MARTIN MARIETTA ENERGY SYSTEMS LIBRARIES This report was prepared as an account of Gevernment sponsored work. Ne .ifher:_fh‘e: United States, nor the Commission, nor any person"acting on behalf of the Commission: [T e A, Mgkes any .wurrdnfy or representation, express or implied, with respect to fhe},nccu‘rfici; completeness, or usefulness of the information co_ntcj:ivned-_ .in this report, or thot the use of. any information, apparatus, method, or prqcess;crl'is_c_:losed in this report may net infringe privately owned rights; or ) ' L : Assumes any ligbilities with respect to the use of, or for damages resulting from the use of any information, cppfi;fi;us, method, or process disclosed in this report. . : As used in the above, '‘person acting on behalf of the Commission’ includes any employes or it qx s : . contractor of the Commission to the extent that such employee or contractor prepares, handles < Cor distributes, or provides access to, any informotiqn:puréuunt to h_is_._ employment -or contract. with the Commission, L R P R Ry, i 2k i L SR e, L R N A i I ploly]3 I | N o C.84 —~ Reactors-Special Features of Aircraft Reactors ORNL-2387, Parts 1-5 This document consists of 360 pages. Copy/'l'y-:)f 273 copies. Series A, ~ Contract No. W-7405-eng-26 AIRCRAFT NUCLEAR PROPULSION PROJECT QUARTERLY PROGRESS REPORT For Period Ending September 30, 1957 W..H. Jordan, Director S. J. Cromer, Co-Director A. J. Miller, Ass.i stant Director DATE ISSUED FEB 41958 0AK RIDGE NATIONAL LABORATORY ’ :Operated by . UNION CARBIDE NUCLEAR COMPANY DIVISIOI‘! of Union Carbide Corporation . Post Office Box X _ Ook R.dge, Tennessae MARTIN MARIETTA ENERGY SYSTEMS LIBRARIE i | 3 4456 0251034 3 Loy o b o el o WACIENENL o I, S BERR L S A i i ey Gl . sl bl i ki e i s 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 ORNL-1816 ORNL.-1864 ORNL-1896 ORNL-1947 ORNL-2012 ORNL-2061 ORNL.-2106 ORNL-2157 ORNL-2221 ORNL-2274 ORNL-2340 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 March 10, 1951 Period Ending June 10, 1951 Period Ending September 10, 1951 Period Ending December 10, 1951 Period Ending March 10, 1952 Period .Ern.c.ling June 10, 1952 Period Ending September 10, 1952 Period Ending December 10, 1952 Period Ending March 10, 1953 Period Ending June 10, 1953 Period Ending September 10, 1953 Period Ending December 10, 1953 Period Ending Mafch 10, 1954 " Period Ending June 10, 1954 ~ Period Ending September 10, 1954 Period Ending December 10, 1954 Period Ending March 10, 1955 Period Ending June 10, 1955 Period Ending September 10, 1955 Period Ending December 10, 1955 Period Ending March 10, 1956 Period Ending June 10, 1956 Period Ending September 10, 1956 Period Ending December 31, 1956 Period Ending March 31, 1957 Period Ending June 30, 1957 £ ik ‘::‘: [L] £ 0 1. R 2. C 3. M 4, D. 5. F. 6. E. 7. C. 8. W. 9. G. 10. M, 1. E. 12. W, 13. F. 14. A. 15. D. 16. C. 17. R. 18. R. 19. C. 20. J. 21. W. 22. R. 23. F. 24. D. 25. J. 26. L. 27. D. 28. W. 29. L. - 30. D. 31. A. 32, J. 33. W. . R. 35. A, . W, 37. A. . E. ' 39. C. 40, M. 41. E. 42, H 43. A 44, A 45, J. . G. Affel . J. Barton Bender . Billington . Blankenship . Blizard . Borkowski . Boudreau . Boyd . Bredig . Breeding . Browning . Bruce . Center (K~25) . Charpie o . Clark . Clifford Coobs . Cottrell Crouse . Culler . Cuneo DeVan . Doney . Douglas . Eister . Ferguson . Fraas Frye . Furgerson Gray - . Gresky R. Grimes G. Grindell Guth S. Harrill R. Hill E. Hoffman A A T OMO AP T IOrFrOYe IMrF>PrME=00OM-">MN-TN® . W. Hoffman . Hollaender . S. Householder T. Howe . Callihan . e C-84 ~ Reactors-Special Features of Aircraft Reactors INTERNAL DISTRIBUT ION . Emlet (K-25) ORNL.-2387, Parts 1-5 46. W. H. Jordan 47. G. W. Keilholtz 48. C. FE Keim 49, F. L. Keller 50. M. T. Kelley 51. F. Kertesz 52. J. J. Keyes 53-54. J. A. Lane 55. R. B. Lindauer 56. R. S. Livingston 57. R.N. Lyon 58. H. G. MacPherson gt B 59. R. E. MacPherson T !’{ 3 60, F, C. Maienschein ; ‘ 61. W. D. Manly v 62. E.R. Mann 63. L. A. Mann 64. W. B. McDonald 65. J. R. McNally 66. F. R. McQuilkin ~67. R.V, Meghreblian 68. R. P. Milford 69. A. J. Miller 70. R. E. Moore 71. J. G. Morgan 72. K. Z. Morgan 73. E. J. Murphy 74. J. P. Murray (Y-12) 75. M. L. Nelson 76. G. J. Nessle 77. R. B. Oliver 78. L. G. Overholser 79. P. Patriarca 80. S. K. Penny 81. A. M. Perry 82. D. Phillips 83. J. C. Pigg 84, P. M. Reyling 85. A. E. Richt 86. M. T. Robinsen '87. H. W. Savage 88. A.W. Savolainen 89. R. D. Schultheiss 90. D. Scott 91. J. L. Scott 94, 95. 9. 97. 98. 100. - 101, - 102. 103. 104. E. D. Shipley A. Simon 0. Sisman J. Sites M. J. Skinner A. H. Snell J. A, Swartout E. H Taylor R. E. Thoma D. B. Trauger D. K. Trubey G. M. Watson 105. 106. 107. 108. 109. 110. 1. 112, 113-115. 116-122. 123. 124-126. Wigner (consultant) Williams Wilson C. E. Vinters W. Zobel ORNL - Y-12 Technical Library, Document Reference Section Laboratory Records Department Laboratory Records, ORNL R.C. A. M. J.C. G. D. Whitman E. P. G. C. J. C. 127-129. 130-131, 132, 133. 134. 135-137. 138. 139-140, 141, 142. 143-144, ol 145, e 146. 2 147. T 148-161. 162. 163-165. 166. 167. 168. 169. 170, 171-176. 177. 178. 179-180. 181. 182, 183. 184, 185. 186. 187. 188, EXTERNAL DISTRIBUTION Air Force Ballistic Missile Division AFPR, Boeing, Seattle AFPR, Boeing, Wichita AFPR, Curtiss-Wright, Clifton AFPR, Douglas, Long Beach AFPR, Douglas, Santa Monica AFPR, Lockheed, Burbank AFPR, Lockheed, Marietta AFPR, North American, Canoga Park AFPR, North American, Downey Air Force Special Weapons Center Air Materiel Command Central Research Library Air Research and Development Command (RDGN) Air Research and Development Command (RDTAPS) Air Research and Development Command (RDZPSP) Air Technical Intelligence Center ANP Project Office, Convair, Fort Worth Albuquerque Operations Office Argonne National Laboratory Armed Forces Special Weapons Project, Sandia Armed Forces Special Weapons Project, Washington Assistant Secretary of the Air Force, R&D Atomic Energy Commission, Washington Atomics International Battelle Memorial Institute Bettis Plant (WAPD) Bureau of Aeronautics Bureau of Aeronautics General Representative BAR, Aerojet-General, Azusa BAR, Convair, San Diego BAR, Glenn L. Martin, Baltimore BAR, Grumman Aircraft, Bethpage Bureau of Yards and Docks Chicago Operations Office x) | 216-219. 189. 190, S 192--195, 1%, 197. 198. 199 200, 201, o202, 203, 204, 205, 206, 207. 209. 210, 211, - 212, - 213, 214, 215. 220. 221. 222, 1223, 224. 225, - 226, 227. o 228—229 . 230-247, 248-272. o 273 Chicago Patent Group Curtiss-Wright Corporation Engineer Research and Development Laboratories - General Electric Company (ANPD) General Nuclear Engineering Corporation Hartford Area Office Ildaho Operations Office Knolls Atomic Power Laboratory Lockland Area Office Los Alamos Scientific Laboratory Marquardt Aircraft Compony - Martin Company - - : ‘National Advisory Committee for Aeronauhcs, Cleveland- National Advisory Committee for Aeronautics, Washmgton o Naval Air Development Center _Naval Air Material Center 208. Naval Air Turbine Test Station Naval Research Laboratory New York Operations Office ‘Nuclear Development Corporotlon of Amerlco Nuclear Metals, Inc. Office of Naval Research Office of the Chief of Naval Operations (OP-361) Patent Branch, Washington Pratt and Whimey Aircraft Division San Francisco Operations Office Sandia Corporation : -Scheol of Aviation Medlcme Sylvania-Corning Nuclear Corporation Technical Research Group USAF Headquarters USAF Project RAND U.S. Naval Radiological Defense Lnborotory University of California Radiation Laboratory, Livermore - Wright -Aie Development Center (WCOSI-3) ‘Technical Information Service Extension, Oak Ridge .DIVISIOn of Reseorch cmd Development AEC, ORO 5 FOREWORD This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL records the technical progress of the research on circulating-fuel reactors and other ANP research at the Laboratory under its Contract W-7405-eng-26. The report is divided into five major parts: 1. Aircraft Reactor Engineering, 2. Chemistry, 3. Metallurgy, 4. Radiation Damage, and 5. Reactor Shielding. | With the suspension of work on circulating-fuel reactors as of September 1957, program emphasis will shift to research in support of the work of other organizations participating in the national ANP effort. The major fields of the future ORNL effort will be research studies of shielding, materials, and radiation damage and experimental investigations of systems and components of power plants designed for the nuclear propulsion of aircraft. ({ G vii St v b b PRI -l év Sl ey vk (AR ¥ 3 o L tm S ! » . - - ; i ' i | | £ ; 1 - i i . w LIt S e R ER A e ARy e WYL WL S e « TR nte T ce N Im T - 'CONTENTS FOREWORD ...ttt eensisiacatese s rensnrste e sasssssasserstasssss nssssssssssossssmassnsssassassens st st sssssassesassansssssnsnseses vii SUMMRY ............................................................................................... .‘...‘.....‘.,.'...' ........................................ xvii 1.1 1.2 PART 1. AIRCRAFT REACTOR'ENGINEERING REACTOR AND FACILITY CONSTRUCTION........... reresiaeensaanss eetereeseee e st et bar b aRasnetrestontensaneans 3 Methods Development ........ccvvreeceninses cevnerannns veesarenen ravisresssasnsnenbiseantan asneta s brnasee s dent s ner e sansanestanearsrass 3 WEld T @St cccriirecrerieciriiecmnrmsesesscsnensstesesasnssssssssarsnsssssessasnsssineasass evenesteesnsetsresees pesvesessenessnsnennants 3 Equipment Inspection................. rreisesnnanasiirens veesenips st sasesi ssts s shn s ena s b e bR R s er e e sR bR st b r s 4 Component Fabrication and Assembly ........cccicciivninniccsenivenionnes eanereanenaranis resessssammensasens resseeessesesasne 5 Main Fuel-to-NaK Heat Exchanger........coueucunnccs ceuereranesaesessaeasesresnnensasarereres rreseverenerbeneseressrasans 5 Sodium-to-NaK Heat Exchangers ......... erevebreeranire e s b ase revesesaeseeaee s vaveransinsan reresersererseneanesnens - ART-ETU Radiators......cuveervamrcvcorecnsns sterescreisonsisiennnsss tanan eeevassasesaiansresnaressesaesanasarersresnansaeasrase sanans S Main Pressure Shetl and Liner ............ resserssanebivtpeenninneisinnanaataens sanes reneseinensiesatentasessbentsessessseantons 9 Beryllium Reflector-Moderator Outer Shell ...t 9 Boron Shield Containers — Shells IV and V ........... 9 Inner and Outer Core Shells ...ciinmiiiviiimiiimmsroensistaresinens reeebeasssereteresnsentasmoases 13 "Reactor North Head ...t pideesnenes veeresneressnsereanee e seastbers st bPRaOsSa L e Loy asen s e e et 13 Strut Load Ring.......... vorersaseserenmissonnsase ssernas revnrsssares e raseseresararasssnerasanan et anstnsens aeresenenas resseeenanas 15 Neutron Shielding .......cccoeeuureurenceusen. eesnensessspes eerermesreessness s et sanesRsranas vesesrsesesanes preseasessenestenns 16 Lead and Tungsten Shielding ......cccceuuusn. eessestnseninsssssnsreRest s R AR RO g AR R Sur s SrvSRS e radUS R SRR A SRS 18 Miscellaneous COMPONENntS ... iiirmmmismsmsiiessssmssmastanssistusssassessasssesssesss smssasssssasessasessens 19 Reactor Assembly .........cccceveivvcnnse estemenoasnsansensarieas ivesserreeinenrsuenss sesnrenertesnessnrresase tenistsbonsessstasnbestanees 21 - ETU Facility ......... donieienseseiisassibnnansia sares euiceessuisenissssnesidinseessusernssenen sire dosers seerespinasresessnsresssanasne senrs sasees 22 Reactor Support Structure .......... ikt usiesiai e seinsanniasu b sesan s SosssLan e aussesa s s RR LB s Re e b1 22 NaK Piping, Associated Supports, and Components ..... rsteseebessensesseeassrasse resmenaseanenneenesressnsnennive B Electrical Equipment .....ciiivnviiccinvinininsn s innsiaseses mssssesaases reeeearteeterenresetantenassbere siet ot besbebattens 25 Auxiliary Services ......oocneennes Crsesnensnenssbane seperanriainER R RO 4 SRS NS KON SRS S AR IR LAS LSRR SRR SRR ARSI AR R SRS RS A 25 ART FQCiliy cuvveeereresmserassnssiesssassssssssmssssasessssssassesssnens reesertaraeraioseemitresstsonsestsinben nsisassiomase s sesaesssreeses 25 Design Activity...ciciinmcsiniinnee rerassssnsvenis seescesiasassises esmsisssnsrens Sestsumsensiasaisiosnatnsesninsiensirasssasansarenssns 25 CONSHUCHON PrOGIESS muviiiecsssiosiaivsmiseiusicssminsessissisminirssarssmmssssessssamesssansssssssrssisssssisssssssssasaisassans 27 Insta“ahon P!annmg................;.;..,_...‘;...;.,.',;'..,.;;.; ........ rersesivsssnssninens easssossassasens reseersessanisenssrsssninsres 27 ART DiSASSEMBY vovvuveresereervssasressserescsssissammsesssssssnisinssebens sssessssmsinsssiessssiisssmssiss eisssasemsisessssmasssssssonsoses 31 Reactor Disassembly e nainsimbinaesiasansbs Sssapsamasint nesats s s smistessiestesessasestsReIIASISY SRS SRS SRR SR CR LR bR - 31 Cutting Methods Evaluation ... cimiseniinssssssssens cersmssessosnsasmas sarasssasaneses wesesrrssessasinnes 31 Facfllfy Invesflgatlons eseiens ..... .............. esasiaeasaenrassaesrs sesiases 31 COMPONENT DEVELOPMENT AND TESTING serersie s siaebis s sEe s R RRs bR s e TR nes vemsenis 32 Pump Development TestS vmmrimininin eeseerdenrie s setn s s daege R iRt R beos SRR SR RSO RRS R PSR SRR sk s R SRR s 32 ' Bearing, Seal, and Lubricant Tests.-’....;....,..;.;..';L ............. eerersai s seiaraittsessbaaseana et snsariasiseannnashens 32 — Aluminum North Head Water Tests . immiussmimmmssssimmmssisesinssimmmsosssassmiss it issmssssss W 32 ~ Fuel Pump H|gh-Temperature-Performonce Tests eoviiodivieanintmes idhe s senaasebasnsas ssssne rerseesnrensrrseninsenson 34 Fuel Pump Endurance Tests . ....coiminivinmmsiescsrsss iises et sespiesensassiaeseb s st sssassonseserenestResseaeen vienrans 35 ancry and Auxiliary NaK Pump Development.........; ........... verirener et ereasnssseses sasseseresas vessrsranasaes 35 Reactor Component Development Tests ......cccoviiiiiusininoiiinmin s sessssstssssssssmase snsaes 38 Heaf Exchanger and Radiator Development Tests ........................... 38 /4 ' Valve Development Tests .....cccveveeenrreeniivecnennes reierenceneeseeensensesseneerans 43 Sodium Circuit Water Flow Tests ... iiiiiininsientssninnesssssssssssmsssmsasssmmsressstnsorasisasasess 44 Outer Core Shell Thermal Stability Test.... ettt cstesncsssssenrennss 4D Liquid-Metal-Vapor Condensers ......ccocveeececcivnnnncee. , ‘ Y 1. Zirconium Fluoride Vapor Traps ...ceeoceveerrennccnrse e . - 48 ISIAND BelIOWS T@Stuuiiiiritirieieeiieteivesecesetetsecsasaesestsssasssbesssasnssassenssnsssosestan sssesssese s sansnesssass ssssns snsssas 48 Fuel Fill-and-Drain Tank Test.....coioicigiimimmsimssesiissssssamessissssgessssssessosssassassassssssmssansss 52 1.3. INSTRUMENT AND CONTROLS DEVELOPMENT ........... verseeessiseensrennaas cesesneees 54 ART Control Rod Drive Test. .. ettt s s sonsnss s ssssssnsss sessnese 54 Fuel-Expansion-Tank Liquid-Level Indicator. ... sencnnserssssensssensssssessasssessssons o 54 Bubbler Plugging Tests. ...t s s snscsctenn s sesassscnsoses senens ssesesssssesasssassassasass 54 Aluminum NorthAHeod Expansion Tank Tests woneirccrinciniinnnins 55 ON-Off LeVel Probes ... oeeeccierrereeecsenne cenesnsennsecansnsassnse ovsnsarssssensasosest sasssssasesnessansrsssnssnsssssessns senons 58 Magnetic: Flowmeters .................. eeestenestetesasestessrasaneeasste et eaesaenaensenee st rns sneaneeasanenns ' ..... 58 Liquid-Metal-Level Transducers............... reretetstesetesseenesasnee e ress b sensbaenes 59 ART ThermOCOUPIES «uvvereeeiriirsaaecireessesreesssmasssasessassassasssses ssssssssssassnsssras sstssassstessnse s ssens amansssansesssesssosns 61 Sheathed TherMOCOUPLIES ... et ceeteccecceeesecertesnesnecsssnsbassessesaranesesanonnensnns veererenes eeeninsennnnes 61 We Il ThermOCOUP ES o et te e tearrescree e snn e sasess e sseesasneesssssesessassasanenasasssaransases e 66 SUrface ThermOCOUPIES ... uiieeiciecccneietrtcreeetesteesaenesssasesessmasassessasantrsssnesassnsrsessassesnsssssntasassnss 68 Apparatus for Checking Thermocouples ........riveeeervnccrnnsnennas : 71 1.4. ENGINEERING DESIGN STUDIES ..covuceveeeitncnnennenesmsensssisessensensessns s sass s sssesesssssssssasssassssostasass 72 Applied Mechanics and Stress AnGlysis ...t reads s neaens 72 Thermal Stress Cycling Tests of Beryllium ...t ctscrissssssenassans 72 Thermal-Cycling Tests of a Welded Core Shell Model ...t 73 Stress Analysis of Island Bellows ... .ttt s rssae e e 73 Stress Analysis of Shield Support Structure ...................... 75 Stress Analysis of a Pressure-Measuring Instrument .........ocieviviiviiinininnicsmicsssissnncssssenseas 77 Apparatus for ETR In-Pile Tests of Moderator Materials .....c..ccveievevnnercrerccisennnans corvesranaseinsrarseasasns 78 1.5, DESIGN PHYSICS .....cooerreeteererenrneeesentensssesnenessssesssesssnsnssssseesesssssansesaessnssass sassessssssssssesassssssnsssasee e snsens - 80 ART Fill-and-Drain Tank Shielding ....cccccciiieiriine et ccescreecstestesnsseseanessasse snssnssessssssssnsesssssann 80 PeRetratIONS ..o.eeeeeecicceenereeereee s me s seaessesesesseesase s saesessssansasseseesassesssssesesenssseserssssssssasensessesen sanesensases 80 Overlap in Shield at Removable Top Section ........ccceovrrrecmeeerrererrinersnesressnnesens renerenesessrsnsasaenes 81 Incompleted Work ................. vrereessesessreeraiaeass eeevsesessseriseestasasssensessreesssassannnassnnssens sasssnn nnonanssnissasans 81 Dose Rate in Region of ART PUMPS ..cccceeerrreeninececcccnnenccncseessnne e ssenes erseseieeseneaeaeretatns saaseans 81 Decay of Gamma Activity in U235.Containing ART Fuel for Short Times . ' After Shutdown ........ eteerseeteestetsseaetesatareaaetetesenas s s ehebasa s bens Sese bR se st R bR ere s R arR s s eannrann eevesereseseesenenses sasenes 82 Decay-Gamma Heating in ART Fuel Drain Valves...... et eeneresesene rrererebesebeneseresrsrestrasrasiens e 82 L.6. MATERIALS AND COMPONENTS INSPECTION .......oovceerrerrrarreannnns restensnssissnaseesaessarassenanes eeisrenanenens 83 Material INSPECHiON ....cocvveerererecverncrreesrrnresmraeesvaseneresssessssessneass reveveresatersasasresneesare asresssperaneesesbnberaes N 83 TUDING oottt e sese s ar s ssne st e as s nsts e s s sasnstene s nsse sasrneresa st e s hn e saesesnanesben raninesaranens 83 Pipe, Plate, Sheet, Rod, and Fittings cc..ccerirrcrenenrencrreeeserssrsivanessesnsassesssesssesssnnssees rveenessenes 83 Embrittlement of Inconel by Penetrants ......cvoicioncrnrinirnnnrsessesensseseseessessesnsssssnoreinaseens 84 Chemical Analyses of Materials.....ovnininnininninns caessesases arnis IR ESORESUFSRUSORES 4 PO L AR SRS 0 renneaeeanes . 8 Electrical Calrods .......... ' ........................................................... cereersssaseans reeeeanesrepesaans 84 [ C- Weld Inspection .......ccocveccueveenennenns Maeeesesesissesersresanares fereeersaseeeneareaneesnerenans eresasesimsrsssas sessstessisnsessasisansay Inspection of Reactor Shells ..... eeestersesesaesesara s ettt et e saes st et et she s e e a s s R sebebeeReraeRbena e e arereseneane nareitseres Inner Core Shells (Shell 1)..eiiiiviiciireee fetaresttet st te et et en sttt e rnser s beennt stk e rasbeua st sorace Outer Core Shells (Shell 11 ..ot sesbrresmasa e sessenass sssssessisssuessans sereen Shell Hl et vervensteneresaanaes Lt e e bt st sttt et s seressr s sabasesareseae reeuensieereeeaenens Shells 1V and V............. reestea et e ensnans veerenneesnsaterans eerens feensetesaseinanantstes esanentesaeretsstesnrsananbereessntrees Inspection of Components Received from Vendors......... S veeeereesarenes treeereesetessaeesrrestesserebesrae penessen York Corp. Radiators and Radiator Materials...........cinieuesvecnirsieeenesenesesssessesesssessenessnees Griscom-Russell Co. Heat Exchangers ................. reverasanesaene Black, Sivalls & Bryson Heat EXchangers ...t cssnsnesssseseeenesesssnssnsersssses Midwest Piping Company Forgings and Weldments .................... reereniesenestenressnans resreere e snaeae Process Engineering, Inc., Tanks and Bellows Expansion Jomts......_....; ................................... Fulton-Sylphon Company Bellows ......... eebeeesseereesasanotessteteee et rs st e e ba s eRbe s e s e s b bR e n s eusbeabes a0t bbae - Hoke, Inc., Inconel and Stainless Steel Valves ...t Metal ldentlflcahon MEter coveieeiereeecsieeereeeeeaesoien tretesieresaerersssterenaneneresssssarasttreren reeveneneeeeeseasaerasaessnsessas Instrumentation and Controls Inspechons sebaisesansassresusestiastans e eeserbesRON e RO TSRO SRR SRS SRR RSO PR SO R RS O RS Re e b OO 0O Nickel-Copper Welds in Liquid-Level Probes ...........ccommveinicrnrcincnr e sreneesissens s - Thermocouple Welds...........c..coee.ee. vedvaeeseseratteesresReranesr et eaes ot tanas e At et eeaseshes aenr e se bt e seraransans nries Procedures for Radiographic |nspechon of Weldments on ART lIsland and - REFIEEION woriiieierreeircrercre sttt redeseiaimn e e e ses sassnssesaresssbes e astnssrssesensessassansas teverereeestensentenerenesnsenens 1.7. HEAT TRANSFER STUDIES......coiinercenpensicsnessscssenssssseesenssenaas reess sttt ar e e s s s rse s asaes Thermal-Cycling Research.......ccccveerecvevievnnnenes eeveneeseerirassereesnenaraisasaseest sens s anetes uesaasternsesasasssrasassrananes Pressurized System ......... veereriueeeteerenane sessaenes veveresaresasnnstesnsssen ieesrsess b bbb s e tene Pulse-Pump System ......ccovervrnene, ceavesesnessssssasansene Fesspersesnesninsareresssnenasesess beresaserssansessrtsessnsnssnnenens Thermocouple Development......_.....-. ....... eereraeeraeserasnnsaen eterisaisseniesraar besase st seaaaste s se et e et e e e et ane ART Hydrodynamics c.ueerermnssssesesisessesesssssnessssmsnnns cres sttt s s s st et s banesa e nes renrerorisererseenss Full-Scale Core Studies..........conuen... deeneeesertsssrenasaises bnnestissensrasne satassesinares asenaratensesseseesueanesesens s sueren - QUArtEr-SCale Core STUAIES ... vt esisesssseesare sereesasabesssness e sasssasastsesassnensensaneesssss Instantaneous Velocity Profile Measurements ...........ccoreieriiivinveiceseevenvnnnneenssseriseessvessnmene sevese Fused Salt Heat Transfer .. iciinniarivenee rsisesscssesssssesstesesorns eveeresssnesesesessesnnn reevesseetaessrenssenrenssesens Heat Transfer Experiments .......cooccivrimnionnecsisesensenes vevereereresnrssnenrarenasaneres vetreserer s s s st ART-Type Core with Screens ...... eeeensavanas sreeesiesrtsenearerstan et ne e ser st st e antson bas s res s aSRT Rt RS e e R bes Vortex Tube...cccevrnrranene vorresrran vevaeiesanatens \ibdeberineesires saasusre sanineneaest sesioenen teeenerersentiesra e et e nrasenes Liquid-Metal Volume-Heaf-Source Expenmenf ceitessii e ri s sassrsaneee cerrrteseseamenns erereeserenesnsraenees Mass Transfer .t rrensnsasarsntnasens versieassar s asasssas sresans Physncal Properties ..........ummermmmssmesisens sersesbemsiensenstasas st e e iR e s ssabene terreessenseseass veeiesasesnnasyonsessaseas "Enthalpy and Heat Capacnty vere s rreneaaerneas veeveuiresnereasessbesseninai aseareniensasrrrnne asensseroras s ansansaseasen Thermal Conductivity.. . uummmmsiviicrismnmssissmsmssssssnsssssss e vertoraeuss saosssente sesesraessesesbeseatsanens vovones Propemes of Zirconium Fluorlde Vapor Deposifs ................................. cersearsanessasersennensrenes - PART 2. CH EMlSTRY 2.1, PHASE EQUILIBRIUM STUDIES .c.ocercorsncssissmstmssmsissrsssisesiossssiesmsscs | The System KF-UF , .......cco.c... eeeeesririneeen etae e AR 818 R A SRRk AR SRR The System NaF HIF ; oonre it veeseons 93 96 96 96 xi 2.2 2.3. 2.4. 2.5. xii The System KBF (sNaBF | w.c.oucenieec i, YHrium FIluoride Systems ... iicicreeeccrctcenentetornnsssmststnssessassssstossssisssssssnssssessmsisssorssnsassaseases CHEMICAL REACTIONS IN MOLTEN SALTS ....conminnrimcsenenisans et senae bt sare R e R sesstassensessaens Equilibrium Reduction of NiF, by H,, in NaF PERIOD ENDING SEPTEMBER 30, 1957 Fig. 1.1.23. Strut Load Ring Assembly. UNCLASSIFIED PHOTO 29703 UNCL ASSIFLED PHOTO 29260 FT T T T T T T T T T T T T 2 - 3 INCHES o 1 Fig. 1.1.24, Boron Carbide Tile Blank and Finished Part. 4 S & 17 ANP PROJECT PROGRESS REPORT UNCLASSIFIED PHOTO 29261 2 3 4 INCHES ,lllllll,1|l|||1'||llrlli||||'|.|.l.‘; 'l‘l'l" 5434 pieces is required for the ETU and ART Fig. 1.1.25. Boron Carbide Tile for Load Ring Assembly. involve 48 different shapes. Work is progressing reactors; the first 6 pieces have been received and are being inspected. One of the tile, can, and lid assemblies is shown in Fig. 1.1.26. Stainless-Steel-Clad Copper-B,C Cermets. -~ Approximately 15% of the rolled cermet sheet material being fabricated by Allegheny-Ludlum Steel Corp. has been completed ot a price of $1.07/in.2 for 0.100-in.-thick material. An esti- mated 80,000 in.2 will be required for the entire job which includes a small amount of 0.250- and 0.312-in.-thick material ot a cost proportional to the increased thickness. The ERCO Division of American Cor & Foundry Co. is fabricating the required cermet shapes from the rolled cermet sheet material. Tooling for this job is 95% com- plete. Seventy fabricated pieces have been re- ceived and accepted. The 1050 pieces required 18 satisfactorily on this job. The controlling factor in the completion of this order is the rate of cutting of the material by the Elox (electrical- discharge) process. _ The electrical discharge method has been the only practical way found to trim the cermet parts accurately, and the work has been scheduled on a multishift basis to expedite delivery of the parts. Some of the cermet shapes required in the load ring assembly are shown in Figs. 1.1.27 and 1.1.28. Lead and Tungsten Shielding A. M. Smith Details of the cooling coils in the equaterial section of the lead shield have been determined, and the design of the steel inner shell support " Fign ]. 1! 27. Plates. PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED PHOTD 29796 I l'l'lllll'l'_‘_'"|rT“" o 1 INCHES Fig. 1.1.26. Tile, Can, and Lid Assembly. 520 ANGLASSIFIED B9z Load Ring Cermet and Inconel Cover assembly has been modified by increasing the number of support ribs to reduce the stress level at operating conditions. Changes have been made in the method of welding these ribs to the shell to avoid buckling problems. A vendor is now fabri- ‘cating this support assembly. Work is continuing on the design of the north and south lead shield sections. It has been decided to support the tungsten shielding from the NaK manifolds, and the prehmmary layouts have been developed P Miecelluneous Components A M. Smith ¥. E. Thomas Reactor Spacer Rings.,- Parts for two spacer ring assemblies have been machined. One set of - parts that is ready for ossembly is shown in Fig. 7'1 1.29. Recctor Filler Plates. -Two sets of flfler plates f'are in various stages of completlon._. .One design - change was made to prov;de a better flow profile. The ETU filler plates, with the exception of plate ~No. 6, have been finish-machined on the mating ‘surfaces, the sodium passages hove been cut, and- ‘the bolt ond dowel holes have been drilled. The plates have been bolted together and set up for contour machining of the inner and outer surfaces, as shown in Fig. 1.1.30. Plate No. 6 for the ETU 19 ANP PROJECT PROGRESS REPORT UNCLASSIFIED PHOTO 29695 T T T T T T T T . 1 2 4 5 6 INCHES Fig. 1.1.28. Load Ring Cermet Pieces. UNCLASSIFIED PHOTO 29741 Fig. 1.1.29. Spacer Rings. has been finish-machined on the inner surface only. All ART filler plates, except No. 6, have been finish-machined on mating surfaces and the sodium passages have been cut. NaK Manifolds. — Design of the NaK manifolds for the ETU has been completed, and the drawings were transmitted to the vendor. Several minor 20 changes of ART manifold drawings are being mede. Sufficient material for the manifold nozzles was obtained despite the unusually close dimen- sional tolerances specified as a means of reducing stresses. It is presently excepted that a prototype of one of the ETU manifolds will be fabricated by October 1, 1957. i i i i i i i i i (] PERIOD ENDING SEPTEMBER 30, 1957 UNCL ASSIFIED PHOTO 29704 Fig. 1.1.30. Filler Plate Assembly. ETU Cell end Accessories. - Layouts o'f-f-he- equipment to be installed inside the cell are esti- mated to be 75% complete. A ‘‘negater’’ spring arrangement to aid in support of the dump valve actuators has been ordered. Preliminary drawings have been prepared of the mockup shielding and of the balconies or platforms that will be provided within the cell. Fobrication of the fuel fill-and- drain tank supports was completed. Other equip- ment such as . ZrF -vapor trops, ‘the auxiliary sodium exponsion tank and its associated piping, and the fuel overflow - line are being designed. Estimates of cost aond material needed for the fabrication of the various units have been made. | ~cold-positioning the island. REACTOR ASSEMBLY G. D. Whitman W. E. Thomas ~ The top and bottom neutron shielding cans for the top of the island have been fabricated and assembled for installation in the ETU reactor. ‘Two design changes have been initiated which will make final assembly easier. Welds on the sodium expansion tank are to be displaced to a point more distant from the sodium pump barrels in order to eliminate the need for boring the pump barrels after final assembly. A second change is being made to eliminate a possible hazard while In the previous 21 ANP PROJECT PROGRESS REPORT arrangement, a nut and bolt located between the pressure shell and the filler plates were to be used, but, if the nut and bolt were to become in-. operative during assembly, it would be necessary to grind out the pressure shell equatorial weldment and replace the threaded parts of the support struc- ture. The redesign eliminates the hidden nut and bolt by providing threads on the pressure shell and a plug and plunger arrangement that can be re- moved and replaced externally. E.TU FACILITY G. D. Whitman P. A. Gnadt A. M. Smith Installation work on the ETU facility is pro- gressing at the rate originally predicted. The status of construction as of August 15, 1957, may be seen in Fig. 1.1.31, which shows the control room enclosure, the NaK pipe main support columns and cell mockup, and the main piping of the furnace and isothermal NaK circuits, with the stationary units of the four pumps mounted at the top of the structure. In the background just above the center of the support steel is a portion of the main air duct in which one radiator is already installed, In the foreground are the two reactor support pedestals. cross bracing and walkways on the structure. The induction regulators and associated conduit installed in the basement area may be seen in Fig. 1.1.32, and the heater distribution panels, transformers, and associated switches may be seen in Fig. 1.1.33, with the induction regulators in the background. Present estimates show an expected ETU facility completion date, exclusive of the reactor and its associated cell equipment, of September 1, 1958. Completion of the cell equipment is scheduled for December 1, 1958. A description of the construction and design status of the facility as of September 1, 1957, is given below. Reactor Support Structure A. M. Smith Details of the reactor support platform were completed, and fabrication was started on both the ETU ond the ART structures. Plans have been made for the machining of the columns and the stress-relieving of the entire support platform. 22 The workmen are shown installing Column footings . for both ETU and ART support _ columns hclve been completed ~ | NGK Piping, Assb‘ciufed Supporié,‘dr'lacq‘:mfipn‘enfs P. A. Gnadt Suppo_rf- Structures and Shielding. ~ All main supporting steel has been set in place. Installation of cross bracing, gratings, shielding supports, and - hanger supports is approximately 65% complete. Work is now in progress on this portion of the facility. It is planned that the shielding plates will be installed during the summer of 1958. Design drawings are complete for this work, except for minor modifications. The shielding plate design is scheduled for completion early in 1958. NaK Piping. — The installation of the furnace circuit main piping is complete to the cell wall. Pumps, furnaces, throttle valves, and flowmeters are installed. One cold trap has been installed. Present plans are to install the remaining cold- trap circuit, thermocouples, liquid-level-measuring devices, fill-and-drain tanks, and associated lines and valves after completion of the auxiliary and isothermal NaK-circuit main piping. All material for this work is on hand, except the drain valves, drain tenks, and continuous-level-measuring de- vices. The drain valves are scheduled for delivery late in 1957. The drain tanks and continuous- level-measuring devices are scheduled for delivery October 1, 1957. The main piping for the two isothermal NaK circuits is 95% installed. Pumps, throttle valves, and one flowmeter are in place. The status of material for the remaining work is the same as noted above for the furnace circuits. A radiator has been installed for one heat dump circuit. The remaining radiator is scheduled for shipment approximately September 1, 1957, and it will be installed in the air duct as soon as it is received. The two pump bowls for these circuits are scheduled for delivery September 15, 1957. The main piping for these two radiator circuits is preformed ond will be installed as soon as the pump bowls and remaining radiators are delivered. The flowmeters are on hand and will be installed with the pump bowls and the radiators. The status of the cold-trap circuits, thermocouples, level devices, drain valves, and drain tanks is the same as mentioned above for the furnace and isothermal circuits. PERIOD ENDING SEPTEMBER 30, 1957 o w w w3 w < — Q z 3 PHOTO 29616 et 5T T R oy % WWH 7. (Seeret-with-saptiom 195 Status of ETU Facility Construction as of August 15 1L.3L 1. Figu 23 ANP PROJECT PROGRESS REPORT Fig. 1.1.32. Induction Regulators and Associated Conduit Installed in Buselfienf Area of ETU Facility, {(Gem Fig. 1.1.33. Heater Distribution Panels, of ETU Facility, (Confidertiel-with-copriom 24 - AW UNCL ASSIFIED 2 Fho Transformers, and Associated Switches Installed TO 29613 in Basement Area % f The cold-trap and plug-indicator circuits for the six NaK systems are prefabricated and ready for welding .into the main piping circuits. It has been decided, however, that the plug-indicator circuits can be eliminated, and they will not be installed. Main Air Duct. — All major components for this system are on hand. The section of the duct for housing the two radiators is installed. The re- maining sections will be installed following the installation of the radiators. Electrical Equipment P. A. Gnadt The emergency electrical system, consisting of the 300-kw diesel generator unit and associated switchgear and supply cables is 95% complete. Design work for the normal electrical distribution system is 90% complete, and will be 100% com- pleted by November 1, 1957. All completed design drawings were released to a cost-plus—flxed fee contractor on June 1, 1957. : The contracter has installed all vohoge regu- lators, transformers, motor control centers, heater distribution equipment, and 85% of the cable trays in the basement area. Installation of cables be- tween the induction regulators and the distribution cabinets is approximately 50% complete. Accept- ance tests for approximately 125 variable trans- formers revealed excessive winding temperatures at the specified ratings. These units are being returned to the vendor for modification or replace- ment. These variable transformers are for use as voltage controllers for o portion of the heater circuits. The cost-plus-fixed-fee contractor has completed the control room enclosure, installed the cable trays ubove the control room; installed all conduit entrances to the control room through the roof and floor, and is presently setting the steel support structures for the NaK motor con- trollers, resistors, and associated cable troughs. The variable-frequency drives for the reactor pumps have been ordered, and delivery is expected late in calendar year: 1957 ‘The motor-generator sets ussocmted with the pump drive motors will be ‘installed by the cost- plus-flxed fee contractor. The pump’ drive units will be installed after the reactor - and pumps have: been mstulled in the_'.- facility. De5|gn ‘work on pcmel onouts ‘and controls has ‘ been started by the Instrumentation and Controls Division. Preliminary ponel layout drawings and PERIOD ENDING SEPTEMBER 30, 1957 a set of preliminary block diagrams describing ‘contro} functions have been prepared. The control panels are to be available by approxlmately August 1, 1958. Details of the heater designs are bemg prepared, .and installation and wiring-connection drawings for the Calrod ond clamshell heaters are to be completed by December 16, 1957. Heaters and surface-type thermocouples will be installed after completion of the NoK piping assembly. Auxiliary Services A. M, Smith Piping drawings for the auxiliory services (helium, lube oil, air, and water) in the basement area have been revised. The installation work for this piping will be done by a cost-plus-fixed-fee contractor. The removal of some of the existing pipe and installation of the modified pipe runs were delayed until August 28, 1957, to allow the elec- trical forces to complete their overhead work in this area. Layouts of the auxiliary piping above the track floor have been prepared. A preliminary bill of material for items required on this part of the system have been prepared, and procurement of some items is now in progress. After considerable difficulty in obtaining a leak- tight system, the type R lube package (Figs. 1.1.34 and 1.1.35) delivery was made on June 28, 1957. This equipment is stored in the basement area and will be installed by the cost-plus-fixed- fee contractor. Failure of the subcontractor to deliver leak-tight pumps to the lube package vendor - has delayed the completion of the type K lube package. Delivery of this package is now scheduled “for October 1, 1957, ART FACILITY F. R. McQuilkin Design Achvntyf | Ma|or design work on components ‘outside the . cell of the ART facility is to be completed by April 1, 1958. The work outlined by these designs is to be done by ORNL forces. Design work is ‘proceeding on the special equipment room (equip- .ment for NaK coolant system fuel fill-and-drain tank), the radiator pit, and the radiator-penthouse - area (NaK coolant system equipment for fuel and sodium systems). Incomplete l/‘z-scclle models of 25 ANP 26 PROJECT PROGRESS REPORT ——— o = Fig. 1.1.35. Type R Lube Package for ETU Facility. i@l unCLASSIFIED PHOTO 29614 n._.. ittt e © # these areas, which are being made;to assist in detail design and to demonstrate equipment, as- sembly, and change-out procedures, are shown in Figs. 1.1.36, 1.1.37, and 1.1.38. Figure 1.1.36 is an elevation view taken looking southeast at the radiator pit. The circular pattern in the lower left corner represents the cell; the special equipment room is in the background. The air duct, which has not been assembled, will be located in the upper right corner above the radiator pit. The tank at the right is the drain tank for the radiator drain pans. The five vertical tanks are the main and aquxiliary NaK dump tanks.. The horizontal tank is the special NaK dump tank. The equipment structure to the left .of the drain tank ‘is the NaK purification system for the main and auxiliary coolant system and consists of the cold traps, plug indicators, flowineters, valves, and piping. The rack will be revised to omit the plug indicators. In the ceiling of the radiator pit are the main and duxiliary NaK piping, with the flowmeters and the heat-barrier-door operators. Figure 1.1.37 is a view looking down into the radiater pit with the air duct removed. The cell wall is at the left; the NaK piping and off-gas penetrations may be seen. The smaller boxes represent the NaK flowmeters and the larger boxes represent the heat-barrier-door operators. - The hot NaK piping penetrations for routing to the radiators are located between the heat-barrier-door operators. Figure 1.1.38 is a view looking south from the cell into the radiator pit at the right and the special equipment room at the left. The special NaK pump and motor, along with the motor cooling duct, are located at the top. The two large tubes - at the center of the special ‘equipment room-are . the spectrometer tubes. Below the tubes in the foreground is located the special NaK pUl'lflCGhOfl. system.” The. specml heat dump annulus and main "blowers, wnth trcmsmon ond ducf are’. locoted in S ,the background ' - S " The major |tems of plping and eqmpment for fl'le'_ = specml equnpmenj‘ room have been ossembled A prehmmury demonstrahon by use. “of “this model,-- indicates - fhat “the reqmred equipment change-out funchons can -be accompllshed Meanwhile the equipment _is belng rearranged in the radiator pit - “to provrde befler uccess for eqmpment mmntenance_ and change-out. Design work is in progress on the cell-evucucflon system, main blower back-flow louvers, smoke- PERIOD ENDING SEPTEMBER 30, 1957 generator equipment, water filter, control-room lighting, and additional switchgear. Detailed schedules for design and construction of package 3 (all werk that is remaining exterior to the cell) are being prepared. The remaining capital work, which is scheduled for completion by June 30, 1958, consists primarily of additions and medifi- cations to electrical facilities, cable trays, and instrument-room air conditioning. Shop fabrication of noncapital items has been started in preparation for construction work to begin at the facility " November 1, 1957. ltems scheduled to be started at an early date consist of the NaK purification equipment, the NaK dump tank installation, valve rack equipment, air duct structure, and NaK piping in the special equipment room and radiator pit. Construction Progress Lump sum contract work on the ART facility has been completed. This work was contracted in four stages, as determined by the design program. The first stage, package 1, included major building alterations, additions to existing buildings, cell installation, installation of main air duct, instal- lation of 1500-kva substation, and installation of 480-v main switchgear. The second stage, package A, included installation of auxiliary service piping. The- third stage, package 2, in- cluded installation of dlesel generators and facility, electrical motor control centers, spectrometer room electrical system, and the spectrometer air-con- ditioning system. The fourth stage, package 3A, “which was completed during this report period, ~included installation of an electrical system for - -supplying power to the pipe and equipment heaters, " a dry-air plant and facility, NaK pump ‘motor con- -trollers, a lube-oil fill-and-waste system, and a : hydrauhc system for Iouver operahon Insfnflnflon Planmng F’rellmmary plunnlng is in progress for instal- . lation of facnllty equipment for the ART. A revised . schedule of time estimates and completion dates - has been prepared and is being reviewed with regurd to the design und procurement programs. The - revused schedule proposes that installation - of Inconel piping ‘and -auxiliary equ:pmenf in -the “main, auxiliary, and special heat dump areas should commence during November 1957. It is estimated that 250 man-months of craft labor will be expended 27 8z Fig. 1.1.36. One-Twelfth.Scale Model of ART Radiater Pit. (Swcretr-wirhcaptiond ~; 140dIY SSITA20Ad L23r0dd dNV \ . 0!3\K RIDGE NATIONAL LABORATCRY LLAl“lIJ-.!l.lljlllillll?ll]?lll-lrlll?l]l?lll"l)ill m o 1.1.37. Model of Radiator Pit with Air Duct Removed, ~(Seeret.with captien) UNCLASSIFIED PHOTO 41475 192 Lelil LS6L ‘05 ¥IIWILJAS ONIANT @013 d 0t 1Y0d3Y SSIYV0dd LD33roAdd dNV 1 ] ”,-‘_-' ", X ) | 4 ! w i 4 y to accomplish the work planned ‘for fiscal year 1958. ART DISASSEMB LY M. Bender ; _ F. R McQunIkln, A A Abbaflello Reactor Dlsussembly ' Methods for taking accurate measurements fhrough ' hot-cell windows were mveshgafed further. A test performed with a low- power cathetometer gave accuracy in an acceptable range and no appreciable - varigtions over the surface area could be defected These results indicated ‘that such measurement methods should be explored in detail. Accordingly, optical tooling was studied and greater familiarity with these instruments and their adjustment was obtained by attending an optical tooling course. A design has been evolved which utilizes standard equipment for taking measurements. of radioactive - parts from outside a hot-cell window. A full-scale experiment and demonstration of optical tooling has been planned. The use of an optically flat reference bar mounted adjacent to and parallel with the exterior window surface will permit aligning the instruments with adequate precision. the accuracy of the measurements. The preliminary tests with a cathetometer indicate that small vari- ations may be acceptable, but further tests will be required to establish limits on the v::riat‘"io'ns.' Optical tooling has the advantage of requiring no _ space wflhm the hot cell. ' s:mulraneously controlled for operation in a hot cell. Tests at the manufacturer's laboratory indicate that these optical tools are: capable of meeting the accuracy requirements, but vcmqtlons _ in the glass of the hot-cell windows may reduce PERIOD ENDING SEPTEMBER 30, 1957 Further, since the instruments will be outside the hot cell, they will “not become contaminated. Cutting Methods Evaluation Further experimentd work was done with cutting _ methods for hot-cell use. Two methods were tested and discarded: the ultrasonic process proved to be extremely slow, and the Elox process was some- what erratic in operation on Inconel, although it -might be .applied to hard, elre'ctrica“y conducting .. materials. " A Heliarc cutting torch was found to be capable of severing multiple layers of metal Such @ torch could be remotely However, the -amount of activity released would probably ~limif its use to the reactor outer shells which have " -the greatest bulk but where the activity level is relatively low. Other cutting equipment has been received and is to be tested. Among these items are a portable bandsaw and a metallizing gun for more- complete evaluation of metal replication techniques. A stud gun has also been received for full-scale testing of the cnrmdge-acfuuted sealant injector. Facility Investigations Equ:pment was added to the / ,-scale model of the hot cell, such as the reccfor supports, the water bag, miscellaneous tools, and handling equipment. The model has been extremely valuable in visualization of the handling problems. A list was prepared of the tools and fixtures that will be required for removal of the ART from the cell. 'Tl"le_ list includes items that must be built ~into the cell and items that are portable. 31 ANP PROJECT PROGRESS REPORT 1.2. COMPONENT DEVELOPMENT AND TESTING | ~ H. W. Savage PUMP DEVELOPMENT TESTS W. B. McDoneald A. G. Grindell Bearing, Seal, and Lubricant Tests D. L. Gray W. K. Stair? An additional seal test was conducted to determine the effect of increasing the reactor sodium pump speed from 3000 to 4000 rpm, as necessitated by an increase in the required sodium head. In a 1000-hr period at 4000 rpm, the seal leakage of reactor pump rotary assembly was never greater than that previously observed ot the lower speed, and inspection revealed no deleterious effects. A plastic catch basin was fabricated for the reactor pump rotary assembly test, and a mechani- cal shakedown stand was fitted with glass ports to facilitate observation of the oil-sparging system. At a helium flow of 2500 liters/day down the shaft annulus, an oil carryover occurred which increased when the oil-sparge-line gas flow was stopped. When the shaoft ennulus flow was reduced to 500 liters/day, no oil carryover was observed when the oil sparge gas was flowing, and oil leckage was adequately sparged by a continuocus flow of 50 liters/day through the oil sparge line, which effected a 23-ft vertical lift of the oil. With a lower shaft ennulus flow of 500 liters/day and a low oil leakage rate, satisfactory sparging was achieved by intermittent gas flow. Work is con- tinuing to establish a detailed sparging procedure for ART and ETU operation. The reactor pump rotary seal assemblies have consistently passed the static 50-psi gas-pressure room-temperature test which is imposed prior to installation in a rotary element. However, the Graphitar No. 39 seal nose mountings in brass, SAE 1020 steel, and type 316 stainless steel rings have been found to leak at operating temperatures between 200 and 280°F, with no appreciable dif- ference being noted for the different materials. It is now apparent that further improvement in the mechanics of assembly and more severe acceptance tests are required. A redesigned positioning collar of the lower Durametallic NaK pump seal has functioned satis- 1Consultant from the University of Tennessee. 32 factorily in a series of tests simulating an ART reactor cell catastrophe. Various conditions were simulated in pressure tests of the NaK pump bearing housing and seal assembly, including oil system pressures from 0 to 250 psig, both with oil and with gas in the bearing housing. No failures occurred with gas pressures of up to 250 psig for either stationary or rotating seal elements. In one test with a stationary seal, a 200-psig oil pressure was applied in the lubrication system without leakage to the pump tank, which was at 15 psig. The reactor fuel pump rotary element, aitered for irradiation testing, was installed in a new gamma- irradiation facility at the MTR. The facility con- sists of an 8-in.-dia tube surrounded by racks of fuel elements immersed in the reactor canal. The header used for the experiment may be seen above the cancl parapet in Fig. 1.2.1. After approxi- mately one week of shakedown operation of the rotary element, partially spent MTR fuel elements were placed in the grid to supply the test radiation. Operation of the experimental equipment was stopped temporarily for repair of a defective ion chamber, and a general inspection revealed the equipment to be in good condition. Testing has been resumed at the rated pump speed of 2700 rpm, and the average dose rate is 108 rep/hr. Aluminum North Head ther Tests J. W. Cooke? H. Gilkey? Fuel System. — Operation of the twin fuel pumps installed in the aluminum mockup of the ART north head has demonstrated effective removal of air from the test loop over a wide range of operating conditions. Degassing times for removal of 0.1 scfm of air ranged from 8.5 min with 3’/2 in. of fuel in the expansion tank and the pumps operating at 2100 rpm to 1.5 min with 2 in. of fuel in the ex- pansion tank and the pumps operating at 3000 rpm. No ingassing occurred at speeds from 0 to 3000 rpm when the pump speeds were equal and some variation in relative speeds could be tolerated above 500 rpm. ' 20n assignment from Pratt & Whitney Aircraft. f “ i’ PERIOD ENDING SEPTEMBER 30, 1957 An eight-channel pressure-recording instrument fluctuations increased with an increase in the was used to measure the fluctuations of the dis- pump speeds and o decrease in the expansion-tank charge and suction pressure taps and the flowmeter liquid level. The maximum recorded pressure orifice taps for both pumps. As was expected, the fluctuations at 3000 rpm and a flow rate of 780 gpm FIED] | NRTS-57-3919 "4 P s L ST, Fig. 1.2.1. Reactor Fuel Pump Rotary Element Irrediation Installation in MTR Canal. 33 ANP PROJECT PROGRESS REPORT per pump, with l/ in. of fuel in the expansion tank, were 10.5 psi for the pump discharges and 0.8 psu for the pump suctions. The performance of one of the fuel pumps was comparable with that obtained in single-pump tests, but the head obtained with the other pump was 7% (2.5 to 3 f1) below that for the single pump at 2700 rpm and 645 gpm. This difference can be explained by minor dimensional differences in the impeller and volute. Sodium System. — Fabrication of the twin sodium pump loop for testing in the aluminum north head mockup was completed. Preliminary data showed that the system readily degassed with the pumps at equal speeds above 3000 rpm and with the liquid level in the expansion tank above the ports that connect the expansion tank to the pump impeller regions. The pumps shared the pumping lead equally when operating ot equal speeds. The loop flow resistance was considerably higher than anticipated, with most of the higher resistance being attributable to the sodium-to-NaK heat ex- changer pressure drop. Examination revealed that the high heat exchanger pressure drop was caused by blockage of inlet and flow passages by flakes of an epoxy resin that had been applied to the inner surfaces of aluminum components. Fuel Pump High-Temperature-Performance Tests P. G. Smith The fuel pump high-temperature-performance test loop in which NaF-ZrF,-UF, (50-46-4 mole %, fuel 30) is circulated was placed in operation again on June 26, 1957, and has since been operating continuously. For this series of tests a nuclear radiation thermal barrier was assembled with the pump rotary element in order to simulate, as nearly as practicable, the temperature distribution at re- actor zero power operation during priming and purge tests; a bubble-type liquid-level indicator was installed in the pump fuel-expansion tank; and four of the seven Moore pressure-measuring devices were replaced with Taylor instruments. The head-vs-flow performance tests at 2400, 2700, ond 3000 rpm were repeated, and the dis- parity in total head between the tests with fuel and with water was again found to be 1 ft, as previously reported.3 The total head obtained 3p. G. Smith and H. C. Young, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 34. 34 with fuel ot flow rates of 450 to 950 gpm ranged from 2 ft below that obtained with water to 3 ft above. The head-vs-flow performance curves ob- tained with fuel are presented in Fig. 1.2.2. UNCLASSIFIED . ORNL~-LR-DWG 25799 60 3000 rpm . 50 | N 40 S 2400 rpm \' \- NN 20 \ HEAD (ft) o o {0 o 500 600 700 800 900 1000 FLOW.(gpm) Fig. 1.22. Head-vs-Flow Performance Curves for an ART Fuel Pump Circulating Fuel 30 in a Test Loop. s ol ror) Cavitation tests at different operating conditions (600, 645, and 700 gpm ot 2400, 2700, and 3000 rpm) made during the previous quarter were re- peated. From the date obtained, curves showing the minimum surge tank gas pressure and the minimum suction pressure vs flow for cavitation- free operation were plotted (Fig. 1.2.3). Priming tests were made with the fuel at a tem- perature of 1200°F after having filled the loop to levels of 0, 9/1 , 2, and 3 in. gbove the expansion tank floor with the pumps stopped. In all cases in which the fuel was above the floor of the expansion tank, the pump primed and gave full head and flow performance. Without fuel above the floor of the tank the pump head was about 4 ft low. During these tests the effects of ZrF, vapor on pump performance were investigated at various liquid levels in the expansion tank and for various shaft- annulus helium-purge flow rates. With the pump stationary, a purge of 500/liters of helium per day down the pump shaft was maintained for 46 hr without fuel above the expansion tank floor, @ purge of 50 liters/day was maintained for 71 hr L8 9 MINIMUM SUCTION PRESSURE (psig) " MINIMUM SURGE PRESSURE (psig) with the fuel / in. above the floor of the expan- sion tank, und a purge of 50 liters/day was mgin- tained for 24 hr with. fuel at the 2- and 3-in. levels in the tank. During each of these tests the pump was rotated by hand one turn each hour. No detri- mental effects on pump mechanical or hydraulic performance were noted that could be attributed to ZrF , vapor. UNCLASSIFIED . ORNL—LR—-DWG 25800 32 -1 2, 30 6' . . . . - - : . 60 - e - - - 700 . . FLOW({gpm} . = | Fig. 1 2.3. 'Cur"fee "Shovririnrg ' Minim(mi “Surge and Suction Pressures vs Flow for Cavitation-Free Oper- ation of an ART Fuel Pump with Fuel 30. {Seeret-with £Laption” PERIOD ENDING SEPTEMBER 30, 1957 The bubble type of liquid-level indicator was used intermittently during these tests, and the level indications agreed with those of two spark plug probes to within 1/16 in. Fuel Pump Endurance Tests P. G. Smith The fuel pump being tested for endurance with fuel 30 as the circulated fiuid was stopped on June 4, 1957, after 3550 hr of operation. The wetted portions of the pump had been thermally cycled 652 times between 1100 and 1400°F, The pump lower-seal oil-leakage rate continued to be 5 em3/day, and there was no measurable leakage from the upper seal. Examination of the impeller revealed that it had been operating in a region of cavitation (Fig: 1.2.4). Data on input power at 2700 rpm vs pump surge tank helium pressure, ~which were obtained at the midpoint and at the end of the test, are plotted in Fig. 1.2.5. The slight difference in the maximum power input in these two runs may be attributed to the nonrepro- ducibility of the instruments used for obtaining the power data. The 4- to 6-psig range of surge tank pressures during this test was below that required for cavitation-free operation. Disassembly of the hydraulic drive motor revealed failure of the output shaft bearings prior to termi- - 'nation of the test. The pump and the test stand will be put back into operation for further endurance testing in the near future. " Primary and Auxiliary NaK Pump Development H. C. Youhg 5 ‘The primary NaK pump test that was terminated J. N. Simpson ) _durmg the previous. ‘quarter® when the NaK level - 'in the pump tank rose and flooded the oil cutch o fbasm has been exammed in order to determlne the oo -cause” of the- -operating difficulties encountered. -_-_";':'_The exammoflon of the lower seal has’ mdtcated - that grease formed by the interaction of NaK vapor -f;-und_ oil clogged the seal loading springs and _:,;,eventuall'y" lnhiBifed their operation. seal faces separated possibly because of normal . pump . wbrcmons or because of an increase in the ~ When the 4P G Smnh ANP Quar Prog Rep. June 30, 1957, 'ORNL.-2340, p 35. 50n assignment from Pratt & Whitney Aircraft. 4. c. Young and J. N. Simpson, ANP Quar. Prog. Rep. June 30, 1957 ORNL-2340, p 37. 35 9€ Flg. 1.2.4. Impelier of Fuel Pump That Was Enduronce Te e T s . e UNCLASSIFIED ; b : PHOTO 28996 sted with Fuel 30 at 1100 to 1400°F for 3550 hr. {SecTErwithrception} 13043y SSIFYJ03d LI23r0dd dNV 9 * "GNCLASSIFIED ORNL-LR~DWG 2580 es e o ) e 8 28 ./ DATA TAKEN 3-19-57 ) A A A Amd i - Pl N ' : ‘ / "~ DATA TAKEN 6-3-57 o ® = / & 27 A = Q a .— = % A Q. = > a b—'—l \\SURGE PRESSURE RANGE (4 TO 26 6 psig) DURING OPERATION FROM — l 1-8-57 TO 6-4-57 (3550 hr) f | | ~ PUMP SPEED: 2700 rpm 0 4 8 12 48 20 PUMP SURGE PRESSURE (psigy Fig. 1.2.5. Input Power vs Surge Pressure for Fuel Pump Operating with Fuel 30. fs'ecm-whl‘r*cupfl‘un) pump tank gas pressure to obove the lube o:l_‘fi_ pressure as a result of voponzohon of the oif that feaked into the NaK, the clogged sprlngs were unable to reseat ‘the seal faces and thus more. oil “leaked into the NaK.. Further: development work' on _the seal region of the pump is planned in an -~ effort to eliminate these seal problems. o After removal of the pump from the’ ‘test |oop b (No. l) a cover plote was installed on the - pump' tank opening, the atmosphere was changed fo argon, . and the loop was left for three weeks before it was An explosion occurred inside the loop - approximately 45 sec dfter removal of the throttle NaK sprayed through the throttle - valve opened. valve. opening and caused a lost-time injury. As a result detailed loop cleaning procedures were prepared PERIOD ENDING SEPTEMBER 30, 1957 for this and subsequent loop cleaning. The results of tests of the reaction of NaK and lubricating oil that were made in order to investigate the cause of the explosion are reported in Chap. 2.5, ‘‘Ana- lytical Chemistry."’ Since the oil leakage might have carburized the Inconel components of the test loop, Inconel samples were token from several areas and sub- mitted for metallurgical examination. The maximum depth of carburization was found to be 0.005 in., which was not considered sufficient to be harmful in further test operations. The loop has been cleaned and reassembled and is to be put back into operation. Operation of NaK pump hot test loop No. 2 was stopped for removal of four of the first set of 3'/- in.-IPS electromagnetic flowmeters inserted for calibration and installation of four uncalibrated flowmeters. When an attempt was made to resume operation of the loop, it was found that the main loop flow was restricted to very low values, that is, 100 to 150 gpm at 3550 rpm with the throttle valve wide open. Heating of the loop to nearly 1200°F finally restored full flow. The only subse- quent interruption of loop operation has been an unscheduled shutdown for replacement of the cold trap because water was leaking from a ruptured cooling coil into the loop drip pan. It is believed that the thermal shock that occurred when a full stream of cold water (60°F) was admitted to the cooling coil while the cold trap was at a fairly high temperature (450°F) hastened cooling coil failure. A nozzle was therefore devised to admit a controlled water-air mixture to the cooling coil. The nozzle has proved useful in effecting a gradual _transfer of the cold-trap cooling load from air to ~ water on all NaK pump hot test loops, and its use has ‘been suggested for the ETU cold traps. The total operating time for the pump in loop No. 2 is 1850 hr, which brings the total running time for 5pr|mory NoK pumps operating with NaK at elevated temperatures to 6250 hr. " "Auxiliary NaK pump hot test loop No. 1. was converted to ‘operation with hot NaK (56% Na) - during the previous quarter, The initial operation . of ‘this foop has been devoted to a study of the "oxide level in.a new NaK loop operating ot an elevated temperature. The test loop is constructed mainly of 4-in. sched-40 Inconel pipe, and it holds approximately 42 gal of NaK, For the oxide-level study the loop was heated to 1500°F without cold 37 ANP PROJECT PROGRESS REPORT trapping of the NaK, the cold trap being insulated and uncooled, and plug-indicator break tempera- tures were recorded as a measure of the oxide level in the NaK. Data on plug-indicator break temperatures were also tcken to observe changes in the oxide level caused by draining the 1500°F NaK into the system dump tank, cooling the NaK and the loop to room temperature, slowly refilling the loop, and reheating the NaK and the loop to 1500°F. This process was repeated twice. The data obtained in these tests are presented in Fig. 1.2.6. The plug-indicator electromagnetic flowmeter only occasionally showed a sharp de- crease (break) in flow for a particular plugging screen temperature during an oxide determination run, and thus the break temperatures are shown as ranges in Fig. 1.2.6 rather than as points. De- creases in the break temperatures with time may be seen at loop NaK temperatures of 1200 and 1500°F and for each dump-tank cooling process. The initial operation at a NaK temperature of 1500°F showed an average break temperature of 1250°F (extrapolated to 1350 ppm O,) and the final test run at 1500°F showed a break temperature of 900°F (575 ppm O,). The first break temperature for operation at a NaK temperature of 1200°F was approximately 1175°F (1275 ppm O,) and the last break temperature for operation at 1200°F was 700°F (225 ppm O,). The process of draining the hot NaK into the dump tank and cooling it there to room temperature reduced the oxide concentration in the system NoK. The oxide level decreased with time during the initial operation at 1500°F and increased with time during operation at 1500°F ofter the 346th hr of operation. Since the com- pletion of the oxide-level study, 500 hr of high- temperature. operation has been accumulated with this pump test loop. Head, flow, speed, power, and cavitation data for the pump will be taken during the next quarter. Auxiliary NaK pump hot test loop No. 2 was placed in hot operation during the quarter, and calibration of the first set of 2-in.-IPS electro- magnetic flowmeters for the ETU and the ART was completed. Four of the calibrated flowmeters were removed and replaced by four uncalibrated flow- meters. During the shutdown a defective liquid- level probe was removed for replacement and carbon deposits were found on it. The pump was therefore removed for inspection, and a lower seal was found to be contaminated in o manner similar to that 38 found on the primary NaK pump; that is, the seal- face loading springs were clogged with grease. The pump was cleaned, reinstalied, and calibration of the second set of 2-in.-IPS electromagnetic flowmeters will begin early next quarter. At shut- down for replacement of the flowmeters, 1070 hr of hot operation had been logged on this test loop that brought the total hot operating time for auxiliary NaK pumps to 1370 hr. REACTOR COMPONENT DEVELOPMENT TESTS D. B. Trauger Heat Exchanger and Radiator Develepment Tests J. C. Amos R. L. Senn D. R. Ward A summary of heat exchanger and radiator test operations during the quarter is presented in Table 1.2.1. The small heat exchanger stands were shut down for installation of fest pieces. The semicircular heat exchanger shown in Fig. 1.2.7, which was designed to simulate the tube stresses of the ART main heat exchanger, is currently being instalied in small heat exchanger test stand B, ond test operation is to start early next quarter. A description of this heat exchanger and the test objectives were presented previously.” A 25-tube heat exchanger, type SHE-7, 8 fabricated by the Process Engineering Corp. and designated No. 4, has been installed for testing with the fuel mixture NaF-ZrF 4-UF 4 (56-39-5 mole %, fuel 70). This test will be a repetition of the previous test which was interrupted prematurely after 1438 hr of operation by failure of York Corp. radiator No. 16 (for details see Chap. 3.6, this report). ' A Black, Sivalls & Bryson intermediate heat exchanger No. 3 (type IHE-8), which was installed in test stand B, had survived 168 thermal cycles and 421 hr of high-temperature operation when the test was terminated by failure of the furnace inlet line. Daily spectroscopic analyses of fuel samples for potassium from the NaK system, which was at a higher pressure than that of the fuel system, revealed no evidence of leakage. A first-approxi- mation thermal-stress calculation had predicted a 7). C. Amos et al., ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 43. 8 . H.Devlin and J. G. Turner, ANP Quar. Prog. Rep Sept. 10, 1956, ORNL-2157, p 50, Fig. 1.4.5. 4 o ) O UNCLASSIFIED ORNL-LR-DWG 25802 1800 # 4 e o * 4 l PUMP STOPPED AND NaK e . DUMPED AT 16:50, 7/30/57 . PUMP STOPPED AND NaK DUMPED PUMP STOPPED AT 22:05, 7/26/57, [ 1600 . TO SUMP AT 16:30, T/17/57 BECAUSE OF FROZEN COLD TRAP VALVE ~ COLD TRAP DPENED AT T o - \ - \ '/ / 16:45, 8/8/57 1400 . - ; o PUMP STOPPED AT 45:25, fi © 7/22/57 FOR CALROD _ f : , REPLACEMENT\ _ B j ey _ 1200 , * : o . = J & REMOVED INSULATION FROM COLD TRAP AND BEGAN COLD - T T TRAPPING NaX AT 09:45, 8/9/57 - ‘ ‘ < 1000 : T I : T ! & = - « . o L W . ui _ fl & T W 800 1 { }BREAK TEMPERATURE RANGE ON PLUG INDICATOR; POINT : INDICATES MOST SIGNIFICANT BREAK. 600 IA 400 - LOOP FILLED SLOWLY (Bhr] AT NaK . TEMPERATURE OF 200°F‘ ON 7/20/57 STARTED PUMP AT 09:2'5, _ . 