be dumped and the fuel will be reprocessed. The anMCIpated cooling period before repro- ‘:"'cessmg will be 10 days to allow decay of short- lived activities. The minimum decontamination factor requ;red for the process would be approxi- mately 100 for poison removal only. The other steps in fuel makeup can easily be handled remotely. The following essential steps are used in the chemical processing (Fig. 11.2). First, the UF in the molten fuel mixture is fluorinated to volatile UF, by introducing a 10-fold excess of elemental fluorine to achieve separation and partial decon- tamination from the other fuel components and fission products; second, the UF, is reduced to UF, with hydrogen in the gas phase in the Y reactor, as designed by K-25; and third, the re- sulting UF, is added to 2 moles of NaF to prepare a fuel concentrate for subsequent return to the reactor. Presentknowledge of this systemindicates that all steps are adaptable for radicactive remote operation. Considerable engineering development and operational experience have been obtained with the second and third steps, while extended laboratory development has indicated the feasibility of the dlrec’r fluormatlon of the molten fuel mixture. Heferogeneous Reactor Fuels The separation of U233 from zirconium alloy fuel elements is currently the most promising - 'applicafi'onrdf"the_ fluoride process to fixed fuel ’élemenf reactors. The dissolution rate of zirconium Cwith HF in the NaF- ZrF fused salt is very high, that is, 22 to 35 mlls/hr (Table 11.4). The range is probably due to metallurgical differences. 138 Although the ratio of zirconium to uranium is very high in STR elements, the cost of HF to dissolve zirconium will be only 5¢ per gram of U235 A probable method for zirconium separation as a guide for process development is indicated by the flow diagram shown in Fig. 11.3. This flow sheet is similar to that for fluid-fuel processing, but it “includes hydrofluorination of the fuel element in a TABLE 11.4, RATES OF HF PENETRATION OF VARIOUS METALS AND ALLOYSIN A TYPICAL FUSED FLUORIDE SALT BATH Bath composition: NeF-KF-ZrF4 (7-48. 5-44 5 mole %) HF flow rate: 250 em>/min Temperature: 675°C Nitrogen or argonblanket incases ofopen test vessels Penetration Rate Material {mils/hr) Vanadium Not detected Silicon Not detected Nickel 0.0001 Monel .02 Molybdenum 0.03 Tungsten 0.06 Silicon carbide 2* Type 304 stainless steel 4 Type 347 Nb stainless steel 7 Niobium 7 Tantalum 8 Manganese 10 Mild steel (Unistrut) 13 Thorium, 1/8-in. plate 14 Uronium 17 Zirconium 22 to 35%* Chromium 31 Titanium 31 Zircdloy-2 22 to 46** 95 wi %.uranium-—s wt % 50 ' zirconium Tin Sompledissolvéd instantly Zinc Sample dissolved instantly *Material disintegrated ond left suspendé&ufi‘firticles. o **Range is due to metallurgical differences in indi- vidual specimens, 2 o w3 ot o 7 - - . e NS 1 3 - } © . N . 2 2 & v & & S = - @ ® » > 0 (2) PERIOD ENDING DECEMBER 10, 1954 : Fa.F.RFy 2HF(H,) e — = WASTE WASTE | H L Y o, UFg UFg 2 2 B d.f.>102 B d.f.>10° Yy , Y Y 7 e U, — // Y REACTOR UFq SPENT FUEL _ -80°C(TO 50°C) COLD TRAP 650°¢C FLUORINATION WASTE STORAGE [ NaF - ZrF, WASTE <99% F.P. <0.02% U ~800°C REDUCTION SEeREt ORNL-LR-DWG 4387 2NaF (3) Na,UFg FUEL CONCENTRATE REACTOR FUEL Bl 3.2% U?