g 3 " BN g Jree &~ l'_‘-: I AFEA fpraer] :Ffitl a‘.“.fl Ewli*'.“ Tt ey ..':1} Ei}:.-:]:.m 15'-'1-.-[“}1 l.o'-t:i.} i:fl-..?i‘ ".m'nh:'.l T 1R ORNL-1864 Progress /44 FA " fi.'k"-r. ¢ i (@, | ‘&'ea. AIRCRAFT NUCLEAR PROPULSION PROJECT QUARTERLY PROGRESS REPORT -l 3 N FOR PERIOD ENDING MARCH 10, 1955 mgsmmnun Cru NGED T,-. | i by - — g = — | BY AUTHOR 'r ov... LEC f, /&‘éd N ot i J_ =X iy B!_._.... -.,_----._---_-_G../?l..éé-— -4'.‘) —'.“ Sery F -‘ o = N 3 445k 02509968 9 OAK RIDGE NATIONAL LABORATORY OPERATED BY CARBIDE AND CARBON CHEMICALS COMPANY 5" A DIVISION OF UNION CARBIDE AND CARBON CORPORATION (T3 POST OFFICE BOX P OAK RIDGE. TENNESSEE ORNL.-1864 This document consists of 205 pages. Capy//? of 207 copies. Series A, Contract No, W+7405-eng-26 AIRCRAFT NUCLEAR PROPULSION PROJECT Y QUARTERLY PROGRESS REPORT For Period Ending March 10, 1955 W. H. Jordan, Director S. J. Cromer, Co-Director R. 1. Strough, Associate Director A. J. Miller, Assistant Director A. W. Savolainen, Editor DATE RECEIVED BY INFORMATION AND REPORTS DIVISION (MARCH 25, 1955) e (AR Post Office Box P 3 quI:; DESDCIE!B g Qak Ridge, Tennessee PN GA LN~ 0O CMMEOOEANCPPIEILPOMMANTEIOOAOON-UPNZIOMMUMONDO . Adamson . Affel . Baldock Barton . Bettis Biltington . Blankenship . Blizard . Boyd . Bredig . Bruce . Callihan . Cardweli Cathcart . Center (K-25) . Chapman . Charpie . Clewett . Clifford . Cottrell . Cowen omer Crouse . Culler . Emlet (K.25) . Ferguson . Fraas Frye . Furgerson . Grimes . Hoffman . Hollaender . S. Householder 7. Howe . W. Johnson H. Jordan . W, Keilholtz P. Keim . T. Kelley . Kertesz . M. King - MOA LT TMErwOQooMIPAMSIETOPMOUMPLNE DO . A, Lane . E. Larson "l‘|.‘ ORNL-1864 Progress INTERNAL DISTRIBUTION 44, 45. 46, 47. 48, 49, 50. 51. 52. 53. 54. 53. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68, 69. 70. 71, 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83-92. 93-112, 113. 114-116. 4 . LaVerne . Livingston . Lyon . Maienschein . Manly . Mann . MeDonald . McQuitken . Miller . Morgan . Murphy . Murray (Y-12) . Nessle . Oliver . Patriarca . F. Poppendiek M. Reyling W. Savage W. Savelainen . D. Shipley Sisman P. Smith . H. Snell . |. Strough . Susano . Swartout . Taylor Trice . Yan Artsdalen . YonderL.age Warde . Weinberg White . Whitman . Wigner (consultant) . Williams Wilson . Winters . Zerby X-10 Document Reference L ibrary (Y-12) Laboratory Records Department |_aboratory Records, ORNL R.C. Central Research L ibrary WUV NCLERP>»POOZYM NOE-OMOE=PE-TMEMENADP>OOMP> LVITVTOAOSMAP>PTMIEICENDIZ omOONwoOzTNnaPIrog . e ‘; X S, s ) o "f‘:&r“:‘fi% o ek 117. 118-119. 120. 121. 122.132. 133-137. 138. 139. 140. 141.144, 145, 146-149, 150-151. 152-153. 154. 155. 156-157. 158, 159. 160. 161.170, 171, 172. 173-178. 179-190. 191-205. 206. 207. EXTERNAL DISTRIBUTION Air Force Plant Representative, Burbank Air Force Plant Representative, Seattle Air Force Plant Representative, Wood-Ridge ANP Project Office, Fort Worth Argonne National Laboratory Atomic Energy Commission, Washington (L.t. Col. T, A, Redfield) Bureau of Aeronautics (Grant) Chief of Naval Research Convair, San Diego (C, H. Helms) General Electric Company, ANPD Glen L. Martin Company (T. F. Nagey) Knolls Atomic Power Laboratory Lockland Area Office Los Alamos Scientific Laboratory Materials Laboratory (WADC) (Col. P, L. Hill) National Advisory Committee for Aeronautics, Cleveland North American Aviation, Inc, Nuclear Development Associates, Inc. Patent Branch, Washington Powerplant Laboratory (WADC) (A, M. Nelson) Pratt & Whitney Aircraft Division (Fox Project) USAF Project Rand USAF Headquarters Westinghouse Electric Corporation (Bettis Laboratories) Wright Air Development Center (Lt. John F. Wett, Jr,, WCOSI-3) Technical Information Service, Qak Ridge Division of Research and Medicine, AEC, ORO Atomic Energy Commission — East Hartford Area s Reports previously issued in this series are as follows: ORNL-528 ORNL-629 ORNL-768 ORNL-858 ORNL-919 ANP-60 ANP-.65 ORNL-1154 ORNL-1170 ORNL-1227 ORNL.-1294 ORNL-1375 ORNL.1439 ORNL-1515 ORNL-1556 ORNL-1609 ORNL-1649 ORNL-1692 ORNL-1729 ORNL-1771 ORNL-1816 Period Ending November 30, 1949 Period Ending February 28, 1950 Period Ending May 31, 1950 Period Ending August 31, 1950 Period Ending December 10, 1950 Period Ending March 10, 1951 Period Ending June 10, 1951 Period Ending September 10, 1951 Period Ending December 10, 1951 Period Ending March 10, 1952 Period Ending June 10, 1952 Period Ending September 10, 1952 Period Ending December 10, 1952 Period Ending March 10, 1953 Period Ending June 10, 1953 Period Ending September 10, 1953 Period Ending December 10, 1953 Period Ending March 10, 1954 Period Ending June 10, 1954 Period Ending September 10, 1954 Period Ending December 10, 1954 FOREWORD This quarterly progress report of the Aircraft Nuclear Propulsion Project at ORNL re- cords the technical progress of the research on circulating-fuel reactors and all other ANP research at the Laboratory under its Contract W-7405-eng-26. The report is divided into three major parts: |. Reactor Theory, Component Development, and Construction, il. Materials Research, and [ll. Shielding Research. The ANP Project is comprised of about 400 technical and scientific personnel engaged in many phases of research directed toward the achievement of nuclear propulsion of air- craft. A considerable portion of this research is performed in support of the work of other organizations participating in the national ANP effort. However, the buik of the ANP research at ORNL is directed toward the development of a circulating-fuel type of reactor. The effort on circulating-fuel reactors was, until recently, centered upon the Aircraft Reactor Experiment. This experiment has now been completed, and the analyses of the results of the operating experience are presented in Section 1 of Part |, The design, construction, and operation of the Aircraft Reactor Test (ART), with the cooperation of the Pratt & Whitney Aircraft Division, are now the specific long-range objectives. The ART is to be a power plant system that will include a 60-Mw circulating- fuel reflector-moderated reactor and adequate means for heat disposal. Operation of the system will be for the purpose of determining the feasibility, and the problems associated with the design, construction, and operation, of a high-power, circulating-fuel, reflector- moderated aircraft reactor system. The design work, as well as the supporting research on materials and problems peculiar to the ART (previously included in the subject sections), is now reported as a subsection of Part |, Section 2, ‘“‘Reflector-Moderated Reactor.”’ FOREWORD ..o SUMMARY oo CONTENTS ........................................................................................................................ ........................................................................................................................ PART I. REACTOR THECRY, COMPONENT DEVELOPMENT, AND CONSTRUCTION 1. CIRCULATING-FUEL AIRCRAFT REACTOR EXPERIMENT ..o Analyses of the Aircraft Reactor Experiment ..., Dismantling of the ARE ...... ........................................................................................................................ Fission Product Investigaiions ..ottt e Xenon Poisoning of the ARE ... Neutron energy distribution ... Xe137 cross section inm the ARE ...ttt 2. REFLECTOR-MODERATED REACTOR oo Reactor Design .......ccccocee. Reactor Physics ................ ........................................................................................................................ ........................................................................................................................ Activity of ART components after shutdown ... Gamma-ray heating........... ........................................................................................................................ Control FOd CONSIAOIIi OIS )\p = 5 x 10-4 sec-l This value is consistent with the observed be- havior of Cs'37 also. From the $r®? data reported above it appears that krypton isotopes have larger valyues of A_ than do xenon isotopes. Several lines of investigation are suggested by the results reported here. In particular, the effects of ruthenium on the physical properties of Inconel and on corrosion by the fluoride fuels should be studied. If all Ru'®® is removed from the fuel by the walls, and if the ruthenium “‘plate’’ is uniform over the entire reactor, the approximate rate of deposition of Ru'®® is 0.7 (P/A) p/hr, where P is the total reactor power in Mw and A is the surface ared in cm2. XENON POISONING OF THE ARE W. K. Ergen Aircraft Reactor Engineering Division Neutron Energy Distribution R. R. Coveyou R. K. Osborn Aircraft Reactor Engineering Division R. R. Bate United States Air Force If neutrons slow down in an infinite moderator of small absorption cross sections, their equilibrium velocity distribution will be nearly Maxweilian and will correspond to the moderator temperature. |f IOM. T. Robinson, Release of Xenon from Fluoride Fuels: Proposal for an Experimental Program, ORNL CF-54-6-4 {June 2, 1954). 15 ANP PROJECT PROGRESS REPORT the moderator has an appreciable absorption, the velocity distribution will be such as to favor higher energies because some of the neutrons will be absorbed before they reach thermal equilibrium with the moderator. This phenomenon has been investigated quantitatively by using the Monte Carlo method (flrst used for a similar purpose by G. F. von Dardel'") on the Oracle. The scattering cross section was assumed to be constant and the absorption cross section ~1/v. The higher ab- sorption at low velocities further emphasizes the shift of the neutron spectrum to higher energies. It appears that the neutron velocity distribution is well represented by a Maxwellian distribution, corresponding to an ‘“‘effective temperature,’’ T, if 22 1 = < 0.06 , 1) “ . where the macroscopic absorption cross section E is measyred at the moderator temperature, T, and % _ is the macroscopic scattering cross section. The effective temperature is (2) T, =T, (1 + aAK) where a2 is a constant, approximately equal to 0.9, and A is the atomic weight of the moderator. Corre- lation of this work with the results obtained by Y16, F. v. Dardel, Phys. Rev. 94, 1272 (1954). 16 Pratt & Whitney Aircraft'? John'? is in progress. Xe'33 Cross Section in the ARE W. K. Ergen H. W. Bertini Aircraft Reactor Engineering Division and by Brown and St. By using the results given above and computing 2 ond %_ by homogeneously distributing the constituents of the ARE core, the ARE effective neufron femperature was found to be 1.43 times the moderator temperature, both temperatures being measured on the absolute scale. Since the moder- ator temperature was 1400°F or 1033°%K, the effec- tive neutron temperature was 1477°K or 2200°F. If the Maxwellion distribution corresponding to this temperature is convoluted with the xenon cross section, an effective xenon absorption cross section of 1.37 x 10% barns is obtained, as com- pared with 1.76 x 10° barns at 1400°F and about twice that much at room temperature. Even by using the reduced cross section, it is seen that the xenon poisoning of the ARE would have been much larger than the upper limit of the observed value if the xenon had not been removed by the off-gas system. ]2P ratt & Whitney Aircraft, Nuclear Propulsion Program Engineering Progress Report No, 14, Oct. 1, 1954—-Dec. 3] 1954 PWAC-544. H D. Brown and D. S. St. John, Neutron Energy Spectrum in D 50, DP-33 (Feb. 1954). PERIOD ENDING MARCH 10, 1955 2, REFLECTOR-MODERATED REACTOR A. P, Fraas W. K. Ergen Aircraft Reactor Engineering Division REACTOR DESIGN A. P, Fraas Aircraft Reactor Engineering Division The preliminary design of the Aircraft Reactor Test (ART) installation has been completed and a careful examination of the hazards associated with reactor operation has been made. The instal- lation proposed makes use of a sealed reactor cell installed in an extension of the present ARE building, The hazards analysis disclosed that no nuclear explosion could occur that would damage the reactor cell and that the cell was more than adequate to contain the worst conceivable accident that might be experienced with the installation. The cell was described in the previous quarterly report, ! together with the other major components of the proposed installation. The reactor hazards report has been completed and a presentation made to the Reactor Safeguards Committee. Work is now beginning on the detailed design of the facility. Work on the reactor layout is proceeding. A half-scale model of the reactor has been completed to disclose the major problems associated with the assembly operation. The pump bearing and seal layout has been modified to incorporate a number of madifications to facilitate fabrication and assembly., A number of possible heat ex- changer header detail designs are being investi- gated from both the fabricability and the stress analysis standpoints. Preliminary load deflection curves have been obtained on a model of the first of these systems. A model of the most promising of the others is under construction. A detailed stress analysis for the pressure shell is being carried out, with particulor attention being given to the complex pump and header tank region at the top. The problem is greatly com- plicated by the need for evaluating the gamma heating of the Inconel and the attendant unusual temperature distribution, Much has been accomplished in the past quarter on the detailed design and fabrication of component test units., A full-scale model of the fuel pump ]A. P. Fraas and F. R. McQuilkin, ANP Quar. Prog. Rep., Dec, 10, 1954, ORNL-1816, p 29. has been completed and is nearly ready for testing. This unit will be used for determination of the performance characteristics of the impeller, in- cluding the cavitation limit, Four sets of core shells are being fabricated by Pratt & Whitney Aircraft for dimensional stability tests at temper- ature and for fabrication investigations. Two test rigs for investigating the flow characteristics of full-scale cores have been fabricated and set-up work is nearly complete. An intensive test pro- gram is planned for these units. A pump—expansion tank configuration designed to remove xenon has been evolved which performs well hydrodynamically; however, power required for its operation is higher than is considered desirable. A careful re- examination of each of the elements in this system is being made so that a revised arrangement can be designed and tested during the coming quarter. The performance of this unit will be evaluated analytically by using the data obtained on xenon removal in the ARE and information from the chemistry and physics programs, Several heat exchanger test rigs are being completed and should be ready for testing by April. One of these will give information on the heat transfer performance of this unusual con- figuration with water, A second is designed to yield performance data with NaK and the fluoride mixture. The former rig is sufficiently flexible so that various tube-spacing arrangements can be employed (cf., sec. 8, ‘“‘Heat Transfer and Physical Properties’’). This will, of course, not be pos« sible with the welded-up test unit for operation with fluoride mixture and NaK. Design-study models have been built to show the preliminary layouts for both the reactor assembly and the facility, The model of the reactor, pump, heat exchanger, and pressure shell assembly is shown in Fig. 2.1. A 12.in. scale was placed ot the bottom of the half-scale model to give an indication of the key dimensions. (The full-scale pressure shell will be approximately 54 in, in diameter.) The NaK outlet pipes project axially from the lower part of the pressure shell, and the NaK inlet pipes enter the upper part of the pressure shell radially. The casings for the 17 ANP PROJECT PROGRESS REPORT PHOTO 23368 SODIUM PUMPS FUEL PUMPS Nak INLET PIPES ) PRESSURE SHELL * NaK QUTLET PIPES Fig. 2.1. Preliminary Model of ART Reactor, Pump, Heat Exchanger, and Pressure Shell Assembly. 18 two fuel and two sodium pumps project vertically upward from the ‘‘north head'’ of the reactor. The fuel pump at the left is sectioned to show the bearings and seal in the removable pump body in the upper portion of the pump well. The pump and expansion tank region is shown in more detail in . SODIUM PUMP FUEL PUMPS | 'SODIUM PUMP VOLUTE PERIOD ENDING MARCH 10, 1955 Fig. 2.2, for which one of the sodium pump casings was removed. The cylindrical fuel expansion tank is at the center between the two fuel pumps, while the sodium pump velute can be seen in the foreground. A section cut through this region just below the top plate, or ‘“‘upper pump deck,”’ pHOTO THE70 Fig. 2.2. Detailed Yiew of ART Pump and Expansion Tank Region. 19 ANP PROJECT PROGRESS REPORT is shown in Fig. 2.3, The fuel pump volutes are arranged to discharge tangentially into the core inlet at the center, The sodium-to-NaK heat ex- changer tube bundles are along the upper and lower quadrants at the periphery. Sodium returns upward from the reflector through the circular openings in the ‘“lower pump deck’’ between the core inlet and the sodium-to-NaK heat exchangers. The mss ¥ rising through the return opening in the foreground flows to the leff, enters the sodium heat exchanger near the fuel pump volute, passes to the right through the heat exchanger, flows back to the left around the right end of the baffle enclosing the heat exchanger, and rises into the sodium pump inlet (see Fig. 2.2). The NaK inlet pipes to the sodium-to-NaK heat exchangers project radially from the assembly in the lower right and upper left in Fig. 2.3, The NaK outlet NaK INLET PIPE o FUEL PUMP VOLUTES "NaK OUTLET PIPE SODIUM = TO ~NaK HEAT EXCHANGER pipes are at the lower left and upper right, A section through the island, core, reflector, and heat exchanger is shown in Fig. 2.4, The control rod can be seen at the center of the island, The rifle-drilled holes for the cooling passages through the reflector and island can be seen in the beryllium regions, The preliminary l/iz-scule model of the reactor test facility is shown in Figs., 2.5 and 2.6, The full-scale cell will be 24 ft in diameter. The 10- ft-dia reactor and shield assembly can be seen to the right of the center of the cell in Fig. 2.5. The NaK outlet pipes pass from the lower portion of the shield, through a bulkhead in the cell wall, and then upward to the radiator cores. The NaK leaving the radiators rises to the four NaK pumps at the top, from which it is returned to the upper portion of the reactor. Four axial-flow blowers ..... N PHOTO 23372 ° NakK QUTLET PIPE e NoK INLET PIPE Fig. 2.3. Section Through Fuel Pump Volutes and Sodium-to-NaK Heat Exchanger Region. 20 PERIOD ENDING MARCH 10, 1955 ey PHOTO 23369 SODIUM PUMPS FUELL PUMPS EXPANSION TANK T B SODIUM -TO~NaK HEAT EXCHANGER § BERYL.LIUM REFLECTOR- MODERATOR PRESSURE SHELL NEEI R & | O CORE FUEL ANNULUS CONTROL ROD == e SODIUM COOLING PASSAGES BERYLLIUM iSLAND FUEL~TO-NgK HEAT EXCHANGER REGION { HEAT EXCHANGER NOT SHOWN ) Fig. 2.4. Section Through the Core, Reflector, and Island of the ART. 21 ANP PROJECT PROGRESS REPORT i GROUND -FLOCR LEVEL Nq_K PUMP DR?VE‘ MOTORS REACTOR CELL i _ NaK EXPANSION TANK ; NaK __PUMP COOLING BLOWERS % B ok OUTLET PIPES \ REACTOR AND SHIELD ASSEMBLY Fig. 2.5. Preliminary Scale Model of ART Facility Showing Radiator Cores in Foreground. 22 £ MOTORS | G S MaK EXPANSION Fig. 2.6. Preliminary Scale Model of ART Facility Showing Stack in Foreground. BN il PHOTQ 23366 REACTOR CELL SS61 ‘01 HOMYW ONIGONF a0l¥3d ANP PROJECT PROGRESS REPORT at the right force air through the radiators and out the stack at the rear. The black line around the upper part of the outer tank represents the ground-floor level. The NaK drain tanks can be seen in Fig. 2.6 to the left of the lower part of the reactor cell. Note that four separate NaK systems are used, each with its own pump, expansion tank, dump tank, ete. The NaK expansion tanks are just below and to the left of their respective pump drive motors in Fig. 2.6. REACTOR PHYSICS W. K. Ergen Aircraft Reactor Engineering Division Activity of ART Components After Shutdown H. W, Bertini Aircrtaft Reactor Engineering Division The octivities of the main components of the ART as a function of time after shutdown are 109 ORNL_L,R-D,‘,N,G 5833 5 2 108 5 2 - @ 5 0o & g E 5 c = = o £, > = 5 "\ 1000 hr < 106 . . _ 5 2 . {00 hr 5 10 5 » ] PREDOMINANT ACTMITY ngs, NS ——— 4 10 0 50 100 150 200 250 300 TIME AFTER SHUTDOWN (days) Fig. 2.7. Activity of the ART Fuel After Opera- tion at 60 Mw for 100 and 1000 hr. 24 given in Figs. 2.7 through 2,11. The activities are given in gamma curies for 100 and 1000 hr of operation at 60 Mw. Curtiss-Wright multigroup 2,3 were used as a basis for the calculations, The detailed work? was done For 100 hr of operation, and the curves for ]OOOéhr were obtained by multiplying the approximately flat portions of the 100-hr curves by 10, This method is valid for all curves except the fuel curve, Fig. 2.7. Comparison with GE-ANP data® indicates that the 1000-hr fuel-activity curve may be high by a factor of 2 or 3, calculations 2H. Reese, Jr., et al., Geomet?l Study for an ANP Circulating- Fuel Reactor, WAD-1901 (Sept. 1, 1954). Personal communication with S. Strauch, Curtiss- Wright Corporation. 4H. W. Bertini, unpublished memorandum. 3. Motett, Miscellaneous Data for Shielding Calcu- lations, p 66, APEX-176 (Dec. 1, 1954). ORNL-LR-DWG 5834 INDICATES PREDOMINANT CACTIVITY ACTIVITY {gamma curies} 1000 hr cB0cr® ]z cef0 o 5¢ 100 150 200 250 300 TIME AFTER SHUTDOWN (days) Fig. 2.8. Activity of the ART Sodium Coolant After Operation at 60 Mw for 100 and 1000 hr. T ORNL—LR-DWG 5835 INDICATES PREDOMINANT ACTIVITY ACTIVITY (gamma curies ) 41000 hr 100 hr 0 50 100 4506 200 250 300 350 TIME AFTER SHUTDOWN (days) Fig. 2.9. Activity of the ART NaK Coolant After Operation at 60 Mw for 100 and 1000 hr. For extrapolation purgo¥es, the following rules may be considered to give approximate activities, For times after shutdown greater than four days, the height of the fuel curve can be considered to be linear with operating time up to 1000 hr., The 1000-hr curve is approximately the curve for infinite operating time, The heights of the approximately flat portions of the other curves can be considered to be linear with time for operating times up to 1 year. All portions of all the curves are directly proportional to the reactor power, The activities of the B'? layer around the re- flector and of the lead shield were not examined in detail, but estimates of their activities follow. The activity of the B'® layer would be due mainly to the sodium impurity.® This activity would be about 5 x 10~3 of the activity due to the sodium impurity in beryllium, The value for beryllium is 6Spectrographic report on sample 481(a) of H3E503 gives sodium present as 0.31 wt % in boron. PERIOD ENDING MARCH 10, 1955 ORNL-LR-DWG 5836 INDICATES PREISOMINANT -AKCTIVITY — _ f_fl\_lCONEL_CORE SHELLS {1000 hr) ACTIVITY (gamma curies) —_ —— —_— e — — NEL CORE SHELLS {100 fir) oS0 M MODERATOR {100 hr) CoGO T 0 50 100 150 200 250 300 TIME AFTER SHUTDOWN {days) Fig. 2.10. Activity of the Beryllium and the Inconel Core Shells After Operation at 60 Mw for 100 and 1000 hr. given in Fig. 2.10. The maximum activity of the lead shield would be about 30 curies at shutdown, At about two days, and beyond, the activity would be a fraction of a curie. The activities at time zero, that is, at shutdown, are intended to be maximum possible activities, The inclusion of this point in the curves has introduced slight distortions in the interval from one to five days after shutdown, I|f more accurate activities are required in this interval, the values are available in tabular form.4 Gamma-Ray Heating L. T. Anderson, Consultant Calculations have been made to get better estimates of the gamma heating in the fuel pump and expansion chamber regions of the ART., The gemma enetgy flow at the axial-surface points of a seties of circular cylinders of delayed-gamma- 25 ANP PROJECT PROGRESS REPORT 4”*-32’*‘ QRNL-LR-DWG 5837 10 10° 5 2 7 2 § 10 o o E 5 £ o e > £ 2 s = g o I NCONEL {00 hr) > _ HEAT EXCHANGER TUBES 1000 2 e — INCONEL PRESSURE o R ESSuRe (G0 hr 5 : FEAT EXCHANGER TUBES (100 hr) 2 107! 0 50 100 {50 200 250 300 350 TIME AFTER SHUTDOWN (days) Fig. 2.11. Activity of the Inconel Heat Ex- changer Tubes and Pressure Shell After Operation at 60 Mw for 100 and 1000 hr. emitting, fpel was computed. A gamma interaction coefficient of 0.13 em=! and a gamma activity of 10 watts/cm? were assigned to the fuel. Buildup was not taken into account. Figure 2,12 presents the results of these computations, The radidl dependence of gamma energy flow was also com- puted for a cylinder of 15 ecm height and diameter and is shown in Fig. 2.13. Control Rod Considerations W. K. Ergen H. W. Bertini Aircraft Reactor Engineering Division The burnup in a control rod can be computed if it is assumed that a rod, worth Ak, absorbs Ak neutrons per fission. The number of grams burned up during ¢ hours of operation at P megawatts is [A%k (atoms destroyed/fission) % 3.1 % 1019 (fissions/w-sec) x P x 108 (w) x 3600 # {sec) x M {g/g-atom)] /10.602 x 1024 (atoms/g-atom)] = 1.85x 10=4 Ak PtM (g) . For Ak = 5%, P = 60, ¢+ = 1000, and the atomic weight M = 150 (corresponding to the rare earth region), this amounts to 83g. For a density of 7, representative of a rare earth metal, this would correspond to 12 ecm3. A rod 7/16 in. in diometer ORNL-LR-DWG 5838 A=15.24 cm - A=12.70cem h=1016cm- v - ENERGY FLOW {w/cm?. sec) CYLINDER RADIUS (cm) Fig. 2.12. Emitting Fuel vs Cylinder Radius and Height. 26 Gomma Energy Flow at Axial-Surface Points of Circular Cylinders of Delayed-Gamma- ane— ORNL-LR-DWG 5839 ENERGY FLOW (w/cm2-sec) 0 1 2 3 4 5 & 7 8 DISTANCE FROM AXIS (cm) Fig. 2.13. Gamma Energy Flow at Plane-Surface Points of 15-cm Square Cylinder of Delayed- Gamma-Emitting Fuel vs Distance from Cylinder Axis. 3 and a rod and 12 in. long would have 30 cm 1 in. in diameter and 20 in. long would have 258 e¢cm3. Thus, if all the rod consisted of ‘‘burn- able’’ material, the smaller rod would suffer con- siderable burnup, but the larger one would survive with little change. In considering the use of PERIOD ENDING MARCH 10, 1955 samarium for the rod it was found that only 14% of the element is made up of the large cross section isotope; therefore 12/0.14, or 86 cm?, would burn out, which is more than the total volume of the small rod and aconsiderable fraction of even the large rod. Furthermore, if the easily obtainable oxide were used instead of the metal and if a binder were used to give the rod me- chanical rigidity, even the large rod would become of marginal usefulness. From the heating viewpoint, a power density of 20 w/cm® and a thermal conductivity of the rare earth oxide of 0.003 cal/sec.cm:°C would yield a temperature difference of 1200°F between the center of a l-in. rod and the surface. Although both the thermal conductivity and the power density are considerably in doubt, the heating problem appears to make impossible the use of a solid oxide rod 1 in. in diameter. The use of a hollow rod might present problems because of burnup unless an element such as europium, which is 100% large absorption cross section isotopes, were used, 27 ANP PROJECT PROGRESS REPORT 3. EXPERIMENTAL REACTOR-ENGINEERING M. W. Savage E. S. Bettis Aircraft Reactor Engineering Division Design work on the in-pile loop proceeded as scheduled. A flux-measuring loop for obtaining information on the flux to be expected is being fabricated, and 1000 hr of trouble-free operation of the horizontal-shaft sump pump was concluded. A double-walled heat exchanger was found to be acceptable for in-pile use. Performance tests on the ARE-type sump pump with fluoride fuel were completed, and tests with water and with hot NaK (1400°F) indicated the suitability of this type of pump for the heat ex- changer test loop. Mechanical shokedown tests of ART pump rotary elements are under way. Three high-velocity, high-temperature-ditferential loop tests of corrosion and mass transfer in fused- salt-Inconel systems were completed, as well as a fourth test of sodium in a beryllium-Inconel system. The test loop for the intermediate heat exchanger test No. 2 is being fabricated, and work is under way on a small heat exchanger loop for testing a bundle of 20 tubes. Tests were made of the integrity of the cell for containing the ART, IN-PILE LOOP COMPONENT DEVELOPMENT D. B, Trauger Aircraft Reactor Engineering Division The in-pile loop design specifications were revised to make the first test a more conservative one. The power density was reduced from 2.5 kw/cm?® to approximately 1 kw/em?® and the temperature differential from 300 to 200°F; the maximum temperature remains 1500°F. An extra turn of tubing in the ‘‘nose’ section, approxi- mately 2 ft of developed length, is necessary to meet the temperature differential specification at this power density. The total power is expected to be about 22 kw. The fuel mixture NaF-ZrF - UF4 (53.5-40-6.5 mole %) will be circulated at « Reynolds number of 5000. These conditions, with the exception of the temperature differential, are at least as severe as those expected in the ART, The design of the loop has progressed through the layout stage and is now ready for detailing. Figure 3.1 shows the present basic design. A full-scale wooden model, shown in Fig. 3.2, was 28 prepared to aid in solving the difficult assembly problems and for instructional purposes. The numbers on the base indicate distance, in inches, from the reactor lattice face. The calrod pre- heaters and the fill tank had not been installed when the photograph wos made. The water jacket, which will completely encase the loop forward of the rear header, is represented, in part, by the plastic tube. Instrumentation R. A, Affel P. A. Gnadt Aircraft Reactor Engineering Division Instrumentation design for the in-pile loop is nearly complete, and fabrication and testing of most components are now under way, A simplified diagram for the instrumentation and control system is shown in Fig. 3.3. Since the power generated in the loop is dependent on the reactor flux, it is subject to sudden and unexpected changes. These sudden and unexpected changes would come about as a result of power failures, experimental troubles, etc. The main control therefore is by regulation of the cooling air in response to a thermocouple signal at the exit end of the heat exchanger. This signal will also energize the preheat calrod circuits if danger of freezing exists. The pump is driven by a Vickers hydraulic oil motor, which, in turn, is activated by a pump unit. Manual speed control will be provided by varying the output of the drive unit. An electromagnetic tachometer will operate from a toothed gear on the pump shaft te indicate and record the pump speed. Activity monitors on all fluid lines leaving the loop will be connected to alarms or reactor scrams, depending upon the potential hazard involved. Melt-down Hazard C. W. Cunningham Aircraft Reactor Engineering Division The principle hazard to the experiment would be a pump failure or other sudden stoppage of flow. In the event of a sudden flow stoppage, the fuel temperature would rise rapidly, and the fuel tube would be melted in a few seconds. The 6L FACE OF LATTICE . ¥ - i . ‘/F\LLE u’ pil L . - WATER OUT T WATER IN Tt~ —-THERMOCOUPLE LEADS e .- OUTER WATER JACKET e .~ INNER WATER JACKET 7 i ~-— — HYDRAULIC OIL OUT //’ el e -~ HYDRAULIC OIL IN ) 7 . -~ L HYDRAULIC MOTOR /J‘/ - e g . - -BEARING HOUSING . P o ,,‘6‘"‘GRF‘P " ““-THERMOGOUPLE LEADS il ek r“"LLfiD . -~ PUMP ! ! . “FILL LINE ““FUEL LINE “VENT UNE " FILL TANK S WATER JACKET OUTLINE T HEAT EXCHANGER T SNIFFER LINE - NOSE LOOP Fig. 3.1. Loop for Circulating Fluoride Fuel in MTR Horizontal Beam Hole. ORNL-LR-DWG 5663 o - 20 T AR OUT T T AIR IN SS61 ‘0L HOYVW ONIGNIT do1y¥3d ANP PROJECT PROGRESS REPORT BEARING HOUSING R . UNCLASSIFIED ACKET | . PHOTO 23005 8-ft BARYTES CONCRETE SHIELD - 7 T i STEEL SHOT SHIELD Fig. 3.2, Model of In-Pile Loop. ORNL Mathematics Panel has investigated this problem with the use of the Oracle. A predicted chronology of events follows: At 0.00 sec the flow is assumed to stop and the salt temperature is rising at 70°F per 0.10 sec. After 1.20 sec a thermocouple on the outside of the tube will have indicated a 100°F rise in temperature, which is considered to be adequate to provide a reactor scram signal. At 1.45 sec the salt at the tube center is at its boiling point, 2430°F. The center of the Inconel wall, however, will have reached only 1650°F and will have sufficient strength to resist considerable pressure if the salt should superheat and suddenly boil, If the assumption is made that the salt temper- ature will continue to rise with no boiling, the chronology continues, as follows: At 4.2 sec the salt next to the inner wall will have reached 2500°F and the Inconel tube will begin to melt. The entire tube would melt in 7.5 sec. Calculations and circuit tests indicate that the reactor can be shut down, on the basis of thermo- 30 couple or tachometer signals, in time to prevent melting of the nose tube. Although it is unlikely that a scram can be effected from a thermocouple signal in time to keep the tube below temperatures where crystal grain growth will become very rapid, a signal from the tachometer should be fast enough to avoid this type of damage to the loop. If the scram fails, it seems likely that boiling will oceur, aleng with subsequent displacement of the salt into the pump sump and fill tank. This displace- ment should result in additional time for secondary protective circuits to function and shut down the reactor. If a melt-down or a sudden leak should occur, it would be somewhat hazardous for molten salt to reach the stainless steel water jacket wall., A molybdenum shield which will allow no line of sight to the jacket has been provided as protection in this event. Molybdenum has been demonstrated to resist corrosion by the salt at boiling temper- atures for 30 min or longer. . * r T § € UNCLASSIFIED ORNL-LR-DWG 5670 VAB T v19,20,21,22,23,24 » F Ve FILTER FILTE v-36 INSTRUMENT LTE ILTER PLANT AIR SUPPLY AIR SUPPLY NITROGEN V-26 " Nl UNTREATED WATER DEMINERALIZED PROCESS A FOR EMERGENCY USE WATER SUPPLY ELECTRIC AR e MOTOR - COMPRESSOR V.33 v-32 v-34 N e L 2000 3. L] }_ W L e | ‘ o V=55 b= - ‘l TO STACK : ' Y “ 1 HELIUM | 1500 | & — — l AA DN A w TO STACK L= | T Y % _ e \ _l = e ' PUMP BEARING\ ‘ 1400 o« | e U LIQUID NITROGEN W < | T YTy 3 COOLED < ‘ HYDRAULIC « N MOTOR | 1300 OlL HOLD £ v-43 v-44 |4 - — BY-PASS OIL > { RETURN BLOWER N < | - - | ’ 4 TO RETURN LINE LEGEND : v-48 v-49 . — A TEMPERATURE {°F) D PRESSURE (psig) jlb-—J—— TO RETURN LINE TO RETU O FLOW (gpm) () FLOW (scth) / © LINE AN 70 Hy0 A30\ @ FLOW {scfm) 1/\ . l TO HG-5 l VoW VoW I AR HEAT BEAM HOLE A @ 125 EXCHANGER — MAIN LOOP ——— PROCESS PIPING L€ ———— MECHANICAL CONNECTION Fig. 3.3. Flow Diagram of In-Pile Loop. SS6L ‘0L HOYYW ONIONZ a0ld3d ANP PROJECT PROGRESS REPORT Fission-Gas Holdup D, W. Magnuson Aircraft Reactor Engineering Division Fission gases will be purged from the pump sump and from the pump bearing regions, with purging from the bearing region serving to remove leakage past the rotating face seal. These purge gases will pass through liguid-nitrogen-cooled charcoal traps and then through filters. The design of the fission-gas holdup system is based on adsorption data of the Linde Air Products Company.! The equations of Jury? were used to predict the performance of adsorption traps in this gas holdup system, and it was found that a 250-g activated- carbon trap ot ~170°C wos five times more ef- fective than required to give a reduction in molecular density of more than 1 x 108, This should constitute adequate decontamination. How- ever, radioactive decay will further increase the decontamination factor, and a second trap is provided, in series, as an additional safety factor. The use of an extremely small-pore metal filter will prevent particles from reaching the MTR stack. Flux-Measuring Loop D. M. Haines Pratt & Whitney Aircraft One of the greatest uncertainties that might affect the successful operation of this in-pile loop is in the estimation of the neutron flux, The thermal flux will be “‘depressed’’ by the materials of construction of the loop and by the fuel. Several estimates have been made and an analog computation was performed by Prott & Whitney Aircraft. These estimates indicate that the ef- fective flux will be about one half that of the undisturbed beam hole. However, such estimates are difficult to make, since the loop geometry is quite complicated and the unperturbed flux pattern is not well known. Accordingly, a flux-measuring loop is being fabricated for early irradiation at the MTR. This loop consists of the identical water jacket and other components of the forward end of the loop. The heat exchanger section will be . N Burdick, Adsorption of Krypton and Xenon, ORO-118 (Oct. 17, 1951). 25. H. Jury, Design of Percolators, ORNL CF-51-7-41 (July 9, 1951}, 32 mocked up with approximately the correct ma- terials. Cobalt foils placed inside and outside the fuel tube and in a graphite bar along the heat exchanger tubes will be counted after irradiation to determine the flux. An exposure time of less than 1 hr is planned. Since this loop will not measure the flux de- pression caused by the salt, a second and inde- pendent experiment (cf., sec. 9, ‘‘Radigtion Damage’’) is also being performed by the Solid State Division. Small tubes, 0.269 in. ID by 7 in. long, containing foil will be irradiated in the MTR “rabbit’’ facility. The tubes will be filled with various mixtures of material to match the effective cross section of the fuel mixture NaF-ZrF -UF, (53.5-40-6.5 mole %). Horizontal-Shaft Sump Pump J. A. Conlin Aircraft Reactor Engineering Division The first fused-salt test model of the in-pile pump, which was described previously,® has com- pleted a trouble-free 1000-hr endurance test. The pump circulated 1400°F NaF-ZrF ,-UF, (53.5- 40-6.5 mole %) at 1 gpm in an isothermal loop similar to the proposed in-pile loop. Immediately after filling and priming, the pump was operated at flows from 0.54 to 1.64 gpm and found to be quite satisfactory. The pump, seal, and drive motor are basically the same as those proposed for in-pile use. However, in the final in-pile pump design, the shaoft is cooled with helium rather than oil, and the seals and bearings are drop-lubricated. These changes were in- corporated to minimize the problem of oil disposal in the loop assembly. A prototype pump is now fully designed. The priming difficulty previously reported® has been solved by drilling a %z-in. vent hole between the impeller cavity and sump in the vicinity of the shaft. This priming difficulty was attributed to a seal caused by the surface tension of the water and the close radial clearance, 0.007 in., between the shaft and the pump casing. The seal was strong enough to prevent venting of gas from the pump and loop along the shaft and into the sump. A system has been proposed and tested with water which will permit automatic filling and pump- 3). A, Conlin, ANP Quar. Prog. Rep., Dec. 10, 1954, ORNL-1816, p 41. sump:lewaly,control. It consists of a fill tank connected to the pump sump by two tubes., One tube, for filling, connects the bottom of the tank to the bottom of the sump. A second tube, for venting, extends from the top of the fill tank to the pump. This latter tube enters the side of the pump sump and is directed downward into the sump, ending at the normal sump fluid operating level. In operation, the pump and fill tank are heated, with the fill line remaining frozen. The fill line is then heated and fuel flows from the tank into the sump and displaces gas to the fill tank through the vent tube. When the sump level reaches the vent tube, fuel is forced up the vent until it balances the head in the fill tank, at which time all flow stops. As the sump level lowers, the bottom of the vent tube is uncovered, the pressure balance is upset, and the sump again fills. In the test model it has been possible to control sump level to within ]/] in. The operation of this fill system is particularly sensitive to* vent-tube design. The tube must have o gradual downward slope from the fill tank to the sump, and, for best results, have a sharp-edged flared end in the sump. A unit is now being set up to test the system with fused salts. In operation the fill line will be frozen after transfer to eliminate the possibility of additional fuel entering the sump. Heat Exchanger o L. P. Carpenter Aircraft Reactor Engineering Division The problems of sealing and shielding a reliable, helium-recirculating cooling system for removal of the heat generated in the in-pile loop are so great that an air-cooling system appears to be the least expensive in time and money.? Air will flow from a compressor, through the loop heat ex- changer, and discharge into the MTR pebble zone through a second, little-used beam hole. A double- walled heat exchanger is planned in which a tube will be shrunk thermally around the salt-carrying tube of the loop, and the tubes will be brazed together at the ends. A third tube will form the annular air passage, as in the original exchanger. The double-walled tube is a safety feature in that in the event of a break of the salt tube within the 4D. F. Salmon and L. P. Carpenter, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 43. PERIOD ENDING MARCH 10, 1955 heat exchanger, the radioactivity will be contained within the second tube and will not enter the high- A helical groove cut in the inside wall of the second tube will direct any gas leakage from the salt tube to an activity monitor. Three tests of in-pile heat exchangers have been run in which lengths equivalent to one half the actual heat exchanger were used, Two exchangers were of the single-tube-and-shell type that was designed for helium cooling, and the third was a double-walled tube type. The results of the tests indicate that the required heat removal can be accomplished with the double-walled tube, al- though there is about a 20% loss in efficiency in comparison with the single tube. The double- walled heat exchanger removed 15 to 20 kw with 0.4 to 0.7 |Ib/sec of air and a temperature rise of 120 to 150°F. In the in-pile loop, the mass rate of air flow will be about the same as that used in the test, but the temperature rise at the exit will be about twice that in the test, and therefore heat-removal capacity will be about 30 to 40 kw. velocity air stream. PUMP DEVYELOPMENT ARE-Type Sump Pumps A. G. Grindell Aircraft Rzactor Engineering Division Performance tests of the ARE-type sump pump were concluded. In these tests the pump charac- teristics were obtained over the speed range 600 to 1400 rpm with the fluoride mixture NaF-ZrF - UF4 (53.5-40-6.5 mole %) at 1300°F;> the critical speed range was determined to be between 2850 and 3250 rpm; and the zirconium fluoride vapor trap was demonstrated to be effective. The re- liability test was terminated, despite continued trouble-free operation, at a total operating time of 3748 hr. The only interruptions in operation were the shutdown for inspection at 2000 hr, and a motor bearing failure at about 3000 hr (downtime, 21,5 hr). [t may be concluded from the tests that the ARE-type sump pump, when used in con- junction with a suitable vapor trap, may be expected to operate with NaF-ZrF4-UF4 {53.5- 40-6.5 mole %) for about 4000 hr without trouble at its design conditions: 1350°F, 1500 rpm, and 40 gpm. 5w, G. Cobb, A. G. Grindell, and W. R. Huntley, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, Fig. 3.3, p 44. 33 ANP PROJECT PROGRESS REPORT Tests of the ARE-type sump pump were also made to determine its suitability for delivering NaK at 1400°F, 184 gpm, 300-ft head in the ART intermediate heat exchanger tests. A preliminary water test at 100°F indicated that the desired conditions could be reached with the ARE-type pump at a speed of about 3800 rpm (Fig. 3.4). A hot test with NaK verified the preliminary test data (Fig. 3.5). No life test of the pump at the intermediate heat exchanger test conditions was performed, but previous experiences indicate that no serious trouble would be encountered during the 1000-hr duration of the test. However, the V-belt drive would be operating beyond its rated speed limit, and it would probably have to be replaced pericdically during the test. Mechanical Shakedown Tests of ART Pump (Model 1) Rotary Elements A. G. Grindell Aircraft Reactor Engineering Division W. C. Snapp Pratt & Whitney Aircraft Included in the program for development and testing of pumps for the ART circuits is a me- chanical shakedown test of each rotary element to be used. The objectives of the test are to determine adequate assembly techniques, the ap- propriate clearances, and the reliability of bearings and seals and their lubricating and cooling equip- ment. Particular emphasis is placed on assuring that seal parts have no manufacturing defects and that they have been properly assembled with respect to the pump shaft and impeller. No load is to be imposed on the impeller. A mechanical shakedown test stand has been erected, and two rotary elements have been tested. Both elements were run at a constant 3800 rpm with lubricating oil supplied at 2.0 to 2.5 gpm at 150°F, One rotary assembly (No. 1) was run for 96 hr without vibration, bearing heating, or seal leakage, and the test was shut down for disassembly and inspection of the rotary element. No damage was detected, The element was then reassembled and the test restarted. A slight amount of oil leakage past the lower seal was detected 22 hr after the operation was resumed, and the test was termi- nated at the end of 67 hr when the leakage rate became excessive. ORNL-LR-DWG 58Tt 400 3900 rpm 3600 rpm 300 e S e INTERMEDIATE HEAT EXCHANGER TEST DESIGN POINT 3300rpm ON T ; 3000rpm o 200 < w T 2700rpm 2400rpm 100 o 100 200 300 FLOW RATE (gpm) Fig. 3.4. Performance Characteristics of ARE- Type Sump Pump in Tests with Water at 100°F, ORNL-LR-DWG 5672 400 i 4000 rpm 36C0rpm 300 S g o =4 £ 200 Q o oy 100 e —— - _— - - —_——— ] 0 ;, — - J__——L o 50 {00 150 200 250 FLOW RATE {rpm) Fig. 3.5. Performance Characteristics of ARE- Type Sump Pump in Tests with NaK at 1400°F. Inspection of the seal parts revealed that dis- tortion of the bellows nosepiece had destroyed the surface flatness of the seal face. A study revealed that the material and hardness specifi- cations for the bellows nosepiece precluded the possibility of heat treatment for stress relief, and the &'Portion was therefore attributed to stress relief induced by test conditions. The bellows nosepiece specification was rewritten, and new seals were ordered. The results of a test of a second rotary as- sembly (No. 2) with seals manufactured to the original specification verified those obtained in the test of assembly No. 1. The total operating time was 200 hr. Both these rotary elements will be retested when seals manufactured to the revised specifications are received. DESIGN AND OPERATION OF FORCED- CIRCULATION CORROSION AND MASS TRANSFER TESTS Operation of Fused-Salt«lnconel Loops W. B. McDeoneld P. G. Smith Aircraft Reactor Engineering Division J. J. Milich R. A. Dreisbach Pratt & Whitney Aircraft Three fused-salt-Inconel forced-circulation loops with large temperature differentials were terminated after 774, 625, and 521 hr of operation, respectively. In each case operation was termi- nated short of the scheduled 1000 hr because a leak developed; however, the operating periods were long enough to provide valuable corrosion and mass transfer data. The fuel used was Nch-ZrF“-UF4 (53.5-40-6.5 mole %). The design of the loop was illustrated in the previous report.® The loop that operated 774 hr had a peak fuel temperature of 1500°F, a maximum tube wall temperature of 1740°F, a temperature differential of 200°F, and a Reynolds number of 10,000. The loop had to be terminated when a small leak developed near one of the heater terminals. The cause of this failure is not known. Because of a plant power failure and a pump motor circuit failure, the fuel had to be dumped twice during this test. The loop that operated 625 hr had a peak fuel temperature of 1500°F, a temperature differential 5, A Mann, W. B. McDonald, and W. C. Tunnell, ANP Quar. Prog. Rep. Dec, 10, 1954, ORNL-1816, Fig. 3.4, p 45. PERIOD ENDING MARCH 10, 1955 of 300°F, and a Reynolds number of 10,000. The maximum tube wall temperature in this test was 1610°F. The loop had to be terminated when the pump (LFB model) seized because of a bearing failure. The third loop was terminated after 521 hr because of a failure in the heater section. The peck fuel temperature was 1500°F; the maximum tube wall temperature was 1610°F; the Reynolds number was about 15,000; and the temperature differential was 200°F, During one period of the test, the loop had to be operated isothermally for 69 hr because of a failure in the power supply control. The results of examination of these loops are reported in Sec. 6, ‘‘Corrosion Re- search.” The results obtained from these tests indicate an undesirable temperature distribution in the heater section. At the tube bends between the heating elements, higher temperatures (as much as 100°F) occurred on the inside of the bend than on the outside. To determine the effect of the uneven temperature distribution on corrosion, a new loop in which the heating is accomplished entirely in a straight section of tubing was put into operation. A second loop incorporating the old, coiled heater section was also started. A zirconium-base fuel containing both UF4 and UF is being circulated in these loops. The unit with no bends in the heater is operating at a Reynolds number of 15,000, a temperature differential of 200°F, and a maximum fuel temperature of 1500°F, The other loop has a Reynolds number of 10,000, a temperature differential of 300°F, and a maximum fuel temperature of 1500°F. In order to determine the effect of the method of heating the loop on corrosion and mass transfer, a gas-heated loop and an electrical-resistance heated loop are being operated on a comparative basis. They are both operating at a Reynolds number of 1000, o temperature differential of 300°F, and maximum fuel temperature of 1500°F; they are circulating the fluoride mixture NaF- Zer-UI"'4 (53.5-40-6.5 mole %). Sodium in Multimetal Loops D. R. Ward W. B. McDonald Aircraft Reactor Engineering Division The fourth of a series of tests’ of sodium in beryllium-inconel systems was completed. The ’D. R. Ward, L. A. Mann, and W. B. McDonald, ANP Quar, Prog. Rep. Dec. 10, 1954, ORNL-1816, p 45. 