“l‘”l\‘i““““\.“w—‘l“{‘\\‘lw‘ HNM\‘ I 3 yy56 D3b44LE 1 ORNL-3215 UC-80 — Reactor Technology MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT FOR PERIOD FROM MARCH 1 TO AUGUST 31, 1961 | o CENTRAL RESEARCH LIBRARY =3 > DOCUMENT COLLECTION LIBRARY LOAN COPY DO NOT TRANSFER TO ANOTHER PERSON 3 . : If you wish someone else to see this ] document, send in name with document and the library will arrange a loan. OAK RIDGE NATIONAL LABORATORY operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION L oxrel T= $2.75 Office of Technical Services Printed in USA. Price . Available from the Department of Commerce Washington 25, D.C. LEGAL NOTICE This report was prepared as an account of Government sponsored work. Neither the United States, nor the Commission, nor any person acting on behalf of the Commissi A. Mokes any warranty or represéntation, expressed or implied, with respect to the accuracy, s of the information contained in this report, or that the use of completeness, or usefuln ony information, apparatus, method, or process disclosed in this report moy not infringe privately owned rights; or e with respect 1o the use of, or for domages resulting from the use of B. Assumes any liabi any information, opporatus, method, or process disclosed in this report. As used in the above, “‘person acting on behalf of the Commission includes any employee or contractor of the Commizsion, or employee of such contractor, 1o the extent that such employee or contractor of the Commission, or employee of such contractor propares, disseminates, or provides occess to, any information pursuant 1o his employment or controct with the Commission, or his employment with such contractor. ORNL-3215 UC-80 — Reactor Technology TID-4500 (16th ed.) Contract No. W-TLO5-eng-26 MOLTEN-SALT REACTOR PROGRAM PROGRESS REPORT FOR PERIOD FROM MARCH 1 TO AUGUST 31, 1961 R. B. Briggs, Program Director Date Issued (At J N T A ‘@2 B | g SN L O u3@ OAK RIDGE NATIONAL LABORATORY Oak Ridge, Tennessee operated by UNION CARBIDE CORPORATION for the U.S. ATOMIC ENERGY COMMISSION MARTIN MARIETTA ENERGY § LT 3 Y456 O3LYyypg 1 RS SUMMARY PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE Design In order to facilitate graphite sampling and to make possible the use of solid control rods, the reactor layout was reviewed, and the pump location was moved from the top of the reactor to a position which left the top of the reactor accessible from above. This change, while not of a basic nature, necessitated a considerable amount of new design work. A "tee" section on top of the reactor vessel was designed as a port of entry for control rods and graphite samples and as an outlet for fuel. A frozen salt seal and a gasketed flange type of gas seal were designed for the entry port. Also thimbles and rods together with rod drives and a cooling system had to be designed for the control system. Moving the pump required designing a pump support structure which would allow movement of the pump as the suction and discharge lines changed length with thermal cycling of the system., OStress analysis of the entire piping system was redone, Auxiliary systems for the pump, off-gas handling system, electrical heating system layout, and all penetration designs were modified somewhat as a result of the cell piping changes. Containment cell penetrations were designed and checked. The design of the top plugs was modified to increase the minimum design pressure for the contain- ment to 40 psig. The drain-tank cooling system was simplified and improved with respect to mechanical complexity and steam flows. Major building modification design was completed and sent to prospective bidders. Component designs were completed, and a design report for these components was 1ssued. A design package consisting of the reactor vessel, storage tanks, radiator, and heat exchanger was sent to prospective bidders, Thirteen instrument-application drawings were issued for comment. Prepara- tion of these drawings is approximately 85% complete. Preparation of instrument- application tabulations is approximately 60% complete. Twelve thermocouple- locations drawings were issued for approval, Control circuit design is in a pre- liminary stage., Layout of the instrumentation and controls system is proceeding as layout of the building and equipment become firm and as instrumentation requirements become known., A front-elevation layout of the main control board and a proposal for layout of the main control area have been prepared. iii A study of MSRE data-handling requirements 1is nearing completion. A study is under way to determine the most reliable and least expensive method of providing single-~point temperature alarm channels for monitoring the operating status of freeze flanges and valves. There are no areas of design which pose unresolved problems, and the design of the MSRE is essentially on schedule. 2. Component Development Design of a 5-in. freeze flange incorporating a buffered ring-joint gas seal was completed and procurement initiated for INOR-8 forgings and soft nickel rings. A 6-in, freeze flange of Inconel was fabricated and installed in the thermal- cycle and gas-seal test facility, using a L0O-kc induction coil in the bore to produce the temperature distribution. A study of fabrication and inspection methods for the flanges and rings was initiated. Thermal cycle and seal tests for a 3-1/2-in. Inconel flange and gold-plated Inconel gasket using a spring clamp indicated that successful seals can be made by this method. Freeze valves opened by Calrod heaters and incorporating siphon breaks were installed in the engineering test loop and successfully tested under extreme operating conditions. A two-plece pipe heater insulation unit was tested and discarded because of excessive heat loss. A modified heater, better adapted to maintenance from above and designed for less heat loss, was accepted for testing. Tests conducted on the core-heater prototype verified the calculated heat loss; however, insulation shrinkage caused some trouble. The system was operated for 3000 hr without difficulty and with no apparent deterioration of the metal components. A firm proposal for the sampler enricher was accepted for testing and detall design was started. The solder freeze valve was less reliable than had been predicted and was replaced in the sampler design with a buffered-seat gate valve. The sample cap- sule development progressed to the stage of a hot-cell crushing and unloading study. A sample transport container was designed and constructed for testing a system of atmospheric control during shipment of the deoxidized capsule to the site and return of the sample to the hot cell, The drain-tank cooler was mocked up and testing was started. Flow tests in the lower head of the one~fifth-scale core model indicate a shift in the flow characteristics with reduced flow. Fuel age and heat transfer coefficient and characteristics of solid distribution were measured in this area. Efforts to produce oscillatory disturbances in the core were unsuccessful, Cormponents for the full-scale-core-model test loop were fabricated and in- stalled in preparation for the arrival of the core internmals. Helium purification tests, using titanium sponge as the oxygen getter, were used as the basis for the design of a full-scale MSRE helium purification system. The engineering test loop was operated continuously for 1300 hr; the oxide bulldup history was observed during a startup-flushing operation. Revisions made to the bubble-type level indicator made the system very reliable under engineering- test loop conditions. The design of the graphite sectlon of the salt loop to- gether with the graphite-handling dry box was completed and fabrication started. A maintenance plan using the most appropriate maintenance method for each operation was evolved. The one-twelfth-scale model was completed to the existing design stage. A full-scale maintenance mockup is being constructed for study of special problems such as freeze-flange operation. The program to develop a remotely brazed Jjoint for l~l/2-in. pipe progressed to the bench-demonstration stage, and design for tools to produce remotely brazed Joints was started. OSeveral mechanical disconnects were demonstrated which could be used in auxiliary lines, A program was initiated for the study of the steam generator problems. Development of a continuous-level element, for use in measurement of molten- salt level in the MSRE fuel and coolant salt pump bowl, is continuing. A graphite float assembly was fabricated, and a test stand was constructed. A recent transformer design for the level element was operated for more than two weeks at temperatures in excess of 1200°F without evidence of insulation failure. Performance data obtained from this transformer at temperature is very encouraging. A new design concept for a single-point conductivity-type level probe was developed and tested. A conceptual design was developed for a two-point level probe, based on the new design concept and compatible with the physical geometry of the MSRE storage tanks. Extensive measurements were made of the resistivity of a molten salt at various voltages, frequencies, and temperatures. A temperature-scanning system is being developed to a present profile dis- play of approximately 250 thermocouples attached to the reactor pipes and components. The development of techniques and procedures for attaching sheathed thermo- couples to INOR-8 pipes and vessels is continuing. Sample quantities of sheathed thermocouple wire have been procured, and several methods of attachment are being investigated. 3. Reactor Engineering Analysis The MSRE temperature coefficients of reactivity were calculated and found to be -2.8 x 1077 Ak./°F for the fuel salt, and -6.0 x 10 ° Ake/°F for the graphite. The reactivity worth of all three MSRE control rods was calculated to be 6.T7% Mke. The worth of individual rods varied between 2.8 and 2.9% Ake, while the reactivity worth of pairs of rods varied between 4.9 and 5.3% JAY: 9N The addition of rod thimbles (no control rods) lowered the peak-to-average power density ratio by about 8% of the homogeneous reactor value and moved the position of peak power density to a point sbout 6 in. away from the reactor center line, The peak gamma-ray heating rate in the core was calculated to be 2.5 w/ce with the reactor at 10 Mw; the maximum fast-neutron heating in the rod thimbles (rods out) was sbout 0.1 w/cc. Vi The critical fuel concentration with the use of 4LO-mil-thick INOR-8 fuel tubes in the MSRE was calculated to be nearly double that associated with no cladding present. An improved estimate of the activity acssociated with the Flg(n,a)N16 reaction was obtained. The N'© activity in the fuel salt ranged from 0.27 X 101° ais sec™t cc™l of fuel at the pressure-vessel inlet to 0.65 x 10'° at the exit from the core proper. The activity associated with the pump bowl was calculated on the basis that the daughter products of xenon and krypton plate out in the bowl. The associated dose rate 10 ft from the pump bowl was 10° r/hr after 10 days' cooling time following 1 year's operation at 10 Mw. For the same conditions, the resid- ual activity in the heat exchanger gave a dose rate of 2 x 10% r/hr, based on the assumption that all the isotopes which might plate out on INOR-8 do so in the heat exchanger. The gamma-ray dose rates above the top shield (3.5 ft of barytes concrete : plus 3.5 ft of ordinary concrete) during 10-Mw operation were calculated to be about 15 mr/hr for a solid shield and sbout 80 mr/hr if the ordinary concrete has 1/2-in.-thick slits; the neutron dose rates for these conditions were about . 2 mr/hr and 4.6 r/hr, respectively. Filler material and additional shielding will reduce the dose rates below the tolerance value. Estimates were made of the dose rates outside the side shield from individual sources within the reactor cell. The primary radiation source during 10-Mw power operation was the neutron-capture gammas from the iron in the thermal shield. With 7 £t of ordinary concrete the total dose rate was about 45 mr/hr; addition of 1 ft of barytes concrete block reduced the dose rate to about 1 mr/hr, PART II. MATERTALS STUDIES 4, Metallurgy Examination of the final eight INOR-8 forced-convection loops was completed, and summary information for the program is reported. In general, maximum corro- sion rates of INOR-8 by fused salts at 1300°F ranged from 1/2 to 1 mil in 20,000 hr for both LiF-BeFs; and NaF-BeFo systems. The attack was in the form of a pitted surface layer. The compatibility of molybdenum sheet with the materials in the MSRE system has been tested in thermal convection loops. No significant attack was observed on the metal parts of the system; however, a deterioration of the mechanical properties of the molybdenum was noted. Corrosion studies were started in order to test the effect of the oxidizing impurities in fused-salt mixtures. Tests designed to establish the effect of moisture were completed and are being examined. Work was continued in order to determine the solubility limits of chromium plus lron in nickel«base alloys contalning 18% molybdenum over the temperature range of 900 to 2000°F. A phase boundary for this metal system has been established. The 'temperature range of melting was investigated for various heats of INOR-3, and data are presented that show solidus temperatures to be higher than nil duc- tility temperatures reported by Rensselaer Polytechnic Institute, Specific heat of annealed INOR-8 was determined by direct calorimetric measurements and the data reported. An anomalous rise of about 20% was observed at approximately 600°C. The total hemispherical emittance (t.h.e.) was determined for INOR-8 in the bright-finished, matte, and oxidized conditions. INOR-8 in the oxidized condition was found to have a t.h.e. at 600°C of 0.76, compared with 0.24 for an unoxidized surface value. Studies were conducted to circumvent the problem of cracking and microfissur- ing observed in the welds of certain heats of INOR-8 material. The problem is associated with improper melting practices used by vendors in pouring the initial ingots. By using base metal and weld metal originally poured under staisfactory conditions, sound, crack-free welds possessing good mechanical properties can be made. An investigation is under way to develop brazing procedures suitable for remote fabrication operations. Brazed joints possessing good shear strengths at elevated temperatures were made, and a joint design suitable for remote operation was developed. A program has been started to determine the strain fatigue behavior of INOR-8. Data at 1300 and 1500°F are reported with a plot of Coffin's equation. Molten-salt permeation tests with different grades of graphite indicated that increasing the diameter of a graphite rod or fabricating it in the shape of a pipe can decrease its resistance to impregnation by molten salts. Tests showed that oxygen contamination can be removed from a moderately permeable grade of graphite by exposing it for 20 hr to the thermal decomposition products of NH4F:HF at temperatures as low as 930°F. The tensile specimens of 0.040~in.-thick INOR-8 exposed to this same oxygen-purging atmosphere developed a reaction layer <0.0005-in. thick. The reaction layer did not alter the prop- erties of the INOR-8. 5. In-Pile Tests A molten-salt-fuel capsule experiment, ORNL-MIR-4T-3, has been operating at the Materials Testing Reactor from May 5 to July 24 and is now at Battelle Memorial Institute for postirradiation examination., The four capsules contained fuel in AGOT, fuel impregnated, or R~-0025 unimpregnated graphite "boats". Samples of molybdenum, pyrolytic carbon, and INOR-8 were irradiated in contact with the fuel to temperatures to 900°C. 6. Chemistry The phase diagram for the ternary system NaF-ThF4-UF4 has been finished; this completes the systems limiting the quaternary Nal'-BeF,~-ThF4-UF4 and provides interesting comparisons with LiFF-BeF,-ThF4-UF4, in which solid solutions arising from the interchangeability of UF4 and ThF4 are much less cormon. Studies of the crystallization of the MSRE fuel show that with fast cooling a nonequilibrium path is followed along which the equilibrium primary phase, 6LiF«BeFo«ZrF4, fails to nucleate. Slow cooling leads to considerable segregation of the MSRE fuel (LiF-BeFp-ZrF4-ThF4-UFy, 70-23-5-1-1 mole %), with UF4 concen- trated in the last liquid to freeze. Since the first phases that freeze out are viii rich in ZrF4, the depleted residual liquid can, under some conditions, deposit UO0s. Small segregated regions are produced vwhen frozen plugs of fuel are used as freeze valves. Much difficulty has been encountered with fuel studies that involve the sampling and the measuring of oxide content in the range 100 to 1000 ppm. A suitable resolution of the problem of oxide analyses is being sought. An spparatus for measuring the surface tension of fluoride melts has been constructed for use in studying the wetting behavior of salts with respect to graphite. Graphite withstood at least a mild exposure to cesium vapor without noticeable alteration of its interfacial behavior toward fuel. A large-scale move from laboratories in the Y-12 Plant to new quarters at ORNI: interrupted much work on MSRE problems in both the Analytical Chemistry and Reactor Chemistry Division. Methods for analyses of the MSRE cover gas are under development, as are improved treatments for fuel and coolant purificationm. T. Engineering Research The enthalpy of the coolant mixture LiF-BeF, (68-32 mole %) was determined over the range 50 to 820°C. For the liquid, the heat capacity varied from 0.48 cal/ge°C at 500°C to 0.66 cal/g+°C at 800°C. The solid-liquid transition was not sha;ply defined; the heat of fusion, evaluated between 360 and 480°C, was 151.k cal/z. In order to further clarify the heat-balance discrepancy noted in the heat- transfer studies with the LiF-BeFo-ThF4-UF4 (67-18.5-14-0.5 mole %) mixture, the enthalpy of the liquid was redetermined, using a sample of circulated salt. Despite significant differences in composition, the enthalpies of circulated and uncirculated salt samples were equal within 3.5%. In contrast, the heat capaci- ties showed a deviation of as much as 10% at the extremes of the temperature range (550 to 800°C); mean values of the heat capacity were identical (Eb = 0.335 cal/g-°C). Interpretation of the data obtained in the study of heat transfer with the LiF-BeF o=UF4-ThFg (67-18.5-0.5-1L mole %) salt mixture was continued, with pri- mary emphasis on the evaluation of the abnormal heat balances observed. Exami- nation of the data and the analytical procedures suggests that additional measure- ments of the heat capacity and density of this salt mixture are needed to resolve possible errors in the convective heat gain. 8. Tuel Processing Compounds of SbFs with KF, AgF, and SrF, were prepared by reacting the com- ponents in anhydrous HF. Products were AgSbFg, KSbFg, and a compound that may not have been stoichiometric in the case of SrFo. The material prepared by reacting NaF with MoFg in HF had the approximate composition MoFg+5NaF plus some HF; attempts to remove the HF by evacuating the container appeared to remove some of the MoFg as well, Reactions of LiF, NaF, and KF with UFg, all in HF solutions, yielded yellow or orange solids on evaporation of the solvent, The solids contained HF and much less UFg than the anticipated complexes, probably because of evaporation of UFg during the HF evaporation. A single experiment to test the possibility of separating rare earths from ThF4 in MSBR blanket salt by dissolving the rare earths in HF contalning SbhFs was unsuccessful; the HF-5bFs solution dissolved neither rare earths nor thorium. CONTENTS SUMMARY vovuvveesonaseannsanee teseecans D 111 PART T. MSRE DESIGN, COMPONENT DEVELCOPMENT, AND ENGINEERING ANALYSIS 1. MORE DESIGN suveveceososesoaosasossstosonsosssossssssscssnsnsosesssansnsaas Introduction s.veevvriecicscrsanenssssrsnonsnns ceseane vesesasones 1l L Reactor Core and Vessel ...civveene cessans Cetsssreresrasrensans e 2 Primary Heat Exchanger ........ Gt eesresessessstesessaarseraorenas A A 4 PFHHRRRHREHERF O =3 W Fw o Radiator ...veevveenns e Fuel-52l1t Drain Tanks iteveeessvsaessvosestonssnssssssossaasssssas Equipment Layoul sueveeeesecssocsessavsosnessssssasssascososssssnses 8 Cover-Gas SYSTEIM .eesrsreevisosssstsosssesectosesnssnsscnsssssns s 11 System Heaters tveevertesecsnssscsossotascccsnnsss tesserevacsseanses 12 Design Status of Remote-Maintenance Systems ..cevecectirsecncreosss 13 1.9.1 Maintenance-Design Systems ....... tasecssserserrecsannans 13 1.9.2 Remote Maintenance by ManipulatorsS ..eeeeesecoessocsssces 13 1.9.3 Remote Maintenance by Manual Operations .eeeevseccecssos 14 1.9.4 Assembly Jigs and FixXBUTesS .eeeeverevosasocorasenossonss 15 1.9.5 Graphite-Sample Removal .uvveeeerornoscssrcsssoscsssnssonsse 15 1.10 Reactor-Control DesSigil teeeeressesccrescerescosscascsasossassocnsess 16 1.11 Decsign Status of Building and Site .vieeeceveessesocasecscnncsess 19 1.12 Reactor Procurement and Installation ..eeeeseseessevencveossveose 19 1.12,1 Demolition and Minor Alteration Work to Bullding 7503 teeeeesasessctssescsensetscasenancnes 19 Major Modifications to Building 7503 ..... cesecsacseaa .o 20 Procurement of Materials .......... tevecennasarscasrsacs 20 Procurement of Components ...... tessessescesseraneserens 20 Instrumentation and Controls Desigll veeseevrescsescocoone 21 Instrument Application Diagram ..eeesvecocsrseosceronanes 21 Electrical Control Circuitry s.veeserereecasosscaccnsras 2L LayOUL tivreeeneorsvsonnsseosessscsosssscosossssssnsssnsnes 21 Main Control Board ...eeeesesseossosassssossssssssssnasas 22 MSRE Data-Handling Study seesesesccceassssarsssnossscsss 24 Single-Point Temperature Alarm SYyStemM ceeeesesrssscacssse 25 s o FRRFRHERPRED P& PRRHRPRRFO o) wWwwwww et oD OV FWNRR WD 1.13 2. COMPONENT DEVELOPMENT aeeeevsscosocosessssssscessossassssnssasascssassca 28 2.1 Freeze-Flange DeveloPmMENt .veieeesrsesncncsasscsssesvococesasssnassne 28 2.1.1 MSRE 5-in. Flanges tueevsvescccscacsosssssevsscossnacanas 28 2.1.2 Freeze-Flange-Seal Test Facility ceeecececeencessosoasan 28 2.2 Freeze ValVeS sueeesserscscscsstosnsesasstsssancsssssstssccasavancs 32 2.3 Heater TeStS ctesreearterrascocssensesssesosssnsacsocsssssscsnssossssas 33 2.3.1 Pipe Heaters ..cveeeseresarsseossnens fetecesesereserennnase 33 2.3.2 Core Heaters teveesesessessssrssoscosssssssassoscsnsananas 33 2.4 Drain-Tank COOLET'S «eeteecrsessorsosesotsosossoecssososnssosassences 37 2.5 Sampler-Enricher Development ......... tereaeaene tesasensncaseases 37 2.5.1 Sampler-Enricher CONCEDPt vuveverrcrrsosessoscosccasasonas 37 2.5.2 So0lder-Freeze VAlVE .ceeeeetcssresrsosvsosoccscsacsesasacacacas 37 2.5.3 Sample CapSUle .uieesscescsosveraorrasssosnasonssnsans ces 40 2.5.4 Sample-Transport Container and Removal Seal ....eeeees.. 0 2.5.5 Sampler-Enricher DeSifll seeveetesrsrserecessasassoasssascs 40 x| 3. , 2.6 2,11 xit MSRE Core Development .c.ccceeveesecenrsacsanss cessesercsedsnsranns 2.6.1 2.6.2 2.6.3 2.6.4 2.6.5 Helium Purification ....... e eseas s s s es st trasat e st us et s ea Pump Development ....coceeeseevetssscsssssesrscsvnacsons ceseesnan 2 2 2 T 8. 8.2 8 1 .3 Heat Transfer Coefficients in the Lower Head of the Reactor Vessel ........ cecsccsresarenrsenne Fuel-Age Measurements in the Lower Head of the Reactor Vessel ...ceeecitervanscsssssscense cees Studies on the Disposition of Fine Particles in the Lower Head of the Reactor Vessel ........ PN Fluid-Dynamic-Induced Power Oscillations ...... cevesaane Full-Scale MSRE Core Model ...iciieevrsacassescosossancss MSRE F‘uelPLm]P.'l....l.l.l.l.l.l.l.'.t....l-..I....Clll Water Test of MSRE Coolant PUMD ceeeecvoressccscacecanes Advanced Molten-salt P‘-lmps e d & & 4 & 8 B B &0 & PSS AP R F R C SRS Graphite-Molybdenum Compatibility Test ....... e MSRE Engineering-Test LOOD ceeececceacerroressascscsssssaaceonsros .10.1 Oxide-Flush Run s.cvececcecresoncscesrscsnacens seseesnas .10.2 Freeze-Valve Operation ...ceceecececesscssccaas ceveesnae 2.11.4 .10.3 Level Indicator for Molten Salt (Test TOOD) eceveesseanns .10.4 Graphite-Handling Facility .cecevveceecennnronensocacns ‘e Maintenance Development ....... Cetetseserentasnasasneerens - 1.1 Maintenance Plan ...c.ccc... cessisessssrsessesessnrenans 1.2 Remote-Maintenance Mockup ...ccevevecscrornsnsrnsrcncsnns 2.11.3 MORE MOd€l ..vevevesncccssecscanasasesoscsrsssossnsscsssas . Portable Maintenance Shield ....ccceeoveenscososcssnersss Brazed-Joint Development ..eecvcecn. Cieeesectsasatsseseannans e Mechanical-Joint Development ...ceccceecacse et esesssenss s n s Steam Generator tececssrtsseseesssssscsssesossssssssassssenssssese MSRE Instrument Development ...eeeevsrececscccsascrscssssscseacasna .15.1 Pump Bowl Level Indicator ..ecceciesscasssnscsscrsnnsonns .15.2 Single-Point Level Indicaltor ..ccecececssscsransosocssss 5.3 Temperature SCANNEY ..c.ieveseeerrosrosssansvrsssscsssssacss 2 2 2 2 .1 .1 5.k Thermocouple Attachments and End Seals ...coveecencccsns REACTOR :ENG]:NEERING ANALYSIS ¢ * 8 80800 ® 2 0 " 8 B &k 2 PN O SN S PEE SRS EN S S e esbe ReaCtOr Physics 'E N EEREEEEE RN NN NI NI I B R R IR IR R I IR R IR I I I I A B 3.1.1 3.1 -F"L.AJI\) }EIIAIJH]RGY ® 8 ¢ % & 068 080 LB P T LR PSS PR S S b1 3.1.2 3.1.3 3.1.4 Nitrogen-16 Activity in Fuel Salt ...cecvvveneneen. eceiasasariae Residual Activity in Pump Bowl and Heat Exchanger ......... ceesas Reactor-Cell Shielding ..... Gt eceteesser et asus s et nses s s tane b 3.4.1 3.4.2 Analysis of MSRE Temperature Coefficient Of Reactivity ceeeesererrencsesscscssosncsrscansnansens Rod Worth, Flux, and Power Density Distributions in MOSRE s.tiviesesenseneccccnnssncesanans cesasaeranse Gamma -Ray and Fast-Neutron Heating in Thimbles ......... Reactors with INOR-8 Fuel TUDES sesscvaseosansscssoscnns TOP Shield N N NI N R A B R I R Y R I R I R Y R I T I I I I IR R I R B L Side Shield- * % & 8 & B OB F O B E S S AP A PRSP O NS SES eSO EPSE SR PART II. MATERTALS STUDIES mnamic-corrosion Stud.ies P I B B B B I I T T I R R BN Y R R Y I RN B N RN B BN BN N BN E S Examination of INOR-8 Forced-Convection LoOPS +vivsseess Molybdenum-Graphite Compatibility Tests .eeeecevsvansans Compatibility of INOR-8 — 2% Nb and Molten Fluorides ..cececeicescccscccseocorosnerennnssanns Fluoride-~Salt Contamination Studies ......iccevevenseess 77 95 5. 6. 7. 8. L.,2 =& = - —3 O\ IN-PILE TESTS 5.1 INOR-8 Development ......... ® s 8 9 v 0 9O r e e s 4 2 0 8 8 s 4 s 4 a8 h.2.1 Structural Stability of Nlckel Based 18% Molybdenum Alloys h.2.2 Temperature Range of Melting for INOR-8 ...ieeeveeersene 4,2.3 Specific Heat of TNOR-8 ..evereronnrorrrorosonosnrenenons Total Hemispherical Emittance of TNOR-8 ..veeerrveerrnrresonensns Welding and Brazing STudies ..ceevievrerenesescnsrennsssssserasnns Lh.li,l Welding of INOR-8 tuvuivienoroenconssssosensassasaacsnnes L.k b, L.k = .2 Metallographic Examination and Bend Tests ON WEldS suvevevesvesnencsasosscaconseascesosesssasnanns Transverse Tensile and Creep Tests .vieierersvcnsncncass Welding of INOR-8 to Stainless Steel and Inconel ...cieveineenanancsons tesesesarssasnaseana L,k,5 Remote BrazZiNg ceeeeeeeseeseoceroosoresstonssenesersaoess Mechanical Properties of TNOR=8 .iiveveereronrnsrersesaonsrsnonsas Impregnation of Graphite by Molten Salts ...iiivieoennnterennonees Ammonium Bifluoride as an Oxygen-Purging Agent for Graphite ....... h,7.1 Removal of Oxygen from Graphite ....eeeevvescsrcsoseanss h.7.2 Effects of the Thermal-Decomposition Products of Ammonium Bifluoride on INOR-8 ......cuve.s Graphite-Fuel Capsule Experiments ...vieeeesessscesssasossnsenene CHEMISTRY [ R N A A DR B I I B N I O A LT A I I T I I B I N Y I R Y BN B B R B I B I Y B B I R I R Y I I B B I I IR ] 6.1 6.2 6.3 6.k 6.5 6.6 ENGINEERING RESEARCH T.1 7.2 FUEL PROCESSING Phase-Equilibrium Studies .. 6.1.1 Systems Involving NaF with ZrF,, ThF4, and UFg and Other Constituents ..cvesevevrrsvrocsersrenens 6£.1.2 The System NaF-ThF4-UF4 . veeeeeeeeneresososnasensnnnosoes 6.1.3 Stannous Fluoride as a Component of Reactor Fuels ...... Oxide Behavior in Fuels .... ¢ # o o 0 s OO b e bbb SO E S PSS #8008 09 00e 6.2.1 Behavior of MSRE Fuel on Freezing and Effect of Segregation on Oxide Precipitation ........cecuenne 6.2.2 Oxide Content of Fluoride MeltsS ....eeevececreercnnncans Physical and Chemical PropertiesS .s.veecereereroscososessarsconsss 6.3.1 Surface Tension Apparatus for Molten Fluorides ......... 6.3.2 Volatilization of Iodine from the MSRE Fuel .....vevev0. Graphite Compatibility seeeserersvesrsereerseascesorssosrescacsanenss 6.4.1 Effect of Cesium Vapor on Graphite ...e..eeveeerensnnsse Fuel ProdUCtion seeveesreecstressasessscsosnsossossrsssssacssncsnssss 6.5.1 Purification Treatments ....eeeeeeeeeseecececaccaeacanas Analytical Chemistry ....... 