°°F, 46.4% ZrF, 50.4 % NaF (1) UF,+10 F,—> UF+ 9 F, (EXCESS F, REQUIRED FOR COMPLETE UFg DISTILLATION). (2) F, DISPOSAL REQUIRED; RECOVERY OR RECYCLE NOT ECONOMICAL (3) REPRESENTS PREFERRED FORM OF CONCENTRATE FOR ADDITION TQO REACTOR. Fig. 11.2. Tentative Flow Diagram for Aircraft Reactor Fuel Reprocessing. 600°C NaF ABSORBER d.f. 50 F2 Ty FRR 2HF (H,) 2y X F 2 —_———— — —— COLD e WASTE | Thip WASTE HE Fs, UFg, F.EF, i i, WASTE y v § el Fz(s) Mg 800°C Ufs, F.PF, / UFg / NoF'2 } NoF ABSORBER 41 =104 Y b s 508 REFABRICATION 1 31 : 650°C % FUEL DISSOLVER . ELEMENT (HYDRO- 650°C -80°C(T0 50°C) 0=-50°C (5 ~800°C o . FLUORINATION) FLUORINATION COLD TRAP DISTILLATION REDUCTION UoR o ' £.PF. UR, NaF —Z¢F, RECYCLE WASTE STORAGE |me T ORNL-LR-DWG 4338 L/ =D STR FUEL ELEMENT |-t NoF — ZrF, WASTE >99% F.P <002% U {3 ASSEMBLIES) 560¢g U ’ 71,0009 Zr 2,000 g Sn ~309F.R S OVE ASSUMES NO PRIOR’ MECHANICAL SEPARATION OF U AND Zr}. gy F2 DISPOSAL REQUIRED; RECOVERY OR RECYCLE OF F, NOT ECONOMICAL. (5} PROCEDURE UNKNOWN ; DEVELOPMENT REQUIRED ) (3 ): U, +10'F, == UR; + 9F, (EXCESS 'R} REGUIRED “FOR COMPLETE UR; DISTILLATION). ot—— 7 | RCALOY 139 T T coa bl fused salt bath to permit dissolution at a pene- tration rate of 22 to 35 mils/hr, passage of the L'Jl'-'fi‘«l"':2 mixture through a bed of NaF at 650°C to obtain additional decontamination, and final puri- fication of the UF ; by distillation, 140 Recovery of the uranium by fluoride volatilization will be easy, although methods for obtaining a decontamination factor of about 106 will have to be developed in order to allow metallurgical pro- cessing of the uranium, | | - & Part K2 SHIELDING RESEARCH » " " d i T s . - < . ¢ ; - - = » - = 2 < - = - . - - - = = - - w -’ - * -~ = . N ~ R B SR aR . LR L e & S e o Ry .‘ * FRACTION' OF ENERGY TRANSMITTED o oot L . > 12, SHIELDING ANALYSIS E. P. Blizard F. H. Murray C. D. Zerby Physics Division H. E. Stern Consolldcted Vulfee Alrcroff Corporation S. Auslender Pratt & Whitney Aircraft The Monte Carlo method was used for two calcu- lations: the penetration of gamma rays through composite slab shields and the heating in beryllium slabs resulting from gomma rays in an adjacent source. In addition, a three-parameter mathematical formulation of the variation of radiation-induced injury to pilots of nuclear-powered aircraft with time after exposure was derived, and analyses were made of the preliminary differential experi- ments and of the first experiments with the GE-ANP R-1 reactor shield mockup at the Tower Shielding Facility (TSF). SLANT PENETRATION OF COMPOSITE SLAB SHIEL DS BY GAMMA RAYS C. D. Zerby A Monte Carlo calculation of the penetrcmon of gamma rays through the side of a crew-compartment shield has been completed, The study was ex- S e UNCLASSIFIED S - 2-01-059-10 ANGLE. OF ENCE, 60deg in. POLY- ETHYLENE 3 in. POLY- ETHYLENE 3in. POLY- ETHYLENE o o o o 0.05 0.02 . ANGLE OF INCIDENCE, 30deg ANGLE OF | INCIDENCE, Odeg - \.\»Ov:-“:fhfl 0’2 we 0.4.