35 ANP PROJECT PROGRESS REPORT loop, which was made of Inconel and contained a beryllium insert, was operated continuously for 1000 hr. The sodium was circulated at a maximum temperature of 1300°F (at the beryllium insert) and a temperature differential of 300°F. The Reynolds number ot the minor diameter of the beryllium insert was approximately 440,000. A similar loop that is all Inconel is being operated with sodium to provide a comparison with loops containing other materials. Also, a loop made of Inconel and type 316 stainless steel is being operated. The steel is inserted in the cold leg of the loop. The operating conditions for both loops are a maximum temperature of 1500°F, a minimum temperature on the cold leg of 1200°F, and the Reynolds number in excess of 15,000. The loops are scheduled to operate 1000 hr. HEAT EXCHANGER TESTS Intermediate Heat Exchanger Test No. 2 R. E. MacPherson Aircraft Reactor Engineering Division Design work is progressing on a heat exchanger test loop to provide data on corrosion, mass transfer, and reliability of a fuel-to-NaK-to-air system operating under conditions comparable to those postulated for the ART. The design will incorporate two heat exchanger tube bundles con- taining 100 tubes each in a regenerative type of circuit. Component procurement and fabrication are well under way, and it is intended that the testing will begin in May. The heat exchanger is basically similar to that used in the first intermediate heat exchanger test.® The NaK will flow from a 1-Mw gas-fired heater through one tube bundle to a radiator. The stream will then go to the liquid-metal pump and back through the second tube bundle to the heater. The fuel mixture NaF-ZrF4—UF4 (50-46-4 mole %) will be circulated outside the tubes countercurrent to the NaK flow and will be alternately heated and cooled by the NaK stream. A single tube bundle is composed of 100 Inconel tubes 3/] in, in outside diameter, 0.017 in. in wall thiciness, and approximately 6 ft in length. These tubes are arranged in a 10 by 10 matrix and contained in a square channel on 0.210-in. centers with a 0.011-in. tube-to-wall clearance. 8'R. E. MacPherson and H. J. Stumpf, ANP Quar. Prog. Rep. Dec. 10, 1953, ORNL-1649, Fig. 2.2, p 29. 36 The operating conditions are indicated on the flow diagram, Fig. 3.6. The fuel mixture specified for this test will restrict the attainable fluoride- side Reynolds number to approximately 26G0. o ORNL-LR-DWG 5673 80" F 700° F AIR IN — ) ——————— - AR OUT 154 psia - Mw 15 psio NaK-TO- AR 1099° F RADIATOR 1279°F “- 20 psic 30 psia A ol & e | o o3 | = Ztl ' E FUEL PUMP w o - —_— - » a 1 37 qpm wj{ u|o | uie o 3 % | @ ol wla gl Slo @fs Tl SR \' | - Lo | | — s 1] wl g8 | | & l 'U < I ZZ ) 4 concentration in the various loadings. The first assembly consisted of an essentially spherical fuel region surrounded by the reflector. In the second, of which there have been three variations, the fuel surrounded a core, or island, of ‘berylium, also spherical, and was enclosed by the reflector. The fuel was separated from the island and from the reflector by metal core shells to provide stability and, in the more recent experi- ments, to represent the reactor structure. This three-region assembly has been modified in the A. D. Callihan et al.,, ANP Quarn Pro§' Re ., ORNL-1692, p 45; ORNL-U?] p 44; ORNL-1816, third experiment by the addition of a column of beryllium to each side of the island and of cylindri- cal shells of fuel to the former spherical fuel shell to simulate the so-called ““end ducts.”” A cutaway view of the assembly with the end ducts is shown in Fig. 4.1. These end-duct additions represent the inlet and outlet flow channels of the reactor, The data from the first experiment and some of the critical parameters of the second experiment have been presented.! The results from the second experiment and preliminary data from the assembly with the end ducts are presented here. The dimensions and contents of all the assemblies are summarized in Table 4.1, and a comparison is made with the results of critical mass calculations by a multigroup method.?2 The present fuel loading of the assembly with the end ducts is excessive by an amount too great to extrapolate with certainty to a critical system with all poison rods removed. The immediate program includes a reduction of the loading so that an evaluation of the critical mass can be made, The distributions of neutrons detected by indium activation, with and without cadmium covers, along a vertical traverse in the midplane of the three- region assembly with aluminum core shells (CA-20q) is shown in Fig. 4.2, The cadmium fraction, that is, the fraction of the neutrons having energies below those absorbed by 0.02-in.-thick cadmium, is also shown as a function of the radius. These data have not been corrected for the absorption of indium-resonance neutrons by cadmium. The calcu- lated? bare and cadmium-covered indium activation traverses are also shown. Figure 4.3 gives the neutron distributions obtained from activated gold toils also located on a vertical traverse in the midplane of the assembly with the aluminum core shells (CA-20a) and of the assembly with the ]/s-in.-thick Inconel shells (CA-20¢c). From these data the fraction of neutrons with energies below the cadmium cut-off in the center of the fuel region of CA-20a is about 0.2, The corresponding value for CA-20¢, with considerable uncertainty, is 0.1, W. E. Kinney, private communication. 41 ANP PROJECT PROGRESS REPORT ALUMINUM SHEETS ALUMINUM SHEETS BERYLLIUM CORE SHELLS ALUMINUM EXTRUSIONS . ORNL-LR-DWG 3689A ALUMINUM SHEETS REGION NN REGION TS AT \\ ) ».‘\“ N N N \\\\ \\\\\\\ A ¥ “-PLANE OF TABLE SEPARATION Fig. 4.1. Longitudinal Cross Section of Three-Region Critical AssemBIy with End Ducts. The fission-rate distributions across the fuel in the three modifications of the three-region assembly are given in Fig. 4.4, The data were obtained from the activity of fission products collected on aluminum foils in contact with fuel sheets. Pairs of points at the same abscissa value give the results from opposite sides of a single 0.004-in.- thick foil. From these and similar data from cadmium-covered uranium-aluminum foil combi- nations, it has been possible to plot the cadmium fraction (that is, the fraction of all fissions pro- 42 duced by neutrons of energy below about 0.5 ev) across the fuel for the assembiies with the aluminum core shells and the ]/B-En.-thick Inconel shelis, The lower values for the latter assembly (CA-20c) are a consequence of the nuclear properties of the Inconel and of the higher uranium density. This variation in fission rate in the fuel sheets is also shown in Fig. 4.5, where the results of self- shielding measurements within the 0.004-in.-thick uranium foils are plotted. A fuel sheet near the beryllium island was replaced by four sheets, each TABLE 4.1, PERIOD ENDING MARCH 10, 1955 COMPOSITION OF REFLECTOR-MODERATED CRITICAL ASSEMBLIES Three Region with Core Shells Three Region wi th l’B-in.-Thick Two Region ]/lé-in.TThick ]/lé-in.-Thick ) «in.~Thick Core Shell's and Aluminum Inconel Inconel End Ducts ¥ Assembly number CA-19 CA-204q CA-20b CA-20c CA-21 % Beryllium island * Volume, ft3 None 0.37 0.37 0.37 1.27 Average radius, in. 5.18 3,18 5.18 4,20 (duct core radius) Mass, kg 19.4 19.4 19.4 67.0 Fuel region (excluding shells and interface plates) Volume, 3 1.05 1.78 1.78 1.72 2.06 liters 29,7 50.4 50.4 48.8 58.3 Average radius, in. Inside 5.24 .5,24 5.31 4,33 Outside 7.48 9.51 9,51 9. 44 5.28}("“‘ radii) Distance between fuel 0.173 0.639 0.284 0.142 0.142 sheets, in. Mass of components, kg Teflon 54,46 99.38 99.27 94.37 108.88 Uranium loading 11.88 5.00 11.74 22,07 26,02 U233 1oading 11.07 4.66 10,94 20.56 24,24 U233 density,* ¢/cm’ 0.372 0.092 0.217 0.421 0.416 Uranium coating 0.11 0.05 0.11 0.20 0.25 material Scotch tape 0.09 0.1 0.11 0.1 0.15 Core shells and interface plates Mass of components, kg Aluminum 0.92 5.85 1.10 1.10 1.10 Inconel 0 0 13.68 27.73 53.02 Reflector Volume, f° 21.24 22,22 22,22 22,22 20,88 Minimum thickness, in, 12,9 11.5 11.5 1.5 11.5 Mass of components, kg Beryllium 1102.5 1155.0 1155.0 1155.0 1094.1 Aluminum 16.8 29.2 29.2 29.2 29.2 Excess reactivity as loaded, % 0 0.9 0.3 0.4 ~3 Critical mass,** U235, kg Experimental 11.Q7*** 4,35 10.8 19.8 19 + 2 Cal culated 9.5 to 4.5 11,3 10.3 *y235 s per unit volume of fuel region. **Mass required for a critical system with poison rods removed. ***|t was necessary to increase the reflector thickness slightly to make this mass critical, 43 RELATIVE ACTIVITY (arbitrary units) ORNL—LR—DWG 5778 28 22 N Q N o 32 BERYLLIUM FUEL_ REGION BERYLLIUM 26 ——r e EXPERIMENTAL) 24 e — 4}, : i | CALCULATED BARE INDIUM ACTIVATION | I | 18 | - 16 - - ‘ CADMIUM - COVERED INDIUM 4 ACTIVATION (EXPERIMENTAL) / t.0 8 0.8 6 0.6 4 0.4 CADMIUM FRACTION 2 - 0.2 CALCULATED CADMIUM-COVERED INDIUM ACTIVA 0 0 C 1 2 3 4 5 6 T B 9 10 1 12 13 14 15 16 17 18 19 20 21 22 DISTANCE FROM REACTOR CENTER {in.) Fig. 4.2, Radial Neutron Distribution in Midplane of Three-Region Assembly with Aluminum Shells (CA-20aq). 140d3Y SS§3YI0Yd LDIFr0y¥d dNV CADMIUM FRACTION RELATIVE ACTIVITY (arbitrary units) RELATIVE POWER (arbitrary units) PERIOD ENDING MARCH 10, 1955 ORNL—LR—-DWG 5779 ] } ‘ 1 e —SERYLLIUM ISLAND —3=—|=€—FUEL ANNULUS ~meftt———— e BERYLLIUM REFLECTOR - | : — — — Vg-in=THICK 1/ ~in-THICK ALUMINUM INCONEL SHELLS SHELLS o0 BARE FOILS O ° CADMIUM-COVERED FOILS A A ® | : | 0.8 ‘ P - %\ 06 \ T ‘l\ * \\ ~ y/4 e \ . \ N X ° ° 2 0.2 A®, AL vl —O o ®. -/‘/ .\ — -'-l—‘-/ 0 2 4 6 8 10 12 14 16 18 20 22 DISTANCE FROM REACTOR CENTER (in.) Fig. 4.3. Radial Neutron Distribution in Midplane of Assembly from Gold Foil Activation. 05 ‘ l | ‘ ORNL-LR-DWG 5780 TR — BERTLILM FUEL REGION ir 8ERYLLIUM REFLECTOR | | ‘ | | FISSION O 'ig-in-THICK ALUMINUM CORE SHELLS ‘ al RATE 0 Yg-in.-THICK INCONEL CORE SHELLS A i DISTRIBUTION| A Yg-in.- THICK INCONEL CORE SHELLS 10— @ CADMIUM |— 'g-in-THICK ALUMINUM CORE SHELLS ] | FRACTION |—— Yg-in - THICK INGONEL CORE SHELLS ? NOTE: THE HALF SYMBOLS INDICATE THAT MEASUREMENTS / WERE MADE ON BOTH SIDES OF SOME URANIUM FOILS, | \, THE OPEN SIDE OF THE SYMBOL BEING DIRECTED xe‘;/‘/( TOWARD THE FOIL. i | ] 0.5 A A | 1 _ 3 /| i \ : a_/fl/ \ 8 ‘%\ v _ 2 o | ] 5 1 : 10 < L - | ‘ ----- — 0.8 & L N TTTH— 0/ — P . — 106 s — ~ L | i } : - .4 g \_\_ | ,_—/r ‘ ‘ ! ‘ ‘ — 0.2 A 0 T : . 0 g 0 1 2 3 4 5 6 DISTANCE FROM INNER EDGE OF FUEL ANNULUS (in.) Fig. 4.4. Fission-Rate Distribution Across Fuel of Three-Region Assemblies. 45 0.001-iM%#htck, and the catcher-foil activity was measured at each sheet. The experiment was ORNL-LR-DWG 5784 | | ® DATA OBTAINED 043 in. FROM ISLAND-CORE INTERFACE i ——r-otp— - 4 DATA CBTAINED 171 in. FROM ISLAND-CORE INTERFACE | | . 1 t . N O 1 2 3 4 5 5 7 URANIUM THICKNESS {(mils) RELATIVE ACTIVITY (arbitrary units) Fig. 4.5. Self-Shielding in 0.004-in.-Thick Urani- uvm Foil of the Aluminum Shell Assembly (CA-20q). Wt 2o 46 repeated in a plane near the center of the fuel region, and a more uniformly distributed fission rate through the centrally located fuel was ob- tained. These results, which agree at least quali- tatively with the neutron-flux distribution measure- ments in Fig. 4.3, are indicative of the spectral differences in the neutrons in the fuel, Some preliminary investigations have been made of the suggested use of the rare-earth elements as neutron poisons in reactor control rods, The reactivity coefficients of several wafers® of Gd,0, and Sm,0,, each about 0.01 in. thick and 4 em 2in area, have been measured at two places in the ]/g-in.-ihick Inconel shell assembly (CA-20¢), and little difference between the two materials has been observed, At the center of the island, where the cadmium fraction measured with gold foils is 0.5, the reactivity coefficients range from 0.10 ¢/mg/em? for a 0.0l-in.-thick sample to 0.04 ¢/mg/cm? for a 0.04-in.-thick sample. At the edge of the island, where the cadmium fraction is 0,35, the corresponding values are about 0.015 and 0.009 ¢/mg/cm?. The measurements show that a sample of these materials a few thousandths of an inch thick is 85% ‘‘black’’ to neutrons having the spectrum they have in this experiment. 3F‘repc|red by J. R. Johnson of theMetallurgy Division. Part |1 MATERIALS RESEARCH PR 5. CHEMISTRY OF MOLTEN MATERIALS W. R, Grimes Materials Chemistry Division Solid phase studies in the NQF-UF4 and NaF- ZrF ,-UF, systems were continued, and phase relationships and UF, solubility in the BeF,- bearing systems were studied. The usefulness of LaF, as a ‘‘stand-in” for UF, was demon- strated. Extensive experimental studies of the preparation and stability of UF, in alkali-metal- and zirconium-base fluorides are reported. All measurements made to date on the stability of UF, in molten systems and on the feasibility of prepa- ration of UF, by reduction of UF, have shown poor reproducibility, The 250-lb production facility, which was taken out of service on January 1, was reactivated to supply increased demands for zirconium-bearing fluoride mixtures. The pilot-scale facility has been used for developing a process for preparing BeF ,-bearing mixtures and for the preparation of various UF -bearing mixtures. The charge ma- terial for an in-pile loop was also prepared. PHASE EQUILIBRIUM STUDIES C. J. Barton Materials Chemistry Division H. Insley Consultant Solid Phase Studies in the I"lch-UF4 and NuF-ZrF4-UF4 Systems R. E. Moore L. M. Bratcher R. E. Thoma Materials Chemistry Division NoF-UF,. The gradient quenching technique prevmusly descrsbedl was used to study phase relationships in the NaF- UF4 system. Although this system had been studied earlier in this laboratory? and by others® by using thermal analysis, visual observation, and quenching tech- niques, it was never felt that o satisfactory under- ]C. J. Barton et al., 1954, ORNL-18164, p 57. 2C. J. Barton et al., ANP Quar. Prog. Reps., ORNL- 858, p 14; ANP-60, p 128, ORNL-1692, p 52; ORNL- 1729, p 41; ORNL-1771, p 55, 3c. A, Kraus, Phase Diagrams of Some Complex Salts of Uranium with Halides of the Alkali and Alkaline Earth Metals, M-251 (July 1, 1943). ANP Quar. Prog. Rep. Dec. 10, standing of phase relationships in this system had been achieved, particularly in the 25 to 50 mole % UF, region. As a result of x-ray and petrographic studies of quenched samples it was concluded that the cubic phase that Zachariasen* postulated to be a high-temperature form of Na,UF,, with a homogeneity range extending to 40 mole %, is actually a new compound Na U.F._ having « rather narrow temperature range of stability. Below 630°C it decomposes to form (3,-Na,UF, and Nao U F,,. At 672°C it melts incongruently to give fi\la U F.,. and liquid. A revised diagram that mc(udes the results of recent studies, as well as some of the earlier thermal analyses and visual observation results, is shown in Fig. 5.1. Liquidus temperatures determined by the quenching technique for mixtures having 27 to 40 mole % UF, were in excellent agreement with the recently presented partial phase diagram for this region.’ Two crystalline forms of Na,UF, reported by Zachariasen? and observed in earfler quenching work with this composition have not been observed in recent studies; since their tfemperature range of stability is not known, they were omitted from the diagram. NuF-ZrF4-UF4. The recent studies of the NaF-UF, system have contributed materially to a better understanding of phase relationships in the ternary system. A number of gradient quenches were carried out in an effort to define primary phase fields more accurately. The results of these experiments, presented in Table 5.1, seem to indicate an invariant point near the composition 66 mole % NaF-9 mole % ZrF,-25 mole % UF, where the primary phase fields of Na U F”, N03U(Zr)F7 solid solution, and Na,U( r)6 11 solid solution meet. Invariant pomts are also indicated near the composition 69 mole % NaF -4 mole % ZrF4—27 mole % UF4 where the N02UF6, Na U(Zr) solid solution, and Na U,F . primary phase fields meet and near 65 mole % NaF-28 mole % ZrF4—7 mole % UF4 where the Nazsz6, Na,Zr(U)F_, and Na Zr(U)6 41 Primary phase 7! 4W. H. Zachariasen, J. Am. Chem Soc. 70, 2147 {1948). 5('.'. J. Barton et al., ANP Quar. Prog. Rep. Sept. 10 1954, ORNL-1771, p 55. 49 ANP PROJECT PROGRESS REPORT ORNL-LR-DWG 5856 1100 l ! 1000 — — 900 |- —_——— — — , LBy=Na,UFg + LIQUID ; i i / ! UF, + LIQUID ! - - L . 800 e X - 7 NagUzFyp+ B3-NayUFg —— _ NaF + LIQUID | / ‘ o \ 3 f Nogll3F(, + LIQUID | | | w i L Na,UgFyy + LIQUID j i | E 00l % | /o o ] g 700 : a-NagUF, + LIQUID | | . [ ; \ - {‘107U6F3' + ‘ | ' : No U.F i 5-3 17 600 | ——— ————fa-NagUF, 1 T | - | —— e __‘_/ NaF + a-Na,UF, \ + ‘ i ‘ w\ £3,-Na,UF ‘ ‘ 3 2¥' 6 ! | - B5-Na,UF | | + ‘ i -Na,UF 500 |[——— - - B-nazuFy L No,UgF s, — - MU o 4 UF — NoF + 35-NazUF, + c7UgF 3 4 | | B5~Na, UF | | | 400 | ———— - 1 s S e Naf + 3,-Na,UfF g W W wo w : :)m :)N :}m | Dh ° Q ] : o ! z = 2 z ; 300 l | J ‘ E l i NaF 10 20 30 40 50 60 70 80 90 UF, UF, (mole Yo Fig. 5.1. Phase Diagram of the NaF-UF, System. fields meet. The NaBZr4F19 primary phase field boundaries have not been accurately located, as yet. It is expected that a revised phase diagram for this system will be completed in the near future. Phase Relationships and UF; Solubility in BeF ,-Bearing Systems L. M. Bratcher B. H. Clampitt R. J. Sheil R. E. Thoma Materials Chemistry Division NaF-BeF .LiF. Data on the solubility of UF, in NaF-BeF. mixtures obtained at another instal- {ation® showed that it is necessary to go to low concentrations of BeF. in this system before usefully high concentrations of UF, dissolved in 50 the melt at 600°C can be obtained. It has also been demonstrated that lower viscosities are obtained as the BeF, concentration is reduced. These developments encouroged a re-examination of the ternary system, with emphasis upon the LiF-LizBeF4-NazBeF4-NoF section of the system. Phase diagrams for the LiF-BeF, and NaF-BeF, binary systems have been published.7:8:% Un- published thermal analysis data for the ternary 6Private communication, J. F. Eichelberger, Mound Laboratories, to W. R, Grimes. ’D. M. Roy, R. Roy, and E, F, Osborn, . Am. Ceram. Soc. 36, 185 (1953); op cit., 37, 300 (1954). 8A. V. Noveselova and M. E. Levina, J. Gen. Chem. U.S.5.R. 14, 385 (1944). 7E. Thilo and H. A, Lehmann, Z. anorg, Chem. 258, 332 (1949). PERIOD ENDING MARCH 10, 1955 TABLE 5.1, RESULTS OF QUENCHING EXPERIMENTS IN THE NGF-ZrF4-UF4 SYSTEM Compos tion (mole %) Liquidus Secondary Temperature Peimary Phase Temperature Secondary Phase NaF ZeF, UF, °C) Q) 66.7 10.3 23.0 640 Na,U(Zr)F,* 66.7 6.3 27.0 646 Na UsF 640 Na ,U(Zr)F * 66.7 2.3 31.0 657 Na U,F .. 648 Na,UF 61.5 6.5 32.0 667 Na,U(Zr)F4,* 624 Na UsF (s 64.0 6.5 29.5 654 Na U.F, 613 Na U(Ze)F 4, * 64.0 11.0 25.0 640 Na,U(Zr) Fyy* 627 Na UsF 62.5 11.0 26.5 669 Na,U(Zr) F 4 p* 635 Na UsF o5 63.5 31.5 5.0 577 Na,ZrF, 567 Na, Ze(U)F 44 * 61.5 32.5 6.0 575 Na,Zr(U)F 4 * 547 Na, ZrF 59.5 33.5 7.0 580 Na_Zr(U) F 5 * 66.7 30.3 3.0 644 Na,Zr (U)F ,* 66.7 26.3 7.0 660 Na,Zr(U)F* 45.0 53.0 2.0 528 Na,Zr F o 513 Na,Zr(U)F 4 * 41.0 55.0 4.0 530 NayZr,F g 42.0 48.0 10.0 600 Zr(U)F > 553 Na,Zr(U)F 42.0 50.0 8.0 621 Ze(U)F 42.0 51.0 7.0 623 Zr(U)F * *Solid solution. system were obtained earlier in this laborctory.!? These data showed low thermal effects, below 300°C, for a number of compositions, but eutectic compositions could not be determined. Published data on the binary systems show that compositions in the ternary system near the LiF-Li BeF, eu- tectic (3] mole % BeF2; melting point, 460 * 5°C) offer the most promise of low melting points with low BeF . concentrations. The compound Na_ Bef melts at about 595°C, while the l\lql'—'-l\lc:,BeF4 eutectic (30 mole % BeF,) has a reported? melting point of 570°C. Recently published datal! on the ternary system show the existence of three ternary compounds, NuLiBeF4, Noa(BeF4)2, and NazLiBezFr The first compound apparently does ]OJ. P. Blakely, L. M, Bratcher, and C. J. Barton, unpublished data. ”W. Jahn, Z. anorg. u. allgem. Chem. 276, 113 (1954); op cit., 276, 274 (1954). not exist at liquidus temperatures, while the last was reported to separate from the melt, together with LiF, at a eutectic temperature of 320°C. Preliminary thermal analysis results for ternary compositions with 31 mole % BeF_, or less, seem to confirm earlier indications that conventional thermal analysis techniques do not give reliabie liquidus temperatures in this system. Other tech- niques, such as visual observation, filtration and quenching, will be applied to the study of these mixtures as soon as possible. LiF-BeF -UF,. Attempts to prepare the LiF- BeF,-UF, mixture, either by adding UF,; to purified iiF-BeF2 compositions or by reduction of UF, with excess uranium metal in similar solvents, have yielded material with considerable, and rather variable, concentrations of UF4. The experimental evidence suggests that alloying of uranium with the nickel equipment and the conse- quent lack of control of the uranium activity may 31 ANP PROJECT PROGRESS REPORT be responsible for the lack of reproducibility of the data, The considerable scatter in the data precludes firm statements as to UF, solubility in this medium. It appears, however, that the LiF-BeF mixture containing 31 mole % BeF, will dissolve at least 2 wt % of U3* at 600°C. Petrographic examination of these materials shows the UF_ to be present as large well-formed crystals, which appear to have been deposited from solution. No complex compounds of UF3 appear. lhermal data on such mixtures suggest very low solubility of UF, at the solidus temperatures., BeF -UF,. A 50-50 mole % mixture of BeF, and UF, is the only mixture of these materials that has been examined to date. No distinct thermal effects were found on the cocoling curve, the maximum temperature being about 900°C. Petrographic examination of the preparation showed UF3, crystalline colorless BeF2, and BeF, with a yellowish color. Chemical analysis of the material showed 95% of the uranium to be in the trivalent form. The remaining tetravalent uranium is presumed to be due to oxidation of a part of the U3* by oxidizing impurities in the BeF,, such as H20 and BeSO,. It appears from this preliminary experiment that UF, and BeF, do not form a compound and that solid solution, if it occurs, is very limited, PbFz-Ber. A phase diagram for the PbF,- BeF, system, published recently,” showed that these components form two compounds, 3PbF, - BeF, and PbF,-BeF,. It was shown also that the latter compound forms extensive solid so- lutions with Ber. The statement was made that mixtures in this system are quite fluid, even with as much as 95 mole % BeF.,. This suggested that the fluidity of alkali fluoride-BeF, mixtures might be usefully increased by the addition of PbF, if the resulting mixtures were compatible with structural materials at high temperatures. As o preliminary test of compatibility, mixtures of PbF, and BeF, containing 50 and 75 mole % BeF, were prepared by heating the compounds to about 800°C in nickel crucibles equipped with nickel stirrers. The breaks on the cooling curves agreed fairly well with the values reported for these compositions. The absence of metallic lead in the resulting melts suggests that the activity of the Pb*" jons in the melts was sufficiently low that very little reduction of PbF, by the nickel 52 walls occurred. The compatibility of these ma- terials with Inconel containers will be tested in the near future, NaF-Ber-UF . Three filtrations were carried out to determine the solubility of U3* in the Na,BeF ,-NaBeF, eutectic composition (43 mole % BeF,). The U3* values obtained with samples filtered at 600 = 10°C were 1.16, 1.22, and 0.82 wt %. These results are in agreement with the findings at Mound Laboratory® that UF, solubility in this system is very low except at low Ber concentrations. This behavior might be expected, since BeF, and UF, do not form a compound, while NaF and UI:3 form a compound believed to be NoUFd. Phase Relationships in LaF ;- and UF -Bearing Systems L. M. Bratcher R. E. Thoma Materials Chemistry Division The usefulness of LaF, as a ‘‘stand-in”’ for UF, was discussed in the previous quarterly report.'2 These studies have been continued, and it was found that there is o close corre- spondence in thermal effects between L0F3 ond UF, systems with components that form simple eutectics (for example, LiF oand UF4). This probably means that LaF, and UF_ have about the same melting point. LiF-LaF,. The LiF-LaF; system appears to be a eutectic system with a eutectic temperature of 770°C, the same as that reported for the LiF-UF, system.'® The eutectic composition is probably about the same also, that is, approxi- mately 28 mole % LaFa. UF,-LaF,. The components UF,-LaF, also form a eutectic that melts at about 865°C. Insuf- ficient data were obtained to locate the eutectic composition, but, since UF, melts at a much lower temperature than LGF3 (estimated melting point, 1425°C), the eutectic composition probably con- tains much more than 50 mole % UF,. Cooling curves with UFA--UI'—'3 mixtures have shown a spread from about 835 to 875°C in thermal effects, but the higher value is probably nearer to the correct eutectic temperature. 12R. E. Moore and R. E. Thoma, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 59; P. A. Agron and M. A. Bredig, loc. cit., p 72. 13¢, J. Barton et al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 59. NaF-KF-Lqu. A mixture containing 25 mole % NaF-25 mole % KF-50 mole % LaF, was found to contain a single phase. This indicates solid solution formation between NalaF, and l'(LnF4 and confirms the belief that NaUFfl and KUF4 form solid solutions. RbF-LuFa. Dergunov’s thermal data'? for the FbF-LaF, system were essentially confirmed. The melting point of 585°C for the eutectic at 20 mole % Lcll:3 is the lowest melting point in alkali fluoride-L.aF; systems, and the corresponding RbF-UF3 eutectic would be of interest if oxidation and disproportionation of UF3 in RbF mixtures could be avoided. LaF, and UF, in ZrF4-Beuring Melts. A mixture having 2 moles of ZrF, per mole of LaF; con- tained some free Lan, while a mixture with a 3:1 molar ratio was essentially a single phase that was believed to be a compound composition. Therefore the compound previously designated QZrFA-UF may actually have the composition 3ZrF ,-UF,. Free UF; was reported to be present in the best preparation of the ZZrF4-UF3 compo- sition; failure to find free UF_ in other prepa- rations of this composition may have been due to the existence of part of the uranium in the tetra- valent state. Petrographic and x-ray studies of a number of ZrF,-UF,-LaF, compositions were made in on effort to identi?y more conclusively the two compounds believed to exist in the ZrFA- UF4-UF3 system, but phase relationships in this system are not well understood and will receive further study when time permits. A brief and incomplete thermal analysis survey of the RbF- LaFa-ZrF4 system failed to show usefully low observed melting points except at very low LaF, concen- trations. PREPARATION AND STABILITY OF UF ,~BEARING MELTS F. F. Blankenship C. J. Barton Materials Chemistry Division Reduction of UF , with Uranium in Alkali Fiuorides R. J. Sheil B. H. Clampitt Materials Chemistry Division Data on the reaction of metallic uranium with Ul:4 dissolved in alkali fluorides have been e p, Dergunov, Doklady Akad. Nauk S.5.5.R. 60, 1185 (1948). PERIOD ENDING MARCH 10, 1955 obtained in the past mainly with preparations ranging from 0.5 to 2.0 kg in weight.!3:16 |4 appeared desirable, however, to study the reaction with smaller scale equipment in order to facilitate investigation of the different variables involved in the reaction. Therefore most of the recent experiments have been performed with small nickel reactors fitted with nickel filter sticks, and 20 ¢ of material has been used. The main advantages of the small equipment are the larger number of experiments that can be performed per man-hour, smaller amount of materials required, and the faster rates of heating and cooling that can be achieved. The principal disadvantage of the small-scale filtration apparatus is that it does not lend itself readily to the use of the purification procedures routinely practiced with larger prepa- rations. In general, previously purified alkali fluorides are ground and loaded into a reactor in a dry box with the necessary amount of UF and freshly cleaned uranium metal. After equilibration and filtration at the desired temperature, the reactor is opened in a dry box and all the filtrate and unfiltered residue are ground for analysis. The data obtained to date are not sufficiently complete or reproducible to accurately evaluate the effect of all the variables. Effect of Nickel Surface Area. Early efforts to reduce UF, dissolved in NaF-LiF eutectic with metallic uranium in small-scale filtration equip- ment were uniformly unsuccessful, although the reaction could be readily carried out in sealed capsules in phase study apparatus and in large- scale preparations. The filtration equipment used in these experiments had the filter medium in contact with the melt during the reaction period (usually 2 hr or more). This apparatus was modi- fied to permit the filter medium to be in contact with the melt only during the actual filtration time and for a short time thereafter while the reactor was being cooled as rapidly as possible. This change resulted in a large increase in the degree of reduction of uranium in the filtrate. Measurements of the surface area of the nickel filter medium (0.0004-in. pore size) by the Carman permeability method'” showed a surface area of 135G, M. Watson et al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, o 77. 164 A, Friedman, ANP Quar. Prog. Rep. Dec. 19, 1954, ORNL-1816, p 61. 17p, C. Carman and J. C. Amell, Can. J. Res. 26, Sec. A, 128 (1948). 53 ANP PROJECT PROGRESS REPORT 2.9 x 10° cm? per cubic centimeter of filter ma- terial. For the standard Y%-in. sheet thickness, this is equivalent to about 900 c¢m? per square centimeter of apparent area. The reason for the marked effect of this large surface area upon the degree of reduction of uranium is not clearly under- stood, but it can be explained by assuming that disproportionation of UF, in alkali fluorides is a heterogeneous reaction that occurs at metal surfaces. Further studies of this variable are under way. Effect of Uranium Surface Area. Increasing the surface area of the metallic uranium by converting it to the hydride brought about a significant increase in the reduction of UF, dissolved in NoF-KF-LiF eutectic (11.5-42-46.5 mole %). The effect of uranium surface area was larger at lower temperatures, probably because the UF, produced at the metal-salt interface is more soluble at the higher temperatures. Conversion of the uranium metal to the hydride was adopted as o standard procedure. Effect of Temperature. Varying the temperature for the reaction between NaF-KF-LiF-UF, mix- tures and metallic uranium from 550 to 750°C produced no significant change in degree of re- duction. The ratio of U3* to total uranium for five experiments in this temperature range varied from 0.57 to 0.70. Effect of Time. Increasing the time of reaction of the NaF-KF-LiF-UF, mixture with the metallic uranium at 550°C from 15 min to 2 hr produced an increase from 0.47 to 0.58 in the ratio of U3* to total uranium in the filtrate. This difference may not be significant in view of the poor repro- ducibility of these experiments. The reaction time for NaF-LiF-UF, mixtures with metallic uranium at 750°C varied from 15 min to 6.7 hr. Again, the variation of the ratio of U3 to total uranium was of doubtful significance, but the highest ratio (0.75) was obtained with the shortest reaction time and the lowest ratic (0.57) with the longest reaction time. The data obtained to date show that the reaction of UF , dissolved in alkali fluorides with finely divided uranium metal is very rapid and thet it may possibly be followed by disproportionation at a rather slow rate. Effect of Excess Uranium. Parallel studies of the reaction of NaF-KF-LiF-UF, mixtures with 8y, c. Whitley and R. J. Sheil, ANP Quar. Prog. Rep. Dec. IO: 1954: ORNL']B]éJ P 60' 54 the theoretical amount of uranium required to satisfy the equation ‘4U°+ 3/4UF4 = UF, and 1.2 times the theoretical amount showed very little difference in the ratio of U* to total uranium in most cases. Where there was a ditference, the preparation with the theoretical amount of uranium gave the lower degree of reduction. It seems likely that uranium can react with oxidizing impurities in the fused salts which would other- wise oxidize a part of the UF . Stability of UF; in Alkali Fluorides R, J. Sheil B. H. Clampitt C. J. Barton Materials Chemistry Division The study of the stability of UF, in KF, which was discussed in the previous repor’r,]8 was continued and extended to other alkali fluorides. The wvariables investigated were temperature, solvent, container material, and concentration of UF3. Some experiments were performed to in- vestigote the effect of several variables simul- taneously and the results are reported according fo experiment type. Sealed-Capsule Tests, A number of sealed nickel, Inconel, and type 316 stainless steel capsules containing KF-UF, and RbF-UF, mix- tures were prepared. After t?'\e capsules had been welded shut with a helium atmosphere over the fluoride mixture, they were heated for 90 min at the temperatures shown in Table 5.2, After cooling, the capsules were opened in a dry box. A part of the fused salt was removed by drilling, and as much as possible of the remainder of the material was removed by chipping and was added to the drillings for analysis. The capsule walls were carefully cleaned and then drilled in air to provide a sample of the wall material that had been in contact with the fused salt for chemical analysis. The drillings were weighed so that the total amount of uranium in the walls could be calculated. It was recognized that recovery of metallic uranium alloyed with the walls was probably not complete, and therefore the values given in the last column of Table 5.2 are more likely to be low than high. The percentages were calculated on the basis of the amount of uranium in the original mixture. PERIOD ENDING MARCH 10, 1955 TABLE 5.2, SEALED-CAPSULE UF3 STABILITY TESTS WITH KF-UF3 AND RhF-UF3 MIXTURES Fused Salt Analysis o Test Ratio of U3+ Total Uranium Composition Temperature Container Material Total y3t to in Walls (mole %) °C) Uranium* (wt %) Total Uranium (%) (wt %) 85 KF-15 UF3 780 Nickel 35.6 8.2 0.23 5.2 Inconel 34.8 8.8 0.25 4.6 Type 316 stainless 36.4 8.5 0.23 3.4 steel 800 Nickel 37.8 11.5 0.30 4.0 Inconel 35.6 11.5 0.32 4.0 Type 316 stainless 35.2 13.8 0.39 steel 850 Nickel 35.2 7.2 0.20 3.5 Inconel 36.0 12.6 0.35 3.8 Type 316 stainless 34.3 5.3 steel 200 Nickel 36.2 19.2 0.53 2.6 85 RbF~15 UI::3 800 Nickel 27.9 13.1 0.47 3.5 Incone! 28.9 14.9 0.52 4.4 Type 316 stainless 28.6 12.6 0.44 9.2 steel 200 Nickel 23.7 8.5 0.36 5.0 Inconel 22.8 7.4 0.32 6.2 Type 316 stainless 5.1 steel *The theoretical total uranium for KF-UF3 mixtures was 38.1 wt %; for RbF-UF3 mixtures it was 27.6 wt %. The resuits given in Table 5.2 show no sig- nificant effect of temperature or container material on the ratio of U3 to total uranium in the fused salt, There was also no reproducible difference between KF-UF, and RbF-UF3 mixtures. In- spection of the data indicates that the repro- ducibility of results is not good enough to show anything except large effects. There appears to be a discrepancy between the ratio of the U™ to total uranium for the KI':-UF3 mixtures in Table 5.2 and that previously reported'® for an Incone! capsule heated to 1000°C, which showed 70% of the uranium in trivalent form. In that case, the material analyzed consisted only of the drillings from the fused salt mixture. A fused salt layer approximately 90 mils thick adjacent to the tube walls was not sampled by this method, whereas an attempt was made to get all the fused salt out of the capsules in the later experiments. The copsules were not agitated during the heating period, and, therefore, if disproportionation of UF occurred at the tube walls, it is possible that the concentration of tetravalent uranium next to the walls was higher than in the center of the capsules. This possibility will be further in- vestigated. Open-Capsule Tests. The effect of varying the concentration of UF | in NaF-KF-LiF eutectic was tested in open Inconel and molybdenum capsules heated in a helium atmosphere for 90 min at 750°C. Molybdenum was used because it does not alloy appreciably with uranium at this temperature. The capsule walls were not analyzed for uranium in these experiments, because it is difficult to effect more or less complete removal of the fused sait from the capsules without leaving the walls in a poor shape for sampling, porticularly with the 35 ANP PROJECT PROGRESS REPORT TABLE 5.3. OPEN-CAPSULE UF3 STABILITY TESTS WITH I(F-UF3 AND Nch-KF«LiF-UF’3 MIXTURES . 3+ . Ratio of U Th tical Initial Composition Container Total Uranium ud* ate @ eore u:-q le Material (w % (wt %) to Total Uranium (mole %) atertd wt %) e Total Uranium (mole %) 85 KF-15 UF3 Inconel 34.8 17.5 0.50 38.1 Molybdenum 33.8 12.1 0.31 38.1 95 {NaF-KF-LiF*)-5 UF3 Inconel 19.7 12.2 0.62 22.1 Meolybdenum 19.9 10.6 0.53 221 80 (NaF-KF.LiF*)_20 UF3 Inconel 48.5 41.3 0.85 51.7 Malybdenum 49.5 44.0 0.89 51.7 60 (NaF-KF-LiF*)-40 UF; Inconel 64.2 56.4 0.88 66.65 Molybdenum 66.0 60.8 0.92 66.65 *11.5-42-46.5 mole %. brittle molybdenum capsules. The KF-UF; mix- tures were included in these experiments for com- parison with similar mixtures tested in sealed capsules. The results of chemical analyses of the fused salt mixtures are given in Table 5.3. It appears from the results in Table 5.3 that alloying of the metallic uranium formed by dispro- portionation of UF3 hos little effect on the extent of disproportionation. A part of the U3 results are not in agreement with the values calculated from total uranium values on the basis that one gram-atom of uranium metal is produced when 4 gram-atoms of U3* disproportionate. The possi- bility of inhomogeneity in the samples cannot be ruled out, even though they were ground to —60 mesh in dry air. Solubility of UF, in NaF-RbF.LiF Mixtures R. J. Sheil Materials Chemistry Division Thermal analysis data for UF, dissolved in NaF-RbF-LiF mixtures were reported earlier,!? Since it had been demonstrated that thermal analysis was not a reliable method for determining solubility of UF, in such mixtures,?? it seemed advisable to check the thermal analysis data for this system by other techniques. Visual obser- vation of liquidus temperatures and chemical analysis of filtered samples gave the values shown in Table 5.4 for the solubility of UF, in a mixture near the ternary eutectic composition. It appears from the data in Table 5.4 that UF is significantly less soluble in the RbF-containing 56 TABLE 5.4. SOLUBILITY OF UF4 IN NaF-RbF-LiF (10-50-40 mole %) Temperature Uranium Content UF4 Content (°C) (wt %) (mole %) 500 7.4 ~ 2.5 600 13.1 ~ 4.1 650 18.8 ~6.3 mixture than in the NaF-KF-LiF eutectic (11.5- 42-46.5 mole %) and that the thermal effect ot 425°C observed earlier with the NaF-RbF-LiF mixture containing 2.5 mole % UF, represented the solidus temperature for this composition rather than the liquidus temperature. Solubility of UF, in NaF-KF-ZrF, B. H. Clampitt Materials Chemistry Division An attempt was made to determine the solubility of UF; in a low melting NaF-KF-ZrF, mixture (5-52-43 mole %; melting point, 410°C). This mixture was recently reported to have low vis- cosity.2! After adding sufficient UF, to give 19y, P. Blakely, L. M. Bratcher, and C. J. Barton, ANP Quar. Prog. Rep. Dec. 10, 1951, ORNL-1170, p 87. 20R, ). Sheil, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 59. 215 |, Cohen and T. N. Jones, Preliminary Measure- ments of the Viscosity of Compesition 20, ORNL CF- 55-2-20 (Feb. 2, 1955}, 4.0 mole % concentration to a hydrofluorinated preparation of the ternary:composition, the mixture was heated to 600°C for 2 hr and then filtered at that temperature. The filtrate analyzed 0.48 wt % U3* and 1.55 wt % total uranium, while the corre- sponding values for the unfiltered residue were 13.1 and 14.0% respectively. These results show a yvery low ust solubiljty at 600°C in this solvent. Reduction of UF, in Molten Liasz7 and LiZZrF6 M. A. Friedman Materials Chemistry Division A mixture of LizZrF, (mp, 645°C) with sufficient UF, to make the mixture 15 wt % uranium was treated at 800°C with o 50% excess of zirconium metal and stirred by bubbling hydrogen. The ap- parent melting point of the mixture so obtained was 630°C. A filtrate obtained at 645°C showed 3.9 and 4.0% U3 by two ditferent analytical methods and 4.66% total uranium. Since UF, is quite soluble in this melt the insoluble uranium compound is certainly trivalent; about 95% of the UF, charged was reduced to UF,; about 85% of the uranium in the filtrate was trivalent. When o similar experiment with Li,ZrF, was performed the filtrate obtained at 600°C showed only 1.1% U* and 2.89% total uranium. Preparation of UF, in NaF-ZrF, F. P. Boody Materials Chemistry Division Experiments on a 500-g scale demonstrated that attempts to completely reduce UF, to UF, in NaF-ZrF, produced mixtures containing both UF, and UF,. Therefore a request for 3-kg quantities reduced as completely as possible by ZrH, pro- vided an opportunity to study the effect of batch size on the UF,-to-UF, ratio. A preparation was carried out in which 5 wt % uranium as UF, was used in 3 kg of NaF-ZrF, (53-47 mole %) and 275% of the stoichiometric zirconium metal. The metallic zirconium was added as smail chips after a preliminary purification of the melt with HF and hydrogen. After 24 hr of equilibration while bubbling with hydrogen at 800°C, the melt was filtered at 710°C. The product contained 5.17 wt % uranium, including 3.13 wt % U3*. In view of the possibility that an incomplete reaction had occurred because of precipitation of UF,; on PERIOD ENDING MARCH 10, 1955 the zirconium metal, another trial was made in which two 6- by ‘/2-in. zirconium bars were used in addition to the chips. The bars were propped up in the body of the melt to afford good surface contact. After 5 hr of equilibration at 800°C, the filtered product contained 5.06 wt % total uranium, including 2.86 wt % U3*. Since it appeared that the equilibrium amount of UF, had not been in- creased, another method was tried in which 7 wt % uranium metal was added to 3 kg of NaF-ZrF, (53-47 mole %). Complete reaction between the uranium metal and the ZrF, would change the NaF-to-ZrF, ratio from 53:47 to 55:45. After 6 hr of equilibration at 800°C and filtration at 780°C, a product containing 4.68 wt % total uranium and 3.30 wt % U3" was obtained. The ratio of U 1o total uranium in these trials was 0.605, 0.565, and 0.705, respectively. It was concluded that the last and highest figure was representative of equilibrium at 800°C, but that reproducibility could not be expected because of uncontrolled variation in the activity of solid uranium metal and zirconium metal. CHEMICAL REACTIONS IN MOLTEN SALTS F. F. Blankenship L. G. Overholser W. R. Grimes Materials Chemistry Division The Equilibrium FeF, + H,=—Fe° + 2HF in NaZrF, ot 800°C C. M. Blood Materials Chemistry Division During the past year numerous attempts have been made to determine the equilibrium HF con- centrations resulting from the reduction of FeF, solutions in NaZrF; by hydrogen. In addition to the importance of this reaction in purification procedures, there was interest in the possibility of measuring the activity coefficient of FeF,. Most of the attempts have been efforts to determine the equilibrium HF concentration by extrapolation to zero flow rate in a system with hydrogen bub- bling through a melt containing FeF .. During the past quarter several methods of improving the contact between hydrogen bubbles and the melt have been tried, and as the contact has been improved the measured HF concentrations associated with a given FeF, concentration have 57 ANP PROJECT PROGRESS REPORT In each case, however, the HF concentration at very low flow rates has been about three times the value at flow rates of 200 ml/min in the same system. The shapes of the concentration vs flow-rate curves have been identical, and it has not been possible to obtain a satisfactory extrapolation to zero flow rate be- cause of the steep slope at low flow rates. By using the best extrapolation that could be made with each system and assuming the activity of the reduced metallic iron to be unity, successive improvements in contact of hydrogen with the melt have increased the values of consistently increased. 2 PouF P HZ X FeF2 from 2 to 8, where X is mole fraction and P is pressure in atmospheres. Since K, = 8 is about four times the K, expected from the thermodynamic estimates, it is possible that the assumption of unit activity for the metallic iron is not valid. The reduced iron may alloy with the nickel container. The largest HF values are found in systems having the largest surface area of nickel. It had been recognized that the nickel surface was probably acting as a catalyst for the heterogeneous reaction, and it was hoped that a definite equilibrium could be reached in a system containing nickel mesh baffles arranged so that a bubble required 6 sec to rise through 10 in. of melt. When a slow flow rate was resumed after the sys- tem containing the baffles had been cllowed to “‘rest’’ over night, an increase instead of the expected decrease in HF concentration was noted during the next few hours. This same apparatus was also used in matching the HF concentration in the inlet and outlet streams by adjusting the inlet HF concentration. The results again agreed with the values of K found by extrapolation to zero flow rate, that is, K= 8 at 800°C. Typical results for various systems are shown in Table 5.5. Reduction of UF, by Structural Metals J. D. Redman C. F. Weaver Materials Chemistry Division Apparatus and techniques for experimental de- termination of equilibrium constants for the re- actions ‘ Cr° + 2UF, ==CrF, + 2UF, and Fe® + 2UF, —=FeF, + 2UF, were described in previous reports.??2 Equilibrium 22), D. Redman and C. F. Weaver, ANP Quar. Prog. Reps,, ORNL-1692, p 56; ORNL-1729, p 50; ORNL- 1771, p 60; ORNL-1816, p 64. TABLE 5.5. EFFECT OF BUBBLE PATH ON EQUILIBRIUM HF CONCENTRATION IN THE REACTION H, + FeF, ——=Fe® + 2HF IN NoZrF_ AT 800°C Calculated Activity Conditions of Contact Between Gas and Melt K _* Coefficient** for FeF2 Bubbles from 3/B-in.- tube rising through 5 in. of melt 2.1 1.3 Bubbles from 3/B-in. tube rising through 10 in. of melt 4.4 2.7 Bubbles from 3/8-in. tube rising through zigzag baffles 8 5 in 10 in. of melt 2 T *Kx = , where P is pressure in atmospheres and X is mole fraction. X FeF2 H2 1 Ka **K}/ = - — ;AF°=RT In K_; at 800°C, Ka = 1.6, if solid FeF2 in the standard state and an activity of metallic iron of unity are assumed. 58 data were presented for these reactions in NaZrF, with pure metallic chromium and metallic iron used as the reducing agents. In addition, some experi- ments with Inconel as the reducing agent and some preliminary observations with chromium metal and the NaF-KF-LiF eutectic as the reaction medium have been presented. The results of some recent studies on the re- action of UF, with metallic iron in NaF-KF-LiF (11.5-42-46.5 mole %) at 600 and 800°C are given in Table 5.6. In these experiments, 2 g of hydro- gen-fired iron wire was reacted with UF, (15 wt %) in about 21 g of the NaF-KF-LiF eutectic contained in nickel. The iron concentrations reported in Table 5.6 are very nearly the same as those found when NaF-ZrF, was used as the solvent. The iron concentrations in the latter case were obtained after reaction with 11.8 wt % UF, as compared with 15 wt %, but this difference in UF, content would result in only slightly higher values for iron at the higher UF, concentration. It may be seen from Table 5.6 that higher iron concentrations result at 600°C than at 800°C, and this is also in agreement with what was found in NaF-ZrF,. No attempt has been made to calculate equilibrium constants from these data, since the valence state of the iron is not known with certainty. Similar studies have been made of the reaction of UF, with chromium in NaF-KF-LiF (11.5-42- TABLE 5.6. EQUILIBRIUM DATA FOR THE REACTION OF UF (15 wt %) WITH IRON IN MOLTEN NaF-KF-LiF (11.5-42-46.5 mole %) AT 600 AND 800°C Conditions of Found in Filtrate Equilibration Total Total Total Temperature Time Uranium lren* Nickel (°C) (hr) {wt %) {ppm) (ppm) 600 3 12.7 655 100 3 12.6 750 125 5 11.7 685 85 5 10.8 680 95 800 3 11.9 470 80 3 12.2 425 90 5 11.2 525 130 5 11.9 485 110 *Blank of 80 ppm of iron at 800°C. PERIOD ENDING MARCH 10, 1955 456.5 mole %) at 600 and 800°C. In these experi- ments 2 g of hydrogen-fired metallic chromium, 21 g of the solvent, and the desired amount of UF, were equilibrated in nickel apparatus. The total uranium concentration was kept constant at 11.4 wt % but the ratio of UF, to UF; was varied. The results are presented in Table 5.7. The chromium concentrations given in Table 5.7 for 600°C are very much lower than the 2200 ppm of chromium found when NaF-ZrF, was the solvent, On the other hand, the values at 800°C when no UF, was added are approximately equal to those found in NaF-ZrF,. The larger differences in chromium concentrations at 600 and 800°C in molten NaF-KF-LiF than in NoF-ZrF, suggest that mass transfer of chromium will take place much more readily in NaF-KF-LiF. The addition of UF, to the system decreased the equilibrium chromium concentration markedly and indicated that corrosion might be greatly reduced by using a mixture of UF, and UF4. However, the total uranium concen- trations found for those runs in which UF,; was added were decreased significantly during the test, probably because of disproportionation of the UF,. Uncertainty in the valence state of the chromium precludes any evaluation of equilibrium constants at present. Previous studies of the reaction between UF, and Inconel in molten NaZrF,, as expected, gave equilibrium chromium concentrations that were much lower than those resulting from the reaction of UF, with chromium. Similar studies with the NaF-KF-LiF mixture as solvent for reacting hydro- gen-fired Inconel turnings with 14.7 wt % UF , have been made in apparatus of nickel. The data are shown in Table 5.8. The values given in Table 5.8 are somewhat more than threefold lower than those found when UF, reacts with metaltlic chromium. However, there is some doubt as to whether equi- librium was attained in the runs reported in Table 5.8, since the increase in chromium concentration from 5 to 12 hr at both temperatures is small enough that it may or may not be significant. Some studies have also been made on the re- action of metallic molybdenum with UF, in molten NaF-ZrF, contained in nickel ot 600 and 800°C. The results given in Table 5.9 were obtained at a UF, concentration of 11.4 wt %. The results show that UF4 is stable in contact with metallic molyb- denum under the conditions used and confirm the low molybdenum concentrations reported for similar 59 ANP PROJECT PROGRESS REPORT TABLE 5.7. EQUILIBRIUM DATA FOR THE REACTION OF UF, WITH METALLIC CHROMIUM IN MOLTEN NaF-KF-LiF (11.5-42-46.5 mole %) AT 600 AND 800°C Conditions of Equilibration Original Concentration Found in Filtrate Temperature Time UF4 UF3 Total Uranium Total Chromium* Total Nickel (°C) (hr) (Wt %) (wt %) (wt %) (ppm) (ppm) 600 3 15.0 12.2 1190 30 3 15.0 11.8 1075 75 5 15.0 11.3 960 75 5 15.0 12.0 1130 40 800 3 15.0 1.2 2690 20 3 15.0 11.2 2600 50 5 15.0 11.9 2700 35 5 15.0 11.4 2810 15 5 10.0 4.7 11.3 55 95 5 10.0 4.7 10.6 30 70 5 10.0 4.7 10.7 25 45 5 7.5 7.1 10.1 120 70 5 7.5 7.1 10.4 130 80 5 7.5 7.1 10.2 175 75 5 7.5 7.1 10.4 135 135 *Blank of 500 ppm chromium at 800°C. TABLE 5.8. REACTION OF !.JF4 WITH INCONEL IN MOLTEN NaF-KF.LiF (11.5-42-46.5 mole %) AT 600 AND 800°C Conditions of Equilibration Found in Filtrate Temperature Time Total Uranium Total Chromium* Tetal lron Total Nickel e {hr) (wt %) (ppm) (ppm) (ppm) 600 5 11.3 350 195 90 5 10.9 355 165 75 12 11.0 385 195 550 800 5 10.9 640 90 145 5 10.9 735 125 100 12 11.0 995 110 230 12 10.9 810 95 125 *Blank of 200 ppm of chromium at 800°C. melts that had been circulated in thermal-con- vection loops fabricated of molybdenum or of Hastelloy B. The data obtained for similar runs in NaF-KF- LiF (11.5-42-46.5 mole %) at 600 and 800°C with 15 wt % UF, are given in Table 5.10. The results indicate that the reaction of UF, with metallic 60 molybdenum proceeds in this solvent to a greater extent than in NaF-ZrF,. They show that the molybdenum concentration is lower at 800 than at 600°C and suggest that it decreases with time at 800°C. There is no apparent explanation for either the high uranium values or for the poor precision of the experimental results. PERIOD ENDING MARCH 10, 1955 TABLE 5.9. REACTION OF UF, WITH METALLIC MOLYBDENUM IN MOLTEN NaF-ZrF, AT 600 AND 800°C Conditions of Equilibration Found in Filtrate Temperature Time Tatal Uranium Total Melybdenum* Total Nickel (°C) {(hr) (wt %) (ppm) (ppm) 600 3 8.4 7 155 3 8.6 7 135 5 8.5 7 230 5 8.6 9 215 800 3 8.5 8 90 3 8.4 8 85 5 8.5 9 30 5 8.6 11 85 5 8.6 9 80 5 8.6 9 10 *Blank of 20 ppm of molybdenum at 800°C, TABLE 5.10. REACTION OF UF, WITH METALLIC MOLYBDENUM IN MOLTEN MNaF-KF-LiF (11,5.42-46,5 mole %) AT 600 AND 800°C Conditions of Equilibration Found in Filtrate Temperature Time Total Uranium Total Molybdenum* Total Nickel (°Q) (hn) (wt %) (ppm) (ppm) 600 3 13.3 210 85 5 14.0 200 110 5 1.8 325 170 800 3 13.5 130 170 3 13.9 105 85 5 14,4 55 205 5 13.9 65 145 *Blank of 5 ppm of molybdenum at 800°C and 30 ppm at 600°C. Stability of Chromous and lron Fluerides in Molten Fluorides J. D. Redman C. F. Weaver Materials Chemistry Division Previous studies?