6.6.1 Analyses of MSRE Fuel MixXture .eeeeeviecscasscorsnncnsss 6.6.2 Analyses of MSRE COVEr GB5 +cvvevevsorcsacsnncancananans LI B I B I Y I B I B I B I B R I R BN BN R R R BN AN BN I R N Y B DN A B L B B B RN B R B R AR B A Physical-Property Measurementis .ieeceiscerascsssocssorssssscssensns Heat—TranSfer Studies ® B & & 3 8 4 & 0 & 8 S0P P ORI R ERNSEEEE S ER SR 0 8 0 8 & 4 8 0 0 8 F PR PSSR PRSP P AN PSSR SRS S s Preparation and Analysis of Complex Fluorides of SbFs with KF, AgF, and SrFpo seeeecetcevsorescscosssnoascssssssesanss Preparation and Analysis of the NaF-MoFg CompleX ..cceveersceccecs Preparation and Analysis of Complexes of UFg with LiF, NaF, and KF seveveveecses . 0 & @ & 0 P 2 & & v é P B AR RSN Rr e Attempt to Separate Rare Earth Iluorides from MSBR Blanket Salt by SbFs in HF ® & 8 0 P8PSR T DN N LSRN e 136 136 PART I. MSRE DESIGN, COMPONENT DEVELOPMENT, AND ENGINEERING ANALYSIS 1. MSRE DESIGN 1.1 TINTRODUCTION Just as the semiannual report for the period ending February 28, 1961 was being issued, it was decided to review the fuel-circuit design with regard to two guestions: (1) should the reactor vessel have a flanged top head, and (2) should the pump be removed from the top of the reactor. The additional design studies indicated that it was not practical to flange the reactor head but that it was probably worth the additional complexity in pump mounting to remove the pump from the top of the reactor vessel. It was believed that the only advantage provided by flanging the reactor top was in the ability to remove the core independent of the reactor vessel. This was thought not to be of sufficient Importance to warrant the additional complexity of a large flanged seal. Moving the pump off the reactor, however, was thought to be quite desirable in order to facilitate graphite sampling and control-rod operation. The com- plexity of a more involved pump mount was thought to be justified by the accessibility to graphite and rods which it provided. Accordingly, the design was changed to include the new fuel-pump location. The liquid-poison tubes described in the last report were eliminated. With the pump moved off the reactor and with the top exposed, the more conventional and tested solid control rods were preferable. In all other basic concepts the MSRE system has remalned the same, and detailing of building alterations and component design has proceeded. The result is that the building-alterations designs and component designs, together with specifiications for these designs, have been issued to prospective bidders. The electrical heater design is not yet finished, but work is proceeding without difficulty. Instrumentation is also incomplete but is being carried forward as rapidly as necessary to meet construction and installation schedules. Design of both these phases can be accomplished during the building-alteration period and the component-fabrication period. There will be a decline in design manpower effort as of October 1, 1961, and for the remainder of the fiscal year the design manpower will be carried at a reduced rate. This effort will consist in effecting necessary minor changes and completing auxiliary-systems designs. Major items of design are discussed in more detail in the paragraphs which follow. 1.2 REACTOR CORE AND VESSEL The general configuration of the core and contalner vessel has not been changed, but some changes have resulted because of moving the pump off the top of the reactor. These changes involve: (1) the reactor discharge line, {2) the upper -head neutron shield, (3) the graphite-sampling access, and {4) the control rod penetrations. The reactor i1s shown in Fig. 1.1. The reactor discharge is through a 10-in. pipe, rising vertically from the center of the top head. This 10-in. pipe has a 5-1in. slde outlet leading off to the pump suction. The 10-in.-diam vertical section terminates in a flanged top with a metallic O-ring gasket. Into the 10-in. pipe, and extending down to a point above the 5-in. discharge tee, is a hollow plug which is welded to the mating flange. The purpose of this plug is to provide removable penetrations for control rod thimbles and a graphite-sample port. There is an annulus between the plug and the vertical pipe and fuel entering this annulus can be frozen to form a salt seal. In this manner the ring-seal flange 1s never required to hold molten salt but becomes a gas seal only. Air 1s circulated on both sides of the annular space. When the air is unheated it serves as a coolant to establish this freeze plug. The air passes through an electrical Turnace so that it can be heated, when desired, to thaw the plug after the reactor has been drained so that the plug can be removed. Removing the pump off the top of the reactor vessel made it unnecessary to put the massive INOR-8 shield plug in the top plenum of the reactor vessel. This plug has been eliminated completely, and the resulting space will be filled with fuel. Provision has now been made for control rods and graphite samples in the center of the reactor core. The four graphite stringer positions at the corners of the exact center of the core are omitted from the core matrix. INOR-8 thimbles are inserted in three of these positions. The upper ends of the thimbles are welded in the bottom of the plug which fills the 10-in.-diam vertical pipe on top of the reactor. These thimbles then provide penetrations into the core into which solid contrel rods can be inserted for control of the reactor. The fourth position in the core is arranged to accommodate four 7/8-in.-diam graphite rods. Any one of these rods (which extend only down to the midpoint of the reactor, the lower half containing a standard stringer of half length) can be withdrawn from the reactor for examination. In order to remove the sample graphite a 3-in.-diam access tube extends up- ward from the base of the flanged plug. A bolted flange seal on top of this tube can be removed for access to the core. This 3-in. riser tube has a plug, similar in principle to the one used on the 10-in.-diam pipe, with a similar frozen-salt annular seal. The plug in the graphite-sampling port provides a holddown of the four samples. When the flange is unbolted and the plug is removed, the samples are free and can be removed by the graphite-sampling device. In the event that the condition of the sample warrants more extensive examination, main graphite stringers can be remcved by opening the 10-in. flange and removing the entire control-rod-thimble assembly. 3 UNCLASSIFIED ORNL-LR-DWG 61097 AIR QUTLET AR INLET _ FLEXIBLE CONDUIT TO : CONTROL ROD DRIVE {3) SMALL GRAPHITE SAMPLE ACCESS PORT CONTROL ROD COGCLING AIR INLETS CONTROL ROD COOLING AIR OQUTLETS COOLING JAGKET AIR INLETS COOLING JACKET AIR QUTLETS AGCESS PORT COOLING JACKETS FUEL OUTLET REACTOR ACCESS PCRT SMALL GRAPHITE SAMPLES CONTROL ROD THIMBLES (3) HOLD DOWN ROD LARGE GRAPHITE SAMPLES (5) CORE CENTERING GRID FLOW DISTRIBUTOR STRINGER GRAPHITE-MODERATOR \ FUEL INLET / 5 REAGTOR GORE GAN — | REACTOR VESSEL — ] I “' T i,l # l (=) 111 | TE l ,fi’l L | . ?‘;I! / > | L I 4 6 _ = ' 20 ‘ ANTI-SWIRL VANES MODERATOR VESSEL DRAIN LINE SUPPORT GRiD Fig. 1.1. Cutaway Drawing of MSRE Core and Core Vessel, The tubes for liquid poison, previously considered for reactor control, were abandoned in favor of rods similar in design to those used successfully in the Aircraft Reactor Experiment. The control design is discussed in Sec. 1.10. 1.3 PRIMARY HEAT EXCHANGER Heat transfer and pressure-drop calculations for this component, along with preliminary stress analysis, have been reported in the component design report.l Drawings showing configuration and support and tiepoint locations have Leen prepared and were included with the component package released for bidding. 1.4 RADIATOR ] The design of the air-cooled radiator c0il? and enclosure was approved, and drawings and specifications were released for bidding. - The door drive mechanism and superstructure, which were not included in the coil-and-enclosure bid package, were redesigned. The single counterweight for the doors was eliminated. Each door was suspended from a drive shaft by means of a wire rope and sheave assembly. A Tflywheel was mounted on each drive shaft to prevent damage to the door by causing it to fall slowly when released. An over-running clutch was provided for each flywheel to allow the flywheel energy to be dissipated through friction. A magnetic clutch and a magnetic brake were placed on each drive shaft to permit individual raising, lowering, and positioning of the radiator doors. A chain drive system was retained in a modified form. The redesigned radiator assembly is shown in Fig. 1.2. 1.5 FUEL-SALT DRAIN TANKS Several changes were made in the design of the fuel-salt drain tanks. The most significant one was the method of connecting the steam and water lines from the bayonet heat-removal units to the steam dome and water supply. The number of bayonets was reduced from 40 to 32, and the steam dome diame- ter wag increased to 48 in. The steam lines enter the steam dome through the lower head, and the bayonet water supply tubes are concentric with the steam lines except at their inlets inside the steam dome. The water supply tube in- lets are through the wall of the steam outlet nozzles. The steam dome serves as an intermediate water reservoir during operation. The water inlets are staggered on two elevations to permit use of only half the bayonets, if desirable, by controlling the condensate return rate. The condensate returns to the steam dome from a reservoir located with the condenser outside the drain-tank cell. The bayonets rest on a support plate which is attached to the steam dome. This permits removal of the steam dome and bayonets without removal of the tank. The steam dome has a penetration through its center, through which passes a nozzle from the salt tank. This nozzle provides access for inspection and sampling and for inserting level probes in the salt tank. Figures 1.3 and 1.4 depict the new arrangement of the bayonet heat-exchanger units and the complete fuel-salt drain tank, respectively. This arrangement * DRIVE CHAIN [/ DOOR DRIVE MOTOR AND GEAR REDUCER —._ ', ‘ oo WIRE ROPE SHEAVEj—w\\ s T s i‘ . LY SUPPORTING STEEL 7 WIRE DOOR CAM GUIDE ——T BLDG. 7503, FIRST FLOOR {ELEV. 852 ft - Qin.) -7 | INLET DOOR | AIR BAFFLES / \ / \ AIR INLET DUCT | AN /AR DUCT FLANGE—//f & | - P MAIN AIR BYPASS DUCT / / £ [~ Fig. 1.2. MSRE Radiator Coil and Enclosure. UNCLASSIFIED ORNL-LR-DWG 55841R PENTHQUSE .« T~AIR OUTLET DUCT RADIATOR ENCLOSURE B RADIATOR TUBES UNCLASSIFIED ORNL-LR-DWG &0B3BA !' STEAM OUTLET h WATER INLET —— / N STEAM DOME FLEXIBLE HOSE LOWER HEAD BAYONET SUPPORT PLATE DRAIN TANK HEAD Fig. 1.3. Bayonet Cooling Thimble. UNCLASSIFIED ORNL-LR-DWG 61719 INSPECTION, SAMPLER, AND LEVEL PROBE ACCESS STEAM QUTLET STEAM DOME CONDENSATE RETURN WATER DOWNGOMER INLETS BAYONET SUPPORT PLATE CORRUGATED FLEXIBLE HOSE STEAM RISER 0 0 T STRIP WOUND FLEXIBLE BAYONET SUPPORT PLATE HOSE WATER DOWNCOMER HANGER CABLE GAS PRESSURIZATION gar AND VENT LINES INSTRUMENT THIMBLE FUEL SALT SYSTEM FILL AND DRAIN LiNE SUPPORT RING 1 | | FUEL SALT DRAIN TANK BAYONET HEAT EXCHANGER ‘ '| | THIMBLES (32) TANK FILL LINE 3 | | 2q | ' 8 ON Q \\\\r lN 8 L . CH ™y E i S g THIMBLE POSITIONING RINGS FUEL SALT SYSTEM FILL AND DRAIN LINE TANK FILL LINE Fig. 1.4. Primary Drain and Fill Tank for MSRE, simplifies the unit, lowers its fabrication cost, improves the water supply sys- tem, facilitates maintenance, and reduces maintenance cost. Other minor dimensional changes were made. 1.6 EQUIPMENT LAYOUT One maJjor change in layout design has been made. The fuel pump has been moved off the top of the reactor, and, therefore, the pump mount must provide Tlexibility in order to allow for movement imposed by thermal expansion of the lines. Figures 1.5 and 1.6 show the plan and elevation of the layout of the salt circuits. The fuel system now permits the removal of any component, although, to remove the reactor or pump bowl, shifting of some other components will be required. The piping system for the revised layout has been analyzed for stresses and has been found to be conservatively loaded by a good margin. The coolant piping system has also been analyzed by the same computer code and was found to te satisfactory. A satisfactory layout has been accomplished for the secondary piping (for 0il, gas, and cooling water) for the fuel pump. The final design makes use of a ring-joint flange for each line. These flanges are grouped around the pump in a circular pattern to facilitate remote make and break of these containers. A satisfactory layout of service disconnects plus spares has been finished for the electrical heaters and the thermocouples. Details of containment penetrations for all services (gas, electric power, lubrication, and instrumentation) have been designed. The reinforcement of the vessel wall for these penetrations was designed and analyzed for stresses. The nuclear instrumentation penetrations have been reduced to one large tube in which will be housed all the radiation counters to be used on the equip- ment . The cooling air which is used to establish and maintain the freeze valves and which alsc cools the control rods and top surface of the fuel pump is supplied from positive-displacement air pumps located in the special equipment cell. This air is recirculated, and the heat is removed by the space coolers within the reactor contalmment vessel. A1) structural members for component mounting within the containment vessel have been designed. This includes the reactor thermal shield, which is used as the structural support for the reactor. AUX. EQUIP ROOM FUEL PUMP LUBE OIL PK. NEUTRON INST. TUBES AIR DUCT SPECIAL EQUIP. ROOM FUEL PUMP HEAT EXCHANGER REACTOR COOLANT PUMP COOLANT DRAIN TANK RADIATOR hSRLE L -any. ) . X L STACK BLOWER HOUSE WASTE TREATMENT CELL . EQUIP, STORAGE CELL x DECON. CELL FUEL TRANS. CELL SPENT FUEL STORAGE TANK{;‘E DRAIN TANK NO. 1 . DRAIN TANK NO. 2 FLUSH TANK AUX. INST. ROOM | L A i P LT A | o UNCLASSIFIED ORNL-LR-DWG 60680 Fig. 1.5. MSRE Plant Layout, Plan, UNCLASSIFIED ORNL-LR-DWG 60681 1. 30 TON CRANE 8. WATER/SAND ANNULUS 15. HOT STORAGE 2. 7 & 3 TON CRANE 9. CONTAINMENT VESSEL 16. DECON. CELL 3. MAINTENANCE CONTROL ROOM 10, FUEL DRAIN LINE 17. SHIPPING CASK 4. COOLANT PUMP 11. RADIATOR 18. FUEL STORAGE TANK 5. FUEL PUMP 12, BY-PASS DUCT 19. DRAIN TANK NO. 1 6. HEAT EXCHANGER 13, COOLANT DRAIN TANK 20. DRAIN TANK NO, 2 D) 7. REACTOR VESSEL 14. STACK 21. FLUSH TANK N || a ® | I 1l L.J - o - "~ ' : : : :, ) .~ — =, In o Fig. 1.6. MSRE Plant Layout, Elevation. (I RN YIRS TR S P AR . s N et At o 7 EEa Y A TR AT T e S g ot 11 1.7 COVER-GAS SYSTEM The original design of the cover-gas system provided the alternatives of recycling or discarding contaminated helium at the option of the operator. Since the equipment for cleaning and recycling the gas requires some additional development as well as additional expense, installation of a recycle system has been deferred. The conceptual design of the components of the supply-and-discard system was completed, and the construction drawings were started for the charcoal beds, the cell blowers, and the helium dryers. A drawing showing the general location of components and the arrangement of interconnecting piping was issued for review and comment. Design memoranda on the charcoal beds and the leak detector system were issued. The only basic change made in the off-gas system was in the xenon holdup time provided by the charcoal beds. The design now calls for a xenon holdup time of 9C days at a sweep-gas flow rate of 4.2 liters/min (6000 liters/ day). The leak-detector system will be used to monitor flanged joints in the fuel- circulating system and in various auxiliary lines. Helium pressure, higher than system pressure, will be applied against joint-sealing surfaces. Leakage, if any, will be into the salt system, and leaks will be detected by loss of helium pressure. (See Figs. 1.7 and 1.8). UNCLASSIFIED ORNL-LR-DWG 63241 LEAK-DETECTOR LINE FROM LEAK-DETECTOR STATION O-RING GASKET w /% Z Z & ?_, SE’F‘{%*NG ‘\LEAK DETECTOR LINE SURFACE TO NEXT FLANGE, OR MAY BE CAPPED OFF. l TR . FLANGES ANY LINE REQUIRING LEAK-DETECTOR SERVICE REACTOR COMPONENT REMOTE-DISCONNECT FLANGE ] I - - i 3 FROM L ————w— | EAK-DETECTOR STATION Fig. 1.7. Leak-Detector Flange and Method for Using One Leak-Detector Line to Serve Two Flanges in Series. 12 UNCLASSIFIED ORNL-LR-DWG 63242 TO PRESSURE INDICATOR AND PRESSURE ALARM T IN CONTROL ROOM jw—— 250-psig HELIUM HEADER | PE -l O MINIATURE BELLOWS-SEALED VALVE, SIMILAR TO HOKE 480 SERIERS OR EQUAL TUBING, ¥g-in.0D x ’//’ooss—m.meL HIGH-SENSITIVITY {0.05 psi) D/p CELL (LOCAL) REFERENCE-PRESSURE LEG = 1/2 in. SCHED 40 AVERAGE OF 50 ft —{><}~ EQUALIZING VALVE RANGE ; 0-5C cm3/min MAX AP; Spsig —D<— {50 psi / P9 —D><~ TO LEAK-DETECTOR S~ - < JOINTS — 1/4 in. SCHED 40 | ] | VOLUME OF SYSTEM DOWNSTREAM — <~ OF THIS VALVE TO BE KEPT AS SMALL AS POSSIBLE TO INCREASE [ SENSITIVITY — TO OTHER LEAK-DETECTOR STATIONS L Fig. 1.8. Leak-Detector Station. 1.8 SYSTEM HEATERS The design of a typical line-heating unit has been changed from that previ- ously reported. The present design is shown in Fig. 1.9; remote-maintenance practices dictated this modification in the unit. The heater as shown can be removed from the pipe by breaking the electrical disconnect and engaging the lifting eye with a tool. A simple lifting motion will then remove the unit. The unit consists of heating elements of resistance coils (Nichrome V) embedded in fused alumina, thermal insulation, and a metal container or housing. The thermal insulation is a light-weight felt of alumina and silica. Reflective foils are placed at intervals through the insulation. The horizontal piece does not contain heating elements. It serves only as thermal insulation and as a support for the top unit. Process-piping supports penetrate it. Heating elements are provided in the radiator enclosure. Flat, ceramic- embedded heating elements are mounted where possible around the radiator coil, These elements have a combined capability of 30 kw and will be used for pre- heating and during periods of zero-power operation. At all other times they will be energized at a reduced voltage. Tubular heating units are placed parallel to the tubes and are located between the tubes and the radiator-enclosure doors. The tubular elements will have their full-rated voltage applied when the doors are closed or dropped. The surface temperature of these units will reach 100C°F in 1.5 min, and the heaters are installed to prevent freezing of the outer row of tubes in a lossg-of-flow incident. The tubular heaters can also be used in preheating and during barren-salt operation. UNCLASSIFIED ORNL-LR—-DWG 63243 LIFTING BAIL-——~__ Y HIGH-TEMPERATURE THERMAL INSULATION > HEATING ELEMENT —— \ SECTION A-A Fig. 1.9. Typical Pipe-Heater Unit. ELEVATION 1.9 DESIGN STATUS OF REMOTE-MATNTENANCE SYSTEMS 1.9.1 Maintenance-Design Systems Two distinct systems of maintenance are provided, one of which permits working directly through the shielding roof plugs for replacement of small com- ponents or the preparation for removal of large components, and the second requires operation of the crane and a manipulator from a remote control room. Although it is theoretically possible to do all maintenance tasks by the fully remote method, the use of semidirect methods will greatly reduce the time required to perform many small jobs. Furthermore, the capital cost of Building 503 shielding will be greatly reduced since it will not be necessary to provide protection to control room and office areas for long periocds of time while all roof plugs are removed from above the reactor cell for remote maintenance. 1.9.2 Remote Maintenance by Manipulators The completely remote maintenance system, to be used for the replacement of those major components which require removal of major porticns of the cell shielding, requires that personnel, operating in a shielded area, can see and manipulate the cell shielding, the component disconnects, and the components. The elements of this system have been designed or specified. The maintenance control room, above and to the right of the reactor cell (see Fig. 2.28), pro- vides a shielded work area with windows for direct vision of the reactor cell, drain cell, and crane bay. Located in it are controls for the 7.5-ton and 30- ton cranes, the manipulator, air-operated tools, and the television system, The 7.5- and 50-ton cranes will be modified to permit control of five speeds and three speeds, respectively, in all directions, from the maintenance control room. The 30-ton crane will be further modified by the addition of a motorized, 360°-rotating hook and a remote-reading load cell. The manipulator and bridge have been specif'ied and the rails designed to permit the manipulator to be used over the reactor cell and the maintenance- practice cell. 14 The television viewing system provides two pairs of stereo cameras, each pair with its own pan-and-tilt mechanism and vehicle to carry it around the track. Specifications have been written for automatic 1ift tongs (Fig. 1.10) that will engage, l1ift, lower, and disengage the lower shield beams and shield support beams, using only the remotely operated crane. UNCLASSIFIED ORNL-LR-DWG 61193 A\ / MOTORIZED CRANE HOOK SHIELDING PLUG—" Fig. 1.10. Remotely Operated Lifting Tongs for Shielding Plugs. 1.9.3% Remote Maintenance by Manual Operations This semidirect maintenance system permits personnel to use manually operated tools through the cell shielding and to view directly through small windows or with optical aids. Conceptual design of and procurement for the portable mainte- nance shield has been completed. The shield covers the opening left by the removal of two lower shield beams from the reactor or drain cell and provides a movable opening through which tools, windows, and lights can be inserted. Several hand tools have been designed for use with the shield. - 15 Several special problem areas have had to be considered separately. A fixture has been designed for the removal and precise replacement of the rela- tively heavy sampler-enricher spoocl piece with expansion bellows. The mechanical tools for making a remotely brazed joint have been designed and detailed for construction. These include machines for holding the pipe, cutting it, taper- machining the male stub, assembling the Joint, and holding it during the brazing cycle, 1.9.4 Assembly Jigs and Fixtures The design of the jigs and fixtures for the assembly of the reactor, fuel circulating pump, heat exchanger, and associated piping is progressing. The conceptual work is almost complete, and detail design has been started. A small model has been built. The jig (Fig. 1.11) will consist of four basic parts plus additional supports and bracing. After the completion of the fabrication of the initial system, the jig can be broken down into four parts for storage. If it becomes necessary to fabricate replacement components, only two parts of the jig will be required for the duplication of any given component . UNCLASSIFIED ORNL-LR-DWG 611524 FREEZE FLANGE (TYPICAL) MOTOR ;! COOLANT PIPE = \\\<\ PUMP BOWL SECONDARY FLANGE aS AT CLAMPING SECTION YA ) { DETACHABLE) \ = L \ Ne COOLANT PIPE L &— PUMP ASSEMBLY SECTION (DETACHABLE ) +_—— THERMAL SHIELD HEAT EXCHANGER ASSEMBLY SECTION { DETACHABLE) % REACTOR VESSEL BASIC FRAME REACTOR SUPPORT {FOR FINAL JIGGING) NOTES: DRAIN LINE 1. JIGGING FOR AUXILIARY LINES, CLAMPS FOR PIPE AND PIPE FLANGES, LEVELING DEVICES AND CERTAIN FRAME BRACES ARE NOT SHOWN. 2. STRESS RELIEF AND THERMAL CYCLING ARE DONE BEFORE REMOVING FINISHED COMPONENTS. Fig. 1.11. Concept of Fabrication Jig for Primary System, 1.9.5 Graphite-Sample Removal The preliminary design of a proposed method for removing the small (7/8-in.- diam) graphite core samples has been completed. The graphite sampler consists of': 16 1. An eccentric steel plug 32 in. in diameter and 18 in. thick [Fig. 1.12 (3)] fitted with seals and buffer gas to prevent the escape of activity or the introduction of air to the reactor system while samples are being removed and replaced. 2. A standpipe [Fig. 1.12 (0)] bolted to the sample access flange and to the lower shield plug for containment purposes. An exhaust line will be attached to the lower end of the standpipe, and nitrogen will be introduced at the upper end [Fig. 1.12 (5)]. Tnside the standpipe will be located two containers for the graphite samples, one containing the new graphite sample to be installed and one for the graphite sample that is to be removed. In addition, a bracket [Fig. 1.12 (8)] for holding the sample-access flange and holddown assembly [Fig. 1.12 (7}] while samples are being changed will be mounted on the side of the standpipe. Viewing will be through a lead glass plug, and general lighting will be pro- vided by a floodlight [Fig. 1.12 (17)] inserted in the eccentric steel plug. A zirconia light will be used to project a light beam inside the sample-access pipe for handling the graphite samples. 1.10 REACTOR-CONTROL DESIGN After a preliminary assessment of MGRE hazardsbr it was concluded that neither the amount nor the rate of addition of excess reactivity from any source would create a hazard. Therefore, the need for a fast-acting, multichannel safety system is eliminated. Control rods have been designed to shim for reactivity changes as follows: 1. ZXenon 0.013% 2. Fuel-pump speed 0.002 3., Power coefficient 0.002 4. Burnup between fuel additions 0.002 5. Temperature control (~300°F) and fuel penetration 0.027 Total reactivity C.0LO Three control rods are provided. Each contrecl rod has a maximum worth, when inserted with all other rods withdrawn, of 0.025 8k/k. Their combined worth is 0.0LO 8k/k. The maximum rate of withdrawal for a single rod is 0.0002 8k/(k-sec). With three rods inserting as a group, the maximum rate of poisoning is 0.0005 &k/(k-sec) . All rods are identical, and any one, but only one at a time, may be used as a servo-operated shim for automatic control purposes. Control rod design has been altered by substituting solid, mechanically driven rods for the liquid poison tubes originally considered. Fach poison rod consists of a series of short tubes of B,C which are sheathed with INOR-C and mounted concentrically on a flexible drive cable (Fig. 1.13). The resulting configuration is the same as a so0lid rod 1 in. in diameter. These rods move inside of vertical thimbles located centrally in the core (Fig. 1.14). The assembly of three rods is part of a larger flange- mounted assembly mounted on top of the reactor vessel. The rods are air cooled, but loss of air cooling will not render the rods inoperative. UNCLASSIFIED ORNL-LR-DWG 61027A FLOOD LIGHT INSERT Fig. 1.12. Schematic of Graphite-Sample Removal System. ] 18 fmmt—— 4,360 in. DIA (REF) ———= 5ft 5in. | 1% in. ——————————=] L i e ] Wi, Y N Y v N \ W NZZ N 7777777 77 77 777777 7 77 7 7 7 A N 7 7 AN A PC 7 NUT Y Fig. 1.13. Vertical Section of Poison Control Rod. UNCLASSIFIED ORNL-LR-DWG 6i152A | —— PC 8 SUPPORT SPIDER GUIDE TUBE 1% -in. OD x 0.065-in, WALL ——— PC 10 SEAL RING | —s— THIMBLE 2-in. OD x 0.065-in. WALL INOR-8 /PC 9 FLEXIBLE INCONEL HOSE L —— AIR FLOW: 3.3 cfm | __—— PC { B4C CONTAINED IN INOR-8 CONTAINER—1.040-in. OD —— PC 8 SUPPORT SPIDER |_—GUIDE BARS, 4 AT 90° — PC 6 FLEXIBLE DRIVE CABLE 19 UNCLASSIFIED RNL-LR-DWG A F\’EACTOF\’ q:_ ORNL-LR-D 60845 f .\ ‘Q§§\ GUIDE TUBE N \ - GUIE BAR FuETLYEEA“NLNEU /'\\ \i/// . \‘ \\\\\ el 7t AN 4 GRAPHITE lRRADlATION V/\\\“’///// SAMPLES ( 7/g-in.DIA) > J g in. TYPICAL // -t 2.00in —»= Fig. 1.14. MSRE Contro!l Rod Arrangement. Since ultimate reactor safety will not be vested in the control rods, the number of nuclear penetrations has been reduced from three to one. This re- maining penetration, a large water-filled tube, will house two fission chambers and two neutron-sensitive ion chambers. These are sufficient to cover the full range of reactor operation and will serve as control system input sensors. A preliminary list of those situations calling for either automatic shut- down by draining or for contreol rod insertion has been prepared. 1.11 DESIGN STATUS OF BUILDING AND SITE All major design on the building has been completed, and drawings and specifications were sent out to prospective bidders in July. The packages designated A and B (construction in reactor, radiator, and containment areas, drain-tank cells and maintenance control room) will be bid on September 6, 1961. These packages, now combined and currently referred to as Major Building Modi- fication, have had three addenda issued. The addenda were made in time to be included in the September & bid date. The remaining site modification design, consisting of Minor Building Modi- fication (office ventilation, service-air compressors, etc.) and Exterior Site (cooling tower, filter house, charcoal filter columns, etc.) will be completed in October. 