- e emi S NS e e o ] " Fig. 12.1. “Energy Resulting from Slant-Inei 2#m062 Photons Transmitted Through Composite Slabs of Polyethylene Backed by Lead. “LEAD THiCKNESS () T T tended to include an investigation of the effect of the variation of the parameters involved so that the penetration of any complex spectrum of incident gamma-ray flux could be determined by integration of the resulting data, Results of a calculation with one set of boundary conditions were given pre- viously,! are avo:lable. The results are presented graphically in Figs. 12.1 and 12.2 as the fraction of incident energy penetrating composite slabs. The exponential IC. D, Zerby, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 157. 2¢. b Zerby, Preliminary Report on the Penetration of Composite Slabs by Slant Incident Gamma Radiation, ORNL CF-54-9-120 (Sept. 21, 1954). UNCLASSIFIED 2-01-059-9 1.0 POLYETHYLENE 3 POLYETHYLENE 3 in, POLYETHYLENE 05 9 0.2 ot 205 ANGLE OF 30 deg FRACTION OF ENERGY TRANSMITTED ANGL.E OF INCIDENCE, 60 deg 0o 02 04 0 02 04 -~ LEAD THICKNESS (in. 6-m0c2 Photons Transmitted Through Composite Slabs of Polyethylene Backed by Lead. 143 cnd summary tabulations of the results. 12.2. Energy Resulting from Slant-Incident T i ot L R ANP QUARTERLY PROGRESS REPORT character of the data is evident from the straight- line fit to the points on semilogarithm paper. ENERGY ABSORPTION RESULTING FROM INCIDENT GAMMA RADIATION AS A FUNCTION ' OF THICKNESS OF MATERIALS WITH SLAB GEOMETRY C. D. Zerby The interaction of radiation with the atomic particles of matter results in an energy transfer to ‘the particles or in the creation of secondary radia- tion. The secondary radiation, in turn, is usually absorbed close to the primary interaction. With " high radiation intensity, the absorption of radiation energy in biological shields or in structural members can be important, Since essentially all the energy absorbed becomes heat, problems of induced thermal stresses in already stressed structural members may well determine design limitations, The heating effect may also be a limiting design factor in some cases 'in which material strength is seriously de- pendent on temperature. In all cases some knowl- edge of the energy absorption distribution within a material is necessary, Only a preliminary study of heating in beryllium has been completed thus far, and results are reported here. Most radiation absorption problems can be at least approximated by known analytical methods. For the gomma radiation absorption distribution, however, solution of the Boltzmann transport equa- tion is necessary for accuracy. At NDA, extensive calculations have been carried out® to solve the transport equation for the penetration of gamma rays, The NDA calculations included the energy absorption as a function of position for a point isotropic source,* but it was not possible to in- clude the absorption distribution for plane slabs, At ORNL, the total energy absorption of incident gamma radiation in the different materials of finitely thick composite slabs was also calculated,? but, again, the calculation did not include the absorption distribution, In this study the Monte Carlo method is being used for the calculation of the energy absorption ~ distribution in finitely thick slabs, and the calcu- lctions are progrommed for solution on the ORACLE 3H Goldstein and J. E. Wilkins, Jr., Calculations o the Penetration of Gamma Rays, NYO0-3075 (also ND ]5C-4'|) (June 30, 1954). Ibzci, Tables 7.113, 7.116, 7.119, 7.122, and 7.127, 144 at ORNL, For a monodirectional beam of mono- energetic photons on a slab of material of finite