7/30/57 (COLD TRAP OPEN) LOCP FILLED SLOWLY (9hr, 40 min) : ‘ | N AT NaK TEMPERATURE OF 100°F ON 200 |————— 1 SUMP FILLED AT 09:00, 7/12/57 8/6/57 WITH USE OF VENT BUBBLER — LOOP FILLED AT 15:00, 7/12/57 STARTED PUMP AT 08:30, :\J AFTER SUMP HAD COOLED 160 hr S | - ‘ 7/25/57 (COLD TRAP CLOSED) | : | ~w—5TA AT 08: 7 * —— PUMP STARTED AT {3:40, “W——STARTED PUMP AT 09:00, ?COESDT::;"ELJS%%‘;O' 8/8/5 7/16/57 (COLD TRAP OPEN) 7/22/57 (COLD TRAP CLOSED) l o "l. . L A A l_ 14 A b -_4'A. /l':l. 120 130 240 250 320 “340 350 440 610 620 630 0 100 : Ho Fig. 1.2,6. Results of a Study of the Oxide Level in the NaK Used in the Startup of a New Inconel Loop for High-Temperature Pump Tests with NaK. Auxiliary pump test loop No. 1. TIME FROM FILLING OF SUMP {hr) 640 LS61 ‘08 ¥IIWILHIS ONIAND aoly3d ANP PROJECT PROGRESS REPORT Table 1.2.1. Summary of Heat Exchanger and Radiator Operations (As of September 4, 1957) Howrs of Operation Howrs of Number of Test Unit Test at ART Design Nonisothermol Total Hours Thermol Status of Test Stand Temperature or Above Operation of Operation Cycles Black, Sivalls & Bryson IHE-B 671 800 2122 173 Terminated because of heat exchanger No. 2 NaK furnace faojlure (type HE-8) Black, Sivalls & Bryson IHE-B 168 253 1131 168 Terminated because of heat exchanger No. 3 NaK furnace failure {type IHE-8) York Corp. 500-kw IHE-B -800 2122 173 Terminated because of roc'!iciors Nos. 11 NaK furnace failure and 12 York Corp. ART test IHE-C 73 480 870 9 Terminated because of radiator No. 1 radiator foiluwre York Corp. ART test tHE-C 100 140 285 39 Test continuing radiater No. 2 : life of 35 to 45 cycles for this test, which was phase |l of a program previcusly described.’ Each thermal cycle consisted of the following con- ditions, as achieved experimentally: 42 min of steady-state operation at 1200°%, isothermal; transition to power over a 50-min period; 40 min of steady-state operation at power; and traensition to 1200°F isothermal operation over en 18-min pericd. Conditions during power operation were the same as for phase |, and the traonsition to power was at a constant rate of temperature change. The transition from power operation to isothermal operation was accomplished as fast as the thermal inertia of the test stand would permit, and the resultant rate of tempercture chonge ot the NaK inlet header of heat exchanger No. 3 ranged from 71.5°F/min starting at 1700°F to a final rate of 12.5°F/min between 1300 and 1200°F. Black, Sivalls & Bryson heat exchanger No. 2, which also was operated during phase | of this test, had been thermally cycled 173 times., Phase | included 503 hr of steady-state operation with & meximum NaK temperature of 1700°F. At the time of failure, the NaK furnace had operated for a total of 5329 hr, including 671 hr at @ maximum tube temperature of 1750°F. Since the furnace was designed for 3000 hr %). C. Amos et al., ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 46; ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 39. 40 at ¢ maximum NaK temperature of 1590°F, it must be replaced before further testing can be done in this stand. Because of the ensuing time delay, the heat exchangers are being removed for ex- amination and will not be tested to failure. The test of York Corp. ART test radiator No. 1 was terminated after 870 hr because of a leak in the tube matrix opproximately ¥ in. from the NaK inlet header in a tube near the center of the fourth row from the air outlet side of the radiator, as shown in Fig. 1.2.8. The leak was noted immedi- ately following the first power cycle, and rapid shutdown was effected to limit damage to the unit. Even so, the tube that lecked was damaged suf- ficiently by fire to destroy evidence that would indicate the nature of the failure. Metallurgical examination of 16 tubes, four of which were in the in the immediate vicinity of the tube that failed, did not reveal a cause for the leak. There were no incipient cracks and no defects in damaged areas, and no impurities were evident {for details of the examination see Chap. 3.6, this report). Inspection of assembled ART prototype radiators Nos. 2 and 3 will be even more rigid that the quite thorough inspection of unit No. 1. Heat transfer performance, gir pressure drop, and the increase in the NaK circuit pressure drop for radiator No. 1, as well as for radiater No. 2, which is now being tested, are in agreement with design values. ly UNCL ASSIFIED PHOTO 29722 FLUID=-NAK HEAT EXCHANGER 20 TUBES MATERIAL- INCONEL MANUFACTURER - ORM Fig. 1.2.7. Twenty-Tube Semicircular Heat Exchonger Type SHE.9. LS61 ‘0E YIIWILJ3S ONIGNI dOI¥3d ANP PROJECT PROGRESS REPORT Fig. 1.2.8. ART Test Radiator No. 1. {(Serrfrdmniialaith-cupTion UNCLASSIFIED PHOTO 29280 * " During radiator removal and replacement, extreme care was exercised to maintain an inert atmasphere on the test loop, The ART-type circulating cold trap was drained but not cleaned: In order to prevent oxide remaining in the cold trap from re- turning to the system, the cold trap was water- cooled at the design NaK flow during the heatup pericd, This procedure is contrary to the previous startup procedure of allowmg the cold-trap temper- ature to increase with the system temperature, cooling the cold trap with air until water can be introduced safely, and then bringing the cold-trap temperature down. There was no indication of the oxide plugging formerly observed when the cold- trap system was kept cold during preheatmg_of the system, - When the system reached 1200°F, a plug- indicator reading indicated a systemoxide satura- tion level of approxurnately 600°F; 48 hr later this level had been reduced to below the sensitivity ~of the plug indicator (300°F). - This startup pro- cedure demonstrated that if ‘adequate care is exercised to prevent OXIdGl’IOI'I of residual NaK - when o system is cut open, the - cold-trap circuit need not be replaced prior to restartlng system" operation, The excellent pertormance of the cold trap in ~ the HE-C test stand has led to elimination of plug indicators from the ETU ond. ART. -During both the initial startup and the restarting of the system, the temperature llmltatlons imposed on.the ART by the permissible berylhum temperatures were adhered to closely. This included establishing o temper- ature gradient in the radiater before the system temperature necessary for oxide cleanup of the new piping had been reached. planning the steps in the operatlon. _ Elimination of these. devsces ond their assocuated _equipment “and instrumentation should effect o substantlal- -saving for the reactor faclllttes. e Valve Development Tests L J A Canlm l T Dudley rvalve deslgnated ORNL-1, which .was described prewously, were terminated after 1500 hr ot 1300°F and 910 hr at 1500°F. During the test at - Pl ug-lndlcator , readings were taken to confirm that low oxlde_')) levels had- been achleved but were not utnl:zed in- A G Sm1th Jr.“.o_ ucceptably large. Although small scars have been Fuel Dump Valve.'.---' Tests of the fuel dump PERIOD ENDING SEPTEMBER 30, 1957 1500°F a steady increase was noted in the force required to open the valve. The force required increased from 865 |b immediately after the 500-hr closure ‘period to 2100 ib following 400 additional hours of cycling at 24-hr intervals. There was no measurable leakage during the test at 1500°F. Examination showed that galling had occurred between the Stellite-coated valve stem and the Inconel bonnet at the outer stem guide where the surfaces were near 1500°F and were exposed to the atmosphere. The valve was in excellent con- dition in all other respects, including the Kentanium 151A seat and plug, and is considered to be entirely ~ satisfactory for service at 1300°F, A second test valve (designated ORNL-2) which also incorporates the guided-plug feature, has been assembled, This valve, except for the valve plug cermet configuration and the omission of the sodium cooling jacket, is identical to the proposed ART fuel dump valve. The plug and seat, respectively, -are the cermets Kentanium KM and K-162B, the materials now ordered for the ART valves; however, ‘the plug cermet does not incorporate the mechanical ‘retaining feature of the ART valve, since the new pieces are not yet available. The valve includes o double bellows to minimize the chance of fission- gas release in the event of a bellows rupture in ‘the ART and o K-162B bonnet guide against ‘Stellite-coated stem to eliminate the galling which ‘occurred in the ORNL-1 valve. A test is scheduled to start early next quarter. The pressure drop characteristics of this type of valve when fully open are given in Fig. 1.2.9; the data presented were obtained from water tests. The ART valve * pressure drop - characteristics should be the same as those given in Fig. 1.2,9. ‘NaK- Dump Valve. = The first prototype NaK * valve!! received from Black, Sivalls & Bryson was tight when tested with water, but it leaked ~ badly when tested in NaK, The valve was dis- assembled and stress relieved, but there was no improvement - until it was rapped with a hammer during a stress-relieving cycle. 1t was then tight . with water aond initially with NaK at-the fest temperature . of - 1000°F. The leakage increased rapidly with time, however, and was . soon un- 109, assignment from Pratt & Whitney Aircraft. "y A. Conlin, 1. T. Dudley, and M. H. Cooper, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 41. ANP PROJECT PROGRESS REPORT SONTHEENTHre ORNL—LR—DWG 25803 100 - 50 20 NoK VALVE 10 HEAD LOSS (ft) FUEL VALVE {0 20 50 100 200 FLOW (gpm) Fig. 1.29. Pressure Drop Characteristics Observed in Water Flow Tests of ART Prototype NaK and Fuel Dump Valves. {(Serretwith-eeption) observed on the seat and poppet, the erratic results appear to be due to weld or thermal stress dis- tortions. The fuel valve, which had not exhibited this type of behavior, has a more massive end connection and o separate seat insert which is brazed to the body in a configuration which should minimize strains, The NaK valve is being rede- signed to provide similar relief as well as to accommodate cermet seats if this proves to be necessary. Water flow resistance was also measured for this valve, as shown in Fig. 1.2.9. “ln.Line’’ Vealve Actuater Test, — The NaK valve *‘in-line’’ actuator!! operated satisfactorily throughout a 3000-cycle test ofter revision to pro- vide for locking of the spring load adjustment nuts and correction of other minor defects. NaK Cold Trap Throttle Valves. — The NaK cold-trap throttle valve!! configuration has been 44 established, ond the required valves were delivered for the ETU. Modification of the remaining valves required for the ETU and ART is in progress. ~ Sodium Circuit Water Flow-Te:si's J. A, Conlin S. Kress'2 Two tests run during this quarter brought to com- pletion the water flow distribution and head loss examinations of various regions of the ART sodium circuit. A test was conducted in order to determine the proper size for ‘an erifice in the control red sodium cooling passage circuit to limit flow through that circuit to 0.032 cfs with the island flow at its design value of 0.697 cfs. The test piece simulated the control rod passage below the thimble and included the Inconel shoulder which screws into a threaded recess at the bottom of the island. Four s/1 -in.-thick erifices were tested which had inside diameters of 0,250, 0.375, 0.500, and 0.625 in. ond which were machined to fit snugly into a 1.422-in.-dia section at a transition to a 0.9«in.-ID passage. The head loss vs flow relationship for each orifice was correlated in terms of a loss coefficient, K, defined from the relationship AH = KV2/2g where AH is the head loss (in ft) across the orifice, V the velocity (in fps) through the 0.9-in.-ID passage, and g is the proportionality constant relating force to mass and acceleration. The loss coefficients are shown in Fig. 1.2.10 as a function of orifice diameter. | The second water test was a study of the sodium flow divider in the volute elbows located at each ART sodium pump discharge nozzle. These dividers split the sodium flow from the pumps, direct a portion downward into the reflector, and direct the remainder up to the island entrance region. It is estimated that to obtain the design reflector-to- island flow ratio of 2.2 at a sodium pump flow of 1.106 cfs, the head loss on the island leg of the flow divider should be 8 ft greater than that on the reflector leg. The purpose of this test was to determine the proper position for the flow divider tongue and to determine the magnitude of the head loss through each leg. The test piece, shown in Fig. 1.2.11, duplicated the ART component dimen- sionally, except for the flow divider tongue adjust- ment. A diagram of the test piece as installed in 129, assignment from Pratt & Whitney Aircraft. " UNCLASSIFIED ORNL-LR-DWG 25804 1000 500 g 3 1] Q N o LOSS COEFFICIENT, & (DIMENSIONLESS) S 0 0.4 0.2 03 04 05 06 07 ‘ ORIFICE DIAMETER (in.) ' Fig. 1.210. Results of Tests for Determining Proper Size of Orifice in ART Control Rod Coollng Passage. & rigd Hom) the test stond is shown in Fig. 1.2.12. Data were obtained at each flow divider tongue position (A, B, C, and D of Fig. 1.2.11) to determine the total head loss through each leg of the test piece, the flow ratio between the reflector and island, and the o entrance velocity head. ~The head losses were correlated - with the ‘entrance velocity by loss coefficients defined os stated above. The loss - - coefficients cobtained are plotted in Figs. 1.2,13° ond 1.2.14 .as functions of the flow raho and fher tongue pos:flon. CoT PERIOD ENDING SEPTEMBER 30, 1957 WONRTENTI ORNL-LR-DWG 25805 TO REFLECTCR FLOW DIVIDER TONGUE A B c D Fig. 1.211. ART Sodium Pump Volute Test Section Showing the Test Positions of the Flow Divider (A, B, 'C, D). Divider shown in pesition A. required -head difference and also results in the iowest net head loss. - : _Outer Core Shell Thermal Stability Tesi ' J. C. Amos - R. L. Senn The second test of a one-fourth-scule outer core shell. mode! for determining dlmens:onol stability “under thermal cycling conditions was terminated ofter 339 cycles, The test period was extended e R J _ - beyond the scheduled 300 cycles since no changes The head 'alffeféfic'e' between the reflector and island legs of the flow divider ot the design flow - and flow ratio is shown as @ function of the flow divider tongue position in Fig. 1.2,15. Position A ~ of the flow divider tongue gives approxlmutely the were noted in fluid circuit resistances. Loss of lubricating oil flow to the cold sodium loop pump "necessltated shutdown of the system, and the test ‘piece was removed for examination. Visudl inspec- tion has disclosed no damage ‘to the core shell, 45 o b b it ANP PROJECT PROGRESS REPORT WATER INLET - E _ —— . - TAPS MANIFOLDED TO OBTAIN HEAD LOSS ISLAND CIRCUIT R NOZZLE AND GAGE FOR MEASURING ISLAND CIRCUIT FLOW WATER OUTLET FLOW DIVIDER VALVES TO VARY FLOW RATIO GNP Tid ORNL-LR-DWG 25806 TAPS MANIFOLDED TO OBTAIN HEAD LOSS REFLECTOR CIRCUIT NOZZLE AND GAGE FOR MEASURING REFLECTOR CIRCUIT FLOW \ 4 WATER OUTLET Fig. 1.2.12. Apparctus for Water Flow Tests of Volute of ART Sodium Pump. and it is currently being measured to determine whether any dimensional changes occurred during the test. The core shell model and the test con- ditions were described previously.'3¢14 The shell will next be subjected to an extended creep buckling test at 1500°F by using an externul helium gas pressure of 52 psi. Liquid-Metal-Vapor Condensers J. A. Conlin A. G. Smith, Jr. The sodium vaepor condenser for the sodium-pum;; ‘purge-gas vent system has satisfactorily completed a 1500-hr performance test in a system in which 1000 liters of helium per day was being bubbled 46 through sodium held ot 1200°F. The condenser consists of 7 ft of 3/4--in. pipe inclined at an angle of 5 deg to the sodium-pump helium vent line. The condenser temperatures, which were lower than those used previously in order to cbtain more com- plete vapor removal, were 730°F at the inlet, 330°F at the midpoint, and 90°F at the outlet end. In the previous tests these temperatures were 810, 470, and 90, respectively. Upon sectioning of the ~13g. p. Whitman, A. M. Smith, and R. Curry, ANP - Quar. Prog. Rep. March 10, 1956, ORNL.-2061, p 58. 145, C. Amos and L. H. Devlin, ANP Quar Prog. Rep. March 31, 1957, ORNL-2274 p 50. o X OIS NTA L ORNL-LR—DWG 25807 6 —— 1 POSITION A 5 a4 POSITION 8 ~——-—9 POSITION C A . ===-=18 POSITION D e [+ ‘,a" g ’a’P” a7 l&:“ 1 ,4’ 14 ,’ 2 /~ /.o—"' ; 3 "’ — - i / L 2 J e . " / wi A /' S > ,-' / @ / / 3 |/ ' ¢ q ;" ‘/ ) |l i / /"—-_- I "l /' // 4 * 5 "-p-—v # - il 0 0 { 2 3 4 5 REFLECTOR-TO- ISLAND FLOW RATIO Fig. 1.213, Reflector Loss Coefficient vs Reflector- to-1stand Flow Ratic and Flow Divider Tongue Position. condenser, it was found that only small droplets of sodium had reached the outlet end. The inlet end contained some condensed sodium that had been frozen and trapped in the condenser. The NaK vapor condenser for the NaK pump purge vent, a. vertical 2:ft section of 2-in. pipe filled with Demister pocklng, was . sansfactoniy'-;r : tested with NaK for 1500 hr under the same - flow: = and temperature condmons as those used for -the test of the sodium pump condenser, Upon. termi- ‘nation of the test, a very small quantity of NaK " - was ‘observed in the condenser outlet gas line. The . condenser pocklng was - covered - wnth very small droplets ond fme strmgers of NaK The prototype NuK dump tank vent condenser, : "consisting of a vertical-12-in. section of 2-in, pipe " at the inlet ‘or_lower end followed by .o 12-in. - ~section -of 6-m. pipe filled with Demister packing at the outlet or upper end, which was tested under - simulated NaK dumps ot 1200°F ond proved to be - HEAD DIFFERENCE BETWEEN REFLECTOR AND ISLAND LEGS AT TEE EXIT (ff) PERIOD ENDING SEPTEMBER 30, 1957 SONFIGLMNTITY ORNL-LR—DWG 25808 12 " 10 \' \ s % A\ \\G ~———¢ POSITION A A POSITION B =—-=—0 POSITION C =====08 POSITION D LOSS COEFFICIENT FOR ISLAND o» 5 4 3 2 ! S —] : "'"-..,‘____' -""‘l-._ Ty —— 0 0 { 2 3 4 5 REFLECTOR-TO-ISLAND FLOW RATIO Fig. 1.214. Island Loss Coefficient vs Reflector-to- Istand Flow Ratio, and Flow Divider Tongue Position. . BSNMTENT e ORNL-LR-DWG 25809 20 N v N AN A B . ¢ D o ~ FLOW DIVIDER TONGUE Posmou 1 o . i " N o . Flg. l 2.15. Head D|flerenee Befween VTee Exils vs Flow - leder Tongue Position for a Reflector-to-Istand Flow Ratio of 2.2 and an Entronce Velocity Head of 11.43 #. 47 ANP PROJECT PROGRESS REPORT 99.9% efficient in removing NaK vapor,!® was sectioned and found to have fine droplets of NaK on the Demister packing similar to those found in the NaK pump condenser. As a result of these - tests the development work on the NaK and sodium pump and NaK dump tank condensers is considered to be complete. The NaK pump condenser is presently in use in conjunction with NaK pump tests to obtain further information on its performance in actual service. Zirconium Fluoride Vapor Traps J. A. Conlin A. G. Smith, Jr, Plugging of the zirconium fluoride vapor trap inlet'® was alleviated by increasing the outer- wall temperature of the inlet pipe to 1600°F, and the test was continued to completion. The test included 500 hr of continuous operation at a helium flow rate of 5000 liters/day from o fuel sump maintained at 1200°F and 50 cycles of 30 sec duration at ¢ helium flow rate of 10 scfm from fuel at 1300°F. There was no observable carryover of ZrF, to the off-gas line downstream of the trap. Light powdery ZrF, deposits were found on the water-cooled coils in the condenser inlet section and inside the tubes of the rear heat exchanger section, as shown in Figs. 1.2.16 ond 1.2,17. In the rear section of the trap, all the tubes were coated with vapor deposits on the inner surface down to and covering the surface of the Demister packing. A section of the packing removed from one of the 1%-in.-dia tubes is shown in Fig. 1.2.18, and the depth of penetration of the ZrF ¢ deposit into the packing may be seen. Although none of the tubes were plugged, the !é-in. tubes offered considerabie resistance to flow, and the pressure drop across the large tubes was sufficient during simulated dumps to move the packing te the rear of the trap. The quentity and physical properties of the deposits were considerably different from those of the deposits found previously. A total of 758 g was measured as compared with the 3400 g expected on the basis of previous tests. Densities ranged from 0.55 to 0.88 g/cm® compared with previous deposit densities of 4 g/em3; similarly, the thermal conductivity decreased to 0.04 Btu/hr.°F-ft2 from 15M. H. Cooper and A. G. Smith, Jr., ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 47. 165, A. Conlin and M. H. Cooper, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 48. 48 the previous 0.15 Btu/hr.F.ft2, The small quantity of ZrF, collected indicates that the purge gas was not fully saturated, and modifications are being made in the sump tanks to ensure more com- plete vapor saturation. The inlet to the trap is also being changed to a conical section to reduce thermal losses and to minimize the need for in- creased heating of this region, since the tempera- ture range is limited for the ART sodium system, which provides heating and cooling, as required. Additional thermocouples are included in the new test trap to facilitate more precise measurement of the thermal conductivity of the deposited ma- terial. This is particularly important, since the final design of the ART trap cannot be established until the volume and thickness of the deposit are established. When the reactor is operating at full power the surfaces wiil be subject to considerable beta heating from fission gas decay which will limit the deposit thickness as a function of its thermal conductivity. ' ' Other tests are in progress to simulate the effect of beta heating by radiant heating of deposit samples in order to determine the effect the beta heating may have on the physical properties of the deposited material. Island Bellows Test W. B. McDonald W. H. Kelley, Jr. A. S, Olson A test of the ART island bellows was undertaken because of the difficulty of accurately predicting by calculational methods the stresses to be ex- pected in this compeonent under service conditions. The test consisted of cycling the bellows under conditions of strain, temperature, pressure, and environment that simulated those of the reactor, In the test the bellows was compressed from normal to 0.090 in. in 5 min, held compressed for 25 min, returned to normal in 5 min, and left in the normal position for 25 min, The test temperature was 1250°F, and the pressure on the outside of the bellows was 3 psig for 50 cycles and 20 psig for the remainder of the test. The inside of the bellows was at atmospheric pressure. The fuel mixture NaF-ZrF -UF‘ (50-46-4 mole %, fuel 30) contacted the bellows on the outer surface, and the inner surface was in a helium atmosphere. Sodium was not used inside the bellows because it would have added little to the value of the test ” (e PERIOD ENDING SEPTEMBER 30, 1957 "UNCLASSIFIED 7 PHOTO 29050 B | \ | - U Fig. 1.2.16. Inlet End of ZtF ;~Vapor Trap Tested Under Simulated Fuel Dump Conditions. {Semerat with.oaptiom 49 ANP PROJECT PROGRESS REPORT o, o L& “2

4 C. S. Walker R. G. Affel The rod test system has been dismantled, and samples of the test sodium have been tcken for analysis. The Lindsay ‘‘mix'’ rare-earth-oxide control-rod slugs are clso being examined, os are the walls of the sodium contaginer. The water side of the heat exchanger is being examined for scale formation. [f the results of these examina- tions are satisfactory, it may be concluded that the ART regulating rod and its actuating equipment have met all the specifications that can reasonably be tested without the actual nuclear tests. FUEL-EXPANSION-TANK LIQUID-LEVEL INDICATOR R. F. Hyland Bubbler Plugging Tests The test apparatus was completed that was being constructed for determining the couse of the plugging of the helium bubbler tubes of the fuel- expansion-tank liquid-level indicator. The op- paratus and operating conditions were described previously.?2 The system successfully passed a mass spectrometer leak check, and preliminary tests of the oxygen- and moisture-removal systems are under way. The moisture content of the 200-psi building helium supply was found to be surprisingly vari- able. Heretofore, spot checks made on the header with an Alnor model 7300 dew-point meter showed an average H,O content of about 0.08 ppm by volume. On the new apparatus, however, a Beck- man mode! 179 electrolytic hygrometer is used, and in one month of continuous monitoring the moisture of the building helium was found to vary between 4.0 and 28.0 ppm by volume. The Beck- man instrument responded very well to liquid- nitrogen cold trapping of the helium at ofl H,0 concentrations, while spot checks with the Alnor unit showed no change during the entire monitering period. 2r. F. Hyland, ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 23. n The oxygen content of the 200-psi supply has likewise been continuously monitored for a one- month period with a Baker model $$-2 super sensi- tive Deoxo indicator, The oxygen content has been found to vary between 12.0 and 14.0 ppm by volume. The internal calibration of this unit was checked with an external electrolytic cell, and the two units agreed to within 2 ppm. Attempts to remove the small amount of oxygen with hot metal getters have been unsuccessful thus far. One type of getter that consists of copper turnings maintained ot 1200°F appears to be promising. A 50.0 ppm concentration of O, was introduced into the system and was reduced to 14.0 ppm by this getter. Aluminum North Head Expansion Tank Tests Additional tests of the helium-bubbler type of liquid-level indicator were run in the fuel expan- sion tank of the aluminum mockup of the ART north head in order to further investigate the pre- viously noted? level changes that were associated with changes in pump speed. Three bubblers were installed as indicated in Fig. 1.3.1. Typical dota obtained during ‘tests with these bubblers are shown in Figs. 1.3.2 through 1.3.5. When the 3R. F. Hyland, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 50. PERIOD ENDING SEPTEMBER 30, 1957 pump speed was increased, the level at bubbler No. 1 increased and the levels at the bubblers Nos. 2 and 3 decreased. Thus the previously postulated *‘dishing’’ effect on the surface in the expansion tank is further confirmed. There is an actual change in level, as indicated by the meas- uring system. A decrease in pump speed created an opposite and approximately equal effect. These tests were all conducted with a static level of 1.5 in. of water in the expansion tank and with the throttle valve set for design flow ot design pump speed, Attempts to confirm these results by using high- speed photography were inconclusive because of the difficulty of photographing the level through a small viewing port on top the expansion tank. The general direction of the level change, as observed photographically, agreed with that indicated by the measuring system, but it was impossible to- determine the magnitude of the change. Improved photographic techniques are to be used in further experiments, Tests were also run to determine how accurately the level measuring system would respond to a change in level with both pumps at design speed (2700 rpm) and with the system at design flow (645 gpm). Water was accurately metered into and out of the system under static conditions and the SR ORNL-LR-DWG 25822 o _%7.35in.—~——\ © BUBBLERNO. - ‘ 7L "~ {REACTOR) Fig. 1.3.1. Diagram of Fuel Expansion Tank Showing Locations of Three Helium-Bubbler-Type Liquid-Level Indicators. 55 ANP PROJECT PROGRESS REPORT UNCLASSIFIED UNCLASSIFIED ORNL—-LR—DWG 25823 ORNL—LR—DWG 25825 6 : I 6 | | S00rpm 3000 rpm:g-— 5 5 {000 rpm 2500 rpm—f 4 _ 4 . E K' 'E 2000 rpm%— w 3 =500 rpm w 3 — = S’ | g TEST NO.2 1|500 rpm - BUBBLER NO.{ j 2 oo —| 2 1000 rpm | TEST NO.Y | 500 r BUBBLER NO.! 2500 em q pm____ | { : N 3000 rpm 0 . 0 O i 2 3 0 i 2 3 FLUID LEVEL (in. H,0) FLUID LEVEL (in. H,0) Fig. 1.3.4. Fluid Level vs Time at Bubbler No. 1 Fig. 1.3.2 Fluid Level vs Time at Bubbler No. 1 During Second Test. During First Test. UNCLASSIFIED UNCLASSIFIED _ ORNL—-LR-DWG 25824 ORNL—LR-—-DWG 25826 6 amem | | —500 rpm 6 TEST NO. 1 ' 3000 rpm 5 f==w- BUBBLER NO.2 {3 5 : ——BUBBLER NO. 3_1::L1000 rom {?’ 3 _[i—‘2500 rpm 4 ':[ 4 - E 1500 rpm E 2000 rpm W 3 {4 | w 3 @ = f—E— 2000 rpm = ! - i l - ‘ 1500 rpm 2 T ., §—2500 rpm 2 "' {000 rpm ¢ : l !.' ] . 500 rpm { S 1 { 5 ! s 3000 rpm : TEST NO.2 < l { |~---BUBBLERNO.2 0 { 0 { |——BUBBLERNO. 3 0 ! 2 3 0 { 2 3 FLUID LEVEL (in. H,0) FLUID LEVEL (in. H,0) Fig. 1.3.3. Fluid Level vs Time at Bubblers Nos. 2 Fig. 1.3,5. Fluid Leve! vs Time at Bubblers Nos. 2 and 3 During First Test. and 3 During Second Test. 56 quantity required to vary the level 1 in. was de- termined. The same quantity was added or re- moved with the system at design speed and flow, and typical results are shown in Figs. 1.3.6 and 1.3.7. A peculiarity noted in these tests was' that while bubbler No. 1 yielded consistent results, bubbler Nos. 2 and 3 gave erratic and inconsistent results. _ o - it was felt that severe ingassing of the system might seriously affect the accuracy of the bubbler because of its fluid density dependence.. Con- sequently, a test was run in which the system was purposely ingassed, and high-speed photography was used to check the level indicated by the measuring system. The ingassing was accom- plished by operating both pumps at 1000 rpm and then shutting off pump No. 2. The results ob- tained are shown in Figs. 1.3.8 and 1.3.9. The indicated increase in level averaged approximately 1.2 in.,, whereas the increase determined photo- graphically was approximately 0.8 in. No density error is apparent here because the decrease in density caused by the ingassing should have produced a low level indication. Instead, the in- dicated level was 0.4 in. higher than the photo- UNCLASSIFIED . ORNL—-LR-DWG 25827 6 - > T~ 4 Y—1in. OF H,0 | - REMOVED N START TO REMOVE H L | DESIGN SPEED—2700 rpm DESIGN FLOW—645 gpm TIME (min) n o n LY i J | TESTNO. 6 _ L (I fBUB_BL‘ERrIN();q\ Y iy ~ FLUID LEVEL (in. Hy0) Fig. 1.3.6. Fluid Level vs Time ot Bubbler No. 1 When Water Was Being Removed Durlng Sixth Test. PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL—LR—DWG 25828 6 . | 1in.OF H,0 ADDED—'% ° p 4 § 4 = | TESTNo.6 £ BUBBLER NO. 1 / w3 ' i . i J Z | DESIGN SPEED— / 5 2700 rpm ~ DESIGN FLOW— - START TO 645 gpm ? ADD H,0 0 } 0 1 2 3 FLUID LEVEL (in. H,0) Fig. 1.3.7. Fluid Level vs Time at Bubbler No. 1 When Water Was Being Added During Sixth Test. UNCLASSIFIED ORNL~-LR—DWG 25829 6 t ) (‘ 5 ‘ TEST NO.3 ----BUBBLERNO.2 ¥ 4 | ——BUBBLERNO.3 <&z - | BOTH PUMPS | % £ TO 2700 _lrpm, DEGASSIING 1 . E_ N A w 3 SYSTEM INGASSING —F 2 i =T | PUMP NO. 2 OFF —37] : 1000 rpm —j 1 ' 3 4 0 4 2 3 FLUID LEVEL (in. H,0) Fig. 1.3.8. Fluid Level vs Time at Bubblers Nos. 2 and 3 During Third Test. 57 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL—LR—DWG 25830 © 1 BOTH PUMPS i TO 2700 rpm, DEGASSING 4 - SYSTEM INGASSING — = W > [ PUMP NO.2 OFF — = 1 C 2 1000 rpm TEST NO.3 BUBBLER NO.1 \ 1 o 3 0 ' > 3 FLUID LEVEL (in. Hy0) Fig. 1.3.9. Fluid Level vs Time ot Bubbler No. 1 During Third Test. graphed level. The reason for this error is un- known at this time. ON-OFF LEYEL PROBES G. H. Burger R. E. Pidgeon, Jr.4 In an attempt to eliminate failures® of the on-off probes used in NeK pump bowls, the intemal copper wire was aluminized to prevent oxidation. However, these level probes continued to fail at high NaK temperatures. Further examination of the defective units showed that the probes failed because of internal oxidation of the copper wires and that such oxidation occurred when the copper- to-Inconel welds were made. To eliminate this type of failure, the probe was redesigned so that the copper and Inconel junction could be brazed instead of being welded. No probes of this design have yet been tested in service. There are sixteen *‘on-off’' level probes now in use in the gas pots of engineering test loops. 40n loan from Radio Corp. of America. S5r. E. Pidgeon, Jr., and G. H. Burger, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 51. 58 The probes in one auxiliary NaK pump test loop have operated 1640 hr and those in a primary NaK pump test loop have operated 2162 hr without failure. These probes are installed where the NaK temperature is usually below BOO°F. Six probes are in use in the NaK pump bowls of two of the pumps being tested. As mentioned above, a number of these probes have failed and have been replaced. It is planned to replace all these probes with the redesigned units as soon as they become available following complete testing. MAGNETIC FLOWMETERS G. H. Burger C. L. Pearce, Jr.* Calibration and testing of the 2- and 3!5‘5"- magnetic flowmeters for use in the ART and ETU were continued according to the program described ‘previously.® Six 3’4-in. units in lot 1 were op- erated for 770 hr and six 2-in. units in lot 1 for 1078 hr in NaK pump test loops PKP-2 and PKA-2, After termination of the calibration test runs on lot 1, four flowmeters were removed from each loop and replaced by four uncalibrated flowmeters (lot 2). Two 3!(‘,- and 2-in. units from lot 1 were left in the loops for continued testing and to provide reference to lot 1. Thus far the 3‘4-in. vnits of lot 2 have operated for 1408 hr ond the 2-in. units for 431 hr. The two 3!5-in. reference units have operated 2178 hr ond the two 2-in. reference units for 1469 hr. The units which were removed from the loops were cleaned, rechamfered, and inspected in preparation for installation in the ETU and the ART. Eight magnetic flowmeters required for the ETU have been delivered and five units have been installed. Analyses of the experimental data from the tests of lot 1 are being made. In general, the results of these tests with respect to accuracy and reliability of the units are in agreement with the results presented previously.® In tests of the lot 2 units, which were instclled ofter the loops were opened and rewelded, the oxide level of the NaK appeared to be higher, as evidenced by erratic loop be- havior at temperatures below 800°F, and a drop in sensitivity of the flowmeters at low flows was observed which had not been evident to such an extent in the tests of lot 1. The earlier calibration 60. H. Burger and C. L. Pearce, Jr., ANP Quar. Prog, Rep. June 30, 1957, ORNL-2340, p 51. © showed the 3'/2-in. units to be linear to better than 1% over the flow range from 800 to 1600 gpm. From 600 to 800 gpm at fluid temperatures below 800°F, the sensitivity appeared to drop as much as 2% in some of the runs. For the 2-in. units of lot 1, no such decreases in sensitivity were evident. The effect was inveshgated in the 2-in. lot 2 units by running the loop at a high oxide level without the cold trap. The high oxide level in the loop did not affect the readings at the higher flow rates (450 to 600 gpm), but the oxide level or some other factor produced an apparent change in sensitivity of as much as 4% at lower flow rates (250 to 450 gpm). The sensitivity decreased with a flow decrease. The use of the cold trap did not correct this condition. Effort is being directed toward determining the exact mechanism by which the change in sensitivity occurs. All the :’%-in. mognetic flowmeters required for the ETU have been delivered end several have been installed. The units were delivered without the magnets attached to permit easier installation and handling. The temperature tests of the units have been completed and the results were satis- factory., The tests indicated that the units will perform in a satisfactory manner without continuous temperature and magnet flux monitoring. A curve of flowmeter output signal vs flow will be supplied with each individual flowmeter. LIQUID-METAL-LEVEL TRANSDUCERS G. H. Burger R. E. Pidgeon, Jr. Construction work was continued on the ORNL resistance-type level elements for use in the NaK ‘pump bowls. Twenty of these units have been completed. The first four units completed con- tained a sand-type (coarse-grained MgO) - of in- sulating material for the Inconel wires within the Inconel tube. Additional units were constructed' with’ fme-gromed MgO powder, -however, in an at- ~ tempt to get more dense packing of the material. Construction: of the level element containing the | fine-grained MgO was complicated by the tendency ~ of the packing tool-and the Inconel wires to gall - and ‘cause improper spacing of the Inconel wires. After experimenting with packing techniques, it was decided to return to- the use of the sand-type MgO for insulation. Some investigation is con- tinving of different insulating materials for ap- plications similar to this resmtunce -type level probe. PERIOD ENDING SEPTEMBER 30, 1957 Two of the ORNL level probes are in operation in the NaK pump bowls in the PKA-1 and PKA-2 test loops. These probes have operated a total of 835 and 1590 hr (plus 500 hr each in the test rig prior to installation in the pump bowls). Two level probes manufactured by the General Electric Company are in operation in the PKP-1 and PKP-2 test loops, and they have accumulated 4740 and 2162 hr of operation, respectively. Eight addi- tional units have been tested in the NaK-level test rig for the total operating times given below: Nc. of Units Operating Time (hr) 2 3060 2 2340 1 875 1 700 1 480 1 320 Thus far all failures of these resistance-type level probes while in operation have been caused by weld failures, and no failures have occurred during this quarter, Recorder readings of the NaK level taken at NaK temperatures of 1000, 1200, and 1400°F are shown in Fig. 1.3.10. All the values given lie UNCLASSIFIED ORNL-LR-DWG 25834 4 NaK * {O00°F 10 —eo 4200°F 9 & {400°F ] 8 27 o6 o s -4 3 2 1 K O - - 0O 20 40 60 80 {00 RECORDER READING Fig. 1.3.10, NaK-Level Calibration Test of @ Resistance-Type Level Transducer. 59 ANP PROJECT PROGRESS REPORT within 0.1in. (1.4% of full scale) of the line drawn for the data taken at 1200°F. The variation in the zero-level reading with NaK temperature is shown in Fig. 1.3.11. These data were taken while the level was being cycled from above to below the level probe. As may be seen the zero- level reading corresponds to the maximum output voltage of the level transducer. Data taken during this type of test indicate a zero-level shift of less than 3% over the temperature range of from below 300 to above 1200°F. UNCLASSIFIED ORNL—LR—DWG 25832 LY CYCLING ! wmloowm ZERO-LEVEL SHIFT {FROM CALI TESTS) | {~in-LEVEL SHIFT {FROM CALIBRATION TESTS) LEVEL CHANGE (% OF FULL SCALE) | rnlolw 300 400 500 600 700 800 900 {000 110C 1200 {1300 4400 NaK TEMPERATURE (°F) Fig. ‘.3"‘. Transducer as a Function of NaK Temperature. Variation in Zerc-Level Reading of The shift in the maximum level reading is es- sentially independent of NaK temperature, as would be expected, since the output voltage of the level transducer for the maximum level is only 1 to 2% of the full output voltage. During the 3000-hr test, the maximum level reading remained essenti- ally the same after the probe was wetted. Four level probes were operated for 54 days consecutively with the NaK temperature being held at 1200°F and the NaK level being cycled from below the level probe to near the top of the probe. The recorded zero-level reading was in- spected for drift which would indicate a change in the calibration of the transducer. This reading 60 drifted, but the drift was masked by a drift in the recorder calibration. A transformer used in the modification of the recorder was sensitive to tem- perdture, and, as a result, the voltage span of the recorder changed by an average of about 10.5% daily because of ambient temperature changes. Most of the drift in the zero-level reading was therefore caused by drift in the recorder calibra- tion. The recorder modification has been cor- rected so that the recorder calibration does not change.’ In order to observe the wetting effect, a new level transducer was installed and the NaK was heated to 300°F with the NaK level below the level transducer. The probe was then immersed in the NaK, a series of step changes in level was made, and finally the usual output vs level cali- bration run was made. This procedure was re- peated for NaK temperatures of 400, 500, and 600°F. In the 300°F test, the indicated level continued to rise after the step changes in level had been made. The indicated level reached to within 5% of its final value after 10 min and to within 2% of its final value after 2 hr, The re- sponse to the level changes at NaK temperatures of 400°F and up was good. The indicated output was about 1% of its final value immediately after the leve! changes. More tests under more rigidly controlled condi- tions will be required for a final evaluation of the wetting phenomenon. It is planned to continue the tests to determine the exact wetting effects on the measured level accuracy at various fill tempera- tures and to find, if possible, a solution to the wetting problem by using other probe material or by applying a special coating or plating to the level probe. Further testing of the NaK level probes has been delayed by repairs and modifications of the test rig. The level probes are installed in the test rig by clamping them to the NaK pots and sealing them with O-rings. The seals used leaked vapor, and, as a result, the test rig had to be rebuilt. . Construction was begun on ORNL resistance- type level elements for use in the sodium expan- sion tank. None of these units have been com- pleted. Construction was also started on the rlg for testing these level elements. The design was completed of the level element for continuous level measurements in the furnace- circuit drain tank of the ETU. This probe will be mounted in the bottom of the tank and will have a 30‘{,—in. range. The sensing element will be made from a single 1J‘,--in_. sched-40 Inconel pipe. The design of the probe for the auxiliary sodium expansion tank was completed, and fabrication work was started. This probe will be mounted in the tank bottom and will have a linear range of 25/8 in. The unit will be of the single-tube type, rather than the J configuration. A total of nine units will be required for tests, operational units, and spares. ART THERMOCOUPLES -~ J. T. De Lorenzo Sheathed Thgrmocouple‘s | Inconel-sheathed thermocouples (0.250-in.-0D sheath, MgO insulation) with hot-junction closure welds made by the Heliarc welding process are still being tested. Approximately 200 hr of addi- tional exposure time has been accumulated since the previous report. = Eighteen Chromel-Alumel thermocouples with closure welds thot passed dye-penetrant and x-ray inspection have now com- pleted over 6000 hr of operation at 1500°F in sodium.” The operating temperature has been reduced weekly to 1300 and to 1100°F for readings at these temperatures. No major shifts in drift {1100°F MANUFACTURER'S 1300°*F 1500°F _DEVIATION IN' TEMPERATURE (°F) o ® TURER’'S LOWER TOLERANCE ! o -6 11000 hr 2000 hr 3000 hr =16 ) - - LOWER TOLERANCE PERIOD ENDING SEPTEMBER 30, 1957 were observed dfiring the first 2800 hr of operation, but after that time four of the sheathed assemblies showed sharp downward trends and finally ex- ceeded the vendor's lower tolerance limit of minus 3‘4% at approximately 3800 hr. Since that time all four thermocouples have reversed their downward trend and are drifting back into tolerance, as shown in Fig. 1.3.12. A fifth couple that showed a sharp downward drift that stayed within the manufacturer’s - tolerance has leveled off and started o slight vpward drift. The remaining thirteen couples have, almost equally, shown either a slight upward drift or a slight downward “drift. Typical plots showing these slight upward and downward drifts are presented in Figs. 1.3.13 end 1.3.14. The total drifts for these thirteen couples have ranged from 2 to 8 degrees over the ~ entire test period, and it seems that the drift is independent of the test temperature. The data for all 18 thermocouples are summarized in Table ].3.]. Drift dota for similar thermocouples tested in air revealed upward drifts of all couples tested. 7). T. De Lorenzo, ANP Quar. Prog. Rep. June 30, 1957, ORNL.-2340, p 56. The earlier report cited data on 19 units; but one unit, which was suspected of having o leak, was pulled out of the apparatus after about 5000 hr of operation. UNCLASSIFIED ORNL~-LR-DWG 25833 46— — 4 MANUFACTURER'S LOWER TOLERANCE 4000 5000 hr Wog V243 V2ng M Yoy 2 2% M3 s M3 Yes U3 e Y2 %27 e Ter B Bne C O TIME Fig. 1.3.12, Drift Curves of 0.250-In.-OD Inconel-Sheathed Chromel-Alumel Thermocouple Showing Downward Drift Starting at About 3000 hr. 61 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL~LR-DWG 25834 OO°F MANUFACTURER’S UPPER TOLERANCE 12 1300°F - MANUFACTURER'S UPPER TQLEF\;ANCE 2 : MANUFACTURER’S UPPER TOLERANCE DEVIATION IN TEMPERATURE (°F) {500°F . 4 0 1000 hr 2000 hr 3000 hr 4000 hr 5000 hr Yo 243 Y2 Y2 Yor M Phe Uz Y Y3 Yes Uz % %2 G2 T r By %% TIME Fig. 1.3.13. Drift Curves Typical of Those Obtained for 0.250-in.-OD Inconel-Shecthed Chromel-Alumel Thermo- couples That Showed Slight Upward Drifts. ' UNCLASSIFIED QRNL-LR-DWG 25835 1100°F MANUFACTURER’S UPPER TOLERANCE 12 1300°F MANUFACTURER’S UPPER TOLERANCE 1500°F MANUFACTURER’S UPPER T R DEVIATION IN TEMPERATURE (°F) a _ 0 1000 hr 2000 hr 3000 hr 4000 hr 5000 hr " 2/, 1 4 "4y - ' g ‘243 2he Yo Yor P Phe U3 b Y3 Yhe 3 e %% S T Tor By TIME ' Fig. 1.3.14. Drift Curves Typlcal of Those Obtained for 0.250-in.-0D Inconel-Sheathed Thermocouples That Showed Slight Downward Drifts. ' 62 fe: W L It was believed earlier that leaks in sheaths of the thermocouples tested in fuel and in sodium could account for the downward drifts, and there- fore one sheathed couple which had exhibited a sharp downward drift was removed from the sodium Table 1.3.1. Results of Drift Tests of Inconel-Sheathed Chromel-Alume! Thermocouples in Sodium Test Thermocouple Rea‘din‘g (CF)¢ Period — = (hr) Deviation Spread® Test Temperature: 1500°F 0 52 3.5 1000 5.1 4.5 2000 5.4 6.5 3000 4.3 5.5 4000 4.4 : 7.3 5000 24.0(4. o 18.5(8.5)% 6000 1.9(13.5)¢ 30.5(13,5)¢ Test Temperature: 1300°F 0 7.3 X 1000 . 8.1 . 50 2000 57 . 60 - 3000 - 47(5.6° 12.3(5.0)¢ 4000 s 19373y 5000 - 3.8(5.5¢ 15.7(9.7)¢ 6000 . 2.7(4.6¢ . 19.5(9.0)¢ Test Temperature: 1100°F 0 68 26 1000 68 .34 2000 62 "'4*6" 0 6y . 10 4000 4.26.2)° 23,5(7.7)91; 5000 4963 “14.0(9.9¢ 14; 5(9 5,4, ‘ - gsooo . ,_-'-_:4.6(6-.2)“’ tained two fhermocouples cged 24 hr ot 135 pnor to testing. - -®Deviation of the average of the ihermocouple feodings : _from the test temperofure. e “Maximum spread of the fhlfty-elght readings obtamed.“fl“- 14 25,31, 87, and 91, eYalues in pufenthesee exclude two assembhes, Nos 14 and 31. fValues in parentheses exclude three assembltes, Nos 14, 31, and 91. PERIOD ENDING SEPTEMBER 30, 1957 for -examination, Metallographic examination re- vealed no evidence of a leak; in fact, the unit appeared to be remarkably sound even after the 5000 hr of operation in sodium at 1500°F. |t is now believed that the temperature cycling of the fuel and the sodium could have produced ‘‘breath- ing'’ of the magnesium oxide on the unsealed connector end and thereby induced conditions which might possibly cause the downward drifts. The temperatures were not cycled during the drift tests in air. Additional drift tests with sealed samples are being proposed in an effort to de- finitely establish the cause of the downward drifts. Thirteen similar specimens and three platinum, platinum=10% rhodium units have completed over 4000 hr of operation at 1500°F in NaF-ZrF ,-UF, (50-46-4 mole %, fuel 30). Weekly readlngs were also taken at 1300 and 1100°F. After the first 1400 hr of operation, three Chromel-Alumel units started showing sharp downward drifts similar to those shown in the first 4000 hr in Fig. 1.3.12, ‘and one unit finally exceeded the vendor's lower tolerance limit at 2000 hr. The drifts of the other two units hove leveled off, and slight upward trends have started. The remaining ten Chromel- - Alumel assemblies showed drift performances similar to those shown in Figs. 1.3.13 and 1.3.14 over the first 4000 hr. For comparison, the typical performance of the platinum, platinum—10% rhodium units as they aged for over 4000 hr in fuel 30 is shown in Fig. 1.3.15. The data on these 16 thermocouples are summarized in Table 1.3.2. A cross section of the hot-junction end of a - sheathed Chromel-Alumel thermocouple is shown . in Fig. 1.3.16. - The closure consists of a four- "+ hole Inconel plug which is welded to the sheath " and wires with the Heliorc welding technique. "A - new, improved closure technique is being investi- gated, in which Coast Metals brazing alloy No. 52 - will be used to. form a brazed joint that will leave . a minimum void below the tip and will permit Nmteen assemblies were tested eoch a%;T:Ihye;:'c:’r;_x ' : accurate rpositioning'_df_‘_fhe‘ thermocouple junction. This new type of closure.is itlustrated in Figs. 3j:f.— 1.3.17 .and 1.3.18." The test samples for the addi- - tional - tests will 'be""_'fa!:ric_p?ed- with the new ch!ues in purentheses exclude flve nssemblles, N°"."7 brazmg technlque. | 7 A 50-kw Megatherm mducflon ‘heating unit is being used to rapidly heat samples of these Inconel-sheathed thermocouples in order to deter- mine the effect of heating rate on the integrity of 63 ANP PROJECT PROGRESS REPORT " UNCLASSIFIED ORNL-LR-DWG 25836 1100°F 0 -8 MANUFACTURER'S LOWER TOLERANCE -6 & & ® | o E' 1300°F c o0 o = i - -8 Z — z MANUFACTURER'S LOWER TOLERANCE = ) : g > ur O 8 ) -8 ACTURER'S LOWER TOLERANCE -16 1000 hr 2000 hr 3000hr 4000 hr oy g Y3 Y Y3 g Y3 O %> 17 o B 8y TIME Fig. 1.3.15. Drift Curves Typical of Those Obtained for 0.250-in.-0D Inconel-Sheathed Platinum, Platinum-10% Rhodium Thermocouples. UNCLASSIFIED Fig. 1.3.16. Cross Section of Helicrc-Welded Junction of a 0.250-in.-OD Inconel-Sheathed Chromel-Alumel Thermocouple . Showing Inconel Filler Tip, Weld to Sheath, Thermocouple Wire, Void, and Magnesium Oxide In- sulation, 64 i ® the couple. The test sample is mserted in a NaK- filled Inconel charge container made of /,6-m.-0D 0.030-in.-wall tubing. One sample failed after 240 heating cycles between 800 and 1200°F at a heating rate of 100°F/sec. In this failure, both Chromel wires were broken, but the Alumel wires remained intact. In addition to supplying informa- tion regarding the physical soundness of these sheathed couples, the induction heating unit provides a simple means for obtaining aging data PERIOD ENDING SEPTEMBER 30, 1957 The study of drift as a function of time and tem- perature of similar sheathed thermocouples ond conventional beaded thermocouples operating in air has continved. In these tests, all the thermo- couples were opercted continuously at one tem- perature. Three temperatures were investigated, 1800, 1600, and 1300°F. The 1800°F test was terminated after 5000 hr; the test at 1300°F at 6000 hr; and the test at 1600°F is still under way, with data having been accumulated for over 6000 hr. All thermocouples that were tested at under various degrees of cyclic thermal shock. Table 1.3.2, Results of Drift Tests of Sheathed Thermocouples in Fuel Chromel-Alumel Chromel-Alumel Pt, P+—10% Rh PTe?fd Assemblies? Aged 24 hr Assemblies® Aged 200 hr Assemblies® erio (ht) Deviation? Spread® Deviation? Spread® Deviation? Spread® (°F) °F) - P (°F) (°F) (°F) Test Temperature: 1500°F 0 1.9 45 2.9 0.2 -1.0 5.5 1000 3.5 3.8 - 5.1 0.7 0.4 1.0 2000 2.4 7.0 7.3 0.5 ~0.6 2.0 3000 0.2(1.35) 23.5(1y 8.4 0.8 ~1.4 5.2 4000 0.7(2.2)f 22.0(8.5) 8.0 1.0 0.7 3.0 Test Temperature: 1300°F 0 7.0 | 3.0 6.9 0.1 -0.1 2.7 1000 4.3 8.1 6.9 0.9 ~1.2 2.1 2000 2.5(3.2). 14.2(9.0Y 7.8 , 0.4 -0.6 4.2 3000 1.9(3.3) 20.5(10.5) 8.8 0.7 -0.8 4.2 4000 3.8(4.8) 13.7(6.4) 8.8 1.2 -0.6 5.3 ' Tei!_ T_eniperofure: "I'IOOOF o . 42 24 4.6 0.3 ~2.1 4.0 1000 52 43 b4 0.8 -0.3 1.3 ) 2.8(4.8) 1320240 77 1.0 ~0.5" 27 3000 2.4(3.6Y 18. _5(10,0)f 83 06 -0.3 3.2 4000 2.7(3.9')’ 17.3(_7.4)[ ;3.1‘_ 1.1 0o 38 aElghf assemblies were tested' each ussembly comained two thermocouples oged 24 hr at 1350°F in helium prior to testing. PTwo assemblies were tested each essembly contamed two fhermocouples uged 200 hr at 1350°F in helium prior to testmg CThree essemb!:es were tested each assembly contained two thermocouples. dDeviohon of the average of the ihermocouple readmgs from the 1est temperature. ' €Maximum spfeud of the thermocouple reodmgs. . ' ' fvaives in parentheses exclude assembly No. 143. 65 ANP PROJECT PROGRESS REPORT UNCLASSIFIED Yo Fig. 1.3.17. Parts for Completing Brazed Closure at Hot Junction of a 0.250-in.-0OD [Inconel-Sheathed Thermocouple. In this view the sheath has been cut " to show the wires. The four-hole Inconel plug, the Inconel end plug, and the Coast Metals No. 52 brazing alloy ring are also shown. 1800 and 1300°F are being prepared for metallo- graphic examination. The dato token during these tests are summarized in Table 1.3.3. Well Thermocouples The design of a well thermocouple for tempera- ture measurement of high-velocity, high-tempera- ture (1400 gpm, 1100 to 1500°F) liquid metals in Inconel piping has been completed, and several units were constructed. The thermocouples (two per well) employed in these units are fabricated from 20 AWG, solid, Chromel-Alumel wire (ac- curacy, 13/8 of 1% over range 530 to 2300°F) and are welded with Heliarc welding techniques to the bottom of the well. All the ceramic beads used contain a minimum of 96% aluminum oxide. A typical unit is shown before and after assembly in Fig. 1.3.19. 66 Fig. 1.3.18, Cross Section of Brazed Hot Junction. The light gray areas are brazing alloy; the dark gray areas are MgO insulation. Note pores in braze material above the thermocouple wire and the lack of voids below the closure. The well immersion depth was arbitrarily chosen as one-third the inside diameter of the pipe line. The well-body material selected for use with 2'/2-in. and larger pipes was 3/B-in. sched-40 Inconel pipe, and ‘/4-in. sched-40 pipe was selected for wells in 2-in. pipes. The immersed portion of the well body was turned down to a wall thickness of 0.050 in. This well-body design was tested for pressure drop effects on a water loop at velocities up to 1200 gpm. The permanent pressure drop was less than 2.0 psi ot 1200 gpm. The lag length of the well body did not extend beyond the 4 in. of pipe insulation in an attempt to minimize heat losses and the selective oxidation effect that is common in long, narrow thermocouple wells used at elevated temperatures. The thickness of the well bottom was determined on the basis of stresses at operating pressures and temperatures. A value of 0.109 in. was selected on the basis of a stress analysis. o i » o f In order to ensure a reliable and easily re- producible attachment of the thermocouple junction to the well bottom, a technique was developed which permitted the coating of the thermocouple junction with Inconel without undue burning of the wires. This ““wetting'’ of the junction simplitied PERIOD ENDING SEPTEMBER 30, 1957 the subsequent Heliarc welding to the well bottom. The diameter of the spherical junction produced can be controlled to within £0.005 in. with very little difficulty. It was discovered that careful cleaning of oxide from the thermocouple wires facilitates the formation of a reliable junction. Table 1.3.3. Results of Drift Tests of Sheathed and Beaded Thermoeouples in Air ot 1300, 1600, and 1800°F Cl;ll;omel-Alumel Specicl Begded b Chromel-Alumel Sheathed Assemblies? Test Period Assemblies Chromel-Alumel Normal Beaded Assemblies® Pt, Pt~10% Rh Sheathed Assemblies® Chromel-Alumel Sheathed Assemblies? (hr) Deviation! Spread® Deviation! Spread® Deviation! Spread® Deviation/ Spread® Deviation/ Spread® CH R CH CF CF) CH CAH A CH R Test Temperature: 1300°F 0 5.0 3.3 -7.0 2.1 5.3 1000 5.5 3.7 -4.0 3.1 -4.4 2000 5.3 2.4 -3.8 3.5 -4.1 3000 60 25 30 - 4.5 -3.0 4000 6.2 2.0 -2.0 4.5 -2.0 5000 7.8 2.0 0.6 5.0 0 1.1 Measurement error 2.2 0.2 2.0 6.5 -0.5 2.0 5.0 ' ‘ =3.5 5.0 6.0 -3.0 4.0 5.0 Test Temperature: 1600°F 0 3.8 2.8 6.0 2,0 -7 1000 6.4 2.6 3.9 2.5 -2.2 2000 7.0 2.7 6.2 5.5 =0.7 3000 6.0 3.0 120 . 6.0 5.0 4000 13.0 4.0 16.0 5.0 7.0 5000 12.0 7.0 17.3 4.0 0.6 6000 9.9 4.0 15.5 4.0 0.3 3.5 4.7 4.3 6.1 5.6 3.1 10.7 14.0 4.0 14.0 17.0 4.0 18.0 12.0 5.0 22,0 12.0 6.0 24.0 Test Temperature: 1800°F c 10 2.6 ~7.5 4.3 1000 7.7 15.1 5.2 19.2 2000 © 80 . 16.2 26.2 9.5 3000 120 200 30.0 4.0 4000 140 1800 . 300 150 15000 - 161 20 330 150 7' 103 8.7 nz 18.0 21.0 "‘34'.‘_0' ) 2.1 5.5 4.2 Measurement error 14.3 - 10.2 10.5 0.2 4.0 19.2 140 - 140 -9.0 6.5 29.0 16.0 12.0 46.0 140 12,0 68.0 20.0 15.0 “Sn: assemblies were tested cnch ussembly coniuined two thermocouples nged 24 hr ot 1350°F in helium pruor to teshng. bsix assemblles were tesfed ecch nssembly contalned fwo thermocouples fcbricoted wlth especmlly cleaned wire (GCCUI’GCY. +3 é of 1% over range 530 to 2300°F) in ccreful Iy cleaned wells., €Same as speciol ussemblies except that they received no speclul cleaning. - dScme as nssembl:es descrlhed in footnote @ except fhaf they were aged for 200 hr ot 1350°F in helium, - €Five ossemblies were tested; eoch asumbly contained two thermocoupies. fDevmhog of the average of the readings of the fhermocouples tested from the standard test temperature of 1300, 1600, or 1800 EMaximum spreod of the readings obtained. 67 ANP PROJECT PROGRESS REPORT UNCLASSIFIED 859 T T T T UL | EE Jd T " e e 8 SO et S B Fig. 1.3.19. Well Thermocouple Before and After Assembly. The unassembled unit shows the copper chill block for junction formation, the Inconel plate used in making the weld junction at the well bottom, and the plastic tubing used to protect the unit during hendling. The cross sections of junctions shown in Figs. 1.3.20 and 1.3.21 reveal the effect of careful cleaning of the wires. Extreme care was taken in the assembly of the well thermocouples. All ceramic beads were care- fully washed with distilled water and methyl alcohol and baked at 1500°F for 1 hr. All wires were washed with methyl alcohol and air dried. All well-body pieces were degreased, washed with methyl -alcohol, and baked at 400°F for 1 hr. All subsequent handling after the cleaning was done with clean white gloves. After installation o clean, coil-spring type of flexible extension will be provided for each well thermocouple to eliminate 68 any possibility of bending near the region of high- temperature gradients. Drift tests on five sample units mstolled flowing NaK have been started. The test proce- dure will be similar to that used in tests of sheafhed couples. Surface Thermocouples Thermocouples for surface attachment to Inconel piping and components will be fubrlcated in a manner identical to that used for the fabrication of the well couples. A reliable and simple Heliarc method of welding these surface-type couples to " b PERIOD ENDING SEPTEMBER 30, 1957 ¥ UNCL ASSIFIED ] T-10307 i -Alumel Thermocouple Wires. Coated Chromel sn of Inconel- Welded Juncii Cross Section of Heiicn; 20. Chromium oxide deposits may be seen in the Chromel wire. - 1.3 Fig. on © 69 ANP PROJECT PROGRESS REPORT BEY UNCLASSIFIED T-10731 aum, 1 ) fivwwwfl.ré .hu: PN gt & A “.u.. e, w35 A AR o % mw e -Coated Chromel-Alumel Thermocouple Wires Cross Section of Hellarc-Welded Junction of Inconel That Were Carefully Cleaned Before Welding. Fig. 1.3.21. Note absence of oxide deposits. 70 the various Inconel surfaces has been devised by the Metallurgy Division. Apparatus for Checking Thermocouples A melting-point apparatus filled with NaF-CaF, (68-32 mole %) has been completed that is being PERIOD ENDING SEPTEMBER 30, 1957 used for weekly spot checks of the three Pt, Pt-10% Rh couples that have been used as the standards for all the thermocouple tests. The melting point of the NaF-CaF, mixture as meas- ured with these three couples appears to be 1487.5°F and is reproducible to within £1°F. 71 ANP PROJECT PROGRESS REPORT . 1.4. ENGINEERING DESIGN STUDIES A. P. Fraas APPLIED MECHAN!CS AND STRESS ANALYSIS 'R, V. Meghreblian Thermal Stress Cycling Tests of Beryllium D. L. Platus V. H. Kelley J. R. Tallackson The beryllium reflector in the ART will ex- perience thermal strain cycling whenever the power level of the reactor is changed. The most severe strains in the beryllium will occur in the region beneath the load ring. These strains will be produced by the gross radial temperature gradient across the reflector combined with the 30 deg TYPICAL 15 deg TYPICAL ‘9/32—“‘1. R 23%5-in. R 29/32—in. R 1945 -in. R 129%,-in. R / v 1.996 in. local gradients in the vicinity of the coolant holes. Superimposed on this strain field will be the strain concentrations arising from the presence of the holes themselves. Over any extended period of operation these strains will be cyclic in character and can lead to a fatigue crack. A small-scale experiment has been designed to test the resistance of beryllium to fatigue cracking under the tempercture conditions anticipated in the ART. The test specimen is a short, thick- walled cylinder with an axial pattern of small holes (Fig. 1.4.1). The size and locations of the ‘/u-in.-dic holes simulate the cooling-hole pattern in the reactor, and the radial temperature UNCLASSIFIED ORNL-LR~-DWG 25889 NaK INLET FILLER PLUG TO INCREASE FLOW VELOCITY AND IMPROVE HEAT TRANSFER NaK INLET Fig. 1.4.1. Beryllium Thermal-Cycling Test Specimen. 72 gradient across the cylinder has been calculated to produce a gross strain variation similar to that in the reflector. A forced-circulation loop (Fig. 1.4.2) is used to produce the thermal cycles. The beryllium specimen is housed (Figs. 1.4,2 and 1.4.3) so that both the inner and the outer cylindrical surfaces are exposed to flowing sodium. Thermal deformations are produced by the radial temper- ature gradient across the cylinder. Sodium temper- atures are controlled so that the test specimen is cycled between the isothermal and temperature- gradient conditions at Y-hr intervals, as shown in Fig. 1.4.4, The first test will be for a period of 500 hr and will produce 500 complete strain cycles. A detailed study of the stress (elastic) distri- bution in the specimen is under way., Relaxation methods are being used for this study at present, but analytical techniques are being considered. Thermal-Cycling Test of a Welded Core Shell Model B. L. Creenstreet Shells | and Il (the inner and outer fuel annulus shells) for the ART and ETU are to be made from rolled plate weldments which will be machined to the final contour. This method of fabrication replaces that of spinning, The weld joint lo- cations are shown in Figs. 1.4.5 and 1.4.6. As may be seen, each half of shell | has both circum- ferential and longitudinal welds, whereas fhe halves of shell Il have two longlfudmul ones. The matenul in the heat-affected. zones of the_' welds will- exhtblt re!atlvely large grain structure, This material will permit preferential deformation, since it will be weaker under steady load. than “the rest.of the shell.. This strain mtens:flcahon, coupled with a lower resistance to strain cycling " damage, ‘may. result in premature fcnlure. From . this- sfandpomf the most critical regions are the locations " at which the c:rcumferenhcl and longl- - tudmol welds mtersect.r, R cycling condlhons. , the weld pattern. This model is to be tested in PERIOD ENDING SEPTEMBER 30, 1957 ~ the existing core-shell test rig and will be sub- jected to simulated service condmons identical to those applied to the two /4-sca|e specimens tested previously.! This test will establish the the adequacy of the welded reactor shells. Stress Anclfsis of Island Bellows S. E. Moore ~ The geometric configuration of the island ex- pansion bellows of the ART consists of a fourfold repetition of the unit shown in Fig. 1.4.7. The unit is composed of three elements, a concave flange, a very shallow cone, and a convex flange. Adequate methods are presently available for analyzing these shapes,?3 and IBM 650 and 704 digital-computor subroutines have been written for these and other types of rotational shell elements. These routines have been applied to investigate the stress distribution and load- deflection curve for the complete configuration, The design condition for the bellows is 300 cycles of 90-mil axial deflection at a temperature of 1250°F. The stress analysis of the present design revealed that under these conditions the bellows would be subjected to considerable plastic deformation at the outer bend of the convolutions. Correlation with strain-cycling data for Inconel indicated that the fatigue life of the design was less than the proposed 300 cycles, This was verified in a recent test when the ‘bellows failed after 80 cycles.4 The genera! subroutines mentioned. above have " been applied to obtain designs of several bellows which would withstand the deflection and external -pressure (20 psi) onticipated within the reactor. A design has been selected from this group which - consists of four convolutions of 0,062-in.-thick (,fmctérial . The outer diameter of the new bellows will be: % f in, greafer thun that of the bellows of i;,v_an. 1 4 o W “Bell,” ANP Quar. Prog 'Rep. Sept. 10, 195, A Yescale model of shell Il made - with thei{;’ .°R"L 547, p51-52. " shell l weld pattern wull be tested under sframj.,:, ORNL CF-56-12.37 (Aug. 22, 1957). It is ‘expected that the difference in contour between the two shells will have little effect on the distortions arising from. M. Essl inger, Stauc Calc'ulauons of Bo:ler Bottoms, 3See, for example, S. Timoshenko, Theory of Plates and Shells, McGraw-Hill, New York, f940. 4W‘ H. Kelley, ART Island Bellows Test, ORNL CF-57-8-123 (Aug. 29, 1957). 73 ANP PROJECT PROGRESS REPORT o w L vy vy < od Q z 3 - PHOTO 29724 @ «Cycling Experiment Test Rig. Beryllium Thermal 4.2, F'g- 74 9 PERIOD ENDING SEPTEMBER 30, 1957 = ¢ UNCLASSIFIED T.29723 Figs 1.4.3. Beryllium Specimen in Housing. UNCLASSIFIED . ORNL-LR-DWG 25820 1CYCLE ———=1_ BERYLLIUM R _-———4/2hr e iyghy ] " TEST PIECE =~ 1250 — P £ 1200 | — N wl i = 2 Wt F 900 Fig. 1.4.4. Cycle of Temperature vs Time Used for Beryllium Thermal-Cycling Tests. 73 Stress Analysls of Shield Support Structure J L Leggett The two prmcnpal structural components of the "ART ' shield are the lead support structure and the water bag. ‘A detailed stress analysm of the lead suppcrt has been concluded, and “an ‘analysis of the water bag has been initiated. The mu|or structurol member - for SUpportmg the lead is the equatorial shell which fits around the girth of the reactor and carries the equatorial lead (13,937 Ib), The shell is to be fabricated from a ;’lé-in. carbon steel plate to which the lead will be bonded. In addition to providing support 75 ANP PROJECT PROGRESS REPORT . "BECRET ORNL~-LR-DWG 25894 EQUATOR l 10in. (£ Y in) " Fig. 1+4.5. Shell | Weld Pattern. | SEERET ORNL—LR~DWG 26453 EQUATOR Fig. 1.4.6. Shell 1l Weld Pattern. for the lead, the shell will also transmit the weight loads from the north head lead shield (4678 Ib), the south head lead (5475 Ib), and the down-drain lead (6210 Ib). examined in the stress analysis of this structure: (1) the bolts that will transfer the load of the equatorial lead shield to the overhead reactor _support structure, (2) the lugs that will transfer loads to these bolts, (3) the steel backup shell 76 Five creas were UNCLASSIFIED ORNL-LR -DWG 25892 0.501 in.— (D CONCAVE FLANGE (® CIRCULAR FLAT PLATE (@ CONVEX FLANGE Fig. 1.4.7. Elementary Unit of Island Expansion Bellows. (Secret with caption) ' to which the lead will be bonded, (4) the north ring of the equatorial shielding shell, and (5) the stresses in the steel and lead at the bond surface as a result of dissimilar coefficients of thermal expansion, Eight bolts unsymmetrically placed around the north ring of the shell will transfer a load of approximately 30,300 Ib to the reactor support. The stress in the most heavily loaded bolt was checked on the basis of the net cross-sectional area at the root of the threads. Stresses in the welds connecting the lugs to the steel ring and the ribs of the steel backup shell were also examined. The investigation of the steel backup shell was based on the primary assumption that the combi- nation of the steel shell and ribs will carry all the weight of the lead and the steel. The lead was not considered to be carrying any load. The critical point on the shell was assumed to be located 10'/l in. above the equator of the reactors where accommodation of one of the coolant tubes would require that the rib be notched (see Fig. 124 1 ) 1.4.8). The analysns indicated that o %-in. rib combined with a /l -in. shell would suffice, The north ring of the equatorial shielding shell was investigated from the standpoint of bending and torsion. Here it was assumed that the weight of lead and steel would be applied along one edge of the right-angle cross section as a uniformly distributed load. Stresses in the steel and lead as a result of changes in shield temperature were investigated at the bond surface. Two conditions were considered: a slow increase in temperature of 180°F (that would result in ¢ bond temperature of 250°F) to simulate the approach of the reactor to power operation and @ sudden decrease of 120°F (a bond temperature of 130°F) to simulate shutdown of the reactor. An elastic analysis based on these operating conditions revealed that the stresses in both the steel and the lead ot the bond would remain within acceptable limits. An analysis of the aluminum water bag is presently under way. Since the north head is made of 10 or 12 segments and the equatorial section of 5 segments, it will be necessary to examine the integrity of the individual pieces as well as the whole unit, Preliminary analyses UNCLASSIFIED ORNL-LR-DWG 25893 . 