® had indicated that Fe*" was relatively stable in NaF-KF-LiF and that the solubility of FeF, was 12 and 19 wt % at 600 and 800°C, respectively; however, it was found that Cr*? was not stable in this solvent and that it apparently underwent disproportionation to metallic 23y, D. Redman and C. F. Weaver, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 63. chromium and Cr3*. Experiments have now been performed in which ferrous, ferric, and chromous fluorides were added to NaF-KF-LiF (11,5-42-46.5 mole %) to determine their stability. The results of a number of runs in which FeF, was added to the NaF-KF-LiF mixture and equilibrated ot 600 and 800°C Table 5.11, The nickel and ferrous concentrations found in the filtrate suggest that the reaction Ni + 2FeF,==NiF, + 2FeF, in nickel equipment are given in occurs under these conditions. However, the ratio of Fe** present to that theoretically resulting from 61 ANP PROJECT PROGRESS REPORT TABLE 5.11, STABILITY OF FeFy IN MOLTEN NaF-KF-LiF (11,5-42.46.5 mole %) Conditions of Equilibration Found in Filtrate FeF; Added s Temperature Time (% Fe) Fe Tetal Fe Total Ni (°C) (hr) (wt %) (wt %) {wt %) 600 5 4.9 0.72 1.57 0.53 10.0 0.73 1.72 0.53 800 10.0 5.1 9.3 3.5 15 5.4 0.24 4.6 1.10 5.4 0.17 4.7 1.60 10.8 0.58 9.2 2.7 10.8 0.50 9.0 2.9 the Ni** content is only 0.75 at 600°C and about 0.1 at 800°C. This behavior indicates that al- though Fe™™ is formed by the reaction of Ni with Fe®* the resulting Fe** is not stable and probably disproportionates as 3Fe** ==Fe° + 2Fe3* The relatively large amounts of finely divided iron observed in all the runs heated for 15 hr sub- stantiate the belief that disproportionation occurs. In an attempt to reconcile the results presented in Table 5.11 with the earlier observation that FeF, is stable in the NaF-KF-LiF mixture, some runs were made in which FeF, was equilibrated in the presence of NiF,, These experiments were carried out in nickel, and the results given in Table 5.12 show that Fe*" is not stable in the presence of Ni**, and it must be assumed that Ni** enters into some reaction which at 600°C removes nearly all the Ni** from the melt. The high nickel concentrations of the residues sug- gested that metallic nickel was present. The reactions involved may be the following: 3Fe*t ==2Fe3*t 4 Fe° L+ N 44 ) Fe® + Ni —Fe + Ni° Some studies with metallic iron and NiF, as re- actants are to be made in an attempt to evaluate the equilibrium concentrations for the reactions. Previous studies?® have shown that CrFy is stable in the molten NaF-KF-LiF mixture contained in nickel at 600 and 800°C and data also were presented which indicated that CrF, is not stable 62 TABLE 5.12. STABILITY OF FeF, IN MOLTEN NaF-KF«LiF (11.5-42.46,5 mole %) IN PRESENCE OF NiF, FeF2 Odded: 6-0 wit % Fe NiF2 added: 2.1 wt % Ni Equilibrated 5 hr Equilibratien Found in Filtrate Temperature Fe'? Total Fe Total Ni (°C) (wt %) (wt %) (wt %) 600 2.4 2.9 0.01 2.5 2,9 0.02 800 2.9 5.9 0.7 2.7 5.8 1.1 under these conditions but, rather, undergoes disproportionation (3CrF2—> Cr® + 2CrF3). Additional data that support this belief are given in Table 5,13, Analytical data obtained for the residues indicate that the solubility of CrF, was not exceeded in any of the runs and that the solubility of CrF, was exceeded in all runs at 600°C but not at 800°C. Although the precision of these measure- ments is not of a high order, the data have been used to obtain an approximate value for the equi- librium constant of the reaction 3CrF2:—f— 2CrF3 + Cr° For concentrations expressed in mole fractions, the value for K_ is of the order of 10% at 600°C. TABLE 5.13. STABILITY OF CrF, IN MOL TEN NaF-K F-LiF (11.5-42.46,5 mole %) AT 600°C Equilibrated 5 hr at 600°C Found in Filtrate CrF, Added *+ 34 % Co (52 %) (E: %) 2.9 0.30 1.08 2.9 0,32 0.74 5.8 0.30 0.91 5.8 0.48 1.08 The value for K _ for this reaction at 600°C (calcu- lated from the standard free energy of formation of CtF, and CrF3) is about 103, PRODUCTION OF PURIFIED MOLTEN FLUORIDES F. F. Blankenship G. J. Nessle L. G. Overholser Materials Chemistry Division Production-Scale Operations F. L. Daley J. P. Blakely Materials Chemistry Division A total of 2077.6 kg of processed fluorides was produced in the 250-lb facility during the quarter. The various compositions and amount of each processed are listed below. Amount Processed Composition (kg) NcF-ZrF4—U Fy (50-46-4 mole %) 496.8 NaF-ZrF 4-U Fy4 (53.5-40-6.5 mole %) 451.7 NaF-—ZrF4-U Fyu (56-39-5 mole %) 790.4 NaF-—Zer (54.1-45.9 mole %) 338.7 Totol 2077.6 The 250-Ib facility was taken out of service on January 1, at which time a total of 2545 kg of processed fiuoride was available for the various testing programs. |t was anticipated that this stockpile would be adequate for 4 to 6 months. It now appears that increased demands resulting from the general speed-up in the ANP program and the increased interest in UF j-bearing materials PERIOD ENDING MARCH 10, 1955 will require operation of the production facility briefly in March and at frequent intervals there- after. This facility is in standby condition and can be returned to service on approximately a two-day notice, Pilot-Scale Preparation of BeF, Mixtures J. P. Blakely C. R. Croft 4. Truitt Materials Chemistry Division A general process development program has been instituted to determine the best method for proc- essing fluoride melts containing BeF, in the pilot- scale facility. A preliminary batch of NaF-BeF, (57.43 mole %) was processed with no treatment except melting and mixing under a helium atmos- phere. This material could not be filtered. Visudl observation of a melted sample revealed consider~ able amounts of an insoluble white powder, which was assumed to be BeO. Later analyses of ¢ sample of this batch which had been used for physical property studies showed approximately 5% BeO. This value may, perhaps, be high due to imperfect handling of the batch, but it is cbvious that more rigorous purification is required, Accordingly, the standard procedure for the preparation of zirconium~base fuels was applied to mixtures containing BeF,. Briefly, the pro- cedure consists of melting the batch under a HF atmosphere, treating the melt with hydrogen for 1 hr at 800°C, hydrofluorinating it at 800°C for 90 min, and finally stripping it with hydrogen to a value of 1.5 x 10=4 moles of HF per liter of exit gas. It was immediately apparent that the sulfur content of the BeF, was high; the odor of H,S was readily detectable in the out-gas. Table 5.14 shows pertinent analyses for several samples of the available BeF,. The analyses indicate that both sulfur and chromium are present in considerably higher con- centration than previosuly found in ZrF,. Since chromium is removed very slowly from the ZrF - bearing mixtures with hydrogen, some trouble from high CrF, concentrations in the product was ex- pected. Chemical analyses are at present avail- able for only three of the nine batches processed by this method, and these results are presented in Table 5,15, The data are, as yet, insufficient for generali- zation, but some differences in behavior between BeF,- and ZrF,-bearing mixtures are apparent. 63 ANP PROJECT PROGRESS REPORT TABLE 5,14 PURITY OF BeF, RAWMATERIAL Major Constituents (wt %) Container No. Minor Constituents {ppm) Be F Fe Cr Ni 5 1 19.3 76.7 0.032 230 150 30 2350 3 19.1 78.1 0.049 250 165 35 2205 5 19.1 76.7 0.029 250 135 35 2270 7 19.2 80.1 0.035 215 145 30 1810 8 19.3 80.4 0.038 245 135 35 1930 TABLE 5.15. FINAL PURITY OF NaF-BeF, MIXTURES Major Constituents Minor Constituents B atch Na. (wt %) {ppm} Be F Fe Cr Ni S EE330 8.70 59.9 235 12 1 19 EE331 8.87 59.4 225 50 120 55 EE332* 8.79 58.3 215 35 165 320 *This was batch EE330 broken up and reprocessed. The CrF, content of the melt is appreciably low- ered by using the zirconium-base fuel process, but the iron concentration is not, There is consider- able inconsistency in the sulfur analyses, but there is no reason to doubt that the sulfur is readily removable. It was observed that the reactor dip-line tended to plug as HF concentrations approached 1 x 10~4 moles of HF per liter of exit gas. Consequently about 50% of the baiches processed were termi- nated at readings of approximately 1.5 x 10—4 moles of HF per liter. It was also noted that occasionally the HF level of a sample rapidly decreased to a value near 1 x 10~3 moles HF per liter, remained there for 2 to 4 hr, and then gradu- ally rose to nearly 2 x 104 moles HF per liter before decreasing again. Successful filtrations of these preparations have been accomplished with sintered nickel filters of 0.0024., 0.0015., and 0.004-in. pore diameter. The finest of these filters is presently being used on a routine basis. Pilot-Scale Preparation of UF ;-Bearing Mixtures G. J. Nessle J. P. Blakely Materials Chemistry Division A series of eight preparations was completed that contained NaF-KF-LiF and a constant total uranium concentration with wvarious U3+/UF4 ratios. The process used in the preparation of these batches was identical with that used previ- ously in which separate purification steps for the main constituents were utilized. The NaF-KF-LiF eutectic (11.5-42-46.5 mole %) is first heated and stripped with hydrogen at approximately 800°C. The melt is then cooled and the UF, is added. The heating and stripping are again repeated; the melt is cooled, and the uranium metal is added. The final heating and stripping are carried out, and the batch is transferred to a storage can, In a previous study of the reaction U + 3UF, —> 4UF, in this environment, it was demonstrated that the reaction does not proceed to completion. In the previous series of preparations, sufficient uranium metal was added to yield approximately 11 wt % U3*, and fairly consistent values of around 5 wt % U3* were obtained. In the present series of prepa- rations, approximately 2.5 wt % U3* was desired, so the uranium-to-UF , ratio was set to yield theo- retical U3* values of 6 wt %. The results ob- tained are presented in Table 5.16. Although the level of U3* desired was not generally attained, these batches have been released for corrosion studies in thermal~-convection loops. For another series of eight preparations con- taining 50 mole % NaF, 46 mole % ZrF,, about 3 mole % UF,, and 1 mole % UF,, the proper quantities of NaF and ZrF containing the desired PERIOD ENDING MARCH 10, 1955 amount of UF, were treated with HF and hydrogen in the normal manner until less than 1 x 10~% moles of HF per liter of exit gas was indicated. The melt was then cooled to room temperature and the UF; was added carefully under a helium atmos- phere. The melt was again heated and agitated with hydrogen until the temperature of the melt reached 600°C, held at temperature and agitated for 2 hr, and transferred through the filter to the storage receiver in the normal fashion. The general reproducibility of this type of preparation is indi- cated in Table 5,17, 1t is believed that experience with this procedure will make it generally useful for preparation of several types of UFj-bearing mixtures. TABLE 5,16, CHEMICAL ANALYSES OF UF,-BEARING ALKALI FLUORIDE MIXTURES Barch No. Einal U3+ Total Uranium Minor Con stituents (ppm) (wt %) {wt %) Fe Cr Ni EE323 2.0 1.5 50 40 70 EE325 1.2 111 80 30 35 EE326 3.4 13.3 100 30 105 EE327 1.6 11.3 115 45 30 EE328 1.6 11.0 105 45 30 EE410 1.4 1.9 70 35 30 EE411] 1.6 12,3 50 50 35 EE412 3.0 12,2 95 40 55 EE413 .7 10,6 65 S0 75 TABLE 5.17. CHEMICAL ANALYSIS OF UF;-BEARING NaF-ZrF ,-U Fy MIXTURES 5 atch No. Total Uranium Final U%" Minor Constituents (ppm) (wt %) (wt %) Fe Cr Ni EE420 11.8 0.98 260 125 40 EE421 7.4 1.87 50 70 295 EE422 8.5 1.60 65 55 105 EE423 8.7 2.19 95 175 120 EE424 9.4 1.66 40 45 30 EE426 2,00 EE427 1.08 65 ANP PROJECT PROGRESS REPORT Special Services J. E. Eorgan J. P. Blakely Materials Chemistry Division Hazards Experiments, Two 300-Ib batches of molten NaF-KF-LiF eutectic were prepared for hazards experimentation. The first test consisted in dumping 500 Ib of the molten mixture at 1500°F into a large steel tank whose hemispherical bottom was externally cooled by immersion in water. The second test involved injection of the 1500°F batch of salt into a tank of water well below the water line.. The transfers of 500 Ib of molten salt through l1-in. lines required approximately 1 min in each case; the transfers were effected by using 25 psig of helium to force the salt from the storage reser- voir into the test tank. The results of these tests are described in sec. 3, ‘“Experimental Reactor Engineering.’’ Charge Material for In-Pile Loop. A request for 1500 g of a mixture containing 63 mole % NaF, 25 mole % ZrF,, and 12 mole % UF, was received from the radiation damage group. Since this ma- terial was to be used for an in-pile loop, enriched uranium was required, and the processing had to be done in the facility in Bldg. 9212 at Y-12. Assembly and repair of the processing unit in Building 9212 were completed on January 9, 1955, and the batch was finished on January 11, 1955, Samples were submitted for analysis to both the X-10 and Y-12 analytical laboratories. The ac- countability transfer of the material was done on the reported values of the Y-12 laboratories, with the X-10 results used as checks. A considerable amount of repair and parts re- placement had to be done on this processing unit before the material could be prepared. Complete remodeling of the unit will be required before new batches can be processed safely again. Five thousand pounds of low-hafnium ZrCl, was received for conversion intfo ZrF, at Building 9211 in Y-12. It is esti- mated that this stockpile of ZrF , will be sufficient to meet all production demands of zirconium-base fuels unless the present rate of consumption in- creases or a new composition is needed. ZrF, Processing. Preparation of Various Fluorides B. J. Sturm E. E. Ketchen Materials Chemistry Division The preparation and purification of various fluorides have continued. Increasing demands for 66 hydrofluorinated UF, and pure UF, made it neces- sary to devote most of the effort to these materials, although substantial amounts of the structural metal fluorides and LaF, were also prepared. Chemical analysis supported by x-ray and petro- graphic examination has been used to establish the identity and purity of the moterials, Approximately 12 kg of UF, was purified by hydrofluorination at 600°C to meet the requirements for the preparation of UF, and various experi- mental uses. Three batches of FeF, were pre- pared by hydrofluorination of anhydrous FeCl, at 450°C followed by a helium flush at this temper- ature. Approximately 8 Ib of (NH,),CrF, was synthesized by heating an excess of NH, HF, with CrF,3%H,0 at 200°C. A portion of the (NH);CrF, was thermally decomposed under helivm at 700°C to yield CrF ;, and another portion, after being decomposed to CrF, at 600°C, was reduced to CrF, by heating under hydrogen at 800°C. Approximately 4 1b of LaF; was prepared from La,0; by converting the oxide to an aqueous solution of LaCl,, precipitating the fluoride by decantation and centrifugation, and finally drying at 150°C in a silver vessel. The need for large quantities of pure UF, for various studies has been met through a concerted effort on the part of this group. The method being used is essentially the same as that described previously24 in which finely divided uranium metal and UF, are reacted at 900°C while being rotated in a steel capsule containing steel balls. Some modifications were made in the apparatus which permit the capsules to be [oaded and unloaded in the vacuum dry-box, and, also, the heating period has been extended from 24 to 32 hr. There is some evidence that these changes have been instru- mental in producing a product of higher purity than previously obtained. The composition of nine batches (740 g per batch), as determined by chemical, and petrographic examination, is given in Table 5.18, All runs were made with 1.3% excess of uranium metal. Runs 2 through 8 were heated for 24 hr and 9 and 10 for 32 hr. An evaluation of the purity of the UF; is very difficult. The x-ray method will not detect the small amounts of UF, that are present in most cases, and the petrographic examination, while being more sensitive to the presence of UF,, cannot give any quantitative values, The chemical X=ray, 24y, C. Whitley and C. J. Barton, ANP Quar. Prog. Rep. Sept. 10, 1951, ORNL-1154, p 159. PERIOD ENDING MARCH 10, 1955 TABLE 5,18, COMPOSITION OF UF, Chemical Analysis (wt %) Run No. Petrographic Examination X-ray Data Total U UF, F 2 81.1 95 18.8 Some UF, Found UF,, UO, 3 79.6 99 19.9 Trace of UF, Only UF, 4 80.5 99 19.5 Seme UF Found UF 5 80.2 93* 19.8 Only UF, Only UF, 6 79.8 97 20,7 Trace of UF, Only UF4 7 79.9 97 19.7 Trace of UF, Only UF, 8 80.1 5% 20.0 Some UF, Only UF4 9 80.6 100 19.6 Trace of UF, 10 79.1 98 19.1 Seme UF *Found 97% for sample submitted six weeks later, **Found 98% for sample submitted four weeks later. method vused is probably not capable of giving values that are more accurate than 2%, It should be noted that there is rather poor agreement be- tween the chemical results and the petrographic results in the case of run No, 5, It is not known why the reaction is practically complete in some cases and not so in others. No UO, has been detected in any of the batches except No. 2, and this suggests that the cause of low UF, contents is incomplete reaction rather than oxidation, FUNDAMENTAL CHEMISTRY OF FUSED SALTS Electrochemistry of Fused Salts L. E. Topol Materials Chemistry Division Measurements are being made in an attempt to find a reversible fluoride electrode for melts, It is known that Ag/AgCl serves as a reproducible reference electrode for molten chlorides, and therefore the half-cell Ag/AgF is one possibility for fluorides. However, preliminary experiments with AgF indicate the salt to be very unstable in the presence of traces of moisture., Even with careful handling in a dry box the AgF heated at 650°C for 5 hr in helium appeared to have hydro- lyzed almost completely to yield metallic silver (silver oxide decomposes above 300°C), Further 25¢, Wagner and D. Balz, Z. Elektrochem. 56, 574-9 (1952). work with AgF has therefore been abandoned for the time being. Another possibility for a fluoride electrode is the half-cell Ni/NiF2 (saturated in other fluorides). Wagner and Balz25 have studied the system KF-NiF,, and have found that two compounds exist = K,NiF, and KNiF,. According to their data, solutions between 9.1 and 33.3 mole % NiF, in KF in the temperature range of 797 to 930°C would result in a solid phase containing K,NiF, that would precipitate upon saturation. Thus two half cells containing different concentrations of NiF, in the above limits should produce an emf of zero. The emf apparatus consists of a fairly gas-tight can of suitable size to fit in a dry box. Electrodes of grade A nickel rod of 1/8--in. diameter are welded to 40-mil nickel wire leads. These are insulated from the can by Morganite recrystallized.alumina thermocouple beads through which the wire fits tightly. A gas inlet, an outlet, and a thermocouple well complete the cell. The two half-cells con- tained in nickel or Morganite crucibles sit on a Morganite plate and are joined electrically by a slightly porous ZrO, bridge previously impregnated with an alkali fluoride (the same as that to be used in the experiment as the solvent). All nickel electrodes and vessels are annealed before each run by heating in hydrogen at 770°C for 1 hr. A helium atmosphere purified by a liquid-nitrogen, activated-charcoal trap is used throughout. 67 ANP PROJECT PROGRESS REPORT Measurements have been made, to date, in KF, LiF-KF (50-50 mole % eutectic), and the eutectic mixture NaF-KF-LiF (11.5-42-46.5 mole %)}. Half- cells containing equal concentrations of NiF, in the NaF-KF-LiF or LiF-KF eutectics give emf’s of the order of 2 to 10 mv at 600 to 750°C, How- ever, with differing concentrations of NiF,_ in the two half-cells, emf's much higher than the above were obtained (20 to 60 mv). Even though most cells were run for two days, equilibrium conditions may not have been obtained, especially in the formation and solubility of the complex. X-Ray Diffraction Studies in the NoF-ZrF , System P. A. Agron M. A. Bredig Chemistry Division The use of high-temperature x-ray diffraction techniques was continued in the study of the polymorphous transitions of the compounds Nazsz6 and Noasz7 in the NoF-ZrF4 system. Several runs were made up to 550°C with the 30 mole % ZrF4 composition to assist in locating the various Na,ZrF, transitions and to determine the extent of solid solution occurring in the Na,ZrF, phase. A marked phase transition took place at 520 to 525°C which did not reverse on cooling and holding for 1 hr at 495°C. The ex- tensive solid solution in the Na Zr[:7 phase (Table 5.19) even at temperatures :Leiow 550°C reduced the amount of Na,ZrF, phase and made the identification of the latter rather difficult. Table 5.19 lists the lattice parameters of the body-centered tetragonal phase (isomorphous with the body-centered tetragonal form of Na UF.) at two temperatures and compositions. etro- graphic examination?® of sample b-2 showed the presence of two major phases, along with a minor amount of a finely divided constituent that is possibly an oxidation product. One phase is index of refraction uniaxial negative with an slightly tower than 1.404, and the second is biaxial positive with on index slightly higher than 1.404. the second phase are close to those?? belonging to phase 4 of Na,ZrF , but these phases differ in x-ray diffraction pattern. On the basis of the resemblance of the x-ray pattern of this phase to that of the lower temperature form2® of N°3UF7’ tentative orthorhombic cell dimensions are pro- posed as belonging to a lower temperature NOBZrF phase. The petrographic characteristics of 7 Physical Chemistry E. R. Yan Artsdalen Chemistry Division The density and electrical conductivity of the pure fused salts (1) potassium bromide, (2) sodium iodide, (3) cesium chloride, and (4) rubidium 264, 1n sley, Consultant. 27 ANP Quar. Prog. Rep. June 10, 1954, ORNL-1729, p 40, 28)51e ported data. TABLE 5.19. LATTICE PARAMETERS IN THE COMPOUND Na;ZrF, OF THE NaF-ZrF, SYSTEM NagZrF, Sample Composition Temperature Cell Dimensions Designation (mole % ZrF ,) (°C) Crystal Symmetry (z) a 25 25 Body-centered tetragonal a = 533 c = 10,53 b-1 30 524 Body-centered tetragenal a = 5,45 c = 10.90 b-2* 30 25 Body-centered tetragonal a = 537 c = 10.75 Orthorhombic = 8,36 = 573 c = 10,90 *Sample cooled in high-temperature furnace from sample b-1. 68 PERIOD ENDING MARCH 10, 1955 bromide were measured as functions of temperature. Specific conductances of these salts, respectively, may be expressed by the equations: (1) K = =3.226; + 1012, x 107% - 4.827, x 105/ (© {(range 740 to 960°C) (2) K = =0.820, + 5.9404 x 1073 ~ 1976, x 10-6,2 (¢ = 0.002,) (range 675 to 915°C) (3) K = —1.346, + 4,537, x 1073 - 1106, x 1078/ (0 = 0,002,) (range 650 to 9205°C) (4) K = —3.030 + 9.103, x 10~3: - 4.509, x 10~6:2 (@ = 0.002.) 5 9 6 2 (range 695 to 905°C) where ¢t is in °C and o represents standard deviation. The densities of these salts, respectively, are given by: (1) p = 2733, (2) p = 3.3683 (3) p = 3.482, (4) p = 3.4464 The applicable temperature ranges - 0.8252; x 103 (o = 0.0004,) -3 - ~ 0.94905 x 10~3; (¢ = 0,0004)) - 1061 x 10~3; (0 = 0.0004) — 1.0718 x 10~3; (@ = 0.0005) are the same as those for «. 69 ANP PROJECT PROGRESS REPORT 6. CORROSION RESEARCH W. D. Manly G. M, Adamson Metallurgy Division W. R. Grimes F. Kertesz Materials Chemistry Division Additional thermal-convection loop studies have been made of Inconel and type 316 stainless steel exposed to alkali-metal-base fluorides with various proportions of UF; and UF , added, Inconel exposed to NaF-ZrF ,-UF, with ceramic contamination, Inconel exposed to the zirconium fluoride base mixture with 25 mole % UF,, Hastelloy B and molybdenum exposed to NaF-ZrF -UF,, Hastelloy B exposed to sodium, and beryllium exposed to sodium in Inconel. Preliminary examinations have been made of three Inconel forced-circulation cor- rosion and mass transfer tests, and the indications are that the depths of attack are only two to three times those found in thermal-convection loops operated for comparable periods. General corrosion studies were continued that included high-temperature tests of molybdenum, tests of brazing alloys on Inconel and stainless steel, a study of dissimilar metal mass transfer in the zirconium—type 304 stainless steel-sodium system, tests of the diffusion of sodium into beryl- lium in a beryllium-sodium-Inconel system, beryl- lium-Inconel spacer tests, tests of the corrosion resistance of cermets in NaF-ZrF -UF,, and a study of the solid-phase bonding of cermets. The investigation of mass transfer in liquid lead has been completed and summarized, and additionadl information on the fundamental properties of fused hydroxides is presented. In chemical studies of corrosion, the addition of chromium metal to NaF-ZrF, and NaF-ZrF -UF melts exposed to Inconel in tilting~furnace tests was found to reduce attack on the Inconel. The metallic chromium reduces the cotrosive chromic ion to the noncerrosive chromous state. THERMAL-CONVECTION LOOP CORROSION STUDIES G. M. Adamson, Metallurgy Division V. P. Treciokas, Pratt & Whitney Aircraft Alkali-Metal-Base Mixtures with UF; and UF, Several additional Inconel and type 316 stainless steel thermal-convection loops have circulated the 70 alkali-metal-base flyoride mixture NaF-KF-LiF (11.5-42-46.5 mole %) with various proportions of UF; and UF, added. The fluoride mixtures were prepared and handled with more precise procedures than those used previously, ! and therefore results that can be more accurately analyzed were ob- tained, The results obtained from these loops, which were operated for 500 hr with a hot-leg temperature of 1500°F, substantiate the previous observation that wuranium fluoride can dispro- portionate in these systems. Four Inconel and two type 316 stainless steel loops operated with fluoride mixtures containing more than 4.83 wt % UF, showed visible and microscopic evidence of a vranium-rich layer on the entire surface of the loop. The layers were about 1 mil in thickness and were thicker in the hot legs than in the cold legs. No corrosive attack was noted in any of the loops and no difficulty was encountered with plugging of the stainless steel loops. The absence of attack in these loops compared with the attack previously reported is thought to be, primarily, the result of the closer control of the UF, concen- tration; however, the materials probably also con- tained fewer impurities. The metallographic ex- of the loops have not yet been completed, but the data obtained thus far are presented in Table 6.1. The results of chemical analyses of the fluoride mixtures used are pre- sented in Table 6.2. A typical hot leg of an Inconel loop (loop 590) is shown in Fig. 6.1, and Fig. 6.2 shows a typical hot leg of a type 316 stainless steel loop (loop 193). The results of the chemical analyses {(Table 6.2) of the batch and the fill samples should have been identical since they were both taken from the aminations original material, but discrepancies existed and That prac- tically no UF; was left in the Inconel loops after additional portionation took place. therefore both results were reported. operation is evidence that dispro- Less disproportionation was evident in the stainiess steel loops than in 'G. M. Adamson and A. Taboada, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 76. PERIOD ENDING MARCH 10, 1955 TABLE 46.1. RESULTS OF METALLOGRAPHIC EXAMINATION OF THERMAL-CONVECTION LOOPS AFTER CIRCULATING NaF-KF-LiF (11,5-42-46.5 mole %) CONTAINING UF3 AND UF, Initial Uranium Maxi Loop Loop Content (wt %) aximum No. Material ° Atff:ck Metallographic Notes U udt (mils) 589 Inconel 8.09 5.4 None Uranium metal layer to a thickness of 0.5 mil throughout loop 590 Inconel 10.9 4.8 None Uranium metal layer to a thickness of 0.5 mil throughout loop 595 Inconel 10.5 5.6 None Uranium metal layer to a thickness of 0.5 mil in cold leg and 1.0 mil in hot leg 596 Inconel 11.5 5.0 None Uranium metal layer to a thickness of 0.5 mil in cold leg and 1.0 mil in hot leg 193 Type 316 stain- 12.5 66 2 Thin uvranium metal layer throughout loop; less steel attack in form of pits 194 Type 316 stain- 12.3 4.9 None Thin uranium metal layer throughout loop less steel TABLE 6.2. RESULTS OF CHEMICAL ANALYSES OF THE FLUORIDE MIXTURES CIRCULATED Loo Uranium {wt %) U3+ (wt %) Nickel Chromium Iron . P (ppm) (ppm)} (ppm) Ce Batch Fill Final Batch Fill Final Fill Final Fill Fina! Fill Final 589 8.1 7.9 7.7 5.4 4.7 0.7 30 45 40 55 95 125 590 10.9 10.8 10.1 4,8 7.7 1.1 75 39 55 55 125 210 595 10.5 11.0 10.1 5.6 5.3 1.3 30 75 35 20 75 140 596 11.5 10.7 10.7 5.0 5.8 1.5 110 140 40 17 130 140 193 12.5 11.7 11.6 6.6 6.4 4.4 65 70 85 55 130 335 194 12.3 10.8 11.4 4.9 6.0 3.3 35 55 45 18 75 160 the Inconel loops. The nickel and iron impurities increased more in these loops than in loops which circulated mixtures containing UF, and no UF,. The chromium contents were comparable. In addition to the loops described above, six other Inconel loops have been operated, but the results of examination have not yet been received. Visually, all six loops showed evidence of metallic deposits similar to those reported metallographi- cally in Table 6.1. These loops inciuded one which circulated a fluoride mixture containing as little as 1.24 wt % UF,. Ceramic Contamination in NaF-ZrF ,-UF, on Inconel The depths of attack have recently increased on Inconel thermal-convection loops operated as controls for 500 hr at a hot-leg temperature of 1500°F with NaF-ZrF,-UF, (50-46-4 mole %), and, in addition, more erratic results have been obtained. Previously operated loops showed attack to a depth of 8 mils, whereas the last three control loops have shown attack to a depth of 11 mils. One possible cause of the increased attack has been thought to be the ceramic spacers 71 ANP PROJECT PROGRESS REPORT that are occasionally used on the filling level probes. These probes are about 18 in. long, and the operators occasionally place a few ceramic beads near the lower end to prevent shorting on the side walls of the loops. The ceramic beads are the type used on thermocouples, and they are not completely resistant to attack by fluoride mixtures, The fluoride mixture is always removed from the probe as soon as possible, but even in a short time some contamination could occur. The loop design has now been changed so that « shorter probe may be used. Five Inconel loops have been operated for 500 hr at 1500°F to determine what effect the UNCLASSIFIED ¥-7205 ‘NI OO0 X082 Fig. 6.1. Hot Leg of Inconel Thermal-Convection Loop After Circulating NaF.KF-LiF.-UF ,-UF, for 500 hr at 1500°F. Loop 590. 250X. Reduced 30%. ceramic beads may have had on the rate of cor- rosion. Loops 609 and 610 were operated without ceramic beads as controls; loops 611 and 612 were operated with a string of ceramic beads on the probes; and loop 613 was operated with a fluoride mixture that had been contaminated with ceramic beads placed in the fill pot. The results of examination of these loops are presented in Table 6.3, Although essentially no differences in depth of attack were noted for control loop 609 compared with loop 613 and for control loop 610 compared with loops 611 and 612, larger and more numerous voids were observed on the loops operated with fluoride mixtures. The UNCL ASSIFIED T-7209 ceramic-contaminated . ’\3 ‘NI 0b0'0 X0§Z —#——-{ Fig. 6.2. Hot Leg of Type 316 Stainless Steel Thermal-Convection Loop After Circulating NaF- KF-LiF-UF;-UF, for 500 hr at 1500°F. Loop 193. 250X, Reduced 30%, TABLE 6.3. RESULTS FROM INCONEL THERMAL-CONVECTION LOOPS OPERATED WITH Nch-ZrF4--UF4 (50-46-4 mole %) CONTAMINATED WITH CERAMIC BEADS Loop Metallographic Notes N Variable o Hot-Leg Appearance Cold+Leg Appearance 609 Control Subsurface voids to 7.5 mils No deposit 610 Control Subsurface voids to 8 to 11 mils No deposit 611 Ceramic beads on probe Subsurface voids to 11 to 13 mils; No deposit surface rough 612 Ceramic beads on probe Subsurface voids to 11 to 12 mils No deposit 613 Ceramic beads in fill pot Subsurface voids to 7.5 to 8 mils Intermittent small metallic particles 72 surface of loop 613 was also much rougher than the surfaces of the other loops. the results for the two control loops indicates that, while the ceramic may have some effect, there are other unknown variables that affect the corrosion rate. A variation in High Uranium Content in Fluoride Mixture Determination of any unusual corrosive effects of fluoride mixtures with higher uranium con- centrations than those normally used was re- quested by the Solid State Division for evaluation of in-pile tests for which o high uranium con- centration is required. Subsurface void formations to depths of 10 and 15 mils were found in the hot legs of Inconel thermal-convection loops after circulating NaF-ZrF -UF, (62.5-12.5-25.0 mole %) for 500 hr ot 1500°F. The maximum depth of attack found in any loops in which the standard fluoride mixture NaF-ZrF,-UF, (50-46-4 mole %) has been circulated has been 11 mils. The higher uranium concentration does therefore appear to result in a small increase in depth of attack. Hastelloy B Loops Hastelloy B thermal-convection loops in both the as-fabricated and the dry-hydrogen-cleaned condition have been operated for 1000 hr af o hot-leg temperature of 1500°F with NaF-ZrF,-UF, (50-46-4 mole %) as the circulated fluid. The dry- hydrogen cleaning was for the removal of surface oxide layers. No appreciable reduction in cor- rosive attack was noted in the dry-hydrogen- cleaned loops as compared with the as-fabricated loops. The metallographic data for these loops, presented in Table 6.4, are in agreement with those previously reported for similar Hastelloy 3 PERIOD ENDING MARCH 10, 1955 loops.! The chemical analyses of the mixtures circulated in these loops have not yet been received. Hastelloy B loop 186 was operated with NaF-ZrF,-UF, (50-46-4 mole %) for 1000 hr with a hot-leg temperature of 1650°F. constructed with the drain, which acts as a cold This loop was trap during operation, in direct line with the cold leg. The lower loop joint was a sharp angle of 75 deg rather than the smooth bend used previously. Visval examination showed a considerable deposit of needle-like, magnetic crystals around the top of the trap and on both sides of the lower joint. This loop is to be compared with Hastelloy B loops 161 and 162, operated previously at 1650°F with off-set traps, which showed only small scattered deposits. Molybdenum Loops Three loops constructed of molybdenum jacketed with type 310 stainless steel for oxidation re- sistance have been operated at a hot-leg tempera- ture of 1500°F for periods up to 1000 hr with NaF-ZrF,-UF, (50-46-4 mole %) as the circulated tluid. subsurface void formation, or intergranular attack No evidence of mass-transferred particles, has been found. A rough surface, with depressions to a maximum depth of 2 mils, was noted but is believed to be associated with fabrication pro- A thin, aitered surface or metallic- appearing layer was also observed on the inner surface in both the as-received condition and following loop operation. The hot leg of loop 185, which operated for 1000 hr, is shown in Fig. 6.3. cedures. Sodium in Hastelloy B Loops Two Hastelloy B loops were cleaned with dry TABLE 6.4. CORROSION FOUND IN HASTELLOY B THERMAL-CONVECTION LOOPS AFTER CIRCULATING NoF-ZrF,-UF, (50-46-4 mole %) FOR 1000 hr AT 1500°F Metallographic Notes L.oocp No. Condition Hot-Leg Appearance Cold-l.eg Appearance 163 Dry hydrogen cleaned Pitting to a depth of 2 mils and subsurface voids No deposit 164 Dry hydrogen cleaned Pitting to a depth of 2 mils and subsurface voids No depesit 179 As-fabricated Subsurface voids to o depth of 4 mils Small metallic deposit 181 As-fabricated Subsurface voids to a depth of 1 mil No deposit 182 As-fabricated Subsurface voids to a depth of 2 mils No deposit 73 ANP PROJECT PROGRESS REPORT UNCLASSIFIED T-8944 i 3 t ‘NE OO0 X0§2 Fig. 6.3, HotLeg of Motybdenum Loop After Circulating NaF-ZrF ,«UF, for 1000 hr ot 1500°F. Loop 185. 250X. Reduced 30%, hydrogen and operated at a hot-leg temperature of 1500°F with sodium as the circulated fluid. Loop 188 was operated for 500 hr and loop 189 for 1000 hr. In both loops, deposits of magnetic crystals were found in the lower coid leg and in the hot horizontal section. The metallographic examination of loop 188 showed a rough surface, with pits and subsurface voids to a depth of 2 mils. These results confirm those presented in the previous reporf.] The susceptibility of Hastelloy B to the mass transfer of nickel inthe circulated sodium prompted the operation with sodium of three loops con- structed from A-nickel. Two of the loops were dry hydrogen cleaned prior to operation, while the third was operated in the as-fabricated con- dition. Dendritic, magnetic crystals were found in the cold legs of all three loops after operation for periods of 3500 and 1000 hr at a hot-leg temperature of 1500°F. Sodium in Inconel Loops with Beryllium Inserts G. M. Adamson A. Taboada Metallurgy Division Six Inconel thermal-convection loops with beryl- lium inserts were operated with sodium as the circulated fluid in a second series of sodium- beryllium-lnconel compatibility tests similar to 74 those previously discussed. ! Loops were operated for 500 hr at 1300 and 1500°F and for 1000 hr at 1000, 1200, and 1350°F. High-purity sodium from the same batch was used in all the loops. Metallographically, no attack and no deposits could be found in any of the cold legs or in any of the Inconel parts of hot legs that operated below 1350°F. The Inconel hot legs, including the sleeves, of loops operated at 1350 and 1500°F showed maximum intergranular attack to 2 mils. The beryllium inserts showed subsurface-void attack on the outer surface, as tabulated in Table 6.5. TABLE 6.5. DEPTH OF ATTACK ON QUTSIDE OF BERYLLIUM INSERTS FROM INCONEL LOOPS AFTER CIRCULATING SODIUM Temperature of Operating Maximum Loop No. Hot Leg Time Attack (°F) (hr) (mils) 560 1000 1000 0 559 1200 1000 3 557* 1200 1000 4 555 1300 500 1.5 558* 1350 1000 9 556 1500 500 21 *These two loops may have been reversed in cutting. Visual examination of the inserts reveaied no general attack; however, large streaks or rough areas were observed near the top or down one side of all inserts. Both the inside and the outside of the insert from loop 559 are shown in Fig. 6.4. The spiral down the center of this insert was found in all inserts except the one from the loop operated at 1000°F. Microscopic examination showed no change in the surface and no attack under the spirals. The spirals were accompanied by a dark nonmetallic deposit that could not be identified by diffraction or seen under the microscope. Neither of these areas of visible attack seems to correspond to any expected flow pattern. Loops are now being started to determine whether the areas of attack correspond to machining variables. UNCLASSIFIED T-6547 DIRECTION OF FLOW T e eesbrerarn L 3 OPNTHES Fig. 6.4. Surface of Beryllium Insert from Inconel Loop 559 After Circulating Sodium for 1000 hr ot 1200°F. FORCED-CIRCULATION CORROSION AND MASS TRANSFER G. M. Adamson R. S. Crouse Metallurgy Division Preliminary corrosion results have been received on the first three Inconel forced-circulation cor- rosion and mass transfer testing loops operated by Experimental Engineering (cf., sec. 3, “‘Experi- All these loops were terminated before their scheduled time, but mental Reactor Engineering’’). they operated long enough to provide some cor- rosion data. Similar mixtures of NaF-ZrF,-UF, (53.5-40-6.5 mole %) were circulated in all these loops. The operating data provided by Experi- mental Engineering and the corrosion data now available are presented in Table 6.6. Despite the turbulent flow and the increase in the number of cycles, the depths of attack found in these loops are only two to three times the depths found in thermal-convection loops operated for comparable periods. A typical hot-leg section from loop 4690 is shown in Fig. 6.5. The data are still too meager to permit any conclusions to be drawn, but several results are worth noting. A large difference was found between the depth of attack on the inside of the bends and that on the outside. Where a depth of attack on the inside of a bend was 16 mils, directly opposite the attack would be much lighter in intensity and to a depth of only 5 mils. The inside and outside surfaces of a bend from loop 4690 are shown in Fig. 6.6. In loops 4696-C and 4694 differences in wall PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED 7-7130 X ¢ e . e ® e i * & 2 . o ;,".1 h i » - .. / ¢ .//’V_, % \\" . T - > ‘\ g LA e o v o . . ,, ‘ :~p & .4 o » - . » .. . - .O & o - - s e 2 - z Fig. 6.5. Hot Leg of Forced-Circulation Inconel Loop 4690 After Circulating NaF-ZrF ,-UF . 250X, Reduced 31%. thickness in the bends of from 42 to 47 mils were found; but, as yet, no large differences have been found in loop 4690. These loops were electrical- resistance heated, and therefore more current and higher temperatures existed where the walls were thicker on the inside of the bends. |n loop 4694 differences were found thickness in straight sections, and corresponding differences in attack were noted. It has been shown (cf., sec., 8, '"Heat Transfer and Physical Properties'”) that variations in flow also occur in the bends that would cause higher temperatures on the inside. in wall A complete bend from loop 4694 was sectioned longitudinally and examined. The depth of attack was the same on both sides of the straight section before the bend, but the difference increased very rapidly as soon as any curvature was noted. The difference remained approximately the same all around the bend. At the exit end the difference did not stop suddenly but continued into the straight area beyond. The depth gradually de- creased from the end of the curvature but some difference was still present several inches beyond the bend. This pattern of attack was predicted by the flow studies. These differences seem to show that mass-transfer is even more temperature sensi- tive than it was thought to be. One other dis- crepancy noted in the data was that in loops 4696-C and 4694 the deepest attack occurred in the first leg of the heater rather than in the second leg where the maximum temperature was supposed 75 ANP PROJECT PROGRESS REPORT TABLE 6.6, DATA ON OPERATION AND CORROSION OF INCONEL FORCED-CIRCULATION CORROSION AND MASS TRANSFER TESTING LOOPS Loop Designation 4690 4696-C 4694 Operating Dota Time, hr 774 625 519 Maximum fluoride temperature, °F 1500 1500 1500 Maximum wall temperature, °F 1740 1610 ' 1580 Temperature drop, °F 200 300 200 Approximate Reynolds number 10,000 10,000 15,000 Heater length First section, ft 4 6 5 Second section, ft 4 8 7 Loop length, ft 48 54 52 Velocity, fps 6.67 5.94 9.19 Cause of termination Leak Pump bearing failure Leak Corrosion Data Maximum depth of attack, mils First heater section Straight 12 6 16* Bend 12 15 20 Second heater section Straight 15 7 8 Bend 17 11 11 *This straight section showed variations in wall thickness. & * - ~ -;‘ i .- a8, '.,% ¢ . - . - . ™. .». o@* - ‘ * * - # s q - ' L v o .- 1 . . b . . e 5 * o b4 . o . o ' O i x . & el ."""‘0 \ - ¥ (@) ‘ . - (B) 4 - Fig. 6.6. Inner (a) and Outer (b) Surfaces of a Bend from Forced-Circulation Inconel Loop 4690 After Circulating NoF-ZrF ,-UF,. 250X. Reduced 23%. 76 to have occurred. However, the two heater legs were found not to be of equal length, and therefore the power input was higher in the first leg than in the second. Chemical analysis data have been received for loop 4690. The batch analysis was reported to contain 14.8 wt % U, 7 ppm Ni, 45 ppm Cr, and 60 ppm Fe. Two samples taken from the sump tank after the loop was drained show the following: 15.1 and 12.2 wt % U, 100 and 40 ppm Ni, 690 ppm Cr in both samples, and 115 ppm Fe in both samples. The only cold-leg deposits found so far were in loop 4696-C. Metallographically, a deposit, as yet unidentified, was found in a single section near the end of the cooling coil of this loop. In loop 4690 two deposits were found in the pump that were not present in either of the other two loops. There was a deposit of loose magnetic crystals on the back of the impelier. These crystals ana- lyzed 4.52 wt % Ni, 9.54 wt % Cr, and 1.18 wt % Fe or (corrected to 100 % metal) 29.6 wt % Ni, 62.7 wt % Cr, and 7.7 wt % Fe. The second deposit was at the liquid level on the sleeve around the pump shaft. This deposit was not discrete crys- tals, and, under the microscope, it was shown to be a thin film of metal crystals. A diffraction examination reported the deposit to be a mixture of nickel and slightly altered fluorides. Two different spectrographic samples were submitted, and the resuits, which surrounding fluoride follow, gave the first indication of nickel mass transfer, Sample 1: Metal plus Fluoride Mixture Cr 3 ppm Fe 1 ppm Ni Sppm Co 0.1 ppm Na Trace U Trace Zr Trace Sample 1: Fluoride Mixture Cr 2 ppm Fe <0.02 ppm Ni <0.05 ppm Co <0.05 ppm Na Trace U Trace Zr Trace PERIOD ENDING MARCH 10, 1955 Sample 2: Magnetiec Phase Ni >5 ppm Cr <0.4 ppm Fe >10 ppm Sample 2: MNonmagnetic Phase Ni <0.05 ppm Cr >5 ppm Fe <0.02 ppm GENERAL CORROSION STUDIES E. E. Hoffman W. H. Cook C. F. Leitten, Jr. Metallurgy Division High-Temperature Tests of Molybdenum in Contact with NoF-ZrF ,-UF An attempt is being made to devise protection for the cooling jacket of the fused salt—Inconel in-pile forced-circulation loop experiment which is scheduled for insertion in the MTR. A loop failure would undoubtedly occur in the vicinity of the nose of the loop if pumping of the fused salt were interrupted for a short period. It is possible that in the event of a pump failure, the fused salt might reach a temperature as high as 2430°F, which has been given as the approximate boiling point of the fluoride fuel mixture to be used. It has been proposed that a sheath of molybdenum around the nose of the loop would afford sufficient protection for the cooling jacket, and it was arbi- trarily concluded that 30 min at 2430°F would be more than adequate to determine the suitability of A molybdenum specimen was therefore placed in a molybdenum container, and the container was then filled with NaF-ZrF,-UF, (53.5-40-6.5 mole %). In one test a molybdenum plug was welded into the top of the capsule, while in a second test, the top was left open. molybdenum for this service. These test containers were sealed first in Hastelloy B and then in quartz to prevent oxi- dation of the molybdenum. There was no weight change of either specimen during the test, and metaliographic examination showed no attack (Fig. 6.7). However, the effects of alloying be- tween the molybdenum and Hastelloy B capsules may be seen in Fig. 6.8. 77 ANP PROJECT PROGRESS REPORT Brazing Alloys on Inconel and Stainless Steel in Sodium and in Fuel Mixtures Static corrosion tests have been completed on a series of T-joints which were submitted by the Wall Colmonoy Corporation. The purpose of these tests was to find a brazing alloy which had good corrosion resistance to both sodium and the fuel INCH UNCLASSIFIED ¥-14263 0.003 Fig. 6.7. Surface of Molybdenum Specimen After Exposure for 30 min at 2430°F to NaF.ZrF - UF,. Etched with NH,OH + H,0,. 