thickness, the energy absorption was calculated in every interval of 0.10 in, through the slab. The absorption included the energy transmitted to the material in a scattering collision as well as that transmitted in an absorption collision; in every case the photon was followed until it was reflected from the slab, was absorbed, or had penetrated the slab. The scattering cross sections were calcy- lated with the use of the Klein-Nishina scattering- cross-section formula,® while the absorption cross sections were obtained by empirical fits to published data,® A 9-in.-thick slab of beryllium was used, and the incident photons were considered to have energies of 0.5, 1.5, and 5.0 m0c2 and to be normally incident. For the 1.5-m0c2 case, angles of incidence of 30, 60, and 85 deg were also investigated (1 m062 = 0.51 Mev). The results obtained for the normally incident 0.5-m0c2 photons on the slab of beryllium are shown in Fig. 12.3, The histogram indicates the statistical nature of the Monte Carlo solution, while the smooth curve is a fit, by eye, to the data, The values given on the histogram are the averages of the values for the 0.10-in, intervals. In the future it is intended that a rigorous method of fitting the data will be employed. The area under the curve is the total energy absorption. Sw. Heitler, The Quantum Theory of Radiation, 3d ed., Eq.39, p 219, Oxford University Press, New York, 1954, 6. R. White, X-Ray Attenuation Coefficients from 10 Kev to 100 Mev, Table |, NBS-]OOS (May 13, 1952). UNCLASSIFIED 2-01—059—13 . 0.2(; &= ¢ O REFLECTED 0187 =% 06 ABSORBED 0.726 L= o PENETRATING ~ 0.087 x B o2 STATISTICAL ERROR % yli=d) =g 00 J/=FRACTION Lch 008 zQ 006 £ < 004 g 002 o o 1 2 3 4 5 & T 8 9 10 THICKNESS (in.) Fig. 12.3. Radiation Heating of a 9-in.-Thick Beryllium Slab by Normally Incident 0.5--mnc2 Photons. Smooth fit to statistical data shown. L oy s u) . i\ '_ Vthrs is a smooth fit, by eye, to the statistical data. crews of nuclear«powered mrcroff some : should be taken of the recovery from ‘radiation- UNCLASSIFIED - . ) . 2—01-059-15 S 030 — : ——— : - , = 028 'Npi’é’%“J FRACTION OF ENERGY - B 0.26 | —{ENERGY (mgc?)| REFLECTED | ABSORBED PENETRATING || Ro2a— 05 0.187 0726 0.087 || g 022 — 15 0.065 0.763 o172z | %.;3 02004 50’ 0012 0595 0393 || 018 Lt A JTO=F — % g STATIST!CAL ERROR VL) s ® 016 J/=FRACTION Z 014 >§\ | — \< E,,/O.Smoc 5 012 e 2 010 \‘\mjc Z 008 - % 006 [— z,>“~—-\:\\ O ' 5.0 moc . “T—.é % 004 [ - ] = \ 5 002 <[ e 0 : 0 1 2 3 4 5 6 7 8 9 THICKNESS (in.) Fig. 12.4. Radiation Heating of a 9-in,-Thick Beryllium Slab by Normally Incident Photons. In- Fig. 12.4 the smooth fits to the statistical dota are presented for normally incident photons of energy 0.5, 1.5, and 5.0 m, c2, The effect of the reflected photons on the energy absorption curve is readily seen at the entrance face of the slab. The value of the fraction of energy absorbed per mch at the entrance face for the incident photons can be calculated by multiplying e the scattering cross section, by f, the average fraction of energy absorbed by the material.” These values for 0.5, 1.5, and 5.0 myc? are, respectively, 0.118, 0.123, and 0.089. The large contribution of the reflected component is especially evident for the 0.5-mqc? photons at the entrance face as the difference be- tween 0.118 and the frccnon 0. 