287 in, 47in, J y _ } ¥ ¥ i NOTCH £ @ = _ Q | - | ~EQUATOR Yo"\l 5d - - 29in. - — Ny 329 - o / : 30.3in. : - N - 43-inLEAD\ [fis70] E-) . \,b /16 in. STEEL: y o . ' . 7 € REACTOR - 0/03 1-inRIB A | Vv r bt 24,38 in.—————— e———————— 28.75 in, —————( 188010 Fige 1.4.8. Steel Backup Shell for Equutfiricl Lead Shield. {(Eonfidenticlwith-eapticr) PERIOD ENDING SEPTEMBER 30, 1957 indicate that it may be necessary to fit the segments together carefully in order for the assembly to act as o single structure, This will also require shear attachments along the joints between segments on both the inner and the outer surfaces of the bag. ' Stress Anclysis of a Pressure-Measurmg Instrument B. Y. Cotton An analytical study was made to determine the elastic stresses in a pressure-measuring device manufactured by the Taylor Instrument Company (Fig. 1.4.9), The instrument consists of a riser, two flat circular plates, a cylindrical portion containing the weldment, and a pressure-trans- mitting capillary. A sandwich-type diaphragm of five layers of 0.005-in,-thick Inconel sheet sepa- rates the fluid being measured from the trans- mitting fluid, Under the specified operating temperatures, creep is on important consideration, and the pressure forces on the circular flat plates will tend to distort them into spherical shapes. If the pressure forces are applied for extended periods of time, the creep deformation will cause an appreciable increase in the internal volume of the instrument and will result in instability or drift. The anticipated operating conditions for the instrument vary widely, since the instruments are to be employed in various test loops in different fluid environments, pressures, and tem- peratures. The results of the stress analysis are summarized in Teble 1.4,1 in order to provide a basis for evaluating particular applications. The _indicated estimated life is the time required to - Table 1.4,1. Life Span of a Pressure-Measuring - Device Based on 0.2% Creep Deformation Internal Pressure Maximum Stress Estimated Life . . at 1300°F (psi) : (psi) (hr) 25 | 3,546 2000 50 S 7,092 . 50 100 14,185 0 200 28,371 0 77 ANP PROJECT PROGRESS REPORT i il ks S A A— — ——— — ————— ——— UNCLASSIFIED ORNL—-LR-DWG 25894 SANDWICH-TYPE DIAPHRAGM TRANSMITTING FLUID \ HIGH-STRESS REGION s~ FLEXIBLE TUBING CAPILLARY Fig. 1:4.9. Pressure-Measuring Device. exceed 0,2% creep elongation. The analysis indicated that this device should not be used at 1500°F and above under the stress conditions indicated in Table 1.4.1. In applications where the dimensional stability of the unit is not the most important criterion, testing would be de- sirable in order to demonstrate the extent of the useful life. APPARATUS FOR ETR IN-PILE TESTS OF MODERATOR MATERIALS G. Samuels A preliminary layout of an apparatus for in-pile testing of moderator material in the temperature range of 1500 to 2000°F has been completed. The apparatus is designed for testing six moderator 78 specimens at one time. The specimens will be 3 in. long ond will vary from ‘/2 to 1 in. in diameter. The initial moderator materials to be tested are beryllium oxide, yttrium hydride, and beryllium metal. The purpose of the tests of yttrium hydride and beryllium oxide is to determine their maximum allowable thermal stress under various reactor operating conditions. In the case of the beryllium specimens, the thermal stress consideration is secondary to the study of compatibility with various canning materials. The moderator materials will be canned in Incone! jockets, and, where necessary, a third material will be used between the moderator and jacket. The beryllium will be separated from > L} the Inconel by either chromium, molybdenum, or an oxide coating on the beryllium. For yttrium hydride, the intermediate material will be mo- lybdenum. Beryllium oxide will not need the buffer material. The jacketed moderator specimens will be centered in a water-cooled nicke! sleeve. The clearance between the specimen and sleeve will vary from 0.015 to 0,060 in. depending on the moderator material, the diameter of the specimen, the location in the reactor core, and the desired minimum surface of the specimen. The heat from the specimen will be removed by radiation and conduction to the nickel sleeve. Helium or o mixture of helium ond orgon will be bled slowly through the gas space. For normal operating temperatures {about 1500°F), 80% or more of the PERIOD ENDING SEPTEMBER 30, 1957 heat flux is carried by conduction so that the surface temperature will be very sensitive to the argon content of the gas, which will thus provide a convenient means of controlling the temperature, Two thermocouples will be spot-welded to the jacket of each specimen at equal distances from each end and 180 deg apart. In addition each BeO specimen will have one thermocouple located in the center of the specimen at a distance of 3’8 to 3, in. from the end. Only one specimen each of yttrium hydride and beryllium will contain a " center thermocouple during any test run, Knowl- edge of the tempercture difference between the center and the surface of the specimens will facilitate the determination of the internal hect generation rate and thermal stress, W o 56 i 79 ANP PROJECT PROGRESS REPORT 1.5. DESIGN PHYSICS CAM. Perry ART FILL-AND-DRAIN TANK SHIELDING H. W. Bertini The general method of attack on the problem of shielding the ART fuel fill-and-drain tank was to choose a field point on the surface of the shield and calculate the dose at this point from radiation emitted from the surface of the tank. ' The sources at the surface of the tank were considered to be point sources of magnitude Sv ' - "~ Mev — cos 0 x,A(———-——) , 4p sec-steradian S = Mev/sec.cm? of fuel volume, g = total linear gamma-ray cross section of the fuel (cm="), 6 = angle between the normal to the dump tank surface and the line between the field point and the point source, A = area of the dump tank surface represented by a point source at its center. The (S, /4mu) cos 6 term is the surface source strength at the interface between a thick slab source and a thick slab shield that will give, essentially, the correct dose rate at the surface of the shield. To the extent to which it was completed, the lead shield was designed so that the dose rate at any point on its surface was less than 0.2 r/hr after 100 hr of operation at 60 Mw, 9 days after shutdown. Under these conditions, S was calculated for a fuel volume of 2,5 x 10° em®, and the curves of decay energy after shutdown given by Moteff! were used. The 1,65 and 2,55-Mev groups were the only ones used in these calculations, and the p volues for the fuel were taken from ORNL-2113 (ref 2) for these energies. In all cases, cos O was set equal to 1. J -Moteff, M:sce!laneous Data for Shielding Calcu- latzons, APEX-176 (Dec. 1, 1954). 2H. W. Bertini et al., Basic Gamma-R Data {or ART Heat Defosz'tion Calculations, ORNL-2 3, p 61 {Sept. 17, 1956). 80 Penetrations ‘The problem of penetrations in the shield is aggrovated by the need to allow for relative motion of the tank, all pipes connected to'it, and the tank shield. All penetrations through the shield are therefore larger than would ordmurlly be anticipated. ' In the region where the lead shield for the drain line meshes with that of the tank the dose rates at 16 field points from 6 point sources were calculated by using the technique described above. In addition, the dose rate ot a point at the upper- most surface of the tank shield from single scattering in the lead around the drain line was calculated and found to be within design limits. At another point in this general region, where only 3 in. of lead separates it from about :?4 in. of unshielded drain line, the dose rate was cal- culated and found to be acceptable, The valve actuator penetration through the lead allowed the dose rate in this vicinity to be rather high. A 2-in.-thick, 8-in.-dia disk of lead fitted around the actuator arms in this region would reduce the dose rate to acceptable limits, The pipes for carrying the NaK to heat and cool the fuel cannot be allowed to penetrate straight through the shield, because the dose rate at such penetrations would be prohibitive. A feeling for the magnitude of the problem can be obtairi¢d by noting that the dose rate at the surface of the tank is of the order of 2 x 10° r/hr 9 days after shutdown from operation for 100 hr at 60 Mw. By repeated calculations of the direct and reflected radiation and the radiation incident at slant angles it was found that the geometry sketched in Fig. 1.5.1 would meet the shielding requirements for penetrations in the tank. The direct radiation was calculated by using the point source concept outlined above. The reflected radiation was calculated by using the data given by Rockwell® to estimate the angular distribution 37, Rockwell, NI (ed.), Reactor Shielding Design Manual, T|D-7004, P 330 (March 1956). ™ i 4. ~ dose rate was ca_lcul_ated to be about 2 x 108 t/he. SESREF ORNL —LR—DWG 25835 NoK LINES LEAD SHIELD ——\[ = ——TANK MANN Fig. 1.5.1. Top of ART Fill-and-Drain Taenk and Shiefd Showing Recommended Configuration for NaK Lines Penetration, of reflected radiation and the data given by Zerby* to estimate the fraction of the incident radiation that is reflected. The data of Rockwell do not correspond to ART fuel fill-and-drain tank ganima-- ray energies or material, and the data of Zerby do not correspond to ART fuel fill-and-drain tank gamma-ray energies. This calculation appears to be about all that can be done at this time, how- ever, and the calculated dose rate is probably in error by a factor of the order of 2. The dose rate for radiation incident at slont angles was col- culated from the work of Peebles.® v At the bottom of the reactor, the fill-and-drain tank support penetrates the lead shield. About 2 in. below this penetration the dose rate was calculated to be ‘about 2.5 x 104 r/hr._ In the_ vicinity of the top_of the air cylmder for the tank support, about. 29 in. below the penetration, the Overlcp m Shleld at Removcble Top Section B In order ‘o facnhtate assembly ‘and dlsassembly, it was necessary to deSIgn the tank shleld '$0 4C.D.Zerb Transm:sszon o[ Obl:quely lnczdent Gamma-Rad:‘atz‘on Tbrousb Stratified - Slab’ Bamers,- ORNL-2224 vol 1 (Nov. 1957). 56. H. Peebles, Gamma-Ray 'I'ransm:sszon Through Finite Slabs, R-240 (Dec. 1, 1952). 9, 1956) and vol 2 (Jan. 2 PERIOD ENDING SEPTEMBER 30, 1957 that the top portions of the lead would be re- movable in sections. The problem was to de- termine whether any overlap of these sections would be required when they were in place on the tank. - By using the method descrlbed above, the dose rate at a point opposite a /l -in. crack in the lead which would expose the tank surface was calculated to be about 120 r/hr. Therefore a clearance of this size. separating the sections of the lead would be much too large. | If the sections of the lead were brought togefher with no c!earance, but in such a manner that the %fi-m. steel cladding layers around each lead section touched and formed a / -in. slab of iren between the adjoining lead sechons, the dose rate at the outer surface of the slab of iron would be about 2 r/hr, which is also not acceptable, it was therefore concluded that all adjoining pieces should be overlapped ot about their mid- thickness. Incompleted Work Additional calculations are required for the shielding of the small lines which connect with the upper region of the foce of the tank and penetrate the shield. Calculations are also required to determine the dose rate near the bottom of the shield as a result of scattered and direct radiation through the penetration in the lead required for the tank support. DOSE RATE IN REGION OF ART PUMPS H. W. Bertini- C. M. Copenhaver In order to ascertain the feas:blhty of having ‘a man_enter the ART ‘reactor cell to unbolt an ‘inoperative fuel or sodium pump for removal, the _dose .rate was calculated at a point, P, 62.5 in. ' _above the reactor midpoint and 22.5 in. from the axis, such point being in the vicinity of the bottom of the pump motor. The condition chosen was 9 days cfter shutdown followmg 100 hr of reactor operation at 60 Mw.. ‘The sodium activation contnbutlon was found ~:-to ‘be -0.23 r/hr, .of which 0.05 ¢/hr. would be ~ through lead, 0.13 r/hr would be through NaK pPipe - penetrations to the sqdium-to-NaK_' heat ex- changers, and 0.05 r/he would be through other penetrations. The corresponding sodium dose rate at shutdown is 5000 r/hr. 81 ANP PROJECT PROGRESS REPORT The fuel contribution would be principally from residual fuel pockets whose probable volumes and locations are presently not well known. The fuel pumps, the fuel onnulus around the control rod in the sodium expansion tank, etc., would con- tribute about 3 r/hr, with about 2 r/hr being through the fuel pump . penetrations. On the assumption that 2% of the total fuel in the reactor system would remain uniformly distributed in the main fuel-to-NaK heat exchanger upon draining of the reactor, the dose through the NaK pipe penetrations would be about 500 r/hr. The off-gas line will pass very close to one of the fuel pumps and part of it will be visible through the pump penetration. A total of 3 x 1015 Ba'4® nuclides per second® was assumed to pass into the off-gas line and deposit uniformly on its surfaces. If it is assumed that a 3-cm length of off-gas line will be directly visible through the penetration, the dose rate was cal- culated to be 260 r/hr from this source. A summary of this preliminary calculated dose rate at P is given below: Source Dose Rate (¢/hr) Sodium 0.2 Fuel pockets 3 Residual fuel in heat exchanger 500 Off-gas line 260 DECAY OF GAMMA ACTIVITY IN U235 CONTAINING ART FUEL FOR SHORT TIMES AFTER SHUTDOWN J. Foster The rates of decay-gamma energy emission from U235 fuel for the first 180 sec after reactor shutdown following 500 hr of operation were calculated for six gamma-ray energy groups. The results (for the six energy groups) were plotted (Fig. 1.5.2) as Mev/sec for an operating level of one fission per second as a function of time after shutdown. Although data have been available for the rate of decay of gamma activity for times of a few minutes or more after shutdown, no data have been available up to this time for the period 66, Samuels, private communication to H. W. Bertini. 82 Py -LR-DWG 25896 DECAY ENERGY [(Mev/sec)/(fission/sec)] 2 5 2 5 ¢ 2 5 10 2 5 2x10 TIME AFTER SHUTDOWN (sect Fig. 1.5.2. Decay-Gamma Energy vs Time After Shut- down for Uzss-c‘\miulning Fuel Operated for 500 hr at 60 Mw in the ART. Decay energy calculated for the six gamma-ray energy groups indicated. immediately ofter shutdown. These short-time decay rates are important in calculating heating rates and cooling requirements for such system components as the fuel drain valves. Results of this study yield a total decay-gamma energy of 5.83 Mev fission at 10=3 hr after shutdown.” DECAY-GAMMA HEATING IN ART FUEL DRAIN YALVES J. Foster Calculations were made of the heat deposition rates in the bellows and stem of the ART fuel drain valves by decay-gamma energy originating in fuel draining through the valve immediately after shutdown following 500 hr of 60-Mw oper- ation. ' : The total decay-gamma energy deposition rate at the instant of shutdown was calculated to be 2.5 w/cm® of Inconel, which corresponds to a rate of temperature rise of about 1°F/sec. |f it is assumed that fuel remains in the valve for 3 min, which is the maximum allowable duration of the fuel draining operation, and the decrease in gamma activity during this period is aliowed for, a total temperature rise of 106°F is obtained. No allowance was made in the calculation for heat loss from the Inconel piping during the 3-min pericd. ' : 7The basic experimental data were taken from interim unpublished work of W. Zobel oand T. A. Love cf ORNL and were modified by C. Copenhaver of ORNL, ' y W PERIOD ENDING SEPTEMBER 30, 1957 1.6. MATERIALS AND COMPONENTS INSPECTION A. Taboada MATERIAL INSPECTION A. Taboada G. M. Tolson Tubing Approximately 3 miles of tubing, 15,827 ft, was inspected during the quarter, and, of this omount, 4126 ft (26%) was rejected. The rejected tubing included 1109 ft that can be salvaged by pickling; 1093 ft of the rejected tubing was returned to the vendor for credit. material was downgraded for Iess critical ap- plications. included in the. tublng mspecfed was 128 ft- of INOR-8 tubing of which 5 ft (4%) was rejected. Typical defects found on the INOR-8 tubing are- shown in Figs. 1.6.1 and 162 ond a defect not previously observed is shown in Fig. 1.6.3. As mentioned previously,! a phosphoric acid pickling method has been developed for removing ‘G M. Tolson, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 75. 0.25in. L ¥ 1 ‘The remaining rejected . magnetic particles stuck to the inside of CX-900 subing. In order to determine whether there would be any harmful effects from the pickling solution, a sample of Inconel tubing was immersed in phosphoric acid for 160 hr and then examined metallographically. No attack was found, and it ‘therefore appears that the pickling method may be used satisfactorily for the removal of magnetic particles, A large number. of oufer surfuce scratches are found during the inspection process, and, although the scratches are usually superficial, they interfere with subsequent ultrasonic tests. Investigations have revedled that intermittent polishing and re- “polishing can reduce the rejection rate substan- tially. Tubing rejected previously by ultrasonic inspection is being re-evaluated. Pipe, Plate, Sheet, Rod, and Fittings A total of 532 ft of Inconel pipe was inspected, ond 73 ft (14%) was rejected. Also, 1027 f? of UNCLASSIFIED T-12859 Fig. 1.6.1. Photomacrograph of INOR-8 Tubing Showing Two Cracks Which Penetrate Completely Through the Wall of the Tube. The cracks were found by radiographic inspection, 5X. 83 ANP PROJECT PROGRESS REPORT 0.25 in. y ~ UNCLASSIFIED ! T-12860 Flg. 1.6.2 Ph'dfbmacrogruph of VGouge Found on Inside Surface of INOR-8 Tubing by Radiographic Inspection. 5X. plate and sheet was examined, of which 47 2 (4.6%) was rejected. The inspection of 81 ft of rod resulted in the rejection of 18 ft (22%). There were 41 f{ittings inspected, and 5 were rejected. Embrittlement of Inconel by Penetrants An experiment, described previc:msly,2 was con- ducted to determine whether the solutions used in penetrant inspection would couse embrittlement of Inconel at high temperctures. Metallographic examinations of samples which had been heated at 1500°F for ¢ period of 50 hr showed no embrit- tlement or attack as a result of exposure to the penetrants. ' ' Chemical Analyses of Materials Chemical anclyses of materials received for inspection are now being obtained by spectro- graphic techniques. A comparison of spectrographic results with the standards established by wet chemical analysis indicates an accuracy of +10%. 21bid., p 80. 84 Electrical Calrods Radiegraphic inspection was made on 1964 ft of electrical Calrods in order to determine the positioning of the conduit. Those units which showed improper positioning or faulty conduit were either rejected or downgraded. WELD INSPECTION A. Taboada R. L. Heestand Welds fabricated on ART and ETU components and on experimental test loops by the inert-gas shielded-arc method were inspected. The defective conditions found by inspection were porosity, cracks, tungsten inclusions, lack of fusion, and iack of penetration. Several welds were examined which had been fabricated on core shells for test purposes. In general, these welds were found to be acceptable for the intended use; however, in some instances defects were noted that would have been cause for rejection or repair if they had been found on ETU or ART parts. PERIOD ENDING SEPTEMBER 30, 1957 0.25 in. UNCLASSIFIED T-12868 " Fig. 1.6.3. Photomacrograph (a) and Photomicrograph (5) of. Defective INOR-8 Tubing Located by Radiographic lnspeeti!fiou. (a) 8X. (&) S0X. Etchant: giyg:e;egid."””- 7 85 ANP PROJECT PROGRESS REPORT INSPECTIONS OF REACTOR SHELLS A. Taboada G. M. Tolson Inner Core Shells (Shel_l 1) Radiographs of inner core shell sections made by forming, welding, and machining were found to have *‘white lines’’ in the weldments. Metal- lographic examinations showed the lines to be associated with large dendritic grains that ex- tended through the entire weldment. Outer Core Shells (Shell 1) Inspections of spun outer core shells described previously as “‘in process’’ were completed. The results of all the inspections are presented in Table 1.6.1. It is to be noted that none of these shells are to be used in the ETU and ART. Complete in- spection was performed in many cases for checking manufacturing techniques and quality of the finished parts and for obtaining data for use in other tests. Shell 1il The bottom section of core shell Il was in- spected by radiography following an intermediate machining operation. One weldment was found to contain pores and areas that were contaminated. No repairs were made on the weldment because the poor areas may be removed during subsequent machining. The shell will be inspected again in the final form. A Shells IV and V Shell material formed by deep-drawing in the Y-12 shops was examined to investigate the method that is to be used by Kaiser Metal Products in the fabrication of shells IV and V. Specimens consisting both of weld deposits and of the parent material were found to be acceptable. Additional specimens of as-received material and of annealed material (1800°F for 10 min) which haed ruptured during fabrication ‘were received from Kaiser Metal Products. Rockwell hardness tests were performed (Table 1.6.2), and the ductility of the material in various stages of the Table 1.6.1. Results of Examinations of Spun Outer Core Shells Metallographic Nonde structive Lot No. Piece No. Examination of Sample Inspection of Shell 1 1 Not required Rejected for surface defects 1 2 Not required Rejected for defects 1 5 Rejected because of Not required overpickling* 4 12 Accepted* Not reguired 3 | 1 Rejected Thin wall areas 4 Accepted Rejected by visual inspection because of overpickled con- dition 4 8 Accepted Not required 4 n Accepted Not required 4 10 Rejected Thin wall areas 4 4 Accepted* Not required 4 7 Rejected Rejected on visual and dye- penetrant inspections; thin wall areas *Metallographic specimens obtained from cut-up shell. 86 Table 1.6.2 Hardness and Ductility of Inconel Specimens Recelved from Kaiser Metal Products Company = . Hardness, Estimated Elongation Specimen Rg (% in 2 in,) As received* 80 41 As drawn* = 93 . 22 Annealed | . 73 ‘ . 43 *A control sample of this material has been taken for subsequent comparisons. shell-forming process was extrapolated from the - hardness data. Although data of this type are reasonably accurate, tensile and elongation tests are also being performed, since ‘large variations. in mechanical properties are usudlly found in different heats of the material and occasionally" in the same heat. It may be noted from Table 1.6.2 that the hardness of the annealed material is not so low as that sometimes found in deep- drawn good-quality Inconel. Therefore it has been recommended that the material be annealed at 1800°F for 20 min in the future. INSPECTION OF COMPONENTS RECEIVED FROM YENDORS R. L. Heestand G. M. Tolson York Corp. Radiators und Radiator Matenuls ‘Radiegraphs of - the tube-to-header ‘welds of - finned-section units 9, 11, 12, and 14 of the main - ETU NoK-to-gir radiators bemg supphed by the York Corp. were reviewed to assure ‘that they' had been: properly interpreted. Rejection of units 9 ond 11 by the ORNL inspector stationed at the York .Corp. was affirmed; unit 12 was found o be acceptable; and recommendations - were made - for. the repair of two porous.welds on unit 14, Radiator - units 18 and 19 were inspected and " found to.be- ucceptable, and unit 20 is bemg ; .mspected _ = - Two " l-Mw test rad:ufors, units 2 und 3, were recelved for inspection, Unit 2 was found to be acceptuble, but it was found that the return bends of five of the tubes of unit 3 had been domaged during installation of a sawing fixture for the PERIOD ENDING SEPTEMBER 30, 1957 slitting of the fins. The deepest defect was repaired by welding. A header piece for the south head of the main - fuel-to-NaK heat exchanger that had been fabricated for stress onalysis tests was also inspected. Several pinhole leaks found in the brazed tube- to-header joints were repaired by the ORNL Welding and Brazing Laboratory by flowing Coast Metals brazing alloy No. 52 over the pinholes with on acetylene torch. The necessary inlet and outlet tubes were also installed, and the unit was submitted for testing. inspections of samples of stainless-steel-clad copper fins on which the cladding appeared to be blistered revealed a thin layer of oxide on ‘the stainless steel surface that was of sufficient thickness to prevent bonding. It was recommended “that all fin material thot showed defects of this . type be rejected. ' Control samples, which duplicated units 5, 6, 7, 8, and 12, were examined metallographically before and after exposure to air at 1500°F for 100 hr. Control samples 5, 6, 7, and 8 were acceptable, but sample 12 showed lack of flow of the brazing alloy, and 60% of the fin lips were found to be unprotected after oxidation. ‘Recommendations were made to the York Corp. for the use of jigs and fixtures in making the -welds of tube-to-header joints in order to prevent the braze cracks found on ETU radiator unit 1. . The eracks resulted from excessive weld shrinkage. ‘A procedure for the repair of the cracked unit ~was also recommended. Spacers for use during brazing were prepared and sent to York Corp. These Inconel X spacers were fabricated by rolling and processing in a hydrogen atmosphere for 30 ~min at 1950°F; 200 spacers were prepared. .Griscom'-Rus'séll Co. Heat Exchangers Sodlum-to-NoK heat exchanger vnits 1 and 2 '._‘Zfabrlcated by Griscom-Russell Co. were submitted ~ for .inspection prior to .installation in the ETU - north. head.. Root passes of the welds on each head of both heat exchangers showed lack of _ penetration because of shrinkage and inadequate - repair. -work. Cleanup passes will be made on future “units to eliminate defects of this type. ‘Since no defects serious enough to cause failure of the units were found, they were accepted for use in the ETU system. 87 ANP PROJECT PROGRESS REPORT. A K-fin radiator job sample was examined and ~ rejected because of a manufacturing defect imposed on the tube by the winding of the fin. It was re- quested that the tube-sheet design for the K-fin radiator be revised because there are areas in the present design which would be stressed beyond the yield strength of the Inconel. The foliowing procedures and samples were examined and opproved: (1) o tube-to-header weld and braze sample with eight tubes; (2) three tube- to-header welder-qualification samples; (3) a fuel- to-NaK heat-exchanger-header assembly sequence; (4) a dummy sodium-to-NaK heat-exchanger-header assembly; (5) a joint change of a sodium-to-NaK heat-exchanger nozzle-to-header weld; and () a sodium-to-NaK heat-exchanger header for stress- analysis testing. Black, Sivalls & Brysoh Heat Exchengers A resistance technique for bending the fuel-to- NaK heat exchanger tubes was approved after destructive and nondestructive exomingtion of ten tube samples submitted by Black, Sivalls & Bryson. North and south header stress analysis samples were received, and after repairs of several cracks, the north head sample was accepted for testing. The south head stress analysis sample is being inspected. The following reports, procedures, and samples were examined and approved: (1) radiographic techniques for examination of tube-to-header welds; (2) a report covering the modification of the channel-to-web weld joint; (3) samples of and a report on the bulkhead weld; (4) a report on tube- to-header brazing covering the use of rings con- toured to fit the header; (5) braze samples 1 and 2 that illustrated vertical and horizontal brazing; (6) proposed welding techniques for the channel, flute filling, inlet pipes, outlet pipes, and the bulkhead-to-tube-sheet joints. Midwest Piping Company Forgings end Weldments Nondestructive and destructive examinations were completed on samples of hot-forged Inconel re- ducers which were fabricated at temperatures of 1800 to 1850°F by the Midwest Piping Company. Metallographic examination revealed the grain size to be no greater than No. 5, and therefore the fabrication method was approved. An Inconel pipe saddle weldment was examined and approved for design and weld integrity. Also 88 a 90-deg bend in a piece of 3‘/2-in. Inconel pipe was checked for wall thinning, flattening, and other defective conditions and was found to be satisfactory. ' o ’ : Process Engineering, Inc., Tanks and Bellows Expansion Joints Inspection was completed on standard NaK tank No. 2 received from Process Engineering, Inc., and it was found to be acceptable for the intended use. Four other tanks have been received and are being inspected. Twelve bellows expansion joints for the ART were also received for in- spection. ' Fulton-Sylphon Company Bel lows A sample ETU sodium expansion bellows was received from the Fulton-Sylphon Company for evaluation of fabrication techniques, Examination - revealed that the seam welds on each end of the unit were porous. Metallographic examination re- vealed dirt imbedded between each of the three plys of the bellows. In the weld metal, pores up to 0.005 in. in diameter were found. It was recom- mended that more care be taken in the fabrication of the bellows to eliminate sources of dirt and contamination. Hoke, Inc., Inconel and Stainless Steel Valves Fifteen Inconel and fifteen type 316 stainless steel valves, which have two welds each, were received from Hoke, Inc., for acceptance of the welds. Inspection resulted in the detection of five defective welds., These welds were repaired and were found to be acceptable upon re-examination. METAL IDENTIFICATION METER A. Taboada G. M. Tolson The examination with the metal identification meter of an ART-type fuel pump after test operation indicated that a material other than Inconel had been used. Further tests with the metal identifi- cation meter and a Spotchek chemical kit revealed that the labyrinth seal was fabricated of a 300- series stainless steel. Since material other than Inconel has been detected in other components by this equipment,® two additional meters have 3G, M. Tolson and R, L. Heestand, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 78. i ] been ordered for adequate inspection to ensure that only Inconel is used in the ETU and ART. The meter now available is being used to check all components being received from vendors. The second meter will be used in the Y-12 Plont General Shops in order to check all incoming materials and to check all components that are fabricated there. The third meter will be used to check all material that is welded in the welding shop established for the assembly of the ETU and ART. INSTRUMENTATION AND CONT ROLS INSPECTIONS Nlckel-Copper Welds in quuld-Level Probes Several ‘failures- have occurred in the nickel< copper welds in liquid-level probes. A microscopic - examination of one of the failed welds revealed - stringers of copper oxide adjacent to the ‘failure zone. Radiographs of several other probe welds showed large pores, whlch indicate high. oxygen content of the copper. failed are being examined to obtain comparative information that will be useful in further investi- gation of the cause of failure. A liquid-level indicator which cracked and Ieuked in five areas was found to have failed because of ratcheting of the ceramic insulators. The tube failures were typical of tube-burst test failures under biaxial loading. A second liquid-level in- dicator failed because of leakage through a pin- hole. The initial radiographs taken of this area were re-examined, but no indication of the defect was found. Since the unit had also passed dye- penetrant inspection, the pinhole probably origi- nated in the inner lead ~weld which- ‘was’ not accessible for dye'Penetrcmt inspection. ~ Thermocouple Welds A test fhermocouple which failed durlng testmg - in NaK at 1500F was examined mefollographucall; The failure was found to have occurred in the closure weld of the thermocouple sheath and was due to lack of penetration of the weld. - At ‘the point of fallure ‘the magnesium- oxide insulating material had been leached to a 'depth of approxi- mately Y% in. during the 5000 hr of exposure to NaK ot 1500°F, and NaK had diffused through the insulating material approximately 10 in. No Probes which have not = PERIOD ENDING SEPTEMBER 30, 1957 deterioration of the metal parts of the unit could be seen. The failure zone is shown in Fig. 1.6.4, and the thermocouple assembly is shown in Fig. 1.6.5. AR LUNCLASSIFIED T-12895 Fig. 1.6.4, Sectioned Thermocouple Assemblies After Service in NaK at 1500°F. operated satisfactorily for 5000 hr, The unit shown The unit shown in (a) in (b) failed because of a crack in the closure weld on the sheath., Note removal of insulation by reaction with NaK. The dark insulation visible in (b) is satyrated with NaK. UNCLASSIFIED T-13094 \‘<\\ FAILURE AREA SHOWN IN Fig.1.6.4 2 ' & WNEA W\ kN %‘\ ‘P:\\O‘A‘ g\\b\}\ 6 N ?E‘A \\ o S o Fig. 1.6.5. Thermocouple Assembly Which Failed During Exposure to NaK for 5000 hr at 1500°F, 89 ANP PROJECT PROGRESS REPORT PROCEDURES FOR RADIOGRAPHIC INSPECTION OF WELDMENTS ON ART ISLAND AND REFLECTOR A. Taboada G. M. Tolson There are seven critical welds on the ART and ETU which can be radiographically inspected only through sections of beryllium. Normal radiographic techniques .cannot be used because beryllium scatters the .incident x-’rqy beam and causes general fogging of the film. Experiments are being 90 run in order to determine techniques for obtaining maximum sensitivity. o Several radiographic fechniques ‘were used in the inspection of the equatorial weld on the ETU island shell assembly, but it was impossible to estimate the radiographic sensitivity obtainable by these methods. A mockup of the assembly with known discontinuities was therefore obtained, and the sensitivities of the inspection methods are being investigated. ' PERIOD ENDING SEPTEMBER 30, 1957 1.7. HEAT TRANSFER STUDIES H. W. Hoffman THERMAL.CYCLING RESEARCH Pressurized System H. W. Hoffman D. P. Gregory! The experimental investigation of the effect of thermal-stress cycling on Inconel tubes filled with flowing NaF-ZrF -UF, (50-46-4 mole %) was continued with the use of the pressurized-flow system. Data were obtained at temperatures near 1600°F for tube bends, tube welds, and straight 10n assignment from Pratt & Whitney Aircraft. tubes cycled at frequencies of 0.4 and 1.0 cps. The results of the experiments performed during the quarter are presented in Tables 1.7.1 ond 1.7.2 and are summarized in Fig. 1.7.1. Straight tubes were tested at a cyclic frequency of 1 cps for comparison with the data of experi- ments made at frequencies of 0.01 and 0.4 cps. As previously reported,? the tests ot 0.01 and 0.4 cps frequencies indicated no significant effect of cycling frequency on depth of attack. However, 24, W. Hoffman and D. P. Gregory, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 82. Table 1.7.1. Results of Thermul-CycIIng Test of Inconel Tubing Containing Flowing NaF-ZrF -UF, (50-46-4 mole %) Cycling tate: 0«4 cps Tubing size: 1{‘«tn. 0D, 0.035+in. wall Heater length: 8 in. Test Conditions ET-27% _ ET-29® ET-30 ET-31 ET-32 ET-33 ET-34 ET-35 ET-36 ET-37 Duration of run, hr 4 n 65 30 25 54 19 45 11.5 15 Heater section Outer wall temperature, °F Average 1597 1795 1725 1769 1767 1763 1716 1700 1664 1648 Fluetuation g5 116 197 197 196 tg9 177 180 170 Maximum depth of attack, mils General subsurface veids 2 2 4 6 intergranular attack 3 4 4 5 105 & 8 Test section Outer wa!l temperature, °F : B . : ~ Average 1597 1623 1604 1590 1613 1620 1598 1626 1590 1584 Fluctuation - 3 - #1314 16 113 t15 15 e e Maximum dqfith of attack, mils e Generel subsurface voids - 2 5 4 4.5 7 3 35 . Intergranular attack 3 . 5. 4 35 7 6 - 35 inlet mixed=mean fused~salt 1597 ‘1573 1578 1573 | 1574 1573 1542 1598 1555 1562 temperature, °F ' ' _ Cuu§§ of test termihatlon Voluntary® %) sothermals bHigh-frequency cycling rate: 1 cpse €All other runs were terminated by instrument failure. 9 ANP PROJECT PROGRESS REPORT Table 1.7.2. Results of Thermul-CicHnfi Tests of Bent Inconel Tubes Containing Flowing NuF-ZrF(UF‘ (50-46-4 mole %) Cycled at a Rate of 0.4 cps Attack in Straight Pertion Attack in Bends Comments | Run Description _— s ‘ Depth Type of De::th TAY pe :f {mils) Attack (mils) Hac ET-30 Prior creep, fine grained, 3.5 General 4 General in all ‘Equ'él.atfocl_: on both tension and four 90-deg bends bends compression sides of each bend ET-31 No prior stress, fine 3.5 lnfergrcnulér 4 Intergranular Deepest attack found ne&r center - grained, four 90-deg in all bends of each bend with no difference bends " between tension and come pression sides of bends ET-32 No prior stress, fine 4.5 Intergranular Tendency toward slightly deeper grained : ' intergranular attack on tension Two 30-deg bends 6 intergranular sides of bends Two 150-deg bends 5 Heavy inter- & granular ET-33 No prior stress, coarse 7 General 8 General in ail Equal ottack on both tension and grained, four 90=deg bends compression sides of bends bends ET-34 No prior stress, coarse 3 General 3.5 General in all Egqual attack on both tension and grained, two 30-deg bends, two 150-deg bends ' bends compression sides of bends UNCLASSIFIED ORNL~-LR-DWG 25897 90-deg 8?7 o 0O.4cps //‘)c' - 30-deq /’0 BEND-~® o Ve o pEN o .go-deg /< & i q'(ao—deg BEND HO; = + ;{ PRIOR CREEP I ~1SOTHERMAL { 1.0cps~COARSE GRAINED O O.4¢ps-COARSE GRAINED — ® 0O.4cps- FINE GRAINED O I1SOTHERMAL -COARSE GRAINED & ISOTHERMAL- FINE GRAINED [ | 40 60 80 100 120 140 TEST DURATION (hr) MAXIMUM DEPTH OF ATTACK (mils) o n o Fig. 1.7.1. Summary of Data Obtained in Thermal- Cycling Experiments with Incone! Tubing Filled with Flowing NnF-ZrF‘-UF‘ (50-46-4 mole %), (Seeretauith “apsicn) 92 preliminary results of @ run (ET-29, Table 1.7.1) at 1 cps showed test-section attack that was somewhat greater than that observed at the lower frequencies. The low temperature amplitude (13°F) at the outside wall of the test section resulted from the increased cycle frequency. The observed intergranular attack of 5 mils with a temperature tluctuation of only +3°F indicates that the extent of the attack on Inconel is frequency dependent, with a critical frequency in the vicinity of ‘é to 1 cps. Additional data are currently being ob- tained at the 1-cps frequency in order to verify this conclusion. The effect of prior strain on the extent of the corrosive attack on Inconel under thermal cycling conditions was investigated in the series of runs ET-30 through ET-34 (Table 1.7.2). For run ET-30, a fine-grained tube was prestressed to give o fixed strain, allowed to relax for 24 hr at 1300°F, and then bent, as shown in Fig. 1.7.22. Because of o UNCLASSIFIED ORNL-LR-OWG 25898 t Nl . i: H e 2 i1, — Il oy ‘ ELECTR i “,//— ODESNI & T - | , e ' | e ili ' i!’ S *. -9 il HEATER SECTION it ‘é‘ . l . LF— 8in. : = TEST SECTION {a) HEATER SECTION P (5) TEST SECTION Fig. 1.7.2. Test Section Configurations Used in the Determination of the Effect of Thermal Cycling on Tube Bends. {2) 90-deg bend. (b) 30-deg bend. the crudeness of the creep apparatus, the exact strain imparted is not known. However, it is felt that the results give at least a qualitative estimate of the effect of stress. Runs ET-31 through ET-34- were performed with annealed (as-received) tubing. Two of these tubes, ET-31 (fine grained) and ET-33 (coarse grained), were bent as shown in Fige 1.7.2a; while the other two tubes, ET-32 (fine grained) and ET-34 (coarse grained), were - bent as shown in Fig. 1.7.2b. It may be noted from Table 1.7.2 that in all cases the attack was greater . in the bends than in the straight sections of the . tube. However, no difference was detecied between the attack on the tension and compression sides of the bends. On the basns of data obtamed by the'_ Experimental Engineering Group in tests of tube “bundles -in ‘a flucride fuel environment, this resuh! is somewhat unexpecfed.‘, A" duplicate series of - | ‘thermal-cycling tests -with bent tubes-. will be - ‘initiated shorfly thaf wnll be scheduled for 100 hr_-,:' rof operahon. o ' lt may also be noted in Table ].7.2 that, as.in __'prewous ‘tests, the fine-grained tubes showed - intergranular attack and the coarse-gramed tuhes_'i_;,i showed genera! attack. Photomicrographs showmg . the attack in two tube bends are presented in Figo 1.7030 - short . exposures. PERIOD ENDING SEPTEMBER 30, 1957 Since it is planned to use welded and machined shells in contact with the fluoride salt in the ART, a study of the effect of thermal cycling on machined welds was undertaken. The test sections were formed by butt-welding two pieces of Inconel tubing - {0.25-in. OD, 0.035-in. wall thickness). The inside of the weld was reamed to the original inside tube diameter, and the outside was machined back to the original outside tube diameter. The weld was located approximately 1 in. downstream from the -heater. The results of these tests, runs ET-35 - through ET-37, are given in Table 1.7.1. Photo- micrographs of the weld areas of the specimens are not yet available. Run ET-35 was made with a fine-grained tube, while coarse-grained tubes were used for runs ET-36 and ET-37. Further, the specimen used in run ET-37 was constructed by building up ‘4 in. of weld metal on the ends of ~the tubes, machining the weld metal to the original tube dimensions, and welding the weld-metal - pieces together. The specimen used in run ET-35 showed the same depth of attack in the weld region as in the unwelded portion of the tube, with perhaps somewhat denser subsurface void formations in the weld. The specimens used in runs ET-36 and ET-37 both failed in the weld areas after relatively In all cases the welds were fabricated by a qualified Inconel welder. However, since the welds were not critically inspected, the possibility of initial weld porosity exists. Ad- ditional tests of this type are to be run with welds which have satisfactorily passed inspection. Pulse-Pump System J. J. Keyes, Jr. A. . Krakoviak A. G, Smith, Jr.1 The high-frequency pulse-pump thermal-cyclmg loop has been successfully operated with the fuel “mixture NaF-ZrF -UF, (56-39-5 mole -%) as- the fluid medium. The initial run extended over ap- “proximately 72 hr of intermittent isothermal and ~eyelic temperature-dlfference operation. The mean fluid temperature was in the range 1300 to 1350°F, and hot-_to cold- leg temperature differences of as . ‘much as I75°F were achieved. Much difficulty was ;}‘A}:experlenced however, -in controlling the levels of “the six free liquid surfaces incorporated in the ':i;lédp.a At the end 6f the' 'ihifiai run, the loop was 35, E. Mott and As G Smith, Jr., ANP Quar. Prog. Re .Dec. 31, 1936, 0RNL-222 p 54~58. For details of oop, see esp Figs. 1.4.27 anJ 1+4.28. 93 v6 UNCLASSIFIED T.13299 oo e o1 00T ey | ooz .00 ! -5- -‘i‘-n -II:I:‘- 2 z 2 .008 008 | oon, 2oL uoer. .00t 008 008 | 008 LRAe | 00® oon 01 Lou N by T | A A S Y G {g) STRAIGHT PORTION (&) 90-deg BEND~TENSION SIDE (¢) 90-deg BEND—COMPRESSION SID RUN ET=3{: NO PRIOR STRESS, FINE GRAINED, FOUR 90-deg BENDS B UNCLASSIFIED EREB EEE Y |NC|.'|ES T INC';IES 008 .ooe. L.oo? oot 008 008 Looe oo et i 91 , X . o1, : . / el ,w.: [ 3 AL e, AR . R ’ (d)} STRAIGHT PORTIO () 90-deg BEND-TENSION SIDE (#) 90-deg BEND—-COMPRESSION SIDE .RUN ET-33: NO PRIOR STRESS, COARSE GRAINED, FOUR 90~deg BENDS Fig. 1.7.3. Specimens of Inconel Tubing Exposed to NaF-ZeF,-UF, (50-46-4 mole %) ot 1600°F in Thermol Cycling Tests of Bends ot 0.4 cps. Etchant: modified aqua regia. 250X. (Secret-with-eaption) ‘I . . L] ¥ . ® e + - W . v 1¥0d3Y $S3A90Ud 1I370d¥d INV L1 cooled to allow for some modifications of the liquid-level control system. For the second run there was a goal of 100 hr of operation at a fluid flow rate of 8 gpm with a frequency of 1 cps and a temperature differential of 200°F. Difficulties again occurred with liquid- level control, and, following another 72 hr of intermittent operation, the system was shut down for general repairs and further modifications of the level controls. »After installation of a new test section, the third run completed the 100-hr test peried at a fluid flow rate of 8 gpm, and the test was dis- PERIOD ENDING SEPTEMBER 30, 1957 continued without incident. The pulse frequency was 1 cps, the hot-leg temperature was 1500°F, and the cold-leg temperature was 1220°F. The test section (Fig. 1.7.4) was fabricated from an 11ein. length of 5/s-in.-()D Inconel tubing with a &5-milthick wall. The central 9 in. of the test section -was machined to a wall thickness of 29 mils (11 mil). The two ends of the test section were left at the original wall thickness to facilitate welding of the unit into the system. Chromel- Alume! thermocouples (5-mil wires) were attached to the outer surface of the tube by resistance- welding. The total amplitude of the outer wall -UNCLASSIFIED PHOTO-29668 Fig. 1.7.4, Typical Test Section After a Run in the High-Frequency Pulse-Pump Thermal-Cycling System. 95 ANP PROJECT PROGRESS REPORT temperature oscillation as indicated by these thermocouples was 127°F at the test section inlet and 103°F at the outlet. The inner surface temper- ature oscillation was estimated to be between +67 and 195°F. This variation exists because of uncertainties in establishing the boundary con- dition at the outer surface in the equation describing the decay of the radial temperature wave. The metallurgical examination of this test section has not yet been completed; however, standard dye-stain and radicgraphic examinations show no superficial damage. Future tests will again be made at a cyclic frequency of 1 cps, but the amplitude of the temperature fluctuation will be varied. Thérmocouple Development Jo E- MO"' An improved technique was used for further measurements of the surface temperature of a thick- walled pipe in contact with a fluid whose temper- ature is a periodic function of time. An Inconel- nickel ‘‘gunbarrel’’ thermocouple® was shrunk-fit into the pipe wall, the inside of the pipe was reamed so that the junction would be flush with the wall, and a thin (approximately 1-ux) layer of nickel was deposited by vacuum evaporation. This nickel layer introduced negligible resistance to heat flow, and thus enabled instantaneous measurement of the interface temperature. The thinness of the nickel layer ensured negligible perturbation of the fluid flow in the vicinity of the thermocouple. Typical resuvits for a test section fabricated from a 0.47-in.-ID, 0.25-in..wall Inconel pipe exposed to sinusoidal temperature oscillations in water are shown in Fig. 1.7.5 for a film heat transfer coef- ficient of 3700 Btu/hr-ft2.°F, As may be seen, the results are in reasonable agreement with the Jakob equation# for an infinitely thick plane wall. An Inconel-nickel ‘‘gunbarrei’’ thermocouple (not attached to a tube) was tested in static NaF-ZrF -UF, (50-46-4 mole %) at 1200°F for 300 hr. During this time the thermocouple indicated a steady temperature of about 1230°F. Metallurgical examination of the thermocouple, of which o longitudinal secticn is shown in Fig. 1.7.6, re- vealed, however, that the metallic junction between the central nickel wire and the Inconel had been 4M, Jakob, Heat Transfer, vol 1, p 298, Wiley, New York, 1949, 96 UNCLASSIFIED ORNL-LR-DWG 25899 07 0.5 e THEORETICAL CURVE o~ BY JAKOS h\.\_‘ 03 P 0.2 oA ) 2 -4 6 8 {0 FREQUENCY (cps) Fig. 1.7.5. Ratic of Wall Temperature Amplitude to Fluid Temperature Amplitude, 7, as a Function of the Temperature Cycling Frequency for a Thick-Walled Inconel Tube., Film coefficient: 3700 Bfu/hflfiz-oF. removed by the fluoride salt.? This is shown more clearly in the photomicrograph of Fig. 1.7.7. lt is concluded that the conductivity of the melt was sufficient to provide electrical continuity. Since the fluoride salt was found to have penetrated some distance along the nickel-oxide insulation, it is evident that this type of thermocouple cannot be relied upon for long-term accurate measurement of surface temperatures in a fluoride salt environment. ART HYDRODYNAMICS FullsScale Core Studies W. J. Stelzman The investigation of auxiliary flow guidance for improving hydredynamic stability in the ART core was continved, and two new configurctions were SRe Jo Gray and J. He DeVan, Metallograpbic Ex- amination of a Gunbarrel Thermocouple, Metallograpby Specimen No. 16203~Metallography Report No. 294, ORNL CF-57-6-71 (June 17, 1957). o PERIOD ENDING SEPTEMBER 30, 1957 B UNCLASSIFIED: " ¥.22959 Fig. 1.7.6. Longitudinal Section Through “Gunbarre!’’ Tfiermocouple After 300 hr of Exposure to Static NaF- ZcF ,-UF, (50-46-4 mole %) ot 1200°F. Oblique lighting causes the nickel wire to appear dark and the insulating gap, white. (Secret-witireeption) . : el UNCLASSIFIED ¥-22960. Fig. 1.7.7. -Photomlcrogmph of End ;f "Gunfiérté-.l"' 'Th-e'l"mldcouplé 'Af;er 300 hr of Exposure to Static NaF-ZrF . UF, (50-46-4 mole %) at 1200°F. Note the absence of the nickel _pla_te between the end of the central nickel wire and the Inconel on either side. {Serretwith-eaption) 97 ANP PROJECT PROGRESS REPORT tested. One configuration consisted of a pair of inserts and ramps mounted at each pump discharge port. The insert, one side of which maintained the circular contour of the core entrance passage, while the opposite side provided a gredual area transition for flow from the pump discharge, was installed at the junction of the center volute and pump-discharge passage and occupied a 38-deg circumferential sector. It extended the full height of the volute. The ramp was located on the roof - of the center volute and occupied a 50-deg radial sector adjacent to the insert. The shapes and the positions of the inserts and ramps are shown in Fig- ]07-8'- For the second configuration, a pair of thin (approximately l/a-in.) Plexiglas baffles, similar in contour to the actual core shells, were suspended in the annular passage between the island and reflector shells, as described by Platus,$ to form narrow (approximately 7/a-in.--wide) annuli adjacent to the island and reflector walls. The baffles extended approximately three-fourths of the length of the core. The first configuration (inserts and ramps) was designed to increase the tangential component of pump-discharge flow into the center volute. At the same time it was intended to provide more uniform flow distribution circumferentially into the core annulus by increasing the velocity of pump discharge ond by providing a downward velocity component to the swirling fluid mass. Observation of the flow pattern revealed no major improvement in stability even when used in con- junction with the GS-2 guide vanes. However, additional tests are needed to properly evaluate this design. The second configuration (baffles) was proposed primarily to improve the flow stability adjacent to the core shells, to absorb some of the thermal oscillations in the main stream, and to decrease the peak fluid temperature near the inner and outer wails. Preliminary work has been concerned with obtaining proper flow in the three parallel channels formed by the baffles. Qualitative observations indicated that satisfactory flow could be achieved by controlling the area of the inlet passages. Many more experiments will be required to optimize the design and to ascertain the effectiveness of these baffles. p, H. Platus, ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 10-12. - 98 Quarter-Scale Core Studies G. L. Muller? Velocity and pressure loss data were obtained in water flow tests of the "’(“-scule model of the ART 21-in. core with an ART-type entrance header, a core inlet collimator, and five core screens. The core and header system are shown in Fig. 1.7.9. The collimator was fabricated from a 3/8-in.-thick perforated plate of 0.42 solidity. The screens were cut from 20 x 20 mesh, commercial, woven- wire screening. The first four screens had a 0.385 solidity, while the final screen was of 0.510 solidity. , : The velocity distributions at six axial positions clong the core are shown in Fig. 1.7.10 for a mid-plane Reynolds modulus of 6020. These profiles were cbtained from analysis of photographs of the . velocity profiles visualized at these positions by the phosphorescent-particle technique (see Fig. 1.7.13 of the next section). Because of the entrance collimator, the core flow was nontotational. The results are in good agreement with previously reported qualitative data on this system.? The profiles at positions C, D, and E indicate that comparatively low velocities exist near the outer wall in the mid-plane region. A prelimi- nary examination of velocity data obtained at NRe,mia = 25,000 for the mid-plane region showed no major deviations from the profiles of Fig.1.7.10. The experimental variation of average fluid velocity with axial position is given by the data points of Fig. 1.7.11. The positions A°and D’ are located in a vertical plane rotated 90 deg from the plane containing positions A through F. The curves shown were calculated on the basis of flow continuity by using the data at positions A and A’ for base points. The reliability of the experimental data is indicated by the excellent agreement be- tween the measured average velocity at position F and that predicted by the calculated curves. The average fluid velocities are equal at positions A and A’ however, a 10% deviation exists at positions D and D’. This difference lies within the experimental error at this position. Thus, the data indicate a core flow which is peripherally symmetrical despite a large pressure variation in the core header. The extent of this 7G. L. Muller and F. E. Lynch, Effects of Screen Packing in the ART 2l- 3 '— J 60 —g 10 W \ 40 k—-..._ 4 \'fil__ 20 0 2 x 10? 5 10’ NRe, mid Fig. 1.7.12. Relative Pressure Difference in Mode! of ART Inlet Header. 102 C p) . Instantaneous Velocity Profile Measurements ' F. E. Lynch Velocity profile photographs were obtained for flow through the 19/ 14-Scale model of the ART with an ART-type entrance header, a core inlet colli- mator, and five core screens. The experimental system is as described in the previous section (see Fig. 1.7.9). Photographs were taken at the six axial positions A through F and at positions A’ ~and D’ for twin-pump operation and ot positions A and B for single-pump operation. Typical velocity profile photographs are shown in Fig. 1.7.13 at positions C and E. The lack of sharpness in the photographs results partially from the fluid mixing brought about by the screens and partially from the slow shutter speed (% . sec) of the camera. This study will be repeated with a 35-mm comera (faster shutter speed) and a Reynolds modulus higher than the 6000 to 9000 used for the present test in order to obtain more sharply defined profile photographs. Screens of various solidities will also be investigated. : In order to obtain quantitative velocity infor- mation, it is necessary to obtain a photograph of FERIOD ENDING SEPTEMBER 30, 1957 a grid placed in the plane of the visualized proefile, as shown in Fig. 1.7.14. The distortion of the grid lines by the curvature of the Plexiglas shell can be clearly seen. "FUSED SALT HE‘AT'TRANSFER H. W. Hoffman ~ S. L. Cohen D. P, Gregory Forced-convection heat transfer. studies with NaNO,-NaNO,-KNO, (40-7-53 wt %) flowing through heated tubes I::ave been completed. The data cover the Reynolds modulus range of 5000 te 25,000 and show a heat transfer coefficient variation from 800 to 2900 Btu/hr+ft2.°F, The final results are presented in Fig. 1.7.15 in terms of the heat transfer parameter, N /N%;‘. For comparison, earlier data obtained with this nitrate-nitrite mixture, both in this laboratory® and by other investigators,? are presented. All the data of - 8H. W, Hoffman, Physical Properties and Heat Transfer Characteristics of an Alkali Nitrate~Nitrite Salt Mixture, ORNL CF-55-7-52 (July 21, 1955). 9W. E. Kirst, Wo M. Nagle, and J. B. Castner, Trans. Am. Inst. Chem. Engrs. 36, 371 (1940). SECRET PHOTO 42241 (a) POSITION C (b) POSITION E Fig, 1.7.13. Photographs of Visualized Velocity Frofiles in Screen-Filled 10/44-Scale Model of 21«in. ART Core. 103 ANP PROJECT PROGRESS REPORT S PHO TO 42242 Fig. 1.7.14. Photograph of Coordinate Grid Placed in Plane of Visualized Velocity Profile. Fig. 1.7.15 have been adjusted by using the value for the thermal conductivity of this salt that was recently determined by Powers (see subsequent section of this chapter on *‘Physical Properties®’). The experimental results are in good agreement with the empirical equation, Ny /N34 = 0.023 NQ'S, which describes forced-convecnon heat transfer in ducts containing ordinary fluids.!0 Studies have been initiated of forced-convection heat transfer with KCI-LiCl (41.2-58.8 mole %) flowing through a type 347 stainless steel tube. The experimental system is nearly identical to that used for the NaNO -NaNO,-KNO, (40-7-53 wt %) investigation. Prehmmary results fall below the standard correlations, but this discrepancy is probably due to uncertainty in the data on the thermal conductivity of this chloride mixture. l‘)Ol'c:llm:n'y fluids are defined as those fluids whose Prandtl moduli lie in range 0.5 < Np_ < 100. 104 UNCLASSIFIED ORNL-LR-DWG 25904 04 r {00 © KIRST, NAGLE, AND CASTNER ® HOFFMAN AND 4 PRESENT RESULTS 5 50 20 N, N - 0.8 o = 0.023 Ny, NPr 10 5 10* 2 5 10° REYNOLDS MODULUS, NRe ™ 0 10 HEAT TRANSFER PARAMETER, Ay, /W, Fig. 1.7.15. Hecat Transfer with NaNOz-NcNO -KNO3 _ (40-7-53 wt %), The possibility of an interfacial film is also being investigated by using both chemical and x-ray diffraction techniques. The existence of a film ~ will be further checked in experiments with an ~Inconel tube. | HEAT TRANSFER EXPERIMENTS N. D. Green W. R. Gambill ART-Type Core with Screens Initial experiments have been completed on the half-scale volume-heat-source ART core model containing four 0.342 solidity screens and an inlet collimator. Both steady-state and fluctuating temperatures were measured at the inner and outer core walls. The transient data are compared in Fig. 1.7.16 with datc obtained in previous measure- ments1! of a vaned entrance system. While the magnitude of the outer wall temperature fluctuations peak is appreciably above the corresponding value for the vaned core, the inner core fluctuations have been reduced by an order of magnitude. In- Fig. 1.7.17 the corresponding steady-state temper- ature distributions are presented. The inner wall temperatures lie consistently below the fluid mixed-mean temperatures and indicate uniform high-velocity flow along the inner wall. These results are in good agreement with predictions made on the basis of the velocity profiles experimentally determined by Muller and Lynch? (see also, Fig. 1.7.10, this chapter). Thus, some hydredynamic instability in the mid-plane region on the outer wall Ny, F. Poppendiek et ali, Analytical and Experi- mental Studies of the Temperature Siructure Within the ART Core, ORNL-2198 (Jan. 31, 1957). , fw UNCLASSIFIED ORNL-LR -DWG 25905 NRE" 82,000 AT =175°C W = UNIFORM P = 0425 OF ~EDDYING JN ¥;-SCALE VANED OUTER WALL INNER o] 2 4 6 8 10 12 14 16 18 AXIAL DISTANCE FROM INLET (in.} Fig. 1.7.16. Tmnslehf 'Tem.percture Fluctuations in the Screen-Filled One-Half-Scale Volume-Heat-Source Model of the ART Core. (‘S'e'eoe*wi-rh‘m‘p'rl'on’) UNCLASSIFIED ORKL-LR-DWG 25906 2.0 SCREEN POSITIONS € [+ E =z 2| z 1.8 F-). ] hj el &l /='—- S8 5 = & TN S2 48 3 | D B g 8 g & B g F 8 ] 0, w | 14 1% ] Nge™ 82,000 (AXIAL) AT =175°C 1.2 P =0.425mw i / W = UNIFORM e / ST 10 ‘ : : o = hE /(~oun-:n WALL 1 0.8 | 1T~ . ‘ vd > FLUID MIXED-MEAN;/ / 06 / W / INNER WALL. 0.4 / /"/ / ~ 0 AXIAL DISTANCE FROM INLET (In) = - Fig. 17.17. Steudy;Sffife Wuli Teini:ieratfires’ in the . Screen-Filled One-Half-Scule Volume-Heuf-Source Model o of Ihe ART Core. (Gmfl-wfirtvphm}- *is'réflected in thek observed peaking of the transient . ' temperature fluctuotlons on _the outer wall of the ~ volume-heat-source “system. - Slmxlurly, the flat. : -,_velocuty prof:le (approxlmatlng the shape of the " power profile) near the -inner. wall indicates a low - surface temperature. - Conversely, a hsgh outer wall " temperature, as shown in Fig. 1.7.17, was to be expected. A possible method of altering the core flow to increase the flow along the outer wall without 0 2 4 6 B - 10 12 14 16 - 18 - " MODIFIED COLBURN FACTOR, / o PERIOD ENDING SEPTEMBER 30, 1957 materially affecting the inner wall flow lies in introducing ‘screens of variable radial solidities. Screens of this type would have high pressure drops (high solidity) through the central portions that would force greater peripheral flow. Several such experimenta! screens have been designed and are now being fabricated. Vortex Tube Preliminary experimental determinations have been made of the heat transfer coefficients for air in source vortex flow. The measurements were made by using metal Hilsch-Ranque tubes having length-to-diameter ratios of 20. The tubes were heated by the axial passage of a high-amperage electrical current through the tube wall. The data are shown in Fig. 1.7.18, in which a meodified Colburn factor is plotted as o function of the energy required per square foot of heat transfer surfoce. For tubes of identical geometry and the same over-all pressure drop, heat transfer in straight turbulent flow and in source vortex flow can be directly compared. On this basis, the experimental data of Fig. 1.7.18 show vortex flow to be a factor of 5 better than straight flow. The correction to the vortex flow data to account for the pressure drop associated with the momentum increase of the air in passing from the tube inlet to exit has not been included in this analysis. The magnitude of this correction will be experi- mentally determined in the 3-in. vortex currently under construction. This tube will allow radial and axial traverses of fluid velocity, temperature, UNCLASSIFIED . - © ORNL-LR-DWG 25907 100 w o N o . ™ 04 02 05 4 2 5 40 20 50 400 ENERGY REQUIRED PER UNIT AREA (hp/f%) Fig. 1.7.18. Experimental Heat Transfer to Air in Source-Yortex Flow ond in Straight, Turbulent Flow Through Pipes. 105 ANP PROJECT PROGRESS REPORT and pressure. Optimization of the heat transfer efficiency as related to such factors as tube length-to-diameter ratio and entrance region design is being investigated. LIQUID-METAL VOLUME-HEAT.SOURCE ' EXPERIMENT G. L. Muller An investigation of heat transfer in a liquid metal (mercury)system with internal heat generation is in progress. This study is aimed at a better understanding of the heat transfer mechanism so that more accurate calculation of the thermal structures in circulating-fuel reactors can be made. Discrepancies between results from experimental fiquid-metal forced-circulation heat transfer studies and theoretical calculations have raised questions as to the validity of corresponding analyses for the volume-heat-source case. In particular, the “*interface thermal resistance’’ theory postulated to explain the reduced values of the experimentally measured Nusselt modulus when heat is transferred through the tube wall will be tested in this study. A diagram of the experimental system now being constructed is shown in Fig. 1.7.19. Mercury wil! flow from the sump into the Moyno pump, through a flow=measuring orifice, through the test section and ccoler, and then return to the sump. In an additional flow path, the mercury will bypass the test section and cooler. The test section is a l(“-in. glass tube interrupted by three electrodes for volume electrical heating. The arrangement is such that no heat is generated outside the test section. The wall temperatures of the test section will be measured by copper- constantan thermocouples attached to the outside of the glass tube. To prevent heat loss through the test section wall, external insulation and guard heating will be provided. Fluid mixed-mean inlet and exit temperatures will also be measured. In this apparatus Reynolds moduli of 500,000 can be attained. A second test section of 1%-in.-ID glass tubing will be available with which to make radial temperature and velocity traverses. MASS TRANSFER J. J. Keyes, Jr. A comparison with experimental results12.13 of the amount of mass transfer to be expected in alkali metal-alloy systems was made for two assumed diffusion mechanisms.'4 The results are presented in Table 1.7.3. In the first mecha- nism, the boundary layer was assumed to be satu- rated and the rate of transfer was limited by the rate of diffusion of solute into the liquid. A general expression was derived that included effects of variation in temperature and liquid composition around a closed loop and which simplified under certain conditions to W(O) = k, Sty — t.)0, where W(@) is the amount of mass transfer in time 0, k; is the conventional mass transfer coefficient, 125, H, DeVan and J. B. West, A Brie[ Review of Thermal Gradient Mass Transfer in Sodium and NaK Systems, ORNL CF+57-2-146 (Feb. 11, 1957). 13), H. DeVan and R. S. Crouse, ANP Quar. Prog. Rep. Dec. 31, 1956, ORNL2221, p 175; ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 154. 145, 3, Keyes, Jr., Some Calculations of Diffusion Controlled Thermal Gradient Mass Transfer, ORNL. CF-57-7-115 (July 22, 1957). Table 1.7.3. Summary of Calculations and Comparison with Data for Two Postulated Mechanisms of Mass Transfer in Sodium=inconel Systems Exposure time: 1000 he Temperature differential: 300°F Mass Transfer Estimate, W (g) Hot-Zone Temperature Experimental Mass Transfer (°F) LiquideDiffusion Limiting Wall-Diffusion Limiting (g) from Forced=Circulation Lo'ops- ' Method - Methed R : 1500 400-1000 2.-25 13 1-8 <0.5 1200 300-~700 106 { »y o 4 FLOW CONTROL VALVE BYPASS - MERCURY FLOW DIRECTION zone = temperatures, respectively. forced-crrculahon loop conditions (tH ="1500°F, ture dependent, in contradiction to observation, . In the second mechanism, the rate of mass transfer was assumed to be limited by the rate of SECTION PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR—DWG 25908 MIXING CHAMBER AND THERMOCOUPLE WELL i POWER j TRANSFORMER GLASS TEST \MQMQJ‘—I \\SATURABLE-CO RE REACTOR \THIN-PLATE ORIFICE N o | PUMP DRIVE MOTOR AND SPEED SELECTOR POSITIVE DISPLACEMENT MQYNO PUMP Flg. 1.7. 19. . Schematic Diugrom of Voluma-Heaf-Source Experimenful Appcrurus in Which Mercury will be Circu- iufed. L : : S is the solublltty of the soiute (Nl) in the !:quld'i?' (No or NaK), and -2, and tC are the hot- and ‘cold- - For- typical diffusion of a component (for example, Ni) of the -solid alloy (for -example, Inconel) to the solid- . liquid interface. derwed for W(G) was . =1200°F, 0 = 1000 hr), this mechanism. pred:cts , to be between 400 ond 1000 g, whereas: the‘-__-, observed value is about 13 g.