1.12 REACTOR PROCUREMENT AND INSTALLATTION 1.12.,1 Demclition and Minor Alteration Work to Building 7503 The demclition and minor alteration work performed on a cost-plus-fixed- fee contract by the H. K. Ferguson Company was completed on schedule in June. The work consisted of stripping structural steel from the 24-ft-diam containment 20 ) vessel, partial excavation of the drain-tank cell, rehabilitation of the storage cells, stripping of obsolete conduit and piping which interfered with MSRE installation, excavation for the nuclear instrumentation tube, removal of a section of the penthouse wall, construction of an outside stairway for an emergency exit, painting of offices, control room and hallways, and general cleanup. 1.12.2 Major Modifications to Building 7503 Drawings and specifications for the first bid package of work, Modifiication of Building 7503, have been sent to prospective bidders, and bids for performing this work are to be opened September 6. This package includes all work associated with the modification of the 2L4-ft-diam reactor containment vessel, the radiator cell, drain-tank cells, the secondary containment walls, and the remote mainte- nance control room. Involved are structural, piping, and electrical work, &all of which will be accomplished by lump-sum contract. Other site-preparation work will be advertised for lump-sum contracts in October, when design work is scheduled to be completed. All site-modification work is scheduled for completion prior to June 30, 1962. 1.12.% Procurement of Materials Contracts were awarded on a formal bid basis for the fabrication of INOR-8 plate, sheet, rod, and weld wire for the MSRE. Promised delivery date for this material is September 15. Bids received for fabricating pipe, tubing, and fittings of INOR-8 are being reviewed and evaluated. It is expected that contracts will be awarded to successful bidders in August. Graphite manufacturers have been invited toc submit bids for furnishing the - MSRE moderator graphite, completely machined to specifications and tolerances. Bid closing date is August 14, and the purchase contract is expected to be awarded tc the successful bidder by September. 1.12.4 Procurement of Components A request for bids on a package of MSRE major components has been sent to prospective fabricators. This package consists of the reactor vessel, including the internal support structure, the radiator, the primary heat exchanger, and the fill-and-drain tanks for both the fuel and coolant systems. All INOR-8 mgterial will be furnished by the Project. A pre-bid conference with fabricators interested in bidding for this work is scheduled for August 29. Because INOR-8 is a new alloy it is expected that scme fabricators may not be familiar with its working properties, therefore sample material produced to MSRE specifications will be furnished to each qualified prospective fabricator requesting it in order that he may become familiar with the materigl before bidding. Closing date for bids to fabricate these components is October 6. . 21 1.15 RBEACTOR INSTRUMENTATION AND CONTROLS DESIGN 1.15.1 TInstrument Application Diagrams During the past report period the major effort was directed toward the completion of an acceptable set of instrument application flow diagrams. A total of 15 drawings is scheduled. Of this number, the 13 listed below were issued for comment at least once. D-AA-B-40500 Fuel-Salt Circuit D-AA-B-L0501 Coolant-Salt System D-AA-B-L0502 Fuel, Flush, and Drain-Tank System D-AA-B-40503 Cover-Gas System D-AA-B-4050L Fuel-Salt-Pump Lubricating 0il System D-AA-B-L0505 Fuel-Sampler-Enricher System D-AA-B-40506 Liquid-Waste System D-AA-B-L0508 Coolant-Salt-Pump Lubricating 0il System D-AA-B-L0509 Water System D-AA-B-L40510 Off-Gas System D-AA-B-40513% Fuel Fill and Transfer System D-AA-B-40514 Instrument-Air and Service-Air Systems D-AA-B-L0515 Containment Air Revisions are now being made to all these drawings, bringing them up to date with current thought. It is estimated that this basic part of the instrumentation ef'fort is 85% ccmplete. A tabulation of all instruments shown on the flow dia- grams, giving identifying numbers, location, function in the process, and a brief description is also being developed along with the application drawings. This tabulation is approximately 60% complete. Twelve thermocouple-location drawings were issued "For Approval," but some additional revisions will be required before final approval is obtained. A tabulation suggesting the method of readout for all thermocouples was also completed. 1.1%.2 ZElectrical Control Circuitry Control-circuit design is in a preliminary stage. The first series of conferenceg to establish control-circuit requirements was completed. A function- al listing” of the most important interlocks was prepared. This listing is preliminary and many control 1limits have not yet been established. 1.15.3 Layout Layout of the Instrumentation and Controls (I and C) system is proceeding as layout of the building and equipment become firm and as instrumentation requirements become known. Two major instrument areas have been designated. A main control area will be located at elevation £52 in the northeast corner of the building, and a transmitter room will be located on the 840 level, adjacent to the reactor. Some instrumentation will be located on auxiliary panels outside these areas. However, an effort i1s being made to centralize instrumentation, and field panels will be used only where necessary or where the nature of the operation dictates that controls and instruments be located in the field. The T and C system layout is being designed to permit all routine operations to be performed in the main control-room area. A proposed arrangement of the main control area is shown in Fig. 1.15. All thermocouple leads will be brought through a patch panel located in the auxiliary area adjacent to the main control rocm. Safety-control-circuit relays will be mounted in a special cabinet also 22 located in the auxiliary control area. Data-logging equipment will be located in a room adjacent to the main control area. UNCLASSIFIED ORNL-LR-DWG 63244 1 OBSERVATION DECK - A f'/\ S NEW GLASS WA MAIN CONTROL THERMOCOQUPLE ‘ @ LL BOARD . PATCH PANEL —- . o ~ i CONSOLE@ t[2]3]a] | i FRONT T AUXILIARY CONTROL ROOM INSTRUMENT SHOP AREA 18 x17 ft (306 112) DATA AND SUPERVISORY ROCOM AREA 18 x14 ft (252 f12) REL AY apEsobnaen NEW GLASS WALL = -- 18 % 49 ft (BA2 £12) AUXILIARY CONTROL BOARDS I e ELEVATION 852 ft = = ] C 2 4 6 8 10 SCALE IN FEET Fig. 1.15. Proposed Operational Area Layout. A1l electrical and pneumatic signal lines originating within the contained areas will be brought out through specially designed lesktight penetrations in the wall of the containment vessel. .Thermocouple and electrical signal leads will be terminated in pressurized junction boxes located in a tunnel adjacent to the reactor. Standard control cable and thermocouple lead wire will be run in open trays from the tunnel to the main and auxiliary control areas. Pneu- matic signal lines will be brought through a tunnel to the transmitter room adjacent to the reactor. Some of these lines will terminate at equipment such as weigh-system control panels, solencid valves, and pneumatic receivers mounted on auxiliary panels in this area. Other lines will continue through the trans- mitter room to the main control area. Detail layout of equipment in these areas will begin soon. A survey of requirements for pneumatic tubing, thermocouple cable, and electrical signal cable for the contained areas was made in order to establish the number and size of penetrations needed. A preliminary drawing showing proposed routing of cable, trays, conduits, and tubing throughout the 7503 area was prepared and issued "For Comment." Existing facilities will be utilized wherever possible. 1.13.4 Main Contrcl Board Figure 1.16 shows a front elevation layout of the proposed MSRE control panel. This design is based on the principle that only that instrumentation necessary for operation of the reactor system will be located on the main con- trol board (MCB). Other information will be read out either on a data logger or on instruments located in an auxiliary area adjacent to the main control room. Those controls and instruments associated with routine adjustment of reactor power will be located on a console. Although the layout shown is incomplete with respect to details and the number of instruments, comments indicate that it is acceptable from the standpoint of general arrangement, method of presentation, and systems shown. a4 — s g— 1 T4 - [ HHEEEE ] IEEGGEEEN HELEEE ||| M 23 UNCLASSIFIED ORNL-LR-DWG 63245 [ BEHETE | EEEEE [ REFEDE [ EREENE | EEEER | BEROE | JUMPER BHOARD JUMPELR BOARD JUMPER BOARD AUKILIARY PUNMP & COMPREDSOR P UH-ESUTTOMNS RAULTIPONIT TEMPERATURE (NDICATOR ADLILIARY PUMP & COMPRESS,OR PUSHBUTTOM®S PAMNEL MYM.CB. PANEL N% M.CB.2 ; i e bi - L 3N - 3 i LR FP-A, Dt (D IprC 3224 & 4 — YA AT ] PANEL MN*MC B3I PANEL NOM.C.B 4 i KU I POIRIT TE W APERATURE CECORDER | PAEL NTMCBS DANEL N'MCH & PANLL N*IC BT PAMEL NEMCEB S PAMEL NOMCBI PAMNEL. N MEBG PAMEL 2 MC By | | 2eto’ Fig. 1.16. Composite Control Board Layout. 24 1.15.5 MSRE Data-Handling Study A study of the MSRE data-handling requirements is nearing completion. The objective 1s to determine those requirements and to propose one or more data systems to meet them. Studies have indicated that approximately 612 input signals, excluding those transmitted to field-mounted instruments and indicator lamps, are to be collected by a data display system. A summary of these inputs is shown below. Radiation 37 Pressure ) 2l Flow 17 Level L Speed 1 Weight 5 Tlectrical (voltage, current, frequency) 7 Temperature 517 Of the 517 thermocouple inputs, 245 are used primarily during the reactor heating and coocling operation and will be handled by a separate temperature- scanning system. All information necessary for reactor operation will be displayed on the reactor conscle and on main and auxiliary panel boards by conventional recorders and indicators. An automatic data-handling system would implement this system by recording data in a convenient form. Using this approach to information dis- play, a data system should have a capacity to handle 250 to 500 inputs. The major problem associated with the conventicnal display system is the long-term storage and retrieval of information. This problem results from the storage of information on numerous single-point recorder charts and the subsequent job of retrieving and correlating information from several charts at a common time base. Another major disadvantage of the conventional data-display system is the production on charts of large guantities of static data. This static data recording is necessitated by the requirement to record transients should they occur. Using strip charts, the data are continually recorded, even during static system operation. A data-handling system would alleviate these problems and provide further improvement in data handling by: 1. providing high- and/or low-1limit alarm detection for each input and providing printout of the signal value when an alarm occurs; 2. logging data on a typewriter at periodic intervals or by ocperator demand, in a format such that interaction between variables can be seen; %. 7providing for long-term data storage, by recording on magnetic tape in a format for entry into an external computer; 4., displaying data directly in engineering units; 5. providing calculations necessary to perform some on-line date analysis; and 25 6. providing flexibility so that the input-scanning sequence and output logging can be altered merely by a program change. A1l these features will aid in the collection and handling of data. The data will be more accurate, and analytical results will be available much guicker. This will result in reducing manpower requirements for data processing and in reducing over-all experiment costs by providing information in time to influence the planning of subsequent experiments. The complete results of the data study will be contained in the report to be issued to the MSRE proJect in the latter part of August cr early September. 1.1%3.5 Single-Point Temperature Alarm System A study is under way to determine the most reliable and least expensive method of monitoring the operating status of freeze flanges and freeze valves. Both the flanges and valves use temperature measurements to indicate their operating status. The flanges are equipped with four thermocouples and four spares and the valves with three thermocouples and three spares. Low tempera- ture of the flanges indicates normal operation, and low temperature of the valves indicates closed position. High temperature of the freeze flanges indicates possible seal leaskage, and high temperature of the valves indicates valve-open position. Definite operating temperature ranges are designated for both the flanges and valves. The operating status of both these components must be known during the fill and drain and normal reactor operation. The freeze-flange status will be determined by monitoring two thermocouples, by single-channel devices. An alarm will be produced when either thermocouple exceeds a preset high limit. Single-channel monitors are required to produce the required reliability. In addition to these monitors, one thermocouple will be logged to provide an operating history. The freeze-valve operation is determined by monitoring three thermocouples by single-channel devices and logging one spare thermocouple for operating history. Two of the single-channel monitors will produce a visual indication when the temperature reaches either high or low setpoints. The visual indication will indicate valve position. A visual signal on low temperature indicates valve closed, and on high temperature, valve open. Control circuits will be utilized with these monitors to produce audible and visual alarm signals should operating malfunctions occur. The third single-channel monitor will be used to provide control interlocks or permissives during certailn phases of reactor operation. This unit produces high- or low-limit signals which are used for this purpose. One of the systems being considered for this application utilizes bistable magnetic amplifiers equipped with high/low alarm detectors. The units are com- pletely solid-state devices. A block diagram of the system is shown in Figs. 1.17 and 1.18. The control module furnishes power to operate up to 22 slarm modules. It contains a circuit common to all alarm modules to provide central indication of high- or low~alarm condition when any module in the system is triggered. The input signal from a thermocouple is fed to an isolated winding on the alarm-module magnetic amplifier. The magnetic amplifier is adjusted so that an 26 UNCLASSIFIED ORNL-LR-DWG 63246 LIMIT REMOTE QUTPUT SET MAGNETIC AMPLIFIER HIGH-LIMIT REDOH[GH-LIMiT THERMOCOUPLE DETECTOR INDICATOR LAMP < 113 _j SET CALIBRATE SIGNAL LOW-LIMIT AMBERO LOW-LIMIT DETECTOR INDICATOR LAMP od 1] TEST METER REMOTE QUTPUT Fig. 1.17. Block Diagram of Alarm Module ond Single-Channel Temperature Alarm System. UNCLASSIFIED ORNL-LR-DWG 63247 [ METER METER LINES FROM D MASTER ALARM MODULES RED LIGHTS o L RED - OR AMBER LIGHTS »| CATES MASTER AMBER CARRIER POWER CARRIER - TO MODULES OSCILLATOR AND POWER AMPLIFIER DC ' ‘ POWER SUPPLY POWER SPEAKER FAILURE DETECTOR w1 LIGHT ] \ CALIBRATOR i l BATTERY Fig. 1.18. Block Diagram of Control Module and Single-Channel Temperature Alarm System. input in its operating range produces a differential output into the limit- detecting circuits. If this cutput exceeds the levels set by the high- or low- limit set controls, the magnetic amplifier is triggered into a "high" or "low" operating condition and drives the proper alarm light or relay. The magnetic amplifier is reset by the "reset" button, but if the off-normal condition still exists the magnetic amplifier will revert to its alarm condition after reset. 27 The system is designed for fail-safe operation and is equipped with power- failure-detection circuits. The design of the alarm modules is such that mal- functions can be detected by the alarm circuits. To provide the best reliability, the system would be operated with two control modules, each module supplying one alarm module at each freeze flange and valve. In this manner the failure of one power supply would not cause the logs of the complete information or control action. REFERENCES 1. E. S. Bettis, MSRE Component Design Report, MSR-01-0( (June 20, 1961). 2. MSRP Prog. Rep. Feb. 28, 1961, ORNL-3122, Table 1.%, p O. %, Ibid., p 16. 4L, S. E. Beall, W. L. Breazeale, and B. W. Kinyon, MSRE Preliminary Hazards Report, ORNL CF-61-2-46 (Feb. 28, 1901). 5. Addendum to MSRE Preliminary Hazards Report, ORNL CP-01-2-46, (Aug. 14, 1961), p 29. 2. COMPONENT DEVELOPMENT 2.1 FREEZE-FLANGE DEVELOPMENT 2.1.1 MSRE 5-in. Ilanges The freeze flanges for use in the primary and secondary salt systems are now sized for 5-in. sched-40 pipe connections. The over-all dimensions, with the ex- ception of the pipe connection size, are the same as those previously reported.l The gas seal was changed from a buffered double-conical-gasket type to a buffered ring-joint type because of the following considerations: 1. The conical-gasket type of seal is believed to be susceptible to gasket damage if excessive angular misalignment exists during assembly. It is believed that the ring-joint type of seal is less sensitive to misalignment as the mating flanges are belng brought together. 2. Additional confidence in the ability of ring-joint seals to meet the leaktightness requirements has been gained through the freeze-flange thermal cycling testing described in Sec. Z2.1.2 and the remote-disconnect testing described in another I'eport.2 The revised flange design is shown in Fig. Z2.1. Two sets of clamps were purchased for developmental use with the 5-in. flanges. However, they were returned to the vendor for heat treatment, which nhad been overlooked during fabrication. Two sets of rough-machined INOR-8 flange forgings were ordered for development use. Four nickel ring-joint gaskets are also on order. These flanges and gaskets will be used for testing the full- size gas seals and for thermal cycling to check the dimensional stability and thermal-fatigue life of the flanges. A system for installing and removing the clamps was designed and is described in Sec. Z.11. A study is under way to determine means of obtaining consistent over-all thicknesses of the flange-gasket assemblies. Methods of closely controlling the effective height of the ring gasket and the effective depth of the flange grooves will be reguired. This requirement is imposed by the steep load-deflection characteristic of the spring clamp. 2.1.2 Freeze-Flange-Seal Test Facility The equipment for testing freeze-flange seals was completed and placed in operation. The facility, with a 3-1/2-in. and a 6-in. freeze flange installed, is shown in Fig. 2.2. In heating the 3—1/2-in. flange, it was necessary to install a hollow copper heat collector that eacily clears the inside of the pipe but closely fits the bore of the flange. This collector receives heat from the entire length of the 28 29 UNCLASSIFIED ORNL-LR-DWG 63248 ¢ ~ \ FLANGE / fin. —~=0,030-in. \ CLAMP ———» * GAP WIDTH \ l {in. (TYP) | 2-in. TAPER / 0400in./in. . 5 ey, (TYP) ks BUFFER / \ % CONNECTION {a=in. TAPER 0.050 in./in. MODIFIED R-68 (TYP) RING GASKET —] . _ Y g L ! t%-in-R (TYP) 9-in. SCHED-40C PIPE SIZE Fig. 2.1. Five-inch Molten-Salt Freeze Flange. silicon carbide heater and conducts a large part of the heat to the flange. By appropriate adjustment of the heater power and suitable insulation thickness on the pipe stubs, the desired temperature distribution was obtained. The 3-1/2-in. flange was assembled, using a gold~plated Inconel gasket and a resilient clamp, and each of the two bolts was torqued to 300 ft-1b. The gasket was lubricated with a coating of graphite-alcohol suspension before in- stallation. During and after torqueing, each clamp block was rapped sharply on each side several times with a small hammer, with noticeable immediate reduction in leak rate. The flange was then cycled 21 times between 200 and 1300°F. Results of leak-rate readings made during the cycling are listed in Table 2.1. No measurable leakage occurred after the fifth cycle. Fig. 2.2. Freeze Flange Seal Test Facility Showing 6-in. Flange (left) and 3‘/2-in. Flange. SRR UnCLASSIFIED . PHOTO 36959 - L 9939 5 0€ 31 Table 2.1 Helium Leak Rates (std cc/sec) of 3—1/2—in. I'reeze Flange with Gold-Plated Inconel Ring-Gasket During Cycling Between 200 and 1300°F Cycle No. Male Qutside Male Inside Female Outside Female Inside Initial Makeup NDL* 3.5 x 10'8 NDL 5.6 x 107 B-1 Hot 1 x 10'8 1.5 x 1078 NDL NDL B-1 Cold NDL NDL NDL NDL B-2 Hot 7 X 1077 6 x 1077 NDL NDL B-2 Cold NDL 5 x 1077 NDL NDL B-3 Hot NDL 2 x 1077 NDL NDL B-3 Cold 2 x 1077 1.3 x 1070 NDL NDL B-4 Hot NDL b x 1077 NDL NDL B-4 Cold NDL 6 x 1077 NDL NDL B-5 Hot NDL b x 1077 NDL NDL B-5 Cold NDL NDL DL NDL B-6 Hot NDL NDL DL NDL B-11 Hot NDL NDL NDL NDL B-15 Hot NDL NDL NDL NDL B-19 Cold NDL NDL NDL NDL B-21 Hot NDL NDL NDL NDL B-21 Cold NDL NDL NDL NDL Disassembled, then Reassembled with Same Gasket Initial Makeup 1.4 x 1077 NDL NDL NDL B-22 Hot 5 x 107 NDL NDL NDL B-22 Cold 1.5 x 1077 NDL DL NDL B-26 Hot NDL NDL NDL NDL B-26 Cold 1.8 x 1077 NDL NDL NDL *No detectable leak. The 3—1/2-in. flange was then disassembled, the gasket was again coated with the graphite-alcohol suspension, and the assembly was remade with hammering. As shown in Table Z.1, all seals were nearly perfect during the hot part of thermal cycling, but one of the seals persisted in leaking when cold and showed no im- provement after five thermal cycles. A solid-nickel ring gasket has been installed and will undergo similar test- ing. Additional nickel rings, two of which are grooved on each "nose" to in- crease resiliency, are on hand for testing. It is expected that the nickel rings will form tight seals without requiring the use of gold plate. 32 The 6-in. flange-pair in the seal-test facility is essentially the flange design except that the bore is slightly greater and Inconel is used as the flange material stead of INOR- Efforts to heat the bore of this flange to 1400°F with the silicon carbide bar heater at the center line and with an additional Nichrome wire heater at the bore were uccessful. Resistance heaters were installed on the outside of the pipe stubs adjacent to the flange, and an induction coil was placed in the bore of the flange to supply the required heat 1 distribution. Temperature distributions along the gap between the flanges and on the outside surface will be determined. It is intended, then, to thermal cycle the flanges and check their dimensional stability. 2.2 FREEZE VALVES Two Calrod-heated valves,3 of the type proposed for MSRE operutional valves, were operated in the RE engineering-test loop (Sec. 2.10). A photograph of one of the valves is own in Fig. 2.3. eir arrangement in the loop was shown previously (see Fi 5 in ORNL- ). The valves are located in the test loop 5o that salt may be diverted to either of two drain tank; cservoirs are placed on the vertical risers to ensure thot the freeze zones will not run dry following a drain. Helium was permitted to flow through one valve and reservoir at the rate of 2.2 cfm to check the operation of the reservoir after a system dump. The other valve was closed during this operation. The gas pressure was then equalized 7 UNCLASSIFIED PHOTO " 36713 PIPE HEATER FREEZE ZONE I Fig. 2.3. Freeze Valve. 33 between the drain tank and circulating system, the valve was frozen, and x-ray photographs were taken of the valve assembly. The photographs indicated that sufficient salt remained in the valve to make a seal. The valve proved to be gas tight when refrozen. The valve to the flush tank was operated with cooclant salt through 40 cycles without difficulty. The melt time varied from about 3 min at 1.5 to 2 kw of input to about 15 min when the only heat provided was from losses from the loop. 2.3 HEATER TESTS 2.3.1 Pipe Heaters A prototype, clamshell heater-insulation unit for the MSRE piping was built and tested (see pp 28 and 29 in ORNL-3122). When the pipe wall was maintained at the design temperature of 12000F, there was a heat loss of 550 w from a 1-f% length of 4-in. pipe. The average surface temperature of the insulation was 2LOOF except at the latch closure, where the temperature was 450C0F., Heaters of this type were eliminated from the MSRE design because of their high heat loss and the difficulty in their removal and installation with remote maintenance tools. A heater of the type shown in Fig. 1.9 is being buillt for testing. 2.3.2 Core Heaters The reactor vessel is to be surrounded by a furnace containing 63 hairpin resistance heaters.u The heaters are nested in nine banks of 2-in. stainless steel tubes; each bank of tubes is replaceable separately. An Inconel-sheet reflector and ceramic insulation are provided to minimize heat losses from the furnace. A test mockup of a full-size bank of heaters was set up in order to evaluate heater and furnace performance and replaceability. Figure 2.4 shows a photo- graph of one of the test Inconel heaters, which has a resistance of 0.2 ohm. Both Inconel and stainless steel heaters were used. TFigures 2.5 and 2.6 show the test facility which contains nine heaters, containment tubes, reflector high- temperature insulation, and a heat sink replacing the core vessel. The system was operated for 3000 hr without difficulty. Visual inspection was made of the heaters after 576 hr and again at 3000 hr. There was little change in the appearance of the equipment. The entire inside of the furnace, heaters included, had oxidized to a soft black, but there was no indication of severe attack in the form of flaking. The Inconel heaters appeared in better condition than the stainless steel heaters. The stainless steel containment tubes had a brassy-to-black mottled appearance, compared with a uniform black on the Inconel. The reflector and heat sink faces were also uniformly blackened. The organic binder in the Cerafelt (Johns-Manville Co.) insulation had been burned out to a depth of 1—1/2 to 2 in. at the hot face. The insulation had little remaining structural strength, although the insulating qualities were un- affected. Shrinkage of the insulating material was 1.5 to 2%. Shrinkage of the insulation on a system of this size (9 ft high and 4 ft wide) causes serious heat losses unless some provision is made to avold air leakage. Leakage on the test stand was stopped by plugging the cracks with Fiberfrax, a loose-wool high- temperature lnsulation (product of Carborundum Company ). 1" ft 34 UNCLASSIFIED PHOTO 36550 ‘/%—in, INCONEL ROD ELECTRICAL CONNECTOR INSULATION FOR THERMAL NEUTRON COVER SHIELD PENETRATION [ E 'fi,/%—in. ROD TO ¥g-in.- TUBE TRANSITION \%~in.-on, 0.035-in.-WALL INCONEL TUBE HAIRPIN HEATING ELEMENT ' INSULATORS Fig. 2.4. Core Heater Element. 35 UNCLASSIFIED ORNL-LR—DWG 63249 0.031-in. INCONEL REFLECTOR— 4| f"fl 7in. : |1 TUBE SHEET — 777 INCONEL TUBE HAIRPIN HEATER IE/ (3/g-in.— 0D, 0.035—in.—~WALL) - «—HEAT SINK (0. 425 —in- INCONEL) b\ — COOLING TUBES 4-in, CERAFELT INSULATION (JOHNS-MANVILLE CO) — ™ 304L SS CONTAINMENT TUBE (2-in.- 0D, 0.0GS—in.—WALL)/U:U: 3in——» }<— INSULATION (HI-TEMP) REFLECTOR CERAFELT INSULATION HEAT SINK AND COOLING TUBES CONTAINMENT TUBE SHEET INSUL ATION ¥ _INSULATION INSULATION T0P Fig. 2.5. Profile of Core Heater Test, a Full-Scale Mockup of One-Seventh of the Core Heaters. UNCLASSIFIED PHOTO 36551 Fig. 2.6. Core-Heater Test Assembly. 37 Tests of commercial high-temperature insulations are in progress; at least one of the samples does not shrink at operating temperature. [The Fiberfrax ceramic material without organic binder (2300°F)]. When the experimental apparatus was designed, the total heat loss was esti- mated to be 5 kw with the heat-sink at 1000°F. Heat losses caused by insulation shrinkage, warping of the test stand, etc., made it necessary initially to in- crease the input power to 14.6 kw to achieve 1000°F operation of the heat sink. After the air leaks were stopped, the average input power was 5.1 kw for 1000°F at the heat sink, with no heat being removed by the coolant tubes in the sink. Runs were made alsc with various rates of heat withdrawal, but these data have not yet been analyzed. Preliminary estimates indicate 15 kw can be supplied with the heaters operating at a maximum temperature of 1400°F and the heat sink at 975°F. 2.4 DRAIN TANK COOLERS The system for removing after-heat from each MSRE drain tank has 32 thimbles projecting intc the molten salt. Inside each 1.5-in. thimble is a l-in. bayonet boiler tube in which steam is generated at low pressure. There is air in the gap between the thimble and the bayonet tube. Three full-scale thimbles (Fig. 2.7) were installed into a tank to be filled with a mixture of carbonate salts for testing. Preliminary tests without salt are in progress to obtain the system heat losses and the radiant-heat transfer coefficients in the cooling tubes. After salt is added, the cooling capacity of each tube will be determined, control of the steaming rate will be investigated, and the ability of the system to withstand thermal shocks will be tested. 