500X. Re- duced 42.5%. mixture NaF-ZrF4-UF4 (53.5-40-6.5 mole %). Several of the brazing alloys listed in Table 6.7 as having been tested on Inconel have also been tested on type 304 stainless steel. The results Metallo- graphic results indicate that only the brazing ailoys 1-10 and B-11 of this series have fair cor- rosion resistance to both sodium and to the fused salt fuel mixture. An Inconel T-joint brazed with alloy B-11 (10.8% P-9.2% Si—-80% Ni) is shown in Fig. 6.9 in the as-received condition. may be seen along the fillet surface of this brazing alloy. Figure 6.10 shows brazing alloy B-11 after exposure to NaF.ZrF -UF, (53.5-40-6.5 mole %) for 100 hr at 1500°F; only nonuniform attack to a depth of 1 mil is noticeable along the fillet surface. Figure 6.11 shows the same brazing alloy after exposure to sodium for 100 hr at 1500°F, The very erratic surface attack to a depth of 4 mils can be seen, along with several subsurface voids to a depth of 9 mils. A brazing alloy such as B-11 might be used in a heat exchange system as a back-braze material where the surface of the braze alloy would normally be exposed to the of these tests were reported previously.2 Several small voids 2g, E, Hoffman, W. H, Cook, and C. F. Leitten, ANP Quar, Prog. Rep. Dec. 10, 1954, ORNL-1816, p 80. UNCL ASSIFIED Y-14017 Fig. 6.8, Molybdenum-Hasteiloy B Capsule After Exposure for 30 min at 2430°F to NaF-ZrF -UF ,. Bottom of capsule at left. Note alloying of inner molybdenum capsule with outer Hastelloy B container. 78 PERIOD ENDING MARCH 10, 1955 TABLE 6.7. RESULTS OF STATIC TESTS OF BRAZING ALLOYS IN SODIUM AND IN NuF-ZrF4-UF4 (53.5-40-6.5 mole %) AT 1500°F FOR 100 hr Alloy Alloy Weicht Ch Designa- Composition Joint Material Bath eight Lhange Metallographic Notes tion (wt %) (g) (%) A-10 12 P—88 Ni Type 304 stain- Sodium -0.0015 ~0.141 Attack along entire fillet surface less steel to a depth of & mils Flueride -0.0031 -0.30 No attack on surface of fillet mixture A-10 12 P88 Ni Inconel Sodium —0.0002 -0.028 Attack on entire fillet syrface to a depth of 13 mils Fluoride ~-0.0008 -0.108 Surface attack along fillet to a mixture depth of 0.5 mil B-11 10.8 P-9,2 Inconel Sodium ~0.0010 —-0.098 Erratic surface attack to a depth S5i—80 Ni of 4 mils, subsurface voids to a depth of 9 mils Fluoride -0.0018 —-0.166 Fillet surface attacked to a mixture depth of 1 mil E-11 13 Si-87 Ni Type 304 stain~ Sodium -~0.0007 —-0.068 No attack present along fillet less steel Fluoride -0.0036 -0.358 Fillet completely attacked mixture F-11 9 Siw17.8 Cr— Type 304 stain- Sodium 0.0 0.0 No attack on fillet 73.2 Ni less steel Flueride -0.0052 -0.52 Attack on braze joint to o depth mixture of & mils H-10 10 P-4.3 Mo— Inconel Soedium -0.0011 ~0.107 Uniform surface attack along 85.7 Ni fillet to a depth of 9 mils Fluoride —0.0016 —-0.154 Attack along fillet surface to a mixture depth of 0.5 mil I-10 11.6 P=6.25 Inconel Sodium ~0.0014 -0.135 Erratic attack along fillet to a Mn—-82.2 Ni depth of 4 mils Flueride ~0.0016 —0.157 Attack along fillet surface to a mixture depth of 0.5 mil fluoride mixture and would be in contact with sodium only in the event of a tube-to-header weld failure. Dissimilar Metal Mass Transfer in the System Zirconium=Type 304 Stainless Steel-Sodium The problem of dissimilar metal mass transfer of zirconium to type 304 stainless steel in sodium has been studied in seesaw tests in the tempera- ture range 1000 to 1500°F. The test containers were capsules made of type 304 stainless steel, and the zirconium specimens were retained in the hot zone by partially crimping the capsule wall. Each specimen of zirconium was carefully cleaned and weighed both before and after testing in order to obtain weight-change data. Upon completion of the test, each capsule was sectioned and samples of each section were submitted for metallographic and spectrographic analysis. Six samples were cut from each container that taken together were representative of all portions of the container. Table 6.8 gives the conditions of the tests. In all three tests a seesaw furnace with a speed of 1 cpm was used, and the duration of each test was 100 hr. 79 ANP PROJECT PROGRESS REPORT Fig. 6.9. " UNCLASSIFIED Y382 0.05 INCH 0.040 0.015 0.020 0.025 Inconel T-Joint Brazed with 10.8% P-9.2% Si=80% Ni in the As-Brazed Condition. Note small voids at fillet surface. Etched with glyceria regia. 150X. [ UNCLASSIFIED - Y1360 0.015 0.020 A g L fx? L £ L T et Fig. 6,10, Inconel T-Joint Brazed with 10.8% P-9.2% Si-80% Ni After Exposure to Static NaF-ZsF ,-UF, ot 1500°F for 100 hr. Note slight attack along fillet surface. Etched with aqua regia, 150X. Reduced 39%. 80 Rk 2l 0 UNCLASSIFIED Y8 L in.0s NICKEL PLATE. INCH 0.010 0.045 0.020 0.025 Fig. 6.11, P -9.2% $i~80% Ni After Exposure to Static Sodium for 100 hr at 1500°F. Note attack at fillet surface and subsurface voids. Etched with glyceria regia. 150X. Reduced 37%. Inconel T-Joint Brozed with 10.8% TABLE 6.8. TEST CONDITIONS FOR STUDYING DISSIMILAR METAL MASS TRANSFER IN THE SYSTEM ZIRCONIUM-TYPE 304 STAINLESS STEEL-SODIUM Hot-Zone Cold-Zone Temperature Test No. Temperature Temperature Gradient (°F) (°F) (°F) 1 1000 , 531 469 2 1200 700 500 3 1500 1030 470 It appeared that no thermal-gradient mass trans- fer occurred in any of the tests, because there was no deposit of zirconium found in the cold zones. The cold-zone section of the type 304 stainless steel capsule used in test No. 3 is shown in Fig. 6.12. Metallographic examination showed a layer of fine particles in the hot-zone section of each capsule. nation of the surface of a specimen sectioned from the hot zone (Fig. 6.13) of the capsule used in test No. 3 showed no trace of zirconium or However, x-ray exami- compounds. The precipitated layer found in the hot zone reached a maximum thickness of 2 mils, and it wos very similar to the layers observed on type 304 stainless steel samples carburized in the presence of sodium. Spectrogrephic analysis revecled only a slight trace of zirconium in the hot-zone section of the capsule used in test No. 2. An increase in zir- conium content was found in the capsule used in test No. 3, but the amount could be considered to be negligible, since it was of the order of 10~ 3%. It was also noted in test No. 3 that a slightly larger concentration of zirconium appeared in the cold zone of the capsule than in the hot zone. cases the zirconium samples gained X-ray analysis revealed zirconium In all weight during the tests. a layer of zirconium oxide on each sampie. The following tabulation gives the weight changes found: T N Zirconium Sample Weight est Re. Change (g) +0.0041 2 +0.0089 3 +0.0150 PERIOD ENDING MARCH 10, 1955 = UNGLASHFIED - ¥-14370 : NICKEL PLATE INCH 7 o008 0.002 0.003 e 10.004 = 10.005 ¢ 0.006 Fig. 6.12, Cold Zone of Type 304 Stainless Steel Capsule After Exposure to Sodium in Seesaw Apparatus for 100 hr at 1500°F, Etched with 10% oxalic acid. 500X. Reduced 37%. NOLASSIFIED NICKEL. PLATE I e INGH | L o ov . = ‘,“ C e ] ¥ . 0.005 0.005 Fig. 6.13. Hot Zone of Type 304 Stainless Steel Capsule After Exposure to Sodium in Seesaw Apparatus for 100 hr at 1500°F. Etched with 10% oxalic acid. 500X, Reduced 42%, The original weight of each sample was 17 g. |t was apparent that under the conditions of these tests, zirconium exhibits negligible amounts of dissimilar metal transfer to type 304 stainless steel. mass Diffusion of Sodium into Beryllium Tests were run to determine the extent to which sodium penetrates beryllium metal in a beryllium- 81 ANP PROJECT PROGRESS REPORT sodium-inconel static system. The beryliium specimens and the sodium were sealed in Inconel capsules and maintained at 1200 or 1500°F for 1000 hr. Five cuts 10 mils thick were then machined from one surface of each specimen. The sides of the specimen were machined off to a depth of 50 mils to avoid sodiym contamination from the edges. The turnings were carefully collected and submitted for spectrographic sodium analysis. The results of these analyses are presented in Table 6.9. TABLE 6.9. SODIUM CONCENTRATION IN BERYLLIUM TURNINGS AFTER EXPOSURE OF BERYLLIUM SPECIMEN TO MOLTEN SODIUM FOR 1000 hr IN AN INCONEL CAPSULE Depth of Sodium Concentration of Beryllium Layers Be Layer (mg of Na/g of Be) (mils) After Test at 1200°F After Test ot 1500°F 0to 10 0.2 135 10 te 20 0.2 29.4 20 to 30 0.6 30 to 40 0.3 0.1 40 to 50 0.02 0.6* *This opparent increase in sodium content is not con- sidered to be significant. It appears from the data in Table 6.9 that very little penetration of beryllium by sodium will occur at a temperature of 1200°F. As may be seen in Fig. 6.14, the beryllium specimen in the test at a temperature of 1200°F was attacked irregularly to a maximum depth of 5 mils. In the test at 1500°F the specimen was very heavily attacked to a maximum depth of 20 mils (Fig. 6.15), and a 3- to 4-mil porous metallic layer covered the surface of the beryllium specimen. This layer is anisotropic and therefore has been identified as either beryllium metal or a beryllium-rich beryllium-nickel solid solution. No surface layers could be found on the walls of the Inconel capsules used in these tests; however, there was quite a bit of fine precipitate along the surface to a depth of 2 to 3 mils. This precipitate may be either BeNi or Be,;Ni, particles. Beryllium-Inconel Spacer Tests Tests revealed previously3 that dissimilar metal mass transfer of beryllium metal to Inconel across 82 LNCLASSIFIED Y-14189 Fig. 6.14- After Exposure to Static Sodium for 1000 hr at Surface of a Beryllium Specimen 1200°F. Large voids are due to attack by sodium, Unetched, 250X. Reduced 35%. small sodium gaps is a serious problem at tempera- tures in excess of 1200°F. Therefore a study is under way to determine the effect of temperature and spacer distance on the alloying of beryllium with Inconel. Layers of the compounds BeNi and Be, ,Ni,, both of which are very hard and brittle, have been found on Inconel plumbing in past tests. Thermal-convection loop tests® at a hot-zone temperature of 1300°F for 1000 hr revealed a Be,Ni; layer opproximately 20 mils thick where an Inconel pipe and a beryllium insert were in direct contact. In the same test in areas where a 6-mil clearance was present between the lnconel and the beryilium, a 0.5-mil layer of the BeNi compound was found on the surface of the Inconel. The tests for determining the optimum spacing between beryllium and Inconel in a sodium environ- ment have been conducted with the sodium static, because the maximum attack on beryllium speci- mens and the only beryllium-nickel compound layers on Inconel in thermal-convection loop tests have been found in areas where the sodium was fairly stagnant. In the tests completed to date, spaces of 0, 5, and 20 mils were used between the Inconel and the beryllium, and the specimens were exposed to sodium for 1000 hr at 1200°F. The appearance of the specimens after testing may be seen in 3G. M. Adamson et al., ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 78. UNCLASSIFIED PERIOD ENDING MARCH 10, 1955 3 UNCLASSIFIED Y-14190 Y-14191 | 0.007 0.008 0009 0.040 0.041 0.012 0.013 0.044 0.045 Fig. 6.15. Surface of a Beryllium Specimen After Exposure to Static Sodium for 1000 hr at 1500°F, (z) Bright-field illumination. (b) Polarized light — layer which formed on the surface is either pure beryllium or a berytlium-rich beryllium-nickel solid solution. Unetched, 250X, Figs. 6.16, 6.17, and 6.18. alloying occurred where the specimens were in As may be seen, direct contact. Metallographic examination of the surface of the Inconel specimen separated from the beryllium by the 5-mil space revealed a maximum of 0.2 mil of beryllium-nickel compound formation. The surface of the Inconel specimen separated from beryllium by the 20-mil space had no beryl- lium-nickel compound layers; however, there was an excessive amount of precipitate in the Inconel grains to a depth of 1 mil. Beryllium and inconel specimens which were held in direct contact during testing are shown in Figs. 6.19 and 6.20, with the 6-mil layer of Be,,Ni, and BeNi which formed on the surface of the Inconel during this 1000-hr test being shown in Fig. 6.20. Spectrographic analyses of drillings from the surface of the Inconel re- vealed a beryllium concentration of 3.49 mg/cmz, which is equivalent to a layer of pure beryllium approximately 1 mil in thickness. The beryllium insert and the Inconel sleeve which surrounded it in a recent thermal-convection BERYLLIUM 20~ MIL SPACE INCONE L 5-MIL SPACE BERYLLIUM Inconel and Beryllium Specimens in Fig. 6.16. Positions Occupied During Exposure to Sodium for 1000 hr at 1200°F. loop test with sodium as the circulated fluid are shown in Fig. 6.21. This test operated for 1500 hr with a hot-zone temperature of 1300°F, and ex- cessive alloying occurred between the sleeve and the beryllium specimen. These surfaces were separated by a space of approximately 6 mils at the beginning of the test. 83 ANP PROJECT PROGRESS REPORT UNCLASSIFIED -13863 -—— BERYLLIUM ~g——— INCONEL Fig. 6.17. Surfaces of Beryllium and Inconel Specimens Separated by the 20-mil Space During Exposure to Sodium for 1000 hr ot 1200°F, BUNCLASSIFIED Y-12864 -«—— INCONEL ~——— BERYLLIUM Fig. 6.18. Specimens Separated by 5-mil Space During Ex- posure to Sodium for 1000 hr at 1200°F, Gray and black areas on beryllium indicate formation of beryllium oxide, while the dark areas on the lnconel indicate Disregard areas of direct contact, Surfaces of Beryllivm and Inconel beryllium-nickel compound formation. Cermets in NuF-ZrF4-U F4 The normally inert and refractory nature of cermets makes them attractive for applications in which solid-phase bonding (*‘self-bonding’’) is a problem. Valve stems and valve seats are especially sensi- tive areas where such bonding cannot be tolerated. VYery clean surfaces promote solid-phase bonding, and in fused fluoride salt systems at elevated 84 UNCLASSIFIED 8 Y-1395 —ag——— BERYLLIUM -a—— |NCONEL Fig. 6.19. Surfaces of Beryllium and Inconel Specimens Held in Contact During Exposure to Static Sodium for 1000 hr at 1200°F. temperatures any surface contaminants are rapidly reduced. Therefore, further evaluations have been made of the corrosion resistance in seesaw appa- ratys of Kennametal, Inc., cermet specimens K150A, K151A, K152B, K162B, and D4675 in contact with NaF-ZrF ,-UF, (53.5-40-6.5 mole %) for 100 hr at a hot-zone temperature of 1500°F. These specimens had previously shown promising corrosion resistance in NaF-ZrF,-UF, (50-46-4 mole %), a somewhat less severe corroding medium, but there were some variations in duplicate tests.? The results of the recent tests are given in Table 6.10, together with the complete compositions of the specimens. Comparisons of the as-received with the tested specimens under a metallographic microscope did not reveal attack on any of the specimens shown in Table 6.10. In consideration of the medium used and the temperature difference between the hot and cold zones, the set of specimens from the tests with the cold-zone temperature of 1230°F should have had the most severe corrosion. The discrepancies between past and present fused salt corrosion attack on these Kennametal specimens are believed to be due to the improved purity of the fused fluoride salts used aond to improved metallographic polishing techniques., The metallo- graphie polishing of cermets to enable high magni- fication examination of edges within accuracies of 1 to 2 mils is very difficult. Special techniques are being developed for polishing these very hard materials. 4E. E. Hoffman et al., ANP Quar. Prog. Rep. June 10, 1954, ORNL-1729, p 68. PERIOD ENDING MARCH 10, 1955 UMNCL ASSIFIED Y-14365 INCH 0.001 0.002 0.003 0.004 0.005. Fig. 6.20. The BeNi and Be,|Ni; Loyers Formed on Inconel Specimen Shown in Fig. 6.19 During Ex- posure to Static Sodium for 1000 hr ot 1200°F. Unetched. 500X, TABLE 6.10. RESULTS OF SEESAW TESTS OF TWO SAMPLES OF EACH OF SEVERAL CERMETS IN CONTACT WITH NaF-ZrF4-UF4 (53.5-40.0-6.5 mole %) FOR 100 hr AT A HOT-ZONE TEMPERATURE OF 1500°F Approximate Sample Cha ngea (%) Sample , Cold-Zone Desi ) Composition (wt %) T esignation emperature Dimensionclb Weight (°F) K150A 80 TiC-10 Ni-10 Nl:;Tc:TiC3 1385 +1.0 + 2.5 1230 +0.6 +0.05 K151A 70 TiC-20 Ni-10 NbTuTiC3 1385 +0.2 +0.1 1230 +0.2 +0.09 K152B 64 TiC-30 Ni~é Nch:TiC3 1385 +0.1 0.0 1230 +1.0 +0.09 K162B 64 TiC=25 Ni-5 Mo-6 NbTaTiC3 1385 +0.1 0.0 1230 +0.1 -0.02 D4675 97.5 WC-2.5 Co 1385 +0.2 +0.1 1230 +Q.1 +0.04 %Values based on figures to four decimal places; the dimensional measurements were made with a bench micrometer and the weights were obtained with a Gram-atic balance. These percentu?e values should not be used to determine relative corrosion resistance of different specimens; they are given here only to show their relative magnitudes. Average percentage changes in thickness, width, and length, 85 ANP PROJECT PROGRESS REPORT BERYLLIUM INSERT —————"> Fig. 6.21. ————— |[NCONEL SLEEVE UNCLASSIFIED Y-14407 Beryllium Insert and Inconel Sleeve from Thermal-Convection Loop in Which Sodium Was Circulated for 1500 hr at a Hot-Zone Temperature of 1300°F. Clearance between Inconel and sleeve was 6 mils. Note alloying of beryllium and Inconel, The as-received and tested specimens of K151A, K162B, and D4675 shown in Figs. 6.22, 6.23, and 6.24 illustrate typical specimen structures and the absence of corrosion. In Figs. 6.22 and 6.23 the TiC particles are the larger and darker colored material and the lighter material is the binding metal. Single-Crystal Specimens of Magnesium Oxide in Lithium and in Lead Single-crystal specimens of magnesium oxide were tested in static lithium and in static lead in Globe- iron containers at 1500°F for 100 hr. These cor- rosion tests supplement those previously reported for single-crystal magnesium oxide tested in sodium and in NaF-ZrF ;-UF (53.5-40-6.5 mole %).° SE. E. Hoffman et al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 94. 86 Two magnesium oxide specimens, nominally, 0.10 by 0.23 by 0.24 in., were cleaved from a piece of synthetic magnesium oxide crystal. The speci- men tested in lithium had a weight loss of 66.4%, and it tock on the shape of the Globeiron container in regions of contact. There was no attack on the magnesium oxide specimen tested in lead. FUNDAMENTAL CORROSION RESEARCH G. P. Smith Metallurgy Division Mass Transfer in Liquid Lead J. V. Catheart Metallurgy Division It was the purpose of this investigation to survey the mass transfer properties in liquid lead of a Y-13585 i % i PERIOD ENDING MARCH 10, 1955 UNCL ASSIFIED | INCH 0.002 UNCLASSIFIED Y-13584 0.003 Fig. 6.22, Cemet K151A (70% LiC~20% Ni=10% NbTaTiC,) As-Received (a) and (b) After Exposure to NaF-Z¢F,-UF, for 100 hr 1230°F in Seesaw Apparatus. Unetched. { UNCLASSIFIED [ TY13586 at a Hot-Zone Temperature of 1500°F and a Cold-Zone Temperature of 1000X. Reduced 5.5%. INCH 0.002 | UNCLASSIFIED yaser b 0.003] 1000 X 0.004 Fig. 6.23. Cemmet K162B (64% TiC~25% Ni-5% Mo=6% NbTaTiC,) As-Received (a) and (b) After Exposure to NaF-ZrF -UF, for 100 hr at a HotZone Temperature of 1500°F and a Cold-Zone Tempera- ture of 1230°F in Seesaw Apparatus. Unetched. 1000X. Reduced 6.5%. 87 ANP PROJECT PROGRESS REPORT UNCLASSIFIED | Y-12679 f INCH 0.002 UNCLASSIFIED Y-12680 0.003 Fig. 6.24. Cermet D4675 (97.5% WC-2.5% Co) As-Received (a) and (b} After Exposure to NaF-ZrF .- UF, for 100 hr at a Hot-Zone Temperature of 1500°F and o Cold-Zone Temperature of 1385°F in Seesaw Apparatus. Etched with KOH-KsFe(CN)é. 1000X. Reduced 2%. variety of metals and alioys with the aim of ob- taining an insight into the roles of various important alloying materials in fixing the resistance to mass transfer of given alloys. Thus pure metals such as chromium, columbium, and nickel were tested in addition to alloys having practical structural im- portance. Work on this project has now been terminated. All tests were performed in small quartz thermal- convection loops, the test specimens being fastened in the hot and cold legs of the loops. In all cases a temperature gradient of about 300°C was main- tained across the loop, and the hot-leg temperature was 800 to 810°C. Details of the construction and operation of the loops, as well as of the experi- mental results, may be found in earlier reporfs.6 The only new material tested during the past quarter was type 310 stainless steel. A loop con- taining this alloy plugged after 65 hr of operation with hot- and cold-leg temperatures of 805 and 500°C, respectively. A transverse section of the ©J. V. Cathcart, ANP Quar. Prog. Reps. ORNL-1439, p 148; ORNL-1771, p 100; ORNL-1816, p 88. 88 hot-leg specimen is shown in Fig, 6.25. The results of this research project are summa- rized in Fig. 6.26. The numbers at the ends of the bars in the chart represent the time period of the test, Except as noted on the chart, every loop was operated until it plugged. The pure metals tested, as a rule, showed rather poor resistance to mass transfer, with the excep- tions being molybdenum and columbium. These {atter two metals were the only materials studied which suffered virtually no mass transfer or cor- rosion under the test conditions. At the other extreme, nickel was found to be highly subject to mass transfer; the nickel loops plugged in about 2 hr. Loops containing chromium and iron plugged in about 100 and 250 hr, respectively, and there- fore can be regarded as being intermediate between molybdenum and columbium on the one hand and nickel on the other with respect to resistance to mass transfer. The alloys studied could be conveniently grouped on the basis of whether their resistance to mass transfer was about the same as or greater than the LEAD—STAINLESS STEE - Pt b ¥ 2. e s T - & . 3T ‘*\ Fy‘ - - 5 s . : CORRODED REG!ON . “ i e . S, ; : . P LS ; Tl T~ ;K . o fi’w # - L #}"f # o # £ - W . + e P ) A . Yo - e S £ » . . S f .‘\5‘ i { ‘ L INTERFACE — + vy UNCLASSIFIED Y-14253 T . L Z r 0.001 T S 7 : 0.002 1 * - “Ey ) 3 . . "' "E."" 3 * \ " . F‘) < sw c 0003 4 \,\ =.>\ . . "_: . ' L S ~: L 3 .. £T ; ' . N i .’ / x 3 vty me ¥ ». # P h * 0.004 ‘f”’/ ) { 4 ! {’L N ial 3 P k'//{’\ ~‘Ab‘. . awfii s 2 &.S’;{' Lo Lo oo " _ o \ . 0.005 i - iw * e :?’ ? \N“\: ur e: :Xf* : "f:'.e N ) ‘ f. "?Vf - | L‘ ) * *fifl 0.006 b L x # * Lt o "; ; -~ O S » F 0 - i, ., ;; &\‘ PERIOD ENDING MARCH 10, 1955 Fig. 6.25. Transverse Section of Type 310 Stainless Steel Specimen Exposed to Liquid Lead in Hot Leg (805°C) of Quartz Thermal-Convection Loop. 500X. Reduced 5%. average of their pure metal components. Examples of the former category are furnished by the 300 series stainless steels, Inconel, and Nichrome V. On the other hand, loops containing the 400 series stainless steels and alloys such as 45% Cr-55% Co and Hastelloy B required much longer times for plugging than would have been predicted on the basis of the data for their pure metal components. The results obtained with the 45% Cr~55% Co alloy were particularly striking, more than 750 hr being required for the plugging of a loop containing this material. Plugging times of 100 and 80 hr were recorded for comparable chromium and cobalt loops, respectively. A survey of the phase diagrams of the alloys tested showed that those alloys having a greater than expected resistance to mass transfer all con- tain either a phase which is an intermetallic com- pound or else the composition is relatively close to that which would produce an intermetallic com- Temperature gradient in the loop was about 300°C, pound. Conversely, for the other alloys, either no intermetallic compounds exist or the compositions are far removed from those of possible compounds, On the basis of the data available, it seems reasonable to conclude that alloys containing an intermetallic compound will, in general, possess a higher resistance to mass transfer than those in which compound formation is not possible. The only exception to this rule among the alloys tested was the 50% Cr-50% Fe alloy. This composition corresponds closely to the theoretical composition for the sigma-phase region inthe iron-chromiumdia- gram, and it was thought that this alloy would pos- sess arelatively great resistance to mass transfer. However, the average plugging time for two such loops was 3% hr. No completely satisfactory expla- nation for the apparent anomaly has been found, but it was thought that on account of the extreme fria- bility of the specimens, the very short plugging times might have been due to @ mechanical failure 89 06 COLUMBIUM MOLYBDENUM TYPE 446 STAINLESS STEEL TYPE 410 STAINLESS STEEL 2% Si-14% Cr-84, Fe HASTELLOY B (5% Fe ~28%, Mo -67 %, Ni) 25% Mo - 75%, Ni 45% Cr=55% Co 50% Mo -50%, Fe 16 /o Ni~37%, Cr-47%, Fe (AUSTENITE AND SIGMA PHASE) =] 380 hr 50 76 Cr-50% Fe (SIGMA PHASE ), 38 hr BERYLLIUM IRON CHROMIUM, 100 hr COBALT, 80 hr TITANIUM, 5 hr NICKEL, 2hr 30 % Ni— 70, Fe 275 hr 16 7o Ni-37 %5 Cr-47 Y Fe (AUSTENITE AND FERRITE) TYPE 347 STAINLESS STEEL, {40 hr TYPE 304 STAINLESS STEEL, 100 hr INCONEL, 90 hr TYPE 310 STAINLESS STEEL, 65 hr NICHROME V, 12 hr 100 200 300 400 300 TIME (hr) ? UNCLASSIFIED ORNL-LR-DWG 5857 (/774 NO MASS TRANSFER EE==] USUALLY LITTLE MASS TRANSFER GROUP { GROUP 2 6Rour 3 [ HEavy mass TRansFER A—EXPERIMENT TERMINATED BEFORE COMPLETE PLUGGING OF LOOPS 194 hr 600 700 800 Fig. 6.26. Summary of Tests of Mass Transfer of Various Materials in Liquid Lead. LYOdIY SSTID0dd LIFrodd dNv of the specimens.’ A topical report covering this research project is being prepared. Mass Transfer and Corrosion in Fused Hydroxides M. E. Steidlitz W. H. Bridges Metallurgy Division Work that has been done during the past yec:lr8 on the mass transfer of nickel in sodium hydroxide has indicated a possibility of finding limiting top temperatures and temperature gradients at which mass transfer will not occur. In order to investi- gate this more thoroughly and to extend the work to other potential container materials, a new cor- rosion test apparatus has been constructed. The new system is essentially a multicontainer duplicate of the single model used in the earlier work. Five steel pots are connected through ap- propriate valving to vacuum, purified helium, and purified hydrogen lines and through a bubbler to the exhaust line which extends up into a hood. The sodium hydroxide is contained in a bucket fabricated of the material being studied. The bucket is hung from the cold finger, which is also made of the test material, The pot, heated by an external furnace, heats the bucket and the hy- droxide, and an air jet, blown down the inside of the tubular cold finger, provides the thermal gradient. The hydroxide is circulated by thermal- convection currents established between the bucket and cold finger. Temperatures are measured on the bucket and inside of the cold finger. A thermo- couple outside the pot controls the furnace. A few pilot runs have been made and scheduied test- ing should begin shortly. Spectrophotometry of Fused Hydroxides C. R. Boston Metallurgy Division Construction of a high-temperature spectropho- tometer for studying the fundamental nature of fused salts has been completed. Approximate absorption spectra have been measured for fused sodium hy- droxide at various temperatures up to 700°C in air with sodium hydroxide at 350°C as a reference. Limitations of the instrument restricted measure- ments to the wave length range 400 to 650 mpu. 7). V. Cathcart, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 91. ) 8w. H. Bridges, Met. Semiann. Prog. Rep. Apr. 10, 1954, ORNL-1727, p 52. PERIOD ENDING MARCH 10, 1955 The experimental procedure consisted briefly of placing a periclase {magnesium oxide crystal) cell containing sodium hydroxide in the light path and at a given temperature measuring the photocell response at various wave lengths relative to the response at 600 my, which is the wave length of maximum response at 350°C. Repetition of this procedure at various temperatures gave a series of curves that intersected at 600 my and, by their general form, gave some indication of how the absorption was changing with temperature. With increasing temperature a rapid increase in absorption appears at the shorter wave lengths, until, at 700°C, the absorption at 400 mu reverses the trend and drops markedly; however, at 440 my, the absorption again increases and results in a sharp absorption peak at 440 mu. The long wave length end of the spectrum does not appear to be significantly influenced by temperature change. These preliminary measurements are approximate and will serve mainly as a guide in designing future apparatus and experiments. Improvements in experimental design are now in progress. These include the fabrication of a deeper cell to prevent the hydroxide from creeping out as it frequently does with the present cell. In ad- dition, auxiliary apparatus will be constructed to permit maintenance of a controiled atmosphere over the melt. Fused Hydroxides as Acid-Base Analog Systems G. P. Smith Metallurgy Division In the preceding quarterly’ an outline was given of the application of acid-base analog concepts in a qualitative way to reactions in fused hydroxide systems. Some aspects of this application have now been considered quantitatively. It was previ- ously shown that acid-base analog reactions in fused hydroxide mediums were controlled by the following anionic self-decomposition reaction, 20H= = H,0 + 07~ As previously shown, the equilibrium constant for this reaction is determined in part by the cations which do not enter into the stoichiometry of the reaction, that is, the alkali and alkaline earth metal cations, because of the ability of these cations to polarize the various species in the %. p. Smith, ANP Quar. Prog. Rep. Dec. 10, 1954, 91 ANP PROJECT PROGRESS REPORT reaction without significantly increasing the cation- anion bond energy through resonance energy due to covalent structures. Expressions have now been derived which show quantitatively the way in which these nonreactive cations control this self-decomposition equilibrium. First, it should be pointed out that fused alkali-metal carbonates and sulfates when acting as reaction media form acid-base systems which are closely analogous to fused alkali-metal hydroxides. The corresponding anionic self-decomposition equilibria are the fol- lowing: co,”" = COo, + 07~ , 0,77 = 30, + 077 This correspondence will be useful inasmuch as the literature contains very much more information concerning reactions in fused carbonates and sulfates thon in fused hydroxides. For a reaction medium, whether hydroxide, car- bonate, or sulfate, containing a single honreactive cationic species, it has been shown thot there exists a linear empirical relationship between the ionic potential of the cation and the self-decompo- sition constant K for the medium. Second, for a reaction medium which contains two or more non- reactive cationic species, ¢ thermodynamic formula has been derived which relates the self-decompo- sition constant K, of the medium to its cationic composition. These two expressions can be com- bined, in principle, to give a complete phenome- nological description of the role of nonreactive cations in acid-base analog reactions in the mediums under consideration. In 1951, Cart- gave simple empirical formulas relating the heats of oxysalt formation to the ionic potentials of the cations based on their corrected univalent radii. However, his relationships are not quite those required. In 1952, Ramberg'' found that there exist arbitrary parameters for each of the alkali and atkaline earth metals which, when plotted against the heats of oxysalt formation per two equivalents, give a straight line for each kind of oxysalt. Empirical Relationships for K ;. ledge'? Ramberg’s work was therefore used as a starting point to obtain the empirical relationships shown in Fig. 6.27. (1;20, H. Cartledge, J. Phys. & Colloid Chem. 55, 248 1}. V14, Ramberg, J. Chem. Phys. 20, 1532 (1952). 92 UNCL ASSIFIED ORNL-LR-DWG 5083 100 | I T | ! \ 1 | Be = | z (B ™o o \ 4 20 40 [30] 80 100 120 140 160 180 AH (kcal) Fige. 6.27. Heats of Self 4,0 + H,0 A quantity of melt containing two gram-atoms of OH is converted into one mole of water in its standard state and solution of oxides. The free energy change accompanying this process defines, in terms of the reaction isotherm, an equilibrium constant which is the self-decomposition constant K, of the melt AOH: AF] = —RT InK In the second process, d A,0 + H,0—ZN,(X)),0 + ENI.’(YI.)O + H20 , and thus the product of the preceding process is separated into component oxides, each in its standard state: — NN o SN S o AFg = XNAP ) o + ENJAFSy jg - EN/F(y 5o 0 ! where AFO(X) o represents the standard free "2 i energy of formation of (X )20, and fi(x) o repre- 2 sents the partial molal free energy of {X.),0 in the solution of oxides A,0. By combining terms 93 ANP PROJECT PROGRESS REPORT in the above equation and applying the definition of activity, there is obtained + 3N’lna AF2 = RT [):Ni In a(Xf)2° ; (Y)O J ’ where @, ) o is the activity of (X,),0 in the so- lution of oxides A.,,O. In the third process, EN(X),0 + EN/(¥,)O + H,0 —> 32N (X,JOH + EN/(Y )OH), , and thus the products of the preceding process are reacted individually with the water to yield the corresponding hydroxides in their standard reference states. For the free energy change for this process, the following expression is obtained upon collecting like terms: AF, = SN, {QAFO(Xi)OH - APy o - AFOHzo] + LN/ [AFO(YZ.)(OH) - AFO(YI-)O - AF°H20 The quantities in brackets define, by means of the reaction isotherm, two sets of equilibrium con- stants, K, and Ki', which are the same as the self- decomposition constants of the constituent hy- droxides, (Xz.)OH and (Yz.)(OH)2, respectively, as defined in an earlier progress report.” Thus, AF, = RT(EN, InK, + 2N In K}) 1 1 1 3 In the fourth process, 22N (X,)JOH + ENI(Y )(OH),— 240H , and thus the products of the preceding process are mixed to form the starting solution. The free energy change is AF = E2Nz’F(Xi)0H + EN;F(Yi)(OH)z - Z2’\71'A1E'@()<2.)OH - EN{AFO(YI.)(OH)Q : Then, proceeding as for the second process, 3 = 2 + ENZ.' In a(yi)(oH)z , 94 where a(Xz.)OH is the activity of (X )OH in the solution of hydroxides AOH. The sum of the free energy changes for the steps of the isothermal cycle must vanish. Therefore, ~AF = AF, + AF; + AF,. Writing these terms out, changing the logarithm base to 10, dividing through by RT, and collecting some of the like terms, give (1) logK, = EI_VFI. log K, + ):1-\71.' log K “r o + 2N log —m8— i a(Yz.)(OH)z %x,),0 + 2N. log ! 2 “(x,)OH where the activity terms are not the activities of the various constituents in a hydroxide melt at equilibrium with its self-decomposition products; they are activities of oxides in a solution con- taining nothing but oxides, and activities of hy- droxides in a solution containing nothing but hydroxides. Equation 1 is a thermodynamically exact formula. Next, it is assumed that the solutions consist of jons and Temkin's rules'> are applied to the activity terms. Equation 1 reduces to (2) log K, = EN—I. log K; + EN; log Kz.’ Yy Yo s i + ZN'T: iog———z—— ; sz.YOH 2 Yx 7o —_ 1 + ENz‘ log 2 2 inVOH where the y are ionic activity coefficients. In particular, Yxu Yy and Yo are the activity coef- 7 7 ficients of Xz.+, Yz.++, and 077, respectively, in a solution of oxides only, and yxz_, sz_, and ¥4 are the activity coefficients of Xz.+, YI.+)r OH™ in a solution of hydroxides only, The notation may be simplified by changing the subscripts for the alkaline earth metals from Yo Yo Ygooito¥ 00 Y ot Y s where m is the number of kinds of alkali metals and 7 is the number of kinds of alkaline earth metals. Thus, , and 2/v _ I yl. }/0 (3) |ong=EN.logK.+ZN.|og-——-—_ , 1 1 1 2/7/ Y Y . OH 15M.. Temkin, Acta Physicochim. U.R.S5.S. 20, 411 (1945). where the summation is from unity to m + n and v is the charge on the cation. Two models of an ideal mixture of electrolytes have been proposed as approximations to reality, one model by Temkin'> and one by Herasymenko. '® If Temkin’s model is a good approximation, then the activity coefficients in Eq. 3 will approximate unity to the same degree. |f Herasymenko's model is a good approximation, then the above activity coefficients will by no means approximate unity or even constants. A considerable volume of compu- tations published since 1945 not only overwhelm- ingly supports Temkin's model as the better of the two, but indicates that Temkin's model is a satis- factory first approximation for a surprising variety of fused electrolytes. Therefore, the equation (4) log Kd = EN—z’ log Kz. Furthermore, Eq. 4 should be a better approximation than the model of an ideal solution on which it is based, inasmuch as the activity coefficients in Eq. 2 occur as ratios of similar terms which probably tend to cancel. is taken as a first approximation. CHEMICAL STUDIES OF CORROSION F. Kertesz Materials Chemistry Division Effect of Chromium Metal Additions on Corrosion of Inconel by CrF, in Molten Fluorides H. J. Buttram R. E. Meadows Materials Chemistry Division The addition of chromic fiuoride causes strong attack of inconel exposed to NaF-ZrF, and NaF-ZrF ,-UF, while chromous fluoride apparently has little effect. Therefore experi- ments have been carried out to determine whether melts, by addition of metailic chromium the chromic ion can be reduced in the melt to the noncorrosive chromous state and thus provide a protective environment for the Inconel. Inconel capsules containing the fluoride mixture and CrF; additions with and without chromium pellets were studied in customary tilting-furnace type of test. The chromium concentration in the melt after the test, when plotted as a function of CtF, added before the test, was found to increase at a faster rate when added metallic chromium was Tép, Herasymenko, Trans. Faraday Soc. 34, 1245 (1938). PERIOD ENDING MARCH 10, 1955 present than when the only chromium available as the metal was that contained in the Inconel capsule walls. Metallographic examination showed strong attack (4 mils) when 3.3 wt % CrF; was odded to NaF-ZrF, (50-50 mole %). When the same experi- ment was run with the addition of chromium pellets, the examination revealed only light to moderate subsurface void formation to a depth of 0.5 mil. Therefore the reaction Cr + CrF3—> :.’:Crl'_'2 probably takes place with the chromium peilets and protects the Inconel surface. Protective Action of Small Chromium Metal Additions H. J. Buttram R. E. Meadows Materials Chemistry Division placed in Inconel capsules containing 30-g quantities of NaF-ZrF, (50-50 mole %) or NaF-ZrF,-UF, (53.5- 40-6.5 mole %) to determine the protective effect of small chromium additions. The thicknesses of plating utilized were 1, 2, and 3 miis, which, on 6-in.-long, 40-mil, copper wire, introduced 32, 107, and 213 mg of chromium. After tilting-furnace tests, analyses of the capsule contents revealed very low concentrations of copper (20 to 30 ppm) in all the mixtures. Therefore, it is reasonable to assume that even in the experiment in which the l-mil-coated wire was used, all the chromium was not used. These small additions were sufficient to limit the attack on the Inconel to a depth of 0.5 mil. Chromium-plated copper wire was Effect of Fission Products H. J. Buttram R. E. Meadows Materials Chemistry Division The effect of simulated fission products on the corrosion of Inconel in fluoride melts was reported previously.17 A repeat experiment in which commercial-grade YF., was used as an oddition to NoF-ZrF4-UF3 (53.5-40-6.5 mole %) tested in a tilting furnace revealed that less than stoichio- metric amounts of chromium were released from the Inconel by the YF,. This is not surprising, since consideration of the free energies of the compounds involved indicates that no appreciable 174, J. Buttram and R. E. Meadows, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 109. 935 ANP PROJECT PROGRESS REPORT reaction of YF, (or other rare earth fluorides) should occur. Time Dependence of Corrosion in Tilting-Furnace Tests H. J. Buttram N. V. Smith Materials Chemistry Division Attempts to correlate corrosion data obtained from tilting-furnace tests with those obtained from thermal-convection loops have included some experiments in which Inconel capsules filled with NaF-ZrF,-UF, (53.5-40-6.5 mole %) were cycled (4 cpm) at 800°C on the hot end and approximately 600°C on the cold end for 25, 100, 1000, and 2000 hr. Chemical analyses did not reveal any substantial changes in amounts of chromium re- moved from the capsule walls as a function of time. Metallographic examination showed nearly identical attack of 1 to 2 mils, with some evidence that attack was more intergranular in the longer exposures. It seems apparent from these data that only the reactions of chromium with the » equilibration according to the reaction “impurities’’ and 96 Cr + 2UF, = 2UF,; + CrF, are important in this type of test. Mass transfer in tilting-furnace tests with these materials is, apparently, so slight as to be imperceptible. Effect of Yalence State of Iron on Corrosion of Inconel by Fluoride Mixtures H. J. Buttram R. E. Meadows Materials Chemistry Division The effect of the valence state of iron on the corrosion of Inconel by fluoride mixtures was studied by adding FeF, and FeF; to Inconel capsules containing NaF-ZrF, -UF, (53.5-40- 6.5 mole %) and running the customary tilting- furnace tests. The tests showed that the amount of chromium removed from the Inconel was nearly constant regardless of the quantity of FeF, added, while it increased as a direct function of size of the FeF, additions. Metallographic observations agreed with the chemical analyses; the depth of attack did not exceed 2.5 mils with any gquantity of FeF, tested (up to 6 wt %), while the FeF, was found to cause heavy subsurface void for- mation to a depth of 8 mils. PERIOD ENDING MARCH 10, 1955 7. METALLURGY AND CERAMICS W. D. Manly J. M. Warde Metallurgy Division Investigations were continued in the study of the properties of nickel-base alloys containing 15 to 32% molybdenum, ternary alloys with a nickel-molybdenum base, and vacuum-melted Hastelloy B. The studies include fabrication experiments, tests of oxidation and oxidation protection, stress-rupture examinations, and the development of techniques for welding and brazing the materials. An extensive study of the effects of high-temperature aging on the microstructure and physical properties of these materials is also under way. Stress-rupture design curves are presented for Inconel tested in fused salts and in argon, for comparison, at 1300°F and at 1500°F. Preliminary data obtained at 1650°F are also given. A summary of the results of tests of the oxidation resistance of dry-hydrogen-brazed inconel T-joints is presented. The tests made to date indicate that most of the nickel and nickel-chromium-base brazing alloys are suitable for service in an oxidizing atmosphere at 1500°F, and several are suitable at 1700°F, Work also continues on the development of fabricational techniques and their application to the production of radiators, heat exchangers, and other special items, DEVELOPMENT OF NICKEL-MOLYBDENUM BASE ALLQYS H. Inouye J. H. Coobs Metallurgy Division M. R. D'Amore Pratt & Whitney Aircraft Fabrication Studies Hastelloy B. The variables affecting the quality of extrusions of billets of Hastelloy B were studied. Hastelioy B melting-stock pellets were used to prepare the forged billets supplied by the Haynes Stellite Company. The Hastelloy B pellets were vacuum melted and cast at pressures of 6 to 20 p Hg. The vacuum-melting technique was used to eliminate volatile constituents in the ]H. Inouye and J. H. Coobs, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 100, commercial pellets in an effort to improve the hot working characteristics of the alloy. Neutron- activation, spectroscopic, and vacuum-fusion methods of analysis of the alloy revealed that the trace elements had not been effectively removed. The alloys tested were vacuum-melted Hastelloy B, vacuum-melted Hastelloy B plus 0.1% cerium, airmelted and forged Hastelloy B, air-melted Hastelloy B deoxidized with a titanium-nickel- aluminum-manganese master alloy, and vacvum- melted Hastelloy B decarburized with FeO and deoxidized with calcium, In the tests of these alloys the effects of homogeneizing the cast billet, of various extrusion temperatures, of several die designs, and of canning were studied. The experiments have shown that temperature is the variable that affects the extrudability of the alloys. Consistently high recovery of sound metal is obtained at an extrusion temperature of 2000°F, Both tubing and rod have been made with fair success, showed some promise and will be continued, since, as shown in Canning experiments the previous repor'r,1 higher extrusion ratios are attainable with the canned material because higher temperatures are permissible. The importance of increasing the deformation is indicated, since during tube reducing ot room temperature the tube blank made from a forged billet was reduced suc- cessfully, while the tube blank extruded from a cast billet fractured with no recovery of sound metal. Attempts to hot roll the extruded rods to sheet were unsuccessful because of edge cracking and center splitting during rolling. Center splitting is characterized by the formation of two pieces of approximately the same thickness while the hot billet is being rolled. Rolling temperatures be- tween 1950 and 2200°F were investigated at reductions of 5 to 10% per pass. The alloy was reheated in air between each pass, Hastelloy B + 0.03% Ce. experiments described above, melts of Hastelloy B were being made. One ingot was made with a 0.03% cerium addition in the form of an 11% Ce—89% Al master alloy. Spectro- graphic analysis revealed the residual cerium Concurrent with the smaller vacuum 97 ANP PROJECT PROGRESS REPORT TABLE 7, 1. HASTELLOY B PLUS 0.03% CERIUM (NOMINAL) IN AIR Specimens annealed 1 hr at 2100°F in hydrogen Strain rate — 0.05 in./min ELEVATED TEMPERATURE TENSILE TESTS OF ) Test Yield Point, ] | ) Specimen Temperature 0.2% Offset Tensile Strength Elongation Number ©F) (psi) {psi) (% in 2 in.) vVT-9-9 1100 58,700 108,500 20 vT-9-8 1300 49,700 79,400 14 VT-9-7 1500 50,200 58,400 15 VT-9-6 1600 41,100 44,200 30 VYT-9-5 1650 41,800 43,000 38 TABLE 7.2, ROOM TEMPERATURE BEND TESTS OF 0.065-in.-SHEET HASTELLOY B PLUS 0.03% CERIUM Specimen annealed 1 hr at 2100°F in hydrogen Strainrate — 1 in./min Time Aged at 1500°F Bend Angle Results (hr) (deg) 0 180 Broke, bent on itself 5 89 Did not break 50 103 Did not break 100 103 Did not break 200 85 Broke 500 106 Did not break content to be 57 x 1073%. The ingot was successfully hot rolled at 2100°F, with a moderate amount of edge cracking, and sufficient material was obtained for some tensile and bend tests. The results of these tests are given in Tables 7.1 and 7.2, The bend angles listed in Table 7.2 indicate that the alloy tested does not show brittleness after aging at 1500°F, The ductilities of vacuum-melted Hastelloy B contdining cerium are compared in Table 7.3 with the average values for commercial Hastelloy B, as reported by Haynes Stellite Company. Impact tests of commercial Hastelloy B were also made at room temperature to determine the sensitivity 98 TABLE 7.3. COMPARISON OF THE DUCTILITIES OF COMMERCIAL HASTELLOY B AND VACUUM-MELTED HASTELLOY B CONTAINING 0.03% CERIUM Elongation (%) Test Yacuum-Melted Temperature Commercial CF - llov B Hastelloy B astelloy Containing Cerium 1100 22 20 1300 9 14 1500 13 15 1600 17 30 1650 18 38 of the test in showing the effects of aging. The results are presented in Table 7.4. Impact values generally decrease with aging time, and the em- brittling effect of aging appears more pronounced at 1300°F, 20% Mo-80% Ni. During this period, two 30-1b vacuum melts and six extrusions of tube blanks and rod were made of the 20% Mo-80% Ni alloy. The tube blanks were drawn to 0.500.in.-dig, 0.055-in.-wall tubing and used for the fabrication of three standard size thermal-convection loops. The extrusion data were reported previously. ! The extruded rods were rolled to sheet at 2100°F, and the reductions per pass were as hign as 30%. The alloy showed no tendency toward edge cracking, and no deleterious effects of aging were evident from room-temperature tensile tests. The tensile test data are tabulated in Table 7.5, Elevated-temperature tensile tests were per- formed with the alloy in the solution-annealed condition and after aging for 500 hr at the testing temperature. The ductilities of the specimens at elevated temperature were disappointingly small compared with the elongations obtained in room- temperature tests. The elevated-temperature data for this alloy are presented in Table 7.6. The oxidation rate, weldability characteristics, creep- rupture data, work hardening, and recrystallization temperatures for this alloy were reported previ- ously.] 24% Mo-76% Ni, Three 30-Ib vacuum melts of a 24% Mo-76% Ni alloy were prepared, and tubing for three thermal-convection loops and a quantity TABLE 7.4. STANDARD CHARPY IMPACT TEST OF COMMERCIAL HASTELLOY B (R-340, HEAT NO. 1235) AT ROOM TEMPERATURE Impact Values (ft-lb) PERIOD ENDING MARCH 10, 1955 of sheet were obtained from these melts, The fabricability and oxidation rate of this alloy were reported previously,! Room-temperature bend and tensile tests and elevated-temperature tensile and stress-rupture tests have now been completed. Bend tests of specimens aged at 1500°F for periods up to 500 hr showed no deleterious effects from aging at this temperature, Since the alloy is a single-phase at these temperatures, impairment of the ductility would indicate impurity effects, tests showed the embrittling effect of the precipitation of beta phase from the alpha-phase solid solution at 1300°F (Table 7.7). The elevated-temperature tensile properties of this alloy have been de- termined for both the annealed and aged con- Room-temperature tensile ditions. Low ductilities were obtained between 1300 and 1600°F. The data are summarized in Table 7.8, 32% Mo=-68% Ni. The extrusion properties and the oxidation rate of the 32% Mo—68% Ni alloy were described previously.! It is more difficult to extrude this alloy than alloys containing less Aging Time molybdenum, since it is tougher and lower ex- (hr) Aged at 1300°F Aged at 1500°F trusion temperatures must be used because it is - hot short, The alloyis readily hot rolled ot 2100°F 5}’2 112 and at room temperature, Sufficient tubing for one 25 38 49 thermal-convection loop for corrosion testing and 50 43 a quantity of sheet for mechanical property testing were prepared. 