- Furthermore, this mechanism predicts W to be only slightly tempera- - The approximate expression 1/2 ' o i D6 x(ty = te) : W(@) = 2pw ——77 ' —-———-—t ’ where D is the diffusivity of nicke! in the solid wall, x is the bulk solid concentration, and p , is the solid density. - 107 ANP PROJECT PROGRESS REPORT Based on extrapolation of some data for the diffusivity of nickel, as reported in the literature, W(@) was estimated to be between 2 and 25 g for ty = 1500°F, ¢, = 1200°F, 6 = 1000 hr. As may be seen, this estimate is in better agreement with the observed mass transfer accumulation than was the estimate based on liquid diffusion. On the other hand, the wall diffusion mechanism does not correctly account for the dependence of the mass transfer (W) on flow rate and time and does not predict a sharp break in mass transfer at cbout 1350°F. It is concluded that ¢ more general hypothesis is needed for combining the mechanism of hot-zone attack with o nucleation-deposition mechanism in the cold zone. PHYSICAL PROPERTIES W.D. Powers Enthalpy and Heat Capacity Studies were made to determine the enthalpies and heat capacities of the eutectic mixtures LiCl- SrCl, (22.5-77.5 wt %) and LiCl-BaCl, (32.2-67.8 wt %). Preliminary analyses of the dzata yielded values for the heat capacities of 0.21 Btu/1b.°F for the barium-containing mixture and 0.24 Btu/1b.°F for the strontium-lithium salt. The reliability of these data is in some doubt, since wide dis- crepancies in enthalpy were observed between duplicate test samples. This may be due to the large density differences between the salts. Further studies with this pair of chloride eutectics will be made as soon as new samples can be prepared. Thermal Conductivity Modification of the variable-gap thermal conduc- tivity device to include a large guard heating ring surrounding the heat meter has been completed. This guard heater ensures lower heat losses from the heat meter and at the same time provides a heat flow path of cross section more nearly equal to the cross-sectional area of the sample heater ond its guard. The results obtained with the salt mixture NaNO,-NaNO,-KNO, (40-7-53 wt %) show increased consistency ioth within runs and between runs. The results of a number of typical experimental runs are shown in Fig. 1.7.20. In this plot, the slope of the data line is the reciprocal of the thermal conductivity of the liquid sample. The 108 thermal conductivity values obtained in this study of the nitrate-nitrite salt are given below: Temperature Thermal Conductivity (°F) (Btu/he-#1°F) 689 0.35 667 0.34 487 0.34 468 0.33 All temperature effects were within experimental error. The results compare favorably with data obtsined by Deem!'5 (0.33 Btu/hr-ft-°F between 400 and 900°F) at Battelle Memorial Institute. UNCLASSIFIED ORNL-LR-DWG 25909 06 as A~ x | 0.4 L] S a 203 < e 353°% o 255°C 02 o 242°C 7 o / F Q 0050 0.400 0.150 GAP THICKNESS (in.) 0.200 Fig. 1.7,20, Experimental Data Obtained in Measure- ments of the Thermal Cenductivity of NcNOz-NcNO3- KNO 4 (40-7-53 wt %) The thermal conductivity of the fuel mixture NaF+Z¢F UF, (50-46-4 mole %) has been de- termined by several laboratories. The results are compared in Table 1.7.4. No trend with tempera- ature was noted. This salt mixture will be re- investigated with the use of the modified variable- gap device. The variation between the Mound Laboratory and the ORNL and BM! data has not yet been resolved. 154, w. Deem, unpublished data. Properties of Zirconium Fluoride Vapor Deposits The density of a sample of zirconium fluoride powder collected in a "*snow’’ trap was measured and found to vary between 0.55 and 0.8 g/cm3. The samples were collected by pushing a thine walled, sharp-edged tube through the ZrF, and into contact with the metal surface to which the powder was adherings. The thickness of the sample, whose area was defined by the inside cross-sectional area of the tube, was determined by averaging the depth ‘indicated by a fine-wire probe inserted at four positions immediately adjacent to the sample. The sample thicknesses. varied from 0.05 to 0.4 in. It was found, for sample thicknesses greater than 0.1 in., that the density varied inversely with the sample thickness, as shown in Fig. 1.7.21. Tests have also been initiaoted to determine the effect of subsequent heating on the apparent density of the ZrF , deposit. An attempt was also made to measure the thermal conductivity of the ZrF , deposit. A sample of the material was ploced in o cylindrical annulus surrounding a central heated tube. The conductivity of the material around the *‘hot wire’’ is determined from the temperature rise of the electrically heated wire. A value of 0.45 Btu/hr-ft-°F was obtained; PERIOD ENDING SEPTEMBER 30, 1957 however, this measurement was made on a ‘*disturbed’’ sample, and it is planned to mount @ tube in @ snow trap and allow the ZrF , to collect on the tube. UNCLASSIFIED CRNL—-LR-DWG 25910 1.0 P 'h-.--..__---. e — ® ., @ " ® '--.__. ® £ . .--—.'-. 'gx ® ™ - = e E 0.5 w Z w O 0 0 0.1 0.2 0.3 0.4 0.5 SAMPLE THICKNESS (in.} Fig. 1.7.21. Experimental Measurements of Apparent Density of Zirconium Fluoride Vapor Deposits. (Gom frckerrtirtwrith sien) Table 1.7.4. Thermal Conductivity of Melten NuF-ZrF4-UF 4 (50-46-4 mole %) Investigator Experimental System Thermal Conductivity (Btu/ hr+ft°F) Powers, ORNL Variable gap 142=1.5 Fixed gap 0:8~146 Deem at BMI* Variable gap 1.2 - Jordan at Mqunfl Laboratory** Calériméfr'ic device | 049 *H. W, 'Deém, unpublished datas _ ** Aircraft Prbpulsz'bn Reaétéf,s_, ~‘Mound "Lc‘b'o'ra'féry' Memorandum Report No. 57-7-40, p 5 (Aug. 5, 1957). 109 » Part 2 CHEMISTRY W. R. Grimes # A 2.1. PHASE EQUILIBRIUM STUDIES' C. J. Barton THE SYSTEM KF-UF H. A. Friedman Further equilibrium studies of the system KF-UF have confirmed the liquidus curve presented previously.? Results of thermal-gradient quenching experiments have shown, however, several differ- ences from the phase behavior proposed earlier on the basis of thermal analysis and optical and x- ray-diffraction examinations of slowly cooled melts. Of the eight KF-UF , phases reported by Zacharia- sen,® two phases, B-2KF-UF, and KF.2UF,, have been shown to be equilibrium phases in the system. |t has not been definitely determined whether 2KF.UF, displays distinct temperature ranges for the ordered (B) and disordered (8°) forms. Further, no evidence has been found that the phases designated by Zachariasen as o-3KF.UF,, B-3KF.UF,, o-2KF.UF,, KF.UF, (which occurs, rather, as 7KF «6UF ), KF- 3UF4, and KF.6UF, occur in slowly cooled melts or in quenched samples purified by hydrofluorination and subsequently protected against exposure to the atmosphere. 1f, however, such samples are pre- pared exactly as described by Zachariasen,* that is, exposed to air during heating and cooling, the phases designated an3KF-UF4, B-3KF'-UF4, and KF.6UF, can be formed in large amounts. No phase that fits the description of the compound KF.:3UF, has been found under either type of exper:mental condition. In thermal-gradient quenching experiments the phuses a and B-3KF.UF, are apparently stable if initially present in 1he samples. Protected purified melts of the composition 3KF.UF, show a biaxial-negative blue-green phase upon solld:ftcahon, and thermal- gradient quenching experiments and - “thermal analysis indicate that this is the only equilibrium phase for the compound 3KF UF4. No snmllar ]The pefrogmphlc ‘examinations _re orted here were performed by T. N. McYay, Consultant, and R. A. Strehlow, Chemistry Division. The x-ray examinations were performed by Ri W, Thoma, Chemistry Division. H. A. Frledmcn, ANP Quar. Prog. Rep. Dec. 10, 1955, ORNL-2012, p 78 and Fig. 4.4, p 82. 3w. H. Zachoriasen, J. Am. Chem, Soc. 70, 2147-2151 (1948). 4W. H. Zachariasen, Crystal Structure Studies at the Systems KF-UF4, KF-TbF4. and KFe-LaF,, CC-3426 (Feb. 9, 1946). R. E. Moore R. E. Thoma phase has been reported by Zachariasen. The phases a- and B-3KF.UF, and KF-6UF, may be either metastable phases in the system KF-UF, that are stable against inversion to equrhbnum forms or oxygen-containing phases. Zachariasen has reported a face-centered cubic form (a) and two hexagonal forms (8, end B,) for the compound 2KF-UF ,. If the alpha form were to display phase behavior similor to that of the isomorphous phase® reported to be a-2NaF-UF 4 it would be observed as a primary phase at fhe 35 to 40 mole % UF, region, and its formula would be 5KF-3UF,. No such primary phase for a cubic material has been observed. The possibility of a subsolidus existence for the cubic phase is not precluded, although the phase has not been ob- served within the 50°C temperature range below the solidus. No observations of 2KF-UF , samples have yet permitted a distinction of the ordered and dis- ordered hexagonal forms. No well-crystallized samples of 2KF:UF , have become available. The compound 2KF.UF, has a lower limit of stability of 620°C, at which temperature it decomposes into 3KF«UF, and 7KF-6UF,. Some unexplained low- temperature thermal effects in the 2KF-UF, com- position region require that this decomposition temperature be considered to be tentative, The equilibrium phases in the system KF-UF, and their melting choracteristics are given in Table 2.1.1. 5This phase has been identified as 5KF'3UF4. Table 2,1.1. Equilibrium Phases in the System KF-UF‘ Melting Melti Formula Point ch e*tu-\g'. ©0) aracteristics 3KF.UF, 955 Melts congruently 2I(F'_UF4 _ 760 Melts incongruently to 3KF-UF and liquid; decomposes at 620°C 7|‘(F'6UF4 790 Melts congruemly KF-ZUF4 765 Melts incongruently to UF ; and liquid “()Ay 13 ANP PROJECT PROGRESS REPORT THE SYSTEM NuF-HfF‘ H. A. Friedman Samples of pure, sublimed HfF , became available during the quarter that were pure enough to per- mit refined phase equilibrium studies of the system NaF-HfF . Thermal analysis data obtained with the pure materwl are in substantial agreement with data obtained in thermal-gradient quenching experi~ ments, and thus some of the phase equilibrium results based on previously reported® . thermal analysis data are presently in doubt. Precision in ‘the determination of liquidus and solidus temper- atures and temperatures of polymorphic tronsitions has been improved to +3°C, in several cases, by using sublimed HfF,. A tentative phase equi- librium diagrom of the system NaF-HfF , is shown in Fig. 2.1.1. Areas outlined by dashed curves are not yet well-established. UNCLASSIFIED ORNL-LR-DWG 25916 TEMPERATURE (°C) 3NaF -HfF, 5NaF - 2HfF; 2NaF - 6HfF, -HIF, 7NaF - 6HF, NaF 10 20 30 40 50 60 70 80 90 Hfl';', HfF, (mole %) Fig. 21.1. The System NaF-HfF (Tentative). Seven NaF-HfF, phases are isostructural with NaF-ZrF, phases of corresponding stoichiometric formulas. The analogous series of compounds in the two systems are not continuous, however, s is shown in Table 2.1,2; there are fewer phases in the NaF-HfF , system then in the NaF-ZrF system. The appeorcmce of the phase NaF- HfF4 as a congruently melting stable compound is the only ‘8B, S. Landau, H. A. Friedman, and R. E. Thoma, _ANP Quar, Prog. Rep. March 31, 1957, ORNL-2274, p 93. 114 marked deviation of phase béhavior foran NaF-HfF analog of an NaF-ZrF , phase. THE SYSTEM KBF ,-NaBF, R. E. Moore A quenching study of the system KBF (NaBF was undertaken during the past quarter as part of cl search for low-melting salt mixtures for use as reactor coolants, Thermal data obtained pre- viously?¢® defined a liquidus curve with a mini- mum at about 90 mole % NaBF, and 360°C, but there were no thermal effects on the cooling curves that corresponded to solidus temperatures, Thermal effects that presymably represented solid transitions were found. in the range 180 to 280°C. ~Petrographic examinations of quenched samples gave liquidus values that were in good agreement with the thermal data. The liquidus temperatures at 50, 70, and 75 mole % NaBF , were 441, 392, and 387°C, respectively, with KBF, as the pri- mary phase; the solidus temperatures were 356, 360, aond 355°C, respectively. The secondary phase was identified as NaBF in each case. At 80 mole % NaBF , the primary phuse is KBF,, and the solidus temperoture is 358°C. The composition containing 90 mole % NaBF, is very near the eutectic. The primary phase is NaBF,; the liquidus temperature is 355°C; and the solidus temperature is 351°C. X-ray diffraction and petrographic examinations of previously equilibrated quenched samples provided no indication of the existence of either compounds or solid solutions in the system. |t is a simple binary system with a eutectic near 90 mole % NaBF , and about 375°C. YTTRIUM FLUORIDE SYSTEMS R. J. Sheil Petrographic and x-ray-diffraction examinations of slowly cooled samples of LiF-MgF,-YF, and LiF-ZnF o~YF, mixtures, which are of mferest in the investigation of the production of oxygen- free yttrium metal,? have revealed the same un- identified compound in both systems, This sug- gests the possibility of an LiF-YF; compound, 7). G. Surak, R. E. Moore, and C. J. Barton, ANP Quar. Prog. Rep. Sept. 10, 1952, ORNL-1375, p 82. J. G. Surak, ANP Quar. Prog. Rep. Dec. 10, 1952, ORNL-1439, p i10. 9 Rc Jo Shel'l ANP 2are Pfogn Re . ure 30' 1957] ORNL-2340, p 128. : e pe " Gl observed ( % s v » v * " 0 ! ® Table 2,1,2, Properties of Analogous Phases in the SYstef'ns NuF-Z_rF4 and Nc::F-HfF4 NGF-ZI'F4 ‘ NoF-HlfF“ ‘ Phase Change | Phase Change Phus_e'.. - Temperature Typ °. of : Phusel " Temperature TTVPQ_ ?f Formula (°C_)- Transition Formula (°C) ransition 3N¢:F-ZrF4 o 850 | Congruent melting point 3NaF-HfF4' _ 855 Congruent melting point SNuFQZrF4 : 639 . |nc§ng‘ruent melting point SNuF-ZIHl‘F4 606 Incongruent melting point SNOF'Z.Zl;F4‘ : 525 'Iriver;ipn of a-SNoF-ZZrF"‘ 5N¢:F-2HI"F4 533 Inversion of G-SNoF-ZHfF4 R to 5‘5““’:'22":4 | to fiSNc:iF'ZHI‘F4 2NuF-IZrF.4' 544 ' lncdngruent melting point 2NaF - HfF, 593 Incongruent melting point - of 3,-2NaF ZrF , of 3,-2NaF -HfF 2Na F'ZrF4 533 Inversion of B2-2NaF'ZrF4 ' No B2-2N0F-HfF4 . ‘ to fi3-2N-aF-ZrF 4 | observed 2NaFZrF 505 Inversion of B3-2NaF ZeF 2 NoF *HfF 520 + 15 Inversion of fi3-2NoF-HfF4 : to [3“-2Nch-ZrF4 to 54'2N0F'H”:4 3Nch-22rF4 487 Upper stability limit of No 3NoF2HfF ‘ ' 3NaF«2Zr F4 observed ' 7NflF°GZrF4 525 Congruent melting point 7NaF'6HfF4 515 10 Incongruent melting point Nch-ZrF4 Metastable phase formation NoF-HfF4 545 Congruent melting point 3N U0, + NiF, or 4UOF, + 2Ni—> U,0, + 2NiF, + UF, It seemed evident also that UO_F, is more stable in contact with platinum, as might be expected, - but that a portion of the compound present was reduced by some means or underwent thermal decomposition. In order to investigate the likelihood of thermal decomposition, a portion of the UO,F, was heated alone in an inert atmos- phere in the visuval-thermal apparatus with o platinum crucible as the container. When the temperature was increased to 120°C, a large thermal effect was noted, and the exposed surface of the material gradually darkened. The temper- ature was further increased to the maximum temper- ature of 470°C, and the sample was then allowed to cool to room temperature, placed in a tared screw-cap bottle, and weighed immediately after removal from the inert atmosphere box. The compound had lost 0.4% in weight, and it was quite hygroscopic. It gained about 1.1% in weight after standing in the screw-capped bottle for 10 days. The analyses of the material before and after heat treatment are given in Table 2.2.27. The exact nature of the chemical change in the UO,F, used in these experiments is not clear, Table 2.2.27. Chemical Anclyses of UO,F, Samples U F H,0 (Wt %) (wt%) (wt %) Sompled before heating 75.3 12.3 6.37 Sampled after heating* 7642 12.4 6:16 Theotetical content 774 12.3 *Analyzed after standing in screw-capped bottle for 10 days. 142 but it seems obvious that hydrolysis did not occur, because the fluorine content did not change on heating. It has been reported39 that hydrated uranyl fluoride can be dehydrated at 120°C without serious decomposition and also that uranyl fluoride formed at low temperatures is very hygroscopic.3! This experiment seems to confirm both obser- vations, except that the observed weight loss was much less than would be expected from the reported water content of the starting material. OXIDATION OF MIXTURES COMPOSED OF SODIUM AND POTASSIUM E. E. Ketchen The results of some preliminary studies of the oxidation of a sodium-potassium mixture at room temperature were reported previously.32 The data indicated that a sodium-to-potassium ratio of about 22 was obtained in the oxides formed by the oxidation of a mixture of 50.4 wt % sodium with 49.6 wt % potassium. Additional experiments have shown, however, that the values obtained were probably inaccurate and that the precise results were due either to a very careful repro- duction of experimental conditions or, more probably, pure chance. ‘ The apparatus used in the recent studies is a modification of the glass apparatus described previously.32 The upper portion of the apparatus was enlarged to a diameter of 2% in. to eliminate the need for shaking the mixture during the oxidation. The oxidation and filtration are per- formed in the dry box in order to reduce the amount of contamination by water ond the resulting hydroxide formation. At present, three capsules are loaded in the dry box, and each is oxidized by intreducing dry air at a pressure of 1.5 to 2 cm of Hg, evacuating, and repeating this cycle wuntil four portions of air have been used. Filtration is achieved by applying helium pressure to the upper portion of the apparatus, forcing the alkali metals through the glass filter disk, and leaving the oxides with a small amount of metal in the upper chamber. The residual alkali metals are removed 3°G. R. Dean, R. Bradt, and M. S. Katz, Chemical Research-P-9. Report for Period Ending January 5— March 1, 1944, CC-1382, p 20, ' 31), J. Katz and E. Rabinowitch, Natl. Nuclear Energy Ser. Div. VIII 5, 569 (1951). 32E, E. Ketchen, ANP Quar. Prog. R . . y . £ ep. ]une 30 1957, ORNL-2340, p 156. P ' W » ”» from the oxides by mercury according to the procedure described previously. The results obtained by using this method for the room-temper- ature oxidation of two different sodium-potassium mixtures are given in Table 2.2.28. The lack of preciseness of the results suggests that it is extremely difficult to control all the variables which offect the oxidation process. Traces of moisture should tend to reduce the sodium-to-potassium ratio, and consequently the PERIOD ENDING SEPTEMBER 30, 1957 larger ratios are probably closest to the correct valve. It may be noted that the values given in Table 2.2.28 that were obtained for the 51 wt % Na~49 wt % K mixture are significantly larger than the value of 22 obtained previously. The lack of precision in the results makes it difficult to ascertain the effect of composition of the alkali metal mixture on the sodium-to-potassium ratio in the oxides, although it appears that the ratio de- creases with increasing potassium concentration. Toeble 2.2.28. Compesition of Oxides Formed by Oxidation of Sodium-Petassium Mixiures'uf 30 t;a 3s°cC Composition of Na-K Mixtures Composition of Oxides Ratio of Na to K (wt %) Na (mg) K (mg) : in the Oxides 51 Na—49 K 13.8 0.47 29 13.5 0.30 45 10.9 0.27 - 40 7.3 0.17 43 6.7 0.18 ' 37 10.4 0.29 36 8.5 0.18 47 22 Na-78 K 8.4 0.32 26 8.9 0.29 31 10.1 0.29 ' 35 8.1 0.27 30 7.4 0.23 32 86 0.31 28 143 ANP PROJECT PROGRESS REPORT 2.3. PHYSICAL PROPERTIES OF MOLTEN MATERIALS F. F. Blankenship ' VAPOR PRESSURE OF l'lfl'-'4 S. Cantor An appreciable difference in properties between analogous compounds of zirconium and hafnium can be of considerable importance in the separa- tion of the two elements. Therefore the sublima- tion pressures of HfF, were determined, by using the Rodebush-Dixon method,! for comparison with the known sublimation pressure of ZrF, (ref 2). The data obtained, as plotted in Fig. 2.3.1 and given in Table 2.3.1, fit the equation 12,011 T (°K) ° As may be seen from Fig. 2.3.1, the sublimation pressure of ZrF , is approximately three times that of HfF, at o given temperature. This factor of 3 difference between the vapor pressures of these Iogrp (mm Hg) = 12,631 - UNCLASSIFIED ORNL—~LR~-DWG 25923 TEMPERATURE (°C) 915 900 850 800 750 1000 800 600 400 200 100 80 60 40 PRESSURE (mm Hg) 20 10 8 6 a : 084 086 088 080 092 094 096 098 (.00 10/T (°K) Fig. 2.3.1. Sublimation Pressures of HfF‘ and ZeF . 144 two fluorides could possibly be used as @ basis for separating zirconium from hafium in a high- temperature (above 925°C) process. It also might be useful in the production of ZrF, from a starting mixture of zirconium and hafnium oxides. The oxides would be hydrofluorinated and ZrF, would be separated by heating ond preferential subliming. "YAPOR PRESSURE OF Ber S. Cantor An early determination of the vapor pressures of BeF, was reported by Moore, but the scatter in the measurements was larger thon is now expected with improved techniques. Therefore new meas- urements were made by using the Rodebush-Dixon method. The values obtained, as given in Table 2.3.2, fit the equation 10,967 T (°K) ~ log p (mm Hg) = 10.487 - The vapor pressures of BeF, in the range 800 to 1025°C have also been measured by Sense et al, %> by using the carrier gas method and assuming that the vapor species was monomeric. Their most recent results? are in very good agreement with the data given in Table 23.2. Comparisons of three quantities derived from the vapor pressures are presented in Table 2.3.3, The agreement between the values obtained by the direct and the carrier-gas measurements indi- cates that BeF, vapor is monomeric, Accordingly it should be possible to measure activities from vapor pressures over BeF,-rich solutions, where the vapor could be proved to be predominantly BeF,. A svitable system for an investigation of this type is LiF-BeF .. 1 « H. Rodebush and A. L. Dixon, Phys. Rev. 26, W 851 (1925). 2 S. Cantor, ANP Quar. Prog. Rep. June 10, 1956, ORNL-2106, p 113. Q & Rep J 3R. E. Moore, ANP Quar. Prog. Rep. Sept. 10, 1952, "ORNL-1294, p 151. 4K. A. Sense, M. J. Snyder, and J. W. Clegg, J. Phys. Chem. 58, 223 (1954). K. A. Sense, R. W. Stone, and R. B. Filbert, Jr., Vapor Pressure and Equilibrium Studies of the Sodium fgg%ide-Beryllium Fluoride System, BMI-1186 (May 27, ”n n Table 2.3.1. Vapor Pressures of HF, PERIOD ENDING SEPTEMBER 30, 1957 Pressure {mm Hg) Table 2,3.2, Vapor Pressures of BeF, Temperature Temperature Pressure (mm Hg) (?C) -Observed Calculated (°C) Observed Calculated 730.9 4.65 4.54 872.4 8.05 8.20 - 739.3 5,65 5.70 896.7 13.0 13.0 754.4 8.30 8.59. 912.5 17.1 17.3 755.5 8.75 8.81 917.6 19.2 18.9 764.1 10,9 1.0 936.9 26.5 26.6 775.9 14.5 46 " 958,5 38.4 38.3 778.0 153 15.7 ' 978.1 53.7 52,8 786.8 19,4 19.6 982.4 57.4 56.6 798.8 25.5 26.5 1000 76.2 74.8 806.6 32 318 : 1009 86.8 85.9 808.3 - 328 33.2 1024 - 108.6 107.9 818.4 - 43,1 42,1 1026 112.5 111.2 827.8 50.6 522 1038 133.4 132.4 830.7 554 560 1050 156.3 158.1 8319 56.4 515 - 1081 182.9 | 180.7 839.3 65.6 679 1070 210.9 209.9 840.2 - 67.3 69.5 - 1081 245.8 244.9 849.1 827 - 847 . 1088 267.5 269.8 859.5 1097 1062 1099 305.2 3119 8647 _120.2 . e 881.1 - - 206.8 1954 - 310 3027 - 909.7 ' Tnb].e‘ 2.3.3. Com-pdflso_n of Quantities Derived from Ber Vapor Pressures From Work of-’Sen_sE et al. This Work Vap_or pr:ejs‘su':re' 'eé;udfion_ e ‘Heat of vaporizaflon S ) Ex'l'ropolofed bolilng polni - 10 943 | _Iogp = 10.466 -——T— - 50 1 kcal/mole o | "_-_71170°c S ~ logp = 10.487 - 10,967 | :50.2 kcal/mole ) ‘!169°C 145 ANP PROJECT PROGRESS REPORT 2.4. PRODUCTION OF PURIFIED MIXTURES J. P. Blakely PREPARATION OF VARIOUS SPECIAL FLUORIDES B. J. Sturm Preparation of YF, The amount of YF, required by the ORNL Metal- lurgy Division in the experimental production of ytrium metal has continved to increase. Plans are being considered for changing the method of preparing YF; from the wet process to the direct hydrofluorination of Y,0, in order to meet the increasing demands. Approximately 6000 g of YF, was prepared by dissolving Y,0, in aqueous HCI solution, filter- ing, precipitating the fluoride by the addition of aqueous HF (48%), washing by decantation, re- covering the solid by centrifugation, and drying at 150°C. The partially dried material was given the standard hydrogen=hydrogen fluoride treatment following the addition of the desired quantities of MgF, and LiF. Approximately 8000 g of the processed YF,;-MgF,-LiF mixture was unioaded from the receivers, crushed to a suitable size, and loaded into charging vessels supplied by the Metaliurgy Division. Preparation of Other Fluorides Approximately 5 Ib of CrF, was prepared by treating hydrated chromic fluoride with NH,F.-HF and decomposing the resulting (NH,),CrF, at 450°C. Most of the CrF; was converted to CrF, by hydrogen reduction at 700 to 800°C. A silver reactor was used in a recent hydrogen reduction run, but the resulting CrF, has not been analyzed. The complete dissolution of CrF, frequently is G. J. Nessle L. G. Overholser difficult to obtain, and, as a result, the analytical values are not as reliable as desired. A suitable flux that will not oxidize CrF, but will render it more readily soluble is being sought. One run, in which equimolar quantities of ZrF, and CrF, were fused at 900°C in a sealed capsule, yielded a product that was readily soluble. ' Additional work is necessary before the stability of GF, in this medium can be established. A small batch of VF, was prepared by con- densing HF on ¢ V,0, powder, heating to remove the water and excess HF, and then hydrofluorinat- ing at 200°C. PRODucnoN",jA-»_ND DISPENSING OF FUSED SALTS AND LIQUID METALS Experimental Batch Production C. R. Croft - J. Truitt Experimental batches of various compositions were prepared for use inphase equilibrium studies, physical properties testing, and corrosion testing. A total of 442 kg of halide salts was prepared in 41 batches. Seventeen of the 41 batches were NoF-KF-LiF-UF, (11.2-41-45.3-2.5 mole %, fuel 107) for use by the Metallurgy Division in studies of nickel-molybdenum base alloys. These 17 batches averaged 20 kg each. Conversion of 10 Ib of low-zirconium HfCl, to HfF, was completed with a total recovery of 96.3%. The total input was 4300 g; the total _output was 3288 g; and the theoretical output was 3415 g. Chemical analyses of the three batches in which the material was converted are presented in Table 2.4.1. Table 2.4.1. Chemical Analyses of Three Batches of HfF ; Prepared from Low-Zirconium HfCI, Botch Major Constituents (wt %) Contaminants {(ppm) Number Hf F Cl Ni ‘ Cr Fe 1 68.7 30.0 0.30 175 115 1295 2 69.4 29.4 0.27 175 160 1225 3 69.3 29.5 0.29 125 175 1270 146 ] s 0 Four batches of LiCI-KCl were prepared for corrosion and physical properties testing, Three batches of YF,-LiF-MgF, and seven batches of KF-BeF ,-ZrF , were prepared for chemical studies, In addition, batches of BeF, and NaF.BeF, were prepared by over-the-surface treatment (non- bubbling) with HF. Dispensing and Servicing Opgidfiqns F. A. Doss D. C. Wood During the quarter, 100 filling and draining op- erations were performed that involved 785 kg of liquid metals and 1229 kg of salts. The 1229 kg of salts was made up of 419 kg of new material and 810 kg of salvaged material. The salvaged material is being used in testing of the fluoride- volatility process for the recovery of uranium. This salvaged material is not included in the mo- terial balance given in the following section. A series of experiments for determining the amount of fuel which will be held up on various internal surfaces of the ART was initiated. The surfaces to be studied are various sizes of tubing held at various angles to the horizontal, machined flat plates held at various angles to the horizontal, a section of the heat-exchanger tube-to-header junction, a mockup of a tube bundle, a 3-in.-dia bellows, and various right-angle welded joints. The pieces being studied are appropriately sup- ported in a small flange-top receiver, the receiver is filled to immerse the piece in NaF-ZrF ,-UF, (56-39-5 mole %, fuel 70) at 1500°C for 2 hr, the temperature is allowed to drop to 1200°F (the draining temperature of the ART), and the receiver "PERIOD ENDING SEPTEMBER 30, 1957 is emptied through a dip line to simulate draining. Tests have been completed on five of the 17 pieces to be examined. Materials Available J. E. Eorgan F. A. Doss D. C. Wood “Upon termination of operation of the production- scale facility at the end of June 1957, the mo- terials on hand consisted of 11,559 Ib of NaF.ZrF ,, 1810 Ib of low-hafnium ZrF,, 1685 Ib of normal ZrF,, 200 Ib of NaF, two new 58-in. reaction vessels, two reworked 48-in. reaction vessels, one new furnace liner, and seven new storage containers. The higher-than-expected amount of NaF.ZrF, on hand resulted from the receipt of 10,840 ib from Kawecki Chemical Company, who took advantage of the 4000-1b overage allowed in a purchase order for 30,000 Ib of material. The types and emounts of purified fluoride mix- tures on hand on July 30, 1957, are listed below: ‘Composi- Amount tion No. Type (kg) 30 NaF-ZrF ,-UF 4 (50-46-4 mole %) 5860 31 NaF-ZrF, (50-50 mole %) 933 70 NaF-ZrF ,-UF 4 (56-39-5 mole %) 900 107 NaF-KF-LiF-UF, (11.2-41-45.3-2.5 119 mole %) VYarious special mixtures 552 Total '8_3-6:4— 147 ANP PROJECT PROGRESS REPORT 2.5. ANALYTICAL CHEMISTRY J. C. White DETERMINATION OF TRACES OF NaK IN AIR A. S. Meyer, Jr. J. P. Young Additional modifications were made in the optical system of the instrument for the detection of NaK in air by measurement of the absorption of sodium resonance radiation by sodium atoms.! These modifications, which were first discussed in the previous report in this series,? include the insertion of an additional light stop which is designed to exclude the continuum radiation of the heated furnace from the photomultiplier tube. The continuum radiation caused a near-saturation phenomenon with respect to the light seen by the phototube and thereby severely reduced the sensitivity of the instrument when the temperature of the furnace exceeded 700°C. It was suggested® that the light stop consist of a plate with two small apertures approximately 0.08 in. in diometer. The aperture plate (shown in Fig. 2.5.1) is placed in the focal plane of 1A, s. Meyer, Jr., and J. P. Young, ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 137. 2. s, Meyer, Jr., and J. P. Young, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 166. 3G. K. Werner, private communication to A. S. Meyer, Jr, INLET AIR SAMPLE FURNACE QUARTZ WINDOWS the image of the quartz end windows of the ab- sorption cells; the image is formed by the con- densing lens. The two circular holes in this plate correspond precisely to the images of the end windows. With this arrangement, all radiation except that from the heated quartz windows is masked by the light stop. An additional condensing lens is required to refocus the beams on the photocell. The positions of both the aperture plate and the small. condensing lens can be adjusted along the axis of the optical system to provide sharp focusing. A multilayer interference filter and several possible com- binations of glass and gelatin color filters have also been included to isolate more effectively the resonance radiation at 589 my. Since the openings in the added cperture plate are relatively small, much more precise focusing and alignment of the optical system are necessary than were required for the original system. Ac- cordingly, o base assembly is being fabricated which provides for precise adjustments of the positions of both the chopper and detector com- ponents. The arrangement is designed according to the principles delineated by Strong,4 and it 4J Strong, Procedures in Experimental Physics, p 585-592, Prentice-Hall, New York, 1938. UNCLASSIFIED ORNL-LR-DWG 25924 ABSORPTION CELLS SODIUM VAPOR LAMP INTERFERENCE FILTER APERTURE V 7 //// // //4 APERTURE~ / @ Z:f_"___f__ ANALYZER e FURNACE CHOPPER EFFLUENT ASSEMBLY AIR SAMPLE QUARTZ WINDOWS —y— DETECTOR ASSEMBLY Fig. 2.5.1. Instrument for the Detection of Submicrogrom Quantities of NaK in Air. 148 of =} L @ provides for movement of the optical components in any desired direction with a single degree of freedom. All the components are mounted on -a 7-ft section of 18-in. channel iron, which serves as a rigid optical bench. Each end of the channel iron is -slotted to accommodate bushings which locate the base plates for the chopper and detector assemblies. These base plates can be moved along the optical axis, for coarse focusing, or can be rotated about a vertical axis. A second plate is supported on each base plate by ball bearings in opposing V.grooves. The second plate is spring-loaded with respect to the base plate and is driven by o screw-drive mechanism to provide motion at right angles to the optical axis of the system. The chopper and detector sections are each supported by three vertically adjustable - screws whose spherical ends rest in radial V-grooves in the second plates. Ad- justments of these screws can produce tilting motions -in any desired direction with negligible translational motion. Fabrication and gssembly of the modified instrument are about 90% completed. DETERMINATION OF OXYGEN IN METALLIC LITHIUM A. S. Meyer, Jr. 'R. E. Feathers Further studies of the butyl iodide method for the determination of oxygen in metallic lithium have revealed that several unexpected reactions occur during the dissolution of the sample in "ethereal solutions of iodine and butyl iodide and the subsequent extraction of the inorganic compounds into -aqueous. solutions for titration. It was demonstrated that oxygen, “when present ‘as LiOH, cannot be. defermmed by this. method. No ucud was consumed in the titration ‘of the . ‘aqueous extract of- the soluhon which was formed‘ : “ by the reactlon of 500 mg of LiOH with the ethereal reagents. The following reaction of ._L!OH with iodine is proposed to explain the elimination of the hydrox:de durlng the dissolution reaction: L|0H + 1, -—>L|I + HOI The hypo:odous acud is oo weak for tltrcmon cnd' is probably decomposed to -water, iodine, and oxygen in the nonaqueous medium. The similar - - reaction which may occur with Li,0, Li20 + I2~—-—>Li0| + Lil PERIOD ENDING SEPTEMBER 30, 1957 does not interfere with the determination of oxygen as Li,0. In the titration with acid, the hypoiodite consumes a quantity of acid which is equivalent to that required to titrate the oxide from which it is formed: - - L4 ' oI + I~ + 2H 7——>I2+H2O The butyl iodide method appears therefore to be essentially specific for the determination of . 'Li20' The elimination of the alkalinity of Li 3N and Li,C, by reaction with iodine was dlscussed prevuously. ~ Lithium carbonate is the only re- maining alkaline constituent which might be titrated. It is unlikely, however, that Li,CO, will be present in sngnlflcant concentrahons m samples which have been heated to elevated temperatures, since it is not compatible with metallic lithium. A small negative bias has been observed in determinations by the butyl iodide method that is believed to be caused by traces of acid which are liberated by the reaction H20 + Iz-% HI + HIO which occurs during the extraction of the ethereal solution with water. The equilibrium constant for this reaction has been reported® to be approxi- mately 10~13, On the basis of this constant, less than one microequivalent of acid would be liberated by this reaction under the conditions of the titration. When ether is present as a third phase, however, the equilibrium is displaced to the right and hydrogen ions are liberated in ‘quantities - equivalent to 50 to 100 ppm for a ~1-g sample. The magnitude of this error is a “function of the amount of excess iodine which remains after the dissolution of the sample and _'fhe'refore 'is not easily corrected for by blank ~ determinations. The only explanation which has been postulated for the increase in the hydrolysis " reaction ‘is that hypoiodous acid is complexed or preferenhally extracted to a high degree by ether. The error can be eliminated by removing the 'free -iodine before the organic solution is ex- “tracted with water. A distillation method for the SA. S. Meyer, Jr., et al., ANP Quar. Prog. Rep. Sept. 10, 1956, ORNL-2157, p 127. SN. V. Sidgwick, The Chemical Eléménts and Their Compounds, vol 1l, p 1213, Clarendon Press, Oxford, 1950. 149 ANP PROJECT PROGRESS REPORT removal of the iodine and the ether was tested and was found to be impractical. The iodine was rapidly removed, however, by reduction with finely divided silver or with mercury. Mercury was found to be the more convenient reagent, since an exfensive preparation procedure must be carried out to obtain silver in o sufficiently finely divided form for ropid reaction. The effec- tiveness of this procedure in eliminating the negative interference was demonstrated by titration of reagent blanks end by carrying 500-1g sarnples, of Li, O through -the procedure. - Approxlmutely-' 80% of the odded Li, 0 was titrated. . In view. of the possible confumlnaflon with moisture during” transfer of small quantities of Li,O, a recovery of 80% was considered to be satisfactory. in the presence of iodide, the excess silver or mercury is oxidized rapidly by atmospheric oxygen, These reactions,’ indicated in ‘the fol- lowing equations, can consume hydrogen’ ions during titration and introduce: posmve errors: 4Ag + 41 + O2 + 4H* ——>4Ag| + 2H20 2Hg + 81~ + 0, + 4H*-.->2Hg| == + 2H,0 The positive - errors can. of course be - avoxded" by excluding. oxygen durmg fhc ‘extraction of The errors can be eliminated more convemently, however, by the addition of enhydrous HgCI after the reduction of the iodine. The HgCI, complexes the iodide according to the equcfion 217 + HgCl,—> Hgl, + 2Cl~ The extraction and titration are then, in effect, carried out in a chloride system, A proposed procedure based on the considerations discussed above is now being tested on samples of metallic lithium. The dissolution is carried out in the opparatus shown in Fig. 2.5.2. With the cooling jacket of the condenser drained and an empty reaction vessel in position, the con- denser is flushed by transferring 25 ml of anhy- drous ether to the reaction flask. lodine {15 g) and a magnetic stirring bar are transferred to a second reaction vessel immediately after it is removed from a drying oven, and, with a positive pressure of helium in the condenser, the reaction vessel containing the iodine is substituted for that containing the flushing. 150 solution, The system is then evacuated and purged with helium; ond a 100-ml wvolume of n-butyl iodide, which is stored over activated aluming, is transferred to the reaction vessel, along with 100 ml of diethyl ether, which is stored over metallic sodium. The solution is then stirred until the iodine is mixed thoroughly. The sample of 1 to 2 g of metallic_ lithium is added through the condenser from a Jomesbury “valve sampler.” The solution is heated at reflux . temperatures and stirred until dissolution of the metal is completed. An excess of iodine should be evident at this time. A 10-ml aliquot of the organic solution is withdrawn, and the con- centration of iodide in the solution is determined by titration with AgNO,. sample is calculated from this titration. A 10-ml portion of mercury is then added, and the solution is stirred until the color of the iodine is discharged. - A 30-g quentity of anhydrous HgCl, is added,” and the solution in the flask is snrred for 5 min to permit complexing of the iodide ion. The solution is transferred from the flask to o separotory funnel and extracted with recently boiled distilled water which has been pretitrated to a pH of approximately 5. The . aqueous extract is then 'ritroted with 0.01 or - 0.002 N'HCI. the organic solution .and . dunng _the fitrgtion. There has, as yet, . been no opportunlty to test this modification of the. procedure by analyzing 'hlghly purified samples of metallic lithium. Excellent recovery was attained when a 0.5-mg ~sample of Li,O was added to a simulated sample solution which was prepared from anhydrous Lil, butyl iodide, ether, and iodine. As soon as a sufficiently reproducible series of samples is available the results of determinations by this procedure will be compared with unolyses carried out by other methods DETERMINATION OF OXYGEN IN.FLUORIDE SALTS AND IN METALS A. S. Meyer, Jr. G. Goldberg The apparatus for the determination of oxygen as oxides in fluoride salts and in metals by the 7A. s, Meyer, Jr., R. E. Feathers, and G, Goldberg, Ahlli' Quar. Prog. Rep. March 31, 1957, ORNL.-2274, p BA, S. Meyer, Jr., G. Goldberg, and B. L. McDowell, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, - p 167 The weight of the u} - GLASS-TO-METAL SWAGELOK FITTING - PRESSURE R ELl EF BELLOWS VALVE HELIUM ——== 'ETHER WATER-COOLED JACKET— "~ SERUM . Blliter FLASK - FRITTED FILTER STOPPER _fi UNCLASSIFIED ORNL—LR—DWG 25925 SAMPLE PORT FITTED TO JAMESBURY-VALVE TRANSFER CHAMBER INLET FOR HELIUM, VACUUM 14Y5 in. r\l- wfijw\ sTO PPER‘—\ ™~ WATER- q SERUM COOLED | CONDENSER - BUTYL IODIDE REACTION VESSEL ™~—MAGNETIC STIRRER FLASK 3-liter FRITTED FILTER Fig. 2.5.2. Apporatus for the Determination of Oxygen in Metallic Lithium, 1S1 £561 0 ¥IGWILJ3IS ONIANI Q0Id3d ANP PROJECT PROGRESS REPORT fluorination of the oxides with KBrF, was as- sembled. This apparatus, which is adapted from the work of Hoekstra and Katz,? is shown in Fig. 2.5.3. A detciled procedure was formulated and is presently being given an initial test. Briefly, the reagent, KBrF,, is formed by the addition of BrF; to KF in the reaction tube. Sufficient KBrF, is made to permit the andlysis of a number of samples without. recharging the system with fresh reagent for each determination. The reaction tube is cooled to the temperature of liquid nitrogen, and, with argon flowing into the tube, the sample is quickly added to the tube, which is then caopped. The temperature of the reactor is then raised to operating temperatures, about - 400°C, so that the KBrF, can react quanmotlve!y‘ with the metallic oxides present and liberate oxy- gen gas. This gas.is passed through @ series of traps to remove bromine and excess recgent. "An * automatic Toepler pump is used to transfer the oxygen fto either of two pressure-measuring de- vices - a mercury manometer or, if greater sen- sitivity is desired, an oil manometer. . Since the volume of the measuring system cannot be obtained directly, the system must be calibrated by carrying out the procedure with samples of known oxygen content. Zirconium oxide has been selected as a standard for calibration. On the basis of estimated wvolumes of the measuring system, quantities of oxygen as low as 5 pg can be detected. The range of the apparatus caon be extended upward by severdl orders of magnitude by increasing the volume of the measuring system. DETERMINATION OF NICKEL IN METALLIC SODIUM A. S. Meyer, Jr. B. L. McDowell The method for the determination of nickel in which the absorbance of its chelate complex with 4-isopropyl-1,2-cyclohexanedionedioxime in xy- lene'® is measured was applied to -onalyses of samples from studies which are being carried out by the Metallurgy Division in order to deter- mine the solubility of nickel in molten sodium. 4. R. Hoekstra and J. J. Katz, Anal. Chem. 25, 1608 (1953). 104, 5. Meyer, Jr., and B. L. McDowell, ANP. Quar ‘ “""is converted to a compression fitting by facing Prog ‘Rep. June 30, 1957, ORNL-2340, p 168." 152 The samples are prepared by heating metallic sodium under an atmosphere of helium in a nickel crucible which is contained in a welded Inconel capsule. After the sample hds been maintained at the desired temperature for a period sufficient to establish equilibrium, the molten metal is transferred to an opposing molybdenum bucket by inverting the Inconel capsule. The sample is submitted for analysis in the molybdenum bucket.. The sample includes the contents of . the bucket together with ‘any nickel deposrfed on the inner walls of the bucket. - Initial determinations were carried out by dis- solving the metallic sodium in 50 m! of methanol. An aliquot of the methanol solution was titrated with standard acid to determine the sample weight. The moiybdenum ‘bucket was -rinsed with warm 1:3 HCI, ond the washings were combined with the methanol solution, which was acidified and - ‘evaporated toa volume of approximately 25 ml to remove the methanol. In the determinations, an unusual interference by molybdenum, which was dissolved from the con- tainer, -was experienced. It was found that molyb- ~denum -(an orange-colored extract) .interfered only when methanol or some contaminant thereof was present and when cyanide and Sulfi-Down, added _as -masking agents, had been used. No colored extract was observed when the chelating agent, 4-isopropyl-1,2-cyclohexanedionedioxime, was ab- sent, and the interference appeared to be in- tensified by the presence of nickel. Since the interfering reactions appeared to be exceedingly complex and could be eliminated by dissolving the samples in water rather than in methanol, the nature of the reactions has not yet been studied in detail. Because the outer walls of the molybdenum bucket may be contaminated with nickel as a result of contact with the Inconel capsule, it is necessary to exclude from the sample the water used for the washing of the outer surfaces of the bucket. This requirement presented special problems when the samples were dissolved in water and required the design of the new ap- paratys, fabricated from fiuorothene, which is shown in Fig. 2.5.4. The apparatus is con- structed by threading a modified ‘/2-ih. flared . fluorothene fitting through the bottom of a 250-ml ' heavy-walled fluorothene beaker, The flared fitting ‘l gsl VACUUM GAGE UNCLASSIFIED ORNL—LR—-DWG 25926 VACUUM . le 53 PLvdb ety rr by rrrrer MANOMETER ey o=t I' & oI ManoMETER 1P i hs; S N / blFFERENTlAL MANOMETER N TOEPLER PUMP MANIFOLD GLASS-TO-METAL SEAL V—INCONEL DIAPHRAGM VALVES (HOKE, INC) S—HIGH-VACUUM STOPCOCKS '. Flg. 25.3. Apparatus for the Determination of Oxygen In Fluoride Salts and in Metals, =0 {?S. PRESSURE REGULATOR FOR HELJUM SUPPLY (¥ 24, PYREX TRAP LS61 ‘0c ¥IIWILHIS ONIANI goIyad ANP PROJECT PROGRESS REPORT BRASS NUT UNCLASSIFIED ORNL—~LR-~DWG 25927 INERT GAS INLET P RuBBER STOPPER FLUOROTHENE BEAKER MODIFIED FLARE FITTING RUBBER SLEEVE RETAINING RING MOLYBDENUM BUCKET APPROXIMATELY ‘/3 FULL OF SODIUM Fig. 2.5.4. Apparatus for Dissolution of Sodium Samples in Water, off approximm‘el¥ ‘/8 in. of the tapered end ond turning into it a /2-in.-!D recess 7 in. deep. The molybdenum bucket is held in place in this recess and sealed by a Y-in. section of 3/8-in. rubber tubing. The rubber is backed by o brass ring (back ferrule of a Swagelok fitting) and compressed by the brass nut of a flared fitting. The apparatus can easily be modified to accommodate cylindrical containers of smaller diameters. In the dissolution of the samples the apporatus is loosely covered with a rubber stopper which ig fitted with on inlet for inert gas. An inert gas is passed ropidly through the fitting for approximately 30 sec, after which the sample bucket -is inserted and locked into position. After the bucket is cooled in an ice water bath, approximately 25 g of ice, which is prepared by freezing demineralized water in plastic tubes, is added to the becker. While the flow of inert gas is continued, the stopper is removed suffi- ciently for water to be added cautiously until 154 the sample storts to react. The beaker is again covered, and an atmosphere of inert gas is main- tained until the dissolution is completed. Approxi- mately 5 min is required to complete the reaction. FORMATION OF CARBIDES BY THE REACTION OF PUMP LUBRICANTS AND NaK AT ELEVATED TEMPERATURES A. S. Meyer, Jr. G. Goldberg During the dissolution of alkali metdls and alkali metal residues, an odor that is commonly associated with acetylene has frequently been detected. This phenomenon has been most pro- nounced during the chemical examination of cold traps by the method described previously,'! When water was added to cold traps from NaK systems after most of the residual alkali metals A, s. Meyer, Jr., and G. Goldberg, ANP Quar. Prog. Rep. June 10, 1956, ORNL-2106, p 128. " had been converted to the halides by reacting them with butyl bromide, an intense odor was detected and minor explosions were experienced, even though an attempt was made to maintain an inert atmosphere within the system. The presence of acetylene in the evolved gases was sub- stantiated when precipitates of Ag,C, were formed upon passage of the gases through solutions of AgNO Recenfly an explosion occurred when an ex- perimental NaK system was opened after operation had been terminated because of a leak in the pump seals. In view of the well-known acetylene explosion hazard, it was deemed necessary to investigate the possibility that carbides had formed as a result of a reaction between the pump lubricant and the alkali metals, with the subsequent liberation of acetylene. In earlier tests of the compatibility of lubricants with molten alkali metals,!? there was no ob- servable reaction between the alkali metals and the lubricants. The tests were carried out, how- ever, to determine whether a reaction would occur if alkali metals at 1400°F leaked into the {ubricant at 400°F. The conditions of these tests did not approximate, therefore, the temperatures which would prevail during a leak of the pump oil into the NaK being pumped. In order to test for the formation of carbides, the reaction was investigated in the apparatus shown in Fig. 2.5.5. The reaction tube is « section of 3/ -in.-OD, 0.065-in.-wall Inconel tubing, which can be heated in a vertical tube furnace. A thermocouple is spot-welded to the tubing. The upper section of the apparatus, which is fitted with a cooling coil, provides an expansion chamber and reflux surfaces for the lubricant and prevents the formohon of dangerous pressures, A pressure gage is fitted fo the ‘/4-|n. Swagelok - fitting.- A bellows valve is fitted to the 3/8-|n.' Swagelok fitting after the reochon components are added to the apparatus. = For the test,2 ml of NaK and 2 ml of Gulfspln-35 lubricant were added ‘to the reactor, “which was “then evocuuted ond purged with helivm. . The bellows valve was. closed and the lower section of the reaction tube was heated to a temperature of 650°C. No. sugmhca_nt increase in pressure 12G, Goldberg, A. S. Meyer, Jr., and J. C. White, Compatibility of Pump Lubricants with Alkali Metals and Molten Fluoride Salts, ORNL-2168 (Dec. 28, 1956). PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR-DWG 25928 a—3/g-in. SWAGELOK FITTING Yfa=in, SWAGELOK FITTING ~* & COPPER COQOLING COIL. e _——— ) 2in =L )\|~ ri z ‘ REACTION TUBE ———= 4in. Fig. 25.5. Appoaratus for Sfudying the Reaction of NaK with Pump Lubricants. was observed wuntil the temperature reached approximately 600°C, at which time the pressure increased abruptly to 450 psi. The gases were later analyzed mass spectrographically and found to consist primarily of methane, with a trace of hydrogen. After the reaction tube had been main- tained at a temperature of 650°C for 4 hr, it was cooled and frozen with dry ice. A 3-in. section of the tubing was then cut off and placed "in‘ a stainless steel bomb which contained 200 ml - of frozen water. The bomb was sealed and allowed to warm to room temperature. When the reaction was completed, the ewvolved gases were passed 'through a 1.5 M solution of AgClO,, according’ -to the procedure which was developed for the determmcmon of Li C2 in metallic lithium.13 Copious quantities of Agzc were precipitated. The acetylene present was not determined quan- titatively because part of the precipitate was lost when a portion of the Ag, C, exploded. It is estimated that approxumcrfely 2% of the alkali 137, W. Gilbert, Jr., A. S. Meyer, Jr., and J. C. White, Spectrophotometric Determination of Lithium Carbide in Metallic Lithium as the Acetylene-Silver Perchlorate Complex, ORNL CF-56-12-111 (Dec. 27, 1956). 155 ANP PROJECT PROGRESS REPORT metal was present as the carbide. A heavy deposit of carbonaceous material remained in the reaction tube. Negligible quantities of carbides were found in a blank run in which the test was repeated with only NoK added to the reaction tube. The test is being repeated in order to obtain quantitative measurements of the carbides, and tests with sodium are proposed. DETERMINATION OF ALUMINUM IN MIXTURES OF FLUORIDE SALTS J. P. Young J.R. French An investigation of a spectrophotometric method for the determination of trace amounts of aluminum' in mixtures of fluoride salts with pyrocatechol violet was continued. The repro- ducibility of the data was improved by washing all glassware with ¢ dilute solution of pyrocatechol violet prior to use and by the addition of thiogly- colic acid to the test solutions. These modifi- cations lowered the coefficient of variation of the method to approximately 2%. A study was made of interferences that would normally be encountered in corrosion tests with NaF-KF-LiF-UF, and the newly developed nickel- molybdenum-base alloys. These alloys contain trace amounts of iron plus a number of nonferrous materials. Since pyrocatechol violet is a sensitive reagent for iron, essentially complete removal of iron is necessary before application of the reagent in the determination of aluminum. Removal of the iron is best accomplished by means of the mercury cathode. The interference by titanium is partially eliminated by measuring the absorbance of the solution at both 582 and 700 mp and correcting for the contribution at 582 mu from ftitanium. The simultaneous determination of titanium and aluminum in the scme test portion appears to be possible and is to be the subject of further work. Nickel, chromium, molybdenum, and niobium do not interfere in determinations by this method. Initial results obtained in the determination of aluminum in synthetic solutions of samples from corrosion tests were about 10% high. These results indicate possible accumulative inter- ferences, ond therefore further work is planned to improve the reliability of the determinations. 14, p. Young and J. R. French, ANP Quar. Prog. " Rep. June 30, 1957, ORNL-2340, p 168. 156 SOLVENT EXTRACTION OF TITANIUM AND NIOBIUM FROM ACIDIC SOLUTIONS WITH TRI-z-OCTYLPHOSPHINE OXIDE Studies of the application of tri-n-octylphosphine oxide (TOPO) in the solvent extraction of metals of interest in the ANP program were continued. Since sulfuric acid is commonly used in the dis- solution of fluoride salts, the extraction char- acteristics of sulfuric acid by TOPO were also investigated. Extraction of Sulfuric Acid T. W. Gilbert Solutions of 1 to 8 M sulfuric acid were ex- tracted with 0.1 M tri-n-octylphosphine oxide (TOPO) in cyclohexone. The quontity of sulfuric acid in the organic phase was determined directly by titration with alcoholic potassium hydroxide ond also by back-extraction of the organic phase with water. The results were the same by either method. The quantity of sulfuric acid extracted was found to increase almost linearly with in- creasing concentration of sulfuric acid. From 7.0 M acid, 0.77 mmole of sulfuric acid was extracted by 1.0 mmole of TOPO. A third phase formed when 8.0 M acid was extracted. The third phase was analyzed for total acid and for phosphorus. The ratio of H,50, to TOPO obtained was 1.05. Extraction of Titanium T. W. Gilbert Tri-n-octylphosphine oxide (TOPO) dissolved in cyclohexane has proved to be an efficient ex- tractant for titanium from solutions of sulfuric acid. The efficiency of extraction was found to be highly dependent on the concentration of sulfuric acid in the aqueous phase. In the ex- traction of 0.16 mmole of titanium with 1.0 mmole of TOPO, only 6% of the titanium was extracted from 1 M sulfuric acid. From 7 M sulfuric acid, however, 99% of the titanium was extracted. From 8 M acid, 99% of the titanium was removed from the aqueous phase, but a third liquid phase was formed. The extraction of titanium from solutions 0.5 M in tartaric acid in the presence of varying amounts of sulfuric acid was also investigated. The resuits were very similar to those obtained in the presence of sulfuric acid alone; only a slight improvement in the extraction of titanium was observed. -, A study of the extracthn of increasing amounts of titanium from 6 M sulfuric acid with 1.0 mmole of TOPO showed that saturation of the organic phase occurred when 0.5 mmole of titanium had been extracted. This result, when combined with the results of an acidimetric study of the sodium fluoride solutions used to strip the titanium from the organic phase, indicates that the principal species present in the cyclohexane phase at saturation is TiOSO,.2TOPO. Extraction of Niobium C. A. Horton Niobium(V) extracts best into TOPO in cyclo- hexane from hydrochloric acid solution, fairly well from sulfuric acid, and' almost not ot ali from nitric or perchloric acid, with approximate extraction coefficients for approximately 4 M acid of 200, 20, 0.02, and 0.01, respectively. For stable acidic solutions of over 0.2 mg of niobium, the presence of citric, lactic, tartaric, or oxalic acids or thiocyonate, peroxide, or catechol are required to keep the niobium in solution. For 500 pg of niobium in 0.2 M tartaric acid, maximum extraction is obtained from 4 M hydrochloric acid, as shown in Fig. 2.5.6. The complexing agents must be added before the inorganic acids in order to keep the niobium in solution. The niobium in the organic layer can be estimated by adding catechol and butyl acetate to an aliquot of the extract and measuring the absorbance at 360 mu. : INSTALLATION OF A MICROBALANCE IN A YVACUUM DRYBOX A. S. Meyér, Jr. R. E. Feathers The welghlng compartment of a Cahn Electro- balance 13 has been installed in a vacuum drybox. This’ electrobalance can be adjusted for weighing ~ quontities of . the - order of 5, 10, 20, and 50 mg. - The balance operates on the principle of equating - the- torque of the sample with an opposmg electro- magnehc torque. The components which are mstulled w:fhm the drybox mclude the weighing compartment, four - connecting leads, a ‘terminal strip, and the torque ISCuhn Electrobalance, Model M-10, Cahn Instrument Company, Downey, Calif. EXTRACTION COEFFICIENT (orbitrary units) PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR-DWG 25929 200 100 80 60 40 20 o i 2 3 4 5 & T 8 9 10 CONCENTRATION OF HCI (M) Fig. 2.5.6. Effect of HCl Concentration on Extraction of Niobium with Tri-n-octylphosphine Oxide. motor and balance beam, which are similar in construction to the movement and pointer of o galvanometer. All these components can be subjected to evacuation without damage and do not offer any problems of excessive outgassing. Other components, which cannot be conveniently evacuated, including batteries, switches, and Helipot precision resistors, are retained in the original instrument cabinet outside the drybox. All balancing operations are thus carried out remotely by adjusting the currents to the torque motor. Electrical connections are effected through four insulated conductors, which are sealed through the walls of the drybox. This balance will be used principaily for the ~preparation of synthetic samples to which trace quantities of a constituent must be added with- out exposure to atmospheric contamination or moisture. It has already proved invaluable in the preparation of simulated somples for tests of the proposed method for the determination of ~oxygen in metallic lithium (see preceding section of this chapter). In these studies sub- milligram quantities of Li,O were combined with 20-g samples of anhydrous Lil. 157 Part 3 METALLURGY 1 S¥ B 3.1. NICKEL-MOLYBDENUM ALLOY DEVELOPMENT STUDIES MATERIAL DEVELOPMENT T. K. Roche J. H. Coobs - H. Inouye The developmental work on a new container alloy for the liquid fuel, NaF-KF-LiF-UF , (11.2-41-45.3- 2.5 mole %, fuel 107), being considered for use in odvanced. circulating-fuel reactors has been con- cerned primarily with three major problems. First, studies have been made for determining the optimum: alloy composition for service temperatures up to 1800°F. The second problem involves determining whether the most promising alloy, INOR-8, can be produced as a ‘‘commercial’’ item. = Since all nickel-base alloys exhibit mass transfer in sodium at the service temperatures of interest, the third problem is the fabrication of duplex tubing of - INOR-8 clad with stainless steel for use in heat - exchangers, The research and developmental studies of the - alloys had consisted of a- nickel-molybdenum _ screening program until recently, when sufficient data had been accumulated to indicate that INOR-8 was the most sat_isfdctory' of the alloys studied . in terms of the original objectives. The data on which the selection was based were obtained for - INOR-8 and seven other INOR-type alloys at tem- peratures up to 1500°F. Since forced-circulation loop tests, both in-pile and out-of-pile, have n_bt ‘ been completed, and, in addition, comprehensive weldability and strength evaluations remain to be made, two other compositions, INOR-8 modified and INOR-9, are aiso being studied as possible alternates if INOR-8 shows excessive corrosion ~attack. “or poor creep strength at 1800°F, The composi- tions of INOR-8, INOR-8 modified, and INOR-9 are given in Table 3.1,1. No data on INOR-8 modified are as yet available. -~ Tests of thermal-convection loops fabricated of INOR-8 in which fuel 107 was circulated showed ‘acceptable corrosion resistance of the alloy at 1500°F, but no tests have been run at 1800°F. It may be assumed, however, that if excessive cor- rosion occurs at 1800°F it will be due to the chromium content of the alloy, since in the tests at 1500°F the chromium content of the fuel in- creased with increases in the depth of corrosive Therefore INOR-9, which contains no chromium but is strengthened with niobium, is being studied as a possible substitute in the event that INOR-8 is not sufficiently corrosion resistant at 1800°F. On the other hand, INOR-8 may prove to be satisfactory from the corrosion standpoint and unsatisfactory in terms of creep strength, and - therefore the possibility of increasing its creep ‘strength with niobium additions, such as in INOR-8 modified, and with carbon additions is being in- vestigated. Although it is known that more than 0,10 wt % carbon usually prevents the production of tubing because of splitting along carbide boundaries, present indications are that more than 0.10 wt % carbon can be tolerated in the fabrication of plate and sheet. Thus it is important that the effect of “the carbon content on the creep strength of these alloys be determined. ~ content is found to be beneficial, thin-walled tubing If increasing the carbon Table 3.1.1. -Comppiitiohs of INOR-Type ‘Alloys- Being Considered as Container Materlals for Fuel 107 in the Tempercture R&ngg of 1500 to 1800°F - : ;'Allo)"f L - Nominal ‘Cc‘:mp‘osition.(wt %) ~ Designation” Mo Cr " Fe Nb Ni INOR-8 15-19 68 4-7 0 Balance ~INOR-8 modified 5-19 - 68 4-7 2-4 Balance INOR-9 4-7 2-5 Balance 12-19 0 ({89 (Co) 18 ANP PROJECT PROGRESS REPORT of a suitable carbon content may be produced by carburizing tubing containing 0.10% or less carbon. Properties of INOR-8 Physical Properties. — The density of INOR-8 has been determined to be 8.79 g/cm® or 0.317 Ib/in.3, and its modulus of elasticity at room tem- perature in the annealed condition is 32 x 10% psi. Data on its thermal expansion over various tem- perature ranges are given below: Tempercture Range Expansion CF) (pin./in. °F) 212--1832 8.6 212-752 7.0 752-1112 8.4 1112-1800 9.9 Fabricability. — The principal difficulty that has been experienced in the fabrication of INOR-8 has been the formation of large shrinkage cavities in cast ingots. While the tendency for ingots to form such cavities is inherent in the nature of freezing of all molten metals, the tendency seems to be somewhat accentuated in the case of INOR-8. It is not known at present whether this casting diffi- culty is attributable to the casting variables, the expulsion of dissolved gases, or the temperature range of freezing of the alloy. Irrespective of their cause, cavities in an ingot can, in the worst case, cause center bursts during forging, or they can, in less drastic situations, cause center-line defects that are detectable by ultrasonic inspec- tion. Present approaches to a solution of this problem include changes in casting practice, con- sumable arc melting, and subjecting the molten metal to a carbon boil. In all other respects, the forgeability of the INOR-8 alloy is excellent. Oxidation Resistance. — The oxidation resistance of the nickei-molybdenum alloys in air is strongly influenced by the chromium content. The studies made thus far have indicated that no chromium is better than a slight amount and that unless a critical amount is present the oxidation rate is not decreased to a level that can be considered to be acceptable. As the data in Table 3.1.2 indi- cate, the critical chromium content is approxi- mately 6 wt %. . Corrosion Resistance. ~ Tests of thermal-convec- tion loops fabricated of INOR-8 in which fuel 107 was circulated at @ maximum temperature of 1500°F 162 Table 3.1.2. Oxidation of Nickel-Molybdenum (17 wt %) Alloys at 1800°F as a Function of the Chromium Content of the Alloy Chromium Weight Gain of Content of Alloy Specimen in 168 hr (wt %) (mg/cm?)* 0 10.44 2.83 16.46 4.61 1245 5.43 17.83 6.19 0.44 6.86 (INOR-8) 0.60 7.97 0.31 *One mil of oxidation corresponds to a weight increase of about 4 mg/cmz. revealed corrosive attack to a depth of 1 to 3 mils in 500 hr, and mass transfer was observed when sodium was the circulated fluid. Forced-circulation loops fabricated of INOR-8 are to be operated in order to determine the effect of various amounts of iron and chromium on the corrosion resistance of the alloy when exposed to fuel 107. In-pile cap- sule tests have also been initiated in order to study the effect of radiation on the corrosion re- sistance of the INOR-8 alloy. | Mechanical Properties. — Details of the studies being made of the stress-rupture characteristics of the INOR-8 alloys are presented in a subsequent section of this chapter. Most of the tests thus far have been made ot a stress of 8000 psi ot 1500°F in fuel 107, The results of two tests at 1800°F and 3000 psi in fuel 107 are presented in Table 3.1.3. Data for Inconel tested at 1650°F are included for comparison. Properties of INOR-9 Alloys in the INOR-9 classification have a nickel base, 12 to 17 wt % molybdenum, up to 5 wt % niobium, and up to 7 wt % iron. The constituents of the INOR-9 alloy were chosen on the basis of corrosion resistance to fuel 107, Preliminary tests have indicated that the oxidation resistance of INOR-9 is about equal to that of Hastelloy B; the creep properties are about the same as those of INOR-8; and it is attacked by fuel 107 to a depth of 1 to 3 mils in 1000 hr ot 1500°F. In contrast to " » PERIOD ENDING SEPTEMBER 30, 1957 Table 3.1.3. Stress-Rupture Properties of INOR-8 Tested at a Stress of 3000 psi ot 1800°F in Fuel 107 Time to Specified Strain (hr) Time to Elongation Heat No. Source Rupture at Rupture 1% 2% 5% 10% (hr) (%) SP-16 Haynes Stellite Company 9 40 68 132 30-34 ORNL 3.4 8 24 50 94 20 Inconel* 6 16 40 130 *Tested at 1650°F (rather than at 1800 in NaF-ZrF“-UF4 {(50-46-4 mole %, fuel 30). the attack on the chromium-bearing alloys, such as INOR-8, the attack on the INOR-9 alloys does not increase when the exposure time is increased from 500 to 1000 hr, Further, the rate of attack does not seem to be affected by the niobium content up to 5% niobium, Studies at University of Tennessee! Studies of the solubility of chromium and molyb- denum in nickel have been made with the use of x-ray and metallographic methods. The composi- tions studied ranged from 20 to 25 wt % molybdenum and 3 to 15 wt % chromium. The beta phase, Ni Mo, of the nickel-molybdenum system was not observed in any of the ternary alioy specimens. It was concluded that the addition of chromium suppressed the formation of the beta phase. As the alloying content was increased, evidences of the existence of the gamma phase, Ni,Mo; the delta phase, NiMo; and the primitive orthorhombic *‘p'’ phase were observed. A report in the form of a thesis has been prepared by T.S. Lundy of the Umversnfy of Tennessee on this sub- ject. : Studies at Bafielle Memorial Institute 2. | The semicnnual ANP materials conference was held at Battelle Mernorlal Institute on June 17, 1957, and reports on. progress made during the | preceding six months were presented. A summary - ceived from Superior Tube Company for creep- of the information presented has been published.? In general, the work at BMI has ‘stressed the de- velopment of alloys with superior strength; thus lUnder subcontract 582, 2Under subcontract 979. 