2.5 SAMPLER-ENRICHER DEVELOPMENT 2.5.1 Sampler-Enricher Concept A schematic layout of the MSRKRE sampler-enricher system for the primary loop is shown in Fig. 2.8. The components and operating techniques are similar to those reported previously.5;6 When the location of the MSRE primary pump Wwas changed, the angle of inclination of the transfer tube which connects the pump bowl with the dry box was increased from 45° to 54-1/29; this aids in inserting the capsule. The location of the dry box was also changed from the south side of the reactor pit to the east side. 2.5.2 Sclder-Freeze Valve In order to repair the sampler-enricher system components located outside the reactor cell, a reliable maintenance valve must be provided. A solder freeze valve was being considered because of its potential reliability. In initial tests, a 1l-in. valve of this type appeared to perform as expected, with an acceptable leak rate of < 108 ce of helium per second.! A full-sized l-l/z-in. single-seal INOR-E valve using the same solder was then constructed and thermal cycled 22 times. The valve was at the operating temperature of lEOOOF for a total of 29.5 hr. Three cycles failed completely to seal, and the remaining cycles were not consistent in the measured leak rates. This apparent unreliability has not been explained, but examination is continuing. UNCLASSIFIED PHOTO 37231 STEAM CHEST N . - 1-in. COOLING BAYONETS (PARTLY WITHDRAWN) N\ = \ N\ N\ ) | N B ~ . 1Yo ~in. CONTAINMENT THIMBLES Fig. 2.7. Drain-Tank Cooler Test Assembly. 39 ORNL-LR-Ivg. 55006 CONTAINMENT/ VESSEL SHELL Unclassified PLUG EXHAUS] HOOD \P” TRANSPORT CASK $$0 AREA 4 3 TO YACUUM PUMP | / : (S stuie TRANSPORT CONTAINER wr i' QEMOVAL VA\LV/\E (ByH ACCESS DoO CELLILLUMINATOR ‘ B WELDED ACCESS BI(R ] 7 2 A / N e b ;“e.l — ZINT eAPSULE , B L% £ pepiscope pocess Br s v T - | T0 vacuum -F’_liMP [INAARRRRT fi@ AREA] T—LATCH S 1c L — CAPSULE p =, MANIPULETO o SAMPLE TRANSPORT — D= % AREA 3 CONTAINER 70 VACUUM - Jic—a [M]) ['SEAL PUMP To Watuum puMp =—=p2/ = AREA 2¢ DISCONNECT FlTANGE ACCESS Dol [ Yo OPERATIONAL VALVE —/WANGE LANGE FOR REMOVAL rragr EONTAINII:!ENT SHELL or\_ r' [ AREA 7b MAINEENANCE |"’"1 AQEA Ib VALY —r 3 AC il proéfu?s HEATER A ELEV. 852 | BLDG, 7503 A A A AranAi ey A " . Pttty g?;{ Ll INDICATES BUFFER ZONE DISCONNECT FLANGES ouTeR STEEL sHELL EXPANSION TR STEEL S JOINT LATCH STOP T CAPSULE GUIDE PUMP BOWL Fig. 2.8. Schematic Layout of MSRE Sampler-Enricher System. 40 For reactor use, a buffered double-sealing valve would be required. DBecause of the apparent unreliability noted above and the complications produced by the gas controls necessary to make the double-solder seal, an alternative design is being prepared. This uses a buffered double-sealing gate valve. 2.5.3 Sample Capsule In order to prevent oxygen contamination of the sample being isolated and of the circulating fuel, the oxide film on the copper sample capsule must be removed. A pretreatment consisting of 4 hr of hydrogen firing at 14759F removed the oxide film from the capsule. On the basis that molten fluoride salt does not wet the surface of oxide- free copper metal, an attempt was made to prepare a re-usable capsule from which a single slug of salt would be removed. However, the salt adheres to the metal surface tightly enough to prevent its removal without damage to the capsule. A capsule was then designed from which the salt can be completely removed by crushing the capsule. The salt is then removed in the form of a coarse powder. Capsules of two diameters (1/2 in. and 3/& in.) have been tested. Preliminary crushing tests in a hot cell indicated that the salt would crush out of the larger capsule more easily. The upper portion of the smaller one tended to col- lapse, trapping salt in the lower part. 2.5.4 Sample-Transport Container and Removal Seal A transport container was designed which will seal the hydrogen-fired cap- sule 1in a dry, inert atmosphere for shipment to the sampler dry box and for trans- porting the salt-filled capsule to the hot cells. The transport container is inserted into the dry box through a buffered double-sealing section located above the removal valve (see Fig. 2.8). The seal provides a purge zone to pre- vent the contamination of the dry box with oxygen and the release of fission- product gases from the dry box to the atmosphere. The leak rate of the container and the seal section will be tested after each installation has been completed. 2.5.5 Sampler-Enricher Design Detail design of the various components of the sampler-enricher system is in progress. Preliminary design of the transfer tube, the maintenance valve, and the operational valve are complete. A preliminary instrumentation flow sheet 1is complete. 2.6 MSRE CORE DEVELOPMENT Additional tests were conducted in the cne-fifth-scale MSRE core model.8 A report was issued summarizing the results of this program. 2.6.1 Heat Transfer Coefficients in Lower Head of the Reactor Vessel Heat transfer coefficients were determined experimentally in the lower head of the one-fifth-scale core model at the reactor center line. Figure 2.9 is a plot of these coefficients predicted for the MSRKRE as a function of the reactor flow rate. The lines in Fig. 2.9 represent the theoretical slopes for laminar and turbulent heat transfer. Irom this it can be seen that the character of flow in the lower head changes from turbulent at the rated flow rate to laminar at a reduced value of the flow rate. 4] UNCLASSIFIED CRNL—LR— DWG 63250 — & S 600 1 | T T T T T h SLOPE = 0.8 FOR TURBULENT HEAT TRANSFER : D [ o« ~ 400 o el (£ 'GL a /é/ - B < o 29 L fran) (o] T, 7 W £ 200 e =7 S jm} a a—] o V/ . | 5 /O | O o w x n ’<\SLOPE = ’/3 FOR LAMINAR HEAT TRANSFER I 100 200 500 1000 2000 3000 REACTOR FUEL FLOW RATE (gpm) 100 Fig. 2.9. Predicted Wall-Heat Transfer Coefficients at the Center of the MSRE Reactor Vessel Head, Based on Measurements in the One-Fifth-Scale Model. 2.6.2 Tuel-Age Measurements in the Lower Head of the Reactor Vessel Fuel-age measurements were determined throughout the entire lower head of the one-fifth-scale model; however, most of these measurements were made with a reactor configuration which has since been modified slightly. From these ecarlier measurements it was concluded that the most critical region, that is to say, the region with the highest combirnation of fuel age, power dengity, and importance was at the reactor center line. For the present core configuration, the fuel ages were redetermined along the center line. Figure 2.10 is a plot of these age measurements as a function of the distance from the wall and with parameters of reactor flow rate. It i1s not apparent why the fuel age is at a maximum at about 2 in. from the wall. It may indicate a zone of recirculation of the fuel. Another possibility is that the mere existence of the probe in the model has changed the flow characteristics. At any rate, the point of greatest interest is that next to the wall. Figure 2.1l is a cross-plot of the reciprocal of the flow rate vs the fuel age at the wall. One would expect that the fuel age is related to the flow rate as follows: constant {units of volume) volumetric flow rate fuel age = Figure 2.11 secems to substantiate this except at low flow rates, where the age increases less rapidly than would be predicted. 1f, however, the character of flow has changed from turbulent to laminar when the flow rate is reduced as pro- posed above, then it seems reasonable to expect the constant in the above equation to change. This could also explain the radical change of shape of the curve of lowest flow rate in Fig. 2.10. Based on heat transfer coefficients, fuel ages and power-density distributions in the fuel and vessel wall, the outside-wall temperature of the lower head of the reactor vessel at the center line position has been estimated to be 1215°F, which is 40°F above the iniet temperature of the reactor. This wall temperature rises to 1226°F at 50% of design flow rate. This estimate is with the reactor operating at design conditions and with no solids deposited at the wall. 42 UNCLASSIFIED ORNL—LR-DWG 63251 140 ‘ FLOW RATE OF REACTOR FUEL (gpm) 120 A 322 ~ 100 LAMINAR FLOW > \ 80‘ /U, | ‘\\\fi 561 60 FUEL AGE AT POSITION INDICATED (sec) 40 o ~7Y 1120 — ° TURBULENT FLOW T—4 20 pD—m 1400 0 0 1 2 3 4 5 (3] DISTANCE FROM WALL IN LOWER HEAD OF REACTOR VESSEL (in.) Fig. 2.10. Fuel Age Predicted at the Center Line of the Lower Head of the - Reactor Vessel. It should be noted that after the above measurements were made, the design of the reactor drain line was modified. The drain line now projects above the wall at the center line of the head, and some minor changes will probably be pro- duced in the flow distributions reported above. 2.6.3 Studies on the Disposition of Fine Particles in the Lower Head of the Reactor Vessel Experiments were conducted with the one-fifth-scale model to determine the disposition of particulate matter in the core. The particles could result from at least two sources: corrosion of the contaimment material and oxidaticn of the fuel. For these tests, molybdenum powder with an average diameter of about 5 p was used. This simulated 25-p particles with a specific gravity of about 10 in the fuel salt. 1In one typical run, 70 g of molybdenum powder was added to the lower head of the model, and the model was operated at the design flow rate for a total of 4O hr. After the first 21 hr the model was disassembled and it was found that about two-thirds of the powder had been swept out and that the rest - UNCLASSIFIED ORNL—LR—DWG 63252 FLOW RATE (gpm) 1400 1120 842 561 322 80 . T ] A 70 - A c0 ! | | ! 3 50| - i —J - / -15 6- / O’8 = CIRCUMFERENTIAL PRINCIPAL STRESS -20 / O‘¢ = MERIDIONAL PRINCIPAL STRESS -25 0 1.87 3.74 5.62 7.4% MERIDIONAL POSITION (in.) 9.36 Fig.2.19. Thermal Stress Profile — Sphere. Ten-Mw operation; about 200 acfm of cooling air. UNCLASSIFIED ORNL—LR—DWG 63259 - SPHERE JUNCIHON TOP FLANGE - VOLUTE l Ox INSIDE\\// - 1 / o, INSIDE g, OUTSIDE v e —— L - - [’k ~d 7 ol S -~ — - -..-‘ I g QUTSIDE / o, = AXIAL PRINCIPAL STRESS oy = CIRCUMFERENTIAL PRINCIPAL STRESS oy OUTSIDE = —o, INSIDE Ox INSIDE | x -6 AXITAL 2 Fig. 2.20. Thermal Stress Profile — Cylinder. acfm of cooling air. POSITION (in.) Ten-Mw operation; about 200 54 Thermal stress calculations indicated that a small quantity of air cooling will be required on the fuel-pump tank during zero-power operation to avoid ex- cessive strain cycling during reactor power changes. The calculations also indi- cated that forced air cooling of the coolant pump will not be required since the stress range between zero-power and 1l0-Mw coperation is relatively small. Calculations have been initiated to determine the necessity for variable control of cooling air flow for the fuel-pump tank under all operating conditions from zero to full power. The stresses shown on Figs. 2.19 and 2.20 were calcu- lated by using an ORACLE code which assumes the spherical shell to be at constant temperature. The effect of the thermal gradient on the thermal stresses in the sphere are being determined with the IBM 7090 code which uses tangent cone geometry in place of the sphere. 2.8.2 Water Test of MSRE Coolant Pump Tests to determine the feasibility of conducting the coolant pumpl9 hot tests in the fuel-pump volute have been completed. The tests indicated that the radial hydraulic imbalance would cause rubbing between rotary and stationary parts. It is planned to conduct the hot-proof test of the rotary element for the MSRE coolant pumps in the hot-test stand for the fuel pump, using a fuel-pump impeller. The water-test rig is being modified to permit hydraulic tests of the coolant pump impeller-and-volute combination. 2.8.3 Advanced Molten-Salt Pumps Pump Equipped with One Molten-Salt-Lubricated Bearing The pump containing one salt-lubricated journal bearingzo continued to oper- ate, circulating LiF-BeFp-UF), (62-37-1 mole %) at 1225°F, 1200 rpm, and 75 gpm. The pump has operated for 11,772 hr and has been started 62 times thus far. There has been no visible leakage of oil from the shaft seal. 2.9 GRAPHITE-MOLYBDENUM COMPATIBILITY TEST The two natural convectilion loops were shut down and examined, as reported in the metallurgy section of this report. 2.10 MSRE ENGINEERING-TEST LOOP The engineering-test loop21 began operation April 20 with MSRE coolant salt (LiF-BeF5, 66-34 mole %). The purpose of the first test was to evaluate the effectiveness of the flushing operation by following the oxide content of the salt immediately after filling the system for the first time. 2.10.1 Oxide-Flush Run The loop remained in operation, isothermally at llOOOF, for approximately 1300 hr while 50 samples were removed from the pump bowl and drain tank. The samples were removed in a hydrogen-fired copper dip tube by the use of a ball valve and sliding seal arrangement. Both the dip tube and the sample were kept in a helium atmosphere until loaded into the analyst's dry box. The results of the oxygen analyses were so scattered that it was not possible to determine how much oxygen had been dissolved or whether the solubility had been exceeded. As seen in PFig. 2.21, the variation reported in oxygen content was from 300 to 1100 ppm. 55 UNCLASSIFIED ORNL—LR—DWG 63260 1200 , [ 1000 BeO ADDED 5 800 . . . = L J = 3 ° = N : = H g 600 i © L 2 . [ ] o ° H o ° - S I O 3 L 400 o - . o e °83 . [ [ 2 e *° % 200 0 0 200 400 600 800 1000 1200 OPERATION TIME (hr) Fig. 2.21. Results of Oxygen Analyses, Engineering Test Loop. Between the 950th and 1100th hr (see Fig. 2.21), approximately 35 g of BeO pellets were added to the system through the pump bowl. This was sufficient to raise the entire oxygen content approximately 200 ppm; such a difference did not show up in the analyses, however. From this and other tests on the accuracy of the oxygen analysis, a re- evaluation of eguipment and technique was begun by the analysts. The engineering- test loop was dumped after 1300 hr of operation to begin other tests., The chro- mium content of the salt increased from 170 ppm to 370 ppm during the above oper- ation. This represents leaching of all the chromium from about a l/Z-mil thickness of the metal surfaces of the loop. The chromium pickup 1s believed to be reason- able for the initial operation of an Inconel system. 2.10.2 TFreeze-Valve Operation The loop freeze valves and thelr operating characteristics are described in Sec., 2.2. The first attempt to dump the loop after the 1300-hr operation required over an hour, which is excessive. The difficulty is thought not to have been in the freeze valve but in the drain line. The alternate valve (fuel freeze valve), which had seen the same conditions, opened in normal time. The partial plug may have been caused by the settlement of undissolved BeO particles added to the loop a week earlier. An attempt to establish a similar plug will be made by adding additional BeO in the same manner. 56 2.10.3 Level Indicator for Molten Salt (Test Loop) The differential pressure or bubble-tube type of level indicator system was installed on the test loop along with the standard spark plug type. A helium flow rate of approximately 1.8 scfh was used continuously through a 1/h—in.-OD tube inserted into the liquid of the DANA pump bowl. The level was recorded from the output of a O to 20 in. of water differential-pressure transmitter. Flow was supplied by a constant-pressure-differential type of regulator. The tube has plugged twice, once during a draining operation and again when attempts were made to duplicate the original conditions. It was later realized that the pressure in the pump bowl could increase faster than the pressure in the bubble tube (considering the volume of the bubble tube and the helium flow rate), causing salt to enter the tube.22 The system volume of the bubble level tube was therefore decreased by a factor of 3, and, with the same flow rate and pressure change rate, the tube did not plug. Since then the bubble tube has been used for 11 fill and drain oper- ations and has been valved off (equalized so that salt rises in the tube to pump level) for 700 hr and valved on for 300 hr of level indication, all without diffi- culty. 2.10.4 (Craphite-Handling Facility An 8-in.-dia vertical graphite container with a longitudinal frozen-salt seal access has been designed and is being fabricated of INOR-G plate. The graphite, on order, will have the same dimensions and properties as that for the MSRE core. The installation will allow the examination of the graphite after various operations, without contamination by air. As shown in Fig. 2.22 a dry box will be lowered over the uccess flange to the freeze seal after the test loop has been drained and cooled. Conventional gasket materials make connection between the dry box handling facility and the graphite container while the box is evacuated and purged of Op. The access flange is opened through the use of the glove ports and the hoist. Two sealed graphite carriers are provided so that a full length (65-1/L in.) sample may be added or removed in an inert atmos- phere. The access flange contains a cone-shaped plug, with cooling-air coils to form the frozen-salt seal in an annulus. A standard metal oval ring gasket is used for the gas seal. The main purpose of the test will be for verification of laboratory-scale tests of the removal of Oz from the graphite, the ease of graphite removal from the container, wetting characteristics, completeness of draining, etc. 2.11 MSRE MAINTENANCE DEVELOPMENT 2.11.] Maintenance Plan The dual maintenance approach, using both semidirect and remote techniques as described in the design section (Chap. 1 of this report), is being implemented. Figure 2.23 shows the building model with the primary-cell upper-shield beams in place. Access to a small component, using the portable maintenance shield and remote mamual techniques, will require the removal of three or four of these beams. Replacement of a large component by use of remote maintenance techniquesz3 will require the removal of all the beams. They will be handled by the building 30-ton crane and a directly-operated 1lift fixture. 57 UNCLASSIFIED ORNL-LR-DWG 63261 —=——GRAPHITE-HANDLING DRY BOX {INSTALLED AFTER LOOP IS DRAINED AND COOLED) ACCESS FLANGE TO NEOPRENE GLOVE GRAPHITE CONTAINER-] TMETAL OVAL RING GASKET kY ! \ A Q : COOLING AIR I= fi 9 (BOTH EXTERNAL AND INTERNAL) : o ! i i l FROZEN-SALT SEAL IN ANNULUS g STORAGE CYLINDER FOR ACCESS COVER | | ~ | e TO GRAPHITE CONTAINER | VACUUM CONNECTION OUTLET A — WL ‘:‘N}p« L] ¥ a2 & —_— D | | GRAPHITE —I7 BB = | HOLDDOWN (DURING VACUUM) i - NN ] ‘ ‘_l " . —_— = & f.. ° T“FLOOR LEVEL E‘L.““— — ) . —= Y GRAPHITE CONTAINER —=| (8-in. DIA) CONTAINER SUPPORT GRAPHITE NN INOR-8 ’-rfl/z—in. SCHED-40 INLET Fig, 2.22. Graphite-Handling Facility for the Engineering Test Loop. | | - UNCLASSIFIED VanTenance covrro. roow I CooLm ceL B REACTOR CELL ¢ / SPENT FUEL STORAGE CELL TOP SHIELD BEAMS 4 / ; WASTE TREATMENT CELL Fig. 2.23. One-Twelfth-Scale Model of Building 7503, Column 3 to Column 9, Sections A, B, and C. Figure 2.2h shows the primary cell with the lower shield beams in place. To obtain access to a small component, the seal membrane (not shown) will be cut away from the two beams to be removed and the portable maintenance shield set over them and opened. The shield beams will then be removed, using the 30-ton crane and an automatic 1ifting tong which can attach to and release from the = beams without direct attendance, and the portable maintenance shield will be closed. The operations of removing the lower shield beams and closing the port- able meintenance shield will be conducted from the shielded maintenance control room visible to the right and above the primary cell in Fig. 2.23. Maintenance to or replacement of the component will be accomplished by work- ing directly through the portable shield. To replace a large component the entire seal membrane will be removed. The manipulator and telev: be set up over the primary cell. From this point until the lowe are replaced, all work must be conducted from the maintenance control room. All the lower shield beams and the two support beams will be removed. Maintenance or replacement will be accomplished with the manipulator, with vision provided by the stereo-television system and through shielded windows in the maintenance control room. 2.11.2 Remote-Maintenance Mockup The remote-maintenance mockup of the MSRE is being constructed to develop, test, and improve maintenance tools and techniques.2% It occupies the space - 59 UNCLASSIFIED == PHOTO 37263 i o PUMP MOTOR PIPE HEATER PUMP P PUMP BOWL HEAT EXCHANGER FREEZE FLANGE Fig. 2.24. One-Twelfth-Scale Model of Primary Cell. 60 left vacant following dismantling of the remote-maintenance demonstration facility in the southwest corner of Building 9201-3. Most of the new structural steel was erected, and the pump, reactor thermal shield, and some piping were installed. Shop-fabricated models of the heat exchanger and primary loop piping are being installed. Heater, component, and instrument replacement appear to be straightforward maintenance operations, except in areas of crowding, such as around the pump. Extensive practice in the closely-duplicated pump mockup will resolve any problem in this area. At present, it appears that the inventory of special long-handled tools will be small because many operations have been reduced to the rotation of a standard hex nut or to the vertical translation of a standard eyebolt. Making and breaking the freeze flanges, which is the primary concern of the mockup, includes the following: (1) removal of the freeze-flange clamp, (2) stowing the clamp, (3) jacking the piping to separate the faces of the flange for gusket replacement or to replace a component, and (4) providing force and visibility to re-position misaligned flanges. Test equipment for this process was designed (Fig. 2.25) and is being fabricated. UNCLASSIFIED ORNL-LR—DWG 63262 LIFT HOOK (@) REMOVABLE um PIPE HEATER Al (@ I SOCKET WRENCH LIFT HOOK AU FLANGE CLAMP m’ N FREEZE g | I3 FLANGE - \. N * JACK INSTALLATION CABLE TO DISENGAGE JACK HOOK PERMANENT INSULATION \ PIPE SUPPORT STRUCTURAL STEEL Fig. 2.25. Equipment for Jacking Flanges Apart in the MSRE. 61 2.11.3 MBRE Model The one-twelfth scale model of Building 7503 follows closely the reactor design and has aided in working out several design changes. The building and the primary cell are shown separately in Figs. 2.23 and 2.24, respectively. 2.11.4 Portable Maintenance Shield Conceptual layout drawings for a portable MSRE maintenance facility were com- pleted. Orders were placed for approximately 80% of all materials required for fabrication; most of these items have been received. Component detailing was started. 2.1l2 BRAZED-JOLINT DEVELOPMENT To standardize tools and techniques, all points in the salt-system piping that will require a braze joint in the event of component replacement will be 1—1/2—in. sched-40 pipe run on a slope 3° from horizontal. A clear or readily clearable space 3 ft along the axis of the pipe and 1 ft on either side; a perma- nent base-plate 3 ft long and 2 ft wide, mounted 1 ft below and parallel to the axis of the pipe; and access from above are required at each point. After testing several designs, it was tentatively decided to use a sleeve joint with a 1-in. engagement and a 3° taper, shown disassembled in Fig. 2.26 and assembled and sectioned in Fig. 2.27. The male stub remaining in the system will be machined in place to the required taper. The replacement component will carry the female sleeve with the 82% Au - 18% Ni braze metal preplaced. The present test pieces use two braze-wire rings preplaced in grooves in the sleeve (Fig. 2.26). The Metallurgy Division is developing a technique to "tin" the sleeve with about 0.005 in. of braze metal so that only cone surface will have to wet during brazing. The brazing will be accomplished by heating the joint to about 1800°F in an induction furnace similar to that of Fig. 2.26. The furnace requires electrical leads, cooling water, and purge gas. The furnace will be preplaced with the new component and will be destroyed upon completion of the braze. Figure 2.27 shows a completed, sectioned joint. Although braze metal did not flow the entire length of the anmulus, an adequate seal was formed. Development efforts are continuing in order to ensure wetting of the entire annulus. A photo- micrograph and discussion of the Joint is in the Metallurgy section of this report (Part II, Materials Studies). Tools for use in the reactor have been designed to hold the pipe, cut it, machine the taper, and assemble the joint. These will be tested in simulated reactor geometries. UNCLASSIFIED PHOTO 36994 TNGHES e 7 Fig. 2.26. Disassembled Test Furnace and Joint. UNCLASSIFIED T-20322 ONE INCH Fig. 2.27. Assembled and Sectioned Brazed Joint. 64 2.13 MECHANICAL-JOINT DEVELOPMENT The molten-salt reactor program requires simple, leak-detectable, radiation- resistant joints to permit component replacement and maintenance, and several types of disconnects have been developed (Fig. 2.28). The integral dual-seal ring joint25 closed by steel spring clamps offers a satisfactory disconnect for many reactor applications in the process temperature range 100 to 1500°F. Remote operation is facilitated because only low torque is required to operate the two bolts on the clamp. Use of the metal-to-metal seal eliminates material radiation damage problems. Integral ring joints were selected in preference to separate rings to prevent ring buckling when the connecting flange halves are slightly cocked on makeup, to reduce the number of seal contacts from four to two, to reduce the probability of foreign particles into seal surfaces, and to eliminate the need to handle gaskets during remote operation of the disconnect. Two integral ring-joint spring-clamp flange sets were designed, built, and tested for coupling E?z- and l—l/z-in. pipe, respectively. Final helium leak rates with the 1/2-in. unit following eight assemblies and & total of 1020 thermal cycles between 100 and 1200 to 1500°F were 9 x 10-6 atmospheric cc/sec at room temperature, and 2 x 10-7 atmospheric cc/sec at 1500°F. Following fourteen assem- blies and 69 steam cycles up to 400°F with the 1-1/2-in. unit, no detectable readings above background were observed in a helium mass-spectrometer test on both UNCLASSIFIED PHOTO 369984 INTEGRAL DUAL-SEAL BLOCKS, YOKE MOUNTED RING-JOINT SPRING-CLAMP FOR E HIGH-TEMPERATURE SERVICE COMBINATION RING- JOINT FLEXITALLIC FLANGE | SINGLE-CONE SEAL BLOCKS, FOR YOKE MOUNTING Fig. 2.28. MSRE Disconnects. - 65 inner and ocuter seals, assuring a seal tightness in excess of 1.3 x 10-8 atmos- pheric cc/sec. The joint was also thermal cycled (250°F AT) with bending moments of 4200 in.-1b applied both parallel and perpendicular to the clamp separation line. ©No detectable leakage was noted. A combination ring-joint - Flexitallic disconnect20 was developed to serve as a multiple disconnect in crowded spaces. The joint contains a tight R-26 ring at the periphery, and four l/Z—in. lines, each with a Flexitallic gasket. It 1s most useful where small amounts of interline leakage can be tolerated but leakage to the outside cannot be accepted. Farly difficulties in achieving tight ring seals were eliminated by modifying the usual flange-assembly procedure. In the successful procedure the center of the joint is loaded temporarily with a C-ciamp before the flange-bolts are tight- ened. Seals can then be consistently made with bolt-torque loads well within the maximum limits recommended for the bolt threads. Proper seals can be obtained by using this procedure, even with rings manufactured only to commercial specifi- cations. The addition of the Flexitaliic inserts appears to add no further diffi- culties. No leakages in excess of 1 x 10-6 ¢td cc/sec were apparent for ring seals (and 1 x 10-3 sta cc/sec for Flexitallic seals) in tests run with oval and octag- onal rings and four l/Z—in. 150-1b Flexitallic inserts reused for three successive assemblies. The flange was successfully steam-cycled to 250°F with 500-psi helium pressurization in the buffer zone of the ring. Development testing with cone-seat disconnects®! demonstrated that such Joints can be made to adhere to the low leak rates required in reactor service. Both single and multiple units were tested. Cone-seat units represent a great simplification over other commonly used metal-to-metal seals for guick disconnects. Low-torque single bolt loading, compactness, minimum space required for assembly (l/H—in. travel), interchangeability of parts, and inexpensive construction are some of the major advantages of this seal. The cone-seat disconnects are well suited for reactor leak-detector-iine disconnects. Development work with ycke-mounted l/fi— to l/B—in. pipe dual-seal block dis- connects revealed that seals fashioned to variations of the ring-joint gecmetry work reliably repeatedly. Test biocks with both solid and slotted integral oval rings operating within 60°-included-angie grooves were operated in excess of 12 assemblies and 30 thermal cycles from 80 to 3259F, with leak rates of less than 8 x 10~7 std cc/sec as measured with a helium mass spectrometer. Additional test- ing 1s planned. Two commercial quick-disconnect ccocuplings were received from the On-Mark Company. Preliminary testing indicated good sealing action at room temperatures. On thermal cycling up to 3250F, however, one unit failed to seal on the tenth downcyecle. Visual inspection revealed considerable peeling of the seal's silver plating. Additional testing is scheduled. 