100 43 Room-temperature bend tests show that aging is 150 32.5 rapid at 1500°F and that the alloy is character- 200 0.5 istically brittle after aging times greater than ' about 50 hr, as shown in Table 7.9. Plates of 500 17 35 beta plus gamma phase are found in this alloy 1000 29 after aging from the solution-annealed condition. Aging of a 72% cold-rolled alloy at 1500°F shows TABLE 7.5. ROOM TEMPERATURE TENSILE TESTS OF A 20% Mo—-80% Ni ALLOY Specimen annealed 1 hr at 2100°F in hydrogen and aged 96 hr at temperature Aging Yield Point, Specimen Temperature 0.2% Offset Tensile Strength Elongation Number (°F) (psi) {psi) (% in2in.) DP1-26-1 Unaged 36,900 104,600 62 .2 1100 36,200 106,500 61 -3 1300 39,700 109,800 58 .4 1500 36,500 106,400 63 99 ANP PROJECT PROGRESS REPORT TABLE 7.6. ELEVATED-TEMPERATURE TENSILE TESTS OF A 20% Mo-80% Ni ALLOY Strain rate — 0.05 in./min Speci Test Yield Point, T e S N El . :Iec:;nEn Temperature 0.2% Offset en5|(e -;rengf (Vo'ngzuf'lor; umber ©F) (psi) psi 2 in2 in. Annealed 1 hr at 2100°F in hydrogen DP1-26-29 1100 24,100 53,600 20 -30 1300 21,200 41,700 14 -31 1500 21,800 36,100 8.8 -32 1600 19,400 28,800 7.5 -33 1650 21,000 25,400 9 Aged 500 hr at testing temperature -34 1100 22,500 45,100 1.5 -37 1500 21,500 34,800 6.5 -36 1600 20,900 29,000 7.5 TABLE 7.7. ROOM-TEMPERATURE TENSILE TESTS OF A 24% Mo-76% Ni ALLOY (HEAT NO. VT-11) Specimens annealed 1 hr at 2100°F in hydrogen and aged 284 hr at temperature Aging Yield Point, Tensile El . Temperature 0.2% Offset Strength ongs .ion o _ ) (% in2 in.) (" F) (psi) (psi) Unaged 42,400 109,000 62 1300 88,500 123,000 4 1500 47,800 114,600 47 1650 42,000 110,000 47 equiaxed grains of beta plus gamma phase. The properties of the alloy in this condition are not known, except that the hardness remains high (~500 VPN) up to aging times of 200 hr. Room- temperature tensile tests of the alloy confirm the brittle nature, as indicated in Table 7,10, The elevated-temperature tensile tests of the alloy show strengths comparable to those of Hastelloy B; but the ductilities are low., Notch sensitivity of the alloy was indicated by the rupture of several specimens in the pin hole during 100 tensile tests. Plates spot welded onto the shoulder to strengthen the area of the pin hole did not help; the specimens failed in the heat-affected zone. Adequate ductilities were obtained at 1100°F and at 1650°F (the alloy transforms from beta plus gamma phase to alpha plus delta phase at 1600°F). Data for this alloy in both the so- lutionsannealed condition and after 500 hr of aging at the test temperature are shown in Table 7.11, Cb-Mo-Ni. The columbium-molybdenum-nickel ternary alloy system was partially investigated. The alloys studied thus far contained 2, 5, or 10% columbium and 20% molybdenum; the balance of each alloy was nickel, The 3-lb vacuum melts of the 2 and 5% columbium alloys were readily fabricated into sheet, but the 10% columbium alloy fractured severely upon rolling at 2100°F. The oxidation rate at 1500°F of alloys con- taining 5 and 10% columbium was less than that for the 20% Mo—80% Ni alloy but slightly more than that for Hastelloy B. The scale that formed in oxidation tests was nonadherent. The 2% columbium alloy has not yet been tested. The 2% columbium alloy is a single-phase be- tween 1300 and 2100°F; however, Widmanstatten structure appears in both the 5 and 10% columbium alloys at 1300 and 1500°F. The 5% columbium alloy shows no increase in hardness after 20 hr PERIOD ENDING MARCH 10, 1955 TABLE 7.8. ELEVATED-TEMPERATURE TENSILE TESTS OF A 24% Mo-76% Ni ALLOY Specimen annealed 1 hr at 2100°F in hydrogen Strain rate — 0.05 in./min Specimen Test Yield Point, Tensile Strength Elongation . Tem;z)erature 0.2% O.ffse'r (oon G i (" F) {psi) DP1-25-19 1100 28,300 49,300 17.5 -20 1300 43,800 5 -21 1500 25,100 38,100 6.5 -22 1600 24,000 32,400 7.5 .23 1650 24,800 31,200 .3 -24* 1100 37,800 55,000 " .27* 1500 24,200 38, 100 8 -26* 1600 25,200 35,000 10 *Aged 500 hr at test temperature. TABLE 7.9. ROOM-TEMPERATURE BEND TESTS OF A 32% Mo—68% Ni ALLOY (HEAT NO. YT-10) AFTER AGING AT 1500°F Specimens annealed for 1 hr at 2100°F in hydrogen TABLE 7.10. ROOM-TEMPERATURE TENSILE TESTS OF A 32% Mo-68% Ni ALLOY (HEAT NG, VT-10) Specimen annealed for ]/é hr at 2100°F in hydrogen and aged for 288 hr at temperature Aging Time Bend Angle Hardness Resuire (hr) (deg) (VPN) Unaged 75 233 Did not break 5 27 265 Cracked 50 0 516 Broke 100 0 498 Broke 200 0 594 Broke 500 1.5 509 Broke Aging Yield Point, Tensile El ) Temperature 0.2% Offset Strength (?O.ng;f.lor)] (°F) (psi) (psi) st Unaged 66,400 135,600 33 1300 122,500 1.0 1500 85,200 1.0 1650 121,800 2.5 of aging at either 1300 or 1500°F. The 10% Cb-20% Mo-70% Ni alloy ages af both 1300 and 1500°F, as indicated in Table 7.12, The room-temperature tensile data for the 2 and 5% columbium alloys were determined after aging at temperatures of 1300, 1500, and 1650°F. The results are shown in Tables 7.13 and 7.14. The iow ductility of the 5% columbium specimen tested after aging at 1650°F may be due to hydrogen embrittiement, Several other ternary alloys based on molybde- num and nickel will be screened by creep, oxi- dation, tensile, and fabrication tests. The stress- rupture properties of the nickel-molybdenum and the nickel-molybdenum-columbium alloys are reported below. Cr-Mo-Ni. The effect of chromium additions to nickel-molybdenum base alloys is also being investigated as a means of improving the corrosion resistance of these alloys in sodium. [f small chromium additions {up to 5%} are effective, these alloys may also be useful in the fluoride salts, since the small percentages of chromium involved would not cause appreciable chromium mass trans- fer. Also, the oxidation resistance, strength, and 101 ANP PROJECT PROGRESS REPORT TABLE 7,11, ELEVATED-TEMPERATURE TENSILE TESTS OF A 32% Mo~-68% Ni ALLOY (HEAT NO, DP1-24) Specimens annealed 1 hr at 2100°F in hydrogen Test Yield Point, Tensile Elongation Temperature 0.2% Offset Strength (% in2 in) (°F {psi) (psi) 1100 44,800 113,400 32.5 1300 51,200 77,700* 1500 59,700* 1600 35,200 37,300* 1.0 1650 30,700 39,400* 11.3 1650** 30,400 41,800 17.0 1100 *** 53,700 85,500 18.8 150Q*** 82,700 1.3 1600*** 61,500 2.0 TABLE 7.13. ROOM-TEMPERATURE TENSILE TESTS OF 2% Cb-20% Mo-78% Ni ALLOY (HEAT NO. VYT-12) Specimens annealed 1 hr at 2100°F in hydrogen and aged 284 hr at temperature Aging Yield Point, Tensile £l ' Temperature 0.2% Offset Strength (70?9;?0'; °P (psi) (psi) o tm& n Unaged 38,600 102,500 60.5 1300 39,400 104,600 65.5 1500 37,900 103,800 70.0 TABLE 7.14. ROOM-TEMPERATURE TENSILE TESTS OF 5% Cb=20% Mo-75% Ni ALLOY (HEAT NO, YT.5) Specimens annealed ]/2 hr at 2000°F in hydrogen and aged 284 hr at temperature *Specimens which failed in pin hole, **Rerun of 1650°F specimen. ***Specimens aged 500 hr at testing temperature. TABLE 7,12. AGING EFFECTS IN A 10% Cb-20% Mo-70% Ni ALLOY (HEAT NO. AC.4) Specimens annealed 1 hr at 2100°F in hydrogen Hardness Mumber (VPN) Aging Yield Point, Tensile ' Temperature 0.2% Offset Strength Elolngat.uon °F) (psi) (ps) (Pin2in) Unaged 37,800 126,000 61.3 1300~ 128,500 2.5 1500 89,500 152,000 18.8 1650 65,200 102,500 6.5 Aging Time (hr) Aged at 1300°F Aged at 1500°F Unaged 193 5 332 216 25 353 2565 50 353 286 130 371 329 500 N 341 ductility of the nickel-molybdenum base alloys at temperatures above 1500°F may be improved by alloying with chromium, Fourteen 30-lb vacuum-melted ingots of chro- mium-molybdenum-nickel alloys have been ob- tained, These alloys all have a nicke! base, and they contain 20% molybdenum; the chromium con- tent varies from 3 to 10%. All the ingots have been machined into 3- by 3-in. extrusion billets. 102 *In vacuum. Thirteen of the 14 alloy compositions have been extruded into tube and/or rod. The extruded tubes are being reduced to 0.50 in. in outside diameter with 0.055-in. walls for use in thermal-convection loops, The extruded rods are being rolled to 0.065-in.-thick sheet. The sheet is to be machined into test specimens for determining strength prop- erties of the alloy. The alloy compositions and extrusion data are tabulated in Table 7,15, The extruded rods showed severe edge cracking during hot rolling at temperatures between 2000 and 2250°F. A chemical analysis of the electro- lytic chromium used in the melting of these alloys revealed an oxygen content of 0.45%. Therefore to investigate the role of oxygen contamination of the chromium on the fabricability of the alloys, 100-g arc-melt buttons containing high purity TABLE 7.15. PERIOD ENDING MARCH 10, 1955 ALLOY COMPOSITIONS AND EXTRUSION DATA FOR CHROMIUM-MOL YBDENUM-NICKEL ALLOYS Atmosphere: Houghton's salt bath No. 1550 Heating Time: tube blanks, 30 min; rod, 45 min Billets: 3 in. in diameter and 3 in. long Die: 30- to 45-deg cone, alloy CHW Mandrel: Lubrication: glass wool in container, Fiberglas sleeving on mandrel stem 1-in. straight stem, alloy LPD Dummy Block: Al + graphite for tube extrusions; 70-3Q brass + graphite for tube extrusions Pressure Requirements {on 3-in. ram): 70 to 90 tons/in.2 Alloy Composition (wt %) Extrusion Data Soaking Reduction E dabili Cr Mo Ni Other Form No. Ratio Temperature xtrudability (°F) 3 20 77 Tube 1 7.5:1 2200 Good Rod 1 6:1 2200 Good 5 20 75 Tube 2 2300 Good Rod 1 6:1 2200 Good 7 20 73 Tube 2 7.5:1 2250 Poor Rod 1 6:1 2125 Good 10 20 70 Tube 2 7.5:1 2300 Poor Rod 1 6:1 2200 Good 3 20 76.5 6.5 Cb Tube 2 5:1 2200 Goad Rod 1 6:1 2125 Good 5 20 74.5 0.5 Cb Tube 1 7.5:1 2200 Shattered Red 1 6:1 2200 Good 7 20 72.5 0.5 Cb Tube 2 7.5:1 2175 Good Rod 1 6:1 2200 Good 10 20 695 0.5 Cb Tube 2 5:1 2175 One good; one shattered Rod 1 6:1 2200 Goaod 3 20 76.9 0.1 Ce Tube 2 5:1 2200 Shattered 5 20 74.9 0.1 Ce Tube 2 5:1 2200 Shattered 7 20 72.9 0.1 Ce Tube 2 5:1 2250 Shattered Red 1 6:1 2125 Good 10 20 69.9 0.1 Ce Tube 1 5:1 2125 Shattered 5 20 74.75 0.25 Ce Tube 2 7.5:1 2125 Good Rod 1 6:1 2125 Goed 7 20 72.75 0.25 Ce Mot extruded 103 ANP PROJECT PROGRESS REPORT chromium were prepared, The oxygen content of the chromium was 0,011%. The buttons were hot rolled at 2150°F from a thickness of 0.45 in, to a thickness of 0.15 in, The alloy compositions and the results of hotrolling are given inthe following: Hot Rolling Characteristics Alloy Composition {wt %) 3 Cr—-20 Mo-=77 Ni 5 Cr=20 Mo=75 Ni 7 Cr—20 Mo~73 Ni 10 Cr—20 Mo—=70 Ni Slight edge cracking Severe edge cracking No edge cracking No edge cracking The edge cracking during rolling of the alloys containing 3 and 5% Cr was attributed to melting From the results obtained with the two alloys of higher chromium content, it is apparent that the difficulties encountered in hot rolling are due to oxygen contamination of the chromium. Therefore small amounts of cerium and columbium were added to the melts for deoxidation purposes. The columbium additions had no apparent effect, but cerium added in a ratio of approximately 0.1% cerium for each 3% chromium improves the hot rolling properties of these alloys, It has not been definitely determined whether the beneficial effect of cerium on the hot rolling properties is due to a deoxidizing action or to alloying. Arc-melts of chromium-molybdenum-nickel alloys with various percentages of aluminum or titanium added are being prepared for studying the effect of known deoxidizers on these alloys. Many of the alloys tested thus far have shattered during extrusion in the 2200°F temperature range. The shattering is probably caused by too high an extrusion temper- ature or by flaws within the extrusion billet. practice. Oxidation Studies H. Inouye Metallurgy Division M. R. D’ Amore Pratt & Whitney Aircraft Oxidation tests at 1500°F of 168 hr duration have been completed for the following alloys: 3% Cr-20% Mo-77% Ni, 5% Cr-20% Mo-75% Ni, 7% Cr-20% Mo—73% Ni, 10% Cr-20% Mo-70% Ni. The data obtained are plotted in Fig. 7.1, and oxidation curves for commercial and vacuum- melted Hastelloy B are included for comparison. The curve for vacuum-melted Hastelloy B is the 104 UNCLASSIFIED ORNL-LR-DWG 5782 4.0 | i 35— - — — NOMINAL COMPOSITION (wt %) O 30r-20Mo-T7 Ni 3.0 ® S5Cr-20Mo-75Ni _ _ __ _ — A 7 Cr-20 Mo-73 Ni & 5 A 10 Cr-20 Mo-7C Ni s S | O HASTELLOY B-VAGUUM MELTED g B HASTELLOY B-COMMERGIAL = < o l._ x @ o = 0 50 100 150 200 ELAPSED TIME (hr) Fig. 7.1. Base Alloys at 1500°F in Static Air. The Oxidation of Nickel-Molybdenum The data reported for the remaining alloys are the results of single tests. Additional testing is planned to supple- ment the present data. The oxide scale formed on Hastelloy B and the 3% Cr-20% Mo-77% Ni ailoy spalled completely during cooling to room temper- average of two oxidation tests. ature at completion of the test. Moderate spalling of the oxide scale was observed on the alloys containing 5, 7, and 10% chromium. From the limited data it appears that nickel-molybdenum base alloys containing over 5% chromium are more resistant to oxidation than Hastelloy B at 1500°F. STRESS-RUPTURE STUDIES OF NICKEL-MOLYBDENUM BASE ALLOYS R. B. Oliver D. A, Douglas J. H. DeVan J. W. Woods Metallurgy Division An expansion of mechanical testing facilities has recently been accomplished to provide ad- ditional equipment for investigating the strengths of Hastelloy B and related nickel-molybdenum alloys under conditions encountered in a circu- lating-fuel Six new lever-arm type of machines (Fig. 7.2) have been added for testing materials in fused salts, as well as four direct-loading or dead-load type of units (Fig. 7.3) for testing materials in argon, hydrogen, Twelve new tube-burst units (Fig. 7.4) reactor. stress-rupture or air. PERIOD ENDING MARCH 10, 1955 | UNCLASSIFIED | Y3952 Fig. 7.2. LeverrArm Stress-Rupture Machines. 105 ANP PROJECT PROGRESS REPORT - A UNCELASSIFIED a 106 ines, Rupture Mach Dead-Load Stress- 3 ige 7 F PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED Y-13946 Fig. 7.4, Tube Burst Test Unit, 107 ANP PROJECT PROGRESS REPORT were also installed for the study of multiaxial stress systems, In order to provide space for this additional equipment, it was necessary to outfit a new labo- ratory, and thus there was some interruption of the testing program during the past two quarters. In addition to the machines mentioned above, the new laboratory also includes six lever-arm units and four tube-burst units used previously for testing materials in fused salts and liquid metals. Each testing machine and associated furnace has its own temperature-recorder-controller, which is mounted in the central control panel board shown in Fig, 7.5. The results obtained in the Hastelloy B testing program are quite preliminary; however, a series of creep-rupture tests on solution-annealed ma- terial have been completed in an argon atmosphere at 1500 and 1650°F, and a similar series of tests is being run in fused salts. Based on the limited data availeble, the corrosive action of the fused fluorides has not adversely affected the creep- rupture properties at 1500 and 1650°F. Rupture times in the fluorides are compcrable to or greater than those found in argon under similar stress conditions. In all cases the results found for Hastelloy B show marked improvement in strength compared with Inconel for the same temperature conditions. A comparison of the stress-rupture properties found for Hastelloy B, type 316 stain- less steel, and Inconel in argon at 1500°F is presented in Fig. 7.6. In conjunction with the testing of Hastelloy B, tests are being conducted on the modified nickel- molybdenum alloys described above. Stress- rupture results at 1500°F in argon have been ob- tained for the alloys listed in Table 7.16. By comparison with the data presented in Fig. 7.6, it can be seen that these alloys are inferior in TABLE 7.16. STRESS-RUPTURE PROPERTIES OF NICKEL-MOLYBDENUM BASE ALLOYS AT 1500°F IN ARGON Time to Stress {psi) Elongation Rupture °p (%) (hr) Composition (wt %) 78 Ni—20 Mo-2 Cb 8000 133 8 76 Ni-~24 Mo 8000 69 3 5000 358 2 68 Ni—32 Mo 8000 332 8 Fig. 7.5. Central Control Panel. 108 20,000 ~ i ) i HASTELLOY B, ANNEALED AT 2100°F 10,000 5000 STRESS (psi) 2000 1000 10 20 50 100 200 Figo 7.6. Inconel Tested in Argon at 1500 °F, strength to Hastelloy B when tested under similar conditions. The final elongations are also less than those obtained for the commercial Hastelloy B. Tube burst tests, which provide a study of multiaxial stress systems, are presently under way on %-in.-OD Inconel and Hastelloy B tubing with 0.020- to 0.040-in. walls. A ratio of axial- to-tangential stress of 1:2 and a test temperature of 1500°F are being used. The tube is internally pressurized with argon and is surrounded with fluorides. Variable stress ratio tests, described previously,? have been initiated to determine how changes in tangential-to-axial stress ratios affect tube failure. In conjunction with these tests, an analysis will be made of the stress systems which are present in pressurized tubes and of the rate of creep deformation and failure which can be expected to take place as a result of such stress patterns. 2R. B. Oliver et al., ANP Quar. Prog. Rep, Sept. 10, 1954, ORNL-1771, p 112. PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED ORNL—LR~-DWG 5783 TYPE 316 STAINLESS STEEL, AS RECEIVED INCONEL., ANNEALED AT 2050°F 500 1000 2000 5000 10,000 TIME (hr) Stress-Rupture Curves for Sheet Specimens of Hasteiloy B, Type 316 Stainless Steel, and WEL DING AND BRAZING STUDIES OF HASTELLOY B P. Patriarca K. W. Reber R. E. Clausing G. M. Slaughter Metallurgy Division J. M, Cisar Aircraft Reactor Engineering Division R. L. Heestand Pratt & Whitney Aircraft Radiators Additional Hastelloy B radiator test components have been fabricated, as described previously,® to determine the feasibility of using this materidl in applications involving thermal shock and high- temperature oxidation. |t was thought that the 3p. Patriarca et al., ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 103. 109 ANP PROJECT PROGRESS REPORT characteristic aging of Hastelloy B with its ac- companying loss in ductility might result in fis- sures in the tubes or tube-to-header joints when they were subjected to simulated cyclic service. Table 7.17 presents a history of test radiators built to date for evaluating these assemblies under Typical components of a radiator are shown in Fig. 7.7 before welding of the split headers and after completion. Radiator No. 1 (Table 7.17) differed from the other radiators in that only a single row of five tubes was used. The stress distribution onthe tubes in this radiator was not comparable to that on a full-scale radiator, and therefore this test was not run for the full 500-hr life test, The tests of radiators No, 2 and No. 3 differed in that radiator No. 2 was subjected to severdl water quenches from 1500°F to simulate a more severely stressed condition. Radiator No. 3 was several service conditions. carefully examined visually after 12 aqir cools from 1500°F, since this was thought to be a real- istic number of thermal cycles to which an actuadl radiator might be subjected. As can be seen from Fig. 7.8 the fin distortion was not severe enough to cause a material increase in air-pressure drop through the core. A testing temperature of 1200°F was chosen for radiator No. 5 to obtain an experi- mental check of the evidence obtained from micro- structural studies that aging proceeded very rap- idly at this temperature. The test of radiator No, 4 was terminated after 69 hr at 1650°F because the severe, accelerated oxidation caused leaks. (All specimens were leak tested periodically by means of a helium leak detector to determine soundness.,) Radiator No. 4 is shown in Fig. 7.9 after termination of the test. Since the severe localized attack suggested the S, UNCLASSIFIED | E v.13913 Fig. 7.7. Hastelloy B Radiator Before Welding of Split Headers and Completed Test Radiator. 110 TABLE 7.17. PERIOD ENDING MARCH 10, 1955 HISTORY OF HASTELLOY B TEST RADIATORS WITH STAINLESS-STEEL- OR INCONEL-CLAD COPPER HIGH-CONDUCTIVITY FINS Braze Alloy: Coast Metals Alioy No. 52 Test Atmosphere: Air Radiator Test Number 1 2 3 4 5 Number of tubes 5 10 10 10 10 Fin cladding material* Type 310 stainless Type 310 stainless Inconel Inconel Type 310 stainless steel stee! steel Service temperature, °F 1500 1500 1500 1650 1200 Service time, hr 350 500 511 69 500 Number of eycles 8 27 265 8 191 Type of quench Water Water Air Air Air Type of failure Nene Nene None Oxidized None *Fifteen fins per inch. INCHES UNCLASSIFIED Y-13930 Fig. 7.8. Hastelloy B Radiator No. 3 After 12 Air Cools from 1500°F, Note mild fin distortion. 111 ANP PROJECT PROGRESS REPORT L INCHES , UNCLASSIFIED Y-13928 Fig. 7.9. Hastelloy B Radiator No. 4 After 6% hr at 1650°F in Air. influence of a self-fluxing oxide, a series of ex- periments was conducted to investigate the effect of several oxides on the susceptibility of Hastelloy B to this phenomenon., Each specimen was tested for 24 hr at 1650°F and for 100 hr at 1500°F, The results of these experiments are given in Ta- ble 7.18. Very serious accelerated oxidation occurred both with Hastelloy B and Hastelloy C in contact with copper oxide, A comparison of Hastelloy B in contact with copper oxide for 24 hr at 1650°F in air with an uncontaminated control specimen of Hastelloy B tested under the same conditions is shown in Fig. 7.10. Even though the copper oxide was placed only in the area of pitting, which has been brushed clean of scale, the entire surface of the specimen shows a different type of scaling than that on the control specimen. Since the scaling was not observed on specimens exposed 112 to copper oxide at 1500°F, tests were carried out at intermediate temperatures to determine the flow point of the composite oxide, and it was found to have a flow point of around 1525°F, Since the Inconel-clad copper fins were also adversely offected in radiator No, 4, it is sus- pected that a volatile oxide, MoO,, condensed on the fins near the headers with a consequent ac- celeration in the rate of oxidation, As shown in Figs. 7.9 and 7.11, the fins close to the headers were completely oxidized, and the others were only distorted. It is thought that the high velocity air may have carried the volatile oxide away with a consequent reduction in the deleterious effects on the fins farthest from the headers. These experiments illustrate that protection of the copper exposed when the clad- copper high-conductivity fins are punched out is of extreme importance and that brazing techniques complete TABLE 7.18. EFFECTS OF VARIOUS OXIDES ON THE FLUXING OF PERIOD ENDING MARCH 10, 1955 HASTELLOY B AND HASTELLOY C OXIDE Type of Specimen Time (hr) Temperature (°F) Result Hastelloy B + A1203 24 1650 Slight oxidation 100 1500 Slight oxidation Hastelloy B + Cu0O 24 1650 Pitting to 40 mils 100 1500 Pitting to 15 mils Hastelloy C + CuO 24 1650 Pitting to 40 mils 100 1500 Slight oxidation Hastelloy B + SiO 24 1650 Slight oxidation 100 1500 Slight oxidation Hastelloy B + FeO 24 1650 Slight oxidation 100 1500 Slight oxidation Hastelloy B + type 316 stainless steel 24 1650 Slight oxidation 100 1500 Slight oxidation UNCL ASSIFIED Y-14038 Fig. 7.10. (a) Hastelloy B Oxidized at 1650°F. Note characteristic scale. {b) Hastelloy B + CuO Oxidized at 1650°F. Note pitting and change in scale characteristics. 113 ANP PROJECT PROGRESS REPORT UNCLASSIFIED Y-140%94 Fig. 7.11. Finned-Tube Section of Hastelioy B Radiator No. 4. should be developed to ensure complete coverage by the brazing alloy. This problem is further discussed below under the heading ‘‘High-Conduc- tivity Fin Radiator,”’ Welding of Thick Sections The fabrication of a pressure shell from Hast- elloy B will present welding problems of an en- tirely different nature from those encountered in the tube-to-header welding of this material, As a means of evaluating the weldability of the ma- terial and to determine suitable procedures for its fabrication, a heavy-welding investigation is under way in which the metallic arc and the semi- automatic heliarc welding processes are being used, The mefc”ic-orc welds were made on as-re- ceived / by 3- by 4-in. pl afes by using a 90-deg lncluded-ongle bevel, with a /] -in. land, and a /16"” spacing between plates in all cases, except 114 plate No. 4 (Table 7.19). The welds were made with both l/é- and 5/32-in. Hastelloy B coated elec- frodes. All plates were tack-welded to a strong- back to simulate a severely restrained condition. Tests on the plates listed in Table 7.19 con- sisted of 180-deg guided-root and face bends in the as-welded and aged conditions. Tensile specimens were also cut from plate No. 3 for testing in the as-welded and aged condition. All aged specimens were treated at 1500°F for 200 hr, It may be noted that the bend-test results pre- sented in Table 7.19 are somewhat inconsistent. The plate No. 1 failures may be attributed to incomplete fusion in the root pass., Plate No, 2 was welded in much the same manner as plate No. 1, but particular care was used in the de- position of the root pass. All bend tests on this plate were made in the as-welded condition, and no failures were observed., Plate No. 3 was tested PERIOD ENDING MARCH 10, 1955 TABLE 7.19. RESULTS OF BEND TESTS OF WELDED HASTELLOY B PLATES (?'/8 BY 3 BY 6 in.) i Bend Tests b ter of Plate No. Current {amp) El lcmeder ? N;mber of Bevel ectrode (in.) asses Number of Tests Results 1 80 ]/8 Single 3 root Two welds failed 3 face None failed 2 20 ]/8 Single 2 root None failed 1 face None failed 3 100 5/32 Single 1 root None failed 1 face None failed 1 root, aged Failed 1 face, aged Failed 4 A 95 % 7 Double 2 None failed 100 562 7 2 One failed to determine the effect of aging on the bend-test resuits, Several tensile and guided-bend test specimens are shown in Fig. 7.12. It can be seen that speci- mens 4 and 5 are typical as-welded face and root bends, while 6 and 7 are the aged face and root bends, respectively, from plate No. 3. An as- welded tensile specimen and a welded-and-aged tensile specimen were also prepared from plate No. 3. The as-welded specimen (specimen 1 in Fig. 7.12) exhibited good ductility in the parent material but fracturing occurred in the weld. The aged specimen (specimen 2 in Fig. 7.12) also showed good ductility in the parent material, but, in this case, the fracture occurred in the parent metal. One-half of plate No. 4 was prepared with a 1/S-in.-dio coated electrode, while the remaining half was welded with a 5/32-in.-c|ic| coated elec- trode. Radiographic inspection showed that all welds were sound. Two bends from the fillet for which the lé-in.-dia rod was used exhibited good ductility, while one bend from the fillet made by using 5/32-in.-<:|ic1 rod fractured during testing. Inspection of the fracture showed no apparent flaws, The inconsistencies in the bend-test data could possibly be attributed yielding observed in the weid zone. The root-bend to the inhomogeneous specimen (specimen 5 in Fig. 7.12) shows the extent to which such yielding was present. Speci- men 3 in Fig. 7.12 was machined from a double- beveled specimen to determine the effect of bending axially with the weld. Some parting was noticed along the fusion line, but there was no fracture. It is evident that yielding in this longi- tudinal test is more homogeneous than in the transverse bend test. Several weld beads have been prepared with the semiautomatic Aircomatic ma- chine available in the Welding Laboratory. These welds were deposited on }/z—in.-thick Hastelloy B plate with 0.060-in.-dia Hastelloy B wire as the filler metal, experimental In this process, the filler wire is continuously fed from a coiled spool. Conditions have been established which produce good bead- on-plate welds as determined from visual obser- These conditions are listed in the follow- ing to present an indication of the rapid travel speed and deposition rate which can be obtained by this process: arc current, 325 amp; welding speed, 325 in./min; rate of wire feed, 200 in./min; argon consumption, 40 cfth. Metallo- graphic examination of these welds indicated that vation. travel porosity would not be a major problem when welding with the semiautomatic equipment. How- ever, further experiments will be conducted to evaluate the physical properties of these welds in 115 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ©Y-14351 Fig. 7.12. Metallic-Arc Welded Hastelloy B Bend and Tensile Specimens. (1) As-welded — broken in weld zone, (2) welded and aged — broken in parent material, (3) as-welded — bent along axis of welding, (4) as-welded face bend, (5) as-welded root bend, (6) welded-and-aged face bend — broken along fusion zone, {7) welded-and-aged root bend — broken through center of root. order to compare them with the properties of welds made by the conventional, manual, heliarc process. High-Temperature Aging It has been realized that at service temperatures ranging upward from 1000°F copious quantities of precipitated phases would be formed throughout the microstructure of Hastelloy B. Preliminary observations also indicated that the degree of precipitation in weld deposits and in weld heat affected zones might be noticeably different from that found in wrought structures. An investigation has therefore been initiated to determine the effect of several variables on the type and quantity of precipitate, as determined by metallographic examination. An attempt is also being made to 116 correlate the physical properties of the material with the observed microstructures, The variables being investigated in this study include aging temperature, time at this temper- ature, prior thermal history, degree of residual cold deformation, base metal composition, original and aging environment, The relatively homogeneous microstructure of a section of 9’]6-in. Hastelloy B plate after a solution heat treatment of 2 hr at 2150°F is shown in Fig. 7.13. A similar sample after aging for 1500 hr at 1500°F is shown in Fig. 7.14, which illustrates the typical appearance of the precipitate formed at this temperature after extended aging times. A photo- micrograph of the same sample at a higher magni- fication (Fig. 7.15) shows the diverse nature of microstructure, the precipitated phases. The grain boundaries are clearly outlined by the more massive particles. Aging at 1300°F for prolonged periods produces the characteristic microstructure shown in Fig. 7.16. The marked difference in structures ob- Fig. 7.13. Microstructure of 3/16-in. Hastelloy B Plate After a Solution Heat Treatment of 2 hr at 2150°F. Etched with chrome regia. 150X, Re- duced 30.5%. PERIOD ENDING MARCH 10, 1955 tained at 1300°F and at 1500°F would be expected to produce appreciable wvariations in physical properties. Hot tensile specimens have been subjected to long-time aging treatments and are now being tested to confirm this and other pre- dictions based upon this metallographic evaluation. The quantity of precipitate formed was found to increase proportionately with increasing time at temperature within the limits of this investigation, A 1950°F ‘“‘overaging’’ heat treatment, recom- mended by the Hoynes Stellite Company, noticeably reduced the rate of precipitation after short-time aging, with the effect still somewhat evident up to 1500 hr, Residual cold work in Hastelloy B produces a striking effect upon the rate of precipitation during aging, as evidenced by the excessive quantities of precipitate shown in Fig. 7.17, a specimen cold worked (20% deformation) and held for 100 hr ot 1200°F. Evidence indicates that the induced stresses produced during water quenching from above 1950°F may increase this precipitation rate. The influence of base metal composition is now being studied, and aging treatments will be performed on specimens of selected chemical analysis. Extreme variations in microstructure e RS NCLASSIFIED SN Y40ss | o e N \\”*4”‘ }*’4% x \ ,‘? )';_'(.;f'& :'?‘“t Fig. 7.14. Sample Similar to That Shown in Fig. 7.13 After Aging for 1500 hr at 1500°F, Etched with chrome regia. 150X. 117 ANP PROJECT PROGRESS REPORT o Tt N ey o - Loy ML . . - : o7 LAk o Mty s Ao - UNCLASSIFIED A # i ¢ - o -~ i ¥ . o Cog A LT e e AT T Y4058 -~ Ly 1 T P 0t . yfi B ¥ ! 5 ) \ i (P, ‘ L = }1() :“% I‘ R A ‘ 22 L }"{1 v i \ .y L\Eg Y * \ oo L ~ .\"s‘ : ?‘x ©w & \ VT . o ‘."Xia/f P % ey i 3 k. e . i - Ar/(f N s J”f't‘n ~ bk L = f; ofly . n7 F A % jff > ” 3 o o Nl % 3"‘ e e G l) . 1 Y *on " Ay e 3 PR . ::bmi‘ .7 : - ‘;7 ” ' f g o ‘i‘ %:‘”s Y A oY ok Jab N AL ’ ;"gf‘\ ’ . (f{,-”s““ o . o \ F \’? . idd;”“ ' / e,,"(/ W% ; 3 ) N4 e ’»}I.{ * 'f “?“, ) :-"""P 3 E i.‘ “\_hrg“%fia*—fi‘w?“ S Lo * L e T Te / Bt L e ’&?\ ; f I oo ? o, ek - R W - \« }},,{ # o s Ji - : G L - \ R it .. . W\fii .«-«""”‘f‘ ‘ je i . e Fig. 7.15. Higher Magnification (750X) of Fig. 7.14. Note the several types of precipitate present and especially, the more massive particles outlining the grain boundaries. Etched with chrome regia. r Fig. 7.16. Hastelloy B Aged at 1300°F for Prolonged Periods. Note characteristic precipitate, Etched with chrome regic. 150X, 118 ~ e smpons . UNCLASSIFIED Fig. 7.17. Hastelloy B Cold Reduced 20% and Subsequently Aged 100 hr ot 1200°F. Note ex- cessive precipitation. Etched electrolytically. 150X. Reduced 30.5%. are prevalent in welded joints, as would be ex- pected since they are susceptible to both segre- gation in the fusion zone and fo stresses from the welding process. The original microstructure, as developed by fabrication techniques or thermal treatments, has a decided effect upon the aging process, but each individual case should be evaluated separately. Except under corrosive or oxidizing conditions, the environment does not appear to be important, Early experiments indicated that it would be desirable to develop a heat treatment which would result in improved ductility ot room temperature and at high temperatures. Since a spheroidized structure should be beneficial, a heat treatment has been developed to produce it. This structure, which is shown in Fig. 7.18 at a magnification of 150 diameters and in Fig. 7.19 at a magnification of 2000 diameters, can be obtained by a heat treatment4 which is not deemed practical for structural components. However, physical test specimens containing the spheroidized micro- structure are being prepared to evaluate the effect of spheroidization onductility and other properties. If the results are promising, further work will be aimed at developing a more practical heat treat- ment, 4Aging of a 20% cold reduced specimen at 1200°F for 200 hr followed by spheroidization at 1540°F for 150 hr. PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED ¥.i3se8 Fig. 7.18. Microstructure of Hastelloy B Ob- tained by the Experimental Spheroidizing Heat Treatment. Etched with chrome regia., 150X, Reduced 30.5%. UNCLASSIFIED ;‘*21 * i‘d i * & Y L i . aanu :.'.\'gs P :%i & & G * | » o oo ® il ”~ * o 35 s ¥ G € ¥ 2, i, e &s'f&i.g Fig. 7.19. Fig. 7.18. Reduced 31%. Higher Magnification (2000X) of Etched with chrome regia. 2000X. As a part of the correlation of physical proper- ties with microstructure, hot tensile tests have been completed on wrought specimens aged for 100 hr at five different temperatures. The yield strengths and elongations were determined for specimens aged’’ conditions and are shown in Fig. 7.20. |t can be seen that a minimum elongation exists at 1300°F, while the yield strength does not appear to be as seriously affected. in the solution annealed and ‘‘over- 119 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 5784 80 L LT o] T 70 24 SPECIMENS SOLUTION-ANNEALED FOR 2 hr AT ——- 4 7Q 2150°F AND OVERAGED FOR 70 hr AT 1950°F | — BEFORE 100-hr AGING AND SUBSEQUENT TESTING —] 80 f | | \ ! i I | 60 [-#0 SPECIMENS ANNEALED FOR 2 hr AT 2450°F BEFORE # 60 100~ hr AGING AND SUBSEQUENT TESTING _ o ! Q . 0 —~ 50 < ] 50 = & ] | | E < - ) | 2 — T z | | A = 2 a0 — YIELD STRENGTH — 1= “I f— 40 © E | T / i o r" } o Z S / . S ! ; s 0 w 30 - 4 30 1 1| & T ,/ > | —— g 20 . : C—t™ — 74 20 § - s\\ J_,// ] 1} \..‘\ //,‘ 40 ‘ T 3 o S = 1 ! 10 ] DUCTILITY 4{ - | ‘. _ . BEREEENEEE . 1000 1400 4200 1300 1400 {1500 4800 4700 TEMPERATURE (°F) Fig. 7.20. Strength and Elongation of Hastelloy B as a Function of Testing Temperature After Aging for 100 hr at the Test Temperature. All physical testing performed at the aging temperature, Long-time aging of hot tensile test specimens is now in progress to permit a more complete analysis of the precipitation effects which would occur in Hastelloy B during service at these elevated temperatures. An apparatus for the hot bend testing of wrought and weld metal specimens is also under construction. The determination of aging rates by hardness measurements has also been proposed as a means of supplementing the data accumulated by other methods. STRESS-RUPTURE CESIGN CURVES FOR INCONEL R. B. Oliver D. A. Douglas J. H. DeVan J. W, Woods Metallurgy Division The program to obtain design data for lnconel at 1300 and at 1500°F under reactor conditions is and some data at 1650°F are A series of tests at an extremely low range of stresses has begun that will provide data on rupture times in the 5,000 to 10,000-hr range, which is longer than the present period assumed for aircraft reactor operation. nearly complete, available, The low stresses 120 involved in these tests are, however, representa- tive of the principal stresses with which the reactor design engineers are concerned. The to date, in fused salts and, for comparison, in argon are presented as a series of design curves shown in Figs. 7.21 through 7.30. The times to 0.5, 1, 2, 5, and 10% elongations and to rupture are plotted against stress for both annealed and as-received sheet results obtained for Inconel, specimens. In the case of Inconel specimens annealed at 2050°F and tested in argon at 1500°F, it should be pointed out that considerable scatter exists in the data as a result of a precipitation hardening phenomenon which sometimes occurs. The nature of the precipitate and the causes of its occurrence are presently uncertain; since it may or may not appear in entirely similar material under identical stress and temperature conditions. The precipi- tation tokes place at from 300 to 4000 hr and significantly strength of the alloy. earlier the creep and rupture The higher the stress, the in the test the precipitation effect is observed. Because of the unpredictable behavior of this precipitate, the design curves which are presented represent those tests in which no pre- cipitation hardening was observed. improves DEVELOPMENT OF BRAZIMG ALLOYS P. Patriarca K. W, Reber R. E. Clausing G. M. Slaughter Metallurgy Division J. M, Cisar Aircraft Reactor Engineering Division R. L. Heestand Pratt & Whitney Aircraft The resistance of brazing alloys to high-temper- ature oxidation is a primary factor to consider in the choice of these materials for use in the fabri- cation of liquid-metal-to-air radiators. ation program is therefore being conducted to determine the suitability of 28 potential high- temperature alloys for this application. Inconel T-joints were prepared with these materials and small smaples of these joints were subjected to static air at 1500 and at 1700°F for various times. The results of these tests, as determined by me- tallographic examination in the as-polished con- dition, are presented in Table 7.20. Some of the information found in this table has appeared in a An evalu- PERIOD ENDING MARCH 10, 1955 TABLE 7,20, OXIDATION RESISTANCE OF DRY-HYDROGEN-BRAZED INCONEL T-JOINTS Brazing Oxidation in Static Air* Brazing Alloy Composition (wt %) Temperature At 1500°F At 1700°F °F) For 200 hr For 500 hr For 1300 hr For 200 hr For 500 hr Commercial Alloys Nicrobraz 70 Ni~14 Cr—6 Fe—5 B~4 Si—1C 2150 Slight Slight Slight Slight Slight Low-melting Nicrobraz 80 Ni-5 Cr—6 Fe=3 B-5 Si-1 C 1950 Slight Slight Slight Slight Slight Coast Metals No. 50 93 Ni—-3.5 §i—-2.5 B-1 Fe 2050 Slight Slight Slight Slight Slight 51 92 Ni-4.5 5i-3 B-0.5 Fe 2050 Slight Slight Slight Slight Slight 52 89 Ni-5 5i~4 B~2 Fe 1840 Slight Slight Slight Slight Slight 53 81 Ni-4 Si—-4 B-8B Cr-3 Fe 1950 Slight Slight Slight Slight Slight NP 50 Ni-12 $i—28 Fe—4 Mo—-4.5 2050 Slight Slight Slight Slight Moderate P-1 Mn-0.5 Cr Mond Ni Co. Alloy 64 Ag—33 Pd-3 Mn 2150 Severe Severe Severe Complete Copper 100 Cu 2050 Complete Complete Experimental Nickel-Base Alloys G-E No. 62 69 Ni=20 Cr-11 Si 2150 Slight Slight Slight Slight Moderate 81 66 Ni=19 Cr—10 Si—4 Fe—1 Mn 2150 Slight Slight Slight Slight Moderate Mi-Cr-5i 73.5 Ni-16.5 Cr—10 Si 2150 Slight Slight Slight Slight Moderate Ni-Si 88 Ni—12 §i 2200 Slight Slight Slight Slight Slight Ni-Ge 75 Ni-25 Ge 2150 Slight Slight Slight Moderate Severe Ni-Ge-Cr 65 Ni—-25 Ge—~10 Cr 2130 Slight Slight Slight Slight Moderate Electroless Ni-P 88 Ni-12 P 1740 Slight Slight Slight Above melting point of alloy Ni-P-Cr 80 Ni—=10 P-10 Cr 1830 Slight Slight Slight Above sclidus of alloy Ni-Mo-Ge 50 Ni-=25 Mo~25 Ge 2150 Slight Stight Slight Moderate Severe MNi-5n 68 Ni-32 5n 2150 Slight Moderate Severe Severe Compiete Ni-Mn 40 Ni—-60 Mn 1950 Complete Complete Ni-Mn-Cr 35 Ni-55 Mn—10 Cr 2050 Severe Severe Complete Severe Complete Experimental Precious-Metal Base Alloys Pd-Ni 60 Pd—40 Ni 2300 Very Slight Moderate Very Slight slight slight Pd-Ni-5i 60 Pd—37 Ni-3 Si 2150 Very Slight Moderate Slight Moderate slight Pd-Al 92 Pd-8 Al 2020 Very Very Very Very Slight slight slight slight slight Pd-Ge 90 Pd—10 Ge 2050 Very Slight Severe Complete slight Au-Ni 82 Au-18 Ni 1830 Very Very Slight Moderate Moderate slight slight Au-Co 90 Au—10 Co 1830 Very Very Moderote Slight Severe slight slight Au-Cu 80 Au-20 Cu 1740 Moderate Complete Complete *Very slight, less than 1 mil of penetration; slight, 1 to 2 mils of penetration; moderate, 2 to 5 mils of penetration; severe, greater than § mils of penetration; complete, fillet completely destroyed. 121 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-DWG 5785 20,000 10,000 5000 STRESS (psi) 2000 1000 - : : 100 200 1000 2000 5000 10,000 TIME Lhr} Fig. 7.21. Stress-Rupture Characteristics of As-Received Inconel Tested in Argon at 1300 °F. UNCLASSIFIED ORNL—1.R-DWG 5786 20,000 \ 10,000 5000 STRESS (psi) 2000 1000 1 2 5 10 20 50 100 200 500 1000 2000 5000 10,000 TIME. (k) Fig. 7.22. Stress-Rupture Characteristics of As-Received Inconel Tested in Fused Salt at 1300°F, 122 PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED ORNL-LR-DWG 5787 30,000 20,000 RUPTURE 10 , 10,000 5000 STRESS {psi) 2000 1000 2000 5000 10,000 5 10 20 50 100 zZ00 500 TIME {hr) 1000 . Fig. 7.23. Stress-Rupture Characteristics of Annealed Inconel Tested in Argon at 1300°F. UNCLASSIFIED CRNL-LR—-DWG 5788 20,000 10,000 'E £ 5000 o wn L o f 2000 1000 \ 2 5 10 20 50 100 200 500 1000 2000 5000 10,000 TIME (hr) Fig. 7.24. Stress-Rupture Characteristics of As-Received Inconel Tested in Argon at 1500°F. 123 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL-LR-OWG 6789 T |o,ooor T 5000 |- _ | “w e w w 3 ul o '_ n 2000 Lo . N \ il 1000 i _ | f 2 5 10 20 100 200 500 100G 2000 5000 10,000 TIME (hr) Fig. 7.25. Stress-Rupture Characteristics of As-Received Inconel Tested in Fused Salt at 1500°F. 10,000 5000 | STRESS {psi) UNCLASSIFIED ORNL-LR-DWG 5790 UPTURE 100 200 500 1000 2000 5000 TIME (hr) 10,000 Fig. 7.26. Stress-Rupture Characteristics of Annealed Inconel Tested in Argon at 1500 °F. previous quarterly repmt,5 but it is presented again in a more complete form to permit a compre- hensive and detailed evaluation. The void formation along the braze metal-Inconel interface, which was noted on the samples brazed with G-E No. 62, G-E No. 81, and the 73.5% Ni-16.5% Cr-10.0% Si alloys, and which was previously thought to result from internal oxi- 5P. Patriarca et al.,, ANP Quar. Prog. Rep. June 10, 1954, ORNL-1729, p 94. 124 dation, is now believed to be associated with a diffusion phenomenon. Check samples, which were fested in vacuum under identical conditions of time and temperature, contained voids at similar locations along the interface. Although the nature of the void formation is not yet completely under- stood, it has been noted that the quantity of one constituent appears to increase with increasing time at temperature and with increasing temper- ature of test. The identification of this con- stituent, which is probably a complex intermetallic 10,000 5000 [~ - - STRESS (psi) 2000 1000 PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED ORNL-LR-DWG 5794 ! i 100 200 500 1000 2000 5000 10,000 TIME {hr} Fig. 7.27. Stress-Rupture Characteristics of Annealed Inconel Tested in Fused Salt at 1500°F. 10,000 5000 STRESS {psi) 2000 | | 2 5 10 [4e] 50 10001 UNCLASSIFIED ORNL-LR--DWG 5792 100 200 500 000 2000 53000 10,000 TIME (hr} Fig. 7.28. Stress-Rupture Characteristics of As-Received Inconel Tested in Argon at 1650°F. compound, is to be attempted by x-ray techniques, since knowledge of its composition is required for the determination of the diffusion mechanism. This investigation of brazing alloys has shown that a majority of the nickel and nickel-chromium base alloys are suitable for service in an oxi~ dizing atmosphere at 1500°F, and several are suitable ot 1700°F. Oxidized Inconel joints brazed with two of the alloys of prime interest for liquid-metal-to-air radiators are shown inFigs.7.31 and 7.32. The excellent resistance to attack of alloy No. 81 after 500 hr at 1500°F in static air is illustrated in Fig. 7.31. This alloy has been used for the fabrication of units containing austen- itic stainless steel or Inconel fin materials, while Coast Metals alloy No. 52, shown in Fig. 7.32, is of interest in the fabrication of high-conduc- tivity-fin assemblies. Only minor oxidation is evident on the joint brazed with Coast Metals alloy No. 52 after being subjected to the oxidizing 125 ANP PROJECT PROGRESS REPORT 10,000 5000 0.5% 1% 5% STRESS (psi) 2000 1000 ra o ) 20 50 UNCLASSIFIED ORNL-LR-DWG 5793 RUPTURE 100 200 500 1000 2000 5000 10,000 TIME {hr) Fig. 7.29. Stress-Rupture Characteristics of As) V(’ ‘:W X o S X S e Fig. 7.31. Inconel T-Joint Brazed with G-E Brazing Alloy No. 81 and Tested in Static Air for 500 hr at 1500 °F, Note negligible attack. Etched with electrolytic oxalic acid. 100X, nickel-chromium-germanium-silicon types may be of interest in fluoride-to~sodium heat exchangers. In order to conduct further research on these alloys and to provide enough material for tests on heat exchanger components, l-lb quantities of each alloy and of the 75% Ni-25% Ge binary alloy were arc-melted. Several techniques have been investi- gated for reducing these arc-melted ingots to the more desirable powder form, but no suitable labo- ratory method has been obtained. Grinding with a mortar and pestle or with a ball mill has not been satisfactory because the alloy possesses suf- ficient inherent ductility to prevent easy powdering. Experiments in which the alloy was induction melted in a quartz tube and the molten metal then permitted to drop from a small hole into water have not been promising. However, 50-g samples have been submitted to a commercial manufacturer of brazing alloy powders for experimentation in Other possibilities for obtaining powder lie in the fabrication and instal- powder-making equipment. lation of a small atomizing machine or the use of larger atomizers available in industry. Methods for the production of brazing alloy powders from of elemental powders are now being developed, and preliminary have indicated that this technique may be very promising. The elemental powders are carefully mixed and sintered in dry hydrogen for several hours at a temperature very near that required to fuse the lowest melting point eutectic in the system. This sintered material may then be crushed easily to provide a fine powder suitable for preplacement on heat exchanger or radiator components. Experiments for determining the optimum time and temperature required to obtain sufficient diffusion without seriously impairing the friability of the compact are being conducted, and the possibilities of applying this presintering method to other alloy systems of interest are being studied. sintered mixtures experiments 127 ANP PROJECT PROGRESS REPORT | UNCLASSIF 1ED . INCH 0.02 0.03 Fig. 7.32. lInconel T-Joint Brazed with Coast Metals Brazing Alloy No. 52 and Tested in Static Air for 500 hr at 1500°F. Only very slight oxidation can be seen. Etched with electrolytic oxalicacid. 100X, FABRICATION OF TEST COMPONENTS P. Patriarca K. W. Reber R. E. Clausing G. M. Slaughter Metallurgy Division J. M. Cisar Aircraft Reactor Engineering Division R. L. Heestand Pratt & Whitney Aircraft High-Conductivity-Fin Radiator The fabrication of a sodium-to-air radiator with 6 in. of type 430 stainless-steel-clad copper high- conductivity fins was described in a previous report.® The tube-to-fin joints were brazed with Coast Metals alloy No. 52 and the tube-to-header welds were made by the semiautomatic heliarc welding technique. The welded tube-to-header joints were back brazed as a precaution against the formation of leaks during service. During the fabrication of this radiator it was b, Patriarca ez al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 120. 128 found that it was necessary to maintain critical control over the quantity of brazing alloy placed on each tube-to-fin joint. A sufficient amount of alloy was required for the oxidation protection of the exposed copper at the punched holes of the high-conductivity fins. The presence of braze metal in excess of that required for the protection of the copper might result in “‘puddling,”” that is, concentration of the excess on the bottom fins, which would be undesirable because the air pas- sages between the fins might be sealed and lo- calized solution of the base metal might occur. The use of extruded brazing alloy wires as a means of obtaining controlled quantities has been investigated extensively, but the lack of a suitable binder makes their use somewhat unatiractive at the present time. An acrylic binder material produces wires which are relatively easy to handle when freshly extruded, but they become embrittled upon aging for a few hours at room temperature. Thus the assembly of large complicated radiators is seriously hampered by crumbling and subse- quent movement of the brazing alloy from the desired tube-to-fin location. Another binder ma- terial, Castolite, produces wires that are extremely ductile and weak at room temperature and therefore require meticulous care during handling. A dry-powder method of brazing alloy preplace- ment has been developed which provides a prom- ising means for obtaining controlled amounts of alloy on each tube-to-fin joint. A stainless steel template containing holes that have been precision drilled is placed over a sheared fin so that each hole in the template is centered over a punched hole in the fins. Since the drilled hole is larger than the punched hole, the template fits securely against the flat portions of the fin. The dry powder is then applied, and the excess powder is After careful removal of the template, the powder is secured to the fin with a methyl- acrylate cement and allowed to dry. A 36-hole fin with brazing alloy preplaced by this technique is shown in Fig. 7.33. The oxidation protection of the exposed copper on the sheared edges of the fin is also required in order to minimize oxidation and thus overcome the severe fin distortion that would result from the volume changes that would occur during formation of the oxide. An aluminizing process has been developed in which edge protection is obtained by the formation of a highly oxidation-resistant copper-aluminum alloy. A procedure has also been removed, PERIOD ENDING MARCH 10, 1955 devised to permit the aluminizing of large numbers of type 310 stainless-steel-clad copper sheet fins. This procedure, which requires that all fins be precision sheared to an exact size, is described in the following: chlorethylene or another suitable organic solvent, the fins are stacked in groups of approximately 200 and securely clamped together with ]/4-in.