175 ’rcken from Flg. 12.4. ‘T‘he effec In “the Specn‘ncqnon of the tolerance dose for " The fcctor fwas obfamed as 1 minus the value given in a table by U. Fano, Nucleonics 11, Table 3, p 11 " (Aug. 1953). PERIOD ENDING DECEMBER 10, 1954 " UNCLASSIFIED 2—=01—059-14 % 060 i £ 056 FRACTION OF ENERGY i 052 ’ PENETRATING a 0.48 0.065 Q763 oi7e D 044 0.079 0.785 0136 @ 040 0.135 0.813 0.052 < 436 0.277 0.704 0.019 ;‘5 0.32 STATISTICAL ERROR + ./(1 J) W J= Z 028 = -'deg ANGLE OF £ 024 - & 020 QO Z ole & 012 g 008 = G 004 E 0 0 { 2 3 4 5 6 7 8 9 THICKNESS (in.) Fig. 12.5. Radiation Heating of a 9-in.-Thick Beryllium Slab by Slant-Incident '|.5-moc2 Fhotons. induced injury with time. Just recently the ANP Medical Advisory Group (ANP-MAG) issued recom- mendations® which made some concessions in this direction in that increased total doses were allowed for schedules which distributed the radiation over longer times. It has been demonstrated® that the ANP-MAG recommendations can be expressed to within 4% by a mathematical formulation with three parameters based on a model of partly irreparable injury and partly injury which recovers exponen- tially with time, The model was previously sug- gested by Blair.'=12 The mathematical formula- tion is much more flexible than the Biological Planning Chart which expresses the ANP-MAG "'recommendahons since it glves a umque value of “Radiation, ORNL CF-54-9-119 (S to the Guznea Pzg, Rat and Dog, _U R—207 (July phofons is rshO.Wn in Fig.” 12,5 Aga;n,“""""’ R 8“Mlnu'res of the 3rd ANP-MAG Meehng — 11-12 May ) ]954 Scbool of szatzon Medzcme, 5-18 4.9E P thord Tbe sze Varzatzon {or In,'ury /rom epf.2] 954) ]OH A Blctr A Formulatzon of the In]ury, sze Span, i Dose Relations Féo:v' lonizing Radiations — I Applzcatzon oLte tbe Mouse, U 206 (May 13, 1952). H, A Blcur, A Formulatzon of tbe In]ury, sze Spcm, Dose Relations for lonizing Radiations — 11, g %zggzt;on A 2H A. Blmr, Recovery from Rczdzatzon In]ury in Mice and Its Effect on LDSO for Durations of Exposure up to Several Weeks, UR-312 (Feb. 10, 1954). 145 :f’."'n "“.QM 2 Yo i 148 where ANP QUARTERLY PROGRESS REPORT injury for any radiation schedule no matter how irregular, The expression is t - dD - I(t) = [fi 1.08 x 1073 o dt”’ Jo & . +1.05 x 1072 e°°'713(f"‘")}dt” , I(t) = injury a;r' time ¢ in units of maximum safe . injury (MSI) which is just that allowed for ©" . training by the ANP-MAG (z 2 t’), a7 | ' 2 [ . — = dose rate, rem per unit time, at time :” dt (0 was found to contain the known 550 + é-kev transition, with a 1.05 * 0.05-min period, and bombardment of Rb87 produced the known Rb3® activity. The apparatus described above was also used in the discovery of a number of new isomeric transitions with half periods? between 0.05 and 5 sec. One of these is Zr?%”, which yields a 2.30 + 0.03-Mev gamma ray with a half life of about 0.83 * 0.03 sec. This isomer is produced by inelastic neutron scattering.> 2E, C. Campbell, R, W. Peelle, and F. C. Maienschein, Phbys. Semiann. Prog. Rep. Sept. 10, 1954 (unclassified), ORNL-1798, p 36. 