3E. M. Simmons, Semiannual ANP Meeting, Battelle Memorial Institute, ORNL CF-57-6-77 (June 17, 1957). ~ studies have been made of alloys containing alumi- num and molybdenum as the principal strengtheners. Several compositions were evaluated, and heat B-3277 (nominal composition: 20% Mo-7% Cr- 1.5% Al-2% Nb-0.15% C) showed outstanding creep strength, Thermal-convection loop tests of this alloy with fuel 107 as the circulated fluid showed that it was prone to attack by the salt be- cause of its aluminum and chromium contents, Attempts to operate forced-circulation loops fabri- cated of alloys of this type were unsuccessful because of the poor quality of available tubing. This contract has been terminated and a final report is being prepared. Studies of Stress-Rupture Properties at New Englond Materials Laboratory* Specimens from the production heats of INOR-1 through -6 (ref 5), which were prepared by the International Nickel Company, were tested to obtain their stress-rupture characteristics at 1350, 1500, and 1650°F in air at the New England Materials Laboratory. The results of the tests are summarized in Table 3,1.4. A final report on the work done under this subcontract is being prepared, INOR-8 Alloys Produced by Supenor Tube | Company® A series of alioys of the INOR-8 type was re- tupture evaluation. These alloys were air-melted at Superior and subsequently rolled to 0.065-in.- thick strip. Creep-rupture tests are to be run in 4Unr.ler subcontract 584. SFor compositions of these alloys see ANP Quar, Prog. Rep. March 31, 1957, ORNL.-2274, p 177. 6-Un_r:!er subcontract 1112, 163 ANP PROJECT PROGRESS REPORT fuel 107, A limited amount of 0.500-in.-OD, 0.045-in.-wall tubing was also processed from these alloys that is sufficient for the fabrication of three thermal-convection loops for operation with fuel 107, Fabrication of the loops will be started soon if inspection shows the quality of the material to be satisfactory., Descriptions of the alloys and the quantities awvailable are pre- sented in Table 3.1.5, These heats were prepared primarily for studies of the effect of zirconium additions on the strength and corrosion resistance of the INOR-8 alloys. The work at Superior Tube Company has also been concerned with the processing of tubing from material supplied by ORNL, International Nickel Company, Haynes Stellite Company, and Westing- house Electric Corporation. This work is described below under the contributions of the individual companies to the over-all alloy development program. Status of Preduction Heats Westinghouse Electric Corporation. ~ The work accomplished to date on the production of pilot heats of INOR-8 by Westinghouse was recently reviewed and their first periodic progress report on this development work was submitted to ORNL. Table 3.1.4. Summary of Results of Stres#-Rupfure Tests of INOR-Type Alloys 1 hr at 2000°F, air cooled 16 hr at 1500°F, air cooled Heat treatment: Stress to Rupture in 1000 hr (psi)* Elongation at Rupture in 100 hr (%)** _:‘:’:’: At 1350°F At 1500 At 1650°F At 1350°F At 1500°F At 1650°F INOR-1 13,000 3,900 1800 18 19 19 INOR-2 8,000 3,600 1600 1 12 20 INOR-3 15,000 5,100 2800 S 19 9 INOR-4 17,000 6,300 3400 1 5 n INOR-5 15,000 4,500 1700 22 26 32 INOR-6 18,000 11,000 2600 S 8 . 20 *Extrapolated values. **|nterpolated values. Table 3.1.5. Air-Melted INOR-8 Alloys Prepared by Superior Tube Company Nominal Composition (wt %) Number of 0.065-in.-Thick Total Length of 0.500-in.- Heat No. Creep-Rupture Specimen 0D, 0.045-in.-Wall Mo C Fe C Zr Ti Ni Blanks Received Tubing Received 5712 16.99 6.64 3.14 0.070 0.08 * Bal 5 14§ 2in. 5713 17.20 6.66 3.08 0.064 0.07 * Bal 5 18 ft 5714 15.13 6.71 2.18 0.032 0.03 Ba! 4 5715 1527 4.61 2.16 0.024 0.02 Bal 7 | 5716 16.60 6.64 5.34 0.046 Bal 2 2 1in. 5717 16.13 6.79 5.34 0.094 0.02 Bol 3 5718 15.88 5.33 2.27 0.058 0.03 Bal 1 6ft 1in. *Titanium present in heats 5712 and 5713, but analysis not yet complete. 164 v In summary, 1hree heats of the composmons specu- fied below have been prepared by lnduchon meltmg in air; . Amounts", - .Cem.ponenf:s. _ L (wt %) Molybdenum 1517 Chromium ' ) 6-8 . } Cdren e et S Carbon - o _.0-‘04-_""0-08‘.:7 Silicon 0.5 (max) Manganese 0.8 (mox) | Nickel ‘Balance The first heat (INOR-8M, ~5200 Ib) was poured into a 3000-1bmold, a 1500-1b mold, and two 300-ib molds. The composition of the alloy was within~ the specified range, except for the carbon content, which was found to be 0.13%, and: therefore forging studies were carried out on the maferlal One 300-1b ingot was worked on a 1000-ton press forge and the other onan 18,000-1b hammer forge. Both forging ‘methods worked the metal; but the press. forge moved the material easier than did the hammer - forge. The press forge worked one 300-Ib ingot to - approximately 4 in. in diameter, and the hammer forge made a 4-in. square.. Both ,forg_l_ngs were cooled and machined to 3-in.-dia bars. The bars were inspected by ultrasonic techniques and were found, for the most part, to be unsound, . The forger ingots were then sectioned and were also found to - contain internal flaws, The unsoundness of the | forgings was therefore ctmbufed to shrmkage_ cavities in the or:gmai mgots. S Two pOSSIb|e methods for. prevenhng 1he formu-_'_,.,_, R tion of cavities ‘during solldlflcuhon were proposed el Fiest, it was ‘suggested thut the mold be surrounded_'-;,._i__.‘~__"-7L':‘fi:]~ T by an insulating jocket so that the .ingot would ~ cool slowly and - 1haf the mold be desugned o' ‘vide~'an ingot with a. smaller. length-io-wndth ratio, The- second method mvolved mcreaslng the fhlck : " ness of the mold’ walls, accentuahng the taper of '-vdcuum-remelhng on fhe creep-rupture character- istics of this hlgh-corbon Westinghouse heat, a feeding the molten core“of the ingot and assure "~ 3-Ib ingot was prepared from the air-melted ma- ~- terial and tested, The results of the tests, which the mold -and chill- castmg the ingot.on a copper stool. Both of these methods ‘should facilitate progressive sohcl:flcaflon from the boflom to :lhe top of the ingot. The second heat (iNOR-SMZ ~500 |b) was cust' into a mold of the type suggested in the first PERIOD ENDING SEPTEMBER 30, 1957 “method proposed above, but- faulty pouring resulted ~in ‘mold-wash ‘and contamination of the alloy with - _cast iron. This ingot was not forged. ~ The “third heat. (INOR-8M3 ~2000 Ib) was cast into-a.mold of the second proposed type and was .subsequently press-forged at 2150°F from a 15-in, . square to a 9Q-m. round. Ultrasonic inspection indicated internal flaws in the forging, but it was o possible to obtam three 8-in.-dia extrusion billets " from the maferlal These billets have been sent - to ‘the In$ernuhonal Nickel Company to be extruded ~ to tube blanks, ~ Work for the immediate future at Westmghouse “will be concerned with (1) the preparation of an ingot of the initial high-carbon heat (M-1327) by “the consumable-electrode method for determining “ingot soundness, forgeability, and strength proper- ties relative to those of the air-melted material; (2) the preparation of plate, sheet, bar, and wire products of air-melted INOR-8; and (3) an evalua- tion of the casting characteristics of the alloy, ~ with -possible application in the shell molding of pump impellers, -As indicated .previously,” an evaluation of the high-carbon air-melted heat (M-1327) of INOR-8 is . in progress. The results of thermal-expansion measurements made on the material from room ~temperature to 1832°F, in cooperation with the Ceramics Group, are shown in Fig. 3.1.1. The ‘average coefficients of thermal expansion, a, ob- tained from the data in Fig. 3.1. l are, as a func- tion of temperature: Temperuture Range ' a (°F) - {in./in.+°F) o xwd _‘.2'].2_7”5_2. e e 0.70 Co7s2-1m2 - © o 0.84 T izossr N X ‘212-133'2 . 0.86 ln order to obtam an’ mdlcahon of the effect of 7H. Inouye, T. K. Roche, and J. H. Coobs, ANP Quar, ~Prog. Rep, June 30, 1957, ORNL-2340, p 175. 165 P ANP PROJECT PROGRESS REPORT were carried out on the air melt end the vacuum .melt under identical conditions, are given in Table 3.1.6 for comparison. The data showed that a definite improvement in rupture life and elonga- ‘tion was obtained by vacuum remelting, However, the creep rate of the air-melted alloy was superior to that of the vacuum-remelted material, Haynes Stellite Company. — A 10,000-Ib air- mglted heat® of INOR-8 has been purchased from__ 8Designafed Haynes Stellite experimental dliby No. ' 8284 (heat SP-16). UNCLASSIFIED ORNL—LR—DWG 21929 TEMPERATURE (*F) 392 - ™2 e 1472 1832 0.0i800 0.01400 0.01200 TRUE EXPANSION & 0.01000 FOR S ~ 0.989-in. LENGTH z o 2 aoosoo & » ul . - PLOTTED DATA S FOR 0.989-in. g 000600 0.00400 000200 INOR—8 COMPOSITION (wi %): Mo Cr Fe c Mn Ni 16.90 421 04 0.4 023 BAL 0 200 400 600 800 1000 1200 TEMPERATURE (°C) Fig. 31.1. Lineoar Thermal Expansion of INOR-8 as Determined with the Use of the Vacuum ‘'Atcotran’’ Differential Transformer Recorder. the Haynes Stellite Company to obtain the material needed. for corrosion, welding, and creep testing. The composition of this heat is given below: Amount Components (wt %) Nickel 70.50 Molybdenum 15.82 Chromium 6.99 ~ lren . 4.85 . Carbon 0.02 Tungsten 0.35 Silicon | 0.32 Cobalt 0.51 Manganese 0.34 Copper 0.03 ‘Phesphorus - 0.009 Sulfur o 0.014 Boron = - 0.04 Casting ond fabrication of the alloy by hot and cold working fo bar, plate, sheet, and strip products were accomplished with little or no difficulty. To date, the various shapes of this materia! “which have been received include: a 3-in.-dia forged bar; | 1-, ‘/2-, and ‘/4-in.-thick hot-rolled plates; 0.063-in.-thick hot-rolled sheet; ?(‘-in.-dia . * hot-rolled bar; six production-size extrusion billets; and rolled strip for the processing of weldrawn tubing. The extrusion billets will be sent to Allegheny Ludlum Steel Corporation for conversion to type 316 stainless steei—INOR-8 composite tubing. Superior Tube Company has received the rolled strip ond is currently processing the alloy into small-diameter weldrawn tubing. Upon com- pletion of the orders by the various fabricators involved, enough material should be received to fabricate approximately 30 forced-circulation loops, one in-pile test loop, and specimens for a rigorous program of creep and weld tests, Table 3.1.6. Creep-Rupture Characteristics of the Westinghouse High-Carbon INOR-8 Heat M-1327 Tested at 1500°F in an Argon Atmosphere at a Stress of 10,000 psi Time to Specified Strain (hr) Type of Specimen Time to Rupture - Elongation : ' 1% 5% 10% (hr) (%) As air-melted material 32 _ 120 135 6.25 Vacuum-remelted material 25 80 125 214 54.5 166 U International MNickel Company. — As reported previously,’ eight 9-in.-dio, - 12-in.-long billets, two each of INOR-1, -2, and -5 ond cne each of INOR-3 and -6, were extruded to 3)-in.-OD, 0.500-in.-wall tube shells at the International Nicke!l Company. Recently the reducing of these tube shells was completed at INCO, and the re- duced shells have been sent to the Superior Tube Company for processing to 0.500-in, -OD 0. 045-m. : wall seamless tubing. Considerable delay was experlenced in completmg the tube reducing of these shells at INCO for various reasons. Primarily, the delay was caused by unexpected difficulty in annealing the tube shells following cold working on the tube reducer. It was found that the alloys were subject to thermal splitting when a total reduction of 30% or less was given prior to annealing. This behavior was en- countered at annealing temperatures between 1700 and 2150°F. However, by increasing the total cold reduction between anneals to over 50%, the alloys were successfully annealed at 2150°F without splitting. The tube shells, which were reduced to 1.66-in.-0D, 0.148-in.-wall shells at INCO,. have been sent to Superior Tube Company, Descrip- tions of the tube blanks sent to Superior Tube Company for further processing are given in Table 3.1.7. Composite Tubing | All the nickel-base alloys exhibit mass trnnsfer when used as containers for flowing sodium at H. Inouye and T. K. Roche, ANP Quar._Prog. Rep March 31, 1957, ORNL 2274, P 134. . - Table 3 I 7. |NOR-Type Tube Sbells Produced by Internuhonal Nlckel Compuny Approxtmufe Lengfhs of H'eatf,No.‘ Alloy 1.66-in.-0D; 0.148in.- o °="9"°"°" _ Wall Tube Sheils (ft) Y-8195 INOR-1- 64 Y-8197. INOR2 64 Y-8196 INOR3 . . 32 Y8200 INOR5 64 ° Y.8199 INOR-6 32 *Products from 4800-1b air melts. PERIOD ENDING SEPTEMBER 30, 1957 high temperatures. The extent of mass transfer is -temperature dependent, being slight at 1300°F and tolerable -at 1500°F, For service temperatures of 1650°F and ‘above, it appears that a composite ‘tube, composed of a nickel-base alloy and type 316 stainless - steel, is desuroble and, possibly, necessary. : There are three basic problems assocmted wi th the use of such composite tubing in fluoride fuel- to-NaK heat exchangers. The first problem is the fabrication of the composite tubing. Experiments indicate that tubing of acceptable quality can be made by following the practice of coextruding composite billets. The high temperatures and pressures used during coextrusion result in the formation of metallurgical bonds. Dimensional control is maintained by a suitable choice of the starting composite and by conditioning of the tube shell after extrusion. The outlook is optimistic, since it is believed that composites of type 316 stainless steel and INOR-8 should not be more difficult to prepare than composites of Inconel and type 316 stainless steel, which have been made in substantial quantities. The second problem is the welding of the com- posite tubing. The mechonical properties of the welds and the resistance of the weld nugget to sodium mass transfer must be investigated. Con- siderations of the properties of alloys intermediate between INOR-8 and type 316 stainless steel may provide information on these points. The six intermediate alloys described in Table 3.1.8 were cast, fabricated, and aged to determine _the equilibrium phases and the effect of various : T_u'ble 3.1.8. Compositions of Alloys Intermediate -Between INOR-8 and Type 31§ Stainless Steel Alloy ~ Nominal Compéshion’ (wt %) Designation Mo Fe Cr Ni ~ INOR-8 7 5 7 71 Atloy-1 15 n 8 66 a2 13 22 9 56 -3 1 33 1 45 4 9 44 12 35 -5 6 56 13 25 -6 3.5 67 14 15 Type 316 25 70 15 12 stainless stee! 167 ANP PROJECT PROGRESS REFPORT aging femperatures on their tensile properties. Thus far only the microstructures of the cast-and- aged. alloys have been determined. It was found that the alloys 1 and 6 were solid solutions, and therefore they should not exhibit undesirable mechanical properties. The other four alloys showed small quantities of a precipitate. The effects aond nature of the precipitate have not yet been determined. Based upon the preliminary study, the configuration for a tube header sheet should be one in which the stainless steel is: diluted by a minimum omount of the nickel-base alloy after welding. ' ' The third problem involves an investigation of the diffusion of the elements of one alloy into the other, Since diffusion is a temperature-dependent phenomenon, its effect is expected to be greatest ot high temperatures. ' Of prime importance is the - determination of the depth to which the various elements will penetrate each other under a variety - of conditions, since this will determine the thick- nesses of the layers comprising the composite. Since the nickel-base alloy is stronger thaon the stainless steel at high temperatures, the composite tube should contain as large o percentage as pos- sible of the nickel-base alioy. The diffusion effects have been studied in a series of tensile tests of a composite composed of 50% |INOR-8 clad on both sides with type 316 stainless steel. The data obtained are shown in Table 3.1.9. JOINING EXPERIMENTAL INOR-8 ALLOYS R. E. Clausing G. M. Slgughter P. Patriarca Ten heats of experimental nickel-molybdenum alloys similar to INOR-8 have been tested to de- termine whether these alloys can be successfully fabricated into complex components by means of brazing and inert-gas-shielded arc welding and whether, once fabricated, the material will have suitable mechanical properties. Since the compo- sition of Battelie Memorial Institute heat 3766 was closest to the composition range of INOR-8, the experimental results obtained in the study of speci- mens of heat 3766 are summarized here. The analysis of heat 3766 is compared with the compo- sition range of INOR-8 in Table 3.1.10, It was readily determined that specimens from heat 3766 could be brazed in a dry-hydrogen atmos- phere with Coast Metals brozing alloy No., 52 and similar brazing alloys by -using conventional techniques. The determination of weldability, however, was inherently more complex and time- consuming. Many tests were required to determine Table 3.1.10. Comparison of Analysis of Battelle Memorial Institute Heat 3766 with the Composition Range of INOR.8 Quantity Present (wt %) Components INOR-8 Heat 3766 Molybdenum 15-19 13.3 Chromium 6-8 6.05 Iron 4--6 6.07 Carbon 0.04-0.10 0.065 Manganese 0.78 Siticon 0.43 Titanium 0.11 Magnesium 0.05 Nicket Balance Balance Table 3.1.9. Room-Temperature Tensile Properties of INOR-8 Clad on Both Sides with Type 316 Stainless Steel Heat Treatment Tensile Strength Yield Strength at 0.2% Offset Elongation (psi) (psi) (%) ~ Annealed 89,400 33,600 59 500 hr at 1300°F 94,400 36,800 51 500 hr at 1500%F 93,600 34,600 42 500 hr ot 1650°F 83,000 32,400 - 28 500 hr ot 1800°F 79,800 31,800 43 168 \'5. ‘mental Studies of Inconel, (1) that sound weld deposits could be made, that is, deposits free of cracks and pores, as well as excessive segregation, (2) that there would be no deleterious regions in the heat-affected zone, (3) that the base metal would not be subject to hot tearing, and (4) that the mechanical properties of the joint would be satisfoctory, both in the as-welded and the aged conditions. A test specimen was designed which permitted nondestructive determination of the soundness and freedom of cracking of the weld metal and the heat-affected zone. This specimen also provided material for metallographic and hardness speci- mens, as well as the bend- and tensile-test speci- mens needed for determining the mechanical properties of the welded joint. The specimens were machined from plates that had been welded under a high degree of restraint. ~ The methods used for obtaining the specimens were similar to those shown in Fig. 3.2.18 of Chap. 3.2, '‘Develop- " this report. Examinations of many specimens of weld deposits made by inert-gas-shielded tungsten-arc welding procedures on BM!| 3766 base metal with filler wire of the some material failed to reveal any significont porosity and only one small root crack. On the basis of these examinations, which were made with dye-penetrant, x-ray, and metallographic methods, it may be concluded that it should not be difficult to make sound weld deposits in BMI 3766 joints, Hardness measurements indicated that very little, if any, hardening occurs either in the weld metal or in the heat-affected zone as a result of aging for 200 hr at temperatures from 1100 to 1700°F, Aging at 1700°F for 200 hr reduced the hardness of the_ weld metal to that of the base metal, while aging at 1500°F for 200 hr results in considerable soften- ing of the weld metal, Aging at 1100, 1200, and 1300°F for 200 hr did not appreciably change the hardness. Metallographic examination of hardness specimens indicated that the eutectic material - which was present in the welds initially was 170 seemed fo have nearly identical base-metal and partially spheroidized upon aging at 1500°F and completely spherocdrzed upon agmg at 1700°F for i 200 hr, , Bend test specimens were eut from the - Welded joints and tested in the as-welded and aged condi- tions at temperatures ranging from 1100 to 1700°F in order to evaluate the mechanical properties of ‘the joints. It is possible in a bend test to test not PERIOD ENDING SEPTEMBER 30, 1957 only the joint as a whole but also the various com- ponents of the joint. For instance, the deformation can be made to occur in a selected area, such as the base-mefal—weld-metal interface, which might not otherwise deform in the tensile test. The data obtained from these tests are presented in Table 3.1.11 in terms of deflection when the load had dropped to two-thirds of its maximum value. The deflection values multiplied by a factor of 100 roughly correspond to the elongation values ob- tained in tensile testing (that is, a 0.100-in, de- flection corresponds to approximately 10% elonga- tion in the standard 0.252-in.-dia tensile specimen). - The maximum possible deflection with the particu- lar machine used is 0.3 in, Therefore, when the reported deflection is 0.3 in., the specimen is not considered to have failed. An analysis of the data in Table 3.1.11 reveals that none of the specimens tested at 1100°F failed, that the specimens tested at 1300°F in the aged condition exhibited con- siderably less ductility than those tested in the unaged condition, and that the specimens tested at 1500°F in the aged condition showed considerably more ductility than specimens tested in the unaged condition, This may be correlated with the ob- served change in microstructure, that is, the dis- appearance of eutectic material. Further, the speci- mens tested at 1700°F also exhibited more ductility in the aged condition than in the unaged condition. This is again associated with the disappearance of eutectic material, At 1300 and 1500°F the heat- offected zone showed a loss in ductility when compared with the base metal. However, the ductility was equal to or greater than that of the -Vweld metal, A number of tensile specimens cut from the weldments were tested at temperatures ranging from room temperature to 1700°F. The results of these tensile tests, as shown in Table 3.1.12, . agree very well with the results of the bend tests (Table 3.1.11). It may be noted that the speci- ~ mens which were aged for 200 hr at 1500°F and at 1700°F and were tested at the aging temperatures This is indicated by the weld-metal properties. - fact that one specimen failed in the weld at each temperature, while duplicate specimens failed in the base metal, with similar strengths and elonga- tions. It may also be noted that very few of the failures occurred at the gage marks, and thus it appears that this material is not particularly notch 169 ANP PROJECT PROGRESS REPORT Table 3.1.11. Results of Bend Tests of Specimens Cut from Welded Joints of BMI 3766 Filler Wire on Base Metcl of the Source Material Test Temperature Area Tested Final Deflection when Load Dropped to Two-Thirds of (F) Maximum Value {in.) 1100 Weld centerline 0.300+ Weld centerline 0.300+ Weld centerline* 0.300+ Weld centerline* 0.300+ 1300 Boase metal 0.290 Heat-affected zone and weld interface 0.220** Weld centerline 0.240 Weld centerline 0.220 Weld centerline* 0.120 Weld centerline* 0.145 1500 Base metal 0.300 Heat-affected zone and weld interface 0.165 Weld centerline 0.080 Weld centerline 0.070 Weld centerline* 0.200 Weld centerline* 0.190 1700 Weld centerline 0.200 Weld centerline 0.100 Weld centerline* 0.300 Weld centerline* 0.300 *Specimens aged 200 hr ot the test temperature prior to testing. **Specimen broke in base metal. sensitive. Specimens aged at 1300 and at 1400°F failed in the welds, and it seems that the weld metal is weaker than the base metal, at least in the aged condition. Unfortunately, the solution heat treatments of the base materials tested at room temperature and at 1500°F were not identical with the solution treatment given the base material used in making these welds. Therefore, a direct comparison between the properties of the base- metal tensile specimens and the base metal in the transverse-test specimens is not truly valid. In order to make such a comparison, base-metal tensile tests should be made of specimens given the same heat treatment as that given to the base metal used in the tronsverse tensile tests, The base-metal specimens should then be tested at all the temperatures listed in Table 3.1.12, It would also be desirable to test unaged tensile specimens at 1300°F in order to verify the results of bend 170 tests, which indicate that aging decreases the ductility at this temperature. |t should also be noted that the room-temperature results indicate that failures should occur in the base metal, A comparison of the tensile properties of weld- ments of Inconel, Hastelloy B, and the experimental INOR-8 alloys is given in Table 3.1.13. Paort of the data were interpolated and therefore may be subject to some deviation from the actual values, and the Inconel data for aged specimens are actually data for unaged specimens, since it is thought that any changes which would occur on aging for 200 hr at the specified temperatures would not appreciably alter the date. It may be noted that at nearly all temperatures the values for the INOR-8 specimens are between those for Inconel and those for Hastel- loy B. In general, the strength is less than that of Hastelloy B but greater than that of Inconel, while the ductility lies between that of Inconel b 9 Qs‘o PERIOD ENDING SEPTEMBER 30, 1957 ~ Table 3.1.12. Results of Tensile Teits of BMI 3766 Weldments Test Type of Tensile Strength at Reduction Temperature Tensile Test Strength ~ 0.2% Offset Elongation of Area Location of Fracture (°F) Specimen (psi) (psi) (%) (%) Room All weld metal 108,500 77,800 40 34 Weld metal; no defect All weld metal 107,900 64,600 40 32 Through small inclusion : ' in weld metal Transverse® 108,800 = . 57,200 42 44 Base metal Transverse 108,600 - - 58,700 43 37 Base metal Base metal® 107,100 41,800 38 Base metal Base metal® 105,500 39,400 50 38 Base metal 1300 Transverse® 56,400 . 37,300 12 13 Weld; no defect Transverse® 59,900 _ _27,000 13 13 Weld; no defect 1400 Transverse® . 59,000 37,600 15 14 Weld; no defect Tronsverse® 57,000 41,700 13 8 Weld; no defect 1500 All weld metal 51,500 36,800 12 7 Weld metal; no defect All weld metal 51,700 .37,900 12 5 Weld metal; no defect Transverse 51,500 | 36,800 12 6 Weld; ot gage mark Trnhéversg 51,700 37,900 12 6 Weld; no defect Transver;ec _ 46,700 31,400 23 27 Weld; ot gage mark " Transverse® 45,800 | 32,200 35 - 34 Base metal Base metal® 47,700 25,100 40 35 Base metal Base metal? 47,100 - 26,300 40 32 Base metal 1700 Transverse® 28,0000 . 27,400 30 ' 25 Weld; no defect 32 15 Base metal Transverse® 27,2_00 26,700 “Specimens were cuf to contain weld metul bose metal, and the weld-metal-base-metal interface within a uniform gage length, This material was given a longer soluhon treatment than was the joint-base material. €Specimens were aged 200 hr at the test femperutufe prior to testing. and that 6frHc»x‘é;fe'liloyi'B_.';'-:J-Tfié" lowest ;:!uc't_ility value reported for INOR-8 is above 10% at 1300°F, . In addition, numerous- fl'termal-convechon loops L by these methods will not be subject to any un- “usval restrictions as a result of the fubrlccmon Qprocesses. have been fabricated of olloys similar to INOR-B'-"'-_ without difficulty. Corrosion tests of many of ihese. e loops have been completed, and there have been no. md:cahons of any deleterious effects near or “in the welds.l Although the filler metal used in “ali these loops was Hastelloy. W rather than being the same as the base metal, it is thought that the - successful fabrication and operation of these loops 'is a good indication that these materials should “not be difficult to fabricate by welding. . In summary, - it appears that material of the INOR-8 composition will be readily fabricable by both welding and brazing techniques and that structures fabricated _ STRESS-RUPTURE PROPERTIES OF NICKEL- .. 'MOLYBDENUM BASE ALLOYS IN A FUEL ENVIRONMENT J.W Woods ' - D, A, Douglas Evuluuhon studles of rhe creep characterlsflcs _of the nlckei-molybdenum alloys under a constant load .in the fuel mixture NaF-KF-LiF- UF, (11.2- ~ 41-45,3-2.5 mole %, fuel 107) were conhnued The results of tests made during the quarter are pre- sented in Table 3.1,14. 171 ANP PROJECT PROGRESS REPORT Table 3,1.13. Comparative Tensile Properties of Inconel, Hastelloy B, and INOR-8 Weldments As Welded Aged 200 hr at Test Temperature Tempercture . Tensile Yield Strength Tensile Yield Strength Material Elongation Elongati (°F) Strength at 0.2% Offset gatio Strength ot 0.2% Offset gaiton (psi) (psi) %) (psi) (psi) (%) Room {nconel 88,000 50,100 - 40 Inconel 90,700 52,700 40 Hastelloy B 118,300 21 Hastelloy B 124,500 23 INOR-8 108,800 57,200 43 INOR-8 108,600 58,700 1300 Inconel 50,000* 35,000* 40* Hastelloy B 108,300 5.0 Hastelloy B 109,100 2.0 INOR-8 56,400 37,300 12 INOR-8 59,900 27,000 13 1500 Inconel 37,000 28,000 45 Inconel 37,000 28,000 45 37,000* 28,000* 45* Hastelloy B 66,700 9 67,200 : 19.5 Hastelloy B 62,300 12 63,800 10 INOR-8 51,500 36,800 12 46,700 31,400 23 INOR-8 51,700 37,900 12 45,800 32,200 35 1700 Inconel 22,000* 20,000* 50* Hastelloy B 40,000** 30** INOR-8 28,000 27,400 30 27,200 26,700 32 *Data for as-welded specimens. **These are conservative figures based on data obtained ot 1650°F. Tests of specimens from Battelle Memorial Institute alloy heats B-3874 through B-3879 were completed. As previously reported,'® no data could be obtained on alloys B-3874, B-3875, and B-3876 because of their poor welding characteris- tics. Alloys B-3877, B-3878, and B-3879 ex- hibited good stress-rupture characteristics but the ductility was poor. A metallographic examination of the specimen from heat B-3879, which was tested in the annealed condition, revealed a small grain size and a heavy concentration of carbides in the grain boundaries (Fig. 3.1.2). In the past, the alloys that have exhibited the best creep characteristics have also shown this concentration 105, W. Woods and D. A. Douglas, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 179. 172 of carbides in the grain boundaries. Little, if any, corrosive attack is evident. Even though this alloy looks promising from the stress-rupture standpoint, dynamic corrosion studies have indicated con- siderable corrosive attack by fuel 107 of all me- terials containing substantial quantities of alumi- num, In an effort to determine the effect of annealing temperature on the creep properties, a series of specimens from ORNL vacuum-melted heat 30-63 (13% Mo-5% Fe-0.06% C-5% Nb-—bal Ni) were tested at 8000 psi and 1500°F in fuel 107. The data obtained are presented in Table 3.1.15, There is little difference in the creep properties of ‘the material annealed at 2100°F and that annealed at 2200°F. However, when these annealed speci- mens were aged at 1300°F prior to testing, the * & PERIOD ENDING SEPTEMBER 30, 1957 Table 3.1.14. Stress-Rupture Properties of Nickel-Base Molybdenum Alloys Tested at 1500°F and o Stress of 8000 psi Exposed to NcF-KF-LiF-UF‘ (11.2-41-45.3-2.5 Mole %, Fuel 107) Heat No. Composition (wt %), Bfilance Nickel Rl;_p:;:e Elongation Heat Mo Cr Fe C Al" Mn Ti Other (he) (%) Treatment* Battelle Memorial Institute Alloys, Air Melted B-3877 174 6.9 09 007 21 09 006 1.3Nb 904 16 4 B-3878 19.9 4.5 0.06 1.5 0.9 0.9 Nb 751 5 3 B-3878 19.9 4.5 0.06 .5 0.9 0.9 Nb 801 | 6 4 B-3879 19.8. 6.9 43 0.05 1.5 09 008 1.0Nb 1641 14 3 B-3879 19.8 69 43 0.05 1.5 0.9 008 1.0Nb 2034 11 4 ORNL Alloys, Vacuum Melted 30-62 16 7 5 0.06 0.5 0S5 0.15 Zr 492 53 1 30-62 16 7 5 0.06 0.5 0.5 0.15 Zr 1189** 66 2 VT-66 17 7 4 0.14 0.8 195%* 51 1 VT-66 17 7 4 0.14 0.8 330** 51 3 7 Haynes Stellite Company Alloy, Air Melted SP-16 158 7 4.8 0.02 0.3 530 13 ) SP-16 0.02 0.3 536 12 2 15.8 7 4.8 *Heat treatment No. 1: Heat treatment No, 2: Heat treotment No. 3: Heat treatment No. 4: **Test discontinued before rupture occurred. solution annealed at 2000°F rupture life and ductility were decreased, The specimen annealed at 2200°F before aging was less affected than the one annealed at the Iower temperafure. ' In order to obtain an indication of the effecf of vacuum remelting on the creep properties of the “high-carbon-content Westinghouse Electric Corpo- . | a vacvum-melted “3-1b casting, VT-66 was' prepared from the air-melted - . material. casting were tested in fuel 107, The ductility of_ , ration - heat of INOR-8, . - Stress-rupture specimens - from -this ‘the remelted material was supeflor to that of the original melt, and there was a smuH increase in the -rupture life; however, the creep rate was not improved over that ‘obtained for the alr-melted'. The VT-66 specimen, which - Wesh_ngh_ous._e heat. was tested in the annealed condition, is shown in Fig. 3.1.3. In view of the high-carbon content solution annealed at 2000°F for 1 hr. for 'I hr and aged at 1300°F for 50 hr. solution annealed at 2100°F for 1 hr. solution annealed ot 2100°F for 1 hr and aged at 1300°F for 50 hr. (0.14%), surprisingly few carbide precipitates are present. Those present are dispersed randomly throughout the sample, with no concentrations in the grain boundaries. The lack of carbides prob- ably accounts for the poor stress-rupture properties. ~There was little conosive attack. Two tests were conducted on sheet material obtumed “from a 10,000-1b qir-melted heat of the INOR-8 -composition, heat SP-16, alloy 8284, pro- - duced by the Haynes Stellite Company. The alloy exhibited a shorter time-to-rupture than was ex- pecfed but fl'ns could be attributed to the low carbon content (0.02%). The ‘specimen tested in ~ the annealed condition is shown in Fig. 3.1.4. As was expected, there are few carbide precipitates, and those present in the grain boundaries do not form a continuous network, The properties of the Haynes Stellite heat are compared with those of 173 ANP PROJECT PROGRESS REPORT RN e m*‘ UNCLASSIFIED & s"l-i_p ¢ TRE - ._.ns — e A} s > ox o g o 2 2 INCHES Fig. 3.1.2. Stressed (8000 psi) and Unstressed Sections of a Speélmen of Alloy B-3879 Tested in the Annealed Condition at 1500°F in Fuel 107. Etchant: chrome regia. 500X. Seerativith-captton) Table 3.1.15. Results of Study of Effect of Prior Heat Treatment on Creep Properties of Specimens from ORNL Vacuum-Melted Heat 30-63 Tested at 8000 psi and 1500°F in Fuel 107 Rupture Life Elongation Heat Treatment (hr) (%) Annealed at 2100°F for 1 hr 750 50 Annealed at 2200 for 1 hr 740 25 Annealed at 2100°F for 1 hr and aged at 1300°F for 50 hr 350 15 Annealed ot 2200°F for 1 hr and aged at 1300°F for 50 hr 670 13 other heats of INOR-8 in Table 3.1.16, and the compositions of the heats are given in Table 3.1.17, It may be seen that the Haynes alloy, 8284, has the best creep properties of all the alloys listed. It is felt that with an increase in carbon content, this alloy will have a rupture life of 1000 hr or more at 1500°F and 8000 psi. 174 Specimens from the vacuum-melted heat 30-62 were tested to determine the effect on the stress- rupture behavior of adding zirconium to the INOR-8 clloy. Previously, specimens of 30-lb heats of INOR-8 had failed in 680 hr with 30% elongatien, but testing of the zirconium-bearing specimen was discontinved ot 1189 hr and the elongation was PERIOD ENDING SEPTEMBER 30, 1957 4 * -« -«% o ‘..: - ’V : gfl { AL " & o ! o o ‘ o ' . 33 S 500X 3 S g 3 § INCHES Fig. 3.1.3. Stressed (8000 psi) and Unstressed Sections of Specimen from Vacuum-Melted Heat VT-66 Tested in " the Annealed Condition at 1500°F in Fuel 107, Etchont: chrome regia. 500X. (Seeret-with-cuption) » . - lo ) l . e .- i ooolet 100X 0 o - o e o INCHES - '; ‘ . a . Sy . - w n : ) J : u Fig. 3.1.4 Stressed (8000 psi) and Unstressed Sections of Specimen of Hayes Alloy 8284 Tested in the : Annealed Condition ot 1500°F in Fuel 107. Etchant: chrome regia. 100X, {Seeret-with captien) 175 ANP PROJECT PROGRESS REPORT Table 3.1.16. Comparison of Creep Properties of Several Heats of INOR-8 Tested at 1500°F and a Stress of 8000 psi in Fuel 107 Time to Specified Strain (hr) ‘Rupture Total Alloy Designation Source Life Elongation ) : 1% 2% . 5% 10% (hf) (%) 8284 (heat SP-16) Haynes Stell.ite 36 100 280 510 530 13 Company M-1327 " Westinghouse Electric’ 14 32 64 92 110 16 : Corporation 30-62 ORNL 29 72 180 340 1200 66 VT-51 ORNL 28 62 200 450 1243 49 75 180 340 670 26 30-34 ~ ORNL 36 Table 3,1.17. Compositions of Yarious INOR-8 Heats Nominal Composition (wt %), Balance Nickel Alloy Designation Mo Cr Fe C Other 8284 (heat SP-16) - 16 5 0.02 0.35 W M-1327 | 17 4 0.14 30-62 16 7 5 0.06 0.15 Zr VT-51 17 10 7 0.06 30-34 15 6 5 0.06 66%. The creep rates of both specimens were similar, and thus the only difference was the in- creased rupture life of the zirconium-bearing alloy, which may be ctiributed to an increase in ductility, The zirconium-bearing specimen, which was tested in the annealed and aged condition is shown in Fig. 3.1.5. Bands of various grain sizes ranging from coarse at the surface to fine at the center of the specimen may be seen. The usual heavy concentration of carbide particles in the grain boundaries that is associated with all the high-strength nickel-molybdenum alloy systems is also present. The specimen shows corrosion attack to a depth of approximately 0.5 mil. Although the INOR-8 alloy does not represent the ultimate in mechanical properties which can be cbtained from the nickel-molybdenum alloy systems, it is superior to type 316 stainless steel and 176 Inconel at 1500°F in a fuel environment, as is shown by the representative creep data in Table 3..18, Tensile tests of alloy 30-62 in air at 1500°F indicated a tensile strength of 46,000 psi with 20% elongation, The maximum corrosive attack observed corresponds to the corrosion found for Hastelloy B in the same fuel environment. CORROSION STUDIES J. H. DeVan Forced-Circulation Loop Test of a Nickel- Molybdenum Base Alloy Exposed to Fuel 107 J. H. DeVan R. S. Crouse The first forced-circulation loop fabricated of an experimental nickel-molybdenum base ciloy com- pleted 1000 hr of operation with fuel 107 at a maximum fuel-to-metal interface temperature of a4 1760°F, The loop alloy had the nominal ¢omposi- tion 17% Mo—-6% Fe-bal Ni. Other operating conditions for the loop, 7641-9, are given below: Maximum bulk fuel temperature 1610°F Fuel temperature drop 300°F Reynolds No. ' 10,000 ., Flow rate 1.4 gpm Ratio of surface area of heated 2.2 il1.2/in.3 section to total loop volume PERIOD ENDING SEPTEMBER 30, 1957 An examination of a hot-leg section of this loop showed attack by the fuel to be confined almost entirely to grain boundaries. The attack reached a maximum depth of 4 mils at the point of maximum wall temperature. ‘As may be seen in Fig. 3.1.6, the attack was, in general, in the form of small, discontinuous voids, although shallow surface pits are also evident along the exposed surface, Con- siderable oxidation occurred on the outer surface of the hot leg, which was in contact with air during operation of the loop. In addition to forming a Flg. 3.'! 5. Speclmen of AI on 30-62 Tufed In the Annnaled and Aged Condlflon at 8000 psi and 1500°F in Fuel '|07. _ Etchanf' ' chrome regiu. soox. (&cmHh-eep'hon) chlo 3. 'l 18. Repruentuflve Creep T est Data Obfclned at 'ISDO°F LT : B g B . , . ' Rrar.lptrure‘ Total L : sl T : _ Time te Specified Strain (hr) ) “Alloy Stress - Test ' : . - Life Elongation : (psi) Envi_ronrnen'f_r 0.5% _1%, 2% 5% 10% (hr) ' (%) " CAlley'30-62 < 8000 Fuel107 7 29 72 180 340 1200 66 Type 316 . - 6300 Fuel 30 4 20 95 300 - 540 o 977 55 stainless steel - : ) : : Inconel 3500 Fuel 30 22 70 180 600 1000 1000 10 u ‘ ' 8000 Fuel 30 3.5 15 30 47 177 ANP PROJECT PROGRESS REPORT heavy, uniform oxide film, the oxidation had pro- ceeded preferentially along grain boundaries to depths of as much as 9 mils. Examinations of cold-leg sections showed numer- ous surface pits and some areas of intergranular attack to a depth of 2 mils. A very thin deposit of tiny metal crystals was present over most of the cold-leg surface, and, in addition, widely dispersed clumps of metal deposits up to 6 mils in thickness were found. The heaviest deposits found are shown in Fig. 3.1.7. The deposits interfered with drain- _ age of the fuel from the walls of the loop along these portions of the cold leg, and therefore the deposits were partially covered with a heavy fluoride-mixture film, Samples of this film which contained sizeable quantities of metal crystals were analyzed and were found to contain the fol- lowing elements: | Fuel Quantity Found Constituents (wt %) K 5-10 Na 2-5 Li 2-5 ) 10 Quantity Found Contaminants (wt %) Al 0.05 Fe 0.3 Cr 0.1 Mo _ 0.1 Ni | 5 I+ was possible to make a separation of the metal crystals from the residual fuel by using an ammonium oxalate solution, and analyses of the metal particles are in progress. The separated metal particles were quite magnetic, and, on the basis of the above cnalysis, it would appear that they contain large percentages of nickel, with some iron, \ Before-test and after-test samples of fuel from this loop showed no change in nickel and iron contamination, but a slight increase in chromium contamination was noted in the after-test sample. It is assumed that Hastelloy W, which was used as a filler wire in making the loop welds, served as the source of chromium, 178 Fig. 3.1.6. Hot-Leg Attack in Nickel-Molybdenum Base Alloy Forced-Circulation Loop 7641-9 Operated with Fuel 107. Etchant: 250X, (Gecres with-voptior BRI UNCLASSIFIED. R T-13256 aqua regia. Fig. 3.1.7Z. Cold-Leg Sections of Nickel-Molybdenum Base Alloy Forced-Circulation Loop 7641-9 Operated with Fuel 107. Unetched. 250X. G&m‘flfihfl% Thermal-Convection Loop Tests of Nickel- - Molybdenum Base Alloys Exposed to Fuel 107 D. A. Stoneburnerl' J. R. DiStefano Two '_thermal_-convecztion loops constructed of experimental nickel-molybdenum base alloys were tested with the fuel 107 at 1500°F for 1000 hr, The compositions of the alloys from which these loops (Nos. 1136 ond 1155) were fabricated are presented in Table 3.1.19, dlong with the results of chemical analyses of the fuel circulated in these and three previously operated loops. " Metallographic examination of loop 1136 revealed heavy surface roughening and pits along the hot- leg surface to a maximum depth of 5 mils. In addition, very large and irregular-shaped voids were present at depths as great as 9 mils below the surface. As may be seen in Fig. 3.1.8, these voids were elongated in. the direction of the tubing axis and did not appear to be connected with the attacked areas above, They probably reflect fab- rication defects rather than actual areas of attack, The cold leg of loop 1136 showed light surface roughening with no evidence of deposits. . The analysis of the fuel after the test, as shown in Table 3.1.19, indicated considerable reaction of the fuel with aluminum and a slight reaction with titanium, : : lLoop 1155 showed heavy surface roughening and surface pits, with heavy intergranular subsurface void formation to a depth of 3 mils in the hot-leg 1D, H. DeVan and D. A. Stoneburner, ANP Quar., Prog. Rep. June 10, 1957, ORNL-2340, p 185, PERIOD ENDING SEPTEMBER 30, 1957 section. The cold leg showed moderate surface roughness, with second-phase material concen- trated near the surface. No cold-leg deposits or layers were visible, Fabrication defects in the form of voids and cracks were found in the samples examined. Such defects occurred generally to a depth of 7 mils, although one crack extended from the outer surface into the sample to a depth of 20 mils. Fig. 3.1.8. Hot-Leg Attack in Nickel Molybdenum Base Alloy Thermal-Convection Loop 1136 Which Operoted with Fuel 107 for 1000 hr at 1500°F. Etchant: copper regia. 100X, (&eeret-with copiien) Table 3.1.19. Anglyses of Corrosion Products in Fuel 107 After Circulation in 'Nickel-Molybdenum Base Alloys at 1500°F : o ' ' ' 70pe;rufing Loop - Alloy Composition Time Quantity of Alloy Constituents Ne.. O (mR) (hr) in After-Test Fuel Samples 1125 17 Mo—d V=bal Ni 1000 210 ppm 1123 17 Mo—4 V—bal Ni 500 290 ppm "1135 16 Mo—6 Cr—1 Nb—1 Al-bal Ni 1000 455 ppm Cr, 20 ppm Nb, 1455 ppm Al 1136 16 Mo—2 Al=1.5 Ti=ba! Ni 1000 1400 ppm Al, 370 ppm Ti | 1155 INOR-8 plus 0.5 Al-5 Fe~6 Cr-0.5 Mn--0.06 C 1000 760 ppm Cr, 415 ppm Al 179 ANP PROJECT PROGRESS REPORT A summary of thermal-convection loop test re- sults has been prepared to provide a comparison of the corrosion properties of the various nickel- molybdenum base alloys tested thus far in fuel 107. In Tables 3.1.20 ond 3.1.21, the alloys are grouped according to the level of attack found after 500- and 1000-hr tests at 1500°F. As may be seen in Table 3.1.20, the maximum attack did not ex- ceed 3 mils for any of the alloys tested, There- fore it appears that none of the additions seriously affect corrosion resistance from the standpoint of total depth of attack. . In particular, additions of chromium in amounts up to 9% and niobium in amounts up to 5% had little effect on the depth of Table 3.1.20. Comparison of Attack of Nickel-Molybden&m Base Alloy Thermal-Convection Loops Operoted for 500 hr with Fuel 107 Alloys with Very Limited Attack of <1 mil, Nomina! Composition Alloys with Limited Attack of 1 to 2 mils, Nominal Composition (wt %) {wt %) Alloys with Noticeable Attack of 2 to 3 mils, Nominal Composition (wt %) 17 Mo=2 Ti-bal Ni 17 Mo~2 V-bal Ni 17 Mo—3 Nb-—bal Ni 17 Mo--5 Nb—bal Ni 17 Mo=5 Cr—bal Ni 17 Mo—4 Fe-bal Ni 17 Mo—7 Cr-bal Ni 15 Mo-0.5 Al-3 Nb=3 W—bal Ni 20 Mo—7 Cr-bal Ni 16 Mo=9 Cr—bal Ni 17 Mo-2 W=bal Ni 17 Mo—4 W-bal Ni 17 Mo=3 Cr—bal Ni 20 Mo-3 Cr-bal Ni 17 Mo-0.5 Al-bal Ni 17 Mo-2 Al-boal Ni 16 Mo—~2 Al-1.5 Ti-bal Ni 16 Mo—6 Cr—1 Nb—1 Al-bal Ni - 16 Mo—6 Cr--1 Nb—1 Al-bal Ni 15 Mo-5 Cr—3 Nb-3 W-0.5 Al-bal Ni 17 Mo—4 V~bal Ni 17 Mo~1 Al-1.5 Ti-bal Ni 18 Mo—1 Al-1.5 Ti-bal Ni 20 Mo-7 Cr—2 Nb—-1 Fe~bal Ni 20 Mo—-7 Cr~1 Al-2 Nb—1 Fe—bal Ni 20 Mo—1 Nb—-2 Ti—0.8 Mn—bal Ni Table 3.1.21. Comparison of Attack of Nickel-Molybdenum Base Alloy Thermal-Convection Loops Operated for 1000 hr with Fuel 107 Alloys with Limited Attack of 1 to 2 mils, Nominal Composition Alloys with Noticeable Attack of 2 to 3 mils, Nominal Composition Alloys with Heaviest Attack of 3 to 5 mils, Nominal Composition (wt %) (wt %) (wt %) 15 Mo-85 Ni 17 Mo—-5 Cr—=bal Ni 17 Mo~-2 Ti-bal Ni 17 Mo—4 W-bal Ni 17 Mo—3 Nb—bal Ni 11 Mo—-2 Al-bal Ni 20 Mo—~1 Nb—2 Ti—8 Mn—bal Ni 20 Mo—~7 Cr—2 Nb-1 Al-1 Fe-=bal Ni 17 Mo—-2 Al=bal Ni 17 Mo—4 Y—ba! Ni 17 Mo-5 Nb-bal Ni 16 Mo—6 Cr—1 Nb-1 Al-bal Ni 16 Mo—2 Al-1.5 Ti-bal Ni 15 Mo—6 Cr-0.5 Al-5 Fe—-0.5 Mn—bal Ni 180 W corrosion; however, the presence of these elements as corrosion products in the fuel increased notice- ably as the percentage of the element in the base metal increased. It is interesting fo note that in no case did alloys containing 20% Mo appear in the column of noticeable attack (2 to 3 mils), whereas alloys with similar additives but with a decreased percentage of molybdenum did show notaceable. attack. Increasing the time of the test from 500to 1000 hr caused the attack to increase 1 to 2 mils, as shown in Table 3.1.2]. The maximum attack recorded was 4 mils, and agcin none of the alloy additions seriously affected the corrosion resistance, Compatibility of Nickel-Molybdenum Base Alloys with Molybdenum It is expected that the use of nickel-molybdenum alloys as materials of construction in molten fluoride salt reactors will make possible a considerable in- crease in the operating temperature attainable in - The operating temperature of the such systems, moderator material will, of course, increase cor- respondingly. At the temperatures being con- sidered, moderating materials such as beryliium oxide and the hydrides of zirconium ‘and yttrium must be protectively clad with molybdenum. It is necessary therefore to determine whether the molyb- denum cladding will be compatible with nickel- molybdenum alloys in a common flucride fuel circuit, If it is found that-excessive dissimilar metal mass transfer occurs between molybdenum and these alloys, an outer cladding of a nickel- molybdenum alloy over the molybdenUm w:ll become an added requirement, .. . Preliminary compatibility studles of molybdenum : with nickel-molybdenum alioys are being made by - ~using Hostelloy W thermal-convection loops. with - molybdenum inserts contained in the hot leg, One ~ such loop has been operated with fuel 107 at o maximum temperature of 1500°F for. 1000 hr,: anda second loop has operated at. 1650°F for 517 hr. The latter loop was terrmnated prior to fhe scheduled > “ately roughened, PERIOD ENDING SEPTEMBER 30, 1957 1000 hr when a leak developed in the hot leg. Metallographic examinations revealed no evidence of attack or metallurgical change of any kind along the inner surface of molybdenum samples taken from the top, middle, and bottom of the inserts from both loops. Numerous fabrication defects were found along. the inner surface of the insert from the loop operated at 1500°F, and the wall thickness of the insert varied considerably. The outer surfaces of the molybdenum inserts revealed no attack, but the insert from the loop operated at 1500°F had an extremely thin metallic layer in some areas. The samples from the bottom of the insert showed less of the deposited layer than did the middle and top insert samples, The Hastelloy W surfoces that were exposed to fuel showed heavy surface pits and shallow void formation to a depth of 3 mils after the test at 1650°F, and 2 mils after the test ot 1500°F, The Hastelloy W tested at 1500°F was attacked in some areas and yet other areas were completely free of attack. Also, the metallic layer on the molybdenum insert was found opposite the area of least attack of the Hastelloy W. The Hastelloy W tested ot 1650°F had a very thin, discontinuous line of particles immediately above some areas of attacked surface. This line appeared to be associated with foreign phases originally present in the tubing which withstood attack by the fuel. The Hastelloy W cold-leg surfaces were moder- and in the loop operated at 1500°F there were moderate surface pits and some general subsurface voids to o depth of 1.5 mils. - No mass transfer deposits or layers were found. The COrquion'rate' and appearance of the after- test samples showed no evidence of serious inter- ~ action between molybdenum and Hastelloy W under ‘these ' test conditions in either loop, ~samples of the molybdenum insert and the Hastel- - loy W sleeve have been prepared for spectrographic “examination to further check for alloying between 'r-'these metals. ' However, 181 ANP PROJECT PROGRESS REPORT 3.2. DEVELOPMENT STUDIES OF INCONEL STRAIN-CYCLING AND STRESS-RE LAXATION ' STUDIES C. R. Kennedy' D. A. Douglas The studies in o program for investigation of the strain-cycle properties of fine- and coarse-grained Inconel were continued, with a series of tests being made to determine the effect of strain re- versals upon the creep properties of Inconel. Also, other studies were initiated to defermine the strain- cycle properties of Inconel weldments, of cor- burized Inconel, and of reactor-grade beryllium, The tests to be performed, with the exception of the tests of carburized Inconel, will provide engineering data to be utilized in determining the life expectancy of structural members subjected to strain reversals, The carburized-Inconel test pro- gram is being run in conjunction with the alloy development program (Chap. 3.1, ‘‘Nickel-Moiyb- denum Alloy Development Studies’’) to determine whether corburization is deleterious to the strain- cycle properties of Inconel. Results of strain-cycle tests of fine- and coarse- grained Inconel tubular specimens at 1600°F in NaF-ZrF,-UF, (50-46-4 mole %, fuel 30) are pre- sented in Fig. 3.2.1, in which the plastic strain- lOri assignment from Pratt & Whitney Aircraft. UNCLASSIFIED ORNL~-LR-DWG 25930 S w [¢.] AT TESTED AT 0% o S S m n COARSE - GRAINED INCONEL- PLASTIC STRAIN PER CYCLE, o o o . 0.2 of : 0102 05 1 2 5 t0 2 5 1022 5 100°2 5 0 NUMBER OF GYCLES TO FAILURE, ¥ Fig. 321 Stru‘ln-Cycle Properties of Fine- and Coarse-Grained Inconel Tubular Specimens Tested in NuF-ZrF l.ll‘-'4 (50-46-4 mole %, fuel 30) ot 1500 and 1600°F. (Se-sr.gt_w.uth-snflnn-). 182 per-cycle (€,) is plotted vs the number of cycles to failure (I\f Data from previous similar tests? at 1500°F are also shown for comparison. As may be seen in Fig. 3.2.1, the data indicate that the strain-cycle properties ot 1600°F in fuel 30 are very similar to those at 1500°F and again show that coarse-grained Inconel is greatly affected by this environment, The evaluation study of the effect of prior strain cycling upon the creep properties of Inconel is under way. The test procedure includes strain cycling the specimens at 0.83% plastic strain per cycle for different numbers of cycles at 1500°F and then creep testing the specimen at 7000 psi. The test results obtained thus far are shown in Fig. 3.2.2 as creep curves of specimen with 0, 100, and 300 strain cycles prior to creep testing. ' The curves indicate that the creep strength and rupture life are immediately reduced by the first 100 cycles. The effect of 300 cycles does not further reduce the creep strength; however, the total elongation and rupture life are less than after the 100-cycle test. The type of plot that will be produced when the program is completed is shown in Fig. 3.2.3, where the number of cycles prior to creep testing is plotted vs the percentage loss of rupture life 2c. R, Kennedy and D. A. Douglas, ANP Quar, Prog. Rep. June 30, 1957, ORNL-2340, p 190. UNCLASSIFIED - ORNL-LR-DWG 25934 100 NOQ PRIOR 50 FOR. PRIOR CYCLES 20 RUPTURE FOR 300 PRIOR CYCLES 10 3 500 PRIOR CYCLES ‘é 5 o = n 2 4 05 a2 04 04 1 10 00 . 1000 TIME (hr) - Fig. .22, Creep Curves of Inconel Tubes Tested at 1500°F in Acrgon After Prior Strain Cycling at 1500°F. o O and percentage loss of creep strength, The creep strength value used is the time to 1.0% strain, This is the first attempt to study the interrelation of strain cycling and simple creep, and these early results emphasize the importance of the inter- dependence. Apparently the effects can be con- sidered to _be additive in that the consumption of UNCLASSIFIED ORNL-LR-DWG 25932 \ 8 T 20 \ 5 = i 40 \RUPTURE LIFE v * . g . \ TIME TO 4% ELONGATION W 60 ‘\ < \ S ~— @ 80 e — STRAIN-CYCLE FAILURE 3 FOR 0.85% «, 100 0 200 400 600 800 1000 4200 440Q 1600 800 NUMBER OF STRAIN CYCLES PRIORTO CREEP TESTING Fig. . 3.2.3, The Effect of Prior Strain Cycling ot 1500°F on the Creep Properties of Inconel ot 1500°F. Specimens creep tested at 7000 psi in argon. PERIOD ENDING SEPTEMBER 30, 1957 life expectancy by one of these factors also re- sults in reduced life under the other condition. This implies that design calculations based on only one of these conditions will be quite optimistic for cases where both strain cycling and creep are present, A test program has been initiated for evaluating the strain-cycle properties of Inconel weldments in both argon and in fuel 30. Two tubular specimens of conventional size were machined from all-weld- metal bars, The as-received weld metal is shown in Fig. 3.2.4, The results of two tests at 1500°F are listed in Table 3.2.1. The data from the test Table 3.2.1. Results of Strain-Cycling Tests of All-Weld-Metal Inconel Specimens at 1500°F Plastic Strai asfic Strain Number of Cycles Environment per Cycle to Failure (%) Argon 1.0 870 FUBI 30 0‘94 45 : / | UNCLASSIFIED o T Yes3s o o o —w-— T 0 =z 0.02 |0.03 e -o- o Fig. 3.24. As-Received All-Weld-Metal inconel Specimen. Etchont: 10% oxalic acid. 100X, 183 ANP PROJECT PROGRESS REPORT run in argon, as shown in the table, compare favorably with the standard tubular test data, The all-weld-metal specimen tested in fuel 30 was given a 100-hr soak in this environment prior to testing to allow corrosion to occur. As shown in the table, the fuel environment seriocusly reduced the number of cycles to failure. A metallographic specimen, shown in Fig., 3.2.5, does not indicate, however, that corrosion caused the premature failure. In fact, it appears that the major inter- granular cracks initiated at the inner surface where no corrosion had occurred. |t theréfore appears that the loss of life expectancy may have been caused by defects in the weld or unfavorable orienfation of the dendritic grains. Future tests are proposed to check these findings. During the operation of a reactor with e beryllium reflector, the beryllium will be subjected to thermal gradients which wiil resolve into strain, There- fore, a test program was begun to investigate the strain-cycle properties of hot-pressed, reactor- grade beryllium. Strain-cycle tests at 1250°F were completed and the results are shown in Fig. 3.2.6 and compared with results obtained for Inconel at 1500°F, The data shown illustrate the magnitude of the difference between the strain- cycle characteristics of a brittle metal and those of a ductile alloy. UNCLASSIFIED ORNL-LR-DWG 25033 - o ~n Q’"o' »w g N o™ o o . N PLASTIC STRAN PER CYCLE, o n o4 0102 05 { 2 5 1©© 2 8 2 s 2 BixioY NUMBER OF CYCLES TO FAILURE, ¥ Fig. 3.2.6. StraineCycle Properties of ReactorGrade Beryllium Tested at 1250°F in Argon Compared with Those of Incene! at 1500°F, "UNCLASSIFIED | v} = tu)— I [ 3] z .02 8l lo.o3 > "o- S Fig. 3.25. All-Weld-Metal Inconel Specimen After Strain-Cycle Testing ot 1500°F in Fuel 30. Etchant: 10% oxalic acid. 100X, (Secretwith captian). 184 |" In conjunction with the alloy development pro- gram, a series of strain-cycling tests of carburized Inconel at 1500°F have also been run, The results of these tests are shown in Fig. 3.2.7, where the data are compared with the data for coarse- and fine-grained Inconel, These results show that for the level of carburization of the specimens fested there was no impairment of the strain-cycle properties., The program under wcly -at the University of - previous tests. Alabama for exploring the effect pof thermally- induced strain cycles is pregressing, A series of tests in which the specimen is thermally strain cycled about a mean temperature of 1300°F has been completed, and the test results are shown in Fig. 3.2.8 and compared with the mechenicaliy- UNCLASSIFIED 2 ORNL-LR-DWG 25934 ® CARBURIZED PLASTIC STRAIN PER CYCLE, &, (%) C102 05 4 2 5 10 2 S 1 2 5 10 2 saod NUMBER OF CYCLES TO FAILURE, & Fig. 3.27. Sirdifi-Cycle Properties of Carburized Inconel Tubes Tested at 1500°F in Argon Compured with Fine- and Course-Grulned Inconel, S UNCLASSIFIED - ORNL-LR-DWG 25935 * @ MECHANICAL STRAIN CYCLE TEST - PLASTIC STRAIN PER CYCLE, «5 (%) 2 5 2 540% 2 510° 2 5 w02 :mo“ " NUMBER OF CYCLES TO FAILURE N - Fig. 3.2.8. Thermal Sircin—Cycle Properties of lnconel Tested at a Mean “Temperature of 1300°F. - PERIOD ENDING SEPTEMBER 30, 1957 induced strain cycle data obtained at 1300°F under isothermal conditions at ORNL. As may be seen, .the correlation of the results is excellent. A similar series of tests at a mean temperature of 1500°F is in progress. . Relaxation tests of fine-grained Inconel at 1100°F have been completed, and the results are shown in Fig. 3.2.2. The test procedure was described pre- Yiéygly,a The results obtained this quarter agree well with the values expected on the basis of the BIAXIAL CREEP STUDIES C. R, Kennedy D. A. Douglas The performance of a particular type of test does not always reliably indicate the manner in which material will react under operating conditions, The uniaxial creep test yields quantities of useful information; however, the applicability of the data to multiaxial stress conditions is questionable, Therefore, in an attempt to achieve a better under- standing of the behavior of metals under multiaxial stress conditions, bioxial creep studies were initiated. This program has as its aim the corre- lation of the strain rate and the rupture life data obtained in tests under uniaxial stress conditions with data obtained under biaxial stress conditions, A convenient way to vary the principal stress ratios is to superimpose an axial load on a tubular specimen, Thus, the circumferential and radial stresses will be dependent upon the internal pres- sure, and the axial stress will depend upon both the pressure and the axial load. The machines for strain cycling tubular specimens, which were described previously,? are used for these tests. This equipment provides a simple means for ob- taining strain measurements in the axial direction ~ with-the use of a. dial ‘gage fastened to the pull rod. Rupture is determined by a drop of internal pressure, which indicates the first complete crack - through the wall of the tube. .The strain in the tangential direction cannot be measured during the test and can only be roughly obtained after failure. .. The . tangential ~-stresses - applied- in. ‘various “ biaxial creep tests were calculated by considering that the stresses are elastic and that their distri- bution is represented by fhederne’ equations. * The 3¢. R Kennedy, ANP Quar. Prog Rep. March 31, 1957. ORNL-2274 p 216. J. R. Weir, Jr., ANP Quar. Prog. Rep. Dec. 31 1956, ORNL~2221, p 246. 185 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 25936 - P A NO PRIOR STRAIN 22,000 = 8,000 psi AND 0.4% STRAIN B 0.04% PRIOR STRAIN _ T C 008% OR GREATER PRIOR STRAIN 20,000 (] D 045% OR GREATER PRIOR STRAIN \\\ ~ 48,000 ~ 16,000 P N 3 " @ 14000 | 13:800 psi AND 0.05% STRAIN N W o< 13,700 psi AND 0.05% STRAIN \ - "-n—.__ ! : l 12,000 < 13,600 psi AND 0.05% STRAIN -._,___"' Sw > T~ T~ "~ r— \K . c 10,000 ~ M., [~~~ TR \\ NN \\-.“ , \ ) \\\ 7 8,000 ™t ™ =~ 6,000 X s B \\.. 4,000 002 004 006 008 Of 2 5 100 20 50 100 200 500 1000 TIME (hr) Fig. 3.29. Reloxation Characteristics of Inconel Tested at 1100°F and Stressed to Produce a Constant Strain. distribution of the stresses is indicated in Fig. 3.2.10, where 0, 0,, and g, are the radial, axial, and tangential stresses, respectively. It should be pointed out that o, + 0, = constant for all values of r , (at)a - (al)b =p, | (**thin-wall”’ formula) , (0)ey = £ 57— o constant for all values of r . z The latter condition results from the axial strain, €,, being constant for all values of r and o, + o, being constant. The actual pressure applied internally was determined by the ‘‘thin-wall’ formula. It must be realized, however, that, when the material ceases to behave elastically, the stress distribution will change. This distribution must 186 ~ Fig. 3.2.11 {ref 5). be calculated before o correlation of the uniaxial stress properties with the multiaxial stress properties can be made, The two most important factors governing the plastic-stress distribution are {a) the stress-strain-time properties of the material, such as creep and relaxation, and (%) the portion of the stress system which directly con- tributes to plasticity, For example, in a closed cylinder made of a perfectly plastic material (under a noncreep condition) in which yielding follows the Henky-von Mises criterion, M (o, - av_)2 + (o, - oz)2 + (o, - crt)2 =. 2002 , where g, is the yield stress in a simple fension test, the stresses are distributed as shown in It may be seen in Fig. 3.2.11 that maximum valves of o, and o, occur on the 5a. Nddai, Theory of Flow and Fracture in Sol:ds. vol 1, 2d ad., McGraw-Hill, New York, 1950. 3 gy b2+02‘ 7 # | p———-— PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR-DWG 25937 -/ L ] —— —- Fig. 3.210. Elastic Distribution of Stresses in a Thick-Walled Cylinder Under Interncl Pressure. outside surface of the pressure vessel.. The variaticn of 0, across the . wall “thickness o), ~ (ot) 1 equafs the magnltude of the internal pressure, The g, 0, . and o, vs r curves are equs- 3 _dnstnnt loganthmnc curves. ;’ S | | “For &" moderqfely fluck-wailed cylmder (us opposed to a heaw-wclled cylmder) the distribu- . tion of all’ stresses across the wall of the cylinder i " can_be assumed to.vary Imearly rather than. loga- ST ,nthmscally. "Also, for the condmon of creep, the ’ .. material will probobly ‘behave . more like the per- - fectly . ploshc ‘material than an eiasflc material, since the strom rate’ across - the wall of the tube . lls constanf. : By usmg these - assumphons, the' stress dlstrlbutlon in the tubular creep specimens was found to be that shown in Fig. 3.2.12, for either the Henky-von Mises or the maximum shear o 3{fi] o[- e stress criteria for yielding. The octahedral shear stress (Henky-von Mises), T _,, or the maximum shear - stress, T, should occur slmuitaneously over the wall of the tube. The value of Tocy i as follows: | 172 The value of T, ., Will depend upon the average o, relchve to o, und o as follows- - ©Jov 5 _ (3) _(az > 7, > ar) Tmux = 2 +_4' ' ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 25938 ) o P > : ’ 8 g, 1[ 0 o ] F %z € A cro/ \/'5' , | o i Ir = b Fig. 3.211. Plastic Distribution of Stresses in ¢ Thick-Walled Cylinder Under Internal Pressure. p b +a (4) (01>‘7z>0') Tmox =-Z b—a ’ p/ a (az)av (3 (ot >0, > oz) Tmux =—2(b _a) - ) ° For the specific case in which the inside diameter of the tube is 0,843 in. and the outside diameter is 0.963 in., the values of T, _, and T, are given in Table 3.2.2 for the tests reported here, As indi- cated in Table 3.2,2, a series of tests at 1500°F and 4000 psi has been completed. The Inconel specimens were machined from %-in. pipe to have e 0,843-in. ID end o 0,963-in. OD with a 2.5-in. gage length, All the specimens were annealed in 188 hydrogen for 2 hr at 1950°F after machining, The axial creep curves obtained from these tests are shown in Figs. 3.2.13 and 3.2.14, where the axial strain, positive or negative, is plotted vs time, As shewn in Figs. 3.2.13 and 3.2.14, for the tests in which the stress ratio of g, to o, was greater than 1/2, the strain was positive, and, for those in which the stress ratio was less than 1/2, the sirain was negative, The specimen in which the stress ratio of o_ to 0, was 1/2 did not experience strain in the axial direction.” The total strain at rupture (axial, tangential, and radial) is plotted vs the stress ratio in Fig. 3.2,15, and the rupture life of the biaxial creep specimens is plotted vs the stress ratio in Fig, 3.2.16. All stresses shown in o> AXIAL STRAIN (%) PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL~-LR-DWG 25939 Fig. 3.212 Plastic Distribution of Stresses in a Moderately Thick-Walled Cylinder Under Internal Pressure. UNCLASSIFED ORNL-LR~-DWG 25940 20 STRESS (psi), o /oy 05 02 | — 04 " TIME (he) 'Fig. 3.2.13. Creep Curves of Inconel Tubes Bilaxially Stressed and Tested in Argon ot 1500°F. 0 2 5 0 .2 5 0 2 5 w0 UNCLASSIFIED ORNL-LR-DWG 25941 AXIAL STRAIN ~COMPRESSION (%) - 04 1 2 5 w0 2 5 w02 2 5 10° - : TIME (hr) ) - Fig; 3.2.14. Creep Curves of Inconel Tubes Biaxiolly Stressed and Tested In Argon at 1500°F. 189 ANP PROJECT PROGRESS REPORT Tan 190 UNCLASSIFIED ORNL—LR—-DWG 25942 20 - €, _ STRAIN, € / N\ N / \, r—— -2 —14 —16 —18 _20 * —4000/4000 —2990/4000 O/acoo 2999%/4000 4990/s000 4990/2000 4990/ STRESS (psi), 0, /0, Fig. 3.215. Total Stroin at Rupture of Inconel Tubes Creep Tested at 1500°F Under Combined A'x'iul and . | gential Stresses, ' ‘ LF' ‘chle. 3.2.2. Octohedral and Maximum Shear Stresses of Biaxial Creep Specimens Tested at 1500°F in Argon QOctahedral PERIOD ENDING SEPTEMBER 30, 1957 the figures were calculated by the elastic ‘“‘thin- wall’’ formula. These data show some very interesting aspects In Fig. 1500°F, 200 - < , Maximum ‘of high-temperature flow characteristics. "”’/("_‘ ) Sheor Stress, Shear Stress, 3.2,13, the creep curves demonstrate that the o S Toey psl) o T, (psi) axial strain rate is independent of the state of _ _ stress and dependent only upon the stress in the 4000/0 - - 1890 - 2000 axial direction in those tests in which the stress 4000/1000 - 1400 2030 ratio, 0, fo o, is greater than 1/2, It appears N » from Fig. 3.2.15 that if the greatest extension 4000/2000 - 1560 £ 2070 -were taken as a measure of deformation, the con- 4000/3000 1790 2100 clusion would be reached that the state of stress 4000/4000 2020 2140 seriously uffects‘the umou'nt of deformation v:'n , . rupture. There is no basis, however, for this 5000/4000 1820 2140 selection of o measure of deformation; a better 2000/4000 1750 2140 measuwe is the sum of the three strains without 1000/4000 790 2140 regard .to s.im.' 