- thick stainless steel end plates. The exposed copper edges are then sprayed with three coats of Kestron acrylic spray to seal the cracks be- tween fins and prevent the flow of the aluminum- bearing slurry onto the stainless steel cladding. A coat of slurry, consisting of 100 cm?® of acrylic resin to 40 g of atomized aluminum powder (~325 mesh), is then applied evenly to the fin edges, After sufficient drying the clamped fins are heated in helium at 750°F for 2‘/2 hr to accelerate the formation of the oxidation-resistant aluminum bronze. A high helium flow rate (80 cfh) should be maintained for the first hour, but it can then be reduced to 40 cth for the remainder of the heat treatment. After cooling, the excess aluminum can be easily removed from the fin edges with the aid of a brass wire brush. These fins can then be punched by using the conventional techniques. Test specimens of fins with this type of edge protection have been metallographically examined after 100, 200, 500, and 1000 hr in air at 1500°F, After degreasing in poly- UNCLASSIFIED Y-14294 Fig. 7.33. High-Conductivity Fin for Radiator with Brazing Alloy Preplaced by Dry-Powder Technique. 129 ANP PROJECT PROGRESS REPORT A protected specimen tested in air for 1000 hr at 1500°F is compared in Fig. 7.34 with an un- protected fin tested under the same conditions. The small voids near the aluminized edge of the protected fin are atiributed to the migration of copper into the aluminumerich surface metal. Cyclic tests are also being conducted on alumi- nized fins to evaluate their stability under simu- lated aircraft service conditions., After 200 hr at 1500°F and 70 air-cools to room temperature, no adverse effects were evident, In preparation for the subsequent assembly, welding, and brazing of two 500-kw NaK-to-air radiators for heat exchange experiments, 2200 stainless-steel-clad copper high-conductivity fins have been sheared to the desired 2.in. by 8-in. final dimensions. They were degreased, inspected for surface imperfections, and aluminum-bronze edge-protected by the procedure described above. They will be punched and ossembled with the tubes by utilizing the dry-powder preplacement technique. The use of a l/lé-in.-thick template containing holes 0.246 in. in diameter will provide the minimum quantity of brazing alloy required per joint to permit good coverage of the exposed copper on the punched lips. To further reduce the tendency toward the undesirable accumulation of excess alloy as a result of normal variations in the quantity deposited, two thin sheets of Inconel are to be placed tightly together at 4-in. intervals along the 12 in. of stacked fins. The capillary joint between these two Inconel sheets will act as a sump to remove any excess brazing alloy that may be present. This technique is applicable to the fabrication of the 500-kw radiators because l/Ié-in.-thick Inconel sheets are required at 4-in. intervals for structural support. UNCLASSIFIED Y¥-14334 . Fig. 7.34. Type 310 Stainless-Steel-Clad Copper High-Conductivity Fin After Oxidation for 1000 hr at 1500°F. The top fin illustrates the excellent edge protection afforded by the aluminizing treatment. The bottom fin illustrates the severe attack of the unprotected fin. 130 Difficulties have been encountered in the use of the 6% aluminum-bronze high-conductivity fin material when brazing the fins to Inconel tubes. The aluminum oxide film present on the fins prevents wetting by the conventional brazing alloys. An electroplate of nickel on the fin did not act as an adequate diffusion barrier to permit wetting by the alloys. A chromium electroplate served as a diffusion barrier, but chromium is extremely difficult to wet in the hydrogen dew- points readily available in large-scale brazing operations. However, with an electroplate of iron on the aluminum bronze, tube-to-fin joints brazed with Coast Metals alloy No. 52 have been obtained that exhibit excellent wettability, Fig. 7.35. Joints brazed in this manner have also shown excellent resistance to oxidation after 100 hr at 1500°F. The iron pickup during brazing does not appreciably reduce the oxidation resistance of the nickel-silicon-boron brazing alloy. The excess iron can be removed from the fins after brazing by Fi g. 70350 PERIOD ENDING MARCH 10, 1955 a suitable pickling operation, or it can be re- moved early in service upon the formation of a nontenacious iron oxide. In view of these prom- ising developments, a quantity of commercially available 5% aluminum bronze has been punched and is being iron-plated for the subsequent fabri- cation of a large-scale radiator. Intermediate Heat Exchanger No. 2 The fabrication of the fluoride-to-sodium inter- mediate heat exchanger No. 2 requires the heliarc welding of 400 tube-to-header joints and the back- brazing of these welds with a suitable corrosion- resistant alloy, such as Nicrobraz or nickel- chromium-germanium. The welding of approximately 200 of these joints has been satisfactorily com- pleted by the semiautomatic rotating-arc method, but prior to initiation of the actual fabrication, a set of experiments was conducted to determine the optimum combination of welding conditions. Several of the sample joints were examined under UNCLASSIFIED U NT14098 lron-Plated Aluminum Bronze High-Conductivity Fins After Brazing with Coast Metals Brazing Alloy No. 52. Good wettability of the fins was obtained. 60X. 131 ANP PROJECT PROGRESS REPORT high magnification to determine the presence of weld microfissuring, but this type of cracking was not found to constitute a problem. In these experiments, two Inconel test headers, of the same size and physical shape as those on the actual unit, were machined, and sample tube- to-header welds were made under various con- trolled conditions. It was immediately evident that the preparation of the header surface after insertion of the tubes, but prior to welding, should definitely not be done by abrasive grinding. En- trapped abrasive in the joint caused severe arc instability and inconsistent welds. A more reliable method consists of preshaping the tubes to conform to the curvature of the header before assembly, The tube can then be heliarc tack- welded to the header and expanded with a special tool before final welding. With the method of header preparation prescribed, welding variables were evaluated to ascertain those which gave the most consistent weld penetration and which were least likely to result in excessive hole constric- tion or in undersirable preferential melting of the tube wall, Prior experience had shown that an arc distance of 0.050 in. and a weld time of approximately 6 sec produced consistently good welds when joining 3/] ¢ in-0D, 0.017-in. wall tubing to relatively thick headers. These values were therefore used for the determination of the optimum diameter of electrode rotation and the proper welding current. A rotation diameter of between 0.21 and 0.22 in. was found to be desirable, since diameters of less than this often resulted in preferential melting of the tube wall and greater diameters produced the maximum weld penetration in the header plate rather than at the joint where it is of most impor- tance. An arc current of 60 amp at an arc voltage of 10 produced consistently satisfactory welds with penetrations of approximately twice the tube wall thickness, A typical weld produced under these optimum conditions is shown in Fig. 7.36. No weld porosity or cracks are evident and excellent penetration was achieved. The two-pass effect evident in the nugget occurs because a weld over- lap of one-fourth revolution is used after the com- plete peripheral weld has been made. As the weld overlap is being made the weld current is gradually decreased to prevent the formation of undersirable arc craters. The uniformity of welds made under these conditions can be seen in Fig. 7.37, which 132 shows a 100-weld header section of the partially completed intermediate heat exchanger No. 2. The fabrication of approximately 50 ‘‘comb" spacers for this unit has also been completed. The necessary jigs were prepared and cone-arc plug welding conditions were determined for the heliarc welding of the 0.020-in. by 0.040-in. wire spacers into the 0.010-in. Inconel strip headers. After these spacers are attached to the assembly and the tube-to-header joints are back-brazed, the unit will be ready for insertion in the heavy Inconel pressure shell, Radiator for Cornell Aeronautical Laboratory The fabrication of a full-scale liquid-metal-to- air radiator designed by the Cornell Aeronautical Laboratory has been undertaken. The design incorporates integral-helical-finned tubing com- pletely machined from type 316 stainless steel bar stock. The tubing is machined to 0.200 in. in outside diameter with a 0.037-in. wall, and the header plates and other structural components of the radiator are machined from 1!/1 - and Y-in. type 316 stainless steel plate. The partially completed unit is shown in Fig. 7.38. After assembly of the tube-to-header section of the radiator, the 312 tube welds were manually heliarc welded by qualified operators using pre- scribed procedures. Longitudinal grooves were machined in the header plate before welding to simulate trepanning. The grooves in the header aid in minimizing microfissures because they substantially reduce the strain restraint near the weld, A ‘‘skip sequence’’ was also employed during welding to equalize the heat distribution in the header plate, The filler plates that can be seen in Fig. 7.38 were welded to the side plates to occupy most of the space between these plates and the outer rows of tubes. To prevent severe distortion when welding the side plates to the headers, the side plates were reinforced with 1-in.-thick stainless steel strong-backs, as shown in Fig. 7.38, and then welded to the unit, Figure 7.38 shows the unit after completion of the root passes. The re- maining heavy welding and back-brazing with an alloy such as Coast Metals alloy No. 50 should be completed in the next few weeks. The choice of an alloy for use in back-brazing the welded tube-to-header joints in this radiator depends to a large extent upon the ability of the PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED Y-14185 Q.01 X O 2 0.02 0.03 Q.04 0.05 Fig. 7.36. Inconel Tube-to-Header Weld Prepared Under Optimum Welding Conditions. Excellent pene- tration and weld quality obtained. Etched with electrolytic oxalic acid, 75X. UNCLASSIFIED ¥-14355 A 100.Weld Header Section of the Fig. 7.37. Intermediate Heat Exchanger No. 2 Welded with the Semiautomatic Heliarc Welding Equipment, Note weld uniformity. alloy to withstand the severe strain imposed upon cooling as a result of the heavy sections involved. Although several high-temperature brazing alloys possess adequate resistance to high-temperature oxidation, many of them, such as Nicrobraz, G-E alloy No. 81, and Coast Metals alloy No. 52, crack upon furnace cooling from the brazing temperature. However, Coast Metals alloy No. 50 and the nickel-chromium-germanium alloy show no evidence of cracking under these circumstances. Since Coast Metals alloy No. 50 can be purchased in a powder form and it possesses good resistance to both oxidation and to liquid sodium, it will be seriously considered for this application. MHow- ever, the extent and effects of boron diffusion from the braze into the stainless steel base material will be determined before it is used. The large mass of the unit suggests that a 12-hr-heating and 12-hr-cooling cycle should be used during brazing, 133 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ¥-14282 Fig. 7.38. Cotnell Aeronautical Laboratory Radiator After Welding of the Tube-to-Header Joints and Deposition of the Heliarc Root Passes in the Side Plate-to-Header Joints, The heavy stainless steel strongbacks were used to prevent severe distortion of the side plates during welding. and metallographic test specimens have been prepared to test this cycle under simulated condi- tions, The extent of diffusion will be determined in the as-brazed condition and after heating for 25, 50, 100, and 500 hr at 1650°F. Also, nine tensile-test specimens have been coated with a slurry and brazed under conditions approximating those to be used on the actual radiator. These specimens were submitted to Cornell Aeronautical Laboratory for longitudinal drilling and testing to simulate tests on tubing. Two uncoated tubes wete subjected to the some thermal cycle as the brazed samples to serve as controls. If the re- sults of these metallographic and physical tests are promising, Coast Metals alloy No. 50 will be used for the brazing operation. Otherwise, an 134 alloy of the nickel-chromium-germanium type will probably be used. Sodium-Betryllium-Inconel Compatibility Testing Apparatus A second sodium-beryllium-Inconel compatibility testing apparatus was fabricated; the components are shown in Fig. 7.39 in an exploded view and in Fig. 7.40 after completion, The unit comprises a cylindrical beryllium insert inside an lnconel housing. The Inconel housing, the sodium pots, and the thermocouple housings were manually heliarc welded into position, and nine thermocouple assemblies were brazed with G-E alloy No. 62. All welded and brazed joints were leak-tight, as determined by testing with a helium leak-detecting apparatus. Heot Exchanger Brazes It was reported previously? that the 82% Au-18% Ni brazing alloy should be considered for use in the fabrication of fluoride-to-air heat exchangers, and it has now been established that the solid- state migration of gold from this alloy into Inconel tubing is minor after an extended peried at an elevated temperature. Thus corrosion tests of the tube interior after service would not be obscured by the presence of localized gold-rich areas. Recent tests with copper as a brazing alloy for fluoride-to-helium heat exchangers have shown similar promising results. A typical joint was held at 1500°F for 900 hr and then examined for copper diffusion, Microspark spectrographic techniques coupled with a metallographic exami- nation revealed that a maximum of 0.003 in. of copper diffusion had occurred. | Dynamic corrosion loops are also tobe fabricated for studying the corrosion of joints brazed with several alloys and tested in intimate contact with circulating fluoride fuel, These loops will in- corporate several sleeve-type sections in the hot leg to assist in the evaluation of brazing alloys for in-pile loop applications. Cermet-to-Metal Brazing The joining of cermets to Inconel will be required if cermets are to be used for applications such as valve seats, and, in evaluation studies of cermets, 7P. Patriarca et al., ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 109. Fi Ge 7-39¢ PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED - Y-14142 Unassembled Sodium-Beryllium-Inconel Compatibility Testing Apparatus No. 2 Showing Beryllium insert, Incone! Housing, Sodium Pots, and Thermocouple Assemblies. joins between these two materials are now re- quired for the preparation of samples for cermet self-welding tests. A study of cermet-to-Inconel brazing is therefore being conducted in an attempt to develop suitable techniques for the production of these joins. Preliminary tests have consisted of determining the wettability of certain cermets with fluoride- resistant alloys such as copper, gold-nickel, and nickel-phosphorus, Although the nickel binder is easily bonded, it appears that wetting of the non- metallic portion of the cermet may be much more difficult. Also of prime importance in the success- ful production of these joins will be the solution of the problem posed by the unequal thermal ex- pansions of the two materials, The cermets of immediate interest have thermal expansion coeffi- cients ranging from 4.0 x 10=% to 5,7 x 10-6 in./in..°F as compared with 10,2 x 10~¢ in./in..°F for Inconel in the range 1000 to 1400°F. The best solution to this problem seems to lie in the use of a combination of a ductile brazing alloy and a ductile pad of metal between the cermet and the Inconel. A technique now being investigated consists of the deposition on the cermet of a 0.0001-in. layer of ‘'electroless’’ nickel-phosphorys, which wets both the metallic and nonmetailic portions of the cermet, When this layer is covered with a 0.005- in-thick acid copper plate and then joined to Inconel by heating to above the melting point of both the nickel-phosphorus layer and the copper, a join that possesses fair mechanical properties is produced. The phosphorus apparently aids in wetting the cermet and then diffuses into the sur- rounding metal sufficiently topermit some ductility. Physical tests on this type of join will be con- ducted and screening tests on other alloys will be 135 ANP PROJECT PROGRESS REPORT UNCLASSIFIED Y-14179 Fig. 7.40, Assembled Sodium-Beryllium-Inconel Compatibility Testing Apparatus No. 2, performed in an attempt to obtain more satisfactory brazing material for this opplication. Methods for bonding ceramics, such as the manganese-molyb- denum and titanium hydride techniques, will also be investigated. SPECIAL MATERIALS FABRICATION H. Inocuye J. H. Coobs Metallurgy Division M. R. D'Amore Pratt & Whitney Aircraft Duplex Tubing The study of flow patterns in impact-extruded duplex and three-ply tubing was continued, Four extrusions of composite stainless steel—carbon steel tubing have been completed. An examination of the layers of the extruded tubes revealed that fairly close control of the final layer thickness was obtained by adjusting the layer thickness of the extrusion billet, The extruded tubes were 136 subject to poor material recovery because of metal loss at both ends of the tube. An attempt to im- prove the material recovery by varying the ex- trusion billet design is planned, G-E Control Rods The 34 control rods ordered for the GE-ANP Project have been completed and delivered for reactivity testing. They were prepared by filling tubes with a mixture of 50% aluminum and 50% boron carbide by tamping with a pneumatic hammer. The tubes were then evacuated while being heated at 300°C to remove the paraffin binder and cold swaged to the final diameter. The tubes, as loaded, had a boron density of about 0.82 g/cm3 that was increased during the final swaging operating to more than 0.85 g/cm?, which was considerably more than the 0.70 g/cm? minimum specified, Tubular Fuel Elements Twelve more plate-type fuel elements were prepared by using either elemental or prealloyed type 304 stainless steel in combination with either high-fired or steam oxidized UO, in the cores. These elements have cores 2 in. wide and 5 in. long, and they will be formed into single, seamed tubes for hot and cold drawing experi- ments. Several more tubular assemblies are being prepared by using seamless materials for reduction by hot swaging or hot rod rolling techniques. Boron Shield for ART Samples of an 85% B C-15% SiC composition with a boron density of 1,26 g/cm3 were received from the Carborundum Company, This composition is reputed to be easy to fabricate, with very close dimensional control, by cold pressing and firing. Nicholson of the Carborundum Company stated that the boron content could be increased somewhat by the addition of boron or boron nitride to the composition. Metal lographic examination of samples from a recent test designed for comparing the reactions of boron and of boron carbide with Inconel has been completed, In the test, three samples of boron-containing materials were in contact with Inconel in a helium atmosphere for 100 hr at 1500°F. The boron-containing materials were commercial, hot-pressed B ,C (Norton Company’s ““Norbide’’), hot-pressed boron, and cold-pressed Total depth of penetra- tion in Inconel of the reaction with these materials was 0,004, 0.006, and 0.010 in., respectively. The diffusion of boron from these materials into Inconel seems to proceed along grain bound- aries, and @ second phase which is boron-rich is gradually formed. In the case of severe reac- tion, as from the amorphous boron powder, the second phase forms a continuous layer at the amorphous boron powder. interface; the layer separates slightly from the inconel during cooling. Thermal cycling of a system in which considerable reactionhad occurred could thus result in spalling of the layer and an increase in the rate of attack. Another compatibility test has been completed that was designed to compare the reactions of hot- pressed B,C and the B ,C-SiC composition with Inconel. In this test the samples were exposed in helium at 1600°F for 100 hr. In an attempt to in- hibit the reaction, one surface of each of two pieces of the B C-SiC was sprayed with a sus- pension of Norton alumina (Grade 38-900) in acetone before assembly for testing., Visual exami- nation showed some reaction, even on those sur- faces protected with alumina. Samples are now being prepared for metallographic examination. Al-UO, Fuel Plates for Shielding Experiment A series of fuel elements was prepared with both 525 and 24S aluminum alioys for cladding in an attempt to make aluminum-clad sandwiches con- taining UO, mixed with aluminum that will be strong enough to withstand the high stresses expected in the proposed shielding experiments that mock up the reflector-moderated reactor (cf., sec. 13, ““Lid Tank Shielding Facility’). In the test the plates will be cycled at speeds ranging up to 20 fps on a link chain running on 3.5-in.-dia sprocket wheels. No difficulties were encountered in fabricating the cores for these plates or in maintaining uniform fuel distribution. The cores were prepared by mixing 57 wt % steam-oxidized UO, in -140 mesh atomized aluminum and then cold pressing at 33 tsi. The sandwiches were reduced about 92% in 10 roll passes with no evidence of edge cracking or separation in the cores. In preparing the 525 aluminum-clad plates, much difficulty was encountered in obtaining bonding PERIOD ENDING MARCH 10, 1955 between the cladding and the frame during hot rolling. Therefore a second set was prepared with 2S5 aluminum frames and 525 aluminum cladding. These plates bonded satisfactorily. They were strain-hardened by cold rolling to the required thickness and finally bent slightly on a 5]/5-in. radius to straighten them and add ridigity, How- ever, attempts to bend them to a smaller radius, to achieve greater ridigity, resulted in cracking of the cladding and exposure of the UO -bearing core, and therefore these plates also were unsatisfactory. Next, a series of plates was prepared with 245 These plates were rolled successfully, with good bonding and fuel distribution, and they could be solution annealed and bent cold to a cross sectional radius as small as 1]/2 in. without cracking the cladding. After age hardening, these plates were tested and found to be very close to the required strength. A set of 17 such plates has been prepared for an endurance test. Each plate contains the required 91 g of UO, with 245 aluminum cladding. They were bent to a ]%—in. radius and age-hardened to the T-81 condition. Fabrication of enriched plates will probably begin as soon as results of the test are available, aluminum as the cladding material, CERAMIC RESEARCH C. E. Curtis J. A, Griffin J. R. Johnson Metallurgy Division Oxidation Reactions of U02 and of UO2 in BeO Uranium oxide in the form of a powder or bars previously sintered at 2900°F in hydrogen gained up to 2%in weight when fired for 66]/2 hr at 2500°F in still air. Under the same conditions, up to 14% was lost from beryllium oxide tubes containing 16.2 wt % of UO, previously sintered at 2900°F. The accelerated volatilization may have been due to the formation of a compound between uranium oxide, beryllium oxide, and possibly water vapor, The losses in UO, were much lower than those obtained recently by H, C, Brassfield with similar tubes heated in moving air. However, all results point to the necessity of stabilizing UO, against oxidation under these conditions. For this purpose either the formation of a stable compound of UO or use of a protective glaze may be found to be feasible. Both approaches are being investigated. 137 ANP PROJECT PROGRESS REPORT Fabrication of Rare Earth Oxide Wafers for Critical Experiments Five wafers each of gadolinium and samarium oxides (?93 in. in diameter and about 10 mils thick) were supplied for testing at the Critical Experi- ment Facility (cf., sec. 4, *"Critical Experi- 138 ments’’). About 32 slugs of a mixture of Sm,0, and Gd,0, are being fabricated for subsequent canning in an Inconel rod, The slugs will be 0.45 in. in diameter and 0.75 in. in length. These slugs will also be used at the Critical Experiment Facility, Some physical property measurements will be made before delivery. PERIOD ENDING MARCH 10, 1955 8. HEAT TRANSFER AND PHYSICAL PROPERTIES H. F. Poppendiek Reactor Experimental Engineering Division Additional heat transfer experiments have been performed with NaF-ZrF ,-UF, flowing in Inconel tubes. Some quantitative velocity measurements were obtained for the 18-in. ART core, and the nature of fluid flow in a simple separation region was studied. Analytical studies of heat and ve- locity distribution in the high-temperature-differ- ential, high-velocity loops were made, A preliminary value was obtained for the heat capacity of NaF-ZrF ,-UF, in the liquid state, and the viscosities of NaF-KF.ZrF,, NaF-BeF,, and LiF-BeF, were determined, A study was made of the influence on ART heat transfer of the physical properties of three types of fluoride fuels, FUSED SALT HEAT TRANSFER H. W. Hoffman Reactor Experimental Engineering Division Further heat transfer experiments have been performed with the fuel mixture NaF-ZrF,-UF, (53.5-40-6.5 mole %) flowing in an Inceonel tube. The data obtained were for the period of operation from O to 2 hr and were in agreement with the previously reported results for 24 to 115 hr of operation. Failure of the apparatus caused termi- nation of this sequence of experiments before data for longer operational times could be obtained. The results of the heat transfer measurements in the system NaF-ZrF ,-UF ;-Inconel are summarized in Fig. 8.1, and a comparison is made with the general correlation for ordinary fluids, as well as with the results for the NaF-KF-LiF (11.5-42-46.5 mole %)-Inconel system in which corrosion deposits were found on the tube walls. Surface deposits did not occur with NaF-ZrF ,-UF , in |lnconel, but vapor blanketing on the inside tube surface could, perhaps, cause the 24% difference between the general correlation and the current data. Therefore, a system is being readied for studying the pressureedrop characteristics of NaF-ZrF ,-UF ; in circular tubes with electrical ly heated wall s, Photomicrographs (Figs. 8.2 and 8.3) of the two Inconel tubes used in the NaF-ZrF,-UF, heat transfer experiments were made by R, Crouse of the Metallurgy Division. The conditions prevailing during exposure are presented in Table 8.1, In tube 1, extensive subsurface void formation oc- curred to a depth of 5 mils. However, the total void volume was not large enough to affect the thermal conductivity of this 5-mil thick region. This is further indicated by the agreement between the heat transfer results of tube 1 and tube 2; no B ORNL—LR—DWG 4 0.010 %0 & a = 0005 ?"’ " GENERAL CORRELATION, /=0023 Wp02 | 1 ; T “-” -0.‘._l T = //’{ 3 : e o 5 & o s -~ T o ko | = o NoF - ZrF, —Ur, (53.5~40 ~6.5 mole %) & go0z s \ IN INCONEL | f m o 2 dinys o NaF — KF—LiF (11.5-42—46.5mole %) IN INCONEL L | Q.00 ! - 2000 5000 10,000 50,000 NR: Fig. 8.1. Comparison of Current Heat Transfer Measurements on NaF.ZrF UF, (53.5-40.6.5 mole %) in Inconel with the General Correlation for Ordinary Fluids and the Data for NaF-KF.LiF (11.5-42.46.5 mole %) in lnconel. Fig. 8.2. NoF-ZrF,-UF, for 115 hr. 1300°F. Average velocity: 10.5 fps. Inconel Tube Exposed to Flowing Average temperature: 139 ANP PROJECT PROGRESS REPORT TABLE 8.1. CONDITIONS PREVAILING DURING HEAT TRANSFER EXPERIMENTS ON NaF-ZrF ,-UF Tub Exposure Temperature Reynolds Velocity uhe (hr) (°F) Number {fp s) 1 24 1200 5,500 to 6,500 8.5 to 10,0 85 1300 ~ 8,500 10.5 6 1400 9,000 to 10,000 10.8 2 3 1420 ~ 8,500 10.5 Fig. 8.3. NaF.ZeF,-UF, for 3 hr. 1420°F. Average velocity: 10.5 fps. Inconel Tube Exposed to Flowing Average temperature: apparent void formation occurred in tube 2. Ad- ditional heat transfer studies with NaF-ZrF -UF, are planned in a system modified so that the cause of previous apparatus failures is avoided. Since the ART fuel-to-NaK heat exchanger has been designed to operate in the transition flow regiomy-experiments to define more precisely fused salt heat transfer in both the transition and the laminar zones are now under way. Experiments with NaF-KF-LiF eutectic in stainless steel and a zirconium-base fuel in Inconel will be conducted. REACTOR CORE HYDRODYNAMICS J. O, Bradfute L. D. Palmer F. E. Lynch Reactor Experimental Engineering Division G. L. Muller Pratt & Whimey Aircraft Preliminary, quantitative velocity data for a model of an 18-in. core for a reflector-moderated reactor were obtained by the photographic tech- 140 nique. This core has straight-through flow and no entrance vanes. Qualitative velocity infor- mation was obtained by using the visualization technique in which a phosphorescent material is employed.‘ This technique revealed a very large region of flow separation along the outer core wall starting at a point close to the entrance and con- tinving past the midplane, Figure 8.4 shows an estimate of several velocity profiles in the flow channel. Additional smaller separation regions were observed next to the island wall, but they were much less distinct. The results of these experiments are described in detail in a separate repori'.2 The design of a one-quarter scale model of a proposed 2l-in, core has been undertaken. It is planned to utilize the flow facilities and the lighting and photographic equipment assembled for the smaller core model to describe the velocity pro- files. The phosphorescentparticle and the photo- graphic techniques will again be used. A vaned section to impart a rotational component to the flow immediately upstream from the core model has been fabricated (Fig., 8.5). The effect of rotation on the large separation region will be studied qualitatively by using the visuvalization technique. A variable-angle divergent channel system was designed, fabricated, and installed in the flow- visualization system so that the nature of flow in separation regions could be studied. Thick, laminar-like flow layers were noted next to the wall where flow separation occurred under turbulent L. D. Palmer and G. M. Winn, A Feasibility Study of Flow Visualization Using a Phosphorescent Particle Method, ORNL CF-54-4-205 (Apr. 30, 1954), 2J. 0. Bradfute, Qualitative Velocity Information Regarding the ART Core: Status Report No. 4, ORNL CF-54-12-116 (Dec. 14, 1954). SECRET ORNL-LR-DWG 5619 Ut l VERY LOW VELOCITY, \ PERHAPS NEGATIVE \ | VERY LOW VELOCITY, PERHAPS NEGATIVE VAN A L———VERY LOW VELCCITY, /T PERHAPS NEGATIVE N1y SEPARATION REGION, LARGE-SCALE TURBULENCE LIMIT OF SEPARATION REGION (OBSCURE DUE TO INTENSE TURBULENCE) | v/ Fig. 8.4. Qualitative Velocity Profiles in a Model of an 18-in. Core of a Reflector-Moderated Reactor for a Reynolds Number of 3000. PERIOD ENDING MARCH 10, 1955 (high Reynolds number) conditions. Also, regions of partial flow stagnation were observed next to this laminar-like layer, Experiments dre now in progress in which these separation regions are being wiped out by the insertion of very fine screens | ocated upstream. ELECTRICAL HEATING AND FLOW IN TUBE BENDS H. W. Hoffman L. D. Palmer N. D. Greene Reactor E xperimental Engineering Division It has been observed that peripheral corrosion on the inside surface of the 180-deg bends of a high-temperature-differential, high-velocity loop is not uniform. The preliminary resuvlts of the experi- ments (cf., sec. 6, ‘‘Corrosion Research’’) indi- cated the attack on the short-radius side of the bend to be 2 to 3 times that on the long-radius side. If it is presumed that the corrosive attack is related to the temperature of the surface and that, probably, it is greater at a high surface temperature, two possible mechanisms by which the short-radius side temperotures could be greater than the long-radius side temperatures can be hypothesized: (1) nonuniform heat generation in the wall and (2} nonuniform fluid velocity distri- butions. UNGLASSIFIED PHOTO - 23481 Fig. 8.5. Vaned Entrance Section for Obtaining Flow with a Rotational Component. 141 ANP PROJECT PROGRESS REPORT It was shown that for a tube bend with an elec- trical current passing through the tube walls, the heat generation in the wall on the shorteradius side of the bend is greater than that on the long- radius side. For the specific dimensions of the bends in the loops used in these experiments (tube OD, 0.5 in.; wall thickness, 0.045 in.; radius of curvature, 3.5 in.), the short-radius side heat generation was 30% greater than that of the long- radius side. Therefore the surface temperature on the short-radius side was higher than that on the long-radius side. Experimental verification of this conclusion was obtained by instrumenting a typical bend with thermocouples and observing the temperatures as the power generation was varied while the outer surface was being cooled by uniform natural convection and there was no flow on the inside of the tube. Velocity profile observations were made of the flow through a 180-deg glass bend having the same ratio of radius of curvature to tube radius as that in the experimental loop. The phosphorescent flow visualization method was used for these observations.! |t was found that the maximum fluid velocity occurred close to the long radius of the bend and that there was a large difference in the amount of wall cooling between the short- and long-radius sides because of the large differ- ences in fluid velocity, Thus the difference between the wall temperature on the short-radius side of the tube and that on the long-radius side was further increased by the difference in wall cooling. Indeed, it is probable that the effect of nonuniform fluid flow distribution is more important in yielding high wall temperatures than is the effect of the nonuniform heat flux in the tube wall. Estimotes of the combination of the heating by the two mechanisms have yielded differences in temperature between the short- and long-radius sides of the order of 100°F. Some measurements have been made that tend to confirm this rough estimate (cf.,, sec. 3, ‘‘Experimental Reactor Engineering’’). ART FUEL-TO«NaK HEAT EXCHANGER J. L. Wantland H. W. Hoffman Reactor Experimental Engineering Division In the current design of the ART, the flow on the fuel side of the heat exchanger is in the middle of the transition flow region (Reynolds number ranges from 2000 to about 5000). The fuel will 142 flow parallel to a bundle of tubes through which NaK will be flowing, and, for a zirconium-base fuel in this system, the Nusselt number may vory from about 4.4 to 35 through this narrow Reynolds number range. Thus, a study has been initiated to determine the heat transfer and friction charac- teristics of the heat exchanger in the transition flow region, An experiment has been designed which utilizes a full-scale heat exchanger tube bundle containing 100 tubes. Water is to be used as the heat trans. fer fluid at a temperature level that will give Prandtl numbers and kinematic viscosities similar to those of the fuel. The heat transfer charac- teristics will be determined by two different methods. First, measurements will be made with the system operating as a water-to-water heat exchanger with high fluid flow rates through the tubes to yield low and calculable thermal re- sistances in the flowing water inside the tubes, In the second experiment, the tube bundle will be resistance heated by passing an electric current through it, and it will be cooled by water flow between the tubes. Pressure-drop measurements will also be made for the system. These experi- mental studies will yield the variations of Nusselt number and the friction characteristics throughout the transition flow region. REACTOR CORE HEAT TRANSFER N. D, Greene J. A. Russell Reactor Experimental Engineering Division A study of the feasibility and desired objectives of o volume heat source experiment for investi- gating uniform volumetric heat generation within divergent and convergent channels has been com- pleted. The design of several components for this experiment, for example, the heat exchanger and the test channel, has been established, and con- struction of the divergent-flow test channel .is nearing completion. Construction of the associated heat exchanger is awaitingreceipt of the necessary metal. The 300-kw, 460-v power supply, which is required both for the volume heat source experi- ment and the ART heat exchanger experiment, has been ordered, and all necessary power instru- mentation has been designed to adapt this power source to both experiments. Means for recording rapid temperature changes in the fluid, which will be at an electrical potential of 440-v (alternating current), are being studied. A saturable reactor is being tested to determine its ability to control large amounts of alternating current. The charac- teristics of the reactor will be evaluated as soon as a source of direct current (for control purposes) is available,? TRANSIENT BOILING STUDIES M. W. Rosenthal Reactor Experimental Engineering Division The phase of the transient boiling program which involved a series of delay-time and superheat measurements was completed. A number of runs with exponential increases in power were under- taken, and several experiments were performed with step increases and with [inear increases in power. Reduction of the data and analysis of the results are continuing. HEAT CAPACITY W. D. Powers Reactor Experimental Engineering Division Two copper calorimeters have been installed and are now in operation. Because of the much greater precision that can be obtained, these two calo- rimeters are expected to furnish enthalpies and heat copacities at least as fast as did the five ice calorimeters previously in use. - Preliminary results for the heat capacity of liquid NaF-ZrF -UF, (56-39-5 mole %) were ob- tained. The heat capacity was found to be 0.256 + 0.004 cal/g.°C over the temperature range of 570 to 890°C. YISCOSITY S. I. Cohen Reactor Experimental Engineering Division The program of refining viscometry techniques has been substantially completed. The specific changes made during the course of this program were described previously.* Results of the program suggest that subsequent measurements will yield viscosities somewhat lower than those given by early measurements on the same compo- sitions, 3Most of the power supply development and instru- mentation work is being done by J. A. Russell of the Instrumentation and Control Group of the Reactor Ex- perimental Engineering Division, 4s. I. Cohen and T. N, Jones, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 117, PERIOD ENDING MARCH 10, 1955 Measurements have been made with the refined techniques on three fluoride mixtures. Data were taken on NaF-KF-Z¢F , (5-52-43 mole %) with both the Brookfield and capillary viscometers, and the results yielded by the two devices were in good agreement (Fig. 8.6); the viscosity varied from about 7.9 cp at 550°C to about 3.5 cp at 750°C and may be represented throughout this range by po= 0.1233 3427/T where T is in °K, UNCLASSIFIED ORNL-LR-DWG 5605 TEMPERATURE (°K) 700 800 200 1000 100 1200 3427/T(°K) ——=p=01233 e © BROOKFIELD VISCOMETER ® CAPILLARY VISCOMETER 8 }_ = oy ) Q 0 > i 400 500 800 700 800 900 1000 TEMPERATURE (°C) Fig. 8.6. Viscosity of NoF.KF.ZrF, (5-52.43 mole %) Obtained by Both the Brookfield and the Capillary Viscometers. Measurements were mode on two beryllium- bearing mixtures in a small, disposable drybox fabricated expecially for beryllium work from « 20-gal drum. The data were taken with capillary viscometers, The viscosity of NaF-BeF, (57-43 mole %) varied from about 18,5 cp at 550°C to about 5.2 cp at 750°C (Fig. 8.7) and may be represented throughout this range by g = 0.0308 5240/T | 143 ANP PROJECT PROGRESS REPORT UNCLASSIFIED ORNL~LR—DWG 5606 TEMPERATURE {°K) 700 800 200 1000 1100 1200 1300 ] . T T T TTTTITTIT T T T T T - | _ 1ol ~ 1 NeF —Bef,(57—43 mole %) 4\ N 2=00308 £ 5240/T (oK)~ : \ o =10 —— l \g —L ' roF NN H I | - N g [ [ —1 NN Bl = 5 A ‘ | . _ < | i 1 i L|F BeF2(69 31 mole "Fo) —‘ p= —0.148¢3650/T(°K) ¥ 5 , \ L 0 CAP\LLARY VISCOMETER NO 27 \ ® CAPILLARY VISCOMETER NOQ. 28 \ L] N 300 400 500 600 TOO 800 90Q 1000 TEMPERATURE (°C) Fig. 8.7. Viscosities of NciF-BeF;2 (57-43 mole %) and LiF.BeF, (69-31 mole %). where T is in °K. The viscosity of LiF-BeF, (69-31 mole %) varied from about 10 cp at 550°C to about 3.5 ¢p at 800°C (Fig. 8.7) and may be represented throughout this range by = 0.118 6’3650/T where T is in °K. It was expected that the vis- cosity of the LiF-BeF, mixture would be lower than that of the NaF.BeF, mixture because the former has a lower density. However, the vis- cosities of both are higher than would be predicted from the general trend for other fluorides, which indicates a proportionality between the viscosities and densities. The beryllium-bearing mixtures are apparently glass-like in character. THERMAL CONDUCTIVITY W. D. Powers Reactor Experimental Engineering Division A radial thermal conductivity apparatus has been fobricated, and a drybox has been designed and fabricated to be used with this apparatus. The sample to be studied in this apparatus will be contained between concentric cylinders. A known amount of heat will be generated in the inner cylinder, and the temperature difference across the sample will be measured. The thermal conduc- tivity may then be calculated from these measure- 144 ments and the physical dimensions of the system. Measurements were made on water for a number of different heat flows. For low heat flows the calculated thermal conductivity is constant, but when the heat flow reaches a critical value the conductivity increases with increasing heat flows. Above the critical value, heat is transferred not only by conduction, but also by free convection. A plot of the observed conductivity of water vs the heat flow measured in watts per inch of sample is presented in Fig. 8.8; the results are in good agreement with the known values. A radial thermal conductivity apparatus is being designed for de- termining the thermal conductivity of lithium hydride. UNCLASSIFIED QRNL-LR-OWG 5807 0.6 l : ‘ T ‘ ’ ‘j R R R N ~ ’ o * ' c. = . . L. @ N r S LITERATURE VALUE | i E e —— l_—f,‘_—,‘ 5 \ \ ‘ > a ‘ ‘ _ Z | | 8 . ‘ : ‘ ! Q2 ——— — - —_— . - _ — ' 2 ‘ s > | < | ® EXPERIMENTAL DETERMINATION ot F— ‘ - ' O S R R —— 0 2 a 6 8 10 12 HEAT FLOW (w/in.) Fig. 8.8. Thermal Conductivities of Water Measured by the Radial Device and Compared with the Literature Yalue. ELECTRICAL CONDUCTIVITY N. D. Greene Reactor Experimental Engineering Division The platinum electrical conductivity cell has been completed and standardized at room temper- ature. An attempt to standardize the cell at high temperatures was forestalled by electrode ex- pansion and binding of adjacent components. The machine work necessary to alleviate this problem is now being undertaken. INFLUENCES OF THE PHYSICAL PROPERTIES ON REACTOR HEAT TRANSFER H. F. Poppendiek Reactor Experimental Engineering Division It is of interest to compare the heat and momentum transfer characteristics of the ART for the three different types of fluoride fuels now being con- sidered, namely, lithium-base, zirconium-base, and beryllium-base fuels. The core, as well as the fuel-to-NaK heat exchanger, must be considered. In the case of the core, the wall-to-fluid temper- ature difference which must be reduced by cooling to prevent excessive wall temperatures is of prime importance. For a given core geometry, power density, and system pressure drop the wall-to-fluid temperature difference can be related analytically to the physical properties. Calculations for the three different types of fuels indicate that the wall-to-fluid temperature difference for the zir- conium~base fuel will be about two times as great as that for the lithium-base fuel and that the wall-to-fluid temperature difference for the be- ryllium-base fuel (based on somewhat limited physical property data, at present) will be about 1.3 times as great as that for the lithium-base fuel. PERIOD ENDING MARCH 10, 71955 In the case of the heat exchanger, a comparison has been made on the basis that the sum of the radial temperature difference plus one half the axial fluid temperature difference is equal to a constant; if all other parameters remained the same, this condition would always yield the same air outlet temperature in the radiator. An analysis was made for a typical ART heat exchanger by the method described previously.® The results indicate that the zirconium-base fuel has a Reynolds number of about 3000 and a corresponding pressure drop about 1.6 times as great as that for the lithium-base fuel and that the beryllium-base fuel has a Reynolds number of about 2300 and a corresponding pressure drop at least 1.4 times as great as that for the lithium-bose fuel. The Reynolds number for the lithium-base fuel is about 4000. The Reynolds number of the beryllium-base fuel is low because its viscosity is relatively high and its density relatively low (high kinematic viscosity). 5M. W. Rosenthal, H. F. Poppendiek, and M. R. Burnett, A Method for Evaluating the Heat Transfer Effectiveness of Reactor Coolants, ORNL CF-534-11-63 (Nov, 4, 1954). 145 ANP PROJECT PROGRESS REPORT 9. RADIATION DAMAGE D. S. Billington J. B, Trice Solid State Division The program of irradiating Inconel capsules con- taining fluoride fuels in the MTR is continuing. A new method for resistance welding thermo- couples to the capsule surface coupled with a new control system has resulted in much improved temperature control during the irradiations. Design work has been completed on the miniature in-pile loop, and many of the parts have been fabricated, The miniature sump pump for the loop has been tested and found to be satisfactory. The horizontal-beam hole in-pile loop was operated in the LITR and is now being disassembled for metallographic examination. A pneumatic flux- measuring device is being used for preliminary measurements of the thermal-neutron flux to be expected in the fuel region of the MTR in-pile loop. Approval was obtained for irradiation of the stress-corrosion apparatus in the LITR, and examinations of the first irradiated specimen are under way. The MTR creep apparatus has been shipped to NRTS, MTR STATIC CORROSION TESTS W. E. Browning G. W, Keilholtz Solid State Division H. L. Hemphitl Analytical Chemistry Division irradiations in the MTR of Inconel capsules con- taining fluoride fuels are continuing but have been somewhat retarded because of interruptions in the MTR operating schedule. ments Three recent improve- in the MTR capsule irradiation facility, which is being used to study static corrosion and chemical stability in ANP fuel-container systems, have consisted of a method for assembly of the air annulus onto the capsule prior to insertion, a controller with faster temperature and air flow response, and an improved method for fabricating thermocouples and for attaching them to the fuel container walls. The revised capsule—qir annulus assembly, which is already being used in the MTR, was shown schematically in the previous report.! The arrangement has the advantages that the annulus alignment can be performed and visually inspected outside the reactor and that all the thermocouples 146 can be attached to the capsule assembly before insertion into the reactor. Also, the problem of maintaining thermocouples in the permanently installed part of the facility to indicate capsule alignment has been eliminated. sleeve is easily removable in the hot cells in the post-irradiation examination of the capsules. The new system of control for maintaining a steady capsule temperature through the proper metering of cooling air to the capsule has de- rivative action plus proportional control and fast reset. This combination of controls has been shown in bench tests to quench thermal oscil- lations caused by inherent instabilities in the system and to request an increase in cooling air from the controller with sufficient speed to handle any foreseeable sudden increases or decreases in reactor power. It is now operating successfully in the MTR capsule facility. Two improved methods for welding thermocouples to the Inconel capsules for the present series of tests were made by R. J. Fox of the Central Machine Shops. One method (illustrated in Fig. 9.1) involves crossing the chromel and alumel wires, pressing them against the capsule with a copper electrode, and passing current through them to weld them to the capsule by resistance heating. The other method uses the same welding tech- nique, but the wires are laid parallel along the surface before the current is applied. In each case spot-welding parameters were optimized for producing a thermocouple bead which appeared, under the microscope, to have the best shape for good heat transfer and mechanical strength. These methods are better than the previous method, in which the thermocouple was fused to the Inconel surface with an arc discharge, in that a thermo- couple junction is obtained which lies closer to the Inconel capsule surface and therefore yields a better measure of the capsule wall temperature. A comparison of the crossed-wire type of re- sistance-welded thermocouple and the previous spark-welded type of thermocouple is presented The annulus w. E. Browning, G. W. Keitholtz, and H. L. Hemphill, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 120. UNCLASSIFIED PHOTO 13864 TRIMMED THERMOCOUPLE -« DIRECTION OF AIR FLOW AS-WELDED THERMOCOUPLE |, ¢ REACTOR HL_’_’ JLA—IZJA‘ 5 L Fig. 9.3. Installation of Vertical In-Pile Loop in Position C.48 of the LITR. in the event the in-pile loop wall ruptured and allowed all the fuel to spill into the bottom of the stainless steel jacket of the loop. The bottom of the loop will be in a very high thermal-neutron flux region and, hence, in a very sensitive region with respect to influence on reactivity of the LITR. Since flux calculations indicated that an excess reactivity increase of 2 to 4% might occur, it was decided to mock up the situation experi- mentally to obtain a more accurate prediction of the increase. For the experiment the fuel mixture was contained in an annulus around a long rod, as shown in Fig. 9.6. This assembly was placed in an outer aluminum jacket, which formed an air annulus, and then lowered into the C-48 fuel element position in the LITR. The fuel mixture used contained 120 g of U233, which is within 12 g of the amount which will be used in the first test. The annular geometry was chosen to repre- sent the conditions which would produce the highest increase in Ak/k. UNCLASSIFIED 55D0~B-10444 ORNL-LR-~DWG 3222A T .l ol PUMP SPEED, 5000 rpm TN t% -in--dia STRAIGHT -VANE 46 | e e SN IMPELLER HEAD (ft OF H,0) o 14 R l Il I k | I i | ! I | | : [ [ i | | I ! | e - 6 p— - — o G fommemm e e f——DESIGN POINT: 1480 cm¥/min = \ 2 : 4 fps (0.200-in.