2.14 STEAM GENERATOR Preliminary design of a high-pressure, electrically heated, single-tube, steam-generator - superheater was initiated. The generator would be a bayonet- tube unit of the type proposed for the experimental molten-salt fueled 30-Mw power reactor.20 Completion of the design was postponed until the feasibility of a separator that will fit the small available space is determined. A test chamber was designed and is under construction to test separators for use in the steam generator. TIigure 2.29 is a schematic drawing of the test UNCLASSIFIED ORNL-LR-DWG 63263 §§“\———TEST SEPARATOR AIR AND "CARRYOVER NN i ol . # WATER AND "CARRYUNDER Fig. 2.29. Separator Test Chamber, 67 chamber. Alr and water will be used in the tests. Water carryover in the air and alr carryunder in the return water will be determined for several air and water flow rates. 2.15 MSRE INSTRUMENT DEVELOPMENT 2.15.1 Pump Bowl Level Indicator Development of a continuous level-element®? for use in measurement of molten salt level in the MSRKE fuel and coolant salt pump bowl 1is continuing. As reported previouslycY a graphite float assembly will be used in the level element of the reactor pump bowl, and an INOR-8 hollow-ball float will be used in the level ele- ment of the coolant salt pump bowl. Design of the graphite float and core assembly has been completed, and two prototype units have been fabricated. This assembly (see p 66 in ORNL-3122) has been designed to provide sufficient buoyancy, after absorption of fuel salt by the graphite, to support the core assembly and to over- come frictional forces. A molten-salt-level test stand for use in testing prototype level transmitter was designed. Construction of this stand, which is shown in Fig. 2.30, 1s nearing completion. The major effort in this program has been devoted to the development of a differential transformer and core assembly which will be capable of contlnuous operation in the range of 1200 to 1300CF and which will not exhibit excessive shifts in zero, span, and phase of the transmitted signal as the temperature varies over the operating range of 250 to 13009F. Several transformer designs have been tested during the course of this work. A transformer design using a ceramic-insulated nickel wire, manufactured by the Secon Metals Corporation, with Fiberfrax paper insulation between the windings, chowed some initial promise of success but falled when operated at elevated temperatures. Two transformers of this type were tested. One developed intermittent shorts after 12 hr operation at 1100°F. This failure produced rapid and erratic changes in the output signal. Subsequent examination revealed that the insulation had cracked off the wire in places, and several turns were shorted. Another transformer of this type failed in the same manner at QOOOF. Figure 2.31 shows a recent transformer design which promises success. Insula- tion ig provided by three concentric, unfired, grade A lava sleevec. Bare nickel wire is wound into threaded grooves machined into the inner two sleeves, forming the primary and secondary of the transformer. The threcads act as insulation for the wire., The external sleeve servegs to keep the wire in place. DSufficient clearances are allowed to permit differential expansion between the wire and the sleeve. A transformer of this type has accumulated more than two weeks of operation at tem- peratures in excess of 1200°F., This period included 24 hr of operation at 1500°F and several periods when the temperature was reduced to room temperature. 5So far, there has been no evidence of insulation failure. Calibration data obtained from this transformer are also very encouraging. Figures 2.32 and 2.33 show calibration curves obtained at various temperatures with constant current excitation and with constant vollage excitation. Note that the sensitivity of the transformer, as indicated by the slope of the curves, in- creagses with temperature when constant current excitation is used and remains essentially constant when constant voltage excitation is used. This result was entirely unexpected. It had been expected that the increase in transformer losses would result in & decrease in sensitivity as the temperature was increased when constant voltage excitation was used. A probable explanation for the increase in 8 UNCLASSIFIED PHOTO 37084 | Fig. 2.30. Molten-Salt-Level Test Stand. ) LR R R S R R Dl B ll P R E 5 i AR lLo 1 2 3 a INCHES Fig. 2.31. Lava-Insulated Differential Transformer. UNCLASSIFIED PHOTO 37261 69 70 UNCLASSIFIED ORNL-LR-DWG 63264 50 1200°F ‘ 15.5 mv/in. / N | \»4 900°F ] p i0.2 in, 100°F S mvn 12.6 mv/in. /. - ® ™ ~ ° — ® -~ > 30 A £ ° o 5 - I z - 2 © 20 . cb/.’ — 10 — ROOM TEMPERATURE 5.6t mv/in, . ‘ [ . H 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 DISPLACEMENT {in.) Fig. 2.32. Calibration Curves for Lava-Insulated Transformer, Constant Current Excitation. UNCLASSIFIED ORNL-LR-DWG 63265 QUTPUT (mv) 500-mv AC EXCITATION /‘ 0/ /‘. CONSTANT VOLTAGE , __ IRON CORE / / I o 7' F = 1000 cps 8] 0.5 1.0 1.5 2.0 2.5 3.0 3.5 DISPLACEMENT [in.) Fig. 2.33. Calibration Curves for Lava-insulated Transformer, Constant Volt- age Excitation. 71 sensitivity isc found in the data presented in Fig. 2.34. These data, extracted from the literature,3C indicate that the intrinsic induction (permeability) of Armco iron increases with temperature at Low excitation levels. Thic increase in permeability would result in an increase in coupling, with & resultant increase in transformer sensitivity. The fact that the increase in sensitivity due to the permeability effect almost exactly offsets the decrease due to transformer losses 1is apparently a fortunate coincidence. The offset between the curves in Figs. 2.32 and 2.33 is due to residual voltage capacitively coupled through the transformer. This offset can be elimi- nated by appropriate circultry in the receiving instrument. Figure 2.35 shows a calibration curve obtained by using a Foxboro dynalog recorder, with special input circuitry, as the receiving instrument. Figure 2.34 indicates that the curie point of the Armco iron core occurs at 1L0CPF. Thic effect has been verified experimentally. Tests have shown that the transformer becomes insensitive, and signal output repeatedly falls to zero when the temperature is raised above 1400°F. This characteristic is not believed to be objecticnable since the point where the failure occurs is above the expected operating temperature range and since the failure is in a safe direction. How- ever, an attempt is being made to extend the useable range of the transformer. UNCLASSIFIED ORNL-LR-DWG 63266 3.0 e 20— 4 | T H=0.3 INTRINSIC INDUCTION (kilcgauss) 200 400 600 800 1000 1200 1400 1600 TEMPERATURE (°F) Fig. 2.34. Armco Iron Magnetization vs Temperature. 72 UNCLASSIFIED ORNL-LR-DWG 63267 5 4 £ o 3 ] = o = W x 0} S a 2 b= >— O 1000°F 1 R . _ — 1250°F 0 O 1 2 3 4 5 13) ACTUAL POSITION (in.) Fig. 2.35. Calibration Curves for Lava Transformer with Dynalog Readout. A cobalt core, which has a higher curie point, has been fabricated. Tests will be run to determine the feasibility of using this core with the present trans- former design. The lava-insulated transformer assembly is being mounted on an Inconel coil form which will give mechanical protection to the coil. The complete transformer assembly will then be tested to determine whether the transformer characteristics are adversely affected by the coil form. If the results are satisfactory, a prototype level transmitter will be assembled and tested on the molten-salt-level test stand. 2.15.2 Single-Point Level Indicator As previously reported,3l a conductivity type of level indicator is being developed for use in the single-point measurement of the molten-salt level in the MSRE storage tanks. Extensive measurements of the resistivity of a molten salt have been made at various voltages and frequencies, a new probe-design concept has been developed and tested, and a conceptual design has been made for a two-point level probe suitable for use in the MSRE application. The salt resistivity was investigated to determine the reason for and nature of the observed changes,32;33 in the apparent reslstivity of the salt with changes in excitation voltage and to determine whether any barrier or interface effects existed which would require the excitation voltage to be above some mini- mum value. A spark-plug probe was used for these measurements. With the salt in contact with the probe, a 60-cycle ac excitation voltage was applied between the probe and the salt pot. 73 The excitation voltage was varied from 2 mv ac to 2 v ac, and the probe current was plotted as a function of excitation voltage. The plotted curve was linear over the entire range, thus indicating that there are no interface effects in the region of interest when ac excitation is used. These tests were repeated, using dec excitation and with the sparik plug positive in one case and negatlve in another. The plotted curves were not linear and thelr shape changed considerably when the polarity of the excitation voltage was reversed. From these tests it was concluded that a definite polarization effect exists when de excitation is used and that the use of de excitation ic undesirable. Additional tests were run in which the temperature of the salt was varied; the results have yet to be interpreted. Figure 2.36 shows a new design concept for a conductivity probe. In previous probe designs both the potential and excitation leads were connected at the bottom of the probe, and the signal level, when the salt was below the plug tip, was a funection of the excitation current. This produced a floating-zero effect and required the detection of a small change in a relatively large voltage. In the new design, the excitation lead is connected at a point above the probe tip, and the signal is obtained by measuring the potential between the point where the excltation lead is connected and the probe tip. Since no current flows to the probe tip when the salt is below the probe, a true zero characterictic is obtalned, and the signal is a small voltage deviation from zero. This detected. A model of this probe was constructed and tested. can be amplified and It has operated UNCLASSIFIED ORNL—LR—DWG 63268 ( ) N | B S EXCITATION ] READOUT TEST UNIT NO.2 AS FABRICATED SCHEMATIC | Fig. 2.36. Single-Point Level Indicator. | MOLTEN SALT SALT POT 74 several hundred hours at temperature in both fuel 130 (LiF-BeF2-UF),, 62-37-1 mole %) and MSRE reactor salt (LiF-BeF;-ZrF),-ThFj-UF), 70-23-5-1-1 mole %) without apparent damage. Figure 2.37 shows a calibration curve cbtained using MSRE salt. The UNCLASSIFIED ORNL-LR-DWG 63269 MSRE REACTOR SALT TEMPERATURE: 1369°F / 2.3 amp AT 1-kc EXITATION TEST PROBE NO. 2. ’ T - ] 35 30 25 - - : e - — | 20 -4 // s — SIGNAL {mv, AC) o] 1 2 3 SURFACE AREA OF PROBE IMMERSED IN SALT (in.2) Fig. 2.37. Calibration Curve, Single-Point Level Indicator, curve was obtalned by immersing the probe, which is 3/8 in. in diameter, to a known depth and recording the output signal level. This signal level was then plotted as a function of the calculated surface area exposed to the salt. Although this curve is linear, it 1s not expected that the probe would be useable as a con- tinuous level indicator unless some means can be devised to compensate for changes in the resistivity of the molten salt with time. The curves shown in Fig. 2.38 were obtained using various excitation frequencies. From these curves it can be seen that the sensitivity of the probe is increased by increasing the excitation frequency. The use of higher excitation frequencies has the additional advantage that stray 60-cycle pickup voltages can be rejected with appropriate filter circuitry. Figure 2.39 shows a conceptual design for a two-level conductance probe, based on the new design concept and compatible with the configuration of the MSRE storage tanks. Potential lead 1 is connected to the high-level contact plate. 75 UNCLASSIFIED ORNL-LR-DWG 63270 80 I T ; | | ; | 70 T / MSRE REACTOR SALT TEMPERATURE: 1369°F EXCITATION CURRENT: 2.3 amp 1ke ) // 50 / O < v e £ @ S 500cps} w -J -J T = © ") / 30 / 20 - 10 100cps - / .—-—_'—___ 50 ¢cps 0 0 1 2 3 4 5 6 SALT LEVEL (in.} Fig. 2.38. Signal vs Frequency Characteristics for Single-Point Level Device. Potential lead 3 and the excitation lead are connected to the excitation ring. Potential lead 2 is connected to the high-level contact plate. Excitation voltage is applied between the excitation terminal and the mounting flange. The low-level signal is developed between potential leads 1 and 3, and the high-level signal isg developed between potential leads 2 and 3. All leads are brought out through l/M-in. INOR-8 tubes and are insulated from contact with the probe assembly except at the point of connection. The assembly is all welded, and INOR-8 is the only material which will be in contact with the salt. The assembly will be instailed through a 3-in. ring-joint flange located at top center of the storage tank. 76 UNCLASSIFIED ORNL-LR-DWG 63271 POTENTIAL LEAD | / POTENTIAL LEAD 2 POTENTIAL LEAD 3 EXCITATION TERMINAL MOUNTING FLANGE HIGH-LEVEL CONTACT PLATE LOW-LEVEL CONTACT PLATE Fig. 2.39. Conceptual Design, MSRE Conductance Level Probe, Two Level. 77 2.15.3 Temperature Scanner A temperature-scanning system is being developed to display the output of approximately 250 thermocouples attached to the reactor system pipes and components. These thermocouples will be used to indicate the temperature of the pipes and com- ponents during system startup and shutdown operations. The display of the tempera- tures will be used to prevent excessive thermal stresses of the pipes and components during these operations. An operator will observe the temperature profile of the system and adjust heaters to keep the entire system at a common temperature. The scanner will consist of three separate systems, each capable of handling 100 thermocouples. The thermocouple signals from the pipes and components in one part of the reactor system will be switched at a 2000-point-per-second rate and compared to a reference thermocouple attached to the pipe or component having the slowest thermal response. If a temperature difference, elther high or low, exists between the commutated thermocouples and the reference couple, an alarm signal is produced. The difference signal is simultaneously transmitted to a 17 in. oscil- loscope for display. If all signals are of the same value a straight line will be . seen on the scope. A block diagram of a 100-point system is shown in Fig. 2.LO. UNCLASSIFIED ORNL—LR-DWG 58910R{ 1200 RPM MOTOR | | \ (e} o D’j o) Pam— Vo) | \n : W » X_nj 1 fl: | | i ] i < ) o ! o — : > o 00000 | ] < i . S ! AMPLIFIER 17" 0SCILLOSCOPE W 1 o o \ > | 2§ : Sk ] clo ' x| ONZ ' = S ol - ! L o~ o - ALARM MERCURY JET SWITCH {100 CONTACTS) \ JDISCRIMINATOR] < /) — REFERENCE THERMOCOUPLE AMPLIFIER ANNUNCIATOR \_ )_ [aag o Zz 100 |— - (W) > < b3 w — [fa) o = 'C_) .\ P50 \,O,‘, __‘_ _ _ _ _ - - i ‘ | O 0 0.020 0.040 0.060 0.080 INOR-8 WALL THICKNESS {in.) Fig. 3.5. System Inventory for Reactors with INOR-8 Fuel Tubes. 3.2 NITROGEN=16 ACTIVITY IN FUEL SALT A preliminarg estimate® of the radioactive sources within the reactor cell showed that the N'€ activity from the Flg(n,a)N16 reaction in the fuel salt was one of the major contributors. This estimate was based on a fission-spectrum-~ averaged cross section and the conservative assumption that the total reactor flux had a fission spectrum. The K© activity has been re-estimated by integrating the reacticon cross section obtained from BNL=-325, with the fast neutron fluxes obtained from rmltigroup reactor calculations.”® A lower value was obtained, but N® activ- ity still gives an appreciable dose. At a reactor power of 10 Mw(th), the average production rates obtained were 1.8 x 10° atoms of N8 per sec per cc in the headers and coolant annulus, and 1.1 X 10%° in the core proper. Assuming these production rates constant over the respective regions gave the activities listed in Table 3.2. 90 Table 3.2. Nitrogen-16 Activity in Fuel Salt System Location 0 Activity, (dis sec™ cec™ of fuel) x 101° Pressure vessel entrance 0.27 Core proper entrance 0.3U Core proper exit 0.65 Pressure vessel exit 0.5 3.3 RESIDUAL ACTIVITY IN PUMP BOWL AND HEAT EXCHANGER Two of the major sources of residual activity in the reactor cell will be the pump bowl and the heat exchanger if appreciable quantities of fission prod- ucts plate out in either of these components. The magnitude of these sources was estimated, making certain assumptions about the behavior of the fission products. In addition, the average energy release associated with the decay of these prod- ucts was estimated. The residual activity in the pump bowl was based on the assumption that all the daughter products of xenon and krypton plate out at the point of formation. The method of Stevenson® was used to determine the formation rate of daughter products. Considering the gamma activity to be equivalent to l-Mev photons emit~ ted from a point source, the calculated dose rate 10 £t from the pump bowl was 1000 and 400 r/hr after 10 and 30 days cooling time, respectively, following 1 year's operation at 10 Mw. The residual activity in the heat exchanger was based on the assumption that 100% of the isotopes which might plate out on INOR-8 (ref 7) plated out in the heat exchanger. Assuming that both parents and daughters remained in the - heat exchanger and that the gammas could be represented as ~l-=Mev photons origi- nating from a point source, the dose rate at a distance of 10 ft was 2 x 10% and 1.4 x 10* r/hr after 10 and 30 days cooling time, respectively, following 1 year's operation at 10 Mw. The estimated beta- and gamma~energy emission rates in the heat exchanger following a year's operation at 10 Mw were, respectively, 18.4 e"7'8lt + 9.56 o 100t 5.19 e'0'0u38t, 10° Q Btu/hr , B 10.0t -0.896t . 20.0 e~ + 9.01 e e'o'00692t, Q lO3 Btu/hr , y 13.9 n where t = cooling time in hours: O St < 24 hr. It was assumed that the heat loss from the primary heat exchanger through the in- sulation is linear with the average temperature of the metal in the heat exchanger, that there is 5-kw loss at 1200°F (no coolant circulation), and that all the energy emitted by the fission products is absorbed by the metal., Then, for 2000 1lb of metal in the heat exchanger, the peak temperature was ~1350°F and occurred about L hr following reactor shutdown and stoppage of coolant flow. . ?1 3.4 REACTOR-CELL SHIFLDING 3.4.1 Top Shield The dose rates to be expected above the reactor cell have been re-estimated, using the new values of the N'® activity in the fuel salt (see sec 3.2) and the revised layout of the fuel-salt system. Basic changes in tne layocut pertinent to shielding considerations consisted of the removal of the pump from directly over the reactor vessel and the removal of the INOR-8 shield located in the upper header. The resulting 2-ft-diam hole in the top of the reactor shield is to be used for removing the graphite sample and control equipment (the effect of this equipment on shielding requirements was neglected in the shielding calculations). Both a solid shield and the existence of a 1/2-in. crack between the upper snield beams has been considered. Dose-rate estimates above the crack were based on tihe uncollided photon flux, whereas the sclid shield estimates included bulld- up. Self-shielding by the source and source container were included in all cases. In addition to the H'® activity, other radiation sources considered were fission product activity, tnermal-neutron capture gammas, prompt-fission gammas, and prompt- fission neutrons. Table 3.3 gives the calculated dose rates directly over thie reactor vessel for the present shield thickness (upper shield beams, 3.5 £t of ordinary concrete; lower shield beams, 3.5 £t of barytes concrete). Results for the balance of the shield have been reported,® and, except for the region above the reactor vessel, the general radiation level is less than 2.5 rw/hr for the solid sihield. It is presently planned to insert a filler material in the cracks between the upper shield beams. Should the measured dose rate during reactor operation still be excessive, provision has been made for placing additional shield material as needed. Table 3.3. Dose Rates Above the Reactor Top Saield (10-Mw Reactor Power; Shield Consists of 3.5 It of Barytes and 3.5 £t of Ordinary Concrete ) Dose Ratcs (mr/hr) With Solid Shield With 1/2-in. Crack Between Lource Upper Snield Beams Gamma Rays from: Core 7.8 51 Iron capture 5 14 o 0.7 k.9 Concrete capture 1.3 I ,6 Neutrons from: Core 1.8 L4600 92 3.4.2 $ide Shield Estimates of the dose rates outside the side shield of the reactor were made considering reactor operation at 10 Mw for one year and various times after shut- down, assuming that the fuel system was not drained. In obtaining these estimates, the fuel-salt-system components were considered individually. An IBM 7090 shielding- calculation code prepared by E. D. Arnold® was used for the calculations. Activity due to the decay of fission products was obtained from the IBM Internuc code ,+° Other sources®® consisted of the saturated fission-product activity and capture gammas in the reactor shield. The calculations included the effect of 2 in. of steel (containment-vessel wall) and the self-shielding associated with the source medium and source container. Typical results obtained are given in Table 3.4. Table 3.4. Dose Rates Outside the Side Shield Due to Individual Fuel-Salt-System Components (20-ft Attenuation Distance Between the Source and Dose Point) Dose Rate (mr/hr) Component 7 ££-10 Mw® L ££-105 sec cooling® - Thermal shield 40 0 Heat exchanger, side view 3 1.5 Heat exchanger, end view 0.k 0.3 Pipe (6 ft£ 5 in.), side view 0.6 0.5 Pump bowl, side view 0.5 Q.4 aT ft of ordinary concrete, 10-Mw reactor power, 1 year's operation. bu 't of ordinary concrete, 10° sec cooling time following 1 year of operation at 10 liw. REFERENCES 1. W. L. Breazeale, MSRE Design Data Sheets, MSR-61-100 (Aug. 15, 1961). » 2, T. B. Fowler and M. L. Tobias, personal commnication. 3. M. L, Tobias, D. R. Vondy, and M. P. Lietzke, Nightmare - An IBM 7090 Code . for the Calculation of Gamma Heating in Cylindrical Geometry, ORNL report (in preparation). 4. D. W, Vroom, Preliminary MSRE Gamma Ray Source and Biological Shielding Survey, ORNL CF-61-4-97 (Apr. 20, 1961). 5« Ca. W. Nestor, personal communication, May 1961. 6. R. B. Stevenson, Radiation Source Strengths in the Expansion Chamber and Off-Gas System of the ART, ORNL CF-57-7=-17, (Nov. 18, 1957). 7. I. Spiewak, Proposed MSR Small Pump Loop Program for FY 1962, MSR-61-12, (Feb. 10, 1961}, 8. E. S. Bettis, MSRE Component Design Report, Section VIII, MSR-61-67, (June 20, 19617). 9. E. D. Arnold, Shielding Design Calculation Code, ORNL-~304l, to be published. 10, D. M. Shapiro, EAPPR, Internuclear Co., TM=-DMS-58-1, (Dec. 31, 1958). PART Il. MATERIALS STUDIES 4. METALLURGY . 4.1 DYNAMIC-CORROSION STUDIES 4.1.1 Examination of INOR-8 Forced-Convection Loops Examinations of eilght additional INOR-8 forced-convection loops were completed during this period, concluding the investigation of long-term corrosion effects in fluoride systems. 1In all, 24 loops (9 of Inconel and 15 of INOR-8) were operated under this program. Table L.l gives a summary of the operating times of these tests. Table L.1. Summary of Forced-Convection Loop Operation Length of Test TNOR-8 Inconel (hour ) No. of Total Hours No. of Total Hours Tests Operated Tests Operated 3,000 to 5,000 2 6, 440 . 6,000 to 10,000 6 52, 680 5 4L, 760 11,000 to 15,000 L 54,370 2 28,190 16,000 to 20,000 5 99,950 Total hours accumulated 207, 000 79,390 A summary of the operating conditions and test findings for the 24 loops is presented in Table 4.2. The last eight loops in this table represent those examined during this period. With the exception of loops 9354-3 and MSRP-16, corrosive attack was 1imited to the hottest segments of the loops and was in the form of a pitted surface layer. Metallographic examinations of four 20,000-hr loops (MSRP 6, 7, 10, 11) showed corrosion depths generally in the range of 1/2 to 2 mils for systems based on LiF-BeF, and NaF-BeF,. Figure 4.1 shows the general appearance of attack in these loops. In loop MSRP-6, one of those which circulated a salt . similar to the proposed MSRE fuel mixture but without ZrF), the depth of attack was greater, reaching a maximum of 2 mils, as shown in Fig. 4.2. . 93 Table 4.2, Operating Conditions of Forced-Convection Loops and Results of Metallographic Examinations of Loop Materials Maximim Duration Fluid-Metal Flow Loop of Test Interface AT Reynolds Rate Number Material (hr) Salt Mixture> Temp. {°F) (°F) Number (gzpm) Results of Metallographic Exemination 9075-1 Inconel 8,801 122 1250 100 5000 2.0 Intergranular penetrations to 1-1/2 milsP 93k4-1 Inconel 8,892 123 1300 200 3250 2.0 Heavy intergranular voids to 38 mils® 93442 TInconel 8,735 12 1200 100 8200 2.5 Intergranular voids to 8 mils® 9354-1 INOR-8 14,563 126 1300 200 2000 2.5 Heavy surface roughening and pitting to 1-1/2 milsd 9354-3 INOR-8 19,942 Bl 1200 100 3000 2.0 No attack, slight trace of metallic deposit in cooler coil 935L-4 INOR-8 15,140 130 1300 200 3000 2.5 No attack®s® 9354-5 INOR-8 14,503 130 1300 200 3000 2.5 No attackS 3T77=-1 Inconel 3,390 126 1300 200 1600 2.0 Intergranular voids to T m:i.ls-b 9377-2 Inconel 3,046 130 1300 200 3000 2,0 Intergranular voids to 8 mils 9377-3 Inconel 8, 76k 131 1300 200 3Loo 2.0 Intergranular and general voids to 1k mils® 9377-4+ Inconel 9,57L 130 1300 200 2600 1.75 Intergranular and general voids to 1lb mils 9377-5 Inconel 15,038 13k 1300 200 2300 1.8 Intergranular voids to 24 mils 9377-6 Inconel 13,155 133 1300 200 3100 1.8 Intergranular voids to 13 mils? MSRP-6 INOR-8 20,000 134 1300 200 2300 1.5 Pitted surface layer to 2 mils MSRP-7 INOR-8 20,000 133 1300 200 3100 1.8 Pitted surface layer to 1 mil MSRP-8 INOR-8 9,633 124 1300 200 000 2.0 No attack MSRP-9 INOR-8 9,687 13L 1300 200 2300 1.8 No attackd MSRP-10 INOR-8 20,000 135 1300 200 3400 2.0 Pitted surface layer to 1/2 mil MSRP-11 INOR-8 20,000 123 1300 200 3200 2.0 Pitted surface layer to 1 mil MSRP-12 INOR-8 14,498 13k 1300 200 2300 1.8 No attack MSRP-13 INOR-8 8,085 136 1300 200 3900 2.0 THeavy surface roughening and pitting MSRP-1L4 INOR-8 9,800 Bult 1+0.5 U 1300 200 Pitted surface layer to 1/2 mil MSRP-15 INOR-8 10,200 Bult 1%4+0.5 U 1400 200 Pitted surface layer to 2/3 mil MSRP-16 INOR-8 6,500 Bult 14+0.5 U 1500 200 Moderate subsurface void formation to I mils 85alt Compositions: bMSRP-Quar. Prog. Rep. Oct. 31, 1958, ORNL-2626, p 5L4-55. éfi fiii-iig-gzégfléi-gglgshioigli * °MSRP Quar. Prog, Rep. Oct, 31, 1959, ORNL-2890, p 35-i2. 122 NoF-Zrf),-UP), 57-h2-1 mole % “MSRP Quar, Prog. Rep. Jan. 31 and April 30, 1960, 3 NaF-BeFo-UF),, 53-46-1 mole % e e 124 NoF-BeFo-ThF),, 58-35-T moled% -2913, p 36-3%. 126 LiF-BeF,-UF), 53-46-1 mole % e 130 LiF_Bng_UFJ, EQ_BE'i oL ? MSRP Quar. Prog. Rep. July 31, 1960, ORNL-3014, p 55-58. 131 LiF-BeF,-UF,, 60-36-4 mole % £ 133 LiF_Bng_Thfiu, T1216-13 mole % MSRP Proz. Rep. Feb. 28, 1961, ORNL-3122, p T7-81. 