'l:hus it can be seen tfmt the ' elongation is distributed between the axial and 0/4000 1960 . 2]40 transverse directions; what is lost in one direction ~1000/4000 2210 2500 is gained by the other. If the three strains are —2000/4000 2520 3000 summed with respect to sign., the quantity is | reasonably close to zero and indicates the con- -3000/4000 2880 3500 stancy of volume for high-temperature flow, - ~4000/4000 3270 4000 A correlation of rupture life as a function of the stress system has not yet been accomplished. UNCLASSIFIED 40-00/ ORNL-LR-DWG 25943 0 & e 4000/, . . —1 2000 . /‘ LA _ 400%/060 (— —— =/i T 2 2000/4000 prit—d oL " oo T e Peip o s s LW e e e e oo . %4000 [ -1 20004000 / - =4000/3000 - — —— = 1. . e 600~ . 800 . . 4000 4200 1400 . 1600 . - 1800 2000 2200 RUPTURE LIFE {hr) Fig. 3.2.16. Stress-Rupture Properties of Inconel Tubes .Biuxiully_Sfressed’.and Cr&ep Tested in Argon at 191 ANP PROJECT PROGRESS REPORT At high temperatures where intergranular fracture is the cause of failure, there are two factors commonly accepted as contributing to failure: first, the consolidation of vacancies at boundaries, and, second, stress concentrations arising at boundaries because of relative motion of the crystals, The first factor may well be a function of the shear stresses, either the maximum shear stress or the octchedral shear stress (maximum shear stress deviator). In the case of the second factor, the greatest principal stress present in the system appears likely to be a measure of the rate of propagation of the failure, Neither of these two -factors alone appears to govern the time to failure. The correlation is further complicated by the apparent difference in rupture life in tests in which the axial stress is greater than the tangential stress and the rupture life in tests in which the tangential stress was the greater. This behavior may be caused by the material being isotropic; however, the symmetry noted in Fig. 3.2.15 gives an indication that this is probably not the case. This same phenomenon has been reported by other investigators3 in results of tests of tubes biaxially stressed at low temperature, It is their explanation that this anomaly is caused by an instability of the uniform mode of deformation; this same instability can also gradually develop under biaxial stress systems in a simple tension test, RELATIVE TENSILE PROPERTIES OF INCONEL PLATE AND INCONEL WELD DEPOSITS R. E. Clausing P. Patriarca Knowledge of the relative strengths of weld metal and base metal is required for the accurate prediction of the behavior of welded assemblies. In many instances it is necessary only to determine that the weld is as strong or stronger than the base material for the conditions of operation. Unless the weld metal is subjected to unusual cenditions, such as corrosion, impact loads, or loading which will result in excessive weld-metal deformation, failures which occur should be confined to the base metal or heat-affected zones. If the weld proves to be weaker than the base material, this factor must be considered in the design of welded components, Many times welds may be placed in low-stress areas or may be mechanically reinforced to provide the necessary strength. In certain types of assemblies it is desirable to match the properties of the weld with the properties 192 of the base material as nearly as possible.. Matched mechanical and physical properties are important in structures of uniform cross section which may be expected to deform during operation. Unequal strengths in the proximity of the welds will result in unequal deformation and may result in stress concentrations that will lead to early failure of the component, Welds which are more resistant to deformation than the parent plate may be as un- - desirable as welds which deform more readily than the base material, Although in many instances it is not possible to test welded assemblies in simulated service con- ditions, short-time tensile tests provide much valuable information regarding comparative proper- ties of welds and parent material, Early tensile tests of Inconel weldments indicated that, at room temperature, failures occurred in the weld metal, whereas, at elevated temperatures, the failures occurred in the base material. Two specimens that illustrate these conditions are shown in Fig, 3.2.17; one was tested at room temperature and the other was tested ot 1500°F, The experimental work reported here was undertaken in order to determine the relative strengths and elongations at failure of Inconel weld metal and base material. A number of 0,252-in,-dia tensile specimens were machined from an Inconel weldment, as shown in Fig. 3.2.18. The weld was deposited as a fillet by using INCO No. 62 weld wire and PS-1 (ref 6) welding procedures, Preliminary testing revealed, as indicated previously, that the specimens were not likely to fail in the weld metal when tested at the elevated temperatures of interest, It was decided, therefore, to reduce the weld portion of the specimen to a stondard size which would ensure an all-weld-metal test section. The 0.252-in.-dia specimens were etched lightly to accurately locate the weld area, which was subsequently reduced to a diameter of 0.126 in. with o gage length of 1/2 in. A drawing of the resulting specimen is shown in Fig. 3.2.19. Specimens of this type were used to obtain the data reported here. Base-metal tensile specimens were also prepared in the some manner. The base metal had an ASTM grain size of between 6 and 7 and could be classified as fine-grained material. 6ps-1 is the designation of an ORNL. welding specifi- cation, s "a » Tested af Room Temperature PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED Y-23246 Tested at 1500°F Fig. 3.217, Tensile Test Specimens Machined from Inconel Weldments. Fracture occurred in weld at room temperature and in base metal at 1500°F, Y-23310 UNCLASSIFIED ORNL-LR-DWG 22383R ‘fiz-in.-THICK INCONEL PLATE INCONEL BACKING STRIP Fig. 3.218, Location of 0. 252.in.-dia’ Tenslle Specl-_' mens Relative to the Weld Deposh. - Two weld specimens and one base-metal speci- " men were tested at each .temperature.. The base material was tested in the mzll-annealed condmon and the welded specimens were tested i the as- welded "condition.. - A cross-head speed of - 005 - in./min produced a strain rate of 0.1 in./in.smin - on the l/2-in. gage length. The appearance after fracture of the 0.126-in. standard specimen . is Y-23309 1/2 —in. UNCLASSIFIED ORNL-LR-DWG 22384R GAGE LENGTH 04262 0,00t in. 'WELD METAL AREA Fig. 3.219. Modified, Transverse, Weld-Metal Tensile Specimen. : "illliisfrated in Fig. 3.2.20. Although the test speci- men geometry used in this investigation was un- usual, the consistency of the data was quite satis- factory. The experimental data obtained are listed in . Table 3.2,3 aond presented graphically in Fig. 3.2.21. | , As may be seen, both the yield and ultimate - strengths of the weld metal are greater than those ~of the parent metal at elevated temperatures, ~while the ductility of the weld metal is lower. If a component is subjected to plastic deformation, - the 0,2%-offset strength is of paramount importance, 193 ANP PROJECT PROGRESS REPORT . UNCLASSIFIED : Y-23245 Weld Metal at 1600°F Fig. 3.2220. Typical Reduced-Sectian 0,126-in.-dia. Tensile Specimens. Table 3.2.3. Tensile Properties of Inconel Weld Metal and Base Metal T . ost Specimen Yield Strength at Ultimate Strength Elongation at Failure Temperature T 0.2% Offset (psi) % e 0 (°F) ve (psi) ° ) Room Base metal 41,532 99,516 52 Weld metal 50,074 87,982 40 Weld metal 52,670 90,652 40 1400 Base metal 23,790 34,395 52 Weld metal 32,796 44,672 : 45 Weld metal 47,834 40 1600 Base metal 18,870 21,643 ' 68 Weld metal 25,902 27,225 60 Weld metal 26,361 27,624 34 1800 Base metal 9,475 12,500 80 Weld metal 16,617 17,314 56 Weld metal 16,279 16,455 48 194 {l.‘ PERIOD ENDING SEPTEMBER 30, 1957 Y-23311 3 UNCLASSIFIED (x10%) _ ORNL ~LR-DWG 22382R8 80 | | /h' ” BASE METAL // O ELONGATION / A ULTIMATE STRENGTH /0< 20 | O YIELD STRENGTH //e\;\\ . 70 ‘ AT 0.2% OFFSET ot I‘ _ 7”7 WELD METAL // B ELONGATION 7 60 |————— A ULTIMATE STRENGTH 7z 60 ® YIELD STRENGTH // ‘ AT 0.2% OFFSET 7 » + . l = gLONGX * O / 0 @ 40 40 e 0 <1 o S > S 2 Q g Ll -l 5 30 30 =2 ) = 20 20 10 10 oL — — : ‘ — 0 1200 1300 1400 - 1500 1600 - {700 1800 1900 2000 TEMPERATURE (°F) - | ng. _3.2;21.- fensllfi Proficrfles of Inconel Weldments. "lt may ulso be noted rhat at 1400°F the weld meful' : ‘has a yield strength which is approxlmate!y 30% greater than that ‘of the base material; at 1600°F - ably. stronger in tension thon the base material at the yield strength of the weld is 35% greater; and - at 1800°F it is more than 70% greater, It is readlly' ' apparent, ‘therefore, that the weld should not de- form appreciably at these temperatures in structures wuth uniform stress distribution if serwce ‘condi- tions comparable to the test conditions are assumed. In summary, Inconel welds appear to be consider- temperatures - from 1400 to 1800°F, a condition which may be desirable or undesirable depending on the stress distribution in service, 195 ANP PROJECT PROGRESS REPORT BRAZING ALLOYS FOR INCONEL JOINTS Oxi_ddtion Studies . G, M. Slaughter P. Patriarca . The results of static and cyclic oxidation studies, at 1500°F and at 1700°F, of joints brazed with q large number of alloys have been reported.” Other commercial and developmental alloys have since - become of interest, and the results of similar static 7E. E. Hoffman et al., An Evaluation of the Corrosion and Oxidation Resistance of High-Temperature Brazing Alloys, ORNL-1934 (Oct. 23, 1956). studies of these newer alloys are presented in Table 3.2.4. The static test results are compared in Table 3.2.5 with the results of cyclic tests. It is evident that the oxidation resistance of most of these alloys is excellent. Corrosion Studies D. H. Jansen E. E. Hoffman In an effort to find brfizing alloys that could be used in fabricating a NaK-to-fuel heat exchanger, the remainder of a series of high-nickel-content brazing alloys have been corrosion tested in the fuel mixture NaF-ZrF -UF, (53.5-40-6.5 mole %, Table 3.2.4. Oxidation Resistance of Dry-Hydrogen-Brozed Inconel T-Joints Brazing Alloy Composition (wt %) Oxidation in Static Air ot 1700°F* For 200 hr For 500 hr For 1300 hr For 200 hr For 500 hr Oxidation in Static Air at 1500°F* Handy & Harman No, 93 Handy & Harman No, 91 Handy & Harman No. 82 Handy & Harman No. 72 Rexweld No. 64 Haynes No. 40 Haynes No. 8244 Coast Metals No. 52 with added boren Special Nicrobraz No. 50 Coast Metals No, 53 + 1% Li Nicrobraz No. 130 Colmonoy LM Commercial GE No. 81 Experimental lron-Base Experimental Palladium- Base Self-Fluxing Nicrobraz No. 150 Nicrobraz No. 60 Nicrobraz No. 10 Nicrobraz No. 45 L. C. Nicrobraz Cobalt-Modified Coast Metals No, 52 93,3 Ni—3.5 Si—1.9 B-bal Fe and C Slight Slight Slight Moderate Moderate 91.3 Ni—4,5 5i—2.9 B-bal Fe and C Slight Slight Slight Slight Slight 82 Ni-4.55i-2.9 B-7 Cr—bal Feond C Slight Slight Slight Slight Slight 72.5 Ni-5$i-3.5B-16 Cr-bal Fe and C Slight Slight Slight Slight Slight 72.5Ni-4 5i-3.5 B-15Cr—4 Fe-1C Slight Slight Slight Slight Slight 73 Ni—4 Si—3 B—14 Cr—0.4 C-bal others Slight Slight Slight Slight Slight 74.5 Ni=3.5 Si—2 B=9.5 Cr—bol others Slight Slight Shight Slight Slight 91.25 Ni—4.5 $i=3.25+ B~bal others Slight Slight Slight Slight Slight 77 Ni=13 Cr—-10 P + wetting agent Slight Slight Slight Complete Complete 82.1 Ni-4.5 5i-2.9 B~7 Cr-3 Fe-0.5 Moderate Moderate Severe Severe Severs others + 1% Li (voids) {voids) (voids) {voids) (voids) 4.5 5i~3.5 B=0.2 C=bal Ni Slight Slight Slight Slight Slight 55i—3 B—6 Cr=2.5 Fe~0.15C—bal Ni Slight Slight Slight Slight Slight 10.2 5i-19 Cr=5 Fe-0.25 C~bal Ni Slight Slight Slight Slight Moderate 88.3 Fe-4.8 5i~2.8 B-4.1 Cv Slight Slight Moderate Moderate Moderate Pd-Li Slight Slight Slight Severe 7 Complete 3.4 B-15 Cr-0.15 C-bal Ni Slight Slight Slight Slight Stight 9 Si-15 Mn=bal Ni Slight Slight Slight Slight Slight 11 P=bal Ni Slight Moderate Moderate Corfiplete Complete - 6 P-bal Ni Moderate Moderate Moderate ~ Severe Severs 4.55i-3.5 B=13,5 Cr-4,5 Fe-0.15C- Slight Slight Slight Moderate Moderate bal Ni Coast Metals No. 52 + 20% Co Slight Slight Slight Moderate Moderate *Slight, 1 to 2 mils of penetration; moderate, 2 to 5 mils of penetration; severe, greater than 5 mils of penetration, 196 ' t‘} o b fuel 44) and in NaK (56-44 wt %) in a seesaw furnace oapparatus at 1500°F for 100 hr. The specimens used for the tests were brazed Inconel tube-to-header joints. The alloys, their compo- sitions, and the results of the individual tests are listed in Table 3.2.6. The Nicrobraz No. 130 alloy, which exhibited the best corrosion resistance to both mediums, is shown in Fig., 3.2.22, The only bad feature of this alloy is that numerous voids were formed. This void formation is attributed to the brazing cycle, since the voids were present in the as-received specimens prior fo etching. The composition of this alloy is nearly the same as the compositions PERIOD ENDING SEPTEMBER 30, 1957 of two other brazing alloys (Handy and Harman Nos. 91 and 93) tested previously; the boron con- tent differed slightly. It appears that with the alloys containing 91 to 93% Ni, the porosity in the fillet is related to the amount of boron in the alloy. No porosity was found in Handy and Harman alloy No. 91 (1.9% B), very little was observed in Handy and Harman alloy No. 93 (2.9% B), but, as shown in Fig. 3.2.22, a considerable amount of porosity was found in the Nicrobraz No. 130 (3.5% B). In an effort to lower the melting point of the Coast Metals No. 52 alloy, 1 wt % boron was added to the alloy mixture, The additional boron produced a Table 3.2.5. Comparison of Static and Cyclic Oxidation Resistance of Dry-Hydrogen-Brazed Inconel T-Joints Brazing Alloy Composition (wt %) Tested at 1500°F for Tested at 1700°F for 500 hr 500 hr Static Test Cyclic Test Static Test Cyclic Test Handy & Harman No. 93 Handy & Harman No. 91 Handy & Harman No. 82 Handy & Harman No, 72 Rexweld No, 64 Haynes No., 40 Haynes No, 8244 Coast Metals No, 52 with added boron Special Nicrobraz No. 50 Coast Metals No, 53 + 1% Li Nicrobraz No. 130 Colmonoy LM Commercial GE No. 81 Experimental Iron-Base Experiméntfl Palladium-Base Seif-Fluxing 7 Nicrobraz No. 150 Nierobraz No. 60 Nicrobraz VN.ca. 0 Nicrobraz No, 45 L.C. Nicrobraz - Cobalt-Modified Coast Metals No, 52 93,3 Ni=3.5 $i—1.9 B-bal Fe and C Slight Slight Moderate Moderate 91,3 Ni=4.5 $i—2.9 Bbal Fe and C Slight Slight Slight Slight 82 Ni~4.5 $i-2.9 B~7 Crbal Fe and C Slight Slight Slight Moderate 72,5 Ni=5 $i-3.5 B_16 Cr—bal Fe and C Slight Slight Slight Stight 72.5 Ni—4 Si-3.5 B=15 Cr—4 Fe—1C Slight Slight Slight Slight 73 Ni—4 $i-3 B=14 Cr—0.4 C—bal others Slight Moderate Stight Moderate 74.5 Ni~3.5 5i=2 B-9.5 Cr—bal others Slight Shight Slight Slight 91,25 Ni—4.5 $i3.25+ B-bal others Slight Slight Slight Slight 77 Ni-13 Cr—-10 P + wetting agent Slight Slight Complete Complete 82,1 Ni—4.5 $i—2.9 B-7 Cr—3 Fe-0.5 Moderate Moderate Severe Severe others + 1% Li . - (voids) (voids) (voids) (voids) 4.5 $i~3,5 B-0.2 C~bal Ni Slicht Slight Slight Slight 5 $i—3 B=6 Cr—2.5 Fe—0.15 C~bal Ni Slight Slight Slight Moderate 102 $i-19 Cr=5 Fe-0.25 C—bal Ni Slight Slight Moderate Moderate ' 88.3 Fe—4.85i-2.8 B—4.1 Cu Stight Moderate Moderate ~ Moderate Pd-Li Shight Slight Complete Complete 3.4 B15 Cr—0.15 C-bal Ni Stight Slight Slight Stight 98i-15Mabal Ni Stight Siight Slight Slight 11Pbal Ni Moderate Moderate Complete Complete 6 P-bal Ni B | Moderate Severe Severe Severe 4.5 5i=3.5 B-13.5 Cr-4.5 Fe-0.15 C=bal Ni Slight Moderate Moderate Moderate Coast Metals No. 52 + 20% Co Slight Slight Moderate Moderate 197 ANP PROJECT PROGRESS REPORT Table 3.2.6. Resulis of Corrosion Tests in Seesaw Furnaces of Brazed Inconel T-Joints Exposed to Ne F-ZrF4-UF‘ (53.5-40-6.5 mole %, fuel 44) ond to NaK (56-44 wt %) Test period: 100 hr Test temperature: 1500°F in hot zone Tested in Fuel 44 Tested in NaK Brozing Alloy Used : i razing Alloy Use Weight Loss Metallographic Notes Weight Loss Metallographic Notes (%) (%) L.ow-Melting Nicrobraz 0.05 Scattered subsurface voids to 0.04 Depleted to o depth of 3 mils; (83% Ni—6% Cr—5% Si-3% B-3% Fe) a depth of 2 mils; alloy de- subsurface woids to a depth pleted to same depth of 3 mils » Rexweld No. 64 0.07 Depleted to a depfh of 3 mils; 0.06 Particle leaching and void (72.5% Ni-15% Cr—4% Fe—4% Si— subsurface voids to a depth formation to a depth of 2 3.5% B-1% C) of 2 mils mils General Electric No. 81 0.17 Attack to maximum depth of Uniform attack to a depth of (65% Ni=19% Cr-10% Si-5% Fe-1% Mn) 7 mils 3 mils Nicrobraz No. 50 0.08 No attack or depleted alloy Attacked to a depth of 5 mils (74% Ni-13% Cr-10% P=3% others) found; voids in fillet Nicrobraz No, 130 0.01 Light, subsurface voids to a 0.03 Depleted to maximum depth of (92% Ni-4.5% 5i-3.5% B) depth of 1 mil; numercus 3 mils voids in body of fillet Coast Metals No, 52 plus boron 0.007 Porous areas in fillet; fillet 0.03 No attack or depletion; fillet (89% Ni-5% Si—4% B-2% Fe) + 1% B severely cracked severely cracked Coast Metals No. 53 plus lithium 0.10 Alloy depleted to depth of 0.06 Depleted to o depth of 2 mils (81% Ni—8% Cr—4% Si—4% B-3% Fe) + 3 mils; voids present 1 mil 1% Li from surfoce; porous areas in body of fillet UNCLASSIFIED iy o, |’ YRS NGk ‘9,01 BRRINN UNCL ASSIFIED R Y.22251 - ® - %(a,*' ‘a ! ”f > v v o s Fig. 3.222. Nicrobrax No, 130 Brazing Alloy (92% Ni-4.5% Si=3.5% B) After Exposure to (a) NaF-ZrFiUF‘ (53.5-40-6.5 mole %, fuel 44) and to (b) NaK (56-44 wt %) for 100 hr at 1500°F in Seesaw Furnace Apparatus. Note the, limited attack and the porous dreas formed throughout the fillet during brazing. Etchant: electrolytic oxalic acid. 75X. Reduced 16%. {(Seecred-with-cuptiom) 198 brittle alloy that cracked during the brazing opera- tion. The alloy obtained with the boron addition showed the same corrosion resistance to fuel 44 and to NaK as did the orlgmal Coast Metals No. 52 alloy. A Coast Metals No. 53 alloy with 1 wt % lithium added to increase flowability during the brazing operation also showed essentially the same cor- rosion resistance as the non-lithium-bearing Coast Metals No. 53 alloy when tested in fuel 44 and in NaK in the seesaw furnace apparatus. : The Handy and Harman alloys Nos. 91 and 93, which showed good resistance to the fuel 44,8 have been tested in NaF-KF-LiF-UF, (11.2-41- 45.3-2.5 mole %, fuel 107) at a hot-zone tempera- ture of 1500°F for 100 hr in the seesaw furnace apparatus. Both alloys showed about the same resistance to fuel 107 under these conditions. Corrosion was limited to 3 to 4 mils of depletion of the minor constituents from the exposed edge 8p. H. Jansen and E. E. Hoffman, ANP Quar. Prog. Rep. June 30, 1957, 0RNL-2340, p 230. UNCLASSIFIED Y-23064 ,fl “”“ "*(a)g, PERIOD ENDING SEPTEMBER 30, 1957 and the formation of small subsurface voids in the ‘depleted region, as may be seen in Fig. 3.2.23. ‘Coast Metals alloy No. 52 was also tested in fuel 107 under the same conditions. Subsurface void formation was more extensive in this alloy than in the Handy and Harman alloys, as shown in Fig. 3.2.24, EFFECT OF GRAIN SIZE ON CORROSION OF INCONEL BY FUEL 30 J. H. Devan R. S. Crouse During the examination of the fuel circuit of the NaK-to-fuel heat exchanger ORNL-1, type IHE-3, unusually deep intergranular voids were found in tubes near the tube-to-header joints of the NaK outlet header.? The Inconel tubing in this area had relatively large grains because of the heat freatment required to back-braze the tube-to- header joints. It was therefore of interest to de- termine whether the metallurgical changes accom- panying the grain-coarsening heat treatment of Inconel could affec_t the corrosion resistance of 9G. M. Slaughter, ANP Quar. Prog. Rep. June 10, 1956, ORNL-2106, Fig. 3.4,36, p 198. UNCLASSIFIED Y-22836 T oy L5) Flg. 3.2.23. Bruzlng Alloys .(a) Hufidy & Harman No. 91 and (b) Hondy & Harman No. 93 After Exposure to - NaF-KF-LiF-UF (11,2-41-45.3-2.5 mole %, fuel 107) at 1500°F for 100 ht in Seesaw Furnace Apparatus, Etchant: 10% oxalic acid, electrolytic. 150X. Reduced 11%. Sosretawith-eeption) 199 ANP PROJECT PROGRESS REPORT Y "y Xml ! i ity SR i i 5\] l.‘_b\ ) im‘ ST g J\ \ i Wn N BEEREN (INCLASSIFIED B Y.22835 Fig, 3.2.24. Cocst Metals Brazing Alloy No. 52 (89% Ni=5% Si-4% B-2% Fe) After Exposure for 100 hr ot 1500°F to NaF-KF-LIF-Ufi (11:2+41045.3+2.5 mole %, fuel 107) In a Seesaw Furnace Apparatus. Etchant: 10% oxallc acid, electrolytices 150X: (Sacredwrith-esption) this alloy in molten flucrides, Accordingly, an Inconel forced-circulation loop, 7641-50, was prepared with hot-leg sections which had been annealed at 1940°F for 2 hr in hydrogen. The annealing treatment was intended to produce grains similar to those formed during the furnace brazing of heat exchanger and radiator tubes, The loop was operated with NaF-ZrF -UF , (50-46-4 mole %, fuel 30) for 1000 hr with @ maximum fuel-to-metal interface temperature of 1600°F and a minimum fuel temperature of 1300°F. Under similar conditions, a standard Inconel loop (ASTM. grain size 4-8) would show 5 to 7 mils of general and intergranular subsurface void formation, The sizes of the grains of the annealed sections, measured in both before-test and after-test samples, were very random, ranging from ASTM 5 to greater than ASTM 1 within a single sample. After-test samples tcken from the point of maximum wall temperature revealed heavy general and inter- granular void formation to a depth of 5.5 mils. In addition, a few scattered intergranular voids were visible to @ depth of 12 mils, as shown in Fig. 3.2.25. Thus preferential grain-boundary attack, 200 . UNCLASSIFIED - T-12848 ERE ¥ mer!a EEFEEREE Fig. 3.225. Intergranuvlar Voids Formed During Operaticn of Inconel Forced-Circulation Loop 7641.50 for 1000 hr with Fuel 30. Heated sections of loop had been annealed at 1940°F for 2 hr prior to operation. Etchant: coqua regia, 250X, Reduced 29%. (beerat with-copticn), 4% such as that found in the exariination of heat exchanger units, occurred to a limited extent in the coarse-grained sections of this Inconel locop. Thermal-convection loop tests were run several years ago to evaluate grain-size effects, but little difference in the depth of attack was noted between as-received and annealed specimens.'® In re- 10 G. M. Adamson, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL 1771, p98- ; br eb PERIOD ENDING SEPTEMBER 30, 1957 viewing " the previous tests, however, it appears that subsurface voids were much more concen- trated in the grain boundaries of coarse-grained tubing than in the grain boundaries of fine-grained tubing, While the depth of the intergranular voids in the coarse-grained tubing did not greatly exceed the depth of voids in the grain matrix of fine- grained samples, such tests indicate the possibility of relatively deep attack under certain conditions of grain-boundary orientation in coarse-grained material, 201 ANP PROJECT PROGRESS REPORT 3.3. WELDING AND BRAZING STUDIES P. Patriarca EXAMINATION OF ALL-WELDED INCONEL IMPELLER THAT CRACKED IN SERVICE G. M. Slaughter The sump-type centrifugal pump (model DANA), which was used for circulating the fuel mixture NaF-ZrF“-UF‘ (50-46-4 mole %, fuel 30) in small heat exchanger test stand SHE-C, was found, during a routine examination aofter termination of a test, to have a cracked impeller. The all-welded Inconel impeller had operated for 2350 hr in the temperature range of 1200 to 1500°F, with the bulk of the operation at 1300°F. The location and the extent of the cracks in the welded vane- to-plate joint are shown in Fig. 3.3.1, and o photomicrograph of one of the cracks is presented in Fig. 3.3.2. The impeller vane-to-plate welds were fabricated according to the weld joint design shown in UNCLASSIFIED Y-22570 F'g- 3-30]0 Showing Location and Extent of Cracks Found After Operation for 2350 hr in the Temperature Ronge of 1200 to 1500°F with Fuel 30. (SecreTwith—cupriom AlleWelded Inconel Pump Impeller NCLASSIFIED Fig. 3.3.2. Photomicrograph of Weld Fracture Shown in Figs 3.3.1. 100X. Etchant: electrolytic oxalic acid, 202 ¥ Fig. 3.3.3. It is evident that the fabrication procedure provided only o limited joint length where the vanes were welded to the top plate. Furthermore, the joint design and lack of accessi- bility permitted only very shallow weld pene- tration. This condition is exhibited in the upper left of Fig. 3.3.4, which shows a polished and macroetched section of the impeller. A modified fabrication procedure is now being used in which the joints are brazed for additional reinforcement and to improve reliability. BRAZED-JOINT CRACKING TESTS G. M. Slaughter | Preliminary tests have been conducted to de- termine the extent of braze cracking during the fabrication of high-conductivity-fin NaK-to-air radiators at the York Corp. The cracks have resulted from the transverse shrinkage and dis- tortion of the tube sheets during the deposition of the longitudinal welds of the headers. As a means of simulating the conditions under which the cracking occurred, tube-to-header broze FILLET WELD AS MUCH AS POSSIBLE FROM OUTSIDE,; WELD LENGTH X AND Y TO BE SYMMETRICAL FOR ALL VANES SECTION BB 'SECTION BB IS TYPICAL ‘OF WELDS AT TIPS OF 5 VANES FOR APPROXIMATE DISTANCE SHOWN PERIOD ENDING SEPTEMBER 30, 1957 fillets were prepared and then cracked by bending the headers. The degree of deformation obtained upon bending is apparent in Fig. 3.3.5, which shows the samples before and after bending. Bending of the header to the extent indicated produced cracks which were considered to be- visually representative of those found in the radiators fabricated at the York Corp. 'Metallographic examinations of the cracked joints revealed the cracks to be in the braze fillet. In no case did the cracks penetrate into the tubes or the tube sheet. A typical crack in the braze fillet is shown in Fig. 3.3.6. ‘Sample joints were then rebrazed with and without the use of additional Coast Metals No. 52 brazing slurry. The cracks appeared to completely heal in. all cases. The addition of slurry to the fillet before rebrazing appeared to be advan- tageous, however, in that it permitted a more uniform fillet and provided an additional source " of alloy for healing the cracks. UNCLASSIFIED ORNL—LR—DWG 25944 SECTION AA Fig. 3.3.3. Weld Deiullfi of Impeller Vaneoto-Plate Joints. 203 ANP PROJECT PROGRESS REPORT BN UNCLASSIFIED Y-22571 | - Figs 3.3.4. Vane-to-Plate Welds of All-Welded Inconel Impeller Fabricated According to Design Shown in Fig. 30303- FABRICATION OF ART FUEL FILL-AND-DRAIN TANK E. A. Franco-Ferreira G. M. Slaughter As reported previously,! a study of some of the problems associated with fabrication of the ART fuel fill-and-drain tank is under way. Specifically, tests are being run to establish suitable welding and brazing procedures for making the tube-to-tube sheet joints in the tank heads. As a result of the high degree of restraint exerted by the 1)y-in.-thick tube sheet, the tube- to-tube sheet welds, when made with the stendard fusion procedure used for thinner tube sheets, have consistently exhibited root cracks of the type illustrated in Fig. 3.3.7. In order to produce ‘E. A. Franco-Ferreira, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 225. 204 ‘UNCLASSIFIED - Y.23346 Fig. 3.3.5. Simulated Tube-to-Header Joints for Braze-Cracking Tests. joints free of such root cracks, a low restraint, trepanned type of joint was designed. The joint design and the welding procedure for obtaining crack-free welds are presented in Fig. 3.3.8. A machine welding method is preferred for the fabrication of these joints in order to assure consistency and reliability. Welds made according to the prescribed procedure are shown in Fig. 3.3.9. | Brazing experiments were also conducted in order to determine the proper amount of preplaced brazing alloy and the type of preplacement pro- cedure to be used, as well as to evolve a suitable brazing cycle. Calculations were made to de- termine the amount of Coast Metals brazing alloy No. 52 required to fill a 0.003- to 0.004-in.-wide onnulus between the %-in.-OD tube and the I‘/z-in.-thick tube sheet. The results indicated that two brazing rings, each weighing approxi- mately 1.3 g, would be required for the brazing of each tube. : The configuration of the specimens, that is, the thick tube sheet and the comparatively thin- " walled tubes, posed additional problems.- When the assembly is being heated to the brazing temperature (1920°F), the temperature of the massive tube sheet lags behind the temperature of the tubes. Several joints may be seen in i i 1 2= e PERIOD ENDING VSEP TEMBER 30, 1957 UNCLASSIFIED Fig. 3.3.:6. Typica!l Crack in Braze Fillets 100X. Ase-polished. Fig. 3.3.7. Root Crack in o TubestosTube Sheet Joint Made Under Conditions of High Restraint. Etchant: electrolytic oxalic ccids 50X. Fig. 3.3.10 in which the brazing alloy failed to fillet because the ring melted around the hot tube ond did not wet the cooler tube sheet. The use of radiation shielding, which minimized direct radiation on tubes from the heat source, produced the results shown in Fig. 3.3.11. Such shielding definitely improved the reliability of filleting, but the use of shielding would be impractical for the fabrication of the full-scale heads. Therefore a relicble brazing methed has been developed which depends upon keeping the brazing alloy away ~from the tubes until the tube sheet has become hot enough .to melt the alloy. In order to ac- ‘complish this, the brazing alloy is contained in ~annular sumps machined in the tube sheet around each tube. ~ When the tube sheet reaches the brazing temperature, the brazing alloy flows into “the ‘space between the tubes ond tube sheet ~through three holes which are drilled from each sump into the tube hole. A detail of the sump -arrangement.is shown in Fig. 3.3.8. "Both dry powder and cast rings were used successfully for the preplacement of the brazing alloy on the test specimens. When powder was used, the sumps were filled level with the tube 205 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 25945 ~ JOINT .DESIGN: WELDING SEQUENCE : t = 0060 in., TREPAN DEPTH = 2/ TUBE, 0.625-in. OD ! ”“I ) MAXIMUM - _ OVERLAP - . MINIMUM £ CURRENT RS TAPER { Y5 in. EXPAND -] DRILL AND REAM TO 0.004- 1 0.001-in. CLEARANCE ON DIAMETER ALLOY FEED HOLE SUMP (2/ DEEP) (¥/32-in. DRILL) BRAZING ALLOY SUMP DETAIL WELDING CONDITIONS : MACHINE MADE WELD CUP SIZE — ¥g-in. OPENING ELECTRODE—- THORIATED TUNGSTEN %3, in., POINTED ELECTRODE PROJECTION — ¥4 in. ELECTRODE SWING — 0.640-in. DIAMETER ARC DISTANCE — 0.050 in. ARGON FLOW THROUGH TORCH—45 TO 50 cfh HELIUM BACKUP GAS FLOW—15 cfh ARC CURRENT—58 TO 60 amp ELECTRODE TRAVEL SPEED (PERIPHERAL)—5 in./min Fige 3.3.8. Joint and Welding Procedure Details for Tubé-to-Tube_Sheef Welds of ART Filleand«Drain Tpnk- (Sectatunithcaption) 206 UNCLASSIFIED Y297 Test No. 4 RN Fig. 3-3.9; Tube-to-Tube Sheet Welds Made by Using the Optimum Welding Procedure Prescribed for Such Welds in the Heads of the ART Fue! Fill-and=-Drain Tank. @eersiauith caption) UNCLASSIFIED Y-215% Fig. 3.3.10. Tube-to-Header Joints Made ot o Rate of Rise to Brozing Temperafure of 300° F/hr Without Radiation Shielding. PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED Y-21 %1 Fige 3.3.11. TubestoeHeader Joints Showing Good Fillets Obtained By Using Radiation Shielding. sheet, and then the powder was wetted with Nicrobraz cement. The use of rings simplifies the prepfacement procedure, but it is desirable to prime the rings with a small amount of powdered alloy wetted with Nicrobraz cement. The brazing cycle finally decided upon provides for a rate of temperature rise of 300°F/hr to the brazing temperature of 1920°F, 1 hr at that temperature, and cooling ot a rate of 300°F/hr. The relatively low rate of temperature rise is required to avoid brazing alloy liquation before ‘the massive header sheet has reached the brazing temperature. CORRELATION OF RADIOGRAPHIC - AND METALLOGRAPHIC DETERMINATIONS OF POROSITY IN TUBE-TO-HEADER WELDS G. M. Sluughter The determination of porosity in tube-to-header welds by radiography requires the use of carefully developed procedures ond skillful interpretation ~ of the radiographic film. - Further the geometry of tube-to-header joints : containing - small-diameter thin-walled tubes is so complex that a precise correlation of the results of radiographic and 207 ANP PROJECT PROGRESS REPORT metallographic examinations would be extremely beneficial. A preliminary study has therefore been conducted to obtain the basic information needed for such a correlation. For this investigation, approximately 60 welds were made by inert-gas-shielded arc welding of short lengths of Inconel tubing, 9’“ in. OD with 0.025-in. walls, into holes drilled on the circum- ference of sections of 3-in. sched-40 Inconel pipe. Each weld was given o designation before it was radiographed, and, ofter the films had been carefully examined, the locations of pores were noted on the weld specimens. Five welds that contained defects were then removed and the pores were located by alternately grinding and polishing, with a microscope being used frequently as a visual aid. Gross porosity was found in every case where it was revealed in the radiographic examination. The sizes of the voids detected are summarized in Table 3.3.1. It may be seen that the pores in these five welds were relatively large ond thus were readily detectable by careful radiographic examination. A similar investigation will be Table 3.3.1. Sizes of Voids Detected in Tube-tc-Header Welds by Radicgraphic Examination Void Dimensions® Weld No. Defect Designation (in.) 1 P-3-1-Wé 0.013 x 0.011 2 P-3-1-W8 0.012 x 0.008 3 P-9-1.W§ 0.009 x 0.009 4 P-16-2-W12 0.013 x 0.010 5 P-9-2-W12 0.010 x 0.008 *Void diameter was measured in two directions with a Filar eyepiece. conducted in which finer pores, in the range of 0.003 to 0.005 in., will be studied in order to determine the limitations of detection by the radiographic method. FABRICATION OF SODIUM-TO-NcK HEAT EXCHANGERS G. M. Slaughter The brazing of the two sodium-to-NaK heat exchangers recently received for installation in 208 the ETU north head was observed at the Ferrotherm Corp. The brazing equipment, pro- cedures, and general furnace cycles utilized were described previously in connection wuth the fabri- cation of radiators at the York Corp.?2 ' During the brazing of heat exchanger No. 1, the exit dew point from the retort was approximately ~100°F throughout the brazing cycle. The spread of the readings of the four thermocouples at the start of the hold peried was 6°F, from 1914 to 1920°F, and at the end of the hold period was 11°F, from 1921 to 1932°F. All thermocouples reached 1920°F ot some point in the brazing cycle, but the moximum temperature attained was 1932°F.. During the brazing of unit No. 2, the exit dew point from the retort was approximately ~95°F throughout the brazing cycle. The spread of the four thermocouples at the start of the hold period was 5°F, from 1918 to 1923°F, and ot the end of the hold period it was 4°F, from 1920 to 1924°F. The maximum temperature aflumed was 1927°F. Upon opening the retorts, the units were found to be clean and bright, oand good filleting was evident at all tube-to-header joints. Minor brazing alloy run-off was noted at one header, but it is thought that this could be minimized in the future by the use of commercially available stop-off compounds. The two job samples, which were brazed at the saome time as the complete units, were evaluated at ORNL and found to be satis- factory. INVESTIGATION OF MATERIAL FOR RADIATOR HEADERS G. M. Slaughter Several NaK-to-air radiator headers have been completely machined at York Corp. and cre ready for welding to the tubes. York has reported, however, that these headers were fabricated from material which is difficult to weld. The welds made thus far have cracked and have been found to be porous. Further, there has been inconsistent penetration. Consequently samples of the header material were transmitted to ORNL for examination and evaluation. Preliminary conventional machine 26, M. Slaughter, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 223. i ,! welding tests were conducted, and distinctly different welding behavior from that of normal Inconel was evident. The tube-to-header welds of the York material had a *‘grainy’’ appearance. Low-power examination and dye-penetrant in- spection also revealed cracks in the welds. In an attempt to determine the reason for the observed differences in welding characteristics, o metallographic comparison was made between the York material and some lnconel plates of known satisfactory welding characteristics. All samples of the questionable material were found to contain a quantity -of inclusions (possibly oxides) distributed randomly as stringers. The samples of known satisfactory ~characteristics were relatively free of inclusions.: The distri- bution and appearance of the stringers of the York material are shown in Figs. 3.3.12 end 3.3.13, and a typical area in material which has exhibited satisfactory welding behavior is shown in Fig. 3.3.14. The inclusions which appear .as dots are probably titanium nitrides ond were evident in all samples examined. Such inclusions are apparently unobjectionable. Efforts to positively establish the difference between these materials PERIOD ENDING SEPTEMBER 30, 1957 by conventional chemical analysis were unsuc- cessful. Welding tests are being conducted on other lots of inspected Inconel pipe in an effort to ' provide the necessary quantity of suitable material for York. FABRICATION OF A SMALL SEMICIRCULAR HEAT EXCHANGER E. A. Franco-Ferreira - The construction of a semicircular heat ex- changer for experimental tests was undertaken. A description of the heat exchanger and the test objectives was presented previously.3 The initial planning and procedura! experimentation have been completed, and it is expected that the first of the two heat exchangers being fabricated will be available soon. The initial experiments, which were made in order to devise suitable welding and brazing procedures for the tube bundle, were performed on scrap material. An example of the experimental 3J. C. Amos et al., ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 43. INCHES 0.02 0.03 Fig. 3.3.12. Stringers in Inconel Header Mnferriulk Available at York Corp. Asepolished. 100X. 209 ANP PROJECT PROGRESS REPORT Fige 3.3.14. Inconel Which Exhibited Satisfactory Welding Characteristics. Note relative freedom from large stringerss Asepolished. 100X. 210 .n results is shown in Fig. 3.3.15. Work was not started on the actual heat exchanger until all procedures had been carefully planned. The components of the heat exchanger are shown in Fig. 3.3.16. All parts were minutely checked before the assembly sequence was started. Experimentation had ~ shown that successful assembly of the tube bundle into the tube sheets, and subsequently, into the channel would depend ‘primarily upon extreme accuracy in- the bending - of the tubes. Consequently the jig shown in Fig. 3.3.17 was built so that ‘the tubes could be individually checked for conformity to the required dimensions. - Thus, each tube in the heat ex- changer was hand-picked for assembly by checkmg _ it as shown in Fig. 3.3.18. In thé course of os_sei'nblyk of the tubes into the tube sheets it was found that small inaccuracies in bending, even though within the specified tolerances, gave rise to undue stresses under the restraint of the comblike tube spacers. As a result, a decision was made to stress-relieve the tube bundle in the channe!l before welding and brazing the tube sheets. The tubes were as- sembled into the tube sheets, and the tube bundle PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED PHOTO 40469 Fig. 3.3.15." Prelimincry Welded and Brazed Semi- circular Heat Exchanger Test Headers As-polished. 100X+ Reduced 31%. UNCLASSIFIED PHOTO 40670 Fige 3-3-16. Components for Semicircular Heat Exchanger. 211 ANP PROJECT PROGRESS REPORT Fig. 3.3.18. Tube Being Checked in Jig. UNCLASSIFIED PHOTO. 41101- UNCLASSIFIED PHOTO. 41100 & was placed into the channel with shim spacers between tubes. The resulting assembly, shown in Fig. 3.3.19, was sfress-relievse,cvl_. in dry helium at 1500°F for 1 hr. After the stress-relief treat- ment, the comb spacers were welded into the tube bundle, ond the tube sheet welds were made. Assembly methods dictated that the tube ends protrude randomly beyond the tube - sheets, as shown in Fig. 3.3.20. Since it is of prime im- portance that the distence between the face of the tube sheet and the first bend in each tube be known accurately, the distance that each tube end protruded beyond the tube sheet was measured accurately with a depth micrometer before the ends were cut off flush with the tube sheet. This measurement was correlated with the over-all measurement of each tube before assembly, which was obtained by the photo-lofting technique illus- trated in Fig. 3.3.21. The length of the cut-off portion was subtracted from the original length to obtain the required dimension. The NaK inlet and outlet nozzles were welded to the tube sheets before the tube sheets were. brazed. The assembly and the preplaced Coast Metals brazing alloy No. 52 rings are shown in Fig. 3.3.22. The tube sheets were then back brazed, one at a time, in a special, curved retort PERIOD ENDING SEPTEMBER 30, 1957 which was placed in @ pit furnace, as shown in Fig. 3.3.23. The remaining components are to be assembled by welding. UNCLASSIFIED PHOTO 41103 Fig. 3.3.20. Tube Sheet with Tube Ends Protruding ot Random Lengths. UNCLASSIFIED PHOTO 41102 Fige 3.3.19. Assembled Heat Exchanger Jigged for Stress Relieving. 213 ANP PROJECT PROGRESS REPORT UNCLASSIFIED: PHOTO 40885 - * . Ve D P .%&w;wm R 3 .m...m ! SR ; z, i 17 3k o 51 Method Used for Accurately Measuring Dimensions of Tube Bends. Fig- 3-3021. Header Assembly with Nozzles and 3.22. 3. Preplaced Coast Metals Brazing Alloy No. 5 - 52 Rings. UNCLASSIFIED PHOTO 40889 I S Al A Fige. 3.3.23. Special Curved Retort for Brazing the Tube Sheets of a Semicircular Heat Exchanger. 214 PERIOD ENDING SEPTEMBER 30, 1957 '3.4. CORROSION AND MASS TRANSFER STUDIES CORROSION OF YANADIUM IN FL-UORIDE FUEL AND IN LITHIUM D. H. Jansen E. E. Hoffman The possibility of usmg vanadium as a con- stituent of a container material for the fuel mixture NaF-KF-LiF-UF (11.2-41-45.3-2.5 mole %, fuel 107) ond for lithium is being studied. In order to investigate the effect of fuel 107, o vanadium capsule clad with type 347 stainless steel was filled to one-third its- volume with the fuel and inserted in a seesaw furnace apparatus for 100 hr with the hot zone of the capsule at 1500°F, A considerable amount of mass transfer was found after the test in the cold zone of the capsule, as shown in Fig. 3.4.1. The crystals were found by chemical analysis ‘to ‘contain, in - addition to vanadium, 0.95 wt % Fe, 0, 21 wt % Ni, and <0.50 wt % Cr. Thus it appeared that ‘the vanadium . capsule material was - not- - pure. An’ analysis of the material from which. the cnpsule, - was fabricated showed traces of the same . im- purities with the same relative magnitudes. ~The fuel picked up approximately 700 ppm of voncdlum : during the test. The surface urec-to-volume ratio of the capsule was 10 in.2/in.3. Thickness measurements of the . capsule - dicated uniform removal - of vanadlum to @ depth“ of 1.5 mils from the hot zone of ,ihe capsule, - UNCLASSIFIED TUess TYPE 347 Fig. 3.4.1. Cold VZon;- of ~Vdnd_diufn' Capsule Exposed . to Fuel 107 for 100 hr in Seesaw Fumace Appurcfus with e Hot-Zone Temperoture of 1500°F and a Cold-Zone. Temperature of 1200°F, 5X. Reduced 35%. ('561:?!1' " rion) : and there was intergranular attack to a depth of 1 mil, as shown in Fig.‘3.4.2l. A static vanadium CQpédl‘é.-"-f&i‘ll.éc‘i' with lithium was also tested. for' 100 hr at 1500°F. Slight surface roughenmg on the walls of the bath zone ~ was ' the extent of the attack found after this test, | ";.MOLYBDENUM AND NIOBIUM IN CONTACT WITH LITHIUM E. E. Hoffman Mdiybdenum and niobium were tested in contact with lithium in o seesaw-furnace test for 500 hr with o hot-zone temperature of 1500°F and a cold-zone temperature of 950°F. The test capsule ‘,_t:‘q_ns'is‘ted of & molybdenum or a niobium cup enclosed in a jocket container of type 316 stain- less steel, as shown in Fig. 3.4.3. The two types of copsules were loaded with lithium, placed in seesaw-furnace apparatus, and cycled " at arate of 1 cpm. The lithium bath was alter- _ ‘thquf in contact with the refractory metal at approximately 1500°F and with the stainless- “steel cold zone at approximately 950°F. Following - the test, the lithium was removed and the copsules were examined. 'Mefu'llographfc inspection revealed no attack . 7 ... on either the molybdenum or the niobium hot-zone - VANADIUM - walls, as shown in Figs. 3.4.4 and 3.4.5. Type | 34 2_5-31'6 stainless-steel specimens from the cold-zone . STAINLESS - - STEEL - . “walls showed less than 0.0005 in. of surface ‘roughemng and less than 0.0005 in. of scattered - mass-transfer crystals. These “small crystcls - were similar to crystals seen in previous experi- - _ments with lithium in all-stainless-steel systems. Microspark spectrographic examination of the stainless-steel - wall indicated no detectable in- ~crease of molybdenum or niobium in comparison _with standard samples of type 316 stainless steel. The results of these: tests indicate that under the conditions of these experiments, there was little tendency for dissimilar-metal mass transfer or temperature-gradient mass transfer of molyb- denum or niobium to the stainless-steel wall. 215 ANP PROJECT PROGRESS REPORT. L UNCLASSIFIED | Y.22939 ¢ ' 23 @ Wl I... £ < QU2 G033 _094 005 _.'gwcé > O O—. W NN Fig. 3.4.2. ' Intergranular ‘Attack In Hot Zene (1500°F) of Vonadiom Seesaw Test Capsule After Exposure for 100 hr to NaF-KF—L!F-UF‘ (11.2-41-45.3-2.5 Mole %, Fuel 107), 500X. Etchant: NH,OH + H202.- Seeretwitr taptienl NPT TUNGSTEN-NICKEL-COPPER ALLOY ORNL-LR-DWG 25946 |N SOD"JM ’ W. H. Cook E. E. Hoffman Two specimens of the alloy Mallory 1000 (90% W-6% Ni-4% Cu), manufactured by P. R. Maliory and Company, Inc., were exposed to- sodium for 100 hr in the hot zones (1500°F) of Inconel containers inserted in a seesaw-fumace apparatus cycled at a rate of 1 cpm. Maliory 1000 is being considered for use as a transition layer between tungsten - carbide—cobalt cermets and Inconel in disk and seat brazed joints of sodium or NaK valves for use at high temperatures. One of the specimens was tested as-received and the other was heat treated at 1920°F for ‘é hr in a furnace with heating and cooling rates of 300% /hr. . This heat treatment simulated a fab- rication operation. that would be involved in construction of the valves. ' AR ' Metailographic examinations indicated that the Fig. 3.4.3. Seesaw Fumace Test Capsule. specimens were equally attacked. The untested TYPE 346 STAINLESS STEEL {1in. SCHEDULE 40) COLD~ZONE {950°F) THERMOCOUPLE 216 » PERIOD ENDING SEPTEMBER 30, 1957 8l UNCLASSIFIED e, Ya23334. | o X00§—| |e——— YouI00°0 Fig. 3.4.4, Wall of Molybdenum Capsule Following Exposure for 500 hr to Lithium in the Hot Zone of a Seesaw Furnace at 1500°F. Etchant: 50% H,0, + 50% NH,OH. 500X. {Seeretsmiticuption) X008 ——=| |e—— youig00°0 ‘Fig. 3'4_5. Wall -of_i Niobium :Cupmle F_o“dwipg Exposure for 500 hr to Lithium in the Hot Zone of a Seesaw Furnace ot 1500°F. Etchant: HF + H,0 + H,S0, + HNO,. 500X. (Sestetwith-capTion) - ' 217 ANP PROJECT PROGRESS REPORT and tested specimens are shown in Fig. 3.4.6. ‘Sodium had completely penetrated each specimen {nominal dimensions: :/?2 X ‘4 X '/2 in.). The - untested specimens had pore-space areas that constituted approximately 1% of the microscopic field of view, while the pore-space areas of the tested specimens constituted approximately 4% of the field. The attack appeared to have enlarged existing pores and created new ones. Each speci- men had a weight loss of 1.3%. Apparently, the quantity of copper present in Mallory 1000 is sufficient to cause this material to be of doubtful value for long exposure to sodium or NaK at 1500°F. TUNGSTEN CARBIDE- AND TITARIUM CARBIDE-BASE CERMETS IN SODIUM - W. H. Cock E. E. Hoffman Three types of tungsten carbide- and titanium carbide-base cermets were exposed for 100 hr to sodium at 1200°F in Inconel capsules inserted ¢ UNCLASSIFIED . Mgargdley Y-23268 ;\," Y % LX . ' ' : SO Ty oy YA SO o ffi T ) ¢ L'l A% i i 4 e 4 M )"‘ "»" :f 5 Yt v AL b{: et IR b o 7 + 4 Y (o v G TN &N ; . g " 1 "L > < £ " N et 3. > Ty 4 < "L <) AL £ . iy - ur - : / Y " 2 2 « - - ) L, L i Lot > ;o ; - * kY o & - - s - 3 ol 3 - . ’ W . Ty Ll . e 5 e z . J : YIS 3 : “; “"155;\}" Lo w beyih 410,03 v . 3 3, gus ] S '\l’d! ('\.\‘-;r A ¥ i ¢ i {{a) " ie Ny SO ey o 1A 3 R * ~ s e _in a seesaw-furnace apparatus, The designations and compositions of the cermets are given in Table 3.4.1. These corrosion tests were part of over-all evaluation tests to - determine the suitability of these materials as seat components of valves for sodium or NaK service at 1000°F. The specimens were tested at 1200°F rather than at 1000°F in an oftempt to make the corrosion, if any, severe enough so that the differences in corrosion resistance would be more apparent; however, metallographic examinations of the tested specimens indicated that they were not attacked. : It is interesting to note that K162B was attacked by sodium to a depth of 0.0005 in. in a similar test in which the temperature' was 1500°F. w. H. Cook, ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 171. UNCLASSIFIED < ¥e23269 0 il fl(C‘_ — 15, T Y L ¥ wa i 3 }/‘1 et Y KR 43 " TN S N 3 P A o 5 a2, - i A 1 § =By C Ty Y AN A £ s o HE RS % . - _ - o R Ve . ‘. i = Loy % i g ;4 o ' e . - 4 ; St Y . e T A PN K i T o e ‘. E F St LE o Sk e . i o AN Lo B oY a7 \ v F - v £y T o . i ’ e s e | s A P o . { ; ‘_/ s : 3 z =‘\ 4 c ‘ 9 , A h 7 4 {,\ ol S et Jl ..\ - ; : 4, Fig. 3.4,6. Mallory 1000 (90% W-6% Ni-4% Cu) (c) As-Received and (b) After Exposfire for 100 hr to Sodium at 1500°F in an Inconel Container in Seesaw Furnace Apparatus. The globular particles are tungsten; the black areas are voids. 100X. Unetched. 218 @ & PERIOD ENDING SEPTEMBER 30, 1957 Table 3.4.1. Designations and Compositions of Cermets Tested in Sodium at 1200°F in Seesaw Furnace Apparatus ~ Chemical Composition {(wt %) Designation* Primary Component Co Ni Mo W Ti Ta Nb C K94 - WC 11.75 80.6 = 1.4 0.6 5.4 KM | WTIC, 10.75 685 67 54 17 67 K162B TiC 25 5 521 03 45 13 *These cermets cre manurfo-c:'tur.ed by Kennametal, stated by the manufacturer, EXPERIMENTAL STUDIES WITH MOLTEN LITHIUM E. E. Hoffman T. Hikido? Reports have been received from NDA cdver.mg- | of corrosion by lithium at high temperatures. The. | . principal reports ;ssued during the year are listed the fiscal year 1957 subcontract studies? below: 1. Determination of Nztrogen in thbmm by N . Sax, N. Chu, R. W Miles, and R. H Miles, - NDA-38; 2. Purzfzcatzo-n- of thbwm by Vacuum Dtstzllatzou. _ by W. Arbiter and S. Lazerus,. NDA-39; - 3. Effect of Nitrogen on Corrosion by thb:ém by J. M. McKee, NDA-40 4., A Method for Determmmg the Solutzon Rate of ‘ Container Metals in Lithium by M. Kolodney and B. Minushkin, NDA-41; 5. Progress on Lithium Experzmental ‘Program During June, 1957, NDA Memo 75-18. A brief summary of the recent work is presented here to brmg up o date the - progress repofled ’ prevnously. An analyhcal procedure for the determmatlon s of nitrogen was developed- that is sahsfactory ‘_ for concentrations ‘as low ds 10 ppm. No com- pletely satisfactory method has yet been found, however, -for measuring low concentrations of . oxygen in lithim despite considerable effort. 2Orl _assignment from fhe USAF. 3This Nuclear Development Corporcfion of Amerlcu L {NDA) subcontract is coordinated at ORNL by E. " Hoffman and T Hikido, o 4T. Hikido and E. E, Hoffman, ANP Quar Prog Rep June 30, 1957, ORNL-2340, p 235. Inc., and the designofions and compositions given are those Solution rate studies of stainless steel speci- mens (principally type 304 stainless steel) have ‘continued. The base-line solution rates observed in this study at 1600°F (1.1 to 0.48 mg/in.2.hr) - were of the same order of magnitude as the rate of weight loss of specimens suspended in the legs of thermal-convection loops (0.83 mg/in.2:hr) in which lithium was circulated for ~approximately 100 hr at a hot-leg temperature of 1600°F, Additions of nitrogen or oxygen (opproximately 1000 ppm) were found to increase ~ "the solution rate by factors that varied from 1.5 to 3 in 4-hr tests. The test apparatus is particularly suitable for the screening of cor- ~ rosion-resistant - materials and for comparing the effects of additives and lithium impurities on the solution process. The main limitation of the apparatus is the lack of an inert container in which to perform the solution rate studies. It -is hoped, however, that molybdenum will be a satisfactory inert container. - Several - type . 316 stainless steel-lithium ~‘thermal-convection loop tests have been com- . pleted. . These tests were conducted with a hot- "leg temperature of 1600°F, a cold-leg temperature “at 1100°F, and a lithium velocity of 7 to 8 fpm. ~ The results of the tests have indicated that, - -in" general, the rate of loss of hot-leg weight - increased with .increased nitrogen ‘contamination . of the lithium.. The lowest rate of weight loss ~ “in the 11 loop tests was approximately 60 mg/in.?2 " in 100 hr, which represents quite heavy and rapid attack. - It was concluded . that further efforts to reduce this type of corrosion by still further elimination of nitrogen from the lithium were not justified. Only two of eleven loops com- pleted the scheduled 500-hr test without plugging. 219 ANP PROJECT PROGRESS REPORT The nine loops plugged in times that ranged from 31 to 457 hr. The weights of the crystals found in the loops varied from opproxlmutely 1to4g : Mass-transfer crystals were token from five locations in each of eight different loops for analysis. The compositions of the crystals varied considerably from sample to sample, but the compositions are fairly well represented by the over-all average composition shown in Table 3.4.2. As may be seen from the composition, the crystals were high in chromium, nickel, and manganese content in comparison with the base material. Present plcns call for continuation of this study with primary. emphasns on refractory metals and corrosion inhibitors. INERT ATMOSPHERE CHAMBER FOR ~ THERMAL-CONVECTION LOOP TESTING OF NONOXIDATION-RESISTANT MATERIALS E. E. Hoffman L. R. Trotter A stainless steel inert-gtmosphere chamber in which thermal-convection loop tests may be conducted was recently constructed. This device was built so that nonoxidatien-resistant materials, such as the refractory metals, could be tested in the unclad condition, since attempts to clad molybdenum and niocbium with oxidation-resistant alloys have been generally unsatisfactory, Almost invariably, the refractory metals have cracked either during the cladding operation or during the corrosion test, and therefore the correding medium (sodium, lithium, or fused salt) has .contacted both the refractory metal and the cladding metal during the test. A typical crack is shown in Fig. 3.4.7. Other methods of protecting niobium from oxidation are being studied, but none have been perfected to date. : - The essential features of the inert-atmosphere chamber are shown in Fig. 3.4.8. Thermal- convection loops may be operated in the chamber . either under o vacuum of less than 5 u or with an atmosphere of purified helium or argon. A stainless steel liner is incorporated in the chamber that serves as a thermal-radiation shield for the outer walls. The liner would also simplify cleaning of the chamber walls if a loop leaked. Several preliminary tests of the chamber have been conducted in order to determine the effec- " tiveness of the system in protecting several UNCLASSIFIED ¥-23348 Fig. 3.4.7. Soddle-Weld Section of a Type 310 Stain- less-Steel-Clad Niobium Thermal-Convection Loop After Exposure to Lithium for 500 hr. The stainless steel cladding of the saddle weld was stripped off after com- pletion of the test. The crack occurred during startup or operation of the loop. Table 3.4.2. Averages of Analyses of Mass-Transfer Crystals Found In Eight Lithium=Type 316 Stainless Steel Thermal-Convection Loops Compared with Anclysis of As-Received Stainless Steel Pipe Average Content (wt %) Cr Ni - Mn , Mo Fe Mass-transfer crystals* 55.9 22,5 18.0 33 . 12 As-received pipe 64.9 17.4 12.1 e 22 (June 14, 1957). 220 :‘”f*These data were compiled from a report by J. M. McKee, Effect of Nitrogen on Corrosion by Lithium, NDA-40 " DIFFUSION e PUMP— T | W AP TO ROUGHING 4 Il PUMP 2 2 I W 4o & PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR-DWG 22162b L 9, COOLANT OUTLET H.' I A~ COGLING COILS A R — IR N COOLANT INLET 5 10 5 20 INCHES Fig. 3.4.8. Inert-Atmosphere Chomber for Testing Thermal-Convection Loops Constructed of Nonoxidaotion- .Resistant Mu!*erlqls. refractory metals from oxidation - ot elevated temperatures. In the first test, specimens of. molybdenum ond "riiq:::bium ~were suspended in @ furnace inside the chamber- (argon atmosphere) and heated to 1700°F for 23 hr and 2100°F for 1 hr. The specimens showed weight gains of 0.1 and 0.5 mg/in.2, respectively. In a second, more comprehensive experiment, specimens of four refractory metals were heated to 1700°F for 24 hr in evacuated Inconel capsules and in the inert- atmosphere chamber filled with argon at a pressure . of 14,7 psi. Hardness measurements of these specimens are compared with the hardness of the as-received material in. Table 3.4.3. The hardness values indicate that in all cases the specimens tested in the chamber were softer than as-received material but not substantially -different from the specimens tested in the evacuated capsules. The weight-change datq, which were obtained only for the specimens tested in argon, indicate that the argon atmosphere was very pure. e , Since these results were encouraging, two attempts were made to operate unclad niobium thermal-convection - loops in the chamber. In both tests, leaks developed in the saddle-weld areas of the loops. These welds had been made in a dry box in an atmosphere of high-purity argon. 221 ANP PROJECT PROGRESS REPORT Table 3.4.3. Hardness of Various Refractory Metals As Received and After Being Heated for 24 hr at 1700°F in an Evacuated Inconel Capsule or in an Inert-Atmosphere Chamber Diamond Pyramid Hardness Weight Change (mg/in.z) (500-g load) Niobium As received 212 Heated in&ocuurfi ' . 152 Heated in argfin 160 0 Molybdenum As received . 215 Heated in vacuum 207 Heated in argon 210 -0.1 Titanium As received _ 246 Heated in vacuum 253 Heated in argon 227 +35.4 Zirconium As received 185 Heated in vacuum 19 Heated in argon 122 +2.5 Further tests of niobium will be attempted with loops fabricated by Fansteel Metallurgical Corporation. The saddle welds of these loops are fabricated according to the unique design shown in Fig. 3.4.9, and it is hoped they will not crack. FORCED-CIRCULATION LOOP TESTS OF SODIUM IN HASTELLOYS, INCONEL, AND STAINLESS STEEL J.H.DeVan R. S. Crouse Two forced-circulation loops, one fabricated of Inconel (loop 7426-26) and the other of Hastelloy B (loop 7642-3), were operated with sodium for 2000 hr in order to evaluate the comparative aemounts of mass transfer that would occur in these systems over relatively long periods of time.”The maximum sodium temperature in both loops was 1500°F, and a 300°F temperoture drop was maintained between hot- and cold-leg sections. The loops also contained by-pass cold traps to 222 UNCLASSIFIED ¥-23350 Fig. 3.4.9. Niobium Tubing Saddle Weld. ¥ —-in - Table 3.4.4. establish and maintain a low level of oxide contamination in the sodium. ‘ The crystalline deposits found in both loops after shutdown were weighed, and the amounts were practically the same, being 21 g for the Hastelloy B loop and 20 g for the Inconel loop. The value of 21 g for the Hastelloy B loop may be compared with a value of 17 g for a similar Hastelioy B .loop operated for - 1000 hr under similar conditions.> It is evident that mass transfer deposits form in a Haostelloy B system in which sodium is circulated at a substantially faster rate during an initial 1000-hr operating period than during a subsequent 1000-hr period. A similar situation exists in the case of Inconel systems, although the decrease in the rate of mass transfer during the second 1000-hr period is somewhat less for Inconel than for Hastelloy B. A plot showing the variation in mass transfer buildup as a function of time for both sysfems is presented in Fig. 3.4.10. Ll UNCLASSIFIED ORNL—LR-DWG 214094 25 R — 20 hsTELLO‘L? e . - /’ A" 3 s r WCON® Gi — 0 % / o 10 g |~ w / . . _ . = _____7_/ LOOPS OPERATED AT {500°F WITH 300°F A7 ___| 5 // COLD TRAPS MAINTAINED AT 30Q°F . 7 ' 0 0 500 1000 500 2000 AT OPERATION (hr) o Fig. 3.4.10. Mass Transfe_r_Byfldup as a Fukngfion_of_-;‘ 7 Time in Foi-ced-Clrcfildtion Loops '_Operut_ed with iSod‘lu'm. _; The resuhs of chemlcal analyses of the cold- leg deposits formed in the 2000-hr tests are given - In. both loops, the cold-leg deposits contained greater percentages of nickel - than were present in the base metals. ~The inner surfaces -of the ‘tubing comprising the hot legs of both loops revealed intergranular attack to a depth of 2 mils. The deposit in the cold-leg secfion of the Haste_lloy B loop reached a maximUm 5J. H. DeVan, R. S. Crouse, and D. A. Stoneburner, Alggs Quar. Prog. Rep. June 30, 1957, ORNL-2340, p conditions; PERIOD ENDING SEPTEMBER 30, 1957 Table 3.4.4 Analyses of Cold-Leg Deposits from Hastelloy B and Incone! Forced-Circulation Loops “Operated for 2000 hr with Sodium Elements Present in Cold-Leg Loop Moferi-al Deposits (wt %) Ni Cr Fe Mo ~ Hastelloy B 97.4 0.53 0.50 Inconel 85.6 12.8 0.70 thickness of 24 mils, whereas the deposit in the Inconel loop had a maximum thickness of 30 mils. A Hastelloy W forced-circulation loop, 7642-4, which was also operated with sodium, was ex- amined following 1000 hr of operation at a hot- leg temperature of 1500°F. The mass transfer deposits found in this loop were approximately 17% greater by weight than the deposits found in the Hastelloy B loop operated under identical 5 namely, 20 g compared with 17 g. A metallographic examination revealed a spongy, heavily attacked area approximately 1.5 mils deep along the hot-leg surfaces exposed to sodium. = These corroded areas, which are illustrated in Fig. 3.4.11, seemed to be quite brittle, and in many cases had pulled away from the metal underneath them. The cold-leg deposits in this loop were found by analysis to contain 97% Ni and 3% Cr. SR E] |NC|:'IE5 " Fig. 3.4.11. Hot-Leg Surface of 'Ha.sféllo'y- W Forced- Circulation Loop Which Operated 1000 hr with Sodium at 1500°F. Etchant: H2Cr04-H Cl. 250X, Reduced 31%. 223 ANP PROJECT PROGRESS REPORT A type 347 stainless steel forced-circulation loop also completed 1000 hr of operation with sodium at @ maximum temperature of 1500°F and a temperature drop of 300°F. The loop (7426-27) also .contained an oxide cold trap, which was maintained at 300°F. Visual examination of this loop revealed only slight traces of mass transfer in the cooled portions of the loop. The deposits were similar in extent ond appearance to those which have been observed in types 316 and 304 stainless steel loops operated under similar canditions, and were much less in amount than the deposits found in Inconel loops operated under similar conditions. A typical cold-leg section from the type 347 stainless steel loop is shown in Fig. 3.4.12. The hot leg of the loop showed light surface pits and wvoid formation to a depth of 2 mils. N e g o T ¥ > v . ? - L‘ A P v » 0 o .- .‘ € ay L F.o* ¥ & A LT e . - v . . 57 L ,a. . L ) 0 - : » ‘:\k-:‘e Lrwoc v Wi . v gt il e @, o My P e E e R Fig. 3.4.12. Mass Transfer Deposits in Cold-Leg Section of Type 347 Stainless Steel Forced-Circulation Loop Operated with Sodium for 1000 hr ot ¢ Maximum Temperature of 1500°F, Etchant: glycerol regia. 750X. Reduced 32%. DIFFUSION OF NICKEL IN LIQUID LEAD J. L. Scott H. N. Leavenworth, Jr.% Studies of the effects of temperature and con- centration on the diffusion of nickel in liquid lead were continued as a part of an over-all fundamental investigation of the mass-transfer process. The initial work and the experimental method were described previously, ond the dif- 60n assignment from Pratt & Whitney Aircraft. 224 fusion coefficients for saturated solutions of various compositions were presented.” Additional studies have now been conducted in order to determine the effect of temperature on the diffusion coefficient at constant composition. A constant composition for a series of tests. was obtained by saturating the lead, which filled the capillary, with nickel at a given temperature, and then making the diffusion measurements at a higher temperature. Saturation temperatures of 363°C and 481°C were used for these tests. The results are shown in Fig. 3.4.13, which is a plot of the logarithm of the diffusion coefficient, D, vs the reciprocal of the absolute temperature. The curve for the saturation temperature of 481°C is slightly above that for the saturation tem- perature of 363°C, but the two curves are roughly parallel. In order to determine whether the spread in these curves was a result of random variation, a statistical anclysis of the data was made. These calculations showed that the curve for 481°C differed from the curve for saturated solutions at the 50% confidence level, but not at the 95% level. The data for the curve for 363°C were found to be significantly different from the satu- ration data at the 95% confidence level. Thus, it appears that the diffusion coefficient for nickel in lead is a function of composition, even though the saturation composition is less than 0.5 wt % 7J. L. Scott and H. W. Leavenworth, Jr., ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 236. 5 UNCLASSIFIED x10 ORNL—LR—DWG 25947 200 100 50 0O SATURATED AT TEMPERATURE OF RUN A SATURATED AT 481 °C 20 C SATURATED AT 363 °C == DIFFUSION OF NICKEL IN SATURATED LiQUID 10 =~==DIFFUSICN OF NICKEL iN UNSATURATED LIQUID SELF-DIFFUSION OF LEAD DIFFUSION COEFFICIENT, Otem™/sec) M L5 12 125 13 435 14 145 15 455 46 + x 10° oK) Fig. 3.4.13, Diffusion in Liquid Lead, at the highest temperature studied. It also appears that a change in composition affects the entropy of activation for diffusion but not the enthalpy of activation. These results are only tentative because the range of temperatures studied and the amount of data obtained at each temperature Jnsufficient for a complete analysis. Higher Tatures will be- attained through the use of quartz diffus cells, “which.. will _replace the less expensive pyrex vells used in these tests, At the same time, a secohd\method of chemical analysis for nickel in lead, namely, activation analysis, will be utilized in addition fo the polarographic method, so that results from in- dependent anolyhcol methods €an be compored PURIFICATION OF somum ' J. L. Scott H. N. Leovenworth Jr. The vacuum shll which was desngned for purifying small batches of sodlum for use in fundamental studies, has now - ‘been. fobncated and installed. A picture of the - stlll ‘before the insulation was installed is shown in Flg. 3.4.14, The basic components of the ‘still are ‘the “sodium pot, the receiver for the . dlshlled mefol ‘and @ storage tank for undlstllled meml “The pot and receiver are each 8 in. in dlometer ond about 12 in. high. Sodium is. introduced into the system and withdrawn through the tubes below the receiver. It is pumped from contomer to contomer with the use of puflfied orgon un&er pressure. An odvanfoge of this sysfem is thof sodlum can be transferred from the receiver to the pot “and redistilled any number of times without opening the system. The heat sources for distillation and for momtommg the sodium in the molten state are Calrod units, as shown, Heatmg tape will be utilized -to maintain the tronsfer tubes above the meltmg poinf of sodmm. Cooling for _condensation of - sodium vapors is- provided by “coils through wbich efhyiene glycol is pumped. - These “coils “are |n3|de the tube between -the pot und the “receiver. - “The level of the sodium - in the recelver is determmed by spark-plug probes' L ~that pass through ‘the receiver flange. S - Safety is provuded through the use of the: regu- - lating system which .opens the volve shown in ~ Fig. 3.4.14. to introduce purlfled argon when the pressure ‘in the system rises to above 1 p.