-1D TUBE) \ | 0 | 0 1000 2000 3000 4000 5000 FLOW (cm3/min) Fig. 9.4. Head vs Flow Characteristics of Miniature Sump Pump for In-Pile Loop. 58D-A-1040a 104 ORNL-LR-DWG 3221A T T T T T T T T T - @ EXPERIMENTAL RESULTS IN METERED VOLUME OF FUEL — ™ 0 EXPERIMENTAL RESULTS IN FUEL WITH - — VENTURI FLOWMETER o — - EXPERIMENTAL RESULTS IN WATER 1~ =" : -~ -~ . 5 _ . 9’4 D . € 0e /.‘/// & ”* e - o + Ta f// A ol l - T N o - 2 - 547 = /\fi’/// S 2f - e | T o EXPECTED LIMITS OF ACTUAL HEAD 7 (CENTRIFUGAL PUMP) | - // 10° | I 1 2 5 10 20 30 CUTOFF HEAD (f) Fig. 9.5. Head vs Pump Speed of Miniature Sump Pump for In-Pile Loop. PERIOD ENDING MARCH 10, 1955 The excess reactivity resulting from this simu- lated accident was determined by measuring the UNCLASSIFIED SSD-C-11t44 ORNL-LR-DWG 50704 FILL TUBES PINCHED OFF AND WELDED AFTER ANNULUS IS FILLED\ ANNULUS\ 0.203in —* Fig. 2.6. Excess Reactivity Apparatus. 149 ANP PROJECT PROGRESS REPORT LITR control rod position with the mockup in and with the mockup out of the lattice. The LITR rod calibrations were checked by following the known xenon buildup in the reactor. After a correction was made for buildup of xenon during the experi- ment, it was found that the simulated accident would odd less than 0.2% excess reactivity, which can be handled easily by the LITR control rods. LITR HORIZONTAL-BEAM-HOLE FLUORIDE FUEL LOOP J4. G. Morgan M. T. Morgan 0. Sisman W. E. Brundage C. D. Baumann A. S, Olson R. M. Carroll W. W. Parkinson Solid State Division The circulating-fluoride-fuel experiment has been successfully conducted in hole HB-2 of the LITR, and the loop is being disassembled for metallo- graphic examination in the hot cell of Building 3025. The general description of the apparatus and facility, as presented previously,Z applies to this test, with a few modifications. All loop welding was supervised by the Metal- lurgy Division, and the recommended specifi- cations were met. The schematic drawing shown in Fig. 9.7 gives the details of the construction of the loop. By x-raying the cairod heaters to determine their effective heating lengths, it was possible to position them in such a way as to avoid hot spots on the fuel tubes. After the loop was assembled, it was leak tested at 1500°F with a helium leak detector, and then the loop and the pump were brought to 1500°F under a helium atmosphere. The molten salt NuF-Zer (50-50 mole %) was then charged into the loop to the level of the lower probe in the pump bowl. After further leak checks the loop was operated for 7]/2 hr at 1500°F with flow rates of 5 to 15 fps based on the 0.225-in.-ID nose piece section. The loop was then drained and cooled under g helium atmosphere. The filling and draining methods used are shown schematically in Fig. 9.8. In filling the locp all salt lines were heated to 1500°F under a purified helium atmosphere. With the pump idling, the salt tank was pressurized and the salt forced into the loop until it made With contact with the probe in the pump bowl. 2W. E. Brundage et al., Solid State Semiann. Prog. Rep. Aug. 30, 1954, ORNL-1762, p 21, 150 the pump turned off, all pressures were equalized and the fill line was frozen at point B, In draining the loop, flow was stopped, and a portion of the loop on the inlet side of the pump was frozen off. By using the draining-pressure system, the salt was then forced back into the salt tank until the fill line was flushed empty, The loop was inserted into LITR hole HB-2 on December 7, 1954, with the reactor shut down, and electrical and piping connections were made before the shielding was completed. After bringing the loop to temperature it was again filled with the barren salt mixture, operated for 8% hr, and drained. The enriched fuel NaF-ZrF -UF, (62.5- 12.5-25 mole %) was then charged into the loop and the fill line was sealed off. When 4.86 kg of the enriched mixture had been added, the loop was filled to the bottom proke of the pump bowl. With the fuel circulating at temperature and the reactor at zero power, the remainder of the external shielding was added. A cutaway sketch of the external shielding required is shown in Fig. 9.9, Because of the difficulty in placing the paraffin close to the hot pump enclosure, more gamma shielding was required around the periphery than was originally planned. Additional concrete block walls were placed in front of the instrument panel for protection of the operating personnel. Full reactor power was reached ot 4:30 PM on December 11, and the electric heaters were ad- justed to maintain the desired operating tempera- ture. Flow during the entire run was maintained at 8 to 10 fps (Reynolds number of 5000 te 6500). A composite plot of reactor power, loop tempera- tufe, and salt Reynolds number vs time is shown in Fig. 9.10. The total power generation of the fuel under reactor flux was determined by a series of heat balances with the reactor off and at full power. Equilibrium conditions were established with the reactor at full power, and then upon reactor shutdown the electric heaters were increased to match the same temperature conditions along the loop. This increase in electric power to duplicate thermal conditions under flux was taken as the power added to the system by the fissioning fuel. Several such measurements gave an average of 2800 w, or about one-third the anticipated power. An experiment to duplicate the perturbed flux pattern is being conducted to determine the highest wattage densities obtained in the loop LGt 0L IN THROUGH SHAFT oren s s06s HELIUM EXHAUST- ~JACKET ASSEMBLY _--~SPARK PLUG PROBE ClL INLET FOR RING COOLING —~{1 4 RADIATOR _.—CENTRIFUGAL PUMP MODEL LFA I _—-——CONTROL GAS -~ FUEL TUBING KOVAR SEAL .. ) FITTINGS —* - PUMP ENCLOSURE _ _NOSE COVER DRAIN TUBE SEAL-. e ALSMAG 202 "} ~HELIUM ATMOSPHERE zd HEATER CORE EXPANSION BELLOWS -, i T NICHROME V . CORE HEATER e [ ‘*—f% . ! SECTION A-A WATER OUTLET EXPANSION BELLOWS-~_ N JACKET ASSEMBLY-. HEAT EXCHANGER W ' LINER SEAL CADMIUM SHEILD~ _Vs-in. DIA SUPPORT ROD N [ {AND FLANGE HELIUM PURGE LINE - N / _FUEL TUBING 3 ~CALRORS 1oy e . 0" RING \ ) CFUELTUBING | ‘. \ /o (OUTS | \ \33 P AR INLET v \ - 5 RADIATOR - y | ANNULLS FOR / \ VENTURI- i/ ' | BORON SHIELD~ /. WATER FLOW-~ “TRANSITION N3, “-FUEL TUBING {IN) “KOVAR SEAL ,_/ PLATE /a-in_DIA x FITTINGS NOSE COVER SUPPORT ROD FUEL LEADS *PRESSURE N TO DIAPHRAGM” TRANSMITTER CELLS %~ ~WATER INLET HELIUM LEADS TO 20 in. REF — ———————— : DIAPHRAGM “———12ft 5in. REF ~———————— e Fig. 9.7. Schematic Diagram of LITR Horizontal-Beam-Hole Fluoride-Fuel Loop. §S61 ‘0L HOY¥YW ONION3 doiy3d ZGi UNCLASSIFIED SSD~-B— 11134 ORNL-LR—-DWG 5065A CHECK VALVE FREEZE-OFF LINE =" B - LOOP OIL-FILLED BURBLER - / < —3 ¢ FILL LINE — \ . >> A/ (N { | '[ | \ CHARGE TANK | I | MELT-0UT FURNACE — =i HIGH-PURITY HELIUM DRAINING HIGH - PURITY HELIUM PRESSURE SYSTEM Fig. 9.8. Fill-and-Drain System for LITR Horizontal-Beam-Hole Fluoride-Fuel Loop. L30d3Y §534908d LID3F0dd ANV test and to see why the total power was lower than expected. On January 7 the drive belt to the pump motor became inoperable, and, with the reactor shut down, a portion of the external shielding was removed and the belt replaced. It was found to be impossible to get circulation started again because of a cold region in the loop, and therefore the test was terminated. The loop was withdrawn from the reactor and taken to the hot cells, Building 3025, for disassembly. The tubing will be sectioned for metallographic examination. The loop had operated for a total of 645 hr, with 475 hr at full reactor power. CONCRETE BLOCK SHIELD 6 ft HIGH PERIOD ENDING MARCH 10, 1955 The flow-measuring device described previously? proved to be very successful and enabled con- tinuous monitoring of the salt velocity to be recorded. Design changes were made (Fig. 9.11) that permitted the transmitter to withstand a pressure as high as 10 psi across the diaphragm without causing a pressure shift. The transmitters were subjected to pressures as high as 40 psig during operation with the first filling of npon- uranium-bearing salt. Each cell has a sensitivity of about 0.1 psi. k 3W. E. Brundage et al., Solid State Semiann. Prog. Rep. Aug. 31, 1953, ORNL-1606, p 29. PR SSD-B-1121A ORNL-LR-DWG 53424 HB-2 LITR SHIELD 400 mr/hr 7 mr/hr HB-3 EQUIPMENT SHIELD L PUMP % 16in. —m=t 3S5Smr/hr &0 mr/hr 3-in. PARAFFIN \H1GH~DENS|TY BLOCK walLL gft HIGH /___ CONTROL PANEL Fig. 9.9. Horizontal Section Through Shielding Required During Operation of LITR Horizontal-Beam- Hole Fluoride-Fuel Loop. Gamma fluxes indicated. 153 ¥Sl 1500 1450 1400 NOSE TEMPERATURE (°F) 6500 6000 5500 5000 43500 REYNOLDS NUMBER 4000 REACTOR POWER ( Mw) TIME {days) Fig. 9.10. Reactor Power, Flyoride Fuel Reynolds Number, Operating Time. ORNL—LR—DWG 5343 A and Temperature of Nose of Loop vs LY0dIY SSIYI03d LDIrodd dNV UNCLASSIFIED 550-A—-1t09A4 ORNL-LR-DWG 5065A HELIUM GAS LINE TO CONTROL PANEL MODIFIED CHAMPION V-1 SPARK PLUG WITH PLATINUM TIP PLATINUM CONTACT WELD LIQUID LINE TO VENTUR| Fig. 9.11. Pressure-Transmitter Cell. FLUX-DEPRESSION EXPERIMENTS IN MTR J. B. Trice H. V. Klaus Solid State Division R. H. Lewis Phillips Petroleum Company Preliminary measurements of the thermal-neutron flux to be expected in the fuel region of the first in-pile loop scheduled to go into hole HB-3 of the MTR are being made by using the pneumatic flux-measuring device which is presently in hole HB-3. The decrease in the flux of the MTR that will be caused by the fuel and the fuel container, as well as by the auxiliary equipment of the in-pile loop, is being determined. The loop, which will be made of lnconel, was mocked up by using a straight Inconel tube filled with a simulated fuel, which consisted of a mixture of cadmium and magnesium for one series of tests and a mixture of aluminum and boron carbide for another series. The preliminary results of the flux-depression experiment indicate that for the fuel which is to be used in the first in-pile loop tests, the de- pression will be on the order of 70%, which means, in terms of loop power, that the power of the presently conceived loop design may be expected to be 5 to 10 kw rather than the desired power of 15 to 30 kw. The nose of the loop is therefore being modified to increase the expected power. Before the loop is inserted in the MTR, it is PERIOD ENDING MARCH 10, 1955 planned that a full-scale mockup of the loop will be placed in the MTR to determine the expected power (cf., sec. 3, ‘‘Experimental Reactor Engi- neering’'). CREEP AND STRESS-CORROSION TESTS W. W. Davis J. C, Wilsen N. E. Hinkle J. C. Zukas Solid State Division Approval by the ORNL Experiment Review Com- mittee for irradiation of the stress-corrosion appa- ratus previously described? was received shortly after the completion of bench tests involving compatibility of component parts in case of sodium leakage. Alarm circuits to signal sodium leakage, water in-leakage, and excessive temperatures were installed in the apparatus and it was inserted in HB-3 of the LITR. The Inconel fuel chamber, which contained 0.52 g of Nc:l"'—ZrF“-UF4 (63- 25-12 mole %), was surrounded by approximately 25 g of sodium. Because of the rapid power changes in the furnace required to maintain the specimen control temperature at 1500°F during reactor startup or shutdown periods, a Leeds & Northrup Speedomax and air-controller combination was used. Thermocouples in wells in the sodium chamber recorded fluctuations that did not exceed 10°F at the outset of the test; the fluctuations gradually decreased to one-half this value after 135 hr of test, and the temperature remained steady thereafter, The chamber was ot control tempera- ture for 1120 hr, during which time the reactor was up to power for approximately 700 hr. Periodic checks of the resistance between the stressing weight and the weight probe throughout the test indicated that no gross increase in creep raote was caused by irradiation at a stress of 1000 psi. The rig was removed to a shield to decay sufficiently to permit handling and dis- secting. The transverse cross section of the specimen below the fuel level will be polished and etched for metallographic examination. A companion bench test is now in operation. Several attempts to change or eliminate the baffle ar- rangement now in use to simplify both assembly and sectioning of the apparatus reintroduced the temperature excursions which were so troublesome at the outset of design of the rig. 4J. C. Wilson er al.,, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 142. 155 ANP PROJECT PROGRESS REPORT Operation of the present apparatus in the LITR is satisfactory, but its performance in the MTR is not assured because of gamma heating in the relatively massive apparatus required to achieve smooth temperature control, Stress-corrosion data are vurgently needed at fuel power densities of 1000 w/cm® in the MTR, and therefore a possible short-cut stress-corrosion apparatus is being mocked up. The specimen tube consists of a cylinder stressed by gas pressure on the inside and in contact with fused salts on the outside. The salt is in an annular space around the specimen tube; the outside of the salt annulus is the inside of a container tube that is finned on the outside to transfer heat to helium in a water-jacketed con. To preserve a low surface- to-volume ratio it was suggested (by R. G. Berggren) that the salt be allowed to remain solid at the container tube wall. This cuts the surface- to-volume ratio approximately in half if any inter- action between the solid salt and the container wall is ignored. Whether good heat transfer can be obtained between the solid salt and the wall remains to be determined. The experiment is unique in that the hottest part of the molten salt is in contact with the specimen tube (at 1500°F). In capsule tests the inside of the salt volume has frequently been several hundred degrees hotter than the salt-metal interface. Since rupture of one of the metal members is a possibility, cooling in a gas stream is probably unsafe from a reactor operations standpoint. Con- vection cooling by fins to helium in a water jacket or conduction by fins to a water-cooled heat sink appears to be attractive. Experiments on con- vective heat transfer from fins have shown that there is a good chance of success. [f this is not successful, conductive heat transfer will achieve the results. Empirical heat transfer data are being obtained for longitudinally finned, vertical cyl- inders, and this work will be extended to cover conducting fins. Generation of sufficient heat in small cylinders to simulate the fission heat has 156 proved to be difficult, Passage of an electric current through the finned tube works well for low-conductivity fins, and carbon-arc and platinum- radiation heaters are being tried. As soon as. dota can be reliably extrapolated a test will be put in the LITR. The MTR creep apparatus and accompanying instrumentation have been shipped to the MTR. Details to be supplied for approval of the stress- corrosion rig have been assembled but cannot be completed until the first LITR test has been analyzed. Tests have begun on an electromagnetic trans- ducer that is advertised to be operable at 1300°F. Similar to a microformer in principle, it is ‘‘canned’’ in stainless steel and has ceramic insulation on all windings. The device will be tested under irradiation upon completion of bench tests. If operable in-pile, the transducer will be suitable for strain measurements in the bending- type stress-corrosion apparatus. An extensometer for the internally pressurized ‘‘tube burst’’ speci- men is possible, but no development work is being carried out. For strain data, the internal diameters of the specimen tube will be measured with o pneumatic goge after irradiation, and multiple specimens will be irradiated, if neces- sary, to obtain a strain-time curve. Two bench tests of Inconel tubes in bending in a helium atmosphere with NaF-ZrF -UF, (53.5- 40-6.5 mole %) were completed. After 432 hr at 1500°F a few subsurface voids were visible that were distributed about equally around the periphery of the tube. Another test was operated for 866 hr, and the number of subsurface voids per unit length of periphery was greater by a factor of 2 at the tension and compression sides of the tube than at the neutral axis. This tentatively confirms the hypothesis that stress-corrosion in the usual sense does not take place in these tests and that the phenomenon should be called “*strain-rate’’ or **strain’’ corrosion. PERIOD ENDING MARCH 10, 1955 10. ANALYTICAL STUDIES OF REACTOR MATERIALS C. D. Susano Analytical Chemistry Division J. M. Warde Metallurgy Division Research on the determination of frivalent uranium and uranium metal in fluoride-base reactor fuels was continued. Further studies were made of the methods for determining oxygen as oxides in fluoride salts, A bromination procedure was applied to the determination of oxygen in uranium and beryllium compounds. Modifications were developed of methods for determining beryllium, potassium, and lithium in fluoride fuels. High- temperature x-ray spectrometer studies of fluoride mixtures were initiated as an aid in the determi- nation of phase diagrams. ANALYTICAL CHEMISTRY OF REACTOR MATERIALS J. C. White Analytical Chemistry Division Determination of Trivalent Uranium in Fluoride Fuels A. S. Meyer, Jr. D. L. Manning W. J. Ross Analytical Chemistry Division Oxidation of Trivalent Uranium by Methylene Blue. Evaluation of the data obtained for the determination of trivalent uranium in fluoride fuels by the methylene-blue, one-step oxidation method’ has shown that the coefficient of variation for NaF-KF-LiF-base materials is 2%. The coeffi- cient of variation of the hydrogen-evolution method? for similar materials, including NaF-ZrF -base eutectics, is 4%. In LiF, NaF, and a mixture which contains NaF-KF-LiF, the agreement be- tween the results for trivalent uranium obtained by both methods is, in general, satisfactory. The results of the determination of trivalent uranium by the use of methylene blue in compositions of NaF-ZrF ,-UF ,-UF ; and KF-UF ;-UF, are difficult to reproduce and are significantly lower than those ]A. 5. Mevyer, Jr., et al., ANP Quar. Prog. Rep. Dec, 10, 1954, ORNL-1816, p 130. 2D. L. Manning, W. K. Miller, and R. Rowan, Jr., Methods of Determination of Uranium Trifluoride, ORNL-1279 (Apr. 25, 1952). obtained by the hydrogen-evolution method.. Since the major portion of the samples received for analysis are from the NaF-ZrF, system, an effort has been made to adapt the methylene-blue method to these samples. The low values obtained by this method may result from incomplete dissolution of the samples or from partial oxidation of the trivalent uranium by ionic hydrogen. It has been found that these samples can be dissolved if stirred for 2 hr in a methylene-blue solution that is almost completely saturated with AICI,. Al- though more reproducible results are obtained when complete dissolution is attained, the in- stability and heterogeneity of the samples make impractical a direct comparison of the results of the modified methylene-blue procedure with those of the hydrogen-evolution method., Accordingly, dissolutionis now being carried out in an apparatus in which any hydrogen that is formed by the reaction UF, +H" — U4" +3F- + L H, can be measured., Preliminary experiments on the dissolution of samptes of UCl, have indicated that when the samples are dissolved in solutions of methylene blue that are 3 M in HCl and saturated with AICI, a small quantity of hydrogen is liberated but that the volume of hydrogen is negligible when the con- centration of HCl in the reagent is 1.5 M. In solutions of lower acidity the oxidation of trivalent uranium is carried beyond the tetravalent state. Dissolution tests in solutions of intermediate acid concentration are now being performed. Simultoneous Determination of Trivalent and Total Uranium. Further efforts were made to develop a procedure for the determination of the total uranium content of these samples by direct titration of the tetravalent uranium in the solution after the determination of the trivalent uranium by the methylene-blue method. Potentiometric titra- tions of the reduced solutions have been carried out with K2Cr207, Ce(504)2, KMnO4,Gnd Fe2(504)3 as oxidants. The titration with each of the above reagents is too slow for application to analytical 157 ANP PROJECT PROGRESS REPORT procedures. Titration was not improved by re- ducing the chloride ion concentration or by adding H,S0, or H,PO,. The solutions could not be titrated at elevated temperatures because the methylene blue decomposed rapidly when heated. When the solutions were treated with an excess of oxidant and back-titrated with a standard solution of ferrous sulfate, a portion of the excess reagent was reduced, either by chloride ion or by methylene blue. During the course of the above investigations, solutions of UO,‘_,SO4 in an excess of methylene blue were titrated with a solution of CrS0,. Definite potentiometric end points were obtained that were similar to those reported ! for the titration of solutions of tetravalent uranium in methylene white with K,Cr,0,. The volume of titrant con- sumed at the first end point, which coincided with the decoloration of the methylene blue, corres- ponded to the reduction of the methylene blue to methylene white plus a one-electron reduction of the hexavalent uranium. An additional equivalent of Cr50, per mole of uranium was required to titrate to the second end point. When solutions which contained a molar excess of U02504 over methylene blue were titrated, a similar titration curve was obtained in which two equivalents of titrant per equivalent of methylene blue were re- quired to titrate to the first end point. In order to eliminote the green color of trivalent chromium the titrations were repeated with solutions of trivalent titanium. At the first end point a colorless solu- tion was obtained, and on the addition of an excess of titrant the green color of tetravalent wuranium was developed. The absorption spectra of the solutions that had been titrated to the first end point did not correspond to those of solutions of any combination of tetra- and hexavalent uranium. These results are consistent with the postulation that an interaction species of pentavalent uvranium and methylene white is stable in aqueous solu- tions. A more detailed study will be necessary before the existence of such a complex can be established. The only procedure which appears to be practical for the determination of total uranium in these solutions consists of destroying the methylene blue and removing the chloride by a wet-oxidation procedure; the uranium content of the resulting solution can then be determined by a conventional titration method. Methylene blue was found to be rapidly oxidized by fuming 158 with HNO, and HCIO,,, Determination of Trivalent Uranium by the Karl Fischer Modification of the Hydrogen-Evolution Method. A procedure is being tested in which the hydrogen that is evolved by the reaction between UF, and an acid solution is converted to water, and the water is determined by a modified Karl Fischer titration. The hydrogen-evolution method? is limited to samples which contain enough tri- valent uranium to liberate a volume of hydrogen that is sufficient for precise volumetric measure- ment (1 mg of U liberates ~0.04 ml of H,, STP). An additional limitation of this method is that all gases which are insoluble in solutions of KOH are measured as hydrogen. In the analysis of samples which yield only a few tenths of a milli- liter of hydrogen, adsorbed gases and gaseous con- taminants of the reagents and the sweep gas may introduce serious error., In the procedure being investigated the sample is dissolved in an 80% solution of HCI. The hydrogen liberated is passed through concentrated H,50, and two drying towers of MgClO, by a stream of purified helium. The dried gases are then passed over CuO at a temperature of 500°C to convert the hydrogen to water, which is then absorbed in a solution of Karl Fischer reagent in ethylene glycol and finally titrated coulometrically with iodine. When samples of UF, that contained from 4 to 20 mg of trivalent uranium were analyzed by this procedure, the titrations corresponded to 96% of the theoretical trivalent uranium content, with a coefficient of variation of 5%. The precision of the method appears to be limited by the large and variable blank titrations which are probably a result of a mixing between the solutions in the cathode and ancde compartments of the coulometric titration cell. The cell is being modified in an attempt to reduce this source of error. Determination of Uranium Metal in Fluoride Salt Mixtures A. S. Meyer, Jr. B. L. McDowell Analytical Chemistry Division Further studies were carried out to improve the method for the determination of uranium metal in fluoride fuels by converting the metal to the hydride and measuring the quantity of hydrogen evolved by the thermal decomposition of UH,. Although results of satisfactory precision and accuracy can be obtained by the procedured in which the decomposition of the hydride is carried out under an atmosphere of CO,, it was found that for some samples of UF ., periods in excess of 8 hr were required for the complete evolution of hydrogen. In addition, no completely dependable method has been found for removing the CO formed when the CO, is reduced by uranium metal and by trivalent uronium., Although the CO can be quantitatively oxidized to CO, by passing the gaseous reaction products over a mixture of 1,0, reagent and powdered pumice stone at a tempera- ture of 150 °C, the 120 reagent has been found to become inactivated after limited, but varied, periods of service., The reagent is, therefore, of doubtful value when extended ignition periods are required to complete the decomposition of the hydride. In order to eliminate the possibility of interference by CO and to reduce the ignition time, modifications investigated in which the decomposition of the hydride was carried out under atmospheres of HCl and NH.,. When NH, was used, the hydrogen was measured over a solution of H2504 instead of KOH. In the presence of each of the gases the volumes of hydrogen obtained from the decomposition of UH, were not reproduc- ible and were less than those predicted by the postulated reactions: UH, + 3HCl —> UCI, +3H, 2UH, + 4NH, —> 2UN, + 9H, As an additional complication the decomposition of NH, is catalyzed by the residue of uranium nitrides to produce a slow evolution of insoluble gases which continves for several hours after the initial, rapid evolution of hydrogen from the reaction between NH, and UH.,. A modification is now being investigated in which UH, is ignited in a stream of oxygen, and the effluent gases are passed over heated CuO to ensure the conversion of hydrogen to water, which is then measured volumetricaily at reduced pres- sure. The reaction between UH; and oxygen has been reported? to be rapid and quantitative when applied to the determination of hydrogen in large samples of UH,. A modification of the method of were 3A. S. Meyer, Jr,, and B. L. McDowell, ANP Quar. Prog, Rep. Dec. 10, 1954, ORNL-1816, p 129, 4. C. Warf, The Composition of Uranium Hydride and Its Decomposition at 250°C, CC-105% (Oct. 9, 1943). PERIOD ENDING MARCH 10, 1955 Naughton and Frodyma® for the microdetermination of carbon and hydrogen will be used for the deter- mination of microgram quantities of hydrogen as UH,. In this method the water produced by the reaction between UH, and O, is first isclated in an evacuated freezeout trap. It is then allowed to volatilize into an evacuated vessel of known volume, and the equivalent quantity of UH, is calculated from the pressure which is measured on a mercury-sealed, oil manometer. The apparatus has been constructed and is now being calibrated by using samples of BaCi,.2H,0 as a standard for hydrogen in the form of water. On the basis of design calculations, it would appear that less than 1 ug of hydrogen can be determined by this technique. Determination of Oxygen in Fluoride Fuels A. S. Meyer, Jr. J. M. Peele Analytical Chemistry Division Further tests of the procedure® for the determi- nation of oxygen as oxides were carried out with new components which were incorporated in the apparatus to prevent the carryover of nonvolatile electrolytes during the transfer of HF to the con- ductivity cell. The new components include a larger reactor and a splash trap in the transfer line. The experiments indicate that, although the conductivity method for the determination of the water produced by the reaction of metallic oxides with KHF, is theoretically applicable, the method is not practical for the determination of microgram quantities of oxygen. It has been found that repeated distillations with HF are required to transfer the H,0 from the KHF, solution of the sample to the conductivity cell. In the course of these distillations, a sufficient quantity of KF is, apparently, carried into the cell to mask the in- crease in conductivity that would be produced by small quantities of water. An alternate procedure has been proposed for the determination of the H,O that is formed on dissolution of the oxides in KHF ,. Tests of this procedure are now being carried out in parallel with some further studies of the original method. 5J. J. Naughton and M. M. Frodyma, Anal. Chem, 22, 711 (1950). 6A. S. Meyer, Jr., and J. M. Peele, ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 148. 159 ANP PROJECT PROGRESS REPORT The new method is based on the electrolysis of the water in the fused bifluoride melt to yield oxygen which ¢an then be separated from the other reaction products and measured, Melts of KHF, have been reported to be readily dehydrated by carrying out the electrolysis until rapid evolution of fluorine occurs.” During this drying period, hydrogen is liberated at the cathode, while O,, OF,, and F, are liberated at the anode. Analysis of the product obtained during the initial electrolysis of a commercial fluorine generator® indicates that the H,O is removed during the early stages of the electrolysis and that less than 10% of the oxygen is evolved as OF,. The OF, is generated only from relatively wet melts, and negligible quantities are formed after the concen- tration of H,O in the electrolyte is reduced to a few tenths of a per cent. Tests of the method are now being carried out by using samples which contain milligram quantities of oxygen. For these samples the oxygen can be measured volumetrically by sweeping it into an azotometer with CO,. Hydrogen is eliminated from the electrolysis products by adding AgF to the melts so that metallic silver rather than hydrogen is deposited at the cathode. Two addi- tional advantages from the Ag* ions in the elec- trolyte may be anticipated. The standard oxidation- reduction potentials? in acid solution of pertinent couples are tabulated below: Reaction E° (v) U4+ 2H,0 — U0, ™ + 4H" + 2¢~ ~0.41 Ag —> A'g+ +e” —-0.80 2H,0 —> 0, + 4H" + 4~ 1.298 Agt — Agtt i e- 1.98 2F~ — F, + 2¢™ 2.85 The potential of the first reaction was obtained from the Handbook of Chemistry and Physics, 10 7'J. H. Simons (ed.), Fluorine Chemistry, |, 227, Academic Press, New York, 1950. BR. C. Dawning etal., Ind. Eng. Chem. 39, 259 (1947). 9W. M. Latimer, The Qxidation States of the Elements and Their Potentials in Aqueocus Solutions, p 296, Prentice-Hall, New York, 1938. IOC. D. Hodgman {Editer-in-Chief), Handbook o( Chemistry and Physics, 34th ed., p 1552, Chemica Rubber Publishing Co., Cleveland, 1952. 160 If the potentials of the above couples in the KF-HF system are of similar relative order, tetravalent uranium would be oxidized to the hexavalent state by Ag* ions. In the absence of the silver salt it would be necessary to oxidize the wuranium electrolytically before the oxygen could be quantitatively evolved. Since the method is to be applied to samples which contain tri- or tetravalent uranium as major constituents and only traces of oxygen, the time required for the elec- trolytic oxidation of the uranium would be prohibi- tive. Furthermore, if Ag** forms a stable solution in the fused electrolyte, it would be expected to act as a fluorine carrier and thereby increase the efficiency of the generation of O,. Under ideal conditions Ag** could serve as a coulometrically generated reagent for hydrogen in the fused salts. A silver-lined, nickel cell has been assembled for the purpose of studying this reaction, The cell is charged with sufficient KF, AgF, and HF, purified by distillation, to produce approximately 40 g of a composition of KF-2HF containing 5 g of AgF. The electrolyte is fused by heating the cell to 100°C, Air is removed from the cell by bubbling CO, through the melt, The effluent gas from the cell is freed of HF and fluorine by passing it through NaF and mercury, and the insoluble gaseous components are then collected and measured over KOH in an azotometer. The charge is electrolyzed until no more insoluble gas is obtained. The sample is then added to the cell and the gas that is obtained on electrolysis is measured as oxygen. A recovery of 96% of the oxygen present in the H,O produced when a sample of Na,CO, was dissolved in the electrolyte was attained, Difficulties in electrolysis have been experienced because of corrosion of the electrodes. When a platinum anode is used, only a small quantity of fluorine is liberated before the evolution of oxygen is completed. The anode is, however, corroded rapidly, particularly in the final stages of the electrolysis. When nickel ancdes are used, corro- sion is less severe, but the evolution of fluorine is so rapid that it cannot be conveniently removed by a mercury scrubber, An additional problem is presented in that after extensive periods of elec- trolysis, dendritic deposits of silver accumulate and grow to such lengths that they provide a metallic electrical circuit between the electrodes, it is believed that this problem will not be of major importance when samples which contain only microgram quantities of oxygen are analyzed. Determination of Oxygen in Metallic Oxides by Bromination J. C. White G. Goldberg J. P. Young Analytical Chemistry Division A simple, precise method for the determination of oxygen present as oxides in metals has leng been desired. Recently, Codell and Norwitz!! reported on the successful application of a bromi- nation procedure for the determination of oxygen in fitanium and titanium alloys. These investi- gators passed bromine at 815°C over a sample of titanium (which had been intimately mixed with graphite) to form CO, which was, in turn, con- verted to CO,, absorbed, and weighed. The reaction time was of the order of 2 hr. Experi- ments are now being made in an attempt to apply this technique to the determination of oxygen in metals of interest to the ANP program. The progress of experiments on uranium ond beryflium is presented here, Uranium. The apparatus for the determination of oxygen in uranium is shown in Fig. 10.1. Bromine vapor is carried by helium over a boat (maintained YIM. Codell and G. Norwitz, '‘New Method for De- termining Oxygen in Titanium and Alloys,”’ Chem. Eng. News 32, 4564 (1954). TYGON TUBING HELIUM DEHYDRITE ASCARITE X1 TUBE FURNACE /? 300-mi FLASK SCRUBBER BROMINE BUBBLER DEWAR FLASK ALCOHOL AND DRY ICE QUARTZ SAMPLE BOAT QUARTZ TUBE (14in. BY 1in.} g YQ |/‘_\"/lx:x:z::x PERIOD ENDING MARCH 10, 1955 at about 950°C) that contains a mixture of a uranium compound and spectrographically pure graphite in a weight ratio of about 1 to 10, The volatile reaction products are condensed in two dry-ice—alcohol baths and the COis passed through CuO, where it is converted to CO, and then absorbed in a standard solution of Ba(OH),. The reactions involved when UO, is the standard material are UO2 + 2Br, + 2C —> UBr, + 2C0O CO + Cu0 — CO;2 + Cu CO, + Ba(OH), —> BaCO, + H,0 The data indicate that for quantities of UO, of the order of 100 mg, the first reaction is quantita- tive in 2.5 hr at 950°C. A blank of 0.4 mg of oxygen per hour must, however, be subtracted from the total recovered, At lower temperatures, the rate of reaction is considerably decreased. A precision of approximately 1% is indicated. Future work will involve a study of essentially micro quantities of oxygen with direct measurement of CO by 1,0, or by other methods. Beryllium. The bromination procedure has also been applied to the determination of oxygen in beryllium. Preliminary experiments showed that the rate of reaction between BeO, graphite, and bromine at 950°C is far too slow for analytical application. For example the reaction was only 5% complete after 1.5 hr. If a flux of Na,FeF is added to the mixture, however, the rate of reaction ] ORNL-LR-DWG 5608 TUBE FURNACE TUBING {Smm 1D) xfl) SN = 250-mi FLASK CuD TUBE {20mm [D AND 14in. LONG) Ba(OH)o BUBBLER ™~ DEWAR FLASK Fig. 10.1. Apparatus for the Determination of Oxygen in Uranium Compounds. 161 ANP PROJECT PROGRESS REPORT is increased substantially, The probable reactions are 3BeO + 6NaF + 2FeF3 —_ Fe203 + 3NazBeF4 Fe203 + 3Br2 +3C —> 2FeBr3 + 3CO The reagents are mixed in a platinum boat which is placed in a quartz tube encased with a platinum shield to prevent attack by fluoride on the quartz. The experiments made to date indicate that at temperatures of the order of 700°C the reaction is 80% complete within 2 hr, It is anticipated that higher temperatures will accelerate the reaction rate. The determination of CO is completed as previously described. Future work will involve the application of this procedure to the determi- nation of oxygen in NaF-BeF2 mixtures, Differential Spectrophotometric Determination of Beryllium A. S. Meyer, Jr. D. L. Manning Analytical Chemistry Division The determination of beryllium in NaF-LiF-BeF,- UF ,-UF, samples is required. The method of Vinci'? in which the absorption of the colored lake that is developed when alkaline solutions of beryllium are treated with p-nitrobenzeneazoorcinol is essentially specific for beryllium, but it is not sufficiently precisg to permit application to the determination of beryllium in reactor fuels, When the technique of differential spectrophotometry was applied to this determination a precision com- parable to that of the titrimetric method!3 was found.. In the differential procedure the absorbance of the solution at 510 mu is measured against a blank which contains 10 pg/ml of beryllium. On the basis of replicate determinations of standards and of a limited number of fuel samples, the coefficient of variation of the method is less than 1% for solutions which contain between 10 and 16 pg/ml of beryllium. Citrate ion is added to the solutions to prevent the precipitation of interfering ions, principally U02++. The presence of U02++ introduces a small positive error which is approximately a linear When the function of the uranium concentration. 26 A, Vinci, Anal. Chem. 25, 1581 (1953), 13}, C. White, ORNL Master Analytical Manual, Method No. 9 00711070, ORNL CE-53-1-235, Vol. 1. 162 weight ratio of uranium to beryllium is 10 to 1, the error is only 3%. When the ratio is increased to 25 to 1, uranium is precipitated. Determination of Lithium in NaF-Ber-LiF and NuF-ZrFA-LiF Base Fuels J. C. White G. Goldberg Analytical Chemistry Division The method of White and Goldberg'4 was applied to the determination of lithium in NaF-BeF,-LiF and NaF-ZrF -LiF base fuels, In this procedure the sulfate solution of the alkali metals is con- verted to a chloride solution by passage through an anion-exchange resin in the chloride form. The LiCl is then extracted with 2-ethyl-1-hexanol, and the chloride ion is titrated in the nonaqueous medium. If the acidity of the sulfate solution is adjusted to approximately 1 N, the zirconium, like the wuranium, resin as the anionic sulfate complex and is thereby separated from lithium. Beryllium does not form an anionic complex with sulfate ion and will pass through the column with lithium and the other alkali metals. Moreover, BeCl, is soluble in 2-ethyl-1-hexanol and ac- companies the LiCl through the procedure. The extraction of BeCl, was shown to be quantitative. The described method for lithium is thus used to determine the sum of lithium and beryllium. The beryllium concentration is then determined by the fluoride titration method, '3 while the lithium con- tent is obtained by difference, is retained on an anion-exchange Determination of Potassium in Fluoride Fuels C. R. Williams Analytical Chemistry Division A rapid, direct method for the determination of potassium in fluoride fuels has been developed. This procedure is based on the work of Wittig!S and his co-workers, who found that the potassium salt of tetraphenyl boron is very slightly soluble in acidic solution, in contrast to the sodium and lithium salts, Investigations have shown that of the cations of concern in the ANP fuel program, M, C. White and G. Goldberg, Application of the Volbard Titration to the 2-Ethyl-1-Hexanol Separation Method for the Determination of Lithium, ORNL-1827 {to be published). 13G. Wittig et al., Ann. 563, 11026 (1949). only nickel forms a slightly soluble precipitate with the reagent, Those elements that hydrolyze in the acid solution, such as zirconium, beryllium, and uranium, are held in solution by complexing with fluoride or citrate ions (for hexavalent uranium). The method can also be applied in sulfate solutions and thus offers a distinct advan- tage over the perchlorate method'¢ for deter- mining potassium. The standard deviation of the method is approximately 0.5%. X-RAY SPECTROMETER INVESTIGATIONS OF FLUORIDE FUEL G. D. White, Metallurgy Division T. N. McVay, Consuyltant In addition to routine petrographic examination of fuel samples, several samples were x-rayed on the high-temperature x-ray spectrometer. This work was inaugurated as an additional aid in the determination of phase diagrams, To date, the composition 2NaF-ZrF, has been x-rayed at temperatures from 400 to 600°C. The samples were held by a nickel holder and were heated in a vacuum of from 1 to 2 microns or in a purified helium atmosphere. The atmospheres achieved so far cause slight oxidation which becomes excessive only after heating for over 4 hr. High-temperature x-ray patterns have been obtained on two of the five polymorphs of Na,ZrF . wT. R. Phillips, ORNL Master Analytical Manual, Method No. 9 00716450, ORNL CF-53-1-235, Vol. |, PERIOD ENDING MARCH 10, 1955 ANP SERVICE LABORATORY J. G, White W. F. Yaughan C. R. Williams Analytical Chemistry Division The determination of beryllium in fluoride fuel mixtures was resumed during this quarter. The mixture of fluoride salts is dissolved in H,50, and the solution is heated until copious fumes of 30, are evident. This process of fuming is re- peated in order to remove the fluotide completely from the solution. Uranium as UQ,50, is re- moved from the solution by sorption on an anion- exchange resin. Beryllium is then determined by the modified volumetric method!3® of McClure and Banks., The major portion of the work continues to be analyses of fluoride salts, mixtures of fluoride salts, and alkali metal hydroxides. The total of 1110 samples analyzed involved 7905 determinations. The backlog at the end of the quarter consisted of 335 samples. A breakdown of the work load is given in Table 10.1. TABLE 10.1. SUMMARY OF SERVICE ANALYSES REPORTED Number Number of of Samples Determinations Reactor Chemistry 793 5642 Experimental Engineering 304 2172 Miscellaneous 13 91 Total 1110 7905 163 ANP PROJECT PROGRESS REPORT 11. RECOVERY AND REPROCESSING OF REACTOR FUEL D. E. Ferguson M. R. Bennett G. I. Cathers J. T. Long R. P. Milford S. H., Stainker Chemical Technology Division PILOT PLANT DESIGN Design of the pilot plant to recover, in seven batches, the 65 kg of U?3% in the 2500 Ib of ARE fuel by a fused salt—fluoride volatility process is in progress. Completion of construction by December 31, 1955, is planned, The January 1955 flow sheet calls for a ninefold excess of fluotine to be passed through the molten fuel at 650°C. The UF, and volatile fission-product fluorides formed will pass from the fluorinator into a bed of 20- to 40-mesh NaF at 650°C, where the volatile fission-product fluorides will be absorbed, The UF, will collect in a series of three cold traps ot +4, —40, and ~60°C, respec- tively, |f sufficient decontamination has occurred, the traps will be isolated and heated electrically in order to liquefy the UF,, which can then be drawn off into receivers, [f further decontami- nation is required, the UF, will be volatilized out of the cold traps into another system con- sisting of an NaF absortber and three cold traps identical with those mentioned. An aqueous KOH scrubbing system will be required for disposing of excess fluorine. A schematic flow sheet for the plant is shown in Fig. 11.1. The preliminary cost estimate for the plant, including a 20% contingency factor, is $285,000, PROCESS DEVELOPMENT Laboratory-scale studies were made on the efficiency of fluorine usage and the corrosion of the reaction vessel in the fluorination step. During the first few minutes of the reaction, fluorine was absorbed completely but no UF ; was evolved. Fluorine continued to be absorbed completely until 80% of the UF6 had been volatilized off; after this, fluorine was detectable in the exit gas. The data indicated that the UF, was essentially all evolved by the time a fivefold excess of fluorine had been used. Therefore the ninefold excess shown in the flow sheet is believed to provide a sufficiently large safety factor. The 164 average cortosion rate of nickel during the whole fluorination step was about 0.1 mph, but the rate appeared to be considerably higher than the average during the time that UF6 was present. Two methods were used in studying the reaction kinetics of the direct fluorination of ARE-type fuel. The first consisted in passing the exit gas from the fluorination vessel through two absorbent traps filled, respectively, with NaF and NaCl. These wete weighed before and after each ex- periment., The NaF trap, held at about 100°C, was capable of trapping an equal weight of UF by absorption. The NaCl frap lost weight as the NaCl was converted to NaF by the fluorine. Weighing these traps at 5-min intervals provided a means of following the progress of the fluo- rination, The second method, Fig. 11.2, con- sisted in using three flowmeters! to measure the fluorine input flow rate, the combined UF, and fluorine output flow rate, and the fluorine output flow rate after all UF, in the gas stream had been absorbed. Kinetic data obtained by this method were more precise than those cbtained by the first method, since a Brown 5-mv recorder was used to obtain flow readings at 20-sec intervals, By means of the NaF-NaCl absorptionstrap method, calibration curves for UF,, N,, and F, were prepared which show the flow in cubic minute corresponding to the voltage readings obtained in the flowmeter method. The procedure usually used with both techniques centimeters per was to fluorinate, in a l-in,-dia nickel reactor, 70 g of ARE-type fuel (NaF-ZrF ,-UF,, 52-44-4 mole %) at 650°C with a fluorine input rate of about 60 ml/min, The variation in flow rate with timein one run in which thermal flowmeters were used is shown in 'K. 0. Johnson, W. F. Peed, and G. H. Clewett, A Thermal Type Flow Meter jor Low Flow Rates of An- bydrous Hydrofluoric Acid, C-5.355.8 (June 24, 1946). PERIOD ENDING MARCH 10, 1955 5001R QOOo00000DO000CO0QO0O0QQ 0 B B = P2 ot a i_L o = T © 2 3 = 3 o 5 oy 3 > g oy 2 a i <1 g T 9 Lt 5 3 T & o] = - |'1TLan B O |} 2 > ity o S v fo 0 2 23 o 8y : T g8 ] © 2 = TR Iz w oo . 223 [ % W W W%M_ < mv_ ANH %N.u n\un_umm - o S @ a5 N L 4w > oY o w oo w <~ S = e CF b, £ o of - STy @ - T T ey ye? = B 2 2 |19 u - o £ @ £ v v L] — — — - 2 i, T 5 . o [ad] = ML 03.‘,60% ol O < 0= > =zfy e " o @mm f///////////////////m w EHE | <1 cgs [ A T 5% W 1 /] . 