3, c. Campbell, R. W. Peelle, F. C. Maienschein, and P, H, Stelson, Decay and Fast-Neutron Excitation of ngo , papet to be presented at the meeting of the American Physical Society, New York, January 1955, m i ""‘ e 7,; | i _@ ,;' j 59 N -y N » In summary, it appears that no previously un- known shorteperiod activities which might be of importance in the shielding of circulating-fuel reactors may be induced in K39 or Rb by reactor PERIOD ENDING DECEMBER 10, 1954 neutron bombardment. The newly discovered 2.3- Mev isomeric transition excited by inelastic scattering in Zr’® may be of importance in design considerations for a circulating-fuel reactor. " 2 4 7 D o TR i " 157 7 60 { f b ~ ANP QUARTERLY PROGRESS REPORT 15. TOWER SHIELDING FACILITY C. E. Clifford T. V. Blosser - L. B. Holland J. L. Hull F. N. Watson Physics Division D. L. Gilliland, General Electric Company M. F. Valerino, NACA, Cleveland J. Van Hoomissen, Boeing Airplane Company © Tests on the GE-ANP R-1 divided shield mockup began with measurements of the radiation around “* " the reactor shield in the Tower Shielding Facility " (TSF) reactor handling pool. The experimentation ~thus far has also included thermal-neutron flux and some gamma-ray dose rate measurements in the . TSF detector tank located a horizontal distance of 2264 ft from the reactor shield, This work was in- errupted for a period of two and one-half weeks so S ,,,that_ the TSF could participate in an Air Force . project in which a group of monkeys were exposed “to massive neutron radiation doses. The work on the G-E experiment has now been resumed. Analyses of some aspects of the TSF data have been completed and are presented in Sec. 12, “Shielding Analysis.”’ TSF EXPERIMENT WITH THE MOCKUP OF THE GE-ANP R-1 SHIELD DESIGN T. V. Blosser J. YVan Hoomissen ,D' L. Gilliland F. N. Watson The mockup of the GE-ANP R-1 reactor shield design (Fig. 15.1) is being used at the TSF for a series of measurements, Fast-neutron and gamma- ray dose rates were measured around the reactor shield section while it was submerged in water in the reactor handling pool. In addition, thermal- neutron flux and gamma-ray dose rate measurements have been made in the detector tank located a fixed horizontal distance from the reactor shield. The reactor loading for this experiment (Fig. 15.2) was a 5 x 7 fuel element array which gave a symmetrical power distribution throughout the re- < actor. In ofder to avoid excess reactivity, a fuel - element was removed from the center of each of " the two seven-element faces, The measurements mthe ‘pool "at the points indicated in Fig. 15.1 i H. E. fil"\‘l‘Ufl’gerf.orc;{, Bulk Shielding F'acz'lfty Tests on “the GE-ANP R-1 Divided Shield Mockup, ORNL CF-54-8-94 (to be issued); see also ANP Quar. Prog. Rep. Mar. 10, 1954, ORNL-1692, p 124. were compared with similar measureme'nfs‘nfl'n'ade earlier at the BSF.! L In general, the BSF and TSF data were in agree- ment, as is indicated in Figs. 15.3 and 15.4. How- ever, they were not in complete agreement, largely because of differences in the experimental setups at the two facilities. For example, while the TSF had a symmetrical power density distribution, the fuel element loading at the BSF gave an asym- metrical distribution which had a maximum 8.1 ecm from the center line of the reactor grid. Also, only compartments A and D contained borated water at the BSF, whereas at the TSF all compartments contained borated water for the