13k LiF-~BeF,-UF),-ThF),, 62-36.5-0.5-1 mole % 135 NaF-BeFo-UF),-ThF), 53-45.5-0.5-1 mole % 136 LiF-BeF2-UF), T0-10-20 mole % Bult 1%0.5 U LiF-BeFp-ThF),~UF),, 67-18.5-14-0.5 mole % 1Z0) 95 UNCLASSIFIED T-20553 Fig. 4.1. Appearance of Specimen Removed from Point of Maximum Salt-Metal Interface Temperature Region (1300°F) of INOR-8 Forced-Convection Loop MSRP-10. The loop was operated for 20,000 hr. UNCLASSIFIED T-20038 Fig. 4.2. Appearance of Specimen Removed from Point of Maximum Salt-Metal Interface Temperature Region (1300°F) of INOR-8 Forced-Convection Loop MSRP-6. The loop was operated for 20,000 hr. 96 Examination of loop 9354-3, which circulated a coolant salt of the system NaF-LiF-BeF, for 19,942 hr, revealed no evidence of attack in either the hot or cold areas of the loop. However, macroscopic examination of specimens removed from the end of the cooler coil did reveal the presence of a small amount of metallic deposit. Analyses of the deposit have not been completed. The last three loops in Table 4.2 afford an evaluation of temperature effects on corrosion by the system based on LiF-BeF,. Two of the loops, which operated at maximum temperatures of 1300CF and ll&OOgF, respectively, sustained approximately the same depths of attack (1/2 to 2/3 mil) after 10,000 hr. Loop MSRP-16, which operated at 1500°F, exhibited a considerably higher rate of attack (~l4 mils) in only 6500 hr. Corrosion at 1500°F was manifested by sub- surface void formation, and the surface film that is characteristic of the corrosion process at lower temperatures was conspicuously absent, as shown in Fig. b4.3. A final report covering the results of all forced-convection loops operated under the MSRE program is being written. UNCLASSIFIED | @ T-20428 Gz o £ Fig. 4.3. Appearance of Specimen Removed from Point of Maximum Salt-Metal Interface Temperature Region (1500°F) of INOR-8 Forced-Convection Loop MSRP-16. The loop was operated for 20,000 hr. 4.1.2 Molybdenum-Graphite Compatibility Tests As previously reported,]‘ tests have been in progress to investigate the compatibility of molybdenum in contact with graphite, INOR-8, and the proposed MSRE fuel mixture (LiF-BeFy-ZrF)-ThFy-UF), 70-23-5-1-1 mole %). The tests were carried out in INOR-8 thermal convection loops which incorporated test sections containing type RO025 graphite, molybdenum strips, and INOR-8.2 The graphite, in the form of cylinders 6 in. long, was encased in a 1-1/4-in. sched-80 INOR-8 pipe section and mounted in the hot leg of the thermal convection loops. The first test (loop 1250) contained two graphite cylinders, and the second (loop 1251) contained four cylinders. The salt-to-graphite volume ratios of the two 97 loops were approximately 2:1 and 1:1, respectively. Table 4.3 summarizes the operating conditions of these tests. Table 4.3. Operating Conditions for Graphite-Molybdenum Compatibility Tests Molybdenum: Universal-Cyclops arc-cast Mo + 0.5 T alloy used in as-rolled condition. Graphite: ©National Carbon Grade R-0025. Fluoride mixture used: Proposed MSRE fuel — LiF-BeFy-ZrF) -ThF) -UF), (70~23-5-1-1 mole %). Condition Loop 1250 Loop 1251 Max. salt — metal interface temp. (°F) 1350 1300 Max. salt — graphite interface temp. (°F) 1360 £ 5 Max. salt temp. (OF) 1300 1300 Min. salt temp. (OF) 1140 1140 Loop AT (°F) 210 160 Operating time (hr) 5585 5545 Salt-to-graphite volume ratio 2:1 1:1 The tests were terminated after 5585 and 5545 hr, respectively, and results of examination have been obtained for the longer-duration loop. A "scratch" test of the molybdenum strips contained in loop 1250 showed that the surface of the strip, where in direct contact with graphite, was significantly harder than the surfaces in contact with either the salt or the INOR-8. Metallographic examination of the strips showed a layer ~0.1l-mil thick on the surface exposed to the graphite; otherwise, no apparent changes in microstructure were found as a result of the exposure conditions, as shown in Figs. 4.4 and 4.5, Tensile tests of the molybdenum strips in both the as-received and as- tested condition showed gqualitatively that the ultimate strength and per cent reduction in area of the molybdenum were reduced as a result of exposure. However, there was no apparent difference in the hardness of the as-tested and as-received molybdenum. Table 4.4 gives the results of the tensile tests and hardness measurements. (Tensile results must be regarded as qualitative because of the necessarily small specimen size). Table 4.4, Mechanical Properties of Molybdenum (Qualitative Only) Ultimate Reduction Hardness — Strength in Area Four Determinations (psi) (%) (DPH) As Received 102, 600 50 267 105,200 49,7 As Tested 79, 650 3.5 273 JUNCLASSIFIED T-203014 Fig. 4.4. As-Received Appearance of Molybdenum Strips Contained in INOR-8 Thermal-Convection Loop 1250, Etchant: 1 part NH,OH, 1 part H,0,. UNCLASSIFIED: T-20302 Fig. 4.5. As-Tested Appearance of Molybdenum Strips Contained in INOR-8 Thermal-Convection Loop 1250. Etchant: 1 part NH,OH, 1 part H,0,. Results of chemical analyses made on both the as-received and as-tested molybdenum are shown in Table 4.5. An increase in the chromium concentration of the molybdenum during test is clearly indicated. 99 Table 4.5. Results of Chemical Analyses of Molybdenum Strips Constituent (ppm) Sample Ni Cr Fe As-received 50 210 L20 As-tested 250 1600 860 Metallographic examination of INOR-8 specimens removed from the loop reveal- ed relatively heavy surface roughening and pitting; however, examination of the as-received tubing revealed a similar surface condition. Examination of one of the INOR-& spacers, which held the molybdenum in contact with the graphite, revealed attack in the form of surface roughening and pitting to a maximum depth of 1 mil along the inner surface, which was exposed to the salt. The outer surfaces of the spacers, which were in intimate contact with both the graphite and molybdenum, were found to have been carburized to a depth of >10 mils. Considering the difference in thermal expansion between INOR-8 and graphite, it is likely that there was a relatively large contact pressure between the graphite and INOR-8 along the carburized areas. Chemical analysis made on the as-received and as-tested salt showed the chromium content to have increased from an initial value of 550 ppm to a final value of 960; no other changes were apparent in the salt analyses. 4.1.3 Compatibility of INOR-8 — 2% Nb and Molten Fluorides As a consequence of recent developments to improve the weldability of INOR-8,3 corrosion tests have been scheduled to evaluate the resistance to fluoride attack of INOR-8 that has been modified by the addition of 2% Nb, In addition to corrosion data on INOR-8 — 2% Nb, these loop tests will also supply data on the corrosion properties of various types of weld junctions listed below. 1. TINOR-8 welded with INOR-8 weld rod. 2. INOR-8 welded with INOR-8 — 2% Nb weld rod. 3. INOR-8 —~ 2% Nb welded with INOR-8 — 2% Nb weld rod. Corrosion tests will be conducted in two INOR-8 thermal-convection loops, each loop to contain two tubular inserts in the hot leg. One of the inserts will be fabricated from INOR-8 modified with 2% Nb, while the other will be of standard INOR-8. The inserts will be separated from each other, and the loop by spacers. They will be protected from the atmosphere by an outer jacket welded to the loop. The spacers will also function as the weld-test specimens. Tentative operating condltions for both loops are as follows: Salt composition LiF-BeF,-ZrF) -UF), - ThF), (70-23-5-1-1 mole %) Maximum salt temp. 1300°CF AT 170°F Operating time One year h,1.4 Fluoride-Salt Contamination Studies Corrosion studies of fused fluoride mixtures have conventionally been conducted under closely controlled test conditions to maintain a high degree 100 of salt purity. However, maintaining such stringent control becomes problemati- cal in the case of a large-scale reactor experiment, and there is a probability that small amounts of certain impurities, which act as strong oxidants, will be encountered throughout normal reactor operation. Examples of these impurities include moisture, dry air, oil, and combinations thereof. Accordingly, a test program was started to systematically investigate the effects of such contami- nants on the corrosion behavior of fused fluoride mixtures and to define the limits of the various contaminants which can be safely allowed to enter the salt. The first series of tests was designed to establish the effect of moisture on the corrosion of INOR-8 by fused-salt mixtures. It is known that water immediately reacts with the salt to form HF and oxides of the melt through hydrolysis; thus, this phase of the program is being divided into two series of tests. In the first series of tests, salt mixtures containing controlled amounts of HF will be circulated to study both the mechanism and rates of attack as a function of HF concentration. In order to translate these results into an understanding of the corrosive effect of H,0, a second series of tests utilizing small static pots will be operated to determine the rate of HF buildup in salts exposed to wet helium. The initial thermal-convection test under this program was terminated prematurely after 255 hr as a result of a sudden flow restriction. The test circulated an NaF-ZrF, (50-50-mole %) salt mixture to which HF had been added immediately before operation. The salt mixture was a standard production batch obtained from the Reactor Chemistry Division. Before beilng placed in the loop, the mixture was repurified through the use of pure zirconium shavings and a hydrogen purge. Radicactive FeF, was used to monitor the progress of the repurification. Chemical analysis of the salt before and after repurification showed the nickel and iron content to be reduced, whereas the chromium content increased. After the repurification process, the loop was filled, and the HF was introduced through a dip leg attached to a salt pot at the top of the loop. The HF-bearing salt was clirculated through the loop under approximately a 550F AT. Determination of the amount of HF introduced to the salt was rendered impossible because of several leaks which occurred in the gas lines. However, chemical analysis of the salt circulated will provide some measure of the amount of HF absorbed. Metallographic and chemical analyses are in progress. 4.2 INOR-8 DEVELOPMENT 4.2.1 Structural Stability of Nickel-Based, 18% Molybdenum Alloys Work is continuing in order to determine the solubility limit of chromium plus iron in nickel-base alloys containing 18% Mo over the temperature range 900 to 2000°F. The purpose is to confirm that INOR-8 will not be subject to the precipitation of intermetallic compounds when the maximum amount of alloy- ing additions allowed by the specification are present, that is, 18% Mo — 8% Cr — 5% Fe. It is of further interest to determine the location of the solid-solution phase boundary known to exist near the upper limit of the INOR-8 compositional specification in order to establish what deviations from specifications might be permitted. 101 The previous reportl'L summarized the results obtained from a study of the 900-t0~-2000°F microstructures of a series of 18% Mo alloys encompassing the maximum compositional limit of INOR-8. The interpretation of these results wvas complicated by the presence of a fine grain-boundary precipitate in the alloys over the range 1100 to 1650°F. This precipitate, thought to be a carbide, redissolved at 1800°F. Two alloys, both containing iron and chromium contents in the range 8 to 10%, retained a precipitate at lBOOOF, indicating that this phase was an intermetallic compound. Based on these results, an 1800°F phase boundary was proposed, as shown in Fig. L.6. UNCLASSIFIED ORNL-~LR-DWG 63277 14 r VT—154|| Ll 12 I 1 \ PREVIOUSLY \ NEW BOUNDARY 10 PROPOSED //“ BOUNDARY —1~ \ \ \ . VT—-156 \ < 8 [ Y@ VT-157 — \ = \ ~ \ = o '\\ x g i \-\ ~ SINGLE - PHASE FIELD VT-158 . TWO-PHASE FIELD . | ~ VT-159 Tw VT-160 4 - > ¢ ~ ~ ~ ~ Mo VT-162 2 . 0 0 2 4 6 8 10 12 14 16 CHROMIUM (wt %) Fig. 4.6. 1800°F Phase Boundary for 18% Mo—Balance Ni Alloys Containing Varying Chromium and Iron Contents. Recent effort has been concerned with confirming the proposed phase boundary (alloys VI-154 through VI-160, Fig. L4.6). Prior to heat treating samples of these alloys within the range 900 to 2000°F, a decarburization heat treatment was carried out at 22000F in dry hydrogen to eliminate as much carbon impurity as possible from the materials. Metallographic examination of samples receiving 100-hr heat treatments at 1800°F showed in the construction 102 of a new phase boundary, as shown in Fig. L4.6. Again it is thought that carbon contamination presents a problem in so far as all the alloys generally showed a small amount of fine grain-boundary precipitate in the range 1100 to 1650°F. This precipitate dissolved at 1800CF in those alloys located in the single- phase region of the diagram shown in Fig. L4.6. Particles cobserved at 18000F in those alloys located in the two-phase region are thus believed to be an intermetallic compound. According to this interpretation, the decarburization treatment given the alloys was insufficient in lowering the carbon content below the solubility limit at temperatures up to 165CCF even though subsequent carbon analyses for the alloys were in the range of 0.005 to 0.01%. 4.2.2 Temperature Range of Melting for INOR-8 Apparatus has been set up to determine the temperature range of melting for INOR-8. The purpose is twofold: (1) +to establish the melting range of the alloy more accurately, and (2) to compare the melting range of the various heats of the alloy with their respective nil ductility temperature as reported by Rensselaer Polytechnic Institute. The latter purpose is related to the incidence of base-metal cracking found in certain heats of INOR-8 when welded under restraint. The particular heats involved were found (by means of the TPI Hot-Ductility Apparatus) to exhibit inferior ductility behavior at high temperatures; therefore, a comparison of the nil-ductility temperature with the solidus temperature might establish whether the presence of liquid was the cause of the poor elevated-temperature ductility. Differences in the solidus temperatures between various heats of the alloy could be related to varying amounts of trace impurities in the material. The data obtained to date are presented in Table 4.6. Although no definite conclusions are yet possible, it does appear that solidus tempera- tures are somewhat higher than nil-ductility temperatures, presenting some doubt that liquid phases are the cause of the elevated-temperature brittleness reported. Teble 4.6. Comparison of Nil-Ductility Temperature with Melting-Temperature Range for Various Heats of INOR-8 Melting-Temperature Range Nil-Ductility Liquidus Solidus Heat Temperature Temperature Temperature Nunber Vendor (°F) (o) (°F) Y-8488 INCO 2300 2L62-2498 NI-5055 Haynes 22hs5 2hob-2Lhl SP-19 Haynes 2L00 2L62-2498 SP-16 Haynes 2200 SP-20 Haynes 2350 SP-26 Haynes 2557—2561, SP-37 Haynes 2453—2471. DI-0L0OO Haymes 2h53-24 71 DI-OLO1l KJ Haynes 2453-2L71. M-1 Westinghouse 2250 M-1566 Westinghouse 2400 M-1666 Westinghouse 2552 103 4.2.3 Specific Heat of INOR-& The specific heat of annealed INOR-8 has been determined by direct calorimetric measurements at the University of Tennessee. The value was observed to increase uniformly from 0.409 joule/g, ©C at 60°C to 0.483 joule/g, ©C at 5L0OPC. At 550°C, an anomalous rise of about 20% in the specific heat was observed, reaching a maximum value of 0.586 joule/g, OC at 610°C. Above this temperature, the specific heat decreased to 0.578 joule/g, OC at 700°C. The phenomenon has been attributed to an ordering reaction. The values for the specific heat are presznted in Table 4.7, Table 4.7. Specific Heat of TINOR-8 Temperature c_(3oule/g, ©C) Temperature c (joule/g, ©C) (°c) o (°c) P 60 0.4090 LLo 0.4730 &0 0.Lk152 LG0 0.47h9 100 0.4207 480 0.4766 120 0.4257 500 0.4781 140 0.4303 520 0.4794 160 0.4340 530 0.4805 160 0.4370 540 0.4833 200 0.4403 550 0.hkoz2 220 0.4439 560 0.5060 2Lo 0.L475 570 0.5226 260 0.4510 580 0.54202 280 0.4539 590 0.5637 300 0.4566 600 0.5808 320 0.4592 610 0.5858 340 0.4616 620 0.5848 360 0.46k41 640 0.5825 380 0.4666 660 0.5808 LOO 0.4688 680 0.5792 L20 0.4710 700 0.5778 4,3 TOTAL HEMISPHERICAL EMITTANCE OF INOR-O The temperature dependence of the total hemispherical emittance of platinum, matte and bright-finished INOR-8, and several oxidized INOR-8 specimens was measured between 100 and 700°C. The method employed an instrumented thin-strip specimen heated electrically in a black-body vacuum chamber which was held at a constant known temperatures A steady state, a heat balance on the specimen yields the following: VI E = Ac(TSh - T, h) ’ 104 where E = total hemispherical emittance (t.h.e.), A = surface area of the specimen between the voltage taps, g = Stefan-Boltzmann constant, T_= average specimen temperature, CK, S Te= average black-body vacuum-chamber temperature, CK, VI = electrical power per unit length of the specimen. Heat gains or losses considered in obtalning this equation are: (1) the electrical power dissipated by the specimen, (2) the radiant energy incident upon the specimen, and (3) the radiant energy emitted by the specimen. Heat gains or losses which are neglected as being cumulatively less than 4% are: (1) radiant energy emitted by the specimen that is reflected by the chamber walls back to the specimen surfaces, (2) the heat loss by atmospheric convection and conduction, (3) the heat loss by conduction through the specimen to the input electrodes, and (4) the heat loss by conduction through the thermocouple leads. The t.h.e. of platinum measured by this method serves as a guide to the validity of the measurements on the INOR-8 specimens. The platinum data as seen in Table 4.8 and Fig. 4.7 compare quite favorably with other literature values. The authors' data for platinum are approximately 16% higher than data obtained by a similar method (run II, Table 4.8) at 500°C. A large part of this percent- age difference is because the platinum tested did not have a bright polish. Table 4.8. Total Hemispherical Emittance at Selected Temperatures Temperature Specimen Run No. Condition Prior to Run 100°c 200°c 300°c Loo°c 500°C 600°Cc 700°C B0O°C INOR-8 No. 1 1 Matte finish 0.220 ©.210 0.214% 0.226 0.242 0.260 0.275 0.2852 2 1 hr at 8309 0.412 0.451 0.502 0.561 0.625 3 0.5 hr at T30°C plus 0.%417 0.468 0.528 0.579 0.607 five l-min cycles to 730°C Iy 5 hr at 730vC G.436 0.496 0.570 0.622 0.6L5 5 5 hr at 730°C 0.465 0.478 0.513 0.567 0.605 INOR-8 No. 2 6 Bright Finish 0.2012 0.202 0.212 0.226 0.239 0.239 0.226 7 25 l-min cycles to 730°C 0.311 0.34% 0.398 o0.L46L 0,503 8 50 cycles to T30°C 0.31% 0.3k 0.4L18 0.497 0.525 9 75 cycles to T30°C 0.293 0.315 0.34% 0.389 0.433 0.509 10 100 cycles to T730°C 0.286% 0.318 0.353 0.394 0.his5 0.L88 11 1 hr at 830°¢ 0.349 0.39C 0.434 0.481 0.528 INOR-8 No. 3 12 Bright finish 0.194% 0.208 0.221 0.232 0.2L2 0.24s5 13 1 hr at 1000°C 0.502 0.523 0,630 0.704 0.757% Platinum 1k Bright finish 0.127 0.109 0.106 0.110 0.119 0.129 ©0.1k2 (§.156 ks 0.071 0.087 0,102 0.116 0.130 0.1k4 II¢ 0.062 0.070 0.083 0.103 ©0.122 (©.lk2 rrrd 0.061 0.070 o0.080 0.088 0.103 0©.l22 aExtrapolated. Phavisson and Weeks, J. Opt. Soc. Am. 8, 585 (1924). °A. H. Sully et al., Brit. J. Appl. Phys. 3, 97 (1952) 9. D. Foote, N.E.S. Bull. 11, 607 (191k) The data on three of the INOR-8 specimens reflect the importance of surface finish. Runs 6 and 12 were on different specimens having a bright polish, and the results are remarkably close. Run 1 was on a matte-finish INOR-8 and is 7.1% higher than runs 6 and 12 at 500°C. 105 UNCLASSIFIED ORNL-LR-DWG 63278 0.8 TOTAL HEMISPHERICAL EMITTANCE p—""9 1 O A e 0.2 12 6 b - 14.F§-‘““*flb—— p 0.1 = & —— O C 1C0 200 300 400 500 600 700 800 TEMPERATURE (°C) Fig. 4.7. Total Hemispherical Emittance vs Temperature for Platinum, Matte and Bright-Finished INOR-8, and Several Oxidized INOR-8 Specimens. (See Table 4.8 for meaning of numbers.) Following the initial runs, the matte-finished and bright-polished specimens were given the treatments listed in Table 4.8. These treatments resulted in an increase in the t.h.e. which was insensitive to thermal cycling effects. The increase in t.h.e. was greatest for the 1000°C treatment and least for the T730°C treatments. Based on these measurements, i1f INOR-8 is given an oxidizing heat treatment at 1000°C, its t.h.e. at 600°C will be 0.76 as compared to an unoxidized surface value of 0.24; and at 5000C values of 0.705 and 0.225, respectively, are to be expected. Oxidation at temperatures below 1000°C will not yield as high a t.h.e. 106 4.4 WELDING AND BRAZING STUDIES 4.4,1 Welding of INOR-8 A program is under way to investigate and circumvent weld-cracking which has been observed in some heats of INOR-8 material. Cracking and microfissuring have been reported in test welds made with certain base-metal and filler-metal heats.” Since the melting practice used by the material supplier in pouring the initlal ingots appeared to influence the cracking behavior markedly, the major effort has been directed at investigation of this wvariable. Since different vendors use their own melting practice or practices, quantities of l-in.-thick material were obtained from three vendors. 1In some cases, the vendor furnished individual heats of material melted according to different proprietary practices. Test welds were then made with these materials under conditions of high restraint and were then carefully examined for cracking by visual, dye-penetrant, and metallographic inspection. Mechanical tests on these welds were conducted at room and elevated tempera- tures, and, in addition, hot-ductility testing6 is being performed and evaluated on a subcontract at Rensselaer Polytechnic Institute. Complete chemical analyses on all heats used 1n the study are being obtained in an attempt to correlate composition and melting practice with performance. A discussion of the different phases of the study, together with the results to date, are presented in the following paragraphs. 44,2 Metallographic Examination and Bend Tests on Welds An extensive metallographic examination was performed on 14 weld-test plates fabricated from 9 different heats of INOR-8 base metal from three different vendors. When available, the filler wire used was of the same heat of INOR-8 as the base plate. However, in several cases, filler wire of the desired heat was not available, and, another heat was used in the interest of expediency. A compilation of the results of the metallographic examination of typical test plates is presented in Table 4.9. The results of side-bend tests made on these plates are also shown for comparative purposes. It will be noted that those welds containing cracks and porosity were fabricated from INOR-8 produced with certain air-melting procedures. Heat-affected-zone cracking and porosity typical of these welds are shown in Fig. 4.8 (a photomicrograph of the fusion-line area of weld No. SP-37-Wl). A sound microstructure, typical of welds of the improved heats of material, is that shown in Fig. 4.9, the fusion-line area of a weld in heat Y-8488. 4.4.3 Transverse Tensile and Creep Tests Transverse tensile and creep tests were performed on welds of seven of the heats included in the program. The results of the room- and elevated-temperature tensile tests are presented in Table 4.10, while creep-test results performed at 1300°F at a stress of 27,500 psi are presented in Table 4.1l. From the testing conducted thus far, it appears that the vacuum-melted heat, SP-OLO1, and the air-melted heat, Y-8488, exhibit the best over-all weldability. The air-melted heat, SP-5055, did not exhibit any cracking or porosity, but the creep and tensile properties were generally somewhat lower. Table 4.9. Results of Metallographic Examination and Side-Bend Tests on INOR-8 Weldability Samples Metallozraphic Results S5ide Bend Results Heat-Affected Heat-Affected Weld Base-Plate Base-Plate Filler-Metal Zone of Weld Zone of Weld Plate No. Heat No. Melting Practice Heat No. Base Metal Metal Base Plate Metal C = Cracked; P = Porous; 0.K. = No cracking or Porosity SP-19-W1 SP=-19 INOR-8 air-melting Haynes SP-19 C C C C procedure (vendor A) SP-20-W1 SP-20 INOR-8 air-melting Haynes SP-19 C C C C procedure {vendor A) SP-23-W1 SP-23 TNOR-8 air-melting Haynes SP-19 P,C C C C procedure (vendor A) SP-26-W1 SP-26 INOR-8 air-melting INCO Y-8LE8 P,C 0.K. 0.K. 0.K. procedure (vendor A) SP-37-Wl SP-37 INOR-8 air-melting Haynes SP-37 P,C C C C procedure (vendor A) SP-5055~W1 SP-5055 Hastelloy-W air- Haynes SP-0401 0.K 0.K. 0.K 0.K melting procedure (vendor A) SP-0401-W2 SP-0LO1 Vacuum-melting Haynes SP-0401 0.K. 0.K. 0.K 0.K. procedure (vendor A) W-16T71-W1 W-1671 INOR-8 ajir-melting Haynes SP-19 0.K. 0.X. C C procedure (vendor B) Y-8488-W1 Y-8488 INOR-8 air-melting INCO Y-8488 0.K. 0.X. 0.K. 0.K. procedure (vendor C) LOL 108 Fig. 4.8. Heat-Affected-Zone Porosity and Cracking in an INOR-8 Butt Weld (No. SP-37-W1) Made Under High Restraint. Etchant: CrO;-H,0-HCI. Note that the cracking and porosity seen in the metallographic examination are generally reflected in the transverse tensile tests by markedly reduced ductilities (particularly at 1300CF) and by slightly lower tensile strengths. The good tensile properties of the weld in SP-26 material are attributed to the fact that the filler wire (heat Y-8488, see Table 4.9) possessed excellent weldability. In the creep tests, the weld porosity and microfissuring were reflected in short rupture times. In most cases, lowered ductilities were also observed. Note however that the weld in SP-26 material responded rather poorly in the creep test, indicating that this test is really a more selective criterion for true joint integrity. L.4.4 Welding of INOR-8 to Stainless Steel and Inconel The qualification of tests for tungsten-arc welding of INOR-8 to stainless steel and Inconel has been completed. Procedure specifications for INOR-8-to- stainless steel (PS-35) and INOR-8-to-Inconel (PS-33) have been prepared. As a result of the findings of the weldability study being performed on INOR-8, the existing specification for the welding of INOR-8 alloy pipe, plate, and fittings is being revised. The revisions will incorporate into the welding procedure the requirement that the interpass temperature does not exceed 200°F. Experience with this type of material indicated that interpass-temperature “UNCLASSIFIED Y-42433 7 S Fig. 4.9. Fusion-Line Area of Weld No. Y-8488-#1. No cracks or porosity is evident. Etchant: CrO,- H,0-HCI. 2 control was desirable to ensure high joint integrity when welding this material in thick sections under high restraint. k.4.5 Remote Brazing If removal of any MSRE primary components is required, provisions must be made for remotely disconnecting and subsequently rejoining the necessary pipe- work. For this reason, a study is being conducted in cooperation with the Reactor Division Remote Maintenance Group to develop procedures for constructing these remotely brazed joints. The design of such a joint is influenced by both brazing and remote- maintenance requirements. Maintenance procedures are simplified by use of relatively loose fits and a minimum of remote machining, but optimum brazing procedures require relatively tight fits and very clean surfaces. Accordingly, a tentative joint design was adopted (see Figs. 2.26 and 2.27, Chap. 2). The major features of this joint are the provisions for brazing alloy preplacement and a radial clearance of 0.0025 to 0.0035 in. (0.005 to 0.007 in. in diameter). This clearance is considered optimum in terms of ease of fitup and reliable brazing-alloy flowability. Preliminary work with the brazing of I]\IOR-B7 led to the selection of a . brazing alloy composed of 82 Au — 18 Ni (wt %). The alloy is ductile and very 110 Table 4.10. TINOR-8 Transverse Tensile Tests on Butt Welds in l-in. Plate % Test Tensile Yield Elongation Weld- Temp. Strength Strength in Location Plate No. (°F) (psi) (psi) 1-1/2 in. of Failure SP-23-W1 Room 105, 500 67,800 28 Weld metal 1300 59, 700 43,000 8 Heat-affected zone SP-26-W1 Room 116,300 69, 800 L7 Weld metal 1300 70, 400 47,500 o2 Heat-affected zone SP-37-WL Room 107, 600 63, 200 b1 Weld metal 1300 54,900 41,800 7 Weld metal SP-5055-W1 Room 118, 200 68, 900 39 Weld metal 1300 73,200 49, 600 17 Base metal SP-O4O1-w2 Roon 115,900 67, 600 43 Base metal 1300 71, 400 48, 600 18 Base metal W-1671-W1 Room 11k, 400 75,600 31 Weld metal 1300 67,000 53, 600 10 Weld metal Y-8L488-wl Room 117,000 67,200 5h Weld metal 1300 69, 400 L4, 300 22 Weld metal Table 4.11. INOR-8 Transverse Creep Tests on Butt Welds in l-in. Plate Tests conducted at l300°F at 27,500-psi stress Rupture Total Time Strain Weld Plate No. (hr) (%) SP-23-Wl 6 1.0 SP-26-W1 41 9.0 SP-27-Wl 13 ~0,2 SP-5055-W1 127 L.5 SP-OU01-W2 493 25.0 W~1671-W1 19 k.0 Y-8488-wW1 370 15.0 111 resistant to both molten fluoride-sglt corrosion and high-temperature oxidation. Miller-Peaslee shear-test specimens® of INOR-8 brazed with this alloy exhibit good shear strengths at MSRE-operating temperatures, as shown in Table 4.12. Joints were brazed in the flat position in both inert- and dry-hydrogen atmospheres with gaps ranging from contact to 0.010 in. Testing at 13C0°F in air showed the joints to have a minimum shear strength of 12,500 psi and an average of 18,100 psi. The joints brazed in helium exhibited slightly higher strengths than did equivalent hydrogen-brazed Jjoints, and attempts are being made to determine the reason. Additional shear-test specimens have been brazed and will be tested to provide room-temperature shear-strength data as well as data on specimens aged at 1300°F for 1000 and 10,000 hr. Table 4.12. Miller-Peaslee Shear-Strength Data for Brazed Joints Base metal: INOR-8 Brazing alloy: 82 Au — 18 Wi (wt %) Brazing temperature: 1830°F Brazing time: 10 min Testing temperature: 1300CF Testing atmosphere: Air Average Shear Brazing Gap Brazing Number of Strength (mils) Atmosphere Specimens (psi) 0 Hy 3 16, 800 0 He 2 23,100 1.5 Hy 3 19, 800 1.5 He 2 19, 800 2 Hy 3 15,300 2 He 2 17, 400 L Ho 3 18, 200 Ly He 2 20, 400 6 Ho 3 16,100 6 He 2 19,000 10 Ho, 2 15,700 10 He 2 17,900 Long-time diffusion studies of INOR-8 lap joints brazed with the gold- nickel alloy have also been conducted. Specimens were aged from 1000 tec 10,000 hr at both 1200°F and 1500°F in air. The maximum amount of diffusion of brazing alloy into the base metal, as detected by microhardness measurements was only about 2 mils at 1200°F and 7 mils at 1500°F. This minor amount of diffusion is expected to have little effect on joint strength and over-all base-metal properties. To date, four Jjoints of the type shown in Figs. 2,26 and 2.27 have been brazed in the laboratory to determine the quality of joint. Ultrasonic inspection of the Jjoints indicated a few unbonded areas ranging from very small scattered voids to occasional regions up to one-fourth the length of 112 the joint. This Jjoint was greatly overdesigned to accommodate occasional defects and, no doubt, would be adequate in a molten-salt system. Never- theless, a slightly tapered joint design is being investigated; it is expected to eliminate nearly all nonbonded areas and to make remote assembly somewhat easier. 