- This device, ‘which operates off a thermocouple gage, also turns off the power to the still and the vacuum components. The vacuum is provided by a Megavac-25 pump and a Consolidated Vacuum MC '500 dlffuslon pump. .seporoted from the stiil by the valve and a cold from supersoturoted ‘solutions. ocud is shghtly soluble in water and has a high temperature coeff:cnent of solubility. As the PERIOD ENDING SEPTEMBER 30, 1957 These components are trap, which is not shown. The pressure in the system is measured with a National Research type of ion gage. Preliminary tests of the system have shown that a vacuum of 5 x 10~5 4 is readily ‘.oftomoble Operation of this unit will be started .as _soon as the electrical wiring and instrumen- tation are completed. “MASS TRANSFER IN AQUEOUS SOLUTIONS J L. Scott P. Y. Jackson® Thermol-convechon loop tests in which water is “circulated through a cylinder of solid benzoic _ocud locufed in the hot zone have been initiated in the hope that this “system will simulate a hqmd-metol system for the analysis of nucleation Solid benzoic benzoic acid dlssolves, it s transferred by essentially free convection to the cold zone, which is_ maintained ot a temperature approxi- mately 20°C lower than the temperature of the hot zone. The process may be monitored by taking somples from the loop for titration and visually studying the nuclecmon ‘and growth of crystals in the cold zone. * Although insufficient - 'data for complete analysis have been obtained, some ‘trends have been ‘observed. Titration of samples of solution show ‘that the solid dissolves at a rate which is nearly constaont, since saturation is not approached in the hot zone. A considerable degree of supersaturation is attained in the cold . zone before crystallization ‘becomes visible. The condition of the cold-zone sutface, whether clean or dirty, smooth or etched, or straight or formed _in a_series of bulbs, affects the rate of formation ~of crystal nuclei ond the rate of attachment of these nuclei to the wall. When the cold zone is provided with a collar of,crysfols upon which _new melecules can deposit, the concentration of the acid in the water levels off, as shown in .run 3 of Fig. 3.4,15, with little tendency toward supersaturation. In the absence of crystals, the concentration rises to a value correspondmg to ‘some degree of supersaturation in the cold zone, ~ but after recrystallization has begun, the con- "centration falls rapidly to a value corresponding 8 Summer participant, St. Peter’s College, Jersey City, 225 ANP PROJECT PROGRESS REPORT RECEIVER FOR[ STILLED SODIUM} PRESSURE AND POWER™™ REGULATING SYSTEM | 226 Fig. 3.4.14. Sodium Still. UNCL ASSI Y2338 FIED 3 a to the saturation level, run 8 of Fig. 3.4.15. The hot-zone saturation curve was calculated, from the following equation, to show the solution rate at the hot-zone temperature of 60°C for the given surface area A and loop volume V: (- where C = C (1 - ~AT/V) C = concentration at time T, C, = saturation concentration, a = the solution rate constant, which is evaluated from the solution rate at times close to zero. ' Further tests will be made to determine the effects of changes in the hot-zone temperature, cold-zone geometry, and the surface conditions on the mass- transfer process. PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR-DWG 25948 T g9 SATURATION = P CONCENTRATION « 8 12 g/liter AT 60°C p g7 o PLUGGED] z J [ —FIRST CRYSTALS VISIBLE Q6 g 7 '%fi: PLUGGED | | o5 £ e =] “SATURATION CONCENTRATION _| 2 / = CORRESPONDING TO MIXED MEAN S ) COLD ZONE TEMPERATURE OF 40°C w4 T 1 @ / \SATURATION CONGENTRATION CORRESPONDING TO 5 3 / CONDENSER WALL TEMPERATURE OF ~30°C____| g, | g ® COMPUTED SATURATION CURVE FOR HOT ZONE | £ ]_© NO CRYSTALS IN COLD ZONE AT BEGINNING OF TEST-RUN 8 _| & A EXCESS OF CRYSTALS IN COLD ZONE AT BEGINNING OF TEST- g o RUN 3 R D 0 50 100 150 200 250 TIME (min) Fig. 3.4.15. Concentration vs Time Curve for a Benzoic Acid-Water System in a Thermal-Convection Loop. 227 ANP PROJECT PROGRESS REPORT < . : T YTTRIUM METAL PRODUCTION T. Hikido! R. E. McDonald? Jo Ho CQObS Experimental studies of the production of high- purity ytirium metal were continued. In the process being developed, lithium is used to reduce a mix- ture of YF ;-MgF ,-LiF to yield an yttrium-magnesium alloy. The fluoride mixtures used are prepared by the Chemistry Division with the use of techniques developed for fluoride-fuel processing (see Chap. 2.4, **Production of Purified Fluoride Mixtures''). The results of the experiments completed thus far are summarized in Table 3.5.1. The low yields obtained in runs L-2 and L-3 have been attributed to segregation in the crucible; that is, the lithium is separated from the YF ,-MgF ,-L.iF layer by the LiF slag, which has a density inter- mediate between the densities of the two reactants, In run L4, the reaction retort was tilted at a 45-deg angle to increase the areas of the interface layers and also to promote thermal-convection mixing. The yield was increased from 67 to 83% by this procedure. 10n assignment from USAF. 20n assignment from Pratt & Whitney Aircraft. 3.5. MATERIALS FABRICATION RESEARCH The lithium was added in the first three runs in the form of sticks cut into short lengths. In run L4, the lithium was transferred into the reaction retort in the molten state by differential argon-gas pressure. The molten lithium was filtered through a porous stainless steel filter in this transfer process. Experiments have been conducted on vacuum heat treatment of the ytiriumemagnesium alloy to distill off the magnesium and excess lithium. A 20-g sample of the yttrium sponge produced in this manner from run L-1 was arc melted with the use of a tungsten electrode under an argon atmosphere. The oxygen content of the arc-melted button was determined to be 1200 to 1300 ppm by a. vacuum- - fusion analysis performed by the Analytical Chemistry Division, A 30-g button from run L-2 has been arc melted and is now being analyzed. HYDROGEN DIFFUSION THROUGH METALS R. E. McDonald T. Hikido J. H. Coobs Equipment with which to measure the rate of diffusion of hydrogen through potential cladding materials for hydride moderators is being built, The apparatus is shown schematically in Fig. 3.5.1. The test specimen for the basic diffusion measure- ments will be a heavy walled tube of the material Table 3.5.1. Summary of Results cf Ytirium Reduction Experiments Quantity Yield of Y-Mg Alloy Run of Lithium (9) Yield Ne. Charge Reductant Added (% of theoretical) Theoretical Actual (9) L-1 591 g of YF3, 151 . 450 375 83 (10% excess) 232 g of MgF,, 174 g of LiF L-2 1000 g of YF3-M9F2- 137 411 246 60 LiF mixture (53.9- (10% excess) 21.1-25.0 wt %) L-3 200 g of mixture used 290 822 551 67 in run L~2 (15% excess) L-4 1000 g of mixture used 150 411 340 : 83 in run L=2 (15% excess) 228 4 i o FURNACE MUFFLE WATER QUT WATER JACKET HOT ZONE PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL-LR-DWG 25962 DIFFUSION WINDOW { THINNED TO 0.010in.} QUTER GAS TUBE TEST SPECIMEN 2-THERMOCOUPLE WELLS ~ INSULATOR /WATER JACKET END PLUG SF = ~—=-— WATER IN PALLADIUM VALVE .~ WITH HEATER FURNACE CONTROLLER \, - ~ailf— "O'LRING - SEALS THERMAL-CONDUCTIVITY CELL ARGON FLOWMETER SCALE ~—mg OF H, FURNACE PROGRAMMER CONSTANT-PRESSURE § - ‘ HYDROGEN REGULATOR \ e ARGON + HYDROGEN GAS " BROWN RECORDER Fig. 3.5.1. Diagram of Apparatus for Measuring the Rate of Hydrogen Diffusion Through Metals, being studied, with a section reduced in thickness. This thin **diffusion-window’’ section will be held in the hot zone of the tube furnace. Hydrogen gas will be maintained at a constant pressure inside the tube, while the outside of the tube will be purged continuously with high-purity argon gas. The purge ‘gas will be passed through a' thermal con- duetivity ‘cell which will detect changes in con- ductivity of the gas as a result of small additions = The factor of 10 dif- ference between the thermal conductivity of hydrogen (0.416 x 10~3 cal/cm+sec:°C at 0°C) and that of of hydrogen to the argon. argon (0.039 x 10~3 at 0°C) wili make possible the detection of quite small quantities of hydrogen. The impulse from the cell will be transmitted to a strip-type Brown recorder with a scale marked to present the hydrogen content. “The same eqmpment WI" be used to measure fhe_ o rate of loss of hydrogen from clad hydrlded-metalJ assemblies. It will also be useful for monitoring . hydrogen losses during thermal cycling and hydrogen migration tests, VOLATILITY OF BeO IN STEAM AT MODERATOR TEMPERATURES A. G, Tharp The major reported work on the volatility of BeO in steom is found in publications by Grossweiner and Seifert® and by Elliott.# The paper by Elliott, which was a doctorate thesis, reviews all the work -on volatile oxides in water vapor. " As is usual in these studies it is very difficult to “prove ‘an identification of the volatile molecule. The best summation of the present data indicates that the major reaction is -~ xBeO(s) + HZO(g)——> xBeO *H,0(g) There is evidence that one or more volah le species ~exist which are dependent on oxygen pressure. If 3L. Grossweiner and R. L. Selferf, ] Am. Cbem. Soc. 74, 2701 (1952) and The Reaction of Beryllium Oxide - with Water Vapor, AECU-1573 (July 30, 1951). 4. R. B. Elliott, Gaseous Hydrated Oxides, Hy- droxides, and Other Hydrated Molecules, UCRL-1831 (June 1952). 229 ANP PROJECT PROGRESS REPORT the reaction given above is assumed to be essen- tially the correct one, the data summarized in Table 3.5.2 are applicable. The dota of Table 3.5.2 are for equnhbrlum con- ditions or as near to equilibrium conditions as can be attained in an experiment of this nature. The samples of BeO used were in powder form, and Grossweinerand Seifert used water vapor in helium, Elliott used H,0.0, and H,0-H,. It should be pointed out that inactual practice it is not possible to obtain a greater volatility than that indicated by the equilibrium constant; that is, no concentration or pressure can exceed the equilibrium. . The dynamic conditions of the BeO(s) + H,O(g) re- action would determine the amount of BeO carried by the steam. Another important factor would be the nature of the BeQ, that is, density, crystal size, external surface area, etc. Another important factor could be that another molecular species of hydrated BeO would become the controlling species under different experimental conditions. The available data, which are probably quite accurate, indicate that under some conditions it would be entirely possible to corrode and transport large quontities of BeO in steam in relatively short times. FABRICATICON OF HIGH-DENSITY BeO R. L. Hamner It has been found in the process of fabricating BeO by hot pressing that the “Luckey S.P.”’- grade oxide produced by Brush Beryllium Company is equivalent to the ‘‘flucrescent’’ grade when densities of 95% or greater are desired. Such densities -are obtained by pressing at 1900 to 1950°C with a pressure of 2000 .to 2500 psi. A quantity of high-purity BeSO, has been received for experimental attempts to lower the temperature of hot pressing and to obtain high-density material by cold pressing and/or extrusion followed by sintering ot about 1500°C. This experimental approach is based on the production at Battelie Memorial Institute® of a highly sinterable material by carefully controlling the calcination temper- ature to obtain an extremely fine oxide powder. The hot-pressed BeQ specimens described below ~ were fabricated for physical tests: 1. five BeO cylinders, 2 x 2 x 1/2 in. ID; density, 97%; two cylinders damaged in machining; 2. four BeO blocks, 2 x 4 x 1 in., with four J '4-in.- dia holes running Iongltudlnally, one BeO block, 1 x 1 x 6% in.; density, 80%, one BeO block, 1 x ‘?’ x 6% in.; density, 90%; one BeO block, 1 x ‘/ X 67, m., density, 96%, one BeQ cylinder, 2/2 in. OD, / in. 1D, 2/2 in, long; density, 80%; 7. one BeO-UO, (50-50 mole %) block, /16 X /16 9in.; densny, 98%. oA w S.S‘ummary of Progress Report on an Invéstzgatzon of the Manufacture of High Density Beryllium Oxide. f;g;ls) by the Brush Beryllium Co., PWAC-167 {April 15, Table 3.5.2. Data on the Reaction of BeO with Steam Ratio of BeO-H20 {9) Flow Rate Pressure to Temperature H,0 BeO (liters/min) Steam Hydrogen Oxygen Pressure of H,0 (g) (°C) Passed Collected at Room Pressure Pressure Pressure —— = (moles) (moles) Temperature (atm) (atm) (atm) Grossweiner P Y . EHiott and Seifert x 10—4 x 1073 x 103 1270 1.95 0.3 1.3 0.87 0.13 1.0 6.9 1285 1.83 0.84 1.7 0.74 0.26 8.4 7.6 1310 3.44 1.52 1.5 0.95 0.05 4.4 9.8 1310 2.27 1.12 19 0.91 0.09 4.9 9.8 1310 3.9 3.40 1.7 0.74 0.26 8.7 9.8 1254 3.88 1.56 1.4 0.88 0.12 4.0 5.9 230 P The physical properties to be investigated include modulus of rupture, elastic modulus, internal friction, thermal expansion, thermal conductivity, total emissivity, and erosion at temperatures up to 1650°C. In addition, UO o-BeO diffusion studies are to be made. BeO THERMOCOUPLE INSERTS AND IRRADIATION TEST SPECIMENS R. L. Hamner Special BeO thermocouple inserts and shells for determining volume temperature fluctuations in the ART haoif-scale core model are being fabricated. Hot-pressed BeO blocks are to be ground to cyl- inders /8 }/2 in. for the 12 inserts and to cylinders '4 X l/2 in. for the accompcmymg shells, Holes will then be made in the ]/4 X / -in, cylinders by the activation technique to form the shells to accommodate the inserts, and parallel grooves 180-deg apart will be made in the same manner along the periphery of the inserts for holding the thermo- couple wires. For good thermal contact the base of the shell and the mating surface of the insert are to be platinized. Twenty-seven l-in.-long high-density-BeQ cyl- indrical specimens with diameters of from 0.44 to 0.94 in. are being fabricated for irradiation studies. These specimens are to be canned in Inconel and irradiated in the ETR. THERMAL STRESS RESISTANCE OF CERAMIC CYLINDERS R. A. Potter An apparatus is being designed for the deter- mination of the heat throughput required to cause initial failure in ceramic cylinders. This in-. formation will aid in determining the thermal stress resistance of various ceramic materials. * It will - also provide a means of sfudymg the effects of varying the physical properties of the muterlals on their. ability. to withstand thermol shocks. Con- sideration is being given to various methods of heating hollow cylmders mternolly and coolmg. them exiernaliy. FABRICATlON OF METAL HYDRIDES ' R A. Poh‘er | ' Nlne groups of dense zirconium hydrlde samples have been prepared in the hydriding apparatus.® SR. A. Potter, R. E. McDonald, and T. Hikido, ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274, p 193. PERIOD ENDING SEPTEMBER 30, 1957 The samples are rods approximately 3 in. in length and '/2. in. in diameter, and they contain various amounts of hydrogen, as shown in Table 3.5.3. These pieces were made as part of an effort to establish procedures for hydriding massive yttrium metal, Modifications have been made in the equipment, which is now ready for production operation as yttrium metal becomes available for hydriding. The procedure to be used is based on a process, developed at the General Electric Company, in which a mixture of hydrogen and helium is passed over the metal. Toble 3,5.3. Hydrogen Contents of Derise Zirconium Hydride Samples Sample 5 | Hydrogen Group ?mP e_ Content* No. Designation (% of total weight) | a 0.19 b 0.21 il a 0.36 b 0.41 it a 0.54 b 0.56 v a 0.83 b 0.83 Y a 0.86 b 0.92 vi a 0.95 b 0.95 Vi a 05 Vi a 1.10 AX - o 1.15 - b 1.15 *Determined on the basis of weight gain. 231 ANP PROJECT PROGRESS REPORT 3.6, METALLOGRAPHIC EXAMINATIONS OF ENGINEERING TEST COMPONENTS AFTER SERVICE ART PROTOTYPE TEST RADIATOR R. J. Gray J. E. Yan Cleve, Jr. A metallographic examination was made of the York Corp. ART prototype test radiator No. |, which operated a total of 870 hr under the following conditions: |sothermal operation at 1200°F 390 hr Operation with varicus temperature dif- 47 hr ferentials and a meximum NaK inlet temperature of 1200°F Operction with varicus temperature dif- 360 hr ferentials and a maximum NaK inlet temperature of 1500°F Operation at the design condition of a 73 hr NeaK inlet temperature of 1500°F and an outlet temperature of 1070°F Thermal cycles, total 9]/2 Slow cycles at a maximum NaK tem- 7 percture of 1200°F Slow cycles ot a maximum NaK tem- 2 perature of 1500°F Fast cycle (50°F/min) to a NaK tem- I/i_, perature of 1525°F The radiator failed upon reaching full power in the first controlled thermal cycle. The NaK inlet temperature had been increased from 1350°F to 1525°F at a rate of 50°F/min and the outlet tem- perature had remained constant at about 1070°F. After the radiator had been cleaned it was sectioned for examination, The section containing the tube that failed, which was located by pres- surizing the tube with wafer and observing the leak, is shown in Fig. 3.6.1. The arrow paints toward the tube that leaked, A close view of the tube that failed is shown in Fig. 3.6.2. The numbered arrows indicate the readily visible holes, which appear as spots because of internal lighting during photography. The metaliographic examination did not reveal evidence of complete penetration of the tube wall at either the circum- ferential or the longitudinal fissures. The some tube is shown in Fig. 3.6.3 rotated approximately 90 deg. The dark area in the tube is hole No. 1 in Fig. 3.6,2., The curvature of the tube as it enters the header is evident. The good condition of the tube may be attributed to the promptness with which the fire that occurred as a result of the leak was extinguished, Hole No. 1, which was the largest of the three holes, is shown in Fig. 3.6.4. The opening in the tube wall in this plane measures 0.030 in. More of the tube wall that appears at the extreme right edge of Fig. 3.6.4 may be seen in Fig. 3.6.5. Hole No. 2 is shown at a magnification of 100 in Fig. 3.6.6 and at a magnification of 500 in UNCLASSIFIED Y-13342 Figs 3.6.1. ART Prototype Test Radiator No. 1. (SeTfet with-cuptign) 232 & - PERIOD ENDING SEPTEMBER 30, 1957 Fige 3.6¢3. Tube Shown in Fig. 3.6.2 After Being fiofuted 90 deg. UNCLASSIFIED O Ye23343 UNCLASSIFIED 3341 233 ANP PROJECT PROGRESS REPORT Figs 3:6¢4. Hole No. 1 of the Tube That Feileds The opening in the tube wall is 0.030 In. Etchant: 10% oxalic acids 100X. INCHES 0.02 0.02 Figs 3.6:5. Tube Wall Shown at the Right Edge of Fig: 3.6:4. Note that tfie internal pressure in the tube produced an outer llp at the point of fallura_- 100X. 234 W Fige 346¢6. Hole Nos 2 of Tube That Failed. Etchant: 10% oxalic acids 100X. Fig. 3.6.7. As may be seen the shape of the metal bordering the hole is simitar to that of the metal bordering hole No. 1. Figure 3.6.7 shows that there was .no grain-boundary void formation, and there is no evidence of incipient failure, even very near the area that failed, The microstructures of the three tubes surroundmg the tube that foiled are shown in Figs. 3.6.8, 3.6.9, and 3.6.10. Again, there is no evidence of incipient failure in any respect. The inside diameter, outside diameter, and wall o thickness of the fube that failed were measured at various times ‘during polishing ot the specimens, - Variations in the ocutside diameter and in the wall thickness were noted in several instances, but all done by the fire. There was little or no variation - in - the inside diameter whlch would indicate the - sresence of a stress high enough to produce plastic flow in the tube, YORK CQ,RP.,RADIATOR NO, 16 R, J. Gray J. E. Yan Cleve, Jr. York Corp. radiator No. 16 failed with a NaK-to-air leak after 1438 hr of operation, including 1044 hr PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED Y-23354 Note that the shape of the metal is the same as in Figs 3.645, under thermal gradient conditions. The NaK inlet temperature during thermal-gradient operation was 1500°F or above for 541 hr, and the unit was thermally cycled 21 times. The fire damage was extensive, and therefore no attempt was made to evaluate the integrity of the brazed fin-to-tube joints. Of the 72 tubes attached to the NaK inlet header, 29 were burned through " and 5 more were found to be plugged. Four tubes contained partial plugs, and one was completely plugged. The plug was drilled out of the tube that was completely plugged, and the plug material was . found to be green in color, The partial plugs were removed by scraping the tube walls and were found ~to -be brown to black in color, Chemical analysis the variations could be dlrectly related to damage . showed the green muterlcl to contain 3.51 wt % ‘wranivm and 4.19 wt % zirconium, - The darker moterial - contained 7 wt % U, 47 wt % Ni, 7 wt % Cr,and -3 wt % Fe. The difference between the iwo plugs therefore was that the dark-colored naterial contained considerable mass-transfer leposits. The presence of uranium indicates that there was a leak of fuel into the NaK in the fuel- to-NaK heat exchanger included in the circuit in which this NaK-to-air radiator was operated. The 235 ANP PROJECT PROGRESS REPORT UNCLASSIFIED . Y-23353 Figs 3.6.7. Higher Magnification of Fig. 3.6.6. No evidence of incipient follbrg may be seen. Etchant: 10% oxalic acide 500X, ' - 'INCHES 1 1 o Q - T s P Gy, @Oy E - - . w Fige 3.6.8. Microstructure of Tube Adjacent to the Tube That Failed. No evidence of incipient failure was found. Etchant: 10% oxalic acids 200X, : 236 6 £ i i 1 i i 1 1 TR . PERIOD ENDING SEPTEMBER 30, 1957 CLASSIFIEOH [~ ¥.23339 - | . Fige 3.6.9. | Microstruclure of Tube Adlccentto the ':Tl;hfl V‘I_'hcf ‘Flc_:_lled. | 'No.evidcrice of inciplent failure was founds Etchant: _-l_(_).%_fq“'xo|ic acide 200X, . - 3 D Figs 3.76.10‘. Mierostructure of Tube Adiccenf to the Tube That Falleds No evidence of incipient failure was founds Etchant: 10% oxalic acids 200X. ' 237 ANP PROJECT PROGRESS REPORT examination of the heot exchanger is: descrlbed in the following: sectlon. PROCESS"ENGINEERING CORP. HEAT EXCHANGER NO. 3, TYPE SHE-7 R. J. Gray J. E. Van Cleve, Jr. A small heat exchanger of type SHE-7, which was fabricated by the Process Engineering Corp. and is designated No. 3, was ferminated after 1438 hr of operation with the fuel mixture NaF- ZrF UF ;. (56-39-5 mole %, fuel 70). The unit operated under thermal-gradient conditions for 1044 hr and was thermally cycled 21 times. Opera- tion of this unit was terminated because of failure of the radiator in the system (see preceding section of fhi'_s chapter). Thirty-one samples were removed from the tube ‘bundle and header areas and mounted for metallo- graphic examination to determine the depth of corrosion on both the fuel and the NoK sides. The fuel-side corrosion was found to range from a minimum of 0.001 in. to a maximum of complete penetration. The NaK-side corrosion was found to vary from general roughening of the surface to penetration to a maximum depth of 0.004 in. An area from the hot-header section is shown in Fig. 3.6.11, As may be seen, the voids completely penetrate the tube wall, but there is little or no general corrosion. Adjacent areas in which the same conditions exist are shown in Figs. 3.6.12 and 3.6.13. A metallic layer found on the fuel side of a tube from the hot-header area is shown in Fig. 3.6.14, Spectrographic analysis showed the metallic layer to be 6.3 wt % Cr, 8,5 wt % Fe, and 76 wt % Ni. This metallic layer was not similar to the gold-colored deposit found in the cool section of a previously examined heat exchanger. The gold-colcred deposit was found to have resulted from corrosion of the header region. The examination. of the heat exchanger did not show any open cracks through the tube walls. The presence of the uranium- and zirconium-bearing plugs in York radiator No. 16 proved, however, that a leck was present, Fuel probably leaked through several of the tubes which showed com- plete grain-boundary penetration, such as those shown in Figs. 3.6.11, 3.6.12, and 3.6.13, 238 FEE] BN 131N CLASSIFIED RN Y-23431 T 1 t CHES 1 - r: & S s o ol EEEEE B e > 2 & BEEE 00X 1 i f [ P’ ' : . P B . - T by g .022 § Cow s At ° ) .oa . i . G & . 2025 | [ T A . ) Fig. 3.6.11. Area from o Tube Near the Hot-Header Section of @ NaK-to-Fuel Heat Exchanger Operated at High Temperatures. Note complete grain-boundary penetration and the wide fissures in both surfaces, with little or no general corrosion. Etchant: 10% oxalic acid. 200X, Reduced 24.5%. (Lomfidemttot—with vhrory) , PROCESS ENGINEERING CORP. HEAT EXCHANGER NO, 2, TYPE SHE-7 R. J. Gray J. E. Van Cleve, Jr, Process Engineering Corp. heat exchanger No. 2, type SHE-7, which operated for a total time of 1845 hr and for 1645 hr with NaF-ZrF ,-UF , (50-46-4 mole %, fuel 30) was also exammed Thls small heat exchanger was operated under thermal- . "gradient conditions for 1456 hr and was thermally cycled 29 times. The fuel-inlet temperature was above 1600°F for 552 hr. When operation was terminated because of o failure of the economizer in the circulating cold trap that was operating in " o - EEE] Y 1 INCHES £ Fig. 3.6.12. Tube Adjacent to the Tube Shown in Fig. 3.6.11. Complete grain-boundary penetration, with little general corrosion, may be seen. oxalic acid. 200X. Reduced 24.5%. Etchant: 10% the NaK system, the fuel circuit would not dumpr"‘ and the removal of samples was dlfflculf : Thirty-four samples were removed from vthe tube bundle and header areas.and mounted for metallo- graphic examination. - The fuel-side corrosion was found to range from a minimum of 0,001 in, to a maximum of complete: penefrahon. “The NaK-side corrosion varied- from ‘general ‘roughening of the ‘exposed - surface to a maximum depth of 0,010 in. The general corrosion and gram-boundclry voids found " in “the ‘wall’ exposed to fuel ore shown in Fig. 3.6.15, - The general corrosion in Fig. 3.6.15 reaches a- depth of 0.005 in., -and beneath this ‘extends 0. 005 in, of gram-boundary voids, fThus“:” the depth - of corrosion on “the fuel . srde was - 0,010 in., or approximately two- fifths of the tube wall. PERIOD ENDING SEPTEMBER 30, 1957 Tube Adjacent to Tube Shown In Fig. Complete grain-boundary penetration, with "Etchant: 10% - Fig. 3,613, 3.6.12, “lttle general corrosion, may be seen. oxalic acid, 200X, The NaK-side of the tube -'wull'_shdwn in Fig. - 3.6.‘]_5—is shown in Fig. 3.6,16, The corrosion, “which ¢an, in this i'insicnée,"be' 'atfiibu'ted to the 'NaK, extends to a depth of 0,004 in. This depth of attack added to ‘that by the fuel yieids 0,014 in. of corrosion attack; that is, slightly more than one-half the tube wall was attacked. 239 ANP PROJECT PROGRESS REPORT LIL LB o n (=] Bl 1] 1 ISCP( m‘u’pfim‘,o B } UNCLASSIFIED . Y-22558 Fig. 3.6.15. Fuel Side of Tube from Hot-Hecder Reglon of Process Engineering Corp. NaKeto-Fuel Heat Ex- changer No. 2, SHE=7. General corrosion may be seen to a depth of 0.005 in. plus 0.005 in. of grain-boundary voids. Estchant: 10% oxalic acide 200X. 240 ") O PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED - Y-22557 ' { ; i )\ | ; :’ f Sy ' 9 ¢ Figs 3.6.16. NaK-Side Surface of Tfii:e‘f:oivn in Figs 3.6:15. Etchant: 10% oxalic acid. 200X. Grain-boundary * voids that extended through a tube that was in the hot-header region are shown in Fig. 3.6.17, The wvoids progressed to such an extent that they cannot be definitely attributed to either fuel or NaK, even though the general corrosion is 0,008 in, in depth, The usual gold- colored metal deposit was found along the surface exposed to the fuel in the cool section. INCONEL CENTRIFUGAL PUMP THAT CIRCULATED NaK : J. H. Dchm R.S. Crouse - An endurance test of an Inconel centnfugal pump that circulated NoK was _terminated following upproxumately 4400 hr of operation primarily at 1200°F, -For some time before the ¢lose .of the - test, there were indications that lubncatlng ol had been leaking past the lower shafi sea| and into the NaK stream. ‘Upon shutdown, extensive earbon deposits were found lmmedm:ely below the lower - pump seal; “and all wetted parts of the pump were dark ond tornished. Metallographic examination was therefore made of some of the wetted parts to determine the nature of metallurgical changes which might have occurred during the test, The - examination revealed a heavy carbide precipitate " to a depth of approximately 5 mils along all Inconel surfaces exposed to NaK. The extent of carburi- zation, which is illustrated in Fig. 3.6.18, was not sufficient to have measurably affected the me- chanical properties of heavier Inconel sections, although it was necessary to replace all thinner - sections, such as bellows and diaphragms. | WELDS OF TEST COMPONENTS - G, M. Slaughter | Thermul-Convechon Loop Welds - P, Patriarca A metallographlc ‘study of the fused-salt corro- snon resistance of saddle welds removed from a Iclrge number of thermol-convecflon loops has been madé. Brief operating histories of these loops are - presented in Table 3.6,1.. No evidence of inter- granulorf corrosion of improperly oriented grains of the type found in the welds of Black, Sivalls & Bryson heot exchanger No. 1 (type - IHE- 8) was - 'observed E. A. Franco-Ferreira, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 247. 241 ANP PROJECT PROGRESS REPORT UNCLASSIFIED L Y-23494 . - . L W - . T gy . gt L . v . . el g - Fig. 3.6.17. °Tube Wall from Hot-Header Region. Complete penetration of the tube wall was found. Etchant: 10% oxalic acid. 200X. Forced-Circulation Loop Welds Segments of an electric-resistance-heater lug removed from a nickel-molybdenum alloy forced- circulation loop were obtained and examined metal- lographicatly for weld-metal corrosion. The alloy (17% Mo—7% Fe—bal Ni) tubing had been welded to a Hastelloy B lug adapter with Hastelloy W filler wire. The loop operated for 1000 hr with fuel 107. 242 RN UNCLASSIFIED - Fig. 3.6.18, Curb-uriiufion Which Occurred on Inconel Surfaces Exposed to NaK in Centrifugal Pfimp Which Circulated NaK for About 4400 hr at Approximately 1200°F. The NaK was contaminated with oil following a leak in the pump seal. Etchant: aqua regia. 250X. Welds from both the upstream and downstream ends of the loop were mounted and examined. The samples were sectioned along the longitudinal axis of the tubing to permit a composite evaluation of the joints. No measurable corrosion of the welds was evident in any of the samples examined, Heat Exchan ger Welds The results of metallogrophic examination of fused-salt corrosion of welds and brazes,and the fused salt and NaK corrosion of tubes in the Black, Sivalls & Bryson heat exchangers types IHE-8 were reported previously.2 Tube-to-header welds from both the NaK inlet and outlet header have now been examined to determine the extent of NaK corrosion. Attack to a depth of approximately 0.003 in. was prevalent in the NaK inlet header, and in some cases attack to a depth of 0.005 in. R. J. Gray, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 244. & el o was observed. Typical weld-metal corrosion is shown in Fig. 3.6.19, and the extent of the corro- sion in an adjocent tube wall is shown in Fig. 3.6.20. Similar specimens taken from the NaK cut- PERIOD ENDING SEPTEMBER 30, 1957 let header are shown in Figs. 3.6.21 and 3.6.22. The extent of corrosion is, of course, much less in the cooler regions. The amount of mass transfer was not determined in this study. Table 3.6.1. bpémflng Histories of Thermal-Convection Loops Examined for Weld-Metal Corrosion oo Loop Moteri Fusedaltt Operoting Temperatre T 765 Inconel | 30 1500 1500 779 Inconel 30 1400-1600 2000 ' (cycle) 876 Inconel + Inconel 30 1500 500 cast insert 1010 Inconel 30 1500 1000 792 lnconel | 130 + 0.2% ZiH, 1500 500 788 Inconel 304 0.5%ZrH, 1500 500 789 Inconel S 30 + 0.6% ZrH, - 1500 500 790 Inconel - 30 + 0.7% ZrH, 1500 500 791 Inconel 30 + 0.8% ZrH, 1500 500 802 Inconel : * 30 under N, atm 1500 500 786 Inconel | 730 + 2.96% addition 1500 1500 785 Inconel 30 + 2.23% addition 1500 500 830 Inconel 4 1500 500 831 Inconel A 1500 500 734 Inconel 82 1500 500 735 Inconel 82 1500 1500 841 Inconel 82 1500 500 842 Inconel - 82 1500 500 843 lnconel o9 1500 500 844 lInconel o 1500 500 847 lnconel . 93 1500 500 845 lnconel 94 1500 500 846 ' lnconel o4 1500 500 839 Inconel 95 1500 500 840 Inconet 95 | 1500 500 798 Inconel 107 + 0.94% addition 1500 500 243 ANP PROJECT PROGRESS REPORT Table 3.6.1 {continued) * Fused-Salt* Time LoopNo. Loop Material Crotaned R 797 inconel 107 + 1.5%.alddition 1500 500 793 lnconel 107 + 1.83% addition 1500 500 799 Inconel 107 + 1.83% addition 1500 - 500 - 796 Inconel 107 + 3.22% addition 1500 500 806 Monel 107 1500 500 808 Monel 107 1500 1000 N Hastelloy B 30 1500 1500 866 Hastelloy B 30 1625 1000 814 Hastelloy B 107 1500 500 816 Hastelloy B 107 1500 2000 826 Hastelloy B 107 + 2% addition 1500 1000 873 Hastelloy W 30 1500 1000 857 Hasteloy X 30 1500 1000 1014 76% Ni-24% Mo 30 1500 831 1015 76% Ni-24% Mo 30 1500 1000 1004 74% Ni-26% Mo 30 1500 791 1007 75% Ni-20% Mo—5% Cr 30 1500 1000 1008 75% Ni-20% Mo 5% Cr 30 1500 240 1006 77% Ni—20% Mo~3% Cr 30 1500 1000 1013 85% Ni-15% Mo 30 1500 1000 ®*Fused-salt composition: No. 30 - Na F--ZrF[UF4 (50-46-4 mole %). No. 41 - NaF-ZrF ,-UF , (63-25-12 mole %). No. 82 - NaF-LiF-ZrF4-UF4 (20-55-21-4 mole %). No. 91 - NaF-LiF-ZrF‘-UFA (53-35-8-4 mole %). No. 93 — LiF-ZrF ,-UF , (50-46-4 mole %). No. 94 ~ KF-ZrF",-UF4 (50-46-4 mole %). No. 95 — Rl:F-ZrF4-UI=4 (50-45-4 mole %). No. 107 = NaF-KF-LiF-U F 4 (11.2-41-45.3-2.5 mole %). 244 PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED mcm;s,’ 0.02 . jc.03 Fig. 3.6.19. - NeK. Corroslon in Tube-tc-Heuder Weld of NoK Inlet Header of Black, Sivalls & Bryson NaK-to-Fuyel * Heat Exchcnger Type lHE-& ‘Etchants eleciro!yfic oxclic ccid. IOOX- Confidential with-coptiony o [ 23] e x O z 0.02 » ./ Figs 3.6.20. NcK Corrosion in Tube Wall Adjacent to Weld Shown in Fig. 3.6.17. 100X« Etchant: electrolytic oxalic acid. | 245 ANP PROJECT PROGRESS REPORT S i Py Figs 3.6:21s NaK Corrosion in TubestorHeader Weld of NaK Outlet Header of Bleck, Sivalls & Bryson NaKeto=Fuel Heat Exchanger Type IHE-8. Etchant: electrolytic oxalic acids 100X, (Cenfidentietwithcaption) UNCLASSIFIED " = 10§ T T = 0.02 ’ Tl ; C : v . e aner ok wi oL St R S A S AT lo.o3 | ! 1 * e © o , e *5si : A T e ’ S e e S DRI s AT R N .&-.?s‘\i\f.qfi_‘? A Fige 3.6:22. NaK Corrosion in Tube Woll Adjacent to Weld Shown in Fig. 3.6.19. Etchant: electrolytic oxalic acids 100X, 246 PERIOD ENDING SEPTEMBER 30, 1957 3.7. NONDESTRUCTIVE TESTING R. B. Oliver EDDY-CURRENT MEASUREMENTS OF METAL THICKNESS ' R. B. Oliver | J. W. Allen R. A. Nance The Metal Identification Meter (MIM-1) developed' for the sorting of Inconel, austenitic stainless steels, and the common Hastelloys was success- fully adapted to the thickness measurement of thin sections of these alloys. A simple circuit change allows adjustment of the scale range so that finite thicknesses of these alloys, up to 0040-|n., can be measured with a maximum error of £0,001 in. The instrument was used to monitor the wall thick- ness of a type 347 stainless steel cylinder while - the wall was being lathe-turned from the original ‘/4 in. to 0, 020 in. The stainless steel cylinder, which was 3/2 ft long and 6 ‘in. in diameter, was fitted onto a wooden mandrel. - Both the Metal Identification Meter and a resonance ultrasound thickness gage showed the finished wall thickness to vary from 0.013 to 0,40 in. The readings made with the two instruments agreed to within 0.001 in. at all times. . The Metal ldentification Meter and the resonance ultrasound instrument are both ex- - cellent instruments for measuring metal thickness when only one side of the material is accessible. When applied to metal sections thinner than 0.020 in., the eddy- t method embodied in . " o eddy-current methor emboclec. in:. "“make a similar.analysis of the impedance of a probe “coil ‘in proximity with a flat metal plate have not, ~ in the past, been successful, but by making several - the Metal ldenhflcahon Meter is more accurate than the resonance ultrasound method, and, when applied to section flucknesses greater than 0,060 or. 0 070 - ine, the resonance ultrasound method is. ‘has been completed. -stable than its predecessor,? more accurate. The meter is compact, portable, and simple to operate, and it requires no coupling between the probe and the work surface. The values of several components in the circuitry of the eddy-current instrument can be changed so that measurements can be made of thicknesses of metals having conductivities beyond the range of detection of the present model of the Metal ldentification Meter. ‘A new instrument, illustrated in Fig. 3.7.1, for the measurement of sheet and cladding thicknesses This instrument is more being nearly free from drift, and is more sensitive and versatile in that it will accommodate a wider range of probe- coil types and of test frequencies., The new in- strument is being used in studies of the optimum test parameters for measuring various combinations of cladding and core materials. MATHEMATICAL ANALYSIS OF EDDY-CURRENT PROBE COILS J. W, Allen M. O. Chester® Numerous mathematical studies of the impedance of a coil surrounding a cylindrical red or a tubular object have been made that form a rational basis for the selection of test parameters. Attempts to drastic '_assumptions regqrding- _the field of the 2. w, Allen and R. A. Nance, ANP Quar. Prog. Rep. R. B. Oliver, J. W. Allen, dnd R. A. Nance, ANP March 31, 1957, ORNL-2274, p 227. - Quar. -Prog,: Rep. ]ur_:e’i _30,.' 1957, ‘_ORNL-2340,'-_p '25§.’ o 3St.lmmer employee from Comell University. 247 ANP PROJECT PROGRESS REPORT L W T A IS TV L) . ) L R T ETY . - EDDY-GURRENY = B 1103 LYY Marx & B Y N D.RNL METALLURGY DIVIBION Fig. 3.7.1. Eddy-Current Thickness Gage, Mark 1l, Serial 1. 248 UNCLASSIFIED - Y-:23439 " probe coil a semi-empirical mathematical expres- sion has been developed for a single-turn coil near an infinitely thick metal plate. The equation PERIOD ENDING SEPTEMBER 30, 1957 The curves of Fig. 3.7.2 were plotted from Eq. 1 by using the terms k222 and v as separate param- eters even though they are related. |t should R = a-c res:stanceof conl © = 2r times the frequency of the coul L = inductance of the co:l in the presence of B the metal oL = the reactance of the coil in the presence of the metal, ' wL, = the reactance of the coil in air, j= J—-_]-l‘ | o k2 = Jopryo, po = 4mx 10~ -7, = permeoblhty of nonrnagnehc’ materials in mks units, . o = the mhos/meter, a = the radws of fhe equwalent smgie-turn - - eoil, v.= I/a: , L o "1 = the “lift-off'* or spoc:ng between equw- - alent single-turn co:l und fhe metul rsurfc:ce. : : : | The term K is determmed by the relohonshlp of‘ | the followmg equahon' ' @ fo(Kd>.=[. A2 ” r v+ 4 in which Jq is the Bessel function of zerd‘erder. and of the first kind, conductivity of the l'\:léfql,"ini,n;». is: also be noted that these curves are universal, R+jolL 1 | L (1) L "7 1+ — =( @iy (ka)2 : /2 | | (Ka)? o 1 "1 va [ g 2 v+ 2+ 1172 S ll=—=+— = [1+— + In | L | _4 .vz_‘ o v2 v L2\2 | 1/2 1 | 1 1o . 2 3 + teom g — - q - " . [ wka)? ]2 | 1 1 — - —{—+ In — (Ka)? 2 2 o - J ~in whlch ~since . the parameters are nondimensional and may -be calculated for any given test condition. For practical testing purposes, the equivalent “‘single- turn radii’’ and the “‘single-turn lift-off’’ values ‘may be calculated by a graphical technique on the curves shown in Fig, 3.7.2. In general, very good agreement has been obtained between a con- - siderable amount of experimental data and the curves. More experimental data will be obtained and compared in order to increase the confidence - level in applying these calculations. The .problems involved in obtacining a mathe- matical expression for the impedance of probe - coils in the presence of plates haying finite thick- ness and in the presence of a cladding-core couple _are much more difficuit and more complex than those presented by this case of an infinitely thick - plate. There is some possibility, however, that ~ these cases can also be solved. The results of - the theoretical probe-coil studies should be vuluoble in the future application of eddy-current _probe coils to the measurement of metal thickness r_cnd of the thickness of cladding over a core, INSPECTlON OF TUB!NG R. W, McClung - 'R. B. Oliver " Both the encircling-coil eddy-current method and the immersed-ultrasound pulse-echo method were used to inspect approximately 18,500 ft of Inconel tubing during the quarter, The tubing ranged in 249 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 23097 1.00 60 19 09 | 2 33 kY Juper 0.90 =4¢ X0 (PERMEABILITY) _] . #o 47 OF NONMAGNETIC MATERIAL) w= ANGULAR APPLIED 7 FREQUENCY - 080 hO J ' o= CONDUCTIVITY {mhos/meter] | g Vo 25 2= "EQUIVALENT" COIL RADIUS 3 O | {meters) o O - _ z £ 14.2 4 070 Y A= "EQUIVALENT“COIL LIFT- —] ? 0?‘0 190 OFF (meters) 5 28.5 = % o« o, Vara 060 & 105.2-—1k2 02| o | 364.5 ¥ 050 15 0.40 ‘ o 0. 0.2 03 04 05 06 RESISTIVE COMPONENT Fig. 3.7.2. Calcu’lofgd Reldiive Impedance of a Probe Coil Above a Conductor with Infinite Thickness. size from ;16 in. OD with a 0.025-in. wall to '/2 in, OD with 0.065-in. wall. The rejection rates were dependent on both the quality required and on the minimum usable length, The average re- jection rate was about 5%, with variations for individual batches running from as low as 2% to more than 10%. Factors that confribute to the lower average rejection rate in comparison with the rates previously reported were re-polishing and re-inspection of ultrasonic rejects and the removal of magnetic particles by a phosphoric acid freatment, INSPECTION OF PIPE J. K. White R. W. McClung Approximately 650 ft of Inconel pipe was in- spected by the immersed ultrasound method during the quarter with less than 1% rejection. About 500 ft of the pipe, both 2}, and 3% in. sched 40, was inspected with resonance ultrasound methods in order to determine the wall thickness. For a specified tolerance of +5% of the nominal wall thickness, the rejection rate was approximately 90%; but for a standard tolerance of +10%, the rejection rate decreased to approximately 10% of " the total quantity of pipe. 250 . PLATE INSPECTION R. W. McClung The immersed ultrasound method was used for the inspection of more than 625 ft2 of CX-900 Inconel plate in thicknesses ranging from % to 1 in. The rejection rate was less than 1%. Since the inspected plate was, in general, larger than the required size and rejected areas could be left in the trim, it is difficult to accurately estimate the true rejection as a result of defects, CORE SHELL INSPECTION R. W. McClung Immersed ultrasound methods were developed for the inspection of the Inconel reactor core shells. The core shells are inspected both for the presence of laminations and for the presence of radial cracks. For the detection of laminations, 5 mega- cycle ultrasound is directed perpendicularly to the surface and along the radius of a shell, and *‘ringing’’ within the thin wall is established, Any damping of the ‘’ringing’’ pattern is indicative either of a change of wall thickness or of the presence of a lamination, For the detection of cracks, the 5 megacycle ultrasound is introduced into the wall of the shell with an incident angle of approximately 17 deg; any crack lying approxi- mately perpendicular to the path of the ultrasound as it travels through the metal annulus will reflect the signal and indicate the presence of that crack. Areas on a core shell that indicate damping of the “’ringing’’ of the ultrasound are also examined with resonance ultrasound techniques to assure measurement of the shell thickness with an accu- racy of approximately 0.001 in, and to differentiote between possible laminations in the wall and sections of wall thinner than specified. The instaliation for the continucus medsurement of the wall thickness of a core shell is illustrated in Fig. 3.7.3, and the facility for ultrasonic examina- tion for the detection of laminations and cracks is shown in Fig. 3.7.4. DETECTION OF UNBONDED AREAS IN PLATES WITH RESONANCE ULTRASOUND ‘R. B. Oliver Several of the boron-containing plates for the ART shielding, which were fabricated by Allegheny- Ludlum Steel Corporation, blistered when they were ] » PERIOD ENDING SEPTEMBER 30, 1957 = UNCLASSIFIED Y-23440 Fig. 3.7.3. Facility for Wall Thickness Measurements of Core Shellis by the Ultrasonic-Resonance Method. (Socres-sith-capion) heated for formmg. The boron carblde-copper core was separated from the type 430 stainless steel cladding- by @ copper diffusion barrler,‘and the blisters indicated large . unbonded areas at the -mterface ‘between the cladding .and the diffusion’ bacrier, Since only. large unbonded areas formed blnsters, a nondesfruchve inspection method was', needed in order to detect all unbonded areas before further processmg., The resononce-ultrasound e method was selected as the method most suited for measurement of these thin sections, In order to - use the ulirasomc ‘method, the #ansducer ‘is sup= - plied wnth a frequency-modulcted sngnal and resonance is established for each case that the instantanecus frequency satisfies the equation: nV- Tz where - f = frequency (106 cyc Ies/sec),' ' V = velocity of sound (105 in /sec), t = fhe ihlckness of the ob|ect (m.), n ‘= asmall, whole integer. " Either the fundumenfal frequéncy n = l) or -the dlfference between any two successive harmonic frequencies fixes the thickness of the test object, 251 ANP PROJECT PROGRESS REPORT UNCLASSIFIED Y¥-23476 Fig. 3.7.4. Facility for Ultrasonic Inspection of Core Shells for Laminations and Cracks. (-Gemvm-fl-rfl'phan') 252 L] and any change in thickness will cause a shift in these resonance frequencies.. A lamination or an unbonded area will appear as a change in thickness and may be readily detected if the defect area is at least one-fourth the transducer area. The in- strument presents the amplitude of the resonance pecks on the vertical sweep of a cathode-ray os- cilloscope, and the frequency values, or thickness dimensions, are displayed on the horizontal sweep of the oscilloscope. A modulation range of 2 to 4 megacycles was chosen, since this range would produce two harmonic resonances in the full 0.100-in. plate thickness but would be too low to produce a reso- nance in the 0.010-in. cladding layer. Also, in this frequency range, on unbonded area in the interface near the transducer results in a complete loss of signal; while an unbonded area in the interface on the opposite side from the transducer results in a pronounced shift in the location and spacing of the resonance peaks. This shift in PERIOD ENDING SEPTEMBER 30, 1957 the resonance frequencies is equivalent to a change in thickness equal to that of the cladding layer on the far side of the plate. If the unbonded area approaches the area of the transducer, or is greater, the set of rescnance lines representing the full plate thickness will disappear oand a new set of lines will appear; if the unbonded area is less than the area of the transducer, both pairs of resonance lines will be on the screen, A signal gate is available on the instrument to aliow the incorporation of visible and audible signals to indicate the occurrence of an unbonded area, ‘Several of the larger unbonded areas detected with resonance ultrasound and confirmed by stripping the cladding layer from the plate are shown in Fig. 3.7.5. The correlation observed between the nondestructive ond destructive tests indicates that the resonance-uvltrasound method is a simple ond reliable way to locate unbonded areas in this type of plate. Fig. 3.7.5. Unbonded Areas in Stainless-Steel-Clad B ,C-Cu Plates. 253 Part 4 RADIATION DAMAGE G. W. Keilholtz 254 4.1. RADIATION DAMAGE - G.W. Keilholtz EXAMINATIONS OF IRRADIATED - COMPONENTS AND MATERIALS A. E. Richt N. A, Carter W. B. Parsley C. Ellis E. D. Sims E. J. Manthos R. M. Wallace MTR In-Pile Loops Disassembly of MTR in-pile loop No. 6, described previously,! has been completed except for re- moval of the pump impeller from the impeller housing. Thirty metallographic specimens were obtained from the fuel circuit, and 27 of the speci- mens have been examined. |t appears from the results obtained thus far that the maximum depth of corrosion is 3 mils. Specimens from the outlet side of the nose coil showed the greatest pene- tration. Specimen 672, which was taken from the outlet side of the nose coil approximately 1 in. downstream from the location of thermocouples 7 and 8, is shown in Figs. 4.1.1 and 4,1.2, Speci- men 651, which was taken at the location of thermo- couples 9, 10, and 22, on the nose outlet side, is shown in Figs. 4.1.3.and 4.1.4. Specimens 672 and 651 both show the maximum depth of attack of 3 mils... A typical specimen which showed little or no corrosive attack is shown in Flgs. 4,1 5 and 4.1.6. This specimen, No. 641, was taken approxi- mately 13 in. from the face of the pump on the nose inlet side, A remotely operated lathe will be used to' open- the impeller housings of the pumps removed from in-pile loops Nos. 4, 5, and 6 so that the impellers can be examined. ' The impeller housing of the pump from foop No. 5 will be opened first, since its activity level is much lower than the activity levels of the other two smpeller housmgs. . Moderal‘or Materiul s | The three Cc:psules__:rrudtated in the MTR high-" | temperature moderatorematerials testing assembly were examined,? ‘The test assembly, after removal C. C. Bolta et al,, ANP uar. Prog. Rep. Sept. 10, 1956 ORNL-2157, p 81. ¢ P P 2E0r details of test conditions see J. A. Conlin, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 107. from the container, but with the capsules in place, is shown in Fig. 4.1.7. The largest capsule con- sists of a graphite slug enclosed in a nickel can, and the smallest capsule consists of ZrH jacketed in molybdenum. The remaining capsule consnsts of three BeO slugs in an Inconel can. The thermo- couples on the BeQO capsule were all intact, but some of the thermocouples on the graphite and / ZrH capsules were broken off or damaged. Di- ‘fl‘i‘énsions of the three capsules, taken before and after irradiation, are compored in Table 4.1.1. The three BeO slugs were removed from the capsule by cutting the welds at the end caps and then pushing the slugs out of the can. Dimensions of the BeO slugs before and after irradiation are compared in Table 4,1.2. No evidences of swelling Table 4. 1.1, Dimensions of Moderator«Material Capsules Before and After Irradiation in the MTR Diameter of Can* (in.) Before After Irradiation irradiation Capsule - Z?H" in a molybdenum can Not available 0.5366 0.5367 0.5360 0.5363 1.0412-1,0425 1.0417 1.0404 1.0402 1.0404 BeO in an Inconel can Graphite in o nickel can 1.289-1.290 1.2914 1.2905 1.2900 1.2897 1.2894 1.2885 *Measurements were tocken at evenly spaced points along the length of the capsule, - &y 257 s {‘\o ANP PROJECT PROGRESS REPORT Fig. 4.1:2. Specimen 672 After Etching. 250X. : UNCLASSIFIED il RMG 1815 UNCLASSIFIED . RMG 1816 PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED -~ RMG 1817 . [ ] i - . F . - ; :: . : ?a- ! . f * & 'z' [T ‘e . * . #* B 95 . > . - ‘ . " . » & . e " S ] - . & ; Flg- 4.1.3. § | x UNCLASSIFIED ' RMG 1818 » -~ | ! Fig. 4.1.4. Specimen 651 After Etching. 250X, 259 ANP PROJECT PROGRESS REPORT Fig. 4.1.6. Specimen 641 After Etching. 250X, 260 8 UNCLASSIFIED. | - 3 PHOTO 41095 o Fig. 4.1.7. Copsules Containing Moderator Materials Shown During Disassembly After Irradiation in the MTR. O~ LS6L ‘08 Y39WIL4IS ONIAGNI A0id3 d ANP PROJECT PROGRESS REPORT or cracking were noted on the BeO slugs, which are shown in Figs. 4.1.8 and 4.1.9. Slug No. 1 was at the top of the capsule and thus was farthest from the reactor. An attempt was made to remove the graphite slug from its can in a similar manner, but it could not be pushed out easily. Therefore, the can on the graphite slug will be slit open when the ZrH,- containing capsule is sectioned on the remote milling machine. Table 4.1.2. Dimensions of Three BeO Slugs Before and After lrradiation in the MTR Length - Diameter Slug (in) (in.) Number Before After Before After Irradiation liradiation Irradiation Irradiation 1 0.8325 0.8297 0.9964 0.9953 0.8298 0.9957 2 0.8318 0.8315 0.9950 0.9938 0.8316 0.9940 3 0.832 0.8317 0.997 0.9967 0.8317 0.9964 UNCLASSIFIED RMG 1821 (a) (&) ARE Specimens Six specimens from the ARE have been examined. Specimen R8, a section from the pressure shell wall, including a weld, is shown in Fig. 4.1.10. A crack may be seen at the weld junction, and there are several voids in the weld area. A section from a serpentine bend in a fuel tube in the center of the reactor, specimen R7, is shown in Figs. 4.1.11 and 4,1.12. There was subsurface void formation on the interior wall to a maximum depth of 3.5 mils, but the penetration was not ~uniform. Some parts of the wall showed deeper and more . dense penetration than others. The outer wall of the serpentine bend, which was in contact with sodium, showed what appeared to be a mass transfer deposit (Fig. 4.1.13), in addition to some subsurface void formation, The deposit had plated on the wall to a maximum thickness of 1 mil. Specimen R3, which was also taken from a ser- ~pentine bend ina fuel tube in the center of the core, also showed some subsurface void formation; however, the density and depth of penetration were not so great as for specimen R7. The areas of attack were localized, and some areas of the wall showed no attack, as may be seen in Figs. 4.1.14 and 4.1.15. No deposit similar to that found on specimen R7 was noted on the exterior wall, Specimen F5, which was taken from the inlet of a fuel-to-helium heat exchanger, was cast in epoxy resin so that the fins on the tube would not be UNCLASSIFIED RMG 1823 UNCLASSIFIED RMG 1822 {c) Fige 4.1.8. BeO Slugs After lrradiation in MTR. (a) Slug No. 1, which was farthest from reactore (b) Slug No. 2, {c¢) Slug No. 3, which was closest to reactors 2X. Reduced 5.5%. 262 i i 1 i i ! 2 g PERIOD ENDING SEPTEMBER 30, 1957 3 . UNCLASSIFIED Fig. 4, 1.9. Opposite Ends of BeO Slu§$ 1, 2, and 3 After Irradiation in the MTR. 2X. Reduced 32%. UNCLASSIFIED - RMG-1795 . Fig. 4.1.10. Section from ARE Pressure Shell, Including o Weld. 5X. {Secretwithcoptiory ' 263 ANP PROJECT PROGRESS REPORT UNCLASSIFIED - RMG-1828 . - © (Eorrat with-captiond UNCLASSIFIED i RMG-1827 - Fig. 4.1.12, ‘Section Shown In Fig. 4.1.11 After Etching. 250X. (Secretwithreeption) 264 PERIOD ENDING SEPTEMBER 30, 1957 Fige 4.1.13. Mass Transfer Deposit on Outer Wall of Fuel Tube at Serpentine Bend. This surface was in contact with sodium during ARE operation. As polisheds 500X. (Sectet with.seption) Fige 4.1.14. Section of Specimen R3 Taken from Serpentine Bend In Fuel Tube in ARE Core. Etcheds 250X. {Sesret-wiirtuption)y— 265 ANP PROJECT PROGRESS REPORT B UNCLASSIFIED | "RMG-1830 Fige 4.1.15. Another Section of Specimen Shown in Fig. 4.1.14. Etched. 250X. (Secarwmmrrasusny damaged during cutting. The fin-to-tube wall junction is shown in Fig. 4,1.16. (In the ARE fuel- to-helium heat exchangers the fins were helical strips placed in grooves on the tubes. The strips were not brazed to the tubes.) The interior wall of the tube showed subsurface voids to a depth of 4 mils, as shown in Figs. 4.1.17 and 4.1.18. Specimen F9, which was taken from the middle of a bend in a fuel-to-helium heat exchanger, also showed subsurface voids to o depth of 4 mils, as shown in Figs. 4,1.19 and 4,1.20. Specimen S6, the bottom bellows from sodium valve U-23, was also cast in epoxy resin. No cracks in the bellows folds were found, and only a slight roughening of the inside surface was noted, as shown in Fig. 4.1.21. CREEP AND STRESS-RUPTURE TESTS OF INCONEL _ J. C. Wilson W. E. Brundage N. E. Hinkle W. W, Davis J. C. Zukas MTR Experiments An eight-specimen tube-burst creep test at 1500°F in the MTR, as reported previously,’ resulted in 266 times to rupture that were much shorter than ex- pected, and overheating of the specimens under stress during reactor startup was thought to be a possible explanation for the shortened rupture life. Subsequent out-of-pile tests condu,c‘:ted‘-'by Douglas and East of the Metallurgy Division in other apparatus but under conditions that duplicated the time-stress-temperature history of the in-pile specimens have indicated that the time to rupture was not appreciably affected by the overheating cycle. _ Another out-of-pile test in which the MTR dppd-. ratus is duplicated is currently being conducted by the Solid State Division. The time-temperature- stress history of the in-pile specimens is again being duplicated exactly. The apparatus used for the Solid State Division test differs from the appa- ratus used by the Metallurgy Division in that the heaters and thermocouples are in the same chamber as the specimens and contamination of the atmos- phere is a greater problem. ' The creep apparatus irradiated in the MTR has been cut from the irradiation plug and returned to J. C. Wilson et al., ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, » 268, 2" g. Rep. ] PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED RMG-1832 o f’ Lttt L X fare ) M Section of Specimen F5 Showing Fin-to-Tube Wall Junction in ARE Fuel-to-Helium Heat Exchanger. Fig. 4.1.16. Fins were wound helically in grooves and were not brazed to the tubes. . 100X. Etched D A . g2 <9 J NR u, Interior Wall of Tube Shown in Fig. 4.1.16. As polished. 250X. (mflnupflmfi— L.17. 4, Figo 267 ANP PROJECT PROGRESS REPORT Fig. 4.1.18. Section Shown in Fig. 4.1,17 After Etching. 250X. (fzTret~with-S-ect.I6tr| ofrs.peclmen. Sé 'I-'a.ken frorfi the Bottom Bellows of Sodium Valve U-23, Etch’do 250 x. 4.1.21. F‘gn ) 269 ANP PROJECT PROGRESS REPORT ORNL. Examination of the specimens. for fracture location and surface - contomlnoflon is scheduled- in the hot cells. | Temperature control in a recent MTR |rrad|cmon'3 of the elgh_t-spec_:lmen apparatus was . found ‘to ‘be - difficult because of the steep gamma.ray heating gradient in specimens oriented with their long - axes along the direction of the gradient (specimens parallel to beam hole uxis),' An out-of-pile mockup of an apparatus with specimens mounted perpen- dicular to the beam hole axis is being tested. Only four specimens (instead of eight) can be. accommodated in this apparatus, but improved temperature control should result during startup and shutdown of the reactor. LITR Experiments Leaks of undetermined origin caused three stress- corrosion experiments in the LITR to be discon- tinved. Inconel specimens exposed to NaF-ZrF o UF, (62.5-12.5-25 mole %, fuel 46) were being tested at 1500°F and a stress of 2000 psi. One specimen _had been in the reactor for 100 hr after startup, but the other two were irradiated for less than 24 hr. Since all the apparatus was leaktight (as determined with a helium leck detector) before irradiation, thermal stresses that occurred during irradiation are believed to have caused opening of brazed joints or tubing connections. The appa- ratus now being constructed is being instrumented to check the point of failure. The specimens that failed are to be examined in the hot cells. The apparatus that failed showed that improved furnace construction had cllowed control of the temper- ature at 1500°F for a wider variety of fuel loadings and neutron fluxes than was possible earlier. An experiment was carried out in beam hole HB-3 of the LITR to investigate the effect of irradiation on several kinds of thermocouple in- sulation. Glass-insulated thermocouple pairs; sheathed, swaged, MgO-insulated pairs; ceramic insulators; and metal-sheathed, unimpregnated, gloss-insulated pairs were tested. The specimens were tested in air and in helium at temperatures up to 1500°F. Measurements of electrical leakage between the wires of a pair were made in each case as a functlon of time, flux, mtegrated flux, and occasnonal variations in temperature during irradiation and during shutdown periods. The data have not yet been correlated completely, but it has been established that the insulating materials 270 .and practices used in past (and projected) LITR and “MTR cpporctus “have not ‘caused errors .in - temperature measurements. The only insulation - that failed during the tests wasona metal-sheathed, " unimpregnated, glass-insulated pair. The electrical _ potential used during the tests was about 20 v. A - metal-sheathed, swaged, MgO-insulated pair of the type intended for use in the ART was included in - these tests, buf a more definitive experiment that ~ will include a study of the effects of irradiation on the resistivity of Inconel is to be conducted. LITR YERTICAL IN-PILE LOOP W, E. Browning J. E. Lee, Jr. H. E. Robertson M. F. Osborne R. P. Shields The forced-circulation loop which was operated in a vertical hole in the LITR for 235 hr, as de- scribed previously,4 is being disassembled for ex- amination and analysis of the fluoride fuel mixture circulated therein. The loop has been removed from the outer container, and the major subassemblies have been detached. Examination of these sub- assemblies is in progress. The supposition that the pump became inoperative because of beormg seizure was substantiated. The two pump-shaft bearings were removed by cutting laterally through the pump and shaft above and below each bearing. The shoft in the bearing nearest to the fuel could not be turned by the manip- ulator, A black porous deposit was found on the wall near the bearing that was, presumably, radio- lytic decomposition products from the grease, The shaft in the bearing just below the motor could be tumed, but only a small amount at a time, and the bearing above the motor turned freely. -All three bearings had been lubricated with the same radiation- resistant lubricant. The difference between the behavior of the bearings is attributed to_the differ- ence in dosages from the decay -of short-lived fission gases as they diffused upward through fhe pump. Of the 55 thermocouples in the Ioop, 22 fctled Seven of the failures were due to an open. circuit in the platinum-rhodium leads. One thermocouple failed because the bead came loose - from the Inconel fuel tube, and 14 failed because both the platinum and the platinum-rhodium lead wires had ‘w. E. Browning et al., ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340, p 275. i broken. |llustrations of thermocouples that failed are presented in Figs. 4.1,22, 4.1.23, ond 4.1.24. These thermocouples were located on the fuel tubing in the high-temperature region of the loop. Three beads were used for each thermocouple in- stallation, as shown in Figs. 4.1.22 and 4.1.24. Both leads had pulled away from one bead shown in Fig. 4.1.22, and only one lead remains attached to another of the beads. The pair of wires that pulled away from the bead during operation are shown in Fig. 4.1.23. The tapered ends are typical of ductile failure in tension. Wires that broke about Y to 1/2 in. from the bead are shown in Fig. 4.1.24. The ends of these wires also show the tapered effect illustrated in Fig. 4.1.23. The mechanical damage of the lead wires probably occurred during thermal expansion of the loop. - The platinum resistance heaters and their lead wires were examined closely to determine the cause of failure of 4 of the 22 heater circuits. The heating elements were in good condition, and UNCL ASSIFIED RMG-1877a PERIOD ENDING SEPTEMBER 30, 1957 . UNCLASSIFIED RMG-1877b Fig. 4.1.23, WIres"Tfiut Pulled Aweay _fror.n. Thermo= couple Bead Shown in Fig. 4,1.22, the failures appecred to have been in the lead wires,’ " Particular ottention was given to the d:sfrlbuhon of fuel and fuel vapor deposits in the pump. The pump was sectioned longitudinally along the shaft, ~_ond -one-half is shown in Fig. 4.1.25. No vapor . deposits were found in the region above the Graph- itar vapor baffle, - splashing of the fuel in the sump chamber. It is . .inferesting ‘to compare the vapor deposfls with - those observed in the out-of-pile pump,’ in which ,much Inrger deposns of ZrF, were found, The ,dnfference may be. attributed to the’ longer operating ~life of the out-of-plle test pump of over 3500 hr. " Microscopic examination - of the - Graphitar vapor There- was no evidence of ' baffle- showed. fhat the shaft had not. fubbed against the baffle; the original tool marks on:the Graphitar . "-::'-’r_..:-we'f.e undisturbed. The general appearance of the Fig. 4.1.22, Thermocouple That Failed During Op- eration of a Vertical In-Pile Loop in the LITR. 5W E. Browning et al., ANP Quar. Prog. Rep. March 31d 19]5;.4 ORNL-2274, p 255, esp Figs. 4.1.22, 4.1.23, and 4.1,24, 271 ANP PROJECT PROGRESS REPORT Fig. 4.1.24. Thermocouple That Failed During Op- eration of a Vertical In-Pile Loop in the LITR Showing Broken Lead Wires. loop, as observed during disassembly, was goed. The quartz tape that was used to insulate and bind the wires was not discolored or damaged. One qir annulus was slightly bowed. Some of the structure of the loop may be seen in Fig. 4.1.26. The heater section shown in the figure is in good condition. No discoloration was found on any of the ceramic heater forms. Tests are being prepared of bearings designed to withstand fission-gas radiation for longer operating times and thermocouple installations designed to prevent damage of the thermocouple leads during thermal expansion of the fuel tubes. Samples of minute deposits found at various locations in the pump are being analyzed chemically and radio- chemically in order to determine the distribution of fission products in the fuel. Specimens of the Inconel tubing are being taken from the fuel circuit for metallographic examination to determine the extent of corrosion in the loop. 272 MTR STATIC CORROSION TESTS "H. L. Hemphill “Two Inconel capsules containing the fuel mixture W. E. Browning - NaF-ZrF -UF, (53.5-40-6.5 mole %, fuel 44) were withdrawn from the MTR after 720 hr of irradiation at 1500°F under static conditions. These capsules will be returned to ORNL for examination. Two Hastelloy B capsules were inserted in the MTR, but “operational difficulties made it necessary to delay temperature cycling, FLUX MONIfORING OF MTR STATIC CORROSION CAPSULES D. E. Guss The average thermaleneutron-flux exposures of several static corrosion capsules irradiated in the MTR have been calculated from the Cof? disin- tegration rate of the Inconel container material, The average fluxes for four of the capsules were also determined by the yields of Zr?5 and Cs'37 from U235 fission in the fuel contained in the capsules. The values are presented in Table 4,.1.3. Revised flux values based on corrected irradiation times are given for three capsules. RELATIVE CONTRIBUTION FROM THE N16%(n,p)Co%® REACTION TO THE Co%® DISINTEGRATION RATE IN IRRADIATED INCONEL D. E. Guss The contribution to the Co%? disintegration rate in irradiated Inconel arising from the nickel con- stituent through the Ni%%(n,)Co%0 reaction was investigated further,® In order to obtain a measure of this contribution the Co®° disintegration rate was determined for several pieces of vacuum- melted carbonyl nickei (~0,002% Co) which had - received @ 75-min irradiation in the lattice (hole .C-46) of the LITR in September 1955. The Co%0 disintegration rate per gram of nickel was equiva- lent to that which would have been produced by 1 x 103 g of cobalt. Even if all this activity . were attributed to the Ni®%(n,p)Co80 reaction, that is, if there were assumed to be absolutely no 6p, .E. Guss, ANP ar. Prog. Rep. June 30, 1957, ORNL-2340, p 3760 2" ! PERIOD ENDING SEPTEMBER 30, 1957 PHOTO 41397 - | " REGIONS WHERE VAPOR “4E DEPOSITS COULD FORM | . GRAPHITAR. | VAPOR BAFFLE it FUEL SURFACE g IMPELLER ’ FUEL OUT ~ FUEL IN fi Fig. 4,1.25, Sectioned Pump from Vertical In-Pile Leop Operated in the LITR. 273 vLe e L6 AIR TUBES TO LEGS OF LOOP s -v"(j %‘5?’ i I%%%’&%g % o i o 5 é}&" B . {k;m . A P T i G oy P R 2 T g Fig. 4.1,26. View of Partially Disassembled L.ITR Vertical In-Pile Loop. L30d3Y $SIN90¥d L1D3Irodd dNV " PERIOD ENDING SEPTEMBER 30, 1957 Table 4,1.3. Thermal-Neutron Flux Yalues Obtained from Flux Monitoring . of MTR Static Corrosion Capsules Average Flux Average Flux Capsule Discharge Irrod.iation ' Cal.cu.rlated from Co%0 Cfll:t.:lufed from Number Date Time Disintegration Rate Fission-Product _ (Mwd) in Inconel Capsule Yields (n/cmz-sec) (n/cmzcsec) x 1014 x 1014 227 12-29-53 367 1.39 232 11-18-54 1282 1.44 237 3-28-55 C 990 1.42 245 6-20-55 | 939 0.80* 252 7-7-55 535 2.36 253 3-28-55 990 1.60 274 6-20-55 | 939 0.78* 275 | 6-20-55 939 1.10% 316 - 7-5:56 1980 177 1.97 317 7-5-56 1980. 1.12 1.19 333 9-24-56 510 0.78 0.76 345 9-24-56 - s10 0.59 0.67 *Revised values based on corrected irradiation times. cobalt present in the nickel, it may be seen that the contribution to the Co®® activity in irradiated Inconel arising from the Ni%%(n,p)Co%? reaction would be negligible.” The Inconel used in fabricating corrosion-test capsules contains about 1,5 x 10-3 g of cobalt per gram of Inconel, and thus the con- - tribution to the total Co%9 activity from the nickel would amount to less than 0.5%. Further, the tests’ are run under conditions in which the ratio ofthe fast-neutron flux to the fhermal-neutron flux |s- ' clbout unity, " ETR.'IRRAD'IATION' OF MODERATOR MATERIALS W E. Brownmg G. Somuels - R. P. Shields Preparqhons have continued for urudlahon mi the ETR of moderator materials for use in high- ~ temperature reactors. * The irradiation request was approved by the AEC and design details are being worked out with ETR personnel. Fabrication of the BeO test specimens has been started, and arrange- ments have been made for fabrication of yttrium hydride test specimens at GE-ANP. (For details of design of equipment for the irradiations see Chap. 1.4, “Engineering Design Studies.”’) ... EFFECT OF IRRADIATION ON ~ THERMAL-NEUTRON SHIELD MATERIALS J. G. Morgan ~ M. T. Morgan P. E. Reagan | ‘Boron Nitride Boron nitride is similar to graphite in many re- - spects, but it is white in color and much more re- sistant to oxidation at high temperatures. Favor- able results have ‘been obtained from radiation damage tests conducted on boron nitride as a _neutron shielding material. The tests were made, primarily, to explore the general integrity of this relatively new ceramic under irradiation and, 275 ANP PROJECT PROGRESS REPORT specifically, to determine weight, density, and dimensional changes, Boron nitride was obtained from two manufacturers (Norton Company and The Carborundum Company) in powder form and 3/1 gmine~ oD, l'/z-in.-long test specimens were hot pressed from the powder. The Norton samples are designated 6002, and the Carborundum samples are designated Cl. The samples were irradiated in an inert atmosphere either in quartz ampoules or in Inconel capsules. High-temperature irradiations were made in the Inconel capsules, and the quartz ampoules were irradiated at the temperature of the test reactor water, The high-temperature tests were monitored by thermocouples pressed against the outer surface of the sample with leads extending through a ceramic seal in the capsule top. The thermal- neutron dosage was calculated from the activity of a cobalt foil located against the sample. Surface darkening of the samples duringirradiation is thought to have been caused by a metallic deposit. The coating was found to be primarily magnesium, with some silicon and traces of calcium and copper. The irradiated samples retained good crystallinity, The results of x-ray diffraction ex- ominations of samples irradiated in the MTR are presented in Table 4.1.4. The conditions of the tests and the physical property changes resulting from irradiation are described in Table 4.1.5, and irradiated and unirradiated samples are compared in Fig. 4.1.27. The dimensicnal changes that occurred during the tests were small, and the changes that were detected could not be definitely attributed to radiation damage, since boron nitride is a soft material’ and the micrometers used scraped material from the surface during measurements of the samples. Both the weight and the density changes were less than 1%. The burnup values shown in Table 4.1.5 are average values calcu- lated on the basis of neutron flux and irradiation time. Boron isotopic analyses are being made. These tests indicate that boron nitride is un- damaged by irradiation at a temperature of 1800°F to a thermal-neutron dosage of 1.3 x 1020 neutrons/cm?. Physical property measurements are now being made on the four samples irradiated at 70°F. The gas evolved during irradiation will alsobe analyzed. 7The samples had a Mohs' scale hardness of 2 before lrradluhon, which increased to 2.5 during irradiation. 276 Table 4, 1,4, Results of X-Ray Diffraction ' Examinations of Irradiated Boron Nitride Lattice Spqcing Dimension (A) Unirradiated Plane | Indices Original 6002-1 6002-4 Powder Hot-Pressed Sample (002) 3.32 3.33 3.37 3.33 - (100) 217 2,17 2.17 2.17 (101) 2.06 2.06 2.06 . (102) 1.81 .82 1.82 1.82 (004) 1.66 1.68 1.66 (104) 1.32 1.32 (110) 1.25 1.25 1.25 (12) 117 137 1.7 Hexagonal Structure Axes A 2.50 : 2.50 C 6.64 6.74* = *The expansion in C is 1.6%. UNCLASSIFIED RMG- 1842 (a) o)y » Fig. 4.1.27. Hotk ** Average B'0 burnup, % '5.0." 48 48 4.0 None 3.8 3.8 *Weight-change measurement not made because chips were |osi when sample broke. **Physical properties yet to be determined. " Boren Carbide Irradiation tests of two hotspressed boron carbide tiles that were prototype samples of ART tiles were completed. A sample designated J-1 was irradiated .in the LITR C-39 facility for 1148 hr at the reactor water temperature to a cajculated average burnup of 1,4%. A sample designated J-3 . _ was irradiated in the LITR C-42 facility for 910 hr .. . at 1500°F to an average bumup of 2.2%.. Post- - urud:uhon exammahons of these scmples are in,,. ‘ _ -+ than the CaB,-Fe cermet, and there was some ~_ separation - of the core from the cladding. No ~ significant changes in structure of the core or the cladding ‘were found. There was no noticeable ‘increase in fragmentation or porosity. . Longitudinal progre ss. . An_ ou?-of-plle test ‘was conducted in’ whlch a- third sample (J-2) was cycled 18 times: from 200 to - 1600°F, "~ The . sample was ‘then "examined under @ microscope af magnifications of 20 and 40, and '_ - no cracks were observed Ce rmeis in evacuated quartz ampoules cooled by the LITR reactor water and examined after two different bumups. The boron in both cermets was natural boron. There were no structural changes as a re- sult of irradiation to an average B'0 burnup of . 19%, as shown in Table 4.1,6, but both materials were damaged by irradiation to an average burnup of 38%. " The CaB,-Fe cefmet showed slight cracks "throughouf the core matrix after the irradiation to ~'an average bumup of 38%, as shown in Fig. 4.1.28. The BN-Ni cermet was more severely cracked cracks in the sample irradiated to an average burn- ‘up of 38% are shown in Fig. 4.1.29, and the two irradiated samples are compared at hngher magmfi- e e . = . cation in Fig. 4,1,30. Plates of BNoNl (103 wt % BN) and CaB -Fe- (7.6 wt % CaB,) cermets clad with type 304 stum-fi 3 less steel have been irradiated at less than 345°F Stainless-steel-clad B C-Cu cermets ‘have also been studied. The first three specimens irradiated “were fabricated by the ORNL Metallurgy Division. They were 0.5 in. long, 0.1875 in. wide, and - 0.102 in. thick. The cores of these specimens 277 ANP PROJECT PROGRESS REPORT Takle 4.1.6. Effect of Irradiation on Dimensions and Hardness of BN 104 photons/cm?2 for the 4 ohm-cm material diode “and approximately 7 x 1014 photons/cm? for the 12 ohm-cm rna_terlal_dlode. Further, carrier re- moval is ‘manifested as a reduction of current, 9).'W. Cleland, J. H. Crawford, Jr., and D. K. Holmes, Phys. Rev. 102, 722 (1956). 289 ANP PROJECT PROGRESS REPORT ;6 UNCLASSIFIED = - (;10 ) ¢ HNL-LR-D\VIG'ZEQSG "|CK=-708~-12 ¢ CK-708-4 ¢ ° ® /— 3 / e AL V2E° ~108 AR 40“‘/ } | NO. 4179 -~ \ NO.11 ® a &, s pROBABLE CURVE OF NOS, WARDHe — — : — - ) — — T 4 — /NORMAL!ZED NOS. 41 AND 4.4 2 ) 0 i00 200 300 400 500 600 700 800 2900 3000 IRRADIATION TIME {min) Fig. 4.1.42. Reverse Current (1-v Bias) Through CK-708 Diodes vs Gamma Irradiation Time. while in these samples the current increases, It is evident that other factors are involved. There would be, of course, a point at which the bulk effects would begin to show, but higher dosages then those given to these samples would be re- quired. The dosage required can be estimated by referring to previous work done on gamma-ray bombardment of diodes in which the reverse current at 1 v continved to rise until a dosage of 1,75 x 107 r was reached and fell thereafter.!0 Two components of the current which composed the total current through the CK-708 diodes — the bulk current, represented by the saturation current, and the surface leakage, represented by the satu- ration slope -~ are shown in Figs. 4.1.43 aond 4.1.44. In the present model of a diode, these currents are in parallel, and the sum of the currents represents the total current through the diode. Such is not the case, however, and this discrepancy is part of the reason for the speculation regarding the need for a new model. Annealing of the samples is showrin Fig. 4.1.45, but since there is an instance of a high-resistivity. sample annealing more rapidly thana low-resistivity 10, c. Pigg, Solid State Semiann, Prog. Rep, Feb, 29, 1956, ORNL-2051, p 59. 290 sample and an instance of a low-resistivity sample annealing more rapidly than o high-resistivity sample, no conclusion can be drawn from these - curves, Annealing saturation currents and satu- ration slopes have not yet been analyzed and are not presented here. ' The study of the germanium junction as a function of both temperature and radiation is being hampered by instrumentation difficulties. The study is to be as precise as possible, and the sample currents are to be measured at voltage biases as low as 10 mv. Difficulty hos been encountered with a-c electrical pickup in the servo-controlled power supplies and in providing high-impedance inputs to a Brown recorder, which has a normal impedance of 500 ohms. The servo power supply has been com- pletely redesigned and now incorporates a number of advanced features not ordinarily found in re- cording systems of this type. An input impedance of 180,000 ohms has been obtained. A block diagram of the system is shown in Fig. 4.1.46. The recorder is an advonced model of the recorder described in ORNL-1199 (ref 11). The more recent ", c Pigg, An Automatic Multira}zge Recording De- vice for Measuring Varying Potentials, ORNL-1199 (April 18, 1952). PERIOD ENDING SEPTEMBER 30, 1957 -6 UNCLASSIFIED (’;'O } ORNL-LR-DWG 22982 3 I E2 CK-708-4 o ™ . ° CK-T708-12 . ' o / \ / had 0 0 100 200 300 400 500 600 700 800 IRRADIATION TIME (min) Fig. 4.1.43. Bulk Current, Represented by Reverse Saturation Current, Through CK-708 Diodes vs Gomma Irradiation Time. (’é' 0% ORNL-LR-OWG 22083 5 CK-708—4 o I —5 ° CK—708—12 A Lol ~ 3 2 0 . 100 200 300 400 500 600 700 800 IRRADIATION TIME (min) Fig. 4.1,44. Surface Leakage, Represented by Reverse Saturation Slope, of CK-708 Diodes vs Irradiation Time. 291 ANP PROJECT PROGRESS REPORT _ UNCLASSIFIED (“g ) ORNL~LR-DWG 22953 ® A \ \. : - 7 N ~ et \ v = /" NO.4 L " i \___-.. ’/ et = § 5 ~ S\ NN RN A a %"’-\ ~ -. e R S .~NO. 4.4 . =T \ '-_. / - "'N..,___. . —"_.-______ AV @ 3 NO.§ 2 o 500 {000 ’ 1500 4800 2000 ANNEALING TIME (min) Fig. 4.1.45. Reverse Current (1-v Bias) Through CK-708 Diodes vs Annealing Time. modifications of the servo power supply for the system are the high-impedance input, the variable- gain amplifier, the velocity-damping network, and the variable-phase chopper power supply. Recent modifications of the Brown recorder servo circuit are the high-impedance input, the variable-phase chopper power supply, the positive-current feedback- damping network, and jitter filter. Other improve- ments in the system have included replacement of certain relays to reduce the current through selectro switches, and replacements of certain relays which were shown to be incapable of carrying the necessary currents. The samples for this experi- ment are IN-91 alloy junction diodes. For com- parison, a germanium bulk sample is bombarded with the diodes. The lifetimes and the barrier heights of the diodes are measured following irradiations. Changes in the resistance of the bulk sample provide a check on the neutron dosage to which the samples are subjected. Both the dicdes .and the bulk sample have been bombarded at ~78°C in a facility that has a fast- 292 neutron flux component of 1.66 x 10% neu- trons/cmZ.sec., After irradiation, the characteristics curves of the diodes were recorded at -~78°C, and the diodes are to be warmed to 0°C. Sample characteristics will be recorded while the samples are being warmed. Bombardments up to and in- cluding a dosage of 10! neutrons/cm? have caused little change in the diode samples. After a bom- bardment to a dosage of 102 neutrons/cm?, a flattening of the forward curve at —78°C has been observed. The reverse current is not observable at this temperature. The samples are being held at ~78°C pending replacement of faulty relays so that characteristics may be recorded during the wormup period. Experimental Refinements It has been noted that the changes observed in the irradiations performed both in this program ond in other programs cre, for the most part, opposite to those expected from the results of bulk or sur- face studies. Therefore efforts have been made to B PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED ORNL~LR—-DWG 22954 _ SERVO HIGH IMPEDANCE | \ AMPLIFIER INPUT SAMPLE - ' REFERENCE l VOLTAGE SWITCH | Aflfiamfiyc HIGH IMPEDANCE - BROWN SHUNT INPUT RECORDER THERMOCOUPLE COLD JUNCTION Fig. 4.1.46. Block Diagrom of Multipoint Recorder. compile quantitative information concerning the behavior of barriers or rectlfyang devices in the | presence of radiation fields so that, by comparison with the results of bulk and surface studies, those chcrccter:shcs unique to the barrier can be defined. Further refinements of these studies are being made. Several samples of barriers of known charac- teristics have been obtained and will be studied. The equipment necessary for fhese stud|es is bemg , designed, built, or purchcsed ' ‘A probe mechanism is bemg constructed that consists basically ofa fine screw-driven base which will carry a scmple under a probe for determination of its resistivity as a function of sample displace- . ment and to determine location and ‘change of the junction in a sample. ' As mentioned before, the surface conditions or surrounding environment of a device can play a more significant part in the behavior of the device than would be expected from the results of bulk or - surface studies. In order to eliminate or at least minimize surface effects it has been decided to " prepare samples for irradiation in a stondard atmosphere. A vacuum has been proposed and may be tried, but, for the present,.helium will be used, - since it will provide a heat transfer medium in the event gamma-ray heating becomes a problem, A dry box has been obtained, and ‘it will be meodified ‘to provide a helium atmosphere, To further assist _in sample preparation, a small chemical hood has been designed and is now being built. This will permit the safe use of etches in the preparation of the samples and in changing the surfaces after the samples have been bombarded. : In order to increase the range of voltages which can be applied to a sample and thereby obtain 293 information about break-through characteristics and to enhance the accuracy of saturation slope data, an *“X-Y'’ recorder with increased ranges is being obtained. A sample power supply that will “be compatible with the characteristics of the recorder is to be designed. The recorder can measure applied voltages up to 100 v on either scale, and thus the range is beyond the maximum voltage at which most diodes operate. With proper . shunts it will be possible to extend considerably the range and the types of units which can be studied. Another multipoint recorder has become available with which to constantly follow changes in a sample 294 over a period of days. A controlled-voltage power supply is also available, and it is thought that with moderate modifications it will be suitable for use with the recorder. A recorder of this type willi do much to clarify such changes as the initial drop in current of the 12 ohm-cm dicde, as shown in Fig. 4.1.42, Clarification of such changes is necessary for a conclusive understanding of the mechanism of changes in semiconductor devices, The design of the fast-flux irradiation facility discussed previously is now being detailed. No date for completion of the design has been set, but the werk is proceeding satisfactorily. n Part 5 REACTOR SHIELDING E. P. Blizard 5.1. LID TANK SHIELDING FACILITY W. Zobel RADIATION ATTENUATION MEASUREMENTS IN PLAIN WATER, BORATED WATER, AND OIL D. W. Cady! E. A. Warman? A comparison of the attenuation of fast neutrons, thermal neutrons, ond gomma rays in various liquids has been made through the use of the Lid Tank Shielding Facility {LTSF) at ORNL. The liquids investigated were - plain water, borated water, and transformer oil. The borated water used in these tests contained 1.34 wt % natural boron and had a density of 1.05 g/cm3. The transformer oil was composed of 86.7 wt % carbon and 12.7 wt % hydrogen cnd had a denslty of 0.87 g/ecm® at 20°C. In order to avoid mlxmg of the borated wuter or oil with the plain water in the lid tank, these materials were contained in the usuval *‘configura- tion'’ tank which was positioned as close to the source as was physically possible. The con- figuration tank is constructed of mild steel and has dimensions of 654 x 72 x 7]'/ in. The pro- duction of capture gamma rays in the tenk wall near the source was reduced by the insertion of a %-in.-thick aluminum window in the side of the tank adjacent to the source. A 2-cm gap between the source and the window was filled with air contained in a thin plastic bag. Thermal-neutron, fast-neutron, ond gamma-ray measurements taken in the three media contained in the configuration tank are shown in Figs. 5.1.1, 5.1.2, and 5.1.3, respectively. Before the data for plain water can be. compared with previously published measurements,3:4 which were taken in the lid tank rather than the configuration tank, the new curves must be shifted 3 cm to the left. The On usmgnment from ‘Wright Alf Development Center, Dayton, Ohio. . 20n osslgnment from Pratt & Whitney Alrcrah, East Hartford, Connecticut. 3D. R. Otis and J. R. Smolen, Prelimina LTSF Data Report on Experiments with Advanced Shielding Ma- terials (LiH, Zr, Be, Stainless Steel, W, Hevzmet, Pb, and v (Expt 69), ORNL CF-57-2-8 (Feb . 1957). 4R, W. Peelle et al., ANP Quar. Prog. Rep. Dec. 31, 1956, ORNL-2221 (Part 1-5), p 332, Fig. 5.1.1. agreement between a thermal-neutron traverse in the lid tank and a traverse in the configuration tank, with the latter data shifted on the graph a distance of 3 cm to the left, is illustrated in Fig. 5.1.4. The previously reported curves for thermal- neutron fluxes in oil and inborated water represent data taken with the instrument response normal- ized to gold-foil fluxes in plain water. This normalizing procedure is not particularly valid, as is indicated by a comparison of counts-to-flux conversion factors used over a period of time. For example, the current conversion factor used with a ]2'/2-in. BF, counter in plain water, ob- tained by normalizing to the data taken with gold foils in plain water, is 1.66 x 10=3. Previously, this same factor would have been used for this counter for all media. The newly determined con- version factor for borated water is 9.78 x 10=4, a change of about 41%. Similarly, the conversion UNCLASSIFIED 2-01-057-0-383% %Y 0% - -t - - - ‘thoN' ofl Uob U% GC’O o ' THERMAL-NEUTRON FLUX (neutrons/cmE-sec-wott) o o> O 20 40 60 80 100 120 140 160 2, DISTANCE FROM SOURCE PLATE {cm) Fig. 5.1.1. Thermal-Neutron Fluxes in Water, Borated Water, and Oil in the Configuration Tank. 297 j/ 297 ANP PROJECT PROGRESS REPORT UNGLASSIFIED 2-01-057-0-366 10* ad uqa uq FAST-NEUTRON TiSSUE DOSE RATE (ergs/g-hr-watt) - 0 20 40 60 B8O 100 120 a0 160 Z,, DISTANCE FROM SOURGE PLATE (cm) Fig. 5.1.2. Fast-Neutron Tissue Dose Rates in Woter, Borated Water, and Oil in the Configuration Tank. UNCLASSIFED 2+01-057-0-367 GAMMA-RAY TISSUE DOSE RATE (ergs/q-hr-wolt) o 20 40 60 80 100 120 140 160 Z,, DISTANCE FROM SOURCE PLATE {cm) Fig. 5.1.3. Gamma-Ray Tissue Dose Rates in Water, Borated Water, and Oil in the Configuration Tank. factors for the 8-in. BF, counter, the 3-in. fission chamber, and the I;g-in. fission chamber have been reduced by 26, 37, and 16%, respectively. Cor- responding changes were made in the conversion factors for oil. The data presented here are opproximately 5% higher than the earlier data because of the re- placement of the old !fi-in.-thick boral shutter on 298 UNCLASSIFIED 2-01-057-0—-368 0 20 40 60 80 100 120 140 160 #o. DISTANCE FROM SOURCE PLATE {cm) Fige 5.1.4. Thermal-Neutron Fluxes in Water in the Lid Tank and in the Configuration Tank. the source plate by a I4'-in.-thick shutter. Lid tank data are obtained by determining the dif- ference between the shutter-open aond shutter- closed measurements. The old Y%-in. shutter did not completely shield the source plate from the reactor thermal flux; therefore, even with the shutter closed, there was fissioning in the source plate. When the shutter-closed readings are sub- tracted from the shutter-open readings, a small percentage of the source-plate power is eliminated as background. Thus, an increase in the shutter thickness essentially increased the effective source-plate power. This increase is small enough to fall within the uncertainty in source-plate power (15%), and therefore a more exact deter- mination of the new effective source-plate power will be made. Thermal-neutron measurements in wdter are-now being corrected for the flux depression that re- sults from the presence of the gold foils and for self-cbsorption and *‘self-shielding®’. in the foils. These corrections were not made on previously reported curves.3% Although it is commonly assumed that the flux depression for *‘thin'’ foils is negligible, calcula- tion of the flux depression caused by 2-mil-thick, 1-em2 gold foils in water, oil, and borated water were made. Several methods of calculation were attempted, 78 and each produced a different re- sult, The figures reported here are based ona method suggested by W. R. Burrus.® The prelimi- nary values obtained were 8.5% for water, 5.9% for oil, and 5.0% for borated water. Little flux depression is normally expected in borated water, but the amount of boration in the LTSF water is ~so slight, 1,34 wt %, that the 5.0% value is reasonable. it should be noted here that cadmium- difference flux measurements in borated water are of dubious validity. The cadmium ratio is so small, about 1.5:1, that the errors inherent in taking the difference between two large, neorly equal, values are considerable. The self-absorption correction is a correction to the measured activity of the foil after exposure. Since the foil cannot be infinitely thin, there will be scattering and absorption of the radiation in: the activated foil itself. Hence the true activity of the foil, or counting rate, will be reduced by some factor which is dependent on the foil thick- ness. The dependence of the self-absorption on the foil thickness can be calculated roughly by assuming an exponential absorption and an ab- " sorption coefficient _independent of the depth in which the particles are em:fled Based on the assumption that one-half the thickness of the foil contributed to the self-absorpnon, the calculat:on yields a self-absorptlon factor for a 2—m|l-thnck : gold foil of approx:mately 1%. 5W VZOBei“et ali ANP Qufir Prog Rep ]une 30 S 1957 ORNL-2340 (Part 1-5), p 285 ' T. Chapman et al., ANP Quar Prog. Rep ]une - : 10 1955, 0RNL-]896, P 194 : _ ’p, 4. Hughes, * Pile Neutron Researcb Adduon- - Wesley, Cambridge, 1953. 8W R.- Burrus, USAF, Wright Air Developrnent Center, Dayton, Ohio, pnvate communication. R, Aronson et als, Penetration of Neutrons Point Isotropic Fission Source in Water, NY 6267 (Sept. 22, 1954). " THERMAL~NEUTRON FLUX {neutrons/cm?- sec -watt) a - if a foil is quite thick, enough neutrons will be absorbed so that the nuclei in the center of the foil will **see’’ a lower thermal-neutron flux than can be seen by those at the surface. if the ratio of the absorption to the scattering for the foil is large, the ‘‘self-protection,’’ or *‘self-shielding," can be calculated because the incident flux suf- fers an exponential decrease in the foil.” For a 2-mil-thick gold foil this caleulation yields a ‘‘self-protection’’ factor of 1.5%. The effect on the configuration tank water curve of considering the 11% correction for flux depres- sion, self-absorption, and self-shielding is shown in Fig. 5.1.5. Both curves shown in Fig. 5.1.5 already contain the 5% correction for the change in thickness of the boral shutter. Therefore, the UNCLASSIFIED 2-01-057-0-369 - 08 0% 0l 08 0l 0 od, o %0 oL 5, N -n - - 10 0O 20 40 € B8O 100 120 440 160 Z, DISTANCE FROM SOURCE PLATE {cm) . Fig. 5.1, 5. Thermal-Neutron Fluxes In Water in the Conflgurcflon Tank: Compurlson of Uncorrected Meos- urements wlfh Those Corrected for Flux Depression, Self—Absorphon, and Self-ShieIding. 299 ANP PROJECT PROGRESS REPORT difference between the two curves is due only to this 11% correction. There have been efforts to predict theorehcally the attenuation of gamma rays and neutrons in water.?=11 The data presented here will be of value to any future work on this problem. It should be noted that nc attempt has been made to predict attenuation in the oil or the borated water. STUDY OF ADYANCED SHIELDING MATERIALS E. A. Warman?2 The thermol-neutron data that were omitted from the previous report® on the group of tests per- formed to aid the ANP shield design effort of Pratt & Whitney Aircraft are reported here. These data’ present detector response measurements normalized to cadmium-difference gold-foil fluxes. The gold-foil fluxes have been corrected for flux depression by the 2-mil-thick foils and for self- absorption and self-shielding in the foils, as described in the preceding paper. The configurations tested included Hevimet, stainless steel, boral, and lithium hydride, pre- ceded by a beryllium reflector-moderator and backed by borated water to mock up borated alkylbenzene. The materials and their experimental arrangements were described previously;5 however, subsequent chemical analysis of the borated water has shown that the boron content of the water was 1.34 wt % rather than the 1.46 wt % reported previously. A comparison of the thermal-neutron measure- ments taken in water and in borated water (refer to Fig. 5.1.1 in preceding section) showed that the addition of the boron was quite effective in re- ducing the thermal-neutron flux. The borated water curve had a slightly steeper slope than that of the water curve. The thermal-neutron flux in the borated water was a factor of 29 below that in plain water 15 cm from the source plate and a factor of 36 below the flux in water at 120 em. It should also be noted that the hump in the curve in the forward region was not as pronounced in borated water because of the greater percentage of Ve P, Blizard and T. A, Welton, The Shieldin e of Mobile Reactors, ORNL-1133 (Part 2) (June 30, 195 ”D R. Otis, Neutron and Gamma-Ray Attenuation for a Fission Source in Water — Comparison of Theory with LTSF Measurements, ORNL CF-57-3-48 (March 12, 1957); ANP OQuar. Prog Rep. Dec. 31, 1956, ORNL- 222% (Part 1=5), p 331, 300 absorption of the neutrons being thermalized in the first few centimeters of water. A short discussion of the measurements made in the borated water beyond the various configura- tions is presented below, although a complete analysis has not yet been made. N Configurations 25, 26, and 27. ~ Thermal- neutron flux measurements beyond conflguruhon 25 (Fig. 5.1.6) show the effect of replacing 4/ in, of borated water with ‘/ in. of Li-Mg alloy plus 4 in. of beryllium. The fhermul-neutron flux was raised, as was expected, because of the addition of the beryllium. Conflgurahon 26 consisted of 4 in. of Hevimet preceded by / in. of Li-Mg and 4 in. of beryllium, The slope of a plot of the thermal-neutron flux curve for this configuration is considerably altered from that for configuration 25. - This is possibly due to the Hevimet not being os effective as borated water as a thermal-neutron shield, and, as a result, there is a higher thermal-neutron flux at VR 2-01-057-70-376 7 10 5 2 10° 5 ONLY (134 wi % 2 10° . 5 ' . 4 in.OF Be +4 in.OF HEVIMET + 2 10t TION 27: ¥4 in OF Li-Mg 5 +4 inOF Be+% in.OF +4 inOF HEVIMET + B-H,0 2 S W % Li-Mg ALLOY 4+ 4 in. OF Be +B- ONN\ 18] THERMAL-NEUTRON FLUX {neutrons/cm?.sec- watt) N =N O N o < N o <3I nN 0 {0 20 30 40 50 60 VO B8O 90 {00 {10 120 430 Zo, DISTANGE FROM SOURGE {cm) Fig. 5.1.6. Thermal-Neutron Fluxes Beyond Con- figurations 25, 26, and 27, (Secrety short distances beyond the Hevimet. However, the attenuation of fast neutrons in the Hevimet resulted in a reduction of the thermal-neutron flux at greater distances. ' In configuration 27 a lé-in.-thicl«: boral curtain was placed between the beryllium and the Hevimet. The thermal-neutron flux was somewhat lower than that beyond configuration 26 for short distances beyond the Hevimet but was raised so that curves of the data for the two configurations approxi- mately coincided at greater distances. This pos- sibly indicates that the effect of the boral be- comes negligible at these distances. Configurations 28, 29, 30, and 31. — The thermal- neutron flux measurements beyond configuration 28 (Fig. 5.1.7) show the effect of adding 12 in. of lithium hydride behind configuration 27. The flux was lowered by about a factor of 2 immediately following the lithium hydride and approximately a factor of 2.7 at a distance of 100 cm from the source plate. This indicates a softer neutron spectrum. The thermal-neutron measurements beyond a configuration having an additional 12-in. SEenTT™ 10° 5 : Yy in, OF Li-Mg ALLOY + o 4in. OF Be + Y2 in. OF + 4in, OF HEVIMET + B~ 10* ’g 5 g 2 % 103 o § 5 v 5 2 3 10° o = 5 5 J e = 10 g s & CONFIGURATION 28: 4 in.OF @ 2] ALLOY +4in.OF Be + Y2 in. OF . T | BORAL + 4in. OF HEVIMET + 3 12in. OF LiH + B-Hp0 = 5 & = A o CONFIGURATION 29: Y, in. OF Li-Mg S ALLOY + 4in.OF Be +¥,in. OF BORAL +4in. OF HEVIMET +24in. OF LiH + B~H,0 2 - 2 30 40 50 60 70 8C 90 {00 110 120 130 © z,, DISTANCE FROM SOURCE (cm) o 10 'Fig.r 5.71.77.' Thermal-Neutron Fluxes Beyond Con- figurations 27, 28, and 29. {Secred) PERIOD ENDING SEPTEMBER 30, 1957 thickness of lithium hydride (configuration 29) appear to be inconclusive owing to the decreased thermal-neutron flux and to instrumentation dif- ficulties; no thermal-neutron data were obtainable for configurations 30 and 31. Configurations 32, 33, and 34. — The effect of placing a l-in.-thick stainless steel pressure shell beyond configuration 27 was studied in configuration 32, The change in the slope of the thermal-neutron flux curve may be partially ex- plained by the fact that steel is less effective as a thermal-neutron shield than is borated water. Also, the iron in the steel has a dip in its cross- section curve for 23-kev neutrons, which allows a large number of neutrons of this energy to escape into the borated water where they are subsequently thermalized. These effects would cause an in- crease in the thermal-neutron flux in the forward region of the curve. On the other hand, scattering of fast neutrons in the steel increases the prob- ability for their absorption in the region further from the source and causes the curve to go below that of configuration 27 in this region. In configuration 33 the effect of increasing the thickness of the beryilium reflector-moderator used in configuration 32 was studied. The addition of 4 in. of beryllium resulted in an increase in the thermal-neutron flux near the shield surface be- cause of the large percentage of thermalization of neutrons in the beryllium. The flux dropped in the region further away from the source because of the capture of these same neutrons by the boron and hydrogen. Configuration 34 was an attempt to determine the effect of reducing the thickness of the Hevimet gaomma-ray shield used in configuration 32. Sub- tracting 2 in. of Hevimet allowed on additional 2 in. of borated water to be added to the back of the shield, which resulted in a decrease of the thermal-neutron flux at short distances out in the borated water. The flux increased, however, at greater distances and was slightly above that of configuration 32. The results for configurations 32, 33, and 34 are compared with those for con- figuration 27 in Fig. 5.1.8. 301 ANP PROJECT PROGRESS REPORT 302 SECRER 2-01-057-70-378 [¢2] 10 5 ATION 34: Y%4in. OF Li-Mg ALLOY + 4in.OF Be + Y% in.OF BORAL + 2in. OF HEVIMET % +in. OF STAINLESS STEEL + B-H,0 10 CONFIGURATION 27: Y, in.OF Li-Mg ALLOY S +4in. OF Be + Yin. OF BORAL + 4in. OF = 2 HEVIMET + B- S . 4 210 O 8 9 CONFIGURATION 32: Y in. OF Li-Mg o ALLOY + 4in.OF Be +% in. OF BORAL S 3 +4in.OF HEVIMET +1in. OF ¢ 10 STAINLESS STEEL + B-H,0 o = ..:_’ ~ - 2 % 10 2 e . J = e 2 5 10 Lt < 5 g g 2 oo L E 5 CONFIGURATION 33: ¥, in. OF Li-Mg ALLOY 2 + 8in. OF Be + Y%, in. OF BORAL + 4in.OF 10! = HEVIMET + tin. OF STAINLESS STEEL + B-H,0 5 2 1072 0O 10 20 30 40 50 60 70 80 90 {00 HO {20 130 2y, DISTANCE FROM SOURCGE (cm) Fig. 5.1.8. Thermal-Neutron Fluxes Beyond Configurations 27, 32, 33, ond 34. PERIOD ENDING SEPTEMBER 30, 1957 5.2. BULK SHIELDING FACILITY F. C. Maienschein THE SUPPRESSION OF CAPTURE GAMMA RAYS BY LITHIUM IN LEAD J. D. Kington One of the problems that confront reactor shield designers is that of shielding against secondary gamma rays produced by the capture of thermal neutrons in lead. A search of the litercture re- vealed that a process for fabricating a lithium- lead alloy was patented in Germany in the 1930’s and that the eutectic contained 0.67 wt % lithium. A lead sample containing 0.67 wt % lithium was then fabricated by the Metallurgy Division, and a preliminary experiment at the Bulk Shielding Facility was performed in order to investigate the effectiveness of the lithium in reducing the cap- ture gamma-ray production in the lead. It was found that lithium distributed throughout lead was approximately as effective as boral covers on a pure lead shield. 303 ANP PROJECT PROGRESS REPORT 5.3. TOWER SHIELDING REACTOR-Il C. E. Clifford MECHANICAL DESIGN' A method of suspension for the TSR-Il that will allow a beam of radiation from a collimator through its shield to sweep both a vertical plane and o horizontal plane has been chosen and the design is being examined from structural and stability standpoints. In order to permit rotation of the beam in the vertical direction it was necessary to reduce the length of the reactor tank from that used in previous designs. The relative dimensions of the TSR-Hl, as now designed, are shown in Fig. 5.3.1. NUCLEAR CALCULATIONS? As was reported previously,3 a nuclear parameter study was carried out with the 3G3R code on the Oracle to establish the dimensions of the TSR-II. The resulting configuration had a 17.5-in.-dia internal water reflector, a 5.5-in.-thick core region, a 7.875-in.-thick external water region, and an aluminum-to-water volume ratio of 0.57. Similar calculations have now been performed on the UNIVAC with the Murine code. A comparison of the values of concentration vs reactivity from the two calculations revealed that @ change of 0.3% in the thermal-neutron macroscopic absorption cross section, % o for the Oracle calculations was all that was required to obtain complete agreement. Results of the Oracle calculations with the revised value of X are compared with the UNIVAC cal- culations in Fig. 5.3.2. The agreement is such that. control problems and the effects of various shells in the design are to be investigated almost entirely with the Oracle 3G3R code, and only the final results will be checked by a repeat calcula- tion with another code. Calculations have also been performed to de- termine the total amount of reactivity that could be controlled with a boral shell in an internal water reflector. An ideal case was used with the dimen- sions given above. A boral shell was placed at Mhe en ineering design of the TSR-1l is being per- formed by the ORNL Engineering Department. 2The nuclear calculations for the TSR-Il are being performed by M. E. LaVerne. 3¢. E. Clifford and L. B. Holland, ANP Quar. Prog. Rep. March 31, 1957, ORNL-2274 (Part 1=5), p 294. 304 L. B. Holland the core-internal reflector interface and the trans- mission of thermal neutrons through the shell was varied from 0 to 100%. With zero transmission of thermal neutrons the reactivity that could be con- trolled was found to be 11%. For these calcula- tions the transmission of fast neutrons was as- sumed to be 100%, and thus the result should be conservative, The change in the thermal-neutron flux produced by the insertion of the zero trans- mission shell is shown in Fig. 5.3.3. FUEL ELEMENT DEVELOPMENT? The fuel element development work is near com- pletion. Of the 71 different fuel plate sizes re- quired for the two types of elements, 60 plates with depleted fuel have been rolled to the correct shape. It is necessary to determine the proper rolling sequences to produce plates with the proper thickness and with length and width both held closer than the usual MTR tolerances. Delivery of a three-cylinder hand roller sufficiently long to accommodate the longer fuel plates of the annular elements has made it possible to fabricate several dummy annular elements. Inspection of these elements shows them to be well within the toler- ance necessary for operation in the TSR-1l. One of the annular elements is shown in Fig. 5.3.4. Two views of a quarter section of the reactor core formed with the dummy aluminum elements are shown in Fig. 5.3.5. DEVELOPMENT OF THE CONTROL DEVICE® Improvements were made in the TSR-Il prototype control mechanism and a new method was employed to measure the distance traveled by the control plate under scram conditions as a function of time after scram. The design changes were of a minor nature but tended to reduce friction in the posi- tioning device and guiding system. The effect of the changes was to bring the operating pressures more in line with the calculated values. As 4The fuel element development work is being per- formed by the ORNL. Metallurgy Division. 5The control system for the TSR-1l is being developed by the Reactor Controls Group. (‘ | 14 " CONTROL MECHANISM FISSION CHAMBER POSITIONING MOTCR CONTROL MECHANISM POSITIONING DRIVE MOTOR HANDLING LUGS FISSION CHAMBER ——— ] .:i CONTROL MECHANISM POSITIONING DEVICE A ALUMINUM BORAL Fig. 5.3.1. Proposed Tower Shielding Reactor T 1] | q L) ok Fo b i i I K3 ‘ =1 PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED 2-04-060-1-R2 WATER INLET | WATER OUTLET [ | ] | ¥ D] o R R :.\\:s:a"\\.\ Ry —=——REACTOR TANK H—— IONIZATION CHAMBER INTERNAL REFLEGTOR CORE REGION LEAD I1 {(Vertical Section). 305 ANP PROJECT PROGRESS REPORT 19 - UNCLASSIFIED {(x 10" 2—04—~060—27 '2 T T e UNIVAC CALCULATIONS @ — ® ORACLE CALCULATIONS [ ———— ~§ P e 6 / REACTOR GEOMETRY: 0.57 i " ALUMINUM-TO-WATER VOLUME 5 - RATIO, 445-cm~dia INTERNAL —| - WATER _REF LECTOR, _lS-cm-THICK CORE, '25-cm-THICK EXTERNAL WATER REFLECTOR . / 06 [o)rg 08 09 10 14 12 3 REACTIVITY, & U23® CONCENTRATION (atoms /cc) Fig. 5.3.2. Concentration of U235 in the Proposed TSR-1l as a Function of Reactivity. Reactor geometry: 0.57 aluminum-to-water volume ratio, 44.5cm-dia internal * water reflector, 15-cm-thick core, 25-cm-thick external water reflector. previously reported,® a 66-psi line pressure should be sufficient to hold the poppet against the posi- tioning device for normal operation; reducing the pressure below 50 psi should allow the spring force to overcome the pressure in the cylinder, and the mechanism would move away from the posi- tioning device. (In the reactor the control plates would then move toward the fuel.) The actual pressures observed with the design improvements were 72 psi to operate the mechanism satisfactorily ond less than 53 psi to permit the spring to over- come the hydraulic force. 8¢, E. Clifford and L. B, Holland, ANP Quar. Prog. Rep. June 30, 1957, ORNL-2340 (Part 1-5), p 323. 306 UNCLASSIFIED - 2=-04~—060—25 BN | INTERNAL CORE _+exTERNAL__{- | B ™ REFLECTOR ] REGION | REFLECTOR |- S ......... RADIUS {(cm) (@) No Boral Shell. ' INTERNAL ’ CORE , EXTERNAL_‘_i : REFLECTOR REGION . REFLECTOR : .ave. ..... ......... .................. RADIUS (cm) (6) Thermolly Black Shell at Interface of Core and Internal Reflector. Fig. 5.3.3. Effect of a Boral Shell at the Core ~Internal Reflector Interface on the Neutron Flux in the Proposed TSR.Il. 3 PERIOD ENDING SEPTEMBER 30, 1957 8 UNCLASSIFIED 57 ¥ Fig. 5.3.4: TSR-11 Dummy Annular Fuel Element. The control-plate travel as a function of time under scram conditions was obtained by observing, with an oscilloscope which has a persistent screen, the time required for the piston and control plate to traverse a known part of its travel. The time ~was defined by having the control. p!ate break an electrical ‘contact when motion started and then make electrical contact ‘when the desired. distance had been _covered. peared as pips on the sweep sngnul moving across the’ oscnlloscope screen. The. values of distonce traveled as a function of time thus obtained are - shown in Fig. 5.3.6, fogether with the same values obtained - by an’ iterative solution “of a force- ‘balance equafion that was written to design the ‘prototype. A full-scale model is being desngned on the same prmmples. ' The change in contact ap- CALCULATION OF THE GAMMA-RAY HEATING IN THE LEAD SHIELD W. W. Dunn’ D. K. Trubey Calculations have been carried out in order to - predict the gamma-ray heating in the lead region around the TSR-ll core. For, these calculations, the core was assumed to contain 9 kg of U233 and to have an aluminum-to-water core volume ratio of 0.57. The caleulations involved two steps: (1) o “caleulation of the heating in the core at the core- shield interface, and (2) a Monte Carlo Oracle calculation of the heating in_arbitrarily chosen regions {as shown in Fig. 5.3.7) throughout the lead shleld The calculated heating in the core at 70n assignment from USAF, 307 ANP PROJECT PROGRESS REPORT Rl UNCLASSIFIED| ¥-23553 () INSIDE VIEW [ERE. N JHCLASSIFIED ¥-23548 o L3 itlt”lzll'llll ) (5) EXTERNAL VIEW Fig. 5.3.5. Dummy Fuel Elements for the Proposed TSR-I1, Assembled in Quarter Sphere. the core-shield interface, coupled with an as- sumption for the angular distribution at the inter- face, was used as the source term for the Monte Carlo calculation. The two steps are described in detail below. Heating at the Core-Shield Interface For the calculation of the core heating at the core-shield interface, the core was divided into seven spherical shells (see Fig. 5.3.7) chosen so that a reasonable average flux could be assumed for each shell. The source distribution was as- sumed to be proportional to the thermal-neutron flux as calculated by the Oracle 3G3R code for an aluminum-to-water volume ratio of 0.707 and core dimensions of 20 and 40 cm. This flux plot was scaled to the assumed core dimensions (22.225-cm inner radius and 37.46-cm outer radius) and was 308 vused for the calculation, although the aluminum- to-water volume ratio in the calculation was specified as 0.57. The gomma-ray spectrum from each source shell was then divided into six energy groups, which were arbitrarily chosen as 0.5, 1, 2, 4, 6, and 8 Mev. According to the usual transformation from a spherical-shell source to an equivalent infinite plane source (neglecting the back plane),® the core heating at the core-shield interface as a function of the initial energy E, leaving the source 8This transformation has been described by E. P. Blizard; see, for example, S. Glasstone, Principles of Nuclear Reactor Engineering, p 592, Van Nostrand, New York, 1955. In summary, H(shell) = (r/ro) X [H(infinite plane ot rg = 1) ~ H(infinite plane at rg + Il o 9 UNCLASSIFIED 2—01—-060-30 0.9 //—"//’ 0.8 / 0.7 // 0.6 CALCULATIONS ~ / ]/ 0.5 / [ DISPLACEMENT (in.} ~~ LABORATORY DATA e / i/ 0.2 / s ~ 0 5 t0 5 20 25 30 35 40 TIME {(msec) Fig. 5.3.6. Piston Displacement in Control Mechanism Test for Proposed TSR-Il as a Function of Tirne. plane was determined by the equation ' Sr - H(shell,Eo) := -".; [21TKS(E0) Eqy Fa(EO)] X x [ B(REq) GIR,Eg) R dR 3 where rq = distance - from center - of core (or center of internal water reflecfor) to . core-shield interface, cm, : r = distance from center of core to . spherical source shell cm, ' .z = o — r,cm, S _S(Eo) - . - gamma rays of energy E, generated per square cennmeter per: hssron' : | ‘per second, - | - (Ey) = energy ubsorphon | coefficient» for gamma rays of energy E, in the core, source term for each sheH number of, PERIOD ENDING SEPTEMBER 30, 1957 "R = distance from the source point on the spherical shell to the core-shield interface, cm, B(R,E,) = energy absorption buildup factor in the core as a function of distance R and source energy E,, G(R,Ey) = oftenuation kernel for a point iso- fropic source = e-#'R/47TR2, #, = total absorption coefficient, K = term to convert to watts of heating per cubic centimeter of material per megawatt of reactor power. The correction for the heating in the 7, + r infinite plane, that is, the back plane, was made in the actual calculation, but it amounted to only a small per cent and then only in the innermost core source planes. The source term, S(E;), which was ussumed to be proportional to the average thermal-neutron flux in that shell, can be given by the expression @ S(Eg) = Ayt [N + o) u) + + Nfp) ZA2) + N(AI) Z(A)) + + NH) X H)] , where A = factor for normalization to one fis- . sion per second in the core, $,, = average thermal-neutron flux in the source shell, t = shell thickness, cm, N(p + c) = number of prompt fission and capture gamma rays inthe energy group about E, produced in uranium per fission, 2/(") = macroscopic fission cross section of U235, _ number of fission-product decay gamma rays in the energy group ‘about E, per fission, N(Al) = number of aluminum capture and de- ' cay gamma rays in the energy group about Eo per capture, 2 (A1) = macroscopic absorption cross sec- ' tion of aluminum, : N(fp) 309 ANP PROJECT PROGRESS REPORT BORAL INTERNAL WATER REFLECTOR 315‘“\- UNCLASSIFIED 2—01—-060-28 123 4 5 67135 iom._v o2 . 46 11 NOT TO SCALE IN " Y v . ‘ 1 ‘ _ /a-in. Jq—in. Yo-in. CORE REGION SPACES SPACES SPACES Fig. 5.3.7. Shield Configurction Used to Calculate Gamma-Ray Heating in the TSR-Il. N(H) = number of hydrogen capture gamma rays in the energy group about E, per capture, Z (H) = macroscopic absorption cross sec- tion of hydrogen. The values used for the N's as a function of E, are listed in Table 5.3.1. 310 The linear and energy absorption coefficients of the core as a function of E, were determined by the following expression, where p, is the fractional density of the ith material, ' K Ficore) = E(_) Pi - i \P /i o " Toble 5.3.1. Equivalent Number of Gamma Rays In the Core for Each Energy Group E, PERIOD ENDING SEPTEMBER 30, 1957 (i?') N(p + c)/fission? N(fp)/fission® N(H)/captureb N(Al)/capfureb 0.5 3.31 2.972 0 0 1 2.55 1.864 0 0.10 2 1.687 0.89 1L.115 1.53 4 0.368 0.11465 0 0.77 6 0.050 0.0084 o 0.21 8 0.0067 0.0006 0 0.35 “H. W. Bertini et al., Basic Gamma-Ray Data for ART Heat Depo&ition Calculations, ORNL-2113, pp 8,16 (Sept. 17, 1956). b4, C. Claiborne and T. B. Fowler, The Calculation of Gamma Heating in Reactors of Rectanguloid Geometry, ORNL CF-56-7-97, p 20 (July 20, 1956). The energy absorption buildup factors for each E, in the core were obtained by fitting the ex- pression: B(R) = 1 + du) + blun)? to the tabulated NDA values for aluminum and water.? Where possible, the empirical expression was weighted as the core density ratio, two-thirds aluminum to one-third water. The values of 4 and b for each value of E,, which are good up to ap- proximately 7 mean free paths, ore given in Table 5.3.2. The core heating at the core-shield interface was obtained by summing the result of Eq. 1 over all seven core planes. The results are shown in the second column of Table 5.3.3. Because anofher method of cclculaflon was used - for. the second step, it was felt that the heating in some overlappmg region should be computed by ~ both methods for purposes of comparison. There- fore, the method described above was also used to. _determine the heating in the shield at the first 'alummum ‘shell and in. the initial layer of lead. ~This was uccompllshed for the aluminum shell by assuming that the flux entering the aluminum was ,'th"e"éqm'e_ usfi thct-whi ch left ihe cbre. The oluminum 9H. Go!dstem und J E Wllkms, Jr., Calculauons o/. the Penetrations o Gamma Rays. Final Report, NYO- 3075 (June 30, 1954). - Toble 5,3,2. Values of the Coefficlents a and b Used in Determination of B(R) for the Core Eo a b 0.5 1.10 0.45 1 0.955 0.185 2 0.78 0.05 4 0.57 0.0 6 0.42 0.0 8 0.33 0.0 Table 5.3.3. Heat Contributions from Eqy Energy Groups Heat Generation (fiafis/cma-Mw)_ In First Ey | Ar In First (Mev) Core-Shield Aluminum Lead ' ' interface Shell interval 0.5 0.034 0.047 0.649 1 0.042 0.063 0.389 2 0.050 0.074 - 0.360 . 0.017 0.026 0.251 6 0.003 . 0.005 0.055 8 0.002 0.003 0.025 Total 0.148 . 0.218 - 1L729 31 ANP PROJECT PROGRESS REPORT heating was then the product of the heating.per energy group in the core and the ratio of the aluminum-to-core energy absorption coefficients. The results are tabulated in the third column of Table 5.3.3. For the lead heating, Eq. 1 was resolved for the heating in the aluminum at the aluminum-lead interface, by using the new distances involved and by including the effect of the thin shell of alumi- num on the linear and energy absorption coef- ficients. The buildups in the core and aluminum were assumed to be the same. The resulting heat- ing as a function of the energy group (which agreed to within several per cent of that estimated in the preceding paragraph) was multiplied by the ratio of the lead-to-aluminum energy absorption coefficients to give the results tabulated in the fourth column of Table 5.3.3. Naturally, it is not expected that this estimate for lead will be as good as that for the aluminum shell, and no reli- ance is placed on it other than for comparison with the calculation discussed below. ) Hicollided) = [¢4-Me§ (4 Mev) + &9 y0 collided collided Since the Monte Carlo code yields gamma-ray heating as a function of initial energy and angle of incidence, it was necessary to find the energy and angular distribution of the flux incident upon the shield. The energy flux incident upon the shield was estimated by dividing the calculated heating for each initial energy group at the core-aluminum interface into collided and uncollided fluxes. The uncollided flux for each E, was determined by solving Eq. 1 without buildup factors or absorption coefficients. The total collided heating for each E in the core was found by: Hicollided) = Hiuncollided) [B(R) = 11 . This was assumed to be made up of collided fluxes of the energy group E, under consideration plus collided fluxes of each lower energy group E cor- responding in size, except for the upper group, to the initial groups E,. For example, for E, at 4 Mev: Ho(2 Mev) + &y oy #5(1 Mev) + collided + ¢0-5-Mev ua(O.S Mev) + @9 1 .Mev £ (0-1 Mev) | K . Heating Within the Shield The Oracle Monte Carlo code'® for the heat de- position resuiting from the penetration of gamma rays through stratified slab shields was used to calculate the heating in various TSR-1l shield layers which were arbitrarily chosen as follows: 1. the first layer of aluminum, 2. the six equal regions in the first layer of lead, 3. the second aluminum layer, the water layer, the third aluminum layer, the five equal regions in the second lead layer, the two equal regions in the final borated aluminum layers which were considered all aluminum, These layers are shown in pictorial form in Fig. 5.3.7. N A 10¢c. D. Zerby and S. Auslender, ANP Quar. Prog. Rep. Dec. 31, 1955, ORNL-2012 (Part 1-3), p 203. 312 collided collided The relative values of the collided fluxes in the above equation were obtained from the ratios of the areas under the differential energy spectrum curve for each E,, as given by Goldstein and Wilkins? for a point isotropic source after passage through one mean free path of water. This was possible since the relative spectrum for each E; is nearly independent of the distance from the source for a number of mean free paths. Figure 5.3.8 illustrates such a curve for an E, of 4 Mev. With the ratios of the collided fluxes for each E known, it was pos- sible to solve Eq. 3 and obtain the collided num- ber flux for each energy group. The sum of the collided and uncollided fluxes for each E gave the total incident flux for each energy group. It was assumed for the uncollided fluxes that the radiation was isotropic in the source planes and, since the back plane correction is unimportant, the angular distribution should also be nearly the same as for infinite plane sources. This angulor 9 41rrze"°rlo PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED 175 i" 2-01{-060-34 ® 150 - - T o o WATER o ALUMINUM .25 A WATER ! A ALUMINUM [ fror=2 flN / ’/ A 1.00 ' / | /f/ I ../ Hor =1 0.50 I \ =T ' \‘ /:fl?’ nfil;-,. ] #:’:‘:—-"—’:fi a a == 0.25 g % {=-Mev 2-Mev 4-Mev |l —— jentlf ——— — Sttt ————— - —— T é “"g > GROUP GROUP GROUP -~ 0 o o 0 l L 1 0 05 1.0 1.5 2.0 2.5 3.0 35 4.0 45 5.0 £ (Mev) Fig. 5.3.8. NDA Moments Method Calculation of Collided Gamma & & Z 0400 g 0400 - g 60° g 60° = Q 2 oo7s 2 0075 5 5 = 2 2 ' 2 70° § 0.050 o G 0.050 g & 0.025 80° 0025 80° ace ot 90° .'A . ,Elg.,".5.4.l.=:- F'lh'c'(:t_iorln of ,lr‘:cfrdént_' Energy Tfuhsmi"ed (Collided) per Unit Solid Angle: 0.662-Mev Gomma Rays No_rrnul_ly’ incident on _d 3-in.-T_hIc Sltab. ' - k Aluminum ~ Fig. 5.4.2. Fraction of Incident Enefgf Trensmitted (Collided) per Unit Solid Angle: Slab. 0.662-Mev Gamma Rays Normally Incident on o &in.-Thick Aluminum 319 ANP PROJECT PROGRESS REPORT UNCLASSIFIED 2-01-059-226 i0* 20° 30° 40° o Q.030 0.025 g O FRACTION OF INCIDENT ENERGY o o o o o & 0.005 80 20 Fig. 5.4.3. Fracticn of Incident Energy Tronsmitted (Ceollided) per Unit Solid Angle: 0.662-Mev Gamma Roys Normally Incident on a 9-in.-Thick Aluminum S|ub. where u = cos 0 and N = the total collided energy transmitted. Equation 2 reduces to 2rA 3) b + 1 and Eq. 1 reduces to (4) NE*' — b = ¢, i + 1) or $i, i + 1) min = w7 - N for py = 1, py4 = 0; therefore =V G, i+ 1 5 b =1-F q“__rvll- =1 , forj# 1,13 . Equation 5 may then be used to get 11 estimates of b, It was found that b did not vary greatly, and so an average was used to determine A from Eq. 3, The resulting equations, ¢ = A cos® 6, are also shown in Figs. 5.4.1 through 5.4.3 and indicate no change in the angular distribution with slab thick- ness. 320 MONTE CARLO CALCULATION OF THE GAMMA- RAY PENETRATION OF LEAD-WATER SLABS L. A, Bowman? D. K. Trubey The DO2C55 series of calculations of the gamma- ray penetration of lead-water slab shields, described previously,! has been continued, and a total of 512 problems has now been computed. All combi- nations of the following parameters have been used: slab thicknesses of 1, 2, 4, and 6 mean free paths, consisting of eight varying percentages of lead preceding or following water; initial gamma-ray energies of 1, 3, 6, and 10 Mev; and angles of gamma-ray incidence of 0, 60, 70.5, and 75.5 deg from the normal. The results include information on heating, dose rates, and energy fluxes through- out the slabs; energy and angular distributions reflected back by the slabs; and energy and angular distributions transmitted through the slabs. The data from all these problems are now being analyzed. Typical plots are shown in Figs. 5.4.4 through 5.4,7, which represent the heating caused by 3-Mev gamma rays incident at various angles on lead slabs 4 mean free paths thick., Note that the heating values in Figs. 5.4.6 and 5.4.7 are re- duced by a factor of 2 for ease in plotting. MONTE CARLO CALCULATION OF GAMMA-RAY DOSE RATE BUILDUP FACTORS FOR LEAD AND WATER SHIELDS L. A, Bowman? D. K. Trubey The DO2C55 series of Monte Carlo calculations discussed in the preceding paper also included calculations of dose-rate buildup factors for lead and water shields, sketches of which are presented in Fig. 5.4.8, Only the results for gamma rays incident normal to the slabs are reported here. In the first group of calculations buildup factors were determined for all-lead slabs which were 1, 2, 4, and 6 mean free paths in thickness at the initial gamma-ray energy. The initial energies were 1, 3, 6, and 10 Mev. The results are compared in Fig. 5.4.9 with the buildup factors computed at NDA by the moments method for an infinitely thick slab, The results are in agreement except near the rear of the slabs, where the Monte Carlo curves are lower, This is due to the reduction in back- scattering in the slabs having finite thicknesses, 2On assignment from USAF. i 2.00 1.80 1.60 1.40 1.20 1.00 0.80 0.60 0.40 J, FRACTION OF TOTAL INCIDENT GAMMA~RAY ENERGY ABSORBED PER MEAN FREE PATH 0.20 0 PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED 2—-01-059-249 CODE NO.: 3414 REFLECTED: 0.056 TRANSMITTED: 3.869 ABSORBED:96.075 Eo = 3 Mev 8= 0 deg LEAD 0 o 2 3 4 o, NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.4. Energy Absorption in a Lead Shield as a Function of the Shield Thickness: 3-Mev Gamma Rays Normally Incident, 321 ANP PROJECT PROGRESS REPORT UNCLASSIFIED 2-01-059-220 2.00 CODE NO : 3442 REFLECTED: 0.759 ) TRANSMITTED: 0.148 1.80 ABSORBED: 99.093 Eo= 3 Mev 6= 60 deg 1.60 _ LEAD 1.40 .20 1.00 0.80 0.60 0.40 J, FRACTION OF TOTAL INCIDENT GAMMA-RAY ENERGY ABSORBED PER MEAN FREE PATH 0.20 0 0 1 2 3 4 Lor, NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.5. Energy Absorption in a Lead Shield as a Function of the Shield Thickness: 3-Mev Gamma Rays Incident at 60 deg. ‘ j 322 2.00 1.8 o O o ~RAY ENERGY ABSORBED PER MEAN FREE PATH/2 o 0.8 0.6 J/5, FRACTION OF TOTAL INCIDENT GAMMA 0.2 0 N o D o 0 0 0.40 O, o PERIOD ENDING SEPTEMBER 30, 1957 UNCLASSIFIED 2-01-059-224 CODE NO.: 3443 REFLECTED:1.697 TRANSMITTED: 0.053 ABSORBED:98.250 Eo = 3 Mev 8 =70.5 deg LEAD 0 1 - 2 3 4 Hor, NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.6. Energy Absorption in a Lead Shield as a Function of the Shield Thickness: 3-Mev Gamma Rays incident at 70.5 deg. 323 e sl AL 12 | ANP PROJECT PROGRESS REPORT UNCLASSIFIED 2-01-059-222 2.00 CODE NO : 3414 REFLECTED: 3.350 TRANSMITTED: 0.040 1.80 ABSORBED: 96.610 Ep= - 3 Mev 8 =75.5deg 0 o LEAD 1.40 .20 {.00 0.80 0.60 0.40 J/Z, FRACTION OF TOTAL INCIDENT GAMMA-RAY ENERGY ABSORBED PER MEAN FREE PATH,, 0.20 0 0 1 2 3 4 por, NORMAL THICKNESS OF COMPOSITE SLAB IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.7. Energy Absorption in a Lead Shield as a Function of the Shield Thickness: 3-Mev Gamma Rays Incident at 75.5 deg. : S 324 ’) UNCLASSIFIED 2-04-059-238 SLAB CONFIGURATION NO. 4 H,0 H,0 TOTAL THICKNESS T gt PARAMETERS: INCIDENT ENERGY = 4,3, 6, AND 40 Mev TOTAL THICKNESS 7 = 4,2,4,AND 6 MEAN FREE PATHS sec8=14,2,3, AND 4 - Fig. 5.4.8. Lead and Water Siab Shield Configurations Used for Monte Carlo Calculations. UNCLASSIFIED 2-01-059-249 {0 ’ - - v . [ =——emme MOMENTS METHOD (NDA) FOR INFINITELY THICK SLAB MONTE CARLO METHOD (ORNL) FOR SLABS WITH FINITE THICKNESSES S 8=0deg £= INITIAL ENERGY 1 { 3 Mev_l — VERTICAL LINE INDICATES - — REAR OF FINITE SLAB—————} ELOI‘/W“-"’ oo & 8,, GAMMA-RAY DOSE-RATE BUILD-UP FACTOR 0 ¢ 2 3 4 5 6 7 8 9 0 # 12 13 ' 7, SLAB THICKNESS (cm) | Flg.lhrs.4.r9. Comparison of Dése—Rute Buildup Factors for Lead Slabs Computed by Monte Carlo Method with Those Computed by Moments Method. Normally incident gamma rays. L—‘/ /.-" 2 — 5 "-' 6Mev] ' / /‘/ —”.ijb/‘::q ‘ /I/fi/;_‘r 10 Mev S 10 PERIOD ENDING SEPTEMBER 30, 1957 The Monte Carlo buildup factors for all-water slabs of finite thickness are compared with NDA buildup factors for on infinitely thick slab in Fig. 5.4.10. In this case the Monte Carlo calcu- lations always predicted higher values for the buildup factors. Values of the Monte Carlo dose-rate buildup factors at the rear of composite lead and water slab shields are plotted in Figs. 5.4.11 through 3.4.14 for total composite slab thicknesses of 1, 2, 4, and 6 mean free paths. The effect of alter- nating the position of the lead and water is apparent from these figures. : All the buildup factors determined for composite slabs in this series have been compared with UNCLASSIFIED 50 2—-01-059-250 | | l [ | =——aae MOMENTS METHOD (NDA) FOR INFINITELY ___| THICK SLAB | MONTE CARLO METHOD (ORNL) FOR SLABS — | WITH FINITE THICKNESSES 8=0 20 Eo= INITIAL ENERGY . . e o II/VERTICAL LINE INDICATES : REAR OF FINITE SLAB =2 g 10 —AE0= ! Mev 2 o p et [+ w 8 - 3M :[- ev T a o 1 = ]%7{' € Mev - N o = % =" | 10 Mev - o 45 . 90 135 180 225 270 1,SLAB THICKNESS {cm) Fig. 5.4.10. Coméuri;on of Dose-Rate Bulldup Factors for Woter Slabs Computed by Monte Carlo Method with Those Computed by Moments Method. Normally incident gamma rays. 325 ANP PROJECT PROGRESS REPORT UNCLASSIFIED 2-01-059-254 H20 Hz0 Pb Eq, INITIAL 7= tmfp ENERGY O ~m===ee-- { Mev 3 Mev A ————— § Mev & ———-—= 40 Mev T ={imfp By, GAMMA-RAY DOSE-RATE BUILD-UP FACTOR 4 O 025 050 075 100 100 075 050 025 O ptpy/T, NORMAL THICKNESS IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.11. Monte Corlo Dose-Rate Buildup Factors at the Rear of Composite Lead-Water Slab Shields 1 Mean Free Path Thick. Normally incident gamma rays. UNCLASSIFIED 2-01—-059-253 H,0 L’T = 4mfp 3 h ----- == { Mev . 3 Mev b =———— G Mev A ——==—10 Mev 8,, GAMMA-RAY DOSE-RATE BUILD-UP FACTOR n { 0 025 050 075 +t00 1.00 075 050 025 O 2oy / 7, NORMAL THICKNESS (N MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.13. Monte Carlo Dose-Rate Buildup Factor at the Reor of Composite Lead-Water Slab Shields 4 Mean Free Paths Thick. Normally incident gamma Rays. 326 UNCLASSIFIED 2-01-059-2%2 S Hp0 o ’ ENERGY 0 =eemmme==n | Mev ® 3 Mev A =——= 6 Mev A ————-—— {0 Mev N By, GAMMA-RAY DOSE-RATE BUILD-UP FACTOR \ i 0 025 050 075 100 100 075 050 025 O pfpb/r, NORMAL THICKNESS IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.12. Monte Carlo Dose-Rate Buildup Factor at the Rear of Composite Lead-Water Slab Shields 2 Mean Free Paths Thick. Normally incident gamma rays. " UNCLASSIFIED 2—01-059—254 8 =0deg H,0 H,0 10 _ £, INITIAL I___ 7=6mb= ENERGY T=6 ommm=-== | Mev o—— 3 Mev 5 6 Mev \ A=——-—10 Mev 5., GAMMA-RAY DOSE-RATE BUILD-UP FACTOR 1 0 025 050 075 1.00 100 075 050 025 O Bl /T, NORMAL THICKNESS IN MEAN FREE PATHS AT INITIAL ENERGY Fig. 5.4.14. Monte Carlo Dose-Rate Buildup Factor at the Rear of Composite Lead-Water Slgb Shields 6 Mean Free Paths Thick. Normally incident gamma rays. ) C PERIOD ENDING SEPTEMBER 30, 1957 values obtained by use of o formula, proposed by X, X, = thickness in mean free paths of the M. H. Kalos of NDA3 in which the buildup factors first and second materials, respec- independently computed for lead and water are tively. ' combined to determine the buildup factor for a . i . For a water-lead shield the formula is: composite slab consisting of the two materials, B,(x,) -1 - (flcs/p') - 1 1.7X, 4 = __._.........'I (1 - e Xz) [Bz(Xl +X2) - Bg(x])] ’ B(X.,X,) = B.,(X.,) + | ————— ¢ 1402 2v" 2 : B,y(X,) -1 | (/1) where For a lead-water slab ('IhO'l’ is, lead followed by Bes = Compfon scattering cross secfion, water) the formula is written as follows: i . p, = total cross section, B(X],Xz) = BZ(XZ) + Tables 5.4.6 and 5.4.7 show the buildup factors determined both by Oracle calculations and by the Kales formula, It should be pointed out that the numbers representing the Oracle calculations are . taken from the curves of Figs. 5.4.11 through where B o , 5.4.14 and do not represent actual data points, B,(X,) - 1 : +W[Bz(xl + X,) - Bz(xz)] ' B,,B, = gamma-ray dose-rate buildup factors 3M. H. Kalos, Some Theoretical Methods and Results . , s in Gamma Ray and Neutron Shielding, Shielding Sym- for the first and second materials, re- posium held at the Naval Radiologicuf Defense Labora- spectively, - tory, Oct. 1719, 1956, 12NDP 1965, p 93-94. Table 5.4.6, Monte Carlo Gamma-Ray Dose Rote Buildup Factors at Rear of Lead=-Water Slab Shields: Comparison of Oracle Calculations with Values Obtained with Kalos Formula Incident Shield (mfp) B Gamma-Ray —_— _ T Ratio of B, (Calculated) Energy Lead Water _Oracle Kalos to B_ (Kalos) - Cadlculations Formula r {(Mev) 1 1 1 233 2.18 1.069 2 3.30 3.13 1.054 4 5.81 - 5,60 1.038 | 5 7.51 7.14 1.052 2 1 2.72 2,51 1.084 2 3,90 : 3.47 1.124 3 533 4.73 o aZ , 4 695 6,06 , 147 3 1 300 3.00 1.033 2 4.50 4,05 1.1 3 6,22 5.36 1.160 4 2 327 ANP PROJECT PROGRESS REPORT Table 5.4.6 (continued) Incident Shield (mfp) B, e SR Gamma-Ray _ . Ratio of B, (Calculated) - Energy Lead Water Oracle Kalos to B, (Kalos) (Mev) Calculations Formula . 3 1 1 1.95 1.90 1.026 ' 2 . 2,58 2,54 1.016 3 3.30 3.10 1.065 4 3.80 3.60 1,056 5 4.2 4.07 1.034 2 ] 2.47 2.45 1.004 2 3.15 3.00 1,050 3 3.69 3.52 1.048 4 4.18 4,03 1.b37 3 ] 2,80 2.85 0.982 2 3.44 2.89 1.190 3 4.00 3.47 1.153 6 1 1 1.60 1.55 1.032 2 2.03 1.98 1.025 3 2.50 2.43 1.029 4 2.66 2.67 0.996 5 2.71 2.74 0.989 2 1 1.98 1.80 1.100 2 2.34 2.14 1.093 3 2.55 2.50 1.020 4 2.62 2.70 0.970 3 1 2.11 1.96 1.077 2 2.31 2.18 1.060 3 2.41 2.52 0.956 10 1 1 1.41 1.38 1.022 2 1.68 1.64 1.024 3 1.90 1.88 1.0Nn 4 2.04 2.03 1.005 5 2.14 2.12 1.009 2 1 1.55 1.50 1.033 2 1.78 1.72 1.035 3 1.97 1.93 1.021 4 2.10 2,07 1.014 3 1 1.65 1.65 1.000 2 1.81 1.84 0.984 3 1.94 2.02 0.960 328 g [ 3 Fbovom s PERIOD ENDING SEPTEMBER 30, 1957 Toble 5.4.7. Monte Carlo Gamma-Ray Dose Rate Buildup Factors at Rear of Water-Lead Slab Shields: Comparison of Oracle Calculations with Valuves Obtained with Kalos Formula Incident Shield (mfp) B, Gamma-Ray Ratio of B, (Calculated) Energy Water Lead 0"‘":'? Kalos to B, (Kalos) (Mev) Calculations Formula 1 1 1 1.70 1.66 1.024 2 1.86 1.92 0.969 3 2.05 2.12 0.967 4 2.22 2.35 0.945 5 2.42 2.55 0.949 2 1 2.20 1.96 1.122 2 2.18 2.17 1.005 3 2.31 2.45 0.943 4 2.52 2,63 0.958 3 1 2.50 2.27 1.101 2 2,46 2.50 0.984 3 2.72 2.73 0.996 4 2 3 1 1 1.74 1.62 1.074 2 1.92 1.88 1.021 3 2.15 2.16 0.995 4 2.39 2.44 0.980 5 2.66 2.68 0.993 2 1 2.28 1.96 1.163 2 2.39 2.24 1.067 3 2.51 2.54 0.988 4 2.70 2.78 0.971 3 1 2.80 2.33 1.202 2 2.88 2.60 1.108 3 2.97 2.88 1.031 6 1 1 1.55 1.50 1.033 2 1.75 1.79 0.978 3 2.01 2.04 0.985 4 2.20 2.27 0.969 5 2.40 2.56 - 0.938 2 1 2.09 1.89 1.106 2 2.24 2.23 1.004 3 2.38 2.50 0.952 4 2.48 2.82 0.879 3 1 2.48 2,31 1.074 2 2.58 2.68 0.963 3 2.61 3.04 0.859 wa - . e, T = o 4 e 329 ‘Table 5.4.7 (continved) hcident B Shield (mfp) , Gamma-Ray Ratic of B (Calculated) Energy Water Lead Oracle Kales ‘to B, (Kalos) (Mev) Calculations Formula 10 1 1 1.3% 1.41 0.986 2 1.52 - 1.55 0.981 3 1.65 1.70 0.971 4 1.77 1.92 0.922 S 1.90 2.07 0.918 2 ] 1.70 1.72 0.988 2 1.79 1.92 0.932 3 1.88 2.16 0.870 4 1.96 2,39 0.820 3 1 1.95 2.00 0.975 2 1.99 2,34 0.850 3 2.08 2,61 0.797 - 330 )