first in-pool meas- urements, Later, plain water was used in Com- partment D. The BSF boration was 1.1 wt %, while the TSF boration was 0.85 wt %, For the measure- ments off the rear section at the BSF, no lead or iron side shielding was present, while at the TSF the side shielding shown in Fig. 15.1 was in position, For the measurements off the front and side sections, the side shield at the BSF had « total of 1in. less steel than the side shield at the TSF. Measurements of the neutrons scattered into the side of the detector tank (located a horizontal distance of 70.8 ft from the G-E reactor shield) were made as a function of altitude.. The flux (Fig. 15.5) is a slowly varying function of the altitude, and it exhibits a peak at about 35 ft. A decrease of about 6% in the readings between 150 and 200 ft indicates that the ground-scattered neufrons are still observable at these altitudes for this particular reactor-shield combination. The shape of the curve is quite different from that ob- tained in the differential shielding experiments with the water tank,? because the G-E shield 2C. E. Clifford et al., Preliminary Study of Fast Neutron Ground and Air Scattering at the Tower Shielding Facility, ORNL CF-54-8-95 (Aug. 23, 1954), 651 —SEcRE 2-04-056-3-T6A '35.2510n. 38.00 in, .50-in. GASKET 38in. STEEL 0.50in. STEEL 0.62in. STEEL .381in. STEEL 012in. STEEL 40 cm . I 20 ¢cm l-‘ T TSF ) REAR r— REACTOR COMPARTMENT F 0.75in. STEEL FRONT COMPARTMENT A 2.05in. LEAD (3 PIECES) 381in. STEEL! 1.5in. STEEL -1.5in. LEAD 3in. STEEL COMPARTMENT D 0.62in.STEEL 0.121in. ALUMINUM in. STEEL 3in X 3in. LEAD VOID COMPARTMENT C COMPARTMENT B 0.62in.STEEL COMPARTMENT E § 22.7 _L em ! 10.cm S . ¥ [ 20¢ ; m X ~ POSITIONS OF GEOMETRICAL )'(4i_ CENTERS OF COUNTERS Fig. 15.1. Mockup of the GE-ANP R-1 Reactor Shield. pS61°0L 439W3ID3Q ONIANI QOIY3d TR T T R ANP QUARTERLY PROGRESS REPORT UNCLASSIFIED 2-01-056 -0-Ti4R{ - STANDARD FUEL ELEMENT RALF FUEL L ] ELEMENT | . AFETY RODS fffff J W %3 i “ - OPEN LATTICE = POSITION R E@\ REGULATING ROD s = NOTE S AND R ARE HALF ELEMENTS | ) | L)) | | T T Fig. 15.2, Tower Shielding Reactor Loading No. 20 “SEeRET. 2—O01-056-3-T4A 9 1072 104 3 ol 1S REAR AND SIDE FAST NEUTRON DOSE RATE (mrep/hr/wm‘t)' ' 9 B 1078 FRONT FAST NEUTRON DOSE RATE (mrep/hr/watt) 0 10 20 30 40 50 &0 70 DISTANCE FROM SHIELD SURFACE (cm) | Flg. 15.3. .qus't»Neuh"on‘ DoseA Rate Measure- ‘ments in Water Near GE-ANP R-1 Mockup. 160 4 2-01-056-3-T3A O BSF O TSF 2 & TSF (NO BORATED WATER IN COMPARTMENT D) GAMMA DOSE RATE (r/hr/watt) 1678 0 10 20 30 40 50 60 70 DISTANCE FROM SHIELD SURFACE (cm) Fig. 15.4. Gamma-Ray Dose Rate Measurements in Water Near GE-ANP R-1 Mockup. mockup emits neutrons in a direction more favorable for ground scattering. The intensity from the G-E shield peaks at about 55 and 120 deg from the reactor-detector center line and is symmetrical about the reactor-detector axis, Measurements were also made of the thermal- neutron flux and the gomma-ray dose rates within the detector tank at an altitude of 195 ft. The G-E reactor shield was a horizontal distance of 64 ft from the detector tank, and, for these meas- urements, five l-in.