4.5 MECHANICAL PROPERTIES OF INOR-8 A program has been started to determine the strain-fatigue behavior of INOR-8, and, to date, four tests, two at 1300 and two at 1500°F, were performed on rod specimens of heat SP-25. These test results are shown in Fig. 4.10 with a plot of Coffin's equation, 1 F] 1l - Nf ep = g_ef vhere Ne is the number of cycles to failure, €, 1s the plastic strain range and €r is the true fracture ductility. The tensile ductility at Ny = 1/4 was used as the intercept for these curves. UNCLASSIFIED ORNL-LR-DWG 63279 100 — 11 ® HEAT SP-25: 1300°F O HEAT SP-25; t500°F B | - . ! 40 cycles per hour - L1 R I I S : | NN | i “h !, 4.7 77 I I 1 Ll I + : | L | 4 | ! I/IH €5, TRUE PLASTIC-STRAIN RANGE (%) . o | | ‘ A 0.1 1 10 100 N, ,CYCLES TO FAILURE Fig. 4.10. Strain Fatigue Curves for INOR-8. Further strain-fatigue tests, both mechanical and thermal, on tubular specimens from another heat of material are in progress. A number of tensile tests have been performed on specimens subjected to creep loading for various times. No deleterious effect of prior creep was observed in the tensile properties. 113 Creep tests on a weak, low-ductility heat material, M-1566, indicate that soaking at 1600 and 1800°F for 2 hr increased the rupture life and ductility. 4,6 TIMPREGNATION OF GRAPHITE BY MOLTEN SALTS The use of unclad graphite in the molten-salt reactor requires that a minirmum of the graphite pore space be filled with salt. A test program is in progress to study the impregnation of different grades and sizes of graphite by LiF-BeF,-ThF), -UF), (67-18.5-14-0.5 mole %) at 1300°F for various times and pressures. Three additional grades of graphites failed to meet the low salt-impregna- tion requirements in the standard screening test in which the exposure was for 100 hr under a salt pressure of 150 psig. However, the tests indicated that increasing the "as-fabricated" diameter of a rod of graphite or fabricating it in the shape of a pipe made it more "open" to the molten salt. The results of the tests are sumarized in Table 4.13. Table 4.13. Impregnation of Various Grades of Graphite by LiF-BeF,-ThF)-UF), (67-18.5-14-0.5 mole %) Test conditions: Temperature: 1300°F Pressure: 150 psig Time: 100 hr Nominal dimensions of test specimens: 0.5 in. in diam and 1.5 in. long Grade of Nominal Dimensions Bulk Density Bulk Volume Graphite of Graphite, As- of Graphite of Graphite Fabricated (g/cc) Impregnated (in.) by Salt (%)2 Rods (diameter only) GB-3.6 3-5/8 1.91 3.8 GP-2 2 1.87 h.1 GP-3.4 3-3/8 1.88 k.9 B-1-B-1 1-1/4 1.8k 6.2 B-1-B-2 2 1.76 8.5 B-1-B-3 3-1/2 1.70 13.3 Tube B-1-B-4 3-1/16 0D x 2 ID x 3 long 1.67 16.4 %gach value is an average of three. 114 Grades GB and GP were extruded and impregnated types of graphite. Grade B-1-B was the base stock for a low-permeability graphite, grade B-1l. (Base stock as used here means a graphitized graphite shape that has not had its open pore spaces filled and/or sealed by a series of impregnations with carbon- base materials and graphitizing treatments to make it into a low-permeability graphite.) Most of the rod-shaped graphite was fabricated with relatively large diameters. L,7 AMMONTIUM BIFLUORIDE AS AN OXYGEN-PURGING AGENT FOR GRAPHITE Tests were run to determine the effects of temperature on removing oxygen and oxide contamination from graphite by the thermal decomposition products of NH,F«HF. Purging temperatures varied from 1300 to 930°F, but times were held constant at 20 hr. INOR-8 specimens were also included in these test runs to observe the effect of the decomposition products on this alloy., Test conditions are shown in Table 4,1k, Table L.1Lk. Effect of Temperature on Purging Oxygen and Oxide Contamination from Graphite by Means of the Thermal Decomposition Products of NHhF-HF INOR-8 Reaction Time Exposed to Test System Purging Conditions During Purging b Grade of ) .a Weight Payer Salt W1?hout ) romite vewl (S T Peswe (ple) G Memess 00 Preeiptiacing AGOT 1300 20 86 22 2000 AGOT 1110 20 83 58 2000 AGOT 930 20 €0 ho 2000 R-0025 INCR-8 1300 20 90 35 +0.137 <0.5 2000 R-0025 INOR-8 1110 20 80 46 +0.130 1000 R-0025 INCR-8 930 20 70 Ly +0.002 o¢ 2000 aPressures at the purging temperature. bLiF-Bng-UFh (62-37-1 mole %). CReaction layer too thin to measure. 4.7.1 Removal of Oxygen from Graphite The LiF-BeF,-UF) (62-37-1 mole %) salt is a sensitive oxygen detector and in the presence of oxygen at 1300°F readily precipitates part of its uranium as UO,. This reaction is easily detected using radiography. Therefore, to determine %he effectiveness of the purging of graphite crucibles with the thermal decomposition product of NHhF-HF, the purged crucibles were charged with the LiF-BeF -UFh salt and heated to 1300°F. At intervals of time the crucibles containing the salt were radiographically monitored for a UO precipitate. Results shown in Table L.1l4 indicated that purging at tempera- tures of 930, 1110, and 1300°F was equally effective in removing oxygen from AGOT and R-0025 graphites. All the purged graphite crucibles contained the LiF-BeFe-UFh salt at 1300°F for more than 2000 hr without U'O2 being detected. 115 There has been no indication, at this time, that the different porosities of AGOT and R-0025 graphites affected the purging action. It is expected that porosity might be a factor in a more impermeable grade of graphite. 4,7.2 Effects of the Thermal Decomposition Products of NH)F-HF on INOR-& To secure additional data on the reaction of the thermal decomposition products of the NH),F-HF with INOR-8,9 0.040-1in.-thick sheet-type tensile specimens were included in the purging test using grade R-0025 graphite. A layer was formed on the INOR-8 at all purging temperatures. This layer is being identified. INOR-8 reaction data observed as weight changes are summarized in Table 4.1hk. The thickest of the reaction layers apparently did not seriously harm the tensile properties of the INOR-8. These data are shown in Table L4.15. Table 4.15. Tension Test Results® on INOR-8 Treated with the Thermal Decomposition Products of NHMF-HF for 20 hr at 1300°F Yield Strength Extension (psi, (%) 2% Offset) At Room Temperature Controls® 43.5 45,000 Treated specimens 38.5 45,600 At 1250°F Controlsb 19.2 33, 400 Treated specimens £2.3 33, 600 %Fach value is an average of three. All specimens were nominally 0.040O-in, thick. bControl specimens had the same thermal history as the treated specimens but were exposed to an atmosphere of pure argon. One-hundred-hour accelerated corrosion tests at 1832°F in the LiF-Belo- ZrF) - ThF),~UF (70-23-5-1-1 mole %) salt were made on control and exposed (with reaction layers) specimens from the 1300°F purge to determine the effects of this layer. The reaction layer was removed by the salt but there was no difference in corrosion attack (primarily intergranular to a depth of 6 mils) between the control specimens and those with reaction layers. More detailed studies will be made to determine the effect of the reaction layer on INOR-8 corrosion resistance to molten fluorides. 116 REFERENCES MSRP Prog. Rep. Feb. 28, 1961, ORNI-3122, p Ll-L6. Ibid., p 81. Ibid., p 81-84. Ibid., p 89. J. C. Richmond and W. N. Harrison, "Equipment and Procedures for Evaluation of Total Hemispherical Emittance,"” J. Am. Ceram. Soc., 39 (11), 668-T73 (1960). - E., F. Nippes et al., "An Investigation of the Hot Ductility of High- Temperature Alloys," Welding J. 34 (L), 183-85 (1955). MSRP Quar. Prog. Rep. July 31, 1960, ORNI~301L4, p 63-6k. R. L. Peaslee and F. M. Miller, Proposed Industry Wide Standard Procedure for Testing the Shear Strength of Brazed Joints, Wall Colmonoy Corp., Detroit 3, Michigan. MSRP Prog. Rep. Feb. 28, 1961, ORNI~3122, p 98. 5. IN-PILE TESTS 5.1 GRAPHITE-FUEL CAPSULE EXPERIMENTS A molten-salt-fuel capsule irradiation experiment, ORNL-MIR-4T7-3, has been operated at the Materials Test Reactor and is now undergoing postirradiation examination. The experiment was designed to determine the effect of irradiation on: (1) graphite at or in excess of MSRE temperatures and heating conditions, (2) the wetting characteristics of the fuel with graphite at high temperature, and (3) the compatibility of the principal MSRE structural materials with the molten salt. The experiment contained four capsules with fuel [UF4(93%UZ°%)-ThF4-ZrF,- BeFo-Li7F (1.4-1.2-5.2-23,2-69.0 mole %)] in graphite boats, as described previously.l The graphite and fuel were encapsulated in cans of INOR-8 which were irmersed in molten sodium for convenience in heat removal and temperature measurement, Two capsules contained AGOT graphite impregnated prior to irradi- | ation with fuel salt, and two contained R-0025 graphite, unimpregnated. Also, included in each capsule were specimens of molybdenum, pyrolytic carbon, and INOR-8 in contact with the fuel. The temperature history of the capsules during the irradiation period, May 5 to July 24, is summarized in Table 5.1. The over-all time-averaged temperatures Table 5.1. Temperature History of Graphite-Salt Interfaces in ORNL-MIR-4T-3 Capsules Type of Sample . Temperature Unimpregnated Impregnated Impregnated Interval Graphite Graphite Graphite (°c) (15, 16)% (8)2 (3)% i} Time at Steady Temperature (hr) 0-50 318 318 318 50-750 33 33 27.5 750-800 803 803 L 800-850 702 702 0.5 850-900 601 900-950 89l 950-1000 11 Total: 1856 1856 1856 Time at Non-Steady Temperature (hr) 56 56 56 Total Irradiation Time 1594 159k 1594 aCapsule identification number. 117 118 of the graphite-to-salt interfaces were about T90°C for the capsules with unimpreg- nated R-0025 graphite and 810 and 900°C for the capsules with impregnated AGOT graphite. The difference between the temperatures of the two impregnated capsules corresponds to the differences between the local ambient temperatures in the sodium. These temperatures conservatively are above the design temperature of 1300°F proposed for the MSRE reactor. They are, however, below the proposed tem- perature for the capsules.! This is attributable largely to the fact that the impregnated graphite contained only 9 g of fuel instead of the 17 g assumed in the design. It also is probable that the changes made in the MIR core, namely, the removal of ANP experiments and the relocation of control rods, reduced the thermal neutron flux at the HB-3 beam hole and also contributed to the lower temperatures. The steady-state temperatures of the capsules were constant within *10°C. The accumilated time during non-steady state operation, listed in Table 5.1, includes only the transient times for temperature changes of 30°C or more. The experiment was removed from the MIR, partially disassembled, and is now at Battelle Memorial Institute, where further dismantling and postirradiation examination will be conducted. As an aid to the interpretation of the irradi- ation experiments, out-of-plile experiments have been initiated by the Reactor Chemistry Division to duplicate the temperature history, including transients, on control sample capsules. REFERENCE 1. MSRP Prog. Rep. Feb. 28, 1961, ORNL-3122 (June 20, 1961). 6. CHEMISTRY 6.1 PHASE-EQUILIBRIUM STUDIES 6.1.1 Systems Involving NaF with ZrF4, ThF4, and UFg and. Other Constituents During the past four years the phase relationship for one after another of the systems limiting the quaternary systems LiF-BeF,-ThF4-UF4 and NalF'-BeFs-ThPg- UF4 have been completely explored and described. Systems involving LiF have had priority because of their direct bearing on fuel, blanket, and coolant materials for the MSRE, and the results have been summarized previously.l For comparison, and because of the possibility that NaF-based materials will be useful in the future, a parallel investigation of the phase relationships in the binary and ternary systems limiting the quaternary system NaF-BeF,-ThF4-UF4 has been pur- sued; diagrams of these limiting systems have been reported® % as they were completed, and the last one, for NaF-ThF4-UF4, is described below. DBefore the inclusion of 7ZrF4 in the MSRE fuel was found necessary, only preliminary infor- mation was available regarding the systems LiF-BeFs-ZrFs and NaF-BeFo-7rFg4; brief studies have been made for evaluating the usefulness of NaF-BeF, and LiF- BeF- mixtures as solvents for zirconium alloy fuel elements in the Fluoride Volatility Process. Current investigations involve the systems LiF-BeFs-ZrF4, NaF-BeFo-iZrFg, and composition~temperature studies of several complicated mixtures whose compo- sitions are similar to the MSRE fuel, LiF-BeFo-ZrFy-ThFs-UFy (70-23-5-1-1 mole %). 6.1.2 The System NaF-ThF,~UFg4 With but few exceptions, melts containing ThF,a, UF4, and NaF crystallize on cooling to form complex compounds in which the Th4* and ust positions are to some extent interchangeable. In all but two composition areas of the ternary, as shown in Fig. 6.1, the final products of crystallization are a mixture of solid solutions containing both ThF4 and UF4. Mixtures of NaF-ThF,-UF4 containing more than 66.7 mole % NaF produce pure NaF among the final products of equilibrium crystallization; solid solutions of iNaF:ThF4 and 3NaF.UF4 are not stable equi- librium phases at low temperatures and decompose to yield NaF. The only Nal-ThF, or NaF-UF, complex compound in which at least partial interchangeability of UFs4 and ThF4 does not occur is in NaF-ThF4. Continuous solid solutions occur across the composition section ThF4-UFs and in the composition-temperature range in which 3NaF.2ThF4--5NaF+3UF4 and f-4NaF«ThFs--f-3NaF-UF4 solid solutions are stable. Extensive limited solid solutions are formed along the constant NaF compositions 33.3, 46.2, 66.7, 75, and 80 mole %. Of these sections, limited solubility is observed between the compound pairs NaF.2ThF4--NaF:2UF4, TNaF-.6ThF4--TNaF.6UF,, oNaF «ThF4--2NaF.UFy, in @-3NaF.UFs, and in Q-4NaF.ThF,. The principal difference between the systems NaF-ThF4-UF4 and LiF-ThF4-UF4 in respect to segregation of uranium- and thorium-bearing phases is that 1n the NaF system limited solid solutions are the rule, while in the LiF-based system they are the exception. Solubility of a small amount of either ThF4 or UF4 in 119 120 UNCLASSIFIED ORNL-LR-DWG 45512R ThE, 1 @ NaF a—4NaF - Thfy ss (C) B-4NaF -Thiy — §-3NaF-Ufg ss (D) a-3NaF-UFgss (E) Ba-2NoF - Thigss ® 8 -2NaF - UFsss (G) 3NoF - 2ThF, -5NoF - 3UF, s (H) a-NoF-Thr, (1) 7NoF -6ThF4-7Naf - 6UF4 55 @ NaF - 2ThFsss ® NaF - 2UF4ss COMPOSITION IN mole Y L) ThE-Ufgss TEMPERATURE IN °C 100 NoF-2Thf, 834 NaF-ThFy 3NaF-2Thfy” 742 £630~ m.p. 705 ZNGF-ThE" ‘ &O o, / B £ 618 m. p. 651 4NaF-ThF,- 4 g @ £632° \\ 710 ) ) O\ o 706} (e N 3 Wss N ©)] 692 ) \ o) %b@oo & @ ® ® A)Oo = [@) % @ \\ N7 e 2 \ ) N \ - X KON ALV . 995 £618 £l Fe4B P 673 m.p. 78 £ 680 NaF-2UF, m.p. 629 2NoF-UF" SNoF-3UF" 7NaF-6UR, BNGF'UE‘ Fig. 6.1. The System NaF-ThF,-UF,. NaF-~-Thl'4-UF4 solld-sclution phases containing a much larger amount of the other tetrafluoride is practically complete. Solid-phase segregation in mixtures con- taining nearly equal amounts of ThF, and UF4 yields solid solutions moderately richer in one of the tetrafluorides, though in such processes the residual liquids do not become appreciably more concentrated in either ThFs4 or UF4 than the original composition. The composition sections UF4-ThF$, 2NaF.ThF4--2NaF-UF,;, and 4NaF.ThF,-- 3NaF.ThF, were reported previously.s" The composition sections TNaF.6UF4- - TNaF+6ThF, and NaF.2ThF4--NaF«2UF4 are shown in Figs. 6.2 and 6.3, respectively. The temperatures and compositions of invariant equilibria and singular points in the system NaF-ThF4-UF, are given in Table 6.1. TEMPERATURE (°C) 121 UNCLASSIFIED ORNL-LR-DWG 64964 750 [ | 11+41:2 ss+ LIQUID 1:2 ss + LIQUID _J 1:2 ss+(7:6—3:2) ss + LIQUID Mmoo\ - /- 1:2 ss+7:6 55+ LIQUID i e — 7:6 ss + LIQUID [ I —\\ A S [ o / ! 114+4{(7:6-3:2)ss + LIQUID > (e} ] i ?20 — + - i T P P — :/ 7:6 ss+ LIQUID ‘ 710 - M'G*}Z)SS'F ! f LIQuID 700 1 : 11+(3:2-7:6)ss + LIQUID “:\‘h—(B:Z—?:G)ss+(7:6—3:2)ss+1:1 | 690 i ————— ] | | 1:2 = NaF - 2ThF, s : 7:6 = 7NaF - 6ThFy— 7NaF - 6UF4 55 680 I 3:2=23NaF - 2ThF;ss Lo - l 1:1 = a—NgF - Thf [ | | I 1 670 L ' 0 2 4 o 8 10 12 14 Uk, {mole %) Fig. 6.2. Composition Section 7NaF-6UF ,—7NaF.6ThF , (Partial). 122 UNCLASSIFIED ORNL-LR-DWG 61554 950 T \ LIQUID 900 - ‘ —— 850 | R R . 83 1 ThF4-UF4 ss + LIQUID : ~ THF4 ~UF, ss + NaF -+ 2ThFg4 ss+ /7 LiQuip | S ] NaF -2UF4 ss + Thfg-UFg s5 + LIQuo —_ 800 |- 750 | P ThFg—UFy ss+ 7NaF- 6 (Th,U)F, + uoum\ L— TEMPERATURE (°C) | 700 - ---1»——— g | x! | | NaoF - 2ThF, ss + I : NoF-2UFg ss— | | ThFy~UFg 55+ | T t 7NaF - 6(Th, U) Fq 650 [ | : ! = | ! NoF - 2UF4 55+ \| ! | 7NoF - 6(Th, U) g+ ) | : I| ThF, - UF, ss | 600 e et - ' [ : L NeF-2UFg ss | | | | 550 : . 0 10 20 30 40 50 60 UF, (mole 7o) Fig. 6.3. Composition Section NaF.2ThF ,~NaF.2UF,. 6.1.3 Stannous Fluoride as a Component of Reactor Fuels Low-melting nonvolatile fluorides such as SnFp and PbF, have sometimes been proposed as possible additives to lower the melting point of fuels. Unfortunately they are too corrosive to contain in anything except graphite and possibly molyb- denum. For concentrations of UFs4 or ThFs over 10 mole %, the use of LiF-BeF; as a solvent gives a lower melting mixture than found for the SnF, - quadrivalent fluoride binary even though pure SnFp melts at only 213°C,°B Table 6.1. Invariant and Singular Points in the System NaF-ThF4-UF4 Composition (mole %) Temp. Type of X Nab Thi s UFs (°c) Equilibrium S50l1lid Phases Present 5.5 10.5 1k 588 Eutectic B-UNaF«ThF4--B-3NaF«UFy ss, Bo-2NaF-ThFy ss, 6-2NaF.UFy4 ss 66.7 23.3 10 688 Maximum on boundary Bo-2NaF.ThFy ss, 3-2NaF.UFy4 ss curve 64 27 9 655 Futectic Bo-2NaF+ThF, ss, ©0-2NaF-UF4 ss, 5NaF «3UF4-3NaF.2ThF4 ss 6k 3k 12 665 Maximim on boundary 3NaF.2ThF4-5NaF«3UFs ss, 0-2NaF:UFy4 ss curve 59 37 y T12 Maximun on boundary 3NaF.2ThF4-5NaF.30Fs ss, TNaF.6ThFg- curve TNaF.6ThF, ss 58.5 Lo 1. TO3 Eutectic 3NaF«2ThFs-5NaF-3UFy ss, O-NaF-ThFg, 'TNaF-6UF4-TNaF-6ThF4 55 5k.5 k3.5 2 732 Peritectic 0-NaF.ThF4, NaF-2ThFs ss, TNaF:6UFg- TNaF.6ThF, ss 54 41.5 4 T34 Maximum on boundary TNaF:6UF4-TNaF-6ThFy4 ss, NaF.2ThFy ss curve 48 12 L0 T06 Peritectic TNaF«6UF4-TNaF.6ThFy ss, NaF.2ThF4 ss, NeF.2UF4 ss b7 11 Lo 692 Peritectic TNaF.6UF 4 -TNaF«6ThFy ss, NaF«2UF4 ss, ThiF'g-UFs4 sSs 46 1k Lo 710 Peritectic NaF.2ThFs ss, NaF«2UF4 ss, ThF4-UFy4 ss £zl 124 > 6.2 OXIDE BEHAVIOR IN FUELS 6.2.1 Behavior of MSRE Fuel on Freezing and Effect of Segregation on Oxide Precipitation Much effort has been devoted over a period of years to locating the lowest- melting regions in phase diagrams of salts suitable for use in nuclear reactors. But in practice, the lowest-melting compositions, which characteristically solidify without change in composition of the liquid during the freezing process, are rarely suitable for reactor use; somewhat altered compositions which retain the desirable feature of a relatively low melting point are used. As a conse- quence, fuels of practical interest are prone to segregate on freezing. 1In phase-study terminology an important aspect of the segregation behavior is re- ferred to as the crystallization path of the fuel. Complex mixtures such as the MSRE fuel can follow numerous crystallization paths depending on the conditions of freezing. Slow cooling, which favors the equilibrium crystallization path, can greatly enhance segregation into strata of different crystalline compounds. A fast rate of cooling, besides giving rise to some different crystalline com- . pounds and to altered proportions of those that form in any case, can often virtually eliminate stratification. Freeze valves, which are frozen plugs in salt lines to stop flow or to allow flow on thawing, have given generally satisfactory results in engineering tests. The presence of a frozen zone does, however, provide an opportunity for segre- gation and for the "cold trapping" of high melting or insoluble precipitates. Consequently attention has been paid to gaining a better understanding of the crystallization path of the fuel during freeze valve operation, To provide an example of the worst case (from the standpoint of segregation), a vertically aligned freeze valve® was slowly cooled to shut off a very slow down- ward flow of the MSRE fuel (LiF-BeFp-ZrF4-ThF,-UFy; T0-23-5-1-1 mole %). When sectioned and examined, a sharply segregated layer of 6LiF-BeFs-ZrF; (the equi- librium primary phase) and 2LiF:BeF, (the equilibrium secondary phase) was clearly marked and visible, due to the absence of green color associated with UFg. The following sequence of events was surmised to have occurred during - freezing. The primary and secondary phases began to precipitate on the coolest metal - and formed a layer which grew inward and upward, finally forming a complete arch or dome which stopped the flow. TFurther growth of this layer depleted the ZrFa4 content of the overlying stagnant melt, and as the temperature fell to the point that all the liquid in the frozen zone solidified, the last liquid to freeze, found as a mixture of crystals too fine to be readily identifiable, was enriched in UF4. In this portion, depleted in ZrF4 and enriched in UF4, some UO, was found. Evidently the fuel originally contained a small amount of oxide which was held in solution by complexing with Zr4+ ion, but which formed a precipitate of U0z when the ZrF4-rich phases were frozen out. Several days after the frozen plug had formed, finely divided BeO was added to the still-molten fuel in the upper tank for the purpose of forming ZrO- in order to check on the extent of cold trapping of Zr0s in the cold zone in the vicinity of the frozen plug. A precipitate of ZrQOp formed, as expected, but it appeared to have accumulated predominantly on the walls in the upper parts of the system in crystals up to 20 p in size, rather than growing preferentially in the cold zone of the freeze valve. Had the system been allowed to age for a long time, presumably more ZrOz would have been cold trapped, but under the conditions of this experiment no indications of a plug of ZrQ0s in the cold zone were found. 125 The main conclusion reached was that segregation in a stagnant fuel can cause localized depletion of ZrF4 with resultant precipitation of small amounts A contrasting and more typical case of ZrOp precipitation, without the for- mation of UO,, was afforded by a melt which became contaminated with oxide from insufficiently pure helium used as a protective atmosphere. 1In this case Zr0Os accurulated in a layer at the bottom of the melt, and no UOp was found. Crystals of ZrOs that grow in the MSRE fuel at 600 to T0O0°C are usually 20 to 30 p in size, while those which are formed during freezing, by precipitation of soluble oxide, are only 1 or 2 p in maximum diameter. An illustration of an interesting deviation from the equilibrium crystal- lization path of the MSRE has repeatedly been encountered. Slowly-cooled melts correspond to the equilibrium crystallization path in which 6LiF.BeFs-ZrF4 forms at 442°C, followed by LizBeF4 at 435°C. However, with cooling rates of 2 or 3°C per minute or faster, the nucleation of 6LiF:BeFz+ZrF4 frequently does not occur, and the supercooled melt yields both B~3LiF.ZrF4 and well-formed crystals of 2LiF.ZrF4, the result of metastable crystallization to phases which are not present in the equilibrium sequence. Temperature-time curves for one example of rapid freezing showed supercooling to 435°C followed by crystallization at 438°C, perhaps of B-3LiF:ZrF4, and then another phase, presumably 2LiF:7ZrF4, at 431°C after supercooling to 421°C. 1In all cases the major phase present is LigBeFq4, which in the rapidly cooled samples appears as crystallites too small for recog- nition by petrographic microscopy but which can, however, be identified by x-ray diffraction. In the LiF-ZrF4 binary system, B-3LiF*ZrF4 has a lower limit of stability at 4TO°C; thus its appearance in the MSRE fuel is even more surprising. The formation of a frozen zone, if carried out slowly enough,would produce a plug of the primary phase throughout the region maintained below the liquidus temperature. In actual practice, the zone is formed quickly enough that a mixture of phases is deposited, and no extensive region of segregated high-melting crystals occurs. In the temperature gradient leading away from the coldest portion, however, there will be a contour corresponding to the liquidus tempera- ture of the MSRE fuel., Here there should be a crust of the primary phase, 6LiF+BeFs+ZrFs. Fortunately, this crust has apparently always been thin enough to dissolve rapidly on raising the temperature, and sluggish responses attribut- able to high-melting phases, even in other compositions, have not been noted. 6.2,2 0Oxide Content of Fluoride Melts When an attempt was made to use oxide analyses in the 1l00- to 1lOOC-ppm range on a routine basis in several extensive experiments on fuels, a greater- than-expected variance in the oxide content among replicate samples was en- countered. The analysis of trace amounts of oxides has long been recognized as exceedingly dif‘ficulti but a method based on a fluorination procedure had shown considerable promise.® In view, however, of the frequent occurrence of appar- ently irreconcilable results, a statistical survey was carried out in a search for clues to the source of the difficulty. Analyses of T5 replicate samples from a single composite source, or standard mixture, were accumulated for statistical treatment to determine the differences attributable to two different but supposedly identical assemblies, each involving three fluorination cells as employed by two different operators. Temporal factors were also included as a parameter. A mean value of 500 ppm oxide was obtained, and 56 of the T5 analyses were distributed between 300 and 750 ppm. No significant source of variance was attributable to any of the nonprocedural 126 factors which were tested. With this point estabiished, attention is currently being turned to procedural problems. When an atterpt was made to calibrate the analytical method on UzOg, which provides a relatively large oxygen content, gocd results were obtained. The results for eighteen samples ranged from 1h4.6 to 15.6% oxide with an average of 15.10%, whereas the theoretical content was 15.18%h. The effect of using larger samples of salt will be studied next even though there are disadvantages associ- ated with the use of large samples. Also, the influence of particle size, which may affect both sampling and the completeness of the reaction, is to be investi- gated in experiments making use of a factorial design and sequential analysis. Consideration 1s being given to the feasibility of an alternative method based on the isotopic dilution of 0'® introduced for equilibration as Hgol8 vapor in a cover gas containing HF. 6.3 PHYSICAL AND CHEMICAI PROPERTIES 6.3.1 Surface Tension Apparatus for Molten Fluorides Apparatus has been designed, the major components constructed, and assembly nearly completed, for the measurement of the surface tension of fused fluoride mixtures by the bubble pressure method. Surface tensions have been measured for very few fluoride mixtures; the measurements are expected to throw light on changes in the structure and bonding with changes in the composition of molten salt mixtures. The surface tension of breeder fuels is also one of the important factors determining the wetting or permeation of graphite by fuel, a matter of considerable interest for the MSRE. With the recently constructed apparatus a search for strongly surface active additives in pure fluorides and in LiF-BeFo- based melts will be made, although no examples are known in the literature on molten salts. In conjunction with the concurrent work on the permeation of graphite, the surface tension results will permit estimation of the individual contributions of surface tension and contact angle effects. (The combined effect is measured in permeation tests,) The bubble pressure method was chosen for the surface tension measurements because it is generally accepted to be the most accurate and most generally applicable method for molten salts. Since many fluorides are sensitive to moisture and air, the bubble tube and crucible of salt were enclosed in a gas- tight flanged nickel reactor which can be evacuated and filled with a thoroughly dried inert atmosphere, Provision was made for checking the water content of the inert gas blanket and the bubble gas with a Meeco electrolytic water analyzer, Other principal innovations include the use of a special controllable leak valve (used on the mass spectrometer) for regulating the bubble flow, and the use of a cistern-~type dibutyl phthalate manometer constructed from precision~bore tubing in both arms. The procedure and apparatus for meking each measurement needed for the surface tension determination were designed to yleld results as accurate as any in the literature of molten salts. 