1 2 pEaaa W 7 TR z =9 =C —_ 5 <4 < o3 £z = \/ - £\ 165 ANP PROJECT PROGRESS REPORT v ORNL-LR-DWG 5609 F, F, F, + UFg Rrufg [ ] R F 5 ——=| FLOWMETER »| REACTOR = FLOWMETER w{ NGF ABSORBER [—pm| FLOWMETER |——m OUTUET Fig. 11.2. Schematic Diagram of Flowmeter Method of Measuring Gas Flow Rates in Fused Salt- Fluoride Volatility Process. Fig. 11.3a. |n this run three curves were obtained from the three flowmeters. The first curve shows the fluorine input rate, which remained practically constant during the entire run, The second curve gives the flow of gas from the reactor to the NaF trap; the UF, calibration curve was used to in- terpret this flow. However, the first part of this curve is somewhat in error, since the system originally contained nitrogen, and nitrogen or a mixture of nitrogen and UF_ initially passed through the second flowmeter. The third curve gives the output flow from the NaF trap. The gas passing through this meter was initially nitrogen; it was fluorine after 31 min, at which time nitrogen displacement was complete and a chemical test for fluorine was first obtained. The hump in the nitrogen part of the third flowmeter curve results from displacement of nitrogen gas by UF, from the apparatus between the reactor and flowmeter. The initial flow of about 5 ml/min was probably the result of some inert impurity in the fluorine supply. The flowmeter readings are interpreted in Fig. 11.35 to show the UF, and fluorine evolution as a function of time. During the first 14 min, fluorine was completely absorbed. Material balance calcu- lations indicated that only about one half this amount of fluorine could be attributed to the reaction UF, + F,—> UF, . The remainder is assumed to have been utilized in corrosion. At the end of 14 min, UF6 evolution began and rapidly increased to 35 to 40 ml/min after about 20 min had elapsed. The UF ; evolution remained essentially constant over the next 10min and began to drop off simultaneously with the breakthrough of flucrine, which occurred after 31 min, The fluorine absorption was essentially com- plete until 85% of the UF, had been evolved. Based on the amount of fluctine absorbed during 166 this period, an average corresion rate of 4 mph was calculated, Qualitative work reported previously? on the solubility of UF, in a NaF-ZrF, mixture had suggested the existence of a stable NaF-UF6 complex that is soluble in molten NaF-ZrF,. The induction period and the subsequent plateau in UF production shown in Fig. 11.3a2 and b are also indicative of this, The reaction mechanism is believed to be: UF, (soln in NaF-ZrF,) + F, (g)—> chlI:-UF6 (soln in NaF-ZrF ) —> UF6(g) . The induction period is the result of both fluorine consumption in corrosion and of a buildup in the concentration of the NaF-UF . complex until the saturation solubility is reached, Then UF, begins to evolve as fast as fluorine is supplied in excess of the rate of utilization in corrosion, The pos- sibility that oxides dissolved in the fused salt could account for the delay in UF, generation wos discounted by the results of a run in which the ARE fuel was sparged with HF for 45 min and then with nitrogen for 10 min. Any oxides present would have been eliminated by this treat- ment, but the induction period was the same as that in the run plotted in Fig. 11,32 and 5, The sodium-to-zirconium atom ratioc in ARE- type fuel (4 mole % UF,) is about 5:4. Addition of more ZrF, to change this ratic to 4:5 had practically no effect on the curves shown in Fig. 11.32 and 4. This suggests that the NaF-ZrF, complex is relatively weak in comparison to the product of interaction between NaF and UF,, The main effect of a change in the sodium-to-zirconium ratio is on the melting point of the salt mixture, The corrosion work to date has consisted mainly of gravimetric tests on metal coupons and deter- minations of nickel in the fused salt residues 2p. E. Ferguson et al,, ANP Quar. Prog. Rep. Dec. 10, 1954, ORNL-1816, p 134. 70 &80 W b wm o O o FLOW RATE {ml/min, STP) n O 70 60 50 40 30 FLOW RATE {ml/min, STP) 20 PERIOD ENDING MARCH 10, 1955 ORNL-LR—DWG 5610 s~ - FLOWMETER (F2 FLOW TO REACTOR)i A A—A_A VESSEL . NICKEL ——— TEMPERATURE: 650°C | | FUEL: 70 g of NaF-ZrF,—UF, (52-44-4 mole %) FLOWMETER 2 (UF6 FLOW FROM REACTOR) FLOWMETER 3 —__ (N2 DISPLACEMENT BY UFg AND F, CUTPUT ) . ® ./ o’o o0 ® / / ° 50 TIME {min) / 20 30 TIME {min) Fig. 11.3. (o) Gos Flow Characteristics in ARE-Type Fuel Fluorination Process. (b) Kinetics of Fuel Fluorination. Dotted lines indicate uncertainty resulting from mixing of gas with nitrogen. 167 ANP PROJECT PROGRESS REPORT from fluorinatiem studies (Table 1L1). In the gravimetric tests the corrosion rate was generally less than 0.1 mph for nickel and somewhat higher for Inconel and Monel. The cortosion rate was higher in the runs in which UF, was present than in the others. In these five runs there was a UF ,-te-UF, con- version period of less than 1 hr, followed by a period of several hours during which fluorine was passed through the salt. That the average cor- rosion rate was, in general, greater in the short runs is believed to be the result of the very high corrosion during the time that UF, was present. The lower average rate obtained in long runs is due to less corrosion occurring after the vranium has all been evolved. The corrosion rates determined from salt anal- yses were obtained in five fluorination runs made with simulated ARE fuel. The test periods were probably greater than specified, since the time spent in starting or stopping each experiment is not accurately known., The internal surface area of the nickel reactor used in each run was about 66 cm? The greatest corrosion rate, 0.66 mph, was again obtained in the run of shorte st duration, TABLE 11.1. CORROSION IN ARE-TYPE FUEL FUSED SALT-FLUORIDE VOLATILITY PROCESS AT 650°C Gravimetric studies: corrosion coupons were cut from round rod, weighed, immersed for time shown in fused salt through which fluorine was being passed, and reweighed Salt analyses studies: fused salt residues from fluorination studies carried out in nickel reactors were analyzed for nickel Sale* Test Period Corrosion Rate {mph) (hr) Ni Inconel Mone! Gravimetric Method Nch--ZrF4 (56-44 mole %) 14** 0.031 0.021 0.016 14** 0.36 0.40 0.76 Na F-ZrF4-UF4 (52-44-4 mole %) grxx 0.056 0.063 0.12 5,5%** 0.081 0.017 0.11 6, 5xx* 0.066 0.1 0.15 gxx* 0.077 0.064 0.48 | Rkl 0.37 0.56 0.44 chF-ZrF’4 {56~44 mole %) Jrxx 0.058 0.10 0.10 Salt Analyses NuF-ZrF4-UF4 (52-44-4 mole %) 1.5 0.66 3 0.33 44 0.06 5 0.25 S 0.21 *Two runs were made with the first batch of salt, run. In all other experiments a fresh batch of salt was used in each * % . . ; Same corrosion samples used in the two successive runs, %k . . ) . Same corrosion samples used in the six successive runs. 168 Some information on corrosion can also be ob- tained from the kinetic data in Fig. 11.3a and 6. The discrepancy between the maximum UF, flow rate of 40 ml/min and the flucrine flow rate of - 62 ml/min represents a nickel cortosion rate of about 1.8 mph. This is perhaps effective only during the period of existence of the NaF-UF, complex. However, an even more rapid average corrosion rate of 4 mph during the period UF, was present was indicated by the fluorine material balance. The low corrosion rates obtained in the gravimetric test and fused salt analyses are possibly obtained only over the process as a PERIOD ENDING MARCH 10, 1955 whole, that is, when the UF ,-to-UF, conversion period is accompanied by several hours of fluorine sparging fo remove the last traces of vuranium. In all laboratory fluorination tuns, 50% or more of the fluorine was consumed in cotrosion of the nickel reactor. This need not be true in scaling- up the process to a pilot plant level, It is esti- mated that the consumption of fluorine in cor- rosion for a 10-kg uranium batch will be less than one-tenth that in a batch of 6 g. This is a consequence of the greater volume-to-surface ratio in the 10-kg case. 169 Part |l SHIELDING RESEARCH 12. SHIELDING ANALYSIS E. P. Blizard F. H. Murray C. D. Zerby Physics Division S. Auslender Pratt & Whitney Aircraft H. E. Stern Consolidated Vultee Aircraft Corporation Air-scattering measurements made at the Tower Shielding Facility indicate that for optimized divided shields the multiple scattering of radiation in air is quite important — more so than in many shield designs considered heretofore. Consequently a large fraction of the analysis effort has been pointed toward more complete calculations of air scattering. Two approaches are used. In one, certain concessions are made in the physical hy- potheses, such as assuming constant neutron velocity, to enable analytical treatment of the problem. In the other, a stochastic {Monte Carlo) method is adopted, with as nearly exact data being used as are available., Neither method has yet been completely developed or applied. An offshoot of a Monte Carlo calculation of slant penetration of gamma rays in crew-shield side walls has been the development of a similar calculation of gamma heating in a multilayered region, This method will be applied to the presently conceived reactor designs to find the gamma heating to be expected in regions near the core. ANISOTROPIC SCATTERING OF NEUTRONS IN A UNIFORM MEDIUM WITH BEAM SOURCES F. H. Murray Formulas useful in the analysis of problems of multiple scattering of neutrons in an unbounded medium have been derived. These formulas are being applied in some problems of scattering in stratified mediums. The particle velocity is as- sumed to be constant in the analysis, and the scattering law is assumed to depend only on the angle between the initial and final directions of the scattered particle. _ This analysis is based directly on the Fourier transform, and the spatial distribution of sources and the intensity os a function of angle from any point may be arbitrary. The formulas and their derivations are presented in a separate report.1 ENERGY ABSORPTION RESULTING FROM INCIDENT GAMMA RADIATION AS A FUNCTION OF THICKNESS OF MATERIALS WITH SLAB GEOMETRY C. D. Zerby S. Auslender Initial calculations of the heat generation in berytllium slabs as a result of the transport of slant incident gamma radiation were described in a previous report,2 The method used for these initial calculations is being extended to include photons with energy up to approximately 10 Mev, The coded caiculation for the Oracle, from which the initial results were obtained, is also being revised to make it more versatile. It will be pos- sible in the future to study the heat generation resulting from the transport of gamma radiation through homogeneous slabs of materials containing any number of elements. The code for the calcu- lation is being designed for simplicity of operation so that cases of interest can be calculated with a minimum of effort and time. In particular it is anticipated that the method will be applicable to caleculations of heat deposition in the laminations around the reflector-moderated reactor. ENERGY AND ANGULAR DISTRIBUTION OF AIR-SCATTERED NEUTRONS FROM A MONOENERGETIC, MONODIRECTIONAL POINT SOURCE C. D. Zerby A calculation of the energy and angular distribu- tion of air-scattered neutrons from a monoenergetic, monodirectional point source is being made to pro- vide data from which the energy and direction of the neutron flux at an aircraft crew compartment 1F. H. Murray, Anisotropic Scattering of Neutrons in a Uniform Medium with Beam Sources, ORNL CF-54-11-83 {to be published). 2c. b. Zerby, ANP Quar, Prog. Rep. Dec. 10, 1954, ORNL-1816, p 144, 173 ANP PROJECT PROGRESS REPORT shield resuvlting from air-scattered neutrons from an arbitrary source can be determined at any altitude. The Monte Carlo method is being used for this calculation, and it is presently being coded for the Oracle, The cross sections being used were taken from a previous report® and include the complete resonance 3Neutron Cross Sections, AECU-2040 {(May 15, 1952). 174 structure, as reported. The scattering will initially be taken as isotropic in the center-of-mass system; however, later, it is intended that the anisotropy, as determined from experiments or calculations, will be included. For comparison with approximate solutions to this problem, the energy and angular distribution of single, double, triple, and all other multiply scattered neutrons will be separately. recorded PERIOD ENDING MARCH 10, 1955 13. LID TANK SHIELDING FACILITY G. T. Chapman J. M. Miller D. K. Trubey' Physics Division J. B. Dee W, J, McCool H. C. Woodsum J. Smolen Pratt & Whitney Aircraft The GE-ANP helical air duct experimentation has been completed, and an analysis of the data is presented, A new attempt to correlate the removal cross section data with other properties of the atom has resulted in a plot of the ratio of the macro- scopic removal cross section to the density of the material against the atomic weight., These data are compared with the total cross section at 8 Mev. Final preparations for the second series of tests on the reflector-moderated reactor and shield mockup have been completed. GE-ANP HELICAL AIR DUCT EXPERIMENTATION J. M. Miller All experimental work has been completed at the Lid Tank Shielding Facility (LTSF) on the GE- ANP helical air ducts and a final report on the work is being written, As described previously,? the ducts were fabricated by shaping 3-in.-ID flexible tubing around a 9-in. core. After removal of the core, the ducts were stiffened with Fiberglas wrapping. The projected length of each duct along the z axis was 46.5 in., and the duct arrays were arranged so that there was 5 in. between the duct center lines, ' Radiation measurements were made beyond a single duct, a three-duct array, and a 35-duct array in plain water. Measurements were also mode beyond the 35-duct array in a medium of Raschig rings (hollow steel cylinders, ]/4 in. in diamefer and ]/2 in. long) and borated water, The Raschig rings were packed so that the medium around the ducts was 33% steel and 67% borated water. Various thicknesses of lead were also placed alongside the 35-duct array in the Raschig ring—borated water medium to mock up a side shield for the ducts, Some of the results of the experimental work are given in Table 13.1. 1On leave to attend ORSORT. 2J. M. Miller, ANP Quar. Prog. Rep., ORNL-1771, p 166; ORNL-1816, p 153. REMOVAL CROSS SECTIONS G. T. Chapman Previous attempts to correlate the effective neutron-removal cross section data obtained at the LTSF with more rigorously defined properties of the atoms, such as atomic weight and total cross section, have been limited because of the scarcity of data. However, a plot of the data has been made that may prove useful in shield calculations. The plot is presented in Fig. 13.1. The ratic of the macroscopic removal cross section to the density of the material (/p, cm?/g), a number proportional to the removal cross section per nucleon, is plotted against the atomic weight. These data are com- pared with the total cross section at 8 Mev, as 2-01-057-0-93 ———— T T ] TOTAL CROSS-SECTION VALUE AT 8 Mev _ {TAKEN FROM REF, 3) Lo REMOVAL CROSS-SECTION VALUE DERIVED — FROM MEASUREMENTS BEHIND COMPOUND 1T} 2| OF THE ELEMENT SRR o FEMOVAL CROSS-SECTION VALUE FROM PE MEASUREMENTS BEHIND ELEMENT " £ = REMOVAL CROSS SECTION OR TOTAL ~ 051 CROSS SECTION AS DESIGNATED S p = DENSITY OF MATERIAL g ‘ I | =~ 02 , P ot o : | i P N\ @ 1 | ‘ \ o ;] P ————3% i & vos | T e | | .i ' 00852 | 4 54 o | N Vi o :‘?j? : . + A 3 T - !O T 002 |—— e S~k Co . I 'Q-ng\‘“ S ool | 1 L Mo T ; | S| I 1 i - 1 1 0.005 = - 12 5 40 20 50 400 200 500 4000 A, ATOMIC WEIGHT Fig. 13.1. Neutron Shielding Ability per Unit Weight of a Material as o Function of Atomic Weight. 175 ANP PROJECT PROGRESS REPORT TABLE 13.1, INTENSITY INCREASES RESULTING FROM THE REDUCED DENSITY EFFECT AND THE PRESENCE OF G.E HELICAL AIR DUCTS Calculated Factor Total Factor Increase in lncrease in Duct Non-Shield 'UC Shield* Quantity Measured en-snte Intensity Resulting Intensity Resulting Configuration Component* from Reduced from Presence Density Effect of Ducts Single duct Plain water Thermal-neutron flux ~2 Gamma-ray dose rate ~1 Fast-neutron dose rate o Qrk Three ducts Plain water Thermal-neutron flux ~6 Gamma-ray dose rate ~ 1.4 Fast-neutron dose rate ~4 35 duets Plain water Thermal-neutron flux 43.5% air 210 ~ 480 Gamma-ray dose rate 49% ducts 2, 1%%* ~2.4 Fast-neutron dose rate ~ 200 ~ 400 Raschig rings in Thermal-neutron flux 43.5% air 770 4000 borated water Gamma-ray dose rate 49% ducts 175 ~160 *Does not include water thickness between end of duct array and center of detection. **Poor statistics. ***Corrected for iron in ducts. obtained by N. G. Nereson et al.® at the Los Alamos Scientific Laboratory, All removal cross section measurements at the LTSF have been compiled and will be published in a separate repor’r.4 REFLECTOR-MODERATED REACTOR AND SHIELD MOCKUP TESTS J. B. Dee The second series of experiments® with mockups of the circulating-fuel reflector-moderated reactor (RMR) and shield has been started ot the LTSF, The final preparations involved some changes from the plans previously published.® The 1I/B-in.-i'hiczl-c Inconel window originally built into the source side N G. Nereson et al, Survey of Hverage Neutron Total Cross Sections from 3 to 13 Mev, LA-1655 (July 13, 1954). ‘. T. Chapman and C. L.. Storrs, Effective Neutron Removal Cross Sections for Shielding, ORNL-1843 (to be published). sFor first series see C. L. Storrs et al., ANP Quar Prog. Rep. Sept, 10, 1953, ORNL-1609, p 128. 6). B. Dee et al., ANP Quar. Prog. Rep., ORNL-1771, p 164; ORNL-1816, p 155, 176 of the large tank for holding all the dry components of the configurations has been replaced with a % _in.-thick aluminum window. With the removal of the lIncone!l and, consequently, the high-energy nickel capture gamma ray, a more accurate siud),r can be made of the smaller effects resulting from in- elastic scattering and capture gammas in the other components of the shield mockup. As originally planned, Inconel will be used to simulate the RMR core shell, but it will be in the form of a removable slab placed immediately behind the aluminum window of the dry tank., An advontage of this ar- rangement is that it will be possible to place the [nconel inside the fuel belt during the dynamic tests and thus provide a more realistic mockup for this series of tests, The solid Al-UO, plates, which will simulate the fuel in the dynamic fission-source tests, were mounted in @ continuous series on a sprocket-driven chain-link belt and tested under operating condi- tions in a special rig outside the LTSF (Fig. 13.2). The belt, which will move continuously from the neutron window to the heat exchanger region and back to the neutron window, was successfully oper- PERIOD ENDING MARCH 10, 1955 UNCLASSIFIED PHOTO 13853 Fig. 13.2. Apparatus for Testing the Chain-Driven Fuel Elements for the RMR-Shield Mockup. ated at speeds up to 1050 fpm, which corresponds to a transit time of 0.73 sec in an actual mockup; the maximum design speed for the mockup is 900 fpm or a transit time of (.83 sec. As mentioned previously,® this second series of RMR experiments will consist of two sets of tests: the static-source tests and the dynamic- source tests. In the static tests, the first measure- ments of which are now being made, the primary and secondary sources of radiation that reach the outside of the shield will be determined. This will be accomplished by analysis of differences in dose curves obtained by varying the thicknesses of the mockup regions, as follows: (1) an Inconel core shell % in. thick, (2) a beryllium reflector 8 to 16 in. thick, (3) a boron curtain consisting of 2!{‘ in. of boral, (4) a heat exchanger consisting of one to four ]1/2-in. slabs of NaF contained in nickel, (5) a second boron curtain, (6) a pressure shell consisting of 0 to 2 in. of nickel, and (7) a shield * termined, made up of lead, 0 to 10]/2 in. thick, and water, usually borated. In some tests, bismuth and copper will be substituted for part of the beryllium re- flector for determining the effect of placing a gamma shield closer to the reactor. These tests will not include radiations resulting from fuel circulation, but the sodium activation within the heat exchanger region will be studied. For the dynamic tests a typical mockup will be used, primarily, and the LTSF source place will be replaced with the continuous series of Al-UO, plates mounted on a movable belt. Several transit times will be used — the minimum will be 0.66 sec — and the sodium activation in the heat ex- changer caused by delayed neutrons will be de- The fuel belt will also give a source of fresh fission products. The attenuation of the gamma rays resulting from these fission products by various thicknesses of lead will be studied. 177 ANP PROJECT PROGRESS REPORT 14. TOWER SHIELDING FACILITY C. E. Clifford T. V. Blosser J. L. Hull L. B. Holland F. N, Watson Physics Division D. L. Gilliland, General Electric Company M. F. Valerino, NACA, Cleveland J. Yan Hoomissen, Boeing Airplane Company The first experiment with the mockup of the GE-ANP R-1 reactor shield has been completed, and the first series of differential experiments has been started. A specially designed gamma-ray dosimeter with which it is possible to simulate the addition of small thicknesses of lead to an aircraft crew compartment has been calibrated. Some results of the exposure of primates to high fast-neutron dose rates in a program initiated by the U.S. Air Force are reported. TSF EXPERIMENT WITH THE MOCKUP OF THE GE-ANP R-1 SHIELD DESIGN T.V. Blosser M. F. Yalerino D. L. Gilliland J. Van Hoomissen F. N. Watson The first experiment with the mockup of the GE-ANP R-1 reactor shield has been completed at the TSF with measurements of the gamma-ray dose rates along the x, y, and z axes of the detector tank. The experimental arrangement for these measurements was the same as that for the earlier thermal-neutron flux measurements.! The reactor-detector altitude (b) was 195 ft and the G-E shield was a horizontal distance (d) of 64 ft from the detector tank. Five 1-in.-thick lead slabs were installed 1 ft from the rear face (reactor side) of the detector tank to simulate the shielding in the crew compartment. In order to gain a knowledge of the dose rates from gamma rays which will be incident on the various faces of the crew shield mockup and of the gamma-ray attenuation characteristics of the hydrogenous shielding of the mockup, the gamma- ray dose rates were measured with both plain and borated water in the detector tank. From the borated water traverses the reduction of the ]C. E. Clifford es al., ANP Quar, Prog. Rep. Dec. 10, 1954, ORNL-1816, p 158. 178 gamma-ray dose rate caused by the suppression of the gamma rays resulting from the capture of thermal neutrons in the water could be determined. The y traverse (Fig. 14.1) measured in plain water indicates an average relaxation length of 19.7 e¢m to the rear of the lead. The shape of the curve in front of the lead indicates that the gamma dose is due to the gamma rays which enter the tank from the sides rather than those which penetrate the [ead. Therefore, an effective relax- ation length for the lead attenuation cannot be obtained from these data. The reduction of the gamma dose in the center of the tank which resulted from the boration (0.4 wt %) of the water indicates the presence of substantial quantities of secondary gammas from captures in hydrogen. The effective relaxation length at the front of the tank along the y axis (reactor—detector tank axis) was found to be 9.6 c¢m and was obtained by subtracting from the curve the gamma dose pene- trating the sides of the tank. The relaxation length obtained at the sides of the tank (Figs. 14.2 and 14.3) in a similar manner was 10.3 cm and was apparently a single exponential for pene- tration of at least 25 cm of water. An error in the calibration of the gamma-ray detector caused by the neglect of the decay of the cobalt source introduces a correction factor of 0.94 for all gamma-ray dose rates measured in the detector tank during this experiment. THE DIFFERENTIAL EXPERIMENTS AT THE TSF: PHASE | T.V. Blosser M. F. Valerino L. B, Holland J. Van Hoomissen J. L. Hull F. N, Watson The experimental program at the TSF is now being arranged so that the mockup experiments are interspersed with differential experiments (those in which the 12-ft-dia reactor tank and the 2-01-056-3—~18+24-82 h =195 ft d =64 ft ———— i - CENTER OF SCINTILLATION CRYSTAL AT: x£=0cm _T ' ' ‘ y=VARIABLE | | Z=0cm PLAIN WATER GAMMA —RAY DOSE RATE (mr/hr/watt) MULTIPLY ORDINATE BY 0.94 . BORATED WATER 0 12 24 36 48 60 72 84 96 108 120 132 144 ¥, DISTANCE FROM REAR FACE OF DETECTOR TANK TO DETECTOR CENTER {(cm) Fig. 14.1. Gamma-Ray Dose Rate Along the y Axis of the Detector Tank. 641 10 156 104 -7 GAMMA —RAY DOSE RATE (mr/hr/watt) S561 ‘0L HOYVW ONIONZ QOId3d 081 GAMMA—RAY DOSE RATE (mr/hr/watt) < 0 O )— m [T} - 1 — o ~ —J D = GAMMA DOSE RATE (mr/hr/watt) 1077 -78 -66 -54 —42 -30 -8 -6 0 6 18 30 42 54 z, VERTICAL DISTANCE FROM & AXIS TO DETECTOR CENTER (cm) Fig. 14.3. Gamma-Ray Dose Rate Along the z Axis of the Detector Tank. 18t 2-01—-056-3-21-88 66 78 $S61 ‘01 HOYVYW 9NIAON3I @old3d ANP PROJECT PROGRESS REPORT detector tank are used). A series? of differential experiments has now been started with emphasis on obtaining the fast-neutron dose rate distribution within the detector tank as a function of 6, the angle between the axis of symmetry of the beam from the reactor tank and the source-detector axis. The thickness (p) of the water layer shielding the reactor, as measured from the edge of the reactor tank, is being held constant at 45 em. This thickness was chosen because it is in the region of interest of the side shielding on an gircraft reactor, and it also allows sufficient intensity in the detector tank for accurate measure- ments. Further, the prediction of the dose within the detector tank for a variation of p is felt to be less uncertain than for the variation of 4. The quantities to be determined are the magni- tude of the doses impinging on the front, side and rear of the detector tank and the rate at which these doses are attenuated. It is hoped that the rate can be expressed in terms of the relaxation length of a simple exponential function. The results of these measurements will be compared with single-scatter calculations which are now being coded for solution on the Oracle. The calculations are being devised so that the doses arriving at the various faces of the tank can be calculated separately, and, for the 0- and 90-deg case, the angular distribution of the dose arriving at the faces will also be calculated. The calcu- lations will be sufficiently general that any angular distribution at the source may be used. The measurements which have been made to date have been planned as a quick survey to indicate the most interesting regions for the more detailed measurements to follow. CALIBRATION OF THE REVALET, A REMOTELY VARIABLE LEAD-TRANSMISSION GAMMA-RAY DOSIMETER D. L. Gilliland The optimization of gamma-ray shielding around the crew compartment is one of the major problems in the design of nuclear aircraft shields. To permit optimization the attenuation of lead for the gammg-ray dose penetrating the crew com- partment must be determined as a function of the lead disposition within the crew compartment. A 250me preliminary differential experiments were re- ported previously; C. E. Clifford e al., ANP Quar. Prog. Rep. Sept. 10, 1954, ORNL-1771, p 175. 182 full-scale experiment to obtain this information by varying the lead thickness inside an actual crew shield mockup would be both ‘difficult and time consuming; therefore, a method has been developed in which it is hoped the crew com- partment gamma shield can be simulated with lead thicknesses that can be varied with ease. This has been accomplished by enclosing an gnthracene scintillation counter in a thick lead shield which has an aperture in one side that can be covered with lead disks of varying thicknesses from 0 to 0.7 in. Some of the details of the instrument, known as the Revalet, are shown in Fig. 14.4. In order to ensure the validity of the measure- ments to be made at the TSF with the Revalet, an experiment with a known geometry and a source of known energy (Co®%) was devised in which the angle of incidence of photons on the lead absorber disk could be varied. The measured lead attenu- ation for the various angles of incidence (0 to 60 deg) was in good agreement with Monte Carlo calculations of slant penetration.® A comparison of the experimental measurements with the Monte Carlo calculation is shown in Fig. 14.5. The sensitivity of the counter as a function of angle of incidence is shown in Fig. 14.6 for the open- aperture case. The details of the instrument are given in a separate repori’.4 THE PROJECT ORANGE PRIMATE EXPOSURE AT THE TSF T.V. Blosser L. B. Holland D. L. Gilliland J. L. Hull F. N. Watson A preliminary experiment® in which primates were exposed to mgssive doses of radiation has been completed at the TSF as part of a program initiated by the U.S, Air Force. Prior to these measurements the only experimental data available had been obtained with massive gamma-ray doses from barium-lanthanum sources. A more realistic approximation of the doses that would be received from an atomic warhead would, of course, include neutron as well as gamma-ray doses. It would 3c. b Zerby, Preliminary Report on the Penetration of Composite Slabs by Slant Incident Gamma Radiation, ORNL CF-54-9-120 (Sept. 21, 1954). AD. L. Gilliland, Calibration of the Revalet, a Re- motely Variable Lead-Transmission Gamma-Ray Dosime- ter, ORNL CF-55-2-111 (to be published). SA future experiment involving several hundred animals is planned. Lt 7T s FRONT VIEW == LEAD DISK IN ABSORPTION WHEEL {0.1-in. GRADATION IN THICKNESS) ¥ Y, X Y4-in. ANTHRACENE CRYSTAL PERIOD ENDING MARCH 10, 1955 2-01{-056-8-D1C4 GENEVA-LOC ACTUATOR TO DRIVE LEAD-ABSORPTION WHEEL\\ \ ) X oo W ‘ L | by —COUNTERBALANCE N ~POTENTIOMETER . ! PREAMPLIFIER Ll | ~ PN Y, % Y, ~in. ANTHRACENE N - 12 4 : P CRYSTAL & LEAD SHIELD | :..,‘ _ N“ LRI - . ; = A DUMONT 6292 PHOTOMULTIPLIER TUBE LUCITE LIGHT PIPE SIDE VIEW Fig. 14.4. The Revalet, a Remotely Variable Lead-Transmission Gamma-Ray Dosimeter. 2~01-056-8-3~105 | ! " ¢ CALCULATED VALUE, WITH STANDARD CEVIATION INDICATED FRACTION OF INCIDENT ENERGY PENETRATING SLAB EXPERIMENTAL ANGLE OF INCIDENCE | 0 Odeg 30 deg a A 60 deg A 75 deg 0 eXi 0.2 0.3 0.4 0.5 0.6 0.7 LEAD THICKNESS (in.} Fig. 14.5. Lead-Absorption Measurements with the Revalet, also include a more diverse spectrum of the gamma-ray energies. Since high radiation levels exist outside the shielding of the TSF reactor, it was chosen as a convenient source for further experiments. This preliminary experiment, desig- nated Project ORANGE by the USAF, involved the exposure of 25 rhesus monkeys. Since little was known of the effect of fast- neutron doses, emphasis was placed on maxi- mizing the fast-neutron dose rate and minimizing the gamma-ray dose contamination. The maximum neutron dose achieved was 30,000 rep for an interval of approximately 1Y% min, the gamma-ray contamination being approximately 1 r/7.5 rep. The experimental arrangement for the exposures is shown in Figs. 14.7 and 14.8. Two monkeys were exposed simultaneously in the two cy- lindrical Lucite cages shown in Fig. 14.7. The cages were placed on a motorized turntable which was rotated at 1 rpm, and they were separated from the reactor tank by a 3-in.-thick lead shield and @ ¥ -in.-thick boron-impregnated plastic sheet. The horizontal midplane of the reactor coincided with the bottom edge of the white stripe around the reactor tank, shown in Fig. 14.7. The front face of the reactor was 5.5 cm from the tank wall. (Before the exposures were actually made the pine two-by-four running across the top half of the 183 ANP PROJECT PROGRESS REPORT g ST S 2-01-056~8-5-106 ANGLE OF INOC!DENCE (deg) 30 OPEN APERTURE CASE B 90 3 2 4 60 90 INTENSITY {(arbitrary units) Fig. 14.6. Solid Angle of Detection of the Revalet. cages, Fig. 14.7, was removed. The three boron carbide stabs in front of, behind, and under the turntable were also removed.) Measurements of the fast-neutron doses were made both with chemical dosimeters and with a Hurst-type fast-neutron dosimeter. The chemical dosimeters were attached to the outside of the cages and on the monkeys. In some instances cylindrical polyethylene containers filled with a solution of sugar, water, and urea to simulate the body composition of a monkey were used as phantoms in the cages and chemical dosimeters 184 were attached to the phantoms. The Hurst do- simeter measurements were made inside the cages with and without the phantoms., The gamma-ray dose measurements were also made with chemical dosimeters placed on the cages, monkeys, and phanfoms. Measurements with the 900-cm® ion chamber were made inside the cages without the phantoms. The total radiation doses to which the animals were exposed are given in Table 14.1. The results of these preliminary exposure studies will be reported by the USAF, G8l Fig. ST 1-01-056~-20 14.7. TSF Experimental Arrangement for Monkey lrradiation Showing Lucite Cages in Position Near Reactor Tank. §S61 ‘01 HOMVW ONIGNI dol¥3d ANP PROJECT PROGRESS REPORT 2-01-056-4~-D107 AR R INSIDE SURFACE COVERED WITH BLOTTING PAPER \ PLASTIC CAGE ’ TSR | CAGE HOLDER (- 924 in. CAGE ROTATJO_N// MECHANISM ; WOODEN STAND —— | Ya-in- THICK PLYWOOD Yo-in-THICK LEAD SUPPORT CHANNEL 1-in.~THICK LEAD 1" SUPPCRT FRAME REACTOR TANK 7 1 3 L‘* =~—rooL EDcE A IR NOTE: ALL DIMENSIONS ARE APPROXIMATE Fig. 14.8. TSF Experimental Arrangement for Monkey Irradiation (Side View). 186 PERIOD ENDING MARCH 10, 1955 TABLE 14.1. TOTAL RADIATION DOSES TO WHICH MONKEYS WERE EXPOSED Total Neutron Dose (rep) Tota! Gamma Dose (r) Exposure NJ;T: of FTIil"’: ot Chemical Chemical Number ye vl Fower Hurst Dosimeters lon Dosimeters Exposed (min) Dosimeter _— Chamber B Cage A Cage B Cage A Cage B 1 2 1.5 31,900 28,846 28,461 4280 7763 7894 2 2 1.5 31,100 , 26,154 27,692 4190 6578 6842 3 2 1.5 31,300 23,077 24,615 4210 7105 7026 5 2 1.5 2,180 2,300 2,153 292 150 263 6 2 1.5 2,240 2,305 2,164 299 180 296 7 2 1.5 12,300 10,000 10,770 1650 2368 2237 8 2 1.5 4,500 5,153 4,615 602 o921 789 9 2 1.5 11,600 10,000 11,540 1600 1973 2000 10 2 1.5 4,400 4,386 4,769 589 723 723 11 2 1.5 11,500 10,000 13,077 1550 1842 1973 12 2 1.5 4,420 3,946 3,846 591 592 592 13 1* 1.5 2,200 2,154 2,154 295 263 263 14 2 2.0 14,700 15,381 16,153 1960 2194 2389 *Plus 1 phantom, 187 Part IV APPENDIX REPORT NO. CF-54-9-160 CF-54-11-188 CF-54-12-120 CF-54-12-209 CF-54-9-111 CF-54-10-106 CF-54-11-69 CF-54-12-154 CF:55-2-16 ORNL-1721 ORNL-1835 CF-54-9-24 CF-54-10-97 CF-54-10-119 CF-54-10-168 CF-54-11-33 CF-54-11-150 CF-54-12-21 15. TITLE OF REPORT I. Aircraft Reactor Experiment ARE Instrumentation List Preliminary Report — Operation of the Aircraft Reactor Experiment The Amount of Na22 in the Ne Coolant of the ARE After Operation at 2 Mw for Two Days Examination of the ARE II. Reflector-Moderated Reactor On Gamma Ray Heating in the Reflector-Moderated Reactor Thermal Stresses in Beryllium — Test No, 1 Temperature-Time History and Tube Stress Study of the Intermediote Heat Exchanger Test Shield Weights for the CFRE Preliminary Evaluation of Possible Poisons for Use in the ART Control Rod ORNL Aircraft Nuclear Power Plant Designs Aircraft Reactor Test Hazards Summary Report ), Experimental Engineering Program for Continuing Study of Cavitation Phenomena in Sodium The Design of @ Small Forced Circulation Corrasion Loop IV. Critical Experiments Reflector Moderated Critical Assembly Experimental Program — Part [I Experimental Program for Reflector Moderated Critical Assemblies The First Assembly of the Three-Region Reflector Moderated Reactor The Second Assembly of the Three-Region Reflector Moderated Reactor Preliminary Critical Assembly for Supercritical Water Reactor, Part I LIST OF REPORTS ISSUED FROMSEPTEMBER 1954 TO MARCH 1955 AUTHOR Ro Go Aff3| Jl L- Meem W. B. Cottrell H. W. Bertini W- Dl Monly P. H. Pitkanen R- W- BUSSOI’d R. E. MacPherson R. In Gl’oy J. B. Dee, Jr. H. C. Woodsum J. W. Noaks A. P. Fraas A. W. Savolainen W. B. Cottrell ez al, Je Mv Trummel W. K. Stair B. L. Greenstreet B. L. Greenstreet R. M. Spencer R. M. Spencer B. L. Greenstreet J. W. Noaks Jo Sc Crude|e DATE ISSUED 9-20-54 11-30-54 12-23-54 12-30-54 9-14-54 10-25-54 11-30-54 12-20-54 2-2-55 12-3-54 1-19-55 9-3-54 10-11-54 10-19-54 10-29-54 11-5-54 11-24-54 12-1-54 191 ANP PROJECT PROGRESS REPORT ‘REPORT NO. CF-55-1-123 ORNL-1770 CF-54-9-98 CF-55-1-54 CF-54-10-139 CF-54-10-140 CF-54-11-37 CF-54-11-63 CF-54-12-110 CF-55-2-20 ORNL-1769 ORNL-1777 CF-54-9-36 CF-54-12-65 CF-54-9-119 CF-54-9-120 CF-54-11-3 CF-54-11-113 192 TITLE OF REPORT Three-Region Reflector Moderated Critical Assembly R. with ]/“5 ins Inconel Core Shells Preliminary Critical Assemblies of the Reflector R. Moderated Reactar V. Metallurgy Examination of Sodium, Beryllium, Inconel Pump Loops G. Numbers 1 and 2 E. Materials Handbook L ¥I. Heat Transfer ond Physical Properties Measurement of the Thermal Conductivity of Molten Fluoride Mixture No. 44 Heat Capacity of Composition No. 40 W. G. A Laminar Forced-Convection Solution for Pipes Ducting H. Liquids Having Volume Heat Sources and Large Radial Differences in Viscosity A Method for Evaluating the Heat Transfer Effectiveness M. H. F. M. of Reactor Coolants Qualitative Velocity Information Regarding the ART Core: Status Report No. 4 Preliminary Measurements of the Viscosity of Composi- tion 20 Free-Convection in Fluids Having a Volume Heat Source AUTHOR M. Spencer M. Spencer M. Adamson Long D. Manly W. D. Powers S' Ju C|aiborne D. Powers Blalock Peoppendiek Rosenthal Poppendiek Burnett J. O. Bradfute S. 1. Cohen T. N. Jones Du Cu Hdmilfon el al- Fused Salt Heat Transfer — Part |I: Forced Convection He W. Hoffman Heat Transfer in Circular Tubes Containing NaF-KF- J. Lones LiF Eutectic V1. Radiation Damage Removal of Xenon from Fluoride Fuels: Preliminary M. Robinson Design of In-Pile Equipment Volatilization of Fission Products from Fluoride Fuels M. Robinson VIIl. Shielding The Time Variation for Injury from Radiation E. Blizard Preliminary Report on the Penetration of Composite Ce Zerby Slabs by Slant Incident Gamma Radiation Measurement of an Effective Neutron Removal Cross G. Section of Lithium at the Lid Tank Shielding Facility Fraction of Biological Dose Due to Thermal Neutrons in E. Aircraft Reactor Shields Chapman et al. Blizard DATE ISSUED 1-21-55 11-22-54 9-13-54 1-5-55 10-26-54 10-26-54 11-5-54 11-4-54 12-14-54 2-2-55 11-15-54 2-1-55 9-3-54 12-10-54 9-21-54 9-21-54 11-2-54 11-19-54 REPORT NO. ORNL-1682 CF-54-10-20 CF-54-10-48 CF-54-10-49 CF-54-10-138 ORNL-1771 ORNL-1816 TITLE OF REPORT Reactivity Measurements with the Bulk Shielding Reactor IX. Miscellaneous ANP Information Meeting of August 18, 1954 ANP Research Conference of September 28, 1954 ANP Research Conference of September 7, 1954 ANP Research Conference of Qctober 26, 1954 Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending September 10, 1954 Aircraft Nuclear Propulsion Project Quarterly Progress Report for Period Ending December 10, 1954 PERIOD ENDING MARCH 10, 1955 A. A, A. A. A. A. AUTHOR G. Cochran et al. W. Savolainen W. Savolainen W. Savolainen W. Savolainen W. Savolainen (ed.) W. Savolainen (ed.) DATE ISSUED 11-19-54 10-6-54 10-11-54 10-11-54 10-26-54 10-21-54 1-20-55 193 THE AIRCRAFT NUCLEAR PROPULSION PROJECT THE OAK RIDGE NATIONAL LABORATORY MARCH 1, 1955 ANP PROJECT DIRECTOR W. H CO.DIRECTOR 5. 4. CROMER RD ASSOCIATE DIRECTOR R. I Al ASSISTANT DIRECTOR . J. MILLER RD D. HILYER, SEC. RD PRATT & WHITHEY AIRCRAET REPORTS R. I. STROUGH, PROJECT ENGINEER PWA A. W, SAVOLAINEN ARE A, GIANGREGORID, ADM. ASS'T. PWA P. HARMAN, SEC. ARE 0. A. BRADEN, $EC. ARE LITERATURE SEARCHES H. C. GRAY, DESIGN INEER Pwa . ENGIKE A. L. DAVIS ARE ARCRAFT REACTOR ENGINE ERIMG DLYISION 5. J. CROMER, DIRECTOR RD H. MCFATRIDGE, SEC. ARE SUPPORTING RESEARCH W. H. JORDAN ASSISTANT 70 DIRECTOR A. J. MILLER R. 5. CARLSMITH ARE J. P. LANE ARE REACTOR PHYSICS ENGINEERING DESIGN REACTOR CONSTRUCTION POWER PLANT ENGINEERING EXPERIMENTAL ENGINEERING STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT STAFF ASSISTANT W. K. ERGEN ARE H, C, GRAY PWA E. 5. BETTIS ARE A. P. FRAAS ARE H. W, SAVAGE ARE W, R. GRIMES W W. D, MANLY M E. P. BLIZARD P D. S, BILLINGTON 5§ H. F. POPPENDIEK REE w, K, ERGEN ARE P. HARMAN, SEC. ARE 1. ZASLER ARE J. PARKER, SEC. ARE A. 5. THOMPSON PMC D. ALEXANDER, SEC, ARE L. E. FERGUSON, SEC. ARE P, HARMAN, SEC. ARE H. W, BERTINI ARE s, J. FOSTER, CLERK ARE PUMPS PLANNING AND CONTROL. w, E. KINNEY ARE A.G. . ADMINISTRATION LR, DY R, K. OSBORN ARE A. A, ABBATIELLO ARE . R ae wLscored ARE B B T PSON . CHEMISTRY WETALLURGY SHIELDING RADLATION DAMAGE HEAT TRAMSFER AND PHYSICAL REACTOR CALCULATIONS A, H. ANDERSON ARE W, R. GRIMES MC ¥. D, MANLY M o 5. 5. BILLIMGTON 5 PROPERTIES RESEARCH R, A. CHARPIE RO MATHEMATICIANS D, €. BORDEN PHA CORE HYDRODYNAMICS AND THERMODYHAMICS ART COMPONENT TEST FACILITIES b, R. CUNEQ MC G. M. ADAMSON® M HIELDING RESEARCH 18, TRICE 55 H. F. POPPENDIEK REE i A. FORBES ARE ¥. G, COBB ARE G. . WHITMAN ARE . T. FURGERSON ARE E.STORTO ARE D. E. CALDWELL, SEC. He #. H. BRIDGES* M E. P. BLIZARD F T. K. CARLSMITH, SEC. REE R, R, COVEYOU* ARE M, TSAGARIS ARE J, ¥, ESTABROOK ARE R. L. BREWSTER ARE M. E. LACKEY ARE J. CROWLEY ARE M. J. WHITMAN USAF R, RICKMAN, SEC. P . D. BAUMANN 5% . D' GIVEN* WP ::. ::;ms P¥A WEAT EXCHANGER G, SAMUELS ARE A. G. GRINDELL* ARE SPECIAL, PROBLEMS 1. THOMAS, SEC. M L. 5. ABBOTT " W, E. BROWNING $5 HEAT TRANSFER W. C. SANGREN* WP CONSUL TANTS > 3 PWA W, ). STELZMAN ARE R, E. MACPHERSDN* ARE R. F. NEWTON RD S, ALSLENDER PWA W. E, BRUNDAGE 88 J. 0. BRADFUTE REE L R. E. HELMS ARE R. E. MACPHERSON® ARE M. W. HINES, COMPUTER ARE L. A, MANN" ARE BENERAL CORROSION F. H. MURRAY P R, W, CARROLL. 88 B. C. HAMILTON REE A m”"“hwfim A TECH, T. L. HUDSON ARE J. €, AMOS ARE 1. 4, TUDOR, DRAFTSMAN ARE C. H, WELLS P& PHASE EQUILIBRIUM STUDIES E. E. HOFFMAN M H. E. STERN oy W.W. DAYIS 5§ H W, HOFFMAN REE A, D, JARRATT PWA L, H, DEVLIN PHA €. J. BARTON e W, H. COOK o C.D. ZERBY P M. J. FELDMAN 55 F. £, LYNCH REE s ':zERLRE erioer ;;;E R. D, PEAK PWA APPLIED MECHANICS AND PHYSICS IN-PILE LOOP TESTS L. M. BRATCHER M C. F. LEITTEN M D. E. GUSS USAF G L MULLER PHA VE A 1. W, KINGSLEY AR D. 8. TRAUGER LID TANK SHIELDING FAC JE. oL Fo Ly MAGLEY ARE € M.E. LAYERNE e ARE B. H. CLAMPITT e TECHNICIANS Ty M. E. HIBKLE 55 L.D. PALMER REE CRITICAL EXPERIMENTS et DRIVES AND AUKILIARIES 7. J. BALLES PWA S. EUBANKS, STENO. ARE R. E. MOORE WC 6. T. CHAPMAN P H. L. HEWPHILL AC 1. L. WANTLAND REE ¥ NELs'BN m: C.5. BURTNETTE USAF R. G. AFFEL ARE R. J. SHEIL WC 3. €. POPE M ). B. DEE PWA G. W, KEILHOLTZ 55 - A 0. CALLIHAN® P b P 5. M. DECAMP ARE F. A. FIELD USAF L. P. CARPENTER ARE D. L. STOCKTON ME L.R. TROTTER M ¥, J. MCCOTL YA H. V. KLAUS I TECHNICIANS . L. RUEFF,* SEC, P N C. B, THOMPSON MH J. A CONLIN ARE J. M, MILLER P R, E. MAYES PwA NES REE :. :. :x PWA CONTROLS W, I. GALYON, DRAFTSMAN ARE %, CUNRINGHAM i FUEL PREPARATION RESEARCH DYNAMIC CORRCSION 1 SMOLEN PHA J. G, MORGAN 55 w CLer REE J. 5. CRUDELE PWA s FyA J- EASTHAN . HEAT TRANSFER P, A, GNADT ARE AND PHASE BEHAVIOR G. M, ADAMSON" M D. K. TRUBEY P M, T. MORGAN 8 G, M. WINN REE 3. FLELLIS® v DRAFTSMEN J. W, ALLEN ARE D. M, HAINES PHA F. F. BLANKENSHIP We T, PRICE FWA H. €. WOCOSUM PHA A5, OLSON 5§ B. L. GREENSTREET ARE E. R. MANN 1 R. B, SCHUL THEISS ARE D, W, HAGHUSON ARE ) WC A, TABOADA " M. F. CSBORNE 3 PHYSICAL PROPERTIES ¥, G, HARNESS* P ¥, C. GEORGE ARE S, C. SHUFORD PWA M. M. YAROSH ARE o ¥. P. TRECIOKAS PWA TECHNICIANS W, W, PARKINSON 88 ‘e F. P. B0ODY MC 5. |, COHEN REE %, G. KENNEDY PWA G. R. HICKS ARE A. L. SOUTHERN ARE TECHNICIANS R E CLEARY PWa E. BECKHAM p H. E. ROBERTSON 55 GREENE REE LT v G, L. HOLLIS AR SPECIAL PROBLEMS | - E. TECHNICIANS . M T, ROBINSON - N. D. GRE: L J.LYN . L E C. J. GREEN ARE H, A, FRIEDMAN MC H. JARVIS 1c "ot W. D, POWERS REE J. H. MARABLE* P FILL AND DRAIN AND BR L H. J. R. LARRABEE PWA ENRICFMENT SYSTEM . 8. COTTRELL ARE F. M, GRIZZELL ARE 5. LANGER MC G, O, BRADY M 4. R, TAYLOR P E. 5. SCHRARTZ 55 1. W, NDAKS PWA g- G. MICHELSON ARE Ca . ARE H. E. HUNGERFORD ARE C. C. NANCE ARE R. P. METCALF MC R. M. EVANS " J. W. WAMPLER p f-é'Sfir‘-;:PLETON g: TECHNICIANS E. R. ROHRER® P - A MILLS ARE R CMR"R“! J. K, LESLIE ARE B, L. JOHNSON ARE F. W, MILES MC PHYSICAL CHEMISTRY OF LIQUID METALS BULK SHIELDING FACILITY s G.C. BLALOCK REE E. V. SANDIN PWA C, F, SALES AHE . Cu PWA SPECIAL PROBLENS I F. J. SCHAFER ARE R. E. THOMA Me C.C. WEBSTER 5 R M. BURNETT REE D, SCOTT, JR. ARE C. E. STEVENSON ARE G. P. SMITH " F. C. MAIENSCHEIN P R. A, WEEKS 58 L 5, SNYDER PWA SPECIAL PROBLEMS 5. J. CLAIBORNE REE 3 R. W. BUSSARD ARE CORROSION STUDIES C. R. BOSTON ] G, M. ESTABRODK e W R.WILLIS 5 REE R. M, SPENCER USAF W. €. TUNNELL ARE R. 1. GRAY PWA PROCUREMENT E. KERTESZ MC J, ¥, CATHCART M 1B, FLYNN & 1. C, WILSON $5 T. . JONES D. V. P. WILLIAMS* P H, €. HOPKINS PWA W. F. BOUDREAU® K.25 " K. 1. BUTTRAM MC G. F. PETERSEN M WP HAYDON® P +. C. ZUKAS S E. L. ZIMMERMAN® P R, CLARKE® K-25 F. A. KNOX MC M. £. STEIDLITZ " K. 4. HENRY P TECHNICIANS s ART FACILITY CONSUL TANTS 4. R, EVANS ARE R. E. MEADOWS M TECHNICIANS E. B, JOHMSON P A. H. FOX, UMION COLLEGE W, R. O$BORN ARE N, V. SMITH MC K. M. JOHNSON PWA J. R. DUCKWORTH ARE ¥. G. PIPER* K25 W. LOWEN, UMION COLLEGE R. . ANDERSON I W, ZOBEL P C.ELLIS $5 R. L. MAXWELL, CLERKS TECHN!CIAN L.l HALL M ' ¢. BOUNDS, SEC. P T, £, CRABTREE ARE UNIVERSITY OF TENNESSEE G. C, JENKINS ARE J. M. DIDLAKE MC WECHANICAL PROPERTIES ’ . CONSULTANT ¥. J. KELLEGHAN ARE C. F. WISLICENUS, F. LEWIS ARE TECHNICIANS D. F. WEEKES, TEXAS A & M COLLEGE £. R. MCQUILKIN ARE PENNSYLVANIA STATE UNIVERSITY E. 8. PERRIN ARE PRODUCTIGN OF PURIFIED FUELS R.B. DLIVER M D. J. KIRBY P G. C. ROBINSON ARE D. STOREY ARE 2. H. DEVAN " 1. SELLERS c G. J. NESSLE We D. A. DOUGLAS M - SELL. ! C. F. VEST ARE 1. F. BLAKELY MC 1. ¥, WOODS W R, M, SIMMONS P TEST SERYICES C. R. CROFT e . S, SMIDDIE Ic W. B. MACDONALD ARE F. L. DALEY W TECHNICIANS G, G. 5TouT Ic D. HARRIS, STENO. ARE 1. €. EDRGAN Mec J.T. EAST W H. WEAVER P FUEL REPROCESSING A. MONTGOMERY, CL. TYP. ARE 3 TRUITT HC G. M. GONZALEZ M — . N C. P COUGHLEN ARE AN S O HUDSON . TOWER SHIELDING FACILITY F. R, BRUCE cT R. A. DREISBACK PwA V. G, LANE M €. E. CLIFFORD p R. HEL TON ARE F. A. DOSS ARE B. MCNABE W .. BLOSSER P CHEMICAL DEVELOPMENT E. M. LEES ARE B, F., HITCH MC C. K. THOMAS M L, B, HOLLAND B D, E. FERGUSON* T J. 1. MILICH PWA W, JENNINGS MC C, W, WALKER M J L. HULL P G. I. CATHERS T W, J. MUENZER ARE G, A, PALMER MC w. F. VALERING NACA M. R. BENNETT cT ¢ E. MURPHY ARE B. C. THOMAS MC FABRICATION J. VAN HOOMISSEN BAC TECHNICIAN P, G. SMITH ARE R. G. WILEY ARE 1. H, CODBS " F. N, WATSON P CHRIC L DY . M. DUFF cr D. R. WARD ARE CHEMICAL EQUILIBRIA n :?NODU¢E‘DRE SWA E. MCBEE, SEC. " Rt TECHNICIANS L. G. OVERHOLSER WC TECHNICIANS UNIT OPERATIONS NOTE: THIS CHART SHOWS ONLY THE LINES OF TECHNICAL COORDINATION OF THE ANP PROJECT. THE ), ADKINS ARE E. E. KETCHEM MC TECHNICEAN E. D. CARROLL IC W, K, EISTER* cT VARIOUS INDIVIDUALS AND GROUPS OF PEOPLE LISTED ARE ENGAGED EITHER WHOLLY OR PART TIME ON .5, ADDISON ARE . g- ;;g“f“ MC R. W, JOHNSON " J. N, MONEY P N R cr RESEARCH AND DESIGH WHICH IS COORDINATED FOR THE BENEFIT OF THE ANP PROJECT IN THE MANNER R wELL e PN s WELDING AND BRAZING o o e - < JNDICATED GN THE CHART, EACH GROUP, HOWEYER, IS ALSC RESPONSIBLE TO ITS DIVISION DIRECTOR G. 5. CHILTON ARE L. E. TOPOL MC P. PATRIARCA ') - DESIGH FOR THE DETAILED PROGRESS OF ITS RESEARCH AND FOR ADMINISTRATIVE MATTERS, 1. M. COBURN ARE C. F. WEAVER MC :- E fié’é??k':fo *;w A CONSUL TANT H. E. GOELLER" cT J. R. CROLEY ARE L L. R. P. MILFORD cT AL HE, R ERSIT THE KEY TO THE ABBREVIATIONS USED IS GIVEN BELOW. 7. M. CUNNINGHAM ARE SUPPORTING STUDIES K. W, REBER W H. A, BETHE, CORNELL UNIVERSITY R.E. DIAL ARE P. A. AGRON c G. M. SLAUGHTER " CONTRACTORS AC ANALYTICAL CHEMISTRY DIVISION — ORNL W. H. DUCKWORTH ARE J- E. SUTHERLAND < TECHNICIANS YORIDES, INC ALC AMERICAN LOCOMOTIVE COMPANY 1. P. EUBANKS ARE Tt DoV bhe) W. K. R. FINNELL ARE CONSULTANTS 0. E. CONNOR M NUCLEAR DEVELOPMENT ARE AIRCRAFT REACTOR ENGINEERING DIVISION = ORNL H. FOUST ARE ). 4. CARTER, CARTER LABORATORIES B. MCDOWEL L M ASSOCIATES, INC. BAC BOEING AIRPLANE COMPANY W. D. GHORMLEY ARE B G FILL. DUKE UNIVERSITY g. €., SHUBERT M P BENDIX PRODUCTS, DIVISION OF BENDIX AVIATION CORPORATION il o H. INSLEY - . SHOOSTER M c CHEMISTRY DIVISION - ORNL C: G: HENLEY ARE T. N. MCYAY, UNIVERSITY OF ALABAMA COMSULTANTS AHP STEERING COMMITTEE cr CHEMICAL TECHNOLOGY DIVISION — ORNL D. E, MCCARTY ARE N. CABRERA, UNIVERSITY OF VIRGINIA cv CONSOLIDATED YULTEE AIRCRAFT CORPORATION G.E. MILLS ARE CoMTRACTORS M. J. GRANT, MASSACHUSETTS METALLOGRAPHY W. H. JORDAN, CHAIRMAN B. H. MONTGOMERY ARE AMES LABORATORY INSTITUTE OF TECHNOLOGY 2. ). GRAY™ " e : GE GENERAL ELECTRIC COMPANY 1. ). PARSONS ARE BATTELLE MEMORIAL INSTITUTE J, L, GREGG, CORNELL UMIVERSITY - E. 5, BETTIS Ic INSTRUMENTATION AND CONTROLS DIVISION = ORNL M. A. REDDEN ARE CARTER LABORATOR(ES W, D, JORDAN, UNIVERSITY OF ALABAMA R, 5. CROLSE M D. 5. BHLLINGTON " METALLURGY DIVISION ORNL ®. REID ARE METAL HYDRIDES, INC. E. F, NIPPES$, RENSSELAER T M KEGLEY M E. P, 8LIZARD A G. TOWNS ARE MOUND LABORATORY POLYTECHMIC INSTITUTE E L LONG M 80D e MATERIALS CHEMISTRY DIVISION - QRNL B. C. WILLIAMS ARE A. R, NICHOLS, W. F. SAVAGE, RENSSELAER £ 3 MANTHOS W §. J. CROMER ME MERCK & COMPANY, INC, CRAFT COORDINATION SAN DIEGO STATE COLLEGE POLYTECHNIC INSTITUTE s :. l;- ::fii: . SISTARE, HANDY & HARMON TECHNICIANS - o Mid WINNEAPOL [$-HONE YWELL REGULATOR COMPANY . L. MATTHEWS Y12 P. €. SHARRAH, UNIVERSITY DF ARKANSAS €. R BOYD " . R. GRIMES WP MATHEMATICS PANEL ~ ORNL D. L. CLARK ARE F. G, TATNALL, B, F, DAY M G e HACA NATIDNAL ADVISORY COMM TTEE FOR AERONAUTICS D. L. GRAY ARE cb Sug:::;f"“" CHEMISTRY AC BALDWIN-LIMA-HAMIL TON CORPORATION R. L. FITZGERALD M : ?’ ,::?_T'E: P ARYSICS DIVISION — ORNL || O N e s A E. C. WRIGHT, UNIVERSITY OF ALABAMA B C. LESLIE M b b R RENDIEK PG PACKARD MOTOR CAR COMPANY CONSULTANTS J.C. WHITE AC CONTRACTORS H. W. SAVAGE PWA PRATT & WHITNEY AIRCRAFT, DIVISION OF LUAC J. F. BAILEY, UNIVERSITY OF TENNESSEE E. M. ZARZECKI,* SEC AC BATTELLE MEWORIAL INSTITUTE :' ?.Ssr;’gbg RD RESEARCH DIRECTOR'S DEPARTMENT ~ ORNL :’ :_‘- S:M"‘E‘:ERS RESCARCH 22;’%?:2,:;"'”” COMPANY CERAMIC RESEARCH 2. A, SHARTOUT REE REACTOR EXPERIMENTAL ENGINEERING DIVISION — ORNL ¥. K. STAIR, UNIVERSITY OF TENNESSEE A 5. MEYER, JR, AC GLENN L. MARTIN COMPANY D s+ SEC : A. M, REINBERS 8 STABLE ISGTOPE RESEARCH AND PRODUCTION DIVISION =~ ORNL D, L, MANNING AC METAL HYgRID:i'Tlng P : HOBBS, - B, L, McDOWELL AC NEW ENGLAND 5 $OLID STATE DIVISION — ORNL 1M PEELE AC TESTING LABORATORY €. E. CURTIS: M USAF UNITED STATES AIR FORCE W, J. ROSS AC RENSSELAER POLYTECHNIC INSTITUTE L. M. DONEY W s N 5. D, FULKERSOR® M “PART TIME DEVELOPMENT UPERIOR TUBE COMPANY 5. . FULKERSONY M 1 P YOUNG ac UNIVERSITY OF TENNESSEE Lo AGH A G. GOLDBERG AC M, P, HAYDON® P SPECTROGRAPHIC ANALYSIS M. A MARLER AC J. R, JOHNSON* » LR eRALLY + H SERVICE &, J. TAYLOR* " - Rohcl ) W, F. VAUGHAN ac G. D. WHITE® » J. A, NORRIS* AND OTHERS 5t €. R WILLIAMS ac TECHNICIAN R. F. APPLE AC . " MASS SPECTROMETRY D. E. CARPENTER AC . A4 GRIFFIN C. R, BALDOCK" st TECHNICIANS CONSULTANTS 1. R.SITES* si R, C. BRYANT AC ™. T. CORY, GRAPHITE SPECIALTIES L. R, HALL ac CORPORATION L. E. IDOM AC H. INSLEY A, H. MATTHEWS AC T. N, MCVAY, UNIVERSITY OF ALABAMA C. E. PRATHER AC T, 5. SHEYL IN, OHIO STATE UMIVERSITY A. D. WILSON AC H. THURNAUER, AMERICAN LAYA C. M. WILSON AC CORPORATION e ~7 sm= the Atomic