-thick lead slabs were installed 1 ft from the rear face (reactor side) of the tank. The lead slabs simulated the shielding in the crew compartment, A knowledge of the thermal-neutron flux distri- bution in the detector tank is useful in extending the fast-neutron dose rate measurements into regions where the intensity is too low to be meas- ured with the dosimeter. This distribution is also necessary for predicting the capture g'amma-rgjy infensity to be expected in the crew shield. The 4= . *) .‘} . ‘,\ o 4 U @) RS a €« ) ‘»flgr5 - RELATIVE THERMAL NEUTRON FLUX (ARBITRARY SCALE) flux (Fig. 15.6) along the y axis (coincident with the reactor-detector axis) exhibits an average re- " laxation length of 5.0 e¢m in the water between the rear face and the lead. This indicates that these neutrons are largely air-scattered neutrons, as would be expected, since the thick neutron shadow shield eliminates the direct beam. The apparent transparency of the lead is suspect, since the possibility exists that neutrons were streaming in through the lead from the side of the detector tank. An x traverse (Fig. 15.7) was taken in order to indicate the attenuation to be expected from the _. < e L, S F;g.'|5.5 Thermal-Neutron Flux at Side of Detector Tank as a Function of Reactor- Reactor in GE-ANP R-1 Mockup. PERIOD ENDING DECEMBER 10, 1954 crew compartment side shielding, The apparent relaxation length, which is not constant, is very short, 4.2 cm, in the first few centimeters of pene- tration and increases to 6.0 cm between 25 and 40 cm of penetration. This short relaxation length is probably indicative of the large number of low- energy neutrons escaping through the air ducts in the reactor shield. In a y traverse (Fig. 15.8) the gamma relaxation length is approximately 13 cm at the front of the detector tank and approximately 17 ¢m at the rear of the tank. SECRTT 1-01-056-3-3CD CENTER OF TRIPLET BFy CHAMBER AT: X —67.5cm ¥y 62.0c¢m z 0.0c¢m Dirrw 161 e T i, o, i L e i s B ~ ANP QUARTERLY PROGRESS REPORT O - o - B 1-01-056-3-4(-)2 "R - ' 1-01-056-3—142 (-1 0 = h 195 ft —— CENTER OF TRIPLET BF; CHAMBER AT: S d 64 ft — ¥ 0.0 cm z 0.0cm " a O0deg y VARIABLE Wy z = _| < ,A-3 g '° < a ! W v Y ; L. 107 , _ _ | 0 25 50 75 100 125 150 ¥, DISTANCE FROM REAR FACE OF DETECTOR TANK TO DETECTOR CENTER (cm) - F:g.'|5.6. f.lgl.termal-Ne.ufron Flux Along y Axis of Detector Tank; Reactor in GE-ANP R-'lMockup. o b _ o 162 i . 3 ~NY S o [ L b, b s el ks .. il e i e o s > n 5 -~ o . & ~ 2 - z > a T N - * RELATIVE THERMAL NEUTRON FLUX (ARBITRARY SCALE) < S 6! o PERIOD ENDING DECEMBER 10, 1954 h 195 fi d 64 ft p —90 deg x VARIABLE y 90.0cm HORIZONTAL DISTANGE FROM & AXIS TO DETECTOR CENTER (cm B SEGRET {-01-056-3-7(-)1R{ z 0.0 ¢m 1 Moc fl(u.]’.'n-. CENTER OF TRIPLET BFy CHAMBER AT: 163 o B i ok e A B B A i B Bk b ale Rl nmfl’a; b SECRET 4 2-04—056-3—TI12R{A 10 CENTER OF 900-cc ION CHAMBER AT! x 0.0cm ¥ VARIABLE Z 00cm h 195 ft d 65ft ¢ 0O deg 1073 FIVE f=in. LEAD SLABS WITH Y4—in. TRAPPED WATER 10 GAMMA DOSE RATE (mr hr'/wcn”r' front of tank) 5 5 2 2 40"7 40“5 148.8 428.8 108.8 88.8 ©8.8 48.8 28.8 8.8 y, DISTANCE FROM REAR FACE OF DETECTOR TANK TO DETECTOR CENTER (cm) Fig. 15.8. Gamma Dose Rate Along y Axis of Detector Tank; Reactor in GE-ANP R-1 Mockup. 164 £ T ey Y % el r 2 i R e iliiin i — Pl g it g sty R Si_— : % ey e e e nr Tt e S . - o ¥ - = o o Y O - [&] o e = il g , } = = £ Ll '.—.