6.3.2 Volatilization of Iodine from the MSRE Fuel Consideration of proposed experiments to ascertain the extent of liberation of fission product iodine from the MSRE fuel has led to a review of the factors controlling the release. 127 As predicted previously,ll elemental iodine is not expected in the MSRE off- gas in appreciable amounts because of the reducing action of chromium in the INOR-8 used to contain the fuel; the detailed mechanism involves the agency of a small concentration of UF3 in equilibrium with the chromium and UF4. The vapor pressure of iodide salts from the fuel is expected to be below 10" mm of Hg at reactor temperature,l? In case of a spill involving contact of the salt with air or water there will be some release of iodine activity. Oxygen in air can displace iodide ions in the fuel to produce elemental iodine, and water hydrolyzes iodide salts to give volatile HI, which also reacts with oxygen. Both direct oxidation and hydrolysis come to virtual cessation as soon as the fuel freezes, but it is difficult to predict the extent of reaction prior to solidification. Assuming that the spill is quenched by flooding with water, hydrolysis to produce HI might be expected to proceed at roughly the same rate per ion of halide as hydrolysis to produce HF, Hence if less than 10% of the fluoride leaves as HF, which should cover most cases, then a release of 10% or less of the iodine content could be predicted. 6.4t GRAPHITE COMPATIBILITY 6.4.1 Effect of Cesium Vapor on Graphite In the MSRE some of the xenon produced by fission is expected to diffuse into the graphite moderator and there be converted into cesium by decay or by neutron capture. Cesium is known to intercalate with graphite.l Changes pro- duced in the graphite surface by the cesium might conceivably change the wetting behavior, To evaluate this, graphite treated with cesium, along with untreated specimens as controls were immersed in molten salt and compared. Assuming the use of graphite with a DXe of 1075 cme/sec, it was calculatedl® that in one year of operation of the MSRE at 10 Mw, 80 g of cesium would be de- posited in the graphite, nearly all of it near the 1.4 x 10®° cm® of surface, corresponding to an average of 0.057 g of cesium per 1000 cm® of graphite. Three graphite (AGOT) cylinders, 12.77 rm in diam and 25.3 mm long, weighing 5.4 g each, were outgassed at 200°C by pumping a vacuum for 3-1/4 hr, and then treated with 26.5 liters (measured at room temperature) of helium saturated with cesium at 200°C, According to the vapor pressure data of Kubaschewski and Evans,l5 0.032 g of cesium was exposed to the graphite. This is about 1.25 g of cesium per 1000 em® of graphite. These treated specimens, and three similar graphite specimens that had merely been outgassed, were kept immersed in molten MSRE fuel at 650°C for 41 hr. No wetting appeared to have occurred and none of the specimens was changed in appearance by the experiment. Those without the cesium treatment lost 0.002 g each, probably due to outgassing. The specimen nearest the cesium inlet gained 0.012 g. The other treated specimens lost 0.001 g each. It was concluded that the exposure to cesium vapor obtained in this experi- ment was too mild to alter the graphite detectibly. 128 6.5 FUEL PRODUCTION 6.5.1 Purification Treatments Continued improvements on processing methods have been made at the facility for the production of purified fluoride mixtures. Current procedure for fluoride mixtures which melt below 800°C uses flowing hydrogen and anhydrous hydrogen fluoride to achieve purification with a minimum of corrosion of the equipment. During the period of this report approximately 885 kg of mixtures comprised of LiF, BeFs, ZrF4, ThF4, and UF4 were processed for the molten-salt reactor program. An additional 2000 kg of the mixture NaF-LiF-ZrFs (37.5-37.5-25.0 mole %) were processed for the Chemical Technology Division. The current production rate for purifying molten fluorides is approximately 1.3 £° per week for each of two units. This value is derived from a single- shift operation during a normal five-day week. However, by employing a multi- * shift operation and a seven-day week this production rate can probably be increased to approximately 1.9 £ per week for each unit. - The ant1c1pated operation of the MSRE will require the preparation of approXimately 155 £+° of fluoride mixtures for the flush, fuel, and coolant media of the reactor. From considerations of methods for transportlng these purified fluoride mixtures to the reactor, the use of some of about 50 existing cans, now in storage and filled with NaF-ZrF4-UF4 melts, appears feasible. To be suitable, these containers should be reconditioned to remove residual salts from previous usage. However, the costs incurred by this operation can be partially offset by incorporating a modification to these vessels which will increase their capacity. By lengthening each shipping container by approximately 12 in., the full capacity of the purification treatment vessels can be utilized. This modification would effectively 1ncrease the maximum weekly production rate of fluoride mixtures to approximately 3.5 £t2 for each of two operational units. With these proposed alterations the fluoride mixtures required by the MSRE could be processed during a minimum operating time of approximately six months. Eight months of lead time should allow for contingencies and unforeseen difficulties. 6.6 ANALYTICAL CHEMISTRY 6.6.1 Analyses of MSRE Fuel Mixture Thorium The evaluation of the proposed amperometric method® for the determination of thorium in the MSRE fuel mixture was completed. This method is particularly applicable to remote operations with radiocactive fuel mixtures. Zirconium An indirect method has been developed for estimating the zirconium content of fuels; it is based on a method involving the determination of total thorium and zirconium. In this method, an excess of ethylenediaminetetraacetic acid (EDTA) is added to an aliquot of the solution containing both thorium and zir- conium. After the addition of sodium fluoride, the acidity of the solution is adjusted to pH b with sodium acetate, and the excess EDTA 1s back-titrated with - 129 a solution of vanadyl sulfate. The equivalence point of the titration is de- tected amperometrically wilth a platinum electrode at an applied potential of 0.6 v vs S.C.E. At this potential, the addition of excess vanadyl sulfate re- sults in an increased rate of current flow. Calculations based on this titration yield the sum of thorium and zirconium. Since the thorium content is determined independently, as mentioned above, the zirconium content can be estimated by difference. nggen The equipment which is proposed for use in the determination of oxygen in concentrations of the order of parts per million in highly radioactive fuel was assembled and cursory tests were initiated. The number of such tests which have been completed are not sufficient to permit a proper evaluation of the equipment at this time. 6.6.2 Analyses of MSRE Cover Gas Tests, which have been described previously,'® were carried out to determine the effectiveness of titanium sponge as a "getter" for the removal of oxygen and nitrogen from helium, the MSRE cover gas., Preheated helium that was contaminated with oxygen and nitrogen at levels from 0.0l to 2% was passed over beds (10- and 20-in. deep) of titanium metal sponge in a heated quartz tube (30-mm ID). In these tests, the flow rates were varied from about 100 cc/min to 4 liters/min, and the temperature of the titanium was varied from 600 to 1200°F. In order to increase the sensitivity of the chromatographic method for the determination of oxygen and nitrogen in helium, a method was devised in which these contaminants are concentrated from a 1/2-liter sample of the gas onto a bed of molecular sieves at the temperature of liquid nitrogen. In this method, the exit gas, from the purification apparatus, flowing at a rate of 100 cc/min, is passed for 5 min through a tube (1/4-in. OD and 1 ft long) that is packed with thoroughly dried, Linde, type 5A, molecular sieves. During this operation, the trap assembly is submerged in a liquid-nitrogen bath. At the end of the 5-min period, the molecular-sieve trap is inserted into a stream of specially purified, helium carrier gas by means of a six-way valve, after which the liquid-nitrogen bath is removed and replaced with warm water. The oxygen and nitrogen which are trapped quantitatively from the sample are released immediately thereby and then routed through a molecular sieve colummn (1/4-in. OD and 4 ft long) in a Greenbrier model-050 chromatograph. The concentrations of the resolved gases in the eluate from the chromatographic column are determined in the conventional manner by com- paring the thermal conductivity of the eluate wlth that of the pure carrier-gas stream. Calibrations were performed by carrying out the same operations on a sample of the inlet gas to the purification train. The fraction of a contaminant passing through the purification train is thus given by the ratio of the area of a contaminant peak in the exit-gas chromatogram to the area of the corresponding peak in the inlet-gas chromatogram. : Concentrations of nitrogen as low as 0.1 ppm can be measured easily by this method. The sensitivity can obviously be increased by trapping the contaminants for periods longer than 5 min. On the basis of repeated calibration experiments, it has been found that the precision (relative standard deviation) of such measurements is about 3% for concentrations of the order of 100 ppm. The sensi- tivity of the method for the determination of oxygen actually exceeds that for the determination of nitrogen; however, the measurement of oxygen 1s complicated by the presence of argon, which accompanies the oxygen when separations are per- formed on colurms of molecular sieves at room temperature. Since the oxygen, from which contaminated helium samples have been prepared, contains almost 1% 130 of argon, the exact measurement of oxygen becomes uncertain when more than 9% of the oxygen is removed by the purification process. At present, the measure- ments of oxygen are corrected by subtracting the contribution of argon on the measured oxygen-argon peak. This method is not, however, completely satisfactory, since under some conditions only a few tenths of a percent of the original oxygen escapes capture in the titanium purifier. Future experiments will be carried out with contaminated helium samples prepared from argon-free oxygen; the oxygen will be generated by the thermal decomposition of lead dioxide. REFERENCES 1. C. F. Weaver, R. E, Thoma, H., Insley, and H. A. Friedman, Phase Equilibria in Molten Salt Breeder Reactor Fuels, ORNL-2896 (Dec. 1960). 2, R. E. Thoma (ed.), Phase Diagrams of Nuclear Reactor Materials, ORNL-2548 (Nov. 1959), p 110. 3. MSRP Prog. Rep. Feb. 28, 1961, ORNL-3122, p 111. 4k, C. F. Weaver et al., J. Am. Ceram. Soc. Lk, 146 (1961). . C. F. Weaver et al., J. Am, Ceram. Soc. 43, 213 (1960). . MSRP Quar. Prog. Rep. July 31, 1959, ORNL-2799, p T5. > 6 7. MSRP Quar. Prog. Rep. Apr. 30, 1960, ORNL-2973, p 67. 8 B. J. Sturm, Stannous Fluoride as a Component of Molten-5alt Reactor Fuels, ORNL CF-60-10-133 (October 1960). 9. MSRP Prog. Rep. Feb. 28, 1961, ORNL-3122, p 27. 10. MSRP Prog. Rep. Feb, 28, 1961, ORNL-3122, p 132. 11. F. F. Blankenship, B. J. Sturm, and R. F. Newton, Predictions Concerning Volatilization of Free Todine from the MSRE, MSR-60-4 (Sept. 29, 1960). 12, B. J. Sturm, Calculated Vapor Pressure of Iodide from MSRE Fuel, MSR-60-6 (Oct. 11, 1960). 13. W. Rudorf, "Graphite Intercalation Compounds," in Advances in Organic Chemistry and Radiochemistry, (ed. by H. J. Emeleus and A. G. Sharpe) vol I, pp 22L-66, Academic Press, New York (1959). 14, I. Spiewsk, Cesium Distribution in MSRE Graphite, MSR-60-40 (Nov. 23, 1960). 15. 0. Kubaschewskl and E. Evans, Metallurgical Thermochemistry, 2nd ed., Wiley, New York (1956). 16. MSRP Prog. Rep. Feb. 28, 1961, ORNL-3122, pp 130-3%. 7. ENGINEERING RESEARCH 7.1 PHYSICAL-PROPERTY MEASUREMENTS The determination of the enthalpy of the coolant mixture LiF-BeFz (68-32 mole %) was completed. The data, shown in Fig. 7.1, can be represented by the following equations: () solid (50 to 360°C), H, - H30 = -6.22 + 0.225 t + (k.24 x 10~%) tZ ; (1) (b) Liquid (480 to 8620°C), H, - H,. = 123.90 + 0.185 t + (2.97 x 10-*%) t% . (2) The enthalpy, H, is in cal/g for t in °C. As seen in Fig. 7.1, the solid- liquid transition was not sharply defined and extended over the broad span of 360 to LB0°C; there are suggestions of two-phase transformations in this region. The heat of fusion, evaluated between 360 and 480°C, was 151.h4 cal/g. UNCLASSIFIED ORNL-LR-DWG 63280 500 | i "/‘./irwn/, 400 fr——f— b o - 5 k S - el - 3 OO Y IS — e == - L - . . . | —_— . p— N rg \ .| : I | :>/£fi’ 5 4N A, By S 250 | # - o 2 036 %fi}( - —_ - 2 ///‘fiu - N o 3 E N 2 s R B« S o034 b L CIRCULATED ____| = prad ) s 7 = / \ o N w ? M, = 200 —- b e e T E 0.32 e S x UNCIRCULATED T N i - -._. o 175 L o 0.30 I i f ~ ‘ | 150 . \ 0.28 500 600 700 800 500 600 700 800 TEMPERATURE (°C) TEMPERATURE (°C) Fig. 7.2. Comparison of Enthalpies and Heat Capacities of LiF-BeF ,-UF ,-ThF, (67-18.5-0.5-14 mole %) Mixtures. 133 viously;a sample analyses are reproduced in Table 7.2. The mean value for all three samples over the range 550 to 800°C is 0.335 cal/g-°C; this exact agree- ment in the mean values suggests that the observed variations in c,, may lie within the experimental precision associated with these measurements. This aspect of the enthalpy program will be investigated further. Average values estimated by using Eq. (3) agree within 7% with the experimental mean (sample A, 0.347 cal/g-°C; circulated, 0.312 cal/g-°C). Table 7.2. Composition of LiF-BeFz2-UF4-ThF4 Salt Mixture Samples Constituent (mole %) Sample LiF BeFz UF 4 ThF 4 Nominal 67.0 18.5 0.5 14.0 A 70.3 15.4 0.6 13.8 Uncirculated 70.8 16.6 0.5 12.1 Circulated” 66.8 15.8 0.6 16.8 aAverage of three samples taken from loop sump. 7.2 HEAT-TRANSFER STUDILES Interpretation of the data obtained in the study of heat transfer with the LiF-BeFz-UF4-ThF4 (67-18.5-0.5-14 mole %) salt mixture has been continued, with primary emphasis on the resolution of the abnormal heat balances observed.Z Since the convective heat gain by the fluid is calculated from the equation Qp = W Cp At (4) the experimental measurements for the flow rate (w), heat capacity (cp), and fluid temperature rise {at) must all be re-examined. The fluid-flow rate was determined by use of a turbine-type flowmeter for which the primary data were displayed as the rate of turbine rotation. This was converted to 1b/hr of salt flow, using a water-flow calibratlion corrected for the specific-gravity difference between water and the salt. Since experi- mental data on the density of the LiF-BeFz-ThF4-UFs (67-18.5-14-0.5 mole %) mixture were not available, recourse was had to an empirical correlation devel- oped from measurements on other BeFz-containing mixtures. This yielded the relation (based on the nominal composition of the salt mixture): o =3.9% - 0.00087 t (5) where p is in g/cm8 for t in °C. However, as noted from Table 7.2, there is a great deal of uncertainty as to the actual composition of the salt circulated in the loop; this has a significant effect on the predicted density. A compari- son of estimated densities (at 1399°F) is given in Table 7.3; also shown are values obtained by using a stochastic correlation developed from data on the 134 Table 7.3. Estimated Densities. for an LiF-BeF2-ThF4-UF4 Salt Mixture Com ositionb Correlation D Nominal Uncirculated Circulated 0, g/cm’ Empirical 3.30 3.15 3.61 Stochastic 3.49 3.38 3.73 ®Run MF-2, t = 1399°F. bAs given in Table 7.2. molar volumes of LiF-BeFz mixtures.® For a typical run (MB-2), the heat bal- ance ratio, /qe (where 9. i1s the measured electrical energy input), varies from 1.05, using the empirically derived density value for the uncirculated salt, to 1.24, with the stochastic estimate of the circulated-salt density. An empirical determination of the density of this salt mixture is planned. A second possible error relating to the rate of salt flow derives from a lack of information as to the influence of fluid viscosity on the output of the turbine meter. SGince preliminary inquiries suggest that this may be signifi- cant, a study is in progress using liquids of wvarious viscosities [water, eth- ylene glycol, and possibly a noncorrosive salt such as NaNOz2-NaNO3-KNOs (L40-7- 53 wt %)] to calibrate a turbine-type flowmeter., A further uncertainty in the convective heat gain relates to the discrep- ancy in heat capacities as measured with circulated and uncirculated salt sam- ples (see Table 7.1). Thus (considering run MB-2, with an average salt temper- ature of 1408°C in the INOR-O6 section), the heat-capacity ratio (circulated to uncirculated) is 1.06. As noted in the discussion above on the enthalpy of the LiF-BeFa-UF4-ThFs (67-18.5-0.5-14 mole %) mixture, the influence of this uncer- tainty in ¢, is & minimum near 1247°F (values are equal at this temperature). Again, furtger measurements on the heat capacity are indicated. A re-examination of the calibrations for the thermocouples used in obtain- ing the fluid mean temperatures shows that, while a systemic error exists in the emf output, the readings from the thermocouple pairs (inlet and outlet) are sufficiently close to disallow any effect on the value of Qp- REFERENCES 1. MSRP Quar. Prog. Rep. July 31, 1960, ORNL-301k, p 85. MSRP Prog. Rep. Feb. 28, 1961, ORNL-3122, pp 1l40-42. 3‘ Ibid-j pp 123-250 8. FUEL PROCESSING 8.1 PREPARATION AND ANALYSIS OF COMPLEX FLUORIDES OF SbFs WITH KF, AgF, AND SrFs The complex fluoride salts of SbFs with KF, AgF, and SrF» were prepared by dissolving the KF, AgF, or SrFs in anhydrous hydrogen fluoride and adding a small excess of SbFs liquid. In each case a preclpitate formed, showing that the com- plex is less soluble than the simple fluoride. The KF-S5bFg complex is soluble in anhydrous hydrogen fluoride at 20°C to 150 mg per gram of solution, or 34 mg of KF per gram of solution. Qualitatively, the AgF-SbFg complex is rather ine soluble, probably less than 1 mg per gram of solution, and the SrFz-SbFg salt is intermediate in solubility. The compounds were purified by recrystallization from anhydrous hydrogen fluoride after two anhydrous HF washes of the original precipitate. The KI-SbFg compound was so soluble that a rather concentrated solution was formed during evaporation of the excess HF, and solids tended to form a surface film; good single crystals were not obtained., The SrFo-SbFg compound formed crystals on the sides of the bottle, above the liquid phase. The AgF-SbFs compound formed granular brown crystals of reasonably good quality for crystal structure determination. Analyses showed that the KF salt was KSbFg, with K/Sb mole ratios of 1.07 and 1.14 in two preparations and F/(K + 5Sb) ratios of 1.07 and 1.00. Analyses for the other two salts were not entirely consistent. The AgF-S5bFs compound was probably AgSbFg since the Ag/Sb ratio was 1.10, but the total of Ag, Sb, and F added up to llO%, probably because of a high F determination. For the S5rFz-SbFg compound the constituents added up to only 8T% (compared with 96 to 98% gener- ally). The Sr/Sb mole ratio was 0.71, and the F/(2Sr + 5Sb) ratio was 1.5; the compound apparently contained considerable HF. 8.2 PREPARATION AND ANALYSIS OF THE NaF-MoFg COMPLEX The NaF-MoFg complex was prepared by mixing solutions of NaF and MoFg, both in anhydrous HF. There was little if any heat of reaction, and no precipitate formed. The solution was boiled to dryness, both HF and MoFg being evolved, and the boiling point increased slowly to about 60°C before any solids appeared. The salt was further dried by heating to about 80°C under a stream of helium, The resulting solid had an Na/Mo ratio of 5.15 and an F/(Na + 6Mo) ratio of 1.05, in agreement within analytical error with the formula MoFg-5NalF. Later prepa- rations were de-solvated by pumping off the excess HF solvent with a vacuum pump for several hours at a little above room temperature. These samples generally had Na/Mo ratios of 30 or greater, indicating that the pumping removed most of the MoFg. There was an indication that the solids contained 0.8 to 0.9 mole of HF per mole NaF. Apparently MoFg and HF compete for the NaF, with HF probably being complexed more strongly at these temperatures. 135 136 8.3 PREPARATION AND ANALYSIS OF COMPLEXES OF UFg WITH LiF, NaF, AND KF The preparation of Ufg complexes with LiF, NaF, and KF was attempted in a manner similar to that for the MoFg complexes. All were quite soluble in HF, and the solids obtained upon evaporation of the excess HF were yellow for LiF and NaF and orange for KF. The LiF preparation had an Li/U mole ratio of 14 and contained 0.6 mole of HF per mole of LiF. Three NaF-UFg preparations, with suc- cessively more vigorous treatment to remove HF, had Na/U ratios of 10, 17, and 39, and Na/HF ratios of 1 to 2. Apparently UFg was removed rather than HF by the de-solvation treatment, The KF complex had a K/U mole ratio of 8.5 and contained 2 moles of HF per mole of KF. Thus, all the compounds contained less UFg than expected, and there was some indication that UFg was removed by the procedures used to remove excess HF solvent., It appears that HF and UFg are complexed com- petitively by the alkali metal fluorides and that all the complexes contain some HF. They are therefore different from the complexes formed by reaction of UFg gas with solid NaF. It appears, further, that the HF is held more tightly than 3 UFg, at least at room temperature. This suggests the possibility of separating some of the volatile fluorides by displacement, for example, by displacing UFg - from an NaF-UFg bed with HF, This would require a knowledge of the vapor pressure of all these compounds as a function of temperature and composition. 8.4 ATTEMPT TO SEPARATE RARE EARTH FLUORIDES FROM MSBR BLANKET SALT BY SbFs IN HF Analytical results from one experiment in which SbFs in HF was used to extract rare earths from an MSBR blanket salt indicate that the rare earths were not separated from the ThF4. The salt (LiF-BeFo-ThF4-UF4, 67-18.5-14-0.5 mole % containing 0.05 mole % rare earth fluorides and trace fission products) was first contacted twice with 95% HF - 5% water to dissolve the LiF and BeF». The remaining salt, containing the ThF4, UF4, and rare earths, was then contacted twice with HF containing 5 vol % SbFs. This solvent dissolved very little of anything and, in particular, did not dissolve a significant portion of the rare earths. However, it should be pointed out that the analytical results were not consistent with the known composition of the salt, so a large error could be present. These results suggest that the rare earths may be present in the salt in the form of a compound, probably with LiF, which is insoluble in HF-~SbFs solutions, at least under the conditions of this test. - AUGUST 34, 196! R. B. BRIGGS, DIRECTOR o P MOLTEN SALT ADYANCED DEYELOPMENT REACTOR EXP ERIMENT A. L. BOCH, PROJECT ENGINEER R ~ ENGINEERING RESEARCH COMPONENT DEVELOPMENT FUEL REPROCESSING METALLURGY REACTOR CHEMISTRY H. W. HOF FMAN R I. SPIEWAK* R D. E. FERGUSON cT A. TABDADA® " W, R. GRIMES RC R. G. DONNELLY " H. F. MCDUFFIE* RC T. K. ROCHE " F. F. BLANKENSHIP® RC C. M. BLOOD RC J. E. EQRGAN RC G. M. HEBERT RC W. F. SCHAFFER RC R. E. THOMA RC B. F.HITCH RC W. M. JOHNSON RC DESIGN METALLURGY REACTOR CHEMISTRY RADIATION TESTING COMPONENT DEVELOPMENT CONSTRUCTION OPERATIONS E. S BETTIS R A. TABOADA* M H. F. MCDUFFIE* RC 1 A.CONLIN R I SPIEWAK® R W. B. MCDONALD R 5. E. BEALL %. H. Coox M F. F. BLANKENSHIP* RC C. A. BRANDON R D. SCOTT, JR. R C. K. MCGLOTHL AN R R. G. GILLILAND M F. A. DOSS RC B. H. WEBSTER R W. ). LEONARD M 5. 5 KIRSLIS RC R.C. SCHULZE M B. 5. WEAVER RC C. H. WODTKE M W. K. R. FINNELL RC 3. L. GRIFFITH M ¥ JENNINGS RC M. 4. REDDEN M R. G. WILEY RC ‘DUAL CAPACITY CT CHEMICAL TECHNOLOGY DIVISION EBM ENGINEERING AND MECHANICAL DiVISION IBC INSTRUMENTATION AND CONTROLS DIVISION M METALLURGY DIVISION BUILDING AND FACILITIES COMPOMENTS SYSTEMS INSTRUMENTATION AND CONTROLS ANAL YIS RC REACTOR CHEMISTRY DIVISION R J. R. TALLACKSON 186 C. W. NESTOR R R REACTOR DIVISICN L. N, HOWELL EAM J. H. WESTSIK R L. F. PARSLY . R. A LW N. E. DUNWQODY EAM C. H. GABBARD R s. E. BOLT R R. L. MOORE 1&C D. W. YROOM. R B&R BURNS AND ROE C. E. BETTIS E&M 1. C. MOYERS* R P, H, HARLEY R J. R. BROWN 18C W. L. BREAZEALE L L LOCKHEED 0. T. GIBSON EAM K. R. PAYNE R R. 8. LINDAUER cT 0. W. BURKE Egm . H. FORD L D DIRECTOR'S DIVISION C. B. HAYNIE £8M P. G. SMITH R R. C. ROBERTSON R P. G. HERNDON 1&C y R, W. KNIGHT E&M W. C. ULRICH R A, N, SMITH* R 1. W, KREWSON I&C T) TECHNICAL INFORMATION DIVISION . 0. LEWIS E&M L. V. WILSON R R, £, RAMSEY* B&R B. LEVENSON L G. H. LLEWELLYN &M R. E. RAMSEY™ B&R A, J. MADAIO B&R 8. SQUIRES BAR T. E. NORTHUP E&M P. B ANAS BAR 5. A. SCHILKE BAR R. . BORGERS 18C G, W. RENFRD E&M H. L. BEAUVAIS B&R L. J. SCIRICA B&R B. 1. JONES 1&C W. E. SALLEE E&H C. A MILLS R L. J. ZOLA B&R C. D. WILLER B&R D. C. WATKIN E8M F. L. ROUSER R T. A WATSON EAM C. N.SRSICH B&AR W. D. HOYLE £8M J. W.XIDD E&M M. ROSENBAUM ERM F. J. UNDERWOOD E&M FLOW MODELS REMOTE MAINTENANCE FUEL HANDLING AND HEATERS LOOPS AND TANKS GAS SYSTEMS R. J. KEDL R E. C. HISE R J. €. MOYERS* R J. L. CROWLEY R AN, SMITH® R B. J. YOUNG R R. BLUMBERG R M. RICHARDSON R R. B. GALLAHER R P. P.HOLZ R H. 5. WEBER R R. E. CARNES R R. F. BENSON R G. A, WINDLE R W, H. DUCKWORTH R C. JONES R J. R. SHUGART R 137 \NolNooNENANCANIG B AR UUAR \O TN o QY =EpDrrrdlEEQrans Qg O0gHdgoEgsydoHdogEHOOOnepHHOHCOn O Adamson Beall Bettis Bettis Blizard Boch Bolt Borkowski Boyd Breeding Briggs Bruce Burke Campbell Cobb Conlin Cook Cristy Crowley Culler DeVan Douglas English Epler Ergen Fraas Frye, Jr, Gabbard Gall Gallaher Grimes Grindel Harrill Hill Hise Hoffman . Holz Hollaender HEEQwonaoaydhogo@mnodRYdp O epHEQQOEdWagEHGEHPP DR DN EHEQE Howell Jordan Jordan Kasten Kedl Keilholtz =g oom=Em INTERNATL DISTRIBUTION Alexander Billington Blankenship Householder G OHPROEPEQEOEHOOrOdN s Udbddbrs Q2RO E 2 0naQqg ol = ORNL-3215 UC-80 — Reactor Technology TID-4500 (l6th ed., Rev.) Kelley Kinyon Knight Lane Larson Lincoln Lind Lindauver Livingston Lundin MacPherson Manly Mann McDonald McGlothlan Miller Moore Morgan Moyers Murray (K-25) Nelson Nestor Northup Csborn Parsly Patriarca R. Payne Phillips B. Pike Richardson C. Robertson K. Roche W. Savage W. Savolainen Scott E. Seagren D. Shipley Sisman J. Skinner M, Slaughter N. Smith G. Smith H. Snell Spilewak D, Susanc A, Swartout Taboada R. Tallackson . - - - - - . - - - - - - - . - . - - . . - L] - L] . . - [ ] . - - - » L] - - - HoHsnmaNnmperORdoyaoaHNE QPR = =43 140 97. E. H., Taylor 107. Biology Library 98. R. E. Thoma 108-109, Reactor Division Library 99. D. B. Trauger 110-113. ORNL - Y-12 Technical Library, 100, W, C. Ulrich Document Reference Section 101, D. C. Watkin 114-158. Laboratory Records Department 102. A, M. Weinberg 159, Laboratory Records, ORNL R.C. 103. J. H. Westsik 160-162. Central Research Library 104. L. V. Wilson 105. C. E, Winters 106. C. H., Wodtke EXTERNAL DISTRIBUTION 163-164, D. F. Cope, AEC, CRO 165. R. A. Duval, AEC, Washington 166. T. W. McIntosh, AEC, Washington 167. R. W. McNamee, Manager, Research Administration, UCC, New York 168, Division of Research and Development, AEC, ORO 169-770, Given distribution as shown in TID-4500 (16th ed., Rev.) under Reactor Technology category (75 copies - OTS)